Re: [RFC PATCH] cifs: Fix possible deadlock with cifs and work queues

[Jeff Layton <jlayton@redhat.com>]

On Fri, 21 Mar 2014 08:17:06 -0400
Steven Rostedt <rostedt@goodmis.org> wrote:

> On Fri, 21 Mar 2014 12:32:12 +0400
> Pavel Shilovsky <piastry@etersoft.ru> wrote:
> 
> 
> > Read and write codepaths both obtain lock_sem for read and then wait
> > for cifsiod_wq to complete and release lock_sem. They don't do any
> > lock_sem operations inside their work task queued to cifsiod_wq. But
> > oplock code can obtain/release lock_sem in its work task. So, that's
> > why I agree with Jeff and suggest to move the oplock code to a
> > different work queue (cifsioopd_wq?) but leave read and write
> > codepaths use cifsiod_wq.
> 
> OK, how about I submit a second patch that moves the reader and writer
> to its own "safe" workqueue?
> 
> -- Steve
> 

That'd probably work fine too. The main point is to make sure oplock
breaks run on a different workqueue from where read or write completion
jobs run since they are operating on the lock_sem. The other jobs that
get queued to cifsiod_wq don't touch the lock_sem and shouldn't be a
problem.

-- 
Jeff Layton <jlayton@redhat.com>