Re: [PATCH 5/6] xfs: convert the xfsaild threads to a workqueue

From: Dave Chinner <david@fromorbit.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 5/6] xfs: convert the xfsaild threads to a workqueue
Date: Fri, 18 Mar 2011 15:06:48 +1100	[thread overview]
Message-ID: <20110318040648.GG30195@dastard> (raw)
In-Reply-To: <20110310174818.GE19609@infradead.org>

On Thu, Mar 10, 2011 at 12:48:18PM -0500, Christoph Hellwig wrote:
> >  STATIC int __init
> > +xfs_init_workqueues(void)
> 
> > +STATIC void __exit
> > +xfs_destroy_workqueues(void)
> 
> I don't think these helpers are overly useful.

I'm thinking of adding a few more workqueues, so I though I'd split
them out like zone initialisation at the outset....

> 
> > +	xfs_ail_wq = alloc_workqueue("xfsail", WQ_CPU_INTENSIVE, 8);
> 
> > +}
> 
> 
> > +static void
> > +xfs_ail_push_queue(
> > +	struct xfs_ail		*ailp,
> > +	xfs_lsn_t		threshold_lsn,
> > +	int			tout)
> > +{
> > +	if (XFS_LSN_CMP(threshold_lsn, ailp->xa_target) > 0) {
> > +		ailp->xa_target = threshold_lsn;
> > +		queue_delayed_work(xfs_syncd_wq, &ailp->xa_work, tout);
> 
> tout is always one in the only caller and thus doesn't need to be
> passed.  But I think you really want a timeout of 0 here to queue it up
> ASAP (it translates to a direct queue_work() call internally).
> 
> Also this function could simply be merged into it's only and relatively
> simple caller.

It gets used by a second caller in the next patch that uses a
timeout of zero. The idea of adding a delay to a normal push is to
rate limit the number of times we do work so we always work on
batches rather a few items at a time in multiple executions of the
work.

I'll see if it's simpler to just do this work directly in teh
callers, though.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs