public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: Dave Chinner <david@fromorbit.com>
To: Christoph Hellwig <hch@infradead.org>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 09/11] xfs: remove the i_new_size field in struct xfs_inode
Date: Wed, 14 Dec 2011 10:16:11 +1100	[thread overview]
Message-ID: <20111213231611.GG3179@dastard> (raw)
In-Reply-To: <20111208155919.025710280@bombadil.infradead.org>

On Thu, Dec 08, 2011 at 10:58:04AM -0500, Christoph Hellwig wrote:
> Now that we use the VFS i_size field throughout XFS there is no need for the
> i_new_size field any more given that the VFS i_size field gets updated
> in ->write_end before unlocking the page, and thus is a) always uptodate when
> writeback could see a page.  Removing i_new_size also has the advantage that
> we will never have to trim back di_size during a failed buffered write,
> given that it never gets updated past i_size.
> 
> Note that currently the generic direct I/O code only updates i_size after
> calling our end_io handler, which requires a small workaround to make
> sure di_size actually makes it to disk.  I hope to fix this properly in
> the generic code.

I think there's a couple of issues with the work around - the
generic code marks the inode dirty when it updates the inode size.
With your early setting of the size, it will no longer do this
because it doesn't see that it needs to update the inode size.
This may not be a problem if we mark the inode dirty elsewhere, but
I'm not sure we do in the direct IO path.

Apart from that, I don't see anything wrong with the change. I'll
wait for you to clarify whether we need to mark the inode dirty
before signing off on it, though....

> A downside is that we lose the support for parallel non-overlapping O_DIRECT
> appending writes that recently was added.  I don't think keeping the complex
> and fragile i_new_size infrastructure for this is a good tradeoff - if we
> really care about parallel appending writers we should investigate turning
> the iolock into a range lock, which would also allow for parallel
> non-overlapping buffered writers.

Yeah, i think we can live without that for the moment. Range locks
are a much better idea for many reasons (like concurrent buffered
write support).

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

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

  reply	other threads:[~2011-12-13 23:16 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-12-08 15:57 [PATCH 00/11] inode shrink and misc updates Christoph Hellwig
2011-12-08 15:57 ` [PATCH 01/11] xfs: remove xfs_itruncate_data Christoph Hellwig
2011-12-13 21:23   ` Dave Chinner
2011-12-08 15:57 ` [PATCH 02/11] xfs: cleanup xfs_iomap_eof_align_last_fsb Christoph Hellwig
2011-12-13 21:25   ` Dave Chinner
2011-12-08 15:57 ` [PATCH 03/11] xfs: remove the unused dm_attrs structure Christoph Hellwig
2011-12-08 15:57 ` [PATCH 04/11] xfs: remove the if_ext_max field in struct xfs_ifork Christoph Hellwig
2011-12-13 21:59   ` Dave Chinner
2011-12-08 15:58 ` [PATCH 05/11] xfs: make i_flags an unsigned long Christoph Hellwig
2012-01-13 19:07   ` Ben Myers
2012-01-24 18:03     ` Christoph Hellwig
2011-12-08 15:58 ` [PATCH 06/11] xfs: replace i_flock with a sleeping bitlock Christoph Hellwig
2011-12-13 22:19   ` Dave Chinner
2011-12-18 18:11     ` Christoph Hellwig
2011-12-08 15:58 ` [PATCH 07/11] xfs: replace i_pin_wait with a bit waitqueue Christoph Hellwig
2011-12-13 22:21   ` Dave Chinner
2011-12-08 15:58 ` [PATCH 08/11] xfs: remove the i_size field in struct xfs_inode Christoph Hellwig
2011-12-13 22:58   ` Dave Chinner
2011-12-08 15:58 ` [PATCH 09/11] xfs: remove the i_new_size " Christoph Hellwig
2011-12-13 23:16   ` Dave Chinner [this message]
2011-12-14 13:30     ` Christoph Hellwig
2011-12-08 15:58 ` [PATCH 10/11] xfs: always return with the iolock held from xfs_file_aio_write_checks Christoph Hellwig
2011-12-13 23:20   ` Dave Chinner
2011-12-14 13:27     ` Christoph Hellwig
2011-12-08 15:58 ` [PATCH 11/11] xfs: cleanup xfs_file_aio_write Christoph Hellwig
2011-12-13 23:28   ` Dave Chinner
  -- strict thread matches above, loose matches on Subject: below --
2011-12-18 20:00 [PATCH 00/11] inode shrink and misc updates V2 Christoph Hellwig
2011-12-18 20:00 ` [PATCH 09/11] xfs: remove the i_new_size field in struct xfs_inode Christoph Hellwig
2011-12-18 22:13   ` Dave Chinner
2012-01-16 22:41   ` Ben Myers
2012-01-17 20:14   ` Ben Myers

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=20111213231611.GG3179@dastard \
    --to=david@fromorbit.com \
    --cc=hch@infradead.org \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox