From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 2420 invoked by alias); 6 May 2006 15:19:17 -0000 Received: (qmail 2386 invoked by uid 22791); 6 May 2006 15:19:16 -0000 X-Spam-Check-By: sourceware.org Received: from nevyn.them.org (HELO nevyn.them.org) (66.93.172.17) by sourceware.org (qpsmtpd/0.31.1) with ESMTP; Sat, 06 May 2006 15:19:12 +0000 Received: from drow by nevyn.them.org with local (Exim 4.54) id 1FcOYi-0006J2-Sm; Sat, 06 May 2006 11:19:04 -0400 Date: Sat, 06 May 2006 15:19:00 -0000 From: Daniel Jacobowitz To: Mark Kettenis Cc: davem@davemloft.net, uweigand@de.ibm.com, gdb-patches@sourceware.org Subject: Re: [committed] Fix gcore crashes on s390 Message-ID: <20060506151904.GB24060@nevyn.them.org> Mail-Followup-To: Mark Kettenis , davem@davemloft.net, uweigand@de.ibm.com, gdb-patches@sourceware.org References: <200605060118.k461IpKO009920@d12av02.megacenter.de.ibm.com> <20060505.183205.74231759.davem@davemloft.net> <200605060825.k468P3BX023503@elgar.sibelius.xs4all.nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200605060825.k468P3BX023503@elgar.sibelius.xs4all.nl> User-Agent: Mutt/1.5.8i X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2006-05/txt/msg00147.txt.bz2 On Sat, May 06, 2006 at 10:25:03AM +0200, Mark Kettenis wrote: > Yes, targets are allowed to not implement the collect_regset functions > if they only implement reading core dumps and don't need it for > something else (like writing core dumps with gcore or fiddling with > threads). All Linux targets now fall in the second category, but some > probably think they fall in the first. So either we should: > > 1. Deal gracefully with the collect_regset function pointer being > NULL. > > 2. Put in a gdb_assert() to check it's not null before it's used. > > I'm thinking that we should try option #2 for a while to get people to > implement the functions for the other Linux targets too, just like > Ulrich did for s390. Oh - sorry, I didn't see this before I approved David's change. I'd rather not do it this way, especially with GDB 6.5 upcoming and a lot of the Linux targets getting somewhat sporadic test coverage. Is that all right with you? If we want to make it mandatory later, someone would have to go through the targets and fix them. -- Daniel Jacobowitz CodeSourcery