From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 15062 invoked by alias); 17 Sep 2008 14:35:26 -0000 Received: (qmail 13504 invoked by uid 48); 17 Sep 2008 14:34:04 -0000 Date: Wed, 17 Sep 2008 14:35:00 -0000 Message-ID: <20080917143404.13503.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug tree-optimization/37102] [4.3/4.4 Regression] out-of-SSA is broken In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "amacleod at redhat dot com" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2008-09/txt/msg02019.txt.bz2 ------- Comment #13 from amacleod at redhat dot com 2008-09-17 14:34 ------- I was in the middle of updating this PR and taking possesion :-) Upon further reflection, I don't think it is acceptable for out-of-ssa to generate incorrect code simply because an optimization wasn't run before it. So, there are 2 basic choices. ONnis as you state, only there is an additional bit required during the rewrite verification. The other option is a very mini dead phi removal done at the same time as virtual phis are eliminated. its only a few lines of code and simple removes PHIs with no uses, and takes care of the ripple effect as well. I have a patch for it as well, its just not completely tested yet. I'll attach both in a minute, I just wanted you to know I was looking at it :-) -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37102