Re: [PATCH] Fix shrink_dcache_parent() against shrink_dcache_memory() race (updated patch)

[Jan Blunck <jblunck@suse.de>]

On Thu, Mar 09, Kirill Korotaev wrote:

> >Thanks! I'll send the corrected patch.
> >So, everythings fine now?
> looks so! Will be glad to Ack/Sign or whatever needed :)))
> 

Ok.

> >>>	d_free(dentry);
> >>>	if (parent != dentry)
> >>>		dput(parent);
> >>>	spin_lock(&dcache_lock);
> >>>+	sb->s_prunes--;
> >>>+	if (likely(!sb->s_prunes))
> >>
> >><<< Is it possibe to do something like:
> >>if (unlikely(!sb->s_root && !sb->s_prunes))
> >>?
> >
> >
> >Uh, I forgot about that one. You already complained about that before :(
> But I'm not sure it is that simple... s_root is set to NULL w/o locks, 
> so I wonder whether it is safe to check it here or we can miss some 
> wakeups...

No, it's not. We need to down_read(&sb->s_umount) for that which is
deadlocking because we down_write() it before calling ->kill_sb(). So this
isn't safe. For now I'll keep it like before and live with the overhead of
calling wake_up() on an empty wait-queue.

Regards,
	Jan

-- 
Jan Blunck                                               jblunck@suse.de
SuSE LINUX AG - A Novell company
Maxfeldstr. 5                                          +49-911-74053-608
D-90409 Nürnberg                                      http://www.suse.de