From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 21866 invoked by alias); 3 Jul 2004 01:00:41 -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 21855 invoked by alias); 3 Jul 2004 01:00:40 -0000 Date: Sat, 03 Jul 2004 01:00:00 -0000 Message-ID: <20040703010040.21854.qmail@sourceware.org> From: "wilson at specifixinc dot com" To: gcc-bugs@gcc.gnu.org In-Reply-To: <20040702031629.16325.wilson@gcc.gnu.org> References: <20040702031629.16325.wilson@gcc.gnu.org> Reply-To: gcc-bugzilla@gcc.gnu.org Subject: [Bug target/16325] value profiling clobbers gp on mips X-Bugzilla-Reason: CC X-SW-Source: 2004-07/txt/msg00200.txt.bz2 List-Id: ------- Additional Comments From wilson at specifixinc dot com 2004-07-03 01:00 ------- Subject: Re: value profiling clobbers gp on mips On Thu, 2004-07-01 at 21:09, pinskia at gcc dot gnu dot org wrote: > This this the case the work STARTING_FRAME_OFFSET done too ealy? Or in the same line of thought, maybe we are doing the profiling instrumentation too late. Moving either one of these across the other would solve the problem. I doubt that we want to move the virtual register instantiation later, but moving the profiling support earlier might be feasible. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16325