2008-05-12 13:20:42 -06:00
|
|
|
/*
|
|
|
|
* Infrastructure for profiling code inserted by 'gcc -pg'.
|
|
|
|
*
|
|
|
|
* Copyright (C) 2007-2008 Steven Rostedt <srostedt@redhat.com>
|
|
|
|
* Copyright (C) 2004-2008 Ingo Molnar <mingo@redhat.com>
|
|
|
|
*
|
|
|
|
* Originally ported from the -rt patch by:
|
|
|
|
* Copyright (C) 2007 Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
|
|
*
|
|
|
|
* Based on code in the latency_tracer, that is:
|
|
|
|
*
|
|
|
|
* Copyright (C) 2004-2006 Ingo Molnar
|
|
|
|
* Copyright (C) 2004 William Lee Irwin III
|
|
|
|
*/
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
#include <linux/stop_machine.h>
|
|
|
|
#include <linux/clocksource.h>
|
|
|
|
#include <linux/kallsyms.h>
|
2008-05-12 13:20:43 -06:00
|
|
|
#include <linux/seq_file.h>
|
|
|
|
#include <linux/debugfs.h>
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
#include <linux/kthread.h>
|
|
|
|
#include <linux/hardirq.h>
|
2008-05-12 13:20:42 -06:00
|
|
|
#include <linux/ftrace.h>
|
2008-05-12 13:20:43 -06:00
|
|
|
#include <linux/uaccess.h>
|
2008-05-12 13:20:43 -06:00
|
|
|
#include <linux/sysctl.h>
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
#include <linux/hash.h>
|
2008-05-12 13:20:43 -06:00
|
|
|
#include <linux/ctype.h>
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
#include <linux/list.h>
|
|
|
|
|
|
|
|
#include "trace.h"
|
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
int ftrace_enabled;
|
|
|
|
static int last_ftrace_enabled;
|
2008-05-12 13:20:43 -06:00
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static DEFINE_SPINLOCK(ftrace_lock);
|
2008-05-12 13:20:43 -06:00
|
|
|
static DEFINE_MUTEX(ftrace_sysctl_lock);
|
|
|
|
|
2008-05-12 13:20:42 -06:00
|
|
|
static struct ftrace_ops ftrace_list_end __read_mostly =
|
|
|
|
{
|
|
|
|
.func = ftrace_stub,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct ftrace_ops *ftrace_list __read_mostly = &ftrace_list_end;
|
|
|
|
ftrace_func_t ftrace_trace_function __read_mostly = ftrace_stub;
|
|
|
|
|
|
|
|
/* mcount is defined per arch in assembly */
|
|
|
|
EXPORT_SYMBOL(mcount);
|
|
|
|
|
|
|
|
notrace void ftrace_list_func(unsigned long ip, unsigned long parent_ip)
|
|
|
|
{
|
|
|
|
struct ftrace_ops *op = ftrace_list;
|
|
|
|
|
|
|
|
/* in case someone actually ports this to alpha! */
|
|
|
|
read_barrier_depends();
|
|
|
|
|
|
|
|
while (op != &ftrace_list_end) {
|
|
|
|
/* silly alpha */
|
|
|
|
read_barrier_depends();
|
|
|
|
op->func(ip, parent_ip);
|
|
|
|
op = op->next;
|
|
|
|
};
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
* clear_ftrace_function - reset the ftrace function
|
2008-05-12 13:20:42 -06:00
|
|
|
*
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
* This NULLs the ftrace function and in essence stops
|
|
|
|
* tracing. There may be lag
|
2008-05-12 13:20:42 -06:00
|
|
|
*/
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
void clear_ftrace_function(void)
|
2008-05-12 13:20:42 -06:00
|
|
|
{
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
ftrace_trace_function = ftrace_stub;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int notrace __register_ftrace_function(struct ftrace_ops *ops)
|
|
|
|
{
|
|
|
|
/* Should never be called by interrupts */
|
|
|
|
spin_lock(&ftrace_lock);
|
2008-05-12 13:20:42 -06:00
|
|
|
|
|
|
|
ops->next = ftrace_list;
|
|
|
|
/*
|
|
|
|
* We are entering ops into the ftrace_list but another
|
|
|
|
* CPU might be walking that list. We need to make sure
|
|
|
|
* the ops->next pointer is valid before another CPU sees
|
|
|
|
* the ops pointer included into the ftrace_list.
|
|
|
|
*/
|
|
|
|
smp_wmb();
|
|
|
|
ftrace_list = ops;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
if (ftrace_enabled) {
|
|
|
|
/*
|
|
|
|
* For one func, simply call it directly.
|
|
|
|
* For more than one func, call the chain.
|
|
|
|
*/
|
|
|
|
if (ops->next == &ftrace_list_end)
|
|
|
|
ftrace_trace_function = ops->func;
|
|
|
|
else
|
|
|
|
ftrace_trace_function = ftrace_list_func;
|
|
|
|
}
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
|
|
|
spin_unlock(&ftrace_lock);
|
2008-05-12 13:20:42 -06:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static int notrace __unregister_ftrace_function(struct ftrace_ops *ops)
|
2008-05-12 13:20:42 -06:00
|
|
|
{
|
|
|
|
struct ftrace_ops **p;
|
|
|
|
int ret = 0;
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
spin_lock(&ftrace_lock);
|
2008-05-12 13:20:42 -06:00
|
|
|
|
|
|
|
/*
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
* If we are removing the last function, then simply point
|
|
|
|
* to the ftrace_stub.
|
2008-05-12 13:20:42 -06:00
|
|
|
*/
|
|
|
|
if (ftrace_list == ops && ops->next == &ftrace_list_end) {
|
|
|
|
ftrace_trace_function = ftrace_stub;
|
|
|
|
ftrace_list = &ftrace_list_end;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (p = &ftrace_list; *p != &ftrace_list_end; p = &(*p)->next)
|
|
|
|
if (*p == ops)
|
|
|
|
break;
|
|
|
|
|
|
|
|
if (*p != ops) {
|
|
|
|
ret = -1;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
*p = (*p)->next;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
if (ftrace_enabled) {
|
|
|
|
/* If we only have one func left, then call that directly */
|
|
|
|
if (ftrace_list == &ftrace_list_end ||
|
|
|
|
ftrace_list->next == &ftrace_list_end)
|
|
|
|
ftrace_trace_function = ftrace_list->func;
|
|
|
|
}
|
2008-05-12 13:20:42 -06:00
|
|
|
|
|
|
|
out:
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
spin_unlock(&ftrace_lock);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_DYNAMIC_FTRACE
|
|
|
|
|
2008-05-12 13:20:44 -06:00
|
|
|
static struct task_struct *ftraced_task;
|
|
|
|
static DECLARE_WAIT_QUEUE_HEAD(ftraced_waiters);
|
|
|
|
static unsigned long ftraced_iteration_counter;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
enum {
|
|
|
|
FTRACE_ENABLE_CALLS = (1 << 0),
|
|
|
|
FTRACE_DISABLE_CALLS = (1 << 1),
|
|
|
|
FTRACE_UPDATE_TRACE_FUNC = (1 << 2),
|
|
|
|
FTRACE_ENABLE_MCOUNT = (1 << 3),
|
|
|
|
FTRACE_DISABLE_MCOUNT = (1 << 4),
|
|
|
|
};
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static int ftrace_filtered;
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static struct hlist_head ftrace_hash[FTRACE_HASHSIZE];
|
|
|
|
|
|
|
|
static DEFINE_PER_CPU(int, ftrace_shutdown_disable_cpu);
|
|
|
|
|
|
|
|
static DEFINE_SPINLOCK(ftrace_shutdown_lock);
|
|
|
|
static DEFINE_MUTEX(ftraced_lock);
|
2008-05-12 13:20:43 -06:00
|
|
|
static DEFINE_MUTEX(ftrace_filter_lock);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
struct ftrace_page {
|
|
|
|
struct ftrace_page *next;
|
|
|
|
int index;
|
|
|
|
struct dyn_ftrace records[];
|
|
|
|
} __attribute__((packed));
|
|
|
|
|
|
|
|
#define ENTRIES_PER_PAGE \
|
|
|
|
((PAGE_SIZE - sizeof(struct ftrace_page)) / sizeof(struct dyn_ftrace))
|
|
|
|
|
|
|
|
/* estimate from running different kernels */
|
|
|
|
#define NR_TO_INIT 10000
|
|
|
|
|
|
|
|
static struct ftrace_page *ftrace_pages_start;
|
|
|
|
static struct ftrace_page *ftrace_pages;
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static int ftraced_trigger;
|
|
|
|
static int ftraced_suspend;
|
|
|
|
|
|
|
|
static int ftrace_record_suspend;
|
|
|
|
|
|
|
|
static inline int
|
|
|
|
notrace ftrace_ip_in_hash(unsigned long ip, unsigned long key)
|
|
|
|
{
|
|
|
|
struct dyn_ftrace *p;
|
|
|
|
struct hlist_node *t;
|
|
|
|
int found = 0;
|
|
|
|
|
|
|
|
hlist_for_each_entry(p, t, &ftrace_hash[key], node) {
|
|
|
|
if (p->ip == ip) {
|
|
|
|
found = 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return found;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void notrace
|
|
|
|
ftrace_add_hash(struct dyn_ftrace *node, unsigned long key)
|
|
|
|
{
|
|
|
|
hlist_add_head(&node->node, &ftrace_hash[key]);
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static notrace struct dyn_ftrace *ftrace_alloc_dyn_node(unsigned long ip)
|
2008-05-12 13:20:43 -06:00
|
|
|
{
|
|
|
|
if (ftrace_pages->index == ENTRIES_PER_PAGE) {
|
|
|
|
if (!ftrace_pages->next)
|
|
|
|
return NULL;
|
|
|
|
ftrace_pages = ftrace_pages->next;
|
|
|
|
}
|
|
|
|
|
|
|
|
return &ftrace_pages->records[ftrace_pages->index++];
|
|
|
|
}
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static void notrace
|
2008-05-12 13:20:43 -06:00
|
|
|
ftrace_record_ip(unsigned long ip)
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
{
|
|
|
|
struct dyn_ftrace *node;
|
|
|
|
unsigned long flags;
|
|
|
|
unsigned long key;
|
|
|
|
int resched;
|
|
|
|
int atomic;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
if (!ftrace_enabled)
|
|
|
|
return;
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
resched = need_resched();
|
|
|
|
preempt_disable_notrace();
|
|
|
|
|
|
|
|
/* We simply need to protect against recursion */
|
|
|
|
__get_cpu_var(ftrace_shutdown_disable_cpu)++;
|
|
|
|
if (__get_cpu_var(ftrace_shutdown_disable_cpu) != 1)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
if (unlikely(ftrace_record_suspend))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
key = hash_long(ip, FTRACE_HASHBITS);
|
|
|
|
|
|
|
|
WARN_ON_ONCE(key >= FTRACE_HASHSIZE);
|
|
|
|
|
|
|
|
if (ftrace_ip_in_hash(ip, key))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
atomic = irqs_disabled();
|
|
|
|
|
|
|
|
spin_lock_irqsave(&ftrace_shutdown_lock, flags);
|
|
|
|
|
|
|
|
/* This ip may have hit the hash before the lock */
|
|
|
|
if (ftrace_ip_in_hash(ip, key))
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* There's a slight race that the ftraced will update the
|
2008-05-12 13:20:43 -06:00
|
|
|
* hash and reset here. If it is already converted, skip it.
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
*/
|
2008-05-12 13:20:43 -06:00
|
|
|
if (ftrace_ip_converted(ip))
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
node = ftrace_alloc_dyn_node(ip);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
if (!node)
|
|
|
|
goto out_unlock;
|
|
|
|
|
|
|
|
node->ip = ip;
|
|
|
|
|
|
|
|
ftrace_add_hash(node, key);
|
|
|
|
|
|
|
|
ftraced_trigger = 1;
|
|
|
|
|
|
|
|
out_unlock:
|
|
|
|
spin_unlock_irqrestore(&ftrace_shutdown_lock, flags);
|
|
|
|
out:
|
|
|
|
__get_cpu_var(ftrace_shutdown_disable_cpu)--;
|
|
|
|
|
|
|
|
/* prevent recursion with scheduler */
|
|
|
|
if (resched)
|
|
|
|
preempt_enable_no_resched_notrace();
|
|
|
|
else
|
|
|
|
preempt_enable_notrace();
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
#define FTRACE_ADDR ((long)(&ftrace_caller))
|
2008-05-12 13:20:43 -06:00
|
|
|
#define MCOUNT_ADDR ((long)(&mcount))
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static void notrace
|
|
|
|
__ftrace_replace_code(struct dyn_ftrace *rec,
|
|
|
|
unsigned char *old, unsigned char *new, int enable)
|
|
|
|
{
|
|
|
|
unsigned long ip;
|
|
|
|
int failed;
|
|
|
|
|
|
|
|
ip = rec->ip;
|
|
|
|
|
|
|
|
if (ftrace_filtered && enable) {
|
|
|
|
unsigned long fl;
|
|
|
|
/*
|
|
|
|
* If filtering is on:
|
|
|
|
*
|
|
|
|
* If this record is set to be filtered and
|
|
|
|
* is enabled then do nothing.
|
|
|
|
*
|
|
|
|
* If this record is set to be filtered and
|
|
|
|
* it is not enabled, enable it.
|
|
|
|
*
|
|
|
|
* If this record is not set to be filtered
|
|
|
|
* and it is not enabled do nothing.
|
|
|
|
*
|
|
|
|
* If this record is not set to be filtered and
|
|
|
|
* it is enabled, disable it.
|
|
|
|
*/
|
|
|
|
fl = rec->flags & (FTRACE_FL_FILTER | FTRACE_FL_ENABLED);
|
|
|
|
|
|
|
|
if ((fl == (FTRACE_FL_FILTER | FTRACE_FL_ENABLED)) ||
|
|
|
|
(fl == 0))
|
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If it is enabled disable it,
|
|
|
|
* otherwise enable it!
|
|
|
|
*/
|
|
|
|
if (fl == FTRACE_FL_ENABLED) {
|
|
|
|
/* swap new and old */
|
|
|
|
new = old;
|
|
|
|
old = ftrace_call_replace(ip, FTRACE_ADDR);
|
|
|
|
rec->flags &= ~FTRACE_FL_ENABLED;
|
|
|
|
} else {
|
|
|
|
new = ftrace_call_replace(ip, FTRACE_ADDR);
|
|
|
|
rec->flags |= FTRACE_FL_ENABLED;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
|
|
|
|
if (enable)
|
|
|
|
new = ftrace_call_replace(ip, FTRACE_ADDR);
|
|
|
|
else
|
|
|
|
old = ftrace_call_replace(ip, FTRACE_ADDR);
|
|
|
|
|
|
|
|
if (enable) {
|
|
|
|
if (rec->flags & FTRACE_FL_ENABLED)
|
|
|
|
return;
|
|
|
|
rec->flags |= FTRACE_FL_ENABLED;
|
|
|
|
} else {
|
|
|
|
if (!(rec->flags & FTRACE_FL_ENABLED))
|
|
|
|
return;
|
|
|
|
rec->flags &= ~FTRACE_FL_ENABLED;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
failed = ftrace_modify_code(ip, old, new);
|
|
|
|
if (failed)
|
|
|
|
rec->flags |= FTRACE_FL_FAILED;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void notrace ftrace_replace_code(int enable)
|
2008-05-12 13:20:43 -06:00
|
|
|
{
|
|
|
|
unsigned char *new = NULL, *old = NULL;
|
|
|
|
struct dyn_ftrace *rec;
|
|
|
|
struct ftrace_page *pg;
|
|
|
|
int i;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
if (enable)
|
2008-05-12 13:20:43 -06:00
|
|
|
old = ftrace_nop_replace();
|
|
|
|
else
|
|
|
|
new = ftrace_nop_replace();
|
|
|
|
|
|
|
|
for (pg = ftrace_pages_start; pg; pg = pg->next) {
|
|
|
|
for (i = 0; i < pg->index; i++) {
|
|
|
|
rec = &pg->records[i];
|
|
|
|
|
|
|
|
/* don't modify code that has already faulted */
|
|
|
|
if (rec->flags & FTRACE_FL_FAILED)
|
|
|
|
continue;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
__ftrace_replace_code(rec, old, new, enable);
|
2008-05-12 13:20:43 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static notrace void ftrace_shutdown_replenish(void)
|
|
|
|
{
|
|
|
|
if (ftrace_pages->next)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/* allocate another page */
|
|
|
|
ftrace_pages->next = (void *)get_zeroed_page(GFP_KERNEL);
|
|
|
|
}
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static notrace void
|
2008-05-12 13:20:43 -06:00
|
|
|
ftrace_code_disable(struct dyn_ftrace *rec)
|
2008-05-12 13:20:43 -06:00
|
|
|
{
|
|
|
|
unsigned long ip;
|
|
|
|
unsigned char *nop, *call;
|
|
|
|
int failed;
|
|
|
|
|
|
|
|
ip = rec->ip;
|
|
|
|
|
|
|
|
nop = ftrace_nop_replace();
|
2008-05-12 13:20:43 -06:00
|
|
|
call = ftrace_call_replace(ip, MCOUNT_ADDR);
|
2008-05-12 13:20:43 -06:00
|
|
|
|
|
|
|
failed = ftrace_modify_code(ip, call, nop);
|
|
|
|
if (failed)
|
|
|
|
rec->flags |= FTRACE_FL_FAILED;
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static int notrace __ftrace_modify_code(void *data)
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
{
|
2008-05-12 13:20:43 -06:00
|
|
|
unsigned long addr;
|
|
|
|
int *command = data;
|
|
|
|
|
|
|
|
if (*command & FTRACE_ENABLE_CALLS)
|
|
|
|
ftrace_replace_code(1);
|
|
|
|
else if (*command & FTRACE_DISABLE_CALLS)
|
|
|
|
ftrace_replace_code(0);
|
|
|
|
|
|
|
|
if (*command & FTRACE_UPDATE_TRACE_FUNC)
|
|
|
|
ftrace_update_ftrace_func(ftrace_trace_function);
|
|
|
|
|
|
|
|
if (*command & FTRACE_ENABLE_MCOUNT) {
|
|
|
|
addr = (unsigned long)ftrace_record_ip;
|
|
|
|
ftrace_mcount_set(&addr);
|
|
|
|
} else if (*command & FTRACE_DISABLE_MCOUNT) {
|
|
|
|
addr = (unsigned long)ftrace_stub;
|
|
|
|
ftrace_mcount_set(&addr);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static void notrace ftrace_run_update_code(int command)
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
{
|
2008-05-12 13:20:43 -06:00
|
|
|
stop_machine_run(__ftrace_modify_code, &command, NR_CPUS);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static ftrace_func_t saved_ftrace_func;
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static void notrace ftrace_startup(void)
|
|
|
|
{
|
2008-05-12 13:20:43 -06:00
|
|
|
int command = 0;
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
mutex_lock(&ftraced_lock);
|
|
|
|
ftraced_suspend++;
|
2008-05-12 13:20:43 -06:00
|
|
|
if (ftraced_suspend == 1)
|
|
|
|
command |= FTRACE_ENABLE_CALLS;
|
|
|
|
|
|
|
|
if (saved_ftrace_func != ftrace_trace_function) {
|
|
|
|
saved_ftrace_func = ftrace_trace_function;
|
|
|
|
command |= FTRACE_UPDATE_TRACE_FUNC;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!command || !ftrace_enabled)
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
goto out;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
ftrace_run_update_code(command);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
out:
|
|
|
|
mutex_unlock(&ftraced_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void notrace ftrace_shutdown(void)
|
|
|
|
{
|
2008-05-12 13:20:43 -06:00
|
|
|
int command = 0;
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
mutex_lock(&ftraced_lock);
|
|
|
|
ftraced_suspend--;
|
2008-05-12 13:20:43 -06:00
|
|
|
if (!ftraced_suspend)
|
|
|
|
command |= FTRACE_DISABLE_CALLS;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
if (saved_ftrace_func != ftrace_trace_function) {
|
|
|
|
saved_ftrace_func = ftrace_trace_function;
|
|
|
|
command |= FTRACE_UPDATE_TRACE_FUNC;
|
|
|
|
}
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
if (!command || !ftrace_enabled)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
ftrace_run_update_code(command);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
out:
|
|
|
|
mutex_unlock(&ftraced_lock);
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static void notrace ftrace_startup_sysctl(void)
|
|
|
|
{
|
2008-05-12 13:20:43 -06:00
|
|
|
int command = FTRACE_ENABLE_MCOUNT;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_lock(&ftraced_lock);
|
2008-05-12 13:20:43 -06:00
|
|
|
/* Force update next time */
|
|
|
|
saved_ftrace_func = NULL;
|
2008-05-12 13:20:43 -06:00
|
|
|
/* ftraced_suspend is true if we want ftrace running */
|
|
|
|
if (ftraced_suspend)
|
2008-05-12 13:20:43 -06:00
|
|
|
command |= FTRACE_ENABLE_CALLS;
|
|
|
|
|
|
|
|
ftrace_run_update_code(command);
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_unlock(&ftraced_lock);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void notrace ftrace_shutdown_sysctl(void)
|
|
|
|
{
|
2008-05-12 13:20:43 -06:00
|
|
|
int command = FTRACE_DISABLE_MCOUNT;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_lock(&ftraced_lock);
|
|
|
|
/* ftraced_suspend is true if ftrace is running */
|
|
|
|
if (ftraced_suspend)
|
2008-05-12 13:20:43 -06:00
|
|
|
command |= FTRACE_DISABLE_CALLS;
|
|
|
|
|
|
|
|
ftrace_run_update_code(command);
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_unlock(&ftraced_lock);
|
|
|
|
}
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static cycle_t ftrace_update_time;
|
|
|
|
static unsigned long ftrace_update_cnt;
|
|
|
|
unsigned long ftrace_update_tot_cnt;
|
|
|
|
|
|
|
|
static int notrace __ftrace_update_code(void *ignore)
|
|
|
|
{
|
|
|
|
struct dyn_ftrace *p;
|
|
|
|
struct hlist_head head;
|
|
|
|
struct hlist_node *t;
|
2008-05-12 13:20:43 -06:00
|
|
|
int save_ftrace_enabled;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
cycle_t start, stop;
|
|
|
|
int i;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
/* Don't be recording funcs now */
|
|
|
|
save_ftrace_enabled = ftrace_enabled;
|
|
|
|
ftrace_enabled = 0;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:46 -06:00
|
|
|
start = ftrace_now(raw_smp_processor_id());
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
ftrace_update_cnt = 0;
|
|
|
|
|
|
|
|
/* No locks needed, the machine is stopped! */
|
|
|
|
for (i = 0; i < FTRACE_HASHSIZE; i++) {
|
|
|
|
if (hlist_empty(&ftrace_hash[i]))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
head = ftrace_hash[i];
|
|
|
|
INIT_HLIST_HEAD(&ftrace_hash[i]);
|
|
|
|
|
|
|
|
/* all CPUS are stopped, we are safe to modify code */
|
|
|
|
hlist_for_each_entry(p, t, &head, node) {
|
2008-05-12 13:20:43 -06:00
|
|
|
ftrace_code_disable(p);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
ftrace_update_cnt++;
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:46 -06:00
|
|
|
stop = ftrace_now(raw_smp_processor_id());
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
ftrace_update_time = stop - start;
|
|
|
|
ftrace_update_tot_cnt += ftrace_update_cnt;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
ftrace_enabled = save_ftrace_enabled;
|
2008-05-12 13:20:42 -06:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
static void notrace ftrace_update_code(void)
|
|
|
|
{
|
|
|
|
stop_machine_run(__ftrace_update_code, NULL, NR_CPUS);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int notrace ftraced(void *ignore)
|
|
|
|
{
|
|
|
|
unsigned long usecs;
|
|
|
|
|
|
|
|
set_current_state(TASK_INTERRUPTIBLE);
|
|
|
|
|
|
|
|
while (!kthread_should_stop()) {
|
|
|
|
|
|
|
|
/* check once a second */
|
|
|
|
schedule_timeout(HZ);
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_lock(&ftrace_sysctl_lock);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
mutex_lock(&ftraced_lock);
|
2008-05-12 13:20:43 -06:00
|
|
|
if (ftrace_enabled && ftraced_trigger && !ftraced_suspend) {
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
ftrace_record_suspend++;
|
|
|
|
ftrace_update_code();
|
|
|
|
usecs = nsecs_to_usecs(ftrace_update_time);
|
|
|
|
if (ftrace_update_tot_cnt > 100000) {
|
|
|
|
ftrace_update_tot_cnt = 0;
|
|
|
|
pr_info("hm, dftrace overflow: %lu change%s"
|
|
|
|
" (%lu total) in %lu usec%s\n",
|
|
|
|
ftrace_update_cnt,
|
|
|
|
ftrace_update_cnt != 1 ? "s" : "",
|
|
|
|
ftrace_update_tot_cnt,
|
|
|
|
usecs, usecs != 1 ? "s" : "");
|
|
|
|
WARN_ON_ONCE(1);
|
|
|
|
}
|
|
|
|
ftraced_trigger = 0;
|
|
|
|
ftrace_record_suspend--;
|
|
|
|
}
|
2008-05-12 13:20:44 -06:00
|
|
|
ftraced_iteration_counter++;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
mutex_unlock(&ftraced_lock);
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_unlock(&ftrace_sysctl_lock);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
2008-05-12 13:20:44 -06:00
|
|
|
wake_up_interruptible(&ftraced_waiters);
|
|
|
|
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
ftrace_shutdown_replenish();
|
|
|
|
|
|
|
|
set_current_state(TASK_INTERRUPTIBLE);
|
|
|
|
}
|
|
|
|
__set_current_state(TASK_RUNNING);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static int __init ftrace_dyn_table_alloc(void)
|
|
|
|
{
|
|
|
|
struct ftrace_page *pg;
|
|
|
|
int cnt;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
/* allocate a few pages */
|
|
|
|
ftrace_pages_start = (void *)get_zeroed_page(GFP_KERNEL);
|
|
|
|
if (!ftrace_pages_start)
|
|
|
|
return -1;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocate a few more pages.
|
|
|
|
*
|
|
|
|
* TODO: have some parser search vmlinux before
|
|
|
|
* final linking to find all calls to ftrace.
|
|
|
|
* Then we can:
|
|
|
|
* a) know how many pages to allocate.
|
|
|
|
* and/or
|
|
|
|
* b) set up the table then.
|
|
|
|
*
|
|
|
|
* The dynamic code is still necessary for
|
|
|
|
* modules.
|
|
|
|
*/
|
|
|
|
|
|
|
|
pg = ftrace_pages = ftrace_pages_start;
|
|
|
|
|
|
|
|
cnt = NR_TO_INIT / ENTRIES_PER_PAGE;
|
|
|
|
|
|
|
|
for (i = 0; i < cnt; i++) {
|
|
|
|
pg->next = (void *)get_zeroed_page(GFP_KERNEL);
|
|
|
|
|
|
|
|
/* If we fail, we'll try later anyway */
|
|
|
|
if (!pg->next)
|
|
|
|
break;
|
|
|
|
|
|
|
|
pg = pg->next;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
enum {
|
|
|
|
FTRACE_ITER_FILTER = (1 << 0),
|
|
|
|
FTRACE_ITER_CONT = (1 << 1),
|
|
|
|
};
|
|
|
|
|
|
|
|
#define FTRACE_BUFF_MAX (KSYM_SYMBOL_LEN+4) /* room for wildcards */
|
|
|
|
|
|
|
|
struct ftrace_iterator {
|
|
|
|
loff_t pos;
|
|
|
|
struct ftrace_page *pg;
|
|
|
|
unsigned idx;
|
|
|
|
unsigned flags;
|
|
|
|
unsigned char buffer[FTRACE_BUFF_MAX+1];
|
|
|
|
unsigned buffer_idx;
|
|
|
|
unsigned filtered;
|
|
|
|
};
|
|
|
|
|
|
|
|
static void notrace *
|
|
|
|
t_next(struct seq_file *m, void *v, loff_t *pos)
|
|
|
|
{
|
|
|
|
struct ftrace_iterator *iter = m->private;
|
|
|
|
struct dyn_ftrace *rec = NULL;
|
|
|
|
|
|
|
|
(*pos)++;
|
|
|
|
|
|
|
|
retry:
|
|
|
|
if (iter->idx >= iter->pg->index) {
|
|
|
|
if (iter->pg->next) {
|
|
|
|
iter->pg = iter->pg->next;
|
|
|
|
iter->idx = 0;
|
|
|
|
goto retry;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
rec = &iter->pg->records[iter->idx++];
|
|
|
|
if ((rec->flags & FTRACE_FL_FAILED) ||
|
|
|
|
((iter->flags & FTRACE_ITER_FILTER) &&
|
|
|
|
!(rec->flags & FTRACE_FL_FILTER))) {
|
|
|
|
rec = NULL;
|
|
|
|
goto retry;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
iter->pos = *pos;
|
|
|
|
|
|
|
|
return rec;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void *t_start(struct seq_file *m, loff_t *pos)
|
|
|
|
{
|
|
|
|
struct ftrace_iterator *iter = m->private;
|
|
|
|
void *p = NULL;
|
|
|
|
loff_t l = -1;
|
|
|
|
|
|
|
|
if (*pos != iter->pos) {
|
|
|
|
for (p = t_next(m, p, &l); p && l < *pos; p = t_next(m, p, &l))
|
|
|
|
;
|
|
|
|
} else {
|
|
|
|
l = *pos;
|
|
|
|
p = t_next(m, p, &l);
|
|
|
|
}
|
|
|
|
|
|
|
|
return p;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void t_stop(struct seq_file *m, void *p)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
static int t_show(struct seq_file *m, void *v)
|
|
|
|
{
|
|
|
|
struct dyn_ftrace *rec = v;
|
|
|
|
char str[KSYM_SYMBOL_LEN];
|
|
|
|
|
|
|
|
if (!rec)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
kallsyms_lookup(rec->ip, NULL, NULL, NULL, str);
|
|
|
|
|
|
|
|
seq_printf(m, "%s\n", str);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct seq_operations show_ftrace_seq_ops = {
|
|
|
|
.start = t_start,
|
|
|
|
.next = t_next,
|
|
|
|
.stop = t_stop,
|
|
|
|
.show = t_show,
|
|
|
|
};
|
|
|
|
|
|
|
|
static int notrace
|
|
|
|
ftrace_avail_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct ftrace_iterator *iter;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
iter = kzalloc(sizeof(*iter), GFP_KERNEL);
|
|
|
|
if (!iter)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
iter->pg = ftrace_pages_start;
|
|
|
|
iter->pos = -1;
|
|
|
|
|
|
|
|
ret = seq_open(file, &show_ftrace_seq_ops);
|
|
|
|
if (!ret) {
|
|
|
|
struct seq_file *m = file->private_data;
|
2008-05-12 13:20:46 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
m->private = iter;
|
2008-05-12 13:20:46 -06:00
|
|
|
} else {
|
2008-05-12 13:20:43 -06:00
|
|
|
kfree(iter);
|
2008-05-12 13:20:46 -06:00
|
|
|
}
|
2008-05-12 13:20:43 -06:00
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
int ftrace_avail_release(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct seq_file *m = (struct seq_file *)file->private_data;
|
|
|
|
struct ftrace_iterator *iter = m->private;
|
|
|
|
|
|
|
|
seq_release(inode, file);
|
|
|
|
kfree(iter);
|
2008-05-12 13:20:46 -06:00
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void notrace ftrace_filter_reset(void)
|
|
|
|
{
|
|
|
|
struct ftrace_page *pg;
|
|
|
|
struct dyn_ftrace *rec;
|
|
|
|
unsigned i;
|
|
|
|
|
|
|
|
/* keep kstop machine from running */
|
|
|
|
preempt_disable();
|
|
|
|
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 &= ~FTRACE_FL_FILTER;
|
|
|
|
}
|
|
|
|
pg = pg->next;
|
|
|
|
}
|
|
|
|
preempt_enable();
|
|
|
|
}
|
|
|
|
|
|
|
|
static int notrace
|
|
|
|
ftrace_filter_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct ftrace_iterator *iter;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
iter = kzalloc(sizeof(*iter), GFP_KERNEL);
|
|
|
|
if (!iter)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
mutex_lock(&ftrace_filter_lock);
|
|
|
|
if ((file->f_mode & FMODE_WRITE) &&
|
|
|
|
!(file->f_flags & O_APPEND))
|
|
|
|
ftrace_filter_reset();
|
|
|
|
|
|
|
|
if (file->f_mode & FMODE_READ) {
|
|
|
|
iter->pg = ftrace_pages_start;
|
|
|
|
iter->pos = -1;
|
|
|
|
iter->flags = FTRACE_ITER_FILTER;
|
|
|
|
|
|
|
|
ret = seq_open(file, &show_ftrace_seq_ops);
|
|
|
|
if (!ret) {
|
|
|
|
struct seq_file *m = file->private_data;
|
|
|
|
m->private = iter;
|
|
|
|
} else
|
|
|
|
kfree(iter);
|
|
|
|
} else
|
|
|
|
file->private_data = iter;
|
|
|
|
mutex_unlock(&ftrace_filter_lock);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static ssize_t notrace
|
|
|
|
ftrace_filter_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 notrace
|
|
|
|
ftrace_filter_lseek(struct file *file, loff_t offset, int origin)
|
|
|
|
{
|
|
|
|
loff_t ret;
|
|
|
|
|
|
|
|
if (file->f_mode & FMODE_READ)
|
|
|
|
ret = seq_lseek(file, offset, origin);
|
|
|
|
else
|
|
|
|
file->f_pos = ret = 1;
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
enum {
|
|
|
|
MATCH_FULL,
|
|
|
|
MATCH_FRONT_ONLY,
|
|
|
|
MATCH_MIDDLE_ONLY,
|
|
|
|
MATCH_END_ONLY,
|
|
|
|
};
|
|
|
|
|
|
|
|
static void notrace
|
|
|
|
ftrace_match(unsigned char *buff, int len)
|
|
|
|
{
|
|
|
|
char str[KSYM_SYMBOL_LEN];
|
|
|
|
char *search = NULL;
|
|
|
|
struct ftrace_page *pg;
|
|
|
|
struct dyn_ftrace *rec;
|
|
|
|
int type = MATCH_FULL;
|
|
|
|
unsigned i, match = 0, search_len = 0;
|
|
|
|
|
|
|
|
for (i = 0; i < len; i++) {
|
|
|
|
if (buff[i] == '*') {
|
|
|
|
if (!i) {
|
|
|
|
search = buff + i + 1;
|
|
|
|
type = MATCH_END_ONLY;
|
|
|
|
search_len = len - (i + 1);
|
|
|
|
} else {
|
|
|
|
if (type == MATCH_END_ONLY) {
|
|
|
|
type = MATCH_MIDDLE_ONLY;
|
|
|
|
} else {
|
|
|
|
match = i;
|
|
|
|
type = MATCH_FRONT_ONLY;
|
|
|
|
}
|
|
|
|
buff[i] = 0;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* keep kstop machine from running */
|
|
|
|
preempt_disable();
|
|
|
|
ftrace_filtered = 1;
|
|
|
|
pg = ftrace_pages_start;
|
|
|
|
while (pg) {
|
|
|
|
for (i = 0; i < pg->index; i++) {
|
|
|
|
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)
|
|
|
|
rec->flags |= FTRACE_FL_FILTER;
|
|
|
|
}
|
|
|
|
pg = pg->next;
|
|
|
|
}
|
|
|
|
preempt_enable();
|
|
|
|
}
|
|
|
|
|
|
|
|
static ssize_t notrace
|
|
|
|
ftrace_filter_write(struct file *file, const char __user *ubuf,
|
|
|
|
size_t cnt, loff_t *ppos)
|
|
|
|
{
|
|
|
|
struct ftrace_iterator *iter;
|
|
|
|
char ch;
|
|
|
|
size_t read = 0;
|
|
|
|
ssize_t ret;
|
|
|
|
|
|
|
|
if (!cnt || cnt < 0)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
mutex_lock(&ftrace_filter_lock);
|
|
|
|
|
|
|
|
if (file->f_mode & FMODE_READ) {
|
|
|
|
struct seq_file *m = file->private_data;
|
|
|
|
iter = m->private;
|
|
|
|
} else
|
|
|
|
iter = file->private_data;
|
|
|
|
|
|
|
|
if (!*ppos) {
|
|
|
|
iter->flags &= ~FTRACE_ITER_CONT;
|
|
|
|
iter->buffer_idx = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = get_user(ch, ubuf++);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
read++;
|
|
|
|
cnt--;
|
|
|
|
|
|
|
|
if (!(iter->flags & ~FTRACE_ITER_CONT)) {
|
|
|
|
/* skip white space */
|
|
|
|
while (cnt && isspace(ch)) {
|
|
|
|
ret = get_user(ch, ubuf++);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
read++;
|
|
|
|
cnt--;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
if (isspace(ch)) {
|
|
|
|
file->f_pos += read;
|
|
|
|
ret = read;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
iter->buffer_idx = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
while (cnt && !isspace(ch)) {
|
|
|
|
if (iter->buffer_idx < FTRACE_BUFF_MAX)
|
|
|
|
iter->buffer[iter->buffer_idx++] = ch;
|
|
|
|
else {
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
ret = get_user(ch, ubuf++);
|
|
|
|
if (ret)
|
|
|
|
goto out;
|
|
|
|
read++;
|
|
|
|
cnt--;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (isspace(ch)) {
|
|
|
|
iter->filtered++;
|
|
|
|
iter->buffer[iter->buffer_idx] = 0;
|
|
|
|
ftrace_match(iter->buffer, iter->buffer_idx);
|
|
|
|
iter->buffer_idx = 0;
|
|
|
|
} else
|
|
|
|
iter->flags |= FTRACE_ITER_CONT;
|
|
|
|
|
|
|
|
|
|
|
|
file->f_pos += read;
|
|
|
|
|
|
|
|
ret = read;
|
|
|
|
out:
|
|
|
|
mutex_unlock(&ftrace_filter_lock);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:45 -06:00
|
|
|
/**
|
|
|
|
* ftrace_set_filter - set a function to filter on in ftrace
|
|
|
|
* @buf - the string that holds the function filter text.
|
|
|
|
* @len - the length of the string.
|
|
|
|
* @reset - non zero to reset all filters before applying this filter.
|
|
|
|
*
|
|
|
|
* Filters denote which functions should be enabled when tracing is enabled.
|
|
|
|
* If @buf is NULL and reset is set, all functions will be enabled for tracing.
|
|
|
|
*/
|
|
|
|
notrace void ftrace_set_filter(unsigned char *buf, int len, int reset)
|
|
|
|
{
|
|
|
|
mutex_lock(&ftrace_filter_lock);
|
|
|
|
if (reset)
|
|
|
|
ftrace_filter_reset();
|
|
|
|
if (buf)
|
|
|
|
ftrace_match(buf, len);
|
|
|
|
mutex_unlock(&ftrace_filter_lock);
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static int notrace
|
|
|
|
ftrace_filter_release(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct seq_file *m = (struct seq_file *)file->private_data;
|
|
|
|
struct ftrace_iterator *iter;
|
|
|
|
|
|
|
|
mutex_lock(&ftrace_filter_lock);
|
|
|
|
if (file->f_mode & FMODE_READ) {
|
|
|
|
iter = m->private;
|
|
|
|
|
|
|
|
seq_release(inode, file);
|
|
|
|
} else
|
|
|
|
iter = file->private_data;
|
|
|
|
|
|
|
|
if (iter->buffer_idx) {
|
|
|
|
iter->filtered++;
|
|
|
|
iter->buffer[iter->buffer_idx] = 0;
|
|
|
|
ftrace_match(iter->buffer, iter->buffer_idx);
|
|
|
|
}
|
|
|
|
|
|
|
|
mutex_lock(&ftrace_sysctl_lock);
|
|
|
|
mutex_lock(&ftraced_lock);
|
|
|
|
if (iter->filtered && ftraced_suspend && ftrace_enabled)
|
|
|
|
ftrace_run_update_code(FTRACE_ENABLE_CALLS);
|
|
|
|
mutex_unlock(&ftraced_lock);
|
|
|
|
mutex_unlock(&ftrace_sysctl_lock);
|
|
|
|
|
|
|
|
kfree(iter);
|
|
|
|
mutex_unlock(&ftrace_filter_lock);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static 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_filter_fops = {
|
|
|
|
.open = ftrace_filter_open,
|
|
|
|
.read = ftrace_filter_read,
|
|
|
|
.write = ftrace_filter_write,
|
|
|
|
.llseek = ftrace_filter_lseek,
|
|
|
|
.release = ftrace_filter_release,
|
|
|
|
};
|
|
|
|
|
2008-05-12 13:20:44 -06:00
|
|
|
/**
|
|
|
|
* ftrace_force_update - force an update to all recording ftrace functions
|
|
|
|
*
|
|
|
|
* The ftrace dynamic update daemon only wakes up once a second.
|
|
|
|
* There may be cases where an update needs to be done immediately
|
|
|
|
* for tests or internal kernel tracing to begin. This function
|
|
|
|
* wakes the daemon to do an update and will not return until the
|
|
|
|
* update is complete.
|
|
|
|
*/
|
|
|
|
int ftrace_force_update(void)
|
|
|
|
{
|
|
|
|
unsigned long last_counter;
|
|
|
|
DECLARE_WAITQUEUE(wait, current);
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
if (!ftraced_task)
|
|
|
|
return -ENODEV;
|
|
|
|
|
|
|
|
mutex_lock(&ftraced_lock);
|
|
|
|
last_counter = ftraced_iteration_counter;
|
|
|
|
|
|
|
|
set_current_state(TASK_INTERRUPTIBLE);
|
|
|
|
add_wait_queue(&ftraced_waiters, &wait);
|
|
|
|
|
|
|
|
do {
|
|
|
|
mutex_unlock(&ftraced_lock);
|
|
|
|
wake_up_process(ftraced_task);
|
|
|
|
schedule();
|
|
|
|
mutex_lock(&ftraced_lock);
|
|
|
|
if (signal_pending(current)) {
|
|
|
|
ret = -EINTR;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
set_current_state(TASK_INTERRUPTIBLE);
|
|
|
|
} while (last_counter == ftraced_iteration_counter);
|
|
|
|
|
|
|
|
mutex_unlock(&ftraced_lock);
|
|
|
|
remove_wait_queue(&ftraced_waiters, &wait);
|
|
|
|
set_current_state(TASK_RUNNING);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static __init int ftrace_init_debugfs(void)
|
|
|
|
{
|
|
|
|
struct dentry *d_tracer;
|
|
|
|
struct dentry *entry;
|
|
|
|
|
|
|
|
d_tracer = tracing_init_dentry();
|
|
|
|
|
|
|
|
entry = debugfs_create_file("available_filter_functions", 0444,
|
|
|
|
d_tracer, NULL, &ftrace_avail_fops);
|
|
|
|
if (!entry)
|
|
|
|
pr_warning("Could not create debugfs "
|
|
|
|
"'available_filter_functions' entry\n");
|
|
|
|
|
|
|
|
entry = debugfs_create_file("set_ftrace_filter", 0644, d_tracer,
|
|
|
|
NULL, &ftrace_filter_fops);
|
|
|
|
if (!entry)
|
|
|
|
pr_warning("Could not create debugfs "
|
|
|
|
"'set_ftrace_filter' entry\n");
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
fs_initcall(ftrace_init_debugfs);
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
static int __init notrace ftrace_dynamic_init(void)
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
{
|
|
|
|
struct task_struct *p;
|
2008-05-12 13:20:43 -06:00
|
|
|
unsigned long addr;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
int ret;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
addr = (unsigned long)ftrace_record_ip;
|
|
|
|
stop_machine_run(ftrace_dyn_arch_init, &addr, NR_CPUS);
|
|
|
|
|
|
|
|
/* ftrace_dyn_arch_init places the return code in addr */
|
|
|
|
if (addr)
|
|
|
|
return addr;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
ret = ftrace_dyn_table_alloc();
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
p = kthread_run(ftraced, NULL, "ftraced");
|
|
|
|
if (IS_ERR(p))
|
|
|
|
return -1;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
last_ftrace_enabled = ftrace_enabled = 1;
|
2008-05-12 13:20:44 -06:00
|
|
|
ftraced_task = p;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
core_initcall(ftrace_dynamic_init);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
#else
|
2008-05-12 13:20:45 -06:00
|
|
|
# define ftrace_startup() do { } while (0)
|
|
|
|
# define ftrace_shutdown() do { } while (0)
|
|
|
|
# define ftrace_startup_sysctl() do { } while (0)
|
|
|
|
# define ftrace_shutdown_sysctl() do { } while (0)
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
#endif /* CONFIG_DYNAMIC_FTRACE */
|
|
|
|
|
2008-05-12 13:20:42 -06:00
|
|
|
/**
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
* register_ftrace_function - register a function for profiling
|
|
|
|
* @ops - ops structure that holds the function for profiling.
|
2008-05-12 13:20:42 -06:00
|
|
|
*
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
* Register a function to be called by all functions in the
|
|
|
|
* kernel.
|
|
|
|
*
|
|
|
|
* Note: @ops->func and all the functions it calls must be labeled
|
|
|
|
* with "notrace", otherwise it will go into a
|
|
|
|
* recursive loop.
|
2008-05-12 13:20:42 -06:00
|
|
|
*/
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
int register_ftrace_function(struct ftrace_ops *ops)
|
2008-05-12 13:20:42 -06:00
|
|
|
{
|
2008-05-12 13:20:43 -06:00
|
|
|
int ret;
|
|
|
|
|
|
|
|
mutex_lock(&ftrace_sysctl_lock);
|
|
|
|
ret = __register_ftrace_function(ops);
|
2008-05-12 13:20:43 -06:00
|
|
|
ftrace_startup();
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_unlock(&ftrace_sysctl_lock);
|
|
|
|
|
|
|
|
return ret;
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* unregister_ftrace_function - unresgister 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.
|
|
|
|
*/
|
|
|
|
int unregister_ftrace_function(struct ftrace_ops *ops)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_lock(&ftrace_sysctl_lock);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
ret = __unregister_ftrace_function(ops);
|
2008-05-12 13:20:43 -06:00
|
|
|
ftrace_shutdown();
|
2008-05-12 13:20:43 -06:00
|
|
|
mutex_unlock(&ftrace_sysctl_lock);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
notrace int
|
|
|
|
ftrace_enable_sysctl(struct ctl_table *table, int write,
|
2008-05-12 13:20:43 -06:00
|
|
|
struct file *file, void __user *buffer, size_t *lenp,
|
2008-05-12 13:20:43 -06:00
|
|
|
loff_t *ppos)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
mutex_lock(&ftrace_sysctl_lock);
|
|
|
|
|
2008-05-12 13:20:43 -06:00
|
|
|
ret = proc_dointvec(table, write, file, buffer, lenp, ppos);
|
2008-05-12 13:20:43 -06:00
|
|
|
|
|
|
|
if (ret || !write || (last_ftrace_enabled == ftrace_enabled))
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
last_ftrace_enabled = ftrace_enabled;
|
|
|
|
|
|
|
|
if (ftrace_enabled) {
|
|
|
|
|
|
|
|
ftrace_startup_sysctl();
|
|
|
|
|
|
|
|
/* we are starting ftrace again */
|
|
|
|
if (ftrace_list != &ftrace_list_end) {
|
|
|
|
if (ftrace_list->next == &ftrace_list_end)
|
|
|
|
ftrace_trace_function = ftrace_list->func;
|
|
|
|
else
|
|
|
|
ftrace_trace_function = ftrace_list_func;
|
|
|
|
}
|
|
|
|
|
|
|
|
} else {
|
|
|
|
/* stopping ftrace calls (just send to ftrace_stub) */
|
|
|
|
ftrace_trace_function = ftrace_stub;
|
|
|
|
|
|
|
|
ftrace_shutdown_sysctl();
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
mutex_unlock(&ftrace_sysctl_lock);
|
ftrace: dynamic enabling/disabling of function calls
This patch adds a feature to dynamically replace the ftrace code
with the jmps to allow a kernel with ftrace configured to run
as fast as it can without it configured.
The way this works, is on bootup (if ftrace is enabled), a ftrace
function is registered to record the instruction pointer of all
places that call the function.
Later, if there's still any code to patch, a kthread is awoken
(rate limited to at most once a second) that performs a stop_machine,
and replaces all the code that was called with a jmp over the call
to ftrace. It only replaces what was found the previous time. Typically
the system reaches equilibrium quickly after bootup and there's no code
patching needed at all.
e.g.
call ftrace /* 5 bytes */
is replaced with
jmp 3f /* jmp is 2 bytes and we jump 3 forward */
3:
When we want to enable ftrace for function tracing, the IP recording
is removed, and stop_machine is called again to replace all the locations
of that were recorded back to the call of ftrace. When it is disabled,
we replace the code back to the jmp.
Allocation is done by the kthread. If the ftrace recording function is
called, and we don't have any record slots available, then we simply
skip that call. Once a second a new page (if needed) is allocated for
recording new ftrace function calls. A large batch is allocated at
boot up to get most of the calls there.
Because we do this via stop_machine, we don't have to worry about another
CPU executing a ftrace call as we modify it. But we do need to worry
about NMI's so all functions that might be called via nmi must be
annotated with notrace_nmi. When this code is configured in, the NMI code
will not call notrace.
Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2008-05-12 13:20:42 -06:00
|
|
|
return ret;
|
2008-05-12 13:20:42 -06:00
|
|
|
}
|