2005-04-16 16:20:36 -06:00
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/hugetlb.h>
|
|
|
|
#include <linux/mount.h>
|
|
|
|
#include <linux/seq_file.h>
|
2005-09-03 16:55:10 -06:00
|
|
|
#include <linux/highmem.h>
|
2005-09-03 16:54:45 -06:00
|
|
|
#include <linux/pagemap.h>
|
|
|
|
#include <linux/mempolicy.h>
|
2005-09-03 16:55:10 -06:00
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
#include <asm/elf.h>
|
|
|
|
#include <asm/uaccess.h>
|
2005-09-03 16:55:10 -06:00
|
|
|
#include <asm/tlbflush.h>
|
2005-04-16 16:20:36 -06:00
|
|
|
#include "internal.h"
|
|
|
|
|
|
|
|
char *task_mem(struct mm_struct *mm, char *buffer)
|
|
|
|
{
|
|
|
|
unsigned long data, text, lib;
|
[PATCH] mm: update_hiwaters just in time
update_mem_hiwater has attracted various criticisms, in particular from those
concerned with mm scalability. Originally it was called whenever rss or
total_vm got raised. Then many of those callsites were replaced by a timer
tick call from account_system_time. Now Frank van Maarseveen reports that to
be found inadequate. How about this? Works for Frank.
Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros
update_hiwater_rss and update_hiwater_vm. Don't attempt to keep
mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually
by 1): those are hot paths. Do the opposite, update only when about to lower
rss (usually by many), or just before final accounting in do_exit. Handle
mm->hiwater_vm in the same way, though it's much less of an issue. Demand
that whoever collects these hiwater statistics do the work of taking the
maximum with rss or total_vm.
And there has been no collector of these hiwater statistics in the tree. The
new convention needs an example, so match Frank's usage by adding a VmPeak
line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS
(High-Water-Mark or High-Water-Memory).
There was a particular anomaly during mremap move, that hiwater_vm might be
captured too high. A fleeting such anomaly remains, but it's quickly
corrected now, whereas before it would stick.
What locking? None: if the app is racy then these statistics will be racy,
it's not worth any overhead to make them exact. But whenever it suits,
hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under
page_table_lock (for now) or with preemption disabled (later on): without
going to any trouble, minimize the time between reading current values and
updating, to minimize those occasions when a racing thread bumps a count up
and back down in between.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-29 19:16:18 -06:00
|
|
|
unsigned long hiwater_vm, total_vm, hiwater_rss, total_rss;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Note: to minimize their overhead, mm maintains hiwater_vm and
|
|
|
|
* hiwater_rss only when about to *lower* total_vm or rss. Any
|
|
|
|
* collector of these hiwater stats must therefore get total_vm
|
|
|
|
* and rss too, which will usually be the higher. Barriers? not
|
|
|
|
* worth the effort, such snapshots can always be inconsistent.
|
|
|
|
*/
|
|
|
|
hiwater_vm = total_vm = mm->total_vm;
|
|
|
|
if (hiwater_vm < mm->hiwater_vm)
|
|
|
|
hiwater_vm = mm->hiwater_vm;
|
|
|
|
hiwater_rss = total_rss = get_mm_rss(mm);
|
|
|
|
if (hiwater_rss < mm->hiwater_rss)
|
|
|
|
hiwater_rss = mm->hiwater_rss;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
data = mm->total_vm - mm->shared_vm - mm->stack_vm;
|
|
|
|
text = (PAGE_ALIGN(mm->end_code) - (mm->start_code & PAGE_MASK)) >> 10;
|
|
|
|
lib = (mm->exec_vm << (PAGE_SHIFT-10)) - text;
|
|
|
|
buffer += sprintf(buffer,
|
[PATCH] mm: update_hiwaters just in time
update_mem_hiwater has attracted various criticisms, in particular from those
concerned with mm scalability. Originally it was called whenever rss or
total_vm got raised. Then many of those callsites were replaced by a timer
tick call from account_system_time. Now Frank van Maarseveen reports that to
be found inadequate. How about this? Works for Frank.
Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros
update_hiwater_rss and update_hiwater_vm. Don't attempt to keep
mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually
by 1): those are hot paths. Do the opposite, update only when about to lower
rss (usually by many), or just before final accounting in do_exit. Handle
mm->hiwater_vm in the same way, though it's much less of an issue. Demand
that whoever collects these hiwater statistics do the work of taking the
maximum with rss or total_vm.
And there has been no collector of these hiwater statistics in the tree. The
new convention needs an example, so match Frank's usage by adding a VmPeak
line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS
(High-Water-Mark or High-Water-Memory).
There was a particular anomaly during mremap move, that hiwater_vm might be
captured too high. A fleeting such anomaly remains, but it's quickly
corrected now, whereas before it would stick.
What locking? None: if the app is racy then these statistics will be racy,
it's not worth any overhead to make them exact. But whenever it suits,
hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under
page_table_lock (for now) or with preemption disabled (later on): without
going to any trouble, minimize the time between reading current values and
updating, to minimize those occasions when a racing thread bumps a count up
and back down in between.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-29 19:16:18 -06:00
|
|
|
"VmPeak:\t%8lu kB\n"
|
2005-04-16 16:20:36 -06:00
|
|
|
"VmSize:\t%8lu kB\n"
|
|
|
|
"VmLck:\t%8lu kB\n"
|
[PATCH] mm: update_hiwaters just in time
update_mem_hiwater has attracted various criticisms, in particular from those
concerned with mm scalability. Originally it was called whenever rss or
total_vm got raised. Then many of those callsites were replaced by a timer
tick call from account_system_time. Now Frank van Maarseveen reports that to
be found inadequate. How about this? Works for Frank.
Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros
update_hiwater_rss and update_hiwater_vm. Don't attempt to keep
mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually
by 1): those are hot paths. Do the opposite, update only when about to lower
rss (usually by many), or just before final accounting in do_exit. Handle
mm->hiwater_vm in the same way, though it's much less of an issue. Demand
that whoever collects these hiwater statistics do the work of taking the
maximum with rss or total_vm.
And there has been no collector of these hiwater statistics in the tree. The
new convention needs an example, so match Frank's usage by adding a VmPeak
line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS
(High-Water-Mark or High-Water-Memory).
There was a particular anomaly during mremap move, that hiwater_vm might be
captured too high. A fleeting such anomaly remains, but it's quickly
corrected now, whereas before it would stick.
What locking? None: if the app is racy then these statistics will be racy,
it's not worth any overhead to make them exact. But whenever it suits,
hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under
page_table_lock (for now) or with preemption disabled (later on): without
going to any trouble, minimize the time between reading current values and
updating, to minimize those occasions when a racing thread bumps a count up
and back down in between.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-29 19:16:18 -06:00
|
|
|
"VmHWM:\t%8lu kB\n"
|
2005-04-16 16:20:36 -06:00
|
|
|
"VmRSS:\t%8lu kB\n"
|
|
|
|
"VmData:\t%8lu kB\n"
|
|
|
|
"VmStk:\t%8lu kB\n"
|
|
|
|
"VmExe:\t%8lu kB\n"
|
|
|
|
"VmLib:\t%8lu kB\n"
|
|
|
|
"VmPTE:\t%8lu kB\n",
|
[PATCH] mm: update_hiwaters just in time
update_mem_hiwater has attracted various criticisms, in particular from those
concerned with mm scalability. Originally it was called whenever rss or
total_vm got raised. Then many of those callsites were replaced by a timer
tick call from account_system_time. Now Frank van Maarseveen reports that to
be found inadequate. How about this? Works for Frank.
Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros
update_hiwater_rss and update_hiwater_vm. Don't attempt to keep
mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually
by 1): those are hot paths. Do the opposite, update only when about to lower
rss (usually by many), or just before final accounting in do_exit. Handle
mm->hiwater_vm in the same way, though it's much less of an issue. Demand
that whoever collects these hiwater statistics do the work of taking the
maximum with rss or total_vm.
And there has been no collector of these hiwater statistics in the tree. The
new convention needs an example, so match Frank's usage by adding a VmPeak
line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS
(High-Water-Mark or High-Water-Memory).
There was a particular anomaly during mremap move, that hiwater_vm might be
captured too high. A fleeting such anomaly remains, but it's quickly
corrected now, whereas before it would stick.
What locking? None: if the app is racy then these statistics will be racy,
it's not worth any overhead to make them exact. But whenever it suits,
hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under
page_table_lock (for now) or with preemption disabled (later on): without
going to any trouble, minimize the time between reading current values and
updating, to minimize those occasions when a racing thread bumps a count up
and back down in between.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-29 19:16:18 -06:00
|
|
|
hiwater_vm << (PAGE_SHIFT-10),
|
|
|
|
(total_vm - mm->reserved_vm) << (PAGE_SHIFT-10),
|
2005-04-16 16:20:36 -06:00
|
|
|
mm->locked_vm << (PAGE_SHIFT-10),
|
[PATCH] mm: update_hiwaters just in time
update_mem_hiwater has attracted various criticisms, in particular from those
concerned with mm scalability. Originally it was called whenever rss or
total_vm got raised. Then many of those callsites were replaced by a timer
tick call from account_system_time. Now Frank van Maarseveen reports that to
be found inadequate. How about this? Works for Frank.
Replace update_mem_hiwater, a poor combination of two unrelated ops, by macros
update_hiwater_rss and update_hiwater_vm. Don't attempt to keep
mm->hiwater_rss up to date at timer tick, nor every time we raise rss (usually
by 1): those are hot paths. Do the opposite, update only when about to lower
rss (usually by many), or just before final accounting in do_exit. Handle
mm->hiwater_vm in the same way, though it's much less of an issue. Demand
that whoever collects these hiwater statistics do the work of taking the
maximum with rss or total_vm.
And there has been no collector of these hiwater statistics in the tree. The
new convention needs an example, so match Frank's usage by adding a VmPeak
line above VmSize to /proc/<pid>/status, and also a VmHWM line above VmRSS
(High-Water-Mark or High-Water-Memory).
There was a particular anomaly during mremap move, that hiwater_vm might be
captured too high. A fleeting such anomaly remains, but it's quickly
corrected now, whereas before it would stick.
What locking? None: if the app is racy then these statistics will be racy,
it's not worth any overhead to make them exact. But whenever it suits,
hiwater_vm is updated under exclusive mmap_sem, and hiwater_rss under
page_table_lock (for now) or with preemption disabled (later on): without
going to any trouble, minimize the time between reading current values and
updating, to minimize those occasions when a racing thread bumps a count up
and back down in between.
Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-10-29 19:16:18 -06:00
|
|
|
hiwater_rss << (PAGE_SHIFT-10),
|
|
|
|
total_rss << (PAGE_SHIFT-10),
|
2005-04-16 16:20:36 -06:00
|
|
|
data << (PAGE_SHIFT-10),
|
|
|
|
mm->stack_vm << (PAGE_SHIFT-10), text, lib,
|
|
|
|
(PTRS_PER_PTE*sizeof(pte_t)*mm->nr_ptes) >> 10);
|
|
|
|
return buffer;
|
|
|
|
}
|
|
|
|
|
|
|
|
unsigned long task_vsize(struct mm_struct *mm)
|
|
|
|
{
|
|
|
|
return PAGE_SIZE * mm->total_vm;
|
|
|
|
}
|
|
|
|
|
|
|
|
int task_statm(struct mm_struct *mm, int *shared, int *text,
|
|
|
|
int *data, int *resident)
|
|
|
|
{
|
2005-10-29 19:16:05 -06:00
|
|
|
*shared = get_mm_counter(mm, file_rss);
|
2005-04-16 16:20:36 -06:00
|
|
|
*text = (PAGE_ALIGN(mm->end_code) - (mm->start_code & PAGE_MASK))
|
|
|
|
>> PAGE_SHIFT;
|
|
|
|
*data = mm->total_vm - mm->shared_vm;
|
2005-10-29 19:16:05 -06:00
|
|
|
*resident = *shared + get_mm_counter(mm, anon_rss);
|
2005-04-16 16:20:36 -06:00
|
|
|
return mm->total_vm;
|
|
|
|
}
|
|
|
|
|
|
|
|
int proc_exe_link(struct inode *inode, struct dentry **dentry, struct vfsmount **mnt)
|
|
|
|
{
|
|
|
|
struct vm_area_struct * vma;
|
|
|
|
int result = -ENOENT;
|
2006-06-26 01:25:55 -06:00
|
|
|
struct task_struct *task = get_proc_task(inode);
|
|
|
|
struct mm_struct * mm = NULL;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2006-06-26 01:25:55 -06:00
|
|
|
if (task) {
|
|
|
|
mm = get_task_mm(task);
|
|
|
|
put_task_struct(task);
|
|
|
|
}
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!mm)
|
|
|
|
goto out;
|
|
|
|
down_read(&mm->mmap_sem);
|
|
|
|
|
|
|
|
vma = mm->mmap;
|
|
|
|
while (vma) {
|
|
|
|
if ((vma->vm_flags & VM_EXECUTABLE) && vma->vm_file)
|
|
|
|
break;
|
|
|
|
vma = vma->vm_next;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (vma) {
|
2006-12-08 03:36:36 -07:00
|
|
|
*mnt = mntget(vma->vm_file->f_path.mnt);
|
|
|
|
*dentry = dget(vma->vm_file->f_path.dentry);
|
2005-04-16 16:20:36 -06:00
|
|
|
result = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
mmput(mm);
|
|
|
|
out:
|
|
|
|
return result;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void pad_len_spaces(struct seq_file *m, int len)
|
|
|
|
{
|
|
|
|
len = 25 + sizeof(void*) * 6 - len;
|
|
|
|
if (len < 1)
|
|
|
|
len = 1;
|
|
|
|
seq_printf(m, "%*c", len, ' ');
|
|
|
|
}
|
|
|
|
|
2005-09-03 16:55:10 -06:00
|
|
|
struct mem_size_stats
|
|
|
|
{
|
|
|
|
unsigned long resident;
|
|
|
|
unsigned long shared_clean;
|
|
|
|
unsigned long shared_dirty;
|
|
|
|
unsigned long private_clean;
|
|
|
|
unsigned long private_dirty;
|
|
|
|
};
|
|
|
|
|
|
|
|
static int show_map_internal(struct seq_file *m, void *v, struct mem_size_stats *mss)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2006-06-26 01:25:55 -06:00
|
|
|
struct proc_maps_private *priv = m->private;
|
|
|
|
struct task_struct *task = priv->task;
|
2005-09-03 16:55:10 -06:00
|
|
|
struct vm_area_struct *vma = v;
|
|
|
|
struct mm_struct *mm = vma->vm_mm;
|
|
|
|
struct file *file = vma->vm_file;
|
|
|
|
int flags = vma->vm_flags;
|
2005-04-16 16:20:36 -06:00
|
|
|
unsigned long ino = 0;
|
|
|
|
dev_t dev = 0;
|
|
|
|
int len;
|
|
|
|
|
|
|
|
if (file) {
|
2006-12-08 03:36:36 -07:00
|
|
|
struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
|
2005-04-16 16:20:36 -06:00
|
|
|
dev = inode->i_sb->s_dev;
|
|
|
|
ino = inode->i_ino;
|
|
|
|
}
|
|
|
|
|
|
|
|
seq_printf(m, "%08lx-%08lx %c%c%c%c %08lx %02x:%02x %lu %n",
|
2005-09-03 16:55:10 -06:00
|
|
|
vma->vm_start,
|
|
|
|
vma->vm_end,
|
2005-04-16 16:20:36 -06:00
|
|
|
flags & VM_READ ? 'r' : '-',
|
|
|
|
flags & VM_WRITE ? 'w' : '-',
|
|
|
|
flags & VM_EXEC ? 'x' : '-',
|
|
|
|
flags & VM_MAYSHARE ? 's' : 'p',
|
2005-09-03 16:55:10 -06:00
|
|
|
vma->vm_pgoff << PAGE_SHIFT,
|
2005-04-16 16:20:36 -06:00
|
|
|
MAJOR(dev), MINOR(dev), ino, &len);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Print the dentry name for named mappings, and a
|
|
|
|
* special [heap] marker for the heap:
|
|
|
|
*/
|
2005-09-03 16:55:10 -06:00
|
|
|
if (file) {
|
2005-04-16 16:20:36 -06:00
|
|
|
pad_len_spaces(m, len);
|
2006-12-08 03:36:36 -07:00
|
|
|
seq_path(m, file->f_path.mnt, file->f_path.dentry, "\n");
|
2005-04-16 16:20:36 -06:00
|
|
|
} else {
|
[PATCH] vdso: randomize the i386 vDSO by moving it into a vma
Move the i386 VDSO down into a vma and thus randomize it.
Besides the security implications, this feature also helps debuggers, which
can COW a vma-backed VDSO just like a normal DSO and can thus do
single-stepping and other debugging features.
It's good for hypervisors (Xen, VMWare) too, which typically live in the same
high-mapped address space as the VDSO, hence whenever the VDSO is used, they
get lots of guest pagefaults and have to fix such guest accesses up - which
slows things down instead of speeding things up (the primary purpose of the
VDSO).
There's a new CONFIG_COMPAT_VDSO (default=y) option, which provides support
for older glibcs that still rely on a prelinked high-mapped VDSO. Newer
distributions (using glibc 2.3.3 or later) can turn this option off. Turning
it off is also recommended for security reasons: attackers cannot use the
predictable high-mapped VDSO page as syscall trampoline anymore.
There is a new vdso=[0|1] boot option as well, and a runtime
/proc/sys/vm/vdso_enabled sysctl switch, that allows the VDSO to be turned
on/off.
(This version of the VDSO-randomization patch also has working ELF
coredumping, the previous patch crashed in the coredumping code.)
This code is a combined work of the exec-shield VDSO randomization
code and Gerd Hoffmann's hypervisor-centric VDSO patch. Rusty Russell
started this patch and i completed it.
[akpm@osdl.org: cleanups]
[akpm@osdl.org: compile fix]
[akpm@osdl.org: compile fix 2]
[akpm@osdl.org: compile fix 3]
[akpm@osdl.org: revernt MAXMEM change]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Cc: Gerd Hoffmann <kraxel@suse.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 03:53:50 -06:00
|
|
|
const char *name = arch_vma_name(vma);
|
|
|
|
if (!name) {
|
|
|
|
if (mm) {
|
|
|
|
if (vma->vm_start <= mm->start_brk &&
|
2005-09-03 16:55:10 -06:00
|
|
|
vma->vm_end >= mm->brk) {
|
[PATCH] vdso: randomize the i386 vDSO by moving it into a vma
Move the i386 VDSO down into a vma and thus randomize it.
Besides the security implications, this feature also helps debuggers, which
can COW a vma-backed VDSO just like a normal DSO and can thus do
single-stepping and other debugging features.
It's good for hypervisors (Xen, VMWare) too, which typically live in the same
high-mapped address space as the VDSO, hence whenever the VDSO is used, they
get lots of guest pagefaults and have to fix such guest accesses up - which
slows things down instead of speeding things up (the primary purpose of the
VDSO).
There's a new CONFIG_COMPAT_VDSO (default=y) option, which provides support
for older glibcs that still rely on a prelinked high-mapped VDSO. Newer
distributions (using glibc 2.3.3 or later) can turn this option off. Turning
it off is also recommended for security reasons: attackers cannot use the
predictable high-mapped VDSO page as syscall trampoline anymore.
There is a new vdso=[0|1] boot option as well, and a runtime
/proc/sys/vm/vdso_enabled sysctl switch, that allows the VDSO to be turned
on/off.
(This version of the VDSO-randomization patch also has working ELF
coredumping, the previous patch crashed in the coredumping code.)
This code is a combined work of the exec-shield VDSO randomization
code and Gerd Hoffmann's hypervisor-centric VDSO patch. Rusty Russell
started this patch and i completed it.
[akpm@osdl.org: cleanups]
[akpm@osdl.org: compile fix]
[akpm@osdl.org: compile fix 2]
[akpm@osdl.org: compile fix 3]
[akpm@osdl.org: revernt MAXMEM change]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Cc: Gerd Hoffmann <kraxel@suse.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 03:53:50 -06:00
|
|
|
name = "[heap]";
|
|
|
|
} else if (vma->vm_start <= mm->start_stack &&
|
|
|
|
vma->vm_end >= mm->start_stack) {
|
|
|
|
name = "[stack]";
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
[PATCH] vdso: randomize the i386 vDSO by moving it into a vma
Move the i386 VDSO down into a vma and thus randomize it.
Besides the security implications, this feature also helps debuggers, which
can COW a vma-backed VDSO just like a normal DSO and can thus do
single-stepping and other debugging features.
It's good for hypervisors (Xen, VMWare) too, which typically live in the same
high-mapped address space as the VDSO, hence whenever the VDSO is used, they
get lots of guest pagefaults and have to fix such guest accesses up - which
slows things down instead of speeding things up (the primary purpose of the
VDSO).
There's a new CONFIG_COMPAT_VDSO (default=y) option, which provides support
for older glibcs that still rely on a prelinked high-mapped VDSO. Newer
distributions (using glibc 2.3.3 or later) can turn this option off. Turning
it off is also recommended for security reasons: attackers cannot use the
predictable high-mapped VDSO page as syscall trampoline anymore.
There is a new vdso=[0|1] boot option as well, and a runtime
/proc/sys/vm/vdso_enabled sysctl switch, that allows the VDSO to be turned
on/off.
(This version of the VDSO-randomization patch also has working ELF
coredumping, the previous patch crashed in the coredumping code.)
This code is a combined work of the exec-shield VDSO randomization
code and Gerd Hoffmann's hypervisor-centric VDSO patch. Rusty Russell
started this patch and i completed it.
[akpm@osdl.org: cleanups]
[akpm@osdl.org: compile fix]
[akpm@osdl.org: compile fix 2]
[akpm@osdl.org: compile fix 3]
[akpm@osdl.org: revernt MAXMEM change]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Cc: Gerd Hoffmann <kraxel@suse.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 03:53:50 -06:00
|
|
|
} else {
|
|
|
|
name = "[vdso]";
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
[PATCH] vdso: randomize the i386 vDSO by moving it into a vma
Move the i386 VDSO down into a vma and thus randomize it.
Besides the security implications, this feature also helps debuggers, which
can COW a vma-backed VDSO just like a normal DSO and can thus do
single-stepping and other debugging features.
It's good for hypervisors (Xen, VMWare) too, which typically live in the same
high-mapped address space as the VDSO, hence whenever the VDSO is used, they
get lots of guest pagefaults and have to fix such guest accesses up - which
slows things down instead of speeding things up (the primary purpose of the
VDSO).
There's a new CONFIG_COMPAT_VDSO (default=y) option, which provides support
for older glibcs that still rely on a prelinked high-mapped VDSO. Newer
distributions (using glibc 2.3.3 or later) can turn this option off. Turning
it off is also recommended for security reasons: attackers cannot use the
predictable high-mapped VDSO page as syscall trampoline anymore.
There is a new vdso=[0|1] boot option as well, and a runtime
/proc/sys/vm/vdso_enabled sysctl switch, that allows the VDSO to be turned
on/off.
(This version of the VDSO-randomization patch also has working ELF
coredumping, the previous patch crashed in the coredumping code.)
This code is a combined work of the exec-shield VDSO randomization
code and Gerd Hoffmann's hypervisor-centric VDSO patch. Rusty Russell
started this patch and i completed it.
[akpm@osdl.org: cleanups]
[akpm@osdl.org: compile fix]
[akpm@osdl.org: compile fix 2]
[akpm@osdl.org: compile fix 3]
[akpm@osdl.org: revernt MAXMEM change]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Cc: Gerd Hoffmann <kraxel@suse.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 03:53:50 -06:00
|
|
|
}
|
|
|
|
if (name) {
|
2005-04-16 16:20:36 -06:00
|
|
|
pad_len_spaces(m, len);
|
[PATCH] vdso: randomize the i386 vDSO by moving it into a vma
Move the i386 VDSO down into a vma and thus randomize it.
Besides the security implications, this feature also helps debuggers, which
can COW a vma-backed VDSO just like a normal DSO and can thus do
single-stepping and other debugging features.
It's good for hypervisors (Xen, VMWare) too, which typically live in the same
high-mapped address space as the VDSO, hence whenever the VDSO is used, they
get lots of guest pagefaults and have to fix such guest accesses up - which
slows things down instead of speeding things up (the primary purpose of the
VDSO).
There's a new CONFIG_COMPAT_VDSO (default=y) option, which provides support
for older glibcs that still rely on a prelinked high-mapped VDSO. Newer
distributions (using glibc 2.3.3 or later) can turn this option off. Turning
it off is also recommended for security reasons: attackers cannot use the
predictable high-mapped VDSO page as syscall trampoline anymore.
There is a new vdso=[0|1] boot option as well, and a runtime
/proc/sys/vm/vdso_enabled sysctl switch, that allows the VDSO to be turned
on/off.
(This version of the VDSO-randomization patch also has working ELF
coredumping, the previous patch crashed in the coredumping code.)
This code is a combined work of the exec-shield VDSO randomization
code and Gerd Hoffmann's hypervisor-centric VDSO patch. Rusty Russell
started this patch and i completed it.
[akpm@osdl.org: cleanups]
[akpm@osdl.org: compile fix]
[akpm@osdl.org: compile fix 2]
[akpm@osdl.org: compile fix 3]
[akpm@osdl.org: revernt MAXMEM change]
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Cc: Gerd Hoffmann <kraxel@suse.de>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Andi Kleen <ak@muc.de>
Cc: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 03:53:50 -06:00
|
|
|
seq_puts(m, name);
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
}
|
|
|
|
seq_putc(m, '\n');
|
2005-09-03 16:55:10 -06:00
|
|
|
|
|
|
|
if (mss)
|
|
|
|
seq_printf(m,
|
|
|
|
"Size: %8lu kB\n"
|
|
|
|
"Rss: %8lu kB\n"
|
|
|
|
"Shared_Clean: %8lu kB\n"
|
|
|
|
"Shared_Dirty: %8lu kB\n"
|
|
|
|
"Private_Clean: %8lu kB\n"
|
|
|
|
"Private_Dirty: %8lu kB\n",
|
|
|
|
(vma->vm_end - vma->vm_start) >> 10,
|
|
|
|
mss->resident >> 10,
|
|
|
|
mss->shared_clean >> 10,
|
|
|
|
mss->shared_dirty >> 10,
|
|
|
|
mss->private_clean >> 10,
|
|
|
|
mss->private_dirty >> 10);
|
|
|
|
|
|
|
|
if (m->count < m->size) /* vma is copied successfully */
|
|
|
|
m->version = (vma != get_gate_vma(task))? vma->vm_start: 0;
|
2005-04-16 16:20:36 -06:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-09-03 16:55:10 -06:00
|
|
|
static int show_map(struct seq_file *m, void *v)
|
|
|
|
{
|
2005-11-13 17:07:20 -07:00
|
|
|
return show_map_internal(m, v, NULL);
|
2005-09-03 16:55:10 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
static void smaps_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
|
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
struct mem_size_stats *mss)
|
|
|
|
{
|
|
|
|
pte_t *pte, ptent;
|
2005-10-29 19:16:27 -06:00
|
|
|
spinlock_t *ptl;
|
2005-09-03 16:55:10 -06:00
|
|
|
struct page *page;
|
|
|
|
|
2005-10-29 19:16:27 -06:00
|
|
|
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
|
2005-09-03 16:55:10 -06:00
|
|
|
do {
|
|
|
|
ptent = *pte;
|
2005-10-29 19:16:27 -06:00
|
|
|
if (!pte_present(ptent))
|
2005-09-03 16:55:10 -06:00
|
|
|
continue;
|
|
|
|
|
|
|
|
mss->resident += PAGE_SIZE;
|
2006-03-06 16:42:58 -07:00
|
|
|
|
|
|
|
page = vm_normal_page(vma, addr, ptent);
|
|
|
|
if (!page)
|
2005-09-03 16:55:10 -06:00
|
|
|
continue;
|
|
|
|
|
2006-03-06 16:42:58 -07:00
|
|
|
if (page_mapcount(page) >= 2) {
|
2005-09-03 16:55:10 -06:00
|
|
|
if (pte_dirty(ptent))
|
|
|
|
mss->shared_dirty += PAGE_SIZE;
|
|
|
|
else
|
|
|
|
mss->shared_clean += PAGE_SIZE;
|
|
|
|
} else {
|
|
|
|
if (pte_dirty(ptent))
|
|
|
|
mss->private_dirty += PAGE_SIZE;
|
|
|
|
else
|
|
|
|
mss->private_clean += PAGE_SIZE;
|
|
|
|
}
|
|
|
|
} while (pte++, addr += PAGE_SIZE, addr != end);
|
2005-10-29 19:16:27 -06:00
|
|
|
pte_unmap_unlock(pte - 1, ptl);
|
|
|
|
cond_resched();
|
2005-09-03 16:55:10 -06:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline void smaps_pmd_range(struct vm_area_struct *vma, pud_t *pud,
|
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
struct mem_size_stats *mss)
|
|
|
|
{
|
|
|
|
pmd_t *pmd;
|
|
|
|
unsigned long next;
|
|
|
|
|
|
|
|
pmd = pmd_offset(pud, addr);
|
|
|
|
do {
|
|
|
|
next = pmd_addr_end(addr, end);
|
|
|
|
if (pmd_none_or_clear_bad(pmd))
|
|
|
|
continue;
|
|
|
|
smaps_pte_range(vma, pmd, addr, next, mss);
|
|
|
|
} while (pmd++, addr = next, addr != end);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void smaps_pud_range(struct vm_area_struct *vma, pgd_t *pgd,
|
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
struct mem_size_stats *mss)
|
|
|
|
{
|
|
|
|
pud_t *pud;
|
|
|
|
unsigned long next;
|
|
|
|
|
|
|
|
pud = pud_offset(pgd, addr);
|
|
|
|
do {
|
|
|
|
next = pud_addr_end(addr, end);
|
|
|
|
if (pud_none_or_clear_bad(pud))
|
|
|
|
continue;
|
|
|
|
smaps_pmd_range(vma, pud, addr, next, mss);
|
|
|
|
} while (pud++, addr = next, addr != end);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void smaps_pgd_range(struct vm_area_struct *vma,
|
|
|
|
unsigned long addr, unsigned long end,
|
|
|
|
struct mem_size_stats *mss)
|
|
|
|
{
|
|
|
|
pgd_t *pgd;
|
|
|
|
unsigned long next;
|
|
|
|
|
|
|
|
pgd = pgd_offset(vma->vm_mm, addr);
|
|
|
|
do {
|
|
|
|
next = pgd_addr_end(addr, end);
|
|
|
|
if (pgd_none_or_clear_bad(pgd))
|
|
|
|
continue;
|
|
|
|
smaps_pud_range(vma, pgd, addr, next, mss);
|
|
|
|
} while (pgd++, addr = next, addr != end);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int show_smap(struct seq_file *m, void *v)
|
|
|
|
{
|
|
|
|
struct vm_area_struct *vma = v;
|
|
|
|
struct mem_size_stats mss;
|
|
|
|
|
|
|
|
memset(&mss, 0, sizeof mss);
|
2006-03-06 16:42:57 -07:00
|
|
|
if (vma->vm_mm && !is_vm_hugetlb_page(vma))
|
2005-09-03 16:55:10 -06:00
|
|
|
smaps_pgd_range(vma, vma->vm_start, vma->vm_end, &mss);
|
|
|
|
return show_map_internal(m, v, &mss);
|
|
|
|
}
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
static void *m_start(struct seq_file *m, loff_t *pos)
|
|
|
|
{
|
2006-06-26 01:25:55 -06:00
|
|
|
struct proc_maps_private *priv = m->private;
|
2005-04-16 16:20:36 -06:00
|
|
|
unsigned long last_addr = m->version;
|
|
|
|
struct mm_struct *mm;
|
2006-06-26 01:25:55 -06:00
|
|
|
struct vm_area_struct *vma, *tail_vma = NULL;
|
2005-04-16 16:20:36 -06:00
|
|
|
loff_t l = *pos;
|
|
|
|
|
2006-06-26 01:25:55 -06:00
|
|
|
/* Clear the per syscall fields in priv */
|
|
|
|
priv->task = NULL;
|
|
|
|
priv->tail_vma = NULL;
|
|
|
|
|
2005-04-16 16:20:36 -06:00
|
|
|
/*
|
|
|
|
* We remember last_addr rather than next_addr to hit with
|
|
|
|
* mmap_cache most of the time. We have zero last_addr at
|
2005-09-03 16:55:10 -06:00
|
|
|
* the beginning and also after lseek. We will have -1 last_addr
|
|
|
|
* after the end of the vmas.
|
2005-04-16 16:20:36 -06:00
|
|
|
*/
|
|
|
|
|
|
|
|
if (last_addr == -1UL)
|
|
|
|
return NULL;
|
|
|
|
|
2006-06-26 01:25:56 -06:00
|
|
|
priv->task = get_pid_task(priv->pid, PIDTYPE_PID);
|
2006-06-26 01:25:55 -06:00
|
|
|
if (!priv->task)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
mm = get_task_mm(priv->task);
|
2005-04-16 16:20:36 -06:00
|
|
|
if (!mm)
|
|
|
|
return NULL;
|
|
|
|
|
2006-06-26 01:25:55 -06:00
|
|
|
priv->tail_vma = tail_vma = get_gate_vma(priv->task);
|
2005-04-16 16:20:36 -06:00
|
|
|
down_read(&mm->mmap_sem);
|
|
|
|
|
|
|
|
/* Start with last addr hint */
|
2005-09-03 16:55:10 -06:00
|
|
|
if (last_addr && (vma = find_vma(mm, last_addr))) {
|
|
|
|
vma = vma->vm_next;
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2005-09-03 16:55:10 -06:00
|
|
|
* Check the vma index is within the range and do
|
2005-04-16 16:20:36 -06:00
|
|
|
* sequential scan until m_index.
|
|
|
|
*/
|
2005-09-03 16:55:10 -06:00
|
|
|
vma = NULL;
|
2005-04-16 16:20:36 -06:00
|
|
|
if ((unsigned long)l < mm->map_count) {
|
2005-09-03 16:55:10 -06:00
|
|
|
vma = mm->mmap;
|
|
|
|
while (l-- && vma)
|
|
|
|
vma = vma->vm_next;
|
2005-04-16 16:20:36 -06:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (l != mm->map_count)
|
2005-09-03 16:55:10 -06:00
|
|
|
tail_vma = NULL; /* After gate vma */
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
out:
|
2005-09-03 16:55:10 -06:00
|
|
|
if (vma)
|
|
|
|
return vma;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
2005-09-03 16:55:10 -06:00
|
|
|
/* End of vmas has been reached */
|
|
|
|
m->version = (tail_vma != NULL)? 0: -1UL;
|
2005-04-16 16:20:36 -06:00
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
mmput(mm);
|
2005-09-03 16:55:10 -06:00
|
|
|
return tail_vma;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2006-06-26 01:25:55 -06:00
|
|
|
static void vma_stop(struct proc_maps_private *priv, struct vm_area_struct *vma)
|
2005-04-16 16:20:36 -06:00
|
|
|
{
|
2006-06-26 01:25:55 -06:00
|
|
|
if (vma && vma != priv->tail_vma) {
|
2005-09-03 16:55:10 -06:00
|
|
|
struct mm_struct *mm = vma->vm_mm;
|
2005-04-16 16:20:36 -06:00
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
mmput(mm);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void *m_next(struct seq_file *m, void *v, loff_t *pos)
|
|
|
|
{
|
2006-06-26 01:25:55 -06:00
|
|
|
struct proc_maps_private *priv = m->private;
|
2005-09-03 16:55:10 -06:00
|
|
|
struct vm_area_struct *vma = v;
|
2006-06-26 01:25:55 -06:00
|
|
|
struct vm_area_struct *tail_vma = priv->tail_vma;
|
2005-04-16 16:20:36 -06:00
|
|
|
|
|
|
|
(*pos)++;
|
2005-09-03 16:55:10 -06:00
|
|
|
if (vma && (vma != tail_vma) && vma->vm_next)
|
|
|
|
return vma->vm_next;
|
2006-06-26 01:25:55 -06:00
|
|
|
vma_stop(priv, vma);
|
2005-09-03 16:55:10 -06:00
|
|
|
return (vma != tail_vma)? tail_vma: NULL;
|
2005-04-16 16:20:36 -06:00
|
|
|
}
|
|
|
|
|
2006-06-26 01:25:55 -06:00
|
|
|
static void m_stop(struct seq_file *m, void *v)
|
|
|
|
{
|
|
|
|
struct proc_maps_private *priv = m->private;
|
|
|
|
struct vm_area_struct *vma = v;
|
|
|
|
|
|
|
|
vma_stop(priv, vma);
|
|
|
|
if (priv->task)
|
|
|
|
put_task_struct(priv->task);
|
|
|
|
}
|
|
|
|
|
2006-06-26 01:25:48 -06:00
|
|
|
static struct seq_operations proc_pid_maps_op = {
|
2005-04-16 16:20:36 -06:00
|
|
|
.start = m_start,
|
|
|
|
.next = m_next,
|
|
|
|
.stop = m_stop,
|
|
|
|
.show = show_map
|
|
|
|
};
|
2005-09-03 16:54:45 -06:00
|
|
|
|
2006-06-26 01:25:48 -06:00
|
|
|
static struct seq_operations proc_pid_smaps_op = {
|
2005-09-03 16:55:10 -06:00
|
|
|
.start = m_start,
|
|
|
|
.next = m_next,
|
|
|
|
.stop = m_stop,
|
|
|
|
.show = show_smap
|
|
|
|
};
|
|
|
|
|
2006-06-26 01:25:48 -06:00
|
|
|
static int do_maps_open(struct inode *inode, struct file *file,
|
|
|
|
struct seq_operations *ops)
|
|
|
|
{
|
2006-06-26 01:25:55 -06:00
|
|
|
struct proc_maps_private *priv;
|
|
|
|
int ret = -ENOMEM;
|
|
|
|
priv = kzalloc(sizeof(*priv), GFP_KERNEL);
|
|
|
|
if (priv) {
|
2006-06-26 01:25:56 -06:00
|
|
|
priv->pid = proc_pid(inode);
|
2006-06-26 01:25:55 -06:00
|
|
|
ret = seq_open(file, ops);
|
|
|
|
if (!ret) {
|
|
|
|
struct seq_file *m = file->private_data;
|
|
|
|
m->private = priv;
|
|
|
|
} else {
|
|
|
|
kfree(priv);
|
|
|
|
}
|
2006-06-26 01:25:48 -06:00
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int maps_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return do_maps_open(inode, file, &proc_pid_maps_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct file_operations proc_maps_operations = {
|
|
|
|
.open = maps_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
2006-06-26 01:25:55 -06:00
|
|
|
.release = seq_release_private,
|
2006-06-26 01:25:48 -06:00
|
|
|
};
|
|
|
|
|
2005-09-03 16:54:45 -06:00
|
|
|
#ifdef CONFIG_NUMA
|
2006-01-08 02:01:02 -07:00
|
|
|
extern int show_numa_map(struct seq_file *m, void *v);
|
2005-09-03 16:54:45 -06:00
|
|
|
|
2006-06-26 01:25:48 -06:00
|
|
|
static struct seq_operations proc_pid_numa_maps_op = {
|
2006-01-08 02:01:02 -07:00
|
|
|
.start = m_start,
|
|
|
|
.next = m_next,
|
|
|
|
.stop = m_stop,
|
|
|
|
.show = show_numa_map
|
2005-09-03 16:54:45 -06:00
|
|
|
};
|
2006-06-26 01:25:48 -06:00
|
|
|
|
|
|
|
static int numa_maps_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return do_maps_open(inode, file, &proc_pid_numa_maps_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct file_operations proc_numa_maps_operations = {
|
|
|
|
.open = numa_maps_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
2006-06-26 01:25:55 -06:00
|
|
|
.release = seq_release_private,
|
2006-06-26 01:25:48 -06:00
|
|
|
};
|
2005-09-03 16:54:45 -06:00
|
|
|
#endif
|
2006-06-26 01:25:48 -06:00
|
|
|
|
|
|
|
static int smaps_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return do_maps_open(inode, file, &proc_pid_smaps_op);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct file_operations proc_smaps_operations = {
|
|
|
|
.open = smaps_open,
|
|
|
|
.read = seq_read,
|
|
|
|
.llseek = seq_lseek,
|
2006-06-26 01:25:55 -06:00
|
|
|
.release = seq_release_private,
|
2006-06-26 01:25:48 -06:00
|
|
|
};
|