X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=lib%2Flibrte_vhost%2Fvirtio-net.c;h=90da9bafdac15aec22ca9d6476c3e8655143062e;hb=ae19955e7c86;hp=f984dc45959812b535ee9002e01874bfe32c6567;hpb=859b480d5afd2275252b2a6f49af5a16edd3a792;p=dpdk.git

diff --git a/lib/librte_vhost/virtio-net.c b/lib/librte_vhost/virtio-net.c
index f984dc4595..90da9bafda 100644
--- a/lib/librte_vhost/virtio-net.c
+++ b/lib/librte_vhost/virtio-net.c
@@ -55,18 +55,11 @@
 #include "vhost-net.h"
 #include "virtio-net.h"
 
-/*
- * Device linked list structure for configuration.
- */
-struct virtio_net_config_ll {
-	struct virtio_net dev;			/* Virtio device.*/
-	struct virtio_net_config_ll *next;	/* Next dev on linked list.*/
-};
+#define MAX_VHOST_DEVICE	1024
+static struct virtio_net *vhost_devices[MAX_VHOST_DEVICE];
 
 /* device ops to add/remove device to/from data core. */
 struct virtio_net_device_ops const *notify_ops;
-/* root address of the linked list of managed virtio devices */
-static struct virtio_net_config_ll *ll_root;
 
 #define VHOST_USER_F_PROTOCOL_FEATURES	30
 
@@ -74,6 +67,7 @@ static struct virtio_net_config_ll *ll_root;
 #define VHOST_SUPPORTED_FEATURES ((1ULL << VIRTIO_NET_F_MRG_RXBUF) | \
 				(1ULL << VIRTIO_NET_F_CTRL_VQ) | \
 				(1ULL << VIRTIO_NET_F_CTRL_RX) | \
+				(1ULL << VIRTIO_NET_F_GUEST_ANNOUNCE) | \
 				(VHOST_SUPPORTS_MQ)            | \
 				(1ULL << VIRTIO_F_VERSION_1)   | \
 				(1ULL << VHOST_F_LOG_ALL)      | \
@@ -115,77 +109,17 @@ qva_to_vva(struct virtio_net *dev, uint64_t qemu_va)
 }
 
 
-/*
- * Retrieves an entry from the devices configuration linked list.
- */
-static struct virtio_net_config_ll *
-get_config_ll_entry(struct vhost_device_ctx ctx)
-{
-	struct virtio_net_config_ll *ll_dev = ll_root;
-
-	/* Loop through linked list until the device_fh is found. */
-	while (ll_dev != NULL) {
-		if (ll_dev->dev.device_fh == ctx.fh)
-			return ll_dev;
-		ll_dev = ll_dev->next;
-	}
-
-	return NULL;
-}
-
-/*
- * Searches the configuration core linked list and
- * retrieves the device if it exists.
- */
 struct virtio_net *
 get_device(struct vhost_device_ctx ctx)
 {
-	struct virtio_net_config_ll *ll_dev;
-
-	ll_dev = get_config_ll_entry(ctx);
-
-	if (ll_dev)
-		return &ll_dev->dev;
+	struct virtio_net *dev = vhost_devices[ctx.fh];
 
-	RTE_LOG(ERR, VHOST_CONFIG,
-		"(%"PRIu64") Device not found in linked list.\n", ctx.fh);
-	return NULL;
-}
-
-/*
- * Add entry containing a device to the device configuration linked list.
- */
-static void
-add_config_ll_entry(struct virtio_net_config_ll *new_ll_dev)
-{
-	struct virtio_net_config_ll *ll_dev = ll_root;
-
-	/* If ll_dev == NULL then this is the first device so go to else */
-	if (ll_dev) {
-		/* If the 1st device_fh != 0 then we insert our device here. */
-		if (ll_dev->dev.device_fh != 0) {
-			new_ll_dev->dev.device_fh = 0;
-			new_ll_dev->next = ll_dev;
-			ll_root = new_ll_dev;
-		} else {
-			/*
-			 * Increment through the ll until we find un unused
-			 * device_fh. Insert the device at that entry.
-			 */
-			while ((ll_dev->next != NULL) &&
-				(ll_dev->dev.device_fh ==
-					(ll_dev->next->dev.device_fh - 1)))
-				ll_dev = ll_dev->next;
-
-			new_ll_dev->dev.device_fh = ll_dev->dev.device_fh + 1;
-			new_ll_dev->next = ll_dev->next;
-			ll_dev->next = new_ll_dev;
-		}
-	} else {
-		ll_root = new_ll_dev;
-		ll_root->dev.device_fh = 0;
+	if (unlikely(!dev)) {
+		RTE_LOG(ERR, VHOST_CONFIG,
+			"(%"PRIu64") device not found.\n", ctx.fh);
 	}
 
+	return dev;
 }
 
 static void
@@ -206,13 +140,7 @@ cleanup_device(struct virtio_net *dev, int destroy)
 {
 	uint32_t i;
 
-	/* Unmap QEMU memory file if mapped. */
-	if (dev->mem) {
-		munmap((void *)(uintptr_t)dev->mem->mapped_address,
-			(size_t)dev->mem->mapped_size);
-		free(dev->mem);
-		dev->mem = NULL;
-	}
+	vhost_backend_cleanup(dev);
 
 	for (i = 0; i < dev->virt_qp_nb; i++) {
 		cleanup_vq(dev->virtqueue[i * VIRTIO_QNUM + VIRTIO_RXQ], destroy);
@@ -224,43 +152,14 @@ cleanup_device(struct virtio_net *dev, int destroy)
  * Release virtqueues and device memory.
  */
 static void
-free_device(struct virtio_net_config_ll *ll_dev)
+free_device(struct virtio_net *dev)
 {
 	uint32_t i;
 
-	for (i = 0; i < ll_dev->dev.virt_qp_nb; i++)
-		rte_free(ll_dev->dev.virtqueue[i * VIRTIO_QNUM]);
+	for (i = 0; i < dev->virt_qp_nb; i++)
+		rte_free(dev->virtqueue[i * VIRTIO_QNUM]);
 
-	rte_free(ll_dev);
-}
-
-/*
- * Remove an entry from the device configuration linked list.
- */
-static struct virtio_net_config_ll *
-rm_config_ll_entry(struct virtio_net_config_ll *ll_dev,
-	struct virtio_net_config_ll *ll_dev_last)
-{
-	/* First remove the device and then clean it up. */
-	if (ll_dev == ll_root) {
-		ll_root = ll_dev->next;
-		cleanup_device(&ll_dev->dev, 1);
-		free_device(ll_dev);
-		return ll_root;
-	} else {
-		if (likely(ll_dev_last != NULL)) {
-			ll_dev_last->next = ll_dev->next;
-			cleanup_device(&ll_dev->dev, 1);
-			free_device(ll_dev);
-			return ll_dev_last->next;
-		} else {
-			cleanup_device(&ll_dev->dev, 1);
-			free_device(ll_dev);
-			RTE_LOG(ERR, VHOST_CONFIG,
-				"Remove entry from config_ll failed\n");
-			return NULL;
-		}
-	}
+	rte_free(dev);
 }
 
 static void
@@ -268,8 +167,8 @@ init_vring_queue(struct vhost_virtqueue *vq, int qp_idx)
 {
 	memset(vq, 0, sizeof(struct vhost_virtqueue));
 
-	vq->kickfd = -1;
-	vq->callfd = -1;
+	vq->kickfd = VIRTIO_UNINITIALIZED_EVENTFD;
+	vq->callfd = VIRTIO_UNINITIALIZED_EVENTFD;
 
 	/* Backends are set to -1 indicating an inactive device. */
 	vq->backend = -1;
@@ -355,63 +254,56 @@ reset_device(struct virtio_net *dev)
  * initialised and a new entry is added to the device configuration linked
  * list.
  */
-static int
-new_device(struct vhost_device_ctx ctx)
+int
+vhost_new_device(struct vhost_device_ctx ctx)
 {
-	struct virtio_net_config_ll *new_ll_dev;
+	struct virtio_net *dev;
+	int i;
 
-	/* Setup device and virtqueues. */
-	new_ll_dev = rte_zmalloc(NULL, sizeof(struct virtio_net_config_ll), 0);
-	if (new_ll_dev == NULL) {
+	dev = rte_zmalloc(NULL, sizeof(struct virtio_net), 0);
+	if (dev == NULL) {
 		RTE_LOG(ERR, VHOST_CONFIG,
 			"(%"PRIu64") Failed to allocate memory for dev.\n",
 			ctx.fh);
 		return -1;
 	}
 
-	new_ll_dev->next = NULL;
+	for (i = 0; i < MAX_VHOST_DEVICE; i++) {
+		if (vhost_devices[i] == NULL)
+			break;
+	}
+	if (i == MAX_VHOST_DEVICE) {
+		RTE_LOG(ERR, VHOST_CONFIG,
+			"Failed to find a free slot for new device.\n");
+		return -1;
+	}
 
-	/* Add entry to device configuration linked list. */
-	add_config_ll_entry(new_ll_dev);
+	vhost_devices[i] = dev;
+	dev->device_fh   = i;
 
-	return new_ll_dev->dev.device_fh;
+	return i;
 }
 
 /*
  * Function is called from the CUSE release function. This function will
  * cleanup the device and remove it from device configuration linked list.
  */
-static void
-destroy_device(struct vhost_device_ctx ctx)
-{
-	struct virtio_net_config_ll *ll_dev_cur_ctx, *ll_dev_last = NULL;
-	struct virtio_net_config_ll *ll_dev_cur = ll_root;
-
-	/* Find the linked list entry for the device to be removed. */
-	ll_dev_cur_ctx = get_config_ll_entry(ctx);
-	while (ll_dev_cur != NULL) {
-		/*
-		 * If the device is found or
-		 * a device that doesn't exist is found then it is removed.
-		 */
-		if (ll_dev_cur == ll_dev_cur_ctx) {
-			/*
-			 * If the device is running on a data core then call
-			 * the function to remove it from the data core.
-			 */
-			if ((ll_dev_cur->dev.flags & VIRTIO_DEV_RUNNING))
-				notify_ops->destroy_device(&(ll_dev_cur->dev));
-			ll_dev_cur = rm_config_ll_entry(ll_dev_cur,
-					ll_dev_last);
-		} else {
-			ll_dev_last = ll_dev_cur;
-			ll_dev_cur = ll_dev_cur->next;
-		}
-	}
+void
+vhost_destroy_device(struct vhost_device_ctx ctx)
+{
+	struct virtio_net *dev = get_device(ctx);
+
+	if (dev->flags & VIRTIO_DEV_RUNNING)
+		notify_ops->destroy_device(dev);
+
+	cleanup_device(dev, 1);
+	free_device(dev);
+
+	vhost_devices[ctx.fh] = NULL;
 }
 
-static void
-set_ifname(struct vhost_device_ctx ctx,
+void
+vhost_set_ifname(struct vhost_device_ctx ctx,
 	const char *if_name, unsigned int if_len)
 {
 	struct virtio_net *dev;
@@ -433,8 +325,8 @@ set_ifname(struct vhost_device_ctx ctx,
  * This function just returns success at the moment unless
  * the device hasn't been initialised.
  */
-static int
-set_owner(struct vhost_device_ctx ctx)
+int
+vhost_set_owner(struct vhost_device_ctx ctx)
 {
 	struct virtio_net *dev;
 
@@ -448,8 +340,8 @@ set_owner(struct vhost_device_ctx ctx)
 /*
  * Called from CUSE IOCTL: VHOST_RESET_OWNER
  */
-static int
-reset_owner(struct vhost_device_ctx ctx)
+int
+vhost_reset_owner(struct vhost_device_ctx ctx)
 {
 	struct virtio_net *dev;
 
@@ -469,8 +361,8 @@ reset_owner(struct vhost_device_ctx ctx)
  * Called from CUSE IOCTL: VHOST_GET_FEATURES
  * The features that we support are requested.
  */
-static int
-get_features(struct vhost_device_ctx ctx, uint64_t *pu)
+int
+vhost_get_features(struct vhost_device_ctx ctx, uint64_t *pu)
 {
 	struct virtio_net *dev;
 
@@ -487,8 +379,8 @@ get_features(struct vhost_device_ctx ctx, uint64_t *pu)
  * Called from CUSE IOCTL: VHOST_SET_FEATURES
  * We receive the negotiated features supported by us and the virtio device.
  */
-static int
-set_features(struct vhost_device_ctx ctx, uint64_t *pu)
+int
+vhost_set_features(struct vhost_device_ctx ctx, uint64_t *pu)
 {
 	struct virtio_net *dev;
 	uint16_t vhost_hlen;
@@ -527,8 +419,9 @@ set_features(struct vhost_device_ctx ctx, uint64_t *pu)
  * Called from CUSE IOCTL: VHOST_SET_VRING_NUM
  * The virtio device sends us the size of the descriptor ring.
  */
-static int
-set_vring_num(struct vhost_device_ctx ctx, struct vhost_vring_state *state)
+int
+vhost_set_vring_num(struct vhost_device_ctx ctx,
+	struct vhost_vring_state *state)
 {
 	struct virtio_net *dev;
 
@@ -551,70 +444,71 @@ static struct virtio_net*
 numa_realloc(struct virtio_net *dev, int index)
 {
 	int oldnode, newnode;
-	struct virtio_net_config_ll *old_ll_dev, *new_ll_dev = NULL;
-	struct vhost_virtqueue *old_vq, *new_vq = NULL;
+	struct virtio_net *old_dev;
+	struct vhost_virtqueue *old_vq, *vq;
 	int ret;
-	int realloc_dev = 0, realloc_vq = 0;
 
-	old_ll_dev = (struct virtio_net_config_ll *)dev;
-	old_vq = dev->virtqueue[index];
+	/*
+	 * vq is allocated on pairs, we should try to do realloc
+	 * on first queue of one queue pair only.
+	 */
+	if (index % VIRTIO_QNUM != 0)
+		return dev;
+
+	old_dev = dev;
+	vq = old_vq = dev->virtqueue[index];
+
+	ret = get_mempolicy(&newnode, NULL, 0, old_vq->desc,
+			    MPOL_F_NODE | MPOL_F_ADDR);
 
-	ret  = get_mempolicy(&newnode, NULL, 0, old_vq->desc,
-			MPOL_F_NODE | MPOL_F_ADDR);
-	ret = ret | get_mempolicy(&oldnode, NULL, 0, old_ll_dev,
-			MPOL_F_NODE | MPOL_F_ADDR);
+	/* check if we need to reallocate vq */
+	ret |= get_mempolicy(&oldnode, NULL, 0, old_vq,
+			     MPOL_F_NODE | MPOL_F_ADDR);
 	if (ret) {
 		RTE_LOG(ERR, VHOST_CONFIG,
-			"Unable to get vring desc or dev numa information.\n");
+			"Unable to get vq numa information.\n");
 		return dev;
 	}
-	if (oldnode != newnode)
-		realloc_dev = 1;
+	if (oldnode != newnode) {
+		RTE_LOG(INFO, VHOST_CONFIG,
+			"reallocate vq from %d to %d node\n", oldnode, newnode);
+		vq = rte_malloc_socket(NULL, sizeof(*vq) * VIRTIO_QNUM, 0,
+				       newnode);
+		if (!vq)
+			return dev;
+
+		memcpy(vq, old_vq, sizeof(*vq) * VIRTIO_QNUM);
+		rte_free(old_vq);
+	}
 
-	ret = get_mempolicy(&oldnode, NULL, 0, old_vq,
-			MPOL_F_NODE | MPOL_F_ADDR);
+	/* check if we need to reallocate dev */
+	ret = get_mempolicy(&oldnode, NULL, 0, old_dev,
+			    MPOL_F_NODE | MPOL_F_ADDR);
 	if (ret) {
 		RTE_LOG(ERR, VHOST_CONFIG,
-			"Unable to get vq numa information.\n");
-		return dev;
+			"Unable to get dev numa information.\n");
+		goto out;
 	}
-	if (oldnode != newnode)
-		realloc_vq = 1;
-
-	if (realloc_dev == 0 && realloc_vq == 0)
-		return dev;
-
-	if (realloc_dev)
-		new_ll_dev = rte_malloc_socket(NULL,
-			sizeof(struct virtio_net_config_ll), 0, newnode);
-	if (realloc_vq)
-		new_vq = rte_malloc_socket(NULL,
-			sizeof(struct vhost_virtqueue), 0, newnode);
-	if (!new_ll_dev && !new_vq)
-		return dev;
-
-	if (realloc_vq)
-		memcpy(new_vq, old_vq, sizeof(*new_vq));
-	if (realloc_dev)
-		memcpy(new_ll_dev, old_ll_dev, sizeof(*new_ll_dev));
-	(new_ll_dev ? new_ll_dev : old_ll_dev)->dev.virtqueue[index] =
-		new_vq ? new_vq : old_vq;
-	if (realloc_vq)
-		rte_free(old_vq);
-	if (realloc_dev) {
-		if (ll_root == old_ll_dev)
-			ll_root = new_ll_dev;
-		else {
-			struct virtio_net_config_ll *prev = ll_root;
-			while (prev->next != old_ll_dev)
-				prev = prev->next;
-			prev->next = new_ll_dev;
-			new_ll_dev->next = old_ll_dev->next;
+	if (oldnode != newnode) {
+		RTE_LOG(INFO, VHOST_CONFIG,
+			"reallocate dev from %d to %d node\n",
+			oldnode, newnode);
+		dev = rte_malloc_socket(NULL, sizeof(*dev), 0, newnode);
+		if (!dev) {
+			dev = old_dev;
+			goto out;
 		}
-		rte_free(old_ll_dev);
+
+		memcpy(dev, old_dev, sizeof(*dev));
+		rte_free(old_dev);
 	}
 
-	return realloc_dev ? &new_ll_dev->dev : dev;
+out:
+	dev->virtqueue[index] = vq;
+	dev->virtqueue[index + 1] = vq + 1;
+	vhost_devices[dev->device_fh] = dev;
+
+	return dev;
 }
 #else
 static struct virtio_net*
@@ -629,14 +523,14 @@ numa_realloc(struct virtio_net *dev, int index __rte_unused)
  * The virtio device sends us the desc, used and avail ring addresses.
  * This function then converts these to our address space.
  */
-static int
-set_vring_addr(struct vhost_device_ctx ctx, struct vhost_vring_addr *addr)
+int
+vhost_set_vring_addr(struct vhost_device_ctx ctx, struct vhost_vring_addr *addr)
 {
 	struct virtio_net *dev;
 	struct vhost_virtqueue *vq;
 
 	dev = get_device(ctx);
-	if (dev == NULL)
+	if ((dev == NULL) || (dev->mem == NULL))
 		return -1;
 
 	/* addr->index refers to the queue index. The txq 1, rxq is 0. */
@@ -673,12 +567,16 @@ set_vring_addr(struct vhost_device_ctx ctx, struct vhost_vring_addr *addr)
 		return -1;
 	}
 
+	vq->log_guest_addr = addr->log_guest_addr;
+
 	LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address desc: %p\n",
 			dev->device_fh, vq->desc);
 	LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address avail: %p\n",
 			dev->device_fh, vq->avail);
 	LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") mapped address used: %p\n",
 			dev->device_fh, vq->used);
+	LOG_DEBUG(VHOST_CONFIG, "(%"PRIu64") log_guest_addr: %"PRIx64"\n",
+			dev->device_fh, vq->log_guest_addr);
 
 	return 0;
 }
@@ -687,8 +585,9 @@ set_vring_addr(struct vhost_device_ctx ctx, struct vhost_vring_addr *addr)
  * Called from CUSE IOCTL: VHOST_SET_VRING_BASE
  * The virtio device sends us the available ring last used index.
  */
-static int
-set_vring_base(struct vhost_device_ctx ctx, struct vhost_vring_state *state)
+int
+vhost_set_vring_base(struct vhost_device_ctx ctx,
+	struct vhost_vring_state *state)
 {
 	struct virtio_net *dev;
 
@@ -707,8 +606,8 @@ set_vring_base(struct vhost_device_ctx ctx, struct vhost_vring_state *state)
  * Called from CUSE IOCTL: VHOST_GET_VRING_BASE
  * We send the virtio device our available ring last used index.
  */
-static int
-get_vring_base(struct vhost_device_ctx ctx, uint32_t index,
+int
+vhost_get_vring_base(struct vhost_device_ctx ctx, uint32_t index,
 	struct vhost_vring_state *state)
 {
 	struct virtio_net *dev;
@@ -730,8 +629,8 @@ get_vring_base(struct vhost_device_ctx ctx, uint32_t index,
  * The virtio device sends an eventfd to interrupt the guest. This fd gets
  * copied into our process space.
  */
-static int
-set_vring_call(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
+int
+vhost_set_vring_call(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
 {
 	struct virtio_net *dev;
 	struct vhost_virtqueue *vq;
@@ -767,8 +666,8 @@ set_vring_call(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
  * The virtio device sends an eventfd that it can use to notify us.
  * This fd gets copied into our process space.
  */
-static int
-set_vring_kick(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
+int
+vhost_set_vring_kick(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
 {
 	struct virtio_net *dev;
 	struct vhost_virtqueue *vq;
@@ -797,8 +696,8 @@ set_vring_kick(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
  * At that point we remove the device from the data core.
  * The device will still exist in the device configuration linked list.
  */
-static int
-set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
+int
+vhost_set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
 {
 	struct virtio_net *dev;
 
@@ -825,42 +724,6 @@ set_backend(struct vhost_device_ctx ctx, struct vhost_vring_file *file)
 	return 0;
 }
 
-/*
- * Function pointers are set for the device operations to allow CUSE to call
- * functions when an IOCTL, device_add or device_release is received.
- */
-static const struct vhost_net_device_ops vhost_device_ops = {
-	.new_device = new_device,
-	.destroy_device = destroy_device,
-
-	.set_ifname = set_ifname,
-
-	.get_features = get_features,
-	.set_features = set_features,
-
-	.set_vring_num = set_vring_num,
-	.set_vring_addr = set_vring_addr,
-	.set_vring_base = set_vring_base,
-	.get_vring_base = get_vring_base,
-
-	.set_vring_kick = set_vring_kick,
-	.set_vring_call = set_vring_call,
-
-	.set_backend = set_backend,
-
-	.set_owner = set_owner,
-	.reset_owner = reset_owner,
-};
-
-/*
- * Called by main to setup callbacks when registering CUSE device.
- */
-struct vhost_net_device_ops const *
-get_virtio_net_callbacks(void)
-{
-	return &vhost_device_ops;
-}
-
 int rte_vhost_enable_guest_notification(struct virtio_net *dev,
 	uint16_t queue_id, int enable)
 {