* [PATCH] kmemleak: do not leak object after tree insertion error (v2, fixed)
@ 2012-04-02 23:06 Sergey Senozhatsky
2012-04-18 14:40 ` Catalin Marinas
0 siblings, 1 reply; 3+ messages in thread
From: Sergey Senozhatsky @ 2012-04-02 23:06 UTC (permalink / raw)
To: Catalin Marinas; +Cc: linux-mm, linux-kernel
[PATCH] kmemleak: do not leak object after tree insertion error
In case when tree insertion fails due to already existing object
error, pointer to allocated object gets lost due to lookup_object()
overwrite. Free allocated object and return the existing one,
obtained from lookup_object().
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
---
mm/kmemleak.c | 19 ++++++++++++-------
1 file changed, 12 insertions(+), 7 deletions(-)
diff --git a/mm/kmemleak.c b/mm/kmemleak.c
index 45eb621..4177d83 100644
--- a/mm/kmemleak.c
+++ b/mm/kmemleak.c
@@ -516,7 +516,7 @@ static struct kmemleak_object *create_object(unsigned long ptr, size_t size,
int min_count, gfp_t gfp)
{
unsigned long flags;
- struct kmemleak_object *object;
+ struct kmemleak_object *object, *ex_object;
struct prio_tree_node *node;
object = kmem_cache_alloc(object_cache, gfp_kmemleak_mask(gfp));
@@ -578,17 +578,22 @@ static struct kmemleak_object *create_object(unsigned long ptr, size_t size,
if (node != &object->tree_node) {
kmemleak_stop("Cannot insert 0x%lx into the object search tree "
"(already existing)\n", ptr);
- object = lookup_object(ptr, 1);
- spin_lock(&object->lock);
- dump_object_info(object);
- spin_unlock(&object->lock);
+ ex_object = lookup_object(ptr, 1);
+ spin_lock(&ex_object->lock);
+ dump_object_info(ex_object);
+ spin_unlock(&ex_object->lock);
- goto out;
+ goto out_error;
}
list_add_tail_rcu(&object->object_list, &object_list);
-out:
+
write_unlock_irqrestore(&kmemleak_lock, flags);
return object;
+out_error:
+ write_unlock_irqrestore(&kmemleak_lock, flags);
+ object->flags &= ~OBJECT_ALLOCATED;
+ put_object(object);
+ return ex_object;
}
/*
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH] kmemleak: do not leak object after tree insertion error (v2, fixed)
2012-04-02 23:06 [PATCH] kmemleak: do not leak object after tree insertion error (v2, fixed) Sergey Senozhatsky
@ 2012-04-18 14:40 ` Catalin Marinas
2012-04-18 15:29 ` Sergey Senozhatsky
0 siblings, 1 reply; 3+ messages in thread
From: Catalin Marinas @ 2012-04-18 14:40 UTC (permalink / raw)
To: Sergey Senozhatsky; +Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org
On Tue, Apr 03, 2012 at 12:06:56AM +0100, Sergey Senozhatsky wrote:
> [PATCH] kmemleak: do not leak object after tree insertion error
>
> In case when tree insertion fails due to already existing object
> error, pointer to allocated object gets lost due to lookup_object()
> overwrite. Free allocated object and return the existing one,
> obtained from lookup_object().
We really need to return NULL if the tree insertion fails as kmemleak is
disabled in this case (fatal condition for kmemleak). So we could just
call kmem_cache_free(object_cache, object) in the 'if' block.
--
Catalin
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] kmemleak: do not leak object after tree insertion error (v2, fixed)
2012-04-18 14:40 ` Catalin Marinas
@ 2012-04-18 15:29 ` Sergey Senozhatsky
0 siblings, 0 replies; 3+ messages in thread
From: Sergey Senozhatsky @ 2012-04-18 15:29 UTC (permalink / raw)
To: Catalin Marinas; +Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org
On (04/18/12 15:40), Catalin Marinas wrote:
> On Tue, Apr 03, 2012 at 12:06:56AM +0100, Sergey Senozhatsky wrote:
> > [PATCH] kmemleak: do not leak object after tree insertion error
> >
> > In case when tree insertion fails due to already existing object
> > error, pointer to allocated object gets lost due to lookup_object()
> > overwrite. Free allocated object and return the existing one,
> > obtained from lookup_object().
>
> We really need to return NULL if the tree insertion fails as kmemleak is
> disabled in this case (fatal condition for kmemleak). So we could just
> call kmem_cache_free(object_cache, object) in the 'if' block.
>
Good point. Thanks a lot for your review!
I was chasing two bugs and sort of messed things up. I'll send v3 shortly.
Sergey
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2012-04-18 15:30 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-04-02 23:06 [PATCH] kmemleak: do not leak object after tree insertion error (v2, fixed) Sergey Senozhatsky
2012-04-18 14:40 ` Catalin Marinas
2012-04-18 15:29 ` Sergey Senozhatsky
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).