From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 21976 invoked by alias); 13 Dec 2006 22:11:37 -0000 Received: (qmail 21967 invoked by uid 22791); 13 Dec 2006 22:11:36 -0000 X-Spam-Check-By: sourceware.org Received: from 195.22.55.53.adsl.nextra.cz (HELO host0.dyn.jankratochvil.net) (195.22.55.53) by sourceware.org (qpsmtpd/0.31) with ESMTP; Wed, 13 Dec 2006 22:11:30 +0000 Received: from host0.dyn.jankratochvil.net (localhost [127.0.0.1]) by host0.dyn.jankratochvil.net (8.13.8/8.13.8) with ESMTP id kBDMBC3x028470; Wed, 13 Dec 2006 23:11:12 +0100 Received: (from jkratoch@localhost) by host0.dyn.jankratochvil.net (8.13.8/8.13.8/Submit) id kBDMBCEx028469; Wed, 13 Dec 2006 23:11:12 +0100 Date: Wed, 13 Dec 2006 22:11:00 -0000 From: Jan Kratochvil To: Mark Kettenis Cc: gdb-patches@sourceware.org, Daniel Jacobowitz Subject: Re: [PATCH] Re: Unwinding CFI gcc practice of assumed `same value' regs Message-ID: <20061213221111.GA26406@host0.dyn.jankratochvil.net> References: <20061211190300.GA4372@host0.dyn.jankratochvil.net> <20061211224022.AD76E1800E7@magilla.sf.frob.com> <20061212155233.GH29911@devserv.devel.redhat.com> <20061213204603.GA11741@host0.dyn.jankratochvil.net> <20061213210357.GA27039@nevyn.them.org> <23342.82.92.89.47.1166046052.squirrel@webmail.xs4all.nl> <20061213214540.GA28965@nevyn.them.org> <5950.82.92.89.47.1166046772.squirrel@webmail.xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <5950.82.92.89.47.1166046772.squirrel@webmail.xs4all.nl> User-Agent: Mutt/1.4.2.2i 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: 2006-12/txt/msg00183.txt.bz2 On Wed, 13 Dec 2006 22:52:52 +0100, Mark Kettenis wrote: ... > No, sorry, it is me who is confused. I didn't actually look at Jan's patch > and assumed it did the the right thing of marking the return address as > undefined. Therefore what should the patch do? Currently `.cfi_undefined' looks as too radical. GDB may check for `PC == 0 && strcmp (name, "clone") == 0' (more reliably). That `strcmp (name, "clone")' needs to be coded in more trustworthy way. This "outermost" framework is already present in GDB but it would mean to add besides `set backtrace past-main' and `set backtrace past-entry' also some `set backtrace past-clone' and several more others in the future. Thanks, Jan