Home | History | Annotate | Line # | Download | only in kern
vfs_lockf.c revision 1.31
      1 /*	$NetBSD: vfs_lockf.c,v 1.31 2003/05/03 15:02:54 fvdl 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. All advertising materials mentioning features or use of this software
     19  *    must display the following acknowledgement:
     20  *	This product includes software developed by the University of
     21  *	California, Berkeley and its contributors.
     22  * 4. Neither the name of the University nor the names of its contributors
     23  *    may be used to endorse or promote products derived from this software
     24  *    without specific prior written permission.
     25  *
     26  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     27  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     28  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     29  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     30  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     31  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     32  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     33  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     34  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     35  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     36  * SUCH DAMAGE.
     37  *
     38  *	@(#)ufs_lockf.c	8.4 (Berkeley) 10/26/94
     39  */
     40 
     41 #include <sys/cdefs.h>
     42 __KERNEL_RCSID(0, "$NetBSD: vfs_lockf.c,v 1.31 2003/05/03 15:02:54 fvdl Exp $");
     43 
     44 #include <sys/param.h>
     45 #include <sys/systm.h>
     46 #include <sys/kernel.h>
     47 #include <sys/file.h>
     48 #include <sys/proc.h>
     49 #include <sys/vnode.h>
     50 #include <sys/malloc.h>
     51 #include <sys/fcntl.h>
     52 #include <sys/lockf.h>
     53 
     54 MALLOC_DEFINE(M_LOCKF, "lockf", "Byte-range locking structures");
     55 
     56 /*
     57  * This variable controls the maximum number of processes that will
     58  * be checked in doing deadlock detection.
     59  */
     60 int maxlockdepth = MAXDEPTH;
     61 
     62 #ifdef LOCKF_DEBUG
     63 int	lockf_debug = 0;
     64 #endif
     65 
     66 #define NOLOCKF (struct lockf *)0
     67 #define SELF	0x1
     68 #define OTHERS	0x2
     69 
     70 static int lf_clearlock(struct lockf *, struct lockf **);
     71 static int lf_findoverlap(struct lockf *,
     72 	    struct lockf *, int, struct lockf ***, struct lockf **);
     73 static struct lockf *lf_getblock(struct lockf *);
     74 static int lf_getlock(struct lockf *, struct flock *);
     75 static int lf_setlock(struct lockf *, struct lockf **, struct simplelock *);
     76 static void lf_split(struct lockf *, struct lockf *, struct lockf **);
     77 static void lf_wakelock(struct lockf *);
     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  * Use pools for lock allocation.
     91  */
     92 
     93 /*
     94  * If there's a lot of lock contention on a single vnode, locking
     95  * schemes which allow for more paralleism would be needed.  Given how
     96  * infrequently byte-range locks are actually used in typical BSD
     97  * code, a more complex approach probably isn't worth it.
     98  */
     99 
    100 /*
    101  * Do an advisory lock operation.
    102  */
    103 int
    104 lf_advlock(struct vop_advlock_args *ap, struct lockf **head, off_t size)
    105 {
    106 	struct flock *fl = ap->a_fl;
    107 	struct lockf *lock = NULL;
    108 	struct lockf *sparelock;
    109 	struct simplelock *interlock = &ap->a_vp->v_interlock;
    110 	off_t start, end;
    111 	int error;
    112 
    113 	/*
    114 	 * Convert the flock structure into a start and end.
    115 	 */
    116 	switch (fl->l_whence) {
    117 	case SEEK_SET:
    118 	case SEEK_CUR:
    119 		/*
    120 		 * Caller is responsible for adding any necessary offset
    121 		 * when SEEK_CUR is used.
    122 		 */
    123 		start = fl->l_start;
    124 		break;
    125 
    126 	case SEEK_END:
    127 		start = size + fl->l_start;
    128 		break;
    129 
    130 	default:
    131 		return EINVAL;
    132 	}
    133 	if (start < 0)
    134 		return EINVAL;
    135 
    136 	/*
    137 	 * allocate locks before acquire simple lock.
    138 	 * we need two locks in the worst case.
    139 	 */
    140 	switch (ap->a_op) {
    141 	case F_SETLK:
    142 	case F_UNLCK:
    143 		/*
    144 		 * XXX for F_UNLCK case, we can re-use lock.
    145 		 */
    146 		if ((fl->l_type & F_FLOCK) == 0) {
    147 			/*
    148 			 * byte-range lock might need one more lock.
    149 			 */
    150 			MALLOC(sparelock, struct lockf *, sizeof(*lock),
    151 			    M_LOCKF, M_WAITOK);
    152 			if (sparelock == NULL) {
    153 				error = ENOMEM;
    154 				goto quit;
    155 			}
    156 			break;
    157 		}
    158 		/* FALLTHROUGH */
    159 
    160 	case F_GETLK:
    161 		sparelock = NULL;
    162 		break;
    163 
    164 	default:
    165 		return EINVAL;
    166 	}
    167 
    168 	MALLOC(lock, struct lockf *, sizeof(*lock), M_LOCKF, M_WAITOK);
    169 	if (lock == NULL) {
    170 		error = ENOMEM;
    171 		goto quit;
    172 	}
    173 
    174 	simple_lock(interlock);
    175 
    176 	/*
    177 	 * Avoid the common case of unlocking when inode has no locks.
    178 	 */
    179 	if (*head == (struct lockf *)0) {
    180 		if (ap->a_op != F_SETLK) {
    181 			fl->l_type = F_UNLCK;
    182 			error = 0;
    183 			goto quit_unlock;
    184 		}
    185 	}
    186 
    187 	if (fl->l_len == 0)
    188 		end = -1;
    189 	else
    190 		end = start + fl->l_len - 1;
    191 	/*
    192 	 * Create the lockf structure.
    193 	 */
    194 	lock->lf_start = start;
    195 	lock->lf_end = end;
    196 	/* XXX NJWLWP
    197 	 * I don't want to make the entire VFS universe use LWPs, because
    198 	 * they don't need them, for the most part. This is an exception,
    199 	 * and a kluge.
    200 	 */
    201 
    202 	lock->lf_head = head;
    203 	lock->lf_type = fl->l_type;
    204 	lock->lf_next = (struct lockf *)0;
    205 	TAILQ_INIT(&lock->lf_blkhd);
    206 	lock->lf_flags = ap->a_flags;
    207 	if (lock->lf_flags & F_POSIX) {
    208 		KASSERT(curproc == (struct proc *)ap->a_id);
    209 	}
    210 	lock->lf_id = (struct proc *)ap->a_id;
    211 	lock->lf_lwp = curlwp;
    212 
    213 	/*
    214 	 * Do the requested operation.
    215 	 */
    216 	switch (ap->a_op) {
    217 
    218 	case F_SETLK:
    219 		error = lf_setlock(lock, &sparelock, interlock);
    220 		lock = NULL; /* lf_setlock freed it */
    221 		break;
    222 
    223 	case F_UNLCK:
    224 		error = lf_clearlock(lock, &sparelock);
    225 		break;
    226 
    227 	case F_GETLK:
    228 		error = lf_getlock(lock, fl);
    229 		break;
    230 
    231 	default:
    232 		break;
    233 		/* NOTREACHED */
    234 	}
    235 
    236 quit_unlock:
    237 	simple_unlock(interlock);
    238 quit:
    239 	if (lock)
    240 		FREE(lock, M_LOCKF);
    241 	if (sparelock)
    242 		FREE(sparelock, M_LOCKF);
    243 
    244 	return error;
    245 }
    246 
    247 /*
    248  * Set a byte-range lock.
    249  */
    250 static int
    251 lf_setlock(struct lockf *lock, struct lockf **sparelock,
    252     struct simplelock *interlock)
    253 {
    254 	struct lockf *block;
    255 	struct lockf **head = lock->lf_head;
    256 	struct lockf **prev, *overlap, *ltmp;
    257 	static char lockstr[] = "lockf";
    258 	int ovcase, priority, needtolink, error;
    259 
    260 #ifdef LOCKF_DEBUG
    261 	if (lockf_debug & 1)
    262 		lf_print("lf_setlock", lock);
    263 #endif /* LOCKF_DEBUG */
    264 
    265 	/*
    266 	 * Set the priority
    267 	 */
    268 	priority = PLOCK;
    269 	if (lock->lf_type == F_WRLCK)
    270 		priority += 4;
    271 	priority |= PCATCH;
    272 	/*
    273 	 * Scan lock list for this file looking for locks that would block us.
    274 	 */
    275 	while ((block = lf_getblock(lock)) != NULL) {
    276 		/*
    277 		 * Free the structure and return if nonblocking.
    278 		 */
    279 		if ((lock->lf_flags & F_WAIT) == 0) {
    280 			FREE(lock, M_LOCKF);
    281 			return EAGAIN;
    282 		}
    283 		/*
    284 		 * We are blocked. Since flock style locks cover
    285 		 * the whole file, there is no chance for deadlock.
    286 		 * For byte-range locks we must check for deadlock.
    287 		 *
    288 		 * Deadlock detection is done by looking through the
    289 		 * wait channels to see if there are any cycles that
    290 		 * involve us. MAXDEPTH is set just to make sure we
    291 		 * do not go off into neverneverland.
    292 		 */
    293 		if ((lock->lf_flags & F_POSIX) &&
    294 		    (block->lf_flags & F_POSIX)) {
    295 			struct lwp *wlwp;
    296 			struct lockf *waitblock;
    297 			int i = 0;
    298 
    299 			/*
    300 			 * The block is waiting on something.  if_lwp will be
    301 			 * 0 once the lock is granted, so we terminate the
    302 			 * loop if we find this.
    303 			 */
    304 			wlwp = block->lf_lwp;
    305 			while (wlwp && (i++ < maxlockdepth)) {
    306 				waitblock = (struct lockf *)wlwp->l_wchan;
    307 				/* Get the owner of the blocking lock */
    308 				waitblock = waitblock->lf_next;
    309 				if ((waitblock->lf_flags & F_POSIX) == 0)
    310 					break;
    311 				wlwp = waitblock->lf_lwp;
    312 				if (wlwp == lock->lf_lwp) {
    313 					FREE(lock, M_LOCKF);
    314 					return EDEADLK;
    315 				}
    316 			}
    317 			/*
    318 			 * If we're still following a dependancy chain
    319 			 * after maxlockdepth iterations, assume we're in
    320 			 * a cycle to be safe.
    321 			 */
    322 			if (i >= maxlockdepth) {
    323 				FREE(lock, M_LOCKF);
    324 				return EDEADLK;
    325 			}
    326 		}
    327 		/*
    328 		 * For flock type locks, we must first remove
    329 		 * any shared locks that we hold before we sleep
    330 		 * waiting for an exclusive lock.
    331 		 */
    332 		if ((lock->lf_flags & F_FLOCK) &&
    333 		    lock->lf_type == F_WRLCK) {
    334 			lock->lf_type = F_UNLCK;
    335 			(void) lf_clearlock(lock, NULL);
    336 			lock->lf_type = F_WRLCK;
    337 		}
    338 		/*
    339 		 * Add our lock to the blocked list and sleep until we're free.
    340 		 * Remember who blocked us (for deadlock detection).
    341 		 */
    342 		lock->lf_next = block;
    343 		TAILQ_INSERT_TAIL(&block->lf_blkhd, lock, lf_block);
    344 #ifdef LOCKF_DEBUG
    345 		if (lockf_debug & 1) {
    346 			lf_print("lf_setlock: blocking on", block);
    347 			lf_printlist("lf_setlock", block);
    348 		}
    349 #endif /* LOCKF_DEBUG */
    350 		error = ltsleep(lock, priority, lockstr, 0, interlock);
    351 
    352 		/*
    353 		 * We may have been awakened by a signal (in
    354 		 * which case we must remove ourselves from the
    355 		 * blocked list) and/or by another process
    356 		 * releasing a lock (in which case we have already
    357 		 * been removed from the blocked list and our
    358 		 * lf_next field set to NOLOCKF).
    359 		 */
    360 		if (lock->lf_next != NOLOCKF) {
    361 			TAILQ_REMOVE(&lock->lf_next->lf_blkhd, lock, lf_block);
    362 			lock->lf_next = NOLOCKF;
    363 		}
    364 		if (error) {
    365 			FREE(lock, M_LOCKF);
    366 			return error;
    367 		}
    368 	}
    369 	/*
    370 	 * No blocks!!  Add the lock.  Note that we will
    371 	 * downgrade or upgrade any overlapping locks this
    372 	 * process already owns.
    373 	 *
    374 	 * Skip over locks owned by other processes.
    375 	 * Handle any locks that overlap and are owned by ourselves.
    376 	 */
    377 	lock->lf_lwp = 0;
    378 	prev = head;
    379 	block = *head;
    380 	needtolink = 1;
    381 	for (;;) {
    382 		ovcase = lf_findoverlap(block, lock, SELF, &prev, &overlap);
    383 		if (ovcase)
    384 			block = overlap->lf_next;
    385 		/*
    386 		 * Six cases:
    387 		 *	0) no overlap
    388 		 *	1) overlap == lock
    389 		 *	2) overlap contains lock
    390 		 *	3) lock contains overlap
    391 		 *	4) overlap starts before lock
    392 		 *	5) overlap ends after lock
    393 		 */
    394 		switch (ovcase) {
    395 		case 0: /* no overlap */
    396 			if (needtolink) {
    397 				*prev = lock;
    398 				lock->lf_next = overlap;
    399 			}
    400 			break;
    401 
    402 		case 1: /* overlap == lock */
    403 			/*
    404 			 * If downgrading lock, others may be
    405 			 * able to acquire it.
    406 			 */
    407 			if (lock->lf_type == F_RDLCK &&
    408 			    overlap->lf_type == F_WRLCK)
    409 				lf_wakelock(overlap);
    410 			overlap->lf_type = lock->lf_type;
    411 			FREE(lock, M_LOCKF);
    412 			lock = overlap; /* for debug output below */
    413 			break;
    414 
    415 		case 2: /* overlap contains lock */
    416 			/*
    417 			 * Check for common starting point and different types.
    418 			 */
    419 			if (overlap->lf_type == lock->lf_type) {
    420 				FREE(lock, M_LOCKF);
    421 				lock = overlap; /* for debug output below */
    422 				break;
    423 			}
    424 			if (overlap->lf_start == lock->lf_start) {
    425 				*prev = lock;
    426 				lock->lf_next = overlap;
    427 				overlap->lf_start = lock->lf_end + 1;
    428 			} else
    429 				lf_split(overlap, lock, sparelock);
    430 			lf_wakelock(overlap);
    431 			break;
    432 
    433 		case 3: /* lock contains overlap */
    434 			/*
    435 			 * If downgrading lock, others may be able to
    436 			 * acquire it, otherwise take the list.
    437 			 */
    438 			if (lock->lf_type == F_RDLCK &&
    439 			    overlap->lf_type == F_WRLCK) {
    440 				lf_wakelock(overlap);
    441 			} else {
    442 				while ((ltmp = TAILQ_FIRST(&overlap->lf_blkhd))) {
    443 					KASSERT(ltmp->lf_next == overlap);
    444 					TAILQ_REMOVE(&overlap->lf_blkhd, ltmp,
    445 					    lf_block);
    446 					ltmp->lf_next = lock;
    447 					TAILQ_INSERT_TAIL(&lock->lf_blkhd,
    448 					    ltmp, lf_block);
    449 				}
    450 			}
    451 			/*
    452 			 * Add the new lock if necessary and delete the overlap.
    453 			 */
    454 			if (needtolink) {
    455 				*prev = lock;
    456 				lock->lf_next = overlap->lf_next;
    457 				prev = &lock->lf_next;
    458 				needtolink = 0;
    459 			} else
    460 				*prev = overlap->lf_next;
    461 			FREE(overlap, M_LOCKF);
    462 			continue;
    463 
    464 		case 4: /* overlap starts before lock */
    465 			/*
    466 			 * Add lock after overlap on the list.
    467 			 */
    468 			lock->lf_next = overlap->lf_next;
    469 			overlap->lf_next = lock;
    470 			overlap->lf_end = lock->lf_start - 1;
    471 			prev = &lock->lf_next;
    472 			lf_wakelock(overlap);
    473 			needtolink = 0;
    474 			continue;
    475 
    476 		case 5: /* overlap ends after lock */
    477 			/*
    478 			 * Add the new lock before overlap.
    479 			 */
    480 			if (needtolink) {
    481 				*prev = lock;
    482 				lock->lf_next = overlap;
    483 			}
    484 			overlap->lf_start = lock->lf_end + 1;
    485 			lf_wakelock(overlap);
    486 			break;
    487 		}
    488 		break;
    489 	}
    490 #ifdef LOCKF_DEBUG
    491 	if (lockf_debug & 1) {
    492 		lf_print("lf_setlock: got the lock", lock);
    493 		lf_printlist("lf_setlock", lock);
    494 	}
    495 #endif /* LOCKF_DEBUG */
    496 	return 0;
    497 }
    498 
    499 /*
    500  * Remove a byte-range lock on an inode.
    501  *
    502  * Generally, find the lock (or an overlap to that lock)
    503  * and remove it (or shrink it), then wakeup anyone we can.
    504  */
    505 static int
    506 lf_clearlock(struct lockf *unlock, struct lockf **sparelock)
    507 {
    508 	struct lockf **head = unlock->lf_head;
    509 	struct lockf *lf = *head;
    510 	struct lockf *overlap, **prev;
    511 	int ovcase;
    512 
    513 	if (lf == NOLOCKF)
    514 		return 0;
    515 #ifdef LOCKF_DEBUG
    516 	if (unlock->lf_type != F_UNLCK)
    517 		panic("lf_clearlock: bad type");
    518 	if (lockf_debug & 1)
    519 		lf_print("lf_clearlock", unlock);
    520 #endif /* LOCKF_DEBUG */
    521 	prev = head;
    522 	while ((ovcase = lf_findoverlap(lf, unlock, SELF,
    523 					&prev, &overlap)) != 0) {
    524 		/*
    525 		 * Wakeup the list of locks to be retried.
    526 		 */
    527 		lf_wakelock(overlap);
    528 
    529 		switch (ovcase) {
    530 
    531 		case 1: /* overlap == lock */
    532 			*prev = overlap->lf_next;
    533 			FREE(overlap, M_LOCKF);
    534 			break;
    535 
    536 		case 2: /* overlap contains lock: split it */
    537 			if (overlap->lf_start == unlock->lf_start) {
    538 				overlap->lf_start = unlock->lf_end + 1;
    539 				break;
    540 			}
    541 			lf_split(overlap, unlock, sparelock);
    542 			overlap->lf_next = unlock->lf_next;
    543 			break;
    544 
    545 		case 3: /* lock contains overlap */
    546 			*prev = overlap->lf_next;
    547 			lf = overlap->lf_next;
    548 			FREE(overlap, M_LOCKF);
    549 			continue;
    550 
    551 		case 4: /* overlap starts before lock */
    552 			overlap->lf_end = unlock->lf_start - 1;
    553 			prev = &overlap->lf_next;
    554 			lf = overlap->lf_next;
    555 			continue;
    556 
    557 		case 5: /* overlap ends after lock */
    558 			overlap->lf_start = unlock->lf_end + 1;
    559 			break;
    560 		}
    561 		break;
    562 	}
    563 #ifdef LOCKF_DEBUG
    564 	if (lockf_debug & 1)
    565 		lf_printlist("lf_clearlock", unlock);
    566 #endif /* LOCKF_DEBUG */
    567 	return 0;
    568 }
    569 
    570 /*
    571  * Check whether there is a blocking lock,
    572  * and if so return its process identifier.
    573  */
    574 static int
    575 lf_getlock(struct lockf *lock, struct flock *fl)
    576 {
    577 	struct lockf *block;
    578 
    579 #ifdef LOCKF_DEBUG
    580 	if (lockf_debug & 1)
    581 		lf_print("lf_getlock", lock);
    582 #endif /* LOCKF_DEBUG */
    583 
    584 	if ((block = lf_getblock(lock)) != NULL) {
    585 		fl->l_type = block->lf_type;
    586 		fl->l_whence = SEEK_SET;
    587 		fl->l_start = block->lf_start;
    588 		if (block->lf_end == -1)
    589 			fl->l_len = 0;
    590 		else
    591 			fl->l_len = block->lf_end - block->lf_start + 1;
    592 		if (block->lf_flags & F_POSIX)
    593 			fl->l_pid = ((struct proc *)block->lf_id)->p_pid;
    594 		else
    595 			fl->l_pid = -1;
    596 	} else {
    597 		fl->l_type = F_UNLCK;
    598 	}
    599 	return 0;
    600 }
    601 
    602 /*
    603  * Walk the list of locks for an inode and
    604  * return the first blocking lock.
    605  */
    606 static struct lockf *
    607 lf_getblock(struct lockf *lock)
    608 {
    609 	struct lockf **prev, *overlap, *lf = *(lock->lf_head);
    610 
    611 	prev = lock->lf_head;
    612 	while (lf_findoverlap(lf, lock, OTHERS, &prev, &overlap) != 0) {
    613 		/*
    614 		 * We've found an overlap, see if it blocks us
    615 		 */
    616 		if ((lock->lf_type == F_WRLCK || overlap->lf_type == F_WRLCK))
    617 			return overlap;
    618 		/*
    619 		 * Nope, point to the next one on the list and
    620 		 * see if it blocks us
    621 		 */
    622 		lf = overlap->lf_next;
    623 	}
    624 	return NOLOCKF;
    625 }
    626 
    627 /*
    628  * Walk the list of locks for an inode to
    629  * find an overlapping lock (if any).
    630  *
    631  * NOTE: this returns only the FIRST overlapping lock.  There
    632  *	 may be more than one.
    633  */
    634 static int
    635 lf_findoverlap(struct lockf *lf, struct lockf *lock, int type,
    636     struct lockf ***prev, struct lockf **overlap)
    637 {
    638 	off_t start, end;
    639 
    640 	*overlap = lf;
    641 	if (lf == NOLOCKF)
    642 		return 0;
    643 #ifdef LOCKF_DEBUG
    644 	if (lockf_debug & 2)
    645 		lf_print("lf_findoverlap: looking for overlap in", lock);
    646 #endif /* LOCKF_DEBUG */
    647 	start = lock->lf_start;
    648 	end = lock->lf_end;
    649 	while (lf != NOLOCKF) {
    650 		if (((type == SELF) && lf->lf_id != lock->lf_id) ||
    651 		    ((type == OTHERS) && lf->lf_id == lock->lf_id)) {
    652 			*prev = &lf->lf_next;
    653 			*overlap = lf = lf->lf_next;
    654 			continue;
    655 		}
    656 #ifdef LOCKF_DEBUG
    657 		if (lockf_debug & 2)
    658 			lf_print("\tchecking", lf);
    659 #endif /* LOCKF_DEBUG */
    660 		/*
    661 		 * OK, check for overlap
    662 		 *
    663 		 * Six cases:
    664 		 *	0) no overlap
    665 		 *	1) overlap == lock
    666 		 *	2) overlap contains lock
    667 		 *	3) lock contains overlap
    668 		 *	4) overlap starts before lock
    669 		 *	5) overlap ends after lock
    670 		 */
    671 		if ((lf->lf_end != -1 && start > lf->lf_end) ||
    672 		    (end != -1 && lf->lf_start > end)) {
    673 			/* Case 0 */
    674 #ifdef LOCKF_DEBUG
    675 			if (lockf_debug & 2)
    676 				printf("no overlap\n");
    677 #endif /* LOCKF_DEBUG */
    678 			if ((type & SELF) && end != -1 && lf->lf_start > end)
    679 				return 0;
    680 			*prev = &lf->lf_next;
    681 			*overlap = lf = lf->lf_next;
    682 			continue;
    683 		}
    684 		if ((lf->lf_start == start) && (lf->lf_end == end)) {
    685 			/* Case 1 */
    686 #ifdef LOCKF_DEBUG
    687 			if (lockf_debug & 2)
    688 				printf("overlap == lock\n");
    689 #endif /* LOCKF_DEBUG */
    690 			return 1;
    691 		}
    692 		if ((lf->lf_start <= start) &&
    693 		    (end != -1) &&
    694 		    ((lf->lf_end >= end) || (lf->lf_end == -1))) {
    695 			/* Case 2 */
    696 #ifdef LOCKF_DEBUG
    697 			if (lockf_debug & 2)
    698 				printf("overlap contains lock\n");
    699 #endif /* LOCKF_DEBUG */
    700 			return 2;
    701 		}
    702 		if (start <= lf->lf_start &&
    703 		           (end == -1 ||
    704 			   (lf->lf_end != -1 && end >= lf->lf_end))) {
    705 			/* Case 3 */
    706 #ifdef LOCKF_DEBUG
    707 			if (lockf_debug & 2)
    708 				printf("lock contains overlap\n");
    709 #endif /* LOCKF_DEBUG */
    710 			return 3;
    711 		}
    712 		if ((lf->lf_start < start) &&
    713 			((lf->lf_end >= start) || (lf->lf_end == -1))) {
    714 			/* Case 4 */
    715 #ifdef LOCKF_DEBUG
    716 			if (lockf_debug & 2)
    717 				printf("overlap starts before lock\n");
    718 #endif /* LOCKF_DEBUG */
    719 			return 4;
    720 		}
    721 		if ((lf->lf_start > start) &&
    722 			(end != -1) &&
    723 			((lf->lf_end > end) || (lf->lf_end == -1))) {
    724 			/* Case 5 */
    725 #ifdef LOCKF_DEBUG
    726 			if (lockf_debug & 2)
    727 				printf("overlap ends after lock\n");
    728 #endif /* LOCKF_DEBUG */
    729 			return 5;
    730 		}
    731 		panic("lf_findoverlap: default");
    732 	}
    733 	return 0;
    734 }
    735 
    736 /*
    737  * Split a lock and a contained region into
    738  * two or three locks as necessary.
    739  */
    740 static void
    741 lf_split(struct lockf *lock1, struct lockf *lock2, struct lockf **sparelock)
    742 {
    743 	struct lockf *splitlock;
    744 
    745 #ifdef LOCKF_DEBUG
    746 	if (lockf_debug & 2) {
    747 		lf_print("lf_split", lock1);
    748 		lf_print("splitting from", lock2);
    749 	}
    750 #endif /* LOCKF_DEBUG */
    751 	/*
    752 	 * Check to see if spliting into only two pieces.
    753 	 */
    754 	if (lock1->lf_start == lock2->lf_start) {
    755 		lock1->lf_start = lock2->lf_end + 1;
    756 		lock2->lf_next = lock1;
    757 		return;
    758 	}
    759 	if (lock1->lf_end == lock2->lf_end) {
    760 		lock1->lf_end = lock2->lf_start - 1;
    761 		lock2->lf_next = lock1->lf_next;
    762 		lock1->lf_next = lock2;
    763 		return;
    764 	}
    765 	/*
    766 	 * Make a new lock consisting of the last part of
    767 	 * the encompassing lock
    768 	 */
    769 	splitlock = *sparelock;
    770 	*sparelock = NULL;
    771 	memcpy(splitlock, lock1, sizeof(*splitlock));
    772 	splitlock->lf_start = lock2->lf_end + 1;
    773 	TAILQ_INIT(&splitlock->lf_blkhd);
    774 	lock1->lf_end = lock2->lf_start - 1;
    775 	/*
    776 	 * OK, now link it in
    777 	 */
    778 	splitlock->lf_next = lock1->lf_next;
    779 	lock2->lf_next = splitlock;
    780 	lock1->lf_next = lock2;
    781 }
    782 
    783 /*
    784  * Wakeup a blocklist
    785  */
    786 static void
    787 lf_wakelock(struct lockf *listhead)
    788 {
    789 	struct lockf *wakelock;
    790 
    791 	while ((wakelock = TAILQ_FIRST(&listhead->lf_blkhd))) {
    792 		KASSERT(wakelock->lf_next == listhead);
    793 		TAILQ_REMOVE(&listhead->lf_blkhd, wakelock, lf_block);
    794 		wakelock->lf_next = NOLOCKF;
    795 #ifdef LOCKF_DEBUG
    796 		if (lockf_debug & 2)
    797 			lf_print("lf_wakelock: awakening", wakelock);
    798 #endif
    799 		wakeup(wakelock);
    800 	}
    801 }
    802 
    803 #ifdef LOCKF_DEBUG
    804 /*
    805  * Print out a lock.
    806  */
    807 static void
    808 lf_print(char *tag, struct lockf *lock)
    809 {
    810 
    811 	printf("%s: lock %p for ", tag, lock);
    812 	if (lock->lf_flags & F_POSIX)
    813 		printf("proc %d", ((struct proc *)lock->lf_id)->p_pid);
    814 	else
    815 		printf("file 0x%p", (struct file *)lock->lf_id);
    816 	printf(" %s, start %qx, end %qx",
    817 		lock->lf_type == F_RDLCK ? "shared" :
    818 		lock->lf_type == F_WRLCK ? "exclusive" :
    819 		lock->lf_type == F_UNLCK ? "unlock" :
    820 		"unknown", lock->lf_start, lock->lf_end);
    821 	if (TAILQ_FIRST(&lock->lf_blkhd))
    822 		printf(" block %p\n", TAILQ_FIRST(&lock->lf_blkhd));
    823 	else
    824 		printf("\n");
    825 }
    826 
    827 static void
    828 lf_printlist(char *tag, struct lockf *lock)
    829 {
    830 	struct lockf *lf, *blk;
    831 
    832 	printf("%s: Lock list:\n", tag);
    833 	for (lf = *lock->lf_head; lf; lf = lf->lf_next) {
    834 		printf("\tlock %p for ", lf);
    835 		if (lf->lf_flags & F_POSIX)
    836 			printf("proc %d", ((struct proc *)lf->lf_id)->p_pid);
    837 		else
    838 			printf("file 0x%p", (struct file *)lf->lf_id);
    839 		printf(", %s, start %qx, end %qx",
    840 			lf->lf_type == F_RDLCK ? "shared" :
    841 			lf->lf_type == F_WRLCK ? "exclusive" :
    842 			lf->lf_type == F_UNLCK ? "unlock" :
    843 			"unknown", lf->lf_start, lf->lf_end);
    844 		TAILQ_FOREACH(blk, &lf->lf_blkhd, lf_block) {
    845 			if (blk->lf_flags & F_POSIX)
    846 				printf("proc %d",
    847 				    ((struct proc *)blk->lf_id)->p_pid);
    848 			else
    849 				printf("file 0x%p", (struct file *)blk->lf_id);
    850 			printf(", %s, start %qx, end %qx",
    851 				blk->lf_type == F_RDLCK ? "shared" :
    852 				blk->lf_type == F_WRLCK ? "exclusive" :
    853 				blk->lf_type == F_UNLCK ? "unlock" :
    854 				"unknown", blk->lf_start, blk->lf_end);
    855 			if (TAILQ_FIRST(&blk->lf_blkhd))
    856 				 panic("lf_printlist: bad list");
    857 		}
    858 		printf("\n");
    859 	}
    860 }
    861 #endif /* LOCKF_DEBUG */
    862