vfs_lockf.c revision 1.50 1 /* $NetBSD: vfs_lockf.c,v 1.50 2006/05/20 12:02:47 yamt 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 1.50 2006/05/20 12:02:47 yamt 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 #include <sys/kauth.h>
50
51 /*
52 * The lockf structure is a kernel structure which contains the information
53 * associated with a byte range lock. The lockf structures are linked into
54 * the inode structure. Locks are sorted by the starting byte of the lock for
55 * efficiency.
56 *
57 * lf_next is used for two purposes, depending on whether the lock is
58 * being held, or is in conflict with an existing lock. If this lock
59 * is held, it indicates the next lock on the same vnode.
60 * For pending locks, if lock->lf_next is non-NULL, then lock->lf_block
61 * must be queued on the lf_blkhd TAILQ of lock->lf_next.
62 */
63
64 TAILQ_HEAD(locklist, lockf);
65
66 struct lockf {
67 short lf_flags; /* Lock semantics: F_POSIX, F_FLOCK, F_WAIT */
68 short lf_type; /* Lock type: F_RDLCK, F_WRLCK */
69 off_t lf_start; /* The byte # of the start of the lock */
70 off_t lf_end; /* The byte # of the end of the lock (-1=EOF)*/
71 void *lf_id; /* process or file description holding lock */
72 struct lwp *lf_lwp; /* LWP waiting for lock */
73 struct lockf **lf_head; /* Back pointer to the head of lockf list */
74 struct lockf *lf_next; /* Next lock on this vnode, or blocking lock */
75 struct locklist lf_blkhd; /* List of requests blocked on this lock */
76 TAILQ_ENTRY(lockf) lf_block;/* A request waiting for a lock */
77 uid_t lf_uid; /* User ID responsible */
78 };
79
80 /* Maximum length of sleep chains to traverse to try and detect deadlock. */
81 #define MAXDEPTH 50
82
83 POOL_INIT(lockfpool, sizeof(struct lockf), 0, 0, 0, "lockfpl",
84 &pool_allocator_nointr);
85
86 /*
87 * This variable controls the maximum number of processes that will
88 * be checked in doing deadlock detection.
89 */
90 int maxlockdepth = MAXDEPTH;
91
92 #ifdef LOCKF_DEBUG
93 int lockf_debug = 0;
94 #endif
95
96 #define NOLOCKF (struct lockf *)0
97 #define SELF 0x1
98 #define OTHERS 0x2
99
100 /*
101 * XXX TODO
102 * Misc cleanups: "caddr_t id" should be visible in the API as a
103 * "struct proc *".
104 * (This requires rototilling all VFS's which support advisory locking).
105 */
106
107 /*
108 * If there's a lot of lock contention on a single vnode, locking
109 * schemes which allow for more paralleism would be needed. Given how
110 * infrequently byte-range locks are actually used in typical BSD
111 * code, a more complex approach probably isn't worth it.
112 */
113
114 /*
115 * We enforce a limit on locks by uid, so that a single user cannot
116 * run the kernel out of memory. For now, the limit is pretty coarse.
117 * There is no limit on root.
118 *
119 * Splitting a lock will always succeed, regardless of current allocations.
120 * If you're slightly above the limit, we still have to permit an allocation
121 * so that the unlock can succeed. If the unlocking causes too many splits,
122 * however, you're totally cutoff.
123 */
124 int maxlocksperuid = 1024;
125
126 #ifdef LOCKF_DEBUG
127 /*
128 * Print out a lock.
129 */
130 static void
131 lf_print(char *tag, struct lockf *lock)
132 {
133
134 printf("%s: lock %p for ", tag, lock);
135 if (lock->lf_flags & F_POSIX)
136 printf("proc %d", ((struct proc *)lock->lf_id)->p_pid);
137 else
138 printf("file %p", (struct file *)lock->lf_id);
139 printf(" %s, start %qx, end %qx",
140 lock->lf_type == F_RDLCK ? "shared" :
141 lock->lf_type == F_WRLCK ? "exclusive" :
142 lock->lf_type == F_UNLCK ? "unlock" :
143 "unknown", lock->lf_start, lock->lf_end);
144 if (TAILQ_FIRST(&lock->lf_blkhd))
145 printf(" block %p\n", TAILQ_FIRST(&lock->lf_blkhd));
146 else
147 printf("\n");
148 }
149
150 static void
151 lf_printlist(char *tag, struct lockf *lock)
152 {
153 struct lockf *lf, *blk;
154
155 printf("%s: Lock list:\n", tag);
156 for (lf = *lock->lf_head; lf; lf = lf->lf_next) {
157 printf("\tlock %p for ", lf);
158 if (lf->lf_flags & F_POSIX)
159 printf("proc %d", ((struct proc *)lf->lf_id)->p_pid);
160 else
161 printf("file %p", (struct file *)lf->lf_id);
162 printf(", %s, start %qx, end %qx",
163 lf->lf_type == F_RDLCK ? "shared" :
164 lf->lf_type == F_WRLCK ? "exclusive" :
165 lf->lf_type == F_UNLCK ? "unlock" :
166 "unknown", lf->lf_start, lf->lf_end);
167 TAILQ_FOREACH(blk, &lf->lf_blkhd, lf_block) {
168 if (blk->lf_flags & F_POSIX)
169 printf("proc %d",
170 ((struct proc *)blk->lf_id)->p_pid);
171 else
172 printf("file %p", (struct file *)blk->lf_id);
173 printf(", %s, start %qx, end %qx",
174 blk->lf_type == F_RDLCK ? "shared" :
175 blk->lf_type == F_WRLCK ? "exclusive" :
176 blk->lf_type == F_UNLCK ? "unlock" :
177 "unknown", blk->lf_start, blk->lf_end);
178 if (TAILQ_FIRST(&blk->lf_blkhd))
179 panic("lf_printlist: bad list");
180 }
181 printf("\n");
182 }
183 }
184 #endif /* LOCKF_DEBUG */
185
186 /*
187 * 3 options for allowfail.
188 * 0 - always allocate. 1 - cutoff at limit. 2 - cutoff at double limit.
189 */
190 static struct lockf *
191 lf_alloc(uid_t uid, int allowfail)
192 {
193 struct uidinfo *uip;
194 struct lockf *lock;
195 int s;
196
197 uip = uid_find(uid);
198 UILOCK(uip, s);
199 if (uid && allowfail && uip->ui_lockcnt >
200 (allowfail == 1 ? maxlocksperuid : (maxlocksperuid * 2))) {
201 UIUNLOCK(uip, s);
202 return NULL;
203 }
204 uip->ui_lockcnt++;
205 UIUNLOCK(uip, s);
206 lock = pool_get(&lockfpool, PR_WAITOK);
207 lock->lf_uid = uid;
208 return lock;
209 }
210
211 static void
212 lf_free(struct lockf *lock)
213 {
214 struct uidinfo *uip;
215 int s;
216
217 uip = uid_find(lock->lf_uid);
218 UILOCK(uip, s);
219 uip->ui_lockcnt--;
220 UIUNLOCK(uip, s);
221 pool_put(&lockfpool, lock);
222 }
223
224 /*
225 * Walk the list of locks for an inode to
226 * find an overlapping lock (if any).
227 *
228 * NOTE: this returns only the FIRST overlapping lock. There
229 * may be more than one.
230 */
231 static int
232 lf_findoverlap(struct lockf *lf, struct lockf *lock, int type,
233 struct lockf ***prev, struct lockf **overlap)
234 {
235 off_t start, end;
236
237 *overlap = lf;
238 if (lf == NOLOCKF)
239 return 0;
240 #ifdef LOCKF_DEBUG
241 if (lockf_debug & 2)
242 lf_print("lf_findoverlap: looking for overlap in", lock);
243 #endif /* LOCKF_DEBUG */
244 start = lock->lf_start;
245 end = lock->lf_end;
246 while (lf != NOLOCKF) {
247 if (((type == SELF) && lf->lf_id != lock->lf_id) ||
248 ((type == OTHERS) && lf->lf_id == lock->lf_id)) {
249 *prev = &lf->lf_next;
250 *overlap = lf = lf->lf_next;
251 continue;
252 }
253 #ifdef LOCKF_DEBUG
254 if (lockf_debug & 2)
255 lf_print("\tchecking", lf);
256 #endif /* LOCKF_DEBUG */
257 /*
258 * OK, check for overlap
259 *
260 * Six cases:
261 * 0) no overlap
262 * 1) overlap == lock
263 * 2) overlap contains lock
264 * 3) lock contains overlap
265 * 4) overlap starts before lock
266 * 5) overlap ends after lock
267 */
268 if ((lf->lf_end != -1 && start > lf->lf_end) ||
269 (end != -1 && lf->lf_start > end)) {
270 /* Case 0 */
271 #ifdef LOCKF_DEBUG
272 if (lockf_debug & 2)
273 printf("no overlap\n");
274 #endif /* LOCKF_DEBUG */
275 if ((type & SELF) && end != -1 && lf->lf_start > end)
276 return 0;
277 *prev = &lf->lf_next;
278 *overlap = lf = lf->lf_next;
279 continue;
280 }
281 if ((lf->lf_start == start) && (lf->lf_end == end)) {
282 /* Case 1 */
283 #ifdef LOCKF_DEBUG
284 if (lockf_debug & 2)
285 printf("overlap == lock\n");
286 #endif /* LOCKF_DEBUG */
287 return 1;
288 }
289 if ((lf->lf_start <= start) &&
290 (end != -1) &&
291 ((lf->lf_end >= end) || (lf->lf_end == -1))) {
292 /* Case 2 */
293 #ifdef LOCKF_DEBUG
294 if (lockf_debug & 2)
295 printf("overlap contains lock\n");
296 #endif /* LOCKF_DEBUG */
297 return 2;
298 }
299 if (start <= lf->lf_start &&
300 (end == -1 ||
301 (lf->lf_end != -1 && end >= lf->lf_end))) {
302 /* Case 3 */
303 #ifdef LOCKF_DEBUG
304 if (lockf_debug & 2)
305 printf("lock contains overlap\n");
306 #endif /* LOCKF_DEBUG */
307 return 3;
308 }
309 if ((lf->lf_start < start) &&
310 ((lf->lf_end >= start) || (lf->lf_end == -1))) {
311 /* Case 4 */
312 #ifdef LOCKF_DEBUG
313 if (lockf_debug & 2)
314 printf("overlap starts before lock\n");
315 #endif /* LOCKF_DEBUG */
316 return 4;
317 }
318 if ((lf->lf_start > start) &&
319 (end != -1) &&
320 ((lf->lf_end > end) || (lf->lf_end == -1))) {
321 /* Case 5 */
322 #ifdef LOCKF_DEBUG
323 if (lockf_debug & 2)
324 printf("overlap ends after lock\n");
325 #endif /* LOCKF_DEBUG */
326 return 5;
327 }
328 panic("lf_findoverlap: default");
329 }
330 return 0;
331 }
332
333 /*
334 * Split a lock and a contained region into
335 * two or three locks as necessary.
336 */
337 static void
338 lf_split(struct lockf *lock1, struct lockf *lock2, struct lockf **sparelock)
339 {
340 struct lockf *splitlock;
341
342 #ifdef LOCKF_DEBUG
343 if (lockf_debug & 2) {
344 lf_print("lf_split", lock1);
345 lf_print("splitting from", lock2);
346 }
347 #endif /* LOCKF_DEBUG */
348 /*
349 * Check to see if spliting into only two pieces.
350 */
351 if (lock1->lf_start == lock2->lf_start) {
352 lock1->lf_start = lock2->lf_end + 1;
353 lock2->lf_next = lock1;
354 return;
355 }
356 if (lock1->lf_end == lock2->lf_end) {
357 lock1->lf_end = lock2->lf_start - 1;
358 lock2->lf_next = lock1->lf_next;
359 lock1->lf_next = lock2;
360 return;
361 }
362 /*
363 * Make a new lock consisting of the last part of
364 * the encompassing lock
365 */
366 splitlock = *sparelock;
367 *sparelock = NULL;
368 memcpy(splitlock, lock1, sizeof(*splitlock));
369 splitlock->lf_start = lock2->lf_end + 1;
370 TAILQ_INIT(&splitlock->lf_blkhd);
371 lock1->lf_end = lock2->lf_start - 1;
372 /*
373 * OK, now link it in
374 */
375 splitlock->lf_next = lock1->lf_next;
376 lock2->lf_next = splitlock;
377 lock1->lf_next = lock2;
378 }
379
380 /*
381 * Wakeup a blocklist
382 */
383 static void
384 lf_wakelock(struct lockf *listhead)
385 {
386 struct lockf *wakelock;
387
388 while ((wakelock = TAILQ_FIRST(&listhead->lf_blkhd))) {
389 KASSERT(wakelock->lf_next == listhead);
390 TAILQ_REMOVE(&listhead->lf_blkhd, wakelock, lf_block);
391 wakelock->lf_next = NOLOCKF;
392 #ifdef LOCKF_DEBUG
393 if (lockf_debug & 2)
394 lf_print("lf_wakelock: awakening", wakelock);
395 #endif
396 wakeup(wakelock);
397 }
398 }
399
400 /*
401 * Remove a byte-range lock on an inode.
402 *
403 * Generally, find the lock (or an overlap to that lock)
404 * and remove it (or shrink it), then wakeup anyone we can.
405 */
406 static int
407 lf_clearlock(struct lockf *unlock, struct lockf **sparelock)
408 {
409 struct lockf **head = unlock->lf_head;
410 struct lockf *lf = *head;
411 struct lockf *overlap, **prev;
412 int ovcase;
413
414 if (lf == NOLOCKF)
415 return 0;
416 #ifdef LOCKF_DEBUG
417 if (unlock->lf_type != F_UNLCK)
418 panic("lf_clearlock: bad type");
419 if (lockf_debug & 1)
420 lf_print("lf_clearlock", unlock);
421 #endif /* LOCKF_DEBUG */
422 prev = head;
423 while ((ovcase = lf_findoverlap(lf, unlock, SELF,
424 &prev, &overlap)) != 0) {
425 /*
426 * Wakeup the list of locks to be retried.
427 */
428 lf_wakelock(overlap);
429
430 switch (ovcase) {
431
432 case 1: /* overlap == lock */
433 *prev = overlap->lf_next;
434 lf_free(overlap);
435 break;
436
437 case 2: /* overlap contains lock: split it */
438 if (overlap->lf_start == unlock->lf_start) {
439 overlap->lf_start = unlock->lf_end + 1;
440 break;
441 }
442 lf_split(overlap, unlock, sparelock);
443 overlap->lf_next = unlock->lf_next;
444 break;
445
446 case 3: /* lock contains overlap */
447 *prev = overlap->lf_next;
448 lf = overlap->lf_next;
449 lf_free(overlap);
450 continue;
451
452 case 4: /* overlap starts before lock */
453 overlap->lf_end = unlock->lf_start - 1;
454 prev = &overlap->lf_next;
455 lf = overlap->lf_next;
456 continue;
457
458 case 5: /* overlap ends after lock */
459 overlap->lf_start = unlock->lf_end + 1;
460 break;
461 }
462 break;
463 }
464 #ifdef LOCKF_DEBUG
465 if (lockf_debug & 1)
466 lf_printlist("lf_clearlock", unlock);
467 #endif /* LOCKF_DEBUG */
468 return 0;
469 }
470
471 /*
472 * Walk the list of locks for an inode and
473 * return the first blocking lock.
474 */
475 static struct lockf *
476 lf_getblock(struct lockf *lock)
477 {
478 struct lockf **prev, *overlap, *lf = *(lock->lf_head);
479
480 prev = lock->lf_head;
481 while (lf_findoverlap(lf, lock, OTHERS, &prev, &overlap) != 0) {
482 /*
483 * We've found an overlap, see if it blocks us
484 */
485 if ((lock->lf_type == F_WRLCK || overlap->lf_type == F_WRLCK))
486 return overlap;
487 /*
488 * Nope, point to the next one on the list and
489 * see if it blocks us
490 */
491 lf = overlap->lf_next;
492 }
493 return NOLOCKF;
494 }
495
496 /*
497 * Set a byte-range lock.
498 */
499 static int
500 lf_setlock(struct lockf *lock, struct lockf **sparelock,
501 struct simplelock *interlock)
502 {
503 struct lockf *block;
504 struct lockf **head = lock->lf_head;
505 struct lockf **prev, *overlap, *ltmp;
506 static char lockstr[] = "lockf";
507 int ovcase, priority, 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 * Set the priority
516 */
517 priority = PLOCK;
518 if (lock->lf_type == F_WRLCK)
519 priority += 4;
520 priority |= PCATCH;
521 /*
522 * Scan lock list for this file looking for locks that would block us.
523 */
524 while ((block = lf_getblock(lock)) != NULL) {
525 /*
526 * Free the structure and return if nonblocking.
527 */
528 if ((lock->lf_flags & F_WAIT) == 0) {
529 lf_free(lock);
530 return EAGAIN;
531 }
532 /*
533 * We are blocked. Since flock style locks cover
534 * the whole file, there is no chance for deadlock.
535 * For byte-range locks we must check for deadlock.
536 *
537 * Deadlock detection is done by looking through the
538 * wait channels to see if there are any cycles that
539 * involve us. MAXDEPTH is set just to make sure we
540 * do not go off into neverneverland.
541 */
542 if ((lock->lf_flags & F_POSIX) &&
543 (block->lf_flags & F_POSIX)) {
544 struct lwp *wlwp;
545 volatile const struct lockf *waitblock;
546 int i = 0;
547
548 /*
549 * The block is waiting on something. if_lwp will be
550 * 0 once the lock is granted, so we terminate the
551 * loop if we find this.
552 */
553 wlwp = block->lf_lwp;
554 while (wlwp && (i++ < maxlockdepth)) {
555 waitblock = wlwp->l_wchan;
556 /* Get the owner of the blocking lock */
557 waitblock = waitblock->lf_next;
558 if ((waitblock->lf_flags & F_POSIX) == 0)
559 break;
560 wlwp = waitblock->lf_lwp;
561 if (wlwp == lock->lf_lwp) {
562 lf_free(lock);
563 return EDEADLK;
564 }
565 }
566 /*
567 * If we're still following a dependency chain
568 * after maxlockdepth iterations, assume we're in
569 * a cycle to be safe.
570 */
571 if (i >= maxlockdepth) {
572 lf_free(lock);
573 return EDEADLK;
574 }
575 }
576 /*
577 * For flock type locks, we must first remove
578 * any shared locks that we hold before we sleep
579 * waiting for an exclusive lock.
580 */
581 if ((lock->lf_flags & F_FLOCK) &&
582 lock->lf_type == F_WRLCK) {
583 lock->lf_type = F_UNLCK;
584 (void) lf_clearlock(lock, NULL);
585 lock->lf_type = F_WRLCK;
586 }
587 /*
588 * Add our lock to the blocked list and sleep until we're free.
589 * Remember who blocked us (for deadlock detection).
590 */
591 lock->lf_next = block;
592 TAILQ_INSERT_TAIL(&block->lf_blkhd, lock, lf_block);
593 #ifdef LOCKF_DEBUG
594 if (lockf_debug & 1) {
595 lf_print("lf_setlock: blocking on", block);
596 lf_printlist("lf_setlock", block);
597 }
598 #endif /* LOCKF_DEBUG */
599 error = ltsleep(lock, priority, lockstr, 0, interlock);
600
601 /*
602 * We may have been awakened by a signal (in
603 * which case we must remove ourselves from the
604 * blocked list) and/or by another process
605 * releasing a lock (in which case we have already
606 * been removed from the blocked list and our
607 * lf_next field set to NOLOCKF).
608 */
609 if (lock->lf_next != NOLOCKF) {
610 TAILQ_REMOVE(&lock->lf_next->lf_blkhd, lock, lf_block);
611 lock->lf_next = NOLOCKF;
612 }
613 if (error) {
614 lf_free(lock);
615 return error;
616 }
617 }
618 /*
619 * No blocks!! Add the lock. Note that we will
620 * downgrade or upgrade any overlapping locks this
621 * process already owns.
622 *
623 * Skip over locks owned by other processes.
624 * Handle any locks that overlap and are owned by ourselves.
625 */
626 lock->lf_lwp = 0;
627 prev = head;
628 block = *head;
629 needtolink = 1;
630 for (;;) {
631 ovcase = lf_findoverlap(block, lock, SELF, &prev, &overlap);
632 if (ovcase)
633 block = overlap->lf_next;
634 /*
635 * Six cases:
636 * 0) no overlap
637 * 1) overlap == lock
638 * 2) overlap contains lock
639 * 3) lock contains overlap
640 * 4) overlap starts before lock
641 * 5) overlap ends after lock
642 */
643 switch (ovcase) {
644 case 0: /* no overlap */
645 if (needtolink) {
646 *prev = lock;
647 lock->lf_next = overlap;
648 }
649 break;
650
651 case 1: /* overlap == lock */
652 /*
653 * If downgrading lock, others may be
654 * able to acquire it.
655 */
656 if (lock->lf_type == F_RDLCK &&
657 overlap->lf_type == F_WRLCK)
658 lf_wakelock(overlap);
659 overlap->lf_type = lock->lf_type;
660 lf_free(lock);
661 lock = overlap; /* for debug output below */
662 break;
663
664 case 2: /* overlap contains lock */
665 /*
666 * Check for common starting point and different types.
667 */
668 if (overlap->lf_type == lock->lf_type) {
669 lf_free(lock);
670 lock = overlap; /* for debug output below */
671 break;
672 }
673 if (overlap->lf_start == lock->lf_start) {
674 *prev = lock;
675 lock->lf_next = overlap;
676 overlap->lf_start = lock->lf_end + 1;
677 } else
678 lf_split(overlap, lock, sparelock);
679 lf_wakelock(overlap);
680 break;
681
682 case 3: /* lock contains overlap */
683 /*
684 * If downgrading lock, others may be able to
685 * acquire it, otherwise take the list.
686 */
687 if (lock->lf_type == F_RDLCK &&
688 overlap->lf_type == F_WRLCK) {
689 lf_wakelock(overlap);
690 } else {
691 while ((ltmp = TAILQ_FIRST(&overlap->lf_blkhd))) {
692 KASSERT(ltmp->lf_next == overlap);
693 TAILQ_REMOVE(&overlap->lf_blkhd, ltmp,
694 lf_block);
695 ltmp->lf_next = lock;
696 TAILQ_INSERT_TAIL(&lock->lf_blkhd,
697 ltmp, lf_block);
698 }
699 }
700 /*
701 * Add the new lock if necessary and delete the overlap.
702 */
703 if (needtolink) {
704 *prev = lock;
705 lock->lf_next = overlap->lf_next;
706 prev = &lock->lf_next;
707 needtolink = 0;
708 } else
709 *prev = overlap->lf_next;
710 lf_free(overlap);
711 continue;
712
713 case 4: /* overlap starts before lock */
714 /*
715 * Add lock after overlap on the list.
716 */
717 lock->lf_next = overlap->lf_next;
718 overlap->lf_next = lock;
719 overlap->lf_end = lock->lf_start - 1;
720 prev = &lock->lf_next;
721 lf_wakelock(overlap);
722 needtolink = 0;
723 continue;
724
725 case 5: /* overlap ends after lock */
726 /*
727 * Add the new lock before overlap.
728 */
729 if (needtolink) {
730 *prev = lock;
731 lock->lf_next = overlap;
732 }
733 overlap->lf_start = lock->lf_end + 1;
734 lf_wakelock(overlap);
735 break;
736 }
737 break;
738 }
739 #ifdef LOCKF_DEBUG
740 if (lockf_debug & 1) {
741 lf_print("lf_setlock: got the lock", lock);
742 lf_printlist("lf_setlock", lock);
743 }
744 #endif /* LOCKF_DEBUG */
745 return 0;
746 }
747
748 /*
749 * Check whether there is a blocking lock,
750 * and if so return its process identifier.
751 */
752 static int
753 lf_getlock(struct lockf *lock, struct flock *fl)
754 {
755 struct lockf *block;
756
757 #ifdef LOCKF_DEBUG
758 if (lockf_debug & 1)
759 lf_print("lf_getlock", lock);
760 #endif /* LOCKF_DEBUG */
761
762 if ((block = lf_getblock(lock)) != NULL) {
763 fl->l_type = block->lf_type;
764 fl->l_whence = SEEK_SET;
765 fl->l_start = block->lf_start;
766 if (block->lf_end == -1)
767 fl->l_len = 0;
768 else
769 fl->l_len = block->lf_end - block->lf_start + 1;
770 if (block->lf_flags & F_POSIX)
771 fl->l_pid = ((struct proc *)block->lf_id)->p_pid;
772 else
773 fl->l_pid = -1;
774 } else {
775 fl->l_type = F_UNLCK;
776 }
777 return 0;
778 }
779
780 /*
781 * Do an advisory lock operation.
782 */
783 int
784 lf_advlock(struct vop_advlock_args *ap, struct lockf **head, off_t size)
785 {
786 struct proc *p = curproc;
787 struct flock *fl = ap->a_fl;
788 struct lockf *lock = NULL;
789 struct lockf *sparelock;
790 struct simplelock *interlock = &ap->a_vp->v_interlock;
791 off_t start, end;
792 int error = 0;
793
794 /*
795 * Convert the flock structure into a start and end.
796 */
797 switch (fl->l_whence) {
798 case SEEK_SET:
799 case SEEK_CUR:
800 /*
801 * Caller is responsible for adding any necessary offset
802 * when SEEK_CUR is used.
803 */
804 start = fl->l_start;
805 break;
806
807 case SEEK_END:
808 start = size + fl->l_start;
809 break;
810
811 default:
812 return EINVAL;
813 }
814 if (start < 0)
815 return EINVAL;
816
817 /*
818 * allocate locks before acquire simple lock.
819 * we need two locks in the worst case.
820 */
821 switch (ap->a_op) {
822 case F_SETLK:
823 case F_UNLCK:
824 /*
825 * XXX for F_UNLCK case, we can re-use lock.
826 */
827 if ((ap->a_flags & F_FLOCK) == 0) {
828 /*
829 * byte-range lock might need one more lock.
830 */
831 sparelock = lf_alloc(kauth_cred_geteuid(p->p_cred), 0);
832 if (sparelock == NULL) {
833 error = ENOMEM;
834 goto quit;
835 }
836 break;
837 }
838 /* FALLTHROUGH */
839
840 case F_GETLK:
841 sparelock = NULL;
842 break;
843
844 default:
845 return EINVAL;
846 }
847
848 lock = lf_alloc(kauth_cred_geteuid(p->p_cred), ap->a_op != F_UNLCK ? 1 : 2);
849 if (lock == NULL) {
850 error = ENOMEM;
851 goto quit;
852 }
853
854 simple_lock(interlock);
855
856 /*
857 * Avoid the common case of unlocking when inode has no locks.
858 */
859 if (*head == (struct lockf *)0) {
860 if (ap->a_op != F_SETLK) {
861 fl->l_type = F_UNLCK;
862 error = 0;
863 goto quit_unlock;
864 }
865 }
866
867 if (fl->l_len == 0)
868 end = -1;
869 else
870 end = start + fl->l_len - 1;
871 /*
872 * Create the lockf structure.
873 */
874 lock->lf_start = start;
875 lock->lf_end = end;
876 /* XXX NJWLWP
877 * I don't want to make the entire VFS universe use LWPs, because
878 * they don't need them, for the most part. This is an exception,
879 * and a kluge.
880 */
881
882 lock->lf_head = head;
883 lock->lf_type = fl->l_type;
884 lock->lf_next = (struct lockf *)0;
885 TAILQ_INIT(&lock->lf_blkhd);
886 lock->lf_flags = ap->a_flags;
887 if (lock->lf_flags & F_POSIX) {
888 KASSERT(curproc == (struct proc *)ap->a_id);
889 }
890 lock->lf_id = (struct proc *)ap->a_id;
891 lock->lf_lwp = curlwp;
892
893 /*
894 * Do the requested operation.
895 */
896 switch (ap->a_op) {
897
898 case F_SETLK:
899 error = lf_setlock(lock, &sparelock, interlock);
900 lock = NULL; /* lf_setlock freed it */
901 break;
902
903 case F_UNLCK:
904 error = lf_clearlock(lock, &sparelock);
905 break;
906
907 case F_GETLK:
908 error = lf_getlock(lock, fl);
909 break;
910
911 default:
912 break;
913 /* NOTREACHED */
914 }
915
916 quit_unlock:
917 simple_unlock(interlock);
918 quit:
919 if (lock)
920 lf_free(lock);
921 if (sparelock)
922 lf_free(sparelock);
923
924 return error;
925 }
926