xref: /haiku/src/add-ons/kernel/file_systems/ufs2/ufs2.h (revision 899e0ef82b5624ace2ccfa5f5a58c8ebee54aaef)
1 /*-
2  * SPDX-License-Identifier: BSD-3-Clause
3  *
4  * Copyright (c) 1982, 1986, 1993
5  *	The Regents of the University of California.  All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  * 3. Neither the name of the University nor the names of its contributors
16  *    may be used to endorse or promote products derived from this software
17  *    without specific prior written permission.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  *
31  *	@(#)fs.h	8.13 (Berkeley) 3/21/95
32  * $FreeBSD$
33  */
34 
35 #ifndef UFS2_H
36 #define UFS2_H
37 
38 #include "system_dependencies.h"
39 
40 typedef uint32 cluster_t;
41 typedef uint64_t ufs2_daddr_t;
42 
43 #define	SBLOCK_FLOPPY		0
44 #define	SBLOCK_UFS1			8192
45 #define	SBLOCK_UFS2			65536
46 #define	SBLOCK_PIGGY		262144
47 #define	SBLOCKSIZE			8192
48 #define	SBLOCKSEARCH \
49 	{ SBLOCK_UFS2, SBLOCK_UFS1, SBLOCK_FLOPPY, SBLOCK_PIGGY, -1 }
50 
51 /*
52  * Max number of fragments per block. This value is NOT tweakable.
53  */
54 #define	MAXFRAG				8
55 
56 /*
57  * Addresses stored in inodes are capable of addressing fragments
58  * of `blocks'. File system blocks of at most size MAXBSIZE can
59  * be optionally broken into 2, 4, or 8 pieces, each of which is
60  * addressable; these pieces may be DEV_BSIZE, or some multiple of
61  * a DEV_BSIZE unit.
62  *
63  * Large files consist of exclusively large data blocks.  To avoid
64  * undue wasted disk space, the last data block of a small file may be
65  * allocated as only as many fragments of a large block as are
66  * necessary.  The filesystem format retains only a single pointer
67  * to such a fragment, which is a piece of a single large block that
68  * has been divided.  The size of such a fragment is determinable from
69  * information in the inode, using the ``blksize(fs, ip, lbn)'' macro.
70  *
71  * The filesystem records space availability at the fragment level;
72  * to determine block availability, aligned fragments are examined.
73  */
74 
75 /*
76  * MINBSIZE is the smallest allowable block size.
77  * In order to insure that it is possible to create files of size
78  * 2^32 with only two levels of indirection, MINBSIZE is set to 4096.
79  * MINBSIZE must be big enough to hold a cylinder group block,
80  * thus changes to (struct cg) must keep its size within MINBSIZE.
81  * Note that super blocks are always of size SBLOCKSIZE,
82  * and that both SBLOCKSIZE and MAXBSIZE must be >= MINBSIZE.
83  */
84 #define	MINBSIZE			4096
85 
86 /*
87  * The path name on which the filesystem is mounted is maintained
88  * in fs_fsmnt. MAXMNTLEN defines the amount of space allocated in
89  * the super block for this name.
90  */
91 #define	MAXMNTLEN			468
92 
93 /*
94  * The volume name for this filesystem is maintained in fs_volname.
95  * MAXVOLLEN defines the length of the buffer allocated.
96  */
97 #define	MAXVOLLEN			32
98 
99 /*
100  * There is a 128-byte region in the superblock reserved for in-core
101  * pointers to summary information. Originally this included an array
102  * of pointers to blocks of struct csum; now there are just a few
103  * pointers and the remaining space is padded with fs_ocsp[].
104  *
105 * NOCSPTRS determines the size of this padding. One pointer (fs_csp)
106  * is taken away to point to a contiguous array of struct csum for
107  * all cylinder groups; a second (fs_maxcluster) points to an array
108  * of cluster sizes that is computed as cylinder groups are inspected,
109  * and the third points to an array that tracks the creation of new
110  * directories. A fourth pointer, fs_active, is used when creating
111  * snapshots; it points to a bitmap of cylinder groups for which the
112  * free-block bitmap has changed since the snapshot operation began.
113  */
114 #define	NOCSPTRS			((128 / sizeof(void *)) - 4)
115 
116 /*
117  * A summary of contiguous blocks of various sizes is maintained
118  * in each cylinder group. Normally this is set by the initial
119  * value of fs_maxcontig. To conserve space, a maximum summary size
120  * is set by FS_MAXCONTIG.
121  */
122 #define	FS_MAXCONTIG		16
123 
124 /*
125  * MINFREE gives the minimum acceptable percentage of filesystem
126  * blocks which may be free. If the freelist drops below this level
127  * only the superuser may continue to allocate blocks. This may
128  * be set to 0 if no reserve of free blocks is deemed necessary,
129  * however throughput drops by fifty percent if the filesystem
130  * is run at between 95% and 100% full; thus the minimum default
131  * value of fs_minfree is 5%. However, to get good clustering
132  * performance, 10% is a better choice. hence we use 10% as our
133  * default value. With 10% free space, fragmentation is not a
134  * problem, so we choose to optimize for time.
135  */
136 #define	MINFREE				8
137 #define	DEFAULTOPT			FS_OPTTIME
138 
139 /*
140  * Grigoriy Orlov <gluk@ptci.ru> has done some extensive work to fine
141  * tune the layout preferences for directories within a filesystem.
142  * His algorithm can be tuned by adjusting the following parameters
143  * which tell the system the average file size and the average number
144  * of files per directory. These defaults are well selected for typical
145  * filesystems, but may need to be tuned for odd cases like filesystems
146  * being used for squid caches or news spools.
147  */
148 #define	AVFILESIZ			16384	/* expected average file size */
149 #define	AFPDIR				64	/* expected number of files per directory */
150 
151 /*
152  * The maximum number of snapshot nodes that can be associated
153  * with each filesystem. This limit affects only the number of
154  * snapshot files that can be recorded within the superblock so
155  * that they can be found when the filesystem is mounted. However,
156  * maintaining too many will slow the filesystem performance, so
157  * having this limit is a good idea.
158  */
159 #define	FSMAXSNAP			20
160 
161 /*
162  * Used to identify special blocks in snapshots:
163  *
164  * BLK_NOCOPY - A block that was unallocated at the time the snapshot
165  *	was taken, hence does not need to be copied when written.
166  * BLK_SNAP - A block held by another snapshot that is not needed by this
167  *	snapshot. When the other snapshot is freed, the BLK_SNAP entries
168  *	are converted to BLK_NOCOPY. These are needed to allow fsck to
169  *	identify blocks that are in use by other snapshots (which are
170  *	expunged from this snapshot).
171  */
172 #define	BLK_NOCOPY			((ufs2_daddr_t)(1))
173 #define	BLK_SNAP			((ufs2_daddr_t)(2))
174 
175 /*
176  * Sysctl values for the fast filesystem.
177  */
178 #define	FFS_ADJ_REFCNT		 1	/* adjust inode reference count */
179 #define	FFS_ADJ_BLKCNT		 2	/* adjust inode used block count */
180 #define	FFS_BLK_FREE		 3	/* free range of blocks in map */
181 #define	FFS_DIR_FREE		 4	/* free specified dir inodes in map */
182 #define	FFS_FILE_FREE		 5	/* free specified file inodes in map */
183 #define	FFS_SET_FLAGS		 6	/* set filesystem flags */
184 #define	FFS_ADJ_NDIR		 7	/* adjust number of directories */
185 #define	FFS_ADJ_NBFREE		 8	/* adjust number of free blocks */
186 #define	FFS_ADJ_NIFREE		 9	/* adjust number of free inodes */
187 #define	FFS_ADJ_NFFREE		10 	/* adjust number of free frags */
188 #define	FFS_ADJ_NUMCLUSTERS	11	/* adjust number of free clusters */
189 #define	FFS_SET_CWD			12	/* set current directory */
190 #define	FFS_SET_DOTDOT		13	/* set inode number for ".." */
191 #define	FFS_UNLINK			14	/* remove a name in the filesystem */
192 /* Was FFS_SET_INODE		15 */
193 /* Was FFS_SET_BUFOUTPUT	16 */
194 #define	FFS_SET_SIZE		17	/* set inode size */
195 #define	FFS_MAXID			17	/* number of valid ffs ids */
196 
197 /*
198  * Command structure passed in to the filesystem to adjust filesystem values.
199  */
200 #define	FFS_CMD_VERSION		0x19790518	/* version ID */
201 struct fsck_cmd {
202 	int32_t	version;	/* version of command structure */
203 	int32_t	handle;		/* reference to filesystem to be changed */
204 	int64_t	value;		/* inode or block number to be affected */
205 	int64_t	size;		/* amount or range to be adjusted */
206 	int64_t	spare;		/* reserved for future use */
207 };
208 
209 /*
210  * A recovery structure placed at the end of the boot block area by newfs
211  * that can be used by fsck to search for alternate superblocks.
212  */
213 struct fsrecovery {
214 	int32_t	fsr_magic;	/* magic number */
215 	int32_t	fsr_fsbtodb;	/* fsbtodb and dbtofsb shift constant */
216 	int32_t	fsr_sblkno;	/* offset of super-block in filesys */
217 	int32_t	fsr_fpg;	/* blocks per group * fs_frag */
218 	u_int32_t fsr_ncg;	/* number of cylinder groups */
219 };
220 
221 /*
222  * Per cylinder group information; summarized in blocks allocated
223  * from first cylinder group data blocks.  These blocks have to be
224  * read in from fs_csaddr (size fs_cssize) in addition to the
225  * super block.
226  */
227 struct csum {
228 	int32_t	cs_ndir;		/* number of directories */
229 	int32_t	cs_nbfree;		/* number of free blocks */
230 	int32_t	cs_nifree;		/* number of free inodes */
231 	int32_t	cs_nffree;		/* number of free frags */
232 };
233 struct csum_total {
234 	int64_t	cs_ndir;		/* number of directories */
235 	int64_t	cs_nbfree;		/* number of free blocks */
236 	int64_t	cs_nifree;		/* number of free inodes */
237 	int64_t	cs_nffree;		/* number of free frags */
238 	int64_t	cs_numclusters;		/* number of free clusters */
239 	int64_t	cs_spare[3];		/* future expansion */
240 };
241 
242 /*
243  * Super block for an FFS filesystem.
244  */
245 struct ufs2_super_block {
246 	int32_t		fs_firstfield;		/* historic filesystem linked list, */
247 	int32_t		fs_unused_1;		/*     used for incore super blocks */
248 	int32_t		fs_sblkno;		/* offset of super-block in filesys */
249 	int32_t		fs_cblkno;		/* offset of cyl-block in filesys */
250 	int32_t		fs_iblkno;		/* offset of inode-blocks in filesys */
251 	int32_t		fs_dblkno;		/* offset of first data after cg */
252 	int32_t		fs_old_cgoffset;	/* cylinder group offset in cylinder */
253 	int32_t		fs_old_cgmask;		/* used to calc mod fs_ntrak */
254 	int32_t		fs_old_time;		/* last time written */
255 	int32_t		fs_old_size;		/* number of blocks in fs */
256 	int32_t		fs_old_dsize;		/* number of data blocks in fs */
257 	u_int32_t	fs_ncg;		/* number of cylinder groups */
258 	int32_t		fs_bsize;		/* size of basic blocks in fs */
259 	int32_t		fs_fsize;		/* size of frag blocks in fs */
260 	int32_t		fs_frag;		/* number of frags in a block in fs */
261 /* these are configuration parameters */
262 	int32_t		fs_minfree;		/* minimum percentage of free blocks */
263 	int32_t		fs_old_rotdelay;	/* num of ms for optimal next block */
264 	int32_t		fs_old_rps;		/* disk revolutions per second */
265 /* these fields can be computed from the others */
266 	int32_t		fs_bmask;		/* ``blkoff'' calc of blk offsets */
267 	int32_t		fs_fmask;		/* ``fragoff'' calc of frag offsets */
268 	int32_t		fs_bshift;		/* ``lblkno'' calc of logical blkno */
269 	int32_t		fs_fshift;		/* ``numfrags'' calc number of frags */
270 /* these are configuration parameters */
271 	int32_t		fs_maxcontig;		/* max number of contiguous blks */
272 	int32_t		fs_maxbpg;		/* max number of blks per cyl group */
273 /* these fields can be computed from the others */
274 	int32_t		fs_fragshift;		/* block to frag shift */
275 	int32_t		fs_fsbtodb;		/* fsbtodb and dbtofsb shift constant */
276 	int32_t		fs_sbsize;		/* actual size of super block */
277 	int32_t		fs_spare1[2];		/* old fs_csmask */
278 					/* old fs_csshift */
279 	int32_t		fs_nindir;		/* value of NINDIR */
280 	u_int32_t	fs_inopb;		/* value of INOPB */
281 	int32_t		fs_old_nspf;		/* value of NSPF */
282 /* yet another configuration parameter */
283 	int32_t		fs_optim;		/* optimization preference, see below */
284 	int32_t		fs_old_npsect;		/* # sectors/track including spares */
285 	int32_t		fs_old_interleave;	/* hardware sector interleave */
286 	int32_t		fs_old_trackskew;	/* sector 0 skew, per track */
287 	int32_t		fs_id[2];		/* unique filesystem id */
288 /* sizes determined by number of cylinder groups and their sizes */
289 	int32_t		fs_old_csaddr;		/* blk addr of cyl grp summary area */
290 	int32_t		fs_cssize;		/* size of cyl grp summary area */
291 	int32_t		fs_cgsize;		/* cylinder group size */
292 	int32_t		fs_spare2;		/* old fs_ntrak */
293 	int32_t		fs_old_nsect;		/* sectors per track */
294 	int32_t		fs_old_spc;		/* sectors per cylinder */
295 	int32_t		fs_old_ncyl;		/* cylinders in filesystem */
296 	int32_t		fs_old_cpg;		/* cylinders per group */
297 	u_int32_t	fs_ipg;		/* inodes per group */
298 	int32_t		fs_fpg;		/* blocks per group * fs_frag */
299 /* this data must be re-computed after crashes */
300 	struct		csum fs_old_cstotal;	/* cylinder summary information */
301 /* these fields are cleared at mount time */
302 	int8_t		fs_fmod;		/* super block modified flag */
303 	int8_t		fs_clean;		/* filesystem is clean flag */
304 	int8_t		fs_ronly;		/* mounted read-only flag */
305 	int8_t		fs_old_flags;		/* old FS_ flags */
306 	u_char		fs_fsmnt[MAXMNTLEN];	/* name mounted on */
307 	char		fs_volname[MAXVOLLEN];	/* volume name */
308 	u_int64_t	fs_swuid;		/* system-wide uid */
309 	int32_t		fs_pad;		/* due to alignment of fs_swuid */
310 /* these fields retain the current block allocation info */
311 	int32_t		fs_cgrotor;		/* last cg searched */
312 	void		*fs_ocsp[NOCSPTRS];	/* padding; was list of fs_cs buffers */
313 	u_int8_t	*fs_contigdirs;	/* (u) # of contig. allocated dirs */
314 	struct		csum *fs_csp;		/* (u) cg summary info buffer */
315 	int32_t		*fs_maxcluster;		/* (u) max cluster in each cyl group */
316 	u_int		*fs_active;		/* (u) used by snapshots to track fs */
317 	int32_t		fs_old_cpc;		/* cyl per cycle in postbl */
318 	int32_t		fs_maxbsize;		/* maximum blocking factor permitted */
319 	int64_t		fs_unrefs;		/* number of unreferenced inodes */
320 	int64_t		fs_providersize;	/* size of underlying GEOM provider */
321 	int64_t		fs_metaspace;		/* size of area reserved for metadata */
322 	int64_t		fs_sparecon64[13];	/* old rotation block list head */
323 	int64_t		fs_sblockactualloc;	/* byte offset of this superblock */
324 	int64_t		fs_sblockloc;		/* byte offset of standard superblock */
325 	struct		csum_total fs_cstotal;	/* (u) cylinder summary information */
326 	int64_t		fs_time;		/* last time written */
327 	int64_t		fs_size;		/* number of blocks in fs */
328 	int64_t		fs_dsize;		/* number of data blocks in fs */
329 	int64_t		fs_csaddr;		/* blk addr of cyl grp summary area */
330 	int64_t		fs_pendingblocks;	/* (u) blocks being freed */
331 	u_int32_t	fs_pendinginodes;	/* (u) inodes being freed */
332 	uint32_t	fs_snapinum[FSMAXSNAP];/* list of snapshot inode numbers */
333 	u_int32_t	fs_avgfilesize;	/* expected average file size */
334 	u_int32_t	fs_avgfpdir;		/* expected # of files per directory */
335 	int32_t		fs_save_cgsize;	/* save real cg size to use fs_bsize */
336 	int64_t		fs_mtime;		/* Last mount or fsck time. */
337 	int32_t		fs_sujfree;		/* SUJ free list */
338 	int32_t		fs_sparecon32[21];	/* reserved for future constants */
339 	u_int32_t	fs_ckhash;		/* if CK_SUPERBLOCK, its check-hash */
340 	u_int32_t	fs_metackhash;	/* metadata check-hash, see CK_ below */
341 	int32_t		fs_flags;		/* see FS_ flags below */
342 	int32_t		fs_contigsumsize;	/* size of cluster summary array */
343 	int32_t		fs_maxsymlinklen;	/* max length of an internal symlink */
344 	int32_t		fs_old_inodefmt;	/* format of on-disk inodes */
345 	u_int64_t	fs_maxfilesize;	/* maximum representable file size */
346 	int64_t		fs_qbmask;		/* ~fs_bmask for use with 64-bit size */
347 	int64_t		fs_qfmask;		/* ~fs_fmask for use with 64-bit size */
348 	int32_t		fs_state;		/* validate fs_clean field */
349 	int32_t		fs_old_postblformat;	/* format of positional layout tables */
350 	int32_t		fs_old_nrpos;		/* number of rotational positions */
351 	int32_t		fs_spare5[2];		/* old fs_postbloff */
352 					/* old fs_rotbloff */
353 	int32_t		fs_magic;		/* magic number */
354 	bool		IsValid();
355 };
356 
357 /* Sanity checking. */
358 #ifdef CTASSERT
359 CTASSERT(sizeof(struct fs) == 1376);
360 #endif
361 
362 
363 struct file_cookie {
364 	bigtime_t	last_notification;
365 	off_t		last_size;
366 	int 		open_mode;
367 };
368 
369 
370 /*
371  * Filesystem identification
372  */
373 #define	FS_UFS1_MAGIC		0x011954	/* UFS1 fast filesystem magic number */
374 #define	FS_UFS2_MAGIC		0x19540119	/* UFS2 fast filesystem magic number */
375 #define	FS_BAD_MAGIC		0x19960408	/* UFS incomplete newfs magic number */
376 #define	FS_42INODEFMT		-1		/* 4.2BSD inode format */
377 #define	FS_44INODEFMT		2		/* 4.4BSD inode format */
378 
379 /*
380  * Preference for optimization.
381  */
382 #define	FS_OPTTIME			0	/* minimize allocation time */
383 #define	FS_OPTSPACE			1	/* minimize disk fragmentation */
384 
385 /*
386  * Filesystem flags.
387  *
388  * The FS_UNCLEAN flag is set by the kernel when the filesystem was
389  * mounted with fs_clean set to zero. The FS_DOSOFTDEP flag indicates
390  * that the filesystem should be managed by the soft updates code.
391  * Note that the FS_NEEDSFSCK flag is set and cleared by the fsck
392  * utility. It is set when background fsck finds an unexpected
393  * inconsistency which requires a traditional foreground fsck to be
394  * run. Such inconsistencies should only be found after an uncorrectable
395  * disk error. The FS_NEEDSFSCK can also be set when a mounted filesystem
396  * discovers an internal inconsistency such as freeing a freed inode.
397  * A foreground fsck will clear the FS_NEEDSFSCK flag when it has
398  * successfully cleaned up the filesystem. The kernel uses this
399  * flag to enforce that inconsistent filesystems be mounted read-only.
400  *
401  * The FS_METACKHASH flag when set indicates that the kernel maintains
402  * one or more check hashes. The actual set of supported check hashes
403  * is stored in the fs_metackhash field. Kernels that do not support
404  * check hashes clear the FS_METACKHASH flag to indicate that the
405  * check hashes need to be rebuilt (by fsck) before they can be used.
406  *
407  * When a filesystem is mounted, any flags not included in FS_SUPPORTED
408  * are cleared. This lets newer features know that the filesystem has
409  * been run on an older version of the filesystem and thus that data
410  * structures associated with those features are out-of-date and need
411  * to be rebuilt.
412  *
413  * FS_ACLS indicates that POSIX.1e ACLs are administratively enabled
414  * for the file system, so they should be loaded from extended attributes,
415  * observed for access control purposes, and be administered by object
416  * owners.  FS_NFS4ACLS indicates that NFSv4 ACLs are administratively
417  * enabled.  This flag is mutually exclusive with FS_ACLS.  FS_MULTILABEL
418  * indicates that the TrustedBSD MAC Framework should attempt to back MAC
419  * labels into extended attributes on the file system rather than maintain
420  * a single mount label for all objects.
421  */
422 #define	FS_UNCLEAN			0x00000001 /* filesystem not clean at mount */
423 #define	FS_DOSOFTDEP		0x00000002 /* filesystem using soft dependencies */
424 /* filesystem needs sync fsck before mount */
425 #define	FS_NEEDSFSCK		0x00000004
426 #define	FS_SUJ				0x00000008 /* Filesystem using softupdate journal */
427 /* file system has POSIX.1e ACLs enabled */
428 #define	FS_ACLS				0x00000010
429 #define	FS_MULTILABEL		0x00000020 /* file system is MAC multi-label */
430 #define	FS_GJOURNAL			0x00000040 /* gjournaled file system */
431 /* flags have been moved to new location */
432 #define	FS_FLAGS_UPDATED	0x0000080
433 #define	FS_NFS4ACLS			0x00000100 /* file system has NFSv4 ACLs enabled */
434 /* kernel supports metadata check hashes */
435 #define	FS_METACKHASH		0x00000200
436 #define	FS_TRIM				0x00000400 /* issue BIO_DELETE for deleted blocks */
437 /* supported flags, others cleared at mount */
438 #define	FS_SUPPORTED		0x00FFFFFF
439 /*
440  * Things that we may someday support, but currently do not.
441  * These flags are all cleared so we know if we ran on a kernel
442  * that does not support them.
443  */
444 #define	FS_INDEXDIRS		0x01000000 /* kernel supports indexed directories */
445 /* kernel supports variable block sizes */
446 #define	FS_VARBLKSIZE		0x02000000
447 #define	FS_COOLOPT1			0x04000000 /* kernel supports cool option 1 */
448 #define	FS_COOLOPT2			0x08000000 /* kernel supports cool option 2 */
449 #define	FS_COOLOPT3			0x10000000 /* kernel supports cool option 3 */
450 #define	FS_COOLOPT4			0x20000000 /* kernel supports cool option 4 */
451 #define	FS_COOLOPT5			0x40000000 /* kernel supports cool option 5 */
452 #define	FS_COOLOPT6			0x80000000 /* kernel supports cool option 6 */
453 
454 /*
455  * The fs_metackhash field indicates the types of metadata check-hash
456  * that are maintained for a filesystem. Not all filesystems check-hash
457  * all metadata.
458  */
459 #define	CK_SUPERBLOCK		0x0001	/* the superblock */
460 #define	CK_CYLGRP			0x0002	/* the cylinder groups */
461 #define	CK_INODE			0x0004	/* inodes */
462 #define	CK_INDIR			0x0008	/* indirect blocks */
463 #define	CK_DIR				0x0010	/* directory contents */
464 /*
465  * The BX_FSPRIV buffer b_xflags are used to track types of data in buffers.
466  */
467 #define	BX_SUPERBLOCK		0x00010000	/* superblock */
468 #define	BX_CYLGRP			0x00020000	/* cylinder groups */
469 #define	BX_INODE			0x00040000	/* inodes */
470 #define	BX_INDIR			0x00080000	/* indirect blocks */
471 #define	BX_DIR				0x00100000	/* directory contents */
472 
473 /*
474 #define	PRINT_UFS_BUF_XFLAGS "\20\25dir\24indir\23inode\22cylgrp\21superblock"
475 */
476 
477 /*
478  * Macros to access bits in the fs_active array.
479  */
480 #define	ACTIVECGNUM(fs, cg)	((fs)->fs_active[(cg) / (NBBY * sizeof(int))])
481 #define	ACTIVECGOFF(cg)		(1 << ((cg) % (NBBY * sizeof(int))))
482 #define	ACTIVESET(fs, cg)	do {					\
483 	if ((fs)->fs_active)						\
484 		ACTIVECGNUM((fs), (cg)) |= ACTIVECGOFF((cg));		\
485 } while (0)
486 #define	ACTIVECLEAR(fs, cg)	do {					\
487 	if ((fs)->fs_active)						\
488 		ACTIVECGNUM((fs), (cg)) &= ~ACTIVECGOFF((cg));		\
489 } while (0)
490 
491 /*
492  * The size of a cylinder group is calculated by CGSIZE. The maximum size
493  * is limited by the fact that cylinder groups are at most one block.
494  * Its size is derived from the size of the maps maintained in the
495  * cylinder group and the (struct cg) size.
496  */
497 #define	CGSIZE(fs) \
498     /* base cg */	(sizeof(struct cg) + sizeof(int32_t) + \
499     /* old btotoff */	(fs)->fs_old_cpg * sizeof(int32_t) + \
500     /* old boff */	(fs)->fs_old_cpg * sizeof(u_int16_t) + \
501     /* inode map */	howmany((fs)->fs_ipg, NBBY) + \
502     /* block map */	howmany((fs)->fs_fpg, NBBY) +\
503     /* if present */	((fs)->fs_contigsumsize <= 0 ? 0 : \
504     /* cluster sum */	(fs)->fs_contigsumsize * sizeof(int32_t) + \
505     /* cluster map */	howmany(fragstoblks(fs, (fs)->fs_fpg), NBBY)))
506 
507 /*
508  * The minimal number of cylinder groups that should be created.
509  */
510 #define	MINCYLGRPS	4
511 
512 /*
513  * Convert cylinder group to base address of its global summary info.
514  */
515 #define	fs_cs(fs, indx) fs_csp[indx]
516 
517 /*
518  * Cylinder group block for a filesystem.
519  */
520 #define	CG_MAGIC	0x090255
521 struct cg {
522 	int32_t		cg_firstfield;		/* historic cyl groups linked list */
523 	int32_t		cg_magic;		/* magic number */
524 	int32_t		cg_old_time;		/* time last written */
525 	u_int32_t	cg_cgx;		/* we are the cgx'th cylinder group */
526 	int16_t	 	cg_old_ncyl;		/* number of cyl's this cg */
527 	int16_t		cg_old_niblk;		/* number of inode blocks this cg */
528 	u_int32_t	cg_ndblk;		/* number of data blocks this cg */
529 	struct		csum cg_cs;		/* cylinder summary information */
530 	u_int32_t	cg_rotor;		/* position of last used block */
531 	u_int32_t	cg_frotor;		/* position of last used frag */
532 	u_int32_t	cg_irotor;		/* position of last used inode */
533 	u_int32_t	cg_frsum[MAXFRAG];	/* counts of available frags */
534 	int32_t		cg_old_btotoff;	/* (int32) block totals per cylinder */
535 	int32_t		cg_old_boff;		/* (u_int16) free block positions */
536 	u_int32_t	cg_iusedoff;		/* (u_int8) used inode map */
537 	u_int32_t	cg_freeoff;		/* (u_int8) free block map */
538 	u_int32_t	cg_nextfreeoff;	/* (u_int8) next available space */
539 	u_int32_t	cg_clustersumoff;	/* (u_int32) counts of avail clusters */
540 	u_int32_t	cg_clusteroff;		/* (u_int8) free cluster map */
541 	u_int32_t	cg_nclusterblks;	/* number of clusters this cg */
542 	u_int32_t	cg_niblk;		/* number of inode blocks this cg */
543 	u_int32_t	cg_initediblk;		/* last initialized inode */
544 	u_int32_t	cg_unrefs;		/* number of unreferenced inodes */
545 	int32_t		cg_sparecon32[1];	/* reserved for future use */
546 	u_int32_t	cg_ckhash;		/* check-hash of this cg */
547 	int64_t		cg_time;		/* time last written */
548 	int64_t		cg_sparecon64[3];	/* reserved for future use */
549 	u_int8_t	cg_space[1];		/* space for cylinder group maps */
550 /* actually longer */
551 };
552 
553 /*
554  * Macros for access to cylinder group array structures
555  */
556 #define	cg_chkmagic(cgp) ((cgp)->cg_magic == CG_MAGIC)
557 #define	cg_inosused(cgp) \
558     ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_iusedoff))
559 #define	cg_blksfree(cgp) \
560     ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_freeoff))
561 #define	cg_clustersfree(cgp) \
562     ((u_int8_t *)((u_int8_t *)(cgp) + (cgp)->cg_clusteroff))
563 #define	cg_clustersum(cgp) \
564     ((int32_t *)((uintptr_t)(cgp) + (cgp)->cg_clustersumoff))
565 
566 /*
567  * Turn filesystem block numbers into disk block addresses.
568  * This maps filesystem blocks to device size blocks.
569  */
570 #define	fsbtodb(fs, b)	((daddr_t)(b) << (fs)->fs_fsbtodb)
571 #define	dbtofsb(fs, b)	((b) >> (fs)->fs_fsbtodb)
572 
573 /*
574  * Cylinder group macros to locate things in cylinder groups.
575  * They calc filesystem addresses of cylinder group data structures.
576  */
577 #define	cgbase(fs, c)	(((ufs2_daddr_t)(fs)->fs_fpg) * (c))
578 #define	cgdata(fs, c)	(cgdmin(fs, c) + (fs)->fs_metaspace)	/* data zone */
579 #define	cgmeta(fs, c)	(cgdmin(fs, c))				/* meta data */
580 #define	cgdmin(fs, c)	(cgstart(fs, c) + (fs)->fs_dblkno)	/* 1st data */
581 #define	cgimin(fs, c)	(cgstart(fs, c) + (fs)->fs_iblkno)	/* inode blk */
582 #define	cgsblock(fs, c)	(cgstart(fs, c) + (fs)->fs_sblkno)	/* super blk */
583 #define	cgtod(fs, c)	(cgstart(fs, c) + (fs)->fs_cblkno)	/* cg block */
584 #define	cgstart(fs, c)							\
585        ((fs)->fs_magic == FS_UFS2_MAGIC ? cgbase(fs, c) :		\
586        (cgbase(fs, c) + (fs)->fs_old_cgoffset * ((c) & ~((fs)->fs_old_cgmask))))
587 
588 /*
589  * Macros for handling inode numbers:
590  *     inode number to filesystem block offset.
591  *     inode number to cylinder group number.
592  *     inode number to filesystem block address.
593  */
594 #define	ino_to_cg(fs, x)	(((ino_t)(x)) / (fs)->fs_ipg)
595 #define	ino_to_fsba(fs, x)						\
596 	((ufs2_daddr_t)(cgimin(fs, ino_to_cg(fs, (ino_t)(x))) +		\
597 	    (blkstofrags((fs), ((((ino_t)(x)) % (fs)->fs_ipg) / INOPB(fs))))))
598 #define	ino_to_fsbo(fs, x)	(((ino_t)(x)) % INOPB(fs))
599 
600 /*
601  * Give cylinder group number for a filesystem block.
602  * Give cylinder group block number for a filesystem block.
603  */
604 #define	dtog(fs, d)	((d) / (fs)->fs_fpg)
605 #define	dtogd(fs, d)	((d) % (fs)->fs_fpg)
606 
607 /*
608  * Extract the bits for a block from a map.
609  * Compute the cylinder and rotational position of a cyl block addr.
610  */
611 #define	blkmap(fs, map, loc) \
612     (((map)[(loc) / NBBY] >> ((loc) % NBBY)) & (0xff >> (NBBY - (fs)->fs_frag)))
613 
614 /*
615  * The following macros optimize certain frequently calculated
616  * quantities by using shifts and masks in place of divisions
617  * modulos and multiplications.
618  */
619 #define	blkoff(fs, loc)		/* calculates (loc % fs->fs_bsize) */ \
620 	((loc) & (fs)->fs_qbmask)
621 #define	fragoff(fs, loc)	/* calculates (loc % fs->fs_fsize) */ \
622 	((loc) & (fs)->fs_qfmask)
623 #define	lfragtosize(fs, frag)	/* calculates ((off_t)frag * fs->fs_fsize) */ \
624 	(((off_t)(frag)) << (fs)->fs_fshift)
625 #define	lblktosize(fs, blk)	/* calculates ((off_t)blk * fs->fs_bsize) */ \
626 	(((off_t)(blk)) << (fs)->fs_bshift)
627 /* Use this only when `blk' is known to be small, e.g., < UFS_NDADDR. */
628 #define	smalllblktosize(fs, blk)    /* calculates (blk * fs->fs_bsize) */ \
629 	((blk) << (fs)->fs_bshift)
630 #define	lblkno(fs, loc)		/* calculates (loc / fs->fs_bsize) */ \
631 	((loc) >> (fs)->fs_bshift)
632 #define	numfrags(fs, loc)	/* calculates (loc / fs->fs_fsize) */ \
633 	((loc) >> (fs)->fs_fshift)
634 #define	blkroundup(fs, size)	/* calculates roundup(size, fs->fs_bsize) */ \
635 	(((size) + (fs)->fs_qbmask) & (fs)->fs_bmask)
636 #define	fragroundup(fs, size)	/* calculates roundup(size, fs->fs_fsize) */ \
637 	(((size) + (fs)->fs_qfmask) & (fs)->fs_fmask)
638 #define	fragstoblks(fs, frags)	/* calculates (frags / fs->fs_frag) */ \
639 	((frags) >> (fs)->fs_fragshift)
640 #define	blkstofrags(fs, blks)	/* calculates (blks * fs->fs_frag) */ \
641 	((blks) << (fs)->fs_fragshift)
642 #define	fragnum(fs, fsb)	/* calculates (fsb % fs->fs_frag) */ \
643 	((fsb) & ((fs)->fs_frag - 1))
644 #define	blknum(fs, fsb)		/* calculates rounddown(fsb, fs->fs_frag) */ \
645 	((fsb) &~ ((fs)->fs_frag - 1))
646 
647 /*
648  * Determine the number of available frags given a
649  * percentage to hold in reserve.
650  */
651 #define	freespace(fs, percentreserved) \
652 	(blkstofrags((fs), (fs)->fs_cstotal.cs_nbfree) + \
653 	(fs)->fs_cstotal.cs_nffree - \
654 	(((off_t)((fs)->fs_dsize)) * (percentreserved) / 100))
655 
656 /*
657  * Determining the size of a file block in the filesystem.
658  */
659 #define	blksize(fs, ip, lbn) \
660 	(((lbn) >= UFS_NDADDR || (ip)->i_size >= \
661 	    (uint64_t)smalllblktosize(fs, (lbn) + 1)) \
662 	    ? (fs)->fs_bsize \
663 	    : (fragroundup(fs, blkoff(fs, (ip)->i_size))))
664 #define	sblksize(fs, size, lbn) \
665 	(((lbn) >= UFS_NDADDR || (size) >= ((lbn) + 1) << (fs)->fs_bshift) \
666 	  ? (fs)->fs_bsize \
667 	  : (fragroundup(fs, blkoff(fs, (size)))))
668 
669 /*
670  * Number of indirects in a filesystem block.
671  */
672 #define	NINDIR(fs)	((fs)->fs_nindir)
673 
674 /*
675  * Indirect lbns are aligned on UFS_NDADDR addresses where single indirects
676  * are the negated address of the lowest lbn reachable, double indirects
677  * are this lbn - 1 and triple indirects are this lbn - 2.  This yields
678  * an unusual bit order to determine level.
679  */
680 /*static inline int
681 lbn_level(ufs_lbn_t lbn)
682 {
683 	if (lbn >= 0)
684 		return 0;
685 	switch (lbn & 0x3) {
686 	case 0:
687 		return (0);
688 	case 1:
689 		break;
690 	case 2:
691 		return (2);
692 	case 3:
693 		return (1);
694 	default:
695 		break;
696 	}
697 	return (-1);
698 }
699 
700 static inline ufs_lbn_t
701 lbn_offset(struct fs *fs, int level)
702 {
703 	ufs_lbn_t res;
704 
705 	for (res = 1; level > 0; level--)
706 		res *= NINDIR(fs);
707 	return (res);
708 }*/
709 
710 /*
711  * Number of inodes in a secondary storage block/fragment.
712  */
713 #define	INOPB(fs)	((fs)->fs_inopb)
714 #define	INOPF(fs)	((fs)->fs_inopb >> (fs)->fs_fragshift)
715 
716 /*
717  * Softdep journal record format.
718  */
719 
720 #define	JOP_ADDREF	1	/* Add a reference to an inode. */
721 #define	JOP_REMREF	2	/* Remove a reference from an inode. */
722 #define	JOP_NEWBLK	3	/* Allocate a block. */
723 #define	JOP_FREEBLK	4	/* Free a block or a tree of blocks. */
724 #define	JOP_MVREF	5	/* Move a reference from one off to another. */
725 #define	JOP_TRUNC	6	/* Partial truncation record. */
726 #define	JOP_SYNC	7	/* fsync() complete record. */
727 
728 #define	JREC_SIZE	32	/* Record and segment header size. */
729 
730 #define	SUJ_MIN		(4 * 1024 * 1024)	/* Minimum journal size */
731 #define	SUJ_MAX		(32 * 1024 * 1024)	/* Maximum journal size */
732 #define	SUJ_FILE	".sujournal"		/* Journal file name */
733 
734 /*
735  * Size of the segment record header.  There is at most one for each disk
736  * block in the journal.  The segment header is followed by an array of
737  * records.  fsck depends on the first element in each record being 'op'
738  * and the second being 'ino'.  Segments may span multiple disk blocks but
739  * the header is present on each.
740  */
741 struct jsegrec {
742 	uint64_t	jsr_seq;	/* Our sequence number */
743 	uint64_t	jsr_oldest;	/* Oldest valid sequence number */
744 	uint16_t	jsr_cnt;	/* Count of valid records */
745 	uint16_t	jsr_blocks;	/* Count of device bsize blocks. */
746 	uint32_t	jsr_crc;	/* 32bit crc of the valid space */
747 	int64_t		jsr_time;	/* timestamp for mount instance */
748 };
749 
750 /*
751  * Reference record.  Records a single link count modification.
752  */
753 struct jrefrec {
754 	uint32_t	jr_op;
755 	uint32_t	jr_ino;
756 	uint32_t	jr_parent;
757 	uint16_t	jr_nlink;
758 	uint16_t	jr_mode;
759 	int64_t		jr_diroff;
760 	uint64_t	jr_unused;
761 };
762 
763 /*
764  * Move record.  Records a reference moving within a directory block.  The
765  * nlink is unchanged but we must search both locations.
766  */
767 struct jmvrec {
768 	uint32_t	jm_op;
769 	uint32_t	jm_ino;
770 	uint32_t	jm_parent;
771 	uint16_t	jm_unused;
772 	int64_t		jm_oldoff;
773 	int64_t		jm_newoff;
774 };
775 
776 /*
777  * Block record.  A set of frags or tree of blocks starting at an indirect are
778  * freed or a set of frags are allocated.
779  */
780 struct jblkrec {
781 	uint32_t	jb_op;
782 	uint32_t	jb_ino;
783 	int64_t		jb_blkno;
784 	int64_t		jb_lbn;
785 	uint16_t	jb_frags;
786 	uint16_t	jb_oldfrags;
787 	uint32_t	jb_unused;
788 };
789 
790 /*
791  * Truncation record.  Records a partial truncation so that it may be
792  * completed at check time.  Also used for sync records.
793  */
794 struct jtrncrec {
795 	uint32_t	jt_op;
796 	uint32_t	jt_ino;
797 	int64_t		jt_size;
798 	uint32_t	jt_extsize;
799 	uint32_t	jt_pad[3];
800 };
801 
802 union jrec {
803 	struct	jsegrec		rec_jsegrec;
804 	struct	jrefrec		rec_jrefrec;
805 	struct	jmvrec		rec_jmvrec;
806 	struct	jblkrec		rec_jblkrec;
807 	struct	jtrncrec	rec_jtrncrec;
808 };
809 
810 #ifdef CTASSERT
811 CTASSERT(sizeof(struct jsegrec) == JREC_SIZE);
812 CTASSERT(sizeof(struct jrefrec) == JREC_SIZE);
813 CTASSERT(sizeof(struct jmvrec) == JREC_SIZE);
814 CTASSERT(sizeof(struct jblkrec) == JREC_SIZE);
815 CTASSERT(sizeof(struct jtrncrec) == JREC_SIZE);
816 CTASSERT(sizeof(union jrec) == JREC_SIZE);
817 #endif
818 
819 extern int inside[], around[];
820 extern u_char *fragtbl[];
821 
822 /*
823  * IOCTLs used for filesystem write suspension.
824  */
825 #define	UFSSUSPEND	_IOW('U', 1, fsid_t)
826 #define	UFSRESUME	_IO('U', 2)
827 
828 #endif
829 
830 
831 
832