From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 5994 invoked by alias); 5 Aug 2010 20:54:25 -0000 Received: (qmail 5923 invoked by alias); 5 Aug 2010 20:54:03 -0000 Date: Thu, 05 Aug 2010 20:54:00 -0000 Message-ID: <20100805205403.5922.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug bootstrap/44970] [4.6 regression] Revision 162270 failed to bootstrap In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "dave at hiauly1 dot hia dot nrc dot ca" 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: 2010-08/txt/msg00346.txt.bz2 ------- Comment #67 from dave at hiauly1 dot hia dot nrc dot ca 2010-08-05 20:54 ------- Subject: Re: [4.6 regression] Revision 162270 failed to bootstrap > I think initial RTL generation is fine, so it looks like my change has exposed > a latent bug. What seems to happen is that some pass between expand and > combine lengthens the lifetime of register %r25, which holds an incoming > argument, so that it now crosses another call, which clobbers it. > > This seems to happen in fwprop1, Cc Paolo. I can't find any code in that pass > which tries to handle the situation. Yes, the rtl generated by fwprop1 is wrong as r25 is clobbered by the call to pool_alloc. Dave -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44970