lfs.h revision 1.10 1 /* $NetBSD: lfs.h,v 1.10 1998/09/11 21:27:12 pk Exp $ */
2
3 /*-
4 * Copyright (c) 1991, 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. All advertising materials mentioning features or use of this software
16 * must display the following acknowledgement:
17 * This product includes software developed by the University of
18 * California, Berkeley and its contributors.
19 * 4. Neither the name of the University nor the names of its contributors
20 * may be used to endorse or promote products derived from this software
21 * without specific prior written permission.
22 *
23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * SUCH DAMAGE.
34 *
35 * @(#)lfs.h 8.9 (Berkeley) 5/8/95
36 */
37
38 #define LFS_LABELPAD 8192 /* LFS label size */
39 #define LFS_SBPAD 8192 /* LFS superblock size */
40
41 /*
42 * XXX
43 * This is a kluge and NEEDS to go away.
44 *
45 * Right now, ufs code handles most of the calls for directory operations
46 * such as create, mkdir, link, etc. As a result VOP_UPDATE is being
47 * called with waitfor set (since ffs does these things synchronously).
48 * Since LFS does not want to do these synchronously, we treat the last
49 * argument to lfs_update as a set of flags. If LFS_SYNC is set, then
50 * the update should be synchronous, if not, do it asynchronously.
51 * Unfortunately, this means that LFS won't work with NFS yet because
52 * NFS goes through paths that will make normal calls to ufs which will
53 * call lfs with a last argument of 1.
54 */
55 #define LFS_SYNC 0x02
56
57 /* On-disk and in-memory checkpoint segment usage structure. */
58 typedef struct segusage SEGUSE;
59 struct segusage {
60 u_int32_t su_nbytes; /* number of live bytes */
61 u_int32_t su_lastmod; /* SEGUSE last modified timestamp */
62 u_int16_t su_nsums; /* number of summaries in segment */
63 u_int16_t su_ninos; /* number of inode blocks in seg */
64
65 #define SEGUSE_ACTIVE 0x01 /* segment is currently being written */
66 #define SEGUSE_DIRTY 0x02 /* segment has data in it */
67 #define SEGUSE_SUPERBLOCK 0x04 /* segment contains a superblock */
68 u_int32_t su_flags;
69 };
70
71 #define SEGUPB(fs) (1 << (fs)->lfs_sushift)
72 #define SEGTABSIZE_SU(fs) \
73 (((fs)->lfs_nseg + SEGUPB(fs) - 1) >> (fs)->lfs_sushift)
74
75 /* On-disk file information. One per file with data blocks in the segment. */
76 typedef struct finfo FINFO;
77 struct finfo {
78 u_int32_t fi_nblocks; /* number of blocks */
79 u_int32_t fi_version; /* version number */
80 u_int32_t fi_ino; /* inode number */
81 u_int32_t fi_lastlength; /* length of last block in array */
82 ufs_daddr_t fi_blocks[1]; /* array of logical block numbers */
83 };
84
85 /* On-disk super block. */
86 struct dlfs {
87 #define LFS_MAGIC 0x070162
88 u_int32_t dlfs_magic; /* 0: magic number */
89 #define LFS_VERSION 1
90 u_int32_t dlfs_version; /* 4: version number */
91
92 u_int32_t dlfs_size; /* 8: number of blocks in fs */
93 u_int32_t dlfs_ssize; /* 12: number of blocks per segment */
94 u_int32_t dlfs_dsize; /* 16: number of disk blocks in fs */
95 u_int32_t dlfs_bsize; /* 20: file system block size */
96 u_int32_t dlfs_fsize; /* 24: size of frag blocks in fs */
97 u_int32_t dlfs_frag; /* 28: number of frags in a block in fs */
98
99 /* Checkpoint region. */
100 ino_t dlfs_free; /* 32: start of the free list */
101 u_int32_t dlfs_bfree; /* 36: number of free disk blocks */
102 u_int32_t dlfs_nfiles; /* 40: number of allocated inodes */
103 int32_t dlfs_avail; /* 44: blocks available for writing */
104 u_int32_t dlfs_uinodes; /* 48: inodes in cache not yet on disk */
105 ufs_daddr_t dlfs_idaddr; /* 52: inode file disk address */
106 ino_t dlfs_ifile; /* 56: inode file inode number */
107 ufs_daddr_t dlfs_lastseg; /* 60: address of last segment written */
108 ufs_daddr_t dlfs_nextseg; /* 64: address of next segment to write */
109 ufs_daddr_t dlfs_curseg; /* 68: current segment being written */
110 ufs_daddr_t dlfs_offset; /* 72: offset in curseg for next partial */
111 ufs_daddr_t dlfs_lastpseg; /* 76: address of last partial written */
112 u_int32_t dlfs_tstamp; /* 80: time stamp */
113
114 /* These are configuration parameters. */
115 u_int32_t dlfs_minfree; /* 84: minimum percentage of free blocks */
116
117 /* These fields can be computed from the others. */
118 u_int64_t dlfs_maxfilesize; /* 88: maximum representable file size */
119 u_int32_t dlfs_dbpseg; /* 96: disk blocks per segment */
120 u_int32_t dlfs_inopb; /* 100: inodes per block */
121 u_int32_t dlfs_ifpb; /* 104: IFILE entries per block */
122 u_int32_t dlfs_sepb; /* 108: SEGUSE entries per block */
123 u_int32_t dlfs_nindir; /* 112: indirect pointers per block */
124 u_int32_t dlfs_nseg; /* 116: number of segments */
125 u_int32_t dlfs_nspf; /* 120: number of sectors per fragment */
126 u_int32_t dlfs_cleansz; /* 124: cleaner info size in blocks */
127 u_int32_t dlfs_segtabsz; /* 128: segment table size in blocks */
128 u_int32_t dlfs_segmask; /* 132: calculate offset within a segment */
129 u_int32_t dlfs_segshift; /* 136: fast mult/div for segments */
130 u_int32_t dlfs_bshift; /* 140: calc block number from file offset */
131 u_int32_t dlfs_ffshift; /* 144: fast mult/div for frag from file */
132 u_int32_t dlfs_fbshift; /* 148: fast mult/div for frag from block */
133 u_int64_t dlfs_bmask; /* 152: calc block offset from file offset */
134 u_int64_t dlfs_ffmask; /* 160: calc frag offset from file offset */
135 u_int64_t dlfs_fbmask; /* 168: calc frag offset from block offset */
136 u_int32_t dlfs_fsbtodb; /* 176: fsbtodb and dbtofsb shift constant */
137 u_int32_t dlfs_sushift; /* 180: fast mult/div for segusage table */
138
139 int32_t dlfs_maxsymlinklen; /* 184: max length of an internal symlink */
140
141 #define LFS_MIN_SBINTERVAL 5 /* minimum superblock segment spacing */
142 #define LFS_MAXNUMSB 10 /* 188: superblock disk offsets */
143 ufs_daddr_t dlfs_sboffs[LFS_MAXNUMSB];
144
145 /* Checksum -- last valid disk field. */
146 u_int32_t dlfs_cksum; /* 228: checksum for superblock checking */
147 int32_t dlfs_pad[70]; /* 232: round to 512 bytes */
148 };
149
150 /* In-memory super block. */
151 struct lfs {
152 struct dlfs lfs_dlfs; /* on-disk parameters */
153 #define lfs_magic lfs_dlfs.dlfs_magic
154 #define lfs_version lfs_dlfs.dlfs_version
155 #define lfs_size lfs_dlfs.dlfs_size
156 #define lfs_ssize lfs_dlfs.dlfs_ssize
157 #define lfs_dsize lfs_dlfs.dlfs_dsize
158 #define lfs_bsize lfs_dlfs.dlfs_bsize
159 #define lfs_fsize lfs_dlfs.dlfs_fsize
160 #define lfs_frag lfs_dlfs.dlfs_frag
161 #define lfs_free lfs_dlfs.dlfs_free
162 #define lfs_bfree lfs_dlfs.dlfs_bfree
163 #define lfs_nfiles lfs_dlfs.dlfs_nfiles
164 #define lfs_avail lfs_dlfs.dlfs_avail
165 #define lfs_uinodes lfs_dlfs.dlfs_uinodes
166 #define lfs_idaddr lfs_dlfs.dlfs_idaddr
167 #define lfs_ifile lfs_dlfs.dlfs_ifile
168 #define lfs_lastseg lfs_dlfs.dlfs_lastseg
169 #define lfs_nextseg lfs_dlfs.dlfs_nextseg
170 #define lfs_curseg lfs_dlfs.dlfs_curseg
171 #define lfs_offset lfs_dlfs.dlfs_offset
172 #define lfs_lastpseg lfs_dlfs.dlfs_lastpseg
173 #define lfs_tstamp lfs_dlfs.dlfs_tstamp
174 #define lfs_minfree lfs_dlfs.dlfs_minfree
175 #define lfs_maxfilesize lfs_dlfs.dlfs_maxfilesize
176 #define lfs_dbpseg lfs_dlfs.dlfs_dbpseg
177 #define lfs_inopb lfs_dlfs.dlfs_inopb
178 #define lfs_ifpb lfs_dlfs.dlfs_ifpb
179 #define lfs_sepb lfs_dlfs.dlfs_sepb
180 #define lfs_nindir lfs_dlfs.dlfs_nindir
181 #define lfs_nseg lfs_dlfs.dlfs_nseg
182 #define lfs_nspf lfs_dlfs.dlfs_nspf
183 #define lfs_cleansz lfs_dlfs.dlfs_cleansz
184 #define lfs_segtabsz lfs_dlfs.dlfs_segtabsz
185 #define lfs_segmask lfs_dlfs.dlfs_segmask
186 #define lfs_segshift lfs_dlfs.dlfs_segshift
187 #define lfs_bmask lfs_dlfs.dlfs_bmask
188 #define lfs_bshift lfs_dlfs.dlfs_bshift
189 #define lfs_ffmask lfs_dlfs.dlfs_ffmask
190 #define lfs_ffshift lfs_dlfs.dlfs_ffshift
191 #define lfs_fbmask lfs_dlfs.dlfs_fbmask
192 #define lfs_fbshift lfs_dlfs.dlfs_fbshift
193 #define lfs_fsbtodb lfs_dlfs.dlfs_fsbtodb
194 #define lfs_sushift lfs_dlfs.dlfs_sushift
195 #define lfs_maxsymlinklen lfs_dlfs.dlfs_maxsymlinklen
196 #define lfs_sboffs lfs_dlfs.dlfs_sboffs
197 #define lfs_cksum lfs_dlfs.dlfs_cksum
198 /* These fields are set at mount time and are meaningless on disk. */
199 struct segment *lfs_sp; /* current segment being written */
200 struct vnode *lfs_ivnode; /* vnode for the ifile */
201 u_int32_t lfs_seglock; /* single-thread the segment writer */
202 pid_t lfs_lockpid; /* pid of lock holder */
203 u_int32_t lfs_iocount; /* number of ios pending */
204 u_int32_t lfs_writer; /* don't allow any dirops to start */
205 u_int32_t lfs_dirops; /* count of active directory ops */
206 u_int32_t lfs_doifile; /* Write ifile blocks on next write */
207 u_int32_t lfs_nactive; /* Number of segments since last ckp */
208 int8_t lfs_fmod; /* super block modified flag */
209 int8_t lfs_clean; /* file system is clean flag */
210 int8_t lfs_ronly; /* mounted read-only flag */
211 int8_t lfs_flags; /* currently unused flag */
212 u_char lfs_fsmnt[MNAMELEN]; /* name mounted on */
213 };
214
215 /*
216 * Inode 0: out-of-band inode number
217 * Inode 1: IFILE inode number
218 * Inode 2: root inode
219 * Inode 3: lost+found inode number
220 */
221 #define LFS_UNUSED_INUM 0 /* out of band inode number */
222 #define LFS_IFILE_INUM 1 /* IFILE inode number */
223 #define LOSTFOUNDINO 3 /* lost+found inode number */
224 #define LFS_FIRST_INUM 4 /* first free inode number */
225
226 /* Address calculations for metadata located in the inode */
227 #define S_INDIR(fs) -NDADDR
228 #define D_INDIR(fs) (S_INDIR(fs) - NINDIR(fs) - 1)
229 #define T_INDIR(fs) (D_INDIR(fs) - NINDIR(fs) * NINDIR(fs) - 1)
230
231 /* Unassigned disk address. */
232 #define UNASSIGNED -1
233
234 /* Unused logical block number */
235 #define LFS_UNUSED_LBN -1
236
237 typedef struct ifile IFILE;
238 struct ifile {
239 u_int32_t if_version; /* inode version number */
240 #define LFS_UNUSED_DADDR 0 /* out-of-band daddr */
241 ufs_daddr_t if_daddr; /* inode disk address */
242 ino_t if_nextfree; /* next-unallocated inode */
243 };
244
245 /*
246 * Cleaner information structure. This resides in the ifile and is used
247 * to pass information between the cleaner and the kernel.
248 */
249 typedef struct _cleanerinfo {
250 u_int32_t clean; /* K: number of clean segments */
251 u_int32_t dirty; /* K: number of dirty segments */
252 } CLEANERINFO;
253
254 #define CLEANSIZE_SU(fs) \
255 ((sizeof(CLEANERINFO) + (fs)->lfs_bsize - 1) >> (fs)->lfs_bshift)
256
257 /*
258 * All summary blocks are the same size, so we can always read a summary
259 * block easily from a segment.
260 */
261 #define LFS_SUMMARY_SIZE 512
262
263 /* On-disk segment summary information */
264 typedef struct segsum SEGSUM;
265 struct segsum {
266 u_int32_t ss_sumsum; /* check sum of summary block */
267 u_int32_t ss_datasum; /* check sum of data */
268 u_int32_t ss_magic; /* segment summary magic number */
269 #define SS_MAGIC 0x061561
270 ufs_daddr_t ss_next; /* next segment */
271 u_int32_t ss_create; /* creation time stamp */
272 u_int16_t ss_nfinfo; /* number of file info structures */
273 u_int16_t ss_ninos; /* number of inodes in summary */
274
275 #define SS_DIROP 0x01 /* segment begins a dirop */
276 #define SS_CONT 0x02 /* more partials to finish this write*/
277 u_int16_t ss_flags; /* used for directory operations */
278 u_int16_t ss_pad; /* extra space */
279 /* FINFO's and inode daddr's... */
280 };
281
282 /* NINDIR is the number of indirects in a file system block. */
283 #define NINDIR(fs) ((fs)->lfs_nindir)
284
285 /* INOPB is the number of inodes in a secondary storage block. */
286 #define INOPB(fs) ((fs)->lfs_inopb)
287
288 #define blksize(fs, ip, lbn) \
289 (((lbn) >= NDADDR || (ip)->i_ffs_size >= ((lbn) + 1) << (fs)->lfs_bshift) \
290 ? (fs)->lfs_bsize \
291 : (fragroundup(fs, blkoff(fs, (ip)->i_ffs_size))))
292 #define blkoff(fs, loc) ((int)(loc) & (fs)->lfs_bmask)
293 #define fragoff(fs, loc) /* calculates (loc % fs->lfs_fsize) */ \
294 ((int)((loc) & (fs)->lfs_ffmask))
295 #define fsbtodb(fs, b) ((b) << (fs)->lfs_fsbtodb)
296 #define dbtofsb(fs, b) ((b) >> (fs)->lfs_fsbtodb)
297 #define fragstodb(fs, b) ((b) << ((fs)->lfs_fsbtodb - (fs)->lfs_fbshift))
298 #define dbtofrags(fs, b) ((b) >> ((fs)->lfs_fsbtodb - (fs)->lfs_fbshift))
299 #define lblkno(fs, loc) ((loc) >> (fs)->lfs_bshift)
300 #define lblktosize(fs, blk) ((blk) << (fs)->lfs_bshift)
301 #define numfrags(fs, loc) /* calculates (loc / fs->lfs_fsize) */ \
302 ((loc) >> (fs)->lfs_ffshift)
303 #define blkroundup(fs, size) /* calculates roundup(size, fs->lfs_bsize) */ \
304 ((int)(((size) + (fs)->lfs_bmask) & (~(fs)->lfs_bmask)))
305 #define fragroundup(fs, size) /* calculates roundup(size, fs->lfs_fsize) */ \
306 ((int)(((size) + (fs)->lfs_ffmask) & (~(fs)->lfs_ffmask)))
307 #define fragstoblks(fs, frags) /* calculates (frags / fs->lfs_frag) */ \
308 ((frags) >> (fs)->lfs_fbshift)
309 #define blkstofrags(fs, blks) /* calculates (blks * fs->lfs_frag) */ \
310 ((blks) << (fs)->lfs_fbshift)
311 #define fragnum(fs, fsb) /* calculates (fsb % fs->lfs_frag) */ \
312 ((fsb) & ((fs)->lfs_frag - 1))
313 #define blknum(fs, fsb) /* calculates rounddown(fsb, fs->lfs_frag) */ \
314 ((fsb) &~ ((fs)->lfs_frag - 1))
315 #define dblksize(fs, dip, lbn) \
316 (((lbn) >= NDADDR || (dip)->di_size >= ((lbn) + 1) << (fs)->lfs_bshift)\
317 ? (fs)->lfs_bsize \
318 : (fragroundup(fs, blkoff(fs, (dip)->di_size))))
319 #define datosn(fs, daddr) /* disk address to segment number */ \
320 (((daddr) - (fs)->lfs_sboffs[0]) / fsbtodb((fs), (fs)->lfs_ssize))
321 #define sntoda(fs, sn) /* segment number to disk address */ \
322 ((ufs_daddr_t)((sn) * ((fs)->lfs_ssize << (fs)->lfs_fsbtodb) + \
323 (fs)->lfs_sboffs[0]))
324
325 /* Read in the block with the cleaner info from the ifile. */
326 #define LFS_CLEANERINFO(CP, F, BP) { \
327 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \
328 if (bread((F)->lfs_ivnode, \
329 (ufs_daddr_t)0, (F)->lfs_bsize, NOCRED, &(BP))) \
330 panic("lfs: ifile read"); \
331 (CP) = (CLEANERINFO *)(BP)->b_data; \
332 }
333
334 /* Read in the block with a specific inode from the ifile. */
335 #define LFS_IENTRY(IP, F, IN, BP) { \
336 int _e; \
337 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \
338 if ((_e = bread((F)->lfs_ivnode, \
339 (IN) / (F)->lfs_ifpb + (F)->lfs_cleansz + (F)->lfs_segtabsz,\
340 (F)->lfs_bsize, NOCRED, &(BP))) != 0) \
341 panic("lfs: ifile read %d", _e); \
342 (IP) = (IFILE *)(BP)->b_data + (IN) % (F)->lfs_ifpb; \
343 }
344
345 /* Read in the block with a specific segment usage entry from the ifile. */
346 #define LFS_SEGENTRY(SP, F, IN, BP) { \
347 int _e; \
348 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \
349 if ((_e = bread((F)->lfs_ivnode, \
350 ((IN) >> (F)->lfs_sushift) + (F)->lfs_cleansz, \
351 (F)->lfs_bsize, NOCRED, &(BP))) != 0) \
352 panic("lfs: ifile read: %d", _e); \
353 (SP) = (SEGUSE *)(BP)->b_data + ((IN) & ((F)->lfs_sepb - 1)); \
354 }
355
356 /*
357 * Determine if there is enough room currently available to write db
358 * disk blocks. We need enough blocks for the new blocks, the current,
359 * inode blocks, a summary block, plus potentially the ifile inode and
360 * the segment usage table, plus an ifile page.
361 */
362 #define LFS_FITS(fs, db) \
363 ((int32_t)((db + ((fs)->lfs_uinodes + INOPB((fs))) / \
364 INOPB((fs)) + fsbtodb(fs, 1) + LFS_SUMMARY_SIZE / DEV_BSIZE + \
365 (fs)->lfs_segtabsz)) < (fs)->lfs_avail)
366
367 /* Determine if a buffer belongs to the ifile */
368 #define IS_IFILE(bp) (VTOI(bp->b_vp)->i_number == LFS_IFILE_INUM)
369
370 /*
371 * Structures used by lfs_bmapv and lfs_markv to communicate information
372 * about inodes and data blocks.
373 */
374 typedef struct block_info {
375 ino_t bi_inode; /* inode # */
376 ufs_daddr_t bi_lbn; /* logical block w/in file */
377 ufs_daddr_t bi_daddr; /* disk address of block */
378 time_t bi_segcreate; /* origin segment create time */
379 int bi_version; /* file version number */
380 void *bi_bp; /* data buffer */
381 int bi_size; /* size of the block (if fragment) */
382 } BLOCK_INFO;
383
384 /* In-memory description of a segment about to be written. */
385 struct segment {
386 struct lfs *fs; /* file system pointer */
387 struct buf **bpp; /* pointer to buffer array */
388 struct buf **cbpp; /* pointer to next available bp */
389 struct buf **start_bpp; /* pointer to first bp in this set */
390 struct buf *ibp; /* buffer pointer to inode page */
391 struct finfo *fip; /* current fileinfo pointer */
392 struct vnode *vp; /* vnode being gathered */
393 void *segsum; /* segment summary info */
394 u_int32_t ninodes; /* number of inodes in this segment */
395 u_int32_t seg_bytes_left; /* bytes left in segment */
396 u_int32_t sum_bytes_left; /* bytes left in summary block */
397 u_int32_t seg_number; /* number of this segment */
398 ufs_daddr_t *start_lbp; /* beginning lbn for this set */
399
400 #define SEGM_CKP 0x01 /* doing a checkpoint */
401 #define SEGM_CLEAN 0x02 /* cleaner call; don't sort */
402 #define SEGM_SYNC 0x04 /* wait for segment */
403 u_int16_t seg_flags; /* run-time flags for this segment */
404 };
405
406 #define ISSPACE(F, BB, C) \
407 (((C)->cr_uid == 0 && (F)->lfs_bfree >= (BB)) || \
408 ((C)->cr_uid != 0 && IS_FREESPACE(F, BB)))
409
410 #define IS_FREESPACE(F, BB) \
411 ((F)->lfs_bfree > ((F)->lfs_dsize * (F)->lfs_minfree / 100 + (BB)))
412
413 #define ISSPACE_XXX(F, BB) \
414 ((F)->lfs_bfree >= (BB))
415
416 #define DOSTATS
417 #ifdef DOSTATS
418 /* Statistics Counters */
419 struct lfs_stats {
420 u_int segsused;
421 u_int psegwrites;
422 u_int psyncwrites;
423 u_int pcleanwrites;
424 u_int blocktot;
425 u_int cleanblocks;
426 u_int ncheckpoints;
427 u_int nwrites;
428 u_int nsync_writes;
429 u_int wait_exceeded;
430 u_int write_exceeded;
431 u_int flush_invoked;
432 };
433 extern struct lfs_stats lfs_stats;
434 #endif
435