From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:34177) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1dAHGe-0004Tv-JD for qemu-devel@nongnu.org; Mon, 15 May 2017 10:46:37 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1dAHGb-00071B-Dg for qemu-devel@nongnu.org; Mon, 15 May 2017 10:46:36 -0400 Received: from mx1.redhat.com ([209.132.183.28]:16294) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1dAHGb-00070v-4V for qemu-devel@nongnu.org; Mon, 15 May 2017 10:46:33 -0400 Date: Mon, 15 May 2017 17:46:28 +0300 From: "Michael S. Tsirkin" Message-ID: <20170515174044-mutt-send-email-mst@kernel.org> References: <1494478641-24549-1-git-send-email-wei.w.wang@intel.com> <591974EB.2010409@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <591974EB.2010409@intel.com> Subject: Re: [Qemu-devel] [PATCH] virtio-net: keep the packet layout intact List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Wei Wang Cc: jasowang@redhat.com, stefanha@gmail.com, marcandre.lureau@gmail.com, pbonzini@redhat.com, virtio-dev@lists.oasis-open.org, qemu-devel@nongnu.org, jan.scheurich@ericsson.com On Mon, May 15, 2017 at 05:29:15PM +0800, Wei Wang wrote: > Ping for comments, thanks. > > On 05/11/2017 12:57 PM, Wei Wang wrote: > > The current implementation may change the packet layout when > > vnet_hdr needs an endianness swap. The layout change causes > > one more iov to be added to the iov[] passed from the guest, which > > is a barrier to making the TX queue size 1024 due to the possible > > off-by-one issue. It blocks making it 512 but I don't think we can make it 1024 as entries might cross page boundaries and get split. > > > > This patch changes the implementation to remain the packet layout > > intact. In this case, the number of iov[] passed to writev will be > > equal to the number obtained from the guest. > > > > Signed-off-by: Wei Wang As this is at the cost of a full data copy, I don't think this makes sense. We could limit this when sg list does not fit in 1024. But I really think we should just add a max s/g field to virtio and then we'll be free to increase the ring size. > > --- > > hw/net/virtio-net.c | 29 ++++++++++++++--------------- > > 1 file changed, 14 insertions(+), 15 deletions(-) > > > > diff --git a/hw/net/virtio-net.c b/hw/net/virtio-net.c > > index 7d091c9..a51e9a8 100644 > > --- a/hw/net/virtio-net.c > > +++ b/hw/net/virtio-net.c > > @@ -1287,8 +1287,7 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q) > > for (;;) { > > ssize_t ret; > > unsigned int out_num; > > - struct iovec sg[VIRTQUEUE_MAX_SIZE], sg2[VIRTQUEUE_MAX_SIZE + 1], *out_sg; > > - struct virtio_net_hdr_mrg_rxbuf mhdr; > > + struct iovec sg[VIRTQUEUE_MAX_SIZE], mhdr, *out_sg; > > elem = virtqueue_pop(q->tx_vq, sizeof(VirtQueueElement)); > > if (!elem) { > > @@ -1305,7 +1304,10 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q) > > } > > if (n->has_vnet_hdr) { > > - if (iov_to_buf(out_sg, out_num, 0, &mhdr, n->guest_hdr_len) < > > + /* Buffer to copy vnet_hdr and the possible adjacent data */ > > + mhdr.iov_len = out_sg[0].iov_len; > > + mhdr.iov_base = g_malloc0(mhdr.iov_len); > > + if (iov_to_buf(out_sg, out_num, 0, mhdr.iov_base, mhdr.iov_len) < > > n->guest_hdr_len) { > > virtio_error(vdev, "virtio-net header incorrect"); > > virtqueue_detach_element(q->tx_vq, elem, 0); > > @@ -1313,17 +1315,12 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q) > > return -EINVAL; > > } > > if (n->needs_vnet_hdr_swap) { > > - virtio_net_hdr_swap(vdev, (void *) &mhdr); > > - sg2[0].iov_base = &mhdr; > > - sg2[0].iov_len = n->guest_hdr_len; > > - out_num = iov_copy(&sg2[1], ARRAY_SIZE(sg2) - 1, > > - out_sg, out_num, > > - n->guest_hdr_len, -1); > > - if (out_num == VIRTQUEUE_MAX_SIZE) { > > - goto drop; > > - } > > - out_num += 1; > > - out_sg = sg2; > > + virtio_net_hdr_swap(vdev, mhdr.iov_base); > > + /* Copy the first iov where the vnet_hdr resides in */ > > + out_num = iov_copy(sg, 1, &mhdr, 1, 0, mhdr.iov_len); > > + out_num += iov_copy(sg + 1, ARRAY_SIZE(sg) - 1, out_sg + 1, > > + elem->out_num - 1, 0, -1); > > + out_sg = sg; > > } > > } > > /* > > @@ -1345,13 +1342,15 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q) > > ret = qemu_sendv_packet_async(qemu_get_subqueue(n->nic, queue_index), > > out_sg, out_num, virtio_net_tx_complete); > > + if (n->has_vnet_hdr) { > > + g_free(mhdr.iov_base); > > + } > > if (ret == 0) { > > virtio_queue_set_notification(q->tx_vq, 0); > > q->async_tx.elem = elem; > > return -EBUSY; > > } > > -drop: > > virtqueue_push(q->tx_vq, elem, 0); > > virtio_notify(vdev, q->tx_vq); > > g_free(elem);