All new accounts created on Gitlab now require administrator approval. If you invite any collaborators, please let Flux staff know so they can approve the accounts.

vhost.c 40.7 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
void vhost_work_init(struct vhost_work *work, vhost_work_fn_t fn)
68 69 70 71 72 73 74 75
{
	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
/* 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);
}

140
void vhost_work_queue(struct vhost_dev *dev, struct vhost_work *work)
141
{
142 143 144 145 146 147 148 149 150
	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);
151 152
}

153 154 155 156 157
void vhost_poll_queue(struct vhost_poll *poll)
{
	vhost_work_queue(poll->dev, &poll->work);
}

158 159 160 161 162 163 164 165 166 167
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;
168 169
	vq->signalled_used = 0;
	vq->signalled_used_valid = false;
170 171 172
	vq->used_flags = 0;
	vq->log_used = false;
	vq->log_addr = -1ull;
173 174
	vq->vhost_hlen = 0;
	vq->sock_hlen = 0;
175 176 177 178 179 180 181
	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;
182
	vq->log_ctx = NULL;
183 184 185
	vq->upend_idx = 0;
	vq->done_idx = 0;
	vq->ubufs = NULL;
186 187
}

188 189 190 191 192
static int vhost_worker(void *data)
{
	struct vhost_dev *dev = data;
	struct vhost_work *work = NULL;
	unsigned uninitialized_var(seq);
193
	mm_segment_t oldfs = get_fs();
194

195
	set_fs(USER_DS);
196 197
	use_mm(dev->mm);

198 199 200 201 202 203 204 205 206 207 208 209 210 211
	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);
212
			break;
213 214 215 216 217 218 219 220 221 222 223 224 225
		}
		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);
226 227
			if (need_resched())
				schedule();
228 229 230 231
		} else
			schedule();

	}
232
	unuse_mm(dev->mm);
233
	set_fs(oldfs);
234
	return 0;
235 236
}

237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253
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
254 255 256 257
/* Helper to allocate iovec buffers for all vqs. */
static long vhost_dev_alloc_iovecs(struct vhost_dev *dev)
{
	int i;
258
	bool zcopy;
259

Jason Wang's avatar
Jason Wang committed
260 261 262 263 264 265 266
	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);
267 268 269 270 271
		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
272
		if (!dev->vqs[i].indirect || !dev->vqs[i].log ||
273 274
			!dev->vqs[i].heads ||
			(zcopy && !dev->vqs[i].ubuf_info))
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:
280 281
	for (; i >= 0; --i)
		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

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

293 294 295 296
long vhost_dev_init(struct vhost_dev *dev,
		    struct vhost_virtqueue *vqs, int nvqs)
{
	int i;
297

298 299 300 301 302 303 304
	dev->vqs = vqs;
	dev->nvqs = nvqs;
	mutex_init(&dev->mutex);
	dev->log_ctx = NULL;
	dev->log_file = NULL;
	dev->memory = NULL;
	dev->mm = NULL;
305 306 307
	spin_lock_init(&dev->work_lock);
	INIT_LIST_HEAD(&dev->work_list);
	dev->worker = NULL;
308 309

	for (i = 0; i < dev->nvqs; ++i) {
Jason Wang's avatar
Jason Wang committed
310 311 312
		dev->vqs[i].log = NULL;
		dev->vqs[i].indirect = NULL;
		dev->vqs[i].heads = NULL;
313
		dev->vqs[i].ubuf_info = NULL;
314 315 316 317 318
		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,
319
					dev->vqs[i].handle_kick, POLLIN, dev);
320
	}
321

322 323 324 325 326 327 328 329 330 331
	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;
}

332
struct vhost_attach_cgroups_struct {
333 334 335
	struct vhost_work work;
	struct task_struct *owner;
	int ret;
336 337 338 339
};

static void vhost_attach_cgroups_work(struct vhost_work *work)
{
340 341 342 343
	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);
344 345 346 347
}

static int vhost_attach_cgroups(struct vhost_dev *dev)
{
348 349 350 351 352 353 354
	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;
355 356
}

357 358 359
/* Caller should have device mutex */
static long vhost_dev_set_owner(struct vhost_dev *dev)
{
360 361
	struct task_struct *worker;
	int err;
362

363
	/* Is there an owner already? */
364 365 366 367
	if (dev->mm) {
		err = -EBUSY;
		goto err_mm;
	}
368

369 370
	/* No owner, become one */
	dev->mm = get_task_mm(current);
371 372 373 374 375 376 377
	worker = kthread_create(vhost_worker, dev, "vhost-%d", current->pid);
	if (IS_ERR(worker)) {
		err = PTR_ERR(worker);
		goto err_worker;
	}

	dev->worker = worker;
378 379 380
	wake_up_process(worker);	/* avoid contributing to loadavg */

	err = vhost_attach_cgroups(dev);
381 382
	if (err)
		goto err_cgroup;
383

Jason Wang's avatar
Jason Wang committed
384 385 386 387
	err = vhost_dev_alloc_iovecs(dev);
	if (err)
		goto err_cgroup;

388
	return 0;
389 390
err_cgroup:
	kthread_stop(worker);
Michael S. Tsirkin's avatar
Michael S. Tsirkin committed
391
	dev->worker = NULL;
392 393 394 395 396 397
err_worker:
	if (dev->mm)
		mmput(dev->mm);
	dev->mm = NULL;
err_mm:
	return err;
398 399 400 401 402 403 404 405 406 407 408 409
}

/* 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;

410
	vhost_dev_cleanup(dev, true);
411 412

	memory->nregions = 0;
Arnd Bergmann's avatar
Arnd Bergmann committed
413
	RCU_INIT_POINTER(dev->memory, memory);
414 415 416
	return 0;
}

417 418 419 420 421 422 423 424 425 426 427
/* 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) {
428
		if (VHOST_DMA_IS_DONE(vq->heads[i].len)) {
429 430 431 432 433 434 435 436 437 438 439 440
			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;
}

441 442
/* Caller should have device mutex if and only if locked is set */
void vhost_dev_cleanup(struct vhost_dev *dev, bool locked)
443 444
{
	int i;
445

446 447 448 449 450
	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);
		}
451 452 453 454 455 456 457
		/* 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]);

458 459 460 461 462 463 464 465 466 467 468 469
		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
470
	vhost_dev_free_iovecs(dev);
471 472 473 474 475 476 477
	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
478
	kfree(rcu_dereference_protected(dev->memory,
479 480
					locked ==
						lockdep_is_held(&dev->mutex)));
Arnd Bergmann's avatar
Arnd Bergmann committed
481
	RCU_INIT_POINTER(dev->memory, NULL);
482
	WARN_ON(!list_empty(&dev->work_list));
483 484 485 486
	if (dev->worker) {
		kthread_stop(dev->worker);
		dev->worker = NULL;
	}
487 488 489
	if (dev->mm)
		mmput(dev->mm);
	dev->mm = NULL;
490 491 492 493 494
}

static int log_access_ok(void __user *log_base, u64 addr, unsigned long sz)
{
	u64 a = addr / VHOST_PAGE_SIZE / 8;
495

496 497 498
	/* Make sure 64 bit math will not overflow. */
	if (a > ULONG_MAX - (unsigned long)log_base ||
	    a + (unsigned long)log_base > ULONG_MAX)
499
		return 0;
500 501 502 503 504 505 506 507 508 509

	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;
510

Michael S. Tsirkin's avatar
Michael S. Tsirkin committed
511 512
	if (!mem)
		return 0;
513

514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535
	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;
536

537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552
	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;
}

553
static int vq_access_ok(struct vhost_dev *d, unsigned int num,
554 555 556 557
			struct vring_desc __user *desc,
			struct vring_avail __user *avail,
			struct vring_used __user *used)
{
558
	size_t s = vhost_has_feature(d, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0;
559 560
	return access_ok(VERIFY_READ, desc, num * sizeof *desc) &&
	       access_ok(VERIFY_READ, avail,
561
			 sizeof *avail + num * sizeof *avail->ring + s) &&
562
	       access_ok(VERIFY_WRITE, used,
563
			sizeof *used + num * sizeof *used->ring + s);
564 565 566 567 568 569
}

/* 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
570 571 572 573 574
	struct vhost_memory *mp;

	mp = rcu_dereference_protected(dev->memory,
				       lockdep_is_held(&dev->mutex));
	return memory_access_ok(dev, mp, 1);
575 576 577 578
}

/* Verify access for write logging. */
/* Caller should have vq mutex and device mutex */
579 580
static int vq_log_access_ok(struct vhost_dev *d, struct vhost_virtqueue *vq,
			    void __user *log_base)
581
{
Arnd Bergmann's avatar
Arnd Bergmann committed
582
	struct vhost_memory *mp;
583
	size_t s = vhost_has_feature(d, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0;
Arnd Bergmann's avatar
Arnd Bergmann committed
584 585 586 587

	mp = rcu_dereference_protected(vq->dev->memory,
				       lockdep_is_held(&vq->mutex));
	return vq_memory_access_ok(log_base, mp,
588 589 590
			    vhost_has_feature(vq->dev, VHOST_F_LOG_ALL)) &&
		(!vq->log_used || log_access_ok(log_base, vq->log_addr,
					sizeof *vq->used +
591
					vq->num * sizeof *vq->used->ring + s));
592 593 594 595 596 597
}

/* Can we start vq? */
/* Caller should have vq mutex and device mutex */
int vhost_vq_access_ok(struct vhost_virtqueue *vq)
{
598 599
	return vq_access_ok(vq->dev, vq->num, vq->desc, vq->avail, vq->used) &&
		vq_log_access_ok(vq->dev, vq, vq->log_base);
600 601 602 603 604 605
}

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);
606

607 608
	if (copy_from_user(&mem, m, size))
		return -EFAULT;
609 610 611 612 613 614 615 616 617
	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);
618 619
	if (copy_from_user(newmem->regions, m->regions,
			   mem.nregions * sizeof *m->regions)) {
620
		kfree(newmem);
621
		return -EFAULT;
622 623
	}

624 625
	if (!memory_access_ok(d, newmem,
			      vhost_has_feature(d, VHOST_F_LOG_ALL))) {
626
		kfree(newmem);
627
		return -EFAULT;
628
	}
Arnd Bergmann's avatar
Arnd Bergmann committed
629 630
	oldmem = rcu_dereference_protected(d->memory,
					   lockdep_is_held(&d->mutex));
631 632 633 634 635 636 637 638
	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)
{
639 640
	struct file *eventfp, *filep = NULL;
	bool pollstart = false, pollstop = false;
641 642 643 644 645 646 647 648 649 650 651 652
	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;
653
	if (idx >= d->nvqs)
654 655 656 657 658 659 660 661 662 663 664 665 666 667
		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;
		}
668 669
		if (copy_from_user(&s, argp, sizeof s)) {
			r = -EFAULT;
670
			break;
671
		}
672 673 674 675 676 677 678 679 680 681 682 683 684
		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;
		}
685 686
		if (copy_from_user(&s, argp, sizeof s)) {
			r = -EFAULT;
687
			break;
688
		}
689 690 691 692 693 694 695 696 697 698 699
		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;
700 701
		if (copy_to_user(argp, &s, sizeof s))
			r = -EFAULT;
702 703
		break;
	case VHOST_SET_VRING_ADDR:
704 705
		if (copy_from_user(&a, argp, sizeof a)) {
			r = -EFAULT;
706
			break;
707
		}
708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730
		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) {
731
			if (!vq_access_ok(d, vq->num,
732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755
				(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:
756 757
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
758
			break;
759
		}
760
		eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
761 762 763 764
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
765
		if (eventfp != vq->kick) {
766 767
			pollstop = (filep = vq->kick) != NULL;
			pollstart = (vq->kick = eventfp) != NULL;
768 769 770 771
		} else
			filep = eventfp;
		break;
	case VHOST_SET_VRING_CALL:
772 773
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
774
			break;
775
		}
776
		eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
777 778 779 780
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
781 782 783 784 785 786 787 788 789 790
		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:
791 792
		if (copy_from_user(&f, argp, sizeof f)) {
			r = -EFAULT;
793
			break;
794
		}
795
		eventfp = f.fd == -1 ? NULL : eventfd_fget(f.fd);
796 797 798 799
		if (IS_ERR(eventfp)) {
			r = PTR_ERR(eventfp);
			break;
		}
800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856
		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:
857 858
		if (copy_from_user(&p, argp, sizeof p)) {
			r = -EFAULT;
859
			break;
860
		}
861 862 863 864 865 866 867 868 869 870
		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. */
871
			if (vq->private_data && !vq_log_access_ok(d, vq, base))
872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916
				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;
917

918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939
	/* 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;
940

941
	r = get_user_pages_fast(log, 1, 1, &page);
942
	if (r < 0)
943
		return r;
944
	BUG_ON(r != 1);
945
	base = kmap_atomic(page);
946
	set_bit(bit, base);
947
	kunmap_atomic(base);
948 949 950 951 952 953 954 955
	set_page_dirty_lock(page);
	put_page(page);
	return 0;
}

static int log_write(void __user *log_base,
		     u64 write_address, u64 write_length)
{
956
	u64 write_page = write_address / VHOST_PAGE_SIZE;
957
	int r;
958

959 960
	if (!write_length)
		return 0;
961
	write_length += write_address % VHOST_PAGE_SIZE;
962 963
	for (;;) {
		u64 base = (u64)(unsigned long)log_base;
964 965
		u64 log = base + write_page / 8;
		int bit = write_page % 8;
966 967 968 969 970 971 972 973
		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;
974
		write_page += 1;
975 976 977 978 979 980 981 982 983 984
	}
	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. */
985
	smp_wmb();
986 987 988 989 990 991
	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;
992 993 994
		if (!len) {
			if (vq->log_ctx)
				eventfd_signal(vq->log_ctx, 1);
995
			return 0;
996
		}
997 998 999 1000 1001 1002
	}
	/* Length written exceeds what we have stored. This is a bug. */
	BUG();
	return 0;
}

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 1053
		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);
}

Christoph Hellwig's avatar
Christoph Hellwig committed
1054 1055
static int translate_desc(struct vhost_dev *dev, u64 addr, u32 len,
			  struct iovec iov[], int iov_size)
1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
{
	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;
1068
		if (unlikely(ret >= iov_size)) {
1069 1070 1071 1072
			ret = -ENOBUFS;
			break;
		}
		reg = find_region(mem, addr, len);
1073
		if (unlikely(!reg)) {
1074 1075 1076 1077 1078 1079
			ret = -EFAULT;
			break;
		}
		_iov = iov + ret;
		size = reg->memory_size - addr + reg->guest_phys_addr;
		_iov->iov_len = min((u64)len, size);
Christoph Hellwig's avatar
Christoph Hellwig committed
1080
		_iov->iov_base = (void __user *)(unsigned long)
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 1110 1111
			(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;
}

1112 1113 1114 1115 1116
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)
1117 1118 1119 1120 1121 1122
{
	struct vring_desc desc;
	unsigned int i = 0, count, found = 0;
	int ret;

	/* Sanity check */
1123
	if (unlikely(indirect->len % sizeof desc)) {
1124 1125 1126 1127 1128 1129 1130 1131
		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
1132
			     UIO_MAXIOV);
1133
	if (unlikely(ret < 0)) {
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144
		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. */
1145
	if (unlikely(count > USHRT_MAX + 1)) {
1146 1147 1148 1149 1150 1151 1152
		vq_err(vq, "Indirect buffer length too big: %d\n",
		       indirect->len);
		return -E2BIG;
	}

	do {
		unsigned iov_count = *in_num + *out_num;
1153
		if (unlikely(++found > count)) {
1154 1155 1156 1157 1158
			vq_err(vq, "Loop detected: last one at %u "
			       "indirect size %u\n",
			       i, count);
			return -EINVAL;
		}
1159 1160
		if (unlikely(memcpy_fromiovec((unsigned char *)&desc,
					      vq->indirect, sizeof desc))) {
1161 1162 1163 1164
			vq_err(vq, "Failed indirect descriptor: idx %d, %zx\n",
			       i, (size_t)indirect->addr + i * sizeof desc);
			return -EINVAL;
		}
1165
		if (unlikely(desc.flags & VRING_DESC_F_INDIRECT)) {
1166 1167 1168 1169 1170 1171 1172
			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);
1173
		if (unlikely(ret < 0)) {
1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188
			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. */
1189
			if (unlikely(*in_num)) {
1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
				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.
 *
1205 1206 1207 1208 1209 1210 1211
 * 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)
1212 1213 1214 1215 1216 1217 1218 1219
{
	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;
1220
	if (unlikely(__get_user(vq->avail_idx, &vq->avail->idx))) {
1221 1222
		vq_err(vq, "Failed to access avail idx at %p\n",
		       &vq->avail->idx);
1223
		return -EFAULT;
1224 1225
	}

1226
	if (unlikely((u16)(vq->avail_idx - last_avail_idx) > vq->num)) {
1227 1228
		vq_err(vq, "Guest moved used index from %u to %u",
		       last_avail_idx, vq->avail_idx);
1229
		return -EFAULT;
1230 1231 1232 1233 1234 1235 1236
	}

	/* 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. */
1237
	smp_rmb();
1238 1239 1240

	/* Grab the next descriptor number they're advertising, and increment
	 * the index we've seen. */
1241 1242
	if (unlikely(__get_user(head,
				&vq->avail->ring[last_avail_idx % vq->num]))) {
1243 1244 1245
		vq_err(vq, "Failed to read head: idx %d address %p\n",
		       last_avail_idx,
		       &vq->avail->ring[last_avail_idx % vq->num]);
1246
		return -EFAULT;
1247 1248 1249
	}

	/* If their number is silly, that's an error. */
1250
	if (unlikely(head >= vq->num)) {
1251 1252
		vq_err(vq, "Guest says index %u > %u is available",
		       head, vq->num);
1253
		return -EINVAL;
1254 1255 1256 1257 1258 1259 1260 1261 1262 1263
	}

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