lfs.h revision 1.178 1 /* $NetBSD: lfs.h,v 1.178 2015/08/12 18:27:01 dholland Exp $ */
2
3 /* from NetBSD: dinode.h,v 1.22 2013/01/22 09:39:18 dholland Exp */
4 /* from NetBSD: dir.h,v 1.21 2009/07/22 04:49:19 dholland Exp */
5
6 /*-
7 * Copyright (c) 1999, 2000, 2001, 2002, 2003 The NetBSD Foundation, Inc.
8 * All rights reserved.
9 *
10 * This code is derived from software contributed to The NetBSD Foundation
11 * by Konrad E. Schroder <perseant (at) hhhh.org>.
12 *
13 * Redistribution and use in source and binary forms, with or without
14 * modification, are permitted provided that the following conditions
15 * are met:
16 * 1. Redistributions of source code must retain the above copyright
17 * notice, this list of conditions and the following disclaimer.
18 * 2. Redistributions in binary form must reproduce the above copyright
19 * notice, this list of conditions and the following disclaimer in the
20 * documentation and/or other materials provided with the distribution.
21 *
22 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
23 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
24 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
25 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
26 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
27 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
28 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
29 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
30 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
31 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
32 * POSSIBILITY OF SUCH DAMAGE.
33 */
34 /*-
35 * Copyright (c) 1991, 1993
36 * The Regents of the University of California. All rights reserved.
37 *
38 * Redistribution and use in source and binary forms, with or without
39 * modification, are permitted provided that the following conditions
40 * are met:
41 * 1. Redistributions of source code must retain the above copyright
42 * notice, this list of conditions and the following disclaimer.
43 * 2. Redistributions in binary form must reproduce the above copyright
44 * notice, this list of conditions and the following disclaimer in the
45 * documentation and/or other materials provided with the distribution.
46 * 3. Neither the name of the University nor the names of its contributors
47 * may be used to endorse or promote products derived from this software
48 * without specific prior written permission.
49 *
50 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
51 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
52 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
53 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
54 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
55 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
56 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
57 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
58 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
59 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
60 * SUCH DAMAGE.
61 *
62 * @(#)lfs.h 8.9 (Berkeley) 5/8/95
63 */
64 /*
65 * Copyright (c) 2002 Networks Associates Technology, Inc.
66 * All rights reserved.
67 *
68 * This software was developed for the FreeBSD Project by Marshall
69 * Kirk McKusick and Network Associates Laboratories, the Security
70 * Research Division of Network Associates, Inc. under DARPA/SPAWAR
71 * contract N66001-01-C-8035 ("CBOSS"), as part of the DARPA CHATS
72 * research program
73 *
74 * Copyright (c) 1982, 1989, 1993
75 * The Regents of the University of California. All rights reserved.
76 * (c) UNIX System Laboratories, Inc.
77 * All or some portions of this file are derived from material licensed
78 * to the University of California by American Telephone and Telegraph
79 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
80 * the permission of UNIX System Laboratories, Inc.
81 *
82 * Redistribution and use in source and binary forms, with or without
83 * modification, are permitted provided that the following conditions
84 * are met:
85 * 1. Redistributions of source code must retain the above copyright
86 * notice, this list of conditions and the following disclaimer.
87 * 2. Redistributions in binary form must reproduce the above copyright
88 * notice, this list of conditions and the following disclaimer in the
89 * documentation and/or other materials provided with the distribution.
90 * 3. Neither the name of the University nor the names of its contributors
91 * may be used to endorse or promote products derived from this software
92 * without specific prior written permission.
93 *
94 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
95 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
96 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
97 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
98 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
99 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
100 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
101 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
102 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
103 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
104 * SUCH DAMAGE.
105 *
106 * @(#)dinode.h 8.9 (Berkeley) 3/29/95
107 */
108 /*
109 * Copyright (c) 1982, 1986, 1989, 1993
110 * The Regents of the University of California. All rights reserved.
111 * (c) UNIX System Laboratories, Inc.
112 * All or some portions of this file are derived from material licensed
113 * to the University of California by American Telephone and Telegraph
114 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
115 * the permission of UNIX System Laboratories, Inc.
116 *
117 * Redistribution and use in source and binary forms, with or without
118 * modification, are permitted provided that the following conditions
119 * are met:
120 * 1. Redistributions of source code must retain the above copyright
121 * notice, this list of conditions and the following disclaimer.
122 * 2. Redistributions in binary form must reproduce the above copyright
123 * notice, this list of conditions and the following disclaimer in the
124 * documentation and/or other materials provided with the distribution.
125 * 3. Neither the name of the University nor the names of its contributors
126 * may be used to endorse or promote products derived from this software
127 * without specific prior written permission.
128 *
129 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
130 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
131 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
132 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
133 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
134 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
135 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
136 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
137 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
138 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
139 * SUCH DAMAGE.
140 *
141 * @(#)dir.h 8.5 (Berkeley) 4/27/95
142 */
143
144 /*
145 * NOTE: COORDINATE ON-DISK FORMAT CHANGES WITH THE FREEBSD PROJECT.
146 */
147
148 #ifndef _UFS_LFS_LFS_H_
149 #define _UFS_LFS_LFS_H_
150
151 #if !defined(_KERNEL) && !defined(_STANDALONE)
152 #include <stddef.h> /* for offsetof */
153 #endif
154
155 #include <sys/rwlock.h>
156 #include <sys/mutex.h>
157 #include <sys/queue.h>
158 #include <sys/condvar.h>
159 #include <sys/mount.h>
160 #include <sys/pool.h>
161
162 /*
163 * Compile-time options for LFS.
164 */
165 #define LFS_IFIND_RETRIES 16
166 #define LFS_LOGLENGTH 1024 /* size of debugging log */
167 #define LFS_MAX_ACTIVE 10 /* Dirty segments before ckp forced */
168
169 /*
170 * Fixed filesystem layout parameters
171 */
172 #define LFS_LABELPAD 8192 /* LFS label size */
173 #define LFS_SBPAD 8192 /* LFS superblock size */
174
175 #define LFS_UNUSED_INUM 0 /* 0: out of band inode number */
176 #define LFS_IFILE_INUM 1 /* 1: IFILE inode number */
177 /* 2: Root inode number */
178 #define LFS_LOSTFOUNDINO 3 /* 3: lost+found inode number */
179 #define LFS_FIRST_INUM 4 /* 4: first free inode number */
180
181 /*
182 * The root inode is the root of the file system. Inode 0 can't be used for
183 * normal purposes and historically bad blocks were linked to inode 1, thus
184 * the root inode is 2. (Inode 1 is no longer used for this purpose, however
185 * numerous dump tapes make this assumption, so we are stuck with it).
186 */
187 #define ULFS_ROOTINO ((ino_t)2)
188
189 /*
190 * The Whiteout inode# is a dummy non-zero inode number which will
191 * never be allocated to a real file. It is used as a place holder
192 * in the directory entry which has been tagged as a LFS_DT_WHT entry.
193 * See the comments about ULFS_ROOTINO above.
194 */
195 #define ULFS_WINO ((ino_t)1)
196
197
198 #define LFS_V1_SUMMARY_SIZE 512 /* V1 fixed summary size */
199 #define LFS_DFL_SUMMARY_SIZE 512 /* Default summary size */
200
201 #define LFS_MAXNAMLEN 255 /* maximum name length in a dir */
202
203 #define ULFS_NXADDR 2
204 #define ULFS_NDADDR 12 /* Direct addresses in inode. */
205 #define ULFS_NIADDR 3 /* Indirect addresses in inode. */
206
207 /*
208 * Adjustable filesystem parameters
209 */
210 #ifndef LFS_ATIME_IFILE
211 # define LFS_ATIME_IFILE 0 /* Store atime info in ifile (optional in LFSv1) */
212 #endif
213 #define LFS_MARKV_MAXBLKCNT 65536 /* Max block count for lfs_markv() */
214
215 /*
216 * Directories
217 */
218
219 /*
220 * A directory consists of some number of blocks of LFS_DIRBLKSIZ
221 * bytes, where LFS_DIRBLKSIZ is chosen such that it can be transferred
222 * to disk in a single atomic operation (e.g. 512 bytes on most machines).
223 *
224 * Each LFS_DIRBLKSIZ byte block contains some number of directory entry
225 * structures, which are of variable length. Each directory entry has
226 * a struct lfs_direct at the front of it, containing its inode number,
227 * the length of the entry, and the length of the name contained in
228 * the entry. These are followed by the name padded to a 4 byte boundary.
229 * All names are guaranteed null terminated.
230 * The maximum length of a name in a directory is LFS_MAXNAMLEN.
231 *
232 * The macro DIRSIZ(fmt, dp) gives the amount of space required to represent
233 * a directory entry. Free space in a directory is represented by
234 * entries which have dp->d_reclen > DIRSIZ(fmt, dp). All LFS_DIRBLKSIZ bytes
235 * in a directory block are claimed by the directory entries. This
236 * usually results in the last entry in a directory having a large
237 * dp->d_reclen. When entries are deleted from a directory, the
238 * space is returned to the previous entry in the same directory
239 * block by increasing its dp->d_reclen. If the first entry of
240 * a directory block is free, then its dp->d_ino is set to 0.
241 * Entries other than the first in a directory do not normally have
242 * dp->d_ino set to 0.
243 */
244
245 /*
246 * Directory block size.
247 */
248 #undef LFS_DIRBLKSIZ
249 #define LFS_DIRBLKSIZ DEV_BSIZE
250
251 /*
252 * Convert between stat structure types and directory types.
253 */
254 #define LFS_IFTODT(mode) (((mode) & 0170000) >> 12)
255 #define LFS_DTTOIF(dirtype) ((dirtype) << 12)
256
257 /*
258 * The LFS_DIRSIZ macro gives the minimum record length which will hold
259 * the directory entry. This requires the amount of space in struct lfs_direct
260 * without the d_name field, plus enough space for the name with a terminating
261 * null byte (dp->d_namlen+1), rounded up to a 4 byte boundary.
262 */
263 #define LFS_DIRECTSIZ(namlen) \
264 ((sizeof(struct lfs_direct) - (LFS_MAXNAMLEN+1)) + (((namlen)+1 + 3) &~ 3))
265
266 #if (BYTE_ORDER == LITTLE_ENDIAN)
267 #define LFS_DIRSIZ(oldfmt, dp, needswap) \
268 (((oldfmt) && !(needswap)) ? \
269 LFS_DIRECTSIZ((dp)->d_type) : LFS_DIRECTSIZ((dp)->d_namlen))
270 #else
271 #define LFS_DIRSIZ(oldfmt, dp, needswap) \
272 (((oldfmt) && (needswap)) ? \
273 LFS_DIRECTSIZ((dp)->d_type) : LFS_DIRECTSIZ((dp)->d_namlen))
274 #endif
275
276 /* Constants for the first argument of LFS_DIRSIZ */
277 #define LFS_OLDDIRFMT 1
278 #define LFS_NEWDIRFMT 0
279
280 /*
281 * Theoretically, directories can be more than 2Gb in length; however, in
282 * practice this seems unlikely. So, we define the type doff_t as a 32-bit
283 * quantity to keep down the cost of doing lookup on a 32-bit machine.
284 */
285 #define doff_t int32_t
286 #define lfs_doff_t int32_t
287 #define LFS_MAXDIRSIZE (0x7fffffff)
288
289 /*
290 * File types for d_type
291 */
292 #define LFS_DT_UNKNOWN 0
293 #define LFS_DT_FIFO 1
294 #define LFS_DT_CHR 2
295 #define LFS_DT_DIR 4
296 #define LFS_DT_BLK 6
297 #define LFS_DT_REG 8
298 #define LFS_DT_LNK 10
299 #define LFS_DT_SOCK 12
300 #define LFS_DT_WHT 14
301
302 /*
303 * (See notes above)
304 */
305 #define d_ino d_fileno
306 struct lfs_direct {
307 u_int32_t d_fileno; /* inode number of entry */
308 u_int16_t d_reclen; /* length of this record */
309 u_int8_t d_type; /* file type, see below */
310 u_int8_t d_namlen; /* length of string in d_name */
311 char d_name[LFS_MAXNAMLEN + 1];/* name with length <= LFS_MAXNAMLEN */
312 };
313
314 /*
315 * Template for manipulating directories. Should use struct lfs_direct's,
316 * but the name field is LFS_MAXNAMLEN - 1, and this just won't do.
317 */
318 struct lfs_dirtemplate {
319 u_int32_t dot_ino;
320 int16_t dot_reclen;
321 u_int8_t dot_type;
322 u_int8_t dot_namlen;
323 char dot_name[4]; /* must be multiple of 4 */
324 u_int32_t dotdot_ino;
325 int16_t dotdot_reclen;
326 u_int8_t dotdot_type;
327 u_int8_t dotdot_namlen;
328 char dotdot_name[4]; /* ditto */
329 };
330
331 /*
332 * This is the old format of directories, sans type element.
333 */
334 struct lfs_odirtemplate {
335 u_int32_t dot_ino;
336 int16_t dot_reclen;
337 u_int16_t dot_namlen;
338 char dot_name[4]; /* must be multiple of 4 */
339 u_int32_t dotdot_ino;
340 int16_t dotdot_reclen;
341 u_int16_t dotdot_namlen;
342 char dotdot_name[4]; /* ditto */
343 };
344
345 /*
346 * Inodes
347 */
348
349 /*
350 * A dinode contains all the meta-data associated with a LFS file.
351 * This structure defines the on-disk format of a dinode. Since
352 * this structure describes an on-disk structure, all its fields
353 * are defined by types with precise widths.
354 */
355
356 struct ulfs1_dinode {
357 u_int16_t di_mode; /* 0: IFMT, permissions; see below. */
358 int16_t di_nlink; /* 2: File link count. */
359 u_int32_t di_inumber; /* 4: Inode number. */
360 u_int64_t di_size; /* 8: File byte count. */
361 int32_t di_atime; /* 16: Last access time. */
362 int32_t di_atimensec; /* 20: Last access time. */
363 int32_t di_mtime; /* 24: Last modified time. */
364 int32_t di_mtimensec; /* 28: Last modified time. */
365 int32_t di_ctime; /* 32: Last inode change time. */
366 int32_t di_ctimensec; /* 36: Last inode change time. */
367 int32_t di_db[ULFS_NDADDR]; /* 40: Direct disk blocks. */
368 int32_t di_ib[ULFS_NIADDR]; /* 88: Indirect disk blocks. */
369 u_int32_t di_flags; /* 100: Status flags (chflags). */
370 u_int32_t di_blocks; /* 104: Blocks actually held. */
371 int32_t di_gen; /* 108: Generation number. */
372 u_int32_t di_uid; /* 112: File owner. */
373 u_int32_t di_gid; /* 116: File group. */
374 u_int64_t di_modrev; /* 120: i_modrev for NFSv4 */
375 };
376
377 struct ulfs2_dinode {
378 u_int16_t di_mode; /* 0: IFMT, permissions; see below. */
379 int16_t di_nlink; /* 2: File link count. */
380 u_int32_t di_uid; /* 4: File owner. */
381 u_int32_t di_gid; /* 8: File group. */
382 u_int32_t di_blksize; /* 12: Inode blocksize. */
383 u_int64_t di_size; /* 16: File byte count. */
384 u_int64_t di_blocks; /* 24: Bytes actually held. */
385 int64_t di_atime; /* 32: Last access time. */
386 int64_t di_mtime; /* 40: Last modified time. */
387 int64_t di_ctime; /* 48: Last inode change time. */
388 int64_t di_birthtime; /* 56: Inode creation time. */
389 int32_t di_mtimensec; /* 64: Last modified time. */
390 int32_t di_atimensec; /* 68: Last access time. */
391 int32_t di_ctimensec; /* 72: Last inode change time. */
392 int32_t di_birthnsec; /* 76: Inode creation time. */
393 int32_t di_gen; /* 80: Generation number. */
394 u_int32_t di_kernflags; /* 84: Kernel flags. */
395 u_int32_t di_flags; /* 88: Status flags (chflags). */
396 int32_t di_extsize; /* 92: External attributes block. */
397 int64_t di_extb[ULFS_NXADDR];/* 96: External attributes block. */
398 int64_t di_db[ULFS_NDADDR]; /* 112: Direct disk blocks. */
399 int64_t di_ib[ULFS_NIADDR]; /* 208: Indirect disk blocks. */
400 u_int64_t di_modrev; /* 232: i_modrev for NFSv4 */
401 u_int32_t di_inumber; /* 240: Inode number */
402 u_int32_t di_spare[3]; /* 244: Reserved; currently unused */
403 };
404
405 /*
406 * The di_db fields may be overlaid with other information for
407 * file types that do not have associated disk storage. Block
408 * and character devices overlay the first data block with their
409 * dev_t value. Short symbolic links place their path in the
410 * di_db area.
411 */
412 #define di_rdev di_db[0]
413
414 /* Size of the on-disk inode. */
415 #define LFS_DINODE1_SIZE (sizeof(struct ulfs1_dinode)) /* 128 */
416 #define LFS_DINODE2_SIZE (sizeof(struct ulfs2_dinode))
417
418 /* File types, found in the upper bits of di_mode. */
419 #define LFS_IFMT 0170000 /* Mask of file type. */
420 #define LFS_IFIFO 0010000 /* Named pipe (fifo). */
421 #define LFS_IFCHR 0020000 /* Character device. */
422 #define LFS_IFDIR 0040000 /* Directory file. */
423 #define LFS_IFBLK 0060000 /* Block device. */
424 #define LFS_IFREG 0100000 /* Regular file. */
425 #define LFS_IFLNK 0120000 /* Symbolic link. */
426 #define LFS_IFSOCK 0140000 /* UNIX domain socket. */
427 #define LFS_IFWHT 0160000 /* Whiteout. */
428
429 /*
430 * "struct buf" associated definitions
431 */
432
433 /* Unassigned disk addresses. */
434 #define UNASSIGNED -1
435 #define UNWRITTEN -2
436
437 /* Unused logical block number */
438 #define LFS_UNUSED_LBN -1
439
440 /*
441 * "struct inode" associated definitions
442 */
443
444 /* For convenience */
445 #define IN_ALLMOD (IN_MODIFIED|IN_ACCESS|IN_CHANGE|IN_UPDATE|IN_MODIFY|IN_ACCESSED|IN_CLEANING)
446
447 /*
448 * On-disk and in-memory checkpoint segment usage structure.
449 */
450 typedef struct segusage SEGUSE;
451 struct segusage {
452 u_int32_t su_nbytes; /* 0: number of live bytes */
453 u_int32_t su_olastmod; /* 4: SEGUSE last modified timestamp */
454 u_int16_t su_nsums; /* 8: number of summaries in segment */
455 u_int16_t su_ninos; /* 10: number of inode blocks in seg */
456
457 #define SEGUSE_ACTIVE 0x01 /* segment currently being written */
458 #define SEGUSE_DIRTY 0x02 /* segment has data in it */
459 #define SEGUSE_SUPERBLOCK 0x04 /* segment contains a superblock */
460 #define SEGUSE_ERROR 0x08 /* cleaner: do not clean segment */
461 #define SEGUSE_EMPTY 0x10 /* segment is empty */
462 #define SEGUSE_INVAL 0x20 /* segment is invalid */
463 u_int32_t su_flags; /* 12: segment flags */
464 u_int64_t su_lastmod; /* 16: last modified timestamp */
465 };
466
467 typedef struct segusage_v1 SEGUSE_V1;
468 struct segusage_v1 {
469 u_int32_t su_nbytes; /* 0: number of live bytes */
470 u_int32_t su_lastmod; /* 4: SEGUSE last modified timestamp */
471 u_int16_t su_nsums; /* 8: number of summaries in segment */
472 u_int16_t su_ninos; /* 10: number of inode blocks in seg */
473 u_int32_t su_flags; /* 12: segment flags */
474 };
475
476 /*
477 * On-disk file information. One per file with data blocks in the segment.
478 *
479 * The FINFO structure is a header; it is followed by fi_nblocks block
480 * pointers, which are logical block numbers of the file. (These are the
481 * blocks of the file present in this segment.)
482 */
483
484 typedef struct finfo64 FINFO64;
485 struct finfo64 {
486 u_int32_t fi_nblocks; /* number of blocks */
487 u_int32_t fi_version; /* version number */
488 u_int64_t fi_ino; /* inode number */
489 u_int32_t fi_lastlength; /* length of last block in array */
490 u_int32_t fi_pad; /* unused */
491 };
492
493 typedef struct finfo32 FINFO32;
494 struct finfo32 {
495 u_int32_t fi_nblocks; /* number of blocks */
496 u_int32_t fi_version; /* version number */
497 u_int32_t fi_ino; /* inode number */
498 u_int32_t fi_lastlength; /* length of last block in array */
499 };
500
501 typedef union finfo {
502 struct finfo64 u_64;
503 struct finfo32 u_32;
504 } FINFO;
505
506 /*
507 * Index file inode entries.
508 */
509
510 /* magic value for daddrs */
511 #define LFS_UNUSED_DADDR 0 /* out-of-band daddr */
512 /* magic value for if_nextfree */
513 #define LFS_ORPHAN_NEXTFREE (~(u_int32_t)0) /* indicate orphaned file */
514
515 typedef struct ifile64 IFILE64;
516 struct ifile64 {
517 u_int32_t if_version; /* inode version number */
518 u_int32_t if_pad; /* 64-bit alignment padding */
519 int64_t if_daddr; /* inode disk address */
520 u_int64_t if_nextfree; /* next-unallocated inode */
521 u_int32_t if_atime_sec; /* Last access time, seconds */
522 u_int32_t if_atime_nsec; /* and nanoseconds */
523 };
524
525 typedef struct ifile32 IFILE32;
526 struct ifile32 {
527 u_int32_t if_version; /* inode version number */
528 int32_t if_daddr; /* inode disk address */
529 u_int32_t if_nextfree; /* next-unallocated inode */
530 u_int32_t if_atime_sec; /* Last access time, seconds */
531 u_int32_t if_atime_nsec; /* and nanoseconds */
532 };
533
534 typedef struct ifile_v1 IFILE_V1;
535 struct ifile_v1 {
536 u_int32_t if_version; /* inode version number */
537 int32_t if_daddr; /* inode disk address */
538 u_int32_t if_nextfree; /* next-unallocated inode */
539 #if LFS_ATIME_IFILE
540 #error "this cannot work"
541 struct timespec if_atime; /* Last access time */
542 #endif
543 };
544
545 /*
546 * Note: struct ifile_v1 is often handled by accessing the first three
547 * fields of struct ifile32. (XXX: Blah. This should be cleaned up as
548 * it may in some cases violate the strict-aliasing rules.)
549 */
550 typedef union ifile {
551 struct ifile64 u_64;
552 struct ifile32 u_32;
553 struct ifile_v1 u_v1;
554 } IFILE;
555
556 /*
557 * Cleaner information structure. This resides in the ifile and is used
558 * to pass information from the kernel to the cleaner.
559 */
560
561 /* flags for ->flags */
562 #define LFS_CLEANER_MUST_CLEAN 0x01
563
564 typedef struct _cleanerinfo32 {
565 u_int32_t clean; /* 0: number of clean segments */
566 u_int32_t dirty; /* 4: number of dirty segments */
567 int32_t bfree; /* 8: disk blocks free */
568 int32_t avail; /* 12: disk blocks available */
569 u_int32_t free_head; /* 16: head of the inode free list */
570 u_int32_t free_tail; /* 20: tail of the inode free list */
571 u_int32_t flags; /* 24: status word from the kernel */
572 } CLEANERINFO32;
573
574 typedef struct _cleanerinfo64 {
575 u_int32_t clean; /* 0: number of clean segments */
576 u_int32_t dirty; /* 4: number of dirty segments */
577 int64_t bfree; /* 8: disk blocks free */
578 int64_t avail; /* 16: disk blocks available */
579 u_int64_t free_head; /* 24: head of the inode free list */
580 u_int64_t free_tail; /* 32: tail of the inode free list */
581 u_int32_t flags; /* 40: status word from the kernel */
582 u_int32_t pad; /* 44: must be 64-bit aligned */
583 } CLEANERINFO64;
584
585 /* this must not go to disk directly of course */
586 typedef union _cleanerinfo {
587 CLEANERINFO32 u_32;
588 CLEANERINFO64 u_64;
589 } CLEANERINFO;
590
591 /*
592 * On-disk segment summary information
593 */
594
595 /* magic value for ss_magic */
596 #define SS_MAGIC 0x061561
597
598 /* flags for ss_flags */
599 #define SS_DIROP 0x01 /* segment begins a dirop */
600 #define SS_CONT 0x02 /* more partials to finish this write*/
601 #define SS_CLEAN 0x04 /* written by the cleaner */
602 #define SS_RFW 0x08 /* written by the roll-forward agent */
603 #define SS_RECLAIM 0x10 /* written by the roll-forward agent */
604
605 typedef struct segsum_v1 SEGSUM_V1;
606 struct segsum_v1 {
607 u_int32_t ss_sumsum; /* 0: check sum of summary block */
608 u_int32_t ss_datasum; /* 4: check sum of data */
609 u_int32_t ss_magic; /* 8: segment summary magic number */
610 int32_t ss_next; /* 12: next segment */
611 u_int32_t ss_create; /* 16: creation time stamp */
612 u_int16_t ss_nfinfo; /* 20: number of file info structures */
613 u_int16_t ss_ninos; /* 22: number of inodes in summary */
614 u_int16_t ss_flags; /* 24: used for directory operations */
615 u_int16_t ss_pad; /* 26: extra space */
616 /* FINFO's and inode daddr's... */
617 };
618
619 typedef struct segsum32 SEGSUM32;
620 struct segsum32 {
621 u_int32_t ss_sumsum; /* 0: check sum of summary block */
622 u_int32_t ss_datasum; /* 4: check sum of data */
623 u_int32_t ss_magic; /* 8: segment summary magic number */
624 int32_t ss_next; /* 12: next segment (disk address) */
625 u_int32_t ss_ident; /* 16: roll-forward fsid */
626 u_int16_t ss_nfinfo; /* 20: number of file info structures */
627 u_int16_t ss_ninos; /* 22: number of inodes in summary */
628 u_int16_t ss_flags; /* 24: used for directory operations */
629 u_int8_t ss_pad[2]; /* 26: extra space */
630 u_int32_t ss_reclino; /* 28: inode being reclaimed */
631 u_int64_t ss_serial; /* 32: serial number */
632 u_int64_t ss_create; /* 40: time stamp */
633 /* FINFO's and inode daddr's... */
634 };
635
636 typedef struct segsum64 SEGSUM64;
637 struct segsum64 {
638 u_int32_t ss_sumsum; /* 0: check sum of summary block */
639 u_int32_t ss_datasum; /* 4: check sum of data */
640 u_int32_t ss_magic; /* 8: segment summary magic number */
641 u_int32_t ss_ident; /* 12: roll-forward fsid */
642 int64_t ss_next; /* 16: next segment (disk address) */
643 u_int16_t ss_nfinfo; /* 24: number of file info structures */
644 u_int16_t ss_ninos; /* 26: number of inodes in summary */
645 u_int16_t ss_flags; /* 28: used for directory operations */
646 u_int8_t ss_pad[2]; /* 30: extra space */
647 u_int64_t ss_reclino; /* 32: inode being reclaimed */
648 u_int64_t ss_serial; /* 40: serial number */
649 u_int64_t ss_create; /* 48: time stamp */
650 /* FINFO's and inode daddr's... */
651 };
652
653 typedef union segsum SEGSUM;
654 union segsum {
655 struct segsum64 u_64;
656 struct segsum32 u_32;
657 struct segsum_v1 u_v1;
658 };
659
660
661 /*
662 * On-disk super block.
663 *
664 * We have separate superblock structures for the 32-bit and 64-bit
665 * LFS, and accessor functions to hide the differences.
666 *
667 * For lfs64, the format version is always 2; version 1 lfs is old.
668 * For both, the inode format version is 0; for lfs32 this selects the
669 * same 32-bit inode as always, and for lfs64 this selects the larger
670 * 64-bit inode structure we got from ffsv2.
671 *
672 * In lfs64:
673 * - inode numbers are still 32 bit
674 * - segments may not be larger than 4G (counted in bytes)
675 * - there may not be more than 2^32 (or perhaps 2^31) segments
676 * - the total volume size is limited to 2^63 frags and/or 2^63
677 * disk blocks, and probably in practice 2^63 bytes.
678 */
679
680 #define LFS_MAGIC 0x070162
681 #define LFS_MAGIC_SWAPPED 0x62010700
682
683 #define LFS64_MAGIC 0x19620701
684 #define LFS64_MAGIC_SWAPPED 0x01076219
685
686 #define LFS_VERSION 2
687
688 #define LFS_MIN_SBINTERVAL 5 /* min superblock segment spacing */
689 #define LFS_MAXNUMSB 10 /* max number of superblocks */
690
691 /* flags for dlfs_pflags */
692 #define LFS_PF_CLEAN 0x1
693
694 /* Inode format versions */
695 #define LFS_44INODEFMT 0
696 #define LFS_MAXINODEFMT 0
697
698 struct dlfs {
699 u_int32_t dlfs_magic; /* 0: magic number */
700 u_int32_t dlfs_version; /* 4: version number */
701
702 u_int32_t dlfs_size; /* 8: number of blocks in fs (v1) */
703 /* number of frags in fs (v2) */
704 u_int32_t dlfs_ssize; /* 12: number of blocks per segment (v1) */
705 /* number of bytes per segment (v2) */
706 u_int32_t dlfs_dsize; /* 16: number of disk blocks in fs */
707 u_int32_t dlfs_bsize; /* 20: file system block size */
708 u_int32_t dlfs_fsize; /* 24: size of frag blocks in fs */
709 u_int32_t dlfs_frag; /* 28: number of frags in a block in fs */
710
711 /* Checkpoint region. */
712 u_int32_t dlfs_freehd; /* 32: start of the free inode list */
713 int32_t dlfs_bfree; /* 36: number of free frags */
714 u_int32_t dlfs_nfiles; /* 40: number of allocated inodes */
715 int32_t dlfs_avail; /* 44: blocks available for writing */
716 int32_t dlfs_uinodes; /* 48: inodes in cache not yet on disk */
717 int32_t dlfs_idaddr; /* 52: inode file disk address */
718 u_int32_t dlfs_ifile; /* 56: inode file inode number */
719 int32_t dlfs_lastseg; /* 60: address of last segment written */
720 int32_t dlfs_nextseg; /* 64: address of next segment to write */
721 int32_t dlfs_curseg; /* 68: current segment being written */
722 int32_t dlfs_offset; /* 72: offset in curseg for next partial */
723 int32_t dlfs_lastpseg; /* 76: address of last partial written */
724 u_int32_t dlfs_inopf; /* 80: v1: time stamp; v2: inodes per frag */
725
726 /* These are configuration parameters. */
727 u_int32_t dlfs_minfree; /* 84: minimum percentage of free blocks */
728
729 /* These fields can be computed from the others. */
730 u_int64_t dlfs_maxfilesize; /* 88: maximum representable file size */
731 u_int32_t dlfs_fsbpseg; /* 96: frags (fsb) per segment */
732 u_int32_t dlfs_inopb; /* 100: inodes per block */
733 u_int32_t dlfs_ifpb; /* 104: IFILE entries per block */
734 u_int32_t dlfs_sepb; /* 108: SEGUSE entries per block */
735 u_int32_t dlfs_nindir; /* 112: indirect pointers per block */
736 u_int32_t dlfs_nseg; /* 116: number of segments */
737 u_int32_t dlfs_nspf; /* 120: number of sectors per fragment */
738 u_int32_t dlfs_cleansz; /* 124: cleaner info size in blocks */
739 u_int32_t dlfs_segtabsz; /* 128: segment table size in blocks */
740 u_int32_t dlfs_segmask; /* 132: calculate offset within a segment */
741 u_int32_t dlfs_segshift; /* 136: fast mult/div for segments */
742 u_int32_t dlfs_bshift; /* 140: calc block number from file offset */
743 u_int32_t dlfs_ffshift; /* 144: fast mult/div for frag from file */
744 u_int32_t dlfs_fbshift; /* 148: fast mult/div for frag from block */
745 u_int64_t dlfs_bmask; /* 152: calc block offset from file offset */
746 u_int64_t dlfs_ffmask; /* 160: calc frag offset from file offset */
747 u_int64_t dlfs_fbmask; /* 168: calc frag offset from block offset */
748 u_int32_t dlfs_blktodb; /* 176: blktodb and dbtoblk shift constant */
749 u_int32_t dlfs_sushift; /* 180: fast mult/div for segusage table */
750
751 int32_t dlfs_maxsymlinklen; /* 184: max length of an internal symlink */
752 /* 188: superblock disk offsets */
753 int32_t dlfs_sboffs[LFS_MAXNUMSB];
754
755 u_int32_t dlfs_nclean; /* 228: Number of clean segments */
756 u_char dlfs_fsmnt[MNAMELEN]; /* 232: name mounted on */
757 u_int16_t dlfs_pflags; /* 322: file system persistent flags */
758 int32_t dlfs_dmeta; /* 324: total number of dirty summaries */
759 u_int32_t dlfs_minfreeseg; /* 328: segments not counted in bfree */
760 u_int32_t dlfs_sumsize; /* 332: size of summary blocks */
761 u_int64_t dlfs_serial; /* 336: serial number */
762 u_int32_t dlfs_ibsize; /* 344: size of inode blocks */
763 int32_t dlfs_s0addr; /* 348: start of segment 0 */
764 u_int64_t dlfs_tstamp; /* 352: time stamp */
765 u_int32_t dlfs_inodefmt; /* 360: inode format version */
766 u_int32_t dlfs_interleave; /* 364: segment interleave */
767 u_int32_t dlfs_ident; /* 368: per-fs identifier */
768 u_int32_t dlfs_fsbtodb; /* 372: fsbtodb and dbtodsb shift constant */
769 u_int32_t dlfs_resvseg; /* 376: segments reserved for the cleaner */
770 int8_t dlfs_pad[128]; /* 380: round to 512 bytes */
771 /* Checksum -- last valid disk field. */
772 u_int32_t dlfs_cksum; /* 508: checksum for superblock checking */
773 };
774
775 struct dlfs64 {
776 u_int32_t dlfs_magic; /* 0: magic number */
777 u_int32_t dlfs_version; /* 4: version number (2) */
778
779 u_int64_t dlfs_size; /* 8: number of frags in fs (v2) */
780 u_int64_t dlfs_dsize; /* 16: number of disk blocks in fs */
781 u_int32_t dlfs_ssize; /* 24: number of bytes per segment (v2) */
782 u_int32_t dlfs_bsize; /* 28: file system block size */
783 u_int32_t dlfs_fsize; /* 32: size of frag blocks in fs */
784 u_int32_t dlfs_frag; /* 36: number of frags in a block in fs */
785
786 /* Checkpoint region. */
787 u_int32_t dlfs_freehd; /* 40: start of the free inode list */
788 u_int32_t dlfs_nfiles; /* 44: number of allocated inodes */
789 int64_t dlfs_bfree; /* 48: number of free frags */
790 int64_t dlfs_avail; /* 56: blocks available for writing */
791 int64_t dlfs_idaddr; /* 64: inode file disk address */
792 int32_t dlfs_uinodes; /* 72: inodes in cache not yet on disk */
793 u_int32_t dlfs_ifile; /* 76: inode file inode number */
794 int64_t dlfs_lastseg; /* 80: address of last segment written */
795 int64_t dlfs_nextseg; /* 88: address of next segment to write */
796 int64_t dlfs_curseg; /* 96: current segment being written */
797 int64_t dlfs_offset; /* 104: offset in curseg for next partial */
798 int64_t dlfs_lastpseg; /* 112: address of last partial written */
799 u_int32_t dlfs_inopf; /* 120: inodes per frag */
800
801 /* These are configuration parameters. */
802 u_int32_t dlfs_minfree; /* 124: minimum percentage of free blocks */
803
804 /* These fields can be computed from the others. */
805 u_int64_t dlfs_maxfilesize; /* 128: maximum representable file size */
806 u_int32_t dlfs_fsbpseg; /* 136: frags (fsb) per segment */
807 u_int32_t dlfs_inopb; /* 140: inodes per block */
808 u_int32_t dlfs_ifpb; /* 144: IFILE entries per block */
809 u_int32_t dlfs_sepb; /* 148: SEGUSE entries per block */
810 u_int32_t dlfs_nindir; /* 152: indirect pointers per block */
811 u_int32_t dlfs_nseg; /* 156: number of segments */
812 u_int32_t dlfs_nspf; /* 160: number of sectors per fragment */
813 u_int32_t dlfs_cleansz; /* 164: cleaner info size in blocks */
814 u_int32_t dlfs_segtabsz; /* 168: segment table size in blocks */
815 u_int32_t dlfs_bshift; /* 172: calc block number from file offset */
816 u_int32_t dlfs_ffshift; /* 176: fast mult/div for frag from file */
817 u_int32_t dlfs_fbshift; /* 180: fast mult/div for frag from block */
818 u_int64_t dlfs_bmask; /* 184: calc block offset from file offset */
819 u_int64_t dlfs_ffmask; /* 192: calc frag offset from file offset */
820 u_int64_t dlfs_fbmask; /* 200: calc frag offset from block offset */
821 u_int32_t dlfs_blktodb; /* 208: blktodb and dbtoblk shift constant */
822 u_int32_t dlfs_sushift; /* 212: fast mult/div for segusage table */
823
824 /* 216: superblock disk offsets */
825 int64_t dlfs_sboffs[LFS_MAXNUMSB];
826
827 int32_t dlfs_maxsymlinklen; /* 296: max len of an internal symlink */
828 u_int32_t dlfs_nclean; /* 300: Number of clean segments */
829 u_char dlfs_fsmnt[MNAMELEN]; /* 304: name mounted on */
830 u_int16_t dlfs_pflags; /* 394: file system persistent flags */
831 int32_t dlfs_dmeta; /* 396: total number of dirty summaries */
832 u_int32_t dlfs_minfreeseg; /* 400: segments not counted in bfree */
833 u_int32_t dlfs_sumsize; /* 404: size of summary blocks */
834 u_int32_t dlfs_ibsize; /* 408: size of inode blocks */
835 u_int32_t dlfs_inodefmt; /* 412: inode format version */
836 u_int64_t dlfs_serial; /* 416: serial number */
837 int64_t dlfs_s0addr; /* 424: start of segment 0 */
838 u_int64_t dlfs_tstamp; /* 432: time stamp */
839 u_int32_t dlfs_interleave; /* 440: segment interleave */
840 u_int32_t dlfs_ident; /* 444: per-fs identifier */
841 u_int32_t dlfs_fsbtodb; /* 448: fsbtodb and dbtodsb shift constant */
842 u_int32_t dlfs_resvseg; /* 452: segments reserved for the cleaner */
843 int8_t dlfs_pad[52]; /* 456: round to 512 bytes */
844 /* Checksum -- last valid disk field. */
845 u_int32_t dlfs_cksum; /* 508: checksum for superblock checking */
846 };
847
848 /* Type used for the inode bitmap */
849 typedef u_int32_t lfs_bm_t;
850
851 /*
852 * Linked list of segments whose byte count needs updating following a
853 * file truncation.
854 */
855 struct segdelta {
856 long segnum;
857 size_t num;
858 LIST_ENTRY(segdelta) list;
859 };
860
861 /*
862 * In-memory super block.
863 */
864 struct lfs {
865 union { /* on-disk parameters */
866 struct dlfs u_32;
867 struct dlfs64 u_64;
868 } lfs_dlfs_u;
869
870 /* These fields are set at mount time and are meaningless on disk. */
871 unsigned lfs_is64 : 1; /* are we lfs64 or lfs32? */
872
873 struct segment *lfs_sp; /* current segment being written */
874 struct vnode *lfs_ivnode; /* vnode for the ifile */
875 u_int32_t lfs_seglock; /* single-thread the segment writer */
876 pid_t lfs_lockpid; /* pid of lock holder */
877 lwpid_t lfs_locklwp; /* lwp of lock holder */
878 u_int32_t lfs_iocount; /* number of ios pending */
879 u_int32_t lfs_writer; /* don't allow any dirops to start */
880 u_int32_t lfs_dirops; /* count of active directory ops */
881 u_int32_t lfs_dirvcount; /* count of VDIROP nodes in this fs */
882 u_int32_t lfs_doifile; /* Write ifile blocks on next write */
883 u_int32_t lfs_nactive; /* Number of segments since last ckp */
884 int8_t lfs_fmod; /* super block modified flag */
885 int8_t lfs_ronly; /* mounted read-only flag */
886 #define LFS_NOTYET 0x01
887 #define LFS_IFDIRTY 0x02
888 #define LFS_WARNED 0x04
889 #define LFS_UNDIROP 0x08
890 int8_t lfs_flags; /* currently unused flag */
891 u_int16_t lfs_activesb; /* toggle between superblocks */
892 daddr_t lfs_sbactive; /* disk address of current sb write */
893 struct vnode *lfs_flushvp; /* vnode being flushed */
894 int lfs_flushvp_fakevref; /* fake vref count for flushvp */
895 struct vnode *lfs_unlockvp; /* being inactivated in lfs_segunlock */
896 u_int32_t lfs_diropwait; /* # procs waiting on dirop flush */
897 size_t lfs_devbsize; /* Device block size */
898 size_t lfs_devbshift; /* Device block shift */
899 krwlock_t lfs_fraglock;
900 krwlock_t lfs_iflock; /* Ifile lock */
901 kcondvar_t lfs_stopcv; /* Wrap lock */
902 struct lwp *lfs_stoplwp;
903 pid_t lfs_rfpid; /* Process ID of roll-forward agent */
904 int lfs_nadirop; /* number of active dirop nodes */
905 long lfs_ravail; /* blocks pre-reserved for writing */
906 long lfs_favail; /* blocks pre-reserved for writing */
907 struct lfs_res_blk *lfs_resblk; /* Reserved memory for pageout */
908 TAILQ_HEAD(, inode) lfs_dchainhd; /* dirop vnodes */
909 TAILQ_HEAD(, inode) lfs_pchainhd; /* paging vnodes */
910 #define LFS_RESHASH_WIDTH 17
911 LIST_HEAD(, lfs_res_blk) lfs_reshash[LFS_RESHASH_WIDTH];
912 int lfs_pdflush; /* pagedaemon wants us to flush */
913 u_int32_t **lfs_suflags; /* Segment use flags */
914 #ifdef _KERNEL
915 struct pool lfs_clpool; /* Pool for struct lfs_cluster */
916 struct pool lfs_bpppool; /* Pool for bpp */
917 struct pool lfs_segpool; /* Pool for struct segment */
918 #endif /* _KERNEL */
919 #define LFS_MAX_CLEANIND 64
920 daddr_t lfs_cleanint[LFS_MAX_CLEANIND]; /* Active cleaning intervals */
921 int lfs_cleanind; /* Index into intervals */
922 int lfs_sleepers; /* # procs sleeping this fs */
923 int lfs_pages; /* dirty pages blaming this fs */
924 lfs_bm_t *lfs_ino_bitmap; /* Inuse inodes bitmap */
925 int lfs_nowrap; /* Suspend log wrap */
926 int lfs_wrappass; /* Allow first log wrap requester to pass */
927 int lfs_wrapstatus; /* Wrap status */
928 int lfs_reclino; /* Inode being reclaimed */
929 daddr_t lfs_startseg; /* Segment we started writing at */
930 LIST_HEAD(, segdelta) lfs_segdhd; /* List of pending trunc accounting events */
931
932 #ifdef _KERNEL
933 /* ULFS-level information */
934 u_int32_t um_flags; /* ULFS flags (below) */
935 u_long um_nindir; /* indirect ptrs per block */
936 u_long um_lognindir; /* log2 of um_nindir */
937 u_long um_bptrtodb; /* indir ptr to disk block */
938 u_long um_seqinc; /* inc between seq blocks */
939 int um_maxsymlinklen;
940 int um_dirblksiz;
941 u_int64_t um_maxfilesize;
942
943 /* Stuff used by quota2 code, not currently operable */
944 unsigned lfs_use_quota2 : 1;
945 uint32_t lfs_quota_magic;
946 uint8_t lfs_quota_flags;
947 uint64_t lfs_quotaino[2];
948
949 /* Sleep address replacing &lfs_avail inside the on-disk superblock */
950 /* XXX: should be replaced with a condvar */
951 int lfs_availsleep;
952 /* This one replaces &lfs_nextseg... all ditto */
953 int lfs_nextsegsleep;
954 #endif
955 };
956
957 /*
958 * Structures used by lfs_bmapv and lfs_markv to communicate information
959 * about inodes and data blocks.
960 */
961 typedef struct block_info {
962 u_int64_t bi_inode; /* inode # */
963 int64_t bi_lbn; /* logical block w/in file */
964 int64_t bi_daddr; /* disk address of block */
965 u_int64_t bi_segcreate; /* origin segment create time */
966 int bi_version; /* file version number */
967 int bi_size; /* size of the block (if fragment) */
968 void *bi_bp; /* data buffer */
969 } BLOCK_INFO;
970
971 /* Compatibility for 7.0 binaries */
972 typedef struct block_info_70 {
973 u_int32_t bi_inode; /* inode # */
974 int32_t bi_lbn; /* logical block w/in file */
975 int32_t bi_daddr; /* disk address of block */
976 u_int64_t bi_segcreate; /* origin segment create time */
977 int bi_version; /* file version number */
978 void *bi_bp; /* data buffer */
979 int bi_size; /* size of the block (if fragment) */
980 } BLOCK_INFO_70;
981
982 /* Compatibility for 1.5 binaries */
983 typedef struct block_info_15 {
984 u_int32_t bi_inode; /* inode # */
985 int32_t bi_lbn; /* logical block w/in file */
986 int32_t bi_daddr; /* disk address of block */
987 u_int32_t bi_segcreate; /* origin segment create time */
988 int bi_version; /* file version number */
989 void *bi_bp; /* data buffer */
990 int bi_size; /* size of the block (if fragment) */
991 } BLOCK_INFO_15;
992
993 /*
994 * 32/64-bit-clean pointer to block pointers. This points into
995 * already-existing storage; it is mostly used to access the block
996 * pointers following a FINFO.
997 */
998 union lfs_blocks {
999 int64_t *b64;
1000 int32_t *b32;
1001 };
1002
1003 /* In-memory description of a segment about to be written. */
1004 struct segment {
1005 struct lfs *fs; /* file system pointer */
1006 struct buf **bpp; /* pointer to buffer array */
1007 struct buf **cbpp; /* pointer to next available bp */
1008 struct buf **start_bpp; /* pointer to first bp in this set */
1009 struct buf *ibp; /* buffer pointer to inode page */
1010 struct ulfs1_dinode *idp; /* pointer to ifile dinode */
1011 FINFO *fip; /* current fileinfo pointer */
1012 struct vnode *vp; /* vnode being gathered */
1013 void *segsum; /* segment summary info */
1014 u_int32_t ninodes; /* number of inodes in this segment */
1015 int32_t seg_bytes_left; /* bytes left in segment */
1016 int32_t sum_bytes_left; /* bytes left in summary block */
1017 u_int32_t seg_number; /* number of this segment */
1018 union lfs_blocks start_lbp; /* beginning lbn for this set */
1019
1020 #define SEGM_CKP 0x0001 /* doing a checkpoint */
1021 #define SEGM_CLEAN 0x0002 /* cleaner call; don't sort */
1022 #define SEGM_SYNC 0x0004 /* wait for segment */
1023 #define SEGM_PROT 0x0008 /* don't inactivate at segunlock */
1024 #define SEGM_PAGEDAEMON 0x0010 /* pagedaemon called us */
1025 #define SEGM_WRITERD 0x0020 /* LFS writed called us */
1026 #define SEGM_FORCE_CKP 0x0040 /* Force checkpoint right away */
1027 #define SEGM_RECLAIM 0x0080 /* Writing to reclaim vnode */
1028 #define SEGM_SINGLE 0x0100 /* Opportunistic writevnodes */
1029 u_int16_t seg_flags; /* run-time flags for this segment */
1030 u_int32_t seg_iocount; /* number of ios pending */
1031 int ndupino; /* number of duplicate inodes */
1032 };
1033
1034 /* Statistics Counters */
1035 struct lfs_stats { /* Must match sysctl list in lfs_vfsops.h ! */
1036 u_int segsused;
1037 u_int psegwrites;
1038 u_int psyncwrites;
1039 u_int pcleanwrites;
1040 u_int blocktot;
1041 u_int cleanblocks;
1042 u_int ncheckpoints;
1043 u_int nwrites;
1044 u_int nsync_writes;
1045 u_int wait_exceeded;
1046 u_int write_exceeded;
1047 u_int flush_invoked;
1048 u_int vflush_invoked;
1049 u_int clean_inlocked;
1050 u_int clean_vnlocked;
1051 u_int segs_reclaimed;
1052 };
1053
1054 /* Fcntls to take the place of the lfs syscalls */
1055 struct lfs_fcntl_markv {
1056 BLOCK_INFO *blkiov; /* blocks to relocate */
1057 int blkcnt; /* number of blocks (limited to 65536) */
1058 };
1059
1060 #define LFCNSEGWAITALL _FCNR_FSPRIV('L', 14, struct timeval)
1061 #define LFCNSEGWAIT _FCNR_FSPRIV('L', 15, struct timeval)
1062 #define LFCNBMAPV _FCNRW_FSPRIV('L', 16, struct lfs_fcntl_markv)
1063 #define LFCNMARKV _FCNRW_FSPRIV('L', 17, struct lfs_fcntl_markv)
1064 #define LFCNRECLAIM _FCNO_FSPRIV('L', 4)
1065
1066 struct lfs_fhandle {
1067 char space[28]; /* FHANDLE_SIZE_COMPAT (but used from userland too) */
1068 };
1069 #define LFCNREWIND _FCNR_FSPRIV('L', 6, int)
1070 #define LFCNINVAL _FCNR_FSPRIV('L', 7, int)
1071 #define LFCNRESIZE _FCNR_FSPRIV('L', 8, int)
1072 #define LFCNWRAPSTOP _FCNR_FSPRIV('L', 9, int)
1073 #define LFCNWRAPGO _FCNR_FSPRIV('L', 10, int)
1074 #define LFCNIFILEFH _FCNW_FSPRIV('L', 11, struct lfs_fhandle)
1075 #define LFCNWRAPPASS _FCNR_FSPRIV('L', 12, int)
1076 # define LFS_WRAP_GOING 0x0
1077 # define LFS_WRAP_WAITING 0x1
1078 #define LFCNWRAPSTATUS _FCNW_FSPRIV('L', 13, int)
1079
1080 /* Debug segment lock */
1081 #ifdef notyet
1082 # define ASSERT_SEGLOCK(fs) KASSERT(LFS_SEGLOCK_HELD(fs))
1083 # define ASSERT_NO_SEGLOCK(fs) KASSERT(!LFS_SEGLOCK_HELD(fs))
1084 # define ASSERT_DUNNO_SEGLOCK(fs)
1085 # define ASSERT_MAYBE_SEGLOCK(fs)
1086 #else /* !notyet */
1087 # define ASSERT_DUNNO_SEGLOCK(fs) \
1088 DLOG((DLOG_SEG, "lfs func %s seglock wrong (%d)\n", __func__, \
1089 LFS_SEGLOCK_HELD(fs)))
1090 # define ASSERT_SEGLOCK(fs) do { \
1091 if (!LFS_SEGLOCK_HELD(fs)) { \
1092 DLOG((DLOG_SEG, "lfs func %s seglock wrong (0)\n", __func__)); \
1093 } \
1094 } while(0)
1095 # define ASSERT_NO_SEGLOCK(fs) do { \
1096 if (LFS_SEGLOCK_HELD(fs)) { \
1097 DLOG((DLOG_SEG, "lfs func %s seglock wrong (1)\n", __func__)); \
1098 } \
1099 } while(0)
1100 # define ASSERT_MAYBE_SEGLOCK(x)
1101 #endif /* !notyet */
1102
1103 /*
1104 * Arguments to mount LFS filesystems
1105 */
1106 struct ulfs_args {
1107 char *fspec; /* block special device to mount */
1108 };
1109
1110 __BEGIN_DECLS
1111 void lfs_itimes(struct inode *, const struct timespec *,
1112 const struct timespec *, const struct timespec *);
1113 __END_DECLS
1114
1115 #endif /* !_UFS_LFS_LFS_H_ */
1116