1 /*-
2 * modified for Lites 1.1
3 *
4 * Aug 1995, Godmar Back (gback@cs.utah.edu)
5 * University of Utah, Department of Computer Science
6 */
7 /*-
8 * Copyright (c) 1982, 1986, 1989, 1993
9 * The Regents of the University of California. All rights reserved.
10 *
11 * Redistribution and use in source and binary forms, with or without
12 * modification, are permitted provided that the following conditions
13 * are met:
14 * 1. Redistributions of source code must retain the above copyright
15 * notice, this list of conditions and the following disclaimer.
16 * 2. Redistributions in binary form must reproduce the above copyright
17 * notice, this list of conditions and the following disclaimer in the
18 * documentation and/or other materials provided with the distribution.
19 * 4. Neither the name of the University nor the names of its contributors
20 * may be used to endorse or promote products derived from this software
21 * without specific prior written permission.
22 *
23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * SUCH DAMAGE.
34 *
35 * @(#)ffs_alloc.c 8.8 (Berkeley) 2/21/94
36 * $FreeBSD$
37 */
38
39 #include <sys/param.h>
40 #include <sys/systm.h>
41 #include <sys/conf.h>
42 #include <sys/vnode.h>
43 #include <sys/stat.h>
44 #include <sys/mount.h>
45 #include <sys/sysctl.h>
46 #include <sys/syslog.h>
47 #include <sys/buf.h>
48
49 #include <fs/ext2fs/fs.h>
50 #include <fs/ext2fs/inode.h>
51 #include <fs/ext2fs/ext2_mount.h>
52 #include <fs/ext2fs/ext2fs.h>
53 #include <fs/ext2fs/ext2_extern.h>
54
55 static daddr_t ext2_alloccg(struct inode *, int, daddr_t, int);
56 static daddr_t ext2_clusteralloc(struct inode *, int, daddr_t, int);
57 static u_long ext2_dirpref(struct inode *);
58 static void ext2_fserr(struct m_ext2fs *, uid_t, char *);
59 static u_long ext2_hashalloc(struct inode *, int, long, int,
60 daddr_t (*)(struct inode *, int, daddr_t,
61 int));
62 static daddr_t ext2_nodealloccg(struct inode *, int, daddr_t, int);
63 static daddr_t ext2_mapsearch(struct m_ext2fs *, char *, daddr_t);
64
65 /*
66 * Allocate a block in the filesystem.
67 *
68 * A preference may be optionally specified. If a preference is given
69 * the following hierarchy is used to allocate a block:
70 * 1) allocate the requested block.
71 * 2) allocate a rotationally optimal block in the same cylinder.
72 * 3) allocate a block in the same cylinder group.
73 * 4) quadradically rehash into other cylinder groups, until an
74 * available block is located.
75 * If no block preference is given the following hierarchy is used
76 * to allocate a block:
77 * 1) allocate a block in the cylinder group that contains the
78 * inode for the file.
79 * 2) quadradically rehash into other cylinder groups, until an
80 * available block is located.
81 */
82 int
83 ext2_alloc(struct inode *ip, daddr_t lbn, e4fs_daddr_t bpref, int size,
84 struct ucred *cred, e4fs_daddr_t *bnp)
85 {
86 struct m_ext2fs *fs;
87 struct ext2mount *ump;
88 int32_t bno;
89 int cg;
90
91 *bnp = 0;
92 fs = ip->i_e2fs;
93 ump = ip->i_ump;
94 mtx_assert(EXT2_MTX(ump), MA_OWNED);
95 #ifdef INVARIANTS
96 if ((u_int)size > fs->e2fs_bsize || blkoff(fs, size) != 0) {
97 vn_printf(ip->i_devvp, "bsize = %lu, size = %d, fs = %s\n",
98 (long unsigned int)fs->e2fs_bsize, size, fs->e2fs_fsmnt);
99 panic("ext2_alloc: bad size");
100 }
101 if (cred == NOCRED)
102 panic("ext2_alloc: missing credential");
103 #endif /* INVARIANTS */
104 if (size == fs->e2fs_bsize && fs->e2fs->e2fs_fbcount == 0)
105 goto nospace;
106 if (cred->cr_uid != 0 &&
107 fs->e2fs->e2fs_fbcount < fs->e2fs->e2fs_rbcount)
108 goto nospace;
109 if (bpref >= fs->e2fs->e2fs_bcount)
110 bpref = 0;
111 if (bpref == 0)
112 cg = ino_to_cg(fs, ip->i_number);
113 else
114 cg = dtog(fs, bpref);
115 bno = (daddr_t)ext2_hashalloc(ip, cg, bpref, fs->e2fs_bsize,
116 ext2_alloccg);
117 if (bno > 0) {
118 /* set next_alloc fields as done in block_getblk */
119 ip->i_next_alloc_block = lbn;
120 ip->i_next_alloc_goal = bno;
121
122 ip->i_blocks += btodb(fs->e2fs_bsize);
123 ip->i_flag |= IN_CHANGE | IN_UPDATE;
124 *bnp = bno;
125 return (0);
126 }
127 nospace:
128 EXT2_UNLOCK(ump);
129 ext2_fserr(fs, cred->cr_uid, "filesystem full");
130 uprintf("\n%s: write failed, filesystem is full\n", fs->e2fs_fsmnt);
131 return (ENOSPC);
132 }
133
134 /*
135 * Reallocate a sequence of blocks into a contiguous sequence of blocks.
136 *
137 * The vnode and an array of buffer pointers for a range of sequential
138 * logical blocks to be made contiguous is given. The allocator attempts
139 * to find a range of sequential blocks starting as close as possible to
140 * an fs_rotdelay offset from the end of the allocation for the logical
141 * block immediately preceding the current range. If successful, the
142 * physical block numbers in the buffer pointers and in the inode are
143 * changed to reflect the new allocation. If unsuccessful, the allocation
144 * is left unchanged. The success in doing the reallocation is returned.
145 * Note that the error return is not reflected back to the user. Rather
146 * the previous block allocation will be used.
147 */
148
149 static SYSCTL_NODE(_vfs, OID_AUTO, ext2fs, CTLFLAG_RW, 0, "EXT2FS filesystem");
150
151 static int doasyncfree = 1;
152
153 SYSCTL_INT(_vfs_ext2fs, OID_AUTO, doasyncfree, CTLFLAG_RW, &doasyncfree, 0,
154 "Use asychronous writes to update block pointers when freeing blocks");
155
156 static int doreallocblks = 0;
157
158 SYSCTL_INT(_vfs_ext2fs, OID_AUTO, doreallocblks, CTLFLAG_RW, &doreallocblks, 0, "");
159
160 int
161 ext2_reallocblks(struct vop_reallocblks_args *ap)
162 {
163 struct m_ext2fs *fs;
164 struct inode *ip;
165 struct vnode *vp;
166 struct buf *sbp, *ebp;
167 uint32_t *bap, *sbap, *ebap;
168 struct ext2mount *ump;
169 struct cluster_save *buflist;
170 struct indir start_ap[NIADDR + 1], end_ap[NIADDR + 1], *idp;
171 e2fs_lbn_t start_lbn, end_lbn;
172 int soff;
173 e2fs_daddr_t newblk, blkno;
174 int i, len, start_lvl, end_lvl, pref, ssize;
175
176 if (doreallocblks == 0)
177 return (ENOSPC);
178
179 vp = ap->a_vp;
180 ip = VTOI(vp);
181 fs = ip->i_e2fs;
182 ump = ip->i_ump;
183
184 if (fs->e2fs_contigsumsize <= 0)
185 return (ENOSPC);
186
187 buflist = ap->a_buflist;
188 len = buflist->bs_nchildren;
189 start_lbn = buflist->bs_children[0]->b_lblkno;
190 end_lbn = start_lbn + len - 1;
191 #ifdef INVARIANTS
192 for (i = 1; i < len; i++)
193 if (buflist->bs_children[i]->b_lblkno != start_lbn + i)
194 panic("ext2_reallocblks: non-cluster");
195 #endif
196 /*
197 * If the cluster crosses the boundary for the first indirect
198 * block, leave space for the indirect block. Indirect blocks
199 * are initially laid out in a position after the last direct
200 * block. Block reallocation would usually destroy locality by
201 * moving the indirect block out of the way to make room for
202 * data blocks if we didn't compensate here. We should also do
203 * this for other indirect block boundaries, but it is only
204 * important for the first one.
205 */
206 if (start_lbn < NDADDR && end_lbn >= NDADDR)
207 return (ENOSPC);
208 /*
209 * If the latest allocation is in a new cylinder group, assume that
210 * the filesystem has decided to move and do not force it back to
211 * the previous cylinder group.
212 */
213 if (dtog(fs, dbtofsb(fs, buflist->bs_children[0]->b_blkno)) !=
214 dtog(fs, dbtofsb(fs, buflist->bs_children[len - 1]->b_blkno)))
215 return (ENOSPC);
216 if (ext2_getlbns(vp, start_lbn, start_ap, &start_lvl) ||
217 ext2_getlbns(vp, end_lbn, end_ap, &end_lvl))
218 return (ENOSPC);
219 /*
220 * Get the starting offset and block map for the first block.
221 */
222 if (start_lvl == 0) {
223 sbap = &ip->i_db[0];
224 soff = start_lbn;
225 } else {
226 idp = &start_ap[start_lvl - 1];
227 if (bread(vp, idp->in_lbn, (int)fs->e2fs_bsize, NOCRED, &sbp)) {
228 brelse(sbp);
229 return (ENOSPC);
230 }
231 sbap = (u_int *)sbp->b_data;
232 soff = idp->in_off;
233 }
234 /*
235 * If the block range spans two block maps, get the second map.
236 */
237 ebap = NULL;
238 if (end_lvl == 0 || (idp = &end_ap[end_lvl - 1])->in_off + 1 >= len) {
239 ssize = len;
240 } else {
241 #ifdef INVARIANTS
242 if (start_ap[start_lvl - 1].in_lbn == idp->in_lbn)
243 panic("ext2_reallocblks: start == end");
244 #endif
245 ssize = len - (idp->in_off + 1);
246 if (bread(vp, idp->in_lbn, (int)fs->e2fs_bsize, NOCRED, &ebp))
247 goto fail;
248 ebap = (u_int *)ebp->b_data;
249 }
250 /*
251 * Find the preferred location for the cluster.
252 */
253 EXT2_LOCK(ump);
254 pref = ext2_blkpref(ip, start_lbn, soff, sbap, 0);
255 /*
256 * Search the block map looking for an allocation of the desired size.
257 */
258 if ((newblk = (e2fs_daddr_t)ext2_hashalloc(ip, dtog(fs, pref), pref,
259 len, ext2_clusteralloc)) == 0) {
260 EXT2_UNLOCK(ump);
261 goto fail;
262 }
263 /*
264 * We have found a new contiguous block.
265 *
266 * First we have to replace the old block pointers with the new
267 * block pointers in the inode and indirect blocks associated
268 * with the file.
269 */
270 #ifdef DEBUG
271 printf("realloc: ino %d, lbns %jd-%jd\n\told:", ip->i_number,
272 (intmax_t)start_lbn, (intmax_t)end_lbn);
273 #endif /* DEBUG */
274 blkno = newblk;
275 for (bap = &sbap[soff], i = 0; i < len; i++, blkno += fs->e2fs_fpb) {
276 if (i == ssize) {
277 bap = ebap;
278 soff = -i;
279 }
280 #ifdef INVARIANTS
281 if (buflist->bs_children[i]->b_blkno != fsbtodb(fs, *bap))
282 panic("ext2_reallocblks: alloc mismatch");
283 #endif
284 #ifdef DEBUG
285 printf(" %d,", *bap);
286 #endif /* DEBUG */
287 *bap++ = blkno;
288 }
289 /*
290 * Next we must write out the modified inode and indirect blocks.
291 * For strict correctness, the writes should be synchronous since
292 * the old block values may have been written to disk. In practise
293 * they are almost never written, but if we are concerned about
294 * strict correctness, the `doasyncfree' flag should be set to zero.
295 *
296 * The test on `doasyncfree' should be changed to test a flag
297 * that shows whether the associated buffers and inodes have
298 * been written. The flag should be set when the cluster is
299 * started and cleared whenever the buffer or inode is flushed.
300 * We can then check below to see if it is set, and do the
301 * synchronous write only when it has been cleared.
302 */
303 if (sbap != &ip->i_db[0]) {
304 if (doasyncfree)
305 bdwrite(sbp);
306 else
307 bwrite(sbp);
308 } else {
309 ip->i_flag |= IN_CHANGE | IN_UPDATE;
310 if (!doasyncfree)
311 ext2_update(vp, 1);
312 }
313 if (ssize < len) {
314 if (doasyncfree)
315 bdwrite(ebp);
316 else
317 bwrite(ebp);
318 }
319 /*
320 * Last, free the old blocks and assign the new blocks to the buffers.
321 */
322 #ifdef DEBUG
323 printf("\n\tnew:");
324 #endif /* DEBUG */
325 for (blkno = newblk, i = 0; i < len; i++, blkno += fs->e2fs_fpb) {
326 ext2_blkfree(ip, dbtofsb(fs, buflist->bs_children[i]->b_blkno),
327 fs->e2fs_bsize);
328 buflist->bs_children[i]->b_blkno = fsbtodb(fs, blkno);
329 #ifdef DEBUG
330 printf(" %d,", blkno);
331 #endif /* DEBUG */
332 }
333 #ifdef DEBUG
334 printf("\n");
335 #endif /* DEBUG */
336 return (0);
337
338 fail:
339 if (ssize < len)
340 brelse(ebp);
341 if (sbap != &ip->i_db[0])
342 brelse(sbp);
343 return (ENOSPC);
344 }
345
346 /*
347 * Allocate an inode in the filesystem.
348 *
349 */
350 int
351 ext2_valloc(struct vnode *pvp, int mode, struct ucred *cred, struct vnode **vpp)
352 {
353 struct timespec ts;
354 struct inode *pip;
355 struct m_ext2fs *fs;
356 struct inode *ip;
357 struct ext2mount *ump;
358 ino_t ino, ipref;
359 int i, error, cg;
360
361 *vpp = NULL;
362 pip = VTOI(pvp);
363 fs = pip->i_e2fs;
364 ump = pip->i_ump;
365
366 EXT2_LOCK(ump);
367 if (fs->e2fs->e2fs_ficount == 0)
368 goto noinodes;
369 /*
370 * If it is a directory then obtain a cylinder group based on
371 * ext2_dirpref else obtain it using ino_to_cg. The preferred inode is
372 * always the next inode.
373 */
374 if ((mode & IFMT) == IFDIR) {
375 cg = ext2_dirpref(pip);
376 if (fs->e2fs_contigdirs[cg] < 255)
377 fs->e2fs_contigdirs[cg]++;
378 } else {
379 cg = ino_to_cg(fs, pip->i_number);
380 if (fs->e2fs_contigdirs[cg] > 0)
381 fs->e2fs_contigdirs[cg]--;
382 }
383 ipref = cg * fs->e2fs->e2fs_ipg + 1;
384 ino = (ino_t)ext2_hashalloc(pip, cg, (long)ipref, mode, ext2_nodealloccg);
385
386 if (ino == 0)
387 goto noinodes;
388 error = VFS_VGET(pvp->v_mount, ino, LK_EXCLUSIVE, vpp);
389 if (error) {
390 ext2_vfree(pvp, ino, mode);
391 return (error);
392 }
393 ip = VTOI(*vpp);
394
395 /*
396 * The question is whether using VGET was such good idea at all:
397 * Linux doesn't read the old inode in when it is allocating a
398 * new one. I will set at least i_size and i_blocks to zero.
399 */
400 ip->i_flag = 0;
401 ip->i_size = 0;
402 ip->i_blocks = 0;
403 ip->i_mode = 0;
404 ip->i_flags = 0;
405 /* now we want to make sure that the block pointers are zeroed out */
406 for (i = 0; i < NDADDR; i++)
407 ip->i_db[i] = 0;
408 for (i = 0; i < NIADDR; i++)
409 ip->i_ib[i] = 0;
410
411 /*
412 * Set up a new generation number for this inode.
413 */
414 ip->i_gen = arc4random();
415
416 vfs_timestamp(&ts);
417 ip->i_birthtime = ts.tv_sec;
418 ip->i_birthnsec = ts.tv_nsec;
419
420 /*
421 printf("ext2_valloc: allocated inode %d\n", ino);
422 */
423 return (0);
424 noinodes:
425 EXT2_UNLOCK(ump);
426 ext2_fserr(fs, cred->cr_uid, "out of inodes");
427 uprintf("\n%s: create/symlink failed, no inodes free\n", fs->e2fs_fsmnt);
428 return (ENOSPC);
429 }
430
431 /*
432 * Find a cylinder to place a directory.
433 *
434 * The policy implemented by this algorithm is to allocate a
435 * directory inode in the same cylinder group as its parent
436 * directory, but also to reserve space for its files inodes
437 * and data. Restrict the number of directories which may be
438 * allocated one after another in the same cylinder group
439 * without intervening allocation of files.
440 *
441 * If we allocate a first level directory then force allocation
442 * in another cylinder group.
443 *
444 */
445 static u_long
446 ext2_dirpref(struct inode *pip)
447 {
448 struct m_ext2fs *fs;
449 int cg, prefcg, cgsize;
450 u_int avgifree, avgbfree, avgndir, curdirsize;
451 u_int minifree, minbfree, maxndir;
452 u_int mincg, minndir;
453 u_int dirsize, maxcontigdirs;
454
455 mtx_assert(EXT2_MTX(pip->i_ump), MA_OWNED);
456 fs = pip->i_e2fs;
457
458 avgifree = fs->e2fs->e2fs_ficount / fs->e2fs_gcount;
459 avgbfree = fs->e2fs->e2fs_fbcount / fs->e2fs_gcount;
460 avgndir = fs->e2fs_total_dir / fs->e2fs_gcount;
461
462 /*
463 * Force allocation in another cg if creating a first level dir.
464 */
465 ASSERT_VOP_LOCKED(ITOV(pip), "ext2fs_dirpref");
466 if (ITOV(pip)->v_vflag & VV_ROOT) {
467 prefcg = arc4random() % fs->e2fs_gcount;
468 mincg = prefcg;
469 minndir = fs->e2fs_ipg;
470 for (cg = prefcg; cg < fs->e2fs_gcount; cg++)
471 if (fs->e2fs_gd[cg].ext2bgd_ndirs < minndir &&
472 fs->e2fs_gd[cg].ext2bgd_nifree >= avgifree &&
473 fs->e2fs_gd[cg].ext2bgd_nbfree >= avgbfree) {
474 mincg = cg;
475 minndir = fs->e2fs_gd[cg].ext2bgd_ndirs;
476 }
477 for (cg = 0; cg < prefcg; cg++)
478 if (fs->e2fs_gd[cg].ext2bgd_ndirs < minndir &&
479 fs->e2fs_gd[cg].ext2bgd_nifree >= avgifree &&
480 fs->e2fs_gd[cg].ext2bgd_nbfree >= avgbfree) {
481 mincg = cg;
482 minndir = fs->e2fs_gd[cg].ext2bgd_ndirs;
483 }
484 return (mincg);
485 }
486 /*
487 * Count various limits which used for
488 * optimal allocation of a directory inode.
489 */
490 maxndir = min(avgndir + fs->e2fs_ipg / 16, fs->e2fs_ipg);
491 minifree = avgifree - avgifree / 4;
492 if (minifree < 1)
493 minifree = 1;
494 minbfree = avgbfree - avgbfree / 4;
495 if (minbfree < 1)
496 minbfree = 1;
497 cgsize = fs->e2fs_fsize * fs->e2fs_fpg;
498 dirsize = AVGDIRSIZE;
499 curdirsize = avgndir ? (cgsize - avgbfree * fs->e2fs_bsize) / avgndir : 0;
500 if (dirsize < curdirsize)
501 dirsize = curdirsize;
502 maxcontigdirs = min((avgbfree * fs->e2fs_bsize) / dirsize, 255);
503 maxcontigdirs = min(maxcontigdirs, fs->e2fs_ipg / AFPDIR);
504 if (maxcontigdirs == 0)
505 maxcontigdirs = 1;
506
507 /*
508 * Limit number of dirs in one cg and reserve space for
509 * regular files, but only if we have no deficit in
510 * inodes or space.
511 */
512 prefcg = ino_to_cg(fs, pip->i_number);
513 for (cg = prefcg; cg < fs->e2fs_gcount; cg++)
514 if (fs->e2fs_gd[cg].ext2bgd_ndirs < maxndir &&
515 fs->e2fs_gd[cg].ext2bgd_nifree >= minifree &&
516 fs->e2fs_gd[cg].ext2bgd_nbfree >= minbfree) {
517 if (fs->e2fs_contigdirs[cg] < maxcontigdirs)
518 return (cg);
519 }
520 for (cg = 0; cg < prefcg; cg++)
521 if (fs->e2fs_gd[cg].ext2bgd_ndirs < maxndir &&
522 fs->e2fs_gd[cg].ext2bgd_nifree >= minifree &&
523 fs->e2fs_gd[cg].ext2bgd_nbfree >= minbfree) {
524 if (fs->e2fs_contigdirs[cg] < maxcontigdirs)
525 return (cg);
526 }
527 /*
528 * This is a backstop when we have deficit in space.
529 */
530 for (cg = prefcg; cg < fs->e2fs_gcount; cg++)
531 if (fs->e2fs_gd[cg].ext2bgd_nifree >= avgifree)
532 return (cg);
533 for (cg = 0; cg < prefcg; cg++)
534 if (fs->e2fs_gd[cg].ext2bgd_nifree >= avgifree)
535 break;
536 return (cg);
537 }
538
539 /*
540 * Select the desired position for the next block in a file.
541 *
542 * we try to mimic what Remy does in inode_getblk/block_getblk
543 *
544 * we note: blocknr == 0 means that we're about to allocate either
545 * a direct block or a pointer block at the first level of indirection
546 * (In other words, stuff that will go in i_db[] or i_ib[])
547 *
548 * blocknr != 0 means that we're allocating a block that is none
549 * of the above. Then, blocknr tells us the number of the block
550 * that will hold the pointer
551 */
552 e4fs_daddr_t
553 ext2_blkpref(struct inode *ip, e2fs_lbn_t lbn, int indx, e2fs_daddr_t *bap,
554 e2fs_daddr_t blocknr)
555 {
556 int tmp;
557
558 mtx_assert(EXT2_MTX(ip->i_ump), MA_OWNED);
559
560 /*
561 * If the next block is actually what we thought it is, then set the
562 * goal to what we thought it should be.
563 */
564 if (ip->i_next_alloc_block == lbn && ip->i_next_alloc_goal != 0)
565 return ip->i_next_alloc_goal;
566
567 /*
568 * Now check whether we were provided with an array that basically
569 * tells us previous blocks to which we want to stay close.
570 */
571 if (bap)
572 for (tmp = indx - 1; tmp >= 0; tmp--)
573 if (bap[tmp])
574 return bap[tmp];
575
576 /*
577 * Else lets fall back to the blocknr or, if there is none, follow
578 * the rule that a block should be allocated near its inode.
579 */
580 return blocknr ? blocknr :
581 (e2fs_daddr_t)(ip->i_block_group *
582 EXT2_BLOCKS_PER_GROUP(ip->i_e2fs)) +
583 ip->i_e2fs->e2fs->e2fs_first_dblock;
584 }
585
586 /*
587 * Implement the cylinder overflow algorithm.
588 *
589 * The policy implemented by this algorithm is:
590 * 1) allocate the block in its requested cylinder group.
591 * 2) quadradically rehash on the cylinder group number.
592 * 3) brute force search for a free block.
593 */
594 static u_long
595 ext2_hashalloc(struct inode *ip, int cg, long pref, int size,
596 daddr_t (*allocator) (struct inode *, int, daddr_t, int))
597 {
598 struct m_ext2fs *fs;
599 ino_t result;
600 int i, icg = cg;
601
602 mtx_assert(EXT2_MTX(ip->i_ump), MA_OWNED);
603 fs = ip->i_e2fs;
604 /*
605 * 1: preferred cylinder group
606 */
607 result = (*allocator)(ip, cg, pref, size);
608 if (result)
609 return (result);
610 /*
611 * 2: quadratic rehash
612 */
613 for (i = 1; i < fs->e2fs_gcount; i *= 2) {
614 cg += i;
615 if (cg >= fs->e2fs_gcount)
616 cg -= fs->e2fs_gcount;
617 result = (*allocator)(ip, cg, 0, size);
618 if (result)
619 return (result);
620 }
621 /*
622 * 3: brute force search
623 * Note that we start at i == 2, since 0 was checked initially,
624 * and 1 is always checked in the quadratic rehash.
625 */
626 cg = (icg + 2) % fs->e2fs_gcount;
627 for (i = 2; i < fs->e2fs_gcount; i++) {
628 result = (*allocator)(ip, cg, 0, size);
629 if (result)
630 return (result);
631 cg++;
632 if (cg == fs->e2fs_gcount)
633 cg = 0;
634 }
635 return (0);
636 }
637
638 /*
639 * Determine whether a block can be allocated.
640 *
641 * Check to see if a block of the appropriate size is available,
642 * and if it is, allocate it.
643 */
644 static daddr_t
645 ext2_alloccg(struct inode *ip, int cg, daddr_t bpref, int size)
646 {
647 struct m_ext2fs *fs;
648 struct buf *bp;
649 struct ext2mount *ump;
650 daddr_t bno, runstart, runlen;
651 int bit, loc, end, error, start;
652 char *bbp;
653 /* XXX ondisk32 */
654 fs = ip->i_e2fs;
655 ump = ip->i_ump;
656 if (fs->e2fs_gd[cg].ext2bgd_nbfree == 0)
657 return (0);
658 EXT2_UNLOCK(ump);
659 error = bread(ip->i_devvp, fsbtodb(fs,
660 fs->e2fs_gd[cg].ext2bgd_b_bitmap),
661 (int)fs->e2fs_bsize, NOCRED, &bp);
662 if (error) {
663 brelse(bp);
664 EXT2_LOCK(ump);
665 return (0);
666 }
667 if (fs->e2fs_gd[cg].ext2bgd_nbfree == 0) {
668 /*
669 * Another thread allocated the last block in this
670 * group while we were waiting for the buffer.
671 */
672 brelse(bp);
673 EXT2_LOCK(ump);
674 return (0);
675 }
676 bbp = (char *)bp->b_data;
677
678 if (dtog(fs, bpref) != cg)
679 bpref = 0;
680 if (bpref != 0) {
681 bpref = dtogd(fs, bpref);
682 /*
683 * if the requested block is available, use it
684 */
685 if (isclr(bbp, bpref)) {
686 bno = bpref;
687 goto gotit;
688 }
689 }
690 /*
691 * no blocks in the requested cylinder, so take next
692 * available one in this cylinder group.
693 * first try to get 8 contigous blocks, then fall back to a single
694 * block.
695 */
696 if (bpref)
697 start = dtogd(fs, bpref) / NBBY;
698 else
699 start = 0;
700 end = howmany(fs->e2fs->e2fs_fpg, NBBY) - start;
701 retry:
702 runlen = 0;
703 runstart = 0;
704 for (loc = start; loc < end; loc++) {
705 if (bbp[loc] == (char)0xff) {
706 runlen = 0;
707 continue;
708 }
709
710 /* Start of a run, find the number of high clear bits. */
711 if (runlen == 0) {
712 bit = fls(bbp[loc]);
713 runlen = NBBY - bit;
714 runstart = loc * NBBY + bit;
715 } else if (bbp[loc] == 0) {
716 /* Continue a run. */
717 runlen += NBBY;
718 } else {
719 /*
720 * Finish the current run. If it isn't long
721 * enough, start a new one.
722 */
723 bit = ffs(bbp[loc]) - 1;
724 runlen += bit;
725 if (runlen >= 8) {
726 bno = runstart;
727 goto gotit;
728 }
729
730 /* Run was too short, start a new one. */
731 bit = fls(bbp[loc]);
732 runlen = NBBY - bit;
733 runstart = loc * NBBY + bit;
734 }
735
736 /* If the current run is long enough, use it. */
737 if (runlen >= 8) {
738 bno = runstart;
739 goto gotit;
740 }
741 }
742 if (start != 0) {
743 end = start;
744 start = 0;
745 goto retry;
746 }
747 bno = ext2_mapsearch(fs, bbp, bpref);
748 if (bno < 0) {
749 brelse(bp);
750 EXT2_LOCK(ump);
751 return (0);
752 }
753 gotit:
754 #ifdef INVARIANTS
755 if (isset(bbp, bno)) {
756 printf("ext2fs_alloccgblk: cg=%d bno=%jd fs=%s\n",
757 cg, (intmax_t)bno, fs->e2fs_fsmnt);
758 panic("ext2fs_alloccg: dup alloc");
759 }
760 #endif
761 setbit(bbp, bno);
762 EXT2_LOCK(ump);
763 ext2_clusteracct(fs, bbp, cg, bno, -1);
764 fs->e2fs->e2fs_fbcount--;
765 fs->e2fs_gd[cg].ext2bgd_nbfree--;
766 fs->e2fs_fmod = 1;
767 EXT2_UNLOCK(ump);
768 bdwrite(bp);
769 return (cg * fs->e2fs->e2fs_fpg + fs->e2fs->e2fs_first_dblock + bno);
770 }
771
772 /*
773 * Determine whether a cluster can be allocated.
774 */
775 static daddr_t
776 ext2_clusteralloc(struct inode *ip, int cg, daddr_t bpref, int len)
777 {
778 struct m_ext2fs *fs;
779 struct ext2mount *ump;
780 struct buf *bp;
781 char *bbp;
782 int bit, error, got, i, loc, run;
783 int32_t *lp;
784 daddr_t bno;
785
786 fs = ip->i_e2fs;
787 ump = ip->i_ump;
788
789 if (fs->e2fs_maxcluster[cg] < len)
790 return (0);
791
792 EXT2_UNLOCK(ump);
793 error = bread(ip->i_devvp,
794 fsbtodb(fs, fs->e2fs_gd[cg].ext2bgd_b_bitmap),
795 (int)fs->e2fs_bsize, NOCRED, &bp);
796 if (error)
797 goto fail_lock;
798
799 bbp = (char *)bp->b_data;
800 EXT2_LOCK(ump);
801 /*
802 * Check to see if a cluster of the needed size (or bigger) is
803 * available in this cylinder group.
804 */
805 lp = &fs->e2fs_clustersum[cg].cs_sum[len];
806 for (i = len; i <= fs->e2fs_contigsumsize; i++)
807 if (*lp++ > 0)
808 break;
809 if (i > fs->e2fs_contigsumsize) {
810 /*
811 * Update the cluster summary information to reflect
812 * the true maximum-sized cluster so that future cluster
813 * allocation requests can avoid reading the bitmap only
814 * to find no cluster.
815 */
816 lp = &fs->e2fs_clustersum[cg].cs_sum[len - 1];
817 for (i = len - 1; i > 0; i--)
818 if (*lp-- > 0)
819 break;
820 fs->e2fs_maxcluster[cg] = i;
821 goto fail;
822 }
823 EXT2_UNLOCK(ump);
824
825 /* Search the bitmap to find a big enough cluster like in FFS. */
826 if (dtog(fs, bpref) != cg)
827 bpref = 0;
828 if (bpref != 0)
829 bpref = dtogd(fs, bpref);
830 loc = bpref / NBBY;
831 bit = 1 << (bpref % NBBY);
832 for (run = 0, got = bpref; got < fs->e2fs->e2fs_fpg; got++) {
833 if ((bbp[loc] & bit) != 0)
834 run = 0;
835 else {
836 run++;
837 if (run == len)
838 break;
839 }
840 if ((got & (NBBY - 1)) != (NBBY - 1))
841 bit <<= 1;
842 else {
843 loc++;
844 bit = 1;
845 }
846 }
847
848 if (got >= fs->e2fs->e2fs_fpg)
849 goto fail_lock;
850
851 /* Allocate the cluster that we found. */
852 for (i = 1; i < len; i++)
853 if (!isclr(bbp, got - run + i))
854 panic("ext2_clusteralloc: map mismatch");
855
856 bno = got - run + 1;
857 if (bno >= fs->e2fs->e2fs_fpg)
858 panic("ext2_clusteralloc: allocated out of group");
859
860 EXT2_LOCK(ump);
861 for (i = 0; i < len; i += fs->e2fs_fpb) {
862 setbit(bbp, bno + i);
863 ext2_clusteracct(fs, bbp, cg, bno + i, -1);
864 fs->e2fs->e2fs_fbcount--;
865 fs->e2fs_gd[cg].ext2bgd_nbfree--;
866 }
867 fs->e2fs_fmod = 1;
868 EXT2_UNLOCK(ump);
869
870 bdwrite(bp);
871 return (cg * fs->e2fs->e2fs_fpg + fs->e2fs->e2fs_first_dblock + bno);
872
873 fail_lock:
874 EXT2_LOCK(ump);
875 fail:
876 brelse(bp);
877 return (0);
878 }
879
880 /*
881 * Determine whether an inode can be allocated.
882 *
883 * Check to see if an inode is available, and if it is,
884 * allocate it using tode in the specified cylinder group.
885 */
886 static daddr_t
887 ext2_nodealloccg(struct inode *ip, int cg, daddr_t ipref, int mode)
888 {
889 struct m_ext2fs *fs;
890 struct buf *bp;
891 struct ext2mount *ump;
892 int error, start, len;
893 char *ibp, *loc;
894
895 ipref--; /* to avoid a lot of (ipref -1) */
896 if (ipref == -1)
897 ipref = 0;
898 fs = ip->i_e2fs;
899 ump = ip->i_ump;
900 if (fs->e2fs_gd[cg].ext2bgd_nifree == 0)
901 return (0);
902 EXT2_UNLOCK(ump);
903 error = bread(ip->i_devvp, fsbtodb(fs,
904 fs->e2fs_gd[cg].ext2bgd_i_bitmap),
905 (int)fs->e2fs_bsize, NOCRED, &bp);
906 if (error) {
907 brelse(bp);
908 EXT2_LOCK(ump);
909 return (0);
910 }
911 if (fs->e2fs_gd[cg].ext2bgd_nifree == 0) {
912 /*
913 * Another thread allocated the last i-node in this
914 * group while we were waiting for the buffer.
915 */
916 brelse(bp);
917 EXT2_LOCK(ump);
918 return (0);
919 }
920 ibp = (char *)bp->b_data;
921 if (ipref) {
922 ipref %= fs->e2fs->e2fs_ipg;
923 if (isclr(ibp, ipref))
924 goto gotit;
925 }
926 start = ipref / NBBY;
927 len = howmany(fs->e2fs->e2fs_ipg - ipref, NBBY);
928 loc = memcchr(&ibp[start], 0xff, len);
929 if (loc == NULL) {
930 len = start + 1;
931 start = 0;
932 loc = memcchr(&ibp[start], 0xff, len);
933 if (loc == NULL) {
934 printf("cg = %d, ipref = %lld, fs = %s\n",
935 cg, (long long)ipref, fs->e2fs_fsmnt);
936 panic("ext2fs_nodealloccg: map corrupted");
937 /* NOTREACHED */
938 }
939 }
940 ipref = (loc - ibp) * NBBY + ffs(~*loc) - 1;
941 gotit:
942 setbit(ibp, ipref);
943 EXT2_LOCK(ump);
944 fs->e2fs_gd[cg].ext2bgd_nifree--;
945 fs->e2fs->e2fs_ficount--;
946 fs->e2fs_fmod = 1;
947 if ((mode & IFMT) == IFDIR) {
948 fs->e2fs_gd[cg].ext2bgd_ndirs++;
949 fs->e2fs_total_dir++;
950 }
951 EXT2_UNLOCK(ump);
952 bdwrite(bp);
953 return (cg * fs->e2fs->e2fs_ipg + ipref + 1);
954 }
955
956 /*
957 * Free a block or fragment.
958 *
959 */
960 void
961 ext2_blkfree(struct inode *ip, e4fs_daddr_t bno, long size)
962 {
963 struct m_ext2fs *fs;
964 struct buf *bp;
965 struct ext2mount *ump;
966 int cg, error;
967 char *bbp;
968
969 fs = ip->i_e2fs;
970 ump = ip->i_ump;
971 cg = dtog(fs, bno);
972 if ((u_int)bno >= fs->e2fs->e2fs_bcount) {
973 printf("bad block %lld, ino %llu\n", (long long)bno,
974 (unsigned long long)ip->i_number);
975 ext2_fserr(fs, ip->i_uid, "bad block");
976 return;
977 }
978 error = bread(ip->i_devvp,
979 fsbtodb(fs, fs->e2fs_gd[cg].ext2bgd_b_bitmap),
980 (int)fs->e2fs_bsize, NOCRED, &bp);
981 if (error) {
982 brelse(bp);
983 return;
984 }
985 bbp = (char *)bp->b_data;
986 bno = dtogd(fs, bno);
987 if (isclr(bbp, bno)) {
988 printf("block = %lld, fs = %s\n",
989 (long long)bno, fs->e2fs_fsmnt);
990 panic("ext2_blkfree: freeing free block");
991 }
992 clrbit(bbp, bno);
993 EXT2_LOCK(ump);
994 ext2_clusteracct(fs, bbp, cg, bno, 1);
995 fs->e2fs->e2fs_fbcount++;
996 fs->e2fs_gd[cg].ext2bgd_nbfree++;
997 fs->e2fs_fmod = 1;
998 EXT2_UNLOCK(ump);
999 bdwrite(bp);
1000 }
1001
1002 /*
1003 * Free an inode.
1004 *
1005 */
1006 int
1007 ext2_vfree(struct vnode *pvp, ino_t ino, int mode)
1008 {
1009 struct m_ext2fs *fs;
1010 struct inode *pip;
1011 struct buf *bp;
1012 struct ext2mount *ump;
1013 int error, cg;
1014 char *ibp;
1015
1016 pip = VTOI(pvp);
1017 fs = pip->i_e2fs;
1018 ump = pip->i_ump;
1019 if ((u_int)ino > fs->e2fs_ipg * fs->e2fs_gcount)
1020 panic("ext2_vfree: range: devvp = %p, ino = %ju, fs = %s",
1021 pip->i_devvp, (uintmax_t)ino, fs->e2fs_fsmnt);
1022
1023 cg = ino_to_cg(fs, ino);
1024 error = bread(pip->i_devvp,
1025 fsbtodb(fs, fs->e2fs_gd[cg].ext2bgd_i_bitmap),
1026 (int)fs->e2fs_bsize, NOCRED, &bp);
1027 if (error) {
1028 brelse(bp);
1029 return (0);
1030 }
1031 ibp = (char *)bp->b_data;
1032 ino = (ino - 1) % fs->e2fs->e2fs_ipg;
1033 if (isclr(ibp, ino)) {
1034 printf("ino = %llu, fs = %s\n",
1035 (unsigned long long)ino, fs->e2fs_fsmnt);
1036 if (fs->e2fs_ronly == 0)
1037 panic("ext2_vfree: freeing free inode");
1038 }
1039 clrbit(ibp, ino);
1040 EXT2_LOCK(ump);
1041 fs->e2fs->e2fs_ficount++;
1042 fs->e2fs_gd[cg].ext2bgd_nifree++;
1043 if ((mode & IFMT) == IFDIR) {
1044 fs->e2fs_gd[cg].ext2bgd_ndirs--;
1045 fs->e2fs_total_dir--;
1046 }
1047 fs->e2fs_fmod = 1;
1048 EXT2_UNLOCK(ump);
1049 bdwrite(bp);
1050 return (0);
1051 }
1052
1053 /*
1054 * Find a block in the specified cylinder group.
1055 *
1056 * It is a panic if a request is made to find a block if none are
1057 * available.
1058 */
1059 static daddr_t
1060 ext2_mapsearch(struct m_ext2fs *fs, char *bbp, daddr_t bpref)
1061 {
1062 char *loc;
1063 int start, len;
1064
1065 /*
1066 * find the fragment by searching through the free block
1067 * map for an appropriate bit pattern
1068 */
1069 if (bpref)
1070 start = dtogd(fs, bpref) / NBBY;
1071 else
1072 start = 0;
1073 len = howmany(fs->e2fs->e2fs_fpg, NBBY) - start;
1074 loc = memcchr(&bbp[start], 0xff, len);
1075 if (loc == NULL) {
1076 len = start + 1;
1077 start = 0;
1078 loc = memcchr(&bbp[start], 0xff, len);
1079 if (loc == NULL) {
1080 printf("start = %d, len = %d, fs = %s\n",
1081 start, len, fs->e2fs_fsmnt);
1082 panic("ext2_mapsearch: map corrupted");
1083 /* NOTREACHED */
1084 }
1085 }
1086 return ((loc - bbp) * NBBY + ffs(~*loc) - 1);
1087 }
1088
1089 /*
1090 * Fserr prints the name of a filesystem with an error diagnostic.
1091 *
1092 * The form of the error message is:
1093 * fs: error message
1094 */
1095 static void
1096 ext2_fserr(struct m_ext2fs *fs, uid_t uid, char *cp)
1097 {
1098
1099 log(LOG_ERR, "uid %u on %s: %s\n", uid, fs->e2fs_fsmnt, cp);
1100 }
1101
1102 int
1103 cg_has_sb(int i)
1104 {
1105 int a3, a5, a7;
1106
1107 if (i == 0 || i == 1)
1108 return 1;
1109 for (a3 = 3, a5 = 5, a7 = 7;
1110 a3 <= i || a5 <= i || a7 <= i;
1111 a3 *= 3, a5 *= 5, a7 *= 7)
1112 if (i == a3 || i == a5 || i == a7)
1113 return 1;
1114 return 0;
1115 }
Cache object: 7c8b07975169943c96d0e3677882a69c
|