From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH uq/master 0/9] remove event_notifier_get_fd from non-KVM code Date: Thu, 12 Jul 2012 12:30:34 +0300 Message-ID: <4FFE993A.9090600@redhat.com> References: <1341501390-797-1-git-send-email-pbonzini@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: qemu-devel@nongnu.org, mtosatti@redhat.com, kvm@vger.kernel.org, anthony.perard@citrix.com, jan.kiszka@siemens.com, mst@redhat.com, stefano.stabellini@eu.citrix.com To: Paolo Bonzini Return-path: Received: from mx1.redhat.com ([209.132.183.28]:29578 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753453Ab2GLJat (ORCPT ); Thu, 12 Jul 2012 05:30:49 -0400 In-Reply-To: <1341501390-797-1-git-send-email-pbonzini@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On 07/05/2012 06:16 PM, Paolo Bonzini wrote: > This is part 1 of a three-part series that expands usage of EventNotifier > in QEMU (including AIO and the main loop). I started working on this when > playing with the threaded block layer; the part of that work that I hope > will be in 1.2 is generalizing posix-aio-compat.c to be a generic portable > thread pool + porting AIO to Win32 (part 2). On top of this, discard > can be easily made asynchronous (part 3), which is a prerequisite for > enabling it. > > This first part does the necessary changes for porting EventNotifier > to Win32. The Win32 version will not have event_notifier_get_fd, > and thus I want to remove all calls in portable code. Instead, all > functions used in portable code after this series take an EventNotifier; > KVM-specific implementations retrieve the file descriptor internally > (these calls are in hw/ivshmem.c, hw/vhost.c, kvm-all.c). > > Patches 1 to 6 cover ivshmem and the memory API, first adding the > required EventNotifier APIs and then using them. Patches 7 to 9 do the > same with KVM ioeventfd and irqfd, refactoring transport-independent > code in the process from virtio-pci to virtio (the two steps are a bit > hard to separate). Looks good, all that is needed is an explanation for patch 2. -- error compiling committee.c: too many arguments to function