1 files changed, 103 insertions, 148 deletions

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 32266fb89c2..b9e002fccbc 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -1,4 +1,4 @@
-/* A simple network driver using virtio.
+/* A network driver using virtio.
  *
  * Copyright 2007 Rusty Russell <rusty@rustcorp.com.au> IBM Corporation
  *
@@ -47,19 +47,9 @@ struct virtnet_info
 	struct napi_struct napi;
 	unsigned int status;
 
-	/* The skb we couldn't send because buffers were full. */
-	struct sk_buff *last_xmit_skb;
-
-	/* If we need to free in a timer, this is it. */
-	struct timer_list xmit_free_timer;
-
 	/* Number of input buffers, and max we've ever had. */
 	unsigned int num, max;
 
-	/* For cleaning up after transmission. */
-	struct tasklet_struct tasklet;
-	bool free_in_tasklet;
-
 	/* I like... big packets and I cannot lie! */
 	bool big_packets;
 
@@ -77,9 +67,17 @@ struct virtnet_info
 	struct page *pages;
 };
 
-static inline void *skb_vnet_hdr(struct sk_buff *skb)
+struct skb_vnet_hdr {
+	union {
+		struct virtio_net_hdr hdr;
+		struct virtio_net_hdr_mrg_rxbuf mhdr;
+	};
+	unsigned int num_sg;
+};
+
+static inline struct skb_vnet_hdr *skb_vnet_hdr(struct sk_buff *skb)
 {
-	return (struct virtio_net_hdr *)skb->cb;
+	return (struct skb_vnet_hdr *)skb->cb;
 }
 
 static void give_a_page(struct virtnet_info *vi, struct page *page)
@@ -118,17 +116,13 @@ static void skb_xmit_done(struct virtqueue *svq)
 
 	/* We were probably waiting for more output buffers. */
 	netif_wake_queue(vi->dev);
-
-	/* Make sure we re-xmit last_xmit_skb: if there are no more packets
-	 * queued, start_xmit won't be called. */
-	tasklet_schedule(&vi->tasklet);
 }
 
 static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 			unsigned len)
 {
 	struct virtnet_info *vi = netdev_priv(dev);
-	struct virtio_net_hdr *hdr = skb_vnet_hdr(skb);
+	struct skb_vnet_hdr *hdr = skb_vnet_hdr(skb);
 	int err;
 	int i;
 
@@ -139,7 +133,6 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 	}
 
 	if (vi->mergeable_rx_bufs) {
-		struct virtio_net_hdr_mrg_rxbuf *mhdr = skb_vnet_hdr(skb);
 		unsigned int copy;
 		char *p = page_address(skb_shinfo(skb)->frags[0].page);
 
@@ -147,8 +140,8 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 			len = PAGE_SIZE;
 		len -= sizeof(struct virtio_net_hdr_mrg_rxbuf);
 
-		memcpy(hdr, p, sizeof(*mhdr));
-		p += sizeof(*mhdr);
+		memcpy(&hdr->mhdr, p, sizeof(hdr->mhdr));
+		p += sizeof(hdr->mhdr);
 
 		copy = len;
 		if (copy > skb_tailroom(skb))
@@ -163,13 +156,13 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 			skb_shinfo(skb)->nr_frags--;
 		} else {
 			skb_shinfo(skb)->frags[0].page_offset +=
-				sizeof(*mhdr) + copy;
+				sizeof(hdr->mhdr) + copy;
 			skb_shinfo(skb)->frags[0].size = len;
 			skb->data_len += len;
 			skb->len += len;
 		}
 
-		while (--mhdr->num_buffers) {
+		while (--hdr->mhdr.num_buffers) {
 			struct sk_buff *nskb;
 
 			i = skb_shinfo(skb)->nr_frags;
@@ -183,7 +176,7 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 			nskb = vi->rvq->vq_ops->get_buf(vi->rvq, &len);
 			if (!nskb) {
 				pr_debug("%s: rx error: %d buffers missing\n",
-					 dev->name, mhdr->num_buffers);
+					 dev->name, hdr->mhdr.num_buffers);
 				dev->stats.rx_length_errors++;
 				goto drop;
 			}
@@ -204,7 +197,7 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 			skb->len += len;
 		}
 	} else {
-		len -= sizeof(struct virtio_net_hdr);
+		len -= sizeof(hdr->hdr);
 
 		if (len <= MAX_PACKET_LEN)
 			trim_pages(vi, skb);
@@ -222,9 +215,11 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 	dev->stats.rx_bytes += skb->len;
 	dev->stats.rx_packets++;
 
-	if (hdr->flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
+	if (hdr->hdr.flags & VIRTIO_NET_HDR_F_NEEDS_CSUM) {
 		pr_debug("Needs csum!\n");
-		if (!skb_partial_csum_set(skb,hdr->csum_start,hdr->csum_offset))
+		if (!skb_partial_csum_set(skb,
+					  hdr->hdr.csum_start,
+					  hdr->hdr.csum_offset))
 			goto frame_err;
 	}
 
@@ -232,9 +227,9 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 	pr_debug("Receiving skb proto 0x%04x len %i type %i\n",
 		 ntohs(skb->protocol), skb->len, skb->pkt_type);
 
-	if (hdr->gso_type != VIRTIO_NET_HDR_GSO_NONE) {
+	if (hdr->hdr.gso_type != VIRTIO_NET_HDR_GSO_NONE) {
 		pr_debug("GSO!\n");
-		switch (hdr->gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
+		switch (hdr->hdr.gso_type & ~VIRTIO_NET_HDR_GSO_ECN) {
 		case VIRTIO_NET_HDR_GSO_TCPV4:
 			skb_shinfo(skb)->gso_type = SKB_GSO_TCPV4;
 			break;
@@ -247,14 +242,14 @@ static void receive_skb(struct net_device *dev, struct sk_buff *skb,
 		default:
 			if (net_ratelimit())
 				printk(KERN_WARNING "%s: bad gso type %u.\n",
-				       dev->name, hdr->gso_type);
+				       dev->name, hdr->hdr.gso_type);
 			goto frame_err;
 		}
 
-		if (hdr->gso_type & VIRTIO_NET_HDR_GSO_ECN)
+		if (hdr->hdr.gso_type & VIRTIO_NET_HDR_GSO_ECN)
 			skb_shinfo(skb)->gso_type |= SKB_GSO_TCP_ECN;
 
-		skb_shinfo(skb)->gso_size = hdr->gso_size;
+		skb_shinfo(skb)->gso_size = hdr->hdr.gso_size;
 		if (skb_shinfo(skb)->gso_size == 0) {
 			if (net_ratelimit())
 				printk(KERN_WARNING "%s: zero gso size.\n",
@@ -284,8 +279,8 @@ static bool try_fill_recv_maxbufs(struct virtnet_info *vi, gfp_t gfp)
 	bool oom = false;
 
 	sg_init_table(sg, 2+MAX_SKB_FRAGS);
-	for (;;) {
-		struct virtio_net_hdr *hdr;
+	do {
+		struct skb_vnet_hdr *hdr;
 
 		skb = netdev_alloc_skb(vi->dev, MAX_PACKET_LEN + NET_IP_ALIGN);
 		if (unlikely(!skb)) {
@@ -297,7 +292,7 @@ static bool try_fill_recv_maxbufs(struct virtnet_info *vi, gfp_t gfp)
 		skb_put(skb, MAX_PACKET_LEN);
 
 		hdr = skb_vnet_hdr(skb);
-		sg_set_buf(sg, hdr, sizeof(*hdr));
+		sg_set_buf(sg, &hdr->hdr, sizeof(hdr->hdr));
 
 		if (vi->big_packets) {
 			for (i = 0; i < MAX_SKB_FRAGS; i++) {
@@ -320,14 +315,14 @@ static bool try_fill_recv_maxbufs(struct virtnet_info *vi, gfp_t gfp)
 		skb_queue_head(&vi->recv, skb);
 
 		err = vi->rvq->vq_ops->add_buf(vi->rvq, sg, 0, num, skb);
-		if (err) {
+		if (err < 0) {
 			skb_unlink(skb, &vi->recv);
 			trim_pages(vi, skb);
 			kfree_skb(skb);
 			break;
 		}
 		vi->num++;
-	}
+	} while (err >= num);
 	if (unlikely(vi->num > vi->max))
 		vi->max = vi->num;
 	vi->rvq->vq_ops->kick(vi->rvq);
@@ -345,7 +340,7 @@ static bool try_fill_recv(struct virtnet_info *vi, gfp_t gfp)
 	if (!vi->mergeable_rx_bufs)
 		return try_fill_recv_maxbufs(vi, gfp);
 
-	for (;;) {
+	do {
 		skb_frag_t *f;
 
 		skb = netdev_alloc_skb(vi->dev, GOOD_COPY_LEN + NET_IP_ALIGN);
@@ -373,13 +368,13 @@ static bool try_fill_recv(struct virtnet_info *vi, gfp_t gfp)
 		skb_queue_head(&vi->recv, skb);
 
 		err = vi->rvq->vq_ops->add_buf(vi->rvq, sg, 0, 1, skb);
-		if (err) {
+		if (err < 0) {
 			skb_unlink(skb, &vi->recv);
 			kfree_skb(skb);
 			break;
 		}
 		vi->num++;
-	}
+	} while (err > 0);
 	if (unlikely(vi->num > vi->max))
 		vi->max = vi->num;
 	vi->rvq->vq_ops->kick(vi->rvq);
@@ -447,42 +442,26 @@ again:
 	return received;
 }
 
-static void free_old_xmit_skbs(struct virtnet_info *vi)
+static unsigned int free_old_xmit_skbs(struct virtnet_info *vi)
 {
 	struct sk_buff *skb;
-	unsigned int len;
+	unsigned int len, tot_sgs = 0;
 
 	while ((skb = vi->svq->vq_ops->get_buf(vi->svq, &len)) != NULL) {
 		pr_debug("Sent skb %p\n", skb);
 		__skb_unlink(skb, &vi->send);
 		vi->dev->stats.tx_bytes += skb->len;
 		vi->dev->stats.tx_packets++;
-		kfree_skb(skb);
+		tot_sgs += skb_vnet_hdr(skb)->num_sg;
+		dev_kfree_skb_any(skb);
 	}
-}
-
-/* If the virtio transport doesn't always notify us when all in-flight packets
- * are consumed, we fall back to using this function on a timer to free them. */
-static void xmit_free(unsigned long data)
-{
-	struct virtnet_info *vi = (void *)data;
-
-	netif_tx_lock(vi->dev);
-
-	free_old_xmit_skbs(vi);
-
-	if (!skb_queue_empty(&vi->send))
-		mod_timer(&vi->xmit_free_timer, jiffies + (HZ/10));
-
-	netif_tx_unlock(vi->dev);
+	return tot_sgs;
 }
 
 static int xmit_skb(struct virtnet_info *vi, struct sk_buff *skb)
 {
-	int num, err;
 	struct scatterlist sg[2+MAX_SKB_FRAGS];
-	struct virtio_net_hdr_mrg_rxbuf *mhdr = skb_vnet_hdr(skb);
-	struct virtio_net_hdr *hdr = skb_vnet_hdr(skb);
+	struct skb_vnet_hdr *hdr = skb_vnet_hdr(skb);
 	const unsigned char *dest = ((struct ethhdr *)skb->data)->h_dest;
 
 	sg_init_table(sg, 2+MAX_SKB_FRAGS);
@@ -490,108 +469,97 @@ static int xmit_skb(struct virtnet_info *vi, struct sk_buff *skb)
 	pr_debug("%s: xmit %p %pM\n", vi->dev->name, skb, dest);
 
 	if (skb->ip_summed == CHECKSUM_PARTIAL) {
-		hdr->flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
-		hdr->csum_start = skb->csum_start - skb_headroom(skb);
-		hdr->csum_offset = skb->csum_offset;
+		hdr->hdr.flags = VIRTIO_NET_HDR_F_NEEDS_CSUM;
+		hdr->hdr.csum_start = skb->csum_start - skb_headroom(skb);
+		hdr->hdr.csum_offset = skb->csum_offset;
 	} else {
-		hdr->flags = 0;
-		hdr->csum_offset = hdr->csum_start = 0;
+		hdr->hdr.flags = 0;
+		hdr->hdr.csum_offset = hdr->hdr.csum_start = 0;
 	}
 
 	if (skb_is_gso(skb)) {
-		hdr->hdr_len = skb_headlen(skb);
-		hdr->gso_size = skb_shinfo(skb)->gso_size;
+		hdr->hdr.hdr_len = skb_headlen(skb);
+		hdr->hdr.gso_size = skb_shinfo(skb)->gso_size;
 		if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV4)
-			hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
+			hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
 		else if (skb_shinfo(skb)->gso_type & SKB_GSO_TCPV6)
-			hdr->gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
+			hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_TCPV6;
 		else if (skb_shinfo(skb)->gso_type & SKB_GSO_UDP)
-			hdr->gso_type = VIRTIO_NET_HDR_GSO_UDP;
+			hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_UDP;
 		else
 			BUG();
 		if (skb_shinfo(skb)->gso_type & SKB_GSO_TCP_ECN)
-			hdr->gso_type |= VIRTIO_NET_HDR_GSO_ECN;
+			hdr->hdr.gso_type |= VIRTIO_NET_HDR_GSO_ECN;
 	} else {
-		hdr->gso_type = VIRTIO_NET_HDR_GSO_NONE;
-		hdr->gso_size = hdr->hdr_len = 0;
+		hdr->hdr.gso_type = VIRTIO_NET_HDR_GSO_NONE;
+		hdr->hdr.gso_size = hdr->hdr.hdr_len = 0;
 	}
 
-	mhdr->num_buffers = 0;
+	hdr->mhdr.num_buffers = 0;
 
 	/* Encode metadata header at front. */
 	if (vi->mergeable_rx_bufs)
-		sg_set_buf(sg, mhdr, sizeof(*mhdr));
+		sg_set_buf(sg, &hdr->mhdr, sizeof(hdr->mhdr));
 	else
-		sg_set_buf(sg, hdr, sizeof(*hdr));
-
-	num = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;
+		sg_set_buf(sg, &hdr->hdr, sizeof(hdr->hdr));
 
-	err = vi->svq->vq_ops->add_buf(vi->svq, sg, num, 0, skb);
-	if (!err && !vi->free_in_tasklet)
-		mod_timer(&vi->xmit_free_timer, jiffies + (HZ/10));
-
-	return err;
-}
-
-static void xmit_tasklet(unsigned long data)
-{
-	struct virtnet_info *vi = (void *)data;
-
-	netif_tx_lock_bh(vi->dev);
-	if (vi->last_xmit_skb && xmit_skb(vi, vi->last_xmit_skb) == 0) {
-		vi->svq->vq_ops->kick(vi->svq);
-		vi->last_xmit_skb = NULL;
-	}
-	if (vi->free_in_tasklet)
-		free_old_xmit_skbs(vi);
-	netif_tx_unlock_bh(vi->dev);
+	hdr->num_sg = skb_to_sgvec(skb, sg+1, 0, skb->len) + 1;
+	return vi->svq->vq_ops->add_buf(vi->svq, sg, hdr->num_sg, 0, skb);
 }
 
 static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
 {
 	struct virtnet_info *vi = netdev_priv(dev);
+	int capacity;
 
 again:
 	/* Free up any pending old buffers before queueing new ones. */
 	free_old_xmit_skbs(vi);
 
-	/* If we has a buffer left over from last time, send it now. */
-	if (unlikely(vi->last_xmit_skb) &&
-	    xmit_skb(vi, vi->last_xmit_skb) != 0)
-		goto stop_queue;
-
-	vi->last_xmit_skb = NULL;
+	/* Try to transmit */
+	capacity = xmit_skb(vi, skb);
 
-	/* Put new one in send queue and do transmit */
-	if (likely(skb)) {
-		__skb_queue_head(&vi->send, skb);
-		if (xmit_skb(vi, skb) != 0) {
-			vi->last_xmit_skb = skb;
-			skb = NULL;
-			goto stop_queue;
+	/* This can happen with OOM and indirect buffers. */
+	if (unlikely(capacity < 0)) {
+		netif_stop_queue(dev);
+		dev_warn(&dev->dev, "Unexpected full queue\n");
+		if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
+			vi->svq->vq_ops->disable_cb(vi->svq);
+			netif_start_queue(dev);
+			goto again;
 		}
+		return NETDEV_TX_BUSY;
 	}
-done:
 	vi->svq->vq_ops->kick(vi->svq);
-	return NETDEV_TX_OK;
-
-stop_queue:
-	pr_debug("%s: virtio not prepared to send\n", dev->name);
-	netif_stop_queue(dev);
 
-	/* Activate callback for using skbs: if this returns false it
-	 * means some were used in the meantime. */
-	if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
-		vi->svq->vq_ops->disable_cb(vi->svq);
-		netif_start_queue(dev);
-		goto again;
-	}
-	if (skb) {
-		/* Drop this skb: we only queue one. */
-		vi->dev->stats.tx_dropped++;
-		kfree_skb(skb);
+	/*
+	 * Put new one in send queue.  You'd expect we'd need this before
+	 * xmit_skb calls add_buf(), since the callback can be triggered
+	 * immediately after that.  But since the callback just triggers
+	 * another call back here, normal network xmit locking prevents the
+	 * race.
+	 */
+	__skb_queue_head(&vi->send, skb);
+
+	/* Don't wait up for transmitted skbs to be freed. */
+	skb_orphan(skb);
+	nf_reset(skb);
+
+	/* Apparently nice girls don't return TX_BUSY; stop the queue
+	 * before it gets out of hand.  Naturally, this wastes entries. */
+	if (capacity < 2+MAX_SKB_FRAGS) {
+		netif_stop_queue(dev);
+		if (unlikely(!vi->svq->vq_ops->enable_cb(vi->svq))) {
+			/* More just got used, free them then recheck. */
+			capacity += free_old_xmit_skbs(vi);
+			if (capacity >= 2+MAX_SKB_FRAGS) {
+				netif_start_queue(dev);
+				vi->svq->vq_ops->disable_cb(vi->svq);
+			}
+		}
 	}
-	goto done;
+
+	return NETDEV_TX_OK;
 }
 
 static int virtnet_set_mac_address(struct net_device *dev, void *p)
@@ -668,7 +636,7 @@ static bool virtnet_send_command(struct virtnet_info *vi, u8 class, u8 cmd,
 		sg_set_buf(&sg[i + 1], sg_virt(s), s->length);
 	sg_set_buf(&sg[out + in - 1], &status, sizeof(status));
 
-	BUG_ON(vi->cvq->vq_ops->add_buf(vi->cvq, sg, out, in, vi));
+	BUG_ON(vi->cvq->vq_ops->add_buf(vi->cvq, sg, out, in, vi) < 0);
 
 	vi->cvq->vq_ops->kick(vi->cvq);
 
@@ -924,10 +892,6 @@ static int virtnet_probe(struct virtio_device *vdev)
 	vi->pages = NULL;
 	INIT_DELAYED_WORK(&vi->refill, refill_work);
 
-	/* If they give us a callback when all buffers are done, we don't need
-	 * the timer. */
-	vi->free_in_tasklet = virtio_has_feature(vdev,VIRTIO_F_NOTIFY_ON_EMPTY);
-
 	/* If we can receive ANY GSO packets, we must allocate large ones. */
 	if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4)
 	    || virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO6)
@@ -959,11 +923,6 @@ static int virtnet_probe(struct virtio_device *vdev)
 	skb_queue_head_init(&vi->recv);
 	skb_queue_head_init(&vi->send);
 
-	tasklet_init(&vi->tasklet, xmit_tasklet, (unsigned long)vi);
-
-	if (!vi->free_in_tasklet)
-		setup_timer(&vi->xmit_free_timer, xmit_free, (unsigned long)vi);
-
 	err = register_netdev(dev);
 	if (err) {
 		pr_debug("virtio_net: registering device failed\n");
@@ -996,7 +955,7 @@ free:
 	return err;
 }
 
-static void virtnet_remove(struct virtio_device *vdev)
+static void __devexit virtnet_remove(struct virtio_device *vdev)
 {
 	struct virtnet_info *vi = vdev->priv;
 	struct sk_buff *skb;
@@ -1004,9 +963,6 @@ static void virtnet_remove(struct virtio_device *vdev)
 	/* Stop all the virtqueues. */
 	vdev->config->reset(vdev);
 
-	if (!vi->free_in_tasklet)
-		del_timer_sync(&vi->xmit_free_timer);
-
 	/* Free our skbs in send and recv queues, if any. */
 	while ((skb = __skb_dequeue(&vi->recv)) != NULL) {
 		kfree_skb(skb);
@@ -1040,10 +996,9 @@ static unsigned int features[] = {
 	VIRTIO_NET_F_GUEST_ECN, VIRTIO_NET_F_GUEST_UFO,
 	VIRTIO_NET_F_MRG_RXBUF, VIRTIO_NET_F_STATUS, VIRTIO_NET_F_CTRL_VQ,
 	VIRTIO_NET_F_CTRL_RX, VIRTIO_NET_F_CTRL_VLAN,
-	VIRTIO_F_NOTIFY_ON_EMPTY,
 };
 
-static struct virtio_driver virtio_net = {
+static struct virtio_driver virtio_net_driver = {
 	.feature_table = features,
 	.feature_table_size = ARRAY_SIZE(features),
 	.driver.name =	KBUILD_MODNAME,
@@ -1056,12 +1011,12 @@ static struct virtio_driver virtio_net = {
 
 static int __init init(void)
 {
-	return register_virtio_driver(&virtio_net);
+	return register_virtio_driver(&virtio_net_driver);
 }
 
 static void __exit fini(void)
 {
-	unregister_virtio_driver(&virtio_net);
+	unregister_virtio_driver(&virtio_net_driver);
 }
 module_init(init);
 module_exit(fini);