xfs
[Top] [All Lists]

Re: [PATCH 3/8] ext4: Warn if direct reclaim tries to writeback pages

To: Johannes Weiner <jweiner@xxxxxxxxxx>
Subject: Re: [PATCH 3/8] ext4: Warn if direct reclaim tries to writeback pages
From: Mel Gorman <mgorman@xxxxxxx>
Date: Wed, 3 Aug 2011 15:35:27 +0100
Cc: Linux-MM <linux-mm@xxxxxxxxx>, LKML <linux-kernel@xxxxxxxxxxxxxxx>, XFS <xfs@xxxxxxxxxxx>, Dave Chinner <david@xxxxxxxxxxxxx>, Christoph Hellwig <hch@xxxxxxxxxxxxx>, Wu Fengguang <fengguang.wu@xxxxxxxxx>, Jan Kara <jack@xxxxxxx>, Rik van Riel <riel@xxxxxxxxxx>, Minchan Kim <minchan.kim@xxxxxxxxx>
In-reply-to: <20110803140019.GA31026@xxxxxxxxxx>
References: <1311265730-5324-1-git-send-email-mgorman@xxxxxxx> <1311265730-5324-4-git-send-email-mgorman@xxxxxxx> <20110803105819.GA27199@xxxxxxxxxx> <20110803110629.GB27199@xxxxxxxxxx> <20110803134420.GH19099@xxxxxxx> <20110803140019.GA31026@xxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Wed, Aug 03, 2011 at 04:00:19PM +0200, Johannes Weiner wrote:
> On Wed, Aug 03, 2011 at 02:44:20PM +0100, Mel Gorman wrote:
> > On Wed, Aug 03, 2011 at 01:06:29PM +0200, Johannes Weiner wrote:
> > > On Wed, Aug 03, 2011 at 12:58:19PM +0200, Johannes Weiner wrote:
> > > > On Thu, Jul 21, 2011 at 05:28:45PM +0100, Mel Gorman wrote:
> > > > > Direct reclaim should never writeback pages. Warn if an attempt
> > > > > is made.
> > > > > 
> > > > > Signed-off-by: Mel Gorman <mgorman@xxxxxxx>
> > > > 
> > > > Acked-by: Johannes Weiner <jweiner@xxxxxxxxxx>
> > > 
> > > Oops, too fast.
> > > 
> > > Shouldn't the WARN_ON() be at the top of the function, rather than
> > > just warn when the write is deferred due to delalloc?
> > 
> > I thought it made more sense to put the warning at the point where ext4
> > would normally ignore ->writepage.
> > 
> > That said, in my current revision of the series, I've dropped these
> > patches altogether as page migration should be able to trigger the same
> > warnings but be called from paths that are of less concern for stack
> > overflows (or at the very least be looked at as a separate series).
> 
> Doesn't this only apply to btrfs which has no own .migratepage aop for
> file pages?  The others use buffer_migrate_page.
> 

Bah, you're right. It was btrfs I was looking at during the time I
decided to drop the patches and I didn't think it through. I only
needed to drop the btrfs one.

> But if you dropped them anyway, it does not matter :)

I put back in the xfs and ext4 checks. The ext4 check is still in the
same place.

-- 
Mel Gorman
SUSE Labs

<Prev in Thread] Current Thread [Next in Thread>