locks.c 68.3 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 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 57 58 59 60 61 62
/*
 *  linux/fs/locks.c
 *
 *  Provide support for fcntl()'s F_GETLK, F_SETLK, and F_SETLKW calls.
 *  Doug Evans (dje@spiff.uucp), August 07, 1992
 *
 *  Deadlock detection added.
 *  FIXME: one thing isn't handled yet:
 *	- mandatory locks (requires lots of changes elsewhere)
 *  Kelly Carmichael (kelly@[142.24.8.65]), September 17, 1994.
 *
 *  Miscellaneous edits, and a total rewrite of posix_lock_file() code.
 *  Kai Petzke (wpp@marie.physik.tu-berlin.de), 1994
 *  
 *  Converted file_lock_table to a linked list from an array, which eliminates
 *  the limits on how many active file locks are open.
 *  Chad Page (pageone@netcom.com), November 27, 1994
 * 
 *  Removed dependency on file descriptors. dup()'ed file descriptors now
 *  get the same locks as the original file descriptors, and a close() on
 *  any file descriptor removes ALL the locks on the file for the current
 *  process. Since locks still depend on the process id, locks are inherited
 *  after an exec() but not after a fork(). This agrees with POSIX, and both
 *  BSD and SVR4 practice.
 *  Andy Walker (andy@lysaker.kvaerner.no), February 14, 1995
 *
 *  Scrapped free list which is redundant now that we allocate locks
 *  dynamically with kmalloc()/kfree().
 *  Andy Walker (andy@lysaker.kvaerner.no), February 21, 1995
 *
 *  Implemented two lock personalities - FL_FLOCK and FL_POSIX.
 *
 *  FL_POSIX locks are created with calls to fcntl() and lockf() through the
 *  fcntl() system call. They have the semantics described above.
 *
 *  FL_FLOCK locks are created with calls to flock(), through the flock()
 *  system call, which is new. Old C libraries implement flock() via fcntl()
 *  and will continue to use the old, broken implementation.
 *
 *  FL_FLOCK locks follow the 4.4 BSD flock() semantics. They are associated
 *  with a file pointer (filp). As a result they can be shared by a parent
 *  process and its children after a fork(). They are removed when the last
 *  file descriptor referring to the file pointer is closed (unless explicitly
 *  unlocked). 
 *
 *  FL_FLOCK locks never deadlock, an existing lock is always removed before
 *  upgrading from shared to exclusive (or vice versa). When this happens
 *  any processes blocked by the current lock are woken up and allowed to
 *  run before the new lock is applied.
 *  Andy Walker (andy@lysaker.kvaerner.no), June 09, 1995
 *
 *  Removed some race conditions in flock_lock_file(), marked other possible
 *  races. Just grep for FIXME to see them. 
 *  Dmitry Gorodchanin (pgmdsg@ibi.com), February 09, 1996.
 *
 *  Addressed Dmitry's concerns. Deadlock checking no longer recursive.
 *  Lock allocation changed to GFP_ATOMIC as we can't afford to sleep
 *  once we've checked for blocking and deadlocking.
 *  Andy Walker (andy@lysaker.kvaerner.no), April 03, 1996.
 *
 *  Initial implementation of mandatory locks. SunOS turned out to be
 *  a rotten model, so I implemented the "obvious" semantics.
P
Paul Bolle 已提交
63
 *  See 'Documentation/filesystems/mandatory-locking.txt' for details.
L
Linus Torvalds 已提交
64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118
 *  Andy Walker (andy@lysaker.kvaerner.no), April 06, 1996.
 *
 *  Don't allow mandatory locks on mmap()'ed files. Added simple functions to
 *  check if a file has mandatory locks, used by mmap(), open() and creat() to
 *  see if system call should be rejected. Ref. HP-UX/SunOS/Solaris Reference
 *  Manual, Section 2.
 *  Andy Walker (andy@lysaker.kvaerner.no), April 09, 1996.
 *
 *  Tidied up block list handling. Added '/proc/locks' interface.
 *  Andy Walker (andy@lysaker.kvaerner.no), April 24, 1996.
 *
 *  Fixed deadlock condition for pathological code that mixes calls to
 *  flock() and fcntl().
 *  Andy Walker (andy@lysaker.kvaerner.no), April 29, 1996.
 *
 *  Allow only one type of locking scheme (FL_POSIX or FL_FLOCK) to be in use
 *  for a given file at a time. Changed the CONFIG_LOCK_MANDATORY scheme to
 *  guarantee sensible behaviour in the case where file system modules might
 *  be compiled with different options than the kernel itself.
 *  Andy Walker (andy@lysaker.kvaerner.no), May 15, 1996.
 *
 *  Added a couple of missing wake_up() calls. Thanks to Thomas Meckel
 *  (Thomas.Meckel@mni.fh-giessen.de) for spotting this.
 *  Andy Walker (andy@lysaker.kvaerner.no), May 15, 1996.
 *
 *  Changed FL_POSIX locks to use the block list in the same way as FL_FLOCK
 *  locks. Changed process synchronisation to avoid dereferencing locks that
 *  have already been freed.
 *  Andy Walker (andy@lysaker.kvaerner.no), Sep 21, 1996.
 *
 *  Made the block list a circular list to minimise searching in the list.
 *  Andy Walker (andy@lysaker.kvaerner.no), Sep 25, 1996.
 *
 *  Made mandatory locking a mount option. Default is not to allow mandatory
 *  locking.
 *  Andy Walker (andy@lysaker.kvaerner.no), Oct 04, 1996.
 *
 *  Some adaptations for NFS support.
 *  Olaf Kirch (okir@monad.swb.de), Dec 1996,
 *
 *  Fixed /proc/locks interface so that we can't overrun the buffer we are handed.
 *  Andy Walker (andy@lysaker.kvaerner.no), May 12, 1997.
 *
 *  Use slab allocator instead of kmalloc/kfree.
 *  Use generic list implementation from <linux/list.h>.
 *  Sped up posix_locks_deadlock by only considering blocked locks.
 *  Matthew Wilcox <willy@debian.org>, March, 2000.
 *
 *  Leases and LOCK_MAND
 *  Matthew Wilcox <willy@debian.org>, June, 2000.
 *  Stephen Rothwell <sfr@canb.auug.org.au>, June, 2000.
 */

#include <linux/capability.h>
#include <linux/file.h>
A
Al Viro 已提交
119
#include <linux/fdtable.h>
L
Linus Torvalds 已提交
120 121 122 123 124 125 126
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/security.h>
#include <linux/slab.h>
#include <linux/syscalls.h>
#include <linux/time.h>
127
#include <linux/rcupdate.h>
128
#include <linux/pid_namespace.h>
129
#include <linux/hashtable.h>
130 131
#include <linux/percpu.h>
#include <linux/lglock.h>
L
Linus Torvalds 已提交
132

133 134 135
#define CREATE_TRACE_POINTS
#include <trace/events/filelock.h>

L
Linus Torvalds 已提交
136 137 138 139
#include <asm/uaccess.h>

#define IS_POSIX(fl)	(fl->fl_flags & FL_POSIX)
#define IS_FLOCK(fl)	(fl->fl_flags & FL_FLOCK)
140
#define IS_LEASE(fl)	(fl->fl_flags & (FL_LEASE|FL_DELEG))
141
#define IS_OFDLCK(fl)	(fl->fl_flags & FL_OFDLCK)
L
Linus Torvalds 已提交
142

J
J. Bruce Fields 已提交
143 144
static bool lease_breaking(struct file_lock *fl)
{
145 146 147 148 149 150 151 152 153 154
	return fl->fl_flags & (FL_UNLOCK_PENDING | FL_DOWNGRADE_PENDING);
}

static int target_leasetype(struct file_lock *fl)
{
	if (fl->fl_flags & FL_UNLOCK_PENDING)
		return F_UNLCK;
	if (fl->fl_flags & FL_DOWNGRADE_PENDING)
		return F_RDLCK;
	return fl->fl_type;
J
J. Bruce Fields 已提交
155 156
}

L
Linus Torvalds 已提交
157 158 159
int leases_enable = 1;
int lease_break_time = 45;

160
/*
161 162 163 164
 * The global file_lock_list is only used for displaying /proc/locks, so we
 * keep a list on each CPU, with each list protected by its own spinlock via
 * the file_lock_lglock. Note that alterations to the list also require that
 * the relevant i_lock is held.
165
 */
166 167
DEFINE_STATIC_LGLOCK(file_lock_lglock);
static DEFINE_PER_CPU(struct hlist_head, file_lock_list);
168

169
/*
170
 * The blocked_hash is used to find POSIX lock loops for deadlock detection.
171
 * It is protected by blocked_lock_lock.
172 173 174 175 176 177 178
 *
 * We hash locks by lockowner in order to optimize searching for the lock a
 * particular lockowner is waiting on.
 *
 * FIXME: make this value scale via some heuristic? We generally will want more
 * buckets when we have more lockowners holding locks, but that's a little
 * difficult to determine without knowing what the workload will look like.
179
 */
180 181
#define BLOCKED_HASH_BITS	7
static DEFINE_HASHTABLE(blocked_hash, BLOCKED_HASH_BITS);
182

183
/*
184 185
 * This lock protects the blocked_hash. Generally, if you're accessing it, you
 * want to be holding this lock.
186 187 188 189 190 191 192 193 194 195 196 197 198 199
 *
 * In addition, it also protects the fl->fl_block list, and the fl->fl_next
 * pointer for file_lock structures that are acting as lock requests (in
 * contrast to those that are acting as records of acquired locks).
 *
 * Note that when we acquire this lock in order to change the above fields,
 * we often hold the i_lock as well. In certain cases, when reading the fields
 * protected by this lock, we can skip acquiring it iff we already hold the
 * i_lock.
 *
 * In particular, adding an entry to the fl_block list requires that you hold
 * both the i_lock and the blocked_lock_lock (acquired in that order). Deleting
 * an entry from the list however only requires the file_lock_lock.
 */
200
static DEFINE_SPINLOCK(blocked_lock_lock);
L
Linus Torvalds 已提交
201

202
static struct kmem_cache *flctx_cache __read_mostly;
203
static struct kmem_cache *filelock_cache __read_mostly;
L
Linus Torvalds 已提交
204

205 206 207 208 209 210 211 212 213 214 215 216 217
static struct file_lock_context *
locks_get_lock_context(struct inode *inode)
{
	struct file_lock_context *new;

	if (likely(inode->i_flctx))
		goto out;

	new = kmem_cache_alloc(flctx_cache, GFP_KERNEL);
	if (!new)
		goto out;

	INIT_LIST_HEAD(&new->flc_flock);
218
	INIT_LIST_HEAD(&new->flc_posix);
219
	INIT_LIST_HEAD(&new->flc_lease);
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242

	/*
	 * Assign the pointer if it's not already assigned. If it is, then
	 * free the context we just allocated.
	 */
	spin_lock(&inode->i_lock);
	if (likely(!inode->i_flctx)) {
		inode->i_flctx = new;
		new = NULL;
	}
	spin_unlock(&inode->i_lock);

	if (new)
		kmem_cache_free(flctx_cache, new);
out:
	return inode->i_flctx;
}

void
locks_free_lock_context(struct file_lock_context *ctx)
{
	if (ctx) {
		WARN_ON_ONCE(!list_empty(&ctx->flc_flock));
243
		WARN_ON_ONCE(!list_empty(&ctx->flc_posix));
244
		WARN_ON_ONCE(!list_empty(&ctx->flc_lease));
245 246 247 248
		kmem_cache_free(flctx_cache, ctx);
	}
}

M
Miklos Szeredi 已提交
249
static void locks_init_lock_heads(struct file_lock *fl)
M
Miklos Szeredi 已提交
250
{
251
	INIT_HLIST_NODE(&fl->fl_link);
252
	INIT_LIST_HEAD(&fl->fl_list);
M
Miklos Szeredi 已提交
253 254
	INIT_LIST_HEAD(&fl->fl_block);
	init_waitqueue_head(&fl->fl_wait);
M
Miklos Szeredi 已提交
255 256
}

L
Linus Torvalds 已提交
257
/* Allocate an empty lock structure. */
258
struct file_lock *locks_alloc_lock(void)
L
Linus Torvalds 已提交
259
{
M
Miklos Szeredi 已提交
260
	struct file_lock *fl = kmem_cache_zalloc(filelock_cache, GFP_KERNEL);
M
Miklos Szeredi 已提交
261 262

	if (fl)
M
Miklos Szeredi 已提交
263
		locks_init_lock_heads(fl);
M
Miklos Szeredi 已提交
264 265

	return fl;
L
Linus Torvalds 已提交
266
}
267
EXPORT_SYMBOL_GPL(locks_alloc_lock);
L
Linus Torvalds 已提交
268

269
void locks_release_private(struct file_lock *fl)
270 271 272 273 274 275 276
{
	if (fl->fl_ops) {
		if (fl->fl_ops->fl_release_private)
			fl->fl_ops->fl_release_private(fl);
		fl->fl_ops = NULL;
	}

277 278 279 280 281
	if (fl->fl_lmops) {
		if (fl->fl_lmops->lm_put_owner)
			fl->fl_lmops->lm_put_owner(fl);
		fl->fl_lmops = NULL;
	}
282
}
283
EXPORT_SYMBOL_GPL(locks_release_private);
284

L
Linus Torvalds 已提交
285
/* Free a lock which is not in use. */
286
void locks_free_lock(struct file_lock *fl)
L
Linus Torvalds 已提交
287
{
M
Miklos Szeredi 已提交
288
	BUG_ON(waitqueue_active(&fl->fl_wait));
289
	BUG_ON(!list_empty(&fl->fl_list));
M
Miklos Szeredi 已提交
290
	BUG_ON(!list_empty(&fl->fl_block));
291
	BUG_ON(!hlist_unhashed(&fl->fl_link));
L
Linus Torvalds 已提交
292

293
	locks_release_private(fl);
L
Linus Torvalds 已提交
294 295
	kmem_cache_free(filelock_cache, fl);
}
296
EXPORT_SYMBOL(locks_free_lock);
L
Linus Torvalds 已提交
297

298 299 300 301 302 303
static void
locks_dispose_list(struct list_head *dispose)
{
	struct file_lock *fl;

	while (!list_empty(dispose)) {
304 305
		fl = list_first_entry(dispose, struct file_lock, fl_list);
		list_del_init(&fl->fl_list);
306 307 308 309
		locks_free_lock(fl);
	}
}

L
Linus Torvalds 已提交
310 311
void locks_init_lock(struct file_lock *fl)
{
M
Miklos Szeredi 已提交
312 313
	memset(fl, 0, sizeof(struct file_lock));
	locks_init_lock_heads(fl);
L
Linus Torvalds 已提交
314 315 316 317 318 319 320
}

EXPORT_SYMBOL(locks_init_lock);

/*
 * Initialize a new lock from an existing file_lock structure.
 */
321
void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
L
Linus Torvalds 已提交
322 323 324
{
	new->fl_owner = fl->fl_owner;
	new->fl_pid = fl->fl_pid;
325
	new->fl_file = NULL;
L
Linus Torvalds 已提交
326 327 328 329
	new->fl_flags = fl->fl_flags;
	new->fl_type = fl->fl_type;
	new->fl_start = fl->fl_start;
	new->fl_end = fl->fl_end;
330
	new->fl_lmops = fl->fl_lmops;
331
	new->fl_ops = NULL;
332 333 334 335 336

	if (fl->fl_lmops) {
		if (fl->fl_lmops->lm_get_owner)
			fl->fl_lmops->lm_get_owner(new, fl);
	}
337
}
338
EXPORT_SYMBOL(locks_copy_conflock);
339 340 341

void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
{
342 343
	/* "new" must be a freshly-initialized lock */
	WARN_ON_ONCE(new->fl_ops);
344

345
	locks_copy_conflock(new, fl);
346

347
	new->fl_file = fl->fl_file;
L
Linus Torvalds 已提交
348
	new->fl_ops = fl->fl_ops;
349

350 351 352 353
	if (fl->fl_ops) {
		if (fl->fl_ops->fl_copy_lock)
			fl->fl_ops->fl_copy_lock(new, fl);
	}
L
Linus Torvalds 已提交
354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372
}

EXPORT_SYMBOL(locks_copy_lock);

static inline int flock_translate_cmd(int cmd) {
	if (cmd & LOCK_MAND)
		return cmd & (LOCK_MAND | LOCK_RW);
	switch (cmd) {
	case LOCK_SH:
		return F_RDLCK;
	case LOCK_EX:
		return F_WRLCK;
	case LOCK_UN:
		return F_UNLCK;
	}
	return -EINVAL;
}

/* Fill in a file_lock structure with an appropriate FLOCK lock. */
373 374
static struct file_lock *
flock_make_lock(struct file *filp, unsigned int cmd)
L
Linus Torvalds 已提交
375 376 377
{
	struct file_lock *fl;
	int type = flock_translate_cmd(cmd);
378

L
Linus Torvalds 已提交
379
	if (type < 0)
380
		return ERR_PTR(type);
L
Linus Torvalds 已提交
381 382 383
	
	fl = locks_alloc_lock();
	if (fl == NULL)
384
		return ERR_PTR(-ENOMEM);
L
Linus Torvalds 已提交
385 386

	fl->fl_file = filp;
387
	fl->fl_owner = filp;
L
Linus Torvalds 已提交
388 389 390 391 392
	fl->fl_pid = current->tgid;
	fl->fl_flags = FL_FLOCK;
	fl->fl_type = type;
	fl->fl_end = OFFSET_MAX;
	
393
	return fl;
L
Linus Torvalds 已提交
394 395
}

396
static int assign_type(struct file_lock *fl, long type)
L
Linus Torvalds 已提交
397 398 399 400 401 402 403 404 405 406 407 408 409
{
	switch (type) {
	case F_RDLCK:
	case F_WRLCK:
	case F_UNLCK:
		fl->fl_type = type;
		break;
	default:
		return -EINVAL;
	}
	return 0;
}

410 411
static int flock64_to_posix_lock(struct file *filp, struct file_lock *fl,
				 struct flock64 *l)
L
Linus Torvalds 已提交
412 413
{
	switch (l->l_whence) {
414
	case SEEK_SET:
415
		fl->fl_start = 0;
L
Linus Torvalds 已提交
416
		break;
417
	case SEEK_CUR:
418
		fl->fl_start = filp->f_pos;
L
Linus Torvalds 已提交
419
		break;
420
	case SEEK_END:
421
		fl->fl_start = i_size_read(file_inode(filp));
L
Linus Torvalds 已提交
422 423 424 425
		break;
	default:
		return -EINVAL;
	}
426 427 428 429 430
	if (l->l_start > OFFSET_MAX - fl->fl_start)
		return -EOVERFLOW;
	fl->fl_start += l->l_start;
	if (fl->fl_start < 0)
		return -EINVAL;
L
Linus Torvalds 已提交
431 432 433

	/* POSIX-1996 leaves the case l->l_len < 0 undefined;
	   POSIX-2001 defines it. */
434
	if (l->l_len > 0) {
435 436 437 438
		if (l->l_len - 1 > OFFSET_MAX - fl->fl_start)
			return -EOVERFLOW;
		fl->fl_end = fl->fl_start + l->l_len - 1;

439
	} else if (l->l_len < 0) {
440
		if (fl->fl_start + l->l_len < 0)
441
			return -EINVAL;
442 443 444 445 446
		fl->fl_end = fl->fl_start - 1;
		fl->fl_start += l->l_len;
	} else
		fl->fl_end = OFFSET_MAX;

L
Linus Torvalds 已提交
447 448 449 450 451 452 453 454 455 456
	fl->fl_owner = current->files;
	fl->fl_pid = current->tgid;
	fl->fl_file = filp;
	fl->fl_flags = FL_POSIX;
	fl->fl_ops = NULL;
	fl->fl_lmops = NULL;

	return assign_type(fl, l->l_type);
}

457 458 459 460 461
/* Verify a "struct flock" and copy it to a "struct file_lock" as a POSIX
 * style lock.
 */
static int flock_to_posix_lock(struct file *filp, struct file_lock *fl,
			       struct flock *l)
L
Linus Torvalds 已提交
462
{
463 464 465 466 467 468 469 470
	struct flock64 ll = {
		.l_type = l->l_type,
		.l_whence = l->l_whence,
		.l_start = l->l_start,
		.l_len = l->l_len,
	};

	return flock64_to_posix_lock(filp, fl, &ll);
L
Linus Torvalds 已提交
471 472 473
}

/* default lease lock manager operations */
J
Jeff Layton 已提交
474 475
static bool
lease_break_callback(struct file_lock *fl)
L
Linus Torvalds 已提交
476 477
{
	kill_fasync(&fl->fl_fasync, SIGIO, POLL_MSG);
J
Jeff Layton 已提交
478
	return false;
L
Linus Torvalds 已提交
479 480
}

481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
static void
lease_setup(struct file_lock *fl, void **priv)
{
	struct file *filp = fl->fl_file;
	struct fasync_struct *fa = *priv;

	/*
	 * fasync_insert_entry() returns the old entry if any. If there was no
	 * old entry, then it used "priv" and inserted it into the fasync list.
	 * Clear the pointer to indicate that it shouldn't be freed.
	 */
	if (!fasync_insert_entry(fa->fa_fd, filp, &fl->fl_fasync, fa))
		*priv = NULL;

	__f_setown(filp, task_pid(current), PIDTYPE_PID, 0);
}

498
static const struct lock_manager_operations lease_manager_ops = {
J
J. Bruce Fields 已提交
499 500
	.lm_break = lease_break_callback,
	.lm_change = lease_modify,
501
	.lm_setup = lease_setup,
L
Linus Torvalds 已提交
502 503 504 505 506
};

/*
 * Initialize a lease, use the default lock manager operations
 */
507
static int lease_init(struct file *filp, long type, struct file_lock *fl)
L
Linus Torvalds 已提交
508
 {
509 510 511
	if (assign_type(fl, type) != 0)
		return -EINVAL;

512
	fl->fl_owner = filp;
L
Linus Torvalds 已提交
513 514 515 516 517 518 519 520 521 522 523 524
	fl->fl_pid = current->tgid;

	fl->fl_file = filp;
	fl->fl_flags = FL_LEASE;
	fl->fl_start = 0;
	fl->fl_end = OFFSET_MAX;
	fl->fl_ops = NULL;
	fl->fl_lmops = &lease_manager_ops;
	return 0;
}

/* Allocate a file_lock initialised to this type of lease */
525
static struct file_lock *lease_alloc(struct file *filp, long type)
L
Linus Torvalds 已提交
526 527
{
	struct file_lock *fl = locks_alloc_lock();
528
	int error = -ENOMEM;
L
Linus Torvalds 已提交
529 530

	if (fl == NULL)
J
J. Bruce Fields 已提交
531
		return ERR_PTR(error);
L
Linus Torvalds 已提交
532 533

	error = lease_init(filp, type, fl);
534 535
	if (error) {
		locks_free_lock(fl);
J
J. Bruce Fields 已提交
536
		return ERR_PTR(error);
537
	}
J
J. Bruce Fields 已提交
538
	return fl;
L
Linus Torvalds 已提交
539 540 541 542 543 544 545 546 547 548 549 550 551
}

/* Check if two locks overlap each other.
 */
static inline int locks_overlap(struct file_lock *fl1, struct file_lock *fl2)
{
	return ((fl1->fl_end >= fl2->fl_start) &&
		(fl2->fl_end >= fl1->fl_start));
}

/*
 * Check whether two locks have the same owner.
 */
552
static int posix_same_owner(struct file_lock *fl1, struct file_lock *fl2)
L
Linus Torvalds 已提交
553
{
J
J. Bruce Fields 已提交
554
	if (fl1->fl_lmops && fl1->fl_lmops->lm_compare_owner)
L
Linus Torvalds 已提交
555
		return fl2->fl_lmops == fl1->fl_lmops &&
J
J. Bruce Fields 已提交
556
			fl1->fl_lmops->lm_compare_owner(fl1, fl2);
L
Linus Torvalds 已提交
557 558 559
	return fl1->fl_owner == fl2->fl_owner;
}

560
/* Must be called with the i_lock held! */
561
static void locks_insert_global_locks(struct file_lock *fl)
562
{
563 564 565 566
	lg_local_lock(&file_lock_lglock);
	fl->fl_link_cpu = smp_processor_id();
	hlist_add_head(&fl->fl_link, this_cpu_ptr(&file_lock_list));
	lg_local_unlock(&file_lock_lglock);
567 568
}

569
/* Must be called with the i_lock held! */
570
static void locks_delete_global_locks(struct file_lock *fl)
571
{
572 573 574 575 576 577 578 579
	/*
	 * Avoid taking lock if already unhashed. This is safe since this check
	 * is done while holding the i_lock, and new insertions into the list
	 * also require that it be held.
	 */
	if (hlist_unhashed(&fl->fl_link))
		return;
	lg_local_lock_cpu(&file_lock_lglock, fl->fl_link_cpu);
580
	hlist_del_init(&fl->fl_link);
581
	lg_local_unlock_cpu(&file_lock_lglock, fl->fl_link_cpu);
582 583
}

584 585 586 587 588 589 590 591
static unsigned long
posix_owner_key(struct file_lock *fl)
{
	if (fl->fl_lmops && fl->fl_lmops->lm_owner_key)
		return fl->fl_lmops->lm_owner_key(fl);
	return (unsigned long)fl->fl_owner;
}

592
static void locks_insert_global_blocked(struct file_lock *waiter)
593
{
594
	hash_add(blocked_hash, &waiter->fl_link, posix_owner_key(waiter));
595 596
}

597
static void locks_delete_global_blocked(struct file_lock *waiter)
598
{
599
	hash_del(&waiter->fl_link);
600 601
}

L
Linus Torvalds 已提交
602 603
/* Remove waiter from blocker's block list.
 * When blocker ends up pointing to itself then the list is empty.
604
 *
605
 * Must be called with blocked_lock_lock held.
L
Linus Torvalds 已提交
606
 */
607
static void __locks_delete_block(struct file_lock *waiter)
L
Linus Torvalds 已提交
608
{
609
	locks_delete_global_blocked(waiter);
L
Linus Torvalds 已提交
610 611 612 613
	list_del_init(&waiter->fl_block);
	waiter->fl_next = NULL;
}

614
static void locks_delete_block(struct file_lock *waiter)
L
Linus Torvalds 已提交
615
{
616
	spin_lock(&blocked_lock_lock);
L
Linus Torvalds 已提交
617
	__locks_delete_block(waiter);
618
	spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
619 620 621 622 623 624
}

/* Insert waiter into blocker's block list.
 * We use a circular list so that processes can be easily woken up in
 * the order they blocked. The documentation doesn't require this but
 * it seems like the reasonable thing to do.
625
 *
626
 * Must be called with both the i_lock and blocked_lock_lock held. The fl_block
J
Jeff Layton 已提交
627
 * list itself is protected by the blocked_lock_lock, but by ensuring that the
628
 * i_lock is also held on insertions we can avoid taking the blocked_lock_lock
629
 * in some cases when we see that the fl_block list is empty.
L
Linus Torvalds 已提交
630
 */
631 632
static void __locks_insert_block(struct file_lock *blocker,
					struct file_lock *waiter)
L
Linus Torvalds 已提交
633
{
634
	BUG_ON(!list_empty(&waiter->fl_block));
L
Linus Torvalds 已提交
635
	waiter->fl_next = blocker;
636
	list_add_tail(&waiter->fl_block, &blocker->fl_block);
637
	if (IS_POSIX(blocker) && !IS_OFDLCK(blocker))
638 639 640 641 642 643 644
		locks_insert_global_blocked(waiter);
}

/* Must be called with i_lock held. */
static void locks_insert_block(struct file_lock *blocker,
					struct file_lock *waiter)
{
645
	spin_lock(&blocked_lock_lock);
646
	__locks_insert_block(blocker, waiter);
647
	spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
648 649
}

650 651 652
/*
 * Wake up processes blocked waiting for blocker.
 *
653
 * Must be called with the inode->i_lock held!
L
Linus Torvalds 已提交
654 655 656
 */
static void locks_wake_up_blocks(struct file_lock *blocker)
{
657 658 659 660 661
	/*
	 * Avoid taking global lock if list is empty. This is safe since new
	 * blocked requests are only added to the list under the i_lock, and
	 * the i_lock is always held here. Note that removal from the fl_block
	 * list does not require the i_lock, so we must recheck list_empty()
662
	 * after acquiring the blocked_lock_lock.
663 664 665 666
	 */
	if (list_empty(&blocker->fl_block))
		return;

667
	spin_lock(&blocked_lock_lock);
L
Linus Torvalds 已提交
668
	while (!list_empty(&blocker->fl_block)) {
669 670 671
		struct file_lock *waiter;

		waiter = list_first_entry(&blocker->fl_block,
L
Linus Torvalds 已提交
672 673
				struct file_lock, fl_block);
		__locks_delete_block(waiter);
J
J. Bruce Fields 已提交
674 675
		if (waiter->fl_lmops && waiter->fl_lmops->lm_notify)
			waiter->fl_lmops->lm_notify(waiter);
L
Linus Torvalds 已提交
676 677 678
		else
			wake_up(&waiter->fl_wait);
	}
679
	spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
680 681
}

682 683 684 685 686 687 688 689
static void
locks_insert_lock_ctx(struct file_lock *fl, struct list_head *before)
{
	fl->fl_nspid = get_pid(task_tgid(current));
	list_add_tail(&fl->fl_list, before);
	locks_insert_global_locks(fl);
}

690 691
static void
locks_unlink_lock_ctx(struct file_lock *fl)
L
Linus Torvalds 已提交
692
{
693
	locks_delete_global_locks(fl);
694
	list_del_init(&fl->fl_list);
695 696 697 698
	if (fl->fl_nspid) {
		put_pid(fl->fl_nspid);
		fl->fl_nspid = NULL;
	}
L
Linus Torvalds 已提交
699
	locks_wake_up_blocks(fl);
700 701
}

702 703
static void
locks_delete_lock_ctx(struct file_lock *fl, struct list_head *dispose)
704
{
705
	locks_unlink_lock_ctx(fl);
706
	if (dispose)
707
		list_add(&fl->fl_list, dispose);
708 709
	else
		locks_free_lock(fl);
L
Linus Torvalds 已提交
710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757
}

/* Determine if lock sys_fl blocks lock caller_fl. Common functionality
 * checks for shared/exclusive status of overlapping locks.
 */
static int locks_conflict(struct file_lock *caller_fl, struct file_lock *sys_fl)
{
	if (sys_fl->fl_type == F_WRLCK)
		return 1;
	if (caller_fl->fl_type == F_WRLCK)
		return 1;
	return 0;
}

/* Determine if lock sys_fl blocks lock caller_fl. POSIX specific
 * checking before calling the locks_conflict().
 */
static int posix_locks_conflict(struct file_lock *caller_fl, struct file_lock *sys_fl)
{
	/* POSIX locks owned by the same process do not conflict with
	 * each other.
	 */
	if (!IS_POSIX(sys_fl) || posix_same_owner(caller_fl, sys_fl))
		return (0);

	/* Check whether they overlap */
	if (!locks_overlap(caller_fl, sys_fl))
		return 0;

	return (locks_conflict(caller_fl, sys_fl));
}

/* Determine if lock sys_fl blocks lock caller_fl. FLOCK specific
 * checking before calling the locks_conflict().
 */
static int flock_locks_conflict(struct file_lock *caller_fl, struct file_lock *sys_fl)
{
	/* FLOCK locks referring to the same filp do not conflict with
	 * each other.
	 */
	if (!IS_FLOCK(sys_fl) || (caller_fl->fl_file == sys_fl->fl_file))
		return (0);
	if ((caller_fl->fl_type & LOCK_MAND) || (sys_fl->fl_type & LOCK_MAND))
		return 0;

	return (locks_conflict(caller_fl, sys_fl));
}

758
void
759
posix_test_lock(struct file *filp, struct file_lock *fl)
L
Linus Torvalds 已提交
760 761
{
	struct file_lock *cfl;
762
	struct file_lock_context *ctx;
763
	struct inode *inode = file_inode(filp);
L
Linus Torvalds 已提交
764

765 766 767 768 769 770
	ctx = inode->i_flctx;
	if (!ctx || list_empty_careful(&ctx->flc_posix)) {
		fl->fl_type = F_UNLCK;
		return;
	}

771
	spin_lock(&inode->i_lock);
772 773 774 775 776 777 778
	list_for_each_entry(cfl, &ctx->flc_posix, fl_list) {
		if (posix_locks_conflict(fl, cfl)) {
			locks_copy_conflock(fl, cfl);
			if (cfl->fl_nspid)
				fl->fl_pid = pid_vnr(cfl->fl_nspid);
			goto out;
		}
L
Linus Torvalds 已提交
779
	}
780 781
	fl->fl_type = F_UNLCK;
out:
782
	spin_unlock(&inode->i_lock);
783
	return;
L
Linus Torvalds 已提交
784 785 786
}
EXPORT_SYMBOL(posix_test_lock);

787 788 789 790 791
/*
 * Deadlock detection:
 *
 * We attempt to detect deadlocks that are due purely to posix file
 * locks.
L
Linus Torvalds 已提交
792
 *
793 794 795 796 797 798 799
 * We assume that a task can be waiting for at most one lock at a time.
 * So for any acquired lock, the process holding that lock may be
 * waiting on at most one other lock.  That lock in turns may be held by
 * someone waiting for at most one other lock.  Given a requested lock
 * caller_fl which is about to wait for a conflicting lock block_fl, we
 * follow this chain of waiters to ensure we are not about to create a
 * cycle.
L
Linus Torvalds 已提交
800
 *
801 802 803
 * Since we do this before we ever put a process to sleep on a lock, we
 * are ensured that there is never a cycle; that is what guarantees that
 * the while() loop in posix_locks_deadlock() eventually completes.
804
 *
805 806 807 808
 * Note: the above assumption may not be true when handling lock
 * requests from a broken NFS client. It may also fail in the presence
 * of tasks (such as posix threads) sharing the same open file table.
 * To handle those cases, we just bail out after a few iterations.
809
 *
810
 * For FL_OFDLCK locks, the owner is the filp, not the files_struct.
811 812 813 814
 * Because the owner is not even nominally tied to a thread of
 * execution, the deadlock detection below can't reasonably work well. Just
 * skip it for those.
 *
815
 * In principle, we could do a more limited deadlock detection on FL_OFDLCK
816 817
 * locks that just checks for the case where two tasks are attempting to
 * upgrade from read to write locks on the same inode.
L
Linus Torvalds 已提交
818
 */
819 820 821

#define MAX_DEADLK_ITERATIONS 10

822 823 824 825 826
/* Find a lock that the owner of the given block_fl is blocking on. */
static struct file_lock *what_owner_is_waiting_for(struct file_lock *block_fl)
{
	struct file_lock *fl;

827
	hash_for_each_possible(blocked_hash, fl, fl_link, posix_owner_key(block_fl)) {
828 829 830 831 832 833
		if (posix_same_owner(fl, block_fl))
			return fl->fl_next;
	}
	return NULL;
}

834
/* Must be called with the blocked_lock_lock held! */
835
static int posix_locks_deadlock(struct file_lock *caller_fl,
L
Linus Torvalds 已提交
836 837
				struct file_lock *block_fl)
{
838
	int i = 0;
L
Linus Torvalds 已提交
839

840 841
	/*
	 * This deadlock detector can't reasonably detect deadlocks with
842
	 * FL_OFDLCK locks, since they aren't owned by a process, per-se.
843
	 */
844
	if (IS_OFDLCK(caller_fl))
845 846
		return 0;

847 848 849 850 851
	while ((block_fl = what_owner_is_waiting_for(block_fl))) {
		if (i++ > MAX_DEADLK_ITERATIONS)
			return 0;
		if (posix_same_owner(caller_fl, block_fl))
			return 1;
L
Linus Torvalds 已提交
852 853 854 855 856
	}
	return 0;
}

/* Try to create a FLOCK lock on filp. We always insert new FLOCK locks
857
 * after any leases, but before any posix locks.
858 859 860 861
 *
 * Note that if called with an FL_EXISTS argument, the caller may determine
 * whether or not a lock was successfully freed by testing the return
 * value for -ENOENT.
L
Linus Torvalds 已提交
862
 */
863
static int flock_lock_file(struct file *filp, struct file_lock *request)
L
Linus Torvalds 已提交
864
{
865
	struct file_lock *new_fl = NULL;
866 867 868
	struct file_lock *fl;
	struct file_lock_context *ctx;
	struct inode *inode = file_inode(filp);
L
Linus Torvalds 已提交
869
	int error = 0;
870
	bool found = false;
871
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
872

873 874 875 876
	ctx = locks_get_lock_context(inode);
	if (!ctx)
		return -ENOMEM;

877
	if (!(request->fl_flags & FL_ACCESS) && (request->fl_type != F_UNLCK)) {
878
		new_fl = locks_alloc_lock();
879 880
		if (!new_fl)
			return -ENOMEM;
881 882
	}

883
	spin_lock(&inode->i_lock);
884 885 886
	if (request->fl_flags & FL_ACCESS)
		goto find_conflict;

887
	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
L
Linus Torvalds 已提交
888 889
		if (filp != fl->fl_file)
			continue;
890
		if (request->fl_type == fl->fl_type)
L
Linus Torvalds 已提交
891
			goto out;
892 893
		found = true;
		locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
894 895 896
		break;
	}

897 898 899
	if (request->fl_type == F_UNLCK) {
		if ((request->fl_flags & FL_EXISTS) && !found)
			error = -ENOENT;
900
		goto out;
901
	}
L
Linus Torvalds 已提交
902 903 904 905 906

	/*
	 * If a higher-priority process was blocked on the old file lock,
	 * give it the opportunity to lock the file.
	 */
907
	if (found) {
908
		spin_unlock(&inode->i_lock);
909
		cond_resched();
910
		spin_lock(&inode->i_lock);
911
	}
L
Linus Torvalds 已提交
912

913
find_conflict:
914
	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
915
		if (!flock_locks_conflict(request, fl))
L
Linus Torvalds 已提交
916 917
			continue;
		error = -EAGAIN;
918 919 920 921
		if (!(request->fl_flags & FL_SLEEP))
			goto out;
		error = FILE_LOCK_DEFERRED;
		locks_insert_block(fl, request);
L
Linus Torvalds 已提交
922 923
		goto out;
	}
924 925
	if (request->fl_flags & FL_ACCESS)
		goto out;
926
	locks_copy_lock(new_fl, request);
927
	locks_insert_lock_ctx(new_fl, &ctx->flc_flock);
928
	new_fl = NULL;
929
	error = 0;
L
Linus Torvalds 已提交
930 931

out:
932
	spin_unlock(&inode->i_lock);
933 934
	if (new_fl)
		locks_free_lock(new_fl);
935
	locks_dispose_list(&dispose);
L
Linus Torvalds 已提交
936 937 938
	return error;
}

939
static int __posix_lock_file(struct inode *inode, struct file_lock *request, struct file_lock *conflock)
L
Linus Torvalds 已提交
940
{
941
	struct file_lock *fl, *tmp;
942 943
	struct file_lock *new_fl = NULL;
	struct file_lock *new_fl2 = NULL;
L
Linus Torvalds 已提交
944 945
	struct file_lock *left = NULL;
	struct file_lock *right = NULL;
946
	struct file_lock_context *ctx;
947 948
	int error;
	bool added = false;
949
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
950

951 952 953 954
	ctx = locks_get_lock_context(inode);
	if (!ctx)
		return -ENOMEM;

L
Linus Torvalds 已提交
955 956 957
	/*
	 * We may need two file_lock structures for this operation,
	 * so we get them in advance to avoid races.
958 959
	 *
	 * In some cases we can be sure, that no new locks will be needed
L
Linus Torvalds 已提交
960
	 */
961 962 963 964 965 966
	if (!(request->fl_flags & FL_ACCESS) &&
	    (request->fl_type != F_UNLCK ||
	     request->fl_start != 0 || request->fl_end != OFFSET_MAX)) {
		new_fl = locks_alloc_lock();
		new_fl2 = locks_alloc_lock();
	}
L
Linus Torvalds 已提交
967

968
	spin_lock(&inode->i_lock);
969 970 971
	/*
	 * New lock request. Walk all POSIX locks and look for conflicts. If
	 * there are any, either return error or put the request on the
972
	 * blocker's list of waiters and the global blocked_hash.
973
	 */
L
Linus Torvalds 已提交
974
	if (request->fl_type != F_UNLCK) {
975
		list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
L
Linus Torvalds 已提交
976 977 978 979
			if (!IS_POSIX(fl))
				continue;
			if (!posix_locks_conflict(request, fl))
				continue;
980
			if (conflock)
981
				locks_copy_conflock(conflock, fl);
L
Linus Torvalds 已提交
982 983 984
			error = -EAGAIN;
			if (!(request->fl_flags & FL_SLEEP))
				goto out;
985 986 987 988
			/*
			 * Deadlock detection and insertion into the blocked
			 * locks list must be done while holding the same lock!
			 */
L
Linus Torvalds 已提交
989
			error = -EDEADLK;
990
			spin_lock(&blocked_lock_lock);
991 992 993 994
			if (likely(!posix_locks_deadlock(request, fl))) {
				error = FILE_LOCK_DEFERRED;
				__locks_insert_block(fl, request);
			}
995
			spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
996 997 998 999 1000 1001 1002 1003 1004
			goto out;
  		}
  	}

	/* If we're just looking for a conflict, we're done. */
	error = 0;
	if (request->fl_flags & FL_ACCESS)
		goto out;

1005 1006 1007 1008
	/* Find the first old lock with the same owner as the new lock */
	list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
		if (posix_same_owner(request, fl))
			break;
L
Linus Torvalds 已提交
1009 1010
	}

1011
	/* Process locks with this owner. */
1012 1013 1014 1015 1016
	list_for_each_entry_safe_from(fl, tmp, &ctx->flc_posix, fl_list) {
		if (!posix_same_owner(request, fl))
			break;

		/* Detect adjacent or overlapping regions (if same lock type) */
L
Linus Torvalds 已提交
1017
		if (request->fl_type == fl->fl_type) {
O
Olaf Kirch 已提交
1018 1019 1020 1021
			/* In all comparisons of start vs end, use
			 * "start - 1" rather than "end + 1". If end
			 * is OFFSET_MAX, end + 1 will become negative.
			 */
L
Linus Torvalds 已提交
1022
			if (fl->fl_end < request->fl_start - 1)
1023
				continue;
L
Linus Torvalds 已提交
1024 1025 1026
			/* If the next lock in the list has entirely bigger
			 * addresses than the new one, insert the lock here.
			 */
O
Olaf Kirch 已提交
1027
			if (fl->fl_start - 1 > request->fl_end)
L
Linus Torvalds 已提交
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043
				break;

			/* If we come here, the new and old lock are of the
			 * same type and adjacent or overlapping. Make one
			 * lock yielding from the lower start address of both
			 * locks to the higher end address.
			 */
			if (fl->fl_start > request->fl_start)
				fl->fl_start = request->fl_start;
			else
				request->fl_start = fl->fl_start;
			if (fl->fl_end < request->fl_end)
				fl->fl_end = request->fl_end;
			else
				request->fl_end = fl->fl_end;
			if (added) {
1044
				locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
1045 1046 1047
				continue;
			}
			request = fl;
1048
			added = true;
1049
		} else {
L
Linus Torvalds 已提交
1050 1051 1052 1053
			/* Processing for different lock types is a bit
			 * more complex.
			 */
			if (fl->fl_end < request->fl_start)
1054
				continue;
L
Linus Torvalds 已提交
1055 1056 1057
			if (fl->fl_start > request->fl_end)
				break;
			if (request->fl_type == F_UNLCK)
1058
				added = true;
L
Linus Torvalds 已提交
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072
			if (fl->fl_start < request->fl_start)
				left = fl;
			/* If the next lock in the list has a higher end
			 * address than the new one, insert the new one here.
			 */
			if (fl->fl_end > request->fl_end) {
				right = fl;
				break;
			}
			if (fl->fl_start >= request->fl_start) {
				/* The new lock completely replaces an old
				 * one (This may happen several times).
				 */
				if (added) {
1073
					locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
1074 1075
					continue;
				}
1076 1077 1078 1079 1080 1081
				/*
				 * Replace the old lock with new_fl, and
				 * remove the old one. It's safe to do the
				 * insert here since we know that we won't be
				 * using new_fl later, and that the lock is
				 * just replacing an existing lock.
L
Linus Torvalds 已提交
1082
				 */
1083 1084 1085 1086 1087 1088
				error = -ENOLCK;
				if (!new_fl)
					goto out;
				locks_copy_lock(new_fl, request);
				request = new_fl;
				new_fl = NULL;
1089 1090
				locks_insert_lock_ctx(request, &fl->fl_list);
				locks_delete_lock_ctx(fl, &dispose);
1091
				added = true;
L
Linus Torvalds 已提交
1092 1093 1094 1095
			}
		}
	}

1096
	/*
1097 1098 1099
	 * The above code only modifies existing locks in case of merging or
	 * replacing. If new lock(s) need to be inserted all modifications are
	 * done below this, so it's safe yet to bail out.
1100 1101 1102 1103 1104
	 */
	error = -ENOLCK; /* "no luck" */
	if (right && left == right && !new_fl2)
		goto out;

L
Linus Torvalds 已提交
1105 1106
	error = 0;
	if (!added) {
1107 1108 1109
		if (request->fl_type == F_UNLCK) {
			if (request->fl_flags & FL_EXISTS)
				error = -ENOENT;
L
Linus Torvalds 已提交
1110
			goto out;
1111
		}
1112 1113 1114 1115 1116

		if (!new_fl) {
			error = -ENOLCK;
			goto out;
		}
L
Linus Torvalds 已提交
1117
		locks_copy_lock(new_fl, request);
1118
		locks_insert_lock_ctx(new_fl, &fl->fl_list);
L
Linus Torvalds 已提交
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
		new_fl = NULL;
	}
	if (right) {
		if (left == right) {
			/* The new lock breaks the old one in two pieces,
			 * so we have to use the second new lock.
			 */
			left = new_fl2;
			new_fl2 = NULL;
			locks_copy_lock(left, right);
1129
			locks_insert_lock_ctx(left, &fl->fl_list);
L
Linus Torvalds 已提交
1130 1131 1132 1133 1134 1135 1136 1137 1138
		}
		right->fl_start = request->fl_end + 1;
		locks_wake_up_blocks(right);
	}
	if (left) {
		left->fl_end = request->fl_start - 1;
		locks_wake_up_blocks(left);
	}
 out:
1139
	spin_unlock(&inode->i_lock);
L
Linus Torvalds 已提交
1140 1141 1142 1143 1144 1145 1146
	/*
	 * Free any unused locks.
	 */
	if (new_fl)
		locks_free_lock(new_fl);
	if (new_fl2)
		locks_free_lock(new_fl2);
1147
	locks_dispose_list(&dispose);
L
Linus Torvalds 已提交
1148 1149 1150 1151 1152 1153 1154
	return error;
}

/**
 * posix_lock_file - Apply a POSIX-style lock to a file
 * @filp: The file to apply the lock to
 * @fl: The lock to be applied
1155
 * @conflock: Place to return a copy of the conflicting lock, if found.
L
Linus Torvalds 已提交
1156 1157 1158 1159
 *
 * Add a POSIX style lock to a file.
 * We merge adjacent & overlapping locks whenever possible.
 * POSIX locks are sorted by owner task, then by starting address
1160 1161 1162 1163
 *
 * Note that if called with an FL_EXISTS argument, the caller may determine
 * whether or not a lock was successfully freed by testing the return
 * value for -ENOENT.
L
Linus Torvalds 已提交
1164
 */
1165
int posix_lock_file(struct file *filp, struct file_lock *fl,
1166 1167
			struct file_lock *conflock)
{
A
Al Viro 已提交
1168
	return __posix_lock_file(file_inode(filp), fl, conflock);
L
Linus Torvalds 已提交
1169
}
1170
EXPORT_SYMBOL(posix_lock_file);
L
Linus Torvalds 已提交
1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185

/**
 * posix_lock_file_wait - Apply a POSIX-style lock to a file
 * @filp: The file to apply the lock to
 * @fl: The lock to be applied
 *
 * Add a POSIX style lock to a file.
 * We merge adjacent & overlapping locks whenever possible.
 * POSIX locks are sorted by owner task, then by starting address
 */
int posix_lock_file_wait(struct file *filp, struct file_lock *fl)
{
	int error;
	might_sleep ();
	for (;;) {
1186
		error = posix_lock_file(filp, fl, NULL);
1187
		if (error != FILE_LOCK_DEFERRED)
L
Linus Torvalds 已提交
1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201
			break;
		error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
		if (!error)
			continue;

		locks_delete_block(fl);
		break;
	}
	return error;
}
EXPORT_SYMBOL(posix_lock_file_wait);

/**
 * locks_mandatory_locked - Check for an active lock
1202
 * @file: the file to check
L
Linus Torvalds 已提交
1203 1204 1205 1206
 *
 * Searches the inode's list of locks to find any POSIX locks which conflict.
 * This function is called from locks_verify_locked() only.
 */
1207
int locks_mandatory_locked(struct file *file)
L
Linus Torvalds 已提交
1208
{
1209
	int ret;
1210
	struct inode *inode = file_inode(file);
1211
	struct file_lock_context *ctx;
L
Linus Torvalds 已提交
1212 1213
	struct file_lock *fl;

1214 1215 1216 1217
	ctx = inode->i_flctx;
	if (!ctx || list_empty_careful(&ctx->flc_posix))
		return 0;

L
Linus Torvalds 已提交
1218 1219 1220
	/*
	 * Search the lock list for this inode for any POSIX locks.
	 */
1221
	spin_lock(&inode->i_lock);
1222 1223
	ret = 0;
	list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
1224
		if (fl->fl_owner != current->files &&
1225 1226
		    fl->fl_owner != file) {
			ret = -EAGAIN;
L
Linus Torvalds 已提交
1227
			break;
1228
		}
L
Linus Torvalds 已提交
1229
	}
1230
	spin_unlock(&inode->i_lock);
1231
	return ret;
L
Linus Torvalds 已提交
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
}

/**
 * locks_mandatory_area - Check for a conflicting lock
 * @read_write: %FLOCK_VERIFY_WRITE for exclusive access, %FLOCK_VERIFY_READ
 *		for shared
 * @inode:      the file to check
 * @filp:       how the file was opened (if it was)
 * @offset:     start of area to check
 * @count:      length of area to check
 *
 * Searches the inode's list of locks to find any POSIX locks which conflict.
 * This function is called from rw_verify_area() and
 * locks_verify_truncate().
 */
int locks_mandatory_area(int read_write, struct inode *inode,
			 struct file *filp, loff_t offset,
			 size_t count)
{
	struct file_lock fl;
	int error;
1253
	bool sleep = false;
L
Linus Torvalds 已提交
1254 1255 1256 1257 1258 1259

	locks_init_lock(&fl);
	fl.fl_pid = current->tgid;
	fl.fl_file = filp;
	fl.fl_flags = FL_POSIX | FL_ACCESS;
	if (filp && !(filp->f_flags & O_NONBLOCK))
1260
		sleep = true;
L
Linus Torvalds 已提交
1261 1262 1263 1264 1265
	fl.fl_type = (read_write == FLOCK_VERIFY_WRITE) ? F_WRLCK : F_RDLCK;
	fl.fl_start = offset;
	fl.fl_end = offset + count - 1;

	for (;;) {
1266
		if (filp) {
1267
			fl.fl_owner = filp;
1268 1269 1270 1271 1272 1273 1274 1275 1276
			fl.fl_flags &= ~FL_SLEEP;
			error = __posix_lock_file(inode, &fl, NULL);
			if (!error)
				break;
		}

		if (sleep)
			fl.fl_flags |= FL_SLEEP;
		fl.fl_owner = current->files;
1277
		error = __posix_lock_file(inode, &fl, NULL);
1278
		if (error != FILE_LOCK_DEFERRED)
L
Linus Torvalds 已提交
1279 1280 1281 1282 1283 1284 1285
			break;
		error = wait_event_interruptible(fl.fl_wait, !fl.fl_next);
		if (!error) {
			/*
			 * If we've been sleeping someone might have
			 * changed the permissions behind our back.
			 */
1286
			if (__mandatory_lock(inode))
L
Linus Torvalds 已提交
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
				continue;
		}

		locks_delete_block(&fl);
		break;
	}

	return error;
}

EXPORT_SYMBOL(locks_mandatory_area);

1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309
static void lease_clear_pending(struct file_lock *fl, int arg)
{
	switch (arg) {
	case F_UNLCK:
		fl->fl_flags &= ~FL_UNLOCK_PENDING;
		/* fall through: */
	case F_RDLCK:
		fl->fl_flags &= ~FL_DOWNGRADE_PENDING;
	}
}

L
Linus Torvalds 已提交
1310
/* We already had a lease on this file; just change its type */
1311
int lease_modify(struct file_lock **before, int arg, struct list_head *dispose)
L
Linus Torvalds 已提交
1312 1313 1314 1315 1316 1317
{
	struct file_lock *fl = *before;
	int error = assign_type(fl, arg);

	if (error)
		return error;
1318
	lease_clear_pending(fl, arg);
L
Linus Torvalds 已提交
1319
	locks_wake_up_blocks(fl);
1320 1321 1322 1323 1324
	if (arg == F_UNLCK) {
		struct file *filp = fl->fl_file;

		f_delown(filp);
		filp->f_owner.signum = 0;
1325 1326 1327 1328 1329
		fasync_helper(0, fl->fl_file, 0, &fl->fl_fasync);
		if (fl->fl_fasync != NULL) {
			printk(KERN_ERR "locks_delete_lock: fasync == %p\n", fl->fl_fasync);
			fl->fl_fasync = NULL;
		}
1330
		locks_delete_lock_ctx(fl, dispose);
1331
	}
L
Linus Torvalds 已提交
1332 1333 1334 1335
	return 0;
}
EXPORT_SYMBOL(lease_modify);

1336 1337 1338 1339 1340 1341 1342 1343
static bool past_time(unsigned long then)
{
	if (!then)
		/* 0 is a special value meaning "this never expires": */
		return false;
	return time_after(jiffies, then);
}

1344
static void time_out_leases(struct inode *inode, struct list_head *dispose)
L
Linus Torvalds 已提交
1345
{
1346 1347
	struct file_lock_context *ctx = inode->i_flctx;
	struct file_lock *fl, *tmp;
L
Linus Torvalds 已提交
1348

1349 1350
	lockdep_assert_held(&inode->i_lock);

1351
	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, fl_list) {
1352
		trace_time_out_leases(inode, fl);
1353
		if (past_time(fl->fl_downgrade_time))
1354
			lease_modify(&fl, F_RDLCK, dispose);
1355
		if (past_time(fl->fl_break_time))
1356
			lease_modify(&fl, F_UNLCK, dispose);
L
Linus Torvalds 已提交
1357 1358 1359
	}
}

J
J. Bruce Fields 已提交
1360 1361 1362 1363 1364 1365 1366
static bool leases_conflict(struct file_lock *lease, struct file_lock *breaker)
{
	if ((breaker->fl_flags & FL_DELEG) && (lease->fl_flags & FL_LEASE))
		return false;
	return locks_conflict(breaker, lease);
}

1367 1368 1369
static bool
any_leases_conflict(struct inode *inode, struct file_lock *breaker)
{
1370
	struct file_lock_context *ctx = inode->i_flctx;
1371 1372 1373 1374
	struct file_lock *fl;

	lockdep_assert_held(&inode->i_lock);

1375
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
1376 1377 1378 1379 1380 1381
		if (leases_conflict(fl, breaker))
			return true;
	}
	return false;
}

L
Linus Torvalds 已提交
1382 1383 1384
/**
 *	__break_lease	-	revoke all outstanding leases on file
 *	@inode: the inode of the file to return
J
J. Bruce Fields 已提交
1385 1386 1387 1388
 *	@mode: O_RDONLY: break only write leases; O_WRONLY or O_RDWR:
 *	    break all leases
 *	@type: FL_LEASE: break leases and delegations; FL_DELEG: break
 *	    only delegations
L
Linus Torvalds 已提交
1389
 *
1390 1391 1392
 *	break_lease (inlined for speed) has checked there already is at least
 *	some kind of lock (maybe a lease) on this file.  Leases are broken on
 *	a call to open() or truncate().  This function can sleep unless you
L
Linus Torvalds 已提交
1393 1394
 *	specified %O_NONBLOCK to your open().
 */
J
J. Bruce Fields 已提交
1395
int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
L
Linus Torvalds 已提交
1396
{
1397
	int error = 0;
1398
	struct file_lock *new_fl;
1399 1400
	struct file_lock_context *ctx = inode->i_flctx;
	struct file_lock *fl;
L
Linus Torvalds 已提交
1401
	unsigned long break_time;
1402
	int want_write = (mode & O_ACCMODE) != O_RDONLY;
1403
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
1404

1405
	new_fl = lease_alloc(NULL, want_write ? F_WRLCK : F_RDLCK);
1406 1407
	if (IS_ERR(new_fl))
		return PTR_ERR(new_fl);
J
J. Bruce Fields 已提交
1408
	new_fl->fl_flags = type;
L
Linus Torvalds 已提交
1409

1410 1411 1412 1413 1414 1415
	/* typically we will check that ctx is non-NULL before calling */
	if (!ctx) {
		WARN_ON_ONCE(1);
		return error;
	}

1416
	spin_lock(&inode->i_lock);
L
Linus Torvalds 已提交
1417

1418
	time_out_leases(inode, &dispose);
L
Linus Torvalds 已提交
1419

1420
	if (!any_leases_conflict(inode, new_fl))
1421 1422
		goto out;

L
Linus Torvalds 已提交
1423 1424 1425 1426 1427 1428 1429
	break_time = 0;
	if (lease_break_time > 0) {
		break_time = jiffies + lease_break_time * HZ;
		if (break_time == 0)
			break_time++;	/* so that 0 means no break time */
	}

1430
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
J
J. Bruce Fields 已提交
1431 1432
		if (!leases_conflict(fl, new_fl))
			continue;
1433 1434 1435 1436
		if (want_write) {
			if (fl->fl_flags & FL_UNLOCK_PENDING)
				continue;
			fl->fl_flags |= FL_UNLOCK_PENDING;
L
Linus Torvalds 已提交
1437
			fl->fl_break_time = break_time;
1438
		} else {
1439
			if (lease_breaking(fl))
1440 1441 1442
				continue;
			fl->fl_flags |= FL_DOWNGRADE_PENDING;
			fl->fl_downgrade_time = break_time;
L
Linus Torvalds 已提交
1443
		}
J
Jeff Layton 已提交
1444
		if (fl->fl_lmops->lm_break(fl))
1445
			locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
1446 1447
	}

1448
	if (list_empty(&ctx->flc_lease))
J
Jeff Layton 已提交
1449 1450
		goto out;

1451
	if (mode & O_NONBLOCK) {
1452
		trace_break_lease_noblock(inode, new_fl);
L
Linus Torvalds 已提交
1453 1454 1455 1456 1457
		error = -EWOULDBLOCK;
		goto out;
	}

restart:
1458 1459
	fl = list_first_entry(&ctx->flc_lease, struct file_lock, fl_list);
	break_time = fl->fl_break_time;
1460
	if (break_time != 0)
L
Linus Torvalds 已提交
1461
		break_time -= jiffies;
1462 1463
	if (break_time == 0)
		break_time++;
1464
	locks_insert_block(fl, new_fl);
1465
	trace_break_lease_block(inode, new_fl);
1466
	spin_unlock(&inode->i_lock);
1467
	locks_dispose_list(&dispose);
1468 1469
	error = wait_event_interruptible_timeout(new_fl->fl_wait,
						!new_fl->fl_next, break_time);
1470
	spin_lock(&inode->i_lock);
1471
	trace_break_lease_unblock(inode, new_fl);
1472
	locks_delete_block(new_fl);
L
Linus Torvalds 已提交
1473
	if (error >= 0) {
1474 1475 1476 1477
		/*
		 * Wait for the next conflicting lease that has not been
		 * broken yet
		 */
1478 1479 1480 1481
		if (error == 0)
			time_out_leases(inode, &dispose);
		if (any_leases_conflict(inode, new_fl))
			goto restart;
L
Linus Torvalds 已提交
1482 1483 1484
		error = 0;
	}
out:
1485
	spin_unlock(&inode->i_lock);
1486
	locks_dispose_list(&dispose);
1487
	locks_free_lock(new_fl);
L
Linus Torvalds 已提交
1488 1489 1490 1491 1492 1493
	return error;
}

EXPORT_SYMBOL(__break_lease);

/**
1494
 *	lease_get_mtime - get the last modified time of an inode
L
Linus Torvalds 已提交
1495 1496 1497 1498 1499
 *	@inode: the inode
 *      @time:  pointer to a timespec which will contain the last modified time
 *
 * This is to force NFS clients to flush their caches for files with
 * exclusive leases.  The justification is that if someone has an
1500
 * exclusive lease, then they could be modifying it.
L
Linus Torvalds 已提交
1501 1502 1503
 */
void lease_get_mtime(struct inode *inode, struct timespec *time)
{
1504
	bool has_lease = false;
1505 1506
	struct file_lock_context *ctx = inode->i_flctx;
	struct file_lock *fl;
1507

1508
	if (ctx && !list_empty_careful(&ctx->flc_lease)) {
1509
		spin_lock(&inode->i_lock);
1510 1511 1512 1513 1514 1515
		if (!list_empty(&ctx->flc_lease)) {
			fl = list_first_entry(&ctx->flc_lease,
						struct file_lock, fl_list);
			if (fl->fl_type == F_WRLCK)
				has_lease = true;
		}
1516 1517 1518 1519
		spin_unlock(&inode->i_lock);
	}

	if (has_lease)
L
Linus Torvalds 已提交
1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552
		*time = current_fs_time(inode->i_sb);
	else
		*time = inode->i_mtime;
}

EXPORT_SYMBOL(lease_get_mtime);

/**
 *	fcntl_getlease - Enquire what lease is currently active
 *	@filp: the file
 *
 *	The value returned by this function will be one of
 *	(if no lease break is pending):
 *
 *	%F_RDLCK to indicate a shared lease is held.
 *
 *	%F_WRLCK to indicate an exclusive lease is held.
 *
 *	%F_UNLCK to indicate no lease is held.
 *
 *	(if a lease break is pending):
 *
 *	%F_RDLCK to indicate an exclusive lease needs to be
 *		changed to a shared lease (or removed).
 *
 *	%F_UNLCK to indicate the lease needs to be removed.
 *
 *	XXX: sfr & willy disagree over whether F_INPROGRESS
 *	should be returned to userspace.
 */
int fcntl_getlease(struct file *filp)
{
	struct file_lock *fl;
1553
	struct inode *inode = file_inode(filp);
1554
	struct file_lock_context *ctx = inode->i_flctx;
L
Linus Torvalds 已提交
1555
	int type = F_UNLCK;
1556
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
1557

1558 1559 1560 1561 1562 1563
	if (ctx && !list_empty_careful(&ctx->flc_lease)) {
		spin_lock(&inode->i_lock);
		time_out_leases(file_inode(filp), &dispose);
		list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
			if (fl->fl_file != filp)
				continue;
1564
			type = target_leasetype(fl);
L
Linus Torvalds 已提交
1565 1566
			break;
		}
1567 1568
		spin_unlock(&inode->i_lock);
		locks_dispose_list(&dispose);
L
Linus Torvalds 已提交
1569 1570 1571 1572
	}
	return type;
}

1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598
/**
 * check_conflicting_open - see if the given dentry points to a file that has
 * 			    an existing open that would conflict with the
 * 			    desired lease.
 * @dentry:	dentry to check
 * @arg:	type of lease that we're trying to acquire
 *
 * Check to see if there's an existing open fd on this file that would
 * conflict with the lease we're trying to set.
 */
static int
check_conflicting_open(const struct dentry *dentry, const long arg)
{
	int ret = 0;
	struct inode *inode = dentry->d_inode;

	if ((arg == F_RDLCK) && (atomic_read(&inode->i_writecount) > 0))
		return -EAGAIN;

	if ((arg == F_WRLCK) && ((d_count(dentry) > 1) ||
	    (atomic_read(&inode->i_count) > 1)))
		ret = -EAGAIN;

	return ret;
}

1599 1600
static int
generic_add_lease(struct file *filp, long arg, struct file_lock **flp, void **priv)
L
Linus Torvalds 已提交
1601
{
1602
	struct file_lock *fl, *my_fl = NULL, *lease;
1603
	struct dentry *dentry = filp->f_path.dentry;
L
Linus Torvalds 已提交
1604
	struct inode *inode = dentry->d_inode;
1605
	struct file_lock_context *ctx;
J
J. Bruce Fields 已提交
1606
	bool is_deleg = (*flp)->fl_flags & FL_DELEG;
J
J. Bruce Fields 已提交
1607
	int error;
1608
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
1609

1610
	lease = *flp;
1611 1612
	trace_generic_add_lease(inode, lease);

1613 1614 1615 1616
	ctx = locks_get_lock_context(inode);
	if (!ctx)
		return -ENOMEM;

J
J. Bruce Fields 已提交
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629
	/*
	 * In the delegation case we need mutual exclusion with
	 * a number of operations that take the i_mutex.  We trylock
	 * because delegations are an optional optimization, and if
	 * there's some chance of a conflict--we'd rather not
	 * bother, maybe that's a sign this just isn't a good file to
	 * hand out a delegation on.
	 */
	if (is_deleg && !mutex_trylock(&inode->i_mutex))
		return -EAGAIN;

	if (is_deleg && arg == F_WRLCK) {
		/* Write delegations are not currently supported: */
1630
		mutex_unlock(&inode->i_mutex);
J
J. Bruce Fields 已提交
1631 1632 1633
		WARN_ON_ONCE(1);
		return -EINVAL;
	}
1634

1635
	spin_lock(&inode->i_lock);
1636
	time_out_leases(inode, &dispose);
1637 1638
	error = check_conflicting_open(dentry, arg);
	if (error)
1639
		goto out;
1640

L
Linus Torvalds 已提交
1641 1642 1643 1644 1645 1646 1647 1648
	/*
	 * At this point, we know that if there is an exclusive
	 * lease on this file, then we hold it on this filp
	 * (otherwise our open of this file would have blocked).
	 * And if we are trying to acquire an exclusive lease,
	 * then the file is not open by anyone (including us)
	 * except for this filp.
	 */
J
J. Bruce Fields 已提交
1649
	error = -EAGAIN;
1650
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
J
J. Bruce Fields 已提交
1651
		if (fl->fl_file == filp) {
1652
			my_fl = fl;
J
J. Bruce Fields 已提交
1653 1654
			continue;
		}
1655

J
J. Bruce Fields 已提交
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667
		/*
		 * No exclusive leases if someone else has a lease on
		 * this file:
		 */
		if (arg == F_WRLCK)
			goto out;
		/*
		 * Modifying our existing lease is OK, but no getting a
		 * new lease if someone else is opening for write:
		 */
		if (fl->fl_flags & FL_UNLOCK_PENDING)
			goto out;
L
Linus Torvalds 已提交
1668 1669
	}

1670 1671
	if (my_fl != NULL) {
		error = lease->fl_lmops->lm_change(&my_fl, arg, &dispose);
1672 1673 1674
		if (error)
			goto out;
		goto out_setup;
L
Linus Torvalds 已提交
1675 1676 1677 1678 1679 1680
	}

	error = -EINVAL;
	if (!leases_enable)
		goto out;

1681
	locks_insert_lock_ctx(lease, &ctx->flc_lease);
1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
	/*
	 * The check in break_lease() is lockless. It's possible for another
	 * open to race in after we did the earlier check for a conflicting
	 * open but before the lease was inserted. Check again for a
	 * conflicting open and cancel the lease if there is one.
	 *
	 * We also add a barrier here to ensure that the insertion of the lock
	 * precedes these checks.
	 */
	smp_mb();
	error = check_conflicting_open(dentry, arg);
1693 1694 1695 1696
	if (error) {
		locks_unlink_lock_ctx(lease);
		goto out;
	}
1697 1698 1699 1700

out_setup:
	if (lease->fl_lmops->lm_setup)
		lease->fl_lmops->lm_setup(lease, priv);
L
Linus Torvalds 已提交
1701
out:
1702
	spin_unlock(&inode->i_lock);
1703
	locks_dispose_list(&dispose);
J
J. Bruce Fields 已提交
1704 1705
	if (is_deleg)
		mutex_unlock(&inode->i_mutex);
1706
	if (!error && !my_fl)
1707
		*flp = NULL;
L
Linus Torvalds 已提交
1708 1709
	return error;
}
1710

1711
static int generic_delete_lease(struct file *filp)
1712
{
1713
	int error = -EAGAIN;
1714
	struct file_lock *fl, *victim = NULL;
1715 1716
	struct dentry *dentry = filp->f_path.dentry;
	struct inode *inode = dentry->d_inode;
1717
	struct file_lock_context *ctx = inode->i_flctx;
1718
	LIST_HEAD(dispose);
1719

1720 1721 1722 1723 1724
	if (!ctx) {
		trace_generic_delete_lease(inode, NULL);
		return error;
	}

1725
	spin_lock(&inode->i_lock);
1726 1727 1728
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
		if (fl->fl_file == filp) {
			victim = fl;
1729
			break;
1730
		}
1731
	}
1732
	trace_generic_delete_lease(inode, fl);
1733 1734
	if (victim)
		error = fl->fl_lmops->lm_change(&victim, F_UNLCK, &dispose);
1735
	spin_unlock(&inode->i_lock);
1736
	locks_dispose_list(&dispose);
1737
	return error;
1738 1739 1740 1741
}

/**
 *	generic_setlease	-	sets a lease on an open file
1742 1743 1744 1745 1746
 *	@filp:	file pointer
 *	@arg:	type of lease to obtain
 *	@flp:	input - file_lock to use, output - file_lock inserted
 *	@priv:	private data for lm_setup (may be NULL if lm_setup
 *		doesn't require it)
1747 1748 1749 1750
 *
 *	The (input) flp->fl_lmops->lm_break function is required
 *	by break_lease().
 */
1751 1752
int generic_setlease(struct file *filp, long arg, struct file_lock **flp,
			void **priv)
1753 1754 1755 1756 1757
{
	struct dentry *dentry = filp->f_path.dentry;
	struct inode *inode = dentry->d_inode;
	int error;

1758
	if ((!uid_eq(current_fsuid(), inode->i_uid)) && !capable(CAP_LEASE))
1759 1760 1761 1762 1763 1764 1765 1766 1767
		return -EACCES;
	if (!S_ISREG(inode->i_mode))
		return -EINVAL;
	error = security_file_lock(filp, arg);
	if (error)
		return error;

	switch (arg) {
	case F_UNLCK:
1768
		return generic_delete_lease(filp);
1769 1770
	case F_RDLCK:
	case F_WRLCK:
1771 1772 1773 1774
		if (!(*flp)->fl_lmops->lm_break) {
			WARN_ON_ONCE(1);
			return -ENOLCK;
		}
1775
		return generic_add_lease(filp, arg, flp, priv);
1776
	default:
1777
		return -EINVAL;
1778 1779
	}
}
1780
EXPORT_SYMBOL(generic_setlease);
L
Linus Torvalds 已提交
1781

1782
/**
1783
 * vfs_setlease        -       sets a lease on an open file
1784 1785 1786 1787 1788
 * @filp:	file pointer
 * @arg:	type of lease to obtain
 * @lease:	file_lock to use when adding a lease
 * @priv:	private info for lm_setup when adding a lease (may be
 * 		NULL if lm_setup doesn't require it)
1789 1790 1791 1792 1793 1794
 *
 * Call this to establish a lease on the file. The "lease" argument is not
 * used for F_UNLCK requests and may be NULL. For commands that set or alter
 * an existing lease, the (*lease)->fl_lmops->lm_break operation must be set;
 * if not, this function will return -ENOLCK (and generate a scary-looking
 * stack trace).
1795 1796 1797
 *
 * The "priv" pointer is passed directly to the lm_setup function as-is. It
 * may be NULL if the lm_setup operation doesn't require it.
L
Linus Torvalds 已提交
1798
 */
1799 1800
int
vfs_setlease(struct file *filp, long arg, struct file_lock **lease, void **priv)
L
Linus Torvalds 已提交
1801
{
1802
	if (filp->f_op->setlease)
1803
		return filp->f_op->setlease(filp, arg, lease, priv);
1804
	else
1805
		return generic_setlease(filp, arg, lease, priv);
L
Linus Torvalds 已提交
1806
}
1807
EXPORT_SYMBOL_GPL(vfs_setlease);
L
Linus Torvalds 已提交
1808

1809
static int do_fcntl_add_lease(unsigned int fd, struct file *filp, long arg)
L
Linus Torvalds 已提交
1810
{
1811
	struct file_lock *fl;
1812
	struct fasync_struct *new;
L
Linus Torvalds 已提交
1813 1814
	int error;

1815 1816 1817
	fl = lease_alloc(filp, arg);
	if (IS_ERR(fl))
		return PTR_ERR(fl);
L
Linus Torvalds 已提交
1818

1819 1820 1821 1822 1823
	new = fasync_alloc();
	if (!new) {
		locks_free_lock(fl);
		return -ENOMEM;
	}
1824
	new->fa_fd = fd;
1825

1826
	error = vfs_setlease(filp, arg, &fl, (void **)&new);
1827 1828
	if (fl)
		locks_free_lock(fl);
1829 1830
	if (new)
		fasync_free(new);
L
Linus Torvalds 已提交
1831 1832 1833
	return error;
}

1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846
/**
 *	fcntl_setlease	-	sets a lease on an open file
 *	@fd: open file descriptor
 *	@filp: file pointer
 *	@arg: type of lease to obtain
 *
 *	Call this fcntl to establish a lease on the file.
 *	Note that you also need to call %F_SETSIG to
 *	receive a signal when the lease is broken.
 */
int fcntl_setlease(unsigned int fd, struct file *filp, long arg)
{
	if (arg == F_UNLCK)
1847
		return vfs_setlease(filp, F_UNLCK, NULL, NULL);
1848 1849 1850
	return do_fcntl_add_lease(fd, filp, arg);
}

L
Linus Torvalds 已提交
1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863
/**
 * flock_lock_file_wait - Apply a FLOCK-style lock to a file
 * @filp: The file to apply the lock to
 * @fl: The lock to be applied
 *
 * Add a FLOCK style lock to a file.
 */
int flock_lock_file_wait(struct file *filp, struct file_lock *fl)
{
	int error;
	might_sleep();
	for (;;) {
		error = flock_lock_file(filp, fl);
1864
		if (error != FILE_LOCK_DEFERRED)
L
Linus Torvalds 已提交
1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896
			break;
		error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
		if (!error)
			continue;

		locks_delete_block(fl);
		break;
	}
	return error;
}

EXPORT_SYMBOL(flock_lock_file_wait);

/**
 *	sys_flock: - flock() system call.
 *	@fd: the file descriptor to lock.
 *	@cmd: the type of lock to apply.
 *
 *	Apply a %FL_FLOCK style lock to an open file descriptor.
 *	The @cmd can be one of
 *
 *	%LOCK_SH -- a shared lock.
 *
 *	%LOCK_EX -- an exclusive lock.
 *
 *	%LOCK_UN -- remove an existing lock.
 *
 *	%LOCK_MAND -- a `mandatory' flock.  This exists to emulate Windows Share Modes.
 *
 *	%LOCK_MAND can be combined with %LOCK_READ or %LOCK_WRITE to allow other
 *	processes read and write access respectively.
 */
1897
SYSCALL_DEFINE2(flock, unsigned int, fd, unsigned int, cmd)
L
Linus Torvalds 已提交
1898
{
1899
	struct fd f = fdget(fd);
L
Linus Torvalds 已提交
1900 1901 1902 1903 1904
	struct file_lock *lock;
	int can_sleep, unlock;
	int error;

	error = -EBADF;
1905
	if (!f.file)
L
Linus Torvalds 已提交
1906 1907 1908 1909 1910 1911
		goto out;

	can_sleep = !(cmd & LOCK_NB);
	cmd &= ~LOCK_NB;
	unlock = (cmd == LOCK_UN);

1912
	if (!unlock && !(cmd & LOCK_MAND) &&
1913
	    !(f.file->f_mode & (FMODE_READ|FMODE_WRITE)))
L
Linus Torvalds 已提交
1914 1915
		goto out_putf;

1916 1917 1918
	lock = flock_make_lock(f.file, cmd);
	if (IS_ERR(lock)) {
		error = PTR_ERR(lock);
L
Linus Torvalds 已提交
1919
		goto out_putf;
1920 1921
	}

L
Linus Torvalds 已提交
1922 1923 1924
	if (can_sleep)
		lock->fl_flags |= FL_SLEEP;

1925
	error = security_file_lock(f.file, lock->fl_type);
L
Linus Torvalds 已提交
1926 1927 1928
	if (error)
		goto out_free;

A
Al Viro 已提交
1929
	if (f.file->f_op->flock)
1930
		error = f.file->f_op->flock(f.file,
L
Linus Torvalds 已提交
1931 1932 1933
					  (can_sleep) ? F_SETLKW : F_SETLK,
					  lock);
	else
1934
		error = flock_lock_file_wait(f.file, lock);
L
Linus Torvalds 已提交
1935 1936

 out_free:
1937
	locks_free_lock(lock);
L
Linus Torvalds 已提交
1938 1939

 out_putf:
1940
	fdput(f);
L
Linus Torvalds 已提交
1941 1942 1943 1944
 out:
	return error;
}

1945 1946 1947
/**
 * vfs_test_lock - test file byte range lock
 * @filp: The file to test lock for
1948
 * @fl: The lock to test; also used to hold result
1949 1950 1951 1952 1953 1954
 *
 * Returns -ERRNO on failure.  Indicates presence of conflicting lock by
 * setting conf->fl_type to something other than F_UNLCK.
 */
int vfs_test_lock(struct file *filp, struct file_lock *fl)
{
A
Al Viro 已提交
1955
	if (filp->f_op->lock)
1956 1957 1958 1959 1960 1961
		return filp->f_op->lock(filp, F_GETLK, fl);
	posix_test_lock(filp, fl);
	return 0;
}
EXPORT_SYMBOL_GPL(vfs_test_lock);

1962 1963
static int posix_lock_to_flock(struct flock *flock, struct file_lock *fl)
{
1964
	flock->l_pid = IS_OFDLCK(fl) ? -1 : fl->fl_pid;
1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978
#if BITS_PER_LONG == 32
	/*
	 * Make sure we can represent the posix lock via
	 * legacy 32bit flock.
	 */
	if (fl->fl_start > OFFT_OFFSET_MAX)
		return -EOVERFLOW;
	if (fl->fl_end != OFFSET_MAX && fl->fl_end > OFFT_OFFSET_MAX)
		return -EOVERFLOW;
#endif
	flock->l_start = fl->fl_start;
	flock->l_len = fl->fl_end == OFFSET_MAX ? 0 :
		fl->fl_end - fl->fl_start + 1;
	flock->l_whence = 0;
1979
	flock->l_type = fl->fl_type;
1980 1981 1982 1983 1984 1985
	return 0;
}

#if BITS_PER_LONG == 32
static void posix_lock_to_flock64(struct flock64 *flock, struct file_lock *fl)
{
1986
	flock->l_pid = IS_OFDLCK(fl) ? -1 : fl->fl_pid;
1987 1988 1989 1990 1991 1992 1993 1994
	flock->l_start = fl->fl_start;
	flock->l_len = fl->fl_end == OFFSET_MAX ? 0 :
		fl->fl_end - fl->fl_start + 1;
	flock->l_whence = 0;
	flock->l_type = fl->fl_type;
}
#endif

L
Linus Torvalds 已提交
1995 1996 1997
/* Report the first existing lock that would conflict with l.
 * This implements the F_GETLK command of fcntl().
 */
1998
int fcntl_getlk(struct file *filp, unsigned int cmd, struct flock __user *l)
L
Linus Torvalds 已提交
1999
{
2000
	struct file_lock file_lock;
L
Linus Torvalds 已提交
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
	struct flock flock;
	int error;

	error = -EFAULT;
	if (copy_from_user(&flock, l, sizeof(flock)))
		goto out;
	error = -EINVAL;
	if ((flock.l_type != F_RDLCK) && (flock.l_type != F_WRLCK))
		goto out;

	error = flock_to_posix_lock(filp, &file_lock, &flock);
	if (error)
		goto out;

2015
	if (cmd == F_OFD_GETLK) {
2016 2017 2018 2019
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2020
		cmd = F_GETLK;
2021
		file_lock.fl_flags |= FL_OFDLCK;
2022
		file_lock.fl_owner = filp;
2023 2024
	}

2025 2026 2027
	error = vfs_test_lock(filp, &file_lock);
	if (error)
		goto out;
L
Linus Torvalds 已提交
2028
 
2029 2030 2031
	flock.l_type = file_lock.fl_type;
	if (file_lock.fl_type != F_UNLCK) {
		error = posix_lock_to_flock(&flock, &file_lock);
2032
		if (error)
2033
			goto rel_priv;
L
Linus Torvalds 已提交
2034 2035 2036 2037
	}
	error = -EFAULT;
	if (!copy_to_user(l, &flock, sizeof(flock)))
		error = 0;
2038 2039
rel_priv:
	locks_release_private(&file_lock);
L
Linus Torvalds 已提交
2040 2041 2042 2043
out:
	return error;
}

2044 2045 2046 2047 2048
/**
 * vfs_lock_file - file byte range lock
 * @filp: The file to apply the lock to
 * @cmd: type of locking operation (F_SETLK, F_GETLK, etc.)
 * @fl: The lock to be applied
2049 2050 2051 2052 2053 2054 2055 2056
 * @conf: Place to return a copy of the conflicting lock, if found.
 *
 * A caller that doesn't care about the conflicting lock may pass NULL
 * as the final argument.
 *
 * If the filesystem defines a private ->lock() method, then @conf will
 * be left unchanged; so a caller that cares should initialize it to
 * some acceptable default.
2057 2058 2059 2060
 *
 * To avoid blocking kernel daemons, such as lockd, that need to acquire POSIX
 * locks, the ->lock() interface may return asynchronously, before the lock has
 * been granted or denied by the underlying filesystem, if (and only if)
J
J. Bruce Fields 已提交
2061
 * lm_grant is set. Callers expecting ->lock() to return asynchronously
2062 2063
 * will only use F_SETLK, not F_SETLKW; they will set FL_SLEEP if (and only if)
 * the request is for a blocking lock. When ->lock() does return asynchronously,
J
J. Bruce Fields 已提交
2064
 * it must return FILE_LOCK_DEFERRED, and call ->lm_grant() when the lock
2065 2066
 * request completes.
 * If the request is for non-blocking lock the file system should return
2067 2068
 * FILE_LOCK_DEFERRED then try to get the lock and call the callback routine
 * with the result. If the request timed out the callback routine will return a
2069 2070 2071 2072 2073
 * nonzero return code and the file system should release the lock. The file
 * system is also responsible to keep a corresponding posix lock when it
 * grants a lock so the VFS can find out which locks are locally held and do
 * the correct lock cleanup when required.
 * The underlying filesystem must not drop the kernel lock or call
J
J. Bruce Fields 已提交
2074
 * ->lm_grant() before returning to the caller with a FILE_LOCK_DEFERRED
2075
 * return code.
2076
 */
2077
int vfs_lock_file(struct file *filp, unsigned int cmd, struct file_lock *fl, struct file_lock *conf)
2078
{
A
Al Viro 已提交
2079
	if (filp->f_op->lock)
2080 2081
		return filp->f_op->lock(filp, cmd, fl);
	else
2082
		return posix_lock_file(filp, fl, conf);
2083 2084 2085
}
EXPORT_SYMBOL_GPL(vfs_lock_file);

M
Miklos Szeredi 已提交
2086 2087 2088 2089 2090 2091 2092 2093 2094
static int do_lock_file_wait(struct file *filp, unsigned int cmd,
			     struct file_lock *fl)
{
	int error;

	error = security_file_lock(filp, fl->fl_type);
	if (error)
		return error;

2095 2096 2097
	for (;;) {
		error = vfs_lock_file(filp, cmd, fl, NULL);
		if (error != FILE_LOCK_DEFERRED)
M
Miklos Szeredi 已提交
2098
			break;
2099 2100 2101 2102 2103 2104
		error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
		if (!error)
			continue;

		locks_delete_block(fl);
		break;
M
Miklos Szeredi 已提交
2105 2106 2107 2108 2109
	}

	return error;
}

2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125
/* Ensure that fl->fl_filp has compatible f_mode for F_SETLK calls */
static int
check_fmode_for_setlk(struct file_lock *fl)
{
	switch (fl->fl_type) {
	case F_RDLCK:
		if (!(fl->fl_file->f_mode & FMODE_READ))
			return -EBADF;
		break;
	case F_WRLCK:
		if (!(fl->fl_file->f_mode & FMODE_WRITE))
			return -EBADF;
	}
	return 0;
}

L
Linus Torvalds 已提交
2126 2127 2128
/* Apply the lock described by l to an open file descriptor.
 * This implements both the F_SETLK and F_SETLKW commands of fcntl().
 */
2129 2130
int fcntl_setlk(unsigned int fd, struct file *filp, unsigned int cmd,
		struct flock __user *l)
L
Linus Torvalds 已提交
2131 2132 2133 2134
{
	struct file_lock *file_lock = locks_alloc_lock();
	struct flock flock;
	struct inode *inode;
2135
	struct file *f;
L
Linus Torvalds 已提交
2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147
	int error;

	if (file_lock == NULL)
		return -ENOLCK;

	/*
	 * This might block, so we do it before checking the inode.
	 */
	error = -EFAULT;
	if (copy_from_user(&flock, l, sizeof(flock)))
		goto out;

A
Al Viro 已提交
2148
	inode = file_inode(filp);
L
Linus Torvalds 已提交
2149 2150 2151 2152

	/* Don't allow mandatory locks on files that may be memory mapped
	 * and shared.
	 */
2153
	if (mandatory_lock(inode) && mapping_writably_mapped(filp->f_mapping)) {
L
Linus Torvalds 已提交
2154 2155 2156 2157
		error = -EAGAIN;
		goto out;
	}

2158
again:
L
Linus Torvalds 已提交
2159 2160 2161
	error = flock_to_posix_lock(filp, file_lock, &flock);
	if (error)
		goto out;
2162

2163 2164 2165 2166
	error = check_fmode_for_setlk(file_lock);
	if (error)
		goto out;

2167 2168
	/*
	 * If the cmd is requesting file-private locks, then set the
2169
	 * FL_OFDLCK flag and override the owner.
2170 2171
	 */
	switch (cmd) {
2172
	case F_OFD_SETLK:
2173 2174 2175 2176
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2177
		cmd = F_SETLK;
2178
		file_lock->fl_flags |= FL_OFDLCK;
2179
		file_lock->fl_owner = filp;
2180
		break;
2181
	case F_OFD_SETLKW:
2182 2183 2184 2185
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2186
		cmd = F_SETLKW;
2187
		file_lock->fl_flags |= FL_OFDLCK;
2188
		file_lock->fl_owner = filp;
2189 2190
		/* Fallthrough */
	case F_SETLKW:
L
Linus Torvalds 已提交
2191 2192
		file_lock->fl_flags |= FL_SLEEP;
	}
2193

M
Miklos Szeredi 已提交
2194
	error = do_lock_file_wait(filp, cmd, file_lock);
L
Linus Torvalds 已提交
2195

2196 2197 2198 2199
	/*
	 * Attempt to detect a close/fcntl race and recover by
	 * releasing the lock that was just acquired.
	 */
2200 2201 2202 2203 2204 2205 2206 2207 2208
	/*
	 * we need that spin_lock here - it prevents reordering between
	 * update of inode->i_flock and check for it done in close().
	 * rcu_read_lock() wouldn't do.
	 */
	spin_lock(&current->files->file_lock);
	f = fcheck(fd);
	spin_unlock(&current->files->file_lock);
	if (!error && f != filp && flock.l_type != F_UNLCK) {
2209 2210
		flock.l_type = F_UNLCK;
		goto again;
L
Linus Torvalds 已提交
2211 2212
	}

2213
out:
L
Linus Torvalds 已提交
2214 2215 2216 2217 2218 2219 2220 2221
	locks_free_lock(file_lock);
	return error;
}

#if BITS_PER_LONG == 32
/* Report the first existing lock that would conflict with l.
 * This implements the F_GETLK command of fcntl().
 */
2222
int fcntl_getlk64(struct file *filp, unsigned int cmd, struct flock64 __user *l)
L
Linus Torvalds 已提交
2223
{
2224
	struct file_lock file_lock;
L
Linus Torvalds 已提交
2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238
	struct flock64 flock;
	int error;

	error = -EFAULT;
	if (copy_from_user(&flock, l, sizeof(flock)))
		goto out;
	error = -EINVAL;
	if ((flock.l_type != F_RDLCK) && (flock.l_type != F_WRLCK))
		goto out;

	error = flock64_to_posix_lock(filp, &file_lock, &flock);
	if (error)
		goto out;

2239
	if (cmd == F_OFD_GETLK) {
2240 2241 2242 2243
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2244
		cmd = F_GETLK64;
2245
		file_lock.fl_flags |= FL_OFDLCK;
2246
		file_lock.fl_owner = filp;
2247 2248
	}

2249 2250 2251 2252
	error = vfs_test_lock(filp, &file_lock);
	if (error)
		goto out;

2253 2254 2255 2256
	flock.l_type = file_lock.fl_type;
	if (file_lock.fl_type != F_UNLCK)
		posix_lock_to_flock64(&flock, &file_lock);

L
Linus Torvalds 已提交
2257 2258 2259
	error = -EFAULT;
	if (!copy_to_user(l, &flock, sizeof(flock)))
		error = 0;
2260 2261

	locks_release_private(&file_lock);
L
Linus Torvalds 已提交
2262 2263 2264 2265 2266 2267 2268
out:
	return error;
}

/* Apply the lock described by l to an open file descriptor.
 * This implements both the F_SETLK and F_SETLKW commands of fcntl().
 */
2269 2270
int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd,
		struct flock64 __user *l)
L
Linus Torvalds 已提交
2271 2272 2273 2274
{
	struct file_lock *file_lock = locks_alloc_lock();
	struct flock64 flock;
	struct inode *inode;
2275
	struct file *f;
L
Linus Torvalds 已提交
2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287
	int error;

	if (file_lock == NULL)
		return -ENOLCK;

	/*
	 * This might block, so we do it before checking the inode.
	 */
	error = -EFAULT;
	if (copy_from_user(&flock, l, sizeof(flock)))
		goto out;

A
Al Viro 已提交
2288
	inode = file_inode(filp);
L
Linus Torvalds 已提交
2289 2290 2291 2292

	/* Don't allow mandatory locks on files that may be memory mapped
	 * and shared.
	 */
2293
	if (mandatory_lock(inode) && mapping_writably_mapped(filp->f_mapping)) {
L
Linus Torvalds 已提交
2294 2295 2296 2297
		error = -EAGAIN;
		goto out;
	}

2298
again:
L
Linus Torvalds 已提交
2299 2300 2301
	error = flock64_to_posix_lock(filp, file_lock, &flock);
	if (error)
		goto out;
2302

2303 2304 2305 2306
	error = check_fmode_for_setlk(file_lock);
	if (error)
		goto out;

2307 2308
	/*
	 * If the cmd is requesting file-private locks, then set the
2309
	 * FL_OFDLCK flag and override the owner.
2310 2311
	 */
	switch (cmd) {
2312
	case F_OFD_SETLK:
2313 2314 2315 2316
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2317
		cmd = F_SETLK64;
2318
		file_lock->fl_flags |= FL_OFDLCK;
2319
		file_lock->fl_owner = filp;
2320
		break;
2321
	case F_OFD_SETLKW:
2322 2323 2324 2325
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2326
		cmd = F_SETLKW64;
2327
		file_lock->fl_flags |= FL_OFDLCK;
2328
		file_lock->fl_owner = filp;
2329 2330
		/* Fallthrough */
	case F_SETLKW64:
L
Linus Torvalds 已提交
2331 2332
		file_lock->fl_flags |= FL_SLEEP;
	}
2333

M
Miklos Szeredi 已提交
2334
	error = do_lock_file_wait(filp, cmd, file_lock);
L
Linus Torvalds 已提交
2335

2336 2337 2338 2339
	/*
	 * Attempt to detect a close/fcntl race and recover by
	 * releasing the lock that was just acquired.
	 */
2340 2341 2342 2343
	spin_lock(&current->files->file_lock);
	f = fcheck(fd);
	spin_unlock(&current->files->file_lock);
	if (!error && f != filp && flock.l_type != F_UNLCK) {
2344 2345
		flock.l_type = F_UNLCK;
		goto again;
L
Linus Torvalds 已提交
2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360
	}

out:
	locks_free_lock(file_lock);
	return error;
}
#endif /* BITS_PER_LONG == 32 */

/*
 * This function is called when the file is being removed
 * from the task's fd array.  POSIX locks belonging to this task
 * are deleted at this time.
 */
void locks_remove_posix(struct file *filp, fl_owner_t owner)
{
2361
	struct file_lock lock;
2362
	struct file_lock_context *ctx = file_inode(filp)->i_flctx;
L
Linus Torvalds 已提交
2363 2364 2365 2366 2367 2368

	/*
	 * If there are no locks held on this file, we don't need to call
	 * posix_lock_file().  Another process could be setting a lock on this
	 * file at the same time, but we wouldn't remove that lock anyway.
	 */
2369
	if (!ctx || list_empty(&ctx->flc_posix))
L
Linus Torvalds 已提交
2370 2371 2372
		return;

	lock.fl_type = F_UNLCK;
2373
	lock.fl_flags = FL_POSIX | FL_CLOSE;
L
Linus Torvalds 已提交
2374 2375 2376 2377 2378 2379 2380 2381
	lock.fl_start = 0;
	lock.fl_end = OFFSET_MAX;
	lock.fl_owner = owner;
	lock.fl_pid = current->tgid;
	lock.fl_file = filp;
	lock.fl_ops = NULL;
	lock.fl_lmops = NULL;

2382
	vfs_lock_file(filp, F_SETLK, &lock, NULL);
L
Linus Torvalds 已提交
2383 2384 2385 2386 2387 2388 2389

	if (lock.fl_ops && lock.fl_ops->fl_release_private)
		lock.fl_ops->fl_release_private(&lock);
}

EXPORT_SYMBOL(locks_remove_posix);

2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400
static void
locks_remove_flock(struct file *filp)
{
	struct file_lock fl = {
		.fl_owner = filp,
		.fl_pid = current->tgid,
		.fl_file = filp,
		.fl_flags = FL_FLOCK,
		.fl_type = F_UNLCK,
		.fl_end = OFFSET_MAX,
	};
2401
	struct file_lock_context *flctx = file_inode(filp)->i_flctx;
2402

2403
	if (!flctx || list_empty(&flctx->flc_flock))
2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414
		return;

	if (filp->f_op->flock)
		filp->f_op->flock(filp, F_SETLKW, &fl);
	else
		flock_lock_file(filp, &fl);

	if (fl.fl_ops && fl.fl_ops->fl_release_private)
		fl.fl_ops->fl_release_private(&fl);
}

2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432
static void
locks_remove_lease(struct file *filp)
{
	struct inode *inode = file_inode(filp);
	struct file_lock_context *ctx = inode->i_flctx;
	struct file_lock *fl, *tmp;
	LIST_HEAD(dispose);

	if (!ctx || list_empty(&ctx->flc_lease))
		return;

	spin_lock(&inode->i_lock);
	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, fl_list)
		lease_modify(&fl, F_UNLCK, &dispose);
	spin_unlock(&inode->i_lock);
	locks_dispose_list(&dispose);
}

L
Linus Torvalds 已提交
2433 2434 2435
/*
 * This function is called on the last close of an open file.
 */
2436
void locks_remove_file(struct file *filp)
L
Linus Torvalds 已提交
2437
{
2438
	/* remove any OFD locks */
2439
	locks_remove_posix(filp, filp);
2440

2441 2442 2443
	/* remove flock locks */
	locks_remove_flock(filp);

2444 2445
	/* remove any leases */
	locks_remove_lease(filp);
L
Linus Torvalds 已提交
2446 2447 2448 2449 2450 2451 2452 2453
}

/**
 *	posix_unblock_lock - stop waiting for a file lock
 *	@waiter: the lock which was waiting
 *
 *	lockd needs to block waiting for locks.
 */
J
J. Bruce Fields 已提交
2454
int
2455
posix_unblock_lock(struct file_lock *waiter)
L
Linus Torvalds 已提交
2456
{
J
J. Bruce Fields 已提交
2457 2458
	int status = 0;

2459
	spin_lock(&blocked_lock_lock);
2460
	if (waiter->fl_next)
L
Linus Torvalds 已提交
2461
		__locks_delete_block(waiter);
J
J. Bruce Fields 已提交
2462 2463
	else
		status = -ENOENT;
2464
	spin_unlock(&blocked_lock_lock);
J
J. Bruce Fields 已提交
2465
	return status;
L
Linus Torvalds 已提交
2466 2467 2468
}
EXPORT_SYMBOL(posix_unblock_lock);

M
Marc Eshel 已提交
2469 2470 2471 2472 2473 2474 2475 2476 2477
/**
 * vfs_cancel_lock - file byte range unblock lock
 * @filp: The file to apply the unblock to
 * @fl: The lock to be unblocked
 *
 * Used by lock managers to cancel blocked requests
 */
int vfs_cancel_lock(struct file *filp, struct file_lock *fl)
{
A
Al Viro 已提交
2478
	if (filp->f_op->lock)
M
Marc Eshel 已提交
2479 2480 2481 2482 2483 2484
		return filp->f_op->lock(filp, F_CANCELLK, fl);
	return 0;
}

EXPORT_SYMBOL_GPL(vfs_cancel_lock);

2485
#ifdef CONFIG_PROC_FS
2486
#include <linux/proc_fs.h>
2487 2488
#include <linux/seq_file.h>

2489 2490 2491 2492 2493
struct locks_iterator {
	int	li_cpu;
	loff_t	li_pos;
};

2494
static void lock_get_status(struct seq_file *f, struct file_lock *fl,
2495
			    loff_t id, char *pfx)
L
Linus Torvalds 已提交
2496 2497
{
	struct inode *inode = NULL;
2498 2499 2500
	unsigned int fl_pid;

	if (fl->fl_nspid)
2501
		fl_pid = pid_vnr(fl->fl_nspid);
2502 2503
	else
		fl_pid = fl->fl_pid;
L
Linus Torvalds 已提交
2504 2505

	if (fl->fl_file != NULL)
A
Al Viro 已提交
2506
		inode = file_inode(fl->fl_file);
L
Linus Torvalds 已提交
2507

2508
	seq_printf(f, "%lld:%s ", id, pfx);
L
Linus Torvalds 已提交
2509
	if (IS_POSIX(fl)) {
2510
		if (fl->fl_flags & FL_ACCESS)
2511
			seq_puts(f, "ACCESS");
2512
		else if (IS_OFDLCK(fl))
2513
			seq_puts(f, "OFDLCK");
2514
		else
2515
			seq_puts(f, "POSIX ");
2516 2517

		seq_printf(f, " %s ",
L
Linus Torvalds 已提交
2518
			     (inode == NULL) ? "*NOINODE*" :
2519
			     mandatory_lock(inode) ? "MANDATORY" : "ADVISORY ");
L
Linus Torvalds 已提交
2520 2521
	} else if (IS_FLOCK(fl)) {
		if (fl->fl_type & LOCK_MAND) {
2522
			seq_puts(f, "FLOCK  MSNFS     ");
L
Linus Torvalds 已提交
2523
		} else {
2524
			seq_puts(f, "FLOCK  ADVISORY  ");
L
Linus Torvalds 已提交
2525 2526
		}
	} else if (IS_LEASE(fl)) {
2527 2528 2529 2530 2531
		if (fl->fl_flags & FL_DELEG)
			seq_puts(f, "DELEG  ");
		else
			seq_puts(f, "LEASE  ");

J
J. Bruce Fields 已提交
2532
		if (lease_breaking(fl))
2533
			seq_puts(f, "BREAKING  ");
L
Linus Torvalds 已提交
2534
		else if (fl->fl_file)
2535
			seq_puts(f, "ACTIVE    ");
L
Linus Torvalds 已提交
2536
		else
2537
			seq_puts(f, "BREAKER   ");
L
Linus Torvalds 已提交
2538
	} else {
2539
		seq_puts(f, "UNKNOWN UNKNOWN  ");
L
Linus Torvalds 已提交
2540 2541
	}
	if (fl->fl_type & LOCK_MAND) {
2542
		seq_printf(f, "%s ",
L
Linus Torvalds 已提交
2543 2544 2545 2546
			       (fl->fl_type & LOCK_READ)
			       ? (fl->fl_type & LOCK_WRITE) ? "RW   " : "READ "
			       : (fl->fl_type & LOCK_WRITE) ? "WRITE" : "NONE ");
	} else {
2547
		seq_printf(f, "%s ",
J
J. Bruce Fields 已提交
2548
			       (lease_breaking(fl))
2549 2550
			       ? (fl->fl_type == F_UNLCK) ? "UNLCK" : "READ "
			       : (fl->fl_type == F_WRLCK) ? "WRITE" : "READ ");
L
Linus Torvalds 已提交
2551 2552 2553
	}
	if (inode) {
#ifdef WE_CAN_BREAK_LSLK_NOW
2554
		seq_printf(f, "%d %s:%ld ", fl_pid,
L
Linus Torvalds 已提交
2555 2556 2557
				inode->i_sb->s_id, inode->i_ino);
#else
		/* userspace relies on this representation of dev_t ;-( */
2558
		seq_printf(f, "%d %02x:%02x:%ld ", fl_pid,
L
Linus Torvalds 已提交
2559 2560 2561 2562
				MAJOR(inode->i_sb->s_dev),
				MINOR(inode->i_sb->s_dev), inode->i_ino);
#endif
	} else {
2563
		seq_printf(f, "%d <none>:0 ", fl_pid);
L
Linus Torvalds 已提交
2564 2565 2566
	}
	if (IS_POSIX(fl)) {
		if (fl->fl_end == OFFSET_MAX)
2567
			seq_printf(f, "%Ld EOF\n", fl->fl_start);
L
Linus Torvalds 已提交
2568
		else
2569
			seq_printf(f, "%Ld %Ld\n", fl->fl_start, fl->fl_end);
L
Linus Torvalds 已提交
2570
	} else {
2571
		seq_puts(f, "0 EOF\n");
L
Linus Torvalds 已提交
2572 2573 2574
	}
}

2575
static int locks_show(struct seq_file *f, void *v)
L
Linus Torvalds 已提交
2576
{
2577
	struct locks_iterator *iter = f->private;
2578
	struct file_lock *fl, *bfl;
L
Linus Torvalds 已提交
2579

2580
	fl = hlist_entry(v, struct file_lock, fl_link);
L
Linus Torvalds 已提交
2581

2582
	lock_get_status(f, fl, iter->li_pos, "");
L
Linus Torvalds 已提交
2583

2584
	list_for_each_entry(bfl, &fl->fl_block, fl_block)
2585
		lock_get_status(f, bfl, iter->li_pos, " ->");
2586

2587 2588
	return 0;
}
L
Linus Torvalds 已提交
2589

2590
static void *locks_start(struct seq_file *f, loff_t *pos)
2591
	__acquires(&blocked_lock_lock)
2592
{
2593
	struct locks_iterator *iter = f->private;
2594

2595 2596
	iter->li_pos = *pos + 1;
	lg_global_lock(&file_lock_lglock);
2597
	spin_lock(&blocked_lock_lock);
2598
	return seq_hlist_start_percpu(&file_lock_list, &iter->li_cpu, *pos);
2599
}
L
Linus Torvalds 已提交
2600

2601 2602
static void *locks_next(struct seq_file *f, void *v, loff_t *pos)
{
2603 2604 2605 2606
	struct locks_iterator *iter = f->private;

	++iter->li_pos;
	return seq_hlist_next_percpu(v, &file_lock_list, &iter->li_cpu, pos);
2607
}
L
Linus Torvalds 已提交
2608

2609
static void locks_stop(struct seq_file *f, void *v)
2610
	__releases(&blocked_lock_lock)
2611
{
2612
	spin_unlock(&blocked_lock_lock);
2613
	lg_global_unlock(&file_lock_lglock);
L
Linus Torvalds 已提交
2614 2615
}

2616
static const struct seq_operations locks_seq_operations = {
2617 2618 2619 2620 2621
	.start	= locks_start,
	.next	= locks_next,
	.stop	= locks_stop,
	.show	= locks_show,
};
2622 2623 2624

static int locks_open(struct inode *inode, struct file *filp)
{
2625 2626
	return seq_open_private(filp, &locks_seq_operations,
					sizeof(struct locks_iterator));
2627 2628 2629 2630 2631 2632
}

static const struct file_operations proc_locks_operations = {
	.open		= locks_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2633
	.release	= seq_release_private,
2634 2635 2636 2637 2638 2639 2640 2641
};

static int __init proc_locks_init(void)
{
	proc_create("locks", 0, NULL, &proc_locks_operations);
	return 0;
}
module_init(proc_locks_init);
2642 2643
#endif

L
Linus Torvalds 已提交
2644 2645
static int __init filelock_init(void)
{
2646 2647
	int i;

2648 2649 2650
	flctx_cache = kmem_cache_create("file_lock_ctx",
			sizeof(struct file_lock_context), 0, SLAB_PANIC, NULL);

L
Linus Torvalds 已提交
2651
	filelock_cache = kmem_cache_create("file_lock_cache",
M
Miklos Szeredi 已提交
2652 2653
			sizeof(struct file_lock), 0, SLAB_PANIC, NULL);

2654 2655 2656 2657 2658
	lg_lock_init(&file_lock_lglock, "file_lock_lglock");

	for_each_possible_cpu(i)
		INIT_HLIST_HEAD(per_cpu_ptr(&file_lock_list, i));

L
Linus Torvalds 已提交
2659 2660 2661 2662
	return 0;
}

core_initcall(filelock_init);