linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Alex Tomas <alex@clusterfs.com>
To: Eric Sandeen <sandeen@redhat.com>
Cc: Alex Tomas <alex@clusterfs.com>,
	sho@tnes.nec.co.jp, tytso@mit.edu, jack@suse.cz,
	linux-fsdevel@vger.kernel.org, linux-ext4@vger.kernel.org
Subject: Re: [RFC] Ext3 online defrag
Date: Fri, 27 Oct 2006 18:05:57 +0400	[thread overview]
Message-ID: <m364e5svju.fsf@bzzz.home.net> (raw)
In-Reply-To: <45420F76.90106@redhat.com> (Eric Sandeen's message of "Fri, 27 Oct 2006 08:53:58 -0500")

>>>>> Eric Sandeen (ES) writes:

 ES> Alex Tomas wrote:
 >> 3) scalable reservation
 >> required for delayed allocation to avoid -ENOSPC at flush time.
 >> current version uses per-sb spinlock.

 ES> Can you elaborate on this issue?  Shouldn't delayed allocation
 ES> decrement free space immediately, and only the actual block location
 ES> choice is delayed?  Or is this due to potential extra metadata space
 ES> required as blocks are allocated?

exactly. in this case, reservation has nothing to do with allocation
or preallocation of real blocks. this is just a *per-sb counter* of
blocks reserved for allocation at flush time. it includes all
non-allocated-yet blocks and metadata needed to allocate them (bitmaps,
group descriptors, blocks extent tree, etc). the previous version
of mballoc has reservation, but it doesn't scale very well being
a single global counter protected by the spinlock. at least, in many
regular loads I observed the reservation function in top30 of oprofile.



thanks, Alex

  reply	other threads:[~2006-10-27 14:05 UTC|newest]

Thread overview: 56+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-10-27  7:23 [RFC] Ext3 online defrag sho
2006-10-27  7:44 ` Alex Tomas
2006-10-27 13:53   ` Eric Sandeen
2006-10-27 14:05     ` Alex Tomas [this message]
2006-10-27 14:24       ` Eric Sandeen
2006-10-27 14:39         ` Alex Tomas
2006-11-15  9:54   ` Takashi Sato
     [not found] <20061023122710.GA12034@atrey.karlin.mff.cuni.cz>
2006-10-23 14:16 ` Theodore Tso
2006-10-23 14:31   ` Alex Tomas
2006-10-23 14:48     ` Andreas Dilger
2006-10-23 14:55       ` Jan Kara
2006-10-23 14:51     ` Jan Kara
2006-10-23 15:01     ` Eric Sandeen
2006-10-24  4:14     ` Jeff Garzik
2006-10-24 13:59       ` David Chinner
2006-10-24 14:51         ` Dave Kleikamp
2006-10-24 16:01           ` David Chinner
2006-10-24 16:26             ` Dave Kleikamp
2006-10-25  1:18               ` David Chinner
2006-10-25  2:30                 ` Barry Naujok
2006-10-25  2:42                   ` Jeff Garzik
2006-10-25  4:27                     ` David Chinner
2006-10-25  4:48                       ` Jeff Garzik
2006-10-25  5:38                         ` David Chinner
2006-10-25  6:01                           ` Jeff Garzik
2006-10-25  8:11                             ` David Chinner
2006-10-25 17:00                               ` Jeff Garzik
2006-10-26  1:40                                 ` David Chinner
2006-10-26  3:33                                   ` Theodore Tso
2006-10-26  6:36                                     ` David Chinner
2006-10-26 13:37                                       ` Theodore Tso
2006-10-26 14:40                                         ` Dave Kleikamp
2006-10-26 11:37                                   ` Jan Kara
2006-10-27  1:32                                     ` David Chinner
2006-10-24 14:52         ` Eric Sandeen
2006-10-24 19:44         ` Theodore Tso
2006-10-24 20:31           ` Russell Cattelan
2006-10-24 23:00           ` Andreas Dilger
2006-10-25 14:54             ` Jan Kara
2006-10-25 17:02               ` Jeff Garzik
2006-10-25 17:58                 ` Jan Kara
2006-10-25 18:08                   ` Jeff Garzik
2006-10-25 18:25                     ` Jan Kara
2006-10-25 18:33                       ` Jeff Garzik
2006-10-26  9:30               ` Andreas Dilger
2006-10-25  2:09           ` David Chinner
2006-10-23 14:45   ` Jan Kara
2006-10-23 15:14   ` Andreas Dilger
2006-10-23 16:03     ` Jan Kara
2006-10-23 17:29       ` Andreas Dilger
2006-10-25 18:36         ` Jan Kara
2006-10-25 18:41           ` Jeff Garzik
2006-10-26 15:25             ` Jörn Engel
2006-10-24  4:13 ` Jeff Garzik
2006-10-24  4:21 ` Chris Wedgwood
2006-10-24 10:09   ` Jan Kara

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=m364e5svju.fsf@bzzz.home.net \
    --to=alex@clusterfs.com \
    --cc=jack@suse.cz \
    --cc=linux-ext4@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=sandeen@redhat.com \
    --cc=sho@tnes.nec.co.jp \
    --cc=tytso@mit.edu \
    /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).