From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH 5/6] KVM/userspace: Device Assignment: Support for assigning PCI devices to guests Date: Sun, 26 Oct 2008 14:54:02 +0200 Message-ID: <4904686A.9030705@redhat.com> References: <1224861988-22781-1-git-send-email-amit.shah@redhat.com> <1224861988-22781-2-git-send-email-amit.shah@redhat.com> <1224861988-22781-3-git-send-email-amit.shah@redhat.com> <1224861988-22781-4-git-send-email-amit.shah@redhat.com> <1224861988-22781-5-git-send-email-amit.shah@redhat.com> <1224861988-22781-6-git-send-email-amit.shah@redhat.com> <4901F658.5080501@codemonkey.ws> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Amit Shah , kvm@vger.kernel.org, weidong.han@intel.com, allen.m.kay@intel.com, muli@il.ibm.com, benami@il.ibm.com To: Anthony Liguori Return-path: Received: from mx2.redhat.com ([66.187.237.31]:34517 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751727AbYJZMzH (ORCPT ); Sun, 26 Oct 2008 08:55:07 -0400 In-Reply-To: <4901F658.5080501@codemonkey.ws> Sender: kvm-owner@vger.kernel.org List-ID: Anthony Liguori wrote: > > I don't think having a kvm_enabled() check here is very useful. I > think device-assignment.c should be conditional on USE_KVM, and the > only kvm_enabled() check should be when creating the initial device > assignment. Practically speaking, QEMU is never going to support > device assignment outside of the context of KVM because I strongly > doubt anything like irqhook will make it upstream. Userspace interrupt handlers are actually possible with MSI; we should see if uio is open to adding support for that. -- error compiling committee.c: too many arguments to function