Re: [PATCH v2] fs/nfsd/nfs4callback: Remove deprecated create_singlethread_workqueue

[Jeff Layton]

On Wed, 2016-11-09 at 15:08 +0000, Trond Myklebust wrote:
> On Wed, 2016-11-09 at 08:18 -0500, Jeff Layton wrote:
> > 
> > On Tue, 2016-11-08 at 20:27 -0500, J. Bruce Fields wrote:
> > > 
> > > 
> > > On Tue, Nov 08, 2016 at 05:52:21PM -0500, Tejun Heo wrote:
> > > > 
> > > > 
> > > > 
> > > > Hello, Bruce.
> > > > 
> > > > On Tue, Nov 08, 2016 at 04:39:11PM -0500, J. Bruce Fields wrote:
> > > > > 
> > > > > 
> > > > > 
> > > > > Apologies, just cleaning out old mail and finding some I should
> > > > > have
> > > > > responded to long ago:
> > > > > 
> > > > > On Wed, Aug 31, 2016 at 02:23:48AM +0530, Bhaktipriya Shridhar
> > > > > wrote:
> > > > > > 
> > > > > > 
> > > > > > 
> > > > > > The workqueue "callback_wq" queues a single work item &cb-
> > > > > > > 
> > > > > > > cb_work per
> > > > > > nfsd4_callback instance and thus, it doesn't require
> > > > > > execution ordering.
> > > > > 
> > > > > What's "execution ordering"?
> > > > > 
> > 
> > AIUI, it means that jobs are always run in the order queued and are
> > serialized.
> > 
> > > 
> > > 
> > > > 
> > > > 
> > > > > 
> > > > > 
> > > > > We definitely do depend on the fact that at most one of these
> > > > > is running
> > > > > at a time.
> > > > 
> > 
> > We do?
> > 
> > > 
> > > 
> > > > 
> > > > 
> > > > If there can be multiple cb's and thus cb->cb_work's per
> > > > callback_wq,
> > > > it'd need explicit ordering.ÂÂIs that the case?
> > > 
> > 
> > These are basically client RPC tasks, and the cb_work just handles
> > the
> > submission into the client RPC state machine. Just because we're
> > running
> > several callbacks at the same time doesn't mean that they need to be
> > strictly ordered. The client state machine can certainly handle
> > running
> > these in parallel.
> > 
> > > 
> > > 
> > > Yes, there can be multiple cb_work's.
> > > 
> > 
> > Yes, but each is effectively a separate work unit. I see no reason
> > why
> > we'd need to order them at all.
> > 
> 
> There needs to be serialisation at the session level (i.e. the
> callbacks have to respect the slot limits set by the client) however
> there shouldnât be a need for serialisation at the RPC level.
> 
> Cheers
> Â Trond

Yes, that all happens in nfsd4_cb_prepare, which is the rpc_call_prepare
operation for the callback. That gets run by the rpc state machine in
the context of the rpciod workqueues. None of that happens in the
context of the cb_work here.

If you have a look at nfsd4_run_cb_work, you can see that it just does a
cb_ops->prepare and then submits it to the client rpc engine with
rpc_call_async. None of that should require singlethreaded workqueue
semantics.

--Â
Jeff Layton <jlayton@xxxxxxxxxxxxxxx>