From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH] Use PCI revision field to indicate virtio PCI ABI version Date: Tue, 29 Jan 2008 17:32:20 +0200 Message-ID: <479F4704.8080107@qumranet.com> References: <1201535999-13998-1-git-send-email-aliguori@us.ibm.com> <479EDDCE.8000000@qumranet.com> <479F3528.9040203@us.ibm.com> <200801291529.53808.borntraeger@de.ibm.com> <479F3B58.9000301@qumranet.com> <479F3E8D.8000308@us.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <479F3E8D.8000308@us.ibm.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: virtualization-bounces@lists.linux-foundation.org Errors-To: virtualization-bounces@lists.linux-foundation.org To: Anthony Liguori Cc: Christian Borntraeger , virtualization@lists.linux-foundation.org List-Id: virtualization@lists.linuxfoundation.org Anthony Liguori wrote: > Avi Kivity wrote: >> Christian Borntraeger wrote: >>> Am Dienstag, 29. Januar 2008 schrieb Anthony Liguori: >>> >>>> That's not what I was agreeing too. I don't want to plumb an ABI >>>> interface through virtio for each device. This is what I didn't >>>> like about having an ABI field in the first place. I'm thinking we >>>> should just drop both of these and instead just rely on feature bits. >>>> >>> >>> Me also updating the our prototype code to the latest levels... >>> >>> And I agree with Anthony. Feature bits seems to be a much better >>> solution than ABI versions. >>> >> >> I agree that feature bits are the long term solution; but we need a >> short term solution before the ABI is stabilized. We don't want to >> add feature bits now, since that will encode virtio development >> history into those bits (likely consuming most of them). > > Well then let's stick with the current patches I put out. It gives us > a safe guard where we can gracefully break things. I'm inclined to > think that we should not bother just "breaking" a particular device > but break all of them at once. I don't want to overcomplicate > something that we expect to never use. > Yeah, okay. We'll bump it one last time when 2.6.25 is released. I'll merge the userspace side. -- error compiling committee.c: too many arguments to function