[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <20240321174742.6c5622f4406673629e266db7@linux-foundation.org>
Date: Thu, 21 Mar 2024 17:47:42 -0700
From: Andrew Morton <akpm@...ux-foundation.org>
To: Lokesh Gidra <lokeshgidra@...gle.com>
Cc: linux-mm@...ck.org, linux-kernel@...r.kernel.org, surenb@...gle.com,
kernel-team@...roid.com, aarcange@...hat.com, peterx@...hat.com,
david@...hat.com, axelrasmussen@...gle.com, bgeffon@...gle.com,
jannh@...gle.com, kaleshsingh@...gle.com, ngeoffray@...gle.com,
rppt@...nel.org, hdanton@...a.com,
syzbot+49056626fe41e01f2ba7@...kaller.appspotmail.com
Subject: Re: [PATCH] userfaultfd: fix deadlock warning when locking src and
dst VMAs
On Thu, 21 Mar 2024 16:58:18 -0700 Lokesh Gidra <lokeshgidra@...gle.com> wrote:
> Use down_read_nested() to avoid the warning.
>
> --- a/mm/userfaultfd.c
> +++ b/mm/userfaultfd.c
> @@ -1444,7 +1444,8 @@ static int uffd_move_lock(struct mm_struct *mm,
> */
> down_read(&(*dst_vmap)->vm_lock->lock);
> if (*dst_vmap != *src_vmap)
> - down_read(&(*src_vmap)->vm_lock->lock);
> + down_read_nested(&(*src_vmap)->vm_lock->lock,
> + SINGLE_DEPTH_NESTING);
> }
> mmap_read_unlock(mm);
> return err;
Thanks, I added
Fixes: 867a43a34ff8a ("userfaultfd: use per-vma locks in userfaultfd operations")
Powered by blists - more mailing lists