From owner-freebsd-arch Tue Jun 20 13:25:51 2000 Delivered-To: freebsd-arch@freebsd.org Received: from alcanet.com.au (mail.alcanet.com.au [203.62.196.10]) by hub.freebsd.org (Postfix) with ESMTP id 2D45037C0D4 for ; Tue, 20 Jun 2000 13:25:37 -0700 (PDT) (envelope-from jeremyp@pc0640.alcatel.com.au) Received: by border.alcanet.com.au id <115222>; Wed, 21 Jun 2000 06:25:22 +1000 Date: Tue, 20 Jun 2000 19:58:52 +1000 From: Peter Jeremy Subject: Re: Preemptive kernel on older X86 hardware In-reply-to: <20000601223101.A17391@sharmas.dhs.org>; from adsharma@sharmas.dhs.org on Fri, Jun 02, 2000 at 03:31:46PM +1000 To: Arun Sharma Cc: arch@FreeBSD.ORG Message-Id: <00Jun21.062522est.115222@border.alcanet.com.au> MIME-version: 1.0 X-Mailer: Mutt 1.0i Content-type: text/plain; charset=us-ascii References: <200005250208.TAA78220@apollo.backplane.com> <82645.959243483@localhost> <20000601223101.A17391@sharmas.dhs.org> Sender: owner-freebsd-arch@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Digging up some old e-mail... On Fri, Jun 02, 2000 at 03:31:46PM +1000, Arun Sharma wrote: >And self modifying code also isn't exactly cheap either on that >architecture. My understanding of the proposal is that the lock reference would be compiled as a function call and some padding. The first time that the code is executed, it will be updated in-place with the optimum code for the UP/SMP/processor configuration. This first execution _is_ quite expensive - function call, code update and various cache flushes - but subsequent executions are optimal inline code. Peter To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message