vhost.c 39.1 KB
Newer Older
1 2 3 4 5 6
/* Copyright (C) 2009 Red Hat, Inc.
 * Copyright (C) 2006 Rusty Russell IBM Corporation
 *
 * Author: Michael S. Tsirkin <mst@redhat.com>
 *
 * Inspiration, some code, and most witty comments come from
Rob Landley's avatar
Rob Landley committed
7
 * Documentation/virtual/lguest/lguest.c, by Rusty Russell
8 9 10 11 12 13 14 15
 *
 * This work is licensed under the terms of the GNU GPL, version 2.
 *
 * Generic code for virtio server in host kernel.
 */

#include <linux/eventfd.h>
#include <linux/vhost.h>
16
#include <linux/uio.h>
17
#include <linux/mm.h>
18
#include <linux/mmu_context.h>
19 20 21 22 23 24
#include <linux/miscdevice.h>
#include <linux/mutex.h>
#include <linux/rcupdate.h>
#include <linux/poll.h>
#include <linux/file.h>
#include <linux/highmem.h>
25
#include <linux/slab.h>
26
#include <linux/kthread.h>
27
#include <linux/cgroup.h>
28
#include <linux/module.h>
29 30 31 32 33 34 35 36

#include "vhost.h"

enum {
	VHOST_MEMORY_MAX_NREGIONS = 64,
	VHOST_MEMORY_F_LOG = 0x1,
};

37 38 39
#define vhost_used_event(vq) ((u16 __user *)&vq->avail->ring[vq->num])
#define vhost_avail_event(vq) ((u16 __user *)&vq->used->ring[vq->num])

40 41 42 43 44
static void vhost_poll_func(struct file *file, wait_queue_head_t *wqh,
			    poll_table *pt)
{
	struct vhost_poll *poll;

45
	poll = container_of(pt, struct vhost_poll, table);
46 47 48 49 50 51 52
	poll->wqh = wqh;
	add_wait_queue(wqh, &poll->wait);
}

static int vhost_poll_wakeup(wait_queue_t *wait, unsigned mode, int sync,
			     void *key)
{
53 54
	struct vhost_poll *poll = container_of(wait, struct vhost_poll, wait);

55 56 57
	if (!((unsigned long)key & poll->mask))
		return 0;

58
	vhost_poll_queue(poll);
59 60 61
	return 0;
}

62
void vhost_work_init(struct vhost_work *work, vhost_work_fn_t fn)
63 64 65 66 67 68 69
{
	INIT_LIST_HEAD(&work->node);
	work->fn = fn;
	init_waitqueue_head(&work->done);
	work->flushing = 0;
	work->queue_seq = work->done_seq = 0;
}
70
EXPORT_SYMBOL_GPL(vhost_work_init);
71

72
/* Init poll structure */
73 74
void vhost_poll_init(struct vhost_poll *poll, vhost_work_fn_t fn,
		     unsigned long mask, struct vhost_dev *dev)
75 76 77 78
{
	init_waitqueue_func_entry(&poll->wait, vhost_poll_wakeup);
	init_poll_funcptr(&poll->table, vhost_poll_func);
	poll->mask = mask;
79
	poll->dev = dev;
80
	poll->wqh = NULL;
81

82
	vhost_work_init(&poll->work, fn);
83
}
84
EXPORT_SYMBOL_GPL(vhost_poll_init);
85 86 87

/* Start polling a file. We add ourselves to file's wait queue. The caller must
 * keep a reference to a file until after vhost_poll_stop is called. */
88
int vhost_poll_start(struct vhost_poll *poll, struct file *file)
89 90
{
	unsigned long mask;
91
	int ret = 0;
92

93 94 95
	if (poll->wqh)
		return 0;

96 97 98
	mask = file->f_op->poll(file, &poll->table);
	if (mask)
		vhost_poll_wakeup(&poll->wait, 0, 0, (void *)mask);
99 100 101 102 103 104 105
	if (mask & POLLERR) {
		if (poll->wqh)
			remove_wait_queue(poll->wqh, &poll->wait);
		ret = -EINVAL;
	}

	return ret;
106
}
107
EXPORT_SYMBOL_GPL(vhost_poll_start);
108 109 110 111 112

/* Stop polling a file. After this function returns, it becomes safe to drop the
 * file reference. You must also flush afterwards. */
void vhost_poll_stop(struct vhost_poll *poll)
{
113 114 115 116
	if (poll->wqh) {
		remove_wait_queue(poll->wqh, &poll->wait);
		poll->wqh = NULL;
	}
117
}
118
EXPORT_SYMBOL_GPL(vhost_poll_stop);
119

120 121 122 123
static bool vhost_work_seq_done(struct vhost_dev *dev, struct vhost_work *work,
				unsigned seq)
{
	int left;
124

125 126 127 128 129 130
	spin_lock_irq(&dev->work_lock);
	left = seq - work->done_seq;
	spin_unlock_irq(&dev->work_lock);
	return left <= 0;
}

131
void vhost_work_flush(struct vhost_dev *dev, struct vhost_work *work)
132
{
133 134 135
	unsigned seq;
	int flushing;

136
	spin_lock_irq(&dev->work_lock);
137 138
	seq = work->queue_seq;
	work->flushing++;
139
	spin_unlock_irq(&dev->work_lock);
140
	wait_event(work->done, vhost_work_seq_done(dev, work, seq));
141
	spin_lock_irq(&dev->work_lock);
142
	flushing = --work->flushing;
143
	spin_unlock_irq(&dev->work_lock);
144
	BUG_ON(flushing < 0);
145
}
146
EXPORT_SYMBOL_GPL(vhost_work_flush);
147

148 149 150 151 152 153
/* Flush any work that has been scheduled. When calling this, don't hold any
 * locks that are also used by the callback. */
void vhost_poll_flush(struct vhost_poll *poll)
{
	vhost_work_flush(poll->dev, &poll->work);
}
154
EXPORT_SYMBOL_GPL(vhost_poll_flush);
155

156
void vhost_work_queue(struct vhost_dev *dev, struct vhost_work *work)
157
{
158 159 160 161 162 163
	unsigned long flags;

	spin_lock_irqsave(&dev->work_lock, flags);
	if (list_empty(&work->node)) {
		list_add_tail(&work->node, &dev->work_list);
		work->queue_seq++;
164
		spin_unlock_irqrestore(&dev->work_lock, flags);
165
		wake_up_process(dev->worker);
166 167
	} else {
		spin_unlock_irqrestore(&dev->work_lock, flags);
168
	}
169
}
170
EXPORT_SYMBOL_GPL(vhost_work_queue);
171

172 173 174 175
void vhost_poll_queue(struct vhost_poll *poll)
{
	vhost_work_queue(poll->dev, &poll->work);
}
176
EXPORT_SYMBOL_GPL(vhost_poll_queue);
177

178 179 180 181 182 183 184 185 186 187
static void vhost_vq_reset(struct vhost_dev *dev,
			   struct vhost_virtqueue *vq)
{
	vq->num = 1;
	vq->desc = NULL;
	vq->avail = NULL;
	vq->used = NULL;
	vq->last_avail_idx = 0;
	vq->avail_idx = 0;
	vq->last_used_idx = 0;
188 189
	vq->signalled_used = 0;
	vq->signalled_used_valid = false;
190 191 192 193 194 195 196 197 198 199
	vq->used_flags = 0;
	vq->log_used = false;
	vq->log_addr = -1ull;
	vq->private_data = NULL;
	vq->log_base = NULL;
	vq->error_ctx = NULL;
	vq->error = NULL;
	vq->kick = NULL;
	vq->call_ctx = NULL;
	vq->call = NULL;
200
	vq->log_ctx = NULL;
201 202
}

203 204 205 206 207
static int vhost_worker(void *data)
{
	struct vhost_dev *dev = data;
	struct vhost_work *work = NULL;
	unsigned uninitialized_var(seq);
208
	mm_segment_t oldfs = get_fs();
209

210
	set_fs(USER_DS);
211 212
	use_mm(dev->mm);

213 214 215 216 217 218 219 220 221 222 223 224 225 226
	for (;;) {
		/* mb paired w/ kthread_stop */
		set_current_state(TASK_INTERRUPTIBLE);

		spin_lock_irq(&dev->work_lock);
		if (work) {
			work->done_seq = seq;
			if (work->flushing)
				wake_up_all(&work->done);
		}

		if (kthread_should_stop()) {
			spin_unlock_irq(&dev->work_lock);
			__set_current_state(TASK_RUNNING);
227
			break;
228 229 230 231 232 233 234 235 236 237 238 239 240
		}
		if (!list_empty(&dev->work_list)) {
			work = list_first_entry(&dev->work_list,
						struct vhost_work, node);
			list_del_init(&work->node);
			seq = work->queue_seq;
		} else
			work = NULL;
		spin_unlock_irq(&dev->work_lock);

		if (work) {
			__set_current_state(TASK_RUNNING);
			work->fn(work);
241 242
			if (need_resched())
				schedule();
243 244 245 246
		} else
			schedule();

	}
247
	unuse_mm(dev->mm);
248
	set_fs(oldfs);
249
	return 0;
250 251
}

252 253 254 255 256 257 258 259 260 261
static void vhost_vq_free_iovecs(struct vhost_virtqueue *vq)
{
	kfree(vq->indirect);
	vq->indirect = NULL;
	kfree(vq->log);
	vq->log = NULL;
	kfree(vq->heads);
	vq->heads = NULL;
}

Jason Wang's avatar
Jason Wang committed
262 263 264
/* Helper to allocate iovec buffers for all vqs. */
static long vhost_dev_alloc_iovecs(struct vhost_dev *dev)
{
265
	struct vhost_virtqueue *vq;
Jason Wang's avatar
Jason Wang committed
266
	int i;
267

Jason Wang's avatar
Jason Wang committed
268
	for (i = 0; i < dev->nvqs; ++i) {
269 270 271 272 273 274
		vq = dev->vqs[i];
		vq->indirect = kmalloc(sizeof *vq->indirect * UIO_MAXIOV,
				       GFP_KERNEL);
		vq->log = kmalloc(sizeof *vq->log * UIO_MAXIOV, GFP_KERNEL);
		vq->heads = kmalloc(sizeof *vq->heads * UIO_MAXIOV, GFP_KERNEL);
		if (!vq->indirect || !vq->log || !vq->heads)
Jason Wang's avatar
Jason Wang committed
275 276 277
			goto err_nomem;
	}
	return 0;
278

Jason Wang's avatar
Jason Wang committed
279
err_nomem:
280
	for (; i >= 0; --i)
281
		vhost_vq_free_iovecs(dev->vqs[i]);
Jason Wang's avatar
Jason Wang committed
282 283 284 285 286 287
	return -ENOMEM;
}

static void vhost_dev_free_iovecs(struct vhost_dev *dev)
{
	int i;
288

289
	for (i = 0; i < dev->nvqs; ++i)
290
		vhost_vq_free_iovecs(dev->vqs[i]);
Jason Wang's avatar
Jason Wang committed
291 292
}

Zhi Yong Wu's avatar
Zhi Yong Wu committed
293
void vhost_dev_init(struct vhost_dev *dev,
294
		    struct vhost_virtqueue **vqs, int nvqs)
295
{
296
	struct vhost_virtqueue *vq;
297
	int i;
298

299 300 301 302 303 304 305
	dev->vqs = vqs;
	dev->nvqs = nvqs;
	mutex_init(&dev->mutex);
	dev->log_ctx = NULL;
	dev->log_file = NULL;
	dev->memory = NULL;
	dev->mm = NULL;
306 307 308
	spin_lock_init(&dev->work_lock);
	INIT_LIST_HEAD(&dev->work_list);
	dev->worker = NULL;
309 310

	for (i = 0; i < dev->nvqs; ++i) {
311 312 313 314 315 316 317 318 319 320
		vq = dev->vqs[i];
		vq->log = NULL;
		vq->indirect = NULL;
		vq->heads = NULL;
		vq->dev = dev;
		mutex_init(&vq->mutex);
		vhost_vq_reset(dev, vq);
		if (vq->handle_kick)
			vhost_poll_init(&vq->poll, vq->handle_kick,
					POLLIN, dev);
321 322
	}
}
323
EXPORT_SYMBOL_GPL(vhost_dev_init);
324 325 326 327 328 329 330

/* Caller should have device mutex */
long vhost_dev_check_owner(struct vhost_dev *dev)
{
	/* Are you the owner? If not, I don't think you mean to do that */
	return dev->mm == current->mm ? 0 : -EPERM;
}
331
EXPORT_SYMBOL_GPL(vhost_dev_check_owner);
332

333
struct vhost_attach_cgroups_struct {
334 335 336
	struct vhost_work work;
	struct task_struct *owner;
	int ret;
337 338 339 340
};

static void vhost_attach_cgroups_work(struct vhost_work *work)
{
341 342 343 344
	struct vhost_attach_cgroups_struct *s;

	s = container_of(work, struct vhost_attach_cgroups_struct, work);
	s->ret = cgroup_attach_task_all(s->owner, current);
345 346 347 348
}

static int vhost_attach_cgroups(struct vhost_dev *dev)
{
349 350 351 352 353 354 355
	struct vhost_attach_cgroups_struct attach;

	attach.owner = current;
	vhost_work_init(&attach.work, vhost_attach_cgroups_work);
	vhost_work_queue(dev, &attach.work);
	vhost_work_flush(dev, &attach.work);
	return attach.ret;
356 357
}

358 359 360 361 362
/* Caller should have device mutex */
bool vhost_dev_has_owner(struct vhost_dev *dev)
{
	return dev->mm;
}
363
EXPORT_SYMBOL_GPL(vhost_dev_has_owner);
364

365
/* Caller should have device mutex */
Asias He's avatar
Asias He committed
366
long vhost_dev_set_owner(struct vhost_dev *dev)
367
{
368 369
	struct task_struct *worker;
	int err;
370

371
	/* Is there an owner already? */
372
	if (vhost_dev_has_owner(dev)) {
373 374 375
		err = -EBUSY;
		goto err_mm;
	}
376

377 378
	/* No owner, become one */
	dev->mm = get_task_mm(current);
379 380 381 382 383 384 385
	worker = kthread_create(vhost_worker, dev, "vhost-%d", current->pid);
	if (IS_ERR(worker)) {
		err = PTR_ERR(worker);
		goto err_worker;
	}

	dev->worker = worker;
386 387 388
	wake_up_process(worker);	/* avoid contributing to loadavg */

	err = vhost_attach_cgroups(dev);
389 390
	if (err)
		goto err_cgroup;
391

Jason Wang's avatar
Jason Wang committed
392 393 394 395
	err = vhost_dev_alloc_iovecs(dev);
	if (err)
		goto err_cgroup;

396
	return 0;
397 398
err_cgroup:
	kthread_stop(worker);
399
	dev->worker = NULL;
400 401 402 403 404 405
err_worker:
	if (dev->mm)
		mmput(dev->mm);
	dev->mm = NULL;
err_mm:
	return err;
406
}
407
EXPORT_SYMBOL_GPL(vhost_dev_set_owner);
408

409
struct vhost_memory *vhost_dev_reset_owner_prepare(void)
410
{
411 412
	return kmalloc(offsetof(struct vhost_memory, regions), GFP_KERNEL);
}
413
EXPORT_SYMBOL_GPL(vhost_dev_reset_owner_prepare);
414

415 416 417
/* Caller should have device mutex */
void vhost_dev_reset_owner(struct vhost_dev *dev, struct vhost_memory *memory)
{
418
	vhost_dev_cleanup(dev, true);
419

420
	/* Restore memory to default empty mapping. */
421
	memory->nregions = 0;
Arnd Bergmann's avatar
Arnd Bergmann committed
422
	RCU_INIT_POINTER(dev->memory, memory);
423
}
424
EXPORT_SYMBOL_GPL(vhost_dev_reset_owner);
425

426
void vhost_dev_stop(struct vhost_dev *dev)
427 428
{
	int i;
429 430

	for (i = 0; i < dev->nvqs; ++i) {
431 432 433
		if (dev->vqs[i]->kick && dev->vqs[i]->handle_kick) {
			vhost_poll_stop(&dev->vqs[i]->poll);
			vhost_poll_flush(&dev->vqs[i]->poll);
434
		}
435 436
	}
}
437
EXPORT_SYMBOL_GPL(vhost_dev_stop);
438

439 440
/* Caller should have device mutex if and only if locked is set */
void vhost_dev_cleanup(struct vhost_dev *dev, bool locked)
441 442
{
	int i;
443

444
	for (i = 0; i < dev->nvqs; ++i) {
445 446 447 448 449 450 451 452 453 454 455
		if (dev->vqs[i]->error_ctx)
			eventfd_ctx_put(dev->vqs[i]->error_ctx);
		if (dev->vqs[i]->error)
			fput(dev->vqs[i]->error);
		if (dev->vqs[i]->kick)
			fput(dev->vqs[i]->kick);
		if (dev->vqs[i]->call_ctx)
			eventfd_ctx_put(dev->vqs[i]->call_ctx);
		if (dev->vqs[i]->call)
			fput(dev->vqs[i]->call);
		vhost_vq_reset(dev, dev->vqs[i]);
456
	}
Jason Wang's avatar
Jason Wang committed
457
	vhost_dev_free_iovecs(dev);
458 459 460 461 462 463 464
	if (dev->log_ctx)
		eventfd_ctx_put(dev->log_ctx);
	dev->log_ctx = NULL;
	if (dev->log_file)
		fput(dev->log_file);
	dev->log_file = NULL;
	/* No one will access memory at this point */
Arnd Bergmann's avatar
Arnd Bergmann committed
465
	kfree(rcu_dereference_protected(dev->memory,
466 467
					locked ==
						lockdep_is_held(&dev->mutex)));
Arnd Bergmann's avatar
Arnd Bergmann committed
468
	RCU_INIT_POINTER(dev->memory, NULL);
469
	WARN_ON(!list_empty(&dev->work_list));
470 471 472 473
	if (dev->worker) {
		kthread_stop(dev->worker);
		dev->worker = NULL;
	}
474 475 476
	if (dev->mm)
		mmput(dev->mm);
	dev->mm = NULL;
477
}
478
EXPORT_SYMBOL_GPL(vhost_dev_cleanup);
479 480 481 482

static int log_access_ok(void __user *log_base, u64 addr, unsigned long sz)
{
	u64 a = addr / VHOST_PAGE_SIZE / 8;
483

484 485 486
	/* Make sure 64 bit math will not overflow. */
	if (a > ULONG_MAX - (unsigned long)log_base ||
	    a + (unsigned long)log_base > ULONG_MAX)
487
		return 0;
488 489 490 491 492 493 494 495 496 497

	return access_ok(VERIFY_WRITE, log_base + a,
			 (sz + VHOST_PAGE_SIZE * 8 - 1) / VHOST_PAGE_SIZE / 8);
}

/* Caller should have vq mutex and device mutex. */
static int vq_memory_access_ok(void __user *log_base, struct vhost_memory *mem,
			       int log_all)
{
	int i;
498

Michael S. Tsirkin's avatar
Michael S. Tsirkin committed
499 500
	if (!mem)
		return 0;
501

502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
	for (i = 0; i < mem->nregions; ++i) {
		struct vhost_memory_region *m = mem->regions + i;
		unsigned long a = m->userspace_addr;
		if (m->memory_size > ULONG_MAX)
			return 0;
		else if (!access_ok(VERIFY_WRITE, (void __user *)a,
				    m->memory_size))
			return 0;
		else if (log_all && !log_access_ok(log_base,
						   m->guest_phys_addr,
						   m->memory_size))
			return 0;
	}
	return 1;
}

/* Can we switch to this memory table? */
/* Caller should have device mutex but not vq mutex */
static int memory_access_ok(struct vhost_dev *d, struct vhost_memory *mem,
			    int log_all)
{
	int i;
524

525 526
	for (i = 0; i < d->nvqs; ++i) {
		int ok;
527
		mutex_lock(&d->vqs[i]->mutex);
528
		/* If ring is inactive, will check when it's enabled. */
529 530
		if (d->vqs[i]->private_data)
			ok = vq_memory_access_ok(d->vqs[i]->log_base, mem,
531 532 533
						 log_all);
		else
			ok = 1;
534
		mutex_unlock(&d->vqs[i]->mutex);
535 536 537 538 539 540
		if (!ok)
			return 0;
	}
	return 1;
}

541
static int vq_access_ok(struct vhost_dev *d, unsigned int num,
542 543 544 545
			struct vring_desc __user *desc,
			struct vring_avail __user *avail,
			struct vring_used __user *used)
{
546
	size_t s = vhost_has_feature(d, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0;
547 548
	return access_ok(VERIFY_READ, desc, num * sizeof *desc) &&
	       access_ok(VERIFY_READ, avail,
549
			 sizeof *avail + num * sizeof *avail->ring + s) &&
550
	       access_ok(VERIFY_WRITE, used,
551
			sizeof *used + num * sizeof *used->ring + s);
552 553 554 555 556 557
}

/* Can we log writes? */
/* Caller should have device mutex but not vq mutex */
int vhost_log_access_ok(struct vhost_dev *dev)
{
Arnd Bergmann's avatar
Arnd Bergmann committed
558 559 560 561 562
	struct vhost_memory *mp;

	mp = rcu_dereference_protected(dev->memory,
				       lockdep_is_held(&dev->mutex));
	return memory_access_ok(dev, mp, 1);
563
}
564
EXPORT_SYMBOL_GPL(vhost_log_access_ok);
565 566 567

/* Verify access for write logging. */
/* Caller should have vq mutex and device mutex */
568 569
static int vq_log_access_ok(struct vhost_dev *d, struct vhost_virtqueue *vq,
			    void __user *log_base)
570
{
Arnd Bergmann's avatar
Arnd Bergmann committed
571
	struct vhost_memory *mp;
572
	size_t s = vhost_has_feature(d, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0;
Arnd Bergmann's avatar
Arnd Bergmann committed
573 574 575 576

	mp = rcu_dereference_protected(vq->dev->memory,
				       lockdep_is_held(&vq->mutex));
	return vq_memory_access_ok(log_base, mp,
577 578 579
			    vhost_has_feature(vq->dev, VHOST_F_LOG_ALL)) &&
		(!vq->log_used || log_access_ok(log_base, vq->log_addr,
					sizeof *vq->used +
580
					vq->num * sizeof *vq->used->ring + s));
581 582 583 584 585 586
}

/* Can we start vq? */
/* Caller should have vq mutex and device mutex */
int vhost_vq_access_ok(struct vhost_virtqueue *vq)
{
587 588
	return vq_access_ok(vq->dev, vq->num, vq->desc, vq->avail, vq->used) &&
		vq_log_access_ok(vq->dev, vq, vq->log_base);
589
}
590
EXPORT_SYMBOL_GPL(vhost_vq_access_ok);
591 592 593 594 595

static long vhost_set_memory(struct vhost_dev *d, struct vhost_memory __user *m)
{
	struct vhost_memory mem, *newmem, *oldmem;
	unsigned long size = offsetof(struct vhost_memory, regions);
596
	int i;
597

598 599
	if (copy_from_user(&mem, m, size))
		return -EFAULT;
600 601 602 603 604 605 606 607 608
	if (mem.padding)
		return -EOPNOTSUPP;
	if (mem.nregions > VHOST_MEMORY_MAX_NREGIONS)
		return -E2BIG;
	newmem = kmalloc(size + mem.nregions * sizeof *m->regions, GFP_KERNEL);
	if (!newmem)
		return -ENOMEM;

	memcpy(newmem, &mem, size);
609 610
	if (copy_from_user(newmem->regions, m->regions,
			   mem.nregions * sizeof *m->regions)) {
611
		kfree(newmem);
612
		return -EFAULT;
613 614
	}

615 616
	if (!memory_access_ok(d, newmem,
			      vhost_has_feature(d, VHOST_F_LOG_ALL))) {
617
		kfree(newmem);
618
		return -EFAULT;
619
	}
Arnd Bergmann's avatar
Arnd Bergmann committed
620 621
	oldmem = rcu_dereference_protected(d->memory,
					   lockdep_is_held(&d->mutex));
622
	rcu_assign_pointer(d->memory, newmem);
623 624 625 626 627 628 629 630

	/* All memory accesses are done under some VQ mutex.
	 * So below is a faster equivalent of synchronize_rcu()
	 */
	for (i = 0; i < d->nvqs; ++i) {
		mutex_lock(&d->vqs[i]->mutex);
		mutex_unlock(&d->vqs[i]->mutex);
	}
631 632 633 634
	kfree(oldmem);
	return 0;
}

635
long vhost_vring_ioctl(struct vhost_dev *d, int ioctl, void __user *argp)
636
{
637 638
	struct file *eventfp, *filep = NULL;
	bool pollstart = false, pollstop = false;
639 640 641 642 643 644 645 646 647 648 649 650
	struct eventfd_ctx *ctx = NULL;
	u32 __user *idxp = argp;
	struct vhost_virtqueue *vq;
	struct vhost_vring_state s;
	struct vhost_vring_file f;
	struct vhost_vring_addr a;
	u32 idx;
	long r;

	r = get_user(idx, idxp);
	if (r < 0)
		return r;
651
	if (idx >= d->nvqs)
652 653
		return -ENOBUFS;

654
	vq = d->vqs[idx];
655 656 657 658 659 660 661 662 663 664 665

	mutex_lock(&vq->mutex);

	switch (ioctl) {
	case VHOST_SET_VRING_NUM:
		/* Resizing ring with an active backend?
		 * You don't want to do that. */
		if (vq->private_data) {
			r = -EBUSY;
			break;
		}
666 667
		if (copy_from_user(&s, argp, sizeof s)) {
			r = -EFAULT;
668
			break;
669
		}
670 671 672 673 674 675 676 677 678 679 680 681 682
		if (!s.num || s.num > 0xffff || (s.num & (s.num - 1))) {
			r = -EINVAL;
			break;
		}
		vq->num = s.num;
		break;
	case VHOST_SET_VRING_BASE:
		/* Moving base with an active backend?
		 * You don't want to do that. */
		if (vq->private_data) {
			r = -EBUSY;
			break;
		}
683 684
		if (copy_from_user(&s, argp, sizeof s)) {
			r = -EFAULT;
685
			break;
686
		}
687 688 689 690 691 692 693 694 695 696 697
		if (s.num > 0xffff) {
			r = -EINVAL;
			break;
		}
		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;
698 699
		if (copy_to_user(argp, &s, sizeof s))
			r = -EFAULT;
700 701
		break;
	case VHOST_SET_VRING_ADDR:
702 703
		if (copy_from_user(&a, argp, sizeof a)) {
			r = -EFAULT;
704
			break;
705
		}
706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728
		if (a.flags & ~(0x1 << VHOST_VRING_F_LOG)) {
			r = -EOPNOTSUPP;
			break;
		}
		/* For 32bit, verify that the top 32bits of the user
		   data are set to zero. */
		if ((u64)(unsigned long)a.desc_user_addr != a.desc_user_addr ||
		    (u64)(unsigned long)a.used_user_addr != a.used_user_addr ||
		    (u64)(unsigned long)a.avail_user_addr != a.avail_user_addr) {
			r = -EFAULT;
			break;
		}
		if ((a.avail_user_addr & (sizeof *vq->avail->ring - 1)) ||
		    (a.used_user_addr & (sizeof *vq->used->ring - 1)) ||
		    (a.log_guest_addr & (sizeof *vq->used->ring - 1))) {
			r = -EINVAL;
			break;
		}

		/* We only verify access here if backend is configured.
		 * If it is not, we don't as size might not have been setup.
		 * We will verify when backend is configured. */
		if (vq->private_data) {
729
			if (!vq_access_ok(d, vq->num,
730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753
				(void __user *)(unsigned long)a.desc_user_addr,
				(void __user *)(unsigned long)a.avail_user_addr,
				(void __user *)(unsigned long)a.used_user_addr)) {
				r = -EINVAL;
				break;
			}

			/* Also validate log access for used ring if enabled. */
			if ((a.flags & (0x1 << VHOST_VRING_F_LOG)) &&
			    !log_access_ok(vq->log_base, a.log_guest_addr,
					   sizeof *vq->used +
					   vq->num * sizeof *vq->used->ring)) {
				r = -EINVAL;
				break;
			}
		}

		vq->log_used = !!(a.flags & (0x1 << VHOST_VRING_F_LOG));
		vq->desc = (void __user *)(unsigned long)a.desc_user_addr;
		vq->avail = (void __user *)(unsigned long)a.avail_user_addr;
		vq->log_addr = a.log_guest_addr;
		vq->used = (void __user *)(unsigned long)a.used_user_addr;
		break;
	case VHOST_SET_VRING_KICK:
754 755
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
756
			break;
757
		}
758
		eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
759 760 761 762
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
763
		if (eventfp != vq->kick) {
764 765
			pollstop = (filep = vq->kick) != NULL;
			pollstart = (vq->kick = eventfp) != NULL;
766 767 768 769
		} else
			filep = eventfp;
		break;
	case VHOST_SET_VRING_CALL:
770 771
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
772
			break;
773
		}
774
		eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
775 776 777 778
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
779 780 781 782 783 784 785 786 787 788
		if (eventfp != vq->call) {
			filep = vq->call;
			ctx = vq->call_ctx;
			vq->call = eventfp;
			vq->call_ctx = eventfp ?
				eventfd_ctx_fileget(eventfp) : NULL;
		} else
			filep = eventfp;
		break;
	case VHOST_SET_VRING_ERR:
789 790
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
791
			break;
792
		}
793
		eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
794 795 796 797
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819
		if (eventfp != vq->error) {
			filep = vq->error;
			vq->error = eventfp;
			ctx = vq->error_ctx;
			vq->error_ctx = eventfp ?
				eventfd_ctx_fileget(eventfp) : NULL;
		} else
			filep = eventfp;
		break;
	default:
		r = -ENOIOCTLCMD;
	}

	if (pollstop && vq->handle_kick)
		vhost_poll_stop(&vq->poll);

	if (ctx)
		eventfd_ctx_put(ctx);
	if (filep)
		fput(filep);

	if (pollstart && vq->handle_kick)
820
		r = vhost_poll_start(&vq->poll, vq->kick);
821 822 823 824 825 826 827

	mutex_unlock(&vq->mutex);

	if (pollstop && vq->handle_kick)
		vhost_poll_flush(&vq->poll);
	return r;
}
828
EXPORT_SYMBOL_GPL(vhost_vring_ioctl);
829 830

/* Caller must have device mutex */
831
long vhost_dev_ioctl(struct vhost_dev *d, unsigned int ioctl, void __user *argp)
832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854
{
	struct file *eventfp, *filep = NULL;
	struct eventfd_ctx *ctx = NULL;
	u64 p;
	long r;
	int i, fd;

	/* If you are not the owner, you can become one */
	if (ioctl == VHOST_SET_OWNER) {
		r = vhost_dev_set_owner(d);
		goto done;
	}

	/* You must be the owner to do anything else */
	r = vhost_dev_check_owner(d);
	if (r)
		goto done;

	switch (ioctl) {
	case VHOST_SET_MEM_TABLE:
		r = vhost_set_memory(d, argp);
		break;
	case VHOST_SET_LOG_BASE:
855 856
		if (copy_from_user(&p, argp, sizeof p)) {
			r = -EFAULT;
857
			break;
858
		}
859 860 861 862 863 864 865
		if ((u64)(unsigned long)p != p) {
			r = -EFAULT;
			break;
		}
		for (i = 0; i < d->nvqs; ++i) {
			struct vhost_virtqueue *vq;
			void __user *base = (void __user *)(unsigned long)p;
866
			vq = d->vqs[i];
867 868
			mutex_lock(&vq->mutex);
			/* If ring is inactive, will check when it's enabled. */
869
			if (vq->private_data && !vq_log_access_ok(d, vq, base))
870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892
				r = -EFAULT;
			else
				vq->log_base = base;
			mutex_unlock(&vq->mutex);
		}
		break;
	case VHOST_SET_LOG_FD:
		r = get_user(fd, (int __user *)argp);
		if (r < 0)
			break;
		eventfp = fd == -1 ? NULL : eventfd_fget(fd);
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
		if (eventfp != d->log_file) {
			filep = d->log_file;
			ctx = d->log_ctx;
			d->log_ctx = eventfp ?
				eventfd_ctx_fileget(eventfp) : NULL;
		} else
			filep = eventfp;
		for (i = 0; i < d->nvqs; ++i) {
893 894 895
			mutex_lock(&d->vqs[i]->mutex);
			d->vqs[i]->log_ctx = d->log_ctx;
			mutex_unlock(&d->vqs[i]->mutex);
896 897 898 899 900 901 902
		}
		if (ctx)
			eventfd_ctx_put(ctx);
		if (filep)
			fput(filep);
		break;
	default:
903
		r = -ENOIOCTLCMD;
904 905 906 907 908
		break;
	}
done:
	return r;
}
909
EXPORT_SYMBOL_GPL(vhost_dev_ioctl);
910 911 912 913 914 915

static const struct vhost_memory_region *find_region(struct vhost_memory *mem,
						     __u64 addr, __u32 len)
{
	struct vhost_memory_region *reg;
	int i;
916

917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938
	/* linear search is not brilliant, but we really have on the order of 6
	 * regions in practice */
	for (i = 0; i < mem->nregions; ++i) {
		reg = mem->regions + i;
		if (reg->guest_phys_addr <= addr &&
		    reg->guest_phys_addr + reg->memory_size - 1 >= addr)
			return reg;
	}
	return NULL;
}

/* TODO: This is really inefficient.  We need something like get_user()
 * (instruction directly accesses the data, with an exception table entry
 * returning -EFAULT). See Documentation/x86/exception-tables.txt.
 */
static int set_bit_to_user(int nr, void __user *addr)
{
	unsigned long log = (unsigned long)addr;
	struct page *page;
	void *base;
	int bit = nr + (log % PAGE_SIZE) * 8;
	int r;
939

940
	r = get_user_pages_fast(log, 1, 1, &page);
941
	if (r < 0)
942
		return r;
943
	BUG_ON(r != 1);
944
	base = kmap_atomic(page);
945
	set_bit(bit, base);
946
	kunmap_atomic(base);
947 948 949 950 951 952 953 954
	set_page_dirty_lock(page);
	put_page(page);
	return 0;
}

static int log_write(void __user *log_base,
		     u64 write_address, u64 write_length)
{
955
	u64 write_page = write_address / VHOST_PAGE_SIZE;
956
	int r;
957

958 959
	if (!write_length)
		return 0;
960
	write_length += write_address % VHOST_PAGE_SIZE;
961 962
	for (;;) {
		u64 base = (u64)(unsigned long)log_base;
963 964
		u64 log = base + write_page / 8;
		int bit = write_page % 8;
965 966 967 968 969 970 971 972
		if ((u64)(unsigned long)log != log)
			return -EFAULT;
		r = set_bit_to_user(bit, (void __user *)(unsigned long)log);
		if (r < 0)
			return r;
		if (write_length <= VHOST_PAGE_SIZE)
			break;
		write_length -= VHOST_PAGE_SIZE;
973
		write_page += 1;
974 975 976 977 978 979 980 981 982 983
	}
	return r;
}

int vhost_log_write(struct vhost_virtqueue *vq, struct vhost_log *log,
		    unsigned int log_num, u64 len)
{
	int i, r;

	/* Make sure data written is seen before log. */
984
	smp_wmb();
985 986 987 988 989 990
	for (i = 0; i < log_num; ++i) {
		u64 l = min(log[i].len, len);
		r = log_write(vq->log_base, log[i].addr, l);
		if (r < 0)
			return r;
		len -= l;
991 992 993
		if (!len) {
			if (vq->log_ctx)
				eventfd_signal(vq->log_ctx, 1);
994
			return 0;
995
		}
996 997 998 999 1000
	}
	/* Length written exceeds what we have stored. This is a bug. */
	BUG();
	return 0;
}
1001
EXPORT_SYMBOL_GPL(vhost_log_write);
1002

1003 1004 1005
static int vhost_update_used_flags(struct vhost_virtqueue *vq)
{
	void __user *used;
1006
	if (__put_user(vq->used_flags, &vq->used->flags) < 0)
1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023
		return -EFAULT;
	if (unlikely(vq->log_used)) {
		/* Make sure the flag is seen before log. */
		smp_wmb();
		/* Log used flag write. */
		used = &vq->used->flags;
		log_write(vq->log_base, vq->log_addr +
			  (used - (void __user *)vq->used),
			  sizeof vq->used->flags);
		if (vq->log_ctx)
			eventfd_signal(vq->log_ctx, 1);
	}
	return 0;
}

static int vhost_update_avail_event(struct vhost_virtqueue *vq, u16 avail_event)
{
1024
	if (__put_user(vq->avail_idx, vhost_avail_event(vq)))
1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052
		return -EFAULT;
	if (unlikely(vq->log_used)) {
		void __user *used;
		/* Make sure the event is seen before log. */
		smp_wmb();
		/* Log avail event write */
		used = vhost_avail_event(vq);
		log_write(vq->log_base, vq->log_addr +
			  (used - (void __user *)vq->used),
			  sizeof *vhost_avail_event(vq));
		if (vq->log_ctx)
			eventfd_signal(vq->log_ctx, 1);
	}
	return 0;
}

int vhost_init_used(struct vhost_virtqueue *vq)
{
	int r;
	if (!vq->private_data)
		return 0;

	r = vhost_update_used_flags(vq);
	if (r)
		return r;
	vq->signalled_used_valid = false;
	return get_user(vq->last_used_idx, &vq->used->idx);
}
1053
EXPORT_SYMBOL_GPL(vhost_init_used);
1054

1055 1056
static int translate_desc(struct vhost_dev *dev, u64 addr, u32 len,
			  struct iovec iov[], int iov_size)
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
{
	const struct vhost_memory_region *reg;
	struct vhost_memory *mem;
	struct iovec *_iov;
	u64 s = 0;
	int ret = 0;

	rcu_read_lock();

	mem = rcu_dereference(dev->memory);
	while ((u64)len > s) {
		u64 size;
1069
		if (unlikely(ret >= iov_size)) {
1070 1071 1072 1073
			ret = -ENOBUFS;
			break;
		}
		reg = find_region(mem, addr, len);
1074
		if (unlikely(!reg)) {
1075 1076 1077 1078 1079
			ret = -EFAULT;
			break;
		}
		_iov = iov + ret;
		size = reg->memory_size - addr + reg->guest_phys_addr;
1080
		_iov->iov_len = min((u64)len - s, size);
1081
		_iov->iov_base = (void __user *)(unsigned long)
1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112
			(reg->userspace_addr + addr - reg->guest_phys_addr);
		s += size;
		addr += size;
		++ret;
	}

	rcu_read_unlock();
	return ret;
}

/* Each buffer in the virtqueues is actually a chain of descriptors.  This
 * function returns the next descriptor in the chain,
 * or -1U if we're at the end. */
static unsigned next_desc(struct vring_desc *desc)
{
	unsigned int next;

	/* If this descriptor says it doesn't chain, we're done. */
	if (!(desc->flags & VRING_DESC_F_NEXT))
		return -1U;

	/* Check they're not leading us off end of descriptors. */
	next = desc->next;
	/* Make sure compiler knows to grab that: we don't want it changing! */
	/* We will use the result as an index in an array, so most
	 * architectures only need a compiler barrier here. */
	read_barrier_depends();

	return next;
}

1113 1114 1115 1116 1117
static int get_indirect(struct vhost_dev *dev, struct vhost_virtqueue *vq,
			struct iovec iov[], unsigned int iov_size,
			unsigned int *out_num, unsigned int *in_num,
			struct vhost_log *log, unsigned int *log_num,
			struct vring_desc *indirect)
1118 1119 1120 1121 1122 1123
{
	struct vring_desc desc;
	unsigned int i = 0, count, found = 0;
	int ret;

	/* Sanity check */
1124
	if (unlikely(indirect->len % sizeof desc)) {
1125 1126 1127 1128 1129 1130 1131 1132
		vq_err(vq, "Invalid length in indirect descriptor: "
		       "len 0x%llx not multiple of 0x%zx\n",
		       (unsigned long long)indirect->len,
		       sizeof desc);
		return -EINVAL;
	}

	ret = translate_desc(dev, indirect->addr, indirect->len, vq->indirect,
Jason Wang's avatar
Jason Wang committed
1133
			     UIO_MAXIOV);
1134
	if (unlikely(ret < 0)) {
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
		vq_err(vq, "Translation failure %d in indirect.\n", ret);
		return ret;
	}

	/* We will use the result as an address to read from, so most
	 * architectures only need a compiler barrier here. */
	read_barrier_depends();

	count = indirect->len / sizeof desc;
	/* Buffers are chained via a 16 bit next field, so
	 * we can have at most 2^16 of these. */
1146
	if (unlikely(count > USHRT_MAX + 1)) {
1147 1148 1149 1150 1151 1152 1153
		vq_err(vq, "Indirect buffer length too big: %d\n",
		       indirect->len);
		return -E2BIG;
	}

	do {
		unsigned iov_count = *in_num + *out_num;
1154
		if (unlikely(++found > count)) {
1155 1156 1157 1158 1159
			vq_err(vq, "Loop detected: last one at %u "
			       "indirect size %u\n",
			       i, count);
			return -EINVAL;
		}
1160 1161
		if (unlikely(memcpy_fromiovec((unsigned char *)&desc,
					      vq->indirect, sizeof desc))) {
1162 1163 1164 1165
			vq_err(vq, "Failed indirect descriptor: idx %d, %zx\n",
			       i, (size_t)indirect->addr + i * sizeof desc);
			return -EINVAL;
		}
1166
		if (unlikely(desc.flags & VRING_DESC_F_INDIRECT)) {
1167 1168 1169 1170 1171 1172 1173
			vq_err(vq, "Nested indirect descriptor: idx %d, %zx\n",
			       i, (size_t)indirect->addr + i * sizeof desc);
			return -EINVAL;
		}

		ret = translate_desc(dev, desc.addr, desc.len, iov + iov_count,
				     iov_size - iov_count);
1174
		if (unlikely(ret < 0)) {
1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
			vq_err(vq, "Translation failure %d indirect idx %d\n",
			       ret, i);
			return ret;
		}
		/* If this is an input descriptor, increment that count. */
		if (desc.flags & VRING_DESC_F_WRITE) {
			*in_num += ret;
			if (unlikely(log)) {
				log[*log_num].addr = desc.addr;
				log[*log_num].len = desc.len;
				++*log_num;
			}
		} else {
			/* If it's an output descriptor, they're all supposed
			 * to come before any input descriptors. */
1190
			if (unlikely(*in_num)) {
1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205
				vq_err(vq, "Indirect descriptor "
				       "has out after in: idx %d\n", i);
				return -EINVAL;
			}
			*out_num += ret;
		}
	} while ((i = next_desc(&desc)) != -1);
	return 0;
}

/* This looks in the virtqueue and for the first available buffer, and converts
 * it to an iovec for convenient access.  Since descriptors consist of some
 * number of output then some number of input descriptors, it's actually two
 * iovecs, but we pack them into one and note how many of each there were.
 *
1206 1207 1208 1209 1210 1211 1212
 * This function returns the descriptor number found, or vq->num (which is
 * never a valid descriptor number) if none was found.  A negative code is
 * returned on error. */
int vhost_get_vq_desc(struct vhost_dev *dev, struct vhost_virtqueue *vq,
		      struct iovec iov[], unsigned int iov_size,
		      unsigned int *out_num, unsigned int *in_num,
		      struct vhost_log *log, unsigned int *log_num)
1213 1214 1215 1216 1217 1218 1219 1220
{
	struct vring_desc desc;
	unsigned int i, head, found = 0;
	u16 last_avail_idx;
	int ret;

	/* Check it isn't doing very strange things with descriptor numbers. */
	last_avail_idx = vq->last_avail_idx;
1221
	if (unlikely(__get_user(vq->avail_idx, &vq->avail->idx))) {
1222 1223
		vq_err(vq, "Failed to access avail idx at %p\n",
		       &vq->avail->idx);
1224
		return -EFAULT;
1225 1226
	}

1227
	if (unlikely((u16)(vq->avail_idx - last_avail_idx) > vq->num)) {
1228 1229
		vq_err(vq, "Guest moved used index from %u to %u",
		       last_avail_idx, vq->avail_idx);
1230
		return -EFAULT;
1231 1232 1233 1234 1235 1236 1237
	}

	/* If there's nothing new since last we looked, return invalid. */
	if (vq->avail_idx == last_avail_idx)
		return vq->num;

	/* Only get avail ring entries after they have been exposed by guest. */
1238
	smp_rmb();
1239 1240 1241

	/* Grab the next descriptor number they're advertising, and increment
	 * the index we've seen. */
1242 1243
	if (unlikely(__get_user(head,
				&vq->avail->ring[last_avail_idx % vq->num]))) {
1244 1245 1246
		vq_err(vq, "Failed to read head: idx %d address %p\n",
		       last_avail_idx,
		       &vq->avail->ring[last_avail_idx % vq->num]);
1247
		return -EFAULT;
1248 1249 1250
	}

	/* If their number is silly, that's an error. */
1251
	if (unlikely(head >= vq->num)) {
1252 1253
		vq_err(vq, "Guest says index %u > %u is available",
		       head, vq->num);
1254
		return -EINVAL;