All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ben Myers <bpm@sgi.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: Paul Anderson <pha@umich.edu>,
	Sean Thomas Caron <scaron@umich.edu>,
	Mark Tinguely <tinguely@sgi.com>,
	xfs@oss.sgi.com
Subject: Re: [PATCH 2/2 v2] xfs: log all dirty inodes in xfs_fs_sync_fs
Date: Fri, 23 Dec 2011 15:47:03 -0600	[thread overview]
Message-ID: <20111223214703.GW29840@sgi.com> (raw)
In-Reply-To: <20111220200841.GA2788@infradead.org>

Hey Christoph,

On Tue, Dec 20, 2011 at 03:08:41PM -0500, Christoph Hellwig wrote:
> Since Linux 2.6.36 the writeback code has introduces various measures for
> live lock prevention during sync().  Unfortunately some of these are
> actively harmful for the XFS model, where the inode gets marked dirty for
> metadata from the data I/O handler.
> 
> The older_than_this checks that are now more strictly enforced since
> 
>     writeback: avoid livelocking WB_SYNC_ALL writeback
> 
> by only calling into __writeback_inodes_sb and thus only sampling the

Do you mean __writeback_inodes_wb, or perhaps writeback_sb_inodes?

So far I'm not seeing the connection with the commit you've referenced
above, which seems to be related to nr_to_write.  It looks like you're
referring to older fs/fs-writeback.c..

This one seems like it might be relevant:

commit 7624ee7
mm: avoid resetting wb_start after each writeback round

Anyway.. in the 3.2-rcX code it appears that older_than_this is only set
at the start of wb_writeback (excluding for_kupdate being set)...

> current cut off time once.  But on a slow enough devices the previous
> asynchronous sync pass might not have fully completed yet, and thus XFS
> might mark metadata dirty only after that sampling of the cut off time for
> the blocking pass already happened. 

Are you referring to sync_filesystem() calling
__sync_filesystem(sb, 0)
and then
__sync_filesystem(sb, 1 /* wait */)?

Ah... and with that I think I understand what you're after here:  After
__sync_filesystem calls sync_inodes_sb and waits for completion... which
will dirty more inodes in completion handlers... you have the
opportunity to clean up those dirty inodes in .sync_fs.  If that's what
you intend:  I'd say this looks good.

Reviewed-by: Ben Myers <bpm@sgi.com>

Mark also reviewed this.

Reviewed-by: Mark Tinguely <tinguely@sgi.com>

THanks,
	Ben

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

  parent reply	other threads:[~2011-12-23 21:46 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20111218154936.GA17626@infradead.org>
2011-12-18 15:49 ` [PATCH 1/1] xfs: log the inode in ->write_inode calls for kupdate Christoph Hellwig
2011-12-20 21:19   ` Dave Chinner
2011-12-23 15:55   ` Mark Tinguely
2011-12-23 17:58   ` Ben Myers
2011-12-28 21:35   ` Hans-Peter Jansen
2011-12-28 21:39     ` Christoph Hellwig
2011-12-18 15:50 ` [PATCH 2/2] xfs: log all dirty inodes in xfs_fs_sync_fs Christoph Hellwig
2011-12-18 20:03   ` Sean Thomas Caron
2011-12-18 20:09     ` Christoph Hellwig
2011-12-18 22:17   ` Dave Chinner
2011-12-18 22:32     ` Christoph Hellwig
2011-12-20 20:08   ` [PATCH 2/2 v2] " Christoph Hellwig
2011-12-20 21:21     ` Dave Chinner
2011-12-23 15:55     ` Mark Tinguely
2011-12-23 21:47     ` Ben Myers [this message]
2011-12-26 12:13       ` Dave Chinner
2011-12-29 15:42         ` Ben Myers
2011-12-29 21:44           ` Dave Chinner
2012-01-03 15:48             ` Mark Tinguely
2011-12-21 17:40 ` sync fixes Christoph Hellwig

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=20111223214703.GW29840@sgi.com \
    --to=bpm@sgi.com \
    --cc=hch@infradead.org \
    --cc=pha@umich.edu \
    --cc=scaron@umich.edu \
    --cc=tinguely@sgi.com \
    --cc=xfs@oss.sgi.com \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.