fanotify_user.c 20.0 KB
Newer Older
1
#include <linux/fanotify.h>
2
#include <linux/fcntl.h>
3
#include <linux/file.h>
4
#include <linux/fs.h>
5
#include <linux/anon_inodes.h>
6
#include <linux/fsnotify_backend.h>
7
#include <linux/init.h>
E
Eric Paris 已提交
8
#include <linux/mount.h>
9
#include <linux/namei.h>
E
Eric Paris 已提交
10
#include <linux/poll.h>
11 12
#include <linux/security.h>
#include <linux/syscalls.h>
T
Tejun Heo 已提交
13
#include <linux/slab.h>
14
#include <linux/types.h>
E
Eric Paris 已提交
15 16 17
#include <linux/uaccess.h>

#include <asm/ioctls.h>
18

19 20
#define FANOTIFY_DEFAULT_MAX_EVENTS	16384

21
extern const struct fsnotify_ops fanotify_fsnotify_ops;
22

23
static struct kmem_cache *fanotify_mark_cache __read_mostly;
24 25 26 27 28 29 30
static struct kmem_cache *fanotify_response_event_cache __read_mostly;

struct fanotify_response_event {
	struct list_head list;
	__s32 fd;
	struct fsnotify_event *event;
};
31

E
Eric Paris 已提交
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56
/*
 * Get an fsnotify notification event if one exists and is small
 * enough to fit in "count". Return an error pointer if the count
 * is not large enough.
 *
 * Called with the group->notification_mutex held.
 */
static struct fsnotify_event *get_one_event(struct fsnotify_group *group,
					    size_t count)
{
	BUG_ON(!mutex_is_locked(&group->notification_mutex));

	pr_debug("%s: group=%p count=%zd\n", __func__, group, count);

	if (fsnotify_notify_queue_is_empty(group))
		return NULL;

	if (FAN_EVENT_METADATA_LEN > count)
		return ERR_PTR(-EINVAL);

	/* held the notification_mutex the whole time, so this is the
	 * same event we peeked above */
	return fsnotify_remove_notify_event(group);
}

57
static int create_fd(struct fsnotify_group *group, struct fsnotify_event *event)
E
Eric Paris 已提交
58 59 60 61 62 63
{
	int client_fd;
	struct dentry *dentry;
	struct vfsmount *mnt;
	struct file *new_file;

64
	pr_debug("%s: group=%p event=%p\n", __func__, group, event);
E
Eric Paris 已提交
65 66 67 68 69

	client_fd = get_unused_fd();
	if (client_fd < 0)
		return client_fd;

70
	if (event->data_type != FSNOTIFY_EVENT_PATH) {
E
Eric Paris 已提交
71 72 73 74 75 76 77 78 79
		WARN_ON(1);
		put_unused_fd(client_fd);
		return -EINVAL;
	}

	/*
	 * we need a new file handle for the userspace program so it can read even if it was
	 * originally opened O_WRONLY.
	 */
80 81
	dentry = dget(event->path.dentry);
	mnt = mntget(event->path.mnt);
E
Eric Paris 已提交
82 83 84 85
	/* it's possible this event was an overflow event.  in that case dentry and mnt
	 * are NULL;  That's fine, just don't call dentry open */
	if (dentry && mnt)
		new_file = dentry_open(dentry, mnt,
86
				       group->fanotify_data.f_flags | FMODE_NONOTIFY,
E
Eric Paris 已提交
87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103
				       current_cred());
	else
		new_file = ERR_PTR(-EOVERFLOW);
	if (IS_ERR(new_file)) {
		/*
		 * we still send an event even if we can't open the file.  this
		 * can happen when say tasks are gone and we try to open their
		 * /proc files or we try to open a WRONLY file like in sysfs
		 * we just send the errno to userspace since there isn't much
		 * else we can do.
		 */
		put_unused_fd(client_fd);
		client_fd = PTR_ERR(new_file);
	} else {
		fd_install(client_fd, new_file);
	}

104
	return client_fd;
E
Eric Paris 已提交
105 106 107 108 109 110 111 112 113 114 115
}

static ssize_t fill_event_metadata(struct fsnotify_group *group,
				   struct fanotify_event_metadata *metadata,
				   struct fsnotify_event *event)
{
	pr_debug("%s: group=%p metadata=%p event=%p\n", __func__,
		 group, metadata, event);

	metadata->event_len = FAN_EVENT_METADATA_LEN;
	metadata->vers = FANOTIFY_METADATA_VERSION;
116
	metadata->mask = event->mask & FAN_ALL_OUTGOING_EVENTS;
117
	metadata->pid = pid_vnr(event->tgid);
118
	metadata->fd = create_fd(group, event);
E
Eric Paris 已提交
119

120
	return metadata->fd;
E
Eric Paris 已提交
121 122
}

123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199
#ifdef CONFIG_FANOTIFY_ACCESS_PERMISSIONS
static struct fanotify_response_event *dequeue_re(struct fsnotify_group *group,
						  __s32 fd)
{
	struct fanotify_response_event *re, *return_re = NULL;

	mutex_lock(&group->fanotify_data.access_mutex);
	list_for_each_entry(re, &group->fanotify_data.access_list, list) {
		if (re->fd != fd)
			continue;

		list_del_init(&re->list);
		return_re = re;
		break;
	}
	mutex_unlock(&group->fanotify_data.access_mutex);

	pr_debug("%s: found return_re=%p\n", __func__, return_re);

	return return_re;
}

static int process_access_response(struct fsnotify_group *group,
				   struct fanotify_response *response_struct)
{
	struct fanotify_response_event *re;
	__s32 fd = response_struct->fd;
	__u32 response = response_struct->response;

	pr_debug("%s: group=%p fd=%d response=%d\n", __func__, group,
		 fd, response);
	/*
	 * make sure the response is valid, if invalid we do nothing and either
	 * userspace can send a valid responce or we will clean it up after the
	 * timeout
	 */
	switch (response) {
	case FAN_ALLOW:
	case FAN_DENY:
		break;
	default:
		return -EINVAL;
	}

	if (fd < 0)
		return -EINVAL;

	re = dequeue_re(group, fd);
	if (!re)
		return -ENOENT;

	re->event->response = response;

	wake_up(&group->fanotify_data.access_waitq);

	kmem_cache_free(fanotify_response_event_cache, re);

	return 0;
}

static int prepare_for_access_response(struct fsnotify_group *group,
				       struct fsnotify_event *event,
				       __s32 fd)
{
	struct fanotify_response_event *re;

	if (!(event->mask & FAN_ALL_PERM_EVENTS))
		return 0;

	re = kmem_cache_alloc(fanotify_response_event_cache, GFP_KERNEL);
	if (!re)
		return -ENOMEM;

	re->event = event;
	re->fd = fd;

	mutex_lock(&group->fanotify_data.access_mutex);
200 201 202 203 204 205 206 207

	if (group->fanotify_data.bypass_perm) {
		mutex_unlock(&group->fanotify_data.access_mutex);
		kmem_cache_free(fanotify_response_event_cache, re);
		event->response = FAN_ALLOW;
		return 0;
	}
		
208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244
	list_add_tail(&re->list, &group->fanotify_data.access_list);
	mutex_unlock(&group->fanotify_data.access_mutex);

	return 0;
}

static void remove_access_response(struct fsnotify_group *group,
				   struct fsnotify_event *event,
				   __s32 fd)
{
	struct fanotify_response_event *re;

	if (!(event->mask & FAN_ALL_PERM_EVENTS))
		return;

	re = dequeue_re(group, fd);
	if (!re)
		return;

	BUG_ON(re->event != event);

	kmem_cache_free(fanotify_response_event_cache, re);

	return;
}
#else
static int prepare_for_access_response(struct fsnotify_group *group,
				       struct fsnotify_event *event,
				       __s32 fd)
{
	return 0;
}

static void remove_access_response(struct fsnotify_group *group,
				   struct fsnotify_event *event,
				   __s32 fd)
{
245
	return;
246 247 248
}
#endif

E
Eric Paris 已提交
249 250 251 252 253
static ssize_t copy_event_to_user(struct fsnotify_group *group,
				  struct fsnotify_event *event,
				  char __user *buf)
{
	struct fanotify_event_metadata fanotify_event_metadata;
254
	int fd, ret;
E
Eric Paris 已提交
255 256 257

	pr_debug("%s: group=%p event=%p\n", __func__, group, event);

258 259 260 261 262 263 264
	fd = fill_event_metadata(group, &fanotify_event_metadata, event);
	if (fd < 0)
		return fd;

	ret = prepare_for_access_response(group, event, fd);
	if (ret)
		goto out_close_fd;
E
Eric Paris 已提交
265

266
	ret = -EFAULT;
E
Eric Paris 已提交
267
	if (copy_to_user(buf, &fanotify_event_metadata, FAN_EVENT_METADATA_LEN))
268
		goto out_kill_access_response;
E
Eric Paris 已提交
269 270

	return FAN_EVENT_METADATA_LEN;
271 272 273 274 275 276

out_kill_access_response:
	remove_access_response(group, event, fd);
out_close_fd:
	sys_close(fd);
	return ret;
E
Eric Paris 已提交
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346
}

/* intofiy userspace file descriptor functions */
static unsigned int fanotify_poll(struct file *file, poll_table *wait)
{
	struct fsnotify_group *group = file->private_data;
	int ret = 0;

	poll_wait(file, &group->notification_waitq, wait);
	mutex_lock(&group->notification_mutex);
	if (!fsnotify_notify_queue_is_empty(group))
		ret = POLLIN | POLLRDNORM;
	mutex_unlock(&group->notification_mutex);

	return ret;
}

static ssize_t fanotify_read(struct file *file, char __user *buf,
			     size_t count, loff_t *pos)
{
	struct fsnotify_group *group;
	struct fsnotify_event *kevent;
	char __user *start;
	int ret;
	DEFINE_WAIT(wait);

	start = buf;
	group = file->private_data;

	pr_debug("%s: group=%p\n", __func__, group);

	while (1) {
		prepare_to_wait(&group->notification_waitq, &wait, TASK_INTERRUPTIBLE);

		mutex_lock(&group->notification_mutex);
		kevent = get_one_event(group, count);
		mutex_unlock(&group->notification_mutex);

		if (kevent) {
			ret = PTR_ERR(kevent);
			if (IS_ERR(kevent))
				break;
			ret = copy_event_to_user(group, kevent, buf);
			fsnotify_put_event(kevent);
			if (ret < 0)
				break;
			buf += ret;
			count -= ret;
			continue;
		}

		ret = -EAGAIN;
		if (file->f_flags & O_NONBLOCK)
			break;
		ret = -EINTR;
		if (signal_pending(current))
			break;

		if (start != buf)
			break;

		schedule();
	}

	finish_wait(&group->notification_waitq, &wait);
	if (start != buf && ret != -EFAULT)
		ret = buf - start;
	return ret;
}

347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
static ssize_t fanotify_write(struct file *file, const char __user *buf, size_t count, loff_t *pos)
{
#ifdef CONFIG_FANOTIFY_ACCESS_PERMISSIONS
	struct fanotify_response response = { .fd = -1, .response = -1 };
	struct fsnotify_group *group;
	int ret;

	group = file->private_data;

	if (count > sizeof(response))
		count = sizeof(response);

	pr_debug("%s: group=%p count=%zu\n", __func__, group, count);

	if (copy_from_user(&response, buf, count))
		return -EFAULT;

	ret = process_access_response(group, &response);
	if (ret < 0)
		count = ret;

	return count;
#else
	return -EINVAL;
#endif
}

374 375 376
static int fanotify_release(struct inode *ignored, struct file *file)
{
	struct fsnotify_group *group = file->private_data;
377
	struct fanotify_response_event *re, *lre;
378 379 380

	pr_debug("%s: file=%p group=%p\n", __func__, file, group);

381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398
#ifdef CONFIG_FANOTIFY_ACCESS_PERMISSIONS
	mutex_lock(&group->fanotify_data.access_mutex);

	group->fanotify_data.bypass_perm = true;

	list_for_each_entry_safe(re, lre, &group->fanotify_data.access_list, list) {
		pr_debug("%s: found group=%p re=%p event=%p\n", __func__, group,
			 re, re->event);

		list_del_init(&re->list);
		re->event->response = FAN_ALLOW;

		kmem_cache_free(fanotify_response_event_cache, re);
	}
	mutex_unlock(&group->fanotify_data.access_mutex);

	wake_up(&group->fanotify_data.access_waitq);
#endif
399 400 401 402 403 404
	/* matches the fanotify_init->fsnotify_alloc_group */
	fsnotify_put_group(group);

	return 0;
}

E
Eric Paris 已提交
405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429
static long fanotify_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
	struct fsnotify_group *group;
	struct fsnotify_event_holder *holder;
	void __user *p;
	int ret = -ENOTTY;
	size_t send_len = 0;

	group = file->private_data;

	p = (void __user *) arg;

	switch (cmd) {
	case FIONREAD:
		mutex_lock(&group->notification_mutex);
		list_for_each_entry(holder, &group->notification_list, event_list)
			send_len += FAN_EVENT_METADATA_LEN;
		mutex_unlock(&group->notification_mutex);
		ret = put_user(send_len, (int __user *) p);
		break;
	}

	return ret;
}

430
static const struct file_operations fanotify_fops = {
E
Eric Paris 已提交
431 432
	.poll		= fanotify_poll,
	.read		= fanotify_read,
433
	.write		= fanotify_write,
434 435
	.fasync		= NULL,
	.release	= fanotify_release,
E
Eric Paris 已提交
436 437
	.unlocked_ioctl	= fanotify_ioctl,
	.compat_ioctl	= fanotify_ioctl,
438
	.llseek		= noop_llseek,
439 440
};

441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493
static void fanotify_free_mark(struct fsnotify_mark *fsn_mark)
{
	kmem_cache_free(fanotify_mark_cache, fsn_mark);
}

static int fanotify_find_path(int dfd, const char __user *filename,
			      struct path *path, unsigned int flags)
{
	int ret;

	pr_debug("%s: dfd=%d filename=%p flags=%x\n", __func__,
		 dfd, filename, flags);

	if (filename == NULL) {
		struct file *file;
		int fput_needed;

		ret = -EBADF;
		file = fget_light(dfd, &fput_needed);
		if (!file)
			goto out;

		ret = -ENOTDIR;
		if ((flags & FAN_MARK_ONLYDIR) &&
		    !(S_ISDIR(file->f_path.dentry->d_inode->i_mode))) {
			fput_light(file, fput_needed);
			goto out;
		}

		*path = file->f_path;
		path_get(path);
		fput_light(file, fput_needed);
	} else {
		unsigned int lookup_flags = 0;

		if (!(flags & FAN_MARK_DONT_FOLLOW))
			lookup_flags |= LOOKUP_FOLLOW;
		if (flags & FAN_MARK_ONLYDIR)
			lookup_flags |= LOOKUP_DIRECTORY;

		ret = user_path_at(dfd, filename, lookup_flags, path);
		if (ret)
			goto out;
	}

	/* you can only watch an inode if you have read permissions on it */
	ret = inode_permission(path->dentry->d_inode, MAY_READ);
	if (ret)
		path_put(path);
out:
	return ret;
}

494 495 496
static __u32 fanotify_mark_remove_from_mask(struct fsnotify_mark *fsn_mark,
					    __u32 mask,
					    unsigned int flags)
497 498 499 500
{
	__u32 oldmask;

	spin_lock(&fsn_mark->lock);
501 502 503 504 505 506 507
	if (!(flags & FAN_MARK_IGNORED_MASK)) {
		oldmask = fsn_mark->mask;
		fsnotify_set_mark_mask_locked(fsn_mark, (oldmask & ~mask));
	} else {
		oldmask = fsn_mark->ignored_mask;
		fsnotify_set_mark_ignored_mask_locked(fsn_mark, (oldmask & ~mask));
	}
508 509 510 511 512 513 514 515
	spin_unlock(&fsn_mark->lock);

	if (!(oldmask & ~mask))
		fsnotify_destroy_mark(fsn_mark);

	return mask & oldmask;
}

516
static int fanotify_remove_vfsmount_mark(struct fsnotify_group *group,
517 518
					 struct vfsmount *mnt, __u32 mask,
					 unsigned int flags)
519 520
{
	struct fsnotify_mark *fsn_mark = NULL;
521
	__u32 removed;
522

523 524 525
	fsn_mark = fsnotify_find_vfsmount_mark(group, mnt);
	if (!fsn_mark)
		return -ENOENT;
526

527
	removed = fanotify_mark_remove_from_mask(fsn_mark, mask, flags);
528 529 530 531 532 533
	fsnotify_put_mark(fsn_mark);
	if (removed & mnt->mnt_fsnotify_mask)
		fsnotify_recalc_vfsmount_mask(mnt);

	return 0;
}
534

535
static int fanotify_remove_inode_mark(struct fsnotify_group *group,
536 537
				      struct inode *inode, __u32 mask,
				      unsigned int flags)
538 539 540 541 542
{
	struct fsnotify_mark *fsn_mark = NULL;
	__u32 removed;

	fsn_mark = fsnotify_find_inode_mark(group, inode);
543 544 545
	if (!fsn_mark)
		return -ENOENT;

546
	removed = fanotify_mark_remove_from_mask(fsn_mark, mask, flags);
547
	/* matches the fsnotify_find_inode_mark() */
548
	fsnotify_put_mark(fsn_mark);
549 550
	if (removed & inode->i_fsnotify_mask)
		fsnotify_recalc_inode_mask(inode);
551

552 553 554
	return 0;
}

555 556 557
static __u32 fanotify_mark_add_to_mask(struct fsnotify_mark *fsn_mark,
				       __u32 mask,
				       unsigned int flags)
558 559 560 561
{
	__u32 oldmask;

	spin_lock(&fsn_mark->lock);
562 563 564 565 566 567
	if (!(flags & FAN_MARK_IGNORED_MASK)) {
		oldmask = fsn_mark->mask;
		fsnotify_set_mark_mask_locked(fsn_mark, (oldmask | mask));
	} else {
		oldmask = fsn_mark->ignored_mask;
		fsnotify_set_mark_ignored_mask_locked(fsn_mark, (oldmask | mask));
568 569
		if (flags & FAN_MARK_IGNORED_SURV_MODIFY)
			fsn_mark->flags |= FSNOTIFY_MARK_FLAG_IGNORED_SURV_MODIFY;
570
	}
571 572 573 574 575
	spin_unlock(&fsn_mark->lock);

	return mask & ~oldmask;
}

576
static int fanotify_add_vfsmount_mark(struct fsnotify_group *group,
577 578
				      struct vfsmount *mnt, __u32 mask,
				      unsigned int flags)
579 580
{
	struct fsnotify_mark *fsn_mark;
581
	__u32 added;
582

583 584 585 586
	fsn_mark = fsnotify_find_vfsmount_mark(group, mnt);
	if (!fsn_mark) {
		int ret;

587 588
		fsn_mark = kmem_cache_alloc(fanotify_mark_cache, GFP_KERNEL);
		if (!fsn_mark)
589
			return -ENOMEM;
590

591 592
		fsnotify_init_mark(fsn_mark, fanotify_free_mark);
		ret = fsnotify_add_mark(fsn_mark, group, NULL, mnt, 0);
593
		if (ret) {
594
			fanotify_free_mark(fsn_mark);
595
			return ret;
596 597
		}
	}
598
	added = fanotify_mark_add_to_mask(fsn_mark, mask, flags);
599
	fsnotify_put_mark(fsn_mark);
E
Eric Paris 已提交
600 601 602
	if (added & ~mnt->mnt_fsnotify_mask)
		fsnotify_recalc_vfsmount_mask(mnt);

603
	return 0;
604 605
}

606
static int fanotify_add_inode_mark(struct fsnotify_group *group,
607 608
				   struct inode *inode, __u32 mask,
				   unsigned int flags)
609 610
{
	struct fsnotify_mark *fsn_mark;
611
	__u32 added;
612 613

	pr_debug("%s: group=%p inode=%p\n", __func__, group, inode);
614

615 616 617 618 619 620 621 622 623 624
	/*
	 * If some other task has this inode open for write we should not add
	 * an ignored mark, unless that ignored mark is supposed to survive
	 * modification changes anyway.
	 */
	if ((flags & FAN_MARK_IGNORED_MASK) &&
	    !(flags & FAN_MARK_IGNORED_SURV_MODIFY) &&
	    (atomic_read(&inode->i_writecount) > 0))
		return 0;

625
	fsn_mark = fsnotify_find_inode_mark(group, inode);
626
	if (!fsn_mark) {
627
		int ret;
628

629 630
		fsn_mark = kmem_cache_alloc(fanotify_mark_cache, GFP_KERNEL);
		if (!fsn_mark)
631
			return -ENOMEM;
632

633 634
		fsnotify_init_mark(fsn_mark, fanotify_free_mark);
		ret = fsnotify_add_mark(fsn_mark, group, inode, NULL, 0);
635
		if (ret) {
636
			fanotify_free_mark(fsn_mark);
637
			return ret;
638 639
		}
	}
640
	added = fanotify_mark_add_to_mask(fsn_mark, mask, flags);
641
	fsnotify_put_mark(fsn_mark);
E
Eric Paris 已提交
642 643
	if (added & ~inode->i_fsnotify_mask)
		fsnotify_recalc_inode_mask(inode);
644
	return 0;
645
}
646

647
/* fanotify syscalls */
648
SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags)
649
{
650 651 652
	struct fsnotify_group *group;
	int f_flags, fd;

653 654
	pr_debug("%s: flags=%d event_f_flags=%d\n",
		__func__, flags, event_f_flags);
655 656

	if (!capable(CAP_SYS_ADMIN))
657
		return -EPERM;
658 659 660 661

	if (flags & ~FAN_ALL_INIT_FLAGS)
		return -EINVAL;

662
	f_flags = O_RDWR | FMODE_NONOTIFY;
663 664 665 666 667 668 669 670 671 672
	if (flags & FAN_CLOEXEC)
		f_flags |= O_CLOEXEC;
	if (flags & FAN_NONBLOCK)
		f_flags |= O_NONBLOCK;

	/* fsnotify_alloc_group takes a ref.  Dropped in fanotify_release */
	group = fsnotify_alloc_group(&fanotify_fsnotify_ops);
	if (IS_ERR(group))
		return PTR_ERR(group);

673
	group->fanotify_data.f_flags = event_f_flags;
E
Eric Paris 已提交
674 675 676 677 678
#ifdef CONFIG_FANOTIFY_ACCESS_PERMISSIONS
	mutex_init(&group->fanotify_data.access_mutex);
	init_waitqueue_head(&group->fanotify_data.access_waitq);
	INIT_LIST_HEAD(&group->fanotify_data.access_list);
#endif
679 680 681 682 683 684 685 686 687 688 689 690 691 692
	switch (flags & FAN_ALL_CLASS_BITS) {
	case FAN_CLASS_NOTIF:
		group->priority = FS_PRIO_0;
		break;
	case FAN_CLASS_CONTENT:
		group->priority = FS_PRIO_1;
		break;
	case FAN_CLASS_PRE_CONTENT:
		group->priority = FS_PRIO_2;
		break;
	default:
		fd = -EINVAL;
		goto out_put_group;
	}
E
Eric Paris 已提交
693

694 695 696 697 698 699 700 701
	if (flags & FAN_UNLIMITED_QUEUE) {
		fd = -EPERM;
		if (!capable(CAP_SYS_ADMIN))
			goto out_put_group;
		group->max_events = UINT_MAX;
	} else {
		group->max_events = FANOTIFY_DEFAULT_MAX_EVENTS;
	}
702

703 704 705 706 707 708 709 710 711
	fd = anon_inode_getfd("[fanotify]", &fanotify_fops, group, f_flags);
	if (fd < 0)
		goto out_put_group;

	return fd;

out_put_group:
	fsnotify_put_group(group);
	return fd;
712
}
713

714 715 716
SYSCALL_DEFINE(fanotify_mark)(int fanotify_fd, unsigned int flags,
			      __u64 mask, int dfd,
			      const char  __user * pathname)
717
{
718 719
	struct inode *inode = NULL;
	struct vfsmount *mnt = NULL;
720 721 722 723 724 725 726 727 728 729 730 731
	struct fsnotify_group *group;
	struct file *filp;
	struct path path;
	int ret, fput_needed;

	pr_debug("%s: fanotify_fd=%d flags=%x dfd=%d pathname=%p mask=%llx\n",
		 __func__, fanotify_fd, flags, dfd, pathname, mask);

	/* we only use the lower 32 bits as of right now. */
	if (mask & ((__u64)0xffffffff << 32))
		return -EINVAL;

732 733
	if (flags & ~FAN_ALL_MARK_FLAGS)
		return -EINVAL;
E
Eric Paris 已提交
734
	switch (flags & (FAN_MARK_ADD | FAN_MARK_REMOVE | FAN_MARK_FLUSH)) {
735 736
	case FAN_MARK_ADD:
	case FAN_MARK_REMOVE:
E
Eric Paris 已提交
737
	case FAN_MARK_FLUSH:
738 739 740 741
		break;
	default:
		return -EINVAL;
	}
742 743 744
#ifdef CONFIG_FANOTIFY_ACCESS_PERMISSIONS
	if (mask & ~(FAN_ALL_EVENTS | FAN_ALL_PERM_EVENTS | FAN_EVENT_ON_CHILD))
#else
745
	if (mask & ~(FAN_ALL_EVENTS | FAN_EVENT_ON_CHILD))
746
#endif
747 748 749 750 751 752 753 754 755 756
		return -EINVAL;

	filp = fget_light(fanotify_fd, &fput_needed);
	if (unlikely(!filp))
		return -EBADF;

	/* verify that this is indeed an fanotify instance */
	ret = -EINVAL;
	if (unlikely(filp->f_op != &fanotify_fops))
		goto fput_and_out;
757 758 759 760 761 762 763 764 765 766
	group = filp->private_data;

	/*
	 * group->priority == FS_PRIO_0 == FAN_CLASS_NOTIF.  These are not
	 * allowed to set permissions events.
	 */
	ret = -EINVAL;
	if (mask & FAN_ALL_PERM_EVENTS &&
	    group->priority == FS_PRIO_0)
		goto fput_and_out;
767 768 769 770 771 772

	ret = fanotify_find_path(dfd, pathname, &path, flags);
	if (ret)
		goto fput_and_out;

	/* inode held in place by reference to path; group by fget on fd */
773
	if (!(flags & FAN_MARK_MOUNT))
774 775 776
		inode = path.dentry->d_inode;
	else
		mnt = path.mnt;
777 778

	/* create/update an inode mark */
E
Eric Paris 已提交
779
	switch (flags & (FAN_MARK_ADD | FAN_MARK_REMOVE | FAN_MARK_FLUSH)) {
780
	case FAN_MARK_ADD:
781
		if (flags & FAN_MARK_MOUNT)
782
			ret = fanotify_add_vfsmount_mark(group, mnt, mask, flags);
783
		else
784
			ret = fanotify_add_inode_mark(group, inode, mask, flags);
785 786
		break;
	case FAN_MARK_REMOVE:
787
		if (flags & FAN_MARK_MOUNT)
788
			ret = fanotify_remove_vfsmount_mark(group, mnt, mask, flags);
789
		else
790
			ret = fanotify_remove_inode_mark(group, inode, mask, flags);
791
		break;
E
Eric Paris 已提交
792 793 794 795 796 797
	case FAN_MARK_FLUSH:
		if (flags & FAN_MARK_MOUNT)
			fsnotify_clear_vfsmount_marks_by_group(group);
		else
			fsnotify_clear_inode_marks_by_group(group);
		break;
798 799 800
	default:
		ret = -EINVAL;
	}
801 802 803 804 805 806 807

	path_put(&path);
fput_and_out:
	fput_light(filp, fput_needed);
	return ret;
}

808 809 810 811 812 813 814 815 816 817 818
#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
asmlinkage long SyS_fanotify_mark(long fanotify_fd, long flags, __u64 mask,
				  long dfd, long pathname)
{
	return SYSC_fanotify_mark((int) fanotify_fd, (unsigned int) flags,
				  mask, (int) dfd,
				  (const char  __user *) pathname);
}
SYSCALL_ALIAS(sys_fanotify_mark, SyS_fanotify_mark);
#endif

819 820 821 822 823 824 825 826
/*
 * fanotify_user_setup - Our initialization function.  Note that we cannnot return
 * error because we have compiled-in VFS hooks.  So an (unlikely) failure here
 * must result in panic().
 */
static int __init fanotify_user_setup(void)
{
	fanotify_mark_cache = KMEM_CACHE(fsnotify_mark, SLAB_PANIC);
827 828
	fanotify_response_event_cache = KMEM_CACHE(fanotify_response_event,
						   SLAB_PANIC);
829 830

	return 0;
831
}
832
device_initcall(fanotify_user_setup);