From owner-cvs-all Tue Jun 26 17:56:53 2001 Delivered-To: cvs-all@freebsd.org Received: from netau1.alcanet.com.au (ntp.alcanet.com.au [203.62.196.27]) by hub.freebsd.org (Postfix) with ESMTP id 2F11337B405; Tue, 26 Jun 2001 17:56:47 -0700 (PDT) (envelope-from jeremyp@gsmx07.alcatel.com.au) Received: from mfg1.cim.alcatel.com.au (mfg1.cim.alcatel.com.au [139.188.23.1]) by netau1.alcanet.com.au (8.9.3 (PHNE_22672)/8.9.3) with ESMTP id KAA25911; Wed, 27 Jun 2001 10:56:35 +1000 (EST) Received: from gsmx07.alcatel.com.au by cim.alcatel.com.au (PMDF V5.2-32 #37645) with ESMTP id <01K59C2EG5M8VNZ1K1@cim.alcatel.com.au>; Wed, 27 Jun 2001 10:56:25 +1000 Received: (from jeremyp@localhost) by gsmx07.alcatel.com.au (8.11.1/8.11.1) id f5R0uUD94300; Wed, 27 Jun 2001 10:56:31 +1000 (EST envelope-from jeremyp) Content-return: prohibited Date: Wed, 27 Jun 2001 10:56:30 +1000 From: Peter Jeremy Subject: Re: Inline optimized bzero (was Re: cvs commit: src/sys/netinet tcp_subr.c) In-reply-to: <200106241549.f5OFn6J78347@earth.backplane.com>; from dillon@earth.backplane.com on Sun, Jun 24, 2001 at 08:49:06AM -0700 To: Matt Dillon Cc: Bruce Evans , cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Mail-Followup-To: Matt Dillon , Bruce Evans , cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Message-id: <20010627105630.B93007@gsmx07.alcatel.com.au> MIME-version: 1.0 Content-type: text/plain; charset=us-ascii Content-disposition: inline User-Agent: Mutt/1.2.5i References: <200106241549.f5OFn6J78347@earth.backplane.com> Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On 2001-Jun-24 08:49:06 -0700, Matt Dillon wrote: > I suppose the bzero() inline could implemented in the machine-dependant > code section. It seems a shame to waste mostly portable code though. > I'll mess around with it a bit. Anecdotal evidence (which isn't very trustworthy) suggests that a Unix kernel spends a significant amount of time copying or zeroing bytes. If this is true, it would be worthwhile having tweaked MD routines for bcopy() and bzero(). One advantage being that at least some architectures have instructions which are designed to copy or zero large blocks of data without totally wiping L1 (and maybe L2) cache. Peter To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message