locks.c 68.5 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)
C
Christoph Hellwig 已提交
140
#define IS_LEASE(fl)	(fl->fl_flags & (FL_LEASE|FL_DELEG|FL_LAYOUT))
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
 * 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
164
 * the relevant flc_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
 *
 * 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,
192
 * we often hold the flc_lock as well. In certain cases, when reading the fields
193
 * protected by this lock, we can skip acquiring it iff we already hold the
194
 * flc_lock.
195 196
 *
 * In particular, adding an entry to the fl_block list requires that you hold
197 198
 * both the flc_lock and the blocked_lock_lock (acquired in that order).
 * Deleting an entry from the list however only requires the file_lock_lock.
199
 */
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
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;

217
	spin_lock_init(&new->flc_lock);
218
	INIT_LIST_HEAD(&new->flc_flock);
219
	INIT_LIST_HEAD(&new->flc_posix);
220
	INIT_LIST_HEAD(&new->flc_lease);
221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243

	/*
	 * 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));
244
		WARN_ON_ONCE(!list_empty(&ctx->flc_posix));
245
		WARN_ON_ONCE(!list_empty(&ctx->flc_lease));
246 247 248 249
		kmem_cache_free(flctx_cache, ctx);
	}
}

M
Miklos Szeredi 已提交
250
static void locks_init_lock_heads(struct file_lock *fl)
M
Miklos Szeredi 已提交
251
{
252
	INIT_HLIST_NODE(&fl->fl_link);
253
	INIT_LIST_HEAD(&fl->fl_list);
M
Miklos Szeredi 已提交
254 255
	INIT_LIST_HEAD(&fl->fl_block);
	init_waitqueue_head(&fl->fl_wait);
M
Miklos Szeredi 已提交
256 257
}

L
Linus Torvalds 已提交
258
/* Allocate an empty lock structure. */
259
struct file_lock *locks_alloc_lock(void)
L
Linus Torvalds 已提交
260
{
M
Miklos Szeredi 已提交
261
	struct file_lock *fl = kmem_cache_zalloc(filelock_cache, GFP_KERNEL);
M
Miklos Szeredi 已提交
262 263

	if (fl)
M
Miklos Szeredi 已提交
264
		locks_init_lock_heads(fl);
M
Miklos Szeredi 已提交
265 266

	return fl;
L
Linus Torvalds 已提交
267
}
268
EXPORT_SYMBOL_GPL(locks_alloc_lock);
L
Linus Torvalds 已提交
269

270
void locks_release_private(struct file_lock *fl)
271 272 273 274 275 276 277
{
	if (fl->fl_ops) {
		if (fl->fl_ops->fl_release_private)
			fl->fl_ops->fl_release_private(fl);
		fl->fl_ops = NULL;
	}

278 279 280 281 282
	if (fl->fl_lmops) {
		if (fl->fl_lmops->lm_put_owner)
			fl->fl_lmops->lm_put_owner(fl);
		fl->fl_lmops = NULL;
	}
283
}
284
EXPORT_SYMBOL_GPL(locks_release_private);
285

L
Linus Torvalds 已提交
286
/* Free a lock which is not in use. */
287
void locks_free_lock(struct file_lock *fl)
L
Linus Torvalds 已提交
288
{
M
Miklos Szeredi 已提交
289
	BUG_ON(waitqueue_active(&fl->fl_wait));
290
	BUG_ON(!list_empty(&fl->fl_list));
M
Miklos Szeredi 已提交
291
	BUG_ON(!list_empty(&fl->fl_block));
292
	BUG_ON(!hlist_unhashed(&fl->fl_link));
L
Linus Torvalds 已提交
293

294
	locks_release_private(fl);
L
Linus Torvalds 已提交
295 296
	kmem_cache_free(filelock_cache, fl);
}
297
EXPORT_SYMBOL(locks_free_lock);
L
Linus Torvalds 已提交
298

299 300 301 302 303 304
static void
locks_dispose_list(struct list_head *dispose)
{
	struct file_lock *fl;

	while (!list_empty(dispose)) {
305 306
		fl = list_first_entry(dispose, struct file_lock, fl_list);
		list_del_init(&fl->fl_list);
307 308 309 310
		locks_free_lock(fl);
	}
}

L
Linus Torvalds 已提交
311 312
void locks_init_lock(struct file_lock *fl)
{
M
Miklos Szeredi 已提交
313 314
	memset(fl, 0, sizeof(struct file_lock));
	locks_init_lock_heads(fl);
L
Linus Torvalds 已提交
315 316 317 318 319 320 321
}

EXPORT_SYMBOL(locks_init_lock);

/*
 * Initialize a new lock from an existing file_lock structure.
 */
322
void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
L
Linus Torvalds 已提交
323 324 325
{
	new->fl_owner = fl->fl_owner;
	new->fl_pid = fl->fl_pid;
326
	new->fl_file = NULL;
L
Linus Torvalds 已提交
327 328 329 330
	new->fl_flags = fl->fl_flags;
	new->fl_type = fl->fl_type;
	new->fl_start = fl->fl_start;
	new->fl_end = fl->fl_end;
331
	new->fl_lmops = fl->fl_lmops;
332
	new->fl_ops = NULL;
333 334 335 336 337

	if (fl->fl_lmops) {
		if (fl->fl_lmops->lm_get_owner)
			fl->fl_lmops->lm_get_owner(new, fl);
	}
338
}
339
EXPORT_SYMBOL(locks_copy_conflock);
340 341 342

void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
{
343 344
	/* "new" must be a freshly-initialized lock */
	WARN_ON_ONCE(new->fl_ops);
345

346
	locks_copy_conflock(new, fl);
347

348
	new->fl_file = fl->fl_file;
L
Linus Torvalds 已提交
349
	new->fl_ops = fl->fl_ops;
350

351 352 353 354
	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
355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373
}

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. */
374 375
static struct file_lock *
flock_make_lock(struct file *filp, unsigned int cmd)
L
Linus Torvalds 已提交
376 377 378
{
	struct file_lock *fl;
	int type = flock_translate_cmd(cmd);
379

L
Linus Torvalds 已提交
380
	if (type < 0)
381
		return ERR_PTR(type);
L
Linus Torvalds 已提交
382 383 384
	
	fl = locks_alloc_lock();
	if (fl == NULL)
385
		return ERR_PTR(-ENOMEM);
L
Linus Torvalds 已提交
386 387

	fl->fl_file = filp;
388
	fl->fl_owner = filp;
L
Linus Torvalds 已提交
389 390 391 392 393
	fl->fl_pid = current->tgid;
	fl->fl_flags = FL_FLOCK;
	fl->fl_type = type;
	fl->fl_end = OFFSET_MAX;
	
394
	return fl;
L
Linus Torvalds 已提交
395 396
}

397
static int assign_type(struct file_lock *fl, long type)
L
Linus Torvalds 已提交
398 399 400 401 402 403 404 405 406 407 408 409 410
{
	switch (type) {
	case F_RDLCK:
	case F_WRLCK:
	case F_UNLCK:
		fl->fl_type = type;
		break;
	default:
		return -EINVAL;
	}
	return 0;
}

411 412
static int flock64_to_posix_lock(struct file *filp, struct file_lock *fl,
				 struct flock64 *l)
L
Linus Torvalds 已提交
413 414
{
	switch (l->l_whence) {
415
	case SEEK_SET:
416
		fl->fl_start = 0;
L
Linus Torvalds 已提交
417
		break;
418
	case SEEK_CUR:
419
		fl->fl_start = filp->f_pos;
L
Linus Torvalds 已提交
420
		break;
421
	case SEEK_END:
422
		fl->fl_start = i_size_read(file_inode(filp));
L
Linus Torvalds 已提交
423 424 425 426
		break;
	default:
		return -EINVAL;
	}
427 428 429 430 431
	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
432 433 434

	/* POSIX-1996 leaves the case l->l_len < 0 undefined;
	   POSIX-2001 defines it. */
435
	if (l->l_len > 0) {
436 437 438 439
		if (l->l_len - 1 > OFFSET_MAX - fl->fl_start)
			return -EOVERFLOW;
		fl->fl_end = fl->fl_start + l->l_len - 1;

440
	} else if (l->l_len < 0) {
441
		if (fl->fl_start + l->l_len < 0)
442
			return -EINVAL;
443 444 445 446 447
		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
448 449 450 451 452 453 454 455 456 457
	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);
}

458 459 460 461 462
/* 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
463
{
464 465 466 467 468 469 470 471
	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
472 473 474
}

/* default lease lock manager operations */
J
Jeff Layton 已提交
475 476
static bool
lease_break_callback(struct file_lock *fl)
L
Linus Torvalds 已提交
477 478
{
	kill_fasync(&fl->fl_fasync, SIGIO, POLL_MSG);
J
Jeff Layton 已提交
479
	return false;
L
Linus Torvalds 已提交
480 481
}

482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
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);
}

499
static const struct lock_manager_operations lease_manager_ops = {
J
J. Bruce Fields 已提交
500 501
	.lm_break = lease_break_callback,
	.lm_change = lease_modify,
502
	.lm_setup = lease_setup,
L
Linus Torvalds 已提交
503 504 505 506 507
};

/*
 * Initialize a lease, use the default lock manager operations
 */
508
static int lease_init(struct file *filp, long type, struct file_lock *fl)
L
Linus Torvalds 已提交
509
 {
510 511 512
	if (assign_type(fl, type) != 0)
		return -EINVAL;

513
	fl->fl_owner = filp;
L
Linus Torvalds 已提交
514 515 516 517 518 519 520 521 522 523 524 525
	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 */
526
static struct file_lock *lease_alloc(struct file *filp, long type)
L
Linus Torvalds 已提交
527 528
{
	struct file_lock *fl = locks_alloc_lock();
529
	int error = -ENOMEM;
L
Linus Torvalds 已提交
530 531

	if (fl == NULL)
J
J. Bruce Fields 已提交
532
		return ERR_PTR(error);
L
Linus Torvalds 已提交
533 534

	error = lease_init(filp, type, fl);
535 536
	if (error) {
		locks_free_lock(fl);
J
J. Bruce Fields 已提交
537
		return ERR_PTR(error);
538
	}
J
J. Bruce Fields 已提交
539
	return fl;
L
Linus Torvalds 已提交
540 541 542 543 544 545 546 547 548 549 550 551 552
}

/* 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.
 */
553
static int posix_same_owner(struct file_lock *fl1, struct file_lock *fl2)
L
Linus Torvalds 已提交
554
{
J
J. Bruce Fields 已提交
555
	if (fl1->fl_lmops && fl1->fl_lmops->lm_compare_owner)
L
Linus Torvalds 已提交
556
		return fl2->fl_lmops == fl1->fl_lmops &&
J
J. Bruce Fields 已提交
557
			fl1->fl_lmops->lm_compare_owner(fl1, fl2);
L
Linus Torvalds 已提交
558 559 560
	return fl1->fl_owner == fl2->fl_owner;
}

561
/* Must be called with the flc_lock held! */
562
static void locks_insert_global_locks(struct file_lock *fl)
563
{
564 565 566 567
	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);
568 569
}

570
/* Must be called with the flc_lock held! */
571
static void locks_delete_global_locks(struct file_lock *fl)
572
{
573 574
	/*
	 * Avoid taking lock if already unhashed. This is safe since this check
575
	 * is done while holding the flc_lock, and new insertions into the list
576 577 578 579 580
	 * also require that it be held.
	 */
	if (hlist_unhashed(&fl->fl_link))
		return;
	lg_local_lock_cpu(&file_lock_lglock, fl->fl_link_cpu);
581
	hlist_del_init(&fl->fl_link);
582
	lg_local_unlock_cpu(&file_lock_lglock, fl->fl_link_cpu);
583 584
}

585 586 587 588 589 590 591 592
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;
}

593
static void locks_insert_global_blocked(struct file_lock *waiter)
594
{
595
	hash_add(blocked_hash, &waiter->fl_link, posix_owner_key(waiter));
596 597
}

598
static void locks_delete_global_blocked(struct file_lock *waiter)
599
{
600
	hash_del(&waiter->fl_link);
601 602
}

L
Linus Torvalds 已提交
603 604
/* Remove waiter from blocker's block list.
 * When blocker ends up pointing to itself then the list is empty.
605
 *
606
 * Must be called with blocked_lock_lock held.
L
Linus Torvalds 已提交
607
 */
608
static void __locks_delete_block(struct file_lock *waiter)
L
Linus Torvalds 已提交
609
{
610
	locks_delete_global_blocked(waiter);
L
Linus Torvalds 已提交
611 612 613 614
	list_del_init(&waiter->fl_block);
	waiter->fl_next = NULL;
}

615
static void locks_delete_block(struct file_lock *waiter)
L
Linus Torvalds 已提交
616
{
617
	spin_lock(&blocked_lock_lock);
L
Linus Torvalds 已提交
618
	__locks_delete_block(waiter);
619
	spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
620 621 622 623 624 625
}

/* 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.
626
 *
627 628 629 630
 * Must be called with both the flc_lock and blocked_lock_lock held. The
 * fl_block list itself is protected by the blocked_lock_lock, but by ensuring
 * that the flc_lock is also held on insertions we can avoid taking the
 * blocked_lock_lock in some cases when we see that the fl_block list is empty.
L
Linus Torvalds 已提交
631
 */
632 633
static void __locks_insert_block(struct file_lock *blocker,
					struct file_lock *waiter)
L
Linus Torvalds 已提交
634
{
635
	BUG_ON(!list_empty(&waiter->fl_block));
L
Linus Torvalds 已提交
636
	waiter->fl_next = blocker;
637
	list_add_tail(&waiter->fl_block, &blocker->fl_block);
638
	if (IS_POSIX(blocker) && !IS_OFDLCK(blocker))
639 640 641
		locks_insert_global_blocked(waiter);
}

642
/* Must be called with flc_lock held. */
643 644 645
static void locks_insert_block(struct file_lock *blocker,
					struct file_lock *waiter)
{
646
	spin_lock(&blocked_lock_lock);
647
	__locks_insert_block(blocker, waiter);
648
	spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
649 650
}

651 652 653
/*
 * Wake up processes blocked waiting for blocker.
 *
654
 * Must be called with the inode->flc_lock held!
L
Linus Torvalds 已提交
655 656 657
 */
static void locks_wake_up_blocks(struct file_lock *blocker)
{
658 659
	/*
	 * Avoid taking global lock if list is empty. This is safe since new
660 661 662
	 * blocked requests are only added to the list under the flc_lock, and
	 * the flc_lock is always held here. Note that removal from the fl_block
	 * list does not require the flc_lock, so we must recheck list_empty()
663
	 * after acquiring the blocked_lock_lock.
664 665 666 667
	 */
	if (list_empty(&blocker->fl_block))
		return;

668
	spin_lock(&blocked_lock_lock);
L
Linus Torvalds 已提交
669
	while (!list_empty(&blocker->fl_block)) {
670 671 672
		struct file_lock *waiter;

		waiter = list_first_entry(&blocker->fl_block,
L
Linus Torvalds 已提交
673 674
				struct file_lock, fl_block);
		__locks_delete_block(waiter);
J
J. Bruce Fields 已提交
675 676
		if (waiter->fl_lmops && waiter->fl_lmops->lm_notify)
			waiter->fl_lmops->lm_notify(waiter);
L
Linus Torvalds 已提交
677 678 679
		else
			wake_up(&waiter->fl_wait);
	}
680
	spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
681 682
}

683
static void
684
locks_insert_lock_ctx(struct file_lock *fl, struct list_head *before)
685 686 687 688 689 690
{
	fl->fl_nspid = get_pid(task_tgid(current));
	list_add_tail(&fl->fl_list, before);
	locks_insert_global_locks(fl);
}

691
static void
692
locks_unlink_lock_ctx(struct file_lock *fl)
L
Linus Torvalds 已提交
693
{
694
	locks_delete_global_locks(fl);
695
	list_del_init(&fl->fl_list);
696 697 698 699
	if (fl->fl_nspid) {
		put_pid(fl->fl_nspid);
		fl->fl_nspid = NULL;
	}
L
Linus Torvalds 已提交
700
	locks_wake_up_blocks(fl);
701 702
}

703
static void
704
locks_delete_lock_ctx(struct file_lock *fl, struct list_head *dispose)
705
{
706
	locks_unlink_lock_ctx(fl);
707
	if (dispose)
708
		list_add(&fl->fl_list, dispose);
709 710
	else
		locks_free_lock(fl);
L
Linus Torvalds 已提交
711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
}

/* 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));
}

759
void
760
posix_test_lock(struct file *filp, struct file_lock *fl)
L
Linus Torvalds 已提交
761 762
{
	struct file_lock *cfl;
763
	struct file_lock_context *ctx;
764
	struct inode *inode = file_inode(filp);
L
Linus Torvalds 已提交
765

766 767 768 769 770 771
	ctx = inode->i_flctx;
	if (!ctx || list_empty_careful(&ctx->flc_posix)) {
		fl->fl_type = F_UNLCK;
		return;
	}

772
	spin_lock(&ctx->flc_lock);
773 774 775 776 777 778 779
	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
780
	}
781 782
	fl->fl_type = F_UNLCK;
out:
783
	spin_unlock(&ctx->flc_lock);
784
	return;
L
Linus Torvalds 已提交
785 786 787
}
EXPORT_SYMBOL(posix_test_lock);

788 789 790 791 792
/*
 * Deadlock detection:
 *
 * We attempt to detect deadlocks that are due purely to posix file
 * locks.
L
Linus Torvalds 已提交
793
 *
794 795 796 797 798 799 800
 * 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
801
 *
802 803 804
 * 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.
805
 *
806 807 808 809
 * 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.
810
 *
811
 * For FL_OFDLCK locks, the owner is the filp, not the files_struct.
812 813 814 815
 * 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.
 *
816
 * In principle, we could do a more limited deadlock detection on FL_OFDLCK
817 818
 * 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
819
 */
820 821 822

#define MAX_DEADLK_ITERATIONS 10

823 824 825 826 827
/* 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;

828
	hash_for_each_possible(blocked_hash, fl, fl_link, posix_owner_key(block_fl)) {
829 830 831 832 833 834
		if (posix_same_owner(fl, block_fl))
			return fl->fl_next;
	}
	return NULL;
}

835
/* Must be called with the blocked_lock_lock held! */
836
static int posix_locks_deadlock(struct file_lock *caller_fl,
L
Linus Torvalds 已提交
837 838
				struct file_lock *block_fl)
{
839
	int i = 0;
L
Linus Torvalds 已提交
840

841 842
	/*
	 * This deadlock detector can't reasonably detect deadlocks with
843
	 * FL_OFDLCK locks, since they aren't owned by a process, per-se.
844
	 */
845
	if (IS_OFDLCK(caller_fl))
846 847
		return 0;

848 849 850 851 852
	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
853 854 855 856 857
	}
	return 0;
}

/* Try to create a FLOCK lock on filp. We always insert new FLOCK locks
858
 * after any leases, but before any posix locks.
859 860 861 862
 *
 * 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
863
 */
864
static int flock_lock_file(struct file *filp, struct file_lock *request)
L
Linus Torvalds 已提交
865
{
866
	struct file_lock *new_fl = NULL;
867 868 869
	struct file_lock *fl;
	struct file_lock_context *ctx;
	struct inode *inode = file_inode(filp);
L
Linus Torvalds 已提交
870
	int error = 0;
871
	bool found = false;
872
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
873

874 875 876 877
	ctx = locks_get_lock_context(inode);
	if (!ctx)
		return -ENOMEM;

878
	if (!(request->fl_flags & FL_ACCESS) && (request->fl_type != F_UNLCK)) {
879
		new_fl = locks_alloc_lock();
880 881
		if (!new_fl)
			return -ENOMEM;
882 883
	}

884
	spin_lock(&ctx->flc_lock);
885 886 887
	if (request->fl_flags & FL_ACCESS)
		goto find_conflict;

888
	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
L
Linus Torvalds 已提交
889 890
		if (filp != fl->fl_file)
			continue;
891
		if (request->fl_type == fl->fl_type)
L
Linus Torvalds 已提交
892
			goto out;
893
		found = true;
894
		locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
895 896 897
		break;
	}

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

904
find_conflict:
905
	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
906
		if (!flock_locks_conflict(request, fl))
L
Linus Torvalds 已提交
907 908
			continue;
		error = -EAGAIN;
909 910 911 912
		if (!(request->fl_flags & FL_SLEEP))
			goto out;
		error = FILE_LOCK_DEFERRED;
		locks_insert_block(fl, request);
L
Linus Torvalds 已提交
913 914
		goto out;
	}
915 916
	if (request->fl_flags & FL_ACCESS)
		goto out;
917
	locks_copy_lock(new_fl, request);
918
	locks_insert_lock_ctx(new_fl, &ctx->flc_flock);
919
	new_fl = NULL;
920
	error = 0;
L
Linus Torvalds 已提交
921 922

out:
923
	spin_unlock(&ctx->flc_lock);
924 925
	if (new_fl)
		locks_free_lock(new_fl);
926
	locks_dispose_list(&dispose);
L
Linus Torvalds 已提交
927 928 929
	return error;
}

930
static int __posix_lock_file(struct inode *inode, struct file_lock *request, struct file_lock *conflock)
L
Linus Torvalds 已提交
931
{
932
	struct file_lock *fl, *tmp;
933 934
	struct file_lock *new_fl = NULL;
	struct file_lock *new_fl2 = NULL;
L
Linus Torvalds 已提交
935 936
	struct file_lock *left = NULL;
	struct file_lock *right = NULL;
937
	struct file_lock_context *ctx;
938 939
	int error;
	bool added = false;
940
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
941

942 943 944 945
	ctx = locks_get_lock_context(inode);
	if (!ctx)
		return -ENOMEM;

L
Linus Torvalds 已提交
946 947 948
	/*
	 * We may need two file_lock structures for this operation,
	 * so we get them in advance to avoid races.
949 950
	 *
	 * In some cases we can be sure, that no new locks will be needed
L
Linus Torvalds 已提交
951
	 */
952 953 954 955 956 957
	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
958

959
	spin_lock(&ctx->flc_lock);
960 961 962
	/*
	 * New lock request. Walk all POSIX locks and look for conflicts. If
	 * there are any, either return error or put the request on the
963
	 * blocker's list of waiters and the global blocked_hash.
964
	 */
L
Linus Torvalds 已提交
965
	if (request->fl_type != F_UNLCK) {
966
		list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
L
Linus Torvalds 已提交
967 968 969 970
			if (!IS_POSIX(fl))
				continue;
			if (!posix_locks_conflict(request, fl))
				continue;
971
			if (conflock)
972
				locks_copy_conflock(conflock, fl);
L
Linus Torvalds 已提交
973 974 975
			error = -EAGAIN;
			if (!(request->fl_flags & FL_SLEEP))
				goto out;
976 977 978 979
			/*
			 * Deadlock detection and insertion into the blocked
			 * locks list must be done while holding the same lock!
			 */
L
Linus Torvalds 已提交
980
			error = -EDEADLK;
981
			spin_lock(&blocked_lock_lock);
982 983 984 985
			if (likely(!posix_locks_deadlock(request, fl))) {
				error = FILE_LOCK_DEFERRED;
				__locks_insert_block(fl, request);
			}
986
			spin_unlock(&blocked_lock_lock);
L
Linus Torvalds 已提交
987 988 989 990 991 992 993 994 995
			goto out;
  		}
  	}

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

996 997 998 999
	/* 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
1000 1001
	}

1002
	/* Process locks with this owner. */
1003 1004 1005 1006 1007
	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
1008
		if (request->fl_type == fl->fl_type) {
O
Olaf Kirch 已提交
1009 1010 1011 1012
			/* 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
1013
			if (fl->fl_end < request->fl_start - 1)
1014
				continue;
L
Linus Torvalds 已提交
1015 1016 1017
			/* 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
1018
			if (fl->fl_start - 1 > request->fl_end)
L
Linus Torvalds 已提交
1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034
				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) {
1035
				locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
1036 1037 1038
				continue;
			}
			request = fl;
1039
			added = true;
1040
		} else {
L
Linus Torvalds 已提交
1041 1042 1043 1044
			/* Processing for different lock types is a bit
			 * more complex.
			 */
			if (fl->fl_end < request->fl_start)
1045
				continue;
L
Linus Torvalds 已提交
1046 1047 1048
			if (fl->fl_start > request->fl_end)
				break;
			if (request->fl_type == F_UNLCK)
1049
				added = true;
L
Linus Torvalds 已提交
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063
			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) {
1064
					locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
1065 1066
					continue;
				}
1067 1068 1069 1070 1071 1072
				/*
				 * 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
1073
				 */
1074 1075 1076 1077 1078 1079
				error = -ENOLCK;
				if (!new_fl)
					goto out;
				locks_copy_lock(new_fl, request);
				request = new_fl;
				new_fl = NULL;
1080 1081
				locks_insert_lock_ctx(request, &fl->fl_list);
				locks_delete_lock_ctx(fl, &dispose);
1082
				added = true;
L
Linus Torvalds 已提交
1083 1084 1085 1086
			}
		}
	}

1087
	/*
1088 1089 1090
	 * 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.
1091 1092 1093 1094 1095
	 */
	error = -ENOLCK; /* "no luck" */
	if (right && left == right && !new_fl2)
		goto out;

L
Linus Torvalds 已提交
1096 1097
	error = 0;
	if (!added) {
1098 1099 1100
		if (request->fl_type == F_UNLCK) {
			if (request->fl_flags & FL_EXISTS)
				error = -ENOENT;
L
Linus Torvalds 已提交
1101
			goto out;
1102
		}
1103 1104 1105 1106 1107

		if (!new_fl) {
			error = -ENOLCK;
			goto out;
		}
L
Linus Torvalds 已提交
1108
		locks_copy_lock(new_fl, request);
1109
		locks_insert_lock_ctx(new_fl, &fl->fl_list);
L
Linus Torvalds 已提交
1110 1111 1112 1113 1114 1115 1116 1117 1118 1119
		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);
1120
			locks_insert_lock_ctx(left, &fl->fl_list);
L
Linus Torvalds 已提交
1121 1122 1123 1124 1125 1126 1127 1128 1129
		}
		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:
1130
	spin_unlock(&ctx->flc_lock);
L
Linus Torvalds 已提交
1131 1132 1133 1134 1135 1136 1137
	/*
	 * Free any unused locks.
	 */
	if (new_fl)
		locks_free_lock(new_fl);
	if (new_fl2)
		locks_free_lock(new_fl2);
1138
	locks_dispose_list(&dispose);
L
Linus Torvalds 已提交
1139 1140 1141 1142 1143 1144 1145
	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
1146
 * @conflock: Place to return a copy of the conflicting lock, if found.
L
Linus Torvalds 已提交
1147 1148 1149 1150
 *
 * 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
1151 1152 1153 1154
 *
 * 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
1155
 */
1156
int posix_lock_file(struct file *filp, struct file_lock *fl,
1157 1158
			struct file_lock *conflock)
{
A
Al Viro 已提交
1159
	return __posix_lock_file(file_inode(filp), fl, conflock);
L
Linus Torvalds 已提交
1160
}
1161
EXPORT_SYMBOL(posix_lock_file);
L
Linus Torvalds 已提交
1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176

/**
 * 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 (;;) {
1177
		error = posix_lock_file(filp, fl, NULL);
1178
		if (error != FILE_LOCK_DEFERRED)
L
Linus Torvalds 已提交
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192
			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
1193
 * @file: the file to check
L
Linus Torvalds 已提交
1194 1195 1196 1197
 *
 * Searches the inode's list of locks to find any POSIX locks which conflict.
 * This function is called from locks_verify_locked() only.
 */
1198
int locks_mandatory_locked(struct file *file)
L
Linus Torvalds 已提交
1199
{
1200
	int ret;
1201
	struct inode *inode = file_inode(file);
1202
	struct file_lock_context *ctx;
L
Linus Torvalds 已提交
1203 1204
	struct file_lock *fl;

1205 1206 1207 1208
	ctx = inode->i_flctx;
	if (!ctx || list_empty_careful(&ctx->flc_posix))
		return 0;

L
Linus Torvalds 已提交
1209 1210 1211
	/*
	 * Search the lock list for this inode for any POSIX locks.
	 */
1212
	spin_lock(&ctx->flc_lock);
1213 1214
	ret = 0;
	list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
1215
		if (fl->fl_owner != current->files &&
1216 1217
		    fl->fl_owner != file) {
			ret = -EAGAIN;
L
Linus Torvalds 已提交
1218
			break;
1219
		}
L
Linus Torvalds 已提交
1220
	}
1221
	spin_unlock(&ctx->flc_lock);
1222
	return ret;
L
Linus Torvalds 已提交
1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243
}

/**
 * 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;
1244
	bool sleep = false;
L
Linus Torvalds 已提交
1245 1246 1247 1248 1249 1250

	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))
1251
		sleep = true;
L
Linus Torvalds 已提交
1252 1253 1254 1255 1256
	fl.fl_type = (read_write == FLOCK_VERIFY_WRITE) ? F_WRLCK : F_RDLCK;
	fl.fl_start = offset;
	fl.fl_end = offset + count - 1;

	for (;;) {
1257
		if (filp) {
1258
			fl.fl_owner = filp;
1259 1260 1261 1262 1263 1264 1265 1266 1267
			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;
1268
		error = __posix_lock_file(inode, &fl, NULL);
1269
		if (error != FILE_LOCK_DEFERRED)
L
Linus Torvalds 已提交
1270 1271 1272 1273 1274 1275 1276
			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.
			 */
1277
			if (__mandatory_lock(inode))
L
Linus Torvalds 已提交
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
				continue;
		}

		locks_delete_block(&fl);
		break;
	}

	return error;
}

EXPORT_SYMBOL(locks_mandatory_area);

1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
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
1301
/* We already had a lease on this file; just change its type */
1302
int lease_modify(struct file_lock *fl, int arg, struct list_head *dispose)
L
Linus Torvalds 已提交
1303 1304 1305 1306 1307
{
	int error = assign_type(fl, arg);

	if (error)
		return error;
1308
	lease_clear_pending(fl, arg);
L
Linus Torvalds 已提交
1309
	locks_wake_up_blocks(fl);
1310 1311 1312 1313 1314
	if (arg == F_UNLCK) {
		struct file *filp = fl->fl_file;

		f_delown(filp);
		filp->f_owner.signum = 0;
1315 1316 1317 1318 1319
		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;
		}
1320
		locks_delete_lock_ctx(fl, dispose);
1321
	}
L
Linus Torvalds 已提交
1322 1323 1324 1325
	return 0;
}
EXPORT_SYMBOL(lease_modify);

1326 1327 1328 1329 1330 1331 1332 1333
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);
}

1334
static void time_out_leases(struct inode *inode, struct list_head *dispose)
L
Linus Torvalds 已提交
1335
{
1336 1337
	struct file_lock_context *ctx = inode->i_flctx;
	struct file_lock *fl, *tmp;
L
Linus Torvalds 已提交
1338

1339
	lockdep_assert_held(&ctx->flc_lock);
1340

1341
	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, fl_list) {
1342
		trace_time_out_leases(inode, fl);
1343
		if (past_time(fl->fl_downgrade_time))
1344
			lease_modify(fl, F_RDLCK, dispose);
1345
		if (past_time(fl->fl_break_time))
1346
			lease_modify(fl, F_UNLCK, dispose);
L
Linus Torvalds 已提交
1347 1348 1349
	}
}

J
J. Bruce Fields 已提交
1350 1351
static bool leases_conflict(struct file_lock *lease, struct file_lock *breaker)
{
C
Christoph Hellwig 已提交
1352 1353
	if ((breaker->fl_flags & FL_LAYOUT) != (lease->fl_flags & FL_LAYOUT))
		return false;
J
J. Bruce Fields 已提交
1354 1355 1356 1357 1358
	if ((breaker->fl_flags & FL_DELEG) && (lease->fl_flags & FL_LEASE))
		return false;
	return locks_conflict(breaker, lease);
}

1359 1360 1361
static bool
any_leases_conflict(struct inode *inode, struct file_lock *breaker)
{
1362
	struct file_lock_context *ctx = inode->i_flctx;
1363 1364
	struct file_lock *fl;

1365
	lockdep_assert_held(&ctx->flc_lock);
1366

1367
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
1368 1369 1370 1371 1372 1373
		if (leases_conflict(fl, breaker))
			return true;
	}
	return false;
}

L
Linus Torvalds 已提交
1374 1375 1376
/**
 *	__break_lease	-	revoke all outstanding leases on file
 *	@inode: the inode of the file to return
J
J. Bruce Fields 已提交
1377 1378 1379 1380
 *	@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
1381
 *
1382 1383 1384
 *	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
1385 1386
 *	specified %O_NONBLOCK to your open().
 */
J
J. Bruce Fields 已提交
1387
int __break_lease(struct inode *inode, unsigned int mode, unsigned int type)
L
Linus Torvalds 已提交
1388
{
1389
	int error = 0;
1390
	struct file_lock *new_fl;
1391 1392
	struct file_lock_context *ctx = inode->i_flctx;
	struct file_lock *fl;
L
Linus Torvalds 已提交
1393
	unsigned long break_time;
1394
	int want_write = (mode & O_ACCMODE) != O_RDONLY;
1395
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
1396

1397
	new_fl = lease_alloc(NULL, want_write ? F_WRLCK : F_RDLCK);
1398 1399
	if (IS_ERR(new_fl))
		return PTR_ERR(new_fl);
J
J. Bruce Fields 已提交
1400
	new_fl->fl_flags = type;
L
Linus Torvalds 已提交
1401

1402 1403 1404 1405 1406 1407
	/* typically we will check that ctx is non-NULL before calling */
	if (!ctx) {
		WARN_ON_ONCE(1);
		return error;
	}

1408
	spin_lock(&ctx->flc_lock);
L
Linus Torvalds 已提交
1409

1410
	time_out_leases(inode, &dispose);
L
Linus Torvalds 已提交
1411

1412
	if (!any_leases_conflict(inode, new_fl))
1413 1414
		goto out;

L
Linus Torvalds 已提交
1415 1416 1417 1418 1419 1420 1421
	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 */
	}

1422
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
J
J. Bruce Fields 已提交
1423 1424
		if (!leases_conflict(fl, new_fl))
			continue;
1425 1426 1427 1428
		if (want_write) {
			if (fl->fl_flags & FL_UNLOCK_PENDING)
				continue;
			fl->fl_flags |= FL_UNLOCK_PENDING;
L
Linus Torvalds 已提交
1429
			fl->fl_break_time = break_time;
1430
		} else {
1431
			if (lease_breaking(fl))
1432 1433 1434
				continue;
			fl->fl_flags |= FL_DOWNGRADE_PENDING;
			fl->fl_downgrade_time = break_time;
L
Linus Torvalds 已提交
1435
		}
J
Jeff Layton 已提交
1436
		if (fl->fl_lmops->lm_break(fl))
1437
			locks_delete_lock_ctx(fl, &dispose);
L
Linus Torvalds 已提交
1438 1439
	}

1440
	if (list_empty(&ctx->flc_lease))
J
Jeff Layton 已提交
1441 1442
		goto out;

1443
	if (mode & O_NONBLOCK) {
1444
		trace_break_lease_noblock(inode, new_fl);
L
Linus Torvalds 已提交
1445 1446 1447 1448 1449
		error = -EWOULDBLOCK;
		goto out;
	}

restart:
1450 1451
	fl = list_first_entry(&ctx->flc_lease, struct file_lock, fl_list);
	break_time = fl->fl_break_time;
1452
	if (break_time != 0)
L
Linus Torvalds 已提交
1453
		break_time -= jiffies;
1454 1455
	if (break_time == 0)
		break_time++;
1456
	locks_insert_block(fl, new_fl);
1457
	trace_break_lease_block(inode, new_fl);
1458
	spin_unlock(&ctx->flc_lock);
1459
	locks_dispose_list(&dispose);
1460 1461
	error = wait_event_interruptible_timeout(new_fl->fl_wait,
						!new_fl->fl_next, break_time);
1462
	spin_lock(&ctx->flc_lock);
1463
	trace_break_lease_unblock(inode, new_fl);
1464
	locks_delete_block(new_fl);
L
Linus Torvalds 已提交
1465
	if (error >= 0) {
1466 1467 1468 1469
		/*
		 * Wait for the next conflicting lease that has not been
		 * broken yet
		 */
1470 1471 1472 1473
		if (error == 0)
			time_out_leases(inode, &dispose);
		if (any_leases_conflict(inode, new_fl))
			goto restart;
L
Linus Torvalds 已提交
1474 1475 1476
		error = 0;
	}
out:
1477
	spin_unlock(&ctx->flc_lock);
1478
	locks_dispose_list(&dispose);
1479
	locks_free_lock(new_fl);
L
Linus Torvalds 已提交
1480 1481 1482 1483 1484 1485
	return error;
}

EXPORT_SYMBOL(__break_lease);

/**
1486
 *	lease_get_mtime - get the last modified time of an inode
L
Linus Torvalds 已提交
1487 1488 1489 1490 1491
 *	@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
1492
 * exclusive lease, then they could be modifying it.
L
Linus Torvalds 已提交
1493 1494 1495
 */
void lease_get_mtime(struct inode *inode, struct timespec *time)
{
1496
	bool has_lease = false;
1497 1498
	struct file_lock_context *ctx = inode->i_flctx;
	struct file_lock *fl;
1499

1500
	if (ctx && !list_empty_careful(&ctx->flc_lease)) {
1501
		spin_lock(&ctx->flc_lock);
1502 1503 1504 1505 1506 1507
		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;
		}
1508
		spin_unlock(&ctx->flc_lock);
1509 1510 1511
	}

	if (has_lease)
L
Linus Torvalds 已提交
1512 1513 1514 1515 1516 1517 1518 1519 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
		*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;
1545
	struct inode *inode = file_inode(filp);
1546
	struct file_lock_context *ctx = inode->i_flctx;
L
Linus Torvalds 已提交
1547
	int type = F_UNLCK;
1548
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
1549

1550
	if (ctx && !list_empty_careful(&ctx->flc_lease)) {
1551
		spin_lock(&ctx->flc_lock);
1552 1553 1554 1555
		time_out_leases(file_inode(filp), &dispose);
		list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
			if (fl->fl_file != filp)
				continue;
1556
			type = target_leasetype(fl);
L
Linus Torvalds 已提交
1557 1558
			break;
		}
1559
		spin_unlock(&ctx->flc_lock);
1560
		locks_dispose_list(&dispose);
L
Linus Torvalds 已提交
1561 1562 1563 1564
	}
	return type;
}

1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575
/**
 * 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
C
Christoph Hellwig 已提交
1576
check_conflicting_open(const struct dentry *dentry, const long arg, int flags)
1577 1578 1579 1580
{
	int ret = 0;
	struct inode *inode = dentry->d_inode;

C
Christoph Hellwig 已提交
1581 1582 1583
	if (flags & FL_LAYOUT)
		return 0;

1584 1585 1586 1587 1588 1589 1590 1591 1592 1593
	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;
}

1594 1595
static int
generic_add_lease(struct file *filp, long arg, struct file_lock **flp, void **priv)
L
Linus Torvalds 已提交
1596
{
1597
	struct file_lock *fl, *my_fl = NULL, *lease;
1598
	struct dentry *dentry = filp->f_path.dentry;
L
Linus Torvalds 已提交
1599
	struct inode *inode = dentry->d_inode;
1600
	struct file_lock_context *ctx;
J
J. Bruce Fields 已提交
1601
	bool is_deleg = (*flp)->fl_flags & FL_DELEG;
J
J. Bruce Fields 已提交
1602
	int error;
1603
	LIST_HEAD(dispose);
L
Linus Torvalds 已提交
1604

1605
	lease = *flp;
1606 1607
	trace_generic_add_lease(inode, lease);

1608 1609 1610 1611
	ctx = locks_get_lock_context(inode);
	if (!ctx)
		return -ENOMEM;

J
J. Bruce Fields 已提交
1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624
	/*
	 * 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: */
1625
		mutex_unlock(&inode->i_mutex);
J
J. Bruce Fields 已提交
1626 1627 1628
		WARN_ON_ONCE(1);
		return -EINVAL;
	}
1629

1630
	spin_lock(&ctx->flc_lock);
1631
	time_out_leases(inode, &dispose);
C
Christoph Hellwig 已提交
1632
	error = check_conflicting_open(dentry, arg, lease->fl_flags);
1633
	if (error)
1634
		goto out;
1635

L
Linus Torvalds 已提交
1636 1637 1638 1639 1640 1641 1642 1643
	/*
	 * 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
1644
	error = -EAGAIN;
1645
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
1646 1647
		if (fl->fl_file == filp &&
		    fl->fl_owner == lease->fl_owner) {
1648
			my_fl = fl;
J
J. Bruce Fields 已提交
1649 1650
			continue;
		}
1651

J
J. Bruce Fields 已提交
1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663
		/*
		 * 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
1664 1665
	}

1666
	if (my_fl != NULL) {
1667
		error = lease->fl_lmops->lm_change(my_fl, arg, &dispose);
1668 1669 1670
		if (error)
			goto out;
		goto out_setup;
L
Linus Torvalds 已提交
1671 1672 1673 1674 1675 1676
	}

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

1677
	locks_insert_lock_ctx(lease, &ctx->flc_lease);
1678 1679 1680 1681 1682 1683 1684 1685 1686 1687
	/*
	 * 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();
C
Christoph Hellwig 已提交
1688
	error = check_conflicting_open(dentry, arg, lease->fl_flags);
1689
	if (error) {
1690
		locks_unlink_lock_ctx(lease);
1691 1692
		goto out;
	}
1693 1694 1695 1696

out_setup:
	if (lease->fl_lmops->lm_setup)
		lease->fl_lmops->lm_setup(lease, priv);
L
Linus Torvalds 已提交
1697
out:
1698
	spin_unlock(&ctx->flc_lock);
1699
	locks_dispose_list(&dispose);
J
J. Bruce Fields 已提交
1700 1701
	if (is_deleg)
		mutex_unlock(&inode->i_mutex);
1702
	if (!error && !my_fl)
1703
		*flp = NULL;
L
Linus Torvalds 已提交
1704 1705
	return error;
}
1706

1707
static int generic_delete_lease(struct file *filp, void *owner)
1708
{
1709
	int error = -EAGAIN;
1710
	struct file_lock *fl, *victim = NULL;
1711 1712
	struct dentry *dentry = filp->f_path.dentry;
	struct inode *inode = dentry->d_inode;
1713
	struct file_lock_context *ctx = inode->i_flctx;
1714
	LIST_HEAD(dispose);
1715

1716 1717 1718 1719 1720
	if (!ctx) {
		trace_generic_delete_lease(inode, NULL);
		return error;
	}

1721
	spin_lock(&ctx->flc_lock);
1722
	list_for_each_entry(fl, &ctx->flc_lease, fl_list) {
1723 1724
		if (fl->fl_file == filp &&
		    fl->fl_owner == owner) {
1725
			victim = fl;
1726
			break;
1727
		}
1728
	}
1729
	trace_generic_delete_lease(inode, fl);
1730
	if (victim)
1731
		error = fl->fl_lmops->lm_change(victim, F_UNLCK, &dispose);
1732
	spin_unlock(&ctx->flc_lock);
1733
	locks_dispose_list(&dispose);
1734
	return error;
1735 1736 1737 1738
}

/**
 *	generic_setlease	-	sets a lease on an open file
1739 1740 1741 1742 1743
 *	@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)
1744 1745 1746 1747
 *
 *	The (input) flp->fl_lmops->lm_break function is required
 *	by break_lease().
 */
1748 1749
int generic_setlease(struct file *filp, long arg, struct file_lock **flp,
			void **priv)
1750 1751 1752 1753 1754
{
	struct dentry *dentry = filp->f_path.dentry;
	struct inode *inode = dentry->d_inode;
	int error;

1755
	if ((!uid_eq(current_fsuid(), inode->i_uid)) && !capable(CAP_LEASE))
1756 1757 1758 1759 1760 1761 1762 1763 1764
		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:
1765
		return generic_delete_lease(filp, *priv);
1766 1767
	case F_RDLCK:
	case F_WRLCK:
1768 1769 1770 1771
		if (!(*flp)->fl_lmops->lm_break) {
			WARN_ON_ONCE(1);
			return -ENOLCK;
		}
C
Christoph Hellwig 已提交
1772

1773
		return generic_add_lease(filp, arg, flp, priv);
1774
	default:
1775
		return -EINVAL;
1776 1777
	}
}
1778
EXPORT_SYMBOL(generic_setlease);
L
Linus Torvalds 已提交
1779

1780
/**
1781
 * vfs_setlease        -       sets a lease on an open file
1782 1783 1784 1785 1786
 * @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)
1787 1788 1789 1790 1791 1792
 *
 * 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).
1793 1794 1795
 *
 * 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
1796
 */
1797 1798
int
vfs_setlease(struct file *filp, long arg, struct file_lock **lease, void **priv)
L
Linus Torvalds 已提交
1799
{
1800
	if (filp->f_op->setlease)
1801
		return filp->f_op->setlease(filp, arg, lease, priv);
1802
	else
1803
		return generic_setlease(filp, arg, lease, priv);
L
Linus Torvalds 已提交
1804
}
1805
EXPORT_SYMBOL_GPL(vfs_setlease);
L
Linus Torvalds 已提交
1806

1807
static int do_fcntl_add_lease(unsigned int fd, struct file *filp, long arg)
L
Linus Torvalds 已提交
1808
{
1809
	struct file_lock *fl;
1810
	struct fasync_struct *new;
L
Linus Torvalds 已提交
1811 1812
	int error;

1813 1814 1815
	fl = lease_alloc(filp, arg);
	if (IS_ERR(fl))
		return PTR_ERR(fl);
L
Linus Torvalds 已提交
1816

1817 1818 1819 1820 1821
	new = fasync_alloc();
	if (!new) {
		locks_free_lock(fl);
		return -ENOMEM;
	}
1822
	new->fa_fd = fd;
1823

1824
	error = vfs_setlease(filp, arg, &fl, (void **)&new);
1825 1826
	if (fl)
		locks_free_lock(fl);
1827 1828
	if (new)
		fasync_free(new);
L
Linus Torvalds 已提交
1829 1830 1831
	return error;
}

1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844
/**
 *	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)
1845
		return vfs_setlease(filp, F_UNLCK, NULL, (void **)&filp);
1846 1847 1848
	return do_fcntl_add_lease(fd, filp, arg);
}

L
Linus Torvalds 已提交
1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861
/**
 * 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);
1862
		if (error != FILE_LOCK_DEFERRED)
L
Linus Torvalds 已提交
1863 1864 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
			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.
 */
1895
SYSCALL_DEFINE2(flock, unsigned int, fd, unsigned int, cmd)
L
Linus Torvalds 已提交
1896
{
1897
	struct fd f = fdget(fd);
L
Linus Torvalds 已提交
1898 1899 1900 1901 1902
	struct file_lock *lock;
	int can_sleep, unlock;
	int error;

	error = -EBADF;
1903
	if (!f.file)
L
Linus Torvalds 已提交
1904 1905 1906 1907 1908 1909
		goto out;

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

1910
	if (!unlock && !(cmd & LOCK_MAND) &&
1911
	    !(f.file->f_mode & (FMODE_READ|FMODE_WRITE)))
L
Linus Torvalds 已提交
1912 1913
		goto out_putf;

1914 1915 1916
	lock = flock_make_lock(f.file, cmd);
	if (IS_ERR(lock)) {
		error = PTR_ERR(lock);
L
Linus Torvalds 已提交
1917
		goto out_putf;
1918 1919
	}

L
Linus Torvalds 已提交
1920 1921 1922
	if (can_sleep)
		lock->fl_flags |= FL_SLEEP;

1923
	error = security_file_lock(f.file, lock->fl_type);
L
Linus Torvalds 已提交
1924 1925 1926
	if (error)
		goto out_free;

A
Al Viro 已提交
1927
	if (f.file->f_op->flock)
1928
		error = f.file->f_op->flock(f.file,
L
Linus Torvalds 已提交
1929 1930 1931
					  (can_sleep) ? F_SETLKW : F_SETLK,
					  lock);
	else
1932
		error = flock_lock_file_wait(f.file, lock);
L
Linus Torvalds 已提交
1933 1934

 out_free:
1935
	locks_free_lock(lock);
L
Linus Torvalds 已提交
1936 1937

 out_putf:
1938
	fdput(f);
L
Linus Torvalds 已提交
1939 1940 1941 1942
 out:
	return error;
}

1943 1944 1945
/**
 * vfs_test_lock - test file byte range lock
 * @filp: The file to test lock for
1946
 * @fl: The lock to test; also used to hold result
1947 1948 1949 1950 1951 1952
 *
 * 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
1953
	if (filp->f_op->lock)
1954 1955 1956 1957 1958 1959
		return filp->f_op->lock(filp, F_GETLK, fl);
	posix_test_lock(filp, fl);
	return 0;
}
EXPORT_SYMBOL_GPL(vfs_test_lock);

1960 1961
static int posix_lock_to_flock(struct flock *flock, struct file_lock *fl)
{
1962
	flock->l_pid = IS_OFDLCK(fl) ? -1 : fl->fl_pid;
1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976
#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;
1977
	flock->l_type = fl->fl_type;
1978 1979 1980 1981 1982 1983
	return 0;
}

#if BITS_PER_LONG == 32
static void posix_lock_to_flock64(struct flock64 *flock, struct file_lock *fl)
{
1984
	flock->l_pid = IS_OFDLCK(fl) ? -1 : fl->fl_pid;
1985 1986 1987 1988 1989 1990 1991 1992
	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
1993 1994 1995
/* Report the first existing lock that would conflict with l.
 * This implements the F_GETLK command of fcntl().
 */
1996
int fcntl_getlk(struct file *filp, unsigned int cmd, struct flock __user *l)
L
Linus Torvalds 已提交
1997
{
1998
	struct file_lock file_lock;
L
Linus Torvalds 已提交
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
	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;

2013
	if (cmd == F_OFD_GETLK) {
2014 2015 2016 2017
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2018
		cmd = F_GETLK;
2019
		file_lock.fl_flags |= FL_OFDLCK;
2020
		file_lock.fl_owner = filp;
2021 2022
	}

2023 2024 2025
	error = vfs_test_lock(filp, &file_lock);
	if (error)
		goto out;
L
Linus Torvalds 已提交
2026
 
2027 2028 2029
	flock.l_type = file_lock.fl_type;
	if (file_lock.fl_type != F_UNLCK) {
		error = posix_lock_to_flock(&flock, &file_lock);
2030
		if (error)
2031
			goto rel_priv;
L
Linus Torvalds 已提交
2032 2033 2034 2035
	}
	error = -EFAULT;
	if (!copy_to_user(l, &flock, sizeof(flock)))
		error = 0;
2036 2037
rel_priv:
	locks_release_private(&file_lock);
L
Linus Torvalds 已提交
2038 2039 2040 2041
out:
	return error;
}

2042 2043 2044 2045 2046
/**
 * 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
2047 2048 2049 2050 2051 2052 2053 2054
 * @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.
2055 2056 2057 2058
 *
 * 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
2059
 * lm_grant is set. Callers expecting ->lock() to return asynchronously
2060 2061
 * 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
2062
 * it must return FILE_LOCK_DEFERRED, and call ->lm_grant() when the lock
2063 2064
 * request completes.
 * If the request is for non-blocking lock the file system should return
2065 2066
 * 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
2067 2068 2069 2070 2071
 * 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
2072
 * ->lm_grant() before returning to the caller with a FILE_LOCK_DEFERRED
2073
 * return code.
2074
 */
2075
int vfs_lock_file(struct file *filp, unsigned int cmd, struct file_lock *fl, struct file_lock *conf)
2076
{
A
Al Viro 已提交
2077
	if (filp->f_op->lock)
2078 2079
		return filp->f_op->lock(filp, cmd, fl);
	else
2080
		return posix_lock_file(filp, fl, conf);
2081 2082 2083
}
EXPORT_SYMBOL_GPL(vfs_lock_file);

M
Miklos Szeredi 已提交
2084 2085 2086 2087 2088 2089 2090 2091 2092
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;

2093 2094 2095
	for (;;) {
		error = vfs_lock_file(filp, cmd, fl, NULL);
		if (error != FILE_LOCK_DEFERRED)
M
Miklos Szeredi 已提交
2096
			break;
2097 2098 2099 2100 2101 2102
		error = wait_event_interruptible(fl->fl_wait, !fl->fl_next);
		if (!error)
			continue;

		locks_delete_block(fl);
		break;
M
Miklos Szeredi 已提交
2103 2104 2105 2106 2107
	}

	return error;
}

2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123
/* 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
2124 2125 2126
/* Apply the lock described by l to an open file descriptor.
 * This implements both the F_SETLK and F_SETLKW commands of fcntl().
 */
2127 2128
int fcntl_setlk(unsigned int fd, struct file *filp, unsigned int cmd,
		struct flock __user *l)
L
Linus Torvalds 已提交
2129 2130 2131 2132
{
	struct file_lock *file_lock = locks_alloc_lock();
	struct flock flock;
	struct inode *inode;
2133
	struct file *f;
L
Linus Torvalds 已提交
2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145
	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
2146
	inode = file_inode(filp);
L
Linus Torvalds 已提交
2147 2148 2149 2150

	/* Don't allow mandatory locks on files that may be memory mapped
	 * and shared.
	 */
2151
	if (mandatory_lock(inode) && mapping_writably_mapped(filp->f_mapping)) {
L
Linus Torvalds 已提交
2152 2153 2154 2155
		error = -EAGAIN;
		goto out;
	}

2156
again:
L
Linus Torvalds 已提交
2157 2158 2159
	error = flock_to_posix_lock(filp, file_lock, &flock);
	if (error)
		goto out;
2160

2161 2162 2163 2164
	error = check_fmode_for_setlk(file_lock);
	if (error)
		goto out;

2165 2166
	/*
	 * If the cmd is requesting file-private locks, then set the
2167
	 * FL_OFDLCK flag and override the owner.
2168 2169
	 */
	switch (cmd) {
2170
	case F_OFD_SETLK:
2171 2172 2173 2174
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2175
		cmd = F_SETLK;
2176
		file_lock->fl_flags |= FL_OFDLCK;
2177
		file_lock->fl_owner = filp;
2178
		break;
2179
	case F_OFD_SETLKW:
2180 2181 2182 2183
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2184
		cmd = F_SETLKW;
2185
		file_lock->fl_flags |= FL_OFDLCK;
2186
		file_lock->fl_owner = filp;
2187 2188
		/* Fallthrough */
	case F_SETLKW:
L
Linus Torvalds 已提交
2189 2190
		file_lock->fl_flags |= FL_SLEEP;
	}
2191

M
Miklos Szeredi 已提交
2192
	error = do_lock_file_wait(filp, cmd, file_lock);
L
Linus Torvalds 已提交
2193

2194 2195 2196 2197
	/*
	 * Attempt to detect a close/fcntl race and recover by
	 * releasing the lock that was just acquired.
	 */
2198 2199
	/*
	 * we need that spin_lock here - it prevents reordering between
2200
	 * update of i_flctx->flc_posix and check for it done in close().
2201 2202 2203 2204 2205 2206
	 * 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) {
2207 2208
		flock.l_type = F_UNLCK;
		goto again;
L
Linus Torvalds 已提交
2209 2210
	}

2211
out:
L
Linus Torvalds 已提交
2212 2213 2214 2215 2216 2217 2218 2219
	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().
 */
2220
int fcntl_getlk64(struct file *filp, unsigned int cmd, struct flock64 __user *l)
L
Linus Torvalds 已提交
2221
{
2222
	struct file_lock file_lock;
L
Linus Torvalds 已提交
2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236
	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;

2237
	if (cmd == F_OFD_GETLK) {
2238 2239 2240 2241
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2242
		cmd = F_GETLK64;
2243
		file_lock.fl_flags |= FL_OFDLCK;
2244
		file_lock.fl_owner = filp;
2245 2246
	}

2247 2248 2249 2250
	error = vfs_test_lock(filp, &file_lock);
	if (error)
		goto out;

2251 2252 2253 2254
	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
2255 2256 2257
	error = -EFAULT;
	if (!copy_to_user(l, &flock, sizeof(flock)))
		error = 0;
2258 2259

	locks_release_private(&file_lock);
L
Linus Torvalds 已提交
2260 2261 2262 2263 2264 2265 2266
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().
 */
2267 2268
int fcntl_setlk64(unsigned int fd, struct file *filp, unsigned int cmd,
		struct flock64 __user *l)
L
Linus Torvalds 已提交
2269 2270 2271 2272
{
	struct file_lock *file_lock = locks_alloc_lock();
	struct flock64 flock;
	struct inode *inode;
2273
	struct file *f;
L
Linus Torvalds 已提交
2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285
	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
2286
	inode = file_inode(filp);
L
Linus Torvalds 已提交
2287 2288 2289 2290

	/* Don't allow mandatory locks on files that may be memory mapped
	 * and shared.
	 */
2291
	if (mandatory_lock(inode) && mapping_writably_mapped(filp->f_mapping)) {
L
Linus Torvalds 已提交
2292 2293 2294 2295
		error = -EAGAIN;
		goto out;
	}

2296
again:
L
Linus Torvalds 已提交
2297 2298 2299
	error = flock64_to_posix_lock(filp, file_lock, &flock);
	if (error)
		goto out;
2300

2301 2302 2303 2304
	error = check_fmode_for_setlk(file_lock);
	if (error)
		goto out;

2305 2306
	/*
	 * If the cmd is requesting file-private locks, then set the
2307
	 * FL_OFDLCK flag and override the owner.
2308 2309
	 */
	switch (cmd) {
2310
	case F_OFD_SETLK:
2311 2312 2313 2314
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2315
		cmd = F_SETLK64;
2316
		file_lock->fl_flags |= FL_OFDLCK;
2317
		file_lock->fl_owner = filp;
2318
		break;
2319
	case F_OFD_SETLKW:
2320 2321 2322 2323
		error = -EINVAL;
		if (flock.l_pid != 0)
			goto out;

2324
		cmd = F_SETLKW64;
2325
		file_lock->fl_flags |= FL_OFDLCK;
2326
		file_lock->fl_owner = filp;
2327 2328
		/* Fallthrough */
	case F_SETLKW64:
L
Linus Torvalds 已提交
2329 2330
		file_lock->fl_flags |= FL_SLEEP;
	}
2331

M
Miklos Szeredi 已提交
2332
	error = do_lock_file_wait(filp, cmd, file_lock);
L
Linus Torvalds 已提交
2333

2334 2335 2336 2337
	/*
	 * Attempt to detect a close/fcntl race and recover by
	 * releasing the lock that was just acquired.
	 */
2338 2339 2340 2341
	spin_lock(&current->files->file_lock);
	f = fcheck(fd);
	spin_unlock(&current->files->file_lock);
	if (!error && f != filp && flock.l_type != F_UNLCK) {
2342 2343
		flock.l_type = F_UNLCK;
		goto again;
L
Linus Torvalds 已提交
2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358
	}

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)
{
2359
	struct file_lock lock;
2360
	struct file_lock_context *ctx = file_inode(filp)->i_flctx;
L
Linus Torvalds 已提交
2361 2362 2363 2364 2365 2366

	/*
	 * 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.
	 */
2367
	if (!ctx || list_empty(&ctx->flc_posix))
L
Linus Torvalds 已提交
2368 2369 2370
		return;

	lock.fl_type = F_UNLCK;
2371
	lock.fl_flags = FL_POSIX | FL_CLOSE;
L
Linus Torvalds 已提交
2372 2373 2374 2375 2376 2377 2378 2379
	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;

2380
	vfs_lock_file(filp, F_SETLK, &lock, NULL);
L
Linus Torvalds 已提交
2381 2382 2383 2384 2385 2386 2387

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

EXPORT_SYMBOL(locks_remove_posix);

2388
/* The i_flctx must be valid when calling into here */
2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399
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,
	};
2400
	struct file_lock_context *flctx = file_inode(filp)->i_flctx;
2401

2402
	if (list_empty(&flctx->flc_flock))
2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413
		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);
}

2414
/* The i_flctx must be valid when calling into here */
2415 2416 2417 2418 2419 2420 2421 2422
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);

2423
	if (list_empty(&ctx->flc_lease))
2424 2425
		return;

2426
	spin_lock(&ctx->flc_lock);
2427
	list_for_each_entry_safe(fl, tmp, &ctx->flc_lease, fl_list)
2428 2429
		if (filp == fl->fl_file)
			lease_modify(fl, F_UNLCK, &dispose);
2430
	spin_unlock(&ctx->flc_lock);
2431 2432 2433
	locks_dispose_list(&dispose);
}

L
Linus Torvalds 已提交
2434 2435 2436
/*
 * This function is called on the last close of an open file.
 */
2437
void locks_remove_file(struct file *filp)
L
Linus Torvalds 已提交
2438
{
2439 2440 2441
	if (!file_inode(filp)->i_flctx)
		return;

2442
	/* remove any OFD locks */
2443
	locks_remove_posix(filp, filp);
2444

2445 2446 2447
	/* remove flock locks */
	locks_remove_flock(filp);

2448 2449
	/* remove any leases */
	locks_remove_lease(filp);
L
Linus Torvalds 已提交
2450 2451 2452 2453 2454 2455 2456 2457
}

/**
 *	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
2458
int
2459
posix_unblock_lock(struct file_lock *waiter)
L
Linus Torvalds 已提交
2460
{
J
J. Bruce Fields 已提交
2461 2462
	int status = 0;

2463
	spin_lock(&blocked_lock_lock);
2464
	if (waiter->fl_next)
L
Linus Torvalds 已提交
2465
		__locks_delete_block(waiter);
J
J. Bruce Fields 已提交
2466 2467
	else
		status = -ENOENT;
2468
	spin_unlock(&blocked_lock_lock);
J
J. Bruce Fields 已提交
2469
	return status;
L
Linus Torvalds 已提交
2470 2471 2472
}
EXPORT_SYMBOL(posix_unblock_lock);

M
Marc Eshel 已提交
2473 2474 2475 2476 2477 2478 2479 2480 2481
/**
 * 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
2482
	if (filp->f_op->lock)
M
Marc Eshel 已提交
2483 2484 2485 2486 2487 2488
		return filp->f_op->lock(filp, F_CANCELLK, fl);
	return 0;
}

EXPORT_SYMBOL_GPL(vfs_cancel_lock);

2489
#ifdef CONFIG_PROC_FS
2490
#include <linux/proc_fs.h>
2491 2492
#include <linux/seq_file.h>

2493 2494 2495 2496 2497
struct locks_iterator {
	int	li_cpu;
	loff_t	li_pos;
};

2498
static void lock_get_status(struct seq_file *f, struct file_lock *fl,
2499
			    loff_t id, char *pfx)
L
Linus Torvalds 已提交
2500 2501
{
	struct inode *inode = NULL;
2502 2503 2504
	unsigned int fl_pid;

	if (fl->fl_nspid)
2505
		fl_pid = pid_vnr(fl->fl_nspid);
2506 2507
	else
		fl_pid = fl->fl_pid;
L
Linus Torvalds 已提交
2508 2509

	if (fl->fl_file != NULL)
A
Al Viro 已提交
2510
		inode = file_inode(fl->fl_file);
L
Linus Torvalds 已提交
2511

2512
	seq_printf(f, "%lld:%s ", id, pfx);
L
Linus Torvalds 已提交
2513
	if (IS_POSIX(fl)) {
2514
		if (fl->fl_flags & FL_ACCESS)
2515
			seq_puts(f, "ACCESS");
2516
		else if (IS_OFDLCK(fl))
2517
			seq_puts(f, "OFDLCK");
2518
		else
2519
			seq_puts(f, "POSIX ");
2520 2521

		seq_printf(f, " %s ",
L
Linus Torvalds 已提交
2522
			     (inode == NULL) ? "*NOINODE*" :
2523
			     mandatory_lock(inode) ? "MANDATORY" : "ADVISORY ");
L
Linus Torvalds 已提交
2524 2525
	} else if (IS_FLOCK(fl)) {
		if (fl->fl_type & LOCK_MAND) {
2526
			seq_puts(f, "FLOCK  MSNFS     ");
L
Linus Torvalds 已提交
2527
		} else {
2528
			seq_puts(f, "FLOCK  ADVISORY  ");
L
Linus Torvalds 已提交
2529 2530
		}
	} else if (IS_LEASE(fl)) {
2531 2532 2533 2534 2535
		if (fl->fl_flags & FL_DELEG)
			seq_puts(f, "DELEG  ");
		else
			seq_puts(f, "LEASE  ");

J
J. Bruce Fields 已提交
2536
		if (lease_breaking(fl))
2537
			seq_puts(f, "BREAKING  ");
L
Linus Torvalds 已提交
2538
		else if (fl->fl_file)
2539
			seq_puts(f, "ACTIVE    ");
L
Linus Torvalds 已提交
2540
		else
2541
			seq_puts(f, "BREAKER   ");
L
Linus Torvalds 已提交
2542
	} else {
2543
		seq_puts(f, "UNKNOWN UNKNOWN  ");
L
Linus Torvalds 已提交
2544 2545
	}
	if (fl->fl_type & LOCK_MAND) {
2546
		seq_printf(f, "%s ",
L
Linus Torvalds 已提交
2547 2548 2549 2550
			       (fl->fl_type & LOCK_READ)
			       ? (fl->fl_type & LOCK_WRITE) ? "RW   " : "READ "
			       : (fl->fl_type & LOCK_WRITE) ? "WRITE" : "NONE ");
	} else {
2551
		seq_printf(f, "%s ",
J
J. Bruce Fields 已提交
2552
			       (lease_breaking(fl))
2553 2554
			       ? (fl->fl_type == F_UNLCK) ? "UNLCK" : "READ "
			       : (fl->fl_type == F_WRLCK) ? "WRITE" : "READ ");
L
Linus Torvalds 已提交
2555 2556 2557
	}
	if (inode) {
#ifdef WE_CAN_BREAK_LSLK_NOW
2558
		seq_printf(f, "%d %s:%ld ", fl_pid,
L
Linus Torvalds 已提交
2559 2560 2561
				inode->i_sb->s_id, inode->i_ino);
#else
		/* userspace relies on this representation of dev_t ;-( */
2562
		seq_printf(f, "%d %02x:%02x:%ld ", fl_pid,
L
Linus Torvalds 已提交
2563 2564 2565 2566
				MAJOR(inode->i_sb->s_dev),
				MINOR(inode->i_sb->s_dev), inode->i_ino);
#endif
	} else {
2567
		seq_printf(f, "%d <none>:0 ", fl_pid);
L
Linus Torvalds 已提交
2568 2569 2570
	}
	if (IS_POSIX(fl)) {
		if (fl->fl_end == OFFSET_MAX)
2571
			seq_printf(f, "%Ld EOF\n", fl->fl_start);
L
Linus Torvalds 已提交
2572
		else
2573
			seq_printf(f, "%Ld %Ld\n", fl->fl_start, fl->fl_end);
L
Linus Torvalds 已提交
2574
	} else {
2575
		seq_puts(f, "0 EOF\n");
L
Linus Torvalds 已提交
2576 2577 2578
	}
}

2579
static int locks_show(struct seq_file *f, void *v)
L
Linus Torvalds 已提交
2580
{
2581
	struct locks_iterator *iter = f->private;
2582
	struct file_lock *fl, *bfl;
L
Linus Torvalds 已提交
2583

2584
	fl = hlist_entry(v, struct file_lock, fl_link);
L
Linus Torvalds 已提交
2585

2586
	lock_get_status(f, fl, iter->li_pos, "");
L
Linus Torvalds 已提交
2587

2588
	list_for_each_entry(bfl, &fl->fl_block, fl_block)
2589
		lock_get_status(f, bfl, iter->li_pos, " ->");
2590

2591 2592
	return 0;
}
L
Linus Torvalds 已提交
2593

2594
static void *locks_start(struct seq_file *f, loff_t *pos)
2595
	__acquires(&blocked_lock_lock)
2596
{
2597
	struct locks_iterator *iter = f->private;
2598

2599 2600
	iter->li_pos = *pos + 1;
	lg_global_lock(&file_lock_lglock);
2601
	spin_lock(&blocked_lock_lock);
2602
	return seq_hlist_start_percpu(&file_lock_list, &iter->li_cpu, *pos);
2603
}
L
Linus Torvalds 已提交
2604

2605 2606
static void *locks_next(struct seq_file *f, void *v, loff_t *pos)
{
2607 2608 2609 2610
	struct locks_iterator *iter = f->private;

	++iter->li_pos;
	return seq_hlist_next_percpu(v, &file_lock_list, &iter->li_cpu, pos);
2611
}
L
Linus Torvalds 已提交
2612

2613
static void locks_stop(struct seq_file *f, void *v)
2614
	__releases(&blocked_lock_lock)
2615
{
2616
	spin_unlock(&blocked_lock_lock);
2617
	lg_global_unlock(&file_lock_lglock);
L
Linus Torvalds 已提交
2618 2619
}

2620
static const struct seq_operations locks_seq_operations = {
2621 2622 2623 2624 2625
	.start	= locks_start,
	.next	= locks_next,
	.stop	= locks_stop,
	.show	= locks_show,
};
2626 2627 2628

static int locks_open(struct inode *inode, struct file *filp)
{
2629 2630
	return seq_open_private(filp, &locks_seq_operations,
					sizeof(struct locks_iterator));
2631 2632 2633 2634 2635 2636
}

static const struct file_operations proc_locks_operations = {
	.open		= locks_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
2637
	.release	= seq_release_private,
2638 2639 2640 2641 2642 2643 2644 2645
};

static int __init proc_locks_init(void)
{
	proc_create("locks", 0, NULL, &proc_locks_operations);
	return 0;
}
module_init(proc_locks_init);
2646 2647
#endif

L
Linus Torvalds 已提交
2648 2649
static int __init filelock_init(void)
{
2650 2651
	int i;

2652 2653 2654
	flctx_cache = kmem_cache_create("file_lock_ctx",
			sizeof(struct file_lock_context), 0, SLAB_PANIC, NULL);

L
Linus Torvalds 已提交
2655
	filelock_cache = kmem_cache_create("file_lock_cache",
M
Miklos Szeredi 已提交
2656 2657
			sizeof(struct file_lock), 0, SLAB_PANIC, NULL);

2658 2659 2660 2661 2662
	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
2663 2664 2665 2666
	return 0;
}

core_initcall(filelock_init);