From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 14929 invoked by alias); 23 Jul 2009 11:36:41 -0000 Received: (qmail 14828 invoked by uid 48); 23 Jul 2009 11:36:26 -0000 Date: Thu, 23 Jul 2009 11:36:00 -0000 Message-ID: <20090723113626.14827.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug rtl-optimization/40667] [4.4/4.5 Regression] stack frames are generated even with -fomit-frame-pointer 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: 2009-07/txt/msg01853.txt.bz2 ------- Comment #17 from jakub at gcc dot gnu dot org 2009-07-23 11:36 ------- Incorrect? Why do you think so? double on ix86 has 8 byte alignment, and unlike long long it has also performance effects when you misalign it (in theory, we could handle double the same as long long with -Os though). When gcc needs to decide whether to align the stack or not, it doesn't yet know whether any temporaries of that type will need to be spilled to stack. -mpreferred-stack-boundary=2 is typically only used in the kernel anyway and that one doesn't have floating point stuff, as this option affects the Linux ABI (you can't call libc functions from code compiled that way among other things). -- jakub at gcc dot gnu dot org changed: What |Removed |Added ---------------------------------------------------------------------------- Status|REOPENED |RESOLVED Resolution| |FIXED http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40667