linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Christoph Hellwig <hch@infradead.org>
To: Brian Foster <bfoster@redhat.com>
Cc: Dave Chinner <david@fromorbit.com>, linux-xfs@vger.kernel.org
Subject: Re: [PATCH 8/7] xfs: delalloc -> unwritten COW fork allocation can go wrong
Date: Tue, 20 Nov 2018 08:33:49 -0800	[thread overview]
Message-ID: <20181120163349.GB7805@infradead.org> (raw)
In-Reply-To: <20181120134537.GB47711@bfoster>

On Tue, Nov 20, 2018 at 08:45:38AM -0500, Brian Foster wrote:
> >  		 * Filling in all of a previously delayed allocation extent.
> > -		 * The right neighbor is contiguous, the left is not.
> > +		 * The right neighbor is contiguous, the left is not. Take care
> > +		 * with delay -> unwritten extent allocation here because the
> > +		 * delalloc record we are overwriting is always written.
> >  		 */
> >  		PREV.br_startblock = new->br_startblock;
> >  		PREV.br_blockcount += RIGHT.br_blockcount;
> > +		PREV.br_state = new->br_state;
> >  
> >  		xfs_iext_next(ifp, &bma->icur);
> >  		xfs_iext_remove(bma->ip, &bma->icur, state);
> 
> Fix looks sane to me, though I'm a little curious why this doesn't
> follow the "contiguous extent extension" pattern that most of the other
> contig cases seem to follow. For example, something like the following
> for the right contig case:
> 
>                 old = RIGHT;
>                 RIGHT.br_startoff = new->br_startoff;
>                 RIGHT.br_startblock = new->br_startblock;
>                 RIGHT.br_blockcount += new.br_blockcount;
> 
>                 xfs_iext_remove(bma->ip, &bma->icur, state); /* PREV */
>                 xfs_iext_next(ifp, &bma->icur);
>                 xfs_iext_update_extent(bma->ip, state, &bma->icur, &RIGHT);
> 
> ... and change the subsequent btree update to use old/RIGHT. Maybe
> Christoph has thoughts on that.

Not sure the above looks much better.  If I had to do something
different I'd apply something like this (untested) on top of Daves patch:

diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 91b94c5c0cae..efe9e554fd0d 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -1698,14 +1698,12 @@ xfs_bmap_add_extent_delay_real(
 		 * with delay -> unwritten extent allocation here because the
 		 * delalloc record we are overwriting is always written.
 		 */
-		PREV.br_startblock = new->br_startblock;
-		PREV.br_blockcount += RIGHT.br_blockcount;
-		PREV.br_state = new->br_state;
+		new->br_blockcount += RIGHT.br_blockcount;
 
 		xfs_iext_next(ifp, &bma->icur);
 		xfs_iext_remove(bma->ip, &bma->icur, state);
 		xfs_iext_prev(ifp, &bma->icur);
-		xfs_iext_update_extent(bma->ip, state, &bma->icur, &PREV);
+		xfs_iext_update_extent(bma->ip, state, &bma->icur, new);
 
 		if (bma->cur == NULL)
 			rval = XFS_ILOG_DEXT;

  reply	other threads:[~2018-11-21  3:03 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-11-19 21:04 [PATCH 0/7] xfs: various fixes for 4.20 Dave Chinner
2018-11-19 21:04 ` [PATCH 1/7] xfs: zero length symlinks are not valid Dave Chinner
2018-11-20  8:12   ` Christoph Hellwig
2018-11-20 13:44   ` Brian Foster
2018-11-20 21:19     ` Dave Chinner
2018-11-21 12:01       ` Brian Foster
2018-11-19 21:04 ` [PATCH 2/7] xfs: uncached buffer tracing needs to print bno Dave Chinner
2018-11-20  8:12   ` Christoph Hellwig
2018-11-20 22:46   ` Darrick J. Wong
2018-11-19 21:04 ` [PATCH 3/7] xfs: fix transient reference count error in xfs_buf_resubmit_failed_buffers Dave Chinner
2018-11-20  8:13   ` Christoph Hellwig
2018-11-20 22:48   ` Darrick J. Wong
2018-11-19 21:04 ` [PATCH 4/7] xfs: finobt AG reserves don't consider last AG can be a runt Dave Chinner
2018-11-20  8:14   ` Christoph Hellwig
2018-11-20 22:49   ` Darrick J. Wong
2018-11-19 21:04 ` [PATCH 5/7] xfs: extent shifting doesn't fully invalidate page cache Dave Chinner
2018-11-20  8:18   ` Christoph Hellwig
2018-11-20 22:53   ` Darrick J. Wong
2018-11-19 21:04 ` [PATCH 6/7] xfs: don't ENOSPC on writeback when punching holes Dave Chinner
2018-11-20  8:20   ` Christoph Hellwig
2018-11-20  9:50     ` Dave Chinner
2018-11-20 16:28       ` Christoph Hellwig
2018-11-20 21:00         ` Dave Chinner
2018-11-21 18:09   ` Darrick J. Wong
2018-11-22  2:31     ` Dave Chinner
2018-11-19 21:04 ` [PATCH 7/7] xfs: flush removing page cache in xfs_reflink_remap_prep Dave Chinner
2018-11-20  8:32   ` Christoph Hellwig
2018-11-20 22:56   ` Darrick J. Wong
2018-11-20  6:36 ` [PATCH 8/7] xfs: delalloc -> unwritten COW fork allocation can go wrong Dave Chinner
2018-11-20 13:45   ` Brian Foster
2018-11-20 16:33     ` Christoph Hellwig [this message]
2018-11-20 21:08       ` Dave Chinner
2018-11-20 16:32   ` Christoph Hellwig
2018-11-20 22:58   ` Darrick J. Wong

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=20181120163349.GB7805@infradead.org \
    --to=hch@infradead.org \
    --cc=bfoster@redhat.com \
    --cc=david@fromorbit.com \
    --cc=linux-xfs@vger.kernel.org \
    /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 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).