[PATCH net-next v14 0/3] xen networking: add XDP support to xen-netfront

[PATCH net-next v14 0/3] xen networking: add XDP support to xen-netfront

[Denis Kirjanov]

The first patch adds a new extra type to enable proper synchronization
between an RX request/response pair.
The second patch implements BFP interface for xen-netfront.
The third patch enables extra space for XDP processing.

v14:
- fixed compilation warnings

v13:
- fixed compilation due to previous rename

v12:
- xen-netback: rename netfront_xdp_headroom to xdp_headroom

v11:
- add the new headroom constant to netif.h
- xenbus_scanf check
- lock a bulk of puckets in xennet_xdp_xmit()

v10:
- add a new xen_netif_extra_info type to enable proper synchronization
 between an RX request/response pair.
- order local variable declarations

v9:
- assign an xdp program before switching to Reconfiguring
- minor cleanups
- address checkpatch issues

v8:
- add PAGE_POOL config dependency
- keep the state of XDP processing in netfront_xdp_enabled
- fixed allocator type in xdp_rxq_info_reg_mem_model()
- minor cleanups in xen-netback

v7:
- use page_pool_dev_alloc_pages() on page allocation
- remove the leftover break statement from netback_changed

v6:
- added the missing SOB line
- fixed subject

v5:
- split netfront/netback changes
- added a sync point between backend/frontend on switching to XDP
- added pagepool API

v4:
- added verbose patch descriprion
- don't expose the XDP headroom offset to the domU guest
- add a modparam to netback to toggle XDP offset
- don't process jumbo frames for now

v3:
- added XDP_TX support (tested with xdping echoserver)
- added XDP_REDIRECT support (tested with modified xdp_redirect_kern)
- moved xdp negotiation to xen-netback

v2:
- avoid data copying while passing to XDP
- tell xen-netback that we need the headroom space

Denis Kirjanov (3):
  xen: netif.h: add a new extra type for XDP
  xen networking: add basic XDP support for xen-netfront
  xen networking: add XDP offset adjustment to xen-netback

 drivers/net/Kconfig                 |   1 +
 drivers/net/xen-netback/common.h    |   4 +
 drivers/net/xen-netback/interface.c |   2 +
 drivers/net/xen-netback/netback.c   |   7 +
 drivers/net/xen-netback/rx.c        |  15 +-
 drivers/net/xen-netback/xenbus.c    |  34 ++++
 drivers/net/xen-netfront.c          | 336 ++++++++++++++++++++++++++++++++++--
 include/xen/interface/io/netif.h    |  20 ++-
 8 files changed, 407 insertions(+), 12 deletions(-)

-- 
1.8.3.1

[PATCH net-next v14 1/3] xen: netif.h: add a new extra type for XDP

[Denis Kirjanov]

The patch adds a new extra type to be able to diffirentiate
between RX responses on xen-netfront side with the adjusted offset
required for XDP processing.

The offset value from a guest is passed via xenstore.

Signed-off-by: Denis Kirjanov <kda@linux-powerpc.org>
---
 include/xen/interface/io/netif.h | 20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)

diff --git a/include/xen/interface/io/netif.h b/include/xen/interface/io/netif.h
index 4f20dbc..2194322 100644
--- a/include/xen/interface/io/netif.h
+++ b/include/xen/interface/io/netif.h
@@ -161,6 +161,19 @@
  */
 
 /*
+ * "xdp-headroom" is used to request that extra space is added
+ * for XDP processing.  The value is measured in bytes and passed by
+ * the frontend to be consistent between both ends.
+ * If the value is greater than zero that means that
+ * an RX response is going to be passed to an XDP program for processing.
+ * XEN_NETIF_MAX_XDP_HEADROOM defines the maximum headroom offset in bytes
+ *
+ * "feature-xdp-headroom" is set to "1" by the netback side like other features
+ * so a guest can check if an XDP program can be processed.
+ */
+#define XEN_NETIF_MAX_XDP_HEADROOM 0x7FFF
+
+/*
  * Control ring
  * ============
  *
@@ -846,7 +859,8 @@ struct xen_netif_tx_request {
 #define XEN_NETIF_EXTRA_TYPE_MCAST_ADD (2)	/* u.mcast */
 #define XEN_NETIF_EXTRA_TYPE_MCAST_DEL (3)	/* u.mcast */
 #define XEN_NETIF_EXTRA_TYPE_HASH      (4)	/* u.hash */
-#define XEN_NETIF_EXTRA_TYPE_MAX       (5)
+#define XEN_NETIF_EXTRA_TYPE_XDP       (5)	/* u.xdp */
+#define XEN_NETIF_EXTRA_TYPE_MAX       (6)
 
 /* xen_netif_extra_info_t flags. */
 #define _XEN_NETIF_EXTRA_FLAG_MORE (0)
@@ -879,6 +893,10 @@ struct xen_netif_extra_info {
 			uint8_t algorithm;
 			uint8_t value[4];
 		} hash;
+		struct {
+			uint16_t headroom;
+			uint16_t pad[2];
+		} xdp;
 		uint16_t pad[3];
 	} u;
 };
-- 
1.8.3.1

[PATCH net-next v14 2/3] xen networking: add basic XDP support for xen-netfront

[Denis Kirjanov]

The patch adds a basic XDP processing to xen-netfront driver.

We ran an XDP program for an RX response received from netback
driver. Also we request xen-netback to adjust data offset for
bpf_xdp_adjust_head() header space for custom headers.

synchronization between frontend and backend parts is done
by using xenbus state switching:
Reconfiguring -> Reconfigured- > Connected

UDP packets drop rate using xdp program is around 310 kpps
using ./pktgen_sample04_many_flows.sh and 160 kpps without the patch.

Signed-off-by: Denis Kirjanov <kda@linux-powerpc.org>
---
 drivers/net/Kconfig        |   1 +
 drivers/net/xen-netfront.c | 336 +++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 327 insertions(+), 10 deletions(-)

diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig
index 9a49c4c..717abf4 100644
--- a/drivers/net/Kconfig
+++ b/drivers/net/Kconfig
@@ -495,6 +495,7 @@ config XEN_NETDEV_FRONTEND
 	tristate "Xen network device frontend driver"
 	depends on XEN
 	select XEN_XENBUS_FRONTEND
+	select PAGE_POOL
 	default y
 	help
 	  This driver provides support for Xen paravirtual network
diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c
index 482c6c8..468f3f6 100644
--- a/drivers/net/xen-netfront.c
+++ b/drivers/net/xen-netfront.c
@@ -44,6 +44,9 @@
 #include <linux/mm.h>
 #include <linux/slab.h>
 #include <net/ip.h>
+#include <linux/bpf.h>
+#include <net/page_pool.h>
+#include <linux/bpf_trace.h>
 
 #include <xen/xen.h>
 #include <xen/xenbus.h>
@@ -102,6 +105,8 @@ struct netfront_queue {
 	char name[QUEUE_NAME_SIZE]; /* DEVNAME-qN */
 	struct netfront_info *info;
 
+	struct bpf_prog __rcu *xdp_prog;
+
 	struct napi_struct napi;
 
 	/* Split event channels support, tx_* == rx_* when using
@@ -144,6 +149,9 @@ struct netfront_queue {
 	struct sk_buff *rx_skbs[NET_RX_RING_SIZE];
 	grant_ref_t gref_rx_head;
 	grant_ref_t grant_rx_ref[NET_RX_RING_SIZE];
+
+	struct page_pool *page_pool;
+	struct xdp_rxq_info xdp_rxq;
 };
 
 struct netfront_info {
@@ -159,6 +167,10 @@ struct netfront_info {
 	struct netfront_stats __percpu *rx_stats;
 	struct netfront_stats __percpu *tx_stats;
 
+	/* XDP state */
+	bool netback_has_xdp_headroom;
+	bool netfront_xdp_enabled;
+
 	atomic_t rx_gso_checksum_fixup;
 };
 
@@ -265,8 +277,8 @@ static struct sk_buff *xennet_alloc_one_rx_buffer(struct netfront_queue *queue)
 	if (unlikely(!skb))
 		return NULL;
 
-	page = alloc_page(GFP_ATOMIC | __GFP_NOWARN);
-	if (!page) {
+	page = page_pool_dev_alloc_pages(queue->page_pool);
+	if (unlikely(!page)) {
 		kfree_skb(skb);
 		return NULL;
 	}
@@ -560,6 +572,66 @@ static u16 xennet_select_queue(struct net_device *dev, struct sk_buff *skb,
 	return queue_idx;
 }
 
+static int xennet_xdp_xmit_one(struct net_device *dev,
+			       struct netfront_queue *queue,
+			       struct xdp_frame *xdpf)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	struct netfront_stats *tx_stats = this_cpu_ptr(np->tx_stats);
+	int notify;
+
+	xennet_make_first_txreq(queue, NULL,
+				virt_to_page(xdpf->data),
+				offset_in_page(xdpf->data),
+				xdpf->len);
+
+	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&queue->tx, notify);
+	if (notify)
+		notify_remote_via_irq(queue->tx_irq);
+
+	u64_stats_update_begin(&tx_stats->syncp);
+	tx_stats->bytes += xdpf->len;
+	tx_stats->packets++;
+	u64_stats_update_end(&tx_stats->syncp);
+
+	xennet_tx_buf_gc(queue);
+
+	return 0;
+}
+
+static int xennet_xdp_xmit(struct net_device *dev, int n,
+			   struct xdp_frame **frames, u32 flags)
+{
+	unsigned int num_queues = dev->real_num_tx_queues;
+	struct netfront_info *np = netdev_priv(dev);
+	struct netfront_queue *queue = NULL;
+	unsigned long irq_flags;
+	int drops = 0;
+	int i, err;
+
+	if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK))
+		return -EINVAL;
+
+	queue = &np->queues[smp_processor_id() % num_queues];
+
+	spin_lock_irqsave(&queue->tx_lock, irq_flags);
+	for (i = 0; i < n; i++) {
+		struct xdp_frame *xdpf = frames[i];
+
+		if (!xdpf)
+			continue;
+		err = xennet_xdp_xmit_one(dev, queue, xdpf);
+		if (err) {
+			xdp_return_frame_rx_napi(xdpf);
+			drops++;
+		}
+	}
+	spin_unlock_irqrestore(&queue->tx_lock, irq_flags);
+
+	return n - drops;
+}
+
+
 #define MAX_XEN_SKB_FRAGS (65536 / XEN_PAGE_SIZE + 1)
 
 static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev)
@@ -778,23 +850,82 @@ static int xennet_get_extras(struct netfront_queue *queue,
 	return err;
 }
 
+static u32 xennet_run_xdp(struct netfront_queue *queue, struct page *pdata,
+		   struct xen_netif_rx_response *rx, struct bpf_prog *prog,
+		   struct xdp_buff *xdp, bool *need_xdp_flush)
+{
+	struct xdp_frame *xdpf;
+	u32 len = rx->status;
+	u32 act = XDP_PASS;
+	int err;
+
+	xdp->data_hard_start = page_address(pdata);
+	xdp->data = xdp->data_hard_start + XDP_PACKET_HEADROOM;
+	xdp_set_data_meta_invalid(xdp);
+	xdp->data_end = xdp->data + len;
+	xdp->rxq = &queue->xdp_rxq;
+	xdp->frame_sz = XEN_PAGE_SIZE - XDP_PACKET_HEADROOM;
+
+	act = bpf_prog_run_xdp(prog, xdp);
+	switch (act) {
+	case XDP_TX:
+		get_page(pdata);
+		xdpf = xdp_convert_buff_to_frame(xdp);
+		err = xennet_xdp_xmit(queue->info->netdev, 1, &xdpf, 0);
+		if (unlikely(err < 0))
+			trace_xdp_exception(queue->info->netdev, prog, act);
+		break;
+	case XDP_REDIRECT:
+		get_page(pdata);
+		err = xdp_do_redirect(queue->info->netdev, xdp, prog);
+		*need_xdp_flush = true;
+		if (unlikely(err))
+			trace_xdp_exception(queue->info->netdev, prog, act);
+		break;
+	case XDP_PASS:
+	case XDP_DROP:
+		break;
+
+	case XDP_ABORTED:
+		trace_xdp_exception(queue->info->netdev, prog, act);
+		break;
+
+	default:
+		bpf_warn_invalid_xdp_action(act);
+	}
+
+	return act;
+}
+
 static int xennet_get_responses(struct netfront_queue *queue,
 				struct netfront_rx_info *rinfo, RING_IDX rp,
-				struct sk_buff_head *list)
+				struct sk_buff_head *list,
+				bool *need_xdp_flush)
 {
 	struct xen_netif_rx_response *rx = &rinfo->rx;
-	struct xen_netif_extra_info *extras = rinfo->extras;
-	struct device *dev = &queue->info->netdev->dev;
+	int max = XEN_NETIF_NR_SLOTS_MIN + (rx->status <= RX_COPY_THRESHOLD);
 	RING_IDX cons = queue->rx.rsp_cons;
 	struct sk_buff *skb = xennet_get_rx_skb(queue, cons);
+	struct xen_netif_extra_info *extras = rinfo->extras;
 	grant_ref_t ref = xennet_get_rx_ref(queue, cons);
-	int max = XEN_NETIF_NR_SLOTS_MIN + (rx->status <= RX_COPY_THRESHOLD);
+	struct device *dev = &queue->info->netdev->dev;
+	struct bpf_prog *xdp_prog;
+	struct xdp_buff xdp;
+	unsigned long ret;
 	int slots = 1;
 	int err = 0;
-	unsigned long ret;
+	u32 verdict;
 
 	if (rx->flags & XEN_NETRXF_extra_info) {
 		err = xennet_get_extras(queue, extras, rp);
+		if (!err) {
+			if (extras[XEN_NETIF_EXTRA_TYPE_XDP - 1].type) {
+				struct xen_netif_extra_info *xdp;
+
+				xdp = &extras[XEN_NETIF_EXTRA_TYPE_XDP - 1];
+				rx->offset = xdp->u.xdp.headroom;
+			}
+		}
 		cons = queue->rx.rsp_cons;
 	}
 
@@ -827,9 +958,24 @@ static int xennet_get_responses(struct netfront_queue *queue,
 
 		gnttab_release_grant_reference(&queue->gref_rx_head, ref);
 
-		__skb_queue_tail(list, skb);
-
+		rcu_read_lock();
+		xdp_prog = rcu_dereference(queue->xdp_prog);
+		if (xdp_prog) {
+			if (!(rx->flags & XEN_NETRXF_more_data)) {
+				/* currently only a single page contains data */
+				verdict = xennet_run_xdp(queue,
+							 skb_frag_page(&skb_shinfo(skb)->frags[0]),
+							 rx, xdp_prog, &xdp, need_xdp_flush);
+				if (verdict != XDP_PASS)
+					err = -EINVAL;
+			} else {
+				/* drop the frame */
+				err = -EINVAL;
+			}
+		}
+		rcu_read_unlock();
 next:
+		__skb_queue_tail(list, skb);
 		if (!(rx->flags & XEN_NETRXF_more_data))
 			break;
 
@@ -998,6 +1144,7 @@ static int xennet_poll(struct napi_struct *napi, int budget)
 	struct sk_buff_head errq;
 	struct sk_buff_head tmpq;
 	int err;
+	bool need_xdp_flush = false;
 
 	spin_lock(&queue->rx_lock);
 
@@ -1014,7 +1161,8 @@ static int xennet_poll(struct napi_struct *napi, int budget)
 		memcpy(rx, RING_GET_RESPONSE(&queue->rx, i), sizeof(*rx));
 		memset(extras, 0, sizeof(rinfo.extras));
 
-		err = xennet_get_responses(queue, &rinfo, rp, &tmpq);
+		err = xennet_get_responses(queue, &rinfo, rp, &tmpq,
+					   &need_xdp_flush);
 
 		if (unlikely(err)) {
 err:
@@ -1060,6 +1208,8 @@ static int xennet_poll(struct napi_struct *napi, int budget)
 		i = ++queue->rx.rsp_cons;
 		work_done++;
 	}
+	if (need_xdp_flush)
+		xdp_do_flush();
 
 	__skb_queue_purge(&errq);
 
@@ -1261,6 +1411,101 @@ static void xennet_poll_controller(struct net_device *dev)
 }
 #endif
 
+#define NETBACK_XDP_HEADROOM_DISABLE	0
+#define NETBACK_XDP_HEADROOM_ENABLE	1
+
+static int talk_to_netback_xdp(struct netfront_info *np, int xdp)
+{
+	int err;
+	unsigned short headroom;
+
+	headroom = xdp ? XDP_PACKET_HEADROOM : 0;
+	err = xenbus_printf(XBT_NIL, np->xbdev->nodename,
+			    "xdp-headroom", "%hu",
+			    headroom);
+	if (err)
+		pr_warn("Error writing xdp-headroom\n");
+
+	return err;
+}
+
+static int xennet_xdp_set(struct net_device *dev, struct bpf_prog *prog,
+			  struct netlink_ext_ack *extack)
+{
+	unsigned long max_mtu = XEN_PAGE_SIZE - XDP_PACKET_HEADROOM;
+	struct netfront_info *np = netdev_priv(dev);
+	struct bpf_prog *old_prog;
+	unsigned int i, err;
+
+	if (dev->mtu > max_mtu) {
+		netdev_warn(dev, "XDP requires MTU less than %lu\n", max_mtu);
+		return -EINVAL;
+	}
+
+	if (!np->netback_has_xdp_headroom)
+		return 0;
+
+	xenbus_switch_state(np->xbdev, XenbusStateReconfiguring);
+
+	err = talk_to_netback_xdp(np, prog ? NETBACK_XDP_HEADROOM_ENABLE :
+				  NETBACK_XDP_HEADROOM_DISABLE);
+	if (err)
+		return err;
+
+	/* avoid the race with XDP headroom adjustment */
+	wait_event(module_wq,
+		   xenbus_read_driver_state(np->xbdev->otherend) ==
+		   XenbusStateReconfigured);
+	np->netfront_xdp_enabled = true;
+
+	old_prog = rtnl_dereference(np->queues[0].xdp_prog);
+
+	if (prog)
+		bpf_prog_add(prog, dev->real_num_tx_queues);
+
+	for (i = 0; i < dev->real_num_tx_queues; ++i)
+		rcu_assign_pointer(np->queues[i].xdp_prog, prog);
+
+	if (old_prog)
+		for (i = 0; i < dev->real_num_tx_queues; ++i)
+			bpf_prog_put(old_prog);
+
+	xenbus_switch_state(np->xbdev, XenbusStateConnected);
+
+	return 0;
+}
+
+static u32 xennet_xdp_query(struct net_device *dev)
+{
+	unsigned int num_queues = dev->real_num_tx_queues;
+	struct netfront_info *np = netdev_priv(dev);
+	const struct bpf_prog *xdp_prog;
+	struct netfront_queue *queue;
+	unsigned int i;
+
+	for (i = 0; i < num_queues; ++i) {
+		queue = &np->queues[i];
+		xdp_prog = rtnl_dereference(queue->xdp_prog);
+		if (xdp_prog)
+			return xdp_prog->aux->id;
+	}
+
+	return 0;
+}
+
+static int xennet_xdp(struct net_device *dev, struct netdev_bpf *xdp)
+{
+	switch (xdp->command) {
+	case XDP_SETUP_PROG:
+		return xennet_xdp_set(dev, xdp->prog, xdp->extack);
+	case XDP_QUERY_PROG:
+		xdp->prog_id = xennet_xdp_query(dev);
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
 static const struct net_device_ops xennet_netdev_ops = {
 	.ndo_open            = xennet_open,
 	.ndo_stop            = xennet_close,
@@ -1272,6 +1517,8 @@ static void xennet_poll_controller(struct net_device *dev)
 	.ndo_fix_features    = xennet_fix_features,
 	.ndo_set_features    = xennet_set_features,
 	.ndo_select_queue    = xennet_select_queue,
+	.ndo_bpf            = xennet_xdp,
+	.ndo_xdp_xmit	    = xennet_xdp_xmit,
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller = xennet_poll_controller,
 #endif
@@ -1331,6 +1578,7 @@ static struct net_device *xennet_create_dev(struct xenbus_device *dev)
 	SET_NETDEV_DEV(netdev, &dev->dev);
 
 	np->netdev = netdev;
+	np->netfront_xdp_enabled = false;
 
 	netif_carrier_off(netdev);
 
@@ -1419,6 +1667,8 @@ static void xennet_disconnect_backend(struct netfront_info *info)
 		queue->rx_ring_ref = GRANT_INVALID_REF;
 		queue->tx.sring = NULL;
 		queue->rx.sring = NULL;
+
+		page_pool_destroy(queue->page_pool);
 	}
 }
 
@@ -1754,6 +2004,51 @@ static void xennet_destroy_queues(struct netfront_info *info)
 	info->queues = NULL;
 }
 
+
+
+static int xennet_create_page_pool(struct netfront_queue *queue)
+{
+	int err;
+	struct page_pool_params pp_params = {
+		.order = 0,
+		.flags = 0,
+		.pool_size = NET_RX_RING_SIZE,
+		.nid = NUMA_NO_NODE,
+		.dev = &queue->info->netdev->dev,
+		.offset = XDP_PACKET_HEADROOM,
+		.max_len = XEN_PAGE_SIZE - XDP_PACKET_HEADROOM,
+	};
+
+	queue->page_pool = page_pool_create(&pp_params);
+	if (IS_ERR(queue->page_pool)) {
+		err = PTR_ERR(queue->page_pool);
+		queue->page_pool = NULL;
+		return err;
+	}
+
+	err = xdp_rxq_info_reg(&queue->xdp_rxq, queue->info->netdev,
+			       queue->id);
+	if (err) {
+		netdev_err(queue->info->netdev, "xdp_rxq_info_reg failed\n");
+		goto err_free_pp;
+	}
+
+	err = xdp_rxq_info_reg_mem_model(&queue->xdp_rxq,
+					 MEM_TYPE_PAGE_POOL, queue->page_pool);
+	if (err) {
+		netdev_err(queue->info->netdev, "xdp_rxq_info_reg_mem_model failed\n");
+		goto err_unregister_rxq;
+	}
+	return 0;
+
+err_unregister_rxq:
+	xdp_rxq_info_unreg(&queue->xdp_rxq);
+err_free_pp:
+	page_pool_destroy(queue->page_pool);
+	queue->page_pool = NULL;
+	return err;
+}
+
 static int xennet_create_queues(struct netfront_info *info,
 				unsigned int *num_queues)
 {
@@ -1779,6 +2074,14 @@ static int xennet_create_queues(struct netfront_info *info,
 			break;
 		}
 
+		/* use page pool recycling instead of buddy allocator */
+		ret = xennet_create_page_pool(queue);
+		if (ret < 0) {
+			dev_err(&info->xbdev->dev, "can't allocate page pool\n");
+			*num_queues = i;
+			return ret;
+		}
+
 		netif_napi_add(queue->info->netdev, &queue->napi,
 			       xennet_poll, 64);
 		if (netif_running(info->netdev))
@@ -1825,6 +2128,17 @@ static int talk_to_netback(struct xenbus_device *dev,
 		goto out_unlocked;
 	}
 
+	info->netback_has_xdp_headroom = xenbus_read_unsigned(info->xbdev->otherend,
+							      "feature-xdp-headroom", 0);
+	if (info->netback_has_xdp_headroom) {
+		/* set the current xen-netfront xdp state */
+		err = talk_to_netback_xdp(info, info->netfront_xdp_enabled ?
+					  NETBACK_XDP_HEADROOM_ENABLE :
+					  NETBACK_XDP_HEADROOM_DISABLE);
+		if (err)
+			goto out_unlocked;
+	}
+
 	rtnl_lock();
 	if (info->queues)
 		xennet_destroy_queues(info);
@@ -1959,6 +2273,8 @@ static int xennet_connect(struct net_device *dev)
 	err = talk_to_netback(np->xbdev, np);
 	if (err)
 		return err;
+	if (np->netback_has_xdp_headroom)
+		pr_info("backend supports XDP headroom\n");
 
 	/* talk_to_netback() sets the correct number of queues */
 	num_queues = dev->real_num_tx_queues;
-- 
1.8.3.1

[PATCH net-next v14 3/3] xen networking: add XDP offset adjustment to xen-netback

[Denis Kirjanov]

the patch basically adds the offset adjustment and netfront
state reading to make XDP work on netfront side.

Reviewed-by: Paul Durrant <paul@xen.org>
Signed-off-by: Denis Kirjanov <kda@linux-powerpc.org>
---
 drivers/net/xen-netback/common.h    |  4 ++++
 drivers/net/xen-netback/interface.c |  2 ++
 drivers/net/xen-netback/netback.c   |  7 +++++++
 drivers/net/xen-netback/rx.c        | 15 ++++++++++++++-
 drivers/net/xen-netback/xenbus.c    | 34 ++++++++++++++++++++++++++++++++++
 5 files changed, 61 insertions(+), 1 deletion(-)

diff --git a/drivers/net/xen-netback/common.h b/drivers/net/xen-netback/common.h
index 05847eb..ae477f7 100644
--- a/drivers/net/xen-netback/common.h
+++ b/drivers/net/xen-netback/common.h
@@ -281,6 +281,9 @@ struct xenvif {
 	u8 ipv6_csum:1;
 	u8 multicast_control:1;
 
+	/* headroom requested by xen-netfront */
+	u16 xdp_headroom;
+
 	/* Is this interface disabled? True when backend discovers
 	 * frontend is rogue.
 	 */
@@ -395,6 +398,7 @@ static inline pending_ring_idx_t nr_pending_reqs(struct xenvif_queue *queue)
 irqreturn_t xenvif_interrupt(int irq, void *dev_id);
 
 extern bool separate_tx_rx_irq;
+extern bool provides_xdp_headroom;
 
 extern unsigned int rx_drain_timeout_msecs;
 extern unsigned int rx_stall_timeout_msecs;
diff --git a/drivers/net/xen-netback/interface.c b/drivers/net/xen-netback/interface.c
index 0c8a02a..8af49728 100644
--- a/drivers/net/xen-netback/interface.c
+++ b/drivers/net/xen-netback/interface.c
@@ -483,6 +483,8 @@ struct xenvif *xenvif_alloc(struct device *parent, domid_t domid,
 	vif->queues = NULL;
 	vif->num_queues = 0;
 
+	vif->xdp_headroom = 0;
+
 	spin_lock_init(&vif->lock);
 	INIT_LIST_HEAD(&vif->fe_mcast_addr);
 
diff --git a/drivers/net/xen-netback/netback.c b/drivers/net/xen-netback/netback.c
index 315dfc6..6dfca72 100644
--- a/drivers/net/xen-netback/netback.c
+++ b/drivers/net/xen-netback/netback.c
@@ -96,6 +96,13 @@
 module_param_named(hash_cache_size, xenvif_hash_cache_size, uint, 0644);
 MODULE_PARM_DESC(hash_cache_size, "Number of flows in the hash cache");
 
+/* The module parameter tells that we have to put data
+ * for xen-netfront with the XDP_PACKET_HEADROOM offset
+ * needed for XDP processing
+ */
+bool provides_xdp_headroom = true;
+module_param(provides_xdp_headroom, bool, 0644);
+
 static void xenvif_idx_release(struct xenvif_queue *queue, u16 pending_idx,
 			       u8 status);
 
diff --git a/drivers/net/xen-netback/rx.c b/drivers/net/xen-netback/rx.c
index ef58870..ac034f6 100644
--- a/drivers/net/xen-netback/rx.c
+++ b/drivers/net/xen-netback/rx.c
@@ -258,6 +258,19 @@ static void xenvif_rx_next_skb(struct xenvif_queue *queue,
 		pkt->extra_count++;
 	}
 
+	if (queue->vif->xdp_headroom) {
+		struct xen_netif_extra_info *extra;
+
+		extra = &pkt->extras[XEN_NETIF_EXTRA_TYPE_XDP - 1];
+
+		memset(extra, 0, sizeof(struct xen_netif_extra_info));
+		extra->u.xdp.headroom = queue->vif->xdp_headroom;
+		extra->type = XEN_NETIF_EXTRA_TYPE_XDP;
+		extra->flags = 0;
+
+		pkt->extra_count++;
+	}
+
 	if (skb->sw_hash) {
 		struct xen_netif_extra_info *extra;
 
@@ -356,7 +369,7 @@ static void xenvif_rx_data_slot(struct xenvif_queue *queue,
 				struct xen_netif_rx_request *req,
 				struct xen_netif_rx_response *rsp)
 {
-	unsigned int offset = 0;
+	unsigned int offset = queue->vif->xdp_headroom;
 	unsigned int flags;
 
 	do {
diff --git a/drivers/net/xen-netback/xenbus.c b/drivers/net/xen-netback/xenbus.c
index 286054b..7e62a6e 100644
--- a/drivers/net/xen-netback/xenbus.c
+++ b/drivers/net/xen-netback/xenbus.c
@@ -393,6 +393,24 @@ static void set_backend_state(struct backend_info *be,
 	}
 }
 
+static void read_xenbus_frontend_xdp(struct backend_info *be,
+				      struct xenbus_device *dev)
+{
+	struct xenvif *vif = be->vif;
+	u16 headroom;
+	int err;
+
+	err = xenbus_scanf(XBT_NIL, dev->otherend,
+			   "xdp-headroom", "%hu", &headroom);
+	if (err != 1) {
+		vif->xdp_headroom = 0;
+		return;
+	}
+	if (headroom > XEN_NETIF_MAX_XDP_HEADROOM)
+		headroom = XEN_NETIF_MAX_XDP_HEADROOM;
+	vif->xdp_headroom = headroom;
+}
+
 /**
  * Callback received when the frontend's state changes.
  */
@@ -417,6 +435,11 @@ static void frontend_changed(struct xenbus_device *dev,
 		set_backend_state(be, XenbusStateConnected);
 		break;
 
+	case XenbusStateReconfiguring:
+		read_xenbus_frontend_xdp(be, dev);
+		xenbus_switch_state(dev, XenbusStateReconfigured);
+		break;
+
 	case XenbusStateClosing:
 		set_backend_state(be, XenbusStateClosing);
 		break;
@@ -947,6 +970,8 @@ static int read_xenbus_vif_flags(struct backend_info *be)
 	vif->ipv6_csum = !!xenbus_read_unsigned(dev->otherend,
 						"feature-ipv6-csum-offload", 0);
 
+	read_xenbus_frontend_xdp(be, dev);
+
 	return 0;
 }
 
@@ -1036,6 +1061,15 @@ static int netback_probe(struct xenbus_device *dev,
 			goto abort_transaction;
 		}
 
+		/* we can adjust a headroom for netfront XDP processing */
+		err = xenbus_printf(xbt, dev->nodename,
+				    "feature-xdp-headroom", "%d",
+				    provides_xdp_headroom);
+		if (err) {
+			message = "writing feature-xdp-headroom";
+			goto abort_transaction;
+		}
+
 		/* We don't support rx-flip path (except old guests who
 		 * don't grok this feature flag).
 		 */
-- 
1.8.3.1

Re: [PATCH net-next v14 0/3] xen networking: add XDP support to xen-netfront

[David Miller]

From: Denis Kirjanov <kda@linux-powerpc.org>
Date: Mon, 29 Jun 2020 16:13:26 +0300

> The first patch adds a new extra type to enable proper synchronization
> between an RX request/response pair.
> The second patch implements BFP interface for xen-netfront.
> The third patch enables extra space for XDP processing.
 ...

Series applied, thank you.