Home | History | Annotate | Line # | Download | only in lfs
lfs_subr.c revision 1.80.6.1
      1  1.80.6.1     skrll /*	$NetBSD: lfs_subr.c,v 1.80.6.1 2015/09/22 12:06:17 skrll Exp $	*/
      2       1.2       cgd 
      3       1.8  perseant /*-
      4      1.31  perseant  * Copyright (c) 1999, 2000, 2001, 2002, 2003 The NetBSD Foundation, Inc.
      5       1.8  perseant  * All rights reserved.
      6       1.8  perseant  *
      7       1.8  perseant  * This code is derived from software contributed to The NetBSD Foundation
      8       1.8  perseant  * by Konrad E. Schroder <perseant (at) hhhh.org>.
      9       1.8  perseant  *
     10       1.8  perseant  * Redistribution and use in source and binary forms, with or without
     11       1.8  perseant  * modification, are permitted provided that the following conditions
     12       1.8  perseant  * are met:
     13       1.8  perseant  * 1. Redistributions of source code must retain the above copyright
     14       1.8  perseant  *    notice, this list of conditions and the following disclaimer.
     15       1.8  perseant  * 2. Redistributions in binary form must reproduce the above copyright
     16       1.8  perseant  *    notice, this list of conditions and the following disclaimer in the
     17       1.8  perseant  *    documentation and/or other materials provided with the distribution.
     18       1.8  perseant  *
     19       1.8  perseant  * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
     20       1.8  perseant  * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
     21       1.8  perseant  * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
     22       1.8  perseant  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
     23       1.8  perseant  * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
     24       1.8  perseant  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
     25       1.8  perseant  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
     26       1.8  perseant  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
     27       1.8  perseant  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
     28       1.8  perseant  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
     29       1.8  perseant  * POSSIBILITY OF SUCH DAMAGE.
     30       1.8  perseant  */
     31       1.1   mycroft /*
     32       1.1   mycroft  * Copyright (c) 1991, 1993
     33       1.1   mycroft  *	The Regents of the University of California.  All rights reserved.
     34       1.1   mycroft  *
     35       1.1   mycroft  * Redistribution and use in source and binary forms, with or without
     36       1.1   mycroft  * modification, are permitted provided that the following conditions
     37       1.1   mycroft  * are met:
     38       1.1   mycroft  * 1. Redistributions of source code must retain the above copyright
     39       1.1   mycroft  *    notice, this list of conditions and the following disclaimer.
     40       1.1   mycroft  * 2. Redistributions in binary form must reproduce the above copyright
     41       1.1   mycroft  *    notice, this list of conditions and the following disclaimer in the
     42       1.1   mycroft  *    documentation and/or other materials provided with the distribution.
     43      1.43       agc  * 3. Neither the name of the University nor the names of its contributors
     44       1.1   mycroft  *    may be used to endorse or promote products derived from this software
     45       1.1   mycroft  *    without specific prior written permission.
     46       1.1   mycroft  *
     47       1.1   mycroft  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
     48       1.1   mycroft  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
     49       1.1   mycroft  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
     50       1.1   mycroft  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
     51       1.1   mycroft  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
     52       1.1   mycroft  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
     53       1.1   mycroft  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
     54       1.1   mycroft  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
     55       1.1   mycroft  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
     56       1.1   mycroft  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
     57       1.1   mycroft  * SUCH DAMAGE.
     58       1.1   mycroft  *
     59       1.6      fvdl  *	@(#)lfs_subr.c	8.4 (Berkeley) 5/8/95
     60       1.1   mycroft  */
     61      1.20     lukem 
     62      1.20     lukem #include <sys/cdefs.h>
     63  1.80.6.1     skrll __KERNEL_RCSID(0, "$NetBSD: lfs_subr.c,v 1.80.6.1 2015/09/22 12:06:17 skrll Exp $");
     64       1.1   mycroft 
     65       1.1   mycroft #include <sys/param.h>
     66       1.3  christos #include <sys/systm.h>
     67       1.1   mycroft #include <sys/namei.h>
     68       1.1   mycroft #include <sys/vnode.h>
     69       1.1   mycroft #include <sys/buf.h>
     70       1.1   mycroft #include <sys/mount.h>
     71       1.1   mycroft #include <sys/malloc.h>
     72       1.1   mycroft #include <sys/proc.h>
     73      1.61  perseant #include <sys/kauth.h>
     74       1.1   mycroft 
     75      1.78  dholland #include <ufs/lfs/ulfs_inode.h>
     76       1.1   mycroft #include <ufs/lfs/lfs.h>
     77  1.80.6.1     skrll #include <ufs/lfs/lfs_accessors.h>
     78      1.80  dholland #include <ufs/lfs/lfs_kernel.h>
     79       1.1   mycroft #include <ufs/lfs/lfs_extern.h>
     80       1.1   mycroft 
     81      1.31  perseant #include <uvm/uvm.h>
     82      1.31  perseant 
     83      1.50  perseant #ifdef DEBUG
     84      1.53  christos const char *lfs_res_names[LFS_NB_COUNT] = {
     85      1.31  perseant 	"summary",
     86      1.31  perseant 	"superblock",
     87      1.52  perseant 	"file block",
     88      1.31  perseant 	"cluster",
     89      1.31  perseant 	"clean",
     90      1.52  perseant 	"blkiov",
     91      1.31  perseant };
     92      1.31  perseant #endif
     93      1.31  perseant 
     94      1.31  perseant int lfs_res_qty[LFS_NB_COUNT] = {
     95      1.31  perseant 	LFS_N_SUMMARIES,
     96      1.31  perseant 	LFS_N_SBLOCKS,
     97      1.31  perseant 	LFS_N_IBLOCKS,
     98      1.31  perseant 	LFS_N_CLUSTERS,
     99      1.31  perseant 	LFS_N_CLEAN,
    100      1.52  perseant 	LFS_N_BLKIOV,
    101      1.31  perseant };
    102      1.31  perseant 
    103      1.31  perseant void
    104      1.31  perseant lfs_setup_resblks(struct lfs *fs)
    105      1.31  perseant {
    106      1.31  perseant 	int i, j;
    107      1.31  perseant 	int maxbpp;
    108      1.31  perseant 
    109      1.51  perseant 	ASSERT_NO_SEGLOCK(fs);
    110  1.80.6.1     skrll 	fs->lfs_resblk = malloc(LFS_N_TOTAL * sizeof(res_t), M_SEGMENT,
    111  1.80.6.1     skrll 				M_WAITOK);
    112      1.31  perseant 	for (i = 0; i < LFS_N_TOTAL; i++) {
    113      1.31  perseant 		fs->lfs_resblk[i].inuse = 0;
    114      1.31  perseant 		fs->lfs_resblk[i].p = NULL;
    115      1.31  perseant 	}
    116      1.31  perseant 	for (i = 0; i < LFS_RESHASH_WIDTH; i++)
    117      1.31  perseant 		LIST_INIT(fs->lfs_reshash + i);
    118      1.31  perseant 
    119      1.31  perseant 	/*
    120      1.31  perseant 	 * These types of allocations can be larger than a page,
    121      1.31  perseant 	 * so we can't use the pool subsystem for them.
    122      1.31  perseant 	 */
    123      1.31  perseant 	for (i = 0, j = 0; j < LFS_N_SUMMARIES; j++, i++)
    124  1.80.6.1     skrll 		fs->lfs_resblk[i].size = lfs_sb_getsumsize(fs);
    125      1.31  perseant 	for (j = 0; j < LFS_N_SBLOCKS; j++, i++)
    126      1.34  perseant 		fs->lfs_resblk[i].size = LFS_SBPAD;
    127      1.31  perseant 	for (j = 0; j < LFS_N_IBLOCKS; j++, i++)
    128  1.80.6.1     skrll 		fs->lfs_resblk[i].size = lfs_sb_getbsize(fs);
    129      1.31  perseant 	for (j = 0; j < LFS_N_CLUSTERS; j++, i++)
    130      1.34  perseant 		fs->lfs_resblk[i].size = MAXPHYS;
    131      1.31  perseant 	for (j = 0; j < LFS_N_CLEAN; j++, i++)
    132      1.34  perseant 		fs->lfs_resblk[i].size = MAXPHYS;
    133      1.52  perseant 	for (j = 0; j < LFS_N_BLKIOV; j++, i++)
    134      1.52  perseant 		fs->lfs_resblk[i].size = LFS_MARKV_MAXBLKCNT * sizeof(BLOCK_INFO);
    135      1.34  perseant 
    136      1.34  perseant 	for (i = 0; i < LFS_N_TOTAL; i++) {
    137      1.34  perseant 		fs->lfs_resblk[i].p = malloc(fs->lfs_resblk[i].size,
    138      1.34  perseant 					     M_SEGMENT, M_WAITOK);
    139      1.34  perseant 	}
    140      1.31  perseant 
    141      1.31  perseant 	/*
    142      1.31  perseant 	 * Initialize pools for small types (XXX is BPP small?)
    143      1.31  perseant 	 */
    144      1.46    simonb 	pool_init(&fs->lfs_clpool, sizeof(struct lfs_cluster), 0, 0, 0,
    145      1.68        ad 		"lfsclpl", &pool_allocator_nointr, IPL_NONE);
    146      1.46    simonb 	pool_init(&fs->lfs_segpool, sizeof(struct segment), 0, 0, 0,
    147      1.68        ad 		"lfssegpool", &pool_allocator_nointr, IPL_NONE);
    148  1.80.6.1     skrll 	maxbpp = ((lfs_sb_getsumsize(fs) - SEGSUM_SIZE(fs)) / sizeof(int32_t) + 2);
    149  1.80.6.1     skrll 	maxbpp = MIN(maxbpp, lfs_segsize(fs) / lfs_sb_getfsize(fs) + 2);
    150      1.46    simonb 	pool_init(&fs->lfs_bpppool, maxbpp * sizeof(struct buf *), 0, 0, 0,
    151      1.68        ad 		"lfsbpppl", &pool_allocator_nointr, IPL_NONE);
    152      1.31  perseant }
    153      1.31  perseant 
    154      1.31  perseant void
    155      1.31  perseant lfs_free_resblks(struct lfs *fs)
    156      1.31  perseant {
    157      1.31  perseant 	int i;
    158      1.31  perseant 
    159      1.31  perseant 	pool_destroy(&fs->lfs_bpppool);
    160      1.31  perseant 	pool_destroy(&fs->lfs_segpool);
    161      1.31  perseant 	pool_destroy(&fs->lfs_clpool);
    162      1.31  perseant 
    163      1.72        ad 	mutex_enter(&lfs_lock);
    164      1.31  perseant 	for (i = 0; i < LFS_N_TOTAL; i++) {
    165      1.39  perseant 		while (fs->lfs_resblk[i].inuse)
    166      1.72        ad 			mtsleep(&fs->lfs_resblk, PRIBIO + 1, "lfs_free", 0,
    167      1.72        ad 				&lfs_lock);
    168      1.31  perseant 		if (fs->lfs_resblk[i].p != NULL)
    169      1.31  perseant 			free(fs->lfs_resblk[i].p, M_SEGMENT);
    170      1.31  perseant 	}
    171      1.31  perseant 	free(fs->lfs_resblk, M_SEGMENT);
    172      1.72        ad 	mutex_exit(&lfs_lock);
    173      1.31  perseant }
    174      1.31  perseant 
    175      1.31  perseant static unsigned int
    176      1.31  perseant lfs_mhash(void *vp)
    177      1.31  perseant {
    178      1.31  perseant 	return (unsigned int)(((unsigned long)vp) >> 2) % LFS_RESHASH_WIDTH;
    179      1.31  perseant }
    180      1.31  perseant 
    181      1.31  perseant /*
    182      1.31  perseant  * Return memory of the given size for the given purpose, or use one of a
    183      1.31  perseant  * number of spare last-resort buffers, if malloc returns NULL.
    184      1.49     perry  */
    185      1.31  perseant void *
    186      1.31  perseant lfs_malloc(struct lfs *fs, size_t size, int type)
    187      1.31  perseant {
    188      1.31  perseant 	struct lfs_res_blk *re;
    189      1.31  perseant 	void *r;
    190      1.31  perseant 	int i, s, start;
    191      1.31  perseant 	unsigned int h;
    192      1.31  perseant 
    193      1.51  perseant 	ASSERT_MAYBE_SEGLOCK(fs);
    194      1.34  perseant 	r = NULL;
    195      1.34  perseant 
    196      1.31  perseant 	/* If no mem allocated for this type, it just waits */
    197      1.34  perseant 	if (lfs_res_qty[type] == 0) {
    198      1.34  perseant 		r = malloc(size, M_SEGMENT, M_WAITOK);
    199      1.34  perseant 		return r;
    200      1.34  perseant 	}
    201      1.31  perseant 
    202      1.31  perseant 	/* Otherwise try a quick malloc, and if it works, great */
    203      1.34  perseant 	if ((r = malloc(size, M_SEGMENT, M_NOWAIT)) != NULL) {
    204      1.31  perseant 		return r;
    205      1.34  perseant 	}
    206      1.31  perseant 
    207      1.31  perseant 	/*
    208      1.31  perseant 	 * If malloc returned NULL, we are forced to use one of our
    209      1.31  perseant 	 * reserve blocks.  We have on hand at least one summary block,
    210      1.31  perseant 	 * at least one cluster block, at least one superblock,
    211      1.31  perseant 	 * and several indirect blocks.
    212      1.31  perseant 	 */
    213      1.51  perseant 
    214      1.72        ad 	mutex_enter(&lfs_lock);
    215      1.31  perseant 	/* skip over blocks of other types */
    216      1.31  perseant 	for (i = 0, start = 0; i < type; i++)
    217      1.31  perseant 		start += lfs_res_qty[i];
    218      1.31  perseant 	while (r == NULL) {
    219      1.31  perseant 		for (i = 0; i < lfs_res_qty[type]; i++) {
    220      1.31  perseant 			if (fs->lfs_resblk[start + i].inuse == 0) {
    221      1.31  perseant 				re = fs->lfs_resblk + start + i;
    222      1.31  perseant 				re->inuse = 1;
    223      1.31  perseant 				r = re->p;
    224      1.34  perseant 				KASSERT(re->size >= size);
    225      1.31  perseant 				h = lfs_mhash(r);
    226      1.31  perseant 				s = splbio();
    227      1.31  perseant 				LIST_INSERT_HEAD(&fs->lfs_reshash[h], re, res);
    228      1.31  perseant 				splx(s);
    229      1.72        ad 				mutex_exit(&lfs_lock);
    230      1.31  perseant 				return r;
    231      1.31  perseant 			}
    232      1.31  perseant 		}
    233      1.51  perseant 		DLOG((DLOG_MALLOC, "sleeping on %s (%d)\n",
    234      1.51  perseant 		      lfs_res_names[type], lfs_res_qty[type]));
    235      1.72        ad 		mtsleep(&fs->lfs_resblk, PVM, "lfs_malloc", 0,
    236      1.72        ad 			&lfs_lock);
    237      1.51  perseant 		DLOG((DLOG_MALLOC, "done sleeping on %s\n",
    238      1.51  perseant 		      lfs_res_names[type]));
    239      1.31  perseant 	}
    240      1.31  perseant 	/* NOTREACHED */
    241      1.72        ad 	mutex_exit(&lfs_lock);
    242      1.31  perseant 	return r;
    243      1.31  perseant }
    244      1.31  perseant 
    245      1.31  perseant void
    246      1.65  christos lfs_free(struct lfs *fs, void *p, int type)
    247      1.31  perseant {
    248      1.31  perseant 	int s;
    249      1.31  perseant 	unsigned int h;
    250      1.31  perseant 	res_t *re;
    251      1.32      yamt #ifdef DEBUG
    252      1.32      yamt 	int i;
    253      1.32      yamt #endif
    254      1.31  perseant 
    255      1.51  perseant 	ASSERT_MAYBE_SEGLOCK(fs);
    256      1.31  perseant 	h = lfs_mhash(p);
    257      1.72        ad 	mutex_enter(&lfs_lock);
    258      1.31  perseant 	s = splbio();
    259      1.31  perseant 	LIST_FOREACH(re, &fs->lfs_reshash[h], res) {
    260      1.31  perseant 		if (re->p == p) {
    261      1.32      yamt 			KASSERT(re->inuse == 1);
    262      1.31  perseant 			LIST_REMOVE(re, res);
    263      1.31  perseant 			re->inuse = 0;
    264      1.31  perseant 			wakeup(&fs->lfs_resblk);
    265      1.31  perseant 			splx(s);
    266      1.72        ad 			mutex_exit(&lfs_lock);
    267      1.31  perseant 			return;
    268      1.31  perseant 		}
    269      1.31  perseant 	}
    270      1.32      yamt #ifdef DEBUG
    271      1.32      yamt 	for (i = 0; i < LFS_N_TOTAL; i++) {
    272      1.32      yamt 		if (fs->lfs_resblk[i].p == p)
    273      1.34  perseant 			panic("lfs_free: inconsistent reserved block");
    274      1.32      yamt 	}
    275      1.32      yamt #endif
    276      1.31  perseant 	splx(s);
    277      1.72        ad 	mutex_exit(&lfs_lock);
    278      1.51  perseant 
    279      1.31  perseant 	/*
    280      1.31  perseant 	 * If we didn't find it, free it.
    281      1.31  perseant 	 */
    282      1.31  perseant 	free(p, M_SEGMENT);
    283      1.31  perseant }
    284       1.1   mycroft 
    285       1.1   mycroft /*
    286       1.1   mycroft  * lfs_seglock --
    287       1.1   mycroft  *	Single thread the segment writer.
    288       1.1   mycroft  */
    289      1.31  perseant int
    290      1.18  perseant lfs_seglock(struct lfs *fs, unsigned long flags)
    291       1.1   mycroft {
    292       1.1   mycroft 	struct segment *sp;
    293      1.49     perry 
    294      1.72        ad 	mutex_enter(&lfs_lock);
    295       1.7   thorpej 	if (fs->lfs_seglock) {
    296      1.58  perseant 		if (fs->lfs_lockpid == curproc->p_pid &&
    297      1.58  perseant 		    fs->lfs_locklwp == curlwp->l_lid) {
    298       1.1   mycroft 			++fs->lfs_seglock;
    299       1.1   mycroft 			fs->lfs_sp->seg_flags |= flags;
    300      1.74   mlelstv 			mutex_exit(&lfs_lock);
    301      1.31  perseant 			return 0;
    302      1.38  perseant 		} else if (flags & SEGM_PAGEDAEMON) {
    303      1.72        ad 			mutex_exit(&lfs_lock);
    304      1.31  perseant 			return EWOULDBLOCK;
    305      1.51  perseant 		} else {
    306      1.51  perseant 			while (fs->lfs_seglock) {
    307      1.72        ad 				(void)mtsleep(&fs->lfs_seglock, PRIBIO + 1,
    308      1.72        ad 					"lfs_seglock", 0, &lfs_lock);
    309      1.51  perseant 			}
    310      1.51  perseant 		}
    311       1.7   thorpej 	}
    312      1.49     perry 
    313       1.1   mycroft 	fs->lfs_seglock = 1;
    314       1.1   mycroft 	fs->lfs_lockpid = curproc->p_pid;
    315      1.58  perseant 	fs->lfs_locklwp = curlwp->l_lid;
    316      1.72        ad 	mutex_exit(&lfs_lock);
    317      1.36  perseant 	fs->lfs_cleanind = 0;
    318      1.36  perseant 
    319      1.51  perseant #ifdef DEBUG
    320      1.51  perseant 	LFS_ENTER_LOG("seglock", __FILE__, __LINE__, 0, flags, curproc->p_pid);
    321      1.51  perseant #endif
    322      1.27  perseant 	/* Drain fragment size changes out */
    323      1.66        ad 	rw_enter(&fs->lfs_fraglock, RW_WRITER);
    324      1.27  perseant 
    325      1.31  perseant 	sp = fs->lfs_sp = pool_get(&fs->lfs_segpool, PR_WAITOK);
    326      1.31  perseant 	sp->bpp = pool_get(&fs->lfs_bpppool, PR_WAITOK);
    327       1.1   mycroft 	sp->seg_flags = flags;
    328       1.1   mycroft 	sp->vp = NULL;
    329      1.26  perseant 	sp->seg_iocount = 0;
    330       1.1   mycroft 	(void) lfs_initseg(fs);
    331      1.49     perry 
    332       1.1   mycroft 	/*
    333       1.1   mycroft 	 * Keep a cumulative count of the outstanding I/O operations.  If the
    334       1.1   mycroft 	 * disk drive catches up with us it could go to zero before we finish,
    335       1.1   mycroft 	 * so we artificially increment it by one until we've scheduled all of
    336       1.1   mycroft 	 * the writes we intend to do.
    337       1.1   mycroft 	 */
    338      1.72        ad 	mutex_enter(&lfs_lock);
    339       1.1   mycroft 	++fs->lfs_iocount;
    340  1.80.6.1     skrll 	fs->lfs_startseg = lfs_sb_getcurseg(fs);
    341      1.72        ad 	mutex_exit(&lfs_lock);
    342      1.31  perseant 	return 0;
    343       1.1   mycroft }
    344       1.8  perseant 
    345      1.31  perseant static void lfs_unmark_dirop(struct lfs *);
    346      1.31  perseant 
    347      1.31  perseant static void
    348      1.31  perseant lfs_unmark_dirop(struct lfs *fs)
    349      1.31  perseant {
    350      1.31  perseant 	struct inode *ip, *nip;
    351      1.31  perseant 	struct vnode *vp;
    352      1.40  perseant 	int doit;
    353      1.31  perseant 
    354      1.51  perseant 	ASSERT_NO_SEGLOCK(fs);
    355      1.72        ad 	mutex_enter(&lfs_lock);
    356      1.40  perseant 	doit = !(fs->lfs_flags & LFS_UNDIROP);
    357      1.40  perseant 	if (doit)
    358      1.40  perseant 		fs->lfs_flags |= LFS_UNDIROP;
    359      1.51  perseant 	if (!doit) {
    360      1.72        ad 		mutex_exit(&lfs_lock);
    361      1.40  perseant 		return;
    362      1.51  perseant 	}
    363      1.40  perseant 
    364      1.31  perseant 	for (ip = TAILQ_FIRST(&fs->lfs_dchainhd); ip != NULL; ip = nip) {
    365      1.31  perseant 		nip = TAILQ_NEXT(ip, i_lfs_dchain);
    366      1.31  perseant 		vp = ITOV(ip);
    367      1.77  perseant 		if ((ip->i_flag & (IN_ADIROP | IN_CDIROP)) == IN_CDIROP) {
    368      1.31  perseant 			--lfs_dirvcount;
    369      1.59  perseant 			--fs->lfs_dirvcount;
    370      1.71        ad 			vp->v_uflag &= ~VU_DIROP;
    371      1.31  perseant 			TAILQ_REMOVE(&fs->lfs_dchainhd, ip, i_lfs_dchain);
    372      1.31  perseant 			wakeup(&lfs_dirvcount);
    373      1.31  perseant 			fs->lfs_unlockvp = vp;
    374      1.72        ad 			mutex_exit(&lfs_lock);
    375      1.76   hannken 			vrele(vp);
    376      1.72        ad 			mutex_enter(&lfs_lock);
    377      1.31  perseant 			fs->lfs_unlockvp = NULL;
    378      1.77  perseant 			ip->i_flag &= ~IN_CDIROP;
    379      1.72        ad 		}
    380      1.31  perseant 	}
    381      1.40  perseant 
    382      1.40  perseant 	fs->lfs_flags &= ~LFS_UNDIROP;
    383      1.51  perseant 	wakeup(&fs->lfs_flags);
    384      1.72        ad 	mutex_exit(&lfs_lock);
    385      1.31  perseant }
    386      1.31  perseant 
    387      1.31  perseant static void
    388      1.31  perseant lfs_auto_segclean(struct lfs *fs)
    389      1.31  perseant {
    390      1.48  perseant 	int i, error, s, waited;
    391      1.31  perseant 
    392      1.51  perseant 	ASSERT_SEGLOCK(fs);
    393      1.31  perseant 	/*
    394      1.31  perseant 	 * Now that we've swapped lfs_activesb, but while we still
    395      1.31  perseant 	 * hold the segment lock, run through the segment list marking
    396      1.31  perseant 	 * the empty ones clean.
    397      1.31  perseant 	 * XXX - do we really need to do them all at once?
    398      1.31  perseant 	 */
    399      1.48  perseant 	waited = 0;
    400  1.80.6.1     skrll 	for (i = 0; i < lfs_sb_getnseg(fs); i++) {
    401      1.31  perseant 		if ((fs->lfs_suflags[0][i] &
    402      1.31  perseant 		     (SEGUSE_ACTIVE | SEGUSE_DIRTY | SEGUSE_EMPTY)) ==
    403      1.31  perseant 		    (SEGUSE_DIRTY | SEGUSE_EMPTY) &&
    404      1.31  perseant 		    (fs->lfs_suflags[1][i] &
    405      1.31  perseant 		     (SEGUSE_ACTIVE | SEGUSE_DIRTY | SEGUSE_EMPTY)) ==
    406      1.31  perseant 		    (SEGUSE_DIRTY | SEGUSE_EMPTY)) {
    407      1.31  perseant 
    408      1.48  perseant 			/* Make sure the sb is written before we clean */
    409      1.72        ad 			mutex_enter(&lfs_lock);
    410      1.48  perseant 			s = splbio();
    411      1.48  perseant 			while (waited == 0 && fs->lfs_sbactive)
    412      1.72        ad 				mtsleep(&fs->lfs_sbactive, PRIBIO+1, "lfs asb",
    413      1.72        ad 					0, &lfs_lock);
    414      1.48  perseant 			splx(s);
    415      1.72        ad 			mutex_exit(&lfs_lock);
    416      1.48  perseant 			waited = 1;
    417      1.48  perseant 
    418      1.31  perseant 			if ((error = lfs_do_segclean(fs, i)) != 0) {
    419      1.50  perseant 				DLOG((DLOG_CLEAN, "lfs_auto_segclean: lfs_do_segclean returned %d for seg %d\n", error, i));
    420      1.31  perseant 			}
    421      1.31  perseant 		}
    422      1.31  perseant 		fs->lfs_suflags[1 - fs->lfs_activesb][i] =
    423      1.31  perseant 			fs->lfs_suflags[fs->lfs_activesb][i];
    424      1.31  perseant 	}
    425      1.31  perseant }
    426      1.31  perseant 
    427       1.1   mycroft /*
    428       1.1   mycroft  * lfs_segunlock --
    429       1.1   mycroft  *	Single thread the segment writer.
    430       1.1   mycroft  */
    431       1.1   mycroft void
    432      1.18  perseant lfs_segunlock(struct lfs *fs)
    433       1.1   mycroft {
    434       1.1   mycroft 	struct segment *sp;
    435       1.1   mycroft 	unsigned long sync, ckp;
    436      1.22  perseant 	struct buf *bp;
    437      1.40  perseant 	int do_unmark_dirop = 0;
    438      1.49     perry 
    439      1.15  perseant 	sp = fs->lfs_sp;
    440      1.15  perseant 
    441      1.72        ad 	mutex_enter(&lfs_lock);
    442      1.72        ad 	KASSERT(LFS_SEGLOCK_HELD(fs));
    443      1.15  perseant 	if (fs->lfs_seglock == 1) {
    444      1.77  perseant 		if ((sp->seg_flags & (SEGM_PROT | SEGM_CLEAN)) == 0)
    445      1.40  perseant 			do_unmark_dirop = 1;
    446      1.72        ad 		mutex_exit(&lfs_lock);
    447       1.1   mycroft 		sync = sp->seg_flags & SEGM_SYNC;
    448       1.1   mycroft 		ckp = sp->seg_flags & SEGM_CKP;
    449      1.57  perseant 
    450      1.57  perseant 		/* We should have a segment summary, and nothing else */
    451      1.57  perseant 		KASSERT(sp->cbpp == sp->bpp + 1);
    452      1.57  perseant 
    453      1.57  perseant 		/* Free allocated segment summary */
    454  1.80.6.1     skrll 		lfs_sb_suboffset(fs, lfs_btofsb(fs, lfs_sb_getsumsize(fs)));
    455      1.57  perseant 		bp = *sp->bpp;
    456      1.57  perseant 		lfs_freebuf(fs, bp);
    457       1.8  perseant 
    458      1.31  perseant 		pool_put(&fs->lfs_bpppool, sp->bpp);
    459      1.18  perseant 		sp->bpp = NULL;
    460      1.36  perseant 
    461      1.36  perseant 		/*
    462      1.36  perseant 		 * If we're not sync, we're done with sp, get rid of it.
    463      1.36  perseant 		 * Otherwise, we keep a local copy around but free
    464      1.36  perseant 		 * fs->lfs_sp so another process can use it (we have to
    465      1.36  perseant 		 * wait but they don't have to wait for us).
    466      1.36  perseant 		 */
    467      1.26  perseant 		if (!sync)
    468      1.31  perseant 			pool_put(&fs->lfs_segpool, sp);
    469      1.18  perseant 		fs->lfs_sp = NULL;
    470       1.1   mycroft 
    471       1.1   mycroft 		/*
    472       1.1   mycroft 		 * If the I/O count is non-zero, sleep until it reaches zero.
    473       1.1   mycroft 		 * At the moment, the user's process hangs around so we can
    474       1.1   mycroft 		 * sleep.
    475       1.1   mycroft 		 */
    476      1.72        ad 		mutex_enter(&lfs_lock);
    477      1.63  christos 		if (--fs->lfs_iocount == 0) {
    478      1.44      yamt 			LFS_DEBUG_COUNTLOCKED("lfs_segunlock");
    479      1.63  christos 		}
    480      1.36  perseant 		if (fs->lfs_iocount <= 1)
    481      1.22  perseant 			wakeup(&fs->lfs_iocount);
    482      1.72        ad 		mutex_exit(&lfs_lock);
    483       1.1   mycroft 		/*
    484      1.26  perseant 		 * If we're not checkpointing, we don't have to block
    485      1.26  perseant 		 * other processes to wait for a synchronous write
    486      1.26  perseant 		 * to complete.
    487      1.26  perseant 		 */
    488      1.26  perseant 		if (!ckp) {
    489      1.51  perseant #ifdef DEBUG
    490      1.51  perseant 			LFS_ENTER_LOG("segunlock_std", __FILE__, __LINE__, 0, 0, curproc->p_pid);
    491      1.51  perseant #endif
    492      1.72        ad 			mutex_enter(&lfs_lock);
    493      1.26  perseant 			--fs->lfs_seglock;
    494      1.26  perseant 			fs->lfs_lockpid = 0;
    495      1.58  perseant 			fs->lfs_locklwp = 0;
    496      1.72        ad 			mutex_exit(&lfs_lock);
    497      1.26  perseant 			wakeup(&fs->lfs_seglock);
    498      1.26  perseant 		}
    499      1.26  perseant 		/*
    500       1.1   mycroft 		 * We let checkpoints happen asynchronously.  That means
    501       1.1   mycroft 		 * that during recovery, we have to roll forward between
    502       1.1   mycroft 		 * the two segments described by the first and second
    503       1.1   mycroft 		 * superblocks to make sure that the checkpoint described
    504       1.1   mycroft 		 * by a superblock completed.
    505       1.1   mycroft 		 */
    506      1.72        ad 		mutex_enter(&lfs_lock);
    507      1.74   mlelstv 		while (ckp && sync && fs->lfs_iocount) {
    508      1.72        ad 			(void)mtsleep(&fs->lfs_iocount, PRIBIO + 1,
    509      1.72        ad 				      "lfs_iocount", 0, &lfs_lock);
    510      1.74   mlelstv 			DLOG((DLOG_SEG, "sleeping on iocount %x == %d\n", fs, fs->lfs_iocount));
    511      1.74   mlelstv 		}
    512      1.26  perseant 		while (sync && sp->seg_iocount) {
    513      1.72        ad 			(void)mtsleep(&sp->seg_iocount, PRIBIO + 1,
    514      1.72        ad 				     "seg_iocount", 0, &lfs_lock);
    515      1.50  perseant 			DLOG((DLOG_SEG, "sleeping on iocount %x == %d\n", sp, sp->seg_iocount));
    516      1.26  perseant 		}
    517      1.72        ad 		mutex_exit(&lfs_lock);
    518      1.26  perseant 		if (sync)
    519      1.31  perseant 			pool_put(&fs->lfs_segpool, sp);
    520      1.36  perseant 
    521       1.1   mycroft 		if (ckp) {
    522       1.1   mycroft 			fs->lfs_nactive = 0;
    523       1.8  perseant 			/* If we *know* everything's on disk, write both sbs */
    524      1.33  perseant 			/* XXX should wait for this one	 */
    525      1.21       chs 			if (sync)
    526  1.80.6.1     skrll 				lfs_writesuper(fs, lfs_sb_getsboff(fs, fs->lfs_activesb));
    527  1.80.6.1     skrll 			lfs_writesuper(fs, lfs_sb_getsboff(fs, 1 - fs->lfs_activesb));
    528      1.48  perseant 			if (!(fs->lfs_ivnode->v_mount->mnt_iflag & IMNT_UNMOUNT)) {
    529      1.35  perseant 				lfs_auto_segclean(fs);
    530      1.48  perseant 				/* If sync, we can clean the remainder too */
    531      1.48  perseant 				if (sync)
    532      1.48  perseant 					lfs_auto_segclean(fs);
    533      1.48  perseant 			}
    534       1.8  perseant 			fs->lfs_activesb = 1 - fs->lfs_activesb;
    535      1.51  perseant #ifdef DEBUG
    536      1.51  perseant 			LFS_ENTER_LOG("segunlock_ckp", __FILE__, __LINE__, 0, 0, curproc->p_pid);
    537      1.51  perseant #endif
    538      1.72        ad 			mutex_enter(&lfs_lock);
    539      1.26  perseant 			--fs->lfs_seglock;
    540      1.26  perseant 			fs->lfs_lockpid = 0;
    541      1.58  perseant 			fs->lfs_locklwp = 0;
    542      1.72        ad 			mutex_exit(&lfs_lock);
    543      1.26  perseant 			wakeup(&fs->lfs_seglock);
    544       1.1   mycroft 		}
    545      1.27  perseant 		/* Reenable fragment size changes */
    546      1.66        ad 		rw_exit(&fs->lfs_fraglock);
    547      1.40  perseant 		if (do_unmark_dirop)
    548      1.40  perseant 			lfs_unmark_dirop(fs);
    549       1.1   mycroft 	} else if (fs->lfs_seglock == 0) {
    550      1.72        ad 		mutex_exit(&lfs_lock);
    551       1.1   mycroft 		panic ("Seglock not held");
    552       1.1   mycroft 	} else {
    553       1.1   mycroft 		--fs->lfs_seglock;
    554      1.72        ad 		mutex_exit(&lfs_lock);
    555       1.1   mycroft 	}
    556      1.41      yamt }
    557      1.41      yamt 
    558      1.41      yamt /*
    559      1.69  perseant  * Drain dirops and start writer.
    560      1.69  perseant  *
    561      1.69  perseant  * No simple_locks are held when we enter and none are held when we return.
    562      1.41      yamt  */
    563      1.41      yamt int
    564      1.41      yamt lfs_writer_enter(struct lfs *fs, const char *wmesg)
    565      1.41      yamt {
    566      1.41      yamt 	int error = 0;
    567      1.41      yamt 
    568      1.51  perseant 	ASSERT_MAYBE_SEGLOCK(fs);
    569      1.72        ad 	mutex_enter(&lfs_lock);
    570      1.41      yamt 
    571      1.41      yamt 	/* disallow dirops during flush */
    572      1.41      yamt 	fs->lfs_writer++;
    573      1.41      yamt 
    574      1.41      yamt 	while (fs->lfs_dirops > 0) {
    575      1.49     perry 		++fs->lfs_diropwait;
    576      1.72        ad 		error = mtsleep(&fs->lfs_writer, PRIBIO+1, wmesg, 0,
    577      1.72        ad 				&lfs_lock);
    578      1.49     perry 		--fs->lfs_diropwait;
    579      1.41      yamt 	}
    580      1.41      yamt 
    581      1.41      yamt 	if (error)
    582      1.41      yamt 		fs->lfs_writer--;
    583      1.41      yamt 
    584      1.72        ad 	mutex_exit(&lfs_lock);
    585      1.41      yamt 
    586      1.41      yamt 	return error;
    587      1.41      yamt }
    588      1.41      yamt 
    589      1.41      yamt void
    590      1.41      yamt lfs_writer_leave(struct lfs *fs)
    591      1.41      yamt {
    592      1.67   thorpej 	bool dowakeup;
    593      1.41      yamt 
    594      1.51  perseant 	ASSERT_MAYBE_SEGLOCK(fs);
    595      1.72        ad 	mutex_enter(&lfs_lock);
    596      1.41      yamt 	dowakeup = !(--fs->lfs_writer);
    597      1.72        ad 	mutex_exit(&lfs_lock);
    598      1.41      yamt 	if (dowakeup)
    599      1.41      yamt 		wakeup(&fs->lfs_dirops);
    600       1.1   mycroft }
    601      1.57  perseant 
    602      1.57  perseant /*
    603      1.57  perseant  * Unlock, wait for the cleaner, then relock to where we were before.
    604      1.57  perseant  * To be used only at a fairly high level, to address a paucity of free
    605      1.57  perseant  * segments propagated back from lfs_gop_write().
    606      1.57  perseant  */
    607      1.57  perseant void
    608      1.57  perseant lfs_segunlock_relock(struct lfs *fs)
    609      1.57  perseant {
    610      1.57  perseant 	int n = fs->lfs_seglock;
    611      1.57  perseant 	u_int16_t seg_flags;
    612      1.61  perseant 	CLEANERINFO *cip;
    613      1.61  perseant 	struct buf *bp;
    614      1.57  perseant 
    615      1.57  perseant 	if (n == 0)
    616      1.57  perseant 		return;
    617      1.57  perseant 
    618      1.57  perseant 	/* Write anything we've already gathered to disk */
    619      1.57  perseant 	lfs_writeseg(fs, fs->lfs_sp);
    620      1.57  perseant 
    621      1.61  perseant 	/* Tell cleaner */
    622      1.61  perseant 	LFS_CLEANERINFO(cip, fs, bp);
    623  1.80.6.1     skrll 	lfs_ci_setflags(fs, cip,
    624  1.80.6.1     skrll 			lfs_ci_getflags(fs, cip) | LFS_CLEANER_MUST_CLEAN);
    625      1.61  perseant 	LFS_SYNC_CLEANERINFO(cip, fs, bp, 1);
    626      1.61  perseant 
    627      1.57  perseant 	/* Save segment flags for later */
    628      1.57  perseant 	seg_flags = fs->lfs_sp->seg_flags;
    629      1.57  perseant 
    630      1.57  perseant 	fs->lfs_sp->seg_flags |= SEGM_PROT; /* Don't unmark dirop nodes */
    631      1.57  perseant 	while(fs->lfs_seglock)
    632      1.57  perseant 		lfs_segunlock(fs);
    633      1.57  perseant 
    634      1.57  perseant 	/* Wait for the cleaner */
    635      1.60  perseant 	lfs_wakeup_cleaner(fs);
    636      1.72        ad 	mutex_enter(&lfs_lock);
    637      1.57  perseant 	while (LFS_STARVED_FOR_SEGS(fs))
    638  1.80.6.1     skrll 		mtsleep(&fs->lfs_availsleep, PRIBIO, "relock", 0,
    639      1.72        ad 			&lfs_lock);
    640      1.72        ad 	mutex_exit(&lfs_lock);
    641      1.57  perseant 
    642      1.57  perseant 	/* Put the segment lock back the way it was. */
    643      1.57  perseant 	while(n--)
    644      1.57  perseant 		lfs_seglock(fs, seg_flags);
    645      1.57  perseant 
    646      1.61  perseant 	/* Cleaner can relax now */
    647      1.61  perseant 	LFS_CLEANERINFO(cip, fs, bp);
    648  1.80.6.1     skrll 	lfs_ci_setflags(fs, cip,
    649  1.80.6.1     skrll 			lfs_ci_getflags(fs, cip) & ~LFS_CLEANER_MUST_CLEAN);
    650      1.61  perseant 	LFS_SYNC_CLEANERINFO(cip, fs, bp, 1);
    651      1.61  perseant 
    652      1.57  perseant 	return;
    653      1.57  perseant }
    654      1.60  perseant 
    655      1.60  perseant /*
    656      1.60  perseant  * Wake up the cleaner, provided that nowrap is not set.
    657      1.60  perseant  */
    658      1.60  perseant void
    659      1.60  perseant lfs_wakeup_cleaner(struct lfs *fs)
    660      1.60  perseant {
    661      1.60  perseant 	if (fs->lfs_nowrap > 0)
    662      1.60  perseant 		return;
    663      1.60  perseant 
    664  1.80.6.1     skrll 	wakeup(&fs->lfs_nextsegsleep);
    665      1.60  perseant 	wakeup(&lfs_allclean_wakeup);
    666      1.60  perseant }
    667