]>
git.saurik.com Git - apple/xnu.git/blob - bsd/ufs/ufs/ufs_lockf.c
2 * Copyright (c) 2000 Apple Computer, Inc. All rights reserved.
4 * @APPLE_LICENSE_HEADER_START@
6 * The contents of this file constitute Original Code as defined in and
7 * are subject to the Apple Public Source License Version 1.1 (the
8 * "License"). You may not use this file except in compliance with the
9 * License. Please obtain a copy of the License at
10 * http://www.apple.com/publicsource and read it before using this file.
12 * This Original Code and all software distributed under the License are
13 * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
14 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
15 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
16 * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT. Please see the
17 * License for the specific language governing rights and limitations
20 * @APPLE_LICENSE_HEADER_END@
22 /* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
24 * Copyright (c) 1982, 1986, 1989, 1993
25 * The Regents of the University of California. All rights reserved.
27 * This code is derived from software contributed to Berkeley by
28 * Scooter Morris at Genentech Inc.
30 * Redistribution and use in source and binary forms, with or without
31 * modification, are permitted provided that the following conditions
33 * 1. Redistributions of source code must retain the above copyright
34 * notice, this list of conditions and the following disclaimer.
35 * 2. Redistributions in binary form must reproduce the above copyright
36 * notice, this list of conditions and the following disclaimer in the
37 * documentation and/or other materials provided with the distribution.
38 * 3. All advertising materials mentioning features or use of this software
39 * must display the following acknowledgement:
40 * This product includes software developed by the University of
41 * California, Berkeley and its contributors.
42 * 4. Neither the name of the University nor the names of its contributors
43 * may be used to endorse or promote products derived from this software
44 * without specific prior written permission.
46 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
47 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
48 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
49 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
50 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
51 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
52 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
53 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
54 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
55 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58 * @(#)ufs_lockf.c 8.4 (Berkeley) 10/26/94
61 #include <sys/param.h>
62 #include <sys/systm.h>
63 #include <sys/kernel.h>
66 #include <sys/vnode.h>
67 #include <sys/malloc.h>
68 #include <sys/fcntl.h>
69 #include <sys/quota.h>
71 #include <ufs/ufs/lockf.h>
72 #include <ufs/ufs/quota.h>
73 #include <ufs/ufs/inode.h>
74 #include <ufs/ufs/ufs_extern.h>
77 * This variable controls the maximum number of processes that will
78 * be checked in doing deadlock detection.
80 int maxlockdepth
= MAXDEPTH
;
84 #include <sys/sysctl.h>
86 struct ctldebug debug4
= { "lockf_debug", &lockf_debug
};
89 #define NOLOCKF (struct lockf *)0
94 * Set a byte-range lock.
98 register struct lockf
*lock
;
100 register struct lockf
*block
;
101 struct inode
*ip
= lock
->lf_inode
;
102 struct lockf
**prev
, *overlap
, *ltmp
;
103 static char lockstr
[] = "lockf";
104 int ovcase
, priority
, needtolink
, error
;
108 lf_print("lf_setlock", lock
);
109 #endif /* LOCKF_DEBUG */
115 if (lock
->lf_type
== F_WRLCK
)
119 * Scan lock list for this file looking for locks that would block us.
121 while (block
= lf_getblock(lock
)) {
123 * Free the structure and return if nonblocking.
125 if ((lock
->lf_flags
& F_WAIT
) == 0) {
130 * We are blocked. Since flock style locks cover
131 * the whole file, there is no chance for deadlock.
132 * For byte-range locks we must check for deadlock.
134 * Deadlock detection is done by looking through the
135 * wait channels to see if there are any cycles that
136 * involve us. MAXDEPTH is set just to make sure we
137 * do not go off into neverland.
139 if ((lock
->lf_flags
& F_POSIX
) &&
140 (block
->lf_flags
& F_POSIX
)) {
141 register struct proc
*wproc
;
142 register struct lockf
*waitblock
;
145 /* The block is waiting on something */
146 wproc
= (struct proc
*)block
->lf_id
;
147 while (wproc
->p_wchan
&&
148 (wproc
->p_wmesg
== lockstr
) &&
149 (i
++ < maxlockdepth
)) {
150 waitblock
= (struct lockf
*)wproc
->p_wchan
;
151 /* Get the owner of the blocking lock */
152 waitblock
= waitblock
->lf_next
;
153 if ((waitblock
->lf_flags
& F_POSIX
) == 0)
155 wproc
= (struct proc
*)waitblock
->lf_id
;
156 if (wproc
== (struct proc
*)lock
->lf_id
) {
157 _FREE(lock
, M_LOCKF
);
163 * For flock type locks, we must first remove
164 * any shared locks that we hold before we sleep
165 * waiting for an exclusive lock.
167 if ((lock
->lf_flags
& F_FLOCK
) &&
168 lock
->lf_type
== F_WRLCK
) {
169 lock
->lf_type
= F_UNLCK
;
170 (void) lf_clearlock(lock
);
171 lock
->lf_type
= F_WRLCK
;
174 * Add our lock to the blocked list and sleep until we're free.
175 * Remember who blocked us (for deadlock detection).
177 lock
->lf_next
= block
;
178 TAILQ_INSERT_TAIL(&block
->lf_blkhd
, lock
, lf_block
);
180 if (lockf_debug
& 1) {
181 lf_print("lf_setlock: blocking on", block
);
182 lf_printlist("lf_setlock", block
);
184 #endif /* LOCKF_DEBUG */
185 if (error
= tsleep((caddr_t
)lock
, priority
, lockstr
, 0)) {
187 * We may have been awakened by a signal (in
188 * which case we must remove ourselves from the
189 * blocked list) and/or by another process
190 * releasing a lock (in which case we have already
191 * been removed from the blocked list and our
192 * lf_next field set to NOLOCKF).
195 TAILQ_REMOVE(&lock
->lf_next
->lf_blkhd
, lock
,
197 _FREE(lock
, M_LOCKF
);
202 * No blocks!! Add the lock. Note that we will
203 * downgrade or upgrade any overlapping locks this
204 * process already owns.
206 * Skip over locks owned by other processes.
207 * Handle any locks that overlap and are owned by ourselves.
213 if (ovcase
= lf_findoverlap(block
, lock
, SELF
, &prev
, &overlap
))
214 block
= overlap
->lf_next
;
219 * 2) overlap contains lock
220 * 3) lock contains overlap
221 * 4) overlap starts before lock
222 * 5) overlap ends after lock
225 case 0: /* no overlap */
228 lock
->lf_next
= overlap
;
232 case 1: /* overlap == lock */
234 * If downgrading lock, others may be
235 * able to acquire it.
237 if (lock
->lf_type
== F_RDLCK
&&
238 overlap
->lf_type
== F_WRLCK
)
239 lf_wakelock(overlap
);
240 overlap
->lf_type
= lock
->lf_type
;
242 lock
= overlap
; /* for debug output below */
245 case 2: /* overlap contains lock */
247 * Check for common starting point and different types.
249 if (overlap
->lf_type
== lock
->lf_type
) {
250 _FREE(lock
, M_LOCKF
);
251 lock
= overlap
; /* for debug output below */
254 if (overlap
->lf_start
== lock
->lf_start
) {
256 lock
->lf_next
= overlap
;
257 overlap
->lf_start
= lock
->lf_end
+ 1;
259 lf_split(overlap
, lock
);
260 lf_wakelock(overlap
);
263 case 3: /* lock contains overlap */
265 * If downgrading lock, others may be able to
266 * acquire it, otherwise take the list.
268 if (lock
->lf_type
== F_RDLCK
&&
269 overlap
->lf_type
== F_WRLCK
) {
270 lf_wakelock(overlap
);
272 while (ltmp
= overlap
->lf_blkhd
.tqh_first
) {
273 TAILQ_REMOVE(&overlap
->lf_blkhd
, ltmp
,
275 TAILQ_INSERT_TAIL(&lock
->lf_blkhd
,
280 * Add the new lock if necessary and delete the overlap.
284 lock
->lf_next
= overlap
->lf_next
;
285 prev
= &lock
->lf_next
;
288 *prev
= overlap
->lf_next
;
289 _FREE(overlap
, M_LOCKF
);
292 case 4: /* overlap starts before lock */
294 * Add lock after overlap on the list.
296 lock
->lf_next
= overlap
->lf_next
;
297 overlap
->lf_next
= lock
;
298 overlap
->lf_end
= lock
->lf_start
- 1;
299 prev
= &lock
->lf_next
;
300 lf_wakelock(overlap
);
304 case 5: /* overlap ends after lock */
306 * Add the new lock before overlap.
310 lock
->lf_next
= overlap
;
312 overlap
->lf_start
= lock
->lf_end
+ 1;
313 lf_wakelock(overlap
);
319 if (lockf_debug
& 1) {
320 lf_print("lf_setlock: got the lock", lock
);
321 lf_printlist("lf_setlock", lock
);
323 #endif /* LOCKF_DEBUG */
328 * Remove a byte-range lock on an inode.
330 * Generally, find the lock (or an overlap to that lock)
331 * and remove it (or shrink it), then wakeup anyone we can.
335 register struct lockf
*unlock
;
337 struct inode
*ip
= unlock
->lf_inode
;
338 register struct lockf
*lf
= ip
->i_lockf
;
339 struct lockf
*overlap
, **prev
;
345 if (unlock
->lf_type
!= F_UNLCK
)
346 panic("lf_clearlock: bad type");
348 lf_print("lf_clearlock", unlock
);
349 #endif /* LOCKF_DEBUG */
351 while (ovcase
= lf_findoverlap(lf
, unlock
, SELF
, &prev
, &overlap
)) {
353 * Wakeup the list of locks to be retried.
355 lf_wakelock(overlap
);
359 case 1: /* overlap == lock */
360 *prev
= overlap
->lf_next
;
361 FREE(overlap
, M_LOCKF
);
364 case 2: /* overlap contains lock: split it */
365 if (overlap
->lf_start
== unlock
->lf_start
) {
366 overlap
->lf_start
= unlock
->lf_end
+ 1;
369 lf_split(overlap
, unlock
);
370 overlap
->lf_next
= unlock
->lf_next
;
373 case 3: /* lock contains overlap */
374 *prev
= overlap
->lf_next
;
375 lf
= overlap
->lf_next
;
376 _FREE(overlap
, M_LOCKF
);
379 case 4: /* overlap starts before lock */
380 overlap
->lf_end
= unlock
->lf_start
- 1;
381 prev
= &overlap
->lf_next
;
382 lf
= overlap
->lf_next
;
385 case 5: /* overlap ends after lock */
386 overlap
->lf_start
= unlock
->lf_end
+ 1;
393 lf_printlist("lf_clearlock", unlock
);
394 #endif /* LOCKF_DEBUG */
399 * Check whether there is a blocking lock,
400 * and if so return its process identifier.
404 register struct lockf
*lock
;
405 register struct flock
*fl
;
407 register struct lockf
*block
;
411 lf_print("lf_getlock", lock
);
412 #endif /* LOCKF_DEBUG */
414 if (block
= lf_getblock(lock
)) {
415 fl
->l_type
= block
->lf_type
;
416 fl
->l_whence
= SEEK_SET
;
417 fl
->l_start
= block
->lf_start
;
418 if (block
->lf_end
== -1)
421 fl
->l_len
= block
->lf_end
- block
->lf_start
+ 1;
422 if (block
->lf_flags
& F_POSIX
)
423 fl
->l_pid
= ((struct proc
*)(block
->lf_id
))->p_pid
;
427 fl
->l_type
= F_UNLCK
;
433 * Walk the list of locks for an inode and
434 * return the first blocking lock.
438 register struct lockf
*lock
;
440 struct lockf
**prev
, *overlap
, *lf
= lock
->lf_inode
->i_lockf
;
443 prev
= &lock
->lf_inode
->i_lockf
;
444 while (ovcase
= lf_findoverlap(lf
, lock
, OTHERS
, &prev
, &overlap
)) {
446 * We've found an overlap, see if it blocks us
448 if ((lock
->lf_type
== F_WRLCK
|| overlap
->lf_type
== F_WRLCK
))
451 * Nope, point to the next one on the list and
452 * see if it blocks us
454 lf
= overlap
->lf_next
;
460 * Walk the list of locks for an inode to
461 * find an overlapping lock (if any).
463 * NOTE: this returns only the FIRST overlapping lock. There
464 * may be more than one.
467 lf_findoverlap(lf
, lock
, type
, prev
, overlap
)
468 register struct lockf
*lf
;
471 struct lockf
***prev
;
472 struct lockf
**overlap
;
481 lf_print("lf_findoverlap: looking for overlap in", lock
);
482 #endif /* LOCKF_DEBUG */
483 start
= lock
->lf_start
;
485 while (lf
!= NOLOCKF
) {
486 if (((type
& SELF
) && lf
->lf_id
!= lock
->lf_id
) ||
487 ((type
& OTHERS
) && lf
->lf_id
== lock
->lf_id
)) {
488 *prev
= &lf
->lf_next
;
489 *overlap
= lf
= lf
->lf_next
;
494 lf_print("\tchecking", lf
);
495 #endif /* LOCKF_DEBUG */
497 * OK, check for overlap
502 * 2) overlap contains lock
503 * 3) lock contains overlap
504 * 4) overlap starts before lock
505 * 5) overlap ends after lock
507 if ((lf
->lf_end
!= -1 && start
> lf
->lf_end
) ||
508 (end
!= -1 && lf
->lf_start
> end
)) {
512 printf("no overlap\n");
513 #endif /* LOCKF_DEBUG */
514 if ((type
& SELF
) && end
!= -1 && lf
->lf_start
> end
)
516 *prev
= &lf
->lf_next
;
517 *overlap
= lf
= lf
->lf_next
;
520 if ((lf
->lf_start
== start
) && (lf
->lf_end
== end
)) {
524 printf("overlap == lock\n");
525 #endif /* LOCKF_DEBUG */
528 if ((lf
->lf_start
<= start
) &&
530 ((lf
->lf_end
>= end
) || (lf
->lf_end
== -1))) {
534 printf("overlap contains lock\n");
535 #endif /* LOCKF_DEBUG */
538 if (start
<= lf
->lf_start
&&
540 (lf
->lf_end
!= -1 && end
>= lf
->lf_end
))) {
544 printf("lock contains overlap\n");
545 #endif /* LOCKF_DEBUG */
548 if ((lf
->lf_start
< start
) &&
549 ((lf
->lf_end
>= start
) || (lf
->lf_end
== -1))) {
553 printf("overlap starts before lock\n");
554 #endif /* LOCKF_DEBUG */
557 if ((lf
->lf_start
> start
) &&
559 ((lf
->lf_end
> end
) || (lf
->lf_end
== -1))) {
563 printf("overlap ends after lock\n");
564 #endif /* LOCKF_DEBUG */
567 panic("lf_findoverlap: default");
573 * Split a lock and a contained region into
574 * two or three locks as necessary.
577 lf_split(lock1
, lock2
)
578 register struct lockf
*lock1
;
579 register struct lockf
*lock2
;
581 register struct lockf
*splitlock
;
584 if (lockf_debug
& 2) {
585 lf_print("lf_split", lock1
);
586 lf_print("splitting from", lock2
);
588 #endif /* LOCKF_DEBUG */
590 * Check to see if spliting into only two pieces.
592 if (lock1
->lf_start
== lock2
->lf_start
) {
593 lock1
->lf_start
= lock2
->lf_end
+ 1;
594 lock2
->lf_next
= lock1
;
597 if (lock1
->lf_end
== lock2
->lf_end
) {
598 lock1
->lf_end
= lock2
->lf_start
- 1;
599 lock2
->lf_next
= lock1
->lf_next
;
600 lock1
->lf_next
= lock2
;
604 * Make a new lock consisting of the last part of
605 * the encompassing lock
607 MALLOC(splitlock
, struct lockf
*, sizeof *splitlock
, M_LOCKF
, M_WAITOK
);
608 bcopy((caddr_t
)lock1
, (caddr_t
)splitlock
, sizeof *splitlock
);
609 splitlock
->lf_start
= lock2
->lf_end
+ 1;
610 TAILQ_INIT(&splitlock
->lf_blkhd
);
611 lock1
->lf_end
= lock2
->lf_start
- 1;
615 splitlock
->lf_next
= lock1
->lf_next
;
616 lock2
->lf_next
= splitlock
;
617 lock1
->lf_next
= lock2
;
624 lf_wakelock(listhead
)
625 struct lockf
*listhead
;
627 register struct lockf
*wakelock
;
629 while (wakelock
= listhead
->lf_blkhd
.tqh_first
) {
630 TAILQ_REMOVE(&listhead
->lf_blkhd
, wakelock
, lf_block
);
631 wakelock
->lf_next
= NOLOCKF
;
634 lf_print("lf_wakelock: awakening", wakelock
);
635 #endif /* LOCKF_DEBUG */
636 wakeup((caddr_t
)wakelock
);
646 register struct lockf
*lock
;
649 printf("%s: lock 0x%lx for ", tag
, lock
);
650 if (lock
->lf_flags
& F_POSIX
)
651 printf("proc %d", ((struct proc
*)(lock
->lf_id
))->p_pid
);
653 printf("id 0x%x", lock
->lf_id
);
654 printf(" in ino %d on dev <%d, %d>, %s, start %d, end %d",
655 lock
->lf_inode
->i_number
,
656 major(lock
->lf_inode
->i_dev
),
657 minor(lock
->lf_inode
->i_dev
),
658 lock
->lf_type
== F_RDLCK
? "shared" :
659 lock
->lf_type
== F_WRLCK
? "exclusive" :
660 lock
->lf_type
== F_UNLCK
? "unlock" :
661 "unknown", lock
->lf_start
, lock
->lf_end
);
662 if (lock
->lf_blkhd
.tqh_first
)
663 printf(" block 0x%x\n", lock
->lf_blkhd
.tqh_first
);
668 lf_printlist(tag
, lock
)
672 register struct lockf
*lf
, *blk
;
674 printf("%s: Lock list for ino %d on dev <%d, %d>:\n",
675 tag
, lock
->lf_inode
->i_number
,
676 major(lock
->lf_inode
->i_dev
),
677 minor(lock
->lf_inode
->i_dev
));
678 for (lf
= lock
->lf_inode
->i_lockf
; lf
; lf
= lf
->lf_next
) {
679 printf("\tlock 0x%lx for ", lf
);
680 if (lf
->lf_flags
& F_POSIX
)
681 printf("proc %d", ((struct proc
*)(lf
->lf_id
))->p_pid
);
683 printf("id 0x%x", lf
->lf_id
);
684 printf(", %s, start %d, end %d",
685 lf
->lf_type
== F_RDLCK
? "shared" :
686 lf
->lf_type
== F_WRLCK
? "exclusive" :
687 lf
->lf_type
== F_UNLCK
? "unlock" :
688 "unknown", lf
->lf_start
, lf
->lf_end
);
689 for (blk
= lf
->lf_blkhd
.tqh_first
; blk
;
690 blk
= blk
->lf_block
.tqe_next
) {
691 printf("\n\t\tlock request 0x%lx for ", blk
);
692 if (blk
->lf_flags
& F_POSIX
)
694 ((struct proc
*)(blk
->lf_id
))->p_pid
);
696 printf("id 0x%x", blk
->lf_id
);
697 printf(", %s, start %d, end %d",
698 blk
->lf_type
== F_RDLCK
? "shared" :
699 blk
->lf_type
== F_WRLCK
? "exclusive" :
700 blk
->lf_type
== F_UNLCK
? "unlock" :
701 "unknown", blk
->lf_start
, blk
->lf_end
);
702 if (blk
->lf_blkhd
.tqh_first
)
703 panic("lf_printlist: bad list");
708 #endif /* LOCKF_DEBUG */