From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 11836 invoked by alias); 10 Dec 2008 17:13:11 -0000 Received: (qmail 11334 invoked by uid 48); 10 Dec 2008 17:11:48 -0000 Date: Wed, 10 Dec 2008 17:13:00 -0000 Message-ID: <20081210171148.11333.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug tree-optimization/32044] [4.3/4.4 Regression] final value replacement too aggressive for e.g. targets with no native div/mod insns In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "mmitchel at gcc dot gnu dot org" 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: 2008-12/txt/msg01043.txt.bz2 ------- Comment #52 from mmitchel at gcc dot gnu dot org 2008-12-10 17:11 ------- Steven, thanks for your comments on this issue. I agree with Richard G. that this is P2 -- but I also agree with you that it's a serious issue. Unfortunately, Bugzilla doesn't want to show me the test case attached to PR38453, so I can't look at that. Is the issue there that we don't take advantage of the ARM __aeabi_uidivmod function that gives both the quotient and modulus? Or is it that we're not giving CSE enough information to figure out that the value we have is already available? Zdenek, it would certainly be helpful to have the original justification for your change available. It does sound a bit like we've traded one set of problems for another, and if those are our only practical options, we need to decide which set of problems is worse. Thanks, -- Mark -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32044