From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 13725 invoked by alias); 19 Oct 2004 22:23:45 -0000 Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Received: (qmail 13716 invoked from network); 19 Oct 2004 22:23:44 -0000 Received: from unknown (HELO dberlin.org) (68.164.203.246) by sourceware.org with SMTP; 19 Oct 2004 22:23:44 -0000 Received: from [192.168.1.7] (account dberlin HELO [192.168.1.7]) by dberlin.org (CommuniGate Pro SMTP 4.2.1) with ESMTP-TLS id 7400389; Tue, 19 Oct 2004 18:23:44 -0400 In-Reply-To: <1098223558.2915.131.camel@localhost.localdomain> References: <200410182303.52997.stevenb@suse.de> <1098218905.2915.123.camel@localhost.localdomain> <200410192328.28252.stevenb@suse.de> <1098223558.2915.131.camel@localhost.localdomain> Mime-Version: 1.0 (Apple Message framework v619) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <896EAE9A-221D-11D9-9D21-000A95DA505C@dberlin.org> Content-Transfer-Encoding: 7bit Cc: dnovillo@redhat.com, gcc-patches@gcc.gnu.org, amacleod@redhat.com, Steven Bosscher From: Daniel Berlin Subject: Re: Fix a tcb crash and a potential bug on mainline Date: Tue, 19 Oct 2004 22:27:00 -0000 To: law@redhat.com X-SW-Source: 2004-10/txt/msg01647.txt.bz2 > >> >>>> This may be something Andrew may want to look into as part >>>> of his immediate uses work?? >>> >>> It's something we've pondered and talked about a little. Basically >>> with the immediate use work we would know where all the uses are >>> and we could probably arrange to delete those uses. >> >> How would you delete those uses? You don't know if those statements >> are dead, so you really can't do that. > If they are using the result of a deleted PHI or some other deleted > statement, then those statements must also be dead. It's one of > the nice SSA properties. > Only if your pass performs *all* insertions and updates to statements before performing *any* removals of statements. I believe some passes don't do this right now. Otherwise, in the intermediate step, we would think some statements are dead when they are only dead because we haven't fixed them up yet :) --Dan