From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [Android-virt] [PATCH v4 05/10] ARM: KVM: Inject IRQs and FIQs from userspace Date: Tue, 09 Aug 2011 14:37:38 +0300 Message-ID: <4E411C02.6070008@redhat.com> References: <20110806103821.27198.41231.stgit@localhost6.localdomain6> <20110806103933.27198.91264.stgit@localhost6.localdomain6> <4E4106D8.3070305@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Christoffer Dall , tech@virtualopensystems.com, android-virt@lists.cs.columbia.edu, kvm@vger.kernel.org To: Christoffer Dall Return-path: Received: from mx1.redhat.com ([209.132.183.28]:59930 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753343Ab1HILh4 (ORCPT ); Tue, 9 Aug 2011 07:37:56 -0400 In-Reply-To: Sender: kvm-owner@vger.kernel.org List-ID: On 08/09/2011 02:27 PM, Christoffer Dall wrote: > > > >> + > >> + if (irq_level->level) { > >> + vcpu->arch.virt_irq |= mask; > >> + vcpu->arch.wait_for_interrupts = 0; > >> + } else > >> + vcpu->arch.virt_irq&= ~mask; > >> + > > > > This seems to be non-smp-safe? Do you need atomic ops and barriers > > here? And a wakeup? > > The whole thing is not SMP tested yet, so I took some shortcuts. I only recently got hold of a SMP model and SMP support will be a focus area for the next series. Thanks for pin-pointing this though. Note even a single vcpu guest on an smp host needs this. -- error compiling committee.c: too many arguments to function