From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 30283 invoked by alias); 17 Oct 2007 00:36:14 -0000 Received: (qmail 30275 invoked by uid 22791); 17 Oct 2007 00:36:14 -0000 X-Spam-Check-By: sourceware.org Received: from NaN.false.org (HELO nan.false.org) (208.75.86.248) by sourceware.org (qpsmtpd/0.31) with ESMTP; Wed, 17 Oct 2007 00:36:12 +0000 Received: from nan.false.org (localhost [127.0.0.1]) by nan.false.org (Postfix) with ESMTP id 5E25598338; Wed, 17 Oct 2007 00:36:10 +0000 (GMT) Received: from caradoc.them.org (22.svnf5.xdsl.nauticom.net [209.195.183.55]) by nan.false.org (Postfix) with ESMTP id EF11898100; Wed, 17 Oct 2007 00:36:09 +0000 (GMT) Received: from drow by caradoc.them.org with local (Exim 4.68) (envelope-from ) id 1IhwtN-0000Mp-3d; Tue, 16 Oct 2007 20:36:09 -0400 Date: Wed, 17 Oct 2007 01:02:00 -0000 From: Daniel Jacobowitz To: Ulrich Weigand Cc: gdb-patches@sourceware.org Subject: Re: [patch] gdbserver support for e500 / altivec Message-ID: <20071017003609.GA1336@caradoc.them.org> Mail-Followup-To: Ulrich Weigand , gdb-patches@sourceware.org References: <20071005180207.GA4545@caradoc.them.org> <200710162209.l9GM9WsD014935@d12av02.megacenter.de.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200710162209.l9GM9WsD014935@d12av02.megacenter.de.ibm.com> User-Agent: Mutt/1.5.15 (2007-04-09) X-IsSubscribed: yes 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: 2007-10/txt/msg00434.txt.bz2 On Wed, Oct 17, 2007 at 12:09:32AM +0200, Ulrich Weigand wrote: > Dan Jacobowitz wrote: > > > * configure.srv: Use alternate register sets for powerpc64-*-linux* > > with AltiVec, powerpc-*-linux* with AltiVec, and powerpc-*-linux* > > with SPE. > > This uses a configure-time check whether __ALTIVEC__ is defined. However, > GCC for PowerPC by default does not define __ALTIVEC__, it does so only > when using the -maltivec flag. In any case, this is rather independent of > whether the target system supports AltiVec or not ... Not quite independent: if you built gdbserver with -maltivec (and -mabi=altivec normally) then gdbserver will require altivec to run. The other direction is a problem though. > Why does this need to be a configure check in the first place? Wouldn't > it be much preferable if you could build a single gdbserver binary that > would report the properties of the system it happens to be running on? I agree. But to do it that way, you need to make gdbserver support multiple register cache layouts; multiple potential XML descriptions to return as target.xml; and runtime detection of which register set to use. It's definitely doable (GDB does it). I haven't found the time yet, and this was easy. So consider it a missing feature, or a challenge :-) -- Daniel Jacobowitz CodeSourcery