Deleted Added
sdiff udiff text old ( 243245 ) new ( 243250 )
full compact
1/*-
2 * Copyright (c) 1982, 1986, 1993
3 * The Regents of the University of California. All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
13 * 4. Neither the name of the University nor the names of its contributors
14 * may be used to endorse or promote products derived from this software
15 * without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE.
28 *
29 * @(#)fs.h 8.13 (Berkeley) 3/21/95
30 * $FreeBSD: head/sys/ufs/ffs/fs.h 243245 2012-11-18 18:57:19Z trasz $
31 */
32
33#ifndef _UFS_FFS_FS_H_
34#define _UFS_FFS_FS_H_
35
36#include <sys/mount.h>
37
38/*
39 * Each disk drive contains some number of filesystems.
40 * A filesystem consists of a number of cylinder groups.
41 * Each cylinder group has inodes and data.
42 *
43 * A filesystem is described by its super-block, which in turn
44 * describes the cylinder groups. The super-block is critical
45 * data and is replicated in each cylinder group to protect against
46 * catastrophic loss. This is done at `newfs' time and the critical
47 * super-block data does not change, so the copies need not be
48 * referenced further unless disaster strikes.
49 *
50 * For filesystem fs, the offsets of the various blocks of interest
51 * are given in the super block as:
52 * [fs->fs_sblkno] Super-block
53 * [fs->fs_cblkno] Cylinder group block
54 * [fs->fs_iblkno] Inode blocks
55 * [fs->fs_dblkno] Data blocks
56 * The beginning of cylinder group cg in fs, is given by
57 * the ``cgbase(fs, cg)'' macro.
58 *
59 * Depending on the architecture and the media, the superblock may
60 * reside in any one of four places. For tiny media where every block
61 * counts, it is placed at the very front of the partition. Historically,
62 * UFS1 placed it 8K from the front to leave room for the disk label and
63 * a small bootstrap. For UFS2 it got moved to 64K from the front to leave
64 * room for the disk label and a bigger bootstrap, and for really piggy
65 * systems we check at 256K from the front if the first three fail. In
66 * all cases the size of the superblock will be SBLOCKSIZE. All values are
67 * given in byte-offset form, so they do not imply a sector size. The
68 * SBLOCKSEARCH specifies the order in which the locations should be searched.
69 */
70#define SBLOCK_FLOPPY 0
71#define SBLOCK_UFS1 8192
72#define SBLOCK_UFS2 65536
73#define SBLOCK_PIGGY 262144
74#define SBLOCKSIZE 8192
75#define SBLOCKSEARCH \
76 { SBLOCK_UFS2, SBLOCK_UFS1, SBLOCK_FLOPPY, SBLOCK_PIGGY, -1 }
77
78/*
79 * Max number of fragments per block. This value is NOT tweakable.
80 */
81#define MAXFRAG 8
82
83/*
84 * Addresses stored in inodes are capable of addressing fragments
85 * of `blocks'. File system blocks of at most size MAXBSIZE can
86 * be optionally broken into 2, 4, or 8 pieces, each of which is
87 * addressable; these pieces may be DEV_BSIZE, or some multiple of
88 * a DEV_BSIZE unit.
89 *
90 * Large files consist of exclusively large data blocks. To avoid
91 * undue wasted disk space, the last data block of a small file may be
92 * allocated as only as many fragments of a large block as are
93 * necessary. The filesystem format retains only a single pointer
94 * to such a fragment, which is a piece of a single large block that
95 * has been divided. The size of such a fragment is determinable from
96 * information in the inode, using the ``blksize(fs, ip, lbn)'' macro.
97 *
98 * The filesystem records space availability at the fragment level;
99 * to determine block availability, aligned fragments are examined.
100 */
101
102/*
103 * MINBSIZE is the smallest allowable block size.
104 * In order to insure that it is possible to create files of size
105 * 2^32 with only two levels of indirection, MINBSIZE is set to 4096.
106 * MINBSIZE must be big enough to hold a cylinder group block,
107 * thus changes to (struct cg) must keep its size within MINBSIZE.
108 * Note that super blocks are always of size SBLOCKSIZE,
109 * and that both SBLOCKSIZE and MAXBSIZE must be >= MINBSIZE.
110 */
111#define MINBSIZE 4096
112
113/*
114 * The path name on which the filesystem is mounted is maintained
115 * in fs_fsmnt. MAXMNTLEN defines the amount of space allocated in
116 * the super block for this name.
117 */
118#define MAXMNTLEN 468
119
120/*
121 * The volume name for this filesystem is maintained in fs_volname.
122 * MAXVOLLEN defines the length of the buffer allocated.
123 */
124#define MAXVOLLEN 32
125
126/*
127 * There is a 128-byte region in the superblock reserved for in-core
128 * pointers to summary information. Originally this included an array
129 * of pointers to blocks of struct csum; now there are just a few
130 * pointers and the remaining space is padded with fs_ocsp[].
131 *
132 * NOCSPTRS determines the size of this padding. One pointer (fs_csp)
133 * is taken away to point to a contiguous array of struct csum for
134 * all cylinder groups; a second (fs_maxcluster) points to an array
135 * of cluster sizes that is computed as cylinder groups are inspected,
136 * and the third points to an array that tracks the creation of new
137 * directories. A fourth pointer, fs_active, is used when creating
138 * snapshots; it points to a bitmap of cylinder groups for which the
139 * free-block bitmap has changed since the snapshot operation began.
140 */
141#define NOCSPTRS ((128 / sizeof(void *)) - 4)
142
143/*
144 * A summary of contiguous blocks of various sizes is maintained
145 * in each cylinder group. Normally this is set by the initial
146 * value of fs_maxcontig. To conserve space, a maximum summary size
147 * is set by FS_MAXCONTIG.
148 */
149#define FS_MAXCONTIG 16
150
151/*
152 * MINFREE gives the minimum acceptable percentage of filesystem
153 * blocks which may be free. If the freelist drops below this level
154 * only the superuser may continue to allocate blocks. This may
155 * be set to 0 if no reserve of free blocks is deemed necessary,
156 * however throughput drops by fifty percent if the filesystem
157 * is run at between 95% and 100% full; thus the minimum default
158 * value of fs_minfree is 5%. However, to get good clustering
159 * performance, 10% is a better choice. hence we use 10% as our
160 * default value. With 10% free space, fragmentation is not a
161 * problem, so we choose to optimize for time.
162 */
163#define MINFREE 8
164#define DEFAULTOPT FS_OPTTIME
165
166/*
167 * Grigoriy Orlov <gluk@ptci.ru> has done some extensive work to fine
168 * tune the layout preferences for directories within a filesystem.
169 * His algorithm can be tuned by adjusting the following parameters
170 * which tell the system the average file size and the average number
171 * of files per directory. These defaults are well selected for typical
172 * filesystems, but may need to be tuned for odd cases like filesystems
173 * being used for squid caches or news spools.
174 */
175#define AVFILESIZ 16384 /* expected average file size */
176#define AFPDIR 64 /* expected number of files per directory */
177
178/*
179 * The maximum number of snapshot nodes that can be associated
180 * with each filesystem. This limit affects only the number of
181 * snapshot files that can be recorded within the superblock so
182 * that they can be found when the filesystem is mounted. However,
183 * maintaining too many will slow the filesystem performance, so
184 * having this limit is a good idea.
185 */
186#define FSMAXSNAP 20
187
188/*
189 * Used to identify special blocks in snapshots:
190 *
191 * BLK_NOCOPY - A block that was unallocated at the time the snapshot
192 * was taken, hence does not need to be copied when written.
193 * BLK_SNAP - A block held by another snapshot that is not needed by this
194 * snapshot. When the other snapshot is freed, the BLK_SNAP entries
195 * are converted to BLK_NOCOPY. These are needed to allow fsck to
196 * identify blocks that are in use by other snapshots (which are
197 * expunged from this snapshot).
198 */
199#define BLK_NOCOPY ((ufs2_daddr_t)(1))
200#define BLK_SNAP ((ufs2_daddr_t)(2))
201
202/*
203 * Sysctl values for the fast filesystem.
204 */
205#define FFS_ADJ_REFCNT 1 /* adjust inode reference count */
206#define FFS_ADJ_BLKCNT 2 /* adjust inode used block count */
207#define FFS_BLK_FREE 3 /* free range of blocks in map */
208#define FFS_DIR_FREE 4 /* free specified dir inodes in map */
209#define FFS_FILE_FREE 5 /* free specified file inodes in map */
210#define FFS_SET_FLAGS 6 /* set filesystem flags */
211#define FFS_ADJ_NDIR 7 /* adjust number of directories */
212#define FFS_ADJ_NBFREE 8 /* adjust number of free blocks */
213#define FFS_ADJ_NIFREE 9 /* adjust number of free inodes */
214#define FFS_ADJ_NFFREE 10 /* adjust number of free frags */
215#define FFS_ADJ_NUMCLUSTERS 11 /* adjust number of free clusters */
216#define FFS_SET_CWD 12 /* set current directory */
217#define FFS_SET_DOTDOT 13 /* set inode number for ".." */
218#define FFS_UNLINK 14 /* remove a name in the filesystem */
219#define FFS_SET_INODE 15 /* update an on-disk inode */
220#define FFS_SET_BUFOUTPUT 16 /* set buffered writing on descriptor */
221#define FFS_MAXID 16 /* number of valid ffs ids */
222
223/*
224 * Command structure passed in to the filesystem to adjust filesystem values.
225 */
226#define FFS_CMD_VERSION 0x19790518 /* version ID */
227struct fsck_cmd {
228 int32_t version; /* version of command structure */
229 int32_t handle; /* reference to filesystem to be changed */
230 int64_t value; /* inode or block number to be affected */
231 int64_t size; /* amount or range to be adjusted */
232 int64_t spare; /* reserved for future use */
233};
234
235/*
236 * Per cylinder group information; summarized in blocks allocated
237 * from first cylinder group data blocks. These blocks have to be
238 * read in from fs_csaddr (size fs_cssize) in addition to the
239 * super block.
240 */
241struct csum {
242 int32_t cs_ndir; /* number of directories */
243 int32_t cs_nbfree; /* number of free blocks */
244 int32_t cs_nifree; /* number of free inodes */
245 int32_t cs_nffree; /* number of free frags */
246};
247struct csum_total {
248 int64_t cs_ndir; /* number of directories */
249 int64_t cs_nbfree; /* number of free blocks */
250 int64_t cs_nifree; /* number of free inodes */
251 int64_t cs_nffree; /* number of free frags */
252 int64_t cs_numclusters; /* number of free clusters */
253 int64_t cs_spare[3]; /* future expansion */
254};
255
256/*
257 * Super block for an FFS filesystem.
258 */
259struct fs {
260 int32_t fs_firstfield; /* historic filesystem linked list, */
261 int32_t fs_unused_1; /* used for incore super blocks */
262 int32_t fs_sblkno; /* offset of super-block in filesys */
263 int32_t fs_cblkno; /* offset of cyl-block in filesys */
264 int32_t fs_iblkno; /* offset of inode-blocks in filesys */
265 int32_t fs_dblkno; /* offset of first data after cg */
266 int32_t fs_old_cgoffset; /* cylinder group offset in cylinder */
267 int32_t fs_old_cgmask; /* used to calc mod fs_ntrak */
268 int32_t fs_old_time; /* last time written */
269 int32_t fs_old_size; /* number of blocks in fs */
270 int32_t fs_old_dsize; /* number of data blocks in fs */
271 u_int32_t fs_ncg; /* number of cylinder groups */
272 int32_t fs_bsize; /* size of basic blocks in fs */
273 int32_t fs_fsize; /* size of frag blocks in fs */
274 int32_t fs_frag; /* number of frags in a block in fs */
275/* these are configuration parameters */
276 int32_t fs_minfree; /* minimum percentage of free blocks */
277 int32_t fs_old_rotdelay; /* num of ms for optimal next block */
278 int32_t fs_old_rps; /* disk revolutions per second */
279/* these fields can be computed from the others */
280 int32_t fs_bmask; /* ``blkoff'' calc of blk offsets */
281 int32_t fs_fmask; /* ``fragoff'' calc of frag offsets */
282 int32_t fs_bshift; /* ``lblkno'' calc of logical blkno */
283 int32_t fs_fshift; /* ``numfrags'' calc number of frags */
284/* these are configuration parameters */
285 int32_t fs_maxcontig; /* max number of contiguous blks */
286 int32_t fs_maxbpg; /* max number of blks per cyl group */
287/* these fields can be computed from the others */
288 int32_t fs_fragshift; /* block to frag shift */
289 int32_t fs_fsbtodb; /* fsbtodb and dbtofsb shift constant */
290 int32_t fs_sbsize; /* actual size of super block */
291 int32_t fs_spare1[2]; /* old fs_csmask */
292 /* old fs_csshift */
293 int32_t fs_nindir; /* value of NINDIR */
294 u_int32_t fs_inopb; /* value of INOPB */
295 int32_t fs_old_nspf; /* value of NSPF */
296/* yet another configuration parameter */
297 int32_t fs_optim; /* optimization preference, see below */
298 int32_t fs_old_npsect; /* # sectors/track including spares */
299 int32_t fs_old_interleave; /* hardware sector interleave */
300 int32_t fs_old_trackskew; /* sector 0 skew, per track */
301 int32_t fs_id[2]; /* unique filesystem id */
302/* sizes determined by number of cylinder groups and their sizes */
303 int32_t fs_old_csaddr; /* blk addr of cyl grp summary area */
304 int32_t fs_cssize; /* size of cyl grp summary area */
305 int32_t fs_cgsize; /* cylinder group size */
306 int32_t fs_spare2; /* old fs_ntrak */
307 int32_t fs_old_nsect; /* sectors per track */
308 int32_t fs_old_spc; /* sectors per cylinder */
309 int32_t fs_old_ncyl; /* cylinders in filesystem */
310 int32_t fs_old_cpg; /* cylinders per group */
311 u_int32_t fs_ipg; /* inodes per group */
312 int32_t fs_fpg; /* blocks per group * fs_frag */
313/* this data must be re-computed after crashes */
314 struct csum fs_old_cstotal; /* cylinder summary information */
315/* these fields are cleared at mount time */
316 int8_t fs_fmod; /* super block modified flag */
317 int8_t fs_clean; /* filesystem is clean flag */
318 int8_t fs_ronly; /* mounted read-only flag */
319 int8_t fs_old_flags; /* old FS_ flags */
320 u_char fs_fsmnt[MAXMNTLEN]; /* name mounted on */
321 u_char fs_volname[MAXVOLLEN]; /* volume name */
322 u_int64_t fs_swuid; /* system-wide uid */
323 int32_t fs_pad; /* due to alignment of fs_swuid */
324/* these fields retain the current block allocation info */
325 int32_t fs_cgrotor; /* last cg searched */
326 void *fs_ocsp[NOCSPTRS]; /* padding; was list of fs_cs buffers */
327 u_int8_t *fs_contigdirs; /* (u) # of contig. allocated dirs */
328 struct csum *fs_csp; /* (u) cg summary info buffer */
329 int32_t *fs_maxcluster; /* (u) max cluster in each cyl group */
330 u_int *fs_active; /* (u) used by snapshots to track fs */
331 int32_t fs_old_cpc; /* cyl per cycle in postbl */
332 int32_t fs_maxbsize; /* maximum blocking factor permitted */
333 int64_t fs_unrefs; /* number of unreferenced inodes */
334 int64_t fs_providersize; /* size of underlying GEOM provider */
335 int64_t fs_sparecon64[15]; /* old rotation block list head */
336 int64_t fs_sblockloc; /* byte offset of standard superblock */
337 struct csum_total fs_cstotal; /* (u) cylinder summary information */
338 ufs_time_t fs_time; /* last time written */
339 int64_t fs_size; /* number of blocks in fs */
340 int64_t fs_dsize; /* number of data blocks in fs */
341 ufs2_daddr_t fs_csaddr; /* blk addr of cyl grp summary area */
342 int64_t fs_pendingblocks; /* (u) blocks being freed */
343 u_int32_t fs_pendinginodes; /* (u) inodes being freed */
344 uint32_t fs_snapinum[FSMAXSNAP];/* list of snapshot inode numbers */
345 u_int32_t fs_avgfilesize; /* expected average file size */
346 u_int32_t fs_avgfpdir; /* expected # of files per directory */
347 int32_t fs_save_cgsize; /* save real cg size to use fs_bsize */
348 ufs_time_t fs_mtime; /* Last mount or fsck time. */
349 int32_t fs_sujfree; /* SUJ free list */
350 int32_t fs_sparecon32[23]; /* reserved for future constants */
351 int32_t fs_flags; /* see FS_ flags below */
352 int32_t fs_contigsumsize; /* size of cluster summary array */
353 int32_t fs_maxsymlinklen; /* max length of an internal symlink */
354 int32_t fs_old_inodefmt; /* format of on-disk inodes */
355 u_int64_t fs_maxfilesize; /* maximum representable file size */
356 int64_t fs_qbmask; /* ~fs_bmask for use with 64-bit size */
357 int64_t fs_qfmask; /* ~fs_fmask for use with 64-bit size */
358 int32_t fs_state; /* validate fs_clean field */
359 int32_t fs_old_postblformat; /* format of positional layout tables */
360 int32_t fs_old_nrpos; /* number of rotational positions */
361 int32_t fs_spare5[2]; /* old fs_postbloff */
362 /* old fs_rotbloff */
363 int32_t fs_magic; /* magic number */
364};
365
366/* Sanity checking. */
367#ifdef CTASSERT
368CTASSERT(sizeof(struct fs) == 1376);
369#endif
370
371/*
372 * Filesystem identification
373 */
374#define FS_UFS1_MAGIC 0x011954 /* UFS1 fast filesystem magic number */
375#define FS_UFS2_MAGIC 0x19540119 /* UFS2 fast filesystem magic number */
376#define FS_BAD_MAGIC 0x19960408 /* UFS incomplete newfs magic number */
377#define FS_OKAY 0x7c269d38 /* superblock checksum */
378#define FS_42INODEFMT -1 /* 4.2BSD inode format */
379#define FS_44INODEFMT 2 /* 4.4BSD inode format */
380
381/*
382 * Preference for optimization.
383 */
384#define FS_OPTTIME 0 /* minimize allocation time */
385#define FS_OPTSPACE 1 /* minimize disk fragmentation */
386
387/*
388 * Filesystem flags.
389 *
390 * The FS_UNCLEAN flag is set by the kernel when the filesystem was
391 * mounted with fs_clean set to zero. The FS_DOSOFTDEP flag indicates
392 * that the filesystem should be managed by the soft updates code.
393 * Note that the FS_NEEDSFSCK flag is set and cleared only by the
394 * fsck utility. It is set when background fsck finds an unexpected
395 * inconsistency which requires a traditional foreground fsck to be
396 * run. Such inconsistencies should only be found after an uncorrectable
397 * disk error. A foreground fsck will clear the FS_NEEDSFSCK flag when
398 * it has successfully cleaned up the filesystem. The kernel uses this
399 * flag to enforce that inconsistent filesystems be mounted read-only.
400 * The FS_INDEXDIRS flag when set indicates that the kernel maintains
401 * on-disk auxiliary indexes (such as B-trees) for speeding directory
402 * accesses. Kernels that do not support auxiliary indicies clear the
403 * flag to indicate that the indicies need to be rebuilt (by fsck) before
404 * they can be used.
405 *
406 * FS_ACLS indicates that POSIX.1e ACLs are administratively enabled
407 * for the file system, so they should be loaded from extended attributes,
408 * observed for access control purposes, and be administered by object
409 * owners. FS_NFS4ACLS indicates that NFSv4 ACLs are administratively
410 * enabled. This flag is mutually exclusive with FS_ACLS. FS_MULTILABEL
411 * indicates that the TrustedBSD MAC Framework should attempt to back MAC
412 * labels into extended attributes on the file system rather than maintain
413 * a single mount label for all objects.
414 */
415#define FS_UNCLEAN 0x0001 /* filesystem not clean at mount */
416#define FS_DOSOFTDEP 0x0002 /* filesystem using soft dependencies */
417#define FS_NEEDSFSCK 0x0004 /* filesystem needs sync fsck before mount */
418#define FS_SUJ 0x0008 /* Filesystem using softupdate journal */
419#define FS_ACLS 0x0010 /* file system has POSIX.1e ACLs enabled */
420#define FS_MULTILABEL 0x0020 /* file system is MAC multi-label */
421#define FS_GJOURNAL 0x0040 /* gjournaled file system */
422#define FS_FLAGS_UPDATED 0x0080 /* flags have been moved to new location */
423#define FS_NFS4ACLS 0x0100 /* file system has NFSv4 ACLs enabled */
424#define FS_INDEXDIRS 0x0200 /* kernel supports indexed directories */
425#define FS_TRIM 0x0400 /* issue BIO_DELETE for deleted blocks */
426
427/*
428 * Macros to access bits in the fs_active array.
429 */
430#define ACTIVECGNUM(fs, cg) ((fs)->fs_active[(cg) / (NBBY * sizeof(int))])
431#define ACTIVECGOFF(cg) (1 << ((cg) % (NBBY * sizeof(int))))
432#define ACTIVESET(fs, cg) do { \
433 if ((fs)->fs_active) \
434 ACTIVECGNUM((fs), (cg)) |= ACTIVECGOFF((cg)); \
435} while (0)
436#define ACTIVECLEAR(fs, cg) do { \
437 if ((fs)->fs_active) \
438 ACTIVECGNUM((fs), (cg)) &= ~ACTIVECGOFF((cg)); \
439} while (0)
440
441/*
442 * The size of a cylinder group is calculated by CGSIZE. The maximum size
443 * is limited by the fact that cylinder groups are at most one block.
444 * Its size is derived from the size of the maps maintained in the
445 * cylinder group and the (struct cg) size.
446 */
447#define CGSIZE(fs) \
448 /* base cg */ (sizeof(struct cg) + sizeof(int32_t) + \
449 /* old btotoff */ (fs)->fs_old_cpg * sizeof(int32_t) + \
450 /* old boff */ (fs)->fs_old_cpg * sizeof(u_int16_t) + \
451 /* inode map */ howmany((fs)->fs_ipg, NBBY) + \
452 /* block map */ howmany((fs)->fs_fpg, NBBY) +\
453 /* if present */ ((fs)->fs_contigsumsize <= 0 ? 0 : \
454 /* cluster sum */ (fs)->fs_contigsumsize * sizeof(int32_t) + \
455 /* cluster map */ howmany(fragstoblks(fs, (fs)->fs_fpg), NBBY)))
456
457/*
458 * The minimal number of cylinder groups that should be created.
459 */
460#define MINCYLGRPS 4
461
462/*
463 * Convert cylinder group to base address of its global summary info.
464 */
465#define fs_cs(fs, indx) fs_csp[indx]
466
467/*
468 * Cylinder group block for a filesystem.
469 */
470#define CG_MAGIC 0x090255
471struct cg {
472 int32_t cg_firstfield; /* historic cyl groups linked list */
473 int32_t cg_magic; /* magic number */
474 int32_t cg_old_time; /* time last written */
475 u_int32_t cg_cgx; /* we are the cgx'th cylinder group */
476 int16_t cg_old_ncyl; /* number of cyl's this cg */
477 int16_t cg_old_niblk; /* number of inode blocks this cg */
478 u_int32_t cg_ndblk; /* number of data blocks this cg */
479 struct csum cg_cs; /* cylinder summary information */
480 u_int32_t cg_rotor; /* position of last used block */
481 u_int32_t cg_frotor; /* position of last used frag */
482 u_int32_t cg_irotor; /* position of last used inode */
483 u_int32_t cg_frsum[MAXFRAG]; /* counts of available frags */
484 int32_t cg_old_btotoff; /* (int32) block totals per cylinder */
485 int32_t cg_old_boff; /* (u_int16) free block positions */
486 u_int32_t cg_iusedoff; /* (u_int8) used inode map */
487 u_int32_t cg_freeoff; /* (u_int8) free block map */
488 u_int32_t cg_nextfreeoff; /* (u_int8) next available space */
489 u_int32_t cg_clustersumoff; /* (u_int32) counts of avail clusters */
490 u_int32_t cg_clusteroff; /* (u_int8) free cluster map */
491 u_int32_t cg_nclusterblks; /* number of clusters this cg */
492 u_int32_t cg_niblk; /* number of inode blocks this cg */
493 u_int32_t cg_initediblk; /* last initialized inode */
494 u_int32_t cg_unrefs; /* number of unreferenced inodes */
495 int32_t cg_sparecon32[2]; /* reserved for future use */
496 ufs_time_t cg_time; /* time last written */
497 int64_t cg_sparecon64[3]; /* reserved for future use */
498 u_int8_t cg_space[1]; /* space for cylinder group maps */
499/* actually longer */
500};
501
502/*
503 * Macros for access to cylinder group array structures
504 */
505#define cg_chkmagic(cgp) ((cgp)->cg_magic == CG_MAGIC)
506#define cg_inosused(cgp) \
507 ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_iusedoff))
508#define cg_blksfree(cgp) \
509 ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_freeoff))
510#define cg_clustersfree(cgp) \
511 ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_clusteroff))
512#define cg_clustersum(cgp) \
513 ((int32_t *)((uintptr_t)(cgp) + (cgp)->cg_clustersumoff))
514
515/*
516 * Turn filesystem block numbers into disk block addresses.
517 * This maps filesystem blocks to device size blocks.
518 */
519#define fsbtodb(fs, b) ((daddr_t)(b) << (fs)->fs_fsbtodb)
520#define dbtofsb(fs, b) ((b) >> (fs)->fs_fsbtodb)
521
522/*
523 * Cylinder group macros to locate things in cylinder groups.
524 * They calc filesystem addresses of cylinder group data structures.
525 */
526#define cgbase(fs, c) (((ufs2_daddr_t)(fs)->fs_fpg) * (c))
527#define cgdmin(fs, c) (cgstart(fs, c) + (fs)->fs_dblkno) /* 1st data */
528#define cgimin(fs, c) (cgstart(fs, c) + (fs)->fs_iblkno) /* inode blk */
529#define cgsblock(fs, c) (cgstart(fs, c) + (fs)->fs_sblkno) /* super blk */
530#define cgtod(fs, c) (cgstart(fs, c) + (fs)->fs_cblkno) /* cg block */
531#define cgstart(fs, c) \
532 ((fs)->fs_magic == FS_UFS2_MAGIC ? cgbase(fs, c) : \
533 (cgbase(fs, c) + (fs)->fs_old_cgoffset * ((c) & ~((fs)->fs_old_cgmask))))
534
535/*
536 * Macros for handling inode numbers:
537 * inode number to filesystem block offset.
538 * inode number to cylinder group number.
539 * inode number to filesystem block address.
540 */
541#define ino_to_cg(fs, x) (((ino_t)(x)) / (fs)->fs_ipg)
542#define ino_to_fsba(fs, x) \
543 ((ufs2_daddr_t)(cgimin(fs, ino_to_cg(fs, (ino_t)(x))) + \
544 (blkstofrags((fs), ((((ino_t)(x)) % (fs)->fs_ipg) / INOPB(fs))))))
545#define ino_to_fsbo(fs, x) (((ino_t)(x)) % INOPB(fs))
546
547/*
548 * Give cylinder group number for a filesystem block.
549 * Give cylinder group block number for a filesystem block.
550 */
551#define dtog(fs, d) ((d) / (fs)->fs_fpg)
552#define dtogd(fs, d) ((d) % (fs)->fs_fpg)
553
554/*
555 * Extract the bits for a block from a map.
556 * Compute the cylinder and rotational position of a cyl block addr.
557 */
558#define blkmap(fs, map, loc) \
559 (((map)[(loc) / NBBY] >> ((loc) % NBBY)) & (0xff >> (NBBY - (fs)->fs_frag)))
560
561/*
562 * The following macros optimize certain frequently calculated
563 * quantities by using shifts and masks in place of divisions
564 * modulos and multiplications.
565 */
566#define blkoff(fs, loc) /* calculates (loc % fs->fs_bsize) */ \
567 ((loc) & (fs)->fs_qbmask)
568#define fragoff(fs, loc) /* calculates (loc % fs->fs_fsize) */ \
569 ((loc) & (fs)->fs_qfmask)
570#define lfragtosize(fs, frag) /* calculates ((off_t)frag * fs->fs_fsize) */ \
571 (((off_t)(frag)) << (fs)->fs_fshift)
572#define lblktosize(fs, blk) /* calculates ((off_t)blk * fs->fs_bsize) */ \
573 (((off_t)(blk)) << (fs)->fs_bshift)
574/* Use this only when `blk' is known to be small, e.g., < NDADDR. */
575#define smalllblktosize(fs, blk) /* calculates (blk * fs->fs_bsize) */ \
576 ((blk) << (fs)->fs_bshift)
577#define lblkno(fs, loc) /* calculates (loc / fs->fs_bsize) */ \
578 ((loc) >> (fs)->fs_bshift)
579#define numfrags(fs, loc) /* calculates (loc / fs->fs_fsize) */ \
580 ((loc) >> (fs)->fs_fshift)
581#define blkroundup(fs, size) /* calculates roundup(size, fs->fs_bsize) */ \
582 (((size) + (fs)->fs_qbmask) & (fs)->fs_bmask)
583#define fragroundup(fs, size) /* calculates roundup(size, fs->fs_fsize) */ \
584 (((size) + (fs)->fs_qfmask) & (fs)->fs_fmask)
585#define fragstoblks(fs, frags) /* calculates (frags / fs->fs_frag) */ \
586 ((frags) >> (fs)->fs_fragshift)
587#define blkstofrags(fs, blks) /* calculates (blks * fs->fs_frag) */ \
588 ((blks) << (fs)->fs_fragshift)
589#define fragnum(fs, fsb) /* calculates (fsb % fs->fs_frag) */ \
590 ((fsb) & ((fs)->fs_frag - 1))
591#define blknum(fs, fsb) /* calculates rounddown(fsb, fs->fs_frag) */ \
592 ((fsb) &~ ((fs)->fs_frag - 1))
593
594/*
595 * Determine the number of available frags given a
596 * percentage to hold in reserve.
597 */
598#define freespace(fs, percentreserved) \
599 (blkstofrags((fs), (fs)->fs_cstotal.cs_nbfree) + \
600 (fs)->fs_cstotal.cs_nffree - \
601 (((off_t)((fs)->fs_dsize)) * (percentreserved) / 100))
602
603/*
604 * Determining the size of a file block in the filesystem.
605 */
606#define blksize(fs, ip, lbn) \
607 (((lbn) >= NDADDR || (ip)->i_size >= smalllblktosize(fs, (lbn) + 1)) \
608 ? (fs)->fs_bsize \
609 : (fragroundup(fs, blkoff(fs, (ip)->i_size))))
610#define sblksize(fs, size, lbn) \
611 (((lbn) >= NDADDR || (size) >= ((lbn) + 1) << (fs)->fs_bshift) \
612 ? (fs)->fs_bsize \
613 : (fragroundup(fs, blkoff(fs, (size)))))
614
615/*
616 * Number of indirects in a filesystem block.
617 */
618#define NINDIR(fs) ((fs)->fs_nindir)
619
620/*
621 * Indirect lbns are aligned on NDADDR addresses where single indirects
622 * are the negated address of the lowest lbn reachable, double indirects
623 * are this lbn - 1 and triple indirects are this lbn - 2. This yields
624 * an unusual bit order to determine level.
625 */
626static inline int
627lbn_level(ufs_lbn_t lbn)
628{
629 if (lbn >= 0)
630 return 0;
631 switch (lbn & 0x3) {
632 case 0:
633 return (0);
634 case 1:
635 break;
636 case 2:
637 return (2);
638 case 3:
639 return (1);
640 default:
641 break;
642 }
643 return (-1);
644}
645
646static inline ufs_lbn_t
647lbn_offset(struct fs *fs, int level)
648{
649 ufs_lbn_t res;
650
651 for (res = 1; level > 0; level--)
652 res *= NINDIR(fs);
653 return (res);
654}
655
656/*
657 * Number of inodes in a secondary storage block/fragment.
658 */
659#define INOPB(fs) ((fs)->fs_inopb)
660#define INOPF(fs) ((fs)->fs_inopb >> (fs)->fs_fragshift)
661
662/*
663 * Softdep journal record format.
664 */
665
666#define JOP_ADDREF 1 /* Add a reference to an inode. */
667#define JOP_REMREF 2 /* Remove a reference from an inode. */
668#define JOP_NEWBLK 3 /* Allocate a block. */
669#define JOP_FREEBLK 4 /* Free a block or a tree of blocks. */
670#define JOP_MVREF 5 /* Move a reference from one off to another. */
671#define JOP_TRUNC 6 /* Partial truncation record. */
672#define JOP_SYNC 7 /* fsync() complete record. */
673
674#define JREC_SIZE 32 /* Record and segment header size. */
675
676#define SUJ_MIN (4 * 1024 * 1024) /* Minimum journal size */
677#define SUJ_MAX (32 * 1024 * 1024) /* Maximum journal size */
678#define SUJ_FILE ".sujournal" /* Journal file name */
679
680/*
681 * Size of the segment record header. There is at most one for each disk
682 * block in the journal. The segment header is followed by an array of
683 * records. fsck depends on the first element in each record being 'op'
684 * and the second being 'ino'. Segments may span multiple disk blocks but
685 * the header is present on each.
686 */
687struct jsegrec {
688 uint64_t jsr_seq; /* Our sequence number */
689 uint64_t jsr_oldest; /* Oldest valid sequence number */
690 uint16_t jsr_cnt; /* Count of valid records */
691 uint16_t jsr_blocks; /* Count of device bsize blocks. */
692 uint32_t jsr_crc; /* 32bit crc of the valid space */
693 ufs_time_t jsr_time; /* timestamp for mount instance */
694};
695
696/*
697 * Reference record. Records a single link count modification.
698 */
699struct jrefrec {
700 uint32_t jr_op;
701 uint32_t jr_ino;
702 uint32_t jr_parent;
703 uint16_t jr_nlink;
704 uint16_t jr_mode;
705 int64_t jr_diroff;
706 uint64_t jr_unused;
707};
708
709/*
710 * Move record. Records a reference moving within a directory block. The
711 * nlink is unchanged but we must search both locations.
712 */
713struct jmvrec {
714 uint32_t jm_op;
715 uint32_t jm_ino;
716 uint32_t jm_parent;
717 uint16_t jm_unused;
718 int64_t jm_oldoff;
719 int64_t jm_newoff;
720};
721
722/*
723 * Block record. A set of frags or tree of blocks starting at an indirect are
724 * freed or a set of frags are allocated.
725 */
726struct jblkrec {
727 uint32_t jb_op;
728 uint32_t jb_ino;
729 ufs2_daddr_t jb_blkno;
730 ufs_lbn_t jb_lbn;
731 uint16_t jb_frags;
732 uint16_t jb_oldfrags;
733 uint32_t jb_unused;
734};
735
736/*
737 * Truncation record. Records a partial truncation so that it may be
738 * completed at check time. Also used for sync records.
739 */
740struct jtrncrec {
741 uint32_t jt_op;
742 uint32_t jt_ino;
743 int64_t jt_size;
744 uint32_t jt_extsize;
745 uint32_t jt_pad[3];
746};
747
748union jrec {
749 struct jsegrec rec_jsegrec;
750 struct jrefrec rec_jrefrec;
751 struct jmvrec rec_jmvrec;
752 struct jblkrec rec_jblkrec;
753 struct jtrncrec rec_jtrncrec;
754};
755
756#ifdef CTASSERT
757CTASSERT(sizeof(struct jsegrec) == JREC_SIZE);
758CTASSERT(sizeof(struct jrefrec) == JREC_SIZE);
759CTASSERT(sizeof(struct jmvrec) == JREC_SIZE);
760CTASSERT(sizeof(struct jblkrec) == JREC_SIZE);
761CTASSERT(sizeof(struct jtrncrec) == JREC_SIZE);
762CTASSERT(sizeof(union jrec) == JREC_SIZE);
763#endif
764
765extern int inside[], around[];
766extern u_char *fragtbl[];
767
768/*
769 * IOCTLs used for filesystem write suspension.
770 */
771#define UFSSUSPEND _IOW('U', 1, fsid_t)
772#define UFSRESUME _IO('U', 2)
773
774#endif