Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1325108 > unrolled thread

[PATCH] mm/workingset: do not forget to unlock page

Started bySergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
First post2016-02-03 11:00 +0100
Last post2016-02-04 01:20 +0100
Articles 6 — 3 participants

Back to article view | Back to linux.kernel

Contents

  [PATCH] mm/workingset: do not forget to unlock page Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com> - 2016-02-03 11:00 +0100
    Re: [PATCH] mm/workingset: do not forget to unlock page Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com> - 2016-02-03 11:50 +0100
      Re: [PATCH] mm/workingset: do not forget to unlock page Johannes Weiner <hannes@cmpxchg.org> - 2016-02-03 17:30 +0100
        Re: [PATCH] mm/workingset: do not forget to unlock page Andrew Morton <akpm@linux-foundation.org> - 2016-02-03 22:20 +0100
          Re: [PATCH] mm/workingset: do not forget to unlock page Johannes Weiner <hannes@cmpxchg.org> - 2016-02-03 23:10 +0100
            Re: [PATCH] mm/workingset: do not forget to unlock page Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com> - 2016-02-04 01:20 +0100

#1325108 — [PATCH] mm/workingset: do not forget to unlock page

FromSergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Date2016-02-03 11:00 +0100
Subject[PATCH] mm/workingset: do not forget to unlock page
Message-ID<qY2o9-hJ-9@gated-at.bofh.it>
Do not leave page locked (and RCU read side locked) when
return from workingset_activation() due to disabled memcg
or page not being a page_memcg().

Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
---
 mm/workingset.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/mm/workingset.c b/mm/workingset.c
index 14522ed..54138a9 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -315,8 +315,10 @@ void workingset_activation(struct page *page)
 	 * XXX: See workingset_refault() - this should return
 	 * root_mem_cgroup even for !CONFIG_MEMCG.
 	 */
-	if (!mem_cgroup_disabled() && !page_memcg(page))
+	if (!mem_cgroup_disabled() && !page_memcg(page)) {
+		unlock_page_memcg(page);
 		return;
+	}
 	lruvec = mem_cgroup_zone_lruvec(page_zone(page), page_memcg(page));
 	atomic_long_inc(&lruvec->inactive_age);
 	unlock_page_memcg(page);
-- 
2.7.0

[toc] | [next] | [standalone]

#1325185

FromSergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Date2016-02-03 11:50 +0100
Message-ID<qY3ay-Nu-15@gated-at.bofh.it>
In reply to#1325108
On (02/03/16 18:58), Sergey Senozhatsky wrote:
> 
> Do not leave page locked (and RCU read side locked) when
> return from workingset_activation() due to disabled memcg
> or page not being a page_memcg().

d'oh... sorry, the commit message is simply insane.


apparently the patch fixes a new code
	mm-workingset-per-cgroup-cache-thrash-detection.patch added to -mm tree
	mm-simplify-lock_page_memcg.patch added to -mm tree


so if there is an option to fold this patch into mm-simplify-lock_page_memcg,
for example, as a -fix, then I wouldn't mind at all.



a better commit message
===8<====8<====

From 1d6315221f2f81c53c99f9980158f8ae49dbd582 Mon Sep 17 00:00:00 2001
From: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Date: Wed, 3 Feb 2016 18:49:16 +0900
Subject: [PATCH] mm/workingset: do not forget to unlock_page in workingset_activation

Do not return from workingset_activation() with locked rcu and page.

Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
---
 mm/workingset.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/mm/workingset.c b/mm/workingset.c
index 14522ed..54138a9 100644
--- a/mm/workingset.c
+++ b/mm/workingset.c
@@ -315,8 +315,10 @@ void workingset_activation(struct page *page)
 	 * XXX: See workingset_refault() - this should return
 	 * root_mem_cgroup even for !CONFIG_MEMCG.
 	 */
-	if (!mem_cgroup_disabled() && !page_memcg(page))
+	if (!mem_cgroup_disabled() && !page_memcg(page)) {
+		unlock_page_memcg(page);
 		return;
+	}
 	lruvec = mem_cgroup_zone_lruvec(page_zone(page), page_memcg(page));
 	atomic_long_inc(&lruvec->inactive_age);
 	unlock_page_memcg(page);
-- 
2.7.0

[toc] | [prev] | [next] | [standalone]

#1325611

FromJohannes Weiner <hannes@cmpxchg.org>
Date2016-02-03 17:30 +0100
Message-ID<qY8tz-4kE-1@gated-at.bofh.it>
In reply to#1325185
On Wed, Feb 03, 2016 at 07:41:36PM +0900, Sergey Senozhatsky wrote:
> From 1d6315221f2f81c53c99f9980158f8ae49dbd582 Mon Sep 17 00:00:00 2001
> From: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
> Date: Wed, 3 Feb 2016 18:49:16 +0900
> Subject: [PATCH] mm/workingset: do not forget to unlock_page in workingset_activation
> 
> Do not return from workingset_activation() with locked rcu and page.
> 
> Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>

Thanks Sergey. Even though I wrote this function, my brain must have
gone "it can't be locking anything when it returns NULL, right?" It's
a dumb interface. Luckily, that's fixed with follow-up patches in -mm.

As for this one:

Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Fixes: mm: workingset: per-cgroup cache thrash detection

Andrew, can you please fold this?

[toc] | [prev] | [next] | [standalone]

#1325954

FromAndrew Morton <akpm@linux-foundation.org>
Date2016-02-03 22:20 +0100
Message-ID<qYd0e-7f0-13@gated-at.bofh.it>
In reply to#1325611
On Wed, 3 Feb 2016 11:24:00 -0500 Johannes Weiner <hannes@cmpxchg.org> wrote:

> On Wed, Feb 03, 2016 at 07:41:36PM +0900, Sergey Senozhatsky wrote:
> > From 1d6315221f2f81c53c99f9980158f8ae49dbd582 Mon Sep 17 00:00:00 2001
> > From: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
> > Date: Wed, 3 Feb 2016 18:49:16 +0900
> > Subject: [PATCH] mm/workingset: do not forget to unlock_page in workingset_activation
> > 
> > Do not return from workingset_activation() with locked rcu and page.
> > 
> > Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
> 
> Thanks Sergey. Even though I wrote this function, my brain must have
> gone "it can't be locking anything when it returns NULL, right?" It's
> a dumb interface. Luckily, that's fixed with follow-up patches in -mm.
> 
> As for this one:
> 
> Acked-by: Johannes Weiner <hannes@cmpxchg.org>
> Fixes: mm: workingset: per-cgroup cache thrash detection
> 
> Andrew, can you please fold this?

Yup.  I turned it into a fix against
mm-workingset-per-cgroup-cache-thrash-detection.patch, which is where
the bug was added.  And I did the goto thing instead, so the final
result will be

void workingset_activation(struct page *page)
{
	struct lruvec *lruvec;

	lock_page_memcg(page);
	/*
	 * Filter non-memcg pages here, e.g. unmap can call
	 * mark_page_accessed() on VDSO pages.
	 *
	 * XXX: See workingset_refault() - this should return
	 * root_mem_cgroup even for !CONFIG_MEMCG.
	 */
	if (!mem_cgroup_disabled() && !page_memcg(page))
		goto out;
	lruvec = mem_cgroup_zone_lruvec(page_zone(page), page_memcg(page));
	atomic_long_inc(&lruvec->inactive_age);
out:
	unlock_page_memcg(page);
}

[toc] | [prev] | [next] | [standalone]

#1325996

FromJohannes Weiner <hannes@cmpxchg.org>
Date2016-02-03 23:10 +0100
Message-ID<qYdMC-7MT-21@gated-at.bofh.it>
In reply to#1325954
On Wed, Feb 03, 2016 at 01:19:39PM -0800, Andrew Morton wrote:
> Yup.  I turned it into a fix against
> mm-workingset-per-cgroup-cache-thrash-detection.patch, which is where
> the bug was added.  And I did the goto thing instead, so the final
> result will be
> 
> void workingset_activation(struct page *page)
> {
> 	struct lruvec *lruvec;
> 
> 	lock_page_memcg(page);
> 	/*
> 	 * Filter non-memcg pages here, e.g. unmap can call
> 	 * mark_page_accessed() on VDSO pages.
> 	 *
> 	 * XXX: See workingset_refault() - this should return
> 	 * root_mem_cgroup even for !CONFIG_MEMCG.
> 	 */
> 	if (!mem_cgroup_disabled() && !page_memcg(page))
> 		goto out;
> 	lruvec = mem_cgroup_zone_lruvec(page_zone(page), page_memcg(page));
> 	atomic_long_inc(&lruvec->inactive_age);
> out:
> 	unlock_page_memcg(page);
> }

LGTM, thank you.

[toc] | [prev] | [next] | [standalone]

#1326313

FromSergey Senozhatsky <sergey.senozhatsky.work@gmail.com>
Date2016-02-04 01:20 +0100
Message-ID<qYfOq-Yj-13@gated-at.bofh.it>
In reply to#1325996
On (02/03/16 17:02), Johannes Weiner wrote:
> On Wed, Feb 03, 2016 at 01:19:39PM -0800, Andrew Morton wrote:
> > Yup.  I turned it into a fix against
> > mm-workingset-per-cgroup-cache-thrash-detection.patch, which is where
> > the bug was added.  And I did the goto thing instead, so the final
> > result will be
> > 
> > void workingset_activation(struct page *page)
> > {
> > 	struct lruvec *lruvec;
> > 
> > 	lock_page_memcg(page);
> > 	/*
> > 	 * Filter non-memcg pages here, e.g. unmap can call
> > 	 * mark_page_accessed() on VDSO pages.
> > 	 *
> > 	 * XXX: See workingset_refault() - this should return
> > 	 * root_mem_cgroup even for !CONFIG_MEMCG.
> > 	 */
> > 	if (!mem_cgroup_disabled() && !page_memcg(page))
> > 		goto out;
> > 	lruvec = mem_cgroup_zone_lruvec(page_zone(page), page_memcg(page));
> > 	atomic_long_inc(&lruvec->inactive_age);
> > out:
> > 	unlock_page_memcg(page);
> > }
> 
> LGTM, thank you.

Thanks!

	-ss

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web