From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 8227 invoked by alias); 12 Jan 2009 16:19:12 -0000 Received: (qmail 6606 invoked by uid 48); 12 Jan 2009 16:18:52 -0000 Date: Mon, 12 Jan 2009 16:19:00 -0000 Message-ID: <20090112161852.6605.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug rtl-optimization/38245] [4.4 Regression] stack corruption when a call is removed but not the outgoing argument pushes In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "jakub 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: 2009-01/txt/msg01222.txt.bz2 ------- Comment #21 from jakub at gcc dot gnu dot org 2009-01-12 16:18 ------- Created an attachment (id=17078) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=17078&action=view) gcc44-pr38245.patch Patch I'm playing with. I don't see why changing CALL_INSN_FUNCTION_USAGE to have only at most word sized stores would simplify it much. The patch is still overly conservative, on the other side some stack arguments (e.g. large struct by value) still aren't added to CALL_INSN_FUNCTION_USAGE. If such an argument is first or last, this patch still won't catch it. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38245