Skip to contentSkip to conversation

virtio_net: add XDP meta data support

PATCH Applied Yuya Kusakabe 4 participants 8 messages Feb 25, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

v6 cover virtio_net: add XDP meta data support

This patch series has 2 patches.

Patch 1/2: keep vnet header zeroed if XDP is loaded for small buffer With this fix, we do not need to care about vnet header in XDP meta data support for small buffer, even though XDP meta data uses in front of packet as same as vnet header.
It would be best if this patch goes into stable.
This patch is based on the feedback by Jason Wang and Michael S. Tsirkin.
https://lore.kernel.org/netdev/[email protected]/
https://lore.kernel.org/netdev/[email protected]/

Patch 2/2: add XDP meta data support

Thanks to Jason Wang, Daniel Borkmann and Michael S. Tsirkin for the feedback.

Yuya Kusakabe (2):
virtio_net: keep vnet header zeroed if XDP is loaded for small buffer
virtio_net: add XDP meta data support

 drivers/net/virtio_net.c | 56 ++++++++++++++++++++++++----------------
 1 file changed, 34 insertions(+), 22 deletions(-)
Signature · 1 line
2.24.1
#1 lore raw

PATCH v6 1/2 virtio_net: keep vnet header zeroed if XDP is loaded for small buffer

We do not want to care about the vnet header in receive_small() if XDP is loaded, since we can not know whether or not the packet is modified by XDP.

Fixes f6b10209b90d ("virtio-net: switch to use build_skb() for small buffer")
---
 drivers/net/virtio_net.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/virtio_net.c +2 -2
index 2fe7a3188282..f39d0218bdaa 100644
@@ -735,10 +735,10 @@ static struct sk_buff *receive_small(struct net_device *dev,
 	}
 	skb_reserve(skb, headroom - delta);
 	skb_put(skb, len);
-	if (!delta) {
+	if (!xdp_prog) {
 		buf += header_offset;
 		memcpy(skb_vnet_hdr(skb), buf, vi->hdr_len);
-	} /* keep zeroed vnet hdr since packet was changed by bpf */
+	} /* keep zeroed vnet hdr since XDP is loaded */
 
 err:
 	return skb;
Signature · 1 line
2.24.1
#2 lore raw

PATCH v6 2/2 virtio_net: add XDP meta data support

Implement support for transferring XDP meta data into skb for virtio_net driver; before calling into the program, xdp.data_meta points to xdp.data, where on program return with pass verdict, we call into skb_metadata_set().

Tested with the script at
https://github.com/higebu/virtio_net-xdp-metadata-test.

---
 drivers/net/virtio_net.c | 52 ++++++++++++++++++++++++----------------
 1 file changed, 32 insertions(+), 20 deletions(-)
drivers/net/virtio_net.c +32 -20
index f39d0218bdaa..12d115ef5e74 100644
@@ -371,7 +371,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 				   struct receive_queue *rq,
 				   struct page *page, unsigned int offset,
 				   unsigned int len, unsigned int truesize,
-				   bool hdr_valid)
+				   bool hdr_valid, unsigned int metasize)
 {
 	struct sk_buff *skb;
 	struct virtio_net_hdr_mrg_rxbuf *hdr;
@@ -393,6 +393,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 	else
 		hdr_padded_len = sizeof(struct padded_vnet_hdr);
 
+	/* hdr_valid means no XDP, so we can copy the vnet header */
 	if (hdr_valid)
 		memcpy(hdr, p, hdr_len);
 
@@ -405,6 +406,11 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 		copy = skb_tailroom(skb);
 	skb_put_data(skb, p, copy);
 
+	if (metasize) {
+		__skb_pull(skb, metasize);
+		skb_metadata_set(skb, metasize);
+	}
+
 	len -= copy;
 	offset += copy;
 
@@ -450,10 +456,6 @@ static int __virtnet_xdp_xmit_one(struct virtnet_info *vi,
 	struct virtio_net_hdr_mrg_rxbuf *hdr;
 	int err;
 
-	/* virtqueue want to use data area in-front of packet */
-	if (unlikely(xdpf->metasize > 0))
-		return -EOPNOTSUPP;
-
 	if (unlikely(xdpf->headroom < vi->hdr_len))
 		return -EOVERFLOW;
 
@@ -644,6 +646,7 @@ static struct sk_buff *receive_small(struct net_device *dev,
 	unsigned int delta = 0;
 	struct page *xdp_page;
 	int err;
+	unsigned int metasize = 0;
 
 	len -= vi->hdr_len;
 	stats->bytes += len;
@@ -683,8 +686,8 @@ static struct sk_buff *receive_small(struct net_device *dev,
 
 		xdp.data_hard_start = buf + VIRTNET_RX_PAD + vi->hdr_len;
 		xdp.data = xdp.data_hard_start + xdp_headroom;
-		xdp_set_data_meta_invalid(&xdp);
 		xdp.data_end = xdp.data + len;
+		xdp.data_meta = xdp.data;
 		xdp.rxq = &rq->xdp_rxq;
 		orig_data = xdp.data;
 		act = bpf_prog_run_xdp(xdp_prog, &xdp);
@@ -695,6 +698,7 @@ static struct sk_buff *receive_small(struct net_device *dev,
 			/* Recalculate length in case bpf program changed it */
 			delta = orig_data - xdp.data;
 			len = xdp.data_end - xdp.data;
+			metasize = xdp.data - xdp.data_meta;
 			break;
 		case XDP_TX:
 			stats->xdp_tx++;
@@ -740,6 +744,9 @@ static struct sk_buff *receive_small(struct net_device *dev,
 		memcpy(skb_vnet_hdr(skb), buf, vi->hdr_len);
 	} /* keep zeroed vnet hdr since XDP is loaded */
 
+	if (metasize)
+		skb_metadata_set(skb, metasize);
+
 err:
 	return skb;
 
@@ -760,8 +767,8 @@ static struct sk_buff *receive_big(struct net_device *dev,
 				   struct virtnet_rq_stats *stats)
 {
 	struct page *page = buf;
-	struct sk_buff *skb = page_to_skb(vi, rq, page, 0, len,
-					  PAGE_SIZE, true);
+	struct sk_buff *skb =
+		page_to_skb(vi, rq, page, 0, len, PAGE_SIZE, true, 0);
 
 	stats->bytes += len - vi->hdr_len;
 	if (unlikely(!skb))
@@ -793,6 +800,7 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 	unsigned int truesize;
 	unsigned int headroom = mergeable_ctx_to_headroom(ctx);
 	int err;
+	unsigned int metasize = 0;
 
 	head_skb = NULL;
 	stats->bytes += len - vi->hdr_len;
@@ -839,8 +847,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 		data = page_address(xdp_page) + offset;
 		xdp.data_hard_start = data - VIRTIO_XDP_HEADROOM + vi->hdr_len;
 		xdp.data = data + vi->hdr_len;
-		xdp_set_data_meta_invalid(&xdp);
 		xdp.data_end = xdp.data + (len - vi->hdr_len);
+		xdp.data_meta = xdp.data;
 		xdp.rxq = &rq->xdp_rxq;
 
 		act = bpf_prog_run_xdp(xdp_prog, &xdp);
@@ -848,24 +856,27 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 
 		switch (act) {
 		case XDP_PASS:
+			metasize = xdp.data - xdp.data_meta;
+
 			/* recalculate offset to account for any header
-			 * adjustments. Note other cases do not build an
-			 * skb and avoid using offset
+			 * adjustments and minus the metasize to copy the
+			 * metadata in page_to_skb(). Note other cases do not
+			 * build an skb and avoid using offset
 			 */
-			offset = xdp.data -
-					page_address(xdp_page) - vi->hdr_len;
+			offset = xdp.data - page_address(xdp_page) -
+				 vi->hdr_len - metasize;
 
-			/* recalculate len if xdp.data or xdp.data_end were
-			 * adjusted
+			/* recalculate len if xdp.data, xdp.data_end or
+			 * xdp.data_meta were adjusted
 			 */
-			len = xdp.data_end - xdp.data + vi->hdr_len;
+			len = xdp.data_end - xdp.data + vi->hdr_len + metasize;
 			/* We can only create skb based on xdp_page. */
 			if (unlikely(xdp_page != page)) {
 				rcu_read_unlock();
 				put_page(page);
-				head_skb = page_to_skb(vi, rq, xdp_page,
-						       offset, len,
-						       PAGE_SIZE, false);
+				head_skb = page_to_skb(vi, rq, xdp_page, offset,
+						       len, PAGE_SIZE, false,
+						       metasize);
 				return head_skb;
 			}
 			break;
@@ -921,7 +932,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 		goto err_skb;
 	}
 
-	head_skb = page_to_skb(vi, rq, page, offset, len, truesize, !xdp_prog);
+	head_skb = page_to_skb(vi, rq, page, offset, len, truesize, !xdp_prog,
+			       metasize);
 	curr_skb = head_skb;
 
 	if (unlikely(!curr_skb))
Signature · 1 line
2.24.1
maintainer Acked-by on v6 2/2
#4 lore raw
In reply to Yuya Kusakabe
Quoted 10 lines · Yuya Kusakabe

Implement support for transferring XDP meta data into skb for virtio_net driver; before calling into the program, xdp.data_meta points to xdp.data, where on program return with pass verdict, we call into skb_metadata_set().

Tested with the script at
https://github.com/higebu/virtio_net-xdp-metadata-test.

---

Quoted 162 lines
 drivers/net/virtio_net.c | 52 ++++++++++++++++++++++++----------------
 1 file changed, 32 insertions(+), 20 deletions(-)
drivers/net/virtio_net.c +32 -20
index f39d0218bdaa..12d115ef5e74 100644
@@ -371,7 +371,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 				   struct receive_queue *rq,
 				   struct page *page, unsigned int offset,
 				   unsigned int len, unsigned int truesize,
-				   bool hdr_valid)
+				   bool hdr_valid, unsigned int metasize)
 {
 	struct sk_buff *skb;
 	struct virtio_net_hdr_mrg_rxbuf *hdr;
@@ -393,6 +393,7 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 	else
 		hdr_padded_len = sizeof(struct padded_vnet_hdr);
 
+	/* hdr_valid means no XDP, so we can copy the vnet header */
 	if (hdr_valid)
 		memcpy(hdr, p, hdr_len);
 
@@ -405,6 +406,11 @@ static struct sk_buff *page_to_skb(struct virtnet_info *vi,
 		copy = skb_tailroom(skb);
 	skb_put_data(skb, p, copy);
 
+	if (metasize) {
+		__skb_pull(skb, metasize);
+		skb_metadata_set(skb, metasize);
+	}
+
 	len -= copy;
 	offset += copy;
 
@@ -450,10 +456,6 @@ static int __virtnet_xdp_xmit_one(struct virtnet_info *vi,
 	struct virtio_net_hdr_mrg_rxbuf *hdr;
 	int err;
 
-	/* virtqueue want to use data area in-front of packet */
-	if (unlikely(xdpf->metasize > 0))
-		return -EOPNOTSUPP;
-
 	if (unlikely(xdpf->headroom < vi->hdr_len))
 		return -EOVERFLOW;
 
@@ -644,6 +646,7 @@ static struct sk_buff *receive_small(struct net_device *dev,
 	unsigned int delta = 0;
 	struct page *xdp_page;
 	int err;
+	unsigned int metasize = 0;
 
 	len -= vi->hdr_len;
 	stats->bytes += len;
@@ -683,8 +686,8 @@ static struct sk_buff *receive_small(struct net_device *dev,
 
 		xdp.data_hard_start = buf + VIRTNET_RX_PAD + vi->hdr_len;
 		xdp.data = xdp.data_hard_start + xdp_headroom;
-		xdp_set_data_meta_invalid(&xdp);
 		xdp.data_end = xdp.data + len;
+		xdp.data_meta = xdp.data;
 		xdp.rxq = &rq->xdp_rxq;
 		orig_data = xdp.data;
 		act = bpf_prog_run_xdp(xdp_prog, &xdp);
@@ -695,6 +698,7 @@ static struct sk_buff *receive_small(struct net_device *dev,
 			/* Recalculate length in case bpf program changed it */
 			delta = orig_data - xdp.data;
 			len = xdp.data_end - xdp.data;
+			metasize = xdp.data - xdp.data_meta;
 			break;
 		case XDP_TX:
 			stats->xdp_tx++;
@@ -740,6 +744,9 @@ static struct sk_buff *receive_small(struct net_device *dev,
 		memcpy(skb_vnet_hdr(skb), buf, vi->hdr_len);
 	} /* keep zeroed vnet hdr since XDP is loaded */
 
+	if (metasize)
+		skb_metadata_set(skb, metasize);
+
 err:
 	return skb;
 
@@ -760,8 +767,8 @@ static struct sk_buff *receive_big(struct net_device *dev,
 				   struct virtnet_rq_stats *stats)
 {
 	struct page *page = buf;
-	struct sk_buff *skb = page_to_skb(vi, rq, page, 0, len,
-					  PAGE_SIZE, true);
+	struct sk_buff *skb =
+		page_to_skb(vi, rq, page, 0, len, PAGE_SIZE, true, 0);
 
 	stats->bytes += len - vi->hdr_len;
 	if (unlikely(!skb))
@@ -793,6 +800,7 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 	unsigned int truesize;
 	unsigned int headroom = mergeable_ctx_to_headroom(ctx);
 	int err;
+	unsigned int metasize = 0;
 
 	head_skb = NULL;
 	stats->bytes += len - vi->hdr_len;
@@ -839,8 +847,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 		data = page_address(xdp_page) + offset;
 		xdp.data_hard_start = data - VIRTIO_XDP_HEADROOM + vi->hdr_len;
 		xdp.data = data + vi->hdr_len;
-		xdp_set_data_meta_invalid(&xdp);
 		xdp.data_end = xdp.data + (len - vi->hdr_len);
+		xdp.data_meta = xdp.data;
 		xdp.rxq = &rq->xdp_rxq;
 
 		act = bpf_prog_run_xdp(xdp_prog, &xdp);
@@ -848,24 +856,27 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 
 		switch (act) {
 		case XDP_PASS:
+			metasize = xdp.data - xdp.data_meta;
+
 			/* recalculate offset to account for any header
-			 * adjustments. Note other cases do not build an
-			 * skb and avoid using offset
+			 * adjustments and minus the metasize to copy the
+			 * metadata in page_to_skb(). Note other cases do not
+			 * build an skb and avoid using offset
 			 */
-			offset = xdp.data -
-					page_address(xdp_page) - vi->hdr_len;
+			offset = xdp.data - page_address(xdp_page) -
+				 vi->hdr_len - metasize;
 
-			/* recalculate len if xdp.data or xdp.data_end were
-			 * adjusted
+			/* recalculate len if xdp.data, xdp.data_end or
+			 * xdp.data_meta were adjusted
 			 */
-			len = xdp.data_end - xdp.data + vi->hdr_len;
+			len = xdp.data_end - xdp.data + vi->hdr_len + metasize;
 			/* We can only create skb based on xdp_page. */
 			if (unlikely(xdp_page != page)) {
 				rcu_read_unlock();
 				put_page(page);
-				head_skb = page_to_skb(vi, rq, xdp_page,
-						       offset, len,
-						       PAGE_SIZE, false);
+				head_skb = page_to_skb(vi, rq, xdp_page, offset,
+						       len, PAGE_SIZE, false,
+						       metasize);
 				return head_skb;
 			}
 			break;
@@ -921,7 +932,8 @@ static struct sk_buff *receive_mergeable(struct net_device *dev,
 		goto err_skb;
 	}
 
-	head_skb = page_to_skb(vi, rq, page, offset, len, truesize, !xdp_prog);
+	head_skb = page_to_skb(vi, rq, page, offset, len, truesize, !xdp_prog,
+			       metasize);
 	curr_skb = head_skb;
 
 	if (unlikely(!curr_skb))
Signature · 1 line
2.24.1
Acked-by on v6 1/2
#5 lore raw

Re: [PATCH bpf-next v6 1/2] virtio_net: keep vnet header zeroed if XDP is loaded for small buffer

In reply to Yuya Kusakabe
Quoted 17 lines · Yuya Kusakabe

We do not want to care about the vnet header in receive_small() if XDP is loaded, since we can not know whether or not the packet is modified by XDP.

Fixes f6b10209b90d ("virtio-net: switch to use build_skb() for small buffer")
---
  drivers/net/virtio_net.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/virtio_net.c +0 -0
index 2fe7a3188282..f39d0218bdaa 100644
@@ -735,10 +735,10 @@ static struct sk_buff *receive_small(struct net_device *dev,
  	}
  	skb_reserve(skb, headroom - delta);

drivers/net/virtio_net.c static struct sk_buff *receive_small(struct net_device *dev, · line 744

Quoted 10 lines
drivers/net/virtio_net.c +2 -2
  	skb_put(skb, len);
-	if (!delta) {
+	if (!xdp_prog) {
  		buf += header_offset;
  		memcpy(skb_vnet_hdr(skb), buf, vi->hdr_len);
-	} /* keep zeroed vnet hdr since packet was changed by bpf */
+	} /* keep zeroed vnet hdr since XDP is loaded */
 
  err:
  	return skb;
maintainer Acked-by on v6 1/2
#6 lore raw

Re: [PATCH bpf-next v6 1/2] virtio_net: keep vnet header zeroed if XDP is loaded for small buffer

In reply to Yuya Kusakabe
Quoted 6 lines · Yuya Kusakabe

We do not want to care about the vnet header in receive_small() if XDP is loaded, since we can not know whether or not the packet is modified by XDP.

Fixes f6b10209b90d ("virtio-net: switch to use build_skb() for small buffer")
Quoted 23 lines
---
 drivers/net/virtio_net.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/virtio_net.c +2 -2
index 2fe7a3188282..f39d0218bdaa 100644
@@ -735,10 +735,10 @@ static struct sk_buff *receive_small(struct net_device *dev,
 	}
 	skb_reserve(skb, headroom - delta);
 	skb_put(skb, len);
-	if (!delta) {
+	if (!xdp_prog) {
 		buf += header_offset;
 		memcpy(skb_vnet_hdr(skb), buf, vi->hdr_len);
-	} /* keep zeroed vnet hdr since packet was changed by bpf */
+	} /* keep zeroed vnet hdr since XDP is loaded */
 
 err:
 	return skb;
Signature · 1 line
2.24.1
maintainer Applied
#7 lore raw
In reply to Yuya Kusakabe
Quoted 13 lines · Yuya Kusakabe

This patch series has 2 patches.

Patch 1/2: keep vnet header zeroed if XDP is loaded for small buffer With this fix, we do not need to care about vnet header in XDP meta data support for small buffer, even though XDP meta data uses in front of packet as same as vnet header.
It would be best if this patch goes into stable.
This patch is based on the feedback by Jason Wang and Michael S. Tsirkin.
https://lore.kernel.org/netdev/[email protected]/
https://lore.kernel.org/netdev/[email protected]/

Patch 2/2: add XDP meta data support

Quoted 9 lines

Thanks to Jason Wang, Daniel Borkmann and Michael S. Tsirkin for the feedback.

Yuya Kusakabe (2):
virtio_net: keep vnet header zeroed if XDP is loaded for small buffer
virtio_net: add XDP meta data support

  drivers/net/virtio_net.c | 56 ++++++++++++++++++++++++----------------
  1 file changed, 34 insertions(+), 22 deletions(-)

Applied, thanks!