[<prev] [next>] [day] [month] [year] [list]
Message-ID: <CAHc6FU4iHhbJ-mfOR=sJMNXfMjLLfr7fnrRxQ1cmj9VaT9yQFg@mail.gmail.com>
Date: Thu, 20 May 2021 13:29:02 +0200
From: Andreas Gruenbacher <agruenba@...hat.com>
To: Hillf Danton <hdanton@...a.com>
Cc: cluster-devel <cluster-devel@...hat.com>,
LKML <linux-kernel@...r.kernel.org>,
Bob Peterson <rpeterso@...hat.com>,
syzkaller-bugs@...glegroups.com,
syzbot <syzbot+34ba7ddbf3021981a228@...kaller.appspotmail.com>
Subject: Re: [PATCH] gfs2: fix uaf in glock shrinker
On Wed, May 19, 2021 at 8:08 AM Hillf Danton <hdanton@...a.com> wrote:
> The GLF_LRU flag is checked under lru_lock in gfs2_glock_remove_from_lru() to
> remove glock from the lru list in __gfs2_glock_put().
>
> On the shrink scan path, the same flag is cleared under lru_lock but because
> of cond_resched_lock(&lru_lock) in gfs2_dispose_glock_lru() progress on the
> put side can be made without deleting glock from lru.
>
> Keep GLF_LRU across the race window opened by cond_resched_lock(&lru_lock) to
> ensure correct behaviors on both sides - clear GLF_LRU after list_del under
> lru_lock. That is the fix to uaf reported by the smart syzbot.
Thanks, I've added this to for-next.
Andreas
Powered by blists - more mailing lists