From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Michael S. Tsirkin" Subject: Re: [RFC][PATCH v2 0/3] Provide a zero-copy method on KVM virtio-net. Date: Thu, 22 Apr 2010 12:19:43 +0300 Message-ID: <20100422091943.GA30532@redhat.com> References: <1270193100-6769-1-git-send-email-xiaohui.xin@intel.com> <20100414152519.GA10792@redhat.com> <97F6D3BD476C464182C1B7BABF0B0AF5C18969CC@shzsmsx502.ccr.corp.intel.com> <20100415100546.GA17035@redhat.com> <20100419102118.GA16198@redhat.com> <20100421083507.GA30855@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "netdev@vger.kernel.org" , "kvm@vger.kernel.org" , "linux-kernel@vger.kernel.org" , "mingo@elte.hu" , "jdike@linux.intel.com" , "davem@davemloft.net" To: "Xin, Xiaohui" Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On Thu, Apr 22, 2010 at 04:57:56PM +0800, Xin, Xiaohui wrote: > Michael, > > >Yes, I think this packet split mode probably maps well to mergeable buffer > >support. Note that > >1. Not all devices support large packets in this way, others might map > > to indirect buffers better > > Do the indirect buffers accord to deal with the skb->frag_list? We currently use skb->frags. > > So we have to figure out how migration is going to work > Yes, different guest virtio-net driver may contain different features. > Does the qemu migration work with different features supported by virtio-net > driver now? For now, you must have identical feature-sets for migration to work. And long as we manage the buffers in software, we can always make features match. > >2. It's up to guest driver whether to enable features such as > > mergeable buffers and indirect buffers > > So we have to figure out how to notify guest which mode > > is optimal for a given device > Yes. When a device is binded, the mp device may query the capabilities from driver. > Actually, there is a structure now in mp device can do this, we can add some field > to support more. > > >3. We don't want to depend on jumbo frames for decent performance > > So we probably should support GSO/GRO > GSO is for the tx side, right? I think driver can handle it itself. > For GRO, I'm not sure it's easy or not. Basically, the mp device now > we have support is doing what raw socket is doing. The packets are not going to host stack. See commit bfd5f4a3d605e0f6054df0b59fe0907ff7e696d3 (it doesn't currently work with vhost net, but that's a separate story). > -- > MST