1 files changed, 299 insertions, 164 deletions

diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c
index 7d389c2cc902..94dd6edd1800 100644
--- a/drivers/net/xen-netfront.c
+++ b/drivers/net/xen-netfront.c
@@ -121,21 +121,17 @@ struct netfront_queue {
 
 	/*
 	 * {tx,rx}_skbs store outstanding skbuffs. Free tx_skb entries
-	 * are linked from tx_skb_freelist through skb_entry.link.
-	 *
-	 *  NB. Freelist index entries are always going to be less than
-	 *  PAGE_OFFSET, whereas pointers to skbs will always be equal or
-	 *  greater than PAGE_OFFSET: we use this property to distinguish
-	 *  them.
+	 * are linked from tx_skb_freelist through tx_link.
 	 */
-	union skb_entry {
-		struct sk_buff *skb;
-		unsigned long link;
-	} tx_skbs[NET_TX_RING_SIZE];
+	struct sk_buff *tx_skbs[NET_TX_RING_SIZE];
+	unsigned short tx_link[NET_TX_RING_SIZE];
+#define TX_LINK_NONE 0xffff
+#define TX_PENDING   0xfffe
 	grant_ref_t gref_tx_head;
 	grant_ref_t grant_tx_ref[NET_TX_RING_SIZE];
 	struct page *grant_tx_page[NET_TX_RING_SIZE];
 	unsigned tx_skb_freelist;
+	unsigned int tx_pend_queue;
 
 	spinlock_t   rx_lock ____cacheline_aligned_in_smp;
 	struct xen_netif_rx_front_ring rx;
@@ -146,6 +142,9 @@ 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];
+
+	unsigned int rx_rsp_unconsumed;
+	spinlock_t rx_cons_lock;
 };
 
 struct netfront_info {
@@ -161,6 +160,9 @@ struct netfront_info {
 	struct netfront_stats __percpu *rx_stats;
 	struct netfront_stats __percpu *tx_stats;
 
+	/* Is device behaving sane? */
+	bool broken;
+
 	atomic_t rx_gso_checksum_fixup;
 };
 
@@ -169,33 +171,25 @@ struct netfront_rx_info {
 	struct xen_netif_extra_info extras[XEN_NETIF_EXTRA_TYPE_MAX - 1];
 };
 
-static void skb_entry_set_link(union skb_entry *list, unsigned short id)
-{
-	list->link = id;
-}
-
-static int skb_entry_is_link(const union skb_entry *list)
-{
-	BUILD_BUG_ON(sizeof(list->skb) != sizeof(list->link));
-	return (unsigned long)list->skb < PAGE_OFFSET;
-}
-
 /*
  * Access macros for acquiring freeing slots in tx_skbs[].
  */
 
-static void add_id_to_freelist(unsigned *head, union skb_entry *list,
-			       unsigned short id)
+static void add_id_to_list(unsigned *head, unsigned short *list,
+			   unsigned short id)
 {
-	skb_entry_set_link(&list[id], *head);
+	list[id] = *head;
 	*head = id;
 }
 
-static unsigned short get_id_from_freelist(unsigned *head,
-					   union skb_entry *list)
+static unsigned short get_id_from_list(unsigned *head, unsigned short *list)
 {
 	unsigned int id = *head;
-	*head = list[id].link;
+
+	if (id != TX_LINK_NONE) {
+		*head = list[id];
+		list[id] = TX_LINK_NONE;
+	}
 	return id;
 }
 
@@ -351,7 +345,7 @@ static int xennet_open(struct net_device *dev)
 	unsigned int i = 0;
 	struct netfront_queue *queue = NULL;
 
-	if (!np->queues)
+	if (!np->queues || np->broken)
 		return -ENODEV;
 
 	for (i = 0; i < num_queues; ++i) {
@@ -373,41 +367,62 @@ static int xennet_open(struct net_device *dev)
 	return 0;
 }
 
-static void xennet_tx_buf_gc(struct netfront_queue *queue)
+static bool xennet_tx_buf_gc(struct netfront_queue *queue)
 {
 	RING_IDX cons, prod;
 	unsigned short id;
 	struct sk_buff *skb;
 	bool more_to_do;
+	bool work_done = false;
+	const struct device *dev = &queue->info->netdev->dev;
 
 	BUG_ON(!netif_carrier_ok(queue->info->netdev));
 
 	do {
 		prod = queue->tx.sring->rsp_prod;
+		if (RING_RESPONSE_PROD_OVERFLOW(&queue->tx, prod)) {
+			dev_alert(dev, "Illegal number of responses %u\n",
+				  prod - queue->tx.rsp_cons);
+			goto err;
+		}
 		rmb(); /* Ensure we see responses up to 'rp'. */
 
 		for (cons = queue->tx.rsp_cons; cons != prod; cons++) {
-			struct xen_netif_tx_response *txrsp;
+			struct xen_netif_tx_response txrsp;
+
+			work_done = true;
 
-			txrsp = RING_GET_RESPONSE(&queue->tx, cons);
-			if (txrsp->status == XEN_NETIF_RSP_NULL)
+			RING_COPY_RESPONSE(&queue->tx, cons, &txrsp);
+			if (txrsp.status == XEN_NETIF_RSP_NULL)
 				continue;
 
-			id  = txrsp->id;
-			skb = queue->tx_skbs[id].skb;
-			if (unlikely(gnttab_query_foreign_access(
-				queue->grant_tx_ref[id]) != 0)) {
-				pr_alert("%s: warning -- grant still in use by backend domain\n",
-					 __func__);
-				BUG();
+			id = txrsp.id;
+			if (id >= RING_SIZE(&queue->tx)) {
+				dev_alert(dev,
+					  "Response has incorrect id (%u)\n",
+					  id);
+				goto err;
+			}
+			if (queue->tx_link[id] != TX_PENDING) {
+				dev_alert(dev,
+					  "Response for inactive request\n");
+				goto err;
+			}
+
+			queue->tx_link[id] = TX_LINK_NONE;
+			skb = queue->tx_skbs[id];
+			queue->tx_skbs[id] = NULL;
+			if (unlikely(!gnttab_end_foreign_access_ref(
+				queue->grant_tx_ref[id], GNTMAP_readonly))) {
+				dev_alert(dev,
+					  "Grant still in use by backend domain\n");
+				goto err;
 			}
-			gnttab_end_foreign_access_ref(
-				queue->grant_tx_ref[id], GNTMAP_readonly);
 			gnttab_release_grant_reference(
 				&queue->gref_tx_head, queue->grant_tx_ref[id]);
 			queue->grant_tx_ref[id] = GRANT_INVALID_REF;
 			queue->grant_tx_page[id] = NULL;
-			add_id_to_freelist(&queue->tx_skb_freelist, queue->tx_skbs, id);
+			add_id_to_list(&queue->tx_skb_freelist, queue->tx_link, id);
 			dev_kfree_skb_irq(skb);
 		}
 
@@ -417,13 +432,22 @@ static void xennet_tx_buf_gc(struct netfront_queue *queue)
 	} while (more_to_do);
 
 	xennet_maybe_wake_tx(queue);
+
+	return work_done;
+
+ err:
+	queue->info->broken = true;
+	dev_alert(dev, "Disabled for further use\n");
+
+	return work_done;
 }
 
 struct xennet_gnttab_make_txreq {
 	struct netfront_queue *queue;
 	struct sk_buff *skb;
 	struct page *page;
-	struct xen_netif_tx_request *tx; /* Last request */
+	struct xen_netif_tx_request *tx;      /* Last request on ring page */
+	struct xen_netif_tx_request tx_local; /* Last request local copy*/
 	unsigned int size;
 };
 
@@ -439,7 +463,7 @@ static void xennet_tx_setup_grant(unsigned long gfn, unsigned int offset,
 	struct netfront_queue *queue = info->queue;
 	struct sk_buff *skb = info->skb;
 
-	id = get_id_from_freelist(&queue->tx_skb_freelist, queue->tx_skbs);
+	id = get_id_from_list(&queue->tx_skb_freelist, queue->tx_link);
 	tx = RING_GET_REQUEST(&queue->tx, queue->tx.req_prod_pvt++);
 	ref = gnttab_claim_grant_reference(&queue->gref_tx_head);
 	WARN_ON_ONCE(IS_ERR_VALUE((unsigned long)(int)ref));
@@ -447,34 +471,37 @@ static void xennet_tx_setup_grant(unsigned long gfn, unsigned int offset,
 	gnttab_grant_foreign_access_ref(ref, queue->info->xbdev->otherend_id,
 					gfn, GNTMAP_readonly);
 
-	queue->tx_skbs[id].skb = skb;
+	queue->tx_skbs[id] = skb;
 	queue->grant_tx_page[id] = page;
 	queue->grant_tx_ref[id] = ref;
 
-	tx->id = id;
-	tx->gref = ref;
-	tx->offset = offset;
-	tx->size = len;
-	tx->flags = 0;
+	info->tx_local.id = id;
+	info->tx_local.gref = ref;
+	info->tx_local.offset = offset;
+	info->tx_local.size = len;
+	info->tx_local.flags = 0;
+
+	*tx = info->tx_local;
+
+	/*
+	 * Put the request in the pending queue, it will be set to be pending
+	 * when the producer index is about to be raised.
+	 */
+	add_id_to_list(&queue->tx_pend_queue, queue->tx_link, id);
 
 	info->tx = tx;
-	info->size += tx->size;
+	info->size += info->tx_local.size;
 }
 
 static struct xen_netif_tx_request *xennet_make_first_txreq(
-	struct netfront_queue *queue, struct sk_buff *skb,
-	struct page *page, unsigned int offset, unsigned int len)
+	struct xennet_gnttab_make_txreq *info,
+	unsigned int offset, unsigned int len)
 {
-	struct xennet_gnttab_make_txreq info = {
-		.queue = queue,
-		.skb = skb,
-		.page = page,
-		.size = 0,
-	};
+	info->size = 0;
 
-	gnttab_for_one_grant(page, offset, len, xennet_tx_setup_grant, &info);
+	gnttab_for_one_grant(info->page, offset, len, xennet_tx_setup_grant, info);
 
-	return info.tx;
+	return info->tx;
 }
 
 static void xennet_make_one_txreq(unsigned long gfn, unsigned int offset,
@@ -487,35 +514,27 @@ static void xennet_make_one_txreq(unsigned long gfn, unsigned int offset,
 	xennet_tx_setup_grant(gfn, offset, len, data);
 }
 
-static struct xen_netif_tx_request *xennet_make_txreqs(
-	struct netfront_queue *queue, struct xen_netif_tx_request *tx,
-	struct sk_buff *skb, struct page *page,
+static void xennet_make_txreqs(
+	struct xennet_gnttab_make_txreq *info,
+	struct page *page,
 	unsigned int offset, unsigned int len)
 {
-	struct xennet_gnttab_make_txreq info = {
-		.queue = queue,
-		.skb = skb,
-		.tx = tx,
-	};
-
 	/* Skip unused frames from start of page */
 	page += offset >> PAGE_SHIFT;
 	offset &= ~PAGE_MASK;
 
 	while (len) {
-		info.page = page;
-		info.size = 0;
+		info->page = page;
+		info->size = 0;
 
 		gnttab_foreach_grant_in_range(page, offset, len,
 					      xennet_make_one_txreq,
-					      &info);
+					      info);
 
 		page++;
 		offset = 0;
-		len -= info.size;
+		len -= info->size;
 	}
-
-	return info.tx;
 }
 
 /*
@@ -562,13 +581,22 @@ static u16 xennet_select_queue(struct net_device *dev, struct sk_buff *skb,
 	return queue_idx;
 }
 
+static void xennet_mark_tx_pending(struct netfront_queue *queue)
+{
+	unsigned int i;
+
+	while ((i = get_id_from_list(&queue->tx_pend_queue, queue->tx_link)) !=
+		TX_LINK_NONE)
+		queue->tx_link[i] = TX_PENDING;
+}
+
 #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)
 {
 	struct netfront_info *np = netdev_priv(dev);
 	struct netfront_stats *tx_stats = this_cpu_ptr(np->tx_stats);
-	struct xen_netif_tx_request *tx, *first_tx;
+	struct xen_netif_tx_request *first_tx;
 	unsigned int i;
 	int notify;
 	int slots;
@@ -577,6 +605,7 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
 	unsigned int len;
 	unsigned long flags;
 	struct netfront_queue *queue = NULL;
+	struct xennet_gnttab_make_txreq info = { };
 	unsigned int num_queues = dev->real_num_tx_queues;
 	u16 queue_index;
 	struct sk_buff *nskb;
@@ -584,6 +613,8 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
 	/* Drop the packet if no queues are set up */
 	if (num_queues < 1)
 		goto drop;
+	if (unlikely(np->broken))
+		goto drop;
 	/* Determine which queue to transmit this SKB on */
 	queue_index = skb_get_queue_mapping(skb);
 	queue = &np->queues[queue_index];
@@ -634,21 +665,24 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
 	}
 
 	/* First request for the linear area. */
-	first_tx = tx = xennet_make_first_txreq(queue, skb,
-						page, offset, len);
-	offset += tx->size;
+	info.queue = queue;
+	info.skb = skb;
+	info.page = page;
+	first_tx = xennet_make_first_txreq(&info, offset, len);
+	offset += info.tx_local.size;
 	if (offset == PAGE_SIZE) {
 		page++;
 		offset = 0;
 	}
-	len -= tx->size;
+	len -= info.tx_local.size;
 
 	if (skb->ip_summed == CHECKSUM_PARTIAL)
 		/* local packet? */
-		tx->flags |= XEN_NETTXF_csum_blank | XEN_NETTXF_data_validated;
+		first_tx->flags |= XEN_NETTXF_csum_blank |
+				   XEN_NETTXF_data_validated;
 	else if (skb->ip_summed == CHECKSUM_UNNECESSARY)
 		/* remote but checksummed. */
-		tx->flags |= XEN_NETTXF_data_validated;
+		first_tx->flags |= XEN_NETTXF_data_validated;
 
 	/* Optional extra info after the first request. */
 	if (skb_shinfo(skb)->gso_size) {
@@ -657,7 +691,7 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
 		gso = (struct xen_netif_extra_info *)
 			RING_GET_REQUEST(&queue->tx, queue->tx.req_prod_pvt++);
 
-		tx->flags |= XEN_NETTXF_extra_info;
+		first_tx->flags |= XEN_NETTXF_extra_info;
 
 		gso->u.gso.size = skb_shinfo(skb)->gso_size;
 		gso->u.gso.type = (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6) ?
@@ -671,12 +705,12 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
 	}
 
 	/* Requests for the rest of the linear area. */
-	tx = xennet_make_txreqs(queue, tx, skb, page, offset, len);
+	xennet_make_txreqs(&info, page, offset, len);
 
 	/* Requests for all the frags. */
 	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
 		skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
-		tx = xennet_make_txreqs(queue, tx, skb, skb_frag_page(frag),
+		xennet_make_txreqs(&info, skb_frag_page(frag),
 					skb_frag_off(frag),
 					skb_frag_size(frag));
 	}
@@ -684,6 +718,8 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
 	/* First request has the packet length. */
 	first_tx->size = skb->len;
 
+	xennet_mark_tx_pending(queue);
+
 	RING_PUSH_REQUESTS_AND_CHECK_NOTIFY(&queue->tx, notify);
 	if (notify)
 		notify_remote_via_irq(queue->tx_irq);
@@ -723,6 +759,38 @@ static int xennet_close(struct net_device *dev)
 	return 0;
 }
 
+static void xennet_destroy_queues(struct netfront_info *info)
+{
+	unsigned int i;
+
+	for (i = 0; i < info->netdev->real_num_tx_queues; i++) {
+		struct netfront_queue *queue = &info->queues[i];
+
+		if (netif_running(info->netdev))
+			napi_disable(&queue->napi);
+		netif_napi_del(&queue->napi);
+	}
+
+	kfree(info->queues);
+	info->queues = NULL;
+}
+
+static void xennet_uninit(struct net_device *dev)
+{
+	struct netfront_info *np = netdev_priv(dev);
+	xennet_destroy_queues(np);
+}
+
+static void xennet_set_rx_rsp_cons(struct netfront_queue *queue, RING_IDX val)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&queue->rx_cons_lock, flags);
+	queue->rx.rsp_cons = val;
+	queue->rx_rsp_unconsumed = RING_HAS_UNCONSUMED_RESPONSES(&queue->rx);
+	spin_unlock_irqrestore(&queue->rx_cons_lock, flags);
+}
+
 static void xennet_move_rx_slot(struct netfront_queue *queue, struct sk_buff *skb,
 				grant_ref_t ref)
 {
@@ -741,7 +809,7 @@ static int xennet_get_extras(struct netfront_queue *queue,
 			     RING_IDX rp)
 
 {
-	struct xen_netif_extra_info *extra;
+	struct xen_netif_extra_info extra;
 	struct device *dev = &queue->info->netdev->dev;
 	RING_IDX cons = queue->rx.rsp_cons;
 	int err = 0;
@@ -757,26 +825,24 @@ static int xennet_get_extras(struct netfront_queue *queue,
 			break;
 		}
 
-		extra = (struct xen_netif_extra_info *)
-			RING_GET_RESPONSE(&queue->rx, ++cons);
+		RING_COPY_RESPONSE(&queue->rx, ++cons, &extra);
 
-		if (unlikely(!extra->type ||
-			     extra->type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
+		if (unlikely(!extra.type ||
+			     extra.type >= XEN_NETIF_EXTRA_TYPE_MAX)) {
 			if (net_ratelimit())
 				dev_warn(dev, "Invalid extra type: %d\n",
-					extra->type);
+					 extra.type);
 			err = -EINVAL;
 		} else {
-			memcpy(&extras[extra->type - 1], extra,
-			       sizeof(*extra));
+			extras[extra.type - 1] = extra;
 		}
 
 		skb = xennet_get_rx_skb(queue, cons);
 		ref = xennet_get_rx_ref(queue, cons);
 		xennet_move_rx_slot(queue, skb, ref);
-	} while (extra->flags & XEN_NETIF_EXTRA_FLAG_MORE);
+	} while (extra.flags & XEN_NETIF_EXTRA_FLAG_MORE);
 
-	queue->rx.rsp_cons = cons;
+	xennet_set_rx_rsp_cons(queue, cons);
 	return err;
 }
 
@@ -784,7 +850,7 @@ static int xennet_get_responses(struct netfront_queue *queue,
 				struct netfront_rx_info *rinfo, RING_IDX rp,
 				struct sk_buff_head *list)
 {
-	struct xen_netif_rx_response *rx = &rinfo->rx;
+	struct xen_netif_rx_response *rx = &rinfo->rx, rx_local;
 	struct xen_netif_extra_info *extras = rinfo->extras;
 	struct device *dev = &queue->info->netdev->dev;
 	RING_IDX cons = queue->rx.rsp_cons;
@@ -793,7 +859,6 @@ static int xennet_get_responses(struct netfront_queue *queue,
 	int max = XEN_NETIF_NR_SLOTS_MIN + (rx->status <= RX_COPY_THRESHOLD);
 	int slots = 1;
 	int err = 0;
-	unsigned long ret;
 
 	if (rx->flags & XEN_NETRXF_extra_info) {
 		err = xennet_get_extras(queue, extras, rp);
@@ -824,8 +889,13 @@ static int xennet_get_responses(struct netfront_queue *queue,
 			goto next;
 		}
 
-		ret = gnttab_end_foreign_access_ref(ref, 0);
-		BUG_ON(!ret);
+		if (!gnttab_end_foreign_access_ref(ref, 0)) {
+			dev_alert(dev,
+				  "Grant still in use by backend domain\n");
+			queue->info->broken = true;
+			dev_alert(dev, "Disabled for further use\n");
+			return -EINVAL;
+		}
 
 		gnttab_release_grant_reference(&queue->gref_rx_head, ref);
 
@@ -842,7 +912,8 @@ next:
 			break;
 		}
 
-		rx = RING_GET_RESPONSE(&queue->rx, cons + slots);
+		RING_COPY_RESPONSE(&queue->rx, cons + slots, &rx_local);
+		rx = &rx_local;
 		skb = xennet_get_rx_skb(queue, cons + slots);
 		ref = xennet_get_rx_ref(queue, cons + slots);
 		slots++;
@@ -855,7 +926,7 @@ next:
 	}
 
 	if (unlikely(err))
-		queue->rx.rsp_cons = cons + slots;
+		xennet_set_rx_rsp_cons(queue, cons + slots);
 
 	return err;
 }
@@ -897,10 +968,11 @@ static int xennet_fill_frags(struct netfront_queue *queue,
 	struct sk_buff *nskb;
 
 	while ((nskb = __skb_dequeue(list))) {
-		struct xen_netif_rx_response *rx =
-			RING_GET_RESPONSE(&queue->rx, ++cons);
+		struct xen_netif_rx_response rx;
 		skb_frag_t *nfrag = &skb_shinfo(nskb)->frags[0];
 
+		RING_COPY_RESPONSE(&queue->rx, ++cons, &rx);
+
 		if (skb_shinfo(skb)->nr_frags == MAX_SKB_FRAGS) {
 			unsigned int pull_to = NETFRONT_SKB_CB(skb)->pull_to;
 
@@ -908,20 +980,21 @@ static int xennet_fill_frags(struct netfront_queue *queue,
 			__pskb_pull_tail(skb, pull_to - skb_headlen(skb));
 		}
 		if (unlikely(skb_shinfo(skb)->nr_frags >= MAX_SKB_FRAGS)) {
-			queue->rx.rsp_cons = ++cons + skb_queue_len(list);
+			xennet_set_rx_rsp_cons(queue,
+					       ++cons + skb_queue_len(list));
 			kfree_skb(nskb);
 			return -ENOENT;
 		}
 
 		skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags,
 				skb_frag_page(nfrag),
-				rx->offset, rx->status, PAGE_SIZE);
+				rx.offset, rx.status, PAGE_SIZE);
 
 		skb_shinfo(nskb)->nr_frags = 0;
 		kfree_skb(nskb);
 	}
 
-	queue->rx.rsp_cons = cons;
+	xennet_set_rx_rsp_cons(queue, cons);
 
 	return 0;
 }
@@ -1008,17 +1081,28 @@ static int xennet_poll(struct napi_struct *napi, int budget)
 	skb_queue_head_init(&tmpq);
 
 	rp = queue->rx.sring->rsp_prod;
+	if (RING_RESPONSE_PROD_OVERFLOW(&queue->rx, rp)) {
+		dev_alert(&dev->dev, "Illegal number of responses %u\n",
+			  rp - queue->rx.rsp_cons);
+		queue->info->broken = true;
+		spin_unlock(&queue->rx_lock);
+		return 0;
+	}
 	rmb(); /* Ensure we see queued responses up to 'rp'. */
 
 	i = queue->rx.rsp_cons;
 	work_done = 0;
 	while ((i != rp) && (work_done < budget)) {
-		memcpy(rx, RING_GET_RESPONSE(&queue->rx, i), sizeof(*rx));
+		RING_COPY_RESPONSE(&queue->rx, i, rx);
 		memset(extras, 0, sizeof(rinfo.extras));
 
 		err = xennet_get_responses(queue, &rinfo, rp, &tmpq);
 
 		if (unlikely(err)) {
+			if (queue->info->broken) {
+				spin_unlock(&queue->rx_lock);
+				return 0;
+			}
 err:
 			while ((skb = __skb_dequeue(&tmpq)))
 				__skb_queue_tail(&errq, skb);
@@ -1035,7 +1119,9 @@ err:
 
 			if (unlikely(xennet_set_skb_gso(skb, gso))) {
 				__skb_queue_head(&tmpq, skb);
-				queue->rx.rsp_cons += skb_queue_len(&tmpq);
+				xennet_set_rx_rsp_cons(queue,
+						       queue->rx.rsp_cons +
+						       skb_queue_len(&tmpq));
 				goto err;
 			}
 		}
@@ -1059,7 +1145,8 @@ err:
 
 		__skb_queue_tail(&rxq, skb);
 
-		i = ++queue->rx.rsp_cons;
+		i = queue->rx.rsp_cons + 1;
+		xennet_set_rx_rsp_cons(queue, i);
 		work_done++;
 	}
 
@@ -1135,17 +1222,18 @@ static void xennet_release_tx_bufs(struct netfront_queue *queue)
 
 	for (i = 0; i < NET_TX_RING_SIZE; i++) {
 		/* Skip over entries which are actually freelist references */
-		if (skb_entry_is_link(&queue->tx_skbs[i]))
+		if (!queue->tx_skbs[i])
 			continue;
 
-		skb = queue->tx_skbs[i].skb;
+		skb = queue->tx_skbs[i];
+		queue->tx_skbs[i] = NULL;
 		get_page(queue->grant_tx_page[i]);
 		gnttab_end_foreign_access(queue->grant_tx_ref[i],
 					  GNTMAP_readonly,
 					  (unsigned long)page_address(queue->grant_tx_page[i]));
 		queue->grant_tx_page[i] = NULL;
 		queue->grant_tx_ref[i] = GRANT_INVALID_REF;
-		add_id_to_freelist(&queue->tx_skb_freelist, queue->tx_skbs, i);
+		add_id_to_list(&queue->tx_skb_freelist, queue->tx_link, i);
 		dev_kfree_skb_irq(skb);
 	}
 }
@@ -1220,34 +1308,79 @@ static int xennet_set_features(struct net_device *dev,
 	return 0;
 }
 
-static irqreturn_t xennet_tx_interrupt(int irq, void *dev_id)
+static bool xennet_handle_tx(struct netfront_queue *queue, unsigned int *eoi)
 {
-	struct netfront_queue *queue = dev_id;
 	unsigned long flags;
 
+	if (unlikely(queue->info->broken))
+		return false;
+
 	spin_lock_irqsave(&queue->tx_lock, flags);
-	xennet_tx_buf_gc(queue);
+	if (xennet_tx_buf_gc(queue))
+		*eoi = 0;
 	spin_unlock_irqrestore(&queue->tx_lock, flags);
 
+	return true;
+}
+
+static irqreturn_t xennet_tx_interrupt(int irq, void *dev_id)
+{
+	unsigned int eoiflag = XEN_EOI_FLAG_SPURIOUS;
+
+	if (likely(xennet_handle_tx(dev_id, &eoiflag)))
+		xen_irq_lateeoi(irq, eoiflag);
+
 	return IRQ_HANDLED;
 }
 
-static irqreturn_t xennet_rx_interrupt(int irq, void *dev_id)
+static bool xennet_handle_rx(struct netfront_queue *queue, unsigned int *eoi)
 {
-	struct netfront_queue *queue = dev_id;
-	struct net_device *dev = queue->info->netdev;
+	unsigned int work_queued;
+	unsigned long flags;
+
+	if (unlikely(queue->info->broken))
+		return false;
 
-	if (likely(netif_carrier_ok(dev) &&
-		   RING_HAS_UNCONSUMED_RESPONSES(&queue->rx)))
+	spin_lock_irqsave(&queue->rx_cons_lock, flags);
+	work_queued = RING_HAS_UNCONSUMED_RESPONSES(&queue->rx);
+	if (work_queued > queue->rx_rsp_unconsumed) {
+		queue->rx_rsp_unconsumed = work_queued;
+		*eoi = 0;
+	} else if (unlikely(work_queued < queue->rx_rsp_unconsumed)) {
+		const struct device *dev = &queue->info->netdev->dev;
+
+		spin_unlock_irqrestore(&queue->rx_cons_lock, flags);
+		dev_alert(dev, "RX producer index going backwards\n");
+		dev_alert(dev, "Disabled for further use\n");
+		queue->info->broken = true;
+		return false;
+	}
+	spin_unlock_irqrestore(&queue->rx_cons_lock, flags);
+
+	if (likely(netif_carrier_ok(queue->info->netdev) && work_queued))
 		napi_schedule(&queue->napi);
 
+	return true;
+}
+
+static irqreturn_t xennet_rx_interrupt(int irq, void *dev_id)
+{
+	unsigned int eoiflag = XEN_EOI_FLAG_SPURIOUS;
+
+	if (likely(xennet_handle_rx(dev_id, &eoiflag)))
+		xen_irq_lateeoi(irq, eoiflag);
+
 	return IRQ_HANDLED;
 }
 
 static irqreturn_t xennet_interrupt(int irq, void *dev_id)
 {
-	xennet_tx_interrupt(irq, dev_id);
-	xennet_rx_interrupt(irq, dev_id);
+	unsigned int eoiflag = XEN_EOI_FLAG_SPURIOUS;
+
+	if (xennet_handle_tx(dev_id, &eoiflag) &&
+	    xennet_handle_rx(dev_id, &eoiflag))
+		xen_irq_lateeoi(irq, eoiflag);
+
 	return IRQ_HANDLED;
 }
 
@@ -1258,12 +1391,17 @@ static void xennet_poll_controller(struct net_device *dev)
 	struct netfront_info *info = netdev_priv(dev);
 	unsigned int num_queues = dev->real_num_tx_queues;
 	unsigned int i;
+
+	if (info->broken)
+		return;
+
 	for (i = 0; i < num_queues; ++i)
 		xennet_interrupt(0, &info->queues[i]);
 }
 #endif
 
 static const struct net_device_ops xennet_netdev_ops = {
+	.ndo_uninit          = xennet_uninit,
 	.ndo_open            = xennet_open,
 	.ndo_stop            = xennet_close,
 	.ndo_start_xmit      = xennet_start_xmit,
@@ -1477,9 +1615,10 @@ static int setup_netfront_single(struct netfront_queue *queue)
 	if (err < 0)
 		goto fail;
 
-	err = bind_evtchn_to_irqhandler(queue->tx_evtchn,
-					xennet_interrupt,
-					0, queue->info->netdev->name, queue);
+	err = bind_evtchn_to_irqhandler_lateeoi(queue->tx_evtchn,
+						xennet_interrupt, 0,
+						queue->info->netdev->name,
+						queue);
 	if (err < 0)
 		goto bind_fail;
 	queue->rx_evtchn = queue->tx_evtchn;
@@ -1507,18 +1646,18 @@ static int setup_netfront_split(struct netfront_queue *queue)
 
 	snprintf(queue->tx_irq_name, sizeof(queue->tx_irq_name),
 		 "%s-tx", queue->name);
-	err = bind_evtchn_to_irqhandler(queue->tx_evtchn,
-					xennet_tx_interrupt,
-					0, queue->tx_irq_name, queue);
+	err = bind_evtchn_to_irqhandler_lateeoi(queue->tx_evtchn,
+						xennet_tx_interrupt, 0,
+						queue->tx_irq_name, queue);
 	if (err < 0)
 		goto bind_tx_fail;
 	queue->tx_irq = err;
 
 	snprintf(queue->rx_irq_name, sizeof(queue->rx_irq_name),
 		 "%s-rx", queue->name);
-	err = bind_evtchn_to_irqhandler(queue->rx_evtchn,
-					xennet_rx_interrupt,
-					0, queue->rx_irq_name, queue);
+	err = bind_evtchn_to_irqhandler_lateeoi(queue->rx_evtchn,
+						xennet_rx_interrupt, 0,
+						queue->rx_irq_name, queue);
 	if (err < 0)
 		goto bind_rx_fail;
 	queue->rx_irq = err;
@@ -1542,7 +1681,7 @@ static int setup_netfront(struct xenbus_device *dev,
 			struct netfront_queue *queue, unsigned int feature_split_evtchn)
 {
 	struct xen_netif_tx_sring *txs;
-	struct xen_netif_rx_sring *rxs;
+	struct xen_netif_rx_sring *rxs = NULL;
 	grant_ref_t gref;
 	int err;
 
@@ -1562,21 +1701,21 @@ static int setup_netfront(struct xenbus_device *dev,
 
 	err = xenbus_grant_ring(dev, txs, 1, &gref);
 	if (err < 0)
-		goto grant_tx_ring_fail;
+		goto fail;
 	queue->tx_ring_ref = gref;
 
 	rxs = (struct xen_netif_rx_sring *)get_zeroed_page(GFP_NOIO | __GFP_HIGH);
 	if (!rxs) {
 		err = -ENOMEM;
 		xenbus_dev_fatal(dev, err, "allocating rx ring page");
-		goto alloc_rx_ring_fail;
+		goto fail;
 	}
 	SHARED_RING_INIT(rxs);
 	FRONT_RING_INIT(&queue->rx, rxs, XEN_PAGE_SIZE);
 
 	err = xenbus_grant_ring(dev, rxs, 1, &gref);
 	if (err < 0)
-		goto grant_rx_ring_fail;
+		goto fail;
 	queue->rx_ring_ref = gref;
 
 	if (feature_split_evtchn)
@@ -1589,22 +1728,28 @@ static int setup_netfront(struct xenbus_device *dev,
 		err = setup_netfront_single(queue);
 
 	if (err)
-		goto alloc_evtchn_fail;
+		goto fail;
 
 	return 0;
 
 	/* If we fail to setup netfront, it is safe to just revoke access to
 	 * granted pages because backend is not accessing it at this point.
 	 */
-alloc_evtchn_fail:
-	gnttab_end_foreign_access_ref(queue->rx_ring_ref, 0);
-grant_rx_ring_fail:
-	free_page((unsigned long)rxs);
-alloc_rx_ring_fail:
-	gnttab_end_foreign_access_ref(queue->tx_ring_ref, 0);
-grant_tx_ring_fail:
-	free_page((unsigned long)txs);
-fail:
+ fail:
+	if (queue->rx_ring_ref != GRANT_INVALID_REF) {
+		gnttab_end_foreign_access(queue->rx_ring_ref, 0,
+					  (unsigned long)rxs);
+		queue->rx_ring_ref = GRANT_INVALID_REF;
+	} else {
+		free_page((unsigned long)rxs);
+	}
+	if (queue->tx_ring_ref != GRANT_INVALID_REF) {
+		gnttab_end_foreign_access(queue->tx_ring_ref, 0,
+					  (unsigned long)txs);
+		queue->tx_ring_ref = GRANT_INVALID_REF;
+	} else {
+		free_page((unsigned long)txs);
+	}
 	return err;
 }
 
@@ -1620,6 +1765,7 @@ static int xennet_init_queue(struct netfront_queue *queue)
 
 	spin_lock_init(&queue->tx_lock);
 	spin_lock_init(&queue->rx_lock);
+	spin_lock_init(&queue->rx_cons_lock);
 
 	timer_setup(&queue->rx_refill_timer, rx_refill_timeout, 0);
 
@@ -1627,13 +1773,15 @@ static int xennet_init_queue(struct netfront_queue *queue)
 	snprintf(queue->name, sizeof(queue->name), "vif%s-q%u",
 		 devid, queue->id);
 
-	/* Initialise tx_skbs as a free chain containing every entry. */
+	/* Initialise tx_skb_freelist as a free chain containing every entry. */
 	queue->tx_skb_freelist = 0;
+	queue->tx_pend_queue = TX_LINK_NONE;
 	for (i = 0; i < NET_TX_RING_SIZE; i++) {
-		skb_entry_set_link(&queue->tx_skbs[i], i+1);
+		queue->tx_link[i] = i + 1;
 		queue->grant_tx_ref[i] = GRANT_INVALID_REF;
 		queue->grant_tx_page[i] = NULL;
 	}
+	queue->tx_link[NET_TX_RING_SIZE - 1] = TX_LINK_NONE;
 
 	/* Clear out rx_skbs */
 	for (i = 0; i < NET_RX_RING_SIZE; i++) {
@@ -1747,22 +1895,6 @@ error:
 	return err;
 }
 
-static void xennet_destroy_queues(struct netfront_info *info)
-{
-	unsigned int i;
-
-	for (i = 0; i < info->netdev->real_num_tx_queues; i++) {
-		struct netfront_queue *queue = &info->queues[i];
-
-		if (netif_running(info->netdev))
-			napi_disable(&queue->napi);
-		netif_napi_del(&queue->napi);
-	}
-
-	kfree(info->queues);
-	info->queues = NULL;
-}
-
 static int xennet_create_queues(struct netfront_info *info,
 				unsigned int *num_queues)
 {
@@ -1838,6 +1970,9 @@ static int talk_to_netback(struct xenbus_device *dev,
 	if (info->queues)
 		xennet_destroy_queues(info);
 
+	/* For the case of a reconnect reset the "broken" indicator. */
+	info->broken = false;
+
 	err = xennet_create_queues(info, &num_queues);
 	if (err < 0) {
 		xenbus_dev_fatal(dev, err, "creating queues");