virtio,vhost,vdpa: bugfixes

A bunch of fixes all over the place
 
 Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
 -----BEGIN PGP SIGNATURE-----
 
 iQFDBAABCAAtFiEEXQn9CHHI+FuUyooNKB8NuNKNVGkFAmSDUGkPHG1zdEByZWRo
 YXQuY29tAAoJECgfDbjSjVRpw6gH+wbopniPKDrxNwTpDFx3jix3QDTzVMY4Bq4k
 QdwPfjAZ1aDZXYHV1CdFXeKTA+ZkWHIREZSr+E/2/jeI55Exc2AeFptZrUesSg29
 jMN1MPs00CCy8Qi9BiCZIQkFkIKHNA2PY8wIA0oIXhIaG7pBtYQ14CnAFqn41ev5
 II20h389KMthe0lwm4ni/qHVZzG/2qP/JXLKf35proDEnU5WWM1rQZ1666EFMaIR
 6QExqwbPubxfv44Kl3mMkanGj6MmtLtFa2XlMLbEfLrU5/Xz+CywqSFHTUerrh3I
 eTNyqz4Oyj6UpRq264rqQBJmpSn8LWFBZXQlJ6Y+ef/h8Mhdewk=
 =G8CT
 -----END PGP SIGNATURE-----

Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost

Pull virtio bug fixes from Michael Tsirkin:
 "A bunch of fixes all over the place"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  tools/virtio: use canonical ftrace path
  vhost_vdpa: support PACKED when setting-getting vring_base
  vhost: support PACKED when setting-getting vring_base
  vhost: Fix worker hangs due to missed wake up calls
  vhost: Fix crash during early vhost_transport_send_pkt calls
  vhost_net: revert upend_idx only on retriable error
  vhost_vdpa: tell vqs about the negotiated
  vdpa/mlx5: Fix hang when cvq commands are triggered during device unregister
  tools/virtio: Add .gitignore for ringtest
  tools/virtio: Fix arm64 ringtest compilation error
  vduse: avoid empty string for dev name
  vhost: use kzalloc() instead of kmalloc() followed by memset()

drivers/vdpa/mlx5/net/mlx5_vnet.c

@@ -3349,10 +3349,10 @@ static void mlx5_vdpa_dev_del(struct vdpa_mgmt_dev *v_mdev, struct vdpa_device *
 	mlx5_vdpa_remove_debugfs(ndev->debugfs);
 	ndev->debugfs = NULL;
 	unregister_link_notifier(ndev);
+	_vdpa_unregister_device(dev);
 	wq = mvdev->wq;
 	mvdev->wq = NULL;
 	destroy_workqueue(wq);
-	_vdpa_unregister_device(dev);
 	mgtdev->ndev = NULL;
 }
 

drivers/vdpa/vdpa_user/vduse_dev.c

@@ -1685,6 +1685,9 @@ static bool vduse_validate_config(struct vduse_dev_config *config)
 	if (config->vq_num > 0xffff)
 		return false;
 
+	if (!config->name[0])
+		return false;
+
 	if (!device_is_allowed(config->device_id))
 		return false;
 

drivers/vhost/net.c

@@ -935,13 +935,18 @@ static void handle_tx_zerocopy(struct vhost_net *net, struct socket *sock)
 
 		err = sock->ops->sendmsg(sock, &msg, len);
 		if (unlikely(err < 0)) {
+			bool retry = err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS;
+
 			if (zcopy_used) {
 				if (vq->heads[ubuf->desc].len == VHOST_DMA_IN_PROGRESS)
 					vhost_net_ubuf_put(ubufs);
-				nvq->upend_idx = ((unsigned)nvq->upend_idx - 1)
+				if (retry)
-					% UIO_MAXIOV;
+					nvq->upend_idx = ((unsigned)nvq->upend_idx - 1)
+						% UIO_MAXIOV;
+				else
+					vq->heads[ubuf->desc].len = VHOST_DMA_DONE_LEN;
 			}
-			if (err == -EAGAIN || err == -ENOMEM || err == -ENOBUFS) {
+			if (retry) {
 				vhost_discard_vq_desc(vq, 1);
 				vhost_net_enable_vq(net, vq);
 				break;

drivers/vhost/vdpa.c

@@ -407,7 +407,10 @@ static long vhost_vdpa_set_features(struct vhost_vdpa *v, u64 __user *featurep)
 {
 	struct vdpa_device *vdpa = v->vdpa;
 	const struct vdpa_config_ops *ops = vdpa->config;
+	struct vhost_dev *d = &v->vdev;
+	u64 actual_features;
 	u64 features;
+	int i;
 
 	/*
 	 * It's not allowed to change the features after they have
@@ -422,6 +425,16 @@ static long vhost_vdpa_set_features(struct vhost_vdpa *v, u64 __user *featurep)
 	if (vdpa_set_features(vdpa, features))
 		return -EINVAL;
 
+	/* let the vqs know what has been configured */
+	actual_features = ops->get_driver_features(vdpa);
+	for (i = 0; i < d->nvqs; ++i) {
+		struct vhost_virtqueue *vq = d->vqs[i];
+
+		mutex_lock(&vq->mutex);
+		vq->acked_features = actual_features;
+		mutex_unlock(&vq->mutex);
+	}
+
 	return 0;
 }
 
@@ -594,7 +607,14 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 		if (r)
 			return r;
 
-		vq->last_avail_idx = vq_state.split.avail_index;
+		if (vhost_has_feature(vq, VIRTIO_F_RING_PACKED)) {
+			vq->last_avail_idx = vq_state.packed.last_avail_idx |
+					     (vq_state.packed.last_avail_counter << 15);
+			vq->last_used_idx = vq_state.packed.last_used_idx |
+					    (vq_state.packed.last_used_counter << 15);
+		} else {
+			vq->last_avail_idx = vq_state.split.avail_index;
+		}
 		break;
 	}
 
@@ -612,9 +632,15 @@ static long vhost_vdpa_vring_ioctl(struct vhost_vdpa *v, unsigned int cmd,
 		break;
 
 	case VHOST_SET_VRING_BASE:
-		vq_state.split.avail_index = vq->last_avail_idx;
+		if (vhost_has_feature(vq, VIRTIO_F_RING_PACKED)) {
-		if (ops->set_vq_state(vdpa, idx, &vq_state))
+			vq_state.packed.last_avail_idx = vq->last_avail_idx & 0x7fff;
-			r = -EINVAL;
+			vq_state.packed.last_avail_counter = !!(vq->last_avail_idx & 0x8000);
+			vq_state.packed.last_used_idx = vq->last_used_idx & 0x7fff;
+			vq_state.packed.last_used_counter = !!(vq->last_used_idx & 0x8000);
+		} else {
+			vq_state.split.avail_index = vq->last_avail_idx;
+		}
+		r = ops->set_vq_state(vdpa, idx, &vq_state);
 		break;
 
 	case VHOST_SET_VRING_CALL:

drivers/vhost/vhost.c

@@ -235,7 +235,7 @@ void vhost_dev_flush(struct vhost_dev *dev)
 {
 	struct vhost_flush_struct flush;
 
-	if (dev->worker) {
+	if (dev->worker.vtsk) {
 		init_completion(&flush.wait_event);
 		vhost_work_init(&flush.work, vhost_flush_work);
 
@@ -247,7 +247,7 @@ EXPORT_SYMBOL_GPL(vhost_dev_flush);
 
 void vhost_work_queue(struct vhost_dev *dev, struct vhost_work *work)
 {
-	if (!dev->worker)
+	if (!dev->worker.vtsk)
 		return;
 
 	if (!test_and_set_bit(VHOST_WORK_QUEUED, &work->flags)) {
@@ -255,8 +255,8 @@ void vhost_work_queue(struct vhost_dev *dev, struct vhost_work *work)
 		 * sure it was not in the list.
 		 * test_and_set_bit() implies a memory barrier.
 		 */
-		llist_add(&work->node, &dev->worker->work_list);
+		llist_add(&work->node, &dev->worker.work_list);
-		vhost_task_wake(dev->worker->vtsk);
+		vhost_task_wake(dev->worker.vtsk);
 	}
 }
 EXPORT_SYMBOL_GPL(vhost_work_queue);
@@ -264,7 +264,7 @@ EXPORT_SYMBOL_GPL(vhost_work_queue);
 /* A lockless hint for busy polling code to exit the loop */
 bool vhost_has_work(struct vhost_dev *dev)
 {
-	return dev->worker && !llist_empty(&dev->worker->work_list);
+	return !llist_empty(&dev->worker.work_list);
 }
 EXPORT_SYMBOL_GPL(vhost_has_work);
 
@@ -341,6 +341,8 @@ static bool vhost_worker(void *data)
 
 	node = llist_del_all(&worker->work_list);
 	if (node) {
+		__set_current_state(TASK_RUNNING);
+
 		node = llist_reverse_order(node);
 		/* make sure flag is seen after deletion */
 		smp_wmb();
@@ -456,7 +458,8 @@ void vhost_dev_init(struct vhost_dev *dev,
 	dev->umem = NULL;
 	dev->iotlb = NULL;
 	dev->mm = NULL;
-	dev->worker = NULL;
+	memset(&dev->worker, 0, sizeof(dev->worker));
+	init_llist_head(&dev->worker.work_list);
 	dev->iov_limit = iov_limit;
 	dev->weight = weight;
 	dev->byte_weight = byte_weight;
@@ -530,47 +533,30 @@ static void vhost_detach_mm(struct vhost_dev *dev)
 
 static void vhost_worker_free(struct vhost_dev *dev)
 {
-	struct vhost_worker *worker = dev->worker;
+	if (!dev->worker.vtsk)
-
-	if (!worker)
 		return;
 
-	dev->worker = NULL;
+	WARN_ON(!llist_empty(&dev->worker.work_list));
-	WARN_ON(!llist_empty(&worker->work_list));
+	vhost_task_stop(dev->worker.vtsk);
-	vhost_task_stop(worker->vtsk);
+	dev->worker.kcov_handle = 0;
-	kfree(worker);
+	dev->worker.vtsk = NULL;
 }
 
 static int vhost_worker_create(struct vhost_dev *dev)
 {
-	struct vhost_worker *worker;
 	struct vhost_task *vtsk;
 	char name[TASK_COMM_LEN];
-	int ret;
 
-	worker = kzalloc(sizeof(*worker), GFP_KERNEL_ACCOUNT);
-	if (!worker)
-		return -ENOMEM;
-
-	dev->worker = worker;
-	worker->kcov_handle = kcov_common_handle();
-	init_llist_head(&worker->work_list);
 	snprintf(name, sizeof(name), "vhost-%d", current->pid);
 
-	vtsk = vhost_task_create(vhost_worker, worker, name);
+	vtsk = vhost_task_create(vhost_worker, &dev->worker, name);
-	if (!vtsk) {
+	if (!vtsk)
-		ret = -ENOMEM;
+		return -ENOMEM;
-		goto free_worker;
-	}
 
-	worker->vtsk = vtsk;
+	dev->worker.kcov_handle = kcov_common_handle();
+	dev->worker.vtsk = vtsk;
 	vhost_task_start(vtsk);
 	return 0;
-
-free_worker:
-	kfree(worker);
-	dev->worker = NULL;
-	return ret;
 }
 
 /* Caller should have device mutex */
@@ -1614,17 +1600,25 @@ long vhost_vring_ioctl(struct vhost_dev *d, unsigned int ioctl, void __user *arg
 			r = -EFAULT;
 			break;
 		}
-		if (s.num > 0xffff) {
+		if (vhost_has_feature(vq, VIRTIO_F_RING_PACKED)) {
-			r = -EINVAL;
+			vq->last_avail_idx = s.num & 0xffff;
-			break;
+			vq->last_used_idx = (s.num >> 16) & 0xffff;
+		} else {
+			if (s.num > 0xffff) {
+				r = -EINVAL;
+				break;
+			}
+			vq->last_avail_idx = s.num;
 		}
-		vq->last_avail_idx = s.num;
 		/* Forget the cached index value. */
 		vq->avail_idx = vq->last_avail_idx;
 		break;
 	case VHOST_GET_VRING_BASE:
 		s.index = idx;
-		s.num = vq->last_avail_idx;
+		if (vhost_has_feature(vq, VIRTIO_F_RING_PACKED))
+			s.num = (u32)vq->last_avail_idx | ((u32)vq->last_used_idx << 16);
+		else
+			s.num = vq->last_avail_idx;
 		if (copy_to_user(argp, &s, sizeof s))
 			r = -EFAULT;
 		break;
@@ -2563,12 +2557,11 @@ EXPORT_SYMBOL_GPL(vhost_disable_notify);
 /* Create a new message. */
 struct vhost_msg_node *vhost_new_msg(struct vhost_virtqueue *vq, int type)
 {
-	struct vhost_msg_node *node = kmalloc(sizeof *node, GFP_KERNEL);
+	/* Make sure all padding within the structure is initialized. */
+	struct vhost_msg_node *node = kzalloc(sizeof(*node), GFP_KERNEL);
 	if (!node)
 		return NULL;
 
-	/* Make sure all padding within the structure is initialized. */
-	memset(&node->msg, 0, sizeof node->msg);
 	node->vq = vq;
 	node->msg.type = type;
 	return node;

drivers/vhost/vhost.h

@@ -92,13 +92,17 @@ struct vhost_virtqueue {
 	/* The routine to call when the Guest pings us, or timeout. */
 	vhost_work_fn_t handle_kick;
 
-	/* Last available index we saw. */
+	/* Last available index we saw.
+	 * Values are limited to 0x7fff, and the high bit is used as
+	 * a wrap counter when using VIRTIO_F_RING_PACKED. */
 	u16 last_avail_idx;
 
 	/* Caches available index value from user. */
 	u16 avail_idx;
 
-	/* Last index we used. */
+	/* Last index we used.
+	 * Values are limited to 0x7fff, and the high bit is used as
+	 * a wrap counter when using VIRTIO_F_RING_PACKED. */
 	u16 last_used_idx;
 
 	/* Used flags */
@@ -154,7 +158,7 @@ struct vhost_dev {
 	struct vhost_virtqueue **vqs;
 	int nvqs;
 	struct eventfd_ctx *log_ctx;
-	struct vhost_worker *worker;
+	struct vhost_worker worker;
 	struct vhost_iotlb *umem;
 	struct vhost_iotlb *iotlb;
 	spinlock_t iotlb_lock;

kernel/vhost_task.c

@@ -28,10 +28,6 @@ static int vhost_task_fn(void *data)
 	for (;;) {
 		bool did_work;
 
-		/* mb paired w/ vhost_task_stop */
-		if (test_bit(VHOST_TASK_FLAGS_STOP, &vtsk->flags))
-			break;
-
 		if (!dead && signal_pending(current)) {
 			struct ksignal ksig;
 			/*
@@ -48,11 +44,17 @@ static int vhost_task_fn(void *data)
 				clear_thread_flag(TIF_SIGPENDING);
 		}
 
-		did_work = vtsk->fn(vtsk->data);
+		/* mb paired w/ vhost_task_stop */
-		if (!did_work) {
+		set_current_state(TASK_INTERRUPTIBLE);
-			set_current_state(TASK_INTERRUPTIBLE);
+
-			schedule();
+		if (test_bit(VHOST_TASK_FLAGS_STOP, &vtsk->flags)) {
+			__set_current_state(TASK_RUNNING);
+			break;
 		}
+
+		did_work = vtsk->fn(vtsk->data);
+		if (!did_work)
+			schedule();
 	}
 
 	complete(&vtsk->exited);

tools/virtio/ringtest/.gitignore

@@ -0,0 +1,7 @@
+# SPDX-License-Identifier: GPL-2.0-only
+/noring
+/ptr_ring
+/ring
+/virtio_ring_0_9
+/virtio_ring_inorder
+/virtio_ring_poll

tools/virtio/ringtest/main.h

@@ -8,6 +8,7 @@
 #ifndef MAIN_H
 #define MAIN_H
 
+#include <assert.h>
 #include <stdbool.h>
 
 extern int param;
@@ -95,6 +96,8 @@ extern unsigned ring_size;
 #define cpu_relax() asm ("rep; nop" ::: "memory")
 #elif defined(__s390x__)
 #define cpu_relax() barrier()
+#elif defined(__aarch64__)
+#define cpu_relax() asm ("yield" ::: "memory")
 #else
 #define cpu_relax() assert(0)
 #endif
@@ -112,6 +115,8 @@ static inline void busy_wait(void)
 
 #if defined(__x86_64__) || defined(__i386__)
 #define smp_mb()     asm volatile("lock; addl $0,-132(%%rsp)" ::: "memory", "cc")
+#elif defined(__aarch64__)
+#define smp_mb()     asm volatile("dmb ish" ::: "memory")
 #else
 /*
  * Not using __ATOMIC_SEQ_CST since gcc docs say they are only synchronized
@@ -136,10 +141,16 @@ static inline void busy_wait(void)
 
 #if defined(__i386__) || defined(__x86_64__) || defined(__s390x__)
 #define smp_wmb() barrier()
+#elif defined(__aarch64__)
+#define smp_wmb() asm volatile("dmb ishst" ::: "memory")
 #else
 #define smp_wmb() smp_release()
 #endif
 
+#ifndef __always_inline
+#define __always_inline inline __attribute__((always_inline))
+#endif
+
 static __always_inline
 void __read_once_size(const volatile void *p, void *res, int size)
 {

tools/virtio/virtio-trace/README

@@ -95,7 +95,7 @@ Run
 
 1) Enable ftrace in the guest
  <Example>
-	# echo 1 > /sys/kernel/debug/tracing/events/sched/enable
+	# echo 1 > /sys/kernel/tracing/events/sched/enable
 
 2) Run trace agent in the guest
  This agent must be operated as root.

tools/virtio/virtio-trace/trace-agent.c

@@ -18,8 +18,9 @@
 #define PIPE_DEF_BUFS		16
 #define PIPE_MIN_SIZE		(PAGE_SIZE*PIPE_DEF_BUFS)
 #define PIPE_MAX_SIZE		(1024*1024)
-#define READ_PATH_FMT	\
+#define TRACEFS 		"/sys/kernel/tracing"
-		"/sys/kernel/debug/tracing/per_cpu/cpu%d/trace_pipe_raw"
+#define DEBUGFS 		"/sys/kernel/debug/tracing"
+#define READ_PATH_FMT		"%s/per_cpu/cpu%d/trace_pipe_raw"
 #define WRITE_PATH_FMT		"/dev/virtio-ports/trace-path-cpu%d"
 #define CTL_PATH		"/dev/virtio-ports/agent-ctl-path"
 
@@ -120,9 +121,12 @@ static const char *make_path(int cpu_num, bool this_is_write_path)
 	if (this_is_write_path)
 		/* write(output) path */
 		ret = snprintf(buf, PATH_MAX, WRITE_PATH_FMT, cpu_num);
-	else
+	else {
 		/* read(input) path */
-		ret = snprintf(buf, PATH_MAX, READ_PATH_FMT, cpu_num);
+		ret = snprintf(buf, PATH_MAX, READ_PATH_FMT, TRACEFS, cpu_num);
+		if (ret > 0 && access(buf, F_OK) != 0)
+			ret = snprintf(buf, PATH_MAX, READ_PATH_FMT, DEBUGFS, cpu_num);
+	}
 
 	if (ret <= 0) {
 		pr_err("Failed to generate %s path(CPU#%d):%d\n",