From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 16537 invoked by alias); 16 Mar 2004 17:25:52 -0000 Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-owner@gcc.gnu.org Received: (qmail 16523 invoked from network); 16 Mar 2004 17:25:51 -0000 Received: from unknown (HELO sadr.equallogic.com) (66.155.203.134) by sources.redhat.com with SMTP; 16 Mar 2004 17:25:51 -0000 Received: from sadr.equallogic.com (localhost.localdomain [127.0.0.1]) by sadr.equallogic.com (8.12.8/8.12.8) with ESMTP id i2GHPmPx015441 for ; Tue, 16 Mar 2004 12:25:48 -0500 Received: from deneb.dev.equallogic.com (deneb [172.16.1.99]) by sadr.equallogic.com (8.12.8/8.12.8) with ESMTP id i2GHPmM5015436; Tue, 16 Mar 2004 12:25:48 -0500 Received: from pkoning.equallogic.com (localhost.localdomain [127.0.0.1]) by deneb.dev.equallogic.com (8.11.6/8.11.6) with ESMTP id i2GHPjl15389; Tue, 16 Mar 2004 12:25:46 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <16471.14489.831595.320450@gargle.gargle.HOWL> Date: Tue, 16 Mar 2004 17:25:00 -0000 From: Paul Koning To: zack@codesourcery.com Cc: ian@wasabisystems.com, gcc@gcc.gnu.org Subject: Re: GCC Status Report (2004-03-09) References: <200403091809.i29I9P04020607@sirius.codesourcery.com> <200403111030.14616.ebotcazou@act-europe.fr> <40573108.4080602@codesourcery.com> <16471.12911.55000.733792@gargle.gargle.HOWL> <87oeqw7ljg.fsf@egil.codesourcery.com> X-SW-Source: 2004-03/txt/msg00833.txt.bz2 >>>>> "Zack" == Zack Weinberg writes: Zack> Ian Lance Taylor writes: >> Paul Koning writes: >>> Perhaps we need asm("":::"barrier")? >> Seconded. For high performance device manipulation, it would be >> good to have a way to say "do not schedule instructions across >> this point" without also saying "clobber everything". Zack> __builtin_fence() (possibly with an argument to request various Zack> sorts of memory barriers that various architectures have) seems Zack> preferable to me over asm()... Good point. You'd want to be able to ask for memory barrier instructions, or not -- in some cases all you need is to block scheduler actions, and the ordering rules of the target machine are otherwise sufficient for the purpose. paul