X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fvirtio%2Fvirtio_ethdev.c;h=41042cb2343c7d355e10aa1e2b0fd18b38b3d8fc;hb=caccf8b318cafcdafe39faa3c5ce3eef67007621;hp=e6b537b17f80e05168a29d02c677ebf33f97c7ff;hpb=a2ffb87b1b60ded74a5e73be69503065faa827c7;p=dpdk.git

diff --git a/drivers/net/virtio/virtio_ethdev.c b/drivers/net/virtio/virtio_ethdev.c
index e6b537b17f..41042cb234 100644
--- a/drivers/net/virtio/virtio_ethdev.c
+++ b/drivers/net/virtio/virtio_ethdev.c
@@ -8,17 +8,18 @@
 #include <errno.h>
 #include <unistd.h>
 
-#include <rte_ethdev.h>
+#include <rte_ethdev_driver.h>
 #include <rte_ethdev_pci.h>
 #include <rte_memcpy.h>
 #include <rte_string_fns.h>
 #include <rte_memzone.h>
 #include <rte_malloc.h>
-#include <rte_atomic.h>
 #include <rte_branch_prediction.h>
 #include <rte_pci.h>
 #include <rte_bus_pci.h>
 #include <rte_ether.h>
+#include <rte_ip.h>
+#include <rte_arp.h>
 #include <rte_common.h>
 #include <rte_errno.h>
 #include <rte_cpuflags.h>
@@ -26,6 +27,7 @@
 #include <rte_memory.h>
 #include <rte_eal.h>
 #include <rte_dev.h>
+#include <rte_cycles.h>
 
 #include "virtio_ethdev.h"
 #include "virtio_pci.h"
@@ -65,7 +67,7 @@ static int virtio_mac_addr_add(struct rte_eth_dev *dev,
 				struct ether_addr *mac_addr,
 				uint32_t index, uint32_t vmdq);
 static void virtio_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index);
-static void virtio_mac_addr_set(struct rte_eth_dev *dev,
+static int virtio_mac_addr_set(struct rte_eth_dev *dev,
 				struct ether_addr *mac_addr);
 
 static int virtio_intr_enable(struct rte_eth_dev *dev);
@@ -80,6 +82,9 @@ static int virtio_dev_queue_stats_mapping_set(
 int virtio_logtype_init;
 int virtio_logtype_driver;
 
+static void virtio_notify_peers(struct rte_eth_dev *dev);
+static void virtio_ack_link_announce(struct rte_eth_dev *dev);
+
 /*
  * The set of PCI devices this driver supports
  */
@@ -273,17 +278,6 @@ virtio_dev_queue_release(void *queue __rte_unused)
 	/* do nothing */
 }
 
-static int
-virtio_get_queue_type(struct virtio_hw *hw, uint16_t vtpci_queue_idx)
-{
-	if (vtpci_queue_idx == hw->max_queue_pairs * 2)
-		return VTNET_CQ;
-	else if (vtpci_queue_idx % 2 == 0)
-		return VTNET_RQ;
-	else
-		return VTNET_TQ;
-}
-
 static uint16_t
 virtio_get_nr_vq(struct virtio_hw *hw)
 {
@@ -397,8 +391,8 @@ virtio_init_queue(struct rte_eth_dev *dev, uint16_t vtpci_queue_idx)
 		     size, vq->vq_ring_size);
 
 	mz = rte_memzone_reserve_aligned(vq_name, vq->vq_ring_size,
-					 SOCKET_ID_ANY,
-					 0, VIRTIO_PCI_VRING_ALIGN);
+			SOCKET_ID_ANY, RTE_MEMZONE_IOVA_CONTIG,
+			VIRTIO_PCI_VRING_ALIGN);
 	if (mz == NULL) {
 		if (rte_errno == EEXIST)
 			mz = rte_memzone_lookup(vq_name);
@@ -423,8 +417,8 @@ virtio_init_queue(struct rte_eth_dev *dev, uint16_t vtpci_queue_idx)
 		snprintf(vq_hdr_name, sizeof(vq_hdr_name), "port%d_vq%d_hdr",
 			 dev->data->port_id, vtpci_queue_idx);
 		hdr_mz = rte_memzone_reserve_aligned(vq_hdr_name, sz_hdr_mz,
-						     SOCKET_ID_ANY, 0,
-						     RTE_CACHE_LINE_SIZE);
+				SOCKET_ID_ANY, RTE_MEMZONE_IOVA_CONTIG,
+				RTE_CACHE_LINE_SIZE);
 		if (hdr_mz == NULL) {
 			if (rte_errno == EEXIST)
 				hdr_mz = rte_memzone_lookup(vq_hdr_name);
@@ -779,46 +773,6 @@ static const struct eth_dev_ops virtio_eth_dev_ops = {
 	.mac_addr_set            = virtio_mac_addr_set,
 };
 
-static inline int
-virtio_dev_atomic_read_link_status(struct rte_eth_dev *dev,
-				struct rte_eth_link *link)
-{
-	struct rte_eth_link *dst = link;
-	struct rte_eth_link *src = &(dev->data->dev_link);
-
-	if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
-			*(uint64_t *)src) == 0)
-		return -1;
-
-	return 0;
-}
-
-/**
- * Atomically writes the link status information into global
- * structure rte_eth_dev.
- *
- * @param dev
- *   - Pointer to the structure rte_eth_dev to read from.
- *   - Pointer to the buffer to be saved with the link status.
- *
- * @return
- *   - On success, zero.
- *   - On failure, negative value.
- */
-static inline int
-virtio_dev_atomic_write_link_status(struct rte_eth_dev *dev,
-		struct rte_eth_link *link)
-{
-	struct rte_eth_link *dst = &(dev->data->dev_link);
-	struct rte_eth_link *src = link;
-
-	if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
-					*(uint64_t *)src) == 0)
-		return -1;
-
-	return 0;
-}
-
 static void
 virtio_update_stats(struct rte_eth_dev *dev, struct rte_eth_stats *stats)
 {
@@ -1102,7 +1056,7 @@ virtio_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index)
 	virtio_mac_table_set(hw, uc, mc);
 }
 
-static void
+static int
 virtio_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr)
 {
 	struct virtio_hw *hw = dev->data->dev_private;
@@ -1118,9 +1072,14 @@ virtio_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr)
 		ctrl.hdr.cmd = VIRTIO_NET_CTRL_MAC_ADDR_SET;
 
 		memcpy(ctrl.data, mac_addr, ETHER_ADDR_LEN);
-		virtio_send_command(hw->cvq, &ctrl, &len, 1);
-	} else if (vtpci_with_feature(hw, VIRTIO_NET_F_MAC))
-		virtio_set_hwaddr(hw);
+		return virtio_send_command(hw->cvq, &ctrl, &len, 1);
+	}
+
+	if (!vtpci_with_feature(hw, VIRTIO_NET_F_MAC))
+		return -ENOTSUP;
+
+	virtio_set_hwaddr(hw);
+	return 0;
 }
 
 static int
@@ -1223,9 +1182,101 @@ virtio_negotiate_features(struct virtio_hw *hw, uint64_t req_features)
 	return 0;
 }
 
+int
+virtio_dev_pause(struct rte_eth_dev *dev)
+{
+	struct virtio_hw *hw = dev->data->dev_private;
+
+	rte_spinlock_lock(&hw->state_lock);
+
+	if (hw->started == 0) {
+		/* Device is just stopped. */
+		rte_spinlock_unlock(&hw->state_lock);
+		return -1;
+	}
+	hw->started = 0;
+	/*
+	 * Prevent the worker threads from touching queues to avoid contention,
+	 * 1 ms should be enough for the ongoing Tx function to finish.
+	 */
+	rte_delay_ms(1);
+	return 0;
+}
+
 /*
- * Process Virtio Config changed interrupt and call the callback
- * if link state changed.
+ * Recover hw state to let the worker threads continue.
+ */
+void
+virtio_dev_resume(struct rte_eth_dev *dev)
+{
+	struct virtio_hw *hw = dev->data->dev_private;
+
+	hw->started = 1;
+	rte_spinlock_unlock(&hw->state_lock);
+}
+
+/*
+ * Should be called only after device is paused.
+ */
+int
+virtio_inject_pkts(struct rte_eth_dev *dev, struct rte_mbuf **tx_pkts,
+		int nb_pkts)
+{
+	struct virtio_hw *hw = dev->data->dev_private;
+	struct virtnet_tx *txvq = dev->data->tx_queues[0];
+	int ret;
+
+	hw->inject_pkts = tx_pkts;
+	ret = dev->tx_pkt_burst(txvq, tx_pkts, nb_pkts);
+	hw->inject_pkts = NULL;
+
+	return ret;
+}
+
+static void
+virtio_notify_peers(struct rte_eth_dev *dev)
+{
+	struct virtio_hw *hw = dev->data->dev_private;
+	struct virtnet_rx *rxvq;
+	struct rte_mbuf *rarp_mbuf;
+
+	if (!dev->data->rx_queues)
+		return;
+
+	rxvq = dev->data->rx_queues[0];
+	rarp_mbuf = rte_net_make_rarp_packet(rxvq->mpool,
+			(struct ether_addr *)hw->mac_addr);
+	if (rarp_mbuf == NULL) {
+		PMD_DRV_LOG(ERR, "failed to make RARP packet.");
+		return;
+	}
+
+	/* If virtio port just stopped, no need to send RARP */
+	if (virtio_dev_pause(dev) < 0) {
+		rte_pktmbuf_free(rarp_mbuf);
+		return;
+	}
+
+	virtio_inject_pkts(dev, &rarp_mbuf, 1);
+	virtio_dev_resume(dev);
+}
+
+static void
+virtio_ack_link_announce(struct rte_eth_dev *dev)
+{
+	struct virtio_hw *hw = dev->data->dev_private;
+	struct virtio_pmd_ctrl ctrl;
+
+	ctrl.hdr.class = VIRTIO_NET_CTRL_ANNOUNCE;
+	ctrl.hdr.cmd = VIRTIO_NET_CTRL_ANNOUNCE_ACK;
+
+	virtio_send_command(hw->cvq, &ctrl, NULL, 0);
+}
+
+/*
+ * Process virtio config changed interrupt. Call the callback
+ * if link state changed, generate gratuitous RARP packet if
+ * the status indicates an ANNOUNCE.
  */
 void
 virtio_interrupt_handler(void *param)
@@ -1248,6 +1299,11 @@ virtio_interrupt_handler(void *param)
 						      NULL);
 	}
 
+	if (isr & VIRTIO_NET_S_ANNOUNCE) {
+		virtio_notify_peers(dev);
+		if (hw->cvq)
+			virtio_ack_link_announce(dev);
+	}
 }
 
 /* set rx and tx handlers according to what is supported */
@@ -1384,6 +1440,11 @@ virtio_init_device(struct rte_eth_dev *eth_dev, uint64_t req_features)
 	/* Reset the device although not necessary at startup */
 	vtpci_reset(hw);
 
+	if (hw->vqs) {
+		virtio_dev_free_mbufs(eth_dev);
+		virtio_free_queues(hw);
+	}
+
 	/* Tell the host we've noticed this device. */
 	vtpci_set_status(hw, VIRTIO_CONFIG_STATUS_ACK);
 
@@ -1695,6 +1756,7 @@ virtio_dev_configure(struct rte_eth_dev *dev)
 {
 	const struct rte_eth_rxmode *rxmode = &dev->data->dev_conf.rxmode;
 	struct virtio_hw *hw = dev->data->dev_private;
+	uint64_t rx_offloads = rxmode->offloads;
 	uint64_t req_features;
 	int ret;
 
@@ -1707,14 +1769,11 @@ virtio_dev_configure(struct rte_eth_dev *dev)
 			return ret;
 	}
 
-	/* The name hw_ip_checksum is a bit confusing since it can be
-	 * set by the application to request L3 and/or L4 checksums. In
-	 * case of virtio, only L4 checksum is supported.
-	 */
-	if (rxmode->hw_ip_checksum)
+	if (rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
+			   DEV_RX_OFFLOAD_TCP_CKSUM))
 		req_features |= (1ULL << VIRTIO_NET_F_GUEST_CSUM);
 
-	if (rxmode->enable_lro)
+	if (rx_offloads & DEV_RX_OFFLOAD_TCP_LRO)
 		req_features |=
 			(1ULL << VIRTIO_NET_F_GUEST_TSO4) |
 			(1ULL << VIRTIO_NET_F_GUEST_TSO6);
@@ -1726,14 +1785,15 @@ virtio_dev_configure(struct rte_eth_dev *dev)
 			return ret;
 	}
 
-	if (rxmode->hw_ip_checksum &&
+	if ((rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
+			    DEV_RX_OFFLOAD_TCP_CKSUM)) &&
 		!vtpci_with_feature(hw, VIRTIO_NET_F_GUEST_CSUM)) {
 		PMD_DRV_LOG(ERR,
 			"rx checksum not available on this host");
 		return -ENOTSUP;
 	}
 
-	if (rxmode->enable_lro &&
+	if ((rx_offloads & DEV_RX_OFFLOAD_TCP_LRO) &&
 		(!vtpci_with_feature(hw, VIRTIO_NET_F_GUEST_TSO4) ||
 		 !vtpci_with_feature(hw, VIRTIO_NET_F_GUEST_TSO6))) {
 		PMD_DRV_LOG(ERR,
@@ -1745,9 +1805,10 @@ virtio_dev_configure(struct rte_eth_dev *dev)
 	if (vtpci_with_feature(hw, VIRTIO_NET_F_CTRL_VQ))
 		virtio_dev_cq_start(dev);
 
-	hw->vlan_strip = rxmode->hw_vlan_strip;
+	if (rx_offloads & DEV_RX_OFFLOAD_VLAN_STRIP)
+		hw->vlan_strip = 1;
 
-	if (rxmode->hw_vlan_filter
+	if ((rx_offloads & DEV_RX_OFFLOAD_VLAN_FILTER)
 	    && !vtpci_with_feature(hw, VIRTIO_NET_F_CTRL_VLAN)) {
 		PMD_DRV_LOG(ERR,
 			    "vlan filtering not available on this host");
@@ -1762,10 +1823,12 @@ virtio_dev_configure(struct rte_eth_dev *dev)
 			return -EBUSY;
 		}
 
+	rte_spinlock_init(&hw->state_lock);
+
 	hw->use_simple_rx = 1;
 	hw->use_simple_tx = 1;
 
-#if defined RTE_ARCH_ARM64 || defined CONFIG_RTE_ARCH_ARM
+#if defined RTE_ARCH_ARM64 || defined RTE_ARCH_ARM
 	if (!rte_cpu_get_flag_enabled(RTE_CPUFLAG_NEON)) {
 		hw->use_simple_rx = 0;
 		hw->use_simple_tx = 0;
@@ -1776,7 +1839,8 @@ virtio_dev_configure(struct rte_eth_dev *dev)
 		hw->use_simple_tx = 0;
 	}
 
-	if (rxmode->hw_ip_checksum)
+	if (rx_offloads & (DEV_RX_OFFLOAD_UDP_CKSUM |
+			   DEV_RX_OFFLOAD_TCP_CKSUM))
 		hw->use_simple_rx = 0;
 
 	return 0;
@@ -1873,47 +1937,47 @@ virtio_dev_start(struct rte_eth_dev *dev)
 
 static void virtio_dev_free_mbufs(struct rte_eth_dev *dev)
 {
+	struct virtio_hw *hw = dev->data->dev_private;
+	uint16_t nr_vq = virtio_get_nr_vq(hw);
+	const char *type __rte_unused;
+	unsigned int i, mbuf_num = 0;
+	struct virtqueue *vq;
 	struct rte_mbuf *buf;
-	int i, mbuf_num = 0;
-
-	for (i = 0; i < dev->data->nb_rx_queues; i++) {
-		struct virtnet_rx *rxvq = dev->data->rx_queues[i];
-
-		PMD_INIT_LOG(DEBUG,
-			     "Before freeing rxq[%d] used and unused buf", i);
-		VIRTQUEUE_DUMP(rxvq->vq);
+	int queue_type;
 
-		PMD_INIT_LOG(DEBUG, "rx_queues[%d]=%p", i, rxvq);
-		while ((buf = virtqueue_detatch_unused(rxvq->vq)) != NULL) {
-			rte_pktmbuf_free(buf);
-			mbuf_num++;
-		}
+	if (hw->vqs == NULL)
+		return;
 
-		PMD_INIT_LOG(DEBUG, "free %d mbufs", mbuf_num);
-		PMD_INIT_LOG(DEBUG,
-			     "After freeing rxq[%d] used and unused buf", i);
-		VIRTQUEUE_DUMP(rxvq->vq);
-	}
+	for (i = 0; i < nr_vq; i++) {
+		vq = hw->vqs[i];
+		if (!vq)
+			continue;
 
-	for (i = 0; i < dev->data->nb_tx_queues; i++) {
-		struct virtnet_tx *txvq = dev->data->tx_queues[i];
+		queue_type = virtio_get_queue_type(hw, i);
+		if (queue_type == VTNET_RQ)
+			type = "rxq";
+		else if (queue_type == VTNET_TQ)
+			type = "txq";
+		else
+			continue;
 
 		PMD_INIT_LOG(DEBUG,
-			     "Before freeing txq[%d] used and unused bufs",
-			     i);
-		VIRTQUEUE_DUMP(txvq->vq);
+			"Before freeing %s[%d] used and unused buf",
+			type, i);
+		VIRTQUEUE_DUMP(vq);
 
-		mbuf_num = 0;
-		while ((buf = virtqueue_detatch_unused(txvq->vq)) != NULL) {
+		while ((buf = virtqueue_detach_unused(vq)) != NULL) {
 			rte_pktmbuf_free(buf);
 			mbuf_num++;
 		}
 
-		PMD_INIT_LOG(DEBUG, "free %d mbufs", mbuf_num);
 		PMD_INIT_LOG(DEBUG,
-			     "After freeing txq[%d] used and unused buf", i);
-		VIRTQUEUE_DUMP(txvq->vq);
+			"After freeing %s[%d] used and unused buf",
+			type, i);
+		VIRTQUEUE_DUMP(vq);
 	}
+
+	PMD_INIT_LOG(DEBUG, "%d mbufs freed", mbuf_num);
 }
 
 /*
@@ -1928,25 +1992,27 @@ virtio_dev_stop(struct rte_eth_dev *dev)
 
 	PMD_INIT_LOG(DEBUG, "stop");
 
+	rte_spinlock_lock(&hw->state_lock);
 	if (intr_conf->lsc || intr_conf->rxq)
 		virtio_intr_disable(dev);
 
 	hw->started = 0;
 	memset(&link, 0, sizeof(link));
-	virtio_dev_atomic_write_link_status(dev, &link);
+	rte_eth_linkstatus_set(dev, &link);
+	rte_spinlock_unlock(&hw->state_lock);
 }
 
 static int
 virtio_dev_link_update(struct rte_eth_dev *dev, __rte_unused int wait_to_complete)
 {
-	struct rte_eth_link link, old;
+	struct rte_eth_link link;
 	uint16_t status;
 	struct virtio_hw *hw = dev->data->dev_private;
+
 	memset(&link, 0, sizeof(link));
-	virtio_dev_atomic_read_link_status(dev, &link);
-	old = link;
 	link.link_duplex = ETH_LINK_FULL_DUPLEX;
 	link.link_speed  = ETH_SPEED_NUM_10G;
+	link.link_autoneg = ETH_LINK_FIXED;
 
 	if (hw->started == 0) {
 		link.link_status = ETH_LINK_DOWN;
@@ -1967,9 +2033,8 @@ virtio_dev_link_update(struct rte_eth_dev *dev, __rte_unused int wait_to_complet
 	} else {
 		link.link_status = ETH_LINK_UP;
 	}
-	virtio_dev_atomic_write_link_status(dev, &link);
 
-	return (old.link_status == link.link_status) ? -1 : 0;
+	return rte_eth_linkstatus_set(dev, &link);
 }
 
 static int
@@ -1977,9 +2042,10 @@ virtio_dev_vlan_offload_set(struct rte_eth_dev *dev, int mask)
 {
 	const struct rte_eth_rxmode *rxmode = &dev->data->dev_conf.rxmode;
 	struct virtio_hw *hw = dev->data->dev_private;
+	uint64_t offloads = rxmode->offloads;
 
 	if (mask & ETH_VLAN_FILTER_MASK) {
-		if (rxmode->hw_vlan_filter &&
+		if ((offloads & DEV_RX_OFFLOAD_VLAN_FILTER) &&
 				!vtpci_with_feature(hw, VIRTIO_NET_F_CTRL_VLAN)) {
 
 			PMD_DRV_LOG(NOTICE,
@@ -1990,7 +2056,7 @@ virtio_dev_vlan_offload_set(struct rte_eth_dev *dev, int mask)
 	}
 
 	if (mask & ETH_VLAN_STRIP_MASK)
-		hw->vlan_strip = rxmode->hw_vlan_strip;
+		hw->vlan_strip = !!(offloads & DEV_RX_OFFLOAD_VLAN_STRIP);
 
 	return 0;
 }
@@ -2003,7 +2069,6 @@ virtio_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
 
 	dev_info->speed_capa = ETH_LINK_SPEED_10G; /* fake value */
 
-	dev_info->pci_dev = dev->device ? RTE_ETH_DEV_TO_PCI(dev) : NULL;
 	dev_info->max_rx_queues =
 		RTE_MIN(hw->max_queue_pairs, VIRTIO_MAX_RX_QUEUES);
 	dev_info->max_tx_queues =
@@ -2016,18 +2081,21 @@ virtio_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
 	};
 
 	host_features = VTPCI_OPS(hw)->get_features(hw);
-	dev_info->rx_offload_capa = 0;
+	dev_info->rx_offload_capa = DEV_RX_OFFLOAD_VLAN_STRIP;
 	if (host_features & (1ULL << VIRTIO_NET_F_GUEST_CSUM)) {
 		dev_info->rx_offload_capa |=
 			DEV_RX_OFFLOAD_TCP_CKSUM |
 			DEV_RX_OFFLOAD_UDP_CKSUM;
 	}
+	if (host_features & (1ULL << VIRTIO_NET_F_CTRL_VLAN))
+		dev_info->rx_offload_capa |= DEV_RX_OFFLOAD_VLAN_FILTER;
 	tso_mask = (1ULL << VIRTIO_NET_F_GUEST_TSO4) |
 		(1ULL << VIRTIO_NET_F_GUEST_TSO6);
 	if ((host_features & tso_mask) == tso_mask)
 		dev_info->rx_offload_capa |= DEV_RX_OFFLOAD_TCP_LRO;
 
-	dev_info->tx_offload_capa = 0;
+	dev_info->tx_offload_capa = DEV_TX_OFFLOAD_MULTI_SEGS |
+				    DEV_TX_OFFLOAD_VLAN_INSERT;
 	if (hw->guest_features & (1ULL << VIRTIO_NET_F_CSUM)) {
 		dev_info->tx_offload_capa |=
 			DEV_TX_OFFLOAD_UDP_CKSUM |
@@ -2058,10 +2126,10 @@ RTE_INIT(virtio_init_log);
 static void
 virtio_init_log(void)
 {
-	virtio_logtype_init = rte_log_register("pmd.virtio.init");
+	virtio_logtype_init = rte_log_register("pmd.net.virtio.init");
 	if (virtio_logtype_init >= 0)
 		rte_log_set_level(virtio_logtype_init, RTE_LOG_NOTICE);
-	virtio_logtype_driver = rte_log_register("pmd.virtio.driver");
+	virtio_logtype_driver = rte_log_register("pmd.net.virtio.driver");
 	if (virtio_logtype_driver >= 0)
 		rte_log_set_level(virtio_logtype_driver, RTE_LOG_NOTICE);
 }