From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH] kvm: work around inability of older kvm modules to destroy memory regions Date: Wed, 17 Dec 2008 14:26:20 +0200 Message-ID: <4948EFEC.2020407@redhat.com> References: <20081209161057.DC28A25006D@cleopatra.tlv.redhat.com> <493FCD5A.7010308@linux.vnet.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: kvm , Hollis Blanchard , Carsten Otte , Christian Borntraeger To: Christian Ehrhardt Return-path: Received: from mx2.redhat.com ([66.187.237.31]:36915 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750958AbYLQM1F (ORCPT ); Wed, 17 Dec 2008 07:27:05 -0500 In-Reply-To: <493FCD5A.7010308@linux.vnet.ibm.com> Sender: kvm-owner@vger.kernel.org List-ID: Christian Ehrhardt wrote: > Hi, this patch breaks all non x86 architectures as libkvm/libkvm-x86.c > has the only implementation of the alias functionality. > Until now only qemu-kvm-x86 has called that functions, but since this > patch the generic qemu-kvm.c calls them which leads to unresolved > symbols for powerpc, s390 and surely ia64 too. > > Well we could insert stubs for these call, but when looking on the > kernel side x86 is also the only implementer of the > KVM_SET_MEMORY_ALIAS ioctl. Until more arch support that there is no > reason to create these functions for non-x86 in libkvm. Also the > assumptions which addresses must be aliased base on hardware specific > assumptions e.g. vga card -> arch specific too. > > For now I hold a no-op stub in my private queue to test powerpc, but > eventually this mechanism should be arch dependent and this > implementation x86 only. > Avi could you modify your patch to work for the other arch's too ? Your band aid should be fine. Yes, it's ugly, but this will be in flux as we merge with upstream qemu. Please resend it with a signoff. -- error compiling committee.c: too many arguments to function