lfs_subr.c revision 1.79 1 /* $NetBSD: lfs_subr.c,v 1.79 2013/06/18 18:18:58 christos Exp $ */
2
3 /*-
4 * Copyright (c) 1999, 2000, 2001, 2002, 2003 The NetBSD Foundation, Inc.
5 * All rights reserved.
6 *
7 * This code is derived from software contributed to The NetBSD Foundation
8 * by Konrad E. Schroder <perseant (at) hhhh.org>.
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 *
19 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
21 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
22 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
23 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
29 * POSSIBILITY OF SUCH DAMAGE.
30 */
31 /*
32 * Copyright (c) 1991, 1993
33 * The Regents of the University of California. All rights reserved.
34 *
35 * Redistribution and use in source and binary forms, with or without
36 * modification, are permitted provided that the following conditions
37 * are met:
38 * 1. Redistributions of source code must retain the above copyright
39 * notice, this list of conditions and the following disclaimer.
40 * 2. Redistributions in binary form must reproduce the above copyright
41 * notice, this list of conditions and the following disclaimer in the
42 * documentation and/or other materials provided with the distribution.
43 * 3. Neither the name of the University nor the names of its contributors
44 * may be used to endorse or promote products derived from this software
45 * without specific prior written permission.
46 *
47 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
48 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
49 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
50 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
51 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
52 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
53 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
54 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
55 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
56 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
57 * SUCH DAMAGE.
58 *
59 * @(#)lfs_subr.c 8.4 (Berkeley) 5/8/95
60 */
61
62 #include <sys/cdefs.h>
63 __KERNEL_RCSID(0, "$NetBSD: lfs_subr.c,v 1.79 2013/06/18 18:18:58 christos Exp $");
64
65 #include <sys/param.h>
66 #include <sys/systm.h>
67 #include <sys/namei.h>
68 #include <sys/vnode.h>
69 #include <sys/buf.h>
70 #include <sys/mount.h>
71 #include <sys/malloc.h>
72 #include <sys/proc.h>
73 #include <sys/kauth.h>
74
75 #include <ufs/lfs/ulfs_inode.h>
76 #include <ufs/lfs/lfs.h>
77 #include <ufs/lfs/lfs_extern.h>
78
79 #include <uvm/uvm.h>
80
81 #ifdef DEBUG
82 const char *lfs_res_names[LFS_NB_COUNT] = {
83 "summary",
84 "superblock",
85 "file block",
86 "cluster",
87 "clean",
88 "blkiov",
89 };
90 #endif
91
92 int lfs_res_qty[LFS_NB_COUNT] = {
93 LFS_N_SUMMARIES,
94 LFS_N_SBLOCKS,
95 LFS_N_IBLOCKS,
96 LFS_N_CLUSTERS,
97 LFS_N_CLEAN,
98 LFS_N_BLKIOV,
99 };
100
101 void
102 lfs_setup_resblks(struct lfs *fs)
103 {
104 int i, j;
105 int maxbpp;
106
107 ASSERT_NO_SEGLOCK(fs);
108 fs->lfs_resblk = (res_t *)malloc(LFS_N_TOTAL * sizeof(res_t), M_SEGMENT,
109 M_WAITOK);
110 for (i = 0; i < LFS_N_TOTAL; i++) {
111 fs->lfs_resblk[i].inuse = 0;
112 fs->lfs_resblk[i].p = NULL;
113 }
114 for (i = 0; i < LFS_RESHASH_WIDTH; i++)
115 LIST_INIT(fs->lfs_reshash + i);
116
117 /*
118 * These types of allocations can be larger than a page,
119 * so we can't use the pool subsystem for them.
120 */
121 for (i = 0, j = 0; j < LFS_N_SUMMARIES; j++, i++)
122 fs->lfs_resblk[i].size = fs->lfs_sumsize;
123 for (j = 0; j < LFS_N_SBLOCKS; j++, i++)
124 fs->lfs_resblk[i].size = LFS_SBPAD;
125 for (j = 0; j < LFS_N_IBLOCKS; j++, i++)
126 fs->lfs_resblk[i].size = fs->lfs_bsize;
127 for (j = 0; j < LFS_N_CLUSTERS; j++, i++)
128 fs->lfs_resblk[i].size = MAXPHYS;
129 for (j = 0; j < LFS_N_CLEAN; j++, i++)
130 fs->lfs_resblk[i].size = MAXPHYS;
131 for (j = 0; j < LFS_N_BLKIOV; j++, i++)
132 fs->lfs_resblk[i].size = LFS_MARKV_MAXBLKCNT * sizeof(BLOCK_INFO);
133
134 for (i = 0; i < LFS_N_TOTAL; i++) {
135 fs->lfs_resblk[i].p = malloc(fs->lfs_resblk[i].size,
136 M_SEGMENT, M_WAITOK);
137 }
138
139 /*
140 * Initialize pools for small types (XXX is BPP small?)
141 */
142 pool_init(&fs->lfs_clpool, sizeof(struct lfs_cluster), 0, 0, 0,
143 "lfsclpl", &pool_allocator_nointr, IPL_NONE);
144 pool_init(&fs->lfs_segpool, sizeof(struct segment), 0, 0, 0,
145 "lfssegpool", &pool_allocator_nointr, IPL_NONE);
146 maxbpp = ((fs->lfs_sumsize - SEGSUM_SIZE(fs)) / sizeof(int32_t) + 2);
147 maxbpp = MIN(maxbpp, lfs_segsize(fs) / fs->lfs_fsize + 2);
148 pool_init(&fs->lfs_bpppool, maxbpp * sizeof(struct buf *), 0, 0, 0,
149 "lfsbpppl", &pool_allocator_nointr, IPL_NONE);
150 }
151
152 void
153 lfs_free_resblks(struct lfs *fs)
154 {
155 int i;
156
157 pool_destroy(&fs->lfs_bpppool);
158 pool_destroy(&fs->lfs_segpool);
159 pool_destroy(&fs->lfs_clpool);
160
161 mutex_enter(&lfs_lock);
162 for (i = 0; i < LFS_N_TOTAL; i++) {
163 while (fs->lfs_resblk[i].inuse)
164 mtsleep(&fs->lfs_resblk, PRIBIO + 1, "lfs_free", 0,
165 &lfs_lock);
166 if (fs->lfs_resblk[i].p != NULL)
167 free(fs->lfs_resblk[i].p, M_SEGMENT);
168 }
169 free(fs->lfs_resblk, M_SEGMENT);
170 mutex_exit(&lfs_lock);
171 }
172
173 static unsigned int
174 lfs_mhash(void *vp)
175 {
176 return (unsigned int)(((unsigned long)vp) >> 2) % LFS_RESHASH_WIDTH;
177 }
178
179 /*
180 * Return memory of the given size for the given purpose, or use one of a
181 * number of spare last-resort buffers, if malloc returns NULL.
182 */
183 void *
184 lfs_malloc(struct lfs *fs, size_t size, int type)
185 {
186 struct lfs_res_blk *re;
187 void *r;
188 int i, s, start;
189 unsigned int h;
190
191 ASSERT_MAYBE_SEGLOCK(fs);
192 r = NULL;
193
194 /* If no mem allocated for this type, it just waits */
195 if (lfs_res_qty[type] == 0) {
196 r = malloc(size, M_SEGMENT, M_WAITOK);
197 return r;
198 }
199
200 /* Otherwise try a quick malloc, and if it works, great */
201 if ((r = malloc(size, M_SEGMENT, M_NOWAIT)) != NULL) {
202 return r;
203 }
204
205 /*
206 * If malloc returned NULL, we are forced to use one of our
207 * reserve blocks. We have on hand at least one summary block,
208 * at least one cluster block, at least one superblock,
209 * and several indirect blocks.
210 */
211
212 mutex_enter(&lfs_lock);
213 /* skip over blocks of other types */
214 for (i = 0, start = 0; i < type; i++)
215 start += lfs_res_qty[i];
216 while (r == NULL) {
217 for (i = 0; i < lfs_res_qty[type]; i++) {
218 if (fs->lfs_resblk[start + i].inuse == 0) {
219 re = fs->lfs_resblk + start + i;
220 re->inuse = 1;
221 r = re->p;
222 KASSERT(re->size >= size);
223 h = lfs_mhash(r);
224 s = splbio();
225 LIST_INSERT_HEAD(&fs->lfs_reshash[h], re, res);
226 splx(s);
227 mutex_exit(&lfs_lock);
228 return r;
229 }
230 }
231 DLOG((DLOG_MALLOC, "sleeping on %s (%d)\n",
232 lfs_res_names[type], lfs_res_qty[type]));
233 mtsleep(&fs->lfs_resblk, PVM, "lfs_malloc", 0,
234 &lfs_lock);
235 DLOG((DLOG_MALLOC, "done sleeping on %s\n",
236 lfs_res_names[type]));
237 }
238 /* NOTREACHED */
239 mutex_exit(&lfs_lock);
240 return r;
241 }
242
243 void
244 lfs_free(struct lfs *fs, void *p, int type)
245 {
246 int s;
247 unsigned int h;
248 res_t *re;
249 #ifdef DEBUG
250 int i;
251 #endif
252
253 ASSERT_MAYBE_SEGLOCK(fs);
254 h = lfs_mhash(p);
255 mutex_enter(&lfs_lock);
256 s = splbio();
257 LIST_FOREACH(re, &fs->lfs_reshash[h], res) {
258 if (re->p == p) {
259 KASSERT(re->inuse == 1);
260 LIST_REMOVE(re, res);
261 re->inuse = 0;
262 wakeup(&fs->lfs_resblk);
263 splx(s);
264 mutex_exit(&lfs_lock);
265 return;
266 }
267 }
268 #ifdef DEBUG
269 for (i = 0; i < LFS_N_TOTAL; i++) {
270 if (fs->lfs_resblk[i].p == p)
271 panic("lfs_free: inconsistent reserved block");
272 }
273 #endif
274 splx(s);
275 mutex_exit(&lfs_lock);
276
277 /*
278 * If we didn't find it, free it.
279 */
280 free(p, M_SEGMENT);
281 }
282
283 /*
284 * lfs_seglock --
285 * Single thread the segment writer.
286 */
287 int
288 lfs_seglock(struct lfs *fs, unsigned long flags)
289 {
290 struct segment *sp;
291
292 mutex_enter(&lfs_lock);
293 if (fs->lfs_seglock) {
294 if (fs->lfs_lockpid == curproc->p_pid &&
295 fs->lfs_locklwp == curlwp->l_lid) {
296 ++fs->lfs_seglock;
297 fs->lfs_sp->seg_flags |= flags;
298 mutex_exit(&lfs_lock);
299 return 0;
300 } else if (flags & SEGM_PAGEDAEMON) {
301 mutex_exit(&lfs_lock);
302 return EWOULDBLOCK;
303 } else {
304 while (fs->lfs_seglock) {
305 (void)mtsleep(&fs->lfs_seglock, PRIBIO + 1,
306 "lfs_seglock", 0, &lfs_lock);
307 }
308 }
309 }
310
311 fs->lfs_seglock = 1;
312 fs->lfs_lockpid = curproc->p_pid;
313 fs->lfs_locklwp = curlwp->l_lid;
314 mutex_exit(&lfs_lock);
315 fs->lfs_cleanind = 0;
316
317 #ifdef DEBUG
318 LFS_ENTER_LOG("seglock", __FILE__, __LINE__, 0, flags, curproc->p_pid);
319 #endif
320 /* Drain fragment size changes out */
321 rw_enter(&fs->lfs_fraglock, RW_WRITER);
322
323 sp = fs->lfs_sp = pool_get(&fs->lfs_segpool, PR_WAITOK);
324 sp->bpp = pool_get(&fs->lfs_bpppool, PR_WAITOK);
325 sp->seg_flags = flags;
326 sp->vp = NULL;
327 sp->seg_iocount = 0;
328 (void) lfs_initseg(fs);
329
330 /*
331 * Keep a cumulative count of the outstanding I/O operations. If the
332 * disk drive catches up with us it could go to zero before we finish,
333 * so we artificially increment it by one until we've scheduled all of
334 * the writes we intend to do.
335 */
336 mutex_enter(&lfs_lock);
337 ++fs->lfs_iocount;
338 fs->lfs_startseg = fs->lfs_curseg;
339 mutex_exit(&lfs_lock);
340 return 0;
341 }
342
343 static void lfs_unmark_dirop(struct lfs *);
344
345 static void
346 lfs_unmark_dirop(struct lfs *fs)
347 {
348 struct inode *ip, *nip;
349 struct vnode *vp;
350 int doit;
351
352 ASSERT_NO_SEGLOCK(fs);
353 mutex_enter(&lfs_lock);
354 doit = !(fs->lfs_flags & LFS_UNDIROP);
355 if (doit)
356 fs->lfs_flags |= LFS_UNDIROP;
357 if (!doit) {
358 mutex_exit(&lfs_lock);
359 return;
360 }
361
362 for (ip = TAILQ_FIRST(&fs->lfs_dchainhd); ip != NULL; ip = nip) {
363 nip = TAILQ_NEXT(ip, i_lfs_dchain);
364 vp = ITOV(ip);
365 if ((ip->i_flag & (IN_ADIROP | IN_CDIROP)) == IN_CDIROP) {
366 --lfs_dirvcount;
367 --fs->lfs_dirvcount;
368 vp->v_uflag &= ~VU_DIROP;
369 TAILQ_REMOVE(&fs->lfs_dchainhd, ip, i_lfs_dchain);
370 wakeup(&lfs_dirvcount);
371 fs->lfs_unlockvp = vp;
372 mutex_exit(&lfs_lock);
373 vrele(vp);
374 mutex_enter(&lfs_lock);
375 fs->lfs_unlockvp = NULL;
376 ip->i_flag &= ~IN_CDIROP;
377 }
378 }
379
380 fs->lfs_flags &= ~LFS_UNDIROP;
381 wakeup(&fs->lfs_flags);
382 mutex_exit(&lfs_lock);
383 }
384
385 static void
386 lfs_auto_segclean(struct lfs *fs)
387 {
388 int i, error, s, waited;
389
390 ASSERT_SEGLOCK(fs);
391 /*
392 * Now that we've swapped lfs_activesb, but while we still
393 * hold the segment lock, run through the segment list marking
394 * the empty ones clean.
395 * XXX - do we really need to do them all at once?
396 */
397 waited = 0;
398 for (i = 0; i < fs->lfs_nseg; i++) {
399 if ((fs->lfs_suflags[0][i] &
400 (SEGUSE_ACTIVE | SEGUSE_DIRTY | SEGUSE_EMPTY)) ==
401 (SEGUSE_DIRTY | SEGUSE_EMPTY) &&
402 (fs->lfs_suflags[1][i] &
403 (SEGUSE_ACTIVE | SEGUSE_DIRTY | SEGUSE_EMPTY)) ==
404 (SEGUSE_DIRTY | SEGUSE_EMPTY)) {
405
406 /* Make sure the sb is written before we clean */
407 mutex_enter(&lfs_lock);
408 s = splbio();
409 while (waited == 0 && fs->lfs_sbactive)
410 mtsleep(&fs->lfs_sbactive, PRIBIO+1, "lfs asb",
411 0, &lfs_lock);
412 splx(s);
413 mutex_exit(&lfs_lock);
414 waited = 1;
415
416 if ((error = lfs_do_segclean(fs, i)) != 0) {
417 DLOG((DLOG_CLEAN, "lfs_auto_segclean: lfs_do_segclean returned %d for seg %d\n", error, i));
418 }
419 }
420 fs->lfs_suflags[1 - fs->lfs_activesb][i] =
421 fs->lfs_suflags[fs->lfs_activesb][i];
422 }
423 }
424
425 /*
426 * lfs_segunlock --
427 * Single thread the segment writer.
428 */
429 void
430 lfs_segunlock(struct lfs *fs)
431 {
432 struct segment *sp;
433 unsigned long sync, ckp;
434 struct buf *bp;
435 int do_unmark_dirop = 0;
436
437 sp = fs->lfs_sp;
438
439 mutex_enter(&lfs_lock);
440 KASSERT(LFS_SEGLOCK_HELD(fs));
441 if (fs->lfs_seglock == 1) {
442 if ((sp->seg_flags & (SEGM_PROT | SEGM_CLEAN)) == 0)
443 do_unmark_dirop = 1;
444 mutex_exit(&lfs_lock);
445 sync = sp->seg_flags & SEGM_SYNC;
446 ckp = sp->seg_flags & SEGM_CKP;
447
448 /* We should have a segment summary, and nothing else */
449 KASSERT(sp->cbpp == sp->bpp + 1);
450
451 /* Free allocated segment summary */
452 fs->lfs_offset -= lfs_btofsb(fs, fs->lfs_sumsize);
453 bp = *sp->bpp;
454 lfs_freebuf(fs, bp);
455
456 pool_put(&fs->lfs_bpppool, sp->bpp);
457 sp->bpp = NULL;
458
459 /*
460 * If we're not sync, we're done with sp, get rid of it.
461 * Otherwise, we keep a local copy around but free
462 * fs->lfs_sp so another process can use it (we have to
463 * wait but they don't have to wait for us).
464 */
465 if (!sync)
466 pool_put(&fs->lfs_segpool, sp);
467 fs->lfs_sp = NULL;
468
469 /*
470 * If the I/O count is non-zero, sleep until it reaches zero.
471 * At the moment, the user's process hangs around so we can
472 * sleep.
473 */
474 mutex_enter(&lfs_lock);
475 if (--fs->lfs_iocount == 0) {
476 LFS_DEBUG_COUNTLOCKED("lfs_segunlock");
477 }
478 if (fs->lfs_iocount <= 1)
479 wakeup(&fs->lfs_iocount);
480 mutex_exit(&lfs_lock);
481 /*
482 * If we're not checkpointing, we don't have to block
483 * other processes to wait for a synchronous write
484 * to complete.
485 */
486 if (!ckp) {
487 #ifdef DEBUG
488 LFS_ENTER_LOG("segunlock_std", __FILE__, __LINE__, 0, 0, curproc->p_pid);
489 #endif
490 mutex_enter(&lfs_lock);
491 --fs->lfs_seglock;
492 fs->lfs_lockpid = 0;
493 fs->lfs_locklwp = 0;
494 mutex_exit(&lfs_lock);
495 wakeup(&fs->lfs_seglock);
496 }
497 /*
498 * We let checkpoints happen asynchronously. That means
499 * that during recovery, we have to roll forward between
500 * the two segments described by the first and second
501 * superblocks to make sure that the checkpoint described
502 * by a superblock completed.
503 */
504 mutex_enter(&lfs_lock);
505 while (ckp && sync && fs->lfs_iocount) {
506 (void)mtsleep(&fs->lfs_iocount, PRIBIO + 1,
507 "lfs_iocount", 0, &lfs_lock);
508 DLOG((DLOG_SEG, "sleeping on iocount %x == %d\n", fs, fs->lfs_iocount));
509 }
510 while (sync && sp->seg_iocount) {
511 (void)mtsleep(&sp->seg_iocount, PRIBIO + 1,
512 "seg_iocount", 0, &lfs_lock);
513 DLOG((DLOG_SEG, "sleeping on iocount %x == %d\n", sp, sp->seg_iocount));
514 }
515 mutex_exit(&lfs_lock);
516 if (sync)
517 pool_put(&fs->lfs_segpool, sp);
518
519 if (ckp) {
520 fs->lfs_nactive = 0;
521 /* If we *know* everything's on disk, write both sbs */
522 /* XXX should wait for this one */
523 if (sync)
524 lfs_writesuper(fs, fs->lfs_sboffs[fs->lfs_activesb]);
525 lfs_writesuper(fs, fs->lfs_sboffs[1 - fs->lfs_activesb]);
526 if (!(fs->lfs_ivnode->v_mount->mnt_iflag & IMNT_UNMOUNT)) {
527 lfs_auto_segclean(fs);
528 /* If sync, we can clean the remainder too */
529 if (sync)
530 lfs_auto_segclean(fs);
531 }
532 fs->lfs_activesb = 1 - fs->lfs_activesb;
533 #ifdef DEBUG
534 LFS_ENTER_LOG("segunlock_ckp", __FILE__, __LINE__, 0, 0, curproc->p_pid);
535 #endif
536 mutex_enter(&lfs_lock);
537 --fs->lfs_seglock;
538 fs->lfs_lockpid = 0;
539 fs->lfs_locklwp = 0;
540 mutex_exit(&lfs_lock);
541 wakeup(&fs->lfs_seglock);
542 }
543 /* Reenable fragment size changes */
544 rw_exit(&fs->lfs_fraglock);
545 if (do_unmark_dirop)
546 lfs_unmark_dirop(fs);
547 } else if (fs->lfs_seglock == 0) {
548 mutex_exit(&lfs_lock);
549 panic ("Seglock not held");
550 } else {
551 --fs->lfs_seglock;
552 mutex_exit(&lfs_lock);
553 }
554 }
555
556 /*
557 * Drain dirops and start writer.
558 *
559 * No simple_locks are held when we enter and none are held when we return.
560 */
561 int
562 lfs_writer_enter(struct lfs *fs, const char *wmesg)
563 {
564 int error = 0;
565
566 ASSERT_MAYBE_SEGLOCK(fs);
567 mutex_enter(&lfs_lock);
568
569 /* disallow dirops during flush */
570 fs->lfs_writer++;
571
572 while (fs->lfs_dirops > 0) {
573 ++fs->lfs_diropwait;
574 error = mtsleep(&fs->lfs_writer, PRIBIO+1, wmesg, 0,
575 &lfs_lock);
576 --fs->lfs_diropwait;
577 }
578
579 if (error)
580 fs->lfs_writer--;
581
582 mutex_exit(&lfs_lock);
583
584 return error;
585 }
586
587 void
588 lfs_writer_leave(struct lfs *fs)
589 {
590 bool dowakeup;
591
592 ASSERT_MAYBE_SEGLOCK(fs);
593 mutex_enter(&lfs_lock);
594 dowakeup = !(--fs->lfs_writer);
595 mutex_exit(&lfs_lock);
596 if (dowakeup)
597 wakeup(&fs->lfs_dirops);
598 }
599
600 /*
601 * Unlock, wait for the cleaner, then relock to where we were before.
602 * To be used only at a fairly high level, to address a paucity of free
603 * segments propagated back from lfs_gop_write().
604 */
605 void
606 lfs_segunlock_relock(struct lfs *fs)
607 {
608 int n = fs->lfs_seglock;
609 u_int16_t seg_flags;
610 CLEANERINFO *cip;
611 struct buf *bp;
612
613 if (n == 0)
614 return;
615
616 /* Write anything we've already gathered to disk */
617 lfs_writeseg(fs, fs->lfs_sp);
618
619 /* Tell cleaner */
620 LFS_CLEANERINFO(cip, fs, bp);
621 cip->flags |= LFS_CLEANER_MUST_CLEAN;
622 LFS_SYNC_CLEANERINFO(cip, fs, bp, 1);
623
624 /* Save segment flags for later */
625 seg_flags = fs->lfs_sp->seg_flags;
626
627 fs->lfs_sp->seg_flags |= SEGM_PROT; /* Don't unmark dirop nodes */
628 while(fs->lfs_seglock)
629 lfs_segunlock(fs);
630
631 /* Wait for the cleaner */
632 lfs_wakeup_cleaner(fs);
633 mutex_enter(&lfs_lock);
634 while (LFS_STARVED_FOR_SEGS(fs))
635 mtsleep(&fs->lfs_avail, PRIBIO, "relock", 0,
636 &lfs_lock);
637 mutex_exit(&lfs_lock);
638
639 /* Put the segment lock back the way it was. */
640 while(n--)
641 lfs_seglock(fs, seg_flags);
642
643 /* Cleaner can relax now */
644 LFS_CLEANERINFO(cip, fs, bp);
645 cip->flags &= ~LFS_CLEANER_MUST_CLEAN;
646 LFS_SYNC_CLEANERINFO(cip, fs, bp, 1);
647
648 return;
649 }
650
651 /*
652 * Wake up the cleaner, provided that nowrap is not set.
653 */
654 void
655 lfs_wakeup_cleaner(struct lfs *fs)
656 {
657 if (fs->lfs_nowrap > 0)
658 return;
659
660 wakeup(&fs->lfs_nextseg);
661 wakeup(&lfs_allclean_wakeup);
662 }
663