locks.c 77.3 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6 7 8 9 10 11 12 13
/*
 *  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
14
 *
Linus Torvalds's avatar
Linus Torvalds committed
15 16 17
 *  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
18
 *
Linus Torvalds's avatar
Linus Torvalds committed
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
 *  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
44
 *  unlocked).
Linus Torvalds's avatar
Linus Torvalds committed
45 46 47 48 49 50 51 52
 *
 *  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
53
 *  races. Just grep for FIXME to see them.
Linus Torvalds's avatar
Linus Torvalds committed
54 55 56 57 58 59 60 61 62
 *  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.
Paul Bolle's avatar
Paul Bolle committed
63
 *  See 'Documentation/filesystems/mandatory-locking.txt' for details.
Linus Torvalds's avatar
Linus Torvalds committed
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
 *  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.
115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
 *
 * Locking conflicts and dependencies:
 * If multiple threads attempt to lock the same byte (or flock the same file)
 * only one can be granted the lock, and other must wait their turn.
 * The first lock has been "applied" or "granted", the others are "waiting"
 * and are "blocked" by the "applied" lock..
 *
 * Waiting and applied locks are all kept in trees whose properties are:
 *
 *	- the root of a tree may be an applied or waiting lock.
 *	- every other node in the tree is a waiting lock that
 *	  conflicts with every ancestor of that node.
 *
 * Every such tree begins life as a waiting singleton which obviously
 * satisfies the above properties.
 *
 * The only ways we modify trees preserve these properties:
 *
 *	1. We may add a new leaf node, but only after first verifying that it
 *	   conflicts with all of its ancestors.
 *	2. We may remove the root of a tree, creating a new singleton
 *	   tree from the root and N new trees rooted in the immediate
 *	   children.
 *	3. If the root of a tree is not currently an applied lock, we may
 *	   apply it (if possible).
 *	4. We may upgrade the root of the tree (either extend its range,
 *	   or upgrade its entire range from read to write).
 *
 * When an applied lock is modified in a way that reduces or downgrades any
 * part of its range, we remove all its children (2 above).  This particularly
 * happens when a lock is unlocked.
 *
 * For each of those child trees we "wake up" the thread which is
 * waiting for the lock so it can continue handling as follows: if the
 * root of the tree applies, we do so (3).  If it doesn't, it must
 * conflict with some applied lock.  We remove (wake up) all of its children
 * (2), and add it is a new leaf to the tree rooted in the applied
 * lock (1).  We then repeat the process recursively with those
 * children.
 *
Linus Torvalds's avatar
Linus Torvalds committed
155 156 157 158
 */

#include <linux/capability.h>
#include <linux/file.h>
Al Viro's avatar
Al Viro committed
159
#include <linux/fdtable.h>
Linus Torvalds's avatar
Linus Torvalds committed
160 161 162 163 164 165
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/security.h>
#include <linux/slab.h>
#include <linux/syscalls.h>
#include <linux/time.h>
166
#include <linux/rcupdate.h>
167
#include <linux/pid_namespace.h>
168
#include <linux/hashtable.h>
169
#include <linux/percpu.h>
Linus Torvalds's avatar
Linus Torvalds committed
170

171 172 173
#define CREATE_TRACE_POINTS
#include <trace/events/filelock.h>

174
#include <linux/uaccess.h>
Linus Torvalds's avatar
Linus Torvalds committed
175 176 177

#define IS_POSIX(fl)	(fl->fl_flags & FL_POSIX)
#define IS_FLOCK(fl)	(fl->fl_flags & FL_FLOCK)
178
#define IS_LEASE(fl)	(fl->fl_flags & (FL_LEASE|FL_DELEG|FL_LAYOUT))
179
#define IS_OFDLCK(fl)	(fl->fl_flags & FL_OFDLCK)
180
#define IS_REMOTELCK(fl)	(fl->fl_pid <= 0)
Linus Torvalds's avatar
Linus Torvalds committed
181

J. Bruce Fields's avatar
J. Bruce Fields committed
182 183
static bool lease_breaking(struct file_lock *fl)
{
184 185 186 187 188 189 190 191 192 193
	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. Bruce Fields's avatar
J. Bruce Fields committed
194 195
}

Linus Torvalds's avatar
Linus Torvalds committed
196 197 198
int leases_enable = 1;
int lease_break_time = 45;

199
/*
200
 * The global file_lock_list is only used for displaying /proc/locks, so we
201 202 203 204 205
 * keep a list on each CPU, with each list protected by its own spinlock.
 * Global serialization is done using file_rwsem.
 *
 * Note that alterations to the list also require that the relevant flc_lock is
 * held.
206
 */
207 208 209 210 211
struct file_lock_list_struct {
	spinlock_t		lock;
	struct hlist_head	hlist;
};
static DEFINE_PER_CPU(struct file_lock_list_struct, file_lock_list);
212
DEFINE_STATIC_PERCPU_RWSEM(file_rwsem);
213

214
/*
215
 * The blocked_hash is used to find POSIX lock loops for deadlock detection.
216
 * It is protected by blocked_lock_lock.
217 218 219 220 221 222 223
 *
 * 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.
224
 */
225 226
#define BLOCKED_HASH_BITS	7
static DEFINE_HASHTABLE(blocked_hash, BLOCKED_HASH_BITS);
227

228
/*
229 230
 * This lock protects the blocked_hash. Generally, if you're accessing it, you
 * want to be holding this lock.
231
 *
232 233 234
 * In addition, it also protects the fl->fl_blocked_requests list, and the
 * fl->fl_blocker pointer for file_lock structures that are acting as lock
 * requests (in contrast to those that are acting as records of acquired locks).
235 236
 *
 * Note that when we acquire this lock in order to change the above fields,
237
 * we often hold the flc_lock as well. In certain cases, when reading the fields
238
 * protected by this lock, we can skip acquiring it iff we already hold the
239
 * flc_lock.
240
 */
241
static DEFINE_SPINLOCK(blocked_lock_lock);
Linus Torvalds's avatar
Linus Torvalds committed
242

243
static struct kmem_cache *flctx_cache __read_mostly;
244
static struct kmem_cache *filelock_cache __read_mostly;
Linus Torvalds's avatar
Linus Torvalds committed
245

246
static struct file_lock_context *
247
locks_get_lock_context(struct inode *inode, int type)
248
{
249
	struct file_lock_context *ctx;
250

251 252 253
	/* paired with cmpxchg() below */
	ctx = smp_load_acquire(&inode->i_flctx);
	if (likely(ctx) || type == F_UNLCK)
254 255
		goto out;

256 257
	ctx = kmem_cache_alloc(flctx_cache, GFP_KERNEL);
	if (!ctx)
258 259
		goto out;

260 261 262 263
	spin_lock_init(&ctx->flc_lock);
	INIT_LIST_HEAD(&ctx->flc_flock);
	INIT_LIST_HEAD(&ctx->flc_posix);
	INIT_LIST_HEAD(&ctx->flc_lease);
264 265 266 267 268

	/*
	 * Assign the pointer if it's not already assigned. If it is, then
	 * free the context we just allocated.
	 */
269 270 271 272
	if (cmpxchg(&inode->i_flctx, NULL, ctx)) {
		kmem_cache_free(flctx_cache, ctx);
		ctx = smp_load_acquire(&inode->i_flctx);
	}
273
out:
274
	trace_locks_get_lock_context(inode, type, ctx);
275
	return ctx;
276 277
}

278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304
static void
locks_dump_ctx_list(struct list_head *list, char *list_type)
{
	struct file_lock *fl;

	list_for_each_entry(fl, list, fl_list) {
		pr_warn("%s: fl_owner=%p fl_flags=0x%x fl_type=0x%x fl_pid=%u\n", list_type, fl->fl_owner, fl->fl_flags, fl->fl_type, fl->fl_pid);
	}
}

static void
locks_check_ctx_lists(struct inode *inode)
{
	struct file_lock_context *ctx = inode->i_flctx;

	if (unlikely(!list_empty(&ctx->flc_flock) ||
		     !list_empty(&ctx->flc_posix) ||
		     !list_empty(&ctx->flc_lease))) {
		pr_warn("Leaked locks on dev=0x%x:0x%x ino=0x%lx:\n",
			MAJOR(inode->i_sb->s_dev), MINOR(inode->i_sb->s_dev),
			inode->i_ino);
		locks_dump_ctx_list(&ctx->flc_flock, "FLOCK");
		locks_dump_ctx_list(&ctx->flc_posix, "POSIX");
		locks_dump_ctx_list(&ctx->flc_lease, "LEASE");
	}
}

305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320
static void
locks_check_ctx_file_list(struct file *filp, struct list_head *list,
				char *list_type)
{
	struct file_lock *fl;
	struct inode *inode = locks_inode(filp);

	list_for_each_entry(fl, list, fl_list)
		if (fl->fl_file == filp)
			pr_warn("Leaked %s lock on dev=0x%x:0x%x ino=0x%lx "
				" fl_owner=%p fl_flags=0x%x fl_type=0x%x fl_pid=%u\n",
				list_type, MAJOR(inode->i_sb->s_dev),
				MINOR(inode->i_sb->s_dev), inode->i_ino,
				fl->fl_owner, fl->fl_flags, fl->fl_type, fl->fl_pid);
}

321
void
322
locks_free_lock_context(struct inode *inode)
323
{
324 325
	struct file_lock_context *ctx = inode->i_flctx;

326 327
	if (unlikely(ctx)) {
		locks_check_ctx_lists(inode);
328 329 330 331
		kmem_cache_free(flctx_cache, ctx);
	}
}

Miklos Szeredi's avatar
Miklos Szeredi committed
332
static void locks_init_lock_heads(struct file_lock *fl)
Miklos Szeredi's avatar
Miklos Szeredi committed
333
{
334
	INIT_HLIST_NODE(&fl->fl_link);
335
	INIT_LIST_HEAD(&fl->fl_list);
336 337
	INIT_LIST_HEAD(&fl->fl_blocked_requests);
	INIT_LIST_HEAD(&fl->fl_blocked_member);
Miklos Szeredi's avatar
Miklos Szeredi committed
338
	init_waitqueue_head(&fl->fl_wait);
Miklos Szeredi's avatar
Miklos Szeredi committed
339 340
}

Linus Torvalds's avatar
Linus Torvalds committed
341
/* Allocate an empty lock structure. */
342
struct file_lock *locks_alloc_lock(void)
Linus Torvalds's avatar
Linus Torvalds committed
343
{
Miklos Szeredi's avatar
Miklos Szeredi committed
344
	struct file_lock *fl = kmem_cache_zalloc(filelock_cache, GFP_KERNEL);
Miklos Szeredi's avatar
Miklos Szeredi committed
345 346

	if (fl)
Miklos Szeredi's avatar
Miklos Szeredi committed
347
		locks_init_lock_heads(fl);
Miklos Szeredi's avatar
Miklos Szeredi committed
348 349

	return fl;
Linus Torvalds's avatar
Linus Torvalds committed
350
}
351
EXPORT_SYMBOL_GPL(locks_alloc_lock);
Linus Torvalds's avatar
Linus Torvalds committed
352

353
void locks_release_private(struct file_lock *fl)
354 355 356 357 358 359 360
{
	if (fl->fl_ops) {
		if (fl->fl_ops->fl_release_private)
			fl->fl_ops->fl_release_private(fl);
		fl->fl_ops = NULL;
	}

361
	if (fl->fl_lmops) {
362 363 364 365
		if (fl->fl_lmops->lm_put_owner) {
			fl->fl_lmops->lm_put_owner(fl->fl_owner);
			fl->fl_owner = NULL;
		}
366 367
		fl->fl_lmops = NULL;
	}
368
}
369
EXPORT_SYMBOL_GPL(locks_release_private);
370

Linus Torvalds's avatar
Linus Torvalds committed
371
/* Free a lock which is not in use. */
372
void locks_free_lock(struct file_lock *fl)
Linus Torvalds's avatar
Linus Torvalds committed
373
{
Miklos Szeredi's avatar
Miklos Szeredi committed
374
	BUG_ON(waitqueue_active(&fl->fl_wait));
375
	BUG_ON(!list_empty(&fl->fl_list));
376 377
	BUG_ON(!list_empty(&fl->fl_blocked_requests));
	BUG_ON(!list_empty(&fl->fl_blocked_member));
378
	BUG_ON(!hlist_unhashed(&fl->fl_link));
Linus Torvalds's avatar
Linus Torvalds committed
379

380
	locks_release_private(fl);
Linus Torvalds's avatar
Linus Torvalds committed
381 382
	kmem_cache_free(filelock_cache, fl);
}
383
EXPORT_SYMBOL(locks_free_lock);
Linus Torvalds's avatar
Linus Torvalds committed
384

385 386 387 388 389 390
static void
locks_dispose_list(struct list_head *dispose)
{
	struct file_lock *fl;

	while (!list_empty(dispose)) {
391 392
		fl = list_first_entry(dispose, struct file_lock, fl_list);
		list_del_init(&fl->fl_list);
393 394 395 396
		locks_free_lock(fl);
	}
}

Linus Torvalds's avatar
Linus Torvalds committed
397 398
void locks_init_lock(struct file_lock *fl)
{
Miklos Szeredi's avatar
Miklos Szeredi committed
399 400
	memset(fl, 0, sizeof(struct file_lock));
	locks_init_lock_heads(fl);
Linus Torvalds's avatar
Linus Torvalds committed
401 402 403 404 405 406
}
EXPORT_SYMBOL(locks_init_lock);

/*
 * Initialize a new lock from an existing file_lock structure.
 */
407
void locks_copy_conflock(struct file_lock *new, struct file_lock *fl)
Linus Torvalds's avatar
Linus Torvalds committed
408 409 410
{
	new->fl_owner = fl->fl_owner;
	new->fl_pid = fl->fl_pid;
411
	new->fl_file = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
412 413 414 415
	new->fl_flags = fl->fl_flags;
	new->fl_type = fl->fl_type;
	new->fl_start = fl->fl_start;
	new->fl_end = fl->fl_end;
416
	new->fl_lmops = fl->fl_lmops;
417
	new->fl_ops = NULL;
418 419 420

	if (fl->fl_lmops) {
		if (fl->fl_lmops->lm_get_owner)
421
			fl->fl_lmops->lm_get_owner(fl->fl_owner);
422
	}
423
}
424
EXPORT_SYMBOL(locks_copy_conflock);
425 426 427

void locks_copy_lock(struct file_lock *new, struct file_lock *fl)
{
428 429
	/* "new" must be a freshly-initialized lock */
	WARN_ON_ONCE(new->fl_ops);
430

431
	locks_copy_conflock(new, fl);
432

433
	new->fl_file = fl->fl_file;
Linus Torvalds's avatar
Linus Torvalds committed
434
	new->fl_ops = fl->fl_ops;
435

436 437 438 439
	if (fl->fl_ops) {
		if (fl->fl_ops->fl_copy_lock)
			fl->fl_ops->fl_copy_lock(new, fl);
	}
Linus Torvalds's avatar
Linus Torvalds committed
440 441 442
}
EXPORT_SYMBOL(locks_copy_lock);

443 444 445 446 447 448 449 450 451 452 453 454 455
static void locks_move_blocks(struct file_lock *new, struct file_lock *fl)
{
	struct file_lock *f;

	/*
	 * As ctx->flc_lock is held, new requests cannot be added to
	 * ->fl_blocked_requests, so we don't need a lock to check if it
	 * is empty.
	 */
	if (list_empty(&fl->fl_blocked_requests))
		return;
	spin_lock(&blocked_lock_lock);
	list_splice_init(&fl->fl_blocked_requests, &new->fl_blocked_requests);
456
	list_for_each_entry(f, &new->fl_blocked_requests, fl_blocked_member)
457 458 459 460
		f->fl_blocker = new;
	spin_unlock(&blocked_lock_lock);
}

Linus Torvalds's avatar
Linus Torvalds committed
461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
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. */
476
static struct file_lock *
477
flock_make_lock(struct file *filp, unsigned int cmd, struct file_lock *fl)
Linus Torvalds's avatar
Linus Torvalds committed
478 479
{
	int type = flock_translate_cmd(cmd);
480

Linus Torvalds's avatar
Linus Torvalds committed
481
	if (type < 0)
482
		return ERR_PTR(type);
483

484 485 486 487 488 489 490
	if (fl == NULL) {
		fl = locks_alloc_lock();
		if (fl == NULL)
			return ERR_PTR(-ENOMEM);
	} else {
		locks_init_lock(fl);
	}
Linus Torvalds's avatar
Linus Torvalds committed
491 492

	fl->fl_file = filp;
493
	fl->fl_owner = filp;
Linus Torvalds's avatar
Linus Torvalds committed
494 495 496 497
	fl->fl_pid = current->tgid;
	fl->fl_flags = FL_FLOCK;
	fl->fl_type = type;
	fl->fl_end = OFFSET_MAX;
498

499
	return fl;
Linus Torvalds's avatar
Linus Torvalds committed
500 501
}

502
static int assign_type(struct file_lock *fl, long type)
Linus Torvalds's avatar
Linus Torvalds committed
503 504 505 506 507 508 509 510 511 512 513 514 515
{
	switch (type) {
	case F_RDLCK:
	case F_WRLCK:
	case F_UNLCK:
		fl->fl_type = type;
		break;
	default:
		return -EINVAL;
	}
	return 0;
}

516 517
static int flock64_to_posix_lock(struct file *filp, struct file_lock *fl,
				 struct flock64 *l)
Linus Torvalds's avatar
Linus Torvalds committed
518 519
{
	switch (l->l_whence) {
520
	case SEEK_SET:
521
		fl->fl_start = 0;
Linus Torvalds's avatar
Linus Torvalds committed
522
		break;
523
	case SEEK_CUR:
524
		fl->fl_start = filp->f_pos;
Linus Torvalds's avatar
Linus Torvalds committed
525
		break;
526
	case SEEK_END:
527
		fl->fl_start = i_size_read(file_inode(filp));
Linus Torvalds's avatar
Linus Torvalds committed
528 529 530 531
		break;
	default:
		return -EINVAL;
	}
532 533 534 535 536
	if (l->l_start > OFFSET_MAX - fl->fl_start)
		return -EOVERFLOW;
	fl->fl_start += l->l_start;
	if (fl->fl_start < 0)
		return -EINVAL;
Linus Torvalds's avatar
Linus Torvalds committed
537 538 539

	/* POSIX-1996 leaves the case l->l_len < 0 undefined;
	   POSIX-2001 defines it. */
540
	if (l->l_len > 0) {
541 542 543 544
		if (l->l_len - 1 > OFFSET_MAX - fl->fl_start)
			return -EOVERFLOW;
		fl->fl_end = fl->fl_start + l->l_len - 1;

545
	} else if (l->l_len < 0) {
546
		if (fl->fl_start + l->l_len < 0)
547
			return -EINVAL;
548 549 550 551 552
		fl->fl_end = fl->fl_start - 1;
		fl->fl_start += l->l_len;
	} else
		fl->fl_end = OFFSET_MAX;

Linus Torvalds's avatar
Linus Torvalds committed
553 554 555 556 557 558 559 560 561 562
	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);
}

563 564 565 566 567
/* 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)
Linus Torvalds's avatar
Linus Torvalds committed
568
{
569 570 571 572 573 574 575 576
	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);
Linus Torvalds's avatar
Linus Torvalds committed
577 578 579
}

/* default lease lock manager operations */
580 581
static bool
lease_break_callback(struct file_lock *fl)
Linus Torvalds's avatar
Linus Torvalds committed
582 583
{
	kill_fasync(&fl->fl_fasync, SIGIO, POLL_MSG);
584
	return false;
Linus Torvalds's avatar
Linus Torvalds committed
585 586
}

587 588 589 590 591 592 593 594 595 596 597 598 599 600
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;

601
	__f_setown(filp, task_pid(current), PIDTYPE_TGID, 0);
602 603
}

604
static const struct lock_manager_operations lease_manager_ops = {
605 606
	.lm_break = lease_break_callback,
	.lm_change = lease_modify,
607
	.lm_setup = lease_setup,
Linus Torvalds's avatar
Linus Torvalds committed
608 609 610 611 612
};

/*
 * Initialize a lease, use the default lock manager operations
 */
613
static int lease_init(struct file *filp, long type, struct file_lock *fl)
614
{
615 616 617
	if (assign_type(fl, type) != 0)
		return -EINVAL;

618
	fl->fl_owner = filp;
Linus Torvalds's avatar
Linus Torvalds committed
619 620 621 622 623 624 625 626 627 628 629 630
	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 */
631
static struct file_lock *lease_alloc(struct file *filp, long type)
Linus Torvalds's avatar
Linus Torvalds committed
632 633
{
	struct file_lock *fl = locks_alloc_lock();
634
	int error = -ENOMEM;
Linus Torvalds's avatar
Linus Torvalds committed
635 636

	if (fl == NULL)
637
		return ERR_PTR(error);
Linus Torvalds's avatar
Linus Torvalds committed
638 639

	error = lease_init(filp, type, fl);
640 641
	if (error) {
		locks_free_lock(fl);
642
		return ERR_PTR(error);
643
	}
644
	return fl;
Linus Torvalds's avatar
Linus Torvalds committed
645 646 647 648 649 650 651 652 653 654 655 656 657
}

/* 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.
 */
658
static int posix_same_owner(struct file_lock *fl1, struct file_lock *fl2)
Linus Torvalds's avatar
Linus Torvalds committed
659
{
660
	if (fl1->fl_lmops && fl1->fl_lmops->lm_compare_owner)
Linus Torvalds's avatar
Linus Torvalds committed
661
		return fl2->fl_lmops == fl1->fl_lmops &&
662
			fl1->fl_lmops->lm_compare_owner(fl1, fl2);
Linus Torvalds's avatar
Linus Torvalds committed
663 664 665
	return fl1->fl_owner == fl2->fl_owner;
}

666
/* Must be called with the flc_lock held! */
667
static void locks_insert_global_locks(struct file_lock *fl)
668
{
669 670
	struct file_lock_list_struct *fll = this_cpu_ptr(&file_lock_list);

671 672
	percpu_rwsem_assert_held(&file_rwsem);

673
	spin_lock(&fll->lock);
674
	fl->fl_link_cpu = smp_processor_id();
675 676
	hlist_add_head(&fl->fl_link, &fll->hlist);
	spin_unlock(&fll->lock);
677 678
}

679
/* Must be called with the flc_lock held! */
680
static void locks_delete_global_locks(struct file_lock *fl)
681
{
682 683
	struct file_lock_list_struct *fll;

684 685
	percpu_rwsem_assert_held(&file_rwsem);

686 687
	/*
	 * Avoid taking lock if already unhashed. This is safe since this check
688
	 * is done while holding the flc_lock, and new insertions into the list
689 690 691 692
	 * also require that it be held.
	 */
	if (hlist_unhashed(&fl->fl_link))
		return;
693 694 695

	fll = per_cpu_ptr(&file_lock_list, fl->fl_link_cpu);
	spin_lock(&fll->lock);
696
	hlist_del_init(&fl->fl_link);
697
	spin_unlock(&fll->lock);
698 699
}

700 701 702 703 704 705 706 707
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;
}

708
static void locks_insert_global_blocked(struct file_lock *waiter)
709
{
710 711
	lockdep_assert_held(&blocked_lock_lock);

712
	hash_add(blocked_hash, &waiter->fl_link, posix_owner_key(waiter));
713 714
}

715
static void locks_delete_global_blocked(struct file_lock *waiter)
716
{
717 718
	lockdep_assert_held(&blocked_lock_lock);

719
	hash_del(&waiter->fl_link);
720 721
}

Linus Torvalds's avatar
Linus Torvalds committed
722 723
/* Remove waiter from blocker's block list.
 * When blocker ends up pointing to itself then the list is empty.
724
 *
725
 * Must be called with blocked_lock_lock held.
Linus Torvalds's avatar
Linus Torvalds committed
726
 */
727
static void __locks_delete_block(struct file_lock *waiter)
Linus Torvalds's avatar
Linus Torvalds committed
728
{
729
	locks_delete_global_blocked(waiter);
730 731
	list_del_init(&waiter->fl_blocked_member);
	waiter->fl_blocker = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
732 733
}

734 735 736 737 738 739 740 741 742 743 744 745 746 747 748
static void __locks_wake_up_blocks(struct file_lock *blocker)
{
	while (!list_empty(&blocker->fl_blocked_requests)) {
		struct file_lock *waiter;

		waiter = list_first_entry(&blocker->fl_blocked_requests,
					  struct file_lock, fl_blocked_member);
		__locks_delete_block(waiter);
		if (waiter->fl_lmops && waiter->fl_lmops->lm_notify)
			waiter->fl_lmops->lm_notify(waiter);
		else
			wake_up(&waiter->fl_wait);
	}
}

749 750 751 752 753 754 755
/**
 *	locks_delete_lock - stop waiting for a file lock
 *	@waiter: the lock which was waiting
 *
 *	lockd/nfsd need to disconnect the lock while working on it.
 */
int locks_delete_block(struct file_lock *waiter)
Linus Torvalds's avatar
Linus Torvalds committed
756
{
757 758
	int status = -ENOENT;

759 760 761 762 763 764 765 766 767 768 769 770 771
	/*
	 * If fl_blocker is NULL, it won't be set again as this thread
	 * "owns" the lock and is the only one that might try to claim
	 * the lock.  So it is safe to test fl_blocker locklessly.
	 * Also if fl_blocker is NULL, this waiter is not listed on
	 * fl_blocked_requests for some lock, so no other request can
	 * be added to the list of fl_blocked_requests for this
	 * request.  So if fl_blocker is NULL, it is safe to
	 * locklessly check if fl_blocked_requests is empty.  If both
	 * of these checks succeed, there is no need to take the lock.
	 */
	if (waiter->fl_blocker == NULL &&
	    list_empty(&waiter->fl_blocked_requests))
772
		return status;
773
	spin_lock(&blocked_lock_lock);
774 775
	if (waiter->fl_blocker)
		status = 0;
776
	__locks_wake_up_blocks(waiter);
Linus Torvalds's avatar
Linus Torvalds committed
777
	__locks_delete_block(waiter);
778
	spin_unlock(&blocked_lock_lock);
779
	return status;
Linus Torvalds's avatar
Linus Torvalds committed
780
}
781
EXPORT_SYMBOL(locks_delete_block);
Linus Torvalds's avatar
Linus Torvalds committed
782 783 784 785 786

/* 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.
787
 *
788
 * Must be called with both the flc_lock and blocked_lock_lock held. The
789 790 791 792
 * fl_blocked_requests 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_blocked_requests list is empty.
793 794 795 796
 *
 * Rather than just adding to the list, we check for conflicts with any existing
 * waiters, and add beneath any waiter that blocks the new waiter.
 * Thus wakeups don't happen until needed.
Linus Torvalds's avatar
Linus Torvalds committed
797
 */
798
static void __locks_insert_block(struct file_lock *blocker,
799 800 801
				 struct file_lock *waiter,
				 bool conflict(struct file_lock *,
					       struct file_lock *))
Linus Torvalds's avatar
Linus Torvalds committed
802
{
803
	struct file_lock *fl;
804
	BUG_ON(!list_empty(&waiter->fl_blocked_member));
805 806 807 808 809 810 811

new_blocker:
	list_for_each_entry(fl, &blocker->fl_blocked_requests, fl_blocked_member)
		if (conflict(fl, waiter)) {
			blocker =  fl;
			goto new_blocker;
		}
812 813
	waiter->fl_blocker = blocker;
	list_add_tail(&waiter->fl_blocked_member, &blocker->fl_blocked_requests);
814
	if (IS_POSIX(blocker) && !IS_OFDLCK(blocker))
815
		locks_insert_global_blocked(waiter);
816 817 818 819 820 821

	/* The requests in waiter->fl_blocked are known to conflict with
	 * waiter, but might not conflict with blocker, or the requests
	 * and lock which block it.  So they all need to be woken.
	 */
	__locks_wake_up_blocks(waiter);
822 823
}

824
/* Must be called with flc_lock held. */
825
static void locks_insert_block(struct file_lock *blocker,
826 827 828
			       struct file_lock *waiter,
			       bool conflict(struct file_lock *,
					     struct file_lock *))
829
{
830
	spin_lock(&blocked_lock_lock);
831
	__locks_insert_block(blocker, waiter, conflict);
832
	spin_unlock(&blocked_lock_lock);
Linus Torvalds's avatar
Linus Torvalds committed
833 834
}

835 836 837
/*
 * Wake up processes blocked waiting for blocker.
 *
838
 * Must be called with the inode->flc_lock held!
Linus Torvalds's avatar
Linus Torvalds committed
839 840 841
 */
static void locks_wake_up_blocks(struct file_lock *blocker)
{
842 843
	/*
	 * Avoid taking global lock if list is empty. This is safe since new
844
	 * blocked requests are only added to the list under the flc_lock, and
845 846 847
	 * the flc_lock is always held here. Note that removal from the
	 * fl_blocked_requests list does not require the flc_lock, so we must
	 * recheck list_empty() after acquiring the blocked_lock_lock.
848
	 */
849
	if (list_empty(&blocker->fl_blocked_requests))
850 851
		return;

852
	spin_lock(&blocked_lock_lock);
853
	__locks_wake_up_blocks(blocker);
854
	spin_unlock(&blocked_lock_lock);
Linus Torvalds's avatar
Linus Torvalds committed
855 856
}

857
static void
858
locks_insert_lock_ctx(struct file_lock *fl, struct list_head *before)
859 860 861 862 863
{
	list_add_tail(&fl->fl_list, before);
	locks_insert_global_locks(fl);
}

864
static void
865
locks_unlink_lock_ctx(struct file_lock *fl)
Linus Torvalds's avatar
Linus Torvalds committed
866
{
867
	locks_delete_global_locks(fl);
868
	list_del_init(&fl->fl_list);
Linus Torvalds's avatar
Linus Torvalds committed
869
	locks_wake_up_blocks(fl);
870 871
}

872
static void
873
locks_delete_lock_ctx(struct file_lock *fl, struct list_head *dispose)
874
{
875
	locks_unlink_lock_ctx(fl);
876
	if (dispose)
877
		list_add(&fl->fl_list, dispose);
878 879
	else
		locks_free_lock(fl);
Linus Torvalds's avatar
Linus Torvalds committed
880 881 882 883 884
}

/* Determine if lock sys_fl blocks lock caller_fl. Common functionality
 * checks for shared/exclusive status of overlapping locks.
 */
885 886
static bool locks_conflict(struct file_lock *caller_fl,
			   struct file_lock *sys_fl)
Linus Torvalds's avatar
Linus Torvalds committed
887 888
{
	if (sys_fl->fl_type == F_WRLCK)
889
		return true;
Linus Torvalds's avatar
Linus Torvalds committed
890
	if (caller_fl->fl_type == F_WRLCK)
891 892
		return true;
	return false;
Linus Torvalds's avatar
Linus Torvalds committed
893 894 895 896 897
}

/* Determine if lock sys_fl blocks lock caller_fl. POSIX specific
 * checking before calling the locks_conflict().
 */
898 899
static bool posix_locks_conflict(struct file_lock *caller_fl,
				 struct file_lock *sys_fl)
Linus Torvalds's avatar
Linus Torvalds committed
900 901 902 903
{
	/* POSIX locks owned by the same process do not conflict with
	 * each other.
	 */
904
	if (posix_same_owner(caller_fl, sys_fl))
905
		return false;
Linus Torvalds's avatar
Linus Torvalds committed
906 907 908

	/* Check whether they overlap */
	if (!locks_overlap(caller_fl, sys_fl))
909
		return false;
Linus Torvalds's avatar
Linus Torvalds committed
910

911
	return locks_conflict(caller_fl, sys_fl);
Linus Torvalds's avatar
Linus Torvalds committed
912 913 914 915 916
}

/* Determine if lock sys_fl blocks lock caller_fl. FLOCK specific
 * checking before calling the locks_conflict().
 */
917 918
static bool flock_locks_conflict(struct file_lock *caller_fl,
				 struct file_lock *sys_fl)
Linus Torvalds's avatar
Linus Torvalds committed
919 920 921 922
{
	/* FLOCK locks referring to the same filp do not conflict with
	 * each other.
	 */
923
	if (caller_fl->fl_file == sys_fl->fl_file)
924
		return false;
Linus Torvalds's avatar
Linus Torvalds committed
925
	if ((caller_fl->fl_type & LOCK_MAND) || (sys_fl->fl_type & LOCK_MAND))
926
		return false;
Linus Torvalds's avatar
Linus Torvalds committed
927

928
	return locks_conflict(caller_fl, sys_fl);
Linus Torvalds's avatar
Linus Torvalds committed
929 930
}

931
void
932
posix_test_lock(struct file *filp, struct file_lock *fl)
Linus Torvalds's avatar
Linus Torvalds committed
933 934
{
	struct file_lock *cfl;
935
	struct file_lock_context *ctx;
936
	struct inode *inode = locks_inode(filp);
Linus Torvalds's avatar
Linus Torvalds committed
937

938
	ctx = smp_load_acquire(&inode->i_flctx);
939 940 941 942 943
	if (!ctx || list_empty_careful(&ctx->flc_posix)) {
		fl->fl_type = F_UNLCK;
		return;
	}

944
	spin_lock(&ctx->flc_lock);
945 946 947 948 949
	list_for_each_entry(cfl, &ctx->flc_posix, fl_list) {
		if (posix_locks_conflict(fl, cfl)) {
			locks_copy_conflock(fl, cfl);
			goto out;
		}
Linus Torvalds's avatar
Linus Torvalds committed
950
	}
951 952
	fl->fl_type = F_UNLCK;
out:
953
	spin_unlock(&ctx->flc_lock);
954
	return;
Linus Torvalds's avatar
Linus Torvalds committed
955 956 957
}
EXPORT_SYMBOL(posix_test_lock);

958 959 960 961 962
/*
 * Deadlock detection:
 *
 * We attempt to detect deadlocks that are due purely to posix file
 * locks.
Linus Torvalds's avatar
Linus Torvalds committed
963
 *
964 965 966 967 968 969 970
 * 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.
Linus Torvalds's avatar
Linus Torvalds committed
971
 *
972 973 974
 * 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.
975
 *
976 977 978 979
 * 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.
980
 *
981
 * For FL_OFDLCK locks, the owner is the filp, not the files_struct.
982 983 984 985
 * 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.
 *
986
 * In principle, we could do a more limited deadlock detection on FL_OFDLCK
987 988
 * locks that just checks for the case where two tasks are attempting to
 * upgrade from read to write locks on the same inode.
Linus Torvalds's avatar
Linus Torvalds committed
989
 */
990 991 992

#define MAX_DEADLK_ITERATIONS 10

993 994 995 996 997
/* 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;

998
	hash_for_each_possible(blocked_hash, fl, fl_link, posix_owner_key(block_fl)) {
999 1000 1001 1002 1003
		if (posix_same_owner(fl, block_fl)) {
			while (fl->fl_blocker)
				fl = fl->fl_blocker;
			return fl;
		}
1004 1005 1006 1007
	}
	return NULL;
}

1008
/* Must be called with the blocked_lock_lock held! */
1009
static int posix_locks_deadlock(struct file_lock *caller_fl,
Linus Torvalds's avatar
Linus Torvalds committed
1010 1011
				struct file_lock *block_fl)
{
1012
	int i = 0;
Linus Torvalds's avatar
Linus Torvalds committed
1013

1014 1015
	lockdep_assert_held(&blocked_lock_lock);

1016 1017
	/*
	 * This deadlock detector can't reasonably detect deadlocks with
1018
	 * FL_OFDLCK locks, since they aren't owned by a process, per-se.
1019
	 */
1020
	if (IS_OFDLCK(caller_fl))
1021 1022
		return 0;

1023 1024 1025 1026 1027
	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;
Linus Torvalds's avatar
Linus Torvalds committed
1028 1029 1030 1031 1032
	}
	return 0;
}

/* Try to create a FLOCK lock on filp. We always insert new FLOCK locks
1033
 * after any leases, but before any posix locks.
1034 1035 1036 1037
 *
 * 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.
Linus Torvalds's avatar
Linus Torvalds committed
1038
 */
1039
static int flock_lock_inode(struct inode *inode, struct file_lock *request)
Linus Torvalds's avatar
Linus Torvalds committed
1040
{
1041
	struct file_lock *new_fl = NULL;
1042 1043
	struct file_lock *fl;
	struct file_lock_context *ctx;
Linus Torvalds's avatar
Linus Torvalds committed
1044
	int error = 0;
1045
	bool found = false;
1046
	LIST_HEAD(dispose);
Linus Torvalds's avatar
Linus Torvalds committed
1047

1048 1049 1050 1051 1052 1053
	ctx = locks_get_lock_context(inode, request->fl_type);
	if (!ctx) {
		if (request->fl_type != F_UNLCK)
			return -ENOMEM;
		return (request->fl_flags & FL_EXISTS) ? -ENOENT : 0;
	}
1054

1055
	if (!(request->fl_flags & FL_ACCESS) && (request->fl_type != F_UNLCK)) {
1056
		new_fl = locks_alloc_lock();
1057 1058
		if (!new_fl)
			return -ENOMEM;
1059 1060
	}

1061
	percpu_down_read_preempt_disable(&file_rwsem);
1062
	spin_lock(&ctx->flc_lock);
1063 1064 1065
	if (request->fl_flags & FL_ACCESS)
		goto find_conflict;

1066
	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
1067
		if (request->fl_file != fl->fl_file)
Linus Torvalds's avatar
Linus Torvalds committed
1068
			continue;
1069
		if (request->fl_type == fl->fl_type)
Linus Torvalds's avatar
Linus Torvalds committed
1070
			goto out;
1071
		found = true;
1072
		locks_delete_lock_ctx(fl, &dispose);
Linus Torvalds's avatar
Linus Torvalds committed
1073 1074 1075
		break;
	}

1076 1077 1078
	if (request->fl_type == F_UNLCK) {
		if ((request->fl_flags & FL_EXISTS) && !found)
			error = -ENOENT;
1079
		goto out;
1080
	}
Linus Torvalds's avatar
Linus Torvalds committed
1081

1082
find_conflict:
1083
	list_for_each_entry(fl, &ctx->flc_flock, fl_list) {
1084
		if (!flock_locks_conflict(request, fl))
Linus Torvalds's avatar
Linus Torvalds committed
1085 1086
			continue;
		error = -EAGAIN;
1087 1088 1089
		if (!(request->fl_flags & FL_SLEEP))
			goto out;
		error = FILE_LOCK_DEFERRED;
1090
		locks_insert_block(fl, request, flock_locks_conflict);
Linus Torvalds's avatar
Linus Torvalds committed
1091 1092
		goto out;
	}
1093 1094
	if (request->fl_flags & FL_ACCESS)
		goto out;
1095
	locks_copy_lock(new_fl, request);
1096
	locks_move_blocks(new_fl, request);
1097
	locks_insert_lock_ctx(new_fl, &ctx->flc_flock);
1098
	new_fl = NULL;
1099
	error = 0;
Linus Torvalds's avatar
Linus Torvalds committed
1100 1101

out:
1102
	spin_unlock(&ctx->flc_lock);
1103
	percpu_up_read_preempt_enable(&file_rwsem);
1104 1105
	if (new_fl)
		locks_free_lock(new_fl);
1106
	locks_dispose_list(&dispose);
1107
	trace_flock_lock_inode(inode, request, error);
Linus Torvalds's avatar
Linus Torvalds committed
1108 1109 1110
	return error;
}

1111 1112
static int posix_lock_inode(struct inode *inode, struct file_lock *request,
			    struct file_lock *conflock)
Linus Torvalds's avatar
Linus Torvalds committed
1113
{
1114
	struct file_lock *fl, *tmp;
1115 1116
	struct file_lock *new_fl = NULL;
	struct file_lock *new_fl2 = NULL;
Linus Torvalds's avatar
Linus Torvalds committed
1117 1118
	struct file_lock *left = NULL;
	struct file_lock *right = NULL;
1119
	struct file_lock_context *ctx;
1120 1121
	int error;
	bool added = false;
1122
	LIST_HEAD(dispose);
Linus Torvalds's avatar
Linus Torvalds committed
1123

1124
	ctx = locks_get_lock_context(inode, request->fl_type);
1125
	if (!ctx)
1126
		return (request->fl_type == F_UNLCK) ? 0 : -ENOMEM;
1127

Linus Torvalds's avatar
Linus Torvalds committed
1128 1129 1130
	/*
	 * We may need two file_lock structures for this operation,
	 * so we get them in advance to avoid races.
1131 1132
	 *
	 * In some cases we can be sure, that no new locks will be needed
Linus Torvalds's avatar
Linus Torvalds committed
1133
	 */
1134 1135 1136 1137 1138 1139
	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();
	}
Linus Torvalds's avatar
Linus Torvalds committed
1140

1141
	percpu_down_read_preempt_disable(&file_rwsem);
1142
	spin_lock(&ctx->flc_lock);
1143 1144 1145
	/*
	 * New lock request. Walk all POSIX locks and look for conflicts. If
	 * there are any, either return error or put the request on the
1146
	 * blocker's list of waiters and the global blocked_hash.
1147
	 */
Linus Torvalds's avatar
Linus Torvalds committed
1148
	if (request->fl_type != F_UNLCK) {
1149
		list_for_each_entry(fl, &ctx->flc_posix, fl_list) {
Linus Torvalds's avatar
Linus Torvalds committed
1150 1151
			if (!posix_locks_conflict(request, fl))
				continue;
1152
			if (conflock)
1153
				locks_copy_conflock(conflock, fl);
Linus Torvalds's avatar
Linus Torvalds committed
1154 1155 1156
			error = -EAGAIN;
			if (!(request->fl_flags & FL_SLEEP))
				goto out;
1157 1158 1159 1160
			/*
			 * Deadlock detection and insertion into the blocked
			 * locks list must be done while holding the same lock!
			 */
Linus Torvalds's avatar