From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 27158 invoked by alias); 3 Feb 2010 14:46:37 -0000 Received: (qmail 27149 invoked by uid 22791); 3 Feb 2010 14:46:36 -0000 X-SWARE-Spam-Status: No, hits=-2.4 required=5.0 tests=AWL,BAYES_00,SARE_SUB_OBFU_Q1,SPF_PASS X-Spam-Check-By: sourceware.org Received: from mail.codesourcery.com (HELO mail.codesourcery.com) (38.113.113.100) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Wed, 03 Feb 2010 14:46:33 +0000 Received: (qmail 6270 invoked from network); 3 Feb 2010 14:46:31 -0000 Received: from unknown (HELO caradoc.them.org) (dan@127.0.0.2) by mail.codesourcery.com with ESMTPA; 3 Feb 2010 14:46:31 -0000 Date: Wed, 03 Feb 2010 14:46:00 -0000 From: Daniel Jacobowitz To: "H.J. Lu" Cc: GDB Subject: Re: RFC: Support target specific qSupported Message-ID: <20100203144626.GA31389@caradoc.them.org> Mail-Followup-To: "H.J. Lu" , GDB References: <20100203040339.GA24984@lucon.org> <20100203135848.GA27938@caradoc.them.org> <6dc9ffc81002030605k6eadda3me45828f7c8c6a362@mail.gmail.com> <20100203142202.GB28951@caradoc.them.org> <6dc9ffc81002030634wd2976e5t2a42b639c455b850@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <6dc9ffc81002030634wd2976e5t2a42b639c455b850@mail.gmail.com> User-Agent: Mutt/1.5.20 (2009-06-14) Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2010-02/txt/msg00073.txt.bz2 On Wed, Feb 03, 2010 at 06:34:14AM -0800, H.J. Lu wrote: > We have our own remote gdb stub, which needs to talk to both old gdb, > which only understands SSE g/G packet, and new gdb, which understands > AVX g/G packet. Does the target description support negotiation so > that old gdb and new remote gdb stub can use SSE g/G packet? The goal of target descriptions is to not need negotiation. Everything is controlled by the target. But with older GDBs, because x86 did not get target-described register support right away, there's a problem. I suggest adding something to the GDB-side qSupported packet saying that AVX is OK. You don't need anything on the stub side of the qSupported reply; you just need to reply to qXfer:features. -- Daniel Jacobowitz CodeSourcery