From 3871f2ffe53db3cef4fe0c18993ad9e6e0f69408 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Wed, 24 Dec 2008 16:06:57 -0800 Subject: [PATCH 001/413] sysrq: fix ftrace help msg & doc. Impact: update documentation and help messages We have a conventional method of explicitly stating the sysrq action key in a sysrq help message, so change dump-ftrace-buffer to use that method and add it to Documentation/sysrq.txt. Signed-off-by: Randy Dunlap Signed-off-by: Ingo Molnar --- Documentation/sysrq.txt | 2 ++ drivers/char/sysrq.c | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-) diff --git a/Documentation/sysrq.txt b/Documentation/sysrq.txt index 10a0263ebb3f..56b53e005d18 100644 --- a/Documentation/sysrq.txt +++ b/Documentation/sysrq.txt @@ -114,6 +114,8 @@ On all - write a character to /proc/sysrq-trigger. e.g.: 'x' - Used by xmon interface on ppc/powerpc platforms. +'z' - Dump the ftrace buffer + '0'-'9' - Sets the console log level, controlling which kernel messages will be printed to your console. ('0', for example would make it so that only emergency messages like PANICs or OOPSes would diff --git a/drivers/char/sysrq.c b/drivers/char/sysrq.c index 94966edfb44d..785a08ef9a13 100644 --- a/drivers/char/sysrq.c +++ b/drivers/char/sysrq.c @@ -283,7 +283,7 @@ static void sysrq_ftrace_dump(int key, struct tty_struct *tty) } static struct sysrq_key_op sysrq_ftrace_dump_op = { .handler = sysrq_ftrace_dump, - .help_msg = "dumpZ-ftrace-buffer", + .help_msg = "dump-ftrace-buffer(Z)", .action_msg = "Dump ftrace buffer", .enable_mask = SYSRQ_ENABLE_DUMP, }; From 412d0bb553c0227191f1bfd06100f561600bff22 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 24 Dec 2008 01:43:25 +0100 Subject: [PATCH 002/413] tracing/function-graph-tracer: strip ending newlines on comments Impact: tracer output improvement Ending newlines are appended automatically on comments by the function graph tracer because the newline needs to be placed after the "*/" comment characters. So if the user puts an ending newline, we want to strip it. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions_graph.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 4bf39fcae97a..bc7d90850be5 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -592,6 +592,12 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, if (ent->flags & TRACE_FLAG_CONT) trace_seq_print_cont(s, iter); + /* Strip ending newline */ + if (s->buffer[s->len - 1] == '\n') { + s->buffer[s->len - 1] = '\0'; + s->len--; + } + ret = trace_seq_printf(s, " */\n"); if (!ret) return TRACE_TYPE_PARTIAL_LINE; From c47956d9ae3341d2d1998bff26620fa3338c01e4 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 23 Dec 2008 23:24:11 -0500 Subject: [PATCH 003/413] ftrace: remove obsolete print continue functionality Impact: cleanup, remove obsolete code Now that the ring buffer used by ftrace allows for variable length entries, we do not need the 'cont' feature of the buffer. This code makes other parts of ftrace more complex and by removing this it simplifies the ftrace code. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 61 ---------------------------- kernel/trace/trace.h | 7 ---- kernel/trace/trace_functions_graph.c | 3 -- kernel/trace/trace_mmiotrace.c | 3 -- 4 files changed, 74 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index f4bb3800318b..fca0233f1d73 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1765,43 +1765,6 @@ static int task_state_char(unsigned long state) return bit < sizeof(state_to_char) - 1 ? state_to_char[bit] : '?'; } -/* - * The message is supposed to contain an ending newline. - * If the printing stops prematurely, try to add a newline of our own. - */ -void trace_seq_print_cont(struct trace_seq *s, struct trace_iterator *iter) -{ - struct trace_entry *ent; - struct trace_field_cont *cont; - bool ok = true; - - ent = peek_next_entry(iter, iter->cpu, NULL); - if (!ent || ent->type != TRACE_CONT) { - trace_seq_putc(s, '\n'); - return; - } - - do { - cont = (struct trace_field_cont *)ent; - if (ok) - ok = (trace_seq_printf(s, "%s", cont->buf) > 0); - - ftrace_disable_cpu(); - - if (iter->buffer_iter[iter->cpu]) - ring_buffer_read(iter->buffer_iter[iter->cpu], NULL); - else - ring_buffer_consume(iter->tr->buffer, iter->cpu, NULL); - - ftrace_enable_cpu(); - - ent = peek_next_entry(iter, iter->cpu, NULL); - } while (ent && ent->type == TRACE_CONT); - - if (!ok) - trace_seq_putc(s, '\n'); -} - static void test_cpu_buff_start(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; @@ -1834,9 +1797,6 @@ print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu) int S, T; int i; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - test_cpu_buff_start(iter); next_entry = find_next_entry(iter, NULL, &next_ts); @@ -1922,8 +1882,6 @@ print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu) seq_print_ip_sym(s, field->ip, sym_flags); trace_seq_printf(s, ": %s", field->buf); - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); break; } case TRACE_BRANCH: { @@ -1968,9 +1926,6 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - test_cpu_buff_start(iter); comm = trace_find_cmdline(iter->ent->pid); @@ -2076,8 +2031,6 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) seq_print_ip_sym(s, field->ip, sym_flags); trace_seq_printf(s, ": %s", field->buf); - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); break; } case TRACE_GRAPH_RET: { @@ -2124,9 +2077,6 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - ret = trace_seq_printf(s, "%d %d %llu ", entry->pid, iter->cpu, iter->ts); if (!ret) @@ -2187,8 +2137,6 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) trace_assign_type(field, entry); trace_seq_printf(s, "# %lx %s", field->ip, field->buf); - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); break; } } @@ -2217,9 +2165,6 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - SEQ_PUT_HEX_FIELD_RET(s, entry->pid); SEQ_PUT_HEX_FIELD_RET(s, iter->cpu); SEQ_PUT_HEX_FIELD_RET(s, iter->ts); @@ -2283,9 +2228,6 @@ static enum print_line_t print_printk_msg_only(struct trace_iterator *iter) if (!ret) return TRACE_TYPE_PARTIAL_LINE; - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); - return TRACE_TYPE_HANDLED; } @@ -2296,9 +2238,6 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) entry = iter->ent; - if (entry->type == TRACE_CONT) - return TRACE_TYPE_HANDLED; - SEQ_PUT_FIELD_RET(s, entry->pid); SEQ_PUT_FIELD_RET(s, entry->cpu); SEQ_PUT_FIELD_RET(s, iter->ts); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index cc7a4f864036..3a357382cce6 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -16,7 +16,6 @@ enum trace_type { TRACE_FN, TRACE_CTX, TRACE_WAKE, - TRACE_CONT, TRACE_STACK, TRACE_PRINT, TRACE_SPECIAL, @@ -178,7 +177,6 @@ struct trace_power { * NEED_RESCED - reschedule is requested * HARDIRQ - inside an interrupt handler * SOFTIRQ - inside a softirq handler - * CONT - multiple entries hold the trace item */ enum trace_flag_type { TRACE_FLAG_IRQS_OFF = 0x01, @@ -186,7 +184,6 @@ enum trace_flag_type { TRACE_FLAG_NEED_RESCHED = 0x04, TRACE_FLAG_HARDIRQ = 0x08, TRACE_FLAG_SOFTIRQ = 0x10, - TRACE_FLAG_CONT = 0x20, }; #define TRACE_BUF_SIZE 1024 @@ -262,7 +259,6 @@ extern void __ftrace_bad_type(void); do { \ IF_ASSIGN(var, ent, struct ftrace_entry, TRACE_FN); \ IF_ASSIGN(var, ent, struct ctx_switch_entry, 0); \ - IF_ASSIGN(var, ent, struct trace_field_cont, TRACE_CONT); \ IF_ASSIGN(var, ent, struct stack_entry, TRACE_STACK); \ IF_ASSIGN(var, ent, struct userstack_entry, TRACE_USER_STACK);\ IF_ASSIGN(var, ent, struct print_entry, TRACE_PRINT); \ @@ -489,9 +485,6 @@ extern int trace_selftest_startup_branch(struct tracer *trace, extern void *head_page(struct trace_array_cpu *data); extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...); -extern void trace_seq_print_cont(struct trace_seq *s, - struct trace_iterator *iter); - extern int seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags); diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index bc7d90850be5..f261966e5b6c 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -589,9 +589,6 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, if (!ret) return TRACE_TYPE_PARTIAL_LINE; - if (ent->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); - /* Strip ending newline */ if (s->buffer[s->len - 1] == '\n') { s->buffer[s->len - 1] = '\0'; diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index fffcb069f1dc..83f20ae6bd68 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -262,9 +262,6 @@ static enum print_line_t mmio_print_mark(struct trace_iterator *iter) if (!ret) return TRACE_TYPE_PARTIAL_LINE; - if (entry->flags & TRACE_FLAG_CONT) - trace_seq_print_cont(s, iter); - return TRACE_TYPE_HANDLED; } From f0868d1e23a8efec33beb3aa688aab7fdb1ae093 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 23 Dec 2008 23:24:12 -0500 Subject: [PATCH 004/413] ftrace: set up trace event hash infrastructure Impact: simplify/generalize/refactor trace.c The trace.c file is becoming more difficult to maintain due to the growing number of events. There is several formats that an event may be printed. This patch sets up the infrastructure of an event hash to allow for events to register how they should be printed. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/Makefile | 1 + kernel/trace/trace.c | 275 +------------------- kernel/trace/trace.h | 8 +- kernel/trace/trace_boot.c | 1 + kernel/trace/trace_functions_graph.c | 1 + kernel/trace/trace_hw_branches.c | 1 + kernel/trace/trace_mmiotrace.c | 1 + kernel/trace/trace_output.c | 365 +++++++++++++++++++++++++++ kernel/trace/trace_output.h | 43 ++++ kernel/trace/trace_power.c | 1 + 10 files changed, 416 insertions(+), 281 deletions(-) create mode 100644 kernel/trace/trace_output.c create mode 100644 kernel/trace/trace_output.h diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 349d5a93653f..549f93c9b393 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -19,6 +19,7 @@ obj-$(CONFIG_FUNCTION_TRACER) += libftrace.o obj-$(CONFIG_RING_BUFFER) += ring_buffer.o obj-$(CONFIG_TRACING) += trace.o +obj-$(CONFIG_TRACING) += trace_output.o obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o obj-$(CONFIG_SYSPROF_TRACER) += trace_sysprof.o obj-$(CONFIG_FUNCTION_TRACER) += trace_functions.o diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index fca0233f1d73..90ce0c1d437b 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -38,6 +38,7 @@ #include #include "trace.h" +#include "trace_output.h" #define TRACE_BUFFER_FLAGS (RB_FL_OVERWRITE) @@ -330,132 +331,6 @@ __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) tracing_record_cmdline(current); } -/** - * trace_seq_printf - sequence printing of trace information - * @s: trace sequence descriptor - * @fmt: printf format string - * - * The tracer may use either sequence operations or its own - * copy to user routines. To simplify formating of a trace - * trace_seq_printf is used to store strings into a special - * buffer (@s). Then the output may be either used by - * the sequencer or pulled into another buffer. - */ -int -trace_seq_printf(struct trace_seq *s, const char *fmt, ...) -{ - int len = (PAGE_SIZE - 1) - s->len; - va_list ap; - int ret; - - if (!len) - return 0; - - va_start(ap, fmt); - ret = vsnprintf(s->buffer + s->len, len, fmt, ap); - va_end(ap); - - /* If we can't write it all, don't bother writing anything */ - if (ret >= len) - return 0; - - s->len += ret; - - return len; -} - -/** - * trace_seq_puts - trace sequence printing of simple string - * @s: trace sequence descriptor - * @str: simple string to record - * - * The tracer may use either the sequence operations or its own - * copy to user routines. This function records a simple string - * into a special buffer (@s) for later retrieval by a sequencer - * or other mechanism. - */ -static int -trace_seq_puts(struct trace_seq *s, const char *str) -{ - int len = strlen(str); - - if (len > ((PAGE_SIZE - 1) - s->len)) - return 0; - - memcpy(s->buffer + s->len, str, len); - s->len += len; - - return len; -} - -static int -trace_seq_putc(struct trace_seq *s, unsigned char c) -{ - if (s->len >= (PAGE_SIZE - 1)) - return 0; - - s->buffer[s->len++] = c; - - return 1; -} - -static int -trace_seq_putmem(struct trace_seq *s, void *mem, size_t len) -{ - if (len > ((PAGE_SIZE - 1) - s->len)) - return 0; - - memcpy(s->buffer + s->len, mem, len); - s->len += len; - - return len; -} - -#define MAX_MEMHEX_BYTES 8 -#define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) - -static int -trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len) -{ - unsigned char hex[HEX_CHARS]; - unsigned char *data = mem; - int i, j; - -#ifdef __BIG_ENDIAN - for (i = 0, j = 0; i < len; i++) { -#else - for (i = len-1, j = 0; i >= 0; i--) { -#endif - hex[j++] = hex_asc_hi(data[i]); - hex[j++] = hex_asc_lo(data[i]); - } - hex[j++] = ' '; - - return trace_seq_putmem(s, hex, j); -} - -static int -trace_seq_path(struct trace_seq *s, struct path *path) -{ - unsigned char *p; - - if (s->len >= (PAGE_SIZE - 1)) - return 0; - p = d_path(path, s->buffer + s->len, PAGE_SIZE - s->len); - if (!IS_ERR(p)) { - p = mangle_path(s->buffer + s->len, p, "\n"); - if (p) { - s->len = p - s->buffer; - return 1; - } - } else { - s->buffer[s->len++] = '?'; - return 1; - } - - return 0; -} - static void trace_seq_reset(struct trace_seq *s) { @@ -1473,154 +1348,6 @@ static void s_stop(struct seq_file *m, void *p) mutex_unlock(&trace_types_lock); } -#ifdef CONFIG_KRETPROBES -static inline const char *kretprobed(const char *name) -{ - static const char tramp_name[] = "kretprobe_trampoline"; - int size = sizeof(tramp_name); - - if (strncmp(tramp_name, name, size) == 0) - return "[unknown/kretprobe'd]"; - return name; -} -#else -static inline const char *kretprobed(const char *name) -{ - return name; -} -#endif /* CONFIG_KRETPROBES */ - -static int -seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address) -{ -#ifdef CONFIG_KALLSYMS - char str[KSYM_SYMBOL_LEN]; - const char *name; - - kallsyms_lookup(address, NULL, NULL, NULL, str); - - name = kretprobed(str); - - return trace_seq_printf(s, fmt, name); -#endif - return 1; -} - -static int -seq_print_sym_offset(struct trace_seq *s, const char *fmt, - unsigned long address) -{ -#ifdef CONFIG_KALLSYMS - char str[KSYM_SYMBOL_LEN]; - const char *name; - - sprint_symbol(str, address); - name = kretprobed(str); - - return trace_seq_printf(s, fmt, name); -#endif - return 1; -} - -#ifndef CONFIG_64BIT -# define IP_FMT "%08lx" -#else -# define IP_FMT "%016lx" -#endif - -int -seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) -{ - int ret; - - if (!ip) - return trace_seq_printf(s, "0"); - - if (sym_flags & TRACE_ITER_SYM_OFFSET) - ret = seq_print_sym_offset(s, "%s", ip); - else - ret = seq_print_sym_short(s, "%s", ip); - - if (!ret) - return 0; - - if (sym_flags & TRACE_ITER_SYM_ADDR) - ret = trace_seq_printf(s, " <" IP_FMT ">", ip); - return ret; -} - -static inline int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, - unsigned long ip, unsigned long sym_flags) -{ - struct file *file = NULL; - unsigned long vmstart = 0; - int ret = 1; - - if (mm) { - const struct vm_area_struct *vma; - - down_read(&mm->mmap_sem); - vma = find_vma(mm, ip); - if (vma) { - file = vma->vm_file; - vmstart = vma->vm_start; - } - if (file) { - ret = trace_seq_path(s, &file->f_path); - if (ret) - ret = trace_seq_printf(s, "[+0x%lx]", ip - vmstart); - } - up_read(&mm->mmap_sem); - } - if (ret && ((sym_flags & TRACE_ITER_SYM_ADDR) || !file)) - ret = trace_seq_printf(s, " <" IP_FMT ">", ip); - return ret; -} - -static int -seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, - unsigned long sym_flags) -{ - struct mm_struct *mm = NULL; - int ret = 1; - unsigned int i; - - if (trace_flags & TRACE_ITER_SYM_USEROBJ) { - struct task_struct *task; - /* - * we do the lookup on the thread group leader, - * since individual threads might have already quit! - */ - rcu_read_lock(); - task = find_task_by_vpid(entry->ent.tgid); - if (task) - mm = get_task_mm(task); - rcu_read_unlock(); - } - - for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { - unsigned long ip = entry->caller[i]; - - if (ip == ULONG_MAX || !ret) - break; - if (i && ret) - ret = trace_seq_puts(s, " <- "); - if (!ip) { - if (ret) - ret = trace_seq_puts(s, "??"); - continue; - } - if (!ret) - break; - if (ret) - ret = seq_print_user_ip(s, mm, ip, sym_flags); - } - - if (mm) - mmput(mm); - return ret; -} - static void print_lat_help_header(struct seq_file *m) { seq_puts(m, "# _------=> CPU# \n"); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 3a357382cce6..6bd71fa1e1c7 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -30,7 +30,7 @@ enum trace_type { TRACE_HW_BRANCHES, TRACE_POWER, - __TRACE_LAST_TYPE + __TRACE_LAST_TYPE, }; /* @@ -484,12 +484,6 @@ extern int trace_selftest_startup_branch(struct tracer *trace, #endif /* CONFIG_FTRACE_STARTUP_TEST */ extern void *head_page(struct trace_array_cpu *data); -extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...); -extern int -seq_print_ip_sym(struct trace_seq *s, unsigned long ip, - unsigned long sym_flags); -extern ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, - size_t cnt); extern long ns2usecs(cycle_t nsec); extern int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args); diff --git a/kernel/trace/trace_boot.c b/kernel/trace/trace_boot.c index 3ccebde28482..cb2ff3e297b1 100644 --- a/kernel/trace/trace_boot.c +++ b/kernel/trace/trace_boot.c @@ -11,6 +11,7 @@ #include #include "trace.h" +#include "trace_output.h" static struct trace_array *boot_trace; static bool pre_initcalls_finished; diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index f261966e5b6c..f8ac5417afc8 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -12,6 +12,7 @@ #include #include "trace.h" +#include "trace_output.h" #define TRACE_GRAPH_INDENT 2 diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index b6a3e20a49a9..879752b006b3 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -14,6 +14,7 @@ #include #include "trace.h" +#include "trace_output.h" #define SIZEOF_BTS (1 << 13) diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index 83f20ae6bd68..fcec59ff2355 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -11,6 +11,7 @@ #include #include "trace.h" +#include "trace_output.h" struct header_iter { struct pci_dev *dev; diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c new file mode 100644 index 000000000000..1f3f80002b5e --- /dev/null +++ b/kernel/trace/trace_output.c @@ -0,0 +1,365 @@ +/* + * trace_output.c + * + * Copyright (C) 2008 Red Hat Inc, Steven Rostedt + * + */ + +#include +#include +#include + +#include "trace_output.h" + +/* must be a power of 2 */ +#define EVENT_HASHSIZE 128 + +static DEFINE_MUTEX(trace_event_mutex); +static struct hlist_head event_hash[EVENT_HASHSIZE] __read_mostly; + +static int next_event_type = __TRACE_LAST_TYPE + 1; + +/** + * trace_seq_printf - sequence printing of trace information + * @s: trace sequence descriptor + * @fmt: printf format string + * + * The tracer may use either sequence operations or its own + * copy to user routines. To simplify formating of a trace + * trace_seq_printf is used to store strings into a special + * buffer (@s). Then the output may be either used by + * the sequencer or pulled into another buffer. + */ +int +trace_seq_printf(struct trace_seq *s, const char *fmt, ...) +{ + int len = (PAGE_SIZE - 1) - s->len; + va_list ap; + int ret; + + if (!len) + return 0; + + va_start(ap, fmt); + ret = vsnprintf(s->buffer + s->len, len, fmt, ap); + va_end(ap); + + /* If we can't write it all, don't bother writing anything */ + if (ret >= len) + return 0; + + s->len += ret; + + return len; +} + +/** + * trace_seq_puts - trace sequence printing of simple string + * @s: trace sequence descriptor + * @str: simple string to record + * + * The tracer may use either the sequence operations or its own + * copy to user routines. This function records a simple string + * into a special buffer (@s) for later retrieval by a sequencer + * or other mechanism. + */ +int trace_seq_puts(struct trace_seq *s, const char *str) +{ + int len = strlen(str); + + if (len > ((PAGE_SIZE - 1) - s->len)) + return 0; + + memcpy(s->buffer + s->len, str, len); + s->len += len; + + return len; +} + +int trace_seq_putc(struct trace_seq *s, unsigned char c) +{ + if (s->len >= (PAGE_SIZE - 1)) + return 0; + + s->buffer[s->len++] = c; + + return 1; +} + +int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len) +{ + if (len > ((PAGE_SIZE - 1) - s->len)) + return 0; + + memcpy(s->buffer + s->len, mem, len); + s->len += len; + + return len; +} + +int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len) +{ + unsigned char hex[HEX_CHARS]; + unsigned char *data = mem; + int i, j; + +#ifdef __BIG_ENDIAN + for (i = 0, j = 0; i < len; i++) { +#else + for (i = len-1, j = 0; i >= 0; i--) { +#endif + hex[j++] = hex_asc_hi(data[i]); + hex[j++] = hex_asc_lo(data[i]); + } + hex[j++] = ' '; + + return trace_seq_putmem(s, hex, j); +} + +int trace_seq_path(struct trace_seq *s, struct path *path) +{ + unsigned char *p; + + if (s->len >= (PAGE_SIZE - 1)) + return 0; + p = d_path(path, s->buffer + s->len, PAGE_SIZE - s->len); + if (!IS_ERR(p)) { + p = mangle_path(s->buffer + s->len, p, "\n"); + if (p) { + s->len = p - s->buffer; + return 1; + } + } else { + s->buffer[s->len++] = '?'; + return 1; + } + + return 0; +} + +#ifdef CONFIG_KRETPROBES +static inline const char *kretprobed(const char *name) +{ + static const char tramp_name[] = "kretprobe_trampoline"; + int size = sizeof(tramp_name); + + if (strncmp(tramp_name, name, size) == 0) + return "[unknown/kretprobe'd]"; + return name; +} +#else +static inline const char *kretprobed(const char *name) +{ + return name; +} +#endif /* CONFIG_KRETPROBES */ + +static int +seq_print_sym_short(struct trace_seq *s, const char *fmt, unsigned long address) +{ +#ifdef CONFIG_KALLSYMS + char str[KSYM_SYMBOL_LEN]; + const char *name; + + kallsyms_lookup(address, NULL, NULL, NULL, str); + + name = kretprobed(str); + + return trace_seq_printf(s, fmt, name); +#endif + return 1; +} + +static int +seq_print_sym_offset(struct trace_seq *s, const char *fmt, + unsigned long address) +{ +#ifdef CONFIG_KALLSYMS + char str[KSYM_SYMBOL_LEN]; + const char *name; + + sprint_symbol(str, address); + name = kretprobed(str); + + return trace_seq_printf(s, fmt, name); +#endif + return 1; +} + +#ifndef CONFIG_64BIT +# define IP_FMT "%08lx" +#else +# define IP_FMT "%016lx" +#endif + +int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, + unsigned long ip, unsigned long sym_flags) +{ + struct file *file = NULL; + unsigned long vmstart = 0; + int ret = 1; + + if (mm) { + const struct vm_area_struct *vma; + + down_read(&mm->mmap_sem); + vma = find_vma(mm, ip); + if (vma) { + file = vma->vm_file; + vmstart = vma->vm_start; + } + if (file) { + ret = trace_seq_path(s, &file->f_path); + if (ret) + ret = trace_seq_printf(s, "[+0x%lx]", + ip - vmstart); + } + up_read(&mm->mmap_sem); + } + if (ret && ((sym_flags & TRACE_ITER_SYM_ADDR) || !file)) + ret = trace_seq_printf(s, " <" IP_FMT ">", ip); + return ret; +} + +int +seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, + unsigned long sym_flags) +{ + struct mm_struct *mm = NULL; + int ret = 1; + unsigned int i; + + if (trace_flags & TRACE_ITER_SYM_USEROBJ) { + struct task_struct *task; + /* + * we do the lookup on the thread group leader, + * since individual threads might have already quit! + */ + rcu_read_lock(); + task = find_task_by_vpid(entry->ent.tgid); + if (task) + mm = get_task_mm(task); + rcu_read_unlock(); + } + + for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { + unsigned long ip = entry->caller[i]; + + if (ip == ULONG_MAX || !ret) + break; + if (i && ret) + ret = trace_seq_puts(s, " <- "); + if (!ip) { + if (ret) + ret = trace_seq_puts(s, "??"); + continue; + } + if (!ret) + break; + if (ret) + ret = seq_print_user_ip(s, mm, ip, sym_flags); + } + + if (mm) + mmput(mm); + return ret; +} + +int +seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) +{ + int ret; + + if (!ip) + return trace_seq_printf(s, "0"); + + if (sym_flags & TRACE_ITER_SYM_OFFSET) + ret = seq_print_sym_offset(s, "%s", ip); + else + ret = seq_print_sym_short(s, "%s", ip); + + if (!ret) + return 0; + + if (sym_flags & TRACE_ITER_SYM_ADDR) + ret = trace_seq_printf(s, " <" IP_FMT ">", ip); + return ret; +} + +/** + * ftrace_find_event - find a registered event + * @type: the type of event to look for + * + * Returns an event of type @type otherwise NULL + */ +struct trace_event *ftrace_find_event(int type) +{ + struct trace_event *event; + struct hlist_node *n; + unsigned key; + + key = type & (EVENT_HASHSIZE - 1); + + hlist_for_each_entry_rcu(event, n, &event_hash[key], node) { + if (event->type == type) + return event; + } + + return NULL; +} + +/** + * register_ftrace_event - register output for an event type + * @event: the event type to register + * + * Event types are stored in a hash and this hash is used to + * find a way to print an event. If the @event->type is set + * then it will use that type, otherwise it will assign a + * type to use. + * + * If you assign your own type, please make sure it is added + * to the trace_type enum in trace.h, to avoid collisions + * with the dynamic types. + * + * Returns the event type number or zero on error. + */ +int register_ftrace_event(struct trace_event *event) +{ + unsigned key; + int ret = 0; + + mutex_lock(&trace_event_mutex); + + if (!event->type) + event->type = next_event_type++; + else if (event->type > __TRACE_LAST_TYPE) { + printk(KERN_WARNING "Need to add type to trace.h\n"); + WARN_ON(1); + } + + if (ftrace_find_event(event->type)) + goto out; + + key = event->type & (EVENT_HASHSIZE - 1); + + hlist_add_head_rcu(&event->node, &event_hash[key]); + + ret = event->type; + out: + mutex_unlock(&trace_event_mutex); + + return ret; +} + +/** + * unregister_ftrace_event - remove a no longer used event + * @event: the event to remove + */ +int unregister_ftrace_event(struct trace_event *event) +{ + mutex_lock(&trace_event_mutex); + hlist_del(&event->node); + mutex_unlock(&trace_event_mutex); + + return 0; +} diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h new file mode 100644 index 000000000000..1fcc76e1378e --- /dev/null +++ b/kernel/trace/trace_output.h @@ -0,0 +1,43 @@ +#ifndef __TRACE_EVENTS_H +#define __TRACE_EVENTS_H + +#include "trace.h" + +typedef int (*trace_print_func)(struct trace_seq *s, struct trace_entry *entry, + int flags); + +struct trace_event { + struct hlist_node node; + int type; + trace_print_func trace; + trace_print_func latency_trace; + trace_print_func raw; + trace_print_func hex; + trace_print_func binary; +}; + +extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...); +extern int +seq_print_ip_sym(struct trace_seq *s, unsigned long ip, + unsigned long sym_flags); +extern ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, + size_t cnt); +int trace_seq_puts(struct trace_seq *s, const char *str); +int trace_seq_putc(struct trace_seq *s, unsigned char c); +int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len); +int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len); +int trace_seq_path(struct trace_seq *s, struct path *path); +int seq_print_userip_objs(const struct userstack_entry *entry, + struct trace_seq *s, unsigned long sym_flags); +int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, + unsigned long ip, unsigned long sym_flags); + +struct trace_event *ftrace_find_event(int type); +int register_ftrace_event(struct trace_event *event); +int unregister_ftrace_event(struct trace_event *event); + +#define MAX_MEMHEX_BYTES 8 +#define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) + +#endif + diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c index a7172a352f62..b9b13c39b4bb 100644 --- a/kernel/trace/trace_power.c +++ b/kernel/trace/trace_power.c @@ -16,6 +16,7 @@ #include #include "trace.h" +#include "trace_output.h" static struct trace_array *power_trace; static int __read_mostly trace_power_enabled; From f633cef0200bbaec539e2dbb0bc4bed7f022f98b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 23 Dec 2008 23:24:13 -0500 Subject: [PATCH 005/413] ftrace: change trace.c to use registered events Impact: rework trace.c to use new event register API Almost every ftrace event has to implement its output display in trace.c through a different function. Some events did not handle all the formats (trace, latency-trace, raw, hex, binary), and this method does not scale well. This patch converts the format functions to use the event API to find the event and and print its format. Currently, we have a print function for trace, latency_trace, raw, hex and binary. A trace_nop_print is available if the event wants to avoid output on a particular format. Perhaps other tracers could use this in the future (like mmiotrace and function_graph). Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 402 +++---------------------------- kernel/trace/trace_branch.c | 53 ++++ kernel/trace/trace_output.c | 467 ++++++++++++++++++++++++++++++++++++ kernel/trace/trace_output.h | 16 ++ 4 files changed, 574 insertions(+), 364 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 90ce0c1d437b..3f0317586cfd 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1483,15 +1483,6 @@ lat_print_timestamp(struct trace_seq *s, u64 abs_usecs, trace_seq_puts(s, " : "); } -static const char state_to_char[] = TASK_STATE_TO_CHAR_STR; - -static int task_state_char(unsigned long state) -{ - int bit = state ? __ffs(state) + 1 : 0; - - return bit < sizeof(state_to_char) - 1 ? state_to_char[bit] : '?'; -} - static void test_cpu_buff_start(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; @@ -1515,14 +1506,14 @@ print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu) struct trace_seq *s = &iter->seq; unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); struct trace_entry *next_entry; + struct trace_event *event; unsigned long verbose = (trace_flags & TRACE_ITER_VERBOSE); struct trace_entry *entry = iter->ent; unsigned long abs_usecs; unsigned long rel_usecs; u64 next_ts; char *comm; - int S, T; - int i; + int ret; test_cpu_buff_start(iter); @@ -1547,94 +1538,16 @@ print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu) lat_print_generic(s, entry, cpu); lat_print_timestamp(s, abs_usecs, rel_usecs); } - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - trace_assign_type(field, entry); - - seq_print_ip_sym(s, field->ip, sym_flags); - trace_seq_puts(s, " ("); - seq_print_ip_sym(s, field->parent_ip, sym_flags); - trace_seq_puts(s, ")\n"); - break; + event = ftrace_find_event(entry->type); + if (event && event->latency_trace) { + ret = event->latency_trace(s, entry, sym_flags); + if (ret) + return ret; + return TRACE_TYPE_HANDLED; } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = task_state_char(field->prev_state); - comm = trace_find_cmdline(field->next_pid); - trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", - field->prev_pid, - field->prev_prio, - S, entry->type == TRACE_CTX ? "==>" : " +", - field->next_cpu, - field->next_pid, - field->next_prio, - T, comm); - break; - } - case TRACE_SPECIAL: { - struct special_entry *field; - - trace_assign_type(field, entry); - - trace_seq_printf(s, "# %ld %ld %ld\n", - field->arg1, - field->arg2, - field->arg3); - break; - } - case TRACE_STACK: { - struct stack_entry *field; - - trace_assign_type(field, entry); - - for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { - if (i) - trace_seq_puts(s, " <= "); - seq_print_ip_sym(s, field->caller[i], sym_flags); - } - trace_seq_puts(s, "\n"); - break; - } - case TRACE_PRINT: { - struct print_entry *field; - - trace_assign_type(field, entry); - - seq_print_ip_sym(s, field->ip, sym_flags); - trace_seq_printf(s, ": %s", field->buf); - break; - } - case TRACE_BRANCH: { - struct trace_branch *field; - - trace_assign_type(field, entry); - - trace_seq_printf(s, "[%s] %s:%s:%d\n", - field->correct ? " ok " : " MISS ", - field->func, - field->file, - field->line); - break; - } - case TRACE_USER_STACK: { - struct userstack_entry *field; - - trace_assign_type(field, entry); - - seq_print_userip_objs(field, s, sym_flags); - trace_seq_putc(s, '\n'); - break; - } - default: - trace_seq_printf(s, "Unknown type %d\n", entry->type); - } + trace_seq_printf(s, "Unknown type %d\n", entry->type); return TRACE_TYPE_HANDLED; } @@ -1643,13 +1556,12 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) struct trace_seq *s = &iter->seq; unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); struct trace_entry *entry; + struct trace_event *event; unsigned long usec_rem; unsigned long long t; unsigned long secs; char *comm; int ret; - int S, T; - int i; entry = iter->ent; @@ -1671,127 +1583,17 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) if (!ret) return TRACE_TYPE_PARTIAL_LINE; - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - - trace_assign_type(field, entry); - - ret = seq_print_ip_sym(s, field->ip, sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - if ((sym_flags & TRACE_ITER_PRINT_PARENT) && - field->parent_ip) { - ret = trace_seq_printf(s, " <-"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = seq_print_ip_sym(s, - field->parent_ip, - sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - ret = trace_seq_printf(s, "\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; + event = ftrace_find_event(entry->type); + if (event && event->trace) { + ret = event->trace(s, entry, sym_flags); + if (ret) + return ret; + return TRACE_TYPE_HANDLED; } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; + ret = trace_seq_printf(s, "Unknown type %d\n", entry->type); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = task_state_char(field->prev_state); - ret = trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c\n", - field->prev_pid, - field->prev_prio, - S, - entry->type == TRACE_CTX ? "==>" : " +", - field->next_cpu, - field->next_pid, - field->next_prio, - T); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_SPECIAL: { - struct special_entry *field; - - trace_assign_type(field, entry); - - ret = trace_seq_printf(s, "# %ld %ld %ld\n", - field->arg1, - field->arg2, - field->arg3); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_STACK: { - struct stack_entry *field; - - trace_assign_type(field, entry); - - for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { - if (i) { - ret = trace_seq_puts(s, " <= "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - ret = seq_print_ip_sym(s, field->caller[i], - sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - ret = trace_seq_puts(s, "\n"); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_PRINT: { - struct print_entry *field; - - trace_assign_type(field, entry); - - seq_print_ip_sym(s, field->ip, sym_flags); - trace_seq_printf(s, ": %s", field->buf); - break; - } - case TRACE_GRAPH_RET: { - return print_graph_function(iter); - } - case TRACE_GRAPH_ENT: { - return print_graph_function(iter); - } - case TRACE_BRANCH: { - struct trace_branch *field; - - trace_assign_type(field, entry); - - trace_seq_printf(s, "[%s] %s:%s:%d\n", - field->correct ? " ok " : " MISS ", - field->func, - field->file, - field->line); - break; - } - case TRACE_USER_STACK: { - struct userstack_entry *field; - - trace_assign_type(field, entry); - - ret = seq_print_userip_objs(field, s, sym_flags); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_putc(s, '\n'); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - } return TRACE_TYPE_HANDLED; } @@ -1799,8 +1601,8 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; struct trace_entry *entry; + struct trace_event *event; int ret; - int S, T; entry = iter->ent; @@ -1809,86 +1611,26 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) if (!ret) return TRACE_TYPE_PARTIAL_LINE; - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; - - trace_assign_type(field, entry); - - ret = trace_seq_printf(s, "%x %x\n", - field->ip, - field->parent_ip); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; + event = ftrace_find_event(entry->type); + if (event && event->raw) { + ret = event->raw(s, entry, 0); + if (ret) + return ret; + return TRACE_TYPE_HANDLED; } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; + ret = trace_seq_printf(s, "%d ?\n", entry->type); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = entry->type == TRACE_WAKE ? '+' : - task_state_char(field->prev_state); - ret = trace_seq_printf(s, "%d %d %c %d %d %d %c\n", - field->prev_pid, - field->prev_prio, - S, - field->next_cpu, - field->next_pid, - field->next_prio, - T); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_SPECIAL: - case TRACE_USER_STACK: - case TRACE_STACK: { - struct special_entry *field; - - trace_assign_type(field, entry); - - ret = trace_seq_printf(s, "# %ld %ld %ld\n", - field->arg1, - field->arg2, - field->arg3); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - break; - } - case TRACE_PRINT: { - struct print_entry *field; - - trace_assign_type(field, entry); - - trace_seq_printf(s, "# %lx %s", field->ip, field->buf); - break; - } - } return TRACE_TYPE_HANDLED; } -#define SEQ_PUT_FIELD_RET(s, x) \ -do { \ - if (!trace_seq_putmem(s, &(x), sizeof(x))) \ - return 0; \ -} while (0) - -#define SEQ_PUT_HEX_FIELD_RET(s, x) \ -do { \ - BUILD_BUG_ON(sizeof(x) > MAX_MEMHEX_BYTES); \ - if (!trace_seq_putmem_hex(s, &(x), sizeof(x))) \ - return 0; \ -} while (0) - static enum print_line_t print_hex_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; unsigned char newline = '\n'; struct trace_entry *entry; - int S, T; + struct trace_event *event; entry = iter->ent; @@ -1896,47 +1638,10 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) SEQ_PUT_HEX_FIELD_RET(s, iter->cpu); SEQ_PUT_HEX_FIELD_RET(s, iter->ts); - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; + event = ftrace_find_event(entry->type); + if (event && event->hex) + event->hex(s, entry, 0); - trace_assign_type(field, entry); - - SEQ_PUT_HEX_FIELD_RET(s, field->ip); - SEQ_PUT_HEX_FIELD_RET(s, field->parent_ip); - break; - } - case TRACE_CTX: - case TRACE_WAKE: { - struct ctx_switch_entry *field; - - trace_assign_type(field, entry); - - T = task_state_char(field->next_state); - S = entry->type == TRACE_WAKE ? '+' : - task_state_char(field->prev_state); - SEQ_PUT_HEX_FIELD_RET(s, field->prev_pid); - SEQ_PUT_HEX_FIELD_RET(s, field->prev_prio); - SEQ_PUT_HEX_FIELD_RET(s, S); - SEQ_PUT_HEX_FIELD_RET(s, field->next_cpu); - SEQ_PUT_HEX_FIELD_RET(s, field->next_pid); - SEQ_PUT_HEX_FIELD_RET(s, field->next_prio); - SEQ_PUT_HEX_FIELD_RET(s, T); - break; - } - case TRACE_SPECIAL: - case TRACE_USER_STACK: - case TRACE_STACK: { - struct special_entry *field; - - trace_assign_type(field, entry); - - SEQ_PUT_HEX_FIELD_RET(s, field->arg1); - SEQ_PUT_HEX_FIELD_RET(s, field->arg2); - SEQ_PUT_HEX_FIELD_RET(s, field->arg3); - break; - } - } SEQ_PUT_FIELD_RET(s, newline); return TRACE_TYPE_HANDLED; @@ -1962,6 +1667,7 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; struct trace_entry *entry; + struct trace_event *event; entry = iter->ent; @@ -1969,43 +1675,11 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) SEQ_PUT_FIELD_RET(s, entry->cpu); SEQ_PUT_FIELD_RET(s, iter->ts); - switch (entry->type) { - case TRACE_FN: { - struct ftrace_entry *field; + event = ftrace_find_event(entry->type); + if (event && event->binary) + event->binary(s, entry, 0); - trace_assign_type(field, entry); - - SEQ_PUT_FIELD_RET(s, field->ip); - SEQ_PUT_FIELD_RET(s, field->parent_ip); - break; - } - case TRACE_CTX: { - struct ctx_switch_entry *field; - - trace_assign_type(field, entry); - - SEQ_PUT_FIELD_RET(s, field->prev_pid); - SEQ_PUT_FIELD_RET(s, field->prev_prio); - SEQ_PUT_FIELD_RET(s, field->prev_state); - SEQ_PUT_FIELD_RET(s, field->next_pid); - SEQ_PUT_FIELD_RET(s, field->next_prio); - SEQ_PUT_FIELD_RET(s, field->next_state); - break; - } - case TRACE_SPECIAL: - case TRACE_USER_STACK: - case TRACE_STACK: { - struct special_entry *field; - - trace_assign_type(field, entry); - - SEQ_PUT_FIELD_RET(s, field->arg1); - SEQ_PUT_FIELD_RET(s, field->arg2); - SEQ_PUT_FIELD_RET(s, field->arg3); - break; - } - } - return 1; + return TRACE_TYPE_HANDLED; } static int trace_empty(struct trace_iterator *iter) diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 6c00feb3bac7..c15222a01073 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -14,7 +14,9 @@ #include #include #include + #include "trace.h" +#include "trace_output.h" #ifdef CONFIG_BRANCH_TRACER @@ -142,6 +144,49 @@ static void branch_trace_reset(struct trace_array *tr) stop_branch_trace(tr); } +static int +trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct print_entry *field; + + trace_assign_type(field, entry); + + if (seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if (trace_seq_printf(s, ": %s", field->buf)) + goto partial; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static int +trace_branch_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct trace_branch *field; + + trace_assign_type(field, entry); + + if (trace_seq_printf(s, "[%s] %s:%s:%d\n", + field->correct ? " ok " : " MISS ", + field->func, + field->file, + field->line)) + return TRACE_TYPE_PARTIAL_LINE; + + return 0; +} + +static struct trace_event trace_branch_event = { + .type = TRACE_BRANCH, + .trace = trace_branch_print, + .latency_trace = trace_branch_print, + .raw = trace_nop_print, + .hex = trace_nop_print, + .binary = trace_nop_print, +}; + struct tracer branch_trace __read_mostly = { .name = "branch", @@ -154,6 +199,14 @@ struct tracer branch_trace __read_mostly = __init static int init_branch_trace(void) { + int ret; + + ret = register_ftrace_event(&trace_branch_event); + if (!ret) { + printk(KERN_WARNING "Warning: could not register branch events\n"); + return 1; + } + return register_tracer(&branch_trace); } diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 1f3f80002b5e..df0c25cbed30 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -286,6 +286,15 @@ seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) return ret; } +static const char state_to_char[] = TASK_STATE_TO_CHAR_STR; + +static int task_state_char(unsigned long state) +{ + int bit = state ? __ffs(state) + 1 : 0; + + return bit < sizeof(state_to_char) - 1 ? state_to_char[bit] : '?'; +} + /** * ftrace_find_event - find a registered event * @type: the type of event to look for @@ -363,3 +372,461 @@ int unregister_ftrace_event(struct trace_event *event) return 0; } + +/* + * Standard events + */ + +int +trace_nop_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + return 0; +} + +/* TRACE_FN */ +static int +trace_fn_latency(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct ftrace_entry *field; + + trace_assign_type(field, entry); + + if (!seq_print_ip_sym(s, field->ip, flags)) + goto partial; + if (!trace_seq_puts(s, " (")) + goto partial; + if (!seq_print_ip_sym(s, field->parent_ip, flags)) + goto partial; + if (!trace_seq_puts(s, ")\n")) + goto partial; + + return 0; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static int +trace_fn_trace(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct ftrace_entry *field; + + trace_assign_type(field, entry); + + if (!seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if ((flags & TRACE_ITER_PRINT_PARENT) && field->parent_ip) { + if (!trace_seq_printf(s, " <-")) + goto partial; + if (!seq_print_ip_sym(s, + field->parent_ip, + flags)) + goto partial; + } + if (!trace_seq_printf(s, "\n")) + goto partial; + + return 0; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static int +trace_fn_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct ftrace_entry *field; + + trace_assign_type(field, entry); + + if (trace_seq_printf(s, "%x %x\n", + field->ip, + field->parent_ip)) + return TRACE_TYPE_PARTIAL_LINE; + + return 0; +} + +static int +trace_fn_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct ftrace_entry *field; + + trace_assign_type(field, entry); + + SEQ_PUT_HEX_FIELD_RET(s, field->ip); + SEQ_PUT_HEX_FIELD_RET(s, field->parent_ip); + + return 0; +} + +static int +trace_fn_bin(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct ftrace_entry *field; + + trace_assign_type(field, entry); + + SEQ_PUT_FIELD_RET(s, field->ip); + SEQ_PUT_FIELD_RET(s, field->parent_ip); + + return 0; +} + +static struct trace_event trace_fn_event = { + .type = TRACE_FN, + .trace = trace_fn_trace, + .latency_trace = trace_fn_latency, + .raw = trace_fn_raw, + .hex = trace_fn_hex, + .binary = trace_fn_bin, +}; + +/* TRACE_CTX an TRACE_WAKE */ +static int +trace_ctxwake_print(struct trace_seq *s, struct trace_entry *entry, int flags, + char *delim) +{ + struct ctx_switch_entry *field; + char *comm; + int S, T; + + trace_assign_type(field, entry); + + T = task_state_char(field->next_state); + S = task_state_char(field->prev_state); + comm = trace_find_cmdline(field->next_pid); + if (trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", + field->prev_pid, + field->prev_prio, + S, delim, + field->next_cpu, + field->next_pid, + field->next_prio, + T, comm)) + return TRACE_TYPE_PARTIAL_LINE; + + return 0; +} + +static int +trace_ctx_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + return trace_ctxwake_print(s, entry, flags, "==>"); +} + +static int +trace_wake_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + return trace_ctxwake_print(s, entry, flags, " +"); +} + +static int +trace_ctxwake_raw(struct trace_seq *s, struct trace_entry *entry, int flags, + char S) +{ + struct ctx_switch_entry *field; + int T; + + trace_assign_type(field, entry); + + if (!S) + task_state_char(field->prev_state); + T = task_state_char(field->next_state); + if (trace_seq_printf(s, "%d %d %c %d %d %d %c\n", + field->prev_pid, + field->prev_prio, + S, + field->next_cpu, + field->next_pid, + field->next_prio, + T)) + return TRACE_TYPE_PARTIAL_LINE; + + return 0; +} + +static int +trace_ctx_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + return trace_ctxwake_raw(s, entry, flags, 0); +} + +static int +trace_wake_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + return trace_ctxwake_raw(s, entry, flags, '+'); +} + + +static int +trace_ctxwake_hex(struct trace_seq *s, struct trace_entry *entry, int flags, + char S) +{ + struct ctx_switch_entry *field; + int T; + + trace_assign_type(field, entry); + + if (!S) + task_state_char(field->prev_state); + T = task_state_char(field->next_state); + + SEQ_PUT_HEX_FIELD_RET(s, field->prev_pid); + SEQ_PUT_HEX_FIELD_RET(s, field->prev_prio); + SEQ_PUT_HEX_FIELD_RET(s, S); + SEQ_PUT_HEX_FIELD_RET(s, field->next_cpu); + SEQ_PUT_HEX_FIELD_RET(s, field->next_pid); + SEQ_PUT_HEX_FIELD_RET(s, field->next_prio); + SEQ_PUT_HEX_FIELD_RET(s, T); + + return 0; +} + +static int +trace_ctx_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + return trace_ctxwake_hex(s, entry, flags, 0); +} + +static int +trace_wake_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + return trace_ctxwake_hex(s, entry, flags, '+'); +} + +static int +trace_ctxwake_bin(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct ctx_switch_entry *field; + + trace_assign_type(field, entry); + + SEQ_PUT_FIELD_RET(s, field->prev_pid); + SEQ_PUT_FIELD_RET(s, field->prev_prio); + SEQ_PUT_FIELD_RET(s, field->prev_state); + SEQ_PUT_FIELD_RET(s, field->next_pid); + SEQ_PUT_FIELD_RET(s, field->next_prio); + SEQ_PUT_FIELD_RET(s, field->next_state); + + return 0; +} + +static struct trace_event trace_ctx_event = { + .type = TRACE_CTX, + .trace = trace_ctx_print, + .latency_trace = trace_ctx_print, + .raw = trace_ctx_raw, + .hex = trace_ctx_hex, + .binary = trace_ctxwake_bin, +}; + +static struct trace_event trace_wake_event = { + .type = TRACE_WAKE, + .trace = trace_wake_print, + .latency_trace = trace_wake_print, + .raw = trace_wake_raw, + .hex = trace_wake_hex, + .binary = trace_ctxwake_bin, +}; + +/* TRACE_SPECIAL */ +static int +trace_special_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct special_entry *field; + + trace_assign_type(field, entry); + + if (trace_seq_printf(s, "# %ld %ld %ld\n", + field->arg1, + field->arg2, + field->arg3)) + return TRACE_TYPE_PARTIAL_LINE; + + return 0; +} + +static int +trace_special_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct special_entry *field; + + trace_assign_type(field, entry); + + SEQ_PUT_HEX_FIELD_RET(s, field->arg1); + SEQ_PUT_HEX_FIELD_RET(s, field->arg2); + SEQ_PUT_HEX_FIELD_RET(s, field->arg3); + + return 0; +} + +static int +trace_special_bin(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct special_entry *field; + + trace_assign_type(field, entry); + + SEQ_PUT_FIELD_RET(s, field->arg1); + SEQ_PUT_FIELD_RET(s, field->arg2); + SEQ_PUT_FIELD_RET(s, field->arg3); + + return 0; +} + +static struct trace_event trace_special_event = { + .type = TRACE_SPECIAL, + .trace = trace_special_print, + .latency_trace = trace_special_print, + .raw = trace_special_print, + .hex = trace_special_hex, + .binary = trace_special_bin, +}; + +/* TRACE_STACK */ + +static int +trace_stack_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct stack_entry *field; + int i; + + trace_assign_type(field, entry); + + for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { + if (i) { + if (trace_seq_puts(s, " <= ")) + goto partial; + + if (seq_print_ip_sym(s, field->caller[i], flags)) + goto partial; + } + if (trace_seq_puts(s, "\n")) + goto partial; + } + + return 0; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static struct trace_event trace_stack_event = { + .type = TRACE_STACK, + .trace = trace_stack_print, + .latency_trace = trace_stack_print, + .raw = trace_special_print, + .hex = trace_special_hex, + .binary = trace_special_bin, +}; + +/* TRACE_USER_STACK */ +static int +trace_user_stack_print(struct trace_seq *s, struct trace_entry *entry, + int flags) +{ + struct userstack_entry *field; + + trace_assign_type(field, entry); + + if (seq_print_userip_objs(field, s, flags)) + goto partial; + + if (trace_seq_putc(s, '\n')) + goto partial; + + return 0; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static struct trace_event trace_user_stack_event = { + .type = TRACE_USER_STACK, + .trace = trace_user_stack_print, + .latency_trace = trace_user_stack_print, + .raw = trace_special_print, + .hex = trace_special_hex, + .binary = trace_special_bin, +}; + +/* TRACE_PRINT */ +static int +trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct print_entry *field; + + trace_assign_type(field, entry); + + if (seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if (trace_seq_printf(s, ": %s", field->buf)) + goto partial; + + return 0; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static int +trace_print_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +{ + struct print_entry *field; + + trace_assign_type(field, entry); + + if (seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if (trace_seq_printf(s, "# %lx %s", field->ip, field->buf)) + goto partial; + + return 0; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static struct trace_event trace_print_event = { + .type = TRACE_PRINT, + .trace = trace_print_print, + .latency_trace = trace_print_print, + .raw = trace_print_raw, + .hex = trace_nop_print, + .binary = trace_nop_print, +}; + +static struct trace_event *events[] __initdata = { + &trace_fn_event, + &trace_ctx_event, + &trace_wake_event, + &trace_special_event, + &trace_stack_event, + &trace_user_stack_event, + &trace_print_event, + NULL +}; + +__init static int init_events(void) +{ + struct trace_event *event; + int i, ret; + + for (i = 0; events[i]; i++) { + event = events[i]; + + ret = register_ftrace_event(event); + if (!ret) { + printk(KERN_WARNING "event %d failed to register\n", + event->type); + WARN_ON_ONCE(1); + } + } + + return 0; +} +device_initcall(init_events); diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 1fcc76e1378e..ecab4ea4a4fd 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -36,8 +36,24 @@ struct trace_event *ftrace_find_event(int type); int register_ftrace_event(struct trace_event *event); int unregister_ftrace_event(struct trace_event *event); +int +trace_nop_print(struct trace_seq *s, struct trace_entry *entry, int flags); + #define MAX_MEMHEX_BYTES 8 #define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) +#define SEQ_PUT_FIELD_RET(s, x) \ +do { \ + if (!trace_seq_putmem(s, &(x), sizeof(x))) \ + return 0; \ +} while (0) + +#define SEQ_PUT_HEX_FIELD_RET(s, x) \ +do { \ + BUILD_BUG_ON(sizeof(x) > MAX_MEMHEX_BYTES); \ + if (!trace_seq_putmem_hex(s, &(x), sizeof(x))) \ + return 0; \ +} while (0) + #endif From dbd0b4b33074aa6b7832a9d9a5bd985eca5c1aa2 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 28 Dec 2008 20:44:51 -0800 Subject: [PATCH 006/413] tracing/ftrace: provide the base infrastructure for histogram tracing Impact: extend the tracing API The goal of this patch is to normalize and make more easy the implementation of statistical (histogram) tracing. It implements a trace_stat file into the /debugfs/tracing directory where one can print a one-shot output of statistics/histogram entries. A tracer has to provide two basic iterator callbacks: stat_start() => the first entry stat_next(prev, idx) => the next one. Note that it is adapted for arrays or hash tables or lists.... since it provides a pointer to the previous entry and the current index of the iterator. These two callbacks are called to get a snapshot of the statistics at each opening of the trace_stat file because. The values are so updated between two "cat trace_stat". And the tracer is free to lock its datas during the iteration to keep consistent values. Since it is almost always interesting to sort statisticals values to address the problems by priority, this infrastructure provides a "sorting" of the stat entries too if desired. A tracer has just to provide a stat_cmp callback to compare two entries and the stat tracing infrastructure will build a sorted list of the given entries. A last callback, called stat_headers, can be implemented by a tracer to output headers on its trace. If one of these callbacks is changed on runtime, it just have to signal it to the stat tracing API by calling the init_tracer_stat() helper. Changes in V2: - Fix a memory leak if the user opens multiple times the trace_stat file without closing it. Now we always free our list before rebuilding it. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/Makefile | 1 + kernel/trace/trace.c | 3 +- kernel/trace/trace.h | 17 +++ kernel/trace/trace_stat.c | 251 ++++++++++++++++++++++++++++++++++++++ 4 files changed, 271 insertions(+), 1 deletion(-) create mode 100644 kernel/trace/trace_stat.c diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 549f93c9b393..31cd5fbc0eed 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -20,6 +20,7 @@ obj-$(CONFIG_RING_BUFFER) += ring_buffer.o obj-$(CONFIG_TRACING) += trace.o obj-$(CONFIG_TRACING) += trace_output.o +obj-$(CONFIG_TRACING) += trace_stat.o obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o obj-$(CONFIG_SYSPROF_TRACER) += trace_sysprof.o obj-$(CONFIG_FUNCTION_TRACER) += trace_functions.o diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3f0317586cfd..b789c010512c 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2354,6 +2354,7 @@ static int tracing_set_tracer(char *buf) if (ret) goto out; } + init_tracer_stat(t); trace_branch_enable(tr); out: @@ -3206,7 +3207,7 @@ __init static int tracer_alloc_buffers(void) #else current_trace = &nop_trace; #endif - + init_tracer_stat(current_trace); /* All seems OK, enable tracing */ tracing_disabled = 0; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 6bd71fa1e1c7..05fa804d1c16 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -336,6 +336,21 @@ struct tracer { struct tracer *next; int print_max; struct tracer_flags *flags; + + /* + * If you change one of the following on tracing runtime, recall + * init_tracer_stat() + */ + + /* Iteration over statistic entries */ + void *(*stat_start)(void); + void *(*stat_next)(void *prev, int idx); + /* Compare two entries for sorting (optional) for stats */ + int (*stat_cmp)(void *p1, void *p2); + /* Print a stat entry */ + int (*stat_show)(struct seq_file *s, void *p); + /* Print the headers of your stat entries */ + int (*stat_headers)(struct seq_file *s); }; struct trace_seq { @@ -421,6 +436,8 @@ void tracing_start_sched_switch_record(void); int register_tracer(struct tracer *type); void unregister_tracer(struct tracer *type); +void init_tracer_stat(struct tracer *trace); + extern unsigned long nsecs_to_usecs(unsigned long nsecs); extern unsigned long tracing_max_latency; diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c new file mode 100644 index 000000000000..6f194a33a64a --- /dev/null +++ b/kernel/trace/trace_stat.c @@ -0,0 +1,251 @@ +/* + * Infrastructure for statistic tracing (histogram output). + * + * Copyright (C) 2008 Frederic Weisbecker + * + * Based on the code from trace_branch.c which is + * Copyright (C) 2008 Steven Rostedt + * + */ + + +#include +#include +#include +#include "trace.h" + + +/* List of stat entries from a tracer */ +struct trace_stat_list { + struct list_head list; + void *stat; +}; + +static struct trace_stat_list stat_list; + +/* + * This is a copy of the current tracer to avoid racy + * and dangerous output while the current tracer is + * switched. + */ +static struct tracer current_tracer; + +/* + * Protect both the current tracer and the global + * stat list. + */ +static DEFINE_MUTEX(stat_list_mutex); + + +static void reset_stat_list(void) +{ + struct trace_stat_list *node; + struct list_head *next; + + if (list_empty(&stat_list.list)) + return; + + node = list_entry(stat_list.list.next, struct trace_stat_list, list); + next = node->list.next; + + while (&node->list != next) { + kfree(node); + node = list_entry(next, struct trace_stat_list, list); + } + kfree(node); + + INIT_LIST_HEAD(&stat_list.list); +} + +void init_tracer_stat(struct tracer *trace) +{ + mutex_lock(&stat_list_mutex); + current_tracer = *trace; + mutex_unlock(&stat_list_mutex); +} + +/* + * For tracers that don't provide a stat_cmp callback. + * This one will force an immediate insertion on tail of + * the list. + */ +static int dummy_cmp(void *p1, void *p2) +{ + return 1; +} + +/* + * Initialize the stat list at each trace_stat file opening. + * All of these copies and sorting are required on all opening + * since the stats could have changed between two file sessions. + */ +static int stat_seq_init(void) +{ + struct trace_stat_list *iter_entry, *new_entry; + void *prev_stat; + int ret = 0; + int i; + + mutex_lock(&stat_list_mutex); + reset_stat_list(); + + if (!current_tracer.stat_start || !current_tracer.stat_next || + !current_tracer.stat_show) + goto exit; + + if (!current_tracer.stat_cmp) + current_tracer.stat_cmp = dummy_cmp; + + /* + * The first entry. Actually this is the second, but the first + * one (the stat_list head) is pointless. + */ + new_entry = kmalloc(sizeof(struct trace_stat_list), GFP_KERNEL); + if (!new_entry) { + ret = -ENOMEM; + goto exit; + } + + INIT_LIST_HEAD(&new_entry->list); + list_add(&new_entry->list, &stat_list.list); + new_entry->stat = current_tracer.stat_start(); + + prev_stat = new_entry->stat; + + /* + * Iterate over the tracer stat entries and store them in a sorted + * list. + */ + for (i = 1; ; i++) { + new_entry = kmalloc(sizeof(struct trace_stat_list), GFP_KERNEL); + if (!new_entry) { + ret = -ENOMEM; + goto exit_free_list; + } + + INIT_LIST_HEAD(&new_entry->list); + new_entry->stat = current_tracer.stat_next(prev_stat, i); + + /* End of insertion */ + if (!new_entry->stat) + break; + + list_for_each_entry(iter_entry, &stat_list.list, list) { + /* Insertion with a descendent sorting */ + if (current_tracer.stat_cmp(new_entry->stat, + iter_entry->stat) > 0) { + + list_add_tail(&new_entry->list, + &iter_entry->list); + break; + + /* The current smaller value */ + } else if (list_is_last(&iter_entry->list, + &stat_list.list)) { + list_add(&new_entry->list, &iter_entry->list); + break; + } + } + + prev_stat = new_entry->stat; + } +exit: + mutex_unlock(&stat_list_mutex); + return ret; + +exit_free_list: + reset_stat_list(); + mutex_unlock(&stat_list_mutex); + return ret; +} + + +static void *stat_seq_start(struct seq_file *s, loff_t *pos) +{ + struct trace_stat_list *l = (struct trace_stat_list *)s->private; + + /* Prevent from tracer switch or stat_list modification */ + mutex_lock(&stat_list_mutex); + + /* If we are in the beginning of the file, print the headers */ + if (!*pos && current_tracer.stat_headers) + current_tracer.stat_headers(s); + + return seq_list_start(&l->list, *pos); +} + +static void *stat_seq_next(struct seq_file *s, void *p, loff_t *pos) +{ + struct trace_stat_list *l = (struct trace_stat_list *)s->private; + + return seq_list_next(p, &l->list, pos); +} + +static void stat_seq_stop(struct seq_file *m, void *p) +{ + mutex_unlock(&stat_list_mutex); +} + +static int stat_seq_show(struct seq_file *s, void *v) +{ + struct trace_stat_list *l = list_entry(v, struct trace_stat_list, list); + return current_tracer.stat_show(s, l->stat); +} + +static const struct seq_operations trace_stat_seq_ops = { + .start = stat_seq_start, + .next = stat_seq_next, + .stop = stat_seq_stop, + .show = stat_seq_show +}; + +static int tracing_stat_open(struct inode *inode, struct file *file) +{ + int ret; + + ret = seq_open(file, &trace_stat_seq_ops); + if (!ret) { + struct seq_file *m = file->private_data; + m->private = &stat_list; + ret = stat_seq_init(); + } + + return ret; +} + + +/* + * Avoid consuming memory with our now useless list. + */ +static int tracing_stat_release(struct inode *i, struct file *f) +{ + mutex_lock(&stat_list_mutex); + reset_stat_list(); + mutex_unlock(&stat_list_mutex); + return 0; +} + +static const struct file_operations tracing_stat_fops = { + .open = tracing_stat_open, + .read = seq_read, + .llseek = seq_lseek, + .release = tracing_stat_release +}; + +static int __init tracing_stat_init(void) +{ + struct dentry *d_tracing; + struct dentry *entry; + + INIT_LIST_HEAD(&stat_list.list); + d_tracing = tracing_init_dentry(); + + entry = debugfs_create_file("trace_stat", 0444, d_tracing, + NULL, + &tracing_stat_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'trace_stat' entry\n"); + return 0; +} +fs_initcall(tracing_stat_init); From e302cf3f961ceb54c1dd0aff7ba8531df83be07a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 27 Dec 2008 23:25:38 +0100 Subject: [PATCH 007/413] tracing/branch-tracer: adapt to the stat tracing API Impact: refactor the branch tracer This patch adapts the branch tracer to the tracing API. This is a proof of concept because the branch tracer implements two "stat tracing" that were split in two files. So I added an option to the branch tracer: stat_all_branch. If it is set, then trace_stat will output all of the branches entries stats. Otherwise, it will print the annotated branches. Its is a kind of quick trick, waiting for a better solution. By default, the annotated branches stat are sorted by incorrect branch prediction percentage. Ie: correct incorrect % Function File Line ------- --------- - -------- ---- ---- 0 1 100 native_smp_prepare_cpus smpboot.c 1228 0 1 100 hpet_rtc_timer_reinit hpet.c 1057 0 18032 100 sched_info_queued sched_stats.h 223 0 684 100 yield_task_fair sched_fair.c 984 0 282 100 pre_schedule_rt sched_rt.c 1263 0 13414 100 sched_info_dequeued sched_stats.h 178 0 21724 100 sched_info_switch sched_stats.h 270 0 1 100 get_signal_to_deliver signal.c 1820 0 8 100 __cancel_work_timer workqueue.c 560 0 212 100 verify_export_symbols module.c 1509 0 17 100 __rmqueue_fallback page_alloc.c 793 0 43 100 clear_page_mlock internal.h 129 0 124 100 try_to_unmap_anon rmap.c 1021 0 53 100 try_to_unmap_anon rmap.c 1013 0 6 100 vma_address rmap.c 232 0 3301 100 try_to_unmap_file rmap.c 1082 0 466 100 try_to_unmap_file rmap.c 1077 0 1 100 mem_cgroup_create memcontrol.c 1090 0 3 100 inotify_find_update_watch inotify.c 726 2 30163 99 perf_counter_task_sched_out perf_counter.c 385 1 2935 99 percpu_free allocpercpu.c 138 1544 297672 99 dentry_lru_del_init dcache.c 153 8 1074 99 input_pass_event input.c 86 1390 76781 98 mapping_unevictable pagemap.h 50 280 6665 95 pick_next_task_rt sched_rt.c 889 750 4826 86 next_pidmap pid.c 194 2 8 80 blocking_notifier_chain_regist notifier.c 220 36 130 78 ioremap_pte_range ioremap.c 22 1093 3247 74 IS_ERR err.h 34 1023 2908 73 sched_slice sched_fair.c 445 22 60 73 disk_put_part genhd.h 206 [...] It enables a developer to quickly address the source of incorrect branch predictions. Note that this sorting would be better with a second sort on the number of incorrect predictions. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_branch.c | 302 +++++++++++++++++++----------------- 1 file changed, 159 insertions(+), 143 deletions(-) diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index c15222a01073..4785a3b9bc4a 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -18,10 +18,13 @@ #include "trace.h" #include "trace_output.h" +static struct tracer branch_trace; + #ifdef CONFIG_BRANCH_TRACER static int branch_tracing_enabled __read_mostly; static DEFINE_MUTEX(branch_tracing_mutex); + static struct trace_array *branch_tracer; static void @@ -178,6 +181,7 @@ trace_branch_print(struct trace_seq *s, struct trace_entry *entry, int flags) return 0; } + static struct trace_event trace_branch_event = { .type = TRACE_BRANCH, .trace = trace_branch_print, @@ -187,30 +191,6 @@ static struct trace_event trace_branch_event = { .binary = trace_nop_print, }; -struct tracer branch_trace __read_mostly = -{ - .name = "branch", - .init = branch_trace_init, - .reset = branch_trace_reset, -#ifdef CONFIG_FTRACE_SELFTEST - .selftest = trace_selftest_startup_branch, -#endif -}; - -__init static int init_branch_trace(void) -{ - int ret; - - ret = register_ftrace_event(&trace_branch_event); - if (!ret) { - printk(KERN_WARNING "Warning: could not register branch events\n"); - return 1; - } - - return register_tracer(&branch_trace); -} - -device_initcall(init_branch_trace); #else static inline void trace_likely_condition(struct ftrace_branch_data *f, int val, int expect) @@ -236,65 +216,38 @@ void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect) } EXPORT_SYMBOL(ftrace_likely_update); -struct ftrace_pointer { - void *start; - void *stop; - int hit; -}; +extern unsigned long __start_annotated_branch_profile[]; +extern unsigned long __stop_annotated_branch_profile[]; -static void * -t_next(struct seq_file *m, void *v, loff_t *pos) +static int annotated_branch_stat_headers(struct seq_file *m) { - const struct ftrace_pointer *f = m->private; - struct ftrace_branch_data *p = v; - - (*pos)++; - - if (v == (void *)1) - return f->start; - - ++p; - - if ((void *)p >= (void *)f->stop) - return NULL; - - return p; -} - -static void *t_start(struct seq_file *m, loff_t *pos) -{ - void *t = (void *)1; - loff_t l = 0; - - for (; t && l < *pos; t = t_next(m, t, &l)) - ; - - return t; -} - -static void t_stop(struct seq_file *m, void *p) -{ -} - -static int t_show(struct seq_file *m, void *v) -{ - const struct ftrace_pointer *fp = m->private; - struct ftrace_branch_data *p = v; - const char *f; - long percent; - - if (v == (void *)1) { - if (fp->hit) - seq_printf(m, " miss hit %% "); - else - seq_printf(m, " correct incorrect %% "); - seq_printf(m, " Function " + seq_printf(m, " correct incorrect %% "); + seq_printf(m, " Function " " File Line\n" " ------- --------- - " " -------- " " ---- ----\n"); - return 0; - } + return 0; +} + +static inline long get_incorrect_percent(struct ftrace_branch_data *p) +{ + long percent; + + if (p->correct) { + percent = p->incorrect * 100; + percent /= p->correct + p->incorrect; + } else + percent = p->incorrect ? 100 : -1; + + return percent; +} + +static int branch_stat_show(struct seq_file *m, void *v) +{ + struct ftrace_branch_data *p = v; + const char *f; + long percent; /* Only print the file, not the path */ f = p->file + strlen(p->file); @@ -305,11 +258,7 @@ static int t_show(struct seq_file *m, void *v) /* * The miss is overlayed on correct, and hit on incorrect. */ - if (p->correct) { - percent = p->incorrect * 100; - percent /= p->correct + p->incorrect; - } else - percent = p->incorrect ? 100 : -1; + percent = get_incorrect_percent(p); seq_printf(m, "%8lu %8lu ", p->correct, p->incorrect); if (percent < 0) @@ -320,76 +269,143 @@ static int t_show(struct seq_file *m, void *v) return 0; } -static struct seq_operations tracing_likely_seq_ops = { - .start = t_start, - .next = t_next, - .stop = t_stop, - .show = t_show, -}; - -static int tracing_branch_open(struct inode *inode, struct file *file) +static void *annotated_branch_stat_start(void) { - int ret; - - ret = seq_open(file, &tracing_likely_seq_ops); - if (!ret) { - struct seq_file *m = file->private_data; - m->private = (void *)inode->i_private; - } - - return ret; + return __start_annotated_branch_profile; } -static const struct file_operations tracing_branch_fops = { - .open = tracing_branch_open, - .read = seq_read, - .llseek = seq_lseek, -}; +static void * +annotated_branch_stat_next(void *v, int idx) +{ + struct ftrace_branch_data *p = v; + + ++p; + + if ((void *)p >= (void *)__stop_annotated_branch_profile) + return NULL; + + return p; +} + +static int annotated_branch_stat_cmp(void *p1, void *p2) +{ + struct ftrace_branch_data *a = p1; + struct ftrace_branch_data *b = p2; + + long percent_a, percent_b; + + percent_a = get_incorrect_percent(a); + percent_b = get_incorrect_percent(b); + + if (percent_a < percent_b) + return -1; + if (percent_a > percent_b) + return 1; + else + return 0; +} #ifdef CONFIG_PROFILE_ALL_BRANCHES +enum { + TRACE_BRANCH_OPT_ALL = 0x1 +}; + +static struct tracer_opt branch_opts[] = { + { TRACER_OPT(stat_all_branch, TRACE_BRANCH_OPT_ALL) }, + { } +}; + +static struct tracer_flags branch_flags = { + .val = 0, + .opts = branch_opts +}; + extern unsigned long __start_branch_profile[]; extern unsigned long __stop_branch_profile[]; -static const struct ftrace_pointer ftrace_branch_pos = { - .start = __start_branch_profile, - .stop = __stop_branch_profile, - .hit = 1, -}; - -#endif /* CONFIG_PROFILE_ALL_BRANCHES */ - -extern unsigned long __start_annotated_branch_profile[]; -extern unsigned long __stop_annotated_branch_profile[]; - -static const struct ftrace_pointer ftrace_annotated_branch_pos = { - .start = __start_annotated_branch_profile, - .stop = __stop_annotated_branch_profile, -}; - -static __init int ftrace_branch_init(void) +static int all_branch_stat_headers(struct seq_file *m) { - struct dentry *d_tracer; - struct dentry *entry; - - d_tracer = tracing_init_dentry(); - - entry = debugfs_create_file("profile_annotated_branch", 0444, d_tracer, - (void *)&ftrace_annotated_branch_pos, - &tracing_branch_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'profile_annotatet_branch' entry\n"); - -#ifdef CONFIG_PROFILE_ALL_BRANCHES - entry = debugfs_create_file("profile_branch", 0444, d_tracer, - (void *)&ftrace_branch_pos, - &tracing_branch_fops); - if (!entry) - pr_warning("Could not create debugfs" - " 'profile_branch' entry\n"); -#endif - + seq_printf(m, " miss hit %% "); + seq_printf(m, " Function " + " File Line\n" + " ------- --------- - " + " -------- " + " ---- ----\n"); return 0; } -device_initcall(ftrace_branch_init); +static void *all_branch_stat_start(void) +{ + return __start_branch_profile; +} + +static void * +all_branch_stat_next(void *v, int idx) +{ + struct ftrace_branch_data *p = v; + + ++p; + + if ((void *)p >= (void *)__stop_branch_profile) + return NULL; + + return p; +} + +static int branch_set_flag(u32 old_flags, u32 bit, int set) +{ + if (bit == TRACE_BRANCH_OPT_ALL) { + if (set) { + branch_trace.stat_headers = all_branch_stat_headers; + branch_trace.stat_start = all_branch_stat_start; + branch_trace.stat_next = all_branch_stat_next; + branch_trace.stat_cmp = NULL; + } else { + branch_trace.stat_headers = + annotated_branch_stat_headers; + branch_trace.stat_start = annotated_branch_stat_start; + branch_trace.stat_next = annotated_branch_stat_next; + branch_trace.stat_cmp = annotated_branch_stat_cmp; + } + init_tracer_stat(&branch_trace); + } + return 0; +} + +#endif /* CONFIG_PROFILE_ALL_BRANCHES */ + +static struct tracer branch_trace __read_mostly = +{ + .name = "branch", +#ifdef CONFIG_BRANCH_TRACER + .init = branch_trace_init, + .reset = branch_trace_reset, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_branch, +#endif /* CONFIG_FTRACE_SELFTEST */ +#endif /* CONFIG_BRANCH_TRACER */ + .stat_start = annotated_branch_stat_start, + .stat_next = annotated_branch_stat_next, + .stat_show = branch_stat_show, + .stat_headers = annotated_branch_stat_headers, + .stat_cmp = annotated_branch_stat_cmp, +#ifdef CONFIG_PROFILE_ALL_BRANCHES + .flags = &branch_flags, + .set_flag = branch_set_flag, +#endif +}; + +__init static int init_branch_trace(void) +{ +#ifdef CONFIG_BRANCH_TRACER + int ret; + ret = register_ftrace_event(&trace_branch_event); + if (!ret) { + printk(KERN_WARNING "Warning: could not register branch events\n"); + return 1; + } +#endif + + return register_tracer(&branch_trace); +} +device_initcall(init_branch_trace); From f7d48cbde5c0710008caeaf7dbf14f4a9b064940 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 29 Dec 2008 13:02:17 +0100 Subject: [PATCH 008/413] tracing/ftrace: make trace_find_cmdline() generally available Impact: build fix On !CONFIG_CONTEXT_SWITCH_TRACER trace_find_cmdline() is not defined: kernel/trace/trace_output.c: In function 'trace_ctxwake_print': kernel/trace/trace_output.c:499: error: implicit declaration of function 'trace_find_cmdline' kernel/trace/trace_output.c:499: warning: assignment makes pointer from integer without a cast Move it to the generic section in trace.h. Signed-off-by: Ingo Molnar --- kernel/trace/trace.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 05fa804d1c16..a8b624ccd4d6 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -469,10 +469,10 @@ struct tracer_switch_ops { void *private; struct tracer_switch_ops *next; }; - -char *trace_find_cmdline(int pid); #endif /* CONFIG_CONTEXT_SWITCH_TRACER */ +extern char *trace_find_cmdline(int pid); + #ifdef CONFIG_DYNAMIC_FTRACE extern unsigned long ftrace_update_tot_cnt; #define DYN_FTRACE_TEST_NAME trace_selftest_dynamic_test_func From 35995a4d815586bc968a857f7235707940a2f755 Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Tue, 19 Aug 2008 20:43:25 +0300 Subject: [PATCH 009/413] SLUB: Replace __builtin_return_address(0) with _RET_IP_. This patch replaces __builtin_return_address(0) with _RET_IP_, since a previous patch moved _RET_IP_ and _THIS_IP_ to include/linux/kernel.h and they're widely available now. This makes for shorter and easier to read code. [penberg@cs.helsinki.fi: remove _RET_IP_ casts to void pointer] Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- include/linux/slab.h | 8 ++++---- mm/slab.c | 8 ++++---- mm/slub.c | 48 ++++++++++++++++++++++---------------------- 3 files changed, 32 insertions(+), 32 deletions(-) diff --git a/include/linux/slab.h b/include/linux/slab.h index 000da12b5cf0..c97ed28559ec 100644 --- a/include/linux/slab.h +++ b/include/linux/slab.h @@ -253,9 +253,9 @@ static inline void *kmem_cache_alloc_node(struct kmem_cache *cachep, * request comes from. */ #if defined(CONFIG_DEBUG_SLAB) || defined(CONFIG_SLUB) -extern void *__kmalloc_track_caller(size_t, gfp_t, void*); +extern void *__kmalloc_track_caller(size_t, gfp_t, unsigned long); #define kmalloc_track_caller(size, flags) \ - __kmalloc_track_caller(size, flags, __builtin_return_address(0)) + __kmalloc_track_caller(size, flags, _RET_IP_) #else #define kmalloc_track_caller(size, flags) \ __kmalloc(size, flags) @@ -271,10 +271,10 @@ extern void *__kmalloc_track_caller(size_t, gfp_t, void*); * allocation request comes from. */ #if defined(CONFIG_DEBUG_SLAB) || defined(CONFIG_SLUB) -extern void *__kmalloc_node_track_caller(size_t, gfp_t, int, void *); +extern void *__kmalloc_node_track_caller(size_t, gfp_t, int, unsigned long); #define kmalloc_node_track_caller(size, flags, node) \ __kmalloc_node_track_caller(size, flags, node, \ - __builtin_return_address(0)) + _RET_IP_) #else #define kmalloc_node_track_caller(size, flags, node) \ __kmalloc_node(size, flags, node) diff --git a/mm/slab.c b/mm/slab.c index 09187517f9dc..a14787799014 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -3686,9 +3686,9 @@ void *__kmalloc_node(size_t size, gfp_t flags, int node) EXPORT_SYMBOL(__kmalloc_node); void *__kmalloc_node_track_caller(size_t size, gfp_t flags, - int node, void *caller) + int node, unsigned long caller) { - return __do_kmalloc_node(size, flags, node, caller); + return __do_kmalloc_node(size, flags, node, (void *)caller); } EXPORT_SYMBOL(__kmalloc_node_track_caller); #else @@ -3730,9 +3730,9 @@ void *__kmalloc(size_t size, gfp_t flags) } EXPORT_SYMBOL(__kmalloc); -void *__kmalloc_track_caller(size_t size, gfp_t flags, void *caller) +void *__kmalloc_track_caller(size_t size, gfp_t flags, unsigned long caller) { - return __do_kmalloc(size, flags, caller); + return __do_kmalloc(size, flags, (void *)caller); } EXPORT_SYMBOL(__kmalloc_track_caller); diff --git a/mm/slub.c b/mm/slub.c index 7ad489af9561..06da86654875 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -178,7 +178,7 @@ static LIST_HEAD(slab_caches); * Tracking user of a slab. */ struct track { - void *addr; /* Called from address */ + unsigned long addr; /* Called from address */ int cpu; /* Was running on cpu */ int pid; /* Pid context */ unsigned long when; /* When did the operation occur */ @@ -367,7 +367,7 @@ static struct track *get_track(struct kmem_cache *s, void *object, } static void set_track(struct kmem_cache *s, void *object, - enum track_item alloc, void *addr) + enum track_item alloc, unsigned long addr) { struct track *p; @@ -391,8 +391,8 @@ static void init_tracking(struct kmem_cache *s, void *object) if (!(s->flags & SLAB_STORE_USER)) return; - set_track(s, object, TRACK_FREE, NULL); - set_track(s, object, TRACK_ALLOC, NULL); + set_track(s, object, TRACK_FREE, 0UL); + set_track(s, object, TRACK_ALLOC, 0UL); } static void print_track(const char *s, struct track *t) @@ -401,7 +401,7 @@ static void print_track(const char *s, struct track *t) return; printk(KERN_ERR "INFO: %s in %pS age=%lu cpu=%u pid=%d\n", - s, t->addr, jiffies - t->when, t->cpu, t->pid); + s, (void *)t->addr, jiffies - t->when, t->cpu, t->pid); } static void print_tracking(struct kmem_cache *s, void *object) @@ -866,7 +866,7 @@ static void setup_object_debug(struct kmem_cache *s, struct page *page, } static int alloc_debug_processing(struct kmem_cache *s, struct page *page, - void *object, void *addr) + void *object, unsigned long addr) { if (!check_slab(s, page)) goto bad; @@ -906,7 +906,7 @@ static int alloc_debug_processing(struct kmem_cache *s, struct page *page, } static int free_debug_processing(struct kmem_cache *s, struct page *page, - void *object, void *addr) + void *object, unsigned long addr) { if (!check_slab(s, page)) goto fail; @@ -1029,10 +1029,10 @@ static inline void setup_object_debug(struct kmem_cache *s, struct page *page, void *object) {} static inline int alloc_debug_processing(struct kmem_cache *s, - struct page *page, void *object, void *addr) { return 0; } + struct page *page, void *object, unsigned long addr) { return 0; } static inline int free_debug_processing(struct kmem_cache *s, - struct page *page, void *object, void *addr) { return 0; } + struct page *page, void *object, unsigned long addr) { return 0; } static inline int slab_pad_check(struct kmem_cache *s, struct page *page) { return 1; } @@ -1499,8 +1499,8 @@ static inline int node_match(struct kmem_cache_cpu *c, int node) * we need to allocate a new slab. This is the slowest path since it involves * a call to the page allocator and the setup of a new slab. */ -static void *__slab_alloc(struct kmem_cache *s, - gfp_t gfpflags, int node, void *addr, struct kmem_cache_cpu *c) +static void *__slab_alloc(struct kmem_cache *s, gfp_t gfpflags, int node, + unsigned long addr, struct kmem_cache_cpu *c) { void **object; struct page *new; @@ -1584,7 +1584,7 @@ static void *__slab_alloc(struct kmem_cache *s, * Otherwise we can simply pick the next object from the lockless free list. */ static __always_inline void *slab_alloc(struct kmem_cache *s, - gfp_t gfpflags, int node, void *addr) + gfp_t gfpflags, int node, unsigned long addr) { void **object; struct kmem_cache_cpu *c; @@ -1613,14 +1613,14 @@ static __always_inline void *slab_alloc(struct kmem_cache *s, void *kmem_cache_alloc(struct kmem_cache *s, gfp_t gfpflags) { - return slab_alloc(s, gfpflags, -1, __builtin_return_address(0)); + return slab_alloc(s, gfpflags, -1, _RET_IP_); } EXPORT_SYMBOL(kmem_cache_alloc); #ifdef CONFIG_NUMA void *kmem_cache_alloc_node(struct kmem_cache *s, gfp_t gfpflags, int node) { - return slab_alloc(s, gfpflags, node, __builtin_return_address(0)); + return slab_alloc(s, gfpflags, node, _RET_IP_); } EXPORT_SYMBOL(kmem_cache_alloc_node); #endif @@ -1634,7 +1634,7 @@ EXPORT_SYMBOL(kmem_cache_alloc_node); * handling required then we can return immediately. */ static void __slab_free(struct kmem_cache *s, struct page *page, - void *x, void *addr, unsigned int offset) + void *x, unsigned long addr, unsigned int offset) { void *prior; void **object = (void *)x; @@ -1704,7 +1704,7 @@ static void __slab_free(struct kmem_cache *s, struct page *page, * with all sorts of special processing. */ static __always_inline void slab_free(struct kmem_cache *s, - struct page *page, void *x, void *addr) + struct page *page, void *x, unsigned long addr) { void **object = (void *)x; struct kmem_cache_cpu *c; @@ -1731,7 +1731,7 @@ void kmem_cache_free(struct kmem_cache *s, void *x) page = virt_to_head_page(x); - slab_free(s, page, x, __builtin_return_address(0)); + slab_free(s, page, x, _RET_IP_); } EXPORT_SYMBOL(kmem_cache_free); @@ -2659,7 +2659,7 @@ void *__kmalloc(size_t size, gfp_t flags) if (unlikely(ZERO_OR_NULL_PTR(s))) return s; - return slab_alloc(s, flags, -1, __builtin_return_address(0)); + return slab_alloc(s, flags, -1, _RET_IP_); } EXPORT_SYMBOL(__kmalloc); @@ -2687,7 +2687,7 @@ void *__kmalloc_node(size_t size, gfp_t flags, int node) if (unlikely(ZERO_OR_NULL_PTR(s))) return s; - return slab_alloc(s, flags, node, __builtin_return_address(0)); + return slab_alloc(s, flags, node, _RET_IP_); } EXPORT_SYMBOL(__kmalloc_node); #endif @@ -2744,7 +2744,7 @@ void kfree(const void *x) put_page(page); return; } - slab_free(page->slab, page, object, __builtin_return_address(0)); + slab_free(page->slab, page, object, _RET_IP_); } EXPORT_SYMBOL(kfree); @@ -3200,7 +3200,7 @@ static struct notifier_block __cpuinitdata slab_notifier = { #endif -void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, void *caller) +void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, unsigned long caller) { struct kmem_cache *s; @@ -3216,7 +3216,7 @@ void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, void *caller) } void *__kmalloc_node_track_caller(size_t size, gfp_t gfpflags, - int node, void *caller) + int node, unsigned long caller) { struct kmem_cache *s; @@ -3427,7 +3427,7 @@ static void resiliency_test(void) {}; struct location { unsigned long count; - void *addr; + unsigned long addr; long long sum_time; long min_time; long max_time; @@ -3475,7 +3475,7 @@ static int add_location(struct loc_track *t, struct kmem_cache *s, { long start, end, pos; struct location *l; - void *caddr; + unsigned long caddr; unsigned long age = jiffies - track->when; start = -1; From b9ce08c01020eb28bfbfa6faf1c740281c5f418e Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Sun, 10 Aug 2008 20:14:03 +0300 Subject: [PATCH 010/413] kmemtrace: Core implementation. kmemtrace provides tracing for slab allocator functions, such as kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected data is then fed to the userspace application in order to analyse allocation hotspots, internal fragmentation and so on, making it possible to see how well an allocator performs, as well as debug and profile kernel code. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- Documentation/kernel-parameters.txt | 10 + MAINTAINERS | 6 + include/linux/kmemtrace.h | 85 +++++++ init/main.c | 2 + lib/Kconfig.debug | 28 +++ mm/Makefile | 1 + mm/kmemtrace.c | 335 ++++++++++++++++++++++++++++ 7 files changed, 467 insertions(+) create mode 100644 include/linux/kmemtrace.h create mode 100644 mm/kmemtrace.c diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index e0f346d201ed..542c2d8843db 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -49,6 +49,7 @@ parameter is applicable: ISAPNP ISA PnP code is enabled. ISDN Appropriate ISDN support is enabled. JOY Appropriate joystick support is enabled. + KMEMTRACE kmemtrace is enabled. LIBATA Libata driver is enabled LP Printer support is enabled. LOOP Loopback device support is enabled. @@ -1018,6 +1019,15 @@ and is between 256 and 4096 characters. It is defined in the file use the HighMem zone if it exists, and the Normal zone if it does not. + kmemtrace.enable= [KNL,KMEMTRACE] Format: { yes | no } + Controls whether kmemtrace is enabled + at boot-time. + + kmemtrace.subbufs=n [KNL,KMEMTRACE] Overrides the number of + subbufs kmemtrace's relay channel has. Set this + higher than default (KMEMTRACE_N_SUBBUFS in code) if + you experience buffer overruns. + movablecore=nn[KMG] [KNL,X86-32,IA-64,PPC,X86-64] This parameter is similar to kernelcore except it specifies the amount of memory used for migratable allocations. diff --git a/MAINTAINERS b/MAINTAINERS index 618c1ef4a397..e2b3c8555051 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -2565,6 +2565,12 @@ M: jason.wessel@windriver.com L: kgdb-bugreport@lists.sourceforge.net S: Maintained +KMEMTRACE +P: Eduard - Gabriel Munteanu +M: eduard.munteanu@linux360.ro +L: linux-kernel@vger.kernel.org +S: Maintained + KPROBES P: Ananth N Mavinakayanahalli M: ananth@in.ibm.com diff --git a/include/linux/kmemtrace.h b/include/linux/kmemtrace.h new file mode 100644 index 000000000000..2c332010cb4e --- /dev/null +++ b/include/linux/kmemtrace.h @@ -0,0 +1,85 @@ +/* + * Copyright (C) 2008 Eduard - Gabriel Munteanu + * + * This file is released under GPL version 2. + */ + +#ifndef _LINUX_KMEMTRACE_H +#define _LINUX_KMEMTRACE_H + +#ifdef __KERNEL__ + +#include +#include + +enum kmemtrace_type_id { + KMEMTRACE_TYPE_KMALLOC = 0, /* kmalloc() or kfree(). */ + KMEMTRACE_TYPE_CACHE, /* kmem_cache_*(). */ + KMEMTRACE_TYPE_PAGES, /* __get_free_pages() and friends. */ +}; + +#ifdef CONFIG_KMEMTRACE + +extern void kmemtrace_init(void); + +static inline void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags, + int node) +{ + trace_mark(kmemtrace_alloc, "type_id %d call_site %lu ptr %lu " + "bytes_req %lu bytes_alloc %lu gfp_flags %lu node %d", + type_id, call_site, (unsigned long) ptr, + bytes_req, bytes_alloc, (unsigned long) gfp_flags, node); +} + +static inline void kmemtrace_mark_free(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr) +{ + trace_mark(kmemtrace_free, "type_id %d call_site %lu ptr %lu", + type_id, call_site, (unsigned long) ptr); +} + +#else /* CONFIG_KMEMTRACE */ + +static inline void kmemtrace_init(void) +{ +} + +static inline void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags, + int node) +{ +} + +static inline void kmemtrace_mark_free(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr) +{ +} + +#endif /* CONFIG_KMEMTRACE */ + +static inline void kmemtrace_mark_alloc(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags) +{ + kmemtrace_mark_alloc_node(type_id, call_site, ptr, + bytes_req, bytes_alloc, gfp_flags, -1); +} + +#endif /* __KERNEL__ */ + +#endif /* _LINUX_KMEMTRACE_H */ + diff --git a/init/main.c b/init/main.c index 7e117a231af1..be1fe2242a55 100644 --- a/init/main.c +++ b/init/main.c @@ -69,6 +69,7 @@ #include #include #include +#include #ifdef CONFIG_X86_LOCAL_APIC #include @@ -653,6 +654,7 @@ asmlinkage void __init start_kernel(void) enable_debug_pagealloc(); cpu_hotplug_init(); kmem_cache_init(); + kmemtrace_init(); debug_objects_mem_init(); idr_init_cache(); setup_per_cpu_pageset(); diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index b0f239e443bc..78d669b461d2 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -803,6 +803,34 @@ config FIREWIRE_OHCI_REMOTE_DMA If unsure, say N. +config KMEMTRACE + bool "Kernel memory tracer (kmemtrace)" + depends on RELAY && DEBUG_FS && MARKERS + help + kmemtrace provides tracing for slab allocator functions, such as + kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected + data is then fed to the userspace application in order to analyse + allocation hotspots, internal fragmentation and so on, making it + possible to see how well an allocator performs, as well as debug + and profile kernel code. + + This requires an userspace application to use. See + Documentation/vm/kmemtrace.txt for more information. + + Saying Y will make the kernel somewhat larger and slower. However, + if you disable kmemtrace at run-time or boot-time, the performance + impact is minimal (depending on the arch the kernel is built for). + + If unsure, say N. + +config KMEMTRACE_DEFAULT_ENABLED + bool "Enabled by default at boot" + depends on KMEMTRACE + help + Say Y here to enable kmemtrace at boot-time by default. Whatever + the choice, the behavior can be overridden by a kernel parameter, + as described in documentation. + menuconfig BUILD_DOCSRC bool "Build targets in Documentation/ tree" depends on HEADERS_CHECK diff --git a/mm/Makefile b/mm/Makefile index c06b45a1ff5f..3782eb66d4b3 100644 --- a/mm/Makefile +++ b/mm/Makefile @@ -34,3 +34,4 @@ obj-$(CONFIG_MIGRATION) += migrate.o obj-$(CONFIG_SMP) += allocpercpu.o obj-$(CONFIG_QUICKLIST) += quicklist.o obj-$(CONFIG_CGROUP_MEM_RES_CTLR) += memcontrol.o page_cgroup.o +obj-$(CONFIG_KMEMTRACE) += kmemtrace.o diff --git a/mm/kmemtrace.c b/mm/kmemtrace.c new file mode 100644 index 000000000000..83ad1cc71a92 --- /dev/null +++ b/mm/kmemtrace.c @@ -0,0 +1,335 @@ +/* + * Copyright (C) 2008 Pekka Enberg, Eduard - Gabriel Munteanu + * + * This file is released under GPL version 2. + */ + +#include +#include +#include +#include +#include +#include +#include + +#define KMEMTRACE_SUBBUF_SIZE 524288 +#define KMEMTRACE_DEF_N_SUBBUFS 20 + +static struct rchan *kmemtrace_chan; +static u32 kmemtrace_buf_overruns; + +static unsigned int kmemtrace_n_subbufs; +#ifdef CONFIG_KMEMTRACE_DEFAULT_ENABLED +static unsigned int kmemtrace_enabled = 1; +#else +static unsigned int kmemtrace_enabled = 0; +#endif + +/* + * The sequence number is used for reordering kmemtrace packets + * in userspace, since they are logged as per-CPU data. + * + * atomic_t should always be a 32-bit signed integer. Wraparound is not + * likely to occur, but userspace can deal with it by expecting a certain + * sequence number in the next packet that will be read. + */ +static atomic_t kmemtrace_seq_num; + +#define KMEMTRACE_ABI_VERSION 1 + +static u32 kmemtrace_abi_version __read_mostly = KMEMTRACE_ABI_VERSION; + +enum kmemtrace_event_id { + KMEMTRACE_EVENT_ALLOC = 0, + KMEMTRACE_EVENT_FREE, +}; + +struct kmemtrace_event { + u8 event_id; + u8 type_id; + u16 event_size; + s32 seq_num; + u64 call_site; + u64 ptr; +} __attribute__ ((__packed__)); + +struct kmemtrace_stats_alloc { + u64 bytes_req; + u64 bytes_alloc; + u32 gfp_flags; + s32 numa_node; +} __attribute__ ((__packed__)); + +static void kmemtrace_probe_alloc(void *probe_data, void *call_data, + const char *format, va_list *args) +{ + unsigned long flags; + struct kmemtrace_event *ev; + struct kmemtrace_stats_alloc *stats; + void *buf; + + local_irq_save(flags); + + buf = relay_reserve(kmemtrace_chan, + sizeof(struct kmemtrace_event) + + sizeof(struct kmemtrace_stats_alloc)); + if (!buf) + goto failed; + + /* + * Don't convert this to use structure initializers, + * C99 does not guarantee the rvalues evaluation order. + */ + + ev = buf; + ev->event_id = KMEMTRACE_EVENT_ALLOC; + ev->type_id = va_arg(*args, int); + ev->event_size = sizeof(struct kmemtrace_event) + + sizeof(struct kmemtrace_stats_alloc); + ev->seq_num = atomic_add_return(1, &kmemtrace_seq_num); + ev->call_site = va_arg(*args, unsigned long); + ev->ptr = va_arg(*args, unsigned long); + + stats = buf + sizeof(struct kmemtrace_event); + stats->bytes_req = va_arg(*args, unsigned long); + stats->bytes_alloc = va_arg(*args, unsigned long); + stats->gfp_flags = va_arg(*args, unsigned long); + stats->numa_node = va_arg(*args, int); + +failed: + local_irq_restore(flags); +} + +static void kmemtrace_probe_free(void *probe_data, void *call_data, + const char *format, va_list *args) +{ + unsigned long flags; + struct kmemtrace_event *ev; + + local_irq_save(flags); + + ev = relay_reserve(kmemtrace_chan, sizeof(struct kmemtrace_event)); + if (!ev) + goto failed; + + /* + * Don't convert this to use structure initializers, + * C99 does not guarantee the rvalues evaluation order. + */ + ev->event_id = KMEMTRACE_EVENT_FREE; + ev->type_id = va_arg(*args, int); + ev->event_size = sizeof(struct kmemtrace_event); + ev->seq_num = atomic_add_return(1, &kmemtrace_seq_num); + ev->call_site = va_arg(*args, unsigned long); + ev->ptr = va_arg(*args, unsigned long); + +failed: + local_irq_restore(flags); +} + +static struct dentry * +kmemtrace_create_buf_file(const char *filename, struct dentry *parent, + int mode, struct rchan_buf *buf, int *is_global) +{ + return debugfs_create_file(filename, mode, parent, buf, + &relay_file_operations); +} + +static int kmemtrace_remove_buf_file(struct dentry *dentry) +{ + debugfs_remove(dentry); + + return 0; +} + +static int kmemtrace_subbuf_start(struct rchan_buf *buf, + void *subbuf, + void *prev_subbuf, + size_t prev_padding) +{ + if (relay_buf_full(buf)) { + /* + * We know it's not SMP-safe, but neither + * debugfs_create_u32() is. + */ + kmemtrace_buf_overruns++; + return 0; + } + + return 1; +} + +static struct rchan_callbacks relay_callbacks = { + .create_buf_file = kmemtrace_create_buf_file, + .remove_buf_file = kmemtrace_remove_buf_file, + .subbuf_start = kmemtrace_subbuf_start, +}; + +static struct dentry *kmemtrace_dir; +static struct dentry *kmemtrace_overruns_dentry; +static struct dentry *kmemtrace_abi_version_dentry; + +static struct dentry *kmemtrace_enabled_dentry; + +static int kmemtrace_start_probes(void) +{ + int err; + + err = marker_probe_register("kmemtrace_alloc", "type_id %d " + "call_site %lu ptr %lu " + "bytes_req %lu bytes_alloc %lu " + "gfp_flags %lu node %d", + kmemtrace_probe_alloc, NULL); + if (err) + return err; + err = marker_probe_register("kmemtrace_free", "type_id %d " + "call_site %lu ptr %lu", + kmemtrace_probe_free, NULL); + + return err; +} + +static void kmemtrace_stop_probes(void) +{ + marker_probe_unregister("kmemtrace_alloc", + kmemtrace_probe_alloc, NULL); + marker_probe_unregister("kmemtrace_free", + kmemtrace_probe_free, NULL); +} + +static int kmemtrace_enabled_get(void *data, u64 *val) +{ + *val = *((int *) data); + + return 0; +} + +static int kmemtrace_enabled_set(void *data, u64 val) +{ + u64 old_val = kmemtrace_enabled; + + *((int *) data) = !!val; + + if (old_val == val) + return 0; + if (val) + kmemtrace_start_probes(); + else + kmemtrace_stop_probes(); + + return 0; +} + +DEFINE_SIMPLE_ATTRIBUTE(kmemtrace_enabled_fops, + kmemtrace_enabled_get, + kmemtrace_enabled_set, "%llu\n"); + +static void kmemtrace_cleanup(void) +{ + if (kmemtrace_enabled_dentry) + debugfs_remove(kmemtrace_enabled_dentry); + + kmemtrace_stop_probes(); + + if (kmemtrace_abi_version_dentry) + debugfs_remove(kmemtrace_abi_version_dentry); + if (kmemtrace_overruns_dentry) + debugfs_remove(kmemtrace_overruns_dentry); + + relay_close(kmemtrace_chan); + kmemtrace_chan = NULL; + + if (kmemtrace_dir) + debugfs_remove(kmemtrace_dir); +} + +static int __init kmemtrace_setup_late(void) +{ + if (!kmemtrace_chan) + goto failed; + + kmemtrace_dir = debugfs_create_dir("kmemtrace", NULL); + if (!kmemtrace_dir) + goto cleanup; + + kmemtrace_abi_version_dentry = + debugfs_create_u32("abi_version", S_IRUSR, + kmemtrace_dir, &kmemtrace_abi_version); + kmemtrace_overruns_dentry = + debugfs_create_u32("total_overruns", S_IRUSR, + kmemtrace_dir, &kmemtrace_buf_overruns); + if (!kmemtrace_overruns_dentry || !kmemtrace_abi_version_dentry) + goto cleanup; + + kmemtrace_enabled_dentry = + debugfs_create_file("enabled", S_IRUSR | S_IWUSR, + kmemtrace_dir, &kmemtrace_enabled, + &kmemtrace_enabled_fops); + if (!kmemtrace_enabled_dentry) + goto cleanup; + + if (relay_late_setup_files(kmemtrace_chan, "cpu", kmemtrace_dir)) + goto cleanup; + + printk(KERN_INFO "kmemtrace: fully up.\n"); + + return 0; + +cleanup: + kmemtrace_cleanup(); +failed: + return 1; +} +late_initcall(kmemtrace_setup_late); + +static int __init kmemtrace_set_boot_enabled(char *str) +{ + if (!str) + return -EINVAL; + + if (!strcmp(str, "yes")) + kmemtrace_enabled = 1; + else if (!strcmp(str, "no")) + kmemtrace_enabled = 0; + else + return -EINVAL; + + return 0; +} +early_param("kmemtrace.enable", kmemtrace_set_boot_enabled); + +static int __init kmemtrace_set_subbufs(char *str) +{ + get_option(&str, &kmemtrace_n_subbufs); + return 0; +} +early_param("kmemtrace.subbufs", kmemtrace_set_subbufs); + +void kmemtrace_init(void) +{ + if (!kmemtrace_enabled) + return; + + if (!kmemtrace_n_subbufs) + kmemtrace_n_subbufs = KMEMTRACE_DEF_N_SUBBUFS; + + kmemtrace_chan = relay_open(NULL, NULL, KMEMTRACE_SUBBUF_SIZE, + kmemtrace_n_subbufs, &relay_callbacks, + NULL); + if (unlikely(!kmemtrace_chan)) { + printk(KERN_ERR "kmemtrace: could not open relay channel.\n"); + return; + } + + if (unlikely(kmemtrace_start_probes())) + goto probe_fail; + + printk(KERN_INFO "kmemtrace: early init successful.\n"); + + return; + +probe_fail: + printk(KERN_ERR "kmemtrace: could not register marker probes!\n"); + kmemtrace_cleanup(); +} + From aa46a7e0228c0477708ce44a0c5621902b3c157c Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Sun, 10 Aug 2008 20:14:04 +0300 Subject: [PATCH 011/413] kmemtrace: Additional documentation. Documented kmemtrace's ABI, purpose and design. Also includes a short usage guide, FAQ, as well as a link to the userspace application's Git repository, which is currently hosted at repo.or.cz. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- Documentation/ABI/testing/debugfs-kmemtrace | 71 +++++++++++ Documentation/vm/kmemtrace.txt | 126 ++++++++++++++++++++ 2 files changed, 197 insertions(+) create mode 100644 Documentation/ABI/testing/debugfs-kmemtrace create mode 100644 Documentation/vm/kmemtrace.txt diff --git a/Documentation/ABI/testing/debugfs-kmemtrace b/Documentation/ABI/testing/debugfs-kmemtrace new file mode 100644 index 000000000000..a5ff9a64c9ee --- /dev/null +++ b/Documentation/ABI/testing/debugfs-kmemtrace @@ -0,0 +1,71 @@ +What: /sys/kernel/debug/kmemtrace/ +Date: July 2008 +Contact: Eduard - Gabriel Munteanu +Description: + +In kmemtrace-enabled kernels, the following files are created: + +/sys/kernel/debug/kmemtrace/ + cpu (0400) Per-CPU tracing data, see below. (binary) + total_overruns (0400) Total number of bytes which were dropped from + cpu files because of full buffer condition, + non-binary. (text) + abi_version (0400) Kernel's kmemtrace ABI version. (text) + +Each per-CPU file should be read according to the relay interface. That is, +the reader should set affinity to that specific CPU and, as currently done by +the userspace application (though there are other methods), use poll() with +an infinite timeout before every read(). Otherwise, erroneous data may be +read. The binary data has the following _core_ format: + + Event ID (1 byte) Unsigned integer, one of: + 0 - represents an allocation (KMEMTRACE_EVENT_ALLOC) + 1 - represents a freeing of previously allocated memory + (KMEMTRACE_EVENT_FREE) + Type ID (1 byte) Unsigned integer, one of: + 0 - this is a kmalloc() / kfree() + 1 - this is a kmem_cache_alloc() / kmem_cache_free() + 2 - this is a __get_free_pages() et al. + Event size (2 bytes) Unsigned integer representing the + size of this event. Used to extend + kmemtrace. Discard the bytes you + don't know about. + Sequence number (4 bytes) Signed integer used to reorder data + logged on SMP machines. Wraparound + must be taken into account, although + it is unlikely. + Caller address (8 bytes) Return address to the caller. + Pointer to mem (8 bytes) Pointer to target memory area. Can be + NULL, but not all such calls might be + recorded. + +In case of KMEMTRACE_EVENT_ALLOC events, the next fields follow: + + Requested bytes (8 bytes) Total number of requested bytes, + unsigned, must not be zero. + Allocated bytes (8 bytes) Total number of actually allocated + bytes, unsigned, must not be lower + than requested bytes. + Requested flags (4 bytes) GFP flags supplied by the caller. + Target CPU (4 bytes) Signed integer, valid for event id 1. + If equal to -1, target CPU is the same + as origin CPU, but the reverse might + not be true. + +The data is made available in the same endianness the machine has. + +Other event ids and type ids may be defined and added. Other fields may be +added by increasing event size, but see below for details. +Every modification to the ABI, including new id definitions, are followed +by bumping the ABI version by one. + +Adding new data to the packet (features) is done at the end of the mandatory +data: + Feature size (2 byte) + Feature ID (1 byte) + Feature data (Feature size - 4 bytes) + + +Users: + kmemtrace-user - git://repo.or.cz/kmemtrace-user.git + diff --git a/Documentation/vm/kmemtrace.txt b/Documentation/vm/kmemtrace.txt new file mode 100644 index 000000000000..75360b142ba4 --- /dev/null +++ b/Documentation/vm/kmemtrace.txt @@ -0,0 +1,126 @@ + kmemtrace - Kernel Memory Tracer + + by Eduard - Gabriel Munteanu + + +I. Introduction +=============== + +kmemtrace helps kernel developers figure out two things: +1) how different allocators (SLAB, SLUB etc.) perform +2) how kernel code allocates memory and how much + +To do this, we trace every allocation and export information to the userspace +through the relay interface. We export things such as the number of requested +bytes, the number of bytes actually allocated (i.e. including internal +fragmentation), whether this is a slab allocation or a plain kmalloc() and so +on. + +The actual analysis is performed by a userspace tool (see section III for +details on where to get it from). It logs the data exported by the kernel, +processes it and (as of writing this) can provide the following information: +- the total amount of memory allocated and fragmentation per call-site +- the amount of memory allocated and fragmentation per allocation +- total memory allocated and fragmentation in the collected dataset +- number of cross-CPU allocation and frees (makes sense in NUMA environments) + +Moreover, it can potentially find inconsistent and erroneous behavior in +kernel code, such as using slab free functions on kmalloc'ed memory or +allocating less memory than requested (but not truly failed allocations). + +kmemtrace also makes provisions for tracing on some arch and analysing the +data on another. + +II. Design and goals +==================== + +kmemtrace was designed to handle rather large amounts of data. Thus, it uses +the relay interface to export whatever is logged to userspace, which then +stores it. Analysis and reporting is done asynchronously, that is, after the +data is collected and stored. By design, it allows one to log and analyse +on different machines and different arches. + +As of writing this, the ABI is not considered stable, though it might not +change much. However, no guarantees are made about compatibility yet. When +deemed stable, the ABI should still allow easy extension while maintaining +backward compatibility. This is described further in Documentation/ABI. + +Summary of design goals: + - allow logging and analysis to be done across different machines + - be fast and anticipate usage in high-load environments (*) + - be reasonably extensible + - make it possible for GNU/Linux distributions to have kmemtrace + included in their repositories + +(*) - one of the reasons Pekka Enberg's original userspace data analysis + tool's code was rewritten from Perl to C (although this is more than a + simple conversion) + + +III. Quick usage guide +====================== + +1) Get a kernel that supports kmemtrace and build it accordingly (i.e. enable +CONFIG_KMEMTRACE and CONFIG_DEFAULT_ENABLED). + +2) Get the userspace tool and build it: +$ git-clone git://repo.or.cz/kmemtrace-user.git # current repository +$ cd kmemtrace-user/ +$ ./autogen.sh +$ ./configure +$ make + +3) Boot the kmemtrace-enabled kernel if you haven't, preferably in the +'single' runlevel (so that relay buffers don't fill up easily), and run +kmemtrace: +# '$' does not mean user, but root here. +$ mount -t debugfs none /sys/kernel/debug +$ mount -t proc none /proc +$ cd path/to/kmemtrace-user/ +$ ./kmemtraced +Wait a bit, then stop it with CTRL+C. +$ cat /sys/kernel/debug/kmemtrace/total_overruns # Check if we didn't + # overrun, should + # be zero. +$ (Optionally) [Run kmemtrace_check separately on each cpu[0-9]*.out file to + check its correctness] +$ ./kmemtrace-report + +Now you should have a nice and short summary of how the allocator performs. + +IV. FAQ and known issues +======================== + +Q: 'cat /sys/kernel/debug/kmemtrace/total_overruns' is non-zero, how do I fix +this? Should I worry? +A: If it's non-zero, this affects kmemtrace's accuracy, depending on how +large the number is. You can fix it by supplying a higher +'kmemtrace.subbufs=N' kernel parameter. +--- + +Q: kmemtrace_check reports errors, how do I fix this? Should I worry? +A: This is a bug and should be reported. It can occur for a variety of +reasons: + - possible bugs in relay code + - possible misuse of relay by kmemtrace + - timestamps being collected unorderly +Or you may fix it yourself and send us a patch. +--- + +Q: kmemtrace_report shows many errors, how do I fix this? Should I worry? +A: This is a known issue and I'm working on it. These might be true errors +in kernel code, which may have inconsistent behavior (e.g. allocating memory +with kmem_cache_alloc() and freeing it with kfree()). Pekka Enberg pointed +out this behavior may work with SLAB, but may fail with other allocators. + +It may also be due to lack of tracing in some unusual allocator functions. + +We don't want bug reports regarding this issue yet. +--- + +V. See also +=========== + +Documentation/kernel-parameters.txt +Documentation/ABI/testing/debugfs-kmemtrace + From 36555751c6751a5bdfd6d7bdf0648343bb1ef0de Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Sun, 10 Aug 2008 20:14:05 +0300 Subject: [PATCH 012/413] kmemtrace: SLAB hooks. This adds hooks for the SLAB allocator, to allow tracing with kmemtrace. We also convert some inline functions to __always_inline to make sure _RET_IP_, which expands to __builtin_return_address(0), always works as expected. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- include/linux/slab_def.h | 68 +++++++++++++++++++++++++++++++++----- mm/slab.c | 71 ++++++++++++++++++++++++++++++++++++---- 2 files changed, 123 insertions(+), 16 deletions(-) diff --git a/include/linux/slab_def.h b/include/linux/slab_def.h index 39c3a5eb8ebe..7555ce99f6d2 100644 --- a/include/linux/slab_def.h +++ b/include/linux/slab_def.h @@ -14,6 +14,7 @@ #include /* kmalloc_sizes.h needs PAGE_SIZE */ #include /* kmalloc_sizes.h needs L1_CACHE_BYTES */ #include +#include /* Size description struct for general caches. */ struct cache_sizes { @@ -28,8 +29,26 @@ extern struct cache_sizes malloc_sizes[]; void *kmem_cache_alloc(struct kmem_cache *, gfp_t); void *__kmalloc(size_t size, gfp_t flags); -static inline void *kmalloc(size_t size, gfp_t flags) +#ifdef CONFIG_KMEMTRACE +extern void *kmem_cache_alloc_notrace(struct kmem_cache *cachep, gfp_t flags); +extern size_t slab_buffer_size(struct kmem_cache *cachep); +#else +static __always_inline void * +kmem_cache_alloc_notrace(struct kmem_cache *cachep, gfp_t flags) { + return kmem_cache_alloc(cachep, flags); +} +static inline size_t slab_buffer_size(struct kmem_cache *cachep) +{ + return 0; +} +#endif + +static __always_inline void *kmalloc(size_t size, gfp_t flags) +{ + struct kmem_cache *cachep; + void *ret; + if (__builtin_constant_p(size)) { int i = 0; @@ -50,10 +69,17 @@ static inline void *kmalloc(size_t size, gfp_t flags) found: #ifdef CONFIG_ZONE_DMA if (flags & GFP_DMA) - return kmem_cache_alloc(malloc_sizes[i].cs_dmacachep, - flags); + cachep = malloc_sizes[i].cs_dmacachep; + else #endif - return kmem_cache_alloc(malloc_sizes[i].cs_cachep, flags); + cachep = malloc_sizes[i].cs_cachep; + + ret = kmem_cache_alloc_notrace(cachep, flags); + + kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, _THIS_IP_, ret, + size, slab_buffer_size(cachep), flags); + + return ret; } return __kmalloc(size, flags); } @@ -62,8 +88,25 @@ static inline void *kmalloc(size_t size, gfp_t flags) extern void *__kmalloc_node(size_t size, gfp_t flags, int node); extern void *kmem_cache_alloc_node(struct kmem_cache *, gfp_t flags, int node); -static inline void *kmalloc_node(size_t size, gfp_t flags, int node) +#ifdef CONFIG_KMEMTRACE +extern void *kmem_cache_alloc_node_notrace(struct kmem_cache *cachep, + gfp_t flags, + int nodeid); +#else +static __always_inline void * +kmem_cache_alloc_node_notrace(struct kmem_cache *cachep, + gfp_t flags, + int nodeid) { + return kmem_cache_alloc_node(cachep, flags, nodeid); +} +#endif + +static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node) +{ + struct kmem_cache *cachep; + void *ret; + if (__builtin_constant_p(size)) { int i = 0; @@ -84,11 +127,18 @@ static inline void *kmalloc_node(size_t size, gfp_t flags, int node) found: #ifdef CONFIG_ZONE_DMA if (flags & GFP_DMA) - return kmem_cache_alloc_node(malloc_sizes[i].cs_dmacachep, - flags, node); + cachep = malloc_sizes[i].cs_dmacachep; + else #endif - return kmem_cache_alloc_node(malloc_sizes[i].cs_cachep, - flags, node); + cachep = malloc_sizes[i].cs_cachep; + + ret = kmem_cache_alloc_node_notrace(cachep, flags, node); + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, _THIS_IP_, + ret, size, slab_buffer_size(cachep), + flags, node); + + return ret; } return __kmalloc_node(size, flags, node); } diff --git a/mm/slab.c b/mm/slab.c index a14787799014..b6d9b8cdefa9 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -112,6 +112,7 @@ #include #include #include +#include #include #include @@ -568,6 +569,14 @@ static void **dbg_userword(struct kmem_cache *cachep, void *objp) #endif +#ifdef CONFIG_KMEMTRACE +size_t slab_buffer_size(struct kmem_cache *cachep) +{ + return cachep->buffer_size; +} +EXPORT_SYMBOL(slab_buffer_size); +#endif + /* * Do not go above this order unless 0 objects fit into the slab. */ @@ -3613,10 +3622,23 @@ static inline void __cache_free(struct kmem_cache *cachep, void *objp) */ void *kmem_cache_alloc(struct kmem_cache *cachep, gfp_t flags) { - return __cache_alloc(cachep, flags, __builtin_return_address(0)); + void *ret = __cache_alloc(cachep, flags, __builtin_return_address(0)); + + kmemtrace_mark_alloc(KMEMTRACE_TYPE_CACHE, _RET_IP_, ret, + obj_size(cachep), cachep->buffer_size, flags); + + return ret; } EXPORT_SYMBOL(kmem_cache_alloc); +#ifdef CONFIG_KMEMTRACE +void *kmem_cache_alloc_notrace(struct kmem_cache *cachep, gfp_t flags) +{ + return __cache_alloc(cachep, flags, __builtin_return_address(0)); +} +EXPORT_SYMBOL(kmem_cache_alloc_notrace); +#endif + /** * kmem_ptr_validate - check if an untrusted pointer might be a slab entry. * @cachep: the cache we're checking against @@ -3661,23 +3683,47 @@ int kmem_ptr_validate(struct kmem_cache *cachep, const void *ptr) #ifdef CONFIG_NUMA void *kmem_cache_alloc_node(struct kmem_cache *cachep, gfp_t flags, int nodeid) { - return __cache_alloc_node(cachep, flags, nodeid, - __builtin_return_address(0)); + void *ret = __cache_alloc_node(cachep, flags, nodeid, + __builtin_return_address(0)); + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_CACHE, _RET_IP_, ret, + obj_size(cachep), cachep->buffer_size, + flags, nodeid); + + return ret; } EXPORT_SYMBOL(kmem_cache_alloc_node); +#ifdef CONFIG_KMEMTRACE +void *kmem_cache_alloc_node_notrace(struct kmem_cache *cachep, + gfp_t flags, + int nodeid) +{ + return __cache_alloc_node(cachep, flags, nodeid, + __builtin_return_address(0)); +} +EXPORT_SYMBOL(kmem_cache_alloc_node_notrace); +#endif + static __always_inline void * __do_kmalloc_node(size_t size, gfp_t flags, int node, void *caller) { struct kmem_cache *cachep; + void *ret; cachep = kmem_find_general_cachep(size, flags); if (unlikely(ZERO_OR_NULL_PTR(cachep))) return cachep; - return kmem_cache_alloc_node(cachep, flags, node); + ret = kmem_cache_alloc_node_notrace(cachep, flags, node); + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, + (unsigned long) caller, ret, + size, cachep->buffer_size, flags, node); + + return ret; } -#ifdef CONFIG_DEBUG_SLAB +#if defined(CONFIG_DEBUG_SLAB) || defined(CONFIG_KMEMTRACE) void *__kmalloc_node(size_t size, gfp_t flags, int node) { return __do_kmalloc_node(size, flags, node, @@ -3710,6 +3756,7 @@ static __always_inline void *__do_kmalloc(size_t size, gfp_t flags, void *caller) { struct kmem_cache *cachep; + void *ret; /* If you want to save a few bytes .text space: replace * __ with kmem_. @@ -3719,11 +3766,17 @@ static __always_inline void *__do_kmalloc(size_t size, gfp_t flags, cachep = __find_general_cachep(size, flags); if (unlikely(ZERO_OR_NULL_PTR(cachep))) return cachep; - return __cache_alloc(cachep, flags, caller); + ret = __cache_alloc(cachep, flags, caller); + + kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, + (unsigned long) caller, ret, + size, cachep->buffer_size, flags); + + return ret; } -#ifdef CONFIG_DEBUG_SLAB +#if defined(CONFIG_DEBUG_SLAB) || defined(CONFIG_KMEMTRACE) void *__kmalloc(size_t size, gfp_t flags) { return __do_kmalloc(size, flags, __builtin_return_address(0)); @@ -3762,6 +3815,8 @@ void kmem_cache_free(struct kmem_cache *cachep, void *objp) debug_check_no_obj_freed(objp, obj_size(cachep)); __cache_free(cachep, objp); local_irq_restore(flags); + + kmemtrace_mark_free(KMEMTRACE_TYPE_CACHE, _RET_IP_, objp); } EXPORT_SYMBOL(kmem_cache_free); @@ -3788,6 +3843,8 @@ void kfree(const void *objp) debug_check_no_obj_freed(objp, obj_size(c)); __cache_free(c, (void *)objp); local_irq_restore(flags); + + kmemtrace_mark_free(KMEMTRACE_TYPE_KMALLOC, _RET_IP_, objp); } EXPORT_SYMBOL(kfree); From 3eae2cb24a96509e0a38cc48dc1538a2826f4e33 Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Sun, 10 Aug 2008 20:14:07 +0300 Subject: [PATCH 013/413] kmemtrace: SLOB hooks. This adds hooks for the SLOB allocator, to allow tracing with kmemtrace. We also convert some inline functions to __always_inline to make sure _RET_IP_, which expands to __builtin_return_address(0), always works as expected. Acked-by: Matt Mackall Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- include/linux/slob_def.h | 9 +++++---- mm/slob.c | 39 ++++++++++++++++++++++++++++++++------- 2 files changed, 37 insertions(+), 11 deletions(-) diff --git a/include/linux/slob_def.h b/include/linux/slob_def.h index 59a3fa476ab9..0ec00b39d006 100644 --- a/include/linux/slob_def.h +++ b/include/linux/slob_def.h @@ -3,14 +3,15 @@ void *kmem_cache_alloc_node(struct kmem_cache *, gfp_t flags, int node); -static inline void *kmem_cache_alloc(struct kmem_cache *cachep, gfp_t flags) +static __always_inline void *kmem_cache_alloc(struct kmem_cache *cachep, + gfp_t flags) { return kmem_cache_alloc_node(cachep, flags, -1); } void *__kmalloc_node(size_t size, gfp_t flags, int node); -static inline void *kmalloc_node(size_t size, gfp_t flags, int node) +static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node) { return __kmalloc_node(size, flags, node); } @@ -23,12 +24,12 @@ static inline void *kmalloc_node(size_t size, gfp_t flags, int node) * kmalloc is the normal method of allocating memory * in the kernel. */ -static inline void *kmalloc(size_t size, gfp_t flags) +static __always_inline void *kmalloc(size_t size, gfp_t flags) { return __kmalloc_node(size, flags, -1); } -static inline void *__kmalloc(size_t size, gfp_t flags) +static __always_inline void *__kmalloc(size_t size, gfp_t flags) { return kmalloc(size, flags); } diff --git a/mm/slob.c b/mm/slob.c index cb675d126791..55de44ae5d30 100644 --- a/mm/slob.c +++ b/mm/slob.c @@ -65,6 +65,7 @@ #include #include #include +#include #include /* @@ -463,27 +464,38 @@ void *__kmalloc_node(size_t size, gfp_t gfp, int node) { unsigned int *m; int align = max(ARCH_KMALLOC_MINALIGN, ARCH_SLAB_MINALIGN); + void *ret; if (size < PAGE_SIZE - align) { if (!size) return ZERO_SIZE_PTR; m = slob_alloc(size + align, gfp, align, node); + if (!m) return NULL; *m = size; - return (void *)m + align; - } else { - void *ret; + ret = (void *)m + align; - ret = slob_new_page(gfp | __GFP_COMP, get_order(size), node); + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, + _RET_IP_, ret, + size, size + align, gfp, node); + } else { + unsigned int order = get_order(size); + + ret = slob_new_page(gfp | __GFP_COMP, order, node); if (ret) { struct page *page; page = virt_to_page(ret); page->private = size; } - return ret; + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, + _RET_IP_, ret, + size, PAGE_SIZE << order, gfp, node); } + + return ret; } EXPORT_SYMBOL(__kmalloc_node); @@ -501,6 +513,8 @@ void kfree(const void *block) slob_free(m, *m + align); } else put_page(&sp->page); + + kmemtrace_mark_free(KMEMTRACE_TYPE_KMALLOC, _RET_IP_, block); } EXPORT_SYMBOL(kfree); @@ -569,10 +583,19 @@ void *kmem_cache_alloc_node(struct kmem_cache *c, gfp_t flags, int node) { void *b; - if (c->size < PAGE_SIZE) + if (c->size < PAGE_SIZE) { b = slob_alloc(c->size, flags, c->align, node); - else + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_CACHE, + _RET_IP_, b, c->size, + SLOB_UNITS(c->size) * SLOB_UNIT, + flags, node); + } else { b = slob_new_page(flags, get_order(c->size), node); + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_CACHE, + _RET_IP_, b, c->size, + PAGE_SIZE << get_order(c->size), + flags, node); + } if (c->ctor) c->ctor(b); @@ -608,6 +631,8 @@ void kmem_cache_free(struct kmem_cache *c, void *b) } else { __kmem_cache_free(b, c->size); } + + kmemtrace_mark_free(KMEMTRACE_TYPE_CACHE, _RET_IP_, b); } EXPORT_SYMBOL(kmem_cache_free); From 5b882be4e00e53a44f47ad7eb997cac2938848bf Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Tue, 19 Aug 2008 20:43:26 +0300 Subject: [PATCH 014/413] kmemtrace: SLUB hooks. This adds hooks for the SLUB allocator, to allow tracing with kmemtrace. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- include/linux/slub_def.h | 53 ++++++++++++++++++++++++++++++-- mm/slub.c | 65 ++++++++++++++++++++++++++++++++++++---- 2 files changed, 109 insertions(+), 9 deletions(-) diff --git a/include/linux/slub_def.h b/include/linux/slub_def.h index 2f5c16b1aacd..dc28432b5b9a 100644 --- a/include/linux/slub_def.h +++ b/include/linux/slub_def.h @@ -10,6 +10,7 @@ #include #include #include +#include enum stat_item { ALLOC_FASTPATH, /* Allocation from cpu slab */ @@ -204,13 +205,31 @@ static __always_inline struct kmem_cache *kmalloc_slab(size_t size) void *kmem_cache_alloc(struct kmem_cache *, gfp_t); void *__kmalloc(size_t size, gfp_t flags); +#ifdef CONFIG_KMEMTRACE +extern void *kmem_cache_alloc_notrace(struct kmem_cache *s, gfp_t gfpflags); +#else +static __always_inline void * +kmem_cache_alloc_notrace(struct kmem_cache *s, gfp_t gfpflags) +{ + return kmem_cache_alloc(s, gfpflags); +} +#endif + static __always_inline void *kmalloc_large(size_t size, gfp_t flags) { - return (void *)__get_free_pages(flags | __GFP_COMP, get_order(size)); + unsigned int order = get_order(size); + void *ret = (void *) __get_free_pages(flags | __GFP_COMP, order); + + kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, _THIS_IP_, ret, + size, PAGE_SIZE << order, flags); + + return ret; } static __always_inline void *kmalloc(size_t size, gfp_t flags) { + void *ret; + if (__builtin_constant_p(size)) { if (size > PAGE_SIZE) return kmalloc_large(size, flags); @@ -221,7 +240,13 @@ static __always_inline void *kmalloc(size_t size, gfp_t flags) if (!s) return ZERO_SIZE_PTR; - return kmem_cache_alloc(s, flags); + ret = kmem_cache_alloc_notrace(s, flags); + + kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, + _THIS_IP_, ret, + size, s->size, flags); + + return ret; } } return __kmalloc(size, flags); @@ -231,8 +256,24 @@ static __always_inline void *kmalloc(size_t size, gfp_t flags) void *__kmalloc_node(size_t size, gfp_t flags, int node); void *kmem_cache_alloc_node(struct kmem_cache *, gfp_t flags, int node); +#ifdef CONFIG_KMEMTRACE +extern void *kmem_cache_alloc_node_notrace(struct kmem_cache *s, + gfp_t gfpflags, + int node); +#else +static __always_inline void * +kmem_cache_alloc_node_notrace(struct kmem_cache *s, + gfp_t gfpflags, + int node) +{ + return kmem_cache_alloc_node(s, gfpflags, node); +} +#endif + static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node) { + void *ret; + if (__builtin_constant_p(size) && size <= PAGE_SIZE && !(flags & SLUB_DMA)) { struct kmem_cache *s = kmalloc_slab(size); @@ -240,7 +281,13 @@ static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node) if (!s) return ZERO_SIZE_PTR; - return kmem_cache_alloc_node(s, flags, node); + ret = kmem_cache_alloc_node_notrace(s, flags, node); + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, + _THIS_IP_, ret, + size, s->size, flags, node); + + return ret; } return __kmalloc_node(size, flags, node); } diff --git a/mm/slub.c b/mm/slub.c index 06da86654875..4c48a0146afd 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -24,6 +24,7 @@ #include #include #include +#include /* * Lock order: @@ -1613,18 +1614,46 @@ static __always_inline void *slab_alloc(struct kmem_cache *s, void *kmem_cache_alloc(struct kmem_cache *s, gfp_t gfpflags) { - return slab_alloc(s, gfpflags, -1, _RET_IP_); + void *ret = slab_alloc(s, gfpflags, -1, _RET_IP_); + + kmemtrace_mark_alloc(KMEMTRACE_TYPE_CACHE, _RET_IP_, ret, + s->objsize, s->size, gfpflags); + + return ret; } EXPORT_SYMBOL(kmem_cache_alloc); +#ifdef CONFIG_KMEMTRACE +void *kmem_cache_alloc_notrace(struct kmem_cache *s, gfp_t gfpflags) +{ + return slab_alloc(s, gfpflags, -1, _RET_IP_); +} +EXPORT_SYMBOL(kmem_cache_alloc_notrace); +#endif + #ifdef CONFIG_NUMA void *kmem_cache_alloc_node(struct kmem_cache *s, gfp_t gfpflags, int node) { - return slab_alloc(s, gfpflags, node, _RET_IP_); + void *ret = slab_alloc(s, gfpflags, node, _RET_IP_); + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_CACHE, _RET_IP_, ret, + s->objsize, s->size, gfpflags, node); + + return ret; } EXPORT_SYMBOL(kmem_cache_alloc_node); #endif +#ifdef CONFIG_KMEMTRACE +void *kmem_cache_alloc_node_notrace(struct kmem_cache *s, + gfp_t gfpflags, + int node) +{ + return slab_alloc(s, gfpflags, node, _RET_IP_); +} +EXPORT_SYMBOL(kmem_cache_alloc_node_notrace); +#endif + /* * Slow patch handling. This may still be called frequently since objects * have a longer lifetime than the cpu slabs in most processing loads. @@ -1732,6 +1761,8 @@ void kmem_cache_free(struct kmem_cache *s, void *x) page = virt_to_head_page(x); slab_free(s, page, x, _RET_IP_); + + kmemtrace_mark_free(KMEMTRACE_TYPE_CACHE, _RET_IP_, x); } EXPORT_SYMBOL(kmem_cache_free); @@ -2650,6 +2681,7 @@ static struct kmem_cache *get_slab(size_t size, gfp_t flags) void *__kmalloc(size_t size, gfp_t flags) { struct kmem_cache *s; + void *ret; if (unlikely(size > PAGE_SIZE)) return kmalloc_large(size, flags); @@ -2659,7 +2691,12 @@ void *__kmalloc(size_t size, gfp_t flags) if (unlikely(ZERO_OR_NULL_PTR(s))) return s; - return slab_alloc(s, flags, -1, _RET_IP_); + ret = slab_alloc(s, flags, -1, _RET_IP_); + + kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, _RET_IP_, ret, + size, s->size, flags); + + return ret; } EXPORT_SYMBOL(__kmalloc); @@ -2678,16 +2715,30 @@ static void *kmalloc_large_node(size_t size, gfp_t flags, int node) void *__kmalloc_node(size_t size, gfp_t flags, int node) { struct kmem_cache *s; + void *ret; - if (unlikely(size > PAGE_SIZE)) - return kmalloc_large_node(size, flags, node); + if (unlikely(size > PAGE_SIZE)) { + ret = kmalloc_large_node(size, flags, node); + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, + _RET_IP_, ret, + size, PAGE_SIZE << get_order(size), + flags, node); + + return ret; + } s = get_slab(size, flags); if (unlikely(ZERO_OR_NULL_PTR(s))) return s; - return slab_alloc(s, flags, node, _RET_IP_); + ret = slab_alloc(s, flags, node, _RET_IP_); + + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, _RET_IP_, ret, + size, s->size, flags, node); + + return ret; } EXPORT_SYMBOL(__kmalloc_node); #endif @@ -2745,6 +2796,8 @@ void kfree(const void *x) return; } slab_free(page->slab, page, object, _RET_IP_); + + kmemtrace_mark_free(KMEMTRACE_TYPE_KMALLOC, _RET_IP_, x); } EXPORT_SYMBOL(kfree); From 4a80b24bb2ec66a5cb7fa5ff8335907f09288200 Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Tue, 19 Aug 2008 20:43:27 +0300 Subject: [PATCH 015/413] kmemtrace: Fix typos in documentation. Corrected the ABI description and the kmemtrace usage guide. Thanks to Randy Dunlap for noticing these errors. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- Documentation/ABI/testing/debugfs-kmemtrace | 2 +- Documentation/vm/kmemtrace.txt | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/Documentation/ABI/testing/debugfs-kmemtrace b/Documentation/ABI/testing/debugfs-kmemtrace index a5ff9a64c9ee..5e6a92a02d85 100644 --- a/Documentation/ABI/testing/debugfs-kmemtrace +++ b/Documentation/ABI/testing/debugfs-kmemtrace @@ -63,7 +63,7 @@ Adding new data to the packet (features) is done at the end of the mandatory data: Feature size (2 byte) Feature ID (1 byte) - Feature data (Feature size - 4 bytes) + Feature data (Feature size - 3 bytes) Users: diff --git a/Documentation/vm/kmemtrace.txt b/Documentation/vm/kmemtrace.txt index 75360b142ba4..f656cac3558c 100644 --- a/Documentation/vm/kmemtrace.txt +++ b/Documentation/vm/kmemtrace.txt @@ -61,7 +61,7 @@ III. Quick usage guide ====================== 1) Get a kernel that supports kmemtrace and build it accordingly (i.e. enable -CONFIG_KMEMTRACE and CONFIG_DEFAULT_ENABLED). +CONFIG_KMEMTRACE and CONFIG_KMEMTRACE_DEFAULT_ENABLED). 2) Get the userspace tool and build it: $ git-clone git://repo.or.cz/kmemtrace-user.git # current repository From 73cd6af0413225b0ada8b8881c3e0cfd26506dfa Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Tue, 19 Aug 2008 20:43:24 +0300 Subject: [PATCH 016/413] kmemtrace: Better alternative to "kmemtrace: fix printk format warnings". Fix the problem "kmemtrace: fix printk format warnings" attempted to fix, but resulted in marker-probe format mismatch warnings. Instead of carrying size_t into probes, we get rid of it by casting to unsigned long, just as we did with gfp_t. This way, we don't need to change marker format strings and we don't have to rely on other format specifiers like "%zu", making for consistent use of more generic data types (since there are no format specifiers for gfp_t, for example). Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- include/linux/kmemtrace.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/include/linux/kmemtrace.h b/include/linux/kmemtrace.h index 2c332010cb4e..5bea8ead6a6b 100644 --- a/include/linux/kmemtrace.h +++ b/include/linux/kmemtrace.h @@ -33,7 +33,8 @@ static inline void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, trace_mark(kmemtrace_alloc, "type_id %d call_site %lu ptr %lu " "bytes_req %lu bytes_alloc %lu gfp_flags %lu node %d", type_id, call_site, (unsigned long) ptr, - bytes_req, bytes_alloc, (unsigned long) gfp_flags, node); + (unsigned long) bytes_req, (unsigned long) bytes_alloc, + (unsigned long) gfp_flags, node); } static inline void kmemtrace_mark_free(enum kmemtrace_type_id type_id, From 94b528d0564997d88e8bf166d8c9080536ad8bdc Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Sun, 24 Aug 2008 20:49:35 +0300 Subject: [PATCH 017/413] kmemtrace: SLUB hooks for caller-tracking functions. This patch adds kmemtrace hooks for __kmalloc_track_caller() and __kmalloc_node_track_caller(). Currently, they set the call site pointer to the value recieved as a parameter. (This could change if we implement stack trace exporting in kmemtrace.) Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- mm/slub.c | 19 +++++++++++++++++-- 1 file changed, 17 insertions(+), 2 deletions(-) diff --git a/mm/slub.c b/mm/slub.c index 4c48a0146afd..4ce61c86f581 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -3256,6 +3256,7 @@ static struct notifier_block __cpuinitdata slab_notifier = { void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, unsigned long caller) { struct kmem_cache *s; + void *ret; if (unlikely(size > PAGE_SIZE)) return kmalloc_large(size, gfpflags); @@ -3265,13 +3266,20 @@ void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, unsigned long caller) if (unlikely(ZERO_OR_NULL_PTR(s))) return s; - return slab_alloc(s, gfpflags, -1, caller); + ret = slab_alloc(s, gfpflags, -1, caller); + + /* Honor the call site pointer we recieved. */ + kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, (unsigned long) caller, + ret, size, s->size, gfpflags); + + return ret; } void *__kmalloc_node_track_caller(size_t size, gfp_t gfpflags, int node, unsigned long caller) { struct kmem_cache *s; + void *ret; if (unlikely(size > PAGE_SIZE)) return kmalloc_large_node(size, gfpflags, node); @@ -3281,7 +3289,14 @@ void *__kmalloc_node_track_caller(size_t size, gfp_t gfpflags, if (unlikely(ZERO_OR_NULL_PTR(s))) return s; - return slab_alloc(s, gfpflags, node, caller); + ret = slab_alloc(s, gfpflags, node, caller); + + /* Honor the call site pointer we recieved. */ + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, + (unsigned long) caller, ret, + size, s->size, gfpflags, node); + + return ret; } #ifdef CONFIG_SLUB_DEBUG From 2e67624c22321fa40ad3aa89c307c84bd679d9b2 Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Mon, 1 Sep 2008 10:11:54 +0300 Subject: [PATCH 018/413] kmemtrace: remove unnecessary casts Now that we use _RET_IP_ there's no need to cast 'caller' to unsigned long. Signed-off-by: Pekka Enberg --- mm/slub.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/mm/slub.c b/mm/slub.c index 4ce61c86f581..e27472368cc3 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -3269,8 +3269,8 @@ void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, unsigned long caller) ret = slab_alloc(s, gfpflags, -1, caller); /* Honor the call site pointer we recieved. */ - kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, (unsigned long) caller, - ret, size, s->size, gfpflags); + kmemtrace_mark_alloc(KMEMTRACE_TYPE_KMALLOC, caller, ret, size, + s->size, gfpflags); return ret; } @@ -3292,8 +3292,7 @@ void *__kmalloc_node_track_caller(size_t size, gfp_t gfpflags, ret = slab_alloc(s, gfpflags, node, caller); /* Honor the call site pointer we recieved. */ - kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, - (unsigned long) caller, ret, + kmemtrace_mark_alloc_node(KMEMTRACE_TYPE_KMALLOC, caller, ret, size, s->size, gfpflags, node); return ret; From faa97abe6a3673af268abb661c6b663252a911aa Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Fri, 10 Oct 2008 10:57:44 +0300 Subject: [PATCH 019/413] kmemtrace: allow kmemtrace to be enabled after boot The kmemtrace_init() function returns early if kmemtrace is disabled at boot causing kmemtrace_setup_late() to also bail out on NULL channel. This has the unfortunate side effect that none of the debugfs files needed to enable kmemtrace after boot are created. Cc: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- mm/kmemtrace.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/mm/kmemtrace.c b/mm/kmemtrace.c index 83ad1cc71a92..f7a49c077df2 100644 --- a/mm/kmemtrace.c +++ b/mm/kmemtrace.c @@ -307,29 +307,29 @@ early_param("kmemtrace.subbufs", kmemtrace_set_subbufs); void kmemtrace_init(void) { - if (!kmemtrace_enabled) - return; - if (!kmemtrace_n_subbufs) kmemtrace_n_subbufs = KMEMTRACE_DEF_N_SUBBUFS; kmemtrace_chan = relay_open(NULL, NULL, KMEMTRACE_SUBBUF_SIZE, kmemtrace_n_subbufs, &relay_callbacks, NULL); - if (unlikely(!kmemtrace_chan)) { + if (!kmemtrace_chan) { printk(KERN_ERR "kmemtrace: could not open relay channel.\n"); return; } - if (unlikely(kmemtrace_start_probes())) - goto probe_fail; + if (!kmemtrace_enabled) { + printk(KERN_INFO "kmemtrace: disabled. Pass " + "kemtrace.enable=yes as kernel parameter for " + "boot-time tracing."); + return; + } + if (kmemtrace_start_probes()) { + printk(KERN_ERR "kmemtrace: could not register marker probes!\n"); + kmemtrace_cleanup(); + return; + } - printk(KERN_INFO "kmemtrace: early init successful.\n"); - - return; - -probe_fail: - printk(KERN_ERR "kmemtrace: could not register marker probes!\n"); - kmemtrace_cleanup(); + printk(KERN_INFO "kmemtrace: enabled.\n"); } From bf6803d6fd654d9a73cd90308b5225d78655d027 Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Fri, 10 Oct 2008 11:02:59 +0300 Subject: [PATCH 020/413] kmemtrace: remove config option for enabling tracing at boot Users can pass kmemtrace.enabled=yes as a kernel parameter to enable kmemtrace at boot so remove the useless CONFIG_KMEMTRACE_DEFAULT_ENABLED config option. Cc: Eduard - Gabriel Munteanu Signed-off-by: Pekka Enberg --- Documentation/vm/kmemtrace.txt | 2 +- lib/Kconfig.debug | 8 -------- mm/kmemtrace.c | 8 +++----- 3 files changed, 4 insertions(+), 14 deletions(-) diff --git a/Documentation/vm/kmemtrace.txt b/Documentation/vm/kmemtrace.txt index f656cac3558c..a956d9b7f943 100644 --- a/Documentation/vm/kmemtrace.txt +++ b/Documentation/vm/kmemtrace.txt @@ -61,7 +61,7 @@ III. Quick usage guide ====================== 1) Get a kernel that supports kmemtrace and build it accordingly (i.e. enable -CONFIG_KMEMTRACE and CONFIG_KMEMTRACE_DEFAULT_ENABLED). +CONFIG_KMEMTRACE). 2) Get the userspace tool and build it: $ git-clone git://repo.or.cz/kmemtrace-user.git # current repository diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index 78d669b461d2..b5417e23ba94 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -823,14 +823,6 @@ config KMEMTRACE If unsure, say N. -config KMEMTRACE_DEFAULT_ENABLED - bool "Enabled by default at boot" - depends on KMEMTRACE - help - Say Y here to enable kmemtrace at boot-time by default. Whatever - the choice, the behavior can be overridden by a kernel parameter, - as described in documentation. - menuconfig BUILD_DOCSRC bool "Build targets in Documentation/ tree" depends on HEADERS_CHECK diff --git a/mm/kmemtrace.c b/mm/kmemtrace.c index f7a49c077df2..f7704f52c4b1 100644 --- a/mm/kmemtrace.c +++ b/mm/kmemtrace.c @@ -19,11 +19,9 @@ static struct rchan *kmemtrace_chan; static u32 kmemtrace_buf_overruns; static unsigned int kmemtrace_n_subbufs; -#ifdef CONFIG_KMEMTRACE_DEFAULT_ENABLED -static unsigned int kmemtrace_enabled = 1; -#else -static unsigned int kmemtrace_enabled = 0; -#endif + +/* disabled by default */ +static unsigned int kmemtrace_enabled; /* * The sequence number is used for reordering kmemtrace packets From a4900437f3d76761a1646cd90254ccb01714a9ed Mon Sep 17 00:00:00 2001 From: Vegard Nossum Date: Thu, 6 Nov 2008 17:42:18 +0100 Subject: [PATCH 021/413] kmemtrace: add missing newline This was causing artifacts in my dmesg. Acked-by: Eduard - Gabriel Munteanu Signed-off-by: Vegard Nossum Signed-off-by: Pekka Enberg --- mm/kmemtrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/kmemtrace.c b/mm/kmemtrace.c index f7704f52c4b1..2a70a805027c 100644 --- a/mm/kmemtrace.c +++ b/mm/kmemtrace.c @@ -319,7 +319,7 @@ void kmemtrace_init(void) if (!kmemtrace_enabled) { printk(KERN_INFO "kmemtrace: disabled. Pass " "kemtrace.enable=yes as kernel parameter for " - "boot-time tracing."); + "boot-time tracing.\n"); return; } if (kmemtrace_start_probes()) { From a103e2ab7377dbbef2506be59c49a3f2ae10b60b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 29 Dec 2008 15:07:47 +0100 Subject: [PATCH 022/413] tracing/selftest: remove TRACE_CONT reference Impact: build fix TRACE_CONT is gone - fix up the self-test too. Signed-off-by: Ingo Molnar --- kernel/trace/trace_selftest.c | 1 - 1 file changed, 1 deletion(-) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 88c8eb70f54a..5013812578b1 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -9,7 +9,6 @@ static inline int trace_valid_entry(struct trace_entry *entry) case TRACE_FN: case TRACE_CTX: case TRACE_WAKE: - case TRACE_CONT: case TRACE_STACK: case TRACE_PRINT: case TRACE_SPECIAL: From 2a38b1c4f123afa4579cb1d29380b337e9195cbb Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 29 Dec 2008 15:29:04 +0100 Subject: [PATCH 023/413] kmemtrace: move #include lines Impact: avoid conflicts with kmemcheck kmemcheck modifies the same area of slab.c and slub.c - move the include lines up a bit. Signed-off-by: Ingo Molnar --- mm/slab.c | 2 +- mm/slub.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/mm/slab.c b/mm/slab.c index b6d9b8cdefa9..bcf08ea88380 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -102,6 +102,7 @@ #include #include #include +#include #include #include #include @@ -112,7 +113,6 @@ #include #include #include -#include #include #include diff --git a/mm/slub.c b/mm/slub.c index 4cd7bfd2ab2c..cc4001fee7ac 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -16,6 +16,7 @@ #include #include #include +#include #include #include #include @@ -24,7 +25,6 @@ #include #include #include -#include /* * Lock order: From 7a51cffbd10886c0557677dd916c090097c691ef Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 29 Dec 2008 16:03:40 +0100 Subject: [PATCH 024/413] relayfs: replace BUG() with WARN_ON() in relay_late_setup_files() Impact: turn boot crash into boot warning This BUG() can trigger: [ 16.684131] initcall fail_page_alloc_debugfs+0x0/0xc1 returned 0 after 0 usecs [ 16.692035] calling kmemtrace_setup_late+0x0/0xd5 @ 1 [ 16.700087] relay_late_setup_files: CPU 1 has no buffer, it must have! [ 16.704044] ------------[ cut here ]------------ [ 16.708030] kernel BUG at kernel/relay.c:680! [ 16.708030] invalid opcode: 0000 [#1] SMP DEBUG_PAGEALLOC [ 16.708030] last sysfs file: [ 16.708030] [ 16.708030] Pid: 1, comm: swapper Not tainted (2.6.28-tip-03903-g9a39f58-dirty #13207) System Product Name [ 16.708030] EIP: 0060:[] EFLAGS: 00010246 CPU: 1 [ 16.708030] EIP is at relay_late_setup_files+0x8c/0x176 Reduce it to a more reportable WARN_ONCE(). Signed-off-by: Ingo Molnar --- kernel/relay.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/kernel/relay.c b/kernel/relay.c index 09ac2008f77b..d06450670c86 100644 --- a/kernel/relay.c +++ b/kernel/relay.c @@ -675,9 +675,7 @@ int relay_late_setup_files(struct rchan *chan, */ for_each_online_cpu(i) { if (unlikely(!chan->buf[i])) { - printk(KERN_ERR "relay_late_setup_files: CPU %u " - "has no buffer, it must have!\n", i); - BUG(); + WARN_ONCE(1, KERN_ERR "CPU has no buffer!\n"); err = -EINVAL; break; } From 36994e58a48fb8f9651c7dc845a6de298aba5bfc Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 29 Dec 2008 13:42:23 -0800 Subject: [PATCH 025/413] tracing/kmemtrace: normalize the raw tracer event to the unified tracing API Impact: new tracer plugin This patch adapts kmemtrace raw events tracing to the unified tracing API. To enable and use this tracer, just do the following: echo kmemtrace > /debugfs/tracing/current_tracer cat /debugfs/tracing/trace You will have the following output: # tracer: kmemtrace # # # ALLOC TYPE REQ GIVEN FLAGS POINTER NODE CALLER # FREE | | | | | | | | # | type_id 1 call_site 18446744071565527833 ptr 18446612134395152256 type_id 0 call_site 18446744071565585597 ptr 18446612134405955584 bytes_req 4096 bytes_alloc 4096 gfp_flags 208 node -1 type_id 1 call_site 18446744071565585534 ptr 18446612134405955584 type_id 0 call_site 18446744071565585597 ptr 18446612134405955584 bytes_req 4096 bytes_alloc 4096 gfp_flags 208 node -1 type_id 0 call_site 18446744071565636711 ptr 18446612134345164672 bytes_req 240 bytes_alloc 240 gfp_flags 208 node -1 type_id 1 call_site 18446744071565585534 ptr 18446612134405955584 type_id 0 call_site 18446744071565585597 ptr 18446612134405955584 bytes_req 4096 bytes_alloc 4096 gfp_flags 208 node -1 type_id 0 call_site 18446744071565636711 ptr 18446612134345164912 bytes_req 240 bytes_alloc 240 gfp_flags 208 node -1 type_id 1 call_site 18446744071565585534 ptr 18446612134405955584 type_id 0 call_site 18446744071565585597 ptr 18446612134405955584 bytes_req 4096 bytes_alloc 4096 gfp_flags 208 node -1 type_id 0 call_site 18446744071565636711 ptr 18446612134345165152 bytes_req 240 bytes_alloc 240 gfp_flags 208 node -1 type_id 0 call_site 18446744071566144042 ptr 18446612134346191680 bytes_req 1304 bytes_alloc 1312 gfp_flags 208 node -1 type_id 1 call_site 18446744071565585534 ptr 18446612134405955584 type_id 0 call_site 18446744071565585597 ptr 18446612134405955584 bytes_req 4096 bytes_alloc 4096 gfp_flags 208 node -1 type_id 1 call_site 18446744071565585534 ptr 18446612134405955584 That was to stay backward compatible with the format output produced in inux/tracepoint.h. This is the default ouput, but note that I tried something else. If you change an option: echo kmem_minimalistic > /debugfs/trace_options and then cat /debugfs/trace, you will have the following output: # tracer: kmemtrace # # # ALLOC TYPE REQ GIVEN FLAGS POINTER NODE CALLER # FREE | | | | | | | | # | - C 0xffff88007c088780 file_free_rcu + K 4096 4096 000000d0 0xffff88007cad6000 -1 getname - C 0xffff88007cad6000 putname + K 4096 4096 000000d0 0xffff88007cad6000 -1 getname + K 240 240 000000d0 0xffff8800790dc780 -1 d_alloc - C 0xffff88007cad6000 putname + K 4096 4096 000000d0 0xffff88007cad6000 -1 getname + K 240 240 000000d0 0xffff8800790dc870 -1 d_alloc - C 0xffff88007cad6000 putname + K 4096 4096 000000d0 0xffff88007cad6000 -1 getname + K 240 240 000000d0 0xffff8800790dc960 -1 d_alloc + K 1304 1312 000000d0 0xffff8800791d7340 -1 reiserfs_alloc_inode - C 0xffff88007cad6000 putname + K 4096 4096 000000d0 0xffff88007cad6000 -1 getname - C 0xffff88007cad6000 putname + K 992 1000 000000d0 0xffff880079045b58 -1 alloc_inode + K 768 1024 000080d0 0xffff88007c096400 -1 alloc_pipe_info + K 240 240 000000d0 0xffff8800790dca50 -1 d_alloc + K 272 320 000080d0 0xffff88007c088780 -1 get_empty_filp + K 272 320 000080d0 0xffff88007c088000 -1 get_empty_filp Yeah I shall confess kmem_minimalistic should be: kmem_alternative. Whatever, I find it more readable but this a personal opinion of course. We can drop it if you want. On the ALLOC/FREE column, + means an allocation and - a free. On the type column, you have K = kmalloc, C = cache, P = page I would like the flags to be GFP_* strings but that would not be easy to not break the column with strings.... About the node...it seems to always be -1. I don't know why but that shouldn't be difficult to find. I moved linux/tracepoint.h to trace/tracepoint.h as well. I think that would be more easy to find the tracer headers if they are all in their common directory. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- include/linux/slab_def.h | 2 +- include/linux/slub_def.h | 2 +- include/{linux => trace}/kmemtrace.h | 19 +- init/main.c | 2 +- kernel/trace/Kconfig | 22 ++ kernel/trace/Makefile | 1 + kernel/trace/kmemtrace.c | 343 +++++++++++++++++++++++++++ kernel/trace/trace.h | 25 ++ lib/Kconfig.debug | 20 -- mm/kmemtrace.c | 2 +- mm/slob.c | 2 +- mm/slub.c | 2 +- 12 files changed, 401 insertions(+), 41 deletions(-) rename include/{linux => trace}/kmemtrace.h (72%) create mode 100644 kernel/trace/kmemtrace.c diff --git a/include/linux/slab_def.h b/include/linux/slab_def.h index 7555ce99f6d2..455f9affea9a 100644 --- a/include/linux/slab_def.h +++ b/include/linux/slab_def.h @@ -14,7 +14,7 @@ #include /* kmalloc_sizes.h needs PAGE_SIZE */ #include /* kmalloc_sizes.h needs L1_CACHE_BYTES */ #include -#include +#include /* Size description struct for general caches. */ struct cache_sizes { diff --git a/include/linux/slub_def.h b/include/linux/slub_def.h index dc28432b5b9a..6b657f7dcb2b 100644 --- a/include/linux/slub_def.h +++ b/include/linux/slub_def.h @@ -10,7 +10,7 @@ #include #include #include -#include +#include enum stat_item { ALLOC_FASTPATH, /* Allocation from cpu slab */ diff --git a/include/linux/kmemtrace.h b/include/trace/kmemtrace.h similarity index 72% rename from include/linux/kmemtrace.h rename to include/trace/kmemtrace.h index 5bea8ead6a6b..ad8b7857855a 100644 --- a/include/linux/kmemtrace.h +++ b/include/trace/kmemtrace.h @@ -22,28 +22,17 @@ enum kmemtrace_type_id { extern void kmemtrace_init(void); -static inline void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, +extern void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, unsigned long call_site, const void *ptr, size_t bytes_req, size_t bytes_alloc, gfp_t gfp_flags, - int node) -{ - trace_mark(kmemtrace_alloc, "type_id %d call_site %lu ptr %lu " - "bytes_req %lu bytes_alloc %lu gfp_flags %lu node %d", - type_id, call_site, (unsigned long) ptr, - (unsigned long) bytes_req, (unsigned long) bytes_alloc, - (unsigned long) gfp_flags, node); -} + int node); -static inline void kmemtrace_mark_free(enum kmemtrace_type_id type_id, +extern void kmemtrace_mark_free(enum kmemtrace_type_id type_id, unsigned long call_site, - const void *ptr) -{ - trace_mark(kmemtrace_free, "type_id %d call_site %lu ptr %lu", - type_id, call_site, (unsigned long) ptr); -} + const void *ptr); #else /* CONFIG_KMEMTRACE */ diff --git a/init/main.c b/init/main.c index 9711586aa7c9..beca7aaddb22 100644 --- a/init/main.c +++ b/init/main.c @@ -70,7 +70,7 @@ #include #include #include -#include +#include #ifdef CONFIG_X86_LOCAL_APIC #include diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index e2a4ff6fc3a6..27fb74b06b3c 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -264,6 +264,28 @@ config HW_BRANCH_TRACER This tracer records all branches on the system in a circular buffer giving access to the last N branches for each cpu. +config KMEMTRACE + bool "Trace SLAB allocations" + select TRACING + depends on RELAY + help + kmemtrace provides tracing for slab allocator functions, such as + kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected + data is then fed to the userspace application in order to analyse + allocation hotspots, internal fragmentation and so on, making it + possible to see how well an allocator performs, as well as debug + and profile kernel code. + + This requires an userspace application to use. See + Documentation/vm/kmemtrace.txt for more information. + + Saying Y will make the kernel somewhat larger and slower. However, + if you disable kmemtrace at run-time or boot-time, the performance + impact is minimal (depending on the arch the kernel is built for). + + If unsure, say N. + + config DYNAMIC_FTRACE bool "enable/disable ftrace tracepoints dynamically" depends on FUNCTION_TRACER diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 349d5a93653f..513dc86b5dfa 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -33,5 +33,6 @@ obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += trace_functions_graph.o obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o obj-$(CONFIG_HW_BRANCH_TRACER) += trace_hw_branches.o obj-$(CONFIG_POWER_TRACER) += trace_power.o +obj-$(CONFIG_KMEMTRACE) += kmemtrace.o libftrace-y := ftrace.o diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c new file mode 100644 index 000000000000..d69cbe3c2a4b --- /dev/null +++ b/kernel/trace/kmemtrace.c @@ -0,0 +1,343 @@ +/* + * Memory allocator tracing + * + * Copyright (C) 2008 Eduard - Gabriel Munteanu + * Copyright (C) 2008 Pekka Enberg + * Copyright (C) 2008 Frederic Weisbecker + */ + +#include +#include +#include +#include +#include + +#include "trace.h" +#include "trace_output.h" + +/* Select an alternative, minimalistic output than the original one */ +#define TRACE_KMEM_OPT_MINIMAL 0x1 + +static struct tracer_opt kmem_opts[] = { + /* Default disable the minimalistic output */ + { TRACER_OPT(kmem_minimalistic, TRACE_KMEM_OPT_MINIMAL) }, + { } +}; + +static struct tracer_flags kmem_tracer_flags = { + .val = 0, + .opts = kmem_opts +}; + + +static bool kmem_tracing_enabled __read_mostly; +static struct trace_array *kmemtrace_array; + +static int kmem_trace_init(struct trace_array *tr) +{ + int cpu; + kmemtrace_array = tr; + + for_each_cpu_mask(cpu, cpu_possible_map) + tracing_reset(tr, cpu); + + kmem_tracing_enabled = true; + + return 0; +} + +static void kmem_trace_reset(struct trace_array *tr) +{ + kmem_tracing_enabled = false; +} + +static void kmemtrace_headers(struct seq_file *s) +{ + /* Don't need headers for the original kmemtrace output */ + if (!(kmem_tracer_flags.val & TRACE_KMEM_OPT_MINIMAL)) + return; + + seq_printf(s, "#\n"); + seq_printf(s, "# ALLOC TYPE REQ GIVEN FLAGS " + " POINTER NODE CALLER\n"); + seq_printf(s, "# FREE | | | | " + " | | | |\n"); + seq_printf(s, "# |\n\n"); +} + +/* + * The two following functions give the original output from kmemtrace, + * or something close to....perhaps they need some missing things + */ +static enum print_line_t +kmemtrace_print_alloc_original(struct trace_iterator *iter, + struct kmemtrace_alloc_entry *entry) +{ + struct trace_seq *s = &iter->seq; + int ret; + + /* Taken from the old linux/kmemtrace.h */ + ret = trace_seq_printf(s, "type_id %d call_site %lu ptr %lu " + "bytes_req %lu bytes_alloc %lu gfp_flags %lu node %d\n", + entry->type_id, entry->call_site, (unsigned long) entry->ptr, + (unsigned long) entry->bytes_req, (unsigned long) entry->bytes_alloc, + (unsigned long) entry->gfp_flags, entry->node); + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t +kmemtrace_print_free_original(struct trace_iterator *iter, + struct kmemtrace_free_entry *entry) +{ + struct trace_seq *s = &iter->seq; + int ret; + + /* Taken from the old linux/kmemtrace.h */ + ret = trace_seq_printf(s, "type_id %d call_site %lu ptr %lu\n", + entry->type_id, entry->call_site, (unsigned long) entry->ptr); + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + + +/* The two other following provide a more minimalistic output */ +static enum print_line_t +kmemtrace_print_alloc_compress(struct trace_iterator *iter, + struct kmemtrace_alloc_entry *entry) +{ + struct trace_seq *s = &iter->seq; + int ret; + + /* Alloc entry */ + ret = trace_seq_printf(s, " + "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Type */ + switch (entry->type_id) { + case KMEMTRACE_TYPE_KMALLOC: + ret = trace_seq_printf(s, "K "); + break; + case KMEMTRACE_TYPE_CACHE: + ret = trace_seq_printf(s, "C "); + break; + case KMEMTRACE_TYPE_PAGES: + ret = trace_seq_printf(s, "P "); + break; + default: + ret = trace_seq_printf(s, "? "); + } + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Requested */ + ret = trace_seq_printf(s, "%4d ", entry->bytes_req); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Allocated */ + ret = trace_seq_printf(s, "%4d ", entry->bytes_alloc); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Flags + * TODO: would be better to see the name of the GFP flag names + */ + ret = trace_seq_printf(s, "%08x ", entry->gfp_flags); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Pointer to allocated */ + ret = trace_seq_printf(s, "0x%tx ", (ptrdiff_t)entry->ptr); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Node */ + ret = trace_seq_printf(s, "%4d ", entry->node); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Call site */ + ret = seq_print_ip_sym(s, entry->call_site, 0); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + if (!trace_seq_printf(s, "\n")) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t +kmemtrace_print_free_compress(struct trace_iterator *iter, + struct kmemtrace_free_entry *entry) +{ + struct trace_seq *s = &iter->seq; + int ret; + + /* Free entry */ + ret = trace_seq_printf(s, " - "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Type */ + switch (entry->type_id) { + case KMEMTRACE_TYPE_KMALLOC: + ret = trace_seq_printf(s, "K "); + break; + case KMEMTRACE_TYPE_CACHE: + ret = trace_seq_printf(s, "C "); + break; + case KMEMTRACE_TYPE_PAGES: + ret = trace_seq_printf(s, "P "); + break; + default: + ret = trace_seq_printf(s, "? "); + } + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Skip requested/allocated/flags */ + ret = trace_seq_printf(s, " "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Pointer to allocated */ + ret = trace_seq_printf(s, "0x%tx ", (ptrdiff_t)entry->ptr); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Skip node */ + ret = trace_seq_printf(s, " "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Call site */ + ret = seq_print_ip_sym(s, entry->call_site, 0); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + if (!trace_seq_printf(s, "\n")) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +static enum print_line_t kmemtrace_print_line(struct trace_iterator *iter) +{ + struct trace_entry *entry = iter->ent; + + switch (entry->type) { + case TRACE_KMEM_ALLOC: { + struct kmemtrace_alloc_entry *field; + trace_assign_type(field, entry); + if (kmem_tracer_flags.val & TRACE_KMEM_OPT_MINIMAL) + return kmemtrace_print_alloc_compress(iter, field); + else + return kmemtrace_print_alloc_original(iter, field); + } + + case TRACE_KMEM_FREE: { + struct kmemtrace_free_entry *field; + trace_assign_type(field, entry); + if (kmem_tracer_flags.val & TRACE_KMEM_OPT_MINIMAL) + return kmemtrace_print_free_compress(iter, field); + else + return kmemtrace_print_free_original(iter, field); + } + + default: + return TRACE_TYPE_UNHANDLED; + } +} + +/* Trace allocations */ +void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr, + size_t bytes_req, + size_t bytes_alloc, + gfp_t gfp_flags, + int node) +{ + struct ring_buffer_event *event; + struct kmemtrace_alloc_entry *entry; + struct trace_array *tr = kmemtrace_array; + unsigned long irq_flags; + + if (!kmem_tracing_enabled) + return; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), + &irq_flags); + if (!event) + return; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, 0, 0); + + entry->ent.type = TRACE_KMEM_ALLOC; + entry->call_site = call_site; + entry->ptr = ptr; + entry->bytes_req = bytes_req; + entry->bytes_alloc = bytes_alloc; + entry->gfp_flags = gfp_flags; + entry->node = node; + + ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + + trace_wake_up(); +} + +void kmemtrace_mark_free(enum kmemtrace_type_id type_id, + unsigned long call_site, + const void *ptr) +{ + struct ring_buffer_event *event; + struct kmemtrace_free_entry *entry; + struct trace_array *tr = kmemtrace_array; + unsigned long irq_flags; + + if (!kmem_tracing_enabled) + return; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), + &irq_flags); + if (!event) + return; + entry = ring_buffer_event_data(event); + tracing_generic_entry_update(&entry->ent, 0, 0); + + entry->ent.type = TRACE_KMEM_FREE; + entry->type_id = type_id; + entry->call_site = call_site; + entry->ptr = ptr; + + ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + + trace_wake_up(); +} + +static struct tracer kmem_tracer __read_mostly = { + .name = "kmemtrace", + .init = kmem_trace_init, + .reset = kmem_trace_reset, + .print_line = kmemtrace_print_line, + .print_header = kmemtrace_headers, + .flags = &kmem_tracer_flags +}; + +static int __init init_kmem_tracer(void) +{ + return register_tracer(&kmem_tracer); +} + +device_initcall(init_kmem_tracer); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index cc7a4f864036..534505bb39b0 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -9,6 +9,7 @@ #include #include #include +#include enum trace_type { __TRACE_FIRST_TYPE = 0, @@ -29,6 +30,8 @@ enum trace_type { TRACE_GRAPH_ENT, TRACE_USER_STACK, TRACE_HW_BRANCHES, + TRACE_KMEM_ALLOC, + TRACE_KMEM_FREE, TRACE_POWER, __TRACE_LAST_TYPE @@ -170,6 +173,24 @@ struct trace_power { struct power_trace state_data; }; +struct kmemtrace_alloc_entry { + struct trace_entry ent; + enum kmemtrace_type_id type_id; + unsigned long call_site; + const void *ptr; + size_t bytes_req; + size_t bytes_alloc; + gfp_t gfp_flags; + int node; +}; + +struct kmemtrace_free_entry { + struct trace_entry ent; + enum kmemtrace_type_id type_id; + unsigned long call_site; + const void *ptr; +}; + /* * trace_flag_type is an enumeration that holds different * states when a trace occurs. These are: @@ -280,6 +301,10 @@ extern void __ftrace_bad_type(void); TRACE_GRAPH_RET); \ IF_ASSIGN(var, ent, struct hw_branch_entry, TRACE_HW_BRANCHES);\ IF_ASSIGN(var, ent, struct trace_power, TRACE_POWER); \ + IF_ASSIGN(var, ent, struct kmemtrace_alloc_entry, \ + TRACE_KMEM_ALLOC); \ + IF_ASSIGN(var, ent, struct kmemtrace_free_entry, \ + TRACE_KMEM_FREE); \ __ftrace_bad_type(); \ } while (0) diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug index b5417e23ba94..b0f239e443bc 100644 --- a/lib/Kconfig.debug +++ b/lib/Kconfig.debug @@ -803,26 +803,6 @@ config FIREWIRE_OHCI_REMOTE_DMA If unsure, say N. -config KMEMTRACE - bool "Kernel memory tracer (kmemtrace)" - depends on RELAY && DEBUG_FS && MARKERS - help - kmemtrace provides tracing for slab allocator functions, such as - kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected - data is then fed to the userspace application in order to analyse - allocation hotspots, internal fragmentation and so on, making it - possible to see how well an allocator performs, as well as debug - and profile kernel code. - - This requires an userspace application to use. See - Documentation/vm/kmemtrace.txt for more information. - - Saying Y will make the kernel somewhat larger and slower. However, - if you disable kmemtrace at run-time or boot-time, the performance - impact is minimal (depending on the arch the kernel is built for). - - If unsure, say N. - menuconfig BUILD_DOCSRC bool "Build targets in Documentation/ tree" depends on HEADERS_CHECK diff --git a/mm/kmemtrace.c b/mm/kmemtrace.c index 2a70a805027c..0573b5080cc4 100644 --- a/mm/kmemtrace.c +++ b/mm/kmemtrace.c @@ -10,7 +10,7 @@ #include #include #include -#include +#include #define KMEMTRACE_SUBBUF_SIZE 524288 #define KMEMTRACE_DEF_N_SUBBUFS 20 diff --git a/mm/slob.c b/mm/slob.c index 0f1a49f40690..4d1c0fc33b6b 100644 --- a/mm/slob.c +++ b/mm/slob.c @@ -65,7 +65,7 @@ #include #include #include -#include +#include #include /* diff --git a/mm/slub.c b/mm/slub.c index cc4001fee7ac..7bf8cf8ec082 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -16,7 +16,7 @@ #include #include #include -#include +#include #include #include #include From b6ab4afee4ed56d0f69df59485585cff828c327d Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 30 Dec 2008 09:41:04 +0100 Subject: [PATCH 026/413] tracing, kvm: change MARKERS to select instead of depends on Impact: build fix fix: kernel/trace/Kconfig:42:error: found recursive dependency: TRACING -> TRACEPOINTS -> MARKERS -> KVM_TRACE -> RELAY -> KMEMTRACE -> TRACING markers is a facility that should be selected - not depended on by an interactive Kconfig entry. Signed-off-by: Ingo Molnar --- arch/x86/kvm/Kconfig | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/x86/kvm/Kconfig b/arch/x86/kvm/Kconfig index b81125f0bdee..c7da3683f4c5 100644 --- a/arch/x86/kvm/Kconfig +++ b/arch/x86/kvm/Kconfig @@ -55,7 +55,8 @@ config KVM_AMD config KVM_TRACE bool "KVM trace support" - depends on KVM && MARKERS && SYSFS + depends on KVM && SYSFS + select MARKERS select RELAY select DEBUG_FS default n From 3fd4bc015ef879a7d2b955ce97fb125e3a51ba7e Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 30 Dec 2008 12:07:27 +0100 Subject: [PATCH 027/413] tracing/kmemtrace: export kmemtrace_mark_alloc_node() / kmemtrace_mark_free() Impact: build fix Also fix up Kconfig dependencies and include files. Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 3 ++- kernel/trace/kmemtrace.c | 2 ++ mm/slab.c | 2 +- 3 files changed, 5 insertions(+), 2 deletions(-) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 27fb74b06b3c..cc9f91e7daf4 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -267,7 +267,8 @@ config HW_BRANCH_TRACER config KMEMTRACE bool "Trace SLAB allocations" select TRACING - depends on RELAY + select MARKERS + select RELAY help kmemtrace provides tracing for slab allocator functions, such as kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c index d69cbe3c2a4b..2bfdcd326226 100644 --- a/kernel/trace/kmemtrace.c +++ b/kernel/trace/kmemtrace.c @@ -296,6 +296,7 @@ void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, trace_wake_up(); } +EXPORT_SYMBOL(kmemtrace_mark_alloc_node); void kmemtrace_mark_free(enum kmemtrace_type_id type_id, unsigned long call_site, @@ -325,6 +326,7 @@ void kmemtrace_mark_free(enum kmemtrace_type_id type_id, trace_wake_up(); } +EXPORT_SYMBOL(kmemtrace_mark_free); static struct tracer kmem_tracer __read_mostly = { .name = "kmemtrace", diff --git a/mm/slab.c b/mm/slab.c index bcf08ea88380..7f72bb386a09 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -102,7 +102,7 @@ #include #include #include -#include +#include #include #include #include From f09eac9034a4502cce558b0ec4bf7d422b8b355b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 31 Dec 2008 09:43:46 +0100 Subject: [PATCH 028/413] tracing/kmemtrace: fix typo Impact: build fix Signed-off-by: Ingo Molnar --- mm/slab.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/slab.c b/mm/slab.c index 83075f36df7b..c65c52dc78d4 100644 --- a/mm/slab.c +++ b/mm/slab.c @@ -102,7 +102,7 @@ #include #include #include -#include +#include #include #include #include From 723cbe0775514853c22dc45005af59c360916af1 Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Mon, 5 Jan 2009 22:09:58 +0200 Subject: [PATCH 029/413] kmemtrace: Remove the relay version of kmemtrace Impact: cleanup kmemtrace now uses ftrace. This patch removes the relay version. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 2 - mm/Makefile | 1 - mm/kmemtrace.c | 333 ------------------------------------------- 3 files changed, 336 deletions(-) delete mode 100644 mm/kmemtrace.c diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index cc9f91e7daf4..1c0b7504cab3 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -267,8 +267,6 @@ config HW_BRANCH_TRACER config KMEMTRACE bool "Trace SLAB allocations" select TRACING - select MARKERS - select RELAY help kmemtrace provides tracing for slab allocator functions, such as kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected diff --git a/mm/Makefile b/mm/Makefile index c92e8af13206..51c27709cc7c 100644 --- a/mm/Makefile +++ b/mm/Makefile @@ -35,4 +35,3 @@ obj-$(CONFIG_MIGRATION) += migrate.o obj-$(CONFIG_SMP) += allocpercpu.o obj-$(CONFIG_QUICKLIST) += quicklist.o obj-$(CONFIG_CGROUP_MEM_RES_CTLR) += memcontrol.o page_cgroup.o -obj-$(CONFIG_KMEMTRACE) += kmemtrace.o diff --git a/mm/kmemtrace.c b/mm/kmemtrace.c deleted file mode 100644 index 0573b5080cc4..000000000000 --- a/mm/kmemtrace.c +++ /dev/null @@ -1,333 +0,0 @@ -/* - * Copyright (C) 2008 Pekka Enberg, Eduard - Gabriel Munteanu - * - * This file is released under GPL version 2. - */ - -#include -#include -#include -#include -#include -#include -#include - -#define KMEMTRACE_SUBBUF_SIZE 524288 -#define KMEMTRACE_DEF_N_SUBBUFS 20 - -static struct rchan *kmemtrace_chan; -static u32 kmemtrace_buf_overruns; - -static unsigned int kmemtrace_n_subbufs; - -/* disabled by default */ -static unsigned int kmemtrace_enabled; - -/* - * The sequence number is used for reordering kmemtrace packets - * in userspace, since they are logged as per-CPU data. - * - * atomic_t should always be a 32-bit signed integer. Wraparound is not - * likely to occur, but userspace can deal with it by expecting a certain - * sequence number in the next packet that will be read. - */ -static atomic_t kmemtrace_seq_num; - -#define KMEMTRACE_ABI_VERSION 1 - -static u32 kmemtrace_abi_version __read_mostly = KMEMTRACE_ABI_VERSION; - -enum kmemtrace_event_id { - KMEMTRACE_EVENT_ALLOC = 0, - KMEMTRACE_EVENT_FREE, -}; - -struct kmemtrace_event { - u8 event_id; - u8 type_id; - u16 event_size; - s32 seq_num; - u64 call_site; - u64 ptr; -} __attribute__ ((__packed__)); - -struct kmemtrace_stats_alloc { - u64 bytes_req; - u64 bytes_alloc; - u32 gfp_flags; - s32 numa_node; -} __attribute__ ((__packed__)); - -static void kmemtrace_probe_alloc(void *probe_data, void *call_data, - const char *format, va_list *args) -{ - unsigned long flags; - struct kmemtrace_event *ev; - struct kmemtrace_stats_alloc *stats; - void *buf; - - local_irq_save(flags); - - buf = relay_reserve(kmemtrace_chan, - sizeof(struct kmemtrace_event) + - sizeof(struct kmemtrace_stats_alloc)); - if (!buf) - goto failed; - - /* - * Don't convert this to use structure initializers, - * C99 does not guarantee the rvalues evaluation order. - */ - - ev = buf; - ev->event_id = KMEMTRACE_EVENT_ALLOC; - ev->type_id = va_arg(*args, int); - ev->event_size = sizeof(struct kmemtrace_event) + - sizeof(struct kmemtrace_stats_alloc); - ev->seq_num = atomic_add_return(1, &kmemtrace_seq_num); - ev->call_site = va_arg(*args, unsigned long); - ev->ptr = va_arg(*args, unsigned long); - - stats = buf + sizeof(struct kmemtrace_event); - stats->bytes_req = va_arg(*args, unsigned long); - stats->bytes_alloc = va_arg(*args, unsigned long); - stats->gfp_flags = va_arg(*args, unsigned long); - stats->numa_node = va_arg(*args, int); - -failed: - local_irq_restore(flags); -} - -static void kmemtrace_probe_free(void *probe_data, void *call_data, - const char *format, va_list *args) -{ - unsigned long flags; - struct kmemtrace_event *ev; - - local_irq_save(flags); - - ev = relay_reserve(kmemtrace_chan, sizeof(struct kmemtrace_event)); - if (!ev) - goto failed; - - /* - * Don't convert this to use structure initializers, - * C99 does not guarantee the rvalues evaluation order. - */ - ev->event_id = KMEMTRACE_EVENT_FREE; - ev->type_id = va_arg(*args, int); - ev->event_size = sizeof(struct kmemtrace_event); - ev->seq_num = atomic_add_return(1, &kmemtrace_seq_num); - ev->call_site = va_arg(*args, unsigned long); - ev->ptr = va_arg(*args, unsigned long); - -failed: - local_irq_restore(flags); -} - -static struct dentry * -kmemtrace_create_buf_file(const char *filename, struct dentry *parent, - int mode, struct rchan_buf *buf, int *is_global) -{ - return debugfs_create_file(filename, mode, parent, buf, - &relay_file_operations); -} - -static int kmemtrace_remove_buf_file(struct dentry *dentry) -{ - debugfs_remove(dentry); - - return 0; -} - -static int kmemtrace_subbuf_start(struct rchan_buf *buf, - void *subbuf, - void *prev_subbuf, - size_t prev_padding) -{ - if (relay_buf_full(buf)) { - /* - * We know it's not SMP-safe, but neither - * debugfs_create_u32() is. - */ - kmemtrace_buf_overruns++; - return 0; - } - - return 1; -} - -static struct rchan_callbacks relay_callbacks = { - .create_buf_file = kmemtrace_create_buf_file, - .remove_buf_file = kmemtrace_remove_buf_file, - .subbuf_start = kmemtrace_subbuf_start, -}; - -static struct dentry *kmemtrace_dir; -static struct dentry *kmemtrace_overruns_dentry; -static struct dentry *kmemtrace_abi_version_dentry; - -static struct dentry *kmemtrace_enabled_dentry; - -static int kmemtrace_start_probes(void) -{ - int err; - - err = marker_probe_register("kmemtrace_alloc", "type_id %d " - "call_site %lu ptr %lu " - "bytes_req %lu bytes_alloc %lu " - "gfp_flags %lu node %d", - kmemtrace_probe_alloc, NULL); - if (err) - return err; - err = marker_probe_register("kmemtrace_free", "type_id %d " - "call_site %lu ptr %lu", - kmemtrace_probe_free, NULL); - - return err; -} - -static void kmemtrace_stop_probes(void) -{ - marker_probe_unregister("kmemtrace_alloc", - kmemtrace_probe_alloc, NULL); - marker_probe_unregister("kmemtrace_free", - kmemtrace_probe_free, NULL); -} - -static int kmemtrace_enabled_get(void *data, u64 *val) -{ - *val = *((int *) data); - - return 0; -} - -static int kmemtrace_enabled_set(void *data, u64 val) -{ - u64 old_val = kmemtrace_enabled; - - *((int *) data) = !!val; - - if (old_val == val) - return 0; - if (val) - kmemtrace_start_probes(); - else - kmemtrace_stop_probes(); - - return 0; -} - -DEFINE_SIMPLE_ATTRIBUTE(kmemtrace_enabled_fops, - kmemtrace_enabled_get, - kmemtrace_enabled_set, "%llu\n"); - -static void kmemtrace_cleanup(void) -{ - if (kmemtrace_enabled_dentry) - debugfs_remove(kmemtrace_enabled_dentry); - - kmemtrace_stop_probes(); - - if (kmemtrace_abi_version_dentry) - debugfs_remove(kmemtrace_abi_version_dentry); - if (kmemtrace_overruns_dentry) - debugfs_remove(kmemtrace_overruns_dentry); - - relay_close(kmemtrace_chan); - kmemtrace_chan = NULL; - - if (kmemtrace_dir) - debugfs_remove(kmemtrace_dir); -} - -static int __init kmemtrace_setup_late(void) -{ - if (!kmemtrace_chan) - goto failed; - - kmemtrace_dir = debugfs_create_dir("kmemtrace", NULL); - if (!kmemtrace_dir) - goto cleanup; - - kmemtrace_abi_version_dentry = - debugfs_create_u32("abi_version", S_IRUSR, - kmemtrace_dir, &kmemtrace_abi_version); - kmemtrace_overruns_dentry = - debugfs_create_u32("total_overruns", S_IRUSR, - kmemtrace_dir, &kmemtrace_buf_overruns); - if (!kmemtrace_overruns_dentry || !kmemtrace_abi_version_dentry) - goto cleanup; - - kmemtrace_enabled_dentry = - debugfs_create_file("enabled", S_IRUSR | S_IWUSR, - kmemtrace_dir, &kmemtrace_enabled, - &kmemtrace_enabled_fops); - if (!kmemtrace_enabled_dentry) - goto cleanup; - - if (relay_late_setup_files(kmemtrace_chan, "cpu", kmemtrace_dir)) - goto cleanup; - - printk(KERN_INFO "kmemtrace: fully up.\n"); - - return 0; - -cleanup: - kmemtrace_cleanup(); -failed: - return 1; -} -late_initcall(kmemtrace_setup_late); - -static int __init kmemtrace_set_boot_enabled(char *str) -{ - if (!str) - return -EINVAL; - - if (!strcmp(str, "yes")) - kmemtrace_enabled = 1; - else if (!strcmp(str, "no")) - kmemtrace_enabled = 0; - else - return -EINVAL; - - return 0; -} -early_param("kmemtrace.enable", kmemtrace_set_boot_enabled); - -static int __init kmemtrace_set_subbufs(char *str) -{ - get_option(&str, &kmemtrace_n_subbufs); - return 0; -} -early_param("kmemtrace.subbufs", kmemtrace_set_subbufs); - -void kmemtrace_init(void) -{ - if (!kmemtrace_n_subbufs) - kmemtrace_n_subbufs = KMEMTRACE_DEF_N_SUBBUFS; - - kmemtrace_chan = relay_open(NULL, NULL, KMEMTRACE_SUBBUF_SIZE, - kmemtrace_n_subbufs, &relay_callbacks, - NULL); - if (!kmemtrace_chan) { - printk(KERN_ERR "kmemtrace: could not open relay channel.\n"); - return; - } - - if (!kmemtrace_enabled) { - printk(KERN_INFO "kmemtrace: disabled. Pass " - "kemtrace.enable=yes as kernel parameter for " - "boot-time tracing.\n"); - return; - } - if (kmemtrace_start_probes()) { - printk(KERN_ERR "kmemtrace: could not register marker probes!\n"); - kmemtrace_cleanup(); - return; - } - - printk(KERN_INFO "kmemtrace: enabled.\n"); -} - From 3e80680208ba6ce9635ca7c21ad0019442ea166a Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 6 Jan 2009 10:16:35 +0100 Subject: [PATCH 030/413] kmemtrace: add kmemtrace_init() Impact: build fix leftover from the relayfs version - but we want to keep it because this call is the earliest opportunity when we can start kmemtrace tracing. (after kmem_cache_init()). Signed-off-by: Ingo Molnar --- kernel/trace/kmemtrace.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c index 2bfdcd326226..faaa5ae7e75a 100644 --- a/kernel/trace/kmemtrace.c +++ b/kernel/trace/kmemtrace.c @@ -337,6 +337,11 @@ static struct tracer kmem_tracer __read_mostly = { .flags = &kmem_tracer_flags }; +void kmemtrace_init(void) +{ + /* earliest opportunity to start kmem tracing */ +} + static int __init init_kmem_tracer(void) { return register_tracer(&kmem_tracer); From 431aa3fbf5bbe3be79809c7e603c2ed2ac64b015 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 6 Jan 2009 12:43:01 -0500 Subject: [PATCH 031/413] ftrace: convert unsigned index to signed Impact: fix to unsigned compared to less than zero Roel Kluin pointed out that there is a compare of an unsigned number to less than zero. A previous clean up had the unsigned index set to -1 for certain cases, but never converted it to signed. Frederic Weisbecker noticed that another index is used to compare the above index to and it also needs to be converted to signed. [ Converted ftrace_page->index to int from unsigned long as Andrew Morton pointed out that there's no need for it to stay a long. ] Reported-by: Roel Kluin Reported-by: Frederic Weisbecker Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 2f32969c09df..9e54a6ccdb93 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -289,7 +289,7 @@ static DEFINE_MUTEX(ftrace_regex_lock); struct ftrace_page { struct ftrace_page *next; - unsigned long index; + int index; struct dyn_ftrace records[]; }; @@ -786,7 +786,7 @@ enum { struct ftrace_iterator { struct ftrace_page *pg; - unsigned idx; + int idx; unsigned flags; unsigned char buffer[FTRACE_BUFF_MAX+1]; unsigned buffer_idx; From ff288b274a9b383046fdbda4be3067daba4d5fe8 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 6 Jan 2009 21:33:30 +0100 Subject: [PATCH 032/413] tracing/ftrace: fix a memory leak in stat tracing Impact: fix memory leak This patch fixes a memory leak inside reset_stat_list(). The freeing loop iterated only once. Also turn the stat_list into a simple struct list_head, which simplify the code and avoid an unused static pointer. Reported-by: Steven Rostedt Signed-off-by: Frederic Weisbecker Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_stat.c | 39 +++++++++++++++------------------------ 1 file changed, 15 insertions(+), 24 deletions(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index 6f194a33a64a..4cb4ff27646d 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -21,7 +21,7 @@ struct trace_stat_list { void *stat; }; -static struct trace_stat_list stat_list; +static LIST_HEAD(stat_list); /* * This is a copy of the current tracer to avoid racy @@ -39,22 +39,12 @@ static DEFINE_MUTEX(stat_list_mutex); static void reset_stat_list(void) { - struct trace_stat_list *node; - struct list_head *next; + struct trace_stat_list *node, *next; - if (list_empty(&stat_list.list)) - return; - - node = list_entry(stat_list.list.next, struct trace_stat_list, list); - next = node->list.next; - - while (&node->list != next) { + list_for_each_entry_safe(node, next, &stat_list, list) kfree(node); - node = list_entry(next, struct trace_stat_list, list); - } - kfree(node); - INIT_LIST_HEAD(&stat_list.list); + INIT_LIST_HEAD(&stat_list); } void init_tracer_stat(struct tracer *trace) @@ -107,7 +97,7 @@ static int stat_seq_init(void) } INIT_LIST_HEAD(&new_entry->list); - list_add(&new_entry->list, &stat_list.list); + list_add(&new_entry->list, &stat_list); new_entry->stat = current_tracer.stat_start(); prev_stat = new_entry->stat; @@ -130,7 +120,7 @@ static int stat_seq_init(void) if (!new_entry->stat) break; - list_for_each_entry(iter_entry, &stat_list.list, list) { + list_for_each_entry(iter_entry, &stat_list, list) { /* Insertion with a descendent sorting */ if (current_tracer.stat_cmp(new_entry->stat, iter_entry->stat) > 0) { @@ -141,7 +131,7 @@ static int stat_seq_init(void) /* The current smaller value */ } else if (list_is_last(&iter_entry->list, - &stat_list.list)) { + &stat_list)) { list_add(&new_entry->list, &iter_entry->list); break; } @@ -162,7 +152,7 @@ static int stat_seq_init(void) static void *stat_seq_start(struct seq_file *s, loff_t *pos) { - struct trace_stat_list *l = (struct trace_stat_list *)s->private; + struct list_head *l = (struct list_head *)s->private; /* Prevent from tracer switch or stat_list modification */ mutex_lock(&stat_list_mutex); @@ -171,14 +161,14 @@ static void *stat_seq_start(struct seq_file *s, loff_t *pos) if (!*pos && current_tracer.stat_headers) current_tracer.stat_headers(s); - return seq_list_start(&l->list, *pos); + return seq_list_start(l, *pos); } static void *stat_seq_next(struct seq_file *s, void *p, loff_t *pos) { - struct trace_stat_list *l = (struct trace_stat_list *)s->private; + struct list_head *l = (struct list_head *)s->private; - return seq_list_next(p, &l->list, pos); + return seq_list_next(p, l, pos); } static void stat_seq_stop(struct seq_file *m, void *p) @@ -188,8 +178,10 @@ static void stat_seq_stop(struct seq_file *m, void *p) static int stat_seq_show(struct seq_file *s, void *v) { - struct trace_stat_list *l = list_entry(v, struct trace_stat_list, list); - return current_tracer.stat_show(s, l->stat); + struct trace_stat_list *entry = + list_entry(v, struct trace_stat_list, list); + + return current_tracer.stat_show(s, entry->stat); } static const struct seq_operations trace_stat_seq_ops = { @@ -237,7 +229,6 @@ static int __init tracing_stat_init(void) struct dentry *d_tracing; struct dentry *entry; - INIT_LIST_HEAD(&stat_list.list); d_tracing = tracing_init_dentry(); entry = debugfs_create_file("trace_stat", 0444, d_tracing, From e8a9cbf6ae620d9e5ba9cb42001c033287a284a3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 6 Jan 2009 22:02:35 -0500 Subject: [PATCH 033/413] trace: clean up funny line breaks in stat_seq_show Impact: clean up Andrew Morton pointed out that the entry assignment in stat_seq_show did not need to be done in the declaration, causing funny line breaks. This patch makes it a bit more pleasing on the eyes. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_stat.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index 4cb4ff27646d..f110ce9ce7fb 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -178,8 +178,9 @@ static void stat_seq_stop(struct seq_file *m, void *p) static int stat_seq_show(struct seq_file *s, void *v) { - struct trace_stat_list *entry = - list_entry(v, struct trace_stat_list, list); + struct trace_stat_list *entry; + + entry = list_entry(v, struct trace_stat_list, list); return current_tracer.stat_show(s, entry->stat); } From 34a148bf0911a4a1cae85f8ecba57affb4d76aee Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Fri, 9 Jan 2009 12:27:09 -0800 Subject: [PATCH 034/413] kernel/trace/ring_buffer.c: reduce inlining text data bss dec hex filename before: 11320 228 8 11556 2d24 kernel/trace/ring_buffer.o after: 10592 228 8 10828 2a4c kernel/trace/ring_buffer.o Also: free_page(0) is legal. Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 8b0daf0662ef..9542990f515b 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -133,7 +133,7 @@ enum { }; /* inline for ring buffer fast paths */ -static inline unsigned +static unsigned rb_event_length(struct ring_buffer_event *event) { unsigned length; @@ -179,7 +179,7 @@ unsigned ring_buffer_event_length(struct ring_buffer_event *event) EXPORT_SYMBOL_GPL(ring_buffer_event_length); /* inline for ring buffer fast paths */ -static inline void * +static void * rb_event_data(struct ring_buffer_event *event) { BUG_ON(event->type != RINGBUF_TYPE_DATA); @@ -229,10 +229,9 @@ static void rb_init_page(struct buffer_data_page *bpage) * Also stolen from mm/slob.c. Thanks to Mathieu Desnoyers for pointing * this issue out. */ -static inline void free_buffer_page(struct buffer_page *bpage) +static void free_buffer_page(struct buffer_page *bpage) { - if (bpage->page) - free_page((unsigned long)bpage->page); + free_page((unsigned long)bpage->page); kfree(bpage); } @@ -811,7 +810,7 @@ rb_event_index(struct ring_buffer_event *event) return (addr & ~PAGE_MASK) - (PAGE_SIZE - BUF_PAGE_SIZE); } -static inline int +static int rb_is_commit(struct ring_buffer_per_cpu *cpu_buffer, struct ring_buffer_event *event) { @@ -825,7 +824,7 @@ rb_is_commit(struct ring_buffer_per_cpu *cpu_buffer, rb_commit_index(cpu_buffer) == index; } -static inline void +static void rb_set_commit_event(struct ring_buffer_per_cpu *cpu_buffer, struct ring_buffer_event *event) { @@ -850,7 +849,7 @@ rb_set_commit_event(struct ring_buffer_per_cpu *cpu_buffer, local_set(&cpu_buffer->commit_page->page->commit, index); } -static inline void +static void rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer) { /* @@ -896,7 +895,7 @@ static void rb_reset_reader_page(struct ring_buffer_per_cpu *cpu_buffer) cpu_buffer->reader_page->read = 0; } -static inline void rb_inc_iter(struct ring_buffer_iter *iter) +static void rb_inc_iter(struct ring_buffer_iter *iter) { struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer; @@ -926,7 +925,7 @@ static inline void rb_inc_iter(struct ring_buffer_iter *iter) * and with this, we can determine what to place into the * data field. */ -static inline void +static void rb_update_event(struct ring_buffer_event *event, unsigned type, unsigned length) { @@ -964,7 +963,7 @@ rb_update_event(struct ring_buffer_event *event, } } -static inline unsigned rb_calculate_event_length(unsigned length) +static unsigned rb_calculate_event_length(unsigned length) { struct ring_buffer_event event; /* Used only for sizeof array */ @@ -1438,7 +1437,7 @@ int ring_buffer_write(struct ring_buffer *buffer, } EXPORT_SYMBOL_GPL(ring_buffer_write); -static inline int rb_per_cpu_empty(struct ring_buffer_per_cpu *cpu_buffer) +static int rb_per_cpu_empty(struct ring_buffer_per_cpu *cpu_buffer) { struct buffer_page *reader = cpu_buffer->reader_page; struct buffer_page *head = cpu_buffer->head_page; From 67d347245f76a149c45bffb1a10145d31d61d1da Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Fri, 9 Jan 2009 12:27:09 -0800 Subject: [PATCH 035/413] kernel/trace/ring_buffer.c: use DIV_ROUND_UP Instead of open-coding it. Signed-off-by: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 17 +++++------------ 1 file changed, 5 insertions(+), 12 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 9542990f515b..4832ffa5d937 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -123,8 +123,7 @@ void ring_buffer_normalize_time_stamp(int cpu, u64 *ts) EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp); #define RB_EVNT_HDR_SIZE (sizeof(struct ring_buffer_event)) -#define RB_ALIGNMENT_SHIFT 2 -#define RB_ALIGNMENT (1 << RB_ALIGNMENT_SHIFT) +#define RB_ALIGNMENT 4U #define RB_MAX_SMALL_DATA 28 enum { @@ -151,7 +150,7 @@ rb_event_length(struct ring_buffer_event *event) case RINGBUF_TYPE_DATA: if (event->len) - length = event->len << RB_ALIGNMENT_SHIFT; + length = event->len * RB_ALIGNMENT; else length = event->array[0]; return length + RB_EVNT_HDR_SIZE; @@ -937,15 +936,11 @@ rb_update_event(struct ring_buffer_event *event, break; case RINGBUF_TYPE_TIME_EXTEND: - event->len = - (RB_LEN_TIME_EXTEND + (RB_ALIGNMENT-1)) - >> RB_ALIGNMENT_SHIFT; + event->len = DIV_ROUND_UP(RB_LEN_TIME_EXTEND, RB_ALIGNMENT); break; case RINGBUF_TYPE_TIME_STAMP: - event->len = - (RB_LEN_TIME_STAMP + (RB_ALIGNMENT-1)) - >> RB_ALIGNMENT_SHIFT; + event->len = DIV_ROUND_UP(RB_LEN_TIME_STAMP, RB_ALIGNMENT); break; case RINGBUF_TYPE_DATA: @@ -954,9 +949,7 @@ rb_update_event(struct ring_buffer_event *event, event->len = 0; event->array[0] = length; } else - event->len = - (length + (RB_ALIGNMENT-1)) - >> RB_ALIGNMENT_SHIFT; + event->len = DIV_ROUND_UP(length, RB_ALIGNMENT); break; default: BUG(); From 034939b65ad5ff64b9709210b3469a95153c51a3 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 8 Jan 2009 10:03:56 -0800 Subject: [PATCH 036/413] tracing/ftrace: handle more than one stat file per tracer Impact: new API for tracers Make the stat tracing API reentrant. And also provide the new directory /debugfs/tracing/trace_stat which will contain all the stat files for the current active tracer. Now a tracer will, if desired, want to provide a zero terminated array of tracer_stat structures. Each one contains the callbacks necessary for one stat file. It have to provide at least a name for its stat file, an iterator with stat_start/start_next callback and an output callback for one stat entry. Also adapt the branch tracer to this new API. We create two files "all" and "annotated" inside the /debugfs/tracing/trace_stat directory, making the both stats simultaneously available instead of needing to change an option to switch from one stat file to another. The output of these stats haven't changed. Changes in v2: _ Apply the previous memory leak fix (rebase against tip/master) Changes in v3: _ Merge the patch that adapted the branch tracer to this Api in this patch to not break the kernel build. Signed-off-by: Frederic Weisbecker Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.h | 35 +++--- kernel/trace/trace_branch.c | 69 +++++------ kernel/trace/trace_stat.c | 230 ++++++++++++++++++++++++++---------- 3 files changed, 217 insertions(+), 117 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 94ed45e93a80..b3f9ad1b4d84 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -334,6 +334,25 @@ struct tracer_flags { /* Makes more easy to define a tracer opt */ #define TRACER_OPT(s, b) .name = #s, .bit = b +/* + * If you want to provide a stat file (one-shot statistics), fill + * an iterator with stat_start/stat_next and a stat_show callbacks. + * The others callbacks are optional. + */ +struct tracer_stat { + /* The name of your stat file */ + const char *name; + /* Iteration over statistic entries */ + void *(*stat_start)(void); + void *(*stat_next)(void *prev, int idx); + /* Compare two entries for sorting (optional) for stats */ + int (*stat_cmp)(void *p1, void *p2); + /* Print a stat entry */ + int (*stat_show)(struct seq_file *s, void *p); + /* Print the headers of your stat entries */ + int (*stat_headers)(struct seq_file *s); +}; + /* * A specific tracer, represented by methods that operate on a trace array: */ @@ -361,21 +380,7 @@ struct tracer { struct tracer *next; int print_max; struct tracer_flags *flags; - - /* - * If you change one of the following on tracing runtime, recall - * init_tracer_stat() - */ - - /* Iteration over statistic entries */ - void *(*stat_start)(void); - void *(*stat_next)(void *prev, int idx); - /* Compare two entries for sorting (optional) for stats */ - int (*stat_cmp)(void *p1, void *p2); - /* Print a stat entry */ - int (*stat_show)(struct seq_file *s, void *p); - /* Print the headers of your stat entries */ - int (*stat_headers)(struct seq_file *s); + struct tracer_stat *stats; }; struct trace_seq { diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 4785a3b9bc4a..da5cf3e5581b 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -306,19 +306,6 @@ static int annotated_branch_stat_cmp(void *p1, void *p2) } #ifdef CONFIG_PROFILE_ALL_BRANCHES -enum { - TRACE_BRANCH_OPT_ALL = 0x1 -}; - -static struct tracer_opt branch_opts[] = { - { TRACER_OPT(stat_all_branch, TRACE_BRANCH_OPT_ALL) }, - { } -}; - -static struct tracer_flags branch_flags = { - .val = 0, - .opts = branch_opts -}; extern unsigned long __start_branch_profile[]; extern unsigned long __stop_branch_profile[]; @@ -352,28 +339,36 @@ all_branch_stat_next(void *v, int idx) return p; } -static int branch_set_flag(u32 old_flags, u32 bit, int set) -{ - if (bit == TRACE_BRANCH_OPT_ALL) { - if (set) { - branch_trace.stat_headers = all_branch_stat_headers; - branch_trace.stat_start = all_branch_stat_start; - branch_trace.stat_next = all_branch_stat_next; - branch_trace.stat_cmp = NULL; - } else { - branch_trace.stat_headers = - annotated_branch_stat_headers; - branch_trace.stat_start = annotated_branch_stat_start; - branch_trace.stat_next = annotated_branch_stat_next; - branch_trace.stat_cmp = annotated_branch_stat_cmp; - } - init_tracer_stat(&branch_trace); - } - return 0; -} +static struct tracer_stat branch_stats[] = { + {.name = "annotated", + .stat_start = annotated_branch_stat_start, + .stat_next = annotated_branch_stat_next, + .stat_cmp = annotated_branch_stat_cmp, + .stat_headers = annotated_branch_stat_headers, + .stat_show = branch_stat_show}, + {.name = "all", + .stat_start = all_branch_stat_start, + .stat_next = all_branch_stat_next, + .stat_headers = all_branch_stat_headers, + .stat_show = branch_stat_show}, + + { } +}; +#else +static struct tracer_stat branch_stats[] = { + {.name = "annotated", + .stat_start = annotated_branch_stat_start, + .stat_next = annotated_branch_stat_next, + .stat_cmp = annotated_branch_stat_cmp, + .stat_headers = annotated_branch_stat_headers, + .stat_show = branch_stat_show}, + + { } +}; #endif /* CONFIG_PROFILE_ALL_BRANCHES */ + static struct tracer branch_trace __read_mostly = { .name = "branch", @@ -383,16 +378,8 @@ static struct tracer branch_trace __read_mostly = #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_branch, #endif /* CONFIG_FTRACE_SELFTEST */ -#endif /* CONFIG_BRANCH_TRACER */ - .stat_start = annotated_branch_stat_start, - .stat_next = annotated_branch_stat_next, - .stat_show = branch_stat_show, - .stat_headers = annotated_branch_stat_headers, - .stat_cmp = annotated_branch_stat_cmp, -#ifdef CONFIG_PROFILE_ALL_BRANCHES - .flags = &branch_flags, - .set_flag = branch_set_flag, #endif + .stats = branch_stats }; __init static int init_branch_trace(void) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index f110ce9ce7fb..1515f9e7adfc 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -21,37 +21,87 @@ struct trace_stat_list { void *stat; }; -static LIST_HEAD(stat_list); +/* A stat session is the stats output in one file */ +struct tracer_stat_session { + struct tracer_stat *ts; + struct list_head stat_list; + struct mutex stat_mutex; +}; -/* - * This is a copy of the current tracer to avoid racy - * and dangerous output while the current tracer is - * switched. - */ -static struct tracer current_tracer; - -/* - * Protect both the current tracer and the global - * stat list. - */ -static DEFINE_MUTEX(stat_list_mutex); +/* All of the sessions currently in use. Each stat file embeed one session */ +static struct tracer_stat_session **all_stat_sessions; +static int nb_sessions; +static struct dentry *stat_dir, **stat_files; -static void reset_stat_list(void) +static void reset_stat_session(struct tracer_stat_session *session) { struct trace_stat_list *node, *next; - list_for_each_entry_safe(node, next, &stat_list, list) + list_for_each_entry_safe(node, next, &session->stat_list, list) kfree(node); - INIT_LIST_HEAD(&stat_list); + INIT_LIST_HEAD(&session->stat_list); } -void init_tracer_stat(struct tracer *trace) +/* Called when a tracer is initialized */ +static int init_all_sessions(int nb, struct tracer_stat *ts) { - mutex_lock(&stat_list_mutex); - current_tracer = *trace; - mutex_unlock(&stat_list_mutex); + int i, j; + struct tracer_stat_session *session; + + nb_sessions = 0; + + if (all_stat_sessions) { + for (i = 0; i < nb_sessions; i++) { + session = all_stat_sessions[i]; + reset_stat_session(session); + mutex_destroy(&session->stat_mutex); + kfree(session); + } + } + all_stat_sessions = kmalloc(sizeof(struct tracer_stat_session *) * nb, + GFP_KERNEL); + if (!all_stat_sessions) + return -ENOMEM; + + for (i = 0; i < nb; i++) { + session = kmalloc(sizeof(struct tracer_stat_session) * nb, + GFP_KERNEL); + if (!session) + goto free_sessions; + + INIT_LIST_HEAD(&session->stat_list); + mutex_init(&session->stat_mutex); + session->ts = &ts[i]; + all_stat_sessions[i] = session; + } + nb_sessions = nb; + return 0; + +free_sessions: + + for (j = 0; j < i; j++) + kfree(all_stat_sessions[i]); + + kfree(all_stat_sessions); + all_stat_sessions = NULL; + + return -ENOMEM; +} + +static int basic_tracer_stat_checks(struct tracer_stat *ts) +{ + int i; + + if (!ts) + return 0; + + for (i = 0; ts[i].name; i++) { + if (!ts[i].stat_start || !ts[i].stat_next || !ts[i].stat_show) + return -EBUSY; + } + return i; } /* @@ -69,22 +119,19 @@ static int dummy_cmp(void *p1, void *p2) * All of these copies and sorting are required on all opening * since the stats could have changed between two file sessions. */ -static int stat_seq_init(void) +static int stat_seq_init(struct tracer_stat_session *session) { struct trace_stat_list *iter_entry, *new_entry; + struct tracer_stat *ts = session->ts; void *prev_stat; int ret = 0; int i; - mutex_lock(&stat_list_mutex); - reset_stat_list(); + mutex_lock(&session->stat_mutex); + reset_stat_session(session); - if (!current_tracer.stat_start || !current_tracer.stat_next || - !current_tracer.stat_show) - goto exit; - - if (!current_tracer.stat_cmp) - current_tracer.stat_cmp = dummy_cmp; + if (!ts->stat_cmp) + ts->stat_cmp = dummy_cmp; /* * The first entry. Actually this is the second, but the first @@ -97,9 +144,10 @@ static int stat_seq_init(void) } INIT_LIST_HEAD(&new_entry->list); - list_add(&new_entry->list, &stat_list); - new_entry->stat = current_tracer.stat_start(); + list_add(&new_entry->list, &session->stat_list); + + new_entry->stat = ts->stat_start(); prev_stat = new_entry->stat; /* @@ -114,15 +162,16 @@ static int stat_seq_init(void) } INIT_LIST_HEAD(&new_entry->list); - new_entry->stat = current_tracer.stat_next(prev_stat, i); + new_entry->stat = ts->stat_next(prev_stat, i); /* End of insertion */ if (!new_entry->stat) break; - list_for_each_entry(iter_entry, &stat_list, list) { + list_for_each_entry(iter_entry, &session->stat_list, list) { + /* Insertion with a descendent sorting */ - if (current_tracer.stat_cmp(new_entry->stat, + if (ts->stat_cmp(new_entry->stat, iter_entry->stat) > 0) { list_add_tail(&new_entry->list, @@ -131,7 +180,7 @@ static int stat_seq_init(void) /* The current smaller value */ } else if (list_is_last(&iter_entry->list, - &stat_list)) { + &session->stat_list)) { list_add(&new_entry->list, &iter_entry->list); break; } @@ -140,49 +189,49 @@ static int stat_seq_init(void) prev_stat = new_entry->stat; } exit: - mutex_unlock(&stat_list_mutex); + mutex_unlock(&session->stat_mutex); return ret; exit_free_list: - reset_stat_list(); - mutex_unlock(&stat_list_mutex); + reset_stat_session(session); + mutex_unlock(&session->stat_mutex); return ret; } static void *stat_seq_start(struct seq_file *s, loff_t *pos) { - struct list_head *l = (struct list_head *)s->private; + struct tracer_stat_session *session = s->private; /* Prevent from tracer switch or stat_list modification */ - mutex_lock(&stat_list_mutex); + mutex_lock(&session->stat_mutex); /* If we are in the beginning of the file, print the headers */ - if (!*pos && current_tracer.stat_headers) - current_tracer.stat_headers(s); + if (!*pos && session->ts->stat_headers) + session->ts->stat_headers(s); - return seq_list_start(l, *pos); + return seq_list_start(&session->stat_list, *pos); } static void *stat_seq_next(struct seq_file *s, void *p, loff_t *pos) { - struct list_head *l = (struct list_head *)s->private; + struct tracer_stat_session *session = s->private; - return seq_list_next(p, l, pos); + return seq_list_next(p, &session->stat_list, pos); } -static void stat_seq_stop(struct seq_file *m, void *p) +static void stat_seq_stop(struct seq_file *s, void *p) { - mutex_unlock(&stat_list_mutex); + struct tracer_stat_session *session = s->private; + mutex_unlock(&session->stat_mutex); } static int stat_seq_show(struct seq_file *s, void *v) { - struct trace_stat_list *entry; + struct tracer_stat_session *session = s->private; + struct trace_stat_list *l = list_entry(v, struct trace_stat_list, list); - entry = list_entry(v, struct trace_stat_list, list); - - return current_tracer.stat_show(s, entry->stat); + return session->ts->stat_show(s, l->stat); } static const struct seq_operations trace_stat_seq_ops = { @@ -192,15 +241,18 @@ static const struct seq_operations trace_stat_seq_ops = { .show = stat_seq_show }; +/* The session stat is refilled and resorted at each stat file opening */ static int tracing_stat_open(struct inode *inode, struct file *file) { int ret; + struct tracer_stat_session *session = inode->i_private; + ret = seq_open(file, &trace_stat_seq_ops); if (!ret) { struct seq_file *m = file->private_data; - m->private = &stat_list; - ret = stat_seq_init(); + m->private = session; + ret = stat_seq_init(session); } return ret; @@ -212,9 +264,12 @@ static int tracing_stat_open(struct inode *inode, struct file *file) */ static int tracing_stat_release(struct inode *i, struct file *f) { - mutex_lock(&stat_list_mutex); - reset_stat_list(); - mutex_unlock(&stat_list_mutex); + struct tracer_stat_session *session = i->i_private; + + mutex_lock(&session->stat_mutex); + reset_stat_session(session); + mutex_unlock(&session->stat_mutex); + return 0; } @@ -225,17 +280,70 @@ static const struct file_operations tracing_stat_fops = { .release = tracing_stat_release }; + +static void destroy_trace_stat_files(void) +{ + int i; + + if (stat_files) { + for (i = 0; i < nb_sessions; i++) + debugfs_remove(stat_files[i]); + kfree(stat_files); + stat_files = NULL; + } +} + +static void init_trace_stat_files(void) +{ + int i; + + if (!stat_dir || !nb_sessions) + return; + + stat_files = kmalloc(sizeof(struct dentry *) * nb_sessions, GFP_KERNEL); + + if (!stat_files) { + pr_warning("trace stat: not enough memory\n"); + return; + } + + for (i = 0; i < nb_sessions; i++) { + struct tracer_stat_session *session = all_stat_sessions[i]; + stat_files[i] = debugfs_create_file(session->ts->name, 0644, + stat_dir, + session, &tracing_stat_fops); + if (!stat_files[i]) + pr_warning("cannot create %s entry\n", + session->ts->name); + } +} + +void init_tracer_stat(struct tracer *trace) +{ + int nb = basic_tracer_stat_checks(trace->stats); + + destroy_trace_stat_files(); + + if (nb < 0) { + pr_warning("stat tracing: missing stat callback on %s\n", + trace->name); + return; + } + if (!nb) + return; + + init_all_sessions(nb, trace->stats); + init_trace_stat_files(); +} + static int __init tracing_stat_init(void) { struct dentry *d_tracing; - struct dentry *entry; d_tracing = tracing_init_dentry(); - entry = debugfs_create_file("trace_stat", 0444, d_tracing, - NULL, - &tracing_stat_fops); - if (!entry) + stat_dir = debugfs_create_dir("trace_stat", d_tracing); + if (!stat_dir) pr_warning("Could not create debugfs " "'trace_stat' entry\n"); return 0; From 5d2ad3316e29ad218f98d66b9c0ce6d4bcd05b77 Mon Sep 17 00:00:00 2001 From: Pekka Paalanen Date: Sat, 3 Jan 2009 21:09:27 +0200 Subject: [PATCH 037/413] doc: mmiotrace.txt, buffer size control change Impact: prevents confusing the user when buffer size is inadequate The tracing framework offers a resizeable buffer, which mmiotrace uses to record events. If the buffer is full, the following events will be lost. Events should not be lost, so the documentation instructs the user to increase the buffer size. The buffer size is set via a debugfs file. Mmiotrace documentation was not updated the same time the debugfs file was changed. The old file was tracing/trace_entries and first contained the number of entries the buffer had space for, per cpu. Nowadays this file is replaced with the file tracing/buffer_size_kb, which tells the amount of memory reserved for the buffer, per cpu, in kilobytes. Previously, a flag had to be toggled via the debugfs file tracing/tracing_enabled when the buffer size was changed. This is no longer necessary. The mmiotrace documentation is updated to reflect the current state of the tracing framework. Signed-off-by: Pekka Paalanen Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- Documentation/tracers/mmiotrace.txt | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/Documentation/tracers/mmiotrace.txt b/Documentation/tracers/mmiotrace.txt index cde23b4a12a1..5731c67abc55 100644 --- a/Documentation/tracers/mmiotrace.txt +++ b/Documentation/tracers/mmiotrace.txt @@ -78,12 +78,10 @@ to view your kernel log and look for "mmiotrace has lost events" warning. If events were lost, the trace is incomplete. You should enlarge the buffers and try again. Buffers are enlarged by first seeing how large the current buffers are: -$ cat /debug/tracing/trace_entries +$ cat /debug/tracing/buffer_size_kb gives you a number. Approximately double this number and write it back, for instance: -$ echo 0 > /debug/tracing/tracing_enabled -$ echo 128000 > /debug/tracing/trace_entries -$ echo 1 > /debug/tracing/tracing_enabled +$ echo 128000 > /debug/tracing/buffer_size_kb Then start again from the top. If you are doing a trace for a driver project, e.g. Nouveau, you should also From fe6f90e57fd31af8daca534ea01db2e5666c15da Mon Sep 17 00:00:00 2001 From: Pekka Paalanen Date: Sat, 3 Jan 2009 21:23:51 +0200 Subject: [PATCH 038/413] trace: mmiotrace to the tracer menu in Kconfig Impact: cosmetic change in Kconfig menu layout This patch was originally suggested by Peter Zijlstra, but seems it was forgotten. CONFIG_MMIOTRACE and CONFIG_MMIOTRACE_TEST were selectable directly under the Kernel hacking / debugging menu in the kernel configuration system. They were present only for x86 and x86_64. Other tracers that use the ftrace tracing framework are in their own sub-menu. This patch moves the mmiotrace configuration options there. Since the Kconfig file, where the tracer menu is, is not architecture specific, HAVE_MMIOTRACE_SUPPORT is introduced and provided only by x86/x86_64. CONFIG_MMIOTRACE now depends on it. Signed-off-by: Pekka Paalanen Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/Kconfig.debug | 24 ++---------------------- kernel/trace/Kconfig | 23 +++++++++++++++++++++++ 2 files changed, 25 insertions(+), 22 deletions(-) diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index 10d6cc3fd052..e1983fa025d2 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug @@ -174,28 +174,8 @@ config IOMMU_LEAK Add a simple leak tracer to the IOMMU code. This is useful when you are debugging a buggy device driver that leaks IOMMU mappings. -config MMIOTRACE - bool "Memory mapped IO tracing" - depends on DEBUG_KERNEL && PCI - select TRACING - help - Mmiotrace traces Memory Mapped I/O access and is meant for - debugging and reverse engineering. It is called from the ioremap - implementation and works via page faults. Tracing is disabled by - default and can be enabled at run-time. - - See Documentation/tracers/mmiotrace.txt. - If you are not helping to develop drivers, say N. - -config MMIOTRACE_TEST - tristate "Test module for mmiotrace" - depends on MMIOTRACE && m - help - This is a dumb module for testing mmiotrace. It is very dangerous - as it will write garbage to IO memory starting at a given address. - However, it should be safe to use on e.g. unused portion of VRAM. - - Say N, unless you absolutely know what you are doing. +config HAVE_MMIOTRACE_SUPPORT + def_bool y # # IO delay types: diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 1c0b7504cab3..944239296f13 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -323,4 +323,27 @@ config FTRACE_STARTUP_TEST functioning properly. It will do tests on all the configured tracers of ftrace. +config MMIOTRACE + bool "Memory mapped IO tracing" + depends on HAVE_MMIOTRACE_SUPPORT && DEBUG_KERNEL && PCI + select TRACING + help + Mmiotrace traces Memory Mapped I/O access and is meant for + debugging and reverse engineering. It is called from the ioremap + implementation and works via page faults. Tracing is disabled by + default and can be enabled at run-time. + + See Documentation/tracers/mmiotrace.txt. + If you are not helping to develop drivers, say N. + +config MMIOTRACE_TEST + tristate "Test module for mmiotrace" + depends on MMIOTRACE && m + help + This is a dumb module for testing mmiotrace. It is very dangerous + as it will write garbage to IO memory starting at a given address. + However, it should be safe to use on e.g. unused portion of VRAM. + + Say N, unless you absolutely know what you are doing. + endmenu From 173ed24ee2d64f5de28654eb456ec1ee18a142e5 Mon Sep 17 00:00:00 2001 From: Pekka Paalanen Date: Tue, 6 Jan 2009 13:57:11 +0200 Subject: [PATCH 039/413] mmiotrace: count events lost due to not recording Impact: enhances lost events counting in mmiotrace The tracing framework, or the ring buffer facility it uses, has a switch to stop recording data. When recording is off, the trace events will be lost. The framework does not count these, so mmiotrace has to count them itself. Signed-off-by: Pekka Paalanen Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_mmiotrace.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index fcec59ff2355..621c8c3f3139 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -9,6 +9,7 @@ #include #include #include +#include #include "trace.h" #include "trace_output.h" @@ -20,6 +21,7 @@ struct header_iter { static struct trace_array *mmio_trace_array; static bool overrun_detected; static unsigned long prev_overruns; +static atomic_t dropped_count; static void mmio_reset_data(struct trace_array *tr) { @@ -122,11 +124,11 @@ static void mmio_close(struct trace_iterator *iter) static unsigned long count_overruns(struct trace_iterator *iter) { - unsigned long cnt = 0; + unsigned long cnt = atomic_xchg(&dropped_count, 0); unsigned long over = ring_buffer_overruns(iter->tr->buffer); if (over > prev_overruns) - cnt = over - prev_overruns; + cnt += over - prev_overruns; prev_overruns = over; return cnt; } @@ -308,8 +310,10 @@ static void __trace_mmiotrace_rw(struct trace_array *tr, event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq_flags); - if (!event) + if (!event) { + atomic_inc(&dropped_count); return; + } entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, preempt_count()); entry->ent.type = TRACE_MMIO_RW; @@ -336,8 +340,10 @@ static void __trace_mmiotrace_map(struct trace_array *tr, event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq_flags); - if (!event) + if (!event) { + atomic_inc(&dropped_count); return; + } entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, preempt_count()); entry->ent.type = TRACE_MMIO_MAP; From 18c167fd6d8feec5d337bd8fbc3a17da4cc37652 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Mon, 12 Jan 2009 10:00:51 +0800 Subject: [PATCH 040/413] ftrace, ia64: make recordmcount distinct module compile In IA64, module build and kernel build use different option. Make recordmcount.pl differentiate the two cases. Signed-off-by: Shaohua Li Acked-by: Sam Ravnborg Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- scripts/Makefile.build | 13 +++++++------ scripts/recordmcount.pl | 6 +++--- 2 files changed, 10 insertions(+), 9 deletions(-) diff --git a/scripts/Makefile.build b/scripts/Makefile.build index 5d900307de3e..b5efa98994bb 100644 --- a/scripts/Makefile.build +++ b/scripts/Makefile.build @@ -112,13 +112,13 @@ endif # --------------------------------------------------------------------------- # Default is built-in, unless we know otherwise -modkern_cflags := $(CFLAGS_KERNEL) +modkern_cflags = $(if $(part-of-module), $(CFLAGS_MODULE), $(CFLAGS_KERNEL)) quiet_modtag := $(empty) $(empty) -$(real-objs-m) : modkern_cflags := $(CFLAGS_MODULE) -$(real-objs-m:.o=.i) : modkern_cflags := $(CFLAGS_MODULE) -$(real-objs-m:.o=.s) : modkern_cflags := $(CFLAGS_MODULE) -$(real-objs-m:.o=.lst): modkern_cflags := $(CFLAGS_MODULE) +$(real-objs-m) : part-of-module := y +$(real-objs-m:.o=.i) : part-of-module := y +$(real-objs-m:.o=.s) : part-of-module := y +$(real-objs-m:.o=.lst): part-of-module := y $(real-objs-m) : quiet_modtag := [M] $(real-objs-m:.o=.i) : quiet_modtag := [M] @@ -215,7 +215,8 @@ endif ifdef CONFIG_FTRACE_MCOUNT_RECORD cmd_record_mcount = perl $(srctree)/scripts/recordmcount.pl "$(ARCH)" \ "$(if $(CONFIG_64BIT),64,32)" \ - "$(OBJDUMP)" "$(OBJCOPY)" "$(CC)" "$(LD)" "$(NM)" "$(RM)" "$(MV)" "$(@)"; + "$(OBJDUMP)" "$(OBJCOPY)" "$(CC)" "$(LD)" "$(NM)" "$(RM)" "$(MV)" \ + "$(if $(part-of-module),1,0)" "$(@)"; endif define rule_cc_o_c diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index fe831412bea9..282485a22bf8 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -100,14 +100,14 @@ $P =~ s@.*/@@g; my $V = '0.1'; -if ($#ARGV < 6) { - print "usage: $P arch objdump objcopy cc ld nm rm mv inputfile\n"; +if ($#ARGV < 7) { + print "usage: $P arch objdump objcopy cc ld nm rm mv is_module inputfile\n"; print "version: $V\n"; exit(1); } my ($arch, $bits, $objdump, $objcopy, $cc, - $ld, $nm, $rm, $mv, $inputfile) = @ARGV; + $ld, $nm, $rm, $mv, $is_module, $inputfile) = @ARGV; # Acceptable sections to record. my %text_sections = ( From 25aac9dc7c8c73798c1be8aa36141f980d32579e Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Fri, 9 Jan 2009 11:29:40 +0800 Subject: [PATCH 041/413] ftrace, ia64: explictly ignore a file in recordmcount.pl In IA64, a function pointer isn't a 'unsigned long' but a 'struct {unsigned long ip, unsigned long gp}'. MCOUNT_ADDR is determined at link time not compile time, so explictly ignore kernel/trace/ftrace.o in recordmcount.pl. Signed-off-by: Shaohua Li Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 10 +--------- scripts/recordmcount.pl | 5 +++++ 2 files changed, 6 insertions(+), 9 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 9e54a6ccdb93..76bb884b6e16 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -263,14 +263,6 @@ static void ftrace_update_pid_func(void) # error Dynamic ftrace depends on MCOUNT_RECORD #endif -/* - * Since MCOUNT_ADDR may point to mcount itself, we do not want - * to get it confused by reading a reference in the code as we - * are parsing on objcopy output of text. Use a variable for - * it instead. - */ -static unsigned long mcount_addr = MCOUNT_ADDR; - enum { FTRACE_ENABLE_CALLS = (1 << 0), FTRACE_DISABLE_CALLS = (1 << 1), @@ -575,7 +567,7 @@ ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec) ip = rec->ip; - ret = ftrace_make_nop(mod, rec, mcount_addr); + ret = ftrace_make_nop(mod, rec, MCOUNT_ADDR); if (ret) { ftrace_bug(ret, ip); rec->flags |= FTRACE_FL_FAILED; diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 282485a22bf8..070042b4ccd5 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -109,6 +109,11 @@ if ($#ARGV < 7) { my ($arch, $bits, $objdump, $objcopy, $cc, $ld, $nm, $rm, $mv, $is_module, $inputfile) = @ARGV; +# This file refers to mcount and shouldn't be ftraced, so lets' ignore it +if ($inputfile eq "kernel/trace/ftrace.o") { + exit(0); +} + # Acceptable sections to record. my %text_sections = ( ".text" => 1, From f00012074b1a1a67d9c8603617bbbab267347ca6 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Fri, 9 Jan 2009 11:29:42 +0800 Subject: [PATCH 042/413] ftrace, ia64: Add macro for ftrace_caller Define FTRACE_ADDR. In IA64, a function pointer isn't a 'unsigned long' but a 'struct {unsigned long ip, unsigned long gp}'. Signed-off-by: Shaohua Li Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 4 ++++ kernel/trace/ftrace.c | 2 +- 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 677432b9cb7e..054721487574 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -126,6 +126,10 @@ extern int ftrace_update_ftrace_func(ftrace_func_t func); extern void ftrace_caller(void); extern void ftrace_call(void); extern void mcount_call(void); + +#ifndef FTRACE_ADDR +#define FTRACE_ADDR ((unsigned long)ftrace_caller) +#endif #ifdef CONFIG_FUNCTION_GRAPH_TRACER extern void ftrace_graph_caller(void); extern int ftrace_enable_ftrace_graph_caller(void); diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 76bb884b6e16..9f536108d3f3 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -455,7 +455,7 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) unsigned long ip, fl; unsigned long ftrace_addr; - ftrace_addr = (unsigned long)ftrace_caller; + ftrace_addr = (unsigned long)FTRACE_ADDR; ip = rec->ip; From 418071eb6adbfd3980b2f57f7df8e03921e3f1d7 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Fri, 9 Jan 2009 11:29:44 +0800 Subject: [PATCH 043/413] ftrace, ia64: Add recordmcount for ia64 Add recordmcount for ia64. Signed-off-by: Shaohua Li Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- scripts/recordmcount.pl | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 070042b4ccd5..2ded5c8e6b85 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -206,6 +206,13 @@ if ($arch eq "x86_64") { $alignment = 2; $section_type = '%progbits'; +} elsif ($arch eq "ia64") { + $mcount_regex = "^\\s*([0-9a-fA-F]+):.*\\s_mcount\$"; + $type = "data8"; + + if ($is_module eq "0") { + $cc .= " -mconstant-gp"; + } } else { die "Arch $arch is not supported with CONFIG_FTRACE_MCOUNT_RECORD"; } From d3e75ff14bc1453c4762428395aac9953a023efc Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Fri, 9 Jan 2009 11:29:46 +0800 Subject: [PATCH 044/413] ftrace, ia64: IA64 static ftrace support IA64 ftrace suppport. In IA64, below code will be added in each function if -pg is enabled. alloc r40=ar.pfs,12,8,0 mov r43=r0;; mov r42=b0 mov r41=r1 nop.i 0x0 br.call.sptk.many b0 = _mcount;; Signed-off-by: Shaohua Li Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/ia64/Kconfig | 1 + arch/ia64/include/asm/ftrace.h | 15 +++++++++++ arch/ia64/kernel/entry.S | 49 ++++++++++++++++++++++++++++++++++ arch/ia64/kernel/ia64_ksyms.c | 6 +++++ 4 files changed, 71 insertions(+) create mode 100644 arch/ia64/include/asm/ftrace.h diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index 3d31636cbafb..b992ba447c41 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -21,6 +21,7 @@ config IA64 select HAVE_OPROFILE select HAVE_KPROBES select HAVE_KRETPROBES + select HAVE_FUNCTION_TRACER select HAVE_DMA_ATTRS select HAVE_KVM select HAVE_ARCH_TRACEHOOK diff --git a/arch/ia64/include/asm/ftrace.h b/arch/ia64/include/asm/ftrace.h new file mode 100644 index 000000000000..48694b3ba5a8 --- /dev/null +++ b/arch/ia64/include/asm/ftrace.h @@ -0,0 +1,15 @@ +#ifndef _ASM_IA64_FTRACE_H +#define _ASM_IA64_FTRACE_H + +#ifdef CONFIG_FUNCTION_TRACER +#define MCOUNT_INSN_SIZE 32 /* sizeof mcount call */ + +#ifndef __ASSEMBLY__ +extern void _mcount(unsigned long pfs, unsigned long r1, unsigned long b0, unsigned long r0); +#define mcount _mcount + +#endif + +#endif /* CONFIG_FUNCTION_TRACER */ + +#endif /* _ASM_IA64_FTRACE_H */ diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S index d435f4a7a96c..c2f7d798e2a5 100644 --- a/arch/ia64/kernel/entry.S +++ b/arch/ia64/kernel/entry.S @@ -47,6 +47,7 @@ #include #include #include +#include #include "minstate.h" @@ -1404,6 +1405,54 @@ GLOBAL_ENTRY(unw_init_running) br.ret.sptk.many rp END(unw_init_running) +#ifdef CONFIG_FUNCTION_TRACER +GLOBAL_ENTRY(_mcount) + movl r2 = ftrace_stub + movl r3 = ftrace_trace_function;; + ld8 r3 = [r3];; + ld8 r3 = [r3];; + cmp.eq p7,p0 = r2, r3 +(p7) br.sptk.many ftrace_stub + ;; + + alloc loc0 = ar.pfs, 4, 4, 2, 0 + ;; + mov loc1 = b0 + mov out0 = b0 + mov loc2 = r8 + mov loc3 = r15 + ;; + adds out0 = -MCOUNT_INSN_SIZE, out0 + mov out1 = in2 + mov b6 = r3 + + br.call.sptk.many b0 = b6 + ;; + mov ar.pfs = loc0 + mov b0 = loc1 + mov r8 = loc2 + mov r15 = loc3 + br ftrace_stub + ;; +END(_mcount) + +GLOBAL_ENTRY(ftrace_stub) + mov r3 = b0 + movl r2 = _mcount_ret_helper + ;; + mov b6 = r2 + mov b7 = r3 + br.ret.sptk.many b6 + +_mcount_ret_helper: + mov b0 = r42 + mov r1 = r41 + mov ar.pfs = r40 + br b7 +END(ftrace_stub) + +#endif /* CONFIG_FUNCTION_TRACER */ + .rodata .align 8 .globl sys_call_table diff --git a/arch/ia64/kernel/ia64_ksyms.c b/arch/ia64/kernel/ia64_ksyms.c index 6da1f20d7372..2d311864e359 100644 --- a/arch/ia64/kernel/ia64_ksyms.c +++ b/arch/ia64/kernel/ia64_ksyms.c @@ -112,3 +112,9 @@ EXPORT_SYMBOL_GPL(esi_call_phys); #endif extern char ia64_ivt[]; EXPORT_SYMBOL(ia64_ivt); + +#include +#ifdef CONFIG_FUNCTION_TRACER +/* mcount is defined in assembly */ +EXPORT_SYMBOL(_mcount); +#endif From a14a07b8018b714e03a39ff2180c66e307ef4238 Mon Sep 17 00:00:00 2001 From: Shaohua Li Date: Fri, 9 Jan 2009 11:29:49 +0800 Subject: [PATCH 045/413] ftrace, ia64: IA64 dynamic ftrace support IA64 dynamic ftrace support. The original _mcount stub for each function is like: alloc r40=ar.pfs,12,8,0 mov r43=r0;; mov r42=b0 mov r41=r1 nop.i 0x0 br.call.sptk.many b0 = _mcount;; The patch convert it to below for nop: [MII] nop.m 0x0 mov r3=ip nop.i 0x0 [MLX] nop.m 0x0 nop.x 0x0;; This isn't completely nop, as there is one instuction 'mov r3=ip', but it should be light and harmless for code follow it. And below is for call [MII] nop.m 0x0 mov r3=ip nop.i 0x0 [MLX] nop.m 0x0 brl.many .;; In this way, only one instruction is changed to convert code between nop and call. This should meet dyn-ftrace's requirement. But this requires CPU support brl instruction, so dyn-ftrace isn't supported for old Itanium system. Assume there are quite few such old system running. Signed-off-by: Shaohua Li Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/ia64/Kconfig | 2 + arch/ia64/include/asm/ftrace.h | 13 +++ arch/ia64/kernel/Makefile | 5 + arch/ia64/kernel/entry.S | 51 ++++++++ arch/ia64/kernel/ftrace.c | 206 +++++++++++++++++++++++++++++++++ 5 files changed, 277 insertions(+) create mode 100644 arch/ia64/kernel/ftrace.c diff --git a/arch/ia64/Kconfig b/arch/ia64/Kconfig index b992ba447c41..e20c1d45930a 100644 --- a/arch/ia64/Kconfig +++ b/arch/ia64/Kconfig @@ -21,6 +21,8 @@ config IA64 select HAVE_OPROFILE select HAVE_KPROBES select HAVE_KRETPROBES + select HAVE_FTRACE_MCOUNT_RECORD + select HAVE_DYNAMIC_FTRACE if (!ITANIUM) select HAVE_FUNCTION_TRACER select HAVE_DMA_ATTRS select HAVE_KVM diff --git a/arch/ia64/include/asm/ftrace.h b/arch/ia64/include/asm/ftrace.h index 48694b3ba5a8..d20db3c2a656 100644 --- a/arch/ia64/include/asm/ftrace.h +++ b/arch/ia64/include/asm/ftrace.h @@ -8,6 +8,19 @@ extern void _mcount(unsigned long pfs, unsigned long r1, unsigned long b0, unsigned long r0); #define mcount _mcount +#include +/* In IA64, MCOUNT_ADDR is set in link time, so it's not a constant at compile time */ +#define MCOUNT_ADDR (((struct fnptr *)mcount)->ip) +#define FTRACE_ADDR (((struct fnptr *)ftrace_caller)->ip) + +static inline unsigned long ftrace_call_adjust(unsigned long addr) +{ + /* second bundle, insn 2 */ + return addr - 0x12; +} + +struct dyn_arch_ftrace { +}; #endif #endif /* CONFIG_FUNCTION_TRACER */ diff --git a/arch/ia64/kernel/Makefile b/arch/ia64/kernel/Makefile index c381ea954892..ab6e7ec0bba3 100644 --- a/arch/ia64/kernel/Makefile +++ b/arch/ia64/kernel/Makefile @@ -2,6 +2,10 @@ # Makefile for the linux kernel. # +ifdef CONFIG_DYNAMIC_FTRACE +CFLAGS_REMOVE_ftrace.o = -pg +endif + extra-y := head.o init_task.o vmlinux.lds obj-y := acpi.o entry.o efi.o efi_stub.o gate-data.o fsys.o ia64_ksyms.o irq.o irq_ia64.o \ @@ -28,6 +32,7 @@ obj-$(CONFIG_IA64_CYCLONE) += cyclone.o obj-$(CONFIG_CPU_FREQ) += cpufreq/ obj-$(CONFIG_IA64_MCA_RECOVERY) += mca_recovery.o obj-$(CONFIG_KPROBES) += kprobes.o jprobes.o +obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o obj-$(CONFIG_KEXEC) += machine_kexec.o relocate_kernel.o crash.o obj-$(CONFIG_CRASH_DUMP) += crash_dump.o obj-$(CONFIG_IA64_UNCACHED_ALLOCATOR) += uncached.o diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S index c2f7d798e2a5..e0be92a6abb0 100644 --- a/arch/ia64/kernel/entry.S +++ b/arch/ia64/kernel/entry.S @@ -1406,6 +1406,56 @@ GLOBAL_ENTRY(unw_init_running) END(unw_init_running) #ifdef CONFIG_FUNCTION_TRACER +#ifdef CONFIG_DYNAMIC_FTRACE +GLOBAL_ENTRY(_mcount) + br ftrace_stub +END(_mcount) + +.here: + br.ret.sptk.many b0 + +GLOBAL_ENTRY(ftrace_caller) + alloc out0 = ar.pfs, 8, 0, 4, 0 + mov out3 = r0 + ;; + mov out2 = b0 + add r3 = 0x20, r3 + mov out1 = r1; + br.call.sptk.many b0 = ftrace_patch_gp + //this might be called from module, so we must patch gp +ftrace_patch_gp: + movl gp=__gp + mov b0 = r3 + ;; +.global ftrace_call; +ftrace_call: +{ + .mlx + nop.m 0x0 + movl r3 = .here;; +} + alloc loc0 = ar.pfs, 4, 4, 2, 0 + ;; + mov loc1 = b0 + mov out0 = b0 + mov loc2 = r8 + mov loc3 = r15 + ;; + adds out0 = -MCOUNT_INSN_SIZE, out0 + mov out1 = in2 + mov b6 = r3 + + br.call.sptk.many b0 = b6 + ;; + mov ar.pfs = loc0 + mov b0 = loc1 + mov r8 = loc2 + mov r15 = loc3 + br ftrace_stub + ;; +END(ftrace_caller) + +#else GLOBAL_ENTRY(_mcount) movl r2 = ftrace_stub movl r3 = ftrace_trace_function;; @@ -1435,6 +1485,7 @@ GLOBAL_ENTRY(_mcount) br ftrace_stub ;; END(_mcount) +#endif GLOBAL_ENTRY(ftrace_stub) mov r3 = b0 diff --git a/arch/ia64/kernel/ftrace.c b/arch/ia64/kernel/ftrace.c new file mode 100644 index 000000000000..7fc8c961b1f7 --- /dev/null +++ b/arch/ia64/kernel/ftrace.c @@ -0,0 +1,206 @@ +/* + * Dynamic function tracing support. + * + * Copyright (C) 2008 Shaohua Li + * + * For licencing details, see COPYING. + * + * Defines low-level handling of mcount calls when the kernel + * is compiled with the -pg flag. When using dynamic ftrace, the + * mcount call-sites get patched lazily with NOP till they are + * enabled. All code mutation routines here take effect atomically. + */ + +#include +#include + +#include +#include + +/* In IA64, each function will be added below two bundles with -pg option */ +static unsigned char __attribute__((aligned(8))) +ftrace_orig_code[MCOUNT_INSN_SIZE] = { + 0x02, 0x40, 0x31, 0x10, 0x80, 0x05, /* alloc r40=ar.pfs,12,8,0 */ + 0xb0, 0x02, 0x00, 0x00, 0x42, 0x40, /* mov r43=r0;; */ + 0x05, 0x00, 0xc4, 0x00, /* mov r42=b0 */ + 0x11, 0x48, 0x01, 0x02, 0x00, 0x21, /* mov r41=r1 */ + 0x00, 0x00, 0x00, 0x02, 0x00, 0x00, /* nop.i 0x0 */ + 0x08, 0x00, 0x00, 0x50 /* br.call.sptk.many b0 = _mcount;; */ +}; + +struct ftrace_orig_insn { + u64 dummy1, dummy2, dummy3; + u64 dummy4:64-41+13; + u64 imm20:20; + u64 dummy5:3; + u64 sign:1; + u64 dummy6:4; +}; + +/* mcount stub will be converted below for nop */ +static unsigned char ftrace_nop_code[MCOUNT_INSN_SIZE] = { + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, /* [MII] nop.m 0x0 */ + 0x30, 0x00, 0x00, 0x60, 0x00, 0x00, /* mov r3=ip */ + 0x00, 0x00, 0x04, 0x00, /* nop.i 0x0 */ + 0x05, 0x00, 0x00, 0x00, 0x01, 0x00, /* [MLX] nop.m 0x0 */ + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, /* nop.x 0x0;; */ + 0x00, 0x00, 0x04, 0x00 +}; + +static unsigned char *ftrace_nop_replace(void) +{ + return ftrace_nop_code; +} + +/* + * mcount stub will be converted below for call + * Note: Just the last instruction is changed against nop + * */ +static unsigned char __attribute__((aligned(8))) +ftrace_call_code[MCOUNT_INSN_SIZE] = { + 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, /* [MII] nop.m 0x0 */ + 0x30, 0x00, 0x00, 0x60, 0x00, 0x00, /* mov r3=ip */ + 0x00, 0x00, 0x04, 0x00, /* nop.i 0x0 */ + 0x05, 0x00, 0x00, 0x00, 0x01, 0x00, /* [MLX] nop.m 0x0 */ + 0xff, 0xff, 0xff, 0xff, 0x7f, 0x00, /* brl.many .;;*/ + 0xf8, 0xff, 0xff, 0xc8 +}; + +struct ftrace_call_insn { + u64 dummy1, dummy2; + u64 dummy3:48; + u64 imm39_l:16; + u64 imm39_h:23; + u64 dummy4:13; + u64 imm20:20; + u64 dummy5:3; + u64 i:1; + u64 dummy6:4; +}; + +static unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) +{ + struct ftrace_call_insn *code = (void *)ftrace_call_code; + unsigned long offset = addr - (ip + 0x10); + + code->imm39_l = offset >> 24; + code->imm39_h = offset >> 40; + code->imm20 = offset >> 4; + code->i = offset >> 63; + return ftrace_call_code; +} + +static int +ftrace_modify_code(unsigned long ip, unsigned char *old_code, + unsigned char *new_code, int do_check) +{ + unsigned char replaced[MCOUNT_INSN_SIZE]; + + /* + * Note: Due to modules and __init, code can + * disappear and change, we need to protect against faulting + * as well as code changing. We do this by using the + * probe_kernel_* functions. + * + * No real locking needed, this code is run through + * kstop_machine, or before SMP starts. + */ + + if (!do_check) + goto skip_check; + + /* read the text we want to modify */ + if (probe_kernel_read(replaced, (void *)ip, MCOUNT_INSN_SIZE)) + return -EFAULT; + + /* Make sure it is what we expect it to be */ + if (memcmp(replaced, old_code, MCOUNT_INSN_SIZE) != 0) + return -EINVAL; + +skip_check: + /* replace the text with the new text */ + if (probe_kernel_write(((void *)ip), new_code, MCOUNT_INSN_SIZE)) + return -EPERM; + flush_icache_range(ip, ip + MCOUNT_INSN_SIZE); + + return 0; +} + +static int ftrace_make_nop_check(struct dyn_ftrace *rec, unsigned long addr) +{ + unsigned char __attribute__((aligned(8))) replaced[MCOUNT_INSN_SIZE]; + unsigned long ip = rec->ip; + + if (probe_kernel_read(replaced, (void *)ip, MCOUNT_INSN_SIZE)) + return -EFAULT; + if (rec->flags & FTRACE_FL_CONVERTED) { + struct ftrace_call_insn *call_insn, *tmp_call; + + call_insn = (void *)ftrace_call_code; + tmp_call = (void *)replaced; + call_insn->imm39_l = tmp_call->imm39_l; + call_insn->imm39_h = tmp_call->imm39_h; + call_insn->imm20 = tmp_call->imm20; + call_insn->i = tmp_call->i; + if (memcmp(replaced, ftrace_call_code, MCOUNT_INSN_SIZE) != 0) + return -EINVAL; + return 0; + } else { + struct ftrace_orig_insn *call_insn, *tmp_call; + + call_insn = (void *)ftrace_orig_code; + tmp_call = (void *)replaced; + call_insn->sign = tmp_call->sign; + call_insn->imm20 = tmp_call->imm20; + if (memcmp(replaced, ftrace_orig_code, MCOUNT_INSN_SIZE) != 0) + return -EINVAL; + return 0; + } +} + +int ftrace_make_nop(struct module *mod, + struct dyn_ftrace *rec, unsigned long addr) +{ + int ret; + char *new; + + ret = ftrace_make_nop_check(rec, addr); + if (ret) + return ret; + new = ftrace_nop_replace(); + return ftrace_modify_code(rec->ip, NULL, new, 0); +} + +int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr) +{ + unsigned long ip = rec->ip; + unsigned char *old, *new; + + old= ftrace_nop_replace(); + new = ftrace_call_replace(ip, addr); + return ftrace_modify_code(ip, old, new, 1); +} + +/* in IA64, _mcount can't directly call ftrace_stub. Only jump is ok */ +int ftrace_update_ftrace_func(ftrace_func_t func) +{ + unsigned long ip; + unsigned long addr = ((struct fnptr *)ftrace_call)->ip; + + if (func == ftrace_stub) + return 0; + ip = ((struct fnptr *)func)->ip; + + ia64_patch_imm64(addr + 2, ip); + + flush_icache_range(addr, addr + 16); + return 0; +} + +/* run from kstop_machine */ +int __init ftrace_dyn_arch_init(void *data) +{ + *(unsigned long *)data = 0; + + return 0; +} From 002bb86d8d42f18937aef396c3ecd65c7e02e21a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 10 Jan 2009 11:34:13 -0800 Subject: [PATCH 046/413] tracing/ftrace: separate events tracing and stats tracing engine Impact: tracing's Api change Currently, the stat tracing depends on the events tracing. When you switch to a new tracer, the stats files of the previous tracer will disappear. But it's more scalable to separate those two engines. This way, we can keep the stat files of one or several tracers when we want, without bothering of multiple tracer stat files or tracer switching. To build/destroys its stats files, a tracer just have to call register_stat_tracer/unregister_stat_tracer everytimes it wants to. Signed-off-by: Frederic Weisbecker Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 2 - kernel/trace/trace.h | 20 ---- kernel/trace/trace_branch.c | 108 +++++++++++--------- kernel/trace/trace_stat.c | 193 +++++++++++++++--------------------- kernel/trace/trace_stat.h | 31 ++++++ 5 files changed, 173 insertions(+), 181 deletions(-) create mode 100644 kernel/trace/trace_stat.h diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0418fc338b5c..40217fb499ea 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2353,7 +2353,6 @@ static int tracing_set_tracer(char *buf) if (ret) goto out; } - init_tracer_stat(t); trace_branch_enable(tr); out: @@ -3218,7 +3217,6 @@ __init static int tracer_alloc_buffers(void) #else current_trace = &nop_trace; #endif - init_tracer_stat(current_trace); /* All seems OK, enable tracing */ tracing_disabled = 0; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index b3f9ad1b4d84..79c872100dd5 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -334,24 +334,6 @@ struct tracer_flags { /* Makes more easy to define a tracer opt */ #define TRACER_OPT(s, b) .name = #s, .bit = b -/* - * If you want to provide a stat file (one-shot statistics), fill - * an iterator with stat_start/stat_next and a stat_show callbacks. - * The others callbacks are optional. - */ -struct tracer_stat { - /* The name of your stat file */ - const char *name; - /* Iteration over statistic entries */ - void *(*stat_start)(void); - void *(*stat_next)(void *prev, int idx); - /* Compare two entries for sorting (optional) for stats */ - int (*stat_cmp)(void *p1, void *p2); - /* Print a stat entry */ - int (*stat_show)(struct seq_file *s, void *p); - /* Print the headers of your stat entries */ - int (*stat_headers)(struct seq_file *s); -}; /* * A specific tracer, represented by methods that operate on a trace array: @@ -466,8 +448,6 @@ void tracing_start_sched_switch_record(void); int register_tracer(struct tracer *type); void unregister_tracer(struct tracer *type); -void init_tracer_stat(struct tracer *trace); - extern unsigned long nsecs_to_usecs(unsigned long nsecs); extern unsigned long tracing_max_latency; diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index da5cf3e5581b..ca017e0a9a27 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -16,12 +16,12 @@ #include #include "trace.h" +#include "trace_stat.h" #include "trace_output.h" -static struct tracer branch_trace; - #ifdef CONFIG_BRANCH_TRACER +static struct tracer branch_trace; static int branch_tracing_enabled __read_mostly; static DEFINE_MUTEX(branch_tracing_mutex); @@ -191,6 +191,30 @@ static struct trace_event trace_branch_event = { .binary = trace_nop_print, }; +static struct tracer branch_trace __read_mostly = +{ + .name = "branch", + .init = branch_trace_init, + .reset = branch_trace_reset, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_branch, +#endif /* CONFIG_FTRACE_SELFTEST */ +}; + +__init static int init_branch_tracer(void) +{ + int ret; + + ret = register_ftrace_event(&trace_branch_event); + if (!ret) { + printk(KERN_WARNING "Warning: could not register " + "branch events\n"); + return 1; + } + return register_tracer(&branch_trace); +} +device_initcall(init_branch_tracer); + #else static inline void trace_likely_condition(struct ftrace_branch_data *f, int val, int expect) @@ -305,6 +329,29 @@ static int annotated_branch_stat_cmp(void *p1, void *p2) return 0; } +static struct tracer_stat annotated_branch_stats = { + .name = "branch_annotated", + .stat_start = annotated_branch_stat_start, + .stat_next = annotated_branch_stat_next, + .stat_cmp = annotated_branch_stat_cmp, + .stat_headers = annotated_branch_stat_headers, + .stat_show = branch_stat_show +}; + +__init static int init_annotated_branch_stats(void) +{ + int ret; + + ret = register_stat_tracer(&annotated_branch_stats); + if (!ret) { + printk(KERN_WARNING "Warning: could not register " + "annotated branches stats\n"); + return 1; + } + return 0; +} +fs_initcall(init_annotated_branch_stats); + #ifdef CONFIG_PROFILE_ALL_BRANCHES extern unsigned long __start_branch_profile[]; @@ -339,60 +386,25 @@ all_branch_stat_next(void *v, int idx) return p; } -static struct tracer_stat branch_stats[] = { - {.name = "annotated", - .stat_start = annotated_branch_stat_start, - .stat_next = annotated_branch_stat_next, - .stat_cmp = annotated_branch_stat_cmp, - .stat_headers = annotated_branch_stat_headers, - .stat_show = branch_stat_show}, - - {.name = "all", +static struct tracer_stat all_branch_stats = { + .name = "branch_all", .stat_start = all_branch_stat_start, .stat_next = all_branch_stat_next, .stat_headers = all_branch_stat_headers, - .stat_show = branch_stat_show}, - - { } + .stat_show = branch_stat_show }; -#else -static struct tracer_stat branch_stats[] = { - {.name = "annotated", - .stat_start = annotated_branch_stat_start, - .stat_next = annotated_branch_stat_next, - .stat_cmp = annotated_branch_stat_cmp, - .stat_headers = annotated_branch_stat_headers, - .stat_show = branch_stat_show}, - { } -}; -#endif /* CONFIG_PROFILE_ALL_BRANCHES */ - - -static struct tracer branch_trace __read_mostly = +__init static int all_annotated_branch_stats(void) { - .name = "branch", -#ifdef CONFIG_BRANCH_TRACER - .init = branch_trace_init, - .reset = branch_trace_reset, -#ifdef CONFIG_FTRACE_SELFTEST - .selftest = trace_selftest_startup_branch, -#endif /* CONFIG_FTRACE_SELFTEST */ -#endif - .stats = branch_stats -}; - -__init static int init_branch_trace(void) -{ -#ifdef CONFIG_BRANCH_TRACER int ret; - ret = register_ftrace_event(&trace_branch_event); + + ret = register_stat_tracer(&all_branch_stats); if (!ret) { - printk(KERN_WARNING "Warning: could not register branch events\n"); + printk(KERN_WARNING "Warning: could not register " + "all branches stats\n"); return 1; } -#endif - - return register_tracer(&branch_trace); + return 0; } -device_initcall(init_branch_trace); +fs_initcall(all_annotated_branch_stats); +#endif /* CONFIG_PROFILE_ALL_BRANCHES */ diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index 1515f9e7adfc..cb29282b9485 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -10,28 +10,32 @@ #include -#include #include +#include "trace_stat.h" #include "trace.h" /* List of stat entries from a tracer */ struct trace_stat_list { - struct list_head list; - void *stat; + struct list_head list; + void *stat; }; /* A stat session is the stats output in one file */ struct tracer_stat_session { - struct tracer_stat *ts; - struct list_head stat_list; - struct mutex stat_mutex; + struct list_head session_list; + struct tracer_stat *ts; + struct list_head stat_list; + struct mutex stat_mutex; + struct dentry *file; }; /* All of the sessions currently in use. Each stat file embeed one session */ -static struct tracer_stat_session **all_stat_sessions; -static int nb_sessions; -static struct dentry *stat_dir, **stat_files; +static LIST_HEAD(all_stat_sessions); +static DEFINE_MUTEX(all_stat_sessions_mutex); + +/* The root directory for all stat files */ +static struct dentry *stat_dir; static void reset_stat_session(struct tracer_stat_session *session) @@ -44,66 +48,77 @@ static void reset_stat_session(struct tracer_stat_session *session) INIT_LIST_HEAD(&session->stat_list); } -/* Called when a tracer is initialized */ -static int init_all_sessions(int nb, struct tracer_stat *ts) +static void destroy_session(struct tracer_stat_session *session) { - int i, j; - struct tracer_stat_session *session; + debugfs_remove(session->file); + reset_stat_session(session); + mutex_destroy(&session->stat_mutex); + kfree(session); +} - nb_sessions = 0; - if (all_stat_sessions) { - for (i = 0; i < nb_sessions; i++) { - session = all_stat_sessions[i]; - reset_stat_session(session); - mutex_destroy(&session->stat_mutex); - kfree(session); - } +static int init_stat_file(struct tracer_stat_session *session); + +int register_stat_tracer(struct tracer_stat *trace) +{ + struct tracer_stat_session *session, *node, *tmp; + int ret; + + if (!trace) + return -EINVAL; + + if (!trace->stat_start || !trace->stat_next || !trace->stat_show) + return -EINVAL; + + /* Already registered? */ + mutex_lock(&all_stat_sessions_mutex); + list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { + if (node->ts == trace) + return -EINVAL; } - all_stat_sessions = kmalloc(sizeof(struct tracer_stat_session *) * nb, - GFP_KERNEL); - if (!all_stat_sessions) + mutex_unlock(&all_stat_sessions_mutex); + + /* Init the session */ + session = kmalloc(sizeof(struct tracer_stat_session), GFP_KERNEL); + if (!session) return -ENOMEM; - for (i = 0; i < nb; i++) { - session = kmalloc(sizeof(struct tracer_stat_session) * nb, - GFP_KERNEL); - if (!session) - goto free_sessions; + session->ts = trace; + INIT_LIST_HEAD(&session->session_list); + INIT_LIST_HEAD(&session->stat_list); + mutex_init(&session->stat_mutex); + session->file = NULL; - INIT_LIST_HEAD(&session->stat_list); - mutex_init(&session->stat_mutex); - session->ts = &ts[i]; - all_stat_sessions[i] = session; + ret = init_stat_file(session); + if (ret) { + destroy_session(session); + return ret; } - nb_sessions = nb; + + /* Register */ + mutex_lock(&all_stat_sessions_mutex); + list_add_tail(&session->session_list, &all_stat_sessions); + mutex_unlock(&all_stat_sessions_mutex); + return 0; - -free_sessions: - - for (j = 0; j < i; j++) - kfree(all_stat_sessions[i]); - - kfree(all_stat_sessions); - all_stat_sessions = NULL; - - return -ENOMEM; } -static int basic_tracer_stat_checks(struct tracer_stat *ts) +void unregister_stat_tracer(struct tracer_stat *trace) { - int i; + struct tracer_stat_session *node, *tmp; - if (!ts) - return 0; - - for (i = 0; ts[i].name; i++) { - if (!ts[i].stat_start || !ts[i].stat_next || !ts[i].stat_show) - return -EBUSY; + mutex_lock(&all_stat_sessions_mutex); + list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { + if (node->ts == trace) { + list_del(&node->session_list); + destroy_session(node); + break; + } } - return i; + mutex_unlock(&all_stat_sessions_mutex); } + /* * For tracers that don't provide a stat_cmp callback. * This one will force an immediate insertion on tail of @@ -280,63 +295,7 @@ static const struct file_operations tracing_stat_fops = { .release = tracing_stat_release }; - -static void destroy_trace_stat_files(void) -{ - int i; - - if (stat_files) { - for (i = 0; i < nb_sessions; i++) - debugfs_remove(stat_files[i]); - kfree(stat_files); - stat_files = NULL; - } -} - -static void init_trace_stat_files(void) -{ - int i; - - if (!stat_dir || !nb_sessions) - return; - - stat_files = kmalloc(sizeof(struct dentry *) * nb_sessions, GFP_KERNEL); - - if (!stat_files) { - pr_warning("trace stat: not enough memory\n"); - return; - } - - for (i = 0; i < nb_sessions; i++) { - struct tracer_stat_session *session = all_stat_sessions[i]; - stat_files[i] = debugfs_create_file(session->ts->name, 0644, - stat_dir, - session, &tracing_stat_fops); - if (!stat_files[i]) - pr_warning("cannot create %s entry\n", - session->ts->name); - } -} - -void init_tracer_stat(struct tracer *trace) -{ - int nb = basic_tracer_stat_checks(trace->stats); - - destroy_trace_stat_files(); - - if (nb < 0) { - pr_warning("stat tracing: missing stat callback on %s\n", - trace->name); - return; - } - if (!nb) - return; - - init_all_sessions(nb, trace->stats); - init_trace_stat_files(); -} - -static int __init tracing_stat_init(void) +static int tracing_stat_init(void) { struct dentry *d_tracing; @@ -348,4 +307,16 @@ static int __init tracing_stat_init(void) "'trace_stat' entry\n"); return 0; } -fs_initcall(tracing_stat_init); + +static int init_stat_file(struct tracer_stat_session *session) +{ + if (!stat_dir && tracing_stat_init()) + return -ENODEV; + + session->file = debugfs_create_file(session->ts->name, 0644, + stat_dir, + session, &tracing_stat_fops); + if (!session->file) + return -ENOMEM; + return 0; +} diff --git a/kernel/trace/trace_stat.h b/kernel/trace/trace_stat.h new file mode 100644 index 000000000000..202274cf7f3d --- /dev/null +++ b/kernel/trace/trace_stat.h @@ -0,0 +1,31 @@ +#ifndef __TRACE_STAT_H +#define __TRACE_STAT_H + +#include + +/* + * If you want to provide a stat file (one-shot statistics), fill + * an iterator with stat_start/stat_next and a stat_show callbacks. + * The others callbacks are optional. + */ +struct tracer_stat { + /* The name of your stat file */ + const char *name; + /* Iteration over statistic entries */ + void *(*stat_start)(void); + void *(*stat_next)(void *prev, int idx); + /* Compare two entries for stats sorting */ + int (*stat_cmp)(void *p1, void *p2); + /* Print a stat entry */ + int (*stat_show)(struct seq_file *s, void *p); + /* Print the headers of your stat entries */ + int (*stat_headers)(struct seq_file *s); +}; + +/* + * Destroy or create a stat file + */ +extern int register_stat_tracer(struct tracer_stat *trace); +extern void unregister_stat_tracer(struct tracer_stat *trace); + +#endif /* __TRACE_STAT_H */ From e1d8aa9f1dd655a3534b22fcfbecb70cdb125766 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 12 Jan 2009 23:15:46 +0100 Subject: [PATCH 047/413] tracing: add a new workqueue tracer Impact: new tracer The workqueue tracer provides some statistical informations about each cpu workqueue thread such as the number of the works inserted and executed since their creation. It can help to evaluate the amount of work each of them have to perform. For example it can help a developer to decide whether he should choose a per cpu workqueue instead of a singlethreaded one. It only traces statistical informations for now but it will probably later provide event tracing too. Such a tracer could help too, and be improved, to help rt priority sorted workqueue development. To have a snapshot of the workqueues state at any time, just do cat /debugfs/tracing/trace_stat/workqueues Ie: 1 125 125 reiserfs/1 1 0 0 scsi_tgtd/1 1 0 0 aio/1 1 0 0 ata/1 1 114 114 kblockd/1 1 0 0 kintegrityd/1 1 2147 2147 events/1 0 0 0 kpsmoused 0 105 105 reiserfs/0 0 0 0 scsi_tgtd/0 0 0 0 aio/0 0 0 0 ata_aux 0 0 0 ata/0 0 0 0 cqueue 0 0 0 kacpi_notify 0 0 0 kacpid 0 149 149 kblockd/0 0 0 0 kintegrityd/0 0 1000 1000 khelper 0 2270 2270 events/0 Changes in V2: _ Drop the static array based on NR_CPU and dynamically allocate the stat array with num_possible_cpus() and other cpu mask facilities.... _ Trace workqueue insertion at a bit lower level (insert_work instead of queue_work) to handle even the workqueue barriers. Signed-off-by: Frederic Weisbecker Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- include/trace/workqueue.h | 25 +++ kernel/trace/Kconfig | 11 ++ kernel/trace/Makefile | 1 + kernel/trace/trace_workqueue.c | 287 +++++++++++++++++++++++++++++++++ kernel/workqueue.c | 16 +- 5 files changed, 339 insertions(+), 1 deletion(-) create mode 100644 include/trace/workqueue.h create mode 100644 kernel/trace/trace_workqueue.c diff --git a/include/trace/workqueue.h b/include/trace/workqueue.h new file mode 100644 index 000000000000..867829df4571 --- /dev/null +++ b/include/trace/workqueue.h @@ -0,0 +1,25 @@ +#ifndef __TRACE_WORKQUEUE_H +#define __TRACE_WORKQUEUE_H + +#include +#include +#include + +DECLARE_TRACE(workqueue_insertion, + TPPROTO(struct task_struct *wq_thread, struct work_struct *work), + TPARGS(wq_thread, work)); + +DECLARE_TRACE(workqueue_execution, + TPPROTO(struct task_struct *wq_thread, struct work_struct *work), + TPARGS(wq_thread, work)); + +/* Trace the creation of one workqueue thread on a cpu */ +DECLARE_TRACE(workqueue_creation, + TPPROTO(struct task_struct *wq_thread, int cpu), + TPARGS(wq_thread, cpu)); + +DECLARE_TRACE(workqueue_destruction, + TPPROTO(struct task_struct *wq_thread), + TPARGS(wq_thread)); + +#endif /* __TRACE_WORKQUEUE_H */ diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 944239296f13..dde1d46f77e5 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -284,6 +284,17 @@ config KMEMTRACE If unsure, say N. +config WORKQUEUE_TRACER + bool "Trace workqueues" + select TRACING + help + The workqueue tracer provides some statistical informations + about each cpu workqueue thread such as the number of the + works inserted and executed since their creation. It can help + to evaluate the amount of work each of them have to perform. + For example it can help a developer to decide whether he should + choose a per cpu workqueue instead of a singlethreaded one. + config DYNAMIC_FTRACE bool "enable/disable ftrace tracepoints dynamically" diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 05c9182061de..f76d48f3527d 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -36,5 +36,6 @@ obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o obj-$(CONFIG_HW_BRANCH_TRACER) += trace_hw_branches.o obj-$(CONFIG_POWER_TRACER) += trace_power.o obj-$(CONFIG_KMEMTRACE) += kmemtrace.o +obj-$(CONFIG_WORKQUEUE_TRACER) += trace_workqueue.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c new file mode 100644 index 000000000000..f8118d39ca9b --- /dev/null +++ b/kernel/trace/trace_workqueue.c @@ -0,0 +1,287 @@ +/* + * Workqueue statistical tracer. + * + * Copyright (C) 2008 Frederic Weisbecker + * + */ + + +#include +#include +#include "trace_stat.h" +#include "trace.h" + + +/* A cpu workqueue thread */ +struct cpu_workqueue_stats { + struct list_head list; +/* Useful to know if we print the cpu headers */ + bool first_entry; + int cpu; + pid_t pid; +/* Can be inserted from interrupt or user context, need to be atomic */ + atomic_t inserted; +/* + * Don't need to be atomic, works are serialized in a single workqueue thread + * on a single CPU. + */ + unsigned int executed; +}; + +/* List of workqueue threads on one cpu */ +struct workqueue_global_stats { + struct list_head list; + spinlock_t lock; +}; + +/* Don't need a global lock because allocated before the workqueues, and + * never freed. + */ +static struct workqueue_global_stats *all_workqueue_stat; + +/* Insertion of a work */ +static void +probe_workqueue_insertion(struct task_struct *wq_thread, + struct work_struct *work) +{ + int cpu = cpumask_first(&wq_thread->cpus_allowed); + struct cpu_workqueue_stats *node, *next; + unsigned long flags; + + spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + list_for_each_entry_safe(node, next, &all_workqueue_stat[cpu].list, + list) { + if (node->pid == wq_thread->pid) { + atomic_inc(&node->inserted); + goto found; + } + } + pr_debug("trace_workqueue: entry not found\n"); +found: + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); +} + +/* Execution of a work */ +static void +probe_workqueue_execution(struct task_struct *wq_thread, + struct work_struct *work) +{ + int cpu = cpumask_first(&wq_thread->cpus_allowed); + struct cpu_workqueue_stats *node, *next; + unsigned long flags; + + spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + list_for_each_entry_safe(node, next, &all_workqueue_stat[cpu].list, + list) { + if (node->pid == wq_thread->pid) { + node->executed++; + goto found; + } + } + pr_debug("trace_workqueue: entry not found\n"); +found: + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); +} + +/* Creation of a cpu workqueue thread */ +static void probe_workqueue_creation(struct task_struct *wq_thread, int cpu) +{ + struct cpu_workqueue_stats *cws; + unsigned long flags; + + WARN_ON(cpu < 0 || cpu >= num_possible_cpus()); + + /* Workqueues are sometimes created in atomic context */ + cws = kzalloc(sizeof(struct cpu_workqueue_stats), GFP_ATOMIC); + if (!cws) { + pr_warning("trace_workqueue: not enough memory\n"); + return; + } + tracing_record_cmdline(wq_thread); + + INIT_LIST_HEAD(&cws->list); + cws->cpu = cpu; + + cws->pid = wq_thread->pid; + + spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + if (list_empty(&all_workqueue_stat[cpu].list)) + cws->first_entry = true; + list_add_tail(&cws->list, &all_workqueue_stat[cpu].list); + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); +} + +/* Destruction of a cpu workqueue thread */ +static void probe_workqueue_destruction(struct task_struct *wq_thread) +{ + /* Workqueue only execute on one cpu */ + int cpu = cpumask_first(&wq_thread->cpus_allowed); + struct cpu_workqueue_stats *node, *next; + unsigned long flags; + + spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + list_for_each_entry_safe(node, next, &all_workqueue_stat[cpu].list, + list) { + if (node->pid == wq_thread->pid) { + list_del(&node->list); + kfree(node); + goto found; + } + } + + pr_debug("trace_workqueue: don't find workqueue to destroy\n"); +found: + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + +} + +static struct cpu_workqueue_stats *workqueue_stat_start_cpu(int cpu) +{ + unsigned long flags; + struct cpu_workqueue_stats *ret = NULL; + + + spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + + if (!list_empty(&all_workqueue_stat[cpu].list)) + ret = list_entry(all_workqueue_stat[cpu].list.next, + struct cpu_workqueue_stats, list); + + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + + return ret; +} + +static void *workqueue_stat_start(void) +{ + int cpu; + void *ret = NULL; + + for_each_possible_cpu(cpu) { + ret = workqueue_stat_start_cpu(cpu); + if (ret) + return ret; + } + return NULL; +} + +static void *workqueue_stat_next(void *prev, int idx) +{ + struct cpu_workqueue_stats *prev_cws = prev; + int cpu = prev_cws->cpu; + unsigned long flags; + void *ret = NULL; + + spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + if (list_is_last(&prev_cws->list, &all_workqueue_stat[cpu].list)) { + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + for (++cpu ; cpu < num_possible_cpus(); cpu++) { + ret = workqueue_stat_start_cpu(cpu); + if (ret) + return ret; + } + return NULL; + } + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + + return list_entry(prev_cws->list.next, struct cpu_workqueue_stats, + list); +} + +static int workqueue_stat_show(struct seq_file *s, void *p) +{ + struct cpu_workqueue_stats *cws = p; + unsigned long flags; + int cpu = cws->cpu; + + seq_printf(s, "%3d %6d %6u %s\n", cws->cpu, + atomic_read(&cws->inserted), + cws->executed, + trace_find_cmdline(cws->pid)); + + spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + if (&cws->list == all_workqueue_stat[cpu].list.next) + seq_printf(s, "\n"); + spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + + return 0; +} + +static int workqueue_stat_headers(struct seq_file *s) +{ + seq_printf(s, "# CPU INSERTED EXECUTED NAME\n"); + seq_printf(s, "# | | | |\n\n"); + return 0; +} + +struct tracer_stat workqueue_stats __read_mostly = { + .name = "workqueues", + .stat_start = workqueue_stat_start, + .stat_next = workqueue_stat_next, + .stat_show = workqueue_stat_show, + .stat_headers = workqueue_stat_headers +}; + + +int __init stat_workqueue_init(void) +{ + if (register_stat_tracer(&workqueue_stats)) { + pr_warning("Unable to register workqueue stat tracer\n"); + return 1; + } + + return 0; +} +fs_initcall(stat_workqueue_init); + +/* + * Workqueues are created very early, just after pre-smp initcalls. + * So we must register our tracepoints at this stage. + */ +int __init trace_workqueue_early_init(void) +{ + int ret, cpu; + + ret = register_trace_workqueue_insertion(probe_workqueue_insertion); + if (ret) + goto out; + + ret = register_trace_workqueue_execution(probe_workqueue_execution); + if (ret) + goto no_insertion; + + ret = register_trace_workqueue_creation(probe_workqueue_creation); + if (ret) + goto no_execution; + + ret = register_trace_workqueue_destruction(probe_workqueue_destruction); + if (ret) + goto no_creation; + + all_workqueue_stat = kmalloc(sizeof(struct workqueue_global_stats) + * num_possible_cpus(), GFP_KERNEL); + + if (!all_workqueue_stat) { + pr_warning("trace_workqueue: not enough memory\n"); + goto no_creation; + } + + for_each_possible_cpu(cpu) { + spin_lock_init(&all_workqueue_stat[cpu].lock); + INIT_LIST_HEAD(&all_workqueue_stat[cpu].list); + } + + return 0; + +no_creation: + unregister_trace_workqueue_creation(probe_workqueue_creation); +no_execution: + unregister_trace_workqueue_execution(probe_workqueue_execution); +no_insertion: + unregister_trace_workqueue_insertion(probe_workqueue_insertion); +out: + pr_warning("trace_workqueue: unable to trace workqueues\n"); + + return 1; +} +early_initcall(trace_workqueue_early_init); diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 2f445833ae37..1fc2bc20603f 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -33,6 +33,7 @@ #include #include #include +#include /* * The per-CPU workqueue (if single thread, we always use the first @@ -125,9 +126,13 @@ struct cpu_workqueue_struct *get_wq_data(struct work_struct *work) return (void *) (atomic_long_read(&work->data) & WORK_STRUCT_WQ_DATA_MASK); } +DEFINE_TRACE(workqueue_insertion); + static void insert_work(struct cpu_workqueue_struct *cwq, struct work_struct *work, struct list_head *head) { + trace_workqueue_insertion(cwq->thread, work); + set_wq_data(work, cwq); /* * Ensure that we get the right work->data if we see the @@ -259,6 +264,8 @@ int queue_delayed_work_on(int cpu, struct workqueue_struct *wq, } EXPORT_SYMBOL_GPL(queue_delayed_work_on); +DEFINE_TRACE(workqueue_execution); + static void run_workqueue(struct cpu_workqueue_struct *cwq) { spin_lock_irq(&cwq->lock); @@ -284,7 +291,7 @@ static void run_workqueue(struct cpu_workqueue_struct *cwq) */ struct lockdep_map lockdep_map = work->lockdep_map; #endif - + trace_workqueue_execution(cwq->thread, work); cwq->current_work = work; list_del_init(cwq->worklist.next); spin_unlock_irq(&cwq->lock); @@ -765,6 +772,8 @@ init_cpu_workqueue(struct workqueue_struct *wq, int cpu) return cwq; } +DEFINE_TRACE(workqueue_creation); + static int create_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu) { struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 }; @@ -787,6 +796,8 @@ static int create_workqueue_thread(struct cpu_workqueue_struct *cwq, int cpu) sched_setscheduler_nocheck(p, SCHED_FIFO, ¶m); cwq->thread = p; + trace_workqueue_creation(cwq->thread, cpu); + return 0; } @@ -868,6 +879,8 @@ struct workqueue_struct *__create_workqueue_key(const char *name, } EXPORT_SYMBOL_GPL(__create_workqueue_key); +DEFINE_TRACE(workqueue_destruction); + static void cleanup_workqueue_thread(struct cpu_workqueue_struct *cwq) { /* @@ -891,6 +904,7 @@ static void cleanup_workqueue_thread(struct cpu_workqueue_struct *cwq) * checks list_empty(), and a "normal" queue_work() can't use * a dead CPU. */ + trace_workqueue_destruction(cwq->thread); kthread_stop(cwq->thread); cwq->thread = NULL; } From 32632920a788fb13da35b131b77cc4324c38c1c5 Mon Sep 17 00:00:00 2001 From: Uwe Kleine-Koenig Date: Mon, 12 Jan 2009 23:35:50 +0100 Subject: [PATCH 048/413] ftrace, trivial: fix typo "resgister" -> "register" Signed-off-by: Uwe Kleine-Koenig Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 9f536108d3f3..8c1c9c0f4775 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1894,7 +1894,7 @@ int register_ftrace_function(struct ftrace_ops *ops) } /** - * unregister_ftrace_function - unresgister a function for profiling. + * unregister_ftrace_function - unregister a function for profiling. * @ops - ops structure that holds the function to unregister * * Unregister a function that was added to be called by ftrace profiling. From 428aee1460a75197f0190534b4d610450ee59af1 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 14 Jan 2009 12:24:42 -0500 Subject: [PATCH 049/413] trace: print ftrace_dump at KERN_EMERG log level Impact: fix to print out ftrace_dump when expected I was debugging a hard race condition to only find out that after I hit the race, my log level was not at level to show KERN_INFO. The time it took to trigger the race was wasted because I did not capture the trace. Since ftrace_dump is only called from kernel oops (and only when it is set in the kernel command line to do so), or when a developer adds it to their own local tree, the log level of the print should be at KERN_EMERG to make sure the print appears. ftrace_dump is not called by a normal user setup, and will not add extra unwanted print out to the console. There is no reason it should be at KERN_INFO. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 40217fb499ea..408c03f2b8a9 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3076,7 +3076,7 @@ static struct notifier_block trace_die_notifier = { * it if we decide to change what log level the ftrace dump * should be at. */ -#define KERN_TRACE KERN_INFO +#define KERN_TRACE KERN_EMERG static void trace_printk_seq(struct trace_seq *s) From 6f3b34402e7282cde49dff395d7ea462bf33bf50 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 12 Jan 2009 11:06:18 +0800 Subject: [PATCH 050/413] ring_buffer: reset write when reserve buffer fail Impact: reset struct buffer_page.write when interrupt storm if struct buffer_page.write is not reset, any succedent committing will corrupted ring_buffer: static inline void rb_set_commit_to_write(struct ring_buffer_per_cpu *cpu_buffer) { ...... cpu_buffer->commit_page->commit = cpu_buffer->commit_page->write; ...... } when "if (RB_WARN_ON(cpu_buffer, next_page == reader_page))", ring_buffer is disabled, but some reserved buffers may haven't been committed. we need reset struct buffer_page.write. when "if (unlikely(next_page == cpu_buffer->commit_page))", ring_buffer is still available, we should not corrupt it. Signed-off-by: Lai Jiangshan Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 4832ffa5d937..0b9de5a3d699 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1017,12 +1017,8 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, } if (next_page == head_page) { - if (!(buffer->flags & RB_FL_OVERWRITE)) { - /* reset write */ - if (tail <= BUF_PAGE_SIZE) - local_set(&tail_page->write, tail); + if (!(buffer->flags & RB_FL_OVERWRITE)) goto out_unlock; - } /* tail_page has not moved yet? */ if (tail_page == cpu_buffer->tail_page) { @@ -1097,6 +1093,10 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, return event; out_unlock: + /* reset write */ + if (tail <= BUF_PAGE_SIZE) + local_set(&tail_page->write, tail); + __raw_spin_unlock(&cpu_buffer->lock); local_irq_restore(flags); return NULL; From 0ee6b6cf5bdb793b4c68507dd65adf16341aa4ca Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 14 Jan 2009 14:50:19 -0500 Subject: [PATCH 051/413] trace: stop all recording to ring buffer on ftrace_dump Impact: limit ftrace dump output Currently ftrace_dump only calls ftrace_kill that is a fast way to prevent the function tracer functions from being called (just sets a flag and clears the function to call, nothing else). It is better to also turn off any recording to the ring buffers as well. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 408c03f2b8a9..dcb757f70d21 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3110,6 +3110,7 @@ void ftrace_dump(void) dump_ran = 1; /* No turning back! */ + tracing_off(); ftrace_kill(); for_each_tracing_cpu(cpu) { From 4a2b8dda3f8705880ec7408135645602d5590f51 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 14 Jan 2009 13:33:27 -0800 Subject: [PATCH 052/413] tracing/function-graph-tracer: fix a regression while suspend to disk Impact: fix a crash while kernel image restore When the function graph tracer is running and while suspend to disk, some racy and dangerous things happen against this tracer. The current task will save its registers including the stack pointer which contains the return address hooked by the tracer. But the current task will continue to enter other functions after that to save the memory, and then it will store other return addresses, and finally loose the old depth which matches the return address saved in the old stack (during the registers saving). So on image restore, the code will return to wrong addresses. And there are other things: on restore, the task will have it's "current" pointer overwritten during registers restoring....switching from one task to another... That would be insane to try to trace function graphs at these stages. This patch makes the function graph tracer listening on power events, making it's tracing disabled for the current task (the one that performs the hibernation work) while suspend/resume to disk, making the tracing safe during hibernation. Signed-off-by: Frederic Weisbecker Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 27 +++++++++++++++++++++++++++ 1 file changed, 27 insertions(+) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 8c1c9c0f4775..7e9a20b69939 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -17,6 +17,7 @@ #include #include #include +#include #include #include #include @@ -1957,6 +1958,7 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, #ifdef CONFIG_FUNCTION_GRAPH_TRACER static atomic_t ftrace_graph_active; +static struct notifier_block ftrace_suspend_notifier; int ftrace_graph_entry_stub(struct ftrace_graph_ent *trace) { @@ -2035,6 +2037,27 @@ static int start_graph_tracing(void) return ret; } +/* + * Hibernation protection. + * The state of the current task is too much unstable during + * suspend/restore to disk. We want to protect against that. + */ +static int +ftrace_suspend_notifier_call(struct notifier_block *bl, unsigned long state, + void *unused) +{ + switch (state) { + case PM_HIBERNATION_PREPARE: + pause_graph_tracing(); + break; + + case PM_POST_HIBERNATION: + unpause_graph_tracing(); + break; + } + return NOTIFY_DONE; +} + int register_ftrace_graph(trace_func_graph_ret_t retfunc, trace_func_graph_ent_t entryfunc) { @@ -2042,6 +2065,9 @@ int register_ftrace_graph(trace_func_graph_ret_t retfunc, mutex_lock(&ftrace_sysctl_lock); + ftrace_suspend_notifier.notifier_call = ftrace_suspend_notifier_call; + register_pm_notifier(&ftrace_suspend_notifier); + atomic_inc(&ftrace_graph_active); ret = start_graph_tracing(); if (ret) { @@ -2067,6 +2093,7 @@ void unregister_ftrace_graph(void) ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; ftrace_graph_entry = ftrace_graph_entry_stub; ftrace_shutdown(FTRACE_STOP_FUNC_RET); + unregister_pm_notifier(&ftrace_suspend_notifier); mutex_unlock(&ftrace_sysctl_lock); } From 42fab4b2cdc02cf28e2474ccfd75bc9225076590 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 15 Jan 2009 09:30:52 +0800 Subject: [PATCH 053/413] tracing/ftrace: add missing unlock in register_stat_tracer() We should unlock all_stat_sessions_mutex before returning failure. Signed-off-by: Li Zefan Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_stat.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index cb29282b9485..2110cea2ece3 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -73,8 +73,10 @@ int register_stat_tracer(struct tracer_stat *trace) /* Already registered? */ mutex_lock(&all_stat_sessions_mutex); list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { - if (node->ts == trace) + if (node->ts == trace) { + mutex_unlock(&all_stat_sessions_mutex); return -EINVAL; + } } mutex_unlock(&all_stat_sessions_mutex); From 55922173f1f63903b6de03711ab8ff980cbe58d2 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 15 Jan 2009 11:31:21 +0100 Subject: [PATCH 054/413] tracing: trace_stat.c cleanup Impact: cleanup - whitespace / code alignment cleanups - avoid unnecessary forward prototype by reordering functions Signed-off-by: Ingo Molnar --- kernel/trace/trace_stat.c | 147 ++++++++++++++++++-------------------- 1 file changed, 71 insertions(+), 76 deletions(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index 2110cea2ece3..eae9cef39291 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -17,16 +17,16 @@ /* List of stat entries from a tracer */ struct trace_stat_list { - struct list_head list; - void *stat; + struct list_head list; + void *stat; }; /* A stat session is the stats output in one file */ struct tracer_stat_session { struct list_head session_list; - struct tracer_stat *ts; - struct list_head stat_list; - struct mutex stat_mutex; + struct tracer_stat *ts; + struct list_head stat_list; + struct mutex stat_mutex; struct dentry *file; }; @@ -35,7 +35,7 @@ static LIST_HEAD(all_stat_sessions); static DEFINE_MUTEX(all_stat_sessions_mutex); /* The root directory for all stat files */ -static struct dentry *stat_dir; +static struct dentry *stat_dir; static void reset_stat_session(struct tracer_stat_session *session) @@ -56,71 +56,6 @@ static void destroy_session(struct tracer_stat_session *session) kfree(session); } - -static int init_stat_file(struct tracer_stat_session *session); - -int register_stat_tracer(struct tracer_stat *trace) -{ - struct tracer_stat_session *session, *node, *tmp; - int ret; - - if (!trace) - return -EINVAL; - - if (!trace->stat_start || !trace->stat_next || !trace->stat_show) - return -EINVAL; - - /* Already registered? */ - mutex_lock(&all_stat_sessions_mutex); - list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { - if (node->ts == trace) { - mutex_unlock(&all_stat_sessions_mutex); - return -EINVAL; - } - } - mutex_unlock(&all_stat_sessions_mutex); - - /* Init the session */ - session = kmalloc(sizeof(struct tracer_stat_session), GFP_KERNEL); - if (!session) - return -ENOMEM; - - session->ts = trace; - INIT_LIST_HEAD(&session->session_list); - INIT_LIST_HEAD(&session->stat_list); - mutex_init(&session->stat_mutex); - session->file = NULL; - - ret = init_stat_file(session); - if (ret) { - destroy_session(session); - return ret; - } - - /* Register */ - mutex_lock(&all_stat_sessions_mutex); - list_add_tail(&session->session_list, &all_stat_sessions); - mutex_unlock(&all_stat_sessions_mutex); - - return 0; -} - -void unregister_stat_tracer(struct tracer_stat *trace) -{ - struct tracer_stat_session *node, *tmp; - - mutex_lock(&all_stat_sessions_mutex); - list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { - if (node->ts == trace) { - list_del(&node->session_list); - destroy_session(node); - break; - } - } - mutex_unlock(&all_stat_sessions_mutex); -} - - /* * For tracers that don't provide a stat_cmp callback. * This one will force an immediate insertion on tail of @@ -252,10 +187,10 @@ static int stat_seq_show(struct seq_file *s, void *v) } static const struct seq_operations trace_stat_seq_ops = { - .start = stat_seq_start, - .next = stat_seq_next, - .stop = stat_seq_stop, - .show = stat_seq_show + .start = stat_seq_start, + .next = stat_seq_next, + .stop = stat_seq_stop, + .show = stat_seq_show }; /* The session stat is refilled and resorted at each stat file opening */ @@ -275,7 +210,6 @@ static int tracing_stat_open(struct inode *inode, struct file *file) return ret; } - /* * Avoid consuming memory with our now useless list. */ @@ -322,3 +256,64 @@ static int init_stat_file(struct tracer_stat_session *session) return -ENOMEM; return 0; } + +int register_stat_tracer(struct tracer_stat *trace) +{ + struct tracer_stat_session *session, *node, *tmp; + int ret; + + if (!trace) + return -EINVAL; + + if (!trace->stat_start || !trace->stat_next || !trace->stat_show) + return -EINVAL; + + /* Already registered? */ + mutex_lock(&all_stat_sessions_mutex); + list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { + if (node->ts == trace) { + mutex_unlock(&all_stat_sessions_mutex); + return -EINVAL; + } + } + mutex_unlock(&all_stat_sessions_mutex); + + /* Init the session */ + session = kmalloc(sizeof(struct tracer_stat_session), GFP_KERNEL); + if (!session) + return -ENOMEM; + + session->ts = trace; + INIT_LIST_HEAD(&session->session_list); + INIT_LIST_HEAD(&session->stat_list); + mutex_init(&session->stat_mutex); + session->file = NULL; + + ret = init_stat_file(session); + if (ret) { + destroy_session(session); + return ret; + } + + /* Register */ + mutex_lock(&all_stat_sessions_mutex); + list_add_tail(&session->session_list, &all_stat_sessions); + mutex_unlock(&all_stat_sessions_mutex); + + return 0; +} + +void unregister_stat_tracer(struct tracer_stat *trace) +{ + struct tracer_stat_session *node, *tmp; + + mutex_lock(&all_stat_sessions_mutex); + list_for_each_entry_safe(node, tmp, &all_stat_sessions, session_list) { + if (node->ts == trace) { + list_del(&node->session_list); + destroy_session(node); + break; + } + } + mutex_unlock(&all_stat_sessions_mutex); +} From 6c1a99afbda99cd8d8c69d756387041567a13d87 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Thu, 15 Jan 2009 18:05:40 +0800 Subject: [PATCH 055/413] ftrace: fix trace_output Impact: fix bug for handling partial line trace_seq_printf(), seq_print_userip_objs(), ... return 0 -- partial line was written other(>0) -- success duplicate output is also removed in trace_print_raw(). Signed-off-by: Lai Jiangshan Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_output.c | 65 ++++++++++++++++++------------------- kernel/trace/trace_output.h | 4 +-- 2 files changed, 33 insertions(+), 36 deletions(-) diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index df0c25cbed30..4e3ad36b117c 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -440,9 +440,9 @@ trace_fn_raw(struct trace_seq *s, struct trace_entry *entry, int flags) trace_assign_type(field, entry); - if (trace_seq_printf(s, "%x %x\n", - field->ip, - field->parent_ip)) + if (!trace_seq_printf(s, "%x %x\n", + field->ip, + field->parent_ip)) return TRACE_TYPE_PARTIAL_LINE; return 0; @@ -497,14 +497,14 @@ trace_ctxwake_print(struct trace_seq *s, struct trace_entry *entry, int flags, T = task_state_char(field->next_state); S = task_state_char(field->prev_state); comm = trace_find_cmdline(field->next_pid); - if (trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", - field->prev_pid, - field->prev_prio, - S, delim, - field->next_cpu, - field->next_pid, - field->next_prio, - T, comm)) + if (!trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", + field->prev_pid, + field->prev_prio, + S, delim, + field->next_cpu, + field->next_pid, + field->next_prio, + T, comm)) return TRACE_TYPE_PARTIAL_LINE; return 0; @@ -534,14 +534,14 @@ trace_ctxwake_raw(struct trace_seq *s, struct trace_entry *entry, int flags, if (!S) task_state_char(field->prev_state); T = task_state_char(field->next_state); - if (trace_seq_printf(s, "%d %d %c %d %d %d %c\n", - field->prev_pid, - field->prev_prio, - S, - field->next_cpu, - field->next_pid, - field->next_prio, - T)) + if (!trace_seq_printf(s, "%d %d %c %d %d %d %c\n", + field->prev_pid, + field->prev_prio, + S, + field->next_cpu, + field->next_pid, + field->next_prio, + T)) return TRACE_TYPE_PARTIAL_LINE; return 0; @@ -639,10 +639,10 @@ trace_special_print(struct trace_seq *s, struct trace_entry *entry, int flags) trace_assign_type(field, entry); - if (trace_seq_printf(s, "# %ld %ld %ld\n", - field->arg1, - field->arg2, - field->arg3)) + if (!trace_seq_printf(s, "# %ld %ld %ld\n", + field->arg1, + field->arg2, + field->arg3)) return TRACE_TYPE_PARTIAL_LINE; return 0; @@ -697,13 +697,13 @@ trace_stack_print(struct trace_seq *s, struct trace_entry *entry, int flags) for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { if (i) { - if (trace_seq_puts(s, " <= ")) + if (!trace_seq_puts(s, " <= ")) goto partial; - if (seq_print_ip_sym(s, field->caller[i], flags)) + if (!seq_print_ip_sym(s, field->caller[i], flags)) goto partial; } - if (trace_seq_puts(s, "\n")) + if (!trace_seq_puts(s, "\n")) goto partial; } @@ -731,10 +731,10 @@ trace_user_stack_print(struct trace_seq *s, struct trace_entry *entry, trace_assign_type(field, entry); - if (seq_print_userip_objs(field, s, flags)) + if (!seq_print_userip_objs(field, s, flags)) goto partial; - if (trace_seq_putc(s, '\n')) + if (!trace_seq_putc(s, '\n')) goto partial; return 0; @@ -760,10 +760,10 @@ trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) trace_assign_type(field, entry); - if (seq_print_ip_sym(s, field->ip, flags)) + if (!seq_print_ip_sym(s, field->ip, flags)) goto partial; - if (trace_seq_printf(s, ": %s", field->buf)) + if (!trace_seq_printf(s, ": %s", field->buf)) goto partial; return 0; @@ -779,10 +779,7 @@ trace_print_raw(struct trace_seq *s, struct trace_entry *entry, int flags) trace_assign_type(field, entry); - if (seq_print_ip_sym(s, field->ip, flags)) - goto partial; - - if (trace_seq_printf(s, "# %lx %s", field->ip, field->buf)) + if (!trace_seq_printf(s, "# %lx %s", field->ip, field->buf)) goto partial; return 0; diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index ecab4ea4a4fd..b2c14615e0cd 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -45,14 +45,14 @@ trace_nop_print(struct trace_seq *s, struct trace_entry *entry, int flags); #define SEQ_PUT_FIELD_RET(s, x) \ do { \ if (!trace_seq_putmem(s, &(x), sizeof(x))) \ - return 0; \ + return TRACE_TYPE_PARTIAL_LINE; \ } while (0) #define SEQ_PUT_HEX_FIELD_RET(s, x) \ do { \ BUILD_BUG_ON(sizeof(x) > MAX_MEMHEX_BYTES); \ if (!trace_seq_putmem_hex(s, &(x), sizeof(x))) \ - return 0; \ + return TRACE_TYPE_PARTIAL_LINE; \ } while (0) #endif From 5361499101306cfb776c3cfa0f69d0479bc63868 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 15 Jan 2009 19:12:40 -0500 Subject: [PATCH 056/413] ftrace: add stack trace to function tracer Impact: new feature to stack trace any function Chris Mason asked about being able to pick and choose a function and get a stack trace from it. This feature enables his request. # echo io_schedule > /debug/tracing/set_ftrace_filter # echo function > /debug/tracing/current_tracer # echo func_stack_trace > /debug/tracing/trace_options Produces the following in /debug/tracing/trace: kjournald-702 [001] 135.673060: io_schedule <-sync_buffer kjournald-702 [002] 135.673671: <= sync_buffer <= __wait_on_bit <= out_of_line_wait_on_bit <= __wait_on_buffer <= sync_dirty_buffer <= journal_commit_transaction <= kjournald Note, be careful about turning this on without filtering the functions. You may find that you have a 10 second lag between typing and seeing what you typed. This is why the stack trace for the function tracer does not use the same stack_trace flag as the other tracers use. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 26 +++++++---- kernel/trace/trace.h | 7 +++ kernel/trace/trace_functions.c | 84 ++++++++++++++++++++++++++++++++++ 3 files changed, 108 insertions(+), 9 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index dcb757f70d21..3c54cb125228 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -835,10 +835,10 @@ ftrace(struct trace_array *tr, struct trace_array_cpu *data, trace_function(tr, data, ip, parent_ip, flags, pc); } -static void ftrace_trace_stack(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long flags, - int skip, int pc) +static void __ftrace_trace_stack(struct trace_array *tr, + struct trace_array_cpu *data, + unsigned long flags, + int skip, int pc) { #ifdef CONFIG_STACKTRACE struct ring_buffer_event *event; @@ -846,9 +846,6 @@ static void ftrace_trace_stack(struct trace_array *tr, struct stack_trace trace; unsigned long irq_flags; - if (!(trace_flags & TRACE_ITER_STACKTRACE)) - return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq_flags); if (!event) @@ -869,12 +866,23 @@ static void ftrace_trace_stack(struct trace_array *tr, #endif } +static void ftrace_trace_stack(struct trace_array *tr, + struct trace_array_cpu *data, + unsigned long flags, + int skip, int pc) +{ + if (!(trace_flags & TRACE_ITER_STACKTRACE)) + return; + + __ftrace_trace_stack(tr, data, flags, skip, pc); +} + void __trace_stack(struct trace_array *tr, struct trace_array_cpu *data, unsigned long flags, - int skip) + int skip, int pc) { - ftrace_trace_stack(tr, data, flags, skip, preempt_count()); + __ftrace_trace_stack(tr, data, flags, skip, pc); } static void ftrace_trace_userstack(struct trace_array *tr, diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 79c872100dd5..bf39a369e4b3 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -457,6 +457,11 @@ void update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu); void update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu); +void __trace_stack(struct trace_array *tr, + struct trace_array_cpu *data, + unsigned long flags, + int skip, int pc); + extern cycle_t ftrace_now(int cpu); #ifdef CONFIG_FUNCTION_TRACER @@ -467,6 +472,8 @@ void tracing_stop_function_trace(void); # define tracing_stop_function_trace() do { } while (0) #endif +extern int ftrace_function_enabled; + #ifdef CONFIG_CONTEXT_SWITCH_TRACER typedef void (*tracer_switch_func_t)(void *private, diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 9236d7e25a16..3a5fa08cedb0 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -16,6 +16,8 @@ #include "trace.h" +static struct trace_array *func_trace; + static void start_function_trace(struct trace_array *tr) { tr->cpu = get_cpu(); @@ -34,6 +36,7 @@ static void stop_function_trace(struct trace_array *tr) static int function_trace_init(struct trace_array *tr) { + func_trace = tr; start_function_trace(tr); return 0; } @@ -48,12 +51,93 @@ static void function_trace_start(struct trace_array *tr) tracing_reset_online_cpus(tr); } +static void +function_stack_trace_call(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = func_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu; + int pc; + + if (unlikely(!ftrace_function_enabled)) + return; + + /* + * Need to use raw, since this must be called before the + * recursive protection is performed. + */ + local_irq_save(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + + if (likely(disabled == 1)) { + pc = preempt_count(); + /* + * skip over 5 funcs: + * __ftrace_trace_stack, + * __trace_stack, + * function_stack_trace_call + * ftrace_list_func + * ftrace_call + */ + __trace_stack(tr, data, flags, 5, pc); + } + + atomic_dec(&data->disabled); + local_irq_restore(flags); +} + +static struct ftrace_ops trace_stack_ops __read_mostly = +{ + .func = function_stack_trace_call, +}; + +/* Our two options */ +enum { + TRACE_FUNC_OPT_STACK = 0x1, +}; + +static struct tracer_opt func_opts[] = { +#ifdef CONFIG_STACKTRACE + { TRACER_OPT(func_stack_trace, TRACE_FUNC_OPT_STACK) }, +#endif + { } /* Always set a last empty entry */ +}; + +static struct tracer_flags func_flags = { + .val = 0, /* By default: all flags disabled */ + .opts = func_opts +}; + +static int func_set_flag(u32 old_flags, u32 bit, int set) +{ + if (bit == TRACE_FUNC_OPT_STACK) { + /* do nothing if already set */ + if (!!set == !!(func_flags.val & TRACE_FUNC_OPT_STACK)) + return 0; + + if (set) + register_ftrace_function(&trace_stack_ops); + else + unregister_ftrace_function(&trace_stack_ops); + + return 0; + } + + return -EINVAL; +} + static struct tracer function_trace __read_mostly = { .name = "function", .init = function_trace_init, .reset = function_trace_reset, .start = function_trace_start, + .flags = &func_flags, + .set_flag = func_set_flag, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_function, #endif From bb3c3c95f330f7bf16e33b002e48882616089db1 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 15 Jan 2009 20:40:23 -0500 Subject: [PATCH 057/413] ftrace: move function tracer functions out of trace.c Impact: clean up of trace.c The function tracer functions were put in trace.c because it needed to share static variables that were in trace.c. Since then, those variables have become global for various reasons. This patch moves the function tracer functions into trace_function.c where they belong. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 84 ---------------------------------- kernel/trace/trace_functions.c | 84 +++++++++++++++++++++++++++++++++- 2 files changed, 83 insertions(+), 85 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3c54cb125228..2585ffb6c6b5 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1046,65 +1046,6 @@ ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3) local_irq_restore(flags); } -#ifdef CONFIG_FUNCTION_TRACER -static void -function_trace_call_preempt_only(unsigned long ip, unsigned long parent_ip) -{ - struct trace_array *tr = &global_trace; - struct trace_array_cpu *data; - unsigned long flags; - long disabled; - int cpu, resched; - int pc; - - if (unlikely(!ftrace_function_enabled)) - return; - - pc = preempt_count(); - resched = ftrace_preempt_disable(); - local_save_flags(flags); - cpu = raw_smp_processor_id(); - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - - if (likely(disabled == 1)) - trace_function(tr, data, ip, parent_ip, flags, pc); - - atomic_dec(&data->disabled); - ftrace_preempt_enable(resched); -} - -static void -function_trace_call(unsigned long ip, unsigned long parent_ip) -{ - struct trace_array *tr = &global_trace; - struct trace_array_cpu *data; - unsigned long flags; - long disabled; - int cpu; - int pc; - - if (unlikely(!ftrace_function_enabled)) - return; - - /* - * Need to use raw, since this must be called before the - * recursive protection is performed. - */ - local_irq_save(flags); - cpu = raw_smp_processor_id(); - data = tr->data[cpu]; - disabled = atomic_inc_return(&data->disabled); - - if (likely(disabled == 1)) { - pc = preempt_count(); - trace_function(tr, data, ip, parent_ip, flags, pc); - } - - atomic_dec(&data->disabled); - local_irq_restore(flags); -} - #ifdef CONFIG_FUNCTION_GRAPH_TRACER int trace_graph_entry(struct ftrace_graph_ent *trace) { @@ -1162,31 +1103,6 @@ void trace_graph_return(struct ftrace_graph_ret *trace) } #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ -static struct ftrace_ops trace_ops __read_mostly = -{ - .func = function_trace_call, -}; - -void tracing_start_function_trace(void) -{ - ftrace_function_enabled = 0; - - if (trace_flags & TRACE_ITER_PREEMPTONLY) - trace_ops.func = function_trace_call_preempt_only; - else - trace_ops.func = function_trace_call; - - register_ftrace_function(&trace_ops); - ftrace_function_enabled = 1; -} - -void tracing_stop_function_trace(void) -{ - ftrace_function_enabled = 0; - unregister_ftrace_function(&trace_ops); -} -#endif - enum trace_file_type { TRACE_FILE_LAT_FMT = 1, TRACE_FILE_ANNOTATE = 2, diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 3a5fa08cedb0..2dce3c7370d1 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -20,6 +20,7 @@ static struct trace_array *func_trace; static void start_function_trace(struct trace_array *tr) { + func_trace = tr; tr->cpu = get_cpu(); tracing_reset_online_cpus(tr); put_cpu(); @@ -36,7 +37,6 @@ static void stop_function_trace(struct trace_array *tr) static int function_trace_init(struct trace_array *tr) { - func_trace = tr; start_function_trace(tr); return 0; } @@ -51,6 +51,64 @@ static void function_trace_start(struct trace_array *tr) tracing_reset_online_cpus(tr); } +static void +function_trace_call_preempt_only(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = func_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu, resched; + int pc; + + if (unlikely(!ftrace_function_enabled)) + return; + + pc = preempt_count(); + resched = ftrace_preempt_disable(); + local_save_flags(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + + if (likely(disabled == 1)) + trace_function(tr, data, ip, parent_ip, flags, pc); + + atomic_dec(&data->disabled); + ftrace_preempt_enable(resched); +} + +static void +function_trace_call(unsigned long ip, unsigned long parent_ip) +{ + struct trace_array *tr = func_trace; + struct trace_array_cpu *data; + unsigned long flags; + long disabled; + int cpu; + int pc; + + if (unlikely(!ftrace_function_enabled)) + return; + + /* + * Need to use raw, since this must be called before the + * recursive protection is performed. + */ + local_irq_save(flags); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + disabled = atomic_inc_return(&data->disabled); + + if (likely(disabled == 1)) { + pc = preempt_count(); + trace_function(tr, data, ip, parent_ip, flags, pc); + } + + atomic_dec(&data->disabled); + local_irq_restore(flags); +} + static void function_stack_trace_call(unsigned long ip, unsigned long parent_ip) { @@ -90,6 +148,30 @@ function_stack_trace_call(unsigned long ip, unsigned long parent_ip) local_irq_restore(flags); } + +static struct ftrace_ops trace_ops __read_mostly = +{ + .func = function_trace_call, +}; + +void tracing_start_function_trace(void) +{ + ftrace_function_enabled = 0; + + if (trace_flags & TRACE_ITER_PREEMPTONLY) + trace_ops.func = function_trace_call_preempt_only; + else + trace_ops.func = function_trace_call; + + register_ftrace_function(&trace_ops); + ftrace_function_enabled = 1; +} + +void tracing_stop_function_trace(void) +{ + ftrace_function_enabled = 0; + unregister_ftrace_function(&trace_ops); +} static struct ftrace_ops trace_stack_ops __read_mostly = { .func = function_stack_trace_call, From c37abc5515b5ed5b1d2134d2deaead492d9f92a2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 15 Jan 2009 20:50:54 -0500 Subject: [PATCH 058/413] trace: add gcc printf check to trace_seq_printf Andrew Morton suggested adding a printf checker to trace_seq_printf since there are a number of users that have improper format arguments. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_output.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index b2c14615e0cd..1cbab5e3dc99 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -16,7 +16,8 @@ struct trace_event { trace_print_func binary; }; -extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...); +extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...) + __attribute__ ((format (printf, 2, 3))); extern int seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags); From 5e4abc9839191e213965e0f1dbf36e2e44356c3a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 15 Jan 2009 21:00:50 -0500 Subject: [PATCH 059/413] trace: clean up format errors in calls to trace_seq_printf After adding the printf format checking for trace_seq_printf, several warnings now show up. This patch cleans them up. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/kmemtrace.c | 4 ++-- kernel/trace/trace_mmiotrace.c | 13 +++++++------ kernel/trace/trace_output.c | 2 +- 3 files changed, 10 insertions(+), 9 deletions(-) diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c index faaa5ae7e75a..7ebc58cee3bd 100644 --- a/kernel/trace/kmemtrace.c +++ b/kernel/trace/kmemtrace.c @@ -139,12 +139,12 @@ kmemtrace_print_alloc_compress(struct trace_iterator *iter, return TRACE_TYPE_PARTIAL_LINE; /* Requested */ - ret = trace_seq_printf(s, "%4d ", entry->bytes_req); + ret = trace_seq_printf(s, "%4ld ", entry->bytes_req); if (!ret) return TRACE_TYPE_PARTIAL_LINE; /* Allocated */ - ret = trace_seq_printf(s, "%4d ", entry->bytes_alloc); + ret = trace_seq_printf(s, "%4ld ", entry->bytes_alloc); if (!ret) return TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index 621c8c3f3139..ec78e244242e 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -184,21 +184,22 @@ static enum print_line_t mmio_print_rw(struct trace_iterator *iter) switch (rw->opcode) { case MMIO_READ: ret = trace_seq_printf(s, - "R %d %lu.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", + "R %d %u.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", rw->width, secs, usec_rem, rw->map_id, (unsigned long long)rw->phys, rw->value, rw->pc, 0); break; case MMIO_WRITE: ret = trace_seq_printf(s, - "W %d %lu.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", + "W %d %u.%06lu %d 0x%llx 0x%lx 0x%lx %d\n", rw->width, secs, usec_rem, rw->map_id, (unsigned long long)rw->phys, rw->value, rw->pc, 0); break; case MMIO_UNKNOWN_OP: ret = trace_seq_printf(s, - "UNKNOWN %lu.%06lu %d 0x%llx %02x,%02x,%02x 0x%lx %d\n", + "UNKNOWN %u.%06lu %d 0x%llx %02lx,%02lx," + "%02lx 0x%lx %d\n", secs, usec_rem, rw->map_id, (unsigned long long)rw->phys, (rw->value >> 16) & 0xff, (rw->value >> 8) & 0xff, @@ -230,14 +231,14 @@ static enum print_line_t mmio_print_map(struct trace_iterator *iter) switch (m->opcode) { case MMIO_PROBE: ret = trace_seq_printf(s, - "MAP %lu.%06lu %d 0x%llx 0x%lx 0x%lx 0x%lx %d\n", + "MAP %u.%06lu %d 0x%llx 0x%lx 0x%lx 0x%lx %d\n", secs, usec_rem, m->map_id, (unsigned long long)m->phys, m->virt, m->len, 0UL, 0); break; case MMIO_UNPROBE: ret = trace_seq_printf(s, - "UNMAP %lu.%06lu %d 0x%lx %d\n", + "UNMAP %u.%06lu %d 0x%lx %d\n", secs, usec_rem, m->map_id, 0UL, 0); break; default: @@ -261,7 +262,7 @@ static enum print_line_t mmio_print_mark(struct trace_iterator *iter) int ret; /* The trailing newline must be in the message. */ - ret = trace_seq_printf(s, "MARK %lu.%06lu %s", secs, usec_rem, msg); + ret = trace_seq_printf(s, "MARK %u.%06lu %s", secs, usec_rem, msg); if (!ret) return TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 4e3ad36b117c..1a4e144a9f8f 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -440,7 +440,7 @@ trace_fn_raw(struct trace_seq *s, struct trace_entry *entry, int flags) trace_assign_type(field, entry); - if (!trace_seq_printf(s, "%x %x\n", + if (!trace_seq_printf(s, "%lx %lx\n", field->ip, field->parent_ip)) return TRACE_TYPE_PARTIAL_LINE; From 3eb36aa05329a47cbe201c151fd0024a4a3649cd Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 15 Jan 2009 22:21:43 -0500 Subject: [PATCH 060/413] ftrace: combine stack trace in function call Impact: less likely to interleave function and stack traces This patch does replaces the separate stack trace on function with a record function and stack trace together. This will switch between the function only recording to a function and stack recording. Also some whitespace fix ups as well. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions.c | 61 ++++++++++++++++++++-------------- 1 file changed, 36 insertions(+), 25 deletions(-) diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 2dce3c7370d1..61d0b73dabf5 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -133,6 +133,7 @@ function_stack_trace_call(unsigned long ip, unsigned long parent_ip) if (likely(disabled == 1)) { pc = preempt_count(); + trace_function(tr, data, ip, parent_ip, flags, pc); /* * skip over 5 funcs: * __ftrace_trace_stack, @@ -154,24 +155,6 @@ static struct ftrace_ops trace_ops __read_mostly = .func = function_trace_call, }; -void tracing_start_function_trace(void) -{ - ftrace_function_enabled = 0; - - if (trace_flags & TRACE_ITER_PREEMPTONLY) - trace_ops.func = function_trace_call_preempt_only; - else - trace_ops.func = function_trace_call; - - register_ftrace_function(&trace_ops); - ftrace_function_enabled = 1; -} - -void tracing_stop_function_trace(void) -{ - ftrace_function_enabled = 0; - unregister_ftrace_function(&trace_ops); -} static struct ftrace_ops trace_stack_ops __read_mostly = { .func = function_stack_trace_call, @@ -194,6 +177,31 @@ static struct tracer_flags func_flags = { .opts = func_opts }; +void tracing_start_function_trace(void) +{ + ftrace_function_enabled = 0; + + if (trace_flags & TRACE_ITER_PREEMPTONLY) + trace_ops.func = function_trace_call_preempt_only; + else + trace_ops.func = function_trace_call; + + if (func_flags.val & TRACE_FUNC_OPT_STACK) + register_ftrace_function(&trace_stack_ops); + else + register_ftrace_function(&trace_ops); + + ftrace_function_enabled = 1; +} + +void tracing_stop_function_trace(void) +{ + ftrace_function_enabled = 0; + /* OK if they are not registered */ + unregister_ftrace_function(&trace_stack_ops); + unregister_ftrace_function(&trace_ops); +} + static int func_set_flag(u32 old_flags, u32 bit, int set) { if (bit == TRACE_FUNC_OPT_STACK) { @@ -201,10 +209,13 @@ static int func_set_flag(u32 old_flags, u32 bit, int set) if (!!set == !!(func_flags.val & TRACE_FUNC_OPT_STACK)) return 0; - if (set) + if (set) { + unregister_ftrace_function(&trace_ops); register_ftrace_function(&trace_stack_ops); - else + } else { unregister_ftrace_function(&trace_stack_ops); + register_ftrace_function(&trace_ops); + } return 0; } @@ -214,14 +225,14 @@ static int func_set_flag(u32 old_flags, u32 bit, int set) static struct tracer function_trace __read_mostly = { - .name = "function", - .init = function_trace_init, - .reset = function_trace_reset, - .start = function_trace_start, + .name = "function", + .init = function_trace_init, + .reset = function_trace_reset, + .start = function_trace_start, .flags = &func_flags, .set_flag = func_set_flag, #ifdef CONFIG_FTRACE_SELFTEST - .selftest = trace_selftest_startup_function, + .selftest = trace_selftest_startup_function, #endif }; From a225cdd263f340c864febb1992802fb5b08bc328 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 15 Jan 2009 23:06:03 -0500 Subject: [PATCH 061/413] ftrace: remove static from function tracer functions Impact: clean up After reorganizing the functions in trace.c and trace_function.c, they no longer need to be in global context. This patch makes the functions and one variable into static. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 3 --- kernel/trace/trace.h | 10 ---------- kernel/trace/trace_functions.c | 10 ++++++++-- 3 files changed, 8 insertions(+), 15 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 2585ffb6c6b5..7de6a94063dd 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -187,9 +187,6 @@ int tracing_is_enabled(void) return tracer_enabled; } -/* function tracing enabled */ -int ftrace_function_enabled; - /* * trace_buf_size is the size in bytes that is allocated * for a buffer. Note, the number of bytes is always rounded diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index bf39a369e4b3..54b72781e920 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -464,16 +464,6 @@ void __trace_stack(struct trace_array *tr, extern cycle_t ftrace_now(int cpu); -#ifdef CONFIG_FUNCTION_TRACER -void tracing_start_function_trace(void); -void tracing_stop_function_trace(void); -#else -# define tracing_start_function_trace() do { } while (0) -# define tracing_stop_function_trace() do { } while (0) -#endif - -extern int ftrace_function_enabled; - #ifdef CONFIG_CONTEXT_SWITCH_TRACER typedef void (*tracer_switch_func_t)(void *private, diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 61d0b73dabf5..b3a320f8aba7 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -16,8 +16,14 @@ #include "trace.h" +/* function tracing enabled */ +static int ftrace_function_enabled; + static struct trace_array *func_trace; +static void tracing_start_function_trace(void); +static void tracing_stop_function_trace(void); + static void start_function_trace(struct trace_array *tr) { func_trace = tr; @@ -177,7 +183,7 @@ static struct tracer_flags func_flags = { .opts = func_opts }; -void tracing_start_function_trace(void) +static void tracing_start_function_trace(void) { ftrace_function_enabled = 0; @@ -194,7 +200,7 @@ void tracing_start_function_trace(void) ftrace_function_enabled = 1; } -void tracing_stop_function_trace(void) +static void tracing_stop_function_trace(void) { ftrace_function_enabled = 0; /* OK if they are not registered */ From 745b1626dd71ce9661a05ea4db57859ed5c773d2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 15 Jan 2009 23:40:11 -0500 Subject: [PATCH 062/413] trace: set max latency variable to zero on default Impact: trace max latencies on start of latency tracing This patch sets the max latency to zero whenever one of the irq variant tracers or the wakeup tracer is set to current tracer. Most developers expect to see output when starting up a latency tracer. But since the max_latency is already set to max, and it takes a latency greater than max_latency to be recorded, there is no trace. This is not the expected behavior and has even confused myself. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 2 +- kernel/trace/trace_irqsoff.c | 1 + kernel/trace/trace_sched_wakeup.c | 1 + 3 files changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 7de6a94063dd..220c264e3111 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -41,7 +41,7 @@ #define TRACE_BUFFER_FLAGS (RB_FL_OVERWRITE) -unsigned long __read_mostly tracing_max_latency = (cycle_t)ULONG_MAX; +unsigned long __read_mostly tracing_max_latency; unsigned long __read_mostly tracing_thresh; /* diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 7c2e326bbc8b..62a78d943534 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -380,6 +380,7 @@ static void stop_irqsoff_tracer(struct trace_array *tr) static void __irqsoff_tracer_init(struct trace_array *tr) { + tracing_max_latency = 0; irqsoff_trace = tr; /* make sure that the tracer is visible */ smp_wmb(); diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 43586b689e31..42ae1e77b6b3 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -333,6 +333,7 @@ static void stop_wakeup_tracer(struct trace_array *tr) static int wakeup_tracer_init(struct trace_array *tr) { + tracing_max_latency = 0; wakeup_trace = tr; start_wakeup_tracer(tr); return 0; From b43f70933e7753a284733d5ae355f6778bd118ce Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 16 Jan 2009 23:18:31 -0500 Subject: [PATCH 063/413] ftrace: test for running of recordmcount.pl twice on an object Impact: fix failure of dynamic function tracer selftest In a course of development, a developer does several makes on their kernel. Sometimes, the make might do something abnormal. In the case of running the recordmcount.pl script on an object twice, the script will duplicate all the calls to mcount in the __mcount_loc section. On boot up, the dynamic function tracer is careful when it modifies code, and performs several consistency checks. One is to not modify the call site if it is not what it expects it to be. If a function call site is listed twice, the first entry will convert the site to a nop, and the second will fail because it expected to see a call to mcount, but instead it sees a nop. Thus, the function tracer is disabled. Eric Sesterhenn reported seeing: [ 1.055440] ftrace: converting mcount calls to 0f 1f 44 00 00 [ 1.055568] ftrace: allocating 29418 entries in 116 pages [ 1.061000] ------------[ cut here ]------------ [ 1.061000] WARNING: at kernel/trace/ftrace.c:441 [...] [ 1.060000] ---[ end trace 4eaa2a86a8e2da23 ]--- [ 1.060000] ftrace failed to modify [] check_corruption+0x3/0x2d [ 1.060000] actual: 0f:1f:44:00:00 This warning shows that check_corruption+0x3 already had a nop in its place (0x0f1f440000). After compiling another kernel the problem went away. Later Eric Paris notice the same type of issue. Luckily, he saved the vmlinux file that caused it. In the file we found a bunch of duplicate mcount call site records, which lead us to the script. Perhaps this problem only happens to people named Eric. This patch changes the script to test if the __mcount_loc already exists in the object file, and if it does, it will print out an error message and kill the compile. Reported-by: Eric Sesterhenn Reported-by: Eric Paris Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- scripts/recordmcount.pl | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) diff --git a/scripts/recordmcount.pl b/scripts/recordmcount.pl index 2ded5c8e6b85..409596eca124 100755 --- a/scripts/recordmcount.pl +++ b/scripts/recordmcount.pl @@ -101,7 +101,7 @@ $P =~ s@.*/@@g; my $V = '0.1'; if ($#ARGV < 7) { - print "usage: $P arch objdump objcopy cc ld nm rm mv is_module inputfile\n"; + print "usage: $P arch bits objdump objcopy cc ld nm rm mv is_module inputfile\n"; print "version: $V\n"; exit(1); } @@ -275,7 +275,6 @@ if (!$found_version) { "\tDisabling local function references.\n"; } - # # Step 1: find all the local (static functions) and weak symbols. # 't' is local, 'w/W' is weak (we never use a weak function) @@ -343,13 +342,16 @@ sub update_funcs # # Step 2: find the sections and mcount call sites # -open(IN, "$objdump -dr $inputfile|") || die "error running $objdump"; +open(IN, "$objdump -hdr $inputfile|") || die "error running $objdump"; my $text; +my $read_headers = 1; + while () { # is it a section? if (/$section_regex/) { + $read_headers = 0; # Only record text sections that we know are safe if (defined($text_sections{$1})) { @@ -383,6 +385,19 @@ while () { $ref_func = $text; } } + } elsif ($read_headers && /$mcount_section/) { + # + # Somehow the make process can execute this script on an + # object twice. If it does, we would duplicate the mcount + # section and it will cause the function tracer self test + # to fail. Check if the mcount section exists, and if it does, + # warn and exit. + # + print STDERR "ERROR: $mcount_section already in $inputfile\n" . + "\tThis may be an indication that your build is corrupted.\n" . + "\tDelete $inputfile and try again. If the same object file\n" . + "\tstill causes an issue, then disable CONFIG_DYNAMIC_FTRACE.\n"; + exit(-1); } # is this a call site to mcount? If so, record it to print later From 5c5317de147e9b38ea9c4cbdc2d15bed7648d036 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Mon, 19 Jan 2009 10:26:53 +0100 Subject: [PATCH 064/413] x86, ftrace, hw-branch-tracer: support hotplug cpus Support hotplug cpus. Reported-by: Andi Kleen Signed-off-by: Markus Metzger Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_hw_branches.c | 125 ++++++++++++++++++++++++++----- 1 file changed, 108 insertions(+), 17 deletions(-) diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index df21c1e72b95..398195397c75 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -1,7 +1,8 @@ /* * h/w branch tracer for x86 based on bts * - * Copyright (C) 2008 Markus Metzger + * Copyright (C) 2008-2009 Intel Corporation. + * Markus Metzger , 2008-2009 * */ @@ -10,6 +11,9 @@ #include #include #include +#include +#include +#include #include @@ -19,13 +23,31 @@ #define SIZEOF_BTS (1 << 13) +/* The tracer mutex protects the below per-cpu tracer array. + It needs to be held to: + - start tracing on all cpus + - stop tracing on all cpus + - start tracing on a single hotplug cpu + - stop tracing on a single hotplug cpu + - read the trace from all cpus + - read the trace from a single cpu +*/ +static DEFINE_MUTEX(bts_tracer_mutex); static DEFINE_PER_CPU(struct bts_tracer *, tracer); static DEFINE_PER_CPU(unsigned char[SIZEOF_BTS], buffer); #define this_tracer per_cpu(tracer, smp_processor_id()) #define this_buffer per_cpu(buffer, smp_processor_id()) +static int __read_mostly trace_hw_branches_enabled; + +/* + * Start tracing on the current cpu. + * The argument is ignored. + * + * pre: bts_tracer_mutex must be locked. + */ static void bts_trace_start_cpu(void *arg) { if (this_tracer) @@ -43,14 +65,20 @@ static void bts_trace_start_cpu(void *arg) static void bts_trace_start(struct trace_array *tr) { - int cpu; + mutex_lock(&bts_tracer_mutex); - tracing_reset_online_cpus(tr); + on_each_cpu(bts_trace_start_cpu, NULL, 1); + trace_hw_branches_enabled = 1; - for_each_cpu(cpu, cpu_possible_mask) - smp_call_function_single(cpu, bts_trace_start_cpu, NULL, 1); + mutex_unlock(&bts_tracer_mutex); } +/* + * Start tracing on the current cpu. + * The argument is ignored. + * + * pre: bts_tracer_mutex must be locked. + */ static void bts_trace_stop_cpu(void *arg) { if (this_tracer) { @@ -61,20 +89,58 @@ static void bts_trace_stop_cpu(void *arg) static void bts_trace_stop(struct trace_array *tr) { - int cpu; + mutex_lock(&bts_tracer_mutex); - for_each_cpu(cpu, cpu_possible_mask) - smp_call_function_single(cpu, bts_trace_stop_cpu, NULL, 1); + trace_hw_branches_enabled = 0; + on_each_cpu(bts_trace_stop_cpu, NULL, 1); + + mutex_unlock(&bts_tracer_mutex); } +static int __cpuinit bts_hotcpu_handler(struct notifier_block *nfb, + unsigned long action, void *hcpu) +{ + unsigned int cpu = (unsigned long)hcpu; + + mutex_lock(&bts_tracer_mutex); + + if (!trace_hw_branches_enabled) + goto out; + + switch (action) { + case CPU_ONLINE: + case CPU_DOWN_FAILED: + smp_call_function_single(cpu, bts_trace_start_cpu, NULL, 1); + break; + case CPU_DOWN_PREPARE: + smp_call_function_single(cpu, bts_trace_stop_cpu, NULL, 1); + break; + } + + out: + mutex_unlock(&bts_tracer_mutex); + return NOTIFY_DONE; +} + +static struct notifier_block bts_hotcpu_notifier __cpuinitdata = { + .notifier_call = bts_hotcpu_handler +}; + static int bts_trace_init(struct trace_array *tr) { + register_hotcpu_notifier(&bts_hotcpu_notifier); tracing_reset_online_cpus(tr); bts_trace_start(tr); return 0; } +static void bts_trace_reset(struct trace_array *tr) +{ + bts_trace_stop(tr); + unregister_hotcpu_notifier(&bts_hotcpu_notifier); +} + static void bts_trace_print_header(struct seq_file *m) { seq_puts(m, @@ -108,18 +174,34 @@ void trace_hw_branch(struct trace_array *tr, u64 from, u64 to) { struct ring_buffer_event *event; struct hw_branch_entry *entry; - unsigned long irq; + unsigned long irq1, irq2; + int cpu; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq); - if (!event) + if (unlikely(!tr)) return; + + if (unlikely(!trace_hw_branches_enabled)) + return; + + local_irq_save(irq1); + cpu = raw_smp_processor_id(); + if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) + goto out; + + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq2); + if (!event) + goto out; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, from); entry->ent.type = TRACE_HW_BRANCHES; - entry->ent.cpu = smp_processor_id(); + entry->ent.cpu = cpu; entry->from = from; entry->to = to; - ring_buffer_unlock_commit(tr->buffer, event, irq); + ring_buffer_unlock_commit(tr->buffer, event, irq2); + + out: + atomic_dec(&tr->data[cpu]->disabled); + local_irq_restore(irq1); } static void trace_bts_at(struct trace_array *tr, @@ -143,6 +225,11 @@ static void trace_bts_at(struct trace_array *tr, } } +/* + * Collect the trace on the current cpu and write it into the ftrace buffer. + * + * pre: bts_tracer_mutex must be locked + */ static void trace_bts_cpu(void *arg) { struct trace_array *tr = (struct trace_array *) arg; @@ -152,6 +239,9 @@ static void trace_bts_cpu(void *arg) if (!this_tracer) return; + if (unlikely(atomic_read(&tr->data[raw_smp_processor_id()]->disabled))) + return; + ds_suspend_bts(this_tracer); trace = ds_read_bts(this_tracer); if (!trace) @@ -171,17 +261,18 @@ static void trace_bts_cpu(void *arg) static void trace_bts_prepare(struct trace_iterator *iter) { - int cpu; + mutex_lock(&bts_tracer_mutex); - for_each_cpu(cpu, cpu_possible_mask) - smp_call_function_single(cpu, trace_bts_cpu, iter->tr, 1); + on_each_cpu(trace_bts_cpu, iter->tr, 1); + + mutex_unlock(&bts_tracer_mutex); } struct tracer bts_tracer __read_mostly = { .name = "hw-branch-tracer", .init = bts_trace_init, - .reset = bts_trace_stop, + .reset = bts_trace_reset, .print_header = bts_trace_print_header, .print_line = bts_trace_print_line, .start = bts_trace_start, From b1818748b0cf9427e48acf9713295e829a0d715f Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Mon, 19 Jan 2009 10:31:01 +0100 Subject: [PATCH 065/413] x86, ftrace, hw-branch-tracer: dump trace on oops Dump the branch trace on an oops (based on ftrace_dump_on_oops). Signed-off-by: Markus Metzger Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/dumpstack.c | 6 ++++++ include/linux/ftrace.h | 13 +++++++++++++ kernel/trace/trace.h | 1 - kernel/trace/trace_hw_branches.c | 29 ++++++++++++++++++++++------- 4 files changed, 41 insertions(+), 8 deletions(-) diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c index 6b1f6f6f8661..077c9ea655fc 100644 --- a/arch/x86/kernel/dumpstack.c +++ b/arch/x86/kernel/dumpstack.c @@ -14,6 +14,7 @@ #include #include #include +#include #include @@ -195,6 +196,11 @@ unsigned __kprobes long oops_begin(void) int cpu; unsigned long flags; + /* notify the hw-branch tracer so it may disable tracing and + add the last trace to the trace buffer - + the earlier this happens, the more useful the trace. */ + trace_hw_branch_oops(); + oops_enter(); /* racy, but better than risking deadlock. */ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 054721487574..9f7880d87c39 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -496,4 +496,17 @@ static inline int test_tsk_trace_graph(struct task_struct *tsk) #endif /* CONFIG_TRACING */ + +#ifdef CONFIG_HW_BRANCH_TRACER + +void trace_hw_branch(u64 from, u64 to); +void trace_hw_branch_oops(void); + +#else /* CONFIG_HW_BRANCH_TRACER */ + +static inline void trace_hw_branch(u64 from, u64 to) {} +static inline void trace_hw_branch_oops(void) {} + +#endif /* CONFIG_HW_BRANCH_TRACER */ + #endif /* _LINUX_FTRACE_H */ diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 54b72781e920..b96037d970df 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -438,7 +438,6 @@ void trace_function(struct trace_array *tr, void trace_graph_return(struct ftrace_graph_ret *trace); int trace_graph_entry(struct ftrace_graph_ent *trace); -void trace_hw_branch(struct trace_array *tr, u64 from, u64 to); void tracing_start_cmdline_record(void); void tracing_stop_cmdline_record(void); diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index 398195397c75..e56df2c7d679 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -40,6 +40,7 @@ static DEFINE_PER_CPU(unsigned char[SIZEOF_BTS], buffer); #define this_buffer per_cpu(buffer, smp_processor_id()) static int __read_mostly trace_hw_branches_enabled; +static struct trace_array *hw_branch_trace __read_mostly; /* @@ -128,6 +129,8 @@ static struct notifier_block bts_hotcpu_notifier __cpuinitdata = { static int bts_trace_init(struct trace_array *tr) { + hw_branch_trace = tr; + register_hotcpu_notifier(&bts_hotcpu_notifier); tracing_reset_online_cpus(tr); bts_trace_start(tr); @@ -170,8 +173,9 @@ static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) return TRACE_TYPE_UNHANDLED; } -void trace_hw_branch(struct trace_array *tr, u64 from, u64 to) +void trace_hw_branch(u64 from, u64 to) { + struct trace_array *tr = hw_branch_trace; struct ring_buffer_event *event; struct hw_branch_entry *entry; unsigned long irq1, irq2; @@ -204,8 +208,7 @@ void trace_hw_branch(struct trace_array *tr, u64 from, u64 to) local_irq_restore(irq1); } -static void trace_bts_at(struct trace_array *tr, - const struct bts_trace *trace, void *at) +static void trace_bts_at(const struct bts_trace *trace, void *at) { struct bts_struct bts; int err = 0; @@ -220,7 +223,7 @@ static void trace_bts_at(struct trace_array *tr, switch (bts.qualifier) { case BTS_BRANCH: - trace_hw_branch(tr, bts.variant.lbr.from, bts.variant.lbr.to); + trace_hw_branch(bts.variant.lbr.from, bts.variant.lbr.to); break; } } @@ -236,12 +239,15 @@ static void trace_bts_cpu(void *arg) const struct bts_trace *trace; unsigned char *at; - if (!this_tracer) + if (unlikely(!tr)) return; if (unlikely(atomic_read(&tr->data[raw_smp_processor_id()]->disabled))) return; + if (unlikely(!this_tracer)) + return; + ds_suspend_bts(this_tracer); trace = ds_read_bts(this_tracer); if (!trace) @@ -249,11 +255,11 @@ static void trace_bts_cpu(void *arg) for (at = trace->ds.top; (void *)at < trace->ds.end; at += trace->ds.size) - trace_bts_at(tr, trace, at); + trace_bts_at(trace, at); for (at = trace->ds.begin; (void *)at < trace->ds.top; at += trace->ds.size) - trace_bts_at(tr, trace, at); + trace_bts_at(trace, at); out: ds_resume_bts(this_tracer); @@ -268,6 +274,15 @@ static void trace_bts_prepare(struct trace_iterator *iter) mutex_unlock(&bts_tracer_mutex); } +void trace_hw_branch_oops(void) +{ + mutex_lock(&bts_tracer_mutex); + + trace_bts_cpu(hw_branch_trace); + + mutex_unlock(&bts_tracer_mutex); +} + struct tracer bts_tracer __read_mostly = { .name = "hw-branch-tracer", From e23b8ad83430a6fdfbdbfac365f5b0312dd57f10 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Mon, 19 Jan 2009 10:33:31 +0100 Subject: [PATCH 066/413] x86, ftrace, hw-branch-tracer: reset trace buffer on close Reset the ftrace buffer on close. Since we use cyclic buffers, the trace is not contiguous, anyway. Signed-off-by: Markus Metzger Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_hw_branches.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index e56df2c7d679..372b47ac3154 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -274,6 +274,11 @@ static void trace_bts_prepare(struct trace_iterator *iter) mutex_unlock(&bts_tracer_mutex); } +static void trace_bts_close(struct trace_iterator *iter) +{ + tracing_reset_online_cpus(iter->tr); +} + void trace_hw_branch_oops(void) { mutex_lock(&bts_tracer_mutex); @@ -292,7 +297,8 @@ struct tracer bts_tracer __read_mostly = .print_line = bts_trace_print_line, .start = bts_trace_start, .stop = bts_trace_stop, - .open = trace_bts_prepare + .open = trace_bts_prepare, + .close = trace_bts_close }; __init static int init_bts_trace(void) From 11edda06289d412d13ff7c672bd72e043f637e74 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Mon, 19 Jan 2009 10:29:16 +0100 Subject: [PATCH 067/413] x86, ftrace, hw-branch-tracer: change trace format Change the hw-branch-tracer format to be more readable. Signed-off-by: Markus Metzger Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_hw_branches.c | 13 +++++-------- 1 file changed, 5 insertions(+), 8 deletions(-) diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index 372b47ac3154..fff3545fc866 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -146,10 +146,7 @@ static void bts_trace_reset(struct trace_array *tr) static void bts_trace_print_header(struct seq_file *m) { - seq_puts(m, - "# CPU# FROM TO FUNCTION\n"); - seq_puts(m, - "# | | | |\n"); + seq_puts(m, "# CPU# TO <- FROM\n"); } static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) @@ -157,15 +154,15 @@ static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) struct trace_entry *entry = iter->ent; struct trace_seq *seq = &iter->seq; struct hw_branch_entry *it; + unsigned long symflags = TRACE_ITER_SYM_OFFSET; trace_assign_type(it, entry); if (entry->type == TRACE_HW_BRANCHES) { if (trace_seq_printf(seq, "%4d ", entry->cpu) && - trace_seq_printf(seq, "0x%016llx -> 0x%016llx ", - it->from, it->to) && - (!it->from || - seq_print_ip_sym(seq, it->from, /* sym_flags = */ 0)) && + seq_print_ip_sym(seq, it->to, symflags) && + trace_seq_printf(seq, "\t <- ") && + seq_print_ip_sym(seq, it->from, symflags) && trace_seq_printf(seq, "\n")) return TRACE_TYPE_HANDLED; return TRACE_TYPE_PARTIAL_LINE;; From e2ea5399bb4fb7aaafb08f846db453f4eec55160 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Mon, 19 Jan 2009 10:35:58 +0100 Subject: [PATCH 068/413] x86, ftrace, hw-branch-tracer: documentation Document the hw-branch-tracer in the ftrace documentation. Signed-off-by: Markus Metzger Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- Documentation/ftrace.txt | 74 ++++++++++++++++++++++++++++++++++++++++ 1 file changed, 74 insertions(+) diff --git a/Documentation/ftrace.txt b/Documentation/ftrace.txt index 803b1318b13d..758fb42a1b68 100644 --- a/Documentation/ftrace.txt +++ b/Documentation/ftrace.txt @@ -165,6 +165,8 @@ Here is the list of current tracers that may be configured. nop - This is not a tracer. To remove all tracers from tracing simply echo "nop" into current_tracer. + hw-branch-tracer - traces branches on all cpu's in a circular buffer. + Examples of using the tracer ---------------------------- @@ -1152,6 +1154,78 @@ int main (int argc, char **argv) return 0; } + +hw-branch-tracer (x86 only) +--------------------------- + +This tracer uses the x86 last branch tracing hardware feature to +collect a branch trace on all cpus with relatively low overhead. + +The tracer uses a fixed-size circular buffer per cpu and only +traces ring 0 branches. The trace file dumps that buffer in the +following format: + +# tracer: hw-branch-tracer +# +# CPU# TO <- FROM + 0 scheduler_tick+0xb5/0x1bf <- task_tick_idle+0x5/0x6 + 2 run_posix_cpu_timers+0x2b/0x72a <- run_posix_cpu_timers+0x25/0x72a + 0 scheduler_tick+0x139/0x1bf <- scheduler_tick+0xed/0x1bf + 0 scheduler_tick+0x17c/0x1bf <- scheduler_tick+0x148/0x1bf + 2 run_posix_cpu_timers+0x9e/0x72a <- run_posix_cpu_timers+0x5e/0x72a + 0 scheduler_tick+0x1b6/0x1bf <- scheduler_tick+0x1aa/0x1bf + + +The tracer may be used to dump the trace for the oops'ing cpu on a +kernel oops into the system log. To enable this, ftrace_dump_on_oops +must be set. To set ftrace_dump_on_oops, one can either use the sysctl +function or set it via the proc system interface. + + sysctl kernel.ftrace_dump_on_oops=1 + +or + + echo 1 > /proc/sys/kernel/ftrace_dump_on_oops + + +Here's an example of such a dump after a null pointer dereference in a +kernel module: + +[57848.105921] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000 +[57848.106019] IP: [] open+0x6/0x14 [oops] +[57848.106019] PGD 2354e9067 PUD 2375e7067 PMD 0 +[57848.106019] Oops: 0002 [#1] SMP +[57848.106019] last sysfs file: /sys/devices/pci0000:00/0000:00:1e.0/0000:20:05.0/local_cpus +[57848.106019] Dumping ftrace buffer: +[57848.106019] --------------------------------- +[...] +[57848.106019] 0 chrdev_open+0xe6/0x165 <- cdev_put+0x23/0x24 +[57848.106019] 0 chrdev_open+0x117/0x165 <- chrdev_open+0xfa/0x165 +[57848.106019] 0 chrdev_open+0x120/0x165 <- chrdev_open+0x11c/0x165 +[57848.106019] 0 chrdev_open+0x134/0x165 <- chrdev_open+0x12b/0x165 +[57848.106019] 0 open+0x0/0x14 [oops] <- chrdev_open+0x144/0x165 +[57848.106019] 0 page_fault+0x0/0x30 <- open+0x6/0x14 [oops] +[57848.106019] 0 error_entry+0x0/0x5b <- page_fault+0x4/0x30 +[57848.106019] 0 error_kernelspace+0x0/0x31 <- error_entry+0x59/0x5b +[57848.106019] 0 error_sti+0x0/0x1 <- error_kernelspace+0x2d/0x31 +[57848.106019] 0 page_fault+0x9/0x30 <- error_sti+0x0/0x1 +[57848.106019] 0 do_page_fault+0x0/0x881 <- page_fault+0x1a/0x30 +[...] +[57848.106019] 0 do_page_fault+0x66b/0x881 <- is_prefetch+0x1ee/0x1f2 +[57848.106019] 0 do_page_fault+0x6e0/0x881 <- do_page_fault+0x67a/0x881 +[57848.106019] 0 oops_begin+0x0/0x96 <- do_page_fault+0x6e0/0x881 +[57848.106019] 0 trace_hw_branch_oops+0x0/0x2d <- oops_begin+0x9/0x96 +[...] +[57848.106019] 0 ds_suspend_bts+0x2a/0xe3 <- ds_suspend_bts+0x1a/0xe3 +[57848.106019] --------------------------------- +[57848.106019] CPU 0 +[57848.106019] Modules linked in: oops +[57848.106019] Pid: 5542, comm: cat Tainted: G W 2.6.28 #23 +[57848.106019] RIP: 0010:[] [] open+0x6/0x14 [oops] +[57848.106019] RSP: 0018:ffff880235457d48 EFLAGS: 00010246 +[...] + + dynamic ftrace -------------- From ce5e5540c0e839781e7cd134517d5d2e9e819636 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Mon, 19 Jan 2009 10:38:35 +0100 Subject: [PATCH 069/413] x86, ds, bts: cleanup DS configuration Cleanup the cpuid check for DS configuration. Signed-off-by: Markus Metzger Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/ds.c | 31 +++++++++++++++++-------------- 1 file changed, 17 insertions(+), 14 deletions(-) diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c index da91701a2348..169a120587be 100644 --- a/arch/x86/kernel/ds.c +++ b/arch/x86/kernel/ds.c @@ -15,8 +15,8 @@ * - buffer allocation (memory accounting) * * - * Copyright (C) 2007-2008 Intel Corporation. - * Markus Metzger , 2007-2008 + * Copyright (C) 2007-2009 Intel Corporation. + * Markus Metzger , 2007-2009 */ @@ -890,7 +890,7 @@ int ds_set_pebs_reset(struct pebs_tracer *tracer, u64 value) } static const struct ds_configuration ds_cfg_netburst = { - .name = "netburst", + .name = "Netburst", .ctl[dsf_bts] = (1 << 2) | (1 << 3), .ctl[dsf_bts_kernel] = (1 << 5), .ctl[dsf_bts_user] = (1 << 6), @@ -904,7 +904,7 @@ static const struct ds_configuration ds_cfg_netburst = { #endif }; static const struct ds_configuration ds_cfg_pentium_m = { - .name = "pentium m", + .name = "Pentium M", .ctl[dsf_bts] = (1 << 6) | (1 << 7), .sizeof_field = sizeof(long), @@ -915,8 +915,8 @@ static const struct ds_configuration ds_cfg_pentium_m = { .sizeof_rec[ds_pebs] = sizeof(long) * 18, #endif }; -static const struct ds_configuration ds_cfg_core2 = { - .name = "core 2", +static const struct ds_configuration ds_cfg_core2_atom = { + .name = "Core 2/Atom", .ctl[dsf_bts] = (1 << 6) | (1 << 7), .ctl[dsf_bts_kernel] = (1 << 9), .ctl[dsf_bts_user] = (1 << 10), @@ -949,19 +949,22 @@ void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) switch (c->x86) { case 0x6: switch (c->x86_model) { - case 0 ... 0xC: - /* sorry, don't know about them */ - break; - case 0xD: - case 0xE: /* Pentium M */ + case 0x9: + case 0xd: /* Pentium M */ ds_configure(&ds_cfg_pentium_m); break; - default: /* Core2, Atom, ... */ - ds_configure(&ds_cfg_core2); + case 0xf: + case 0x17: /* Core2 */ + case 0x1c: /* Atom */ + ds_configure(&ds_cfg_core2_atom); + break; + case 0x1a: /* i7 */ + default: + /* sorry, don't know about them */ break; } break; - case 0xF: + case 0xf: switch (c->x86_model) { case 0x0: case 0x1: From 3690b5e6fd9daa030039ae9bda69044228bd476d Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 16 Jan 2009 16:32:25 +0800 Subject: [PATCH 070/413] trace_workqueue: use percpu data for workqueue stat Impact: use percpu data instead of a global structure Use: static DEFINE_PER_CPU(struct workqueue_global_stats, all_workqueue_stat); instead of allocating a global structure. percpu data also works well on NUMA. Signed-off-by: Lai Jiangshan Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_workqueue.c | 64 +++++++++++++++------------------- 1 file changed, 29 insertions(+), 35 deletions(-) diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index f8118d39ca9b..4664990fe9c5 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -8,6 +8,7 @@ #include #include +#include #include "trace_stat.h" #include "trace.h" @@ -37,7 +38,8 @@ struct workqueue_global_stats { /* Don't need a global lock because allocated before the workqueues, and * never freed. */ -static struct workqueue_global_stats *all_workqueue_stat; +static DEFINE_PER_CPU(struct workqueue_global_stats, all_workqueue_stat); +#define workqueue_cpu_stat(cpu) (&per_cpu(all_workqueue_stat, cpu)) /* Insertion of a work */ static void @@ -48,8 +50,8 @@ probe_workqueue_insertion(struct task_struct *wq_thread, struct cpu_workqueue_stats *node, *next; unsigned long flags; - spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); - list_for_each_entry_safe(node, next, &all_workqueue_stat[cpu].list, + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + list_for_each_entry_safe(node, next, &workqueue_cpu_stat(cpu)->list, list) { if (node->pid == wq_thread->pid) { atomic_inc(&node->inserted); @@ -58,7 +60,7 @@ probe_workqueue_insertion(struct task_struct *wq_thread, } pr_debug("trace_workqueue: entry not found\n"); found: - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); } /* Execution of a work */ @@ -70,8 +72,8 @@ probe_workqueue_execution(struct task_struct *wq_thread, struct cpu_workqueue_stats *node, *next; unsigned long flags; - spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); - list_for_each_entry_safe(node, next, &all_workqueue_stat[cpu].list, + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + list_for_each_entry_safe(node, next, &workqueue_cpu_stat(cpu)->list, list) { if (node->pid == wq_thread->pid) { node->executed++; @@ -80,7 +82,7 @@ probe_workqueue_execution(struct task_struct *wq_thread, } pr_debug("trace_workqueue: entry not found\n"); found: - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); } /* Creation of a cpu workqueue thread */ @@ -104,11 +106,11 @@ static void probe_workqueue_creation(struct task_struct *wq_thread, int cpu) cws->pid = wq_thread->pid; - spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); - if (list_empty(&all_workqueue_stat[cpu].list)) + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + if (list_empty(&workqueue_cpu_stat(cpu)->list)) cws->first_entry = true; - list_add_tail(&cws->list, &all_workqueue_stat[cpu].list); - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + list_add_tail(&cws->list, &workqueue_cpu_stat(cpu)->list); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); } /* Destruction of a cpu workqueue thread */ @@ -119,8 +121,8 @@ static void probe_workqueue_destruction(struct task_struct *wq_thread) struct cpu_workqueue_stats *node, *next; unsigned long flags; - spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); - list_for_each_entry_safe(node, next, &all_workqueue_stat[cpu].list, + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + list_for_each_entry_safe(node, next, &workqueue_cpu_stat(cpu)->list, list) { if (node->pid == wq_thread->pid) { list_del(&node->list); @@ -131,7 +133,7 @@ static void probe_workqueue_destruction(struct task_struct *wq_thread) pr_debug("trace_workqueue: don't find workqueue to destroy\n"); found: - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); } @@ -141,13 +143,13 @@ static struct cpu_workqueue_stats *workqueue_stat_start_cpu(int cpu) struct cpu_workqueue_stats *ret = NULL; - spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); - if (!list_empty(&all_workqueue_stat[cpu].list)) - ret = list_entry(all_workqueue_stat[cpu].list.next, + if (!list_empty(&workqueue_cpu_stat(cpu)->list)) + ret = list_entry(workqueue_cpu_stat(cpu)->list.next, struct cpu_workqueue_stats, list); - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); return ret; } @@ -172,9 +174,9 @@ static void *workqueue_stat_next(void *prev, int idx) unsigned long flags; void *ret = NULL; - spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); - if (list_is_last(&prev_cws->list, &all_workqueue_stat[cpu].list)) { - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + if (list_is_last(&prev_cws->list, &workqueue_cpu_stat(cpu)->list)) { + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); for (++cpu ; cpu < num_possible_cpus(); cpu++) { ret = workqueue_stat_start_cpu(cpu); if (ret) @@ -182,7 +184,7 @@ static void *workqueue_stat_next(void *prev, int idx) } return NULL; } - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); return list_entry(prev_cws->list.next, struct cpu_workqueue_stats, list); @@ -199,10 +201,10 @@ static int workqueue_stat_show(struct seq_file *s, void *p) cws->executed, trace_find_cmdline(cws->pid)); - spin_lock_irqsave(&all_workqueue_stat[cpu].lock, flags); - if (&cws->list == all_workqueue_stat[cpu].list.next) + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + if (&cws->list == workqueue_cpu_stat(cpu)->list.next) seq_printf(s, "\n"); - spin_unlock_irqrestore(&all_workqueue_stat[cpu].lock, flags); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); return 0; } @@ -258,17 +260,9 @@ int __init trace_workqueue_early_init(void) if (ret) goto no_creation; - all_workqueue_stat = kmalloc(sizeof(struct workqueue_global_stats) - * num_possible_cpus(), GFP_KERNEL); - - if (!all_workqueue_stat) { - pr_warning("trace_workqueue: not enough memory\n"); - goto no_creation; - } - for_each_possible_cpu(cpu) { - spin_lock_init(&all_workqueue_stat[cpu].lock); - INIT_LIST_HEAD(&all_workqueue_stat[cpu].list); + spin_lock_init(&workqueue_cpu_stat(cpu)->lock); + INIT_LIST_HEAD(&workqueue_cpu_stat(cpu)->list); } return 0; From 5bc4564b224c3d9fe6dddafa25f56059bd978231 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 21 Jan 2009 14:36:52 -0500 Subject: [PATCH 071/413] trace: do not disable wake up tracer on output of trace Impact: fix to erased trace output To try not to have the outputing of a trace interfere with the wakeup tracer, it would disable tracing while the output was printing. But if a trace had started when it was disabled, it can show a partial trace. To try to solve this, on closing of the tracer, it would clear the trace buffer. The latency tracers (wakeup and irqsoff) have two buffers. One for recording and one for holding the max trace that is printed. The clearing of the trace above should only affect the recording buffer. But for some reason it would move the erased trace to the print buffer. Probably due to a race with the closing of the trace and the saving ofhe max race. The above is all pretty useless, and if the user does not want the printing of the trace to be traced itself, then the user can manual disable tracing. This patch removes all the code that tries to keep the output of the tracer from modifying the trace. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_sched_wakeup.c | 33 ++----------------------------- 1 file changed, 2 insertions(+), 31 deletions(-) diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 42ae1e77b6b3..e27adef0171a 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -262,12 +262,6 @@ probe_wakeup(struct rq *rq, struct task_struct *p, int success) atomic_dec(&wakeup_trace->data[cpu]->disabled); } -/* - * save_tracer_enabled is used to save the state of the tracer_enabled - * variable when we disable it when we open a trace output file. - */ -static int save_tracer_enabled; - static void start_wakeup_tracer(struct trace_array *tr) { int ret; @@ -306,13 +300,10 @@ static void start_wakeup_tracer(struct trace_array *tr) register_ftrace_function(&trace_ops); - if (tracing_is_enabled()) { + if (tracing_is_enabled()) tracer_enabled = 1; - save_tracer_enabled = 1; - } else { + else tracer_enabled = 0; - save_tracer_enabled = 0; - } return; fail_deprobe_wake_new: @@ -324,7 +315,6 @@ static void start_wakeup_tracer(struct trace_array *tr) static void stop_wakeup_tracer(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; unregister_ftrace_function(&trace_ops); unregister_trace_sched_switch(probe_wakeup_sched_switch); unregister_trace_sched_wakeup_new(probe_wakeup); @@ -350,28 +340,11 @@ static void wakeup_tracer_start(struct trace_array *tr) { wakeup_reset(tr); tracer_enabled = 1; - save_tracer_enabled = 1; } static void wakeup_tracer_stop(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; -} - -static void wakeup_tracer_open(struct trace_iterator *iter) -{ - /* stop the trace while dumping */ - tracer_enabled = 0; -} - -static void wakeup_tracer_close(struct trace_iterator *iter) -{ - /* forget about any processes we were recording */ - if (save_tracer_enabled) { - wakeup_reset(iter->tr); - tracer_enabled = 1; - } } static struct tracer wakeup_tracer __read_mostly = @@ -381,8 +354,6 @@ static struct tracer wakeup_tracer __read_mostly = .reset = wakeup_tracer_reset, .start = wakeup_tracer_start, .stop = wakeup_tracer_stop, - .open = wakeup_tracer_open, - .close = wakeup_tracer_close, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_wakeup, From 97b17efe4537e11bf6669106cfe4ee2c5331b267 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 21 Jan 2009 15:24:56 -0500 Subject: [PATCH 072/413] ring-buffer: do not swap if recording is disabled If the ring buffer recording has been disabled. Do not let swapping of ring buffers occur. Simply return -EAGAIN. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 15 +++++++++++++++ kernel/trace/trace.c | 2 +- 2 files changed, 16 insertions(+), 1 deletion(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 0b9de5a3d699..890020e28a35 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2266,9 +2266,24 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, if (buffer_a->pages != buffer_b->pages) return -EINVAL; + if (ring_buffer_flags != RB_BUFFERS_ON) + return -EAGAIN; + + if (atomic_read(&buffer_a->record_disabled)) + return -EAGAIN; + + if (atomic_read(&buffer_b->record_disabled)) + return -EAGAIN; + cpu_buffer_a = buffer_a->buffers[cpu]; cpu_buffer_b = buffer_b->buffers[cpu]; + if (atomic_read(&cpu_buffer_a->record_disabled)) + return -EAGAIN; + + if (atomic_read(&cpu_buffer_b->record_disabled)) + return -EAGAIN; + /* * We can't do a synchronize_sched here because this * function can be called in atomic context. diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 220c264e3111..757ae6f7e648 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -415,7 +415,7 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu) ftrace_enable_cpu(); - WARN_ON_ONCE(ret); + WARN_ON_ONCE(ret && ret != -EAGAIN); __update_max_tr(tr, tsk, cpu); __raw_spin_unlock(&ftrace_max_lock); From 3244351c31211a8b1ba8b4b34c3de04d5dfa03e4 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 21 Jan 2009 16:24:46 -0500 Subject: [PATCH 073/413] trace: separate out rt tasks from wakeup tracer Impact: add option to trace all tasks or just RT tasks The current wakeup tracer only traces RT task wakeups. This is fine for those interested in wake up timings of RT tasks, but it is useless for those that are interested in the causes of long wakeups for non RT tasks. This patch creates a "wakeup_rt" to implement the tracing of just RT tasks (as the current "wakeup" does). And makes "wakeup" now trace all tasks as an average developer would expect. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_sched_wakeup.c | 34 +++++++++++++++++++++++++++++-- 1 file changed, 32 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index e27adef0171a..f48957886102 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -25,6 +25,7 @@ static int __read_mostly tracer_enabled; static struct task_struct *wakeup_task; static int wakeup_cpu; static unsigned wakeup_prio = -1; +static int wakeup_rt; static raw_spinlock_t wakeup_lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; @@ -224,7 +225,7 @@ probe_wakeup(struct rq *rq, struct task_struct *p, int success) tracing_record_cmdline(p); tracing_record_cmdline(current); - if (likely(!rt_task(p)) || + if ((wakeup_rt && !rt_task(p)) || p->prio >= wakeup_prio || p->prio >= current->prio) return; @@ -321,7 +322,7 @@ static void stop_wakeup_tracer(struct trace_array *tr) unregister_trace_sched_wakeup(probe_wakeup); } -static int wakeup_tracer_init(struct trace_array *tr) +static int __wakeup_tracer_init(struct trace_array *tr) { tracing_max_latency = 0; wakeup_trace = tr; @@ -329,6 +330,18 @@ static int wakeup_tracer_init(struct trace_array *tr) return 0; } +static int wakeup_tracer_init(struct trace_array *tr) +{ + wakeup_rt = 0; + return __wakeup_tracer_init(tr); +} + +static int wakeup_rt_tracer_init(struct trace_array *tr) +{ + wakeup_rt = 1; + return __wakeup_tracer_init(tr); +} + static void wakeup_tracer_reset(struct trace_array *tr) { stop_wakeup_tracer(tr); @@ -360,6 +373,19 @@ static struct tracer wakeup_tracer __read_mostly = #endif }; +static struct tracer wakeup_rt_tracer __read_mostly = +{ + .name = "wakeup_rt", + .init = wakeup_rt_tracer_init, + .reset = wakeup_tracer_reset, + .start = wakeup_tracer_start, + .stop = wakeup_tracer_stop, + .print_max = 1, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_wakeup, +#endif +}; + __init static int init_wakeup_tracer(void) { int ret; @@ -368,6 +394,10 @@ __init static int init_wakeup_tracer(void) if (ret) return ret; + ret = register_tracer(&wakeup_rt_tracer); + if (ret) + return ret; + return 0; } device_initcall(init_wakeup_tracer); From f8ec1062f589cdb1cffcffab1376124a1bc08500 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 21 Jan 2009 17:17:04 -0500 Subject: [PATCH 074/413] wakeup-tracer: show scheduling data in output Impact: better data for wakeup tracer This patch adds the wakeup and schedule calls that are used by the scheduler tracer to make the wakeup tracer more readable. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_sched_wakeup.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index f48957886102..93cecda650b2 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -153,6 +153,7 @@ probe_wakeup_sched_switch(struct rq *rq, struct task_struct *prev, goto out_unlock; trace_function(wakeup_trace, data, CALLER_ADDR1, CALLER_ADDR2, flags, pc); + tracing_sched_switch_trace(wakeup_trace, data, prev, next, flags, pc); /* * usecs conversion is slow so we try to delay the conversion @@ -214,6 +215,7 @@ static void wakeup_reset(struct trace_array *tr) static void probe_wakeup(struct rq *rq, struct task_struct *p, int success) { + struct trace_array_cpu *data; int cpu = smp_processor_id(); unsigned long flags; long disabled; @@ -253,9 +255,12 @@ probe_wakeup(struct rq *rq, struct task_struct *p, int success) local_save_flags(flags); - wakeup_trace->data[wakeup_cpu]->preempt_timestamp = ftrace_now(cpu); - trace_function(wakeup_trace, wakeup_trace->data[wakeup_cpu], - CALLER_ADDR1, CALLER_ADDR2, flags, pc); + data = wakeup_trace->data[wakeup_cpu]; + data->preempt_timestamp = ftrace_now(cpu); + tracing_sched_wakeup_trace(wakeup_trace, data, p, current, + flags, pc); + trace_function(wakeup_trace, data, CALLER_ADDR1, CALLER_ADDR2, + flags, pc); out_locked: __raw_spin_unlock(&wakeup_lock); From 69507c06539332e6e49f83aa478844130233bece Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 21 Jan 2009 18:45:57 -0500 Subject: [PATCH 075/413] ring-buffer: reset timestamps when ring buffer is reset Impact: fix bad times of recent resets The ring buffer needs to reset its timestamps when reseting of the buffer, otherwise the timestamps are stale and might be used to calculate times in the buffer causing funny timestamps to appear. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 890020e28a35..7839280ffcd8 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2166,6 +2166,9 @@ rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer) cpu_buffer->overrun = 0; cpu_buffer->entries = 0; + + cpu_buffer->write_stamp = 0; + cpu_buffer->read_stamp = 0; } /** From 94523e818f283d3c69f621406f633afff46dbf82 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 22 Jan 2009 11:18:06 -0500 Subject: [PATCH 076/413] trace: remove internal irqsoff disabling for trace output Impact: cleanup of duplicate features The trace output disables the ring buffer and prevents tracing to occur. The code in irqsoff to do the same thing is no longer needed. This patch removes it. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_irqsoff.c | 34 ++-------------------------------- 1 file changed, 2 insertions(+), 32 deletions(-) diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 62a78d943534..ed344b022a14 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -353,28 +353,18 @@ void trace_preempt_off(unsigned long a0, unsigned long a1) } #endif /* CONFIG_PREEMPT_TRACER */ -/* - * save_tracer_enabled is used to save the state of the tracer_enabled - * variable when we disable it when we open a trace output file. - */ -static int save_tracer_enabled; - static void start_irqsoff_tracer(struct trace_array *tr) { register_ftrace_function(&trace_ops); - if (tracing_is_enabled()) { + if (tracing_is_enabled()) tracer_enabled = 1; - save_tracer_enabled = 1; - } else { + else tracer_enabled = 0; - save_tracer_enabled = 0; - } } static void stop_irqsoff_tracer(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; unregister_ftrace_function(&trace_ops); } @@ -395,25 +385,11 @@ static void irqsoff_tracer_reset(struct trace_array *tr) static void irqsoff_tracer_start(struct trace_array *tr) { tracer_enabled = 1; - save_tracer_enabled = 1; } static void irqsoff_tracer_stop(struct trace_array *tr) { tracer_enabled = 0; - save_tracer_enabled = 0; -} - -static void irqsoff_tracer_open(struct trace_iterator *iter) -{ - /* stop the trace while dumping */ - tracer_enabled = 0; -} - -static void irqsoff_tracer_close(struct trace_iterator *iter) -{ - /* restart tracing */ - tracer_enabled = save_tracer_enabled; } #ifdef CONFIG_IRQSOFF_TRACER @@ -431,8 +407,6 @@ static struct tracer irqsoff_tracer __read_mostly = .reset = irqsoff_tracer_reset, .start = irqsoff_tracer_start, .stop = irqsoff_tracer_stop, - .open = irqsoff_tracer_open, - .close = irqsoff_tracer_close, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_irqsoff, @@ -459,8 +433,6 @@ static struct tracer preemptoff_tracer __read_mostly = .reset = irqsoff_tracer_reset, .start = irqsoff_tracer_start, .stop = irqsoff_tracer_stop, - .open = irqsoff_tracer_open, - .close = irqsoff_tracer_close, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_preemptoff, @@ -489,8 +461,6 @@ static struct tracer preemptirqsoff_tracer __read_mostly = .reset = irqsoff_tracer_reset, .start = irqsoff_tracer_start, .stop = irqsoff_tracer_stop, - .open = irqsoff_tracer_open, - .close = irqsoff_tracer_close, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_preemptirqsoff, From b06a830183b610c0a88c29a92feb7991a867ab46 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 22 Jan 2009 14:26:15 -0500 Subject: [PATCH 077/413] trace: fix logic to start/stop counting The logic in the tracing_start/stop code prevents the WARN_ON from ever detecting if a start/stop pair was mismatched. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 757ae6f7e648..2129ab9d2a48 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -610,13 +610,12 @@ void tracing_start(void) return; spin_lock_irqsave(&tracing_start_lock, flags); - if (--trace_stop_count) - goto out; - - if (trace_stop_count < 0) { - /* Someone screwed up their debugging */ - WARN_ON_ONCE(1); - trace_stop_count = 0; + if (--trace_stop_count) { + if (trace_stop_count < 0) { + /* Someone screwed up their debugging */ + WARN_ON_ONCE(1); + trace_stop_count = 0; + } goto out; } From 7e49fcce1bdadd723ae6a0b3b324c4daced61563 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 22 Jan 2009 19:01:40 -0500 Subject: [PATCH 078/413] trace, lockdep: manual preempt count adding for local_bh_disable Impact: fix to preempt trace triggering lockdep check_flag failure In local_bh_disable, the use of add_preempt_count causes the preempt tracer to start recording the time preemption is off. But because it already modified the preempt_count to show softirqs disabled, and before it called the lockdep code to handle this, it causes a state that lockdep can not handle. The preempt tracer will reset the ring buffer on start of a trace, and the ring buffer reset code does a spin_lock_irqsave. This calls into lockdep and lockdep will fail when it detects the invalid state of having softirqs disabled but the internal current->softirqs_enabled is still set. The fix is to manually add the SOFTIRQ_OFFSET to preempt count and call the preempt tracer code outside the lockdep critical area. Thanks to Peter Zijlstra for suggesting this solution. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- include/linux/sched.h | 2 ++ kernel/sched.c | 8 ++++---- kernel/softirq.c | 13 ++++++++++++- 3 files changed, 18 insertions(+), 5 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index 4cae9b81a1f8..33085b88f87b 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -137,6 +137,8 @@ extern unsigned long nr_uninterruptible(void); extern unsigned long nr_active(void); extern unsigned long nr_iowait(void); +extern unsigned long get_parent_ip(unsigned long addr); + struct seq_file; struct cfs_rq; struct task_group; diff --git a/kernel/sched.c b/kernel/sched.c index 52bbf1c842a8..c154825ae753 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -4399,10 +4399,7 @@ void scheduler_tick(void) #endif } -#if defined(CONFIG_PREEMPT) && (defined(CONFIG_DEBUG_PREEMPT) || \ - defined(CONFIG_PREEMPT_TRACER)) - -static inline unsigned long get_parent_ip(unsigned long addr) +unsigned long get_parent_ip(unsigned long addr) { if (in_lock_functions(addr)) { addr = CALLER_ADDR2; @@ -4412,6 +4409,9 @@ static inline unsigned long get_parent_ip(unsigned long addr) return addr; } +#if defined(CONFIG_PREEMPT) && (defined(CONFIG_DEBUG_PREEMPT) || \ + defined(CONFIG_PREEMPT_TRACER)) + void __kprobes add_preempt_count(int val) { #ifdef CONFIG_DEBUG_PREEMPT diff --git a/kernel/softirq.c b/kernel/softirq.c index bdbe9de9cd8d..6edfc2c11d99 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -21,6 +21,7 @@ #include #include #include +#include #include #include @@ -79,13 +80,23 @@ static void __local_bh_disable(unsigned long ip) WARN_ON_ONCE(in_irq()); raw_local_irq_save(flags); - add_preempt_count(SOFTIRQ_OFFSET); + /* + * The preempt tracer hooks into add_preempt_count and will break + * lockdep because it calls back into lockdep after SOFTIRQ_OFFSET + * is set and before current->softirq_enabled is cleared. + * We must manually increment preempt_count here and manually + * call the trace_preempt_off later. + */ + preempt_count() += SOFTIRQ_OFFSET; /* * Were softirqs turned off above: */ if (softirq_count() == SOFTIRQ_OFFSET) trace_softirqs_off(ip); raw_local_irq_restore(flags); + + if (preempt_count() == SOFTIRQ_OFFSET) + trace_preempt_off(CALLER_ADDR0, get_parent_ip(CALLER_ADDR1)); } #else /* !CONFIG_TRACE_IRQFLAGS */ static inline void __local_bh_disable(unsigned long ip) From 9005f3ebebfcfe9ccd731d16c468907a35ac1f9a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 22 Jan 2009 17:04:53 -0800 Subject: [PATCH 079/413] tracing/function-graph-tracer: various fixes and features This patch brings various bugfixes: - Drop the first irrelevant task switch on the very beginning of a trace. - Drop the OVERHEAD word from the headers, the DURATION word is sufficient and will not overlap other columns. - Make the headers fit well their respective columns whatever the selected options. Ie, default options: # tracer: function_graph # # CPU DURATION FUNCTION CALLS # | | | | | | | 1) 0.646 us | } 1) | mem_cgroup_del_lru_list() { 1) 0.624 us | lookup_page_cgroup(); 1) 1.970 us | } echo funcgraph-proc > trace_options # tracer: function_graph # # CPU TASK/PID DURATION FUNCTION CALLS # | | | | | | | | | 0) bash-2937 | 0.895 us | } 0) bash-2937 | 0.888 us | __rcu_read_unlock(); 0) bash-2937 | 0.864 us | conv_uni_to_pc(); 0) bash-2937 | 1.015 us | __rcu_read_lock(); echo nofuncgraph-cpu > trace_options echo nofuncgraph-proc > trace_options # tracer: function_graph # # DURATION FUNCTION CALLS # | | | | | | 3.752 us | native_pud_val(); 0.616 us | native_pud_val(); 0.624 us | native_pmd_val(); About features, one can now disable the duration (this will hide the overhead too for convenient reasons and because on doesn't need overhead if it hasn't the duration): echo nofuncgraph-duration > trace_options # tracer: function_graph # # FUNCTION CALLS # | | | | cap_vm_enough_memory() { __vm_enough_memory() { vm_acct_memory(); } } } And at last, an option to print the absolute time: //Restart from default options echo funcgraph-abstime > trace_options # tracer: function_graph # # TIME CPU DURATION FUNCTION CALLS # | | | | | | | | 261.339774 | 1) + 42.823 us | } 261.339775 | 1) 1.045 us | _spin_lock_irq(); 261.339777 | 1) 0.940 us | _spin_lock_irqsave(); 261.339778 | 1) 0.752 us | _spin_unlock_irqrestore(); 261.339780 | 1) 0.857 us | _spin_unlock_irq(); 261.339782 | 1) | flush_to_ldisc() { 261.339783 | 1) | tty_ldisc_ref() { 261.339783 | 1) | tty_ldisc_try() { 261.339784 | 1) 1.075 us | _spin_lock_irqsave(); 261.339786 | 1) 0.842 us | _spin_unlock_irqrestore(); 261.339788 | 1) 4.211 us | } 261.339788 | 1) 5.662 us | } The format is seconds.usecs. I guess no one needs the nanosec precision here, the main goal is to have an overview about the general timings of events, and to see the place when the trace switches from one cpu to another. ie: 274.874760 | 1) 0.676 us | _spin_unlock(); 274.874762 | 1) 0.609 us | native_load_sp0(); 274.874763 | 1) 0.602 us | native_load_tls(); 274.878739 | 0) 0.722 us | } 274.878740 | 0) 0.714 us | native_pmd_val(); 274.878741 | 0) 0.730 us | native_pmd_val(); Here there is a 4000 usecs difference when we switch the cpu. Changes in V2: - Completely fix the first pointless task switch. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions_graph.c | 302 +++++++++++++++++---------- 1 file changed, 197 insertions(+), 105 deletions(-) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 3c545984816f..66fc7b806a38 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -1,7 +1,7 @@ /* * * Function graph tracer. - * Copyright (c) 2008 Frederic Weisbecker + * Copyright (c) 2008-2009 Frederic Weisbecker * Mostly borrowed from function tracer which * is Copyright (c) Steven Rostedt * @@ -21,9 +21,11 @@ #define TRACE_GRAPH_PRINT_CPU 0x2 #define TRACE_GRAPH_PRINT_OVERHEAD 0x4 #define TRACE_GRAPH_PRINT_PROC 0x8 +#define TRACE_GRAPH_PRINT_DURATION 0x10 +#define TRACE_GRAPH_PRINT_ABS_TIME 0X20 static struct tracer_opt trace_opts[] = { - /* Display overruns ? */ + /* Display overruns? (for self-debug purpose) */ { TRACER_OPT(funcgraph-overrun, TRACE_GRAPH_PRINT_OVERRUN) }, /* Display CPU ? */ { TRACER_OPT(funcgraph-cpu, TRACE_GRAPH_PRINT_CPU) }, @@ -31,17 +33,22 @@ static struct tracer_opt trace_opts[] = { { TRACER_OPT(funcgraph-overhead, TRACE_GRAPH_PRINT_OVERHEAD) }, /* Display proc name/pid */ { TRACER_OPT(funcgraph-proc, TRACE_GRAPH_PRINT_PROC) }, + /* Display duration of execution */ + { TRACER_OPT(funcgraph-duration, TRACE_GRAPH_PRINT_DURATION) }, + /* Display absolute time of an entry */ + { TRACER_OPT(funcgraph-abstime, TRACE_GRAPH_PRINT_ABS_TIME) }, { } /* Empty entry */ }; static struct tracer_flags tracer_flags = { /* Don't display overruns and proc by default */ - .val = TRACE_GRAPH_PRINT_CPU | TRACE_GRAPH_PRINT_OVERHEAD, + .val = TRACE_GRAPH_PRINT_CPU | TRACE_GRAPH_PRINT_OVERHEAD | + TRACE_GRAPH_PRINT_DURATION, .opts = trace_opts }; /* pid on the last trace processed */ -static pid_t last_pid[NR_CPUS] = { [0 ... NR_CPUS-1] = -1 }; + static int graph_trace_init(struct trace_array *tr) { @@ -154,17 +161,25 @@ print_graph_proc(struct trace_seq *s, pid_t pid) /* If the pid changed since the last trace, output this event */ static enum print_line_t -verif_pid(struct trace_seq *s, pid_t pid, int cpu) +verif_pid(struct trace_seq *s, pid_t pid, int cpu, pid_t *last_pids_cpu) { pid_t prev_pid; + pid_t *last_pid; int ret; - if (last_pid[cpu] != -1 && last_pid[cpu] == pid) + if (!last_pids_cpu) return TRACE_TYPE_HANDLED; - prev_pid = last_pid[cpu]; - last_pid[cpu] = pid; + last_pid = per_cpu_ptr(last_pids_cpu, cpu); + if (*last_pid == pid) + return TRACE_TYPE_HANDLED; + + prev_pid = *last_pid; + *last_pid = pid; + + if (prev_pid == -1) + return TRACE_TYPE_HANDLED; /* * Context-switch trace line: @@ -232,9 +247,34 @@ trace_branch_is_leaf(struct trace_iterator *iter, return true; } +/* Signal a overhead of time execution to the output */ +static int +print_graph_overhead(unsigned long long duration, struct trace_seq *s) +{ + /* If duration disappear, we don't need anything */ + if (!(tracer_flags.val & TRACE_GRAPH_PRINT_DURATION)) + return 1; + + /* Non nested entry or return */ + if (duration == -1) + return trace_seq_printf(s, " "); + + if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { + /* Duration exceeded 100 msecs */ + if (duration > 100000ULL) + return trace_seq_printf(s, "! "); + + /* Duration exceeded 10 msecs */ + if (duration > 10000ULL) + return trace_seq_printf(s, "+ "); + } + + return trace_seq_printf(s, " "); +} + static enum print_line_t print_graph_irq(struct trace_seq *s, unsigned long addr, - enum trace_type type, int cpu, pid_t pid) + enum trace_type type, int cpu, pid_t pid) { int ret; @@ -242,35 +282,40 @@ print_graph_irq(struct trace_seq *s, unsigned long addr, addr >= (unsigned long)__irqentry_text_end) return TRACE_TYPE_UNHANDLED; - if (type == TRACE_GRAPH_ENT) { - ret = trace_seq_printf(s, "==========> | "); - } else { - /* Cpu */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { - ret = print_graph_cpu(s, cpu); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - } - /* Proc */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, pid); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - - /* No overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = trace_seq_printf(s, " "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - - ret = trace_seq_printf(s, "<========== |\n"); + /* Cpu */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { + ret = print_graph_cpu(s, cpu); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; } + /* Proc */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { + ret = print_graph_proc(s, pid); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + ret = trace_seq_printf(s, " | "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + + /* No overhead */ + ret = print_graph_overhead(-1, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + if (type == TRACE_GRAPH_ENT) + ret = trace_seq_printf(s, "==========>"); + else + ret = trace_seq_printf(s, "<=========="); + + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* Don't close the duration column if haven't one */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) + trace_seq_printf(s, " |"); + ret = trace_seq_printf(s, "\n"); + if (!ret) return TRACE_TYPE_PARTIAL_LINE; return TRACE_TYPE_HANDLED; @@ -289,7 +334,7 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s) sprintf(msecs_str, "%lu", (unsigned long) duration); /* Print msecs */ - ret = trace_seq_printf(s, msecs_str); + ret = trace_seq_printf(s, "%s", msecs_str); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -322,19 +367,15 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s) } -/* Signal a overhead of time execution to the output */ -static int -print_graph_overhead(unsigned long long duration, struct trace_seq *s) +static int print_graph_abs_time(u64 t, struct trace_seq *s) { - /* Duration exceeded 100 msecs */ - if (duration > 100000ULL) - return trace_seq_printf(s, "! "); + unsigned long usecs_rem; - /* Duration exceeded 10 msecs */ - if (duration > 10000ULL) - return trace_seq_printf(s, "+ "); + usecs_rem = do_div(t, 1000000000); + usecs_rem /= 1000; - return trace_seq_printf(s, " "); + return trace_seq_printf(s, "%5lu.%06lu | ", + (unsigned long)t, usecs_rem); } /* Case of a leaf function on its call entry */ @@ -357,16 +398,16 @@ print_graph_entry_leaf(struct trace_iterator *iter, duration = graph_ret->rettime - graph_ret->calltime; /* Overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = print_graph_overhead(duration, s); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + ret = print_graph_overhead(duration, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; /* Duration */ - ret = print_graph_duration(duration, s); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { + ret = print_graph_duration(duration, s); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + } /* Function */ for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { @@ -395,25 +436,17 @@ print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, struct ftrace_graph_ent *call = &entry->graph_ent; /* No overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = trace_seq_printf(s, " "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + ret = print_graph_overhead(-1, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; - /* Interrupt */ - ret = print_graph_irq(s, call->func, TRACE_GRAPH_ENT, cpu, pid); - if (ret == TRACE_TYPE_UNHANDLED) { - /* No time */ + /* No time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { ret = trace_seq_printf(s, " | "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; - } else { - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; } - /* Function */ for (i = 0; i < call->depth * TRACE_GRAPH_INDENT; i++) { ret = trace_seq_printf(s, " "); @@ -434,15 +467,30 @@ print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, static enum print_line_t print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, - struct trace_iterator *iter, int cpu) + struct trace_iterator *iter) { int ret; + int cpu = iter->cpu; + pid_t *last_entry = iter->private; struct trace_entry *ent = iter->ent; + struct ftrace_graph_ent *call = &field->graph_ent; /* Pid */ - if (verif_pid(s, ent->pid, cpu) == TRACE_TYPE_PARTIAL_LINE) + if (verif_pid(s, ent->pid, cpu, last_entry) == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; + /* Interrupt */ + ret = print_graph_irq(s, call->func, TRACE_GRAPH_ENT, cpu, ent->pid); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + + /* Absolute time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { + ret = print_graph_abs_time(iter->ts, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + /* Cpu */ if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { ret = print_graph_cpu(s, cpu); @@ -470,16 +518,25 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, static enum print_line_t print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, - struct trace_entry *ent, int cpu) + struct trace_entry *ent, struct trace_iterator *iter) { int i; int ret; + int cpu = iter->cpu; + pid_t *last_pid = iter->private; unsigned long long duration = trace->rettime - trace->calltime; /* Pid */ - if (verif_pid(s, ent->pid, cpu) == TRACE_TYPE_PARTIAL_LINE) + if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; + /* Absolute time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { + ret = print_graph_abs_time(iter->ts, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + /* Cpu */ if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { ret = print_graph_cpu(s, cpu); @@ -499,16 +556,16 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, } /* Overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = print_graph_overhead(duration, s); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } + ret = print_graph_overhead(duration, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; /* Duration */ - ret = print_graph_duration(duration, s); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { + ret = print_graph_duration(duration, s); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + } /* Closing brace */ for (i = 0; i < trace->depth * TRACE_GRAPH_INDENT; i++) { @@ -542,14 +599,23 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, { int i; int ret; + int cpu = iter->cpu; + pid_t *last_pid = iter->private; + + /* Absolute time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { + ret = print_graph_abs_time(iter->ts, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } /* Pid */ - if (verif_pid(s, ent->pid, iter->cpu) == TRACE_TYPE_PARTIAL_LINE) + if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; /* Cpu */ if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { - ret = print_graph_cpu(s, iter->cpu); + ret = print_graph_cpu(s, cpu); if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; } @@ -566,17 +632,17 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, } /* No overhead */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - ret = trace_seq_printf(s, " "); + ret = print_graph_overhead(-1, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + /* No time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) { + ret = trace_seq_printf(s, " | "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; } - /* No time */ - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - /* Indentation */ if (trace->depth > 0) for (i = 0; i < (trace->depth + 1) * TRACE_GRAPH_INDENT; i++) { @@ -614,13 +680,12 @@ print_graph_function(struct trace_iterator *iter) case TRACE_GRAPH_ENT: { struct ftrace_graph_ent_entry *field; trace_assign_type(field, entry); - return print_graph_entry(field, s, iter, - iter->cpu); + return print_graph_entry(field, s, iter); } case TRACE_GRAPH_RET: { struct ftrace_graph_ret_entry *field; trace_assign_type(field, entry); - return print_graph_return(&field->ret, s, entry, iter->cpu); + return print_graph_return(&field->ret, s, entry, iter); } case TRACE_PRINT: { struct print_entry *field; @@ -636,28 +701,55 @@ static void print_graph_headers(struct seq_file *s) { /* 1st line */ seq_printf(s, "# "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) + seq_printf(s, " TIME "); if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) - seq_printf(s, "CPU "); + seq_printf(s, "CPU"); if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) - seq_printf(s, "TASK/PID "); - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) - seq_printf(s, "OVERHEAD/"); - seq_printf(s, "DURATION FUNCTION CALLS\n"); + seq_printf(s, " TASK/PID "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) + seq_printf(s, " DURATION "); + seq_printf(s, " FUNCTION CALLS\n"); /* 2nd line */ seq_printf(s, "# "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) + seq_printf(s, " | "); if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) - seq_printf(s, "| "); + seq_printf(s, "| "); if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) - seq_printf(s, "| | "); - if (tracer_flags.val & TRACE_GRAPH_PRINT_OVERHEAD) { - seq_printf(s, "| "); - seq_printf(s, "| | | | |\n"); - } else - seq_printf(s, " | | | | |\n"); + seq_printf(s, " | | "); + if (tracer_flags.val & TRACE_GRAPH_PRINT_DURATION) + seq_printf(s, " | | "); + seq_printf(s, " | | | |\n"); } + +static void graph_trace_open(struct trace_iterator *iter) +{ + /* pid on the last trace processed */ + pid_t *last_pid = alloc_percpu(pid_t); + int cpu; + + if (!last_pid) + pr_warning("function graph tracer: not enough memory\n"); + else + for_each_possible_cpu(cpu) { + pid_t *pid = per_cpu_ptr(last_pid, cpu); + *pid = -1; + } + + iter->private = last_pid; +} + +static void graph_trace_close(struct trace_iterator *iter) +{ + percpu_free(iter->private); +} + static struct tracer graph_trace __read_mostly = { .name = "function_graph", + .open = graph_trace_open, + .close = graph_trace_close, .init = graph_trace_init, .reset = graph_trace_reset, .print_line = print_graph_function, From cc2f6d90e950b69ad31d483c19cc1d121bb25c16 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Fri, 23 Jan 2009 13:03:37 -0800 Subject: [PATCH 080/413] kmemtrace: fix printk format warnings Fix kmemtrace printk warnings: kernel/trace/kmemtrace.c:142: warning: format '%4ld' expects type 'long int', but argument 3 has type 'size_t' kernel/trace/kmemtrace.c:147: warning: format '%4ld' expects type 'long int', but argument 3 has type 'size_t' Signed-off-by: Randy Dunlap Acked-by: Eduard - Gabriel Munteanu Signed-off-by: Ingo Molnar --- kernel/trace/kmemtrace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c index 7ebc58cee3bd..72b326bba005 100644 --- a/kernel/trace/kmemtrace.c +++ b/kernel/trace/kmemtrace.c @@ -139,12 +139,12 @@ kmemtrace_print_alloc_compress(struct trace_iterator *iter, return TRACE_TYPE_PARTIAL_LINE; /* Requested */ - ret = trace_seq_printf(s, "%4ld ", entry->bytes_req); + ret = trace_seq_printf(s, "%4zd ", entry->bytes_req); if (!ret) return TRACE_TYPE_PARTIAL_LINE; /* Allocated */ - ret = trace_seq_printf(s, "%4ld ", entry->bytes_alloc); + ret = trace_seq_printf(s, "%4zd ", entry->bytes_alloc); if (!ret) return TRACE_TYPE_PARTIAL_LINE; From 9011262a37cb438f0fa9394b5e83840db8f9680a Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 23 Jan 2009 12:06:23 -0200 Subject: [PATCH 081/413] ftrace: add ftrace_vprintk Impact: new helper function Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 8 ++++++++ kernel/trace/trace.c | 9 +++++++++ 2 files changed, 17 insertions(+) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 9f7880d87c39..7840e718c6c7 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -302,6 +302,9 @@ ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3); extern int __ftrace_printk(unsigned long ip, const char *fmt, ...) __attribute__ ((format (printf, 2, 3))); +# define ftrace_vprintk(fmt, ap) __ftrace_printk(_THIS_IP_, fmt, ap) +extern int +__ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); extern void ftrace_dump(void); #else static inline void @@ -317,6 +320,11 @@ ftrace_printk(const char *fmt, ...) { return 0; } +static inline int +ftrace_vprintk(const char *fmt, va_list ap) +{ + return 0; +} static inline void ftrace_dump(void) { } #endif diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 2129ab9d2a48..2f8ac1f008f5 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2951,6 +2951,15 @@ int __ftrace_printk(unsigned long ip, const char *fmt, ...) } EXPORT_SYMBOL_GPL(__ftrace_printk); +int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) +{ + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + return trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); +} +EXPORT_SYMBOL_GPL(__ftrace_vprintk); + static int trace_panic_handler(struct notifier_block *this, unsigned long event, void *unused) { From c71a896154119f4ca9e89d6078f5f63ad60ef199 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 23 Jan 2009 12:06:27 -0200 Subject: [PATCH 082/413] blktrace: add ftrace plugin Impact: New way of using the blktrace infrastructure This drops the requirement of userspace utilities to use the blktrace facility. Configuration is done thru sysfs, adding a "trace" directory to the partition directory where blktrace can be enabled for the associated request_queue. The same filters present in the IOCTL interface are present as sysfs device attributes. The /sys/block/sdX/sdXN/trace/enable file allows tracing without any filters. The other files in this directory: pid, act_mask, start_lba and end_lba can be used with the same meaning as with the IOCTL interface. Using the sysfs interface will only setup the request_queue->blk_trace fields, tracing will only take place when the "blk" tracer is selected via the ftrace interface, as in the following example: To see the trace, one can use the /d/tracing/trace file or the /d/tracign/trace_pipe file, with semantics defined in the ftrace documentation in Documentation/ftrace.txt. [root@f10-1 ~]# cat /t/trace kjournald-305 [000] 3046.491224: 8,1 A WBS 6367 + 8 <- (8,1) 6304 kjournald-305 [000] 3046.491227: 8,1 Q R 6367 + 8 [kjournald] kjournald-305 [000] 3046.491236: 8,1 G RB 6367 + 8 [kjournald] kjournald-305 [000] 3046.491239: 8,1 P NS [kjournald] kjournald-305 [000] 3046.491242: 8,1 I RBS 6367 + 8 [kjournald] kjournald-305 [000] 3046.491251: 8,1 D WB 6367 + 8 [kjournald] kjournald-305 [000] 3046.491610: 8,1 U WS [kjournald] 1 -0 [000] 3046.511914: 8,1 C RS 6367 + 8 [6367] [root@f10-1 ~]# The default line context (prefix) format is the one described in the ftrace documentation, with the blktrace specific bits using its existing format, described in blkparse(8). If one wants to have the classic blktrace formatting, this is possible by using: [root@f10-1 ~]# echo blk_classic > /t/trace_options [root@f10-1 ~]# cat /t/trace 8,1 0 3046.491224 305 A WBS 6367 + 8 <- (8,1) 6304 8,1 0 3046.491227 305 Q R 6367 + 8 [kjournald] 8,1 0 3046.491236 305 G RB 6367 + 8 [kjournald] 8,1 0 3046.491239 305 P NS [kjournald] 8,1 0 3046.491242 305 I RBS 6367 + 8 [kjournald] 8,1 0 3046.491251 305 D WB 6367 + 8 [kjournald] 8,1 0 3046.491610 305 U WS [kjournald] 1 8,1 0 3046.511914 0 C RS 6367 + 8 [6367] [root@f10-1 ~]# Using the ftrace standard format allows more flexibility, such as the ability of asking for backtraces via trace_options: [root@f10-1 ~]# echo noblk_classic > /t/trace_options [root@f10-1 ~]# echo stacktrace > /t/trace_options [root@f10-1 ~]# cat /t/trace kjournald-305 [000] 3318.826779: 8,1 A WBS 6375 + 8 <- (8,1) 6312 kjournald-305 [000] 3318.826782: <= submit_bio <= submit_bh <= sync_dirty_buffer <= journal_commit_transaction <= kjournald <= kthread <= child_rip kjournald-305 [000] 3318.826836: 8,1 Q R 6375 + 8 [kjournald] kjournald-305 [000] 3318.826837: <= generic_make_request <= submit_bio <= submit_bh <= sync_dirty_buffer <= journal_commit_transaction <= kjournald <= kthread Please read the ftrace documentation to use aditional, standardized tracing filters such as /d/tracing/trace_cpumask, etc. See also /d/tracing/trace_mark to add comments in the trace stream, that is equivalent to the /d/block/sdaN/msg interface. Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- block/blktrace.c | 651 +++++++++++++++++++++++++++++++++++++++++- fs/partitions/check.c | 7 + kernel/trace/trace.h | 1 + 3 files changed, 654 insertions(+), 5 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index b0a2cae886db..630f167f8240 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -25,9 +25,27 @@ #include #include #include +#include <../kernel/trace/trace_output.h> static unsigned int blktrace_seq __read_mostly = 1; +static struct trace_array *blk_tr; +static int __read_mostly blk_tracer_enabled; + +/* Select an alternative, minimalistic output than the original one */ +#define TRACE_BLK_OPT_CLASSIC 0x1 + +static struct tracer_opt blk_tracer_opts[] = { + /* Default disable the minimalistic output */ + { TRACER_OPT(blk_classic, TRACE_BLK_OPT_CLASSIC ) }, + { } +}; + +static struct tracer_flags blk_tracer_flags = { + .val = 0, + .opts = blk_tracer_opts, +}; + /* Global reference count of probes */ static DEFINE_MUTEX(blk_probe_mutex); static atomic_t blk_probes_ref = ATOMIC_INIT(0); @@ -43,6 +61,9 @@ static void trace_note(struct blk_trace *bt, pid_t pid, int action, { struct blk_io_trace *t; + if (!bt->rchan) + return; + t = relay_reserve(bt->rchan, sizeof(*t) + len); if (t) { const int cpu = smp_processor_id(); @@ -90,6 +111,16 @@ void __trace_note_message(struct blk_trace *bt, const char *fmt, ...) unsigned long flags; char *buf; + if (blk_tr) { + va_start(args, fmt); + ftrace_vprintk(fmt, args); + va_end(args); + return; + } + + if (!bt->msg_data) + return; + local_irq_save(flags); buf = per_cpu_ptr(bt->msg_data, smp_processor_id()); va_start(args, fmt); @@ -131,13 +162,14 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, int rw, u32 what, int error, int pdu_len, void *pdu_data) { struct task_struct *tsk = current; + struct ring_buffer_event *event = NULL; struct blk_io_trace *t; unsigned long flags; unsigned long *sequence; pid_t pid; - int cpu; + int cpu, pc = 0; - if (unlikely(bt->trace_state != Blktrace_running)) + if (unlikely(bt->trace_state != Blktrace_running || !blk_tracer_enabled)) return; what |= ddir_act[rw & WRITE]; @@ -150,6 +182,24 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, pid = tsk->pid; if (unlikely(act_log_check(bt, what, sector, pid))) return; + cpu = raw_smp_processor_id(); + + if (blk_tr) { + struct trace_entry *ent; + tracing_record_cmdline(current); + + event = ring_buffer_lock_reserve(blk_tr->buffer, + sizeof(*t) + pdu_len, &flags); + if (!event) + return; + + ent = ring_buffer_event_data(event); + t = (struct blk_io_trace *)ent; + pc = preempt_count(); + tracing_generic_entry_update(ent, 0, pc); + ent->type = TRACE_BLK; + goto record_it; + } /* * A word about the locking here - we disable interrupts to reserve @@ -163,23 +213,33 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, t = relay_reserve(bt->rchan, sizeof(*t) + pdu_len); if (t) { - cpu = smp_processor_id(); sequence = per_cpu_ptr(bt->sequence, cpu); t->magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION; t->sequence = ++(*sequence); t->time = ktime_to_ns(ktime_get()); + t->cpu = cpu; + t->pid = pid; +record_it: t->sector = sector; t->bytes = bytes; t->action = what; - t->pid = pid; t->device = bt->dev; - t->cpu = cpu; t->error = error; t->pdu_len = pdu_len; if (pdu_len) memcpy((void *) t + sizeof(*t), pdu_data, pdu_len); + + if (blk_tr) { + ring_buffer_unlock_commit(blk_tr->buffer, event, flags); + if (pid != 0 && + (blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC) == 0 && + (trace_flags & TRACE_ITER_STACKTRACE) != 0) + __trace_stack(blk_tr, NULL, flags, 5, pc); + trace_wake_up(); + return; + } } local_irq_restore(flags); @@ -888,3 +948,584 @@ static void blk_unregister_tracepoints(void) tracepoint_synchronize_unregister(); } + +/* + * struct blk_io_tracer formatting routines + */ + +static void fill_rwbs(char *rwbs, const struct blk_io_trace *t) +{ + int i = 0; + + if (t->action & BLK_TC_DISCARD) rwbs[i++] = 'D'; + else if (t->action & BLK_TC_WRITE) rwbs[i++] = 'W'; + else if (t->bytes) rwbs[i++] = 'R'; + else rwbs[i++] = 'N'; + + if (t->action & BLK_TC_AHEAD) rwbs[i++] = 'A'; + if (t->action & BLK_TC_BARRIER) rwbs[i++] = 'B'; + if (t->action & BLK_TC_SYNC) rwbs[i++] = 'S'; + if (t->action & BLK_TC_META) rwbs[i++] = 'M'; + + rwbs[i] = '\0'; +} + +static inline +const struct blk_io_trace *te_blk_io_trace(const struct trace_entry *ent) +{ + return (const struct blk_io_trace *)ent; +} + +static inline const void *pdu_start(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent) + 1; +} + +static inline u32 t_sec(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->bytes >> 9; +} + +static inline unsigned long long t_sector(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->sector; +} + +static inline __u16 t_error(const struct trace_entry *ent) +{ + return te_blk_io_trace(ent)->sector; +} + +static __u64 get_pdu_int(const struct trace_entry *ent) +{ + const __u64 *val = pdu_start(ent); + return be64_to_cpu(*val); +} + +static void get_pdu_remap(const struct trace_entry *ent, + struct blk_io_trace_remap *r) +{ + const struct blk_io_trace_remap *__r = pdu_start(ent); + __u64 sector = __r->sector; + + r->device = be32_to_cpu(__r->device); + r->device_from = be32_to_cpu(__r->device_from); + r->sector = be64_to_cpu(sector); +} + +static int blk_log_action_iter(struct trace_iterator *iter, const char *act) +{ + char rwbs[6]; + unsigned long long ts = ns2usecs(iter->ts); + unsigned long usec_rem = do_div(ts, USEC_PER_SEC); + unsigned secs = (unsigned long)ts; + const struct trace_entry *ent = iter->ent; + const struct blk_io_trace *t = (const struct blk_io_trace *)ent; + + fill_rwbs(rwbs, t); + + return trace_seq_printf(&iter->seq, + "%3d,%-3d %2d %5d.%06lu %5u %2s %3s ", + MAJOR(t->device), MINOR(t->device), iter->cpu, + secs, usec_rem, ent->pid, act, rwbs); +} + +static int blk_log_action_seq(struct trace_seq *s, const struct blk_io_trace *t, + const char *act) +{ + char rwbs[6]; + fill_rwbs(rwbs, t); + return trace_seq_printf(s, "%3d,%-3d %2s %3s ", + MAJOR(t->device), MINOR(t->device), act, rwbs); +} + +static int blk_log_generic(struct trace_seq *s, const struct trace_entry *ent) +{ + const char *cmd = trace_find_cmdline(ent->pid); + + if (t_sec(ent)) + return trace_seq_printf(s, "%llu + %u [%s]\n", + t_sector(ent), t_sec(ent), cmd); + return trace_seq_printf(s, "[%s]\n", cmd); +} + +static int blk_log_with_error(struct trace_seq *s, const struct trace_entry *ent) +{ + if (t_sec(ent)) + return trace_seq_printf(s, "%llu + %u [%d]\n", t_sector(ent), + t_sec(ent), t_error(ent)); + return trace_seq_printf(s, "%llu [%d]\n", t_sector(ent), t_error(ent)); +} + +static int blk_log_remap(struct trace_seq *s, const struct trace_entry *ent) +{ + struct blk_io_trace_remap r = { .device = 0, }; + + get_pdu_remap(ent, &r); + return trace_seq_printf(s, "%llu + %u <- (%d,%d) %llu\n", + t_sector(ent), + t_sec(ent), MAJOR(r.device), MINOR(r.device), + (unsigned long long)r.sector); +} + +static int blk_log_plug(struct trace_seq *s, const struct trace_entry *ent) +{ + return trace_seq_printf(s, "[%s]\n", trace_find_cmdline(ent->pid)); +} + +static int blk_log_unplug(struct trace_seq *s, const struct trace_entry *ent) +{ + return trace_seq_printf(s, "[%s] %llu\n", trace_find_cmdline(ent->pid), + get_pdu_int(ent)); +} + +static int blk_log_split(struct trace_seq *s, const struct trace_entry *ent) +{ + return trace_seq_printf(s, "%llu / %llu [%s]\n", t_sector(ent), + get_pdu_int(ent), trace_find_cmdline(ent->pid)); +} + +/* + * struct tracer operations + */ + +static void blk_tracer_print_header(struct seq_file *m) +{ + if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC)) + return; + seq_puts(m, "# DEV CPU TIMESTAMP PID ACT FLG\n" + "# | | | | | |\n"); +} + +static void blk_tracer_start(struct trace_array *tr) +{ + int cpu; + + tr->time_start = ftrace_now(tr->cpu); + + for_each_online_cpu(cpu) + tracing_reset(tr, cpu); + + mutex_lock(&blk_probe_mutex); + if (atomic_add_return(1, &blk_probes_ref) == 1) + if (blk_register_tracepoints()) + atomic_dec(&blk_probes_ref); + mutex_unlock(&blk_probe_mutex); +} + +static int blk_tracer_init(struct trace_array *tr) +{ + blk_tr = tr; + blk_tracer_start(tr); + mutex_lock(&blk_probe_mutex); + blk_tracer_enabled++; + mutex_unlock(&blk_probe_mutex); + return 0; +} + +static void blk_tracer_stop(struct trace_array *tr) +{ + mutex_lock(&blk_probe_mutex); + if (atomic_dec_and_test(&blk_probes_ref)) + blk_unregister_tracepoints(); + mutex_unlock(&blk_probe_mutex); +} + +static void blk_tracer_reset(struct trace_array *tr) +{ + if (!atomic_read(&blk_probes_ref)) + return; + + mutex_lock(&blk_probe_mutex); + blk_tracer_enabled--; + WARN_ON(blk_tracer_enabled < 0); + mutex_unlock(&blk_probe_mutex); + + blk_tracer_stop(tr); +} + +static struct { + const char *act[2]; + int (*print)(struct trace_seq *s, const struct trace_entry *ent); +} what2act[] __read_mostly = { + [__BLK_TA_QUEUE] = {{ "Q", "queue" }, blk_log_generic }, + [__BLK_TA_BACKMERGE] = {{ "M", "backmerge" }, blk_log_generic }, + [__BLK_TA_FRONTMERGE] = {{ "F", "frontmerge" }, blk_log_generic }, + [__BLK_TA_GETRQ] = {{ "G", "getrq" }, blk_log_generic }, + [__BLK_TA_SLEEPRQ] = {{ "S", "sleeprq" }, blk_log_generic }, + [__BLK_TA_REQUEUE] = {{ "R", "requeue" }, blk_log_with_error }, + [__BLK_TA_ISSUE] = {{ "D", "issue" }, blk_log_generic }, + [__BLK_TA_COMPLETE] = {{ "C", "complete" }, blk_log_with_error }, + [__BLK_TA_PLUG] = {{ "P", "plug" }, blk_log_plug }, + [__BLK_TA_UNPLUG_IO] = {{ "U", "unplug_io" }, blk_log_unplug }, + [__BLK_TA_UNPLUG_TIMER] = {{ "UT", "unplug_timer" }, blk_log_unplug }, + [__BLK_TA_INSERT] = {{ "I", "insert" }, blk_log_generic }, + [__BLK_TA_SPLIT] = {{ "X", "split" }, blk_log_split }, + [__BLK_TA_BOUNCE] = {{ "B", "bounce" }, blk_log_generic }, + [__BLK_TA_REMAP] = {{ "A", "remap" }, blk_log_remap }, +}; + +static int blk_trace_event_print(struct trace_seq *s, struct trace_entry *ent, + int flags) +{ + const struct blk_io_trace *t = (struct blk_io_trace *)ent; + const u16 what = t->action & ((1 << BLK_TC_SHIFT) - 1); + int ret; + + if (unlikely(what == 0 || what > ARRAY_SIZE(what2act))) + ret = trace_seq_printf(s, "Bad pc action %x\n", what); + else { + const bool long_act = !!(trace_flags & TRACE_ITER_VERBOSE); + ret = blk_log_action_seq(s, t, what2act[what].act[long_act]); + if (ret) + ret = what2act[what].print(s, ent); + } + + return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; +} + +static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter) +{ + const struct blk_io_trace *t; + u16 what; + int ret; + + if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC)) + return TRACE_TYPE_UNHANDLED; + + t = (const struct blk_io_trace *)iter->ent; + what = t->action & ((1 << BLK_TC_SHIFT) - 1); + + if (unlikely(what == 0 || what > ARRAY_SIZE(what2act))) + ret = trace_seq_printf(&iter->seq, "Bad pc action %x\n", what); + else { + const bool long_act = !!(trace_flags & TRACE_ITER_VERBOSE); + ret = blk_log_action_iter(iter, what2act[what].act[long_act]); + if (ret) + ret = what2act[what].print(&iter->seq, iter->ent); + } + + return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; +} + +static struct tracer blk_tracer __read_mostly = { + .name = "blk", + .init = blk_tracer_init, + .reset = blk_tracer_reset, + .start = blk_tracer_start, + .stop = blk_tracer_stop, + .print_header = blk_tracer_print_header, + .print_line = blk_tracer_print_line, + .flags = &blk_tracer_flags, +}; + +static struct trace_event trace_blk_event = { + .type = TRACE_BLK, + .trace = blk_trace_event_print, + .latency_trace = blk_trace_event_print, + .raw = trace_nop_print, + .hex = trace_nop_print, + .binary = trace_nop_print, +}; + +static int __init init_blk_tracer(void) +{ + if (!register_ftrace_event(&trace_blk_event)) { + pr_warning("Warning: could not register block events\n"); + return 1; + } + + if (register_tracer(&blk_tracer) != 0) { + pr_warning("Warning: could not register the block tracer\n"); + unregister_ftrace_event(&trace_blk_event); + return 1; + } + + return 0; +} + +device_initcall(init_blk_tracer); + +static int blk_trace_remove_queue(struct request_queue *q) +{ + struct blk_trace *bt; + + bt = xchg(&q->blk_trace, NULL); + if (bt == NULL) + return -EINVAL; + + kfree(bt); + return 0; +} + +/* + * Setup everything required to start tracing + */ +static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) +{ + struct blk_trace *old_bt, *bt = NULL; + int ret; + + ret = -ENOMEM; + bt = kzalloc(sizeof(*bt), GFP_KERNEL); + if (!bt) + goto err; + + bt->dev = dev; + bt->act_mask = (u16)-1; + bt->end_lba = -1ULL; + bt->trace_state = Blktrace_running; + + old_bt = xchg(&q->blk_trace, bt); + if (old_bt != NULL) { + (void)xchg(&q->blk_trace, old_bt); + kfree(bt); + ret = -EBUSY; + } + return 0; +err: + return ret; +} + +/* + * sysfs interface to enable and configure tracing + */ + +static ssize_t sysfs_blk_trace_enable_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct hd_struct *p = dev_to_part(dev); + struct block_device *bdev; + ssize_t ret = -ENXIO; + + lock_kernel(); + bdev = bdget(part_devt(p)); + if (bdev != NULL) { + struct request_queue *q = bdev_get_queue(bdev); + + if (q != NULL) { + mutex_lock(&bdev->bd_mutex); + ret = sprintf(buf, "%u\n", !!q->blk_trace); + mutex_unlock(&bdev->bd_mutex); + } + + bdput(bdev); + } + + unlock_kernel(); + return ret; +} + +static ssize_t sysfs_blk_trace_enable_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct block_device *bdev; + struct request_queue *q; + struct hd_struct *p; + int value; + ssize_t ret = -ENXIO; + + if (count == 0 || sscanf(buf, "%d", &value) != 1) + goto out; + + lock_kernel(); + p = dev_to_part(dev); + bdev = bdget(part_devt(p)); + if (bdev == NULL) + goto out_unlock_kernel; + + q = bdev_get_queue(bdev); + if (q == NULL) + goto out_bdput; + + mutex_lock(&bdev->bd_mutex); + if (value) + ret = blk_trace_setup_queue(q, bdev->bd_dev); + else + ret = blk_trace_remove_queue(q); + mutex_unlock(&bdev->bd_mutex); + + if (ret == 0) + ret = count; +out_bdput: + bdput(bdev); +out_unlock_kernel: + unlock_kernel(); +out: + return ret; +} + +static ssize_t sysfs_blk_trace_attr_show(struct device *dev, + struct device_attribute *attr, + char *buf); +static ssize_t sysfs_blk_trace_attr_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count); +#define BLK_TRACE_DEVICE_ATTR(_name) \ + DEVICE_ATTR(_name, S_IRUGO | S_IWUSR, \ + sysfs_blk_trace_attr_show, \ + sysfs_blk_trace_attr_store) + +static DEVICE_ATTR(enable, S_IRUGO | S_IWUSR, + sysfs_blk_trace_enable_show, sysfs_blk_trace_enable_store); +static BLK_TRACE_DEVICE_ATTR(act_mask); +static BLK_TRACE_DEVICE_ATTR(pid); +static BLK_TRACE_DEVICE_ATTR(start_lba); +static BLK_TRACE_DEVICE_ATTR(end_lba); + +static struct attribute *blk_trace_attrs[] = { + &dev_attr_enable.attr, + &dev_attr_act_mask.attr, + &dev_attr_pid.attr, + &dev_attr_start_lba.attr, + &dev_attr_end_lba.attr, + NULL +}; + +struct attribute_group blk_trace_attr_group = { + .name = "trace", + .attrs = blk_trace_attrs, +}; + +static int blk_str2act_mask(const char *str) +{ + int mask = 0; + char *copy = kstrdup(str, GFP_KERNEL), *s; + + if (copy == NULL) + return -ENOMEM; + + s = strstrip(copy); + + while (1) { + char *sep = strchr(s, ','); + + if (sep != NULL) + *sep = '\0'; + + if (strcasecmp(s, "barrier") == 0) + mask |= BLK_TC_BARRIER; + else if (strcasecmp(s, "complete") == 0) + mask |= BLK_TC_COMPLETE; + else if (strcasecmp(s, "fs") == 0) + mask |= BLK_TC_FS; + else if (strcasecmp(s, "issue") == 0) + mask |= BLK_TC_ISSUE; + else if (strcasecmp(s, "pc") == 0) + mask |= BLK_TC_PC; + else if (strcasecmp(s, "queue") == 0) + mask |= BLK_TC_QUEUE; + else if (strcasecmp(s, "read") == 0) + mask |= BLK_TC_READ; + else if (strcasecmp(s, "requeue") == 0) + mask |= BLK_TC_REQUEUE; + else if (strcasecmp(s, "sync") == 0) + mask |= BLK_TC_SYNC; + else if (strcasecmp(s, "write") == 0) + mask |= BLK_TC_WRITE; + + if (sep == NULL) + break; + + s = sep + 1; + } + kfree(copy); + + return mask; +} + +static ssize_t sysfs_blk_trace_attr_show(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct hd_struct *p = dev_to_part(dev); + struct request_queue *q; + struct block_device *bdev; + ssize_t ret = -ENXIO; + + lock_kernel(); + bdev = bdget(part_devt(p)); + if (bdev == NULL) + goto out_unlock_kernel; + + q = bdev_get_queue(bdev); + if (q == NULL) + goto out_bdput; + mutex_lock(&bdev->bd_mutex); + if (q->blk_trace == NULL) + ret = sprintf(buf, "disabled\n"); + else if (attr == &dev_attr_act_mask) + ret = sprintf(buf, "%#x\n", q->blk_trace->act_mask); + else if (attr == &dev_attr_pid) + ret = sprintf(buf, "%u\n", q->blk_trace->pid); + else if (attr == &dev_attr_start_lba) + ret = sprintf(buf, "%llu\n", q->blk_trace->start_lba); + else if (attr == &dev_attr_end_lba) + ret = sprintf(buf, "%llu\n", q->blk_trace->end_lba); + mutex_unlock(&bdev->bd_mutex); +out_bdput: + bdput(bdev); +out_unlock_kernel: + unlock_kernel(); + return ret; +} + +static ssize_t sysfs_blk_trace_attr_store(struct device *dev, + struct device_attribute *attr, + const char *buf, size_t count) +{ + struct block_device *bdev; + struct request_queue *q; + struct hd_struct *p; + u64 value; + ssize_t ret = -ENXIO; + + if (count == 0) + goto out; + + if (attr == &dev_attr_act_mask) { + if (sscanf(buf, "%llx", &value) != 1) { + /* Assume it is a list of trace category names */ + value = blk_str2act_mask(buf); + if (value < 0) + goto out; + } + } else if (sscanf(buf, "%llu", &value) != 1) + goto out; + + lock_kernel(); + p = dev_to_part(dev); + bdev = bdget(part_devt(p)); + if (bdev == NULL) + goto out_unlock_kernel; + + q = bdev_get_queue(bdev); + if (q == NULL) + goto out_bdput; + + mutex_lock(&bdev->bd_mutex); + ret = 0; + if (q->blk_trace == NULL) + ret = blk_trace_setup_queue(q, bdev->bd_dev); + + if (ret == 0) { + if (attr == &dev_attr_act_mask) + q->blk_trace->act_mask = value; + else if (attr == &dev_attr_pid) + q->blk_trace->pid = value; + else if (attr == &dev_attr_start_lba) + q->blk_trace->start_lba = value; + else if (attr == &dev_attr_end_lba) + q->blk_trace->end_lba = value; + ret = count; + } + mutex_unlock(&bdev->bd_mutex); +out_bdput: + bdput(bdev); +out_unlock_kernel: + unlock_kernel(); +out: + return ret; +} diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 6d720243f5f4..01714efdc65a 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -268,6 +268,10 @@ ssize_t part_fail_store(struct device *dev, } #endif +#ifdef CONFIG_BLK_DEV_IO_TRACE +extern struct attribute_group blk_trace_attr_group; +#endif + static DEVICE_ATTR(partition, S_IRUGO, part_partition_show, NULL); static DEVICE_ATTR(start, S_IRUGO, part_start_show, NULL); static DEVICE_ATTR(size, S_IRUGO, part_size_show, NULL); @@ -294,6 +298,9 @@ static struct attribute_group part_attr_group = { static struct attribute_group *part_attr_groups[] = { &part_attr_group, +#ifdef CONFIG_BLK_DEV_IO_TRACE + &blk_trace_attr_group, +#endif NULL }; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index b96037d970df..e603a291134b 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -32,6 +32,7 @@ enum trace_type { TRACE_KMEM_ALLOC, TRACE_KMEM_FREE, TRACE_POWER, + TRACE_BLK, __TRACE_LAST_TYPE, }; From 157f9c00e88529ed84bd7d581a40d411e5414cf0 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 26 Jan 2009 15:00:56 -0200 Subject: [PATCH 083/413] tracing/blktrace: fix up checkpatch reported problems in ftrace plugin patch Also make sure sparse (make C=2 block/blktrace.o) is happy too. Reported-by: Ingo Molnar Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- block/blktrace.c | 40 ++++++++++++++++++++++-------------- fs/partitions/check.c | 5 +---- include/linux/blktrace_api.h | 5 +++++ 3 files changed, 31 insertions(+), 19 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 630f167f8240..1b2267c798b6 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -37,7 +37,7 @@ static int __read_mostly blk_tracer_enabled; static struct tracer_opt blk_tracer_opts[] = { /* Default disable the minimalistic output */ - { TRACER_OPT(blk_classic, TRACE_BLK_OPT_CLASSIC ) }, + { TRACER_OPT(blk_classic, TRACE_BLK_OPT_CLASSIC) }, { } }; @@ -169,7 +169,8 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, pid_t pid; int cpu, pc = 0; - if (unlikely(bt->trace_state != Blktrace_running || !blk_tracer_enabled)) + if (unlikely(bt->trace_state != Blktrace_running || + !blk_tracer_enabled)) return; what |= ddir_act[rw & WRITE]; @@ -192,7 +193,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, sizeof(*t) + pdu_len, &flags); if (!event) return; - + ent = ring_buffer_event_data(event); t = (struct blk_io_trace *)ent; pc = preempt_count(); @@ -234,7 +235,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, if (blk_tr) { ring_buffer_unlock_commit(blk_tr->buffer, event, flags); if (pid != 0 && - (blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC) == 0 && + !(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC) && (trace_flags & TRACE_ITER_STACKTRACE) != 0) __trace_stack(blk_tr, NULL, flags, 5, pc); trace_wake_up(); @@ -955,19 +956,27 @@ static void blk_unregister_tracepoints(void) static void fill_rwbs(char *rwbs, const struct blk_io_trace *t) { - int i = 0; + int i = 0; - if (t->action & BLK_TC_DISCARD) rwbs[i++] = 'D'; - else if (t->action & BLK_TC_WRITE) rwbs[i++] = 'W'; - else if (t->bytes) rwbs[i++] = 'R'; - else rwbs[i++] = 'N'; + if (t->action & BLK_TC_DISCARD) + rwbs[i++] = 'D'; + else if (t->action & BLK_TC_WRITE) + rwbs[i++] = 'W'; + else if (t->bytes) + rwbs[i++] = 'R'; + else + rwbs[i++] = 'N'; - if (t->action & BLK_TC_AHEAD) rwbs[i++] = 'A'; - if (t->action & BLK_TC_BARRIER) rwbs[i++] = 'B'; - if (t->action & BLK_TC_SYNC) rwbs[i++] = 'S'; - if (t->action & BLK_TC_META) rwbs[i++] = 'M'; + if (t->action & BLK_TC_AHEAD) + rwbs[i++] = 'A'; + if (t->action & BLK_TC_BARRIER) + rwbs[i++] = 'B'; + if (t->action & BLK_TC_SYNC) + rwbs[i++] = 'S'; + if (t->action & BLK_TC_META) + rwbs[i++] = 'M'; - rwbs[i] = '\0'; + rwbs[i] = '\0'; } static inline @@ -1049,7 +1058,8 @@ static int blk_log_generic(struct trace_seq *s, const struct trace_entry *ent) return trace_seq_printf(s, "[%s]\n", cmd); } -static int blk_log_with_error(struct trace_seq *s, const struct trace_entry *ent) +static int blk_log_with_error(struct trace_seq *s, + const struct trace_entry *ent) { if (t_sec(ent)) return trace_seq_printf(s, "%llu + %u [%d]\n", t_sector(ent), diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 01714efdc65a..8a17f7edcc74 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -19,6 +19,7 @@ #include #include #include +#include #include "check.h" @@ -268,10 +269,6 @@ ssize_t part_fail_store(struct device *dev, } #endif -#ifdef CONFIG_BLK_DEV_IO_TRACE -extern struct attribute_group blk_trace_attr_group; -#endif - static DEVICE_ATTR(partition, S_IRUGO, part_partition_show, NULL); static DEVICE_ATTR(start, S_IRUGO, part_start_show, NULL); static DEVICE_ATTR(size, S_IRUGO, part_size_show, NULL); diff --git a/include/linux/blktrace_api.h b/include/linux/blktrace_api.h index 1dba3493d520..59b4b2e8ab67 100644 --- a/include/linux/blktrace_api.h +++ b/include/linux/blktrace_api.h @@ -142,6 +142,9 @@ struct blk_user_trace_setup { #ifdef __KERNEL__ #if defined(CONFIG_BLK_DEV_IO_TRACE) + +#include + struct blk_trace { int trace_state; struct rchan *rchan; @@ -192,6 +195,8 @@ extern int blk_trace_setup(struct request_queue *q, char *name, dev_t dev, extern int blk_trace_startstop(struct request_queue *q, int start); extern int blk_trace_remove(struct request_queue *q); +extern struct attribute_group blk_trace_attr_group; + #else /* !CONFIG_BLK_DEV_IO_TRACE */ #define blk_trace_ioctl(bdev, cmd, arg) (-ENOTTY) #define blk_trace_shutdown(q) do { } while (0) From 32c0bd9624115041cfec31c0436995418083090a Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 26 Jan 2009 16:00:40 -0200 Subject: [PATCH 084/413] blktrace: the ftrace interface needs CONFIG_TRACING Impact: build fix Also mention in the help text that blktrace now can be used using the ftrace interface. Reported-by: Ingo Molnar Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- block/Kconfig | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/block/Kconfig b/block/Kconfig index 0cbb3b88b59a..7cdaa1d72252 100644 --- a/block/Kconfig +++ b/block/Kconfig @@ -50,6 +50,8 @@ config BLK_DEV_IO_TRACE select RELAY select DEBUG_FS select TRACEPOINTS + select TRACING + select STACKTRACE help Say Y here if you want to be able to trace the block layer actions on a given queue. Tracing allows you to see any traffic happening @@ -58,6 +60,12 @@ config BLK_DEV_IO_TRACE git://git.kernel.dk/blktrace.git + Tracing also is possible using the ftrace interface, e.g.: + + echo 1 > /sys/block/sda/sda1/trace/enable + echo blk > /sys/kernel/debug/tracing/current_tracer + cat /sys/kernel/debug/tracing/trace_pipe + If unsure, say N. config BLK_DEV_BSG From 890252823766e562301e61340f3187a14033d045 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Mon, 26 Jan 2009 18:28:02 +0300 Subject: [PATCH 085/413] x86: ftrace - simplify wait_for_nmi Get rid of 'waited' stack variable. Signed-off-by: Cyrill Gorcunov Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 1b43086b097a..4d33224c055f 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -133,15 +133,14 @@ void ftrace_nmi_exit(void) static void wait_for_nmi(void) { - int waited = 0; + if (!atomic_read(&in_nmi)) + return; - while (atomic_read(&in_nmi)) { - waited = 1; + do { cpu_relax(); - } + } while(atomic_read(&in_nmi)); - if (waited) - nmi_wait_count++; + nmi_wait_count++; } static int From 700a3dcb9036d17d3a67d0a7ceee9d4373fbb570 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 28 Jan 2009 12:33:56 -0200 Subject: [PATCH 086/413] blktrace: Use tracing_reset_online_cpus Impact: cleanup Use tracing_reset_online_cpus instead of open coding it. Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- block/blktrace.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 1b2267c798b6..04d81d31fd94 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -1109,12 +1109,7 @@ static void blk_tracer_print_header(struct seq_file *m) static void blk_tracer_start(struct trace_array *tr) { - int cpu; - - tr->time_start = ftrace_now(tr->cpu); - - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); + tracing_reset_online_cpus(tr); mutex_lock(&blk_probe_mutex); if (atomic_add_return(1, &blk_probes_ref) == 1) From f04109bf1be7449e27d38ae1bb8465013374bd49 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 28 Jan 2009 13:02:12 -0200 Subject: [PATCH 087/413] trace: Use tracing_reset_online_cpus in more places MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: cleanup Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Frédéric Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_branch.c | 6 +----- kernel/trace/trace_functions_graph.c | 8 ++------ kernel/trace/trace_nop.c | 6 +----- 3 files changed, 4 insertions(+), 16 deletions(-) diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index ca017e0a9a27..1284145c8898 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -133,11 +133,7 @@ static void stop_branch_trace(struct trace_array *tr) static int branch_trace_init(struct trace_array *tr) { - int cpu; - - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); - + tracing_reset_online_cpus(tr); start_branch_trace(tr); return 0; } diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 66fc7b806a38..c97594d826bc 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -52,15 +52,11 @@ static struct tracer_flags tracer_flags = { static int graph_trace_init(struct trace_array *tr) { - int cpu, ret; - - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); - - ret = register_ftrace_graph(&trace_graph_return, + int ret = register_ftrace_graph(&trace_graph_return, &trace_graph_entry); if (ret) return ret; + tracing_reset_online_cpus(tr); tracing_start_cmdline_record(); return 0; diff --git a/kernel/trace/trace_nop.c b/kernel/trace/trace_nop.c index b9767acd30ac..087b6cbf4ea5 100644 --- a/kernel/trace/trace_nop.c +++ b/kernel/trace/trace_nop.c @@ -47,12 +47,8 @@ static void stop_nop_trace(struct trace_array *tr) static int nop_trace_init(struct trace_array *tr) { - int cpu; ctx_trace = tr; - - for_each_online_cpu(cpu) - tracing_reset(tr, cpu); - + tracing_reset_online_cpus(tr); start_nop_trace(tr); return 0; } From b3a8c34886d0e3dd3a24a5b614ee025181da2f41 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 28 Jan 2009 13:08:37 -0200 Subject: [PATCH 088/413] trace_sched_wakeup: Remove unused variable Impact: cleanup Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- kernel/trace/trace_sched_wakeup.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 93cecda650b2..a48c9b4b0c85 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -184,13 +184,10 @@ probe_wakeup_sched_switch(struct rq *rq, struct task_struct *prev, static void __wakeup_reset(struct trace_array *tr) { - struct trace_array_cpu *data; int cpu; - for_each_possible_cpu(cpu) { - data = tr->data[cpu]; + for_each_possible_cpu(cpu) tracing_reset(tr, cpu); - } wakeup_cpu = -1; wakeup_prio = -1; From ecf441b593ac41cb8cd8cd3695110167c42e098c Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Thu, 29 Jan 2009 13:49:45 -0800 Subject: [PATCH 089/413] kmemtrace: fix printk formats, fix Geert Uytterhoeven wrote: > %4zu? Reported-by: Geert Uytterhoeven Signed-off-by: Randy Dunlap Acked-by: Eduard - Gabriel Munteanu Signed-off-by: Ingo Molnar --- kernel/trace/kmemtrace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c index 72b326bba005..f04c0625f1cd 100644 --- a/kernel/trace/kmemtrace.c +++ b/kernel/trace/kmemtrace.c @@ -139,12 +139,12 @@ kmemtrace_print_alloc_compress(struct trace_iterator *iter, return TRACE_TYPE_PARTIAL_LINE; /* Requested */ - ret = trace_seq_printf(s, "%4zd ", entry->bytes_req); + ret = trace_seq_printf(s, "%4zu ", entry->bytes_req); if (!ret) return TRACE_TYPE_PARTIAL_LINE; /* Allocated */ - ret = trace_seq_printf(s, "%4zd ", entry->bytes_alloc); + ret = trace_seq_printf(s, "%4zu ", entry->bytes_alloc); if (!ret) return TRACE_TYPE_PARTIAL_LINE; From b2821ae68b14480bfc85ea1629537163310bc5cd Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Feb 2009 21:38:32 -0500 Subject: [PATCH 090/413] trace: fix default boot up tracer Peter Zijlstra started the functionality to start up a default tracing at bootup. This patch finishes the work. Now if you add 'ftrace=' to the command line, when that tracer is registered on bootup, that tracer is selected and starts tracing. Note, all selftests for tracers that are registered after this tracer is disabled. This prevents the selftests from disturbing the running tracer, or the running tracer from disturbing the selftest. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 60 +++++++++++++++++++++++++++++++++++++++----- 1 file changed, 54 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 2f8ac1f008f5..2c720c79bc60 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -53,6 +53,11 @@ unsigned long __read_mostly tracing_thresh; */ static bool __read_mostly tracing_selftest_running; +/* + * If a tracer is running, we do not want to run SELFTEST. + */ +static bool __read_mostly tracing_selftest_disabled; + /* For tracers that don't implement custom flags */ static struct tracer_opt dummy_tracer_opt[] = { { } @@ -110,14 +115,19 @@ static cpumask_var_t __read_mostly tracing_buffer_mask; */ int ftrace_dump_on_oops; -static int tracing_set_tracer(char *buf); +static int tracing_set_tracer(const char *buf); + +#define BOOTUP_TRACER_SIZE 100 +static char bootup_tracer_buf[BOOTUP_TRACER_SIZE] __initdata; +static char *default_bootup_tracer; static int __init set_ftrace(char *str) { - tracing_set_tracer(str); + strncpy(bootup_tracer_buf, str, BOOTUP_TRACER_SIZE); + default_bootup_tracer = bootup_tracer_buf; return 1; } -__setup("ftrace", set_ftrace); +__setup("ftrace=", set_ftrace); static int __init set_ftrace_dump_on_oops(char *str) { @@ -468,7 +478,7 @@ int register_tracer(struct tracer *type) type->flags->opts = dummy_tracer_opt; #ifdef CONFIG_FTRACE_STARTUP_TEST - if (type->selftest) { + if (type->selftest && !tracing_selftest_disabled) { struct tracer *saved_tracer = current_trace; struct trace_array *tr = &global_trace; int i; @@ -510,8 +520,25 @@ int register_tracer(struct tracer *type) out: tracing_selftest_running = false; mutex_unlock(&trace_types_lock); - lock_kernel(); + if (!ret && default_bootup_tracer) { + if (!strncmp(default_bootup_tracer, type->name, + BOOTUP_TRACER_SIZE)) { + printk(KERN_INFO "Starting tracer '%s'\n", + type->name); + /* Do we want this tracer to start on bootup? */ + tracing_set_tracer(type->name); + default_bootup_tracer = NULL; + /* disable other selftests, since this will break it. */ + tracing_selftest_disabled = 1; +#ifdef CONFIG_FTRACE_STARTUP_TEST + printk(KERN_INFO "Disabling FTRACE selftests due" + " to running tracer '%s'\n", type->name); +#endif + } + } + + lock_kernel(); return ret; } @@ -2245,7 +2272,7 @@ tracing_set_trace_read(struct file *filp, char __user *ubuf, return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); } -static int tracing_set_tracer(char *buf) +static int tracing_set_tracer(const char *buf) { struct trace_array *tr = &global_trace; struct tracer *t; @@ -3163,5 +3190,26 @@ __init static int tracer_alloc_buffers(void) out: return ret; } + +__init static int clear_boot_tracer(void) +{ + /* + * The default tracer at boot buffer is an init section. + * This function is called in lateinit. If we did not + * find the boot tracer, then clear it out, to prevent + * later registration from accessing the buffer that is + * about to be freed. + */ + if (!default_bootup_tracer) + return 0; + + printk(KERN_INFO "ftrace bootup tracer '%s' not registered.\n", + default_bootup_tracer); + default_bootup_tracer = NULL; + + return 0; +} + early_initcall(tracer_alloc_buffers); fs_initcall(tracer_init_debugfs); +late_initcall(clear_boot_tracer); From 79fb0768fbd371f3b94d909f51f587b3a24ab272 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Feb 2009 21:38:33 -0500 Subject: [PATCH 091/413] trace: let boot trace be chosen by command line Now that we have a working ftrace= function, make the boot tracer get activated by it. This way we can turn it on or off without recompiling the kernel, as well as keeping the selftests on. The selftests are disabled whenever a default tracer starts running. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 7 +++---- kernel/trace/trace.c | 5 +---- kernel/trace/trace_boot.c | 11 +++++++---- 3 files changed, 11 insertions(+), 12 deletions(-) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index dde1d46f77e5..28f2644484d9 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -164,9 +164,8 @@ config BOOT_TRACER representation of the delays during initcalls - but the raw /debug/tracing/trace text output is readable too. - ( Note that tracing self tests can't be enabled if this tracer is - selected, because the self-tests are an initcall as well and that - would invalidate the boot trace. ) + You must pass in ftrace=initcall to the kernel command line + to enable this on bootup. config TRACE_BRANCH_PROFILING bool "Trace likely/unlikely profiler" @@ -326,7 +325,7 @@ config FTRACE_SELFTEST config FTRACE_STARTUP_TEST bool "Perform a startup test on ftrace" - depends on TRACING && DEBUG_KERNEL && !BOOT_TRACER + depends on TRACING && DEBUG_KERNEL select FTRACE_SELFTEST help This option performs a series of startup tests on ftrace. On bootup diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 2c720c79bc60..40edef4255c5 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3167,12 +3167,9 @@ __init static int tracer_alloc_buffers(void) trace_init_cmdlines(); register_tracer(&nop_trace); + current_trace = &nop_trace; #ifdef CONFIG_BOOT_TRACER register_tracer(&boot_tracer); - current_trace = &boot_tracer; - current_trace->init(&global_trace); -#else - current_trace = &nop_trace; #endif /* All seems OK, enable tracing */ tracing_disabled = 0; diff --git a/kernel/trace/trace_boot.c b/kernel/trace/trace_boot.c index 0e94b3d091f7..1f07895977a0 100644 --- a/kernel/trace/trace_boot.c +++ b/kernel/trace/trace_boot.c @@ -28,13 +28,13 @@ void start_boot_trace(void) void enable_boot_trace(void) { - if (pre_initcalls_finished) + if (boot_trace && pre_initcalls_finished) tracing_start_sched_switch_record(); } void disable_boot_trace(void) { - if (pre_initcalls_finished) + if (boot_trace && pre_initcalls_finished) tracing_stop_sched_switch_record(); } @@ -43,6 +43,9 @@ static int boot_trace_init(struct trace_array *tr) int cpu; boot_trace = tr; + if (!tr) + return 0; + for_each_cpu(cpu, cpu_possible_mask) tracing_reset(tr, cpu); @@ -132,7 +135,7 @@ void trace_boot_call(struct boot_trace_call *bt, initcall_t fn) unsigned long irq_flags; struct trace_array *tr = boot_trace; - if (!pre_initcalls_finished) + if (!tr || !pre_initcalls_finished) return; /* Get its name now since this function could @@ -164,7 +167,7 @@ void trace_boot_ret(struct boot_trace_ret *bt, initcall_t fn) unsigned long irq_flags; struct trace_array *tr = boot_trace; - if (!pre_initcalls_finished) + if (!tr || !pre_initcalls_finished) return; sprint_symbol(bt->func, (unsigned long)fn); From c4a8e8be2d43cc22b371e8e9c05c253409759d94 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 2 Feb 2009 20:29:21 -0200 Subject: [PATCH 092/413] trace: better manage the context info for events Impact: make trace_event more convenient for tracers All tracers (for the moment) that use the struct trace_event want to have the context info printed before their own output: the pid/cmdline, cpu, and timestamp. But some other tracers that want to implement their trace_event callbacks will not necessary need these information or they may want to format them as they want. This patch adds a new default-enabled trace option: TRACE_ITER_CONTEXT_INFO When disabled through: echo nocontext-info > /debugfs/tracing/trace_options The pid, cpu and timestamps headers will not be printed. IE with the sched_switch tracer with context-info (default): bash-2935 [001] 100.356561: 2935:120:S ==> [001] 0:140:R -0 [000] 100.412804: 0:140:R + [000] 11:115:S events/0 -0 [000] 100.412816: 0:140:R ==> [000] 11:115:R events/0 events/0-11 [000] 100.412829: 11:115:S ==> [000] 0:140:R Without context-info: 2935:120:S ==> [001] 0:140:R 0:140:R + [000] 11:115:S events/0 0:140:R ==> [000] 11:115:R events/0 11:115:S ==> [000] 0:140:R A tracer can disable it at runtime by clearing the bit TRACE_ITER_CONTEXT_INFO in trace_flags. The print routines were renamed to trace_print_context and trace_print_lat_context, so that they can be used by tracers if they want to use them for one of the trace_event callbacks. Signed-off-by: Frederic Weisbecker Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 149 +++++++++--------------------------- kernel/trace/trace.h | 7 +- kernel/trace/trace_output.c | 107 ++++++++++++++++++++++++++ kernel/trace/trace_output.h | 3 + 4 files changed, 151 insertions(+), 115 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 2f8ac1f008f5..5ec49c3c1597 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -227,7 +227,7 @@ static DECLARE_WAIT_QUEUE_HEAD(trace_wait); /* trace_flags holds trace_options default values */ unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK | - TRACE_ITER_ANNOTATE; + TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO; /** * trace_wake_up - wake up tasks waiting for trace input @@ -285,6 +285,7 @@ static const char *trace_options[] = { "userstacktrace", "sym-userobj", "printk-msg-only", + "context-info", NULL }; @@ -1171,8 +1172,8 @@ __find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts) } /* Find the next real entry, without updating the iterator itself */ -static struct trace_entry * -find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts) +struct trace_entry *trace_find_next_entry(struct trace_iterator *iter, + int *ent_cpu, u64 *ent_ts) { return __find_next_entry(iter, ent_cpu, ent_ts); } @@ -1351,57 +1352,6 @@ print_trace_header(struct seq_file *m, struct trace_iterator *iter) seq_puts(m, "\n"); } -static void -lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu) -{ - int hardirq, softirq; - char *comm; - - comm = trace_find_cmdline(entry->pid); - - trace_seq_printf(s, "%8.8s-%-5d ", comm, entry->pid); - trace_seq_printf(s, "%3d", cpu); - trace_seq_printf(s, "%c%c", - (entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : - (entry->flags & TRACE_FLAG_IRQS_NOSUPPORT) ? 'X' : '.', - ((entry->flags & TRACE_FLAG_NEED_RESCHED) ? 'N' : '.')); - - hardirq = entry->flags & TRACE_FLAG_HARDIRQ; - softirq = entry->flags & TRACE_FLAG_SOFTIRQ; - if (hardirq && softirq) { - trace_seq_putc(s, 'H'); - } else { - if (hardirq) { - trace_seq_putc(s, 'h'); - } else { - if (softirq) - trace_seq_putc(s, 's'); - else - trace_seq_putc(s, '.'); - } - } - - if (entry->preempt_count) - trace_seq_printf(s, "%x", entry->preempt_count); - else - trace_seq_puts(s, "."); -} - -unsigned long preempt_mark_thresh = 100; - -static void -lat_print_timestamp(struct trace_seq *s, u64 abs_usecs, - unsigned long rel_usecs) -{ - trace_seq_printf(s, " %4lldus", abs_usecs); - if (rel_usecs > preempt_mark_thresh) - trace_seq_puts(s, "!: "); - else if (rel_usecs > 1) - trace_seq_puts(s, "+: "); - else - trace_seq_puts(s, " : "); -} - static void test_cpu_buff_start(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; @@ -1419,46 +1369,24 @@ static void test_cpu_buff_start(struct trace_iterator *iter) trace_seq_printf(s, "##### CPU %u buffer started ####\n", iter->cpu); } -static enum print_line_t -print_lat_fmt(struct trace_iterator *iter, unsigned int trace_idx, int cpu) +static enum print_line_t print_lat_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); - struct trace_entry *next_entry; struct trace_event *event; - unsigned long verbose = (trace_flags & TRACE_ITER_VERBOSE); struct trace_entry *entry = iter->ent; - unsigned long abs_usecs; - unsigned long rel_usecs; - u64 next_ts; - char *comm; int ret; test_cpu_buff_start(iter); - next_entry = find_next_entry(iter, NULL, &next_ts); - if (!next_entry) - next_ts = iter->ts; - rel_usecs = ns2usecs(next_ts - iter->ts); - abs_usecs = ns2usecs(iter->ts - iter->tr->time_start); + event = ftrace_find_event(entry->type); - if (verbose) { - comm = trace_find_cmdline(entry->pid); - trace_seq_printf(s, "%16s %5d %3d %d %08x %08x [%08lx]" - " %ld.%03ldms (+%ld.%03ldms): ", - comm, - entry->pid, cpu, entry->flags, - entry->preempt_count, trace_idx, - ns2usecs(iter->ts), - abs_usecs/1000, - abs_usecs % 1000, rel_usecs/1000, - rel_usecs % 1000); - } else { - lat_print_generic(s, entry, cpu); - lat_print_timestamp(s, abs_usecs, rel_usecs); + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + ret = trace_print_lat_context(iter); + if (ret) + return ret; } - event = ftrace_find_event(entry->type); if (event && event->latency_trace) { ret = event->latency_trace(s, entry, sym_flags); if (ret) @@ -1476,33 +1404,20 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); struct trace_entry *entry; struct trace_event *event; - unsigned long usec_rem; - unsigned long long t; - unsigned long secs; - char *comm; int ret; entry = iter->ent; test_cpu_buff_start(iter); - comm = trace_find_cmdline(iter->ent->pid); - - t = ns2usecs(iter->ts); - usec_rem = do_div(t, 1000000ULL); - secs = (unsigned long)t; - - ret = trace_seq_printf(s, "%16s-%-5d ", comm, entry->pid); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, "[%03d] ", iter->cpu); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_printf(s, "%5lu.%06lu: ", secs, usec_rem); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - event = ftrace_find_event(entry->type); + + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + ret = trace_print_context(iter); + if (ret) + return ret; + } + if (event && event->trace) { ret = event->trace(s, entry, sym_flags); if (ret) @@ -1525,10 +1440,12 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) entry = iter->ent; - ret = trace_seq_printf(s, "%d %d %llu ", - entry->pid, iter->cpu, iter->ts); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + ret = trace_seq_printf(s, "%d %d %llu ", + entry->pid, iter->cpu, iter->ts); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } event = ftrace_find_event(entry->type); if (event && event->raw) { @@ -1553,9 +1470,11 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) entry = iter->ent; - SEQ_PUT_HEX_FIELD_RET(s, entry->pid); - SEQ_PUT_HEX_FIELD_RET(s, iter->cpu); - SEQ_PUT_HEX_FIELD_RET(s, iter->ts); + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + SEQ_PUT_HEX_FIELD_RET(s, entry->pid); + SEQ_PUT_HEX_FIELD_RET(s, iter->cpu); + SEQ_PUT_HEX_FIELD_RET(s, iter->ts); + } event = ftrace_find_event(entry->type); if (event && event->hex) @@ -1575,7 +1494,7 @@ static enum print_line_t print_printk_msg_only(struct trace_iterator *iter) trace_assign_type(field, entry); - ret = trace_seq_printf(s, field->buf); + ret = trace_seq_printf(s, "%s", field->buf); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -1590,9 +1509,11 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) entry = iter->ent; - SEQ_PUT_FIELD_RET(s, entry->pid); - SEQ_PUT_FIELD_RET(s, entry->cpu); - SEQ_PUT_FIELD_RET(s, iter->ts); + if (trace_flags & TRACE_ITER_CONTEXT_INFO) { + SEQ_PUT_FIELD_RET(s, entry->pid); + SEQ_PUT_FIELD_RET(s, entry->cpu); + SEQ_PUT_FIELD_RET(s, iter->ts); + } event = ftrace_find_event(entry->type); if (event && event->binary) @@ -1643,7 +1564,7 @@ static enum print_line_t print_trace_line(struct trace_iterator *iter) return print_raw_fmt(iter); if (iter->iter_flags & TRACE_FILE_LAT_FMT) - return print_lat_fmt(iter, iter->idx, iter->cpu); + return print_lat_fmt(iter); return print_trace_fmt(iter); } diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index e603a291134b..f0c7a0f08cac 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -405,6 +405,10 @@ void init_tracer_sysprof_debugfs(struct dentry *d_tracer); struct trace_entry *tracing_get_trace_entry(struct trace_array *tr, struct trace_array_cpu *data); + +struct trace_entry *trace_find_next_entry(struct trace_iterator *iter, + int *ent_cpu, u64 *ent_ts); + void tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, int pc); @@ -591,7 +595,8 @@ enum trace_iterator_flags { TRACE_ITER_ANNOTATE = 0x2000, TRACE_ITER_USERSTACKTRACE = 0x4000, TRACE_ITER_SYM_USEROBJ = 0x8000, - TRACE_ITER_PRINTK_MSGONLY = 0x10000 + TRACE_ITER_PRINTK_MSGONLY = 0x10000, + TRACE_ITER_CONTEXT_INFO = 0x20000 /* Print pid/cpu/time */ }; /* diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 1a4e144a9f8f..a5752d4d3c33 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -286,6 +286,113 @@ seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) return ret; } +static void +lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu) +{ + int hardirq, softirq; + char *comm; + + comm = trace_find_cmdline(entry->pid); + + trace_seq_printf(s, "%8.8s-%-5d ", comm, entry->pid); + trace_seq_printf(s, "%3d", cpu); + trace_seq_printf(s, "%c%c", + (entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : + (entry->flags & TRACE_FLAG_IRQS_NOSUPPORT) ? 'X' : '.', + ((entry->flags & TRACE_FLAG_NEED_RESCHED) ? 'N' : '.')); + + hardirq = entry->flags & TRACE_FLAG_HARDIRQ; + softirq = entry->flags & TRACE_FLAG_SOFTIRQ; + if (hardirq && softirq) { + trace_seq_putc(s, 'H'); + } else { + if (hardirq) { + trace_seq_putc(s, 'h'); + } else { + if (softirq) + trace_seq_putc(s, 's'); + else + trace_seq_putc(s, '.'); + } + } + + if (entry->preempt_count) + trace_seq_printf(s, "%x", entry->preempt_count); + else + trace_seq_puts(s, "."); +} + +static unsigned long preempt_mark_thresh = 100; + +static void +lat_print_timestamp(struct trace_seq *s, u64 abs_usecs, + unsigned long rel_usecs) +{ + trace_seq_printf(s, " %4lldus", abs_usecs); + if (rel_usecs > preempt_mark_thresh) + trace_seq_puts(s, "!: "); + else if (rel_usecs > 1) + trace_seq_puts(s, "+: "); + else + trace_seq_puts(s, " : "); +} + +int trace_print_context(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + char *comm = trace_find_cmdline(entry->pid); + unsigned long long t = ns2usecs(iter->ts); + unsigned long usec_rem = do_div(t, USEC_PER_SEC); + unsigned long secs = (unsigned long)t; + + if (!trace_seq_printf(s, "%16s-%-5d ", comm, entry->pid)) + goto partial; + if (!trace_seq_printf(s, "[%03d] ", entry->cpu)) + goto partial; + if (!trace_seq_printf(s, "%5lu.%06lu: ", secs, usec_rem)) + goto partial; + + return 0; + +partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +int trace_print_lat_context(struct trace_iterator *iter) +{ + u64 next_ts; + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent, + *next_entry = trace_find_next_entry(iter, NULL, + &next_ts); + unsigned long verbose = (trace_flags & TRACE_ITER_VERBOSE); + unsigned long abs_usecs = ns2usecs(iter->ts - iter->tr->time_start); + unsigned long rel_usecs; + + if (!next_entry) + next_ts = iter->ts; + rel_usecs = ns2usecs(next_ts - iter->ts); + + if (verbose) { + char *comm = trace_find_cmdline(entry->pid); + trace_seq_printf(s, "%16s %5d %3d %d %08x %08lx [%08lx]" + " %ld.%03ldms (+%ld.%03ldms): ", + comm, + entry->pid, entry->cpu, entry->flags, + entry->preempt_count, iter->idx, + ns2usecs(iter->ts), + abs_usecs/1000, + abs_usecs % 1000, rel_usecs/1000, + rel_usecs % 1000); + } else { + lat_print_generic(s, entry, entry->cpu); + lat_print_timestamp(s, abs_usecs, rel_usecs); + } + + return 0; +} + static const char state_to_char[] = TASK_STATE_TO_CHAR_STR; static int task_state_char(unsigned long state) diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 1cbab5e3dc99..ec2ed90f10f0 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -33,6 +33,9 @@ int seq_print_userip_objs(const struct userstack_entry *entry, int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, unsigned long ip, unsigned long sym_flags); +int trace_print_context(struct trace_iterator *iter); +int trace_print_lat_context(struct trace_iterator *iter); + struct trace_event *ftrace_find_event(int type); int register_ftrace_event(struct trace_event *event); int unregister_ftrace_event(struct trace_event *event); From 2c9b238eb325895d3312dad64e2685783575e474 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 2 Feb 2009 20:30:12 -0200 Subject: [PATCH 093/413] trace: Change struct trace_event callbacks parameter list Impact: API change The trace_seq and trace_entry are in trace_iterator, where there are more fields that may be needed by tracers, so just pass the tracer_iterator as is already the case for struct tracer->print_line. Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- block/blktrace.c | 8 +-- kernel/trace/trace.c | 10 +-- kernel/trace/trace_branch.c | 7 +- kernel/trace/trace_output.c | 139 ++++++++++++++++-------------------- kernel/trace/trace_output.h | 6 +- 5 files changed, 76 insertions(+), 94 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 3f25425ade12..570cd3c40bd1 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -1140,10 +1140,10 @@ static struct { [__BLK_TA_REMAP] = {{ "A", "remap" }, blk_log_remap }, }; -static int blk_trace_event_print(struct trace_seq *s, struct trace_entry *ent, - int flags) +static int blk_trace_event_print(struct trace_iterator *iter, int flags) { - const struct blk_io_trace *t = (struct blk_io_trace *)ent; + struct trace_seq *s = &iter->seq; + const struct blk_io_trace *t = (struct blk_io_trace *)iter->ent; const u16 what = t->action & ((1 << BLK_TC_SHIFT) - 1); int ret; @@ -1153,7 +1153,7 @@ static int blk_trace_event_print(struct trace_seq *s, struct trace_entry *ent, const bool long_act = !!(trace_flags & TRACE_ITER_VERBOSE); ret = blk_log_action_seq(s, t, what2act[what].act[long_act]); if (ret) - ret = what2act[what].print(s, ent); + ret = what2act[what].print(s, iter->ent); } return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 5ec49c3c1597..152d0969adf8 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1388,7 +1388,7 @@ static enum print_line_t print_lat_fmt(struct trace_iterator *iter) } if (event && event->latency_trace) { - ret = event->latency_trace(s, entry, sym_flags); + ret = event->latency_trace(iter, sym_flags); if (ret) return ret; return TRACE_TYPE_HANDLED; @@ -1419,7 +1419,7 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) } if (event && event->trace) { - ret = event->trace(s, entry, sym_flags); + ret = event->trace(iter, sym_flags); if (ret) return ret; return TRACE_TYPE_HANDLED; @@ -1449,7 +1449,7 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) event = ftrace_find_event(entry->type); if (event && event->raw) { - ret = event->raw(s, entry, 0); + ret = event->raw(iter, 0); if (ret) return ret; return TRACE_TYPE_HANDLED; @@ -1478,7 +1478,7 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) event = ftrace_find_event(entry->type); if (event && event->hex) - event->hex(s, entry, 0); + event->hex(iter, 0); SEQ_PUT_FIELD_RET(s, newline); @@ -1517,7 +1517,7 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) event = ftrace_find_event(entry->type); if (event && event->binary) - event->binary(s, entry, 0); + event->binary(iter, 0); return TRACE_TYPE_HANDLED; } diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 1284145c8898..ea62f101e615 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -160,14 +160,13 @@ trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int -trace_branch_print(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_branch_print(struct trace_iterator *iter, int flags) { struct trace_branch *field; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); - if (trace_seq_printf(s, "[%s] %s:%s:%d\n", + if (trace_seq_printf(&iter->seq, "[%s] %s:%s:%d\n", field->correct ? " ok " : " MISS ", field->func, field->file, diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index a5752d4d3c33..c24503b281a0 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -484,19 +484,18 @@ int unregister_ftrace_event(struct trace_event *event) * Standard events */ -int -trace_nop_print(struct trace_seq *s, struct trace_entry *entry, int flags) +int trace_nop_print(struct trace_iterator *iter, int flags) { return 0; } /* TRACE_FN */ -static int -trace_fn_latency(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_fn_latency(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); if (!seq_print_ip_sym(s, field->ip, flags)) goto partial; @@ -513,12 +512,12 @@ trace_fn_latency(struct trace_seq *s, struct trace_entry *entry, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int -trace_fn_trace(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_fn_trace(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); if (!seq_print_ip_sym(s, field->ip, flags)) goto partial; @@ -540,14 +539,13 @@ trace_fn_trace(struct trace_seq *s, struct trace_entry *entry, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int -trace_fn_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_fn_raw(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); - if (!trace_seq_printf(s, "%lx %lx\n", + if (!trace_seq_printf(&iter->seq, "%lx %lx\n", field->ip, field->parent_ip)) return TRACE_TYPE_PARTIAL_LINE; @@ -555,12 +553,12 @@ trace_fn_raw(struct trace_seq *s, struct trace_entry *entry, int flags) return 0; } -static int -trace_fn_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_fn_hex(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); SEQ_PUT_HEX_FIELD_RET(s, field->ip); SEQ_PUT_HEX_FIELD_RET(s, field->parent_ip); @@ -568,12 +566,12 @@ trace_fn_hex(struct trace_seq *s, struct trace_entry *entry, int flags) return 0; } -static int -trace_fn_bin(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_fn_bin(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); SEQ_PUT_FIELD_RET(s, field->ip); SEQ_PUT_FIELD_RET(s, field->parent_ip); @@ -591,20 +589,19 @@ static struct trace_event trace_fn_event = { }; /* TRACE_CTX an TRACE_WAKE */ -static int -trace_ctxwake_print(struct trace_seq *s, struct trace_entry *entry, int flags, - char *delim) +static int trace_ctxwake_print(struct trace_iterator *iter, char *delim) { struct ctx_switch_entry *field; char *comm; int S, T; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); T = task_state_char(field->next_state); S = task_state_char(field->prev_state); comm = trace_find_cmdline(field->next_pid); - if (!trace_seq_printf(s, " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", + if (!trace_seq_printf(&iter->seq, + " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", field->prev_pid, field->prev_prio, S, delim, @@ -617,31 +614,27 @@ trace_ctxwake_print(struct trace_seq *s, struct trace_entry *entry, int flags, return 0; } -static int -trace_ctx_print(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_ctx_print(struct trace_iterator *iter, int flags) { - return trace_ctxwake_print(s, entry, flags, "==>"); + return trace_ctxwake_print(iter, "==>"); } -static int -trace_wake_print(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_wake_print(struct trace_iterator *iter, int flags) { - return trace_ctxwake_print(s, entry, flags, " +"); + return trace_ctxwake_print(iter, " +"); } -static int -trace_ctxwake_raw(struct trace_seq *s, struct trace_entry *entry, int flags, - char S) +static int trace_ctxwake_raw(struct trace_iterator *iter, char S) { struct ctx_switch_entry *field; int T; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); if (!S) task_state_char(field->prev_state); T = task_state_char(field->next_state); - if (!trace_seq_printf(s, "%d %d %c %d %d %d %c\n", + if (!trace_seq_printf(&iter->seq, "%d %d %c %d %d %d %c\n", field->prev_pid, field->prev_prio, S, @@ -654,27 +647,24 @@ trace_ctxwake_raw(struct trace_seq *s, struct trace_entry *entry, int flags, return 0; } -static int -trace_ctx_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_ctx_raw(struct trace_iterator *iter, int flags) { - return trace_ctxwake_raw(s, entry, flags, 0); + return trace_ctxwake_raw(iter, 0); } -static int -trace_wake_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_wake_raw(struct trace_iterator *iter, int flags) { - return trace_ctxwake_raw(s, entry, flags, '+'); + return trace_ctxwake_raw(iter, '+'); } -static int -trace_ctxwake_hex(struct trace_seq *s, struct trace_entry *entry, int flags, - char S) +static int trace_ctxwake_hex(struct trace_iterator *iter, char S) { struct ctx_switch_entry *field; + struct trace_seq *s = &iter->seq; int T; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); if (!S) task_state_char(field->prev_state); @@ -691,24 +681,22 @@ trace_ctxwake_hex(struct trace_seq *s, struct trace_entry *entry, int flags, return 0; } -static int -trace_ctx_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_ctx_hex(struct trace_iterator *iter, int flags) { - return trace_ctxwake_hex(s, entry, flags, 0); + return trace_ctxwake_hex(iter, 0); } -static int -trace_wake_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_wake_hex(struct trace_iterator *iter, int flags) { - return trace_ctxwake_hex(s, entry, flags, '+'); + return trace_ctxwake_hex(iter, '+'); } -static int -trace_ctxwake_bin(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_ctxwake_bin(struct trace_iterator *iter, int flags) { struct ctx_switch_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); SEQ_PUT_FIELD_RET(s, field->prev_pid); SEQ_PUT_FIELD_RET(s, field->prev_prio); @@ -739,14 +727,13 @@ static struct trace_event trace_wake_event = { }; /* TRACE_SPECIAL */ -static int -trace_special_print(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_special_print(struct trace_iterator *iter, int flags) { struct special_entry *field; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); - if (!trace_seq_printf(s, "# %ld %ld %ld\n", + if (!trace_seq_printf(&iter->seq, "# %ld %ld %ld\n", field->arg1, field->arg2, field->arg3)) @@ -755,12 +742,12 @@ trace_special_print(struct trace_seq *s, struct trace_entry *entry, int flags) return 0; } -static int -trace_special_hex(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_special_hex(struct trace_iterator *iter, int flags) { struct special_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); SEQ_PUT_HEX_FIELD_RET(s, field->arg1); SEQ_PUT_HEX_FIELD_RET(s, field->arg2); @@ -769,12 +756,12 @@ trace_special_hex(struct trace_seq *s, struct trace_entry *entry, int flags) return 0; } -static int -trace_special_bin(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_special_bin(struct trace_iterator *iter, int flags) { struct special_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); SEQ_PUT_FIELD_RET(s, field->arg1); SEQ_PUT_FIELD_RET(s, field->arg2); @@ -794,13 +781,13 @@ static struct trace_event trace_special_event = { /* TRACE_STACK */ -static int -trace_stack_print(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_stack_print(struct trace_iterator *iter, int flags) { struct stack_entry *field; + struct trace_seq *s = &iter->seq; int i; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); for (i = 0; i < FTRACE_STACK_ENTRIES; i++) { if (i) { @@ -830,13 +817,12 @@ static struct trace_event trace_stack_event = { }; /* TRACE_USER_STACK */ -static int -trace_user_stack_print(struct trace_seq *s, struct trace_entry *entry, - int flags) +static int trace_user_stack_print(struct trace_iterator *iter, int flags) { struct userstack_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); if (!seq_print_userip_objs(field, s, flags)) goto partial; @@ -860,12 +846,12 @@ static struct trace_event trace_user_stack_event = { }; /* TRACE_PRINT */ -static int -trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_print_print(struct trace_iterator *iter, int flags) { struct print_entry *field; + struct trace_seq *s = &iter->seq; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); if (!seq_print_ip_sym(s, field->ip, flags)) goto partial; @@ -879,14 +865,13 @@ trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int -trace_print_raw(struct trace_seq *s, struct trace_entry *entry, int flags) +static int trace_print_raw(struct trace_iterator *iter, int flags) { struct print_entry *field; - trace_assign_type(field, entry); + trace_assign_type(field, iter->ent); - if (!trace_seq_printf(s, "# %lx %s", field->ip, field->buf)) + if (!trace_seq_printf(&iter->seq, "# %lx %s", field->ip, field->buf)) goto partial; return 0; diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index ec2ed90f10f0..3aeb31f6506b 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -3,8 +3,7 @@ #include "trace.h" -typedef int (*trace_print_func)(struct trace_seq *s, struct trace_entry *entry, - int flags); +typedef int (*trace_print_func)(struct trace_iterator *iter, int flags); struct trace_event { struct hlist_node node; @@ -40,8 +39,7 @@ struct trace_event *ftrace_find_event(int type); int register_ftrace_event(struct trace_event *event); int unregister_ftrace_event(struct trace_event *event); -int -trace_nop_print(struct trace_seq *s, struct trace_entry *entry, int flags); +int trace_nop_print(struct trace_iterator *iter, int flags); #define MAX_MEMHEX_BYTES 8 #define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) From 08a06b83ff8b2779289f733348c669f31cb65d51 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 2 Feb 2009 20:30:40 -0200 Subject: [PATCH 094/413] blkftrace: binary tracing, synthesizing old format Impact: new feature With this and a blkrawverify modified not to verify the sequence numbers we can start using the userspace tools to verify that the data produced with the ftrace plugin works as expected. Example: [root@f10-1 ~]# echo 1 > /sys/block/sda/sda1/trace/enable [root@f10-1 ~]# echo bin > /d/tracing/trace_options [root@f10-1 ~]# echo blk > /d/tracing/current_tracer [root@f10-1 ~]# cat /d/tracing/trace_pipe > sda1.blktrace.0 ^C [root@f10-1 ~]# ./blkrawverify --noseq sda1 Verifying sda1 CPU 0 Wrote output to sda1.verify.out [root@f10-1 ~]# cat sda1.verify.out --------------- Verifying sda1 --------------------- Summary for cpu 0: 1349 valid + 0 invalid (100.0%) processed [root@f10-1 ~]# Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- block/blktrace.c | 40 +++++++++++++++++++++++++++++++++++++--- 1 file changed, 37 insertions(+), 3 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 570cd3c40bd1..4f45b343690a 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -219,9 +219,16 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, t->magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION; t->sequence = ++(*sequence); t->time = ktime_to_ns(ktime_get()); - t->cpu = cpu; - t->pid = pid; record_it: + /* + * These two are not needed in ftrace as they are in the + * generic trace_entry, filled by tracing_generic_entry_update, + * but for the trace_event->bin() synthesizer benefit we do it + * here too. + */ + t->cpu = cpu; + t->pid = pid; + t->sector = sector; t->bytes = bytes; t->action = what; @@ -1086,6 +1093,7 @@ static void blk_tracer_start(struct trace_array *tr) if (blk_register_tracepoints()) atomic_dec(&blk_probes_ref); mutex_unlock(&blk_probe_mutex); + trace_flags &= ~TRACE_ITER_CONTEXT_INFO; } static int blk_tracer_init(struct trace_array *tr) @@ -1100,6 +1108,7 @@ static int blk_tracer_init(struct trace_array *tr) static void blk_tracer_stop(struct trace_array *tr) { + trace_flags |= TRACE_ITER_CONTEXT_INFO; mutex_lock(&blk_probe_mutex); if (atomic_dec_and_test(&blk_probes_ref)) blk_unregister_tracepoints(); @@ -1147,6 +1156,9 @@ static int blk_trace_event_print(struct trace_iterator *iter, int flags) const u16 what = t->action & ((1 << BLK_TC_SHIFT) - 1); int ret; + if (trace_print_context(iter)) + return TRACE_TYPE_PARTIAL_LINE; + if (unlikely(what == 0 || what > ARRAY_SIZE(what2act))) ret = trace_seq_printf(s, "Bad pc action %x\n", what); else { @@ -1159,6 +1171,28 @@ static int blk_trace_event_print(struct trace_iterator *iter, int flags) return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; } +static int blk_trace_synthesize_old_trace(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct blk_io_trace *t = (struct blk_io_trace *)iter->ent; + const int offset = offsetof(struct blk_io_trace, sector); + struct blk_io_trace old = { + .magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION, + .time = ns2usecs(iter->ts), + }; + + if (!trace_seq_putmem(s, &old, offset)) + return 0; + return trace_seq_putmem(s, &t->sector, + sizeof(old) - offset + t->pdu_len); +} + +static int blk_trace_event_print_binary(struct trace_iterator *iter, int flags) +{ + return blk_trace_synthesize_old_trace(iter) ? + TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; +} + static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter) { const struct blk_io_trace *t; @@ -1200,7 +1234,7 @@ static struct trace_event trace_blk_event = { .latency_trace = blk_trace_event_print, .raw = trace_nop_print, .hex = trace_nop_print, - .binary = trace_nop_print, + .binary = blk_trace_event_print_binary, }; static int __init init_blk_tracer(void) From 939b366977d29b5c0d53d1ea3b0b8cefb1e76202 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 3 Feb 2009 11:58:29 -0200 Subject: [PATCH 095/413] blktrace: fix coding style in recent patches Impact: cleanup Reported-by: Ingo Molnar Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- block/blktrace.c | 51 ++++++++++++++++++++++++++++-------------------- 1 file changed, 30 insertions(+), 21 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 4f45b343690a..8f5c37b0f80f 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -24,7 +24,7 @@ #include #include #include -#include +#include #include <../kernel/trace/trace_output.h> static unsigned int blktrace_seq __read_mostly = 1; @@ -148,11 +148,12 @@ static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector, /* * Data direction bit lookup */ -static u32 ddir_act[2] __read_mostly = { BLK_TC_ACT(BLK_TC_READ), BLK_TC_ACT(BLK_TC_WRITE) }; +static u32 ddir_act[2] __read_mostly = { BLK_TC_ACT(BLK_TC_READ), + BLK_TC_ACT(BLK_TC_WRITE) }; /* The ilog2() calls fall out because they're constant */ -#define MASK_TC_BIT(rw, __name) ( (rw & (1 << BIO_RW_ ## __name)) << \ - (ilog2(BLK_TC_ ## __name) + BLK_TC_SHIFT - BIO_RW_ ## __name) ) +#define MASK_TC_BIT(rw, __name) ((rw & (1 << BIO_RW_ ## __name)) << \ + (ilog2(BLK_TC_ ## __name) + BLK_TC_SHIFT - BIO_RW_ ## __name)) /* * The worker for the various blk_add_trace*() types. Fills out a @@ -221,13 +222,13 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, t->time = ktime_to_ns(ktime_get()); record_it: /* - * These two are not needed in ftrace as they are in the - * generic trace_entry, filled by tracing_generic_entry_update, - * but for the trace_event->bin() synthesizer benefit we do it - * here too. - */ - t->cpu = cpu; - t->pid = pid; + * These two are not needed in ftrace as they are in the + * generic trace_entry, filled by tracing_generic_entry_update, + * but for the trace_event->bin() synthesizer benefit we do it + * here too. + */ + t->cpu = cpu; + t->pid = pid; t->sector = sector; t->bytes = bytes; @@ -453,7 +454,8 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, atomic_set(&bt->dropped, 0); ret = -EIO; - bt->dropped_file = debugfs_create_file("dropped", 0444, dir, bt, &blk_dropped_fops); + bt->dropped_file = debugfs_create_file("dropped", 0444, dir, bt, + &blk_dropped_fops); if (!bt->dropped_file) goto err; @@ -535,10 +537,10 @@ EXPORT_SYMBOL_GPL(blk_trace_setup); int blk_trace_startstop(struct request_queue *q, int start) { - struct blk_trace *bt; int ret; + struct blk_trace *bt = q->blk_trace; - if ((bt = q->blk_trace) == NULL) + if (bt == NULL) return -EINVAL; /* @@ -674,12 +676,14 @@ static void blk_add_trace_rq_issue(struct request_queue *q, struct request *rq) blk_add_trace_rq(q, rq, BLK_TA_ISSUE); } -static void blk_add_trace_rq_requeue(struct request_queue *q, struct request *rq) +static void blk_add_trace_rq_requeue(struct request_queue *q, + struct request *rq) { blk_add_trace_rq(q, rq, BLK_TA_REQUEUE); } -static void blk_add_trace_rq_complete(struct request_queue *q, struct request *rq) +static void blk_add_trace_rq_complete(struct request_queue *q, + struct request *rq) { blk_add_trace_rq(q, rq, BLK_TA_COMPLETE); } @@ -716,12 +720,14 @@ static void blk_add_trace_bio_complete(struct request_queue *q, struct bio *bio) blk_add_trace_bio(q, bio, BLK_TA_COMPLETE); } -static void blk_add_trace_bio_backmerge(struct request_queue *q, struct bio *bio) +static void blk_add_trace_bio_backmerge(struct request_queue *q, + struct bio *bio) { blk_add_trace_bio(q, bio, BLK_TA_BACKMERGE); } -static void blk_add_trace_bio_frontmerge(struct request_queue *q, struct bio *bio) +static void blk_add_trace_bio_frontmerge(struct request_queue *q, + struct bio *bio) { blk_add_trace_bio(q, bio, BLK_TA_FRONTMERGE); } @@ -731,7 +737,8 @@ static void blk_add_trace_bio_queue(struct request_queue *q, struct bio *bio) blk_add_trace_bio(q, bio, BLK_TA_QUEUE); } -static void blk_add_trace_getrq(struct request_queue *q, struct bio *bio, int rw) +static void blk_add_trace_getrq(struct request_queue *q, + struct bio *bio, int rw) { if (bio) blk_add_trace_bio(q, bio, BLK_TA_GETRQ); @@ -744,7 +751,8 @@ static void blk_add_trace_getrq(struct request_queue *q, struct bio *bio, int rw } -static void blk_add_trace_sleeprq(struct request_queue *q, struct bio *bio, int rw) +static void blk_add_trace_sleeprq(struct request_queue *q, + struct bio *bio, int rw) { if (bio) blk_add_trace_bio(q, bio, BLK_TA_SLEEPRQ); @@ -752,7 +760,8 @@ static void blk_add_trace_sleeprq(struct request_queue *q, struct bio *bio, int struct blk_trace *bt = q->blk_trace; if (bt) - __blk_add_trace(bt, 0, 0, rw, BLK_TA_SLEEPRQ, 0, 0, NULL); + __blk_add_trace(bt, 0, 0, rw, BLK_TA_SLEEPRQ, + 0, 0, NULL); } } From d9793bd8018f835c64b10f44e278c86cecb8e932 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 3 Feb 2009 20:20:41 -0200 Subject: [PATCH 096/413] trace: judicious error checking of trace_seq results Impact: bugfix and cleanup Some callsites were returning either TRACE_ITER_PARTIAL_LINE if the trace_seq routines (trace_seq_printf, etc) returned 0 meaning its buffer was full, or zero otherwise. But... /* Return values for print_line callback */ enum print_line_t { TRACE_TYPE_PARTIAL_LINE = 0, /* Retry after flushing the seq */ TRACE_TYPE_HANDLED = 1, TRACE_TYPE_UNHANDLED = 2 /* Relay to other output functions */ }; In other cases the return value was not being relayed at all. Most of the time it didn't hurt because the page wasn't get filled, but for correctness sake, handle the return values everywhere. Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- block/blktrace.c | 2 +- kernel/trace/trace.c | 75 ++++++++++------------ kernel/trace/trace_branch.c | 2 +- kernel/trace/trace_output.c | 123 +++++++++++++++--------------------- 4 files changed, 87 insertions(+), 115 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 8f5c37b0f80f..12df27693972 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -1165,7 +1165,7 @@ static int blk_trace_event_print(struct trace_iterator *iter, int flags) const u16 what = t->action & ((1 << BLK_TC_SHIFT) - 1); int ret; - if (trace_print_context(iter)) + if (!trace_print_context(iter)) return TRACE_TYPE_PARTIAL_LINE; if (unlikely(what == 0 || what > ARRAY_SIZE(what2act))) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index bbdfaa2cbdb9..5822ff4e5a3e 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1402,27 +1402,25 @@ static enum print_line_t print_lat_fmt(struct trace_iterator *iter) unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); struct trace_event *event; struct trace_entry *entry = iter->ent; - int ret; test_cpu_buff_start(iter); event = ftrace_find_event(entry->type); if (trace_flags & TRACE_ITER_CONTEXT_INFO) { - ret = trace_print_lat_context(iter); - if (ret) - return ret; + if (!trace_print_lat_context(iter)) + goto partial; } - if (event && event->latency_trace) { - ret = event->latency_trace(iter, sym_flags); - if (ret) - return ret; - return TRACE_TYPE_HANDLED; - } + if (event && event->latency_trace) + return event->latency_trace(iter, sym_flags); + + if (!trace_seq_printf(s, "Unknown type %d\n", entry->type)) + goto partial; - trace_seq_printf(s, "Unknown type %d\n", entry->type); return TRACE_TYPE_HANDLED; +partial: + return TRACE_TYPE_PARTIAL_LINE; } static enum print_line_t print_trace_fmt(struct trace_iterator *iter) @@ -1431,7 +1429,6 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); struct trace_entry *entry; struct trace_event *event; - int ret; entry = iter->ent; @@ -1440,22 +1437,19 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) event = ftrace_find_event(entry->type); if (trace_flags & TRACE_ITER_CONTEXT_INFO) { - ret = trace_print_context(iter); - if (ret) - return ret; + if (!trace_print_context(iter)) + goto partial; } - if (event && event->trace) { - ret = event->trace(iter, sym_flags); - if (ret) - return ret; - return TRACE_TYPE_HANDLED; - } - ret = trace_seq_printf(s, "Unknown type %d\n", entry->type); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (event && event->trace) + return event->trace(iter, sym_flags); + + if (!trace_seq_printf(s, "Unknown type %d\n", entry->type)) + goto partial; return TRACE_TYPE_HANDLED; +partial: + return TRACE_TYPE_PARTIAL_LINE; } static enum print_line_t print_raw_fmt(struct trace_iterator *iter) @@ -1463,29 +1457,25 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) struct trace_seq *s = &iter->seq; struct trace_entry *entry; struct trace_event *event; - int ret; entry = iter->ent; if (trace_flags & TRACE_ITER_CONTEXT_INFO) { - ret = trace_seq_printf(s, "%d %d %llu ", - entry->pid, iter->cpu, iter->ts); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (!trace_seq_printf(s, "%d %d %llu ", + entry->pid, iter->cpu, iter->ts)) + goto partial; } event = ftrace_find_event(entry->type); - if (event && event->raw) { - ret = event->raw(iter, 0); - if (ret) - return ret; - return TRACE_TYPE_HANDLED; - } - ret = trace_seq_printf(s, "%d ?\n", entry->type); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + if (event && event->raw) + return event->raw(iter, 0); + + if (!trace_seq_printf(s, "%d ?\n", entry->type)) + goto partial; return TRACE_TYPE_HANDLED; +partial: + return TRACE_TYPE_PARTIAL_LINE; } static enum print_line_t print_hex_fmt(struct trace_iterator *iter) @@ -1504,8 +1494,11 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) } event = ftrace_find_event(entry->type); - if (event && event->hex) - event->hex(iter, 0); + if (event && event->hex) { + int ret = event->hex(iter, 0); + if (ret != TRACE_TYPE_HANDLED) + return ret; + } SEQ_PUT_FIELD_RET(s, newline); @@ -1544,7 +1537,7 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) event = ftrace_find_event(entry->type); if (event && event->binary) - event->binary(iter, 0); + return event->binary(iter, 0); return TRACE_TYPE_HANDLED; } diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index ea62f101e615..f6b35e162dfa 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -173,7 +173,7 @@ static int trace_branch_print(struct trace_iterator *iter, int flags) field->line)) return TRACE_TYPE_PARTIAL_LINE; - return 0; + return TRACE_TYPE_HANDLED; } diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index c24503b281a0..5b3c914053f2 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -286,55 +286,41 @@ seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags) return ret; } -static void +static int lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu) { int hardirq, softirq; char *comm; comm = trace_find_cmdline(entry->pid); - - trace_seq_printf(s, "%8.8s-%-5d ", comm, entry->pid); - trace_seq_printf(s, "%3d", cpu); - trace_seq_printf(s, "%c%c", - (entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : - (entry->flags & TRACE_FLAG_IRQS_NOSUPPORT) ? 'X' : '.', - ((entry->flags & TRACE_FLAG_NEED_RESCHED) ? 'N' : '.')); - hardirq = entry->flags & TRACE_FLAG_HARDIRQ; softirq = entry->flags & TRACE_FLAG_SOFTIRQ; - if (hardirq && softirq) { - trace_seq_putc(s, 'H'); - } else { - if (hardirq) { - trace_seq_putc(s, 'h'); - } else { - if (softirq) - trace_seq_putc(s, 's'); - else - trace_seq_putc(s, '.'); - } - } + + if (!trace_seq_printf(s, "%8.8s-%-5d %3d%c%c%c", + comm, entry->pid, cpu, + (entry->flags & TRACE_FLAG_IRQS_OFF) ? 'd' : + (entry->flags & TRACE_FLAG_IRQS_NOSUPPORT) ? + 'X' : '.', + (entry->flags & TRACE_FLAG_NEED_RESCHED) ? + 'N' : '.', + (hardirq && softirq) ? 'H' : + hardirq ? 'h' : softirq ? 's' : '.')) + return 0; if (entry->preempt_count) - trace_seq_printf(s, "%x", entry->preempt_count); - else - trace_seq_puts(s, "."); + return trace_seq_printf(s, "%x", entry->preempt_count); + return trace_seq_puts(s, "."); } static unsigned long preempt_mark_thresh = 100; -static void +static int lat_print_timestamp(struct trace_seq *s, u64 abs_usecs, unsigned long rel_usecs) { - trace_seq_printf(s, " %4lldus", abs_usecs); - if (rel_usecs > preempt_mark_thresh) - trace_seq_puts(s, "!: "); - else if (rel_usecs > 1) - trace_seq_puts(s, "+: "); - else - trace_seq_puts(s, " : "); + return trace_seq_printf(s, " %4lldus%c: ", abs_usecs, + rel_usecs > preempt_mark_thresh ? '!' : + rel_usecs > 1 ? '+' : ' '); } int trace_print_context(struct trace_iterator *iter) @@ -346,22 +332,14 @@ int trace_print_context(struct trace_iterator *iter) unsigned long usec_rem = do_div(t, USEC_PER_SEC); unsigned long secs = (unsigned long)t; - if (!trace_seq_printf(s, "%16s-%-5d ", comm, entry->pid)) - goto partial; - if (!trace_seq_printf(s, "[%03d] ", entry->cpu)) - goto partial; - if (!trace_seq_printf(s, "%5lu.%06lu: ", secs, usec_rem)) - goto partial; - - return 0; - -partial: - return TRACE_TYPE_PARTIAL_LINE; + return trace_seq_printf(s, "%16s-%-5d [%03d] %5lu.%06lu: ", + comm, entry->pid, entry->cpu, secs, usec_rem); } int trace_print_lat_context(struct trace_iterator *iter) { u64 next_ts; + int ret; struct trace_seq *s = &iter->seq; struct trace_entry *entry = iter->ent, *next_entry = trace_find_next_entry(iter, NULL, @@ -376,21 +354,22 @@ int trace_print_lat_context(struct trace_iterator *iter) if (verbose) { char *comm = trace_find_cmdline(entry->pid); - trace_seq_printf(s, "%16s %5d %3d %d %08x %08lx [%08lx]" - " %ld.%03ldms (+%ld.%03ldms): ", - comm, - entry->pid, entry->cpu, entry->flags, - entry->preempt_count, iter->idx, - ns2usecs(iter->ts), - abs_usecs/1000, - abs_usecs % 1000, rel_usecs/1000, - rel_usecs % 1000); + ret = trace_seq_printf(s, "%16s %5d %3d %d %08x %08lx [%08lx]" + " %ld.%03ldms (+%ld.%03ldms): ", comm, + entry->pid, entry->cpu, entry->flags, + entry->preempt_count, iter->idx, + ns2usecs(iter->ts), + abs_usecs / USEC_PER_MSEC, + abs_usecs % USEC_PER_MSEC, + rel_usecs / USEC_PER_MSEC, + rel_usecs % USEC_PER_MSEC); } else { - lat_print_generic(s, entry, entry->cpu); - lat_print_timestamp(s, abs_usecs, rel_usecs); + ret = lat_print_generic(s, entry, entry->cpu); + if (ret) + ret = lat_print_timestamp(s, abs_usecs, rel_usecs); } - return 0; + return ret; } static const char state_to_char[] = TASK_STATE_TO_CHAR_STR; @@ -486,7 +465,7 @@ int unregister_ftrace_event(struct trace_event *event) int trace_nop_print(struct trace_iterator *iter, int flags) { - return 0; + return TRACE_TYPE_HANDLED; } /* TRACE_FN */ @@ -506,7 +485,7 @@ static int trace_fn_latency(struct trace_iterator *iter, int flags) if (!trace_seq_puts(s, ")\n")) goto partial; - return 0; + return TRACE_TYPE_HANDLED; partial: return TRACE_TYPE_PARTIAL_LINE; @@ -533,7 +512,7 @@ static int trace_fn_trace(struct trace_iterator *iter, int flags) if (!trace_seq_printf(s, "\n")) goto partial; - return 0; + return TRACE_TYPE_HANDLED; partial: return TRACE_TYPE_PARTIAL_LINE; @@ -550,7 +529,7 @@ static int trace_fn_raw(struct trace_iterator *iter, int flags) field->parent_ip)) return TRACE_TYPE_PARTIAL_LINE; - return 0; + return TRACE_TYPE_HANDLED; } static int trace_fn_hex(struct trace_iterator *iter, int flags) @@ -563,7 +542,7 @@ static int trace_fn_hex(struct trace_iterator *iter, int flags) SEQ_PUT_HEX_FIELD_RET(s, field->ip); SEQ_PUT_HEX_FIELD_RET(s, field->parent_ip); - return 0; + return TRACE_TYPE_HANDLED; } static int trace_fn_bin(struct trace_iterator *iter, int flags) @@ -576,7 +555,7 @@ static int trace_fn_bin(struct trace_iterator *iter, int flags) SEQ_PUT_FIELD_RET(s, field->ip); SEQ_PUT_FIELD_RET(s, field->parent_ip); - return 0; + return TRACE_TYPE_HANDLED; } static struct trace_event trace_fn_event = { @@ -611,7 +590,7 @@ static int trace_ctxwake_print(struct trace_iterator *iter, char *delim) T, comm)) return TRACE_TYPE_PARTIAL_LINE; - return 0; + return TRACE_TYPE_HANDLED; } static int trace_ctx_print(struct trace_iterator *iter, int flags) @@ -644,7 +623,7 @@ static int trace_ctxwake_raw(struct trace_iterator *iter, char S) T)) return TRACE_TYPE_PARTIAL_LINE; - return 0; + return TRACE_TYPE_HANDLED; } static int trace_ctx_raw(struct trace_iterator *iter, int flags) @@ -678,7 +657,7 @@ static int trace_ctxwake_hex(struct trace_iterator *iter, char S) SEQ_PUT_HEX_FIELD_RET(s, field->next_prio); SEQ_PUT_HEX_FIELD_RET(s, T); - return 0; + return TRACE_TYPE_HANDLED; } static int trace_ctx_hex(struct trace_iterator *iter, int flags) @@ -705,7 +684,7 @@ static int trace_ctxwake_bin(struct trace_iterator *iter, int flags) SEQ_PUT_FIELD_RET(s, field->next_prio); SEQ_PUT_FIELD_RET(s, field->next_state); - return 0; + return TRACE_TYPE_HANDLED; } static struct trace_event trace_ctx_event = { @@ -739,7 +718,7 @@ static int trace_special_print(struct trace_iterator *iter, int flags) field->arg3)) return TRACE_TYPE_PARTIAL_LINE; - return 0; + return TRACE_TYPE_HANDLED; } static int trace_special_hex(struct trace_iterator *iter, int flags) @@ -753,7 +732,7 @@ static int trace_special_hex(struct trace_iterator *iter, int flags) SEQ_PUT_HEX_FIELD_RET(s, field->arg2); SEQ_PUT_HEX_FIELD_RET(s, field->arg3); - return 0; + return TRACE_TYPE_HANDLED; } static int trace_special_bin(struct trace_iterator *iter, int flags) @@ -767,7 +746,7 @@ static int trace_special_bin(struct trace_iterator *iter, int flags) SEQ_PUT_FIELD_RET(s, field->arg2); SEQ_PUT_FIELD_RET(s, field->arg3); - return 0; + return TRACE_TYPE_HANDLED; } static struct trace_event trace_special_event = { @@ -801,7 +780,7 @@ static int trace_stack_print(struct trace_iterator *iter, int flags) goto partial; } - return 0; + return TRACE_TYPE_HANDLED; partial: return TRACE_TYPE_PARTIAL_LINE; @@ -830,7 +809,7 @@ static int trace_user_stack_print(struct trace_iterator *iter, int flags) if (!trace_seq_putc(s, '\n')) goto partial; - return 0; + return TRACE_TYPE_HANDLED; partial: return TRACE_TYPE_PARTIAL_LINE; @@ -859,7 +838,7 @@ static int trace_print_print(struct trace_iterator *iter, int flags) if (!trace_seq_printf(s, ": %s", field->buf)) goto partial; - return 0; + return TRACE_TYPE_HANDLED; partial: return TRACE_TYPE_PARTIAL_LINE; @@ -874,7 +853,7 @@ static int trace_print_raw(struct trace_iterator *iter, int flags) if (!trace_seq_printf(&iter->seq, "# %lx %s", field->ip, field->buf)) goto partial; - return 0; + return TRACE_TYPE_HANDLED; partial: return TRACE_TYPE_PARTIAL_LINE; From ae7462b4f1fe1f36b5d562dbd5202a2eba01f072 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Tue, 3 Feb 2009 22:05:50 -0200 Subject: [PATCH 097/413] trace: make the trace_event callbacks return enum print_line_t As they actually all return these enumerators. Reported-by: Frederic Weisbecker Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- block/blktrace.c | 6 +++-- kernel/trace/trace.c | 2 +- kernel/trace/trace_branch.c | 3 ++- kernel/trace/trace_output.c | 52 ++++++++++++++++++++++--------------- kernel/trace/trace_output.h | 5 ++-- 5 files changed, 41 insertions(+), 27 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 12df27693972..c7698d1617a1 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -1158,7 +1158,8 @@ static struct { [__BLK_TA_REMAP] = {{ "A", "remap" }, blk_log_remap }, }; -static int blk_trace_event_print(struct trace_iterator *iter, int flags) +static enum print_line_t blk_trace_event_print(struct trace_iterator *iter, + int flags) { struct trace_seq *s = &iter->seq; const struct blk_io_trace *t = (struct blk_io_trace *)iter->ent; @@ -1196,7 +1197,8 @@ static int blk_trace_synthesize_old_trace(struct trace_iterator *iter) sizeof(old) - offset + t->pdu_len); } -static int blk_trace_event_print_binary(struct trace_iterator *iter, int flags) +static enum print_line_t +blk_trace_event_print_binary(struct trace_iterator *iter, int flags) { return blk_trace_synthesize_old_trace(iter) ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 5822ff4e5a3e..fd51cf0b94c7 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1495,7 +1495,7 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) event = ftrace_find_event(entry->type); if (event && event->hex) { - int ret = event->hex(iter, 0); + enum print_line_t ret = event->hex(iter, 0); if (ret != TRACE_TYPE_HANDLED) return ret; } diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index f6b35e162dfa..7ac72a44b2d3 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -160,7 +160,8 @@ trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int trace_branch_print(struct trace_iterator *iter, int flags) +static enum print_line_t trace_branch_print(struct trace_iterator *iter, + int flags) { struct trace_branch *field; diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 5b3c914053f2..b7380eee9fa1 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -463,13 +463,14 @@ int unregister_ftrace_event(struct trace_event *event) * Standard events */ -int trace_nop_print(struct trace_iterator *iter, int flags) +enum print_line_t trace_nop_print(struct trace_iterator *iter, int flags) { return TRACE_TYPE_HANDLED; } /* TRACE_FN */ -static int trace_fn_latency(struct trace_iterator *iter, int flags) +static enum print_line_t trace_fn_latency(struct trace_iterator *iter, + int flags) { struct ftrace_entry *field; struct trace_seq *s = &iter->seq; @@ -491,7 +492,7 @@ static int trace_fn_latency(struct trace_iterator *iter, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int trace_fn_trace(struct trace_iterator *iter, int flags) +static enum print_line_t trace_fn_trace(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; struct trace_seq *s = &iter->seq; @@ -518,7 +519,7 @@ static int trace_fn_trace(struct trace_iterator *iter, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int trace_fn_raw(struct trace_iterator *iter, int flags) +static enum print_line_t trace_fn_raw(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; @@ -532,7 +533,7 @@ static int trace_fn_raw(struct trace_iterator *iter, int flags) return TRACE_TYPE_HANDLED; } -static int trace_fn_hex(struct trace_iterator *iter, int flags) +static enum print_line_t trace_fn_hex(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; struct trace_seq *s = &iter->seq; @@ -545,7 +546,7 @@ static int trace_fn_hex(struct trace_iterator *iter, int flags) return TRACE_TYPE_HANDLED; } -static int trace_fn_bin(struct trace_iterator *iter, int flags) +static enum print_line_t trace_fn_bin(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; struct trace_seq *s = &iter->seq; @@ -568,7 +569,8 @@ static struct trace_event trace_fn_event = { }; /* TRACE_CTX an TRACE_WAKE */ -static int trace_ctxwake_print(struct trace_iterator *iter, char *delim) +static enum print_line_t trace_ctxwake_print(struct trace_iterator *iter, + char *delim) { struct ctx_switch_entry *field; char *comm; @@ -593,12 +595,13 @@ static int trace_ctxwake_print(struct trace_iterator *iter, char *delim) return TRACE_TYPE_HANDLED; } -static int trace_ctx_print(struct trace_iterator *iter, int flags) +static enum print_line_t trace_ctx_print(struct trace_iterator *iter, int flags) { return trace_ctxwake_print(iter, "==>"); } -static int trace_wake_print(struct trace_iterator *iter, int flags) +static enum print_line_t trace_wake_print(struct trace_iterator *iter, + int flags) { return trace_ctxwake_print(iter, " +"); } @@ -626,12 +629,12 @@ static int trace_ctxwake_raw(struct trace_iterator *iter, char S) return TRACE_TYPE_HANDLED; } -static int trace_ctx_raw(struct trace_iterator *iter, int flags) +static enum print_line_t trace_ctx_raw(struct trace_iterator *iter, int flags) { return trace_ctxwake_raw(iter, 0); } -static int trace_wake_raw(struct trace_iterator *iter, int flags) +static enum print_line_t trace_wake_raw(struct trace_iterator *iter, int flags) { return trace_ctxwake_raw(iter, '+'); } @@ -660,17 +663,18 @@ static int trace_ctxwake_hex(struct trace_iterator *iter, char S) return TRACE_TYPE_HANDLED; } -static int trace_ctx_hex(struct trace_iterator *iter, int flags) +static enum print_line_t trace_ctx_hex(struct trace_iterator *iter, int flags) { return trace_ctxwake_hex(iter, 0); } -static int trace_wake_hex(struct trace_iterator *iter, int flags) +static enum print_line_t trace_wake_hex(struct trace_iterator *iter, int flags) { return trace_ctxwake_hex(iter, '+'); } -static int trace_ctxwake_bin(struct trace_iterator *iter, int flags) +static enum print_line_t trace_ctxwake_bin(struct trace_iterator *iter, + int flags) { struct ctx_switch_entry *field; struct trace_seq *s = &iter->seq; @@ -706,7 +710,8 @@ static struct trace_event trace_wake_event = { }; /* TRACE_SPECIAL */ -static int trace_special_print(struct trace_iterator *iter, int flags) +static enum print_line_t trace_special_print(struct trace_iterator *iter, + int flags) { struct special_entry *field; @@ -721,7 +726,8 @@ static int trace_special_print(struct trace_iterator *iter, int flags) return TRACE_TYPE_HANDLED; } -static int trace_special_hex(struct trace_iterator *iter, int flags) +static enum print_line_t trace_special_hex(struct trace_iterator *iter, + int flags) { struct special_entry *field; struct trace_seq *s = &iter->seq; @@ -735,7 +741,8 @@ static int trace_special_hex(struct trace_iterator *iter, int flags) return TRACE_TYPE_HANDLED; } -static int trace_special_bin(struct trace_iterator *iter, int flags) +static enum print_line_t trace_special_bin(struct trace_iterator *iter, + int flags) { struct special_entry *field; struct trace_seq *s = &iter->seq; @@ -760,7 +767,8 @@ static struct trace_event trace_special_event = { /* TRACE_STACK */ -static int trace_stack_print(struct trace_iterator *iter, int flags) +static enum print_line_t trace_stack_print(struct trace_iterator *iter, + int flags) { struct stack_entry *field; struct trace_seq *s = &iter->seq; @@ -796,7 +804,8 @@ static struct trace_event trace_stack_event = { }; /* TRACE_USER_STACK */ -static int trace_user_stack_print(struct trace_iterator *iter, int flags) +static enum print_line_t trace_user_stack_print(struct trace_iterator *iter, + int flags) { struct userstack_entry *field; struct trace_seq *s = &iter->seq; @@ -825,7 +834,8 @@ static struct trace_event trace_user_stack_event = { }; /* TRACE_PRINT */ -static int trace_print_print(struct trace_iterator *iter, int flags) +static enum print_line_t trace_print_print(struct trace_iterator *iter, + int flags) { struct print_entry *field; struct trace_seq *s = &iter->seq; @@ -844,7 +854,7 @@ static int trace_print_print(struct trace_iterator *iter, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static int trace_print_raw(struct trace_iterator *iter, int flags) +static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) { struct print_entry *field; diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 3aeb31f6506b..551a25a72217 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -3,7 +3,8 @@ #include "trace.h" -typedef int (*trace_print_func)(struct trace_iterator *iter, int flags); +typedef enum print_line_t (*trace_print_func)(struct trace_iterator *iter, + int flags); struct trace_event { struct hlist_node node; @@ -39,7 +40,7 @@ struct trace_event *ftrace_find_event(int type); int register_ftrace_event(struct trace_event *event); int unregister_ftrace_event(struct trace_event *event); -int trace_nop_print(struct trace_iterator *iter, int flags); +enum print_line_t trace_nop_print(struct trace_iterator *iter, int flags); #define MAX_MEMHEX_BYTES 8 #define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) From 268ccda0cb4d1292029d07ee3dbd07117baf6ecb Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 4 Feb 2009 20:16:39 -0200 Subject: [PATCH 098/413] trace: assign defaults at register_ftrace_event Impact: simplification of tracers As all tracers are doing this we might as well do it in register_ftrace_event and save one branch each time we call these callbacks. Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Frederic Weisbecker Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- block/blktrace.c | 2 -- kernel/trace/trace.c | 13 +++++-------- kernel/trace/trace_branch.c | 3 --- kernel/trace/trace_output.c | 13 +++++++++++-- 4 files changed, 16 insertions(+), 15 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index c7698d1617a1..1ebd068061ec 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -1243,8 +1243,6 @@ static struct trace_event trace_blk_event = { .type = TRACE_BLK, .trace = blk_trace_event_print, .latency_trace = blk_trace_event_print, - .raw = trace_nop_print, - .hex = trace_nop_print, .binary = blk_trace_event_print_binary, }; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index fd51cf0b94c7..a5e4c0af9bb0 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1412,7 +1412,7 @@ static enum print_line_t print_lat_fmt(struct trace_iterator *iter) goto partial; } - if (event && event->latency_trace) + if (event) return event->latency_trace(iter, sym_flags); if (!trace_seq_printf(s, "Unknown type %d\n", entry->type)) @@ -1441,7 +1441,7 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) goto partial; } - if (event && event->trace) + if (event) return event->trace(iter, sym_flags); if (!trace_seq_printf(s, "Unknown type %d\n", entry->type)) @@ -1467,7 +1467,7 @@ static enum print_line_t print_raw_fmt(struct trace_iterator *iter) } event = ftrace_find_event(entry->type); - if (event && event->raw) + if (event) return event->raw(iter, 0); if (!trace_seq_printf(s, "%d ?\n", entry->type)) @@ -1494,7 +1494,7 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) } event = ftrace_find_event(entry->type); - if (event && event->hex) { + if (event) { enum print_line_t ret = event->hex(iter, 0); if (ret != TRACE_TYPE_HANDLED) return ret; @@ -1536,10 +1536,7 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) } event = ftrace_find_event(entry->type); - if (event && event->binary) - return event->binary(iter, 0); - - return TRACE_TYPE_HANDLED; + return event ? event->binary(iter, 0) : TRACE_TYPE_HANDLED; } static int trace_empty(struct trace_iterator *iter) diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 7ac72a44b2d3..297deb202b68 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -182,9 +182,6 @@ static struct trace_event trace_branch_event = { .type = TRACE_BRANCH, .trace = trace_branch_print, .latency_trace = trace_branch_print, - .raw = trace_nop_print, - .hex = trace_nop_print, - .binary = trace_nop_print, }; static struct tracer branch_trace __read_mostly = diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index b7380eee9fa1..b6e99af79214 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -435,6 +435,17 @@ int register_ftrace_event(struct trace_event *event) if (ftrace_find_event(event->type)) goto out; + if (event->trace == NULL) + event->trace = trace_nop_print; + if (event->latency_trace == NULL) + event->latency_trace = trace_nop_print; + if (event->raw == NULL) + event->raw = trace_nop_print; + if (event->hex == NULL) + event->hex = trace_nop_print; + if (event->binary == NULL) + event->binary = trace_nop_print; + key = event->type & (EVENT_HASHSIZE - 1); hlist_add_head_rcu(&event->node, &event_hash[key]); @@ -874,8 +885,6 @@ static struct trace_event trace_print_event = { .trace = trace_print_print, .latency_trace = trace_print_print, .raw = trace_print_raw, - .hex = trace_nop_print, - .binary = trace_nop_print, }; static struct trace_event *events[] __initdata = { From 97e5b191ae7dc0f4f5b82b9db29782928b103b4d Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 5 Feb 2009 01:13:36 -0500 Subject: [PATCH 099/413] trace_branch: Remove unused function Impact: cleanup Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace_branch.c | 17 ----------------- 1 file changed, 17 deletions(-) diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 297deb202b68..027e83690615 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -143,23 +143,6 @@ static void branch_trace_reset(struct trace_array *tr) stop_branch_trace(tr); } -static int -trace_print_print(struct trace_seq *s, struct trace_entry *entry, int flags) -{ - struct print_entry *field; - - trace_assign_type(field, entry); - - if (seq_print_ip_sym(s, field->ip, flags)) - goto partial; - - if (trace_seq_printf(s, ": %s", field->buf)) - goto partial; - - partial: - return TRACE_TYPE_PARTIAL_LINE; -} - static enum print_line_t trace_branch_print(struct trace_iterator *iter, int flags) { From 7be421510b91491d5aa5a29fa1005712039b95af Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 5 Feb 2009 01:13:37 -0500 Subject: [PATCH 100/413] trace: Remove unused trace_array_cpu parameter Impact: cleanup Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- block/blktrace.c | 2 +- kernel/trace/trace.c | 47 ++++++++++++------------------- kernel/trace/trace.h | 4 --- kernel/trace/trace_functions.c | 8 +++--- kernel/trace/trace_irqsoff.c | 10 +++---- kernel/trace/trace_sched_switch.c | 4 +-- kernel/trace/trace_sched_wakeup.c | 12 ++++---- 7 files changed, 35 insertions(+), 52 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 1ebd068061ec..d9d7146ee023 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -245,7 +245,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, if (pid != 0 && !(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC) && (trace_flags & TRACE_ITER_STACKTRACE) != 0) - __trace_stack(blk_tr, NULL, flags, 5, pc); + __trace_stack(blk_tr, flags, 5, pc); trace_wake_up(); return; } diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index a5e4c0af9bb0..1d4ff568cc4d 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -776,7 +776,7 @@ tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, } void -trace_function(struct trace_array *tr, struct trace_array_cpu *data, +trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc) { @@ -802,7 +802,6 @@ trace_function(struct trace_array *tr, struct trace_array_cpu *data, #ifdef CONFIG_FUNCTION_GRAPH_TRACER static void __trace_graph_entry(struct trace_array *tr, - struct trace_array_cpu *data, struct ftrace_graph_ent *trace, unsigned long flags, int pc) @@ -826,7 +825,6 @@ static void __trace_graph_entry(struct trace_array *tr, } static void __trace_graph_return(struct trace_array *tr, - struct trace_array_cpu *data, struct ftrace_graph_ret *trace, unsigned long flags, int pc) @@ -856,11 +854,10 @@ ftrace(struct trace_array *tr, struct trace_array_cpu *data, int pc) { if (likely(!atomic_read(&data->disabled))) - trace_function(tr, data, ip, parent_ip, flags, pc); + trace_function(tr, ip, parent_ip, flags, pc); } static void __ftrace_trace_stack(struct trace_array *tr, - struct trace_array_cpu *data, unsigned long flags, int skip, int pc) { @@ -891,27 +888,24 @@ static void __ftrace_trace_stack(struct trace_array *tr, } static void ftrace_trace_stack(struct trace_array *tr, - struct trace_array_cpu *data, unsigned long flags, int skip, int pc) { if (!(trace_flags & TRACE_ITER_STACKTRACE)) return; - __ftrace_trace_stack(tr, data, flags, skip, pc); + __ftrace_trace_stack(tr, flags, skip, pc); } void __trace_stack(struct trace_array *tr, - struct trace_array_cpu *data, unsigned long flags, int skip, int pc) { - __ftrace_trace_stack(tr, data, flags, skip, pc); + __ftrace_trace_stack(tr, flags, skip, pc); } static void ftrace_trace_userstack(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long flags, int pc) + unsigned long flags, int pc) { #ifdef CONFIG_STACKTRACE struct ring_buffer_event *event; @@ -942,20 +936,17 @@ static void ftrace_trace_userstack(struct trace_array *tr, #endif } -void __trace_userstack(struct trace_array *tr, - struct trace_array_cpu *data, - unsigned long flags) +void __trace_userstack(struct trace_array *tr, unsigned long flags) { - ftrace_trace_userstack(tr, data, flags, preempt_count()); + ftrace_trace_userstack(tr, flags, preempt_count()); } static void -ftrace_trace_special(void *__tr, void *__data, +ftrace_trace_special(void *__tr, unsigned long arg1, unsigned long arg2, unsigned long arg3, int pc) { struct ring_buffer_event *event; - struct trace_array_cpu *data = __data; struct trace_array *tr = __tr; struct special_entry *entry; unsigned long irq_flags; @@ -971,8 +962,8 @@ ftrace_trace_special(void *__tr, void *__data, entry->arg2 = arg2; entry->arg3 = arg3; ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - ftrace_trace_stack(tr, data, irq_flags, 4, pc); - ftrace_trace_userstack(tr, data, irq_flags, pc); + ftrace_trace_stack(tr, irq_flags, 4, pc); + ftrace_trace_userstack(tr, irq_flags, pc); trace_wake_up(); } @@ -981,12 +972,11 @@ void __trace_special(void *__tr, void *__data, unsigned long arg1, unsigned long arg2, unsigned long arg3) { - ftrace_trace_special(__tr, __data, arg1, arg2, arg3, preempt_count()); + ftrace_trace_special(__tr, arg1, arg2, arg3, preempt_count()); } void tracing_sched_switch_trace(struct trace_array *tr, - struct trace_array_cpu *data, struct task_struct *prev, struct task_struct *next, unsigned long flags, int pc) @@ -1010,13 +1000,12 @@ tracing_sched_switch_trace(struct trace_array *tr, entry->next_state = next->state; entry->next_cpu = task_cpu(next); ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - ftrace_trace_stack(tr, data, flags, 5, pc); - ftrace_trace_userstack(tr, data, flags, pc); + ftrace_trace_stack(tr, flags, 5, pc); + ftrace_trace_userstack(tr, flags, pc); } void tracing_sched_wakeup_trace(struct trace_array *tr, - struct trace_array_cpu *data, struct task_struct *wakee, struct task_struct *curr, unsigned long flags, int pc) @@ -1040,8 +1029,8 @@ tracing_sched_wakeup_trace(struct trace_array *tr, entry->next_state = wakee->state; entry->next_cpu = task_cpu(wakee); ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - ftrace_trace_stack(tr, data, flags, 6, pc); - ftrace_trace_userstack(tr, data, flags, pc); + ftrace_trace_stack(tr, flags, 6, pc); + ftrace_trace_userstack(tr, flags, pc); trace_wake_up(); } @@ -1064,7 +1053,7 @@ ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3) data = tr->data[cpu]; if (likely(atomic_inc_return(&data->disabled) == 1)) - ftrace_trace_special(tr, data, arg1, arg2, arg3, pc); + ftrace_trace_special(tr, arg1, arg2, arg3, pc); atomic_dec(&data->disabled); local_irq_restore(flags); @@ -1092,7 +1081,7 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) { pc = preempt_count(); - __trace_graph_entry(tr, data, trace, flags, pc); + __trace_graph_entry(tr, trace, flags, pc); } /* Only do the atomic if it is not already set */ if (!test_tsk_trace_graph(current)) @@ -1118,7 +1107,7 @@ void trace_graph_return(struct ftrace_graph_ret *trace) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) { pc = preempt_count(); - __trace_graph_return(tr, data, trace, flags, pc); + __trace_graph_return(tr, trace, flags, pc); } if (!trace->depth) clear_tsk_trace_graph(current); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index f0c7a0f08cac..df627a948694 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -419,14 +419,12 @@ void ftrace(struct trace_array *tr, unsigned long parent_ip, unsigned long flags, int pc); void tracing_sched_switch_trace(struct trace_array *tr, - struct trace_array_cpu *data, struct task_struct *prev, struct task_struct *next, unsigned long flags, int pc); void tracing_record_cmdline(struct task_struct *tsk); void tracing_sched_wakeup_trace(struct trace_array *tr, - struct trace_array_cpu *data, struct task_struct *wakee, struct task_struct *cur, unsigned long flags, int pc); @@ -436,7 +434,6 @@ void trace_special(struct trace_array *tr, unsigned long arg2, unsigned long arg3, int pc); void trace_function(struct trace_array *tr, - struct trace_array_cpu *data, unsigned long ip, unsigned long parent_ip, unsigned long flags, int pc); @@ -462,7 +459,6 @@ void update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu); void __trace_stack(struct trace_array *tr, - struct trace_array_cpu *data, unsigned long flags, int skip, int pc); diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index b3a320f8aba7..d067cea2ccc3 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -78,7 +78,7 @@ function_trace_call_preempt_only(unsigned long ip, unsigned long parent_ip) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) - trace_function(tr, data, ip, parent_ip, flags, pc); + trace_function(tr, ip, parent_ip, flags, pc); atomic_dec(&data->disabled); ftrace_preempt_enable(resched); @@ -108,7 +108,7 @@ function_trace_call(unsigned long ip, unsigned long parent_ip) if (likely(disabled == 1)) { pc = preempt_count(); - trace_function(tr, data, ip, parent_ip, flags, pc); + trace_function(tr, ip, parent_ip, flags, pc); } atomic_dec(&data->disabled); @@ -139,7 +139,7 @@ function_stack_trace_call(unsigned long ip, unsigned long parent_ip) if (likely(disabled == 1)) { pc = preempt_count(); - trace_function(tr, data, ip, parent_ip, flags, pc); + trace_function(tr, ip, parent_ip, flags, pc); /* * skip over 5 funcs: * __ftrace_trace_stack, @@ -148,7 +148,7 @@ function_stack_trace_call(unsigned long ip, unsigned long parent_ip) * ftrace_list_func * ftrace_call */ - __trace_stack(tr, data, flags, 5, pc); + __trace_stack(tr, flags, 5, pc); } atomic_dec(&data->disabled); diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index ed344b022a14..c6b442d88de8 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -95,7 +95,7 @@ irqsoff_tracer_call(unsigned long ip, unsigned long parent_ip) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) - trace_function(tr, data, ip, parent_ip, flags, preempt_count()); + trace_function(tr, ip, parent_ip, flags, preempt_count()); atomic_dec(&data->disabled); } @@ -153,7 +153,7 @@ check_critical_timing(struct trace_array *tr, if (!report_latency(delta)) goto out_unlock; - trace_function(tr, data, CALLER_ADDR0, parent_ip, flags, pc); + trace_function(tr, CALLER_ADDR0, parent_ip, flags, pc); latency = nsecs_to_usecs(delta); @@ -177,7 +177,7 @@ check_critical_timing(struct trace_array *tr, data->critical_sequence = max_sequence; data->preempt_timestamp = ftrace_now(cpu); tracing_reset(tr, cpu); - trace_function(tr, data, CALLER_ADDR0, parent_ip, flags, pc); + trace_function(tr, CALLER_ADDR0, parent_ip, flags, pc); } static inline void @@ -210,7 +210,7 @@ start_critical_timing(unsigned long ip, unsigned long parent_ip) local_save_flags(flags); - trace_function(tr, data, ip, parent_ip, flags, preempt_count()); + trace_function(tr, ip, parent_ip, flags, preempt_count()); per_cpu(tracing_cpu, cpu) = 1; @@ -244,7 +244,7 @@ stop_critical_timing(unsigned long ip, unsigned long parent_ip) atomic_inc(&data->disabled); local_save_flags(flags); - trace_function(tr, data, ip, parent_ip, flags, preempt_count()); + trace_function(tr, ip, parent_ip, flags, preempt_count()); check_critical_timing(tr, data, parent_ip ? : ip, cpu); data->critical_start = 0; atomic_dec(&data->disabled); diff --git a/kernel/trace/trace_sched_switch.c b/kernel/trace/trace_sched_switch.c index df175cb4564f..c4f9add5ec90 100644 --- a/kernel/trace/trace_sched_switch.c +++ b/kernel/trace/trace_sched_switch.c @@ -43,7 +43,7 @@ probe_sched_switch(struct rq *__rq, struct task_struct *prev, data = ctx_trace->data[cpu]; if (likely(!atomic_read(&data->disabled))) - tracing_sched_switch_trace(ctx_trace, data, prev, next, flags, pc); + tracing_sched_switch_trace(ctx_trace, prev, next, flags, pc); local_irq_restore(flags); } @@ -66,7 +66,7 @@ probe_sched_wakeup(struct rq *__rq, struct task_struct *wakee, int success) data = ctx_trace->data[cpu]; if (likely(!atomic_read(&data->disabled))) - tracing_sched_wakeup_trace(ctx_trace, data, wakee, current, + tracing_sched_wakeup_trace(ctx_trace, wakee, current, flags, pc); local_irq_restore(flags); diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index a48c9b4b0c85..96d716485898 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -72,7 +72,7 @@ wakeup_tracer_call(unsigned long ip, unsigned long parent_ip) if (task_cpu(wakeup_task) != cpu) goto unlock; - trace_function(tr, data, ip, parent_ip, flags, pc); + trace_function(tr, ip, parent_ip, flags, pc); unlock: __raw_spin_unlock(&wakeup_lock); @@ -152,8 +152,8 @@ probe_wakeup_sched_switch(struct rq *rq, struct task_struct *prev, if (unlikely(!tracer_enabled || next != wakeup_task)) goto out_unlock; - trace_function(wakeup_trace, data, CALLER_ADDR1, CALLER_ADDR2, flags, pc); - tracing_sched_switch_trace(wakeup_trace, data, prev, next, flags, pc); + trace_function(wakeup_trace, CALLER_ADDR1, CALLER_ADDR2, flags, pc); + tracing_sched_switch_trace(wakeup_trace, prev, next, flags, pc); /* * usecs conversion is slow so we try to delay the conversion @@ -254,10 +254,8 @@ probe_wakeup(struct rq *rq, struct task_struct *p, int success) data = wakeup_trace->data[wakeup_cpu]; data->preempt_timestamp = ftrace_now(cpu); - tracing_sched_wakeup_trace(wakeup_trace, data, p, current, - flags, pc); - trace_function(wakeup_trace, data, CALLER_ADDR1, CALLER_ADDR2, - flags, pc); + tracing_sched_wakeup_trace(wakeup_trace, p, current, flags, pc); + trace_function(wakeup_trace, CALLER_ADDR1, CALLER_ADDR2, flags, pc); out_locked: __raw_spin_unlock(&wakeup_lock); From dac74940289f350c2590bec92737833bad608541 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Feb 2009 01:13:38 -0500 Subject: [PATCH 101/413] trace: code style clean up Ingo Molnar suggested using goto logic to keep the indentation down and to be able to remove the nasty line breaks. This actually makes the code a bit more readable. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 33 +++++++++++++++++---------------- 1 file changed, 17 insertions(+), 16 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 1d4ff568cc4d..3536ef41575d 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -522,23 +522,24 @@ int register_tracer(struct tracer *type) tracing_selftest_running = false; mutex_unlock(&trace_types_lock); - if (!ret && default_bootup_tracer) { - if (!strncmp(default_bootup_tracer, type->name, - BOOTUP_TRACER_SIZE)) { - printk(KERN_INFO "Starting tracer '%s'\n", - type->name); - /* Do we want this tracer to start on bootup? */ - tracing_set_tracer(type->name); - default_bootup_tracer = NULL; - /* disable other selftests, since this will break it. */ - tracing_selftest_disabled = 1; -#ifdef CONFIG_FTRACE_STARTUP_TEST - printk(KERN_INFO "Disabling FTRACE selftests due" - " to running tracer '%s'\n", type->name); -#endif - } - } + if (ret || !default_bootup_tracer) + goto out_unlock; + if (strncmp(default_bootup_tracer, type->name, BOOTUP_TRACER_SIZE)) + goto out_unlock; + + printk(KERN_INFO "Starting tracer '%s'\n", type->name); + /* Do we want this tracer to start on bootup? */ + tracing_set_tracer(type->name); + default_bootup_tracer = NULL; + /* disable other selftests, since this will break it. */ + tracing_selftest_disabled = 1; +#ifdef CONFIG_FTRACE_STARTUP_TEST + printk(KERN_INFO "Disabling FTRACE selftests due to running tracer '%s'\n", + type->name); +#endif + + out_unlock: lock_kernel(); return ret; } From 0a9877514c4fed10a70720293b37213dd172ee3e Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 5 Feb 2009 16:12:56 -0200 Subject: [PATCH 102/413] ring_buffer: remove unused flags parameter MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: API change, cleanup >From ring_buffer_{lock_reserve,unlock_commit}. $ codiff /tmp/vmlinux.before /tmp/vmlinux.after linux-2.6-tip/kernel/trace/trace.c: trace_vprintk | -14 trace_graph_return | -14 trace_graph_entry | -10 trace_function | -8 __ftrace_trace_stack | -8 ftrace_trace_userstack | -8 tracing_sched_switch_trace | -8 ftrace_trace_special | -12 tracing_sched_wakeup_trace | -8 9 functions changed, 90 bytes removed, diff: -90 linux-2.6-tip/block/blktrace.c: __blk_add_trace | -1 1 function changed, 1 bytes removed, diff: -1 /tmp/vmlinux.after: 10 functions changed, 91 bytes removed, diff: -91 Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Frédéric Weisbecker Signed-off-by: Ingo Molnar --- block/blktrace.c | 8 ++--- include/linux/ring_buffer.h | 9 ++--- kernel/trace/kmemtrace.c | 12 +++---- kernel/trace/ring_buffer.c | 9 ++--- kernel/trace/trace.c | 56 ++++++++++++-------------------- kernel/trace/trace_boot.c | 12 +++---- kernel/trace/trace_branch.c | 7 ++-- kernel/trace/trace_hw_branches.c | 6 ++-- kernel/trace/trace_mmiotrace.c | 12 +++---- kernel/trace/trace_power.c | 12 +++---- 10 files changed, 51 insertions(+), 92 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index d9d7146ee023..8e52f24cc8f9 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -165,7 +165,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, struct task_struct *tsk = current; struct ring_buffer_event *event = NULL; struct blk_io_trace *t; - unsigned long flags; + unsigned long flags = 0; unsigned long *sequence; pid_t pid; int cpu, pc = 0; @@ -191,7 +191,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, tracing_record_cmdline(current); event = ring_buffer_lock_reserve(blk_tr->buffer, - sizeof(*t) + pdu_len, &flags); + sizeof(*t) + pdu_len); if (!event) return; @@ -241,11 +241,11 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, memcpy((void *) t + sizeof(*t), pdu_data, pdu_len); if (blk_tr) { - ring_buffer_unlock_commit(blk_tr->buffer, event, flags); + ring_buffer_unlock_commit(blk_tr->buffer, event); if (pid != 0 && !(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC) && (trace_flags & TRACE_ITER_STACKTRACE) != 0) - __trace_stack(blk_tr, flags, 5, pc); + __trace_stack(blk_tr, 0, 5, pc); trace_wake_up(); return; } diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index b3b359660082..3110d92e7d81 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -74,13 +74,10 @@ void ring_buffer_free(struct ring_buffer *buffer); int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size); -struct ring_buffer_event * -ring_buffer_lock_reserve(struct ring_buffer *buffer, - unsigned long length, - unsigned long *flags); +struct ring_buffer_event *ring_buffer_lock_reserve(struct ring_buffer *buffer, + unsigned long length); int ring_buffer_unlock_commit(struct ring_buffer *buffer, - struct ring_buffer_event *event, - unsigned long flags); + struct ring_buffer_event *event); int ring_buffer_write(struct ring_buffer *buffer, unsigned long length, void *data); diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c index f04c0625f1cd..256749d1032a 100644 --- a/kernel/trace/kmemtrace.c +++ b/kernel/trace/kmemtrace.c @@ -272,13 +272,11 @@ void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, struct ring_buffer_event *event; struct kmemtrace_alloc_entry *entry; struct trace_array *tr = kmemtrace_array; - unsigned long irq_flags; if (!kmem_tracing_enabled) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -292,7 +290,7 @@ void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, entry->gfp_flags = gfp_flags; entry->node = node; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); } @@ -305,13 +303,11 @@ void kmemtrace_mark_free(enum kmemtrace_type_id type_id, struct ring_buffer_event *event; struct kmemtrace_free_entry *entry; struct trace_array *tr = kmemtrace_array; - unsigned long irq_flags; if (!kmem_tracing_enabled) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -322,7 +318,7 @@ void kmemtrace_mark_free(enum kmemtrace_type_id type_id, entry->call_site = call_site; entry->ptr = ptr; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); } diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index b36d7374ceef..aee76b3eeed2 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1257,7 +1257,6 @@ static DEFINE_PER_CPU(int, rb_need_resched); * ring_buffer_lock_reserve - reserve a part of the buffer * @buffer: the ring buffer to reserve from * @length: the length of the data to reserve (excluding event header) - * @flags: a pointer to save the interrupt flags * * Returns a reseverd event on the ring buffer to copy directly to. * The user of this interface will need to get the body to write into @@ -1270,9 +1269,7 @@ static DEFINE_PER_CPU(int, rb_need_resched); * If NULL is returned, then nothing has been allocated or locked. */ struct ring_buffer_event * -ring_buffer_lock_reserve(struct ring_buffer *buffer, - unsigned long length, - unsigned long *flags) +ring_buffer_lock_reserve(struct ring_buffer *buffer, unsigned long length) { struct ring_buffer_per_cpu *cpu_buffer; struct ring_buffer_event *event; @@ -1339,15 +1336,13 @@ static void rb_commit(struct ring_buffer_per_cpu *cpu_buffer, * ring_buffer_unlock_commit - commit a reserved * @buffer: The buffer to commit to * @event: The event pointer to commit. - * @flags: the interrupt flags received from ring_buffer_lock_reserve. * * This commits the data to the ring buffer, and releases any locks held. * * Must be paired with ring_buffer_lock_reserve. */ int ring_buffer_unlock_commit(struct ring_buffer *buffer, - struct ring_buffer_event *event, - unsigned long flags) + struct ring_buffer_event *event) { struct ring_buffer_per_cpu *cpu_buffer; int cpu = raw_smp_processor_id(); diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3536ef41575d..eb453a238a6f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -783,14 +783,12 @@ trace_function(struct trace_array *tr, { struct ring_buffer_event *event; struct ftrace_entry *entry; - unsigned long irq_flags; /* If we are reading the ring buffer, don't trace */ if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -798,7 +796,7 @@ trace_function(struct trace_array *tr, entry->ent.type = TRACE_FN; entry->ip = ip; entry->parent_ip = parent_ip; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); } #ifdef CONFIG_FUNCTION_GRAPH_TRACER @@ -809,20 +807,18 @@ static void __trace_graph_entry(struct trace_array *tr, { struct ring_buffer_event *event; struct ftrace_graph_ent_entry *entry; - unsigned long irq_flags; if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) return; - event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, flags, pc); entry->ent.type = TRACE_GRAPH_ENT; entry->graph_ent = *trace; - ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags); + ring_buffer_unlock_commit(global_trace.buffer, event); } static void __trace_graph_return(struct trace_array *tr, @@ -832,20 +828,18 @@ static void __trace_graph_return(struct trace_array *tr, { struct ring_buffer_event *event; struct ftrace_graph_ret_entry *entry; - unsigned long irq_flags; if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) return; - event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, flags, pc); entry->ent.type = TRACE_GRAPH_RET; entry->ret = *trace; - ring_buffer_unlock_commit(global_trace.buffer, event, irq_flags); + ring_buffer_unlock_commit(global_trace.buffer, event); } #endif @@ -866,10 +860,8 @@ static void __ftrace_trace_stack(struct trace_array *tr, struct ring_buffer_event *event; struct stack_entry *entry; struct stack_trace trace; - unsigned long irq_flags; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -884,7 +876,7 @@ static void __ftrace_trace_stack(struct trace_array *tr, trace.entries = entry->caller; save_stack_trace(&trace); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); #endif } @@ -912,13 +904,11 @@ static void ftrace_trace_userstack(struct trace_array *tr, struct ring_buffer_event *event; struct userstack_entry *entry; struct stack_trace trace; - unsigned long irq_flags; if (!(trace_flags & TRACE_ITER_USERSTACKTRACE)) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -933,7 +923,7 @@ static void ftrace_trace_userstack(struct trace_array *tr, trace.entries = entry->caller; save_stack_trace_user(&trace); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); #endif } @@ -950,10 +940,8 @@ ftrace_trace_special(void *__tr, struct ring_buffer_event *event; struct trace_array *tr = __tr; struct special_entry *entry; - unsigned long irq_flags; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -962,9 +950,9 @@ ftrace_trace_special(void *__tr, entry->arg1 = arg1; entry->arg2 = arg2; entry->arg3 = arg3; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); - ftrace_trace_stack(tr, irq_flags, 4, pc); - ftrace_trace_userstack(tr, irq_flags, pc); + ring_buffer_unlock_commit(tr->buffer, event); + ftrace_trace_stack(tr, 0, 4, pc); + ftrace_trace_userstack(tr, 0, pc); trace_wake_up(); } @@ -984,10 +972,8 @@ tracing_sched_switch_trace(struct trace_array *tr, { struct ring_buffer_event *event; struct ctx_switch_entry *entry; - unsigned long irq_flags; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -1000,7 +986,7 @@ tracing_sched_switch_trace(struct trace_array *tr, entry->next_prio = next->prio; entry->next_state = next->state; entry->next_cpu = task_cpu(next); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); ftrace_trace_stack(tr, flags, 5, pc); ftrace_trace_userstack(tr, flags, pc); } @@ -1013,10 +999,8 @@ tracing_sched_wakeup_trace(struct trace_array *tr, { struct ring_buffer_event *event; struct ctx_switch_entry *entry; - unsigned long irq_flags; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) return; entry = ring_buffer_event_data(event); @@ -1029,7 +1013,7 @@ tracing_sched_wakeup_trace(struct trace_array *tr, entry->next_prio = wakee->prio; entry->next_state = wakee->state; entry->next_cpu = task_cpu(wakee); - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); ftrace_trace_stack(tr, flags, 6, pc); ftrace_trace_userstack(tr, flags, pc); @@ -2841,7 +2825,7 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) trace_buf[len] = 0; size = sizeof(*entry) + len + 1; - event = ring_buffer_lock_reserve(tr->buffer, size, &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, size); if (!event) goto out_unlock; entry = ring_buffer_event_data(event); @@ -2852,7 +2836,7 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) memcpy(&entry->buf, trace_buf, len); entry->buf[len] = 0; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); out_unlock: spin_unlock_irqrestore(&trace_buf_lock, irq_flags); diff --git a/kernel/trace/trace_boot.c b/kernel/trace/trace_boot.c index 1f07895977a0..4e08debf662d 100644 --- a/kernel/trace/trace_boot.c +++ b/kernel/trace/trace_boot.c @@ -132,7 +132,6 @@ void trace_boot_call(struct boot_trace_call *bt, initcall_t fn) { struct ring_buffer_event *event; struct trace_boot_call *entry; - unsigned long irq_flags; struct trace_array *tr = boot_trace; if (!tr || !pre_initcalls_finished) @@ -144,15 +143,14 @@ void trace_boot_call(struct boot_trace_call *bt, initcall_t fn) sprint_symbol(bt->func, (unsigned long)fn); preempt_disable(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) goto out; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, 0); entry->ent.type = TRACE_BOOT_CALL; entry->boot_call = *bt; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); @@ -164,7 +162,6 @@ void trace_boot_ret(struct boot_trace_ret *bt, initcall_t fn) { struct ring_buffer_event *event; struct trace_boot_ret *entry; - unsigned long irq_flags; struct trace_array *tr = boot_trace; if (!tr || !pre_initcalls_finished) @@ -173,15 +170,14 @@ void trace_boot_ret(struct boot_trace_ret *bt, initcall_t fn) sprint_symbol(bt->func, (unsigned long)fn); preempt_disable(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) goto out; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, 0); entry->ent.type = TRACE_BOOT_RET; entry->boot_ret = *bt; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 027e83690615..770e52acfc10 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -33,7 +33,7 @@ probe_likely_condition(struct ftrace_branch_data *f, int val, int expect) struct trace_array *tr = branch_tracer; struct ring_buffer_event *event; struct trace_branch *entry; - unsigned long flags, irq_flags; + unsigned long flags; int cpu, pc; const char *p; @@ -52,8 +52,7 @@ probe_likely_condition(struct ftrace_branch_data *f, int val, int expect) if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) goto out; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) goto out; @@ -75,7 +74,7 @@ probe_likely_condition(struct ftrace_branch_data *f, int val, int expect) entry->line = f->line; entry->correct = val == expect; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); out: atomic_dec(&tr->data[cpu]->disabled); diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index fff3545fc866..e720c001db2b 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -175,7 +175,7 @@ void trace_hw_branch(u64 from, u64 to) struct trace_array *tr = hw_branch_trace; struct ring_buffer_event *event; struct hw_branch_entry *entry; - unsigned long irq1, irq2; + unsigned long irq1; int cpu; if (unlikely(!tr)) @@ -189,7 +189,7 @@ void trace_hw_branch(u64 from, u64 to) if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) goto out; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), &irq2); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) goto out; entry = ring_buffer_event_data(event); @@ -198,7 +198,7 @@ void trace_hw_branch(u64 from, u64 to) entry->ent.cpu = cpu; entry->from = from; entry->to = to; - ring_buffer_unlock_commit(tr->buffer, event, irq2); + ring_buffer_unlock_commit(tr->buffer, event); out: atomic_dec(&tr->data[cpu]->disabled); diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index ec78e244242e..104ddebc11d1 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -307,10 +307,8 @@ static void __trace_mmiotrace_rw(struct trace_array *tr, { struct ring_buffer_event *event; struct trace_mmiotrace_rw *entry; - unsigned long irq_flags; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) { atomic_inc(&dropped_count); return; @@ -319,7 +317,7 @@ static void __trace_mmiotrace_rw(struct trace_array *tr, tracing_generic_entry_update(&entry->ent, 0, preempt_count()); entry->ent.type = TRACE_MMIO_RW; entry->rw = *rw; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); } @@ -337,10 +335,8 @@ static void __trace_mmiotrace_map(struct trace_array *tr, { struct ring_buffer_event *event; struct trace_mmiotrace_map *entry; - unsigned long irq_flags; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) { atomic_inc(&dropped_count); return; @@ -349,7 +345,7 @@ static void __trace_mmiotrace_map(struct trace_array *tr, tracing_generic_entry_update(&entry->ent, 0, preempt_count()); entry->ent.type = TRACE_MMIO_MAP; entry->map = *map; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); } diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c index faa6ab7a1f5c..3b1a292d12d2 100644 --- a/kernel/trace/trace_power.c +++ b/kernel/trace/trace_power.c @@ -115,7 +115,6 @@ void trace_power_end(struct power_trace *it) struct ring_buffer_event *event; struct trace_power *entry; struct trace_array_cpu *data; - unsigned long irq_flags; struct trace_array *tr = power_trace; if (!trace_power_enabled) @@ -125,15 +124,14 @@ void trace_power_end(struct power_trace *it) it->end = ktime_get(); data = tr->data[smp_processor_id()]; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) goto out; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, 0); entry->ent.type = TRACE_POWER; entry->state_data = *it; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); @@ -148,7 +146,6 @@ void trace_power_mark(struct power_trace *it, unsigned int type, struct ring_buffer_event *event; struct trace_power *entry; struct trace_array_cpu *data; - unsigned long irq_flags; struct trace_array *tr = power_trace; if (!trace_power_enabled) @@ -162,15 +159,14 @@ void trace_power_mark(struct power_trace *it, unsigned int type, it->end = it->stamp; data = tr->data[smp_processor_id()]; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry), - &irq_flags); + event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); if (!event) goto out; entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, 0); entry->ent.type = TRACE_POWER; entry->state_data = *it; - ring_buffer_unlock_commit(tr->buffer, event, irq_flags); + ring_buffer_unlock_commit(tr->buffer, event); trace_wake_up(); From 51a763dd84253bab1d0a1e68e11a7753d1b702ca Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 5 Feb 2009 16:14:13 -0200 Subject: [PATCH 103/413] tracing: Introduce trace_buffer_{lock_reserve,unlock_commit} MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: new API These new functions do what previously was being open coded, reducing the number of details ftrace plugin writers have to worry about. It also standardizes the handling of stacktrace, userstacktrace and other trace options we may introduce in the future. With this patch, for instance, the blk tracer (and some others already in the tree) can use the "userstacktrace" /d/tracing/trace_options facility. $ codiff /tmp/vmlinux.before /tmp/vmlinux.after linux-2.6-tip/kernel/trace/trace.c: trace_vprintk | -5 trace_graph_return | -22 trace_graph_entry | -26 trace_function | -45 __ftrace_trace_stack | -27 ftrace_trace_userstack | -29 tracing_sched_switch_trace | -66 tracing_stop | +1 trace_seq_to_user | -1 ftrace_trace_special | -63 ftrace_special | +1 tracing_sched_wakeup_trace | -70 tracing_reset_online_cpus | -1 13 functions changed, 2 bytes added, 355 bytes removed, diff: -353 linux-2.6-tip/block/blktrace.c: __blk_add_trace | -58 1 function changed, 58 bytes removed, diff: -58 linux-2.6-tip/kernel/trace/trace.c: trace_buffer_lock_reserve | +88 trace_buffer_unlock_commit | +86 2 functions changed, 174 bytes added, diff: +174 /tmp/vmlinux.after: 16 functions changed, 176 bytes added, 413 bytes removed, diff: -237 Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Frédéric Weisbecker Signed-off-by: Ingo Molnar --- block/blktrace.c | 21 ++----- kernel/trace/kmemtrace.c | 19 ++----- kernel/trace/trace.c | 94 ++++++++++++++++++-------------- kernel/trace/trace.h | 11 ++++ kernel/trace/trace_boot.c | 20 ++----- kernel/trace/trace_branch.c | 7 +-- kernel/trace/trace_hw_branches.c | 7 +-- kernel/trace/trace_mmiotrace.c | 20 +++---- kernel/trace/trace_power.c | 20 ++----- 9 files changed, 102 insertions(+), 117 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 8e52f24cc8f9..834cd84037b2 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -187,19 +187,15 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, cpu = raw_smp_processor_id(); if (blk_tr) { - struct trace_entry *ent; tracing_record_cmdline(current); - event = ring_buffer_lock_reserve(blk_tr->buffer, - sizeof(*t) + pdu_len); + pc = preempt_count(); + event = trace_buffer_lock_reserve(blk_tr, TRACE_BLK, + sizeof(*t) + pdu_len, + 0, pc); if (!event) return; - - ent = ring_buffer_event_data(event); - t = (struct blk_io_trace *)ent; - pc = preempt_count(); - tracing_generic_entry_update(ent, 0, pc); - ent->type = TRACE_BLK; + t = ring_buffer_event_data(event); goto record_it; } @@ -241,12 +237,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, memcpy((void *) t + sizeof(*t), pdu_data, pdu_len); if (blk_tr) { - ring_buffer_unlock_commit(blk_tr->buffer, event); - if (pid != 0 && - !(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC) && - (trace_flags & TRACE_ITER_STACKTRACE) != 0) - __trace_stack(blk_tr, 0, 5, pc); - trace_wake_up(); + trace_buffer_unlock_commit(blk_tr, event, 0, pc); return; } } diff --git a/kernel/trace/kmemtrace.c b/kernel/trace/kmemtrace.c index 256749d1032a..ae201b3eda89 100644 --- a/kernel/trace/kmemtrace.c +++ b/kernel/trace/kmemtrace.c @@ -276,13 +276,12 @@ void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, if (!kmem_tracing_enabled) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_KMEM_ALLOC, + sizeof(*entry), 0, 0); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_KMEM_ALLOC; entry->call_site = call_site; entry->ptr = ptr; entry->bytes_req = bytes_req; @@ -290,9 +289,7 @@ void kmemtrace_mark_alloc_node(enum kmemtrace_type_id type_id, entry->gfp_flags = gfp_flags; entry->node = node; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); + trace_buffer_unlock_commit(tr, event, 0, 0); } EXPORT_SYMBOL(kmemtrace_mark_alloc_node); @@ -307,20 +304,16 @@ void kmemtrace_mark_free(enum kmemtrace_type_id type_id, if (!kmem_tracing_enabled) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_KMEM_FREE, + sizeof(*entry), 0, 0); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - - entry->ent.type = TRACE_KMEM_FREE; entry->type_id = type_id; entry->call_site = call_site; entry->ptr = ptr; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); + trace_buffer_unlock_commit(tr, event, 0, 0); } EXPORT_SYMBOL(kmemtrace_mark_free); diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index eb453a238a6f..8fad3776e843 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -776,6 +776,39 @@ tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, (need_resched() ? TRACE_FLAG_NEED_RESCHED : 0); } +struct ring_buffer_event *trace_buffer_lock_reserve(struct trace_array *tr, + unsigned char type, + unsigned long len, + unsigned long flags, int pc) +{ + struct ring_buffer_event *event; + + event = ring_buffer_lock_reserve(tr->buffer, len); + if (event != NULL) { + struct trace_entry *ent = ring_buffer_event_data(event); + + tracing_generic_entry_update(ent, flags, pc); + ent->type = type; + } + + return event; +} +static void ftrace_trace_stack(struct trace_array *tr, + unsigned long flags, int skip, int pc); +static void ftrace_trace_userstack(struct trace_array *tr, + unsigned long flags, int pc); + +void trace_buffer_unlock_commit(struct trace_array *tr, + struct ring_buffer_event *event, + unsigned long flags, int pc) +{ + ring_buffer_unlock_commit(tr->buffer, event); + + ftrace_trace_stack(tr, flags, 6, pc); + ftrace_trace_userstack(tr, flags, pc); + trace_wake_up(); +} + void trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, @@ -788,12 +821,11 @@ trace_function(struct trace_array *tr, if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_FN, sizeof(*entry), + flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_FN; entry->ip = ip; entry->parent_ip = parent_ip; ring_buffer_unlock_commit(tr->buffer, event); @@ -811,12 +843,11 @@ static void __trace_graph_entry(struct trace_array *tr, if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) return; - event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_ENT, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_GRAPH_ENT; entry->graph_ent = *trace; ring_buffer_unlock_commit(global_trace.buffer, event); } @@ -832,12 +863,11 @@ static void __trace_graph_return(struct trace_array *tr, if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) return; - event = ring_buffer_lock_reserve(global_trace.buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_RET, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_GRAPH_RET; entry->ret = *trace; ring_buffer_unlock_commit(global_trace.buffer, event); } @@ -861,13 +891,11 @@ static void __ftrace_trace_stack(struct trace_array *tr, struct stack_entry *entry; struct stack_trace trace; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_STACK, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_STACK; - memset(&entry->caller, 0, sizeof(entry->caller)); trace.nr_entries = 0; @@ -908,12 +936,11 @@ static void ftrace_trace_userstack(struct trace_array *tr, if (!(trace_flags & TRACE_ITER_USERSTACKTRACE)) return; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_USER_STACK, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_USER_STACK; memset(&entry->caller, 0, sizeof(entry->caller)); @@ -941,20 +968,15 @@ ftrace_trace_special(void *__tr, struct trace_array *tr = __tr; struct special_entry *entry; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_SPECIAL, + sizeof(*entry), 0, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, pc); - entry->ent.type = TRACE_SPECIAL; entry->arg1 = arg1; entry->arg2 = arg2; entry->arg3 = arg3; - ring_buffer_unlock_commit(tr->buffer, event); - ftrace_trace_stack(tr, 0, 4, pc); - ftrace_trace_userstack(tr, 0, pc); - - trace_wake_up(); + trace_buffer_unlock_commit(tr, event, 0, pc); } void @@ -973,12 +995,11 @@ tracing_sched_switch_trace(struct trace_array *tr, struct ring_buffer_event *event; struct ctx_switch_entry *entry; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_CTX, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_CTX; entry->prev_pid = prev->pid; entry->prev_prio = prev->prio; entry->prev_state = prev->state; @@ -986,9 +1007,7 @@ tracing_sched_switch_trace(struct trace_array *tr, entry->next_prio = next->prio; entry->next_state = next->state; entry->next_cpu = task_cpu(next); - ring_buffer_unlock_commit(tr->buffer, event); - ftrace_trace_stack(tr, flags, 5, pc); - ftrace_trace_userstack(tr, flags, pc); + trace_buffer_unlock_commit(tr, event, flags, pc); } void @@ -1000,12 +1019,11 @@ tracing_sched_wakeup_trace(struct trace_array *tr, struct ring_buffer_event *event; struct ctx_switch_entry *entry; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_WAKE, + sizeof(*entry), flags, pc); if (!event) return; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_WAKE; entry->prev_pid = curr->pid; entry->prev_prio = curr->prio; entry->prev_state = curr->state; @@ -1013,11 +1031,7 @@ tracing_sched_wakeup_trace(struct trace_array *tr, entry->next_prio = wakee->prio; entry->next_state = wakee->state; entry->next_cpu = task_cpu(wakee); - ring_buffer_unlock_commit(tr->buffer, event); - ftrace_trace_stack(tr, flags, 6, pc); - ftrace_trace_userstack(tr, flags, pc); - - trace_wake_up(); + trace_buffer_unlock_commit(tr, event, flags, pc); } void @@ -2825,12 +2839,10 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) trace_buf[len] = 0; size = sizeof(*entry) + len + 1; - event = ring_buffer_lock_reserve(tr->buffer, size); + event = trace_buffer_lock_reserve(tr, TRACE_PRINT, size, irq_flags, pc); if (!event) goto out_unlock; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, irq_flags, pc); - entry->ent.type = TRACE_PRINT; entry->ip = ip; entry->depth = depth; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index df627a948694..e03f157c772e 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -403,6 +403,17 @@ int tracing_open_generic(struct inode *inode, struct file *filp); struct dentry *tracing_init_dentry(void); void init_tracer_sysprof_debugfs(struct dentry *d_tracer); +struct ring_buffer_event; + +struct ring_buffer_event *trace_buffer_lock_reserve(struct trace_array *tr, + unsigned char type, + unsigned long len, + unsigned long flags, + int pc); +void trace_buffer_unlock_commit(struct trace_array *tr, + struct ring_buffer_event *event, + unsigned long flags, int pc); + struct trace_entry *tracing_get_trace_entry(struct trace_array *tr, struct trace_array_cpu *data); diff --git a/kernel/trace/trace_boot.c b/kernel/trace/trace_boot.c index 4e08debf662d..7a30fc4c3642 100644 --- a/kernel/trace/trace_boot.c +++ b/kernel/trace/trace_boot.c @@ -143,17 +143,13 @@ void trace_boot_call(struct boot_trace_call *bt, initcall_t fn) sprint_symbol(bt->func, (unsigned long)fn); preempt_disable(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_BOOT_CALL, + sizeof(*entry), 0, 0); if (!event) goto out; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_BOOT_CALL; entry->boot_call = *bt; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); - + trace_buffer_unlock_commit(tr, event, 0, 0); out: preempt_enable(); } @@ -170,17 +166,13 @@ void trace_boot_ret(struct boot_trace_ret *bt, initcall_t fn) sprint_symbol(bt->func, (unsigned long)fn); preempt_disable(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_BOOT_RET, + sizeof(*entry), 0, 0); if (!event) goto out; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_BOOT_RET; entry->boot_ret = *bt; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); - + trace_buffer_unlock_commit(tr, event, 0, 0); out: preempt_enable(); } diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 770e52acfc10..48b2196abe37 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -52,14 +52,13 @@ probe_likely_condition(struct ftrace_branch_data *f, int val, int expect) if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) goto out; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + pc = preempt_count(); + event = trace_buffer_lock_reserve(tr, TRACE_BRANCH, + sizeof(*entry), flags, pc); if (!event) goto out; - pc = preempt_count(); entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, flags, pc); - entry->ent.type = TRACE_BRANCH; /* Strip off the path, only save the file */ p = f->file + strlen(f->file); diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index e720c001db2b..2aa1c9f4c7d8 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -189,16 +189,15 @@ void trace_hw_branch(u64 from, u64 to) if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) goto out; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_HW_BRANCHES, + sizeof(*entry), 0, 0); if (!event) goto out; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, from); - entry->ent.type = TRACE_HW_BRANCHES; entry->ent.cpu = cpu; entry->from = from; entry->to = to; - ring_buffer_unlock_commit(tr->buffer, event); + trace_buffer_unlock_commit(tr, event, 0, 0); out: atomic_dec(&tr->data[cpu]->disabled); diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index 104ddebc11d1..c401b908e805 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -307,19 +307,17 @@ static void __trace_mmiotrace_rw(struct trace_array *tr, { struct ring_buffer_event *event; struct trace_mmiotrace_rw *entry; + int pc = preempt_count(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_MMIO_RW, + sizeof(*entry), 0, pc); if (!event) { atomic_inc(&dropped_count); return; } entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, preempt_count()); - entry->ent.type = TRACE_MMIO_RW; entry->rw = *rw; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); + trace_buffer_unlock_commit(tr, event, 0, pc); } void mmio_trace_rw(struct mmiotrace_rw *rw) @@ -335,19 +333,17 @@ static void __trace_mmiotrace_map(struct trace_array *tr, { struct ring_buffer_event *event; struct trace_mmiotrace_map *entry; + int pc = preempt_count(); - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_MMIO_MAP, + sizeof(*entry), 0, pc); if (!event) { atomic_inc(&dropped_count); return; } entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, preempt_count()); - entry->ent.type = TRACE_MMIO_MAP; entry->map = *map; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); + trace_buffer_unlock_commit(tr, event, 0, pc); } void mmio_trace_mapping(struct mmiotrace_map *map) diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c index 3b1a292d12d2..bfc21f8079ab 100644 --- a/kernel/trace/trace_power.c +++ b/kernel/trace/trace_power.c @@ -124,17 +124,13 @@ void trace_power_end(struct power_trace *it) it->end = ktime_get(); data = tr->data[smp_processor_id()]; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_POWER, + sizeof(*entry), 0, 0); if (!event) goto out; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_POWER; entry->state_data = *it; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); - + trace_buffer_unlock_commit(tr, event, 0, 0); out: preempt_enable(); } @@ -159,17 +155,13 @@ void trace_power_mark(struct power_trace *it, unsigned int type, it->end = it->stamp; data = tr->data[smp_processor_id()]; - event = ring_buffer_lock_reserve(tr->buffer, sizeof(*entry)); + event = trace_buffer_lock_reserve(tr, TRACE_POWER, + sizeof(*entry), 0, 0); if (!event) goto out; entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, 0); - entry->ent.type = TRACE_POWER; entry->state_data = *it; - ring_buffer_unlock_commit(tr->buffer, event); - - trace_wake_up(); - + trace_buffer_unlock_commit(tr, event, 0, 0); out: preempt_enable(); } From b6f11df26fdc28324cf9c9e3b77f2dc985c1bb13 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 5 Feb 2009 18:02:00 -0200 Subject: [PATCH 104/413] trace: Call tracing_reset_online_cpus before tracer->init() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Impact: cleanup To make it easy for ftrace plugin writers, as this was open coded in the existing plugins Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Frédéric Weisbecker Signed-off-by: Ingo Molnar --- block/blktrace.c | 2 -- kernel/trace/trace.c | 8 +++++++- kernel/trace/trace.h | 1 + kernel/trace/trace_branch.c | 1 - kernel/trace/trace_functions.c | 17 +++-------------- kernel/trace/trace_functions_graph.c | 1 - kernel/trace/trace_hw_branches.c | 1 - kernel/trace/trace_nop.c | 1 - kernel/trace/trace_sched_switch.c | 8 +------- kernel/trace/trace_selftest.c | 18 +++++++++--------- kernel/trace/trace_sysprof.c | 14 ++++---------- 11 files changed, 25 insertions(+), 47 deletions(-) diff --git a/block/blktrace.c b/block/blktrace.c index 834cd84037b2..ca6d32061e4f 100644 --- a/block/blktrace.c +++ b/block/blktrace.c @@ -1086,8 +1086,6 @@ static void blk_tracer_print_header(struct seq_file *m) static void blk_tracer_start(struct trace_array *tr) { - tracing_reset_online_cpus(tr); - mutex_lock(&blk_probe_mutex); if (atomic_add_return(1, &blk_probes_ref) == 1) if (blk_register_tracepoints()) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8fad3776e843..ef4dbac95568 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2171,6 +2171,12 @@ tracing_set_trace_read(struct file *filp, char __user *ubuf, return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); } +int tracer_init(struct tracer *t, struct trace_array *tr) +{ + tracing_reset_online_cpus(tr); + return t->init(tr); +} + static int tracing_set_tracer(const char *buf) { struct trace_array *tr = &global_trace; @@ -2195,7 +2201,7 @@ static int tracing_set_tracer(const char *buf) current_trace = t; if (t->init) { - ret = t->init(tr); + ret = tracer_init(t, tr); if (ret) goto out; } diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index e03f157c772e..f2742fb1575a 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -395,6 +395,7 @@ struct trace_iterator { cpumask_var_t started; }; +int tracer_init(struct tracer *t, struct trace_array *tr); int tracing_is_enabled(void); void trace_wake_up(void); void tracing_reset(struct trace_array *tr, int cpu); diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index 48b2196abe37..f8ae2c50e01d 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -131,7 +131,6 @@ static void stop_branch_trace(struct trace_array *tr) static int branch_trace_init(struct trace_array *tr) { - tracing_reset_online_cpus(tr); start_branch_trace(tr); return 0; } diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index d067cea2ccc3..36bf9568ccd9 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -24,32 +24,21 @@ static struct trace_array *func_trace; static void tracing_start_function_trace(void); static void tracing_stop_function_trace(void); -static void start_function_trace(struct trace_array *tr) +static int function_trace_init(struct trace_array *tr) { func_trace = tr; tr->cpu = get_cpu(); - tracing_reset_online_cpus(tr); put_cpu(); tracing_start_cmdline_record(); tracing_start_function_trace(); -} - -static void stop_function_trace(struct trace_array *tr) -{ - tracing_stop_function_trace(); - tracing_stop_cmdline_record(); -} - -static int function_trace_init(struct trace_array *tr) -{ - start_function_trace(tr); return 0; } static void function_trace_reset(struct trace_array *tr) { - stop_function_trace(tr); + tracing_stop_function_trace(); + tracing_stop_cmdline_record(); } static void function_trace_start(struct trace_array *tr) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index c97594d826bc..222f97d336a6 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -56,7 +56,6 @@ static int graph_trace_init(struct trace_array *tr) &trace_graph_entry); if (ret) return ret; - tracing_reset_online_cpus(tr); tracing_start_cmdline_record(); return 0; diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index 2aa1c9f4c7d8..ca4bbcfb9e2c 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -132,7 +132,6 @@ static int bts_trace_init(struct trace_array *tr) hw_branch_trace = tr; register_hotcpu_notifier(&bts_hotcpu_notifier); - tracing_reset_online_cpus(tr); bts_trace_start(tr); return 0; diff --git a/kernel/trace/trace_nop.c b/kernel/trace/trace_nop.c index 087b6cbf4ea5..9aa84bde23cd 100644 --- a/kernel/trace/trace_nop.c +++ b/kernel/trace/trace_nop.c @@ -48,7 +48,6 @@ static void stop_nop_trace(struct trace_array *tr) static int nop_trace_init(struct trace_array *tr) { ctx_trace = tr; - tracing_reset_online_cpus(tr); start_nop_trace(tr); return 0; } diff --git a/kernel/trace/trace_sched_switch.c b/kernel/trace/trace_sched_switch.c index c4f9add5ec90..30e14fe85896 100644 --- a/kernel/trace/trace_sched_switch.c +++ b/kernel/trace/trace_sched_switch.c @@ -185,12 +185,6 @@ void tracing_sched_switch_assign_trace(struct trace_array *tr) ctx_trace = tr; } -static void start_sched_trace(struct trace_array *tr) -{ - tracing_reset_online_cpus(tr); - tracing_start_sched_switch_record(); -} - static void stop_sched_trace(struct trace_array *tr) { tracing_stop_sched_switch_record(); @@ -199,7 +193,7 @@ static void stop_sched_trace(struct trace_array *tr) static int sched_switch_trace_init(struct trace_array *tr) { ctx_trace = tr; - start_sched_trace(tr); + tracing_start_sched_switch_record(); return 0; } diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 5013812578b1..445700e51f6d 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -115,7 +115,7 @@ int trace_selftest_startup_dynamic_tracing(struct tracer *trace, ftrace_set_filter(func_name, strlen(func_name), 1); /* enable tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); goto out; @@ -189,7 +189,7 @@ trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr) ftrace_enabled = 1; tracer_enabled = 1; - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); goto out; @@ -236,7 +236,7 @@ trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr) int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -290,7 +290,7 @@ trace_selftest_startup_preemptoff(struct tracer *trace, struct trace_array *tr) } /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -344,7 +344,7 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * } /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); goto out; @@ -476,7 +476,7 @@ trace_selftest_startup_wakeup(struct tracer *trace, struct trace_array *tr) wait_for_completion(&isrt); /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -537,7 +537,7 @@ trace_selftest_startup_sched_switch(struct tracer *trace, struct trace_array *tr int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; @@ -569,7 +569,7 @@ trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr) int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return 0; @@ -596,7 +596,7 @@ trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr) int ret; /* start the tracing */ - ret = trace->init(tr); + ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); return ret; diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c index eaca5ad803ff..84ca9d81e74d 100644 --- a/kernel/trace/trace_sysprof.c +++ b/kernel/trace/trace_sysprof.c @@ -226,15 +226,6 @@ static void stop_stack_timers(void) stop_stack_timer(cpu); } -static void start_stack_trace(struct trace_array *tr) -{ - mutex_lock(&sample_timer_lock); - tracing_reset_online_cpus(tr); - start_stack_timers(); - tracer_enabled = 1; - mutex_unlock(&sample_timer_lock); -} - static void stop_stack_trace(struct trace_array *tr) { mutex_lock(&sample_timer_lock); @@ -247,7 +238,10 @@ static int stack_trace_init(struct trace_array *tr) { sysprof_trace = tr; - start_stack_trace(tr); + mutex_lock(&sample_timer_lock); + start_stack_timers(); + tracer_enabled = 1; + mutex_unlock(&sample_timer_lock); return 0; } From 304cc6ae1bf7a8e6d00053fbe0b7e2b26cdddda2 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 6 Feb 2009 01:12:02 +0100 Subject: [PATCH 105/413] ring_buffer: remove unused flags parameter, fix Oprofile's ring-buffer use was not considered. Signed-off-by: Ingo Molnar --- drivers/oprofile/cpu_buffer.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/oprofile/cpu_buffer.c b/drivers/oprofile/cpu_buffer.c index e76d715e4342..f0e99d4c066b 100644 --- a/drivers/oprofile/cpu_buffer.c +++ b/drivers/oprofile/cpu_buffer.c @@ -161,7 +161,7 @@ struct op_sample { entry->event = ring_buffer_lock_reserve (op_ring_buffer_write, sizeof(struct op_sample) + - size * sizeof(entry->sample->data[0]), &entry->irq_flags); + size * sizeof(entry->sample->data[0])); if (entry->event) entry->sample = ring_buffer_event_data(entry->event); else @@ -178,8 +178,7 @@ struct op_sample int op_cpu_buffer_write_commit(struct op_entry *entry) { - return ring_buffer_unlock_commit(op_ring_buffer_write, entry->event, - entry->irq_flags); + return ring_buffer_unlock_commit(op_ring_buffer_write, entry->event); } struct op_sample *op_cpu_buffer_read_entry(struct op_entry *entry, int cpu) From 1830b52d0de8c60c4f5dfbac134aa8f69d815801 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 7 Feb 2009 19:38:43 -0500 Subject: [PATCH 106/413] trace: remove deprecated entry->cpu Impact: fix to prevent developers from using entry->cpu With the new ring buffer infrastructure, the cpu for the entry is implicit with which CPU buffer it is on. The original code use to record the current cpu into the generic entry header, which can be retrieved by entry->cpu. When the ring buffer was introduced, the users were convert to use the the cpu number of which cpu ring buffer was in use (this was passed to the tracers by the iterator: iter->cpu). Unfortunately, the cpu item in the entry structure was never removed. This allowed for developers to use it instead of the proper iter->cpu, unknowingly, using an uninitialized variable. This was not the fault of the developers, since it would seem like the logical place to retrieve the cpu identifier. This patch removes the cpu item from the entry structure and fixes all the users that should have been using iter->cpu. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 +- kernel/trace/trace.h | 1 - kernel/trace/trace_hw_branches.c | 3 +-- kernel/trace/trace_output.c | 6 +++--- 4 files changed, 5 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index fd51cf0b94c7..bd4d9f8818fa 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1531,7 +1531,7 @@ static enum print_line_t print_bin_fmt(struct trace_iterator *iter) if (trace_flags & TRACE_ITER_CONTEXT_INFO) { SEQ_PUT_FIELD_RET(s, entry->pid); - SEQ_PUT_FIELD_RET(s, entry->cpu); + SEQ_PUT_FIELD_RET(s, iter->cpu); SEQ_PUT_FIELD_RET(s, iter->ts); } diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index f0c7a0f08cac..5efc4c707f7e 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -45,7 +45,6 @@ enum trace_type { */ struct trace_entry { unsigned char type; - unsigned char cpu; unsigned char flags; unsigned char preempt_count; int pid; diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index fff3545fc866..549238a9b133 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -159,7 +159,7 @@ static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) trace_assign_type(it, entry); if (entry->type == TRACE_HW_BRANCHES) { - if (trace_seq_printf(seq, "%4d ", entry->cpu) && + if (trace_seq_printf(seq, "%4d ", iter->cpu) && seq_print_ip_sym(seq, it->to, symflags) && trace_seq_printf(seq, "\t <- ") && seq_print_ip_sym(seq, it->from, symflags) && @@ -195,7 +195,6 @@ void trace_hw_branch(u64 from, u64 to) entry = ring_buffer_event_data(event); tracing_generic_entry_update(&entry->ent, 0, from); entry->ent.type = TRACE_HW_BRANCHES; - entry->ent.cpu = cpu; entry->from = from; entry->to = to; ring_buffer_unlock_commit(tr->buffer, event, irq2); diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index b7380eee9fa1..463a310b1d3b 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -333,7 +333,7 @@ int trace_print_context(struct trace_iterator *iter) unsigned long secs = (unsigned long)t; return trace_seq_printf(s, "%16s-%-5d [%03d] %5lu.%06lu: ", - comm, entry->pid, entry->cpu, secs, usec_rem); + comm, entry->pid, iter->cpu, secs, usec_rem); } int trace_print_lat_context(struct trace_iterator *iter) @@ -356,7 +356,7 @@ int trace_print_lat_context(struct trace_iterator *iter) char *comm = trace_find_cmdline(entry->pid); ret = trace_seq_printf(s, "%16s %5d %3d %d %08x %08lx [%08lx]" " %ld.%03ldms (+%ld.%03ldms): ", comm, - entry->pid, entry->cpu, entry->flags, + entry->pid, iter->cpu, entry->flags, entry->preempt_count, iter->idx, ns2usecs(iter->ts), abs_usecs / USEC_PER_MSEC, @@ -364,7 +364,7 @@ int trace_print_lat_context(struct trace_iterator *iter) rel_usecs / USEC_PER_MSEC, rel_usecs % USEC_PER_MSEC); } else { - ret = lat_print_generic(s, entry, entry->cpu); + ret = lat_print_generic(s, entry, iter->cpu); if (ret) ret = lat_print_timestamp(s, abs_usecs, rel_usecs); } From 78d904b46a72fcf15ea6a39672bbef92953876b5 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Feb 2009 18:43:07 -0500 Subject: [PATCH 107/413] ring-buffer: add NMI protection for spinlocks Impact: prevent deadlock in NMI The ring buffers are not yet totally lockless with writing to the buffer. When a writer crosses a page, it grabs a per cpu spinlock to protect against a reader. The spinlocks taken by a writer are not to protect against other writers, since a writer can only write to its own per cpu buffer. The spinlocks protect against readers that can touch any cpu buffer. The writers are made to be reentrant with the spinlocks disabling interrupts. The problem arises when an NMI writes to the buffer, and that write crosses a page boundary. If it grabs a spinlock, it can be racing with another writer (since disabling interrupts does not protect against NMIs) or with a reader on the same CPU. Luckily, most of the users are not reentrant and protects against this issue. But if a user of the ring buffer becomes reentrant (which is what the ring buffers do allow), if the NMI also writes to the ring buffer then we risk the chance of a deadlock. This patch moves the ftrace_nmi_enter called by nmi_enter() to the ring buffer code. It replaces the current ftrace_nmi_enter that is used by arch specific code to arch_ftrace_nmi_enter and updates the Kconfig to handle it. When an NMI is called, it will set a per cpu variable in the ring buffer code and will clear it when the NMI exits. If a write to the ring buffer crosses page boundaries inside an NMI, a trylock is used on the spin lock instead. If the spinlock fails to be acquired, then the entry is discarded. This bug appeared in the ftrace work in the RT tree, where event tracing is reentrant. This workaround solved the deadlocks that appeared there. Signed-off-by: Steven Rostedt --- arch/x86/Kconfig | 1 + arch/x86/kernel/ftrace.c | 8 +++---- include/linux/ftrace_irq.h | 10 +++++++- kernel/trace/Kconfig | 8 +++++++ kernel/trace/ring_buffer.c | 48 ++++++++++++++++++++++++++++++++++++-- 5 files changed, 68 insertions(+), 7 deletions(-) diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 73f7fe8fd4d1..a6be725cb049 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -34,6 +34,7 @@ config X86 select HAVE_FUNCTION_TRACER select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACE_MCOUNT_TEST + select HAVE_FTRACE_NMI_ENTER if DYNAMIC_FTRACE || FUNCTION_GRAPH_TRACER select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64) select HAVE_ARCH_KGDB if !X86_VOYAGER select HAVE_ARCH_TRACEHOOK diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 4d33224c055f..4c683587055b 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -113,7 +113,7 @@ static void ftrace_mod_code(void) MCOUNT_INSN_SIZE); } -void ftrace_nmi_enter(void) +void arch_ftrace_nmi_enter(void) { atomic_inc(&in_nmi); /* Must have in_nmi seen before reading write flag */ @@ -124,7 +124,7 @@ void ftrace_nmi_enter(void) } } -void ftrace_nmi_exit(void) +void arch_ftrace_nmi_exit(void) { /* Finish all executions before clearing in_nmi */ smp_wmb(); @@ -376,12 +376,12 @@ int ftrace_disable_ftrace_graph_caller(void) */ static atomic_t in_nmi; -void ftrace_nmi_enter(void) +void arch_ftrace_nmi_enter(void) { atomic_inc(&in_nmi); } -void ftrace_nmi_exit(void) +void arch_ftrace_nmi_exit(void) { atomic_dec(&in_nmi); } diff --git a/include/linux/ftrace_irq.h b/include/linux/ftrace_irq.h index 366a054d0b05..29de6779a963 100644 --- a/include/linux/ftrace_irq.h +++ b/include/linux/ftrace_irq.h @@ -2,7 +2,15 @@ #define _LINUX_FTRACE_IRQ_H -#if defined(CONFIG_DYNAMIC_FTRACE) || defined(CONFIG_FUNCTION_GRAPH_TRACER) +#ifdef CONFIG_FTRACE_NMI_ENTER +extern void arch_ftrace_nmi_enter(void); +extern void arch_ftrace_nmi_exit(void); +#else +static inline void arch_ftrace_nmi_enter(void) { } +static inline void arch_ftrace_nmi_exit(void) { } +#endif + +#ifdef CONFIG_RING_BUFFER extern void ftrace_nmi_enter(void); extern void ftrace_nmi_exit(void); #else diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 28f2644484d9..25131a5d5e4f 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -9,6 +9,9 @@ config USER_STACKTRACE_SUPPORT config NOP_TRACER bool +config HAVE_FTRACE_NMI_ENTER + bool + config HAVE_FUNCTION_TRACER bool @@ -37,6 +40,11 @@ config TRACER_MAX_TRACE config RING_BUFFER bool +config FTRACE_NMI_ENTER + bool + depends on HAVE_FTRACE_NMI_ENTER + default y + config TRACING bool select DEBUG_FS diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index b36d7374ceef..a60a6a852f42 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -4,6 +4,7 @@ * Copyright (C) 2008 Steven Rostedt */ #include +#include #include #include #include @@ -18,6 +19,35 @@ #include "trace.h" +/* + * Since the write to the buffer is still not fully lockless, + * we must be careful with NMIs. The locks in the writers + * are taken when a write crosses to a new page. The locks + * protect against races with the readers (this will soon + * be fixed with a lockless solution). + * + * Because we can not protect against NMIs, and we want to + * keep traces reentrant, we need to manage what happens + * when we are in an NMI. + */ +static DEFINE_PER_CPU(int, rb_in_nmi); + +void ftrace_nmi_enter(void) +{ + __get_cpu_var(rb_in_nmi)++; + /* call arch specific handler too */ + arch_ftrace_nmi_enter(); +} + +void ftrace_nmi_exit(void) +{ + arch_ftrace_nmi_exit(); + __get_cpu_var(rb_in_nmi)--; + /* NMIs are not recursive */ + WARN_ON_ONCE(__get_cpu_var(rb_in_nmi)); +} + + /* * A fast way to enable or disable all ring buffers is to * call tracing_on or tracing_off. Turning off the ring buffers @@ -982,6 +1012,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, struct ring_buffer *buffer = cpu_buffer->buffer; struct ring_buffer_event *event; unsigned long flags; + bool lock_taken = false; commit_page = cpu_buffer->commit_page; /* we just need to protect against interrupts */ @@ -995,7 +1026,19 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, struct buffer_page *next_page = tail_page; local_irq_save(flags); - __raw_spin_lock(&cpu_buffer->lock); + /* + * NMIs can happen after we take the lock. + * If we are in an NMI, only take the lock + * if it is not already taken. Otherwise + * simply fail. + */ + if (unlikely(__get_cpu_var(rb_in_nmi))) { + if (!__raw_spin_trylock(&cpu_buffer->lock)) + goto out_unlock; + } else + __raw_spin_lock(&cpu_buffer->lock); + + lock_taken = true; rb_inc_page(cpu_buffer, &next_page); @@ -1097,7 +1140,8 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, if (tail <= BUF_PAGE_SIZE) local_set(&tail_page->write, tail); - __raw_spin_unlock(&cpu_buffer->lock); + if (likely(lock_taken)) + __raw_spin_unlock(&cpu_buffer->lock); local_irq_restore(flags); return NULL; } From d8b891a2db13c8ed296158d6f8c4e335896d0cef Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Feb 2009 19:54:51 -0500 Subject: [PATCH 108/413] ring-buffer: allow tracing_off to be used in core kernel code tracing_off() is the fastest way to stop recording to the ring buffers. This may be used in places like panic and die, just before the ftrace_dump is called. This patch adds the appropriate CPP conditionals to make it a stub function when the ring buffer is not configured it. Signed-off-by: Steven Rostedt --- include/linux/ring_buffer.h | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index b3b359660082..ac94c066f6e9 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -124,9 +124,18 @@ unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu); u64 ring_buffer_time_stamp(int cpu); void ring_buffer_normalize_time_stamp(int cpu, u64 *ts); +/* + * The below functions are fine to use outside the tracing facility. + */ +#ifdef CONFIG_RING_BUFFER void tracing_on(void); void tracing_off(void); void tracing_off_permanent(void); +#else +static inline void tracing_on(void) { } +static inline void tracing_off(void) { } +static inline void tracing_off_permanent(void) { } +#endif void *ring_buffer_alloc_read_page(struct ring_buffer *buffer); void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data); From 4e6ea1440c67de32d7c89aacf233472dfc3bce82 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Feb 2009 22:30:07 -0500 Subject: [PATCH 109/413] ftrace, x86: rename in_nmi variable Impact: clean up The in_nmi variable in x86 arch ftrace.c is a misnomer. Andrew Morton pointed out that the in_nmi variable is incremented by all CPUS. It can be set when another CPU is running an NMI. Since this is actually intentional, the fix is to rename it to what it really is: "nmi_running" Signed-off-by: Steven Rostedt --- arch/x86/kernel/ftrace.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 4c683587055b..e3fad2ef622c 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -82,7 +82,7 @@ static unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) * are the same as what exists. */ -static atomic_t in_nmi = ATOMIC_INIT(0); +static atomic_t nmi_running = ATOMIC_INIT(0); static int mod_code_status; /* holds return value of text write */ static int mod_code_write; /* set when NMI should do the write */ static void *mod_code_ip; /* holds the IP to write to */ @@ -115,8 +115,8 @@ static void ftrace_mod_code(void) void arch_ftrace_nmi_enter(void) { - atomic_inc(&in_nmi); - /* Must have in_nmi seen before reading write flag */ + atomic_inc(&nmi_running); + /* Must have nmi_running seen before reading write flag */ smp_mb(); if (mod_code_write) { ftrace_mod_code(); @@ -126,19 +126,19 @@ void arch_ftrace_nmi_enter(void) void arch_ftrace_nmi_exit(void) { - /* Finish all executions before clearing in_nmi */ + /* Finish all executions before clearing nmi_running */ smp_wmb(); - atomic_dec(&in_nmi); + atomic_dec(&nmi_running); } static void wait_for_nmi(void) { - if (!atomic_read(&in_nmi)) + if (!atomic_read(&nmi_running)) return; do { cpu_relax(); - } while(atomic_read(&in_nmi)); + } while (atomic_read(&nmi_running)); nmi_wait_count++; } @@ -374,16 +374,16 @@ int ftrace_disable_ftrace_graph_caller(void) * this page for dynamic ftrace. They have been * simplified to ignore all traces in NMI context. */ -static atomic_t in_nmi; +static atomic_t nmi_running; void arch_ftrace_nmi_enter(void) { - atomic_inc(&in_nmi); + atomic_inc(&nmi_running); } void arch_ftrace_nmi_exit(void) { - atomic_dec(&in_nmi); + atomic_dec(&nmi_running); } #endif /* !CONFIG_DYNAMIC_FTRACE */ @@ -475,7 +475,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) &return_to_handler; /* Nmi's are currently unsupported */ - if (unlikely(atomic_read(&in_nmi))) + if (unlikely(atomic_read(&nmi_running))) return; if (unlikely(atomic_read(¤t->tracing_graph_pause))) From 375b38b4214f29109a393ab762d468054bf52354 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 6 Feb 2009 00:51:37 -0500 Subject: [PATCH 110/413] nmi: add generic nmi tracking state This code adds an in_nmi() macro that uses the current tasks preempt count to track when it is in NMI context. Other parts of the kernel can use this to determine if the context is in NMI context or not. This code was inspired by the -rt patch in_nmi version that was written by Peter Zijlstra, who borrowed that code from Mathieu Desnoyers. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- include/linux/hardirq.h | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index f83288347dda..f3cf86e1465b 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -61,6 +61,12 @@ #error PREEMPT_ACTIVE is too low! #endif +#define NMI_OFFSET (PREEMPT_ACTIVE << 1) + +#if NMI_OFFSET >= 0x80000000 +#error PREEMPT_ACTIVE too high! +#endif + #define hardirq_count() (preempt_count() & HARDIRQ_MASK) #define softirq_count() (preempt_count() & SOFTIRQ_MASK) #define irq_count() (preempt_count() & (HARDIRQ_MASK | SOFTIRQ_MASK)) @@ -73,6 +79,11 @@ #define in_softirq() (softirq_count()) #define in_interrupt() (irq_count()) +/* + * Are we in NMI context? + */ +#define in_nmi() (preempt_count() & NMI_OFFSET) + #if defined(CONFIG_PREEMPT) # define PREEMPT_INATOMIC_BASE kernel_locked() # define PREEMPT_CHECK_OFFSET 1 @@ -167,6 +178,8 @@ extern void irq_exit(void); #define nmi_enter() \ do { \ ftrace_nmi_enter(); \ + BUG_ON(in_nmi()); \ + add_preempt_count(NMI_OFFSET); \ lockdep_off(); \ rcu_nmi_enter(); \ __irq_enter(); \ @@ -177,6 +190,8 @@ extern void irq_exit(void); __irq_exit(); \ rcu_nmi_exit(); \ lockdep_on(); \ + BUG_ON(!in_nmi()); \ + sub_preempt_count(NMI_OFFSET); \ ftrace_nmi_exit(); \ } while (0) From 9a5fd902273d01170fd033691bd70b142baa7309 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 6 Feb 2009 01:14:26 -0500 Subject: [PATCH 111/413] ftrace: change function graph tracer to use new in_nmi The function graph tracer piggy backed onto the dynamic ftracer to use the in_nmi custom code for dynamic tracing. The problem was (as Andrew Morton pointed out) it really only wanted to bail out if the context of the current CPU was in NMI context. But the dynamic ftrace in_nmi custom code was true if _any_ CPU happened to be in NMI context. Now that we have a generic in_nmi interface, this patch changes the function graph code to use it instead of the dynamic ftarce custom code. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- arch/x86/Kconfig | 2 +- arch/x86/kernel/ftrace.c | 21 +-------------------- 2 files changed, 2 insertions(+), 21 deletions(-) diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index a6be725cb049..2cf7bbcaed4e 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -34,7 +34,7 @@ config X86 select HAVE_FUNCTION_TRACER select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACE_MCOUNT_TEST - select HAVE_FTRACE_NMI_ENTER if DYNAMIC_FTRACE || FUNCTION_GRAPH_TRACER + select HAVE_FTRACE_NMI_ENTER if DYNAMIC_FTRACE select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64) select HAVE_ARCH_KGDB if !X86_VOYAGER select HAVE_ARCH_TRACEHOOK diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index e3fad2ef622c..918073c6681b 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -367,25 +367,6 @@ int ftrace_disable_ftrace_graph_caller(void) return ftrace_mod_jmp(ip, old_offset, new_offset); } -#else /* CONFIG_DYNAMIC_FTRACE */ - -/* - * These functions are picked from those used on - * this page for dynamic ftrace. They have been - * simplified to ignore all traces in NMI context. - */ -static atomic_t nmi_running; - -void arch_ftrace_nmi_enter(void) -{ - atomic_inc(&nmi_running); -} - -void arch_ftrace_nmi_exit(void) -{ - atomic_dec(&nmi_running); -} - #endif /* !CONFIG_DYNAMIC_FTRACE */ /* Add a function return address to the trace stack on thread info.*/ @@ -475,7 +456,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) &return_to_handler; /* Nmi's are currently unsupported */ - if (unlikely(atomic_read(&nmi_running))) + if (unlikely(in_nmi())) return; if (unlikely(atomic_read(¤t->tracing_graph_pause))) From a81bd80a0b0a405dc0483e2c428332d69da2c79f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 6 Feb 2009 01:45:16 -0500 Subject: [PATCH 112/413] ring-buffer: use generic version of in_nmi Impact: clean up Now that a generic in_nmi is available, this patch removes the special code in the ring_buffer and implements the in_nmi generic version instead. With this change, I was also able to rename the "arch_ftrace_nmi_enter" back to "ftrace_nmi_enter" and remove the code from the ring buffer. Signed-off-by: Steven Rostedt --- arch/x86/kernel/ftrace.c | 4 ++-- include/linux/ftrace_irq.h | 8 ------- kernel/trace/ring_buffer.c | 43 ++++++++++++-------------------------- 3 files changed, 15 insertions(+), 40 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 918073c6681b..d74d75e0952d 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -113,7 +113,7 @@ static void ftrace_mod_code(void) MCOUNT_INSN_SIZE); } -void arch_ftrace_nmi_enter(void) +void ftrace_nmi_enter(void) { atomic_inc(&nmi_running); /* Must have nmi_running seen before reading write flag */ @@ -124,7 +124,7 @@ void arch_ftrace_nmi_enter(void) } } -void arch_ftrace_nmi_exit(void) +void ftrace_nmi_exit(void) { /* Finish all executions before clearing nmi_running */ smp_wmb(); diff --git a/include/linux/ftrace_irq.h b/include/linux/ftrace_irq.h index 29de6779a963..dca7bf8cffe2 100644 --- a/include/linux/ftrace_irq.h +++ b/include/linux/ftrace_irq.h @@ -3,14 +3,6 @@ #ifdef CONFIG_FTRACE_NMI_ENTER -extern void arch_ftrace_nmi_enter(void); -extern void arch_ftrace_nmi_exit(void); -#else -static inline void arch_ftrace_nmi_enter(void) { } -static inline void arch_ftrace_nmi_exit(void) { } -#endif - -#ifdef CONFIG_RING_BUFFER extern void ftrace_nmi_enter(void); extern void ftrace_nmi_exit(void); #else diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index a60a6a852f42..5ee344417cd5 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -8,6 +8,7 @@ #include #include #include +#include #include #include #include @@ -19,35 +20,6 @@ #include "trace.h" -/* - * Since the write to the buffer is still not fully lockless, - * we must be careful with NMIs. The locks in the writers - * are taken when a write crosses to a new page. The locks - * protect against races with the readers (this will soon - * be fixed with a lockless solution). - * - * Because we can not protect against NMIs, and we want to - * keep traces reentrant, we need to manage what happens - * when we are in an NMI. - */ -static DEFINE_PER_CPU(int, rb_in_nmi); - -void ftrace_nmi_enter(void) -{ - __get_cpu_var(rb_in_nmi)++; - /* call arch specific handler too */ - arch_ftrace_nmi_enter(); -} - -void ftrace_nmi_exit(void) -{ - arch_ftrace_nmi_exit(); - __get_cpu_var(rb_in_nmi)--; - /* NMIs are not recursive */ - WARN_ON_ONCE(__get_cpu_var(rb_in_nmi)); -} - - /* * A fast way to enable or disable all ring buffers is to * call tracing_on or tracing_off. Turning off the ring buffers @@ -1027,12 +999,23 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, local_irq_save(flags); /* + * Since the write to the buffer is still not + * fully lockless, we must be careful with NMIs. + * The locks in the writers are taken when a write + * crosses to a new page. The locks protect against + * races with the readers (this will soon be fixed + * with a lockless solution). + * + * Because we can not protect against NMIs, and we + * want to keep traces reentrant, we need to manage + * what happens when we are in an NMI. + * * NMIs can happen after we take the lock. * If we are in an NMI, only take the lock * if it is not already taken. Otherwise * simply fail. */ - if (unlikely(__get_cpu_var(rb_in_nmi))) { + if (unlikely(in_nmi())) { if (!__raw_spin_trylock(&cpu_buffer->lock)) goto out_unlock; } else From 57794a9d48b63e34acbe63282628c9f029603308 Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Fri, 6 Feb 2009 17:33:27 +0800 Subject: [PATCH 113/413] trace: trivial fixes in comment typos. Impact: clean up Fixed several typos in the comments. Signed-off-by: Wenji Huang Signed-off-by: Steven Rostedt --- include/linux/ftrace.h | 2 +- kernel/trace/ftrace.c | 6 +++--- kernel/trace/trace.h | 6 +++--- 3 files changed, 7 insertions(+), 7 deletions(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 7840e718c6c7..5e302d636fc2 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -140,7 +140,7 @@ static inline int ftrace_disable_ftrace_graph_caller(void) { return 0; } #endif /** - * ftrace_make_nop - convert code into top + * ftrace_make_nop - convert code into nop * @mod: module structure if called by module load initialization * @rec: the mcount call site record * @addr: the address that the call site should be calling diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 68610031780b..1796e018fbff 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -465,7 +465,7 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) * it is not enabled then do nothing. * * If this record is not to be traced and - * it is enabled then disabled it. + * it is enabled then disable it. * */ if (rec->flags & FTRACE_FL_NOTRACE) { @@ -485,7 +485,7 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) if (fl == (FTRACE_FL_FILTER | FTRACE_FL_ENABLED)) return 0; - /* Record is not filtered and is not enabled do nothing */ + /* Record is not filtered or enabled, do nothing */ if (!fl) return 0; @@ -507,7 +507,7 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) } else { - /* if record is not enabled do nothing */ + /* if record is not enabled, do nothing */ if (!(rec->flags & FTRACE_FL_ENABLED)) return 0; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 5efc4c707f7e..f92aba52a894 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -616,12 +616,12 @@ extern struct tracer nop_trace; * preempt_enable (after a disable), a schedule might take place * causing an infinite recursion. * - * To prevent this, we read the need_recshed flag before + * To prevent this, we read the need_resched flag before * disabling preemption. When we want to enable preemption we * check the flag, if it is set, then we call preempt_enable_no_resched. * Otherwise, we call preempt_enable. * - * The rational for doing the above is that if need resched is set + * The rational for doing the above is that if need_resched is set * and we have yet to reschedule, we are either in an atomic location * (where we do not need to check for scheduling) or we are inside * the scheduler and do not want to resched. @@ -642,7 +642,7 @@ static inline int ftrace_preempt_disable(void) * * This is a scheduler safe way to enable preemption and not miss * any preemption checks. The disabled saved the state of preemption. - * If resched is set, then we were either inside an atomic or + * If resched is set, then we are either inside an atomic or * are inside the scheduler (we would have already scheduled * otherwise). In this case, we do not want to call normal * preempt_enable, but preempt_enable_no_resched instead. From 2db270a80b8f2238e536876cfb3987af02684df8 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 7 Feb 2009 20:46:45 +0100 Subject: [PATCH 114/413] tracing/blktrace: move the tracing file to kernel/trace Impact: cleanup Move blktrace.c to kernel/trace, also move its config entry. Signed-off-by: Frederic Weisbecker Acked-by: Arnaldo Carvalho de Melo Acked-by: Jens Axboe Signed-off-by: Ingo Molnar --- block/Kconfig | 24 ------------------------ block/Makefile | 1 - kernel/trace/Kconfig | 23 +++++++++++++++++++++++ kernel/trace/Makefile | 1 + {block => kernel/trace}/blktrace.c | 2 +- 5 files changed, 25 insertions(+), 26 deletions(-) rename {block => kernel/trace}/blktrace.c (99%) diff --git a/block/Kconfig b/block/Kconfig index 7cdaa1d72252..e7d12782bcfb 100644 --- a/block/Kconfig +++ b/block/Kconfig @@ -44,30 +44,6 @@ config LBD If unsure, say N. -config BLK_DEV_IO_TRACE - bool "Support for tracing block io actions" - depends on SYSFS - select RELAY - select DEBUG_FS - select TRACEPOINTS - select TRACING - select STACKTRACE - help - Say Y here if you want to be able to trace the block layer actions - on a given queue. Tracing allows you to see any traffic happening - on a block device queue. For more information (and the userspace - support tools needed), fetch the blktrace tools from: - - git://git.kernel.dk/blktrace.git - - Tracing also is possible using the ftrace interface, e.g.: - - echo 1 > /sys/block/sda/sda1/trace/enable - echo blk > /sys/kernel/debug/tracing/current_tracer - cat /sys/kernel/debug/tracing/trace_pipe - - If unsure, say N. - config BLK_DEV_BSG bool "Block layer SG support v4 (EXPERIMENTAL)" depends on EXPERIMENTAL diff --git a/block/Makefile b/block/Makefile index bfe73049f939..e9fa4dd690f2 100644 --- a/block/Makefile +++ b/block/Makefile @@ -13,6 +13,5 @@ obj-$(CONFIG_IOSCHED_AS) += as-iosched.o obj-$(CONFIG_IOSCHED_DEADLINE) += deadline-iosched.o obj-$(CONFIG_IOSCHED_CFQ) += cfq-iosched.o -obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o obj-$(CONFIG_BLOCK_COMPAT) += compat_ioctl.o obj-$(CONFIG_BLK_DEV_INTEGRITY) += blk-integrity.o diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 25131a5d5e4f..4fee43c01942 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -302,6 +302,29 @@ config WORKQUEUE_TRACER For example it can help a developer to decide whether he should choose a per cpu workqueue instead of a singlethreaded one. +config BLK_DEV_IO_TRACE + bool "Support for tracing block io actions" + depends on SYSFS + select RELAY + select DEBUG_FS + select TRACEPOINTS + select TRACING + select STACKTRACE + help + Say Y here if you want to be able to trace the block layer actions + on a given queue. Tracing allows you to see any traffic happening + on a block device queue. For more information (and the userspace + support tools needed), fetch the blktrace tools from: + + git://git.kernel.dk/blktrace.git + + Tracing also is possible using the ftrace interface, e.g.: + + echo 1 > /sys/block/sda/sda1/trace/enable + echo blk > /sys/kernel/debug/tracing/current_tracer + cat /sys/kernel/debug/tracing/trace_pipe + + If unsure, say N. config DYNAMIC_FTRACE bool "enable/disable ftrace tracepoints dynamically" diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index f76d48f3527d..627090bc262d 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -37,5 +37,6 @@ obj-$(CONFIG_HW_BRANCH_TRACER) += trace_hw_branches.o obj-$(CONFIG_POWER_TRACER) += trace_power.o obj-$(CONFIG_KMEMTRACE) += kmemtrace.o obj-$(CONFIG_WORKQUEUE_TRACER) += trace_workqueue.o +obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o libftrace-y := ftrace.o diff --git a/block/blktrace.c b/kernel/trace/blktrace.c similarity index 99% rename from block/blktrace.c rename to kernel/trace/blktrace.c index ca6d32061e4f..3b91da064820 100644 --- a/block/blktrace.c +++ b/kernel/trace/blktrace.c @@ -25,7 +25,7 @@ #include #include #include -#include <../kernel/trace/trace_output.h> +#include "trace_output.h" static unsigned int blktrace_seq __read_mostly = 1; From 7447dce96f2233d250bc39a4a10a42f7c3dd46fc Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 7 Feb 2009 21:33:57 +0100 Subject: [PATCH 115/413] tracing/function-graph-tracer: provide a selftest for the function graph tracer Making it more easy to do a basic regression test for this tracer. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace.h | 2 ++ kernel/trace/trace_functions_graph.c | 3 ++ kernel/trace/trace_selftest.c | 50 ++++++++++++++++++++++++++++ 3 files changed, 55 insertions(+) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index b9838f4a6929..a011ec062225 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -500,6 +500,8 @@ extern int DYN_FTRACE_TEST_NAME(void); #ifdef CONFIG_FTRACE_STARTUP_TEST extern int trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr); +extern int trace_selftest_startup_function_graph(struct tracer *trace, + struct trace_array *tr); extern int trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr); extern int trace_selftest_startup_preemptoff(struct tracer *trace, diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 222f97d336a6..88f8d9d80a93 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -750,6 +750,9 @@ static struct tracer graph_trace __read_mostly = { .print_line = print_graph_function, .print_header = print_graph_headers, .flags = &tracer_flags, +#ifdef CONFIG_FTRACE_SELFTEST + .selftest = trace_selftest_startup_function_graph, +#endif }; static __init int init_graph_trace(void) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 445700e51f6d..0c9aa1457e51 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -13,6 +13,8 @@ static inline int trace_valid_entry(struct trace_entry *entry) case TRACE_PRINT: case TRACE_SPECIAL: case TRACE_BRANCH: + case TRACE_GRAPH_ENT: + case TRACE_GRAPH_RET: return 1; } return 0; @@ -227,6 +229,54 @@ trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr) } #endif /* CONFIG_FUNCTION_TRACER */ + +#ifdef CONFIG_FUNCTION_GRAPH_TRACER +/* + * Pretty much the same than for the function tracer from which the selftest + * has been borrowed. + */ +int +trace_selftest_startup_function_graph(struct tracer *trace, + struct trace_array *tr) +{ + int ret; + unsigned long count; + + ret = tracer_init(trace, tr); + if (ret) { + warn_failed_init_tracer(trace, ret); + goto out; + } + + /* Sleep for a 1/10 of a second */ + msleep(100); + + tracing_stop(); + + /* check the trace buffer */ + ret = trace_test_buffer(tr, &count); + + trace->reset(tr); + tracing_start(); + + if (!ret && !count) { + printk(KERN_CONT ".. no entries found .."); + ret = -1; + goto out; + } + + /* Don't test dynamic tracing, the function tracer already did */ + +out: + /* Stop it if we failed */ + if (ret) + ftrace_graph_stop(); + + return ret; +} +#endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + + #ifdef CONFIG_IRQSOFF_TRACER int trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr) From 1292211058aaf872eeb2a0e2677d237916b4501f Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 7 Feb 2009 22:16:12 +0100 Subject: [PATCH 116/413] tracing/power: move the power trace headers to a dedicated file Impact: cleanup Move the power tracer headers to trace/power.h to keep ftrace.h and power bits more easy to maintain as separated topics. Signed-off-by: Frederic Weisbecker Cc: Arjan van de Ven Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c | 2 +- arch/x86/kernel/process.c | 2 +- include/linux/ftrace.h | 30 ------------------- include/trace/power.h | 35 ++++++++++++++++++++++ kernel/trace/trace.h | 1 + kernel/trace/trace_power.c | 2 +- 6 files changed, 39 insertions(+), 33 deletions(-) create mode 100644 include/trace/power.h diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 4b1c319d30c3..7ed925edf4d2 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -33,7 +33,7 @@ #include #include #include -#include +#include #include #include diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index e68bb9e30864..026819ffcb0c 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -8,7 +8,7 @@ #include #include #include -#include +#include #include #include diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 5e302d636fc2..106b7909d500 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -339,36 +339,6 @@ ftrace_init_module(struct module *mod, unsigned long *start, unsigned long *end) { } #endif -enum { - POWER_NONE = 0, - POWER_CSTATE = 1, - POWER_PSTATE = 2, -}; - -struct power_trace { -#ifdef CONFIG_POWER_TRACER - ktime_t stamp; - ktime_t end; - int type; - int state; -#endif -}; - -#ifdef CONFIG_POWER_TRACER -extern void trace_power_start(struct power_trace *it, unsigned int type, - unsigned int state); -extern void trace_power_mark(struct power_trace *it, unsigned int type, - unsigned int state); -extern void trace_power_end(struct power_trace *it); -#else -static inline void trace_power_start(struct power_trace *it, unsigned int type, - unsigned int state) { } -static inline void trace_power_mark(struct power_trace *it, unsigned int type, - unsigned int state) { } -static inline void trace_power_end(struct power_trace *it) { } -#endif - - /* * Structure that defines an entry function trace. */ diff --git a/include/trace/power.h b/include/trace/power.h new file mode 100644 index 000000000000..c7cefbcdaea4 --- /dev/null +++ b/include/trace/power.h @@ -0,0 +1,35 @@ +#ifndef _TRACE_POWER_H +#define _TRACE_POWER_H + +#include + +enum { + POWER_NONE = 0, + POWER_CSTATE = 1, + POWER_PSTATE = 2, +}; + +struct power_trace { +#ifdef CONFIG_POWER_TRACER + ktime_t stamp; + ktime_t end; + int type; + int state; +#endif +}; + +#ifdef CONFIG_POWER_TRACER +extern void trace_power_start(struct power_trace *it, unsigned int type, + unsigned int state); +extern void trace_power_mark(struct power_trace *it, unsigned int type, + unsigned int state); +extern void trace_power_end(struct power_trace *it); +#else +static inline void trace_power_start(struct power_trace *it, unsigned int type, + unsigned int state) { } +static inline void trace_power_mark(struct power_trace *it, unsigned int type, + unsigned int state) { } +static inline void trace_power_end(struct power_trace *it) { } +#endif + +#endif /* _TRACE_POWER_H */ diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index a011ec062225..1ecfb9d2b365 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -10,6 +10,7 @@ #include #include #include +#include enum trace_type { __TRACE_FIRST_TYPE = 0, diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c index bfc21f8079ab..b1d0d087d3a6 100644 --- a/kernel/trace/trace_power.c +++ b/kernel/trace/trace_power.c @@ -11,7 +11,7 @@ #include #include -#include +#include #include #include From 3861a17bcc0af815f684c6178bc9ec2d790c350e Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 8 Feb 2009 00:04:02 +0100 Subject: [PATCH 117/413] tracing/function-graph-tracer: drop the kernel_text_address check When the function graph tracer picks a return address, it ensures this address is really a kernel text one by calling __kernel_text_address() Actually this path has never been taken.Its role was more likely to debug the tracer on the beginning of its development but this function is wasteful since it is called for every traced function. The fault check is already sufficient. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 7 ------- kernel/extable.c | 4 ++-- kernel/module.c | 2 +- 3 files changed, 3 insertions(+), 10 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index d74d75e0952d..18828aee8781 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -491,13 +491,6 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) return; } - if (unlikely(!__kernel_text_address(old))) { - ftrace_graph_stop(); - *parent = old; - WARN_ON(1); - return; - } - calltime = cpu_clock(raw_smp_processor_id()); if (push_return_trace(old, calltime, diff --git a/kernel/extable.c b/kernel/extable.c index e136ed8d82ba..0df6253730be 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -41,7 +41,7 @@ const struct exception_table_entry *search_exception_tables(unsigned long addr) return e; } -__notrace_funcgraph int core_kernel_text(unsigned long addr) +int core_kernel_text(unsigned long addr) { if (addr >= (unsigned long)_stext && addr <= (unsigned long)_etext) @@ -54,7 +54,7 @@ __notrace_funcgraph int core_kernel_text(unsigned long addr) return 0; } -__notrace_funcgraph int __kernel_text_address(unsigned long addr) +int __kernel_text_address(unsigned long addr) { if (core_kernel_text(addr)) return 1; diff --git a/kernel/module.c b/kernel/module.c index ba22484a987e..22d7379709da 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2735,7 +2735,7 @@ int is_module_address(unsigned long addr) /* Is this a valid kernel address? */ -__notrace_funcgraph struct module *__module_text_address(unsigned long addr) +struct module *__module_text_address(unsigned long addr) { struct module *mod; From b5db03c4355e568f1567758287b30a6a262d5057 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sat, 7 Feb 2009 18:52:59 -0200 Subject: [PATCH 118/413] tracing: handle unregistering the current tracer Impact: simplification Instead of requiring that plugins have the sequence: my_tracer_stop(my_trace_array); unregister_tracer(my_tracer); it should be possible just do a: unregister_tracer(my_tracer); Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 03fbd4c20bc2..93040f1bef13 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -559,6 +559,15 @@ void unregister_tracer(struct tracer *type) found: *t = (*t)->next; + + if (type == current_trace && tracer_enabled) { + tracer_enabled = 0; + tracing_stop(); + if (current_trace->stop) + current_trace->stop(&global_trace); + current_trace = &nop_trace; + } + if (strlen(type->name) != max_tracer_type_len) goto out; From 1dfba05d0f1a9b4245bb242a7c17fe448811a520 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 9 Feb 2009 12:06:54 +0100 Subject: [PATCH 119/413] tracing/blktrace: move the tracing file to kernel/trace, fix Impact: build fix The BLK_DEV_IO_TRACE entry used to be in block/Kconfig - which file itself was dependent on CONFIG_BLOCK. But now the entry is in kernel/trace/Kconfig - which is present even on !CONFIG_BLOCK. So add a 'depends on BLOCK' to BLK_DEV_IO_TRACE. Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 4fee43c01942..3a331289457a 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -305,6 +305,7 @@ config WORKQUEUE_TRACER config BLK_DEV_IO_TRACE bool "Support for tracing block io actions" depends on SYSFS + depends on BLOCK select RELAY select DEBUG_FS select TRACEPOINTS From b91facc367366b3f71375f337eb5997ec9ab4e69 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 6 Feb 2009 18:30:44 +0100 Subject: [PATCH 120/413] tracing/function-graph-tracer: handle the leaf functions from trace_pipe When one cats the trace file, the leaf functions are printed without brackets: function(); whereas in the trace_pipe file we'll see the following: function() { } This is because the ring_buffer handling is not the same between those two files. On the trace file, when an entry is printed, the iterator advanced and then we can check the next entry. There is no iterator with trace_pipe, the current entry to print has been peeked and not consumed. So checking the next entry will still return the current one while we don't consume it. This patch introduces a new value for the output callbacks to ask the tracing core to not consume the current entry after printing it. We need it because we will have to consume the current entry ourself to check the next one. Now the trace_pipe is able to handle well the leaf functions. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 4 +-- kernel/trace/trace.h | 7 ++-- kernel/trace/trace_functions_graph.c | 48 +++++++++++++++++----------- 3 files changed, 36 insertions(+), 23 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 93040f1bef13..5b1e9a9e9906 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2468,8 +2468,8 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, iter->seq.len = len; break; } - - trace_consume(iter); + if (ret != TRACE_TYPE_NO_CONSUME) + trace_consume(iter); if (iter->seq.len >= cnt) break; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 1ecfb9d2b365..7b0518adf6d7 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -63,13 +63,13 @@ struct ftrace_entry { /* Function call entry */ struct ftrace_graph_ent_entry { - struct trace_entry ent; + struct trace_entry ent; struct ftrace_graph_ent graph_ent; }; /* Function return entry */ struct ftrace_graph_ret_entry { - struct trace_entry ent; + struct trace_entry ent; struct ftrace_graph_ret ret; }; extern struct tracer boot_tracer; @@ -309,7 +309,8 @@ extern void __ftrace_bad_type(void); enum print_line_t { TRACE_TYPE_PARTIAL_LINE = 0, /* Retry after flushing the seq */ TRACE_TYPE_HANDLED = 1, - TRACE_TYPE_UNHANDLED = 2 /* Relay to other output functions */ + TRACE_TYPE_UNHANDLED = 2, /* Relay to other output functions */ + TRACE_TYPE_NO_CONSUME = 3 /* Handled but ask to not consume */ }; diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 88f8d9d80a93..782ec0fdf453 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -212,8 +212,8 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, pid_t *last_pids_cpu) return ret; } -static bool -trace_branch_is_leaf(struct trace_iterator *iter, +static struct ftrace_graph_ret_entry * +get_return_for_leaf(struct trace_iterator *iter, struct ftrace_graph_ent_entry *curr) { struct ring_buffer_iter *ring_iter; @@ -222,24 +222,33 @@ trace_branch_is_leaf(struct trace_iterator *iter, ring_iter = iter->buffer_iter[iter->cpu]; - if (!ring_iter) - return false; - - event = ring_buffer_iter_peek(ring_iter, NULL); + /* First peek to compare current entry and the next one */ + if (ring_iter) + event = ring_buffer_iter_peek(ring_iter, NULL); + else { + /* We need to consume the current entry to see the next one */ + ring_buffer_consume(iter->tr->buffer, iter->cpu, NULL); + event = ring_buffer_peek(iter->tr->buffer, iter->cpu, + NULL); + } if (!event) - return false; + return NULL; next = ring_buffer_event_data(event); if (next->ent.type != TRACE_GRAPH_RET) - return false; + return NULL; if (curr->ent.pid != next->ent.pid || curr->graph_ent.func != next->ret.func) - return false; + return NULL; - return true; + /* this is a leaf, now advance the iterator */ + if (ring_iter) + ring_buffer_read(ring_iter, NULL); + + return next; } /* Signal a overhead of time execution to the output */ @@ -376,18 +385,15 @@ static int print_graph_abs_time(u64 t, struct trace_seq *s) /* Case of a leaf function on its call entry */ static enum print_line_t print_graph_entry_leaf(struct trace_iterator *iter, - struct ftrace_graph_ent_entry *entry, struct trace_seq *s) + struct ftrace_graph_ent_entry *entry, + struct ftrace_graph_ret_entry *ret_entry, struct trace_seq *s) { - struct ftrace_graph_ret_entry *ret_entry; struct ftrace_graph_ret *graph_ret; - struct ring_buffer_event *event; struct ftrace_graph_ent *call; unsigned long long duration; int ret; int i; - event = ring_buffer_read(iter->buffer_iter[iter->cpu], NULL); - ret_entry = ring_buffer_event_data(event); graph_ret = &ret_entry->ret; call = &entry->graph_ent; duration = graph_ret->rettime - graph_ret->calltime; @@ -457,7 +463,11 @@ print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, if (!ret) return TRACE_TYPE_PARTIAL_LINE; - return TRACE_TYPE_HANDLED; + /* + * we already consumed the current entry to check the next one + * and see if this is a leaf. + */ + return TRACE_TYPE_NO_CONSUME; } static enum print_line_t @@ -469,6 +479,7 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, pid_t *last_entry = iter->private; struct trace_entry *ent = iter->ent; struct ftrace_graph_ent *call = &field->graph_ent; + struct ftrace_graph_ret_entry *leaf_ret; /* Pid */ if (verif_pid(s, ent->pid, cpu, last_entry) == TRACE_TYPE_PARTIAL_LINE) @@ -504,8 +515,9 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } - if (trace_branch_is_leaf(iter, field)) - return print_graph_entry_leaf(iter, field, s); + leaf_ret = get_return_for_leaf(iter, field); + if (leaf_ret) + return print_graph_entry_leaf(iter, field, leaf_ret, s); else return print_graph_entry_nested(field, s, iter->ent->pid, cpu); From 3c56819b14b00dd449bd776303e61f8532fad09f Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Mon, 9 Feb 2009 08:15:56 +0200 Subject: [PATCH 121/413] tracing: splice support for tracing_pipe Added and implemented tracing_pipe_fops->splice_read(). This allows userspace programs to get tracing data more efficiently. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 136 +++++++++++++++++++++++++++++++++++++++++++ kernel/trace/trace.h | 6 ++ 2 files changed, 142 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 5b1e9a9e9906..9e29fdb0dfe5 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include @@ -364,6 +365,25 @@ ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) return cnt; } +ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt) +{ + int len; + void *ret; + + if (s->len <= s->readpos) + return -EBUSY; + + len = s->len - s->readpos; + if (cnt > len) + cnt = len; + ret = memcpy(buf, s->buffer + s->readpos, cnt); + if (!ret) + return -EFAULT; + + s->readpos += len; + return cnt; +} + static void trace_print_seq(struct seq_file *m, struct trace_seq *s) { @@ -2493,6 +2513,121 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, return sret; } +static void tracing_pipe_buf_release(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + __free_page(buf->page); +} + +static void tracing_spd_release_pipe(struct splice_pipe_desc *spd, + unsigned int idx) +{ + __free_page(spd->pages[idx]); +} + +static struct pipe_buf_operations tracing_pipe_buf_ops = { + .can_merge = 0, + .map = generic_pipe_buf_map, + .unmap = generic_pipe_buf_unmap, + .confirm = generic_pipe_buf_confirm, + .release = tracing_pipe_buf_release, + .steal = generic_pipe_buf_steal, + .get = generic_pipe_buf_get, +}; + +static ssize_t tracing_splice_read_pipe(struct file *filp, + loff_t *ppos, + struct pipe_inode_info *pipe, + size_t len, + unsigned int flags) +{ + struct page *pages[PIPE_BUFFERS]; + struct partial_page partial[PIPE_BUFFERS]; + struct trace_iterator *iter = filp->private_data; + struct splice_pipe_desc spd = { + .pages = pages, + .partial = partial, + .nr_pages = 0, /* This gets updated below. */ + .flags = flags, + .ops = &tracing_pipe_buf_ops, + .spd_release = tracing_spd_release_pipe, + }; + ssize_t ret; + size_t count, rem; + unsigned int i; + + mutex_lock(&trace_types_lock); + + if (iter->trace->splice_read) { + ret = iter->trace->splice_read(iter, filp, + ppos, pipe, len, flags); + if (ret) + goto out; + } + + ret = tracing_wait_pipe(filp); + if (ret <= 0) + goto out; + + if (!iter->ent && !find_next_entry_inc(iter)) { + ret = -EFAULT; + goto out; + } + + /* Fill as many pages as possible. */ + for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) { + pages[i] = alloc_page(GFP_KERNEL); + + /* Seq buffer is page-sized, exactly what we need. */ + for (;;) { + count = iter->seq.len; + ret = print_trace_line(iter); + count = iter->seq.len - count; + if (rem < count) { + rem = 0; + iter->seq.len -= count; + break; + } + if (ret == TRACE_TYPE_PARTIAL_LINE) { + iter->seq.len -= count; + break; + } + + trace_consume(iter); + rem -= count; + if (!find_next_entry_inc(iter)) { + rem = 0; + iter->ent = NULL; + break; + } + } + + /* Copy the data into the page, so we can start over. */ + ret = trace_seq_to_buffer(&iter->seq, + page_address(pages[i]), + iter->seq.len); + if (ret < 0) { + __free_page(pages[i]); + break; + } + partial[i].offset = 0; + partial[i].len = iter->seq.len; + + trace_seq_reset(&iter->seq); + } + + mutex_unlock(&trace_types_lock); + + spd.nr_pages = i; + + return splice_to_pipe(pipe, &spd); + +out: + mutex_unlock(&trace_types_lock); + + return ret; +} + static ssize_t tracing_entries_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -2656,6 +2791,7 @@ static struct file_operations tracing_pipe_fops = { .open = tracing_open_pipe, .poll = tracing_poll_pipe, .read = tracing_read_pipe, + .splice_read = tracing_splice_read_pipe, .release = tracing_release_pipe, }; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 7b0518adf6d7..dbff0207b213 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -353,6 +353,12 @@ struct tracer { ssize_t (*read)(struct trace_iterator *iter, struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos); + ssize_t (*splice_read)(struct trace_iterator *iter, + struct file *filp, + loff_t *ppos, + struct pipe_inode_info *pipe, + size_t len, + unsigned int flags); #ifdef CONFIG_FTRACE_STARTUP_TEST int (*selftest)(struct tracer *trace, struct trace_array *tr); From ff98781bab2735e6c89793034173e0cb5007a7e5 Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Mon, 9 Feb 2009 08:15:55 +0200 Subject: [PATCH 122/413] tracing: Move pipe waiting code out of tracing_read_pipe(). This moves the pipe waiting code from tracing_read_pipe() into tracing_wait_pipe(), which is useful to implement other fops, like splice_read. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 115 ++++++++++++++++++++++++------------------- 1 file changed, 63 insertions(+), 52 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 9e29fdb0dfe5..11fde0ad9cb6 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2388,6 +2388,63 @@ tracing_poll_pipe(struct file *filp, poll_table *poll_table) } } +/* Must be called with trace_types_lock mutex held. */ +static int tracing_wait_pipe(struct file *filp) +{ + struct trace_iterator *iter = filp->private_data; + + while (trace_empty(iter)) { + + if ((filp->f_flags & O_NONBLOCK)) { + return -EAGAIN; + } + + /* + * This is a make-shift waitqueue. The reason we don't use + * an actual wait queue is because: + * 1) we only ever have one waiter + * 2) the tracing, traces all functions, we don't want + * the overhead of calling wake_up and friends + * (and tracing them too) + * Anyway, this is really very primitive wakeup. + */ + set_current_state(TASK_INTERRUPTIBLE); + iter->tr->waiter = current; + + mutex_unlock(&trace_types_lock); + + /* sleep for 100 msecs, and try again. */ + schedule_timeout(HZ/10); + + mutex_lock(&trace_types_lock); + + iter->tr->waiter = NULL; + + if (signal_pending(current)) { + return -EINTR; + } + + if (iter->trace != current_trace) + return 0; + + /* + * We block until we read something and tracing is disabled. + * We still block if tracing is disabled, but we have never + * read anything. This allows a user to cat this file, and + * then enable tracing. But after we have read something, + * we give an EOF when tracing is again disabled. + * + * iter->pos will be 0 if we haven't read anything. + */ + if (!tracer_enabled && iter->pos) + break; + + continue; + } + + return 1; +} + /* * Consumer reader. */ @@ -2413,61 +2470,15 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, } waitagain: - sret = 0; - while (trace_empty(iter)) { - - if ((filp->f_flags & O_NONBLOCK)) { - sret = -EAGAIN; - goto out; - } - - /* - * This is a make-shift waitqueue. The reason we don't use - * an actual wait queue is because: - * 1) we only ever have one waiter - * 2) the tracing, traces all functions, we don't want - * the overhead of calling wake_up and friends - * (and tracing them too) - * Anyway, this is really very primitive wakeup. - */ - set_current_state(TASK_INTERRUPTIBLE); - iter->tr->waiter = current; - - mutex_unlock(&trace_types_lock); - - /* sleep for 100 msecs, and try again. */ - schedule_timeout(HZ/10); - - mutex_lock(&trace_types_lock); - - iter->tr->waiter = NULL; - - if (signal_pending(current)) { - sret = -EINTR; - goto out; - } - - if (iter->trace != current_trace) - goto out; - - /* - * We block until we read something and tracing is disabled. - * We still block if tracing is disabled, but we have never - * read anything. This allows a user to cat this file, and - * then enable tracing. But after we have read something, - * we give an EOF when tracing is again disabled. - * - * iter->pos will be 0 if we haven't read anything. - */ - if (!tracer_enabled && iter->pos) - break; - - continue; - } + sret = tracing_wait_pipe(filp); + if (sret <= 0) + goto out; /* stop when tracing is finished */ - if (trace_empty(iter)) + if (trace_empty(iter)) { + sret = 0; goto out; + } if (cnt >= PAGE_SIZE) cnt = PAGE_SIZE - 1; From 34cd4998d38f9bd04f34b78a7cb0c7f1bee00bd9 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 9 Feb 2009 12:06:29 -0500 Subject: [PATCH 123/413] tracing: clean up splice code Ingo Molnar suggested a series of clean ups for the splice code. This patch implements those suggestions. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 96 +++++++++++++++++++++++++------------------- 1 file changed, 55 insertions(+), 41 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 11fde0ad9cb6..d89821283b47 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2537,15 +2537,49 @@ static void tracing_spd_release_pipe(struct splice_pipe_desc *spd, } static struct pipe_buf_operations tracing_pipe_buf_ops = { - .can_merge = 0, - .map = generic_pipe_buf_map, - .unmap = generic_pipe_buf_unmap, - .confirm = generic_pipe_buf_confirm, - .release = tracing_pipe_buf_release, - .steal = generic_pipe_buf_steal, - .get = generic_pipe_buf_get, + .can_merge = 0, + .map = generic_pipe_buf_map, + .unmap = generic_pipe_buf_unmap, + .confirm = generic_pipe_buf_confirm, + .release = tracing_pipe_buf_release, + .steal = generic_pipe_buf_steal, + .get = generic_pipe_buf_get, }; +static size_t +tracing_fill_pipe_page(struct page *pages, size_t rem, + struct trace_iterator *iter) +{ + size_t count; + int ret; + + /* Seq buffer is page-sized, exactly what we need. */ + for (;;) { + count = iter->seq.len; + ret = print_trace_line(iter); + count = iter->seq.len - count; + if (rem < count) { + rem = 0; + iter->seq.len -= count; + break; + } + if (ret == TRACE_TYPE_PARTIAL_LINE) { + iter->seq.len -= count; + break; + } + + trace_consume(iter); + rem -= count; + if (!find_next_entry_inc(iter)) { + rem = 0; + iter->ent = NULL; + break; + } + } + + return rem; +} + static ssize_t tracing_splice_read_pipe(struct file *filp, loff_t *ppos, struct pipe_inode_info *pipe, @@ -2556,15 +2590,15 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, struct partial_page partial[PIPE_BUFFERS]; struct trace_iterator *iter = filp->private_data; struct splice_pipe_desc spd = { - .pages = pages, - .partial = partial, - .nr_pages = 0, /* This gets updated below. */ - .flags = flags, - .ops = &tracing_pipe_buf_ops, - .spd_release = tracing_spd_release_pipe, + .pages = pages, + .partial = partial, + .nr_pages = 0, /* This gets updated below. */ + .flags = flags, + .ops = &tracing_pipe_buf_ops, + .spd_release = tracing_spd_release_pipe, }; ssize_t ret; - size_t count, rem; + size_t rem; unsigned int i; mutex_lock(&trace_types_lock); @@ -2573,45 +2607,25 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, ret = iter->trace->splice_read(iter, filp, ppos, pipe, len, flags); if (ret) - goto out; + goto out_err; } ret = tracing_wait_pipe(filp); if (ret <= 0) - goto out; + goto out_err; if (!iter->ent && !find_next_entry_inc(iter)) { ret = -EFAULT; - goto out; + goto out_err; } /* Fill as many pages as possible. */ for (i = 0, rem = len; i < PIPE_BUFFERS && rem; i++) { pages[i] = alloc_page(GFP_KERNEL); + if (!pages[i]) + break; - /* Seq buffer is page-sized, exactly what we need. */ - for (;;) { - count = iter->seq.len; - ret = print_trace_line(iter); - count = iter->seq.len - count; - if (rem < count) { - rem = 0; - iter->seq.len -= count; - break; - } - if (ret == TRACE_TYPE_PARTIAL_LINE) { - iter->seq.len -= count; - break; - } - - trace_consume(iter); - rem -= count; - if (!find_next_entry_inc(iter)) { - rem = 0; - iter->ent = NULL; - break; - } - } + rem = tracing_fill_pipe_page(pages[i], rem, iter); /* Copy the data into the page, so we can start over. */ ret = trace_seq_to_buffer(&iter->seq, @@ -2633,7 +2647,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, return splice_to_pipe(pipe, &spd); -out: +out_err: mutex_unlock(&trace_types_lock); return ret; From b85fa01ed958ca59523a2db3c2ee647b98745d6a Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 9 Feb 2009 14:21:14 +0800 Subject: [PATCH 124/413] ring_buffer: fix typing mistake Impact: Fix bug I found several very very curious line. It's so curious that it may be brought by typing mistake. When (cpu_buffer->reader_page == cpu_buffer->commit_page): 1) We haven't copied it for bpage is changed: bpage = cpu_buffer->reader_page->page; memcpy(bpage->data, cpu_buffer->reader_page->page->data + read ... ) 2) We need update cpu_buffer->reader_page->read, but "cpu_buffer->reader_page += read;" is not right. [ This bug was a typo. The commit->reader_page is a page pointer and not an index into the page. The line should have been commit->reader_page->read += read. The other changes by Lai are nice clean ups to the code. - SDR ] Signed-off-by: Lai Jiangshan Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 53ba3a6d16d0..eca282720838 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2406,7 +2406,7 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * to swap with a page in the ring buffer. * * for example: - * rpage = ring_buffer_alloc_page(buffer); + * rpage = ring_buffer_alloc_read_page(buffer); * if (!rpage) * return error; * ret = ring_buffer_read_page(buffer, &rpage, cpu, 0); @@ -2461,18 +2461,17 @@ int ring_buffer_read_page(struct ring_buffer *buffer, */ if (cpu_buffer->reader_page == cpu_buffer->commit_page) { unsigned int read = cpu_buffer->reader_page->read; + unsigned int commit = rb_page_commit(cpu_buffer->reader_page); if (full) goto out; /* The writer is still on the reader page, we must copy */ - bpage = cpu_buffer->reader_page->page; memcpy(bpage->data, cpu_buffer->reader_page->page->data + read, - local_read(&bpage->commit) - read); + commit - read); /* consume what was read */ - cpu_buffer->reader_page += read; - + cpu_buffer->reader_page->read = commit; } else { /* swap the pages */ rb_init_page(bpage); From 667d24125839b6f3363d8177d7ed9fab8a40e45f Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Mon, 9 Feb 2009 14:21:17 +0800 Subject: [PATCH 125/413] ring_buffer: fix ring_buffer_read_page() Impact: change API and init bpage when copy ring_buffer_read_page()/rb_remove_entries() may be called for a partially consumed page. Add a parameter for rb_remove_entries() and make it update cpu_buffer->entries correctly for partially consumed pages. ring_buffer_read_page() now returns the offset to the next event. Init the bpage's time_stamp when return value is 0. Signed-off-by: Lai Jiangshan Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 33 ++++++++++++++++++++------------- 1 file changed, 20 insertions(+), 13 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index eca282720838..10d202ea06f3 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2332,13 +2332,14 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu); static void rb_remove_entries(struct ring_buffer_per_cpu *cpu_buffer, - struct buffer_data_page *bpage) + struct buffer_data_page *bpage, + unsigned int offset) { struct ring_buffer_event *event; unsigned long head; __raw_spin_lock(&cpu_buffer->lock); - for (head = 0; head < local_read(&bpage->commit); + for (head = offset; head < local_read(&bpage->commit); head += rb_event_length(event)) { event = __rb_data_page_index(bpage, head); @@ -2410,8 +2411,8 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * if (!rpage) * return error; * ret = ring_buffer_read_page(buffer, &rpage, cpu, 0); - * if (ret) - * process_page(rpage); + * if (ret >= 0) + * process_page(rpage, ret); * * When @full is set, the function will not return true unless * the writer is off the reader page. @@ -2422,8 +2423,8 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * responsible for that. * * Returns: - * 1 if data has been transferred - * 0 if no data has been transferred. + * >=0 if data has been transferred, returns the offset of consumed data. + * <0 if no data has been transferred. */ int ring_buffer_read_page(struct ring_buffer *buffer, void **data_page, int cpu, int full) @@ -2432,7 +2433,8 @@ int ring_buffer_read_page(struct ring_buffer *buffer, struct ring_buffer_event *event; struct buffer_data_page *bpage; unsigned long flags; - int ret = 0; + unsigned int read; + int ret = -1; if (!data_page) return 0; @@ -2454,24 +2456,29 @@ int ring_buffer_read_page(struct ring_buffer *buffer, /* check for data */ if (!local_read(&cpu_buffer->reader_page->page->commit)) goto out; + + read = cpu_buffer->reader_page->read; /* * If the writer is already off of the read page, then simply * switch the read page with the given page. Otherwise * we need to copy the data from the reader to the writer. */ if (cpu_buffer->reader_page == cpu_buffer->commit_page) { - unsigned int read = cpu_buffer->reader_page->read; unsigned int commit = rb_page_commit(cpu_buffer->reader_page); + struct buffer_data_page *rpage = cpu_buffer->reader_page->page; if (full) goto out; /* The writer is still on the reader page, we must copy */ - memcpy(bpage->data, - cpu_buffer->reader_page->page->data + read, - commit - read); + memcpy(bpage->data + read, rpage->data + read, commit - read); /* consume what was read */ cpu_buffer->reader_page->read = commit; + + /* update bpage */ + local_set(&bpage->commit, commit); + if (!read) + bpage->time_stamp = rpage->time_stamp; } else { /* swap the pages */ rb_init_page(bpage); @@ -2480,10 +2487,10 @@ int ring_buffer_read_page(struct ring_buffer *buffer, cpu_buffer->reader_page->read = 0; *data_page = bpage; } - ret = 1; + ret = read; /* update the entry counter */ - rb_remove_entries(cpu_buffer, bpage); + rb_remove_entries(cpu_buffer, bpage, read); out: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); From 966657883fdc3a2883a5e641ca4ec8f79ffb8ecd Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Feb 2009 11:53:23 -0500 Subject: [PATCH 126/413] tracing, x86: fix constraint for parent variable The constraint used for retrieving and restoring the parent function pointer is incorrect. The parent variable is a pointer, and the address of the pointer is modified by the asm statement and not the pointer itself. It is incorrect to pass it in as an output constraint since the asm will never update the pointer. Signed-off-by: Steven Rostedt --- arch/x86/kernel/ftrace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 18828aee8781..370bafaa43a3 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -468,8 +468,8 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) * ignore such a protection. */ asm volatile( - "1: " _ASM_MOV " (%[parent_old]), %[old]\n" - "2: " _ASM_MOV " %[return_hooker], (%[parent_replaced])\n" + "1: " _ASM_MOV " (%[parent]), %[old]\n" + "2: " _ASM_MOV " %[return_hooker], (%[parent])\n" " movl $0, %[faulted]\n" ".section .fixup, \"ax\"\n" @@ -479,9 +479,8 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) _ASM_EXTABLE(1b, 3b) _ASM_EXTABLE(2b, 3b) - : [parent_replaced] "=r" (parent), [old] "=r" (old), - [faulted] "=r" (faulted) - : [parent_old] "0" (parent), [return_hooker] "r" (return_hooker) + : [old] "=r" (old), [faulted] "=r" (faulted) + : [parent] "r" (parent), [return_hooker] "r" (return_hooker) : "memory" ); From 4543ae7ce1cb8b5ff27a59009e7991ea63791a71 Mon Sep 17 00:00:00 2001 From: Tobias Klauser Date: Mon, 9 Feb 2009 23:09:32 +0100 Subject: [PATCH 127/413] tracing: storage class should be before const qualifier The C99 specification states in section 6.11.5: The placement of a storage-class specifier other than at the beginning of the declaration specifiers in a declaration is an obsolescent feature. Signed-off-by: Tobias Klauser Signed-off-by: Steven Rostedt --- kernel/trace/trace_sysprof.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c index 84ca9d81e74d..4e2de4de1d65 100644 --- a/kernel/trace/trace_sysprof.c +++ b/kernel/trace/trace_sysprof.c @@ -88,7 +88,7 @@ static void backtrace_address(void *data, unsigned long addr, int reliable) } } -const static struct stacktrace_ops backtrace_ops = { +static const struct stacktrace_ops backtrace_ops = { .warning = backtrace_warning, .warning_symbol = backtrace_warning_symbol, .stack = backtrace_stack, From f54fc98aa656f334c1571df6e3ca9178ea223847 Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Tue, 10 Feb 2009 01:02:46 -0500 Subject: [PATCH 128/413] tracing: remove unneeded variable Impact: clean up. Remove the unnecessary variable ret. Signed-off-by: Wenji Huang Signed-off-by: Steven Rostedt --- kernel/trace/trace_branch.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index f8ae2c50e01d..c2e68d440c4d 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -91,8 +91,6 @@ void trace_likely_condition(struct ftrace_branch_data *f, int val, int expect) int enable_branch_tracing(struct trace_array *tr) { - int ret = 0; - mutex_lock(&branch_tracing_mutex); branch_tracer = tr; /* @@ -103,7 +101,7 @@ int enable_branch_tracing(struct trace_array *tr) branch_tracing_enabled++; mutex_unlock(&branch_tracing_mutex); - return ret; + return 0; } void disable_branch_tracing(void) From 810dc73265cd690b2bc6010489b4317bba2cda39 Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Tue, 10 Feb 2009 01:03:05 -0500 Subject: [PATCH 129/413] tracing: provide correct return value after outputting the event This patch is to make the function return early on failure, and give correct return value on success. Signed-off-by: Wenji Huang Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions_graph.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 782ec0fdf453..519a0cab1530 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -186,30 +186,30 @@ verif_pid(struct trace_seq *s, pid_t pid, int cpu, pid_t *last_pids_cpu) ret = trace_seq_printf(s, " ------------------------------------------\n"); if (!ret) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = print_graph_cpu(s, cpu); if (ret == TRACE_TYPE_PARTIAL_LINE) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = print_graph_proc(s, prev_pid); if (ret == TRACE_TYPE_PARTIAL_LINE) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = trace_seq_printf(s, " => "); if (!ret) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = print_graph_proc(s, pid); if (ret == TRACE_TYPE_PARTIAL_LINE) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; ret = trace_seq_printf(s, "\n ------------------------------------------\n\n"); if (!ret) - TRACE_TYPE_PARTIAL_LINE; + return TRACE_TYPE_PARTIAL_LINE; - return ret; + return TRACE_TYPE_HANDLED; } static struct ftrace_graph_ret_entry * From c3706f005c3aaf570e71f0f083fdbb59a5a9fa2e Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Tue, 10 Feb 2009 01:03:18 -0500 Subject: [PATCH 130/413] tracing: fix typos in comments Impact: clean up. Fix typos in the comments. Signed-off-by: Wenji Huang Signed-off-by: Steven Rostedt --- include/linux/ring_buffer.h | 2 +- kernel/trace/ring_buffer.c | 8 ++++---- kernel/trace/trace.c | 2 +- kernel/trace/trace_hw_branches.c | 2 +- 4 files changed, 7 insertions(+), 7 deletions(-) diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index 3c103d636da3..8e6646a54acf 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -8,7 +8,7 @@ struct ring_buffer; struct ring_buffer_iter; /* - * Don't reference this struct directly, use functions below. + * Don't refer to this struct directly, use functions below. */ struct ring_buffer_event { u32 type:2, len:3, time_delta:27; diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 10d202ea06f3..fa64e1f003eb 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -91,7 +91,7 @@ EXPORT_SYMBOL_GPL(tracing_off); * tracing_off_permanent - permanently disable ring buffers * * This function, once called, will disable all ring buffers - * permanenty. + * permanently. */ void tracing_off_permanent(void) { @@ -210,7 +210,7 @@ EXPORT_SYMBOL_GPL(ring_buffer_event_data); struct buffer_data_page { u64 time_stamp; /* page time stamp */ - local_t commit; /* write commited index */ + local_t commit; /* write committed index */ unsigned char data[]; /* data of buffer page */ }; @@ -260,7 +260,7 @@ struct ring_buffer_per_cpu { struct list_head pages; struct buffer_page *head_page; /* read from head */ struct buffer_page *tail_page; /* write to tail */ - struct buffer_page *commit_page; /* commited pages */ + struct buffer_page *commit_page; /* committed pages */ struct buffer_page *reader_page; unsigned long overrun; unsigned long entries; @@ -303,7 +303,7 @@ struct ring_buffer_iter { * check_pages - integrity check of buffer pages * @cpu_buffer: CPU buffer with pages to test * - * As a safty measure we check to make sure the data pages have not + * As a safety measure we check to make sure the data pages have not * been corrupted. */ static int rb_check_pages(struct ring_buffer_per_cpu *cpu_buffer) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index d89821283b47..d7c175a442df 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1963,7 +1963,7 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, struct tracer_opt *trace_opts = current_trace->flags->opts; - /* calulate max size */ + /* calculate max size */ for (i = 0; trace_options[i]; i++) { len += strlen(trace_options[i]); len += 3; /* "no" and space */ diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index e3e7db61c067..0794dd33f27b 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -75,7 +75,7 @@ static void bts_trace_start(struct trace_array *tr) } /* - * Start tracing on the current cpu. + * Stop tracing on the current cpu. * The argument is ignored. * * pre: bts_tracer_mutex must be locked. From 4fd2735881bf4d8bf5e30979f31fc2f1b1d505fa Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Tue, 10 Feb 2009 19:44:12 +0100 Subject: [PATCH 131/413] tracing: fix sparse warnings: make symbols static Impact: make global variables and a global function static The function '__trace_userstack' does not seem to have a caller, so it is commented out. Fix this sparse warnings: kernel/trace/trace.c:82:5: warning: symbol 'tracing_disabled' was not declared. Should it be static? kernel/trace/trace.c:600:10: warning: symbol 'trace_record_cmdline_disabled' was not declared. Should it be static? kernel/trace/trace.c:957:6: warning: symbol '__trace_userstack' was not declared. Should it be static? kernel/trace/trace.c:1694:5: warning: symbol 'tracing_release' was not declared. Should it be static? Signed-off-by: Hannes Eder Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index d7c175a442df..c157ba70f30c 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -80,7 +80,7 @@ static int dummy_set_flag(u32 old_flags, u32 bit, int set) * of the tracer is successful. But that is the only place that sets * this back to zero. */ -int tracing_disabled = 1; +static int tracing_disabled = 1; static DEFINE_PER_CPU(local_t, ftrace_cpu_disabled); @@ -626,7 +626,7 @@ static int cmdline_idx; static DEFINE_SPINLOCK(trace_cmdline_lock); /* temporary disable recording */ -atomic_t trace_record_cmdline_disabled __read_mostly; +static atomic_t trace_record_cmdline_disabled __read_mostly; static void trace_init_cmdlines(void) { @@ -983,10 +983,12 @@ static void ftrace_trace_userstack(struct trace_array *tr, #endif } -void __trace_userstack(struct trace_array *tr, unsigned long flags) +#ifdef UNUSED +static void __trace_userstack(struct trace_array *tr, unsigned long flags) { ftrace_trace_userstack(tr, flags, preempt_count()); } +#endif /* UNUSED */ static void ftrace_trace_special(void *__tr, @@ -1720,7 +1722,7 @@ int tracing_open_generic(struct inode *inode, struct file *filp) return 0; } -int tracing_release(struct inode *inode, struct file *file) +static int tracing_release(struct inode *inode, struct file *file) { struct seq_file *m = (struct seq_file *)file->private_data; struct trace_iterator *iter = m->private; From 5e39841c45cf5e6ea930ede1b0303309e03037a2 Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Tue, 10 Feb 2009 19:44:34 +0100 Subject: [PATCH 132/413] tracing: fix sparse warnings: fix (un-)signedness Fix these sparse warnings: kernel/trace/ring_buffer.c:70:37: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:84:39: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:96:43: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:2475:13: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:2475:13: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:2478:42: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:2478:42: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:2500:40: warning: incorrect type in argument 3 (different signedness) kernel/trace/ring_buffer.c:2505:44: warning: incorrect type in argument 2 (different signedness) kernel/trace/ring_buffer.c:2507:46: warning: incorrect type in argument 2 (different signedness) kernel/trace/trace.c:2130:40: warning: incorrect type in argument 3 (different signedness) kernel/trace/trace.c:2280:40: warning: incorrect type in argument 3 (different signedness) Signed-off-by: Hannes Eder Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 8 ++++---- kernel/trace/trace.c | 6 +++--- 2 files changed, 7 insertions(+), 7 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index fa64e1f003eb..dc18b5b9ccb4 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -59,7 +59,7 @@ enum { RB_BUFFERS_DISABLED = 1 << RB_BUFFERS_DISABLED_BIT, }; -static long ring_buffer_flags __read_mostly = RB_BUFFERS_ON; +static unsigned long ring_buffer_flags __read_mostly = RB_BUFFERS_ON; /** * tracing_on - enable all tracing buffers @@ -2501,7 +2501,7 @@ static ssize_t rb_simple_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - long *p = filp->private_data; + unsigned long *p = filp->private_data; char buf[64]; int r; @@ -2517,9 +2517,9 @@ static ssize_t rb_simple_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { - long *p = filp->private_data; + unsigned long *p = filp->private_data; char buf[64]; - long val; + unsigned long val; int ret; if (cnt >= sizeof(buf)) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c157ba70f30c..e2434990277c 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2147,7 +2147,7 @@ tracing_ctrl_write(struct file *filp, const char __user *ubuf, { struct trace_array *tr = filp->private_data; char buf[64]; - long val; + unsigned long val; int ret; if (cnt >= sizeof(buf)) @@ -2295,9 +2295,9 @@ static ssize_t tracing_max_lat_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) { - long *ptr = filp->private_data; + unsigned long *ptr = filp->private_data; char buf[64]; - long val; + unsigned long val; int ret; if (cnt >= sizeof(buf)) From e7669b8e329255bbcb40af65b38e342825d97a46 Mon Sep 17 00:00:00 2001 From: Hannes Eder Date: Tue, 10 Feb 2009 19:44:45 +0100 Subject: [PATCH 133/413] tracing: fix sparse warning: attribute function with __acquires/__releases Fix this sparse warning: kernel/trace/trace.c:458:9: warning: context imbalance in 'register_tracer' - unexpected unlock Signed-off-by: Hannes Eder Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e2434990277c..95f99a7abf2f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -459,6 +459,8 @@ update_max_tr_single(struct trace_array *tr, struct task_struct *tsk, int cpu) * Register a new plugin tracer. */ int register_tracer(struct tracer *type) +__releases(kernel_lock) +__acquires(kernel_lock) { struct tracer *t; int len; From b22f4858126a6aa852ad745b94f6b25dbdea708e Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 10 Feb 2009 15:49:11 +0100 Subject: [PATCH 134/413] tracing/sysprof: add missing tracing_{start,stop}_record_cmdline() Add the missing pair tracing_{start,stop}_record_cmdline() to record well the cmdline associated with pid. Changes in v2: - fix a build error, the sched_switch tracer is needed to record the cmdline. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 1 + kernel/trace/trace_sysprof.c | 3 +++ 2 files changed, 4 insertions(+) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 3a331289457a..6ff928acd453 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -134,6 +134,7 @@ config SYSPROF_TRACER bool "Sysprof Tracer" depends on X86 select TRACING + select CONTEXT_SWITCH_TRACER help This tracer provides the trace needed by the 'Sysprof' userspace tool. diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c index 84ca9d81e74d..9902c15997ad 100644 --- a/kernel/trace/trace_sysprof.c +++ b/kernel/trace/trace_sysprof.c @@ -238,6 +238,8 @@ static int stack_trace_init(struct trace_array *tr) { sysprof_trace = tr; + tracing_start_cmdline_record(); + mutex_lock(&sample_timer_lock); start_stack_timers(); tracer_enabled = 1; @@ -247,6 +249,7 @@ static int stack_trace_init(struct trace_array *tr) static void stack_trace_reset(struct trace_array *tr) { + tracing_stop_cmdline_record(); stop_stack_trace(tr); } From 00f62f614bb713027b9296068d1879fbca511eb7 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 9 Feb 2009 17:04:06 -0200 Subject: [PATCH 135/413] ring_buffer: pahole struct ring_buffer While fixing some bugs in pahole (built-in.o files were not being processed due to relocation problems) I found out about these packable structures: $ pahole --packable kernel/trace/ring_buffer.o | grep ring ring_buffer 72 64 8 ring_buffer_per_cpu 112 104 8 If we take a look at the current layout of struct ring_buffer we can see that we have two 4 bytes holes. $ pahole -C ring_buffer kernel/trace/ring_buffer.o struct ring_buffer { unsigned int pages; /* 0 4 */ unsigned int flags; /* 4 4 */ int cpus; /* 8 4 */ /* XXX 4 bytes hole, try to pack */ cpumask_var_t cpumask; /* 16 8 */ atomic_t record_disabled; /* 24 4 */ /* XXX 4 bytes hole, try to pack */ struct mutex mutex; /* 32 32 */ /* --- cacheline 1 boundary (64 bytes) --- */ struct ring_buffer_per_cpu * * buffers; /* 64 8 */ /* size: 72, cachelines: 2, members: 7 */ /* sum members: 64, holes: 2, sum holes: 8 */ /* last cacheline: 8 bytes */ }; So, if I ask pahole to reorganize it: $ pahole -C ring_buffer --reorganize kernel/trace/ring_buffer.o struct ring_buffer { unsigned int pages; /* 0 4 */ unsigned int flags; /* 4 4 */ int cpus; /* 8 4 */ atomic_t record_disabled; /* 12 4 */ cpumask_var_t cpumask; /* 16 8 */ struct mutex mutex; /* 24 32 */ struct ring_buffer_per_cpu * * buffers; /* 56 8 */ /* --- cacheline 1 boundary (64 bytes) --- */ /* size: 64, cachelines: 1, members: 7 */ }; /* saved 8 bytes and 1 cacheline! */ We get it using just one 64 bytes cacheline. To see what it did: $ pahole -C ring_buffer --reorganize --show_reorg_steps \ kernel/trace/ring_buffer.o | grep \/ /* Moving 'record_disabled' from after 'cpumask' to after 'cpus' */ Signed-off-by: Arnaldo Carvalho de Melo Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 53ba3a6d16d0..27ef3bf13ed2 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -273,8 +273,8 @@ struct ring_buffer { unsigned pages; unsigned flags; int cpus; - cpumask_var_t cpumask; atomic_t record_disabled; + cpumask_var_t cpumask; struct mutex mutex; From 5a5fb7dbe88dd57dc2bef0f3be9da991e789612d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Feb 2009 10:53:37 -0500 Subject: [PATCH 136/413] preempt-count: force hardirq-count to max of 10 To add a bit in the preempt_count to be set when in NMI context, we found that some archs did not have enough bits to spare. This is due to the hardirq_count being a mask that can hold NR_IRQS. Some archs allow for over 16000 IRQs, and that would require a mask of 14 bits. The sofitrq mask is 8 bits and the preempt disable mask is also 8 bits. The PREEMP_ACTIVE bit is bit 30, and bit 31 would make the preempt_count (which is type int) a negative number. A negative preempt_count is a sign of failure. Add them up 14+8+8+1+1 you get 32 bits. No room for the NMI bit. But the hardirq_count is to track the number of nested IRQs, not the number of total IRQs. This originally took the paranoid approach of setting the max nesting to NR_IRQS. But when we have archs with over 1000 IRQs, it is not practical to think they will ever all nest on a single CPU. Not to mention that this would most definitely cause a stack overflow. This patch sets a max of 10 bits to be used for IRQ nesting. I did a 'git grep HARDIRQ' to examine all users of HARDIRQ_BITS and HARDIRQ_MASK, and found that making it a max of 10 would not hurt anyone. I did find that the m68k expected it to be 8 bits, so I allow for the archs to set the number to be less than 10. I removed the setting of HARDIRQ_BITS from the archs that set it to more than 10. This includes ALPHA, ia64 and avr32. This will always allow room for the NMI bit, and if we need to allow for NMI nesting, we have 4 bits to play with. Signed-off-by: Steven Rostedt --- arch/alpha/include/asm/hardirq.h | 13 --------- arch/avr32/include/asm/hardirq.h | 11 -------- arch/ia64/include/asm/hardirq.h | 10 ------- include/linux/hardirq.h | 48 ++++++++++++++++---------------- 4 files changed, 24 insertions(+), 58 deletions(-) diff --git a/arch/alpha/include/asm/hardirq.h b/arch/alpha/include/asm/hardirq.h index d953e234daa8..88971460fa6c 100644 --- a/arch/alpha/include/asm/hardirq.h +++ b/arch/alpha/include/asm/hardirq.h @@ -14,17 +14,4 @@ typedef struct { void ack_bad_irq(unsigned int irq); -#define HARDIRQ_BITS 12 - -/* - * The hardirq mask has to be large enough to have - * space for potentially nestable IRQ sources in the system - * to nest on a single CPU. On Alpha, interrupts are masked at the CPU - * by IPL as well as at the system level. We only have 8 IPLs (UNIX PALcode) - * so we really only have 8 nestable IRQs, but allow some overhead - */ -#if (1 << HARDIRQ_BITS) < 16 -#error HARDIRQ_BITS is too low! -#endif - #endif /* _ALPHA_HARDIRQ_H */ diff --git a/arch/avr32/include/asm/hardirq.h b/arch/avr32/include/asm/hardirq.h index 267354356f60..015bc75ea798 100644 --- a/arch/avr32/include/asm/hardirq.h +++ b/arch/avr32/include/asm/hardirq.h @@ -20,15 +20,4 @@ void ack_bad_irq(unsigned int irq); #endif /* __ASSEMBLY__ */ -#define HARDIRQ_BITS 12 - -/* - * The hardirq mask has to be large enough to have - * space for potentially all IRQ sources in the system - * nesting on a single CPU: - */ -#if (1 << HARDIRQ_BITS) < NR_IRQS -# error HARDIRQ_BITS is too low! -#endif - #endif /* __ASM_AVR32_HARDIRQ_H */ diff --git a/arch/ia64/include/asm/hardirq.h b/arch/ia64/include/asm/hardirq.h index 140e495b8e0e..d514cd9edb49 100644 --- a/arch/ia64/include/asm/hardirq.h +++ b/arch/ia64/include/asm/hardirq.h @@ -20,16 +20,6 @@ #define local_softirq_pending() (local_cpu_data->softirq_pending) -#define HARDIRQ_BITS 14 - -/* - * The hardirq mask has to be large enough to have space for potentially all IRQ sources - * in the system nesting on a single CPU: - */ -#if (1 << HARDIRQ_BITS) < NR_IRQS -# error HARDIRQ_BITS is too low! -#endif - extern void __iomem *ipi_base_addr; void ack_bad_irq(unsigned int irq); diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index f3cf86e1465b..9841221f53f2 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -15,61 +15,61 @@ * - bits 0-7 are the preemption count (max preemption depth: 256) * - bits 8-15 are the softirq count (max # of softirqs: 256) * - * The hardirq count can be overridden per architecture, the default is: + * The hardirq count can in theory reach the same as NR_IRQS. + * In reality, the number of nested IRQS is limited to the stack + * size as well. For archs with over 1000 IRQS it is not practical + * to expect that they will all nest. We give a max of 10 bits for + * hardirq nesting. An arch may choose to give less than 10 bits. + * m68k expects it to be 8. * - * - bits 16-27 are the hardirq count (max # of hardirqs: 4096) - * - ( bit 28 is the PREEMPT_ACTIVE flag. ) + * - bits 16-25 are the hardirq count (max # of nested hardirqs: 1024) + * - bit 26 is the NMI_MASK + * - bit 28 is the PREEMPT_ACTIVE flag * * PREEMPT_MASK: 0x000000ff * SOFTIRQ_MASK: 0x0000ff00 - * HARDIRQ_MASK: 0x0fff0000 + * HARDIRQ_MASK: 0x03ff0000 + * NMI_MASK: 0x04000000 */ #define PREEMPT_BITS 8 #define SOFTIRQ_BITS 8 +#define NMI_BITS 1 + +#define MAX_HARDIRQ_BITS 10 #ifndef HARDIRQ_BITS -#define HARDIRQ_BITS 12 - -#ifndef MAX_HARDIRQS_PER_CPU -#define MAX_HARDIRQS_PER_CPU NR_IRQS +# define HARDIRQ_BITS MAX_HARDIRQ_BITS #endif -/* - * The hardirq mask has to be large enough to have space for potentially - * all IRQ sources in the system nesting on a single CPU. - */ -#if (1 << HARDIRQ_BITS) < MAX_HARDIRQS_PER_CPU -# error HARDIRQ_BITS is too low! -#endif +#if HARDIRQ_BITS > MAX_HARDIRQ_BITS +#error HARDIRQ_BITS too high! #endif #define PREEMPT_SHIFT 0 #define SOFTIRQ_SHIFT (PREEMPT_SHIFT + PREEMPT_BITS) #define HARDIRQ_SHIFT (SOFTIRQ_SHIFT + SOFTIRQ_BITS) +#define NMI_SHIFT (HARDIRQ_SHIFT + HARDIRQ_BITS) #define __IRQ_MASK(x) ((1UL << (x))-1) #define PREEMPT_MASK (__IRQ_MASK(PREEMPT_BITS) << PREEMPT_SHIFT) #define SOFTIRQ_MASK (__IRQ_MASK(SOFTIRQ_BITS) << SOFTIRQ_SHIFT) #define HARDIRQ_MASK (__IRQ_MASK(HARDIRQ_BITS) << HARDIRQ_SHIFT) +#define NMI_MASK (__IRQ_MASK(NMI_BITS) << NMI_SHIFT) #define PREEMPT_OFFSET (1UL << PREEMPT_SHIFT) #define SOFTIRQ_OFFSET (1UL << SOFTIRQ_SHIFT) #define HARDIRQ_OFFSET (1UL << HARDIRQ_SHIFT) +#define NMI_OFFSET (1UL << NMI_SHIFT) -#if PREEMPT_ACTIVE < (1 << (HARDIRQ_SHIFT + HARDIRQ_BITS)) +#if PREEMPT_ACTIVE < (1 << (NMI_SHIFT + NMI_BITS)) #error PREEMPT_ACTIVE is too low! #endif -#define NMI_OFFSET (PREEMPT_ACTIVE << 1) - -#if NMI_OFFSET >= 0x80000000 -#error PREEMPT_ACTIVE too high! -#endif - #define hardirq_count() (preempt_count() & HARDIRQ_MASK) #define softirq_count() (preempt_count() & SOFTIRQ_MASK) -#define irq_count() (preempt_count() & (HARDIRQ_MASK | SOFTIRQ_MASK)) +#define irq_count() (preempt_count() & (HARDIRQ_MASK | SOFTIRQ_MASK \ + | NMI_MASK)) /* * Are we doing bottom half or hardware interrupt processing? @@ -82,7 +82,7 @@ /* * Are we in NMI context? */ -#define in_nmi() (preempt_count() & NMI_OFFSET) +#define in_nmi() (preempt_count() & NMI_MASK) #if defined(CONFIG_PREEMPT) # define PREEMPT_INATOMIC_BASE kernel_locked() From 45141d4667d208421ca787a3301542b6a5e0b112 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Feb 2009 13:19:48 -0500 Subject: [PATCH 137/413] ring-buffer: rename label out_unlock to out_reset Impact: clean up While reviewing the ring buffer code, I thougth I saw a bug with if (!__raw_spin_trylock(&cpu_buffer->lock)) goto out_unlock; But I forgot that we use a variable "lock_taken" that is set if the spinlock is taken, and only unlock it if that variable is set. To avoid further confusion from other reviewers, this patch renames the label out_unlock with out_reset, which is the more appropriate name. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index dc18b5b9ccb4..f39d7e9a4305 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1017,7 +1017,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, */ if (unlikely(in_nmi())) { if (!__raw_spin_trylock(&cpu_buffer->lock)) - goto out_unlock; + goto out_reset; } else __raw_spin_lock(&cpu_buffer->lock); @@ -1030,7 +1030,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, /* we grabbed the lock before incrementing */ if (RB_WARN_ON(cpu_buffer, next_page == reader_page)) - goto out_unlock; + goto out_reset; /* * If for some reason, we had an interrupt storm that made @@ -1039,12 +1039,12 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, */ if (unlikely(next_page == commit_page)) { WARN_ON_ONCE(1); - goto out_unlock; + goto out_reset; } if (next_page == head_page) { if (!(buffer->flags & RB_FL_OVERWRITE)) - goto out_unlock; + goto out_reset; /* tail_page has not moved yet? */ if (tail_page == cpu_buffer->tail_page) { @@ -1118,7 +1118,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, return event; - out_unlock: + out_reset: /* reset write */ if (tail <= BUF_PAGE_SIZE) local_set(&tail_page->write, tail); From 2a7b8df04c11a70105c1abe67d006455d3bdc944 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Feb 2009 14:16:46 -0500 Subject: [PATCH 138/413] sched: do not account for NMIs Impact: avoid corruption in system time accounting Martin Schwidefsky told me that there was an issue with NMIs and system accounting. The problem is that the accounting code is not reentrant, and if an NMI goes off after an interrupt it can corrupt the accounting. For now, the best we can do is to treat NMIs like SMIs and they are not accounted for. This patch changes nmi_enter to not call __irq_enter and to do the preempt-count and tracing calls directly. Signed-off-by: Steven Rostedt --- include/linux/hardirq.h | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h index 9841221f53f2..faa1cf848bcd 100644 --- a/include/linux/hardirq.h +++ b/include/linux/hardirq.h @@ -175,24 +175,24 @@ extern void irq_enter(void); */ extern void irq_exit(void); -#define nmi_enter() \ - do { \ - ftrace_nmi_enter(); \ - BUG_ON(in_nmi()); \ - add_preempt_count(NMI_OFFSET); \ - lockdep_off(); \ - rcu_nmi_enter(); \ - __irq_enter(); \ +#define nmi_enter() \ + do { \ + ftrace_nmi_enter(); \ + BUG_ON(in_nmi()); \ + add_preempt_count(NMI_OFFSET + HARDIRQ_OFFSET); \ + lockdep_off(); \ + rcu_nmi_enter(); \ + trace_hardirq_enter(); \ } while (0) -#define nmi_exit() \ - do { \ - __irq_exit(); \ - rcu_nmi_exit(); \ - lockdep_on(); \ - BUG_ON(!in_nmi()); \ - sub_preempt_count(NMI_OFFSET); \ - ftrace_nmi_exit(); \ +#define nmi_exit() \ + do { \ + trace_hardirq_exit(); \ + rcu_nmi_exit(); \ + lockdep_on(); \ + BUG_ON(!in_nmi()); \ + sub_preempt_count(NMI_OFFSET + HARDIRQ_OFFSET); \ + ftrace_nmi_exit(); \ } while (0) #endif /* LINUX_HARDIRQ_H */ From b5f9fd0f8a05c9bafb91a9a85b9110938d8e585b Mon Sep 17 00:00:00 2001 From: Jason Baron Date: Wed, 11 Feb 2009 13:57:25 -0500 Subject: [PATCH 139/413] tracing: convert c/p state power tracer to use tracepoints Convert the c/p state "power" tracer to use tracepoints. Avoids a function call when the tracer is disabled. Signed-off-by: Jason Baron Acked-by: Ingo Molnar Signed-off-by: Steven Rostedt --- arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c | 2 + arch/x86/kernel/process.c | 3 + include/trace/power.h | 25 ++- kernel/trace/trace_power.c | 173 ++++++++++++--------- 4 files changed, 119 insertions(+), 84 deletions(-) diff --git a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c index 7ed925edf4d2..c5d737cdb365 100644 --- a/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -70,6 +70,8 @@ struct acpi_cpufreq_data { static DEFINE_PER_CPU(struct acpi_cpufreq_data *, drv_data); +DEFINE_TRACE(power_mark); + /* acpi_perf_data is a pointer to percpu data. */ static struct acpi_processor_performance *acpi_perf_data; diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index 026819ffcb0c..e0d0fd7ab514 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -19,6 +19,9 @@ EXPORT_SYMBOL(idle_nomwait); struct kmem_cache *task_xstate_cachep; +DEFINE_TRACE(power_start); +DEFINE_TRACE(power_end); + int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src) { *dst = *src; diff --git a/include/trace/power.h b/include/trace/power.h index c7cefbcdaea4..2c733e58e89c 100644 --- a/include/trace/power.h +++ b/include/trace/power.h @@ -2,6 +2,7 @@ #define _TRACE_POWER_H #include +#include enum { POWER_NONE = 0, @@ -18,18 +19,16 @@ struct power_trace { #endif }; -#ifdef CONFIG_POWER_TRACER -extern void trace_power_start(struct power_trace *it, unsigned int type, - unsigned int state); -extern void trace_power_mark(struct power_trace *it, unsigned int type, - unsigned int state); -extern void trace_power_end(struct power_trace *it); -#else -static inline void trace_power_start(struct power_trace *it, unsigned int type, - unsigned int state) { } -static inline void trace_power_mark(struct power_trace *it, unsigned int type, - unsigned int state) { } -static inline void trace_power_end(struct power_trace *it) { } -#endif +DECLARE_TRACE(power_start, + TPPROTO(struct power_trace *it, unsigned int type, unsigned int state), + TPARGS(it, type, state)); + +DECLARE_TRACE(power_mark, + TPPROTO(struct power_trace *it, unsigned int type, unsigned int state), + TPARGS(it, type, state)); + +DECLARE_TRACE(power_end, + TPPROTO(struct power_trace *it), + TPARGS(it)); #endif /* _TRACE_POWER_H */ diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c index b1d0d087d3a6..91ce672fb037 100644 --- a/kernel/trace/trace_power.c +++ b/kernel/trace/trace_power.c @@ -21,15 +21,116 @@ static struct trace_array *power_trace; static int __read_mostly trace_power_enabled; +static void probe_power_start(struct power_trace *it, unsigned int type, + unsigned int level) +{ + if (!trace_power_enabled) + return; + + memset(it, 0, sizeof(struct power_trace)); + it->state = level; + it->type = type; + it->stamp = ktime_get(); +} + + +static void probe_power_end(struct power_trace *it) +{ + struct ring_buffer_event *event; + struct trace_power *entry; + struct trace_array_cpu *data; + struct trace_array *tr = power_trace; + + if (!trace_power_enabled) + return; + + preempt_disable(); + it->end = ktime_get(); + data = tr->data[smp_processor_id()]; + + event = trace_buffer_lock_reserve(tr, TRACE_POWER, + sizeof(*entry), 0, 0); + if (!event) + goto out; + entry = ring_buffer_event_data(event); + entry->state_data = *it; + trace_buffer_unlock_commit(tr, event, 0, 0); + out: + preempt_enable(); +} + +static void probe_power_mark(struct power_trace *it, unsigned int type, + unsigned int level) +{ + struct ring_buffer_event *event; + struct trace_power *entry; + struct trace_array_cpu *data; + struct trace_array *tr = power_trace; + + if (!trace_power_enabled) + return; + + memset(it, 0, sizeof(struct power_trace)); + it->state = level; + it->type = type; + it->stamp = ktime_get(); + preempt_disable(); + it->end = it->stamp; + data = tr->data[smp_processor_id()]; + + event = trace_buffer_lock_reserve(tr, TRACE_POWER, + sizeof(*entry), 0, 0); + if (!event) + goto out; + entry = ring_buffer_event_data(event); + entry->state_data = *it; + trace_buffer_unlock_commit(tr, event, 0, 0); + out: + preempt_enable(); +} + +static int tracing_power_register(void) +{ + int ret; + + ret = register_trace_power_start(probe_power_start); + if (ret) { + pr_info("power trace: Couldn't activate tracepoint" + " probe to trace_power_start\n"); + return ret; + } + ret = register_trace_power_end(probe_power_end); + if (ret) { + pr_info("power trace: Couldn't activate tracepoint" + " probe to trace_power_end\n"); + goto fail_start; + } + ret = register_trace_power_mark(probe_power_mark); + if (ret) { + pr_info("power trace: Couldn't activate tracepoint" + " probe to trace_power_mark\n"); + goto fail_end; + } + return ret; +fail_end: + unregister_trace_power_end(probe_power_end); +fail_start: + unregister_trace_power_start(probe_power_start); + return ret; +} static void start_power_trace(struct trace_array *tr) { trace_power_enabled = 1; + tracing_power_register(); } static void stop_power_trace(struct trace_array *tr) { trace_power_enabled = 0; + unregister_trace_power_start(probe_power_start); + unregister_trace_power_end(probe_power_end); + unregister_trace_power_mark(probe_power_mark); } @@ -39,6 +140,7 @@ static int power_trace_init(struct trace_array *tr) power_trace = tr; trace_power_enabled = 1; + tracing_power_register(); for_each_cpu(cpu, cpu_possible_mask) tracing_reset(tr, cpu); @@ -95,74 +197,3 @@ static int init_power_trace(void) return register_tracer(&power_tracer); } device_initcall(init_power_trace); - -void trace_power_start(struct power_trace *it, unsigned int type, - unsigned int level) -{ - if (!trace_power_enabled) - return; - - memset(it, 0, sizeof(struct power_trace)); - it->state = level; - it->type = type; - it->stamp = ktime_get(); -} -EXPORT_SYMBOL_GPL(trace_power_start); - - -void trace_power_end(struct power_trace *it) -{ - struct ring_buffer_event *event; - struct trace_power *entry; - struct trace_array_cpu *data; - struct trace_array *tr = power_trace; - - if (!trace_power_enabled) - return; - - preempt_disable(); - it->end = ktime_get(); - data = tr->data[smp_processor_id()]; - - event = trace_buffer_lock_reserve(tr, TRACE_POWER, - sizeof(*entry), 0, 0); - if (!event) - goto out; - entry = ring_buffer_event_data(event); - entry->state_data = *it; - trace_buffer_unlock_commit(tr, event, 0, 0); - out: - preempt_enable(); -} -EXPORT_SYMBOL_GPL(trace_power_end); - -void trace_power_mark(struct power_trace *it, unsigned int type, - unsigned int level) -{ - struct ring_buffer_event *event; - struct trace_power *entry; - struct trace_array_cpu *data; - struct trace_array *tr = power_trace; - - if (!trace_power_enabled) - return; - - memset(it, 0, sizeof(struct power_trace)); - it->state = level; - it->type = type; - it->stamp = ktime_get(); - preempt_disable(); - it->end = it->stamp; - data = tr->data[smp_processor_id()]; - - event = trace_buffer_lock_reserve(tr, TRACE_POWER, - sizeof(*entry), 0, 0); - if (!event) - goto out; - entry = ring_buffer_event_data(event); - entry->state_data = *it; - trace_buffer_unlock_commit(tr, event, 0, 0); - out: - preempt_enable(); -} -EXPORT_SYMBOL_GPL(trace_power_mark); From a234aa9ecdf47a5461573a21dc0b154278df5ba8 Mon Sep 17 00:00:00 2001 From: Rakib Mullick Date: Sat, 14 Feb 2009 09:36:00 +0600 Subject: [PATCH 140/413] tracing: fix section mismatch in trace_hw_branches.c The function bts_trace_init() references a variable bts_hotcpu_notifier which is marked as __cpuinitdata. Thus causes section mismatch. This patch fixes it. LD kernel/trace/built-in.o WARNING: kernel/trace/built-in.o(.text+0xc90c): Section mismatch in reference from the function bts_trace_init() to the variable .cpuinit.data:bts_hotcpu_notifier The function bts_trace_init() references the variable __cpuinitdata bts_hotcpu_notifier. This is often because bts_trace_init lacks a __cpuinitdata annotation or the annotation of bts_hotcpu_notifier is wrong. WARNING: kernel/trace/built-in.o(.text+0xc92a): Section mismatch in reference from the function bts_trace_reset() to the variable .cpuinit.data:bts_hotcpu_notifier The function bts_trace_reset() references the variable __cpuinitdata bts_hotcpu_notifier. This is often because bts_trace_reset lacks a __cpuinitdata annotation or the annotation of bts_hotcpu_notifier is wrong. Signed-off-by: Rakib Mullick Cc: markus.t.metzger@gmail.com Cc: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/trace/trace_hw_branches.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index 0794dd33f27b..3561aace075c 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -127,7 +127,7 @@ static struct notifier_block bts_hotcpu_notifier __cpuinitdata = { .notifier_call = bts_hotcpu_handler }; -static int bts_trace_init(struct trace_array *tr) +static int __cpuinit bts_trace_init(struct trace_array *tr) { hw_branch_trace = tr; @@ -137,7 +137,7 @@ static int bts_trace_init(struct trace_array *tr) return 0; } -static void bts_trace_reset(struct trace_array *tr) +static void __cpuinit bts_trace_reset(struct trace_array *tr) { bts_trace_stop(tr); unregister_hotcpu_notifier(&bts_hotcpu_notifier); From 0c75a3ed633419d75d823d5dcb05d42924c6ae61 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 16 Feb 2009 11:21:52 -0500 Subject: [PATCH 141/413] ftrace: state that all functions are enabled in set_ftrace_filter Impact: clean up, make set_ftrace_filter less confusing The set_ftrace_filter shows only the functions that will be traced. But when it is empty, it will trace all functions. This can be a bit confusing. This patch makes set_ftrace_filter show: #### all functions enabled #### When all functions will be traced, and we do not filter only a select few. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 1796e018fbff..369fb78bd4ab 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -773,6 +773,7 @@ enum { FTRACE_ITER_CONT = (1 << 1), FTRACE_ITER_NOTRACE = (1 << 2), FTRACE_ITER_FAILURES = (1 << 3), + FTRACE_ITER_PRINTALL = (1 << 4), }; #define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */ @@ -794,6 +795,9 @@ t_next(struct seq_file *m, void *v, loff_t *pos) (*pos)++; + if (iter->flags & FTRACE_ITER_PRINTALL) + return NULL; + /* should not be called from interrupt context */ spin_lock(&ftrace_lock); retry: @@ -834,6 +838,19 @@ static void *t_start(struct seq_file *m, loff_t *pos) struct ftrace_iterator *iter = m->private; void *p = NULL; + /* + * For set_ftrace_filter reading, if we have the filter + * off, we can short cut and just print out that all + * functions are enabled. + */ + if (iter->flags & FTRACE_ITER_FILTER && !ftrace_filtered) { + if (*pos > 0) + return NULL; + iter->flags |= FTRACE_ITER_PRINTALL; + (*pos)++; + return iter; + } + if (*pos > 0) { if (iter->idx < 0) return p; @@ -852,9 +869,15 @@ static void t_stop(struct seq_file *m, void *p) static int t_show(struct seq_file *m, void *v) { + struct ftrace_iterator *iter = m->private; struct dyn_ftrace *rec = v; char str[KSYM_SYMBOL_LEN]; + if (iter->flags & FTRACE_ITER_PRINTALL) { + seq_printf(m, "#### all functions enabled ####\n"); + return 0; + } + if (!rec) return 0; From 265c831cb03d533cbe159af45798ac9fef534260 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Feb 2009 12:43:56 -0500 Subject: [PATCH 142/413] ftrace: add do_for_each_ftrace_rec and while_for_each_ftrace_rec Impact: clean up To iterate over all the functions that dynamic trace knows about it requires two for loops. One to iterate over the pages and the other to iterate over the records within the page. There are several duplications of these loops in ftrace.c. This patch creates the macros do_for_each_ftrace_rec and while_for_each_ftrace_rec to handle this logic, and removes the duplicate code. While making this change, I also discovered and fixed a small bug that one of the iterations should exit the loop after it found the record it was searching for. This used a break when it should have used a goto, since there were two loops it needed to break out from. No real harm was done by this bug since it would only continue to search the other records, and the code was in a slow path anyway. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 202 ++++++++++++++++++++---------------------- 1 file changed, 98 insertions(+), 104 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 369fb78bd4ab..fed1ebc3a133 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -297,6 +297,19 @@ static struct ftrace_page *ftrace_pages; static struct dyn_ftrace *ftrace_free_records; +/* + * This is a double for. Do not use 'break' to break out of the loop, + * you must use a goto. + */ +#define do_for_each_ftrace_rec(pg, rec) \ + for (pg = ftrace_pages_start; pg; pg = pg->next) { \ + int _____i; \ + for (_____i = 0; _____i < pg->index; _____i++) { \ + rec = &pg->records[_____i]; + +#define while_for_each_ftrace_rec() \ + } \ + } #ifdef CONFIG_KPROBES @@ -341,7 +354,6 @@ void ftrace_release(void *start, unsigned long size) struct ftrace_page *pg; unsigned long s = (unsigned long)start; unsigned long e = s + size; - int i; if (ftrace_disabled || !start) return; @@ -349,14 +361,11 @@ void ftrace_release(void *start, unsigned long size) /* should not be called from interrupt context */ spin_lock(&ftrace_lock); - for (pg = ftrace_pages_start; pg; pg = pg->next) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; + do_for_each_ftrace_rec(pg, rec) { + if ((rec->ip >= s) && (rec->ip < e)) + ftrace_free_rec(rec); + } while_for_each_ftrace_rec(); - if ((rec->ip >= s) && (rec->ip < e)) - ftrace_free_rec(rec); - } - } spin_unlock(&ftrace_lock); } @@ -523,41 +532,37 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) static void ftrace_replace_code(int enable) { - int i, failed; + int failed; struct dyn_ftrace *rec; struct ftrace_page *pg; - for (pg = ftrace_pages_start; pg; pg = pg->next) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; + do_for_each_ftrace_rec(pg, rec) { + /* + * Skip over free records and records that have + * failed. + */ + if (rec->flags & FTRACE_FL_FREE || + rec->flags & FTRACE_FL_FAILED) + continue; - /* - * Skip over free records and records that have - * failed. - */ - if (rec->flags & FTRACE_FL_FREE || - rec->flags & FTRACE_FL_FAILED) - continue; - - /* ignore updates to this record's mcount site */ - if (get_kprobe((void *)rec->ip)) { - freeze_record(rec); - continue; - } else { - unfreeze_record(rec); - } - - failed = __ftrace_replace_code(rec, enable); - if (failed && (rec->flags & FTRACE_FL_CONVERTED)) { - rec->flags |= FTRACE_FL_FAILED; - if ((system_state == SYSTEM_BOOTING) || - !core_kernel_text(rec->ip)) { - ftrace_free_rec(rec); - } else - ftrace_bug(failed, rec->ip); - } + /* ignore updates to this record's mcount site */ + if (get_kprobe((void *)rec->ip)) { + freeze_record(rec); + continue; + } else { + unfreeze_record(rec); } - } + + failed = __ftrace_replace_code(rec, enable); + if (failed && (rec->flags & FTRACE_FL_CONVERTED)) { + rec->flags |= FTRACE_FL_FAILED; + if ((system_state == SYSTEM_BOOTING) || + !core_kernel_text(rec->ip)) { + ftrace_free_rec(rec); + } else + ftrace_bug(failed, rec->ip); + } + } while_for_each_ftrace_rec(); } static int @@ -956,22 +961,17 @@ static void ftrace_filter_reset(int enable) struct ftrace_page *pg; struct dyn_ftrace *rec; unsigned long type = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; - unsigned i; /* should not be called from interrupt context */ spin_lock(&ftrace_lock); if (enable) ftrace_filtered = 0; - pg = ftrace_pages_start; - while (pg) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; - if (rec->flags & FTRACE_FL_FAILED) - continue; - rec->flags &= ~type; - } - pg = pg->next; - } + do_for_each_ftrace_rec(pg, rec) { + if (rec->flags & FTRACE_FL_FAILED) + continue; + rec->flags &= ~type; + } while_for_each_ftrace_rec(); + spin_unlock(&ftrace_lock); } @@ -1094,44 +1094,39 @@ ftrace_match(unsigned char *buff, int len, int enable) spin_lock(&ftrace_lock); if (enable) ftrace_filtered = 1; - pg = ftrace_pages_start; - while (pg) { - for (i = 0; i < pg->index; i++) { - int matched = 0; - char *ptr; + do_for_each_ftrace_rec(pg, rec) { + int matched = 0; + char *ptr; - rec = &pg->records[i]; - if (rec->flags & FTRACE_FL_FAILED) - continue; - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); - switch (type) { - case MATCH_FULL: - if (strcmp(str, buff) == 0) - matched = 1; - break; - case MATCH_FRONT_ONLY: - if (memcmp(str, buff, match) == 0) - matched = 1; - break; - case MATCH_MIDDLE_ONLY: - if (strstr(str, search)) - matched = 1; - break; - case MATCH_END_ONLY: - ptr = strstr(str, search); - if (ptr && (ptr[search_len] == 0)) - matched = 1; - break; - } - if (matched) { - if (not) - rec->flags &= ~flag; - else - rec->flags |= flag; - } + if (rec->flags & FTRACE_FL_FAILED) + continue; + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + switch (type) { + case MATCH_FULL: + if (strcmp(str, buff) == 0) + matched = 1; + break; + case MATCH_FRONT_ONLY: + if (memcmp(str, buff, match) == 0) + matched = 1; + break; + case MATCH_MIDDLE_ONLY: + if (strstr(str, search)) + matched = 1; + break; + case MATCH_END_ONLY: + ptr = strstr(str, search); + if (ptr && (ptr[search_len] == 0)) + matched = 1; + break; } - pg = pg->next; - } + if (matched) { + if (not) + rec->flags &= ~flag; + else + rec->flags |= flag; + } + } while_for_each_ftrace_rec(); spin_unlock(&ftrace_lock); } @@ -1452,7 +1447,7 @@ ftrace_set_func(unsigned long *array, int idx, char *buffer) struct dyn_ftrace *rec; struct ftrace_page *pg; int found = 0; - int i, j; + int j; if (ftrace_disabled) return -ENODEV; @@ -1460,27 +1455,26 @@ ftrace_set_func(unsigned long *array, int idx, char *buffer) /* should not be called from interrupt context */ spin_lock(&ftrace_lock); - for (pg = ftrace_pages_start; pg; pg = pg->next) { - for (i = 0; i < pg->index; i++) { - rec = &pg->records[i]; + do_for_each_ftrace_rec(pg, rec) { - if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) - continue; + if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) + continue; - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); - if (strcmp(str, buffer) == 0) { - found = 1; - for (j = 0; j < idx; j++) - if (array[j] == rec->ip) { - found = 0; - break; - } - if (found) - array[idx] = rec->ip; - break; - } + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + if (strcmp(str, buffer) == 0) { + /* Return 1 if we add it to the array */ + found = 1; + for (j = 0; j < idx; j++) + if (array[j] == rec->ip) { + found = 0; + break; + } + if (found) + array[idx] = rec->ip; + goto out; } - } + } while_for_each_ftrace_rec(); + out: spin_unlock(&ftrace_lock); return found ? 0 : -EINVAL; From 7f24b31b01a271b62346d9df084b029e48612163 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Feb 2009 14:37:33 -0500 Subject: [PATCH 143/413] ftrace: rename ftrace_match to ftrace_match_records Impact: clean up ftrace_match is too generic of a name. What it really does is search all records and matches the records with the given string, and either sets or unsets the functions to be traced depending on if the parameter 'enable' is set or not. This allows us to make another function called ftrace_match that can be used to test a single record. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index fed1ebc3a133..f397d7adb62e 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1054,7 +1054,7 @@ enum { }; static void -ftrace_match(unsigned char *buff, int len, int enable) +ftrace_match_records(unsigned char *buff, int len, int enable) { char str[KSYM_SYMBOL_LEN]; char *search = NULL; @@ -1197,7 +1197,7 @@ ftrace_regex_write(struct file *file, const char __user *ubuf, if (isspace(ch)) { iter->filtered++; iter->buffer[iter->buffer_idx] = 0; - ftrace_match(iter->buffer, iter->buffer_idx, enable); + ftrace_match_records(iter->buffer, iter->buffer_idx, enable); iter->buffer_idx = 0; } else iter->flags |= FTRACE_ITER_CONT; @@ -1236,7 +1236,7 @@ ftrace_set_regex(unsigned char *buf, int len, int reset, int enable) if (reset) ftrace_filter_reset(enable); if (buf) - ftrace_match(buf, len, enable); + ftrace_match_records(buf, len, enable); mutex_unlock(&ftrace_regex_lock); } @@ -1286,7 +1286,7 @@ ftrace_regex_release(struct inode *inode, struct file *file, int enable) if (iter->buffer_idx) { iter->filtered++; iter->buffer[iter->buffer_idx] = 0; - ftrace_match(iter->buffer, iter->buffer_idx, enable); + ftrace_match_records(iter->buffer, iter->buffer_idx, enable); } mutex_lock(&ftrace_sysctl_lock); From 9f4801e30ad291e27284e873696da1ead92d68fa Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Feb 2009 15:56:43 -0500 Subject: [PATCH 144/413] ftrace: break up ftrace_match_records into smaller components Impact: clean up ftrace_match_records does a lot of things that other features can use. This patch breaks up ftrace_match_records and pulls out ftrace_setup_glob and ftrace_match_record. ftrace_setup_glob prepares a simple glob expression for use with ftrace_match_record. ftrace_match_record compares a single record with a glob type. Breaking this up will allow for more features to run on individual records. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 115 +++++++++++++++++++++++++++--------------- 1 file changed, 75 insertions(+), 40 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index f397d7adb62e..fcec31323a10 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1053,79 +1053,114 @@ enum { MATCH_END_ONLY, }; -static void -ftrace_match_records(unsigned char *buff, int len, int enable) +/* + * (static function - no need for kernel doc) + * + * Pass in a buffer containing a glob and this function will + * set search to point to the search part of the buffer and + * return the type of search it is (see enum above). + * This does modify buff. + * + * Returns enum type. + * search returns the pointer to use for comparison. + * not returns 1 if buff started with a '!' + * 0 otherwise. + */ +static int +ftrace_setup_glob(unsigned char *buff, int len, char **search, int *not) { - char str[KSYM_SYMBOL_LEN]; - char *search = NULL; - struct ftrace_page *pg; - struct dyn_ftrace *rec; int type = MATCH_FULL; - unsigned long flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; - unsigned i, match = 0, search_len = 0; - int not = 0; + int i; if (buff[0] == '!') { - not = 1; + *not = 1; buff++; len--; - } + } else + *not = 0; + + *search = buff; for (i = 0; i < len; i++) { if (buff[i] == '*') { if (!i) { - search = buff + i + 1; + *search = buff + 1; type = MATCH_END_ONLY; - search_len = len - (i + 1); } else { - if (type == MATCH_END_ONLY) { + if (type == MATCH_END_ONLY) type = MATCH_MIDDLE_ONLY; - } else { - match = i; + else type = MATCH_FRONT_ONLY; - } buff[i] = 0; break; } } } + return type; +} + +static int +ftrace_match_record(struct dyn_ftrace *rec, char *regex, int len, int type) +{ + char str[KSYM_SYMBOL_LEN]; + int matched = 0; + char *ptr; + + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + switch (type) { + case MATCH_FULL: + if (strcmp(str, regex) == 0) + matched = 1; + break; + case MATCH_FRONT_ONLY: + if (strncmp(str, regex, len) == 0) + matched = 1; + break; + case MATCH_MIDDLE_ONLY: + if (strstr(str, regex)) + matched = 1; + break; + case MATCH_END_ONLY: + ptr = strstr(str, regex); + if (ptr && (ptr[len] == 0)) + matched = 1; + break; + } + + return matched; +} + +static void ftrace_match_records(char *buff, int len, int enable) +{ + char *search; + struct ftrace_page *pg; + struct dyn_ftrace *rec; + int type; + unsigned long flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; + unsigned search_len; + int not; + + type = ftrace_setup_glob(buff, len, &search, ¬); + + search_len = strlen(search); + /* should not be called from interrupt context */ spin_lock(&ftrace_lock); if (enable) ftrace_filtered = 1; do_for_each_ftrace_rec(pg, rec) { - int matched = 0; - char *ptr; if (rec->flags & FTRACE_FL_FAILED) continue; - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); - switch (type) { - case MATCH_FULL: - if (strcmp(str, buff) == 0) - matched = 1; - break; - case MATCH_FRONT_ONLY: - if (memcmp(str, buff, match) == 0) - matched = 1; - break; - case MATCH_MIDDLE_ONLY: - if (strstr(str, search)) - matched = 1; - break; - case MATCH_END_ONLY: - ptr = strstr(str, search); - if (ptr && (ptr[search_len] == 0)) - matched = 1; - break; - } - if (matched) { + + if (ftrace_match_record(rec, search, search_len, type)) { if (not) rec->flags &= ~flag; else rec->flags |= flag; } + } while_for_each_ftrace_rec(); spin_unlock(&ftrace_lock); } From 64e7c440618998fd69eee6ab490b042d12248021 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Feb 2009 17:08:48 -0500 Subject: [PATCH 145/413] ftrace: add module command function filter selection This patch adds a "command" syntax to the function filtering files: /debugfs/tracing/set_ftrace_filter /debugfs/tracing/set_ftrace_notrace Of the format: :: The command is optional, and dependent on the command, so are the parameters. echo do_fork > set_ftrace_filter Will only trace 'do_fork'. echo 'sched_*' > set_ftrace_filter Will only trace functions starting with the letters 'sched_'. echo '*:mod:ext3' > set_ftrace_filter Will trace only the ext3 module functions. echo '*write*:mod:ext3' > set_ftrace_notrace Will prevent the ext3 functions with the letters 'write' in the name from being traced. echo '!*_allocate:mod:ext3' > set_ftrace_filter Will remove the functions in ext3 that end with the letters '_allocate' from the ftrace filter. Although this patch implements the 'command' format, only the 'mod' command is supported. More commands to follow. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 115 +++++++++++++++++++++++++++++++++++++++--- 1 file changed, 109 insertions(+), 6 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index fcec31323a10..9e60ae423af9 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1067,7 +1067,7 @@ enum { * 0 otherwise. */ static int -ftrace_setup_glob(unsigned char *buff, int len, char **search, int *not) +ftrace_setup_glob(char *buff, int len, char **search, int *not) { int type = MATCH_FULL; int i; @@ -1100,14 +1100,11 @@ ftrace_setup_glob(unsigned char *buff, int len, char **search, int *not) return type; } -static int -ftrace_match_record(struct dyn_ftrace *rec, char *regex, int len, int type) +static int ftrace_match(char *str, char *regex, int len, int type) { - char str[KSYM_SYMBOL_LEN]; int matched = 0; char *ptr; - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); switch (type) { case MATCH_FULL: if (strcmp(str, regex) == 0) @@ -1131,6 +1128,15 @@ ftrace_match_record(struct dyn_ftrace *rec, char *regex, int len, int type) return matched; } +static int +ftrace_match_record(struct dyn_ftrace *rec, char *regex, int len, int type) +{ + char str[KSYM_SYMBOL_LEN]; + + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + return ftrace_match(str, regex, len, type); +} + static void ftrace_match_records(char *buff, int len, int enable) { char *search; @@ -1165,6 +1171,100 @@ static void ftrace_match_records(char *buff, int len, int enable) spin_unlock(&ftrace_lock); } +static int +ftrace_match_module_record(struct dyn_ftrace *rec, char *mod, + char *regex, int len, int type) +{ + char str[KSYM_SYMBOL_LEN]; + char *modname; + + kallsyms_lookup(rec->ip, NULL, NULL, &modname, str); + + if (!modname || strcmp(modname, mod)) + return 0; + + /* blank search means to match all funcs in the mod */ + if (len) + return ftrace_match(str, regex, len, type); + else + return 1; +} + +static void ftrace_match_module_records(char *buff, char *mod, int enable) +{ + char *search = buff; + struct ftrace_page *pg; + struct dyn_ftrace *rec; + int type = MATCH_FULL; + unsigned long flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; + unsigned search_len = 0; + int not = 0; + + /* blank or '*' mean the same */ + if (strcmp(buff, "*") == 0) + buff[0] = 0; + + /* handle the case of 'dont filter this module' */ + if (strcmp(buff, "!") == 0 || strcmp(buff, "!*") == 0) { + buff[0] = 0; + not = 1; + } + + if (strlen(buff)) { + type = ftrace_setup_glob(buff, strlen(buff), &search, ¬); + search_len = strlen(search); + } + + /* should not be called from interrupt context */ + spin_lock(&ftrace_lock); + if (enable) + ftrace_filtered = 1; + + do_for_each_ftrace_rec(pg, rec) { + + if (rec->flags & FTRACE_FL_FAILED) + continue; + + if (ftrace_match_module_record(rec, mod, + search, search_len, type)) { + if (not) + rec->flags &= ~flag; + else + rec->flags |= flag; + } + + } while_for_each_ftrace_rec(); + spin_unlock(&ftrace_lock); +} + +static int ftrace_process_regex(char *buff, int len, int enable) +{ + char *func, *mod, *command, *next = buff; + + func = strsep(&next, ":"); + + if (!next) { + ftrace_match_records(func, len, enable); + return 0; + } + + /* command fonud */ + + command = strsep(&next, ":"); + + if (strcmp(command, "mod") == 0) { + /* only match modules */ + if (!next) + return -EINVAL; + + mod = strsep(&next, ":"); + ftrace_match_module_records(func, mod, enable); + return 0; + } + + return -EINVAL; +} + static ssize_t ftrace_regex_write(struct file *file, const char __user *ubuf, size_t cnt, loff_t *ppos, int enable) @@ -1232,7 +1332,10 @@ ftrace_regex_write(struct file *file, const char __user *ubuf, if (isspace(ch)) { iter->filtered++; iter->buffer[iter->buffer_idx] = 0; - ftrace_match_records(iter->buffer, iter->buffer_idx, enable); + ret = ftrace_process_regex(iter->buffer, + iter->buffer_idx, enable); + if (ret) + goto out; iter->buffer_idx = 0; } else iter->flags |= FTRACE_ITER_CONT; From e68746a271eb3393a2183840be9e903caddf765b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Feb 2009 20:53:42 -0500 Subject: [PATCH 146/413] ftrace: enable filtering only when a function is filtered on Impact: fix to prevent empty set_ftrace_filter and no ftrace output The function filter is used to only trace a given set of functions. The filter is enabled when a function name is echoed into the set_ftrace_filter file. But if the name has a typo and the function is not found, the filter is enabled, but no function is listed. This makes a confusing situation where set_ftrace_filter is empty but no functions ever get enabled for tracing. For example: # cat /debug/tracing/set_ftrace_filter #### all functions enabled #### # echo bad_name > set_ftrace_filter # cat /debug/tracing/set_ftrace_filter # echo function > current_tracer # cat trace # tracer: nop # # TASK-PID CPU# TIMESTAMP FUNCTION # | | | | | This patch changes that to only enable filtering if a function is set to be filtered on. Now, the filter is not enabled if a bad name is echoed into set_ftrace_filter. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 9e60ae423af9..340f88b68d9e 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1153,8 +1153,6 @@ static void ftrace_match_records(char *buff, int len, int enable) /* should not be called from interrupt context */ spin_lock(&ftrace_lock); - if (enable) - ftrace_filtered = 1; do_for_each_ftrace_rec(pg, rec) { if (rec->flags & FTRACE_FL_FAILED) @@ -1166,7 +1164,12 @@ static void ftrace_match_records(char *buff, int len, int enable) else rec->flags |= flag; } - + /* + * Only enable filtering if we have a function that + * is filtered on. + */ + if (enable && (rec->flags & FTRACE_FL_FILTER)) + ftrace_filtered = 1; } while_for_each_ftrace_rec(); spin_unlock(&ftrace_lock); } @@ -1217,9 +1220,6 @@ static void ftrace_match_module_records(char *buff, char *mod, int enable) /* should not be called from interrupt context */ spin_lock(&ftrace_lock); - if (enable) - ftrace_filtered = 1; - do_for_each_ftrace_rec(pg, rec) { if (rec->flags & FTRACE_FL_FAILED) @@ -1232,6 +1232,8 @@ static void ftrace_match_module_records(char *buff, char *mod, int enable) else rec->flags |= flag; } + if (enable && (rec->flags & FTRACE_FL_FILTER)) + ftrace_filtered = 1; } while_for_each_ftrace_rec(); spin_unlock(&ftrace_lock); From f6180773d90595650e11de0118bb112018290915 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 14 Feb 2009 00:40:25 -0500 Subject: [PATCH 147/413] ftrace: add command interface for function selection Allow for other tracers to add their own commands for function selection. This interface gives a trace the ability to name a command for function selection. Right now it is pretty limited in what it offers, but this is a building step for more features. The :mod: command is converted to this interface and also serves as a template for other implementations. Signed-off-by: Steven Rostedt --- include/linux/ftrace.h | 16 +++++++ kernel/trace/ftrace.c | 106 ++++++++++++++++++++++++++++++++++++----- 2 files changed, 111 insertions(+), 11 deletions(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 106b7909d500..f0a0ecc63b5c 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -95,6 +95,13 @@ stack_trace_sysctl(struct ctl_table *table, int write, loff_t *ppos); #endif +struct ftrace_func_command { + struct list_head list; + char *name; + int (*func)(char *func, char *cmd, + char *params, int enable); +}; + #ifdef CONFIG_DYNAMIC_FTRACE /* asm/ftrace.h must be defined for archs supporting dynamic ftrace */ #include @@ -119,6 +126,9 @@ struct dyn_ftrace { int ftrace_force_update(void); void ftrace_set_filter(unsigned char *buf, int len, int reset); +int register_ftrace_command(struct ftrace_func_command *cmd); +int unregister_ftrace_command(struct ftrace_func_command *cmd); + /* defined in arch */ extern int ftrace_ip_converted(unsigned long ip); extern int ftrace_dyn_arch_init(void *data); @@ -202,6 +212,12 @@ extern void ftrace_enable_daemon(void); # define ftrace_disable_daemon() do { } while (0) # define ftrace_enable_daemon() do { } while (0) static inline void ftrace_release(void *start, unsigned long size) { } +static inline int register_ftrace_command(struct ftrace_func_command *cmd) +{ +} +static inline int unregister_ftrace_command(char *cmd_name) +{ +} #endif /* CONFIG_DYNAMIC_FTRACE */ /* totally disable ftrace - can not re-enable after this */ diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 340f88b68d9e..45a44c402566 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1239,9 +1239,93 @@ static void ftrace_match_module_records(char *buff, char *mod, int enable) spin_unlock(&ftrace_lock); } +/* + * We register the module command as a template to show others how + * to register the a command as well. + */ + +static int +ftrace_mod_callback(char *func, char *cmd, char *param, int enable) +{ + char *mod; + + /* + * cmd == 'mod' because we only registered this func + * for the 'mod' ftrace_func_command. + * But if you register one func with multiple commands, + * you can tell which command was used by the cmd + * parameter. + */ + + /* we must have a module name */ + if (!param) + return -EINVAL; + + mod = strsep(¶m, ":"); + if (!strlen(mod)) + return -EINVAL; + + ftrace_match_module_records(func, mod, enable); + return 0; +} + +static struct ftrace_func_command ftrace_mod_cmd = { + .name = "mod", + .func = ftrace_mod_callback, +}; + +static int __init ftrace_mod_cmd_init(void) +{ + return register_ftrace_command(&ftrace_mod_cmd); +} +device_initcall(ftrace_mod_cmd_init); + +static LIST_HEAD(ftrace_commands); +static DEFINE_MUTEX(ftrace_cmd_mutex); + +int register_ftrace_command(struct ftrace_func_command *cmd) +{ + struct ftrace_func_command *p; + int ret = 0; + + mutex_lock(&ftrace_cmd_mutex); + list_for_each_entry(p, &ftrace_commands, list) { + if (strcmp(cmd->name, p->name) == 0) { + ret = -EBUSY; + goto out_unlock; + } + } + list_add(&cmd->list, &ftrace_commands); + out_unlock: + mutex_unlock(&ftrace_cmd_mutex); + + return ret; +} + +int unregister_ftrace_command(struct ftrace_func_command *cmd) +{ + struct ftrace_func_command *p, *n; + int ret = -ENODEV; + + mutex_lock(&ftrace_cmd_mutex); + list_for_each_entry_safe(p, n, &ftrace_commands, list) { + if (strcmp(cmd->name, p->name) == 0) { + ret = 0; + list_del_init(&p->list); + goto out_unlock; + } + } + out_unlock: + mutex_unlock(&ftrace_cmd_mutex); + + return ret; +} + static int ftrace_process_regex(char *buff, int len, int enable) { - char *func, *mod, *command, *next = buff; + struct ftrace_func_command *p; + char *func, *command, *next = buff; + int ret = -EINVAL; func = strsep(&next, ":"); @@ -1250,21 +1334,21 @@ static int ftrace_process_regex(char *buff, int len, int enable) return 0; } - /* command fonud */ + /* command found */ command = strsep(&next, ":"); - if (strcmp(command, "mod") == 0) { - /* only match modules */ - if (!next) - return -EINVAL; - - mod = strsep(&next, ":"); - ftrace_match_module_records(func, mod, enable); - return 0; + mutex_lock(&ftrace_cmd_mutex); + list_for_each_entry(p, &ftrace_commands, list) { + if (strcmp(p->name, command) == 0) { + ret = p->func(func, command, next, enable); + goto out_unlock; + } } + out_unlock: + mutex_unlock(&ftrace_cmd_mutex); - return -EINVAL; + return ret; } static ssize_t From 52baf11922db7377b580dd5448a07f71c6a35611 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 14 Feb 2009 01:15:39 -0500 Subject: [PATCH 148/413] ftrace: convert ftrace_lock from a spinlock to mutex Impact: clean up The older versions of ftrace required doing the ftrace list search under atomic context. Now all the calls are in non-atomic context. There is no reason to keep the ftrace_lock as a spinlock. This patch converts it to a mutex. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 51 ++++++++++++++++--------------------------- 1 file changed, 19 insertions(+), 32 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 45a44c402566..4771732037ee 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -61,7 +61,7 @@ int function_trace_stop; */ static int ftrace_disabled __read_mostly; -static DEFINE_SPINLOCK(ftrace_lock); +static DEFINE_MUTEX(ftrace_lock); static DEFINE_MUTEX(ftrace_sysctl_lock); static DEFINE_MUTEX(ftrace_start_lock); @@ -134,8 +134,7 @@ static void ftrace_test_stop_func(unsigned long ip, unsigned long parent_ip) static int __register_ftrace_function(struct ftrace_ops *ops) { - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); ops->next = ftrace_list; /* @@ -172,7 +171,7 @@ static int __register_ftrace_function(struct ftrace_ops *ops) #endif } - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); return 0; } @@ -182,8 +181,7 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) struct ftrace_ops **p; int ret = 0; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); /* * If we are removing the last function, then simply point @@ -224,7 +222,7 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) } out: - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); return ret; } @@ -233,8 +231,7 @@ static void ftrace_update_pid_func(void) { ftrace_func_t func; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); if (ftrace_trace_function == ftrace_stub) goto out; @@ -256,7 +253,7 @@ static void ftrace_update_pid_func(void) #endif out: - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); } #ifdef CONFIG_DYNAMIC_FTRACE @@ -358,15 +355,12 @@ void ftrace_release(void *start, unsigned long size) if (ftrace_disabled || !start) return; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); - + mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { if ((rec->ip >= s) && (rec->ip < e)) ftrace_free_rec(rec); } while_for_each_ftrace_rec(); - - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); } static struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip) @@ -803,8 +797,7 @@ t_next(struct seq_file *m, void *v, loff_t *pos) if (iter->flags & FTRACE_ITER_PRINTALL) return NULL; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); retry: if (iter->idx >= iter->pg->index) { if (iter->pg->next) { @@ -833,7 +826,7 @@ t_next(struct seq_file *m, void *v, loff_t *pos) goto retry; } } - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); return rec; } @@ -962,8 +955,7 @@ static void ftrace_filter_reset(int enable) struct dyn_ftrace *rec; unsigned long type = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); if (enable) ftrace_filtered = 0; do_for_each_ftrace_rec(pg, rec) { @@ -971,8 +963,7 @@ static void ftrace_filter_reset(int enable) continue; rec->flags &= ~type; } while_for_each_ftrace_rec(); - - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); } static int @@ -1151,8 +1142,7 @@ static void ftrace_match_records(char *buff, int len, int enable) search_len = strlen(search); - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { if (rec->flags & FTRACE_FL_FAILED) @@ -1171,7 +1161,7 @@ static void ftrace_match_records(char *buff, int len, int enable) if (enable && (rec->flags & FTRACE_FL_FILTER)) ftrace_filtered = 1; } while_for_each_ftrace_rec(); - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); } static int @@ -1218,8 +1208,7 @@ static void ftrace_match_module_records(char *buff, char *mod, int enable) search_len = strlen(search); } - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); + mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { if (rec->flags & FTRACE_FL_FAILED) @@ -1236,7 +1225,7 @@ static void ftrace_match_module_records(char *buff, char *mod, int enable) ftrace_filtered = 1; } while_for_each_ftrace_rec(); - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); } /* @@ -1676,9 +1665,7 @@ ftrace_set_func(unsigned long *array, int idx, char *buffer) if (ftrace_disabled) return -ENODEV; - /* should not be called from interrupt context */ - spin_lock(&ftrace_lock); - + mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) @@ -1699,7 +1686,7 @@ ftrace_set_func(unsigned long *array, int idx, char *buffer) } } while_for_each_ftrace_rec(); out: - spin_unlock(&ftrace_lock); + mutex_unlock(&ftrace_lock); return found ? 0 : -EINVAL; } From e6ea44e9b4c12325337cd1c06103cd515a1c02b2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 14 Feb 2009 01:42:44 -0500 Subject: [PATCH 149/413] ftrace: consolidate mutexes Impact: clean up Now that ftrace_lock is a mutex, there is no reason to have three different mutexes protecting similar data. All the mutex paths are not in hot paths, so having a mutex to cover more data is not a problem. This patch removes the ftrace_sysctl_lock and ftrace_start_lock and uses the ftrace_lock to protect the locations that were protected by these locks. By doing so, this change also removes some of the lock nesting that was taking place. There are still more mutexes in ftrace.c that can probably be consolidated, but they can be dealt with later. We need to be careful about the way the locks are nested, and by consolidating, we can cause a recursive deadlock. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 68 +++++++++++++------------------------------ 1 file changed, 21 insertions(+), 47 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 4771732037ee..157d4f68b0e0 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -62,8 +62,6 @@ int function_trace_stop; static int ftrace_disabled __read_mostly; static DEFINE_MUTEX(ftrace_lock); -static DEFINE_MUTEX(ftrace_sysctl_lock); -static DEFINE_MUTEX(ftrace_start_lock); static struct ftrace_ops ftrace_list_end __read_mostly = { @@ -134,8 +132,6 @@ static void ftrace_test_stop_func(unsigned long ip, unsigned long parent_ip) static int __register_ftrace_function(struct ftrace_ops *ops) { - mutex_lock(&ftrace_lock); - ops->next = ftrace_list; /* * We are entering ops into the ftrace_list but another @@ -171,17 +167,12 @@ static int __register_ftrace_function(struct ftrace_ops *ops) #endif } - mutex_unlock(&ftrace_lock); - return 0; } static int __unregister_ftrace_function(struct ftrace_ops *ops) { struct ftrace_ops **p; - int ret = 0; - - mutex_lock(&ftrace_lock); /* * If we are removing the last function, then simply point @@ -190,17 +181,15 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) if (ftrace_list == ops && ops->next == &ftrace_list_end) { ftrace_trace_function = ftrace_stub; ftrace_list = &ftrace_list_end; - goto out; + return 0; } for (p = &ftrace_list; *p != &ftrace_list_end; p = &(*p)->next) if (*p == ops) break; - if (*p != ops) { - ret = -1; - goto out; - } + if (*p != ops) + return -1; *p = (*p)->next; @@ -221,10 +210,7 @@ static int __unregister_ftrace_function(struct ftrace_ops *ops) } } - out: - mutex_unlock(&ftrace_lock); - - return ret; + return 0; } static void ftrace_update_pid_func(void) @@ -622,13 +608,10 @@ static void ftrace_startup(int command) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); ftrace_start_up++; command |= FTRACE_ENABLE_CALLS; ftrace_startup_enable(command); - - mutex_unlock(&ftrace_start_lock); } static void ftrace_shutdown(int command) @@ -636,7 +619,6 @@ static void ftrace_shutdown(int command) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); ftrace_start_up--; if (!ftrace_start_up) command |= FTRACE_DISABLE_CALLS; @@ -647,11 +629,9 @@ static void ftrace_shutdown(int command) } if (!command || !ftrace_enabled) - goto out; + return; ftrace_run_update_code(command); - out: - mutex_unlock(&ftrace_start_lock); } static void ftrace_startup_sysctl(void) @@ -661,7 +641,6 @@ static void ftrace_startup_sysctl(void) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); /* Force update next time */ saved_ftrace_func = NULL; /* ftrace_start_up is true if we want ftrace running */ @@ -669,7 +648,6 @@ static void ftrace_startup_sysctl(void) command |= FTRACE_ENABLE_CALLS; ftrace_run_update_code(command); - mutex_unlock(&ftrace_start_lock); } static void ftrace_shutdown_sysctl(void) @@ -679,13 +657,11 @@ static void ftrace_shutdown_sysctl(void) if (unlikely(ftrace_disabled)) return; - mutex_lock(&ftrace_start_lock); /* ftrace_start_up is true if ftrace is running */ if (ftrace_start_up) command |= FTRACE_DISABLE_CALLS; ftrace_run_update_code(command); - mutex_unlock(&ftrace_start_lock); } static cycle_t ftrace_update_time; @@ -1502,12 +1478,10 @@ ftrace_regex_release(struct inode *inode, struct file *file, int enable) ftrace_match_records(iter->buffer, iter->buffer_idx, enable); } - mutex_lock(&ftrace_sysctl_lock); - mutex_lock(&ftrace_start_lock); + mutex_lock(&ftrace_lock); if (ftrace_start_up && ftrace_enabled) ftrace_run_update_code(FTRACE_ENABLE_CALLS); - mutex_unlock(&ftrace_start_lock); - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); kfree(iter); mutex_unlock(&ftrace_regex_lock); @@ -1824,7 +1798,7 @@ static int ftrace_convert_nops(struct module *mod, unsigned long addr; unsigned long flags; - mutex_lock(&ftrace_start_lock); + mutex_lock(&ftrace_lock); p = start; while (p < end) { addr = ftrace_call_adjust(*p++); @@ -1843,7 +1817,7 @@ static int ftrace_convert_nops(struct module *mod, local_irq_save(flags); ftrace_update_code(mod); local_irq_restore(flags); - mutex_unlock(&ftrace_start_lock); + mutex_unlock(&ftrace_lock); return 0; } @@ -2016,7 +1990,7 @@ ftrace_pid_write(struct file *filp, const char __user *ubuf, if (ret < 0) return ret; - mutex_lock(&ftrace_start_lock); + mutex_lock(&ftrace_lock); if (val < 0) { /* disable pid tracing */ if (!ftrace_pid_trace) @@ -2055,7 +2029,7 @@ ftrace_pid_write(struct file *filp, const char __user *ubuf, ftrace_startup_enable(0); out: - mutex_unlock(&ftrace_start_lock); + mutex_unlock(&ftrace_lock); return cnt; } @@ -2118,12 +2092,12 @@ int register_ftrace_function(struct ftrace_ops *ops) if (unlikely(ftrace_disabled)) return -1; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); ret = __register_ftrace_function(ops); ftrace_startup(0); - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } @@ -2137,10 +2111,10 @@ int unregister_ftrace_function(struct ftrace_ops *ops) { int ret; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); ret = __unregister_ftrace_function(ops); ftrace_shutdown(0); - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } @@ -2155,7 +2129,7 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, if (unlikely(ftrace_disabled)) return -ENODEV; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); ret = proc_dointvec(table, write, file, buffer, lenp, ppos); @@ -2184,7 +2158,7 @@ ftrace_enable_sysctl(struct ctl_table *table, int write, } out: - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } @@ -2296,7 +2270,7 @@ int register_ftrace_graph(trace_func_graph_ret_t retfunc, { int ret = 0; - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); ftrace_suspend_notifier.notifier_call = ftrace_suspend_notifier_call; register_pm_notifier(&ftrace_suspend_notifier); @@ -2314,13 +2288,13 @@ int register_ftrace_graph(trace_func_graph_ret_t retfunc, ftrace_startup(FTRACE_START_FUNC_RET); out: - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); return ret; } void unregister_ftrace_graph(void) { - mutex_lock(&ftrace_sysctl_lock); + mutex_lock(&ftrace_lock); atomic_dec(&ftrace_graph_active); ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; @@ -2328,7 +2302,7 @@ void unregister_ftrace_graph(void) ftrace_shutdown(FTRACE_STOP_FUNC_RET); unregister_pm_notifier(&ftrace_suspend_notifier); - mutex_unlock(&ftrace_sysctl_lock); + mutex_unlock(&ftrace_lock); } /* Allocate a return stack for newly created task */ From 59df055f1991c9fc0c71a9230663c39188f6972f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 14 Feb 2009 15:29:06 -0500 Subject: [PATCH 150/413] ftrace: trace different functions with a different tracer Impact: new feature Currently, the function tracer only gives you an ability to hook a tracer to all functions being traced. The dynamic function trace allows you to pick and choose which of those functions will be traced, but all functions being traced will call all tracers that registered with the function tracer. This patch adds a new feature that allows a tracer to hook to specific functions, even when all functions are being traced. It allows for different functions to call different tracer hooks. The way this is accomplished is by a special function that will hook to the function tracer and will set up a hash table knowing which tracer hook to call with which function. This is the most general and easiest method to accomplish this. Later, an arch may choose to supply their own method in changing the mcount call of a function to call a different tracer. But that will be an exercise for the future. To register a function: struct ftrace_hook_ops { void (*func)(unsigned long ip, unsigned long parent_ip, void **data); int (*callback)(unsigned long ip, void **data); void (*free)(void **data); }; int register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, void *data); glob is a simple glob to search for the functions to hook. ops is a pointer to the operations (listed below) data is the default data to be passed to the hook functions when traced ops: func is the hook function to call when the functions are traced callback is a callback function that is called when setting up the hash. That is, if the tracer needs to do something special for each function, that is being traced, and wants to give each function its own data. The address of the entry data is passed to this callback, so that the callback may wish to update the entry to whatever it would like. free is a callback for when the entry is freed. In case the tracer allocated any data, it is give the chance to free it. To unregister we have three functions: void unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, void *data) This will unregister all hooks that match glob, point to ops, and have its data matching data. (note, if glob is NULL, blank or '*', all functions will be tested). void unregister_ftrace_function_hook_func(char *glob, struct ftrace_hook_ops *ops) This will unregister all functions matching glob that has an entry pointing to ops. void unregister_ftrace_function_hook_all(char *glob) This simply unregisters all funcs. Signed-off-by: Steven Rostedt --- include/linux/ftrace.h | 18 +++ kernel/trace/ftrace.c | 247 +++++++++++++++++++++++++++++++++++++++++ 2 files changed, 265 insertions(+) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index f0a0ecc63b5c..13918c4400ad 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -106,6 +106,24 @@ struct ftrace_func_command { /* asm/ftrace.h must be defined for archs supporting dynamic ftrace */ #include +struct ftrace_hook_ops { + void (*func)(unsigned long ip, + unsigned long parent_ip, + void **data); + int (*callback)(unsigned long ip, void **data); + void (*free)(void **data); +}; + +extern int +register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, + void *data); +extern void +unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, + void *data); +extern void +unregister_ftrace_function_hook_func(char *glob, struct ftrace_hook_ops *ops); +extern void unregister_ftrace_function_hook_all(char *glob); + enum { FTRACE_FL_FREE = (1 << 0), FTRACE_FL_FAILED = (1 << 1), diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 157d4f68b0e0..0b80e325f296 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -27,6 +27,7 @@ #include #include #include +#include #include @@ -1245,6 +1246,252 @@ static int __init ftrace_mod_cmd_init(void) } device_initcall(ftrace_mod_cmd_init); +#define FTRACE_HASH_BITS 7 +#define FTRACE_FUNC_HASHSIZE (1 << FTRACE_HASH_BITS) +static struct hlist_head ftrace_func_hash[FTRACE_FUNC_HASHSIZE] __read_mostly; + +struct ftrace_func_hook { + struct hlist_node node; + struct ftrace_hook_ops *ops; + unsigned long flags; + unsigned long ip; + void *data; + struct rcu_head rcu; +}; + +static void +function_trace_hook_call(unsigned long ip, unsigned long parent_ip) +{ + struct ftrace_func_hook *entry; + struct hlist_head *hhd; + struct hlist_node *n; + unsigned long key; + int resched; + + key = hash_long(ip, FTRACE_HASH_BITS); + + hhd = &ftrace_func_hash[key]; + + if (hlist_empty(hhd)) + return; + + /* + * Disable preemption for these calls to prevent a RCU grace + * period. This syncs the hash iteration and freeing of items + * on the hash. rcu_read_lock is too dangerous here. + */ + resched = ftrace_preempt_disable(); + hlist_for_each_entry_rcu(entry, n, hhd, node) { + if (entry->ip == ip) + entry->ops->func(ip, parent_ip, &entry->data); + } + ftrace_preempt_enable(resched); +} + +static struct ftrace_ops trace_hook_ops __read_mostly = +{ + .func = function_trace_hook_call, +}; + +static int ftrace_hook_registered; + +static void __enable_ftrace_function_hook(void) +{ + int i; + + if (ftrace_hook_registered) + return; + + for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { + struct hlist_head *hhd = &ftrace_func_hash[i]; + if (hhd->first) + break; + } + /* Nothing registered? */ + if (i == FTRACE_FUNC_HASHSIZE) + return; + + __register_ftrace_function(&trace_hook_ops); + ftrace_startup(0); + ftrace_hook_registered = 1; +} + +static void __disable_ftrace_function_hook(void) +{ + int i; + + if (!ftrace_hook_registered) + return; + + for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { + struct hlist_head *hhd = &ftrace_func_hash[i]; + if (hhd->first) + return; + } + + /* no more funcs left */ + __unregister_ftrace_function(&trace_hook_ops); + ftrace_shutdown(0); + ftrace_hook_registered = 0; +} + + +static void ftrace_free_entry_rcu(struct rcu_head *rhp) +{ + struct ftrace_func_hook *entry = + container_of(rhp, struct ftrace_func_hook, rcu); + + if (entry->ops->free) + entry->ops->free(&entry->data); + kfree(entry); +} + + +int +register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, + void *data) +{ + struct ftrace_func_hook *entry; + struct ftrace_page *pg; + struct dyn_ftrace *rec; + unsigned long key; + int type, len, not; + int count = 0; + char *search; + + type = ftrace_setup_glob(glob, strlen(glob), &search, ¬); + len = strlen(search); + + /* we do not support '!' for function hooks */ + if (WARN_ON(not)) + return -EINVAL; + + mutex_lock(&ftrace_lock); + do_for_each_ftrace_rec(pg, rec) { + + if (rec->flags & FTRACE_FL_FAILED) + continue; + + if (!ftrace_match_record(rec, search, len, type)) + continue; + + entry = kmalloc(sizeof(*entry), GFP_KERNEL); + if (!entry) { + /* If we did not hook to any, then return error */ + if (!count) + count = -ENOMEM; + goto out_unlock; + } + + count++; + + entry->data = data; + + /* + * The caller might want to do something special + * for each function we find. We call the callback + * to give the caller an opportunity to do so. + */ + if (ops->callback) { + if (ops->callback(rec->ip, &entry->data) < 0) { + /* caller does not like this func */ + kfree(entry); + continue; + } + } + + entry->ops = ops; + entry->ip = rec->ip; + + key = hash_long(entry->ip, FTRACE_HASH_BITS); + hlist_add_head_rcu(&entry->node, &ftrace_func_hash[key]); + + } while_for_each_ftrace_rec(); + __enable_ftrace_function_hook(); + + out_unlock: + mutex_unlock(&ftrace_lock); + + return count; +} + +enum { + HOOK_TEST_FUNC = 1, + HOOK_TEST_DATA = 2 +}; + +static void +__unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, + void *data, int flags) +{ + struct ftrace_func_hook *entry; + struct hlist_node *n, *tmp; + char str[KSYM_SYMBOL_LEN]; + int type = MATCH_FULL; + int i, len = 0; + char *search; + + if (glob && (strcmp(glob, "*") || !strlen(glob))) + glob = NULL; + else { + int not; + + type = ftrace_setup_glob(glob, strlen(glob), &search, ¬); + len = strlen(search); + + /* we do not support '!' for function hooks */ + if (WARN_ON(not)) + return; + } + + mutex_lock(&ftrace_lock); + for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { + struct hlist_head *hhd = &ftrace_func_hash[i]; + + hlist_for_each_entry_safe(entry, n, tmp, hhd, node) { + + /* break up if statements for readability */ + if ((flags & HOOK_TEST_FUNC) && entry->ops != ops) + continue; + + if ((flags & HOOK_TEST_DATA) && entry->data != data) + continue; + + /* do this last, since it is the most expensive */ + if (glob) { + kallsyms_lookup(entry->ip, NULL, NULL, + NULL, str); + if (!ftrace_match(str, glob, len, type)) + continue; + } + + hlist_del(&entry->node); + call_rcu(&entry->rcu, ftrace_free_entry_rcu); + } + } + __disable_ftrace_function_hook(); + mutex_unlock(&ftrace_lock); +} + +void +unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, + void *data) +{ + __unregister_ftrace_function_hook(glob, ops, data, + HOOK_TEST_FUNC | HOOK_TEST_DATA); +} + +void +unregister_ftrace_function_hook_func(char *glob, struct ftrace_hook_ops *ops) +{ + __unregister_ftrace_function_hook(glob, ops, NULL, HOOK_TEST_FUNC); +} + +void unregister_ftrace_function_hook_all(char *glob) +{ + __unregister_ftrace_function_hook(glob, NULL, NULL, 0); +} + static LIST_HEAD(ftrace_commands); static DEFINE_MUTEX(ftrace_cmd_mutex); From 988ae9d6b2bc3ebdc1a488490250a6812f85e9d4 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 14 Feb 2009 19:17:02 -0500 Subject: [PATCH 151/413] ring-buffer: add tracing_is_on to test if ring buffer is enabled This patch adds the tracing_is_on() interface to tell if the ring buffer is turned on or not. Signed-off-by: Steven Rostedt --- include/linux/ring_buffer.h | 2 ++ kernel/trace/ring_buffer.c | 9 +++++++++ 2 files changed, 11 insertions(+) diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index 8e6646a54acf..f5e793d69bd3 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -128,10 +128,12 @@ void ring_buffer_normalize_time_stamp(int cpu, u64 *ts); void tracing_on(void); void tracing_off(void); void tracing_off_permanent(void); +int tracing_is_on(void); #else static inline void tracing_on(void) { } static inline void tracing_off(void) { } static inline void tracing_off_permanent(void) { } +static inline int tracing_is_on(void) { return 0; } #endif void *ring_buffer_alloc_read_page(struct ring_buffer *buffer); diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 2b4626ce95d6..8f19f1aa42b0 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -98,6 +98,15 @@ void tracing_off_permanent(void) set_bit(RB_BUFFERS_DISABLED_BIT, &ring_buffer_flags); } +/** + * tracing_is_on - show state of ring buffers enabled + */ +int tracing_is_on(void) +{ + return ring_buffer_flags == RB_BUFFERS_ON; +} +EXPORT_SYMBOL_GPL(tracing_is_on); + #include "trace.h" /* Up this if you want to test the TIME_EXTENTS and normalization */ From 23b4ff3aa479c9e3bb23cb6b2d0a97878399784a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 14 Feb 2009 19:04:24 -0500 Subject: [PATCH 152/413] ftrace: add traceon traceoff commands to enable/disable the buffers This patch adds the new function selection commands traceon and traceoff. traceon sets the function to enable the ring buffers while traceoff disables the ring buffers. You can pass in the number of times you want the command to be executed when the function is hit. It will only execute if the state of the buffers are not already in that state. Example: # echo do_fork:traceon:4 Will enable the ring buffers if they are disabled every time it hits do_fork, up to 4 times. # echo sys_close:traceoff This will disable the ring buffers every time (unlimited) when sys_close is called. Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions.c | 135 +++++++++++++++++++++++++++++++++ 1 file changed, 135 insertions(+) diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 36bf9568ccd9..5c95708b9dc3 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -9,6 +9,7 @@ * Copyright (C) 2004-2006 Ingo Molnar * Copyright (C) 2004 William Lee Irwin III */ +#include #include #include #include @@ -231,9 +232,143 @@ static struct tracer function_trace __read_mostly = #endif }; +#ifdef CONFIG_DYNAMIC_FTRACE +static void +ftrace_traceon(unsigned long ip, unsigned long parent_ip, void **data) +{ + long *count = (long *)data; + + if (tracing_is_on()) + return; + + if (!*count) + return; + + if (*count != -1) + (*count)--; + + tracing_on(); +} + +static void +ftrace_traceoff(unsigned long ip, unsigned long parent_ip, void **data) +{ + long *count = (long *)data; + + if (!tracing_is_on()) + return; + + if (!*count) + return; + + if (*count != -1) + (*count)--; + + tracing_off(); +} + +static struct ftrace_hook_ops traceon_hook_ops = { + .func = ftrace_traceon, +}; + +static struct ftrace_hook_ops traceoff_hook_ops = { + .func = ftrace_traceoff, +}; + +static int +ftrace_trace_onoff_unreg(char *glob, char *cmd, char *param) +{ + struct ftrace_hook_ops *ops; + + /* we register both traceon and traceoff to this callback */ + if (strcmp(cmd, "traceon") == 0) + ops = &traceon_hook_ops; + else + ops = &traceoff_hook_ops; + + unregister_ftrace_function_hook_func(glob, ops); + + return 0; +} + +static int +ftrace_trace_onoff_callback(char *glob, char *cmd, char *param, int enable) +{ + struct ftrace_hook_ops *ops; + void *count = (void *)-1; + char *number; + int ret; + + /* hash funcs only work with set_ftrace_filter */ + if (!enable) + return -EINVAL; + + if (glob[0] == '!') + return ftrace_trace_onoff_unreg(glob+1, cmd, param); + + /* we register both traceon and traceoff to this callback */ + if (strcmp(cmd, "traceon") == 0) + ops = &traceon_hook_ops; + else + ops = &traceoff_hook_ops; + + if (!param) + goto out_reg; + + number = strsep(¶m, ":"); + + if (!strlen(number)) + goto out_reg; + + /* + * We use the callback data field (which is a pointer) + * as our counter. + */ + ret = strict_strtoul(number, 0, (unsigned long *)&count); + if (ret) + return ret; + + out_reg: + ret = register_ftrace_function_hook(glob, ops, count); + + return ret; +} + +static struct ftrace_func_command ftrace_traceon_cmd = { + .name = "traceon", + .func = ftrace_trace_onoff_callback, +}; + +static struct ftrace_func_command ftrace_traceoff_cmd = { + .name = "traceoff", + .func = ftrace_trace_onoff_callback, +}; + +static int __init init_func_cmd_traceon(void) +{ + int ret; + + ret = register_ftrace_command(&ftrace_traceoff_cmd); + if (ret) + return ret; + + ret = register_ftrace_command(&ftrace_traceon_cmd); + if (ret) + unregister_ftrace_command(&ftrace_traceoff_cmd); + return ret; +} +#else +static inline int init_func_cmd_traceon(void) +{ + return 0; +} +#endif /* CONFIG_DYNAMIC_FTRACE */ + static __init int init_function_trace(void) { + init_func_cmd_traceon(); return register_tracer(&function_trace); } device_initcall(init_function_trace); + From 8fc0c701c5b6c0c3e242758c3acef6f9047940a9 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 16 Feb 2009 15:28:00 -0500 Subject: [PATCH 153/413] ftrace: show selected functions in set_ftrace_filter This patch adds output to show what functions have tracer hooks attached to them. # echo 'sys_open:traceon:4' > /debug/tracing/set_ftrace_filter # cat set_ftrace_filter #### all functions enabled #### sys_open:ftrace_traceon:0000000000000004 # echo 'do_fork:traceoff:' > set_ftrace_filter # cat set_ftrace_filter #### all functions enabled #### sys_open:ftrace_traceon:0000000000000002 do_fork:ftrace_traceoff:ffffffffffffffff Note the 4 changed to a 2. This is because The code was executed twice since the traceoff was added. If a cat is done again: #### all functions enabled #### sys_open:ftrace_traceon do_fork:ftrace_traceoff:ffffffffffffffff The number disappears. That is because it will not print a NULL. Callbacks to allow the tracer to pretty print will be implemented soon. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 123 +++++++++++++++++++++++++++++++++++------- 1 file changed, 103 insertions(+), 20 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 0b80e325f296..1e058848cddb 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -45,14 +45,14 @@ ftrace_kill(); \ } while (0) +/* hash bits for specific function selection */ +#define FTRACE_HASH_BITS 7 +#define FTRACE_FUNC_HASHSIZE (1 << FTRACE_HASH_BITS) + /* ftrace_enabled is a method to turn ftrace on or off */ int ftrace_enabled __read_mostly; static int last_ftrace_enabled; -/* set when tracing only a pid */ -struct pid *ftrace_pid_trace; -static struct pid * const ftrace_swapper_pid = &init_struct_pid; - /* Quick disabling of function tracer. */ int function_trace_stop; @@ -248,6 +248,21 @@ static void ftrace_update_pid_func(void) # error Dynamic ftrace depends on MCOUNT_RECORD #endif +/* set when tracing only a pid */ +struct pid *ftrace_pid_trace; +static struct pid * const ftrace_swapper_pid = &init_struct_pid; +static struct hlist_head ftrace_func_hash[FTRACE_FUNC_HASHSIZE] __read_mostly; + +struct ftrace_func_hook { + struct hlist_node node; + struct ftrace_hook_ops *ops; + unsigned long flags; + unsigned long ip; + void *data; + struct rcu_head rcu; +}; + + enum { FTRACE_ENABLE_CALLS = (1 << 0), FTRACE_DISABLE_CALLS = (1 << 1), @@ -750,12 +765,14 @@ enum { FTRACE_ITER_NOTRACE = (1 << 2), FTRACE_ITER_FAILURES = (1 << 3), FTRACE_ITER_PRINTALL = (1 << 4), + FTRACE_ITER_HASH = (1 << 5), }; #define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */ struct ftrace_iterator { struct ftrace_page *pg; + int hidx; int idx; unsigned flags; unsigned char buffer[FTRACE_BUFF_MAX+1]; @@ -763,18 +780,87 @@ struct ftrace_iterator { unsigned filtered; }; +static void * +t_hash_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct ftrace_iterator *iter = m->private; + struct hlist_node *hnd = v; + struct hlist_head *hhd; + + WARN_ON(!(iter->flags & FTRACE_ITER_HASH)); + + (*pos)++; + + retry: + if (iter->hidx >= FTRACE_FUNC_HASHSIZE) + return NULL; + + hhd = &ftrace_func_hash[iter->hidx]; + + if (hlist_empty(hhd)) { + iter->hidx++; + hnd = NULL; + goto retry; + } + + if (!hnd) + hnd = hhd->first; + else { + hnd = hnd->next; + if (!hnd) { + iter->hidx++; + goto retry; + } + } + + return hnd; +} + +static void *t_hash_start(struct seq_file *m, loff_t *pos) +{ + struct ftrace_iterator *iter = m->private; + void *p = NULL; + + iter->flags |= FTRACE_ITER_HASH; + + return t_hash_next(m, p, pos); +} + +static int t_hash_show(struct seq_file *m, void *v) +{ + struct ftrace_func_hook *rec; + struct hlist_node *hnd = v; + char str[KSYM_SYMBOL_LEN]; + + rec = hlist_entry(hnd, struct ftrace_func_hook, node); + + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); + seq_printf(m, "%s:", str); + + kallsyms_lookup((unsigned long)rec->ops->func, NULL, NULL, NULL, str); + seq_printf(m, "%s", str); + + if (rec->data) + seq_printf(m, ":%p", rec->data); + seq_putc(m, '\n'); + + return 0; +} + static void * t_next(struct seq_file *m, void *v, loff_t *pos) { struct ftrace_iterator *iter = m->private; struct dyn_ftrace *rec = NULL; + if (iter->flags & FTRACE_ITER_HASH) + return t_hash_next(m, v, pos); + (*pos)++; if (iter->flags & FTRACE_ITER_PRINTALL) return NULL; - mutex_lock(&ftrace_lock); retry: if (iter->idx >= iter->pg->index) { if (iter->pg->next) { @@ -803,7 +889,6 @@ t_next(struct seq_file *m, void *v, loff_t *pos) goto retry; } } - mutex_unlock(&ftrace_lock); return rec; } @@ -813,6 +898,7 @@ static void *t_start(struct seq_file *m, loff_t *pos) struct ftrace_iterator *iter = m->private; void *p = NULL; + mutex_lock(&ftrace_lock); /* * For set_ftrace_filter reading, if we have the filter * off, we can short cut and just print out that all @@ -820,12 +906,15 @@ static void *t_start(struct seq_file *m, loff_t *pos) */ if (iter->flags & FTRACE_ITER_FILTER && !ftrace_filtered) { if (*pos > 0) - return NULL; + return t_hash_start(m, pos); iter->flags |= FTRACE_ITER_PRINTALL; (*pos)++; return iter; } + if (iter->flags & FTRACE_ITER_HASH) + return t_hash_start(m, pos); + if (*pos > 0) { if (iter->idx < 0) return p; @@ -835,11 +924,15 @@ static void *t_start(struct seq_file *m, loff_t *pos) p = t_next(m, p, pos); + if (!p) + return t_hash_start(m, pos); + return p; } static void t_stop(struct seq_file *m, void *p) { + mutex_unlock(&ftrace_lock); } static int t_show(struct seq_file *m, void *v) @@ -848,6 +941,9 @@ static int t_show(struct seq_file *m, void *v) struct dyn_ftrace *rec = v; char str[KSYM_SYMBOL_LEN]; + if (iter->flags & FTRACE_ITER_HASH) + return t_hash_show(m, v); + if (iter->flags & FTRACE_ITER_PRINTALL) { seq_printf(m, "#### all functions enabled ####\n"); return 0; @@ -1246,19 +1342,6 @@ static int __init ftrace_mod_cmd_init(void) } device_initcall(ftrace_mod_cmd_init); -#define FTRACE_HASH_BITS 7 -#define FTRACE_FUNC_HASHSIZE (1 << FTRACE_HASH_BITS) -static struct hlist_head ftrace_func_hash[FTRACE_FUNC_HASHSIZE] __read_mostly; - -struct ftrace_func_hook { - struct hlist_node node; - struct ftrace_hook_ops *ops; - unsigned long flags; - unsigned long ip; - void *data; - struct rcu_head rcu; -}; - static void function_trace_hook_call(unsigned long ip, unsigned long parent_ip) { From 809dcf29ce4e1723709910878e050bd187617e0e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 16 Feb 2009 23:06:01 -0500 Subject: [PATCH 154/413] ftrace: add pretty print to selected fuction traces This patch adds a call back for the tracers that have hooks to selected functions. This allows the tracer to show better output in the set_ftrace_filter file. Signed-off-by: Steven Rostedt --- include/linux/ftrace.h | 6 ++++++ kernel/trace/ftrace.c | 3 +++ 2 files changed, 9 insertions(+) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 13918c4400ad..b331e216d8a1 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -106,12 +106,18 @@ struct ftrace_func_command { /* asm/ftrace.h must be defined for archs supporting dynamic ftrace */ #include +struct seq_file; + struct ftrace_hook_ops { void (*func)(unsigned long ip, unsigned long parent_ip, void **data); int (*callback)(unsigned long ip, void **data); void (*free)(void **data); + int (*print)(struct seq_file *m, + unsigned long ip, + struct ftrace_hook_ops *ops, + void *data); }; extern int diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 1e058848cddb..6533c1d20155 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -834,6 +834,9 @@ static int t_hash_show(struct seq_file *m, void *v) rec = hlist_entry(hnd, struct ftrace_func_hook, node); + if (rec->ops->print) + return rec->ops->print(m, rec->ip, rec->ops, rec->data); + kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); seq_printf(m, "%s:", str); From e110e3d1eaa0f9628918be67ddd32e8ad65a2871 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 16 Feb 2009 23:38:13 -0500 Subject: [PATCH 155/413] ftrace: add pretty print function for traceon and traceoff hooks This patch adds a pretty print version of traceon and traceoff output for set_ftrace_filter. # echo 'sys_open:traceon:4' > set_ftrace_filter # cat set_ftrace_filter #### all functions enabled #### sys_open:traceon:count=4 Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions.c | 28 ++++++++++++++++++++++++++++ 1 file changed, 28 insertions(+) diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 5c95708b9dc3..f520aa419dff 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -267,14 +267,42 @@ ftrace_traceoff(unsigned long ip, unsigned long parent_ip, void **data) tracing_off(); } +static int +ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, + struct ftrace_hook_ops *ops, void *data); + static struct ftrace_hook_ops traceon_hook_ops = { .func = ftrace_traceon, + .print = ftrace_trace_onoff_print, }; static struct ftrace_hook_ops traceoff_hook_ops = { .func = ftrace_traceoff, + .print = ftrace_trace_onoff_print, }; +static int +ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, + struct ftrace_hook_ops *ops, void *data) +{ + char str[KSYM_SYMBOL_LEN]; + long count = (long)data; + + kallsyms_lookup(ip, NULL, NULL, NULL, str); + seq_printf(m, "%s:", str); + + if (ops == &traceon_hook_ops) + seq_printf(m, "traceon"); + else + seq_printf(m, "traceoff"); + + if (count != -1) + seq_printf(m, ":count=%ld", count); + seq_putc(m, '\n'); + + return 0; +} + static int ftrace_trace_onoff_unreg(char *glob, char *cmd, char *param) { From 97d0bb8dcd8c2812e1927cdb51d7b1f9c98352b5 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 17 Feb 2009 11:47:39 +0100 Subject: [PATCH 156/413] ftrace: fix !CONFIG_FTRACE [un_]register_ftrace_command() prototypes Impact: build fix Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index b331e216d8a1..63281228ce3e 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -238,9 +238,11 @@ extern void ftrace_enable_daemon(void); static inline void ftrace_release(void *start, unsigned long size) { } static inline int register_ftrace_command(struct ftrace_func_command *cmd) { + return -EINVAL; } static inline int unregister_ftrace_command(char *cmd_name) { + return -EINVAL; } #endif /* CONFIG_DYNAMIC_FTRACE */ From 73d3fd96e77745742f3750b7b19ee42204adc210 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 17 Feb 2009 11:48:18 +0100 Subject: [PATCH 157/413] ftrace: fix !CONFIG_DYNAMIC_FTRACE ftrace_swapper_pid definition Impact: build fix Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 6533c1d20155..4e6c87ecf1bf 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -243,14 +243,16 @@ static void ftrace_update_pid_func(void) mutex_unlock(&ftrace_lock); } +/* set when tracing only a pid */ +struct pid *ftrace_pid_trace; +static struct pid * const ftrace_swapper_pid = &init_struct_pid; + #ifdef CONFIG_DYNAMIC_FTRACE + #ifndef CONFIG_FTRACE_MCOUNT_RECORD # error Dynamic ftrace depends on MCOUNT_RECORD #endif -/* set when tracing only a pid */ -struct pid *ftrace_pid_trace; -static struct pid * const ftrace_swapper_pid = &init_struct_pid; static struct hlist_head ftrace_func_hash[FTRACE_FUNC_HASHSIZE] __read_mostly; struct ftrace_func_hook { From 6a24a244cd3a02d5b290293c32fcf2c6e92b4235 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Feb 2009 11:20:26 -0500 Subject: [PATCH 158/413] ftrace: clean up coding style Ingo Molnar pointed out some coding style issues with the recent ftrace updates. This patch cleans them up. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 24 +++++++++++++----------- kernel/trace/trace_functions.c | 1 - 2 files changed, 13 insertions(+), 12 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 4e6c87ecf1bf..af9d95c0e4de 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -460,8 +460,8 @@ static void ftrace_bug(int failed, unsigned long ip) static int __ftrace_replace_code(struct dyn_ftrace *rec, int enable) { - unsigned long ip, fl; unsigned long ftrace_addr; + unsigned long ip, fl; ftrace_addr = (unsigned long)FTRACE_ADDR; @@ -530,9 +530,9 @@ __ftrace_replace_code(struct dyn_ftrace *rec, int enable) static void ftrace_replace_code(int enable) { - int failed; struct dyn_ftrace *rec; struct ftrace_page *pg; + int failed; do_for_each_ftrace_rec(pg, rec) { /* @@ -1208,14 +1208,15 @@ ftrace_match_record(struct dyn_ftrace *rec, char *regex, int len, int type) static void ftrace_match_records(char *buff, int len, int enable) { - char *search; + unsigned int search_len; struct ftrace_page *pg; struct dyn_ftrace *rec; + unsigned long flag; + char *search; int type; - unsigned long flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; - unsigned search_len; int not; + flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; type = ftrace_setup_glob(buff, len, &search, ¬); search_len = strlen(search); @@ -1263,14 +1264,16 @@ ftrace_match_module_record(struct dyn_ftrace *rec, char *mod, static void ftrace_match_module_records(char *buff, char *mod, int enable) { - char *search = buff; + unsigned search_len = 0; struct ftrace_page *pg; struct dyn_ftrace *rec; int type = MATCH_FULL; - unsigned long flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; - unsigned search_len = 0; + char *search = buff; + unsigned long flag; int not = 0; + flag = enable ? FTRACE_FL_FILTER : FTRACE_FL_NOTRACE; + /* blank or '*' mean the same */ if (strcmp(buff, "*") == 0) buff[0] = 0; @@ -1442,8 +1445,8 @@ register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, struct ftrace_func_hook *entry; struct ftrace_page *pg; struct dyn_ftrace *rec; - unsigned long key; int type, len, not; + unsigned long key; int count = 0; char *search; @@ -1623,8 +1626,8 @@ int unregister_ftrace_command(struct ftrace_func_command *cmd) static int ftrace_process_regex(char *buff, int len, int enable) { - struct ftrace_func_command *p; char *func, *command, *next = buff; + struct ftrace_func_command *p; int ret = -EINVAL; func = strsep(&next, ":"); @@ -2392,7 +2395,6 @@ static __init int ftrace_init_debugfs(void) "'set_ftrace_pid' entry\n"); return 0; } - fs_initcall(ftrace_init_debugfs); /** diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index f520aa419dff..021a574c5988 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -397,6 +397,5 @@ static __init int init_function_trace(void) init_func_cmd_traceon(); return register_tracer(&function_trace); } - device_initcall(init_function_trace); From b6887d7916e44c1d8913084fb6aa5004d9473f1a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Feb 2009 12:32:04 -0500 Subject: [PATCH 159/413] ftrace: rename _hook to _probe Impact: clean up Ingo Molnar did not like the _hook naming convention used by the select function tracer. Luis Claudio R. Goncalves suggested using the "_probe" extension. This patch implements the change of calling the functions and variables "_hook" and replacing them with "_probe". Signed-off-by: Steven Rostedt --- include/linux/ftrace.h | 12 +++--- kernel/trace/ftrace.c | 78 +++++++++++++++++----------------- kernel/trace/trace_functions.c | 26 ++++++------ 3 files changed, 58 insertions(+), 58 deletions(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 63281228ce3e..9d224c43e634 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -108,7 +108,7 @@ struct ftrace_func_command { struct seq_file; -struct ftrace_hook_ops { +struct ftrace_probe_ops { void (*func)(unsigned long ip, unsigned long parent_ip, void **data); @@ -116,19 +116,19 @@ struct ftrace_hook_ops { void (*free)(void **data); int (*print)(struct seq_file *m, unsigned long ip, - struct ftrace_hook_ops *ops, + struct ftrace_probe_ops *ops, void *data); }; extern int -register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, +register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, void *data); extern void -unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, +unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, void *data); extern void -unregister_ftrace_function_hook_func(char *glob, struct ftrace_hook_ops *ops); -extern void unregister_ftrace_function_hook_all(char *glob); +unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops); +extern void unregister_ftrace_function_probe_all(char *glob); enum { FTRACE_FL_FREE = (1 << 0), diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index af9d95c0e4de..330a059f6ed7 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -255,9 +255,9 @@ static struct pid * const ftrace_swapper_pid = &init_struct_pid; static struct hlist_head ftrace_func_hash[FTRACE_FUNC_HASHSIZE] __read_mostly; -struct ftrace_func_hook { +struct ftrace_func_probe { struct hlist_node node; - struct ftrace_hook_ops *ops; + struct ftrace_probe_ops *ops; unsigned long flags; unsigned long ip; void *data; @@ -830,11 +830,11 @@ static void *t_hash_start(struct seq_file *m, loff_t *pos) static int t_hash_show(struct seq_file *m, void *v) { - struct ftrace_func_hook *rec; + struct ftrace_func_probe *rec; struct hlist_node *hnd = v; char str[KSYM_SYMBOL_LEN]; - rec = hlist_entry(hnd, struct ftrace_func_hook, node); + rec = hlist_entry(hnd, struct ftrace_func_probe, node); if (rec->ops->print) return rec->ops->print(m, rec->ip, rec->ops, rec->data); @@ -1351,9 +1351,9 @@ static int __init ftrace_mod_cmd_init(void) device_initcall(ftrace_mod_cmd_init); static void -function_trace_hook_call(unsigned long ip, unsigned long parent_ip) +function_trace_probe_call(unsigned long ip, unsigned long parent_ip) { - struct ftrace_func_hook *entry; + struct ftrace_func_probe *entry; struct hlist_head *hhd; struct hlist_node *n; unsigned long key; @@ -1379,18 +1379,18 @@ function_trace_hook_call(unsigned long ip, unsigned long parent_ip) ftrace_preempt_enable(resched); } -static struct ftrace_ops trace_hook_ops __read_mostly = +static struct ftrace_ops trace_probe_ops __read_mostly = { - .func = function_trace_hook_call, + .func = function_trace_probe_call, }; -static int ftrace_hook_registered; +static int ftrace_probe_registered; -static void __enable_ftrace_function_hook(void) +static void __enable_ftrace_function_probe(void) { int i; - if (ftrace_hook_registered) + if (ftrace_probe_registered) return; for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { @@ -1402,16 +1402,16 @@ static void __enable_ftrace_function_hook(void) if (i == FTRACE_FUNC_HASHSIZE) return; - __register_ftrace_function(&trace_hook_ops); + __register_ftrace_function(&trace_probe_ops); ftrace_startup(0); - ftrace_hook_registered = 1; + ftrace_probe_registered = 1; } -static void __disable_ftrace_function_hook(void) +static void __disable_ftrace_function_probe(void) { int i; - if (!ftrace_hook_registered) + if (!ftrace_probe_registered) return; for (i = 0; i < FTRACE_FUNC_HASHSIZE; i++) { @@ -1421,16 +1421,16 @@ static void __disable_ftrace_function_hook(void) } /* no more funcs left */ - __unregister_ftrace_function(&trace_hook_ops); + __unregister_ftrace_function(&trace_probe_ops); ftrace_shutdown(0); - ftrace_hook_registered = 0; + ftrace_probe_registered = 0; } static void ftrace_free_entry_rcu(struct rcu_head *rhp) { - struct ftrace_func_hook *entry = - container_of(rhp, struct ftrace_func_hook, rcu); + struct ftrace_func_probe *entry = + container_of(rhp, struct ftrace_func_probe, rcu); if (entry->ops->free) entry->ops->free(&entry->data); @@ -1439,10 +1439,10 @@ static void ftrace_free_entry_rcu(struct rcu_head *rhp) int -register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, +register_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, void *data) { - struct ftrace_func_hook *entry; + struct ftrace_func_probe *entry; struct ftrace_page *pg; struct dyn_ftrace *rec; int type, len, not; @@ -1453,7 +1453,7 @@ register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, type = ftrace_setup_glob(glob, strlen(glob), &search, ¬); len = strlen(search); - /* we do not support '!' for function hooks */ + /* we do not support '!' for function probes */ if (WARN_ON(not)) return -EINVAL; @@ -1468,7 +1468,7 @@ register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, entry = kmalloc(sizeof(*entry), GFP_KERNEL); if (!entry) { - /* If we did not hook to any, then return error */ + /* If we did not process any, then return error */ if (!count) count = -ENOMEM; goto out_unlock; @@ -1498,7 +1498,7 @@ register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, hlist_add_head_rcu(&entry->node, &ftrace_func_hash[key]); } while_for_each_ftrace_rec(); - __enable_ftrace_function_hook(); + __enable_ftrace_function_probe(); out_unlock: mutex_unlock(&ftrace_lock); @@ -1507,15 +1507,15 @@ register_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, } enum { - HOOK_TEST_FUNC = 1, - HOOK_TEST_DATA = 2 + PROBE_TEST_FUNC = 1, + PROBE_TEST_DATA = 2 }; static void -__unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, +__unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, void *data, int flags) { - struct ftrace_func_hook *entry; + struct ftrace_func_probe *entry; struct hlist_node *n, *tmp; char str[KSYM_SYMBOL_LEN]; int type = MATCH_FULL; @@ -1530,7 +1530,7 @@ __unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, type = ftrace_setup_glob(glob, strlen(glob), &search, ¬); len = strlen(search); - /* we do not support '!' for function hooks */ + /* we do not support '!' for function probes */ if (WARN_ON(not)) return; } @@ -1542,10 +1542,10 @@ __unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, hlist_for_each_entry_safe(entry, n, tmp, hhd, node) { /* break up if statements for readability */ - if ((flags & HOOK_TEST_FUNC) && entry->ops != ops) + if ((flags & PROBE_TEST_FUNC) && entry->ops != ops) continue; - if ((flags & HOOK_TEST_DATA) && entry->data != data) + if ((flags & PROBE_TEST_DATA) && entry->data != data) continue; /* do this last, since it is the most expensive */ @@ -1560,27 +1560,27 @@ __unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, call_rcu(&entry->rcu, ftrace_free_entry_rcu); } } - __disable_ftrace_function_hook(); + __disable_ftrace_function_probe(); mutex_unlock(&ftrace_lock); } void -unregister_ftrace_function_hook(char *glob, struct ftrace_hook_ops *ops, +unregister_ftrace_function_probe(char *glob, struct ftrace_probe_ops *ops, void *data) { - __unregister_ftrace_function_hook(glob, ops, data, - HOOK_TEST_FUNC | HOOK_TEST_DATA); + __unregister_ftrace_function_probe(glob, ops, data, + PROBE_TEST_FUNC | PROBE_TEST_DATA); } void -unregister_ftrace_function_hook_func(char *glob, struct ftrace_hook_ops *ops) +unregister_ftrace_function_probe_func(char *glob, struct ftrace_probe_ops *ops) { - __unregister_ftrace_function_hook(glob, ops, NULL, HOOK_TEST_FUNC); + __unregister_ftrace_function_probe(glob, ops, NULL, PROBE_TEST_FUNC); } -void unregister_ftrace_function_hook_all(char *glob) +void unregister_ftrace_function_probe_all(char *glob) { - __unregister_ftrace_function_hook(glob, NULL, NULL, 0); + __unregister_ftrace_function_probe(glob, NULL, NULL, 0); } static LIST_HEAD(ftrace_commands); diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 021a574c5988..6ea73ed03bfa 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -269,21 +269,21 @@ ftrace_traceoff(unsigned long ip, unsigned long parent_ip, void **data) static int ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, - struct ftrace_hook_ops *ops, void *data); + struct ftrace_probe_ops *ops, void *data); -static struct ftrace_hook_ops traceon_hook_ops = { +static struct ftrace_probe_ops traceon_probe_ops = { .func = ftrace_traceon, .print = ftrace_trace_onoff_print, }; -static struct ftrace_hook_ops traceoff_hook_ops = { +static struct ftrace_probe_ops traceoff_probe_ops = { .func = ftrace_traceoff, .print = ftrace_trace_onoff_print, }; static int ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, - struct ftrace_hook_ops *ops, void *data) + struct ftrace_probe_ops *ops, void *data) { char str[KSYM_SYMBOL_LEN]; long count = (long)data; @@ -291,7 +291,7 @@ ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, kallsyms_lookup(ip, NULL, NULL, NULL, str); seq_printf(m, "%s:", str); - if (ops == &traceon_hook_ops) + if (ops == &traceon_probe_ops) seq_printf(m, "traceon"); else seq_printf(m, "traceoff"); @@ -306,15 +306,15 @@ ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, static int ftrace_trace_onoff_unreg(char *glob, char *cmd, char *param) { - struct ftrace_hook_ops *ops; + struct ftrace_probe_ops *ops; /* we register both traceon and traceoff to this callback */ if (strcmp(cmd, "traceon") == 0) - ops = &traceon_hook_ops; + ops = &traceon_probe_ops; else - ops = &traceoff_hook_ops; + ops = &traceoff_probe_ops; - unregister_ftrace_function_hook_func(glob, ops); + unregister_ftrace_function_probe_func(glob, ops); return 0; } @@ -322,7 +322,7 @@ ftrace_trace_onoff_unreg(char *glob, char *cmd, char *param) static int ftrace_trace_onoff_callback(char *glob, char *cmd, char *param, int enable) { - struct ftrace_hook_ops *ops; + struct ftrace_probe_ops *ops; void *count = (void *)-1; char *number; int ret; @@ -336,9 +336,9 @@ ftrace_trace_onoff_callback(char *glob, char *cmd, char *param, int enable) /* we register both traceon and traceoff to this callback */ if (strcmp(cmd, "traceon") == 0) - ops = &traceon_hook_ops; + ops = &traceon_probe_ops; else - ops = &traceoff_hook_ops; + ops = &traceoff_probe_ops; if (!param) goto out_reg; @@ -357,7 +357,7 @@ ftrace_trace_onoff_callback(char *glob, char *cmd, char *param, int enable) return ret; out_reg: - ret = register_ftrace_function_hook(glob, ops, count); + ret = register_ftrace_function_probe(glob, ops, count); return ret; } From af513098452b8887d7c0e15a39d7cb74479501bd Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Tue, 17 Feb 2009 01:07:28 -0500 Subject: [PATCH 160/413] tracing: use the more proper parameter Pass tsk to tracing_record_cmdline instead of current. Signed-off-by: Wenji Huang Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 95f99a7abf2f..dc61e82faad9 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -336,7 +336,7 @@ __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) data->rt_priority = tsk->rt_priority; /* record this tasks comm */ - tracing_record_cmdline(current); + tracing_record_cmdline(tsk); } static void From d2ef7c2f0f9ab48c25eafc0ebad0df5f7930420b Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Tue, 17 Feb 2009 01:09:47 -0500 Subject: [PATCH 161/413] tracing: fix the return value of trace selftest This patch is to fix the return value of trace_selftest_startup_sysprof and trace_selftest_startup_branch on failure. Signed-off-by: Wenji Huang Signed-off-by: Steven Rostedt --- kernel/trace/trace_selftest.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 0c9aa1457e51..c72e749bcbef 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -622,7 +622,7 @@ trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr) ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); - return 0; + return ret; } /* Sleep for a 1/10 of a second */ @@ -634,6 +634,11 @@ trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr) trace->reset(tr); tracing_start(); + if (!ret && !count) { + printk(KERN_CONT ".. no entries found .."); + ret = -1; + } + return ret; } #endif /* CONFIG_SYSPROF_TRACER */ @@ -661,6 +666,11 @@ trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr) trace->reset(tr); tracing_start(); + if (!ret && !count) { + printk(KERN_CONT ".. no entries found .."); + ret = -1; + } + return ret; } #endif /* CONFIG_BRANCH_TRACER */ From 73d8b8bc4f24a97a406d09c8268ac019f4ac661e Mon Sep 17 00:00:00 2001 From: Wenji Huang Date: Tue, 17 Feb 2009 01:10:02 -0500 Subject: [PATCH 162/413] tracing: fix typing mistake in hint message and comments Impact: cleanup Fix incorrect hint message in code and typos in comments. Signed-off-by: Wenji Huang Signed-off-by: Steven Rostedt --- kernel/trace/trace_irqsoff.c | 2 +- kernel/trace/trace_sched_switch.c | 2 +- kernel/trace/trace_sched_wakeup.c | 2 +- kernel/trace/trace_selftest.c | 4 ++-- kernel/trace/trace_stat.c | 2 +- kernel/trace/trace_sysprof.c | 2 +- 6 files changed, 7 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index c6b442d88de8..9e5ebd844158 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -1,5 +1,5 @@ /* - * trace irqs off criticall timings + * trace irqs off critical timings * * Copyright (C) 2007-2008 Steven Rostedt * Copyright (C) 2008 Ingo Molnar diff --git a/kernel/trace/trace_sched_switch.c b/kernel/trace/trace_sched_switch.c index 30e14fe85896..82fbb5a2df89 100644 --- a/kernel/trace/trace_sched_switch.c +++ b/kernel/trace/trace_sched_switch.c @@ -93,7 +93,7 @@ static int tracing_sched_register(void) ret = register_trace_sched_switch(probe_sched_switch); if (ret) { pr_info("sched trace: Couldn't activate tracepoint" - " probe to kernel_sched_schedule\n"); + " probe to kernel_sched_switch\n"); goto fail_deprobe_wake_new; } diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 96d716485898..276c51aaf314 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -284,7 +284,7 @@ static void start_wakeup_tracer(struct trace_array *tr) ret = register_trace_sched_switch(probe_wakeup_sched_switch); if (ret) { pr_info("sched trace: Couldn't activate tracepoint" - " probe to kernel_sched_schedule\n"); + " probe to kernel_sched_switch\n"); goto fail_deprobe_wake_new; } diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index c72e749bcbef..01415f4edaa5 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -107,9 +107,9 @@ int trace_selftest_startup_dynamic_tracing(struct tracer *trace, func(); /* - * Some archs *cough*PowerPC*cough* add charachters to the + * Some archs *cough*PowerPC*cough* add characters to the * start of the function names. We simply put a '*' to - * accomodate them. + * accommodate them. */ func_name = "*" STR(DYN_FTRACE_TEST_NAME); diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index eae9cef39291..39310e3434ee 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -30,7 +30,7 @@ struct tracer_stat_session { struct dentry *file; }; -/* All of the sessions currently in use. Each stat file embeed one session */ +/* All of the sessions currently in use. Each stat file embed one session */ static LIST_HEAD(all_stat_sessions); static DEFINE_MUTEX(all_stat_sessions_mutex); diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c index 7c9a2d82a7d8..c771af4e8f1a 100644 --- a/kernel/trace/trace_sysprof.c +++ b/kernel/trace/trace_sysprof.c @@ -327,5 +327,5 @@ void init_tracer_sysprof_debugfs(struct dentry *d_tracer) d_tracer, NULL, &sysprof_sample_fops); if (entry) return; - pr_warning("Could not create debugfs 'dyn_ftrace_total_info' entry\n"); + pr_warning("Could not create debugfs 'sysprof_sample_period' entry\n"); } From 35ebf1caa4854ad5ba25f3a72967acc064147994 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Feb 2009 13:12:12 -0500 Subject: [PATCH 163/413] ftrace: show unlimited when traceon or traceoff has no counter Impact: clean up The traceon and traceoff function probes are confusing to developers to what happens when a counter is not specified. This should help clear things up. # echo "*:traceoff" > set_ftrace_filter # cat /debug/tracing/set_ftrace_filter #### all functions enabled #### do_fork:traceoff:unlimited Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 6ea73ed03bfa..4c113a8c466f 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -296,7 +296,9 @@ ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip, else seq_printf(m, "traceoff"); - if (count != -1) + if (count == -1) + seq_printf(m, ":unlimited\n"); + else seq_printf(m, ":count=%ld", count); seq_putc(m, '\n'); From 6eaaa5d57e76c454479833fc8594cd7c3b75c789 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 11 Feb 2009 02:25:00 +0100 Subject: [PATCH 164/413] tracing/core: use appropriate waiting on trace_pipe Impact: api and pipe waiting change Currently, the waiting used in tracing_read_pipe() is done through a 100 msecs schedule_timeout() loop which periodically check if there are traces on the buffer. This can cause small latencies for programs which are reading the incoming events. This patch makes the reader waiting for the trace_wait waitqueue except for few tracers such as the sched and functions tracers which might be already hold the runqueue lock while waking up the reader. This is performed through a new callback wait_pipe() on struct tracer. If none is implemented on a specific tracer, the default waiting for trace_wait queue is attached. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 62 ++++++++++++++++++---------- kernel/trace/trace.h | 25 +++++++++-- kernel/trace/trace_functions.c | 1 + kernel/trace/trace_functions_graph.c | 1 + kernel/trace/trace_sched_switch.c | 1 + kernel/trace/trace_sched_wakeup.c | 1 + 6 files changed, 67 insertions(+), 24 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index dc61e82faad9..881a94474d79 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -499,6 +499,9 @@ __acquires(kernel_lock) else if (!type->flags->opts) type->flags->opts = dummy_tracer_opt; + if (!type->wait_pipe) + type->wait_pipe = default_wait_pipe; + #ifdef CONFIG_FTRACE_STARTUP_TEST if (type->selftest && !tracing_selftest_disabled) { @@ -1064,7 +1067,10 @@ tracing_sched_wakeup_trace(struct trace_array *tr, entry->next_prio = wakee->prio; entry->next_state = wakee->state; entry->next_cpu = task_cpu(wakee); - trace_buffer_unlock_commit(tr, event, flags, pc); + + ring_buffer_unlock_commit(tr->buffer, event); + ftrace_trace_stack(tr, flags, 6, pc); + ftrace_trace_userstack(tr, flags, pc); } void @@ -2392,6 +2398,38 @@ tracing_poll_pipe(struct file *filp, poll_table *poll_table) } } + +void default_wait_pipe(struct trace_iterator *iter) +{ + DEFINE_WAIT(wait); + + prepare_to_wait(&trace_wait, &wait, TASK_INTERRUPTIBLE); + + if (trace_empty(iter)) + schedule(); + + finish_wait(&trace_wait, &wait); +} + +/* + * This is a make-shift waitqueue. + * A tracer might use this callback on some rare cases: + * + * 1) the current tracer might hold the runqueue lock when it wakes up + * a reader, hence a deadlock (sched, function, and function graph tracers) + * 2) the function tracers, trace all functions, we don't want + * the overhead of calling wake_up and friends + * (and tracing them too) + * + * Anyway, this is really very primitive wakeup. + */ +void poll_wait_pipe(struct trace_iterator *iter) +{ + set_current_state(TASK_INTERRUPTIBLE); + /* sleep for 100 msecs, and try again. */ + schedule_timeout(HZ / 10); +} + /* Must be called with trace_types_lock mutex held. */ static int tracing_wait_pipe(struct file *filp) { @@ -2403,30 +2441,14 @@ static int tracing_wait_pipe(struct file *filp) return -EAGAIN; } - /* - * This is a make-shift waitqueue. The reason we don't use - * an actual wait queue is because: - * 1) we only ever have one waiter - * 2) the tracing, traces all functions, we don't want - * the overhead of calling wake_up and friends - * (and tracing them too) - * Anyway, this is really very primitive wakeup. - */ - set_current_state(TASK_INTERRUPTIBLE); - iter->tr->waiter = current; - mutex_unlock(&trace_types_lock); - /* sleep for 100 msecs, and try again. */ - schedule_timeout(HZ/10); + iter->trace->wait_pipe(iter); mutex_lock(&trace_types_lock); - iter->tr->waiter = NULL; - - if (signal_pending(current)) { + if (signal_pending(current)) return -EINTR; - } if (iter->trace != current_trace) return 0; @@ -2442,8 +2464,6 @@ static int tracing_wait_pipe(struct file *filp) */ if (!tracer_enabled && iter->pos) break; - - continue; } return 1; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index dbff0207b213..eed732c151fc 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -337,18 +337,34 @@ struct tracer_flags { #define TRACER_OPT(s, b) .name = #s, .bit = b -/* - * A specific tracer, represented by methods that operate on a trace array: +/** + * struct tracer - a specific tracer and its callbacks to interact with debugfs + * @name: the name chosen to select it on the available_tracers file + * @init: called when one switches to this tracer (echo name > current_tracer) + * @reset: called when one switches to another tracer + * @start: called when tracing is unpaused (echo 1 > tracing_enabled) + * @stop: called when tracing is paused (echo 0 > tracing_enabled) + * @open: called when the trace file is opened + * @pipe_open: called when the trace_pipe file is opened + * @wait_pipe: override how the user waits for traces on trace_pipe + * @close: called when the trace file is released + * @read: override the default read callback on trace_pipe + * @splice_read: override the default splice_read callback on trace_pipe + * @selftest: selftest to run on boot (see trace_selftest.c) + * @print_headers: override the first lines that describe your columns + * @print_line: callback that prints a trace + * @set_flag: signals one of your private flags changed (trace_options file) + * @flags: your private flags */ struct tracer { const char *name; - /* Your tracer should raise a warning if init fails */ int (*init)(struct trace_array *tr); void (*reset)(struct trace_array *tr); void (*start)(struct trace_array *tr); void (*stop)(struct trace_array *tr); void (*open)(struct trace_iterator *iter); void (*pipe_open)(struct trace_iterator *iter); + void (*wait_pipe)(struct trace_iterator *iter); void (*close)(struct trace_iterator *iter); ssize_t (*read)(struct trace_iterator *iter, struct file *filp, char __user *ubuf, @@ -432,6 +448,9 @@ void tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, int pc); +void default_wait_pipe(struct trace_iterator *iter); +void poll_wait_pipe(struct trace_iterator *iter); + void ftrace(struct trace_array *tr, struct trace_array_cpu *data, unsigned long ip, diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c index 4c113a8c466f..c9a0b7df44ff 100644 --- a/kernel/trace/trace_functions.c +++ b/kernel/trace/trace_functions.c @@ -225,6 +225,7 @@ static struct tracer function_trace __read_mostly = .init = function_trace_init, .reset = function_trace_reset, .start = function_trace_start, + .wait_pipe = poll_wait_pipe, .flags = &func_flags, .set_flag = func_set_flag, #ifdef CONFIG_FTRACE_SELFTEST diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 519a0cab1530..0ff5cb661900 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -757,6 +757,7 @@ static struct tracer graph_trace __read_mostly = { .name = "function_graph", .open = graph_trace_open, .close = graph_trace_close, + .wait_pipe = poll_wait_pipe, .init = graph_trace_init, .reset = graph_trace_reset, .print_line = print_graph_function, diff --git a/kernel/trace/trace_sched_switch.c b/kernel/trace/trace_sched_switch.c index 82fbb5a2df89..77132c2cf3d9 100644 --- a/kernel/trace/trace_sched_switch.c +++ b/kernel/trace/trace_sched_switch.c @@ -221,6 +221,7 @@ static struct tracer sched_switch_trace __read_mostly = .reset = sched_switch_trace_reset, .start = sched_switch_trace_start, .stop = sched_switch_trace_stop, + .wait_pipe = poll_wait_pipe, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_sched_switch, #endif diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index 276c51aaf314..db55f7aaa640 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -380,6 +380,7 @@ static struct tracer wakeup_rt_tracer __read_mostly = .reset = wakeup_tracer_reset, .start = wakeup_tracer_start, .stop = wakeup_tracer_stop, + .wait_pipe = poll_wait_pipe, .print_max = 1, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_wakeup, From fa7c7f6e11f70d62505074a8b30a776236850dec Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 11 Feb 2009 02:51:30 +0100 Subject: [PATCH 165/413] tracing/core: remove unused parameter in tracing_fill_pipe_page() Impact: cleanup The struct page *pages parameter is unused. Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 881a94474d79..e1f3b99a2e52 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2571,8 +2571,7 @@ static struct pipe_buf_operations tracing_pipe_buf_ops = { }; static size_t -tracing_fill_pipe_page(struct page *pages, size_t rem, - struct trace_iterator *iter) +tracing_fill_pipe_page(size_t rem, struct trace_iterator *iter) { size_t count; int ret; @@ -2649,7 +2648,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, if (!pages[i]) break; - rem = tracing_fill_pipe_page(pages[i], rem, iter); + rem = tracing_fill_pipe_page(rem, iter); /* Copy the data into the page, so we can start over. */ ret = trace_seq_to_buffer(&iter->seq, From d1f9cbd78841f1a797c77e9117e4882f932c2ef6 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 18 Feb 2009 04:25:25 +0100 Subject: [PATCH 166/413] tracing/function-graph-tracer: fix traces weirdness while absolute time printing Impact: trace output cleanup/reordering When an interrupt occurs and and the abstime option is selected: echo funcgraph-abstime > /debug/tracing/trace_options then we observe broken traces: 30581.025422 | 0) Xorg-4291 | 0.503 us | idle_cpu(); 30581.025424 | 0) Xorg-4291 | 2.576 us | } 30581.025424 | 0) Xorg-4291 | + 75.771 us | } 0) Xorg-4291 | <========== | 30581.025425 | 0) Xorg-4291 | | schedule() { 30581.025426 | 0) Xorg-4291 | | __schedule() { 30581.025426 | 0) Xorg-4291 | 0.705 us | _spin_lock_irq(); With this patch, the interrupts output better adapts to absolute time printing: 414.856543 | 1) Xorg-4279 | 8.816 us | } 414.856544 | 1) Xorg-4279 | 0.525 us | rcu_irq_exit(); 414.856545 | 1) Xorg-4279 | 0.526 us | idle_cpu(); 414.856546 | 1) Xorg-4279 | + 12.157 us | } 414.856549 | 1) Xorg-4279 | ! 104.114 us | } 414.856549 | 1) Xorg-4279 | <========== | 414.856549 | 1) Xorg-4279 | ! 107.944 us | } 414.856550 | 1) Xorg-4279 | ! 137.010 us | } 414.856551 | 1) Xorg-4279 | 0.624 us | _read_unlock(); 414.856552 | 1) Xorg-4279 | ! 140.930 us | } 414.856552 | 1) Xorg-4279 | ! 166.159 us | } Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions_graph.c | 50 ++++++++++++++++------------ 1 file changed, 29 insertions(+), 21 deletions(-) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 6c7738e4f98b..8f4004a00b4e 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -351,16 +351,35 @@ print_graph_overhead(unsigned long long duration, struct trace_seq *s) return trace_seq_printf(s, " "); } +static int print_graph_abs_time(u64 t, struct trace_seq *s) +{ + unsigned long usecs_rem; + + usecs_rem = do_div(t, NSEC_PER_SEC); + usecs_rem /= 1000; + + return trace_seq_printf(s, "%5lu.%06lu | ", + (unsigned long)t, usecs_rem); +} + static enum print_line_t -print_graph_irq(struct trace_seq *s, unsigned long addr, +print_graph_irq(struct trace_iterator *iter, unsigned long addr, enum trace_type type, int cpu, pid_t pid) { int ret; + struct trace_seq *s = &iter->seq; if (addr < (unsigned long)__irqentry_text_start || addr >= (unsigned long)__irqentry_text_end) return TRACE_TYPE_UNHANDLED; + /* Absolute time */ + if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { + ret = print_graph_abs_time(iter->ts, s); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + /* Cpu */ if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { ret = print_graph_cpu(s, cpu); @@ -446,17 +465,6 @@ print_graph_duration(unsigned long long duration, struct trace_seq *s) } -static int print_graph_abs_time(u64 t, struct trace_seq *s) -{ - unsigned long usecs_rem; - - usecs_rem = do_div(t, 1000000000); - usecs_rem /= 1000; - - return trace_seq_printf(s, "%5lu.%06lu | ", - (unsigned long)t, usecs_rem); -} - /* Case of a leaf function on its call entry */ static enum print_line_t print_graph_entry_leaf(struct trace_iterator *iter, @@ -561,7 +569,7 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; /* Interrupt */ - ret = print_graph_irq(s, call->func, TRACE_GRAPH_ENT, cpu, ent->pid); + ret = print_graph_irq(iter, call->func, TRACE_GRAPH_ENT, cpu, ent->pid); if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; @@ -581,7 +589,7 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, /* Proc */ if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, ent->pid); + ret = print_graph_proc(s, pid); if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; @@ -605,11 +613,11 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, int i; int ret; int cpu = iter->cpu; - pid_t *last_pid = iter->private; + pid_t *last_pid = iter->private, pid = ent->pid; unsigned long long duration = trace->rettime - trace->calltime; /* Pid */ - if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) + if (verif_pid(s, pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; /* Absolute time */ @@ -668,7 +676,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } - ret = print_graph_irq(s, trace->func, TRACE_GRAPH_RET, cpu, ent->pid); + ret = print_graph_irq(iter, trace->func, TRACE_GRAPH_RET, cpu, pid); if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; @@ -684,6 +692,10 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, int cpu = iter->cpu; pid_t *last_pid = iter->private; + /* Pid */ + if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + /* Absolute time */ if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { ret = print_graph_abs_time(iter->ts, s); @@ -691,10 +703,6 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } - /* Pid */ - if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - /* Cpu */ if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { ret = print_graph_cpu(s, cpu); From 985ec20ad531f2641ab9d5193e37891fe959fc7d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 18 Feb 2009 06:35:34 +0100 Subject: [PATCH 167/413] tracing/function-graph-tracer: provide documentation for the function graph tracer Update documentation for the function graph tracer. Signed-off-by: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- Documentation/ftrace.txt | 226 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 226 insertions(+) diff --git a/Documentation/ftrace.txt b/Documentation/ftrace.txt index 758fb42a1b68..055bcd2992da 100644 --- a/Documentation/ftrace.txt +++ b/Documentation/ftrace.txt @@ -129,6 +129,10 @@ of ftrace. Here is a list of some of the key files: set_ftrace_pid: Have the function tracer only trace a single thread. + set_graph_function: Select the function where the trace have to start + with the function graph tracer (See the section + "dynamic ftrace" for more details). + available_filter_functions: This lists the functions that ftrace has processed and can trace. These are the function names that you can pass to "set_ftrace_filter" or @@ -143,6 +147,12 @@ Here is the list of current tracers that may be configured. function - function tracer that uses mcount to trace all functions. + function_graph_tracer - similar to the function tracer except that the + function tracer probes the functions on their entry whereas the + function graph tracer traces on both entry and exit of the + functions. It then provides the ability to draw a graph of + function calls like a primitive C code source. + sched_switch - traces the context switches between tasks. irqsoff - traces the areas that disable interrupts and saves @@ -1226,6 +1236,163 @@ kernel module: [...] +function graph tracer +--------------------------- + +This tracer is similar to the function tracer except that it probes +a function on its entry and its exit. +This is done by setting a dynamically allocated stack of return addresses on each +task_struct. Then the tracer overwrites the return address of each function traced +to set a custom probe. Thus the original return address is stored on the stack of return +address in the task_struct. + +Probing on both extremities of a function leads to special features such as + +_ measure of function's time execution +_ having a reliable call stack to draw function calls graph + +This tracer is useful in several situations: + +_ you want to find the reason of a strange kernel behavior and need to see + what happens in detail on any areas (or specific ones). +_ you are experiencing weird latencies but it's difficult to find its origin. +_ you want to find quickly which path is taken by a specific function +_ you just want to see what happens inside your kernel + +# tracer: function_graph +# +# CPU DURATION FUNCTION CALLS +# | | | | | | | + + 0) | sys_open() { + 0) | do_sys_open() { + 0) | getname() { + 0) | kmem_cache_alloc() { + 0) 1.382 us | __might_sleep(); + 0) 2.478 us | } + 0) | strncpy_from_user() { + 0) | might_fault() { + 0) 1.389 us | __might_sleep(); + 0) 2.553 us | } + 0) 3.807 us | } + 0) 7.876 us | } + 0) | alloc_fd() { + 0) 0.668 us | _spin_lock(); + 0) 0.570 us | expand_files(); + 0) 0.586 us | _spin_unlock(); + + +There are several columns that can be dynamically enabled/disabled. +You can use every combination of options you want, depending on your needs. + +_ The cpu number on which the function executed is default enabled. + It is sometimes better to only trace one cpu (see tracing_cpu_mask file) + or you might sometimes see unordered function calls while cpu tracing switch. + + hide: echo nofuncgraph-cpu > /debug/tracing/trace_options + show: echo funcgraph-cpu > /debug/tracing/trace_options + +_ The duration (function's time of execution) is displayed on the closing bracket + line of a function or on the same line than the current function in case of a leaf + one. It is default enabled. + + hide: echo nofuncgraph-duration > /debug/tracing/trace_options + show: echo funcgraph-duration > /debug/tracing/trace_options + +_ The overhead field precedes the duration one in case of reached duration thresholds. + + hide: echo nofuncgraph-overhead > /debug/tracing/trace_options + show: echo funcgraph-overhead > /debug/tracing/trace_options + depends on: funcgraph-duration + + ie: + + 0) | up_write() { + 0) 0.646 us | _spin_lock_irqsave(); + 0) 0.684 us | _spin_unlock_irqrestore(); + 0) 3.123 us | } + 0) 0.548 us | fput(); + 0) + 58.628 us | } + + [...] + + 0) | putname() { + 0) | kmem_cache_free() { + 0) 0.518 us | __phys_addr(); + 0) 1.757 us | } + 0) 2.861 us | } + 0) ! 115.305 us | } + 0) ! 116.402 us | } + + + means that the function exceeded 10 usecs. + ! means that the function exceeded 100 usecs. + + +_ The task/pid field displays the thread cmdline and pid which executed the function. + It is default disabled. + + hide: echo nofuncgraph-proc > /debug/tracing/trace_options + show: echo funcgraph-proc > /debug/tracing/trace_options + + ie: + + # tracer: function_graph + # + # CPU TASK/PID DURATION FUNCTION CALLS + # | | | | | | | | | + 0) sh-4802 | | d_free() { + 0) sh-4802 | | call_rcu() { + 0) sh-4802 | | __call_rcu() { + 0) sh-4802 | 0.616 us | rcu_process_gp_end(); + 0) sh-4802 | 0.586 us | check_for_new_grace_period(); + 0) sh-4802 | 2.899 us | } + 0) sh-4802 | 4.040 us | } + 0) sh-4802 | 5.151 us | } + 0) sh-4802 | + 49.370 us | } + + +_ The absolute time field is an absolute timestamp given by the clock since + it started. A snapshot of this time is given on each entry/exit of functions + + hide: echo nofuncgraph-abstime > /debug/tracing/trace_options + show: echo funcgraph-abstime > /debug/tracing/trace_options + + ie: + + # + # TIME CPU DURATION FUNCTION CALLS + # | | | | | | | | + 360.774522 | 1) 0.541 us | } + 360.774522 | 1) 4.663 us | } + 360.774523 | 1) 0.541 us | __wake_up_bit(); + 360.774524 | 1) 6.796 us | } + 360.774524 | 1) 7.952 us | } + 360.774525 | 1) 9.063 us | } + 360.774525 | 1) 0.615 us | journal_mark_dirty(); + 360.774527 | 1) 0.578 us | __brelse(); + 360.774528 | 1) | reiserfs_prepare_for_journal() { + 360.774528 | 1) | unlock_buffer() { + 360.774529 | 1) | wake_up_bit() { + 360.774529 | 1) | bit_waitqueue() { + 360.774530 | 1) 0.594 us | __phys_addr(); + + +You can put some comments on specific functions by using ftrace_printk() +For example, if you want to put a comment inside the __might_sleep() function, +you just have to include and call ftrace_printk() inside __might_sleep() + +ftrace_printk("I'm a comment!\n") + +will produce: + + 1) | __might_sleep() { + 1) | /* I'm a comment! */ + 1) 1.449 us | } + + +You might find other useful features for this tracer on the "dynamic ftrace" +section such as tracing only specific functions or tasks. + dynamic ftrace -------------- @@ -1427,6 +1594,65 @@ Produces: We can see that there's no more lock or preempt tracing. + +* Dynamic ftrace with the function graph tracer * + + +Although what has been explained above concerns both the function tracer and +the function_graph_tracer, the following concerns only the latter. + +If you want to trace only one function and all of its childs, you just have +to echo its name on set_graph_function: + +echo __do_fault > set_graph_function + +will produce the following: + + 0) | __do_fault() { + 0) | filemap_fault() { + 0) | find_lock_page() { + 0) 0.804 us | find_get_page(); + 0) | __might_sleep() { + 0) 1.329 us | } + 0) 3.904 us | } + 0) 4.979 us | } + 0) 0.653 us | _spin_lock(); + 0) 0.578 us | page_add_file_rmap(); + 0) 0.525 us | native_set_pte_at(); + 0) 0.585 us | _spin_unlock(); + 0) | unlock_page() { + 0) 0.541 us | page_waitqueue(); + 0) 0.639 us | __wake_up_bit(); + 0) 2.786 us | } + 0) + 14.237 us | } + 0) | __do_fault() { + 0) | filemap_fault() { + 0) | find_lock_page() { + 0) 0.698 us | find_get_page(); + 0) | __might_sleep() { + 0) 1.412 us | } + 0) 3.950 us | } + 0) 5.098 us | } + 0) 0.631 us | _spin_lock(); + 0) 0.571 us | page_add_file_rmap(); + 0) 0.526 us | native_set_pte_at(); + 0) 0.586 us | _spin_unlock(); + 0) | unlock_page() { + 0) 0.533 us | page_waitqueue(); + 0) 0.638 us | __wake_up_bit(); + 0) 2.793 us | } + 0) + 14.012 us | } + +You can also select several functions: + +echo sys_open > set_graph_function +echo sys_close >> set_graph_function + +Now if you want to go back to trace all functions + +echo > set_graph_function + + trace_pipe ---------- From 5752674e140db5bce08c6bc60021a9bc3b960800 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 19 Feb 2009 12:54:10 +0100 Subject: [PATCH 168/413] Documentation/ftrace.txt: update - fix typos/grammos and clarify the text - prettify the document some more Cc: Frederic Weisbecker Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- Documentation/ftrace.txt | 922 ++++++++++++++++++++++----------------- 1 file changed, 513 insertions(+), 409 deletions(-) diff --git a/Documentation/ftrace.txt b/Documentation/ftrace.txt index 055bcd2992da..2041ee951c1a 100644 --- a/Documentation/ftrace.txt +++ b/Documentation/ftrace.txt @@ -15,31 +15,31 @@ Introduction Ftrace is an internal tracer designed to help out developers and designers of systems to find what is going on inside the kernel. -It can be used for debugging or analyzing latencies and performance -issues that take place outside of user-space. +It can be used for debugging or analyzing latencies and +performance issues that take place outside of user-space. Although ftrace is the function tracer, it also includes an -infrastructure that allows for other types of tracing. Some of the -tracers that are currently in ftrace include a tracer to trace -context switches, the time it takes for a high priority task to -run after it was woken up, the time interrupts are disabled, and -more (ftrace allows for tracer plugins, which means that the list of -tracers can always grow). +infrastructure that allows for other types of tracing. Some of +the tracers that are currently in ftrace include a tracer to +trace context switches, the time it takes for a high priority +task to run after it was woken up, the time interrupts are +disabled, and more (ftrace allows for tracer plugins, which +means that the list of tracers can always grow). The File System --------------- -Ftrace uses the debugfs file system to hold the control files as well -as the files to display output. +Ftrace uses the debugfs file system to hold the control files as +well as the files to display output. To mount the debugfs system: # mkdir /debug # mount -t debugfs nodev /debug -(Note: it is more common to mount at /sys/kernel/debug, but for simplicity - this document will use /debug) +( Note: it is more common to mount at /sys/kernel/debug, but for + simplicity this document will use /debug) That's it! (assuming that you have ftrace configured into your kernel) @@ -50,94 +50,124 @@ of ftrace. Here is a list of some of the key files: Note: all time values are in microseconds. - current_tracer: This is used to set or display the current tracer - that is configured. + current_tracer: - available_tracers: This holds the different types of tracers that - have been compiled into the kernel. The tracers - listed here can be configured by echoing their name - into current_tracer. + This is used to set or display the current tracer + that is configured. - tracing_enabled: This sets or displays whether the current_tracer - is activated and tracing or not. Echo 0 into this - file to disable the tracer or 1 to enable it. + available_tracers: - trace: This file holds the output of the trace in a human readable - format (described below). + This holds the different types of tracers that + have been compiled into the kernel. The + tracers listed here can be configured by + echoing their name into current_tracer. - latency_trace: This file shows the same trace but the information - is organized more to display possible latencies - in the system (described below). + tracing_enabled: - trace_pipe: The output is the same as the "trace" file but this - file is meant to be streamed with live tracing. - Reads from this file will block until new data - is retrieved. Unlike the "trace" and "latency_trace" - files, this file is a consumer. This means reading - from this file causes sequential reads to display - more current data. Once data is read from this - file, it is consumed, and will not be read - again with a sequential read. The "trace" and - "latency_trace" files are static, and if the - tracer is not adding more data, they will display - the same information every time they are read. + This sets or displays whether the current_tracer + is activated and tracing or not. Echo 0 into this + file to disable the tracer or 1 to enable it. - trace_options: This file lets the user control the amount of data - that is displayed in one of the above output - files. + trace: - trace_max_latency: Some of the tracers record the max latency. - For example, the time interrupts are disabled. - This time is saved in this file. The max trace - will also be stored, and displayed by either - "trace" or "latency_trace". A new max trace will - only be recorded if the latency is greater than - the value in this file. (in microseconds) + This file holds the output of the trace in a human + readable format (described below). - buffer_size_kb: This sets or displays the number of kilobytes each CPU - buffer can hold. The tracer buffers are the same size - for each CPU. The displayed number is the size of the - CPU buffer and not total size of all buffers. The - trace buffers are allocated in pages (blocks of memory - that the kernel uses for allocation, usually 4 KB in size). - If the last page allocated has room for more bytes - than requested, the rest of the page will be used, - making the actual allocation bigger than requested. - (Note, the size may not be a multiple of the page size due - to buffer managment overhead.) + latency_trace: - This can only be updated when the current_tracer - is set to "nop". + This file shows the same trace but the information + is organized more to display possible latencies + in the system (described below). - tracing_cpumask: This is a mask that lets the user only trace - on specified CPUS. The format is a hex string - representing the CPUS. + trace_pipe: - set_ftrace_filter: When dynamic ftrace is configured in (see the - section below "dynamic ftrace"), the code is dynamically - modified (code text rewrite) to disable calling of the - function profiler (mcount). This lets tracing be configured - in with practically no overhead in performance. This also - has a side effect of enabling or disabling specific functions - to be traced. Echoing names of functions into this file - will limit the trace to only those functions. + The output is the same as the "trace" file but this + file is meant to be streamed with live tracing. + Reads from this file will block until new data + is retrieved. Unlike the "trace" and "latency_trace" + files, this file is a consumer. This means reading + from this file causes sequential reads to display + more current data. Once data is read from this + file, it is consumed, and will not be read + again with a sequential read. The "trace" and + "latency_trace" files are static, and if the + tracer is not adding more data, they will display + the same information every time they are read. - set_ftrace_notrace: This has an effect opposite to that of - set_ftrace_filter. Any function that is added here will not - be traced. If a function exists in both set_ftrace_filter - and set_ftrace_notrace, the function will _not_ be traced. + trace_options: - set_ftrace_pid: Have the function tracer only trace a single thread. + This file lets the user control the amount of data + that is displayed in one of the above output + files. - set_graph_function: Select the function where the trace have to start - with the function graph tracer (See the section - "dynamic ftrace" for more details). + trace_max_latency: - available_filter_functions: This lists the functions that ftrace - has processed and can trace. These are the function - names that you can pass to "set_ftrace_filter" or - "set_ftrace_notrace". (See the section "dynamic ftrace" - below for more details.) + Some of the tracers record the max latency. + For example, the time interrupts are disabled. + This time is saved in this file. The max trace + will also be stored, and displayed by either + "trace" or "latency_trace". A new max trace will + only be recorded if the latency is greater than + the value in this file. (in microseconds) + + buffer_size_kb: + + This sets or displays the number of kilobytes each CPU + buffer can hold. The tracer buffers are the same size + for each CPU. The displayed number is the size of the + CPU buffer and not total size of all buffers. The + trace buffers are allocated in pages (blocks of memory + that the kernel uses for allocation, usually 4 KB in size). + If the last page allocated has room for more bytes + than requested, the rest of the page will be used, + making the actual allocation bigger than requested. + ( Note, the size may not be a multiple of the page size + due to buffer managment overhead. ) + + This can only be updated when the current_tracer + is set to "nop". + + tracing_cpumask: + + This is a mask that lets the user only trace + on specified CPUS. The format is a hex string + representing the CPUS. + + set_ftrace_filter: + + When dynamic ftrace is configured in (see the + section below "dynamic ftrace"), the code is dynamically + modified (code text rewrite) to disable calling of the + function profiler (mcount). This lets tracing be configured + in with practically no overhead in performance. This also + has a side effect of enabling or disabling specific functions + to be traced. Echoing names of functions into this file + will limit the trace to only those functions. + + set_ftrace_notrace: + + This has an effect opposite to that of + set_ftrace_filter. Any function that is added here will not + be traced. If a function exists in both set_ftrace_filter + and set_ftrace_notrace, the function will _not_ be traced. + + set_ftrace_pid: + + Have the function tracer only trace a single thread. + + set_graph_function: + + Set a "trigger" function where tracing should start + with the function graph tracer (See the section + "dynamic ftrace" for more details). + + available_filter_functions: + + This lists the functions that ftrace + has processed and can trace. These are the function + names that you can pass to "set_ftrace_filter" or + "set_ftrace_notrace". (See the section "dynamic ftrace" + below for more details.) The Tracers @@ -145,44 +175,66 @@ The Tracers Here is the list of current tracers that may be configured. - function - function tracer that uses mcount to trace all functions. + "function" - function_graph_tracer - similar to the function tracer except that the - function tracer probes the functions on their entry whereas the - function graph tracer traces on both entry and exit of the - functions. It then provides the ability to draw a graph of - function calls like a primitive C code source. + Function call tracer to trace all kernel functions. - sched_switch - traces the context switches between tasks. + "function_graph_tracer" - irqsoff - traces the areas that disable interrupts and saves - the trace with the longest max latency. - See tracing_max_latency. When a new max is recorded, - it replaces the old trace. It is best to view this - trace via the latency_trace file. + Similar to the function tracer except that the + function tracer probes the functions on their entry + whereas the function graph tracer traces on both entry + and exit of the functions. It then provides the ability + to draw a graph of function calls similar to C code + source. - preemptoff - Similar to irqsoff but traces and records the amount of - time for which preemption is disabled. + "sched_switch" - preemptirqsoff - Similar to irqsoff and preemptoff, but traces and - records the largest time for which irqs and/or preemption - is disabled. + Traces the context switches and wakeups between tasks. - wakeup - Traces and records the max latency that it takes for - the highest priority task to get scheduled after - it has been woken up. + "irqsoff" - nop - This is not a tracer. To remove all tracers from tracing - simply echo "nop" into current_tracer. + Traces the areas that disable interrupts and saves + the trace with the longest max latency. + See tracing_max_latency. When a new max is recorded, + it replaces the old trace. It is best to view this + trace via the latency_trace file. - hw-branch-tracer - traces branches on all cpu's in a circular buffer. + "preemptoff" + + Similar to irqsoff but traces and records the amount of + time for which preemption is disabled. + + "preemptirqsoff" + + Similar to irqsoff and preemptoff, but traces and + records the largest time for which irqs and/or preemption + is disabled. + + "wakeup" + + Traces and records the max latency that it takes for + the highest priority task to get scheduled after + it has been woken up. + + "hw-branch-tracer" + + Uses the BTS CPU feature on x86 CPUs to traces all + branches executed. + + "nop" + + This is the "trace nothing" tracer. To remove all + tracers from tracing simply echo "nop" into + current_tracer. Examples of using the tracer ---------------------------- -Here are typical examples of using the tracers when controlling them only -with the debugfs interface (without using any user-land utilities). +Here are typical examples of using the tracers when controlling +them only with the debugfs interface (without using any +user-land utilities). Output format: -------------- @@ -199,16 +251,16 @@ Here is an example of the output format of the file "trace" bash-4251 [01] 10152.583855: _atomic_dec_and_lock <-dput -------- -A header is printed with the tracer name that is represented by the trace. -In this case the tracer is "function". Then a header showing the format. Task -name "bash", the task PID "4251", the CPU that it was running on -"01", the timestamp in . format, the function name that was -traced "path_put" and the parent function that called this function -"path_walk". The timestamp is the time at which the function was -entered. +A header is printed with the tracer name that is represented by +the trace. In this case the tracer is "function". Then a header +showing the format. Task name "bash", the task PID "4251", the +CPU that it was running on "01", the timestamp in . +format, the function name that was traced "path_put" and the +parent function that called this function "path_walk". The +timestamp is the time at which the function was entered. -The sched_switch tracer also includes tracing of task wakeups and -context switches. +The sched_switch tracer also includes tracing of task wakeups +and context switches. ksoftirqd/1-7 [01] 1453.070013: 7:115:R + 2916:115:S ksoftirqd/1-7 [01] 1453.070013: 7:115:R + 10:115:S @@ -217,8 +269,8 @@ context switches. kondemand/1-2916 [01] 1453.070013: 2916:115:S ==> 7:115:R ksoftirqd/1-7 [01] 1453.070013: 7:115:S ==> 0:140:R -Wake ups are represented by a "+" and the context switches are shown as -"==>". The format is: +Wake ups are represented by a "+" and the context switches are +shown as "==>". The format is: Context switches: @@ -232,19 +284,20 @@ Wake ups are represented by a "+" and the context switches are shown as :: + :: -The prio is the internal kernel priority, which is the inverse of the -priority that is usually displayed by user-space tools. Zero represents -the highest priority (99). Prio 100 starts the "nice" priorities with -100 being equal to nice -20 and 139 being nice 19. The prio "140" is -reserved for the idle task which is the lowest priority thread (pid 0). +The prio is the internal kernel priority, which is the inverse +of the priority that is usually displayed by user-space tools. +Zero represents the highest priority (99). Prio 100 starts the +"nice" priorities with 100 being equal to nice -20 and 139 being +nice 19. The prio "140" is reserved for the idle task which is +the lowest priority thread (pid 0). Latency trace format -------------------- -For traces that display latency times, the latency_trace file gives -somewhat more information to see why a latency happened. Here is a typical -trace. +For traces that display latency times, the latency_trace file +gives somewhat more information to see why a latency happened. +Here is a typical trace. # tracer: irqsoff # @@ -271,20 +324,20 @@ irqsoff latency trace v1.1.5 on 2.6.26-rc8 -0 0d.s1 98us : trace_hardirqs_on (do_softirq) +This shows that the current tracer is "irqsoff" tracing the time +for which interrupts were disabled. It gives the trace version +and the version of the kernel upon which this was executed on +(2.6.26-rc8). Then it displays the max latency in microsecs (97 +us). The number of trace entries displayed and the total number +recorded (both are three: #3/3). The type of preemption that was +used (PREEMPT). VP, KP, SP, and HP are always zero and are +reserved for later use. #P is the number of online CPUS (#P:2). -This shows that the current tracer is "irqsoff" tracing the time for which -interrupts were disabled. It gives the trace version and the version -of the kernel upon which this was executed on (2.6.26-rc8). Then it displays -the max latency in microsecs (97 us). The number of trace entries displayed -and the total number recorded (both are three: #3/3). The type of -preemption that was used (PREEMPT). VP, KP, SP, and HP are always zero -and are reserved for later use. #P is the number of online CPUS (#P:2). +The task is the process that was running when the latency +occurred. (swapper pid: 0). -The task is the process that was running when the latency occurred. -(swapper pid: 0). - -The start and stop (the functions in which the interrupts were disabled and -enabled respectively) that caused the latencies: +The start and stop (the functions in which the interrupts were +disabled and enabled respectively) that caused the latencies: apic_timer_interrupt is where the interrupts were disabled. do_softirq is where they were enabled again. @@ -320,12 +373,12 @@ The above is mostly meaningful for kernel developers. latency_trace file is relative to the start of the trace. delay: This is just to help catch your eye a bit better. And - needs to be fixed to be only relative to the same CPU. - The marks are determined by the difference between this - current trace and the next trace. - '!' - greater than preempt_mark_thresh (default 100) - '+' - greater than 1 microsecond - ' ' - less than or equal to 1 microsecond. + needs to be fixed to be only relative to the same CPU. + The marks are determined by the difference between this + current trace and the next trace. + '!' - greater than preempt_mark_thresh (default 100) + '+' - greater than 1 microsecond + ' ' - less than or equal to 1 microsecond. The rest is the same as the 'trace' file. @@ -333,14 +386,15 @@ The above is mostly meaningful for kernel developers. trace_options ------------- -The trace_options file is used to control what gets printed in the trace -output. To see what is available, simply cat the file: +The trace_options file is used to control what gets printed in +the trace output. To see what is available, simply cat the file: cat /debug/tracing/trace_options print-parent nosym-offset nosym-addr noverbose noraw nohex nobin \ - noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj + noblock nostacktrace nosched-tree nouserstacktrace nosym-userobj -To disable one of the options, echo in the option prepended with "no". +To disable one of the options, echo in the option prepended with +"no". echo noprint-parent > /debug/tracing/trace_options @@ -350,8 +404,8 @@ To enable an option, leave off the "no". Here are the available options: - print-parent - On function traces, display the calling function - as well as the function being traced. + print-parent - On function traces, display the calling (parent) + function as well as the function being traced. print-parent: bash-4000 [01] 1477.606694: simple_strtoul <-strict_strtoul @@ -360,15 +414,16 @@ Here are the available options: bash-4000 [01] 1477.606694: simple_strtoul - sym-offset - Display not only the function name, but also the offset - in the function. For example, instead of seeing just - "ktime_get", you will see "ktime_get+0xb/0x20". + sym-offset - Display not only the function name, but also the + offset in the function. For example, instead of + seeing just "ktime_get", you will see + "ktime_get+0xb/0x20". sym-offset: bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0 - sym-addr - this will also display the function address as well as - the function name. + sym-addr - this will also display the function address as well + as the function name. sym-addr: bash-4000 [01] 1477.606694: simple_strtoul @@ -378,35 +433,41 @@ Here are the available options: bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \ (+0.000ms): simple_strtoul (strict_strtoul) - raw - This will display raw numbers. This option is best for use with - user applications that can translate the raw numbers better than - having it done in the kernel. + raw - This will display raw numbers. This option is best for + use with user applications that can translate the raw + numbers better than having it done in the kernel. - hex - Similar to raw, but the numbers will be in a hexadecimal format. + hex - Similar to raw, but the numbers will be in a hexadecimal + format. bin - This will print out the formats in raw binary. block - TBD (needs update) - stacktrace - This is one of the options that changes the trace itself. - When a trace is recorded, so is the stack of functions. - This allows for back traces of trace sites. + stacktrace - This is one of the options that changes the trace + itself. When a trace is recorded, so is the stack + of functions. This allows for back traces of + trace sites. - userstacktrace - This option changes the trace. - It records a stacktrace of the current userspace thread. + userstacktrace - This option changes the trace. It records a + stacktrace of the current userspace thread. - sym-userobj - when user stacktrace are enabled, look up which object the - address belongs to, and print a relative address - This is especially useful when ASLR is on, otherwise you don't - get a chance to resolve the address to object/file/line after the app is no - longer running + sym-userobj - when user stacktrace are enabled, look up which + object the address belongs to, and print a + relative address. This is especially useful when + ASLR is on, otherwise you don't get a chance to + resolve the address to object/file/line after + the app is no longer running - The lookup is performed when you read trace,trace_pipe,latency_trace. Example: + The lookup is performed when you read + trace,trace_pipe,latency_trace. Example: a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0 x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6] - sched-tree - TBD (any users??) + sched-tree - trace all tasks that are on the runqueue, at + every scheduling event. Will add overhead if + there's a lot of tasks running at once. sched_switch @@ -443,18 +504,19 @@ of how to use it. [...] -As we have discussed previously about this format, the header shows -the name of the trace and points to the options. The "FUNCTION" -is a misnomer since here it represents the wake ups and context -switches. +As we have discussed previously about this format, the header +shows the name of the trace and points to the options. The +"FUNCTION" is a misnomer since here it represents the wake ups +and context switches. -The sched_switch file only lists the wake ups (represented with '+') -and context switches ('==>') with the previous task or current task -first followed by the next task or task waking up. The format for both -of these is PID:KERNEL-PRIO:TASK-STATE. Remember that the KERNEL-PRIO -is the inverse of the actual priority with zero (0) being the highest -priority and the nice values starting at 100 (nice -20). Below is -a quick chart to map the kernel priority to user land priorities. +The sched_switch file only lists the wake ups (represented with +'+') and context switches ('==>') with the previous task or +current task first followed by the next task or task waking up. +The format for both of these is PID:KERNEL-PRIO:TASK-STATE. +Remember that the KERNEL-PRIO is the inverse of the actual +priority with zero (0) being the highest priority and the nice +values starting at 100 (nice -20). Below is a quick chart to map +the kernel priority to user land priorities. Kernel priority: 0 to 99 ==> user RT priority 99 to 0 Kernel priority: 100 to 139 ==> user nice -20 to 19 @@ -475,10 +537,10 @@ The task states are: ftrace_enabled -------------- -The following tracers (listed below) give different output depending -on whether or not the sysctl ftrace_enabled is set. To set ftrace_enabled, -one can either use the sysctl function or set it via the proc -file system interface. +The following tracers (listed below) give different output +depending on whether or not the sysctl ftrace_enabled is set. To +set ftrace_enabled, one can either use the sysctl function or +set it via the proc file system interface. sysctl kernel.ftrace_enabled=1 @@ -486,12 +548,12 @@ file system interface. echo 1 > /proc/sys/kernel/ftrace_enabled -To disable ftrace_enabled simply replace the '1' with '0' in -the above commands. +To disable ftrace_enabled simply replace the '1' with '0' in the +above commands. -When ftrace_enabled is set the tracers will also record the functions -that are within the trace. The descriptions of the tracers -will also show an example with ftrace enabled. +When ftrace_enabled is set the tracers will also record the +functions that are within the trace. The descriptions of the +tracers will also show an example with ftrace enabled. irqsoff @@ -499,17 +561,18 @@ irqsoff When interrupts are disabled, the CPU can not react to any other external event (besides NMIs and SMIs). This prevents the timer -interrupt from triggering or the mouse interrupt from letting the -kernel know of a new mouse event. The result is a latency with the -reaction time. +interrupt from triggering or the mouse interrupt from letting +the kernel know of a new mouse event. The result is a latency +with the reaction time. -The irqsoff tracer tracks the time for which interrupts are disabled. -When a new maximum latency is hit, the tracer saves the trace leading up -to that latency point so that every time a new maximum is reached, the old -saved trace is discarded and the new trace is saved. +The irqsoff tracer tracks the time for which interrupts are +disabled. When a new maximum latency is hit, the tracer saves +the trace leading up to that latency point so that every time a +new maximum is reached, the old saved trace is discarded and the +new trace is saved. -To reset the maximum, echo 0 into tracing_max_latency. Here is an -example: +To reset the maximum, echo 0 into tracing_max_latency. Here is +an example: # echo irqsoff > /debug/tracing/current_tracer # echo 0 > /debug/tracing/tracing_max_latency @@ -544,10 +607,11 @@ irqsoff latency trace v1.1.5 on 2.6.26 Here we see that that we had a latency of 12 microsecs (which is -very good). The _write_lock_irq in sys_setpgid disabled interrupts. -The difference between the 12 and the displayed timestamp 14us occurred -because the clock was incremented between the time of recording the max -latency and the time of recording the function that had that latency. +very good). The _write_lock_irq in sys_setpgid disabled +interrupts. The difference between the 12 and the displayed +timestamp 14us occurred because the clock was incremented +between the time of recording the max latency and the time of +recording the function that had that latency. Note the above example had ftrace_enabled not set. If we set the ftrace_enabled, we get a much larger output: @@ -598,24 +662,24 @@ irqsoff latency trace v1.1.5 on 2.6.26-rc8 Here we traced a 50 microsecond latency. But we also see all the -functions that were called during that time. Note that by enabling -function tracing, we incur an added overhead. This overhead may -extend the latency times. But nevertheless, this trace has provided -some very helpful debugging information. +functions that were called during that time. Note that by +enabling function tracing, we incur an added overhead. This +overhead may extend the latency times. But nevertheless, this +trace has provided some very helpful debugging information. preemptoff ---------- -When preemption is disabled, we may be able to receive interrupts but -the task cannot be preempted and a higher priority task must wait -for preemption to be enabled again before it can preempt a lower -priority task. +When preemption is disabled, we may be able to receive +interrupts but the task cannot be preempted and a higher +priority task must wait for preemption to be enabled again +before it can preempt a lower priority task. The preemptoff tracer traces the places that disable preemption. -Like the irqsoff tracer, it records the maximum latency for which preemption -was disabled. The control of preemptoff tracer is much like the irqsoff -tracer. +Like the irqsoff tracer, it records the maximum latency for +which preemption was disabled. The control of preemptoff tracer +is much like the irqsoff tracer. # echo preemptoff > /debug/tracing/current_tracer # echo 0 > /debug/tracing/tracing_max_latency @@ -649,11 +713,12 @@ preemptoff latency trace v1.1.5 on 2.6.26-rc8 sshd-4261 0d.s1 30us : trace_preempt_on (__do_softirq) -This has some more changes. Preemption was disabled when an interrupt -came in (notice the 'h'), and was enabled while doing a softirq. -(notice the 's'). But we also see that interrupts have been disabled -when entering the preempt off section and leaving it (the 'd'). -We do not know if interrupts were enabled in the mean time. +This has some more changes. Preemption was disabled when an +interrupt came in (notice the 'h'), and was enabled while doing +a softirq. (notice the 's'). But we also see that interrupts +have been disabled when entering the preempt off section and +leaving it (the 'd'). We do not know if interrupts were enabled +in the mean time. # tracer: preemptoff # @@ -712,28 +777,30 @@ preemptoff latency trace v1.1.5 on 2.6.26-rc8 sshd-4261 0d.s1 64us : trace_preempt_on (__do_softirq) -The above is an example of the preemptoff trace with ftrace_enabled -set. Here we see that interrupts were disabled the entire time. -The irq_enter code lets us know that we entered an interrupt 'h'. -Before that, the functions being traced still show that it is not -in an interrupt, but we can see from the functions themselves that -this is not the case. +The above is an example of the preemptoff trace with +ftrace_enabled set. Here we see that interrupts were disabled +the entire time. The irq_enter code lets us know that we entered +an interrupt 'h'. Before that, the functions being traced still +show that it is not in an interrupt, but we can see from the +functions themselves that this is not the case. -Notice that __do_softirq when called does not have a preempt_count. -It may seem that we missed a preempt enabling. What really happened -is that the preempt count is held on the thread's stack and we -switched to the softirq stack (4K stacks in effect). The code -does not copy the preempt count, but because interrupts are disabled, -we do not need to worry about it. Having a tracer like this is good -for letting people know what really happens inside the kernel. +Notice that __do_softirq when called does not have a +preempt_count. It may seem that we missed a preempt enabling. +What really happened is that the preempt count is held on the +thread's stack and we switched to the softirq stack (4K stacks +in effect). The code does not copy the preempt count, but +because interrupts are disabled, we do not need to worry about +it. Having a tracer like this is good for letting people know +what really happens inside the kernel. preemptirqsoff -------------- -Knowing the locations that have interrupts disabled or preemption -disabled for the longest times is helpful. But sometimes we would -like to know when either preemption and/or interrupts are disabled. +Knowing the locations that have interrupts disabled or +preemption disabled for the longest times is helpful. But +sometimes we would like to know when either preemption and/or +interrupts are disabled. Consider the following code: @@ -753,11 +820,13 @@ The preemptoff tracer will record the total length of call_function_with_irqs_and_preemption_off() and call_function_with_preemption_off(). -But neither will trace the time that interrupts and/or preemption -is disabled. This total time is the time that we can not schedule. -To record this time, use the preemptirqsoff tracer. +But neither will trace the time that interrupts and/or +preemption is disabled. This total time is the time that we can +not schedule. To record this time, use the preemptirqsoff +tracer. -Again, using this trace is much like the irqsoff and preemptoff tracers. +Again, using this trace is much like the irqsoff and preemptoff +tracers. # echo preemptirqsoff > /debug/tracing/current_tracer # echo 0 > /debug/tracing/tracing_max_latency @@ -793,9 +862,10 @@ preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8 The trace_hardirqs_off_thunk is called from assembly on x86 when -interrupts are disabled in the assembly code. Without the function -tracing, we do not know if interrupts were enabled within the preemption -points. We do see that it started with preemption enabled. +interrupts are disabled in the assembly code. Without the +function tracing, we do not know if interrupts were enabled +within the preemption points. We do see that it started with +preemption enabled. Here is a trace with ftrace_enabled set: @@ -883,40 +953,42 @@ preemptirqsoff latency trace v1.1.5 on 2.6.26-rc8 sshd-4261 0d.s1 105us : trace_preempt_on (__do_softirq) -This is a very interesting trace. It started with the preemption of -the ls task. We see that the task had the "need_resched" bit set -via the 'N' in the trace. Interrupts were disabled before the spin_lock -at the beginning of the trace. We see that a schedule took place to run -sshd. When the interrupts were enabled, we took an interrupt. -On return from the interrupt handler, the softirq ran. We took another -interrupt while running the softirq as we see from the capital 'H'. +This is a very interesting trace. It started with the preemption +of the ls task. We see that the task had the "need_resched" bit +set via the 'N' in the trace. Interrupts were disabled before +the spin_lock at the beginning of the trace. We see that a +schedule took place to run sshd. When the interrupts were +enabled, we took an interrupt. On return from the interrupt +handler, the softirq ran. We took another interrupt while +running the softirq as we see from the capital 'H'. wakeup ------ -In a Real-Time environment it is very important to know the wakeup -time it takes for the highest priority task that is woken up to the -time that it executes. This is also known as "schedule latency". -I stress the point that this is about RT tasks. It is also important -to know the scheduling latency of non-RT tasks, but the average -schedule latency is better for non-RT tasks. Tools like -LatencyTop are more appropriate for such measurements. +In a Real-Time environment it is very important to know the +wakeup time it takes for the highest priority task that is woken +up to the time that it executes. This is also known as "schedule +latency". I stress the point that this is about RT tasks. It is +also important to know the scheduling latency of non-RT tasks, +but the average schedule latency is better for non-RT tasks. +Tools like LatencyTop are more appropriate for such +measurements. Real-Time environments are interested in the worst case latency. -That is the longest latency it takes for something to happen, and -not the average. We can have a very fast scheduler that may only -have a large latency once in a while, but that would not work well -with Real-Time tasks. The wakeup tracer was designed to record -the worst case wakeups of RT tasks. Non-RT tasks are not recorded -because the tracer only records one worst case and tracing non-RT -tasks that are unpredictable will overwrite the worst case latency -of RT tasks. +That is the longest latency it takes for something to happen, +and not the average. We can have a very fast scheduler that may +only have a large latency once in a while, but that would not +work well with Real-Time tasks. The wakeup tracer was designed +to record the worst case wakeups of RT tasks. Non-RT tasks are +not recorded because the tracer only records one worst case and +tracing non-RT tasks that are unpredictable will overwrite the +worst case latency of RT tasks. -Since this tracer only deals with RT tasks, we will run this slightly -differently than we did with the previous tracers. Instead of performing -an 'ls', we will run 'sleep 1' under 'chrt' which changes the -priority of the task. +Since this tracer only deals with RT tasks, we will run this +slightly differently than we did with the previous tracers. +Instead of performing an 'ls', we will run 'sleep 1' under +'chrt' which changes the priority of the task. # echo wakeup > /debug/tracing/current_tracer # echo 0 > /debug/tracing/tracing_max_latency @@ -946,17 +1018,16 @@ wakeup latency trace v1.1.5 on 2.6.26-rc8 -0 1d..4 4us : schedule (cpu_idle) +Running this on an idle system, we see that it only took 4 +microseconds to perform the task switch. Note, since the trace +marker in the schedule is before the actual "switch", we stop +the tracing when the recorded task is about to schedule in. This +may change if we add a new marker at the end of the scheduler. -Running this on an idle system, we see that it only took 4 microseconds -to perform the task switch. Note, since the trace marker in the -schedule is before the actual "switch", we stop the tracing when -the recorded task is about to schedule in. This may change if -we add a new marker at the end of the scheduler. - -Notice that the recorded task is 'sleep' with the PID of 4901 and it -has an rt_prio of 5. This priority is user-space priority and not -the internal kernel priority. The policy is 1 for SCHED_FIFO and 2 -for SCHED_RR. +Notice that the recorded task is 'sleep' with the PID of 4901 +and it has an rt_prio of 5. This priority is user-space priority +and not the internal kernel priority. The policy is 1 for +SCHED_FIFO and 2 for SCHED_RR. Doing the same with chrt -r 5 and ftrace_enabled set. @@ -1013,24 +1084,25 @@ ksoftirq-7 1d..6 49us : _spin_unlock (tracing_record_cmdline) ksoftirq-7 1d..6 49us : sub_preempt_count (_spin_unlock) ksoftirq-7 1d..4 50us : schedule (__cond_resched) -The interrupt went off while running ksoftirqd. This task runs at -SCHED_OTHER. Why did not we see the 'N' set early? This may be -a harmless bug with x86_32 and 4K stacks. On x86_32 with 4K stacks -configured, the interrupt and softirq run with their own stack. -Some information is held on the top of the task's stack (need_resched -and preempt_count are both stored there). The setting of the NEED_RESCHED -bit is done directly to the task's stack, but the reading of the -NEED_RESCHED is done by looking at the current stack, which in this case -is the stack for the hard interrupt. This hides the fact that NEED_RESCHED -has been set. We do not see the 'N' until we switch back to the task's +The interrupt went off while running ksoftirqd. This task runs +at SCHED_OTHER. Why did not we see the 'N' set early? This may +be a harmless bug with x86_32 and 4K stacks. On x86_32 with 4K +stacks configured, the interrupt and softirq run with their own +stack. Some information is held on the top of the task's stack +(need_resched and preempt_count are both stored there). The +setting of the NEED_RESCHED bit is done directly to the task's +stack, but the reading of the NEED_RESCHED is done by looking at +the current stack, which in this case is the stack for the hard +interrupt. This hides the fact that NEED_RESCHED has been set. +We do not see the 'N' until we switch back to the task's assigned stack. function -------- This tracer is the function tracer. Enabling the function tracer -can be done from the debug file system. Make sure the ftrace_enabled is -set; otherwise this tracer is a nop. +can be done from the debug file system. Make sure the +ftrace_enabled is set; otherwise this tracer is a nop. # sysctl kernel.ftrace_enabled=1 # echo function > /debug/tracing/current_tracer @@ -1060,14 +1132,15 @@ set; otherwise this tracer is a nop. [...] -Note: function tracer uses ring buffers to store the above entries. -The newest data may overwrite the oldest data. Sometimes using echo to -stop the trace is not sufficient because the tracing could have overwritten -the data that you wanted to record. For this reason, it is sometimes better to -disable tracing directly from a program. This allows you to stop the -tracing at the point that you hit the part that you are interested in. -To disable the tracing directly from a C program, something like following -code snippet can be used: +Note: function tracer uses ring buffers to store the above +entries. The newest data may overwrite the oldest data. +Sometimes using echo to stop the trace is not sufficient because +the tracing could have overwritten the data that you wanted to +record. For this reason, it is sometimes better to disable +tracing directly from a program. This allows you to stop the +tracing at the point that you hit the part that you are +interested in. To disable the tracing directly from a C program, +something like following code snippet can be used: int trace_fd; [...] @@ -1082,10 +1155,10 @@ int main(int argc, char *argv[]) { } Note: Here we hard coded the path name. The debugfs mount is not -guaranteed to be at /debug (and is more commonly at /sys/kernel/debug). -For simple one time traces, the above is sufficent. For anything else, -a search through /proc/mounts may be needed to find where the debugfs -file-system is mounted. +guaranteed to be at /debug (and is more commonly at +/sys/kernel/debug). For simple one time traces, the above is +sufficent. For anything else, a search through /proc/mounts may +be needed to find where the debugfs file-system is mounted. Single thread tracing @@ -1186,10 +1259,11 @@ following format: 0 scheduler_tick+0x1b6/0x1bf <- scheduler_tick+0x1aa/0x1bf -The tracer may be used to dump the trace for the oops'ing cpu on a -kernel oops into the system log. To enable this, ftrace_dump_on_oops -must be set. To set ftrace_dump_on_oops, one can either use the sysctl -function or set it via the proc system interface. +The tracer may be used to dump the trace for the oops'ing cpu on +a kernel oops into the system log. To enable this, +ftrace_dump_on_oops must be set. To set ftrace_dump_on_oops, one +can either use the sysctl function or set it via the proc system +interface. sysctl kernel.ftrace_dump_on_oops=1 @@ -1198,8 +1272,8 @@ or echo 1 > /proc/sys/kernel/ftrace_dump_on_oops -Here's an example of such a dump after a null pointer dereference in a -kernel module: +Here's an example of such a dump after a null pointer +dereference in a kernel module: [57848.105921] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000 [57848.106019] IP: [] open+0x6/0x14 [oops] @@ -1239,25 +1313,34 @@ kernel module: function graph tracer --------------------------- -This tracer is similar to the function tracer except that it probes -a function on its entry and its exit. -This is done by setting a dynamically allocated stack of return addresses on each -task_struct. Then the tracer overwrites the return address of each function traced -to set a custom probe. Thus the original return address is stored on the stack of return -address in the task_struct. +This tracer is similar to the function tracer except that it +probes a function on its entry and its exit. This is done by +using a dynamically allocated stack of return addresses in each +task_struct. On function entry the tracer overwrites the return +address of each function traced to set a custom probe. Thus the +original return address is stored on the stack of return address +in the task_struct. -Probing on both extremities of a function leads to special features such as +Probing on both ends of a function leads to special features +such as: -_ measure of function's time execution -_ having a reliable call stack to draw function calls graph +- measure of a function's time execution +- having a reliable call stack to draw function calls graph This tracer is useful in several situations: -_ you want to find the reason of a strange kernel behavior and need to see - what happens in detail on any areas (or specific ones). -_ you are experiencing weird latencies but it's difficult to find its origin. -_ you want to find quickly which path is taken by a specific function -_ you just want to see what happens inside your kernel +- you want to find the reason of a strange kernel behavior and + need to see what happens in detail on any areas (or specific + ones). + +- you are experiencing weird latencies but it's difficult to + find its origin. + +- you want to find quickly which path is taken by a specific + function + +- you just want to peek inside a working kernel and want to see + what happens there. # tracer: function_graph # @@ -1282,24 +1365,28 @@ _ you just want to see what happens inside your kernel 0) 0.586 us | _spin_unlock(); -There are several columns that can be dynamically enabled/disabled. -You can use every combination of options you want, depending on your needs. +There are several columns that can be dynamically +enabled/disabled. You can use every combination of options you +want, depending on your needs. -_ The cpu number on which the function executed is default enabled. - It is sometimes better to only trace one cpu (see tracing_cpu_mask file) - or you might sometimes see unordered function calls while cpu tracing switch. +- The cpu number on which the function executed is default + enabled. It is sometimes better to only trace one cpu (see + tracing_cpu_mask file) or you might sometimes see unordered + function calls while cpu tracing switch. hide: echo nofuncgraph-cpu > /debug/tracing/trace_options show: echo funcgraph-cpu > /debug/tracing/trace_options -_ The duration (function's time of execution) is displayed on the closing bracket - line of a function or on the same line than the current function in case of a leaf - one. It is default enabled. +- The duration (function's time of execution) is displayed on + the closing bracket line of a function or on the same line + than the current function in case of a leaf one. It is default + enabled. hide: echo nofuncgraph-duration > /debug/tracing/trace_options show: echo funcgraph-duration > /debug/tracing/trace_options -_ The overhead field precedes the duration one in case of reached duration thresholds. +- The overhead field precedes the duration field in case of + reached duration thresholds. hide: echo nofuncgraph-overhead > /debug/tracing/trace_options show: echo funcgraph-overhead > /debug/tracing/trace_options @@ -1328,8 +1415,8 @@ _ The overhead field precedes the duration one in case of reached duration thres ! means that the function exceeded 100 usecs. -_ The task/pid field displays the thread cmdline and pid which executed the function. - It is default disabled. +- The task/pid field displays the thread cmdline and pid which + executed the function. It is default disabled. hide: echo nofuncgraph-proc > /debug/tracing/trace_options show: echo funcgraph-proc > /debug/tracing/trace_options @@ -1351,8 +1438,9 @@ _ The task/pid field displays the thread cmdline and pid which executed the func 0) sh-4802 | + 49.370 us | } -_ The absolute time field is an absolute timestamp given by the clock since - it started. A snapshot of this time is given on each entry/exit of functions +- The absolute time field is an absolute timestamp given by the + system clock since it started. A snapshot of this time is + given on each entry/exit of functions hide: echo nofuncgraph-abstime > /debug/tracing/trace_options show: echo funcgraph-abstime > /debug/tracing/trace_options @@ -1377,9 +1465,10 @@ _ The absolute time field is an absolute timestamp given by the clock since 360.774530 | 1) 0.594 us | __phys_addr(); -You can put some comments on specific functions by using ftrace_printk() -For example, if you want to put a comment inside the __might_sleep() function, -you just have to include and call ftrace_printk() inside __might_sleep() +You can put some comments on specific functions by using +ftrace_printk() For example, if you want to put a comment inside +the __might_sleep() function, you just have to include + and call ftrace_printk() inside __might_sleep() ftrace_printk("I'm a comment!\n") @@ -1390,8 +1479,9 @@ will produce: 1) 1.449 us | } -You might find other useful features for this tracer on the "dynamic ftrace" -section such as tracing only specific functions or tasks. +You might find other useful features for this tracer in the +following "dynamic ftrace" section such as tracing only specific +functions or tasks. dynamic ftrace -------------- @@ -1399,43 +1489,45 @@ dynamic ftrace If CONFIG_DYNAMIC_FTRACE is set, the system will run with virtually no overhead when function tracing is disabled. The way this works is the mcount function call (placed at the start of -every kernel function, produced by the -pg switch in gcc), starts -of pointing to a simple return. (Enabling FTRACE will include the --pg switch in the compiling of the kernel.) +every kernel function, produced by the -pg switch in gcc), +starts of pointing to a simple return. (Enabling FTRACE will +include the -pg switch in the compiling of the kernel.) At compile time every C file object is run through the recordmcount.pl script (located in the scripts directory). This script will process the C object using objdump to find all the -locations in the .text section that call mcount. (Note, only -the .text section is processed, since processing other sections -like .init.text may cause races due to those sections being freed). +locations in the .text section that call mcount. (Note, only the +.text section is processed, since processing other sections like +.init.text may cause races due to those sections being freed). -A new section called "__mcount_loc" is created that holds references -to all the mcount call sites in the .text section. This section is -compiled back into the original object. The final linker will add -all these references into a single table. +A new section called "__mcount_loc" is created that holds +references to all the mcount call sites in the .text section. +This section is compiled back into the original object. The +final linker will add all these references into a single table. On boot up, before SMP is initialized, the dynamic ftrace code -scans this table and updates all the locations into nops. It also -records the locations, which are added to the available_filter_functions -list. Modules are processed as they are loaded and before they are -executed. When a module is unloaded, it also removes its functions from -the ftrace function list. This is automatic in the module unload -code, and the module author does not need to worry about it. +scans this table and updates all the locations into nops. It +also records the locations, which are added to the +available_filter_functions list. Modules are processed as they +are loaded and before they are executed. When a module is +unloaded, it also removes its functions from the ftrace function +list. This is automatic in the module unload code, and the +module author does not need to worry about it. -When tracing is enabled, kstop_machine is called to prevent races -with the CPUS executing code being modified (which can cause the -CPU to do undesireable things), and the nops are patched back -to calls. But this time, they do not call mcount (which is just -a function stub). They now call into the ftrace infrastructure. +When tracing is enabled, kstop_machine is called to prevent +races with the CPUS executing code being modified (which can +cause the CPU to do undesireable things), and the nops are +patched back to calls. But this time, they do not call mcount +(which is just a function stub). They now call into the ftrace +infrastructure. One special side-effect to the recording of the functions being traced is that we can now selectively choose which functions we -wish to trace and which ones we want the mcount calls to remain as -nops. +wish to trace and which ones we want the mcount calls to remain +as nops. -Two files are used, one for enabling and one for disabling the tracing -of specified functions. They are: +Two files are used, one for enabling and one for disabling the +tracing of specified functions. They are: set_ftrace_filter @@ -1443,8 +1535,8 @@ and set_ftrace_notrace -A list of available functions that you can add to these files is listed -in: +A list of available functions that you can add to these files is +listed in: available_filter_functions @@ -1481,8 +1573,8 @@ hrtimer_interrupt sys_nanosleep -Perhaps this is not enough. The filters also allow simple wild cards. -Only the following are currently available +Perhaps this is not enough. The filters also allow simple wild +cards. Only the following are currently available * - will match functions that begin with * - will match functions that end with @@ -1492,9 +1584,9 @@ These are the only wild cards which are supported. * will not work. -Note: It is better to use quotes to enclose the wild cards, otherwise - the shell may expand the parameters into names of files in the local - directory. +Note: It is better to use quotes to enclose the wild cards, + otherwise the shell may expand the parameters into names + of files in the local directory. # echo 'hrtimer_*' > /debug/tracing/set_ftrace_filter @@ -1540,7 +1632,8 @@ This is because the '>' and '>>' act just like they do in bash. To rewrite the filters, use '>' To append to the filters, use '>>' -To clear out a filter so that all functions will be recorded again: +To clear out a filter so that all functions will be recorded +again: # echo > /debug/tracing/set_ftrace_filter # cat /debug/tracing/set_ftrace_filter @@ -1572,7 +1665,8 @@ hrtimer_get_res hrtimer_init_sleeper -The set_ftrace_notrace prevents those functions from being traced. +The set_ftrace_notrace prevents those functions from being +traced. # echo '*preempt*' '*lock*' > /debug/tracing/set_ftrace_notrace @@ -1595,18 +1689,20 @@ Produces: We can see that there's no more lock or preempt tracing. -* Dynamic ftrace with the function graph tracer * +Dynamic ftrace with the function graph tracer +--------------------------------------------- +Although what has been explained above concerns both the +function tracer and the function-graph-tracer, there are some +special features only available in the function-graph tracer. -Although what has been explained above concerns both the function tracer and -the function_graph_tracer, the following concerns only the latter. +If you want to trace only one function and all of its children, +you just have to echo its name into set_graph_function: -If you want to trace only one function and all of its childs, you just have -to echo its name on set_graph_function: + echo __do_fault > set_graph_function -echo __do_fault > set_graph_function - -will produce the following: +will produce the following "expanded" trace of the __do_fault() +function: 0) | __do_fault() { 0) | filemap_fault() { @@ -1643,23 +1739,24 @@ will produce the following: 0) 2.793 us | } 0) + 14.012 us | } -You can also select several functions: +You can also expand several functions at once: -echo sys_open > set_graph_function -echo sys_close >> set_graph_function + echo sys_open > set_graph_function + echo sys_close >> set_graph_function -Now if you want to go back to trace all functions +Now if you want to go back to trace all functions you can clear +this special filter via: -echo > set_graph_function + echo > set_graph_function trace_pipe ---------- -The trace_pipe outputs the same content as the trace file, but the effect -on the tracing is different. Every read from trace_pipe is consumed. -This means that subsequent reads will be different. The trace -is live. +The trace_pipe outputs the same content as the trace file, but +the effect on the tracing is different. Every read from +trace_pipe is consumed. This means that subsequent reads will be +different. The trace is live. # echo function > /debug/tracing/current_tracer # cat /debug/tracing/trace_pipe > /tmp/trace.out & @@ -1687,38 +1784,45 @@ is live. bash-4043 [00] 41.267111: select_task_rq_rt <-try_to_wake_up -Note, reading the trace_pipe file will block until more input is added. -By changing the tracer, trace_pipe will issue an EOF. We needed -to set the function tracer _before_ we "cat" the trace_pipe file. +Note, reading the trace_pipe file will block until more input is +added. By changing the tracer, trace_pipe will issue an EOF. We +needed to set the function tracer _before_ we "cat" the +trace_pipe file. trace entries ------------- -Having too much or not enough data can be troublesome in diagnosing -an issue in the kernel. The file buffer_size_kb is used to modify -the size of the internal trace buffers. The number listed -is the number of entries that can be recorded per CPU. To know -the full size, multiply the number of possible CPUS with the -number of entries. +Having too much or not enough data can be troublesome in +diagnosing an issue in the kernel. The file buffer_size_kb is +used to modify the size of the internal trace buffers. The +number listed is the number of entries that can be recorded per +CPU. To know the full size, multiply the number of possible CPUS +with the number of entries. # cat /debug/tracing/buffer_size_kb 1408 (units kilobytes) -Note, to modify this, you must have tracing completely disabled. To do that, -echo "nop" into the current_tracer. If the current_tracer is not set -to "nop", an EINVAL error will be returned. +Note, to modify this, you must have tracing completely disabled. +To do that, echo "nop" into the current_tracer. If the +current_tracer is not set to "nop", an EINVAL error will be +returned. # echo nop > /debug/tracing/current_tracer # echo 10000 > /debug/tracing/buffer_size_kb # cat /debug/tracing/buffer_size_kb 10000 (units kilobytes) -The number of pages which will be allocated is limited to a percentage -of available memory. Allocating too much will produce an error. +The number of pages which will be allocated is limited to a +percentage of available memory. Allocating too much will produce +an error. # echo 1000000000000 > /debug/tracing/buffer_size_kb -bash: echo: write error: Cannot allocate memory # cat /debug/tracing/buffer_size_kb 85 +----------- + +More details can be found in the source code, in the +kernel/tracing/*.c files. From 00a8bf859331e349713274825e6fbf20bf2ac15a Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 19 Feb 2009 13:01:37 +0100 Subject: [PATCH 169/413] tracing/function-graph-tracer: fix merge Merge artifact: pid got changed to ent->pid meanwhile. Signed-off-by: Ingo Molnar --- kernel/trace/trace_functions_graph.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 8f4004a00b4e..c009553a8e81 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -589,7 +589,7 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, /* Proc */ if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, pid); + ret = print_graph_proc(s, ent->pid); if (ret == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; From fe1200b63d158b28eef6d4de1e5b5f99c681ba2f Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Tue, 17 Feb 2009 12:05:07 -0500 Subject: [PATCH 170/413] SLUB: Introduce and use SLUB_MAX_SIZE and SLUB_PAGE_SHIFT constants As a preparational patch to bump up page allocator pass-through threshold, introduce two new constants SLUB_MAX_SIZE and SLUB_PAGE_SHIFT and convert mm/slub.c to use them. Reported-by: "Zhang, Yanmin" Tested-by: "Zhang, Yanmin" Signed-off-by: Christoph Lameter Signed-off-by: Pekka Enberg --- include/linux/slub_def.h | 19 ++++++++++++++++--- mm/slub.c | 16 ++++++++-------- 2 files changed, 24 insertions(+), 11 deletions(-) diff --git a/include/linux/slub_def.h b/include/linux/slub_def.h index 2f5c16b1aacd..986e09dcfd8f 100644 --- a/include/linux/slub_def.h +++ b/include/linux/slub_def.h @@ -120,11 +120,24 @@ struct kmem_cache { #define KMALLOC_SHIFT_LOW ilog2(KMALLOC_MIN_SIZE) +/* + * Maximum kmalloc object size handled by SLUB. Larger object allocations + * are passed through to the page allocator. The page allocator "fastpath" + * is relatively slow so we need this value sufficiently high so that + * performance critical objects are allocated through the SLUB fastpath. + * + * This should be dropped to PAGE_SIZE / 2 once the page allocator + * "fastpath" becomes competitive with the slab allocator fastpaths. + */ +#define SLUB_MAX_SIZE (PAGE_SIZE) + +#define SLUB_PAGE_SHIFT (PAGE_SHIFT + 1) + /* * We keep the general caches in an array of slab caches that are used for * 2^x bytes of allocations. */ -extern struct kmem_cache kmalloc_caches[PAGE_SHIFT + 1]; +extern struct kmem_cache kmalloc_caches[SLUB_PAGE_SHIFT]; /* * Sorry that the following has to be that ugly but some versions of GCC @@ -212,7 +225,7 @@ static __always_inline void *kmalloc_large(size_t size, gfp_t flags) static __always_inline void *kmalloc(size_t size, gfp_t flags) { if (__builtin_constant_p(size)) { - if (size > PAGE_SIZE) + if (size > SLUB_MAX_SIZE) return kmalloc_large(size, flags); if (!(flags & SLUB_DMA)) { @@ -234,7 +247,7 @@ void *kmem_cache_alloc_node(struct kmem_cache *, gfp_t flags, int node); static __always_inline void *kmalloc_node(size_t size, gfp_t flags, int node) { if (__builtin_constant_p(size) && - size <= PAGE_SIZE && !(flags & SLUB_DMA)) { + size <= SLUB_MAX_SIZE && !(flags & SLUB_DMA)) { struct kmem_cache *s = kmalloc_slab(size); if (!s) diff --git a/mm/slub.c b/mm/slub.c index bdc9abb08a23..5a5e7f5bf799 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -2475,7 +2475,7 @@ EXPORT_SYMBOL(kmem_cache_destroy); * Kmalloc subsystem *******************************************************************/ -struct kmem_cache kmalloc_caches[PAGE_SHIFT + 1] __cacheline_aligned; +struct kmem_cache kmalloc_caches[SLUB_PAGE_SHIFT] __cacheline_aligned; EXPORT_SYMBOL(kmalloc_caches); static int __init setup_slub_min_order(char *str) @@ -2537,7 +2537,7 @@ static struct kmem_cache *create_kmalloc_cache(struct kmem_cache *s, } #ifdef CONFIG_ZONE_DMA -static struct kmem_cache *kmalloc_caches_dma[PAGE_SHIFT + 1]; +static struct kmem_cache *kmalloc_caches_dma[SLUB_PAGE_SHIFT]; static void sysfs_add_func(struct work_struct *w) { @@ -2658,7 +2658,7 @@ void *__kmalloc(size_t size, gfp_t flags) { struct kmem_cache *s; - if (unlikely(size > PAGE_SIZE)) + if (unlikely(size > SLUB_MAX_SIZE)) return kmalloc_large(size, flags); s = get_slab(size, flags); @@ -2686,7 +2686,7 @@ void *__kmalloc_node(size_t size, gfp_t flags, int node) { struct kmem_cache *s; - if (unlikely(size > PAGE_SIZE)) + if (unlikely(size > SLUB_MAX_SIZE)) return kmalloc_large_node(size, flags, node); s = get_slab(size, flags); @@ -2985,7 +2985,7 @@ void __init kmem_cache_init(void) caches++; } - for (i = KMALLOC_SHIFT_LOW; i <= PAGE_SHIFT; i++) { + for (i = KMALLOC_SHIFT_LOW; i < SLUB_PAGE_SHIFT; i++) { create_kmalloc_cache(&kmalloc_caches[i], "kmalloc", 1 << i, GFP_KERNEL); caches++; @@ -3022,7 +3022,7 @@ void __init kmem_cache_init(void) slab_state = UP; /* Provide the correct kmalloc names now that the caches are up */ - for (i = KMALLOC_SHIFT_LOW; i <= PAGE_SHIFT; i++) + for (i = KMALLOC_SHIFT_LOW; i < SLUB_PAGE_SHIFT; i++) kmalloc_caches[i]. name = kasprintf(GFP_KERNEL, "kmalloc-%d", 1 << i); @@ -3222,7 +3222,7 @@ void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, unsigned long caller) { struct kmem_cache *s; - if (unlikely(size > PAGE_SIZE)) + if (unlikely(size > SLUB_MAX_SIZE)) return kmalloc_large(size, gfpflags); s = get_slab(size, gfpflags); @@ -3238,7 +3238,7 @@ void *__kmalloc_node_track_caller(size_t size, gfp_t gfpflags, { struct kmem_cache *s; - if (unlikely(size > PAGE_SIZE)) + if (unlikely(size > SLUB_MAX_SIZE)) return kmalloc_large_node(size, gfpflags, node); s = get_slab(size, gfpflags); From f9349a8f978929a0c71d2c42ae299f7d462c239d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 19 Feb 2009 21:13:12 +0100 Subject: [PATCH 171/413] tracing/function-graph-tracer: make set_graph_function file support ftrace regex Impact: trace only functions matching a pattern The set_graph_function file let one to trace only one or several chosen functions and follow all their code flow. Currently, only a constant function name is allowed so this patch allows the ftrace_regex functions: - matches all functions that end with "name": echo *name > set_graph_function - matches all functions that begin with "name": echo name* > set_graph_function - matches all functions that contains "name": echo *name* > set_graph_function Example: echo mutex* > set_graph_function 0) | mutex_lock_nested() { 0) 0.563 us | __might_sleep(); 0) 2.072 us | } 0) | mutex_unlock() { 0) 1.036 us | __mutex_unlock_slowpath(); 0) 2.433 us | } 0) | mutex_unlock() { 0) 0.691 us | __mutex_unlock_slowpath(); 0) 1.787 us | } 0) | mutex_lock_interruptible_nested() { 0) 0.548 us | __might_sleep(); 0) 1.945 us | } Signed-off-by: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 56 +++++++++++++++++++++++++++++-------------- 1 file changed, 38 insertions(+), 18 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 7dd5a2bef9cd..cf59f4c54745 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1895,6 +1895,10 @@ static void *g_start(struct seq_file *m, loff_t *pos) mutex_lock(&graph_lock); + /* Nothing, tell g_show to print all functions are enabled */ + if (!ftrace_graph_count && !*pos) + return (void *)1; + p = g_next(m, p, pos); return p; @@ -1913,6 +1917,11 @@ static int g_show(struct seq_file *m, void *v) if (!ptr) return 0; + if (ptr == (unsigned long *)1) { + seq_printf(m, "#### all functions enabled ####\n"); + return 0; + } + kallsyms_lookup(*ptr, NULL, NULL, NULL, str); seq_printf(m, "%s\n", str); @@ -1966,38 +1975,51 @@ ftrace_graph_read(struct file *file, char __user *ubuf, } static int -ftrace_set_func(unsigned long *array, int idx, char *buffer) +ftrace_set_func(unsigned long *array, int *idx, char *buffer) { - char str[KSYM_SYMBOL_LEN]; struct dyn_ftrace *rec; struct ftrace_page *pg; + int search_len; int found = 0; - int j; + int type, not; + char *search; + bool exists; + int i; if (ftrace_disabled) return -ENODEV; + /* decode regex */ + type = ftrace_setup_glob(buffer, strlen(buffer), &search, ¬); + if (not) + return -EINVAL; + + search_len = strlen(search); + mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { + if (*idx >= FTRACE_GRAPH_MAX_FUNCS) + break; + if (rec->flags & (FTRACE_FL_FAILED | FTRACE_FL_FREE)) continue; - kallsyms_lookup(rec->ip, NULL, NULL, NULL, str); - if (strcmp(str, buffer) == 0) { - /* Return 1 if we add it to the array */ - found = 1; - for (j = 0; j < idx; j++) - if (array[j] == rec->ip) { - found = 0; + if (ftrace_match_record(rec, search, search_len, type)) { + /* ensure it is not already in the array */ + exists = false; + for (i = 0; i < *idx; i++) + if (array[i] == rec->ip) { + exists = true; break; } - if (found) - array[idx] = rec->ip; - goto out; + if (!exists) { + array[(*idx)++] = rec->ip; + found = 1; + } } } while_for_each_ftrace_rec(); - out: + mutex_unlock(&ftrace_lock); return found ? 0 : -EINVAL; @@ -2066,13 +2088,11 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, } buffer[index] = 0; - /* we allow only one at a time */ - ret = ftrace_set_func(array, ftrace_graph_count, buffer); + /* we allow only one expression at a time */ + ret = ftrace_set_func(array, &ftrace_graph_count, buffer); if (ret) goto out; - ftrace_graph_count++; - file->f_pos += read; ret = read; From 000ab691172db3921efa3cb7f17fc79235a1de7f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Feb 2009 13:35:06 -0500 Subject: [PATCH 172/413] ftrace: allow archs to preform pre and post process for code modification This patch creates the weak functions: ftrace_arch_code_modify_prepare and ftrace_arch_code_modify_post_process that are called before and after the stop machine is called to modify the kernel text. If the arch needs to do pre or post processing, it only needs to define these functions. [ Update: Ingo Molnar suggested using the name ftrace_arch_code_modify_* over using ftrace_arch_modify_* ] Signed-off-by: Steven Rostedt --- include/linux/ftrace.h | 3 +++ kernel/trace/ftrace.c | 28 ++++++++++++++++++++++++++++ 2 files changed, 31 insertions(+) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 677432b9cb7e..fdb2a89ae543 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -99,6 +99,9 @@ stack_trace_sysctl(struct ctl_table *table, int write, /* asm/ftrace.h must be defined for archs supporting dynamic ftrace */ #include +int ftrace_arch_code_modify_prepare(void); +int ftrace_arch_code_modify_post_process(void); + enum { FTRACE_FL_FREE = (1 << 0), FTRACE_FL_FAILED = (1 << 1), diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index fdf913dfc7e8..72316d9647bd 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -585,6 +585,24 @@ ftrace_code_disable(struct module *mod, struct dyn_ftrace *rec) return 1; } +/* + * archs can override this function if they must do something + * before the modifying code is performed. + */ +int __weak ftrace_arch_code_modify_prepare(void) +{ + return 0; +} + +/* + * archs can override this function if they must do something + * after the modifying code is performed. + */ +int __weak ftrace_arch_code_modify_post_process(void) +{ + return 0; +} + static int __ftrace_modify_code(void *data) { int *command = data; @@ -607,7 +625,17 @@ static int __ftrace_modify_code(void *data) static void ftrace_run_update_code(int command) { + int ret; + + ret = ftrace_arch_code_modify_prepare(); + FTRACE_WARN_ON(ret); + if (ret) + return; + stop_machine(__ftrace_modify_code, &command, NULL); + + ret = ftrace_arch_code_modify_post_process(); + FTRACE_WARN_ON(ret); } static ftrace_func_t saved_ftrace_func; From 91f73f90d97fa67effbb49e0a79c50cf26dfe324 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 20 Feb 2009 17:34:06 +0100 Subject: [PATCH 173/413] tracing/markers: make markers select tracepoints Sometimes it happens that KConfig dependencies are not handled like in the following scenario: - config A bool - config B bool depends on A - config C bool select B If one selects C, then it will select B without checking its dependency to A, if A hasn't been selected elsewhere, it will result in a build failure. This is what happens on the following build error: kernel/built-in.o: In function `marker_update_probe_range': (.text+0x52f64): undefined reference to `tracepoint_probe_register_noupdate' kernel/built-in.o: In function `marker_update_probe_range': (.text+0x52f74): undefined reference to `tracepoint_probe_unregister_noupdate' kernel/built-in.o: In function `marker_update_probe_range': (.text+0x52fb9): undefined reference to `tracepoint_probe_unregister_noupdate' kernel/built-in.o: In function `marker_update_probes': marker.c:(.text+0x530ba): undefined reference to `tracepoint_probe_update_all' CONFIG_KVM_TRACE will select CONFIG_MARKER, but the latter depends on CONFIG_TRACEPOINTS which will not be selected. Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- init/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/init/Kconfig b/init/Kconfig index f068071fcc5d..26b5bab6f6e8 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -945,7 +945,7 @@ config TRACEPOINTS config MARKERS bool "Activate markers" - depends on TRACEPOINTS + select TRACEPOINTS help Place an empty function call at each marker site. Can be dynamically changed for a probe function. From 16239630974516a8879a3695ee9b4dc661f79f96 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Feb 2009 17:57:30 -0500 Subject: [PATCH 174/413] ftrace, x86: make kernel text writable only for conversions Impact: keep kernel text read only Because dynamic ftrace converts the calls to mcount into and out of nops at run time, we needed to always keep the kernel text writable. But this defeats the point of CONFIG_DEBUG_RODATA. This patch converts the kernel code to writable before ftrace modifies the text, and converts it back to read only afterward. The kernel text is converted to read/write, stop_machine is called to modify the code, then the kernel text is converted back to read only. The original version used SYSTEM_STATE to determine when it was OK or not to change the code to rw or ro. Andrew Morton pointed out that using SYSTEM_STATE is a bad idea since there is no guarantee to what its state will actually be. Instead, I moved the check into the set_kernel_text_* functions themselves, and use a local variable to determine when it is OK to change the kernel text RW permissions. [ Update: Ingo Molnar suggested moving the prototypes to cacheflush.h ] Reviewed-by: Andrew Morton Signed-off-by: Steven Rostedt --- arch/x86/include/asm/cacheflush.h | 5 +++++ arch/x86/kernel/ftrace.c | 13 +++++++++++ arch/x86/mm/init_32.c | 35 ++++++++++++++++++++++++++--- arch/x86/mm/init_64.c | 37 ++++++++++++++++++++++++++----- 4 files changed, 82 insertions(+), 8 deletions(-) diff --git a/arch/x86/include/asm/cacheflush.h b/arch/x86/include/asm/cacheflush.h index 2f8466540fb5..6145063cfe0e 100644 --- a/arch/x86/include/asm/cacheflush.h +++ b/arch/x86/include/asm/cacheflush.h @@ -104,6 +104,11 @@ void clflush_cache_range(void *addr, unsigned int size); #ifdef CONFIG_DEBUG_RODATA void mark_rodata_ro(void); extern const int rodata_test_data; +void set_kernel_text_rw(void); +void set_kernel_text_ro(void); +#else +static inline void set_kernel_text_rw(void) { } +static inline void set_kernel_text_ro(void) { } #endif #ifdef CONFIG_DEBUG_RODATA_TEST diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 231bdd3c5b1c..77857d4f7d0f 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -18,6 +18,7 @@ #include #include +#include #include #include #include @@ -26,6 +27,18 @@ #ifdef CONFIG_DYNAMIC_FTRACE +int ftrace_arch_code_modify_prepare(void) +{ + set_kernel_text_rw(); + return 0; +} + +int ftrace_arch_code_modify_post_process(void) +{ + set_kernel_text_ro(); + return 0; +} + union ftrace_code_union { char code[MCOUNT_INSN_SIZE]; struct { diff --git a/arch/x86/mm/init_32.c b/arch/x86/mm/init_32.c index 2cef05074413..3eb2ed188a4c 100644 --- a/arch/x86/mm/init_32.c +++ b/arch/x86/mm/init_32.c @@ -1155,17 +1155,47 @@ static noinline int do_test_wp_bit(void) const int rodata_test_data = 0xC3; EXPORT_SYMBOL_GPL(rodata_test_data); +static int kernel_set_to_readonly; + +void set_kernel_text_rw(void) +{ + unsigned long start = PFN_ALIGN(_text); + unsigned long size = PFN_ALIGN(_etext) - start; + + if (!kernel_set_to_readonly) + return; + + pr_debug("Set kernel text: %lx - %lx for read write\n", + start, start+size); + + set_pages_rw(virt_to_page(start), size >> PAGE_SHIFT); +} + +void set_kernel_text_ro(void) +{ + unsigned long start = PFN_ALIGN(_text); + unsigned long size = PFN_ALIGN(_etext) - start; + + if (!kernel_set_to_readonly) + return; + + pr_debug("Set kernel text: %lx - %lx for read only\n", + start, start+size); + + set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT); +} + void mark_rodata_ro(void) { unsigned long start = PFN_ALIGN(_text); unsigned long size = PFN_ALIGN(_etext) - start; -#ifndef CONFIG_DYNAMIC_FTRACE - /* Dynamic tracing modifies the kernel text section */ set_pages_ro(virt_to_page(start), size >> PAGE_SHIFT); printk(KERN_INFO "Write protecting the kernel text: %luk\n", size >> 10); + kernel_set_to_readonly = 1; + #ifdef CONFIG_CPA_DEBUG printk(KERN_INFO "Testing CPA: Reverting %lx-%lx\n", start, start+size); @@ -1174,7 +1204,6 @@ void mark_rodata_ro(void) printk(KERN_INFO "Testing CPA: write protecting again\n"); set_pages_ro(virt_to_page(start), size>>PAGE_SHIFT); #endif -#endif /* CONFIG_DYNAMIC_FTRACE */ start += size; size = (unsigned long)__end_rodata - start; diff --git a/arch/x86/mm/init_64.c b/arch/x86/mm/init_64.c index e6d36b490250..63fdc531601d 100644 --- a/arch/x86/mm/init_64.c +++ b/arch/x86/mm/init_64.c @@ -986,21 +986,48 @@ void free_initmem(void) const int rodata_test_data = 0xC3; EXPORT_SYMBOL_GPL(rodata_test_data); +static int kernel_set_to_readonly; + +void set_kernel_text_rw(void) +{ + unsigned long start = PFN_ALIGN(_stext); + unsigned long end = PFN_ALIGN(__start_rodata); + + if (!kernel_set_to_readonly) + return; + + pr_debug("Set kernel text: %lx - %lx for read write\n", + start, end); + + set_memory_rw(start, (end - start) >> PAGE_SHIFT); +} + +void set_kernel_text_ro(void) +{ + unsigned long start = PFN_ALIGN(_stext); + unsigned long end = PFN_ALIGN(__start_rodata); + + if (!kernel_set_to_readonly) + return; + + pr_debug("Set kernel text: %lx - %lx for read only\n", + start, end); + + set_memory_ro(start, (end - start) >> PAGE_SHIFT); +} + void mark_rodata_ro(void) { unsigned long start = PFN_ALIGN(_stext), end = PFN_ALIGN(__end_rodata); unsigned long rodata_start = ((unsigned long)__start_rodata + PAGE_SIZE - 1) & PAGE_MASK; -#ifdef CONFIG_DYNAMIC_FTRACE - /* Dynamic tracing modifies the kernel text section */ - start = rodata_start; -#endif - printk(KERN_INFO "Write protecting the kernel read-only data: %luk\n", (end - start) >> 10); set_memory_ro(start, (end - start) >> PAGE_SHIFT); + kernel_set_to_readonly = 1; + /* * The rodata section (but not the kernel text!) should also be * not-executable. From 90c7ac49aa819feb9433b5310089fca6399881c0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 19 Feb 2009 13:32:57 -0500 Subject: [PATCH 175/413] ftrace: immediately stop code modification if failure is detected Impact: fix to prevent NMI lockup If the page fault handler produces a WARN_ON in the modifying of text, and the system is setup to have a high frequency of NMIs, we can lock up the system on a failure to modify code. The modifying of code with NMIs allows all NMIs to modify the code if it is about to run. This prevents a modifier on one CPU from modifying code running in NMI context on another CPU. The modifying is done through stop_machine, so only NMIs must be considered. But if the write causes the page fault handler to produce a warning, the print can slow it down enough that as soon as it is done it will take another NMI before going back to the process context. The new NMI will perform the write again causing another print and this will hang the box. This patch turns off the writing as soon as a failure is detected and does not wait for it to be turned off by the process context. This will keep NMIs from getting stuck in this back and forth of print outs. Signed-off-by: Steven Rostedt --- arch/x86/kernel/ftrace.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 77857d4f7d0f..c56d73894322 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -124,6 +124,10 @@ static void ftrace_mod_code(void) */ mod_code_status = probe_kernel_write(mod_code_ip, mod_code_newcode, MCOUNT_INSN_SIZE); + + /* if we fail, then kill any new writers */ + if (mod_code_status) + mod_code_write = 0; } void ftrace_nmi_enter(void) From 4377245aa93b65b6597e4b7bb460fb9abc48b56b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 19 Feb 2009 13:41:27 -0500 Subject: [PATCH 176/413] ftrace: break out modify loop immediately on detection of error Impact: added precaution on failure detection Break out of the modifying loop as soon as a failure is detected. This is just an added precaution found by code review and was not found by any bug chasing. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 72316d9647bd..11ad796ca049 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -561,8 +561,11 @@ static void ftrace_replace_code(int enable) if ((system_state == SYSTEM_BOOTING) || !core_kernel_text(rec->ip)) { ftrace_free_rec(rec); - } else + } else { ftrace_bug(failed, rec->ip); + /* Stop processing */ + return; + } } } } From 5e01cb695d29619dd551bac7d6aa4ef1dc8ebc95 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 24 Feb 2009 13:55:18 +0100 Subject: [PATCH 177/413] x86, ftrace: fix section mismatch in hw-branch-tracer Fix an invalid memory reference problem when cpu hotplug support is disabled and the hw-branch-tracer is set as current tracer. Initializing the tracer calls bts_trace_init() which has already been freed at this time. Reported-by: Frederic Weisbecker Signed-off-by: Markus Metzger Signed-off-by: Ingo Molnar --- kernel/trace/trace_hw_branches.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index 3561aace075c..3335e807144b 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -127,20 +127,18 @@ static struct notifier_block bts_hotcpu_notifier __cpuinitdata = { .notifier_call = bts_hotcpu_handler }; -static int __cpuinit bts_trace_init(struct trace_array *tr) +static int bts_trace_init(struct trace_array *tr) { hw_branch_trace = tr; - register_hotcpu_notifier(&bts_hotcpu_notifier); bts_trace_start(tr); return 0; } -static void __cpuinit bts_trace_reset(struct trace_array *tr) +static void bts_trace_reset(struct trace_array *tr) { bts_trace_stop(tr); - unregister_hotcpu_notifier(&bts_hotcpu_notifier); } static void bts_trace_print_header(struct seq_file *m) @@ -299,6 +297,7 @@ struct tracer bts_tracer __read_mostly = __init static int init_bts_trace(void) { + register_hotcpu_notifier(&bts_hotcpu_notifier); return register_tracer(&bts_tracer); } device_initcall(init_bts_trace); From 499aa86dcbc3c4daf7d2c59c5c30e1a78220fbc1 Mon Sep 17 00:00:00 2001 From: Markus Metzger Date: Tue, 24 Feb 2009 14:12:34 +0100 Subject: [PATCH 178/413] x86, ptrace: remove CONFIG guards around declarations Remove unnecessary CONFIG guards around type declarations and macro definitions. Reported-by: Cyrill Gorcunov Signed-off-by: Markus Metzger Cc: markus.t.metzger@gmail.com Cc: roland@redhat.com Cc: eranian@googlemail.com Cc: oleg@redhat.com Cc: juan.villacis@intel.com Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ptrace-abi.h | 3 --- 1 file changed, 3 deletions(-) diff --git a/arch/x86/include/asm/ptrace-abi.h b/arch/x86/include/asm/ptrace-abi.h index 8e0f8d199e05..86723035a515 100644 --- a/arch/x86/include/asm/ptrace-abi.h +++ b/arch/x86/include/asm/ptrace-abi.h @@ -80,8 +80,6 @@ #define PTRACE_SINGLEBLOCK 33 /* resume execution until next branch */ -#ifdef CONFIG_X86_PTRACE_BTS - #ifndef __ASSEMBLY__ #include @@ -140,6 +138,5 @@ struct ptrace_bts_config { BTS records are read from oldest to newest. Returns number of BTS records drained. */ -#endif /* CONFIG_X86_PTRACE_BTS */ #endif /* _ASM_X86_PTRACE_ABI_H */ From 7c37730cd31ddb2d3a1da142af9b18c29b8c433b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 24 Feb 2009 12:07:53 -0500 Subject: [PATCH 179/413] tracing: add DEFINE_TRACE_FMT to tracepoint.h This patch creates a DEFINE_TRACE_FMT to map to DECLARE_TRACE. This allows for the developers to place format strings and args in with their tracepoint declaration. A tracer may now override the DEFINE_TRACE_FMT macro and use it to record a default format. Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 3 +++ 1 file changed, 3 insertions(+) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 757005458366..34ae464effff 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -153,4 +153,7 @@ static inline void tracepoint_synchronize_unregister(void) synchronize_sched(); } +#define DEFINE_TRACE_FMT(name, proto, args, fmt) \ + DECLARE_TRACE(name, TPPROTO(proto), TPARGS(args)) + #endif From b77e38aa240c3bd9c55c98b9f7c81541e042eae5 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 24 Feb 2009 10:21:36 -0500 Subject: [PATCH 180/413] tracing: add event trace infrastructure This patch creates the event tracing infrastructure of ftrace. It will create the files: /debug/tracing/available_events /debug/tracing/set_event The available_events will list the trace points that have been registered with the event tracer. set_events will allow the user to enable or disable an event hook. example: # echo sched_wakeup > /debug/tracing/set_event Will enable the sched_wakeup event (if it is registered). # echo "!sched_wakeup" >> /debug/tracing/set_event Will disable the sched_wakeup event (and only that event). # echo > /debug/tracing/set_event Will disable all events (notice the '>') # cat /debug/tracing/available_events > /debug/tracing/set_event Will enable all registered event hooks. Signed-off-by: Steven Rostedt --- include/asm-generic/vmlinux.lds.h | 11 +- kernel/trace/Kconfig | 9 + kernel/trace/Makefile | 1 + kernel/trace/trace_events.c | 280 ++++++++++++++++++++++++++++++ kernel/trace/trace_events.h | 52 ++++++ 5 files changed, 352 insertions(+), 1 deletion(-) create mode 100644 kernel/trace/trace_events.c create mode 100644 kernel/trace/trace_events.h diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index c61fab1dd2f8..0add6b28c366 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -61,6 +61,14 @@ #define BRANCH_PROFILE() #endif +#ifdef CONFIG_EVENT_TRACER +#define FTRACE_EVENTS() VMLINUX_SYMBOL(__start_ftrace_events) = .; \ + *(_ftrace_events) \ + VMLINUX_SYMBOL(__stop_ftrace_events) = .; +#else +#define FTRACE_EVENTS() +#endif + /* .data section */ #define DATA_DATA \ *(.data) \ @@ -81,7 +89,8 @@ *(__tracepoints) \ VMLINUX_SYMBOL(__stop___tracepoints) = .; \ LIKELY_PROFILE() \ - BRANCH_PROFILE() + BRANCH_PROFILE() \ + FTRACE_EVENTS() #define RO_DATA(align) \ . = ALIGN((align)); \ diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 07877f4b5233..999c6a2485df 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -159,6 +159,15 @@ config CONTEXT_SWITCH_TRACER This tracer gets called from the context switch and records all switching of tasks. +config EVENT_TRACER + bool "Trace various events in the kernel" + depends on DEBUG_KERNEL + select TRACING + help + This tracer hooks to various trace points in the kernel + allowing the user to pick and choose which trace point they + want to trace. + config BOOT_TRACER bool "Trace boot initcalls" depends on DEBUG_KERNEL diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 627090bc262d..c7363568b1cf 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -38,5 +38,6 @@ obj-$(CONFIG_POWER_TRACER) += trace_power.o obj-$(CONFIG_KMEMTRACE) += kmemtrace.o obj-$(CONFIG_WORKQUEUE_TRACER) += trace_workqueue.o obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o +obj-$(CONFIG_EVENT_TRACER) += trace_events.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c new file mode 100644 index 000000000000..05bc80ec8d2c --- /dev/null +++ b/kernel/trace/trace_events.c @@ -0,0 +1,280 @@ +/* + * event tracer + * + * Copyright (C) 2008 Red Hat Inc, Steven Rostedt + * + */ + +#include +#include +#include +#include + +#include "trace_events.h" + +void event_trace_printk(unsigned long ip, const char *fmt, ...) +{ + va_list ap; + + va_start(ap, fmt); + tracing_record_cmdline(current); + trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); + va_end(ap); +} + +static void ftrace_clear_events(void) +{ + struct ftrace_event_call *call = (void *)__start_ftrace_events; + + + while ((unsigned long)call < (unsigned long)__stop_ftrace_events) { + + if (call->enabled) { + call->enabled = 0; + call->unregfunc(); + } + call++; + } +} + +static int ftrace_set_clr_event(char *buf, int set) +{ + struct ftrace_event_call *call = (void *)__start_ftrace_events; + + + while ((unsigned long)call < (unsigned long)__stop_ftrace_events) { + + if (strcmp(buf, call->name) != 0) { + call++; + continue; + } + + if (set) { + /* Already set? */ + if (call->enabled) + return 0; + call->enabled = 1; + call->regfunc(); + } else { + /* Already cleared? */ + if (!call->enabled) + return 0; + call->enabled = 0; + call->unregfunc(); + } + return 0; + } + return -EINVAL; +} + +/* 128 should be much more than enough */ +#define EVENT_BUF_SIZE 127 + +static ssize_t +ftrace_event_write(struct file *file, const char __user *ubuf, + size_t cnt, loff_t *ppos) +{ + size_t read = 0; + int i, set = 1; + ssize_t ret; + char *buf; + char ch; + + if (!cnt || cnt < 0) + return 0; + + ret = get_user(ch, ubuf++); + if (ret) + return ret; + read++; + cnt--; + + /* skip white space */ + while (cnt && isspace(ch)) { + ret = get_user(ch, ubuf++); + if (ret) + return ret; + read++; + cnt--; + } + + /* Only white space found? */ + if (isspace(ch)) { + file->f_pos += read; + ret = read; + return ret; + } + + buf = kmalloc(EVENT_BUF_SIZE+1, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + if (cnt > EVENT_BUF_SIZE) + cnt = EVENT_BUF_SIZE; + + i = 0; + while (cnt && !isspace(ch)) { + if (!i && ch == '!') + set = 0; + else + buf[i++] = ch; + + ret = get_user(ch, ubuf++); + if (ret) + goto out_free; + read++; + cnt--; + } + buf[i] = 0; + + file->f_pos += read; + + ret = ftrace_set_clr_event(buf, set); + if (ret) + goto out_free; + + ret = read; + + out_free: + kfree(buf); + + return ret; +} + +static void * +t_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct ftrace_event_call *call = m->private; + struct ftrace_event_call *next = call; + + (*pos)++; + + if ((unsigned long)call >= (unsigned long)__stop_ftrace_events) + return NULL; + + m->private = ++next; + + return call; +} + +static void *t_start(struct seq_file *m, loff_t *pos) +{ + return t_next(m, NULL, pos); +} + +static void * +s_next(struct seq_file *m, void *v, loff_t *pos) +{ + struct ftrace_event_call *call = m->private; + struct ftrace_event_call *next; + + (*pos)++; + + retry: + if ((unsigned long)call >= (unsigned long)__stop_ftrace_events) + return NULL; + + if (!call->enabled) { + call++; + goto retry; + } + + next = call; + m->private = ++next; + + return call; +} + +static void *s_start(struct seq_file *m, loff_t *pos) +{ + return s_next(m, NULL, pos); +} + +static int t_show(struct seq_file *m, void *v) +{ + struct ftrace_event_call *call = v; + + seq_printf(m, "%s\n", call->name); + + return 0; +} + +static void t_stop(struct seq_file *m, void *p) +{ +} + +static int +ftrace_event_seq_open(struct inode *inode, struct file *file) +{ + int ret; + const struct seq_operations *seq_ops; + + if ((file->f_mode & FMODE_WRITE) && + !(file->f_flags & O_APPEND)) + ftrace_clear_events(); + + seq_ops = inode->i_private; + ret = seq_open(file, seq_ops); + if (!ret) { + struct seq_file *m = file->private_data; + + m->private = __start_ftrace_events; + } + return ret; +} + +static const struct seq_operations show_event_seq_ops = { + .start = t_start, + .next = t_next, + .show = t_show, + .stop = t_stop, +}; + +static const struct seq_operations show_set_event_seq_ops = { + .start = s_start, + .next = s_next, + .show = t_show, + .stop = t_stop, +}; + +static const struct file_operations ftrace_avail_fops = { + .open = ftrace_event_seq_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +static const struct file_operations ftrace_set_event_fops = { + .open = ftrace_event_seq_open, + .read = seq_read, + .write = ftrace_event_write, + .llseek = seq_lseek, + .release = seq_release, +}; + +static __init int event_trace_init(void) +{ + struct dentry *d_tracer; + struct dentry *entry; + + d_tracer = tracing_init_dentry(); + if (!d_tracer) + return 0; + + entry = debugfs_create_file("available_events", 0444, d_tracer, + (void *)&show_event_seq_ops, + &ftrace_avail_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'available_events' entry\n"); + + entry = debugfs_create_file("set_event", 0644, d_tracer, + (void *)&show_set_event_seq_ops, + &ftrace_set_event_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'set_event' entry\n"); + + return 0; +} +fs_initcall(event_trace_init); diff --git a/kernel/trace/trace_events.h b/kernel/trace/trace_events.h new file mode 100644 index 000000000000..39342f86db27 --- /dev/null +++ b/kernel/trace/trace_events.h @@ -0,0 +1,52 @@ +#ifndef _LINUX_KERNEL_TRACE_EVENTS_H +#define _LINUX_KERNEL_TRACE_EVENTS_H + +#include +#include "trace.h" + +struct ftrace_event_call { + char *name; + int enabled; + int (*regfunc)(void); + void (*unregfunc)(void); +}; + + +#undef TPFMT +#define TPFMT(fmt, args...) fmt "\n", ##args + +#undef DEFINE_TRACE_FMT +#define DEFINE_TRACE_FMT(call, proto, args, fmt) \ +static void ftrace_event_##call(proto) \ +{ \ + event_trace_printk(_RET_IP_, "(" #call ") " fmt); \ +} \ + \ +static int ftrace_reg_event_##call(void) \ +{ \ + int ret; \ + \ + ret = register_trace_##call(ftrace_event_##call); \ + if (!ret) \ + pr_info("event trace: Could not activate trace point " \ + "probe to " #call); \ + return ret; \ +} \ + \ +static void ftrace_unreg_event_##call(void) \ +{ \ + unregister_trace_##call(ftrace_event_##call); \ +} \ + \ +static struct ftrace_event_call __used \ +__attribute__((section("_ftrace_events"))) event_##call = { \ + .name = #call, \ + .regfunc = ftrace_reg_event_##call, \ + .unregfunc = ftrace_unreg_event_##call, \ +} + +void event_trace_printk(unsigned long ip, const char *fmt, ...); +extern unsigned long __start_ftrace_events[]; +extern unsigned long __stop_ftrace_events[]; + +#endif /* _LINUX_KERNEL_TRACE_EVENTS_H */ From f3fe8e4a38fd19dbb3f8ffb1826aa840ae304a65 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 24 Feb 2009 10:22:57 -0500 Subject: [PATCH 181/413] tracing: add schedule events to event trace This patch changes the trace/sched.h to use the DECLARE_TRACE_FMT such that they are automatically registered with the event tracer. And it also adds the tracing sched headers to kernel/trace/events.c Signed-off-by: Steven Rostedt --- include/trace/sched.h | 49 +-------------------- include/trace/sched_event_types.h | 72 +++++++++++++++++++++++++++++++ kernel/trace/Makefile | 1 + kernel/trace/events.c | 13 ++++++ 4 files changed, 87 insertions(+), 48 deletions(-) create mode 100644 include/trace/sched_event_types.h create mode 100644 kernel/trace/events.c diff --git a/include/trace/sched.h b/include/trace/sched.h index 0d81098ee9fc..4e372a1a29bf 100644 --- a/include/trace/sched.h +++ b/include/trace/sched.h @@ -4,53 +4,6 @@ #include #include -DECLARE_TRACE(sched_kthread_stop, - TPPROTO(struct task_struct *t), - TPARGS(t)); - -DECLARE_TRACE(sched_kthread_stop_ret, - TPPROTO(int ret), - TPARGS(ret)); - -DECLARE_TRACE(sched_wait_task, - TPPROTO(struct rq *rq, struct task_struct *p), - TPARGS(rq, p)); - -DECLARE_TRACE(sched_wakeup, - TPPROTO(struct rq *rq, struct task_struct *p, int success), - TPARGS(rq, p, success)); - -DECLARE_TRACE(sched_wakeup_new, - TPPROTO(struct rq *rq, struct task_struct *p, int success), - TPARGS(rq, p, success)); - -DECLARE_TRACE(sched_switch, - TPPROTO(struct rq *rq, struct task_struct *prev, - struct task_struct *next), - TPARGS(rq, prev, next)); - -DECLARE_TRACE(sched_migrate_task, - TPPROTO(struct task_struct *p, int orig_cpu, int dest_cpu), - TPARGS(p, orig_cpu, dest_cpu)); - -DECLARE_TRACE(sched_process_free, - TPPROTO(struct task_struct *p), - TPARGS(p)); - -DECLARE_TRACE(sched_process_exit, - TPPROTO(struct task_struct *p), - TPARGS(p)); - -DECLARE_TRACE(sched_process_wait, - TPPROTO(struct pid *pid), - TPARGS(pid)); - -DECLARE_TRACE(sched_process_fork, - TPPROTO(struct task_struct *parent, struct task_struct *child), - TPARGS(parent, child)); - -DECLARE_TRACE(sched_signal_send, - TPPROTO(int sig, struct task_struct *p), - TPARGS(sig, p)); +#include #endif diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h new file mode 100644 index 000000000000..a4f662940f4e --- /dev/null +++ b/include/trace/sched_event_types.h @@ -0,0 +1,72 @@ + +/* use instead */ +#ifndef DEFINE_TRACE_FMT +# error Do not include this file directly. +# error Unless you know what you are doing. +#endif + +DEFINE_TRACE_FMT(sched_kthread_stop, + TPPROTO(struct task_struct *t), + TPARGS(t), + TPFMT("task %s:%d", t->comm, t->pid)); + +DEFINE_TRACE_FMT(sched_kthread_stop_ret, + TPPROTO(int ret), + TPARGS(ret), + TPFMT("ret=%d", ret)); + +DEFINE_TRACE_FMT(sched_wait_task, + TPPROTO(struct rq *rq, struct task_struct *p), + TPARGS(rq, p), + TPFMT("task %s:%d", p->comm, p->pid)); + +DEFINE_TRACE_FMT(sched_wakeup, + TPPROTO(struct rq *rq, struct task_struct *p, int success), + TPARGS(rq, p, success), + TPFMT("task %s:%d %s", + p->comm, p->pid, success?"succeeded":"failed")); + +DEFINE_TRACE_FMT(sched_wakeup_new, + TPPROTO(struct rq *rq, struct task_struct *p, int success), + TPARGS(rq, p, success), + TPFMT("task %s:%d", + p->comm, p->pid, success?"succeeded":"failed")); + +DEFINE_TRACE_FMT(sched_switch, + TPPROTO(struct rq *rq, struct task_struct *prev, + struct task_struct *next), + TPARGS(rq, prev, next), + TPFMT("task %s:%d ==> %s:%d", + prev->comm, prev->pid, next->comm, next->pid)); + +DEFINE_TRACE_FMT(sched_migrate_task, + TPPROTO(struct task_struct *p, int orig_cpu, int dest_cpu), + TPARGS(p, orig_cpu, dest_cpu), + TPFMT("task %s:%d from: %d to: %d", + p->comm, p->pid, orig_cpu, dest_cpu)); + +DEFINE_TRACE_FMT(sched_process_free, + TPPROTO(struct task_struct *p), + TPARGS(p), + TPFMT("task %s:%d", p->comm, p->pid)); + +DEFINE_TRACE_FMT(sched_process_exit, + TPPROTO(struct task_struct *p), + TPARGS(p), + TPFMT("task %s:%d", p->comm, p->pid)); + +DEFINE_TRACE_FMT(sched_process_wait, + TPPROTO(struct pid *pid), + TPARGS(pid), + TPFMT("pid %d", pid)); + +DEFINE_TRACE_FMT(sched_process_fork, + TPPROTO(struct task_struct *parent, struct task_struct *child), + TPARGS(parent, child), + TPFMT("parent %s:%d child %s:%d", + parent->comm, parent->pid, child->comm, child->pid)); + +DEFINE_TRACE_FMT(sched_signal_send, + TPPROTO(int sig, struct task_struct *p), + TPARGS(sig, p), + TPFMT("sig: %d task %s:%d", sig, p->comm, p->pid)); diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index c7363568b1cf..664b6c0dc75a 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -39,5 +39,6 @@ obj-$(CONFIG_KMEMTRACE) += kmemtrace.o obj-$(CONFIG_WORKQUEUE_TRACER) += trace_workqueue.o obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o obj-$(CONFIG_EVENT_TRACER) += trace_events.o +obj-$(CONFIG_EVENT_TRACER) += events.o libftrace-y := ftrace.o diff --git a/kernel/trace/events.c b/kernel/trace/events.c new file mode 100644 index 000000000000..38c89eef99ee --- /dev/null +++ b/kernel/trace/events.c @@ -0,0 +1,13 @@ +/* + * This is the place to register all trace points as events. + * Include the trace/.h at the top. + * Include the trace/_event_types.h at the bottom. + */ + +/* trace/.h here */ +#include + +#include "trace_events.h" + +/* trace/_event_types.h here */ +#include From 1473e4417c79f12d91ef91a469699bfa911f510f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 24 Feb 2009 14:15:08 -0500 Subject: [PATCH 182/413] tracing: make event directory structure This patch adds the directory /debug/tracing/events/ that will contain all the registered trace points. # ls /debug/tracing/events/ sched_kthread_stop sched_process_fork sched_switch sched_kthread_stop_ret sched_process_free sched_wait_task sched_migrate_task sched_process_wait sched_wakeup sched_process_exit sched_signal_send sched_wakeup_new # ls /debug/tracing/events/sched_switch/ enable # cat /debug/tracing/events/sched_switch/enable 1 # cat /debug/tracing/set_event sched_switch Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 137 ++++++++++++++++++++++++++++++++++-- kernel/trace/trace_events.h | 7 +- 2 files changed, 137 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 05bc80ec8d2c..3bcb9df93342 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -12,6 +12,11 @@ #include "trace_events.h" +#define events_for_each(event) \ + for (event = __start_ftrace_events; \ + (unsigned long)event < (unsigned long)__stop_ftrace_events; \ + event++) + void event_trace_printk(unsigned long ip, const char *fmt, ...) { va_list ap; @@ -39,15 +44,16 @@ static void ftrace_clear_events(void) static int ftrace_set_clr_event(char *buf, int set) { - struct ftrace_event_call *call = (void *)__start_ftrace_events; + struct ftrace_event_call *call = __start_ftrace_events; - while ((unsigned long)call < (unsigned long)__stop_ftrace_events) { + events_for_each(call) { - if (strcmp(buf, call->name) != 0) { - call++; + if (!call->name) + continue; + + if (strcmp(buf, call->name) != 0) continue; - } if (set) { /* Already set? */ @@ -223,6 +229,67 @@ ftrace_event_seq_open(struct inode *inode, struct file *file) return ret; } +static ssize_t +event_enable_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char *buf; + + if (call->enabled) + buf = "1\n"; + else + buf = "0\n"; + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2); +} + +static ssize_t +event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char buf[64]; + unsigned long val; + int ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + switch (val) { + case 0: + if (!call->enabled) + break; + + call->enabled = 0; + call->unregfunc(); + break; + case 1: + if (call->enabled) + break; + + call->enabled = 1; + call->regfunc(); + break; + + default: + return -EINVAL; + } + + *ppos += cnt; + + return cnt; +} + static const struct seq_operations show_event_seq_ops = { .start = t_start, .next = t_next, @@ -252,10 +319,59 @@ static const struct file_operations ftrace_set_event_fops = { .release = seq_release, }; +static const struct file_operations ftrace_enable_fops = { + .open = tracing_open_generic, + .read = event_enable_read, + .write = event_enable_write, +}; + +static struct dentry *event_trace_events_dir(void) +{ + static struct dentry *d_tracer; + static struct dentry *d_events; + + if (d_events) + return d_events; + + d_tracer = tracing_init_dentry(); + if (!d_tracer) + return NULL; + + d_events = debugfs_create_dir("events", d_tracer); + if (!d_events) + pr_warning("Could not create debugfs " + "'events' directory\n"); + + return d_events; +} + +static int +event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) +{ + struct dentry *entry; + + call->dir = debugfs_create_dir(call->name, d_events); + if (!call->dir) { + pr_warning("Could not create debugfs " + "'%s' directory\n", call->name); + return -1; + } + + entry = debugfs_create_file("enable", 0644, call->dir, call, + &ftrace_enable_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/enable' entry\n", call->name); + + return 0; +} + static __init int event_trace_init(void) { + struct ftrace_event_call *call = __start_ftrace_events; struct dentry *d_tracer; struct dentry *entry; + struct dentry *d_events; d_tracer = tracing_init_dentry(); if (!d_tracer) @@ -275,6 +391,17 @@ static __init int event_trace_init(void) pr_warning("Could not create debugfs " "'set_event' entry\n"); + d_events = event_trace_events_dir(); + if (!d_events) + return 0; + + events_for_each(call) { + /* The linker may leave blanks */ + if (!call->name) + continue; + event_create_dir(call, d_events); + } + return 0; } fs_initcall(event_trace_init); diff --git a/kernel/trace/trace_events.h b/kernel/trace/trace_events.h index 39342f86db27..cb8455b3ac9a 100644 --- a/kernel/trace/trace_events.h +++ b/kernel/trace/trace_events.h @@ -1,11 +1,13 @@ #ifndef _LINUX_KERNEL_TRACE_EVENTS_H #define _LINUX_KERNEL_TRACE_EVENTS_H +#include #include #include "trace.h" struct ftrace_event_call { char *name; + struct dentry *dir; int enabled; int (*regfunc)(void); void (*unregfunc)(void); @@ -39,6 +41,7 @@ static void ftrace_unreg_event_##call(void) \ } \ \ static struct ftrace_event_call __used \ +__attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ .name = #call, \ .regfunc = ftrace_reg_event_##call, \ @@ -46,7 +49,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ } void event_trace_printk(unsigned long ip, const char *fmt, ...); -extern unsigned long __start_ftrace_events[]; -extern unsigned long __stop_ftrace_events[]; +extern struct ftrace_event_call __start_ftrace_events[]; +extern struct ftrace_event_call __stop_ftrace_events[]; #endif /* _LINUX_KERNEL_TRACE_EVENTS_H */ From 2d542cf34264ac92e9e7ac55c0b096b066d569d2 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 25 Feb 2009 08:40:09 +0100 Subject: [PATCH 183/413] tracing/hw-branch-tracing: convert bts-tracer mutex to a spinlock Impact: fix CPU hotplug lockup bts_hotcpu_handler() is called with irqs disabled, so using mutex_lock() is a no-no. All the BTS codepaths here are atomic (they do not schedule), so using a spinlock is the right solution. Cc: Markus Metzger Signed-off-by: Ingo Molnar --- kernel/trace/trace_hw_branches.c | 57 ++++++++++++++++---------------- 1 file changed, 28 insertions(+), 29 deletions(-) diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c index 3335e807144b..7bfdf4c2347f 100644 --- a/kernel/trace/trace_hw_branches.c +++ b/kernel/trace/trace_hw_branches.c @@ -3,17 +3,15 @@ * * Copyright (C) 2008-2009 Intel Corporation. * Markus Metzger , 2008-2009 - * */ - -#include -#include +#include +#include #include #include -#include -#include +#include #include #include +#include #include @@ -23,16 +21,17 @@ #define SIZEOF_BTS (1 << 13) -/* The tracer mutex protects the below per-cpu tracer array. - It needs to be held to: - - start tracing on all cpus - - stop tracing on all cpus - - start tracing on a single hotplug cpu - - stop tracing on a single hotplug cpu - - read the trace from all cpus - - read the trace from a single cpu -*/ -static DEFINE_MUTEX(bts_tracer_mutex); +/* + * The tracer lock protects the below per-cpu tracer array. + * It needs to be held to: + * - start tracing on all cpus + * - stop tracing on all cpus + * - start tracing on a single hotplug cpu + * - stop tracing on a single hotplug cpu + * - read the trace from all cpus + * - read the trace from a single cpu + */ +static DEFINE_SPINLOCK(bts_tracer_lock); static DEFINE_PER_CPU(struct bts_tracer *, tracer); static DEFINE_PER_CPU(unsigned char[SIZEOF_BTS], buffer); @@ -47,7 +46,7 @@ static struct trace_array *hw_branch_trace __read_mostly; * Start tracing on the current cpu. * The argument is ignored. * - * pre: bts_tracer_mutex must be locked. + * pre: bts_tracer_lock must be locked. */ static void bts_trace_start_cpu(void *arg) { @@ -66,19 +65,19 @@ static void bts_trace_start_cpu(void *arg) static void bts_trace_start(struct trace_array *tr) { - mutex_lock(&bts_tracer_mutex); + spin_lock(&bts_tracer_lock); on_each_cpu(bts_trace_start_cpu, NULL, 1); trace_hw_branches_enabled = 1; - mutex_unlock(&bts_tracer_mutex); + spin_unlock(&bts_tracer_lock); } /* * Stop tracing on the current cpu. * The argument is ignored. * - * pre: bts_tracer_mutex must be locked. + * pre: bts_tracer_lock must be locked. */ static void bts_trace_stop_cpu(void *arg) { @@ -90,12 +89,12 @@ static void bts_trace_stop_cpu(void *arg) static void bts_trace_stop(struct trace_array *tr) { - mutex_lock(&bts_tracer_mutex); + spin_lock(&bts_tracer_lock); trace_hw_branches_enabled = 0; on_each_cpu(bts_trace_stop_cpu, NULL, 1); - mutex_unlock(&bts_tracer_mutex); + spin_unlock(&bts_tracer_lock); } static int __cpuinit bts_hotcpu_handler(struct notifier_block *nfb, @@ -103,7 +102,7 @@ static int __cpuinit bts_hotcpu_handler(struct notifier_block *nfb, { unsigned int cpu = (unsigned long)hcpu; - mutex_lock(&bts_tracer_mutex); + spin_lock(&bts_tracer_lock); if (!trace_hw_branches_enabled) goto out; @@ -119,7 +118,7 @@ static int __cpuinit bts_hotcpu_handler(struct notifier_block *nfb, } out: - mutex_unlock(&bts_tracer_mutex); + spin_unlock(&bts_tracer_lock); return NOTIFY_DONE; } @@ -225,7 +224,7 @@ static void trace_bts_at(const struct bts_trace *trace, void *at) /* * Collect the trace on the current cpu and write it into the ftrace buffer. * - * pre: bts_tracer_mutex must be locked + * pre: bts_tracer_lock must be locked */ static void trace_bts_cpu(void *arg) { @@ -261,11 +260,11 @@ static void trace_bts_cpu(void *arg) static void trace_bts_prepare(struct trace_iterator *iter) { - mutex_lock(&bts_tracer_mutex); + spin_lock(&bts_tracer_lock); on_each_cpu(trace_bts_cpu, iter->tr, 1); - mutex_unlock(&bts_tracer_mutex); + spin_unlock(&bts_tracer_lock); } static void trace_bts_close(struct trace_iterator *iter) @@ -275,11 +274,11 @@ static void trace_bts_close(struct trace_iterator *iter) void trace_hw_branch_oops(void) { - mutex_lock(&bts_tracer_mutex); + spin_lock(&bts_tracer_lock); trace_bts_cpu(hw_branch_trace); - mutex_unlock(&bts_tracer_mutex); + spin_unlock(&bts_tracer_lock); } struct tracer bts_tracer __read_mostly = From 886b5b73d71e4027d7dc6c14f5f7ab102201ea6b Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 25 Feb 2009 11:03:44 +0100 Subject: [PATCH 184/413] tracing: remove /debug/tracing/latency_trace Impact: remove old debug/tracing API /debug/tracing/latency_trace is an old legacy format we kept from the old latency tracer. Remove the file for now. If there's any useful bit missing then we'll propagate any useful output bits into the /debug/tracing/trace output. Reported-by: Andrew Morton Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 5 ----- 1 file changed, 5 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e1f3b99a2e52..11ba100f9a9e 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2938,11 +2938,6 @@ static __init int tracer_init_debugfs(void) if (!entry) pr_warning("Could not create debugfs 'tracing_cpumask' entry\n"); - entry = debugfs_create_file("latency_trace", 0444, d_tracer, - &global_trace, &tracing_lt_fops); - if (!entry) - pr_warning("Could not create debugfs 'latency_trace' entry\n"); - entry = debugfs_create_file("trace", 0444, d_tracer, &global_trace, &tracing_fops); if (!entry) From b04cc6b1f6398b0e0b60d37e27ce51b4899672ec Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 25 Feb 2009 03:22:28 +0100 Subject: [PATCH 185/413] tracing/core: introduce per cpu tracing files Impact: split up tracing output per cpu Currently, on the tracing debugfs directory, three files are available to the user to let him extracting the trace output: - trace is an iterator through the ring-buffer. It's a reader but not a consumer It doesn't block when no more traces are available. - trace pretty similar to the former, except that it adds more informations such as prempt count, irq flag, ... - trace_pipe is a reader and a consumer, it will also block waiting for traces if necessary (heh, yes it's a pipe). The traces coming from different cpus are curretly mixed up inside these files. Sometimes it messes up the informations, sometimes it's useful, depending on what does the tracer capture. The tracing_cpumask file is useful to filter the output and select only the traces captured a custom defined set of cpus. But still it is not enough powerful to extract at the same time one trace buffer per cpu. So this patch creates a new directory: /debug/tracing/per_cpu/. Inside this directory, you will now find one trace_pipe file and one trace file per cpu. Which means if you have two cpus, you will have: trace0 trace1 trace_pipe0 trace_pipe1 And of course, reading these files will have the same effect than with the usual tracing files, except that you will only see the traces from the given cpu. The original all-in-one cpu trace file are still available on their original place. Until now, only one consumer was allowed on trace_pipe to avoid racy consuming on the ring-buffer. Now the approach changed a bit, you can have only one consumer per cpu. Which means you are allowed to read concurrently trace_pipe0 and trace_pipe1 But you can't have two readers on trace_pipe0 or trace_pipe1. Following the same logic, if there is one reader on the common trace_pipe, you can not have at the same time another reader on trace_pipe0 or in trace_pipe1. Because in trace_pipe is already a consumer in all cpu buffers in essence. Signed-off-by: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 168 ++++++++++++++++++++++++++++++++++++------- kernel/trace/trace.h | 3 + 2 files changed, 147 insertions(+), 24 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 11ba100f9a9e..aa58b7bc847c 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -98,6 +98,9 @@ static inline void ftrace_enable_cpu(void) static cpumask_var_t __read_mostly tracing_buffer_mask; +/* Define which cpu buffers are currently read in trace_pipe */ +static cpumask_var_t tracing_reader_cpumask; + #define for_each_tracing_cpu(cpu) \ for_each_cpu(cpu, tracing_buffer_mask) @@ -1195,10 +1198,25 @@ __find_next_entry(struct trace_iterator *iter, int *ent_cpu, u64 *ent_ts) { struct ring_buffer *buffer = iter->tr->buffer; struct trace_entry *ent, *next = NULL; + int cpu_file = iter->cpu_file; u64 next_ts = 0, ts; int next_cpu = -1; int cpu; + /* + * If we are in a per_cpu trace file, don't bother by iterating over + * all cpu and peek directly. + */ + if (cpu_file > TRACE_PIPE_ALL_CPU) { + if (ring_buffer_empty_cpu(buffer, cpu_file)) + return NULL; + ent = peek_next_entry(iter, cpu_file, ent_ts); + if (ent_cpu) + *ent_cpu = cpu_file; + + return ent; + } + for_each_tracing_cpu(cpu) { if (ring_buffer_empty_cpu(buffer, cpu)) @@ -1279,6 +1297,7 @@ static void *s_next(struct seq_file *m, void *v, loff_t *pos) static void *s_start(struct seq_file *m, loff_t *pos) { struct trace_iterator *iter = m->private; + int cpu_file = iter->cpu_file; void *p = NULL; loff_t l = 0; int cpu; @@ -1299,9 +1318,12 @@ static void *s_start(struct seq_file *m, loff_t *pos) ftrace_disable_cpu(); - for_each_tracing_cpu(cpu) { - ring_buffer_iter_reset(iter->buffer_iter[cpu]); - } + if (cpu_file == TRACE_PIPE_ALL_CPU) { + for_each_tracing_cpu(cpu) + ring_buffer_iter_reset(iter->buffer_iter[cpu]); + } else + ring_buffer_iter_reset(iter->buffer_iter[cpu_file]); + ftrace_enable_cpu(); @@ -1653,6 +1675,7 @@ static struct seq_operations tracer_seq_ops = { static struct trace_iterator * __tracing_open(struct inode *inode, struct file *file, int *ret) { + long cpu_file = (long) inode->i_private; struct trace_iterator *iter; struct seq_file *m; int cpu; @@ -1672,9 +1695,10 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) if (current_trace && current_trace->print_max) iter->tr = &max_tr; else - iter->tr = inode->i_private; + iter->tr = &global_trace; iter->trace = current_trace; iter->pos = -1; + iter->cpu_file = cpu_file; /* Notify the tracer early; before we stop tracing. */ if (iter->trace && iter->trace->open) @@ -1684,14 +1708,22 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) if (ring_buffer_overruns(iter->tr->buffer)) iter->iter_flags |= TRACE_FILE_ANNOTATE; + if (iter->cpu_file == TRACE_PIPE_ALL_CPU) { + for_each_tracing_cpu(cpu) { - for_each_tracing_cpu(cpu) { + iter->buffer_iter[cpu] = + ring_buffer_read_start(iter->tr->buffer, cpu); + if (!iter->buffer_iter[cpu]) + goto fail_buffer; + } + } else { + cpu = iter->cpu_file; iter->buffer_iter[cpu] = - ring_buffer_read_start(iter->tr->buffer, cpu); + ring_buffer_read_start(iter->tr->buffer, cpu); if (!iter->buffer_iter[cpu]) - goto fail_buffer; + goto fail; } /* TODO stop tracer */ @@ -1715,6 +1747,7 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) if (iter->buffer_iter[cpu]) ring_buffer_read_finish(iter->buffer_iter[cpu]); } +fail: mutex_unlock(&trace_types_lock); kfree(iter); @@ -2325,54 +2358,77 @@ tracing_max_lat_write(struct file *filp, const char __user *ubuf, return cnt; } -static atomic_t tracing_reader; - static int tracing_open_pipe(struct inode *inode, struct file *filp) { + long cpu_file = (long) inode->i_private; struct trace_iterator *iter; + int ret = 0; if (tracing_disabled) return -ENODEV; - /* We only allow for reader of the pipe */ - if (atomic_inc_return(&tracing_reader) != 1) { - atomic_dec(&tracing_reader); - return -EBUSY; + mutex_lock(&trace_types_lock); + + /* We only allow one reader per cpu */ + if (cpu_file == TRACE_PIPE_ALL_CPU) { + if (!cpumask_empty(tracing_reader_cpumask)) { + ret = -EBUSY; + goto out; + } + cpumask_setall(tracing_reader_cpumask); + } else { + if (!cpumask_test_cpu(cpu_file, tracing_reader_cpumask)) + cpumask_set_cpu(cpu_file, tracing_reader_cpumask); + else { + ret = -EBUSY; + goto out; + } } /* create a buffer to store the information to pass to userspace */ iter = kzalloc(sizeof(*iter), GFP_KERNEL); - if (!iter) - return -ENOMEM; + if (!iter) { + ret = -ENOMEM; + goto out; + } if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) { kfree(iter); - return -ENOMEM; + ret = -ENOMEM; + goto out; } - mutex_lock(&trace_types_lock); - /* trace pipe does not show start of buffer */ cpumask_setall(iter->started); + iter->cpu_file = cpu_file; iter->tr = &global_trace; iter->trace = current_trace; filp->private_data = iter; if (iter->trace->pipe_open) iter->trace->pipe_open(iter); - mutex_unlock(&trace_types_lock); - return 0; +out: + mutex_unlock(&trace_types_lock); + return ret; } static int tracing_release_pipe(struct inode *inode, struct file *file) { struct trace_iterator *iter = file->private_data; + mutex_lock(&trace_types_lock); + + if (iter->cpu_file == TRACE_PIPE_ALL_CPU) + cpumask_clear(tracing_reader_cpumask); + else + cpumask_clear_cpu(iter->cpu_file, tracing_reader_cpumask); + + mutex_unlock(&trace_types_lock); + free_cpumask_var(iter->started); kfree(iter); - atomic_dec(&tracing_reader); return 0; } @@ -2911,6 +2967,59 @@ struct dentry *tracing_init_dentry(void) return d_tracer; } +static struct dentry *d_percpu; + +struct dentry *tracing_dentry_percpu(void) +{ + static int once; + struct dentry *d_tracer; + + if (d_percpu) + return d_percpu; + + d_tracer = tracing_init_dentry(); + + if (!d_tracer) + return NULL; + + d_percpu = debugfs_create_dir("per_cpu", d_tracer); + + if (!d_percpu && !once) { + once = 1; + pr_warning("Could not create debugfs directory 'per_cpu'\n"); + return NULL; + } + + return d_percpu; +} + +static void tracing_init_debugfs_percpu(long cpu) +{ + struct dentry *d_percpu = tracing_dentry_percpu(); + struct dentry *entry; + /* strlen(trace_pipe) + MAX(log10(cpu)) + '\0' */ + char filename[17]; + + if (cpu > 999 || cpu < 0) + return; + + /* per cpu trace_pipe */ + sprintf(filename, "trace_pipe%ld", cpu); + + entry = debugfs_create_file(filename, 0444, d_percpu, + (void *) cpu, &tracing_pipe_fops); + if (!entry) + pr_warning("Could not create debugfs '%s' entry\n", filename); + + /* per cpu trace */ + sprintf(filename, "trace%ld", cpu); + + entry = debugfs_create_file(filename, 0444, d_percpu, + (void *) cpu, &tracing_fops); + if (!entry) + pr_warning("Could not create debugfs '%s' entry\n", filename); +} + #ifdef CONFIG_FTRACE_SELFTEST /* Let selftest have access to static functions in this file */ #include "trace_selftest.c" @@ -2920,6 +3029,7 @@ static __init int tracer_init_debugfs(void) { struct dentry *d_tracer; struct dentry *entry; + int cpu; d_tracer = tracing_init_dentry(); @@ -2939,7 +3049,7 @@ static __init int tracer_init_debugfs(void) pr_warning("Could not create debugfs 'tracing_cpumask' entry\n"); entry = debugfs_create_file("trace", 0444, d_tracer, - &global_trace, &tracing_fops); + (void *) TRACE_PIPE_ALL_CPU, &tracing_fops); if (!entry) pr_warning("Could not create debugfs 'trace' entry\n"); @@ -2970,8 +3080,8 @@ static __init int tracer_init_debugfs(void) if (!entry) pr_warning("Could not create debugfs 'README' entry\n"); - entry = debugfs_create_file("trace_pipe", 0644, d_tracer, - NULL, &tracing_pipe_fops); + entry = debugfs_create_file("trace_pipe", 0444, d_tracer, + (void *) TRACE_PIPE_ALL_CPU, &tracing_pipe_fops); if (!entry) pr_warning("Could not create debugfs " "'trace_pipe' entry\n"); @@ -2999,6 +3109,10 @@ static __init int tracer_init_debugfs(void) #ifdef CONFIG_SYSPROF_TRACER init_tracer_sysprof_debugfs(d_tracer); #endif + + for_each_tracing_cpu(cpu) + tracing_init_debugfs_percpu(cpu); + return 0; } @@ -3222,8 +3336,12 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_cpumask, GFP_KERNEL)) goto out_free_buffer_mask; + if (!alloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL)) + goto out_free_tracing_cpumask; + cpumask_copy(tracing_buffer_mask, cpu_possible_mask); cpumask_copy(tracing_cpumask, cpu_all_mask); + cpumask_clear(tracing_reader_cpumask); /* TODO: make the number of buffers hot pluggable with CPUS */ global_trace.buffer = ring_buffer_alloc(trace_buf_size, @@ -3272,6 +3390,8 @@ __init static int tracer_alloc_buffers(void) ret = 0; out_free_cpumask: + free_cpumask_var(tracing_reader_cpumask); +out_free_tracing_cpumask: free_cpumask_var(tracing_cpumask); out_free_buffer_mask: free_cpumask_var(tracing_buffer_mask); diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index eed732c151fc..508235a39da6 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -395,6 +395,8 @@ struct trace_seq { unsigned int readpos; }; +#define TRACE_PIPE_ALL_CPU -1 + /* * Trace iterator - used by printout routines who present trace * results to users and which routines might sleep, etc: @@ -404,6 +406,7 @@ struct trace_iterator { struct tracer *trace; void *private; struct ring_buffer_iter *buffer_iter[NR_CPUS]; + int cpu_file; /* The below is zeroed out in pipe_read */ struct trace_seq seq; From d7350c3f45694104e820041969c8185c5f99e57c Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 25 Feb 2009 06:13:16 +0100 Subject: [PATCH 186/413] tracing/core: make the read callbacks reentrants Now that several per-cpu files can be read or spliced at the same, we want the read/splice callbacks for tracing files to be reentrants. Until now, a single global mutex (trace_types_lock) serialized the access to tracing_read_pipe(), tracing_splice_read_pipe(), and the seq helpers. Ie: it means that if a user tries to read trace_pipe0 and trace_pipe1 at the same time, the access to the function tracing_read_pipe() is contended and one reader must wait for the other to finish its read call. The trace_type_lock mutex is mostly here to serialize the access to the global current tracer (current_trace), which can be changed concurrently. Although the iter struct keeps a private pointer to this tracer, its callbacks can be changed by another function. The method used here is to not keep anymore private reference to the tracer inside the iterator but to make a copy of it inside the iterator. Then it checks on subsequents read calls if the tracer has changed. This is not costly because the current tracer is not expected to be changed often, so we use a branch prediction for that. Moreover, we add a private mutex to the iterator (there is one iterator per file descriptor) to serialize the accesses in case of multiple consumers per file descriptor (which would be a silly idea from the user). Note that this is not to protect the ring buffer, since the ring buffer already serializes the readers accesses. This is to prevent from traces weirdness in case of concurrent consumers. But these mutexes can be dropped anyway, that would not result in any crash. Just tell me what you think about it. Signed-off-by: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 103 +++++++++++++++++++++++++++++++++++-------- kernel/trace/trace.h | 3 +- 2 files changed, 86 insertions(+), 20 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index aa58b7bc847c..d8d899f3bd6f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1294,20 +1294,32 @@ static void *s_next(struct seq_file *m, void *v, loff_t *pos) return ent; } +/* + * No necessary locking here. The worst thing which can + * happen is loosing events consumed at the same time + * by a trace_pipe reader. + * Other than that, we don't risk to crash the ring buffer + * because it serializes the readers. + * + * The current tracer is copied to avoid a global locking + * all around. + */ static void *s_start(struct seq_file *m, loff_t *pos) { struct trace_iterator *iter = m->private; + static struct tracer *old_tracer; int cpu_file = iter->cpu_file; void *p = NULL; loff_t l = 0; int cpu; + /* copy the tracer to avoid using a global lock all around */ mutex_lock(&trace_types_lock); - - if (!current_trace || current_trace != iter->trace) { - mutex_unlock(&trace_types_lock); - return NULL; + if (unlikely(old_tracer != current_trace && current_trace)) { + old_tracer = current_trace; + *iter->trace = *current_trace; } + mutex_unlock(&trace_types_lock); atomic_inc(&trace_record_cmdline_disabled); @@ -1341,7 +1353,6 @@ static void *s_start(struct seq_file *m, loff_t *pos) static void s_stop(struct seq_file *m, void *p) { atomic_dec(&trace_record_cmdline_disabled); - mutex_unlock(&trace_types_lock); } static void print_lat_help_header(struct seq_file *m) @@ -1691,13 +1702,25 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) goto out; } + /* + * We make a copy of the current tracer to avoid concurrent + * changes on it while we are reading. + */ mutex_lock(&trace_types_lock); + iter->trace = kzalloc(sizeof(*iter->trace), GFP_KERNEL); + if (!iter->trace) { + *ret = -ENOMEM; + goto fail; + } + if (current_trace) + *iter->trace = *current_trace; + if (current_trace && current_trace->print_max) iter->tr = &max_tr; else iter->tr = &global_trace; - iter->trace = current_trace; iter->pos = -1; + mutex_init(&iter->mutex); iter->cpu_file = cpu_file; /* Notify the tracer early; before we stop tracing. */ @@ -1747,8 +1770,9 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) if (iter->buffer_iter[cpu]) ring_buffer_read_finish(iter->buffer_iter[cpu]); } -fail: + fail: mutex_unlock(&trace_types_lock); + kfree(iter->trace); kfree(iter); return ERR_PTR(-ENOMEM); @@ -1783,6 +1807,8 @@ static int tracing_release(struct inode *inode, struct file *file) mutex_unlock(&trace_types_lock); seq_release(inode, file); + mutex_destroy(&iter->mutex); + kfree(iter->trace); kfree(iter); return 0; } @@ -2392,10 +2418,21 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) goto out; } - if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) { - kfree(iter); + /* + * We make a copy of the current tracer to avoid concurrent + * changes on it while we are reading. + */ + iter->trace = kmalloc(sizeof(*iter->trace), GFP_KERNEL); + if (!iter->trace) { ret = -ENOMEM; - goto out; + goto fail; + } + if (current_trace) + *iter->trace = *current_trace; + + if (!alloc_cpumask_var(&iter->started, GFP_KERNEL)) { + ret = -ENOMEM; + goto fail; } /* trace pipe does not show start of buffer */ @@ -2403,7 +2440,7 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) iter->cpu_file = cpu_file; iter->tr = &global_trace; - iter->trace = current_trace; + mutex_init(&iter->mutex); filp->private_data = iter; if (iter->trace->pipe_open) @@ -2412,6 +2449,12 @@ static int tracing_open_pipe(struct inode *inode, struct file *filp) out: mutex_unlock(&trace_types_lock); return ret; + +fail: + kfree(iter->trace); + kfree(iter); + mutex_unlock(&trace_types_lock); + return ret; } static int tracing_release_pipe(struct inode *inode, struct file *file) @@ -2428,6 +2471,8 @@ static int tracing_release_pipe(struct inode *inode, struct file *file) mutex_unlock(&trace_types_lock); free_cpumask_var(iter->started); + mutex_destroy(&iter->mutex); + kfree(iter->trace); kfree(iter); return 0; @@ -2497,18 +2542,15 @@ static int tracing_wait_pipe(struct file *filp) return -EAGAIN; } - mutex_unlock(&trace_types_lock); + mutex_unlock(&iter->mutex); iter->trace->wait_pipe(iter); - mutex_lock(&trace_types_lock); + mutex_lock(&iter->mutex); if (signal_pending(current)) return -EINTR; - if (iter->trace != current_trace) - return 0; - /* * We block until we read something and tracing is disabled. * We still block if tracing is disabled, but we have never @@ -2533,6 +2575,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct trace_iterator *iter = filp->private_data; + static struct tracer *old_tracer; ssize_t sret; /* return any leftover data */ @@ -2542,7 +2585,20 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, trace_seq_reset(&iter->seq); + /* copy the tracer to avoid using a global lock all around */ mutex_lock(&trace_types_lock); + if (unlikely(old_tracer != current_trace && current_trace)) { + old_tracer = current_trace; + *iter->trace = *current_trace; + } + mutex_unlock(&trace_types_lock); + + /* + * Avoid more than one consumer on a single file descriptor + * This is just a matter of traces coherency, the ring buffer itself + * is protected. + */ + mutex_lock(&iter->mutex); if (iter->trace->read) { sret = iter->trace->read(iter, filp, ubuf, cnt, ppos); if (sret) @@ -2599,7 +2655,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, goto waitagain; out: - mutex_unlock(&trace_types_lock); + mutex_unlock(&iter->mutex); return sret; } @@ -2676,11 +2732,20 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, .ops = &tracing_pipe_buf_ops, .spd_release = tracing_spd_release_pipe, }; + static struct tracer *old_tracer; ssize_t ret; size_t rem; unsigned int i; + /* copy the tracer to avoid using a global lock all around */ mutex_lock(&trace_types_lock); + if (unlikely(old_tracer != current_trace && current_trace)) { + old_tracer = current_trace; + *iter->trace = *current_trace; + } + mutex_unlock(&trace_types_lock); + + mutex_lock(&iter->mutex); if (iter->trace->splice_read) { ret = iter->trace->splice_read(iter, filp, @@ -2720,14 +2785,14 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, trace_seq_reset(&iter->seq); } - mutex_unlock(&trace_types_lock); + mutex_unlock(&iter->mutex); spd.nr_pages = i; return splice_to_pipe(pipe, &spd); out_err: - mutex_unlock(&trace_types_lock); + mutex_unlock(&iter->mutex); return ret; } diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 508235a39da6..632191770aac 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -405,8 +405,9 @@ struct trace_iterator { struct trace_array *tr; struct tracer *trace; void *private; - struct ring_buffer_iter *buffer_iter[NR_CPUS]; int cpu_file; + struct mutex mutex; + struct ring_buffer_iter *buffer_iter[NR_CPUS]; /* The below is zeroed out in pipe_read */ struct trace_seq seq; From eef62a6826b8ab530cefff5aa55c1661a209c803 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 25 Feb 2009 15:49:52 -0500 Subject: [PATCH 187/413] tracing: rename DEFINE_TRACE_FMT to just TRACE_FORMAT There's been a bit confusion to whether DEFINE/DECLARE_TRACE_FMT should be a DEFINE or a DECLARE. Ingo Molnar suggested simply calling it TRACE_FORMAT. Reported-by: Ingo Molnar Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 2 +- include/trace/sched_event_types.h | 26 +++++++++++++------------- kernel/trace/trace_events.h | 4 ++-- 3 files changed, 16 insertions(+), 16 deletions(-) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 34ae464effff..3de09fa8e01d 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -153,7 +153,7 @@ static inline void tracepoint_synchronize_unregister(void) synchronize_sched(); } -#define DEFINE_TRACE_FMT(name, proto, args, fmt) \ +#define TRACE_FORMAT(name, proto, args, fmt) \ DECLARE_TRACE(name, TPPROTO(proto), TPARGS(args)) #endif diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index a4f662940f4e..a3d3d66a51c8 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -1,72 +1,72 @@ /* use instead */ -#ifndef DEFINE_TRACE_FMT +#ifndef TRACE_FORMAT # error Do not include this file directly. # error Unless you know what you are doing. #endif -DEFINE_TRACE_FMT(sched_kthread_stop, +TRACE_FORMAT(sched_kthread_stop, TPPROTO(struct task_struct *t), TPARGS(t), TPFMT("task %s:%d", t->comm, t->pid)); -DEFINE_TRACE_FMT(sched_kthread_stop_ret, +TRACE_FORMAT(sched_kthread_stop_ret, TPPROTO(int ret), TPARGS(ret), TPFMT("ret=%d", ret)); -DEFINE_TRACE_FMT(sched_wait_task, +TRACE_FORMAT(sched_wait_task, TPPROTO(struct rq *rq, struct task_struct *p), TPARGS(rq, p), TPFMT("task %s:%d", p->comm, p->pid)); -DEFINE_TRACE_FMT(sched_wakeup, +TRACE_FORMAT(sched_wakeup, TPPROTO(struct rq *rq, struct task_struct *p, int success), TPARGS(rq, p, success), TPFMT("task %s:%d %s", p->comm, p->pid, success?"succeeded":"failed")); -DEFINE_TRACE_FMT(sched_wakeup_new, +TRACE_FORMAT(sched_wakeup_new, TPPROTO(struct rq *rq, struct task_struct *p, int success), TPARGS(rq, p, success), TPFMT("task %s:%d", p->comm, p->pid, success?"succeeded":"failed")); -DEFINE_TRACE_FMT(sched_switch, +TRACE_FORMAT(sched_switch, TPPROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next), TPARGS(rq, prev, next), TPFMT("task %s:%d ==> %s:%d", prev->comm, prev->pid, next->comm, next->pid)); -DEFINE_TRACE_FMT(sched_migrate_task, +TRACE_FORMAT(sched_migrate_task, TPPROTO(struct task_struct *p, int orig_cpu, int dest_cpu), TPARGS(p, orig_cpu, dest_cpu), TPFMT("task %s:%d from: %d to: %d", p->comm, p->pid, orig_cpu, dest_cpu)); -DEFINE_TRACE_FMT(sched_process_free, +TRACE_FORMAT(sched_process_free, TPPROTO(struct task_struct *p), TPARGS(p), TPFMT("task %s:%d", p->comm, p->pid)); -DEFINE_TRACE_FMT(sched_process_exit, +TRACE_FORMAT(sched_process_exit, TPPROTO(struct task_struct *p), TPARGS(p), TPFMT("task %s:%d", p->comm, p->pid)); -DEFINE_TRACE_FMT(sched_process_wait, +TRACE_FORMAT(sched_process_wait, TPPROTO(struct pid *pid), TPARGS(pid), TPFMT("pid %d", pid)); -DEFINE_TRACE_FMT(sched_process_fork, +TRACE_FORMAT(sched_process_fork, TPPROTO(struct task_struct *parent, struct task_struct *child), TPARGS(parent, child), TPFMT("parent %s:%d child %s:%d", parent->comm, parent->pid, child->comm, child->pid)); -DEFINE_TRACE_FMT(sched_signal_send, +TRACE_FORMAT(sched_signal_send, TPPROTO(int sig, struct task_struct *p), TPARGS(sig, p), TPFMT("sig: %d task %s:%d", sig, p->comm, p->pid)); diff --git a/kernel/trace/trace_events.h b/kernel/trace/trace_events.h index cb8455b3ac9a..deb95e5006c8 100644 --- a/kernel/trace/trace_events.h +++ b/kernel/trace/trace_events.h @@ -17,8 +17,8 @@ struct ftrace_event_call { #undef TPFMT #define TPFMT(fmt, args...) fmt "\n", ##args -#undef DEFINE_TRACE_FMT -#define DEFINE_TRACE_FMT(call, proto, args, fmt) \ +#undef TRACE_FORMAT +#define TRACE_FORMAT(call, proto, args, fmt) \ static void ftrace_event_##call(proto) \ { \ event_trace_printk(_RET_IP_, "(" #call ") " fmt); \ From 3cdfdf91fcc77cfc82592e2b5c2ab35abe819c41 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 25 Feb 2009 15:54:30 -0500 Subject: [PATCH 188/413] tracing: wrap arguments with PARAMS Peter Zijlstra warned that TPPROTO and TPARGS might become something other than a simple copy of itself. To prevent this from having side effects in the TRACE_FORMAT macro in tracepoint.h, we add a PARAMS() macro to be defined as just a wrapper. Reported-by: Peter Zijlstra Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 3de09fa8e01d..62d13391a240 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -153,7 +153,8 @@ static inline void tracepoint_synchronize_unregister(void) synchronize_sched(); } +#define PARAMS(args...) args #define TRACE_FORMAT(name, proto, args, fmt) \ - DECLARE_TRACE(name, TPPROTO(proto), TPARGS(args)) + DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) #endif From 8656e7a2fa6afcd8682990f804a2a9674568738f Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 26 Feb 2009 00:41:38 +0100 Subject: [PATCH 189/413] tracing/core: make the per cpu trace files in per cpu directories Impact: restructure the VFS layout of per CPU trace buffers The per cpu trace files are all in a single directory: /debug/tracing/per_cpu. In case of a large number of cpu, the content of this directory becomes messy so we create now one directory per cpu inside /debug/tracing/per_cpu which contain each their own trace_pipe and trace files. Ie: /debug/tracing$ ls -R per_cpu per_cpu: cpu0 cpu1 per_cpu/cpu0: trace trace_pipe per_cpu/cpu1: trace trace_pipe Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Lai Jiangshan Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 25 ++++++++++++++----------- 1 file changed, 14 insertions(+), 11 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index d8d899f3bd6f..bdaf60d3d337 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3061,28 +3061,31 @@ struct dentry *tracing_dentry_percpu(void) static void tracing_init_debugfs_percpu(long cpu) { struct dentry *d_percpu = tracing_dentry_percpu(); - struct dentry *entry; - /* strlen(trace_pipe) + MAX(log10(cpu)) + '\0' */ - char filename[17]; + struct dentry *entry, *d_cpu; + /* strlen(cpu) + MAX(log10(cpu)) + '\0' */ + char cpu_dir[7]; if (cpu > 999 || cpu < 0) return; - /* per cpu trace_pipe */ - sprintf(filename, "trace_pipe%ld", cpu); + sprintf(cpu_dir, "cpu%ld", cpu); + d_cpu = debugfs_create_dir(cpu_dir, d_percpu); + if (!d_cpu) { + pr_warning("Could not create debugfs '%s' entry\n", cpu_dir); + return; + } - entry = debugfs_create_file(filename, 0444, d_percpu, + /* per cpu trace_pipe */ + entry = debugfs_create_file("trace_pipe", 0444, d_cpu, (void *) cpu, &tracing_pipe_fops); if (!entry) - pr_warning("Could not create debugfs '%s' entry\n", filename); + pr_warning("Could not create debugfs 'trace_pipe' entry\n"); /* per cpu trace */ - sprintf(filename, "trace%ld", cpu); - - entry = debugfs_create_file(filename, 0444, d_percpu, + entry = debugfs_create_file("trace", 0444, d_cpu, (void *) cpu, &tracing_fops); if (!entry) - pr_warning("Could not create debugfs '%s' entry\n", filename); + pr_warning("Could not create debugfs 'trace' entry\n"); } #ifdef CONFIG_FTRACE_SELFTEST From af39241b90a345556b8884adff87096afe71b050 Mon Sep 17 00:00:00 2001 From: Jason Baron Date: Thu, 26 Feb 2009 10:11:05 -0500 Subject: [PATCH 190/413] tracing, genirq: add irq enter and exit trace events Impact: add new tracepoints Add them to the generic IRQ code, that way every architecture gets these new tracepoints, not just x86. Using Steve's new 'TRACE_FORMAT', I can get function graph trace as follows using the original two IRQ tracepoints: 3) | handle_IRQ_event() { 3) | /* (irq_handler_entry) irq=28 handler=eth0 */ 3) | e1000_intr_msi() { 3) 2.460 us | __napi_schedule(); 3) 9.416 us | } 3) | /* (irq_handler_exit) irq=28 handler=eth0 return=handled */ 3) + 22.935 us | } Signed-off-by: Jason Baron Signed-off-by: Steven Rostedt Acked-by: Peter Zijlstra Acked-by: Masami Hiramatsu Cc: KOSAKI Motohiro Cc: Mathieu Desnoyers Cc: "Frank Ch. Eigler" Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- include/trace/irq.h | 9 +++++++++ include/trace/irq_event_types.h | 17 +++++++++++++++++ kernel/irq/handle.c | 6 ++++++ kernel/trace/events.c | 2 ++ 4 files changed, 34 insertions(+) create mode 100644 include/trace/irq.h create mode 100644 include/trace/irq_event_types.h diff --git a/include/trace/irq.h b/include/trace/irq.h new file mode 100644 index 000000000000..ff5d4495dc37 --- /dev/null +++ b/include/trace/irq.h @@ -0,0 +1,9 @@ +#ifndef _TRACE_IRQ_H +#define _TRACE_IRQ_H + +#include +#include + +#include + +#endif diff --git a/include/trace/irq_event_types.h b/include/trace/irq_event_types.h new file mode 100644 index 000000000000..5d0919fdd2d4 --- /dev/null +++ b/include/trace/irq_event_types.h @@ -0,0 +1,17 @@ + +/* use instead */ +#ifndef TRACE_FORMAT +# error Do not include this file directly. +# error Unless you know what you are doing. +#endif + +TRACE_FORMAT(irq_handler_entry, + TPPROTO(int irq, struct irqaction *action), + TPARGS(irq, action), + TPFMT("irq=%d handler=%s", irq, action->name)); + +TRACE_FORMAT(irq_handler_exit, + TPPROTO(int irq, struct irqaction *action, int ret), + TPARGS(irq, action, ret), + TPFMT("irq=%d handler=%s return=%s", + irq, action->name, ret ? "handled" : "unhandled")); diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 3aba8d12f328..4709a7c870d7 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -17,6 +17,7 @@ #include #include #include +#include #include "internals.h" @@ -316,6 +317,9 @@ irqreturn_t no_action(int cpl, void *dev_id) return IRQ_NONE; } +DEFINE_TRACE(irq_handler_entry); +DEFINE_TRACE(irq_handler_exit); + /** * handle_IRQ_event - irq action chain handler * @irq: the interrupt number @@ -332,7 +336,9 @@ irqreturn_t handle_IRQ_event(unsigned int irq, struct irqaction *action) local_irq_enable_in_hardirq(); do { + trace_irq_handler_entry(irq, action); ret = action->handler(irq, action->dev_id); + trace_irq_handler_exit(irq, action, ret); if (ret == IRQ_HANDLED) status |= action->flags; retval |= ret; diff --git a/kernel/trace/events.c b/kernel/trace/events.c index 38c89eef99ee..3c75623893cc 100644 --- a/kernel/trace/events.c +++ b/kernel/trace/events.c @@ -6,8 +6,10 @@ /* trace/.h here */ #include +#include #include "trace_events.h" /* trace/_event_types.h here */ #include +#include From 6409c4da289d6905f7ae2bd0630438368439bda2 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 12 May 2008 21:21:14 +0200 Subject: [PATCH 191/413] sched: sched_clock() improvement: use in_nmi() make sure we dont execute more complex sched_clock() code in NMI context. Acked-by: Peter Zijlstra Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/sched_clock.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/kernel/sched_clock.c b/kernel/sched_clock.c index a0b0852414cc..db69174b1178 100644 --- a/kernel/sched_clock.c +++ b/kernel/sched_clock.c @@ -29,6 +29,7 @@ #include #include #include +#include /* * Scheduler clock - returns current time in nanosec units. @@ -151,6 +152,13 @@ u64 sched_clock_cpu(int cpu) struct sched_clock_data *scd = cpu_sdc(cpu); u64 now, clock, this_clock, remote_clock; + /* + * Normally this is not called in NMI context - but if it is, + * trying to do any locking here is totally lethal. + */ + if (unlikely(in_nmi())) + return scd->clock; + if (unlikely(!sched_clock_running)) return 0ull; From 14131f2f98ac350ee9e73faed916d2238a8b6a0d Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 26 Feb 2009 18:47:11 +0100 Subject: [PATCH 192/413] tracing: implement trace_clock_*() APIs Impact: implement new tracing timestamp APIs Add three trace clock variants, with differing scalability/precision tradeoffs: - local: CPU-local trace clock - medium: scalable global clock with some jitter - global: globally monotonic, serialized clock Make the ring-buffer use the local trace clock internally. Acked-by: Peter Zijlstra Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- include/linux/trace_clock.h | 19 +++++++ kernel/trace/Makefile | 1 + kernel/trace/ring_buffer.c | 5 +- kernel/trace/trace_clock.c | 101 ++++++++++++++++++++++++++++++++++++ 4 files changed, 123 insertions(+), 3 deletions(-) create mode 100644 include/linux/trace_clock.h create mode 100644 kernel/trace/trace_clock.c diff --git a/include/linux/trace_clock.h b/include/linux/trace_clock.h new file mode 100644 index 000000000000..7a8130384087 --- /dev/null +++ b/include/linux/trace_clock.h @@ -0,0 +1,19 @@ +#ifndef _LINUX_TRACE_CLOCK_H +#define _LINUX_TRACE_CLOCK_H + +/* + * 3 trace clock variants, with differing scalability/precision + * tradeoffs: + * + * - local: CPU-local trace clock + * - medium: scalable global clock with some jitter + * - global: globally monotonic, serialized clock + */ +#include +#include + +extern u64 notrace trace_clock_local(void); +extern u64 notrace trace_clock(void); +extern u64 notrace trace_clock_global(void); + +#endif /* _LINUX_TRACE_CLOCK_H */ diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 664b6c0dc75a..c931fe0560cb 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -19,6 +19,7 @@ obj-$(CONFIG_FUNCTION_TRACER) += libftrace.o obj-$(CONFIG_RING_BUFFER) += ring_buffer.o obj-$(CONFIG_TRACING) += trace.o +obj-$(CONFIG_TRACING) += trace_clock.o obj-$(CONFIG_TRACING) += trace_output.o obj-$(CONFIG_TRACING) += trace_stat.o obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 8f19f1aa42b0..a8c275c01e83 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -4,6 +4,7 @@ * Copyright (C) 2008 Steven Rostedt */ #include +#include #include #include #include @@ -12,7 +13,6 @@ #include #include #include -#include /* used for sched_clock() (for now) */ #include #include #include @@ -112,14 +112,13 @@ EXPORT_SYMBOL_GPL(tracing_is_on); /* Up this if you want to test the TIME_EXTENTS and normalization */ #define DEBUG_SHIFT 0 -/* FIXME!!! */ u64 ring_buffer_time_stamp(int cpu) { u64 time; preempt_disable_notrace(); /* shift to debug/test normalization and TIME_EXTENTS */ - time = sched_clock() << DEBUG_SHIFT; + time = trace_clock_local() << DEBUG_SHIFT; preempt_enable_no_resched_notrace(); return time; diff --git a/kernel/trace/trace_clock.c b/kernel/trace/trace_clock.c new file mode 100644 index 000000000000..2d4953f93560 --- /dev/null +++ b/kernel/trace/trace_clock.c @@ -0,0 +1,101 @@ +/* + * tracing clocks + * + * Copyright (C) 2009 Red Hat, Inc., Ingo Molnar + * + * Implements 3 trace clock variants, with differing scalability/precision + * tradeoffs: + * + * - local: CPU-local trace clock + * - medium: scalable global clock with some jitter + * - global: globally monotonic, serialized clock + * + * Tracer plugins will chose a default from these clocks. + */ +#include +#include +#include +#include +#include +#include + +/* + * trace_clock_local(): the simplest and least coherent tracing clock. + * + * Useful for tracing that does not cross to other CPUs nor + * does it go through idle events. + */ +u64 notrace trace_clock_local(void) +{ + /* + * sched_clock() is an architecture implemented, fast, scalable, + * lockless clock. It is not guaranteed to be coherent across + * CPUs, nor across CPU idle events. + */ + return sched_clock(); +} + +/* + * trace_clock(): 'inbetween' trace clock. Not completely serialized, + * but not completely incorrect when crossing CPUs either. + * + * This is based on cpu_clock(), which will allow at most ~1 jiffy of + * jitter between CPUs. So it's a pretty scalable clock, but there + * can be offsets in the trace data. + */ +u64 notrace trace_clock(void) +{ + return cpu_clock(raw_smp_processor_id()); +} + + +/* + * trace_clock_global(): special globally coherent trace clock + * + * It has higher overhead than the other trace clocks but is still + * an order of magnitude faster than GTOD derived hardware clocks. + * + * Used by plugins that need globally coherent timestamps. + */ + +static u64 prev_trace_clock_time; + +static raw_spinlock_t trace_clock_lock ____cacheline_aligned_in_smp = + (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; + +u64 notrace trace_clock_global(void) +{ + unsigned long flags; + int this_cpu; + u64 now; + + raw_local_irq_save(flags); + + this_cpu = raw_smp_processor_id(); + now = cpu_clock(this_cpu); + /* + * If in an NMI context then dont risk lockups and return the + * cpu_clock() time: + */ + if (unlikely(in_nmi())) + goto out; + + __raw_spin_lock(&trace_clock_lock); + + /* + * TODO: if this happens often then maybe we should reset + * my_scd->clock to prev_trace_clock_time+1, to make sure + * we start ticking with the local clock from now on? + */ + if ((s64)(now - prev_trace_clock_time) < 0) + now = prev_trace_clock_time + 1; + + prev_trace_clock_time = now; + + __raw_spin_unlock(&trace_clock_lock); + + out: + raw_local_irq_restore(flags); + + return now; +} From a8259075074fb09c230b4cd2c8d3ee3c49d6ecd1 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 26 Feb 2009 22:19:12 -0500 Subject: [PATCH 193/413] tracing: add options directory and core option files This patch creates an options directory in the debugfs, that contains the available tracing options. These files contain 1 or 0, where 1 is the option is enabled and 0 it is disabled. Simply echoing in 1 will enable the option and 0 will disable it. This patch only contains the core options, not the tracer options. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 117 +++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 117 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index bdaf60d3d337..40e983ed994f 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3093,6 +3093,121 @@ static void tracing_init_debugfs_percpu(long cpu) #include "trace_selftest.c" #endif +static ssize_t +trace_options_core_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + long index = (long)filp->private_data; + char *buf; + + if (trace_flags & (1 << index)) + buf = "1\n"; + else + buf = "0\n"; + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2); +} + +static ssize_t +trace_options_core_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + long index = (long)filp->private_data; + char buf[64]; + unsigned long val; + int ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + switch (val) { + case 0: + trace_flags &= ~(1 << index); + break; + case 1: + trace_flags |= 1 << index; + break; + + default: + return -EINVAL; + } + + *ppos += cnt; + + return cnt; +} + + +static const struct file_operations trace_options_core_fops = { + .open = tracing_open_generic, + .read = trace_options_core_read, + .write = trace_options_core_write, +}; + +static struct dentry *trace_options_init_dentry(void) +{ + struct dentry *d_tracer; + static struct dentry *t_options; + + if (t_options) + return t_options; + + d_tracer = tracing_init_dentry(); + if (!d_tracer) + return NULL; + + t_options = debugfs_create_dir("options", d_tracer); + if (!t_options) { + pr_warning("Could not create debugfs directory 'options'\n"); + return NULL; + } + + return t_options; +} + +static struct dentry * +create_trace_option_core_file(const char *option, long index) +{ + struct dentry *t_options; + struct dentry *entry; + + t_options = trace_options_init_dentry(); + if (!t_options) + return NULL; + + entry = debugfs_create_file(option, 0644, t_options, (void *)index, + &trace_options_core_fops); + + return entry; +} + +static __init void create_trace_options_dir(void) +{ + struct dentry *t_options; + struct dentry *entry; + int i; + + t_options = trace_options_init_dentry(); + if (!t_options) + return; + + for (i = 0; trace_options[i]; i++) { + entry = create_trace_option_core_file(trace_options[i], i); + if (!entry) + pr_warning("Could not create debugfs %s entry\n", + trace_options[i]); + } +} + static __init int tracer_init_debugfs(void) { struct dentry *d_tracer; @@ -3111,6 +3226,8 @@ static __init int tracer_init_debugfs(void) if (!entry) pr_warning("Could not create debugfs 'trace_options' entry\n"); + create_trace_options_dir(); + entry = debugfs_create_file("tracing_cpumask", 0644, d_tracer, NULL, &tracing_cpumask_fops); if (!entry) From 577b785f55168d5acb3d123ba41bfe8d7981e044 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 26 Feb 2009 23:43:05 -0500 Subject: [PATCH 194/413] tracing: add tracer dependent options to options directory This patch adds the tracer dependent options dynamically to the options directory when the tracer is activated. These options are removed when the tracer is deactivated. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 174 ++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 173 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 40e983ed994f..485c6e7f4465 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2275,8 +2275,17 @@ int tracer_init(struct tracer *t, struct trace_array *tr) return t->init(tr); } +struct trace_option_dentry; + +static struct trace_option_dentry * +create_trace_option_files(struct tracer *tracer); + +static void +destroy_trace_option_files(struct trace_option_dentry *topts); + static int tracing_set_tracer(const char *buf) { + static struct trace_option_dentry *topts; struct trace_array *tr = &global_trace; struct tracer *t; int ret = 0; @@ -2297,7 +2306,12 @@ static int tracing_set_tracer(const char *buf) if (current_trace && current_trace->reset) current_trace->reset(tr); + destroy_trace_option_files(topts); + current_trace = t; + + topts = create_trace_option_files(current_trace); + if (t->init) { ret = tracer_init(t, tr); if (ret) @@ -3093,6 +3107,95 @@ static void tracing_init_debugfs_percpu(long cpu) #include "trace_selftest.c" #endif +struct trace_option_dentry { + struct tracer_opt *opt; + struct tracer_flags *flags; + struct dentry *entry; +}; + +static ssize_t +trace_options_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct trace_option_dentry *topt = filp->private_data; + char *buf; + + if (topt->flags->val & topt->opt->bit) + buf = "1\n"; + else + buf = "0\n"; + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, 2); +} + +static ssize_t +trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct trace_option_dentry *topt = filp->private_data; + unsigned long val; + char buf[64]; + int ret; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + ret = strict_strtoul(buf, 10, &val); + if (ret < 0) + return ret; + + ret = 0; + switch (val) { + case 0: + /* do nothing if already cleared */ + if (!(topt->flags->val & topt->opt->bit)) + break; + + mutex_lock(&trace_types_lock); + if (current_trace->set_flag) + ret = current_trace->set_flag(topt->flags->val, + topt->opt->bit, 0); + mutex_unlock(&trace_types_lock); + if (ret) + return ret; + topt->flags->val &= ~topt->opt->bit; + break; + case 1: + /* do nothing if already set */ + if (topt->flags->val & topt->opt->bit) + break; + + mutex_lock(&trace_types_lock); + if (current_trace->set_flag) + ret = current_trace->set_flag(topt->flags->val, + topt->opt->bit, 1); + mutex_unlock(&trace_types_lock); + if (ret) + return ret; + topt->flags->val |= topt->opt->bit; + break; + + default: + return -EINVAL; + } + + *ppos += cnt; + + return cnt; +} + + +static const struct file_operations trace_options_fops = { + .open = tracing_open_generic, + .read = trace_options_read, + .write = trace_options_write, +}; + static ssize_t trace_options_core_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -3146,7 +3249,6 @@ trace_options_core_write(struct file *filp, const char __user *ubuf, size_t cnt, return cnt; } - static const struct file_operations trace_options_core_fops = { .open = tracing_open_generic, .read = trace_options_core_read, @@ -3174,6 +3276,76 @@ static struct dentry *trace_options_init_dentry(void) return t_options; } +static void +create_trace_option_file(struct trace_option_dentry *topt, + struct tracer_flags *flags, + struct tracer_opt *opt) +{ + struct dentry *t_options; + struct dentry *entry; + + t_options = trace_options_init_dentry(); + if (!t_options) + return; + + topt->flags = flags; + topt->opt = opt; + + entry = debugfs_create_file(opt->name, 0644, t_options, topt, + &trace_options_fops); + + topt->entry = entry; + +} + +static struct trace_option_dentry * +create_trace_option_files(struct tracer *tracer) +{ + struct trace_option_dentry *topts; + struct tracer_flags *flags; + struct tracer_opt *opts; + int cnt; + + if (!tracer) + return NULL; + + flags = tracer->flags; + + if (!flags || !flags->opts) + return NULL; + + opts = flags->opts; + + for (cnt = 0; opts[cnt].name; cnt++) + ; + + topts = kzalloc(sizeof(*topts) * (cnt + 1), GFP_KERNEL); + if (!topts) + return NULL; + + for (cnt = 0; opts[cnt].name; cnt++) + create_trace_option_file(&topts[cnt], flags, + &opts[cnt]); + + return topts; +} + +static void +destroy_trace_option_files(struct trace_option_dentry *topts) +{ + int cnt; + + if (!topts) + return; + + for (cnt = 0; topts[cnt].opt; cnt++) { + if (topts[cnt].entry) + debugfs_remove(topts[cnt].entry); + } + + kfree(topts); +} + static struct dentry * create_trace_option_core_file(const char *option, long index) { From d8e83d26b5ab3b31ee0ff6d093a2627707a1e221 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 26 Feb 2009 23:55:58 -0500 Subject: [PATCH 195/413] tracing: add protection around open use of current_tracer Impact: fix to possible race conditions There's some uses of current_tracer that is not protected by the trace_types_lock. There is a small chance that a sysadmin changes the tracer while the current_tracer is being referenced. If the race is hit, it is unlikely to cause any harm since the tracers are constant and are not freed. But some strang side effects may occur. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 485c6e7f4465..6c89ec6cbd48 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2024,12 +2024,12 @@ static ssize_t tracing_trace_options_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { - int i; + struct tracer_opt *trace_opts; + u32 tracer_flags; + int len = 0; char *buf; int r = 0; - int len = 0; - u32 tracer_flags = current_trace->flags->val; - struct tracer_opt *trace_opts = current_trace->flags->opts; + int i; /* calculate max size */ @@ -2038,6 +2038,10 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, len += 3; /* "no" and space */ } + mutex_lock(&trace_types_lock); + tracer_flags = current_trace->flags->val; + trace_opts = current_trace->flags->opts; + /* * Increase the size with names of options specific * of the current tracer. @@ -2049,8 +2053,10 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, /* +2 for \n and \0 */ buf = kmalloc(len + 2, GFP_KERNEL); - if (!buf) + if (!buf) { + mutex_unlock(&trace_types_lock); return -ENOMEM; + } for (i = 0; trace_options[i]; i++) { if (trace_flags & (1 << i)) @@ -2067,6 +2073,7 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, r += sprintf(buf + r, "no%s ", trace_opts[i].name); } + mutex_unlock(&trace_types_lock); r += sprintf(buf + r, "\n"); WARN_ON(r >= len + 2); @@ -2074,7 +2081,6 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r); kfree(buf); - return r; } @@ -2149,7 +2155,9 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf, /* If no option could be set, test the specific tracer options */ if (!trace_options[i]) { + mutex_lock(&trace_types_lock); ret = set_tracer_option(current_trace, cmp, neg); + mutex_unlock(&trace_types_lock); if (ret) return ret; } From 85a2f9b46f8cd8aaa11c64c715e1ea3ec27ec486 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 00:12:38 -0500 Subject: [PATCH 196/413] tracing: use pointer error returns for __tracing_open Impact: fix compile warning and clean up When I first wrote __tracing_open, instead of passing the error code via the ERR_PTR macros, I lazily used a separate parameter to hold the return for errors. When Frederic Weisbecker updated that function, he used the Linux kernel ERR_PTR for the returns. This caused the parameter return to possibly not be initialized on error. gcc correctly pointed this out with a warning. This patch converts the entire function to use the Linux kernel ERR_PTR macro methods. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 46 +++++++++++++++++++++++--------------------- 1 file changed, 24 insertions(+), 22 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 6c89ec6cbd48..304e02ce39c4 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1684,23 +1684,20 @@ static struct seq_operations tracer_seq_ops = { }; static struct trace_iterator * -__tracing_open(struct inode *inode, struct file *file, int *ret) +__tracing_open(struct inode *inode, struct file *file) { long cpu_file = (long) inode->i_private; + void *fail_ret = ERR_PTR(-ENOMEM); struct trace_iterator *iter; struct seq_file *m; - int cpu; + int cpu, ret; - if (tracing_disabled) { - *ret = -ENODEV; - return NULL; - } + if (tracing_disabled) + return ERR_PTR(-ENODEV); iter = kzalloc(sizeof(*iter), GFP_KERNEL); - if (!iter) { - *ret = -ENOMEM; - goto out; - } + if (!iter) + return ERR_PTR(-ENOMEM); /* * We make a copy of the current tracer to avoid concurrent @@ -1708,10 +1705,9 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) */ mutex_lock(&trace_types_lock); iter->trace = kzalloc(sizeof(*iter->trace), GFP_KERNEL); - if (!iter->trace) { - *ret = -ENOMEM; + if (!iter->trace) goto fail; - } + if (current_trace) *iter->trace = *current_trace; @@ -1750,9 +1746,11 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) } /* TODO stop tracer */ - *ret = seq_open(file, &tracer_seq_ops); - if (*ret) + ret = seq_open(file, &tracer_seq_ops); + if (ret < 0) { + fail_ret = ERR_PTR(ret); goto fail_buffer; + } m = file->private_data; m->private = iter; @@ -1762,7 +1760,6 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) mutex_unlock(&trace_types_lock); - out: return iter; fail_buffer: @@ -1775,7 +1772,7 @@ __tracing_open(struct inode *inode, struct file *file, int *ret) kfree(iter->trace); kfree(iter); - return ERR_PTR(-ENOMEM); + return fail_ret; } int tracing_open_generic(struct inode *inode, struct file *filp) @@ -1815,9 +1812,12 @@ static int tracing_release(struct inode *inode, struct file *file) static int tracing_open(struct inode *inode, struct file *file) { - int ret; + struct trace_iterator *iter; + int ret = 0; - __tracing_open(inode, file, &ret); + iter = __tracing_open(inode, file); + if (IS_ERR(iter)) + ret = PTR_ERR(iter); return ret; } @@ -1825,11 +1825,13 @@ static int tracing_open(struct inode *inode, struct file *file) static int tracing_lt_open(struct inode *inode, struct file *file) { struct trace_iterator *iter; - int ret; + int ret = 0; - iter = __tracing_open(inode, file, &ret); + iter = __tracing_open(inode, file); - if (!ret) + if (IS_ERR(iter)) + ret = PTR_ERR(iter); + else iter->iter_flags |= TRACE_FILE_LAT_FMT; return ret; From 5c6a3ae1b4beebb56e2916b84f1208d96a9e32ff Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 00:22:21 -0500 Subject: [PATCH 197/413] tracing: use newline separator for trace options list Impact: clean up Instead of listing the trace options like: # cat /debug/tracing/trace_options print-parent nosym-offset nosym-addr noverbose noraw nohex nobin noblock nostacktrace nosched-tree ftrace_printk noftrace_preempt nobranch annotate nouserstacktrace nosym-userobj We now list them like: # cat /debug/tracing/trace_options print-parent nosym-offset nosym-addr noverbose noraw nohex nobin noblock nostacktrace nosched-tree ftrace_printk noftrace_preempt nobranch annotate nouserstacktrace nosym-userobj Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 304e02ce39c4..5db7485158df 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2037,7 +2037,7 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, /* calculate max size */ for (i = 0; trace_options[i]; i++) { len += strlen(trace_options[i]); - len += 3; /* "no" and space */ + len += 3; /* "no" and newline */ } mutex_lock(&trace_types_lock); @@ -2050,7 +2050,7 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, */ for (i = 0; trace_opts[i].name; i++) { len += strlen(trace_opts[i].name); - len += 3; /* "no" and space */ + len += 3; /* "no" and newline */ } /* +2 for \n and \0 */ @@ -2062,22 +2062,21 @@ tracing_trace_options_read(struct file *filp, char __user *ubuf, for (i = 0; trace_options[i]; i++) { if (trace_flags & (1 << i)) - r += sprintf(buf + r, "%s ", trace_options[i]); + r += sprintf(buf + r, "%s\n", trace_options[i]); else - r += sprintf(buf + r, "no%s ", trace_options[i]); + r += sprintf(buf + r, "no%s\n", trace_options[i]); } for (i = 0; trace_opts[i].name; i++) { if (tracer_flags & trace_opts[i].bit) - r += sprintf(buf + r, "%s ", + r += sprintf(buf + r, "%s\n", trace_opts[i].name); else - r += sprintf(buf + r, "no%s ", + r += sprintf(buf + r, "no%s\n", trace_opts[i].name); } mutex_unlock(&trace_types_lock); - r += sprintf(buf + r, "\n"); WARN_ON(r >= len + 2); r = simple_read_from_buffer(ubuf, cnt, ppos, buf, r); From 0cfe82451dfa3ebf4e69158f2eb450f2fbb6b715 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 10:51:10 -0500 Subject: [PATCH 198/413] tracing: replace kzalloc with kcalloc Impact: clean up kcalloc is a better approach to allocate a NULL array. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 5db7485158df..9c5987aca74b 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3328,7 +3328,7 @@ create_trace_option_files(struct tracer *tracer) for (cnt = 0; opts[cnt].name; cnt++) ; - topts = kzalloc(sizeof(*topts) * (cnt + 1), GFP_KERNEL); + topts = kcalloc(cnt + 1, sizeof(*topts), GFP_KERNEL); if (!topts) return NULL; From eb594e45f6979cd10b18d87f7b3f02119e00a108 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 17:36:06 -0500 Subject: [PATCH 199/413] tracing: move trace point formats to files in include/trace directory Impact: clean up To further facilitate the ease of adding trace points for developers, this patch creates include/trace/trace_events.h and include/trace/trace_event_types.h. The former file will hold the trace/.h files and the latter will hold the trace/_event_types.h files. To create new tracepoints and to have them automatically appear in the event tracer, a developer makes the trace/.h file which includes and the trace/_event_types.h file. The trace/_event_types.h file will hold the TRACE_FORMAT macros. Then add the trace/.h file to trace/trace_events.h, and add the trace/_event_types.h to the trace_event_types.h file. No need to modify files elsewhere. Signed-off-by: Steven Rostedt --- include/trace/trace_event_types.h | 4 ++++ include/trace/trace_events.h | 4 ++++ kernel/trace/events.c | 10 ++-------- 3 files changed, 10 insertions(+), 8 deletions(-) create mode 100644 include/trace/trace_event_types.h create mode 100644 include/trace/trace_events.h diff --git a/include/trace/trace_event_types.h b/include/trace/trace_event_types.h new file mode 100644 index 000000000000..33c8ed5ccb6c --- /dev/null +++ b/include/trace/trace_event_types.h @@ -0,0 +1,4 @@ +/* trace/_event_types.h here */ + +#include +#include diff --git a/include/trace/trace_events.h b/include/trace/trace_events.h new file mode 100644 index 000000000000..ea2ef2051762 --- /dev/null +++ b/include/trace/trace_events.h @@ -0,0 +1,4 @@ +/* trace/.h here */ + +#include +#include diff --git a/kernel/trace/events.c b/kernel/trace/events.c index 3c75623893cc..46e27ad2487e 100644 --- a/kernel/trace/events.c +++ b/kernel/trace/events.c @@ -1,15 +1,9 @@ /* * This is the place to register all trace points as events. - * Include the trace/.h at the top. - * Include the trace/_event_types.h at the bottom. */ -/* trace/.h here */ -#include -#include +#include #include "trace_events.h" -/* trace/_event_types.h here */ -#include -#include +#include From 6ecc2d1ca39177edb6fbdb7412948b0e9f409d02 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 21:33:02 -0500 Subject: [PATCH 200/413] tracing: add subsystem level to trace events If a trace point header defines TRACE_SYSTEM, then it will add the following trace points into that event system. If include/trace/irq_event_types.h has: #define TRACE_SYSTEM irq at the top and #undef TRACE_SYSTEM at the bottom, then a directory "irq" will be created in the /debug/tracing/events directory. Inside that directory will contain the two trace points that are defined in include/trace/irq_event_types.h. Only adding the above to irq and not to sched, we get: # ls /debug/tracing/events/ irq sched_process_exit sched_signal_send sched_wakeup_new sched_kthread_stop sched_process_fork sched_switch sched_kthread_stop_ret sched_process_free sched_wait_task sched_migrate_task sched_process_wait sched_wakeup # ls /debug/tracing/events/irq irq_handler_entry irq_handler_exit If we add #define TRACE_SYSTEM sched to the trace/sched_event_types.h then the rest of the trace events will be put in a sched directory within the events directory. I've been playing with this idea of the subsystem for a while, but recently Tom Zanussi posted some patches to lkml that included this method. Tom's approach was clean and got me to finally put some effort to clean up the event trace points. Thanks to Tom Zanussi for demonstrating how nice the subsystem method is. Signed-off-by: Steven Rostedt --- kernel/trace/events.c | 4 ++++ kernel/trace/trace_events.c | 48 +++++++++++++++++++++++++++++++++++++ kernel/trace/trace_events.h | 2 ++ 3 files changed, 54 insertions(+) diff --git a/kernel/trace/events.c b/kernel/trace/events.c index 46e27ad2487e..4e4e45860c58 100644 --- a/kernel/trace/events.c +++ b/kernel/trace/events.c @@ -2,6 +2,10 @@ * This is the place to register all trace points as events. */ +/* someday this needs to go in a generic header */ +#define __STR(x) #x +#define STR(x) __STR(x) + #include #include "trace_events.h" diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 3bcb9df93342..19332200c457 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -345,11 +345,59 @@ static struct dentry *event_trace_events_dir(void) return d_events; } +struct event_subsystem { + struct list_head list; + const char *name; + struct dentry *entry; +}; + +static LIST_HEAD(event_subsystems); + +static struct dentry * +event_subsystem_dir(const char *name, struct dentry *d_events) +{ + struct event_subsystem *system; + + /* First see if we did not already create this dir */ + list_for_each_entry(system, &event_subsystems, list) { + if (strcmp(system->name, name) == 0) + return system->entry; + } + + /* need to create new entry */ + system = kmalloc(sizeof(*system), GFP_KERNEL); + if (!system) { + pr_warning("No memory to create event subsystem %s\n", + name); + return d_events; + } + + system->entry = debugfs_create_dir(name, d_events); + if (!system->entry) { + pr_warning("Could not create event subsystem %s\n", + name); + kfree(system); + return d_events; + } + + system->name = name; + list_add(&system->list, &event_subsystems); + + return system->entry; +} + static int event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) { struct dentry *entry; + /* + * If the trace point header did not define TRACE_SYSTEM + * then the system would be called "TRACE_SYSTEM". + */ + if (strcmp(call->system, "TRACE_SYSTEM") != 0) + d_events = event_subsystem_dir(call->system, d_events); + call->dir = debugfs_create_dir(call->name, d_events); if (!call->dir) { pr_warning("Could not create debugfs " diff --git a/kernel/trace/trace_events.h b/kernel/trace/trace_events.h index deb95e5006c8..b015d7b19878 100644 --- a/kernel/trace/trace_events.h +++ b/kernel/trace/trace_events.h @@ -7,6 +7,7 @@ struct ftrace_event_call { char *name; + char *system; struct dentry *dir; int enabled; int (*regfunc)(void); @@ -44,6 +45,7 @@ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ .name = #call, \ + .system = STR(TRACE_SYSTEM), \ .regfunc = ftrace_reg_event_##call, \ .unregfunc = ftrace_unreg_event_##call, \ } From b628b3e629b1436710e59a21cc020fbb04a52ce1 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 23:32:58 -0500 Subject: [PATCH 201/413] tracing: make the set_event and available_events subsystem aware This patch makes the event files, set_event and available_events aware of the subsystem. Now you can enable an entire subsystem with: echo 'irq:*' > set_event Note: the '*' is not needed. Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 43 ++++++++++++++++++++++++++++++++++--- 1 file changed, 40 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 19332200c457..b811eb343522 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -12,6 +12,8 @@ #include "trace_events.h" +#define TRACE_SYSTEM "TRACE_SYSTEM" + #define events_for_each(event) \ for (event = __start_ftrace_events; \ (unsigned long)event < (unsigned long)__stop_ftrace_events; \ @@ -45,14 +47,47 @@ static void ftrace_clear_events(void) static int ftrace_set_clr_event(char *buf, int set) { struct ftrace_event_call *call = __start_ftrace_events; + char *event = NULL, *sub = NULL, *match; + int ret = -EINVAL; + /* + * The buf format can be : + * *: means any event by that name. + * : is the same. + * + * :* means all events in that subsystem + * : means the same. + * + * (no ':') means all events in a subsystem with + * the name or any event that matches + */ + + match = strsep(&buf, ":"); + if (buf) { + sub = match; + event = buf; + match = NULL; + + if (!strlen(sub) || strcmp(sub, "*") == 0) + sub = NULL; + if (!strlen(event) || strcmp(event, "*") == 0) + event = NULL; + } events_for_each(call) { if (!call->name) continue; - if (strcmp(buf, call->name) != 0) + if (match && + strcmp(match, call->name) != 0 && + strcmp(match, call->system) != 0) + continue; + + if (sub && strcmp(sub, call->system) != 0) + continue; + + if (event && strcmp(event, call->name) != 0) continue; if (set) { @@ -68,9 +103,9 @@ static int ftrace_set_clr_event(char *buf, int set) call->enabled = 0; call->unregfunc(); } - return 0; + ret = 0; } - return -EINVAL; + return ret; } /* 128 should be much more than enough */ @@ -200,6 +235,8 @@ static int t_show(struct seq_file *m, void *v) { struct ftrace_event_call *call = v; + if (strcmp(call->system, TRACE_SYSTEM) != 0) + seq_printf(m, "%s:", call->system); seq_printf(m, "%s\n", call->name); return 0; From 0ec2ef1505b3e1f54b07bf64f184c92859c3e13f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 23:41:43 -0500 Subject: [PATCH 202/413] tracing: add subsystem irq for irq events Add the TRACE_SYSTEM irq for the irq events. Signed-off-by: Steven Rostedt --- include/trace/irq_event_types.h | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/include/trace/irq_event_types.h b/include/trace/irq_event_types.h index 5d0919fdd2d4..47a2be1b2d7d 100644 --- a/include/trace/irq_event_types.h +++ b/include/trace/irq_event_types.h @@ -5,6 +5,9 @@ # error Unless you know what you are doing. #endif +#undef TRACE_SYSTEM +#define TRACE_SYSTEM irq + TRACE_FORMAT(irq_handler_entry, TPPROTO(int irq, struct irqaction *action), TPARGS(irq, action), @@ -15,3 +18,5 @@ TRACE_FORMAT(irq_handler_exit, TPARGS(irq, action, ret), TPFMT("irq=%d handler=%s return=%s", irq, action->name, ret ? "handled" : "unhandled")); + +#undef TRACE_SYSTEM From 3d7ba938da8481b4f7f9ed3d943dbae49389b284 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 23:45:52 -0500 Subject: [PATCH 203/413] tracing: add subsystem sched for sched events Add the TRACE_SYSTEM sched for the sched events. Signed-off-by: Steven Rostedt --- include/trace/sched_event_types.h | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index a3d3d66a51c8..2ada206565a3 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -5,6 +5,9 @@ # error Unless you know what you are doing. #endif +#undef TRACE_SYSTEM +#define TRACE_SYSTEM sched + TRACE_FORMAT(sched_kthread_stop, TPPROTO(struct task_struct *t), TPARGS(t), @@ -70,3 +73,5 @@ TRACE_FORMAT(sched_signal_send, TPPROTO(int sig, struct task_struct *p), TPARGS(sig, p), TPFMT("sig: %d task %s:%d", sig, p->comm, p->pid)); + +#undef TRACE_SYSTEM From ef5580d0fffce6e0a01043bac0625128b5d409a7 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 19:38:04 -0500 Subject: [PATCH 204/413] tracing: add interface to write into current tracer buffer Right now all tracers must manage their own trace buffers. This was to enforce tracers to be independent in case we finally decide to allow each tracer to have their own trace buffer. But now we are adding event tracing that writes to the current tracer's buffer. This adds an interface to allow events to write to the current tracer buffer without having to manage its own. Since event tracing has no "tracer", and is just a way to hook into any other tracer. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 14 ++++++++++++++ kernel/trace/trace.h | 6 ++++++ 2 files changed, 20 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 9c5987aca74b..c5e39cd7310d 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -846,6 +846,20 @@ void trace_buffer_unlock_commit(struct trace_array *tr, trace_wake_up(); } +struct ring_buffer_event * +trace_current_buffer_lock_reserve(unsigned char type, unsigned long len, + unsigned long flags, int pc) +{ + return trace_buffer_lock_reserve(&global_trace, + type, len, flags, pc); +} + +void trace_current_buffer_unlock_commit(struct ring_buffer_event *event, + unsigned long flags, int pc) +{ + return trace_buffer_unlock_commit(&global_trace, event, flags, pc); +} + void trace_function(struct trace_array *tr, unsigned long ip, unsigned long parent_ip, unsigned long flags, diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 632191770aac..adf161f6dd11 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -442,6 +442,12 @@ void trace_buffer_unlock_commit(struct trace_array *tr, struct ring_buffer_event *event, unsigned long flags, int pc); +struct ring_buffer_event * +trace_current_buffer_lock_reserve(unsigned char type, unsigned long len, + unsigned long flags, int pc); +void trace_current_buffer_unlock_commit(struct ring_buffer_event *event, + unsigned long flags, int pc); + struct trace_entry *tracing_get_trace_entry(struct trace_array *tr, struct trace_array_cpu *data); From c32e827b25054cb17b79cf97fb5e63ae4ce2223c Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 27 Feb 2009 19:12:30 -0500 Subject: [PATCH 205/413] tracing: add raw trace point recording infrastructure Impact: lower overhead tracing The current event tracer can automatically pick up trace points that are registered with the TRACE_FORMAT macro. But it required a printf format string and parsing. Although, this adds the ability to get guaranteed information like task names and such, it took a hit in overhead processing. This processing can add about 500-1000 nanoseconds overhead, but in some cases that too is considered too much and we want to shave off as much from this overhead as possible. Tom Zanussi recently posted tracing patches to lkml that are based on a nice idea about capturing the data via C structs using STRUCT_ENTER, STRUCT_EXIT type of macros. I liked that method very much, but did not like the implementation that required a developer to add data/code in several disjoint locations. This patch extends the event_tracer macros to do a similar "raw C" approach that Tom Zanussi did. But instead of having the developers needing to tweak a bunch of code all over the place, they can do it all in one macro - preferably placed near the code that it is tracing. That makes it much more likely that tracepoints will be maintained on an ongoing basis by the code they modify. The new macro TRACE_EVENT_FORMAT is created for this approach. (Note, a developer may still utilize the more low level DECLARE_TRACE macros if they don't care about getting their traces automatically in the event tracer.) They can also use the existing TRACE_FORMAT if they don't need to code the tracepoint in C, but just want to use the convenience of printf. So if the developer wants to "hardwire" a tracepoint in the fastest possible way, and wants to acquire their data via a user space utility in a raw binary format, or wants to see it in the trace output but not sacrifice any performance, then they can implement the faster but more complex TRACE_EVENT_FORMAT macro. Here's what usage looks like: TRACE_EVENT_FORMAT(name, TPPROTO(proto), TPARGS(args), TPFMT(fmt, fmt_args), TRACE_STUCT( TRACE_FIELD(type1, item1, assign1) TRACE_FIELD(type2, item2, assign2) [...] ), TPRAWFMT(raw_fmt) ); Note name, proto, args, and fmt, are all identical to what TRACE_FORMAT uses. name: is the unique identifier of the trace point proto: The proto type that the trace point uses args: the args in the proto type fmt: printf format to use with the event printf tracer fmt_args: the printf argments to match fmt TRACE_STRUCT starts the ability to create a structure. Each item in the structure is defined with a TRACE_FIELD TRACE_FIELD(type, item, assign) type: the C type of item. item: the name of the item in the stucture assign: what to assign the item in the trace point callback raw_fmt is a way to pretty print the struct. It must match the order of the items are added in TRACE_STUCT An example of this would be: TRACE_EVENT_FORMAT(sched_wakeup, TPPROTO(struct rq *rq, struct task_struct *p, int success), TPARGS(rq, p, success), TPFMT("task %s:%d %s", p->comm, p->pid, success?"succeeded":"failed"), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, p->pid) TRACE_FIELD(int, success, success) ), TPRAWFMT("task %d success=%d") ); This creates us a unique struct of: struct { pid_t pid; int success; }; And the way the call back would assign these values would be: entry->pid = p->pid; entry->success = success; The nice part about this is that the creation of the assignent is done via macro magic in the event tracer. Once the TRACE_EVENT_FORMAT is created, the developer will then have a faster method to record into the ring buffer. They do not need to worry about the tracer itself. The developer would only need to touch the files in include/trace/*.h Again, I would like to give special thanks to Tom Zanussi for this nice idea. Idea-from: Tom Zanussi Signed-off-by: Steven Rostedt --- kernel/trace/events.c | 6 +- kernel/trace/trace.h | 19 +++ kernel/trace/trace_events.c | 2 +- kernel/trace/trace_events.h | 57 -------- kernel/trace/trace_events_stage_1.h | 34 +++++ kernel/trace/trace_events_stage_2.h | 72 +++++++++ kernel/trace/trace_events_stage_3.h | 219 ++++++++++++++++++++++++++++ 7 files changed, 350 insertions(+), 59 deletions(-) delete mode 100644 kernel/trace/trace_events.h create mode 100644 kernel/trace/trace_events_stage_1.h create mode 100644 kernel/trace/trace_events_stage_2.h create mode 100644 kernel/trace/trace_events_stage_3.h diff --git a/kernel/trace/events.c b/kernel/trace/events.c index 4e4e45860c58..f2509cbaacea 100644 --- a/kernel/trace/events.c +++ b/kernel/trace/events.c @@ -8,6 +8,10 @@ #include -#include "trace_events.h" +#include "trace_output.h" + +#include "trace_events_stage_1.h" +#include "trace_events_stage_2.h" +#include "trace_events_stage_3.h" #include diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index adf161f6dd11..aa1ab0cb80ab 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -726,4 +726,23 @@ static inline void trace_branch_disable(void) } #endif /* CONFIG_BRANCH_TRACER */ +struct ftrace_event_call { + char *name; + char *system; + struct dentry *dir; + int enabled; + int (*regfunc)(void); + void (*unregfunc)(void); + int id; + struct dentry *raw_dir; + int raw_enabled; + int (*raw_init)(void); + int (*raw_reg)(void); + void (*raw_unreg)(void); +}; + +void event_trace_printk(unsigned long ip, const char *fmt, ...); +extern struct ftrace_event_call __start_ftrace_events[]; +extern struct ftrace_event_call __stop_ftrace_events[]; + #endif /* _LINUX_KERNEL_TRACE_H */ diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index b811eb343522..77a5c02bd634 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -10,7 +10,7 @@ #include #include -#include "trace_events.h" +#include "trace.h" #define TRACE_SYSTEM "TRACE_SYSTEM" diff --git a/kernel/trace/trace_events.h b/kernel/trace/trace_events.h deleted file mode 100644 index b015d7b19878..000000000000 --- a/kernel/trace/trace_events.h +++ /dev/null @@ -1,57 +0,0 @@ -#ifndef _LINUX_KERNEL_TRACE_EVENTS_H -#define _LINUX_KERNEL_TRACE_EVENTS_H - -#include -#include -#include "trace.h" - -struct ftrace_event_call { - char *name; - char *system; - struct dentry *dir; - int enabled; - int (*regfunc)(void); - void (*unregfunc)(void); -}; - - -#undef TPFMT -#define TPFMT(fmt, args...) fmt "\n", ##args - -#undef TRACE_FORMAT -#define TRACE_FORMAT(call, proto, args, fmt) \ -static void ftrace_event_##call(proto) \ -{ \ - event_trace_printk(_RET_IP_, "(" #call ") " fmt); \ -} \ - \ -static int ftrace_reg_event_##call(void) \ -{ \ - int ret; \ - \ - ret = register_trace_##call(ftrace_event_##call); \ - if (!ret) \ - pr_info("event trace: Could not activate trace point " \ - "probe to " #call); \ - return ret; \ -} \ - \ -static void ftrace_unreg_event_##call(void) \ -{ \ - unregister_trace_##call(ftrace_event_##call); \ -} \ - \ -static struct ftrace_event_call __used \ -__attribute__((__aligned__(4))) \ -__attribute__((section("_ftrace_events"))) event_##call = { \ - .name = #call, \ - .system = STR(TRACE_SYSTEM), \ - .regfunc = ftrace_reg_event_##call, \ - .unregfunc = ftrace_unreg_event_##call, \ -} - -void event_trace_printk(unsigned long ip, const char *fmt, ...); -extern struct ftrace_event_call __start_ftrace_events[]; -extern struct ftrace_event_call __stop_ftrace_events[]; - -#endif /* _LINUX_KERNEL_TRACE_EVENTS_H */ diff --git a/kernel/trace/trace_events_stage_1.h b/kernel/trace/trace_events_stage_1.h new file mode 100644 index 000000000000..fd3bf9382d37 --- /dev/null +++ b/kernel/trace/trace_events_stage_1.h @@ -0,0 +1,34 @@ +/* + * Stage 1 of the trace events. + * + * Override the macros in to include the following: + * + * struct ftrace_raw_ { + * struct trace_entry ent; + * ; + * [...] + * }; + * + * The is created by the TRACE_FIELD(type, item, assign) + * macro. We simply do "type item;", and that will create the fields + * in the structure. + */ + +#undef TRACE_FORMAT +#define TRACE_FORMAT(call, proto, args, fmt) + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(name, proto, args, fmt, tstruct, tpfmt) \ + struct ftrace_raw_##name { \ + struct trace_entry ent; \ + tstruct \ + }; \ + static struct ftrace_event_call event_##name + +#undef TRACE_STRUCT +#define TRACE_STRUCT(args...) args + +#define TRACE_FIELD(type, item, assign) \ + type item; + +#include diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h new file mode 100644 index 000000000000..3eaaef5f19e1 --- /dev/null +++ b/kernel/trace/trace_events_stage_2.h @@ -0,0 +1,72 @@ +/* + * Stage 2 of the trace events. + * + * Override the macros in to include the following: + * + * enum print_line_t + * ftrace_raw_output_(struct trace_iterator *iter, int flags) + * { + * struct trace_seq *s = &iter->seq; + * struct ftrace_raw_ *field; <-- defined in stage 1 + * struct trace_entry *entry; + * int ret; + * + * entry = iter->ent; + * + * if (entry->type != event_.id) { + * WARN_ON_ONCE(1); + * return TRACE_TYPE_UNHANDLED; + * } + * + * field = (typeof(field))entry; + * + * ret = trace_seq_printf(s, "%s", "\n"); + * if (!ret) + * return TRACE_TYPE_PARTIAL_LINE; + * + * return TRACE_TYPE_HANDLED; + * } + * + * This is the method used to print the raw event to the trace + * output format. Note, this is not needed if the data is read + * in binary. + */ + +#undef TRACE_STRUCT +#define TRACE_STRUCT(args...) args + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign) \ + field->item, + + +#undef TPRAWFMT +#define TPRAWFMT(args...) args + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +enum print_line_t \ +ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ +{ \ + struct trace_seq *s = &iter->seq; \ + struct ftrace_raw_##call *field; \ + struct trace_entry *entry; \ + int ret; \ + \ + entry = iter->ent; \ + \ + if (entry->type != event_##call.id) { \ + WARN_ON_ONCE(1); \ + return TRACE_TYPE_UNHANDLED; \ + } \ + \ + field = (typeof(field))entry; \ + \ + ret = trace_seq_printf(s, tpfmt "%s", tstruct "\n"); \ + if (!ret) \ + return TRACE_TYPE_PARTIAL_LINE; \ + \ + return TRACE_TYPE_HANDLED; \ +} + +#include diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h new file mode 100644 index 000000000000..7a161c49deb4 --- /dev/null +++ b/kernel/trace/trace_events_stage_3.h @@ -0,0 +1,219 @@ +/* + * Stage 3 of the trace events. + * + * Override the macros in to include the following: + * + * static void ftrace_event_(proto) + * { + * event_trace_printk(_RET_IP_, "() " ); + * } + * + * static int ftrace_reg_event_(void) + * { + * int ret; + * + * ret = register_trace_(ftrace_event_); + * if (!ret) + * pr_info("event trace: Could not activate trace point " + * "probe to "); + * return ret; + * } + * + * static void ftrace_unreg_event_(void) + * { + * unregister_trace_(ftrace_event_); + * } + * + * For those macros defined with TRACE_FORMAT: + * + * static struct ftrace_event_call __used + * __attribute__((__aligned__(4))) + * __attribute__((section("_ftrace_events"))) event_ = { + * .name = "", + * .regfunc = ftrace_reg_event_, + * .unregfunc = ftrace_unreg_event_, + * } + * + * + * For those macros defined with TRACE_EVENT_FORMAT: + * + * static struct ftrace_event_call event_; + * + * static void ftrace_raw_event_(proto) + * { + * struct ring_buffer_event *event; + * struct ftrace_raw_ *entry; <-- defined in stage 1 + * unsigned long irq_flags; + * int pc; + * + * local_save_flags(irq_flags); + * pc = preempt_count(); + * + * event = trace_current_buffer_lock_reserve(event_.id, + * sizeof(struct ftrace_raw_), + * irq_flags, pc); + * if (!event) + * return; + * entry = ring_buffer_event_data(event); + * + * ; <-- Here we assign the entries by the TRACE_FIELD. + * + * trace_current_buffer_unlock_commit(event, irq_flags, pc); + * } + * + * static int ftrace_raw_reg_event_(void) + * { + * int ret; + * + * ret = register_trace_(ftrace_raw_event_); + * if (!ret) + * pr_info("event trace: Could not activate trace point " + * "probe to "); + * return ret; + * } + * + * static void ftrace_unreg_event_(void) + * { + * unregister_trace_(ftrace_raw_event_); + * } + * + * static struct trace_event ftrace_event_type_ = { + * .trace = ftrace_raw_output_, <-- stage 2 + * }; + * + * static int ftrace_raw_init_event_(void) + * { + * int id; + * + * id = register_ftrace_event(&ftrace_event_type_); + * if (!id) + * return -ENODEV; + * event_.id = id; + * return 0; + * } + * + * static struct ftrace_event_call __used + * __attribute__((__aligned__(4))) + * __attribute__((section("_ftrace_events"))) event_ = { + * .name = "", + * .regfunc = ftrace_reg_event_, + * .unregfunc = ftrace_unreg_event_, + * .raw_init = ftrace_raw_init_event_, + * .raw_reg = ftrace_raw_reg_event_, + * .raw_unreg = ftrace_raw_unreg_event_, + * } + * + */ + +#undef TPFMT +#define TPFMT(fmt, args...) fmt "\n", ##args + +#define _TRACE_FORMAT(call, proto, args, fmt) \ +static void ftrace_event_##call(proto) \ +{ \ + event_trace_printk(_RET_IP_, "(" #call ") " fmt); \ +} \ + \ +static int ftrace_reg_event_##call(void) \ +{ \ + int ret; \ + \ + ret = register_trace_##call(ftrace_event_##call); \ + if (!ret) \ + pr_info("event trace: Could not activate trace point " \ + "probe to " #call); \ + return ret; \ +} \ + \ +static void ftrace_unreg_event_##call(void) \ +{ \ + unregister_trace_##call(ftrace_event_##call); \ +} \ + + +#undef TRACE_FORMAT +#define TRACE_FORMAT(call, proto, args, fmt) \ +_TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) \ +static struct ftrace_event_call __used \ +__attribute__((__aligned__(4))) \ +__attribute__((section("_ftrace_events"))) event_##call = { \ + .name = #call, \ + .system = STR(TRACE_SYSTEM), \ + .regfunc = ftrace_reg_event_##call, \ + .unregfunc = ftrace_unreg_event_##call, \ +} + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign)\ + entry->item = assign; + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +_TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) \ + \ +static struct ftrace_event_call event_##call; \ + \ +static void ftrace_raw_event_##call(proto) \ +{ \ + struct ring_buffer_event *event; \ + struct ftrace_raw_##call *entry; \ + unsigned long irq_flags; \ + int pc; \ + \ + local_save_flags(irq_flags); \ + pc = preempt_count(); \ + \ + event = trace_current_buffer_lock_reserve(event_##call.id, \ + sizeof(struct ftrace_raw_##call), \ + irq_flags, pc); \ + if (!event) \ + return; \ + entry = ring_buffer_event_data(event); \ + \ + tstruct; \ + \ + trace_current_buffer_unlock_commit(event, irq_flags, pc); \ +} \ + \ +static int ftrace_raw_reg_event_##call(void) \ +{ \ + int ret; \ + \ + ret = register_trace_##call(ftrace_raw_event_##call); \ + if (!ret) \ + pr_info("event trace: Could not activate trace point " \ + "probe to " #call); \ + return ret; \ +} \ + \ +static void ftrace_raw_unreg_event_##call(void) \ +{ \ + unregister_trace_##call(ftrace_raw_event_##call); \ +} \ + \ +static struct trace_event ftrace_event_type_##call = { \ + .trace = ftrace_raw_output_##call, \ +}; \ + \ +static int ftrace_raw_init_event_##call(void) \ +{ \ + int id; \ + \ + id = register_ftrace_event(&ftrace_event_type_##call); \ + if (!id) \ + return -ENODEV; \ + event_##call.id = id; \ + return 0; \ +} \ + \ +static struct ftrace_event_call __used \ +__attribute__((__aligned__(4))) \ +__attribute__((section("_ftrace_events"))) event_##call = { \ + .name = #call, \ + .system = STR(TRACE_SYSTEM), \ + .regfunc = ftrace_reg_event_##call, \ + .unregfunc = ftrace_unreg_event_##call, \ + .raw_init = ftrace_raw_init_event_##call, \ + .raw_reg = ftrace_raw_reg_event_##call, \ + .raw_unreg = ftrace_raw_unreg_event_##call, \ +} From fd99498989f3b3feeab89dcadf537138ba136d24 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 28 Feb 2009 02:41:25 -0500 Subject: [PATCH 206/413] tracing: add raw fast tracing interface for trace events This patch adds the interface to enable the C style trace points. In the directory /debugfs/tracing/events/subsystem/event We now have three files: enable : values 0 or 1 to enable or disable the trace event. available_types: values 'raw' and 'printf' which indicate the tracing types available for the trace point. If a developer does not use the TRACE_EVENT_FORMAT macro and just uses the TRACE_FORMAT macro, then only 'printf' will be available. This file is read only. type: values 'raw' or 'printf'. This indicates which type of tracing is active for that trace point. 'printf' is the default and if 'raw' is not available, this file is read only. # echo raw > /debug/tracing/events/sched/sched_wakeup/type # echo 1 > /debug/tracing/events/sched/sched_wakeup/enable Will enable the C style tracing for the sched_wakeup trace point. Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 7 ++ kernel/trace/trace_events.c | 199 +++++++++++++++++++++++++++++++----- 2 files changed, 181 insertions(+), 25 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index aa1ab0cb80ab..f6fa0b9f83a8 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -726,6 +726,12 @@ static inline void trace_branch_disable(void) } #endif /* CONFIG_BRANCH_TRACER */ +/* trace event type bit fields, not numeric */ +enum { + TRACE_EVENT_TYPE_PRINTF = 1, + TRACE_EVENT_TYPE_RAW = 2, +}; + struct ftrace_event_call { char *name; char *system; @@ -736,6 +742,7 @@ struct ftrace_event_call { int id; struct dentry *raw_dir; int raw_enabled; + int type; int (*raw_init)(void); int (*raw_reg)(void); void (*raw_unreg)(void); diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 77a5c02bd634..1d07f800a9ce 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -44,6 +44,36 @@ static void ftrace_clear_events(void) } } +static void ftrace_event_enable_disable(struct ftrace_event_call *call, + int enable) +{ + + switch (enable) { + case 0: + if (call->enabled) { + call->enabled = 0; + call->unregfunc(); + } + if (call->raw_enabled) { + call->raw_enabled = 0; + call->raw_unreg(); + } + break; + case 1: + if (!call->enabled && + (call->type & TRACE_EVENT_TYPE_PRINTF)) { + call->enabled = 1; + call->regfunc(); + } + if (!call->raw_enabled && + (call->type & TRACE_EVENT_TYPE_RAW)) { + call->raw_enabled = 1; + call->raw_reg(); + } + break; + } +} + static int ftrace_set_clr_event(char *buf, int set) { struct ftrace_event_call *call = __start_ftrace_events; @@ -90,19 +120,8 @@ static int ftrace_set_clr_event(char *buf, int set) if (event && strcmp(event, call->name) != 0) continue; - if (set) { - /* Already set? */ - if (call->enabled) - return 0; - call->enabled = 1; - call->regfunc(); - } else { - /* Already cleared? */ - if (!call->enabled) - return 0; - call->enabled = 0; - call->unregfunc(); - } + ftrace_event_enable_disable(call, set); + ret = 0; } return ret; @@ -273,7 +292,7 @@ event_enable_read(struct file *filp, char __user *ubuf, size_t cnt, struct ftrace_event_call *call = filp->private_data; char *buf; - if (call->enabled) + if (call->enabled || call->raw_enabled) buf = "1\n"; else buf = "0\n"; @@ -304,18 +323,8 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, switch (val) { case 0: - if (!call->enabled) - break; - - call->enabled = 0; - call->unregfunc(); - break; case 1: - if (call->enabled) - break; - - call->enabled = 1; - call->regfunc(); + ftrace_event_enable_disable(call, val); break; default: @@ -327,6 +336,107 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, return cnt; } +static ssize_t +event_type_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char buf[16]; + int r = 0; + + if (call->type & TRACE_EVENT_TYPE_PRINTF) + r += sprintf(buf, "printf\n"); + + if (call->type & TRACE_EVENT_TYPE_RAW) + r += sprintf(buf+r, "raw\n"); + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); +} + +static ssize_t +event_type_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char buf[64]; + + /* + * If there's only one type, we can't change it. + * And currently we always have printf type, and we + * may or may not have raw type. + * + * This is a redundant check, the file should be read + * only if this is the case anyway. + */ + + if (!call->raw_init) + return -EPERM; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + buf[cnt] = 0; + + if (!strncmp(buf, "printf", 6) && + (!buf[6] || isspace(buf[6]))) { + + call->type = TRACE_EVENT_TYPE_PRINTF; + + /* + * If raw enabled, the disable it and enable + * printf type. + */ + if (call->raw_enabled) { + call->raw_enabled = 0; + call->raw_unreg(); + + call->enabled = 1; + call->regfunc(); + } + + } else if (!strncmp(buf, "raw", 3) && + (!buf[3] || isspace(buf[3]))) { + + call->type = TRACE_EVENT_TYPE_RAW; + + /* + * If printf enabled, the disable it and enable + * raw type. + */ + if (call->enabled) { + call->enabled = 0; + call->unregfunc(); + + call->raw_enabled = 1; + call->raw_reg(); + } + } else + return -EINVAL; + + *ppos += cnt; + + return cnt; +} + +static ssize_t +event_available_types_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char buf[16]; + int r = 0; + + r += sprintf(buf, "printf\n"); + + if (call->raw_init) + r += sprintf(buf+r, "raw\n"); + + return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); +} + static const struct seq_operations show_event_seq_ops = { .start = t_start, .next = t_next, @@ -362,6 +472,17 @@ static const struct file_operations ftrace_enable_fops = { .write = event_enable_write, }; +static const struct file_operations ftrace_type_fops = { + .open = tracing_open_generic, + .read = event_type_read, + .write = event_type_write, +}; + +static const struct file_operations ftrace_available_types_fops = { + .open = tracing_open_generic, + .read = event_available_types_read, +}; + static struct dentry *event_trace_events_dir(void) { static struct dentry *d_tracer; @@ -427,6 +548,7 @@ static int event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) { struct dentry *entry; + int ret; /* * If the trace point header did not define TRACE_SYSTEM @@ -435,6 +557,18 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) if (strcmp(call->system, "TRACE_SYSTEM") != 0) d_events = event_subsystem_dir(call->system, d_events); + if (call->raw_init) { + ret = call->raw_init(); + if (ret < 0) { + pr_warning("Could not initialize trace point" + " events/%s\n", call->name); + return ret; + } + } + + /* default the output to printf */ + call->type = TRACE_EVENT_TYPE_PRINTF; + call->dir = debugfs_create_dir(call->name, d_events); if (!call->dir) { pr_warning("Could not create debugfs " @@ -448,6 +582,21 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) pr_warning("Could not create debugfs " "'%s/enable' entry\n", call->name); + /* Only let type be writable, if we can change it */ + entry = debugfs_create_file("type", + call->raw_init ? 0644 : 0444, + call->dir, call, + &ftrace_type_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/type' entry\n", call->name); + + entry = debugfs_create_file("available_types", 0444, call->dir, call, + &ftrace_available_types_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/type' available_types\n", call->name); + return 0; } From 629928041c53771f9902753d50fef6b35f36d33d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 28 Feb 2009 02:47:59 -0500 Subject: [PATCH 207/413] tracing: create the C style tracing for the sched subsystem This patch utilizes the TRACE_EVENT_FORMAT macro to enable the C style faster tracing for the sched subsystem trace points. Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 3 + include/trace/sched_event_types.h | 119 +++++++++++++++++++++++------- 2 files changed, 97 insertions(+), 25 deletions(-) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 62d13391a240..152b2f03fb86 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -157,4 +157,7 @@ static inline void tracepoint_synchronize_unregister(void) #define TRACE_FORMAT(name, proto, args, fmt) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) +#define TRACE_EVENT_FORMAT(name, proto, args, fmt, struct, tpfmt) \ + TRACE_FORMAT(name, PARAMS(proto), PARAMS(args), PARAMS(fmt)) + #endif diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index 2ada206565a3..ba059c10b58a 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -1,6 +1,6 @@ /* use instead */ -#ifndef TRACE_FORMAT +#ifndef TRACE_EVENT_FORMAT # error Do not include this file directly. # error Unless you know what you are doing. #endif @@ -8,70 +8,139 @@ #undef TRACE_SYSTEM #define TRACE_SYSTEM sched -TRACE_FORMAT(sched_kthread_stop, +TRACE_EVENT_FORMAT(sched_kthread_stop, TPPROTO(struct task_struct *t), TPARGS(t), - TPFMT("task %s:%d", t->comm, t->pid)); + TPFMT("task %s:%d", t->comm, t->pid), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, t->pid) + ), + TPRAWFMT("task %d") + ); -TRACE_FORMAT(sched_kthread_stop_ret, +TRACE_EVENT_FORMAT(sched_kthread_stop_ret, TPPROTO(int ret), TPARGS(ret), - TPFMT("ret=%d", ret)); + TPFMT("ret=%d", ret), + TRACE_STRUCT( + TRACE_FIELD(int, ret, ret) + ), + TPRAWFMT("ret=%d") + ); -TRACE_FORMAT(sched_wait_task, +TRACE_EVENT_FORMAT(sched_wait_task, TPPROTO(struct rq *rq, struct task_struct *p), TPARGS(rq, p), - TPFMT("task %s:%d", p->comm, p->pid)); + TPFMT("task %s:%d", p->comm, p->pid), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, p->pid) + ), + TPRAWFMT("task %d") + ); -TRACE_FORMAT(sched_wakeup, +TRACE_EVENT_FORMAT(sched_wakeup, TPPROTO(struct rq *rq, struct task_struct *p, int success), TPARGS(rq, p, success), TPFMT("task %s:%d %s", - p->comm, p->pid, success?"succeeded":"failed")); + p->comm, p->pid, success ? "succeeded" : "failed"), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, p->pid) + TRACE_FIELD(int, success, success) + ), + TPRAWFMT("task %d success=%d") + ); -TRACE_FORMAT(sched_wakeup_new, +TRACE_EVENT_FORMAT(sched_wakeup_new, TPPROTO(struct rq *rq, struct task_struct *p, int success), TPARGS(rq, p, success), TPFMT("task %s:%d", - p->comm, p->pid, success?"succeeded":"failed")); + p->comm, p->pid, success ? "succeeded" : "failed"), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, p->pid) + TRACE_FIELD(int, success, success) + ), + TPRAWFMT("task %d success=%d") + ); -TRACE_FORMAT(sched_switch, +TRACE_EVENT_FORMAT(sched_switch, TPPROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next), TPARGS(rq, prev, next), TPFMT("task %s:%d ==> %s:%d", - prev->comm, prev->pid, next->comm, next->pid)); + prev->comm, prev->pid, next->comm, next->pid), + TRACE_STRUCT( + TRACE_FIELD(pid_t, prev_pid, prev->pid) + TRACE_FIELD(int, prev_prio, prev->prio) + TRACE_FIELD(pid_t, next_pid, next->pid) + TRACE_FIELD(int, next_prio, next->prio) + ), + TPRAWFMT("prev %d:%d ==> next %d:%d") + ); -TRACE_FORMAT(sched_migrate_task, +TRACE_EVENT_FORMAT(sched_migrate_task, TPPROTO(struct task_struct *p, int orig_cpu, int dest_cpu), TPARGS(p, orig_cpu, dest_cpu), TPFMT("task %s:%d from: %d to: %d", - p->comm, p->pid, orig_cpu, dest_cpu)); + p->comm, p->pid, orig_cpu, dest_cpu), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, p->pid) + TRACE_FIELD(int, orig_cpu, orig_cpu) + TRACE_FIELD(int, dest_cpu, dest_cpu) + ), + TPRAWFMT("task %d from: %d to: %d") + ); -TRACE_FORMAT(sched_process_free, +TRACE_EVENT_FORMAT(sched_process_free, TPPROTO(struct task_struct *p), TPARGS(p), - TPFMT("task %s:%d", p->comm, p->pid)); + TPFMT("task %s:%d", p->comm, p->pid), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, p->pid) + ), + TPRAWFMT("task %d") + ); -TRACE_FORMAT(sched_process_exit, +TRACE_EVENT_FORMAT(sched_process_exit, TPPROTO(struct task_struct *p), TPARGS(p), - TPFMT("task %s:%d", p->comm, p->pid)); + TPFMT("task %s:%d", p->comm, p->pid), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, p->pid) + ), + TPRAWFMT("task %d") + ); -TRACE_FORMAT(sched_process_wait, +TRACE_EVENT_FORMAT(sched_process_wait, TPPROTO(struct pid *pid), TPARGS(pid), - TPFMT("pid %d", pid)); + TPFMT("pid %d", pid_nr(pid)), + TRACE_STRUCT( + TRACE_FIELD(pid_t, pid, pid_nr(pid)) + ), + TPRAWFMT("task %d") + ); -TRACE_FORMAT(sched_process_fork, +TRACE_EVENT_FORMAT(sched_process_fork, TPPROTO(struct task_struct *parent, struct task_struct *child), TPARGS(parent, child), TPFMT("parent %s:%d child %s:%d", - parent->comm, parent->pid, child->comm, child->pid)); + parent->comm, parent->pid, child->comm, child->pid), + TRACE_STRUCT( + TRACE_FIELD(pid_t, parent, parent->pid) + TRACE_FIELD(pid_t, child, child->pid) + ), + TPRAWFMT("parent %d child %d") + ); -TRACE_FORMAT(sched_signal_send, +TRACE_EVENT_FORMAT(sched_signal_send, TPPROTO(int sig, struct task_struct *p), TPARGS(sig, p), - TPFMT("sig: %d task %s:%d", sig, p->comm, p->pid)); + TPFMT("sig: %d task %s:%d", sig, p->comm, p->pid), + TRACE_STRUCT( + TRACE_FIELD(int, sig, sig) + TRACE_FIELD(pid_t, pid, p->pid) + ), + TPRAWFMT("sig: %d task %d") + ); #undef TRACE_SYSTEM From f2034f1e1adaac6713a6d48b5a2d4f3aa3e63ccb Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 28 Feb 2009 02:54:39 -0500 Subject: [PATCH 208/413] tracing: create the C style tracing for the irq subsystem This patch utilizes the TRACE_EVENT_FORMAT macro to enable the C style faster tracing for the irq subsystem trace points. Signed-off-by: Steven Rostedt --- include/trace/irq_event_types.h | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) diff --git a/include/trace/irq_event_types.h b/include/trace/irq_event_types.h index 47a2be1b2d7d..65850bc5ea06 100644 --- a/include/trace/irq_event_types.h +++ b/include/trace/irq_event_types.h @@ -8,15 +8,26 @@ #undef TRACE_SYSTEM #define TRACE_SYSTEM irq -TRACE_FORMAT(irq_handler_entry, +TRACE_EVENT_FORMAT(irq_handler_entry, TPPROTO(int irq, struct irqaction *action), TPARGS(irq, action), - TPFMT("irq=%d handler=%s", irq, action->name)); + TPFMT("irq=%d handler=%s", irq, action->name), + TRACE_STRUCT( + TRACE_FIELD(int, irq, irq) + ), + TPRAWFMT("irq %d") + ); -TRACE_FORMAT(irq_handler_exit, +TRACE_EVENT_FORMAT(irq_handler_exit, TPPROTO(int irq, struct irqaction *action, int ret), TPARGS(irq, action, ret), TPFMT("irq=%d handler=%s return=%s", - irq, action->name, ret ? "handled" : "unhandled")); + irq, action->name, ret ? "handled" : "unhandled"), + TRACE_STRUCT( + TRACE_FIELD(int, irq, irq) + TRACE_FIELD(int, ret, ret) + ), + TPRAWFMT("irq %d ret %d") + ); #undef TRACE_SYSTEM From d20e3b03842bfeb9d21817ff19054c277cc3eac0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 10:53:15 -0500 Subject: [PATCH 209/413] tracing: add TRACE_FIELD_SPECIAL to record complex entries Tom Zanussi pointed out that the simple TRACE_FIELD was not enough to record trace data that required memcpy. This patch addresses this issue by adding a TRACE_FIELD_SPECIAL. The format is similar to TRACE_FIELD but looks like so: TRACE_FIELD_SPECIAL(type_item, item, cmd) What TRACE_FIELD gave was: TRACE_FIELD(type, item, assign) The TRACE_FIELD would be used in declaring a structure: struct { type item; }; And later assign it via: entry->item = assign; What TRACE_FIELD_SPECIAL gives us is: In the declaration of the structure: struct { type_item; }; And the assignment: cmd; This change log will explain the one example used in the patch: TRACE_EVENT_FORMAT(sched_switch, TPPROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next), TPARGS(rq, prev, next), TPFMT("task %s:%d ==> %s:%d", prev->comm, prev->pid, next->comm, next->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, prev_pid, prev->pid) TRACE_FIELD(int, prev_prio, prev->prio) TRACE_FIELD_SPECIAL(char next_comm[TASK_COMM_LEN], next_comm, TPCMD(memcpy(TRACE_ENTRY->next_comm, next->comm, TASK_COMM_LEN))) TRACE_FIELD(pid_t, next_pid, next->pid) TRACE_FIELD(int, next_prio, next->prio) ), TPRAWFMT("prev %d:%d ==> next %s:%d:%d") ); The struct will be create as: struct { pid_t prev_pid; int prev_prio; char next_comm[TASK_COMM_LEN]; pid_t next_pid; int next_prio; }; Note the TRACE_ENTRY in the cmd part of TRACE_SPECIAL. TRACE_ENTRY will be set by the tracer to point to the structure inside the trace buffer. entry->prev_pid = prev->pid; entry->prev_prio = prev->prio; memcpy(entry->next_comm, next->comm, TASK_COMM_LEN); entry->next_pid = next->pid; entry->next_prio = next->prio Reported-by: Tom Zanussi Signed-off-by: Steven Rostedt --- include/trace/sched_event_types.h | 7 ++++++- kernel/trace/trace_events_stage_1.h | 2 ++ kernel/trace/trace_events_stage_2.h | 4 ++++ kernel/trace/trace_events_stage_3.h | 14 ++++++++++++++ 4 files changed, 26 insertions(+), 1 deletion(-) diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index ba059c10b58a..a6de5c1601a0 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -71,10 +71,15 @@ TRACE_EVENT_FORMAT(sched_switch, TRACE_STRUCT( TRACE_FIELD(pid_t, prev_pid, prev->pid) TRACE_FIELD(int, prev_prio, prev->prio) + TRACE_FIELD_SPECIAL(char next_comm[TASK_COMM_LEN], + next_comm, + TPCMD(memcpy(TRACE_ENTRY->next_comm, + next->comm, + TASK_COMM_LEN))) TRACE_FIELD(pid_t, next_pid, next->pid) TRACE_FIELD(int, next_prio, next->prio) ), - TPRAWFMT("prev %d:%d ==> next %d:%d") + TPRAWFMT("prev %d:%d ==> next %s:%d:%d") ); TRACE_EVENT_FORMAT(sched_migrate_task, diff --git a/kernel/trace/trace_events_stage_1.h b/kernel/trace/trace_events_stage_1.h index fd3bf9382d37..3830a731424c 100644 --- a/kernel/trace/trace_events_stage_1.h +++ b/kernel/trace/trace_events_stage_1.h @@ -30,5 +30,7 @@ #define TRACE_FIELD(type, item, assign) \ type item; +#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ + type_item; #include diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index 3eaaef5f19e1..dc79fe3a2ecb 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -39,6 +39,10 @@ #define TRACE_FIELD(type, item, assign) \ field->item, +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ + field->item, + #undef TPRAWFMT #define TPRAWFMT(args...) args diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 7a161c49deb4..2ab65e958223 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -147,6 +147,20 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ #define TRACE_FIELD(type, item, assign)\ entry->item = assign; +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign)\ + entry->item = assign; + +#undef TPCMD +#define TPCMD(cmd...) cmd + +#undef TRACE_ENTRY +#define TRACE_ENTRY entry + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ + cmd; + #undef TRACE_EVENT_FORMAT #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ _TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) \ From 11a241a3302277db05561e01477528629d806c4e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 11:49:04 -0500 Subject: [PATCH 210/413] tracing: add protection around modify trace event fields The trace event objects are currently not proctected against reentrancy. This patch adds a mutex around the modifications of the trace event fields. Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 1d07f800a9ce..26069fa6b3b0 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -14,6 +14,8 @@ #define TRACE_SYSTEM "TRACE_SYSTEM" +static DEFINE_MUTEX(event_mutex); + #define events_for_each(event) \ for (event = __start_ftrace_events; \ (unsigned long)event < (unsigned long)__stop_ftrace_events; \ @@ -104,6 +106,7 @@ static int ftrace_set_clr_event(char *buf, int set) event = NULL; } + mutex_lock(&event_mutex); events_for_each(call) { if (!call->name) @@ -124,6 +127,8 @@ static int ftrace_set_clr_event(char *buf, int set) ret = 0; } + mutex_unlock(&event_mutex); + return ret; } @@ -324,7 +329,9 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, switch (val) { case 0: case 1: + mutex_lock(&event_mutex); ftrace_event_enable_disable(call, val); + mutex_unlock(&event_mutex); break; default: From f9520750c4c9924c14325cd951efae5fae58104c Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 14:04:40 -0500 Subject: [PATCH 211/413] tracing: make trace_seq_reset global and rename to trace_seq_init Impact: clean up The trace_seq functions may be used separately outside of the ftrace iterator. The trace_seq_reset is needed for these operations. This patch also renames trace_seq_reset to the more appropriate trace_seq_init. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 17 +++++------------ kernel/trace/trace.h | 8 ++++++++ 2 files changed, 13 insertions(+), 12 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c5e39cd7310d..ea055aa21cd9 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -342,13 +342,6 @@ __update_max_tr(struct trace_array *tr, struct task_struct *tsk, int cpu) tracing_record_cmdline(tsk); } -static void -trace_seq_reset(struct trace_seq *s) -{ - s->len = 0; - s->readpos = 0; -} - ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) { int len; @@ -395,7 +388,7 @@ trace_print_seq(struct seq_file *m, struct trace_seq *s) s->buffer[len] = 0; seq_puts(m, s->buffer); - trace_seq_reset(s); + trace_seq_init(s); } /** @@ -2620,7 +2613,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, if (sret != -EBUSY) return sret; - trace_seq_reset(&iter->seq); + trace_seq_init(&iter->seq); /* copy the tracer to avoid using a global lock all around */ mutex_lock(&trace_types_lock); @@ -2682,7 +2675,7 @@ tracing_read_pipe(struct file *filp, char __user *ubuf, /* Now copy what we have to the user */ sret = trace_seq_to_user(&iter->seq, ubuf, cnt); if (iter->seq.readpos >= iter->seq.len) - trace_seq_reset(&iter->seq); + trace_seq_init(&iter->seq); /* * If there was nothing to send to user, inspite of consuming trace @@ -2819,7 +2812,7 @@ static ssize_t tracing_splice_read_pipe(struct file *filp, partial[i].offset = 0; partial[i].len = iter->seq.len; - trace_seq_reset(&iter->seq); + trace_seq_init(&iter->seq); } mutex_unlock(&iter->mutex); @@ -3631,7 +3624,7 @@ trace_printk_seq(struct trace_seq *s) printk(KERN_TRACE "%s", s->buffer); - trace_seq_reset(s); + trace_seq_init(s); } void ftrace_dump(void) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index f6fa0b9f83a8..cf6ba4181b14 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -395,6 +395,14 @@ struct trace_seq { unsigned int readpos; }; +static inline void +trace_seq_init(struct trace_seq *s) +{ + s->len = 0; + s->readpos = 0; +} + + #define TRACE_PIPE_ALL_CPU -1 /* From 981d081ec8b958b7d962ee40d433581a55d40fc5 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 13:53:59 -0500 Subject: [PATCH 212/413] tracing: add format file to describe event struct fields This patch adds the "format" file to the trace point event directory. This is based off of work by Tom Zanussi, in which a file is exported to be tread from user land such that a user space app may read the binary record stored in the ring buffer. # cat /debug/tracing/events/sched/sched_switch/format field:pid_t prev_pid; offset:12; size:4; field:int prev_prio; offset:16; size:4; field special:char next_comm[TASK_COMM_LEN]; offset:20; size:16; field:pid_t next_pid; offset:36; size:4; field:int next_prio; offset:40; size:4; Idea-from: Tom Zanussi Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 1 + kernel/trace/trace_events.c | 56 ++++++++++++++++++++++++++++- kernel/trace/trace_events_stage_2.h | 52 +++++++++++++++++++++++++++ kernel/trace/trace_events_stage_3.h | 2 ++ 4 files changed, 110 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index cf6ba4181b14..e606633fb498 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -754,6 +754,7 @@ struct ftrace_event_call { int (*raw_init)(void); int (*raw_reg)(void); void (*raw_unreg)(void); + int (*show_format)(struct trace_seq *s); }; void event_trace_printk(unsigned long ip, const char *fmt, ...); diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 26069fa6b3b0..d57a772981c1 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -3,6 +3,9 @@ * * Copyright (C) 2008 Red Hat Inc, Steven Rostedt * + * - Added format output of fields of the trace point. + * This was based off of work by Tom Zanussi . + * */ #include @@ -444,6 +447,42 @@ event_available_types_read(struct file *filp, char __user *ubuf, size_t cnt, return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); } +static ssize_t +event_format_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + struct trace_seq *s; + char *buf; + int r; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + + if (*ppos) + return 0; + + r = call->show_format(s); + if (!r) { + /* + * ug! The format output is bigger than a PAGE!! + */ + buf = "FORMAT TOO BIG\n"; + r = simple_read_from_buffer(ubuf, cnt, ppos, + buf, strlen(buf)); + goto out; + } + + r = simple_read_from_buffer(ubuf, cnt, ppos, + s->buffer, s->len); + out: + kfree(s); + return r; +} + static const struct seq_operations show_event_seq_ops = { .start = t_start, .next = t_next, @@ -490,6 +529,11 @@ static const struct file_operations ftrace_available_types_fops = { .read = event_available_types_read, }; +static const struct file_operations ftrace_event_format_fops = { + .open = tracing_open_generic, + .read = event_format_read, +}; + static struct dentry *event_trace_events_dir(void) { static struct dentry *d_tracer; @@ -602,7 +646,17 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) &ftrace_available_types_fops); if (!entry) pr_warning("Could not create debugfs " - "'%s/type' available_types\n", call->name); + "'%s/available_types' entry\n", call->name); + + /* A trace may not want to export its format */ + if (!call->show_format) + return 0; + + entry = debugfs_create_file("format", 0444, call->dir, call, + &ftrace_event_format_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/format' entry\n", call->name); return 0; } diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index dc79fe3a2ecb..3a80ea4e92cb 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -74,3 +74,55 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ } #include + +/* + * Setup the showing format of trace point. + * + * int + * ftrace_format_##call(struct trace_seq *s) + * { + * struct ftrace_raw_##call field; + * int ret; + * + * ret = trace_seq_printf(s, #type " " #item ";" + * " size:%d; offset:%d;\n", + * sizeof(field.type), + * offsetof(struct ftrace_raw_##call, + * item)); + * + * } + */ + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign) \ + ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ + "offset:%lu;\tsize:%lu;\n", \ + offsetof(typeof(field), item), \ + sizeof(field.item)); \ + if (!ret) \ + return 0; + + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ + ret = trace_seq_printf(s, "\tfield special:" #type_item ";\t" \ + "offset:%lu;\tsize:%lu;\n", \ + offsetof(typeof(field), item), \ + sizeof(field.item)); \ + if (!ret) \ + return 0; + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +int \ +ftrace_format_##call(struct trace_seq *s) \ +{ \ + struct ftrace_raw_##call field; \ + int ret; \ + \ + tstruct; \ + \ + return ret; \ +} + +#include diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 2ab65e958223..c62a4d2a5283 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -101,6 +101,7 @@ * .raw_init = ftrace_raw_init_event_, * .raw_reg = ftrace_raw_reg_event_, * .raw_unreg = ftrace_raw_unreg_event_, + * .show_format = ftrace_format_, * } * */ @@ -230,4 +231,5 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .raw_init = ftrace_raw_init_event_##call, \ .raw_reg = ftrace_raw_reg_event_##call, \ .raw_unreg = ftrace_raw_unreg_event_##call, \ + .show_format = ftrace_format_##call, \ } From 91729ef96661bfa7dc53923746cd90b62d5495cc Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 15:03:01 -0500 Subject: [PATCH 213/413] tracing: add ftrace headers to event format files This patch includes the ftrace header to the event formats files: # cat /debug/tracing/events/sched/sched_switch/format field:unsigned char type; offset:0; size:1; field:unsigned char flags; offset:1; size:1; field:unsigned char preempt_count; offset:2; size:1; field:int pid; offset:4; size:4; field:int tgid; offset:8; size:4; field:pid_t prev_pid; offset:12; size:4; field:int prev_prio; offset:16; size:4; field special:char next_comm[TASK_COMM_LEN]; offset:20; size:16; field:pid_t next_pid; offset:36; size:4; field:int next_prio; offset:40; size:4; A blank line is used as a deliminator between the ftrace header and the trace point fields. Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index d57a772981c1..cdcc3aed76fd 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -13,7 +13,7 @@ #include #include -#include "trace.h" +#include "trace_output.h" #define TRACE_SYSTEM "TRACE_SYSTEM" @@ -447,6 +447,28 @@ event_available_types_read(struct file *filp, char __user *ubuf, size_t cnt, return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); } +#undef FIELD +#define FIELD(type, name) \ + #type, #name, offsetof(typeof(field), name), sizeof(field.name) + +static int trace_write_header(struct trace_seq *s) +{ + struct trace_entry field; + + /* struct trace_entry */ + return trace_seq_printf(s, + "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" + "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" + "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" + "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" + "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" + "\n", + FIELD(unsigned char, type), + FIELD(unsigned char, flags), + FIELD(unsigned char, preempt_count), + FIELD(int, pid), + FIELD(int, tgid)); +} static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -465,6 +487,9 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, if (*ppos) return 0; + /* If this fails, so will the show_format. */ + trace_write_header(s); + r = call->show_format(s); if (!r) { /* From c5e4e19271edfdf1abd4184933d40d646da6a091 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 15:10:02 -0500 Subject: [PATCH 214/413] tracing: add trace name and id to event formats To be able to identify the trace in the binary format output, the id of the trace event (which is dynamically assigned) must also be listed. This patch adds the name of the trace point as well as the id assigned. Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index cdcc3aed76fd..210e71ff82db 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -487,7 +487,11 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, if (*ppos) return 0; - /* If this fails, so will the show_format. */ + /* If any of the first writes fail, so will the show_format. */ + + trace_seq_printf(s, "name: %s\n", call->name); + trace_seq_printf(s, "ID: %d\n", call->id); + trace_seq_printf(s, "format:\n"); trace_write_header(s); r = call->show_format(s); From 96ccd21cd13140221bda74a4fc4e53ffeba7c7d4 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 15:22:21 -0500 Subject: [PATCH 215/413] tracing: add print format to event trace format files This patch adds the internal print format used to print the raw events to the event trace point format file. # cat /debug/tracing/events/sched/sched_switch/format name: sched_switch ID: 29 format: field:unsigned char type; offset:0; size:1; field:unsigned char flags; offset:1; size:1; field:unsigned char preempt_count; offset:2; size:1; field:int pid; offset:4; size:4; field:int tgid; offset:8; size:4; field:pid_t prev_pid; offset:12; size:4; field:int prev_prio; offset:16; size:4; field special:char next_comm[TASK_COMM_LEN]; offset:20; size:16; field:pid_t next_pid; offset:36; size:4; field:int next_prio; offset:40; size:4; print fmt: "prev %d:%d ==> next %s:%d:%d" Signed-off-by: Steven Rostedt --- kernel/trace/trace_events_stage_2.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index 3a80ea4e92cb..b1cebba1d9b4 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -122,6 +122,8 @@ ftrace_format_##call(struct trace_seq *s) \ \ tstruct; \ \ + trace_seq_printf(s, "\nprint fmt: \"%s\"\n", tpfmt); \ + \ return ret; \ } From c79a61f55773d2519fd0525bf58385f7d20752d3 Mon Sep 17 00:00:00 2001 From: Uwe Kleine-Koenig Date: Fri, 27 Feb 2009 21:30:03 +0100 Subject: [PATCH 216/413] tracing: make CALLER_ADDRx overwriteable The current definition of CALLER_ADDRx isn't suitable for all platforms. E.g. for ARM __builtin_return_address(N) doesn't work for N > 0 and AFAIK for powerpc there are no frame pointers needed to have a working __builtin_return_address. This patch allows defining the CALLER_ADDRx macros in and let these take precedence. Because now is included unconditionally in all archs that don't already had this include get an empty one for free. Signed-off-by: Uwe Kleine-Koenig Cc: Peter Zijlstra Cc: Ingo Molnar Reviewed-by: KOSAKI Motohiro Signed-off-by: Steven Rostedt --- arch/alpha/include/asm/ftrace.h | 1 + arch/avr32/include/asm/ftrace.h | 1 + arch/blackfin/include/asm/ftrace.h | 1 + arch/cris/include/asm/ftrace.h | 1 + arch/h8300/include/asm/ftrace.h | 1 + arch/m68k/include/asm/ftrace.h | 1 + arch/mips/include/asm/ftrace.h | 1 + arch/parisc/include/asm/ftrace.h | 1 + arch/um/include/asm/ftrace.h | 1 + arch/xtensa/include/asm/ftrace.h | 1 + include/asm-frv/ftrace.h | 1 + include/asm-m32r/ftrace.h | 1 + include/asm-mn10300/ftrace.h | 1 + include/linux/ftrace.h | 41 +++++++++++++++--------------- 14 files changed, 34 insertions(+), 20 deletions(-) create mode 100644 arch/alpha/include/asm/ftrace.h create mode 100644 arch/avr32/include/asm/ftrace.h create mode 100644 arch/blackfin/include/asm/ftrace.h create mode 100644 arch/cris/include/asm/ftrace.h create mode 100644 arch/h8300/include/asm/ftrace.h create mode 100644 arch/m68k/include/asm/ftrace.h create mode 100644 arch/mips/include/asm/ftrace.h create mode 100644 arch/parisc/include/asm/ftrace.h create mode 100644 arch/um/include/asm/ftrace.h create mode 100644 arch/xtensa/include/asm/ftrace.h create mode 100644 include/asm-frv/ftrace.h create mode 100644 include/asm-m32r/ftrace.h create mode 100644 include/asm-mn10300/ftrace.h diff --git a/arch/alpha/include/asm/ftrace.h b/arch/alpha/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/alpha/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/avr32/include/asm/ftrace.h b/arch/avr32/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/avr32/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/blackfin/include/asm/ftrace.h b/arch/blackfin/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/blackfin/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/cris/include/asm/ftrace.h b/arch/cris/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/cris/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/h8300/include/asm/ftrace.h b/arch/h8300/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/h8300/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/m68k/include/asm/ftrace.h b/arch/m68k/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/m68k/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/mips/include/asm/ftrace.h b/arch/mips/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/mips/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/parisc/include/asm/ftrace.h b/arch/parisc/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/parisc/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/um/include/asm/ftrace.h b/arch/um/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/um/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/arch/xtensa/include/asm/ftrace.h b/arch/xtensa/include/asm/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/arch/xtensa/include/asm/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/include/asm-frv/ftrace.h b/include/asm-frv/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/include/asm-frv/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/include/asm-m32r/ftrace.h b/include/asm-m32r/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/include/asm-m32r/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/include/asm-mn10300/ftrace.h b/include/asm-mn10300/ftrace.h new file mode 100644 index 000000000000..40a8c178f10d --- /dev/null +++ b/include/asm-mn10300/ftrace.h @@ -0,0 +1 @@ +/* empty */ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 847bb3c48dd0..1f69ac7c1587 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -11,6 +11,8 @@ #include #include +#include + #ifdef CONFIG_FUNCTION_TRACER extern int ftrace_enabled; @@ -103,8 +105,6 @@ struct ftrace_func_command { }; #ifdef CONFIG_DYNAMIC_FTRACE -/* asm/ftrace.h must be defined for archs supporting dynamic ftrace */ -#include int ftrace_arch_code_modify_prepare(void); int ftrace_arch_code_modify_post_process(void); @@ -282,24 +282,25 @@ static inline void __ftrace_enabled_restore(int enabled) #endif } -#ifdef CONFIG_FRAME_POINTER -/* TODO: need to fix this for ARM */ -# define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0)) -# define CALLER_ADDR1 ((unsigned long)__builtin_return_address(1)) -# define CALLER_ADDR2 ((unsigned long)__builtin_return_address(2)) -# define CALLER_ADDR3 ((unsigned long)__builtin_return_address(3)) -# define CALLER_ADDR4 ((unsigned long)__builtin_return_address(4)) -# define CALLER_ADDR5 ((unsigned long)__builtin_return_address(5)) -# define CALLER_ADDR6 ((unsigned long)__builtin_return_address(6)) -#else -# define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0)) -# define CALLER_ADDR1 0UL -# define CALLER_ADDR2 0UL -# define CALLER_ADDR3 0UL -# define CALLER_ADDR4 0UL -# define CALLER_ADDR5 0UL -# define CALLER_ADDR6 0UL -#endif +#ifndef HAVE_ARCH_CALLER_ADDR +# ifdef CONFIG_FRAME_POINTER +# define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0)) +# define CALLER_ADDR1 ((unsigned long)__builtin_return_address(1)) +# define CALLER_ADDR2 ((unsigned long)__builtin_return_address(2)) +# define CALLER_ADDR3 ((unsigned long)__builtin_return_address(3)) +# define CALLER_ADDR4 ((unsigned long)__builtin_return_address(4)) +# define CALLER_ADDR5 ((unsigned long)__builtin_return_address(5)) +# define CALLER_ADDR6 ((unsigned long)__builtin_return_address(6)) +# else +# define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0)) +# define CALLER_ADDR1 0UL +# define CALLER_ADDR2 0UL +# define CALLER_ADDR3 0UL +# define CALLER_ADDR4 0UL +# define CALLER_ADDR5 0UL +# define CALLER_ADDR6 0UL +# endif +#endif /* ifndef HAVE_ARCH_CALLER_ADDR */ #ifdef CONFIG_IRQSOFF_TRACER extern void time_hardirqs_on(unsigned long a0, unsigned long a1); From 633ddaa7f471e9db181f993c1458d6f4bae321ca Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 3 Mar 2009 09:43:50 -0500 Subject: [PATCH 217/413] tracing: fix return value to registering events The registering of events had the return value check backwards. A zero returned is success, the check had it as a failure. This patch also fixes a missing "\n" in the warning that the check failed. Reported-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/trace_events_stage_3.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index c62a4d2a5283..041789ffbac1 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -120,9 +120,9 @@ static int ftrace_reg_event_##call(void) \ int ret; \ \ ret = register_trace_##call(ftrace_event_##call); \ - if (!ret) \ + if (ret) \ pr_info("event trace: Could not activate trace point " \ - "probe to " #call); \ + "probe to " #call "\n"); \ return ret; \ } \ \ @@ -195,9 +195,9 @@ static int ftrace_raw_reg_event_##call(void) \ int ret; \ \ ret = register_trace_##call(ftrace_raw_event_##call); \ - if (!ret) \ + if (ret) \ pr_info("event trace: Could not activate trace point " \ - "probe to " #call); \ + "probe to " #call "\n"); \ return ret; \ } \ \ From 41be4da4e85e58520b934040966a6ae919c66c2d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 2 Mar 2009 20:56:48 -0500 Subject: [PATCH 218/413] ring-buffer: reset write field for ring_buffer_read_page Impact: fix ring_buffer_read_page After a page is swapped into the ring buffer, the write field must also be reset. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index a8c275c01e83..9baad7ee4b36 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2492,6 +2492,7 @@ int ring_buffer_read_page(struct ring_buffer *buffer, rb_init_page(bpage); bpage = cpu_buffer->reader_page->page; cpu_buffer->reader_page->page = *data_page; + local_set(&cpu_buffer->reader_page->write, 0); cpu_buffer->reader_page->read = 0; *data_page = bpage; } From ef7a4a161472b952941bf78855a9cd95703c024e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 3 Mar 2009 00:27:49 -0500 Subject: [PATCH 219/413] ring-buffer: fix ring_buffer_read_page The ring_buffer_read_page was broken if it were to only copy part of the page. This patch fixes that up as well as adds a parameter to allow a length field, in order to only copy part of the buffer page. Signed-off-by: Steven Rostedt --- include/linux/ring_buffer.h | 7 ++- kernel/trace/ring_buffer.c | 92 ++++++++++++++++++++++++------------- 2 files changed, 64 insertions(+), 35 deletions(-) diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index f5e793d69bd3..79fcbc4b09d6 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -121,6 +121,9 @@ unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu); u64 ring_buffer_time_stamp(int cpu); void ring_buffer_normalize_time_stamp(int cpu, u64 *ts); +size_t ring_buffer_page_len(void *page); + + /* * The below functions are fine to use outside the tracing facility. */ @@ -138,8 +141,8 @@ static inline int tracing_is_on(void) { return 0; } void *ring_buffer_alloc_read_page(struct ring_buffer *buffer); void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data); -int ring_buffer_read_page(struct ring_buffer *buffer, - void **data_page, int cpu, int full); +int ring_buffer_read_page(struct ring_buffer *buffer, void **data_page, + size_t len, int cpu, int full); enum ring_buffer_flags { RB_FL_OVERWRITE = 1 << 0, diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 9baad7ee4b36..2ad6bae95a3d 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -234,6 +234,11 @@ static void rb_init_page(struct buffer_data_page *bpage) local_set(&bpage->commit, 0); } +size_t ring_buffer_page_len(void *page) +{ + return local_read(&((struct buffer_data_page *)page)->commit); +} + /* * Also stolen from mm/slob.c. Thanks to Mathieu Desnoyers for pointing * this issue out. @@ -2378,8 +2383,8 @@ static void rb_remove_entries(struct ring_buffer_per_cpu *cpu_buffer, */ void *ring_buffer_alloc_read_page(struct ring_buffer *buffer) { - unsigned long addr; struct buffer_data_page *bpage; + unsigned long addr; addr = __get_free_page(GFP_KERNEL); if (!addr) @@ -2387,6 +2392,8 @@ void *ring_buffer_alloc_read_page(struct ring_buffer *buffer) bpage = (void *)addr; + rb_init_page(bpage); + return bpage; } @@ -2406,6 +2413,7 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * ring_buffer_read_page - extract a page from the ring buffer * @buffer: buffer to extract from * @data_page: the page to use allocated from ring_buffer_alloc_read_page + * @len: amount to extract * @cpu: the cpu of the buffer to extract * @full: should the extraction only happen when the page is full. * @@ -2418,7 +2426,7 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * rpage = ring_buffer_alloc_read_page(buffer); * if (!rpage) * return error; - * ret = ring_buffer_read_page(buffer, &rpage, cpu, 0); + * ret = ring_buffer_read_page(buffer, &rpage, len, cpu, 0); * if (ret >= 0) * process_page(rpage, ret); * @@ -2435,71 +2443,89 @@ void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data) * <0 if no data has been transferred. */ int ring_buffer_read_page(struct ring_buffer *buffer, - void **data_page, int cpu, int full) + void **data_page, size_t len, int cpu, int full) { struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; struct ring_buffer_event *event; struct buffer_data_page *bpage; + struct buffer_page *reader; unsigned long flags; + unsigned int commit; unsigned int read; int ret = -1; if (!data_page) - return 0; + return -1; bpage = *data_page; if (!bpage) - return 0; + return -1; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); - /* - * rb_buffer_peek will get the next ring buffer if - * the current reader page is empty. - */ - event = rb_buffer_peek(buffer, cpu, NULL); - if (!event) + reader = rb_get_reader_page(cpu_buffer); + if (!reader) goto out; - /* check for data */ - if (!local_read(&cpu_buffer->reader_page->page->commit)) - goto out; + event = rb_reader_event(cpu_buffer); + + read = reader->read; + commit = rb_page_commit(reader); - read = cpu_buffer->reader_page->read; /* - * If the writer is already off of the read page, then simply - * switch the read page with the given page. Otherwise - * we need to copy the data from the reader to the writer. + * If len > what's left on the page, and the writer is also off of + * the read page, then simply switch the read page with the given + * page. Otherwise we need to copy the data from the reader to the + * writer. */ - if (cpu_buffer->reader_page == cpu_buffer->commit_page) { - unsigned int commit = rb_page_commit(cpu_buffer->reader_page); + if ((len < (commit - read)) || + cpu_buffer->reader_page == cpu_buffer->commit_page) { struct buffer_data_page *rpage = cpu_buffer->reader_page->page; + unsigned int pos = read; + unsigned int size; if (full) goto out; - /* The writer is still on the reader page, we must copy */ - memcpy(bpage->data + read, rpage->data + read, commit - read); - /* consume what was read */ - cpu_buffer->reader_page->read = commit; + if (len > (commit - read)) + len = (commit - read); + + size = rb_event_length(event); + + if (len < size) + goto out; + + /* Need to copy one event at a time */ + do { + memcpy(bpage->data + pos, rpage->data + pos, size); + + len -= size; + + rb_advance_reader(cpu_buffer); + pos = reader->read; + + event = rb_reader_event(cpu_buffer); + size = rb_event_length(event); + } while (len > size); /* update bpage */ - local_set(&bpage->commit, commit); - if (!read) - bpage->time_stamp = rpage->time_stamp; + local_set(&bpage->commit, pos); + bpage->time_stamp = rpage->time_stamp; + } else { /* swap the pages */ rb_init_page(bpage); - bpage = cpu_buffer->reader_page->page; - cpu_buffer->reader_page->page = *data_page; - local_set(&cpu_buffer->reader_page->write, 0); - cpu_buffer->reader_page->read = 0; + bpage = reader->page; + reader->page = *data_page; + local_set(&reader->write, 0); + reader->read = 0; *data_page = bpage; + + /* update the entry counter */ + rb_remove_entries(cpu_buffer, bpage, read); } ret = read; - /* update the entry counter */ - rb_remove_entries(cpu_buffer, bpage, read); out: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); From e3d6bf0a0781a269f34250fd41e0d3dbfe540cf1 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 3 Mar 2009 13:53:07 -0500 Subject: [PATCH 220/413] ring-buffer: replace sizeof of event header with offsetof Impact: fix to possible alignment problems on some archs. Some arch compilers include an NULL char array in the sizeof field. Since the ring_buffer_event type includes one of these, it is better to use the "offsetof" instead, to avoid strange bugs on these archs. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 2ad6bae95a3d..27cf834d8b4e 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -132,7 +132,7 @@ void ring_buffer_normalize_time_stamp(int cpu, u64 *ts) } EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp); -#define RB_EVNT_HDR_SIZE (sizeof(struct ring_buffer_event)) +#define RB_EVNT_HDR_SIZE (offsetof(struct ring_buffer_event, array)) #define RB_ALIGNMENT 4U #define RB_MAX_SMALL_DATA 28 From 474d32b68d6d842f3e710e9ae9fe2568c53339f8 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 3 Mar 2009 19:51:40 -0500 Subject: [PATCH 221/413] ring-buffer: make ring_buffer_read_page read from start on partial page Impact: dont leave holes in read buffer page The ring_buffer_read_page swaps a given page with the reader page of the ring buffer, if certain conditions are set: 1) requested length is big enough to hold entire page data 2) a writer is not currently on the page 3) the page is not partially consumed. Instead of swapping with the supplied page. It copies the data to the supplied page instead. But currently the data is copied in the same offset as the source page. This causes a hole at the start of the reader page. This complicates the use of this function. Instead, it should copy the data at the beginning of the function and update the index fields accordingly. Other small clean ups are also done in this patch. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 43 +++++++++++++++++++++++++++++--------- 1 file changed, 33 insertions(+), 10 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 27cf834d8b4e..f2a163db52f9 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -61,6 +61,8 @@ enum { static unsigned long ring_buffer_flags __read_mostly = RB_BUFFERS_ON; +#define BUF_PAGE_HDR_SIZE offsetof(struct buffer_data_page, data) + /** * tracing_on - enable all tracing buffers * @@ -234,9 +236,16 @@ static void rb_init_page(struct buffer_data_page *bpage) local_set(&bpage->commit, 0); } +/** + * ring_buffer_page_len - the size of data on the page. + * @page: The page to read + * + * Returns the amount of data on the page, including buffer page header. + */ size_t ring_buffer_page_len(void *page) { - return local_read(&((struct buffer_data_page *)page)->commit); + return local_read(&((struct buffer_data_page *)page)->commit) + + BUF_PAGE_HDR_SIZE; } /* @@ -259,7 +268,7 @@ static inline int test_time_stamp(u64 delta) return 0; } -#define BUF_PAGE_SIZE (PAGE_SIZE - offsetof(struct buffer_data_page, data)) +#define BUF_PAGE_SIZE (PAGE_SIZE - BUF_PAGE_HDR_SIZE) /* * head_page == tail_page && head == tail then buffer is empty. @@ -2454,6 +2463,15 @@ int ring_buffer_read_page(struct ring_buffer *buffer, unsigned int read; int ret = -1; + /* + * If len is not big enough to hold the page header, then + * we can not copy anything. + */ + if (len <= BUF_PAGE_HDR_SIZE) + return -1; + + len -= BUF_PAGE_HDR_SIZE; + if (!data_page) return -1; @@ -2473,15 +2491,17 @@ int ring_buffer_read_page(struct ring_buffer *buffer, commit = rb_page_commit(reader); /* - * If len > what's left on the page, and the writer is also off of - * the read page, then simply switch the read page with the given - * page. Otherwise we need to copy the data from the reader to the - * writer. + * If this page has been partially read or + * if len is not big enough to read the rest of the page or + * a writer is still on the page, then + * we must copy the data from the page to the buffer. + * Otherwise, we can simply swap the page with the one passed in. */ - if ((len < (commit - read)) || + if (read || (len < (commit - read)) || cpu_buffer->reader_page == cpu_buffer->commit_page) { struct buffer_data_page *rpage = cpu_buffer->reader_page->page; - unsigned int pos = read; + unsigned int rpos = read; + unsigned int pos = 0; unsigned int size; if (full) @@ -2497,12 +2517,13 @@ int ring_buffer_read_page(struct ring_buffer *buffer, /* Need to copy one event at a time */ do { - memcpy(bpage->data + pos, rpage->data + pos, size); + memcpy(bpage->data + pos, rpage->data + rpos, size); len -= size; rb_advance_reader(cpu_buffer); - pos = reader->read; + rpos = reader->read; + pos += size; event = rb_reader_event(cpu_buffer); size = rb_event_length(event); @@ -2512,6 +2533,8 @@ int ring_buffer_read_page(struct ring_buffer *buffer, local_set(&bpage->commit, pos); bpage->time_stamp = rpage->time_stamp; + /* we copied everything to the beginning */ + read = 0; } else { /* swap the pages */ rb_init_page(bpage); From 2cadf9135eb3b6d84b6427314be827ddd443c308 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 1 Dec 2008 22:20:19 -0500 Subject: [PATCH 222/413] tracing: add binary buffer files for use with splice Impact: new feature This patch creates a directory of files that correspond to the per CPU ring buffers. These are binary files and are made to be used with splice. This is the fastest way to extract data from the ftrace ring buffers. Thanks to Jiaying Zhang for pushing me to get this code fixed, and to Eduard - Gabriel Munteanu for his splice code that helped me debug my code. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 274 +++++++++++++++++++++++++++++++++++++++++-- kernel/trace/trace.h | 1 + 2 files changed, 268 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index ea055aa21cd9..12539f72f4a5 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -11,31 +11,30 @@ * Copyright (C) 2004-2006 Ingo Molnar * Copyright (C) 2004 William Lee Irwin III */ +#include #include +#include +#include #include #include #include +#include #include #include #include #include #include +#include #include #include #include +#include #include #include #include #include #include #include -#include -#include -#include - -#include -#include -#include #include "trace.h" #include "trace_output.h" @@ -3005,6 +3004,246 @@ static struct file_operations tracing_mark_fops = { .write = tracing_mark_write, }; +struct ftrace_buffer_info { + struct trace_array *tr; + void *spare; + int cpu; + unsigned int read; +}; + +static int tracing_buffers_open(struct inode *inode, struct file *filp) +{ + int cpu = (int)(long)inode->i_private; + struct ftrace_buffer_info *info; + + if (tracing_disabled) + return -ENODEV; + + info = kzalloc(sizeof(*info), GFP_KERNEL); + if (!info) + return -ENOMEM; + + info->tr = &global_trace; + info->cpu = cpu; + info->spare = ring_buffer_alloc_read_page(info->tr->buffer); + /* Force reading ring buffer for first read */ + info->read = (unsigned int)-1; + if (!info->spare) + goto out; + + filp->private_data = info; + + return 0; + + out: + kfree(info); + return -ENOMEM; +} + +static ssize_t +tracing_buffers_read(struct file *filp, char __user *ubuf, + size_t count, loff_t *ppos) +{ + struct ftrace_buffer_info *info = filp->private_data; + unsigned int pos; + ssize_t ret; + size_t size; + + /* Do we have previous read data to read? */ + if (info->read < PAGE_SIZE) + goto read; + + info->read = 0; + + ret = ring_buffer_read_page(info->tr->buffer, + &info->spare, + count, + info->cpu, 0); + if (ret < 0) + return 0; + + pos = ring_buffer_page_len(info->spare); + + if (pos < PAGE_SIZE) + memset(info->spare + pos, 0, PAGE_SIZE - pos); + +read: + size = PAGE_SIZE - info->read; + if (size > count) + size = count; + + ret = copy_to_user(ubuf, info->spare + info->read, size); + if (ret) + return -EFAULT; + *ppos += size; + info->read += size; + + return size; +} + +static int tracing_buffers_release(struct inode *inode, struct file *file) +{ + struct ftrace_buffer_info *info = file->private_data; + + ring_buffer_free_read_page(info->tr->buffer, info->spare); + kfree(info); + + return 0; +} + +struct buffer_ref { + struct ring_buffer *buffer; + void *page; + int ref; +}; + +static void buffer_pipe_buf_release(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + struct buffer_ref *ref = (struct buffer_ref *)buf->private; + + if (--ref->ref) + return; + + ring_buffer_free_read_page(ref->buffer, ref->page); + kfree(ref); + buf->private = 0; +} + +static int buffer_pipe_buf_steal(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + return 1; +} + +static void buffer_pipe_buf_get(struct pipe_inode_info *pipe, + struct pipe_buffer *buf) +{ + struct buffer_ref *ref = (struct buffer_ref *)buf->private; + + ref->ref++; +} + +/* Pipe buffer operations for a buffer. */ +static struct pipe_buf_operations buffer_pipe_buf_ops = { + .can_merge = 0, + .map = generic_pipe_buf_map, + .unmap = generic_pipe_buf_unmap, + .confirm = generic_pipe_buf_confirm, + .release = buffer_pipe_buf_release, + .steal = buffer_pipe_buf_steal, + .get = buffer_pipe_buf_get, +}; + +/* + * Callback from splice_to_pipe(), if we need to release some pages + * at the end of the spd in case we error'ed out in filling the pipe. + */ +static void buffer_spd_release(struct splice_pipe_desc *spd, unsigned int i) +{ + struct buffer_ref *ref = + (struct buffer_ref *)spd->partial[i].private; + + if (--ref->ref) + return; + + ring_buffer_free_read_page(ref->buffer, ref->page); + kfree(ref); + spd->partial[i].private = 0; +} + +static ssize_t +tracing_buffers_splice_read(struct file *file, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags) +{ + struct ftrace_buffer_info *info = file->private_data; + struct partial_page partial[PIPE_BUFFERS]; + struct page *pages[PIPE_BUFFERS]; + struct splice_pipe_desc spd = { + .pages = pages, + .partial = partial, + .flags = flags, + .ops = &buffer_pipe_buf_ops, + .spd_release = buffer_spd_release, + }; + struct buffer_ref *ref; + int size, i; + size_t ret; + + /* + * We can't seek on a buffer input + */ + if (unlikely(*ppos)) + return -ESPIPE; + + + for (i = 0; i < PIPE_BUFFERS && len; i++, len -= size) { + struct page *page; + int r; + + ref = kzalloc(sizeof(*ref), GFP_KERNEL); + if (!ref) + break; + + ref->buffer = info->tr->buffer; + ref->page = ring_buffer_alloc_read_page(ref->buffer); + if (!ref->page) { + kfree(ref); + break; + } + + r = ring_buffer_read_page(ref->buffer, &ref->page, + len, info->cpu, 0); + if (r < 0) { + ring_buffer_free_read_page(ref->buffer, + ref->page); + kfree(ref); + break; + } + + /* + * zero out any left over data, this is going to + * user land. + */ + size = ring_buffer_page_len(ref->page); + if (size < PAGE_SIZE) + memset(ref->page + size, 0, PAGE_SIZE - size); + + page = virt_to_page(ref->page); + + spd.pages[i] = page; + spd.partial[i].len = PAGE_SIZE; + spd.partial[i].offset = 0; + spd.partial[i].private = (unsigned long)ref; + spd.nr_pages++; + } + + spd.nr_pages = i; + + /* did we read anything? */ + if (!spd.nr_pages) { + if (flags & SPLICE_F_NONBLOCK) + ret = -EAGAIN; + else + ret = 0; + /* TODO: block */ + return ret; + } + + ret = splice_to_pipe(pipe, &spd); + + return ret; +} + +static const struct file_operations tracing_buffers_fops = { + .open = tracing_buffers_open, + .read = tracing_buffers_read, + .release = tracing_buffers_release, + .splice_read = tracing_buffers_splice_read, + .llseek = no_llseek, +}; + #ifdef CONFIG_DYNAMIC_FTRACE int __weak ftrace_arch_read_dyn_info(char *buf, int size) @@ -3399,6 +3638,7 @@ static __init void create_trace_options_dir(void) static __init int tracer_init_debugfs(void) { struct dentry *d_tracer; + struct dentry *buffers; struct dentry *entry; int cpu; @@ -3471,6 +3711,26 @@ static __init int tracer_init_debugfs(void) pr_warning("Could not create debugfs " "'trace_marker' entry\n"); + buffers = debugfs_create_dir("binary_buffers", d_tracer); + + if (!buffers) + pr_warning("Could not create buffers directory\n"); + else { + int cpu; + char buf[64]; + + for_each_tracing_cpu(cpu) { + sprintf(buf, "%d", cpu); + + entry = debugfs_create_file(buf, 0444, buffers, + (void *)(long)cpu, + &tracing_buffers_fops); + if (!entry) + pr_warning("Could not create debugfs buffers " + "'%s' entry\n", buf); + } + } + #ifdef CONFIG_DYNAMIC_FTRACE entry = debugfs_create_file("dyn_ftrace_total_info", 0444, d_tracer, &ftrace_update_tot_cnt, diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index e606633fb498..561bb5c5d988 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -217,6 +217,7 @@ enum trace_flag_type { */ struct trace_array_cpu { atomic_t disabled; + void *buffer_page; /* ring buffer spare */ /* these fields get copied into max-trace: */ unsigned long trace_idx; From efed792d6738964f399a508ef9e831cd60fa4657 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Wed, 4 Mar 2009 12:32:55 +0100 Subject: [PATCH 223/413] tracing: add lockdep tracepoints for lock acquire/release Augment the traces with lock names when lockdep is available: 1) | down_read_trylock() { 1) | _spin_lock_irqsave() { 1) | /* lock_acquire: &sem->wait_lock */ 1) 4.201 us | } 1) | _spin_unlock_irqrestore() { 1) | /* lock_release: &sem->wait_lock */ 1) 3.523 us | } 1) | /* lock_acquire: try read &mm->mmap_sem */ 1) + 13.386 us | } 1) 1.635 us | find_vma(); 1) | handle_mm_fault() { 1) | __do_fault() { 1) | filemap_fault() { 1) | find_lock_page() { 1) | find_get_page() { 1) | /* lock_acquire: read rcu_read_lock */ 1) | /* lock_release: rcu_read_lock */ 1) 5.697 us | } 1) 8.158 us | } 1) + 11.079 us | } 1) | _spin_lock() { 1) | /* lock_acquire: __pte_lockptr(page) */ 1) 3.949 us | } 1) 1.460 us | page_add_file_rmap(); 1) | _spin_unlock() { 1) | /* lock_release: __pte_lockptr(page) */ 1) 3.115 us | } 1) | unlock_page() { 1) 1.421 us | page_waitqueue(); 1) 1.220 us | __wake_up_bit(); 1) 6.519 us | } 1) + 34.328 us | } 1) + 37.452 us | } 1) | up_read() { 1) | /* lock_release: &mm->mmap_sem */ 1) | _spin_lock_irqsave() { 1) | /* lock_acquire: &sem->wait_lock */ 1) 3.865 us | } 1) | _spin_unlock_irqrestore() { 1) | /* lock_release: &sem->wait_lock */ 1) 8.562 us | } 1) + 17.370 us | } Signed-off-by: Peter Zijlstra Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?T=F6r=F6k?= Edwin Cc: Jason Baron Cc: Frederic Weisbecker LKML-Reference: <1236166375.5330.7209.camel@laptop> Signed-off-by: Ingo Molnar --- include/trace/lockdep.h | 9 ++++++ include/trace/lockdep_event_types.h | 44 +++++++++++++++++++++++++++++ include/trace/trace_event_types.h | 1 + include/trace/trace_events.h | 1 + kernel/lockdep.c | 17 +++++++++++ kernel/trace/trace.c | 14 +++++---- kernel/trace/trace_events_stage_3.h | 4 +-- 7 files changed, 82 insertions(+), 8 deletions(-) create mode 100644 include/trace/lockdep.h create mode 100644 include/trace/lockdep_event_types.h diff --git a/include/trace/lockdep.h b/include/trace/lockdep.h new file mode 100644 index 000000000000..5ca67df87f2a --- /dev/null +++ b/include/trace/lockdep.h @@ -0,0 +1,9 @@ +#ifndef _TRACE_LOCKDEP_H +#define _TRACE_LOCKDEP_H + +#include +#include + +#include + +#endif diff --git a/include/trace/lockdep_event_types.h b/include/trace/lockdep_event_types.h new file mode 100644 index 000000000000..f713d74a82b4 --- /dev/null +++ b/include/trace/lockdep_event_types.h @@ -0,0 +1,44 @@ + +#ifndef TRACE_EVENT_FORMAT +# error Do not include this file directly. +# error Unless you know what you are doing. +#endif + +#undef TRACE_SYSTEM +#define TRACE_SYSTEM lock + +#ifdef CONFIG_LOCKDEP + +TRACE_FORMAT(lock_acquire, + TPPROTO(struct lockdep_map *lock, unsigned int subclass, + int trylock, int read, int check, + struct lockdep_map *next_lock, unsigned long ip), + TPARGS(lock, subclass, trylock, read, check, next_lock, ip), + TPFMT("%s%s%s", trylock ? "try " : "", + read ? "read " : "", lock->name) + ); + +TRACE_FORMAT(lock_release, + TPPROTO(struct lockdep_map *lock, int nested, unsigned long ip), + TPARGS(lock, nested, ip), + TPFMT("%s", lock->name) + ); + +#ifdef CONFIG_LOCK_STAT + +TRACE_FORMAT(lock_contended, + TPPROTO(struct lockdep_map *lock, unsigned long ip), + TPARGS(lock, ip), + TPFMT("%s", lock->name) + ); + +TRACE_FORMAT(lock_acquired, + TPPROTO(struct lockdep_map *lock, unsigned long ip), + TPARGS(lock, ip), + TPFMT("%s", lock->name) + ); + +#endif +#endif + +#undef TRACE_SYSTEM diff --git a/include/trace/trace_event_types.h b/include/trace/trace_event_types.h index 33c8ed5ccb6c..df56f5694be6 100644 --- a/include/trace/trace_event_types.h +++ b/include/trace/trace_event_types.h @@ -2,3 +2,4 @@ #include #include +#include diff --git a/include/trace/trace_events.h b/include/trace/trace_events.h index ea2ef2051762..fd13750ca4ba 100644 --- a/include/trace/trace_events.h +++ b/include/trace/trace_events.h @@ -2,3 +2,4 @@ #include #include +#include diff --git a/kernel/lockdep.c b/kernel/lockdep.c index 02014f7ccc86..cb70c1db85d0 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -42,6 +42,7 @@ #include #include #include +#include #include @@ -2913,6 +2914,8 @@ void lock_set_class(struct lockdep_map *lock, const char *name, } EXPORT_SYMBOL_GPL(lock_set_class); +DEFINE_TRACE(lock_acquire); + /* * We are not always called with irqs disabled - do that here, * and also avoid lockdep recursion: @@ -2923,6 +2926,8 @@ void lock_acquire(struct lockdep_map *lock, unsigned int subclass, { unsigned long flags; + trace_lock_acquire(lock, subclass, trylock, read, check, nest_lock, ip); + if (unlikely(current->lockdep_recursion)) return; @@ -2937,11 +2942,15 @@ void lock_acquire(struct lockdep_map *lock, unsigned int subclass, } EXPORT_SYMBOL_GPL(lock_acquire); +DEFINE_TRACE(lock_release); + void lock_release(struct lockdep_map *lock, int nested, unsigned long ip) { unsigned long flags; + trace_lock_release(lock, nested, ip); + if (unlikely(current->lockdep_recursion)) return; @@ -3090,10 +3099,14 @@ __lock_acquired(struct lockdep_map *lock, unsigned long ip) lock->ip = ip; } +DEFINE_TRACE(lock_contended); + void lock_contended(struct lockdep_map *lock, unsigned long ip) { unsigned long flags; + trace_lock_contended(lock, ip); + if (unlikely(!lock_stat)) return; @@ -3109,10 +3122,14 @@ void lock_contended(struct lockdep_map *lock, unsigned long ip) } EXPORT_SYMBOL_GPL(lock_contended); +DEFINE_TRACE(lock_acquired); + void lock_acquired(struct lockdep_map *lock, unsigned long ip) { unsigned long flags; + trace_lock_acquired(lock, ip); + if (unlikely(!lock_stat)) return; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 12539f72f4a5..c8abbb0c8397 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -623,7 +623,7 @@ static unsigned map_pid_to_cmdline[PID_MAX_DEFAULT+1]; static unsigned map_cmdline_to_pid[SAVED_CMDLINES]; static char saved_cmdlines[SAVED_CMDLINES][TASK_COMM_LEN]; static int cmdline_idx; -static DEFINE_SPINLOCK(trace_cmdline_lock); +static raw_spinlock_t trace_cmdline_lock = __RAW_SPIN_LOCK_UNLOCKED; /* temporary disable recording */ static atomic_t trace_record_cmdline_disabled __read_mostly; @@ -735,7 +735,7 @@ static void trace_save_cmdline(struct task_struct *tsk) * nor do we want to disable interrupts, * so if we miss here, then better luck next time. */ - if (!spin_trylock(&trace_cmdline_lock)) + if (!__raw_spin_trylock(&trace_cmdline_lock)) return; idx = map_pid_to_cmdline[tsk->pid]; @@ -753,7 +753,7 @@ static void trace_save_cmdline(struct task_struct *tsk) memcpy(&saved_cmdlines[idx], tsk->comm, TASK_COMM_LEN); - spin_unlock(&trace_cmdline_lock); + __raw_spin_unlock(&trace_cmdline_lock); } char *trace_find_cmdline(int pid) @@ -3751,7 +3751,7 @@ static __init int tracer_init_debugfs(void) int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) { - static DEFINE_SPINLOCK(trace_buf_lock); + static raw_spinlock_t trace_buf_lock = __RAW_SPIN_LOCK_UNLOCKED; static char trace_buf[TRACE_BUF_SIZE]; struct ring_buffer_event *event; @@ -3773,7 +3773,8 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) goto out; pause_graph_tracing(); - spin_lock_irqsave(&trace_buf_lock, irq_flags); + raw_local_irq_save(irq_flags); + __raw_spin_lock(&trace_buf_lock); len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args); len = min(len, TRACE_BUF_SIZE-1); @@ -3792,7 +3793,8 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) ring_buffer_unlock_commit(tr->buffer, event); out_unlock: - spin_unlock_irqrestore(&trace_buf_lock, irq_flags); + __raw_spin_unlock(&trace_buf_lock); + raw_local_irq_restore(irq_flags); unpause_graph_tracing(); out: preempt_enable_notrace(); diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 041789ffbac1..2c8d76c7dbed 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -5,7 +5,7 @@ * * static void ftrace_event_(proto) * { - * event_trace_printk(_RET_IP_, "() " ); + * event_trace_printk(_RET_IP_, ": " ); * } * * static int ftrace_reg_event_(void) @@ -112,7 +112,7 @@ #define _TRACE_FORMAT(call, proto, args, fmt) \ static void ftrace_event_##call(proto) \ { \ - event_trace_printk(_RET_IP_, "(" #call ") " fmt); \ + event_trace_printk(_RET_IP_, #call ": " fmt); \ } \ \ static int ftrace_reg_event_##call(void) \ From e543ad76914abec1acf6631604a4154cd7a2ca6b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 4 Mar 2009 18:20:36 -0500 Subject: [PATCH 224/413] tracing: add cpu_file intialization for ftrace_dump Impact: fix to ftrace_dump output corruption The commit: b04cc6b1f6398b0e0b60d37e27ce51b4899672ec tracing/core: introduce per cpu tracing files added a new field to the iterator called cpu_file. This was a handle to differentiate between the per cpu trace output files and the all cpu "trace" file. The all cpu "trace" file required setting this to TRACE_PIPE_ALL_CPU. The problem is that the ftrace_dump sets up its own iterator but was not updated to handle this change. The result was only CPU 0 printing out on crash and a lot of "<0>"'s also being printed. Reported-by: Thomas Gleixner Tested-by: Darren Hart Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c8abbb0c8397..ab5cbcae43a1 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3918,8 +3918,10 @@ void ftrace_dump(void) printk(KERN_TRACE "Dumping ftrace buffer:\n"); + /* Simulate the iterator */ iter.tr = &global_trace; iter.trace = current_trace; + iter.cpu_file = TRACE_PIPE_ALL_CPU; /* * We need to stop all tracing on all CPUS to read the From 4f3640f8a358f2183a8c966f299eeb55ca523e06 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 3 Mar 2009 23:52:42 -0500 Subject: [PATCH 225/413] ring-buffer: fix timestamp in partial ring_buffer_page_read If a partial ring_buffer_page_read happens, then some of the incremental timestamps may be lost. This patch writes the recent timestamp into the page that is passed back to the caller. A partial ring_buffer_page_read is where the full page would not be written back to the user, and instead, just part of the page is copied to the user. A full page would be a page swap with the ring buffer and the timestamps would be correct. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index f2a163db52f9..f7473645b9c6 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2461,6 +2461,7 @@ int ring_buffer_read_page(struct ring_buffer *buffer, unsigned long flags; unsigned int commit; unsigned int read; + u64 save_timestamp; int ret = -1; /* @@ -2515,6 +2516,9 @@ int ring_buffer_read_page(struct ring_buffer *buffer, if (len < size) goto out; + /* save the current timestamp, since the user will need it */ + save_timestamp = cpu_buffer->read_stamp; + /* Need to copy one event at a time */ do { memcpy(bpage->data + pos, rpage->data + rpos, size); @@ -2531,7 +2535,7 @@ int ring_buffer_read_page(struct ring_buffer *buffer, /* update bpage */ local_set(&bpage->commit, pos); - bpage->time_stamp = rpage->time_stamp; + bpage->time_stamp = save_timestamp; /* we copied everything to the beginning */ read = 0; From 2dc5d12b1f43134e9bc5037f69f4739cfdfab93e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 4 Mar 2009 19:10:05 -0500 Subject: [PATCH 226/413] tracing: do not return EFAULT if read copied anything Impact: fix trace read to conform to standards Andrew Morton, Theodore Tso and H. Peter Anvin brought to my attention that a userspace read should not return -EFAULT if it succeeded in copying anything. It should only return -EFAULT if it failed to copy at all. This patch modifies the check of copy_from_user and updates the return code appropriately. I also used H. Peter Anvin's short cut rule to just test ret == count. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index ab5cbcae43a1..57155dc53530 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -346,6 +346,9 @@ ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) int len; int ret; + if (!cnt) + return 0; + if (s->len <= s->readpos) return -EBUSY; @@ -353,9 +356,11 @@ ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) if (cnt > len) cnt = len; ret = copy_to_user(ubuf, s->buffer + s->readpos, cnt); - if (ret) + if (ret == cnt) return -EFAULT; + cnt -= ret; + s->readpos += len; return cnt; } @@ -3049,6 +3054,9 @@ tracing_buffers_read(struct file *filp, char __user *ubuf, ssize_t ret; size_t size; + if (!count) + return 0; + /* Do we have previous read data to read? */ if (info->read < PAGE_SIZE) goto read; @@ -3073,8 +3081,10 @@ tracing_buffers_read(struct file *filp, char __user *ubuf, size = count; ret = copy_to_user(ubuf, info->spare + info->read, size); - if (ret) + if (ret == size) return -EFAULT; + size -= ret; + *ppos += size; info->read += size; From e74da5235cec6cb71eb338c987f876ecc793138b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 4 Mar 2009 20:31:11 -0500 Subject: [PATCH 227/413] tracing: fix seq read from trace files The buffer used by trace_seq was updated incorrectly. Instead of consuming what was actually read, it consumed the rest of the buffer on reads. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 57155dc53530..2e53e6f09440 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -361,7 +361,7 @@ ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) cnt -= ret; - s->readpos += len; + s->readpos += cnt; return cnt; } @@ -380,7 +380,7 @@ ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt) if (!ret) return -EFAULT; - s->readpos += len; + s->readpos += cnt; return cnt; } From c032ef64d680717e4e8ce3da65da6419a35f8a2c Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 4 Mar 2009 20:34:24 -0500 Subject: [PATCH 228/413] tracing: add latency output format option With the removal of the latency_trace file, we lost the ability to see some of the finer details in a trace. Like the state of interrupts enabled, the preempt count, need resched, and if we are in an interrupt handler, softirq handler or not. This patch simply creates an option to bring back the old format. This also removes the warning about an unused variable that held the latency_trace file operations. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 24 ++---------------------- kernel/trace/trace.h | 3 ++- 2 files changed, 4 insertions(+), 23 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 2e53e6f09440..55fcbb567950 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -299,6 +299,7 @@ static const char *trace_options[] = { "sym-userobj", "printk-msg-only", "context-info", + "latency-format", NULL }; @@ -1829,26 +1830,12 @@ static int tracing_open(struct inode *inode, struct file *file) iter = __tracing_open(inode, file); if (IS_ERR(iter)) ret = PTR_ERR(iter); - - return ret; -} - -static int tracing_lt_open(struct inode *inode, struct file *file) -{ - struct trace_iterator *iter; - int ret = 0; - - iter = __tracing_open(inode, file); - - if (IS_ERR(iter)) - ret = PTR_ERR(iter); - else + else if (trace_flags & TRACE_ITER_LATENCY_FMT) iter->iter_flags |= TRACE_FILE_LAT_FMT; return ret; } - static void * t_next(struct seq_file *m, void *v, loff_t *pos) { @@ -1927,13 +1914,6 @@ static struct file_operations tracing_fops = { .release = tracing_release, }; -static struct file_operations tracing_lt_fops = { - .open = tracing_lt_open, - .read = seq_read, - .llseek = seq_lseek, - .release = tracing_release, -}; - static struct file_operations show_traces_fops = { .open = show_traces_open, .read = seq_read, diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 561bb5c5d988..12cd119cca32 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -651,7 +651,8 @@ enum trace_iterator_flags { TRACE_ITER_USERSTACKTRACE = 0x4000, TRACE_ITER_SYM_USEROBJ = 0x8000, TRACE_ITER_PRINTK_MSGONLY = 0x10000, - TRACE_ITER_CONTEXT_INFO = 0x20000 /* Print pid/cpu/time */ + TRACE_ITER_CONTEXT_INFO = 0x20000, /* Print pid/cpu/time */ + TRACE_ITER_LATENCY_FMT = 0x40000, }; /* From 5fd73f862468280d4cbb5ba4321502f911f9f89a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 4 Mar 2009 21:42:04 -0500 Subject: [PATCH 229/413] tracing: remove extra latency_trace method from trace structure Impact: clean up The trace and latency_trace function pointers are identical for every tracer but the function tracer. The differences in the function tracer are trivial (latency output puts paranthesis around parent). This patch removes the latency_trace pointer and all prints will now just use the trace output function pointer. Signed-off-by: Steven Rostedt --- kernel/trace/blktrace.c | 1 - kernel/trace/trace.c | 2 +- kernel/trace/trace_branch.c | 1 - kernel/trace/trace_output.c | 32 -------------------------------- kernel/trace/trace_output.h | 1 - 5 files changed, 1 insertion(+), 36 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index e82cb9e930cc..e39679a72a3b 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1231,7 +1231,6 @@ static struct tracer blk_tracer __read_mostly = { static struct trace_event trace_blk_event = { .type = TRACE_BLK, .trace = blk_trace_event_print, - .latency_trace = blk_trace_event_print, .binary = blk_trace_event_print_binary, }; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 55fcbb567950..21b89ecb8480 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1485,7 +1485,7 @@ static enum print_line_t print_lat_fmt(struct trace_iterator *iter) } if (event) - return event->latency_trace(iter, sym_flags); + return event->trace(iter, sym_flags); if (!trace_seq_printf(s, "Unknown type %d\n", entry->type)) goto partial; diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index c2e68d440c4d..aaa0755268b9 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -159,7 +159,6 @@ static enum print_line_t trace_branch_print(struct trace_iterator *iter, static struct trace_event trace_branch_event = { .type = TRACE_BRANCH, .trace = trace_branch_print, - .latency_trace = trace_branch_print, }; static struct tracer branch_trace __read_mostly = diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 9fc815031b09..306fef84c503 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -437,8 +437,6 @@ int register_ftrace_event(struct trace_event *event) if (event->trace == NULL) event->trace = trace_nop_print; - if (event->latency_trace == NULL) - event->latency_trace = trace_nop_print; if (event->raw == NULL) event->raw = trace_nop_print; if (event->hex == NULL) @@ -480,29 +478,6 @@ enum print_line_t trace_nop_print(struct trace_iterator *iter, int flags) } /* TRACE_FN */ -static enum print_line_t trace_fn_latency(struct trace_iterator *iter, - int flags) -{ - struct ftrace_entry *field; - struct trace_seq *s = &iter->seq; - - trace_assign_type(field, iter->ent); - - if (!seq_print_ip_sym(s, field->ip, flags)) - goto partial; - if (!trace_seq_puts(s, " (")) - goto partial; - if (!seq_print_ip_sym(s, field->parent_ip, flags)) - goto partial; - if (!trace_seq_puts(s, ")\n")) - goto partial; - - return TRACE_TYPE_HANDLED; - - partial: - return TRACE_TYPE_PARTIAL_LINE; -} - static enum print_line_t trace_fn_trace(struct trace_iterator *iter, int flags) { struct ftrace_entry *field; @@ -573,7 +548,6 @@ static enum print_line_t trace_fn_bin(struct trace_iterator *iter, int flags) static struct trace_event trace_fn_event = { .type = TRACE_FN, .trace = trace_fn_trace, - .latency_trace = trace_fn_latency, .raw = trace_fn_raw, .hex = trace_fn_hex, .binary = trace_fn_bin, @@ -705,7 +679,6 @@ static enum print_line_t trace_ctxwake_bin(struct trace_iterator *iter, static struct trace_event trace_ctx_event = { .type = TRACE_CTX, .trace = trace_ctx_print, - .latency_trace = trace_ctx_print, .raw = trace_ctx_raw, .hex = trace_ctx_hex, .binary = trace_ctxwake_bin, @@ -714,7 +687,6 @@ static struct trace_event trace_ctx_event = { static struct trace_event trace_wake_event = { .type = TRACE_WAKE, .trace = trace_wake_print, - .latency_trace = trace_wake_print, .raw = trace_wake_raw, .hex = trace_wake_hex, .binary = trace_ctxwake_bin, @@ -770,7 +742,6 @@ static enum print_line_t trace_special_bin(struct trace_iterator *iter, static struct trace_event trace_special_event = { .type = TRACE_SPECIAL, .trace = trace_special_print, - .latency_trace = trace_special_print, .raw = trace_special_print, .hex = trace_special_hex, .binary = trace_special_bin, @@ -808,7 +779,6 @@ static enum print_line_t trace_stack_print(struct trace_iterator *iter, static struct trace_event trace_stack_event = { .type = TRACE_STACK, .trace = trace_stack_print, - .latency_trace = trace_stack_print, .raw = trace_special_print, .hex = trace_special_hex, .binary = trace_special_bin, @@ -838,7 +808,6 @@ static enum print_line_t trace_user_stack_print(struct trace_iterator *iter, static struct trace_event trace_user_stack_event = { .type = TRACE_USER_STACK, .trace = trace_user_stack_print, - .latency_trace = trace_user_stack_print, .raw = trace_special_print, .hex = trace_special_hex, .binary = trace_special_bin, @@ -883,7 +852,6 @@ static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) static struct trace_event trace_print_event = { .type = TRACE_PRINT, .trace = trace_print_print, - .latency_trace = trace_print_print, .raw = trace_print_raw, }; diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 551a25a72217..8a34d688ed63 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -10,7 +10,6 @@ struct trace_event { struct hlist_node node; int type; trace_print_func trace; - trace_print_func latency_trace; trace_print_func raw; trace_print_func hex; trace_print_func binary; From 27d48be84477d2f0a2e2ac3738a3971dece631d5 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 4 Mar 2009 21:57:29 -0500 Subject: [PATCH 230/413] tracing: consolidate print_lat_fmt and print_trace_fmt Impact: clean up Both print_lat_fmt and print_trace_fmt do pretty much the same thing except for one different function call. This patch consolidates the two functions and adds an if statement to perform the difference. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 39 +++++++-------------------------------- 1 file changed, 7 insertions(+), 32 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 21b89ecb8480..d1ef43999d9e 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1468,33 +1468,6 @@ static void test_cpu_buff_start(struct trace_iterator *iter) trace_seq_printf(s, "##### CPU %u buffer started ####\n", iter->cpu); } -static enum print_line_t print_lat_fmt(struct trace_iterator *iter) -{ - struct trace_seq *s = &iter->seq; - unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); - struct trace_event *event; - struct trace_entry *entry = iter->ent; - - test_cpu_buff_start(iter); - - event = ftrace_find_event(entry->type); - - if (trace_flags & TRACE_ITER_CONTEXT_INFO) { - if (!trace_print_lat_context(iter)) - goto partial; - } - - if (event) - return event->trace(iter, sym_flags); - - if (!trace_seq_printf(s, "Unknown type %d\n", entry->type)) - goto partial; - - return TRACE_TYPE_HANDLED; -partial: - return TRACE_TYPE_PARTIAL_LINE; -} - static enum print_line_t print_trace_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; @@ -1509,8 +1482,13 @@ static enum print_line_t print_trace_fmt(struct trace_iterator *iter) event = ftrace_find_event(entry->type); if (trace_flags & TRACE_ITER_CONTEXT_INFO) { - if (!trace_print_context(iter)) - goto partial; + if (iter->iter_flags & TRACE_FILE_LAT_FMT) { + if (!trace_print_lat_context(iter)) + goto partial; + } else { + if (!trace_print_context(iter)) + goto partial; + } } if (event) @@ -1652,9 +1630,6 @@ static enum print_line_t print_trace_line(struct trace_iterator *iter) if (trace_flags & TRACE_ITER_RAW) return print_raw_fmt(iter); - if (iter->iter_flags & TRACE_FILE_LAT_FMT) - return print_lat_fmt(iter); - return print_trace_fmt(iter); } From e9d25fe6eaa2c720bb3ea661b660e58d54fa38bf Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 4 Mar 2009 22:15:30 -0500 Subject: [PATCH 231/413] tracing: have latency tracers set the latency format The latency tracers (irqsoff, preemptoff, preemptirqsoff, and wakeup) are pretty useless with the default output format. This patch makes them automatically enable the latency format when they are selected. They also record the state of the latency option, and if it was not enabled when selected, they disable it on reset. Signed-off-by: Steven Rostedt --- kernel/trace/trace_irqsoff.c | 8 ++++++++ kernel/trace/trace_sched_wakeup.c | 8 ++++++++ 2 files changed, 16 insertions(+) diff --git a/kernel/trace/trace_irqsoff.c b/kernel/trace/trace_irqsoff.c index 9e5ebd844158..b923d13e2fad 100644 --- a/kernel/trace/trace_irqsoff.c +++ b/kernel/trace/trace_irqsoff.c @@ -32,6 +32,8 @@ enum { static int trace_type __read_mostly; +static int save_lat_flag; + #ifdef CONFIG_PREEMPT_TRACER static inline int preempt_trace(void) @@ -370,6 +372,9 @@ static void stop_irqsoff_tracer(struct trace_array *tr) static void __irqsoff_tracer_init(struct trace_array *tr) { + save_lat_flag = trace_flags & TRACE_ITER_LATENCY_FMT; + trace_flags |= TRACE_ITER_LATENCY_FMT; + tracing_max_latency = 0; irqsoff_trace = tr; /* make sure that the tracer is visible */ @@ -380,6 +385,9 @@ static void __irqsoff_tracer_init(struct trace_array *tr) static void irqsoff_tracer_reset(struct trace_array *tr) { stop_irqsoff_tracer(tr); + + if (!save_lat_flag) + trace_flags &= ~TRACE_ITER_LATENCY_FMT; } static void irqsoff_tracer_start(struct trace_array *tr) diff --git a/kernel/trace/trace_sched_wakeup.c b/kernel/trace/trace_sched_wakeup.c index db55f7aaa640..3c5ad6b2ec84 100644 --- a/kernel/trace/trace_sched_wakeup.c +++ b/kernel/trace/trace_sched_wakeup.c @@ -32,6 +32,8 @@ static raw_spinlock_t wakeup_lock = static void __wakeup_reset(struct trace_array *tr); +static int save_lat_flag; + #ifdef CONFIG_FUNCTION_TRACER /* * irqsoff uses its own tracer function to keep the overhead down: @@ -324,6 +326,9 @@ static void stop_wakeup_tracer(struct trace_array *tr) static int __wakeup_tracer_init(struct trace_array *tr) { + save_lat_flag = trace_flags & TRACE_ITER_LATENCY_FMT; + trace_flags |= TRACE_ITER_LATENCY_FMT; + tracing_max_latency = 0; wakeup_trace = tr; start_wakeup_tracer(tr); @@ -347,6 +352,9 @@ static void wakeup_tracer_reset(struct trace_array *tr) stop_wakeup_tracer(tr); /* make sure we put back any tasks we are tracing */ wakeup_reset(tr); + + if (!save_lat_flag) + trace_flags &= ~TRACE_ITER_LATENCY_FMT; } static void wakeup_tracer_start(struct trace_array *tr) From 5e1607a00bd082972629d3d68c95c8bcf902b55a Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 5 Mar 2009 10:24:48 +0100 Subject: [PATCH 232/413] tracing: rename ftrace_printk() => trace_printk() Impact: cleanup Use a more generic name - this also allows the prototype to move to kernel.h and be generally available to kernel developers who want to do some quick tracing. Signed-off-by: Ingo Molnar --- Documentation/ftrace.txt | 6 +++--- include/linux/ftrace.h | 18 +++++++++--------- kernel/trace/trace.c | 8 ++++---- kernel/trace/trace.h | 2 +- 4 files changed, 17 insertions(+), 17 deletions(-) diff --git a/Documentation/ftrace.txt b/Documentation/ftrace.txt index 2041ee951c1a..22614bef6359 100644 --- a/Documentation/ftrace.txt +++ b/Documentation/ftrace.txt @@ -1466,11 +1466,11 @@ want, depending on your needs. You can put some comments on specific functions by using -ftrace_printk() For example, if you want to put a comment inside +trace_printk() For example, if you want to put a comment inside the __might_sleep() function, you just have to include - and call ftrace_printk() inside __might_sleep() + and call trace_printk() inside __might_sleep() -ftrace_printk("I'm a comment!\n") +trace_printk("I'm a comment!\n") will produce: diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 1f69ac7c1587..fbb9c364e166 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -329,11 +329,11 @@ extern void ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3); /** - * ftrace_printk - printf formatting in the ftrace buffer + * trace_printk - printf formatting in the ftrace buffer * @fmt: the printf format for printing * - * Note: __ftrace_printk is an internal function for ftrace_printk and - * the @ip is passed in via the ftrace_printk macro. + * Note: __trace_printk is an internal function for trace_printk and + * the @ip is passed in via the trace_printk macro. * * This function allows a kernel developer to debug fast path sections * that printk is not appropriate for. By scattering in various @@ -341,14 +341,14 @@ ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3); * where problems are occurring. * * This is intended as a debugging tool for the developer only. - * Please refrain from leaving ftrace_printks scattered around in + * Please refrain from leaving trace_printks scattered around in * your code. */ -# define ftrace_printk(fmt...) __ftrace_printk(_THIS_IP_, fmt) +# define trace_printk(fmt...) __trace_printk(_THIS_IP_, fmt) extern int -__ftrace_printk(unsigned long ip, const char *fmt, ...) +__trace_printk(unsigned long ip, const char *fmt, ...) __attribute__ ((format (printf, 2, 3))); -# define ftrace_vprintk(fmt, ap) __ftrace_printk(_THIS_IP_, fmt, ap) +# define ftrace_vprintk(fmt, ap) __trace_printk(_THIS_IP_, fmt, ap) extern int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); extern void ftrace_dump(void); @@ -356,13 +356,13 @@ extern void ftrace_dump(void); static inline void ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3) { } static inline int -ftrace_printk(const char *fmt, ...) __attribute__ ((format (printf, 1, 2))); +trace_printk(const char *fmt, ...) __attribute__ ((format (printf, 1, 2))); static inline void tracing_start(void) { } static inline void tracing_stop(void) { } static inline void ftrace_off_permanent(void) { } static inline int -ftrace_printk(const char *fmt, ...) +trace_printk(const char *fmt, ...) { return 0; } diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index d1ef43999d9e..c0e9c1263393 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -48,7 +48,7 @@ unsigned long __read_mostly tracing_thresh; * We need to change this state when a selftest is running. * A selftest will lurk into the ring-buffer to count the * entries inserted during the selftest although some concurrent - * insertions into the ring-buffer such as ftrace_printk could occurred + * insertions into the ring-buffer such as trace_printk could occurred * at the same time, giving false positive or negative results. */ static bool __read_mostly tracing_selftest_running; @@ -291,7 +291,7 @@ static const char *trace_options[] = { "block", "stacktrace", "sched-tree", - "ftrace_printk", + "trace_printk", "ftrace_preempt", "branch", "annotate", @@ -3768,7 +3768,7 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) } EXPORT_SYMBOL_GPL(trace_vprintk); -int __ftrace_printk(unsigned long ip, const char *fmt, ...) +int __trace_printk(unsigned long ip, const char *fmt, ...) { int ret; va_list ap; @@ -3781,7 +3781,7 @@ int __ftrace_printk(unsigned long ip, const char *fmt, ...) va_end(ap); return ret; } -EXPORT_SYMBOL_GPL(__ftrace_printk); +EXPORT_SYMBOL_GPL(__trace_printk); int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) { diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 12cd119cca32..8beff03fda68 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -115,7 +115,7 @@ struct userstack_entry { }; /* - * ftrace_printk entry: + * trace_printk entry: */ struct print_entry { struct trace_entry ent; From 526211bc58c4b3265352801c5a7f469af5c34711 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 5 Mar 2009 10:28:45 +0100 Subject: [PATCH 233/413] tracing: move utility functions from ftrace.h to kernel.h Make common utility functions such as trace_printk() and tracing_start()/tracing_stop() generally available to kernel code. Cc: Steven Rostedt Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 58 ++---------------------------------------- include/linux/kernel.h | 58 ++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+), 56 deletions(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index fbb9c364e166..5b64303ec9f2 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -318,62 +318,6 @@ static inline void __ftrace_enabled_restore(int enabled) # define trace_preempt_off(a0, a1) do { } while (0) #endif -#ifdef CONFIG_TRACING -extern int ftrace_dump_on_oops; - -extern void tracing_start(void); -extern void tracing_stop(void); -extern void ftrace_off_permanent(void); - -extern void -ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3); - -/** - * trace_printk - printf formatting in the ftrace buffer - * @fmt: the printf format for printing - * - * Note: __trace_printk is an internal function for trace_printk and - * the @ip is passed in via the trace_printk macro. - * - * This function allows a kernel developer to debug fast path sections - * that printk is not appropriate for. By scattering in various - * printk like tracing in the code, a developer can quickly see - * where problems are occurring. - * - * This is intended as a debugging tool for the developer only. - * Please refrain from leaving trace_printks scattered around in - * your code. - */ -# define trace_printk(fmt...) __trace_printk(_THIS_IP_, fmt) -extern int -__trace_printk(unsigned long ip, const char *fmt, ...) - __attribute__ ((format (printf, 2, 3))); -# define ftrace_vprintk(fmt, ap) __trace_printk(_THIS_IP_, fmt, ap) -extern int -__ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); -extern void ftrace_dump(void); -#else -static inline void -ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3) { } -static inline int -trace_printk(const char *fmt, ...) __attribute__ ((format (printf, 1, 2))); - -static inline void tracing_start(void) { } -static inline void tracing_stop(void) { } -static inline void ftrace_off_permanent(void) { } -static inline int -trace_printk(const char *fmt, ...) -{ - return 0; -} -static inline int -ftrace_vprintk(const char *fmt, va_list ap) -{ - return 0; -} -static inline void ftrace_dump(void) { } -#endif - #ifdef CONFIG_FTRACE_MCOUNT_RECORD extern void ftrace_init(void); extern void ftrace_init_module(struct module *mod, @@ -542,6 +486,8 @@ static inline int test_tsk_trace_graph(struct task_struct *tsk) return tsk->trace & TSK_TRACE_FL_GRAPH; } +extern int ftrace_dump_on_oops; + #endif /* CONFIG_TRACING */ diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 7fa371898e3e..08bf5da86676 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -367,6 +367,64 @@ static inline char *pack_hex_byte(char *buf, u8 byte) ({ if (0) printk(KERN_DEBUG pr_fmt(fmt), ##__VA_ARGS__); 0; }) #endif +/* + * General tracing related utility functions - trace_printk(), + * tracing_start()/tracing_stop: + */ +#ifdef CONFIG_TRACING +extern void tracing_start(void); +extern void tracing_stop(void); +extern void ftrace_off_permanent(void); + +extern void +ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3); + +/** + * trace_printk - printf formatting in the ftrace buffer + * @fmt: the printf format for printing + * + * Note: __trace_printk is an internal function for trace_printk and + * the @ip is passed in via the trace_printk macro. + * + * This function allows a kernel developer to debug fast path sections + * that printk is not appropriate for. By scattering in various + * printk like tracing in the code, a developer can quickly see + * where problems are occurring. + * + * This is intended as a debugging tool for the developer only. + * Please refrain from leaving trace_printks scattered around in + * your code. + */ +# define trace_printk(fmt...) __trace_printk(_THIS_IP_, fmt) +extern int +__trace_printk(unsigned long ip, const char *fmt, ...) + __attribute__ ((format (printf, 2, 3))); +# define ftrace_vprintk(fmt, ap) __trace_printk(_THIS_IP_, fmt, ap) +extern int +__ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); +extern void ftrace_dump(void); +#else +static inline void +ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3) { } +static inline int +trace_printk(const char *fmt, ...) __attribute__ ((format (printf, 1, 2))); + +static inline void tracing_start(void) { } +static inline void tracing_stop(void) { } +static inline void ftrace_off_permanent(void) { } +static inline int +trace_printk(const char *fmt, ...) +{ + return 0; +} +static inline int +ftrace_vprintk(const char *fmt, va_list ap) +{ + return 0; +} +static inline void ftrace_dump(void) { } +#endif + /* * Display an IP address in readable format. */ From 0012693ad4f636c720fed3802027f9427962f540 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 5 Mar 2009 01:49:22 +0100 Subject: [PATCH 234/413] tracing/function-graph-tracer: use the more lightweight local clock Impact: decrease hangs risks with the graph tracer on slow systems Since the function graph tracer can spend too much time on timer interrupts, it's better now to use the more lightweight local clock. Anyway, the function graph traces are more reliable on a per cpu trace. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Peter Zijlstra LKML-Reference: <49af243d.06e9300a.53ad.ffff840c@mx.google.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/ftrace.c | 2 +- include/linux/ftrace.h | 13 +++++++------ kernel/trace/trace_functions_graph.c | 2 +- 3 files changed, 9 insertions(+), 8 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 3925ec0184b1..a85da1764b1c 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -436,7 +436,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) return; } - calltime = cpu_clock(raw_smp_processor_id()); + calltime = trace_clock_local(); if (ftrace_push_return_trace(old, calltime, self_addr, &trace.depth) == -EBUSY) { diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 1f69ac7c1587..6ea62acbe4b6 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -1,15 +1,16 @@ #ifndef _LINUX_FTRACE_H #define _LINUX_FTRACE_H -#include -#include -#include -#include -#include -#include +#include #include +#include #include +#include +#include #include +#include +#include +#include #include diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index c009553a8e81..e527f2f66c73 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -112,7 +112,7 @@ unsigned long ftrace_return_to_handler(void) unsigned long ret; ftrace_pop_return_trace(&trace, &ret); - trace.rettime = cpu_clock(raw_smp_processor_id()); + trace.rettime = trace_clock_local(); ftrace_graph_return(&trace); if (unlikely(!ret)) { From 2002c258faaa8f89543df284fdbaa9e4b171547f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Mar 2009 10:35:56 -0500 Subject: [PATCH 235/413] tracing: add tracing_on/tracing_off to kernel.h Impact: cleanup The functions tracing_start/tracing_stop have been moved to kernel.h. These are not the functions a developer most likely wants to use when they want to insert a place to stop tracing and restart it from user space. tracing_start/tracing_stop was created to work with things like suspend to ram, where even calling smp_processor_id() can crash the system. The tracing_start/tracing_stop was used to stop the tracer from doing anything. These are still light weight functions, but add a bit more overhead to be able to stop the tracers. They also have no interface back to userland. That is, if the kernel calls tracing_stop, userland can not start tracing. What a developer most likely wants to use is tracing_on/tracing_off. These are very light weight functions (simply sets or clears a bit). These functions just stop recording into the ring buffer. The tracers don't even know that this happens except that they would receive NULL from the ring_buffer_lock_reserve function. Also, there's a way for the user land to enable or disable this bit. In debugfs/tracing/tracing_on, a user may echo "0" (same as tracing_off()) or echo "1" (same as tracing_on()) into this file. This becomes handy when a kernel developer is debugging and wants tracing to turn off when it hits an anomaly. Then the developer can examine the trace, and restart tracing if they want to try again (echo 1 > tracing_on). This patch moves the prototypes for tracing_on/tracing_off to kernel.h and comments their use, so that a kernel developer will know how to use them. Signed-off-by: Steven Rostedt --- include/linux/kernel.h | 29 ++++++++++++++++++++++++++++- include/linux/ring_buffer.h | 15 --------------- 2 files changed, 28 insertions(+), 16 deletions(-) diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 08bf5da86676..d4614a8a034b 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -369,8 +369,35 @@ static inline char *pack_hex_byte(char *buf, u8 byte) /* * General tracing related utility functions - trace_printk(), - * tracing_start()/tracing_stop: + * tracing_on/tracing_off and tracing_start()/tracing_stop + * + * Use tracing_on/tracing_off when you want to quickly turn on or off + * tracing. It simply enables or disables the recording of the trace events. + * This also corresponds to the user space debugfs/tracing/tracing_on + * file, which gives a means for the kernel and userspace to interact. + * Place a tracing_off() in the kernel where you want tracing to end. + * From user space, examine the trace, and then echo 1 > tracing_on + * to continue tracing. + * + * tracing_stop/tracing_start has slightly more overhead. It is used + * by things like suspend to ram where disabling the recording of the + * trace is not enough, but tracing must actually stop because things + * like calling smp_processor_id() may crash the system. + * + * Most likely, you want to use tracing_on/tracing_off. */ +#ifdef CONFIG_RING_BUFFER +void tracing_on(void); +void tracing_off(void); +/* trace_off_permanent stops recording with no way to bring it back */ +void tracing_off_permanent(void); +int tracing_is_on(void); +#else +static inline void tracing_on(void) { } +static inline void tracing_off(void) { } +static inline void tracing_off_permanent(void) { } +static inline int tracing_is_on(void) { return 0; } +#endif #ifdef CONFIG_TRACING extern void tracing_start(void); extern void tracing_stop(void); diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index 79fcbc4b09d6..b1a0068a5557 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -124,21 +124,6 @@ void ring_buffer_normalize_time_stamp(int cpu, u64 *ts); size_t ring_buffer_page_len(void *page); -/* - * The below functions are fine to use outside the tracing facility. - */ -#ifdef CONFIG_RING_BUFFER -void tracing_on(void); -void tracing_off(void); -void tracing_off_permanent(void); -int tracing_is_on(void); -#else -static inline void tracing_on(void) { } -static inline void tracing_off(void) { } -static inline void tracing_off_permanent(void) { } -static inline int tracing_is_on(void) { return 0; } -#endif - void *ring_buffer_alloc_read_page(struct ring_buffer *buffer); void ring_buffer_free_read_page(struct ring_buffer *buffer, void *data); int ring_buffer_read_page(struct ring_buffer *buffer, void **data_page, From 40ada30f9621fbd831ac2437b9a2a399aad34b00 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 5 Mar 2009 21:19:55 +0100 Subject: [PATCH 236/413] tracing: clean up menu Clean up menu structure, introduce TRACING_SUPPORT switch that signals whether an architecture supports various instrumentation mechanisms. Signed-off-by: Ingo Molnar --- arch/Kconfig | 1 + kernel/trace/Kconfig | 29 ++++++++++++++++------------- 2 files changed, 17 insertions(+), 13 deletions(-) diff --git a/arch/Kconfig b/arch/Kconfig index 550dab22daa1..a092dc77c24d 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -6,6 +6,7 @@ config OPROFILE tristate "OProfile system profiling (EXPERIMENTAL)" depends on PROFILING depends on HAVE_OPROFILE + depends on TRACING_SUPPORT select TRACING select RING_BUFFER help diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 999c6a2485df..5d733da5345a 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -53,12 +53,22 @@ config TRACING select TRACEPOINTS select NOP_TRACER +# +# Minimum requirements an architecture has to meet for us to +# be able to offer generic tracing facilities: +# +config TRACING_SUPPORT + bool + depends on TRACE_IRQFLAGS_SUPPORT + depends on STACKTRACE_SUPPORT + +if TRACING_SUPPORT + menu "Tracers" config FUNCTION_TRACER bool "Kernel Function Tracer" depends on HAVE_FUNCTION_TRACER - depends on DEBUG_KERNEL select FRAME_POINTER select KALLSYMS select TRACING @@ -91,7 +101,6 @@ config IRQSOFF_TRACER default n depends on TRACE_IRQFLAGS_SUPPORT depends on GENERIC_TIME - depends on DEBUG_KERNEL select TRACE_IRQFLAGS select TRACING select TRACER_MAX_TRACE @@ -114,7 +123,6 @@ config PREEMPT_TRACER default n depends on GENERIC_TIME depends on PREEMPT - depends on DEBUG_KERNEL select TRACING select TRACER_MAX_TRACE help @@ -142,7 +150,6 @@ config SYSPROF_TRACER config SCHED_TRACER bool "Scheduling Latency Tracer" - depends on DEBUG_KERNEL select TRACING select CONTEXT_SWITCH_TRACER select TRACER_MAX_TRACE @@ -152,7 +159,6 @@ config SCHED_TRACER config CONTEXT_SWITCH_TRACER bool "Trace process context switches" - depends on DEBUG_KERNEL select TRACING select MARKERS help @@ -161,7 +167,6 @@ config CONTEXT_SWITCH_TRACER config EVENT_TRACER bool "Trace various events in the kernel" - depends on DEBUG_KERNEL select TRACING help This tracer hooks to various trace points in the kernel @@ -170,7 +175,6 @@ config EVENT_TRACER config BOOT_TRACER bool "Trace boot initcalls" - depends on DEBUG_KERNEL select TRACING select CONTEXT_SWITCH_TRACER help @@ -188,7 +192,6 @@ config BOOT_TRACER config TRACE_BRANCH_PROFILING bool "Trace likely/unlikely profiler" - depends on DEBUG_KERNEL select TRACING help This tracer profiles all the the likely and unlikely macros @@ -241,7 +244,6 @@ config BRANCH_TRACER config POWER_TRACER bool "Trace power consumption behavior" - depends on DEBUG_KERNEL depends on X86 select TRACING help @@ -253,7 +255,6 @@ config POWER_TRACER config STACK_TRACER bool "Trace max stack" depends on HAVE_FUNCTION_TRACER - depends on DEBUG_KERNEL select FUNCTION_TRACER select STACKTRACE select KALLSYMS @@ -343,7 +344,6 @@ config DYNAMIC_FTRACE bool "enable/disable ftrace tracepoints dynamically" depends on FUNCTION_TRACER depends on HAVE_DYNAMIC_FTRACE - depends on DEBUG_KERNEL default y help This option will modify all the calls to ftrace dynamically @@ -369,7 +369,7 @@ config FTRACE_SELFTEST config FTRACE_STARTUP_TEST bool "Perform a startup test on ftrace" - depends on TRACING && DEBUG_KERNEL + depends on TRACING select FTRACE_SELFTEST help This option performs a series of startup tests on ftrace. On bootup @@ -379,7 +379,7 @@ config FTRACE_STARTUP_TEST config MMIOTRACE bool "Memory mapped IO tracing" - depends on HAVE_MMIOTRACE_SUPPORT && DEBUG_KERNEL && PCI + depends on HAVE_MMIOTRACE_SUPPORT && PCI select TRACING help Mmiotrace traces Memory Mapped I/O access and is meant for @@ -401,3 +401,6 @@ config MMIOTRACE_TEST Say N, unless you absolutely know what you are doing. endmenu + +endif # TRACING_SUPPORT + From 5e2336a0d47c9661a40cc5ef85135ce1406af6e8 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Mar 2009 21:44:55 -0500 Subject: [PATCH 237/413] tracing: make all file_operations const Impact: cleanup All file_operations structures should be constant. No one is going to change them. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 10 +++++----- kernel/trace/ring_buffer.c | 2 +- kernel/trace/trace.c | 24 ++++++++++++------------ kernel/trace/trace_sysprof.c | 2 +- 4 files changed, 19 insertions(+), 19 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 5a3a06b21eee..d7a06a0d9447 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1869,21 +1869,21 @@ ftrace_notrace_release(struct inode *inode, struct file *file) return ftrace_regex_release(inode, file, 0); } -static struct file_operations ftrace_avail_fops = { +static const struct file_operations ftrace_avail_fops = { .open = ftrace_avail_open, .read = seq_read, .llseek = seq_lseek, .release = ftrace_avail_release, }; -static struct file_operations ftrace_failures_fops = { +static const struct file_operations ftrace_failures_fops = { .open = ftrace_failures_open, .read = seq_read, .llseek = seq_lseek, .release = ftrace_avail_release, }; -static struct file_operations ftrace_filter_fops = { +static const struct file_operations ftrace_filter_fops = { .open = ftrace_filter_open, .read = ftrace_regex_read, .write = ftrace_filter_write, @@ -1891,7 +1891,7 @@ static struct file_operations ftrace_filter_fops = { .release = ftrace_filter_release, }; -static struct file_operations ftrace_notrace_fops = { +static const struct file_operations ftrace_notrace_fops = { .open = ftrace_notrace_open, .read = ftrace_regex_read, .write = ftrace_notrace_write, @@ -2423,7 +2423,7 @@ ftrace_pid_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations ftrace_pid_fops = { +static const struct file_operations ftrace_pid_fops = { .read = ftrace_pid_read, .write = ftrace_pid_write, }; diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index f7473645b9c6..178858492a89 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2606,7 +2606,7 @@ rb_simple_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations rb_simple_fops = { +static const struct file_operations rb_simple_fops = { .open = tracing_open_generic, .read = rb_simple_read, .write = rb_simple_write, diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c0e9c1263393..e6144acf2b75 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1882,14 +1882,14 @@ static int show_traces_open(struct inode *inode, struct file *file) return ret; } -static struct file_operations tracing_fops = { +static const struct file_operations tracing_fops = { .open = tracing_open, .read = seq_read, .llseek = seq_lseek, .release = tracing_release, }; -static struct file_operations show_traces_fops = { +static const struct file_operations show_traces_fops = { .open = show_traces_open, .read = seq_read, .release = seq_release, @@ -1982,7 +1982,7 @@ tracing_cpumask_write(struct file *filp, const char __user *ubuf, return err; } -static struct file_operations tracing_cpumask_fops = { +static const struct file_operations tracing_cpumask_fops = { .open = tracing_open_generic, .read = tracing_cpumask_read, .write = tracing_cpumask_write, @@ -2134,7 +2134,7 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations tracing_iter_fops = { +static const struct file_operations tracing_iter_fops = { .open = tracing_open_generic, .read = tracing_trace_options_read, .write = tracing_trace_options_write, @@ -2167,7 +2167,7 @@ tracing_readme_read(struct file *filp, char __user *ubuf, readme_msg, strlen(readme_msg)); } -static struct file_operations tracing_readme_fops = { +static const struct file_operations tracing_readme_fops = { .open = tracing_open_generic, .read = tracing_readme_read, }; @@ -2927,25 +2927,25 @@ tracing_mark_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations tracing_max_lat_fops = { +static const struct file_operations tracing_max_lat_fops = { .open = tracing_open_generic, .read = tracing_max_lat_read, .write = tracing_max_lat_write, }; -static struct file_operations tracing_ctrl_fops = { +static const struct file_operations tracing_ctrl_fops = { .open = tracing_open_generic, .read = tracing_ctrl_read, .write = tracing_ctrl_write, }; -static struct file_operations set_tracer_fops = { +static const struct file_operations set_tracer_fops = { .open = tracing_open_generic, .read = tracing_set_trace_read, .write = tracing_set_trace_write, }; -static struct file_operations tracing_pipe_fops = { +static const struct file_operations tracing_pipe_fops = { .open = tracing_open_pipe, .poll = tracing_poll_pipe, .read = tracing_read_pipe, @@ -2953,13 +2953,13 @@ static struct file_operations tracing_pipe_fops = { .release = tracing_release_pipe, }; -static struct file_operations tracing_entries_fops = { +static const struct file_operations tracing_entries_fops = { .open = tracing_open_generic, .read = tracing_entries_read, .write = tracing_entries_write, }; -static struct file_operations tracing_mark_fops = { +static const struct file_operations tracing_mark_fops = { .open = tracing_open_generic, .write = tracing_mark_write, }; @@ -3240,7 +3240,7 @@ tracing_read_dyn_info(struct file *filp, char __user *ubuf, return r; } -static struct file_operations tracing_dyn_info_fops = { +static const struct file_operations tracing_dyn_info_fops = { .open = tracing_open_generic, .read = tracing_read_dyn_info, }; diff --git a/kernel/trace/trace_sysprof.c b/kernel/trace/trace_sysprof.c index c771af4e8f1a..91fd19c2149f 100644 --- a/kernel/trace/trace_sysprof.c +++ b/kernel/trace/trace_sysprof.c @@ -314,7 +314,7 @@ sysprof_sample_write(struct file *filp, const char __user *ubuf, return cnt; } -static struct file_operations sysprof_sample_fops = { +static const struct file_operations sysprof_sample_fops = { .read = sysprof_sample_read, .write = sysprof_sample_write, }; From 33b0c229e3abeae00493ed1d6f0b07191977a0a2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Mar 2009 11:45:43 -0500 Subject: [PATCH 238/413] tracing: move print of event format to separate file Impact: clean up Move the macro that creates the event format file to a separate header. This will allow the default ftrace events to use this same macro to create the formats to read those events. Signed-off-by: Steven Rostedt --- kernel/trace/trace_events_stage_2.h | 53 +-------------------------- kernel/trace/trace_format.h | 55 +++++++++++++++++++++++++++++ 2 files changed, 56 insertions(+), 52 deletions(-) create mode 100644 kernel/trace/trace_format.h diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index b1cebba1d9b4..d24a97e74aea 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -75,56 +75,5 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ #include -/* - * Setup the showing format of trace point. - * - * int - * ftrace_format_##call(struct trace_seq *s) - * { - * struct ftrace_raw_##call field; - * int ret; - * - * ret = trace_seq_printf(s, #type " " #item ";" - * " size:%d; offset:%d;\n", - * sizeof(field.type), - * offsetof(struct ftrace_raw_##call, - * item)); - * - * } - */ - -#undef TRACE_FIELD -#define TRACE_FIELD(type, item, assign) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%lu;\tsize:%lu;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item)); \ - if (!ret) \ - return 0; - - -#undef TRACE_FIELD_SPECIAL -#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ - ret = trace_seq_printf(s, "\tfield special:" #type_item ";\t" \ - "offset:%lu;\tsize:%lu;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item)); \ - if (!ret) \ - return 0; - -#undef TRACE_EVENT_FORMAT -#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ -int \ -ftrace_format_##call(struct trace_seq *s) \ -{ \ - struct ftrace_raw_##call field; \ - int ret; \ - \ - tstruct; \ - \ - trace_seq_printf(s, "\nprint fmt: \"%s\"\n", tpfmt); \ - \ - return ret; \ -} - +#include "trace_format.h" #include diff --git a/kernel/trace/trace_format.h b/kernel/trace/trace_format.h new file mode 100644 index 000000000000..53a6b1357116 --- /dev/null +++ b/kernel/trace/trace_format.h @@ -0,0 +1,55 @@ +/* + * Setup the showing format of trace point. + * + * int + * ftrace_format_##call(struct trace_seq *s) + * { + * struct ftrace_raw_##call field; + * int ret; + * + * ret = trace_seq_printf(s, #type " " #item ";" + * " size:%d; offset:%d;\n", + * sizeof(field.type), + * offsetof(struct ftrace_raw_##call, + * item)); + * + * } + */ + +#undef TRACE_STRUCT +#define TRACE_STRUCT(args...) args + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign) \ + ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ + "offset:%lu;\tsize:%lu;\n", \ + offsetof(typeof(field), item), \ + sizeof(field.item)); \ + if (!ret) \ + return 0; + + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ + ret = trace_seq_printf(s, "\tfield special:" #type_item ";\t" \ + "offset:%lu;\tsize:%lu;\n", \ + offsetof(typeof(field), item), \ + sizeof(field.item)); \ + if (!ret) \ + return 0; + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +int \ +ftrace_format_##call(struct trace_seq *s) \ +{ \ + struct ftrace_raw_##call field; \ + int ret; \ + \ + tstruct; \ + \ + trace_seq_printf(s, "\nprint fmt: \"%s\"\n", tpfmt); \ + \ + return ret; \ +} + From 770cb24345c0f6e0d47bd2b94aa6d67bea6f8b54 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 5 Mar 2009 21:35:29 -0500 Subject: [PATCH 239/413] tracing: add format files for ftrace default entries Impact: allow user apps to read binary format of basic ftrace entries Currently, only defined raw events export their formats so a binary reader can parse them. There's no reason that the default ftrace entries can't export their formats. This patch adds a subsystem called "ftrace" in the events directory that includes the ftrace entries for basic ftrace recorded items. These only have three files in the events directory: type : printf available_types : printf format : format for the event entry For example: # cat /debug/tracing/events/ftrace/wakeup/format name: wakeup ID: 3 format: field:unsigned char type; offset:0; size:1; field:unsigned char flags; offset:1; size:1; field:unsigned char preempt_count; offset:2; size:1; field:int pid; offset:4; size:4; field:int tgid; offset:8; size:4; field:unsigned int prev_pid; offset:12; size:4; field:unsigned char prev_prio; offset:16; size:1; field:unsigned char prev_state; offset:17; size:1; field:unsigned int next_pid; offset:20; size:4; field:unsigned char next_prio; offset:24; size:1; field:unsigned char next_state; offset:25; size:1; field:unsigned int next_cpu; offset:28; size:4; print fmt: "%u:%u:%u ==+ %u:%u:%u [%03u]" Signed-off-by: Steven Rostedt --- kernel/trace/Makefile | 1 + kernel/trace/trace_event_types.h | 165 +++++++++++++++++++++++++++++++ kernel/trace/trace_events.c | 12 ++- kernel/trace/trace_export.c | 81 +++++++++++++++ kernel/trace/trace_format.h | 2 +- 5 files changed, 255 insertions(+), 6 deletions(-) create mode 100644 kernel/trace/trace_event_types.h create mode 100644 kernel/trace/trace_export.c diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index c931fe0560cb..f44736c7574a 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -41,5 +41,6 @@ obj-$(CONFIG_WORKQUEUE_TRACER) += trace_workqueue.o obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o obj-$(CONFIG_EVENT_TRACER) += trace_events.o obj-$(CONFIG_EVENT_TRACER) += events.o +obj-$(CONFIG_EVENT_TRACER) += trace_export.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h new file mode 100644 index 000000000000..fb4eba166433 --- /dev/null +++ b/kernel/trace/trace_event_types.h @@ -0,0 +1,165 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM ftrace + +/* + * We cheat and use the proto type field as the ID + * and args as the entry type (minus 'struct') + */ +TRACE_EVENT_FORMAT(function, TRACE_FN, ftrace_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ip, ip) + TRACE_FIELD(unsigned long, parent_ip, parent_ip) + ), + TPRAWFMT(" %lx <-- %lx") +); + +TRACE_EVENT_FORMAT(funcgraph_entry, TRACE_GRAPH_ENT, + ftrace_graph_ent_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, graph_ent.func, func) + TRACE_FIELD(int, graph_ent.depth, depth) + ), + TPRAWFMT("--> %lx (%d)") +); + +TRACE_EVENT_FORMAT(funcgraph_exit, TRACE_GRAPH_RET, + ftrace_graph_ret_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ret.func, func) + TRACE_FIELD(int, ret.depth, depth) + ), + TPRAWFMT("<-- %lx (%d)") +); + +TRACE_EVENT_FORMAT(wakeup, TRACE_WAKE, ctx_switch_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned int, prev_pid, prev_pid) + TRACE_FIELD(unsigned char, prev_prio, prev_prio) + TRACE_FIELD(unsigned char, prev_state, prev_state) + TRACE_FIELD(unsigned int, next_pid, next_pid) + TRACE_FIELD(unsigned char, next_prio, next_prio) + TRACE_FIELD(unsigned char, next_state, next_state) + TRACE_FIELD(unsigned int, next_cpu, next_cpu) + ), + TPRAWFMT("%u:%u:%u ==+ %u:%u:%u [%03u]") +); + +TRACE_EVENT_FORMAT(context_switch, TRACE_CTX, ctx_switch_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned int, prev_pid, prev_pid) + TRACE_FIELD(unsigned char, prev_prio, prev_prio) + TRACE_FIELD(unsigned char, prev_state, prev_state) + TRACE_FIELD(unsigned int, next_pid, next_pid) + TRACE_FIELD(unsigned char, next_prio, next_prio) + TRACE_FIELD(unsigned char, next_state, next_state) + TRACE_FIELD(unsigned int, next_cpu, next_cpu) + ), + TPRAWFMT("%u:%u:%u ==+ %u:%u:%u [%03u]") +); + +TRACE_EVENT_FORMAT(special, TRACE_SPECIAL, special_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, arg1, arg1) + TRACE_FIELD(unsigned long, arg2, arg2) + TRACE_FIELD(unsigned long, arg3, arg3) + ), + TPRAWFMT("(%08lx) (%08lx) (%08lx)") +); + +/* + * Stack-trace entry: + */ + +/* #define FTRACE_STACK_ENTRIES 8 */ + +TRACE_EVENT_FORMAT(kernel_stack, TRACE_STACK, stack_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, caller[0], stack0) + TRACE_FIELD(unsigned long, caller[1], stack1) + TRACE_FIELD(unsigned long, caller[2], stack2) + TRACE_FIELD(unsigned long, caller[3], stack3) + TRACE_FIELD(unsigned long, caller[4], stack4) + TRACE_FIELD(unsigned long, caller[5], stack5) + TRACE_FIELD(unsigned long, caller[6], stack6) + TRACE_FIELD(unsigned long, caller[7], stack7) + ), + TPRAWFMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" + "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") +); + +TRACE_EVENT_FORMAT(user_stack, TRACE_USER_STACK, userstack_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, caller[0], stack0) + TRACE_FIELD(unsigned long, caller[1], stack1) + TRACE_FIELD(unsigned long, caller[2], stack2) + TRACE_FIELD(unsigned long, caller[3], stack3) + TRACE_FIELD(unsigned long, caller[4], stack4) + TRACE_FIELD(unsigned long, caller[5], stack5) + TRACE_FIELD(unsigned long, caller[6], stack6) + TRACE_FIELD(unsigned long, caller[7], stack7) + ), + TPRAWFMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" + "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") +); + +TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ip, ip) + TRACE_FIELD(unsigned int, depth, depth) + TRACE_FIELD_ZERO_CHAR(buf) + ), + TPRAWFMT("%08lx (%d) %s") +); + +TRACE_EVENT_FORMAT(branch, TRACE_BRANCH, trace_branch, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned int, line, line) + TRACE_FIELD_SPECIAL(char func[TRACE_FUNC_SIZE+1], func, func) + TRACE_FIELD_SPECIAL(char file[TRACE_FUNC_SIZE+1], file, file) + TRACE_FIELD(char, correct, correct) + ), + TPRAWFMT("%u:%s:%s (%u)") +); + +TRACE_EVENT_FORMAT(hw_branch, TRACE_HW_BRANCHES, hw_branch_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(u64, from, from) + TRACE_FIELD(u64, to, to) + ), + TPRAWFMT("from: %llx to: %llx") +); + +TRACE_EVENT_FORMAT(power, TRACE_POWER, trace_power, ignore, + TRACE_STRUCT( + TRACE_FIELD(ktime_t, state_data.stamp, stamp) + TRACE_FIELD(ktime_t, state_data.end, end) + TRACE_FIELD(int, state_data.type, type) + TRACE_FIELD(int, state_data.state, state) + ), + TPRAWFMT("%llx->%llx type:%u state:%u") +); + +TRACE_EVENT_FORMAT(kmem_alloc, TRACE_KMEM_ALLOC, kmemtrace_alloc_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(enum kmemtrace_type_id, type_id, type_id) + TRACE_FIELD(unsigned long, call_site, call_site) + TRACE_FIELD(const void *, ptr, ptr) + TRACE_FIELD(size_t, bytes_req, bytes_req) + TRACE_FIELD(size_t, bytes_alloc, bytes_alloc) + TRACE_FIELD(gfp_t, gfp_flags, gfp_flags) + TRACE_FIELD(int, node, node) + ), + TPRAWFMT("type:%u call_site:%lx ptr:%p req:%lu alloc:%lu" + " flags:%x node:%d") +); + +TRACE_EVENT_FORMAT(kmem_free, TRACE_KMEM_FREE, kmemtrace_free_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(enum kmemtrace_type_id, type_id, type_id) + TRACE_FIELD(unsigned long, call_site, call_site) + TRACE_FIELD(const void *, ptr, ptr) + ), + TPRAWFMT("type:%u call_site:%lx ptr:%p") +); + +#undef TRACE_SYSTEM diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 210e71ff82db..4488d90e75ef 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -656,11 +656,13 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) return -1; } - entry = debugfs_create_file("enable", 0644, call->dir, call, - &ftrace_enable_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'%s/enable' entry\n", call->name); + if (call->regfunc) { + entry = debugfs_create_file("enable", 0644, call->dir, call, + &ftrace_enable_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/enable' entry\n", call->name); + } /* Only let type be writable, if we can change it */ entry = debugfs_create_file("type", diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c new file mode 100644 index 000000000000..0fb7be73e31c --- /dev/null +++ b/kernel/trace/trace_export.c @@ -0,0 +1,81 @@ +/* + * trace_export.c - export basic ftrace utilities to user space + * + * Copyright (C) 2009 Steven Rostedt + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "trace_output.h" + +#include "trace_format.h" + +#undef TRACE_FIELD_ZERO_CHAR +#define TRACE_FIELD_ZERO_CHAR(item) \ + ret = trace_seq_printf(s, "\tfield: char " #item ";\t" \ + "offset:%lu;\tsize:0;\n", \ + offsetof(typeof(field), item)); \ + if (!ret) \ + return 0; + + +#undef TPRAWFMT +#define TPRAWFMT(args...) args + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +static int \ +ftrace_format_##call(struct trace_seq *s) \ +{ \ + struct args field; \ + int ret; \ + \ + tstruct; \ + \ + trace_seq_printf(s, "\nprint fmt: \"%s\"\n", tpfmt); \ + \ + return ret; \ +} + +#include "trace_event_types.h" + +#undef TRACE_ZERO_CHAR +#define TRACE_ZERO_CHAR(arg) + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign)\ + entry->item = assign; + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign)\ + entry->item = assign; + +#undef TPCMD +#define TPCMD(cmd...) cmd + +#undef TRACE_ENTRY +#define TRACE_ENTRY entry + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ + cmd; + +#undef TRACE_EVENT_FORMAT +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ + \ +static struct ftrace_event_call __used \ +__attribute__((__aligned__(4))) \ +__attribute__((section("_ftrace_events"))) event_##call = { \ + .name = #call, \ + .id = proto, \ + .system = __stringify(TRACE_SYSTEM), \ + .show_format = ftrace_format_##call, \ +} +#include "trace_event_types.h" diff --git a/kernel/trace/trace_format.h b/kernel/trace/trace_format.h index 53a6b1357116..03f9a4c165ca 100644 --- a/kernel/trace/trace_format.h +++ b/kernel/trace/trace_format.h @@ -40,7 +40,7 @@ #undef TRACE_EVENT_FORMAT #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ -int \ +static int \ ftrace_format_##call(struct trace_seq *s) \ { \ struct ftrace_raw_##call field; \ From 422d3c7a577b15e1384c9d4e72a9540896b685fa Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 6 Mar 2009 10:40:53 +0900 Subject: [PATCH 240/413] tracing: current tip/master can't enable ftrace After commit 40ada30f9621fbd831ac2437b9a2a399aad34b00, "make menuconfig" doesn't display "Tracer" item. Following modification restores it. --- kernel/trace/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 5d733da5345a..058d949a3214 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -61,6 +61,7 @@ config TRACING_SUPPORT bool depends on TRACE_IRQFLAGS_SUPPORT depends on STACKTRACE_SUPPORT + default y if TRACING_SUPPORT From 10dd3ebe213c31bff14b4dae3c5d32a76b1fad7c Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 6 Mar 2009 15:29:04 +0900 Subject: [PATCH 241/413] tracing: fix deadlock when setting set_ftrace_pid Impact: fix deadlock while using set_ftrace_pid Reproducer: # cd /sys/kernel/debug/tracing # echo $$ > set_ftrace_pid then, console becomes hung. Details: when writing set_ftracepid, kernel callstack is following ftrace_pid_write() mutex_lock(&ftrace_lock); ftrace_update_pid_func() mutex_lock(&ftrace_lock); mutex_unlock(&ftrace_lock); mutex_unlock(&ftrace_lock); then, system always deadlocks when ftrace_pid_write() is called. In past days, ftrace_pid_write() used ftrace_start_lock, but commit e6ea44e9b4c12325337cd1c06103cd515a1c02b2 consolidated ftrace_start_lock to ftrace_lock. Signed-off-by: KOSAKI Motohiro Reviewed-by: Lai Jiangshan Cc: Steven Rostedt LKML-Reference: <20090306151155.0778.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index d7a06a0d9447..d33d306bdcf4 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -218,10 +218,8 @@ static void ftrace_update_pid_func(void) { ftrace_func_t func; - mutex_lock(&ftrace_lock); - if (ftrace_trace_function == ftrace_stub) - goto out; + return; func = ftrace_trace_function; @@ -238,9 +236,6 @@ static void ftrace_update_pid_func(void) #else __ftrace_trace_function = func; #endif - - out: - mutex_unlock(&ftrace_lock); } /* set when tracing only a pid */ From af438c0f114b6f731b923b5c07150f6159471502 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 6 Mar 2009 12:47:08 +0100 Subject: [PATCH 242/413] tracing, power-trace: make it build even if the power-tracer is turned off Impact: build fix The 'struct power_trace' definition is needed (for the event tracer) even if the power-tracer plugin is turned off in the .config. Cc: Steven Rostedt LKML-Reference: <20090306104106.GF31042@elte.hu> Signed-off-by: Ingo Molnar --- include/trace/power.h | 2 -- 1 file changed, 2 deletions(-) diff --git a/include/trace/power.h b/include/trace/power.h index 2c733e58e89c..38aca537e497 100644 --- a/include/trace/power.h +++ b/include/trace/power.h @@ -11,12 +11,10 @@ enum { }; struct power_trace { -#ifdef CONFIG_POWER_TRACER ktime_t stamp; ktime_t end; int type; int state; -#endif }; DECLARE_TRACE(power_start, From 0e39ac444636ff5be39b26f1cb56d79594654dda Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Fri, 6 Mar 2009 10:35:52 -0500 Subject: [PATCH 243/413] tracing, Text Edit Lock - Architecture Independent Code This is an architecture independant synchronization around kernel text modifications through use of a global mutex. A mutex has been chosen so that kprobes, the main user of this, can sleep during memory allocation between the memory read of the instructions it must replace and the memory write of the breakpoint. Other user of this interface: immediate values. Paravirt and alternatives are always done when SMP is inactive, so there is no need to use locks. Signed-off-by: Mathieu Desnoyers LKML-Reference: <49B142D8.7020601@redhat.com> Signed-off-by: Ingo Molnar --- include/linux/memory.h | 6 ++++++ mm/memory.c | 10 ++++++++++ 2 files changed, 16 insertions(+) diff --git a/include/linux/memory.h b/include/linux/memory.h index 3fdc10806d31..86a6c0f0518d 100644 --- a/include/linux/memory.h +++ b/include/linux/memory.h @@ -99,4 +99,10 @@ enum mem_add_context { BOOT, HOTPLUG }; #define hotplug_memory_notifier(fn, pri) do { } while (0) #endif +/* + * Kernel text modification mutex, used for code patching. Users of this lock + * can sleep. + */ +extern struct mutex text_mutex; + #endif /* _LINUX_MEMORY_H_ */ diff --git a/mm/memory.c b/mm/memory.c index baa999e87cd2..05fab3bc5b4b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -48,6 +48,8 @@ #include #include #include +#include +#include #include #include #include @@ -99,6 +101,14 @@ int randomize_va_space __read_mostly = 2; #endif +/* + * mutex protecting text section modification (dynamic code patching). + * some users need to sleep (allocating memory...) while they hold this lock. + * + * NOT exported to modules - patching kernel text is a really delicate matter. + */ +DEFINE_MUTEX(text_mutex); + static int __init disable_randmaps(char *s) { randomize_va_space = 0; From 4460fdad85becd569f11501ad5b91814814335ff Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Fri, 6 Mar 2009 10:36:38 -0500 Subject: [PATCH 244/413] tracing, Text Edit Lock - kprobes architecture independent support Use the mutual exclusion provided by the text edit lock in the kprobes code. It allows coherent manipulation of the kernel code by other subsystems. Changelog: Move the kernel_text_lock/unlock out of the for loops. Use text_mutex directly instead of a function. Remove whitespace modifications. (note : kprobes_mutex is always taken outside of text_mutex) Signed-off-by: Mathieu Desnoyers Acked-by: Ananth N Mavinakayanahalli Acked-by: Masami Hiramatsu LKML-Reference: <49B14306.2080202@redhat.com> Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 7ba8cd9845cb..479d4d5672f9 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -43,6 +43,7 @@ #include #include #include +#include #include #include @@ -699,9 +700,10 @@ int __kprobes register_kprobe(struct kprobe *p) goto out; } + mutex_lock(&text_mutex); ret = arch_prepare_kprobe(p); if (ret) - goto out; + goto out_unlock_text; INIT_HLIST_NODE(&p->hlist); hlist_add_head_rcu(&p->hlist, @@ -710,6 +712,8 @@ int __kprobes register_kprobe(struct kprobe *p) if (kprobe_enabled) arch_arm_kprobe(p); +out_unlock_text: + mutex_unlock(&text_mutex); out: mutex_unlock(&kprobe_mutex); @@ -746,8 +750,11 @@ static int __kprobes __unregister_kprobe_top(struct kprobe *p) * enabled and not gone - otherwise, the breakpoint would * already have been removed. We save on flushing icache. */ - if (kprobe_enabled && !kprobe_gone(old_p)) + if (kprobe_enabled && !kprobe_gone(old_p)) { + mutex_lock(&text_mutex); arch_disarm_kprobe(p); + mutex_unlock(&text_mutex); + } hlist_del_rcu(&old_p->hlist); } else { if (p->break_handler && !kprobe_gone(p)) @@ -1280,12 +1287,14 @@ static void __kprobes enable_all_kprobes(void) if (kprobe_enabled) goto already_enabled; + mutex_lock(&text_mutex); for (i = 0; i < KPROBE_TABLE_SIZE; i++) { head = &kprobe_table[i]; hlist_for_each_entry_rcu(p, node, head, hlist) if (!kprobe_gone(p)) arch_arm_kprobe(p); } + mutex_unlock(&text_mutex); kprobe_enabled = true; printk(KERN_INFO "Kprobes globally enabled\n"); @@ -1310,6 +1319,7 @@ static void __kprobes disable_all_kprobes(void) kprobe_enabled = false; printk(KERN_INFO "Kprobes globally disabled\n"); + mutex_lock(&text_mutex); for (i = 0; i < KPROBE_TABLE_SIZE; i++) { head = &kprobe_table[i]; hlist_for_each_entry_rcu(p, node, head, hlist) { @@ -1318,6 +1328,7 @@ static void __kprobes disable_all_kprobes(void) } } + mutex_unlock(&text_mutex); mutex_unlock(&kprobe_mutex); /* Allow all currently running kprobes to complete */ synchronize_sched(); From 3945dab45aa8c89014893bfa8eb1e1661a409cef Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 6 Mar 2009 10:37:22 -0500 Subject: [PATCH 245/413] tracing, Text Edit Lock - SMP alternatives support Use the mutual exclusion provided by the text edit lock in alternatives code. Since alternative_smp_* will be called from module init code, etc, we'd better protect it from other subsystems. Signed-off-by: Masami Hiramatsu LKML-Reference: <49B14332.9030109@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/alternative.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index 4c80f1557433..092a7b8be68d 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -5,6 +5,7 @@ #include #include #include +#include #include #include #include @@ -226,6 +227,7 @@ static void alternatives_smp_lock(u8 **start, u8 **end, u8 *text, u8 *text_end) { u8 **ptr; + mutex_lock(&text_mutex); for (ptr = start; ptr < end; ptr++) { if (*ptr < text) continue; @@ -234,6 +236,7 @@ static void alternatives_smp_lock(u8 **start, u8 **end, u8 *text, u8 *text_end) /* turn DS segment override prefix into lock prefix */ text_poke(*ptr, ((unsigned char []){0xf0}), 1); }; + mutex_unlock(&text_mutex); } static void alternatives_smp_unlock(u8 **start, u8 **end, u8 *text, u8 *text_end) @@ -243,6 +246,7 @@ static void alternatives_smp_unlock(u8 **start, u8 **end, u8 *text, u8 *text_end if (noreplace_smp) return; + mutex_lock(&text_mutex); for (ptr = start; ptr < end; ptr++) { if (*ptr < text) continue; @@ -251,6 +255,7 @@ static void alternatives_smp_unlock(u8 **start, u8 **end, u8 *text, u8 *text_end /* turn lock prefix into DS segment override prefix */ text_poke(*ptr, ((unsigned char []){0x3E}), 1); }; + mutex_unlock(&text_mutex); } struct smp_alt_module { From 78ff7fae04554b49d29226ed12536268c2500d1f Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 6 Mar 2009 10:37:54 -0500 Subject: [PATCH 246/413] x86: implement atomic text_poke() via fixmap Use fixmaps instead of vmap/vunmap in text_poke() for avoiding page allocation and delayed unmapping. At the result of above change, text_poke() becomes atomic and can be called from stop_machine() etc. Signed-off-by: Masami Hiramatsu Acked-by: Mathieu Desnoyers LKML-Reference: <49B14352.2040705@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/fixmap.h | 2 ++ arch/x86/kernel/alternative.c | 24 +++++++++++++++--------- 2 files changed, 17 insertions(+), 9 deletions(-) diff --git a/arch/x86/include/asm/fixmap.h b/arch/x86/include/asm/fixmap.h index 63a79c77d220..81937a5dc77c 100644 --- a/arch/x86/include/asm/fixmap.h +++ b/arch/x86/include/asm/fixmap.h @@ -111,6 +111,8 @@ enum fixed_addresses { #ifdef CONFIG_PARAVIRT FIX_PARAVIRT_BOOTMAP, #endif + FIX_TEXT_POKE0, /* reserve 2 pages for text_poke() */ + FIX_TEXT_POKE1, __end_of_permanent_fixed_addresses, #ifdef CONFIG_PROVIDE_OHCI1394_DMA_INIT FIX_OHCI1394_BASE, diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index 092a7b8be68d..2d903b760ddb 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -13,7 +13,9 @@ #include #include #include +#include #include +#include #define MAX_PATCH_LEN (255-1) @@ -505,15 +507,16 @@ void *text_poke_early(void *addr, const void *opcode, size_t len) * It means the size must be writable atomically and the address must be aligned * in a way that permits an atomic write. It also makes sure we fit on a single * page. + * + * Note: Must be called under text_mutex. */ void *__kprobes text_poke(void *addr, const void *opcode, size_t len) { + unsigned long flags; char *vaddr; - int nr_pages = 2; struct page *pages[2]; int i; - might_sleep(); if (!core_kernel_text((unsigned long)addr)) { pages[0] = vmalloc_to_page(addr); pages[1] = vmalloc_to_page(addr + PAGE_SIZE); @@ -523,14 +526,17 @@ void *__kprobes text_poke(void *addr, const void *opcode, size_t len) pages[1] = virt_to_page(addr + PAGE_SIZE); } BUG_ON(!pages[0]); - if (!pages[1]) - nr_pages = 1; - vaddr = vmap(pages, nr_pages, VM_MAP, PAGE_KERNEL); - BUG_ON(!vaddr); - local_irq_disable(); + set_fixmap(FIX_TEXT_POKE0, page_to_phys(pages[0])); + if (pages[1]) + set_fixmap(FIX_TEXT_POKE1, page_to_phys(pages[1])); + vaddr = (char *)fix_to_virt(FIX_TEXT_POKE0); + local_irq_save(flags); memcpy(&vaddr[(unsigned long)addr & ~PAGE_MASK], opcode, len); - local_irq_enable(); - vunmap(vaddr); + local_irq_restore(flags); + clear_fixmap(FIX_TEXT_POKE0); + if (pages[1]) + clear_fixmap(FIX_TEXT_POKE1); + local_flush_tlb(); sync_core(); /* Could also do a CLFLUSH here to speed up CPU recovery; but that causes hangs on some VIA CPUs. */ From 1427cdf0592368bdec57276edaf714040ee8744f Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 6 Mar 2009 17:21:47 +0100 Subject: [PATCH 247/413] tracing: infrastructure for supporting binary record Impact: save on memory for tracing Current tracers are typically using a struct(like struct ftrace_entry, struct ctx_switch_entry, struct special_entr etc...)to record a binary event. These structs can only record a their own kind of events. A new kind of tracer need a new struct and a lot of code too handle it. So we need a generic binary record for events. This infrastructure is for this purpose. [fweisbec@gmail.com: rebase against latest -tip, make it safe while sched tracing as reported by Steven Rostedt] Signed-off-by: Lai Jiangshan Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt LKML-Reference: <1236356510-8381-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 3 ++ kernel/trace/Kconfig | 6 +++ kernel/trace/Makefile | 1 + kernel/trace/trace.c | 56 +++++++++++++++++++++++ kernel/trace/trace.h | 12 +++++ kernel/trace/trace_bprintk.c | 87 ++++++++++++++++++++++++++++++++++++ kernel/trace/trace_output.c | 75 +++++++++++++++++++++++++++++++ 7 files changed, 240 insertions(+) create mode 100644 kernel/trace/trace_bprintk.c diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 498769425eb2..1c9cdca02580 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -223,6 +223,9 @@ extern int ftrace_make_nop(struct module *mod, */ extern int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr); +#ifdef CONFIG_TRACE_BPRINTK +extern int trace_vbprintk(unsigned long ip, const char *fmt, va_list args); +#endif /* May be defined in arch */ extern int ftrace_arch_read_dyn_info(char *buf, int size); diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 058d949a3214..ad8d3617d0a6 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -97,6 +97,12 @@ config FUNCTION_GRAPH_TRACER This is done by setting the current return address on the current task structure into a stack of calls. +config TRACE_BPRINTK + bool "Binary printk for tracing" + default y + depends on TRACING + select BINARY_PRINTF + config IRQSOFF_TRACER bool "Interrupts-off Latency Tracer" default n diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index f44736c7574a..46557ef4c379 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -22,6 +22,7 @@ obj-$(CONFIG_TRACING) += trace.o obj-$(CONFIG_TRACING) += trace_clock.o obj-$(CONFIG_TRACING) += trace_output.o obj-$(CONFIG_TRACING) += trace_stat.o +obj-$(CONFIG_TRACE_BPRINTK) += trace_bprintk.o obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o obj-$(CONFIG_SYSPROF_TRACER) += trace_sysprof.o obj-$(CONFIG_FUNCTION_TRACER) += trace_functions.o diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e6144acf2b75..ff53509e19f8 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3792,6 +3792,62 @@ int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) } EXPORT_SYMBOL_GPL(__ftrace_vprintk); +/** + * trace_vbprintk - write binary msg to tracing buffer + * + * Caller must insure @fmt are valid when msg is in tracing buffer. + */ +int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) +{ + static DEFINE_SPINLOCK(trace_buf_lock); + static u32 trace_buf[TRACE_BUF_SIZE]; + + struct ring_buffer_event *event; + struct trace_array *tr = &global_trace; + struct trace_array_cpu *data; + struct bprintk_entry *entry; + unsigned long flags; + int resched; + int cpu, len = 0, size, pc; + + if (tracing_disabled || !trace_bprintk_enable) + return 0; + + pc = preempt_count(); + resched = ftrace_preempt_disable(); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + + if (unlikely(atomic_read(&data->disabled))) + goto out; + + spin_lock_irqsave(&trace_buf_lock, flags); + len = vbin_printf(trace_buf, TRACE_BUF_SIZE, fmt, args); + + if (len > TRACE_BUF_SIZE || len < 0) + goto out_unlock; + + size = sizeof(*entry) + sizeof(u32) * len; + event = trace_buffer_lock_reserve(tr, TRACE_BPRINTK, size, flags, pc); + if (!event) + goto out_unlock; + entry = ring_buffer_event_data(event); + entry->ip = ip; + entry->fmt = fmt; + + memcpy(entry->buf, trace_buf, sizeof(u32) * len); + ring_buffer_unlock_commit(tr->buffer, event); + +out_unlock: + spin_unlock_irqrestore(&trace_buf_lock, flags); + +out: + ftrace_preempt_enable(resched); + + return len; +} +EXPORT_SYMBOL_GPL(trace_vbprintk); + static int trace_panic_handler(struct notifier_block *this, unsigned long event, void *unused) { diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 8beff03fda68..0f5077f8f957 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -20,6 +20,7 @@ enum trace_type { TRACE_WAKE, TRACE_STACK, TRACE_PRINT, + TRACE_BPRINTK, TRACE_SPECIAL, TRACE_MMIO_RW, TRACE_MMIO_MAP, @@ -124,6 +125,16 @@ struct print_entry { char buf[]; }; +struct bprintk_entry { + struct trace_entry ent; + unsigned long ip; + const char *fmt; + u32 buf[]; +}; +#ifdef CONFIG_TRACE_BPRINTK +extern int trace_bprintk_enable; +#endif + #define TRACE_OLD_SIZE 88 struct trace_field_cont { @@ -285,6 +296,7 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct stack_entry, TRACE_STACK); \ IF_ASSIGN(var, ent, struct userstack_entry, TRACE_USER_STACK);\ IF_ASSIGN(var, ent, struct print_entry, TRACE_PRINT); \ + IF_ASSIGN(var, ent, struct bprintk_entry, TRACE_BPRINTK);\ IF_ASSIGN(var, ent, struct special_entry, 0); \ IF_ASSIGN(var, ent, struct trace_mmiotrace_rw, \ TRACE_MMIO_RW); \ diff --git a/kernel/trace/trace_bprintk.c b/kernel/trace/trace_bprintk.c new file mode 100644 index 000000000000..1f8e532c3fb9 --- /dev/null +++ b/kernel/trace/trace_bprintk.c @@ -0,0 +1,87 @@ +/* + * trace binary printk + * + * Copyright (C) 2008 Lai Jiangshan + * + */ +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "trace.h" + +/* binary printk basic */ +static DEFINE_MUTEX(btrace_mutex); +static int btrace_metadata_count; + +static inline void lock_btrace(void) +{ + mutex_lock(&btrace_mutex); +} + +static inline void unlock_btrace(void) +{ + mutex_unlock(&btrace_mutex); +} + +static void get_btrace_metadata(void) +{ + lock_btrace(); + btrace_metadata_count++; + unlock_btrace(); +} + +static void put_btrace_metadata(void) +{ + lock_btrace(); + btrace_metadata_count--; + unlock_btrace(); +} + +/* events tracer */ +int trace_bprintk_enable; + +static void start_bprintk_trace(struct trace_array *tr) +{ + get_btrace_metadata(); + tracing_reset_online_cpus(tr); + trace_bprintk_enable = 1; +} + +static void stop_bprintk_trace(struct trace_array *tr) +{ + trace_bprintk_enable = 0; + tracing_reset_online_cpus(tr); + put_btrace_metadata(); +} + +static int init_bprintk_trace(struct trace_array *tr) +{ + start_bprintk_trace(tr); + return 0; +} + +static struct tracer bprintk_trace __read_mostly = +{ + .name = "events", + .init = init_bprintk_trace, + .reset = stop_bprintk_trace, + .start = start_bprintk_trace, + .stop = stop_bprintk_trace, +}; + +static __init int init_bprintk(void) +{ + return register_tracer(&bprintk_trace); +} + +device_initcall(init_bprintk); diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 306fef84c503..4ab71201862e 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -53,6 +53,26 @@ trace_seq_printf(struct trace_seq *s, const char *fmt, ...) return len; } +static int +trace_seq_bprintf(struct trace_seq *s, const char *fmt, const u32 *binary) +{ + int len = (PAGE_SIZE - 1) - s->len; + int ret; + + if (!len) + return 0; + + ret = bstr_printf(s->buffer + s->len, len, fmt, binary); + + /* If we can't write it all, don't bother writing anything */ + if (ret >= len) + return 0; + + s->len += ret; + + return len; +} + /** * trace_seq_puts - trace sequence printing of simple string * @s: trace sequence descriptor @@ -855,6 +875,60 @@ static struct trace_event trace_print_event = { .raw = trace_print_raw, }; +/* TRACE_BPRINTK */ +static enum print_line_t +trace_bprintk_print(struct trace_iterator *iter, int flags) +{ + struct trace_entry *entry = iter->ent; + struct trace_seq *s = &iter->seq; + struct bprintk_entry *field; + + trace_assign_type(field, entry); + + if (!seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if (!trace_seq_puts(s, ": ")) + goto partial; + + if (!trace_seq_bprintf(s, field->fmt, field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static enum print_line_t +trace_bprintk_raw(struct trace_iterator *iter, int flags) +{ + struct trace_entry *entry = iter->ent; + struct trace_seq *s = &iter->seq; + struct bprintk_entry *field; + + trace_assign_type(field, entry); + + if (!trace_seq_printf(s, ": %lx : ", field->ip)) + goto partial; + + if (!trace_seq_bprintf(s, field->fmt, field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static struct trace_event trace_bprintk_event = { + .type = TRACE_BPRINTK, + .trace = trace_bprintk_print, + .raw = trace_bprintk_raw, + .hex = trace_nop_print, + .binary = trace_nop_print, +}; + static struct trace_event *events[] __initdata = { &trace_fn_event, &trace_ctx_event, @@ -863,6 +937,7 @@ static struct trace_event *events[] __initdata = { &trace_stack_event, &trace_user_stack_event, &trace_print_event, + &trace_bprintk_event, NULL }; From 1ba28e02a18cbdbea123836f6c98efb09cbf59ec Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 6 Mar 2009 17:21:48 +0100 Subject: [PATCH 248/413] tracing: add trace_bprintk() Impact: add a generic printk() for tracing, like trace_printk() trace_bprintk() uses the infrastructure to record events on ring_buffer. [ fweisbec@gmail.com: ported to latest -tip, made it work if !CONFIG_MODULES, never free the format strings from modules because we can't keep track of them and conditionnaly create the ftrace format strings section (reported by Steven Rostedt) ] Signed-off-by: Lai Jiangshan Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt LKML-Reference: <1236356510-8381-4-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- include/asm-generic/vmlinux.lds.h | 9 ++++ include/linux/ftrace.h | 21 ++++++++ include/linux/module.h | 5 ++ kernel/module.c | 6 +++ kernel/trace/trace.c | 15 ++++++ kernel/trace/trace_bprintk.c | 87 +++++++++++++++++++++++++++---- 6 files changed, 133 insertions(+), 10 deletions(-) diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 0add6b28c366..48ade3168b13 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -69,6 +69,14 @@ #define FTRACE_EVENTS() #endif +#ifdef CONFIG_TRACING +#define TRACE_PRINTKS() VMLINUX_SYMBOL(__start___trace_bprintk_fmt) = .; \ + *(__trace_printk_fmt) /* Trace_printk fmt' pointer */ \ + VMLINUX_SYMBOL(__stop___trace_bprintk_fmt) = .; +#else +#define TRACE_PRINTKS() +#endif + /* .data section */ #define DATA_DATA \ *(.data) \ @@ -100,6 +108,7 @@ *(__vermagic) /* Kernel version magic */ \ *(__markers_strings) /* Markers: strings */ \ *(__tracepoints_strings)/* Tracepoints: strings */ \ + TRACE_PRINTKS() \ } \ \ .rodata1 : AT(ADDR(.rodata1) - LOAD_OFFSET) { \ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 1c9cdca02580..1cc8ca453a9b 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -225,6 +225,27 @@ extern int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr); #ifdef CONFIG_TRACE_BPRINTK extern int trace_vbprintk(unsigned long ip, const char *fmt, va_list args); +extern int __trace_bprintk(unsigned long ip, const char *fmt, ...) + __attribute__ ((format (printf, 2, 3))); + +static inline void ____trace_bprintk_check_format(const char *fmt, ...) + __attribute__ ((format (printf, 1, 2))); +static inline void ____trace_bprintk_check_format(const char *fmt, ...) {} +#define __trace_bprintk_check_format(fmt, args...) \ +do { \ + if (0) \ + ____trace_bprintk_check_format(fmt, ##args); \ +} while (0) + +#define trace_bprintk(fmt, args...) \ +do { \ + static char *__attribute__((section("__trace_bprintk_fmt"))) \ + trace_bprintk_fmt = fmt; \ + __trace_bprintk_check_format(fmt, ##args); \ + __trace_bprintk(_THIS_IP_, trace_bprintk_fmt, ##args); \ +} while (0) +#else +#define trace_bprintk trace_printk #endif /* May be defined in arch */ diff --git a/include/linux/module.h b/include/linux/module.h index 145a75528cc1..8cbec972d8e7 100644 --- a/include/linux/module.h +++ b/include/linux/module.h @@ -329,6 +329,11 @@ struct module unsigned int num_tracepoints; #endif +#ifdef CONFIG_TRACE_BPRINTK + const char **trace_bprintk_fmt_start; + unsigned int num_trace_bprintk_fmt; +#endif + #ifdef CONFIG_MODULE_UNLOAD /* What modules depend on me? */ struct list_head modules_which_use_me; diff --git a/kernel/module.c b/kernel/module.c index 22d7379709da..2dece104f9a1 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2158,6 +2158,12 @@ static noinline struct module *load_module(void __user *umod, &mod->num_tracepoints); #endif +#ifdef CONFIG_TRACE_BPRINTK + mod->trace_bprintk_fmt_start = section_objs(hdr, sechdrs, secstrings, + "__trace_bprintk_fmt", sizeof(char *), + &mod->num_trace_bprintk_fmt); +#endif + #ifdef CONFIG_MODVERSIONS if ((mod->num_syms && !mod->crcs) || (mod->num_gpl_syms && !mod->gpl_crcs) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index ff53509e19f8..46b3cd7a5752 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3848,6 +3848,21 @@ int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) } EXPORT_SYMBOL_GPL(trace_vbprintk); +int __trace_bprintk(unsigned long ip, const char *fmt, ...) +{ + int ret; + va_list ap; + + if (!fmt) + return 0; + + va_start(ap, fmt); + ret = trace_vbprintk(ip, fmt, ap); + va_end(ap); + return ret; +} +EXPORT_SYMBOL_GPL(__trace_bprintk); + static int trace_panic_handler(struct notifier_block *this, unsigned long event, void *unused) { diff --git a/kernel/trace/trace_bprintk.c b/kernel/trace/trace_bprintk.c index 1f8e532c3fb9..f4c245a5cd33 100644 --- a/kernel/trace/trace_bprintk.c +++ b/kernel/trace/trace_bprintk.c @@ -19,9 +19,21 @@ #include "trace.h" +#ifdef CONFIG_MODULES + /* binary printk basic */ static DEFINE_MUTEX(btrace_mutex); -static int btrace_metadata_count; +/* + * modules trace_bprintk()'s formats are autosaved in struct trace_bprintk_fmt + * which are queued on trace_bprintk_fmt_list. + */ +static LIST_HEAD(trace_bprintk_fmt_list); + +struct trace_bprintk_fmt { + struct list_head list; + char fmt[0]; +}; + static inline void lock_btrace(void) { @@ -33,26 +45,75 @@ static inline void unlock_btrace(void) mutex_unlock(&btrace_mutex); } -static void get_btrace_metadata(void) + +static inline struct trace_bprintk_fmt *lookup_format(const char *fmt) { + struct trace_bprintk_fmt *pos; + list_for_each_entry(pos, &trace_bprintk_fmt_list, list) { + if (!strcmp(pos->fmt, fmt)) + return pos; + } + return NULL; +} + +static +void hold_module_trace_bprintk_format(const char **start, const char **end) +{ + const char **iter; lock_btrace(); - btrace_metadata_count++; + for (iter = start; iter < end; iter++) { + struct trace_bprintk_fmt *tb_fmt = lookup_format(*iter); + if (tb_fmt) { + *iter = tb_fmt->fmt; + continue; + } + + tb_fmt = kmalloc(offsetof(struct trace_bprintk_fmt, fmt) + + strlen(*iter) + 1, GFP_KERNEL); + if (tb_fmt) { + list_add_tail(&tb_fmt->list, &trace_bprintk_fmt_list); + strcpy(tb_fmt->fmt, *iter); + *iter = tb_fmt->fmt; + } else + *iter = NULL; + } unlock_btrace(); } -static void put_btrace_metadata(void) +static int module_trace_bprintk_format_notify(struct notifier_block *self, + unsigned long val, void *data) { - lock_btrace(); - btrace_metadata_count--; - unlock_btrace(); + struct module *mod = data; + if (mod->num_trace_bprintk_fmt) { + const char **start = mod->trace_bprintk_fmt_start; + const char **end = start + mod->num_trace_bprintk_fmt; + + if (val == MODULE_STATE_COMING) + hold_module_trace_bprintk_format(start, end); + } + return 0; } +#else /* !CONFIG_MODULES */ +__init static int +module_trace_bprintk_format_notify(struct notifier_block *self, + unsigned long val, void *data) +{ + return 0; +} +#endif /* CONFIG_MODULES */ + + +__initdata_or_module static +struct notifier_block module_trace_bprintk_format_nb = { + .notifier_call = module_trace_bprintk_format_notify, +}; + /* events tracer */ int trace_bprintk_enable; static void start_bprintk_trace(struct trace_array *tr) { - get_btrace_metadata(); tracing_reset_online_cpus(tr); trace_bprintk_enable = 1; } @@ -61,7 +122,6 @@ static void stop_bprintk_trace(struct trace_array *tr) { trace_bprintk_enable = 0; tracing_reset_online_cpus(tr); - put_btrace_metadata(); } static int init_bprintk_trace(struct trace_array *tr) @@ -81,7 +141,14 @@ static struct tracer bprintk_trace __read_mostly = static __init int init_bprintk(void) { - return register_tracer(&bprintk_trace); + int ret = register_module_notifier(&module_trace_bprintk_format_nb); + if (ret) + return ret; + + ret = register_tracer(&bprintk_trace); + if (ret) + unregister_module_notifier(&module_trace_bprintk_format_nb); + return ret; } device_initcall(init_bprintk); From 769b0441f438c4bb4872cb8560eb6fe51bcc09ee Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 6 Mar 2009 17:21:49 +0100 Subject: [PATCH 249/413] tracing/core: drop the old trace_printk() implementation in favour of trace_bprintk() Impact: faster and lighter tracing Now that we have trace_bprintk() which is faster and consume lesser memory than trace_printk() and has the same purpose, we can now drop the old implementation in favour of the binary one from trace_bprintk(), which means we move all the implementation of trace_bprintk() to trace_printk(), so the Api doesn't change except that we must now use trace_seq_bprintk() to print the TRACE_PRINT entries. Some changes result of this: - Previously, trace_bprintk depended of a single tracer and couldn't work without. This tracer has been dropped and the whole implementation of trace_printk() (like the module formats management) is now integrated in the tracing core (comes with CONFIG_TRACING), though we keep the file trace_printk (previously trace_bprintk.c) where we can find the module management. Thus we don't overflow trace.c - changes some parts to use trace_seq_bprintk() to print TRACE_PRINT entries. - change a bit trace_printk/trace_vprintk macros to support non-builtin formats constants, and fix 'const' qualifiers warnings. But this is all transparent for developers. - etc... V2: - Rebase against last changes - Fix mispell on the changelog V3: - Rebase against last changes (moving trace_printk() to kernel.h) Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt LKML-Reference: <1236356510-8381-5-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 25 --- include/linux/kernel.h | 34 ++- include/linux/module.h | 2 +- kernel/trace/Kconfig | 7 +- kernel/trace/Makefile | 2 +- kernel/trace/trace.c | 212 +++++------------- kernel/trace/trace.h | 14 +- kernel/trace/trace_functions_graph.c | 6 +- kernel/trace/trace_mmiotrace.c | 9 +- kernel/trace/trace_output.c | 74 ++---- kernel/trace/trace_output.h | 2 + .../trace/{trace_bprintk.c => trace_printk.c} | 92 ++++---- 12 files changed, 164 insertions(+), 315 deletions(-) rename kernel/trace/{trace_bprintk.c => trace_printk.c} (66%) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 1cc8ca453a9b..e1583f2639b0 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -223,31 +223,6 @@ extern int ftrace_make_nop(struct module *mod, */ extern int ftrace_make_call(struct dyn_ftrace *rec, unsigned long addr); -#ifdef CONFIG_TRACE_BPRINTK -extern int trace_vbprintk(unsigned long ip, const char *fmt, va_list args); -extern int __trace_bprintk(unsigned long ip, const char *fmt, ...) - __attribute__ ((format (printf, 2, 3))); - -static inline void ____trace_bprintk_check_format(const char *fmt, ...) - __attribute__ ((format (printf, 1, 2))); -static inline void ____trace_bprintk_check_format(const char *fmt, ...) {} -#define __trace_bprintk_check_format(fmt, args...) \ -do { \ - if (0) \ - ____trace_bprintk_check_format(fmt, ##args); \ -} while (0) - -#define trace_bprintk(fmt, args...) \ -do { \ - static char *__attribute__((section("__trace_bprintk_fmt"))) \ - trace_bprintk_fmt = fmt; \ - __trace_bprintk_check_format(fmt, ##args); \ - __trace_bprintk(_THIS_IP_, trace_bprintk_fmt, ##args); \ -} while (0) -#else -#define trace_bprintk trace_printk -#endif - /* May be defined in arch */ extern int ftrace_arch_read_dyn_info(char *buf, int size); diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 7aef15c4645e..4e726b9a71ec 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -423,6 +423,16 @@ extern void ftrace_off_permanent(void); extern void ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3); +static inline void __attribute__ ((format (printf, 1, 2))) +____trace_printk_check_format(const char *fmt, ...) +{ +} +#define __trace_printk_check_format(fmt, args...) \ +do { \ + if (0) \ + ____trace_printk_check_format(fmt, ##args); \ +} while (0) + /** * trace_printk - printf formatting in the ftrace buffer * @fmt: the printf format for printing @@ -439,13 +449,31 @@ ftrace_special(unsigned long arg1, unsigned long arg2, unsigned long arg3); * Please refrain from leaving trace_printks scattered around in * your code. */ -# define trace_printk(fmt...) __trace_printk(_THIS_IP_, fmt) + +#define trace_printk(fmt, args...) \ +do { \ + static const char *trace_printk_fmt \ + __attribute__((section("__trace_printk_fmt"))); \ + trace_printk_fmt = fmt; \ + __trace_printk_check_format(fmt, ##args); \ + __trace_printk(_THIS_IP_, trace_printk_fmt, ##args); \ +} while (0) + extern int __trace_printk(unsigned long ip, const char *fmt, ...) __attribute__ ((format (printf, 2, 3))); -# define ftrace_vprintk(fmt, ap) __trace_printk(_THIS_IP_, fmt, ap) + +#define ftrace_vprintk(fmt, vargs) \ +do { \ + static const char *trace_printk_fmt \ + __attribute__((section("__trace_printk_fmt"))); \ + trace_printk_fmt = fmt; \ + __ftrace_vprintk(_THIS_IP_, trace_printk_fmt, vargs); \ +} while (0) + extern int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); + extern void ftrace_dump(void); #else static inline void @@ -467,7 +495,7 @@ ftrace_vprintk(const char *fmt, va_list ap) return 0; } static inline void ftrace_dump(void) { } -#endif +#endif /* CONFIG_TRACING */ /* * Display an IP address in readable format. diff --git a/include/linux/module.h b/include/linux/module.h index 8cbec972d8e7..22d9878e868c 100644 --- a/include/linux/module.h +++ b/include/linux/module.h @@ -329,7 +329,7 @@ struct module unsigned int num_tracepoints; #endif -#ifdef CONFIG_TRACE_BPRINTK +#ifdef CONFIG_TRACING const char **trace_bprintk_fmt_start; unsigned int num_trace_bprintk_fmt; #endif diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index ad8d3617d0a6..8e4a2a61cd75 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -52,6 +52,7 @@ config TRACING select STACKTRACE if STACKTRACE_SUPPORT select TRACEPOINTS select NOP_TRACER + select BINARY_PRINTF # # Minimum requirements an architecture has to meet for us to @@ -97,12 +98,6 @@ config FUNCTION_GRAPH_TRACER This is done by setting the current return address on the current task structure into a stack of calls. -config TRACE_BPRINTK - bool "Binary printk for tracing" - default y - depends on TRACING - select BINARY_PRINTF - config IRQSOFF_TRACER bool "Interrupts-off Latency Tracer" default n diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 46557ef4c379..c7a2943796eb 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -22,7 +22,7 @@ obj-$(CONFIG_TRACING) += trace.o obj-$(CONFIG_TRACING) += trace_clock.o obj-$(CONFIG_TRACING) += trace_output.o obj-$(CONFIG_TRACING) += trace_stat.o -obj-$(CONFIG_TRACE_BPRINTK) += trace_bprintk.o +obj-$(CONFIG_TRACING) += trace_printk.o obj-$(CONFIG_CONTEXT_SWITCH_TRACER) += trace_sched_switch.o obj-$(CONFIG_SYSPROF_TRACER) += trace_sysprof.o obj-$(CONFIG_FUNCTION_TRACER) += trace_functions.o diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 46b3cd7a5752..cc94f8642485 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1169,6 +1169,67 @@ void trace_graph_return(struct ftrace_graph_ret *trace) } #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + +/** + * trace_vprintk - write binary msg to tracing buffer + * + */ +int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) +{ + static DEFINE_SPINLOCK(trace_buf_lock); + static u32 trace_buf[TRACE_BUF_SIZE]; + + struct ring_buffer_event *event; + struct trace_array *tr = &global_trace; + struct trace_array_cpu *data; + struct print_entry *entry; + unsigned long flags; + int resched; + int cpu, len = 0, size, pc; + + if (unlikely(tracing_selftest_running || tracing_disabled)) + return 0; + + /* Don't pollute graph traces with trace_vprintk internals */ + pause_graph_tracing(); + + pc = preempt_count(); + resched = ftrace_preempt_disable(); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + + if (unlikely(atomic_read(&data->disabled))) + goto out; + + spin_lock_irqsave(&trace_buf_lock, flags); + len = vbin_printf(trace_buf, TRACE_BUF_SIZE, fmt, args); + + if (len > TRACE_BUF_SIZE || len < 0) + goto out_unlock; + + size = sizeof(*entry) + sizeof(u32) * len; + event = trace_buffer_lock_reserve(tr, TRACE_PRINT, size, flags, pc); + if (!event) + goto out_unlock; + entry = ring_buffer_event_data(event); + entry->ip = ip; + entry->depth = depth; + entry->fmt = fmt; + + memcpy(entry->buf, trace_buf, sizeof(u32) * len); + ring_buffer_unlock_commit(tr->buffer, event); + +out_unlock: + spin_unlock_irqrestore(&trace_buf_lock, flags); + +out: + ftrace_preempt_enable(resched); + unpause_graph_tracing(); + + return len; +} +EXPORT_SYMBOL_GPL(trace_vprintk); + enum trace_file_type { TRACE_FILE_LAT_FMT = 1, TRACE_FILE_ANNOTATE = 2, @@ -1564,7 +1625,7 @@ static enum print_line_t print_printk_msg_only(struct trace_iterator *iter) trace_assign_type(field, entry); - ret = trace_seq_printf(s, "%s", field->buf); + ret = trace_seq_bprintf(s, field->fmt, field->buf); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -3714,155 +3775,6 @@ static __init int tracer_init_debugfs(void) return 0; } -int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) -{ - static raw_spinlock_t trace_buf_lock = __RAW_SPIN_LOCK_UNLOCKED; - static char trace_buf[TRACE_BUF_SIZE]; - - struct ring_buffer_event *event; - struct trace_array *tr = &global_trace; - struct trace_array_cpu *data; - int cpu, len = 0, size, pc; - struct print_entry *entry; - unsigned long irq_flags; - - if (tracing_disabled || tracing_selftest_running) - return 0; - - pc = preempt_count(); - preempt_disable_notrace(); - cpu = raw_smp_processor_id(); - data = tr->data[cpu]; - - if (unlikely(atomic_read(&data->disabled))) - goto out; - - pause_graph_tracing(); - raw_local_irq_save(irq_flags); - __raw_spin_lock(&trace_buf_lock); - len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args); - - len = min(len, TRACE_BUF_SIZE-1); - trace_buf[len] = 0; - - size = sizeof(*entry) + len + 1; - event = trace_buffer_lock_reserve(tr, TRACE_PRINT, size, irq_flags, pc); - if (!event) - goto out_unlock; - entry = ring_buffer_event_data(event); - entry->ip = ip; - entry->depth = depth; - - memcpy(&entry->buf, trace_buf, len); - entry->buf[len] = 0; - ring_buffer_unlock_commit(tr->buffer, event); - - out_unlock: - __raw_spin_unlock(&trace_buf_lock); - raw_local_irq_restore(irq_flags); - unpause_graph_tracing(); - out: - preempt_enable_notrace(); - - return len; -} -EXPORT_SYMBOL_GPL(trace_vprintk); - -int __trace_printk(unsigned long ip, const char *fmt, ...) -{ - int ret; - va_list ap; - - if (!(trace_flags & TRACE_ITER_PRINTK)) - return 0; - - va_start(ap, fmt); - ret = trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); - va_end(ap); - return ret; -} -EXPORT_SYMBOL_GPL(__trace_printk); - -int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) -{ - if (!(trace_flags & TRACE_ITER_PRINTK)) - return 0; - - return trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); -} -EXPORT_SYMBOL_GPL(__ftrace_vprintk); - -/** - * trace_vbprintk - write binary msg to tracing buffer - * - * Caller must insure @fmt are valid when msg is in tracing buffer. - */ -int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) -{ - static DEFINE_SPINLOCK(trace_buf_lock); - static u32 trace_buf[TRACE_BUF_SIZE]; - - struct ring_buffer_event *event; - struct trace_array *tr = &global_trace; - struct trace_array_cpu *data; - struct bprintk_entry *entry; - unsigned long flags; - int resched; - int cpu, len = 0, size, pc; - - if (tracing_disabled || !trace_bprintk_enable) - return 0; - - pc = preempt_count(); - resched = ftrace_preempt_disable(); - cpu = raw_smp_processor_id(); - data = tr->data[cpu]; - - if (unlikely(atomic_read(&data->disabled))) - goto out; - - spin_lock_irqsave(&trace_buf_lock, flags); - len = vbin_printf(trace_buf, TRACE_BUF_SIZE, fmt, args); - - if (len > TRACE_BUF_SIZE || len < 0) - goto out_unlock; - - size = sizeof(*entry) + sizeof(u32) * len; - event = trace_buffer_lock_reserve(tr, TRACE_BPRINTK, size, flags, pc); - if (!event) - goto out_unlock; - entry = ring_buffer_event_data(event); - entry->ip = ip; - entry->fmt = fmt; - - memcpy(entry->buf, trace_buf, sizeof(u32) * len); - ring_buffer_unlock_commit(tr->buffer, event); - -out_unlock: - spin_unlock_irqrestore(&trace_buf_lock, flags); - -out: - ftrace_preempt_enable(resched); - - return len; -} -EXPORT_SYMBOL_GPL(trace_vbprintk); - -int __trace_bprintk(unsigned long ip, const char *fmt, ...) -{ - int ret; - va_list ap; - - if (!fmt) - return 0; - - va_start(ap, fmt); - ret = trace_vbprintk(ip, fmt, ap); - va_end(ap); - return ret; -} -EXPORT_SYMBOL_GPL(__trace_bprintk); - static int trace_panic_handler(struct notifier_block *this, unsigned long event, void *unused) { diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 0f5077f8f957..6140922392c8 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -20,7 +20,6 @@ enum trace_type { TRACE_WAKE, TRACE_STACK, TRACE_PRINT, - TRACE_BPRINTK, TRACE_SPECIAL, TRACE_MMIO_RW, TRACE_MMIO_MAP, @@ -120,16 +119,10 @@ struct userstack_entry { */ struct print_entry { struct trace_entry ent; - unsigned long ip; + unsigned long ip; int depth; - char buf[]; -}; - -struct bprintk_entry { - struct trace_entry ent; - unsigned long ip; - const char *fmt; - u32 buf[]; + const char *fmt; + u32 buf[]; }; #ifdef CONFIG_TRACE_BPRINTK extern int trace_bprintk_enable; @@ -296,7 +289,6 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct stack_entry, TRACE_STACK); \ IF_ASSIGN(var, ent, struct userstack_entry, TRACE_USER_STACK);\ IF_ASSIGN(var, ent, struct print_entry, TRACE_PRINT); \ - IF_ASSIGN(var, ent, struct bprintk_entry, TRACE_BPRINTK);\ IF_ASSIGN(var, ent, struct special_entry, 0); \ IF_ASSIGN(var, ent, struct trace_mmiotrace_rw, \ TRACE_MMIO_RW); \ diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index e527f2f66c73..453ebd3b636e 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -742,7 +742,11 @@ print_graph_comment(struct print_entry *trace, struct trace_seq *s, } /* The comment */ - ret = trace_seq_printf(s, "/* %s", trace->buf); + ret = trace_seq_printf(s, "/* "); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + ret = trace_seq_bprintf(s, trace->fmt, trace->buf); if (!ret) return TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index c401b908e805..23e346a734ca 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -254,15 +254,18 @@ static enum print_line_t mmio_print_mark(struct trace_iterator *iter) { struct trace_entry *entry = iter->ent; struct print_entry *print = (struct print_entry *)entry; - const char *msg = print->buf; struct trace_seq *s = &iter->seq; unsigned long long t = ns2usecs(iter->ts); - unsigned long usec_rem = do_div(t, 1000000ULL); + unsigned long usec_rem = do_div(t, USEC_PER_SEC); unsigned secs = (unsigned long)t; int ret; /* The trailing newline must be in the message. */ - ret = trace_seq_printf(s, "MARK %u.%06lu %s", secs, usec_rem, msg); + ret = trace_seq_printf(s, "MARK %u.%06lu ", secs, usec_rem); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + ret = trace_seq_bprintf(s, print->fmt, print->buf); if (!ret) return TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 4ab71201862e..ef8fd661b217 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -53,8 +53,7 @@ trace_seq_printf(struct trace_seq *s, const char *fmt, ...) return len; } -static int -trace_seq_bprintf(struct trace_seq *s, const char *fmt, const u32 *binary) +int trace_seq_bprintf(struct trace_seq *s, const char *fmt, const u32 *binary) { int len = (PAGE_SIZE - 1) - s->len; int ret; @@ -834,54 +833,12 @@ static struct trace_event trace_user_stack_event = { }; /* TRACE_PRINT */ -static enum print_line_t trace_print_print(struct trace_iterator *iter, - int flags) -{ - struct print_entry *field; - struct trace_seq *s = &iter->seq; - - trace_assign_type(field, iter->ent); - - if (!seq_print_ip_sym(s, field->ip, flags)) - goto partial; - - if (!trace_seq_printf(s, ": %s", field->buf)) - goto partial; - - return TRACE_TYPE_HANDLED; - - partial: - return TRACE_TYPE_PARTIAL_LINE; -} - -static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) -{ - struct print_entry *field; - - trace_assign_type(field, iter->ent); - - if (!trace_seq_printf(&iter->seq, "# %lx %s", field->ip, field->buf)) - goto partial; - - return TRACE_TYPE_HANDLED; - - partial: - return TRACE_TYPE_PARTIAL_LINE; -} - -static struct trace_event trace_print_event = { - .type = TRACE_PRINT, - .trace = trace_print_print, - .raw = trace_print_raw, -}; - -/* TRACE_BPRINTK */ static enum print_line_t -trace_bprintk_print(struct trace_iterator *iter, int flags) +trace_print_print(struct trace_iterator *iter, int flags) { struct trace_entry *entry = iter->ent; struct trace_seq *s = &iter->seq; - struct bprintk_entry *field; + struct print_entry *field; trace_assign_type(field, entry); @@ -900,14 +857,13 @@ trace_bprintk_print(struct trace_iterator *iter, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static enum print_line_t -trace_bprintk_raw(struct trace_iterator *iter, int flags) -{ - struct trace_entry *entry = iter->ent; - struct trace_seq *s = &iter->seq; - struct bprintk_entry *field; - trace_assign_type(field, entry); +static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) +{ + struct print_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); if (!trace_seq_printf(s, ": %lx : ", field->ip)) goto partial; @@ -921,12 +877,11 @@ trace_bprintk_raw(struct trace_iterator *iter, int flags) return TRACE_TYPE_PARTIAL_LINE; } -static struct trace_event trace_bprintk_event = { - .type = TRACE_BPRINTK, - .trace = trace_bprintk_print, - .raw = trace_bprintk_raw, - .hex = trace_nop_print, - .binary = trace_nop_print, + +static struct trace_event trace_print_event = { + .type = TRACE_PRINT, + .trace = trace_print_print, + .raw = trace_print_raw, }; static struct trace_event *events[] __initdata = { @@ -937,7 +892,6 @@ static struct trace_event *events[] __initdata = { &trace_stack_event, &trace_user_stack_event, &trace_print_event, - &trace_bprintk_event, NULL }; diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 8a34d688ed63..3b90e6ade1aa 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -18,6 +18,8 @@ struct trace_event { extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...) __attribute__ ((format (printf, 2, 3))); extern int +trace_seq_bprintf(struct trace_seq *s, const char *fmt, const u32 *binary); +extern int seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags); extern ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, diff --git a/kernel/trace/trace_bprintk.c b/kernel/trace/trace_printk.c similarity index 66% rename from kernel/trace/trace_bprintk.c rename to kernel/trace/trace_printk.c index f4c245a5cd33..a50aea22e929 100644 --- a/kernel/trace/trace_bprintk.c +++ b/kernel/trace/trace_printk.c @@ -21,31 +21,20 @@ #ifdef CONFIG_MODULES -/* binary printk basic */ -static DEFINE_MUTEX(btrace_mutex); /* - * modules trace_bprintk()'s formats are autosaved in struct trace_bprintk_fmt + * modules trace_printk()'s formats are autosaved in struct trace_bprintk_fmt * which are queued on trace_bprintk_fmt_list. */ static LIST_HEAD(trace_bprintk_fmt_list); +/* serialize accesses to trace_bprintk_fmt_list */ +static DEFINE_MUTEX(btrace_mutex); + struct trace_bprintk_fmt { struct list_head list; char fmt[0]; }; - -static inline void lock_btrace(void) -{ - mutex_lock(&btrace_mutex); -} - -static inline void unlock_btrace(void) -{ - mutex_unlock(&btrace_mutex); -} - - static inline struct trace_bprintk_fmt *lookup_format(const char *fmt) { struct trace_bprintk_fmt *pos; @@ -60,7 +49,8 @@ static void hold_module_trace_bprintk_format(const char **start, const char **end) { const char **iter; - lock_btrace(); + + mutex_lock(&btrace_mutex); for (iter = start; iter < end; iter++) { struct trace_bprintk_fmt *tb_fmt = lookup_format(*iter); if (tb_fmt) { @@ -77,7 +67,7 @@ void hold_module_trace_bprintk_format(const char **start, const char **end) } else *iter = NULL; } - unlock_btrace(); + mutex_unlock(&btrace_mutex); } static int module_trace_bprintk_format_notify(struct notifier_block *self, @@ -109,46 +99,40 @@ struct notifier_block module_trace_bprintk_format_nb = { .notifier_call = module_trace_bprintk_format_notify, }; -/* events tracer */ -int trace_bprintk_enable; +int __trace_printk(unsigned long ip, const char *fmt, ...) + { + int ret; + va_list ap; -static void start_bprintk_trace(struct trace_array *tr) -{ - tracing_reset_online_cpus(tr); - trace_bprintk_enable = 1; -} + if (unlikely(!fmt)) + return 0; -static void stop_bprintk_trace(struct trace_array *tr) -{ - trace_bprintk_enable = 0; - tracing_reset_online_cpus(tr); -} + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; -static int init_bprintk_trace(struct trace_array *tr) -{ - start_bprintk_trace(tr); - return 0; -} - -static struct tracer bprintk_trace __read_mostly = -{ - .name = "events", - .init = init_bprintk_trace, - .reset = stop_bprintk_trace, - .start = start_bprintk_trace, - .stop = stop_bprintk_trace, -}; - -static __init int init_bprintk(void) -{ - int ret = register_module_notifier(&module_trace_bprintk_format_nb); - if (ret) - return ret; - - ret = register_tracer(&bprintk_trace); - if (ret) - unregister_module_notifier(&module_trace_bprintk_format_nb); + va_start(ap, fmt); + ret = trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); + va_end(ap); return ret; } +EXPORT_SYMBOL_GPL(__trace_printk); -device_initcall(init_bprintk); +int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) + { + if (unlikely(!fmt)) + return 0; + + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + return trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); +} +EXPORT_SYMBOL_GPL(__ftrace_vprintk); + + +static __init int init_trace_printk(void) +{ + return register_module_notifier(&module_trace_bprintk_format_nb); +} + +early_initcall(init_trace_printk); From 9de36825b321fe9fe9cf73260554251af579f4ca Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 6 Mar 2009 17:52:03 +0100 Subject: [PATCH 250/413] tracing: trace_bprintk() cleanups Impact: cleanup Remove a few leftovers and clean up the code a bit. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1236356510-8381-5-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/module.c | 6 ------ kernel/trace/trace.h | 19 ++++++++----------- 2 files changed, 8 insertions(+), 17 deletions(-) diff --git a/kernel/module.c b/kernel/module.c index 2dece104f9a1..22d7379709da 100644 --- a/kernel/module.c +++ b/kernel/module.c @@ -2158,12 +2158,6 @@ static noinline struct module *load_module(void __user *umod, &mod->num_tracepoints); #endif -#ifdef CONFIG_TRACE_BPRINTK - mod->trace_bprintk_fmt_start = section_objs(hdr, sechdrs, secstrings, - "__trace_bprintk_fmt", sizeof(char *), - &mod->num_trace_bprintk_fmt); -#endif - #ifdef CONFIG_MODVERSIONS if ((mod->num_syms && !mod->crcs) || (mod->num_gpl_syms && !mod->gpl_crcs) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 6140922392c8..2bfb7d11fc17 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -119,14 +119,11 @@ struct userstack_entry { */ struct print_entry { struct trace_entry ent; - unsigned long ip; + unsigned long ip; int depth; const char *fmt; - u32 buf[]; + u32 buf[]; }; -#ifdef CONFIG_TRACE_BPRINTK -extern int trace_bprintk_enable; -#endif #define TRACE_OLD_SIZE 88 @@ -199,7 +196,7 @@ struct kmemtrace_free_entry { * trace_flag_type is an enumeration that holds different * states when a trace occurs. These are: * IRQS_OFF - interrupts were disabled - * IRQS_NOSUPPORT - arch does not support irqs_disabled_flags + * IRQS_NOSUPPORT - arch does not support irqs_disabled_flags * NEED_RESCED - reschedule is requested * HARDIRQ - inside an interrupt handler * SOFTIRQ - inside a softirq handler @@ -302,7 +299,7 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct ftrace_graph_ret_entry, \ TRACE_GRAPH_RET); \ IF_ASSIGN(var, ent, struct hw_branch_entry, TRACE_HW_BRANCHES);\ - IF_ASSIGN(var, ent, struct trace_power, TRACE_POWER); \ + IF_ASSIGN(var, ent, struct trace_power, TRACE_POWER); \ IF_ASSIGN(var, ent, struct kmemtrace_alloc_entry, \ TRACE_KMEM_ALLOC); \ IF_ASSIGN(var, ent, struct kmemtrace_free_entry, \ @@ -325,8 +322,8 @@ enum print_line_t { * flags value in struct tracer_flags. */ struct tracer_opt { - const char *name; /* Will appear on the trace_options file */ - u32 bit; /* Mask assigned in val field in tracer_flags */ + const char *name; /* Will appear on the trace_options file */ + u32 bit; /* Mask assigned in val field in tracer_flags */ }; /* @@ -335,7 +332,7 @@ struct tracer_opt { */ struct tracer_flags { u32 val; - struct tracer_opt *opts; + struct tracer_opt *opts; }; /* Makes more easy to define a tracer opt */ @@ -390,7 +387,7 @@ struct tracer { int (*set_flag)(u32 old_flags, u32 bit, int set); struct tracer *next; int print_max; - struct tracer_flags *flags; + struct tracer_flags *flags; struct tracer_stat *stats; }; From 42b40b3d55f5782b00b74d9105c3565fbfa5cb80 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Sat, 7 Mar 2009 23:55:09 +0900 Subject: [PATCH 251/413] ftrace: fix documentation typo s/trace_max_latency/tracing_max_latency/ There isn't a trace_max_latency file, there is tracing_max_latency. Fix it. Signed-off-by: KOSAKI Motohiro Acked-by: Steven Rostedt LKML-Reference: <20090307235409.5A87.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- Documentation/ftrace.txt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/ftrace.txt b/Documentation/ftrace.txt index 22614bef6359..fd9a3e693813 100644 --- a/Documentation/ftrace.txt +++ b/Documentation/ftrace.txt @@ -100,7 +100,7 @@ of ftrace. Here is a list of some of the key files: that is displayed in one of the above output files. - trace_max_latency: + tracing_max_latency: Some of the tracers record the max latency. For example, the time interrupts are disabled. From 888b55dc314d26239d84c3b187dae555a81c1605 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Sun, 8 Mar 2009 13:12:43 +0900 Subject: [PATCH 252/413] ftrace: tracing header should put '#' at the beginning of a line In a recent discussion, Andrew Morton pointed out that tracing header should put '#' at the beginning of a line. Then, we can easily filtered the header by following grep usage: cat trace | grep -v '^#' Wakeup trace also has the same header problem. Comparison of headers displayed: before this patch: # tracer: wakeup # wakeup latency trace v1.1.5 on 2.6.29-rc7-tip-tip -------------------------------------------------------------------- latency: 19059 us, #21277/21277, CPU#1 | (M:desktop VP:0, KP:0, SP:0 HP:0 #P:4) ----------------- | task: kondemand/1-1644 (uid:0 nice:-5 policy:0 rt_prio:0) ----------------- # _------=> CPU# # / _-----=> irqs-off # | / _----=> need-resched # || / _---=> hardirq/softirq # ||| / _--=> preempt-depth # |||| / # ||||| delay # cmd pid ||||| time | caller # \ / ||||| \ | / irqbalan-1887 1d.s. 0us : 1887:120:R + [001] 1644:115:S kondemand/1 irqbalan-1887 1d.s. 1us : default_wake_function <-autoremove_wake_function irqbalan-1887 1d.s. 2us : check_preempt_wakeup <-try_to_wake_up after this patch: # tracer: wakeup # # wakeup latency trace v1.1.5 on 2.6.29-rc7-tip-tip # -------------------------------------------------------------------- # latency: 529 us, #530/530, CPU#0 | (M:desktop VP:0, KP:0, SP:0 HP:0 #P:4) # ----------------- # | task: kondemand/0-1641 (uid:0 nice:-5 policy:0 rt_prio:0) # ----------------- # # _------=> CPU# # / _-----=> irqs-off # | / _----=> need-resched # || / _---=> hardirq/softirq # ||| / _--=> preempt-depth # |||| / # ||||| delay # cmd pid ||||| time | caller # \ / ||||| \ | / sshd-2496 0d.s. 0us : 2496:120:R + [000] 1641:115:S kondemand/0 sshd-2496 0d.s. 1us : default_wake_function <-autoremove_wake_function sshd-2496 0d.s. 1us : check_preempt_wakeup <-try_to_wake_up Signed-off-by: KOSAKI Motohiro Cc: Andrew Morton LKML-Reference: <20090308124421.23C3.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index cc94f8642485..e5b56199e5e0 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1466,11 +1466,11 @@ print_trace_header(struct seq_file *m, struct trace_iterator *iter) total = entries + ring_buffer_overruns(iter->tr->buffer); - seq_printf(m, "%s latency trace v1.1.5 on %s\n", + seq_printf(m, "# %s latency trace v1.1.5 on %s\n", name, UTS_RELEASE); - seq_puts(m, "-----------------------------------" + seq_puts(m, "# -----------------------------------" "---------------------------------\n"); - seq_printf(m, " latency: %lu us, #%lu/%lu, CPU#%d |" + seq_printf(m, "# latency: %lu us, #%lu/%lu, CPU#%d |" " (M:%s VP:%d, KP:%d, SP:%d HP:%d", nsecs_to_usecs(data->saved_latency), entries, @@ -1492,24 +1492,24 @@ print_trace_header(struct seq_file *m, struct trace_iterator *iter) #else seq_puts(m, ")\n"); #endif - seq_puts(m, " -----------------\n"); - seq_printf(m, " | task: %.16s-%d " + seq_puts(m, "# -----------------\n"); + seq_printf(m, "# | task: %.16s-%d " "(uid:%d nice:%ld policy:%ld rt_prio:%ld)\n", data->comm, data->pid, data->uid, data->nice, data->policy, data->rt_priority); - seq_puts(m, " -----------------\n"); + seq_puts(m, "# -----------------\n"); if (data->critical_start) { - seq_puts(m, " => started at: "); + seq_puts(m, "# => started at: "); seq_print_ip_sym(&iter->seq, data->critical_start, sym_flags); trace_print_seq(m, &iter->seq); - seq_puts(m, "\n => ended at: "); + seq_puts(m, "\n# => ended at: "); seq_print_ip_sym(&iter->seq, data->critical_end, sym_flags); trace_print_seq(m, &iter->seq); - seq_puts(m, "\n"); + seq_puts(m, "#\n"); } - seq_puts(m, "\n"); + seq_puts(m, "#\n"); } static void test_cpu_buff_start(struct trace_iterator *iter) From 8a20d84d09ab5d121f989cd99e4fc5f4b49f98ba Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 9 Mar 2009 10:09:06 +0100 Subject: [PATCH 253/413] tracing: trace_printk() fix, move format array to data section Impact: fix kernel crash when using trace_printk() trace_printk_fmt section is defined into the readonly section. But we do: trace_printk_fmt = fmt; to fill in that table of format strings - which is not read-only. Under CONFIG_DEBUG_RODATA=y this crashes ... Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Peter Zijlstra LKML-Reference: <1236356510-8381-5-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- include/asm-generic/vmlinux.lds.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 48ade3168b13..d656b4624024 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -98,6 +98,7 @@ VMLINUX_SYMBOL(__stop___tracepoints) = .; \ LIKELY_PROFILE() \ BRANCH_PROFILE() \ + TRACE_PRINTKS() \ FTRACE_EVENTS() #define RO_DATA(align) \ @@ -108,7 +109,6 @@ *(__vermagic) /* Kernel version magic */ \ *(__markers_strings) /* Markers: strings */ \ *(__tracepoints_strings)/* Tracepoints: strings */ \ - TRACE_PRINTKS() \ } \ \ .rodata1 : AT(ADDR(.rodata1) - LOAD_OFFSET) { \ From 7bffc23e56e92c14b787bf4d95249a32085bfed5 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Mon, 9 Mar 2009 10:11:36 +0100 Subject: [PATCH 254/413] tracing: optimize trace_printk() Impact: micro-optimization trace_printk() does this unconditionally: trace_printk_fmt = fmt; Where trace_printk_fmt is an entry into a global array. This is very SMP-unfriendly. So only write it once per bootup. Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Peter Zijlstra LKML-Reference: <1236356510-8381-5-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- include/linux/kernel.h | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 4e726b9a71ec..7742798c9208 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -454,7 +454,10 @@ do { \ do { \ static const char *trace_printk_fmt \ __attribute__((section("__trace_printk_fmt"))); \ - trace_printk_fmt = fmt; \ + \ + if (!trace_printk_fmt) \ + trace_printk_fmt = fmt; \ + \ __trace_printk_check_format(fmt, ##args); \ __trace_printk(_THIS_IP_, trace_printk_fmt, ##args); \ } while (0) @@ -467,7 +470,10 @@ __trace_printk(unsigned long ip, const char *fmt, ...) do { \ static const char *trace_printk_fmt \ __attribute__((section("__trace_printk_fmt"))); \ - trace_printk_fmt = fmt; \ + \ + if (!trace_printk_fmt) \ + trace_printk_fmt = fmt; \ + \ __ftrace_vprintk(_THIS_IP_, trace_printk_fmt, vargs); \ } while (0) From c3ffc7a40b7e94b094efe1c8ab4e24370a782b65 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Mon, 9 Mar 2009 18:15:34 +0900 Subject: [PATCH 255/413] tracing: Don't use tracing_record_cmdline() in workqueue tracer Impact: improve workqueue tracer output Currently, /sys/kernel/debug/tracing/trace_stat/workqueues can display wrong and strange thread names. Why? Currently, ftrace has tracing_record_cmdline()/trace_find_cmdline() convenience function that implements a task->comm string cache. This can avoid unnecessary memcpy overhead and the workqueue tracer uses it. However, in general, any trace statistics feature shouldn't use tracing_record_cmdline() because trace statistics can display very old process. Then comm cache can return wrong string because recent process overrides the cache. Fortunately, workqueue trace guarantees that displayed processes are live. Thus we can search comm string from PID at display time. % cat workqueues # CPU INSERTED EXECUTED NAME # | | | | 7 431913 431913 kondemand/7 7 0 0 tail 7 21 21 git 7 0 0 ls 7 9 9 cat 7 832632 832632 unix_chkpwd 7 236292 236292 ls Note: tail, git, ls, cat unix_chkpwd are obiously not workqueue thread. % cat workqueues # CPU INSERTED EXECUTED NAME # | | | | 7 510 510 kondemand/7 7 0 0 kmpathd/7 7 15 15 ata/7 7 0 0 aio/7 7 11 11 kblockd/7 7 1063 1063 work_on_cpu/7 7 167 167 events/7 Signed-off-by: KOSAKI Motohiro Cc: Lai Jiangshan Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_workqueue.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index 4664990fe9c5..46c8dc896bd3 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -99,8 +99,6 @@ static void probe_workqueue_creation(struct task_struct *wq_thread, int cpu) pr_warning("trace_workqueue: not enough memory\n"); return; } - tracing_record_cmdline(wq_thread); - INIT_LIST_HEAD(&cws->list); cws->cpu = cpu; @@ -195,11 +193,12 @@ static int workqueue_stat_show(struct seq_file *s, void *p) struct cpu_workqueue_stats *cws = p; unsigned long flags; int cpu = cws->cpu; + struct task_struct *tsk = find_task_by_vpid(cws->pid); seq_printf(s, "%3d %6d %6u %s\n", cws->cpu, atomic_read(&cws->inserted), cws->executed, - trace_find_cmdline(cws->pid)); + tsk ? tsk->comm : "<...>"); spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); if (&cws->list == workqueue_cpu_stat(cpu)->list.next) From 156b5f172a64103bcb13b6d26288388b9019caa3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 6 Mar 2009 10:50:53 -0500 Subject: [PATCH 256/413] tracing: typecast sizeof and offsetof to unsigned int Impact: fix compiler warnings On x86_64 sizeof and offsetof are treated as long, where as on x86_32 they are int. This patch typecasts them to unsigned int to avoid one arch giving warnings while the other does not. Reported-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 15 ++++++++------- kernel/trace/trace_export.c | 10 +++++----- kernel/trace/trace_format.h | 12 ++++++------ 3 files changed, 19 insertions(+), 18 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 4488d90e75ef..fa32ca320767 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -448,8 +448,9 @@ event_available_types_read(struct file *filp, char __user *ubuf, size_t cnt, } #undef FIELD -#define FIELD(type, name) \ - #type, #name, offsetof(typeof(field), name), sizeof(field.name) +#define FIELD(type, name) \ + #type, #name, (unsigned int)offsetof(typeof(field), name), \ + (unsigned int)sizeof(field.name) static int trace_write_header(struct trace_seq *s) { @@ -457,11 +458,11 @@ static int trace_write_header(struct trace_seq *s) /* struct trace_entry */ return trace_seq_printf(s, - "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" - "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" - "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" - "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" - "\tfield:%s %s;\toffset:%lu;\tsize:%lu;\n" + "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" + "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" + "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" + "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" + "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" "\n", FIELD(unsigned char, type), FIELD(unsigned char, flags), diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 0fb7be73e31c..7162ab49d05d 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -18,11 +18,11 @@ #include "trace_format.h" #undef TRACE_FIELD_ZERO_CHAR -#define TRACE_FIELD_ZERO_CHAR(item) \ - ret = trace_seq_printf(s, "\tfield: char " #item ";\t" \ - "offset:%lu;\tsize:0;\n", \ - offsetof(typeof(field), item)); \ - if (!ret) \ +#define TRACE_FIELD_ZERO_CHAR(item) \ + ret = trace_seq_printf(s, "\tfield: char " #item ";\t" \ + "offset:%u;\tsize:0;\n", \ + (unsigned int)offsetof(typeof(field), item)); \ + if (!ret) \ return 0; diff --git a/kernel/trace/trace_format.h b/kernel/trace/trace_format.h index 03f9a4c165ca..97e59a9c82ea 100644 --- a/kernel/trace/trace_format.h +++ b/kernel/trace/trace_format.h @@ -22,9 +22,9 @@ #undef TRACE_FIELD #define TRACE_FIELD(type, item, assign) \ ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%lu;\tsize:%lu;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item)); \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ if (!ret) \ return 0; @@ -32,9 +32,9 @@ #undef TRACE_FIELD_SPECIAL #define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ ret = trace_seq_printf(s, "\tfield special:" #type_item ";\t" \ - "offset:%lu;\tsize:%lu;\n", \ - offsetof(typeof(field), item), \ - sizeof(field.item)); \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ if (!ret) \ return 0; From 2939b0469d04ba9ac791aca9a81625d7eb50662b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 9 Mar 2009 15:47:18 -0400 Subject: [PATCH 257/413] tracing: replace TP with TP_ Impact: clean up The macros TPPROTO, TPARGS, TPFMT, TPRAWFMT, and TPCMD all look a bit ugly. This patch adds an underscore to their names. Signed-off-by: Steven Rostedt --- Documentation/tracepoints.txt | 8 +-- include/linux/tracepoint.h | 6 +- include/trace/block.h | 70 +++++++++--------- include/trace/irq_event_types.h | 16 ++--- include/trace/lockdep_event_types.h | 24 +++---- include/trace/power.h | 12 ++-- include/trace/sched_event_types.h | 98 +++++++++++++------------- include/trace/workqueue.h | 16 ++--- kernel/trace/trace_event_types.h | 28 ++++---- kernel/trace/trace_events_stage_2.h | 6 +- kernel/trace/trace_events_stage_3.h | 8 +-- kernel/trace/trace_export.c | 8 +-- kernel/trace/trace_format.h | 55 --------------- samples/tracepoints/tp-samples-trace.h | 8 +-- 14 files changed, 154 insertions(+), 209 deletions(-) delete mode 100644 kernel/trace/trace_format.h diff --git a/Documentation/tracepoints.txt b/Documentation/tracepoints.txt index 6f0a044f5b5e..4ff43c6de299 100644 --- a/Documentation/tracepoints.txt +++ b/Documentation/tracepoints.txt @@ -45,8 +45,8 @@ In include/trace/subsys.h : #include DECLARE_TRACE(subsys_eventname, - TPPROTO(int firstarg, struct task_struct *p), - TPARGS(firstarg, p)); + TP_PROTO(int firstarg, struct task_struct *p), + TP_ARGS(firstarg, p)); In subsys/file.c (where the tracing statement must be added) : @@ -66,10 +66,10 @@ Where : - subsys is the name of your subsystem. - eventname is the name of the event to trace. -- TPPROTO(int firstarg, struct task_struct *p) is the prototype of the +- TP_PROTO(int firstarg, struct task_struct *p) is the prototype of the function called by this tracepoint. -- TPARGS(firstarg, p) are the parameters names, same as found in the +- TP_ARGS(firstarg, p) are the parameters names, same as found in the prototype. Connecting a function (probe) to a tracepoint is done by providing a diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 152b2f03fb86..3bcc3e171443 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -31,8 +31,8 @@ struct tracepoint { * Keep in sync with vmlinux.lds.h. */ -#define TPPROTO(args...) args -#define TPARGS(args...) args +#define TP_PROTO(args...) args +#define TP_ARGS(args...) args #ifdef CONFIG_TRACEPOINTS @@ -65,7 +65,7 @@ struct tracepoint { { \ if (unlikely(__tracepoint_##name.state)) \ __DO_TRACE(&__tracepoint_##name, \ - TPPROTO(proto), TPARGS(args)); \ + TP_PROTO(proto), TP_ARGS(args)); \ } \ static inline int register_trace_##name(void (*probe)(proto)) \ { \ diff --git a/include/trace/block.h b/include/trace/block.h index 25c6a1fd5b77..25b7068b819e 100644 --- a/include/trace/block.h +++ b/include/trace/block.h @@ -5,72 +5,72 @@ #include DECLARE_TRACE(block_rq_abort, - TPPROTO(struct request_queue *q, struct request *rq), - TPARGS(q, rq)); + TP_PROTO(struct request_queue *q, struct request *rq), + TP_ARGS(q, rq)); DECLARE_TRACE(block_rq_insert, - TPPROTO(struct request_queue *q, struct request *rq), - TPARGS(q, rq)); + TP_PROTO(struct request_queue *q, struct request *rq), + TP_ARGS(q, rq)); DECLARE_TRACE(block_rq_issue, - TPPROTO(struct request_queue *q, struct request *rq), - TPARGS(q, rq)); + TP_PROTO(struct request_queue *q, struct request *rq), + TP_ARGS(q, rq)); DECLARE_TRACE(block_rq_requeue, - TPPROTO(struct request_queue *q, struct request *rq), - TPARGS(q, rq)); + TP_PROTO(struct request_queue *q, struct request *rq), + TP_ARGS(q, rq)); DECLARE_TRACE(block_rq_complete, - TPPROTO(struct request_queue *q, struct request *rq), - TPARGS(q, rq)); + TP_PROTO(struct request_queue *q, struct request *rq), + TP_ARGS(q, rq)); DECLARE_TRACE(block_bio_bounce, - TPPROTO(struct request_queue *q, struct bio *bio), - TPARGS(q, bio)); + TP_PROTO(struct request_queue *q, struct bio *bio), + TP_ARGS(q, bio)); DECLARE_TRACE(block_bio_complete, - TPPROTO(struct request_queue *q, struct bio *bio), - TPARGS(q, bio)); + TP_PROTO(struct request_queue *q, struct bio *bio), + TP_ARGS(q, bio)); DECLARE_TRACE(block_bio_backmerge, - TPPROTO(struct request_queue *q, struct bio *bio), - TPARGS(q, bio)); + TP_PROTO(struct request_queue *q, struct bio *bio), + TP_ARGS(q, bio)); DECLARE_TRACE(block_bio_frontmerge, - TPPROTO(struct request_queue *q, struct bio *bio), - TPARGS(q, bio)); + TP_PROTO(struct request_queue *q, struct bio *bio), + TP_ARGS(q, bio)); DECLARE_TRACE(block_bio_queue, - TPPROTO(struct request_queue *q, struct bio *bio), - TPARGS(q, bio)); + TP_PROTO(struct request_queue *q, struct bio *bio), + TP_ARGS(q, bio)); DECLARE_TRACE(block_getrq, - TPPROTO(struct request_queue *q, struct bio *bio, int rw), - TPARGS(q, bio, rw)); + TP_PROTO(struct request_queue *q, struct bio *bio, int rw), + TP_ARGS(q, bio, rw)); DECLARE_TRACE(block_sleeprq, - TPPROTO(struct request_queue *q, struct bio *bio, int rw), - TPARGS(q, bio, rw)); + TP_PROTO(struct request_queue *q, struct bio *bio, int rw), + TP_ARGS(q, bio, rw)); DECLARE_TRACE(block_plug, - TPPROTO(struct request_queue *q), - TPARGS(q)); + TP_PROTO(struct request_queue *q), + TP_ARGS(q)); DECLARE_TRACE(block_unplug_timer, - TPPROTO(struct request_queue *q), - TPARGS(q)); + TP_PROTO(struct request_queue *q), + TP_ARGS(q)); DECLARE_TRACE(block_unplug_io, - TPPROTO(struct request_queue *q), - TPARGS(q)); + TP_PROTO(struct request_queue *q), + TP_ARGS(q)); DECLARE_TRACE(block_split, - TPPROTO(struct request_queue *q, struct bio *bio, unsigned int pdu), - TPARGS(q, bio, pdu)); + TP_PROTO(struct request_queue *q, struct bio *bio, unsigned int pdu), + TP_ARGS(q, bio, pdu)); DECLARE_TRACE(block_remap, - TPPROTO(struct request_queue *q, struct bio *bio, dev_t dev, - sector_t from, sector_t to), - TPARGS(q, bio, dev, from, to)); + TP_PROTO(struct request_queue *q, struct bio *bio, dev_t dev, + sector_t from, sector_t to), + TP_ARGS(q, bio, dev, from, to)); #endif diff --git a/include/trace/irq_event_types.h b/include/trace/irq_event_types.h index 65850bc5ea06..0147d9eef5f4 100644 --- a/include/trace/irq_event_types.h +++ b/include/trace/irq_event_types.h @@ -9,25 +9,25 @@ #define TRACE_SYSTEM irq TRACE_EVENT_FORMAT(irq_handler_entry, - TPPROTO(int irq, struct irqaction *action), - TPARGS(irq, action), - TPFMT("irq=%d handler=%s", irq, action->name), + TP_PROTO(int irq, struct irqaction *action), + TP_ARGS(irq, action), + TP_FMT("irq=%d handler=%s", irq, action->name), TRACE_STRUCT( TRACE_FIELD(int, irq, irq) ), - TPRAWFMT("irq %d") + TP_RAW_FMT("irq %d") ); TRACE_EVENT_FORMAT(irq_handler_exit, - TPPROTO(int irq, struct irqaction *action, int ret), - TPARGS(irq, action, ret), - TPFMT("irq=%d handler=%s return=%s", + TP_PROTO(int irq, struct irqaction *action, int ret), + TP_ARGS(irq, action, ret), + TP_FMT("irq=%d handler=%s return=%s", irq, action->name, ret ? "handled" : "unhandled"), TRACE_STRUCT( TRACE_FIELD(int, irq, irq) TRACE_FIELD(int, ret, ret) ), - TPRAWFMT("irq %d ret %d") + TP_RAW_FMT("irq %d ret %d") ); #undef TRACE_SYSTEM diff --git a/include/trace/lockdep_event_types.h b/include/trace/lockdep_event_types.h index f713d74a82b4..1f00e8b3543e 100644 --- a/include/trace/lockdep_event_types.h +++ b/include/trace/lockdep_event_types.h @@ -10,32 +10,32 @@ #ifdef CONFIG_LOCKDEP TRACE_FORMAT(lock_acquire, - TPPROTO(struct lockdep_map *lock, unsigned int subclass, + TP_PROTO(struct lockdep_map *lock, unsigned int subclass, int trylock, int read, int check, struct lockdep_map *next_lock, unsigned long ip), - TPARGS(lock, subclass, trylock, read, check, next_lock, ip), - TPFMT("%s%s%s", trylock ? "try " : "", + TP_ARGS(lock, subclass, trylock, read, check, next_lock, ip), + TP_FMT("%s%s%s", trylock ? "try " : "", read ? "read " : "", lock->name) ); TRACE_FORMAT(lock_release, - TPPROTO(struct lockdep_map *lock, int nested, unsigned long ip), - TPARGS(lock, nested, ip), - TPFMT("%s", lock->name) + TP_PROTO(struct lockdep_map *lock, int nested, unsigned long ip), + TP_ARGS(lock, nested, ip), + TP_FMT("%s", lock->name) ); #ifdef CONFIG_LOCK_STAT TRACE_FORMAT(lock_contended, - TPPROTO(struct lockdep_map *lock, unsigned long ip), - TPARGS(lock, ip), - TPFMT("%s", lock->name) + TP_PROTO(struct lockdep_map *lock, unsigned long ip), + TP_ARGS(lock, ip), + TP_FMT("%s", lock->name) ); TRACE_FORMAT(lock_acquired, - TPPROTO(struct lockdep_map *lock, unsigned long ip), - TPARGS(lock, ip), - TPFMT("%s", lock->name) + TP_PROTO(struct lockdep_map *lock, unsigned long ip), + TP_ARGS(lock, ip), + TP_FMT("%s", lock->name) ); #endif diff --git a/include/trace/power.h b/include/trace/power.h index 38aca537e497..ef204666e983 100644 --- a/include/trace/power.h +++ b/include/trace/power.h @@ -18,15 +18,15 @@ struct power_trace { }; DECLARE_TRACE(power_start, - TPPROTO(struct power_trace *it, unsigned int type, unsigned int state), - TPARGS(it, type, state)); + TP_PROTO(struct power_trace *it, unsigned int type, unsigned int state), + TP_ARGS(it, type, state)); DECLARE_TRACE(power_mark, - TPPROTO(struct power_trace *it, unsigned int type, unsigned int state), - TPARGS(it, type, state)); + TP_PROTO(struct power_trace *it, unsigned int type, unsigned int state), + TP_ARGS(it, type, state)); DECLARE_TRACE(power_end, - TPPROTO(struct power_trace *it), - TPARGS(it)); + TP_PROTO(struct power_trace *it), + TP_ARGS(it)); #endif /* _TRACE_POWER_H */ diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index a6de5c1601a0..71b14828a957 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -9,143 +9,143 @@ #define TRACE_SYSTEM sched TRACE_EVENT_FORMAT(sched_kthread_stop, - TPPROTO(struct task_struct *t), - TPARGS(t), - TPFMT("task %s:%d", t->comm, t->pid), + TP_PROTO(struct task_struct *t), + TP_ARGS(t), + TP_FMT("task %s:%d", t->comm, t->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, t->pid) ), - TPRAWFMT("task %d") + TP_RAW_FMT("task %d") ); TRACE_EVENT_FORMAT(sched_kthread_stop_ret, - TPPROTO(int ret), - TPARGS(ret), - TPFMT("ret=%d", ret), + TP_PROTO(int ret), + TP_ARGS(ret), + TP_FMT("ret=%d", ret), TRACE_STRUCT( TRACE_FIELD(int, ret, ret) ), - TPRAWFMT("ret=%d") + TP_RAW_FMT("ret=%d") ); TRACE_EVENT_FORMAT(sched_wait_task, - TPPROTO(struct rq *rq, struct task_struct *p), - TPARGS(rq, p), - TPFMT("task %s:%d", p->comm, p->pid), + TP_PROTO(struct rq *rq, struct task_struct *p), + TP_ARGS(rq, p), + TP_FMT("task %s:%d", p->comm, p->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, p->pid) ), - TPRAWFMT("task %d") + TP_RAW_FMT("task %d") ); TRACE_EVENT_FORMAT(sched_wakeup, - TPPROTO(struct rq *rq, struct task_struct *p, int success), - TPARGS(rq, p, success), - TPFMT("task %s:%d %s", + TP_PROTO(struct rq *rq, struct task_struct *p, int success), + TP_ARGS(rq, p, success), + TP_FMT("task %s:%d %s", p->comm, p->pid, success ? "succeeded" : "failed"), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, p->pid) TRACE_FIELD(int, success, success) ), - TPRAWFMT("task %d success=%d") + TP_RAW_FMT("task %d success=%d") ); TRACE_EVENT_FORMAT(sched_wakeup_new, - TPPROTO(struct rq *rq, struct task_struct *p, int success), - TPARGS(rq, p, success), - TPFMT("task %s:%d", + TP_PROTO(struct rq *rq, struct task_struct *p, int success), + TP_ARGS(rq, p, success), + TP_FMT("task %s:%d", p->comm, p->pid, success ? "succeeded" : "failed"), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, p->pid) TRACE_FIELD(int, success, success) ), - TPRAWFMT("task %d success=%d") + TP_RAW_FMT("task %d success=%d") ); TRACE_EVENT_FORMAT(sched_switch, - TPPROTO(struct rq *rq, struct task_struct *prev, + TP_PROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next), - TPARGS(rq, prev, next), - TPFMT("task %s:%d ==> %s:%d", + TP_ARGS(rq, prev, next), + TP_FMT("task %s:%d ==> %s:%d", prev->comm, prev->pid, next->comm, next->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, prev_pid, prev->pid) TRACE_FIELD(int, prev_prio, prev->prio) TRACE_FIELD_SPECIAL(char next_comm[TASK_COMM_LEN], next_comm, - TPCMD(memcpy(TRACE_ENTRY->next_comm, + TP_CMD(memcpy(TRACE_ENTRY->next_comm, next->comm, TASK_COMM_LEN))) TRACE_FIELD(pid_t, next_pid, next->pid) TRACE_FIELD(int, next_prio, next->prio) ), - TPRAWFMT("prev %d:%d ==> next %s:%d:%d") + TP_RAW_FMT("prev %d:%d ==> next %s:%d:%d") ); TRACE_EVENT_FORMAT(sched_migrate_task, - TPPROTO(struct task_struct *p, int orig_cpu, int dest_cpu), - TPARGS(p, orig_cpu, dest_cpu), - TPFMT("task %s:%d from: %d to: %d", + TP_PROTO(struct task_struct *p, int orig_cpu, int dest_cpu), + TP_ARGS(p, orig_cpu, dest_cpu), + TP_FMT("task %s:%d from: %d to: %d", p->comm, p->pid, orig_cpu, dest_cpu), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, p->pid) TRACE_FIELD(int, orig_cpu, orig_cpu) TRACE_FIELD(int, dest_cpu, dest_cpu) ), - TPRAWFMT("task %d from: %d to: %d") + TP_RAW_FMT("task %d from: %d to: %d") ); TRACE_EVENT_FORMAT(sched_process_free, - TPPROTO(struct task_struct *p), - TPARGS(p), - TPFMT("task %s:%d", p->comm, p->pid), + TP_PROTO(struct task_struct *p), + TP_ARGS(p), + TP_FMT("task %s:%d", p->comm, p->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, p->pid) ), - TPRAWFMT("task %d") + TP_RAW_FMT("task %d") ); TRACE_EVENT_FORMAT(sched_process_exit, - TPPROTO(struct task_struct *p), - TPARGS(p), - TPFMT("task %s:%d", p->comm, p->pid), + TP_PROTO(struct task_struct *p), + TP_ARGS(p), + TP_FMT("task %s:%d", p->comm, p->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, p->pid) ), - TPRAWFMT("task %d") + TP_RAW_FMT("task %d") ); TRACE_EVENT_FORMAT(sched_process_wait, - TPPROTO(struct pid *pid), - TPARGS(pid), - TPFMT("pid %d", pid_nr(pid)), + TP_PROTO(struct pid *pid), + TP_ARGS(pid), + TP_FMT("pid %d", pid_nr(pid)), TRACE_STRUCT( TRACE_FIELD(pid_t, pid, pid_nr(pid)) ), - TPRAWFMT("task %d") + TP_RAW_FMT("task %d") ); TRACE_EVENT_FORMAT(sched_process_fork, - TPPROTO(struct task_struct *parent, struct task_struct *child), - TPARGS(parent, child), - TPFMT("parent %s:%d child %s:%d", + TP_PROTO(struct task_struct *parent, struct task_struct *child), + TP_ARGS(parent, child), + TP_FMT("parent %s:%d child %s:%d", parent->comm, parent->pid, child->comm, child->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, parent, parent->pid) TRACE_FIELD(pid_t, child, child->pid) ), - TPRAWFMT("parent %d child %d") + TP_RAW_FMT("parent %d child %d") ); TRACE_EVENT_FORMAT(sched_signal_send, - TPPROTO(int sig, struct task_struct *p), - TPARGS(sig, p), - TPFMT("sig: %d task %s:%d", sig, p->comm, p->pid), + TP_PROTO(int sig, struct task_struct *p), + TP_ARGS(sig, p), + TP_FMT("sig: %d task %s:%d", sig, p->comm, p->pid), TRACE_STRUCT( TRACE_FIELD(int, sig, sig) TRACE_FIELD(pid_t, pid, p->pid) ), - TPRAWFMT("sig: %d task %d") + TP_RAW_FMT("sig: %d task %d") ); #undef TRACE_SYSTEM diff --git a/include/trace/workqueue.h b/include/trace/workqueue.h index 867829df4571..7626523deeba 100644 --- a/include/trace/workqueue.h +++ b/include/trace/workqueue.h @@ -6,20 +6,20 @@ #include DECLARE_TRACE(workqueue_insertion, - TPPROTO(struct task_struct *wq_thread, struct work_struct *work), - TPARGS(wq_thread, work)); + TP_PROTO(struct task_struct *wq_thread, struct work_struct *work), + TP_ARGS(wq_thread, work)); DECLARE_TRACE(workqueue_execution, - TPPROTO(struct task_struct *wq_thread, struct work_struct *work), - TPARGS(wq_thread, work)); + TP_PROTO(struct task_struct *wq_thread, struct work_struct *work), + TP_ARGS(wq_thread, work)); /* Trace the creation of one workqueue thread on a cpu */ DECLARE_TRACE(workqueue_creation, - TPPROTO(struct task_struct *wq_thread, int cpu), - TPARGS(wq_thread, cpu)); + TP_PROTO(struct task_struct *wq_thread, int cpu), + TP_ARGS(wq_thread, cpu)); DECLARE_TRACE(workqueue_destruction, - TPPROTO(struct task_struct *wq_thread), - TPARGS(wq_thread)); + TP_PROTO(struct task_struct *wq_thread), + TP_ARGS(wq_thread)); #endif /* __TRACE_WORKQUEUE_H */ diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h index fb4eba166433..d94179aa1fc2 100644 --- a/kernel/trace/trace_event_types.h +++ b/kernel/trace/trace_event_types.h @@ -10,7 +10,7 @@ TRACE_EVENT_FORMAT(function, TRACE_FN, ftrace_entry, ignore, TRACE_FIELD(unsigned long, ip, ip) TRACE_FIELD(unsigned long, parent_ip, parent_ip) ), - TPRAWFMT(" %lx <-- %lx") + TP_RAW_FMT(" %lx <-- %lx") ); TRACE_EVENT_FORMAT(funcgraph_entry, TRACE_GRAPH_ENT, @@ -19,7 +19,7 @@ TRACE_EVENT_FORMAT(funcgraph_entry, TRACE_GRAPH_ENT, TRACE_FIELD(unsigned long, graph_ent.func, func) TRACE_FIELD(int, graph_ent.depth, depth) ), - TPRAWFMT("--> %lx (%d)") + TP_RAW_FMT("--> %lx (%d)") ); TRACE_EVENT_FORMAT(funcgraph_exit, TRACE_GRAPH_RET, @@ -28,7 +28,7 @@ TRACE_EVENT_FORMAT(funcgraph_exit, TRACE_GRAPH_RET, TRACE_FIELD(unsigned long, ret.func, func) TRACE_FIELD(int, ret.depth, depth) ), - TPRAWFMT("<-- %lx (%d)") + TP_RAW_FMT("<-- %lx (%d)") ); TRACE_EVENT_FORMAT(wakeup, TRACE_WAKE, ctx_switch_entry, ignore, @@ -41,7 +41,7 @@ TRACE_EVENT_FORMAT(wakeup, TRACE_WAKE, ctx_switch_entry, ignore, TRACE_FIELD(unsigned char, next_state, next_state) TRACE_FIELD(unsigned int, next_cpu, next_cpu) ), - TPRAWFMT("%u:%u:%u ==+ %u:%u:%u [%03u]") + TP_RAW_FMT("%u:%u:%u ==+ %u:%u:%u [%03u]") ); TRACE_EVENT_FORMAT(context_switch, TRACE_CTX, ctx_switch_entry, ignore, @@ -54,7 +54,7 @@ TRACE_EVENT_FORMAT(context_switch, TRACE_CTX, ctx_switch_entry, ignore, TRACE_FIELD(unsigned char, next_state, next_state) TRACE_FIELD(unsigned int, next_cpu, next_cpu) ), - TPRAWFMT("%u:%u:%u ==+ %u:%u:%u [%03u]") + TP_RAW_FMT("%u:%u:%u ==+ %u:%u:%u [%03u]") ); TRACE_EVENT_FORMAT(special, TRACE_SPECIAL, special_entry, ignore, @@ -63,7 +63,7 @@ TRACE_EVENT_FORMAT(special, TRACE_SPECIAL, special_entry, ignore, TRACE_FIELD(unsigned long, arg2, arg2) TRACE_FIELD(unsigned long, arg3, arg3) ), - TPRAWFMT("(%08lx) (%08lx) (%08lx)") + TP_RAW_FMT("(%08lx) (%08lx) (%08lx)") ); /* @@ -83,7 +83,7 @@ TRACE_EVENT_FORMAT(kernel_stack, TRACE_STACK, stack_entry, ignore, TRACE_FIELD(unsigned long, caller[6], stack6) TRACE_FIELD(unsigned long, caller[7], stack7) ), - TPRAWFMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" + TP_RAW_FMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") ); @@ -98,7 +98,7 @@ TRACE_EVENT_FORMAT(user_stack, TRACE_USER_STACK, userstack_entry, ignore, TRACE_FIELD(unsigned long, caller[6], stack6) TRACE_FIELD(unsigned long, caller[7], stack7) ), - TPRAWFMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" + TP_RAW_FMT("\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n" "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") ); @@ -108,7 +108,7 @@ TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, TRACE_FIELD(unsigned int, depth, depth) TRACE_FIELD_ZERO_CHAR(buf) ), - TPRAWFMT("%08lx (%d) %s") + TP_RAW_FMT("%08lx (%d) %s") ); TRACE_EVENT_FORMAT(branch, TRACE_BRANCH, trace_branch, ignore, @@ -118,7 +118,7 @@ TRACE_EVENT_FORMAT(branch, TRACE_BRANCH, trace_branch, ignore, TRACE_FIELD_SPECIAL(char file[TRACE_FUNC_SIZE+1], file, file) TRACE_FIELD(char, correct, correct) ), - TPRAWFMT("%u:%s:%s (%u)") + TP_RAW_FMT("%u:%s:%s (%u)") ); TRACE_EVENT_FORMAT(hw_branch, TRACE_HW_BRANCHES, hw_branch_entry, ignore, @@ -126,7 +126,7 @@ TRACE_EVENT_FORMAT(hw_branch, TRACE_HW_BRANCHES, hw_branch_entry, ignore, TRACE_FIELD(u64, from, from) TRACE_FIELD(u64, to, to) ), - TPRAWFMT("from: %llx to: %llx") + TP_RAW_FMT("from: %llx to: %llx") ); TRACE_EVENT_FORMAT(power, TRACE_POWER, trace_power, ignore, @@ -136,7 +136,7 @@ TRACE_EVENT_FORMAT(power, TRACE_POWER, trace_power, ignore, TRACE_FIELD(int, state_data.type, type) TRACE_FIELD(int, state_data.state, state) ), - TPRAWFMT("%llx->%llx type:%u state:%u") + TP_RAW_FMT("%llx->%llx type:%u state:%u") ); TRACE_EVENT_FORMAT(kmem_alloc, TRACE_KMEM_ALLOC, kmemtrace_alloc_entry, ignore, @@ -149,7 +149,7 @@ TRACE_EVENT_FORMAT(kmem_alloc, TRACE_KMEM_ALLOC, kmemtrace_alloc_entry, ignore, TRACE_FIELD(gfp_t, gfp_flags, gfp_flags) TRACE_FIELD(int, node, node) ), - TPRAWFMT("type:%u call_site:%lx ptr:%p req:%lu alloc:%lu" + TP_RAW_FMT("type:%u call_site:%lx ptr:%p req:%lu alloc:%lu" " flags:%x node:%d") ); @@ -159,7 +159,7 @@ TRACE_EVENT_FORMAT(kmem_free, TRACE_KMEM_FREE, kmemtrace_free_entry, ignore, TRACE_FIELD(unsigned long, call_site, call_site) TRACE_FIELD(const void *, ptr, ptr) ), - TPRAWFMT("type:%u call_site:%lx ptr:%p") + TP_RAW_FMT("type:%u call_site:%lx ptr:%p") ); #undef TRACE_SYSTEM diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index d24a97e74aea..8e2e0f56c2a8 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -20,7 +20,7 @@ * * field = (typeof(field))entry; * - * ret = trace_seq_printf(s, "%s", "\n"); + * ret = trace_seq_printf(s, "%s", "\n"); * if (!ret) * return TRACE_TYPE_PARTIAL_LINE; * @@ -44,8 +44,8 @@ field->item, -#undef TPRAWFMT -#define TPRAWFMT(args...) args +#undef TP_RAW_FMT +#define TP_RAW_FMT(args...) args #undef TRACE_EVENT_FORMAT #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 2c8d76c7dbed..557ca52bcdcd 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -106,8 +106,8 @@ * */ -#undef TPFMT -#define TPFMT(fmt, args...) fmt "\n", ##args +#undef TP_FMT +#define TP_FMT(fmt, args...) fmt "\n", ##args #define _TRACE_FORMAT(call, proto, args, fmt) \ static void ftrace_event_##call(proto) \ @@ -152,8 +152,8 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ #define TRACE_FIELD(type, item, assign)\ entry->item = assign; -#undef TPCMD -#define TPCMD(cmd...) cmd +#undef TP_CMD +#define TP_CMD(cmd...) cmd #undef TRACE_ENTRY #define TRACE_ENTRY entry diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 7162ab49d05d..e62bc10f8103 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -26,8 +26,8 @@ return 0; -#undef TPRAWFMT -#define TPRAWFMT(args...) args +#undef TP_RAW_FMT +#define TP_RAW_FMT(args...) args #undef TRACE_EVENT_FORMAT #define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ @@ -57,8 +57,8 @@ ftrace_format_##call(struct trace_seq *s) \ #define TRACE_FIELD(type, item, assign)\ entry->item = assign; -#undef TPCMD -#define TPCMD(cmd...) cmd +#undef TP_CMD +#define TP_CMD(cmd...) cmd #undef TRACE_ENTRY #define TRACE_ENTRY entry diff --git a/kernel/trace/trace_format.h b/kernel/trace/trace_format.h deleted file mode 100644 index 97e59a9c82ea..000000000000 --- a/kernel/trace/trace_format.h +++ /dev/null @@ -1,55 +0,0 @@ -/* - * Setup the showing format of trace point. - * - * int - * ftrace_format_##call(struct trace_seq *s) - * { - * struct ftrace_raw_##call field; - * int ret; - * - * ret = trace_seq_printf(s, #type " " #item ";" - * " size:%d; offset:%d;\n", - * sizeof(field.type), - * offsetof(struct ftrace_raw_##call, - * item)); - * - * } - */ - -#undef TRACE_STRUCT -#define TRACE_STRUCT(args...) args - -#undef TRACE_FIELD -#define TRACE_FIELD(type, item, assign) \ - ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ - "offset:%u;\tsize:%u;\n", \ - (unsigned int)offsetof(typeof(field), item), \ - (unsigned int)sizeof(field.item)); \ - if (!ret) \ - return 0; - - -#undef TRACE_FIELD_SPECIAL -#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ - ret = trace_seq_printf(s, "\tfield special:" #type_item ";\t" \ - "offset:%u;\tsize:%u;\n", \ - (unsigned int)offsetof(typeof(field), item), \ - (unsigned int)sizeof(field.item)); \ - if (!ret) \ - return 0; - -#undef TRACE_EVENT_FORMAT -#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ -static int \ -ftrace_format_##call(struct trace_seq *s) \ -{ \ - struct ftrace_raw_##call field; \ - int ret; \ - \ - tstruct; \ - \ - trace_seq_printf(s, "\nprint fmt: \"%s\"\n", tpfmt); \ - \ - return ret; \ -} - diff --git a/samples/tracepoints/tp-samples-trace.h b/samples/tracepoints/tp-samples-trace.h index 01724e04c556..dffdc49878af 100644 --- a/samples/tracepoints/tp-samples-trace.h +++ b/samples/tracepoints/tp-samples-trace.h @@ -5,9 +5,9 @@ #include DECLARE_TRACE(subsys_event, - TPPROTO(struct inode *inode, struct file *file), - TPARGS(inode, file)); + TP_PROTO(struct inode *inode, struct file *file), + TP_ARGS(inode, file)); DECLARE_TRACE(subsys_eventb, - TPPROTO(void), - TPARGS()); + TP_PROTO(void), + TP_ARGS()); #endif From 9cc26a261d43e5898287a1f5808132f8f05ceb1c Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 9 Mar 2009 16:00:22 -0400 Subject: [PATCH 258/413] tracing: use generic __stringify Impact: clean up This removes the custom made STR(x) macros in the tracer and uses the generic __stringify macro instead. Signed-off-by: Steven Rostedt --- kernel/trace/events.c | 4 +--- kernel/trace/trace_events_stage_3.h | 4 ++-- kernel/trace/trace_selftest.c | 6 ++---- 3 files changed, 5 insertions(+), 9 deletions(-) diff --git a/kernel/trace/events.c b/kernel/trace/events.c index f2509cbaacea..9fc918da404f 100644 --- a/kernel/trace/events.c +++ b/kernel/trace/events.c @@ -2,9 +2,7 @@ * This is the place to register all trace points as events. */ -/* someday this needs to go in a generic header */ -#define __STR(x) #x -#define STR(x) __STR(x) +#include #include diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 557ca52bcdcd..41b82b93c9c7 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -139,7 +139,7 @@ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ .name = #call, \ - .system = STR(TRACE_SYSTEM), \ + .system = __stringify(TRACE_SYSTEM), \ .regfunc = ftrace_reg_event_##call, \ .unregfunc = ftrace_unreg_event_##call, \ } @@ -225,7 +225,7 @@ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ .name = #call, \ - .system = STR(TRACE_SYSTEM), \ + .system = __stringify(TRACE_SYSTEM), \ .regfunc = ftrace_reg_event_##call, \ .unregfunc = ftrace_unreg_event_##call, \ .raw_init = ftrace_raw_init_event_##call, \ diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 7238646b8723..f907a2b29028 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -1,5 +1,6 @@ /* Include in trace.c */ +#include #include #include @@ -100,9 +101,6 @@ static inline void warn_failed_init_tracer(struct tracer *trace, int init_ret) #ifdef CONFIG_DYNAMIC_FTRACE -#define __STR(x) #x -#define STR(x) __STR(x) - /* Test dynamic code modification and ftrace filters */ int trace_selftest_startup_dynamic_tracing(struct tracer *trace, struct trace_array *tr, @@ -130,7 +128,7 @@ int trace_selftest_startup_dynamic_tracing(struct tracer *trace, * start of the function names. We simply put a '*' to * accommodate them. */ - func_name = "*" STR(DYN_FTRACE_TEST_NAME); + func_name = "*" __stringify(DYN_FTRACE_TEST_NAME); /* filter only on our function */ ftrace_set_filter(func_name, strlen(func_name), 1); From da4d03020c2af32f73e8bfbab0a66620d85bb9bb Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 9 Mar 2009 17:14:30 -0400 Subject: [PATCH 259/413] tracing: new format for specialized trace points Impact: clean up and enhancement The TRACE_EVENT_FORMAT macro looks quite ugly and is limited in its ability to save data as well as to print the record out. Working with Ingo Molnar, we came up with a new format that is much more pleasing to the eye of C developers. This new macro is more C style than the old macro, and is more obvious to what it does. Here's the example. The only updated macro in this patch is the sched_switch trace point. The old method looked like this: TRACE_EVENT_FORMAT(sched_switch, TP_PROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next), TP_ARGS(rq, prev, next), TP_FMT("task %s:%d ==> %s:%d", prev->comm, prev->pid, next->comm, next->pid), TRACE_STRUCT( TRACE_FIELD(pid_t, prev_pid, prev->pid) TRACE_FIELD(int, prev_prio, prev->prio) TRACE_FIELD_SPECIAL(char next_comm[TASK_COMM_LEN], next_comm, TP_CMD(memcpy(TRACE_ENTRY->next_comm, next->comm, TASK_COMM_LEN))) TRACE_FIELD(pid_t, next_pid, next->pid) TRACE_FIELD(int, next_prio, next->prio) ), TP_RAW_FMT("prev %d:%d ==> next %s:%d:%d") ); The above method is hard to read and requires two format fields. The new method: /* * Tracepoint for task switches, performed by the scheduler: * * (NOTE: the 'rq' argument is not used by generic trace events, * but used by the latency tracer plugin. ) */ TRACE_EVENT(sched_switch, TP_PROTO(struct rq *rq, struct task_struct *prev, struct task_struct *next), TP_ARGS(rq, prev, next), TP_STRUCT__entry( __array( char, prev_comm, TASK_COMM_LEN ) __field( pid_t, prev_pid ) __field( int, prev_prio ) __array( char, next_comm, TASK_COMM_LEN ) __field( pid_t, next_pid ) __field( int, next_prio ) ), TP_printk("task %s:%d [%d] ==> %s:%d [%d]", __entry->prev_comm, __entry->prev_pid, __entry->prev_prio, __entry->next_comm, __entry->next_pid, __entry->next_prio), TP_fast_assign( memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN); __entry->prev_pid = prev->pid; __entry->prev_prio = prev->prio; memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN); __entry->next_pid = next->pid; __entry->next_prio = next->prio; ) ); This macro is called TRACE_EVENT, it is broken up into 5 parts: TP_PROTO: the proto type of the trace point TP_ARGS: the arguments of the trace point TP_STRUCT_entry: the structure layout of the entry in the ring buffer TP_printk: the printk format TP_fast_assign: the method used to write the entry into the ring buffer The structure is the definition of how the event will be saved in the ring buffer. The printk is used by the internal tracing in case of an oops, and the kernel needs to print out the format of the record to the console. This the TP_printk gives a means to show the records in a human readable format. It is also used to print out the data from the trace file. The TP_fast_assign is executed directly. It is basically like a C function, where the __entry is the handle to the record. Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 3 + include/trace/sched_event_types.h | 48 ++++++--- kernel/trace/trace.h | 5 - kernel/trace/trace_event_types.h | 3 +- kernel/trace/trace_events.c | 159 +--------------------------- kernel/trace/trace_events_stage_1.h | 30 +++--- kernel/trace/trace_events_stage_2.h | 89 ++++++++++++---- kernel/trace/trace_events_stage_3.h | 36 ++----- kernel/trace/trace_export.c | 23 +++- 9 files changed, 161 insertions(+), 235 deletions(-) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 3bcc3e171443..6b4f1bb3701e 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -160,4 +160,7 @@ static inline void tracepoint_synchronize_unregister(void) #define TRACE_EVENT_FORMAT(name, proto, args, fmt, struct, tpfmt) \ TRACE_FORMAT(name, PARAMS(proto), PARAMS(args), PARAMS(fmt)) +#define TRACE_EVENT(name, proto, args, struct, print, assign) \ + DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) + #endif diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index 71b14828a957..aa77fb754038 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -62,25 +62,41 @@ TRACE_EVENT_FORMAT(sched_wakeup_new, TP_RAW_FMT("task %d success=%d") ); -TRACE_EVENT_FORMAT(sched_switch, +/* + * Tracepoint for task switches, performed by the scheduler: + * + * (NOTE: the 'rq' argument is not used by generic trace events, + * but used by the latency tracer plugin. ) + */ +TRACE_EVENT(sched_switch, + TP_PROTO(struct rq *rq, struct task_struct *prev, - struct task_struct *next), + struct task_struct *next), + TP_ARGS(rq, prev, next), - TP_FMT("task %s:%d ==> %s:%d", - prev->comm, prev->pid, next->comm, next->pid), - TRACE_STRUCT( - TRACE_FIELD(pid_t, prev_pid, prev->pid) - TRACE_FIELD(int, prev_prio, prev->prio) - TRACE_FIELD_SPECIAL(char next_comm[TASK_COMM_LEN], - next_comm, - TP_CMD(memcpy(TRACE_ENTRY->next_comm, - next->comm, - TASK_COMM_LEN))) - TRACE_FIELD(pid_t, next_pid, next->pid) - TRACE_FIELD(int, next_prio, next->prio) + + TP_STRUCT__entry( + __array( char, prev_comm, TASK_COMM_LEN ) + __field( pid_t, prev_pid ) + __field( int, prev_prio ) + __array( char, next_comm, TASK_COMM_LEN ) + __field( pid_t, next_pid ) + __field( int, next_prio ) ), - TP_RAW_FMT("prev %d:%d ==> next %s:%d:%d") - ); + + TP_printk("task %s:%d [%d] ==> %s:%d [%d]", + __entry->prev_comm, __entry->prev_pid, __entry->prev_prio, + __entry->next_comm, __entry->next_pid, __entry->next_prio), + + TP_fast_assign( + memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN); + __entry->prev_pid = prev->pid; + __entry->prev_prio = prev->prio; + memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN); + __entry->next_pid = next->pid; + __entry->next_prio = next->prio; + ) +); TRACE_EVENT_FORMAT(sched_migrate_task, TP_PROTO(struct task_struct *p, int orig_cpu, int dest_cpu), diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 2bfb7d11fc17..c5e1d8865fe4 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -751,12 +751,7 @@ struct ftrace_event_call { int (*regfunc)(void); void (*unregfunc)(void); int id; - struct dentry *raw_dir; - int raw_enabled; - int type; int (*raw_init)(void); - int (*raw_reg)(void); - void (*raw_unreg)(void); int (*show_format)(struct trace_seq *s); }; diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h index d94179aa1fc2..5cca4c978bde 100644 --- a/kernel/trace/trace_event_types.h +++ b/kernel/trace/trace_event_types.h @@ -106,9 +106,10 @@ TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, TRACE_STRUCT( TRACE_FIELD(unsigned long, ip, ip) TRACE_FIELD(unsigned int, depth, depth) + TRACE_FIELD(char *, fmt, fmt) TRACE_FIELD_ZERO_CHAR(buf) ), - TP_RAW_FMT("%08lx (%d) %s") + TP_RAW_FMT("%08lx (%d) fmt:%p %s") ); TRACE_EVENT_FORMAT(branch, TRACE_BRANCH, trace_branch, ignore, diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index fa32ca320767..1880a6438097 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -59,22 +59,12 @@ static void ftrace_event_enable_disable(struct ftrace_event_call *call, call->enabled = 0; call->unregfunc(); } - if (call->raw_enabled) { - call->raw_enabled = 0; - call->raw_unreg(); - } break; case 1: - if (!call->enabled && - (call->type & TRACE_EVENT_TYPE_PRINTF)) { + if (!call->enabled) { call->enabled = 1; call->regfunc(); } - if (!call->raw_enabled && - (call->type & TRACE_EVENT_TYPE_RAW)) { - call->raw_enabled = 1; - call->raw_reg(); - } break; } } @@ -300,7 +290,7 @@ event_enable_read(struct file *filp, char __user *ubuf, size_t cnt, struct ftrace_event_call *call = filp->private_data; char *buf; - if (call->enabled || call->raw_enabled) + if (call->enabled) buf = "1\n"; else buf = "0\n"; @@ -346,107 +336,6 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, return cnt; } -static ssize_t -event_type_read(struct file *filp, char __user *ubuf, size_t cnt, - loff_t *ppos) -{ - struct ftrace_event_call *call = filp->private_data; - char buf[16]; - int r = 0; - - if (call->type & TRACE_EVENT_TYPE_PRINTF) - r += sprintf(buf, "printf\n"); - - if (call->type & TRACE_EVENT_TYPE_RAW) - r += sprintf(buf+r, "raw\n"); - - return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); -} - -static ssize_t -event_type_write(struct file *filp, const char __user *ubuf, size_t cnt, - loff_t *ppos) -{ - struct ftrace_event_call *call = filp->private_data; - char buf[64]; - - /* - * If there's only one type, we can't change it. - * And currently we always have printf type, and we - * may or may not have raw type. - * - * This is a redundant check, the file should be read - * only if this is the case anyway. - */ - - if (!call->raw_init) - return -EPERM; - - if (cnt >= sizeof(buf)) - return -EINVAL; - - if (copy_from_user(&buf, ubuf, cnt)) - return -EFAULT; - - buf[cnt] = 0; - - if (!strncmp(buf, "printf", 6) && - (!buf[6] || isspace(buf[6]))) { - - call->type = TRACE_EVENT_TYPE_PRINTF; - - /* - * If raw enabled, the disable it and enable - * printf type. - */ - if (call->raw_enabled) { - call->raw_enabled = 0; - call->raw_unreg(); - - call->enabled = 1; - call->regfunc(); - } - - } else if (!strncmp(buf, "raw", 3) && - (!buf[3] || isspace(buf[3]))) { - - call->type = TRACE_EVENT_TYPE_RAW; - - /* - * If printf enabled, the disable it and enable - * raw type. - */ - if (call->enabled) { - call->enabled = 0; - call->unregfunc(); - - call->raw_enabled = 1; - call->raw_reg(); - } - } else - return -EINVAL; - - *ppos += cnt; - - return cnt; -} - -static ssize_t -event_available_types_read(struct file *filp, char __user *ubuf, size_t cnt, - loff_t *ppos) -{ - struct ftrace_event_call *call = filp->private_data; - char buf[16]; - int r = 0; - - r += sprintf(buf, "printf\n"); - - if (call->raw_init) - r += sprintf(buf+r, "raw\n"); - - return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); -} - #undef FIELD #define FIELD(type, name) \ #type, #name, (unsigned int)offsetof(typeof(field), name), \ @@ -470,6 +359,7 @@ static int trace_write_header(struct trace_seq *s) FIELD(int, pid), FIELD(int, tgid)); } + static ssize_t event_format_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) @@ -527,13 +417,6 @@ static const struct seq_operations show_set_event_seq_ops = { .stop = t_stop, }; -static const struct file_operations ftrace_avail_fops = { - .open = ftrace_event_seq_open, - .read = seq_read, - .llseek = seq_lseek, - .release = seq_release, -}; - static const struct file_operations ftrace_set_event_fops = { .open = ftrace_event_seq_open, .read = seq_read, @@ -548,17 +431,6 @@ static const struct file_operations ftrace_enable_fops = { .write = event_enable_write, }; -static const struct file_operations ftrace_type_fops = { - .open = tracing_open_generic, - .read = event_type_read, - .write = event_type_write, -}; - -static const struct file_operations ftrace_available_types_fops = { - .open = tracing_open_generic, - .read = event_available_types_read, -}; - static const struct file_operations ftrace_event_format_fops = { .open = tracing_open_generic, .read = event_format_read, @@ -647,9 +519,6 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) } } - /* default the output to printf */ - call->type = TRACE_EVENT_TYPE_PRINTF; - call->dir = debugfs_create_dir(call->name, d_events); if (!call->dir) { pr_warning("Could not create debugfs " @@ -665,21 +534,6 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) "'%s/enable' entry\n", call->name); } - /* Only let type be writable, if we can change it */ - entry = debugfs_create_file("type", - call->raw_init ? 0644 : 0444, - call->dir, call, - &ftrace_type_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'%s/type' entry\n", call->name); - - entry = debugfs_create_file("available_types", 0444, call->dir, call, - &ftrace_available_types_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'%s/available_types' entry\n", call->name); - /* A trace may not want to export its format */ if (!call->show_format) return 0; @@ -704,13 +558,6 @@ static __init int event_trace_init(void) if (!d_tracer) return 0; - entry = debugfs_create_file("available_events", 0444, d_tracer, - (void *)&show_event_seq_ops, - &ftrace_avail_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'available_events' entry\n"); - entry = debugfs_create_file("set_event", 0644, d_tracer, (void *)&show_set_event_seq_ops, &ftrace_set_event_fops); diff --git a/kernel/trace/trace_events_stage_1.h b/kernel/trace/trace_events_stage_1.h index 3830a731424c..edfcbd3a0d1b 100644 --- a/kernel/trace/trace_events_stage_1.h +++ b/kernel/trace/trace_events_stage_1.h @@ -18,19 +18,23 @@ #define TRACE_FORMAT(call, proto, args, fmt) #undef TRACE_EVENT_FORMAT -#define TRACE_EVENT_FORMAT(name, proto, args, fmt, tstruct, tpfmt) \ - struct ftrace_raw_##name { \ - struct trace_entry ent; \ - tstruct \ - }; \ +#define TRACE_EVENT_FORMAT(name, proto, args, fmt, tstruct, tpfmt) + +#undef __array +#define __array(type, item, len) type item[len]; + +#undef __field +#define __field(type, item) type item; + +#undef TP_STRUCT__entry +#define TP_STRUCT__entry(args...) args + +#undef TRACE_EVENT +#define TRACE_EVENT(name, proto, args, tstruct, print, assign) \ + struct ftrace_raw_##name { \ + struct trace_entry ent; \ + tstruct \ + }; \ static struct ftrace_event_call event_##name -#undef TRACE_STRUCT -#define TRACE_STRUCT(args...) args - -#define TRACE_FIELD(type, item, assign) \ - type item; -#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ - type_item; - #include diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index 8e2e0f56c2a8..d91bf4c56661 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -32,23 +32,14 @@ * in binary. */ -#undef TRACE_STRUCT -#define TRACE_STRUCT(args...) args +#undef __entry +#define __entry field -#undef TRACE_FIELD -#define TRACE_FIELD(type, item, assign) \ - field->item, +#undef TP_printk +#define TP_printk(fmt, args...) fmt "\n", args -#undef TRACE_FIELD_SPECIAL -#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ - field->item, - - -#undef TP_RAW_FMT -#define TP_RAW_FMT(args...) args - -#undef TRACE_EVENT_FORMAT -#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ +#undef TRACE_EVENT +#define TRACE_EVENT(call, proto, args, tstruct, print, assign) \ enum print_line_t \ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ { \ @@ -66,14 +57,76 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ \ field = (typeof(field))entry; \ \ - ret = trace_seq_printf(s, tpfmt "%s", tstruct "\n"); \ + ret = trace_seq_printf(s, print); \ if (!ret) \ return TRACE_TYPE_PARTIAL_LINE; \ \ return TRACE_TYPE_HANDLED; \ } - + #include -#include "trace_format.h" +/* + * Setup the showing format of trace point. + * + * int + * ftrace_format_##call(struct trace_seq *s) + * { + * struct ftrace_raw_##call field; + * int ret; + * + * ret = trace_seq_printf(s, #type " " #item ";" + * " size:%d; offset:%d;\n", + * sizeof(field.type), + * offsetof(struct ftrace_raw_##call, + * item)); + * + * } + */ + +#undef TP_STRUCT__entry +#define TP_STRUCT__entry(args...) args + +#undef __field +#define __field(type, item) \ + ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ + if (!ret) \ + return 0; + +#undef __array +#define __array(type, item, len) \ + ret = trace_seq_printf(s, "\tfield:" #type " " #item "[" #len "];\t" \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ + if (!ret) \ + return 0; + +#undef __entry +#define __entry "REC" + +#undef TP_printk +#define TP_printk(fmt, args...) "%s, %s\n", #fmt, #args + +#undef TP_fast_assign +#define TP_fast_assign(args...) args + +#undef TRACE_EVENT +#define TRACE_EVENT(call, proto, args, tstruct, print, func) \ +static int \ +ftrace_format_##call(struct trace_seq *s) \ +{ \ + struct ftrace_raw_##call field; \ + int ret; \ + \ + tstruct; \ + \ + trace_seq_printf(s, "\nprint fmt: " print); \ + \ + return ret; \ +} + #include diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 41b82b93c9c7..8e398d864096 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -144,27 +144,15 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .unregfunc = ftrace_unreg_event_##call, \ } -#undef TRACE_FIELD -#define TRACE_FIELD(type, item, assign)\ - entry->item = assign; - -#undef TRACE_FIELD -#define TRACE_FIELD(type, item, assign)\ - entry->item = assign; - -#undef TP_CMD -#define TP_CMD(cmd...) cmd - -#undef TRACE_ENTRY -#define TRACE_ENTRY entry - -#undef TRACE_FIELD_SPECIAL -#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ - cmd; - #undef TRACE_EVENT_FORMAT -#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, tpfmt) \ -_TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) \ +#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, raw) \ + TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) + +#undef __entry +#define __entry entry + +#undef TRACE_EVENT +#define TRACE_EVENT(call, proto, args, tstruct, print, assign) \ \ static struct ftrace_event_call event_##call; \ \ @@ -185,7 +173,7 @@ static void ftrace_raw_event_##call(proto) \ return; \ entry = ring_buffer_event_data(event); \ \ - tstruct; \ + assign; \ \ trace_current_buffer_unlock_commit(event, irq_flags, pc); \ } \ @@ -226,10 +214,8 @@ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ .name = #call, \ .system = __stringify(TRACE_SYSTEM), \ - .regfunc = ftrace_reg_event_##call, \ - .unregfunc = ftrace_unreg_event_##call, \ .raw_init = ftrace_raw_init_event_##call, \ - .raw_reg = ftrace_raw_reg_event_##call, \ - .raw_unreg = ftrace_raw_unreg_event_##call, \ + .regfunc = ftrace_raw_reg_event_##call, \ + .unregfunc = ftrace_raw_unreg_event_##call, \ .show_format = ftrace_format_##call, \ } diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index e62bc10f8103..23ae78430d58 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -15,7 +15,28 @@ #include "trace_output.h" -#include "trace_format.h" + +#undef TRACE_STRUCT +#define TRACE_STRUCT(args...) args + +#undef TRACE_FIELD +#define TRACE_FIELD(type, item, assign) \ + ret = trace_seq_printf(s, "\tfield:" #type " " #item ";\t" \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ + if (!ret) \ + return 0; + + +#undef TRACE_FIELD_SPECIAL +#define TRACE_FIELD_SPECIAL(type_item, item, cmd) \ + ret = trace_seq_printf(s, "\tfield special:" #type_item ";\t" \ + "offset:%u;\tsize:%u;\n", \ + (unsigned int)offsetof(typeof(field), item), \ + (unsigned int)sizeof(field.item)); \ + if (!ret) \ + return 0; #undef TRACE_FIELD_ZERO_CHAR #define TRACE_FIELD_ZERO_CHAR(item) \ From 12b5fdb8bbb2d2fc31746d7b672c12fd8897aa08 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 9 Mar 2009 23:03:44 -0400 Subject: [PATCH 260/413] tracing: convert the sched trace points to the TRACE_EVENT macros Impact: enhancement This patch converts the rest of the sched trace points to use the new more powerful TRACE_EVENT macro. Signed-off-by: Steven Rostedt --- include/trace/sched_event_types.h | 352 ++++++++++++++++++++++-------- 1 file changed, 261 insertions(+), 91 deletions(-) diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index aa77fb754038..0bbbf410e01f 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -8,59 +8,137 @@ #undef TRACE_SYSTEM #define TRACE_SYSTEM sched -TRACE_EVENT_FORMAT(sched_kthread_stop, +/* + * Tracepoint for calling kthread_stop, performed to end a kthread: + */ +TRACE_EVENT(sched_kthread_stop, + TP_PROTO(struct task_struct *t), + TP_ARGS(t), - TP_FMT("task %s:%d", t->comm, t->pid), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, t->pid) - ), - TP_RAW_FMT("task %d") - ); -TRACE_EVENT_FORMAT(sched_kthread_stop_ret, + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + ), + + TP_printk("task %s:%d", __entry->comm, __entry->pid), + + TP_fast_assign( + memcpy(__entry->comm, t->comm, TASK_COMM_LEN); + __entry->pid = t->pid; + ) +); + +/* + * Tracepoint for the return value of the kthread stopping: + */ +TRACE_EVENT(sched_kthread_stop_ret, + TP_PROTO(int ret), + TP_ARGS(ret), - TP_FMT("ret=%d", ret), - TRACE_STRUCT( - TRACE_FIELD(int, ret, ret) - ), - TP_RAW_FMT("ret=%d") - ); -TRACE_EVENT_FORMAT(sched_wait_task, + TP_STRUCT__entry( + __field( int, ret ) + ), + + TP_printk("ret %d", __entry->ret), + + TP_fast_assign( + __entry->ret = ret; + ) +); + +/* + * Tracepoint for waiting on task to unschedule: + * + * (NOTE: the 'rq' argument is not used by generic trace events, + * but used by the latency tracer plugin. ) + */ +TRACE_EVENT(sched_wait_task, + TP_PROTO(struct rq *rq, struct task_struct *p), + TP_ARGS(rq, p), - TP_FMT("task %s:%d", p->comm, p->pid), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, p->pid) - ), - TP_RAW_FMT("task %d") - ); -TRACE_EVENT_FORMAT(sched_wakeup, - TP_PROTO(struct rq *rq, struct task_struct *p, int success), - TP_ARGS(rq, p, success), - TP_FMT("task %s:%d %s", - p->comm, p->pid, success ? "succeeded" : "failed"), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, p->pid) - TRACE_FIELD(int, success, success) + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, prio ) ), - TP_RAW_FMT("task %d success=%d") - ); -TRACE_EVENT_FORMAT(sched_wakeup_new, + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio), + + TP_fast_assign( + memcpy(__entry->comm, p->comm, TASK_COMM_LEN); + __entry->pid = p->pid; + __entry->prio = p->prio; + ) +); + +/* + * Tracepoint for waking up a task: + * + * (NOTE: the 'rq' argument is not used by generic trace events, + * but used by the latency tracer plugin. ) + */ +TRACE_EVENT(sched_wakeup, + TP_PROTO(struct rq *rq, struct task_struct *p, int success), + TP_ARGS(rq, p, success), - TP_FMT("task %s:%d", - p->comm, p->pid, success ? "succeeded" : "failed"), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, p->pid) - TRACE_FIELD(int, success, success) + + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, prio ) + __field( int, success ) ), - TP_RAW_FMT("task %d success=%d") - ); + + TP_printk("task %s:%d [%d] success=%d", + __entry->comm, __entry->pid, __entry->prio, + __entry->success), + + TP_fast_assign( + memcpy(__entry->comm, p->comm, TASK_COMM_LEN); + __entry->pid = p->pid; + __entry->prio = p->prio; + __entry->success = success; + ) +); + +/* + * Tracepoint for waking up a new task: + * + * (NOTE: the 'rq' argument is not used by generic trace events, + * but used by the latency tracer plugin. ) + */ +TRACE_EVENT(sched_wakeup_new, + + TP_PROTO(struct rq *rq, struct task_struct *p, int success), + + TP_ARGS(rq, p, success), + + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, prio ) + __field( int, success ) + ), + + TP_printk("task %s:%d [%d] success=%d", + __entry->comm, __entry->pid, __entry->prio, + __entry->success), + + TP_fast_assign( + memcpy(__entry->comm, p->comm, TASK_COMM_LEN); + __entry->pid = p->pid; + __entry->prio = p->prio; + __entry->success = success; + ) +); /* * Tracepoint for task switches, performed by the scheduler: @@ -98,70 +176,162 @@ TRACE_EVENT(sched_switch, ) ); -TRACE_EVENT_FORMAT(sched_migrate_task, +/* + * Tracepoint for a task being migrated: + */ +TRACE_EVENT(sched_migrate_task, + TP_PROTO(struct task_struct *p, int orig_cpu, int dest_cpu), + TP_ARGS(p, orig_cpu, dest_cpu), - TP_FMT("task %s:%d from: %d to: %d", - p->comm, p->pid, orig_cpu, dest_cpu), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, p->pid) - TRACE_FIELD(int, orig_cpu, orig_cpu) - TRACE_FIELD(int, dest_cpu, dest_cpu) - ), - TP_RAW_FMT("task %d from: %d to: %d") - ); -TRACE_EVENT_FORMAT(sched_process_free, + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, prio ) + __field( int, orig_cpu ) + __field( int, dest_cpu ) + ), + + TP_printk("task %s:%d [%d] from: %d to: %d", + __entry->comm, __entry->pid, __entry->prio, + __entry->orig_cpu, __entry->dest_cpu), + + TP_fast_assign( + memcpy(__entry->comm, p->comm, TASK_COMM_LEN); + __entry->pid = p->pid; + __entry->prio = p->prio; + __entry->orig_cpu = orig_cpu; + __entry->dest_cpu = dest_cpu; + ) +); + +/* + * Tracepoint for freeing a task: + */ +TRACE_EVENT(sched_process_free, + TP_PROTO(struct task_struct *p), - TP_ARGS(p), - TP_FMT("task %s:%d", p->comm, p->pid), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, p->pid) - ), - TP_RAW_FMT("task %d") - ); -TRACE_EVENT_FORMAT(sched_process_exit, + TP_ARGS(p), + + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, prio ) + ), + + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio), + + TP_fast_assign( + memcpy(__entry->comm, p->comm, TASK_COMM_LEN); + __entry->pid = p->pid; + __entry->prio = p->prio; + ) +); + +/* + * Tracepoint for a task exiting: + */ +TRACE_EVENT(sched_process_exit, + TP_PROTO(struct task_struct *p), - TP_ARGS(p), - TP_FMT("task %s:%d", p->comm, p->pid), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, p->pid) - ), - TP_RAW_FMT("task %d") - ); -TRACE_EVENT_FORMAT(sched_process_wait, + TP_ARGS(p), + + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, prio ) + ), + + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio), + + TP_fast_assign( + memcpy(__entry->comm, p->comm, TASK_COMM_LEN); + __entry->pid = p->pid; + __entry->prio = p->prio; + ) +); + +/* + * Tracepoint for a waiting task: + */ +TRACE_EVENT(sched_process_wait, + TP_PROTO(struct pid *pid), + TP_ARGS(pid), - TP_FMT("pid %d", pid_nr(pid)), - TRACE_STRUCT( - TRACE_FIELD(pid_t, pid, pid_nr(pid)) - ), - TP_RAW_FMT("task %d") - ); -TRACE_EVENT_FORMAT(sched_process_fork, + TP_STRUCT__entry( + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + __field( int, prio ) + ), + + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio), + + TP_fast_assign( + memcpy(__entry->comm, current->comm, TASK_COMM_LEN); + __entry->pid = pid_nr(pid); + __entry->prio = current->prio; + ) +); + +/* + * Tracepoint for do_fork: + */ +TRACE_EVENT(sched_process_fork, + TP_PROTO(struct task_struct *parent, struct task_struct *child), - TP_ARGS(parent, child), - TP_FMT("parent %s:%d child %s:%d", - parent->comm, parent->pid, child->comm, child->pid), - TRACE_STRUCT( - TRACE_FIELD(pid_t, parent, parent->pid) - TRACE_FIELD(pid_t, child, child->pid) - ), - TP_RAW_FMT("parent %d child %d") - ); -TRACE_EVENT_FORMAT(sched_signal_send, - TP_PROTO(int sig, struct task_struct *p), - TP_ARGS(sig, p), - TP_FMT("sig: %d task %s:%d", sig, p->comm, p->pid), - TRACE_STRUCT( - TRACE_FIELD(int, sig, sig) - TRACE_FIELD(pid_t, pid, p->pid) + TP_ARGS(parent, child), + + TP_STRUCT__entry( + __array( char, parent_comm, TASK_COMM_LEN ) + __field( pid_t, parent_pid ) + __array( char, child_comm, TASK_COMM_LEN ) + __field( pid_t, child_pid ) ), - TP_RAW_FMT("sig: %d task %d") - ); + + TP_printk("parent %s:%d child %s:%d", + __entry->parent_comm, __entry->parent_pid, + __entry->child_comm, __entry->child_pid), + + TP_fast_assign( + memcpy(__entry->parent_comm, parent->comm, TASK_COMM_LEN); + __entry->parent_pid = parent->pid; + memcpy(__entry->child_comm, child->comm, TASK_COMM_LEN); + __entry->child_pid = child->pid; + ) +); + +/* + * Tracepoint for sending a signal: + */ +TRACE_EVENT(sched_signal_send, + + TP_PROTO(int sig, struct task_struct *p), + + TP_ARGS(sig, p), + + TP_STRUCT__entry( + __field( int, sig ) + __array( char, comm, TASK_COMM_LEN ) + __field( pid_t, pid ) + ), + + TP_printk("sig: %d task %s:%d", + __entry->sig, __entry->comm, __entry->pid), + + TP_fast_assign( + memcpy(__entry->comm, p->comm, TASK_COMM_LEN); + __entry->pid = p->pid; + __entry->sig = sig; + ) +); #undef TRACE_SYSTEM From d6e2ca4c05be6a5ab16030a9f227301bd6acc9f0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 9 Mar 2009 23:23:30 -0400 Subject: [PATCH 261/413] tracing: convert irq trace points to new macros Impact: enhancement Converted the two irq trace point macros. The entry macro copies the name of the irq handler, thus it is better to simply use the TRACE_FORMAT macro which uses the trace_printk. The return of the handler does not need to record the name, thus the faster C style handler is more approriate. Signed-off-by: Steven Rostedt --- include/trace/irq_event_types.h | 38 +++++++++++++++++++++------------ 1 file changed, 24 insertions(+), 14 deletions(-) diff --git a/include/trace/irq_event_types.h b/include/trace/irq_event_types.h index 0147d9eef5f4..43bcb74dd49f 100644 --- a/include/trace/irq_event_types.h +++ b/include/trace/irq_event_types.h @@ -8,26 +8,36 @@ #undef TRACE_SYSTEM #define TRACE_SYSTEM irq -TRACE_EVENT_FORMAT(irq_handler_entry, +/* + * Tracepoint for entry of interrupt handler: + */ +TRACE_FORMAT(irq_handler_entry, TP_PROTO(int irq, struct irqaction *action), TP_ARGS(irq, action), - TP_FMT("irq=%d handler=%s", irq, action->name), - TRACE_STRUCT( - TRACE_FIELD(int, irq, irq) - ), - TP_RAW_FMT("irq %d") + TP_FMT("irq=%d handler=%s", irq, action->name) ); -TRACE_EVENT_FORMAT(irq_handler_exit, +/* + * Tracepoint for return of an interrupt handler: + */ +TRACE_EVENT(irq_handler_exit, + TP_PROTO(int irq, struct irqaction *action, int ret), + TP_ARGS(irq, action, ret), - TP_FMT("irq=%d handler=%s return=%s", - irq, action->name, ret ? "handled" : "unhandled"), - TRACE_STRUCT( - TRACE_FIELD(int, irq, irq) - TRACE_FIELD(int, ret, ret) + + TP_STRUCT__entry( + __field( int, irq ) + __field( int, ret ) ), - TP_RAW_FMT("irq %d ret %d") - ); + + TP_printk("irq=%d return=%s", + __entry->irq, __entry->ret ? "handled" : "unhandled"), + + TP_fast_assign( + __entry->irq = irq; + __entry->ret = ret; + ) +); #undef TRACE_SYSTEM From 157587d7ac555458da9f682e3250135e468470a6 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 00:15:34 -0400 Subject: [PATCH 262/413] tracing: remove obsolete TRACE_EVENT_FORMAT macro Impact: clean up The TRACE_EVENT_FORMAT macro is no longer used by trace points and only the DECLARE_TRACE, TRACE_FORMAT or TRACE_EVENT macros should be used by them. Although the TRACE_EVENT_FORMAT macro is still used by the internal tracing utility, it should not be used in core kernel code. Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 3 --- include/trace/lockdep_event_types.h | 2 +- include/trace/sched_event_types.h | 2 +- kernel/trace/trace_events_stage_1.h | 3 --- kernel/trace/trace_events_stage_3.h | 6 +----- 5 files changed, 3 insertions(+), 13 deletions(-) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 6b4f1bb3701e..69b56988813d 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -157,9 +157,6 @@ static inline void tracepoint_synchronize_unregister(void) #define TRACE_FORMAT(name, proto, args, fmt) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) -#define TRACE_EVENT_FORMAT(name, proto, args, fmt, struct, tpfmt) \ - TRACE_FORMAT(name, PARAMS(proto), PARAMS(args), PARAMS(fmt)) - #define TRACE_EVENT(name, proto, args, struct, print, assign) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) diff --git a/include/trace/lockdep_event_types.h b/include/trace/lockdep_event_types.h index 1f00e8b3543e..adccfcd2ec8f 100644 --- a/include/trace/lockdep_event_types.h +++ b/include/trace/lockdep_event_types.h @@ -1,5 +1,5 @@ -#ifndef TRACE_EVENT_FORMAT +#ifndef TRACE_FORMAT # error Do not include this file directly. # error Unless you know what you are doing. #endif diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index 0bbbf410e01f..fb37af672c88 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -1,6 +1,6 @@ /* use instead */ -#ifndef TRACE_EVENT_FORMAT +#ifndef TRACE_EVENT # error Do not include this file directly. # error Unless you know what you are doing. #endif diff --git a/kernel/trace/trace_events_stage_1.h b/kernel/trace/trace_events_stage_1.h index edfcbd3a0d1b..15e9bf965a18 100644 --- a/kernel/trace/trace_events_stage_1.h +++ b/kernel/trace/trace_events_stage_1.h @@ -17,9 +17,6 @@ #undef TRACE_FORMAT #define TRACE_FORMAT(call, proto, args, fmt) -#undef TRACE_EVENT_FORMAT -#define TRACE_EVENT_FORMAT(name, proto, args, fmt, tstruct, tpfmt) - #undef __array #define __array(type, item, len) type item[len]; diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 8e398d864096..3ba55d4ab073 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -35,7 +35,7 @@ * } * * - * For those macros defined with TRACE_EVENT_FORMAT: + * For those macros defined with TRACE_EVENT: * * static struct ftrace_event_call event_; * @@ -144,10 +144,6 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .unregfunc = ftrace_unreg_event_##call, \ } -#undef TRACE_EVENT_FORMAT -#define TRACE_EVENT_FORMAT(call, proto, args, fmt, tstruct, raw) \ - TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) - #undef __entry #define __entry entry From 631595fbf4aeac260e664a8a002897e4db6a50dd Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Tue, 10 Mar 2009 13:57:10 +0900 Subject: [PATCH 263/413] doc: add trace_buf_size description to kernel-parameters.txt from early boot tracing view, trace_buf_size parameter is important. it should be documented. Signed-off-by: KOSAKI Motohiro LKML-Reference: <20090310135200.A48B.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- Documentation/kernel-parameters.txt | 2 ++ 1 file changed, 2 insertions(+) diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 454f42b21f16..7643483bdd6a 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -2344,6 +2344,8 @@ and is between 256 and 4096 characters. It is defined in the file tp720= [HW,PS2] + trace_buf_size=nn[KMG] [ftrace] will set tracing buffer size. + trix= [HW,OSS] MediaTrix AudioTrix Pro Format: ,,,,,,,, From bbcd3063597a3824357cd83c501c2a2aa21ef37b Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Tue, 10 Mar 2009 10:49:53 +0900 Subject: [PATCH 264/413] tracing: Don't assume possible cpu list have continuous numbers "for (++cpu ; cpu < num_possible_cpus(); cpu++)" statement assumes possible cpus have continuous number - but that's a wrong assumption. Insted, cpumask_next() should be used. Signed-off-by: KOSAKI Motohiro Cc: Lai Jiangshan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <20090310104437.A480.A69D9226@jp.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_workqueue.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index 46c8dc896bd3..739fdacf873b 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -91,7 +91,7 @@ static void probe_workqueue_creation(struct task_struct *wq_thread, int cpu) struct cpu_workqueue_stats *cws; unsigned long flags; - WARN_ON(cpu < 0 || cpu >= num_possible_cpus()); + WARN_ON(cpu < 0); /* Workqueues are sometimes created in atomic context */ cws = kzalloc(sizeof(struct cpu_workqueue_stats), GFP_ATOMIC); @@ -175,12 +175,12 @@ static void *workqueue_stat_next(void *prev, int idx) spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); if (list_is_last(&prev_cws->list, &workqueue_cpu_stat(cpu)->list)) { spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); - for (++cpu ; cpu < num_possible_cpus(); cpu++) { - ret = workqueue_stat_start_cpu(cpu); - if (ret) - return ret; - } - return NULL; + do { + cpu = cpumask_next(cpu, cpu_possible_mask); + if (cpu >= nr_cpu_ids) + return NULL; + } while (!(ret = workqueue_stat_start_cpu(cpu))); + return ret; } spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); From ce8eb2bf05042452107e489782105d2e235cbdd0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 10:14:35 -0400 Subject: [PATCH 265/413] tracing: fix printk format specifier Impact: clean up The offsetof and sizeof are of type size_t, and instead of typecasting them to unsigned int for printk formatting, one could just use %zu. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 1880a6438097..a0b41cc26f26 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -338,8 +338,7 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, #undef FIELD #define FIELD(type, name) \ - #type, #name, (unsigned int)offsetof(typeof(field), name), \ - (unsigned int)sizeof(field.name) + #type, #name, offsetof(typeof(field), name), sizeof(field.name) static int trace_write_header(struct trace_seq *s) { @@ -347,11 +346,11 @@ static int trace_write_header(struct trace_seq *s) /* struct trace_entry */ return trace_seq_printf(s, - "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" - "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" - "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" - "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" - "\tfield:%s %s;\toffset:%u;\tsize:%u;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" + "\tfield:%s %s;\toffset:%zu;\tsize:%zu;\n" "\n", FIELD(unsigned char, type), FIELD(unsigned char, flags), From 7cf49427042400d40bdc80b5c3399b6b5945afa8 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Mon, 9 Mar 2009 12:40:40 -0400 Subject: [PATCH 266/413] x86: expand irq-off region in text_poke() Expand irq-off region to cover fixmap using code and cache synchronizing. Signed-off-by: Masami Hiramatsu LKML-Reference: <49B54688.8090403@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/alternative.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c index 2d903b760ddb..f57658702571 100644 --- a/arch/x86/kernel/alternative.c +++ b/arch/x86/kernel/alternative.c @@ -526,13 +526,12 @@ void *__kprobes text_poke(void *addr, const void *opcode, size_t len) pages[1] = virt_to_page(addr + PAGE_SIZE); } BUG_ON(!pages[0]); + local_irq_save(flags); set_fixmap(FIX_TEXT_POKE0, page_to_phys(pages[0])); if (pages[1]) set_fixmap(FIX_TEXT_POKE1, page_to_phys(pages[1])); vaddr = (char *)fix_to_virt(FIX_TEXT_POKE0); - local_irq_save(flags); memcpy(&vaddr[(unsigned long)addr & ~PAGE_MASK], opcode, len); - local_irq_restore(flags); clear_fixmap(FIX_TEXT_POKE0); if (pages[1]) clear_fixmap(FIX_TEXT_POKE1); @@ -542,5 +541,6 @@ void *__kprobes text_poke(void *addr, const void *opcode, size_t len) that causes hangs on some VIA CPUs. */ for (i = 0; i < len; i++) BUG_ON(((char *)addr)[i] != ((char *)opcode)[i]); + local_irq_restore(flags); return addr; } From 40e26815fafd3b8c4aced17b1f22e68ef33eb8db Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 11:32:40 -0400 Subject: [PATCH 267/413] tracing: do not allow modifying the ftrace events via the event files Impact: fix to prevent crash on calling NULL function pointer The ftrace internal records have their format exported via the event system under the ftrace subsystem. These are only for exporting the format to allow binary readers to be able to parse them in a binary output. The ftrace subsystem events can only be enabled via the ftrace tracers and do not have a registering function. The event files expect the event record to have registering function and will call it directly. Passing in a ftrace subsystem event will cause the kernel to crash because it will execute a NULL pointer. This patch prevents the ftrace subsystem from being viewable to the event enabling files. Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 18 +++++++++++++++--- 1 file changed, 15 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index a0b41cc26f26..85ec10fbb38d 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -102,7 +102,7 @@ static int ftrace_set_clr_event(char *buf, int set) mutex_lock(&event_mutex); events_for_each(call) { - if (!call->name) + if (!call->name || !call->regfunc) continue; if (match && @@ -207,8 +207,20 @@ t_next(struct seq_file *m, void *v, loff_t *pos) (*pos)++; - if ((unsigned long)call >= (unsigned long)__stop_ftrace_events) - return NULL; + for (;;) { + if ((unsigned long)call >= (unsigned long)__stop_ftrace_events) + return NULL; + + /* + * The ftrace subsystem is for showing formats only. + * They can not be enabled or disabled via the event files. + */ + if (call->regfunc) + break; + + call++; + next = call; + } m->private = ++next; From 2314c4ae1461c9e8b26cf8b9a851f280bc5769e1 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 12:04:02 -0400 Subject: [PATCH 268/413] tracing: add back the available_events file The event directory files type and available_types were no longer needed with the new TRACE_EVENT_FORMAT macros, they were deleted. But by accident the available_events file was also removed. This patch brings it back. Reported-by: KOSAKI Motohiro Signed-off-by: Steven Rostedt --- kernel/trace/trace_events.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 85ec10fbb38d..769dfd00fc85 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -428,6 +428,13 @@ static const struct seq_operations show_set_event_seq_ops = { .stop = t_stop, }; +static const struct file_operations ftrace_avail_fops = { + .open = ftrace_event_seq_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + static const struct file_operations ftrace_set_event_fops = { .open = ftrace_event_seq_open, .read = seq_read, @@ -569,6 +576,13 @@ static __init int event_trace_init(void) if (!d_tracer) return 0; + entry = debugfs_create_file("available_events", 0444, d_tracer, + (void *)&show_event_seq_ops, + &ftrace_avail_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'available_events' entry\n"); + entry = debugfs_create_file("set_event", 0644, d_tracer, (void *)&show_set_event_seq_ops, &ftrace_set_event_fops); From 30a8fecc2d34f086df34fe2f2b926f080e002600 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 12:41:38 -0400 Subject: [PATCH 269/413] tracing: flip the TP_printk and TP_fast_assign in the TRACE_EVENT macro Impact: clean up In trying to stay consistant with the C style format in the TRACE_EVENT macro, it makes more sense to do the printk after the assigning of the variables. Reported-by: Ingo Molnar Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 2 +- include/trace/irq_event_types.h | 8 +-- include/trace/sched_event_types.h | 102 ++++++++++++++-------------- kernel/trace/trace_events_stage_1.h | 2 +- kernel/trace/trace_events_stage_2.h | 4 +- kernel/trace/trace_events_stage_3.h | 2 +- 6 files changed, 60 insertions(+), 60 deletions(-) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 69b56988813d..c7b09452514b 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -157,7 +157,7 @@ static inline void tracepoint_synchronize_unregister(void) #define TRACE_FORMAT(name, proto, args, fmt) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) -#define TRACE_EVENT(name, proto, args, struct, print, assign) \ +#define TRACE_EVENT(name, proto, args, struct, assign, print) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) #endif diff --git a/include/trace/irq_event_types.h b/include/trace/irq_event_types.h index 43bcb74dd49f..214bb928fe9e 100644 --- a/include/trace/irq_event_types.h +++ b/include/trace/irq_event_types.h @@ -31,13 +31,13 @@ TRACE_EVENT(irq_handler_exit, __field( int, ret ) ), - TP_printk("irq=%d return=%s", - __entry->irq, __entry->ret ? "handled" : "unhandled"), - TP_fast_assign( __entry->irq = irq; __entry->ret = ret; - ) + ), + + TP_printk("irq=%d return=%s", + __entry->irq, __entry->ret ? "handled" : "unhandled") ); #undef TRACE_SYSTEM diff --git a/include/trace/sched_event_types.h b/include/trace/sched_event_types.h index fb37af672c88..63547dc1125f 100644 --- a/include/trace/sched_event_types.h +++ b/include/trace/sched_event_types.h @@ -22,12 +22,12 @@ TRACE_EVENT(sched_kthread_stop, __field( pid_t, pid ) ), - TP_printk("task %s:%d", __entry->comm, __entry->pid), - TP_fast_assign( memcpy(__entry->comm, t->comm, TASK_COMM_LEN); __entry->pid = t->pid; - ) + ), + + TP_printk("task %s:%d", __entry->comm, __entry->pid) ); /* @@ -43,11 +43,11 @@ TRACE_EVENT(sched_kthread_stop_ret, __field( int, ret ) ), - TP_printk("ret %d", __entry->ret), - TP_fast_assign( __entry->ret = ret; - ) + ), + + TP_printk("ret %d", __entry->ret) ); /* @@ -68,14 +68,14 @@ TRACE_EVENT(sched_wait_task, __field( int, prio ) ), - TP_printk("task %s:%d [%d]", - __entry->comm, __entry->pid, __entry->prio), - TP_fast_assign( memcpy(__entry->comm, p->comm, TASK_COMM_LEN); __entry->pid = p->pid; __entry->prio = p->prio; - ) + ), + + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio) ); /* @@ -97,16 +97,16 @@ TRACE_EVENT(sched_wakeup, __field( int, success ) ), - TP_printk("task %s:%d [%d] success=%d", - __entry->comm, __entry->pid, __entry->prio, - __entry->success), - TP_fast_assign( memcpy(__entry->comm, p->comm, TASK_COMM_LEN); __entry->pid = p->pid; __entry->prio = p->prio; __entry->success = success; - ) + ), + + TP_printk("task %s:%d [%d] success=%d", + __entry->comm, __entry->pid, __entry->prio, + __entry->success) ); /* @@ -128,16 +128,16 @@ TRACE_EVENT(sched_wakeup_new, __field( int, success ) ), - TP_printk("task %s:%d [%d] success=%d", - __entry->comm, __entry->pid, __entry->prio, - __entry->success), - TP_fast_assign( memcpy(__entry->comm, p->comm, TASK_COMM_LEN); __entry->pid = p->pid; __entry->prio = p->prio; __entry->success = success; - ) + ), + + TP_printk("task %s:%d [%d] success=%d", + __entry->comm, __entry->pid, __entry->prio, + __entry->success) ); /* @@ -162,10 +162,6 @@ TRACE_EVENT(sched_switch, __field( int, next_prio ) ), - TP_printk("task %s:%d [%d] ==> %s:%d [%d]", - __entry->prev_comm, __entry->prev_pid, __entry->prev_prio, - __entry->next_comm, __entry->next_pid, __entry->next_prio), - TP_fast_assign( memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN); __entry->prev_pid = prev->pid; @@ -173,7 +169,11 @@ TRACE_EVENT(sched_switch, memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN); __entry->next_pid = next->pid; __entry->next_prio = next->prio; - ) + ), + + TP_printk("task %s:%d [%d] ==> %s:%d [%d]", + __entry->prev_comm, __entry->prev_pid, __entry->prev_prio, + __entry->next_comm, __entry->next_pid, __entry->next_prio) ); /* @@ -193,17 +193,17 @@ TRACE_EVENT(sched_migrate_task, __field( int, dest_cpu ) ), - TP_printk("task %s:%d [%d] from: %d to: %d", - __entry->comm, __entry->pid, __entry->prio, - __entry->orig_cpu, __entry->dest_cpu), - TP_fast_assign( memcpy(__entry->comm, p->comm, TASK_COMM_LEN); __entry->pid = p->pid; __entry->prio = p->prio; __entry->orig_cpu = orig_cpu; __entry->dest_cpu = dest_cpu; - ) + ), + + TP_printk("task %s:%d [%d] from: %d to: %d", + __entry->comm, __entry->pid, __entry->prio, + __entry->orig_cpu, __entry->dest_cpu) ); /* @@ -221,14 +221,14 @@ TRACE_EVENT(sched_process_free, __field( int, prio ) ), - TP_printk("task %s:%d [%d]", - __entry->comm, __entry->pid, __entry->prio), - TP_fast_assign( memcpy(__entry->comm, p->comm, TASK_COMM_LEN); __entry->pid = p->pid; __entry->prio = p->prio; - ) + ), + + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio) ); /* @@ -246,14 +246,14 @@ TRACE_EVENT(sched_process_exit, __field( int, prio ) ), - TP_printk("task %s:%d [%d]", - __entry->comm, __entry->pid, __entry->prio), - TP_fast_assign( memcpy(__entry->comm, p->comm, TASK_COMM_LEN); __entry->pid = p->pid; __entry->prio = p->prio; - ) + ), + + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio) ); /* @@ -271,14 +271,14 @@ TRACE_EVENT(sched_process_wait, __field( int, prio ) ), - TP_printk("task %s:%d [%d]", - __entry->comm, __entry->pid, __entry->prio), - TP_fast_assign( memcpy(__entry->comm, current->comm, TASK_COMM_LEN); __entry->pid = pid_nr(pid); __entry->prio = current->prio; - ) + ), + + TP_printk("task %s:%d [%d]", + __entry->comm, __entry->pid, __entry->prio) ); /* @@ -297,16 +297,16 @@ TRACE_EVENT(sched_process_fork, __field( pid_t, child_pid ) ), - TP_printk("parent %s:%d child %s:%d", - __entry->parent_comm, __entry->parent_pid, - __entry->child_comm, __entry->child_pid), - TP_fast_assign( memcpy(__entry->parent_comm, parent->comm, TASK_COMM_LEN); __entry->parent_pid = parent->pid; memcpy(__entry->child_comm, child->comm, TASK_COMM_LEN); __entry->child_pid = child->pid; - ) + ), + + TP_printk("parent %s:%d child %s:%d", + __entry->parent_comm, __entry->parent_pid, + __entry->child_comm, __entry->child_pid) ); /* @@ -324,14 +324,14 @@ TRACE_EVENT(sched_signal_send, __field( pid_t, pid ) ), - TP_printk("sig: %d task %s:%d", - __entry->sig, __entry->comm, __entry->pid), - TP_fast_assign( memcpy(__entry->comm, p->comm, TASK_COMM_LEN); __entry->pid = p->pid; __entry->sig = sig; - ) + ), + + TP_printk("sig: %d task %s:%d", + __entry->sig, __entry->comm, __entry->pid) ); #undef TRACE_SYSTEM diff --git a/kernel/trace/trace_events_stage_1.h b/kernel/trace/trace_events_stage_1.h index 15e9bf965a18..82f68443c556 100644 --- a/kernel/trace/trace_events_stage_1.h +++ b/kernel/trace/trace_events_stage_1.h @@ -27,7 +27,7 @@ #define TP_STRUCT__entry(args...) args #undef TRACE_EVENT -#define TRACE_EVENT(name, proto, args, tstruct, print, assign) \ +#define TRACE_EVENT(name, proto, args, tstruct, assign, print) \ struct ftrace_raw_##name { \ struct trace_entry ent; \ tstruct \ diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index d91bf4c56661..1ad9f8d2fe45 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -39,7 +39,7 @@ #define TP_printk(fmt, args...) fmt "\n", args #undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, print, assign) \ +#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ enum print_line_t \ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ { \ @@ -115,7 +115,7 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ #define TP_fast_assign(args...) args #undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, print, func) \ +#define TRACE_EVENT(call, proto, args, tstruct, func, print) \ static int \ ftrace_format_##call(struct trace_seq *s) \ { \ diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 3ba55d4ab073..d6de06b9201a 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -148,7 +148,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ #define __entry entry #undef TRACE_EVENT -#define TRACE_EVENT(call, proto, args, tstruct, print, assign) \ +#define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ \ static struct ftrace_event_call event_##call; \ \ From 823f9124fb2e33eeb624d139978a52089f8a02ae Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 12:58:51 -0400 Subject: [PATCH 270/413] tracing: document TRACE_EVENT macro in tracepoint.h Impact: clean up / comments Kosaki Motohiro asked about an explanation to the TRACE_EVENT macro. Ingo Molnar replied with a nice description. This patch takes the description that Ingo wrote (with some slight modifications) and adds it to the tracepoint.h file. Reported-by: KOSAKI Motohiro Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 103 +++++++++++++++++++++++++++++++++++++ 1 file changed, 103 insertions(+) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index c7b09452514b..119ece224c21 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -157,6 +157,109 @@ static inline void tracepoint_synchronize_unregister(void) #define TRACE_FORMAT(name, proto, args, fmt) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) + +/* + * For use with the TRACE_EVENT macro: + * + * We define a tracepoint, its arguments, its printk format + * and its 'fast binay record' layout. + * + * Firstly, name your tracepoint via TRACE_EVENT(name : the + * 'subsystem_event' notation is fine. + * + * Think about this whole construct as the + * 'trace_sched_switch() function' from now on. + * + * + * TRACE_EVENT(sched_switch, + * + * * + * * A function has a regular function arguments + * * prototype, declare it via TP_PROTO(): + * * + * + * TP_PROTO(struct rq *rq, struct task_struct *prev, + * struct task_struct *next), + * + * * + * * Define the call signature of the 'function'. + * * (Design sidenote: we use this instead of a + * * TP_PROTO1/TP_PROTO2/TP_PROTO3 ugliness.) + * * + * + * TP_ARGS(rq, prev, next), + * + * * + * * Fast binary tracing: define the trace record via + * * TP_STRUCT__entry(). You can think about it like a + * * regular C structure local variable definition. + * * + * * This is how the trace record is structured and will + * * be saved into the ring buffer. These are the fields + * * that will be exposed to user-space in + * * /debug/tracing/events/<*>/format. + * * + * * The declared 'local variable' is called '__entry' + * * + * * __field(pid_t, prev_prid) is equivalent to a standard declariton: + * * + * * pid_t prev_pid; + * * + * * __array(char, prev_comm, TASK_COMM_LEN) is equivalent to: + * * + * * char prev_comm[TASK_COMM_LEN]; + * * + * + * TP_STRUCT__entry( + * __array( char, prev_comm, TASK_COMM_LEN ) + * __field( pid_t, prev_pid ) + * __field( int, prev_prio ) + * __array( char, next_comm, TASK_COMM_LEN ) + * __field( pid_t, next_pid ) + * __field( int, next_prio ) + * ), + * + * * + * * Assign the entry into the trace record, by embedding + * * a full C statement block into TP_fast_assign(). You + * * can refer to the trace record as '__entry' - + * * otherwise you can put arbitrary C code in here. + * * + * * Note: this C code will execute every time a trace event + * * happens, on an active tracepoint. + * * + * + * TP_fast_assign( + * memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN); + * __entry->prev_pid = prev->pid; + * __entry->prev_prio = prev->prio; + * memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN); + * __entry->next_pid = next->pid; + * __entry->next_prio = next->prio; + * ) + * + * * + * * Formatted output of a trace record via TP_printk(). + * * This is how the tracepoint will appear under ftrace + * * plugins that make use of this tracepoint. + * * + * * (raw-binary tracing wont actually perform this step.) + * * + * + * TP_printk("task %s:%d [%d] ==> %s:%d [%d]", + * __entry->prev_comm, __entry->prev_pid, __entry->prev_prio, + * __entry->next_comm, __entry->next_pid, __entry->next_prio), + * + * ); + * + * This macro construct is thus used for the regular printk format + * tracing setup, it is used to construct a function pointer based + * tracepoint callback (this is used by programmatic plugins and + * can also by used by generic instrumentation like SystemTap), and + * it is also used to expose a structured trace record in + * /debug/tracing/events/. + */ + #define TRACE_EVENT(name, proto, args, struct, assign, print) \ DECLARE_TRACE(name, PARAMS(proto), PARAMS(args)) From 0e3d0f0566f3fcf664782f597070bbc669d78454 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 13:12:58 -0400 Subject: [PATCH 271/413] tracing: update comments to match event code macros Impact: clean up / comments The comments that described the ftrace macros to manipulate the TRACE_EVENT and TRACE_FORMAT macros no longer match the code. This patch updates them. Reported-by: KOSAKI Motohiro Signed-off-by: Steven Rostedt --- kernel/trace/trace_events_stage_1.h | 6 ++++-- kernel/trace/trace_events_stage_2.h | 9 ++++----- kernel/trace/trace_events_stage_3.h | 8 ++++---- 3 files changed, 12 insertions(+), 11 deletions(-) diff --git a/kernel/trace/trace_events_stage_1.h b/kernel/trace/trace_events_stage_1.h index 82f68443c556..38985f9b379c 100644 --- a/kernel/trace/trace_events_stage_1.h +++ b/kernel/trace/trace_events_stage_1.h @@ -6,11 +6,13 @@ * struct ftrace_raw_ { * struct trace_entry ent; * ; + * []; * [...] * }; * - * The is created by the TRACE_FIELD(type, item, assign) - * macro. We simply do "type item;", and that will create the fields + * The is created by the __field(type, item) macro or + * the __array(type2, item2, len) macro. + * We simply do "type item;", and that will create the fields * in the structure. */ diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index 1ad9f8d2fe45..ca347afd6aa0 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -20,7 +20,7 @@ * * field = (typeof(field))entry; * - * ret = trace_seq_printf(s, "%s", "\n"); + * ret = trace_seq_printf(s, "\n"); * if (!ret) * return TRACE_TYPE_PARTIAL_LINE; * @@ -76,10 +76,9 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ * int ret; * * ret = trace_seq_printf(s, #type " " #item ";" - * " size:%d; offset:%d;\n", - * sizeof(field.type), - * offsetof(struct ftrace_raw_##call, - * item)); + * " offset:%u; size:%u;\n", + * offsetof(struct ftrace_raw_##call, item), + * sizeof(field.type)); * * } */ diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index d6de06b9201a..6ee1de59f19d 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -56,7 +56,8 @@ * return; * entry = ring_buffer_event_data(event); * - * ; <-- Here we assign the entries by the TRACE_FIELD. + * ; <-- Here we assign the entries by the __field and + * __array macros. * * trace_current_buffer_unlock_commit(event, irq_flags, pc); * } @@ -96,11 +97,10 @@ * __attribute__((__aligned__(4))) * __attribute__((section("_ftrace_events"))) event_ = { * .name = "", + * .system = "", + * .raw_init = ftrace_raw_init_event_, * .regfunc = ftrace_reg_event_, * .unregfunc = ftrace_unreg_event_, - * .raw_init = ftrace_raw_init_event_, - * .raw_reg = ftrace_raw_reg_event_, - * .raw_unreg = ftrace_raw_unreg_event_, * .show_format = ftrace_format_, * } * From ef18012b248b47ec9a12c3a83ca5e99782d39c5d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 14:10:56 -0400 Subject: [PATCH 272/413] tracing: remove funky whitespace in the trace code Impact: clean up There existed a lot of 's in the tracing code. This patch removes them. Signed-off-by: Steven Rostedt --- include/linux/tracepoint.h | 16 ++--- kernel/trace/blktrace.c | 10 +-- kernel/trace/trace.c | 2 +- kernel/trace/trace_branch.c | 2 +- kernel/trace/trace_events_stage_3.h | 92 ++++++++++++++-------------- kernel/trace/trace_export.c | 2 +- kernel/trace/trace_functions_graph.c | 6 +- kernel/trace/trace_output.c | 14 ++--- kernel/trace/trace_workqueue.c | 6 +- 9 files changed, 75 insertions(+), 75 deletions(-) diff --git a/include/linux/tracepoint.h b/include/linux/tracepoint.h index 119ece224c21..d35a7ee7611f 100644 --- a/include/linux/tracepoint.h +++ b/include/linux/tracepoint.h @@ -178,8 +178,8 @@ static inline void tracepoint_synchronize_unregister(void) * * prototype, declare it via TP_PROTO(): * * * - * TP_PROTO(struct rq *rq, struct task_struct *prev, - * struct task_struct *next), + * TP_PROTO(struct rq *rq, struct task_struct *prev, + * struct task_struct *next), * * * * * Define the call signature of the 'function'. @@ -187,7 +187,7 @@ static inline void tracepoint_synchronize_unregister(void) * * TP_PROTO1/TP_PROTO2/TP_PROTO3 ugliness.) * * * - * TP_ARGS(rq, prev, next), + * TP_ARGS(rq, prev, next), * * * * * Fast binary tracing: define the trace record via @@ -229,13 +229,13 @@ static inline void tracepoint_synchronize_unregister(void) * * happens, on an active tracepoint. * * * - * TP_fast_assign( - * memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN); - * __entry->prev_pid = prev->pid; - * __entry->prev_prio = prev->prio; + * TP_fast_assign( + * memcpy(__entry->next_comm, next->comm, TASK_COMM_LEN); + * __entry->prev_pid = prev->pid; + * __entry->prev_prio = prev->prio; * memcpy(__entry->prev_comm, prev->comm, TASK_COMM_LEN); * __entry->next_pid = next->pid; - * __entry->next_prio = next->prio; + * __entry->next_prio = next->prio; * ) * * * diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index e39679a72a3b..bec69d3678c1 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -33,7 +33,7 @@ static struct trace_array *blk_tr; static int __read_mostly blk_tracer_enabled; /* Select an alternative, minimalistic output than the original one */ -#define TRACE_BLK_OPT_CLASSIC 0x1 +#define TRACE_BLK_OPT_CLASSIC 0x1 static struct tracer_opt blk_tracer_opts[] = { /* Default disable the minimalistic output */ @@ -564,7 +564,7 @@ EXPORT_SYMBOL_GPL(blk_trace_startstop); /** * blk_trace_ioctl: - handle the ioctls associated with tracing * @bdev: the block device - * @cmd: the ioctl cmd + * @cmd: the ioctl cmd * @arg: the argument data, if any * **/ @@ -1128,9 +1128,9 @@ static void blk_tracer_reset(struct trace_array *tr) static struct { const char *act[2]; - int (*print)(struct trace_seq *s, const struct trace_entry *ent); + int (*print)(struct trace_seq *s, const struct trace_entry *ent); } what2act[] __read_mostly = { - [__BLK_TA_QUEUE] = {{ "Q", "queue" }, blk_log_generic }, + [__BLK_TA_QUEUE] = {{ "Q", "queue" }, blk_log_generic }, [__BLK_TA_BACKMERGE] = {{ "M", "backmerge" }, blk_log_generic }, [__BLK_TA_FRONTMERGE] = {{ "F", "frontmerge" }, blk_log_generic }, [__BLK_TA_GETRQ] = {{ "G", "getrq" }, blk_log_generic }, @@ -1229,7 +1229,7 @@ static struct tracer blk_tracer __read_mostly = { }; static struct trace_event trace_blk_event = { - .type = TRACE_BLK, + .type = TRACE_BLK, .trace = blk_trace_event_print, .binary = blk_trace_event_print_binary, }; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index cc94f8642485..8c6a902db40a 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -799,7 +799,7 @@ tracing_generic_entry_update(struct trace_entry *entry, unsigned long flags, entry->preempt_count = pc & 0xff; entry->pid = (tsk) ? tsk->pid : 0; - entry->tgid = (tsk) ? tsk->tgid : 0; + entry->tgid = (tsk) ? tsk->tgid : 0; entry->flags = #ifdef CONFIG_TRACE_IRQFLAGS_SUPPORT (irqs_disabled_flags(flags) ? TRACE_FLAG_IRQS_OFF : 0) | diff --git a/kernel/trace/trace_branch.c b/kernel/trace/trace_branch.c index aaa0755268b9..ad8c22efff41 100644 --- a/kernel/trace/trace_branch.c +++ b/kernel/trace/trace_branch.c @@ -157,7 +157,7 @@ static enum print_line_t trace_branch_print(struct trace_iterator *iter, static struct trace_event trace_branch_event = { - .type = TRACE_BRANCH, + .type = TRACE_BRANCH, .trace = trace_branch_print, }; diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 6ee1de59f19d..ae2e323df0c7 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -5,23 +5,23 @@ * * static void ftrace_event_(proto) * { - * event_trace_printk(_RET_IP_, ": " ); + * event_trace_printk(_RET_IP_, ": " ); * } * * static int ftrace_reg_event_(void) * { - * int ret; + * int ret; * - * ret = register_trace_(ftrace_event_); - * if (!ret) - * pr_info("event trace: Could not activate trace point " - * "probe to "); - * return ret; + * ret = register_trace_(ftrace_event_); + * if (!ret) + * pr_info("event trace: Could not activate trace point " + * "probe to "); + * return ret; * } * * static void ftrace_unreg_event_(void) * { - * unregister_trace_(ftrace_event_); + * unregister_trace_(ftrace_event_); * } * * For those macros defined with TRACE_FORMAT: @@ -29,9 +29,9 @@ * static struct ftrace_event_call __used * __attribute__((__aligned__(4))) * __attribute__((section("_ftrace_events"))) event_ = { - * .name = "", - * .regfunc = ftrace_reg_event_, - * .unregfunc = ftrace_unreg_event_, + * .name = "", + * .regfunc = ftrace_reg_event_, + * .unregfunc = ftrace_unreg_event_, * } * * @@ -41,66 +41,66 @@ * * static void ftrace_raw_event_(proto) * { - * struct ring_buffer_event *event; - * struct ftrace_raw_ *entry; <-- defined in stage 1 - * unsigned long irq_flags; - * int pc; + * struct ring_buffer_event *event; + * struct ftrace_raw_ *entry; <-- defined in stage 1 + * unsigned long irq_flags; + * int pc; * - * local_save_flags(irq_flags); - * pc = preempt_count(); + * local_save_flags(irq_flags); + * pc = preempt_count(); * - * event = trace_current_buffer_lock_reserve(event_.id, - * sizeof(struct ftrace_raw_), - * irq_flags, pc); - * if (!event) - * return; - * entry = ring_buffer_event_data(event); + * event = trace_current_buffer_lock_reserve(event_.id, + * sizeof(struct ftrace_raw_), + * irq_flags, pc); + * if (!event) + * return; + * entry = ring_buffer_event_data(event); * - * ; <-- Here we assign the entries by the __field and + * ; <-- Here we assign the entries by the __field and * __array macros. * - * trace_current_buffer_unlock_commit(event, irq_flags, pc); + * trace_current_buffer_unlock_commit(event, irq_flags, pc); * } * * static int ftrace_raw_reg_event_(void) * { - * int ret; + * int ret; * - * ret = register_trace_(ftrace_raw_event_); - * if (!ret) - * pr_info("event trace: Could not activate trace point " - * "probe to "); - * return ret; + * ret = register_trace_(ftrace_raw_event_); + * if (!ret) + * pr_info("event trace: Could not activate trace point " + * "probe to "); + * return ret; * } * * static void ftrace_unreg_event_(void) * { - * unregister_trace_(ftrace_raw_event_); + * unregister_trace_(ftrace_raw_event_); * } * * static struct trace_event ftrace_event_type_ = { - * .trace = ftrace_raw_output_, <-- stage 2 + * .trace = ftrace_raw_output_, <-- stage 2 * }; * * static int ftrace_raw_init_event_(void) * { - * int id; + * int id; * - * id = register_ftrace_event(&ftrace_event_type_); - * if (!id) - * return -ENODEV; - * event_.id = id; - * return 0; + * id = register_ftrace_event(&ftrace_event_type_); + * if (!id) + * return -ENODEV; + * event_.id = id; + * return 0; * } * * static struct ftrace_event_call __used * __attribute__((__aligned__(4))) * __attribute__((section("_ftrace_events"))) event_ = { - * .name = "", + * .name = "", * .system = "", - * .raw_init = ftrace_raw_init_event_, - * .regfunc = ftrace_reg_event_, - * .unregfunc = ftrace_unreg_event_, + * .raw_init = ftrace_raw_init_event_, + * .regfunc = ftrace_reg_event_, + * .unregfunc = ftrace_unreg_event_, * .show_format = ftrace_format_, * } * @@ -138,7 +138,7 @@ _TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ - .name = #call, \ + .name = #call, \ .system = __stringify(TRACE_SYSTEM), \ .regfunc = ftrace_reg_event_##call, \ .unregfunc = ftrace_unreg_event_##call, \ @@ -163,7 +163,7 @@ static void ftrace_raw_event_##call(proto) \ pc = preempt_count(); \ \ event = trace_current_buffer_lock_reserve(event_##call.id, \ - sizeof(struct ftrace_raw_##call), \ + sizeof(struct ftrace_raw_##call), \ irq_flags, pc); \ if (!event) \ return; \ @@ -208,7 +208,7 @@ static int ftrace_raw_init_event_##call(void) \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ - .name = #call, \ + .name = #call, \ .system = __stringify(TRACE_SYSTEM), \ .raw_init = ftrace_raw_init_event_##call, \ .regfunc = ftrace_raw_reg_event_##call, \ diff --git a/kernel/trace/trace_export.c b/kernel/trace/trace_export.c index 23ae78430d58..4d9952d3df50 100644 --- a/kernel/trace/trace_export.c +++ b/kernel/trace/trace_export.c @@ -94,7 +94,7 @@ ftrace_format_##call(struct trace_seq *s) \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ - .name = #call, \ + .name = #call, \ .id = proto, \ .system = __stringify(TRACE_SYSTEM), \ .show_format = ftrace_format_##call, \ diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 453ebd3b636e..d1493b853e41 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -841,12 +841,12 @@ static void graph_trace_close(struct trace_iterator *iter) } static struct tracer graph_trace __read_mostly = { - .name = "function_graph", + .name = "function_graph", .open = graph_trace_open, .close = graph_trace_close, .wait_pipe = poll_wait_pipe, - .init = graph_trace_init, - .reset = graph_trace_reset, + .init = graph_trace_init, + .reset = graph_trace_reset, .print_line = print_graph_function, .print_header = print_graph_headers, .flags = &tracer_flags, diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index ef8fd661b217..491832af9ba1 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -565,7 +565,7 @@ static enum print_line_t trace_fn_bin(struct trace_iterator *iter, int flags) } static struct trace_event trace_fn_event = { - .type = TRACE_FN, + .type = TRACE_FN, .trace = trace_fn_trace, .raw = trace_fn_raw, .hex = trace_fn_hex, @@ -696,7 +696,7 @@ static enum print_line_t trace_ctxwake_bin(struct trace_iterator *iter, } static struct trace_event trace_ctx_event = { - .type = TRACE_CTX, + .type = TRACE_CTX, .trace = trace_ctx_print, .raw = trace_ctx_raw, .hex = trace_ctx_hex, @@ -704,7 +704,7 @@ static struct trace_event trace_ctx_event = { }; static struct trace_event trace_wake_event = { - .type = TRACE_WAKE, + .type = TRACE_WAKE, .trace = trace_wake_print, .raw = trace_wake_raw, .hex = trace_wake_hex, @@ -759,7 +759,7 @@ static enum print_line_t trace_special_bin(struct trace_iterator *iter, } static struct trace_event trace_special_event = { - .type = TRACE_SPECIAL, + .type = TRACE_SPECIAL, .trace = trace_special_print, .raw = trace_special_print, .hex = trace_special_hex, @@ -796,7 +796,7 @@ static enum print_line_t trace_stack_print(struct trace_iterator *iter, } static struct trace_event trace_stack_event = { - .type = TRACE_STACK, + .type = TRACE_STACK, .trace = trace_stack_print, .raw = trace_special_print, .hex = trace_special_hex, @@ -825,7 +825,7 @@ static enum print_line_t trace_user_stack_print(struct trace_iterator *iter, } static struct trace_event trace_user_stack_event = { - .type = TRACE_USER_STACK, + .type = TRACE_USER_STACK, .trace = trace_user_stack_print, .raw = trace_special_print, .hex = trace_special_hex, @@ -879,7 +879,7 @@ static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) static struct trace_event trace_print_event = { - .type = TRACE_PRINT, + .type = TRACE_PRINT, .trace = trace_print_print, .raw = trace_print_raw, }; diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index 4664990fe9c5..e542483df623 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -19,14 +19,14 @@ struct cpu_workqueue_stats { /* Useful to know if we print the cpu headers */ bool first_entry; int cpu; - pid_t pid; + pid_t pid; /* Can be inserted from interrupt or user context, need to be atomic */ - atomic_t inserted; + atomic_t inserted; /* * Don't need to be atomic, works are serialized in a single workqueue thread * on a single CPU. */ - unsigned int executed; + unsigned int executed; }; /* List of workqueue threads on one cpu */ From 6cc3c6e12bb039047974ad2e7e2d46d15a1b762f Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 10 Mar 2009 19:03:43 +0100 Subject: [PATCH 273/413] trace_clock: fix preemption bug Using the function_graph tracer in recent kernels generates a spew of preemption BUGs. Fix this by not requiring trace_clock_local() users to disable preemption themselves. Signed-off-by: Peter Zijlstra Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_clock.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_clock.c b/kernel/trace/trace_clock.c index 2d4953f93560..05b176abfd30 100644 --- a/kernel/trace/trace_clock.c +++ b/kernel/trace/trace_clock.c @@ -27,12 +27,19 @@ */ u64 notrace trace_clock_local(void) { + unsigned long flags; + u64 clock; + /* * sched_clock() is an architecture implemented, fast, scalable, * lockless clock. It is not guaranteed to be coherent across * CPUs, nor across CPU idle events. */ - return sched_clock(); + raw_local_irq_save(flags); + clock = sched_clock(); + raw_local_irq_restore(flags); + + return clock; } /* From 80370cb758e7ca2692cd9fb5e413d970b1f4b2b2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 10 Mar 2009 17:16:35 -0400 Subject: [PATCH 274/413] tracing: use raw spinlocks for trace_vprintk Impact: prevent locking up by lockdep tracer The lockdep tracer uses trace_vprintk and thus trace_vprintk can not call back into lockdep without locking up. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8c6a902db40a..4c97947650ae 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1176,7 +1176,8 @@ void trace_graph_return(struct ftrace_graph_ret *trace) */ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) { - static DEFINE_SPINLOCK(trace_buf_lock); + static raw_spinlock_t trace_buf_lock = + (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; static u32 trace_buf[TRACE_BUF_SIZE]; struct ring_buffer_event *event; @@ -1201,7 +1202,9 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) if (unlikely(atomic_read(&data->disabled))) goto out; - spin_lock_irqsave(&trace_buf_lock, flags); + /* Lockdep uses trace_printk for lock tracing */ + local_irq_save(flags); + __raw_spin_lock(&trace_buf_lock); len = vbin_printf(trace_buf, TRACE_BUF_SIZE, fmt, args); if (len > TRACE_BUF_SIZE || len < 0) @@ -1220,7 +1223,8 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) ring_buffer_unlock_commit(tr->buffer, event); out_unlock: - spin_unlock_irqrestore(&trace_buf_lock, flags); + __raw_spin_unlock(&trace_buf_lock); + local_irq_restore(flags); out: ftrace_preempt_enable(resched); From 73c5162aa362a543793f4a957c6c536dcbaa89ce Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 11 Mar 2009 13:42:01 -0400 Subject: [PATCH 275/413] tracing: keep ring buffer to minimum size till used Impact: less memory impact on systems not using tracer When the kernel boots up that has tracing configured, it allocates the default size of the ring buffer. This currently happens to be 1.4Megs per possible CPU. This is quite a bit of wasted memory if the system is never using the tracer. The current solution is to keep the ring buffers to a minimum size until the user uses them. Once a tracer is piped into the current_tracer the ring buffer will be expanded to the default size. If the user changes the size of the ring buffer, it will take the size given by the user immediately. If the user adds a "ftrace=" to the kernel command line, then the ring buffers will be set to the default size on initialization. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 79 +++++++++++++++++++++++++++++++++----------- 1 file changed, 59 insertions(+), 20 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 4c97947650ae..0c1dc1850858 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -44,6 +44,12 @@ unsigned long __read_mostly tracing_max_latency; unsigned long __read_mostly tracing_thresh; +/* + * On boot up, the ring buffer is set to the minimum size, so that + * we do not waste memory on systems that are not using tracing. + */ +static int ring_buffer_expanded; + /* * We need to change this state when a selftest is running. * A selftest will lurk into the ring-buffer to count the @@ -128,6 +134,8 @@ static int __init set_ftrace(char *str) { strncpy(bootup_tracer_buf, str, BOOTUP_TRACER_SIZE); default_bootup_tracer = bootup_tracer_buf; + /* We are using ftrace early, expand it */ + ring_buffer_expanded = 1; return 1; } __setup("ftrace=", set_ftrace); @@ -2315,6 +2323,40 @@ int tracer_init(struct tracer *t, struct trace_array *tr) return t->init(tr); } +static int tracing_resize_ring_buffer(unsigned long size) +{ + int ret; + + /* + * If kernel or user changes the size of the ring buffer + * it get completed. + */ + ring_buffer_expanded = 1; + + ret = ring_buffer_resize(global_trace.buffer, size); + if (ret < 0) + return ret; + + ret = ring_buffer_resize(max_tr.buffer, size); + if (ret < 0) { + int r; + + r = ring_buffer_resize(global_trace.buffer, + global_trace.entries); + if (r < 0) { + /* AARGH! We are left with different + * size max buffer!!!! */ + WARN_ON(1); + tracing_disabled = 1; + } + return ret; + } + + global_trace.entries = size; + + return ret; +} + struct trace_option_dentry; static struct trace_option_dentry * @@ -2330,6 +2372,13 @@ static int tracing_set_tracer(const char *buf) struct tracer *t; int ret = 0; + if (!ring_buffer_expanded) { + ret = tracing_resize_ring_buffer(trace_buf_size); + if (ret < 0) + return ret; + ret = 0; + } + mutex_lock(&trace_types_lock); for (t = trace_types; t; t = t->next) { if (strcmp(t->name, buf) == 0) @@ -2903,28 +2952,11 @@ tracing_entries_write(struct file *filp, const char __user *ubuf, val <<= 10; if (val != global_trace.entries) { - ret = ring_buffer_resize(global_trace.buffer, val); + ret = tracing_resize_ring_buffer(val); if (ret < 0) { cnt = ret; goto out; } - - ret = ring_buffer_resize(max_tr.buffer, val); - if (ret < 0) { - int r; - cnt = ret; - r = ring_buffer_resize(global_trace.buffer, - global_trace.entries); - if (r < 0) { - /* AARGH! We are left with different - * size max buffer!!!! */ - WARN_ON(1); - tracing_disabled = 1; - } - goto out; - } - - global_trace.entries = val; } filp->f_pos += cnt; @@ -3916,6 +3948,7 @@ void ftrace_dump(void) __init static int tracer_alloc_buffers(void) { struct trace_array_cpu *data; + int ring_buf_size; int i; int ret = -ENOMEM; @@ -3928,12 +3961,18 @@ __init static int tracer_alloc_buffers(void) if (!alloc_cpumask_var(&tracing_reader_cpumask, GFP_KERNEL)) goto out_free_tracing_cpumask; + /* To save memory, keep the ring buffer size to its minimum */ + if (ring_buffer_expanded) + ring_buf_size = trace_buf_size; + else + ring_buf_size = 1; + cpumask_copy(tracing_buffer_mask, cpu_possible_mask); cpumask_copy(tracing_cpumask, cpu_all_mask); cpumask_clear(tracing_reader_cpumask); /* TODO: make the number of buffers hot pluggable with CPUS */ - global_trace.buffer = ring_buffer_alloc(trace_buf_size, + global_trace.buffer = ring_buffer_alloc(ring_buf_size, TRACE_BUFFER_FLAGS); if (!global_trace.buffer) { printk(KERN_ERR "tracer: failed to allocate ring buffer!\n"); @@ -3944,7 +3983,7 @@ __init static int tracer_alloc_buffers(void) #ifdef CONFIG_TRACER_MAX_TRACE - max_tr.buffer = ring_buffer_alloc(trace_buf_size, + max_tr.buffer = ring_buffer_alloc(ring_buf_size, TRACE_BUFFER_FLAGS); if (!max_tr.buffer) { printk(KERN_ERR "tracer: failed to allocate max ring buffer!\n"); From 1852fcce181faa237c010a3dbedb473cf9d4555f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 11 Mar 2009 14:33:00 -0400 Subject: [PATCH 276/413] tracing: expand the ring buffers when an event is activated To save memory, the tracer ring buffers are set to a minimum. The activating of a trace expands the ring buffer size. This patch adds this expanding, when an event is activated. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 20 ++++++++++++++++++++ kernel/trace/trace.h | 3 +++ kernel/trace/trace_events.c | 8 ++++++++ 3 files changed, 31 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 0c1dc1850858..35ee63ae4122 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2357,6 +2357,26 @@ static int tracing_resize_ring_buffer(unsigned long size) return ret; } +/** + * tracing_update_buffers - used by tracing facility to expand ring buffers + * + * To save on memory when the tracing is never used on a system with it + * configured in. The ring buffers are set to a minimum size. But once + * a user starts to use the tracing facility, then they need to grow + * to their default size. + * + * This function is to be called when a tracer is about to be used. + */ +int tracing_update_buffers(void) +{ + int ret = 0; + + if (!ring_buffer_expanded) + ret = tracing_resize_ring_buffer(trace_buf_size); + + return ret; +} + struct trace_option_dentry; static struct trace_option_dentry * diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index c5e1d8865fe4..336324d717f8 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -737,6 +737,9 @@ static inline void trace_branch_disable(void) } #endif /* CONFIG_BRANCH_TRACER */ +/* set ring buffers to default size if not already done so */ +int tracing_update_buffers(void); + /* trace event type bit fields, not numeric */ enum { TRACE_EVENT_TYPE_PRINTF = 1, diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 769dfd00fc85..ca624df73591 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -141,6 +141,10 @@ ftrace_event_write(struct file *file, const char __user *ubuf, if (!cnt || cnt < 0) return 0; + ret = tracing_update_buffers(); + if (ret < 0) + return ret; + ret = get_user(ch, ubuf++); if (ret) return ret; @@ -331,6 +335,10 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, if (ret < 0) return ret; + ret = tracing_update_buffers(); + if (ret < 0) + return ret; + switch (val) { case 0: case 1: From 9aba60fe6eb20453de53a572143bef22fa929fba Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 11 Mar 2009 19:52:30 -0400 Subject: [PATCH 277/413] tracing: fix trace_wait to know to wait on all cpus or just one Impact: fix to task live locking on reading trace_pipe on one CPU The same code is used for both trace_pipe (all CPUS) and the per_cpu trace_pipe file. When there is no data to read, it will check for signals and wait on the trace wait queue. The problem happens with the per_cpu wait. The trace_wait code checks all CPUs. Thus, if there's data in another CPU buffer, then it will exit the wait, without checking for signals or waiting on the wait queue. It would then try to read the empty buffer, and since that will just return nothing, then it will try to wait again. Unfortunately, that will again fail due to there still being data in the other buffers. This ends up with a live lock for the task. This patch fixes the trace_wait to be aware that the iterator may only be waiting on a single buffer. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 35ee63ae4122..e60f4be10d64 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1666,6 +1666,19 @@ static int trace_empty(struct trace_iterator *iter) { int cpu; + /* If we are looking at one CPU buffer, only check that one */ + if (iter->cpu_file != TRACE_PIPE_ALL_CPU) { + cpu = iter->cpu_file; + if (iter->buffer_iter[cpu]) { + if (!ring_buffer_iter_empty(iter->buffer_iter[cpu])) + return 0; + } else { + if (!ring_buffer_empty_cpu(iter->tr->buffer, cpu)) + return 0; + } + return 1; + } + for_each_tracing_cpu(cpu) { if (iter->buffer_iter[cpu]) { if (!ring_buffer_iter_empty(iter->buffer_iter[cpu])) From 554f786e284a6ce859d51f62240d615603944c8e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 11 Mar 2009 22:00:13 -0400 Subject: [PATCH 278/413] ring-buffer: only allocate buffers for online cpus Impact: save on memory Currently, a ring buffer was allocated for each "possible_cpus". On some systems, this is the same as NR_CPUS. Thus, if a system defined NR_CPUS = 64 but it only had 1 CPU, we could have possibly 63 useless ring buffers taking up space. With a default buffer of 3 megs, this could be quite drastic. This patch changes the ring buffer code to only allocate ring buffers for online CPUs. If a CPU goes off line, we do not free the buffer. This is because the user may still have trace data in that buffer that they would like to look at. Perhaps in the future we could add code to delete a ring buffer if the CPU is offline and the ring buffer becomes empty. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 268 ++++++++++++++++++++++++++++++------- kernel/trace/trace.c | 6 - 2 files changed, 217 insertions(+), 57 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 178858492a89..d07c2888396f 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -16,6 +16,7 @@ #include #include #include +#include #include #include "trace.h" @@ -301,6 +302,10 @@ struct ring_buffer { struct mutex mutex; struct ring_buffer_per_cpu **buffers; + +#ifdef CONFIG_HOTPLUG + struct notifier_block cpu_notify; +#endif }; struct ring_buffer_iter { @@ -459,6 +464,11 @@ static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer) */ extern int ring_buffer_page_too_big(void); +#ifdef CONFIG_HOTPLUG +static int __cpuinit rb_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu); +#endif + /** * ring_buffer_alloc - allocate a new ring_buffer * @size: the size in bytes per cpu that is needed. @@ -496,7 +506,8 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) if (buffer->pages == 1) buffer->pages++; - cpumask_copy(buffer->cpumask, cpu_possible_mask); + get_online_cpus(); + cpumask_copy(buffer->cpumask, cpu_online_mask); buffer->cpus = nr_cpu_ids; bsize = sizeof(void *) * nr_cpu_ids; @@ -512,6 +523,13 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) goto fail_free_buffers; } +#ifdef CONFIG_HOTPLUG + buffer->cpu_notify.notifier_call = rb_cpu_notify; + buffer->cpu_notify.priority = 0; + register_cpu_notifier(&buffer->cpu_notify); +#endif + + put_online_cpus(); mutex_init(&buffer->mutex); return buffer; @@ -525,6 +543,7 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) fail_free_cpumask: free_cpumask_var(buffer->cpumask); + put_online_cpus(); fail_free_buffer: kfree(buffer); @@ -541,9 +560,17 @@ ring_buffer_free(struct ring_buffer *buffer) { int cpu; + get_online_cpus(); + +#ifdef CONFIG_HOTPLUG + unregister_cpu_notifier(&buffer->cpu_notify); +#endif + for_each_buffer_cpu(buffer, cpu) rb_free_cpu_buffer(buffer->buffers[cpu]); + put_online_cpus(); + free_cpumask_var(buffer->cpumask); kfree(buffer); @@ -649,16 +676,15 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) return size; mutex_lock(&buffer->mutex); + get_online_cpus(); nr_pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE); if (size < buffer_size) { /* easy case, just free pages */ - if (RB_WARN_ON(buffer, nr_pages >= buffer->pages)) { - mutex_unlock(&buffer->mutex); - return -1; - } + if (RB_WARN_ON(buffer, nr_pages >= buffer->pages)) + goto out_fail; rm_pages = buffer->pages - nr_pages; @@ -677,10 +703,8 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) * add these pages to the cpu_buffers. Otherwise we just free * them all and return -ENOMEM; */ - if (RB_WARN_ON(buffer, nr_pages <= buffer->pages)) { - mutex_unlock(&buffer->mutex); - return -1; - } + if (RB_WARN_ON(buffer, nr_pages <= buffer->pages)) + goto out_fail; new_pages = nr_pages - buffer->pages; @@ -705,13 +729,12 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) rb_insert_pages(cpu_buffer, &pages, new_pages); } - if (RB_WARN_ON(buffer, !list_empty(&pages))) { - mutex_unlock(&buffer->mutex); - return -1; - } + if (RB_WARN_ON(buffer, !list_empty(&pages))) + goto out_fail; out: buffer->pages = nr_pages; + put_online_cpus(); mutex_unlock(&buffer->mutex); return size; @@ -721,8 +744,18 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) list_del_init(&bpage->list); free_buffer_page(bpage); } + put_online_cpus(); mutex_unlock(&buffer->mutex); return -ENOMEM; + + /* + * Something went totally wrong, and we are too paranoid + * to even clean up the mess. + */ + out_fail: + put_online_cpus(); + mutex_unlock(&buffer->mutex); + return -1; } EXPORT_SYMBOL_GPL(ring_buffer_resize); @@ -1528,11 +1561,15 @@ void ring_buffer_record_disable_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + get_online_cpus(); + if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return; + goto out; cpu_buffer = buffer->buffers[cpu]; atomic_inc(&cpu_buffer->record_disabled); + out: + put_online_cpus(); } EXPORT_SYMBOL_GPL(ring_buffer_record_disable_cpu); @@ -1548,11 +1585,15 @@ void ring_buffer_record_enable_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + get_online_cpus(); + if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return; + goto out; cpu_buffer = buffer->buffers[cpu]; atomic_dec(&cpu_buffer->record_disabled); + out: + put_online_cpus(); } EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu); @@ -1564,12 +1605,19 @@ EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu); unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + unsigned long ret = 0; + + get_online_cpus(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return 0; + goto out; cpu_buffer = buffer->buffers[cpu]; - return cpu_buffer->entries; + ret = cpu_buffer->entries; + out: + put_online_cpus(); + + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu); @@ -1581,12 +1629,19 @@ EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu); unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + unsigned long ret = 0; + + get_online_cpus(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return 0; + goto out; cpu_buffer = buffer->buffers[cpu]; - return cpu_buffer->overrun; + ret = cpu_buffer->overrun; + out: + put_online_cpus(); + + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_overrun_cpu); @@ -1603,12 +1658,16 @@ unsigned long ring_buffer_entries(struct ring_buffer *buffer) unsigned long entries = 0; int cpu; + get_online_cpus(); + /* if you care about this being correct, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; entries += cpu_buffer->entries; } + put_online_cpus(); + return entries; } EXPORT_SYMBOL_GPL(ring_buffer_entries); @@ -1626,12 +1685,16 @@ unsigned long ring_buffer_overruns(struct ring_buffer *buffer) unsigned long overruns = 0; int cpu; + get_online_cpus(); + /* if you care about this being correct, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; overruns += cpu_buffer->overrun; } + put_online_cpus(); + return overruns; } EXPORT_SYMBOL_GPL(ring_buffer_overruns); @@ -1663,9 +1726,14 @@ static void rb_iter_reset(struct ring_buffer_iter *iter) */ void ring_buffer_iter_reset(struct ring_buffer_iter *iter) { - struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer; + struct ring_buffer_per_cpu *cpu_buffer; unsigned long flags; + if (!iter) + return; + + cpu_buffer = iter->cpu_buffer; + spin_lock_irqsave(&cpu_buffer->reader_lock, flags); rb_iter_reset(iter); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); @@ -1900,9 +1968,6 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) struct buffer_page *reader; int nr_loops = 0; - if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return NULL; - cpu_buffer = buffer->buffers[cpu]; again: @@ -2028,13 +2093,21 @@ struct ring_buffer_event * ring_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) { struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; - struct ring_buffer_event *event; + struct ring_buffer_event *event = NULL; unsigned long flags; + get_online_cpus(); + + if (!cpumask_test_cpu(cpu, buffer->cpumask)) + goto out; + spin_lock_irqsave(&cpu_buffer->reader_lock, flags); event = rb_buffer_peek(buffer, cpu, ts); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + out: + put_online_cpus(); + return event; } @@ -2071,24 +2144,31 @@ ring_buffer_iter_peek(struct ring_buffer_iter *iter, u64 *ts) struct ring_buffer_event * ring_buffer_consume(struct ring_buffer *buffer, int cpu, u64 *ts) { - struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; - struct ring_buffer_event *event; + struct ring_buffer_per_cpu *cpu_buffer; + struct ring_buffer_event *event = NULL; unsigned long flags; - if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return NULL; + /* might be called in atomic */ + preempt_disable(); + if (!cpumask_test_cpu(cpu, buffer->cpumask)) + goto out; + + cpu_buffer = buffer->buffers[cpu]; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); event = rb_buffer_peek(buffer, cpu, ts); if (!event) - goto out; + goto out_unlock; rb_advance_reader(cpu_buffer); - out: + out_unlock: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + out: + preempt_enable(); + return event; } EXPORT_SYMBOL_GPL(ring_buffer_consume); @@ -2109,15 +2189,17 @@ struct ring_buffer_iter * ring_buffer_read_start(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - struct ring_buffer_iter *iter; + struct ring_buffer_iter *iter = NULL; unsigned long flags; + get_online_cpus(); + if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return NULL; + goto out; iter = kmalloc(sizeof(*iter), GFP_KERNEL); if (!iter) - return NULL; + goto out; cpu_buffer = buffer->buffers[cpu]; @@ -2132,6 +2214,9 @@ ring_buffer_read_start(struct ring_buffer *buffer, int cpu) __raw_spin_unlock(&cpu_buffer->lock); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + out: + put_online_cpus(); + return iter; } EXPORT_SYMBOL_GPL(ring_buffer_read_start); @@ -2224,9 +2309,13 @@ void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; unsigned long flags; + int resched; + + /* Can't use get_online_cpus because this can be in atomic */ + resched = ftrace_preempt_disable(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return; + goto out; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); @@ -2237,6 +2326,8 @@ void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu) __raw_spin_unlock(&cpu_buffer->lock); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + out: + ftrace_preempt_enable(resched); } EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu); @@ -2246,10 +2337,16 @@ EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu); */ void ring_buffer_reset(struct ring_buffer *buffer) { + int resched; int cpu; + /* Can't use get_online_cpus because this can be in atomic */ + resched = ftrace_preempt_disable(); + for_each_buffer_cpu(buffer, cpu) ring_buffer_reset_cpu(buffer, cpu); + + ftrace_preempt_enable(resched); } EXPORT_SYMBOL_GPL(ring_buffer_reset); @@ -2262,12 +2359,17 @@ int ring_buffer_empty(struct ring_buffer *buffer) struct ring_buffer_per_cpu *cpu_buffer; int cpu; + get_online_cpus(); + /* yes this is racy, but if you don't like the race, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; if (!rb_per_cpu_empty(cpu_buffer)) return 0; } + + put_online_cpus(); + return 1; } EXPORT_SYMBOL_GPL(ring_buffer_empty); @@ -2280,12 +2382,20 @@ EXPORT_SYMBOL_GPL(ring_buffer_empty); int ring_buffer_empty_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; + int ret = 1; + + get_online_cpus(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) - return 1; + goto out; cpu_buffer = buffer->buffers[cpu]; - return rb_per_cpu_empty(cpu_buffer); + ret = rb_per_cpu_empty(cpu_buffer); + + out: + put_online_cpus(); + + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_empty_cpu); @@ -2304,32 +2414,37 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, { struct ring_buffer_per_cpu *cpu_buffer_a; struct ring_buffer_per_cpu *cpu_buffer_b; + int ret = -EINVAL; + + get_online_cpus(); if (!cpumask_test_cpu(cpu, buffer_a->cpumask) || !cpumask_test_cpu(cpu, buffer_b->cpumask)) - return -EINVAL; + goto out; /* At least make sure the two buffers are somewhat the same */ if (buffer_a->pages != buffer_b->pages) - return -EINVAL; + goto out; + + ret = -EAGAIN; if (ring_buffer_flags != RB_BUFFERS_ON) - return -EAGAIN; + goto out; if (atomic_read(&buffer_a->record_disabled)) - return -EAGAIN; + goto out; if (atomic_read(&buffer_b->record_disabled)) - return -EAGAIN; + goto out; cpu_buffer_a = buffer_a->buffers[cpu]; cpu_buffer_b = buffer_b->buffers[cpu]; if (atomic_read(&cpu_buffer_a->record_disabled)) - return -EAGAIN; + goto out; if (atomic_read(&cpu_buffer_b->record_disabled)) - return -EAGAIN; + goto out; /* * We can't do a synchronize_sched here because this @@ -2349,7 +2464,11 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, atomic_dec(&cpu_buffer_a->record_disabled); atomic_dec(&cpu_buffer_b->record_disabled); - return 0; + ret = 0; +out: + put_online_cpus(); + + return ret; } EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu); @@ -2464,27 +2583,32 @@ int ring_buffer_read_page(struct ring_buffer *buffer, u64 save_timestamp; int ret = -1; + get_online_cpus(); + + if (!cpumask_test_cpu(cpu, buffer->cpumask)) + goto out; + /* * If len is not big enough to hold the page header, then * we can not copy anything. */ if (len <= BUF_PAGE_HDR_SIZE) - return -1; + goto out; len -= BUF_PAGE_HDR_SIZE; if (!data_page) - return -1; + goto out; bpage = *data_page; if (!bpage) - return -1; + goto out; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); reader = rb_get_reader_page(cpu_buffer); if (!reader) - goto out; + goto out_unlock; event = rb_reader_event(cpu_buffer); @@ -2506,7 +2630,7 @@ int ring_buffer_read_page(struct ring_buffer *buffer, unsigned int size; if (full) - goto out; + goto out_unlock; if (len > (commit - read)) len = (commit - read); @@ -2514,7 +2638,7 @@ int ring_buffer_read_page(struct ring_buffer *buffer, size = rb_event_length(event); if (len < size) - goto out; + goto out_unlock; /* save the current timestamp, since the user will need it */ save_timestamp = cpu_buffer->read_stamp; @@ -2553,9 +2677,12 @@ int ring_buffer_read_page(struct ring_buffer *buffer, } ret = read; - out: + out_unlock: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + out: + put_online_cpus(); + return ret; } @@ -2629,3 +2756,42 @@ static __init int rb_init_debugfs(void) } fs_initcall(rb_init_debugfs); + +#ifdef CONFIG_HOTPLUG +static int __cpuinit rb_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) +{ + struct ring_buffer *buffer = + container_of(self, struct ring_buffer, cpu_notify); + long cpu = (long)hcpu; + + switch (action) { + case CPU_UP_PREPARE: + case CPU_UP_PREPARE_FROZEN: + if (cpu_isset(cpu, *buffer->cpumask)) + return NOTIFY_OK; + + buffer->buffers[cpu] = + rb_allocate_cpu_buffer(buffer, cpu); + if (!buffer->buffers[cpu]) { + WARN(1, "failed to allocate ring buffer on CPU %ld\n", + cpu); + return NOTIFY_OK; + } + smp_wmb(); + cpu_set(cpu, *buffer->cpumask); + break; + case CPU_DOWN_PREPARE: + case CPU_DOWN_PREPARE_FROZEN: + /* + * Do nothing. + * If we were to free the buffer, then the user would + * lose any trace that was in the buffer. + */ + break; + default: + break; + } + return NOTIFY_OK; +} +#endif diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e60f4be10d64..14c98f6a47bc 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1805,17 +1805,11 @@ __tracing_open(struct inode *inode, struct file *file) iter->buffer_iter[cpu] = ring_buffer_read_start(iter->tr->buffer, cpu); - - if (!iter->buffer_iter[cpu]) - goto fail_buffer; } } else { cpu = iter->cpu_file; iter->buffer_iter[cpu] = ring_buffer_read_start(iter->tr->buffer, cpu); - - if (!iter->buffer_iter[cpu]) - goto fail; } /* TODO stop tracer */ From a123c52b46a1f84bcec3dc963351896c6d6afaf7 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 11:21:08 -0400 Subject: [PATCH 279/413] tracing: fix comments about trace buffer resizing Impact: cleanup Some of the comments about the trace buffer resizing is gobbledygook. And I wonder why people question if I'm a native English speaker. This patch makes the comments make a bit more sense. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c3946a6df34e..c61ee85c50bb 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2336,7 +2336,8 @@ static int tracing_resize_ring_buffer(unsigned long size) /* * If kernel or user changes the size of the ring buffer - * it get completed. + * we use the size that was given, and we can forget about + * expanding it later. */ ring_buffer_expanded = 1; @@ -2351,8 +2352,20 @@ static int tracing_resize_ring_buffer(unsigned long size) r = ring_buffer_resize(global_trace.buffer, global_trace.entries); if (r < 0) { - /* AARGH! We are left with different - * size max buffer!!!! */ + /* + * AARGH! We are left with different + * size max buffer!!!! + * The max buffer is our "snapshot" buffer. + * When a tracer needs a snapshot (one of the + * latency tracers), it swaps the max buffer + * with the saved snap shot. We succeeded to + * update the size of the main buffer, but failed to + * update the size of the max buffer. But when we tried + * to reset the main buffer to the original size, we + * failed there too. This is very unlikely to + * happen, but if it does, warn and kill all + * tracing. + */ WARN_ON(1); tracing_disabled = 1; } From 1027fcb206a0fb8348e63aff078c74bdee1c2698 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 11:33:20 -0400 Subject: [PATCH 280/413] tracing: protect ring_buffer_expanded with trace_types_lock Impact: prevent races with ring_buffer_expanded This patch places the expanding of the tracing buffer under the protection of the trace_types_lock mutex. It is highly unlikely that there would be any contention, but better safe than sorry. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c61ee85c50bb..04ab8243a13d 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2391,8 +2391,10 @@ int tracing_update_buffers(void) { int ret = 0; + mutex_lock(&trace_types_lock); if (!ring_buffer_expanded) ret = tracing_resize_ring_buffer(trace_buf_size); + mutex_unlock(&trace_types_lock); return ret; } @@ -2412,6 +2414,8 @@ static int tracing_set_tracer(const char *buf) struct tracer *t; int ret = 0; + mutex_lock(&trace_types_lock); + if (!ring_buffer_expanded) { ret = tracing_resize_ring_buffer(trace_buf_size); if (ret < 0) @@ -2419,7 +2423,6 @@ static int tracing_set_tracer(const char *buf) ret = 0; } - mutex_lock(&trace_types_lock); for (t = trace_types; t; t = t->next) { if (strcmp(t->name, buf) == 0) break; From 59222efe2d184956464abe5b637bc842ff053b93 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 11:46:03 -0400 Subject: [PATCH 281/413] ring-buffer: use CONFIG_HOTPLUG_CPU not CONFIG_HOTPLUG The hotplug code in the ring buffers is for use with CPU hotplug, not generic hotplug. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index d07c2888396f..035b56c3a6c9 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -303,7 +303,7 @@ struct ring_buffer { struct ring_buffer_per_cpu **buffers; -#ifdef CONFIG_HOTPLUG +#ifdef CONFIG_HOTPLUG_CPU struct notifier_block cpu_notify; #endif }; @@ -464,7 +464,7 @@ static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer) */ extern int ring_buffer_page_too_big(void); -#ifdef CONFIG_HOTPLUG +#ifdef CONFIG_HOTPLUG_CPU static int __cpuinit rb_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu); #endif @@ -523,7 +523,7 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) goto fail_free_buffers; } -#ifdef CONFIG_HOTPLUG +#ifdef CONFIG_HOTPLUG_CPU buffer->cpu_notify.notifier_call = rb_cpu_notify; buffer->cpu_notify.priority = 0; register_cpu_notifier(&buffer->cpu_notify); @@ -562,7 +562,7 @@ ring_buffer_free(struct ring_buffer *buffer) get_online_cpus(); -#ifdef CONFIG_HOTPLUG +#ifdef CONFIG_HOTPLUG_CPU unregister_cpu_notifier(&buffer->cpu_notify); #endif @@ -2757,7 +2757,7 @@ static __init int rb_init_debugfs(void) fs_initcall(rb_init_debugfs); -#ifdef CONFIG_HOTPLUG +#ifdef CONFIG_HOTPLUG_CPU static int __cpuinit rb_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) { From 8aabee573dff131a085c63de7667eacd94ba4ccb Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 13:13:49 -0400 Subject: [PATCH 282/413] ring-buffer: remove unneeded get_online_cpus Impact: speed up and remove possible races The get_online_cpus was added to the ring buffer because the original design would free the ring buffer on a CPU that was being taken off line. The final design kept the ring buffer around even when the CPU was taken off line. This is to allow a user to still read the information on that ring buffer. Most of the get_online_cpus are no longer needed since the ring buffer will not disappear from the use cases. Reported-by: KOSAKI Motohiro Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 90 ++++++-------------------------------- 1 file changed, 14 insertions(+), 76 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 035b56c3a6c9..2c36be9fac2e 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1561,15 +1561,11 @@ void ring_buffer_record_disable_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - get_online_cpus(); - if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return; cpu_buffer = buffer->buffers[cpu]; atomic_inc(&cpu_buffer->record_disabled); - out: - put_online_cpus(); } EXPORT_SYMBOL_GPL(ring_buffer_record_disable_cpu); @@ -1585,15 +1581,11 @@ void ring_buffer_record_enable_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - get_online_cpus(); - if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return; cpu_buffer = buffer->buffers[cpu]; atomic_dec(&cpu_buffer->record_disabled); - out: - put_online_cpus(); } EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu); @@ -1605,17 +1597,13 @@ EXPORT_SYMBOL_GPL(ring_buffer_record_enable_cpu); unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - unsigned long ret = 0; - - get_online_cpus(); + unsigned long ret; if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return 0; cpu_buffer = buffer->buffers[cpu]; ret = cpu_buffer->entries; - out: - put_online_cpus(); return ret; } @@ -1629,17 +1617,13 @@ EXPORT_SYMBOL_GPL(ring_buffer_entries_cpu); unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - unsigned long ret = 0; - - get_online_cpus(); + unsigned long ret; if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return 0; cpu_buffer = buffer->buffers[cpu]; ret = cpu_buffer->overrun; - out: - put_online_cpus(); return ret; } @@ -1658,16 +1642,12 @@ unsigned long ring_buffer_entries(struct ring_buffer *buffer) unsigned long entries = 0; int cpu; - get_online_cpus(); - /* if you care about this being correct, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; entries += cpu_buffer->entries; } - put_online_cpus(); - return entries; } EXPORT_SYMBOL_GPL(ring_buffer_entries); @@ -1685,16 +1665,12 @@ unsigned long ring_buffer_overruns(struct ring_buffer *buffer) unsigned long overruns = 0; int cpu; - get_online_cpus(); - /* if you care about this being correct, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; overruns += cpu_buffer->overrun; } - put_online_cpus(); - return overruns; } EXPORT_SYMBOL_GPL(ring_buffer_overruns); @@ -2093,21 +2069,16 @@ struct ring_buffer_event * ring_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) { struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; - struct ring_buffer_event *event = NULL; + struct ring_buffer_event *event; unsigned long flags; - get_online_cpus(); - if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return NULL; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); event = rb_buffer_peek(buffer, cpu, ts); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); - out: - put_online_cpus(); - return event; } @@ -2189,17 +2160,15 @@ struct ring_buffer_iter * ring_buffer_read_start(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - struct ring_buffer_iter *iter = NULL; + struct ring_buffer_iter *iter; unsigned long flags; - get_online_cpus(); - if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return NULL; iter = kmalloc(sizeof(*iter), GFP_KERNEL); if (!iter) - goto out; + return NULL; cpu_buffer = buffer->buffers[cpu]; @@ -2214,9 +2183,6 @@ ring_buffer_read_start(struct ring_buffer *buffer, int cpu) __raw_spin_unlock(&cpu_buffer->lock); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); - out: - put_online_cpus(); - return iter; } EXPORT_SYMBOL_GPL(ring_buffer_read_start); @@ -2309,13 +2275,9 @@ void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer = buffer->buffers[cpu]; unsigned long flags; - int resched; - - /* Can't use get_online_cpus because this can be in atomic */ - resched = ftrace_preempt_disable(); if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return; spin_lock_irqsave(&cpu_buffer->reader_lock, flags); @@ -2326,8 +2288,6 @@ void ring_buffer_reset_cpu(struct ring_buffer *buffer, int cpu) __raw_spin_unlock(&cpu_buffer->lock); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); - out: - ftrace_preempt_enable(resched); } EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu); @@ -2337,16 +2297,10 @@ EXPORT_SYMBOL_GPL(ring_buffer_reset_cpu); */ void ring_buffer_reset(struct ring_buffer *buffer) { - int resched; int cpu; - /* Can't use get_online_cpus because this can be in atomic */ - resched = ftrace_preempt_disable(); - for_each_buffer_cpu(buffer, cpu) ring_buffer_reset_cpu(buffer, cpu); - - ftrace_preempt_enable(resched); } EXPORT_SYMBOL_GPL(ring_buffer_reset); @@ -2359,8 +2313,6 @@ int ring_buffer_empty(struct ring_buffer *buffer) struct ring_buffer_per_cpu *cpu_buffer; int cpu; - get_online_cpus(); - /* yes this is racy, but if you don't like the race, lock the buffer */ for_each_buffer_cpu(buffer, cpu) { cpu_buffer = buffer->buffers[cpu]; @@ -2368,8 +2320,6 @@ int ring_buffer_empty(struct ring_buffer *buffer) return 0; } - put_online_cpus(); - return 1; } EXPORT_SYMBOL_GPL(ring_buffer_empty); @@ -2382,18 +2332,14 @@ EXPORT_SYMBOL_GPL(ring_buffer_empty); int ring_buffer_empty_cpu(struct ring_buffer *buffer, int cpu) { struct ring_buffer_per_cpu *cpu_buffer; - int ret = 1; - - get_online_cpus(); + int ret; if (!cpumask_test_cpu(cpu, buffer->cpumask)) - goto out; + return 1; cpu_buffer = buffer->buffers[cpu]; ret = rb_per_cpu_empty(cpu_buffer); - out: - put_online_cpus(); return ret; } @@ -2416,8 +2362,6 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, struct ring_buffer_per_cpu *cpu_buffer_b; int ret = -EINVAL; - get_online_cpus(); - if (!cpumask_test_cpu(cpu, buffer_a->cpumask) || !cpumask_test_cpu(cpu, buffer_b->cpumask)) goto out; @@ -2466,8 +2410,6 @@ int ring_buffer_swap_cpu(struct ring_buffer *buffer_a, ret = 0; out: - put_online_cpus(); - return ret; } EXPORT_SYMBOL_GPL(ring_buffer_swap_cpu); @@ -2583,8 +2525,6 @@ int ring_buffer_read_page(struct ring_buffer *buffer, u64 save_timestamp; int ret = -1; - get_online_cpus(); - if (!cpumask_test_cpu(cpu, buffer->cpumask)) goto out; @@ -2681,8 +2621,6 @@ int ring_buffer_read_page(struct ring_buffer *buffer, spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); out: - put_online_cpus(); - return ret; } From db526ca329f855510e8ce672332eba3304aed590 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 13:53:25 -0400 Subject: [PATCH 283/413] tracing: show that buffer size is not expanded Impact: do not confuse user on small trace buffer sizes When the system boots up, the trace buffer is small to conserve memory. It is only two pages per online CPU. When the tracer is used, it expands to the default value. This can confuse the user if they look at the buffer size and see only 7, but then later they see 1408. # cat /debug/tracing/buffer_size_kb 7 # echo sched_switch > /debug/tracing/current_tracer # cat /debug/tracing/buffer_size_kb 1408 This patch tries to help remove this confustion by showing that the buffer has not been expanded. # cat /debug/tracing/buffer_size_kb 7 (expanded: 1408) Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 04ab8243a13d..62a63b2b33dd 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2948,10 +2948,18 @@ tracing_entries_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) { struct trace_array *tr = filp->private_data; - char buf[64]; + char buf[96]; int r; - r = sprintf(buf, "%lu\n", tr->entries >> 10); + mutex_lock(&trace_types_lock); + if (!ring_buffer_expanded) + r = sprintf(buf, "%lu (expanded: %lu)\n", + tr->entries >> 10, + trace_buf_size >> 10); + else + r = sprintf(buf, "%lu\n", tr->entries >> 10); + mutex_unlock(&trace_types_lock); + return simple_read_from_buffer(ubuf, cnt, ppos, buf, r); } From 48ead02030f849d011259244bb4ea9b985479006 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 12 Mar 2009 18:24:49 +0100 Subject: [PATCH 284/413] tracing/core: bring back raw trace_printk for dynamic formats strings Impact: fix callsites with dynamic format strings Since its new binary implementation, trace_printk() internally uses static containers for the format strings on each callsites. But the value is assigned once at build time, which means that it can't take dynamic formats. So this patch unearthes the raw trace_printk implementation for the callers that will need trace_printk to be able to carry these dynamic format strings. The trace_printk() macro will use the appropriate implementation for each callsite. Most of the time however, the binary implementation will still be used. The other impact of this patch is that mmiotrace_printk() will use the old implementation because it calls the low level trace_vprintk and we can't guess here whether the format passed in it is dynamic or not. Some parts of this patch have been written by Steven Rostedt (most notably the part that chooses the appropriate implementation for each callsites). Signed-off-by: Frederic Weisbecker Signed-off-by: Steven Rostedt --- include/linux/kernel.h | 40 ++++++++----- kernel/trace/trace.c | 85 ++++++++++++++++++++++++++-- kernel/trace/trace.h | 13 ++++- kernel/trace/trace_event_types.h | 11 +++- kernel/trace/trace_functions_graph.c | 6 +- kernel/trace/trace_mmiotrace.c | 7 +-- kernel/trace/trace_output.c | 57 +++++++++++++++++-- kernel/trace/trace_printk.c | 35 ++++++++++-- 8 files changed, 214 insertions(+), 40 deletions(-) diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 7742798c9208..1daca3b062bb 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h @@ -452,31 +452,45 @@ do { \ #define trace_printk(fmt, args...) \ do { \ - static const char *trace_printk_fmt \ - __attribute__((section("__trace_printk_fmt"))); \ - \ - if (!trace_printk_fmt) \ - trace_printk_fmt = fmt; \ - \ __trace_printk_check_format(fmt, ##args); \ - __trace_printk(_THIS_IP_, trace_printk_fmt, ##args); \ + if (__builtin_constant_p(fmt)) { \ + static const char *trace_printk_fmt \ + __attribute__((section("__trace_printk_fmt"))) = \ + __builtin_constant_p(fmt) ? fmt : NULL; \ + \ + __trace_bprintk(_THIS_IP_, trace_printk_fmt, ##args); \ + } else \ + __trace_printk(_THIS_IP_, fmt, ##args); \ } while (0) +extern int +__trace_bprintk(unsigned long ip, const char *fmt, ...) + __attribute__ ((format (printf, 2, 3))); + extern int __trace_printk(unsigned long ip, const char *fmt, ...) __attribute__ ((format (printf, 2, 3))); +/* + * The double __builtin_constant_p is because gcc will give us an error + * if we try to allocate the static variable to fmt if it is not a + * constant. Even with the outer if statement. + */ #define ftrace_vprintk(fmt, vargs) \ do { \ - static const char *trace_printk_fmt \ - __attribute__((section("__trace_printk_fmt"))); \ + if (__builtin_constant_p(fmt)) { \ + static const char *trace_printk_fmt \ + __attribute__((section("__trace_printk_fmt"))) = \ + __builtin_constant_p(fmt) ? fmt : NULL; \ \ - if (!trace_printk_fmt) \ - trace_printk_fmt = fmt; \ - \ - __ftrace_vprintk(_THIS_IP_, trace_printk_fmt, vargs); \ + __ftrace_vbprintk(_THIS_IP_, trace_printk_fmt, vargs); \ + } else \ + __ftrace_vprintk(_THIS_IP_, fmt, vargs); \ } while (0) +extern int +__ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap); + extern int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap); diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 62a63b2b33dd..dbb077d8a172 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1179,10 +1179,10 @@ void trace_graph_return(struct ftrace_graph_ret *trace) /** - * trace_vprintk - write binary msg to tracing buffer + * trace_vbprintk - write binary msg to tracing buffer * */ -int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) +int trace_vbprintk(unsigned long ip, int depth, const char *fmt, va_list args) { static raw_spinlock_t trace_buf_lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; @@ -1191,7 +1191,7 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) struct ring_buffer_event *event; struct trace_array *tr = &global_trace; struct trace_array_cpu *data; - struct print_entry *entry; + struct bprint_entry *entry; unsigned long flags; int resched; int cpu, len = 0, size, pc; @@ -1219,7 +1219,7 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) goto out_unlock; size = sizeof(*entry) + sizeof(u32) * len; - event = trace_buffer_lock_reserve(tr, TRACE_PRINT, size, flags, pc); + event = trace_buffer_lock_reserve(tr, TRACE_BPRINT, size, flags, pc); if (!event) goto out_unlock; entry = ring_buffer_event_data(event); @@ -1240,6 +1240,60 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) return len; } +EXPORT_SYMBOL_GPL(trace_vbprintk); + +int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) +{ + static raw_spinlock_t trace_buf_lock = __RAW_SPIN_LOCK_UNLOCKED; + static char trace_buf[TRACE_BUF_SIZE]; + + struct ring_buffer_event *event; + struct trace_array *tr = &global_trace; + struct trace_array_cpu *data; + int cpu, len = 0, size, pc; + struct print_entry *entry; + unsigned long irq_flags; + + if (tracing_disabled || tracing_selftest_running) + return 0; + + pc = preempt_count(); + preempt_disable_notrace(); + cpu = raw_smp_processor_id(); + data = tr->data[cpu]; + + if (unlikely(atomic_read(&data->disabled))) + goto out; + + pause_graph_tracing(); + raw_local_irq_save(irq_flags); + __raw_spin_lock(&trace_buf_lock); + len = vsnprintf(trace_buf, TRACE_BUF_SIZE, fmt, args); + + len = min(len, TRACE_BUF_SIZE-1); + trace_buf[len] = 0; + + size = sizeof(*entry) + len + 1; + event = trace_buffer_lock_reserve(tr, TRACE_PRINT, size, irq_flags, pc); + if (!event) + goto out_unlock; + entry = ring_buffer_event_data(event); + entry->ip = ip; + entry->depth = depth; + + memcpy(&entry->buf, trace_buf, len); + entry->buf[len] = 0; + ring_buffer_unlock_commit(tr->buffer, event); + + out_unlock: + __raw_spin_unlock(&trace_buf_lock); + raw_local_irq_restore(irq_flags); + unpause_graph_tracing(); + out: + preempt_enable_notrace(); + + return len; +} EXPORT_SYMBOL_GPL(trace_vprintk); enum trace_file_type { @@ -1628,6 +1682,22 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) return TRACE_TYPE_HANDLED; } +static enum print_line_t print_bprintk_msg_only(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + struct bprint_entry *field; + int ret; + + trace_assign_type(field, entry); + + ret = trace_seq_bprintf(s, field->fmt, field->buf); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + static enum print_line_t print_printk_msg_only(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; @@ -1637,7 +1707,7 @@ static enum print_line_t print_printk_msg_only(struct trace_iterator *iter) trace_assign_type(field, entry); - ret = trace_seq_bprintf(s, field->fmt, field->buf); + ret = trace_seq_printf(s, "%s", field->buf); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -1702,6 +1772,11 @@ static enum print_line_t print_trace_line(struct trace_iterator *iter) return ret; } + if (iter->ent->type == TRACE_BPRINT && + trace_flags & TRACE_ITER_PRINTK && + trace_flags & TRACE_ITER_PRINTK_MSGONLY) + return print_bprintk_msg_only(iter); + if (iter->ent->type == TRACE_PRINT && trace_flags & TRACE_ITER_PRINTK && trace_flags & TRACE_ITER_PRINTK_MSGONLY) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 336324d717f8..cede1ab49d07 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -20,6 +20,7 @@ enum trace_type { TRACE_WAKE, TRACE_STACK, TRACE_PRINT, + TRACE_BPRINT, TRACE_SPECIAL, TRACE_MMIO_RW, TRACE_MMIO_MAP, @@ -117,7 +118,7 @@ struct userstack_entry { /* * trace_printk entry: */ -struct print_entry { +struct bprint_entry { struct trace_entry ent; unsigned long ip; int depth; @@ -125,6 +126,13 @@ struct print_entry { u32 buf[]; }; +struct print_entry { + struct trace_entry ent; + unsigned long ip; + int depth; + char buf[]; +}; + #define TRACE_OLD_SIZE 88 struct trace_field_cont { @@ -286,6 +294,7 @@ extern void __ftrace_bad_type(void); IF_ASSIGN(var, ent, struct stack_entry, TRACE_STACK); \ IF_ASSIGN(var, ent, struct userstack_entry, TRACE_USER_STACK);\ IF_ASSIGN(var, ent, struct print_entry, TRACE_PRINT); \ + IF_ASSIGN(var, ent, struct bprint_entry, TRACE_BPRINT); \ IF_ASSIGN(var, ent, struct special_entry, 0); \ IF_ASSIGN(var, ent, struct trace_mmiotrace_rw, \ TRACE_MMIO_RW); \ @@ -570,6 +579,8 @@ extern int trace_selftest_startup_branch(struct tracer *trace, extern void *head_page(struct trace_array_cpu *data); extern long ns2usecs(cycle_t nsec); extern int +trace_vbprintk(unsigned long ip, int depth, const char *fmt, va_list args); +extern int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args); extern unsigned long trace_flags; diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h index 5cca4c978bde..d0907d746425 100644 --- a/kernel/trace/trace_event_types.h +++ b/kernel/trace/trace_event_types.h @@ -102,7 +102,7 @@ TRACE_EVENT_FORMAT(user_stack, TRACE_USER_STACK, userstack_entry, ignore, "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") ); -TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, +TRACE_EVENT_FORMAT(bprint, TRACE_PRINT, bprint_entry, ignore, TRACE_STRUCT( TRACE_FIELD(unsigned long, ip, ip) TRACE_FIELD(unsigned int, depth, depth) @@ -112,6 +112,15 @@ TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, TP_RAW_FMT("%08lx (%d) fmt:%p %s") ); +TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, + TRACE_STRUCT( + TRACE_FIELD(unsigned long, ip, ip) + TRACE_FIELD(unsigned int, depth, depth) + TRACE_FIELD_ZERO_CHAR(buf) + ), + TP_RAW_FMT("%08lx (%d) fmt:%p %s") +); + TRACE_EVENT_FORMAT(branch, TRACE_BRANCH, trace_branch, ignore, TRACE_STRUCT( TRACE_FIELD(unsigned int, line, line) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 8566c14b3e9a..4c388607ed67 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -684,7 +684,7 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, } static enum print_line_t -print_graph_comment(struct print_entry *trace, struct trace_seq *s, +print_graph_comment(struct bprint_entry *trace, struct trace_seq *s, struct trace_entry *ent, struct trace_iterator *iter) { int i; @@ -781,8 +781,8 @@ print_graph_function(struct trace_iterator *iter) trace_assign_type(field, entry); return print_graph_return(&field->ret, s, entry, iter); } - case TRACE_PRINT: { - struct print_entry *field; + case TRACE_BPRINT: { + struct bprint_entry *field; trace_assign_type(field, entry); return print_graph_comment(field, s, entry, iter); } diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index 23e346a734ca..f095916e477f 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -254,6 +254,7 @@ static enum print_line_t mmio_print_mark(struct trace_iterator *iter) { struct trace_entry *entry = iter->ent; struct print_entry *print = (struct print_entry *)entry; + const char *msg = print->buf; struct trace_seq *s = &iter->seq; unsigned long long t = ns2usecs(iter->ts); unsigned long usec_rem = do_div(t, USEC_PER_SEC); @@ -261,11 +262,7 @@ static enum print_line_t mmio_print_mark(struct trace_iterator *iter) int ret; /* The trailing newline must be in the message. */ - ret = trace_seq_printf(s, "MARK %u.%06lu ", secs, usec_rem); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - ret = trace_seq_bprintf(s, print->fmt, print->buf); + ret = trace_seq_printf(s, "MARK %u.%06lu %s", secs, usec_rem, msg); if (!ret) return TRACE_TYPE_PARTIAL_LINE; diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 491832af9ba1..ea9d3b410c7a 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -832,13 +832,13 @@ static struct trace_event trace_user_stack_event = { .binary = trace_special_bin, }; -/* TRACE_PRINT */ +/* TRACE_BPRINT */ static enum print_line_t -trace_print_print(struct trace_iterator *iter, int flags) +trace_bprint_print(struct trace_iterator *iter, int flags) { struct trace_entry *entry = iter->ent; struct trace_seq *s = &iter->seq; - struct print_entry *field; + struct bprint_entry *field; trace_assign_type(field, entry); @@ -858,9 +858,10 @@ trace_print_print(struct trace_iterator *iter, int flags) } -static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) +static enum print_line_t +trace_bprint_raw(struct trace_iterator *iter, int flags) { - struct print_entry *field; + struct bprint_entry *field; struct trace_seq *s = &iter->seq; trace_assign_type(field, iter->ent); @@ -878,12 +879,55 @@ static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) } +static struct trace_event trace_bprint_event = { + .type = TRACE_BPRINT, + .trace = trace_bprint_print, + .raw = trace_bprint_raw, +}; + +/* TRACE_PRINT */ +static enum print_line_t trace_print_print(struct trace_iterator *iter, + int flags) +{ + struct print_entry *field; + struct trace_seq *s = &iter->seq; + + trace_assign_type(field, iter->ent); + + if (!seq_print_ip_sym(s, field->ip, flags)) + goto partial; + + if (!trace_seq_printf(s, ": %s", field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + +static enum print_line_t trace_print_raw(struct trace_iterator *iter, int flags) +{ + struct print_entry *field; + + trace_assign_type(field, iter->ent); + + if (!trace_seq_printf(&iter->seq, "# %lx %s", field->ip, field->buf)) + goto partial; + + return TRACE_TYPE_HANDLED; + + partial: + return TRACE_TYPE_PARTIAL_LINE; +} + static struct trace_event trace_print_event = { - .type = TRACE_PRINT, + .type = TRACE_PRINT, .trace = trace_print_print, .raw = trace_print_raw, }; + static struct trace_event *events[] __initdata = { &trace_fn_event, &trace_ctx_event, @@ -891,6 +935,7 @@ static struct trace_event *events[] __initdata = { &trace_special_event, &trace_stack_event, &trace_user_stack_event, + &trace_bprint_event, &trace_print_event, NULL }; diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c index a50aea22e929..f307a11e2332 100644 --- a/kernel/trace/trace_printk.c +++ b/kernel/trace/trace_printk.c @@ -99,7 +99,7 @@ struct notifier_block module_trace_bprintk_format_nb = { .notifier_call = module_trace_bprintk_format_notify, }; -int __trace_printk(unsigned long ip, const char *fmt, ...) +int __trace_bprintk(unsigned long ip, const char *fmt, ...) { int ret; va_list ap; @@ -107,6 +107,33 @@ int __trace_printk(unsigned long ip, const char *fmt, ...) if (unlikely(!fmt)) return 0; + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + va_start(ap, fmt); + ret = trace_vbprintk(ip, task_curr_ret_stack(current), fmt, ap); + va_end(ap); + return ret; +} +EXPORT_SYMBOL_GPL(__trace_bprintk); + +int __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap) + { + if (unlikely(!fmt)) + return 0; + + if (!(trace_flags & TRACE_ITER_PRINTK)) + return 0; + + return trace_vbprintk(ip, task_curr_ret_stack(current), fmt, ap); +} +EXPORT_SYMBOL_GPL(__ftrace_vbprintk); + +int __trace_printk(unsigned long ip, const char *fmt, ...) +{ + int ret; + va_list ap; + if (!(trace_flags & TRACE_ITER_PRINTK)) return 0; @@ -118,10 +145,7 @@ int __trace_printk(unsigned long ip, const char *fmt, ...) EXPORT_SYMBOL_GPL(__trace_printk); int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) - { - if (unlikely(!fmt)) - return 0; - +{ if (!(trace_flags & TRACE_ITER_PRINTK)) return 0; @@ -129,7 +153,6 @@ int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) } EXPORT_SYMBOL_GPL(__ftrace_vprintk); - static __init int init_trace_printk(void) { return register_module_notifier(&module_trace_bprintk_format_nb); From 828275574e0161bdddb5817d4bd76a0265ef0470 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 14:14:31 -0400 Subject: [PATCH 285/413] tracing: make bprint event use the proper event id The bprint record is using TRACE_PRINT when it should be TRACE_BPRINT. Signed-off-by: Steven Rostedt --- kernel/trace/trace_event_types.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h index d0907d746425..019915063fe6 100644 --- a/kernel/trace/trace_event_types.h +++ b/kernel/trace/trace_event_types.h @@ -102,7 +102,7 @@ TRACE_EVENT_FORMAT(user_stack, TRACE_USER_STACK, userstack_entry, ignore, "\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n\t=> (%08lx)\n") ); -TRACE_EVENT_FORMAT(bprint, TRACE_PRINT, bprint_entry, ignore, +TRACE_EVENT_FORMAT(bprint, TRACE_BPRINT, bprint_entry, ignore, TRACE_STRUCT( TRACE_FIELD(unsigned long, ip, ip) TRACE_FIELD(unsigned int, depth, depth) From e9fb2b6d5845e24f104713591286b6f39761c027 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 14:19:25 -0400 Subject: [PATCH 286/413] tracing: have event_trace_printk use static tracer Impact: speed up on event tracing The event_trace_printk is currently a wrapper function that calls trace_vprintk. Because it uses a variable for the fmt it misses out on the optimization of using the binary printk. This patch makes event_trace_printk into a macro wrapper to use the fmt as the same as the trace_printks. Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 17 +++++++++++++++++ kernel/trace/trace_events.c | 10 ---------- 2 files changed, 17 insertions(+), 10 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index cede1ab49d07..35cfa7bbaf38 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -773,4 +773,21 @@ void event_trace_printk(unsigned long ip, const char *fmt, ...); extern struct ftrace_event_call __start_ftrace_events[]; extern struct ftrace_event_call __stop_ftrace_events[]; +extern const char *__start___trace_bprintk_fmt[]; +extern const char *__stop___trace_bprintk_fmt[]; + +#define event_trace_printk(ip, fmt, args...) \ +do { \ + __trace_printk_check_format(fmt, ##args); \ + tracing_record_cmdline(current); \ + if (__builtin_constant_p(fmt)) { \ + static const char *trace_printk_fmt \ + __attribute__((section("__trace_printk_fmt"))) = \ + __builtin_constant_p(fmt) ? fmt : NULL; \ + \ + __trace_bprintk(ip, trace_printk_fmt, ##args); \ + } else \ + __trace_printk(ip, fmt, ##args); \ +} while (0) + #endif /* _LINUX_KERNEL_TRACE_H */ diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index ca624df73591..238ea95a4115 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -24,16 +24,6 @@ static DEFINE_MUTEX(event_mutex); (unsigned long)event < (unsigned long)__stop_ftrace_events; \ event++) -void event_trace_printk(unsigned long ip, const char *fmt, ...) -{ - va_list ap; - - va_start(ap, fmt); - tracing_record_cmdline(current); - trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); - va_end(ap); -} - static void ftrace_clear_events(void) { struct ftrace_event_call *call = (void *)__start_ftrace_events; From 7975a2be16dd42df2cab80c80cb6ece382edb6ec Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 14:23:17 -0400 Subject: [PATCH 287/413] tracing: export trace formats to user space The binary printk saves a pointer to the format string in the ring buffer. On output, the format is processed. But if the user is reading the ring buffer through a binary interface, the pointer is meaningless. This patch creates a file called printk_formats that maps the pointers to the formats. # cat /debug/tracing/printk_formats 0xffffffff80713d40 : "irq_handler_entry: irq=%d handler=%s\n" 0xffffffff80713d48 : "lock_acquire: %s%s%s\n" 0xffffffff80713d50 : "lock_release: %s\n" Signed-off-by: Steven Rostedt --- kernel/trace/trace_printk.c | 119 ++++++++++++++++++++++++++++++++++-- 1 file changed, 114 insertions(+), 5 deletions(-) diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c index f307a11e2332..486785214e3e 100644 --- a/kernel/trace/trace_printk.c +++ b/kernel/trace/trace_printk.c @@ -4,18 +4,19 @@ * Copyright (C) 2008 Lai Jiangshan * */ +#include +#include +#include #include #include #include +#include +#include +#include #include #include -#include #include -#include -#include #include -#include -#include #include "trace.h" @@ -153,6 +154,114 @@ int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) } EXPORT_SYMBOL_GPL(__ftrace_vprintk); +static void * +t_next(struct seq_file *m, void *v, loff_t *pos) +{ + const char **fmt = m->private; + const char **next = fmt; + + (*pos)++; + + if ((unsigned long)fmt >= (unsigned long)__stop___trace_bprintk_fmt) + return NULL; + + next = fmt; + m->private = ++next; + + return fmt; +} + +static void *t_start(struct seq_file *m, loff_t *pos) +{ + return t_next(m, NULL, pos); +} + +static int t_show(struct seq_file *m, void *v) +{ + const char **fmt = v; + const char *str = *fmt; + int i; + + seq_printf(m, "0x%lx : \"", (unsigned long)fmt); + + /* + * Tabs and new lines need to be converted. + */ + for (i = 0; str[i]; i++) { + switch (str[i]) { + case '\n': + seq_puts(m, "\\n"); + break; + case '\t': + seq_puts(m, "\\t"); + break; + case '\\': + seq_puts(m, "\\"); + break; + case '"': + seq_puts(m, "\\\""); + break; + default: + seq_putc(m, str[i]); + } + } + seq_puts(m, "\"\n"); + + return 0; +} + +static void t_stop(struct seq_file *m, void *p) +{ +} + +static const struct seq_operations show_format_seq_ops = { + .start = t_start, + .next = t_next, + .show = t_show, + .stop = t_stop, +}; + +static int +ftrace_formats_open(struct inode *inode, struct file *file) +{ + int ret; + + ret = seq_open(file, &show_format_seq_ops); + if (!ret) { + struct seq_file *m = file->private_data; + + m->private = __start___trace_bprintk_fmt; + } + return ret; +} + +static const struct file_operations ftrace_formats_fops = { + .open = ftrace_formats_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +static __init int init_trace_printk_function_export(void) +{ + struct dentry *d_tracer; + struct dentry *entry; + + d_tracer = tracing_init_dentry(); + if (!d_tracer) + return 0; + + entry = debugfs_create_file("printk_formats", 0444, d_tracer, + NULL, &ftrace_formats_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'printk_formats' entry\n"); + + return 0; +} + +fs_initcall(init_trace_printk_function_export); + static __init int init_trace_printk(void) { return register_module_notifier(&module_trace_bprintk_format_nb); From 2da03ecee6308ea174e8a02b92a3c4ec92e886c8 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 18:57:51 -0400 Subject: [PATCH 288/413] tracing: fix stack tracer header The stack tracer use to look like this: # cat /debug/tracing/stack_trace Depth Size Location (57 entries) ----- ---- -------- 0) 5088 16 mempool_alloc_slab+0x16/0x18 1) 5072 144 mempool_alloc+0x4d/0xfe 2) 4928 16 scsi_sg_alloc+0x48/0x4a [scsi_mod] Now it looks like this: # cat /debug/tracing/stack_trace Depth Size Location (57 entries) ----- ---- -------- 0) 5088 16 mempool_alloc_slab+0x16/0x18 1) 5072 144 mempool_alloc+0x4d/0xfe 2) 4928 16 scsi_sg_alloc+0x48/0x4a [scsi_mod] Signed-off-by: Steven Rostedt --- kernel/trace/trace_stack.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index d0871bc0aca5..4564fd94b0cd 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -251,9 +251,9 @@ static int t_show(struct seq_file *m, void *v) int size; if (v == SEQ_START_TOKEN) { - seq_printf(m, " Depth Size Location" + seq_printf(m, " Depth Size Location" " (%d entries)\n" - " ----- ---- --------\n", + " ----- ---- --------\n", max_stack_trace.nr_entries); return 0; } From e447e1df2e568cd43d1918963c9f09fae85aea57 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 19:42:29 -0400 Subject: [PATCH 289/413] tracing: explain why stack tracer is empty If the stack tracing is disabled (by default) the stack_trace file will only contain the header: # cat /debug/tracing/stack_trace Depth Size Location (0 entries) ----- ---- -------- This can be frustrating to a developer that does not realize that the stack tracer is disabled. This patch adds the following text: # cat /debug/tracing/stack_trace Depth Size Location (0 entries) ----- ---- -------- # # Stack tracer disabled # # To enable the stack tracer, either add 'stacktrace' to the # kernel command line # or 'echo 1 > /proc/sys/kernel/stack_tracer_enabled' # Signed-off-by: Steven Rostedt --- kernel/trace/trace_stack.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index 4564fd94b0cd..91ccbf396c9a 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -245,6 +245,17 @@ static int trace_lookup_stack(struct seq_file *m, long i) #endif } +static void print_disabled(struct seq_file *m) +{ + seq_puts(m, "#\n" + "# Stack tracer disabled\n" + "#\n" + "# To enable the stack tracer, either add 'stacktrace' to the\n" + "# kernel command line\n" + "# or 'echo 1 > /proc/sys/kernel/stack_tracer_enabled'\n" + "#\n"); +} + static int t_show(struct seq_file *m, void *v) { long i; @@ -255,6 +266,10 @@ static int t_show(struct seq_file *m, void *v) " (%d entries)\n" " ----- ---- --------\n", max_stack_trace.nr_entries); + + if (!stack_tracer_enabled && !max_stack_size) + print_disabled(m); + return 0; } From 5d592b44b29a1d73e13d5c9e3426eed843bdc359 Mon Sep 17 00:00:00 2001 From: Jason Baron Date: Thu, 12 Mar 2009 14:33:36 -0400 Subject: [PATCH 290/413] tracing: tracepoints for softirq entry/exit - add softirq-to-name array Create a 'softirq_to_name' array, which is indexed by softirq #, so that we can easily convert between the softirq index # and its name, in order to get more meaningful output messages. LKML-Reference: <20090312183336.GB3352@redhat.com> Signed-off-by: Jason Baron Signed-off-by: Steven Rostedt --- include/linux/interrupt.h | 5 +++++ kernel/softirq.c | 9 ++++++++- 2 files changed, 13 insertions(+), 1 deletion(-) diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index 472f11765f60..9b7e9d743476 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -258,6 +258,11 @@ enum NR_SOFTIRQS }; +/* map softirq index to softirq name. update 'softirq_to_name' in + * kernel/softirq.c when adding a new softirq. + */ +extern char *softirq_to_name[NR_SOFTIRQS]; + /* softirq mask and active fields moved to irq_cpustat_t in * asm/hardirq.h to get better cache usage. KAO */ diff --git a/kernel/softirq.c b/kernel/softirq.c index 7571bcb71be4..9f90fdc039f4 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -53,6 +53,12 @@ static struct softirq_action softirq_vec[NR_SOFTIRQS] __cacheline_aligned_in_smp static DEFINE_PER_CPU(struct task_struct *, ksoftirqd); +char *softirq_to_name[NR_SOFTIRQS] = { + "HI_SOFTIRQ", "TIMER_SOFTIRQ", "NET_TX_SOFTIRQ", "NET_RX_SOFTIRQ", + "BLOCK_SOFTIRQ", "TASKLET_SOFTIRQ", "SCHED_SOFTIRQ", "HRTIMER_SOFTIRQ", + "RCU_SOFTIRQ" +}; + /* * we cannot loop indefinitely here to avoid userspace starvation, * but we also don't want to introduce a worst case 1/HZ latency @@ -209,9 +215,10 @@ asmlinkage void __do_softirq(void) h->action(h); if (unlikely(prev_count != preempt_count())) { - printk(KERN_ERR "huh, entered softirq %td %p" + printk(KERN_ERR "huh, entered softirq %td %s %p" "with preempt_count %08x," " exited with %08x?\n", h - softirq_vec, + softirq_to_name[h - softirq_vec], h->action, prev_count, preempt_count()); preempt_count() = prev_count; } From 39842323ceb368d2ea36ab7696aedbe296e13b61 Mon Sep 17 00:00:00 2001 From: Jason Baron Date: Thu, 12 Mar 2009 14:36:03 -0400 Subject: [PATCH 291/413] tracing: tracepoints for softirq entry/exit - tracepoints Introduce softirq entry/exit tracepoints. These are useful for augmenting existing tracers, and to figure out softirq frequencies and timings. [ s/irq_softirq_/softirq_/ for trace point names and Fixed printf format in TRACE_FORMAT macro - Steven Rostedt ] LKML-Reference: <20090312183603.GC3352@redhat.com> Signed-off-by: Jason Baron Signed-off-by: Steven Rostedt --- include/trace/irq_event_types.h | 12 ++++++++++++ kernel/softirq.c | 7 ++++++- 2 files changed, 18 insertions(+), 1 deletion(-) diff --git a/include/trace/irq_event_types.h b/include/trace/irq_event_types.h index 214bb928fe9e..85964ebd47ec 100644 --- a/include/trace/irq_event_types.h +++ b/include/trace/irq_event_types.h @@ -40,4 +40,16 @@ TRACE_EVENT(irq_handler_exit, __entry->irq, __entry->ret ? "handled" : "unhandled") ); +TRACE_FORMAT(softirq_entry, + TP_PROTO(struct softirq_action *h, struct softirq_action *vec), + TP_ARGS(h, vec), + TP_FMT("softirq=%d action=%s", (int)(h - vec), softirq_to_name[h-vec]) + ); + +TRACE_FORMAT(softirq_exit, + TP_PROTO(struct softirq_action *h, struct softirq_action *vec), + TP_ARGS(h, vec), + TP_FMT("softirq=%d action=%s", (int)(h - vec), softirq_to_name[h-vec]) + ); + #undef TRACE_SYSTEM diff --git a/kernel/softirq.c b/kernel/softirq.c index 9f90fdc039f4..a5e81231ca7a 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -24,6 +24,7 @@ #include #include #include +#include #include /* @@ -186,6 +187,9 @@ EXPORT_SYMBOL(local_bh_enable_ip); */ #define MAX_SOFTIRQ_RESTART 10 +DEFINE_TRACE(softirq_entry); +DEFINE_TRACE(softirq_exit); + asmlinkage void __do_softirq(void) { struct softirq_action *h; @@ -212,8 +216,9 @@ asmlinkage void __do_softirq(void) if (pending & 1) { int prev_count = preempt_count(); + trace_softirq_entry(h, softirq_vec); h->action(h); - + trace_softirq_exit(h, softirq_vec); if (unlikely(prev_count != preempt_count())) { printk(KERN_ERR "huh, entered softirq %td %s %p" "with preempt_count %08x," From 889a6c367283709a80dad9413488472596a1a1d2 Mon Sep 17 00:00:00 2001 From: KOSAKI Motohiro Date: Fri, 13 Mar 2009 09:03:04 +0900 Subject: [PATCH 292/413] tracing: Don't use tracing_record_cmdline() in workqueue tracer fix commit c3ffc7a40b7e94b094efe1c8ab4e24370a782b65 "Don't use tracing_record_cmdline() in workqueue tracer" has a race window. find_task_by_vpid() requires task_list_lock(). LKML-Reference: <20090313090042.43CD.A69D9226@jp.fujitsu.com> Signed-off-by: KOSAKI Motohiro Signed-off-by: Steven Rostedt --- kernel/trace/trace_workqueue.c | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index fb5ccac8bbc0..9ab035b58cf1 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -193,12 +193,20 @@ static int workqueue_stat_show(struct seq_file *s, void *p) struct cpu_workqueue_stats *cws = p; unsigned long flags; int cpu = cws->cpu; - struct task_struct *tsk = find_task_by_vpid(cws->pid); + struct pid *pid; + struct task_struct *tsk; - seq_printf(s, "%3d %6d %6u %s\n", cws->cpu, - atomic_read(&cws->inserted), - cws->executed, - tsk ? tsk->comm : "<...>"); + pid = find_get_pid(cws->pid); + if (pid) { + tsk = get_pid_task(pid, PIDTYPE_PID); + if (tsk) { + seq_printf(s, "%3d %6d %6u %s\n", cws->cpu, + atomic_read(&cws->inserted), cws->executed, + tsk->comm); + put_task_struct(tsk); + } + put_pid(pid); + } spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); if (&cws->list == workqueue_cpu_stat(cpu)->list.next) From f28e55765e40450c127e44d00ae65d0cd1a4efec Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 22:00:19 -0400 Subject: [PATCH 293/413] tracing: show event name in trace for TRACE_EVENT created events Unlike TRACE_FORMAT() macros, the TRACE_EVENT() macros do not show the event name in the trace file. Knowing the event type in the trace output is very useful. Instead of: task swapper:0 [140] ==> ntpd:3308 [120] We now have: sched_switch: task swapper:0 [140] ==> ntpd:3308 [120] Signed-off-by: Steven Rostedt --- kernel/trace/trace_events_stage_2.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index ca347afd6aa0..5117c43f5c67 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -57,7 +57,7 @@ ftrace_raw_output_##call(struct trace_iterator *iter, int flags) \ \ field = (typeof(field))entry; \ \ - ret = trace_seq_printf(s, print); \ + ret = trace_seq_printf(s, #call ": " print); \ if (!ret) \ return TRACE_TYPE_PARTIAL_LINE; \ \ From 5cc985488845ec7227a2c5cfd2fd62cf57fb411a Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 12 Mar 2009 22:24:17 -0400 Subject: [PATCH 294/413] ring-buffer: document reader page design In a private email conversation I explained how the ring buffer page worked by using silly ASCII art. Ingo suggested that I add that to the comments of the code. Here it is. Requested-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 68 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 68 insertions(+) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 2c36be9fac2e..58128ad2fde0 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -21,6 +21,74 @@ #include "trace.h" +/* + * The ring buffer is made up of a list of pages. A separate list of pages is + * allocated for each CPU. A writer may only write to a buffer that is + * associated with the CPU it is currently executing on. A reader may read + * from any per cpu buffer. + * + * The reader is special. For each per cpu buffer, the reader has its own + * reader page. When a reader has read the entire reader page, this reader + * page is swapped with another page in the ring buffer. + * + * Now, as long as the writer is off the reader page, the reader can do what + * ever it wants with that page. The writer will never write to that page + * again (as long as it is out of the ring buffer). + * + * Here's some silly ASCII art. + * + * +------+ + * |reader| RING BUFFER + * |page | + * +------+ +---+ +---+ +---+ + * | |-->| |-->| | + * +---+ +---+ +---+ + * ^ | + * | | + * +---------------+ + * + * + * +------+ + * |reader| RING BUFFER + * |page |------------------v + * +------+ +---+ +---+ +---+ + * | |-->| |-->| | + * +---+ +---+ +---+ + * ^ | + * | | + * +---------------+ + * + * + * +------+ + * |reader| RING BUFFER + * |page |------------------v + * +------+ +---+ +---+ +---+ + * ^ | |-->| |-->| | + * | +---+ +---+ +---+ + * | | + * | | + * +------------------------------+ + * + * + * +------+ + * |buffer| RING BUFFER + * |page |------------------v + * +------+ +---+ +---+ +---+ + * ^ | | | |-->| | + * | New +---+ +---+ +---+ + * | Reader------^ | + * | page | + * +------------------------------+ + * + * + * After we make this swap, the reader can hand this page off to the splice + * code and be done with it. It can even allocate a new page if it needs to + * and swap that into the ring buffer. + * + * We will be using cmpxchg soon to make all this lockless. + * + */ + /* * A fast way to enable or disable all ring buffers is to * call tracing_on or tracing_off. Turning off the ring buffers From 4c6ed8f496fe89ff81b3d6e617a531bedfb613e4 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 13 Mar 2009 04:33:17 +0100 Subject: [PATCH 295/413] Merge branches 'tracing/ftrace' and 'linus' into tracing/core From eb1871f34358024acfa3523ef375ef14b7527173 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Mar 2009 00:00:58 -0400 Subject: [PATCH 296/413] tracing: left align location header in stack_trace Ingo Molnar suggested, instead of: Depth Size Location (27 entries) ----- ---- -------- 0) 2880 48 lock_timer_base+0x2b/0x4f 1) 2832 80 __mod_timer+0x33/0xe0 2) 2752 16 __ide_set_handler+0x63/0x65 To have it be: Depth Size Location (27 entries) ----- ---- -------- 0) 2880 48 lock_timer_base+0x2b/0x4f 1) 2832 80 __mod_timer+0x33/0xe0 2) 2752 16 __ide_set_handler+0x63/0x65 Requested-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/trace_stack.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index 91ccbf396c9a..c750f65f9661 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -262,9 +262,9 @@ static int t_show(struct seq_file *m, void *v) int size; if (v == SEQ_START_TOKEN) { - seq_printf(m, " Depth Size Location" + seq_printf(m, " Depth Size Location" " (%d entries)\n" - " ----- ---- --------\n", + " ----- ---- --------\n", max_stack_trace.nr_entries); if (!stack_tracer_enabled && !max_stack_size) From bdc067582b8b71c7771bab076bbc51569c594fb4 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Mar 2009 00:12:52 -0400 Subject: [PATCH 297/413] tracing: add comment for use of double __builtin_consant_p Impact: documentation The use of the double __builtin_contant_p checks in the event_trace_printk can be confusing to developers and reviewers. This patch adds a comment to explain why it is there. Requested-by: KOSAKI Motohiro LKML-Reference: <20090313122235.43EB.A69D9226@jp.fujitsu.com> Signed-off-by: Steven Rostedt --- kernel/trace/trace.h | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 35cfa7bbaf38..67595b8f0f15 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -776,6 +776,11 @@ extern struct ftrace_event_call __stop_ftrace_events[]; extern const char *__start___trace_bprintk_fmt[]; extern const char *__stop___trace_bprintk_fmt[]; +/* + * The double __builtin_constant_p is because gcc will give us an error + * if we try to allocate the static variable to fmt if it is not a + * constant. Even with the outer if statement optimizing out. + */ #define event_trace_printk(ip, fmt, args...) \ do { \ __trace_printk_check_format(fmt, ##args); \ From 7f96f93f02b7637491a1637dee12dcdcd40b9802 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Mar 2009 00:37:42 -0400 Subject: [PATCH 298/413] tracing: move binary buffers into per cpu directory The binary_buffers directory in /debugfs/tracing held the files to read the trace buffers in a binary format. This held one file per CPU buffer. But we also have a per_cpu directory that holds a way to read the pretty-print formats. This patch moves the binary buffers into the per_cpu_directory: # ls /debug/tracing/per_cpu/cpu1/ trace trace_pipe trace_pipe_raw The new name is called "trace_pipe_raw". The binary buffers always acted similar to trace_pipe, except that they produce raw data. Requested-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 26 +++++--------------------- 1 file changed, 5 insertions(+), 21 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index dbb077d8a172..efe3202c0209 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3543,6 +3543,11 @@ static void tracing_init_debugfs_percpu(long cpu) (void *) cpu, &tracing_fops); if (!entry) pr_warning("Could not create debugfs 'trace' entry\n"); + + entry = debugfs_create_file("trace_pipe_raw", 0444, d_cpu, + (void *) cpu, &tracing_buffers_fops); + if (!entry) + pr_warning("Could not create debugfs 'trace_pipe_raw' entry\n"); } #ifdef CONFIG_FTRACE_SELFTEST @@ -3826,7 +3831,6 @@ static __init void create_trace_options_dir(void) static __init int tracer_init_debugfs(void) { struct dentry *d_tracer; - struct dentry *buffers; struct dentry *entry; int cpu; @@ -3899,26 +3903,6 @@ static __init int tracer_init_debugfs(void) pr_warning("Could not create debugfs " "'trace_marker' entry\n"); - buffers = debugfs_create_dir("binary_buffers", d_tracer); - - if (!buffers) - pr_warning("Could not create buffers directory\n"); - else { - int cpu; - char buf[64]; - - for_each_tracing_cpu(cpu) { - sprintf(buf, "%d", cpu); - - entry = debugfs_create_file(buf, 0444, buffers, - (void *)(long)cpu, - &tracing_buffers_fops); - if (!entry) - pr_warning("Could not create debugfs buffers " - "'%s' entry\n", buf); - } - } - #ifdef CONFIG_DYNAMIC_FTRACE entry = debugfs_create_file("dyn_ftrace_total_info", 0444, d_tracer, &ftrace_update_tot_cnt, From 899039e8746bb9a09b6487ddb8ab2275ce9d0256 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 13 Mar 2009 00:43:33 -0400 Subject: [PATCH 299/413] softirq: no need to have SOFTIRQ in softirq name Impact: clean up It is redundant to have 'SOFTIRQ' in the softirq names. Reported-by: Andrew Morton Signed-off-by: Steven Rostedt --- kernel/softirq.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/kernel/softirq.c b/kernel/softirq.c index a5e81231ca7a..65ff3e3961b4 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -55,9 +55,8 @@ static struct softirq_action softirq_vec[NR_SOFTIRQS] __cacheline_aligned_in_smp static DEFINE_PER_CPU(struct task_struct *, ksoftirqd); char *softirq_to_name[NR_SOFTIRQS] = { - "HI_SOFTIRQ", "TIMER_SOFTIRQ", "NET_TX_SOFTIRQ", "NET_RX_SOFTIRQ", - "BLOCK_SOFTIRQ", "TASKLET_SOFTIRQ", "SCHED_SOFTIRQ", "HRTIMER_SOFTIRQ", - "RCU_SOFTIRQ" + "HI", "TIMER", "NET_TX", "NET_RX", "BLOCK", + "TASKLET", "SCHED", "HRTIMER", "RCU" }; /* From ee08c6eccb7d1295516f7cf420fddf7b14e9146f Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 7 Mar 2009 05:52:59 +0100 Subject: [PATCH 300/413] tracing/ftrace: syscall tracing infrastructure, basics Provide basic callbacks to do syscall tracing. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Cc: Lai Jiangshan LKML-Reference: <1236401580-5758-2-git-send-email-fweisbec@gmail.com> [ simplified it to a trace_printk() for now. ] Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 21 +++++++ kernel/trace/Kconfig | 10 +++ kernel/trace/Makefile | 1 + kernel/trace/trace.h | 2 + kernel/trace/trace_syscalls.c | 113 ++++++++++++++++++++++++++++++++++ 5 files changed, 147 insertions(+) create mode 100644 kernel/trace/trace_syscalls.c diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index e1583f2639b0..c146c1021a29 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -503,4 +503,25 @@ static inline void trace_hw_branch_oops(void) {} #endif /* CONFIG_HW_BRANCH_TRACER */ +/* + * A syscall entry in the ftrace syscalls array. + * + * @syscall_nr: syscall number + */ +struct syscall_trace_entry { + int syscall_nr; +}; + +#ifdef CONFIG_FTRACE_SYSCALLS +extern void start_ftrace_syscalls(void); +extern void stop_ftrace_syscalls(void); +extern void ftrace_syscall_enter(struct pt_regs *regs); +extern void ftrace_syscall_exit(struct pt_regs *regs); +#else +static inline void start_ftrace_syscalls(void) { } +static inline void stop_ftrace_syscalls(void) { } +static inline void ftrace_syscall_enter(struct pt_regs *regs) { } +static inline void ftrace_syscall_exit(struct pt_regs *regs) { } +#endif + #endif /* _LINUX_FTRACE_H */ diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 8e4a2a61cd75..95a0ad191f19 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -34,6 +34,9 @@ config HAVE_FTRACE_MCOUNT_RECORD config HAVE_HW_BRANCH_TRACER bool +config HAVE_FTRACE_SYSCALLS + bool + config TRACER_MAX_TRACE bool @@ -175,6 +178,13 @@ config EVENT_TRACER allowing the user to pick and choose which trace point they want to trace. +config FTRACE_SYSCALLS + bool "Trace syscalls" + depends on HAVE_FTRACE_SYSCALLS + select TRACING + help + Basic tracer to catch the syscall entry and exit events. + config BOOT_TRACER bool "Trace boot initcalls" select TRACING diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index c7a2943796eb..c3feea01c3e0 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -43,5 +43,6 @@ obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o obj-$(CONFIG_EVENT_TRACER) += trace_events.o obj-$(CONFIG_EVENT_TRACER) += events.o obj-$(CONFIG_EVENT_TRACER) += trace_export.o +obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index c5e1d8865fe4..3d49daae47dc 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -30,6 +30,8 @@ enum trace_type { TRACE_GRAPH_ENT, TRACE_USER_STACK, TRACE_HW_BRANCHES, + TRACE_SYSCALL_ENTER, + TRACE_SYSCALL_EXIT, TRACE_KMEM_ALLOC, TRACE_KMEM_FREE, TRACE_POWER, diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c new file mode 100644 index 000000000000..66cf97449af3 --- /dev/null +++ b/kernel/trace/trace_syscalls.c @@ -0,0 +1,113 @@ +#include +#include + +#include + +#include "trace_output.h" +#include "trace.h" + +static atomic_t refcount; + +void start_ftrace_syscalls(void) +{ + unsigned long flags; + struct task_struct *g, *t; + + if (atomic_inc_return(&refcount) != 1) + goto out; + + read_lock_irqsave(&tasklist_lock, flags); + + do_each_thread(g, t) { + set_tsk_thread_flag(t, TIF_SYSCALL_FTRACE); + } while_each_thread(g, t); + + read_unlock_irqrestore(&tasklist_lock, flags); +out: + atomic_dec(&refcount); +} + +void stop_ftrace_syscalls(void) +{ + unsigned long flags; + struct task_struct *g, *t; + + if (atomic_dec_return(&refcount)) + goto out; + + read_lock_irqsave(&tasklist_lock, flags); + + do_each_thread(g, t) { + clear_tsk_thread_flag(t, TIF_SYSCALL_FTRACE); + } while_each_thread(g, t); + + read_unlock_irqrestore(&tasklist_lock, flags); +out: + atomic_inc(&refcount); +} + +void ftrace_syscall_enter(struct pt_regs *regs) +{ + int syscall_nr; + + syscall_nr = syscall_get_nr(current, regs); + + trace_printk("syscall %d enter\n", syscall_nr); +} + +void ftrace_syscall_exit(struct pt_regs *regs) +{ + int syscall_nr; + + syscall_nr = syscall_get_nr(current, regs); + + trace_printk("syscall %d exit\n", syscall_nr); +} + +static int init_syscall_tracer(struct trace_array *tr) +{ + start_ftrace_syscalls(); + + return 0; +} + +static void reset_syscall_tracer(struct trace_array *tr) +{ + stop_ftrace_syscalls(); +} + +static struct trace_event syscall_enter_event = { + .type = TRACE_SYSCALL_ENTER, +}; + +static struct trace_event syscall_exit_event = { + .type = TRACE_SYSCALL_EXIT, +}; + +static struct tracer syscall_tracer __read_mostly = { + .name = "syscall", + .init = init_syscall_tracer, + .reset = reset_syscall_tracer +}; + +__init int register_ftrace_syscalls(void) +{ + int ret; + + ret = register_ftrace_event(&syscall_enter_event); + if (!ret) { + printk(KERN_WARNING "event %d failed to register\n", + syscall_enter_event.type); + WARN_ON_ONCE(1); + } + + ret = register_ftrace_event(&syscall_exit_event); + if (!ret) { + printk(KERN_WARNING "event %d failed to register\n", + syscall_exit_event.type); + WARN_ON_ONCE(1); + } + + return register_tracer(&syscall_tracer); +} +device_initcall(register_ftrace_syscalls); From 1b3fa2ce64363c289b3b14723cca7290bf91cfce Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 7 Mar 2009 05:53:00 +0100 Subject: [PATCH 301/413] tracing/x86: basic implementation of syscall tracing for x86 Provide the x86 trace callbacks to trace syscalls. Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt Cc: Lai Jiangshan LKML-Reference: <1236401580-5758-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 1 + arch/x86/include/asm/thread_info.h | 9 ++++++--- arch/x86/kernel/ptrace.c | 7 +++++++ 3 files changed, 14 insertions(+), 3 deletions(-) diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index bdcee12c25ab..b0a638b4199a 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -35,6 +35,7 @@ config X86 select HAVE_FUNCTION_GRAPH_TRACER select HAVE_FUNCTION_TRACE_MCOUNT_TEST select HAVE_FTRACE_NMI_ENTER if DYNAMIC_FTRACE + select HAVE_FTRACE_SYSCALLS select HAVE_KVM select HAVE_ARCH_KGDB select HAVE_ARCH_TRACEHOOK diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h index df9d5f78385e..8820a73ae090 100644 --- a/arch/x86/include/asm/thread_info.h +++ b/arch/x86/include/asm/thread_info.h @@ -94,6 +94,7 @@ struct thread_info { #define TIF_FORCED_TF 24 /* true if TF in eflags artificially */ #define TIF_DEBUGCTLMSR 25 /* uses thread_struct.debugctlmsr */ #define TIF_DS_AREA_MSR 26 /* uses thread_struct.ds_area_msr */ +#define TIF_SYSCALL_FTRACE 27 /* for ftrace syscall instrumentation */ #define _TIF_SYSCALL_TRACE (1 << TIF_SYSCALL_TRACE) #define _TIF_NOTIFY_RESUME (1 << TIF_NOTIFY_RESUME) @@ -115,15 +116,17 @@ struct thread_info { #define _TIF_FORCED_TF (1 << TIF_FORCED_TF) #define _TIF_DEBUGCTLMSR (1 << TIF_DEBUGCTLMSR) #define _TIF_DS_AREA_MSR (1 << TIF_DS_AREA_MSR) +#define _TIF_SYSCALL_FTRACE (1 << TIF_SYSCALL_FTRACE) /* work to do in syscall_trace_enter() */ #define _TIF_WORK_SYSCALL_ENTRY \ - (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_EMU | \ + (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_EMU | _TIF_SYSCALL_FTRACE | \ _TIF_SYSCALL_AUDIT | _TIF_SECCOMP | _TIF_SINGLESTEP) /* work to do in syscall_trace_leave() */ #define _TIF_WORK_SYSCALL_EXIT \ - (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SINGLESTEP) + (_TIF_SYSCALL_TRACE | _TIF_SYSCALL_AUDIT | _TIF_SINGLESTEP | \ + _TIF_SYSCALL_FTRACE) /* work to do on interrupt/exception return */ #define _TIF_WORK_MASK \ @@ -132,7 +135,7 @@ struct thread_info { _TIF_SINGLESTEP|_TIF_SECCOMP|_TIF_SYSCALL_EMU)) /* work to do on any return to user space */ -#define _TIF_ALLWORK_MASK (0x0000FFFF & ~_TIF_SECCOMP) +#define _TIF_ALLWORK_MASK ((0x0000FFFF & ~_TIF_SECCOMP) | _TIF_SYSCALL_FTRACE) /* Only used for 64 bit */ #define _TIF_DO_NOTIFY_MASK \ diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index 3d9672e59c16..99749d6e87a8 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -21,6 +21,7 @@ #include #include #include +#include #include #include @@ -1416,6 +1417,9 @@ asmregparm long syscall_trace_enter(struct pt_regs *regs) tracehook_report_syscall_entry(regs)) ret = -1L; + if (unlikely(test_thread_flag(TIF_SYSCALL_FTRACE))) + ftrace_syscall_enter(regs); + if (unlikely(current->audit_context)) { if (IS_IA32) audit_syscall_entry(AUDIT_ARCH_I386, @@ -1439,6 +1443,9 @@ asmregparm void syscall_trace_leave(struct pt_regs *regs) if (unlikely(current->audit_context)) audit_syscall_exit(AUDITSC_RESULT(regs->ax), regs->ax); + if (unlikely(test_thread_flag(TIF_SYSCALL_FTRACE))) + ftrace_syscall_exit(regs); + if (test_thread_flag(TIF_SYSCALL_TRACE)) tracehook_report_syscall_exit(regs, 0); From b00f0b6dc1773b4c8f538503247da050b5ea631b Mon Sep 17 00:00:00 2001 From: Zhaolei Date: Fri, 13 Mar 2009 17:14:01 +0800 Subject: [PATCH 302/413] ftrace: avoid double-free of dyn_ftrace If dyn_ftrace is freed before ftrace_release(), ftrace_release() will free it again and make ftrace_free_records wrong. Signed-off-by: Zhao Lei Cc: "Steven Rostedt ;" LKML-Reference: <49BA23D9.1050900@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index d33d306bdcf4..26c45aaf6805 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -356,7 +356,8 @@ void ftrace_release(void *start, unsigned long size) mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { - if ((rec->ip >= s) && (rec->ip < e)) + if ((rec->ip >= s) && (rec->ip < e) && + !(rec->flags & FTRACE_FL_FREE)) ftrace_free_rec(rec); } while_for_each_ftrace_rec(); mutex_unlock(&ftrace_lock); From fa9d13cf135efbd454453a53b6299976bea245a9 Mon Sep 17 00:00:00 2001 From: Zhaolei Date: Fri, 13 Mar 2009 17:16:34 +0800 Subject: [PATCH 303/413] ftrace: don't try to __ftrace_replace_code on !FTRACE_FL_CONVERTED rec Do __ftrace_replace_code for !FTRACE_FL_CONVERTED rec will always fail, we should ignore this rec. Signed-off-by: Zhao Lei Cc: "Steven Rostedt ;" LKML-Reference: <49BA2472.4060206@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 26c45aaf6805..08f4a624e31f 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -532,11 +532,12 @@ static void ftrace_replace_code(int enable) do_for_each_ftrace_rec(pg, rec) { /* - * Skip over free records and records that have - * failed. + * Skip over free records, records that have + * failed and not converted. */ if (rec->flags & FTRACE_FL_FREE || - rec->flags & FTRACE_FL_FAILED) + rec->flags & FTRACE_FL_FAILED || + rec->flags & FTRACE_FL_CONVERTED) continue; /* ignore updates to this record's mcount site */ @@ -548,7 +549,7 @@ static void ftrace_replace_code(int enable) } failed = __ftrace_replace_code(rec, enable); - if (failed && (rec->flags & FTRACE_FL_CONVERTED)) { + if (failed) { rec->flags |= FTRACE_FL_FAILED; if ((system_state == SYSTEM_BOOTING) || !core_kernel_text(rec->ip)) { From 850a80cfaa5aec3e626eb3736eff890a80e4fa77 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 13 Mar 2009 17:47:23 +0800 Subject: [PATCH 304/413] ftrace: use seq_read Impact: cleanup VFS layer has tested the file mode, we do not need test it. Signed-off-by: Lai Jiangshan Cc: Steven Rostedt LKML-Reference: <49BA2BAB.6010608@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 26 +++----------------------- 1 file changed, 3 insertions(+), 23 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 08f4a624e31f..bf78a4c75c67 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -1120,16 +1120,6 @@ ftrace_notrace_open(struct inode *inode, struct file *file) return ftrace_regex_open(inode, file, 0); } -static ssize_t -ftrace_regex_read(struct file *file, char __user *ubuf, - size_t cnt, loff_t *ppos) -{ - if (file->f_mode & FMODE_READ) - return seq_read(file, ubuf, cnt, ppos); - else - return -EPERM; -} - static loff_t ftrace_regex_lseek(struct file *file, loff_t offset, int origin) { @@ -1882,7 +1872,7 @@ static const struct file_operations ftrace_failures_fops = { static const struct file_operations ftrace_filter_fops = { .open = ftrace_filter_open, - .read = ftrace_regex_read, + .read = seq_read, .write = ftrace_filter_write, .llseek = ftrace_regex_lseek, .release = ftrace_filter_release, @@ -1890,7 +1880,7 @@ static const struct file_operations ftrace_filter_fops = { static const struct file_operations ftrace_notrace_fops = { .open = ftrace_notrace_open, - .read = ftrace_regex_read, + .read = seq_read, .write = ftrace_notrace_write, .llseek = ftrace_regex_lseek, .release = ftrace_notrace_release, @@ -1992,16 +1982,6 @@ ftrace_graph_open(struct inode *inode, struct file *file) return ret; } -static ssize_t -ftrace_graph_read(struct file *file, char __user *ubuf, - size_t cnt, loff_t *ppos) -{ - if (file->f_mode & FMODE_READ) - return seq_read(file, ubuf, cnt, ppos); - else - return -EPERM; -} - static int ftrace_set_func(unsigned long *array, int *idx, char *buffer) { @@ -2132,7 +2112,7 @@ ftrace_graph_write(struct file *file, const char __user *ubuf, static const struct file_operations ftrace_graph_fops = { .open = ftrace_graph_open, - .read = ftrace_graph_read, + .read = seq_read, .write = ftrace_graph_write, }; #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ From e94142a67f8bad494c593f0a07c9fc2fbec98c0e Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 13 Mar 2009 17:51:27 +0800 Subject: [PATCH 305/413] ftrace: remove struct list_head from struct dyn_ftrace Impact: save memory The struct dyn_ftrace table is very large, this patch will save about 50%. Signed-off-by: Lai Jiangshan Cc: Steven Rostedt LKML-Reference: <49BA2C9F.8020009@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 1 - kernel/trace/ftrace.c | 14 ++++++++------ 2 files changed, 8 insertions(+), 7 deletions(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index c146c1021a29..9d598bbf28a6 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -145,7 +145,6 @@ enum { }; struct dyn_ftrace { - struct list_head list; unsigned long ip; /* address of mcount call-site */ unsigned long flags; struct dyn_arch_ftrace arch; diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index bf78a4c75c67..90d5729afeff 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -272,7 +272,7 @@ enum { static int ftrace_filtered; -static LIST_HEAD(ftrace_new_addrs); +static struct dyn_ftrace *ftrace_new_addrs; static DEFINE_MUTEX(ftrace_regex_lock); @@ -409,8 +409,8 @@ ftrace_record_ip(unsigned long ip) return NULL; rec->ip = ip; - - list_add(&rec->list, &ftrace_new_addrs); + rec->flags = (unsigned long)ftrace_new_addrs; + ftrace_new_addrs = rec; return rec; } @@ -716,19 +716,21 @@ unsigned long ftrace_update_tot_cnt; static int ftrace_update_code(struct module *mod) { - struct dyn_ftrace *p, *t; + struct dyn_ftrace *p; cycle_t start, stop; start = ftrace_now(raw_smp_processor_id()); ftrace_update_cnt = 0; - list_for_each_entry_safe(p, t, &ftrace_new_addrs, list) { + while (ftrace_new_addrs) { /* If something went wrong, bail without enabling anything */ if (unlikely(ftrace_disabled)) return -1; - list_del_init(&p->list); + p = ftrace_new_addrs; + ftrace_new_addrs = (struct dyn_ftrace *)p->flags; + p->flags = 0L; /* convert record (i.e, patch mcount-call with NOP) */ if (ftrace_code_disable(mod, p)) { From f58ba100678f421bdcb000a3c71793f432dfab93 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 13 Mar 2009 15:42:12 +0100 Subject: [PATCH 306/413] tracing/syscalls: support for syscalls tracing on x86 Extend x86 architecture syscall tracing support with syscall metadata table details. (The upcoming core syscall tracing modifications rely on this.) Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1236955332-10133-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/ftrace.h | 7 ++++ arch/x86/kernel/ftrace.c | 63 +++++++++++++++++++++++++++++++++++ 2 files changed, 70 insertions(+) diff --git a/arch/x86/include/asm/ftrace.h b/arch/x86/include/asm/ftrace.h index db24c2278be0..bd2c6511c887 100644 --- a/arch/x86/include/asm/ftrace.h +++ b/arch/x86/include/asm/ftrace.h @@ -28,6 +28,13 @@ #endif +/* FIXME: I don't want to stay hardcoded */ +#ifdef CONFIG_X86_64 +# define FTRACE_SYSCALL_MAX 296 +#else +# define FTRACE_SYSCALL_MAX 333 +#endif + #ifdef CONFIG_FUNCTION_TRACER #define MCOUNT_ADDR ((long)(mcount)) #define MCOUNT_INSN_SIZE 5 /* sizeof mcount call */ diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index a85da1764b1c..1d0d7f42efe3 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -453,3 +453,66 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) } } #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ + +#ifdef CONFIG_FTRACE_SYSCALLS + +extern unsigned long __start_syscalls_metadata[]; +extern unsigned long __stop_syscalls_metadata[]; +extern unsigned long *sys_call_table; + +static struct syscall_metadata **syscalls_metadata; + +static struct syscall_metadata *find_syscall_meta(unsigned long *syscall) +{ + struct syscall_metadata *start; + struct syscall_metadata *stop; + char str[KSYM_SYMBOL_LEN]; + + + start = (struct syscall_metadata *)__start_syscalls_metadata; + stop = (struct syscall_metadata *)__stop_syscalls_metadata; + kallsyms_lookup((unsigned long) syscall, NULL, NULL, NULL, str); + + for ( ; start < stop; start++) { + if (start->name && !strcmp(start->name, str)) + return start; + } + return NULL; +} + +struct syscall_metadata *syscall_nr_to_meta(int nr) +{ + if (!syscalls_metadata || nr >= FTRACE_SYSCALL_MAX || nr < 0) + return NULL; + + return syscalls_metadata[nr]; +} + +void arch_init_ftrace_syscalls(void) +{ + int i; + struct syscall_metadata *meta; + unsigned long **psys_syscall_table = &sys_call_table; + static atomic_t refs; + + if (atomic_inc_return(&refs) != 1) + goto end; + + syscalls_metadata = kzalloc(sizeof(*syscalls_metadata) * + FTRACE_SYSCALL_MAX, GFP_KERNEL); + if (!syscalls_metadata) { + WARN_ON(1); + return; + } + + for (i = 0; i < FTRACE_SYSCALL_MAX; i++) { + meta = find_syscall_meta(psys_syscall_table[i]); + syscalls_metadata[i] = meta; + } + return; + + /* Paranoid: avoid overflow */ +end: + atomic_dec(&refs); +} +#endif From bed1ffca022cc876fb83161d26670e9b5d3cf36b Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 13 Mar 2009 15:42:11 +0100 Subject: [PATCH 307/413] tracing/syscalls: core infrastructure for syscalls tracing, enhancements Impact: new feature This adds the generic support for syscalls tracing. This is currently exploited through a devoted tracer but other tracing engines can use it. (They just have to play with {start,stop}_ftrace_syscalls() and use the display callbacks unless they want to override them.) The syscalls prototypes definitions are abused here to steal some metadata informations: - syscall name, param types, param names, number of params The syscall addr is not directly saved during this definition because we don't know if its prototype is available in the namespace. But we don't really need it. The arch has just to build a function able to resolve the syscall number to its metadata struct. The current tracer prints the syscall names, parameters names and values (and their types optionally). Currently the value is a raw hex but higher level values diplaying is on my TODO list. Signed-off-by: Frederic Weisbecker LKML-Reference: <1236955332-10133-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- include/asm-generic/vmlinux.lds.h | 11 ++- include/linux/ftrace.h | 14 ++- include/linux/syscalls.h | 60 +++++++++++- kernel/trace/trace.h | 17 ++++ kernel/trace/trace_syscalls.c | 146 ++++++++++++++++++++++++++++-- 5 files changed, 234 insertions(+), 14 deletions(-) diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 0e0f39be6c8b..d3bc3c86df6a 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -77,6 +77,14 @@ #define TRACE_PRINTKS() #endif +#ifdef CONFIG_FTRACE_SYSCALLS +#define TRACE_SYSCALLS() VMLINUX_SYMBOL(__start_syscalls_metadata) = .; \ + *(__syscalls_metadata) \ + VMLINUX_SYMBOL(__stop_syscalls_metadata) = .; +#else +#define TRACE_SYSCALLS() +#endif + /* .data section */ #define DATA_DATA \ *(.data) \ @@ -99,7 +107,8 @@ LIKELY_PROFILE() \ BRANCH_PROFILE() \ TRACE_PRINTKS() \ - FTRACE_EVENTS() + FTRACE_EVENTS() \ + TRACE_SYSCALLS() #define RO_DATA(align) \ . = ALIGN((align)); \ diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index c146c1021a29..6dc1c652447e 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -506,13 +506,21 @@ static inline void trace_hw_branch_oops(void) {} /* * A syscall entry in the ftrace syscalls array. * - * @syscall_nr: syscall number + * @name: name of the syscall + * @nb_args: number of parameters it takes + * @types: list of types as strings + * @args: list of args as strings (args[i] matches types[i]) */ -struct syscall_trace_entry { - int syscall_nr; +struct syscall_metadata { + const char *name; + int nb_args; + const char **types; + const char **args; }; #ifdef CONFIG_FTRACE_SYSCALLS +extern void arch_init_ftrace_syscalls(void); +extern struct syscall_metadata *syscall_nr_to_meta(int nr); extern void start_ftrace_syscalls(void); extern void stop_ftrace_syscalls(void); extern void ftrace_syscall_enter(struct pt_regs *regs); diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index f9f900cfd066..0cff9bb80b02 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -65,6 +65,7 @@ struct old_linux_dirent; #include #include #include +#include #define __SC_DECL1(t1, a1) t1 a1 #define __SC_DECL2(t2, a2, ...) t2 a2, __SC_DECL1(__VA_ARGS__) @@ -95,7 +96,46 @@ struct old_linux_dirent; #define __SC_TEST5(t5, a5, ...) __SC_TEST(t5); __SC_TEST4(__VA_ARGS__) #define __SC_TEST6(t6, a6, ...) __SC_TEST(t6); __SC_TEST5(__VA_ARGS__) +#ifdef CONFIG_FTRACE_SYSCALLS +#define __SC_STR_ADECL1(t, a) #a +#define __SC_STR_ADECL2(t, a, ...) #a, __SC_STR_ADECL1(__VA_ARGS__) +#define __SC_STR_ADECL3(t, a, ...) #a, __SC_STR_ADECL2(__VA_ARGS__) +#define __SC_STR_ADECL4(t, a, ...) #a, __SC_STR_ADECL3(__VA_ARGS__) +#define __SC_STR_ADECL5(t, a, ...) #a, __SC_STR_ADECL4(__VA_ARGS__) +#define __SC_STR_ADECL6(t, a, ...) #a, __SC_STR_ADECL5(__VA_ARGS__) + +#define __SC_STR_TDECL1(t, a) #t +#define __SC_STR_TDECL2(t, a, ...) #t, __SC_STR_TDECL1(__VA_ARGS__) +#define __SC_STR_TDECL3(t, a, ...) #t, __SC_STR_TDECL2(__VA_ARGS__) +#define __SC_STR_TDECL4(t, a, ...) #t, __SC_STR_TDECL3(__VA_ARGS__) +#define __SC_STR_TDECL5(t, a, ...) #t, __SC_STR_TDECL4(__VA_ARGS__) +#define __SC_STR_TDECL6(t, a, ...) #t, __SC_STR_TDECL5(__VA_ARGS__) + +#define SYSCALL_METADATA(sname, nb) \ + static const struct syscall_metadata __used \ + __attribute__((__aligned__(4))) \ + __attribute__((section("__syscalls_metadata"))) \ + __syscall_meta_##sname = { \ + .name = "sys"#sname, \ + .nb_args = nb, \ + .types = types_##sname, \ + .args = args_##sname, \ + } + +#define SYSCALL_DEFINE0(sname) \ + static const struct syscall_metadata __used \ + __attribute__((__aligned__(4))) \ + __attribute__((section("__syscalls_metadata"))) \ + __syscall_meta_##sname = { \ + .name = "sys_"#sname, \ + .nb_args = 0, \ + }; \ + asmlinkage long sys_##sname(void) + +#else #define SYSCALL_DEFINE0(name) asmlinkage long sys_##name(void) +#endif + #define SYSCALL_DEFINE1(name, ...) SYSCALL_DEFINEx(1, _##name, __VA_ARGS__) #define SYSCALL_DEFINE2(name, ...) SYSCALL_DEFINEx(2, _##name, __VA_ARGS__) #define SYSCALL_DEFINE3(name, ...) SYSCALL_DEFINEx(3, _##name, __VA_ARGS__) @@ -117,10 +157,26 @@ struct old_linux_dirent; #endif #endif +#ifdef CONFIG_FTRACE_SYSCALLS +#define SYSCALL_DEFINEx(x, sname, ...) \ + static const char *types_##sname[] = { \ + __SC_STR_TDECL##x(__VA_ARGS__) \ + }; \ + static const char *args_##sname[] = { \ + __SC_STR_ADECL##x(__VA_ARGS__) \ + }; \ + SYSCALL_METADATA(sname, x); \ + __SYSCALL_DEFINEx(x, sname, __VA_ARGS__) +#else +#define SYSCALL_DEFINEx(x, sname, ...) \ + __SYSCALL_DEFINEx(x, sname, __VA_ARGS__) +#endif + #ifdef CONFIG_HAVE_SYSCALL_WRAPPERS #define SYSCALL_DEFINE(name) static inline long SYSC_##name -#define SYSCALL_DEFINEx(x, name, ...) \ + +#define __SYSCALL_DEFINEx(x, name, ...) \ asmlinkage long sys##name(__SC_DECL##x(__VA_ARGS__)); \ static inline long SYSC##name(__SC_DECL##x(__VA_ARGS__)); \ asmlinkage long SyS##name(__SC_LONG##x(__VA_ARGS__)) \ @@ -134,7 +190,7 @@ struct old_linux_dirent; #else /* CONFIG_HAVE_SYSCALL_WRAPPERS */ #define SYSCALL_DEFINE(name) asmlinkage long sys_##name -#define SYSCALL_DEFINEx(x, name, ...) \ +#define __SYSCALL_DEFINEx(x, name, ...) \ asmlinkage long sys##name(__SC_DECL##x(__VA_ARGS__)) #endif /* CONFIG_HAVE_SYSCALL_WRAPPERS */ diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 3d49daae47dc..d80ca0d464d9 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -194,6 +194,19 @@ struct kmemtrace_free_entry { const void *ptr; }; +struct syscall_trace_enter { + struct trace_entry ent; + int nr; + unsigned long args[]; +}; + +struct syscall_trace_exit { + struct trace_entry ent; + int nr; + unsigned long ret; +}; + + /* * trace_flag_type is an enumeration that holds different * states when a trace occurs. These are: @@ -306,6 +319,10 @@ extern void __ftrace_bad_type(void); TRACE_KMEM_ALLOC); \ IF_ASSIGN(var, ent, struct kmemtrace_free_entry, \ TRACE_KMEM_FREE); \ + IF_ASSIGN(var, ent, struct syscall_trace_enter, \ + TRACE_SYSCALL_ENTER); \ + IF_ASSIGN(var, ent, struct syscall_trace_exit, \ + TRACE_SYSCALL_EXIT); \ __ftrace_bad_type(); \ } while (0) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 66cf97449af3..c72e599230ff 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -1,6 +1,5 @@ -#include #include - +#include #include #include "trace_output.h" @@ -8,6 +7,90 @@ static atomic_t refcount; +/* Our two options */ +enum { + TRACE_SYSCALLS_OPT_TYPES = 0x1, +}; + +static struct tracer_opt syscalls_opts[] = { + { TRACER_OPT(syscall_arg_type, TRACE_SYSCALLS_OPT_TYPES) }, + { } +}; + +static struct tracer_flags syscalls_flags = { + .val = 0, /* By default: no args types */ + .opts = syscalls_opts +}; + +enum print_line_t +print_syscall_enter(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *ent = iter->ent; + struct syscall_trace_enter *trace; + struct syscall_metadata *entry; + int i, ret, syscall; + + trace_assign_type(trace, ent); + + syscall = trace->nr; + + entry = syscall_nr_to_meta(syscall); + if (!entry) + goto end; + + ret = trace_seq_printf(s, "%s(", entry->name); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + for (i = 0; i < entry->nb_args; i++) { + /* parameter types */ + if (syscalls_flags.val & TRACE_SYSCALLS_OPT_TYPES) { + ret = trace_seq_printf(s, "%s ", entry->types[i]); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + /* parameter values */ + ret = trace_seq_printf(s, "%s: %lx%s ", entry->args[i], + trace->args[i], + i == entry->nb_args - 1 ? ")" : ","); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + } + +end: + trace_seq_printf(s, "\n"); + return TRACE_TYPE_HANDLED; +} + +enum print_line_t +print_syscall_exit(struct trace_iterator *iter, int flags) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *ent = iter->ent; + struct syscall_trace_exit *trace; + int syscall; + struct syscall_metadata *entry; + int ret; + + trace_assign_type(trace, ent); + + syscall = trace->nr; + + entry = syscall_nr_to_meta(syscall); + if (!entry) { + trace_seq_printf(s, "\n"); + return TRACE_TYPE_HANDLED; + } + + ret = trace_seq_printf(s, "%s -> 0x%lx\n", entry->name, + trace->ret); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + void start_ftrace_syscalls(void) { unsigned long flags; @@ -16,6 +99,7 @@ void start_ftrace_syscalls(void) if (atomic_inc_return(&refcount) != 1) goto out; + arch_init_ftrace_syscalls(); read_lock_irqsave(&tasklist_lock, flags); do_each_thread(g, t) { @@ -48,20 +132,63 @@ void stop_ftrace_syscalls(void) void ftrace_syscall_enter(struct pt_regs *regs) { + struct syscall_trace_enter *entry; + struct syscall_metadata *sys_data; + struct ring_buffer_event *event; + int size; int syscall_nr; + int cpu; syscall_nr = syscall_get_nr(current, regs); - trace_printk("syscall %d enter\n", syscall_nr); + cpu = raw_smp_processor_id(); + + sys_data = syscall_nr_to_meta(syscall_nr); + if (!sys_data) + return; + + size = sizeof(*entry) + sizeof(unsigned long) * sys_data->nb_args; + + event = trace_current_buffer_lock_reserve(TRACE_SYSCALL_ENTER, size, + 0, 0); + if (!event) + return; + + entry = ring_buffer_event_data(event); + entry->nr = syscall_nr; + syscall_get_arguments(current, regs, 0, sys_data->nb_args, entry->args); + + trace_current_buffer_unlock_commit(event, 0, 0); + trace_wake_up(); } void ftrace_syscall_exit(struct pt_regs *regs) { + struct syscall_trace_exit *entry; + struct syscall_metadata *sys_data; + struct ring_buffer_event *event; int syscall_nr; + int cpu; syscall_nr = syscall_get_nr(current, regs); - trace_printk("syscall %d exit\n", syscall_nr); + cpu = raw_smp_processor_id(); + + sys_data = syscall_nr_to_meta(syscall_nr); + if (!sys_data) + return; + + event = trace_current_buffer_lock_reserve(TRACE_SYSCALL_EXIT, + sizeof(*entry), 0, 0); + if (!event) + return; + + entry = ring_buffer_event_data(event); + entry->nr = syscall_nr; + entry->ret = syscall_get_return_value(current, regs); + + trace_current_buffer_unlock_commit(event, 0, 0); + trace_wake_up(); } static int init_syscall_tracer(struct trace_array *tr) @@ -77,17 +204,20 @@ static void reset_syscall_tracer(struct trace_array *tr) } static struct trace_event syscall_enter_event = { - .type = TRACE_SYSCALL_ENTER, + .type = TRACE_SYSCALL_ENTER, + .trace = print_syscall_enter, }; static struct trace_event syscall_exit_event = { - .type = TRACE_SYSCALL_EXIT, + .type = TRACE_SYSCALL_EXIT, + .trace = print_syscall_exit, }; static struct tracer syscall_tracer __read_mostly = { - .name = "syscall", + .name = "syscall", .init = init_syscall_tracer, - .reset = reset_syscall_tracer + .reset = reset_syscall_tracer, + .flags = &syscalls_flags, }; __init int register_ftrace_syscalls(void) From ccd50dfd92ea2c4ba9e39531ac55db53393e783e Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 13 Mar 2009 17:02:17 +0100 Subject: [PATCH 308/413] tracing/syscalls: support for syscalls tracing on x86, fix Impact: build fix kernel/built-in.o: In function `ftrace_syscall_exit': (.text+0x76667): undefined reference to `syscall_nr_to_meta' ftrace.o is built: obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o But now a CONFIG_FTRACE_SYSCALLS dependency is needed too. Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Lai Jiangshan LKML-Reference: <1236401580-5758-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/Makefile | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 339ce35648e6..84000eb931ff 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -66,7 +66,8 @@ obj-$(CONFIG_X86_MPPARSE) += mpparse.o obj-y += apic/ obj-$(CONFIG_X86_REBOOTFIXUPS) += reboot_fixups_32.o obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o -obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o +obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o +obj-$(CONFIG_FTRACE_SYSCALLS) += ftrace.o obj-$(CONFIG_KEXEC) += machine_kexec_$(BITS).o obj-$(CONFIG_KEXEC) += relocate_kernel_$(BITS).o crash.o obj-$(CONFIG_CRASH_DUMP) += crash_dump_$(BITS).o From b478b782e110fdb4135caa3062b6d687e989d994 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Fri, 13 Mar 2009 15:10:26 +0800 Subject: [PATCH 309/413] kallsyms, tracing: output more proper symbol name Impact: bugfix, output more reliable symbol lookup result Debug tools(dump_stack(), ftrace...) are like to print out symbols. But it is always print out the first aliased symbol.(Aliased symbols are symbols with the same address), and the first aliased symbol is sometime not proper. # echo function_graph > current_tracer # cat trace ...... 1) 1.923 us | select_nohz_load_balancer(); 1) + 76.692 us | } 1) | default_idle() { 1) ==========> | __irqentry_text_start() { 1) 0.000 us | native_apic_mem_write(); 1) | irq_enter() { 1) 0.000 us | idle_cpu(); 1) | tick_check_idle() { 1) 0.000 us | tick_check_oneshot_broadcast(); 1) | tick_nohz_stop_idle() { ...... It's very embarrassing, it ouputs "__irqentry_text_start()", actually, it should output "smp_apic_timer_interrupt()". (these two symbol are the same address, but "__irqentry_text_start" is deemed to the first aliased symbol by scripts/kallsyms) This patch puts symbols like "__irqentry_text_start" to the second aliased symbols. And a more proper symbol name becomes the first. Aliased symbols mostly come from linker script. The solution is guessing "is this symbol defined in linker script", the symbols defined in linker script will not become the first aliased symbol. And if symbols are found to be equal in this "linker script provided" criteria, symbols are sorted by the number of prefix underscores. Signed-off-by: Lai Jiangshan Acked-by: Sam Ravnborg Reviewed-by: Paulo Marques LKML-Reference: <49BA06E2.7080807@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- scripts/kallsyms.c | 57 ++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 57 insertions(+) diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c index ad2434b26970..6654cbed965b 100644 --- a/scripts/kallsyms.c +++ b/scripts/kallsyms.c @@ -500,6 +500,51 @@ static void optimize_token_table(void) optimize_result(); } +/* guess for "linker script provide" symbol */ +static int may_be_linker_script_provide_symbol(const struct sym_entry *se) +{ + const char *symbol = (char *)se->sym + 1; + int len = se->len - 1; + + if (len < 8) + return 0; + + if (symbol[0] != '_' || symbol[1] != '_') + return 0; + + /* __start_XXXXX */ + if (!memcmp(symbol + 2, "start_", 6)) + return 1; + + /* __stop_XXXXX */ + if (!memcmp(symbol + 2, "stop_", 5)) + return 1; + + /* __end_XXXXX */ + if (!memcmp(symbol + 2, "end_", 4)) + return 1; + + /* __XXXXX_start */ + if (!memcmp(symbol + len - 6, "_start", 6)) + return 1; + + /* __XXXXX_end */ + if (!memcmp(symbol + len - 4, "_end", 4)) + return 1; + + return 0; +} + +static int prefix_underscores_count(const char *str) +{ + const char *tail = str; + + while (*tail != '_') + tail++; + + return tail - str; +} + static int compare_symbols(const void *a, const void *b) { const struct sym_entry *sa; @@ -521,6 +566,18 @@ static int compare_symbols(const void *a, const void *b) if (wa != wb) return wa - wb; + /* sort by "linker script provide" type */ + wa = may_be_linker_script_provide_symbol(sa); + wb = may_be_linker_script_provide_symbol(sb); + if (wa != wb) + return wa - wb; + + /* sort by the number of prefix underscores */ + wa = prefix_underscores_count((const char *)sa->sym + 1); + wb = prefix_underscores_count((const char *)sb->sym + 1); + if (wa != wb) + return wa - wb; + /* sort by initial order, so that other symbols are left undisturbed */ return sa->start_pos - sb->start_pos; } From ac99c58c9e56967037382e31f865b72b10127965 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 15 Mar 2009 22:10:35 +0100 Subject: [PATCH 310/413] tracing/syscalls: fix missing release of tracing Impact: fix 'stuck' syscall tracer The syscall tracer uses a refcounter to enable several users simultaneously. But the refcounter did not behave correctly and always restored its value to 0 after calling start_syscall_tracing(). Therefore, stop_syscall_tracing() couldn't release correctly the tasks from tracing. Also the tracer forgot to reset the buffer when it is released. Drop the pointless refcount decrement on start_syscall_tracing() and reset the buffer when we release the tracer. This fixes two reported issue: - when we switch from syscall tracer to another tracer, syscall tracing continued. - incorrect use of the refcount. Reported-by: Andrew Morton Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker LKML-Reference: <1237151439-6755-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_syscalls.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index c72e599230ff..c5fc1d8880f6 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -96,8 +96,9 @@ void start_ftrace_syscalls(void) unsigned long flags; struct task_struct *g, *t; + /* Don't enable the flag on the tasks twice */ if (atomic_inc_return(&refcount) != 1) - goto out; + return; arch_init_ftrace_syscalls(); read_lock_irqsave(&tasklist_lock, flags); @@ -107,8 +108,6 @@ void start_ftrace_syscalls(void) } while_each_thread(g, t); read_unlock_irqrestore(&tasklist_lock, flags); -out: - atomic_dec(&refcount); } void stop_ftrace_syscalls(void) @@ -116,8 +115,9 @@ void stop_ftrace_syscalls(void) unsigned long flags; struct task_struct *g, *t; + /* There are perhaps still some users */ if (atomic_dec_return(&refcount)) - goto out; + return; read_lock_irqsave(&tasklist_lock, flags); @@ -126,8 +126,6 @@ void stop_ftrace_syscalls(void) } while_each_thread(g, t); read_unlock_irqrestore(&tasklist_lock, flags); -out: - atomic_inc(&refcount); } void ftrace_syscall_enter(struct pt_regs *regs) @@ -201,6 +199,7 @@ static int init_syscall_tracer(struct trace_array *tr) static void reset_syscall_tracer(struct trace_array *tr) { stop_ftrace_syscalls(); + tracing_reset_online_cpus(tr); } static struct trace_event syscall_enter_event = { From 6404434525bb9f8f2239998f30fd7c93f2efa5b3 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 15 Mar 2009 22:10:36 +0100 Subject: [PATCH 311/413] tracing/syscalls: various cleanups Impact: cleanup - Drop unused cpu variable - Fix some errors on comments Signed-off-by: Frederic Weisbecker LKML-Reference: <1237151439-6755-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_syscalls.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index c5fc1d8880f6..26f9a8679d3d 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -7,7 +7,7 @@ static atomic_t refcount; -/* Our two options */ +/* Option to display the parameters types */ enum { TRACE_SYSCALLS_OPT_TYPES = 0x1, }; @@ -18,7 +18,7 @@ static struct tracer_opt syscalls_opts[] = { }; static struct tracer_flags syscalls_flags = { - .val = 0, /* By default: no args types */ + .val = 0, /* By default: no parameters types */ .opts = syscalls_opts }; @@ -135,12 +135,9 @@ void ftrace_syscall_enter(struct pt_regs *regs) struct ring_buffer_event *event; int size; int syscall_nr; - int cpu; syscall_nr = syscall_get_nr(current, regs); - cpu = raw_smp_processor_id(); - sys_data = syscall_nr_to_meta(syscall_nr); if (!sys_data) return; @@ -166,12 +163,9 @@ void ftrace_syscall_exit(struct pt_regs *regs) struct syscall_metadata *sys_data; struct ring_buffer_event *event; int syscall_nr; - int cpu; syscall_nr = syscall_get_nr(current, regs); - cpu = raw_smp_processor_id(); - sys_data = syscall_nr_to_meta(syscall_nr); if (!sys_data) return; From 5be71b61f17b0e3bc8ad0b1a1b7b53ab7d574ebb Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 15 Mar 2009 22:10:37 +0100 Subject: [PATCH 312/413] tracing/syscalls: protect thread flag toggling from races Impact: fix syscall tracer enable/disable race The current thread flag toggling is racy as shown in the following scenario: - task A is the last user of syscall tracing, it releases the TIF_SYSCALL_FTRACE on each tasks - at the same time task B start syscall tracing. refcount == 0 so it sets up TIF_SYSCALL_FTRACE on each tasks. The effect of the mixup is unpredictable. So this fix adds a mutex on {start,stop}_syscall_tracing(). Reported-by: Andrew Morton Reported-by: Ingo Molnar LKML-Reference: <1237151439-6755-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_syscalls.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-) diff --git a/kernel/trace/trace_syscalls.c b/kernel/trace/trace_syscalls.c index 26f9a8679d3d..a2a3af29c943 100644 --- a/kernel/trace/trace_syscalls.c +++ b/kernel/trace/trace_syscalls.c @@ -5,7 +5,11 @@ #include "trace_output.h" #include "trace.h" -static atomic_t refcount; +/* Keep a counter of the syscall tracing users */ +static int refcount; + +/* Prevent from races on thread flags toggling */ +static DEFINE_MUTEX(syscall_trace_lock); /* Option to display the parameters types */ enum { @@ -96,9 +100,11 @@ void start_ftrace_syscalls(void) unsigned long flags; struct task_struct *g, *t; + mutex_lock(&syscall_trace_lock); + /* Don't enable the flag on the tasks twice */ - if (atomic_inc_return(&refcount) != 1) - return; + if (++refcount != 1) + goto unlock; arch_init_ftrace_syscalls(); read_lock_irqsave(&tasklist_lock, flags); @@ -108,6 +114,9 @@ void start_ftrace_syscalls(void) } while_each_thread(g, t); read_unlock_irqrestore(&tasklist_lock, flags); + +unlock: + mutex_unlock(&syscall_trace_lock); } void stop_ftrace_syscalls(void) @@ -115,9 +124,11 @@ void stop_ftrace_syscalls(void) unsigned long flags; struct task_struct *g, *t; + mutex_lock(&syscall_trace_lock); + /* There are perhaps still some users */ - if (atomic_dec_return(&refcount)) - return; + if (--refcount) + goto unlock; read_lock_irqsave(&tasklist_lock, flags); @@ -126,6 +137,9 @@ void stop_ftrace_syscalls(void) } while_each_thread(g, t); read_unlock_irqrestore(&tasklist_lock, flags); + +unlock: + mutex_unlock(&syscall_trace_lock); } void ftrace_syscall_enter(struct pt_regs *regs) From 0ea1c4156bf9e2eb370cc5c6fa6eb112bd844dec Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 15 Mar 2009 22:10:38 +0100 Subject: [PATCH 313/413] tracing/syscalls: select kallsysms Syscall tracing must select kallsysms. The arch code builds a table to find the syscall metadata by syscall number. It needs the syscalls names resolution from the symbol table to know which name found on the syscalls metadatas match a function pointer from the arch sys_call_table. Reported-by: Andrew Morton Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker LKML-Reference: <1237151439-6755-4-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 95a0ad191f19..b0a46f889659 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -182,6 +182,7 @@ config FTRACE_SYSCALLS bool "Trace syscalls" depends on HAVE_FTRACE_SYSCALLS select TRACING + select KALLSYMS help Basic tracer to catch the syscall entry and exit events. From 59f586db98919d7d9c43527b26c8de1cdf9ed912 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 15 Mar 2009 22:10:39 +0100 Subject: [PATCH 314/413] tracing/core: fix missing mutex unlock on tracing_set_tracer() Impact: fix possible locking imbalance In case of ring buffer resize failure, tracing_set_tracer forgot to release trace_types_lock. Fix it. Signed-off-by: Frederic Weisbecker LKML-Reference: <1237151439-6755-5-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index efe3202c0209..c0cf946d42f5 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -2494,7 +2494,7 @@ static int tracing_set_tracer(const char *buf) if (!ring_buffer_expanded) { ret = tracing_resize_ring_buffer(trace_buf_size); if (ret < 0) - return ret; + goto out; ret = 0; } From ac1d52d0b85854958c7e78c8006e39aadb6ce4b8 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 16 Mar 2009 00:32:41 +0100 Subject: [PATCH 315/413] tracing/ftrace: fix double calls to tracing_start() Impact: fix a warning during preemptirqsoff selftests When the preemptirqsoff selftest fails, we see the following warning: [ 6.050000] Testing tracer preemptirqsoff: .. no entries found .. ------------[ cut here ]------------ [ 6.060000] WARNING: at kernel/trace/trace.c:688 tracing_start+0x67/0xd3() [ 6.060000] Modules linked in: [ 6.060000] Pid: 1, comm: swapper Tainted: G [ 6.060000] Call Trace: [ 6.060000] [] warn_slowpath+0xb1/0x100 [ 6.060000] [] ? trace_preempt_on+0x35/0x4b [ 6.060000] [] ? tracing_start+0x31/0xd3 [ 6.060000] [] ? tracing_start+0x31/0xd3 [ 6.060000] [] ? __lock_acquired+0xe6/0x1f2 [ 6.060000] [] ? tracing_start+0x31/0xd3 [ 6.060000] [] tracing_start+0x67/0xd3 [ 6.060000] [] ? irqsoff_tracer_reset+0x2d/0x57 [ 6.060000] [] trace_selftest_startup_preemptirqsoff+0x1c8/0x1f1 [ 6.060000] [] register_tracer+0x12f/0x241 [ 6.060000] [] ? init_irqsoff_tracer+0x0/0x53 [ 6.060000] [] init_irqsoff_tracer+0x3b/0x53 This is because in fail case, the preemptirqsoff tracer selftest calls twice the tracing_start() function: int trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array *tr) { if (!ret && !count) { printk(KERN_CONT ".. no entries found .."); ret = -1; tracing_start(); <----- goto out; } [...] out: trace->reset(tr); tracing_start(); <------ tracing_max_latency = save_max; return ret; } Since it is well handled in the out path, we don't need the conditional one. Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker LKML-Reference: <1237159961-7447-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_selftest.c | 16 ++++++---------- 1 file changed, 6 insertions(+), 10 deletions(-) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index f907a2b29028..a2ca6f0fef9b 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -414,7 +414,7 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * ret = tracer_init(trace, tr); if (ret) { warn_failed_init_tracer(trace, ret); - goto out; + goto out_no_start; } /* reset the max latency */ @@ -432,21 +432,16 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * tracing_stop(); /* check both trace buffers */ ret = trace_test_buffer(tr, NULL); - if (ret) { - tracing_start(); + if (ret) goto out; - } ret = trace_test_buffer(&max_tr, &count); - if (ret) { - tracing_start(); + if (ret) goto out; - } if (!ret && !count) { printk(KERN_CONT ".. no entries found .."); ret = -1; - tracing_start(); goto out; } @@ -475,9 +470,10 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * goto out; } - out: - trace->reset(tr); +out: tracing_start(); +out_no_start: + trace->reset(tr); tracing_max_latency = save_max; return ret; From 2fc1dfbe17e7705c55b7a99da995fa565e26f151 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 16 Mar 2009 01:45:03 +0100 Subject: [PATCH 316/413] tracing/core: fix early free of cpumasks Impact: fix crashes when tracing cpumasks While ring-buffer allocation, the cpumasks are allocated too, including the tracing cpumask and the per-cpu file mask handler. But these cpumasks are freed accidentally just after. Fix it. Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker LKML-Reference: <1237164303-11476-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c0cf946d42f5..ae32d3b99b4b 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4125,7 +4125,8 @@ __init static int tracer_alloc_buffers(void) &trace_panic_notifier); register_die_notifier(&trace_die_notifier); - ret = 0; + + return 0; out_free_cpumask: free_cpumask_var(tracing_reader_cpumask); From 03303549b1695dc024d4a653cc16bd79f78f9750 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 16 Mar 2009 22:41:00 +0100 Subject: [PATCH 317/413] tracing/ftrace: fix the check on nopped sites Impact: fix a dynamic tracing failure Recently, the function and function graph tracers failed to use dynamic tracing after the following commit: fa9d13cf135efbd454453a53b6299976bea245a9 (ftrace: don't try to __ftrace_replace_code on !FTRACE_FL_CONVERTED rec) The patch is right except a mistake on the check for the FTRACE_FL_CONVERTED flag. The code patching is aborted in case of successfully nopped sites. What we want is the opposite: ignore the callsites that haven't been nopped. Signed-off-by: Frederic Weisbecker Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 90d5729afeff..7847806eefef 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -537,7 +537,7 @@ static void ftrace_replace_code(int enable) */ if (rec->flags & FTRACE_FL_FREE || rec->flags & FTRACE_FL_FAILED || - rec->flags & FTRACE_FL_CONVERTED) + !(rec->flags & FTRACE_FL_CONVERTED)) continue; /* ignore updates to this record's mcount site */ From 4ca530852346be239b7c19e7bec5d2b78855bebe Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 16 Mar 2009 19:20:15 -0400 Subject: [PATCH 318/413] tracing: protect reader of cmdline output Impact: fix to one cause of incorrect comm outputs in trace The spinlock only protected the creation of a comm <=> pid pair. But it was possible that a reader could look up a pid, and get the wrong comm because it had no locking. This also required changing trace_find_cmdline to copy the comm cache and not just send back a pointer to it. Signed-off-by: Steven Rostedt --- kernel/trace/blktrace.c | 23 ++++++++++++++++++----- kernel/trace/trace.c | 20 ++++++++++++-------- kernel/trace/trace.h | 2 +- kernel/trace/trace_functions_graph.c | 12 ++++++------ kernel/trace/trace_output.c | 18 ++++++++++++------ 5 files changed, 49 insertions(+), 26 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 1f32e4edf490..b171778e3863 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1027,7 +1027,9 @@ static int blk_log_action_seq(struct trace_seq *s, const struct blk_io_trace *t, static int blk_log_generic(struct trace_seq *s, const struct trace_entry *ent) { - const char *cmd = trace_find_cmdline(ent->pid); + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); if (t_sec(ent)) return trace_seq_printf(s, "%llu + %u [%s]\n", @@ -1057,19 +1059,30 @@ static int blk_log_remap(struct trace_seq *s, const struct trace_entry *ent) static int blk_log_plug(struct trace_seq *s, const struct trace_entry *ent) { - return trace_seq_printf(s, "[%s]\n", trace_find_cmdline(ent->pid)); + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); + + return trace_seq_printf(s, "[%s]\n", cmd); } static int blk_log_unplug(struct trace_seq *s, const struct trace_entry *ent) { - return trace_seq_printf(s, "[%s] %llu\n", trace_find_cmdline(ent->pid), - get_pdu_int(ent)); + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); + + return trace_seq_printf(s, "[%s] %llu\n", cmd, get_pdu_int(ent)); } static int blk_log_split(struct trace_seq *s, const struct trace_entry *ent) { + char cmd[TASK_COMM_LEN]; + + trace_find_cmdline(ent->pid, cmd); + return trace_seq_printf(s, "%llu / %llu [%s]\n", t_sector(ent), - get_pdu_int(ent), trace_find_cmdline(ent->pid)); + get_pdu_int(ent), cmd); } /* diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index efe3202c0209..2796bd2b17e4 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -770,25 +770,29 @@ static void trace_save_cmdline(struct task_struct *tsk) __raw_spin_unlock(&trace_cmdline_lock); } -char *trace_find_cmdline(int pid) +void trace_find_cmdline(int pid, char comm[]) { - char *cmdline = "<...>"; unsigned map; - if (!pid) - return ""; + if (!pid) { + strcpy(comm, ""); + return; + } - if (pid > PID_MAX_DEFAULT) - goto out; + if (pid > PID_MAX_DEFAULT) { + strcpy(comm, "<...>"); + return; + } + __raw_spin_lock(&trace_cmdline_lock); map = map_pid_to_cmdline[pid]; if (map >= SAVED_CMDLINES) goto out; - cmdline = saved_cmdlines[map]; + strcpy(comm, saved_cmdlines[map]); out: - return cmdline; + __raw_spin_unlock(&trace_cmdline_lock); } void tracing_record_cmdline(struct task_struct *tsk) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 56ce34d90b03..b0ecad8ecc34 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -547,7 +547,7 @@ struct tracer_switch_ops { }; #endif /* CONFIG_CONTEXT_SWITCH_TRACER */ -extern char *trace_find_cmdline(int pid); +extern void trace_find_cmdline(int pid, char comm[]); #ifdef CONFIG_DYNAMIC_FTRACE extern unsigned long ftrace_update_tot_cnt; diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 4c388607ed67..6004ccac2dd7 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -190,15 +190,15 @@ print_graph_cpu(struct trace_seq *s, int cpu) static enum print_line_t print_graph_proc(struct trace_seq *s, pid_t pid) { - int i; - int ret; - int len; - char comm[8]; - int spaces = 0; + char comm[TASK_COMM_LEN]; /* sign + log10(MAX_INT) + '\0' */ char pid_str[11]; + int spaces = 0; + int ret; + int len; + int i; - strncpy(comm, trace_find_cmdline(pid), 7); + trace_find_cmdline(pid, comm); comm[7] = '\0'; sprintf(pid_str, "%d", pid); diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index ea9d3b410c7a..6a4c9dea191e 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -309,9 +309,9 @@ static int lat_print_generic(struct trace_seq *s, struct trace_entry *entry, int cpu) { int hardirq, softirq; - char *comm; + char comm[TASK_COMM_LEN]; - comm = trace_find_cmdline(entry->pid); + trace_find_cmdline(entry->pid, comm); hardirq = entry->flags & TRACE_FLAG_HARDIRQ; softirq = entry->flags & TRACE_FLAG_SOFTIRQ; @@ -346,10 +346,12 @@ int trace_print_context(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; struct trace_entry *entry = iter->ent; - char *comm = trace_find_cmdline(entry->pid); unsigned long long t = ns2usecs(iter->ts); unsigned long usec_rem = do_div(t, USEC_PER_SEC); unsigned long secs = (unsigned long)t; + char comm[TASK_COMM_LEN]; + + trace_find_cmdline(entry->pid, comm); return trace_seq_printf(s, "%16s-%-5d [%03d] %5lu.%06lu: ", comm, entry->pid, iter->cpu, secs, usec_rem); @@ -372,7 +374,10 @@ int trace_print_lat_context(struct trace_iterator *iter) rel_usecs = ns2usecs(next_ts - iter->ts); if (verbose) { - char *comm = trace_find_cmdline(entry->pid); + char comm[TASK_COMM_LEN]; + + trace_find_cmdline(entry->pid, comm); + ret = trace_seq_printf(s, "%16s %5d %3d %d %08x %08lx [%08lx]" " %ld.%03ldms (+%ld.%03ldms): ", comm, entry->pid, iter->cpu, entry->flags, @@ -577,14 +582,15 @@ static enum print_line_t trace_ctxwake_print(struct trace_iterator *iter, char *delim) { struct ctx_switch_entry *field; - char *comm; + char comm[TASK_COMM_LEN]; int S, T; + trace_assign_type(field, iter->ent); T = task_state_char(field->next_state); S = task_state_char(field->prev_state); - comm = trace_find_cmdline(field->next_pid); + trace_find_cmdline(field->next_pid, comm); if (!trace_seq_printf(&iter->seq, " %5d:%3d:%c %s [%03d] %5d:%3d:%c %s\n", field->prev_pid, From 6adaad14d7d4d3ef31b4e2dc992b18b5da7c4eb3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 16 Mar 2009 21:57:17 -0400 Subject: [PATCH 319/413] tracing: stop comm recording on tracing off Impact: fix for losing comms in trace The command lines of tasks are cached at sched switch to not need to record them at every trace point. Disabling the tracing on stops the recording of traces, but does not stop the caching of command lines. When the tracing is off the cache may overflow and cause the tracing to show incorrect tasks matching the PIDs. This patch disables prevents updates to the comm cache when the ring buffer is off. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 2796bd2b17e4..8f89690230e6 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -797,7 +797,7 @@ void trace_find_cmdline(int pid, char comm[]) void tracing_record_cmdline(struct task_struct *tsk) { - if (atomic_read(&trace_record_cmdline_disabled)) + if (atomic_read(&trace_record_cmdline_disabled) || !tracing_is_on()) return; trace_save_cmdline(tsk); From c269fc8c537d761f36cb98e637ae934d9331a9d5 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 17 Mar 2009 01:20:59 -0500 Subject: [PATCH 320/413] tracing: fix leak in event_format_read() Impact: fix memory leak If event_format_read() exits early due to nonzero ppos, the previous kmalloc doesn't get freed - might as well do the check before the kmalloc and avoid the problem. Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237270859.8033.141.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 238ea95a4115..c88227b3b9db 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -378,15 +378,15 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, char *buf; int r; + if (*ppos) + return 0; + s = kmalloc(sizeof(*s), GFP_KERNEL); if (!s) return -ENOMEM; trace_seq_init(s); - if (*ppos) - return 0; - /* If any of the first writes fail, so will the show_format. */ trace_seq_printf(s, "name: %s\n", call->name); From 37886f6a9f62d22530ffee8d3f9215c8345b6969 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Mar 2009 17:22:06 -0400 Subject: [PATCH 321/413] ring-buffer: add api to allow a tracer to change clock source This patch adds a new function called ring_buffer_set_clock that allows a tracer to assign its own clock source to the buffer. Signed-off-by: Steven Rostedt --- include/linux/ring_buffer.h | 7 ++-- kernel/trace/ring_buffer.c | 65 ++++++++++++++++++++++--------------- kernel/trace/trace.c | 21 ++++++++---- 3 files changed, 57 insertions(+), 36 deletions(-) diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index b1a0068a5557..9e6052bd1a1c 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -118,8 +118,11 @@ unsigned long ring_buffer_overruns(struct ring_buffer *buffer); unsigned long ring_buffer_entries_cpu(struct ring_buffer *buffer, int cpu); unsigned long ring_buffer_overrun_cpu(struct ring_buffer *buffer, int cpu); -u64 ring_buffer_time_stamp(int cpu); -void ring_buffer_normalize_time_stamp(int cpu, u64 *ts); +u64 ring_buffer_time_stamp(struct ring_buffer *buffer, int cpu); +void ring_buffer_normalize_time_stamp(struct ring_buffer *buffer, + int cpu, u64 *ts); +void ring_buffer_set_clock(struct ring_buffer *buffer, + u64 (*clock)(void)); size_t ring_buffer_page_len(void *page); diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 58128ad2fde0..bbf51922a8ca 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -180,29 +180,6 @@ EXPORT_SYMBOL_GPL(tracing_is_on); #include "trace.h" -/* Up this if you want to test the TIME_EXTENTS and normalization */ -#define DEBUG_SHIFT 0 - -u64 ring_buffer_time_stamp(int cpu) -{ - u64 time; - - preempt_disable_notrace(); - /* shift to debug/test normalization and TIME_EXTENTS */ - time = trace_clock_local() << DEBUG_SHIFT; - preempt_enable_no_resched_notrace(); - - return time; -} -EXPORT_SYMBOL_GPL(ring_buffer_time_stamp); - -void ring_buffer_normalize_time_stamp(int cpu, u64 *ts) -{ - /* Just stupid testing the normalize function and deltas */ - *ts >>= DEBUG_SHIFT; -} -EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp); - #define RB_EVNT_HDR_SIZE (offsetof(struct ring_buffer_event, array)) #define RB_ALIGNMENT 4U #define RB_MAX_SMALL_DATA 28 @@ -374,6 +351,7 @@ struct ring_buffer { #ifdef CONFIG_HOTPLUG_CPU struct notifier_block cpu_notify; #endif + u64 (*clock)(void); }; struct ring_buffer_iter { @@ -394,6 +372,30 @@ struct ring_buffer_iter { _____ret; \ }) +/* Up this if you want to test the TIME_EXTENTS and normalization */ +#define DEBUG_SHIFT 0 + +u64 ring_buffer_time_stamp(struct ring_buffer *buffer, int cpu) +{ + u64 time; + + preempt_disable_notrace(); + /* shift to debug/test normalization and TIME_EXTENTS */ + time = buffer->clock() << DEBUG_SHIFT; + preempt_enable_no_resched_notrace(); + + return time; +} +EXPORT_SYMBOL_GPL(ring_buffer_time_stamp); + +void ring_buffer_normalize_time_stamp(struct ring_buffer *buffer, + int cpu, u64 *ts) +{ + /* Just stupid testing the normalize function and deltas */ + *ts >>= DEBUG_SHIFT; +} +EXPORT_SYMBOL_GPL(ring_buffer_normalize_time_stamp); + /** * check_pages - integrity check of buffer pages * @cpu_buffer: CPU buffer with pages to test @@ -569,6 +571,7 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) buffer->pages = DIV_ROUND_UP(size, BUF_PAGE_SIZE); buffer->flags = flags; + buffer->clock = trace_clock_local; /* need at least two pages */ if (buffer->pages == 1) @@ -645,6 +648,12 @@ ring_buffer_free(struct ring_buffer *buffer) } EXPORT_SYMBOL_GPL(ring_buffer_free); +void ring_buffer_set_clock(struct ring_buffer *buffer, + u64 (*clock)(void)) +{ + buffer->clock = clock; +} + static void rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer); static void @@ -1191,7 +1200,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, cpu_buffer->tail_page = next_page; /* reread the time stamp */ - *ts = ring_buffer_time_stamp(cpu_buffer->cpu); + *ts = ring_buffer_time_stamp(buffer, cpu_buffer->cpu); cpu_buffer->tail_page->page->time_stamp = *ts; } @@ -1334,7 +1343,7 @@ rb_reserve_next_event(struct ring_buffer_per_cpu *cpu_buffer, if (RB_WARN_ON(cpu_buffer, ++nr_loops > 1000)) return NULL; - ts = ring_buffer_time_stamp(cpu_buffer->cpu); + ts = ring_buffer_time_stamp(cpu_buffer->buffer, cpu_buffer->cpu); /* * Only the first commit can update the timestamp. @@ -2051,7 +2060,8 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) case RINGBUF_TYPE_DATA: if (ts) { *ts = cpu_buffer->read_stamp + event->time_delta; - ring_buffer_normalize_time_stamp(cpu_buffer->cpu, ts); + ring_buffer_normalize_time_stamp(buffer, + cpu_buffer->cpu, ts); } return event; @@ -2112,7 +2122,8 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) case RINGBUF_TYPE_DATA: if (ts) { *ts = iter->read_stamp + event->time_delta; - ring_buffer_normalize_time_stamp(cpu_buffer->cpu, ts); + ring_buffer_normalize_time_stamp(buffer, + cpu_buffer->cpu, ts); } return event; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8f89690230e6..3be2f788e10d 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -155,13 +155,6 @@ ns2usecs(cycle_t nsec) return nsec; } -cycle_t ftrace_now(int cpu) -{ - u64 ts = ring_buffer_time_stamp(cpu); - ring_buffer_normalize_time_stamp(cpu, &ts); - return ts; -} - /* * The global_trace is the descriptor that holds the tracing * buffers for the live tracing. For each CPU, it contains @@ -178,6 +171,20 @@ static struct trace_array global_trace; static DEFINE_PER_CPU(struct trace_array_cpu, global_trace_cpu); +cycle_t ftrace_now(int cpu) +{ + u64 ts; + + /* Early boot up does not have a buffer yet */ + if (!global_trace.buffer) + return trace_clock_local(); + + ts = ring_buffer_time_stamp(global_trace.buffer, cpu); + ring_buffer_normalize_time_stamp(global_trace.buffer, cpu, &ts); + + return ts; +} + /* * The max_tr is used to snapshot the global_trace when a maximum * latency is reached. Some tracers will use this to store a maximum From af4617bdba34aa556272b34c3986b0a4d588f568 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Mar 2009 18:09:55 -0400 Subject: [PATCH 322/413] tracing: add global-clock option to provide cross CPU clock to traces Impact: feature to allow better serialized clock This patch adds an option called "global-clock" that will allow the tracer to switch to a slower but more accurate (across CPUs) clock. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 34 ++++++++++++++++++++++++++++++---- kernel/trace/trace.h | 1 + 2 files changed, 31 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 3be2f788e10d..2f994caab0b7 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -315,6 +315,7 @@ static const char *trace_options[] = { "printk-msg-only", "context-info", "latency-format", + "global-clock", NULL }; @@ -2251,6 +2252,34 @@ static int set_tracer_option(struct tracer *trace, char *cmp, int neg) return 0; } +static void set_tracer_flags(unsigned int mask, int enabled) +{ + /* do nothing if flag is already set */ + if (!!(trace_flags & mask) == !!enabled) + return; + + if (enabled) + trace_flags |= mask; + else + trace_flags &= ~mask; + + if (mask == TRACE_ITER_GLOBAL_CLK) { + u64 (*func)(void); + + if (enabled) + func = trace_clock_global; + else + func = trace_clock_local; + + mutex_lock(&trace_types_lock); + ring_buffer_set_clock(global_trace.buffer, func); + + if (max_tr.buffer) + ring_buffer_set_clock(max_tr.buffer, func); + mutex_unlock(&trace_types_lock); + } +} + static ssize_t tracing_trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt, loff_t *ppos) @@ -2278,10 +2307,7 @@ tracing_trace_options_write(struct file *filp, const char __user *ubuf, int len = strlen(trace_options[i]); if (strncmp(cmp, trace_options[i], len) == 0) { - if (neg) - trace_flags &= ~(1 << i); - else - trace_flags |= (1 << i); + set_tracer_flags(1 << i, !neg); break; } } diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index b0ecad8ecc34..26a7a28ca110 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -667,6 +667,7 @@ enum trace_iterator_flags { TRACE_ITER_PRINTK_MSGONLY = 0x10000, TRACE_ITER_CONTEXT_INFO = 0x20000, /* Print pid/cpu/time */ TRACE_ITER_LATENCY_FMT = 0x40000, + TRACE_ITER_GLOBAL_CLK = 0x80000, }; /* From 97e7e4f391cac2b00417b581b432533d245d4fd0 Mon Sep 17 00:00:00 2001 From: Witold Baryluk Date: Tue, 17 Mar 2009 21:15:44 +0100 Subject: [PATCH 323/413] tracing: optimization of branch tracer Impact: better performance for if branch tracer Use an array to count the hit and misses of a conditional instead of using another conditional. This cuts down on saturation of branch predictions and increases performance of modern pipelined architectures. Signed-off-by: Witold Baryluk Signed-off-by: Steven Rostedt --- include/linux/compiler.h | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/include/linux/compiler.h b/include/linux/compiler.h index d95da1020f1c..6faa7e549de4 100644 --- a/include/linux/compiler.h +++ b/include/linux/compiler.h @@ -68,6 +68,7 @@ struct ftrace_branch_data { unsigned long miss; unsigned long hit; }; + unsigned long miss_hit[2]; }; }; @@ -125,10 +126,7 @@ void ftrace_likely_update(struct ftrace_branch_data *f, int val, int expect); .line = __LINE__, \ }; \ ______r = !!(cond); \ - if (______r) \ - ______f.hit++; \ - else \ - ______f.miss++; \ + ______f.miss_hit[______r]++; \ ______r; \ })) #endif /* CONFIG_PROFILE_ALL_BRANCHES */ From 5fec6ddcb43a91aa9a254c8ecf174c803de6f07e Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Mar 2009 19:59:53 -0400 Subject: [PATCH 324/413] tracing: make sched_switch stop/start light weight The stopping and starting of a tracer should be light weight and be able to be called in all contexts. The sched_switch grabbed mutexes in the start/stop functions. This patch changes it to a simple variable, on/off. Signed-off-by: Steven Rostedt --- kernel/trace/trace_sched_switch.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace_sched_switch.c b/kernel/trace/trace_sched_switch.c index 77132c2cf3d9..de35f200abd3 100644 --- a/kernel/trace/trace_sched_switch.c +++ b/kernel/trace/trace_sched_switch.c @@ -18,6 +18,7 @@ static struct trace_array *ctx_trace; static int __read_mostly tracer_enabled; static int sched_ref; static DEFINE_MUTEX(sched_register_mutex); +static int sched_stopped; static void probe_sched_switch(struct rq *__rq, struct task_struct *prev, @@ -28,7 +29,7 @@ probe_sched_switch(struct rq *__rq, struct task_struct *prev, int cpu; int pc; - if (!sched_ref) + if (!sched_ref || sched_stopped) return; tracing_record_cmdline(prev); @@ -193,6 +194,7 @@ static void stop_sched_trace(struct trace_array *tr) static int sched_switch_trace_init(struct trace_array *tr) { ctx_trace = tr; + tracing_reset_online_cpus(tr); tracing_start_sched_switch_record(); return 0; } @@ -205,13 +207,12 @@ static void sched_switch_trace_reset(struct trace_array *tr) static void sched_switch_trace_start(struct trace_array *tr) { - tracing_reset_online_cpus(tr); - tracing_start_sched_switch(); + sched_stopped = 0; } static void sched_switch_trace_stop(struct trace_array *tr) { - tracing_stop_sched_switch(); + sched_stopped = 1; } static struct tracer sched_switch_trace __read_mostly = From 62524d55e5b9ffe36e3bf3dd7a594114f150b449 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 17 Mar 2009 20:58:00 -0400 Subject: [PATCH 325/413] tracing: make power tracer start/stop methods lighter weight The start/stop methods of a tracer should be able to be executed in all contexts. This patch converts the power tracer to do so. Signed-off-by: Steven Rostedt --- kernel/trace/trace_power.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_power.c b/kernel/trace/trace_power.c index 91ce672fb037..bae791ebcc51 100644 --- a/kernel/trace/trace_power.c +++ b/kernel/trace/trace_power.c @@ -122,10 +122,14 @@ static int tracing_power_register(void) static void start_power_trace(struct trace_array *tr) { trace_power_enabled = 1; - tracing_power_register(); } static void stop_power_trace(struct trace_array *tr) +{ + trace_power_enabled = 0; +} + +static void power_trace_reset(struct trace_array *tr) { trace_power_enabled = 0; unregister_trace_power_start(probe_power_start); @@ -188,7 +192,7 @@ static struct tracer power_tracer __read_mostly = .init = power_trace_init, .start = start_power_trace, .stop = stop_power_trace, - .reset = stop_power_trace, + .reset = power_trace_reset, .print_line = power_print_line, }; From 18aecd362a1c991fbf5f7919ae051a77532ba2f8 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 18 Mar 2009 08:56:58 +0100 Subject: [PATCH 326/413] tracing: stop command line recording when tracing is disabled Impact: prevent overwrite of command line entries When the tracer is stopped the command line recording continues to record. The check for tracing_is_on() is not sufficient here as the ringbuffer status is not affected by setting debug/tracing/tracing_enabled to 0. On a non idle system this can result in the loss of the command line information for the stopped trace, which makes the trace harder to read and analyse. Check tracer_enabled to allow further recording. Signed-off-by: Thomas Gleixner Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 1ce6208fd727..7b6043ea256e 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -797,7 +797,8 @@ void trace_find_cmdline(int pid, char comm[]) void tracing_record_cmdline(struct task_struct *tsk) { - if (atomic_read(&trace_record_cmdline_disabled) || !tracing_is_on()) + if (atomic_read(&trace_record_cmdline_disabled) || !tracer_enabled || + !tracing_is_on()) return; trace_save_cmdline(tsk); From 2c7eea4c62ba090b7f4583c3d7337ea0019be900 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 18 Mar 2009 09:03:19 +0100 Subject: [PATCH 327/413] tracing: replace the crude (unsigned) -1 hackery Impact: cleanup The command line recorder uses (unsigned) -1 to mark non mapped entries in the pid to command line maps. The validity check is completely unintuitive: idx >= SAVED_CMDLINES There is no need for such casting games. Use a constant to mark unmapped entries and check for that constant to make the code readable and understandable. Signed-off-by: Thomas Gleixner Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 7b6043ea256e..ca673c475687 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -633,6 +633,7 @@ void tracing_reset_online_cpus(struct trace_array *tr) } #define SAVED_CMDLINES 128 +#define NO_CMDLINE_MAP UINT_MAX static unsigned map_pid_to_cmdline[PID_MAX_DEFAULT+1]; static unsigned map_cmdline_to_pid[SAVED_CMDLINES]; static char saved_cmdlines[SAVED_CMDLINES][TASK_COMM_LEN]; @@ -644,8 +645,8 @@ static atomic_t trace_record_cmdline_disabled __read_mostly; static void trace_init_cmdlines(void) { - memset(&map_pid_to_cmdline, -1, sizeof(map_pid_to_cmdline)); - memset(&map_cmdline_to_pid, -1, sizeof(map_cmdline_to_pid)); + memset(&map_pid_to_cmdline, NO_CMDLINE_MAP, sizeof(map_pid_to_cmdline)); + memset(&map_cmdline_to_pid, NO_CMDLINE_MAP, sizeof(map_cmdline_to_pid)); cmdline_idx = 0; } @@ -753,12 +754,12 @@ static void trace_save_cmdline(struct task_struct *tsk) return; idx = map_pid_to_cmdline[tsk->pid]; - if (idx >= SAVED_CMDLINES) { + if (idx == NO_CMDLINE_MAP) { idx = (cmdline_idx + 1) % SAVED_CMDLINES; map = map_cmdline_to_pid[idx]; - if (map <= PID_MAX_DEFAULT) - map_pid_to_cmdline[map] = (unsigned)-1; + if (map != NO_CMDLINE_MAP) + map_pid_to_cmdline[map] = NO_CMDLINE_MAP; map_pid_to_cmdline[tsk->pid] = idx; @@ -786,7 +787,7 @@ void trace_find_cmdline(int pid, char comm[]) __raw_spin_lock(&trace_cmdline_lock); map = map_pid_to_cmdline[pid]; - if (map >= SAVED_CMDLINES) + if (map == NO_CMDLINE_MAP) goto out; strcpy(comm, saved_cmdlines[map]); From 50d88758a3f9787cbdbdbc030560b815721eab4b Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 18 Mar 2009 08:58:44 +0100 Subject: [PATCH 328/413] tracing: fix trace_find_cmdline() Impact: prevent stale command line output In case there is no valid command line mapping for a pid trace_find_cmdline() returns without updating the comm buffer. The trace dump keeps the previous entry which results in confusing trace output: -0 [000] 280.702056 .... -23456 [000] 280.702080 .... Update the comm buffer with "<...>" when no mapping is found. Signed-off-by: Thomas Gleixner Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index ca673c475687..06c69a260328 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -787,12 +787,11 @@ void trace_find_cmdline(int pid, char comm[]) __raw_spin_lock(&trace_cmdline_lock); map = map_pid_to_cmdline[pid]; - if (map == NO_CMDLINE_MAP) - goto out; + if (map != NO_CMDLINE_MAP) + strcpy(comm, saved_cmdlines[map]); + else + strcpy(comm, "<...>"); - strcpy(comm, saved_cmdlines[map]); - - out: __raw_spin_unlock(&trace_cmdline_lock); } From a635cf0497342978d417cae19d4a4823932977ff Mon Sep 17 00:00:00 2001 From: Carsten Emde Date: Wed, 18 Mar 2009 09:00:41 +0100 Subject: [PATCH 329/413] tracing: fix command line to pid reverse map Impact: fix command line to pid mapping map_cmdline_to_pid[] is checked in trace_save_cmdline(), but never updated. This results in stale pid to command line mappings and the tracer output will associate the wrong comm string. Signed-off-by: Carsten Emde Signed-off-by: Thomas Gleixner Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 06c69a260328..305c562dae2a 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -738,8 +738,7 @@ void trace_stop_cmdline_recording(void); static void trace_save_cmdline(struct task_struct *tsk) { - unsigned map; - unsigned idx; + unsigned pid, idx; if (!tsk->pid || unlikely(tsk->pid > PID_MAX_DEFAULT)) return; @@ -757,10 +756,17 @@ static void trace_save_cmdline(struct task_struct *tsk) if (idx == NO_CMDLINE_MAP) { idx = (cmdline_idx + 1) % SAVED_CMDLINES; - map = map_cmdline_to_pid[idx]; - if (map != NO_CMDLINE_MAP) - map_pid_to_cmdline[map] = NO_CMDLINE_MAP; + /* + * Check whether the cmdline buffer at idx has a pid + * mapped. We are going to overwrite that entry so we + * need to clear the map_pid_to_cmdline. Otherwise we + * would read the new comm for the old pid. + */ + pid = map_cmdline_to_pid[idx]; + if (pid != NO_CMDLINE_MAP) + map_pid_to_cmdline[pid] = NO_CMDLINE_MAP; + map_cmdline_to_pid[idx] = tsk->pid; map_pid_to_cmdline[tsk->pid] = idx; cmdline_idx = idx; From 490362003457f8d387f6f6e73e3a7efbf56c3314 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 17 Mar 2009 22:38:58 +0100 Subject: [PATCH 330/413] tracing/ftrace: stop {irqs, preempt}soff tracers when tracing is stopped Impact: fix a selftest warning In some cases, it's possible to see the following warning on irqsoff tracer selftest: [ 4.640003] Testing tracer irqsoff: <4>------------[ cut here ]------------ [ 4.653562] WARNING: at kernel/trace/trace.c:458 update_max_tr_single+0x9a/0xc4() [ 4.660000] Hardware name: System Product Name [ 4.660000] Modules linked in: [ 4.660000] Pid: 301, comm: kstop/1 Not tainted 2.6.29-rc8-tip #35837 [ 4.660000] Call Trace: [ 4.660000] [<4014b588>] warn_slowpath+0x79/0x8f [ 4.660000] [<402d6949>] ? put_dec+0x64/0x6b [ 4.660000] [<40162b56>] ? getnstimeofday+0x58/0xdd [ 4.660000] [<40162210>] ? clocksource_read+0x3/0xf [ 4.660000] [<4015eb44>] ? ktime_set+0x8/0x34 [ 4.660000] [<4014101a>] ? balance_runtime+0x8/0x56 [ 4.660000] [<405f6f11>] ? _spin_lock+0x3/0x10 [ 4.660000] [<4011f643>] ? ftrace_call+0x5/0x8 [ 4.660000] [<4015d0f1>] ? task_cputime_zero+0x3/0x27 [ 4.660000] [<40190ee7>] ? cpupri_set+0x90/0xcb [ 4.660000] [<405f7208>] ? _spin_lock_irqsave+0x22/0x34 [ 4.660000] [<40190f12>] ? cpupri_set+0xbb/0xcb [ 4.660000] [<405f7151>] ? _spin_unlock_irqrestore+0x23/0x35 [ 4.660000] [<4018493f>] ? ring_buffer_reset_cpu+0x27/0x51 [ 4.660000] [<405f7208>] ? _spin_lock_irqsave+0x22/0x34 [ 4.660000] [<40184962>] ? ring_buffer_reset_cpu+0x4a/0x51 [ 4.660000] [<405f7151>] ? _spin_unlock_irqrestore+0x23/0x35 [ 4.660000] [<4018cc29>] ? trace_hardirqs_off+0x1a/0x1c [ 4.660000] [<405f7151>] ? _spin_unlock_irqrestore+0x23/0x35 [ 4.660000] [<40184962>] ? ring_buffer_reset_cpu+0x4a/0x51 [ 4.660000] [<401850f3>] ? cpumask_next+0x15/0x18 [ 4.660000] [<4018a41f>] update_max_tr_single+0x9a/0xc4 [ 4.660000] [<4014e5fe>] ? exit_notify+0x16/0xf2 [ 4.660000] [<4018cd13>] check_critical_timing+0xcc/0x11e [ 4.660000] [<4014e5fe>] ? exit_notify+0x16/0xf2 [ 4.660000] [<4014e5fe>] ? exit_notify+0x16/0xf2 [ 4.660000] [<4018cdf1>] stop_critical_timing+0x8c/0x9f [ 4.660000] [<4014e5c4>] ? forget_original_parent+0xac/0xd0 [ 4.660000] [<4018ce3a>] trace_hardirqs_on+0x1a/0x1c [ 4.660000] [<4014e5c4>] forget_original_parent+0xac/0xd0 [ 4.660000] [<4014e5fe>] exit_notify+0x16/0xf2 [ 4.660000] [<4014e8a5>] do_exit+0x1cb/0x225 [ 4.660000] [<4015c72b>] ? kthread+0x0/0x69 [ 4.660000] [<4011f61d>] kernel_thread_helper+0xd/0x10 [ 4.660000] ---[ end trace a7919e7f17c0a725 ]--- [ 4.660164] .. no entries found ..FAILED! During the selftest of irqsoff tracer, we do that: /* disable interrupts for a bit */ local_irq_disable(); udelay(100); local_irq_enable(); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ ret = trace_test_buffer(tr, NULL); If a callsite performs a new max delay with irqs off just after tracing_stop, update_max_tr_single() -> ring_buffer_swap_cpu() will be called with the buffers disabled by tracing_stop(), hence the warning, then ring_buffer_swap_cpu() return -EAGAIN and update_max_tr_single() complains. Fix it by also stopping the tracer before stopping the tracing globally. A similar situation can happen with preemptoff and preemptirqsoff tracers where we apply the same fix. Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Acked-by: Steven Rostedt LKML-Reference: <1237325938-5240-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_selftest.c | 26 ++++++++++++++++++++++++++ 1 file changed, 26 insertions(+) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index a2ca6f0fef9b..38856ba78a92 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -315,6 +315,14 @@ trace_selftest_startup_irqsoff(struct tracer *trace, struct trace_array *tr) local_irq_disable(); udelay(100); local_irq_enable(); + + /* + * Stop the tracer to avoid a warning subsequent + * to buffer flipping failure because tracing_stop() + * disables the tr and max buffers, making flipping impossible + * in case of parallels max irqs off latencies. + */ + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ @@ -369,6 +377,14 @@ trace_selftest_startup_preemptoff(struct tracer *trace, struct trace_array *tr) preempt_disable(); udelay(100); preempt_enable(); + + /* + * Stop the tracer to avoid a warning subsequent + * to buffer flipping failure because tracing_stop() + * disables the tr and max buffers, making flipping impossible + * in case of parallels max preempt off latencies. + */ + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ @@ -428,6 +444,13 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * /* reverse the order of preempt vs irqs */ local_irq_enable(); + /* + * Stop the tracer to avoid a warning subsequent + * to buffer flipping failure because tracing_stop() + * disables the tr and max buffers, making flipping impossible + * in case of parallels max irqs/preempt off latencies. + */ + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ @@ -448,6 +471,8 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * /* do the test by disabling interrupts first this time */ tracing_max_latency = 0; tracing_start(); + trace->start(tr); + preempt_disable(); local_irq_disable(); udelay(100); @@ -455,6 +480,7 @@ trace_selftest_startup_preemptirqsoff(struct tracer *trace, struct trace_array * /* reverse the order of preempt vs irqs */ local_irq_enable(); + trace->stop(tr); /* stop the tracing. */ tracing_stop(); /* check both trace buffers */ From f02b8624fedca39886b0eef770dca70c2f0749b3 Mon Sep 17 00:00:00 2001 From: Ananth N Mavinakayanahalli Date: Wed, 18 Mar 2009 17:06:21 +0530 Subject: [PATCH 331/413] kprobes: Fix locking imbalance in kretprobes Fix locking imbalance in kretprobes: ===================================== [ BUG: bad unlock balance detected! ] ------------------------------------- kthreadd/2 is trying to release lock (&rp->lock) at: [] pre_handler_kretprobe+0xea/0xf4 but there are no more locks to release! other info that might help us debug this: 1 lock held by kthreadd/2: #0: (rcu_read_lock){..--}, at: [] __atomic_notifier_call_chain+0x0/0x5a stack backtrace: Pid: 2, comm: kthreadd Not tainted 2.6.29-rc8 #1 Call Trace: [] ? printk+0xf/0x17 [] ? pre_handler_kretprobe+0xea/0xf4 [] print_unlock_inbalance_bug+0xc3/0xce [] ? clocksource_read+0x7/0xa [] ? getnstimeofday+0x5f/0xf6 [] ? register_lock_class+0x17/0x293 [] ? mark_lock+0x1e/0x30b [] ? tick_dev_program_event+0x4a/0xbc [] ? __slab_alloc+0xa5/0x415 [] ? pre_handler_kretprobe+0x28/0xf4 [] ? pre_handler_kretprobe+0xea/0xf4 [] lock_release_non_nested+0xa4/0x1a5 [] ? pre_handler_kretprobe+0xea/0xf4 [] lock_release+0x141/0x166 [] _spin_unlock_irqrestore+0x19/0x50 [] pre_handler_kretprobe+0xea/0xf4 [] kprobe_exceptions_notify+0x1c9/0x43e [] notifier_call_chain+0x26/0x48 [] __atomic_notifier_call_chain+0x37/0x5a [] ? __atomic_notifier_call_chain+0x0/0x5a [] atomic_notifier_call_chain+0xc/0xe [] notify_die+0x2d/0x2f [] do_int3+0x1f/0x71 [] int3+0x2c/0x34 [] ? do_fork+0x1/0x288 [] ? kernel_thread+0x71/0x79 [] ? kthread+0x0/0x60 [] ? kthread+0x0/0x60 [] ? kernel_thread_helper+0x0/0x10 [] kthreadd+0xac/0x148 [] ? kthreadd+0x0/0x148 [] kernel_thread_helper+0x7/0x10 Signed-off-by: Ananth N Mavinakayanahalli Tested-by: Bharata B Rao Cc: Masami Hiramatsu Cc: Jim Keniston Cc: Linus Torvalds Cc: Andrew Morton Cc: [2.6.29.x, 2.6.28.x, 2.6.27.x] LKML-Reference: <20090318113621.GB4129@in.ibm.com> Signed-off-by: Ingo Molnar --- kernel/kprobes.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/kernel/kprobes.c b/kernel/kprobes.c index 479d4d5672f9..5016bfb682b9 100644 --- a/kernel/kprobes.c +++ b/kernel/kprobes.c @@ -919,10 +919,8 @@ static int __kprobes pre_handler_kretprobe(struct kprobe *p, ri->rp = rp; ri->task = current; - if (rp->entry_handler && rp->entry_handler(ri, regs)) { - spin_unlock_irqrestore(&rp->lock, flags); + if (rp->entry_handler && rp->entry_handler(ri, regs)) return 0; - } arch_prepare_kretprobe(ri, regs); From 4acd4d00f716873e27e7b60ae292cbdbfae674dd Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 18 Mar 2009 10:40:24 -0400 Subject: [PATCH 332/413] tracing: give easy way to clear trace buffer There is currently no easy way to clear the trace buffer. Currently the only way is to change the current tracer. This patch lets the user clear the trace buffer by simply writing into the trace files. echo > /debug/tracing/trace or to clear a single cpu (i.e. for CPU 1): echo > /debug/tracing/per_cpu/cpu1/trace Requested-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 41 +++++++++++++++++++++++++++++++++-------- 1 file changed, 33 insertions(+), 8 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index a2d13e8c8fd8..8d981ababc45 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1941,9 +1941,14 @@ int tracing_open_generic(struct inode *inode, struct file *filp) static int tracing_release(struct inode *inode, struct file *file) { struct seq_file *m = (struct seq_file *)file->private_data; - struct trace_iterator *iter = m->private; + struct trace_iterator *iter; int cpu; + if (!(file->f_mode & FMODE_READ)) + return 0; + + iter = m->private; + mutex_lock(&trace_types_lock); for_each_tracing_cpu(cpu) { if (iter->buffer_iter[cpu]) @@ -1969,12 +1974,24 @@ static int tracing_open(struct inode *inode, struct file *file) struct trace_iterator *iter; int ret = 0; - iter = __tracing_open(inode, file); - if (IS_ERR(iter)) - ret = PTR_ERR(iter); - else if (trace_flags & TRACE_ITER_LATENCY_FMT) - iter->iter_flags |= TRACE_FILE_LAT_FMT; + /* If this file was open for write, then erase contents */ + if ((file->f_mode & FMODE_WRITE) && + !(file->f_flags & O_APPEND)) { + long cpu = (long) inode->i_private; + if (cpu == TRACE_PIPE_ALL_CPU) + tracing_reset_online_cpus(&global_trace); + else + tracing_reset(&global_trace, cpu); + } + + if (file->f_mode & FMODE_READ) { + iter = __tracing_open(inode, file); + if (IS_ERR(iter)) + ret = PTR_ERR(iter); + else if (trace_flags & TRACE_ITER_LATENCY_FMT) + iter->iter_flags |= TRACE_FILE_LAT_FMT; + } return ret; } @@ -2049,9 +2066,17 @@ static int show_traces_open(struct inode *inode, struct file *file) return ret; } +static ssize_t +tracing_write_stub(struct file *filp, const char __user *ubuf, + size_t count, loff_t *ppos) +{ + return count; +} + static const struct file_operations tracing_fops = { .open = tracing_open, .read = seq_read, + .write = tracing_write_stub, .llseek = seq_lseek, .release = tracing_release, }; @@ -3576,7 +3601,7 @@ static void tracing_init_debugfs_percpu(long cpu) pr_warning("Could not create debugfs 'trace_pipe' entry\n"); /* per cpu trace */ - entry = debugfs_create_file("trace", 0444, d_cpu, + entry = debugfs_create_file("trace", 0644, d_cpu, (void *) cpu, &tracing_fops); if (!entry) pr_warning("Could not create debugfs 'trace' entry\n"); @@ -3890,7 +3915,7 @@ static __init int tracer_init_debugfs(void) if (!entry) pr_warning("Could not create debugfs 'tracing_cpumask' entry\n"); - entry = debugfs_create_file("trace", 0444, d_tracer, + entry = debugfs_create_file("trace", 0644, d_tracer, (void *) TRACE_PIPE_ALL_CPU, &tracing_fops); if (!entry) pr_warning("Could not create debugfs 'trace' entry\n"); From 09933a108e6730a464a1ab676c9decc11aee0edc Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Wed, 18 Mar 2009 22:18:56 +0530 Subject: [PATCH 333/413] tracing: fix oops in tracepoint_update_probe_range() Change this crash: BUG: unable to handle kernel NULL pointer dereference at (null) IP: [] tracepoint_update_probe_range+0x1f/0x9b PGD 13d5fb067 PUD 13d688067 PMD 0 Oops: 0000 [#1] SMP To a more debuggable WARN_ONCE(). Signed-off-by: Jaswinder Singh Rajput Cc: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1237394936.3132.1.camel@localhost.localdomain> [ moved the check outside the lock and added a WARN_ON(). ] Signed-off-by: Ingo Molnar --- kernel/tracepoint.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c index 79602740bbb5..adf28734fe2d 100644 --- a/kernel/tracepoint.c +++ b/kernel/tracepoint.c @@ -272,12 +272,17 @@ static void disable_tracepoint(struct tracepoint *elem) * * Updates the probe callback corresponding to a range of tracepoints. */ -void tracepoint_update_probe_range(struct tracepoint *begin, - struct tracepoint *end) +void +tracepoint_update_probe_range(struct tracepoint *begin, struct tracepoint *end) { struct tracepoint *iter; struct tracepoint_entry *mark_entry; + if (!begin) { + WARN_ON_ONCE(1); + return; + } + mutex_lock(&tracepoints_mutex); for (iter = begin; iter < end; iter++) { mark_entry = get_tracepoint(iter->name); From ec625cb29e66824f7ce41082617aeb93fa4e42e2 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Wed, 18 Mar 2009 19:54:04 +0100 Subject: [PATCH 334/413] tracepoints: dont update zero-sized tracepoint sections Zero-sized tracepoint sections can occur if tracing is enabled but no tracepoint is defined. Do not emit a warning in that case. Cc: Frederic Weisbecker Cc: Steven Rostedt Cc: Jaswinder Singh Rajput LKML-Reference: <1237394936.3132.1.camel@localhost.localdomain> Signed-off-by: Ingo Molnar --- kernel/tracepoint.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/kernel/tracepoint.c b/kernel/tracepoint.c index adf28734fe2d..1ef5d3a601c7 100644 --- a/kernel/tracepoint.c +++ b/kernel/tracepoint.c @@ -278,10 +278,8 @@ tracepoint_update_probe_range(struct tracepoint *begin, struct tracepoint *end) struct tracepoint *iter; struct tracepoint_entry *mark_entry; - if (!begin) { - WARN_ON_ONCE(1); + if (!begin) return; - } mutex_lock(&tracepoints_mutex); for (iter = begin; iter < end; iter++) { From e9d9df44736d116726f4596f7e2f9ce2764ffc0a Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 18 Mar 2009 16:42:57 +0800 Subject: [PATCH 335/413] ftrace: protect running nmi (V3) When I review the sensitive code ftrace_nmi_enter(), I found the atomic variable nmi_running does protect NMI VS do_ftrace_mod_code(), but it can not protects NMI(entered nmi) VS NMI(ftrace_nmi_enter()). cpu#1 | cpu#2 | cpu#3 ftrace_nmi_enter() | do_ftrace_mod_code() | not modify | | ------------------------|-----------------------|-- executing | set mod_code_write = 1| executing --|-----------------------|-------------------- executing | | ftrace_nmi_enter() executing | | do modify ------------------------|-----------------------|----------------- ftrace_nmi_exit() | | cpu#3 may be being modified the code which is still being executed on cpu#1, it will have undefined results and possibly take a GPF, this patch prevents it occurred. Signed-off-by: Lai Jiangshan LKML-Reference: <49C0B411.30003@cn.fujitsu.com> Signed-off-by: Steven Rostedt --- arch/x86/kernel/ftrace.c | 63 +++++++++++++++++++++++++--------------- 1 file changed, 40 insertions(+), 23 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 1d0d7f42efe3..57b33edb7ce3 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -79,11 +79,11 @@ static unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) * * 1) Put the instruction pointer into the IP buffer * and the new code into the "code" buffer. - * 2) Set a flag that says we are modifying code - * 3) Wait for any running NMIs to finish. - * 4) Write the code - * 5) clear the flag. - * 6) Wait for any running NMIs to finish. + * 2) Wait for any running NMIs to finish and set a flag that says + * we are modifying code, it is done in an atomic operation. + * 3) Write the code + * 4) clear the flag. + * 5) Wait for any running NMIs to finish. * * If an NMI is executed, the first thing it does is to call * "ftrace_nmi_enter". This will check if the flag is set to write @@ -95,9 +95,9 @@ static unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr) * are the same as what exists. */ +#define MOD_CODE_WRITE_FLAG (1 << 31) /* set when NMI should do the write */ static atomic_t nmi_running = ATOMIC_INIT(0); static int mod_code_status; /* holds return value of text write */ -static int mod_code_write; /* set when NMI should do the write */ static void *mod_code_ip; /* holds the IP to write to */ static void *mod_code_newcode; /* holds the text to write to the IP */ @@ -114,6 +114,20 @@ int ftrace_arch_read_dyn_info(char *buf, int size) return r; } +static void clear_mod_flag(void) +{ + int old = atomic_read(&nmi_running); + + for (;;) { + int new = old & ~MOD_CODE_WRITE_FLAG; + + if (old == new) + break; + + old = atomic_cmpxchg(&nmi_running, old, new); + } +} + static void ftrace_mod_code(void) { /* @@ -127,27 +141,39 @@ static void ftrace_mod_code(void) /* if we fail, then kill any new writers */ if (mod_code_status) - mod_code_write = 0; + clear_mod_flag(); } void ftrace_nmi_enter(void) { - atomic_inc(&nmi_running); - /* Must have nmi_running seen before reading write flag */ - smp_mb(); - if (mod_code_write) { + if (atomic_inc_return(&nmi_running) & MOD_CODE_WRITE_FLAG) { + smp_rmb(); ftrace_mod_code(); atomic_inc(&nmi_update_count); } + /* Must have previous changes seen before executions */ + smp_mb(); } void ftrace_nmi_exit(void) { /* Finish all executions before clearing nmi_running */ - smp_wmb(); + smp_mb(); atomic_dec(&nmi_running); } +static void wait_for_nmi_and_set_mod_flag(void) +{ + if (!atomic_cmpxchg(&nmi_running, 0, MOD_CODE_WRITE_FLAG)) + return; + + do { + cpu_relax(); + } while (atomic_cmpxchg(&nmi_running, 0, MOD_CODE_WRITE_FLAG)); + + nmi_wait_count++; +} + static void wait_for_nmi(void) { if (!atomic_read(&nmi_running)) @@ -167,14 +193,9 @@ do_ftrace_mod_code(unsigned long ip, void *new_code) mod_code_newcode = new_code; /* The buffers need to be visible before we let NMIs write them */ - smp_wmb(); - - mod_code_write = 1; - - /* Make sure write bit is visible before we wait on NMIs */ smp_mb(); - wait_for_nmi(); + wait_for_nmi_and_set_mod_flag(); /* Make sure all running NMIs have finished before we write the code */ smp_mb(); @@ -182,13 +203,9 @@ do_ftrace_mod_code(unsigned long ip, void *new_code) ftrace_mod_code(); /* Make sure the write happens before clearing the bit */ - smp_wmb(); - - mod_code_write = 0; - - /* make sure NMIs see the cleared bit */ smp_mb(); + clear_mod_flag(); wait_for_nmi(); return mod_code_status; From ac5f6c96859e9a664ac05b04bc96ed1caad5fe29 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 19 Mar 2009 11:29:23 -0400 Subject: [PATCH 336/413] function-graph: consolidate prologues for output Impact: clean up The prologue of the function graph entry, return and comments all start out pretty much the same. Each of these duplicate code and do so slightly differently. This patch consolidates the printing of the pid, absolute time, cpu and proc (and for entry, the interrupt). Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions_graph.c | 98 ++++++++-------------------- 1 file changed, 29 insertions(+), 69 deletions(-) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 6004ccac2dd7..2d4d185cd0d8 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -554,24 +554,24 @@ print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, } static enum print_line_t -print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, - struct trace_iterator *iter) +print_graph_prologue(struct trace_iterator *iter, struct trace_seq *s, + int type, unsigned long addr) { - int ret; - int cpu = iter->cpu; - pid_t *last_entry = iter->private; struct trace_entry *ent = iter->ent; - struct ftrace_graph_ent *call = &field->graph_ent; - struct ftrace_graph_ret_entry *leaf_ret; + pid_t *last_pid = iter->private; + int cpu = iter->cpu; + int ret; /* Pid */ - if (verif_pid(s, ent->pid, cpu, last_entry) == TRACE_TYPE_PARTIAL_LINE) + if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; - /* Interrupt */ - ret = print_graph_irq(iter, call->func, TRACE_GRAPH_ENT, cpu, ent->pid); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; + if (type) { + /* Interrupt */ + ret = print_graph_irq(iter, addr, type, cpu, ent->pid); + if (ret == TRACE_TYPE_PARTIAL_LINE) + return TRACE_TYPE_PARTIAL_LINE; + } /* Absolute time */ if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { @@ -598,6 +598,20 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, return TRACE_TYPE_PARTIAL_LINE; } + return 0; +} + +static enum print_line_t +print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, + struct trace_iterator *iter) +{ + int cpu = iter->cpu; + struct ftrace_graph_ent *call = &field->graph_ent; + struct ftrace_graph_ret_entry *leaf_ret; + + if (print_graph_prologue(iter, s, TRACE_GRAPH_ENT, call->func)) + return TRACE_TYPE_PARTIAL_LINE; + leaf_ret = get_return_for_leaf(iter, field); if (leaf_ret) return print_graph_entry_leaf(iter, field, leaf_ret, s); @@ -613,38 +627,12 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, int i; int ret; int cpu = iter->cpu; - pid_t *last_pid = iter->private, pid = ent->pid; + pid_t pid = ent->pid; unsigned long long duration = trace->rettime - trace->calltime; - /* Pid */ - if (verif_pid(s, pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) + if (print_graph_prologue(iter, s, 0, 0)) return TRACE_TYPE_PARTIAL_LINE; - /* Absolute time */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { - ret = print_graph_abs_time(iter->ts, s); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - - /* Cpu */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { - ret = print_graph_cpu(s, cpu); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - } - - /* Proc */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, ent->pid); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - /* Overhead */ ret = print_graph_overhead(duration, s); if (!ret) @@ -689,38 +677,10 @@ print_graph_comment(struct bprint_entry *trace, struct trace_seq *s, { int i; int ret; - int cpu = iter->cpu; - pid_t *last_pid = iter->private; - /* Pid */ - if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) + if (print_graph_prologue(iter, s, 0, 0)) return TRACE_TYPE_PARTIAL_LINE; - /* Absolute time */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_ABS_TIME) { - ret = print_graph_abs_time(iter->ts, s); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - - /* Cpu */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_CPU) { - ret = print_graph_cpu(s, cpu); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - } - - /* Proc */ - if (tracer_flags.val & TRACE_GRAPH_PRINT_PROC) { - ret = print_graph_proc(s, ent->pid); - if (ret == TRACE_TYPE_PARTIAL_LINE) - return TRACE_TYPE_PARTIAL_LINE; - - ret = trace_seq_printf(s, " | "); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - } - /* No overhead */ ret = print_graph_overhead(-1, s); if (!ret) From 3bf832ce1fe6988148d392599f34ca0c6a34427d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 19 Mar 2009 14:47:33 +0100 Subject: [PATCH 337/413] tracing/ring-buffer: fix non cpu hotplug case Impact: fix warning with irqsoff tracer The ring buffer allocates its buffers on pre-smp time (early_initcall). It means that, at first, only the boot cpu buffer is allocated and the ring-buffer cpumask only has the boot cpu set (cpu_online_mask). Later, the secondary cpu will show up and the ring-buffer will be notified about this event: the appropriate buffer will be allocated and the cpumask will be updated. Unfortunately, if !CONFIG_CPU_HOTPLUG, the ring-buffer will not be notified about the secondary cpus, meaning that the cpumask will have only the cpu boot set, and only one cpu buffer allocated. We fix that by using cpu_possible_mask if !CONFIG_CPU_HOTPLUG. This patch fixes the following warning with irqsoff tracer running: [ 169.317794] WARNING: at kernel/trace/trace.c:466 update_max_tr_single+0xcc/0xf3() [ 169.318002] Hardware name: AMILO Li 2727 [ 169.318002] Modules linked in: [ 169.318002] Pid: 5624, comm: bash Not tainted 2.6.29-rc8-tip-02636-g6aafa6c #11 [ 169.318002] Call Trace: [ 169.318002] [] warn_slowpath+0xea/0x13d [ 169.318002] [] ? ftrace_call+0x5/0x2b [ 169.318002] [] ? ftrace_call+0x5/0x2b [ 169.318002] [] ? ftrace_call+0x0/0x2b [ 169.318002] [] ? ftrace_modify_code+0xa9/0x108 [ 169.318002] [] ? trace_hardirqs_off+0x25/0x27 [ 169.318002] [] ? _spin_unlock_irqrestore+0x1f/0x2d [ 169.318002] [] ? ring_buffer_reset_cpu+0xf6/0xfb [ 169.318002] [] ? ring_buffer_reset+0x36/0x48 [ 169.318002] [] update_max_tr_single+0xcc/0xf3 [ 169.318002] [] ? sysret_check+0x22/0x5d [ 169.318002] [] stop_critical_timing+0x142/0x204 [ 169.318002] [] trace_hardirqs_on_caller+0x23/0x25 [ 169.318002] [] trace_hardirqs_on_thunk+0x3a/0x3c [ 169.318002] [] ? sysret_check+0x22/0x5d [ 169.318002] ---[ end trace db76cbf775a750cf ]--- Because this tracer may try to swap two cpu ring buffers for an unregistered cpu on the ring buffer. This patch might also fix a fair loss of traces due to unallocated buffers for secondary cpus. Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Acked-b: Steven Rostedt LKML-Reference: <1237470453-5427-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index bbf51922a8ca..384ca5d9d729 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -577,8 +577,17 @@ struct ring_buffer *ring_buffer_alloc(unsigned long size, unsigned flags) if (buffer->pages == 1) buffer->pages++; + /* + * In case of non-hotplug cpu, if the ring-buffer is allocated + * in early initcall, it will not be notified of secondary cpus. + * In that off case, we need to allocate for all possible cpus. + */ +#ifdef CONFIG_HOTPLUG_CPU get_online_cpus(); cpumask_copy(buffer->cpumask, cpu_online_mask); +#else + cpumask_copy(buffer->cpumask, cpu_possible_mask); +#endif buffer->cpus = nr_cpu_ids; bsize = sizeof(void *) * nr_cpu_ids; From 5ef841f6f32dce0b752a4fa0622781ee67a0e874 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 19 Mar 2009 12:20:38 -0400 Subject: [PATCH 338/413] tracing: make print_(b)printk_msg_only global This patch makes print_printk_msg_only and print_bprintk_msg_only global for other functions to use. It also renames them by adding a "trace_" to the beginning to avoid namespace collisions. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 36 ++---------------------------------- kernel/trace/trace_output.c | 32 ++++++++++++++++++++++++++++++++ kernel/trace/trace_output.h | 5 +++++ 3 files changed, 39 insertions(+), 34 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 8d981ababc45..c637cb687cf2 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1694,38 +1694,6 @@ static enum print_line_t print_hex_fmt(struct trace_iterator *iter) return TRACE_TYPE_HANDLED; } -static enum print_line_t print_bprintk_msg_only(struct trace_iterator *iter) -{ - struct trace_seq *s = &iter->seq; - struct trace_entry *entry = iter->ent; - struct bprint_entry *field; - int ret; - - trace_assign_type(field, entry); - - ret = trace_seq_bprintf(s, field->fmt, field->buf); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - return TRACE_TYPE_HANDLED; -} - -static enum print_line_t print_printk_msg_only(struct trace_iterator *iter) -{ - struct trace_seq *s = &iter->seq; - struct trace_entry *entry = iter->ent; - struct print_entry *field; - int ret; - - trace_assign_type(field, entry); - - ret = trace_seq_printf(s, "%s", field->buf); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; - - return TRACE_TYPE_HANDLED; -} - static enum print_line_t print_bin_fmt(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; @@ -1787,12 +1755,12 @@ static enum print_line_t print_trace_line(struct trace_iterator *iter) if (iter->ent->type == TRACE_BPRINT && trace_flags & TRACE_ITER_PRINTK && trace_flags & TRACE_ITER_PRINTK_MSGONLY) - return print_bprintk_msg_only(iter); + return trace_print_bprintk_msg_only(iter); if (iter->ent->type == TRACE_PRINT && trace_flags & TRACE_ITER_PRINTK && trace_flags & TRACE_ITER_PRINTK_MSGONLY) - return print_printk_msg_only(iter); + return trace_print_printk_msg_only(iter); if (trace_flags & TRACE_ITER_BIN) return print_bin_fmt(iter); diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 6a4c9dea191e..b45141748af5 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -19,6 +19,38 @@ static struct hlist_head event_hash[EVENT_HASHSIZE] __read_mostly; static int next_event_type = __TRACE_LAST_TYPE + 1; +enum print_line_t trace_print_bprintk_msg_only(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + struct bprint_entry *field; + int ret; + + trace_assign_type(field, entry); + + ret = trace_seq_bprintf(s, field->fmt, field->buf); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + +enum print_line_t trace_print_printk_msg_only(struct trace_iterator *iter) +{ + struct trace_seq *s = &iter->seq; + struct trace_entry *entry = iter->ent; + struct print_entry *field; + int ret; + + trace_assign_type(field, entry); + + ret = trace_seq_printf(s, "%s", field->buf); + if (!ret) + return TRACE_TYPE_PARTIAL_LINE; + + return TRACE_TYPE_HANDLED; +} + /** * trace_seq_printf - sequence printing of trace information * @s: trace sequence descriptor diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 3b90e6ade1aa..35c422fb51a9 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -15,6 +15,11 @@ struct trace_event { trace_print_func binary; }; +extern enum print_line_t +trace_print_bprintk_msg_only(struct trace_iterator *iter); +extern enum print_line_t +trace_print_printk_msg_only(struct trace_iterator *iter); + extern int trace_seq_printf(struct trace_seq *s, const char *fmt, ...) __attribute__ ((format (printf, 2, 3))); extern int From 2fbcdb35aca614f9529a0e7d340146cf0b71684f Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 19 Mar 2009 13:24:42 -0400 Subject: [PATCH 339/413] function-graph: calculate function depth within function graph tracer Currently, the function graph tracer depends on the trace_printk to record the depth. All the information is already there in the trace to calculate function depth, with the exception of having the printk be the first item. But as soon as a entry or exit is reached, then we know the depth. This patch changes the iter->private data from recording a per cpu last_pid, to a structure that holds both the last_pid and the current depth. This data is used to determine the function depth for the printks. Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions_graph.c | 91 +++++++++++++++++++++------- 1 file changed, 69 insertions(+), 22 deletions(-) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 2d4d185cd0d8..66ea23b64fe6 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -14,6 +14,11 @@ #include "trace.h" #include "trace_output.h" +struct fgraph_data { + pid_t last_pid; + int depth; +}; + #define TRACE_GRAPH_INDENT 2 /* Flag options */ @@ -231,16 +236,16 @@ print_graph_proc(struct trace_seq *s, pid_t pid) /* If the pid changed since the last trace, output this event */ static enum print_line_t -verif_pid(struct trace_seq *s, pid_t pid, int cpu, pid_t *last_pids_cpu) +verif_pid(struct trace_seq *s, pid_t pid, int cpu, struct fgraph_data *data) { pid_t prev_pid; pid_t *last_pid; int ret; - if (!last_pids_cpu) + if (!data) return TRACE_TYPE_HANDLED; - last_pid = per_cpu_ptr(last_pids_cpu, cpu); + last_pid = &(per_cpu_ptr(data, cpu)->last_pid); if (*last_pid == pid) return TRACE_TYPE_HANDLED; @@ -471,6 +476,7 @@ print_graph_entry_leaf(struct trace_iterator *iter, struct ftrace_graph_ent_entry *entry, struct ftrace_graph_ret_entry *ret_entry, struct trace_seq *s) { + struct fgraph_data *data = iter->private; struct ftrace_graph_ret *graph_ret; struct ftrace_graph_ent *call; unsigned long long duration; @@ -481,6 +487,18 @@ print_graph_entry_leaf(struct trace_iterator *iter, call = &entry->graph_ent; duration = graph_ret->rettime - graph_ret->calltime; + if (data) { + int cpu = iter->cpu; + int *depth = &(per_cpu_ptr(data, cpu)->depth); + + /* + * Comments display at + 1 to depth. Since + * this is a leaf function, keep the comments + * equal to this depth. + */ + *depth = call->depth - 1; + } + /* Overhead */ ret = print_graph_overhead(duration, s); if (!ret) @@ -512,12 +530,21 @@ print_graph_entry_leaf(struct trace_iterator *iter, } static enum print_line_t -print_graph_entry_nested(struct ftrace_graph_ent_entry *entry, - struct trace_seq *s, pid_t pid, int cpu) +print_graph_entry_nested(struct trace_iterator *iter, + struct ftrace_graph_ent_entry *entry, + struct trace_seq *s, int cpu) { - int i; - int ret; struct ftrace_graph_ent *call = &entry->graph_ent; + struct fgraph_data *data = iter->private; + int ret; + int i; + + if (data) { + int cpu = iter->cpu; + int *depth = &(per_cpu_ptr(data, cpu)->depth); + + *depth = call->depth; + } /* No overhead */ ret = print_graph_overhead(-1, s); @@ -557,13 +584,13 @@ static enum print_line_t print_graph_prologue(struct trace_iterator *iter, struct trace_seq *s, int type, unsigned long addr) { + struct fgraph_data *data = iter->private; struct trace_entry *ent = iter->ent; - pid_t *last_pid = iter->private; int cpu = iter->cpu; int ret; /* Pid */ - if (verif_pid(s, ent->pid, cpu, last_pid) == TRACE_TYPE_PARTIAL_LINE) + if (verif_pid(s, ent->pid, cpu, data) == TRACE_TYPE_PARTIAL_LINE) return TRACE_TYPE_PARTIAL_LINE; if (type) { @@ -616,7 +643,7 @@ print_graph_entry(struct ftrace_graph_ent_entry *field, struct trace_seq *s, if (leaf_ret) return print_graph_entry_leaf(iter, field, leaf_ret, s); else - return print_graph_entry_nested(field, s, iter->ent->pid, cpu); + return print_graph_entry_nested(iter, field, s, cpu); } @@ -624,11 +651,24 @@ static enum print_line_t print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, struct trace_entry *ent, struct trace_iterator *iter) { - int i; - int ret; - int cpu = iter->cpu; - pid_t pid = ent->pid; unsigned long long duration = trace->rettime - trace->calltime; + struct fgraph_data *data = iter->private; + pid_t pid = ent->pid; + int cpu = iter->cpu; + int ret; + int i; + + if (data) { + int cpu = iter->cpu; + int *depth = &(per_cpu_ptr(data, cpu)->depth); + + /* + * Comments display at + 1 to depth. This is the + * return from a function, we now want the comments + * to display at the same level of the bracket. + */ + *depth = trace->depth - 1; + } if (print_graph_prologue(iter, s, 0, 0)) return TRACE_TYPE_PARTIAL_LINE; @@ -675,8 +715,13 @@ static enum print_line_t print_graph_comment(struct bprint_entry *trace, struct trace_seq *s, struct trace_entry *ent, struct trace_iterator *iter) { - int i; + struct fgraph_data *data = iter->private; + int depth = 0; int ret; + int i; + + if (data) + depth = per_cpu_ptr(data, iter->cpu)->depth; if (print_graph_prologue(iter, s, 0, 0)) return TRACE_TYPE_PARTIAL_LINE; @@ -694,8 +739,8 @@ print_graph_comment(struct bprint_entry *trace, struct trace_seq *s, } /* Indentation */ - if (trace->depth > 0) - for (i = 0; i < (trace->depth + 1) * TRACE_GRAPH_INDENT; i++) { + if (depth > 0) + for (i = 0; i < (depth + 1) * TRACE_GRAPH_INDENT; i++) { ret = trace_seq_printf(s, " "); if (!ret) return TRACE_TYPE_PARTIAL_LINE; @@ -780,19 +825,21 @@ static void print_graph_headers(struct seq_file *s) static void graph_trace_open(struct trace_iterator *iter) { - /* pid on the last trace processed */ - pid_t *last_pid = alloc_percpu(pid_t); + /* pid and depth on the last trace processed */ + struct fgraph_data *data = alloc_percpu(struct fgraph_data); int cpu; - if (!last_pid) + if (!data) pr_warning("function graph tracer: not enough memory\n"); else for_each_possible_cpu(cpu) { - pid_t *pid = per_cpu_ptr(last_pid, cpu); + pid_t *pid = &(per_cpu_ptr(data, cpu)->last_pid); + int *depth = &(per_cpu_ptr(data, cpu)->depth); *pid = -1; + *depth = 0; } - iter->private = last_pid; + iter->private = data; } static void graph_trace_close(struct trace_iterator *iter) From 40ce74f19c28077550646c76d96a075bf312e461 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 19 Mar 2009 14:03:53 -0400 Subject: [PATCH 340/413] tracing: remove recording function depth from trace_printk The function depth in trace_printk was to facilitate the function graph output. Now that the function graph calculates the depth within the trace output, we no longer need to record the depth when the trace_printk is called. Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 8 +++----- kernel/trace/trace.h | 6 ++---- kernel/trace/trace_event_types.h | 2 -- kernel/trace/trace_mmiotrace.c | 2 +- kernel/trace/trace_printk.c | 8 ++++---- 5 files changed, 10 insertions(+), 16 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index c637cb687cf2..f7f359d45823 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -1194,7 +1194,7 @@ void trace_graph_return(struct ftrace_graph_ret *trace) * trace_vbprintk - write binary msg to tracing buffer * */ -int trace_vbprintk(unsigned long ip, int depth, const char *fmt, va_list args) +int trace_vbprintk(unsigned long ip, const char *fmt, va_list args) { static raw_spinlock_t trace_buf_lock = (raw_spinlock_t)__RAW_SPIN_LOCK_UNLOCKED; @@ -1236,7 +1236,6 @@ int trace_vbprintk(unsigned long ip, int depth, const char *fmt, va_list args) goto out_unlock; entry = ring_buffer_event_data(event); entry->ip = ip; - entry->depth = depth; entry->fmt = fmt; memcpy(entry->buf, trace_buf, sizeof(u32) * len); @@ -1254,7 +1253,7 @@ int trace_vbprintk(unsigned long ip, int depth, const char *fmt, va_list args) } EXPORT_SYMBOL_GPL(trace_vbprintk); -int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) +int trace_vprintk(unsigned long ip, const char *fmt, va_list args) { static raw_spinlock_t trace_buf_lock = __RAW_SPIN_LOCK_UNLOCKED; static char trace_buf[TRACE_BUF_SIZE]; @@ -1291,7 +1290,6 @@ int trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args) goto out_unlock; entry = ring_buffer_event_data(event); entry->ip = ip; - entry->depth = depth; memcpy(&entry->buf, trace_buf, len); entry->buf[len] = 0; @@ -3140,7 +3138,7 @@ static int mark_printk(const char *fmt, ...) int ret; va_list args; va_start(args, fmt); - ret = trace_vprintk(0, -1, fmt, args); + ret = trace_vprintk(0, fmt, args); va_end(args); return ret; } diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 38276d1638e3..7c9a0cbf5dca 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -123,7 +123,6 @@ struct userstack_entry { struct bprint_entry { struct trace_entry ent; unsigned long ip; - int depth; const char *fmt; u32 buf[]; }; @@ -131,7 +130,6 @@ struct bprint_entry { struct print_entry { struct trace_entry ent; unsigned long ip; - int depth; char buf[]; }; @@ -598,9 +596,9 @@ extern int trace_selftest_startup_branch(struct tracer *trace, extern void *head_page(struct trace_array_cpu *data); extern long ns2usecs(cycle_t nsec); extern int -trace_vbprintk(unsigned long ip, int depth, const char *fmt, va_list args); +trace_vbprintk(unsigned long ip, const char *fmt, va_list args); extern int -trace_vprintk(unsigned long ip, int depth, const char *fmt, va_list args); +trace_vprintk(unsigned long ip, const char *fmt, va_list args); extern unsigned long trace_flags; diff --git a/kernel/trace/trace_event_types.h b/kernel/trace/trace_event_types.h index 019915063fe6..fd78bee71dd7 100644 --- a/kernel/trace/trace_event_types.h +++ b/kernel/trace/trace_event_types.h @@ -105,7 +105,6 @@ TRACE_EVENT_FORMAT(user_stack, TRACE_USER_STACK, userstack_entry, ignore, TRACE_EVENT_FORMAT(bprint, TRACE_BPRINT, bprint_entry, ignore, TRACE_STRUCT( TRACE_FIELD(unsigned long, ip, ip) - TRACE_FIELD(unsigned int, depth, depth) TRACE_FIELD(char *, fmt, fmt) TRACE_FIELD_ZERO_CHAR(buf) ), @@ -115,7 +114,6 @@ TRACE_EVENT_FORMAT(bprint, TRACE_BPRINT, bprint_entry, ignore, TRACE_EVENT_FORMAT(print, TRACE_PRINT, print_entry, ignore, TRACE_STRUCT( TRACE_FIELD(unsigned long, ip, ip) - TRACE_FIELD(unsigned int, depth, depth) TRACE_FIELD_ZERO_CHAR(buf) ), TP_RAW_FMT("%08lx (%d) fmt:%p %s") diff --git a/kernel/trace/trace_mmiotrace.c b/kernel/trace/trace_mmiotrace.c index f095916e477f..8e37fcddd8b4 100644 --- a/kernel/trace/trace_mmiotrace.c +++ b/kernel/trace/trace_mmiotrace.c @@ -359,5 +359,5 @@ void mmio_trace_mapping(struct mmiotrace_map *map) int mmio_trace_printk(const char *fmt, va_list args) { - return trace_vprintk(0, -1, fmt, args); + return trace_vprintk(0, fmt, args); } diff --git a/kernel/trace/trace_printk.c b/kernel/trace/trace_printk.c index 486785214e3e..eb81556107fe 100644 --- a/kernel/trace/trace_printk.c +++ b/kernel/trace/trace_printk.c @@ -112,7 +112,7 @@ int __trace_bprintk(unsigned long ip, const char *fmt, ...) return 0; va_start(ap, fmt); - ret = trace_vbprintk(ip, task_curr_ret_stack(current), fmt, ap); + ret = trace_vbprintk(ip, fmt, ap); va_end(ap); return ret; } @@ -126,7 +126,7 @@ int __ftrace_vbprintk(unsigned long ip, const char *fmt, va_list ap) if (!(trace_flags & TRACE_ITER_PRINTK)) return 0; - return trace_vbprintk(ip, task_curr_ret_stack(current), fmt, ap); + return trace_vbprintk(ip, fmt, ap); } EXPORT_SYMBOL_GPL(__ftrace_vbprintk); @@ -139,7 +139,7 @@ int __trace_printk(unsigned long ip, const char *fmt, ...) return 0; va_start(ap, fmt); - ret = trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); + ret = trace_vprintk(ip, fmt, ap); va_end(ap); return ret; } @@ -150,7 +150,7 @@ int __ftrace_vprintk(unsigned long ip, const char *fmt, va_list ap) if (!(trace_flags & TRACE_ITER_PRINTK)) return 0; - return trace_vprintk(ip, task_curr_ret_stack(current), fmt, ap); + return trace_vprintk(ip, fmt, ap); } EXPORT_SYMBOL_GPL(__ftrace_vprintk); From 5087f8d2a2f2daff5a913d72d8ea3ad601948e10 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 19 Mar 2009 15:14:46 -0400 Subject: [PATCH 341/413] function-graph: show binary events as comments With the added TRACE_EVENT macro, the events no longer appear in the function graph tracer. This was because the function graph did not know how to display the entries. The graph tracer was only aware of its own entries and the printk entries. By using the event call back feature, the graph tracer can now display the events. # echo irq > /debug/tracing/set_event Which can show: 0) | handle_IRQ_event() { 0) | /* irq_handler_entry: irq=48 handler=eth0 */ 0) | e1000_intr() { 0) 0.926 us | __napi_schedule(); 0) 3.888 us | } 0) | /* irq_handler_exit: irq=48 return=handled */ 0) 0.655 us | runqueue_is_locked(); 0) | __wake_up() { 0) 0.831 us | _spin_lock_irqsave(); The irq entry and exit events show up as comments. Signed-off-by: Steven Rostedt --- kernel/trace/trace_functions_graph.c | 40 +++++++++++++++++++--------- 1 file changed, 28 insertions(+), 12 deletions(-) diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index 66ea23b64fe6..e876816fa8e7 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -712,10 +712,12 @@ print_graph_return(struct ftrace_graph_ret *trace, struct trace_seq *s, } static enum print_line_t -print_graph_comment(struct bprint_entry *trace, struct trace_seq *s, - struct trace_entry *ent, struct trace_iterator *iter) +print_graph_comment(struct trace_seq *s, struct trace_entry *ent, + struct trace_iterator *iter) { + unsigned long sym_flags = (trace_flags & TRACE_ITER_SYM_MASK); struct fgraph_data *data = iter->private; + struct trace_event *event; int depth = 0; int ret; int i; @@ -751,9 +753,26 @@ print_graph_comment(struct bprint_entry *trace, struct trace_seq *s, if (!ret) return TRACE_TYPE_PARTIAL_LINE; - ret = trace_seq_bprintf(s, trace->fmt, trace->buf); - if (!ret) - return TRACE_TYPE_PARTIAL_LINE; + switch (iter->ent->type) { + case TRACE_BPRINT: + ret = trace_print_bprintk_msg_only(iter); + if (ret != TRACE_TYPE_HANDLED) + return ret; + break; + case TRACE_PRINT: + ret = trace_print_printk_msg_only(iter); + if (ret != TRACE_TYPE_HANDLED) + return ret; + break; + default: + event = ftrace_find_event(ent->type); + if (!event) + return TRACE_TYPE_UNHANDLED; + + ret = event->trace(iter, sym_flags); + if (ret != TRACE_TYPE_HANDLED) + return ret; + } /* Strip ending newline */ if (s->buffer[s->len - 1] == '\n') { @@ -772,8 +791,8 @@ print_graph_comment(struct bprint_entry *trace, struct trace_seq *s, enum print_line_t print_graph_function(struct trace_iterator *iter) { - struct trace_seq *s = &iter->seq; struct trace_entry *entry = iter->ent; + struct trace_seq *s = &iter->seq; switch (entry->type) { case TRACE_GRAPH_ENT: { @@ -786,14 +805,11 @@ print_graph_function(struct trace_iterator *iter) trace_assign_type(field, entry); return print_graph_return(&field->ret, s, entry, iter); } - case TRACE_BPRINT: { - struct bprint_entry *field; - trace_assign_type(field, entry); - return print_graph_comment(field, s, entry, iter); - } default: - return TRACE_TYPE_UNHANDLED; + return print_graph_comment(s, entry, iter); } + + return TRACE_TYPE_HANDLED; } static void print_graph_headers(struct seq_file *s) From 23725aeeab10ba02bcf10ec49ad73146b54cb52f Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 19 Mar 2009 20:26:13 +0100 Subject: [PATCH 342/413] ftrace: provide an id file for each event Since not every event has a format file to read the id from, expose it explicitly in a separate file. Signed-off-by: Peter Zijlstra Cc: Paul Mackerras Cc: Steven Rostedt LKML-Reference: <20090319194233.372534033@chello.nl> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index c88227b3b9db..7763db8fd0b3 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -412,6 +412,29 @@ event_format_read(struct file *filp, char __user *ubuf, size_t cnt, return r; } +static ssize_t +event_id_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + struct trace_seq *s; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + trace_seq_printf(s, "%d\n", call->id); + + r = simple_read_from_buffer(ubuf, cnt, ppos, + s->buffer, s->len); + kfree(s); + return r; +} + static const struct seq_operations show_event_seq_ops = { .start = t_start, .next = t_next, @@ -452,6 +475,11 @@ static const struct file_operations ftrace_event_format_fops = { .read = event_format_read, }; +static const struct file_operations ftrace_event_id_fops = { + .open = tracing_open_generic, + .read = event_id_read, +}; + static struct dentry *event_trace_events_dir(void) { static struct dentry *d_tracer; @@ -550,6 +578,14 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) "'%s/enable' entry\n", call->name); } + if (call->id) { + entry = debugfs_create_file("id", 0444, call->dir, call, + &ftrace_event_id_fops); + if (!entry) + pr_warning("Could not create debugfs '%s/id' entry\n", + call->name); + } + /* A trace may not want to export its format */ if (!call->show_format) return 0; From 28bea271e58e429eccfad3d7ee2ad12d6ee015bf Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 19 Mar 2009 20:26:14 +0100 Subject: [PATCH 343/413] ftrace: ensure every event gets an id Impact: widen user-space visibe event IDs to all events Previously only TRACE_EVENT events got ids, because only they generated raw output which needs to be demuxed from the trace. In order to provide a unique ID for each event, register everybody, regardless. Signed-off-by: Peter Zijlstra Cc: Paul Mackerras Cc: Steven Rostedt LKML-Reference: <20090319194233.464914218@chello.nl> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_stage_3.h | 15 ++++++++++++++- kernel/trace/trace_output.c | 5 +++++ 2 files changed, 19 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index ae2e323df0c7..4c26d97b4508 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -130,7 +130,19 @@ static void ftrace_unreg_event_##call(void) \ { \ unregister_trace_##call(ftrace_event_##call); \ } \ - + \ +static struct ftrace_event_call event_##call; \ + \ +static int ftrace_init_event_##call(void) \ +{ \ + int id; \ + \ + id = register_ftrace_event(NULL); \ + if (!id) \ + return -ENODEV; \ + event_##call.id = id; \ + return 0; \ +} #undef TRACE_FORMAT #define TRACE_FORMAT(call, proto, args, fmt) \ @@ -140,6 +152,7 @@ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ .name = #call, \ .system = __stringify(TRACE_SYSTEM), \ + .raw_init = ftrace_init_event_##call, \ .regfunc = ftrace_reg_event_##call, \ .unregfunc = ftrace_unreg_event_##call, \ } diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index b45141748af5..19261fdd2455 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -481,6 +481,11 @@ int register_ftrace_event(struct trace_event *event) mutex_lock(&trace_event_mutex); + if (!event) { + ret = next_event_type++; + goto out; + } + if (!event->type) event->type = next_event_type++; else if (event->type > __TRACE_LAST_TYPE) { From ac199db0189c091f2863312061c0575937f68810 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 19 Mar 2009 20:26:15 +0100 Subject: [PATCH 344/413] ftrace: event profile hooks Impact: new tracing infrastructure feature Provide infrastructure to generate software perf counter events from tracepoints. Signed-off-by: Peter Zijlstra Cc: Paul Mackerras Cc: Steven Rostedt LKML-Reference: <20090319194233.557364871@chello.nl> Signed-off-by: Ingo Molnar --- kernel/trace/Makefile | 1 + kernel/trace/events.c | 1 - kernel/trace/trace.h | 11 ++++++++ kernel/trace/trace_event_profile.c | 31 ++++++++++++++++++++ kernel/trace/trace_events.c | 9 ++---- kernel/trace/trace_events_stage_3.h | 44 +++++++++++++++++++++++++++++ 6 files changed, 89 insertions(+), 8 deletions(-) create mode 100644 kernel/trace/trace_event_profile.c diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index c3feea01c3e0..0e45c206c2f9 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -44,5 +44,6 @@ obj-$(CONFIG_EVENT_TRACER) += trace_events.o obj-$(CONFIG_EVENT_TRACER) += events.o obj-$(CONFIG_EVENT_TRACER) += trace_export.o obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o +obj-$(CONFIG_EVENT_PROFILE) += trace_event_profile.o libftrace-y := ftrace.o diff --git a/kernel/trace/events.c b/kernel/trace/events.c index 9fc918da404f..246f2aa6dc46 100644 --- a/kernel/trace/events.c +++ b/kernel/trace/events.c @@ -12,4 +12,3 @@ #include "trace_events_stage_2.h" #include "trace_events_stage_3.h" -#include diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 7c9a0cbf5dca..7cfb741be200 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -785,12 +785,23 @@ struct ftrace_event_call { int id; int (*raw_init)(void); int (*show_format)(struct trace_seq *s); + +#ifdef CONFIG_EVENT_PROFILE + atomic_t profile_count; + int (*profile_enable)(struct ftrace_event_call *); + void (*profile_disable)(struct ftrace_event_call *); +#endif }; void event_trace_printk(unsigned long ip, const char *fmt, ...); extern struct ftrace_event_call __start_ftrace_events[]; extern struct ftrace_event_call __stop_ftrace_events[]; +#define for_each_event(event) \ + for (event = __start_ftrace_events; \ + (unsigned long)event < (unsigned long)__stop_ftrace_events; \ + event++) + extern const char *__start___trace_bprintk_fmt[]; extern const char *__stop___trace_bprintk_fmt[]; diff --git a/kernel/trace/trace_event_profile.c b/kernel/trace/trace_event_profile.c new file mode 100644 index 000000000000..22cba9970776 --- /dev/null +++ b/kernel/trace/trace_event_profile.c @@ -0,0 +1,31 @@ +/* + * trace event based perf counter profiling + * + * Copyright (C) 2009 Red Hat Inc, Peter Zijlstra + * + */ + +#include "trace.h" + +int ftrace_profile_enable(int event_id) +{ + struct ftrace_event_call *event; + + for_each_event(event) { + if (event->id == event_id) + return event->profile_enable(event); + } + + return -EINVAL; +} + +void ftrace_profile_disable(int event_id) +{ + struct ftrace_event_call *event; + + for_each_event(event) { + if (event->id == event_id) + return event->profile_disable(event); + } +} + diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 7763db8fd0b3..3047b56f6637 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -19,11 +19,6 @@ static DEFINE_MUTEX(event_mutex); -#define events_for_each(event) \ - for (event = __start_ftrace_events; \ - (unsigned long)event < (unsigned long)__stop_ftrace_events; \ - event++) - static void ftrace_clear_events(void) { struct ftrace_event_call *call = (void *)__start_ftrace_events; @@ -90,7 +85,7 @@ static int ftrace_set_clr_event(char *buf, int set) } mutex_lock(&event_mutex); - events_for_each(call) { + for_each_event(call) { if (!call->name || !call->regfunc) continue; @@ -628,7 +623,7 @@ static __init int event_trace_init(void) if (!d_events) return 0; - events_for_each(call) { + for_each_event(call) { /* The linker may leave blanks */ if (!call->name) continue; diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 4c26d97b4508..6b3261ca988c 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -109,6 +109,40 @@ #undef TP_FMT #define TP_FMT(fmt, args...) fmt "\n", ##args +#ifdef CONFIG_EVENT_PROFILE +#define _TRACE_PROFILE(call, proto, args) \ +static void ftrace_profile_##call(proto) \ +{ \ + extern void perf_tpcounter_event(int); \ + perf_tpcounter_event(event_##call.id); \ +} \ + \ +static int ftrace_profile_enable_##call(struct ftrace_event_call *call) \ +{ \ + int ret = 0; \ + \ + if (!atomic_inc_return(&call->profile_count)) \ + ret = register_trace_##call(ftrace_profile_##call); \ + \ + return ret; \ +} \ + \ +static void ftrace_profile_disable_##call(struct ftrace_event_call *call) \ +{ \ + if (atomic_add_negative(-1, &call->profile_count)) \ + unregister_trace_##call(ftrace_profile_##call); \ +} + +#define _TRACE_PROFILE_INIT(call) \ + .profile_count = ATOMIC_INIT(-1), \ + .profile_enable = ftrace_profile_enable_##call, \ + .profile_disable = ftrace_profile_disable_##call, + +#else +#define _TRACE_PROFILE(call, proto, args) +#define _TRACE_PROFILE_INIT(call) +#endif + #define _TRACE_FORMAT(call, proto, args, fmt) \ static void ftrace_event_##call(proto) \ { \ @@ -147,6 +181,7 @@ static int ftrace_init_event_##call(void) \ #undef TRACE_FORMAT #define TRACE_FORMAT(call, proto, args, fmt) \ _TRACE_FORMAT(call, PARAMS(proto), PARAMS(args), PARAMS(fmt)) \ +_TRACE_PROFILE(call, PARAMS(proto), PARAMS(args)) \ static struct ftrace_event_call __used \ __attribute__((__aligned__(4))) \ __attribute__((section("_ftrace_events"))) event_##call = { \ @@ -155,6 +190,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .raw_init = ftrace_init_event_##call, \ .regfunc = ftrace_reg_event_##call, \ .unregfunc = ftrace_unreg_event_##call, \ + _TRACE_PROFILE_INIT(call) \ } #undef __entry @@ -162,6 +198,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ #undef TRACE_EVENT #define TRACE_EVENT(call, proto, args, tstruct, assign, print) \ +_TRACE_PROFILE(call, PARAMS(proto), PARAMS(args)) \ \ static struct ftrace_event_call event_##call; \ \ @@ -227,4 +264,11 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .regfunc = ftrace_raw_reg_event_##call, \ .unregfunc = ftrace_raw_unreg_event_##call, \ .show_format = ftrace_format_##call, \ + _TRACE_PROFILE_INIT(call) \ } + +#include + +#undef _TRACE_PROFILE +#undef _TRACE_PROFILE_INIT + From 505f2b970b2269ce4cb669b3ff4f6479d379cec2 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 20 Mar 2009 11:05:04 +0100 Subject: [PATCH 345/413] tracing, Text Edit Lock - kprobes architecture independent support, nommu fix Impact: build fix on SH !CONFIG_MMU Stephen Rothwell reported this linux-next build failure on the SH architecture: kernel/built-in.o: In function `disable_all_kprobes': kernel/kprobes.c:1382: undefined reference to `text_mutex' [...] And observed: | Introduced by commit 4460fdad85becd569f11501ad5b91814814335ff ("tracing, | Text Edit Lock - kprobes architecture independent support") from the | tracing tree. text_mutex is defined in mm/memory.c which is only built | if CONFIG_MMU is defined, which is not true for sh allmodconfig. Move this lock to kernel/extable.c (which is already home to various kernel text related routines), which file is always built-in. Reported-by: Stephen Rothwell Cc: Paul Mundt Cc: Mathieu Desnoyers LKML-Reference: <20090320110602.86351a91.sfr@canb.auug.org.au> Signed-off-by: Ingo Molnar --- kernel/extable.c | 16 +++++++++++++--- mm/memory.c | 8 -------- 2 files changed, 13 insertions(+), 11 deletions(-) diff --git a/kernel/extable.c b/kernel/extable.c index 0df6253730be..25d39b0c3a1b 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -15,11 +15,21 @@ along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ -#include -#include #include -#include +#include +#include +#include + #include +#include + +/* + * mutex protecting text section modification (dynamic code patching). + * some users need to sleep (allocating memory...) while they hold this lock. + * + * NOT exported to modules - patching kernel text is a really delicate matter. + */ +DEFINE_MUTEX(text_mutex); extern struct exception_table_entry __start___ex_table[]; extern struct exception_table_entry __stop___ex_table[]; diff --git a/mm/memory.c b/mm/memory.c index 05fab3bc5b4b..dfc9e4ea4e8b 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -101,14 +101,6 @@ int randomize_va_space __read_mostly = 2; #endif -/* - * mutex protecting text section modification (dynamic code patching). - * some users need to sleep (allocating memory...) while they hold this lock. - * - * NOT exported to modules - patching kernel text is a really delicate matter. - */ -DEFINE_MUTEX(text_mutex); - static int __init disable_randmaps(char *s) { randomize_va_space = 0; From 09c9e84d474d917d9de5b9011ed2064b03a19677 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 21 Mar 2009 04:33:36 +0100 Subject: [PATCH 346/413] tracing/ring-buffer: don't annotate rb_cpu_notify with __cpuinit Impact: remove a section warning CONFIG_DEBUG_SECTION_MISMATCH raises the following warning on -tip: WARNING: kernel/trace/built-in.o(.text+0x5bc5): Section mismatch in reference from the function ring_buffer_alloc() to the function .cpuinit.text:rb_cpu_notify() The function ring_buffer_alloc() references the function __cpuinit rb_cpu_notify(). This is actually harmless. The code in the ring buffer don't build rb_cpu_notify and other cpu hotplug stuffs when !CONFIG_HOTPLUG_CPU so we have no risk to reference freed memory here (it would even be harmless if we unconditionally build it because register_cpu_notifier would do nothing when !CONFIG_HOTPLUG_CPU. But since ring_buffer_alloc() can be called everytime, we don't want it to be annotated with __cpuinit so we drop the __cpuinit from rb_cpu_notify. This is not a waste of memory because it is only defined and used on CONFIG_HOTPLUG_CPU. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1237606416-22268-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 384ca5d9d729..808b14bbf076 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -535,8 +535,8 @@ static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer) extern int ring_buffer_page_too_big(void); #ifdef CONFIG_HOTPLUG_CPU -static int __cpuinit rb_cpu_notify(struct notifier_block *self, - unsigned long action, void *hcpu); +static int rb_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu); #endif /** @@ -2784,8 +2784,8 @@ static __init int rb_init_debugfs(void) fs_initcall(rb_init_debugfs); #ifdef CONFIG_HOTPLUG_CPU -static int __cpuinit rb_cpu_notify(struct notifier_block *self, - unsigned long action, void *hcpu) +static int rb_cpu_notify(struct notifier_block *self, + unsigned long action, void *hcpu) { struct ring_buffer *buffer = container_of(self, struct ring_buffer, cpu_notify); From 1a17662ea033674a58bad3603531b0b5d42572f6 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 20 Mar 2009 09:47:30 +0800 Subject: [PATCH 347/413] blktrace: fix possible memory leak When we failed to create "block" debugfs dir, we should do some cleanups. Signed-off-by: Li Zefan Acked-by: Frederic Weisbecker Acked-by: Arnaldo Carvalho de Melo Cc: Jens Axboe Cc: Steven Rostedt LKML-Reference: <49C2F5B2.8000800@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index b171778e3863..fb3bc53835dd 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -432,7 +432,7 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, if (!blk_tree_root) { blk_tree_root = debugfs_create_dir("block", NULL); if (!blk_tree_root) - return -ENOMEM; + goto err; } dir = debugfs_create_dir(buts->name, blk_tree_root); From 5006ea73f38caef6065d1136808413813271633f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 20 Mar 2009 09:48:03 +0800 Subject: [PATCH 348/413] blktrace: make blk_tracer_enabled a bool flag It doesn't have to be a counter, and it can be a bool flag instead. Signed-off-by: Li Zefan Acked-by: Arnaldo Carvalho de Melo Cc: Jens Axboe Cc: Steven Rostedt Acked-by: Frederic Weisbecker LKML-Reference: <49C2F5D3.8090104@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 12 +++--------- 1 file changed, 3 insertions(+), 9 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index fb3bc53835dd..73845b7968bb 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -30,7 +30,7 @@ static unsigned int blktrace_seq __read_mostly = 1; static struct trace_array *blk_tr; -static int __read_mostly blk_tracer_enabled; +static bool blk_tracer_enabled __read_mostly; /* Select an alternative, minimalistic output than the original one */ #define TRACE_BLK_OPT_CLASSIC 0x1 @@ -1111,9 +1111,7 @@ static int blk_tracer_init(struct trace_array *tr) { blk_tr = tr; blk_tracer_start(tr); - mutex_lock(&blk_probe_mutex); - blk_tracer_enabled++; - mutex_unlock(&blk_probe_mutex); + blk_tracer_enabled = true; return 0; } @@ -1131,11 +1129,7 @@ static void blk_tracer_reset(struct trace_array *tr) if (!atomic_read(&blk_probes_ref)) return; - mutex_lock(&blk_probe_mutex); - blk_tracer_enabled--; - WARN_ON(blk_tracer_enabled < 0); - mutex_unlock(&blk_probe_mutex); - + blk_tracer_enabled = false; blk_tracer_stop(tr); } From 3c289ba7c320560ee74979a8895141c829046a2d Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 20 Mar 2009 09:48:26 +0800 Subject: [PATCH 349/413] blktrace: remove blk_probe_mutex blk_register_tracepoints() always returns 0, so make it return void, thus we don't need to use blk_probe_mutex to protect blk_probes_ref. Signed-off-by: Li Zefan Acked-by: Frederic Weisbecker Acked-by: Arnaldo Carvalho de Melo Cc: Jens Axboe Cc: Steven Rostedt LKML-Reference: <49C2F5EA.8060606@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 27 +++++---------------------- 1 file changed, 5 insertions(+), 22 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 73845b7968bb..223b92e77b3f 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -47,10 +47,9 @@ static struct tracer_flags blk_tracer_flags = { }; /* Global reference count of probes */ -static DEFINE_MUTEX(blk_probe_mutex); static atomic_t blk_probes_ref = ATOMIC_INIT(0); -static int blk_register_tracepoints(void); +static void blk_register_tracepoints(void); static void blk_unregister_tracepoints(void); /* @@ -256,10 +255,8 @@ static void blk_trace_cleanup(struct blk_trace *bt) free_percpu(bt->sequence); free_percpu(bt->msg_data); kfree(bt); - mutex_lock(&blk_probe_mutex); if (atomic_dec_and_test(&blk_probes_ref)) blk_unregister_tracepoints(); - mutex_unlock(&blk_probe_mutex); } int blk_trace_remove(struct request_queue *q) @@ -471,13 +468,8 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, bt->pid = buts->pid; bt->trace_state = Blktrace_setup; - mutex_lock(&blk_probe_mutex); - if (atomic_add_return(1, &blk_probes_ref) == 1) { - ret = blk_register_tracepoints(); - if (ret) - goto probe_err; - } - mutex_unlock(&blk_probe_mutex); + if (atomic_add_return(1, &blk_probes_ref) == 1) + blk_register_tracepoints(); ret = -EBUSY; old_bt = xchg(&q->blk_trace, bt); @@ -487,9 +479,6 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, } return 0; -probe_err: - atomic_dec(&blk_probes_ref); - mutex_unlock(&blk_probe_mutex); err: if (bt) { if (bt->msg_file) @@ -863,7 +852,7 @@ void blk_add_driver_data(struct request_queue *q, } EXPORT_SYMBOL_GPL(blk_add_driver_data); -static int blk_register_tracepoints(void) +static void blk_register_tracepoints(void) { int ret; @@ -901,7 +890,6 @@ static int blk_register_tracepoints(void) WARN_ON(ret); ret = register_trace_block_remap(blk_add_trace_remap); WARN_ON(ret); - return 0; } static void blk_unregister_tracepoints(void) @@ -1099,11 +1087,8 @@ static void blk_tracer_print_header(struct seq_file *m) static void blk_tracer_start(struct trace_array *tr) { - mutex_lock(&blk_probe_mutex); if (atomic_add_return(1, &blk_probes_ref) == 1) - if (blk_register_tracepoints()) - atomic_dec(&blk_probes_ref); - mutex_unlock(&blk_probe_mutex); + blk_register_tracepoints(); trace_flags &= ~TRACE_ITER_CONTEXT_INFO; } @@ -1118,10 +1103,8 @@ static int blk_tracer_init(struct trace_array *tr) static void blk_tracer_stop(struct trace_array *tr) { trace_flags |= TRACE_ITER_CONTEXT_INFO; - mutex_lock(&blk_probe_mutex); if (atomic_dec_and_test(&blk_probes_ref)) blk_unregister_tracepoints(); - mutex_unlock(&blk_probe_mutex); } static void blk_tracer_reset(struct trace_array *tr) From cbe28296eb1ac441b35cf45804d0ae808add7dd1 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 20 Mar 2009 09:48:47 +0800 Subject: [PATCH 350/413] blktrace: don't increase blk_probes_ref if failed to setup blk trace do_blk_trace_setup() may return EBUSY, but the current code doesn't decrease blk_probes_ref in this case. Signed-off-by: Li Zefan Acked-by: Frederic Weisbecker Acked-by: Arnaldo Carvalho de Melo Cc: Jens Axboe Cc: Steven Rostedt LKML-Reference: <49C2F5FF.80002@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 223b92e77b3f..11e7c8d9d222 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -468,9 +468,6 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, bt->pid = buts->pid; bt->trace_state = Blktrace_setup; - if (atomic_add_return(1, &blk_probes_ref) == 1) - blk_register_tracepoints(); - ret = -EBUSY; old_bt = xchg(&q->blk_trace, bt); if (old_bt) { @@ -478,6 +475,9 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, goto err; } + if (atomic_add_return(1, &blk_probes_ref) == 1) + blk_register_tracepoints(); + return 0; err: if (bt) { From 15152e448b693fa41de40f1e40ffbe717a3aab88 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 20 Mar 2009 09:49:08 +0800 Subject: [PATCH 351/413] blktrace: report EBUSY correctly blk_trace_remove_queue() returns EINVAL if q->blk_trace == NULL, but blk_trace_setup_queue() doesn't return EBUSY if q->blk_trace != NULL. # echo 0 > sdaX/trace/enable # echo 0 > sdaX/trace/enable bash: echo: write error: Invalid argument # echo 1 > sdaX/trace/enable # echo 1 > sdaX/trace/enable (should return EBUSY) Signed-off-by: Li Zefan Acked-by: Arnaldo Carvalho de Melo Acked-by: Frederic Weisbecker Cc: Jens Axboe Cc: Steven Rostedt LKML-Reference: <49C2F614.2010101@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 11e7c8d9d222..14986afdbc1c 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1260,12 +1260,10 @@ static int blk_trace_remove_queue(struct request_queue *q) static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) { struct blk_trace *old_bt, *bt = NULL; - int ret; - ret = -ENOMEM; bt = kzalloc(sizeof(*bt), GFP_KERNEL); if (!bt) - goto err; + return -ENOMEM; bt->dev = dev; bt->act_mask = (u16)-1; @@ -1276,11 +1274,10 @@ static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) if (old_bt != NULL) { (void)xchg(&q->blk_trace, old_bt); kfree(bt); - ret = -EBUSY; + return -EBUSY; } + return 0; -err: - return ret; } /* From cd649b8bb830d65c57c3c8b98d57b5402256d8bd Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 20 Mar 2009 11:33:55 +0800 Subject: [PATCH 352/413] blktrace: remove sysfs_blk_trace_enable_show/store() sysfs_blk_trace_enable_show()/store() share most of code with sysfs_blk_trace_attr_show()/store(). Signed-off-by: Li Zefan Acked-by: Frederic Weisbecker Acked-by: Arnaldo Carvalho de Melo Cc: Jens Axboe Cc: Steven Rostedt LKML-Reference: <49C30EA3.1060004@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 92 ++++++++++------------------------------- 1 file changed, 22 insertions(+), 70 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 14986afdbc1c..dfee6f915179 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1284,72 +1284,6 @@ static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) * sysfs interface to enable and configure tracing */ -static ssize_t sysfs_blk_trace_enable_show(struct device *dev, - struct device_attribute *attr, - char *buf) -{ - struct hd_struct *p = dev_to_part(dev); - struct block_device *bdev; - ssize_t ret = -ENXIO; - - lock_kernel(); - bdev = bdget(part_devt(p)); - if (bdev != NULL) { - struct request_queue *q = bdev_get_queue(bdev); - - if (q != NULL) { - mutex_lock(&bdev->bd_mutex); - ret = sprintf(buf, "%u\n", !!q->blk_trace); - mutex_unlock(&bdev->bd_mutex); - } - - bdput(bdev); - } - - unlock_kernel(); - return ret; -} - -static ssize_t sysfs_blk_trace_enable_store(struct device *dev, - struct device_attribute *attr, - const char *buf, size_t count) -{ - struct block_device *bdev; - struct request_queue *q; - struct hd_struct *p; - int value; - ssize_t ret = -ENXIO; - - if (count == 0 || sscanf(buf, "%d", &value) != 1) - goto out; - - lock_kernel(); - p = dev_to_part(dev); - bdev = bdget(part_devt(p)); - if (bdev == NULL) - goto out_unlock_kernel; - - q = bdev_get_queue(bdev); - if (q == NULL) - goto out_bdput; - - mutex_lock(&bdev->bd_mutex); - if (value) - ret = blk_trace_setup_queue(q, bdev->bd_dev); - else - ret = blk_trace_remove_queue(q); - mutex_unlock(&bdev->bd_mutex); - - if (ret == 0) - ret = count; -out_bdput: - bdput(bdev); -out_unlock_kernel: - unlock_kernel(); -out: - return ret; -} - static ssize_t sysfs_blk_trace_attr_show(struct device *dev, struct device_attribute *attr, char *buf); @@ -1361,8 +1295,7 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev, sysfs_blk_trace_attr_show, \ sysfs_blk_trace_attr_store) -static DEVICE_ATTR(enable, S_IRUGO | S_IWUSR, - sysfs_blk_trace_enable_show, sysfs_blk_trace_enable_store); +static BLK_TRACE_DEVICE_ATTR(enable); static BLK_TRACE_DEVICE_ATTR(act_mask); static BLK_TRACE_DEVICE_ATTR(pid); static BLK_TRACE_DEVICE_ATTR(start_lba); @@ -1447,6 +1380,12 @@ static ssize_t sysfs_blk_trace_attr_show(struct device *dev, if (q == NULL) goto out_bdput; mutex_lock(&bdev->bd_mutex); + + if (attr == &dev_attr_enable) { + ret = sprintf(buf, "%u\n", !!q->blk_trace); + goto out_unlock_bdev; + } + if (q->blk_trace == NULL) ret = sprintf(buf, "disabled\n"); else if (attr == &dev_attr_act_mask) @@ -1457,6 +1396,8 @@ static ssize_t sysfs_blk_trace_attr_show(struct device *dev, ret = sprintf(buf, "%llu\n", q->blk_trace->start_lba); else if (attr == &dev_attr_end_lba) ret = sprintf(buf, "%llu\n", q->blk_trace->end_lba); + +out_unlock_bdev: mutex_unlock(&bdev->bd_mutex); out_bdput: bdput(bdev); @@ -1499,6 +1440,15 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev, goto out_bdput; mutex_lock(&bdev->bd_mutex); + + if (attr == &dev_attr_enable) { + if (value) + ret = blk_trace_setup_queue(q, bdev->bd_dev); + else + ret = blk_trace_remove_queue(q); + goto out_unlock_bdev; + } + ret = 0; if (q->blk_trace == NULL) ret = blk_trace_setup_queue(q, bdev->bd_dev); @@ -1512,13 +1462,15 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev, q->blk_trace->start_lba = value; else if (attr == &dev_attr_end_lba) q->blk_trace->end_lba = value; - ret = count; } + +out_unlock_bdev: mutex_unlock(&bdev->bd_mutex); out_bdput: bdput(bdev); out_unlock_kernel: unlock_kernel(); out: - return ret; + return ret ? ret : count; } + From b125130b22d67f249beba10b71a254558b5279d0 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 20 Mar 2009 10:34:00 +0800 Subject: [PATCH 353/413] blktrace: avoid accessing NULL bdev->bd_disk bdev->bd_disk can be NULL, if the block device is not opened. Try this against an unmounted partition, and you'll see NULL dereference: # echo 1 > /sys/block/sda/sda5/enable Signed-off-by: Li Zefan Acked-by: Frederic Weisbecker Acked-by: Arnaldo Carvalho de Melo Cc: Jens Axboe Cc: Steven Rostedt LKML-Reference: <49C30098.6080107@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index dfee6f915179..108f4f7715a5 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1362,6 +1362,14 @@ static int blk_str2act_mask(const char *str) return mask; } +static struct request_queue *blk_trace_get_queue(struct block_device *bdev) +{ + if (bdev->bd_disk == NULL) + return NULL; + + return bdev_get_queue(bdev); +} + static ssize_t sysfs_blk_trace_attr_show(struct device *dev, struct device_attribute *attr, char *buf) @@ -1376,9 +1384,10 @@ static ssize_t sysfs_blk_trace_attr_show(struct device *dev, if (bdev == NULL) goto out_unlock_kernel; - q = bdev_get_queue(bdev); + q = blk_trace_get_queue(bdev); if (q == NULL) goto out_bdput; + mutex_lock(&bdev->bd_mutex); if (attr == &dev_attr_enable) { @@ -1435,7 +1444,7 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev, if (bdev == NULL) goto out_unlock_kernel; - q = bdev_get_queue(bdev); + q = blk_trace_get_queue(bdev); if (q == NULL) goto out_bdput; From cf586b61f80229491127d3c57c06ed93c9f530d3 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 22 Mar 2009 05:04:35 +0100 Subject: [PATCH 354/413] tracing/function-graph-tracer: prevent hangs during self-tests Impact: detect tracing related hangs Sometimes, with some configs, the function graph tracer can make the timer interrupt too much slow, hanging the kernel in an endless loop of timer interrupts servicing. As suggested by Ingo, this patch brings a watchdog which stops the selftest after a defined number of functions traced, definitely disabling this tracer. For those who want to debug the cause of the function graph trace hang, you can pass the ftrace_dump_on_oops kernel parameter to dump the traces after this hang detection. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1237694675-23509-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 26 +++++++++++++++++++++--- kernel/trace/trace_selftest.c | 38 ++++++++++++++++++++++++++++++++++- 2 files changed, 60 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e3dfefe69348..e6fac0ffe6f0 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -4018,11 +4018,12 @@ trace_printk_seq(struct trace_seq *s) trace_seq_init(s); } -void ftrace_dump(void) +static void __ftrace_dump(bool disable_tracing) { static DEFINE_SPINLOCK(ftrace_dump_lock); /* use static because iter can be a bit big for the stack */ static struct trace_iterator iter; + unsigned int old_userobj; static int dump_ran; unsigned long flags; int cnt = 0, cpu; @@ -4034,14 +4035,17 @@ void ftrace_dump(void) dump_ran = 1; - /* No turning back! */ tracing_off(); - ftrace_kill(); + + if (disable_tracing) + ftrace_kill(); for_each_tracing_cpu(cpu) { atomic_inc(&global_trace.data[cpu]->disabled); } + old_userobj = trace_flags & TRACE_ITER_SYM_USEROBJ; + /* don't look at user memory in panic mode */ trace_flags &= ~TRACE_ITER_SYM_USEROBJ; @@ -4086,10 +4090,26 @@ void ftrace_dump(void) else printk(KERN_TRACE "---------------------------------\n"); + /* Re-enable tracing if requested */ + if (!disable_tracing) { + trace_flags |= old_userobj; + + for_each_tracing_cpu(cpu) { + atomic_dec(&global_trace.data[cpu]->disabled); + } + tracing_on(); + } + out: spin_unlock_irqrestore(&ftrace_dump_lock, flags); } +/* By default: disable tracing after the dump */ +void ftrace_dump(void) +{ + __ftrace_dump(true); +} + __init static int tracer_alloc_buffers(void) { struct trace_array_cpu *data; diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 38856ba78a92..b56dcf7d3566 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -248,6 +248,28 @@ trace_selftest_startup_function(struct tracer *trace, struct trace_array *tr) #ifdef CONFIG_FUNCTION_GRAPH_TRACER + +/* Maximum number of functions to trace before diagnosing a hang */ +#define GRAPH_MAX_FUNC_TEST 100000000 + +static void __ftrace_dump(bool disable_tracing); +static unsigned int graph_hang_thresh; + +/* Wrap the real function entry probe to avoid possible hanging */ +static int trace_graph_entry_watchdog(struct ftrace_graph_ent *trace) +{ + /* This is harmlessly racy, we want to approximately detect a hang */ + if (unlikely(++graph_hang_thresh > GRAPH_MAX_FUNC_TEST)) { + ftrace_graph_stop(); + printk(KERN_WARNING "BUG: Function graph tracer hang!\n"); + if (ftrace_dump_on_oops) + __ftrace_dump(false); + return 0; + } + + return trace_graph_entry(trace); +} + /* * Pretty much the same than for the function tracer from which the selftest * has been borrowed. @@ -259,15 +281,29 @@ trace_selftest_startup_function_graph(struct tracer *trace, int ret; unsigned long count; - ret = tracer_init(trace, tr); + /* + * Simulate the init() callback but we attach a watchdog callback + * to detect and recover from possible hangs + */ + tracing_reset_online_cpus(tr); + ret = register_ftrace_graph(&trace_graph_return, + &trace_graph_entry_watchdog); if (ret) { warn_failed_init_tracer(trace, ret); goto out; } + tracing_start_cmdline_record(); /* Sleep for a 1/10 of a second */ msleep(100); + /* Have we just recovered from a hang? */ + if (graph_hang_thresh > GRAPH_MAX_FUNC_TEST) { + trace->reset(tr); + ret = -1; + goto out; + } + tracing_stop(); /* check the trace buffer */ From 0cf53ff62b3e9e491ff5e5f05b193fb6ce643047 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 22 Mar 2009 15:13:07 +0100 Subject: [PATCH 355/413] tracing: keep the tracing buffer after self-test failure Instead of using ftrace_dump_on_oops, it's far more convenient to have the trace leading up to a self-test failure available in /debug/tracing/trace. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1237694675-23509-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_selftest.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index b56dcf7d3566..08f4eb2763d1 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -299,7 +299,7 @@ trace_selftest_startup_function_graph(struct tracer *trace, /* Have we just recovered from a hang? */ if (graph_hang_thresh > GRAPH_MAX_FUNC_TEST) { - trace->reset(tr); + tracing_selftest_disabled = true; ret = -1; goto out; } From cf027f645e6aee4f0ca6197a6b6a57f327fdb13f Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sun, 22 Mar 2009 03:30:39 -0500 Subject: [PATCH 356/413] tracing: add run-time field descriptions for event filtering This patch makes the field descriptions defined for event tracing available at run-time, for the event-filtering mechanism introduced in a subsequent patch. The common event fields are prepended with 'common_' in the format display, allowing them to be distinguished from the other fields that might internally have same name and can therefore be unambiguously used in filters. Signed-off-by: Tom Zanussi Acked-by: Frederic Weisbecker LKML-Reference: <1237710639.7703.46.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace.h | 30 +++++++++++++------ kernel/trace/trace_events.c | 40 ++++++++++++++++++++++++- kernel/trace/trace_events_stage_2.h | 45 +++++++++++++++++++++++++++++ kernel/trace/trace_events_stage_3.h | 2 ++ 4 files changed, 107 insertions(+), 10 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 7cfb741be200..9288dc7ad14f 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -775,16 +775,26 @@ enum { TRACE_EVENT_TYPE_RAW = 2, }; +struct ftrace_event_field { + struct list_head link; + char *name; + char *type; + int offset; + int size; +}; + struct ftrace_event_call { - char *name; - char *system; - struct dentry *dir; - int enabled; - int (*regfunc)(void); - void (*unregfunc)(void); - int id; - int (*raw_init)(void); - int (*show_format)(struct trace_seq *s); + char *name; + char *system; + struct dentry *dir; + int enabled; + int (*regfunc)(void); + void (*unregfunc)(void); + int id; + int (*raw_init)(void); + int (*show_format)(struct trace_seq *s); + int (*define_fields)(void); + struct list_head fields; #ifdef CONFIG_EVENT_PROFILE atomic_t profile_count; @@ -793,6 +803,8 @@ struct ftrace_event_call { #endif }; +int trace_define_field(struct ftrace_event_call *call, char *type, + char *name, int offset, int size); void event_trace_printk(unsigned long ip, const char *fmt, ...); extern struct ftrace_event_call __start_ftrace_events[]; extern struct ftrace_event_call __stop_ftrace_events[]; diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 3047b56f6637..961b057da28b 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -19,6 +19,34 @@ static DEFINE_MUTEX(event_mutex); +int trace_define_field(struct ftrace_event_call *call, char *type, + char *name, int offset, int size) +{ + struct ftrace_event_field *field; + + field = kmalloc(sizeof(*field), GFP_KERNEL); + if (!field) + goto err; + field->name = kstrdup(name, GFP_KERNEL); + if (!field->name) + goto err; + field->type = kstrdup(type, GFP_KERNEL); + if (!field->type) + goto err; + field->offset = offset; + field->size = size; + list_add(&field->link, &call->fields); + + return 0; +err: + if (field) { + kfree(field->name); + kfree(field->type); + } + kfree(field); + return -ENOMEM; +} + static void ftrace_clear_events(void) { struct ftrace_event_call *call = (void *)__start_ftrace_events; @@ -343,7 +371,8 @@ event_enable_write(struct file *filp, const char __user *ubuf, size_t cnt, #undef FIELD #define FIELD(type, name) \ - #type, #name, offsetof(typeof(field), name), sizeof(field.name) + #type, "common_" #name, offsetof(typeof(field), name), \ + sizeof(field.name) static int trace_write_header(struct trace_seq *s) { @@ -581,6 +610,15 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) call->name); } + if (call->define_fields) { + ret = call->define_fields(); + if (ret < 0) { + pr_warning("Could not initialize trace point" + " events/%s\n", call->name); + return ret; + } + } + /* A trace may not want to export its format */ if (!call->show_format) return 0; diff --git a/kernel/trace/trace_events_stage_2.h b/kernel/trace/trace_events_stage_2.h index 5117c43f5c67..30743f7d4110 100644 --- a/kernel/trace/trace_events_stage_2.h +++ b/kernel/trace/trace_events_stage_2.h @@ -129,3 +129,48 @@ ftrace_format_##call(struct trace_seq *s) \ } #include + +#undef __field +#define __field(type, item) \ + ret = trace_define_field(event_call, #type, #item, \ + offsetof(typeof(field), item), \ + sizeof(field.item)); \ + if (ret) \ + return ret; + +#undef __array +#define __array(type, item, len) \ + ret = trace_define_field(event_call, #type "[" #len "]", #item, \ + offsetof(typeof(field), item), \ + sizeof(field.item)); \ + if (ret) \ + return ret; + +#define __common_field(type, item) \ + ret = trace_define_field(event_call, #type, "common_" #item, \ + offsetof(typeof(field.ent), item), \ + sizeof(field.ent.item)); \ + if (ret) \ + return ret; + +#undef TRACE_EVENT +#define TRACE_EVENT(call, proto, args, tstruct, func, print) \ +int \ +ftrace_define_fields_##call(void) \ +{ \ + struct ftrace_raw_##call field; \ + struct ftrace_event_call *event_call = &event_##call; \ + int ret; \ + \ + __common_field(unsigned char, type); \ + __common_field(unsigned char, flags); \ + __common_field(unsigned char, preempt_count); \ + __common_field(int, pid); \ + __common_field(int, tgid); \ + \ + tstruct; \ + \ + return ret; \ +} + +#include diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 6b3261ca988c..468938f70141 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -252,6 +252,7 @@ static int ftrace_raw_init_event_##call(void) \ if (!id) \ return -ENODEV; \ event_##call.id = id; \ + INIT_LIST_HEAD(&event_##call.fields); \ return 0; \ } \ \ @@ -264,6 +265,7 @@ __attribute__((section("_ftrace_events"))) event_##call = { \ .regfunc = ftrace_raw_reg_event_##call, \ .unregfunc = ftrace_raw_unreg_event_##call, \ .show_format = ftrace_format_##call, \ + .define_fields = ftrace_define_fields_##call, \ _TRACE_PROFILE_INIT(call) \ } From f80d2d7725b04f8225b11b55e43bb2c77c819926 Mon Sep 17 00:00:00 2001 From: Dmitri Vorobiev Date: Sun, 22 Mar 2009 19:11:10 +0200 Subject: [PATCH 357/413] tracing, Text Edit Lock: Fix one sparse warning in kernel/extable.c Impact: cleanup. The global mutex text_mutex if declared in linux/memory.h, so this file needs to be included into kernel/extable.c, where the same mutex is defined. This fixes the following sparse warning: kernel/extable.c:32:1: warning: symbol 'text_mutex' was not declared. Should it be static? Signed-off-by: Dmitri Vorobiev LKML-Reference: <1237741871-5827-3-git-send-email-dmitri.vorobiev@movial.com> Signed-off-by: Ingo Molnar --- kernel/extable.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/extable.c b/kernel/extable.c index 25d39b0c3a1b..b54a6017b6b5 100644 --- a/kernel/extable.c +++ b/kernel/extable.c @@ -16,6 +16,7 @@ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA */ #include +#include #include #include #include From b8b94265337f83b7db9c5f429b1769d463d7da8c Mon Sep 17 00:00:00 2001 From: Dmitri Vorobiev Date: Sun, 22 Mar 2009 19:11:11 +0200 Subject: [PATCH 358/413] tracing: fix four sparse warnings Impact: cleanup. This patch fixes the following sparse warnings: kernel/trace/trace.c:385:9: warning: symbol 'trace_seq_to_buffer' was not declared. Should it be static? kernel/trace/trace_clock.c:29:13: warning: symbol 'trace_clock_local' was not declared. Should it be static? kernel/trace/trace_clock.c:54:13: warning: symbol 'trace_clock' was not declared. Should it be static? kernel/trace/trace_clock.c:74:13: warning: symbol 'trace_clock_global' was not declared. Should it be static? Signed-off-by: Dmitri Vorobiev LKML-Reference: <1237741871-5827-4-git-send-email-dmitri.vorobiev@movial.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 2 +- kernel/trace/trace_clock.c | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e6fac0ffe6f0..ace685c70186 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -382,7 +382,7 @@ ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt) return cnt; } -ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt) +static ssize_t trace_seq_to_buffer(struct trace_seq *s, void *buf, size_t cnt) { int len; void *ret; diff --git a/kernel/trace/trace_clock.c b/kernel/trace/trace_clock.c index 05b176abfd30..b588fd81f7f9 100644 --- a/kernel/trace/trace_clock.c +++ b/kernel/trace/trace_clock.c @@ -18,6 +18,7 @@ #include #include #include +#include /* * trace_clock_local(): the simplest and least coherent tracing clock. From 2d622719f1572ef31e0616444a515eba3094d050 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sun, 22 Mar 2009 03:30:49 -0500 Subject: [PATCH 359/413] tracing: add ring_buffer_event_discard() to ring buffer This patch overloads RINGBUF_TYPE_PADDING to provide a way to discard events from the ring buffer, for the event-filtering mechanism introduced in a subsequent patch. I did the initial version but thanks to Steven Rostedt for adding the parts that actually made it work. ;-) Signed-off-by: Tom Zanussi Acked-by: Frederic Weisbecker Signed-off-by: Ingo Molnar --- include/linux/ring_buffer.h | 11 +++- kernel/trace/ring_buffer.c | 117 ++++++++++++++++++++++++++++++------ 2 files changed, 105 insertions(+), 23 deletions(-) diff --git a/include/linux/ring_buffer.h b/include/linux/ring_buffer.h index 9e6052bd1a1c..e1b7b2173885 100644 --- a/include/linux/ring_buffer.h +++ b/include/linux/ring_buffer.h @@ -18,10 +18,13 @@ struct ring_buffer_event { /** * enum ring_buffer_type - internal ring buffer types * - * @RINGBUF_TYPE_PADDING: Left over page padding - * array is ignored - * size is variable depending on how much + * @RINGBUF_TYPE_PADDING: Left over page padding or discarded event + * If time_delta is 0: + * array is ignored + * size is variable depending on how much * padding is needed + * If time_delta is non zero: + * everything else same as RINGBUF_TYPE_DATA * * @RINGBUF_TYPE_TIME_EXTEND: Extend the time delta * array[0] = time delta (28 .. 59) @@ -65,6 +68,8 @@ ring_buffer_event_time_delta(struct ring_buffer_event *event) return event->time_delta; } +void ring_buffer_event_discard(struct ring_buffer_event *event); + /* * size is in bytes for each per CPU buffer. */ diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 384ca5d9d729..a09027ec1714 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -189,16 +189,65 @@ enum { RB_LEN_TIME_STAMP = 16, }; +static inline int rb_null_event(struct ring_buffer_event *event) +{ + return event->type == RINGBUF_TYPE_PADDING && event->time_delta == 0; +} + +static inline int rb_discarded_event(struct ring_buffer_event *event) +{ + return event->type == RINGBUF_TYPE_PADDING && event->time_delta; +} + +static void rb_event_set_padding(struct ring_buffer_event *event) +{ + event->type = RINGBUF_TYPE_PADDING; + event->time_delta = 0; +} + +/** + * ring_buffer_event_discard - discard an event in the ring buffer + * @buffer: the ring buffer + * @event: the event to discard + * + * Sometimes a event that is in the ring buffer needs to be ignored. + * This function lets the user discard an event in the ring buffer + * and then that event will not be read later. + * + * Note, it is up to the user to be careful with this, and protect + * against races. If the user discards an event that has been consumed + * it is possible that it could corrupt the ring buffer. + */ +void ring_buffer_event_discard(struct ring_buffer_event *event) +{ + event->type = RINGBUF_TYPE_PADDING; + /* time delta must be non zero */ + if (!event->time_delta) + event->time_delta = 1; +} + +static unsigned +rb_event_data_length(struct ring_buffer_event *event) +{ + unsigned length; + + if (event->len) + length = event->len * RB_ALIGNMENT; + else + length = event->array[0]; + return length + RB_EVNT_HDR_SIZE; +} + /* inline for ring buffer fast paths */ static unsigned rb_event_length(struct ring_buffer_event *event) { - unsigned length; - switch (event->type) { case RINGBUF_TYPE_PADDING: - /* undefined */ - return -1; + if (rb_null_event(event)) + /* undefined */ + return -1; + return rb_event_data_length(event); case RINGBUF_TYPE_TIME_EXTEND: return RB_LEN_TIME_EXTEND; @@ -207,11 +256,7 @@ rb_event_length(struct ring_buffer_event *event) return RB_LEN_TIME_STAMP; case RINGBUF_TYPE_DATA: - if (event->len) - length = event->len * RB_ALIGNMENT; - else - length = event->array[0]; - return length + RB_EVNT_HDR_SIZE; + return rb_event_data_length(event); default: BUG(); } @@ -845,11 +890,6 @@ int ring_buffer_resize(struct ring_buffer *buffer, unsigned long size) } EXPORT_SYMBOL_GPL(ring_buffer_resize); -static inline int rb_null_event(struct ring_buffer_event *event) -{ - return event->type == RINGBUF_TYPE_PADDING; -} - static inline void * __rb_data_page_index(struct buffer_data_page *bpage, unsigned index) { @@ -1219,7 +1259,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, if (tail < BUF_PAGE_SIZE) { /* Mark the rest of the page with padding */ event = __rb_page_index(tail_page, tail); - event->type = RINGBUF_TYPE_PADDING; + rb_event_set_padding(event); } if (tail <= BUF_PAGE_SIZE) @@ -1969,7 +2009,7 @@ static void rb_advance_reader(struct ring_buffer_per_cpu *cpu_buffer) event = rb_reader_event(cpu_buffer); - if (event->type == RINGBUF_TYPE_DATA) + if (event->type == RINGBUF_TYPE_DATA || rb_discarded_event(event)) cpu_buffer->entries--; rb_update_read_stamp(cpu_buffer, event); @@ -2052,9 +2092,18 @@ rb_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) switch (event->type) { case RINGBUF_TYPE_PADDING: - RB_WARN_ON(cpu_buffer, 1); + if (rb_null_event(event)) + RB_WARN_ON(cpu_buffer, 1); + /* + * Because the writer could be discarding every + * event it creates (which would probably be bad) + * if we were to go back to "again" then we may never + * catch up, and will trigger the warn on, or lock + * the box. Return the padding, and we will release + * the current locks, and try again. + */ rb_advance_reader(cpu_buffer); - return NULL; + return event; case RINGBUF_TYPE_TIME_EXTEND: /* Internal data, OK to advance */ @@ -2115,8 +2164,12 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) switch (event->type) { case RINGBUF_TYPE_PADDING: - rb_inc_iter(iter); - goto again; + if (rb_null_event(event)) { + rb_inc_iter(iter); + goto again; + } + rb_advance_iter(iter); + return event; case RINGBUF_TYPE_TIME_EXTEND: /* Internal data, OK to advance */ @@ -2163,10 +2216,16 @@ ring_buffer_peek(struct ring_buffer *buffer, int cpu, u64 *ts) if (!cpumask_test_cpu(cpu, buffer->cpumask)) return NULL; + again: spin_lock_irqsave(&cpu_buffer->reader_lock, flags); event = rb_buffer_peek(buffer, cpu, ts); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + if (event && event->type == RINGBUF_TYPE_PADDING) { + cpu_relax(); + goto again; + } + return event; } @@ -2185,10 +2244,16 @@ ring_buffer_iter_peek(struct ring_buffer_iter *iter, u64 *ts) struct ring_buffer_event *event; unsigned long flags; + again: spin_lock_irqsave(&cpu_buffer->reader_lock, flags); event = rb_iter_peek(iter, ts); spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + if (event && event->type == RINGBUF_TYPE_PADDING) { + cpu_relax(); + goto again; + } + return event; } @@ -2207,6 +2272,7 @@ ring_buffer_consume(struct ring_buffer *buffer, int cpu, u64 *ts) struct ring_buffer_event *event = NULL; unsigned long flags; + again: /* might be called in atomic */ preempt_disable(); @@ -2228,6 +2294,11 @@ ring_buffer_consume(struct ring_buffer *buffer, int cpu, u64 *ts) out: preempt_enable(); + if (event && event->type == RINGBUF_TYPE_PADDING) { + cpu_relax(); + goto again; + } + return event; } EXPORT_SYMBOL_GPL(ring_buffer_consume); @@ -2306,6 +2377,7 @@ ring_buffer_read(struct ring_buffer_iter *iter, u64 *ts) struct ring_buffer_per_cpu *cpu_buffer = iter->cpu_buffer; unsigned long flags; + again: spin_lock_irqsave(&cpu_buffer->reader_lock, flags); event = rb_iter_peek(iter, ts); if (!event) @@ -2315,6 +2387,11 @@ ring_buffer_read(struct ring_buffer_iter *iter, u64 *ts) out: spin_unlock_irqrestore(&cpu_buffer->reader_lock, flags); + if (event && event->type == RINGBUF_TYPE_PADDING) { + cpu_relax(); + goto again; + } + return event; } EXPORT_SYMBOL_GPL(ring_buffer_read); From 7ce7e4249921d5073e764f7ff7ad83cfa9894bd7 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sun, 22 Mar 2009 03:31:04 -0500 Subject: [PATCH 360/413] tracing: add per-event filtering This patch adds per-event filtering to the event tracing subsystem. It adds a 'filter' debugfs file to each event directory. This file can be written to to set filters; reading from it will display the current set of filters set for that event. Basically, any field listed in the 'format' file for an event can be filtered on (including strings, but not yet other array types) using either matching ('==') or non-matching ('!=') 'predicates'. A 'predicate' can be either a single expression: # echo pid != 0 > filter # cat filter pid != 0 or a compound expression of up to 8 sub-expressions combined using '&&' or '||': # echo comm == Xorg > filter # echo "&& sig != 29" > filter # cat filter comm == Xorg && sig != 29 Only events having field values matching an expression will be available in the trace output; non-matching events are discarded. Note that a compound expression is built up by echoing each sub-expression separately - it's not the most efficient way to do things, but it keeps the parser simple and assumes that compound expressions will be relatively uncommon. In any case, a subsequent patch introducing a way to set filters for entire subsystems should mitigate any need to do this for lots of events. Setting a filter without an '&&' or '||' clears the previous filter completely and sets the filter to the new expression: # cat filter comm == Xorg && sig != 29 # echo comm != Xorg # cat filter comm != Xorg To clear a filter, echo 0 to the filter file: # echo 0 > filter # cat filter none The limit of 8 predicates for a compound expression is arbitrary - for efficiency, it's implemented as an array of pointers to predicates, and 8 seemed more than enough for any filter... Signed-off-by: Tom Zanussi Acked-by: Frederic Weisbecker LKML-Reference: <1237710665.7703.48.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/Makefile | 1 + kernel/trace/trace.h | 28 +++ kernel/trace/trace_events.c | 77 +++++++ kernel/trace/trace_events_filter.c | 326 ++++++++++++++++++++++++++++ kernel/trace/trace_events_stage_3.h | 4 + 5 files changed, 436 insertions(+) create mode 100644 kernel/trace/trace_events_filter.c diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 0e45c206c2f9..2630f5121ec1 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -45,5 +45,6 @@ obj-$(CONFIG_EVENT_TRACER) += events.o obj-$(CONFIG_EVENT_TRACER) += trace_export.o obj-$(CONFIG_FTRACE_SYSCALLS) += trace_syscalls.o obj-$(CONFIG_EVENT_PROFILE) += trace_event_profile.o +obj-$(CONFIG_EVENT_TRACER) += trace_events_filter.o libftrace-y := ftrace.o diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 9288dc7ad14f..d9eb39e4bb38 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -795,6 +795,7 @@ struct ftrace_event_call { int (*show_format)(struct trace_seq *s); int (*define_fields)(void); struct list_head fields; + struct filter_pred **preds; #ifdef CONFIG_EVENT_PROFILE atomic_t profile_count; @@ -803,8 +804,35 @@ struct ftrace_event_call { #endif }; +#define MAX_FILTER_PRED 8 + +struct filter_pred; + +typedef int (*filter_pred_fn_t) (struct filter_pred *pred, void *event); + +struct filter_pred { + filter_pred_fn_t fn; + u64 val; + char *str_val; + int str_len; + char *field_name; + int offset; + int not; + int or; + int compound; + int clear; +}; + int trace_define_field(struct ftrace_event_call *call, char *type, char *name, int offset, int size); +extern void filter_free_pred(struct filter_pred *pred); +extern int filter_print_preds(struct filter_pred **preds, char *buf); +extern int filter_parse(char **pbuf, struct filter_pred *pred); +extern int filter_add_pred(struct ftrace_event_call *call, + struct filter_pred *pred); +extern void filter_free_preds(struct ftrace_event_call *call); +extern int filter_match_preds(struct ftrace_event_call *call, void *rec); + void event_trace_printk(unsigned long ip, const char *fmt, ...); extern struct ftrace_event_call __start_ftrace_events[]; extern struct ftrace_event_call __stop_ftrace_events[]; diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 961b057da28b..97470c48956e 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -459,6 +459,71 @@ event_id_read(struct file *filp, char __user *ubuf, size_t cnt, loff_t *ppos) return r; } +static ssize_t +event_filter_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + struct trace_seq *s; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + + r = filter_print_preds(call->preds, s->buffer); + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, r); + + kfree(s); + + return r; +} + +static ssize_t +event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct ftrace_event_call *call = filp->private_data; + char buf[64], *pbuf = buf; + struct filter_pred *pred; + int err; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + pred = kzalloc(sizeof(*pred), GFP_KERNEL); + if (!pred) + return -ENOMEM; + + err = filter_parse(&pbuf, pred); + if (err < 0) { + filter_free_pred(pred); + return err; + } + + if (pred->clear) { + filter_free_preds(call); + return cnt; + } + + if (filter_add_pred(call, pred)) { + filter_free_pred(pred); + return -EINVAL; + } + + *ppos += cnt; + + return cnt; +} + static const struct seq_operations show_event_seq_ops = { .start = t_start, .next = t_next, @@ -504,6 +569,12 @@ static const struct file_operations ftrace_event_id_fops = { .read = event_id_read, }; +static const struct file_operations ftrace_event_filter_fops = { + .open = tracing_open_generic, + .read = event_filter_read, + .write = event_filter_write, +}; + static struct dentry *event_trace_events_dir(void) { static struct dentry *d_tracer; @@ -619,6 +690,12 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) } } + entry = debugfs_create_file("filter", 0444, call->dir, call, + &ftrace_event_filter_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/filter' entry\n", call->name); + /* A trace may not want to export its format */ if (!call->show_format) return 0; diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c new file mode 100644 index 000000000000..8e8c5fa25be9 --- /dev/null +++ b/kernel/trace/trace_events_filter.c @@ -0,0 +1,326 @@ +/* + * trace_events_filter - generic event filtering + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (C) 2009 Tom Zanussi + */ + +#include +#include +#include +#include + +#include "trace.h" + +static int filter_pred_64(struct filter_pred *pred, void *event) +{ + u64 *addr = (u64 *)(event + pred->offset); + u64 val = (u64)pred->val; + int match; + + match = (val == *addr) ^ pred->not; + + return match; +} + +static int filter_pred_32(struct filter_pred *pred, void *event) +{ + u32 *addr = (u32 *)(event + pred->offset); + u32 val = (u32)pred->val; + int match; + + match = (val == *addr) ^ pred->not; + + return match; +} + +static int filter_pred_16(struct filter_pred *pred, void *event) +{ + u16 *addr = (u16 *)(event + pred->offset); + u16 val = (u16)pred->val; + int match; + + match = (val == *addr) ^ pred->not; + + return match; +} + +static int filter_pred_8(struct filter_pred *pred, void *event) +{ + u8 *addr = (u8 *)(event + pred->offset); + u8 val = (u8)pred->val; + int match; + + match = (val == *addr) ^ pred->not; + + return match; +} + +static int filter_pred_string(struct filter_pred *pred, void *event) +{ + char *addr = (char *)(event + pred->offset); + int cmp, match; + + cmp = strncmp(addr, pred->str_val, pred->str_len); + + match = (!cmp) ^ pred->not; + + return match; +} + +/* return 1 if event matches, 0 otherwise (discard) */ +int filter_match_preds(struct ftrace_event_call *call, void *rec) +{ + int i, matched, and_failed = 0; + struct filter_pred *pred; + + for (i = 0; i < MAX_FILTER_PRED; i++) { + if (call->preds[i]) { + pred = call->preds[i]; + if (and_failed && !pred->or) + continue; + matched = pred->fn(pred, rec); + if (!matched && !pred->or) { + and_failed = 1; + continue; + } else if (matched && pred->or) + return 1; + } else + break; + } + + if (and_failed) + return 0; + + return 1; +} + +int filter_print_preds(struct filter_pred **preds, char *buf) +{ + ssize_t this_len = 0; + char *field_name; + struct filter_pred *pred; + int i; + + if (!preds) { + this_len += sprintf(buf + this_len, "none\n"); + return this_len; + } + + for (i = 0; i < MAX_FILTER_PRED; i++) { + if (preds[i]) { + pred = preds[i]; + field_name = pred->field_name; + if (i) + this_len += sprintf(buf + this_len, + pred->or ? "|| " : "&& "); + this_len += sprintf(buf + this_len, + "%s ", field_name); + this_len += sprintf(buf + this_len, + pred->not ? "!= " : "== "); + if (pred->str_val) + this_len += sprintf(buf + this_len, + "%s\n", pred->str_val); + else + this_len += sprintf(buf + this_len, + "%llu\n", pred->val); + } else + break; + } + + return this_len; +} + +static struct ftrace_event_field * +find_event_field(struct ftrace_event_call *call, char *name) +{ + struct ftrace_event_field *field; + struct list_head *entry, *tmp; + + list_for_each_safe(entry, tmp, &call->fields) { + field = list_entry(entry, struct ftrace_event_field, link); + if (!strcmp(field->name, name)) + return field; + } + + return NULL; +} + +void filter_free_pred(struct filter_pred *pred) +{ + if (!pred) + return; + + kfree(pred->field_name); + kfree(pred->str_val); + kfree(pred); +} + +void filter_free_preds(struct ftrace_event_call *call) +{ + int i; + + if (call->preds) { + for (i = 0; i < MAX_FILTER_PRED; i++) + filter_free_pred(call->preds[i]); + kfree(call->preds); + call->preds = NULL; + } +} + +static int __filter_add_pred(struct ftrace_event_call *call, + struct filter_pred *pred) +{ + int i; + + if (call->preds && !pred->compound) + filter_free_preds(call); + + if (!call->preds) { + call->preds = kzalloc(MAX_FILTER_PRED * sizeof(pred), + GFP_KERNEL); + if (!call->preds) + return -ENOMEM; + } + + for (i = 0; i < MAX_FILTER_PRED; i++) { + if (!call->preds[i]) { + call->preds[i] = pred; + return 0; + } + } + + return -ENOMEM; +} + +static int is_string_field(const char *type) +{ + if (strchr(type, '[') && strstr(type, "char")) + return 1; + + return 0; +} + +int filter_add_pred(struct ftrace_event_call *call, struct filter_pred *pred) +{ + struct ftrace_event_field *field; + + field = find_event_field(call, pred->field_name); + if (!field) + return -EINVAL; + + pred->offset = field->offset; + + if (is_string_field(field->type)) { + pred->fn = filter_pred_string; + pred->str_len = field->size; + return __filter_add_pred(call, pred); + } + + switch (field->size) { + case 8: + pred->fn = filter_pred_64; + break; + case 4: + pred->fn = filter_pred_32; + break; + case 2: + pred->fn = filter_pred_16; + break; + case 1: + pred->fn = filter_pred_8; + break; + default: + return -EINVAL; + } + + return __filter_add_pred(call, pred); +} + +int filter_parse(char **pbuf, struct filter_pred *pred) +{ + char *tmp, *tok, *val_str = NULL; + int tok_n = 0; + + /* field ==/!= number, or/and field ==/!= number, number */ + while ((tok = strsep(pbuf, " \n"))) { + if (tok_n == 0) { + if (!strcmp(tok, "0")) { + pred->clear = 1; + return 0; + } else if (!strcmp(tok, "&&")) { + pred->or = 0; + pred->compound = 1; + } else if (!strcmp(tok, "||")) { + pred->or = 1; + pred->compound = 1; + } else + pred->field_name = tok; + tok_n = 1; + continue; + } + if (tok_n == 1) { + if (!pred->field_name) + pred->field_name = tok; + else if (!strcmp(tok, "!=")) + pred->not = 1; + else if (!strcmp(tok, "==")) + pred->not = 0; + else { + pred->field_name = NULL; + return -EINVAL; + } + tok_n = 2; + continue; + } + if (tok_n == 2) { + if (pred->compound) { + if (!strcmp(tok, "!=")) + pred->not = 1; + else if (!strcmp(tok, "==")) + pred->not = 0; + else { + pred->field_name = NULL; + return -EINVAL; + } + } else { + val_str = tok; + break; /* done */ + } + tok_n = 3; + continue; + } + if (tok_n == 3) { + val_str = tok; + break; /* done */ + } + } + + pred->field_name = kstrdup(pred->field_name, GFP_KERNEL); + if (!pred->field_name) + return -ENOMEM; + + pred->val = simple_strtoull(val_str, &tmp, 10); + if (tmp == val_str) { + pred->str_val = kstrdup(val_str, GFP_KERNEL); + if (!pred->str_val) + return -ENOMEM; + } + + return 0; +} + + diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 468938f70141..ebf215e87d5e 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -204,6 +204,7 @@ static struct ftrace_event_call event_##call; \ \ static void ftrace_raw_event_##call(proto) \ { \ + struct ftrace_event_call *call = &event_##call; \ struct ring_buffer_event *event; \ struct ftrace_raw_##call *entry; \ unsigned long irq_flags; \ @@ -222,6 +223,9 @@ static void ftrace_raw_event_##call(proto) \ assign; \ \ trace_current_buffer_unlock_commit(event, irq_flags, pc); \ + \ + if (call->preds && !filter_match_preds(call, entry)) \ + ring_buffer_event_discard(event); \ } \ \ static int ftrace_raw_reg_event_##call(void) \ From cfb180f3e71b2a280a254c8646a9ab1beab63f84 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sun, 22 Mar 2009 03:31:17 -0500 Subject: [PATCH 361/413] tracing: add per-subsystem filtering This patch adds per-subsystem filtering to the event tracing subsystem. It adds a 'filter' debugfs file to each subsystem directory. This file can be written to to set filters; reading from it will display the current set of filters set for that subsystem. Basically what it does is propagate the filter down to each event contained in the subsystem. If a particular event doesn't have a field with the name specified in the filter, it simply doesn't get set for that event. You can verify whether or not the filter was set for a particular event by looking at the filter file for that event. As with per-event filters, compound expressions are supported, echoing '0' to the subsystem's filter file clears all filters in the subsystem, etc. Signed-off-by: Tom Zanussi Acked-by: Frederic Weisbecker LKML-Reference: <1237710677.7703.49.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace.h | 15 ++++++ kernel/trace/trace_events.c | 86 +++++++++++++++++++++++++++--- kernel/trace/trace_events_filter.c | 80 +++++++++++++++++++++++++++ 3 files changed, 175 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index d9eb39e4bb38..f267723c3c52 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -804,6 +804,18 @@ struct ftrace_event_call { #endif }; +struct event_subsystem { + struct list_head list; + const char *name; + struct dentry *entry; + struct filter_pred **preds; +}; + +#define events_for_each(event) \ + for (event = __start_ftrace_events; \ + (unsigned long)event < (unsigned long)__stop_ftrace_events; \ + event++) + #define MAX_FILTER_PRED 8 struct filter_pred; @@ -832,6 +844,9 @@ extern int filter_add_pred(struct ftrace_event_call *call, struct filter_pred *pred); extern void filter_free_preds(struct ftrace_event_call *call); extern int filter_match_preds(struct ftrace_event_call *call, void *rec); +extern void filter_free_subsystem_preds(struct event_subsystem *system); +extern int filter_add_subsystem_pred(struct event_subsystem *system, + struct filter_pred *pred); void event_trace_printk(unsigned long ip, const char *fmt, ...); extern struct ftrace_event_call __start_ftrace_events[]; diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 97470c48956e..97d4daaddd9a 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -524,6 +524,71 @@ event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, return cnt; } +static ssize_t +subsystem_filter_read(struct file *filp, char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct event_subsystem *system = filp->private_data; + struct trace_seq *s; + int r; + + if (*ppos) + return 0; + + s = kmalloc(sizeof(*s), GFP_KERNEL); + if (!s) + return -ENOMEM; + + trace_seq_init(s); + + r = filter_print_preds(system->preds, s->buffer); + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, r); + + kfree(s); + + return r; +} + +static ssize_t +subsystem_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, + loff_t *ppos) +{ + struct event_subsystem *system = filp->private_data; + char buf[64], *pbuf = buf; + struct filter_pred *pred; + int err; + + if (cnt >= sizeof(buf)) + return -EINVAL; + + if (copy_from_user(&buf, ubuf, cnt)) + return -EFAULT; + + pred = kzalloc(sizeof(*pred), GFP_KERNEL); + if (!pred) + return -ENOMEM; + + err = filter_parse(&pbuf, pred); + if (err < 0) { + filter_free_pred(pred); + return err; + } + + if (pred->clear) { + filter_free_subsystem_preds(system); + return cnt; + } + + if (filter_add_subsystem_pred(system, pred)) { + filter_free_pred(pred); + return -EINVAL; + } + + *ppos += cnt; + + return cnt; +} + static const struct seq_operations show_event_seq_ops = { .start = t_start, .next = t_next, @@ -575,6 +640,12 @@ static const struct file_operations ftrace_event_filter_fops = { .write = event_filter_write, }; +static const struct file_operations ftrace_subsystem_filter_fops = { + .open = tracing_open_generic, + .read = subsystem_filter_read, + .write = subsystem_filter_write, +}; + static struct dentry *event_trace_events_dir(void) { static struct dentry *d_tracer; @@ -595,18 +666,13 @@ static struct dentry *event_trace_events_dir(void) return d_events; } -struct event_subsystem { - struct list_head list; - const char *name; - struct dentry *entry; -}; - static LIST_HEAD(event_subsystems); static struct dentry * event_subsystem_dir(const char *name, struct dentry *d_events) { struct event_subsystem *system; + struct dentry *entry; /* First see if we did not already create this dir */ list_for_each_entry(system, &event_subsystems, list) { @@ -633,6 +699,14 @@ event_subsystem_dir(const char *name, struct dentry *d_events) system->name = name; list_add(&system->list, &event_subsystems); + system->preds = NULL; + + entry = debugfs_create_file("filter", 0444, system->entry, system, + &ftrace_subsystem_filter_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/filter' entry\n", name); + return system->entry; } diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 8e8c5fa25be9..1ab20cee0e4c 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -181,6 +181,27 @@ void filter_free_preds(struct ftrace_event_call *call) } } +void filter_free_subsystem_preds(struct event_subsystem *system) +{ + struct ftrace_event_call *call = __start_ftrace_events; + int i; + + if (system->preds) { + for (i = 0; i < MAX_FILTER_PRED; i++) + filter_free_pred(system->preds[i]); + kfree(system->preds); + system->preds = NULL; + } + + events_for_each(call) { + if (!call->name || !call->regfunc) + continue; + + if (!strcmp(call->system, system->name)) + filter_free_preds(call); + } +} + static int __filter_add_pred(struct ftrace_event_call *call, struct filter_pred *pred) { @@ -250,6 +271,65 @@ int filter_add_pred(struct ftrace_event_call *call, struct filter_pred *pred) return __filter_add_pred(call, pred); } +static struct filter_pred *copy_pred(struct filter_pred *pred) +{ + struct filter_pred *new_pred = kmalloc(sizeof(*pred), GFP_KERNEL); + if (!new_pred) + return NULL; + + memcpy(new_pred, pred, sizeof(*pred)); + if (pred->str_val) { + new_pred->str_val = kstrdup(pred->str_val, GFP_KERNEL); + new_pred->field_name = kstrdup(pred->field_name, GFP_KERNEL); + if (!new_pred->str_val) { + kfree(new_pred); + return NULL; + } + } + + return new_pred; +} + +int filter_add_subsystem_pred(struct event_subsystem *system, + struct filter_pred *pred) +{ + struct ftrace_event_call *call = __start_ftrace_events; + struct filter_pred *event_pred; + int i; + + if (system->preds && !pred->compound) + filter_free_subsystem_preds(system); + + if (!system->preds) { + system->preds = kzalloc(MAX_FILTER_PRED * sizeof(pred), + GFP_KERNEL); + if (!system->preds) + return -ENOMEM; + } + + for (i = 0; i < MAX_FILTER_PRED; i++) { + if (!system->preds[i]) { + system->preds[i] = pred; + break; + } + if (i == MAX_FILTER_PRED - 1) + return -EINVAL; + } + + events_for_each(call) { + if (!call->name || !call->regfunc) + continue; + + if (!strcmp(call->system, system->name)) { + event_pred = copy_pred(pred); + if (event_pred) + filter_add_pred(call, event_pred); + } + } + + return 0; +} + int filter_parse(char **pbuf, struct filter_pred *pred) { char *tmp, *tok, *val_str = NULL; From fe9f57f250ab4d781b99504caeb218ca2db14c1a Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Sun, 22 Mar 2009 18:41:59 +0100 Subject: [PATCH 362/413] tracing: add run-time field descriptions for event filtering, kfree fix Impact: fix potential kfree of random data in (rare) failure path Zero-initialize the field structure. Reported-by: Frederic Weisbecker Cc: Tom Zanussi LKML-Reference: <1237710639.7703.46.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 97d4daaddd9a..594d78aaa185 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -24,26 +24,31 @@ int trace_define_field(struct ftrace_event_call *call, char *type, { struct ftrace_event_field *field; - field = kmalloc(sizeof(*field), GFP_KERNEL); + field = kzalloc(sizeof(*field), GFP_KERNEL); if (!field) goto err; + field->name = kstrdup(name, GFP_KERNEL); if (!field->name) goto err; + field->type = kstrdup(type, GFP_KERNEL); if (!field->type) goto err; + field->offset = offset; field->size = size; list_add(&field->link, &call->fields); return 0; + err: if (field) { kfree(field->name); kfree(field->type); } kfree(field); + return -ENOMEM; } From 9bd7d099ab3f10dd666da399c064999bae427cd9 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 22 Mar 2009 23:10:43 +0100 Subject: [PATCH 363/413] tracing/events: make the filter files writable We need the filter files to be writable, the current filter file permissions are only set readable. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt Cc: Tom Zanussi LKML-Reference: <1237759847-21025-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 594d78aaa185..19f61dd23219 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -706,7 +706,7 @@ event_subsystem_dir(const char *name, struct dentry *d_events) system->preds = NULL; - entry = debugfs_create_file("filter", 0444, system->entry, system, + entry = debugfs_create_file("filter", 0644, system->entry, system, &ftrace_subsystem_filter_fops); if (!entry) pr_warning("Could not create debugfs " @@ -769,7 +769,7 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) } } - entry = debugfs_create_file("filter", 0444, call->dir, call, + entry = debugfs_create_file("filter", 0644, call->dir, call, &ftrace_event_filter_fops); if (!entry) pr_warning("Could not create debugfs " From 07edf7121374609709ef1b0889f6e7b8d6a62ec1 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 22 Mar 2009 23:10:46 +0100 Subject: [PATCH 364/413] tracing/events: don't use wake up for events Impact: fix hard-lockup with sched switch events Some ftrace events, such as sched wakeup, can be traced while the runqueue lock is hold. Since they are using trace_current_buffer_unlock_commit(), they call wake_up() which can try to grab the runqueue lock too, resulting in a deadlock. Now for all event, we call a new helper: trace_nowake_buffer_unlock_commit() which do pretty the same than trace_current_buffer_unlock_commit() except than it doesn't call trace_wake_up(). Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1237759847-21025-4-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 26 +++++++++++++++++++++----- kernel/trace/trace.h | 2 ++ kernel/trace/trace_events_stage_3.h | 2 +- 3 files changed, 24 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index e6fac0ffe6f0..6bad12819eb6 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -860,15 +860,25 @@ static void ftrace_trace_stack(struct trace_array *tr, static void ftrace_trace_userstack(struct trace_array *tr, unsigned long flags, int pc); -void trace_buffer_unlock_commit(struct trace_array *tr, - struct ring_buffer_event *event, - unsigned long flags, int pc) +static inline void __trace_buffer_unlock_commit(struct trace_array *tr, + struct ring_buffer_event *event, + unsigned long flags, int pc, + int wake) { ring_buffer_unlock_commit(tr->buffer, event); ftrace_trace_stack(tr, flags, 6, pc); ftrace_trace_userstack(tr, flags, pc); - trace_wake_up(); + + if (wake) + trace_wake_up(); +} + +void trace_buffer_unlock_commit(struct trace_array *tr, + struct ring_buffer_event *event, + unsigned long flags, int pc) +{ + __trace_buffer_unlock_commit(tr, event, flags, pc, 1); } struct ring_buffer_event * @@ -882,7 +892,13 @@ trace_current_buffer_lock_reserve(unsigned char type, unsigned long len, void trace_current_buffer_unlock_commit(struct ring_buffer_event *event, unsigned long flags, int pc) { - return trace_buffer_unlock_commit(&global_trace, event, flags, pc); + return __trace_buffer_unlock_commit(&global_trace, event, flags, pc, 1); +} + +void trace_nowake_buffer_unlock_commit(struct ring_buffer_event *event, + unsigned long flags, int pc) +{ + return __trace_buffer_unlock_commit(&global_trace, event, flags, pc, 0); } void diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index f267723c3c52..54fd9bcd0a65 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -483,6 +483,8 @@ trace_current_buffer_lock_reserve(unsigned char type, unsigned long len, unsigned long flags, int pc); void trace_current_buffer_unlock_commit(struct ring_buffer_event *event, unsigned long flags, int pc); +void trace_nowake_buffer_unlock_commit(struct ring_buffer_event *event, + unsigned long flags, int pc); struct trace_entry *tracing_get_trace_entry(struct trace_array *tr, struct trace_array_cpu *data); diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index ebf215e87d5e..9a3bd49b52e5 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -222,7 +222,7 @@ static void ftrace_raw_event_##call(proto) \ \ assign; \ \ - trace_current_buffer_unlock_commit(event, irq_flags, pc); \ + trace_nowake_buffer_unlock_commit(event, irq_flags, pc); \ \ if (call->preds && !filter_match_preds(call, entry)) \ ring_buffer_event_discard(event); \ From 7e6ea92df3fd7cbe74e7985c6f3e40255c44b201 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 22 Mar 2009 23:10:47 +0100 Subject: [PATCH 365/413] tracing/ftrace: make nop-tracer use polling wait for events on pipe Impact: display events when they arrive Now that the events don't use wake_up() anymore, we need the nop tracer to poll waiting for events on the pipe. Especially because nop is useful to look at orphan traces types (traces types that don't rely on specific tracers) because it doesn't produce traces itself. And unlike other tracers that trigger specific traces periodically, nop triggers no traces by itself that can wake him. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1237759847-21025-5-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_nop.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/trace/trace_nop.c b/kernel/trace/trace_nop.c index 9aa84bde23cd..394f94417e2f 100644 --- a/kernel/trace/trace_nop.c +++ b/kernel/trace/trace_nop.c @@ -91,6 +91,7 @@ struct tracer nop_trace __read_mostly = .name = "nop", .init = nop_trace_init, .reset = nop_trace_reset, + .wait_pipe = poll_wait_pipe, #ifdef CONFIG_FTRACE_SELFTEST .selftest = trace_selftest_startup_nop, #endif From b118415bfad6d75792a85ac999e25149db8e6919 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 23 Mar 2009 00:18:39 +0100 Subject: [PATCH 366/413] tracing/events: don't discard an event after commit When we want to filter an event, the filter test is done after the event is commited to the ring-buffer to be discarded later if needed. But a reader could be reading this event while we are trying to discard it. Other kind of racy events can even happen because the event is commited and can be read and/or consumed. What we want is to discard the event before committing it. Reported-by: Steven Rostedt Signed-off-by: Frederic Weisbecker Cc: Tom Zanussi LKML-Reference: <1237763919-21505-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_stage_3.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_events_stage_3.h b/kernel/trace/trace_events_stage_3.h index 9a3bd49b52e5..9d2fa78cecca 100644 --- a/kernel/trace/trace_events_stage_3.h +++ b/kernel/trace/trace_events_stage_3.h @@ -222,10 +222,11 @@ static void ftrace_raw_event_##call(proto) \ \ assign; \ \ - trace_nowake_buffer_unlock_commit(event, irq_flags, pc); \ - \ if (call->preds && !filter_match_preds(call, entry)) \ ring_buffer_event_discard(event); \ + \ + trace_nowake_buffer_unlock_commit(event, irq_flags, pc); \ + \ } \ \ static int ftrace_raw_reg_event_##call(void) \ From 75c8b417526529d0a7072e4d93ec99dbd483a6f4 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Mon, 23 Mar 2009 03:26:28 -0500 Subject: [PATCH 367/413] tracing/filters: use list_for_each_entry_safe Impact: cleanup Use list_for_each_entry_safe instead of list_for_each_entry in find_event_field(). Reported-by: Frederic Weisbecker Signed-off-by: Tom Zanussi Cc: Steven Rostedt LKML-Reference: <1237796788.7527.35.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_filter.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 1ab20cee0e4c..c4a413b70f78 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -147,11 +147,9 @@ int filter_print_preds(struct filter_pred **preds, char *buf) static struct ftrace_event_field * find_event_field(struct ftrace_event_call *call, char *name) { - struct ftrace_event_field *field; - struct list_head *entry, *tmp; + struct ftrace_event_field *field, *next; - list_for_each_safe(entry, tmp, &call->fields) { - field = list_entry(entry, struct ftrace_event_field, link); + list_for_each_entry_safe(field, next, &call->fields, link) { if (!strcmp(field->name, name)) return field; } From ee6cdabc820a29bd607f38d9cb335c3ceddc673b Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Mon, 23 Mar 2009 03:26:42 -0500 Subject: [PATCH 368/413] tracing/filters: fix bug in copy_pred() Impact: fix potential crash on subsystem filter expression freeing When making a copy of the predicate, pred->field_name needs to be duplicated in the copy as well, otherwise bad things can happen due to later multiple frees of the same string. This affects only per-subsystem event filtering. Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237796802.7527.39.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_filter.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index c4a413b70f78..fd01d8022ad1 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -276,11 +276,19 @@ static struct filter_pred *copy_pred(struct filter_pred *pred) return NULL; memcpy(new_pred, pred, sizeof(*pred)); + + if (pred->field_name) { + new_pred->field_name = kstrdup(pred->field_name, GFP_KERNEL); + if (!new_pred->field_name) { + kfree(new_pred); + return NULL; + } + } + if (pred->str_val) { new_pred->str_val = kstrdup(pred->str_val, GFP_KERNEL); - new_pred->field_name = kstrdup(pred->field_name, GFP_KERNEL); if (!new_pred->str_val) { - kfree(new_pred); + filter_free_pred(new_pred); return NULL; } } From c4cff064be678f1e8344d907499f2a81282edc19 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Mon, 23 Mar 2009 03:26:48 -0500 Subject: [PATCH 369/413] tracing/filters: clean up filter_add_subsystem_pred() Impact: cleanup, memory leak fix This patch cleans up filter_add_subsystem_pred(): - searches for the field before creating a copy of the pred - fixes memory leak in the case a predicate isn't applied - if -ENOMEM, makes sure there's no longer a reference to the pred so the caller can free the half-finished filter - changes the confusing i == MAX_FILTER_PRED - 1 comparison previously remarked upon This affects only per-subsystem event filtering. Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237796808.7527.40.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 1 + kernel/trace/trace_events_filter.c | 31 +++++++++++++++++++++++------- 2 files changed, 25 insertions(+), 7 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 19f61dd23219..fdab30d6c835 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -585,6 +585,7 @@ subsystem_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, } if (filter_add_subsystem_pred(system, pred)) { + filter_free_subsystem_preds(system); filter_free_pred(pred); return -EINVAL; } diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index fd01d8022ad1..4117c2ebc245 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -318,22 +318,39 @@ int filter_add_subsystem_pred(struct event_subsystem *system, system->preds[i] = pred; break; } - if (i == MAX_FILTER_PRED - 1) - return -EINVAL; } + if (i == MAX_FILTER_PRED) + return -EINVAL; + events_for_each(call) { + int err; + if (!call->name || !call->regfunc) continue; - if (!strcmp(call->system, system->name)) { - event_pred = copy_pred(pred); - if (event_pred) - filter_add_pred(call, event_pred); - } + if (strcmp(call->system, system->name)) + continue; + + if (!find_event_field(call, pred->field_name)) + continue; + + event_pred = copy_pred(pred); + if (!event_pred) + goto oom; + + err = filter_add_pred(call, event_pred); + if (err) + filter_free_pred(event_pred); + if (err == -ENOMEM) + goto oom; } return 0; + +oom: + system->preds[i] = NULL; + return -ENOMEM; } int filter_parse(char **pbuf, struct filter_pred *pred) From c0f92ba99bdeaf35f9c580291b4e1a657c67fbd4 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 22 Mar 2009 23:10:44 +0100 Subject: [PATCH 370/413] debugfs: function to know if debugfs is initialized Impact: add new debugfs API With ftrace, some tracers are registered in early initcalls and attempt to create files on the debugfs filesystem. Depending on when they are activated, they can try to create their file at any time. Some checks can be done on the tracing area but providing a helper to know if debugfs is registered make it really more easy. Signed-off-by: Frederic Weisbecker Acked-by: Greg Kroah-Hartman Cc: Steven Rostedt LKML-Reference: <1237759847-21025-2-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- fs/debugfs/inode.c | 16 ++++++++++++++++ include/linux/debugfs.h | 8 ++++++++ 2 files changed, 24 insertions(+) diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c index 81ae9ea3c6e1..0662ba6de85a 100644 --- a/fs/debugfs/inode.c +++ b/fs/debugfs/inode.c @@ -30,6 +30,7 @@ static struct vfsmount *debugfs_mount; static int debugfs_mount_count; +static bool debugfs_registered; static struct inode *debugfs_get_inode(struct super_block *sb, int mode, dev_t dev) { @@ -496,6 +497,16 @@ struct dentry *debugfs_rename(struct dentry *old_dir, struct dentry *old_dentry, } EXPORT_SYMBOL_GPL(debugfs_rename); +/** + * debugfs_initialized - Tells whether debugfs has been registered + */ +bool debugfs_initialized(void) +{ + return debugfs_registered; +} +EXPORT_SYMBOL_GPL(debugfs_initialized); + + static struct kobject *debug_kobj; static int __init debugfs_init(void) @@ -509,11 +520,16 @@ static int __init debugfs_init(void) retval = register_filesystem(&debug_fs_type); if (retval) kobject_put(debug_kobj); + else + debugfs_registered = true; + return retval; } static void __exit debugfs_exit(void) { + debugfs_registered = false; + simple_release_fs(&debugfs_mount, &debugfs_mount_count); unregister_filesystem(&debug_fs_type); kobject_put(debug_kobj); diff --git a/include/linux/debugfs.h b/include/linux/debugfs.h index af0e01d4c663..eb5c2ba2f81a 100644 --- a/include/linux/debugfs.h +++ b/include/linux/debugfs.h @@ -71,6 +71,9 @@ struct dentry *debugfs_create_bool(const char *name, mode_t mode, struct dentry *debugfs_create_blob(const char *name, mode_t mode, struct dentry *parent, struct debugfs_blob_wrapper *blob); + +bool debugfs_initialized(void); + #else #include @@ -183,6 +186,11 @@ static inline struct dentry *debugfs_create_blob(const char *name, mode_t mode, return ERR_PTR(-ENODEV); } +static inline bool debugfs_initialized(void) +{ + return false; +} + #endif #endif From 3e1f60b80cafcb5d7e8d3665b35962fbb8fb9efa Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sun, 22 Mar 2009 23:10:45 +0100 Subject: [PATCH 371/413] tracing/ftrace: check if debugfs is registered before creating files Impact: fix a crash with ftrace={nop,boot} parameter If the nop or initcall tracers are launched as boot tracers, they will attempt to create their option directory and files. But these tracers are registered very early and then assigned as "boot tracers" very early if asked to. Since they do this before debugfs has been registered (core initcall), a crash is triggered. Another early tracers could also come later. So we fix it by checking if debugfs is initialized before creating the root tracing directory. Signed-off-by: Frederic Weisbecker Cc: Greg Kroah-Hartman Cc: Steven Rostedt LKML-Reference: <1237759847-21025-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index ace685c70186..f0e1337b1ebd 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3513,6 +3513,9 @@ struct dentry *tracing_init_dentry(void) if (d_tracer) return d_tracer; + if (!debugfs_initialized()) + return NULL; + d_tracer = debugfs_create_dir("tracing", NULL); if (!d_tracer && !once) { From 45b9560895b07a4a09d55d49235c984db512c5aa Mon Sep 17 00:00:00 2001 From: Anton Vorontsov Date: Tue, 24 Mar 2009 01:07:24 +0300 Subject: [PATCH 372/413] tracing: Fix TRACING_SUPPORT dependency for PPC32 commit 40ada30f9621fbd831ac2437b9a2a399aa ("tracing: clean up menu"), despite the "clean up" in its purpose, introduced a behavioural change for Kconfig symbols: we no longer able to select tracing support on PPC32 (because IRQFLAGS_SUPPORT isn't yet implemented). The IRQFLAGS_SUPPORT is not mandatory for most tracers, tracing core has a special case for platforms w/o irqflags (which, by the way, has become useless as of the commit above). Though according to Ingo Molnar, there was periodic build failures on weird, unmaintained architectures that had no irqflags-tracing support and hence didn't know the raw_irqs_save/restore primitives. Thus we'd better not enable irqflags-less tracing for all architectures. This patch restores the old behaviour for PPC32, and thus brings the tracing back. Other architectures can either add themselves to the exception list or (better) implement TRACE_IRQFLAGS_SUPPORT. Signed-off-by: Anton Vorontsov Acked-b: Steven Rostedt Cc: linuxppc-dev@ozlabs.org LKML-Reference: <20090323220724.GA9851@oksana.dev.rtsoft.ru> Signed-off-by: Ingo Molnar --- kernel/trace/Kconfig | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index b0a46f889659..8a4d72931042 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -63,7 +63,11 @@ config TRACING # config TRACING_SUPPORT bool - depends on TRACE_IRQFLAGS_SUPPORT + # PPC32 has no irqflags tracing support, but it can use most of the + # tracers anyway, they were tested to build and work. Note that new + # exceptions to this list aren't welcomed, better implement the + # irqflags tracing for your architecture. + depends on TRACE_IRQFLAGS_SUPPORT || PPC32 depends on STACKTRACE_SUPPORT default y From 1618536961d31f9b3f55767b22d4a897f4204c26 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Mon, 23 Mar 2009 22:17:01 +0100 Subject: [PATCH 373/413] tracing/function-graph-tracer: fix functions call traces imbalance Impact: fix traces output Sometimes one can observe an imbalance in the traces between function calls and function return traces: func1() { } } The curly brace inside func1() is the return of another function nested inside func1. The return trace have been inserted in the buffer but not the entry. We are storing a return address on the function traces stack while we haven't inserted its entry on the buffer, hence the imbalance on the traces. This is because the tracers doesn't check all failures that can happen on buffer insertion. This patch reports the tracing recursion failures and the ring buffer failures. In such cases, we now restore the original return address for the function, giving up its return trace. Signed-off-by: Frederic Weisbecker Cc: Steven Rostedt LKML-Reference: <1237843021-11695-1-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 6bad12819eb6..89f0c2544ad0 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -924,7 +924,7 @@ trace_function(struct trace_array *tr, } #ifdef CONFIG_FUNCTION_GRAPH_TRACER -static void __trace_graph_entry(struct trace_array *tr, +static int __trace_graph_entry(struct trace_array *tr, struct ftrace_graph_ent *trace, unsigned long flags, int pc) @@ -933,15 +933,17 @@ static void __trace_graph_entry(struct trace_array *tr, struct ftrace_graph_ent_entry *entry; if (unlikely(local_read(&__get_cpu_var(ftrace_cpu_disabled)))) - return; + return 0; event = trace_buffer_lock_reserve(&global_trace, TRACE_GRAPH_ENT, sizeof(*entry), flags, pc); if (!event) - return; + return 0; entry = ring_buffer_event_data(event); entry->graph_ent = *trace; ring_buffer_unlock_commit(global_trace.buffer, event); + + return 1; } static void __trace_graph_return(struct trace_array *tr, @@ -1162,6 +1164,7 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) struct trace_array_cpu *data; unsigned long flags; long disabled; + int ret; int cpu; int pc; @@ -1177,15 +1180,18 @@ int trace_graph_entry(struct ftrace_graph_ent *trace) disabled = atomic_inc_return(&data->disabled); if (likely(disabled == 1)) { pc = preempt_count(); - __trace_graph_entry(tr, trace, flags, pc); + ret = __trace_graph_entry(tr, trace, flags, pc); + } else { + ret = 0; } /* Only do the atomic if it is not already set */ if (!test_tsk_trace_graph(current)) set_tsk_trace_graph(current); + atomic_dec(&data->disabled); local_irq_restore(flags); - return 1; + return ret; } void trace_graph_return(struct ftrace_graph_ret *trace) From 1fc2d5c11918082536acf261ce6abb1f5511053f Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 24 Mar 2009 02:14:01 -0500 Subject: [PATCH 374/413] tracing/filters: use list_for_each_entry Impact: cleanup No need to use the safe version here, so use list_for_each_entry instead of list_for_each_entry_safe in find_event_field(). Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237878841.8339.57.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_filter.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 4117c2ebc245..3f0b79f8a4bc 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -147,9 +147,9 @@ int filter_print_preds(struct filter_pred **preds, char *buf) static struct ftrace_event_field * find_event_field(struct ftrace_event_call *call, char *name) { - struct ftrace_event_field *field, *next; + struct ftrace_event_field *field; - list_for_each_entry_safe(field, next, &call->fields, link) { + list_for_each_entry(field, &call->fields, link) { if (!strcmp(field->name, name)) return field; } From 09f1f245c79585383de63e3ca54d0f91824bff3a Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 24 Mar 2009 02:14:11 -0500 Subject: [PATCH 375/413] tracing/filters: free pred when clearing filters Impact: fix (small) per trace filter modification memory leak Free the current pred when clearing the filters via the filter files. Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237878851.8339.58.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index fdab30d6c835..a9381384aa9e 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -516,6 +516,7 @@ event_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, if (pred->clear) { filter_free_preds(call); + filter_free_pred(pred); return cnt; } @@ -581,6 +582,7 @@ subsystem_filter_write(struct file *filp, const char __user *ubuf, size_t cnt, if (pred->clear) { filter_free_subsystem_preds(system); + filter_free_pred(pred); return cnt; } From 4bda2d517bfa3ce3d7044e06988cdddae7adffe2 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 24 Mar 2009 02:14:31 -0500 Subject: [PATCH 376/413] tracing/filters: use trace_seq_printf() to print filters Impact: cleanup Instead of just using the trace_seq buffer to print the filters, use trace_seq_printf() as it was intended to be used. Reported-by: Steven Rostedt Signed-off-by: Tom Zanussi Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237878871.8339.59.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace.h | 3 ++- kernel/trace/trace_events.c | 8 ++++---- kernel/trace/trace_events_filter.c | 25 +++++++++---------------- 3 files changed, 15 insertions(+), 21 deletions(-) diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 54fd9bcd0a65..90a848debcba 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -840,7 +840,8 @@ struct filter_pred { int trace_define_field(struct ftrace_event_call *call, char *type, char *name, int offset, int size); extern void filter_free_pred(struct filter_pred *pred); -extern int filter_print_preds(struct filter_pred **preds, char *buf); +extern void filter_print_preds(struct filter_pred **preds, + struct trace_seq *s); extern int filter_parse(char **pbuf, struct filter_pred *pred); extern int filter_add_pred(struct ftrace_event_call *call, struct filter_pred *pred); diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index a9381384aa9e..d132997ab756 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -481,8 +481,8 @@ event_filter_read(struct file *filp, char __user *ubuf, size_t cnt, trace_seq_init(s); - r = filter_print_preds(call->preds, s->buffer); - r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, r); + filter_print_preds(call->preds, s); + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len); kfree(s); @@ -547,8 +547,8 @@ subsystem_filter_read(struct file *filp, char __user *ubuf, size_t cnt, trace_seq_init(s); - r = filter_print_preds(system->preds, s->buffer); - r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, r); + filter_print_preds(system->preds, s); + r = simple_read_from_buffer(ubuf, cnt, ppos, s->buffer, s->len); kfree(s); diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 3f0b79f8a4bc..9fca8bb1c06b 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -24,6 +24,7 @@ #include #include "trace.h" +#include "trace_output.h" static int filter_pred_64(struct filter_pred *pred, void *event) { @@ -108,16 +109,15 @@ int filter_match_preds(struct ftrace_event_call *call, void *rec) return 1; } -int filter_print_preds(struct filter_pred **preds, char *buf) +void filter_print_preds(struct filter_pred **preds, struct trace_seq *s) { - ssize_t this_len = 0; char *field_name; struct filter_pred *pred; int i; if (!preds) { - this_len += sprintf(buf + this_len, "none\n"); - return this_len; + trace_seq_printf(s, "none\n"); + return; } for (i = 0; i < MAX_FILTER_PRED; i++) { @@ -125,23 +125,16 @@ int filter_print_preds(struct filter_pred **preds, char *buf) pred = preds[i]; field_name = pred->field_name; if (i) - this_len += sprintf(buf + this_len, - pred->or ? "|| " : "&& "); - this_len += sprintf(buf + this_len, - "%s ", field_name); - this_len += sprintf(buf + this_len, - pred->not ? "!= " : "== "); + trace_seq_printf(s, pred->or ? "|| " : "&& "); + trace_seq_printf(s, "%s ", field_name); + trace_seq_printf(s, pred->not ? "!= " : "== "); if (pred->str_val) - this_len += sprintf(buf + this_len, - "%s\n", pred->str_val); + trace_seq_printf(s, "%s\n", pred->str_val); else - this_len += sprintf(buf + this_len, - "%llu\n", pred->val); + trace_seq_printf(s, "%llu\n", pred->val); } else break; } - - return this_len; } static struct ftrace_event_field * From 9f58a159d022c8f2533a27708aa267adf4f0e3ce Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 24 Mar 2009 02:14:42 -0500 Subject: [PATCH 377/413] tracing/filters: disallow integer values for string filters and vice versa Impact: fix filter use boundary condition / crash Make sure filters for string fields don't use integer values and vice versa. Getting it wrong can crash the system or produce bogus results. Signed-off-by: Tom Zanussi Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237878882.8339.61.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events_filter.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c index 9fca8bb1c06b..026be412f356 100644 --- a/kernel/trace/trace_events_filter.c +++ b/kernel/trace/trace_events_filter.c @@ -237,9 +237,14 @@ int filter_add_pred(struct ftrace_event_call *call, struct filter_pred *pred) pred->offset = field->offset; if (is_string_field(field->type)) { + if (!pred->str_val) + return -EINVAL; pred->fn = filter_pred_string; pred->str_len = field->size; return __filter_add_pred(call, pred); + } else { + if (pred->str_val) + return -EINVAL; } switch (field->size) { From e4955c9986a27bb47ddeb6cd55803053f547e2e9 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 24 Mar 2009 16:04:37 +0800 Subject: [PATCH 378/413] blktrace: mark ddir_act[] const Impact: cleanup ddir_act and what2act always stay immutable. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo LKML-Reference: <49C89415.5080503@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 108f4f7715a5..1ffcbd4ac45b 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -147,8 +147,8 @@ static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector, /* * Data direction bit lookup */ -static u32 ddir_act[2] __read_mostly = { BLK_TC_ACT(BLK_TC_READ), - BLK_TC_ACT(BLK_TC_WRITE) }; +static const u32 ddir_act[2] = { BLK_TC_ACT(BLK_TC_READ), + BLK_TC_ACT(BLK_TC_WRITE) }; /* The ilog2() calls fall out because they're constant */ #define MASK_TC_BIT(rw, __name) ((rw & (1 << BIO_RW_ ## __name)) << \ @@ -1116,10 +1116,10 @@ static void blk_tracer_reset(struct trace_array *tr) blk_tracer_stop(tr); } -static struct { +static const struct { const char *act[2]; int (*print)(struct trace_seq *s, const struct trace_entry *ent); -} what2act[] __read_mostly = { +} what2act[] = { [__BLK_TA_QUEUE] = {{ "Q", "queue" }, blk_log_generic }, [__BLK_TA_BACKMERGE] = {{ "M", "backmerge" }, blk_log_generic }, [__BLK_TA_FRONTMERGE] = {{ "F", "frontmerge" }, blk_log_generic }, From 65796348e09880e12b97267d39b8857c758440a6 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 24 Mar 2009 16:05:06 +0800 Subject: [PATCH 379/413] blktrace: fix wrong calculation of RWBS Impact: fix the output of IO type category characters Trace categories are the upper 16 bits, not the lower 16 bits. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo LKML-Reference: <49C89432.8010805@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 1ffcbd4ac45b..9af41430ee54 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -922,23 +922,24 @@ static void blk_unregister_tracepoints(void) static void fill_rwbs(char *rwbs, const struct blk_io_trace *t) { int i = 0; + int tc = t->action >> BLK_TC_SHIFT; - if (t->action & BLK_TC_DISCARD) + if (tc & BLK_TC_DISCARD) rwbs[i++] = 'D'; - else if (t->action & BLK_TC_WRITE) + else if (tc & BLK_TC_WRITE) rwbs[i++] = 'W'; else if (t->bytes) rwbs[i++] = 'R'; else rwbs[i++] = 'N'; - if (t->action & BLK_TC_AHEAD) + if (tc & BLK_TC_AHEAD) rwbs[i++] = 'A'; - if (t->action & BLK_TC_BARRIER) + if (tc & BLK_TC_BARRIER) rwbs[i++] = 'B'; - if (t->action & BLK_TC_SYNC) + if (tc & BLK_TC_SYNC) rwbs[i++] = 'S'; - if (t->action & BLK_TC_META) + if (tc & BLK_TC_META) rwbs[i++] = 'M'; rwbs[i] = '\0'; From e0dc81bec0927fa0c8aabc521793161909eef7a5 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 24 Mar 2009 16:05:51 +0800 Subject: [PATCH 380/413] blktrace: fix t_error() Impact: fix error flag output t_error() should return t->error but not t->sector. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo LKML-Reference: <49C8945F.5020802@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 9af41430ee54..f69f8bd8bef9 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -968,7 +968,7 @@ static inline unsigned long long t_sector(const struct trace_entry *ent) static inline __u16 t_error(const struct trace_entry *ent) { - return te_blk_io_trace(ent)->sector; + return te_blk_io_trace(ent)->error; } static __u64 get_pdu_int(const struct trace_entry *ent) From 093419971e03362a00f499960557c119982ea09f Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 24 Mar 2009 17:43:30 +0800 Subject: [PATCH 381/413] blktrace: print human-readable act_mask Impact: new feature, allow symbolic values in /debug/tracing/act_mask Print stringified act_mask instead of hex value: # cat act_mask read,write,barrier,sync,queue,requeue,issue,complete,fs,pc,ahead,meta, discard,drv_data # echo "meta,write" > act_mask # cat act_mask write,meta Also: - make act_mask accept "ahead", "meta", "discard" and "drv_data" - use strsep() instead of strchr() to parse user input - return -EINVAL if a token is not found in the mask map - fix a bug that 'value' is unsigned, so it can < 0 - propagate error value of blk_trace_mask2str() to userspace, but not always return -ENXIO. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo LKML-Reference: <49C8AB42.1000802@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 103 +++++++++++++++++++++++++--------------- 1 file changed, 65 insertions(+), 38 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index f69f8bd8bef9..6fb274f5f34e 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1316,53 +1316,77 @@ struct attribute_group blk_trace_attr_group = { .attrs = blk_trace_attrs, }; -static int blk_str2act_mask(const char *str) +static const struct { + int mask; + const char *str; +} mask_maps[] = { + { BLK_TC_READ, "read" }, + { BLK_TC_WRITE, "write" }, + { BLK_TC_BARRIER, "barrier" }, + { BLK_TC_SYNC, "sync" }, + { BLK_TC_QUEUE, "queue" }, + { BLK_TC_REQUEUE, "requeue" }, + { BLK_TC_ISSUE, "issue" }, + { BLK_TC_COMPLETE, "complete" }, + { BLK_TC_FS, "fs" }, + { BLK_TC_PC, "pc" }, + { BLK_TC_AHEAD, "ahead" }, + { BLK_TC_META, "meta" }, + { BLK_TC_DISCARD, "discard" }, + { BLK_TC_DRV_DATA, "drv_data" }, +}; + +static int blk_trace_str2mask(const char *str) { + int i; int mask = 0; - char *copy = kstrdup(str, GFP_KERNEL), *s; + char *s, *token; - if (copy == NULL) + s = kstrdup(str, GFP_KERNEL); + if (s == NULL) return -ENOMEM; - - s = strstrip(copy); + s = strstrip(s); while (1) { - char *sep = strchr(s, ','); - - if (sep != NULL) - *sep = '\0'; - - if (strcasecmp(s, "barrier") == 0) - mask |= BLK_TC_BARRIER; - else if (strcasecmp(s, "complete") == 0) - mask |= BLK_TC_COMPLETE; - else if (strcasecmp(s, "fs") == 0) - mask |= BLK_TC_FS; - else if (strcasecmp(s, "issue") == 0) - mask |= BLK_TC_ISSUE; - else if (strcasecmp(s, "pc") == 0) - mask |= BLK_TC_PC; - else if (strcasecmp(s, "queue") == 0) - mask |= BLK_TC_QUEUE; - else if (strcasecmp(s, "read") == 0) - mask |= BLK_TC_READ; - else if (strcasecmp(s, "requeue") == 0) - mask |= BLK_TC_REQUEUE; - else if (strcasecmp(s, "sync") == 0) - mask |= BLK_TC_SYNC; - else if (strcasecmp(s, "write") == 0) - mask |= BLK_TC_WRITE; - - if (sep == NULL) + token = strsep(&s, ","); + if (token == NULL) break; - s = sep + 1; + if (*token == '\0') + continue; + + for (i = 0; i < ARRAY_SIZE(mask_maps); i++) { + if (strcasecmp(token, mask_maps[i].str) == 0) { + mask |= mask_maps[i].mask; + break; + } + } + if (i == ARRAY_SIZE(mask_maps)) { + mask = -EINVAL; + break; + } } - kfree(copy); + kfree(s); return mask; } +static ssize_t blk_trace_mask2str(char *buf, int mask) +{ + int i; + char *p = buf; + + for (i = 0; i < ARRAY_SIZE(mask_maps); i++) { + if (mask & mask_maps[i].mask) { + p += sprintf(p, "%s%s", + (p == buf) ? "" : ",", mask_maps[i].str); + } + } + *p++ = '\n'; + + return p - buf; +} + static struct request_queue *blk_trace_get_queue(struct block_device *bdev) { if (bdev->bd_disk == NULL) @@ -1399,7 +1423,7 @@ static ssize_t sysfs_blk_trace_attr_show(struct device *dev, if (q->blk_trace == NULL) ret = sprintf(buf, "disabled\n"); else if (attr == &dev_attr_act_mask) - ret = sprintf(buf, "%#x\n", q->blk_trace->act_mask); + ret = blk_trace_mask2str(buf, q->blk_trace->act_mask); else if (attr == &dev_attr_pid) ret = sprintf(buf, "%u\n", q->blk_trace->pid); else if (attr == &dev_attr_start_lba) @@ -1424,7 +1448,7 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev, struct request_queue *q; struct hd_struct *p; u64 value; - ssize_t ret = -ENXIO; + ssize_t ret = -EINVAL; if (count == 0) goto out; @@ -1432,13 +1456,16 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev, if (attr == &dev_attr_act_mask) { if (sscanf(buf, "%llx", &value) != 1) { /* Assume it is a list of trace category names */ - value = blk_str2act_mask(buf); - if (value < 0) + ret = blk_trace_str2mask(buf); + if (ret < 0) goto out; + value = ret; } } else if (sscanf(buf, "%llu", &value) != 1) goto out; + ret = -ENXIO; + lock_kernel(); p = dev_to_part(dev); bdev = bdget(part_devt(p)); From 098335215a4921a8a54193829eaed602dca24df5 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Sat, 21 Mar 2009 02:44:50 -0400 Subject: [PATCH 382/413] tracing: fix memory leak in trace_stat If the function profiler does not have any items recorded and one were to cat the function stat file, the kernel would take a BUG with a NULL pointer dereference. Looking further into this, I found that returning NULL from stat_start did not stop the stat logic, and would later call stat_next. This breaks from the way seq_file works, so I looked into fixing the stat code. This is where I noticed that the last next_entry is never freed. It is allocated, and if the stat_next returns NULL, the code breaks out of the loop, unlocks the mutex and exits. We never link the next_entry nor do we free it. Thus it is a real memory leak. This patch rearranges the code a bit to not only fix the memory leak, but also to act more like seq_file where nothing is printed if there is nothing to print. That is, stat_start returns NULL. Signed-off-by: Steven Rostedt --- kernel/trace/trace_stat.c | 23 +++++++++++++---------- 1 file changed, 13 insertions(+), 10 deletions(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index 39310e3434ee..f71b85b22cfe 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -75,7 +75,7 @@ static int stat_seq_init(struct tracer_stat_session *session) { struct trace_stat_list *iter_entry, *new_entry; struct tracer_stat *ts = session->ts; - void *prev_stat; + void *stat; int ret = 0; int i; @@ -85,6 +85,10 @@ static int stat_seq_init(struct tracer_stat_session *session) if (!ts->stat_cmp) ts->stat_cmp = dummy_cmp; + stat = ts->stat_start(); + if (!stat) + goto exit; + /* * The first entry. Actually this is the second, but the first * one (the stat_list head) is pointless. @@ -99,14 +103,19 @@ static int stat_seq_init(struct tracer_stat_session *session) list_add(&new_entry->list, &session->stat_list); - new_entry->stat = ts->stat_start(); - prev_stat = new_entry->stat; + new_entry->stat = stat; /* * Iterate over the tracer stat entries and store them in a sorted * list. */ for (i = 1; ; i++) { + stat = ts->stat_next(stat, i); + + /* End of insertion */ + if (!stat) + break; + new_entry = kmalloc(sizeof(struct trace_stat_list), GFP_KERNEL); if (!new_entry) { ret = -ENOMEM; @@ -114,11 +123,7 @@ static int stat_seq_init(struct tracer_stat_session *session) } INIT_LIST_HEAD(&new_entry->list); - new_entry->stat = ts->stat_next(prev_stat, i); - - /* End of insertion */ - if (!new_entry->stat) - break; + new_entry->stat = stat; list_for_each_entry(iter_entry, &session->stat_list, list) { @@ -137,8 +142,6 @@ static int stat_seq_init(struct tracer_stat_session *session) break; } } - - prev_stat = new_entry->stat; } exit: mutex_unlock(&session->stat_mutex); From 5d1a03dc541dc6672e60e57249ed22f40654ca47 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 23 Mar 2009 23:38:49 -0400 Subject: [PATCH 383/413] function-graph: moved the timestamp from arch to generic code This patch move the timestamp from happening in the arch specific code into the general code. This allows for better control by the tracer to time manipulation. Signed-off-by: Steven Rostedt --- arch/x86/kernel/ftrace.c | 6 +----- include/linux/ftrace.h | 3 +-- kernel/trace/trace_functions_graph.c | 8 +++++--- 3 files changed, 7 insertions(+), 10 deletions(-) diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index 57b33edb7ce3..61df77532120 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -410,7 +410,6 @@ int ftrace_disable_ftrace_graph_caller(void) void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) { unsigned long old; - unsigned long long calltime; int faulted; struct ftrace_graph_ent trace; unsigned long return_hooker = (unsigned long) @@ -453,10 +452,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) return; } - calltime = trace_clock_local(); - - if (ftrace_push_return_trace(old, calltime, - self_addr, &trace.depth) == -EBUSY) { + if (ftrace_push_return_trace(old, self_addr, &trace.depth) == -EBUSY) { *parent = old; return; } diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index db3fed630db3..1141248c84ee 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -369,8 +369,7 @@ struct ftrace_ret_stack { extern void return_to_handler(void); extern int -ftrace_push_return_trace(unsigned long ret, unsigned long long time, - unsigned long func, int *depth); +ftrace_push_return_trace(unsigned long ret, unsigned long func, int *depth); extern void ftrace_pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret); diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c index e876816fa8e7..d28687e7b3a7 100644 --- a/kernel/trace/trace_functions_graph.c +++ b/kernel/trace/trace_functions_graph.c @@ -57,9 +57,9 @@ static struct tracer_flags tracer_flags = { /* Add a function return address to the trace stack on thread info.*/ int -ftrace_push_return_trace(unsigned long ret, unsigned long long time, - unsigned long func, int *depth) +ftrace_push_return_trace(unsigned long ret, unsigned long func, int *depth) { + unsigned long long calltime; int index; if (!current->ret_stack) @@ -71,11 +71,13 @@ ftrace_push_return_trace(unsigned long ret, unsigned long long time, return -EBUSY; } + calltime = trace_clock_local(); + index = ++current->curr_ret_stack; barrier(); current->ret_stack[index].ret = ret; current->ret_stack[index].func = func; - current->ret_stack[index].calltime = time; + current->ret_stack[index].calltime = calltime; *depth = index; return 0; From 05ce5818adee8f8efd0a5ca0d900a6789012516b Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 24 Mar 2009 00:18:31 -0400 Subject: [PATCH 384/413] function-graph: prevent more than one tracer registering Impact: prevent crash due to multiple function graph tracers The function graph tracer can currently only handle a single tracer being registered. If another tracer registers with the function graph tracer it can crash the system. Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 7847806eefef..c81a759fbf76 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2643,6 +2643,12 @@ int register_ftrace_graph(trace_func_graph_ret_t retfunc, mutex_lock(&ftrace_lock); + /* we currently allow only one tracer registered at a time */ + if (atomic_read(&ftrace_graph_active)) { + ret = -EBUSY; + goto out; + } + ftrace_suspend_notifier.notifier_call = ftrace_suspend_notifier_call; register_pm_notifier(&ftrace_suspend_notifier); From 8aef2d2856158a36c295a8d1288281e4839bff13 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 24 Mar 2009 01:10:15 -0400 Subject: [PATCH 385/413] function-graph: ignore times across schedule Impact: more accurate timings The current method of function graph tracing does not take into account the time spent when a task is not running. This shows functions that call schedule have increased costs: 3) + 18.664 us | } ------------------------------------------ 3) -0 => kblockd-123 ------------------------------------------ 3) | finish_task_switch() { 3) 1.441 us | _spin_unlock_irq(); 3) 3.966 us | } 3) ! 2959.433 us | } 3) ! 2961.465 us | } This patch uses the tracepoint in the scheduling context switch to account for time that has elapsed while a task is scheduled out. Now we see: ------------------------------------------ 3) -0 => edac-po-1067 ------------------------------------------ 3) | finish_task_switch() { 3) 0.685 us | _spin_unlock_irq(); 3) 2.331 us | } 3) + 41.439 us | } 3) + 42.663 us | } Signed-off-by: Steven Rostedt --- include/linux/sched.h | 2 ++ kernel/trace/ftrace.c | 36 ++++++++++++++++++++++++++++++++++++ 2 files changed, 38 insertions(+) diff --git a/include/linux/sched.h b/include/linux/sched.h index 89cd308cc7a5..471e36d30123 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1409,6 +1409,8 @@ struct task_struct { int curr_ret_stack; /* Stack of return addresses for return function tracing */ struct ftrace_ret_stack *ret_stack; + /* time stamp for last schedule */ + unsigned long long ftrace_timestamp; /* * Number of functions that haven't been traced * because of depth overrun. diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index c81a759fbf76..0b90364d1a2c 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -29,6 +29,8 @@ #include #include +#include + #include #include "trace.h" @@ -2590,6 +2592,31 @@ static int alloc_retstack_tasklist(struct ftrace_ret_stack **ret_stack_list) return ret; } +static void +ftrace_graph_probe_sched_switch(struct rq *__rq, struct task_struct *prev, + struct task_struct *next) +{ + unsigned long long timestamp; + int index; + + timestamp = trace_clock_local(); + + prev->ftrace_timestamp = timestamp; + + /* only process tasks that we timestamped */ + if (!next->ftrace_timestamp) + return; + + /* + * Update all the counters in next to make up for the + * time next was sleeping. + */ + timestamp -= next->ftrace_timestamp; + + for (index = next->curr_ret_stack; index >= 0; index--) + next->ret_stack[index].calltime += timestamp; +} + /* Allocate a return stack for each task */ static int start_graph_tracing(void) { @@ -2611,6 +2638,13 @@ static int start_graph_tracing(void) ret = alloc_retstack_tasklist(ret_stack_list); } while (ret == -EAGAIN); + if (!ret) { + ret = register_trace_sched_switch(ftrace_graph_probe_sched_switch); + if (ret) + pr_info("ftrace_graph: Couldn't activate tracepoint" + " probe to kernel_sched_switch\n"); + } + kfree(ret_stack_list); return ret; } @@ -2674,6 +2708,7 @@ void unregister_ftrace_graph(void) mutex_lock(&ftrace_lock); atomic_dec(&ftrace_graph_active); + unregister_trace_sched_switch(ftrace_graph_probe_sched_switch); ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; ftrace_graph_entry = ftrace_graph_entry_stub; ftrace_shutdown(FTRACE_STOP_FUNC_RET); @@ -2694,6 +2729,7 @@ void ftrace_graph_init_task(struct task_struct *t) t->curr_ret_stack = -1; atomic_set(&t->tracing_graph_pause, 0); atomic_set(&t->trace_overrun, 0); + t->ftrace_timestamp = 0; } else t->ret_stack = NULL; } From be6f164a02f394675e2ac2077dd354cebef5b4c0 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 24 Mar 2009 11:06:24 -0400 Subject: [PATCH 386/413] function-graph: add option for include sleep times Impact: give user a choice to show times spent while sleeping The user may want to see the time a function spent sleeping. This patch adds the trace option "sleep-time" to allow that. The "sleep-time" option is default on. echo sleep-time > /debug/tracing/trace_options produces: ------------------------------------------ 2) avahi-d-3428 => -0 ------------------------------------------ 2) | finish_task_switch() { 2) 0.621 us | _spin_unlock_irq(); 2) 2.202 us | } 2) ! 1002.197 us | } 2) ! 1003.521 us | } where as, echo nosleep-time > /debug/tracing/trace_options produces: 0) -0 => yum-upd-3416 ------------------------------------------ 0) | finish_task_switch() { 0) 0.643 us | _spin_unlock_irq(); 0) 2.342 us | } 0) + 41.302 us | } 0) + 42.453 us | } Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 7 +++++++ kernel/trace/trace.c | 3 ++- kernel/trace/trace.h | 1 + 3 files changed, 10 insertions(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 0b90364d1a2c..02d2de9d08ba 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2599,6 +2599,13 @@ ftrace_graph_probe_sched_switch(struct rq *__rq, struct task_struct *prev, unsigned long long timestamp; int index; + /* + * Does the user want to count the time a function was asleep. + * If so, do not update the time stamps. + */ + if (trace_flags & TRACE_ITER_SLEEP_TIME) + return; + timestamp = trace_clock_local(); prev->ftrace_timestamp = timestamp; diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index f0e1337b1ebd..67c6a21dd427 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -255,7 +255,7 @@ static DECLARE_WAIT_QUEUE_HEAD(trace_wait); /* trace_flags holds trace_options default values */ unsigned long trace_flags = TRACE_ITER_PRINT_PARENT | TRACE_ITER_PRINTK | - TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO; + TRACE_ITER_ANNOTATE | TRACE_ITER_CONTEXT_INFO | TRACE_ITER_SLEEP_TIME; /** * trace_wake_up - wake up tasks waiting for trace input @@ -316,6 +316,7 @@ static const char *trace_options[] = { "context-info", "latency-format", "global-clock", + "sleep-time", NULL }; diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 7cfb741be200..d7410bbb9a80 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -683,6 +683,7 @@ enum trace_iterator_flags { TRACE_ITER_CONTEXT_INFO = 0x20000, /* Print pid/cpu/time */ TRACE_ITER_LATENCY_FMT = 0x40000, TRACE_ITER_GLOBAL_CLK = 0x80000, + TRACE_ITER_SLEEP_TIME = 0x100000, }; /* From cc59c9e8d0165c632fd056c4a23e36f917507fb4 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 24 Mar 2009 11:03:01 +0800 Subject: [PATCH 387/413] ftrace: show virtual PID Impact: fix PID output under namespaces When current namespace is not the global namespace, pid read from set_ftrace_pid is no correct. # ~/newpid_namespace_run bash # echo $$ 1 # echo 1 > set_ftrace_pid # cat set_ftrace_pid 3756 Since we write virtual PID to set_ftrace_pid, we need get virtual PID when we read it. Signed-off-by: Lai Jiangshan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <49C84D65.9050606@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 02d2de9d08ba..bb377112b1bb 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2264,7 +2264,7 @@ ftrace_pid_read(struct file *file, char __user *ubuf, if (ftrace_pid_trace == ftrace_swapper_pid) r = sprintf(buf, "swapper tasks\n"); else if (ftrace_pid_trace) - r = sprintf(buf, "%u\n", pid_nr(ftrace_pid_trace)); + r = sprintf(buf, "%u\n", pid_vnr(ftrace_pid_trace)); else r = sprintf(buf, "no pid\n"); From ee000b7f9fe429d2470c674ccec8d344f6789e0d Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Tue, 24 Mar 2009 13:38:06 +0800 Subject: [PATCH 388/413] tracing: use union for multi-usages field Impact: cleanup struct dyn_ftrace::ip has different usages in his lifecycle, we use union for it. And also for struct dyn_ftrace::flags. Signed-off-by: Lai Jiangshan Cc: Steven Rostedt Cc: Frederic Weisbecker LKML-Reference: <49C871BE.3080405@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- include/linux/ftrace.h | 12 +++++++++--- kernel/trace/ftrace.c | 8 ++++---- 2 files changed, 13 insertions(+), 7 deletions(-) diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 1141248c84ee..015a3d22cf74 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -145,9 +145,15 @@ enum { }; struct dyn_ftrace { - unsigned long ip; /* address of mcount call-site */ - unsigned long flags; - struct dyn_arch_ftrace arch; + union { + unsigned long ip; /* address of mcount call-site */ + struct dyn_ftrace *freelist; + }; + union { + unsigned long flags; + struct dyn_ftrace *newlist; + }; + struct dyn_arch_ftrace arch; }; int ftrace_force_update(void); diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index bb377112b1bb..7b8722baf153 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -341,7 +341,7 @@ static inline int record_frozen(struct dyn_ftrace *rec) static void ftrace_free_rec(struct dyn_ftrace *rec) { - rec->ip = (unsigned long)ftrace_free_records; + rec->freelist = ftrace_free_records; ftrace_free_records = rec; rec->flags |= FTRACE_FL_FREE; } @@ -379,7 +379,7 @@ static struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip) return NULL; } - ftrace_free_records = (void *)rec->ip; + ftrace_free_records = rec->freelist; memset(rec, 0, sizeof(*rec)); return rec; } @@ -411,7 +411,7 @@ ftrace_record_ip(unsigned long ip) return NULL; rec->ip = ip; - rec->flags = (unsigned long)ftrace_new_addrs; + rec->newlist = ftrace_new_addrs; ftrace_new_addrs = rec; return rec; @@ -731,7 +731,7 @@ static int ftrace_update_code(struct module *mod) return -1; p = ftrace_new_addrs; - ftrace_new_addrs = (struct dyn_ftrace *)p->flags; + ftrace_new_addrs = p->newlist; p->flags = 0L; /* convert record (i.e, patch mcount-call with NOP) */ From 0a5d649018b151cb9331c213a843ac4a3e7e44ab Mon Sep 17 00:00:00 2001 From: Jody McIntyre Date: Tue, 24 Mar 2009 16:00:28 -0400 Subject: [PATCH 389/413] tracing: Documentation / sample code fixes for tracepoints Fix the tracepoint documentation to refer to "tracepoint-sample" instead of "tracepoint-example" to match what actually exists; fix the directory, and clarify how to compile. Change every instance of "example" in the sample tracepoint code to "sample" for consistency. Signed-off-by: Jody McIntyre Acked-by: Mathieu Desnoyers Cc: torvalds@linux-foundation.org LKML-Reference: <20090324200027.GH8294@clouds> Signed-off-by: Ingo Molnar --- Documentation/tracepoints.txt | 13 +++++++------ samples/tracepoints/tracepoint-sample.c | 24 ++++++++++++------------ 2 files changed, 19 insertions(+), 18 deletions(-) diff --git a/Documentation/tracepoints.txt b/Documentation/tracepoints.txt index 4ff43c6de299..c0e1ceed75a4 100644 --- a/Documentation/tracepoints.txt +++ b/Documentation/tracepoints.txt @@ -103,13 +103,14 @@ used to export the defined tracepoints. * Probe / tracepoint example -See the example provided in samples/tracepoints/src +See the example provided in samples/tracepoints -Compile them with your kernel. +Compile them with your kernel. They are built during 'make' (not +'make modules') when CONFIG_SAMPLE_TRACEPOINTS=m. Run, as root : -modprobe tracepoint-example (insmod order is not important) -modprobe tracepoint-probe-example -cat /proc/tracepoint-example (returns an expected error) -rmmod tracepoint-example tracepoint-probe-example +modprobe tracepoint-sample (insmod order is not important) +modprobe tracepoint-probe-sample +cat /proc/tracepoint-sample (returns an expected error) +rmmod tracepoint-sample tracepoint-probe-sample dmesg diff --git a/samples/tracepoints/tracepoint-sample.c b/samples/tracepoints/tracepoint-sample.c index 68d5dc0310e4..9cf80a11e8b6 100644 --- a/samples/tracepoints/tracepoint-sample.c +++ b/samples/tracepoints/tracepoint-sample.c @@ -1,6 +1,6 @@ /* tracepoint-sample.c * - * Executes a tracepoint when /proc/tracepoint-example is opened. + * Executes a tracepoint when /proc/tracepoint-sample is opened. * * (C) Copyright 2007 Mathieu Desnoyers * @@ -16,7 +16,7 @@ DEFINE_TRACE(subsys_event); DEFINE_TRACE(subsys_eventb); -struct proc_dir_entry *pentry_example; +struct proc_dir_entry *pentry_sample; static int my_open(struct inode *inode, struct file *file) { @@ -32,25 +32,25 @@ static struct file_operations mark_ops = { .open = my_open, }; -static int __init example_init(void) +static int __init sample_init(void) { - printk(KERN_ALERT "example init\n"); - pentry_example = proc_create("tracepoint-example", 0444, NULL, + printk(KERN_ALERT "sample init\n"); + pentry_sample = proc_create("tracepoint-sample", 0444, NULL, &mark_ops); - if (!pentry_example) + if (!pentry_sample) return -EPERM; return 0; } -static void __exit example_exit(void) +static void __exit sample_exit(void) { - printk(KERN_ALERT "example exit\n"); - remove_proc_entry("tracepoint-example", NULL); + printk(KERN_ALERT "sample exit\n"); + remove_proc_entry("tracepoint-sample", NULL); } -module_init(example_init) -module_exit(example_exit) +module_init(sample_init) +module_exit(sample_exit) MODULE_LICENSE("GPL"); MODULE_AUTHOR("Mathieu Desnoyers"); -MODULE_DESCRIPTION("Tracepoint example"); +MODULE_DESCRIPTION("Tracepoint sample"); From e6f489013b985b58d096a3091ece0ed579367232 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 25 Mar 2009 16:27:17 +0800 Subject: [PATCH 390/413] trace_stat: don't call seq_printf() in seq_operation->start() Impact: Fix incorrect way using seq_file's API Use SEQ_START_TOKEN instead of calling ->stat_headers() int seq_operation->start(). Signed-off-by: Lai Jiangshan Acked-by: Steven Rostedt Cc: Alexey Dobriyan LKML-Reference: <49C9EAE5.5070202@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_stat.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index f71b85b22cfe..8c129dd480ad 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -163,7 +163,7 @@ static void *stat_seq_start(struct seq_file *s, loff_t *pos) /* If we are in the beginning of the file, print the headers */ if (!*pos && session->ts->stat_headers) - session->ts->stat_headers(s); + return SEQ_START_TOKEN; return seq_list_start(&session->stat_list, *pos); } @@ -172,6 +172,9 @@ static void *stat_seq_next(struct seq_file *s, void *p, loff_t *pos) { struct tracer_stat_session *session = s->private; + if (p == SEQ_START_TOKEN) + return seq_list_start(&session->stat_list, *pos); + return seq_list_next(p, &session->stat_list, pos); } @@ -186,6 +189,9 @@ static int stat_seq_show(struct seq_file *s, void *v) struct tracer_stat_session *session = s->private; struct trace_stat_list *l = list_entry(v, struct trace_stat_list, list); + if (v == SEQ_START_TOKEN) + return session->ts->stat_headers(s); + return session->ts->stat_show(s, l->stat); } From 220ba351dfa57eca4bec5ce0098a276446a47958 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 25 Mar 2009 16:58:39 +0800 Subject: [PATCH 391/413] trace_stat: keep original order Impact: make trace_stat files show items with the original order trace_stat tracer reverse the items, it makes the output looks a little ugly. Example, when we read trace_stat/workqueues, we get cpu#7's stat. at first, and then cpu#6... cpu#0. Signed-off-by: Lai Jiangshan Acked-by: Steven Rostedt Acked-by: Frederic Weisbecker LKML-Reference: <49C9F23F.5040307@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_stat.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) diff --git a/kernel/trace/trace_stat.c b/kernel/trace/trace_stat.c index 8c129dd480ad..acdebd771a93 100644 --- a/kernel/trace/trace_stat.c +++ b/kernel/trace/trace_stat.c @@ -125,23 +125,21 @@ static int stat_seq_init(struct tracer_stat_session *session) INIT_LIST_HEAD(&new_entry->list); new_entry->stat = stat; - list_for_each_entry(iter_entry, &session->stat_list, list) { + list_for_each_entry_reverse(iter_entry, &session->stat_list, + list) { /* Insertion with a descendent sorting */ - if (ts->stat_cmp(new_entry->stat, - iter_entry->stat) > 0) { + if (ts->stat_cmp(iter_entry->stat, + new_entry->stat) >= 0) { - list_add_tail(&new_entry->list, - &iter_entry->list); - break; - - /* The current smaller value */ - } else if (list_is_last(&iter_entry->list, - &session->stat_list)) { list_add(&new_entry->list, &iter_entry->list); break; } } + + /* The current larger value */ + if (list_empty(&new_entry->list)) + list_add(&new_entry->list, &session->stat_list); } exit: mutex_unlock(&session->stat_mutex); From 2f63b840bc8a816ac879ee773b035cf3e433fae4 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 25 Mar 2009 16:59:18 +0800 Subject: [PATCH 392/413] trace_workqueues: fix empty line's output Empty lines separate cpus stat. After previous fix(trace_stat: keep original order) applied, the empty lines are displayed at incorrect position. Signed-off-by: Lai Jiangshan Acked-by: Steven Rostedt Acked-by: Frederic Weisbecker LKML-Reference: <49C9F266.2060706@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- kernel/trace/trace_workqueue.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/kernel/trace/trace_workqueue.c b/kernel/trace/trace_workqueue.c index 9ab035b58cf1..797201e4a137 100644 --- a/kernel/trace/trace_workqueue.c +++ b/kernel/trace/trace_workqueue.c @@ -196,6 +196,11 @@ static int workqueue_stat_show(struct seq_file *s, void *p) struct pid *pid; struct task_struct *tsk; + spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); + if (&cws->list == workqueue_cpu_stat(cpu)->list.next) + seq_printf(s, "\n"); + spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); + pid = find_get_pid(cws->pid); if (pid) { tsk = get_pid_task(pid, PIDTYPE_PID); @@ -208,18 +213,13 @@ static int workqueue_stat_show(struct seq_file *s, void *p) put_pid(pid); } - spin_lock_irqsave(&workqueue_cpu_stat(cpu)->lock, flags); - if (&cws->list == workqueue_cpu_stat(cpu)->list.next) - seq_printf(s, "\n"); - spin_unlock_irqrestore(&workqueue_cpu_stat(cpu)->lock, flags); - return 0; } static int workqueue_stat_headers(struct seq_file *s) { seq_printf(s, "# CPU INSERTED EXECUTED NAME\n"); - seq_printf(s, "# | | | |\n\n"); + seq_printf(s, "# | | | |\n"); return 0; } From 759ee0915dd713361e72facb78b66600b5712d65 Mon Sep 17 00:00:00 2001 From: Lai Jiangshan Date: Wed, 25 Mar 2009 17:06:30 +0800 Subject: [PATCH 393/413] init,cpuset: fix initialize order Impact: cpuset_wq should be initialized after init_workqueues() When I read /debugfs/tracing/trace_stat/workqueues, I got this: # CPU INSERTED EXECUTED NAME # | | | | 0 0 0 cpuset 0 285 285 events/0 0 2 2 work_on_cpu/0 0 1115 1115 khelper 0 325 325 kblockd/0 0 0 0 kacpid 0 0 0 kacpi_notify 0 0 0 ata/0 0 0 0 ata_aux 0 0 0 ksuspend_usbd 0 0 0 aio/0 0 0 0 nfsiod 0 0 0 kpsmoused 0 0 0 kstriped 0 0 0 kondemand/0 0 1 1 hid_compat 0 0 0 rpciod/0 1 64 64 events/1 1 2 2 work_on_cpu/1 1 5 5 kblockd/1 1 0 0 ata/1 1 0 0 aio/1 1 0 0 kondemand/1 1 0 0 rpciod/1 I found "cpuset" is at the earliest. I found a create_singlethread_workqueue() is earlier than init_workqueues(): kernel_init() ->cpuset_init_smp() ->create_singlethread_workqueue() ->do_basic_setup() ->init_workqueues() I think it's better that create_singlethread_workqueue() is called after workqueue subsystem has been initialized. Signed-off-by: Lai Jiangshan Acked-by: Steven Rostedt Acked-by: Frederic Weisbecker Cc: Paul Menage Cc: miaoxie Cc: Li Zefan Cc: Andrew Morton LKML-Reference: <49C9F416.1050707@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- init/main.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/init/main.c b/init/main.c index 20d784ab5ef8..b0097d2b63ae 100644 --- a/init/main.c +++ b/init/main.c @@ -772,6 +772,7 @@ static void __init do_basic_setup(void) { rcu_init_sched(); /* needed by module_init stage. */ init_workqueues(); + cpuset_init_smp(); usermodehelper_init(); driver_init(); init_irq_proc(); @@ -865,8 +866,6 @@ static int __init kernel_init(void * unused) smp_init(); sched_init_smp(); - cpuset_init_smp(); - do_basic_setup(); /* From fee039a1d05c6e0f71b0fe270d847742a02d56c4 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Mon, 23 Mar 2009 10:14:52 -0400 Subject: [PATCH 394/413] x86: kretprobe-booster interrupt emulation code fix Fix interrupt emulation code in kretprobe-booster according to pt_regs update (es/ds change and gs adding). This issue has been reported on systemtap-bugzilla: http://sources.redhat.com/bugzilla/show_bug.cgi?id=9965 | On a -tip kernel on x86_32, kretprobe_example (from samples) triggers the | following backtrace when its retprobing a class of functions that cause a | copy_from/to_user(). | | BUG: sleeping function called from invalid context at mm/memory.c:3196 | in_atomic(): 0, irqs_disabled(): 1, pid: 2286, name: cat Signed-off-by: Masami Hiramatsu Acked-by: Ananth N Mavinakayanahalli Tested-by: Bharata B Rao Cc: systemtap-ml LKML-Reference: <49C7995C.2010601@redhat.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/kprobes.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c index 4558dd3918cf..759095d53a06 100644 --- a/arch/x86/kernel/kprobes.c +++ b/arch/x86/kernel/kprobes.c @@ -638,13 +638,13 @@ static void __used __kprobes kretprobe_trampoline_holder(void) #else " pushf\n" /* - * Skip cs, ip, orig_ax. + * Skip cs, ip, orig_ax and gs. * trampoline_handler() will plug in these values */ - " subl $12, %esp\n" + " subl $16, %esp\n" " pushl %fs\n" - " pushl %ds\n" " pushl %es\n" + " pushl %ds\n" " pushl %eax\n" " pushl %ebp\n" " pushl %edi\n" @@ -655,10 +655,10 @@ static void __used __kprobes kretprobe_trampoline_holder(void) " movl %esp, %eax\n" " call trampoline_handler\n" /* Move flags to cs */ - " movl 52(%esp), %edx\n" - " movl %edx, 48(%esp)\n" + " movl 56(%esp), %edx\n" + " movl %edx, 52(%esp)\n" /* Replace saved flags with true return address. */ - " movl %eax, 52(%esp)\n" + " movl %eax, 56(%esp)\n" " popl %ebx\n" " popl %ecx\n" " popl %edx\n" @@ -666,8 +666,8 @@ static void __used __kprobes kretprobe_trampoline_holder(void) " popl %edi\n" " popl %ebp\n" " popl %eax\n" - /* Skip ip, orig_ax, es, ds, fs */ - " addl $20, %esp\n" + /* Skip ds, es, fs, gs, orig_ax and ip */ + " addl $24, %esp\n" " popf\n" #endif " ret\n"); @@ -691,6 +691,7 @@ static __used __kprobes void *trampoline_handler(struct pt_regs *regs) regs->cs = __KERNEL_CS; #else regs->cs = __KERNEL_CS | get_kernel_rpl(); + regs->gs = 0; #endif regs->ip = trampoline_address; regs->orig_ax = ~0UL; From 2a4efa42450762cbfa5c5712aa4cc9f06924c9fd Mon Sep 17 00:00:00 2001 From: Zhaolei Date: Wed, 25 Mar 2009 12:06:05 +0800 Subject: [PATCH 395/413] ftrace: Using FTRACE_WARN_ON() to check "freed record" in ftrace_release() "Because when we call ftrace_free_rec we change the rec->ip to point to the next record in the chain. Something is very wrong if rec->ip >= s && rec->ip < e and the record is already free." "Note, use FTRACE_WARN_ON() macro. This way it shuts down ftrace if it is hit and helps to avoid further damage later." -- Steven Rostedt Signed-off-by: Zhao Lei Signed-off-by: Steven Rostedt --- kernel/trace/ftrace.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 7b8722baf153..1752a63f37c0 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -358,9 +358,14 @@ void ftrace_release(void *start, unsigned long size) mutex_lock(&ftrace_lock); do_for_each_ftrace_rec(pg, rec) { - if ((rec->ip >= s) && (rec->ip < e) && - !(rec->flags & FTRACE_FL_FREE)) + if ((rec->ip >= s) && (rec->ip < e)) { + /* + * rec->ip is changed in ftrace_free_rec() + * It should not between s and e if record was freed. + */ + FTRACE_WARN_ON(rec->flags & FTRACE_FL_FREE); ftrace_free_rec(rec); + } } while_for_each_ftrace_rec(); mutex_unlock(&ftrace_lock); } From 9a8118baaeb0eaa148913bed77bf9c6335f6ca63 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Thu, 26 Mar 2009 01:24:34 -0500 Subject: [PATCH 396/413] tracing: filter fix for TRACE_EVENT_FORMAT events Impact: fix crash (hang) when using TRACE_EVENT_FORMAT filter files filters are only hooked up to the tracepoint events defined using TRACE_EVENT but not the tracers that use TRACE_EVENT_FORMAT, such as ftrace. Do not display the filter files at all for TRACE_EVENT_FORMAT events for the time being. Cc: Steven Rostedt Cc: =?ISO-8859-1?Q?Fr=E9d=E9ric?= Weisbecker LKML-Reference: <1237878882.8339.61.camel@charm-linux> Signed-off-by: Ingo Molnar --- kernel/trace/trace_events.c | 18 +++++------------- 1 file changed, 5 insertions(+), 13 deletions(-) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index d132997ab756..64ec4d278ffb 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -680,7 +680,6 @@ static struct dentry * event_subsystem_dir(const char *name, struct dentry *d_events) { struct event_subsystem *system; - struct dentry *entry; /* First see if we did not already create this dir */ list_for_each_entry(system, &event_subsystems, list) { @@ -709,12 +708,6 @@ event_subsystem_dir(const char *name, struct dentry *d_events) system->preds = NULL; - entry = debugfs_create_file("filter", 0644, system->entry, system, - &ftrace_subsystem_filter_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'%s/filter' entry\n", name); - return system->entry; } @@ -770,14 +763,13 @@ event_create_dir(struct ftrace_event_call *call, struct dentry *d_events) " events/%s\n", call->name); return ret; } + entry = debugfs_create_file("filter", 0644, call->dir, call, + &ftrace_event_filter_fops); + if (!entry) + pr_warning("Could not create debugfs " + "'%s/filter' entry\n", call->name); } - entry = debugfs_create_file("filter", 0644, call->dir, call, - &ftrace_event_filter_fops); - if (!entry) - pr_warning("Could not create debugfs " - "'%s/filter' entry\n", call->name); - /* A trace may not want to export its format */ if (!call->show_format) return 0; From 548c316137901cc81dea35e26362a9d5ba0b89b6 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Tue, 31 Mar 2009 00:25:23 +0200 Subject: [PATCH 397/413] tracing, Text Edit Lock: cleanup Remove incorrectly introduced headers. Signed-off-by: Ingo Molnar --- mm/memory.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/mm/memory.c b/mm/memory.c index dfc9e4ea4e8b..baa999e87cd2 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -48,8 +48,6 @@ #include #include #include -#include -#include #include #include #include From 6c051ce0307526adec32a847f0daa1af2124f0a9 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 25 Mar 2009 17:18:56 +0800 Subject: [PATCH 398/413] blktrace: fix timestamp in binary output I found the timestamp is wrong: # echo bin > trace_option # echo blk > current_tracer # cat trace_pipe | blkparse -i - 8,0 0 0 0.000000000 504 A W ... ... 8,7 1 0 0.008534097 0 C R ... (should be 8.534097xxx) user-space blkparse expects the timestamp to be nanosecond. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 6fb274f5f34e..ee7a8bb8b1e8 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1168,7 +1168,7 @@ static int blk_trace_synthesize_old_trace(struct trace_iterator *iter) const int offset = offsetof(struct blk_io_trace, sector); struct blk_io_trace old = { .magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION, - .time = ns2usecs(iter->ts), + .time = iter->ts, }; if (!trace_seq_putmem(s, &old, offset)) From b5230b56ee6caeb27cedb7753c0c319646383bb4 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 25 Mar 2009 17:19:33 +0800 Subject: [PATCH 399/413] blktrace: fix a race when creating blk_tree_root in debugfs t1 t2 ------ ------ do_blk_trace_setup() do_blk_trace_setup() if (!blk_tree_root) { if (!blk_tree_root) blk_tree_root = create_dir() blk_tree_root = create_dir(); (now blk_tree_root == NULL) ... dir = create_dir(name, blk_tree_root); Due to this race, t1 will create 'dir' in /debugfs but not /debugfs/block. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index ee7a8bb8b1e8..95f89faca73e 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -426,11 +426,15 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, ret = -ENOENT; + mutex_lock(&blk_tree_mutex); if (!blk_tree_root) { blk_tree_root = debugfs_create_dir("block", NULL); - if (!blk_tree_root) + if (!blk_tree_root) { + mutex_unlock(&blk_tree_mutex); goto err; + } } + mutex_unlock(&blk_tree_mutex); dir = debugfs_create_dir(buts->name, blk_tree_root); From 5554720482a631702146a959db22fe417740e0a6 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 25 Mar 2009 17:21:26 +0800 Subject: [PATCH 400/413] blktrace: fix the original blktrace Currently the original blktrace, which is using relay and is used via ioctl, is broken. You can use ftrace to see the output of blktrace, but user-space blktrace is unusable. It's broken by "blktrace: add ftrace plugin" (c71a896154119f4ca9e89d6078f5f63ad60ef199) - if (unlikely(bt->trace_state != Blktrace_running)) + if (unlikely(bt->trace_state != Blktrace_running || !blk_tracer_enabled)) return; With this patch, both ioctl and ftrace can be used, but of course you can't use both of them at the same time. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 12 +++++------- 1 file changed, 5 insertions(+), 7 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 95f89faca73e..a7f7ff5db38b 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -110,7 +110,7 @@ void __trace_note_message(struct blk_trace *bt, const char *fmt, ...) unsigned long flags; char *buf; - if (blk_tr) { + if (blk_tracer_enabled) { va_start(args, fmt); ftrace_vprintk(fmt, args); va_end(args); @@ -169,7 +169,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, pid_t pid; int cpu, pc = 0; - if (unlikely(bt->trace_state != Blktrace_running || + if (unlikely(bt->trace_state != Blktrace_running && !blk_tracer_enabled)) return; @@ -185,7 +185,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, return; cpu = raw_smp_processor_id(); - if (blk_tr) { + if (blk_tracer_enabled) { tracing_record_cmdline(current); pc = preempt_count(); @@ -235,7 +235,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, if (pdu_len) memcpy((void *) t + sizeof(*t), pdu_data, pdu_len); - if (blk_tr) { + if (blk_tracer_enabled) { trace_buffer_unlock_commit(blk_tr, event, 0, pc); return; } @@ -267,8 +267,7 @@ int blk_trace_remove(struct request_queue *q) if (!bt) return -EINVAL; - if (bt->trace_state == Blktrace_setup || - bt->trace_state == Blktrace_stopped) + if (bt->trace_state != Blktrace_running) blk_trace_cleanup(bt); return 0; @@ -1273,7 +1272,6 @@ static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) bt->dev = dev; bt->act_mask = (u16)-1; bt->end_lba = -1ULL; - bt->trace_state = Blktrace_running; old_bt = xchg(&q->blk_trace, bt); if (old_bt != NULL) { From eb08f8eb0673d9c1e62b69ad1b41593e73c40467 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Tue, 24 Mar 2009 16:05:27 +0800 Subject: [PATCH 401/413] blktrace: fix off-by-one bug 'what' is used as the index of array what2act, so it can't >= the array size. Signed-off-by: Li Zefan Acked-by: Jens Axboe Acked-by: Arnaldo Carvalho de Melo Acked-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index a7f7ff5db38b..d43cdac7c754 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -1152,7 +1152,7 @@ static enum print_line_t blk_trace_event_print(struct trace_iterator *iter, if (!trace_print_context(iter)) return TRACE_TYPE_PARTIAL_LINE; - if (unlikely(what == 0 || what > ARRAY_SIZE(what2act))) + if (unlikely(what == 0 || what >= ARRAY_SIZE(what2act))) ret = trace_seq_printf(s, "Bad pc action %x\n", what); else { const bool long_act = !!(trace_flags & TRACE_ITER_VERBOSE); @@ -1199,7 +1199,7 @@ static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter) t = (const struct blk_io_trace *)iter->ent; what = t->action & ((1 << BLK_TC_SHIFT) - 1); - if (unlikely(what == 0 || what > ARRAY_SIZE(what2act))) + if (unlikely(what == 0 || what >= ARRAY_SIZE(what2act))) ret = trace_seq_printf(&iter->seq, "Bad pc action %x\n", what); else { const bool long_act = !!(trace_flags & TRACE_ITER_VERBOSE); From 35ac51bfe4c293b67ce9f85082ba0b9bc6123c40 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 27 Mar 2009 10:19:46 +0800 Subject: [PATCH 402/413] blktrace: make classic output more classic Impact: fix ftrace plugin timestamp output In the classic user-space blktrace, the output timestamp is sec.nsec not sec.usec. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index d43cdac7c754..5b28f0f119c5 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -994,8 +994,8 @@ static void get_pdu_remap(const struct trace_entry *ent, static int blk_log_action_iter(struct trace_iterator *iter, const char *act) { char rwbs[6]; - unsigned long long ts = ns2usecs(iter->ts); - unsigned long usec_rem = do_div(ts, USEC_PER_SEC); + unsigned long long ts = iter->ts; + unsigned long nsec_rem = do_div(ts, NSEC_PER_SEC); unsigned secs = (unsigned long)ts; const struct trace_entry *ent = iter->ent; const struct blk_io_trace *t = (const struct blk_io_trace *)ent; @@ -1003,9 +1003,9 @@ static int blk_log_action_iter(struct trace_iterator *iter, const char *act) fill_rwbs(rwbs, t); return trace_seq_printf(&iter->seq, - "%3d,%-3d %2d %5d.%06lu %5u %2s %3s ", + "%3d,%-3d %2d %5d.%09lu %5u %2s %3s ", MAJOR(t->device), MINOR(t->device), iter->cpu, - secs, usec_rem, ent->pid, act, rwbs); + secs, nsec_rem, ent->pid, act, rwbs); } static int blk_log_action_seq(struct trace_seq *s, const struct blk_io_trace *t, From 17ba97e347bec9bbc47a0877c7a098708982129d Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 27 Mar 2009 10:20:09 +0800 Subject: [PATCH 403/413] blktrace: fix blk_probes_ref chaos Impact: fix mixed ioctl and ftrace-plugin blktrace use refcount bugs ioctl-based blktrace allocates bt and registers tracepoints when ioctl(BLKTRACESETUP), and do all cleanups when ioctl(BLKTRACETEARDOWN). while ftrace-based blktrace allocates/frees bt when: # echo 1/0 > /sys/block/sda/sda1/trace/enable and registers/unregisters tracepoints when: # echo blk/nop > /debugfs/tracing/current_tracer or # echo 1/0 > /debugfs/tracing/tracing_enable The separatation of allocation and registeration causes 2 problems: 1. current user-space blktrace still calls ioctl(TEARDOWN) when ioctl(SETUP) failed: # echo 1 > /sys/block/sda/sda1/trace/enable # blktrace /dev/sda BLKTRACESETUP: Device or resource busy ^C and now blk_probes_ref == -1 2. Another way to make blk_probes_ref == -1: # plugin sdb && mount sdb1 # echo 1 > /sys/block/sdb/sdb1/trace/enable # remove sdb This patch does the allocation and registeration when writing sdaX/trace/enable. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 5b28f0f119c5..8d6bd12aab10 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -478,7 +478,7 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, goto err; } - if (atomic_add_return(1, &blk_probes_ref) == 1) + if (atomic_inc_return(&blk_probes_ref) == 1) blk_register_tracepoints(); return 0; @@ -1091,8 +1091,6 @@ static void blk_tracer_print_header(struct seq_file *m) static void blk_tracer_start(struct trace_array *tr) { - if (atomic_add_return(1, &blk_probes_ref) == 1) - blk_register_tracepoints(); trace_flags &= ~TRACE_ITER_CONTEXT_INFO; } @@ -1107,15 +1105,10 @@ static int blk_tracer_init(struct trace_array *tr) static void blk_tracer_stop(struct trace_array *tr) { trace_flags |= TRACE_ITER_CONTEXT_INFO; - if (atomic_dec_and_test(&blk_probes_ref)) - blk_unregister_tracepoints(); } static void blk_tracer_reset(struct trace_array *tr) { - if (!atomic_read(&blk_probes_ref)) - return; - blk_tracer_enabled = false; blk_tracer_stop(tr); } @@ -1254,6 +1247,9 @@ static int blk_trace_remove_queue(struct request_queue *q) if (bt == NULL) return -EINVAL; + if (atomic_dec_and_test(&blk_probes_ref)) + blk_unregister_tracepoints(); + kfree(bt); return 0; } @@ -1280,6 +1276,9 @@ static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) return -EBUSY; } + if (atomic_inc_return(&blk_probes_ref) == 1) + blk_register_tracepoints(); + return 0; } From ad5dd5493a55e462796e42e50a49e76df76fdb05 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 27 Mar 2009 10:20:24 +0800 Subject: [PATCH 404/413] blktrace: fix memory leak when freeing struct blk_io_trace Impact: fix mixed ioctl and ftrace-plugin blktrace use memory leak When mixing the use of ioctl-based blktrace and ftrace-based blktrace, we can leak memory in this way: # btrace /dev/sda > /dev/null & # echo 0 > /sys/block/sda/sda1/trace/enable now we leak bt->dropped_file, bt->msg_file, bt->rchan... Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 29 ++++++++++++----------------- 1 file changed, 12 insertions(+), 17 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 8d6bd12aab10..2f21d7761600 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -247,7 +247,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, static struct dentry *blk_tree_root; static DEFINE_MUTEX(blk_tree_mutex); -static void blk_trace_cleanup(struct blk_trace *bt) +static void blk_trace_free(struct blk_trace *bt) { debugfs_remove(bt->msg_file); debugfs_remove(bt->dropped_file); @@ -255,6 +255,11 @@ static void blk_trace_cleanup(struct blk_trace *bt) free_percpu(bt->sequence); free_percpu(bt->msg_data); kfree(bt); +} + +static void blk_trace_cleanup(struct blk_trace *bt) +{ + blk_trace_free(bt); if (atomic_dec_and_test(&blk_probes_ref)) blk_unregister_tracepoints(); } @@ -410,11 +415,11 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, if (buts->name[i] == '/') buts->name[i] = '_'; - ret = -ENOMEM; bt = kzalloc(sizeof(*bt), GFP_KERNEL); if (!bt) - goto err; + return -ENOMEM; + ret = -ENOMEM; bt->sequence = alloc_percpu(unsigned long); if (!bt->sequence) goto err; @@ -483,17 +488,7 @@ int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev, return 0; err: - if (bt) { - if (bt->msg_file) - debugfs_remove(bt->msg_file); - if (bt->dropped_file) - debugfs_remove(bt->dropped_file); - free_percpu(bt->sequence); - free_percpu(bt->msg_data); - if (bt->rchan) - relay_close(bt->rchan); - kfree(bt); - } + blk_trace_free(bt); return ret; } @@ -1091,6 +1086,7 @@ static void blk_tracer_print_header(struct seq_file *m) static void blk_tracer_start(struct trace_array *tr) { + blk_tracer_enabled = true; trace_flags &= ~TRACE_ITER_CONTEXT_INFO; } @@ -1098,18 +1094,17 @@ static int blk_tracer_init(struct trace_array *tr) { blk_tr = tr; blk_tracer_start(tr); - blk_tracer_enabled = true; return 0; } static void blk_tracer_stop(struct trace_array *tr) { + blk_tracer_enabled = false; trace_flags |= TRACE_ITER_CONTEXT_INFO; } static void blk_tracer_reset(struct trace_array *tr) { - blk_tracer_enabled = false; blk_tracer_stop(tr); } @@ -1250,7 +1245,7 @@ static int blk_trace_remove_queue(struct request_queue *q) if (atomic_dec_and_test(&blk_probes_ref)) blk_unregister_tracepoints(); - kfree(bt); + blk_trace_free(bt); return 0; } From b6a4b0c3ad4c09c1d37b1040ac8e3ebd1016e10b Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 27 Mar 2009 10:21:23 +0800 Subject: [PATCH 405/413] blktrace: extract duplidate code Impact: cleanup blk_trace_event_print() and blk_tracer_print_line() share most of the code. text data bss dec hex filename 8605 393 12 9010 2332 kernel/trace/blktrace.o.orig text data bss dec hex filename 8555 393 12 8960 2300 kernel/trace/blktrace.o This patch also prepares for the next patch, that prints out BLK_TN_MESSAGE. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 62 ++++++++++++++++++++--------------------- 1 file changed, 30 insertions(+), 32 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 2f21d7761600..c103b0c20022 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -986,29 +986,31 @@ static void get_pdu_remap(const struct trace_entry *ent, r->sector = be64_to_cpu(sector); } -static int blk_log_action_iter(struct trace_iterator *iter, const char *act) +typedef int (blk_log_action_t) (struct trace_iterator *iter, const char *act); + +static int blk_log_action_classic(struct trace_iterator *iter, const char *act) { char rwbs[6]; unsigned long long ts = iter->ts; unsigned long nsec_rem = do_div(ts, NSEC_PER_SEC); unsigned secs = (unsigned long)ts; - const struct trace_entry *ent = iter->ent; - const struct blk_io_trace *t = (const struct blk_io_trace *)ent; + const struct blk_io_trace *t = te_blk_io_trace(iter->ent); fill_rwbs(rwbs, t); return trace_seq_printf(&iter->seq, "%3d,%-3d %2d %5d.%09lu %5u %2s %3s ", MAJOR(t->device), MINOR(t->device), iter->cpu, - secs, nsec_rem, ent->pid, act, rwbs); + secs, nsec_rem, iter->ent->pid, act, rwbs); } -static int blk_log_action_seq(struct trace_seq *s, const struct blk_io_trace *t, - const char *act) +static int blk_log_action(struct trace_iterator *iter, const char *act) { char rwbs[6]; + const struct blk_io_trace *t = te_blk_io_trace(iter->ent); + fill_rwbs(rwbs, t); - return trace_seq_printf(s, "%3d,%-3d %2s %3s ", + return trace_seq_printf(&iter->seq, "%3d,%-3d %2s %3s ", MAJOR(t->device), MINOR(t->device), act, rwbs); } @@ -1129,22 +1131,25 @@ static const struct { [__BLK_TA_REMAP] = {{ "A", "remap" }, blk_log_remap }, }; -static enum print_line_t blk_trace_event_print(struct trace_iterator *iter, - int flags) +static enum print_line_t print_one_line(struct trace_iterator *iter, + bool classic) { struct trace_seq *s = &iter->seq; - const struct blk_io_trace *t = (struct blk_io_trace *)iter->ent; - const u16 what = t->action & ((1 << BLK_TC_SHIFT) - 1); + const struct blk_io_trace *t; + u16 what; int ret; + bool long_act; + blk_log_action_t *log_action; - if (!trace_print_context(iter)) - return TRACE_TYPE_PARTIAL_LINE; + t = te_blk_io_trace(iter->ent); + what = t->action & ((1 << BLK_TC_SHIFT) - 1); + long_act = !!(trace_flags & TRACE_ITER_VERBOSE); + log_action = classic ? &blk_log_action_classic : &blk_log_action; if (unlikely(what == 0 || what >= ARRAY_SIZE(what2act))) ret = trace_seq_printf(s, "Bad pc action %x\n", what); else { - const bool long_act = !!(trace_flags & TRACE_ITER_VERBOSE); - ret = blk_log_action_seq(s, t, what2act[what].act[long_act]); + ret = log_action(iter, what2act[what].act[long_act]); if (ret) ret = what2act[what].print(s, iter->ent); } @@ -1152,6 +1157,15 @@ static enum print_line_t blk_trace_event_print(struct trace_iterator *iter, return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; } +static enum print_line_t blk_trace_event_print(struct trace_iterator *iter, + int flags) +{ + if (!trace_print_context(iter)) + return TRACE_TYPE_PARTIAL_LINE; + + return print_one_line(iter, false); +} + static int blk_trace_synthesize_old_trace(struct trace_iterator *iter) { struct trace_seq *s = &iter->seq; @@ -1177,26 +1191,10 @@ blk_trace_event_print_binary(struct trace_iterator *iter, int flags) static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter) { - const struct blk_io_trace *t; - u16 what; - int ret; - if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC)) return TRACE_TYPE_UNHANDLED; - t = (const struct blk_io_trace *)iter->ent; - what = t->action & ((1 << BLK_TC_SHIFT) - 1); - - if (unlikely(what == 0 || what >= ARRAY_SIZE(what2act))) - ret = trace_seq_printf(&iter->seq, "Bad pc action %x\n", what); - else { - const bool long_act = !!(trace_flags & TRACE_ITER_VERBOSE); - ret = blk_log_action_iter(iter, what2act[what].act[long_act]); - if (ret) - ret = what2act[what].print(&iter->seq, iter->ent); - } - - return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; + return print_one_line(iter, true); } static struct tracer blk_tracer __read_mostly = { From 18cea4591a98817697017bcb056a848bae1205df Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 27 Mar 2009 10:21:54 +0800 Subject: [PATCH 406/413] blktrace: print out BLK_TN_MESSAGE properly Impact: improve ftrace plugin output Before this patch: # cat trace make-5383 [001] 741.240059: 8,7 P N [make] __trace_note_message: cfq1074 # echo 1 > options/blk_classic # cat trace 8,7 1 0.692221252 0 C W 130411392 + 1024 [0] Bad pc action 6361 Bad pc action 283d # echo 0 > options/blk_classic # echo bin > trace_options # cat trace_pipe | blkparse -i - (can't parse messages generated by blk_add_trace_msg()) After this patch: # cat trace -0 [001] 187.600933: 8,7 C W 145220224 + 8 [0] -0 [001] 187.600946: 8,7 m N cfq1076 complete # echo 1 > options/blk_classic # cat trace 8,7 1 0.256378996 238 I W 113190728 + 8 [pdflush] 8,7 1 0.256378998 238 m N cfq1076 insert_request # echo 0 > options/blk_classic # echo bin > trace_options # cat trace_pipe | blkparse -i - 8,7 1 0 22.973250293 0 C W 102770576 + 8 [0] 8,7 1 0 22.973259213 0 m N cfq1076 complete Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/blktrace.c | 80 +++++++++++++++++++++++++++++++---------- 1 file changed, 61 insertions(+), 19 deletions(-) diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index c103b0c20022..947c5b3f90c4 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -59,22 +59,39 @@ static void trace_note(struct blk_trace *bt, pid_t pid, int action, const void *data, size_t len) { struct blk_io_trace *t; + struct ring_buffer_event *event = NULL; + int pc = 0; + int cpu = smp_processor_id(); + bool blk_tracer = blk_tracer_enabled; + + if (blk_tracer) { + pc = preempt_count(); + event = trace_buffer_lock_reserve(blk_tr, TRACE_BLK, + sizeof(*t) + len, + 0, pc); + if (!event) + return; + t = ring_buffer_event_data(event); + goto record_it; + } if (!bt->rchan) return; t = relay_reserve(bt->rchan, sizeof(*t) + len); if (t) { - const int cpu = smp_processor_id(); - t->magic = BLK_IO_TRACE_MAGIC | BLK_IO_TRACE_VERSION; t->time = ktime_to_ns(ktime_get()); +record_it: t->device = bt->dev; t->action = action; t->pid = pid; t->cpu = cpu; t->pdu_len = len; memcpy((void *) t + sizeof(*t), data, len); + + if (blk_tracer) + trace_buffer_unlock_commit(blk_tr, event, 0, pc); } } @@ -110,14 +127,8 @@ void __trace_note_message(struct blk_trace *bt, const char *fmt, ...) unsigned long flags; char *buf; - if (blk_tracer_enabled) { - va_start(args, fmt); - ftrace_vprintk(fmt, args); - va_end(args); - return; - } - - if (!bt->msg_data) + if (unlikely(bt->trace_state != Blktrace_running && + !blk_tracer_enabled)) return; local_irq_save(flags); @@ -168,9 +179,9 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, unsigned long *sequence; pid_t pid; int cpu, pc = 0; + bool blk_tracer = blk_tracer_enabled; - if (unlikely(bt->trace_state != Blktrace_running && - !blk_tracer_enabled)) + if (unlikely(bt->trace_state != Blktrace_running && !blk_tracer)) return; what |= ddir_act[rw & WRITE]; @@ -185,7 +196,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, return; cpu = raw_smp_processor_id(); - if (blk_tracer_enabled) { + if (blk_tracer) { tracing_record_cmdline(current); pc = preempt_count(); @@ -235,7 +246,7 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, if (pdu_len) memcpy((void *) t + sizeof(*t), pdu_data, pdu_len); - if (blk_tracer_enabled) { + if (blk_tracer) { trace_buffer_unlock_commit(blk_tr, event, 0, pc); return; } @@ -922,6 +933,11 @@ static void fill_rwbs(char *rwbs, const struct blk_io_trace *t) int i = 0; int tc = t->action >> BLK_TC_SHIFT; + if (t->action == BLK_TN_MESSAGE) { + rwbs[i++] = 'N'; + goto out; + } + if (tc & BLK_TC_DISCARD) rwbs[i++] = 'D'; else if (tc & BLK_TC_WRITE) @@ -939,7 +955,7 @@ static void fill_rwbs(char *rwbs, const struct blk_io_trace *t) rwbs[i++] = 'S'; if (tc & BLK_TC_META) rwbs[i++] = 'M'; - +out: rwbs[i] = '\0'; } @@ -1074,6 +1090,17 @@ static int blk_log_split(struct trace_seq *s, const struct trace_entry *ent) get_pdu_int(ent), cmd); } +static int blk_log_msg(struct trace_seq *s, const struct trace_entry *ent) +{ + int ret; + const struct blk_io_trace *t = te_blk_io_trace(ent); + + ret = trace_seq_putmem(s, t + 1, t->pdu_len); + if (ret) + return trace_seq_putc(s, '\n'); + return ret; +} + /* * struct tracer operations */ @@ -1146,6 +1173,13 @@ static enum print_line_t print_one_line(struct trace_iterator *iter, long_act = !!(trace_flags & TRACE_ITER_VERBOSE); log_action = classic ? &blk_log_action_classic : &blk_log_action; + if (t->action == BLK_TN_MESSAGE) { + ret = log_action(iter, long_act ? "message" : "m"); + if (ret) + ret = blk_log_msg(s, iter->ent); + goto out; + } + if (unlikely(what == 0 || what >= ARRAY_SIZE(what2act))) ret = trace_seq_printf(s, "Bad pc action %x\n", what); else { @@ -1153,7 +1187,7 @@ static enum print_line_t print_one_line(struct trace_iterator *iter, if (ret) ret = what2act[what].print(s, iter->ent); } - +out: return ret ? TRACE_TYPE_HANDLED : TRACE_TYPE_PARTIAL_LINE; } @@ -1253,11 +1287,16 @@ static int blk_trace_remove_queue(struct request_queue *q) static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) { struct blk_trace *old_bt, *bt = NULL; + int ret = -ENOMEM; bt = kzalloc(sizeof(*bt), GFP_KERNEL); if (!bt) return -ENOMEM; + bt->msg_data = __alloc_percpu(BLK_TN_MAX_MSG, __alignof__(char)); + if (!bt->msg_data) + goto free_bt; + bt->dev = dev; bt->act_mask = (u16)-1; bt->end_lba = -1ULL; @@ -1265,14 +1304,17 @@ static int blk_trace_setup_queue(struct request_queue *q, dev_t dev) old_bt = xchg(&q->blk_trace, bt); if (old_bt != NULL) { (void)xchg(&q->blk_trace, old_bt); - kfree(bt); - return -EBUSY; + ret = -EBUSY; + goto free_bt; } if (atomic_inc_return(&blk_probes_ref) == 1) blk_register_tracepoints(); - return 0; + +free_bt: + blk_trace_free(bt); + return ret; } /* From bdd6df6af98ce7e70702edfb5fd5dbbd8d1b0453 Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Mon, 23 Mar 2009 15:12:22 +0200 Subject: [PATCH 407/413] tracing: provide trace_seq_reserve() trace_seq_reserve() allows a caller to reserve space in a trace_seq and write directly into it. This makes it easier to export binary data to userspace via the tracing interface, by simply filling in a struct. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Ingo Molnar --- kernel/trace/trace_output.c | 13 +++++++++++++ kernel/trace/trace_output.h | 1 + 2 files changed, 14 insertions(+) diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 19261fdd2455..6595074cbac5 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -167,6 +167,19 @@ int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len) return trace_seq_putmem(s, hex, j); } +void *trace_seq_reserve(struct trace_seq *s, size_t len) +{ + void *ret; + + if (len > ((PAGE_SIZE - 1) - s->len)) + return NULL; + + ret = s->buffer + s->len; + s->len += len; + + return ret; +} + int trace_seq_path(struct trace_seq *s, struct path *path) { unsigned char *p; diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 35c422fb51a9..0ae20b83eec8 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -33,6 +33,7 @@ int trace_seq_puts(struct trace_seq *s, const char *str); int trace_seq_putc(struct trace_seq *s, unsigned char c); int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len); int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len); +void *trace_seq_reserve(struct trace_seq *s, size_t len); int trace_seq_path(struct trace_seq *s, struct path *path); int seq_print_userip_objs(const struct userstack_entry *entry, struct trace_seq *s, unsigned long sym_flags); From f285901bb21355bb47106658ef14eeb6b8ed538f Mon Sep 17 00:00:00 2001 From: Eduard - Gabriel Munteanu Date: Mon, 23 Mar 2009 15:12:23 +0200 Subject: [PATCH 408/413] tracing: add missing 'extern' keywords to trace_output.h Impact: cleanup Many declarations within trace_output.h are missing the 'extern' keyword in an inconsistent manner. This adds 'extern' where it should be. Signed-off-by: Eduard - Gabriel Munteanu Signed-off-by: Ingo Molnar --- kernel/trace/trace_output.h | 33 +++++++++++++++++---------------- 1 file changed, 17 insertions(+), 16 deletions(-) diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 0ae20b83eec8..46fb9612d884 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -29,25 +29,26 @@ seq_print_ip_sym(struct trace_seq *s, unsigned long ip, unsigned long sym_flags); extern ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt); -int trace_seq_puts(struct trace_seq *s, const char *str); -int trace_seq_putc(struct trace_seq *s, unsigned char c); -int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len); -int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len); -void *trace_seq_reserve(struct trace_seq *s, size_t len); -int trace_seq_path(struct trace_seq *s, struct path *path); -int seq_print_userip_objs(const struct userstack_entry *entry, - struct trace_seq *s, unsigned long sym_flags); -int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, - unsigned long ip, unsigned long sym_flags); +extern int trace_seq_puts(struct trace_seq *s, const char *str); +extern int trace_seq_putc(struct trace_seq *s, unsigned char c); +extern int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len); +extern int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len); +extern void *trace_seq_reserve(struct trace_seq *s, size_t len); +extern int trace_seq_path(struct trace_seq *s, struct path *path); +extern int seq_print_userip_objs(const struct userstack_entry *entry, + struct trace_seq *s, unsigned long sym_flags); +extern int seq_print_user_ip(struct trace_seq *s, struct mm_struct *mm, + unsigned long ip, unsigned long sym_flags); -int trace_print_context(struct trace_iterator *iter); -int trace_print_lat_context(struct trace_iterator *iter); +extern int trace_print_context(struct trace_iterator *iter); +extern int trace_print_lat_context(struct trace_iterator *iter); -struct trace_event *ftrace_find_event(int type); -int register_ftrace_event(struct trace_event *event); -int unregister_ftrace_event(struct trace_event *event); +extern struct trace_event *ftrace_find_event(int type); +extern int register_ftrace_event(struct trace_event *event); +extern int unregister_ftrace_event(struct trace_event *event); -enum print_line_t trace_nop_print(struct trace_iterator *iter, int flags); +extern enum print_line_t trace_nop_print(struct trace_iterator *iter, + int flags); #define MAX_MEMHEX_BYTES 8 #define HEX_CHARS (MAX_MEMHEX_BYTES*2 + 1) From b14b70a6a4e394c9630bcde17e07d3bcdcbca27e Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Fri, 27 Mar 2009 10:21:00 +0800 Subject: [PATCH 409/413] trace: make argument 'mem' of trace_seq_putmem() const Impact: fix build warning I passed a const value to trace_seq_putmem(), and I got compile warning. Signed-off-by: Li Zefan Acked-by: Jens Axboe Cc: Arnaldo Carvalho de Melo Cc: Steven Rostedt Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- kernel/trace/trace_output.c | 6 +++--- kernel/trace/trace_output.h | 5 +++-- 2 files changed, 6 insertions(+), 5 deletions(-) diff --git a/kernel/trace/trace_output.c b/kernel/trace/trace_output.c index 6595074cbac5..d72b9a63b247 100644 --- a/kernel/trace/trace_output.c +++ b/kernel/trace/trace_output.c @@ -137,7 +137,7 @@ int trace_seq_putc(struct trace_seq *s, unsigned char c) return 1; } -int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len) +int trace_seq_putmem(struct trace_seq *s, const void *mem, size_t len) { if (len > ((PAGE_SIZE - 1) - s->len)) return 0; @@ -148,10 +148,10 @@ int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len) return len; } -int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len) +int trace_seq_putmem_hex(struct trace_seq *s, const void *mem, size_t len) { unsigned char hex[HEX_CHARS]; - unsigned char *data = mem; + const unsigned char *data = mem; int i, j; #ifdef __BIG_ENDIAN diff --git a/kernel/trace/trace_output.h b/kernel/trace/trace_output.h index 46fb9612d884..e0bde39c2dd9 100644 --- a/kernel/trace/trace_output.h +++ b/kernel/trace/trace_output.h @@ -31,8 +31,9 @@ extern ssize_t trace_seq_to_user(struct trace_seq *s, char __user *ubuf, size_t cnt); extern int trace_seq_puts(struct trace_seq *s, const char *str); extern int trace_seq_putc(struct trace_seq *s, unsigned char c); -extern int trace_seq_putmem(struct trace_seq *s, void *mem, size_t len); -extern int trace_seq_putmem_hex(struct trace_seq *s, void *mem, size_t len); +extern int trace_seq_putmem(struct trace_seq *s, const void *mem, size_t len); +extern int trace_seq_putmem_hex(struct trace_seq *s, const void *mem, + size_t len); extern void *trace_seq_reserve(struct trace_seq *s, size_t len); extern int trace_seq_path(struct trace_seq *s, struct path *path); extern int seq_print_userip_objs(const struct userstack_entry *entry, From 2aad1b76e6b0cc5a2e5d9b95a9f356ddddbfa8a9 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 30 Mar 2009 11:11:28 -0400 Subject: [PATCH 410/413] function-graph: allow unregistering twice Impact: fix to permanent disabling of function graph tracer There should be nothing to prevent a tracer from unregistering a function graph callback more than once. This can simplify error paths. But currently, the counter does not account for mulitple unregistering of the function graph callback. If it happens, the function graph tracer will be permanently disabled. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ftrace.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 1752a63f37c0..f1ed080406c3 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -2719,6 +2719,9 @@ void unregister_ftrace_graph(void) { mutex_lock(&ftrace_lock); + if (!unlikely(atomic_read(&ftrace_graph_active))) + goto out; + atomic_dec(&ftrace_graph_active); unregister_trace_sched_switch(ftrace_graph_probe_sched_switch); ftrace_graph_return = (trace_func_graph_ret_t)ftrace_stub; @@ -2726,6 +2729,7 @@ void unregister_ftrace_graph(void) ftrace_shutdown(FTRACE_STOP_FUNC_RET); unregister_pm_notifier(&ftrace_suspend_notifier); + out: mutex_unlock(&ftrace_lock); } From 2e572895bf3203e881356a4039ab0fa428ed2639 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Mon, 30 Mar 2009 14:03:19 -0400 Subject: [PATCH 411/413] ring-buffer: do not remove reader page from list on ring buffer free Impact: prevent possible memory leak The reader page of the ring buffer is special. Although it points into the ring buffer, it is not part of the actual buffer. It is a page used by the reader to swap with a page in the ring buffer. Once the swap is made, the new reader page is again outside the buffer. Even though the reader page points into the buffer, it is really pointing to residual data. Note, this data is used by the reader. reader page | v (prev) +---+ (next) +----------| |----------+ | +---+ | v v +---+ +---+ +---+ -->| |------->| |------->| |---> <--| |<-------| |<-------| |<--- +---+ +---+ +---+ ^ ^ ^ \ | / ------- Buffer--------- If we perform a list_del_init() on the reader page we will actually remove the last page the reader swapped with and not the reader page itself. This will cause that page to not be freed, and thus is a memory leak. Luckily, the only user of the ring buffer so far is ftrace. And ftrace will not free its ring buffer after it allocates it. There is no current possible memory leak. But once there are other users, or if ftrace dynamically creates and frees its ring buffer, then this would be a memory leak. This patch fixes the leak for future cases. Signed-off-by: Steven Rostedt Signed-off-by: Ingo Molnar --- kernel/trace/ring_buffer.c | 1 - 1 file changed, 1 deletion(-) diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index edce2ff38944..960cbf44c844 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -563,7 +563,6 @@ static void rb_free_cpu_buffer(struct ring_buffer_per_cpu *cpu_buffer) struct list_head *head = &cpu_buffer->pages; struct buffer_page *bpage, *tmp; - list_del_init(&cpu_buffer->reader_page->list); free_buffer_page(cpu_buffer->reader_page); list_for_each_entry_safe(bpage, tmp, head, list) { From a095bdbb136f7bed96b7adf5aa1dd27bb2f839bf Mon Sep 17 00:00:00 2001 From: Stephen Rothwell Date: Fri, 27 Mar 2009 23:08:34 +1100 Subject: [PATCH 412/413] tracing, powerpc: fix powerpc tree and tracing tree interaction Today's linux-next build (powerpc allyesconfig) failed like this: arch/powerpc/kernel/ftrace.c: In function 'prepare_ftrace_return': arch/powerpc/kernel/ftrace.c:612: warning: passing argument 3 of 'ftrace_push_return_trace' makes pointer from integer without a cast arch/powerpc/kernel/ftrace.c:612: error: too many arguments to function 'ftrace_push_return_trace' Caused by commit 5d1a03dc541dc6672e60e57249ed22f40654ca47 ("function-graph: moved the timestamp from arch to generic code") from the tracing tree which (removed an argument from ftrace_push_return_trace()) interacting with commit 6794c78243bfda020ab184d6d578944f8e90d26c ("powerpc64: port of the function graph tracer") from the powerpc tree. Signed-off-by: Stephen Rothwell Cc: Steven Rostedt Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: LKML-Reference: <20090327230834.93d0221d.sfr@canb.auug.org.au> Signed-off-by: Ingo Molnar --- arch/powerpc/kernel/ftrace.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/arch/powerpc/kernel/ftrace.c b/arch/powerpc/kernel/ftrace.c index 5b5d16b2cac8..5455943f16aa 100644 --- a/arch/powerpc/kernel/ftrace.c +++ b/arch/powerpc/kernel/ftrace.c @@ -557,7 +557,6 @@ extern void mod_return_to_handler(void); void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) { unsigned long old; - unsigned long long calltime; int faulted; struct ftrace_graph_ent trace; unsigned long return_hooker = (unsigned long)&return_to_handler; @@ -606,10 +605,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) return; } - calltime = cpu_clock(raw_smp_processor_id()); - - if (ftrace_push_return_trace(old, calltime, - self_addr, &trace.depth) == -EBUSY) { + if (ftrace_push_return_trace(old, self_addr, &trace.depth) == -EBUSY) { *parent = old; return; } From 645dae969c3b8651c5bc7c54a1835ec03820f85f Mon Sep 17 00:00:00 2001 From: Stephen Rothwell Date: Thu, 19 Mar 2009 15:46:17 +1100 Subject: [PATCH 413/413] tracing, net: fix net tree and tracing tree merge interaction Today's linux-next build (powerpc ppc64_defconfig) failed like this: In file included from net/core/skbuff.c:69: include/trace/skb.h:4: error: expected ')' before '(' token include/trace/skb.h:4: error: expected ')' before '(' token [...] Caused by commit 2939b0469d04ba9ac791aca9a81625d7eb50662b ("tracing: replace TP with TP_") from the tracing tree interacting with commit 4893d39e865b2897bf9fcd329697d37032d853a1 ("Network Drop Monitor: Add trace declaration for skb frees") from the net tree. Acked-by: David S. Miller Signed-off-by: Ingo Molnar --- include/trace/skb.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/trace/skb.h b/include/trace/skb.h index a96610f92f69..b66206d9be72 100644 --- a/include/trace/skb.h +++ b/include/trace/skb.h @@ -5,7 +5,7 @@ #include DECLARE_TRACE(kfree_skb, - TPPROTO(struct sk_buff *skb, void *location), - TPARGS(skb, location)); + TP_PROTO(struct sk_buff *skb, void *location), + TP_ARGS(skb, location)); #endif