vhost.c 40.6 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 16 17
 *
 * 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>
#include <linux/virtio_net.h>
#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 29 30 31 32 33 34 35 36 37 38 39

#include <linux/net.h>
#include <linux/if_packet.h>
#include <linux/if_arp.h>

#include "vhost.h"

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

40 41
static unsigned vhost_zcopy_mask __read_mostly;

42 43 44
#define vhost_used_event(vq) ((u16 __user *)&vq->avail->ring[vq->num])
#define vhost_avail_event(vq) ((u16 __user *)&vq->used->ring[vq->num])

45 46 47 48 49
static void vhost_poll_func(struct file *file, wait_queue_head_t *wqh,
			    poll_table *pt)
{
	struct vhost_poll *poll;

50
	poll = container_of(pt, struct vhost_poll, table);
51 52 53 54 55 56 57
	poll->wqh = wqh;
	add_wait_queue(wqh, &poll->wait);
}

static int vhost_poll_wakeup(wait_queue_t *wait, unsigned mode, int sync,
			     void *key)
{
58 59
	struct vhost_poll *poll = container_of(wait, struct vhost_poll, wait);

60 61 62
	if (!((unsigned long)key & poll->mask))
		return 0;

63
	vhost_poll_queue(poll);
64 65 66
	return 0;
}

67 68 69 70 71 72 73 74 75
static void vhost_work_init(struct vhost_work *work, vhost_work_fn_t fn)
{
	INIT_LIST_HEAD(&work->node);
	work->fn = fn;
	init_waitqueue_head(&work->done);
	work->flushing = 0;
	work->queue_seq = work->done_seq = 0;
}

76
/* Init poll structure */
77 78
void vhost_poll_init(struct vhost_poll *poll, vhost_work_fn_t fn,
		     unsigned long mask, struct vhost_dev *dev)
79 80 81 82
{
	init_waitqueue_func_entry(&poll->wait, vhost_poll_wakeup);
	init_poll_funcptr(&poll->table, vhost_poll_func);
	poll->mask = mask;
83 84
	poll->dev = dev;

85
	vhost_work_init(&poll->work, fn);
86 87 88 89 90 91 92
}

/* 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. */
void vhost_poll_start(struct vhost_poll *poll, struct file *file)
{
	unsigned long mask;
93

94 95 96 97 98 99 100 101 102 103 104 105
	mask = file->f_op->poll(file, &poll->table);
	if (mask)
		vhost_poll_wakeup(&poll->wait, 0, 0, (void *)mask);
}

/* 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)
{
	remove_wait_queue(poll->wqh, &poll->wait);
}

106 107 108 109
static bool vhost_work_seq_done(struct vhost_dev *dev, struct vhost_work *work,
				unsigned seq)
{
	int left;
110

111 112 113 114 115 116
	spin_lock_irq(&dev->work_lock);
	left = seq - work->done_seq;
	spin_unlock_irq(&dev->work_lock);
	return left <= 0;
}

117
static void vhost_work_flush(struct vhost_dev *dev, struct vhost_work *work)
118
{
119 120 121
	unsigned seq;
	int flushing;

122
	spin_lock_irq(&dev->work_lock);
123 124
	seq = work->queue_seq;
	work->flushing++;
125
	spin_unlock_irq(&dev->work_lock);
126
	wait_event(work->done, vhost_work_seq_done(dev, work, seq));
127
	spin_lock_irq(&dev->work_lock);
128
	flushing = --work->flushing;
129
	spin_unlock_irq(&dev->work_lock);
130
	BUG_ON(flushing < 0);
131 132
}

133 134 135 136 137 138 139 140 141
/* 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);
}

static inline void vhost_work_queue(struct vhost_dev *dev,
				    struct vhost_work *work)
142
{
143 144 145 146 147 148 149 150 151
	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++;
		wake_up_process(dev->worker);
	}
	spin_unlock_irqrestore(&dev->work_lock, flags);
152 153
}

154 155 156 157 158
void vhost_poll_queue(struct vhost_poll *poll)
{
	vhost_work_queue(poll->dev, &poll->work);
}

159 160 161 162 163 164 165 166 167 168
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;
169 170
	vq->signalled_used = 0;
	vq->signalled_used_valid = false;
171 172 173
	vq->used_flags = 0;
	vq->log_used = false;
	vq->log_addr = -1ull;
174 175
	vq->vhost_hlen = 0;
	vq->sock_hlen = 0;
176 177 178 179 180 181 182
	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;
183
	vq->log_ctx = NULL;
184 185 186
	vq->upend_idx = 0;
	vq->done_idx = 0;
	vq->ubufs = NULL;
187 188
}

189 190 191 192 193 194
static int vhost_worker(void *data)
{
	struct vhost_dev *dev = data;
	struct vhost_work *work = NULL;
	unsigned uninitialized_var(seq);

195 196
	use_mm(dev->mm);

197 198 199 200 201 202 203 204 205 206 207 208 209 210
	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);
211
			break;
212 213 214 215 216 217 218 219 220 221 222 223 224
		}
		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);
225 226
			if (need_resched())
				schedule();
227 228 229 230
		} else
			schedule();

	}
231 232
	unuse_mm(dev->mm);
	return 0;
233 234
}

235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
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;
	kfree(vq->ubuf_info);
	vq->ubuf_info = NULL;
}

void vhost_enable_zcopy(int vq)
{
	vhost_zcopy_mask |= 0x1 << vq;
}

Jason Wang's avatar
Jason Wang committed
252 253 254 255
/* Helper to allocate iovec buffers for all vqs. */
static long vhost_dev_alloc_iovecs(struct vhost_dev *dev)
{
	int i;
256
	bool zcopy;
257

Jason Wang's avatar
Jason Wang committed
258 259 260 261 262 263 264
	for (i = 0; i < dev->nvqs; ++i) {
		dev->vqs[i].indirect = kmalloc(sizeof *dev->vqs[i].indirect *
					       UIO_MAXIOV, GFP_KERNEL);
		dev->vqs[i].log = kmalloc(sizeof *dev->vqs[i].log * UIO_MAXIOV,
					  GFP_KERNEL);
		dev->vqs[i].heads = kmalloc(sizeof *dev->vqs[i].heads *
					    UIO_MAXIOV, GFP_KERNEL);
265 266 267 268 269
		zcopy = vhost_zcopy_mask & (0x1 << i);
		if (zcopy)
			dev->vqs[i].ubuf_info =
				kmalloc(sizeof *dev->vqs[i].ubuf_info *
					UIO_MAXIOV, GFP_KERNEL);
Jason Wang's avatar
Jason Wang committed
270
		if (!dev->vqs[i].indirect || !dev->vqs[i].log ||
271 272
			!dev->vqs[i].heads ||
			(zcopy && !dev->vqs[i].ubuf_info))
Jason Wang's avatar
Jason Wang committed
273 274 275
			goto err_nomem;
	}
	return 0;
276

Jason Wang's avatar
Jason Wang committed
277
err_nomem:
278 279
	for (; i >= 0; --i)
		vhost_vq_free_iovecs(&dev->vqs[i]);
Jason Wang's avatar
Jason Wang committed
280 281 282 283 284 285
	return -ENOMEM;
}

static void vhost_dev_free_iovecs(struct vhost_dev *dev)
{
	int i;
286

287 288
	for (i = 0; i < dev->nvqs; ++i)
		vhost_vq_free_iovecs(&dev->vqs[i]);
Jason Wang's avatar
Jason Wang committed
289 290
}

291 292 293 294
long vhost_dev_init(struct vhost_dev *dev,
		    struct vhost_virtqueue *vqs, int nvqs)
{
	int i;
295

296 297 298 299 300 301 302
	dev->vqs = vqs;
	dev->nvqs = nvqs;
	mutex_init(&dev->mutex);
	dev->log_ctx = NULL;
	dev->log_file = NULL;
	dev->memory = NULL;
	dev->mm = NULL;
303 304 305
	spin_lock_init(&dev->work_lock);
	INIT_LIST_HEAD(&dev->work_list);
	dev->worker = NULL;
306 307

	for (i = 0; i < dev->nvqs; ++i) {
Jason Wang's avatar
Jason Wang committed
308 309 310
		dev->vqs[i].log = NULL;
		dev->vqs[i].indirect = NULL;
		dev->vqs[i].heads = NULL;
311
		dev->vqs[i].ubuf_info = NULL;
312 313 314 315 316
		dev->vqs[i].dev = dev;
		mutex_init(&dev->vqs[i].mutex);
		vhost_vq_reset(dev, dev->vqs + i);
		if (dev->vqs[i].handle_kick)
			vhost_poll_init(&dev->vqs[i].poll,
317
					dev->vqs[i].handle_kick, POLLIN, dev);
318
	}
319

320 321 322 323 324 325 326 327 328 329
	return 0;
}

/* 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;
}

330
struct vhost_attach_cgroups_struct {
331 332 333
	struct vhost_work work;
	struct task_struct *owner;
	int ret;
334 335 336 337
};

static void vhost_attach_cgroups_work(struct vhost_work *work)
{
338 339 340 341
	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);
342 343 344 345
}

static int vhost_attach_cgroups(struct vhost_dev *dev)
{
346 347 348 349 350 351 352
	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;
353 354
}

355 356 357
/* Caller should have device mutex */
static long vhost_dev_set_owner(struct vhost_dev *dev)
{
358 359
	struct task_struct *worker;
	int err;
360

361
	/* Is there an owner already? */
362 363 364 365
	if (dev->mm) {
		err = -EBUSY;
		goto err_mm;
	}
366

367 368
	/* No owner, become one */
	dev->mm = get_task_mm(current);
369 370 371 372 373 374 375
	worker = kthread_create(vhost_worker, dev, "vhost-%d", current->pid);
	if (IS_ERR(worker)) {
		err = PTR_ERR(worker);
		goto err_worker;
	}

	dev->worker = worker;
376 377 378
	wake_up_process(worker);	/* avoid contributing to loadavg */

	err = vhost_attach_cgroups(dev);
379 380
	if (err)
		goto err_cgroup;
381

Jason Wang's avatar
Jason Wang committed
382 383 384 385
	err = vhost_dev_alloc_iovecs(dev);
	if (err)
		goto err_cgroup;

386
	return 0;
387 388
err_cgroup:
	kthread_stop(worker);
389
	dev->worker = NULL;
390 391 392 393 394 395
err_worker:
	if (dev->mm)
		mmput(dev->mm);
	dev->mm = NULL;
err_mm:
	return err;
396 397 398 399 400 401 402 403 404 405 406 407
}

/* Caller should have device mutex */
long vhost_dev_reset_owner(struct vhost_dev *dev)
{
	struct vhost_memory *memory;

	/* Restore memory to default empty mapping. */
	memory = kmalloc(offsetof(struct vhost_memory, regions), GFP_KERNEL);
	if (!memory)
		return -ENOMEM;

408
	vhost_dev_cleanup(dev, true);
409 410

	memory->nregions = 0;
Arnd Bergmann's avatar
Arnd Bergmann committed
411
	RCU_INIT_POINTER(dev->memory, memory);
412 413 414
	return 0;
}

415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
/* In case of DMA done not in order in lower device driver for some reason.
 * upend_idx is used to track end of used idx, done_idx is used to track head
 * of used idx. Once lower device DMA done contiguously, we will signal KVM
 * guest used idx.
 */
int vhost_zerocopy_signal_used(struct vhost_virtqueue *vq)
{
	int i;
	int j = 0;

	for (i = vq->done_idx; i != vq->upend_idx; i = (i + 1) % UIO_MAXIOV) {
		if ((vq->heads[i].len == VHOST_DMA_DONE_LEN)) {
			vq->heads[i].len = VHOST_DMA_CLEAR_LEN;
			vhost_add_used_and_signal(vq->dev, vq,
						  vq->heads[i].id, 0);
			++j;
		} else
			break;
	}
	if (j)
		vq->done_idx = i;
	return j;
}

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

444 445 446 447 448
	for (i = 0; i < dev->nvqs; ++i) {
		if (dev->vqs[i].kick && dev->vqs[i].handle_kick) {
			vhost_poll_stop(&dev->vqs[i].poll);
			vhost_poll_flush(&dev->vqs[i].poll);
		}
449 450 451 452 453 454 455
		/* Wait for all lower device DMAs done. */
		if (dev->vqs[i].ubufs)
			vhost_ubuf_put_and_wait(dev->vqs[i].ubufs);

		/* Signal guest as appropriate. */
		vhost_zerocopy_signal_used(&dev->vqs[i]);

456 457 458 459 460 461 462 463 464 465 466 467
		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);
	}
Jason Wang's avatar
Jason Wang committed
468
	vhost_dev_free_iovecs(dev);
469 470 471 472 473 474 475
	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
476
	kfree(rcu_dereference_protected(dev->memory,
477 478
					locked ==
						lockdep_is_held(&dev->mutex)));
Arnd Bergmann's avatar
Arnd Bergmann committed
479
	RCU_INIT_POINTER(dev->memory, NULL);
480
	WARN_ON(!list_empty(&dev->work_list));
481 482 483 484
	if (dev->worker) {
		kthread_stop(dev->worker);
		dev->worker = NULL;
	}
485 486 487
	if (dev->mm)
		mmput(dev->mm);
	dev->mm = NULL;
488 489 490 491 492
}

static int log_access_ok(void __user *log_base, u64 addr, unsigned long sz)
{
	u64 a = addr / VHOST_PAGE_SIZE / 8;
493

494 495 496
	/* Make sure 64 bit math will not overflow. */
	if (a > ULONG_MAX - (unsigned long)log_base ||
	    a + (unsigned long)log_base > ULONG_MAX)
497
		return 0;
498 499 500 501 502 503 504 505 506 507

	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;
508

Michael S. Tsirkin's avatar
Michael S. Tsirkin committed
509 510
	if (!mem)
		return 0;
511

512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533
	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;
534

535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
	for (i = 0; i < d->nvqs; ++i) {
		int ok;
		mutex_lock(&d->vqs[i].mutex);
		/* If ring is inactive, will check when it's enabled. */
		if (d->vqs[i].private_data)
			ok = vq_memory_access_ok(d->vqs[i].log_base, mem,
						 log_all);
		else
			ok = 1;
		mutex_unlock(&d->vqs[i].mutex);
		if (!ok)
			return 0;
	}
	return 1;
}

551
static int vq_access_ok(struct vhost_dev *d, unsigned int num,
552 553 554 555
			struct vring_desc __user *desc,
			struct vring_avail __user *avail,
			struct vring_used __user *used)
{
556
	size_t s = vhost_has_feature(d, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0;
557 558
	return access_ok(VERIFY_READ, desc, num * sizeof *desc) &&
	       access_ok(VERIFY_READ, avail,
559
			 sizeof *avail + num * sizeof *avail->ring + s) &&
560
	       access_ok(VERIFY_WRITE, used,
561
			sizeof *used + num * sizeof *used->ring + s);
562 563 564 565 566 567
}

/* 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
568 569 570 571 572
	struct vhost_memory *mp;

	mp = rcu_dereference_protected(dev->memory,
				       lockdep_is_held(&dev->mutex));
	return memory_access_ok(dev, mp, 1);
573 574 575 576
}

/* Verify access for write logging. */
/* Caller should have vq mutex and device mutex */
577 578
static int vq_log_access_ok(struct vhost_dev *d, struct vhost_virtqueue *vq,
			    void __user *log_base)
579
{
Arnd Bergmann's avatar
Arnd Bergmann committed
580
	struct vhost_memory *mp;
581
	size_t s = vhost_has_feature(d, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0;
Arnd Bergmann's avatar
Arnd Bergmann committed
582 583 584 585

	mp = rcu_dereference_protected(vq->dev->memory,
				       lockdep_is_held(&vq->mutex));
	return vq_memory_access_ok(log_base, mp,
586 587 588
			    vhost_has_feature(vq->dev, VHOST_F_LOG_ALL)) &&
		(!vq->log_used || log_access_ok(log_base, vq->log_addr,
					sizeof *vq->used +
589
					vq->num * sizeof *vq->used->ring + s));
590 591 592 593 594 595
}

/* Can we start vq? */
/* Caller should have vq mutex and device mutex */
int vhost_vq_access_ok(struct vhost_virtqueue *vq)
{
596 597
	return vq_access_ok(vq->dev, vq->num, vq->desc, vq->avail, vq->used) &&
		vq_log_access_ok(vq->dev, vq, vq->log_base);
598 599 600 601 602 603
}

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);
604

605 606
	if (copy_from_user(&mem, m, size))
		return -EFAULT;
607 608 609 610 611 612 613 614 615
	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);
616 617
	if (copy_from_user(newmem->regions, m->regions,
			   mem.nregions * sizeof *m->regions)) {
618
		kfree(newmem);
619
		return -EFAULT;
620 621
	}

622 623
	if (!memory_access_ok(d, newmem,
			      vhost_has_feature(d, VHOST_F_LOG_ALL))) {
624
		kfree(newmem);
625
		return -EFAULT;
626
	}
Arnd Bergmann's avatar
Arnd Bergmann committed
627 628
	oldmem = rcu_dereference_protected(d->memory,
					   lockdep_is_held(&d->mutex));
629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650
	rcu_assign_pointer(d->memory, newmem);
	synchronize_rcu();
	kfree(oldmem);
	return 0;
}

static long vhost_set_vring(struct vhost_dev *d, int ioctl, void __user *argp)
{
	struct file *eventfp, *filep = NULL,
		    *pollstart = NULL, *pollstop = NULL;
	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)
652 653 654 655 656 657 658 659 660 661 662 663 664 665
		return -ENOBUFS;

	vq = d->vqs + idx;

	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;
		}
763 764 765 766 767 768 769
		if (eventfp != vq->kick) {
			pollstop = filep = vq->kick;
			pollstart = vq->kick = eventfp;
		} 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;
		}
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854
		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)
		vhost_poll_start(&vq->poll, vq->kick);

	mutex_unlock(&vq->mutex);

	if (pollstop && vq->handle_kick)
		vhost_poll_flush(&vq->poll);
	return r;
}

/* Caller must have device mutex */
long vhost_dev_ioctl(struct vhost_dev *d, unsigned int ioctl, unsigned long arg)
{
	void __user *argp = (void __user *)arg;
	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
		}
859 860 861 862 863 864 865 866 867 868
		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;
			vq = d->vqs + i;
			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))
870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914
				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) {
			mutex_lock(&d->vqs[i].mutex);
			d->vqs[i].log_ctx = d->log_ctx;
			mutex_unlock(&d->vqs[i].mutex);
		}
		if (ctx)
			eventfd_ctx_put(ctx);
		if (filep)
			fput(filep);
		break;
	default:
		r = vhost_set_vring(d, ioctl, argp);
		break;
	}
done:
	return r;
}

static const struct vhost_memory_region *find_region(struct vhost_memory *mem,
						     __u64 addr, __u32 len)
{
	struct vhost_memory_region *reg;
	int i;
915

916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937
	/* 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;
938

939
	r = get_user_pages_fast(log, 1, 1, &page);
940
	if (r < 0)
941
		return r;
942
	BUG_ON(r != 1);
943
	base = kmap_atomic(page);
944
	set_bit(bit, base);
945
	kunmap_atomic(base);
946 947 948 949 950 951 952 953
	set_page_dirty_lock(page);
	put_page(page);
	return 0;
}

static int log_write(void __user *log_base,
		     u64 write_address, u64 write_length)
{
954
	u64 write_page = write_address / VHOST_PAGE_SIZE;
955
	int r;
956

957 958
	if (!write_length)
		return 0;
959
	write_length += write_address % VHOST_PAGE_SIZE;
960 961
	for (;;) {
		u64 base = (u64)(unsigned long)log_base;
962 963
		u64 log = base + write_page / 8;
		int bit = write_page % 8;
964 965 966 967 968 969 970 971
		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;
972
		write_page += 1;
973 974 975 976 977 978 979 980 981 982
	}
	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. */
983
	smp_wmb();
984 985 986 987 988 989
	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;
990 991 992
		if (!len) {
			if (vq->log_ctx)
				eventfd_signal(vq->log_ctx, 1);
993
			return 0;
994
		}
995 996 997 998 999 1000
	}
	/* Length written exceeds what we have stored. This is a bug. */
	BUG();
	return 0;
}

1001 1002 1003
static int vhost_update_used_flags(struct vhost_virtqueue *vq)
{
	void __user *used;
1004
	if (__put_user(vq->used_flags, &vq->used->flags) < 0)
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
		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)
{
1022
	if (__put_user(vq->avail_idx, vhost_avail_event(vq)))
1023 1024 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
		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);
}

1052 1053
static int translate_desc(struct vhost_dev *dev, u64 addr, u32 len,
			  struct iovec iov[], int iov_size)
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065
{
	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;
1066
		if (unlikely(ret >= iov_size)) {
1067 1068 1069 1070
			ret = -ENOBUFS;
			break;
		}
		reg = find_region(mem, addr, len);
1071
		if (unlikely(!reg)) {
1072 1073 1074 1075 1076 1077
			ret = -EFAULT;
			break;
		}
		_iov = iov + ret;
		size = reg->memory_size - addr + reg->guest_phys_addr;
		_iov->iov_len = min((u64)len, size);
1078
		_iov->iov_base = (void __user *)(unsigned long)
1079 1080 1081 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
			(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;
}

1110 1111 1112 1113 1114
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)
1115 1116 1117 1118 1119 1120
{
	struct vring_desc desc;
	unsigned int i = 0, count, found = 0;
	int ret;

	/* Sanity check */
1121
	if (unlikely(indirect->len % sizeof desc)) {
1122 1123 1124 1125 1126 1127 1128 1129
		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
1130
			     UIO_MAXIOV);
1131
	if (unlikely(ret < 0)) {
1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142
		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. */
1143
	if (unlikely(count > USHRT_MAX + 1)) {
1144 1145 1146 1147 1148 1149 1150
		vq_err(vq, "Indirect buffer length too big: %d\n",
		       indirect->len);
		return -E2BIG;
	}

	do {
		unsigned iov_count = *in_num + *out_num;
1151
		if (unlikely(++found > count)) {
1152 1153 1154 1155 1156
			vq_err(vq, "Loop detected: last one at %u "
			       "indirect size %u\n",
			       i, count);
			return -EINVAL;
		}
1157 1158
		if (unlikely(memcpy_fromiovec((unsigned char *)&desc,
					      vq->indirect, sizeof desc))) {
1159 1160 1161 1162
			vq_err(vq, "Failed indirect descriptor: idx %d, %zx\n",
			       i, (size_t)indirect->addr + i * sizeof desc);
			return -EINVAL;
		}
1163
		if (unlikely(desc.flags & VRING_DESC_F_INDIRECT)) {
1164 1165 1166 1167 1168 1169 1170
			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);
1171
		if (unlikely(ret < 0)) {
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186
			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. */
1187
			if (unlikely(*in_num)) {
1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
				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.
 *
1203 1204 1205 1206 1207 1208 1209
 * 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)
1210 1211 1212 1213 1214 1215 1216 1217
{
	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;
1218
	if (unlikely(__get_user(vq->avail_idx, &vq->avail->idx))) {
1219 1220
		vq_err(vq, "Failed to access avail idx at %p\n",
		       &vq->avail->idx);
1221
		return -EFAULT;
1222 1223
	}

1224
	if (unlikely((u16)(vq->avail_idx - last_avail_idx) > vq->num)) {
1225 1226
		vq_err(vq, "Guest moved used index from %u to %u",
		       last_avail_idx, vq->avail_idx);
1227
		return -EFAULT;
1228 1229 1230 1231 1232 1233 1234
	}

	/* 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. */
1235
	smp_rmb();
1236 1237 1238

	/* Grab the next descriptor number they're advertising, and increment
	 * the index we've seen. */
1239 1240
	if (unlikely(__get_user(head,
				&vq->avail->ring[last_avail_idx % vq->num]))) {
1241 1242 1243
		vq_err(vq, "Failed to read head: idx %d address %p\n",
		       last_avail_idx,
		       &vq->avail->ring[last_avail_idx % vq->num]);
1244
		return -EFAULT;
1245 1246 1247
	}

	/* If their number is silly, that's an error. */
1248
	if (unlikely(head >= vq->num)) {
1249 1250
		vq_err(vq, "Guest says index %u > %u is available",
		       head, vq->num);
1251
		return -EINVAL;
1252 1253 1254 1255 1256 1257 1258 1259 1260 1261
	}

	/* When we start there are none of either input nor output. */
	*out_num = *in_num = 0;
	if (unlikely(log))
		*log_num = 0;

	i = head;
	do {
		unsigned iov_count = *in_num + *out_num;
1262
		if (unlikely(i >= vq->num)) {
1263 1264
			vq_err(vq, "Desc index is %u > %u, head = %u",
			       i, vq->num, head);
1265
			return -EINVAL;
1266
		}
1267
		if (unlikely(++found > vq->num)) {
1268 1269 1270
			vq_err(vq, "Loop detected: last one at %u "
			       "vq size %u head %u\n",
			       i, vq->num, head);
1271
			return -EINVAL;
1272
		}
1273
		ret = __copy_from_user(&desc, vq->desc + i, sizeof desc);
1274
		if (unlikely(ret)) {
1275 1276
			vq_err(vq, "Failed to get descriptor: idx %d addr %p\n",
			       i, vq->desc + i);
1277
			return -EFAULT;
1278 1279 1280 1281 1282
		}
		if (desc.flags & VRING_DESC_F_INDIRECT) {
			ret = get_indirect(dev, vq, iov, iov_size,
					   out_num, in_num,
					   log, log_num, &desc);
1283
			if (unlikely(ret < 0)) {
1284 1285
				vq_err(vq, "Failure detected "
				       "in indirect descriptor at idx %d\n", i);
1286
				return ret;
1287 1288 1289 1290 1291 1292
			}
			continue;
		}

		ret = translate_desc(dev, desc.addr, desc.len, iov + iov_count,
				     iov_size - iov_count);
1293
		if (unlikely(ret < 0)) {
1294 1295
			vq_err(vq, "Translation failure %d descriptor idx %d\n",
			       ret, i);
1296
			return ret;
1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
		}
		if (desc.flags & VRING_DESC_F_WRITE) {
			/* If this is an input descriptor,
			 * increment that count. */
			*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. */
1310
			if (unlikely(*in_num)) {
1311 1312
				vq_err(vq, "Descriptor has out after in: "
				       "idx %d\n", i);
1313
				return -EINVAL;
1314 1315 1316 1317 1318 1319 1320
			}
			*out_num += ret;
		}
	} while ((i = next_desc(&desc)) != -1);

	/* On success, increment avail index. */
	vq->last_avail_idx++;
1321 1322 1323 1324

	/* Assume notifications from guest are disabled at this point,
	 * if they aren't we would need to update avail_event index. */
	BUG_ON(!(vq->used_flags & VRING_USED_F_NO_NOTIFY));
1325 1326 1327 1328
	return head;
}

/* Reverse the effect of vhost_get_vq_desc. Useful for error handling. */
1329
void vhost_discard_vq_desc(struct vhost_virtqueue *vq, int n)
1330
{
1331
	vq->last_avail_idx -= n;
1332 1333 1334 1335 1336 1337
}

/* After we've used one of their buffers, we tell them about it.  We'll then
 * want to notify the guest, using eventfd. */
int vhost_add_used(struct vhost_virtqueue *vq, unsigned int head, int len)
{
1338
	struct vring_used_elem __user *used;
1339 1340 1341 1342

	/* The virtqueue contains a ring of used buffers.  Get a pointer to the
	 * next entry in that used ring. */
	used = &vq->used->ring[vq->last_used_idx % vq->num];
1343
	if (__put_user(head, &used->id)) {
1344 1345 1346
		vq_err(vq, "Failed to write used id");
		return -EFAULT;
	}
1347
	if (__put_user(len, &used->len)) {
1348 1349 1350 1351
		vq_err(vq, "Failed to write used len");
		return -EFAULT;
	}
	/* Make sure buffer is written before we update index. */
1352
	smp_wmb();
1353
	if (__put_user(vq->last_used_idx + 1, &vq->used->idx)) {
1354 1355 1356 1357 1358
		vq_err(vq, "Failed to increment used idx");
		return -EFAULT;
	}
	if (unlikely(vq->log_used)) {
		/* Make sure data is seen before log. */
1359
		smp_wmb();
1360 1361
		/* Log used ring entry write. */
		log_write(vq->log_base,
1362 1363
			  vq->log_addr +
			   ((void __user *)used - (void __user *)vq->used),
1364 1365 1366 1367 1368
			  sizeof *used);
		/* Log used index update. */
		log_write(vq->log_base,
			  vq->log_addr + offsetof(struct vring_used, idx),
			  sizeof vq->used->idx);
1369 1370 1371 1372
		if (vq->log_ctx)
			eventfd_signal(vq->log_ctx, 1);
	}
	vq->last_used_idx++;
1373 1374 1375 1376 1377 1378
	/* If the driver never bothers to signal in a very long while,
	 * used index might wrap around. If that happens, invalidate
	 * signalled_used index we stored. TODO: make sure driver
	 * signals at least once in 2^16 and remove this. */
	if (unlikely(vq->last_used_idx == vq->signalled_used))
		vq->signalled_used_valid = false;
1379 1380 1381
	return 0;
}

1382 1383 1384 1385 1386
static int __vhost_add_used_n(struct vhost_virtqueue *vq,
			    struct vring_used_elem *heads,
			    unsigned count)
{
	struct vring_used_elem __user *used;
1387
	u16 old, new;
1388 1389 1390 1391
	int start;

	start = vq->last_used_idx % vq->num;
	used = vq->used->ring + start;
1392
	if (__copy_to_user(used, heads, count * sizeof *used)) {
1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404
		vq_err(vq, "Failed to write used");
		return -EFAULT;
	}
	if (unlikely(vq->log_used)) {
		/* Make sure data is seen before log. */
		smp_wmb();
		/* Log used ring entry write. */
		log_write(vq->log_base,
			  vq->log_addr +
			   ((void __user *)used - (void __user *)vq->used),
			  count * sizeof *used);
	}
1405 1406 1407 1408 1409 1410 1411 1412
	old = vq->last_used_idx;
	new = (vq->last_used_idx += count);
	/* If the driver never bothers to signal in a very long while,
	 * used index might wrap around. If that happens, invalidate
	 * signalled_used index we stored. TODO: make sure driver
	 * signals at least once in 2^16 and remove this. */
	if (unlikely((u16)(new - vq->signalled_used) < (u16)(new - old)))
		vq->signalled_used_valid = false;
1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
	return 0;
}

/* After we've used one of their buffers, we tell them about it.  We'll then
 * want to notify the guest, using eventfd. */
int vhost_add_used_n(struct vhost_virtqueue *vq, struct vring_used_elem *heads,
		     unsigned count)
{
	int start, n, r;

	start = vq->last_used_idx % vq->num;
	n = vq->num - start;
	if (n < count) {
		r = __vhost_add_used_n(vq, heads, n);
		if (r < 0)
			return r;
		heads += n;
		count -= n;
	}
	r = __vhost_add_used_n(vq, heads, count);

	/* Make sure buffer is written before we update index. */
	smp_wmb();
	if (put_user(vq->last_used_idx, &vq->used->idx)) {
		vq_err(vq, "Failed to increment used idx");
		return -EFAULT;
	}
	if (unlikely(vq->log_used)) {
		/* Log used index update. */
		log_write(vq->log_base,
			  vq->log_addr + offsetof(struct vring_used, idx),
			  sizeof vq->used->idx);
		if (vq->log_ctx)
			eventfd_signal(vq->log_ctx, 1);
	}
	return r;
}

1451
static bool vhost_notify(struct vhost_dev *dev, struct vhost_virtqueue *vq)
1452
{
1453 1454
	__u16 old, new, event;
	bool v;
1455 1456 1457 1458 1459
	/* Flush out used index updates. This is paired
	 * with the barrier that the Guest executes when enabling
	 * interrupts. */
	smp_mb();