From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S934688Ab0KQJer (ORCPT ); Wed, 17 Nov 2010 04:34:47 -0500 Received: from vpn.id2.novell.com ([195.33.99.129]:33347 "EHLO vpn.id2.novell.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S934588Ab0KQJep convert rfc822-to-8bit (ORCPT ); Wed, 17 Nov 2010 04:34:45 -0500 Message-Id: <4CE3AFC00200007800022B85@vpn.id2.novell.com> X-Mailer: Novell GroupWise Internet Agent 8.0.1 Date: Wed, 17 Nov 2010 09:34:40 +0000 From: "Jan Beulich" To: "Jeremy Fitzhardinge" Cc: "Jeremy Fitzhardinge" , "Eric Dumazet" , , "Peter Zijlstra" , "Nick Piggin" , "Srivatsa Vaddagiri" , "Linux Virtualization" , "Xen-devel" , "Mathieu Desnoyers" , "Avi Kivity" , "Linux Kernel Mailing List" , "H. Peter Anvin" Subject: Re: [PATCH 13/14] x86/ticketlock: add slowpath logic References: <97ed99ae9160bdb6477284b333bd6708fb7a19cb.1289940821.git.jeremy.fitzhardinge@citrix.com> <4CE3A1060200007800022B04@vpn.id2.novell.com> <4CE397C8.9040901@goop.org> <4CE398A0.3040105@goop.org> <4CE39B7A.1080706@goop.org> In-Reply-To: <4CE39B7A.1080706@goop.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 8BIT Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org >>> On 17.11.10 at 10:08, Jeremy Fitzhardinge wrote: > On 11/17/2010 12:56 AM, Jeremy Fitzhardinge wrote: >> On 11/17/2010 12:52 AM, Jeremy Fitzhardinge wrote: >>> But, yes, %z0 sounds interesting. Is it documented anywhere? I think >>> I've tried to use it in the past and run into gcc bugs. >> This one: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39590 >> >> Should be OK in this case because there's no 64-bit values to be seen... > Hm, it fails when __ticket_t is 16 bits: > > /home/jeremy/git/linux/arch/x86/include/asm/spinlock.h: Assembler messages: > /home/jeremy/git/linux/arch/x86/include/asm/spinlock.h:73: Error: suffix > or operands invalid for `or' > > lock; ors $1, 2(%rbx) #, > > > So I don't think that's going to work out... Indeed, it's only with 4.5 that non-float operands are properly supported here. Sad. Jan