From mboxrd@z Thu Jan  1 00:00:00 1970
X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/7385
Path: news.gmane.org!not-for-mail
From: Rich Felker <dalias@libc.org>
Newsgroups: gmane.linux.lib.musl.general
Subject: Re: [PATCH] Use CAS instead of atomic swap to implement
 spinlock
Date: Tue, 14 Apr 2015 19:16:27 -0400
Message-ID: <20150414231627.GR6817@brightrain.aerifal.cx>
References: <1429051493-2821-1-git-send-email-amonakov@ispras.ru>
Reply-To: musl@lists.openwall.com
NNTP-Posting-Host: plane.gmane.org
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Trace: ger.gmane.org 1429053404 19101 80.91.229.3 (14 Apr 2015 23:16:44 GMT)
X-Complaints-To: usenet@ger.gmane.org
NNTP-Posting-Date: Tue, 14 Apr 2015 23:16:44 +0000 (UTC)
To: musl@lists.openwall.com
Original-X-From: musl-return-7398-gllmg-musl=m.gmane.org@lists.openwall.com Wed Apr 15 01:16:43 2015
Return-path: <musl-return-7398-gllmg-musl=m.gmane.org@lists.openwall.com>
Envelope-to: gllmg-musl@m.gmane.org
Original-Received: from mother.openwall.net ([195.42.179.200])
	by plane.gmane.org with smtp (Exim 4.69)
	(envelope-from <musl-return-7398-gllmg-musl=m.gmane.org@lists.openwall.com>)
	id 1YiA4Q-0007qy-Ad
	for gllmg-musl@m.gmane.org; Wed, 15 Apr 2015 01:16:42 +0200
Original-Received: (qmail 1603 invoked by uid 550); 14 Apr 2015 23:16:40 -0000
Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm
Precedence: bulk
List-Post: <mailto:musl@lists.openwall.com>
List-Help: <mailto:musl-help@lists.openwall.com>
List-Unsubscribe: <mailto:musl-unsubscribe@lists.openwall.com>
List-Subscribe: <mailto:musl-subscribe@lists.openwall.com>
Original-Received: (qmail 1582 invoked from network); 14 Apr 2015 23:16:40 -0000
Content-Disposition: inline
In-Reply-To: <1429051493-2821-1-git-send-email-amonakov@ispras.ru>
User-Agent: Mutt/1.5.21 (2010-09-15)
Original-Sender: Rich Felker <dalias@aerifal.cx>
Xref: news.gmane.org gmane.linux.lib.musl.general:7385
Archived-At: <http://permalink.gmane.org/gmane.linux.lib.musl.general/7385>

On Wed, Apr 15, 2015 at 01:44:53AM +0300, Alexander Monakov wrote:
> This should allow spinning without constantly dirtying cache lines holding the
> spinlock value.  On architectures without native atomic swap, musl implement
> a_swap by looping around a_cas.

It won't help on x86, where all lock-prefixed ops perform both a read
and a write even if they fail. But it should help in ll/sc style
archs, and shouldn't hurt on x86.

Rich