All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Zhang, Yanmin" <yanmin_zhang@linux.intel.com>
To: Christoph Lameter <cl@linux-foundation.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
	"hugh.dickins@tiscali.co.uk" <hugh.dickins@tiscali.co.uk>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	akpm@linux-foundation.org, Tejun Heo <tj@kernel.org>,
	Andi Kleen <andi@firstfloor.org>
Subject: Re: [MM] Make mm counters per cpu instead of atomic
Date: Tue, 24 Nov 2009 16:02:33 +0800	[thread overview]
Message-ID: <1259049753.29789.49.camel@localhost> (raw)
In-Reply-To: <alpine.DEB.2.00.0911230830300.26432@router.home>

On Mon, 2009-11-23 at 08:31 -0600, Christoph Lameter wrote:
> On Mon, 23 Nov 2009, Zhang, Yanmin wrote:
> 
> > Another theoretic issue is below scenario:
> > Process A get the read lock on cpu 0 and is scheduled to cpu 2 to unlock. Then
> > it's scheduled back to cpu 0 to repeat the step. eventually, the reader counter
> > will overflow. Considering multiple thread cases, it might be faster to
> > overflow than what we imagine. When it overflows, processes will hang there.
> 
> True.... We need to find some alternative to per cpu data to scale mmap
> sem then.
I ran lots of benchmarks such like specjbb2005/hackbench/tbench/dbench/iozone
/sysbench_oltp(mysql)/aim7 against percpu tree(based on 2.6.32-rc7) on a 4*8*2 logical
cpu machine, and didn't find big result difference between with your patch and without
your patch.




WARNING: multiple messages have this Message-ID (diff)
From: "Zhang, Yanmin" <yanmin_zhang@linux.intel.com>
To: Christoph Lameter <cl@linux-foundation.org>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>,
	"hugh.dickins@tiscali.co.uk" <hugh.dickins@tiscali.co.uk>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	akpm@linux-foundation.org, Tejun Heo <tj@kernel.org>,
	Andi Kleen <andi@firstfloor.org>
Subject: Re: [MM] Make mm counters per cpu instead of atomic
Date: Tue, 24 Nov 2009 16:02:33 +0800	[thread overview]
Message-ID: <1259049753.29789.49.camel@localhost> (raw)
In-Reply-To: <alpine.DEB.2.00.0911230830300.26432@router.home>

On Mon, 2009-11-23 at 08:31 -0600, Christoph Lameter wrote:
> On Mon, 23 Nov 2009, Zhang, Yanmin wrote:
> 
> > Another theoretic issue is below scenario:
> > Process A get the read lock on cpu 0 and is scheduled to cpu 2 to unlock. Then
> > it's scheduled back to cpu 0 to repeat the step. eventually, the reader counter
> > will overflow. Considering multiple thread cases, it might be faster to
> > overflow than what we imagine. When it overflows, processes will hang there.
> 
> True.... We need to find some alternative to per cpu data to scale mmap
> sem then.
I ran lots of benchmarks such like specjbb2005/hackbench/tbench/dbench/iozone
/sysbench_oltp(mysql)/aim7 against percpu tree(based on 2.6.32-rc7) on a 4*8*2 logical
cpu machine, and didn't find big result difference between with your patch and without
your patch.



--
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:[~2009-11-24  8:00 UTC|newest]

Thread overview: 66+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-11-04 19:14 [MM] Make mm counters per cpu instead of atomic Christoph Lameter
2009-11-04 19:14 ` Christoph Lameter
2009-11-04 19:17 ` [MM] Remove rss batching from copy_page_range() Christoph Lameter
2009-11-04 19:17   ` Christoph Lameter
2009-11-04 21:02   ` Andi Kleen
2009-11-04 21:02     ` Andi Kleen
2009-11-04 22:02     ` Christoph Lameter
2009-11-04 22:02       ` Christoph Lameter
2009-11-05  8:27       ` Andi Kleen
2009-11-05  8:27         ` Andi Kleen
2009-11-04 21:01 ` [MM] Make mm counters per cpu instead of atomic Andi Kleen
2009-11-04 21:01   ` Andi Kleen
2009-11-04 23:49 ` Dave Jones
2009-11-04 23:49   ` Dave Jones
2009-11-05 15:04   ` Christoph Lameter
2009-11-05 15:04     ` Christoph Lameter
2009-11-05 15:36     ` [MM] Make mm counters per cpu instead of atomic V2 Christoph Lameter
2009-11-05 15:36       ` Christoph Lameter
2009-11-06  1:11       ` KAMEZAWA Hiroyuki
2009-11-06  1:11         ` KAMEZAWA Hiroyuki
2009-11-06  3:23         ` KAMEZAWA Hiroyuki
2009-11-06  3:23           ` KAMEZAWA Hiroyuki
2009-11-06 17:32           ` Christoph Lameter
2009-11-06 17:32             ` Christoph Lameter
2009-11-06 19:03             ` KAMEZAWA Hiroyuki
2009-11-06 19:03               ` KAMEZAWA Hiroyuki
2009-11-06 19:13               ` Christoph Lameter
2009-11-06 19:13                 ` Christoph Lameter
2009-11-06 19:20                 ` KAMEZAWA Hiroyuki
2009-11-06 19:20                   ` KAMEZAWA Hiroyuki
2009-11-06 19:47                   ` Christoph Lameter
2009-11-06 19:47                     ` Christoph Lameter
2009-11-10 22:44         ` Andrew Morton
2009-11-10 22:44           ` Andrew Morton
2009-11-10 23:20           ` Christoph Lameter
2009-11-10 23:20             ` Christoph Lameter
2009-11-06  4:08       ` KAMEZAWA Hiroyuki
2009-11-06  4:08         ` KAMEZAWA Hiroyuki
2009-11-06  4:15       ` KAMEZAWA Hiroyuki
2009-11-06  4:15         ` KAMEZAWA Hiroyuki
2009-11-05  1:16 ` [MM] Make mm counters per cpu instead of atomic KAMEZAWA Hiroyuki
2009-11-05  1:16   ` KAMEZAWA Hiroyuki
2009-11-05 15:10   ` Christoph Lameter
2009-11-05 15:10     ` Christoph Lameter
2009-11-05 23:42     ` KAMEZAWA Hiroyuki
2009-11-05 23:42       ` KAMEZAWA Hiroyuki
2009-11-17  6:48 ` Zhang, Yanmin
2009-11-17  6:48   ` Zhang, Yanmin
2009-11-17  7:31   ` Zhang, Yanmin
2009-11-17  7:31     ` Zhang, Yanmin
2009-11-17  9:34     ` Zhang, Yanmin
2009-11-17  9:34       ` Zhang, Yanmin
2009-11-17 17:25       ` Christoph Lameter
2009-11-17 17:25         ` Christoph Lameter
2009-11-19  0:48         ` Zhang, Yanmin
2009-11-19  0:48           ` Zhang, Yanmin
2009-11-23  8:51         ` Zhang, Yanmin
2009-11-23  8:51           ` Zhang, Yanmin
2009-11-23 14:31           ` Christoph Lameter
2009-11-23 14:31             ` Christoph Lameter
2009-11-24  8:02             ` Zhang, Yanmin [this message]
2009-11-24  8:02               ` Zhang, Yanmin
2009-11-24 15:17               ` Christoph Lameter
2009-11-24 15:17                 ` Christoph Lameter
2009-11-25  1:23                 ` Zhang, Yanmin
2009-11-25  1:23                   ` Zhang, Yanmin

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=1259049753.29789.49.camel@localhost \
    --to=yanmin_zhang@linux.intel.com \
    --cc=akpm@linux-foundation.org \
    --cc=andi@firstfloor.org \
    --cc=cl@linux-foundation.org \
    --cc=hugh.dickins@tiscali.co.uk \
    --cc=kamezawa.hiroyu@jp.fujitsu.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=tj@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.