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

Pull virtio fixes from Michael Tsirkin:
 "Fixes all over the place.

  This includes a couple of tests that I would normally defer, but since
  they have already been helpful in catching some bugs, don't build for
  any users at all, and having them upstream makes life easier for
  everyone, I think it's ok even at this late stage"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  tools/virtio: Use tools/include/list.h instead of stubs
  tools/virtio: Reset index in virtio_test --reset.
  tools/virtio: Extract virtqueue initialization in vq_reset
  tools/virtio: Use __vring_new_virtqueue in virtio_test.c
  tools/virtio: Add --reset
  tools/virtio: Add --batch=random option
  tools/virtio: Add --batch option
  virtio-mem: add memory via add_memory_driver_managed()
  virtio-mem: silence a static checker warning
  vhost_vdpa: Fix potential underflow in vhost_vdpa_mmap()
  vdpa: fix typos in the comments for __vdpa_alloc_device()

drivers/vdpa/vdpa.c

@@ -63,7 +63,7 @@ static void vdpa_release_dev(struct device *d)
  * @config: the bus operations that is supported by this device
  * @size: size of the parent structure that contains private data
  *
- * Drvier should use vdap_alloc_device() wrapper macro instead of
+ * Driver should use vdpa_alloc_device() wrapper macro instead of
  * using this directly.
  *
  * Returns an error when parent/config/dma_dev is not set or fail to get

drivers/vhost/test.c

@@ -263,9 +263,62 @@ static int vhost_test_set_features(struct vhost_test *n, u64 features)
 	return 0;
 }
 
+static long vhost_test_set_backend(struct vhost_test *n, unsigned index, int fd)
+{
+	static void *backend;
+
+	const bool enable = fd != -1;
+	struct vhost_virtqueue *vq;
+	int r;
+
+	mutex_lock(&n->dev.mutex);
+	r = vhost_dev_check_owner(&n->dev);
+	if (r)
+		goto err;
+
+	if (index >= VHOST_TEST_VQ_MAX) {
+		r = -ENOBUFS;
+		goto err;
+	}
+	vq = &n->vqs[index];
+	mutex_lock(&vq->mutex);
+
+	/* Verify that ring has been setup correctly. */
+	if (!vhost_vq_access_ok(vq)) {
+		r = -EFAULT;
+		goto err_vq;
+	}
+	if (!enable) {
+		vhost_poll_stop(&vq->poll);
+		backend = vhost_vq_get_backend(vq);
+		vhost_vq_set_backend(vq, NULL);
+	} else {
+		vhost_vq_set_backend(vq, backend);
+		r = vhost_vq_init_access(vq);
+		if (r == 0)
+			r = vhost_poll_start(&vq->poll, vq->kick);
+	}
+
+	mutex_unlock(&vq->mutex);
+
+	if (enable) {
+		vhost_test_flush_vq(n, index);
+	}
+
+	mutex_unlock(&n->dev.mutex);
+	return 0;
+
+err_vq:
+	mutex_unlock(&vq->mutex);
+err:
+	mutex_unlock(&n->dev.mutex);
+	return r;
+}
+
 static long vhost_test_ioctl(struct file *f, unsigned int ioctl,
 			     unsigned long arg)
 {
+	struct vhost_vring_file backend;
 	struct vhost_test *n = f->private_data;
 	void __user *argp = (void __user *)arg;
 	u64 __user *featurep = argp;
@@ -277,6 +330,10 @@ static long vhost_test_ioctl(struct file *f, unsigned int ioctl,
 		if (copy_from_user(&test, argp, sizeof test))
 			return -EFAULT;
 		return vhost_test_run(n, test);
+	case VHOST_TEST_SET_BACKEND:
+		if (copy_from_user(&backend, argp, sizeof backend))
+			return -EFAULT;
+		return vhost_test_set_backend(n, backend.index, backend.fd);
 	case VHOST_GET_FEATURES:
 		features = VHOST_FEATURES;
 		if (copy_to_user(featurep, &features, sizeof features))

drivers/vhost/test.h

@@ -4,5 +4,6 @@
 
 /* Start a given test on the virtio null device. 0 stops all tests. */
 #define VHOST_TEST_RUN _IOW(VHOST_VIRTIO, 0x31, int)
+#define VHOST_TEST_SET_BACKEND _IOW(VHOST_VIRTIO, 0x32, int)
 
 #endif

drivers/vhost/vdpa.c

@@ -818,7 +818,7 @@ static int vhost_vdpa_mmap(struct file *file, struct vm_area_struct *vma)
 	struct vdpa_device *vdpa = v->vdpa;
 	const struct vdpa_config_ops *ops = vdpa->config;
 	struct vdpa_notification_area notify;
-	int index = vma->vm_pgoff;
+	unsigned long index = vma->vm_pgoff;
 
 	if (vma->vm_end - vma->vm_start != PAGE_SIZE)
 		return -EINVAL;

drivers/virtio/virtio_mem.c

@@ -101,6 +101,11 @@ struct virtio_mem {
 
 	/* The parent resource for all memory added via this device. */
 	struct resource *parent_resource;
+	/*
+	 * Copy of "System RAM (virtio_mem)" to be used for
+	 * add_memory_driver_managed().
+	 */
+	const char *resource_name;
 
 	/* Summary of all memory block states. */
 	unsigned long nb_mb_state[VIRTIO_MEM_MB_STATE_COUNT];
@@ -414,8 +419,20 @@ static int virtio_mem_mb_add(struct virtio_mem *vm, unsigned long mb_id)
 	if (nid == NUMA_NO_NODE)
 		nid = memory_add_physaddr_to_nid(addr);
 
+	/*
+	 * When force-unloading the driver and we still have memory added to
+	 * Linux, the resource name has to stay.
+	 */
+	if (!vm->resource_name) {
+		vm->resource_name = kstrdup_const("System RAM (virtio_mem)",
+						  GFP_KERNEL);
+		if (!vm->resource_name)
+			return -ENOMEM;
+	}
+
 	dev_dbg(&vm->vdev->dev, "adding memory block: %lu\n", mb_id);
-	return add_memory(nid, addr, memory_block_size_bytes());
+	return add_memory_driver_managed(nid, addr, memory_block_size_bytes(),
+					 vm->resource_name);
 }
 
 /*
@@ -1192,7 +1209,7 @@ static int virtio_mem_mb_plug_any_sb(struct virtio_mem *vm, unsigned long mb_id,
 						VIRTIO_MEM_MB_STATE_OFFLINE);
 	}
 
-	return rc;
+	return 0;
 }
 
 /*
@@ -1890,10 +1907,12 @@ static void virtio_mem_remove(struct virtio_device *vdev)
 	    vm->nb_mb_state[VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL] ||
 	    vm->nb_mb_state[VIRTIO_MEM_MB_STATE_ONLINE] ||
 	    vm->nb_mb_state[VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL] ||
-	    vm->nb_mb_state[VIRTIO_MEM_MB_STATE_ONLINE_MOVABLE])
+	    vm->nb_mb_state[VIRTIO_MEM_MB_STATE_ONLINE_MOVABLE]) {
 		dev_warn(&vdev->dev, "device still has system memory added\n");
-	else
+	} else {
 		virtio_mem_delete_resource(vm);
+		kfree_const(vm->resource_name);
+	}
 
 	/* remove all tracking data - no locking needed */
 	vfree(vm->mb_state);

tools/virtio/linux/kernel.h

@@ -11,6 +11,7 @@
 
 #include <linux/compiler.h>
 #include <linux/types.h>
+#include <linux/list.h>
 #include <linux/printk.h>
 #include <linux/bug.h>
 #include <errno.h>
@@ -135,10 +136,4 @@ static inline void free_page(unsigned long addr)
 	(void) (&_min1 == &_min2);		\
 	_min1 < _min2 ? _min1 : _min2; })
 
-/* TODO: empty stubs for now. Broken but enough for virtio_ring.c */
-#define list_add_tail(a, b) do {} while (0)
-#define list_del(a) do {} while (0)
-#define list_for_each_entry(a, b, c) while (0)
-/* end of stubs */
-
 #endif /* KERNEL_H */

tools/virtio/linux/virtio.h

@@ -11,12 +11,11 @@ struct device {
 struct virtio_device {
 	struct device dev;
 	u64 features;
+	struct list_head vqs;
 };
 
 struct virtqueue {
-	/* TODO: commented as list macros are empty stubs for now.
-	 * Broken but enough for virtio_ring.c
-	 * struct list_head list; */
+	struct list_head list;
 	void (*callback)(struct virtqueue *vq);
 	const char *name;
 	struct virtio_device *vdev;

tools/virtio/virtio_test.c

 // SPDX-License-Identifier: GPL-2.0
 #define _GNU_SOURCE
 #include <getopt.h>
+#include <limits.h>
 #include <string.h>
 #include <poll.h>
 #include <sys/eventfd.h>
@@ -18,6 +19,8 @@
 #include <linux/virtio_ring.h>
 #include "../../drivers/vhost/test.h"
 
+#define RANDOM_BATCH -1
+
 /* Unused */
 void *__kmalloc_fake, *__kfree_ignore_start, *__kfree_ignore_end;
 
@@ -43,6 +46,10 @@ struct vdev_info {
 	struct vhost_memory *mem;
 };
 
+static const struct vhost_vring_file no_backend = { .fd = -1 },
+				     backend = { .fd = 1 };
+static const struct vhost_vring_state null_state = {};
+
 bool vq_notify(struct virtqueue *vq)
 {
 	struct vq_info *info = vq->priv;
@@ -88,6 +95,19 @@ void vhost_vq_setup(struct vdev_info *dev, struct vq_info *info)
 	assert(r >= 0);
 }
 
+static void vq_reset(struct vq_info *info, int num, struct virtio_device *vdev)
+{
+	if (info->vq)
+		vring_del_virtqueue(info->vq);
+
+	memset(info->ring, 0, vring_size(num, 4096));
+	vring_init(&info->vring, num, info->ring, 4096);
+	info->vq = __vring_new_virtqueue(info->idx, info->vring, vdev, true,
+					 false, vq_notify, vq_callback, "test");
+	assert(info->vq);
+	info->vq->priv = info;
+}
+
 static void vq_info_add(struct vdev_info *dev, int num)
 {
 	struct vq_info *info = &dev->vqs[dev->nvqs];
@@ -97,14 +117,7 @@ static void vq_info_add(struct vdev_info *dev, int num)
 	info->call = eventfd(0, EFD_NONBLOCK);
 	r = posix_memalign(&info->ring, 4096, vring_size(num, 4096));
 	assert(r >= 0);
-	memset(info->ring, 0, vring_size(num, 4096));
-	vring_init(&info->vring, num, info->ring, 4096);
-	info->vq = vring_new_virtqueue(info->idx,
-				       info->vring.num, 4096, &dev->vdev,
-				       true, false, info->ring,
-				       vq_notify, vq_callback, "test");
-	assert(info->vq);
-	info->vq->priv = info;
+	vq_reset(info, num, &dev->vdev);
 	vhost_vq_setup(dev, info);
 	dev->fds[info->idx].fd = info->call;
 	dev->fds[info->idx].events = POLLIN;
@@ -116,6 +129,7 @@ static void vdev_info_init(struct vdev_info* dev, unsigned long long features)
 	int r;
 	memset(dev, 0, sizeof *dev);
 	dev->vdev.features = features;
+	INIT_LIST_HEAD(&dev->vdev.vqs);
 	dev->buf_size = 1024;
 	dev->buf = malloc(dev->buf_size);
 	assert(dev->buf);
@@ -152,41 +166,93 @@ static void wait_for_interrupt(struct vdev_info *dev)
 }
 
 static void run_test(struct vdev_info *dev, struct vq_info *vq,
-		     bool delayed, int bufs)
+		     bool delayed, int batch, int reset_n, int bufs)
 {
 	struct scatterlist sl;
-	long started = 0, completed = 0;
-	long completed_before;
+	long started = 0, completed = 0, next_reset = reset_n;
+	long completed_before, started_before;
 	int r, test = 1;
 	unsigned len;
 	long long spurious = 0;
+	const bool random_batch = batch == RANDOM_BATCH;
+
 	r = ioctl(dev->control, VHOST_TEST_RUN, &test);
 	assert(r >= 0);
+	if (!reset_n) {
+		next_reset = INT_MAX;
+	}
+
 	for (;;) {
 		virtqueue_disable_cb(vq->vq);
 		completed_before = completed;
+		started_before = started;
 		do {
-			if (started < bufs) {
+			const bool reset = completed > next_reset;
+			if (random_batch)
+				batch = (random() % vq->vring.num) + 1;
+
+			while (started < bufs &&
+			       (started - completed) < batch) {
 				sg_init_one(&sl, dev->buf, dev->buf_size);
 				r = virtqueue_add_outbuf(vq->vq, &sl, 1,
 							 dev->buf + started,
 							 GFP_ATOMIC);
-				if (likely(r == 0)) {
+				if (unlikely(r != 0)) {
+					if (r == -ENOSPC &&
+					    started > started_before)
+						r = 0;
+					else
+						r = -1;
+					break;
+				}
+
 				++started;
-					if (unlikely(!virtqueue_kick(vq->vq)))
+
+				if (unlikely(!virtqueue_kick(vq->vq))) {
 					r = -1;
+					break;
+				}
 			}
-			} else
+
+			if (started >= bufs)
 				r = -1;
 
+			if (reset) {
+				r = ioctl(dev->control, VHOST_TEST_SET_BACKEND,
+					  &no_backend);
+				assert(!r);
+			}
+
 			/* Flush out completed bufs if any */
-			if (virtqueue_get_buf(vq->vq, &len)) {
+			while (virtqueue_get_buf(vq->vq, &len)) {
 				++completed;
 				r = 0;
 			}
 
+			if (reset) {
+				struct vhost_vring_state s = { .index = 0 };
+
+				vq_reset(vq, vq->vring.num, &dev->vdev);
+
+				r = ioctl(dev->control, VHOST_GET_VRING_BASE,
+					  &s);
+				assert(!r);
+
+				s.num = 0;
+				r = ioctl(dev->control, VHOST_SET_VRING_BASE,
+					  &null_state);
+				assert(!r);
+
+				r = ioctl(dev->control, VHOST_TEST_SET_BACKEND,
+					  &backend);
+				assert(!r);
+
+				started = completed;
+				while (completed > next_reset)
+					next_reset += completed;
+			}
 		} while (r == 0);
-		if (completed == completed_before)
+		if (completed == completed_before && started == started_before)
 			++spurious;
 		assert(completed <= bufs);
 		assert(started <= bufs);
@@ -203,7 +269,9 @@ static void run_test(struct vdev_info *dev, struct vq_info *vq,
 	test = 0;
 	r = ioctl(dev->control, VHOST_TEST_RUN, &test);
 	assert(r >= 0);
-	fprintf(stderr, "spurious wakeups: 0x%llx\n", spurious);
+	fprintf(stderr,
+		"spurious wakeups: 0x%llx started=0x%lx completed=0x%lx\n",
+		spurious, started, completed);
 }
 
 const char optstring[] = "h";
@@ -244,6 +312,16 @@ const struct option longopts[] = {
 		.name = "no-delayed-interrupt",
 		.val = 'd',
 	},
+	{
+		.name = "batch",
+		.val = 'b',
+		.has_arg = required_argument,
+	},
+	{
+		.name = "reset",
+		.val = 'r',
+		.has_arg = optional_argument,
+	},
 	{
 	}
 };
@@ -255,6 +333,8 @@ static void help(void)
 		" [--no-event-idx]"
 		" [--no-virtio-1]"
 		" [--delayed-interrupt]"
+		" [--batch=random/N]"
+		" [--reset=N]"
 		"\n");
 }
 
@@ -263,6 +343,7 @@ int main(int argc, char **argv)
 	struct vdev_info dev;
 	unsigned long long features = (1ULL << VIRTIO_RING_F_INDIRECT_DESC) |
 		(1ULL << VIRTIO_RING_F_EVENT_IDX) | (1ULL << VIRTIO_F_VERSION_1);
+	long batch = 1, reset = 0;
 	int o;
 	bool delayed = false;
 
@@ -289,6 +370,24 @@ int main(int argc, char **argv)
 		case 'D':
 			delayed = true;
 			break;
+		case 'b':
+			if (0 == strcmp(optarg, "random")) {
+				batch = RANDOM_BATCH;
+			} else {
+				batch = strtol(optarg, NULL, 10);
+				assert(batch > 0);
+				assert(batch < (long)INT_MAX + 1);
+			}
+			break;
+		case 'r':
+			if (!optarg) {
+				reset = 1;
+			} else {
+				reset = strtol(optarg, NULL, 10);
+				assert(reset > 0);
+				assert(reset < (long)INT_MAX + 1);
+			}
+			break;
 		default:
 			assert(0);
 			break;
@@ -298,6 +397,6 @@ int main(int argc, char **argv)
 done:
 	vdev_info_init(&dev, features);
 	vq_info_add(&dev, 256);
-	run_test(&dev, &dev.vqs[0], delayed, 0x100000);
+	run_test(&dev, &dev.vqs[0], delayed, batch, reset, 0x100000);
 	return 0;
 }

tools/virtio/vringh_test.c

@@ -307,6 +307,7 @@ static int parallel_test(u64 features,
 		close(to_host[0]);
 
 		gvdev.vdev.features = features;
+		INIT_LIST_HEAD(&gvdev.vdev.vqs);
 		gvdev.to_host_fd = to_host[1];
 		gvdev.notifies = 0;
 
@@ -453,6 +454,7 @@ int main(int argc, char *argv[])
 
 	getrange = getrange_iov;
 	vdev.features = 0;
+	INIT_LIST_HEAD(&vdev.vqs);
 
 	while (argv[1]) {
 		if (strcmp(argv[1], "--indirect") == 0)