[PATCH net-next v3 0/2] virtio-net: avoid conflicts between XDP and GUEST_CSUM

[PATCH net-next v3 0/2] virtio-net: avoid conflicts between XDP and GUEST_CSUM

[Heng Qi]

virtio-net needs to clear the VIRTIO_NET_F_GUEST_CSUM feature when
loading XDP. The main reason for doing this is because
VIRTIO_NET_F_GUEST_CSUM allows to receive packets marked as
VIRTIO_NET_HDR_F_NEEDS_CSUM. Such packets are not compatible with
XDP programs, because we cannot guarantee that the csum_{start, offset}
fields are correct after XDP modifies the packets.

There is also an existing problem, in the same host vm-vm (eg
[vm]<->[ovs vhost-user]<->[vm]) scenario, loading XDP will cause packet loss.

To solve the above problems, we have discussed in the [1] proposal, and
now try to solve it through the method of reprobing fields suggested
by Jason.

[1] https://lists.oasis-open.org/archives/virtio-dev/202305/msg00318.html

---
v2->v3:
  - Use skb_checksum_setup() instead of virtnet_flow_dissect_udp_tcp().
    The two operations are different, skb_checksum_setup() operates as
    skb->data==>iphdr, and virtnet_flow_dissect_udp_tcp() operates as
    skb->data==>ethhdr. Essentially equivalent. @Jason Wang

v1->v2:
  - Squash v1's patch [1/4] and patch [2/4] into v2's patch [1/3]. @Michael S. Tsirkin
  - Some minor modifications.

Heng Qi (2):
  virtio-net: support coexistence of XDP and GUEST_CSUM
  virtio-net: remove GUEST_CSUM check for XDP

 drivers/net/virtio_net.c | 90 +++++++++++++++++++++++++++++++++-------
 1 file changed, 74 insertions(+), 16 deletions(-)

-- 
2.19.1.6.gb485710b

[PATCH net-next v3 1/2] virtio-net: support coexistence of XDP and GUEST_CSUM

[Heng Qi]

We are now re-probing the csum related fields and trying
to have XDP and RX hw checksum capabilities coexist on the
XDP path. For the benefit of:
1. RX hw checksum capability can be used if XDP is loaded.
2. Avoid packet loss when loading XDP in the vm-vm scenario.

Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
---
v2->v3:
  - Use skb_checksum_setup() instead of virtnet_flow_dissect_udp_tcp().
    Essentially equivalent.

 drivers/net/virtio_net.c | 86 ++++++++++++++++++++++++++++++++++------
 1 file changed, 73 insertions(+), 13 deletions(-)

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 5a7f7a76b920..0a715e0fbc97 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1568,6 +1568,44 @@ static void virtio_skb_set_hash(const struct virtio_net_hdr_v1_hash *hdr_hash,
 	skb_set_hash(skb, __le32_to_cpu(hdr_hash->hash_value), rss_hash_type);
 }
 
+static int virtnet_set_csum_after_xdp(struct virtnet_info *vi,
+				      struct sk_buff *skb,
+				      __u8 flags)
+{
+	int err = 0;
+
+	/* When XDP program is loaded, for example, the vm-vm scenario
+	 * on the same host, packets marked as VIRTIO_NET_HDR_F_NEEDS_CSUM
+	 * will travel. Although these packets are safe from the point of
+	 * view of the vm, to avoid modification by XDP and successful
+	 * forwarding in the upper layer, we re-probe the necessary checksum
+	 * related information: skb->csum_{start, offset}, pseudo-header csum
+	 * using skb_chdcksum_setup().
+	 *
+	 * This benefits us:
+	 * 1. XDP can be loaded when there's _F_GUEST_CSUM.
+	 * 2. The device verifies the checksum of packets, especially
+	 *    benefiting for large packets.
+	 * 3. In the same-host vm-vm scenario, packets marked as
+	 *    VIRTIO_NET_HDR_F_NEEDS_CSUM are no longer dropped after being
+	 *    processed by XDP.
+	 */
+	if (flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
+		/* We don't parse SCTP because virtio-net currently doesn't
+		 * support CRC checksum offloading for SCTP.
+		 */
+		err = skb_checksum_setup(skb, true);
+	} else if (flags & VIRTIO_NET_HDR_F_DATA_VALID) {
+		/* We want to benefit from this: XDP guarantees that packets marked
+		 * as VIRTIO_NET_HDR_F_DATA_VALID still have correct csum after they
+		 * are processed.
+		 */
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+	}
+
+	return err;
+}
+
 static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
 			void *buf, unsigned int len, void **ctx,
 			unsigned int *xdp_xmit,
@@ -1576,6 +1614,7 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
 	struct net_device *dev = vi->dev;
 	struct sk_buff *skb;
 	struct virtio_net_hdr_mrg_rxbuf *hdr;
+	__u8 flags;
 
 	if (unlikely(len < vi->hdr_len + ETH_HLEN)) {
 		pr_debug("%s: short packet %i\n", dev->name, len);
@@ -1584,6 +1623,13 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
 		return;
 	}
 
+	/* Save the flags of the hdr before XDP processes the data.
+	 * It is ok to use this for both mergeable and small modes.
+	 * Because that's what we do now.
+	 */
+	if (unlikely(vi->xdp_enabled))
+		flags = ((struct virtio_net_hdr_mrg_rxbuf *)buf)->hdr.flags;
+
 	if (vi->mergeable_rx_bufs)
 		skb = receive_mergeable(dev, vi, rq, buf, ctx, len, xdp_xmit,
 					stats);
@@ -1595,23 +1641,37 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
 	if (unlikely(!skb))
 		return;
 
-	hdr = skb_vnet_hdr(skb);
-	if (dev->features & NETIF_F_RXHASH && vi->has_rss_hash_report)
-		virtio_skb_set_hash((const struct virtio_net_hdr_v1_hash *)hdr, skb);
-
-	if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
+	if (unlikely(vi->xdp_enabled)) {
+		/* Required to do this before re-probing and calculating
+		 * the pseudo-header checksum.
+		 */
+		skb->protocol = eth_type_trans(skb, dev);
+		skb_reset_network_header(skb);
+		if (virtnet_set_csum_after_xdp(vi, skb, flags) < 0) {
+			pr_debug("%s: errors occurred in setting partial csum",
+				 dev->name);
+			goto frame_err;
+		}
+	} else {
+		hdr = skb_vnet_hdr(skb);
+		if (dev->features & NETIF_F_RXHASH && vi->has_rss_hash_report)
+			virtio_skb_set_hash((const struct virtio_net_hdr_v1_hash *)hdr, skb);
+
+		if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
+
+		if (virtio_net_hdr_to_skb(skb, &hdr->hdr,
+					  virtio_is_little_endian(vi->vdev))) {
+			net_warn_ratelimited("%s: bad gso: type: %u, size: %u\n",
+					     dev->name, hdr->hdr.gso_type,
+					     hdr->hdr.gso_size);
+			goto frame_err;
+		}
 
-	if (virtio_net_hdr_to_skb(skb, &hdr->hdr,
-				  virtio_is_little_endian(vi->vdev))) {
-		net_warn_ratelimited("%s: bad gso: type: %u, size: %u\n",
-				     dev->name, hdr->hdr.gso_type,
-				     hdr->hdr.gso_size);
-		goto frame_err;
+		skb->protocol = eth_type_trans(skb, dev);
 	}
 
 	skb_record_rx_queue(skb, vq2rxq(rq->vq));
-	skb->protocol = eth_type_trans(skb, dev);
 	pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
 		 ntohs(skb->protocol), skb->len, skb->pkt_type);
 
-- 
2.19.1.6.gb485710b

[PATCH net-next v3 2/2] virtio-net: remove GUEST_CSUM check for XDP

[Heng Qi]

XDP and GUEST_CSUM no longer conflict now, so we removed the
check for GUEST_CSUM for XDP loading/unloading.

Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
---
v1->v2:
  - Rewrite the commit log.

 drivers/net/virtio_net.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 0a715e0fbc97..2e4bd9a05c85 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -60,7 +60,6 @@ static const unsigned long guest_offloads[] = {
 	VIRTIO_NET_F_GUEST_TSO6,
 	VIRTIO_NET_F_GUEST_ECN,
 	VIRTIO_NET_F_GUEST_UFO,
-	VIRTIO_NET_F_GUEST_CSUM,
 	VIRTIO_NET_F_GUEST_USO4,
 	VIRTIO_NET_F_GUEST_USO6,
 	VIRTIO_NET_F_GUEST_HDRLEN
@@ -3437,10 +3436,9 @@ static int virtnet_xdp_set(struct net_device *dev, struct bpf_prog *prog,
 	        virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_TSO6) ||
 	        virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_ECN) ||
 		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_UFO) ||
-		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_CSUM) ||
 		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_USO4) ||
 		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_USO6))) {
-		NL_SET_ERR_MSG_MOD(extack, "Can't set XDP while host is implementing GRO_HW/CSUM, disable GRO_HW/CSUM first");
+		NL_SET_ERR_MSG_MOD(extack, "Can't set XDP while host is implementing GRO_HW, disable GRO_HW first");
 		return -EOPNOTSUPP;
 	}
 
-- 
2.19.1.6.gb485710b

Re: [PATCH net-next v3 1/2] virtio-net: support coexistence of XDP and GUEST_CSUM

[Michael S. Tsirkin]

On Mon, Jun 26, 2023 at 08:03:00PM +0800, Heng Qi wrote:
> We are now re-probing the csum related fields and trying
> to have XDP and RX hw checksum capabilities coexist on the
> XDP path. For the benefit of:
> 1. RX hw checksum capability can be used if XDP is loaded.
> 2. Avoid packet loss when loading XDP in the vm-vm scenario.
> 
> Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
> Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> ---
> v2->v3:
>   - Use skb_checksum_setup() instead of virtnet_flow_dissect_udp_tcp().
>     Essentially equivalent.
> 
>  drivers/net/virtio_net.c | 86 ++++++++++++++++++++++++++++++++++------
>  1 file changed, 73 insertions(+), 13 deletions(-)
> 
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 5a7f7a76b920..0a715e0fbc97 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -1568,6 +1568,44 @@ static void virtio_skb_set_hash(const struct virtio_net_hdr_v1_hash *hdr_hash,
>  	skb_set_hash(skb, __le32_to_cpu(hdr_hash->hash_value), rss_hash_type);
>  }
>  
> +static int virtnet_set_csum_after_xdp(struct virtnet_info *vi,
> +				      struct sk_buff *skb,
> +				      __u8 flags)
> +{
> +	int err = 0;
> +
> +	/* When XDP program is loaded, for example, the vm-vm scenario
> +	 * on the same host, packets marked as VIRTIO_NET_HDR_F_NEEDS_CSUM
> +	 * will travel. Although these packets are safe from the point of
> +	 * view of the vm, to avoid modification by XDP and successful
> +	 * forwarding in the upper layer,

why do you want tp avoid forwarding? did you mean
"and to allow forwarding"?

> we re-probe the necessary checksum
> +	 * related information: skb->csum_{start, offset}, pseudo-header csum
> +	 * using skb_chdcksum_setup().

typo

> +	 *
> +	 * This benefits us:

Drop "This benefits us:" - benefits compared to what?

> +	 * 1. XDP can be loaded when there's _F_GUEST_CSUM.
> +	 * 2. The device verifies the checksum of packets, especially
> +	 *    benefiting for large packets.
> +	 * 3. In the same-host vm-vm scenario, packets marked as
> +	 *    VIRTIO_NET_HDR_F_NEEDS_CSUM are no longer dropped after being
> +	 *    processed by XDP.

please rewrite so the text makes sense in the final C file,
not for someone reading the diff. In that cotext it does not
matter that we used to drop packets or that we used to
disable _F_GUEST_CSUM unless you explain
why they had to be dropped previously.


> +	 */
> +	if (flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
> +		/* We don't parse SCTP because virtio-net currently doesn't
> +		 * support CRC checksum offloading for SCTP.

what does this refer to? where does it exclude SCTP?

> +		 */
> +		err = skb_checksum_setup(skb, true);
> +	} else if (flags & VIRTIO_NET_HDR_F_DATA_VALID) {
> +		/* We want to benefit from this: XDP guarantees that packets marked
> +		 * as VIRTIO_NET_HDR_F_DATA_VALID still have correct csum after they
> +		 * are processed.

drop "We want to benefit from this: "

> +		 */
> +		skb->ip_summed = CHECKSUM_UNNECESSARY;
> +	}
> +
> +	return err;
> +}
> +
>  static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
>  			void *buf, unsigned int len, void **ctx,
>  			unsigned int *xdp_xmit,
> @@ -1576,6 +1614,7 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
>  	struct net_device *dev = vi->dev;
>  	struct sk_buff *skb;
>  	struct virtio_net_hdr_mrg_rxbuf *hdr;
> +	__u8 flags;
>  
>  	if (unlikely(len < vi->hdr_len + ETH_HLEN)) {
>  		pr_debug("%s: short packet %i\n", dev->name, len);
> @@ -1584,6 +1623,13 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
>  		return;
>  	}
>  
> +	/* Save the flags of the hdr before XDP processes the data.
> +	 * It is ok to use this for both mergeable and small modes.
> +	 * Because that's what we do now.

What does the last sentence mean?
Instead please explain why this is necessary. what can change the
header?

> +	 */
> +	if (unlikely(vi->xdp_enabled))
> +		flags = ((struct virtio_net_hdr_mrg_rxbuf *)buf)->hdr.flags;
> +
>  	if (vi->mergeable_rx_bufs)
>  		skb = receive_mergeable(dev, vi, rq, buf, ctx, len, xdp_xmit,
>  					stats);
> @@ -1595,23 +1641,37 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
>  	if (unlikely(!skb))
>  		return;
>  
> -	hdr = skb_vnet_hdr(skb);
> -	if (dev->features & NETIF_F_RXHASH && vi->has_rss_hash_report)
> -		virtio_skb_set_hash((const struct virtio_net_hdr_v1_hash *)hdr, skb);
> -
> -	if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
> -		skb->ip_summed = CHECKSUM_UNNECESSARY;
> +	if (unlikely(vi->xdp_enabled)) {
> +		/* Required to do this before re-probing and calculating
> +		 * the pseudo-header checksum.

What if checksum was disabled on device? No need for all the
elaborate hacks then, right?
What about disabling by ethtool?


> +		 */
> +		skb->protocol = eth_type_trans(skb, dev);
> +		skb_reset_network_header(skb);
> +		if (virtnet_set_csum_after_xdp(vi, skb, flags) < 0) {
> +			pr_debug("%s: errors occurred in setting partial csum",
> +				 dev->name);
> +			goto frame_err;
> +		}
> +	} else {
> +		hdr = skb_vnet_hdr(skb);
> +		if (dev->features & NETIF_F_RXHASH && vi->has_rss_hash_report)
> +			virtio_skb_set_hash((const struct virtio_net_hdr_v1_hash *)hdr, skb);
> +
> +		if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
> +			skb->ip_summed = CHECKSUM_UNNECESSARY;
> +
> +		if (virtio_net_hdr_to_skb(skb, &hdr->hdr,
> +					  virtio_is_little_endian(vi->vdev))) {
> +			net_warn_ratelimited("%s: bad gso: type: %u, size: %u\n",
> +					     dev->name, hdr->hdr.gso_type,
> +					     hdr->hdr.gso_size);
> +			goto frame_err;
> +		}
>  
> -	if (virtio_net_hdr_to_skb(skb, &hdr->hdr,
> -				  virtio_is_little_endian(vi->vdev))) {
> -		net_warn_ratelimited("%s: bad gso: type: %u, size: %u\n",
> -				     dev->name, hdr->hdr.gso_type,
> -				     hdr->hdr.gso_size);
> -		goto frame_err;
> +		skb->protocol = eth_type_trans(skb, dev);
>  	}
>  
>  	skb_record_rx_queue(skb, vq2rxq(rq->vq));
> -	skb->protocol = eth_type_trans(skb, dev);
>  	pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
>  		 ntohs(skb->protocol), skb->len, skb->pkt_type);
>  
> -- 
> 2.19.1.6.gb485710b

Re: [PATCH net-next v3 2/2] virtio-net: remove GUEST_CSUM check for XDP

[Michael S. Tsirkin]

On Mon, Jun 26, 2023 at 08:03:01PM +0800, Heng Qi wrote:
> XDP and GUEST_CSUM no longer conflict now, so we removed the

removed -> remove

> check for GUEST_CSUM for XDP loading/unloading.
> 
> Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
> Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> ---
> v1->v2:
>   - Rewrite the commit log.
> 
>  drivers/net/virtio_net.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 0a715e0fbc97..2e4bd9a05c85 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -60,7 +60,6 @@ static const unsigned long guest_offloads[] = {
>  	VIRTIO_NET_F_GUEST_TSO6,
>  	VIRTIO_NET_F_GUEST_ECN,
>  	VIRTIO_NET_F_GUEST_UFO,
> -	VIRTIO_NET_F_GUEST_CSUM,
>  	VIRTIO_NET_F_GUEST_USO4,
>  	VIRTIO_NET_F_GUEST_USO6,
>  	VIRTIO_NET_F_GUEST_HDRLEN
> @@ -3437,10 +3436,9 @@ static int virtnet_xdp_set(struct net_device *dev, struct bpf_prog *prog,
>  	        virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_TSO6) ||
>  	        virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_ECN) ||
>  		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_UFO) ||
> -		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_CSUM) ||
>  		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_USO4) ||
>  		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_USO6))) {
> -		NL_SET_ERR_MSG_MOD(extack, "Can't set XDP while host is implementing GRO_HW/CSUM, disable GRO_HW/CSUM first");
> +		NL_SET_ERR_MSG_MOD(extack, "Can't set XDP while host is implementing GRO_HW, disable GRO_HW first");
>  		return -EOPNOTSUPP;
>  	}
>  
> -- 
> 2.19.1.6.gb485710b

Re: [PATCH net-next v3 1/2] virtio-net: support coexistence of XDP and GUEST_CSUM

[Heng Qi]

On Mon, Jun 26, 2023 at 08:14:04AM -0400, Michael S. Tsirkin wrote:
> On Mon, Jun 26, 2023 at 08:03:00PM +0800, Heng Qi wrote:
> > We are now re-probing the csum related fields and trying
> > to have XDP and RX hw checksum capabilities coexist on the
> > XDP path. For the benefit of:
> > 1. RX hw checksum capability can be used if XDP is loaded.
> > 2. Avoid packet loss when loading XDP in the vm-vm scenario.
> > 
> > Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
> > Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> > ---
> > v2->v3:
> >   - Use skb_checksum_setup() instead of virtnet_flow_dissect_udp_tcp().
> >     Essentially equivalent.
> > 
> >  drivers/net/virtio_net.c | 86 ++++++++++++++++++++++++++++++++++------
> >  1 file changed, 73 insertions(+), 13 deletions(-)
> > 
> > diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> > index 5a7f7a76b920..0a715e0fbc97 100644
> > --- a/drivers/net/virtio_net.c
> > +++ b/drivers/net/virtio_net.c
> > @@ -1568,6 +1568,44 @@ static void virtio_skb_set_hash(const struct virtio_net_hdr_v1_hash *hdr_hash,
> >  	skb_set_hash(skb, __le32_to_cpu(hdr_hash->hash_value), rss_hash_type);
> >  }
> >  
> > +static int virtnet_set_csum_after_xdp(struct virtnet_info *vi,
> > +				      struct sk_buff *skb,
> > +				      __u8 flags)
> > +{
> > +	int err = 0;
> > +
> > +	/* When XDP program is loaded, for example, the vm-vm scenario
> > +	 * on the same host, packets marked as VIRTIO_NET_HDR_F_NEEDS_CSUM
> > +	 * will travel. Although these packets are safe from the point of
> > +	 * view of the vm, to avoid modification by XDP and successful
> > +	 * forwarding in the upper layer,
> 
> why do you want tp avoid forwarding? did you mean
> "and to allow forwarding"?
> 

Yes, I mean "to allow forwarding".

> > we re-probe the necessary checksum
> > +	 * related information: skb->csum_{start, offset}, pseudo-header csum
> > +	 * using skb_chdcksum_setup().
> 
> typo
> 

Will fix.

> > +	 *
> > +	 * This benefits us:
> 
> Drop "This benefits us:" - benefits compared to what?

Benefits compared to not having this patch.

> 
> > +	 * 1. XDP can be loaded when there's _F_GUEST_CSUM.
> > +	 * 2. The device verifies the checksum of packets, especially
> > +	 *    benefiting for large packets.
> > +	 * 3. In the same-host vm-vm scenario, packets marked as
> > +	 *    VIRTIO_NET_HDR_F_NEEDS_CSUM are no longer dropped after being
> > +	 *    processed by XDP.
> 
> please rewrite so the text makes sense in the final C file,
> not for someone reading the diff. In that cotext it does not
> matter that we used to drop packets or that we used to
> disable _F_GUEST_CSUM unless you explain
> why they had to be dropped previously.

Reasonable, will be rewritten from the reader's point of view.

> 
> 
> > +	 */
> > +	if (flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
> > +		/* We don't parse SCTP because virtio-net currently doesn't
> > +		 * support CRC checksum offloading for SCTP.
> 
> what does this refer to?

It means that the SCTP packets received by the virtio-net rx side have
complete checksums instead of only partial checksums (pseudo-header
checksums). Because virtio-net does not provide the NETIF_F_SCTP_CRC feature.

> where does it exclude SCTP?

Please see:
--> skb_checksum_setup()
	--> skb_checksum_setup_ipv4()
        --> skb_checksum_setup_ip()
	        --> IPPROTO_TCP/IPPROTO_UDP

    --> skb_checksum_setup_ipv6()
        --> skb_checksum_setup_ip()
	        --> IPPROTO_TCP/IPPROTO_UDP

> 
> > +		 */
> > +		err = skb_checksum_setup(skb, true);
> > +	} else if (flags & VIRTIO_NET_HDR_F_DATA_VALID) {
> > +		/* We want to benefit from this: XDP guarantees that packets marked
> > +		 * as VIRTIO_NET_HDR_F_DATA_VALID still have correct csum after they
> > +		 * are processed.
> 
> drop "We want to benefit from this: "

Ok.

> 
> > +		 */
> > +		skb->ip_summed = CHECKSUM_UNNECESSARY;
> > +	}
> > +
> > +	return err;
> > +}
> > +
> >  static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> >  			void *buf, unsigned int len, void **ctx,
> >  			unsigned int *xdp_xmit,
> > @@ -1576,6 +1614,7 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> >  	struct net_device *dev = vi->dev;
> >  	struct sk_buff *skb;
> >  	struct virtio_net_hdr_mrg_rxbuf *hdr;
> > +	__u8 flags;
> >  
> >  	if (unlikely(len < vi->hdr_len + ETH_HLEN)) {
> >  		pr_debug("%s: short packet %i\n", dev->name, len);
> > @@ -1584,6 +1623,13 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> >  		return;
> >  	}
> >  
> > +	/* Save the flags of the hdr before XDP processes the data.
> > +	 * It is ok to use this for both mergeable and small modes.
> > +	 * Because that's what we do now.
> 
> What does the last sentence mean?

It means that both mergeable and small modes can use the
virtio_net_hdr_mrg_rxbuf* structure to obtain the flags in it, to remind
readers that this sentence works. Because flags are located at the same
position in the virtio_net_hdr_mrg_rxbuf and virtio_net_hdr structures.

Example of existing code:
--> receive_small_xdp():
      struct virtio_net_hdr_mrg_rxbuf *hdr = buf + header_offset;


> Instead please explain why this is necessary. what can change the
> header?

Loaded XDP may overwrite/modify headers.

> 
> > +	 */
> > +	if (unlikely(vi->xdp_enabled))
> > +		flags = ((struct virtio_net_hdr_mrg_rxbuf *)buf)->hdr.flags;
> > +
> >  	if (vi->mergeable_rx_bufs)
> >  		skb = receive_mergeable(dev, vi, rq, buf, ctx, len, xdp_xmit,
> >  					stats);
> > @@ -1595,23 +1641,37 @@ static void receive_buf(struct virtnet_info *vi, struct receive_queue *rq,
> >  	if (unlikely(!skb))
> >  		return;
> >  
> > -	hdr = skb_vnet_hdr(skb);
> > -	if (dev->features & NETIF_F_RXHASH && vi->has_rss_hash_report)
> > -		virtio_skb_set_hash((const struct virtio_net_hdr_v1_hash *)hdr, skb);
> > -
> > -	if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
> > -		skb->ip_summed = CHECKSUM_UNNECESSARY;
> > +	if (unlikely(vi->xdp_enabled)) {
> > +		/* Required to do this before re-probing and calculating
> > +		 * the pseudo-header checksum.
> 
> What if checksum was disabled on device?

When XDP is loaded, it may modify or overwrite the packet information.
If the packet has a partial checksum, virtio-net needs to recalculate
the partial checksum, it doesn't matter if the device has F_GUEST_CSUM
or not.

If the packet has a complete checksum and the device does not have
F_GUEST_CSUM, virtio-net does nothing and it is up to the stack to
verify the checksum.

>No need for all the elaborate hacks then, right?

No, always required when xdp is loaded.

> What about disabling by ethtool?

No way. guest csum is a [fixed] feature.

Thanks.

> 
> 
> > +		 */
> > +		skb->protocol = eth_type_trans(skb, dev);
> > +		skb_reset_network_header(skb);
> > +		if (virtnet_set_csum_after_xdp(vi, skb, flags) < 0) {
> > +			pr_debug("%s: errors occurred in setting partial csum",
> > +				 dev->name);
> > +			goto frame_err;
> > +		}
> > +	} else {
> > +		hdr = skb_vnet_hdr(skb);
> > +		if (dev->features & NETIF_F_RXHASH && vi->has_rss_hash_report)
> > +			virtio_skb_set_hash((const struct virtio_net_hdr_v1_hash *)hdr, skb);
> > +
> > +		if (hdr->hdr.flags & VIRTIO_NET_HDR_F_DATA_VALID)
> > +			skb->ip_summed = CHECKSUM_UNNECESSARY;
> > +
> > +		if (virtio_net_hdr_to_skb(skb, &hdr->hdr,
> > +					  virtio_is_little_endian(vi->vdev))) {
> > +			net_warn_ratelimited("%s: bad gso: type: %u, size: %u\n",
> > +					     dev->name, hdr->hdr.gso_type,
> > +					     hdr->hdr.gso_size);
> > +			goto frame_err;
> > +		}
> >  
> > -	if (virtio_net_hdr_to_skb(skb, &hdr->hdr,
> > -				  virtio_is_little_endian(vi->vdev))) {
> > -		net_warn_ratelimited("%s: bad gso: type: %u, size: %u\n",
> > -				     dev->name, hdr->hdr.gso_type,
> > -				     hdr->hdr.gso_size);
> > -		goto frame_err;
> > +		skb->protocol = eth_type_trans(skb, dev);
> >  	}
> >  
> >  	skb_record_rx_queue(skb, vq2rxq(rq->vq));
> > -	skb->protocol = eth_type_trans(skb, dev);
> >  	pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
> >  		 ntohs(skb->protocol), skb->len, skb->pkt_type);
> >  
> > -- 
> > 2.19.1.6.gb485710b

Re: [PATCH net-next v3 2/2] virtio-net: remove GUEST_CSUM check for XDP

[Heng Qi]

在 2023/6/26 下午8:14, Michael S. Tsirkin 写道:
> On Mon, Jun 26, 2023 at 08:03:01PM +0800, Heng Qi wrote:
>> XDP and GUEST_CSUM no longer conflict now, so we removed the
> removed -> remove

Will modify.

Thanks.

>
>> check for GUEST_CSUM for XDP loading/unloading.
>>
>> Signed-off-by: Heng Qi <hengqi@linux.alibaba.com>
>> Reviewed-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
>> ---
>> v1->v2:
>>    - Rewrite the commit log.
>>
>>   drivers/net/virtio_net.c | 4 +---
>>   1 file changed, 1 insertion(+), 3 deletions(-)
>>
>> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
>> index 0a715e0fbc97..2e4bd9a05c85 100644
>> --- a/drivers/net/virtio_net.c
>> +++ b/drivers/net/virtio_net.c
>> @@ -60,7 +60,6 @@ static const unsigned long guest_offloads[] = {
>>   	VIRTIO_NET_F_GUEST_TSO6,
>>   	VIRTIO_NET_F_GUEST_ECN,
>>   	VIRTIO_NET_F_GUEST_UFO,
>> -	VIRTIO_NET_F_GUEST_CSUM,
>>   	VIRTIO_NET_F_GUEST_USO4,
>>   	VIRTIO_NET_F_GUEST_USO6,
>>   	VIRTIO_NET_F_GUEST_HDRLEN
>> @@ -3437,10 +3436,9 @@ static int virtnet_xdp_set(struct net_device *dev, struct bpf_prog *prog,
>>   	        virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_TSO6) ||
>>   	        virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_ECN) ||
>>   		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_UFO) ||
>> -		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_CSUM) ||
>>   		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_USO4) ||
>>   		virtio_has_feature(vi->vdev, VIRTIO_NET_F_GUEST_USO6))) {
>> -		NL_SET_ERR_MSG_MOD(extack, "Can't set XDP while host is implementing GRO_HW/CSUM, disable GRO_HW/CSUM first");
>> +		NL_SET_ERR_MSG_MOD(extack, "Can't set XDP while host is implementing GRO_HW, disable GRO_HW first");
>>   		return -EOPNOTSUPP;
>>   	}
>>   
>> -- 
>> 2.19.1.6.gb485710b