All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dave Chinner <david@fromorbit.com>
To: Alex Elder <aelder@sgi.com>
Cc: xfs@oss.sgi.com
Subject: Re: [PATCH 20/34] xfs: remove all the inodes on a buffer from the AIL in bulk
Date: Wed, 22 Dec 2010 14:49:02 +1100	[thread overview]
Message-ID: <20101222034902.GF4907@dastard> (raw)
In-Reply-To: <1292984446.2408.358.camel@doink>

On Tue, Dec 21, 2010 at 08:20:46PM -0600, Alex Elder wrote:
> On Tue, 2010-12-21 at 18:29 +1100, Dave Chinner wrote:
> > From: Dave Chinner <dchinner@redhat.com>
> > 
> > When inode buffer IO completes, usually all of the inodes are removed from the
> > AIL. This involves processing them one at a time and taking the AIL lock once
> > for every inode. When all CPUs are processing inode IO completions, this causes
> > excessive amount sof contention on the AIL lock.
> > 
> > Instead, change the way we process inode IO completion in the buffer
> > IO done callback. Allow the inode IO done callback to walk the list
> > of IO done callbacks and pull all the inodes off the buffer in one
> > go and then process them as a batch.
> > 
> > Once all the inodes for removal are collected, take the AIL lock
> > once and do a bulk removal operation to minimise traffic on the AIL
> > lock.
> > 
> > Signed-off-by: Dave Chinner <dchinner@redhat.com>
> > Reviewed-by: Christoph Hellwig <hch@lst.de>
> 
> One question, below.		-Alex
> 
> . . .
> 
> > @@ -861,28 +910,37 @@ xfs_iflush_done(
> >  	 * the lock since it's cheaper, and then we recheck while
> >  	 * holding the lock before removing the inode from the AIL.
> >  	 */
> > -	if (iip->ili_logged && lip->li_lsn == iip->ili_flush_lsn) {
> > +	if (need_ail) {
> > +		struct xfs_log_item *log_items[need_ail];
> 
> What's the worst-case value of need_ail we might see here?

The number of inodes in a cluster. That's 32 for 256 byte inodes
with the current 8k cluster size.

Cheers,

Dave
-- 
Dave Chinner
david@fromorbit.com

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

  reply	other threads:[~2010-12-22  3:47 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-12-21  7:28 [PATCH 00/34] xfs: scalability patchset for 2.6.38 Dave Chinner
2010-12-21  7:28 ` [PATCH 01/34] xfs: provide a inode iolock lockdep class Dave Chinner
2010-12-21 15:15   ` Christoph Hellwig
2010-12-21  7:28 ` [PATCH 02/34] xfs: use KM_NOFS for allocations during attribute list operations Dave Chinner
2010-12-21 15:16   ` Christoph Hellwig
2010-12-21  7:28 ` [PATCH 03/34] lib: percpu counter add unless less than functionality Dave Chinner
2010-12-22  2:20   ` Alex Elder
2010-12-22  3:46     ` Dave Chinner
2010-12-21  7:29 ` [PATCH 04/34] xfs: use generic per-cpu counter infrastructure Dave Chinner
2010-12-21  7:29 ` [PATCH 05/34] xfs: demultiplex xfs_icsb_modify_counters() Dave Chinner
2010-12-21  7:29 ` [PATCH 06/34] xfs: dynamic speculative EOF preallocation Dave Chinner
2010-12-21 15:15   ` Christoph Hellwig
2010-12-21 21:42     ` Dave Chinner
2010-12-21 23:44       ` Dave Chinner
2010-12-22  2:29         ` Alex Elder
2010-12-29 12:56         ` Christoph Hellwig
2010-12-27 14:57   ` Alex Elder
2010-12-27 15:00   ` Alex Elder
2011-01-06 18:16   ` Christoph Hellwig
2010-12-21  7:29 ` [PATCH 07/34] xfs: don't truncate prealloc from frequently accessed inodes Dave Chinner
2010-12-21 16:45   ` Christoph Hellwig
2010-12-21  7:29 ` [PATCH 08/34] xfs: rcu free inodes Dave Chinner
2010-12-21  7:29 ` [PATCH 09/34] xfs: convert inode cache lookups to use RCU locking Dave Chinner
2010-12-21  7:29 ` [PATCH 10/34] xfs: convert pag_ici_lock to a spin lock Dave Chinner
2010-12-21  7:29 ` [PATCH 11/34] xfs: convert xfsbud shrinker to a per-buftarg shrinker Dave Chinner
2010-12-21  7:29 ` [PATCH 12/34] xfs: add a lru to the XFS buffer cache Dave Chinner
2010-12-21  7:29 ` [PATCH 13/34] xfs: connect up buffer reclaim priority hooks Dave Chinner
2010-12-21  7:29 ` [PATCH 14/34] xfs: fix EFI transaction cancellation Dave Chinner
2010-12-21  7:29 ` [PATCH 15/34] xfs: Pull EFI/EFD handling out from under the AIL lock Dave Chinner
2010-12-21  7:29 ` [PATCH 16/34] xfs: clean up xfs_ail_delete() Dave Chinner
2010-12-21  7:29 ` [PATCH 17/34] xfs: bulk AIL insertion during transaction commit Dave Chinner
2010-12-21  7:29 ` [PATCH 18/34] xfs: reduce the number of AIL push wakeups Dave Chinner
2010-12-21  7:29 ` [PATCH 19/34] xfs: consume iodone callback items on buffers as they are processed Dave Chinner
2010-12-21  7:29 ` [PATCH 20/34] xfs: remove all the inodes on a buffer from the AIL in bulk Dave Chinner
2010-12-22  2:20   ` Alex Elder
2010-12-22  3:49     ` Dave Chinner [this message]
2010-12-21  7:29 ` [PATCH 22/34] xfs: use AIL bulk delete function to implement single delete Dave Chinner
2010-12-21  7:29 ` [PATCH 23/34] xfs: convert log grant ticket queues to list heads Dave Chinner
2010-12-21  7:29 ` [PATCH 24/34] xfs: fact out common grant head/log tail verification code Dave Chinner
2010-12-21  7:29 ` [PATCH 25/34] xfs: rework log grant space calculations Dave Chinner
2010-12-21  7:29 ` [PATCH 26/34] xfs: combine grant heads into a single 64 bit integer Dave Chinner
2010-12-21  7:29 ` [PATCH 27/34] xfs: use wait queues directly for the log wait queues Dave Chinner
2010-12-21  7:29 ` [PATCH 28/34] xfs: make AIL tail pushing independent of the grant lock Dave Chinner
2010-12-21  7:29 ` [PATCH 29/34] xfs: convert l_last_sync_lsn to an atomic variable Dave Chinner
2010-12-21  7:29 ` [PATCH 30/34] xfs: convert l_tail_lsn " Dave Chinner
2010-12-29 12:52   ` Christoph Hellwig
2010-12-29 15:49   ` Alex Elder
2010-12-21  7:29 ` [PATCH 31/34] xfs: convert log grant heads to atomic variables Dave Chinner
2010-12-21  7:29 ` [PATCH 32/34] xfs: introduce new locks for the log grant ticket wait queues Dave Chinner
2010-12-21  7:29 ` [PATCH 33/34] xfs: convert grant head manipulations to lockless algorithm Dave Chinner
2010-12-21  7:29 ` [PATCH 34/34] xfs: kill useless spinlock_destroy macro Dave Chinner
2010-12-23  1:15 ` [PATCH 00/34] xfs: scalability patchset for 2.6.38 Dave Chinner

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=20101222034902.GF4907@dastard \
    --to=david@fromorbit.com \
    --cc=aelder@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.