From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 32704 invoked by alias); 23 Apr 2002 17:47:36 -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 32675 invoked from network); 23 Apr 2002 17:47:32 -0000 Received: from unknown (HELO cygnus.com) (205.180.83.203) by sources.redhat.com with SMTP; 23 Apr 2002 17:47:32 -0000 Received: from redhat.com (reddwarf.sfbay.redhat.com [172.16.24.50]) by runyon.cygnus.com (8.8.7-cygnus/8.8.7) with ESMTP id KAA10691; Tue, 23 Apr 2002 10:47:30 -0700 (PDT) Message-ID: <3CC59B63.17E26EBC@redhat.com> Date: Tue, 23 Apr 2002 10:47:00 -0000 From: Michael Snyder Organization: Red Hat, Inc. X-Accept-Language: en MIME-Version: 1.0 To: Richard.Earnshaw@arm.com CC: Michael Snyder , gdb-patches@sources.redhat.com, cagney@redhat.com, rearnsha@arm.com Subject: Re: [RFA] arm_skip_prologue: recognize "str r(0123), [r11, #-nn]" References: <200204230954.KAA01842@cam-mail2.cambridge.arm.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-SW-Source: 2002-04/txt/msg00856.txt.bz2 Richard Earnshaw wrote: > > > > > > > GCC sometimes generates these stores as discrete instructions. > > > > > > 2002-04-22 Michael Snyder > > > > > > * arm-tdep.c (arm_skip_prologue): Recognize str r(0123), [r11, #-nn]. > > > > > > > Are you sure these should really be considered part of the prologue? > > > > > > > > Ok, here's how I think we should consider this. > > If a DWARF-encoded sequence would consider these as part of the prologue > (ie they are described as prologue instructions, then the unwinder should > consider them similarly. If, however, it considers them as part of the > function body then the manual unwinder should also do so. Richard, This disassembly method is only invoked if there is no dwarf info. If we do have line numbers etc, we use those by preference. So when we are disassembling, we cannot rely on hints from the debug info. I have confirmed that, when there is debug info, these instructions are lumped with the prologue. > > If the latter is correct, then we should also handle storing into the > stack so that we can also handle frameless functions. > > R.