lfs.h revision 1.122
1 /* $NetBSD: lfs.h,v 1.122 2007/10/10 20:42:34 ad Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 2000, 2001, 2002, 2003 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Konrad E. Schroder <perseant (at) hhhh.org>. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. All advertising materials mentioning features or use of this software 19 * must display the following acknowledgement: 20 * This product includes software developed by the NetBSD 21 * Foundation, Inc. and its contributors. 22 * 4. Neither the name of The NetBSD Foundation nor the names of its 23 * contributors may be used to endorse or promote products derived 24 * from this software without specific prior written permission. 25 * 26 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 27 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 28 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 29 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 30 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 31 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 32 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 33 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 34 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 35 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 36 * POSSIBILITY OF SUCH DAMAGE. 37 */ 38 /*- 39 * Copyright (c) 1991, 1993 40 * The Regents of the University of California. All rights reserved. 41 * 42 * Redistribution and use in source and binary forms, with or without 43 * modification, are permitted provided that the following conditions 44 * are met: 45 * 1. Redistributions of source code must retain the above copyright 46 * notice, this list of conditions and the following disclaimer. 47 * 2. Redistributions in binary form must reproduce the above copyright 48 * notice, this list of conditions and the following disclaimer in the 49 * documentation and/or other materials provided with the distribution. 50 * 3. Neither the name of the University nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 64 * SUCH DAMAGE. 65 * 66 * @(#)lfs.h 8.9 (Berkeley) 5/8/95 67 */ 68 69 #ifndef _UFS_LFS_LFS_H_ 70 #define _UFS_LFS_LFS_H_ 71 72 #include <sys/rwlock.h> 73 74 /* 75 * Compile-time options for LFS. 76 */ 77 #define LFS_IFIND_RETRIES 16 78 #define LFS_LOGLENGTH 1024 /* size of debugging log */ 79 #define LFS_MAX_ACTIVE 10 /* Dirty segments before ckp forced */ 80 81 /* 82 * Fixed filesystem layout parameters 83 */ 84 #define LFS_LABELPAD 8192 /* LFS label size */ 85 #define LFS_SBPAD 8192 /* LFS superblock size */ 86 87 #define LFS_UNUSED_INUM 0 /* 0: out of band inode number */ 88 #define LFS_IFILE_INUM 1 /* 1: IFILE inode number */ 89 /* 2: Root inode number */ 90 #define LOSTFOUNDINO 3 /* 3: lost+found inode number */ 91 #define LFS_FIRST_INUM 4 /* 4: first free inode number */ 92 93 #define LFS_V1_SUMMARY_SIZE 512 /* V1 fixed summary size */ 94 #define LFS_DFL_SUMMARY_SIZE 512 /* Default summary size */ 95 96 #define LFS_MAX_DADDR 0x7fffffff /* Highest addressable fsb */ 97 98 #define LFS_MAXNAMLEN 255 /* maximum name length in a dir */ 99 100 /* Adjustable filesystem parameters */ 101 #define MIN_FREE_SEGS 20 102 #define MIN_RESV_SEGS 15 103 #ifndef LFS_ATIME_IFILE 104 # define LFS_ATIME_IFILE 0 /* Store atime info in ifile (optional in LFSv1) */ 105 #endif 106 #define LFS_MARKV_MAXBLKCNT 65536 /* Max block count for lfs_markv() */ 107 108 /* Misc. definitions */ 109 #define BW_CLEAN 1 /* Flag for lfs_bwrite_ext() */ 110 #define PG_DELWRI PG_PAGER1 /* Local def for delayed pageout */ 111 112 /* Resource limits */ 113 #define LFS_MAX_RESOURCE(x, u) (((x) >> 2) - 10 * (u)) 114 #define LFS_WAIT_RESOURCE(x, u) (((x) >> 1) - ((x) >> 3) - 10 * (u)) 115 #define LFS_INVERSE_MAX_RESOURCE(x, u) (((x) + 10 * (u)) << 2) 116 #define LFS_MAX_BUFS LFS_MAX_RESOURCE(nbuf, 1) 117 #define LFS_WAIT_BUFS LFS_WAIT_RESOURCE(nbuf, 1) 118 #define LFS_INVERSE_MAX_BUFS(n) LFS_INVERSE_MAX_RESOURCE(n, 1) 119 #define LFS_MAX_BYTES LFS_MAX_RESOURCE(bufmem_lowater, PAGE_SIZE) 120 #define LFS_INVERSE_MAX_BYTES(n) LFS_INVERSE_MAX_RESOURCE(n, PAGE_SIZE) 121 #define LFS_WAIT_BYTES LFS_WAIT_RESOURCE(bufmem_lowater, PAGE_SIZE) 122 #define LFS_MAX_DIROP ((desiredvnodes >> 2) + (desiredvnodes >> 3)) 123 #define SIZEOF_DIROP(fs) (2 * ((fs)->lfs_bsize + DINODE1_SIZE)) 124 #define LFS_MAX_FSDIROP(fs) \ 125 ((fs)->lfs_nclean <= (fs)->lfs_resvseg ? 0 : \ 126 (((fs)->lfs_nclean - (fs)->lfs_resvseg) * (fs)->lfs_ssize) / \ 127 (2 * SIZEOF_DIROP(fs))) 128 #define LFS_MAX_PAGES lfs_max_pages() 129 #define LFS_WAIT_PAGES lfs_wait_pages() 130 #define LFS_BUFWAIT 2 /* How long to wait if over *_WAIT_* */ 131 132 #ifdef _KERNEL 133 int lfs_wait_pages(void); 134 int lfs_max_pages(void); 135 #endif /* _KERNEL */ 136 137 /* How starved can we be before we start holding back page writes */ 138 #define LFS_STARVED_FOR_SEGS(fs) ((fs)->lfs_nclean < (fs)->lfs_resvseg) 139 140 /* 141 * Reserved blocks for lfs_malloc 142 */ 143 144 /* Structure to keep reserved blocks */ 145 typedef struct lfs_res_blk { 146 void *p; 147 LIST_ENTRY(lfs_res_blk) res; 148 int size; 149 char inuse; 150 } res_t; 151 152 /* Types for lfs_newbuf and lfs_malloc */ 153 #define LFS_NB_UNKNOWN -1 154 #define LFS_NB_SUMMARY 0 155 #define LFS_NB_SBLOCK 1 156 #define LFS_NB_IBLOCK 2 157 #define LFS_NB_CLUSTER 3 158 #define LFS_NB_CLEAN 4 159 #define LFS_NB_BLKIOV 5 160 #define LFS_NB_COUNT 6 /* always last */ 161 162 /* Number of reserved memory blocks of each type */ 163 #define LFS_N_SUMMARIES 2 164 #define LFS_N_SBLOCKS 1 /* Always 1, to throttle superblock writes */ 165 #define LFS_N_IBLOCKS 16 /* In theory ssize/bsize; in practice around 2 */ 166 #define LFS_N_CLUSTERS 16 /* In theory ssize/MAXPHYS */ 167 #define LFS_N_CLEAN 0 168 #define LFS_N_BLKIOV 1 169 170 /* Total count of "large" (non-pool) types */ 171 #define LFS_N_TOTAL (LFS_N_SUMMARIES + LFS_N_SBLOCKS + LFS_N_IBLOCKS + \ 172 LFS_N_CLUSTERS + LFS_N_CLEAN + LFS_N_BLKIOV) 173 174 /* Counts for pool types */ 175 #define LFS_N_CL LFS_N_CLUSTERS 176 #define LFS_N_BPP 2 177 #define LFS_N_SEG 2 178 179 /* 180 * "struct buf" associated definitions 181 */ 182 183 /* Unassigned disk addresses. */ 184 #define UNASSIGNED -1 185 #define UNWRITTEN -2 186 187 /* Unused logical block number */ 188 #define LFS_UNUSED_LBN -1 189 190 /* Determine if a buffer belongs to the ifile */ 191 #define IS_IFILE(bp) (VTOI(bp->b_vp)->i_number == LFS_IFILE_INUM) 192 193 # define LFS_LOCK_BUF(bp) do { \ 194 if (((bp)->b_flags & (B_LOCKED | B_CALL)) == 0) { \ 195 simple_lock(&lfs_subsys_lock); \ 196 ++locked_queue_count; \ 197 locked_queue_bytes += bp->b_bufsize; \ 198 simple_unlock(&lfs_subsys_lock); \ 199 } \ 200 (bp)->b_flags |= B_LOCKED; \ 201 } while (0) 202 203 # define LFS_UNLOCK_BUF(bp) do { \ 204 if (((bp)->b_flags & (B_LOCKED | B_CALL)) == B_LOCKED) { \ 205 simple_lock(&lfs_subsys_lock); \ 206 --locked_queue_count; \ 207 locked_queue_bytes -= bp->b_bufsize; \ 208 if (locked_queue_count < LFS_WAIT_BUFS && \ 209 locked_queue_bytes < LFS_WAIT_BYTES) \ 210 wakeup(&locked_queue_count); \ 211 simple_unlock(&lfs_subsys_lock); \ 212 } \ 213 (bp)->b_flags &= ~B_LOCKED; \ 214 } while (0) 215 216 #ifdef _KERNEL 217 218 extern u_long bufmem_lowater, bufmem_hiwater; /* XXX */ 219 220 # define LFS_IS_MALLOC_BUF(bp) (((bp)->b_flags & B_CALL) && \ 221 (bp)->b_iodone == lfs_callback) 222 223 # ifdef DEBUG 224 # define LFS_DEBUG_COUNTLOCKED(m) do { \ 225 if (lfs_debug_log_subsys[DLOG_LLIST]) { \ 226 lfs_countlocked(&locked_queue_count, &locked_queue_bytes, (m)); \ 227 wakeup(&locked_queue_count); \ 228 } \ 229 } while (0) 230 # else 231 # define LFS_DEBUG_COUNTLOCKED(m) 232 # endif 233 234 /* log for debugging writes to the Ifile */ 235 # ifdef DEBUG 236 struct lfs_log_entry { 237 const char *op; 238 const char *file; 239 int pid; 240 int line; 241 daddr_t block; 242 unsigned long flags; 243 }; 244 extern int lfs_lognum; 245 extern struct lfs_log_entry lfs_log[LFS_LOGLENGTH]; 246 # define LFS_BWRITE_LOG(bp) lfs_bwrite_log((bp), __FILE__, __LINE__) 247 # define LFS_ENTER_LOG(theop, thefile, theline, lbn, theflags, thepid) do {\ 248 int _s; \ 249 \ 250 simple_lock(&lfs_subsys_lock); \ 251 _s = splbio(); \ 252 lfs_log[lfs_lognum].op = theop; \ 253 lfs_log[lfs_lognum].file = thefile; \ 254 lfs_log[lfs_lognum].line = (theline); \ 255 lfs_log[lfs_lognum].pid = (thepid); \ 256 lfs_log[lfs_lognum].block = (lbn); \ 257 lfs_log[lfs_lognum].flags = (theflags); \ 258 lfs_lognum = (lfs_lognum + 1) % LFS_LOGLENGTH; \ 259 splx(_s); \ 260 simple_unlock(&lfs_subsys_lock); \ 261 } while (0) 262 263 # define LFS_BCLEAN_LOG(fs, bp) do { \ 264 if ((bp)->b_vp == (fs)->lfs_ivnode) \ 265 LFS_ENTER_LOG("clear", __FILE__, __LINE__, \ 266 bp->b_lblkno, bp->b_flags, curproc->p_pid);\ 267 } while (0) 268 269 /* Must match list in lfs_vfsops.c ! */ 270 # define DLOG_RF 0 /* roll forward */ 271 # define DLOG_ALLOC 1 /* inode alloc */ 272 # define DLOG_AVAIL 2 /* lfs_{,r,f}avail */ 273 # define DLOG_FLUSH 3 /* flush */ 274 # define DLOG_LLIST 4 /* locked list accounting */ 275 # define DLOG_WVNODE 5 /* vflush/writevnodes verbose */ 276 # define DLOG_VNODE 6 /* vflush/writevnodes */ 277 # define DLOG_SEG 7 /* segwrite */ 278 # define DLOG_SU 8 /* seguse accounting */ 279 # define DLOG_CLEAN 9 /* cleaner routines */ 280 # define DLOG_MOUNT 10 /* mount/unmount */ 281 # define DLOG_PAGE 11 /* putpages/gop_write */ 282 # define DLOG_DIROP 12 /* dirop accounting */ 283 # define DLOG_MALLOC 13 /* lfs_malloc accounting */ 284 # define DLOG_MAX 14 /* The terminator */ 285 # define DLOG(a) lfs_debug_log a 286 # else /* ! DEBUG */ 287 # define LFS_BCLEAN_LOG(fs, bp) 288 # define LFS_BWRITE_LOG(bp) VOP_BWRITE((bp)) 289 # define DLOG(a) 290 # endif /* ! DEBUG */ 291 #else /* ! _KERNEL */ 292 # define LFS_BWRITE_LOG(bp) VOP_BWRITE((bp)) 293 #endif /* _KERNEL */ 294 295 #ifdef _KERNEL 296 /* Filehandle structure for exported LFSes */ 297 struct lfid { 298 struct ufid lfid_ufid; 299 #define lfid_len lfid_ufid.ufid_len 300 #define lfid_ino lfid_ufid.ufid_ino 301 #define lfid_gen lfid_ufid.ufid_gen 302 uint32_t lfid_ident; 303 }; 304 #endif /* _KERNEL */ 305 306 /* 307 * "struct inode" associated definitions 308 */ 309 310 /* Address calculations for metadata located in the inode */ 311 #define S_INDIR(fs) -NDADDR 312 #define D_INDIR(fs) (S_INDIR(fs) - NINDIR(fs) - 1) 313 #define T_INDIR(fs) (D_INDIR(fs) - NINDIR(fs) * NINDIR(fs) - 1) 314 315 /* For convenience */ 316 #define IN_ALLMOD (IN_MODIFIED|IN_ACCESS|IN_CHANGE|IN_UPDATE|IN_MODIFY|IN_ACCESSED|IN_CLEANING) 317 318 #define LFS_SET_UINO(ip, flags) do { \ 319 simple_lock(&(ip)->i_lfs->lfs_interlock); \ 320 if (((flags) & IN_ACCESSED) && !((ip)->i_flag & IN_ACCESSED)) \ 321 ++(ip)->i_lfs->lfs_uinodes; \ 322 if (((flags) & IN_CLEANING) && !((ip)->i_flag & IN_CLEANING)) \ 323 ++(ip)->i_lfs->lfs_uinodes; \ 324 if (((flags) & IN_MODIFIED) && !((ip)->i_flag & IN_MODIFIED)) \ 325 ++(ip)->i_lfs->lfs_uinodes; \ 326 (ip)->i_flag |= (flags); \ 327 simple_unlock(&(ip)->i_lfs->lfs_interlock); \ 328 } while (0) 329 330 #define LFS_CLR_UINO(ip, flags) do { \ 331 simple_lock(&(ip)->i_lfs->lfs_interlock); \ 332 if (((flags) & IN_ACCESSED) && ((ip)->i_flag & IN_ACCESSED)) \ 333 --(ip)->i_lfs->lfs_uinodes; \ 334 if (((flags) & IN_CLEANING) && ((ip)->i_flag & IN_CLEANING)) \ 335 --(ip)->i_lfs->lfs_uinodes; \ 336 if (((flags) & IN_MODIFIED) && ((ip)->i_flag & IN_MODIFIED)) \ 337 --(ip)->i_lfs->lfs_uinodes; \ 338 (ip)->i_flag &= ~(flags); \ 339 if ((ip)->i_lfs->lfs_uinodes < 0) { \ 340 panic("lfs_uinodes < 0"); \ 341 } \ 342 simple_unlock(&(ip)->i_lfs->lfs_interlock); \ 343 } while (0) 344 345 #define LFS_ITIMES(ip, acc, mod, cre) \ 346 while ((ip)->i_flag & (IN_ACCESS | IN_CHANGE | IN_UPDATE | IN_MODIFY)) \ 347 lfs_itimes(ip, acc, mod, cre) 348 349 /* 350 * "struct vnode" associated definitions 351 */ 352 353 /* Heuristic emptiness measure */ 354 #define VPISEMPTY(vp) (LIST_EMPTY(&(vp)->v_dirtyblkhd) && \ 355 !(vp->v_type == VREG && (vp)->v_iflag & VI_ONWORKLST) &&\ 356 VTOI(vp)->i_lfs_nbtree == 0) 357 358 #define WRITEINPROG(vp) ((vp)->v_numoutput > 0 || \ 359 (!LIST_EMPTY(&(vp)->v_dirtyblkhd) && \ 360 !(VTOI(vp)->i_flag & (IN_MODIFIED | IN_ACCESSED | IN_CLEANING)))) 361 362 363 /* 364 * On-disk and in-memory checkpoint segment usage structure. 365 */ 366 typedef struct segusage SEGUSE; 367 struct segusage { 368 u_int32_t su_nbytes; /* 0: number of live bytes */ 369 u_int32_t su_olastmod; /* 4: SEGUSE last modified timestamp */ 370 u_int16_t su_nsums; /* 8: number of summaries in segment */ 371 u_int16_t su_ninos; /* 10: number of inode blocks in seg */ 372 373 #define SEGUSE_ACTIVE 0x01 /* segment currently being written */ 374 #define SEGUSE_DIRTY 0x02 /* segment has data in it */ 375 #define SEGUSE_SUPERBLOCK 0x04 /* segment contains a superblock */ 376 #define SEGUSE_ERROR 0x08 /* cleaner: do not clean segment */ 377 #define SEGUSE_EMPTY 0x10 /* segment is empty */ 378 #define SEGUSE_INVAL 0x20 /* segment is invalid */ 379 u_int32_t su_flags; /* 12: segment flags */ 380 u_int64_t su_lastmod; /* 16: last modified timestamp */ 381 }; 382 383 typedef struct segusage_v1 SEGUSE_V1; 384 struct segusage_v1 { 385 u_int32_t su_nbytes; /* 0: number of live bytes */ 386 u_int32_t su_lastmod; /* 4: SEGUSE last modified timestamp */ 387 u_int16_t su_nsums; /* 8: number of summaries in segment */ 388 u_int16_t su_ninos; /* 10: number of inode blocks in seg */ 389 u_int32_t su_flags; /* 12: segment flags */ 390 }; 391 392 #define SEGUPB(fs) (fs->lfs_sepb) 393 #define SEGTABSIZE_SU(fs) \ 394 (((fs)->lfs_nseg + SEGUPB(fs) - 1) / (fs)->lfs_sepb) 395 396 #ifdef _KERNEL 397 # define SHARE_IFLOCK(F) \ 398 do { \ 399 simple_lock(&(F)->lfs_interlock); \ 400 lockmgr(&(F)->lfs_iflock, LK_SHARED, &(F)->lfs_interlock); \ 401 simple_unlock(&(F)->lfs_interlock); \ 402 } while(0) 403 # define UNSHARE_IFLOCK(F) \ 404 do { \ 405 simple_lock(&(F)->lfs_interlock); \ 406 lockmgr(&(F)->lfs_iflock, LK_RELEASE, &(F)->lfs_interlock); \ 407 simple_unlock(&(F)->lfs_interlock); \ 408 } while(0) 409 #else /* ! _KERNEL */ 410 # define SHARE_IFLOCK(F) 411 # define UNSHARE_IFLOCK(F) 412 #endif /* ! _KERNEL */ 413 414 /* Read in the block with a specific segment usage entry from the ifile. */ 415 #define LFS_SEGENTRY(SP, F, IN, BP) do { \ 416 int _e; \ 417 SHARE_IFLOCK(F); \ 418 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \ 419 if ((_e = bread((F)->lfs_ivnode, \ 420 ((IN) / (F)->lfs_sepb) + (F)->lfs_cleansz, \ 421 (F)->lfs_bsize, NOCRED, &(BP))) != 0) \ 422 panic("lfs: ifile read: %d", _e); \ 423 if ((F)->lfs_version == 1) \ 424 (SP) = (SEGUSE *)((SEGUSE_V1 *)(BP)->b_data + \ 425 ((IN) & ((F)->lfs_sepb - 1))); \ 426 else \ 427 (SP) = (SEGUSE *)(BP)->b_data + ((IN) % (F)->lfs_sepb); \ 428 UNSHARE_IFLOCK(F); \ 429 } while (0) 430 431 #define LFS_WRITESEGENTRY(SP, F, IN, BP) do { \ 432 if ((SP)->su_nbytes == 0) \ 433 (SP)->su_flags |= SEGUSE_EMPTY; \ 434 else \ 435 (SP)->su_flags &= ~SEGUSE_EMPTY; \ 436 (F)->lfs_suflags[(F)->lfs_activesb][(IN)] = (SP)->su_flags; \ 437 LFS_BWRITE_LOG(BP); \ 438 } while (0) 439 440 /* 441 * On-disk file information. One per file with data blocks in the segment. 442 */ 443 typedef struct finfo FINFO; 444 struct finfo { 445 u_int32_t fi_nblocks; /* number of blocks */ 446 u_int32_t fi_version; /* version number */ 447 u_int32_t fi_ino; /* inode number */ 448 u_int32_t fi_lastlength; /* length of last block in array */ 449 int32_t fi_blocks[1]; /* array of logical block numbers */ 450 }; 451 /* sizeof FINFO except fi_blocks */ 452 #define FINFOSIZE (sizeof(FINFO) - sizeof(int32_t)) 453 454 /* 455 * Index file inode entries. 456 */ 457 typedef struct ifile IFILE; 458 struct ifile { 459 u_int32_t if_version; /* inode version number */ 460 #define LFS_UNUSED_DADDR 0 /* out-of-band daddr */ 461 int32_t if_daddr; /* inode disk address */ 462 #define LFS_ORPHAN_NEXTFREE (~(u_int32_t)0) /* indicate orphaned file */ 463 u_int32_t if_nextfree; /* next-unallocated inode */ 464 u_int32_t if_atime_sec; /* Last access time, seconds */ 465 u_int32_t if_atime_nsec; /* and nanoseconds */ 466 }; 467 468 typedef struct ifile_v1 IFILE_V1; 469 struct ifile_v1 { 470 u_int32_t if_version; /* inode version number */ 471 int32_t if_daddr; /* inode disk address */ 472 u_int32_t if_nextfree; /* next-unallocated inode */ 473 #if LFS_ATIME_IFILE 474 struct timespec if_atime; /* Last access time */ 475 #endif 476 }; 477 478 /* 479 * LFSv1 compatibility code is not allowed to touch if_atime, since it 480 * may not be mapped! 481 */ 482 /* Read in the block with a specific inode from the ifile. */ 483 #define LFS_IENTRY(IP, F, IN, BP) do { \ 484 int _e; \ 485 SHARE_IFLOCK(F); \ 486 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \ 487 if ((_e = bread((F)->lfs_ivnode, \ 488 (IN) / (F)->lfs_ifpb + (F)->lfs_cleansz + (F)->lfs_segtabsz, \ 489 (F)->lfs_bsize, NOCRED, &(BP))) != 0) \ 490 panic("lfs: ifile ino %d read %d", (int)(IN), _e); \ 491 if ((F)->lfs_version == 1) \ 492 (IP) = (IFILE *)((IFILE_V1 *)(BP)->b_data + \ 493 (IN) % (F)->lfs_ifpb); \ 494 else \ 495 (IP) = (IFILE *)(BP)->b_data + (IN) % (F)->lfs_ifpb; \ 496 UNSHARE_IFLOCK(F); \ 497 } while (0) 498 499 /* 500 * Cleaner information structure. This resides in the ifile and is used 501 * to pass information from the kernel to the cleaner. 502 */ 503 typedef struct _cleanerinfo { 504 u_int32_t clean; /* number of clean segments */ 505 u_int32_t dirty; /* number of dirty segments */ 506 int32_t bfree; /* disk blocks free */ 507 int32_t avail; /* disk blocks available */ 508 u_int32_t free_head; /* head of the inode free list */ 509 u_int32_t free_tail; /* tail of the inode free list */ 510 #define LFS_CLEANER_MUST_CLEAN 0x01 511 u_int32_t flags; /* status word from the kernel */ 512 } CLEANERINFO; 513 514 #define CLEANSIZE_SU(fs) \ 515 ((sizeof(CLEANERINFO) + (fs)->lfs_bsize - 1) >> (fs)->lfs_bshift) 516 517 /* Read in the block with the cleaner info from the ifile. */ 518 #define LFS_CLEANERINFO(CP, F, BP) do { \ 519 SHARE_IFLOCK(F); \ 520 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \ 521 if (bread((F)->lfs_ivnode, \ 522 (daddr_t)0, (F)->lfs_bsize, NOCRED, &(BP))) \ 523 panic("lfs: ifile read"); \ 524 (CP) = (CLEANERINFO *)(BP)->b_data; \ 525 UNSHARE_IFLOCK(F); \ 526 } while (0) 527 528 /* 529 * Synchronize the Ifile cleaner info with current avail and bfree. 530 */ 531 #define LFS_SYNC_CLEANERINFO(cip, fs, bp, w) do { \ 532 simple_lock(&(fs)->lfs_interlock); \ 533 if ((w) || (cip)->bfree != (fs)->lfs_bfree || \ 534 (cip)->avail != (fs)->lfs_avail - (fs)->lfs_ravail - \ 535 (fs)->lfs_favail) { \ 536 (cip)->bfree = (fs)->lfs_bfree; \ 537 (cip)->avail = (fs)->lfs_avail - (fs)->lfs_ravail - \ 538 (fs)->lfs_favail; \ 539 if (((bp)->b_flags & B_GATHERED) == 0) { \ 540 (fs)->lfs_flags |= LFS_IFDIRTY; \ 541 } \ 542 simple_unlock(&(fs)->lfs_interlock); \ 543 (void) LFS_BWRITE_LOG(bp); /* Ifile */ \ 544 } else { \ 545 simple_unlock(&(fs)->lfs_interlock); \ 546 brelse(bp, 0); \ 547 } \ 548 } while (0) 549 550 /* 551 * Get the head of the inode free list. 552 * Always caled with the segment lock held. 553 */ 554 #define LFS_GET_HEADFREE(FS, CIP, BP, FREEP) do { \ 555 if ((FS)->lfs_version > 1) { \ 556 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 557 (FS)->lfs_freehd = (CIP)->free_head; \ 558 brelse(BP, 0); \ 559 } \ 560 *(FREEP) = (FS)->lfs_freehd; \ 561 } while (0) 562 563 #define LFS_PUT_HEADFREE(FS, CIP, BP, VAL) do { \ 564 (FS)->lfs_freehd = (VAL); \ 565 if ((FS)->lfs_version > 1) { \ 566 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 567 (CIP)->free_head = (VAL); \ 568 LFS_BWRITE_LOG(BP); \ 569 simple_lock(&fs->lfs_interlock); \ 570 (FS)->lfs_flags |= LFS_IFDIRTY; \ 571 simple_unlock(&fs->lfs_interlock); \ 572 } \ 573 } while (0) 574 575 #define LFS_GET_TAILFREE(FS, CIP, BP, FREEP) do { \ 576 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 577 *(FREEP) = (CIP)->free_tail; \ 578 brelse(BP, 0); \ 579 } while (0) 580 581 #define LFS_PUT_TAILFREE(FS, CIP, BP, VAL) do { \ 582 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 583 (CIP)->free_tail = (VAL); \ 584 LFS_BWRITE_LOG(BP); \ 585 simple_lock(&fs->lfs_interlock); \ 586 (FS)->lfs_flags |= LFS_IFDIRTY; \ 587 simple_unlock(&fs->lfs_interlock); \ 588 } while (0) 589 590 /* 591 * On-disk segment summary information 592 */ 593 typedef struct segsum_v1 SEGSUM_V1; 594 struct segsum_v1 { 595 u_int32_t ss_sumsum; /* 0: check sum of summary block */ 596 u_int32_t ss_datasum; /* 4: check sum of data */ 597 u_int32_t ss_magic; /* 8: segment summary magic number */ 598 #define SS_MAGIC 0x061561 599 int32_t ss_next; /* 12: next segment */ 600 u_int32_t ss_create; /* 16: creation time stamp */ 601 u_int16_t ss_nfinfo; /* 20: number of file info structures */ 602 u_int16_t ss_ninos; /* 22: number of inodes in summary */ 603 604 #define SS_DIROP 0x01 /* segment begins a dirop */ 605 #define SS_CONT 0x02 /* more partials to finish this write*/ 606 #define SS_CLEAN 0x04 /* written by the cleaner */ 607 #define SS_RFW 0x08 /* written by the roll-forward agent */ 608 u_int16_t ss_flags; /* 24: used for directory operations */ 609 u_int16_t ss_pad; /* 26: extra space */ 610 /* FINFO's and inode daddr's... */ 611 }; 612 613 typedef struct segsum SEGSUM; 614 struct segsum { 615 u_int32_t ss_sumsum; /* 0: check sum of summary block */ 616 u_int32_t ss_datasum; /* 4: check sum of data */ 617 u_int32_t ss_magic; /* 8: segment summary magic number */ 618 int32_t ss_next; /* 12: next segment */ 619 u_int32_t ss_ident; /* 16: roll-forward fsid */ 620 #define ss_ocreate ss_ident /* ident is where create was in v1 */ 621 u_int16_t ss_nfinfo; /* 20: number of file info structures */ 622 u_int16_t ss_ninos; /* 22: number of inodes in summary */ 623 u_int16_t ss_flags; /* 24: used for directory operations */ 624 u_int8_t ss_pad[6]; /* 26: extra space */ 625 u_int64_t ss_serial; /* 32: serial number */ 626 u_int64_t ss_create; /* 40: time stamp */ 627 /* FINFO's and inode daddr's... */ 628 }; 629 630 #define SEGSUM_SIZE(fs) ((fs)->lfs_version == 1 ? sizeof(SEGSUM_V1) : sizeof(SEGSUM)) 631 632 633 /* 634 * On-disk super block. 635 */ 636 struct dlfs { 637 #define LFS_MAGIC 0x070162 638 u_int32_t dlfs_magic; /* 0: magic number */ 639 #define LFS_VERSION 2 640 u_int32_t dlfs_version; /* 4: version number */ 641 642 u_int32_t dlfs_size; /* 8: number of blocks in fs (v1) */ 643 /* number of frags in fs (v2) */ 644 u_int32_t dlfs_ssize; /* 12: number of blocks per segment (v1) */ 645 /* number of bytes per segment (v2) */ 646 u_int32_t dlfs_dsize; /* 16: number of disk blocks in fs */ 647 u_int32_t dlfs_bsize; /* 20: file system block size */ 648 u_int32_t dlfs_fsize; /* 24: size of frag blocks in fs */ 649 u_int32_t dlfs_frag; /* 28: number of frags in a block in fs */ 650 651 /* Checkpoint region. */ 652 u_int32_t dlfs_freehd; /* 32: start of the free list */ 653 int32_t dlfs_bfree; /* 36: number of free disk blocks */ 654 u_int32_t dlfs_nfiles; /* 40: number of allocated inodes */ 655 int32_t dlfs_avail; /* 44: blocks available for writing */ 656 int32_t dlfs_uinodes; /* 48: inodes in cache not yet on disk */ 657 int32_t dlfs_idaddr; /* 52: inode file disk address */ 658 u_int32_t dlfs_ifile; /* 56: inode file inode number */ 659 int32_t dlfs_lastseg; /* 60: address of last segment written */ 660 int32_t dlfs_nextseg; /* 64: address of next segment to write */ 661 int32_t dlfs_curseg; /* 68: current segment being written */ 662 int32_t dlfs_offset; /* 72: offset in curseg for next partial */ 663 int32_t dlfs_lastpseg; /* 76: address of last partial written */ 664 u_int32_t dlfs_inopf; /* 80: v1: time stamp; v2: inodes per frag */ 665 #define dlfs_otstamp dlfs_inopf 666 667 /* These are configuration parameters. */ 668 u_int32_t dlfs_minfree; /* 84: minimum percentage of free blocks */ 669 670 /* These fields can be computed from the others. */ 671 u_int64_t dlfs_maxfilesize; /* 88: maximum representable file size */ 672 u_int32_t dlfs_fsbpseg; /* 96: fsb per segment */ 673 u_int32_t dlfs_inopb; /* 100: inodes per block */ 674 u_int32_t dlfs_ifpb; /* 104: IFILE entries per block */ 675 u_int32_t dlfs_sepb; /* 108: SEGUSE entries per block */ 676 u_int32_t dlfs_nindir; /* 112: indirect pointers per block */ 677 u_int32_t dlfs_nseg; /* 116: number of segments */ 678 u_int32_t dlfs_nspf; /* 120: number of sectors per fragment */ 679 u_int32_t dlfs_cleansz; /* 124: cleaner info size in blocks */ 680 u_int32_t dlfs_segtabsz; /* 128: segment table size in blocks */ 681 u_int32_t dlfs_segmask; /* 132: calculate offset within a segment */ 682 u_int32_t dlfs_segshift; /* 136: fast mult/div for segments */ 683 u_int32_t dlfs_bshift; /* 140: calc block number from file offset */ 684 u_int32_t dlfs_ffshift; /* 144: fast mult/div for frag from file */ 685 u_int32_t dlfs_fbshift; /* 148: fast mult/div for frag from block */ 686 u_int64_t dlfs_bmask; /* 152: calc block offset from file offset */ 687 u_int64_t dlfs_ffmask; /* 160: calc frag offset from file offset */ 688 u_int64_t dlfs_fbmask; /* 168: calc frag offset from block offset */ 689 u_int32_t dlfs_blktodb; /* 176: blktodb and dbtoblk shift constant */ 690 u_int32_t dlfs_sushift; /* 180: fast mult/div for segusage table */ 691 692 int32_t dlfs_maxsymlinklen; /* 184: max length of an internal symlink */ 693 #define LFS_MIN_SBINTERVAL 5 /* minimum superblock segment spacing */ 694 #define LFS_MAXNUMSB 10 /* 188: superblock disk offsets */ 695 int32_t dlfs_sboffs[LFS_MAXNUMSB]; 696 697 u_int32_t dlfs_nclean; /* 228: Number of clean segments */ 698 u_char dlfs_fsmnt[MNAMELEN]; /* 232: name mounted on */ 699 #define LFS_PF_CLEAN 0x1 700 u_int16_t dlfs_pflags; /* 322: file system persistent flags */ 701 int32_t dlfs_dmeta; /* 324: total number of dirty summaries */ 702 u_int32_t dlfs_minfreeseg; /* 328: segments not counted in bfree */ 703 u_int32_t dlfs_sumsize; /* 332: size of summary blocks */ 704 u_int64_t dlfs_serial; /* 336: serial number */ 705 u_int32_t dlfs_ibsize; /* 344: size of inode blocks */ 706 int32_t dlfs_start; /* 348: start of segment 0 */ 707 u_int64_t dlfs_tstamp; /* 352: time stamp */ 708 #define LFS_44INODEFMT 0 709 #define LFS_MAXINODEFMT 0 710 u_int32_t dlfs_inodefmt; /* 360: inode format version */ 711 u_int32_t dlfs_interleave; /* 364: segment interleave */ 712 u_int32_t dlfs_ident; /* 368: per-fs identifier */ 713 u_int32_t dlfs_fsbtodb; /* 372: fsbtodb abd dbtodsb shift constant */ 714 u_int32_t dlfs_resvseg; /* 376: segments reserved for the cleaner */ 715 int8_t dlfs_pad[128]; /* 380: round to 512 bytes */ 716 /* Checksum -- last valid disk field. */ 717 u_int32_t dlfs_cksum; /* 508: checksum for superblock checking */ 718 }; 719 720 /* Type used for the inode bitmap */ 721 typedef u_int32_t lfs_bm_t; 722 723 /* 724 * Linked list of segments whose byte count needs updating following a 725 * file truncation. 726 */ 727 struct segdelta { 728 long segnum; 729 size_t num; 730 LIST_ENTRY(segdelta) list; 731 }; 732 733 /* 734 * In-memory super block. 735 */ 736 struct lfs { 737 struct dlfs lfs_dlfs; /* on-disk parameters */ 738 #define lfs_magic lfs_dlfs.dlfs_magic 739 #define lfs_version lfs_dlfs.dlfs_version 740 #define lfs_size lfs_dlfs.dlfs_size 741 #define lfs_ssize lfs_dlfs.dlfs_ssize 742 #define lfs_dsize lfs_dlfs.dlfs_dsize 743 #define lfs_bsize lfs_dlfs.dlfs_bsize 744 #define lfs_fsize lfs_dlfs.dlfs_fsize 745 #define lfs_frag lfs_dlfs.dlfs_frag 746 #define lfs_freehd lfs_dlfs.dlfs_freehd 747 #define lfs_bfree lfs_dlfs.dlfs_bfree 748 #define lfs_nfiles lfs_dlfs.dlfs_nfiles 749 #define lfs_avail lfs_dlfs.dlfs_avail 750 #define lfs_uinodes lfs_dlfs.dlfs_uinodes 751 #define lfs_idaddr lfs_dlfs.dlfs_idaddr 752 #define lfs_ifile lfs_dlfs.dlfs_ifile 753 #define lfs_lastseg lfs_dlfs.dlfs_lastseg 754 #define lfs_nextseg lfs_dlfs.dlfs_nextseg 755 #define lfs_curseg lfs_dlfs.dlfs_curseg 756 #define lfs_offset lfs_dlfs.dlfs_offset 757 #define lfs_lastpseg lfs_dlfs.dlfs_lastpseg 758 #define lfs_otstamp lfs_dlfs.dlfs_inopf 759 #define lfs_inopf lfs_dlfs.dlfs_inopf 760 #define lfs_minfree lfs_dlfs.dlfs_minfree 761 #define lfs_maxfilesize lfs_dlfs.dlfs_maxfilesize 762 #define lfs_fsbpseg lfs_dlfs.dlfs_fsbpseg 763 #define lfs_inopb lfs_dlfs.dlfs_inopb 764 #define lfs_ifpb lfs_dlfs.dlfs_ifpb 765 #define lfs_sepb lfs_dlfs.dlfs_sepb 766 #define lfs_nindir lfs_dlfs.dlfs_nindir 767 #define lfs_nseg lfs_dlfs.dlfs_nseg 768 #define lfs_nspf lfs_dlfs.dlfs_nspf 769 #define lfs_cleansz lfs_dlfs.dlfs_cleansz 770 #define lfs_segtabsz lfs_dlfs.dlfs_segtabsz 771 #define lfs_segmask lfs_dlfs.dlfs_segmask 772 #define lfs_segshift lfs_dlfs.dlfs_segshift 773 #define lfs_bmask lfs_dlfs.dlfs_bmask 774 #define lfs_bshift lfs_dlfs.dlfs_bshift 775 #define lfs_ffmask lfs_dlfs.dlfs_ffmask 776 #define lfs_ffshift lfs_dlfs.dlfs_ffshift 777 #define lfs_fbmask lfs_dlfs.dlfs_fbmask 778 #define lfs_fbshift lfs_dlfs.dlfs_fbshift 779 #define lfs_blktodb lfs_dlfs.dlfs_blktodb 780 #define lfs_fsbtodb lfs_dlfs.dlfs_fsbtodb 781 #define lfs_sushift lfs_dlfs.dlfs_sushift 782 #define lfs_maxsymlinklen lfs_dlfs.dlfs_maxsymlinklen 783 #define lfs_sboffs lfs_dlfs.dlfs_sboffs 784 #define lfs_cksum lfs_dlfs.dlfs_cksum 785 #define lfs_pflags lfs_dlfs.dlfs_pflags 786 #define lfs_fsmnt lfs_dlfs.dlfs_fsmnt 787 #define lfs_nclean lfs_dlfs.dlfs_nclean 788 #define lfs_dmeta lfs_dlfs.dlfs_dmeta 789 #define lfs_minfreeseg lfs_dlfs.dlfs_minfreeseg 790 #define lfs_sumsize lfs_dlfs.dlfs_sumsize 791 #define lfs_serial lfs_dlfs.dlfs_serial 792 #define lfs_ibsize lfs_dlfs.dlfs_ibsize 793 #define lfs_start lfs_dlfs.dlfs_start 794 #define lfs_tstamp lfs_dlfs.dlfs_tstamp 795 #define lfs_inodefmt lfs_dlfs.dlfs_inodefmt 796 #define lfs_interleave lfs_dlfs.dlfs_interleave 797 #define lfs_ident lfs_dlfs.dlfs_ident 798 #define lfs_resvseg lfs_dlfs.dlfs_resvseg 799 800 /* These fields are set at mount time and are meaningless on disk. */ 801 struct segment *lfs_sp; /* current segment being written */ 802 struct vnode *lfs_ivnode; /* vnode for the ifile */ 803 u_int32_t lfs_seglock; /* single-thread the segment writer */ 804 pid_t lfs_lockpid; /* pid of lock holder */ 805 lwpid_t lfs_locklwp; /* lwp of lock holder */ 806 u_int32_t lfs_iocount; /* number of ios pending */ 807 u_int32_t lfs_writer; /* don't allow any dirops to start */ 808 u_int32_t lfs_dirops; /* count of active directory ops */ 809 u_int32_t lfs_dirvcount; /* count of VDIROP nodes in this fs */ 810 u_int32_t lfs_doifile; /* Write ifile blocks on next write */ 811 u_int32_t lfs_nactive; /* Number of segments since last ckp */ 812 int8_t lfs_fmod; /* super block modified flag */ 813 int8_t lfs_ronly; /* mounted read-only flag */ 814 #define LFS_NOTYET 0x01 815 #define LFS_IFDIRTY 0x02 816 #define LFS_WARNED 0x04 817 #define LFS_UNDIROP 0x08 818 int8_t lfs_flags; /* currently unused flag */ 819 u_int16_t lfs_activesb; /* toggle between superblocks */ 820 daddr_t lfs_sbactive; /* disk address of current sb write */ 821 struct vnode *lfs_flushvp; /* vnode being flushed */ 822 int lfs_flushvp_fakevref; /* fake vref count for flushvp */ 823 struct vnode *lfs_unlockvp; /* being inactivated in lfs_segunlock */ 824 u_int32_t lfs_diropwait; /* # procs waiting on dirop flush */ 825 size_t lfs_devbsize; /* Device block size */ 826 size_t lfs_devbshift; /* Device block shift */ 827 krwlock_t lfs_fraglock; 828 struct lock lfs_iflock; /* Ifile lock */ 829 struct lock lfs_stoplock; /* Wrap lock */ 830 pid_t lfs_rfpid; /* Process ID of roll-forward agent */ 831 int lfs_nadirop; /* number of active dirop nodes */ 832 long lfs_ravail; /* blocks pre-reserved for writing */ 833 long lfs_favail; /* blocks pre-reserved for writing */ 834 res_t *lfs_resblk; /* Reserved memory for pageout */ 835 TAILQ_HEAD(, inode) lfs_dchainhd; /* dirop vnodes */ 836 TAILQ_HEAD(, inode) lfs_pchainhd; /* paging vnodes */ 837 #define LFS_RESHASH_WIDTH 17 838 LIST_HEAD(, lfs_res_blk) lfs_reshash[LFS_RESHASH_WIDTH]; 839 int lfs_pdflush; /* pagedaemon wants us to flush */ 840 u_int32_t **lfs_suflags; /* Segment use flags */ 841 #ifdef _KERNEL 842 struct pool lfs_clpool; /* Pool for struct lfs_cluster */ 843 struct pool lfs_bpppool; /* Pool for bpp */ 844 struct pool lfs_segpool; /* Pool for struct segment */ 845 #endif /* _KERNEL */ 846 #define LFS_MAX_CLEANIND 64 847 int32_t lfs_cleanint[LFS_MAX_CLEANIND]; /* Active cleaning intervals */ 848 int lfs_cleanind; /* Index into intervals */ 849 struct simplelock lfs_interlock; /* lock for lfs_seglock */ 850 int lfs_sleepers; /* # procs sleeping this fs */ 851 int lfs_pages; /* dirty pages blaming this fs */ 852 lfs_bm_t *lfs_ino_bitmap; /* Inuse inodes bitmap */ 853 int lfs_nowrap; /* Suspend log wrap */ 854 int lfs_wrappass; /* Allow first log wrap requester to pass */ 855 int lfs_wrapstatus; /* Wrap status */ 856 LIST_HEAD(, segdelta) lfs_segdhd; /* List of pending trunc accounting events */ 857 }; 858 859 /* NINDIR is the number of indirects in a file system block. */ 860 #define NINDIR(fs) ((fs)->lfs_nindir) 861 862 /* INOPB is the number of inodes in a secondary storage block. */ 863 #define INOPB(fs) ((fs)->lfs_inopb) 864 /* INOPF is the number of inodes in a fragment. */ 865 #define INOPF(fs) ((fs)->lfs_inopf) 866 867 #define blksize(fs, ip, lbn) \ 868 (((lbn) >= NDADDR || (ip)->i_ffs1_size >= ((lbn) + 1) << (fs)->lfs_bshift) \ 869 ? (fs)->lfs_bsize \ 870 : (fragroundup(fs, blkoff(fs, (ip)->i_ffs1_size)))) 871 #define blkoff(fs, loc) ((int)((loc) & (fs)->lfs_bmask)) 872 #define fragoff(fs, loc) /* calculates (loc % fs->lfs_fsize) */ \ 873 ((int)((loc) & (fs)->lfs_ffmask)) 874 #define fsbtodb(fs, b) ((b) << (fs)->lfs_fsbtodb) 875 #define dbtofsb(fs, b) ((b) >> (fs)->lfs_fsbtodb) 876 #define fragstodb(fs, b) ((b) << ((fs)->lfs_blktodb - (fs)->lfs_fbshift)) 877 #define dbtofrags(fs, b) ((b) >> ((fs)->lfs_blktodb - (fs)->lfs_fbshift)) 878 #define lblkno(fs, loc) ((loc) >> (fs)->lfs_bshift) 879 #define lblktosize(fs, blk) ((blk) << (fs)->lfs_bshift) 880 /* Same as above, but named like dbtob(), btodb() */ 881 #define fsbtob(fs, b) ((b) << ((fs)->lfs_bshift - \ 882 (fs)->lfs_blktodb + (fs)->lfs_fsbtodb)) 883 #define btofsb(fs, b) ((b) >> ((fs)->lfs_bshift - \ 884 (fs)->lfs_blktodb + (fs)->lfs_fsbtodb)) 885 #define fsbtofrags(fs, b) ((b) >> ((fs)->lfs_blktodb - (fs)->lfs_fbshift - \ 886 (fs)->lfs_fsbtodb)) 887 #define fragstofsb(fs, b) ((b) << ((fs)->lfs_blktodb - (fs)->lfs_fbshift - \ 888 (fs)->lfs_fsbtodb)) 889 #define btofrags(fs, b) ((b) >> (fs)->lfs_ffshift) 890 #define numfrags(fs, loc) /* calculates (loc / fs->lfs_fsize) */ \ 891 ((loc) >> (fs)->lfs_ffshift) 892 #define blkroundup(fs, size) /* calculates roundup(size, fs->lfs_bsize) */ \ 893 ((off_t)(((size) + (fs)->lfs_bmask) & (~(fs)->lfs_bmask))) 894 #define fragroundup(fs, size) /* calculates roundup(size, fs->lfs_fsize) */ \ 895 ((off_t)(((size) + (fs)->lfs_ffmask) & (~(fs)->lfs_ffmask))) 896 #define fragstoblks(fs, frags) /* calculates (frags / fs->lfs_frag) */ \ 897 ((frags) >> (fs)->lfs_fbshift) 898 #define blkstofrags(fs, blks) /* calculates (blks * fs->lfs_frag) */ \ 899 ((blks) << (fs)->lfs_fbshift) 900 #define fragnum(fs, fsb) /* calculates (fsb % fs->lfs_frag) */ \ 901 ((fsb) & ((fs)->lfs_frag - 1)) 902 #define blknum(fs, fsb) /* calculates rounddown(fsb, fs->lfs_frag) */ \ 903 ((fsb) &~ ((fs)->lfs_frag - 1)) 904 #define dblksize(fs, dp, lbn) \ 905 (((lbn) >= NDADDR || (dp)->di_size >= ((lbn) + 1) << (fs)->lfs_bshift)\ 906 ? (fs)->lfs_bsize \ 907 : (fragroundup(fs, blkoff(fs, (dp)->di_size)))) 908 909 #define segsize(fs) ((fs)->lfs_version == 1 ? \ 910 lblktosize((fs), (fs)->lfs_ssize) : \ 911 (fs)->lfs_ssize) 912 #define segtod(fs, seg) (((fs)->lfs_version == 1 ? \ 913 (fs)->lfs_ssize << (fs)->lfs_blktodb : \ 914 btofsb((fs), (fs)->lfs_ssize)) * (seg)) 915 #define dtosn(fs, daddr) /* block address to segment number */ \ 916 ((uint32_t)(((daddr) - (fs)->lfs_start) / segtod((fs), 1))) 917 #define sntod(fs, sn) /* segment number to disk address */ \ 918 ((daddr_t)(segtod((fs), (sn)) + (fs)->lfs_start)) 919 920 /* 921 * Structures used by lfs_bmapv and lfs_markv to communicate information 922 * about inodes and data blocks. 923 */ 924 typedef struct block_info { 925 u_int32_t bi_inode; /* inode # */ 926 int32_t bi_lbn; /* logical block w/in file */ 927 int32_t bi_daddr; /* disk address of block */ 928 u_int64_t bi_segcreate; /* origin segment create time */ 929 int bi_version; /* file version number */ 930 void *bi_bp; /* data buffer */ 931 int bi_size; /* size of the block (if fragment) */ 932 } BLOCK_INFO; 933 934 /* Compatibility for 1.5 binaries */ 935 typedef struct block_info_15 { 936 u_int32_t bi_inode; /* inode # */ 937 int32_t bi_lbn; /* logical block w/in file */ 938 int32_t bi_daddr; /* disk address of block */ 939 u_int32_t bi_segcreate; /* origin segment create time */ 940 int bi_version; /* file version number */ 941 void *bi_bp; /* data buffer */ 942 int bi_size; /* size of the block (if fragment) */ 943 } BLOCK_INFO_15; 944 945 /* In-memory description of a segment about to be written. */ 946 struct segment { 947 struct lfs *fs; /* file system pointer */ 948 struct buf **bpp; /* pointer to buffer array */ 949 struct buf **cbpp; /* pointer to next available bp */ 950 struct buf **start_bpp; /* pointer to first bp in this set */ 951 struct buf *ibp; /* buffer pointer to inode page */ 952 struct ufs1_dinode *idp; /* pointer to ifile dinode */ 953 struct finfo *fip; /* current fileinfo pointer */ 954 struct vnode *vp; /* vnode being gathered */ 955 void *segsum; /* segment summary info */ 956 u_int32_t ninodes; /* number of inodes in this segment */ 957 int32_t seg_bytes_left; /* bytes left in segment */ 958 int32_t sum_bytes_left; /* bytes left in summary block */ 959 u_int32_t seg_number; /* number of this segment */ 960 int32_t *start_lbp; /* beginning lbn for this set */ 961 962 #define SEGM_CKP 0x01 /* doing a checkpoint */ 963 #define SEGM_CLEAN 0x02 /* cleaner call; don't sort */ 964 #define SEGM_SYNC 0x04 /* wait for segment */ 965 #define SEGM_PROT 0x08 /* don't inactivate at segunlock */ 966 #define SEGM_PAGEDAEMON 0x10 /* pagedaemon called us */ 967 #define SEGM_WRITERD 0x20 /* LFS writed called us */ 968 #define SEGM_FORCE_CKP 0x40 /* Force checkpoint right away */ 969 u_int16_t seg_flags; /* run-time flags for this segment */ 970 u_int32_t seg_iocount; /* number of ios pending */ 971 int ndupino; /* number of duplicate inodes */ 972 }; 973 974 #ifdef _KERNEL 975 struct lfs_cluster { 976 size_t bufsize; /* Size of kept data */ 977 struct buf **bpp; /* Array of kept buffers */ 978 int bufcount; /* Number of kept buffers */ 979 #define LFS_CL_MALLOC 0x00000001 980 #define LFS_CL_SHIFT 0x00000002 981 #define LFS_CL_SYNC 0x00000004 982 u_int32_t flags; /* Flags */ 983 struct lfs *fs; /* LFS that this belongs to */ 984 struct segment *seg; /* Segment structure, for LFS_CL_SYNC */ 985 }; 986 987 /* 988 * Splay tree containing block numbers allocated through lfs_balloc. 989 */ 990 struct lbnentry { 991 SPLAY_ENTRY(lbnentry) entry; 992 daddr_t lbn; 993 }; 994 #endif /* _KERNEL */ 995 996 /* 997 * LFS inode extensions. 998 */ 999 struct lfs_inode_ext { 1000 off_t lfs_osize; /* size of file on disk */ 1001 u_int32_t lfs_effnblocks; /* number of blocks when i/o completes */ 1002 size_t lfs_fragsize[NDADDR]; /* size of on-disk direct blocks */ 1003 TAILQ_ENTRY(inode) lfs_dchain; /* Dirop chain. */ 1004 TAILQ_ENTRY(inode) lfs_pchain; /* Paging chain. */ 1005 #define LFSI_NO_GOP_WRITE 0x01 1006 #define LFSI_DELETED 0x02 1007 #define LFSI_WRAPBLOCK 0x04 1008 #define LFSI_WRAPWAIT 0x08 1009 u_int32_t lfs_iflags; /* Inode flags */ 1010 daddr_t lfs_hiblk; /* Highest lbn held by inode */ 1011 #ifdef _KERNEL 1012 SPLAY_HEAD(lfs_splay, lbnentry) lfs_lbtree; /* Tree of balloc'd lbns */ 1013 int lfs_nbtree; /* Size of tree */ 1014 LIST_HEAD(, segdelta) lfs_segdhd; 1015 #endif 1016 int16_t lfs_odnlink; /* on-disk nlink count for cleaner */ 1017 }; 1018 #define i_lfs_osize inode_ext.lfs->lfs_osize 1019 #define i_lfs_effnblks inode_ext.lfs->lfs_effnblocks 1020 #define i_lfs_fragsize inode_ext.lfs->lfs_fragsize 1021 #define i_lfs_dchain inode_ext.lfs->lfs_dchain 1022 #define i_lfs_pchain inode_ext.lfs->lfs_pchain 1023 #define i_lfs_iflags inode_ext.lfs->lfs_iflags 1024 #define i_lfs_hiblk inode_ext.lfs->lfs_hiblk 1025 #define i_lfs_lbtree inode_ext.lfs->lfs_lbtree 1026 #define i_lfs_nbtree inode_ext.lfs->lfs_nbtree 1027 #define i_lfs_segdhd inode_ext.lfs->lfs_segdhd 1028 #define i_lfs_odnlink inode_ext.lfs->lfs_odnlink 1029 1030 /* 1031 * Macros for determining free space on the disk, with the variable metadata 1032 * of segment summaries and inode blocks taken into account. 1033 */ 1034 /* Estimate number of clean blocks not available for writing */ 1035 #define LFS_EST_CMETA(F) (int32_t)((((F)->lfs_dmeta * \ 1036 (int64_t)(F)->lfs_nclean) / \ 1037 ((F)->lfs_nseg - (F)->lfs_nclean))) 1038 1039 /* Estimate total size of the disk not including metadata */ 1040 #define LFS_EST_NONMETA(F) ((F)->lfs_dsize - (F)->lfs_dmeta - LFS_EST_CMETA(F)) 1041 1042 /* Estimate number of blocks actually available for writing */ 1043 #define LFS_EST_BFREE(F) ((F)->lfs_bfree > LFS_EST_CMETA(F) ? \ 1044 (F)->lfs_bfree - LFS_EST_CMETA(F) : 0) 1045 1046 /* Amount of non-meta space not available to mortal man */ 1047 #define LFS_EST_RSVD(F) (int32_t)((LFS_EST_NONMETA(F) * \ 1048 (u_int64_t)(F)->lfs_minfree) / \ 1049 100) 1050 1051 /* Can credential C write BB blocks */ 1052 #define ISSPACE(F, BB, C) \ 1053 ((((C) == NOCRED || kauth_cred_geteuid(C) == 0) && \ 1054 LFS_EST_BFREE(F) >= (BB)) || \ 1055 (kauth_cred_geteuid(C) != 0 && IS_FREESPACE(F, BB))) 1056 1057 /* Can an ordinary user write BB blocks */ 1058 #define IS_FREESPACE(F, BB) \ 1059 (LFS_EST_BFREE(F) >= (BB) + LFS_EST_RSVD(F)) 1060 1061 /* 1062 * The minimum number of blocks to create a new inode. This is: 1063 * directory direct block (1) + NIADDR indirect blocks + inode block (1) + 1064 * ifile direct block (1) + NIADDR indirect blocks = 3 + 2 * NIADDR blocks. 1065 */ 1066 #define LFS_NRESERVE(F) (btofsb((F), (2 * NIADDR + 3) << (F)->lfs_bshift)) 1067 1068 /* Statistics Counters */ 1069 struct lfs_stats { /* Must match sysctl list in lfs_vfsops.h ! */ 1070 u_int segsused; 1071 u_int psegwrites; 1072 u_int psyncwrites; 1073 u_int pcleanwrites; 1074 u_int blocktot; 1075 u_int cleanblocks; 1076 u_int ncheckpoints; 1077 u_int nwrites; 1078 u_int nsync_writes; 1079 u_int wait_exceeded; 1080 u_int write_exceeded; 1081 u_int flush_invoked; 1082 u_int vflush_invoked; 1083 u_int clean_inlocked; 1084 u_int clean_vnlocked; 1085 u_int segs_reclaimed; 1086 }; 1087 #ifdef _KERNEL 1088 extern struct lfs_stats lfs_stats; 1089 #endif 1090 1091 /* Fcntls to take the place of the lfs syscalls */ 1092 struct lfs_fcntl_markv { 1093 BLOCK_INFO *blkiov; /* blocks to relocate */ 1094 int blkcnt; /* number of blocks */ 1095 }; 1096 1097 #define LFCNSEGWAITALL _FCNR_FSPRIV('L', 0, struct timeval) 1098 #define LFCNSEGWAIT _FCNR_FSPRIV('L', 1, struct timeval) 1099 #define LFCNBMAPV _FCNRW_FSPRIV('L', 2, struct lfs_fcntl_markv) 1100 #define LFCNMARKV _FCNRW_FSPRIV('L', 3, struct lfs_fcntl_markv) 1101 #define LFCNRECLAIM _FCNO_FSPRIV('L', 4) 1102 1103 struct lfs_fhandle { 1104 char space[28]; /* FHANDLE_SIZE_COMPAT (but used from userland too) */ 1105 }; 1106 #define LFCNREWIND _FCNR_FSPRIV('L', 6, int) 1107 #define LFCNINVAL _FCNR_FSPRIV('L', 7, int) 1108 #define LFCNRESIZE _FCNR_FSPRIV('L', 8, int) 1109 #define LFCNWRAPSTOP _FCNR_FSPRIV('L', 9, int) 1110 #define LFCNWRAPGO _FCNR_FSPRIV('L', 10, int) 1111 #define LFCNIFILEFH _FCNW_FSPRIV('L', 11, struct lfs_fhandle) 1112 #define LFCNWRAPPASS _FCNR_FSPRIV('L', 12, int) 1113 # define LFS_WRAP_GOING 0x0 1114 # define LFS_WRAP_WAITING 0x1 1115 #define LFCNWRAPSTATUS _FCNW_FSPRIV('L', 13, int) 1116 /* Compat */ 1117 #define LFCNSEGWAITALL_COMPAT _FCNW_FSPRIV('L', 0, struct timeval) 1118 #define LFCNSEGWAIT_COMPAT _FCNW_FSPRIV('L', 1, struct timeval) 1119 #define LFCNIFILEFH_COMPAT _FCNW_FSPRIV('L', 5, struct lfs_fhandle) 1120 #define LFCNIFILEFH_COMPAT2 _FCN_FSPRIV(F_FSOUT, 'L', 11, 32) 1121 #define LFCNWRAPSTOP_COMPAT _FCNO_FSPRIV('L', 9) 1122 #define LFCNWRAPGO_COMPAT _FCNO_FSPRIV('L', 10) 1123 1124 #ifdef _KERNEL 1125 /* XXX MP */ 1126 #define LFS_SEGLOCK_HELD(fs) \ 1127 ((fs)->lfs_seglock != 0 && \ 1128 (fs)->lfs_lockpid == curproc->p_pid && \ 1129 (fs)->lfs_locklwp == curlwp->l_lid) 1130 #endif /* _KERNEL */ 1131 1132 /* Debug segment lock */ 1133 #ifdef notyet 1134 # define ASSERT_SEGLOCK(fs) KASSERT(LFS_SEGLOCK_HELD(fs)) 1135 # define ASSERT_NO_SEGLOCK(fs) KASSERT(!LFS_SEGLOCK_HELD(fs)) 1136 # define ASSERT_DUNNO_SEGLOCK(fs) 1137 # define ASSERT_MAYBE_SEGLOCK(fs) 1138 #else /* !notyet */ 1139 # define ASSERT_DUNNO_SEGLOCK(fs) \ 1140 DLOG((DLOG_SEG, "lfs func %s seglock wrong (%d)\n", __func__, \ 1141 LFS_SEGLOCK_HELD(fs))) 1142 # define ASSERT_SEGLOCK(fs) do { \ 1143 if (!LFS_SEGLOCK_HELD(fs)) { \ 1144 DLOG((DLOG_SEG, "lfs func %s seglock wrong (0)\n", __func__)); \ 1145 } \ 1146 } while(0) 1147 # define ASSERT_NO_SEGLOCK(fs) do { \ 1148 if (LFS_SEGLOCK_HELD(fs)) { \ 1149 DLOG((DLOG_SEG, "lfs func %s seglock wrong (1)\n", __func__)); \ 1150 } \ 1151 } while(0) 1152 # define ASSERT_MAYBE_SEGLOCK(x) 1153 #endif /* !notyet */ 1154 1155 __BEGIN_DECLS 1156 void lfs_itimes(struct inode *, const struct timespec *, 1157 const struct timespec *, const struct timespec *); 1158 __END_DECLS 1159 1160 #endif /* !_UFS_LFS_LFS_H_ */ 1161
Indexes created Tue Sep 30 06:09:59 GMT 2025