From: Jeremy Fitzhardinge <jeremy@goop.org>
To: Christoph Lameter <cl@linux.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>,
Linus Torvalds <torvalds@linux-foundation.org>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@elte.hu>,
the arch/x86 maintainers <x86@kernel.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Nick Piggin <npiggin@kernel.dk>,
Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Subject: Re: [PATCH 13/15] x86: add cmpxchg_flag() variant
Date: Wed, 24 Aug 2011 09:33:06 -0700 [thread overview]
Message-ID: <4E5527C2.8000801@goop.org> (raw)
In-Reply-To: <alpine.DEB.2.00.1108240852370.24118@router.home>
On 08/24/2011 06:53 AM, Christoph Lameter wrote:
> On Tue, 23 Aug 2011, Jeremy Fitzhardinge wrote:
>
>> However, having prototyped it, I dunno, it doesn't really seem like much
>> of a win for all the extra code it adds. I just can't get too excited
>> about an extra test instruction adjacent to a monster like a locked
>> cmpxchg. The jump variant avoids the test, but gcc still generates some
>> pretty bogus stuff:
> There are also unlocked cmpxchges in use.
I only looked in arch/x86, but I didn't find any that were
straightforward candidates for cmpxchg_flag.
> And if the cmpxchg is a 16 byte
> cmpxchg (cmpxchg_double) then the comparison is getting more expensive.
We're talking about the difference between cmpxchg_flag() - which does
a sete based on the flags set up cmpxchg - and a variant based on "asm
goto" which could, in principle, avoid the need for sete by allowing a
control flow statment to directly use the asm goto's conditional jump.
The performance of both is invariant wrt the cmpxchg argument size.
J
next prev parent reply other threads:[~2011-08-24 16:33 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-08-22 23:14 [PATCH 00/12] x86: Ticket lock cleanup Jeremy Fitzhardinge
[not found] ` <cover.1314054734.git.jeremy.fitzhardinge@citrix.com>
2011-08-22 23:14 ` [PATCH 01/15] x86/ticketlock: clean up types and accessors Jeremy Fitzhardinge
2011-08-22 23:14 ` [PATCH 02/15] x86/ticketlock: convert spin loop to C Jeremy Fitzhardinge
2011-08-22 23:14 ` [PATCH 03/15] x86/ticketlock: Use C for __ticket_spin_unlock Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 04/15] x86/ticketlock: make large and small ticket versions of spin_lock the same Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 05/15] x86/ticketlock: make __ticket_spin_lock common Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 06/15] x86/ticketlock: make __ticket_spin_trylock common Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 07/15] x86: add xadd helper macro Jeremy Fitzhardinge
2011-08-22 23:29 ` H. Peter Anvin
2011-08-22 23:43 ` Jeremy Fitzhardinge
2011-08-23 4:41 ` H. Peter Anvin
2011-08-23 10:59 ` Peter Zijlstra
2011-08-23 16:34 ` Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 08/15] x86/ticketlock: use xadd helper Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 09/15] x86/cmpxchg: linux/alternative.h has LOCK_PREFIX Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 10/15] x86/cmpxchg: move 32-bit __cmpxchg_wrong_size to match 64 bit Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 11/15] x86/cmpxchg: move 64-bit set64_bit() to match 32-bit Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 12/15] x86/cmpxchg: unify cmpxchg into cmpxchg.h Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 13/15] x86: add cmpxchg_flag() variant Jeremy Fitzhardinge
2011-08-23 19:01 ` Christoph Lameter
2011-08-23 19:22 ` H. Peter Anvin
2011-08-23 19:52 ` Christoph Lameter
2011-08-23 21:03 ` H. Peter Anvin
2011-08-23 19:53 ` Jeremy Fitzhardinge
2011-08-23 20:45 ` H. Peter Anvin
2011-08-23 22:15 ` Jeremy Fitzhardinge
2011-08-23 22:43 ` H. Peter Anvin
2011-08-24 13:54 ` Christoph Lameter
2011-08-24 13:53 ` Christoph Lameter
2011-08-24 16:33 ` Jeremy Fitzhardinge [this message]
2011-08-24 19:27 ` Christoph Lameter
2011-08-24 20:15 ` Jeremy Fitzhardinge
2011-08-24 20:31 ` H. Peter Anvin
2011-08-24 20:38 ` Jeremy Fitzhardinge
2011-08-24 23:04 ` H. Peter Anvin
2011-08-24 23:11 ` Linus Torvalds
2011-08-24 23:19 ` Jeremy Fitzhardinge
2011-08-25 14:07 ` Christoph Lameter
2011-08-22 23:15 ` [PATCH 14/15] x86/ticketlocks: use cmpxchg_flag for trylock Jeremy Fitzhardinge
2011-08-22 23:15 ` [PATCH 15/15] x86: use cmpxchg_flag() where applicable Jeremy Fitzhardinge
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=4E5527C2.8000801@goop.org \
--to=jeremy@goop.org \
--cc=cl@linux.com \
--cc=hpa@zytor.com \
--cc=jeremy.fitzhardinge@citrix.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=npiggin@kernel.dk \
--cc=peterz@infradead.org \
--cc=torvalds@linux-foundation.org \
--cc=x86@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox