All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Christoph Hellwig <hch@lst.de>
Cc: Dave Chinner <david@fromorbit.com>, linux-xfs@vger.kernel.org
Subject: Re: [PATCH 6/6] xfs: avoid COW fork extent lookups in writeback if the fork didn't change
Date: Tue, 24 Jul 2018 15:35:23 -0700	[thread overview]
Message-ID: <20180724223523.GO4813@magnolia> (raw)
In-Reply-To: <20180723074940.GA18816@lst.de>

On Mon, Jul 23, 2018 at 09:49:41AM +0200, Christoph Hellwig wrote:
> On Sun, Jul 22, 2018 at 09:23:11AM +1000, Dave Chinner wrote:
> > On Tue, Jul 17, 2018 at 04:24:05PM -0700, Christoph Hellwig wrote:
> > > @@ -333,12 +335,15 @@ xfs_map_blocks(
> > >  	 * COW fork blocks can overlap data fork blocks even if the blocks
> > >  	 * aren't shared.  COW I/O always takes precedent, so we must always
> > >  	 * check for overlap on reflink inodes unless the mapping is already a
> > > -	 * COW one.
> > > +	 * COW one, or the COW fork hasn't changed from the last time we looked
> > > +	 * at it.
> > >  	 */
> > >  	imap_valid = offset_fsb >= wpc->imap.br_startoff &&
> > >  		     offset_fsb < wpc->imap.br_startoff + wpc->imap.br_blockcount;
> > >  	if (imap_valid &&
> > > -	    (!xfs_inode_has_cow_data(ip) || wpc->io_type == XFS_IO_COW))
> > > +	    (!xfs_inode_has_cow_data(ip) ||
> > > +	     wpc->io_type == XFS_IO_COW ||
> > > +	     wpc->cow_seq == ip->i_cowfp->if_seq))
> > >  		return 0;
> > 
> > This seqno check is still racy against concurrent extent tree
> > modification. We aren't holding the ILOCK here at all, the sequence
> > numbers aren't atomic and there are no memory barriers to serialise
> > cross-cpu load/store operations...
> 
> mutex_unlock contains a barrier, and the sequence is a single register
> read.  There is nothing holding a lock here would help us with.

Which mutex_unlock is that?  I took a second look at the i_cowfp access
and realized that we don't take ILOCK_SHARED until after the comparison.
Writeback doesn't take any of the other inode locks AFAIK... so either
there's some locking subtlety here that ought to be explained in a
comment, or is this a bug waiting to happen?

--D

> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

  reply	other threads:[~2018-07-24 23:44 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-07-17 23:23 reduce lookups in the COW extent tree V3 Christoph Hellwig
2018-07-17 23:24 ` [PATCH 1/6] xfs: remove if_real_bytes Christoph Hellwig
2018-07-17 23:24 ` [PATCH 2/6] xfs: simplify xfs_idata_realloc Christoph Hellwig
2018-07-17 23:24 ` [PATCH 3/6] xfs: remove the xfs_ifork_t typedef Christoph Hellwig
2018-07-17 23:24 ` [PATCH 4/6] xfs: introduce a new xfs_inode_has_cow_data helper Christoph Hellwig
2018-07-17 23:24 ` [PATCH 5/6] xfs: maintain a sequence count for inode fork manipulations Christoph Hellwig
2018-07-18 14:40   ` Carlos Maiolino
2018-07-19 16:32     ` Christoph Hellwig
2018-07-19 18:27       ` Darrick J. Wong
2018-07-23 12:11         ` Carlos Maiolino
2018-07-17 23:24 ` [PATCH 6/6] xfs: avoid COW fork extent lookups in writeback if the fork didn't change Christoph Hellwig
2018-07-18 14:51   ` Carlos Maiolino
2018-07-21 23:23   ` Dave Chinner
2018-07-23  7:49     ` Christoph Hellwig
2018-07-24 22:35       ` Darrick J. Wong [this message]
2018-07-27 15:10         ` Christoph Hellwig
2018-08-06  2:37           ` Dave Chinner
2018-08-06 16:45             ` Christoph Hellwig
  -- strict thread matches above, loose matches on Subject: below --
2018-07-12 13:49 reduce lookups in the COW extent tree V2 Christoph Hellwig
2018-07-12 13:49 ` [PATCH 6/6] xfs: avoid COW fork extent lookups in writeback if the fork didn't change Christoph Hellwig
2018-07-13 22:51   ` Darrick J. Wong
2018-07-14  0:03   ` Dave Chinner
2018-07-17 13:37     ` Christoph Hellwig
2018-07-17 23:13       ` Dave Chinner
2018-07-10  6:05 reduce lookups in the COW extent tree Christoph Hellwig
2018-07-10  6:05 ` [PATCH 6/6] xfs: avoid COW fork extent lookups in writeback if the fork didn't change Christoph Hellwig
2018-07-11 17:15   ` Darrick J. Wong
2018-07-11 17:20     ` Christoph Hellwig
2018-07-11 17:31       ` Darrick J. Wong
2018-07-11 17:35         ` 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=20180724223523.GO4813@magnolia \
    --to=darrick.wong@oracle.com \
    --cc=david@fromorbit.com \
    --cc=hch@lst.de \
    --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 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.