2005-04-16 15:20:36 -07:00
|
|
|
/**
|
|
|
|
* @file buffer_sync.c
|
|
|
|
*
|
2008-12-25 09:26:07 -07:00
|
|
|
* @remark Copyright 2002-2009 OProfile authors
|
2005-04-16 15:20:36 -07:00
|
|
|
* @remark Read the file COPYING
|
|
|
|
*
|
|
|
|
* @author John Levon <levon@movementarian.org>
|
2008-07-22 12:08:54 -07:00
|
|
|
* @author Barry Kasindorf
|
2008-12-25 09:26:07 -07:00
|
|
|
* @author Robert Richter <robert.richter@amd.com>
|
2005-04-16 15:20:36 -07:00
|
|
|
*
|
|
|
|
* This is the core of the buffer management. Each
|
|
|
|
* CPU buffer is processed and entered into the
|
|
|
|
* global event buffer. Such processing is necessary
|
|
|
|
* in several circumstances, mentioned below.
|
|
|
|
*
|
|
|
|
* The processing does the job of converting the
|
|
|
|
* transitory EIP value into a persistent dentry/offset
|
|
|
|
* value that the profiler can record at its leisure.
|
|
|
|
*
|
|
|
|
* See fs/dcookies.c for a description of the dentry/offset
|
|
|
|
* objects.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/workqueue.h>
|
|
|
|
#include <linux/notifier.h>
|
|
|
|
#include <linux/dcookies.h>
|
|
|
|
#include <linux/profile.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/fs.h>
|
2007-07-20 12:39:53 -07:00
|
|
|
#include <linux/oprofile.h>
|
Detach sched.h from mm.h
First thing mm.h does is including sched.h solely for can_do_mlock() inline
function which has "current" dereference inside. By dealing with can_do_mlock()
mm.h can be detached from sched.h which is good. See below, why.
This patch
a) removes unconditional inclusion of sched.h from mm.h
b) makes can_do_mlock() normal function in mm/mlock.c
c) exports can_do_mlock() to not break compilation
d) adds sched.h inclusions back to files that were getting it indirectly.
e) adds less bloated headers to some files (asm/signal.h, jiffies.h) that were
getting them indirectly
Net result is:
a) mm.h users would get less code to open, read, preprocess, parse, ... if
they don't need sched.h
b) sched.h stops being dependency for significant number of files:
on x86_64 allmodconfig touching sched.h results in recompile of 4083 files,
after patch it's only 3744 (-8.3%).
Cross-compile tested on
all arm defconfigs, all mips defconfigs, all powerpc defconfigs,
alpha alpha-up
arm
i386 i386-up i386-defconfig i386-allnoconfig
ia64 ia64-up
m68k
mips
parisc parisc-up
powerpc powerpc-up
s390 s390-up
sparc sparc-up
sparc64 sparc64-up
um-x86_64
x86_64 x86_64-up x86_64-defconfig x86_64-allnoconfig
as well as my two usual configs.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-05-20 14:22:52 -07:00
|
|
|
#include <linux/sched.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 01:04:11 -07:00
|
|
|
#include <linux/gfp.h>
|
2007-07-20 12:39:53 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
#include "oprofile_stats.h"
|
|
|
|
#include "event_buffer.h"
|
|
|
|
#include "cpu_buffer.h"
|
|
|
|
#include "buffer_sync.h"
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
static LIST_HEAD(dying_tasks);
|
|
|
|
static LIST_HEAD(dead_tasks);
|
2009-01-10 22:58:09 -07:00
|
|
|
static cpumask_var_t marked_cpus;
|
2005-04-16 15:20:36 -07:00
|
|
|
static DEFINE_SPINLOCK(task_mortuary);
|
|
|
|
static void process_task_mortuary(void);
|
|
|
|
|
|
|
|
/* Take ownership of the task struct and place it on the
|
|
|
|
* list for processing. Only after two full buffer syncs
|
|
|
|
* does the task eventually get freed, because by then
|
|
|
|
* we are sure we will not reference it again.
|
2006-01-08 02:01:35 -07:00
|
|
|
* Can be invoked from softirq via RCU callback due to
|
|
|
|
* call_rcu() of the task struct, hence the _irqsave.
|
2005-04-16 15:20:36 -07:00
|
|
|
*/
|
2008-07-22 12:08:51 -07:00
|
|
|
static int
|
|
|
|
task_free_notify(struct notifier_block *self, unsigned long val, void *data)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
2006-01-08 02:01:35 -07:00
|
|
|
unsigned long flags;
|
2008-07-22 12:08:51 -07:00
|
|
|
struct task_struct *task = data;
|
2006-01-08 02:01:35 -07:00
|
|
|
spin_lock_irqsave(&task_mortuary, flags);
|
2005-04-16 15:20:36 -07:00
|
|
|
list_add(&task->tasks, &dying_tasks);
|
2006-01-08 02:01:35 -07:00
|
|
|
spin_unlock_irqrestore(&task_mortuary, flags);
|
2005-04-16 15:20:36 -07:00
|
|
|
return NOTIFY_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* The task is on its way out. A sync of the buffer means we can catch
|
|
|
|
* any remaining samples for this task.
|
|
|
|
*/
|
2008-07-22 12:08:51 -07:00
|
|
|
static int
|
|
|
|
task_exit_notify(struct notifier_block *self, unsigned long val, void *data)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
/* To avoid latency problems, we only process the current CPU,
|
|
|
|
* hoping that most samples for the task are on this CPU
|
|
|
|
*/
|
2005-06-21 17:14:34 -07:00
|
|
|
sync_buffer(raw_smp_processor_id());
|
2008-07-22 12:08:51 -07:00
|
|
|
return 0;
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* The task is about to try a do_munmap(). We peek at what it's going to
|
|
|
|
* do, and if it's an executable region, process the samples first, so
|
|
|
|
* we don't lose any. This does not have to be exact, it's a QoI issue
|
|
|
|
* only.
|
|
|
|
*/
|
2008-07-22 12:08:51 -07:00
|
|
|
static int
|
|
|
|
munmap_notify(struct notifier_block *self, unsigned long val, void *data)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
unsigned long addr = (unsigned long)data;
|
2008-07-22 12:08:51 -07:00
|
|
|
struct mm_struct *mm = current->mm;
|
|
|
|
struct vm_area_struct *mpnt;
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
down_read(&mm->mmap_sem);
|
|
|
|
|
|
|
|
mpnt = find_vma(mm, addr);
|
|
|
|
if (mpnt && mpnt->vm_file && (mpnt->vm_flags & VM_EXEC)) {
|
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
/* To avoid latency problems, we only process the current CPU,
|
|
|
|
* hoping that most samples for the task are on this CPU
|
|
|
|
*/
|
2005-06-21 17:14:34 -07:00
|
|
|
sync_buffer(raw_smp_processor_id());
|
2005-04-16 15:20:36 -07:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
/* We need to be told about new modules so we don't attribute to a previously
|
|
|
|
* loaded module, or drop the samples on the floor.
|
|
|
|
*/
|
2008-07-22 12:08:51 -07:00
|
|
|
static int
|
|
|
|
module_load_notify(struct notifier_block *self, unsigned long val, void *data)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
#ifdef CONFIG_MODULES
|
|
|
|
if (val != MODULE_STATE_COMING)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* FIXME: should we process all CPU buffers ? */
|
2006-06-25 05:47:33 -07:00
|
|
|
mutex_lock(&buffer_mutex);
|
2005-04-16 15:20:36 -07:00
|
|
|
add_event_entry(ESCAPE_CODE);
|
|
|
|
add_event_entry(MODULE_LOADED_CODE);
|
2006-06-25 05:47:33 -07:00
|
|
|
mutex_unlock(&buffer_mutex);
|
2005-04-16 15:20:36 -07:00
|
|
|
#endif
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
static struct notifier_block task_free_nb = {
|
|
|
|
.notifier_call = task_free_notify,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct notifier_block task_exit_nb = {
|
|
|
|
.notifier_call = task_exit_notify,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct notifier_block munmap_nb = {
|
|
|
|
.notifier_call = munmap_notify,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct notifier_block module_load_nb = {
|
|
|
|
.notifier_call = module_load_notify,
|
|
|
|
};
|
|
|
|
|
|
|
|
int sync_start(void)
|
|
|
|
{
|
|
|
|
int err;
|
|
|
|
|
2009-06-14 23:58:26 -07:00
|
|
|
if (!zalloc_cpumask_var(&marked_cpus, GFP_KERNEL))
|
2009-01-22 06:14:14 -07:00
|
|
|
return -ENOMEM;
|
|
|
|
|
2010-08-13 07:29:04 -07:00
|
|
|
mutex_lock(&buffer_mutex);
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
err = task_handoff_register(&task_free_nb);
|
|
|
|
if (err)
|
|
|
|
goto out1;
|
|
|
|
err = profile_event_register(PROFILE_TASK_EXIT, &task_exit_nb);
|
|
|
|
if (err)
|
|
|
|
goto out2;
|
|
|
|
err = profile_event_register(PROFILE_MUNMAP, &munmap_nb);
|
|
|
|
if (err)
|
|
|
|
goto out3;
|
|
|
|
err = register_module_notifier(&module_load_nb);
|
|
|
|
if (err)
|
|
|
|
goto out4;
|
|
|
|
|
2010-08-13 07:29:04 -07:00
|
|
|
start_cpu_work();
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
out:
|
2010-08-13 07:29:04 -07:00
|
|
|
mutex_unlock(&buffer_mutex);
|
2005-04-16 15:20:36 -07:00
|
|
|
return err;
|
|
|
|
out4:
|
|
|
|
profile_event_unregister(PROFILE_MUNMAP, &munmap_nb);
|
|
|
|
out3:
|
|
|
|
profile_event_unregister(PROFILE_TASK_EXIT, &task_exit_nb);
|
|
|
|
out2:
|
|
|
|
task_handoff_unregister(&task_free_nb);
|
|
|
|
out1:
|
2009-01-22 06:14:14 -07:00
|
|
|
free_cpumask_var(marked_cpus);
|
2005-04-16 15:20:36 -07:00
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void sync_stop(void)
|
|
|
|
{
|
2010-08-13 07:29:04 -07:00
|
|
|
/* flush buffers */
|
|
|
|
mutex_lock(&buffer_mutex);
|
|
|
|
end_cpu_work();
|
2005-04-16 15:20:36 -07:00
|
|
|
unregister_module_notifier(&module_load_nb);
|
|
|
|
profile_event_unregister(PROFILE_MUNMAP, &munmap_nb);
|
|
|
|
profile_event_unregister(PROFILE_TASK_EXIT, &task_exit_nb);
|
|
|
|
task_handoff_unregister(&task_free_nb);
|
2010-08-13 07:29:04 -07:00
|
|
|
mutex_unlock(&buffer_mutex);
|
2010-10-15 06:51:08 -07:00
|
|
|
flush_cpu_work();
|
2010-08-13 07:29:04 -07:00
|
|
|
|
|
|
|
/* make sure we don't leak task structs */
|
|
|
|
process_task_mortuary();
|
|
|
|
process_task_mortuary();
|
|
|
|
|
2009-01-22 06:14:14 -07:00
|
|
|
free_cpumask_var(marked_cpus);
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
|
|
|
|
2008-02-14 20:38:36 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
/* Optimisation. We can manage without taking the dcookie sem
|
|
|
|
* because we cannot reach this code without at least one
|
|
|
|
* dcookie user still being registered (namely, the reader
|
|
|
|
* of the event buffer). */
|
2008-02-14 20:38:36 -07:00
|
|
|
static inline unsigned long fast_get_dcookie(struct path *path)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
unsigned long cookie;
|
2008-02-14 20:38:36 -07:00
|
|
|
|
2008-12-01 01:33:43 -07:00
|
|
|
if (path->dentry->d_flags & DCACHE_COOKIE)
|
2008-02-14 20:38:36 -07:00
|
|
|
return (unsigned long)path->dentry;
|
|
|
|
get_dcookie(path, &cookie);
|
2005-04-16 15:20:36 -07:00
|
|
|
return cookie;
|
|
|
|
}
|
|
|
|
|
2008-02-14 20:38:36 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
/* Look up the dcookie for the task's first VM_EXECUTABLE mapping,
|
|
|
|
* which corresponds loosely to "application name". This is
|
|
|
|
* not strictly necessary but allows oprofile to associate
|
|
|
|
* shared-library samples with particular applications
|
|
|
|
*/
|
2008-07-22 12:08:51 -07:00
|
|
|
static unsigned long get_exec_dcookie(struct mm_struct *mm)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
2005-06-23 22:02:47 -07:00
|
|
|
unsigned long cookie = NO_COOKIE;
|
2008-07-22 12:08:51 -07:00
|
|
|
struct vm_area_struct *vma;
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
if (!mm)
|
|
|
|
goto out;
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
for (vma = mm->mmap; vma; vma = vma->vm_next) {
|
|
|
|
if (!vma->vm_file)
|
|
|
|
continue;
|
|
|
|
if (!(vma->vm_flags & VM_EXECUTABLE))
|
|
|
|
continue;
|
2008-02-14 20:38:36 -07:00
|
|
|
cookie = fast_get_dcookie(&vma->vm_file->f_path);
|
2005-04-16 15:20:36 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
return cookie;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* Convert the EIP value of a sample into a persistent dentry/offset
|
|
|
|
* pair that can then be added to the global event buffer. We make
|
|
|
|
* sure to do this lookup before a mm->mmap modification happens so
|
|
|
|
* we don't lose track.
|
|
|
|
*/
|
2008-07-22 12:08:51 -07:00
|
|
|
static unsigned long
|
|
|
|
lookup_dcookie(struct mm_struct *mm, unsigned long addr, off_t *offset)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
2005-06-23 22:02:47 -07:00
|
|
|
unsigned long cookie = NO_COOKIE;
|
2008-07-22 12:08:51 -07:00
|
|
|
struct vm_area_struct *vma;
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
for (vma = find_vma(mm, addr); vma; vma = vma->vm_next) {
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
if (addr < vma->vm_start || addr >= vma->vm_end)
|
|
|
|
continue;
|
|
|
|
|
2005-06-23 22:02:47 -07:00
|
|
|
if (vma->vm_file) {
|
2008-02-14 20:38:36 -07:00
|
|
|
cookie = fast_get_dcookie(&vma->vm_file->f_path);
|
2005-06-23 22:02:47 -07:00
|
|
|
*offset = (vma->vm_pgoff << PAGE_SHIFT) + addr -
|
|
|
|
vma->vm_start;
|
|
|
|
} else {
|
|
|
|
/* must be an anonymous map */
|
|
|
|
*offset = addr;
|
|
|
|
}
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2005-06-23 22:02:47 -07:00
|
|
|
if (!vma)
|
|
|
|
cookie = INVALID_COOKIE;
|
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
return cookie;
|
|
|
|
}
|
|
|
|
|
2005-06-23 22:02:47 -07:00
|
|
|
static unsigned long last_cookie = INVALID_COOKIE;
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
static void add_cpu_switch(int i)
|
|
|
|
{
|
|
|
|
add_event_entry(ESCAPE_CODE);
|
|
|
|
add_event_entry(CPU_SWITCH_CODE);
|
|
|
|
add_event_entry(i);
|
2005-06-23 22:02:47 -07:00
|
|
|
last_cookie = INVALID_COOKIE;
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
static void add_kernel_ctx_switch(unsigned int in_kernel)
|
|
|
|
{
|
|
|
|
add_event_entry(ESCAPE_CODE);
|
|
|
|
if (in_kernel)
|
2008-07-22 12:08:51 -07:00
|
|
|
add_event_entry(KERNEL_ENTER_SWITCH_CODE);
|
2005-04-16 15:20:36 -07:00
|
|
|
else
|
2008-07-22 12:08:51 -07:00
|
|
|
add_event_entry(KERNEL_EXIT_SWITCH_CODE);
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
static void
|
2008-07-22 12:08:51 -07:00
|
|
|
add_user_ctx_switch(struct task_struct const *task, unsigned long cookie)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
add_event_entry(ESCAPE_CODE);
|
2008-07-22 12:08:51 -07:00
|
|
|
add_event_entry(CTX_SWITCH_CODE);
|
2005-04-16 15:20:36 -07:00
|
|
|
add_event_entry(task->pid);
|
|
|
|
add_event_entry(cookie);
|
|
|
|
/* Another code for daemon back-compat */
|
|
|
|
add_event_entry(ESCAPE_CODE);
|
|
|
|
add_event_entry(CTX_TGID_CODE);
|
|
|
|
add_event_entry(task->tgid);
|
|
|
|
}
|
|
|
|
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
static void add_cookie_switch(unsigned long cookie)
|
|
|
|
{
|
|
|
|
add_event_entry(ESCAPE_CODE);
|
|
|
|
add_event_entry(COOKIE_SWITCH_CODE);
|
|
|
|
add_event_entry(cookie);
|
|
|
|
}
|
|
|
|
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
static void add_trace_begin(void)
|
|
|
|
{
|
|
|
|
add_event_entry(ESCAPE_CODE);
|
|
|
|
add_event_entry(TRACE_BEGIN_CODE);
|
|
|
|
}
|
|
|
|
|
2009-01-05 02:35:31 -07:00
|
|
|
static void add_data(struct op_entry *entry, struct mm_struct *mm)
|
2008-07-22 12:08:54 -07:00
|
|
|
{
|
2009-01-05 02:35:31 -07:00
|
|
|
unsigned long code, pc, val;
|
|
|
|
unsigned long cookie;
|
2008-07-22 12:08:54 -07:00
|
|
|
off_t offset;
|
|
|
|
|
2009-01-05 02:35:31 -07:00
|
|
|
if (!op_cpu_buffer_get_data(entry, &code))
|
|
|
|
return;
|
|
|
|
if (!op_cpu_buffer_get_data(entry, &pc))
|
|
|
|
return;
|
|
|
|
if (!op_cpu_buffer_get_size(entry))
|
2008-12-16 03:01:18 -07:00
|
|
|
return;
|
2008-07-22 12:08:54 -07:00
|
|
|
|
|
|
|
if (mm) {
|
2009-01-05 05:14:04 -07:00
|
|
|
cookie = lookup_dcookie(mm, pc, &offset);
|
2008-07-22 12:08:54 -07:00
|
|
|
|
2009-01-05 05:14:04 -07:00
|
|
|
if (cookie == NO_COOKIE)
|
|
|
|
offset = pc;
|
|
|
|
if (cookie == INVALID_COOKIE) {
|
2008-07-22 12:08:54 -07:00
|
|
|
atomic_inc(&oprofile_stats.sample_lost_no_mapping);
|
2009-01-05 05:14:04 -07:00
|
|
|
offset = pc;
|
2008-07-22 12:08:54 -07:00
|
|
|
}
|
2009-01-05 05:14:04 -07:00
|
|
|
if (cookie != last_cookie) {
|
|
|
|
add_cookie_switch(cookie);
|
|
|
|
last_cookie = cookie;
|
2008-07-22 12:08:54 -07:00
|
|
|
}
|
|
|
|
} else
|
2009-01-05 05:14:04 -07:00
|
|
|
offset = pc;
|
2008-07-22 12:08:54 -07:00
|
|
|
|
|
|
|
add_event_entry(ESCAPE_CODE);
|
|
|
|
add_event_entry(code);
|
|
|
|
add_event_entry(offset); /* Offset from Dcookie */
|
|
|
|
|
2009-01-05 02:35:31 -07:00
|
|
|
while (op_cpu_buffer_get_data(entry, &val))
|
|
|
|
add_event_entry(val);
|
2008-07-22 12:08:54 -07:00
|
|
|
}
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2008-12-29 10:44:21 -07:00
|
|
|
static inline void add_sample_entry(unsigned long offset, unsigned long event)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
add_event_entry(offset);
|
|
|
|
add_event_entry(event);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2008-12-18 11:44:20 -07:00
|
|
|
/*
|
|
|
|
* Add a sample to the global event buffer. If possible the
|
|
|
|
* sample is converted into a persistent dentry/offset pair
|
|
|
|
* for later lookup from userspace. Return 0 on failure.
|
|
|
|
*/
|
|
|
|
static int
|
|
|
|
add_sample(struct mm_struct *mm, struct op_sample *s, int in_kernel)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
unsigned long cookie;
|
|
|
|
off_t offset;
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2008-12-18 11:44:20 -07:00
|
|
|
if (in_kernel) {
|
|
|
|
add_sample_entry(s->eip, s->event);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* add userspace sample */
|
|
|
|
|
|
|
|
if (!mm) {
|
|
|
|
atomic_inc(&oprofile_stats.sample_lost_no_mm);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-07-22 12:08:51 -07:00
|
|
|
cookie = lookup_dcookie(mm, s->eip, &offset);
|
|
|
|
|
2005-06-23 22:02:47 -07:00
|
|
|
if (cookie == INVALID_COOKIE) {
|
2005-04-16 15:20:36 -07:00
|
|
|
atomic_inc(&oprofile_stats.sample_lost_no_mapping);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (cookie != last_cookie) {
|
|
|
|
add_cookie_switch(cookie);
|
|
|
|
last_cookie = cookie;
|
|
|
|
}
|
|
|
|
|
|
|
|
add_sample_entry(offset, s->event);
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2008-07-22 12:08:51 -07:00
|
|
|
|
|
|
|
static void release_mm(struct mm_struct *mm)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
|
|
|
if (!mm)
|
|
|
|
return;
|
|
|
|
up_read(&mm->mmap_sem);
|
|
|
|
mmput(mm);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2008-07-22 12:08:51 -07:00
|
|
|
static struct mm_struct *take_tasks_mm(struct task_struct *task)
|
2005-04-16 15:20:36 -07:00
|
|
|
{
|
2008-07-22 12:08:51 -07:00
|
|
|
struct mm_struct *mm = get_task_mm(task);
|
2005-04-16 15:20:36 -07:00
|
|
|
if (mm)
|
|
|
|
down_read(&mm->mmap_sem);
|
|
|
|
return mm;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static inline int is_code(unsigned long val)
|
|
|
|
{
|
|
|
|
return val == ESCAPE_CODE;
|
|
|
|
}
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
/* Move tasks along towards death. Any tasks on dead_tasks
|
|
|
|
* will definitely have no remaining references in any
|
|
|
|
* CPU buffers at this point, because we use two lists,
|
|
|
|
* and to have reached the list, it must have gone through
|
|
|
|
* one full sync already.
|
|
|
|
*/
|
|
|
|
static void process_task_mortuary(void)
|
|
|
|
{
|
2006-01-08 02:01:35 -07:00
|
|
|
unsigned long flags;
|
|
|
|
LIST_HEAD(local_dead_tasks);
|
2008-07-22 12:08:51 -07:00
|
|
|
struct task_struct *task;
|
|
|
|
struct task_struct *ttask;
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-01-08 02:01:35 -07:00
|
|
|
spin_lock_irqsave(&task_mortuary, flags);
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-01-08 02:01:35 -07:00
|
|
|
list_splice_init(&dead_tasks, &local_dead_tasks);
|
|
|
|
list_splice_init(&dying_tasks, &dead_tasks);
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-01-08 02:01:35 -07:00
|
|
|
spin_unlock_irqrestore(&task_mortuary, flags);
|
|
|
|
|
|
|
|
list_for_each_entry_safe(task, ttask, &local_dead_tasks, tasks) {
|
2005-04-16 15:20:36 -07:00
|
|
|
list_del(&task->tasks);
|
2006-01-08 02:01:35 -07:00
|
|
|
free_task(task);
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static void mark_done(int cpu)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
2009-01-10 22:58:09 -07:00
|
|
|
cpumask_set_cpu(cpu, marked_cpus);
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
for_each_online_cpu(i) {
|
2009-01-10 22:58:09 -07:00
|
|
|
if (!cpumask_test_cpu(i, marked_cpus))
|
2005-04-16 15:20:36 -07:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* All CPUs have been processed at least once,
|
|
|
|
* we can process the mortuary once
|
|
|
|
*/
|
|
|
|
process_task_mortuary();
|
|
|
|
|
2009-01-10 22:58:09 -07:00
|
|
|
cpumask_clear(marked_cpus);
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/* FIXME: this is not sufficient if we implement syscall barrier backtrace
|
|
|
|
* traversal, the code switch to sb_sample_start at first kernel enter/exit
|
|
|
|
* switch so we need a fifth state and some special handling in sync_buffer()
|
|
|
|
*/
|
|
|
|
typedef enum {
|
|
|
|
sb_bt_ignore = -2,
|
|
|
|
sb_buffer_start,
|
|
|
|
sb_bt_start,
|
|
|
|
sb_sample_start,
|
|
|
|
} sync_buffer_state;
|
|
|
|
|
|
|
|
/* Sync one of the CPU's buffers into the global event buffer.
|
|
|
|
* Here we need to go through each batch of samples punctuated
|
|
|
|
* by context switch notes, taking the task's mmap_sem and doing
|
|
|
|
* lookup in task->mm->mmap to convert EIP into dcookie/offset
|
|
|
|
* value.
|
|
|
|
*/
|
|
|
|
void sync_buffer(int cpu)
|
|
|
|
{
|
|
|
|
struct mm_struct *mm = NULL;
|
2008-09-26 14:50:31 -07:00
|
|
|
struct mm_struct *oldmm;
|
2009-01-05 19:56:50 -07:00
|
|
|
unsigned long val;
|
2008-07-22 12:08:51 -07:00
|
|
|
struct task_struct *new;
|
2005-04-16 15:20:36 -07:00
|
|
|
unsigned long cookie = 0;
|
|
|
|
int in_kernel = 1;
|
|
|
|
sync_buffer_state state = sb_buffer_start;
|
2008-07-14 15:10:36 -07:00
|
|
|
unsigned int i;
|
2005-04-16 15:20:36 -07:00
|
|
|
unsigned long available;
|
2008-12-25 09:26:07 -07:00
|
|
|
unsigned long flags;
|
2008-12-29 20:10:46 -07:00
|
|
|
struct op_entry entry;
|
|
|
|
struct op_sample *sample;
|
2005-04-16 15:20:36 -07:00
|
|
|
|
2006-06-25 05:47:33 -07:00
|
|
|
mutex_lock(&buffer_mutex);
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2005-04-16 15:20:36 -07:00
|
|
|
add_cpu_switch(cpu);
|
|
|
|
|
2008-12-24 08:53:53 -07:00
|
|
|
op_cpu_buffer_reset(cpu);
|
|
|
|
available = op_cpu_buffer_entries(cpu);
|
2005-04-16 15:20:36 -07:00
|
|
|
|
|
|
|
for (i = 0; i < available; ++i) {
|
2008-12-29 20:10:46 -07:00
|
|
|
sample = op_cpu_buffer_read_entry(&entry, cpu);
|
|
|
|
if (!sample)
|
oprofile: port to the new ring_buffer
This patch replaces the current oprofile cpu buffer implementation
with the ring buffer provided by the tracing framework. The motivation
here is to leave the pain of implementing ring buffers to others. Oh,
no, there are more advantages. Main reason is the support of different
sample sizes that could be stored in the buffer. Use cases for this
are IBS and Cell spu profiling. Using the new ring buffer ensures
valid and complete samples and allows copying the cpu buffer stateless
without knowing its content. Second it will use generic kernel API and
also reduce code size. And hopefully, there are less bugs.
Since the new tracing ring buffer implementation uses spin locks to
protect the buffer during read/write access, it is difficult to use
the buffer in an NMI handler. In this case, writing to the buffer by
the NMI handler (x86) could occur also during critical sections when
reading the buffer. To avoid this, there are 2 buffers for independent
read and write access. Read access is in process context only, write
access only in the NMI handler. If the read buffer runs empty, both
buffers are swapped atomically. There is potentially a small window
during swapping where the buffers are disabled and samples could be
lost.
Using 2 buffers is a little bit overhead, but the solution is clear
and does not require changes in the ring buffer implementation. It can
be changed to a single buffer solution when the ring buffer access is
implemented as non-locking atomic code.
The new buffer requires more size to store the same amount of samples
because each sample includes an u32 header. Also, there is more code
to execute for buffer access. Nonetheless, the buffer implementation
is proven in the ftrace environment and worth to use also in oprofile.
Patches that changes the internal IBS buffer usage will follow.
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Robert Richter <robert.richter@amd.com>
2008-12-08 17:21:32 -07:00
|
|
|
break;
|
2008-07-22 12:08:51 -07:00
|
|
|
|
2008-12-29 20:10:46 -07:00
|
|
|
if (is_code(sample->eip)) {
|
2008-12-25 09:26:07 -07:00
|
|
|
flags = sample->event;
|
|
|
|
if (flags & TRACE_BEGIN) {
|
|
|
|
state = sb_bt_start;
|
|
|
|
add_trace_begin();
|
|
|
|
}
|
|
|
|
if (flags & KERNEL_CTX_SWITCH) {
|
2005-04-16 15:20:36 -07:00
|
|
|
/* kernel/userspace switch */
|
2008-12-25 09:26:07 -07:00
|
|
|
in_kernel = flags & IS_KERNEL;
|
2005-04-16 15:20:36 -07:00
|
|
|
if (state == sb_buffer_start)
|
|
|
|
state = sb_sample_start;
|
2008-12-25 09:26:07 -07:00
|
|
|
add_kernel_ctx_switch(flags & IS_KERNEL);
|
|
|
|
}
|
2009-01-05 19:56:50 -07:00
|
|
|
if (flags & USER_CTX_SWITCH
|
|
|
|
&& op_cpu_buffer_get_data(&entry, &val)) {
|
2005-04-16 15:20:36 -07:00
|
|
|
/* userspace context switch */
|
2009-01-05 19:56:50 -07:00
|
|
|
new = (struct task_struct *)val;
|
2008-09-26 14:50:31 -07:00
|
|
|
oldmm = mm;
|
2005-04-16 15:20:36 -07:00
|
|
|
release_mm(oldmm);
|
|
|
|
mm = take_tasks_mm(new);
|
|
|
|
if (mm != oldmm)
|
|
|
|
cookie = get_exec_dcookie(mm);
|
|
|
|
add_user_ctx_switch(new, cookie);
|
|
|
|
}
|
2009-01-05 02:35:31 -07:00
|
|
|
if (op_cpu_buffer_get_size(&entry))
|
|
|
|
add_data(&entry, mm);
|
2008-12-18 11:44:20 -07:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (state < sb_bt_start)
|
|
|
|
/* ignore sample */
|
|
|
|
continue;
|
|
|
|
|
2008-12-29 20:10:46 -07:00
|
|
|
if (add_sample(mm, sample, in_kernel))
|
2008-12-18 11:44:20 -07:00
|
|
|
continue;
|
|
|
|
|
|
|
|
/* ignore backtraces if failed to add a sample */
|
|
|
|
if (state == sb_bt_start) {
|
|
|
|
state = sb_bt_ignore;
|
|
|
|
atomic_inc(&oprofile_stats.bt_lost_no_mapping);
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
release_mm(mm);
|
|
|
|
|
|
|
|
mark_done(cpu);
|
|
|
|
|
2006-06-25 05:47:33 -07:00
|
|
|
mutex_unlock(&buffer_mutex);
|
2005-04-16 15:20:36 -07:00
|
|
|
}
|
2008-10-14 16:37:01 -07:00
|
|
|
|
|
|
|
/* The function can be used to add a buffer worth of data directly to
|
|
|
|
* the kernel buffer. The buffer is assumed to be a circular buffer.
|
|
|
|
* Take the entries from index start and end at index end, wrapping
|
|
|
|
* at max_entries.
|
|
|
|
*/
|
|
|
|
void oprofile_put_buff(unsigned long *buf, unsigned int start,
|
|
|
|
unsigned int stop, unsigned int max)
|
|
|
|
{
|
|
|
|
int i;
|
|
|
|
|
|
|
|
i = start;
|
|
|
|
|
|
|
|
mutex_lock(&buffer_mutex);
|
|
|
|
while (i != stop) {
|
|
|
|
add_event_entry(buf[i++]);
|
|
|
|
|
|
|
|
if (i >= max)
|
|
|
|
i = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
mutex_unlock(&buffer_mutex);
|
|
|
|
}
|
|
|
|
|