From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Sender: List-Post: List-Help: List-Unsubscribe: List-Subscribe: Received: from lists.oasis-open.org (oasis-open.org [10.110.1.242]) by lists.oasis-open.org (Postfix) with ESMTP id 7E50B9865F1 for ; Thu, 25 Aug 2022 12:39:18 +0000 (UTC) From: Gavin Li Date: Thu, 25 Aug 2022 15:38:40 +0300 Message-ID: <20220825123840.20239-3-gavinl@nvidia.com> In-Reply-To: <20220825123840.20239-1-gavinl@nvidia.com> References: <20220825123840.20239-1-gavinl@nvidia.com> MIME-Version: 1.0 Subject: [virtio-dev] [PATCH RESEND v2 2/2] virtio-net: use mtu size as buffer length for big packets Content-Type: text/plain Content-Transfer-Encoding: quoted-printable To: stephen@networkplumber.org, davem@davemloft.net, jesse.brandeburg@intel.com, alexander.h.duyck@intel.com, kuba@kernel.org, sridhar.samudrala@intel.com, jasowang@redhat.com, loseweigh@gmail.com, netdev@vger.kernel.org, virtualization@lists.linux-foundation.org, virtio-dev@lists.oasis-open.org, mst@redhat.com Cc: gavi@nvidia.com, parav@nvidia.com List-ID: Currently add_recvbuf_big() allocates MAX_SKB_FRAGS segments for big packets even when GUEST_* offloads are not present on the device. However, if guest GSO is not supported, it would be sufficient to allocate segments to cover just up the MTU size and no further. Allocating the maximum amount of segments results in a large waste of buffer space in the queue, which limits the number of packets that can be buffered and can result in reduced performance. Therefore, if guest GSO is not supported, use the MTU to calculate the optimal amount of segments required. When guest offload is enabled at runtime, RQ already has packets of bytes less than 64K. So when packet of 64KB arrives, all the packets of such size will be dropped. and RQ is now not usable. So this means that during set_guest_offloads() phase, RQs have to be destroyed and recreated, which requires almost driver reload. If VIRTIO_NET_F_CTRL_GUEST_OFFLOADS has been negotiated, then it should always treat them as GSO enabled. Below is the iperf TCP test results over a Mellanox NIC, using vDPA for 1 VQ, queue size 1024, before and after the change, with the iperf server running over the virtio-net interface. MTU(Bytes)/Bandwidth (Gbit/s) Before After 1500 22.5 22.4 9000 12.8 25.9 Signed-off-by: Gavin Li Reviewed-by: Gavi Teitz Reviewed-by: Parav Pandit --- changelog: v1->v2 - Addressed comments from Jason, Michael, Si-Wei. - Remove the flag of guest GSO support, set sg_num for big packets and use it directly - Recalculate sg_num for big packets in virtnet_set_guest_offloads - Replace the round up algorithm with DIV_ROUND_UP --- drivers/net/virtio_net.c | 41 +++++++++++++++++++++++++++------------- 1 file changed, 28 insertions(+), 13 deletions(-) diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index e1904877d461..ec8c135a26d6 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -225,6 +225,9 @@ struct virtnet_info { =09/* I like... big packets and I cannot lie! */ =09bool big_packets; =20 +=09/* number of sg entries allocated for big packets */ +=09unsigned int big_packets_sg_num; + =09/* Host will merge rx buffers for big packets (shake it! shake it!) */ =09bool mergeable_rx_bufs; =20 @@ -1331,10 +1334,10 @@ static int add_recvbuf_big(struct virtnet_info *vi,= struct receive_queue *rq, =09char *p; =09int i, err, offset; =20 -=09sg_init_table(rq->sg, MAX_SKB_FRAGS + 2); +=09sg_init_table(rq->sg, vi->big_packets_sg_num + 2); =20 -=09/* page in rq->sg[MAX_SKB_FRAGS + 1] is list tail */ -=09for (i =3D MAX_SKB_FRAGS + 1; i > 1; --i) { +=09/* page in rq->sg[vi->big_packets_sg_num + 1] is list tail */ +=09for (i =3D vi->big_packets_sg_num + 1; i > 1; --i) { =09=09first =3D get_a_page(rq, gfp); =09=09if (!first) { =09=09=09if (list) @@ -1365,7 +1368,7 @@ static int add_recvbuf_big(struct virtnet_info *vi, s= truct receive_queue *rq, =20 =09/* chain first in list head */ =09first->private =3D (unsigned long)list; -=09err =3D virtqueue_add_inbuf(rq->vq, rq->sg, MAX_SKB_FRAGS + 2, +=09err =3D virtqueue_add_inbuf(rq->vq, rq->sg, vi->big_packets_sg_num + 2, =09=09=09=09 first, gfp); =09if (err < 0) =09=09give_pages(rq, first); @@ -3690,13 +3693,31 @@ static bool virtnet_check_guest_gso(const struct vi= rtnet_info *vi) =09=09virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_UFO)); } =20 +static void virtnet_set_big_packets_fields(struct virtnet_info *vi, const = int mtu) +{ +=09bool guest_gso =3D virtnet_check_guest_gso(vi); + +=09/* If we can receive ANY GSO packets, we must allocate large ones. */ +=09if (mtu > ETH_DATA_LEN || guest_gso) { +=09=09vi->big_packets =3D true; +=09=09/* if the guest offload is offered by the device, user can modify +=09=09 * offload capability. In such posted buffers may short fall of size= . +=09=09 * Hence allocate for max size. +=09=09 */ +=09=09if (virtio_has_feature(vi->vdev, VIRTIO_NET_F_CTRL_GUEST_OFFLOADS)) +=09=09=09vi->big_packets_sg_num =3D MAX_SKB_FRAGS; +=09=09else +=09=09=09vi->big_packets_sg_num =3D DIV_ROUND_UP(mtu, PAGE_SIZE); +=09} +} + static int virtnet_probe(struct virtio_device *vdev) { =09int i, err =3D -ENOMEM; =09struct net_device *dev; =09struct virtnet_info *vi; =09u16 max_queue_pairs; -=09int mtu; +=09int mtu =3D 0; =20 =09/* Find if host supports multiqueue/rss virtio_net device */ =09max_queue_pairs =3D 1; @@ -3784,10 +3805,6 @@ static int virtnet_probe(struct virtio_device *vdev) =09INIT_WORK(&vi->config_work, virtnet_config_changed_work); =09spin_lock_init(&vi->refill_lock); =20 -=09/* If we can receive ANY GSO packets, we must allocate large ones. */ -=09if (virtnet_check_guest_gso(vi)) -=09=09vi->big_packets =3D true; - =09if (virtio_has_feature(vdev, VIRTIO_NET_F_MRG_RXBUF)) =09=09vi->mergeable_rx_bufs =3D true; =20 @@ -3853,12 +3870,10 @@ static int virtnet_probe(struct virtio_device *vdev= ) =20 =09=09dev->mtu =3D mtu; =09=09dev->max_mtu =3D mtu; - -=09=09/* TODO: size buffers correctly in this case. */ -=09=09if (dev->mtu > ETH_DATA_LEN) -=09=09=09vi->big_packets =3D true; =09} =20 +=09virtnet_set_big_packets_fields(vi, mtu); + =09if (vi->any_header_sg) =09=09dev->needed_headroom =3D vi->hdr_len; =20 --=20 2.31.1 --------------------------------------------------------------------- To unsubscribe, e-mail: virtio-dev-unsubscribe@lists.oasis-open.org For additional commands, e-mail: virtio-dev-help@lists.oasis-open.org