From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 5587 invoked by alias); 5 Dec 2007 10:13:07 -0000 Received: (qmail 5460 invoked by uid 48); 5 Dec 2007 10:12:51 -0000 Date: Wed, 05 Dec 2007 10:13:00 -0000 Message-ID: <20071205101251.5459.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug middle-end/33088] [4.1/4.2/4.3 Regression] spurious exceptions with -ffloat-store In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "ebotcazou 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: 2007-12/txt/msg00351.txt.bz2 ------- Comment #11 from ebotcazou at gcc dot gnu dot org 2007-12-05 10:12 ------- We could probably get away with a kludge for -ffloat-store and optimization, but currently the flag comes into play only very late (in TER) and I think it's better to keep this. So I think the approach to solving this is two-pronged: - at -O0, do not promote the partial stores to total stores in the gimplifier, - at -O1 or above, insert dummy initializations for "uninitialized" SSA names, this should be cleaned up at the RTL level (if -ffloat-store is not passed). Andrew, IIRC you extended DECL_COMPLEX_GIMPLE_REG_P to DECL_GIMPLE_REG_P, can vectors be affected by the same issue? -- ebotcazou at gcc dot gnu dot org changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |pinskia at gcc dot gnu dot | |org http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33088