Home | History | Annotate | Line # | Download | only in kern
vfs_lockf.c revision 1.81
      1 /*	$NetBSD: vfs_lockf.c,v 1.81 2023/09/23 18:21:11 ad Exp $	*/
      2 
      3 /*
      4  * Copyright (c) 1982, 1986, 1989, 1993
      5  *	The Regents of the University of California.  All rights reserved.
      6  *
      7  * This code is derived from software contributed to Berkeley by
      8  * Scooter Morris at Genentech Inc.
      9  *
     10  * Redistribution and use in source and binary forms, with or without
     11  * modification, are permitted provided that the following conditions
     12  * are met:
     13  * 1. Redistributions of source code must retain the above copyright
     14  *    notice, this list of conditions and the following disclaimer.
     15  * 2. Redistributions in binary form must reproduce the above copyright
     16  *    notice, this list of conditions and the following disclaimer in the
     17  *    documentation and/or other materials provided with the distribution.
     18  * 3. Neither the name of the University nor the names of its contributors
     19  *    may be used to endorse or promote products derived from this software
     20  *    without specific prior written permission.
     21  *
     22  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     23  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     24  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     25  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     26  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     27  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     28  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     29  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     30  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     31  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     32  * SUCH DAMAGE.
     33  *
     34  *	@(#)ufs_lockf.c	8.4 (Berkeley) 10/26/94
     35  */
     36 
     37 #include <sys/cdefs.h>
     38 __KERNEL_RCSID(0, "$NetBSD: vfs_lockf.c,v 1.81 2023/09/23 18:21:11 ad Exp $");
     39 
     40 #include <sys/param.h>
     41 #include <sys/systm.h>
     42 #include <sys/kernel.h>
     43 #include <sys/file.h>
     44 #include <sys/proc.h>
     45 #include <sys/vnode.h>
     46 #include <sys/kmem.h>
     47 #include <sys/fcntl.h>
     48 #include <sys/lockf.h>
     49 #include <sys/atomic.h>
     50 #include <sys/kauth.h>
     51 #include <sys/uidinfo.h>
     52 
     53 /*
     54  * The lockf structure is a kernel structure which contains the information
     55  * associated with a byte range lock.  The lockf structures are linked into
     56  * the vnode structure.  Locks are sorted by the starting byte of the lock for
     57  * efficiency.
     58  *
     59  * lf_next is used for two purposes, depending on whether the lock is
     60  * being held, or is in conflict with an existing lock.  If this lock
     61  * is held, it indicates the next lock on the same vnode.
     62  * For pending locks, if lock->lf_next is non-NULL, then lock->lf_block
     63  * must be queued on the lf_blkhd TAILQ of lock->lf_next.
     64  */
     65 
     66 TAILQ_HEAD(locklist, lockf);
     67 
     68 struct lockf {
     69 	kcondvar_t lf_cv;	 /* Signalling */
     70 	short	lf_flags;	 /* Lock semantics: F_POSIX, F_FLOCK, F_WAIT */
     71 	short	lf_type;	 /* Lock type: F_RDLCK, F_WRLCK */
     72 	off_t	lf_start;	 /* The byte # of the start of the lock */
     73 	off_t	lf_end;		 /* The byte # of the end of the lock (-1=EOF)*/
     74 	void	*lf_id;		 /* process or file description holding lock */
     75 	struct	lockf **lf_head; /* Back pointer to the head of lockf list */
     76 	struct	lockf *lf_next;	 /* Next lock on this vnode, or blocking lock */
     77 	struct  locklist lf_blkhd; /* List of requests blocked on this lock */
     78 	TAILQ_ENTRY(lockf) lf_block;/* A request waiting for a lock */
     79 	struct	uidinfo *lf_uip; /* Cached pointer to uidinfo */
     80 };
     81 
     82 /* Maximum length of sleep chains to traverse to try and detect deadlock. */
     83 #define MAXDEPTH 50
     84 
     85 static kmutex_t lockf_lock __cacheline_aligned;
     86 static char lockstr[] = "lockf";
     87 
     88 /*
     89  * This variable controls the maximum number of processes that will
     90  * be checked in doing deadlock detection.
     91  */
     92 int maxlockdepth = MAXDEPTH;
     93 
     94 #ifdef LOCKF_DEBUG
     95 int	lockf_debug = 0;
     96 #endif
     97 
     98 #define SELF	0x1
     99 #define OTHERS	0x2
    100 
    101 /*
    102  * XXX TODO
    103  * Misc cleanups: "void *id" should be visible in the API as a
    104  * "struct proc *".
    105  * (This requires rototilling all VFS's which support advisory locking).
    106  */
    107 
    108 /*
    109  * If there's a lot of lock contention on a single vnode, locking
    110  * schemes which allow for more paralleism would be needed.  Given how
    111  * infrequently byte-range locks are actually used in typical BSD
    112  * code, a more complex approach probably isn't worth it.
    113  */
    114 
    115 /*
    116  * We enforce a limit on locks by uid, so that a single user cannot
    117  * run the kernel out of memory.  For now, the limit is pretty coarse.
    118  * There is no limit on root.
    119  *
    120  * Splitting a lock will always succeed, regardless of current allocations.
    121  * If you're slightly above the limit, we still have to permit an allocation
    122  * so that the unlock can succeed.  If the unlocking causes too many splits,
    123  * however, you're totally cutoff.
    124  */
    125 #define MAXLOCKSPERUID (2 * maxfiles)
    126 
    127 #ifdef LOCKF_DEBUG
    128 /*
    129  * Print out a lock.
    130  */
    131 static void
    132 lf_print(const char *tag, struct lockf *lock)
    133 {
    134 
    135 	printf("%s: lock %p for ", tag, lock);
    136 	if (lock->lf_flags & F_POSIX)
    137 		printf("proc %d", ((struct proc *)lock->lf_id)->p_pid);
    138 	else
    139 		printf("file %p", (struct file *)lock->lf_id);
    140 	printf(" %s, start %jd, end %jd",
    141 		lock->lf_type == F_RDLCK ? "shared" :
    142 		lock->lf_type == F_WRLCK ? "exclusive" :
    143 		lock->lf_type == F_UNLCK ? "unlock" :
    144 		"unknown", (intmax_t)lock->lf_start, (intmax_t)lock->lf_end);
    145 	if (TAILQ_FIRST(&lock->lf_blkhd))
    146 		printf(" block %p\n", TAILQ_FIRST(&lock->lf_blkhd));
    147 	else
    148 		printf("\n");
    149 }
    150 
    151 static void
    152 lf_printlist(const char *tag, struct lockf *lock)
    153 {
    154 	struct lockf *lf, *blk;
    155 
    156 	printf("%s: Lock list:\n", tag);
    157 	for (lf = *lock->lf_head; lf; lf = lf->lf_next) {
    158 		printf("\tlock %p for ", lf);
    159 		if (lf->lf_flags & F_POSIX)
    160 			printf("proc %d", ((struct proc *)lf->lf_id)->p_pid);
    161 		else
    162 			printf("file %p", (struct file *)lf->lf_id);
    163 		printf(", %s, start %jd, end %jd",
    164 			lf->lf_type == F_RDLCK ? "shared" :
    165 			lf->lf_type == F_WRLCK ? "exclusive" :
    166 			lf->lf_type == F_UNLCK ? "unlock" :
    167 			"unknown", (intmax_t)lf->lf_start, (intmax_t)lf->lf_end);
    168 		TAILQ_FOREACH(blk, &lf->lf_blkhd, lf_block) {
    169 			if (blk->lf_flags & F_POSIX)
    170 				printf("; proc %d",
    171 				    ((struct proc *)blk->lf_id)->p_pid);
    172 			else
    173 				printf("; file %p", (struct file *)blk->lf_id);
    174 			printf(", %s, start %jd, end %jd",
    175 				blk->lf_type == F_RDLCK ? "shared" :
    176 				blk->lf_type == F_WRLCK ? "exclusive" :
    177 				blk->lf_type == F_UNLCK ? "unlock" :
    178 				"unknown", (intmax_t)blk->lf_start, (intmax_t)blk->lf_end);
    179 			if (TAILQ_FIRST(&blk->lf_blkhd))
    180 				 panic("lf_printlist: bad list");
    181 		}
    182 		printf("\n");
    183 	}
    184 }
    185 #endif /* LOCKF_DEBUG */
    186 
    187 /*
    188  * 3 options for allowfail.
    189  * 0 - always allocate.  1 - cutoff at limit.  2 - cutoff at double limit.
    190  */
    191 static struct lockf *
    192 lf_alloc(int allowfail)
    193 {
    194 	struct uidinfo *uip;
    195 	struct lockf *lock;
    196 	u_long lcnt;
    197 	const uid_t uid = kauth_cred_geteuid(kauth_cred_get());
    198 
    199 	uip = uid_find(uid);
    200 	lcnt = atomic_inc_ulong_nv(&uip->ui_lockcnt);
    201 	if (uid && allowfail && lcnt >
    202 	    (allowfail == 1 ? MAXLOCKSPERUID : (MAXLOCKSPERUID * 2))) {
    203 		atomic_dec_ulong(&uip->ui_lockcnt);
    204 		return NULL;
    205 	}
    206 
    207 	lock = kmem_alloc(sizeof(*lock), KM_SLEEP);
    208 	lock->lf_uip = uip;
    209 	cv_init(&lock->lf_cv, lockstr);
    210 	return lock;
    211 }
    212 
    213 static void
    214 lf_free(struct lockf *lock)
    215 {
    216 
    217 	atomic_dec_ulong(&lock->lf_uip->ui_lockcnt);
    218 	cv_destroy(&lock->lf_cv);
    219 	kmem_free(lock, sizeof(*lock));
    220 }
    221 
    222 /*
    223  * Walk the list of locks for an inode to
    224  * find an overlapping lock (if any).
    225  *
    226  * NOTE: this returns only the FIRST overlapping lock.  There
    227  *	 may be more than one.
    228  */
    229 static int
    230 lf_findoverlap(struct lockf *lf, struct lockf *lock, int type,
    231     struct lockf ***prev, struct lockf **overlap)
    232 {
    233 	off_t start, end;
    234 
    235 	*overlap = lf;
    236 	if (lf == NULL)
    237 		return 0;
    238 #ifdef LOCKF_DEBUG
    239 	if (lockf_debug & 2)
    240 		lf_print("lf_findoverlap: looking for overlap in", lock);
    241 #endif /* LOCKF_DEBUG */
    242 	start = lock->lf_start;
    243 	end = lock->lf_end;
    244 	while (lf != NULL) {
    245 		if (((type == SELF) && lf->lf_id != lock->lf_id) ||
    246 		    ((type == OTHERS) && lf->lf_id == lock->lf_id)) {
    247 			*prev = &lf->lf_next;
    248 			*overlap = lf = lf->lf_next;
    249 			continue;
    250 		}
    251 #ifdef LOCKF_DEBUG
    252 		if (lockf_debug & 2)
    253 			lf_print("\tchecking", lf);
    254 #endif /* LOCKF_DEBUG */
    255 		/*
    256 		 * OK, check for overlap
    257 		 *
    258 		 * Six cases:
    259 		 *	0) no overlap
    260 		 *	1) overlap == lock
    261 		 *	2) overlap contains lock
    262 		 *	3) lock contains overlap
    263 		 *	4) overlap starts before lock
    264 		 *	5) overlap ends after lock
    265 		 */
    266 		if ((lf->lf_end != -1 && start > lf->lf_end) ||
    267 		    (end != -1 && lf->lf_start > end)) {
    268 			/* Case 0 */
    269 #ifdef LOCKF_DEBUG
    270 			if (lockf_debug & 2)
    271 				printf("no overlap\n");
    272 #endif /* LOCKF_DEBUG */
    273 			if ((type & SELF) && end != -1 && lf->lf_start > end)
    274 				return 0;
    275 			*prev = &lf->lf_next;
    276 			*overlap = lf = lf->lf_next;
    277 			continue;
    278 		}
    279 		if ((lf->lf_start == start) && (lf->lf_end == end)) {
    280 			/* Case 1 */
    281 #ifdef LOCKF_DEBUG
    282 			if (lockf_debug & 2)
    283 				printf("overlap == lock\n");
    284 #endif /* LOCKF_DEBUG */
    285 			return 1;
    286 		}
    287 		if ((lf->lf_start <= start) &&
    288 		    (end != -1) &&
    289 		    ((lf->lf_end >= end) || (lf->lf_end == -1))) {
    290 			/* Case 2 */
    291 #ifdef LOCKF_DEBUG
    292 			if (lockf_debug & 2)
    293 				printf("overlap contains lock\n");
    294 #endif /* LOCKF_DEBUG */
    295 			return 2;
    296 		}
    297 		if (start <= lf->lf_start &&
    298 		           (end == -1 ||
    299 			   (lf->lf_end != -1 && end >= lf->lf_end))) {
    300 			/* Case 3 */
    301 #ifdef LOCKF_DEBUG
    302 			if (lockf_debug & 2)
    303 				printf("lock contains overlap\n");
    304 #endif /* LOCKF_DEBUG */
    305 			return 3;
    306 		}
    307 		if ((lf->lf_start < start) &&
    308 			((lf->lf_end >= start) || (lf->lf_end == -1))) {
    309 			/* Case 4 */
    310 #ifdef LOCKF_DEBUG
    311 			if (lockf_debug & 2)
    312 				printf("overlap starts before lock\n");
    313 #endif /* LOCKF_DEBUG */
    314 			return 4;
    315 		}
    316 		if ((lf->lf_start > start) &&
    317 			(end != -1) &&
    318 			((lf->lf_end > end) || (lf->lf_end == -1))) {
    319 			/* Case 5 */
    320 #ifdef LOCKF_DEBUG
    321 			if (lockf_debug & 2)
    322 				printf("overlap ends after lock\n");
    323 #endif /* LOCKF_DEBUG */
    324 			return 5;
    325 		}
    326 		panic("lf_findoverlap: default");
    327 	}
    328 	return 0;
    329 }
    330 
    331 /*
    332  * Split a lock and a contained region into
    333  * two or three locks as necessary.
    334  */
    335 static void
    336 lf_split(struct lockf *lock1, struct lockf *lock2, struct lockf **sparelock)
    337 {
    338 	struct lockf *splitlock;
    339 
    340 #ifdef LOCKF_DEBUG
    341 	if (lockf_debug & 2) {
    342 		lf_print("lf_split", lock1);
    343 		lf_print("splitting from", lock2);
    344 	}
    345 #endif /* LOCKF_DEBUG */
    346 	/*
    347 	 * Check to see if splitting into only two pieces.
    348 	 */
    349 	if (lock1->lf_start == lock2->lf_start) {
    350 		lock1->lf_start = lock2->lf_end + 1;
    351 		lock2->lf_next = lock1;
    352 		return;
    353 	}
    354 	if (lock1->lf_end == lock2->lf_end) {
    355 		lock1->lf_end = lock2->lf_start - 1;
    356 		lock2->lf_next = lock1->lf_next;
    357 		lock1->lf_next = lock2;
    358 		return;
    359 	}
    360 	/*
    361 	 * Make a new lock consisting of the last part of
    362 	 * the encompassing lock
    363 	 */
    364 	splitlock = *sparelock;
    365 	*sparelock = NULL;
    366 	cv_destroy(&splitlock->lf_cv);
    367 	memcpy(splitlock, lock1, sizeof(*splitlock));
    368 	cv_init(&splitlock->lf_cv, lockstr);
    369 
    370 	splitlock->lf_start = lock2->lf_end + 1;
    371 	TAILQ_INIT(&splitlock->lf_blkhd);
    372 	lock1->lf_end = lock2->lf_start - 1;
    373 	/*
    374 	 * OK, now link it in
    375 	 */
    376 	splitlock->lf_next = lock1->lf_next;
    377 	lock2->lf_next = splitlock;
    378 	lock1->lf_next = lock2;
    379 }
    380 
    381 /*
    382  * Wakeup a blocklist
    383  */
    384 static void
    385 lf_wakelock(struct lockf *listhead)
    386 {
    387 	struct lockf *wakelock;
    388 
    389 	while ((wakelock = TAILQ_FIRST(&listhead->lf_blkhd))) {
    390 		KASSERT(wakelock->lf_next == listhead);
    391 		TAILQ_REMOVE(&listhead->lf_blkhd, wakelock, lf_block);
    392 		wakelock->lf_next = NULL;
    393 #ifdef LOCKF_DEBUG
    394 		if (lockf_debug & 2)
    395 			lf_print("lf_wakelock: awakening", wakelock);
    396 #endif
    397 		cv_broadcast(&wakelock->lf_cv);
    398 	}
    399 }
    400 
    401 /*
    402  * Remove a byte-range lock on an inode.
    403  *
    404  * Generally, find the lock (or an overlap to that lock)
    405  * and remove it (or shrink it), then wakeup anyone we can.
    406  */
    407 static int
    408 lf_clearlock(struct lockf *unlock, struct lockf **sparelock)
    409 {
    410 	struct lockf **head = unlock->lf_head;
    411 	struct lockf *lf = *head;
    412 	struct lockf *overlap, **prev;
    413 	int ovcase;
    414 
    415 	if (lf == NULL)
    416 		return 0;
    417 #ifdef LOCKF_DEBUG
    418 	if (unlock->lf_type != F_UNLCK)
    419 		panic("lf_clearlock: bad type");
    420 	if (lockf_debug & 1)
    421 		lf_print("lf_clearlock", unlock);
    422 #endif /* LOCKF_DEBUG */
    423 	prev = head;
    424 	while ((ovcase = lf_findoverlap(lf, unlock, SELF,
    425 	    &prev, &overlap)) != 0) {
    426 		/*
    427 		 * Wakeup the list of locks to be retried.
    428 		 */
    429 		lf_wakelock(overlap);
    430 
    431 		switch (ovcase) {
    432 
    433 		case 1: /* overlap == lock */
    434 			*prev = overlap->lf_next;
    435 			lf_free(overlap);
    436 			break;
    437 
    438 		case 2: /* overlap contains lock: split it */
    439 			if (overlap->lf_start == unlock->lf_start) {
    440 				overlap->lf_start = unlock->lf_end + 1;
    441 				break;
    442 			}
    443 			lf_split(overlap, unlock, sparelock);
    444 			overlap->lf_next = unlock->lf_next;
    445 			break;
    446 
    447 		case 3: /* lock contains overlap */
    448 			*prev = overlap->lf_next;
    449 			lf = overlap->lf_next;
    450 			lf_free(overlap);
    451 			continue;
    452 
    453 		case 4: /* overlap starts before lock */
    454 			overlap->lf_end = unlock->lf_start - 1;
    455 			prev = &overlap->lf_next;
    456 			lf = overlap->lf_next;
    457 			continue;
    458 
    459 		case 5: /* overlap ends after lock */
    460 			overlap->lf_start = unlock->lf_end + 1;
    461 			break;
    462 		}
    463 		break;
    464 	}
    465 #ifdef LOCKF_DEBUG
    466 	if (lockf_debug & 1)
    467 		lf_printlist("lf_clearlock", unlock);
    468 #endif /* LOCKF_DEBUG */
    469 	return 0;
    470 }
    471 
    472 /*
    473  * Walk the list of locks for an inode and
    474  * return the first blocking lock.
    475  */
    476 static struct lockf *
    477 lf_getblock(struct lockf *lock)
    478 {
    479 	struct lockf **prev, *overlap, *lf = *(lock->lf_head);
    480 
    481 	prev = lock->lf_head;
    482 	while (lf_findoverlap(lf, lock, OTHERS, &prev, &overlap) != 0) {
    483 		/*
    484 		 * We've found an overlap, see if it blocks us
    485 		 */
    486 		if ((lock->lf_type == F_WRLCK || overlap->lf_type == F_WRLCK))
    487 			return overlap;
    488 		/*
    489 		 * Nope, point to the next one on the list and
    490 		 * see if it blocks us
    491 		 */
    492 		lf = overlap->lf_next;
    493 	}
    494 	return NULL;
    495 }
    496 
    497 /*
    498  * Set a byte-range lock.
    499  */
    500 static int
    501 lf_setlock(struct lockf *lock, struct lockf **sparelock,
    502     kmutex_t *interlock)
    503 {
    504 	struct lockf *block;
    505 	struct lockf **head = lock->lf_head;
    506 	struct lockf **prev, *overlap, *ltmp;
    507 	int ovcase, needtolink, error;
    508 
    509 #ifdef LOCKF_DEBUG
    510 	if (lockf_debug & 1)
    511 		lf_print("lf_setlock", lock);
    512 #endif /* LOCKF_DEBUG */
    513 
    514 	/*
    515 	 * Scan lock list for this file looking for locks that would block us.
    516 	 */
    517 	while ((block = lf_getblock(lock)) != NULL) {
    518 		/*
    519 		 * Free the structure and return if nonblocking.
    520 		 */
    521 		if ((lock->lf_flags & F_WAIT) == 0) {
    522 			lf_free(lock);
    523 			return EAGAIN;
    524 		}
    525 		/*
    526 		 * We are blocked. Since flock style locks cover
    527 		 * the whole file, there is no chance for deadlock.
    528 		 * For byte-range locks we must check for deadlock.
    529 		 *
    530 		 * Deadlock detection is done by looking through the
    531 		 * wait channels to see if there are any cycles that
    532 		 * involve us. MAXDEPTH is set just to make sure we
    533 		 * do not go off into neverneverland.
    534 		 */
    535 		if ((lock->lf_flags & F_POSIX) &&
    536 		    (block->lf_flags & F_POSIX)) {
    537 			struct lwp *wlwp;
    538 			volatile const struct lockf *waitblock;
    539 			int i = 0;
    540 			struct proc *p;
    541 
    542 			p = (struct proc *)block->lf_id;
    543 			KASSERT(p != NULL);
    544 			while (i++ < maxlockdepth) {
    545 				mutex_enter(p->p_lock);
    546 				if (p->p_nlwps > 1) {
    547 					mutex_exit(p->p_lock);
    548 					break;
    549 				}
    550 				wlwp = LIST_FIRST(&p->p_lwps);
    551 				lwp_lock(wlwp);
    552 				if (wlwp->l_wchan == NULL ||
    553 				    wlwp->l_wmesg != lockstr) {
    554 					lwp_unlock(wlwp);
    555 					mutex_exit(p->p_lock);
    556 					break;
    557 				}
    558 				waitblock = wlwp->l_wchan;
    559 				lwp_unlock(wlwp);
    560 				mutex_exit(p->p_lock);
    561 				/* Get the owner of the blocking lock */
    562 				waitblock = waitblock->lf_next;
    563 				if ((waitblock->lf_flags & F_POSIX) == 0)
    564 					break;
    565 				p = (struct proc *)waitblock->lf_id;
    566 				if (p == curproc) {
    567 					lf_free(lock);
    568 					return EDEADLK;
    569 				}
    570 			}
    571 			/*
    572 			 * If we're still following a dependency chain
    573 			 * after maxlockdepth iterations, assume we're in
    574 			 * a cycle to be safe.
    575 			 */
    576 			if (i >= maxlockdepth) {
    577 				lf_free(lock);
    578 				return EDEADLK;
    579 			}
    580 		}
    581 		/*
    582 		 * For flock type locks, we must first remove
    583 		 * any shared locks that we hold before we sleep
    584 		 * waiting for an exclusive lock.
    585 		 */
    586 		if ((lock->lf_flags & F_FLOCK) &&
    587 		    lock->lf_type == F_WRLCK) {
    588 			lock->lf_type = F_UNLCK;
    589 			(void) lf_clearlock(lock, NULL);
    590 			lock->lf_type = F_WRLCK;
    591 		}
    592 		/*
    593 		 * Add our lock to the blocked list and sleep until we're free.
    594 		 * Remember who blocked us (for deadlock detection).
    595 		 */
    596 		lock->lf_next = block;
    597 		TAILQ_INSERT_TAIL(&block->lf_blkhd, lock, lf_block);
    598 #ifdef LOCKF_DEBUG
    599 		if (lockf_debug & 1) {
    600 			lf_print("lf_setlock: blocking on", block);
    601 			lf_printlist("lf_setlock", block);
    602 		}
    603 #endif /* LOCKF_DEBUG */
    604 		error = cv_wait_sig(&lock->lf_cv, interlock);
    605 
    606 		/*
    607 		 * We may have been awoken by a signal (in
    608 		 * which case we must remove ourselves from the
    609 		 * blocked list) and/or by another process
    610 		 * releasing a lock (in which case we have already
    611 		 * been removed from the blocked list and our
    612 		 * lf_next field set to NULL).
    613 		 */
    614 		if (lock->lf_next != NULL) {
    615 			TAILQ_REMOVE(&lock->lf_next->lf_blkhd, lock, lf_block);
    616 			lock->lf_next = NULL;
    617 		}
    618 		if (error) {
    619 			lf_free(lock);
    620 			return error;
    621 		}
    622 	}
    623 	/*
    624 	 * No blocks!!  Add the lock.  Note that we will
    625 	 * downgrade or upgrade any overlapping locks this
    626 	 * process already owns.
    627 	 *
    628 	 * Skip over locks owned by other processes.
    629 	 * Handle any locks that overlap and are owned by ourselves.
    630 	 */
    631 	prev = head;
    632 	block = *head;
    633 	needtolink = 1;
    634 	for (;;) {
    635 		ovcase = lf_findoverlap(block, lock, SELF, &prev, &overlap);
    636 		if (ovcase)
    637 			block = overlap->lf_next;
    638 		/*
    639 		 * Six cases:
    640 		 *	0) no overlap
    641 		 *	1) overlap == lock
    642 		 *	2) overlap contains lock
    643 		 *	3) lock contains overlap
    644 		 *	4) overlap starts before lock
    645 		 *	5) overlap ends after lock
    646 		 */
    647 		switch (ovcase) {
    648 		case 0: /* no overlap */
    649 			if (needtolink) {
    650 				*prev = lock;
    651 				lock->lf_next = overlap;
    652 			}
    653 			break;
    654 
    655 		case 1: /* overlap == lock */
    656 			/*
    657 			 * If downgrading lock, others may be
    658 			 * able to acquire it.
    659 			 */
    660 			if (lock->lf_type == F_RDLCK &&
    661 			    overlap->lf_type == F_WRLCK)
    662 				lf_wakelock(overlap);
    663 			overlap->lf_type = lock->lf_type;
    664 			lf_free(lock);
    665 			lock = overlap; /* for debug output below */
    666 			break;
    667 
    668 		case 2: /* overlap contains lock */
    669 			/*
    670 			 * Check for common starting point and different types.
    671 			 */
    672 			if (overlap->lf_type == lock->lf_type) {
    673 				lf_free(lock);
    674 				lock = overlap; /* for debug output below */
    675 				break;
    676 			}
    677 			if (overlap->lf_start == lock->lf_start) {
    678 				*prev = lock;
    679 				lock->lf_next = overlap;
    680 				overlap->lf_start = lock->lf_end + 1;
    681 			} else
    682 				lf_split(overlap, lock, sparelock);
    683 			lf_wakelock(overlap);
    684 			break;
    685 
    686 		case 3: /* lock contains overlap */
    687 			/*
    688 			 * If downgrading lock, others may be able to
    689 			 * acquire it, otherwise take the list.
    690 			 */
    691 			if (lock->lf_type == F_RDLCK &&
    692 			    overlap->lf_type == F_WRLCK) {
    693 				lf_wakelock(overlap);
    694 			} else {
    695 				while ((ltmp = TAILQ_FIRST(&overlap->lf_blkhd))) {
    696 					KASSERT(ltmp->lf_next == overlap);
    697 					TAILQ_REMOVE(&overlap->lf_blkhd, ltmp,
    698 					    lf_block);
    699 					ltmp->lf_next = lock;
    700 					TAILQ_INSERT_TAIL(&lock->lf_blkhd,
    701 					    ltmp, lf_block);
    702 				}
    703 			}
    704 			/*
    705 			 * Add the new lock if necessary and delete the overlap.
    706 			 */
    707 			if (needtolink) {
    708 				*prev = lock;
    709 				lock->lf_next = overlap->lf_next;
    710 				prev = &lock->lf_next;
    711 				needtolink = 0;
    712 			} else
    713 				*prev = overlap->lf_next;
    714 			lf_free(overlap);
    715 			continue;
    716 
    717 		case 4: /* overlap starts before lock */
    718 			/*
    719 			 * Add lock after overlap on the list.
    720 			 */
    721 			lock->lf_next = overlap->lf_next;
    722 			overlap->lf_next = lock;
    723 			overlap->lf_end = lock->lf_start - 1;
    724 			prev = &lock->lf_next;
    725 			lf_wakelock(overlap);
    726 			needtolink = 0;
    727 			continue;
    728 
    729 		case 5: /* overlap ends after lock */
    730 			/*
    731 			 * Add the new lock before overlap.
    732 			 */
    733 			if (needtolink) {
    734 				*prev = lock;
    735 				lock->lf_next = overlap;
    736 			}
    737 			overlap->lf_start = lock->lf_end + 1;
    738 			lf_wakelock(overlap);
    739 			break;
    740 		}
    741 		break;
    742 	}
    743 #ifdef LOCKF_DEBUG
    744 	if (lockf_debug & 1) {
    745 		lf_print("lf_setlock: got the lock", lock);
    746 		lf_printlist("lf_setlock", lock);
    747 	}
    748 #endif /* LOCKF_DEBUG */
    749 	return 0;
    750 }
    751 
    752 /*
    753  * Check whether there is a blocking lock,
    754  * and if so return its process identifier.
    755  */
    756 static int
    757 lf_getlock(struct lockf *lock, struct flock *fl)
    758 {
    759 	struct lockf *block;
    760 
    761 #ifdef LOCKF_DEBUG
    762 	if (lockf_debug & 1)
    763 		lf_print("lf_getlock", lock);
    764 #endif /* LOCKF_DEBUG */
    765 
    766 	if ((block = lf_getblock(lock)) != NULL) {
    767 		fl->l_type = block->lf_type;
    768 		fl->l_whence = SEEK_SET;
    769 		fl->l_start = block->lf_start;
    770 		if (block->lf_end == -1)
    771 			fl->l_len = 0;
    772 		else
    773 			fl->l_len = block->lf_end - block->lf_start + 1;
    774 		if (block->lf_flags & F_POSIX)
    775 			fl->l_pid = ((struct proc *)block->lf_id)->p_pid;
    776 		else
    777 			fl->l_pid = -1;
    778 	} else {
    779 		fl->l_type = F_UNLCK;
    780 	}
    781 	return 0;
    782 }
    783 
    784 /*
    785  * Do an advisory lock operation.
    786  */
    787 int
    788 lf_advlock(struct vop_advlock_args *ap, struct lockf **head, off_t size)
    789 {
    790 	struct flock *fl = ap->a_fl;
    791 	struct lockf *lock = NULL;
    792 	struct lockf *sparelock;
    793 	kmutex_t *interlock = &lockf_lock;
    794 	off_t start, end;
    795 	int error = 0;
    796 
    797 	KASSERTMSG(size >= 0, "size=%jd", (intmax_t)size);
    798 
    799 	/*
    800 	 * Convert the flock structure into a start and end.
    801 	 */
    802 	switch (fl->l_whence) {
    803 	case SEEK_SET:
    804 	case SEEK_CUR:
    805 		/*
    806 		 * Caller is responsible for adding any necessary offset
    807 		 * when SEEK_CUR is used.
    808 		 */
    809 		start = fl->l_start;
    810 		break;
    811 
    812 	case SEEK_END:
    813 		if (fl->l_start > __type_max(off_t) - size)
    814 			return EINVAL;
    815 		start = size + fl->l_start;
    816 		break;
    817 
    818 	default:
    819 		return EINVAL;
    820 	}
    821 
    822 	if (fl->l_len == 0)
    823 		end = -1;
    824 	else {
    825 		if (fl->l_len >= 0) {
    826 			if (start >= 0 &&
    827 			    fl->l_len - 1 > __type_max(off_t) - start)
    828 				return EINVAL;
    829 			end = start + (fl->l_len - 1);
    830 		} else {
    831 			/* lockf() allows -ve lengths */
    832 			if (start < 0)
    833 				return EINVAL;
    834 			end = start - 1;
    835 			start += fl->l_len;
    836 		}
    837 	}
    838 	if (start < 0)
    839 		return EINVAL;
    840 
    841 	/*
    842 	 * Allocate locks before acquiring the interlock.  We need two
    843 	 * locks in the worst case.
    844 	 */
    845 	switch (ap->a_op) {
    846 	case F_SETLK:
    847 	case F_UNLCK:
    848 		/*
    849 		 * XXX For F_UNLCK case, we can re-use the lock.
    850 		 */
    851 		if ((ap->a_flags & F_FLOCK) == 0) {
    852 			/*
    853 			 * Byte-range lock might need one more lock.
    854 			 */
    855 			sparelock = lf_alloc(0);
    856 			if (sparelock == NULL) {
    857 				error = ENOMEM;
    858 				goto quit;
    859 			}
    860 			break;
    861 		}
    862 		/* FALLTHROUGH */
    863 
    864 	case F_GETLK:
    865 		sparelock = NULL;
    866 		break;
    867 
    868 	default:
    869 		return EINVAL;
    870 	}
    871 
    872 	switch (ap->a_op) {
    873 	case F_SETLK:
    874 		lock = lf_alloc(1);
    875 		break;
    876 	case F_UNLCK:
    877 		if (start == 0 || end == -1) {
    878 			/* never split */
    879 			lock = lf_alloc(0);
    880 		} else {
    881 			/* might split */
    882 			lock = lf_alloc(2);
    883 		}
    884 		break;
    885 	case F_GETLK:
    886 		lock = lf_alloc(0);
    887 		break;
    888 	}
    889 	if (lock == NULL) {
    890 		error = ENOMEM;
    891 		goto quit;
    892 	}
    893 
    894 	mutex_enter(interlock);
    895 
    896 	/*
    897 	 * Avoid the common case of unlocking when inode has no locks.
    898 	 */
    899 	if (*head == (struct lockf *)0) {
    900 		if (ap->a_op != F_SETLK) {
    901 			fl->l_type = F_UNLCK;
    902 			error = 0;
    903 			goto quit_unlock;
    904 		}
    905 	}
    906 
    907 	/*
    908 	 * Create the lockf structure.
    909 	 */
    910 	lock->lf_start = start;
    911 	lock->lf_end = end;
    912 	lock->lf_head = head;
    913 	lock->lf_type = fl->l_type;
    914 	lock->lf_next = (struct lockf *)0;
    915 	TAILQ_INIT(&lock->lf_blkhd);
    916 	lock->lf_flags = ap->a_flags;
    917 	if (lock->lf_flags & F_POSIX) {
    918 		KASSERT(curproc == (struct proc *)ap->a_id);
    919 	}
    920 	lock->lf_id = ap->a_id;
    921 
    922 	/*
    923 	 * Do the requested operation.
    924 	 */
    925 	switch (ap->a_op) {
    926 
    927 	case F_SETLK:
    928 		error = lf_setlock(lock, &sparelock, interlock);
    929 		lock = NULL; /* lf_setlock freed it */
    930 		break;
    931 
    932 	case F_UNLCK:
    933 		error = lf_clearlock(lock, &sparelock);
    934 		break;
    935 
    936 	case F_GETLK:
    937 		error = lf_getlock(lock, fl);
    938 		break;
    939 
    940 	default:
    941 		break;
    942 		/* NOTREACHED */
    943 	}
    944 
    945 quit_unlock:
    946 	mutex_exit(interlock);
    947 quit:
    948 	if (lock)
    949 		lf_free(lock);
    950 	if (sparelock)
    951 		lf_free(sparelock);
    952 
    953 	return error;
    954 }
    955 
    956 /*
    957  * Initialize subsystem.   XXX We use a global lock.  This could be the
    958  * vnode interlock, but the deadlock detection code may need to inspect
    959  * locks belonging to other files.
    960  */
    961 void
    962 lf_init(void)
    963 {
    964 
    965 	mutex_init(&lockf_lock, MUTEX_DEFAULT, IPL_NONE);
    966 }
    967