[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20120203142119.GM16796@shiny>
Date: Fri, 3 Feb 2012 09:21:19 -0500
From: Chris Mason <chris.mason@...cle.com>
To: Christoph Hellwig <hch@...radead.org>
Cc: Wu Fengguang <fengguang.wu@...el.com>,
Amit Sahrawat <amit.sahrawat83@...il.com>,
Jan Kara <jack@...e.cz>,
Andrew Morton <akpm@...ux-foundation.org>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
Johannes Weiner <jweiner@...hat.com>,
Amit Sahrawat <a.sahrawat@...sung.com>, linux-mm@...ck.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/2] mm: make do_writepages() use plugging
On Fri, Feb 03, 2012 at 09:08:34AM -0500, Christoph Hellwig wrote:
> On Fri, Feb 03, 2012 at 09:38:23PM +0800, Wu Fengguang wrote:
> > On Fri, Feb 03, 2012 at 06:57:06PM +0530, Amit Sahrawat wrote:
> > > This will cover all the invocations for writepages to be called with
> > > plugging support.
> >
> > Thanks. I'll test it on the major filesystems. But would you
> > name a few filesystems that are expected to benefit from it?
> > It's not obvious because some FS ->writepages eventually calls
> > generic_writepages() which already does plugging.
>
> Ant that's exactly where it should stay instead of beeing sprinkled all
> over the VM code.
>
> NAK to the patch.
We've actually had problems with plugging in the higher layers.
writepages is queueing up lots and lots and lots of pages for IO, and
especially on SSDs we want these IOs sent sooner rather than later.
I'm not sure yet how to make the plugs aware of the
please-feed-me-right-away demands of lower storage, but I'd rather not
add more high level plugs that the filesystems can't control.
-chris
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists