linux-kernel - Re: [PATCH -mm] mm, swap: Fix swap space leak in error path of swap_free

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] [day] [month] [year] [list]

Date:   Tue, 25 Apr 2017 15:05:29 -0700
From:   Tim Chen <tim.c.chen@...ux.intel.com>
To:     Andrew Morton <akpm@...ux-foundation.org>,
        "Huang, Ying" <ying.huang@...el.com>
Cc:     linux-mm@...ck.org, linux-kernel@...r.kernel.org,
        Tim Chen <tim.c.chen@...el.com>,
        Rik van Riel <riel@...hat.com>,
        Hugh Dickins <hughd@...gle.com>, Shaohua Li <shli@...nel.org>,
        Minchan Kim <minchan@...nel.org>
Subject: Re: [PATCH -mm] mm, swap: Fix swap space leak in error path of
 swap_free_entries()

On Tue, 2017-04-25 at 14:37 -0700, Andrew Morton wrote:
> On Fri, 21 Apr 2017 20:47:39 +0800 "Huang, Ying" <ying.huang@...el.com> wrote:
> 
> > 
> > From: Huang Ying <ying.huang@...el.com>
> > 
> > In swapcache_free_entries(), if swap_info_get_cont() return NULL,
> > something wrong occurs for the swap entry.  But we should still
> > continue to free the following swap entries in the array instead of
> > skip them to avoid swap space leak.  This is just problem in error
> > path, where system may be in an inconsistent state, but it is still
> > good to fix it.
> > 
> > ...
> > 
> > --- a/mm/swapfile.c
> > +++ b/mm/swapfile.c
> > @@ -1079,8 +1079,6 @@ void swapcache_free_entries(swp_entry_t *entries, int n)
> >  		p = swap_info_get_cont(entries[i], prev);
> >  		if (p)
> >  			swap_entry_free(p, entries[i]);
> > -		else
> > -			break;
> >  		prev = p;
> So now prev==NULL.  Will this code get the locking correct in
> swap_info_get_cont()?  I think so, but please double-check.
> 

There are 4 possible cases, and I checked that the logic
in swap_info_get_cont do the expected:

entries[i]
valid?		prev	 	Expected swap_info_get_cont behavior
---------------------------------------------------------------------
NO		NULL		Return NULL p, Do nothing on lock/unlock
NO		NON-NULL	Return NULL p, Unlock prev 
YES		NULL		Return non-NULL p, lock p
YES		NON-NULL	Return non-NULL p, (p != prev) unlock prev and lock p 
						   (p == prev) do nothing on lock/unlock

Thanks.

Tim

>