Re: [patch 1/3] mm: fix pageref leak in do_swap_page()

[Balbir Singh <balbir@linux.vnet.ibm.com>]

* Johannes Weiner <hannes@cmpxchg.org> [2009-04-21 10:19:03]:

> On Tue, Apr 21, 2009 at 08:44:20AM +0530, Balbir Singh wrote:
> > * Johannes Weiner <hannes@cmpxchg.org> [2009-04-20 22:24:43]:
> > 
> > > By the time the memory cgroup code is notified about a swapin we
> > > already hold a reference on the fault page.
> > > 
> > > If the cgroup callback fails make sure to unlock AND release the page
> > > or we leak the reference.
> > > 
> > > Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
> > > Cc: Balbir Singh <balbir@linux.vnet.ibm.com>
> > 
> > Seems reasonable to me, could you make the changelog more verbose and
> > mention that lookup_swap_cache() gets a reference to the page and we
> > need to release the extra reference.
> 
> Okay, I will add that information.
> 

Thanks

> > BTW, have you had any luck reproducing the issue? How did you catch
> > the problem?
> 
> I reviewed all the exit points when I shuffled code around in there
> for another series that uses a lighter version of do_wp_page() for
> swap write-faults.  I never triggered that problem.
>

Good way of catching a problem, play with the code :) 

-- 
	Balbir

--
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/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>