1 files changed, 205 insertions, 71 deletions

diff --git a/drivers/net/xen-netfront.c b/drivers/net/xen-netfront.c
index 434a009c52d9..8c3f9f041594 100644
--- a/drivers/net/xen-netfront.c
+++ b/drivers/net/xen-netfront.c
@@ -63,6 +63,10 @@ module_param_named(max_queues, xennet_max_queues, uint, 0644);
 MODULE_PARM_DESC(max_queues,
 		 "Maximum number of queues per virtual interface");
 
+static bool __read_mostly xennet_trusted = true;
+module_param_named(trusted, xennet_trusted, bool, 0644);
+MODULE_PARM_DESC(trusted, "Is the backend trusted");
+
 #define XENNET_TIMEOUT  (5 * HZ)
 
 static const struct ethtool_ops xennet_ethtool_ops;
@@ -142,6 +146,9 @@ 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 {
@@ -160,6 +167,9 @@ struct netfront_info {
 	/* Is device behaving sane? */
 	bool broken;
 
+	/* Should skbs be bounced into a zeroed buffer? */
+	bool bounce;
+
 	atomic_t rx_gso_checksum_fixup;
 };
 
@@ -258,7 +268,7 @@ 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);
+	page = alloc_page(GFP_ATOMIC | __GFP_NOWARN | __GFP_ZERO);
 	if (!page) {
 		kfree_skb(skb);
 		return NULL;
@@ -366,12 +376,13 @@ 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));
@@ -388,6 +399,8 @@ static void xennet_tx_buf_gc(struct netfront_queue *queue)
 		for (cons = queue->tx.rsp_cons; cons != prod; cons++) {
 			struct xen_netif_tx_response txrsp;
 
+			work_done = true;
+
 			RING_COPY_RESPONSE(&queue->tx, cons, &txrsp);
 			if (txrsp.status == XEN_NETIF_RSP_NULL)
 				continue;
@@ -408,14 +421,12 @@ static void xennet_tx_buf_gc(struct netfront_queue *queue)
 			queue->tx_link[id] = TX_LINK_NONE;
 			skb = queue->tx_skbs[id];
 			queue->tx_skbs[id] = NULL;
-			if (unlikely(gnttab_query_foreign_access(
-				queue->grant_tx_ref[id]) != 0)) {
+			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;
@@ -431,11 +442,13 @@ static void xennet_tx_buf_gc(struct netfront_queue *queue)
 
 	xennet_maybe_wake_tx(queue);
 
-	return;
+	return work_done;
 
  err:
 	queue->info->broken = true;
 	dev_alert(dev, "Disabled for further use\n");
+
+	return work_done;
 }
 
 struct xennet_gnttab_make_txreq {
@@ -587,6 +600,34 @@ static void xennet_mark_tx_pending(struct netfront_queue *queue)
 		queue->tx_link[i] = TX_PENDING;
 }
 
+struct sk_buff *bounce_skb(const struct sk_buff *skb)
+{
+	unsigned int headerlen = skb_headroom(skb);
+	/* Align size to allocate full pages and avoid contiguous data leaks */
+	unsigned int size = ALIGN(skb_end_offset(skb) + skb->data_len,
+				  XEN_PAGE_SIZE);
+	struct sk_buff *n = alloc_skb(size, GFP_ATOMIC | __GFP_ZERO);
+
+	if (!n)
+		return NULL;
+
+	if (!IS_ALIGNED((uintptr_t)n->head, XEN_PAGE_SIZE)) {
+		WARN_ONCE(1, "misaligned skb allocated\n");
+		kfree_skb(n);
+		return NULL;
+	}
+
+	/* Set the data pointer */
+	skb_reserve(n, headerlen);
+	/* Set the tail pointer and length */
+	skb_put(n, skb->len);
+
+	BUG_ON(skb_copy_bits(skb, -headerlen, n->head, headerlen + skb->len));
+
+	skb_copy_header(n, skb);
+	return n;
+}
+
 #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)
@@ -639,9 +680,13 @@ static netdev_tx_t xennet_start_xmit(struct sk_buff *skb, struct net_device *dev
 
 	/* The first req should be at least ETH_HLEN size or the packet will be
 	 * dropped by netback.
+	 *
+	 * If the backend is not trusted bounce all data to zeroed pages to
+	 * avoid exposing contiguous data on the granted page not belonging to
+	 * the skb.
 	 */
-	if (unlikely(PAGE_SIZE - offset < ETH_HLEN)) {
-		nskb = skb_copy(skb, GFP_ATOMIC);
+	if (np->bounce || unlikely(PAGE_SIZE - offset < ETH_HLEN)) {
+		nskb = bounce_skb(skb);
 		if (!nskb)
 			goto drop;
 		dev_consume_skb_any(skb);
@@ -756,6 +801,38 @@ 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)
 {
@@ -807,7 +884,7 @@ static int xennet_get_extras(struct netfront_queue *queue,
 		xennet_move_rx_slot(queue, skb, ref);
 	} while (extra.flags & XEN_NETIF_EXTRA_FLAG_MORE);
 
-	queue->rx.rsp_cons = cons;
+	xennet_set_rx_rsp_cons(queue, cons);
 	return err;
 }
 
@@ -824,7 +901,6 @@ 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);
@@ -855,8 +931,13 @@ 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);
 
@@ -887,7 +968,7 @@ next:
 	}
 
 	if (unlikely(err))
-		queue->rx.rsp_cons = cons + slots;
+		xennet_set_rx_rsp_cons(queue, cons + slots);
 
 	return err;
 }
@@ -941,7 +1022,8 @@ 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;
 		}
@@ -954,7 +1036,7 @@ static int xennet_fill_frags(struct netfront_queue *queue,
 		kfree_skb(nskb);
 	}
 
-	queue->rx.rsp_cons = cons;
+	xennet_set_rx_rsp_cons(queue, cons);
 
 	return 0;
 }
@@ -1059,6 +1141,10 @@ static int xennet_poll(struct napi_struct *napi, int budget)
 		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);
@@ -1075,7 +1161,9 @@ 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;
 			}
 		}
@@ -1099,7 +1187,8 @@ 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++;
 	}
 
@@ -1261,40 +1350,79 @@ 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 (queue->info->broken)
-		return IRQ_HANDLED;
+	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 (queue->info->broken)
-		return IRQ_HANDLED;
+	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;
 
-	if (likely(netif_carrier_ok(dev) &&
-		   RING_HAS_UNCONSUMED_RESPONSES(&queue->rx)))
+		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;
 }
 
@@ -1315,6 +1443,7 @@ static void xennet_poll_controller(struct net_device *dev)
 #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,
@@ -1495,6 +1624,12 @@ static int netfront_resume(struct xenbus_device *dev)
 	netif_tx_unlock_bh(info->netdev);
 
 	xennet_disconnect_backend(info);
+
+	rtnl_lock();
+	if (info->queues)
+		xennet_destroy_queues(info);
+	rtnl_unlock();
+
 	return 0;
 }
 
@@ -1528,9 +1663,10 @@ 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;
@@ -1558,18 +1694,18 @@ 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;
@@ -1593,7 +1729,7 @@ 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;
 
@@ -1613,21 +1749,21 @@ 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)
@@ -1640,22 +1776,28 @@ 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;
 }
 
@@ -1671,6 +1813,7 @@ 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);
 
@@ -1800,22 +1943,6 @@ 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)
 {
@@ -1871,6 +1998,10 @@ static int talk_to_netback(struct xenbus_device *dev,
 
 	info->netdev->irq = 0;
 
+	/* Check if backend is trusted. */
+	info->bounce = !xennet_trusted ||
+		       !xenbus_read_unsigned(dev->nodename, "trusted", 1);
+
 	/* Check if backend supports multiple queues */
 	max_queues = xenbus_read_unsigned(info->xbdev->otherend,
 					  "multi-queue-max-queues", 1);
@@ -2024,6 +2155,9 @@ static int xennet_connect(struct net_device *dev)
 	err = talk_to_netback(np->xbdev, np);
 	if (err)
 		return err;
+	if (np->bounce)
+		dev_info(&np->xbdev->dev,
+			 "bouncing transmitted data to zeroed pages\n");
 
 	/* talk_to_netback() sets the correct number of queues */
 	num_queues = dev->real_num_tx_queues;