Home | History | Annotate | Line # | Download | only in kern
vfs_lockf.c revision 1.43
      1 /*	$NetBSD: vfs_lockf.c,v 1.43 2005/05/22 15:54:47 christos 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 1.43 2005/05/22 15:54:47 christos 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/pool.h>
     47 #include <sys/fcntl.h>
     48 #include <sys/lockf.h>
     49 
     50 POOL_INIT(lockfpool, sizeof(struct lockf), 0, 0, 0, "lockfpl",
     51     &pool_allocator_nointr);
     52 
     53 /*
     54  * This variable controls the maximum number of processes that will
     55  * be checked in doing deadlock detection.
     56  */
     57 int maxlockdepth = MAXDEPTH;
     58 
     59 #ifdef LOCKF_DEBUG
     60 int	lockf_debug = 0;
     61 #endif
     62 
     63 #define NOLOCKF (struct lockf *)0
     64 #define SELF	0x1
     65 #define OTHERS	0x2
     66 
     67 static int lf_clearlock(struct lockf *, struct lockf **);
     68 static int lf_findoverlap(struct lockf *,
     69 	    struct lockf *, int, struct lockf ***, struct lockf **);
     70 static struct lockf *lf_getblock(struct lockf *);
     71 static int lf_getlock(struct lockf *, struct flock *);
     72 static int lf_setlock(struct lockf *, struct lockf **, struct simplelock *);
     73 static void lf_split(struct lockf *, struct lockf *, struct lockf **);
     74 static void lf_wakelock(struct lockf *);
     75 static struct lockf *lf_alloc(uid_t, int);
     76 static void lf_free(struct lockf *);
     77 
     78 
     79 #ifdef LOCKF_DEBUG
     80 static void lf_print(char *, struct lockf *);
     81 static void lf_printlist(char *, struct lockf *);
     82 #endif
     83 
     84 /*
     85  * XXX TODO
     86  * Misc cleanups: "caddr_t id" should be visible in the API as a
     87  * "struct proc *".
     88  * (This requires rototilling all VFS's which support advisory locking).
     89  */
     90 
     91 /*
     92  * If there's a lot of lock contention on a single vnode, locking
     93  * schemes which allow for more paralleism would be needed.  Given how
     94  * infrequently byte-range locks are actually used in typical BSD
     95  * code, a more complex approach probably isn't worth it.
     96  */
     97 
     98 /*
     99  * We enforce a limit on locks by uid, so that a single user cannot
    100  * run the kernel out of memory.  For now, the limit is pretty coarse.
    101  * There is no limit on root.
    102  *
    103  * Splitting a lock will always succeed, regardless of current allocations.
    104  * If you're slightly above the limit, we still have to permit an allocation
    105  * so that the unlock can succeed.  If the unlocking causes too many splits,
    106  * however, you're totally cutoff.
    107  */
    108 int maxlocksperuid = 1024;
    109 
    110 /*
    111  * 3 options for allowfail.
    112  * 0 - always allocate.  1 - cutoff at limit.  2 - cutoff at double limit.
    113  */
    114 struct lockf *
    115 lf_alloc(uid_t uid, int allowfail)
    116 {
    117 	struct uidinfo *uip;
    118 	struct lockf *lock;
    119 	int s;
    120 
    121 	uip = uid_find(uid);
    122 	UILOCK(uip, s);
    123 	if (uid && allowfail && uip->ui_lockcnt >
    124 	    (allowfail == 1 ? maxlocksperuid : (maxlocksperuid * 2))) {
    125 		UIUNLOCK(uip, s);
    126 		return NULL;
    127 	}
    128 	uip->ui_lockcnt++;
    129 	UIUNLOCK(uip, s);
    130 	lock = pool_get(&lockfpool, PR_WAITOK);
    131 	lock->lf_uid = uid;
    132 	return lock;
    133 }
    134 
    135 void
    136 lf_free(struct lockf *lock)
    137 {
    138 	struct uidinfo *uip;
    139 	int s;
    140 
    141 	uip = uid_find(lock->lf_uid);
    142 	UILOCK(uip, s);
    143 	uip->ui_lockcnt--;
    144 	UIUNLOCK(uip, s);
    145 	pool_put(&lockfpool, lock);
    146 }
    147 
    148 /*
    149  * Do an advisory lock operation.
    150  */
    151 int
    152 lf_advlock(struct vop_advlock_args *ap, struct lockf **head, off_t size)
    153 {
    154 	struct proc *p = curproc;
    155 	struct flock *fl = ap->a_fl;
    156 	struct lockf *lock = NULL;
    157 	struct lockf *sparelock;
    158 	struct simplelock *interlock = &ap->a_vp->v_interlock;
    159 	off_t start, end;
    160 	int error = 0;
    161 
    162 	/*
    163 	 * Convert the flock structure into a start and end.
    164 	 */
    165 	switch (fl->l_whence) {
    166 	case SEEK_SET:
    167 	case SEEK_CUR:
    168 		/*
    169 		 * Caller is responsible for adding any necessary offset
    170 		 * when SEEK_CUR is used.
    171 		 */
    172 		start = fl->l_start;
    173 		break;
    174 
    175 	case SEEK_END:
    176 		start = size + fl->l_start;
    177 		break;
    178 
    179 	default:
    180 		return EINVAL;
    181 	}
    182 	if (start < 0)
    183 		return EINVAL;
    184 
    185 	/*
    186 	 * allocate locks before acquire simple lock.
    187 	 * we need two locks in the worst case.
    188 	 */
    189 	switch (ap->a_op) {
    190 	case F_SETLK:
    191 	case F_UNLCK:
    192 		/*
    193 		 * XXX for F_UNLCK case, we can re-use lock.
    194 		 */
    195 		if ((fl->l_type & F_FLOCK) == 0) {
    196 			/*
    197 			 * byte-range lock might need one more lock.
    198 			 */
    199 			sparelock = lf_alloc(p->p_ucred->cr_uid, 0);
    200 			if (sparelock == NULL) {
    201 				error = ENOMEM;
    202 				goto quit;
    203 			}
    204 			break;
    205 		}
    206 		/* FALLTHROUGH */
    207 
    208 	case F_GETLK:
    209 		sparelock = NULL;
    210 		break;
    211 
    212 	default:
    213 		return EINVAL;
    214 	}
    215 
    216 	lock = lf_alloc(p->p_ucred->cr_uid, ap->a_op != F_UNLCK ? 1 : 2);
    217 	if (lock == NULL) {
    218 		error = ENOMEM;
    219 		goto quit;
    220 	}
    221 
    222 	simple_lock(interlock);
    223 
    224 	/*
    225 	 * Avoid the common case of unlocking when inode has no locks.
    226 	 */
    227 	if (*head == (struct lockf *)0) {
    228 		if (ap->a_op != F_SETLK) {
    229 			fl->l_type = F_UNLCK;
    230 			error = 0;
    231 			goto quit_unlock;
    232 		}
    233 	}
    234 
    235 	if (fl->l_len == 0)
    236 		end = -1;
    237 	else
    238 		end = start + fl->l_len - 1;
    239 	/*
    240 	 * Create the lockf structure.
    241 	 */
    242 	lock->lf_start = start;
    243 	lock->lf_end = end;
    244 	/* XXX NJWLWP
    245 	 * I don't want to make the entire VFS universe use LWPs, because
    246 	 * they don't need them, for the most part. This is an exception,
    247 	 * and a kluge.
    248 	 */
    249 
    250 	lock->lf_head = head;
    251 	lock->lf_type = fl->l_type;
    252 	lock->lf_next = (struct lockf *)0;
    253 	TAILQ_INIT(&lock->lf_blkhd);
    254 	lock->lf_flags = ap->a_flags;
    255 	if (lock->lf_flags & F_POSIX) {
    256 		KASSERT(curproc == (struct proc *)ap->a_id);
    257 	}
    258 	lock->lf_id = (struct proc *)ap->a_id;
    259 	lock->lf_lwp = curlwp;
    260 
    261 	/*
    262 	 * Do the requested operation.
    263 	 */
    264 	switch (ap->a_op) {
    265 
    266 	case F_SETLK:
    267 		error = lf_setlock(lock, &sparelock, interlock);
    268 		lock = NULL; /* lf_setlock freed it */
    269 		break;
    270 
    271 	case F_UNLCK:
    272 		error = lf_clearlock(lock, &sparelock);
    273 		break;
    274 
    275 	case F_GETLK:
    276 		error = lf_getlock(lock, fl);
    277 		break;
    278 
    279 	default:
    280 		break;
    281 		/* NOTREACHED */
    282 	}
    283 
    284 quit_unlock:
    285 	simple_unlock(interlock);
    286 quit:
    287 	if (lock)
    288 		lf_free(lock);
    289 	if (sparelock)
    290 		lf_free(sparelock);
    291 
    292 	return error;
    293 }
    294 
    295 /*
    296  * Set a byte-range lock.
    297  */
    298 static int
    299 lf_setlock(struct lockf *lock, struct lockf **sparelock,
    300     struct simplelock *interlock)
    301 {
    302 	struct lockf *block;
    303 	struct lockf **head = lock->lf_head;
    304 	struct lockf **prev, *overlap, *ltmp;
    305 	static char lockstr[] = "lockf";
    306 	int ovcase, priority, needtolink, error;
    307 
    308 #ifdef LOCKF_DEBUG
    309 	if (lockf_debug & 1)
    310 		lf_print("lf_setlock", lock);
    311 #endif /* LOCKF_DEBUG */
    312 
    313 	/*
    314 	 * Set the priority
    315 	 */
    316 	priority = PLOCK;
    317 	if (lock->lf_type == F_WRLCK)
    318 		priority += 4;
    319 	priority |= PCATCH;
    320 	/*
    321 	 * Scan lock list for this file looking for locks that would block us.
    322 	 */
    323 	while ((block = lf_getblock(lock)) != NULL) {
    324 		/*
    325 		 * Free the structure and return if nonblocking.
    326 		 */
    327 		if ((lock->lf_flags & F_WAIT) == 0) {
    328 			lf_free(lock);
    329 			return EAGAIN;
    330 		}
    331 		/*
    332 		 * We are blocked. Since flock style locks cover
    333 		 * the whole file, there is no chance for deadlock.
    334 		 * For byte-range locks we must check for deadlock.
    335 		 *
    336 		 * Deadlock detection is done by looking through the
    337 		 * wait channels to see if there are any cycles that
    338 		 * involve us. MAXDEPTH is set just to make sure we
    339 		 * do not go off into neverneverland.
    340 		 */
    341 		if ((lock->lf_flags & F_POSIX) &&
    342 		    (block->lf_flags & F_POSIX)) {
    343 			struct lwp *wlwp;
    344 			struct lockf *waitblock;
    345 			int i = 0;
    346 
    347 			/*
    348 			 * The block is waiting on something.  if_lwp will be
    349 			 * 0 once the lock is granted, so we terminate the
    350 			 * loop if we find this.
    351 			 */
    352 			wlwp = block->lf_lwp;
    353 			while (wlwp && (i++ < maxlockdepth)) {
    354 				waitblock = (struct lockf *)wlwp->l_wchan;
    355 				/* Get the owner of the blocking lock */
    356 				waitblock = waitblock->lf_next;
    357 				if ((waitblock->lf_flags & F_POSIX) == 0)
    358 					break;
    359 				wlwp = waitblock->lf_lwp;
    360 				if (wlwp == lock->lf_lwp) {
    361 					lf_free(lock);
    362 					return EDEADLK;
    363 				}
    364 			}
    365 			/*
    366 			 * If we're still following a dependency chain
    367 			 * after maxlockdepth iterations, assume we're in
    368 			 * a cycle to be safe.
    369 			 */
    370 			if (i >= maxlockdepth) {
    371 				lf_free(lock);
    372 				return EDEADLK;
    373 			}
    374 		}
    375 		/*
    376 		 * For flock type locks, we must first remove
    377 		 * any shared locks that we hold before we sleep
    378 		 * waiting for an exclusive lock.
    379 		 */
    380 		if ((lock->lf_flags & F_FLOCK) &&
    381 		    lock->lf_type == F_WRLCK) {
    382 			lock->lf_type = F_UNLCK;
    383 			(void) lf_clearlock(lock, NULL);
    384 			lock->lf_type = F_WRLCK;
    385 		}
    386 		/*
    387 		 * Add our lock to the blocked list and sleep until we're free.
    388 		 * Remember who blocked us (for deadlock detection).
    389 		 */
    390 		lock->lf_next = block;
    391 		TAILQ_INSERT_TAIL(&block->lf_blkhd, lock, lf_block);
    392 #ifdef LOCKF_DEBUG
    393 		if (lockf_debug & 1) {
    394 			lf_print("lf_setlock: blocking on", block);
    395 			lf_printlist("lf_setlock", block);
    396 		}
    397 #endif /* LOCKF_DEBUG */
    398 		error = ltsleep(lock, priority, lockstr, 0, interlock);
    399 
    400 		/*
    401 		 * We may have been awakened by a signal (in
    402 		 * which case we must remove ourselves from the
    403 		 * blocked list) and/or by another process
    404 		 * releasing a lock (in which case we have already
    405 		 * been removed from the blocked list and our
    406 		 * lf_next field set to NOLOCKF).
    407 		 */
    408 		if (lock->lf_next != NOLOCKF) {
    409 			TAILQ_REMOVE(&lock->lf_next->lf_blkhd, lock, lf_block);
    410 			lock->lf_next = NOLOCKF;
    411 		}
    412 		if (error) {
    413 			lf_free(lock);
    414 			return error;
    415 		}
    416 	}
    417 	/*
    418 	 * No blocks!!  Add the lock.  Note that we will
    419 	 * downgrade or upgrade any overlapping locks this
    420 	 * process already owns.
    421 	 *
    422 	 * Skip over locks owned by other processes.
    423 	 * Handle any locks that overlap and are owned by ourselves.
    424 	 */
    425 	lock->lf_lwp = 0;
    426 	prev = head;
    427 	block = *head;
    428 	needtolink = 1;
    429 	for (;;) {
    430 		ovcase = lf_findoverlap(block, lock, SELF, &prev, &overlap);
    431 		if (ovcase)
    432 			block = overlap->lf_next;
    433 		/*
    434 		 * Six cases:
    435 		 *	0) no overlap
    436 		 *	1) overlap == lock
    437 		 *	2) overlap contains lock
    438 		 *	3) lock contains overlap
    439 		 *	4) overlap starts before lock
    440 		 *	5) overlap ends after lock
    441 		 */
    442 		switch (ovcase) {
    443 		case 0: /* no overlap */
    444 			if (needtolink) {
    445 				*prev = lock;
    446 				lock->lf_next = overlap;
    447 			}
    448 			break;
    449 
    450 		case 1: /* overlap == lock */
    451 			/*
    452 			 * If downgrading lock, others may be
    453 			 * able to acquire it.
    454 			 */
    455 			if (lock->lf_type == F_RDLCK &&
    456 			    overlap->lf_type == F_WRLCK)
    457 				lf_wakelock(overlap);
    458 			overlap->lf_type = lock->lf_type;
    459 			lf_free(lock);
    460 			lock = overlap; /* for debug output below */
    461 			break;
    462 
    463 		case 2: /* overlap contains lock */
    464 			/*
    465 			 * Check for common starting point and different types.
    466 			 */
    467 			if (overlap->lf_type == lock->lf_type) {
    468 				lf_free(lock);
    469 				lock = overlap; /* for debug output below */
    470 				break;
    471 			}
    472 			if (overlap->lf_start == lock->lf_start) {
    473 				*prev = lock;
    474 				lock->lf_next = overlap;
    475 				overlap->lf_start = lock->lf_end + 1;
    476 			} else
    477 				lf_split(overlap, lock, sparelock);
    478 			lf_wakelock(overlap);
    479 			break;
    480 
    481 		case 3: /* lock contains overlap */
    482 			/*
    483 			 * If downgrading lock, others may be able to
    484 			 * acquire it, otherwise take the list.
    485 			 */
    486 			if (lock->lf_type == F_RDLCK &&
    487 			    overlap->lf_type == F_WRLCK) {
    488 				lf_wakelock(overlap);
    489 			} else {
    490 				while ((ltmp = TAILQ_FIRST(&overlap->lf_blkhd))) {
    491 					KASSERT(ltmp->lf_next == overlap);
    492 					TAILQ_REMOVE(&overlap->lf_blkhd, ltmp,
    493 					    lf_block);
    494 					ltmp->lf_next = lock;
    495 					TAILQ_INSERT_TAIL(&lock->lf_blkhd,
    496 					    ltmp, lf_block);
    497 				}
    498 			}
    499 			/*
    500 			 * Add the new lock if necessary and delete the overlap.
    501 			 */
    502 			if (needtolink) {
    503 				*prev = lock;
    504 				lock->lf_next = overlap->lf_next;
    505 				prev = &lock->lf_next;
    506 				needtolink = 0;
    507 			} else
    508 				*prev = overlap->lf_next;
    509 			lf_free(overlap);
    510 			continue;
    511 
    512 		case 4: /* overlap starts before lock */
    513 			/*
    514 			 * Add lock after overlap on the list.
    515 			 */
    516 			lock->lf_next = overlap->lf_next;
    517 			overlap->lf_next = lock;
    518 			overlap->lf_end = lock->lf_start - 1;
    519 			prev = &lock->lf_next;
    520 			lf_wakelock(overlap);
    521 			needtolink = 0;
    522 			continue;
    523 
    524 		case 5: /* overlap ends after lock */
    525 			/*
    526 			 * Add the new lock before overlap.
    527 			 */
    528 			if (needtolink) {
    529 				*prev = lock;
    530 				lock->lf_next = overlap;
    531 			}
    532 			overlap->lf_start = lock->lf_end + 1;
    533 			lf_wakelock(overlap);
    534 			break;
    535 		}
    536 		break;
    537 	}
    538 #ifdef LOCKF_DEBUG
    539 	if (lockf_debug & 1) {
    540 		lf_print("lf_setlock: got the lock", lock);
    541 		lf_printlist("lf_setlock", lock);
    542 	}
    543 #endif /* LOCKF_DEBUG */
    544 	return 0;
    545 }
    546 
    547 /*
    548  * Remove a byte-range lock on an inode.
    549  *
    550  * Generally, find the lock (or an overlap to that lock)
    551  * and remove it (or shrink it), then wakeup anyone we can.
    552  */
    553 static int
    554 lf_clearlock(struct lockf *unlock, struct lockf **sparelock)
    555 {
    556 	struct lockf **head = unlock->lf_head;
    557 	struct lockf *lf = *head;
    558 	struct lockf *overlap, **prev;
    559 	int ovcase;
    560 
    561 	if (lf == NOLOCKF)
    562 		return 0;
    563 #ifdef LOCKF_DEBUG
    564 	if (unlock->lf_type != F_UNLCK)
    565 		panic("lf_clearlock: bad type");
    566 	if (lockf_debug & 1)
    567 		lf_print("lf_clearlock", unlock);
    568 #endif /* LOCKF_DEBUG */
    569 	prev = head;
    570 	while ((ovcase = lf_findoverlap(lf, unlock, SELF,
    571 					&prev, &overlap)) != 0) {
    572 		/*
    573 		 * Wakeup the list of locks to be retried.
    574 		 */
    575 		lf_wakelock(overlap);
    576 
    577 		switch (ovcase) {
    578 
    579 		case 1: /* overlap == lock */
    580 			*prev = overlap->lf_next;
    581 			lf_free(overlap);
    582 			break;
    583 
    584 		case 2: /* overlap contains lock: split it */
    585 			if (overlap->lf_start == unlock->lf_start) {
    586 				overlap->lf_start = unlock->lf_end + 1;
    587 				break;
    588 			}
    589 			lf_split(overlap, unlock, sparelock);
    590 			overlap->lf_next = unlock->lf_next;
    591 			break;
    592 
    593 		case 3: /* lock contains overlap */
    594 			*prev = overlap->lf_next;
    595 			lf = overlap->lf_next;
    596 			lf_free(overlap);
    597 			continue;
    598 
    599 		case 4: /* overlap starts before lock */
    600 			overlap->lf_end = unlock->lf_start - 1;
    601 			prev = &overlap->lf_next;
    602 			lf = overlap->lf_next;
    603 			continue;
    604 
    605 		case 5: /* overlap ends after lock */
    606 			overlap->lf_start = unlock->lf_end + 1;
    607 			break;
    608 		}
    609 		break;
    610 	}
    611 #ifdef LOCKF_DEBUG
    612 	if (lockf_debug & 1)
    613 		lf_printlist("lf_clearlock", unlock);
    614 #endif /* LOCKF_DEBUG */
    615 	return 0;
    616 }
    617 
    618 /*
    619  * Check whether there is a blocking lock,
    620  * and if so return its process identifier.
    621  */
    622 static int
    623 lf_getlock(struct lockf *lock, struct flock *fl)
    624 {
    625 	struct lockf *block;
    626 
    627 #ifdef LOCKF_DEBUG
    628 	if (lockf_debug & 1)
    629 		lf_print("lf_getlock", lock);
    630 #endif /* LOCKF_DEBUG */
    631 
    632 	if ((block = lf_getblock(lock)) != NULL) {
    633 		fl->l_type = block->lf_type;
    634 		fl->l_whence = SEEK_SET;
    635 		fl->l_start = block->lf_start;
    636 		if (block->lf_end == -1)
    637 			fl->l_len = 0;
    638 		else
    639 			fl->l_len = block->lf_end - block->lf_start + 1;
    640 		if (block->lf_flags & F_POSIX)
    641 			fl->l_pid = ((struct proc *)block->lf_id)->p_pid;
    642 		else
    643 			fl->l_pid = -1;
    644 	} else {
    645 		fl->l_type = F_UNLCK;
    646 	}
    647 	return 0;
    648 }
    649 
    650 /*
    651  * Walk the list of locks for an inode and
    652  * return the first blocking lock.
    653  */
    654 static struct lockf *
    655 lf_getblock(struct lockf *lock)
    656 {
    657 	struct lockf **prev, *overlap, *lf = *(lock->lf_head);
    658 
    659 	prev = lock->lf_head;
    660 	while (lf_findoverlap(lf, lock, OTHERS, &prev, &overlap) != 0) {
    661 		/*
    662 		 * We've found an overlap, see if it blocks us
    663 		 */
    664 		if ((lock->lf_type == F_WRLCK || overlap->lf_type == F_WRLCK))
    665 			return overlap;
    666 		/*
    667 		 * Nope, point to the next one on the list and
    668 		 * see if it blocks us
    669 		 */
    670 		lf = overlap->lf_next;
    671 	}
    672 	return NOLOCKF;
    673 }
    674 
    675 /*
    676  * Walk the list of locks for an inode to
    677  * find an overlapping lock (if any).
    678  *
    679  * NOTE: this returns only the FIRST overlapping lock.  There
    680  *	 may be more than one.
    681  */
    682 static int
    683 lf_findoverlap(struct lockf *lf, struct lockf *lock, int type,
    684     struct lockf ***prev, struct lockf **overlap)
    685 {
    686 	off_t start, end;
    687 
    688 	*overlap = lf;
    689 	if (lf == NOLOCKF)
    690 		return 0;
    691 #ifdef LOCKF_DEBUG
    692 	if (lockf_debug & 2)
    693 		lf_print("lf_findoverlap: looking for overlap in", lock);
    694 #endif /* LOCKF_DEBUG */
    695 	start = lock->lf_start;
    696 	end = lock->lf_end;
    697 	while (lf != NOLOCKF) {
    698 		if (((type == SELF) && lf->lf_id != lock->lf_id) ||
    699 		    ((type == OTHERS) && lf->lf_id == lock->lf_id)) {
    700 			*prev = &lf->lf_next;
    701 			*overlap = lf = lf->lf_next;
    702 			continue;
    703 		}
    704 #ifdef LOCKF_DEBUG
    705 		if (lockf_debug & 2)
    706 			lf_print("\tchecking", lf);
    707 #endif /* LOCKF_DEBUG */
    708 		/*
    709 		 * OK, check for overlap
    710 		 *
    711 		 * Six cases:
    712 		 *	0) no overlap
    713 		 *	1) overlap == lock
    714 		 *	2) overlap contains lock
    715 		 *	3) lock contains overlap
    716 		 *	4) overlap starts before lock
    717 		 *	5) overlap ends after lock
    718 		 */
    719 		if ((lf->lf_end != -1 && start > lf->lf_end) ||
    720 		    (end != -1 && lf->lf_start > end)) {
    721 			/* Case 0 */
    722 #ifdef LOCKF_DEBUG
    723 			if (lockf_debug & 2)
    724 				printf("no overlap\n");
    725 #endif /* LOCKF_DEBUG */
    726 			if ((type & SELF) && end != -1 && lf->lf_start > end)
    727 				return 0;
    728 			*prev = &lf->lf_next;
    729 			*overlap = lf = lf->lf_next;
    730 			continue;
    731 		}
    732 		if ((lf->lf_start == start) && (lf->lf_end == end)) {
    733 			/* Case 1 */
    734 #ifdef LOCKF_DEBUG
    735 			if (lockf_debug & 2)
    736 				printf("overlap == lock\n");
    737 #endif /* LOCKF_DEBUG */
    738 			return 1;
    739 		}
    740 		if ((lf->lf_start <= start) &&
    741 		    (end != -1) &&
    742 		    ((lf->lf_end >= end) || (lf->lf_end == -1))) {
    743 			/* Case 2 */
    744 #ifdef LOCKF_DEBUG
    745 			if (lockf_debug & 2)
    746 				printf("overlap contains lock\n");
    747 #endif /* LOCKF_DEBUG */
    748 			return 2;
    749 		}
    750 		if (start <= lf->lf_start &&
    751 		           (end == -1 ||
    752 			   (lf->lf_end != -1 && end >= lf->lf_end))) {
    753 			/* Case 3 */
    754 #ifdef LOCKF_DEBUG
    755 			if (lockf_debug & 2)
    756 				printf("lock contains overlap\n");
    757 #endif /* LOCKF_DEBUG */
    758 			return 3;
    759 		}
    760 		if ((lf->lf_start < start) &&
    761 			((lf->lf_end >= start) || (lf->lf_end == -1))) {
    762 			/* Case 4 */
    763 #ifdef LOCKF_DEBUG
    764 			if (lockf_debug & 2)
    765 				printf("overlap starts before lock\n");
    766 #endif /* LOCKF_DEBUG */
    767 			return 4;
    768 		}
    769 		if ((lf->lf_start > start) &&
    770 			(end != -1) &&
    771 			((lf->lf_end > end) || (lf->lf_end == -1))) {
    772 			/* Case 5 */
    773 #ifdef LOCKF_DEBUG
    774 			if (lockf_debug & 2)
    775 				printf("overlap ends after lock\n");
    776 #endif /* LOCKF_DEBUG */
    777 			return 5;
    778 		}
    779 		panic("lf_findoverlap: default");
    780 	}
    781 	return 0;
    782 }
    783 
    784 /*
    785  * Split a lock and a contained region into
    786  * two or three locks as necessary.
    787  */
    788 static void
    789 lf_split(struct lockf *lock1, struct lockf *lock2, struct lockf **sparelock)
    790 {
    791 	struct lockf *splitlock;
    792 
    793 #ifdef LOCKF_DEBUG
    794 	if (lockf_debug & 2) {
    795 		lf_print("lf_split", lock1);
    796 		lf_print("splitting from", lock2);
    797 	}
    798 #endif /* LOCKF_DEBUG */
    799 	/*
    800 	 * Check to see if spliting into only two pieces.
    801 	 */
    802 	if (lock1->lf_start == lock2->lf_start) {
    803 		lock1->lf_start = lock2->lf_end + 1;
    804 		lock2->lf_next = lock1;
    805 		return;
    806 	}
    807 	if (lock1->lf_end == lock2->lf_end) {
    808 		lock1->lf_end = lock2->lf_start - 1;
    809 		lock2->lf_next = lock1->lf_next;
    810 		lock1->lf_next = lock2;
    811 		return;
    812 	}
    813 	/*
    814 	 * Make a new lock consisting of the last part of
    815 	 * the encompassing lock
    816 	 */
    817 	splitlock = *sparelock;
    818 	*sparelock = NULL;
    819 	memcpy(splitlock, lock1, sizeof(*splitlock));
    820 	splitlock->lf_start = lock2->lf_end + 1;
    821 	TAILQ_INIT(&splitlock->lf_blkhd);
    822 	lock1->lf_end = lock2->lf_start - 1;
    823 	/*
    824 	 * OK, now link it in
    825 	 */
    826 	splitlock->lf_next = lock1->lf_next;
    827 	lock2->lf_next = splitlock;
    828 	lock1->lf_next = lock2;
    829 }
    830 
    831 /*
    832  * Wakeup a blocklist
    833  */
    834 static void
    835 lf_wakelock(struct lockf *listhead)
    836 {
    837 	struct lockf *wakelock;
    838 
    839 	while ((wakelock = TAILQ_FIRST(&listhead->lf_blkhd))) {
    840 		KASSERT(wakelock->lf_next == listhead);
    841 		TAILQ_REMOVE(&listhead->lf_blkhd, wakelock, lf_block);
    842 		wakelock->lf_next = NOLOCKF;
    843 #ifdef LOCKF_DEBUG
    844 		if (lockf_debug & 2)
    845 			lf_print("lf_wakelock: awakening", wakelock);
    846 #endif
    847 		wakeup(wakelock);
    848 	}
    849 }
    850 
    851 #ifdef LOCKF_DEBUG
    852 /*
    853  * Print out a lock.
    854  */
    855 static void
    856 lf_print(char *tag, struct lockf *lock)
    857 {
    858 
    859 	printf("%s: lock %p for ", tag, lock);
    860 	if (lock->lf_flags & F_POSIX)
    861 		printf("proc %d", ((struct proc *)lock->lf_id)->p_pid);
    862 	else
    863 		printf("file %p", (struct file *)lock->lf_id);
    864 	printf(" %s, start %qx, end %qx",
    865 		lock->lf_type == F_RDLCK ? "shared" :
    866 		lock->lf_type == F_WRLCK ? "exclusive" :
    867 		lock->lf_type == F_UNLCK ? "unlock" :
    868 		"unknown", lock->lf_start, lock->lf_end);
    869 	if (TAILQ_FIRST(&lock->lf_blkhd))
    870 		printf(" block %p\n", TAILQ_FIRST(&lock->lf_blkhd));
    871 	else
    872 		printf("\n");
    873 }
    874 
    875 static void
    876 lf_printlist(char *tag, struct lockf *lock)
    877 {
    878 	struct lockf *lf, *blk;
    879 
    880 	printf("%s: Lock list:\n", tag);
    881 	for (lf = *lock->lf_head; lf; lf = lf->lf_next) {
    882 		printf("\tlock %p for ", lf);
    883 		if (lf->lf_flags & F_POSIX)
    884 			printf("proc %d", ((struct proc *)lf->lf_id)->p_pid);
    885 		else
    886 			printf("file %p", (struct file *)lf->lf_id);
    887 		printf(", %s, start %qx, end %qx",
    888 			lf->lf_type == F_RDLCK ? "shared" :
    889 			lf->lf_type == F_WRLCK ? "exclusive" :
    890 			lf->lf_type == F_UNLCK ? "unlock" :
    891 			"unknown", lf->lf_start, lf->lf_end);
    892 		TAILQ_FOREACH(blk, &lf->lf_blkhd, lf_block) {
    893 			if (blk->lf_flags & F_POSIX)
    894 				printf("proc %d",
    895 				    ((struct proc *)blk->lf_id)->p_pid);
    896 			else
    897 				printf("file %p", (struct file *)blk->lf_id);
    898 			printf(", %s, start %qx, end %qx",
    899 				blk->lf_type == F_RDLCK ? "shared" :
    900 				blk->lf_type == F_WRLCK ? "exclusive" :
    901 				blk->lf_type == F_UNLCK ? "unlock" :
    902 				"unknown", blk->lf_start, blk->lf_end);
    903 			if (TAILQ_FIRST(&blk->lf_blkhd))
    904 				 panic("lf_printlist: bad list");
    905 		}
    906 		printf("\n");
    907 	}
    908 }
    909 #endif /* LOCKF_DEBUG */
    910