linux-kernel - [PATCH v2 2/6] oom: Get rid of sparse warnings

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 7 Feb 2012 08:20:26 +0400
From:	Anton Vorontsov <anton.vorontsov@...aro.org>
To:	KOSAKI Motohiro <kosaki.motohiro@...il.com>
Cc:	Oleg Nesterov <oleg@...hat.com>,
	Greg KH <gregkh@...uxfoundation.org>,
	Arve Hjønnevåg <arve@...roid.com>,
	San Mehat <san@...gle.com>, Colin Cross <ccross@...roid.com>,
	"Eric W. Biederman" <ebiederm@...ssion.com>,
	"Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>,
	linux-kernel@...r.kernel.org, kernel-team@...roid.com,
	linaro-kernel@...ts.linaro.org
Subject: [PATCH v2 2/6] oom: Get rid of sparse warnings

Sparse flood makes it hard to catch newly-introduced warnings. So let's
fix the the sparse warnings in the oom killer:

  CHECK   mm/oom_kill.c
  mm/oom_kill.c:139:20: warning: context imbalance in
	  '__find_lock_task_mm' - wrong count at exit
  mm/oom_kill.c:771:9: warning: context imbalance in 'out_of_memory' -
	  different lock contexts for basic block

The first one is fixed by assuring sparse that we know that we exit
with the lock held.

The second one is caused by the fact that sparse isn't smart enough
to handle noreturn attribute.

Signed-off-by: Anton Vorontsov <anton.vorontsov@...aro.org>
---

On Mon, Feb 06, 2012 at 04:33:26PM -0500, KOSAKI Motohiro wrote:
> 2012/2/6 Anton Vorontsov <anton.vorontsov@...aro.org>:
> > Sparse flood makes it hard to catch newly-introduced warnings. So let's
> > fix the the sparse warnings in the oom killer:
> >
> >  CHECK   mm/oom_kill.c
> >  mm/oom_kill.c:139:20: warning: context imbalance in
> >          '__find_lock_task_mm' - wrong count at exit
> >  mm/oom_kill.c:771:9: warning: context imbalance in 'out_of_memory' -
> >          different lock contexts for basic block
> >
> > The first one is fixed by assuring sparse that we know that we exit
> > with the lock held.
> >
> > The second one is caused by the fact that sparse isn't smart enough
> > to handle noreturn attribute.
> >
> > Signed-off-by: Anton Vorontsov <anton.vorontsov@...aro.org>
> > ---
> >  mm/oom_kill.c |    9 ++++++++-
> >  1 files changed, 8 insertions(+), 1 deletions(-)
> >
> > diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> > index 0ebb383..49569b6 100644
> > --- a/mm/oom_kill.c
> > +++ b/mm/oom_kill.c
> > @@ -142,8 +142,14 @@ struct task_struct *__find_lock_task_mm(struct task_struct *p)
> >
> >        do {
> >                task_lock(t);
> > -               if (likely(t->mm))
> > +               if (likely(t->mm)) {
> > +                       /*
> > +                        * Shut up sparse: we do know that we exit w/ the
> > +                        * task locked.
> > +                        */
> > +                       __release(&t->alloc_loc);
> 
> task struct only have allock_lock, not alloc_loc.

Funnily, but sparse does not care. :-) __release(foo) will work as
well. Seems like sparse counts locking balance globally.

This is now fixed in the patch down below, thanks for catching.

> Moreover we don't release
> the lock in this code path. Seems odd.

Indeed. That's exactly what sparse seeing is as well. We exit
without releasing the lock, which is bad (in sparse' eyes). So
we lie to sparse, telling it that we do release, so it shut ups.

Usually, we annotate functions with __releases() and __acquires()
when when functions releases/acquires one of its arguments, but in
this case the function returns a locked value, and it seems that
there's no way to properly annotate this case.

 mm/oom_kill.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/mm/oom_kill.c b/mm/oom_kill.c
index 0ebb383..1914367 100644
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -142,8 +142,14 @@ struct task_struct *__find_lock_task_mm(struct task_struct *p)
 
 	do {
 		task_lock(t);
-		if (likely(t->mm))
+		if (likely(t->mm)) {
+			/*
+			 * Shut up sparse: we do know that we exit w/ the
+			 * task locked.
+			 */
+			__release(&t->alloc_lock);
 			return t;
+		}
 		task_unlock(t);
 	} while_each_thread(p, t);
 
@@ -766,6 +772,7 @@ retry:
 		dump_header(NULL, gfp_mask, order, NULL, mpol_mask);
 		read_unlock(&tasklist_lock);
 		panic("Out of memory and no killable processes...\n");
+		return;
 	}
 
 	if (oom_kill_process(p, gfp_mask, order, points, totalpages, NULL,
-- 
1.7.7.6

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/