linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Pekka Enberg <penberg@cs.helsinki.fi>
To: "Zhang, Yanmin" <yanmin_zhang@linux.intel.com>
Cc: Christoph Lameter <cl@linux-foundation.org>,
	Nick Piggin <npiggin@suse.de>, Lin Ming <ming.m.lin@intel.com>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	Andrew Morton <akpm@linux-foundation.org>,
	Linus Torvalds <torvalds@linux-foundation.org>
Subject: Re: [patch] SLQB slab allocator
Date: Thu, 22 Jan 2009 11:33:12 +0200	[thread overview]
Message-ID: <1232616792.14549.19.camel@penberg-laptop> (raw)
In-Reply-To: <1232616638.11429.131.camel@ymzhang>

On Thu, 2009-01-22 at 17:30 +0800, Zhang, Yanmin wrote:
> On Thu, 2009-01-22 at 11:27 +0200, Pekka Enberg wrote:
> > Hi Christoph,
> > 
> > On Mon, 19 Jan 2009, Nick Piggin wrote:
> > > > > > You only go to the allocator when the percpu queue goes empty though, so
> > > > > > if memory policy changes (eg context switch or something), then subsequent
> > > > > > allocations will be of the wrong policy.
> > > > >
> > > > > The per cpu queue size in SLUB is limited by the queues only containing
> > > > > objects from the same page. If you have large queues like SLAB/SLQB(?)
> > > > > then this could be an issue.
> > > >
> > > > And it could be a problem in SLUB too. Chances are that several allocations
> > > > will be wrong after every policy switch. I could describe situations in which
> > > > SLUB will allocate with the _wrong_ policy literally 100% of the time.
> > 
> > On Wed, 2009-01-21 at 19:13 -0500, Christoph Lameter wrote:
> > > No it cannot because in SLUB objects must come from the same page.
> > > Multiple objects in a queue will only ever require a single page and not
> > > multiple like in SLAB.
> > 
> > There's one potential problem with "per-page queues", though. The bigger
> > the object, the smaller the "queue" (i.e. less objects per page). Also,
> > partial lists are less likely to help for big objects because they get
> > emptied so quickly and returned to the page allocator. Perhaps we should
> > do a small "full list" for caches with large objects?
> That helps definitely. We could use a batch to control the list size.

s/full list/empty list/g

That is, a list of pages that could be returned to the page allocator
but are pooled in SLUB to avoid the page allocator overhead. Note that
this will not help allocators that trigger page allocator pass-through.

		Pekka

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2009-01-22  9:33 UTC|newest]

Thread overview: 99+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-01-14  9:04 Nick Piggin
2009-01-14 10:53 ` Pekka Enberg
2009-01-14 11:47   ` Nick Piggin
2009-01-14 13:44     ` Pekka Enberg
2009-01-14 14:22       ` Nick Piggin
2009-01-14 14:45         ` Pekka Enberg
2009-01-14 15:09           ` Nick Piggin
2009-01-14 15:22             ` Nick Piggin
2009-01-14 15:30               ` Pekka Enberg
2009-01-14 15:59                 ` Nick Piggin
2009-01-14 18:40                   ` Christoph Lameter
2009-01-15  6:19                     ` Nick Piggin
2009-01-15 20:47                       ` Christoph Lameter
2009-01-16  3:43                         ` Nick Piggin
2009-01-16 21:25                           ` Christoph Lameter
2009-01-19  6:18                             ` Nick Piggin
2009-01-22  0:13                               ` Christoph Lameter
2009-01-22  9:27                                 ` Pekka Enberg
2009-01-22  9:30                                   ` Zhang, Yanmin
2009-01-22  9:33                                     ` Pekka Enberg [this message]
2009-01-23 15:32                                       ` Christoph Lameter
2009-01-23 15:37                                         ` Pekka Enberg
2009-01-23 15:42                                           ` Christoph Lameter
2009-01-23 15:32                                   ` Christoph Lameter
2009-01-23  4:09                                 ` Nick Piggin
2009-01-23 15:41                                   ` Christoph Lameter
2009-01-23 15:53                                     ` Nick Piggin
2009-01-26 17:28                                       ` Christoph Lameter
2009-02-03  1:53                                         ` Nick Piggin
2009-02-03 17:33                                           ` Christoph Lameter
2009-02-03 18:42                                             ` Pekka Enberg
2009-02-03 18:47                                               ` Pekka Enberg
2009-02-04  4:22                                                 ` Nick Piggin
2009-02-04 20:09                                                   ` Christoph Lameter
2009-02-05  3:18                                                     ` Nick Piggin
2009-02-04 20:10                                               ` Christoph Lameter
2009-02-05  3:14                                                 ` Nick Piggin
2009-02-04  4:07                                             ` Nick Piggin
2009-01-14 18:01             ` Christoph Lameter
2009-01-15  6:03               ` Nick Piggin
2009-01-15 20:05                 ` Christoph Lameter
2009-01-16  3:19                   ` Nick Piggin
2009-01-16 21:07                     ` Christoph Lameter
2009-01-19  5:47                       ` Nick Piggin
2009-01-22  0:19                         ` Christoph Lameter
2009-01-23  4:17                           ` Nick Piggin
2009-01-23 15:52                             ` Christoph Lameter
2009-01-23 16:10                               ` Nick Piggin
2009-01-23 17:09                                 ` Nick Piggin
2009-01-26 17:46                                   ` Christoph Lameter
2009-02-03  1:42                                     ` Nick Piggin
2009-01-26 17:34                                 ` Christoph Lameter
2009-02-03  1:48                                   ` Nick Piggin
2009-01-21 14:30 Nick Piggin
2009-01-21 14:59 ` Ingo Molnar
2009-01-21 15:17   ` Nick Piggin
2009-01-21 16:56   ` Nick Piggin
2009-01-21 17:40     ` Ingo Molnar
2009-01-23  3:31       ` Nick Piggin
2009-01-23  6:14       ` Nick Piggin
2009-01-23 12:56         ` Ingo Molnar
2009-01-21 17:59 ` Joe Perches
2009-01-23  3:35   ` Nick Piggin
2009-01-23  4:00     ` Joe Perches
2009-01-21 18:10 ` Hugh Dickins
2009-01-22 10:01   ` Pekka Enberg
2009-01-22 12:47     ` Hugh Dickins
2009-01-23 14:23       ` Hugh Dickins
2009-01-23 14:30         ` Pekka Enberg
2009-02-02  3:38         ` Zhang, Yanmin
2009-02-02  9:00           ` Pekka Enberg
2009-02-02 15:00             ` Christoph Lameter
2009-02-03  1:34               ` Zhang, Yanmin
2009-02-03  7:29             ` Zhang, Yanmin
2009-02-03 12:18               ` Hugh Dickins
2009-02-04  2:21                 ` Zhang, Yanmin
2009-02-05 19:04                   ` Hugh Dickins
2009-02-06  0:47                     ` Zhang, Yanmin
2009-02-06  8:57                     ` Pekka Enberg
2009-02-06 12:33                       ` Hugh Dickins
2009-02-10  8:56                         ` Zhang, Yanmin
2009-02-02 11:50           ` Hugh Dickins
2009-01-23  3:55   ` Nick Piggin
2009-01-23 13:57     ` Hugh Dickins
2009-01-22  8:45 ` Zhang, Yanmin
2009-01-23  3:57   ` Nick Piggin
2009-01-23  9:00   ` Nick Piggin
2009-01-23 13:34     ` Hugh Dickins
2009-01-23 13:44       ` Nick Piggin
2009-01-23  9:55 ` Andi Kleen
2009-01-23 10:13   ` Pekka Enberg
2009-01-23 11:25   ` Nick Piggin
2009-01-23 11:57     ` Andi Kleen
2009-01-23 13:18       ` Nick Piggin
2009-01-23 14:04         ` Andi Kleen
2009-01-23 14:27           ` Nick Piggin
2009-01-23 15:06             ` Andi Kleen
2009-01-23 15:15               ` Nick Piggin
2009-01-23 12:55   ` Nick Piggin

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=1232616792.14549.19.camel@penberg-laptop \
    --to=penberg@cs.helsinki.fi \
    --cc=akpm@linux-foundation.org \
    --cc=cl@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=ming.m.lin@intel.com \
    --cc=npiggin@suse.de \
    --cc=torvalds@linux-foundation.org \
    --cc=yanmin_zhang@linux.intel.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