2004-09-13 15:58:03

by Kirill Korotaev

[permalink] [raw]
Subject: BUG in writeback_inodes()?

Hello All,

It looks like there is a small race bug in writeback_inodes()
Have a look at this 2 call chains:

writeback_inodes()
{
....
sb->s_count++;
spin_unlock(&sb_lock);
....
spin_lock(&sb_lock);
if (__put_super(sb)) <<< X
goto restart;
}
}

deactivate_super()
{
fs->kill_sb(s);
kill_block_super(sb)
generic_shutdown_super(sb)
spin_lock(&sb_lock);
list_del(&sb->s_list); <<< Y
spin_unlock(&sb_lock);
....
put_super(s);
spin_lock(&sb_lock);
__put_super(sb); <<< Z
spin_unlock(&sb_lock);
}

The problem with it is that writeback_inodes() supposes that if
__put_super() returns 0 then no super block was deleted from the list
and we can safely traverse sb list further.

But as it is obvious from the deactivate_super() it's not actually true.
because at point Y we delete super block from the list and drop the
lock. We do __put_super() very much later... So we can find sb with
poisoned sb->s_list at point X and we won't be the last sb reference
holders. The last reference will be dropped in point Z.

So in case of the following sequence of execution Y -> X -> Z we'll get
an oops after point X in writeback_inodes().

Am I correct with it?

Kirill


2004-09-13 20:34:48

by Chris Mason

[permalink] [raw]
Subject: Re: BUG in writeback_inodes()?

On Mon, 2004-09-13 at 12:01, Kirill Korotaev wrote:

> The problem with it is that writeback_inodes() supposes that if
> __put_super() returns 0 then no super block was deleted from the list
> and we can safely traverse sb list further.
>
> But as it is obvious from the deactivate_super() it's not actually true.
> because at point Y we delete super block from the list and drop the
> lock. We do __put_super() very much later... So we can find sb with
> poisoned sb->s_list at point X and we won't be the last sb reference
> holders. The last reference will be dropped in point Z.
>
> So in case of the following sequence of execution Y -> X -> Z we'll get
> an oops after point X in writeback_inodes().
>
> Am I correct with it?

Hmmm, sure looks that way. Seems like it should be enough to switch to
list_del_init in deactivate_super, and then check for
list_empty(sb->s_list) in writeback_inodes.

-chris