From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 21953 invoked by alias); 6 Dec 2004 11:26:14 -0000 Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org Received: (qmail 21878 invoked by uid 48); 6 Dec 2004 11:26:07 -0000 Date: Mon, 06 Dec 2004 11:26:00 -0000 Message-ID: <20041206112607.21877.qmail@sourceware.org> From: "amodra at bigpond dot net dot au" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20041201060338.18751.stuckey@earthlink.net> References: <20041201060338.18751.stuckey@earthlink.net> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug target/18751] [4.0 Regression] ICE in expr.c:5083, profiledbootstrap X-Bugzilla-Reason: CC X-SW-Source: 2004-12/txt/msg00827.txt.bz2 List-Id: ------- Additional Comments From amodra at bigpond dot net dot au 2004-12-06 11:26 ------- I think this particular problem was introduced with one of Geoff's patches, http://gcc.gnu.org/ml/gcc-patches/2004-08/msg00022.html rs6000_mode_dependent_address returns true for LO_SUM addresses, which in turn makes them not offsettable. Thus movdi_internal32 won't match DImode gpr stores. It's clear that LO_SUM addresses can't be offset (since an offset may bump the final address over a 64k+32k boundary, and the compiler can't know when this might happen). Also, the rtl generated by -fprofile-generate looks OK. I don't think there's anything wrong with Geoff's patch either, but it's boxed us into a corner. A fix that will work (I think) is to limit legitimate_lo_sum_address_p to 32 bit modes, but that could pessimize all loads and stores of 64 bit values. -- What |Removed |Added ---------------------------------------------------------------------------- CC| |geoffk at geoffk dot org http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18751