From: Tejun Heo <tj@kernel.org>
To: Christoph Lameter <cl@linux.com>
Cc: akpm@linux-foundation.org, Pekka Enberg <penberg@cs.helsinki.fi>,
linux-kernel@vger.kernel.org,
Eric Dumazet <eric.dumazet@gmail.com>,
"H. Peter Anvin" <hpa@zytor.com>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Subject: Re: [cpuops cmpxchg V2 5/5] cpuops: Use cmpxchg for xchg to avoid lock semantics
Date: Wed, 15 Dec 2010 17:47:38 +0100 [thread overview]
Message-ID: <4D08F12A.3090107@kernel.org> (raw)
In-Reply-To: <20101214162855.392020353@linux.com>
On 12/14/2010 05:28 PM, Christoph Lameter wrote:
> Use cmpxchg instead of xchg to realize this_cpu_xchg.
>
> xchg will cause LOCK overhead since LOCK is always implied but cmpxchg
> will not.
>
> Baselines:
>
> xchg() = 18 cycles (no segment prefix, LOCK semantics)
> __this_cpu_xchg = 1 cycle
>
> (simulated using this_cpu_read/write, two prefixes. Looks like the
> cpu can use loop optimization to get rid of most of the overhead)
>
> Cycles before:
>
> this_cpu_xchg = 37 cycles (segment prefix and LOCK (implied by xchg))
>
> After:
>
> this_cpu_xchg = 11 cycle (using cmpxchg without lock semantics)
>
> Signed-off-by: Christoph Lameter <cl@linux.com>
It's not a bad idea to keep this patch separate from the original one
but as both are not applied yet, it probably is better to put this
right after the original addition if you end up re-posting the series;
otherwise, I'll just reorder it when I apply.
Thanks.
--
tejun
prev parent reply other threads:[~2010-12-15 16:48 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-12-14 16:28 [cpuops cmpxchg V2 0/5] Cmpxchg and xchg operations Christoph Lameter
2010-12-14 16:28 ` [cpuops cmpxchg V2 1/5] percpu: Generic this_cpu_cmpxchg() and this_cpu_xchg support Christoph Lameter
2010-12-17 14:55 ` Tejun Heo
2010-12-14 16:28 ` [cpuops cmpxchg V2 2/5] x86: this_cpu_cmpxchg and this_cpu_xchg operations Christoph Lameter
2010-12-17 15:22 ` Tejun Heo
2010-12-14 16:28 ` [cpuops cmpxchg V2 3/5] irq_work: Use per cpu atomics instead of regular atomics Christoph Lameter
2010-12-15 16:32 ` Tejun Heo
2010-12-15 16:34 ` H. Peter Anvin
2010-12-15 16:50 ` Peter Zijlstra
2010-12-15 17:04 ` Christoph Lameter
2010-12-15 17:18 ` Peter Zijlstra
2010-12-15 17:31 ` H. Peter Anvin
2010-12-15 17:32 ` Christoph Lameter
2010-12-18 15:32 ` Tejun Heo
2010-12-14 16:28 ` [cpuops cmpxchg V2 4/5] vmstat: User per cpu atomics to avoid interrupt disable / enable Christoph Lameter
2010-12-15 16:45 ` Tejun Heo
2010-12-15 17:01 ` Christoph Lameter
2010-12-14 16:28 ` [cpuops cmpxchg V2 5/5] cpuops: Use cmpxchg for xchg to avoid lock semantics Christoph Lameter
2010-12-14 16:35 ` Mathieu Desnoyers
2010-12-14 16:44 ` Eric Dumazet
2010-12-14 16:55 ` Christoph Lameter
2010-12-14 17:00 ` H. Peter Anvin
2010-12-14 17:19 ` Christoph Lameter
2010-12-14 17:22 ` H. Peter Anvin
2010-12-14 17:29 ` Tejun Heo
2010-12-14 17:35 ` Christoph Lameter
2010-12-15 1:06 ` H. Peter Anvin
2010-12-15 16:29 ` Tejun Heo
2010-12-15 16:35 ` H. Peter Anvin
2010-12-15 16:39 ` Tejun Heo
2010-12-16 16:14 ` Tejun Heo
2010-12-16 18:13 ` x86: Use this_cpu_has for thermal_interrupt Christoph Lameter
2010-12-18 15:35 ` Tejun Heo
2010-12-21 0:56 ` H. Peter Anvin
2010-12-30 11:29 ` Tejun Heo
2010-12-30 18:19 ` H. Peter Anvin
2010-12-31 12:43 ` Tejun Heo
2010-12-16 18:14 ` x86: udelay: Use this_cpu_read to avoid address calculation Christoph Lameter
2010-12-16 18:15 ` gameport: use this_cpu_read instead of lookup Christoph Lameter
2010-12-18 15:34 ` Tejun Heo
2010-12-16 18:16 ` acpi throttling: Use this_cpu_has and simplify code Christoph Lameter
2010-12-18 15:50 ` Tejun Heo
2010-12-21 1:52 ` ykzhao
2010-12-21 22:43 ` Christoph Lameter
2010-12-21 4:28 ` Len Brown
2010-12-16 18:19 ` [cpuops cmpxchg V2 5/5] cpuops: Use cmpxchg for xchg to avoid lock semantics H. Peter Anvin
2010-12-16 18:55 ` Tejun Heo
2010-12-16 20:42 ` H. Peter Anvin
2010-12-15 16:47 ` Tejun Heo [this message]
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=4D08F12A.3090107@kernel.org \
--to=tj@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=cl@linux.com \
--cc=eric.dumazet@gmail.com \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--cc=penberg@cs.helsinki.fi \
/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