On Mon, Dec 4, 2023 at 3:23 PM Xuan Zhuo <xuanzhuo@xxxxxxxxxxxxxxxxx> wrote:
>
> If the xsk is enabling, the xsk tx will share the send queue.
> But the xsk requires that the send queue use the premapped mode.
> So the send queue must support premapped mode.
>
> command: pktgen_sample01_simple.sh -i eth0 -s 16/1400 -d 10.0.0.123 -m 00:16:3e:12:e1:3e -n 0 -p 100
> matchine: ecs.ebmg6e.26xlarge of Aliyun
> cpu: Intel(R) Xeon(R) Platinum 8269CY CPU @ 2.50GHz
> iommu mode: intel_iommu=on iommu.strict=1 iommu=nopt
>
> | iommu off | iommu on
> ----------------------|-----------------------------------------------------
> | 16 | 1400 | 16 | 1400
> ----------------------|-----------------------------------------------------
> Before: |1716796.00 | 1581829.00 | 390756.00 | 374493.00
> After(premapped off): |1733794.00 | 1576259.00 | 390189.00 | 378128.00
> After(premapped on): |1707107.00 | 1562917.00 | 385667.00 | 373584.00
>
> Signed-off-by: Xuan Zhuo <xuanzhuo@xxxxxxxxxxxxxxxxx>
> ---
> drivers/net/virtio/main.c | 116 ++++++++++++++++++++++++++++----
> drivers/net/virtio/virtio_net.h | 2 +
> 2 files changed, 105 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/net/virtio/main.c b/drivers/net/virtio/main.c
> index d81ee75142e0..fa77a85d1d37 100644
> --- a/drivers/net/virtio/main.c
> +++ b/drivers/net/virtio/main.c
> @@ -167,13 +167,32 @@ static struct xdp_frame *ptr_to_xdp(void *ptr)
> return (struct xdp_frame *)((unsigned long)ptr & ~VIRTIO_XDP_FLAG);
> }
>
> +static void virtnet_sq_unmap_buf(struct virtnet_sq *sq, struct virtio_dma_head *dma)
> +{
> + int i;
> +
> + for (i = 0; i < dma->next; ++i)
> + virtqueue_dma_unmap_single_attrs(sq->vq,
> + dma->items[i].addr,
> + dma->items[i].length,
> + DMA_TO_DEVICE, 0);
> + dma->next = 0;
I'd suggest to check vring_use_dma_api() then we can avoid the unnecessary loop.
> +}
> +
> static void __free_old_xmit(struct virtnet_sq *sq, bool in_napi,
> u64 *bytes, u64 *packets)
> {
> + struct virtio_dma_head *dma;
> unsigned int len;
> void *ptr;
>
> - while ((ptr = virtqueue_get_buf(sq->vq, &len)) != NULL) {
> + dma = (typeof(dma))sq->sg;
This seems tricky, people may forget to check the sizeof when they
want to extend the dma in the future.
Can we use union here?
> + dma->num = ARRAY_SIZE(sq->sg);
> + dma->next = 0;
> +
> + while ((ptr = virtqueue_get_buf_ctx_dma(sq->vq, &len, dma, NULL)) != NULL) {
> + virtnet_sq_unmap_buf(sq, dma);
Can we allow passing NULL then core can just don't bother with
collecting DMA information?
> +
> if (!is_xdp_frame(ptr)) {
> struct sk_buff *skb = ptr;
>
> @@ -567,20 +586,79 @@ static void *virtnet_rq_alloc(struct virtnet_rq *rq, u32 size, gfp_t gfp)
> return buf;
> }
>
> -static void virtnet_rq_set_premapped(struct virtnet_info *vi)
> +static void virtnet_set_premapped(struct virtnet_info *vi)
> {
> + struct virtio_dma_head dma;
> + struct virtnet_sq *sq;
> int i;
>
> - /* disable for big mode */
> - if (!vi->mergeable_rx_bufs && vi->big_packets)
> - return;
> + sq = &vi->sq[0];
> +
> + BUG_ON(sizeof(sq->sg) < sizeof(dma) + sizeof(dma.items[0]) * ARRAY_SIZE(sq->sg));
>
> for (i = 0; i < vi->max_queue_pairs; i++) {
> - if (virtqueue_set_dma_premapped(vi->rq[i].vq))
> - continue;
> + if (!virtqueue_set_dma_premapped(vi->sq[i].vq))
> + vi->sq[i].do_dma = true;
> +
> + /* disable for big mode */
> + if (vi->mergeable_rx_bufs || !vi->big_packets) {
> + if (!virtqueue_set_dma_premapped(vi->rq[i].vq))
> + vi->rq[i].do_dma = true;
> + }
> + }
> +}
> +
> +static void virtnet_sq_unmap_sg(struct virtnet_sq *sq, u32 num)
> +{
> + struct scatterlist *sg;
> + u32 i;
> +
> + for (i = 0; i < num; ++i) {
> + sg = &sq->sg[i];
> +
> + virtqueue_dma_unmap_single_attrs(sq->vq,
> + sg->dma_address,
> + sg->length,
> + DMA_TO_DEVICE, 0);
> + }
> +}
> +
> +static int virtnet_sq_map_sg(struct virtnet_sq *sq, u32 num)
> +{
> + struct scatterlist *sg;
> + u32 i;
> +
> + for (i = 0; i < num; ++i) {
> + sg = &sq->sg[i];
> + sg->dma_address = virtqueue_dma_map_single_attrs(sq->vq, sg_virt(sg),
> + sg->length,
> + DMA_TO_DEVICE, 0);
> + if (virtqueue_dma_mapping_error(sq->vq, sg->dma_address))
> + goto err;
> + }
> +
> + return 0;
> +
> +err:
> + virtnet_sq_unmap_sg(sq, i);
> + return -ENOMEM;
> +}
> +
> +static int virtnet_add_outbuf(struct virtnet_sq *sq, u32 num, void *data)
> +{
> + int ret;
>
> - vi->rq[i].do_dma = true;
> + if (sq->do_dma) {
> + ret = virtnet_sq_map_sg(sq, num);
> + if (ret)
> + return -ENOMEM;
> }
> +
> + ret = virtqueue_add_outbuf(sq->vq, sq->sg, num, data, GFP_ATOMIC);
> + if (ret && sq->do_dma)
> + virtnet_sq_unmap_sg(sq, num);
> +
> + return ret;
> }
>
> static void free_old_xmit(struct virtnet_sq *sq, bool in_napi)
> @@ -686,8 +764,7 @@ static int __virtnet_xdp_xmit_one(struct virtnet_info *vi,
> skb_frag_size(frag), skb_frag_off(frag));
> }
>
> - err = virtqueue_add_outbuf(sq->vq, sq->sg, nr_frags + 1,
> - xdp_to_ptr(xdpf), GFP_ATOMIC);
> + err = virtnet_add_outbuf(sq, nr_frags + 1, xdp_to_ptr(xdpf));
> if (unlikely(err))
> return -ENOSPC; /* Caller handle free/refcnt */
>
> @@ -2126,7 +2203,7 @@ static int xmit_skb(struct virtnet_sq *sq, struct sk_buff *skb)
> return num_sg;
> num_sg++;
> }
> - return virtqueue_add_outbuf(sq->vq, sq->sg, num_sg, skb, GFP_ATOMIC);
> + return virtnet_add_outbuf(sq, num_sg, skb);
> }
>
> static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
> @@ -3852,10 +3929,23 @@ static void virtnet_rq_free_unused_buf(struct virtqueue *vq, void *buf)
>
> static void virtnet_sq_free_unused_bufs(struct virtqueue *vq)
> {
> + struct virtnet_info *vi = vq->vdev->priv;
> + struct virtio_dma_head *dma;
> + struct virtnet_sq *sq;
> + int i = vq2txq(vq);
> void *buf;
>
> - while ((buf = virtqueue_detach_unused_buf(vq)) != NULL)
> + sq = &vi->sq[i];
> +
> + dma = (typeof(dma))sq->sg;
> + dma->num = ARRAY_SIZE(sq->sg);
> + dma->next = 0;
> +
> + while ((buf = virtqueue_detach_unused_buf_premapped(vq, dma)) != NULL) {
> + virtnet_sq_unmap_buf(sq, dma);
> +
> virtnet_sq_free_unused_buf(vq, buf);
> + }
> }
>
> static void virtnet_rq_free_unused_bufs(struct virtqueue *vq)
> @@ -4052,7 +4142,7 @@ static int init_vqs(struct virtnet_info *vi)
> if (ret)
> goto err_free;
>
> - virtnet_rq_set_premapped(vi);
> + virtnet_set_premapped(vi);
>
> cpus_read_lock();
> virtnet_set_affinity(vi);
> diff --git a/drivers/net/virtio/virtio_net.h b/drivers/net/virtio/virtio_net.h
> index ebf9f344648a..870855a829a0 100644
> --- a/drivers/net/virtio/virtio_net.h
> +++ b/drivers/net/virtio/virtio_net.h
> @@ -67,6 +67,8 @@ struct virtnet_sq {
>
> /* Record whether sq is in reset state. */
> bool reset;
> +
> + bool do_dma;
Could we have a core helper for this? It can benefit for other driver
like virtio-fs.
Thanks
> };
>
> /* Internal representation of a receive virtqueue */
> --
> 2.32.0.3.g01195cf9f
>
