[PATCH] perf: rework memory accounting in perf_mmap()

[PATCH] perf: rework memory accounting in perf_mmap()

[Song Liu]

perf_mmap() always increases user->locked_vm. As a result, "extra" could
grow bigger than "user_extra", which doesn't make sense. Here is an
example case:

Note: Assume "user_lock_limit" is very small.
| # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
| 0                    | 0                   | 0             |
| 1                    | user_extra          | user_extra    |
| 2                    | 3 * user_extra      | 2 * user_extra|
| 3                    | 6 * user_extra      | 3 * user_extra|
| 4                    | 10 * user_extra     | 4 * user_extra|

Fix this by maintaining proper user_extra and extra.

Reported-by: Hechao Li <hechaol@fb.com>
Cc: Jie Meng <jmeng@fb.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Song Liu <songliubraving@fb.com>
---
 kernel/events/core.c | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 0463c1151bae..a0120bdbce17 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -5585,7 +5585,8 @@ static void perf_mmap_close(struct vm_area_struct *vma)
 	 * undo the VM accounting.
 	 */
 
-	atomic_long_sub((size >> PAGE_SHIFT) + 1, &mmap_user->locked_vm);
+	atomic_long_sub((size >> PAGE_SHIFT) + 1 - mmap_locked,
+			&mmap_user->locked_vm);
 	atomic64_sub(mmap_locked, &vma->vm_mm->pinned_vm);
 	free_uid(mmap_user);
 
@@ -5729,8 +5730,20 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma)
 
 	user_locked = atomic_long_read(&user->locked_vm) + user_extra;
 
-	if (user_locked > user_lock_limit)
+	if (user_locked <= user_lock_limit) {
+		/* charge all to locked_vm */
+	} else if (atomic_long_read(&user->locked_vm) >= user_lock_limit) {
+		/* charge all to pinned_vm */
+		extra = user_extra;
+		user_extra = 0;
+	} else {
+		/*
+		 * charge locked_vm until it hits user_lock_limit;
+		 * charge the rest from pinned_vm
+		 */
 		extra = user_locked - user_lock_limit;
+		user_extra -= extra;
+	}
 
 	lock_limit = rlimit(RLIMIT_MEMLOCK);
 	lock_limit >>= PAGE_SHIFT;
-- 
2.17.1

Re: [PATCH] perf: rework memory accounting in perf_mmap()

[Song Liu]

Hi Peter,

> On Sep 4, 2019, at 2:46 PM, Song Liu <songliubraving@fb.com> wrote:
> 
> perf_mmap() always increases user->locked_vm. As a result, "extra" could
> grow bigger than "user_extra", which doesn't make sense. Here is an
> example case:
> 
> Note: Assume "user_lock_limit" is very small.
> | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
> | 0                    | 0                   | 0             |
> | 1                    | user_extra          | user_extra    |
> | 2                    | 3 * user_extra      | 2 * user_extra|
> | 3                    | 6 * user_extra      | 3 * user_extra|
> | 4                    | 10 * user_extra     | 4 * user_extra|
> 
> Fix this by maintaining proper user_extra and extra.
> 
> Reported-by: Hechao Li <hechaol@fb.com>
> Cc: Jie Meng <jmeng@fb.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Song Liu <songliubraving@fb.com>

Could you please share your feedbacks/comments on this one?

Thanks,
Song

Re: [PATCH] perf: rework memory accounting in perf_mmap()

[Hechao Li]

Song Liu <songliubraving@fb.com> wrote on Mon [2019-Sep-16 12:43:16 -0700]:
> Hi Peter,
> 
> > On Sep 4, 2019, at 2:46 PM, Song Liu <songliubraving@fb.com> wrote:
> > 
> > perf_mmap() always increases user->locked_vm. As a result, "extra" could
> > grow bigger than "user_extra", which doesn't make sense. Here is an
> > example case:
> > 
> > Note: Assume "user_lock_limit" is very small.
> > | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
> > | 0                    | 0                   | 0             |
> > | 1                    | user_extra          | user_extra    |
> > | 2                    | 3 * user_extra      | 2 * user_extra|
> > | 3                    | 6 * user_extra      | 3 * user_extra|
> > | 4                    | 10 * user_extra     | 4 * user_extra|
> > 
> > Fix this by maintaining proper user_extra and extra.
> > 
> > Reported-by: Hechao Li <hechaol@fb.com>
> > Cc: Jie Meng <jmeng@fb.com>
> > Cc: Peter Zijlstra <peterz@infradead.org>
> > Signed-off-by: Song Liu <songliubraving@fb.com>
> 
> Could you please share your feedbacks/comments on this one?
> 
> Thanks,
> Song

The change looks good to me. Thanks, Song.

Reviewed-By: Hechao Li <hechaol@fb.com>

Hechao

Re: [PATCH] perf: rework memory accounting in perf_mmap()

[Song Liu]

> On Sep 16, 2019, at 1:10 PM, Hechao Li <hechaol@fb.com> wrote:
> 
> Song Liu <songliubraving@fb.com> wrote on Mon [2019-Sep-16 12:43:16 -0700]:
>> Hi Peter,
>> 
>>> On Sep 4, 2019, at 2:46 PM, Song Liu <songliubraving@fb.com> wrote:
>>> 
>>> perf_mmap() always increases user->locked_vm. As a result, "extra" could
>>> grow bigger than "user_extra", which doesn't make sense. Here is an
>>> example case:
>>> 
>>> Note: Assume "user_lock_limit" is very small.
>>> | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
>>> | 0                    | 0                   | 0             |
>>> | 1                    | user_extra          | user_extra    |
>>> | 2                    | 3 * user_extra      | 2 * user_extra|
>>> | 3                    | 6 * user_extra      | 3 * user_extra|
>>> | 4                    | 10 * user_extra     | 4 * user_extra|
>>> 
>>> Fix this by maintaining proper user_extra and extra.
>>> 
>>> Reported-by: Hechao Li <hechaol@fb.com>
>>> Cc: Jie Meng <jmeng@fb.com>
>>> Cc: Peter Zijlstra <peterz@infradead.org>
>>> Signed-off-by: Song Liu <songliubraving@fb.com>
>> 
>> Could you please share your feedbacks/comments on this one?
>> 
>> Thanks,
>> Song
> 
> The change looks good to me. Thanks, Song.
> 
> Reviewed-By: Hechao Li <hechaol@fb.com>

Thanks Hechao!

Hi Peter, 

Does this fix look good to you?

Thanks,
Song

Re: [PATCH] perf: rework memory accounting in perf_mmap()

[Peter Zijlstra]

On Wed, Sep 04, 2019 at 02:46:18PM -0700, Song Liu wrote:
> perf_mmap() always increases user->locked_vm. As a result, "extra" could
> grow bigger than "user_extra", which doesn't make sense. Here is an
> example case:
> 
> Note: Assume "user_lock_limit" is very small.
> | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
> | 0                    | 0                   | 0             |
> | 1                    | user_extra          | user_extra    |
> | 2                    | 3 * user_extra      | 2 * user_extra|
> | 3                    | 6 * user_extra      | 3 * user_extra|
> | 4                    | 10 * user_extra     | 4 * user_extra|
> 
> Fix this by maintaining proper user_extra and extra.

Aah, indeed.

Also, this code is unreadable (which is mostly my own fault I suppose)
:/

Re: [PATCH] perf: rework memory accounting in perf_mmap()

[Song Liu]

Hi Peter,

> On Sep 30, 2019, at 2:02 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> 
> On Wed, Sep 04, 2019 at 02:46:18PM -0700, Song Liu wrote:
>> perf_mmap() always increases user->locked_vm. As a result, "extra" could
>> grow bigger than "user_extra", which doesn't make sense. Here is an
>> example case:
>> 
>> Note: Assume "user_lock_limit" is very small.
>> | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
>> | 0                    | 0                   | 0             |
>> | 1                    | user_extra          | user_extra    |
>> | 2                    | 3 * user_extra      | 2 * user_extra|
>> | 3                    | 6 * user_extra      | 3 * user_extra|
>> | 4                    | 10 * user_extra     | 4 * user_extra|
>> 
>> Fix this by maintaining proper user_extra and extra.
> 
> Aah, indeed.

Thanks for the feedback!

> 
> Also, this code is unreadable (which is mostly my own fault I suppose)
> :/

How does this patch look to you? Is it ready to merge?

Thanks,
Song

Re: [PATCH] perf: rework memory accounting in perf_mmap()

[Peter Zijlstra]

On Mon, Oct 07, 2019 at 04:31:37PM +0000, Song Liu wrote:
> Hi Peter,
> 
> > On Sep 30, 2019, at 2:02 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> > 
> > On Wed, Sep 04, 2019 at 02:46:18PM -0700, Song Liu wrote:
> >> perf_mmap() always increases user->locked_vm. As a result, "extra" could
> >> grow bigger than "user_extra", which doesn't make sense. Here is an
> >> example case:
> >> 
> >> Note: Assume "user_lock_limit" is very small.
> >> | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
> >> | 0                    | 0                   | 0             |
> >> | 1                    | user_extra          | user_extra    |
> >> | 2                    | 3 * user_extra      | 2 * user_extra|
> >> | 3                    | 6 * user_extra      | 3 * user_extra|
> >> | 4                    | 10 * user_extra     | 4 * user_extra|
> >> 
> >> Fix this by maintaining proper user_extra and extra.
> > 
> > Aah, indeed.
> 
> Thanks for the feedback!
> 
> > 
> > Also, this code is unreadable (which is mostly my own fault I suppose)
> > :/
> 
> How does this patch look to you? Is it ready to merge?

Yes, I queued it, I'll get it into tip soon. It got held up due to me
working on some other patches, sorry about that.

[tip: perf/urgent] perf/core: Rework memory accounting in perf_mmap()

[tip-bot2 for Song Liu]

The following commit has been merged into the perf/urgent branch of tip:

Commit-ID:     d44248a41337731a111374822d7d4451b64e73e4
Gitweb:        https://git.kernel.org/tip/d44248a41337731a111374822d7d4451b64e73e4
Author:        Song Liu <songliubraving@fb.com>
AuthorDate:    Wed, 04 Sep 2019 14:46:18 -07:00
Committer:     Ingo Molnar <mingo@kernel.org>
CommitterDate: Wed, 09 Oct 2019 12:44:12 +02:00

perf/core: Rework memory accounting in perf_mmap()

perf_mmap() always increases user->locked_vm. As a result, "extra" could
grow bigger than "user_extra", which doesn't make sense. Here is an
example case:

(Note: Assume "user_lock_limit" is very small.)

  | # of perf_mmap calls |vma->vm_mm->pinned_vm|user->locked_vm|
  | 0                    | 0                   | 0             |
  | 1                    | user_extra          | user_extra    |
  | 2                    | 3 * user_extra      | 2 * user_extra|
  | 3                    | 6 * user_extra      | 3 * user_extra|
  | 4                    | 10 * user_extra     | 4 * user_extra|

Fix this by maintaining proper user_extra and extra.

Reviewed-By: Hechao Li <hechaol@fb.com>
Reported-by: Hechao Li <hechaol@fb.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: <kernel-team@fb.com>
Cc: Jie Meng <jmeng@fb.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/20190904214618.3795672-1-songliubraving@fb.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/events/core.c | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index f953dd1..2b8265a 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -5668,7 +5668,8 @@ again:
 	 * undo the VM accounting.
 	 */
 
-	atomic_long_sub((size >> PAGE_SHIFT) + 1, &mmap_user->locked_vm);
+	atomic_long_sub((size >> PAGE_SHIFT) + 1 - mmap_locked,
+			&mmap_user->locked_vm);
 	atomic64_sub(mmap_locked, &vma->vm_mm->pinned_vm);
 	free_uid(mmap_user);
 
@@ -5812,8 +5813,20 @@ accounting:
 
 	user_locked = atomic_long_read(&user->locked_vm) + user_extra;
 
-	if (user_locked > user_lock_limit)
+	if (user_locked <= user_lock_limit) {
+		/* charge all to locked_vm */
+	} else if (atomic_long_read(&user->locked_vm) >= user_lock_limit) {
+		/* charge all to pinned_vm */
+		extra = user_extra;
+		user_extra = 0;
+	} else {
+		/*
+		 * charge locked_vm until it hits user_lock_limit;
+		 * charge the rest from pinned_vm
+		 */
 		extra = user_locked - user_lock_limit;
+		user_extra -= extra;
+	}
 
 	lock_limit = rlimit(RLIMIT_MEMLOCK);
 	lock_limit >>= PAGE_SHIFT;