From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH] KVM: IA64: Using CFLAGS_vcpu.o to compile vcpu.c Date: Thu, 20 Nov 2008 14:23:25 +0200 Message-ID: <492556BD.5090009@redhat.com> References: <706158FABBBA044BAD4FE898A02E4BC219B28A98@pdsmsx503.ccr.corp.intel.com> <492540EC.9060805@redhat.com> <706158FABBBA044BAD4FE898A02E4BC219B28D98@pdsmsx503.ccr.corp.intel.com> <4925446F.10709@redhat.com> <706158FABBBA044BAD4FE898A02E4BC219B28D9E@pdsmsx503.ccr.corp.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: "kvm-ia64@vger.kernel.org" , "kvm@vger.kernel.org" To: "Zhang, Xiantao" Return-path: Received: from mx2.redhat.com ([66.187.237.31]:42154 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754036AbYKTMXQ (ORCPT ); Thu, 20 Nov 2008 07:23:16 -0500 In-Reply-To: <706158FABBBA044BAD4FE898A02E4BC219B28D9E@pdsmsx503.ccr.corp.intel.com> Sender: kvm-owner@vger.kernel.org List-ID: Zhang, Xiantao wrote: > Avi Kivity wrote: > >> Zhang, Xiantao wrote: >> >>>> Can you provide more information on how this was only discovered >>>> now? Was there a change in mainline that removed support for >>>> EXTRA_CFLAGS?] >>>> >>>> >>> It should be an old issue we didn't sense before. To specify single >>> object's compile option, only CFLAGS can work in this case. >>> >> How did we not notice the issue? What is the effect of the missing >> flags? >> It seems like a critical problem that should prevent kvm from working >> at all, so I don't understand how it was discovered so late. >> > > Once missing this flags, compiler can touch the fixed registers and leads to faults. But fortunately,current gcc didn't use them for optimizations, so we didn't find it before. But anyway it should be a potential issue. > Updated one, remove kernel's default range. > > Okay, so it worked by accident? > Subject: [PATCH] KVM: ia64: Fix: Use CFLAGS_vcpu.o to compile vcpu.c > > EXTRA_CFLAGS doesn't work for this case.Without this flag, > it may result in compiler touch fixed registers. > Applied, thanks. -- I have a truly marvellous patch that fixes the bug which this signature is too narrow to contain.