From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 30031 invoked by alias); 12 Aug 2003 17:06:49 -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 30017 invoked from network); 12 Aug 2003 17:06:43 -0000 Received: from unknown (HELO walton.kettenis.dyndns.org) (213.93.115.144) by sources.redhat.com with SMTP; 12 Aug 2003 17:06:43 -0000 Received: from elgar.kettenis.dyndns.org (elgar.kettenis.dyndns.org [192.168.0.2]) by walton.kettenis.dyndns.org (8.12.6p2/8.12.5) with ESMTP id h7CH6Zkf000560; Tue, 12 Aug 2003 19:06:35 +0200 (CEST) (envelope-from kettenis@elgar.kettenis.dyndns.org) Received: from elgar.kettenis.dyndns.org (localhost [127.0.0.1]) by elgar.kettenis.dyndns.org (8.12.6p2/8.12.6) with ESMTP id h7CH6ZcM026193; Tue, 12 Aug 2003 19:06:35 +0200 (CEST) (envelope-from kettenis@elgar.kettenis.dyndns.org) Received: (from kettenis@localhost) by elgar.kettenis.dyndns.org (8.12.6p2/8.12.6/Submit) id h7CH6Yl4026190; Tue, 12 Aug 2003 19:06:34 +0200 (CEST) Date: Tue, 12 Aug 2003 17:06:00 -0000 Message-Id: <200308121706.h7CH6Yl4026190@elgar.kettenis.dyndns.org> From: Mark Kettenis To: kevinb@redhat.com CC: gdb-patches@sources.redhat.com In-reply-to: <1030812163937.ZM20976@localhost.localdomain> (message from Kevin Buettner on Tue, 12 Aug 2003 09:39:37 -0700) Subject: Re: [RFA/i386]: Enable default support for SSE registers References: <1030722184305.ZM4145@localhost.localdomain> <1030729154932.ZM7867@localhost.localdomain> <863cg6riq5.fsf@elgar.kettenis.dyndns.org> <1030812163937.ZM20976@localhost.localdomain> X-SW-Source: 2003-08/txt/msg00198.txt.bz2 Date: Tue, 12 Aug 2003 09:39:37 -0700 From: Kevin Buettner On Aug 12, 5:36pm, Mark Kettenis wrote: > There is one more concern that I do have: If we include > the SSE registers, we somehow should also set their values to > something sensible for targets that don't support them. We already do > this for Linux, but it makes more sense to add some code to > i387_supply_fsave(). I just took a quick look at this. It's not obvious to me what you have in mind. Would you mind taking care of this? I'll take care of it. Also... I just noticed the following bit of code in i387-tdep.c: i387_supply_register (int regnum, char *fsave) { if (fsave == NULL) { supply_register (regnum, NULL); return; } Calling supply_register() with NULL will end up calling memcpy() with NULL as the second argument. If memcpy() is told to copy a non-zero number of characters, it'll SEGV. Huh, If you call support_register() with NULL, it will use memset() to initialize the register with all zeroes. > Do you want it on the 6.0 branch too? It's not critical, but it'd be nice. I'll see. At the very least I'll wait a few days and see whether someone complains about the changed behaviour. Mark