From: Avi Kivity Subject: Re: [PATCH 12/45] KVM: x86: Support for user space injected NMIs Date: Wed, 10 Dec 2008 10:46:33 +0200 Message-ID: <493F81E9.90306@redhat.com> References: <1228736216-15787-1-git-send-email-avi@redhat.com> <1228736216-15787-13-git-send-email-avi@redhat.com> <493D0E0A.2000400@siemens.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org To: Jan Kiszka Return-path: Received: from mx2.redhat.com ([66.187.237.31]:50248 "EHLO mx2.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754701AbYLJIqi (ORCPT ); Wed, 10 Dec 2008 03:46:38 -0500 In-Reply-To: <493D0E0A.2000400@siemens.com> Sender: kvm-owner@vger.kernel.org List-ID: Jan Kiszka wrote: > Avi Kivity wrote: > >> From: Jan Kiszka >> >> Introduces the KVM_NMI IOCTL to the generic x86 part of KVM for >> injecting NMIs from user space and also extends the statistic report >> accordingly. >> > Avi, please consider [1] again. You've already merged the related user > space cleanup, and I was just waiting for your comment on how to deal > with the kernel part, ideally _before_ merging the superfluous > ready_for/request NMI interface into mainline. > > Thanks, > Jan > > [1] http://permalink.gmane.org/gmane.comp.emulators.kvm.devel/25016 > > Do you mean, drop these nmi window bits completely from 2.6.29, keeping just injection? We can do that without breaking released userspace by dropping KVM_CAP_NMI and adding KVM_CAP_NMI2 (better name?) that doesn't have interrupt window support. This means that userspace versions that supported KVM_CAP_NMI will not work with 2.6.29, but I think we can live with that. If this is acceptable, please send a patch that converts KVM_CAP_NMI to KVM_CAP_NMI2, and a corresponding patch for userspace. I will fold the kernel patch into the 2.6.29 submission, and apply both patches to the master branch. -- error compiling committee.c: too many arguments to function