LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Chris Mason <chris.mason@oracle.com>
To: Miklos Szeredi <miklos@szeredi.hu>
Cc: akpm@linux-foundation.org, linux-kernel@vger.kernel.org,
	linux-mm@kvack.org
Subject: Re: dirty balancing deadlock
Date: Mon, 19 Feb 2007 19:16:20 -0500	[thread overview]
Message-ID: <20070220001620.GK6133@think.oraclecorp.com> (raw)
In-Reply-To: <E1HIx6d-0005V4-00@dorka.pomaz.szeredi.hu>

On Mon, Feb 19, 2007 at 02:14:15AM +0100, Miklos Szeredi wrote:
> > > > In general, writepage is supposed to do work without blocking on
> > > > expensive locks that will get pdflush and dirty reclaim stuck in this
> > > > fashion.  You'll probably have to take the same approach reiserfs does
> > > > in data=journal mode, which is leaving the page dirty if fuse_get_req_wp
> > > > is going to block without making progress.
> > > 
> > > Pdflush, and dirty reclaim set wbc->nonblocking to true.
> > > balance_dirty_pages and fsync don't.  The problem here is that
> > > Andrew's patch is wrong to let balance_dirty_pages() try to write back
> > > pages from a different queue.
> > 
> > async or sync, writepage is supposed to either make progress or bail.
> > loopback aside, if the fuse call is blocking long term, you're going to
> > run into problems.
> 
> Hmm, like what?

Something a little different from what you're seeing.  Basically if the
PF_MEMALLOC paths end up waiting on a filesystem transaction, and that
transaction is waiting for more ram, the system will eventually grind to
a halt.  data=journal is the easiest way to hit it, since writepage
always logs at least 4k.

WB_SYNC_NONE and wbc->nonblocking aren't a great test, in reiser I
resorted to testing PF_MEMALLOC.

-chris


  reply	other threads:[~2007-02-20  0:19 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-02-18 18:28 Miklos Szeredi
2007-02-18 20:53 ` Andrew Morton
2007-02-18 21:25   ` Rik van Riel
2007-02-18 22:54     ` Miklos Szeredi
2007-02-18 22:50   ` Miklos Szeredi
2007-02-18 22:59     ` Andrew Morton
2007-02-18 23:22       ` Miklos Szeredi
2007-02-18 23:59         ` Andrew Morton
2007-02-19  0:25           ` Miklos Szeredi
2007-02-19  0:30             ` Miklos Szeredi
2007-02-19  0:45             ` Miklos Szeredi
2007-02-19  0:45             ` Chris Mason
2007-02-19  0:54               ` Miklos Szeredi
2007-02-19  1:01                 ` Chris Mason
2007-02-19  1:14                   ` Miklos Szeredi
2007-02-20  0:16                     ` Chris Mason [this message]
2007-02-20  8:53                       ` Miklos Szeredi
2007-02-19 17:11           ` Miklos Szeredi
2007-02-19 23:12             ` Miklos Szeredi
2007-02-20  0:13             ` Chris Mason
2007-02-20  8:47               ` Miklos Szeredi
2007-02-20 11:30                 ` Chris Mason
2007-02-21 21:36             ` Andrew Morton
2007-02-22  7:42               ` Miklos Szeredi
2007-02-22  7:55                 ` Andrew Morton
2007-02-22  8:02                   ` Miklos Szeredi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20070220001620.GK6133@think.oraclecorp.com \
    --to=chris.mason@oracle.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=miklos@szeredi.hu \
    --subject='Re: dirty balancing deadlock' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).