From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 22852 invoked by alias); 5 Dec 2003 02:13:34 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 22809 invoked from network); 5 Dec 2003 02:13:32 -0000 Received: from unknown (HELO nevyn.them.org) (66.93.172.17) by sources.redhat.com with SMTP; 5 Dec 2003 02:13:32 -0000 Received: from drow by nevyn.them.org with local (Exim 4.24 #1 (Debian)) id 1AS5TC-00036Y-FQ; Thu, 04 Dec 2003 21:13:26 -0500 Date: Fri, 05 Dec 2003 02:13:00 -0000 From: Daniel Jacobowitz To: Ulrich Weigand Cc: Richard Henderson , gdb-patches@sources.redhat.com, uweigand@de.ibm.com Subject: Re: [PATCH] S/390 DWARF-2 CFI frame support Message-ID: <20031205021326.GA11837@nevyn.them.org> Mail-Followup-To: Ulrich Weigand , Richard Henderson , gdb-patches@sources.redhat.com, uweigand@de.ibm.com References: <20031205004756.GA18170@redhat.com> <200312050203.DAA08025@faui1d.informatik.uni-erlangen.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200312050203.DAA08025@faui1d.informatik.uni-erlangen.de> User-Agent: Mutt/1.5.1i X-SW-Source: 2003-12/txt/msg00182.txt.bz2 On Fri, Dec 05, 2003 at 03:03:52AM +0100, Ulrich Weigand wrote: > However, there is one point that requires special consideration > in any case: what if the stack pointer is unspecified? On s390, > treating it as preserved from the inner frame is correct, but > on i386 and other platforms I guess this would be wrong -- it > needs to be set to the CFA there (which is wrong on s390 due > to the CFA bias we have). > > I'm not familiar enough with gdb internals to decide what the > cleanest way of signaling that difference would be -- but there > needs to be *some* way or s390 will not work. (My reggroup > suggestion would provide a way to solve this problem as well.) Just for my information... Do you mean that the stack pointer always either: - is preserved, i.e. UNMODIFIED - or has explicit unwind information? If so then yes, we definitely need to change something. I'm also in favor of killing the warning, as I said at the time - having used a GDB with it in place for some time now, I find it quite annoying. -- Daniel Jacobowitz MontaVista Software Debian GNU/Linux Developer