lfs.h revision 1.141
1 /* $NetBSD: lfs.h,v 1.141 2013/06/06 00:52:14 dholland 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 * 19 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 20 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 21 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 22 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 23 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 24 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 25 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 26 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 27 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 28 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 29 * POSSIBILITY OF SUCH DAMAGE. 30 */ 31 /*- 32 * Copyright (c) 1991, 1993 33 * The Regents of the University of California. All rights reserved. 34 * 35 * Redistribution and use in source and binary forms, with or without 36 * modification, are permitted provided that the following conditions 37 * are met: 38 * 1. Redistributions of source code must retain the above copyright 39 * notice, this list of conditions and the following disclaimer. 40 * 2. Redistributions in binary form must reproduce the above copyright 41 * notice, this list of conditions and the following disclaimer in the 42 * documentation and/or other materials provided with the distribution. 43 * 3. Neither the name of the University nor the names of its contributors 44 * may be used to endorse or promote products derived from this software 45 * without specific prior written permission. 46 * 47 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 48 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 49 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 50 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 51 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 52 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 53 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 54 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 55 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 56 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 57 * SUCH DAMAGE. 58 * 59 * @(#)lfs.h 8.9 (Berkeley) 5/8/95 60 */ 61 62 #ifndef _UFS_LFS_LFS_H_ 63 #define _UFS_LFS_LFS_H_ 64 65 #include <sys/rwlock.h> 66 #include <sys/mutex.h> 67 #include <sys/queue.h> 68 #include <sys/condvar.h> 69 #include <sys/mount.h> 70 #include <sys/pool.h> 71 72 /* 73 * Compile-time options for LFS. 74 */ 75 #define LFS_IFIND_RETRIES 16 76 #define LFS_LOGLENGTH 1024 /* size of debugging log */ 77 #define LFS_MAX_ACTIVE 10 /* Dirty segments before ckp forced */ 78 79 /* 80 * Fixed filesystem layout parameters 81 */ 82 #define LFS_LABELPAD 8192 /* LFS label size */ 83 #define LFS_SBPAD 8192 /* LFS superblock size */ 84 85 #define LFS_UNUSED_INUM 0 /* 0: out of band inode number */ 86 #define LFS_IFILE_INUM 1 /* 1: IFILE inode number */ 87 /* 2: Root inode number */ 88 #define LFS_LOSTFOUNDINO 3 /* 3: lost+found inode number */ 89 #define LFS_FIRST_INUM 4 /* 4: first free inode number */ 90 91 #define LFS_V1_SUMMARY_SIZE 512 /* V1 fixed summary size */ 92 #define LFS_DFL_SUMMARY_SIZE 512 /* Default summary size */ 93 94 #define LFS_MAX_DADDR 0x7fffffff /* Highest addressable fsb */ 95 96 #define LFS_MAXNAMLEN 255 /* maximum name length in a dir */ 97 98 #define ULFS_NXADDR 2 99 #define ULFS_NDADDR 12 /* Direct addresses in inode. */ 100 #define ULFS_NIADDR 3 /* Indirect addresses in inode. */ 101 102 /* 103 * Adjustable filesystem parameters 104 */ 105 #define MIN_FREE_SEGS 20 106 #define MIN_RESV_SEGS 15 107 #ifndef LFS_ATIME_IFILE 108 # define LFS_ATIME_IFILE 0 /* Store atime info in ifile (optional in LFSv1) */ 109 #endif 110 #define LFS_MARKV_MAXBLKCNT 65536 /* Max block count for lfs_markv() */ 111 112 /* Misc. definitions */ 113 #define BW_CLEAN 1 /* Flag for lfs_bwrite_ext() */ 114 #define PG_DELWRI PG_PAGER1 /* Local def for delayed pageout */ 115 116 /* Resource limits */ 117 #define LFS_MAX_RESOURCE(x, u) (((x) >> 2) - 10 * (u)) 118 #define LFS_WAIT_RESOURCE(x, u) (((x) >> 1) - ((x) >> 3) - 10 * (u)) 119 #define LFS_INVERSE_MAX_RESOURCE(x, u) (((x) + 10 * (u)) << 2) 120 #define LFS_MAX_BUFS LFS_MAX_RESOURCE(nbuf, 1) 121 #define LFS_WAIT_BUFS LFS_WAIT_RESOURCE(nbuf, 1) 122 #define LFS_INVERSE_MAX_BUFS(n) LFS_INVERSE_MAX_RESOURCE(n, 1) 123 #define LFS_MAX_BYTES LFS_MAX_RESOURCE(bufmem_lowater, PAGE_SIZE) 124 #define LFS_INVERSE_MAX_BYTES(n) LFS_INVERSE_MAX_RESOURCE(n, PAGE_SIZE) 125 #define LFS_WAIT_BYTES LFS_WAIT_RESOURCE(bufmem_lowater, PAGE_SIZE) 126 #define LFS_MAX_DIROP ((desiredvnodes >> 2) + (desiredvnodes >> 3)) 127 #define SIZEOF_DIROP(fs) (2 * ((fs)->lfs_bsize + DINODE1_SIZE)) 128 #define LFS_MAX_FSDIROP(fs) \ 129 ((fs)->lfs_nclean <= (fs)->lfs_resvseg ? 0 : \ 130 (((fs)->lfs_nclean - (fs)->lfs_resvseg) * (fs)->lfs_ssize) / \ 131 (2 * SIZEOF_DIROP(fs))) 132 #define LFS_MAX_PAGES lfs_max_pages() 133 #define LFS_WAIT_PAGES lfs_wait_pages() 134 #define LFS_BUFWAIT 2 /* How long to wait if over *_WAIT_* */ 135 136 #ifdef _KERNEL 137 int lfs_wait_pages(void); 138 int lfs_max_pages(void); 139 #endif /* _KERNEL */ 140 141 /* How starved can we be before we start holding back page writes */ 142 #define LFS_STARVED_FOR_SEGS(fs) ((fs)->lfs_nclean < (fs)->lfs_resvseg) 143 144 /* 145 * Reserved blocks for lfs_malloc 146 */ 147 148 /* Structure to keep reserved blocks */ 149 typedef struct lfs_res_blk { 150 void *p; 151 LIST_ENTRY(lfs_res_blk) res; 152 int size; 153 char inuse; 154 } res_t; 155 156 /* Types for lfs_newbuf and lfs_malloc */ 157 #define LFS_NB_UNKNOWN -1 158 #define LFS_NB_SUMMARY 0 159 #define LFS_NB_SBLOCK 1 160 #define LFS_NB_IBLOCK 2 161 #define LFS_NB_CLUSTER 3 162 #define LFS_NB_CLEAN 4 163 #define LFS_NB_BLKIOV 5 164 #define LFS_NB_COUNT 6 /* always last */ 165 166 /* Number of reserved memory blocks of each type */ 167 #define LFS_N_SUMMARIES 2 168 #define LFS_N_SBLOCKS 1 /* Always 1, to throttle superblock writes */ 169 #define LFS_N_IBLOCKS 16 /* In theory ssize/bsize; in practice around 2 */ 170 #define LFS_N_CLUSTERS 16 /* In theory ssize/MAXPHYS */ 171 #define LFS_N_CLEAN 0 172 #define LFS_N_BLKIOV 1 173 174 /* Total count of "large" (non-pool) types */ 175 #define LFS_N_TOTAL (LFS_N_SUMMARIES + LFS_N_SBLOCKS + LFS_N_IBLOCKS + \ 176 LFS_N_CLUSTERS + LFS_N_CLEAN + LFS_N_BLKIOV) 177 178 /* Counts for pool types */ 179 #define LFS_N_CL LFS_N_CLUSTERS 180 #define LFS_N_BPP 2 181 #define LFS_N_SEG 2 182 183 /* 184 * "struct buf" associated definitions 185 */ 186 187 /* Unassigned disk addresses. */ 188 #define UNASSIGNED -1 189 #define UNWRITTEN -2 190 191 /* Unused logical block number */ 192 #define LFS_UNUSED_LBN -1 193 194 /* Determine if a buffer belongs to the ifile */ 195 #define IS_IFILE(bp) (VTOI(bp->b_vp)->i_number == LFS_IFILE_INUM) 196 197 # define LFS_LOCK_BUF(bp) do { \ 198 if (((bp)->b_flags & B_LOCKED) == 0 && bp->b_iodone == NULL) { \ 199 mutex_enter(&lfs_lock); \ 200 ++locked_queue_count; \ 201 locked_queue_bytes += bp->b_bufsize; \ 202 mutex_exit(&lfs_lock); \ 203 } \ 204 (bp)->b_flags |= B_LOCKED; \ 205 } while (0) 206 207 # define LFS_UNLOCK_BUF(bp) do { \ 208 if (((bp)->b_flags & B_LOCKED) != 0 && bp->b_iodone == NULL) { \ 209 mutex_enter(&lfs_lock); \ 210 --locked_queue_count; \ 211 locked_queue_bytes -= bp->b_bufsize; \ 212 if (locked_queue_count < LFS_WAIT_BUFS && \ 213 locked_queue_bytes < LFS_WAIT_BYTES) \ 214 cv_broadcast(&locked_queue_cv); \ 215 mutex_exit(&lfs_lock); \ 216 } \ 217 (bp)->b_flags &= ~B_LOCKED; \ 218 } while (0) 219 220 #ifdef _KERNEL 221 222 extern u_long bufmem_lowater, bufmem_hiwater; /* XXX */ 223 224 # define LFS_IS_MALLOC_BUF(bp) ((bp)->b_iodone == lfs_callback) 225 226 # ifdef DEBUG 227 # define LFS_DEBUG_COUNTLOCKED(m) do { \ 228 if (lfs_debug_log_subsys[DLOG_LLIST]) { \ 229 lfs_countlocked(&locked_queue_count, &locked_queue_bytes, (m)); \ 230 cv_broadcast(&locked_queue_cv); \ 231 } \ 232 } while (0) 233 # else 234 # define LFS_DEBUG_COUNTLOCKED(m) 235 # endif 236 237 /* log for debugging writes to the Ifile */ 238 # ifdef DEBUG 239 struct lfs_log_entry { 240 const char *op; 241 const char *file; 242 int pid; 243 int line; 244 daddr_t block; 245 unsigned long flags; 246 }; 247 extern int lfs_lognum; 248 extern struct lfs_log_entry lfs_log[LFS_LOGLENGTH]; 249 # define LFS_BWRITE_LOG(bp) lfs_bwrite_log((bp), __FILE__, __LINE__) 250 # define LFS_ENTER_LOG(theop, thefile, theline, lbn, theflags, thepid) do {\ 251 int _s; \ 252 \ 253 mutex_enter(&lfs_lock); \ 254 _s = splbio(); \ 255 lfs_log[lfs_lognum].op = theop; \ 256 lfs_log[lfs_lognum].file = thefile; \ 257 lfs_log[lfs_lognum].line = (theline); \ 258 lfs_log[lfs_lognum].pid = (thepid); \ 259 lfs_log[lfs_lognum].block = (lbn); \ 260 lfs_log[lfs_lognum].flags = (theflags); \ 261 lfs_lognum = (lfs_lognum + 1) % LFS_LOGLENGTH; \ 262 splx(_s); \ 263 mutex_exit(&lfs_lock); \ 264 } while (0) 265 266 # define LFS_BCLEAN_LOG(fs, bp) do { \ 267 if ((bp)->b_vp == (fs)->lfs_ivnode) \ 268 LFS_ENTER_LOG("clear", __FILE__, __LINE__, \ 269 bp->b_lblkno, bp->b_flags, curproc->p_pid);\ 270 } while (0) 271 272 /* Must match list in lfs_vfsops.c ! */ 273 # define DLOG_RF 0 /* roll forward */ 274 # define DLOG_ALLOC 1 /* inode alloc */ 275 # define DLOG_AVAIL 2 /* lfs_{,r,f}avail */ 276 # define DLOG_FLUSH 3 /* flush */ 277 # define DLOG_LLIST 4 /* locked list accounting */ 278 # define DLOG_WVNODE 5 /* vflush/writevnodes verbose */ 279 # define DLOG_VNODE 6 /* vflush/writevnodes */ 280 # define DLOG_SEG 7 /* segwrite */ 281 # define DLOG_SU 8 /* seguse accounting */ 282 # define DLOG_CLEAN 9 /* cleaner routines */ 283 # define DLOG_MOUNT 10 /* mount/unmount */ 284 # define DLOG_PAGE 11 /* putpages/gop_write */ 285 # define DLOG_DIROP 12 /* dirop accounting */ 286 # define DLOG_MALLOC 13 /* lfs_malloc accounting */ 287 # define DLOG_MAX 14 /* The terminator */ 288 # define DLOG(a) lfs_debug_log a 289 # else /* ! DEBUG */ 290 # define LFS_BCLEAN_LOG(fs, bp) 291 # define LFS_BWRITE_LOG(bp) VOP_BWRITE((bp)->b_vp, (bp)) 292 # define DLOG(a) 293 # endif /* ! DEBUG */ 294 #else /* ! _KERNEL */ 295 # define LFS_BWRITE_LOG(bp) VOP_BWRITE((bp)) 296 #endif /* _KERNEL */ 297 298 #ifdef _KERNEL 299 /* This overlays the fid structure (see fstypes.h). */ 300 struct ulfs_ufid { 301 u_int16_t ufid_len; /* Length of structure. */ 302 u_int16_t ufid_pad; /* Force 32-bit alignment. */ 303 u_int32_t ufid_ino; /* File number (ino). */ 304 int32_t ufid_gen; /* Generation number. */ 305 }; 306 /* Filehandle structure for exported LFSes */ 307 struct lfid { 308 struct ulfs_ufid lfid_ufid; 309 #define lfid_len lfid_ufid.ufid_len 310 #define lfid_ino lfid_ufid.ufid_ino 311 #define lfid_gen lfid_ufid.ufid_gen 312 uint32_t lfid_ident; 313 }; 314 #endif /* _KERNEL */ 315 316 /* 317 * "struct inode" associated definitions 318 */ 319 320 /* Address calculations for metadata located in the inode */ 321 #define S_INDIR(fs) -ULFS_NDADDR 322 #define D_INDIR(fs) (S_INDIR(fs) - NINDIR(fs) - 1) 323 #define T_INDIR(fs) (D_INDIR(fs) - NINDIR(fs) * NINDIR(fs) - 1) 324 325 /* For convenience */ 326 #define IN_ALLMOD (IN_MODIFIED|IN_ACCESS|IN_CHANGE|IN_UPDATE|IN_MODIFY|IN_ACCESSED|IN_CLEANING) 327 328 #define LFS_SET_UINO(ip, flags) do { \ 329 if (((flags) & IN_ACCESSED) && !((ip)->i_flag & IN_ACCESSED)) \ 330 ++(ip)->i_lfs->lfs_uinodes; \ 331 if (((flags) & IN_CLEANING) && !((ip)->i_flag & IN_CLEANING)) \ 332 ++(ip)->i_lfs->lfs_uinodes; \ 333 if (((flags) & IN_MODIFIED) && !((ip)->i_flag & IN_MODIFIED)) \ 334 ++(ip)->i_lfs->lfs_uinodes; \ 335 (ip)->i_flag |= (flags); \ 336 } while (0) 337 338 #define LFS_CLR_UINO(ip, flags) do { \ 339 if (((flags) & IN_ACCESSED) && ((ip)->i_flag & IN_ACCESSED)) \ 340 --(ip)->i_lfs->lfs_uinodes; \ 341 if (((flags) & IN_CLEANING) && ((ip)->i_flag & IN_CLEANING)) \ 342 --(ip)->i_lfs->lfs_uinodes; \ 343 if (((flags) & IN_MODIFIED) && ((ip)->i_flag & IN_MODIFIED)) \ 344 --(ip)->i_lfs->lfs_uinodes; \ 345 (ip)->i_flag &= ~(flags); \ 346 if ((ip)->i_lfs->lfs_uinodes < 0) { \ 347 panic("lfs_uinodes < 0"); \ 348 } \ 349 } while (0) 350 351 #define LFS_ITIMES(ip, acc, mod, cre) \ 352 while ((ip)->i_flag & (IN_ACCESS | IN_CHANGE | IN_UPDATE | IN_MODIFY)) \ 353 lfs_itimes(ip, acc, mod, cre) 354 355 /* 356 * "struct vnode" associated definitions 357 */ 358 359 /* Heuristic emptiness measure */ 360 #define VPISEMPTY(vp) (LIST_EMPTY(&(vp)->v_dirtyblkhd) && \ 361 !(vp->v_type == VREG && (vp)->v_iflag & VI_ONWORKLST) &&\ 362 VTOI(vp)->i_lfs_nbtree == 0) 363 364 #define WRITEINPROG(vp) ((vp)->v_numoutput > 0 || \ 365 (!LIST_EMPTY(&(vp)->v_dirtyblkhd) && \ 366 !(VTOI(vp)->i_flag & (IN_MODIFIED | IN_ACCESSED | IN_CLEANING)))) 367 368 369 /* 370 * On-disk and in-memory checkpoint segment usage structure. 371 */ 372 typedef struct segusage SEGUSE; 373 struct segusage { 374 u_int32_t su_nbytes; /* 0: number of live bytes */ 375 u_int32_t su_olastmod; /* 4: SEGUSE last modified timestamp */ 376 u_int16_t su_nsums; /* 8: number of summaries in segment */ 377 u_int16_t su_ninos; /* 10: number of inode blocks in seg */ 378 379 #define SEGUSE_ACTIVE 0x01 /* segment currently being written */ 380 #define SEGUSE_DIRTY 0x02 /* segment has data in it */ 381 #define SEGUSE_SUPERBLOCK 0x04 /* segment contains a superblock */ 382 #define SEGUSE_ERROR 0x08 /* cleaner: do not clean segment */ 383 #define SEGUSE_EMPTY 0x10 /* segment is empty */ 384 #define SEGUSE_INVAL 0x20 /* segment is invalid */ 385 u_int32_t su_flags; /* 12: segment flags */ 386 u_int64_t su_lastmod; /* 16: last modified timestamp */ 387 }; 388 389 typedef struct segusage_v1 SEGUSE_V1; 390 struct segusage_v1 { 391 u_int32_t su_nbytes; /* 0: number of live bytes */ 392 u_int32_t su_lastmod; /* 4: SEGUSE last modified timestamp */ 393 u_int16_t su_nsums; /* 8: number of summaries in segment */ 394 u_int16_t su_ninos; /* 10: number of inode blocks in seg */ 395 u_int32_t su_flags; /* 12: segment flags */ 396 }; 397 398 #define SEGUPB(fs) (fs->lfs_sepb) 399 #define SEGTABSIZE_SU(fs) \ 400 (((fs)->lfs_nseg + SEGUPB(fs) - 1) / (fs)->lfs_sepb) 401 402 #ifdef _KERNEL 403 # define SHARE_IFLOCK(F) \ 404 do { \ 405 rw_enter(&(F)->lfs_iflock, RW_READER); \ 406 } while(0) 407 # define UNSHARE_IFLOCK(F) \ 408 do { \ 409 rw_exit(&(F)->lfs_iflock); \ 410 } while(0) 411 #else /* ! _KERNEL */ 412 # define SHARE_IFLOCK(F) 413 # define UNSHARE_IFLOCK(F) 414 #endif /* ! _KERNEL */ 415 416 /* Read in the block with a specific segment usage entry from the ifile. */ 417 #define LFS_SEGENTRY(SP, F, IN, BP) do { \ 418 int _e; \ 419 SHARE_IFLOCK(F); \ 420 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \ 421 if ((_e = bread((F)->lfs_ivnode, \ 422 ((IN) / (F)->lfs_sepb) + (F)->lfs_cleansz, \ 423 (F)->lfs_bsize, NOCRED, 0, &(BP))) != 0) \ 424 panic("lfs: ifile read: %d", _e); \ 425 if ((F)->lfs_version == 1) \ 426 (SP) = (SEGUSE *)((SEGUSE_V1 *)(BP)->b_data + \ 427 ((IN) & ((F)->lfs_sepb - 1))); \ 428 else \ 429 (SP) = (SEGUSE *)(BP)->b_data + ((IN) % (F)->lfs_sepb); \ 430 UNSHARE_IFLOCK(F); \ 431 } while (0) 432 433 #define LFS_WRITESEGENTRY(SP, F, IN, BP) do { \ 434 if ((SP)->su_nbytes == 0) \ 435 (SP)->su_flags |= SEGUSE_EMPTY; \ 436 else \ 437 (SP)->su_flags &= ~SEGUSE_EMPTY; \ 438 (F)->lfs_suflags[(F)->lfs_activesb][(IN)] = (SP)->su_flags; \ 439 LFS_BWRITE_LOG(BP); \ 440 } while (0) 441 442 /* 443 * On-disk file information. One per file with data blocks in the segment. 444 */ 445 typedef struct finfo FINFO; 446 struct finfo { 447 u_int32_t fi_nblocks; /* number of blocks */ 448 u_int32_t fi_version; /* version number */ 449 u_int32_t fi_ino; /* inode number */ 450 u_int32_t fi_lastlength; /* length of last block in array */ 451 int32_t fi_blocks[1]; /* array of logical block numbers */ 452 }; 453 /* sizeof FINFO except fi_blocks */ 454 #define FINFOSIZE (sizeof(FINFO) - sizeof(int32_t)) 455 456 /* 457 * Index file inode entries. 458 */ 459 typedef struct ifile IFILE; 460 struct ifile { 461 u_int32_t if_version; /* inode version number */ 462 #define LFS_UNUSED_DADDR 0 /* out-of-band daddr */ 463 int32_t if_daddr; /* inode disk address */ 464 #define LFS_ORPHAN_NEXTFREE (~(u_int32_t)0) /* indicate orphaned file */ 465 u_int32_t if_nextfree; /* next-unallocated inode */ 466 u_int32_t if_atime_sec; /* Last access time, seconds */ 467 u_int32_t if_atime_nsec; /* and nanoseconds */ 468 }; 469 470 typedef struct ifile_v1 IFILE_V1; 471 struct ifile_v1 { 472 u_int32_t if_version; /* inode version number */ 473 int32_t if_daddr; /* inode disk address */ 474 u_int32_t if_nextfree; /* next-unallocated inode */ 475 #if LFS_ATIME_IFILE 476 struct timespec if_atime; /* Last access time */ 477 #endif 478 }; 479 480 /* 481 * LFSv1 compatibility code is not allowed to touch if_atime, since it 482 * may not be mapped! 483 */ 484 /* Read in the block with a specific inode from the ifile. */ 485 #define LFS_IENTRY(IP, F, IN, BP) do { \ 486 int _e; \ 487 SHARE_IFLOCK(F); \ 488 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \ 489 if ((_e = bread((F)->lfs_ivnode, \ 490 (IN) / (F)->lfs_ifpb + (F)->lfs_cleansz + (F)->lfs_segtabsz, \ 491 (F)->lfs_bsize, NOCRED, 0, &(BP))) != 0) \ 492 panic("lfs: ifile ino %d read %d", (int)(IN), _e); \ 493 if ((F)->lfs_version == 1) \ 494 (IP) = (IFILE *)((IFILE_V1 *)(BP)->b_data + \ 495 (IN) % (F)->lfs_ifpb); \ 496 else \ 497 (IP) = (IFILE *)(BP)->b_data + (IN) % (F)->lfs_ifpb; \ 498 UNSHARE_IFLOCK(F); \ 499 } while (0) 500 501 /* 502 * Cleaner information structure. This resides in the ifile and is used 503 * to pass information from the kernel to the cleaner. 504 */ 505 typedef struct _cleanerinfo { 506 u_int32_t clean; /* number of clean segments */ 507 u_int32_t dirty; /* number of dirty segments */ 508 int32_t bfree; /* disk blocks free */ 509 int32_t avail; /* disk blocks available */ 510 u_int32_t free_head; /* head of the inode free list */ 511 u_int32_t free_tail; /* tail of the inode free list */ 512 #define LFS_CLEANER_MUST_CLEAN 0x01 513 u_int32_t flags; /* status word from the kernel */ 514 } CLEANERINFO; 515 516 #define CLEANSIZE_SU(fs) \ 517 ((sizeof(CLEANERINFO) + (fs)->lfs_bsize - 1) >> (fs)->lfs_bshift) 518 519 /* Read in the block with the cleaner info from the ifile. */ 520 #define LFS_CLEANERINFO(CP, F, BP) do { \ 521 SHARE_IFLOCK(F); \ 522 VTOI((F)->lfs_ivnode)->i_flag |= IN_ACCESS; \ 523 if (bread((F)->lfs_ivnode, \ 524 (daddr_t)0, (F)->lfs_bsize, NOCRED, 0, &(BP))) \ 525 panic("lfs: ifile read"); \ 526 (CP) = (CLEANERINFO *)(BP)->b_data; \ 527 UNSHARE_IFLOCK(F); \ 528 } while (0) 529 530 /* 531 * Synchronize the Ifile cleaner info with current avail and bfree. 532 */ 533 #define LFS_SYNC_CLEANERINFO(cip, fs, bp, w) do { \ 534 mutex_enter(&lfs_lock); \ 535 if ((w) || (cip)->bfree != (fs)->lfs_bfree || \ 536 (cip)->avail != (fs)->lfs_avail - (fs)->lfs_ravail - \ 537 (fs)->lfs_favail) { \ 538 (cip)->bfree = (fs)->lfs_bfree; \ 539 (cip)->avail = (fs)->lfs_avail - (fs)->lfs_ravail - \ 540 (fs)->lfs_favail; \ 541 if (((bp)->b_flags & B_GATHERED) == 0) { \ 542 (fs)->lfs_flags |= LFS_IFDIRTY; \ 543 } \ 544 mutex_exit(&lfs_lock); \ 545 (void) LFS_BWRITE_LOG(bp); /* Ifile */ \ 546 } else { \ 547 mutex_exit(&lfs_lock); \ 548 brelse(bp, 0); \ 549 } \ 550 } while (0) 551 552 /* 553 * Get the head of the inode free list. 554 * Always called with the segment lock held. 555 */ 556 #define LFS_GET_HEADFREE(FS, CIP, BP, FREEP) do { \ 557 if ((FS)->lfs_version > 1) { \ 558 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 559 (FS)->lfs_freehd = (CIP)->free_head; \ 560 brelse(BP, 0); \ 561 } \ 562 *(FREEP) = (FS)->lfs_freehd; \ 563 } while (0) 564 565 #define LFS_PUT_HEADFREE(FS, CIP, BP, VAL) do { \ 566 (FS)->lfs_freehd = (VAL); \ 567 if ((FS)->lfs_version > 1) { \ 568 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 569 (CIP)->free_head = (VAL); \ 570 LFS_BWRITE_LOG(BP); \ 571 mutex_enter(&lfs_lock); \ 572 (FS)->lfs_flags |= LFS_IFDIRTY; \ 573 mutex_exit(&lfs_lock); \ 574 } \ 575 } while (0) 576 577 #define LFS_GET_TAILFREE(FS, CIP, BP, FREEP) do { \ 578 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 579 *(FREEP) = (CIP)->free_tail; \ 580 brelse(BP, 0); \ 581 } while (0) 582 583 #define LFS_PUT_TAILFREE(FS, CIP, BP, VAL) do { \ 584 LFS_CLEANERINFO((CIP), (FS), (BP)); \ 585 (CIP)->free_tail = (VAL); \ 586 LFS_BWRITE_LOG(BP); \ 587 mutex_enter(&lfs_lock); \ 588 (FS)->lfs_flags |= LFS_IFDIRTY; \ 589 mutex_exit(&lfs_lock); \ 590 } while (0) 591 592 /* 593 * On-disk segment summary information 594 */ 595 typedef struct segsum_v1 SEGSUM_V1; 596 struct segsum_v1 { 597 u_int32_t ss_sumsum; /* 0: check sum of summary block */ 598 u_int32_t ss_datasum; /* 4: check sum of data */ 599 u_int32_t ss_magic; /* 8: segment summary magic number */ 600 #define SS_MAGIC 0x061561 601 int32_t ss_next; /* 12: next segment */ 602 u_int32_t ss_create; /* 16: creation time stamp */ 603 u_int16_t ss_nfinfo; /* 20: number of file info structures */ 604 u_int16_t ss_ninos; /* 22: number of inodes in summary */ 605 606 #define SS_DIROP 0x01 /* segment begins a dirop */ 607 #define SS_CONT 0x02 /* more partials to finish this write*/ 608 #define SS_CLEAN 0x04 /* written by the cleaner */ 609 #define SS_RFW 0x08 /* written by the roll-forward agent */ 610 #define SS_RECLAIM 0x10 /* written by the roll-forward agent */ 611 u_int16_t ss_flags; /* 24: used for directory operations */ 612 u_int16_t ss_pad; /* 26: extra space */ 613 /* FINFO's and inode daddr's... */ 614 }; 615 616 typedef struct segsum SEGSUM; 617 struct segsum { 618 u_int32_t ss_sumsum; /* 0: check sum of summary block */ 619 u_int32_t ss_datasum; /* 4: check sum of data */ 620 u_int32_t ss_magic; /* 8: segment summary magic number */ 621 int32_t ss_next; /* 12: next segment */ 622 u_int32_t ss_ident; /* 16: roll-forward fsid */ 623 #define ss_ocreate ss_ident /* ident is where create was in v1 */ 624 u_int16_t ss_nfinfo; /* 20: number of file info structures */ 625 u_int16_t ss_ninos; /* 22: number of inodes in summary */ 626 u_int16_t ss_flags; /* 24: used for directory operations */ 627 u_int8_t ss_pad[2]; /* 26: extra space */ 628 u_int32_t ss_reclino; /* 28: inode being reclaimed */ 629 u_int64_t ss_serial; /* 32: serial number */ 630 u_int64_t ss_create; /* 40: time stamp */ 631 /* FINFO's and inode daddr's... */ 632 }; 633 634 #define SEGSUM_SIZE(fs) ((fs)->lfs_version == 1 ? sizeof(SEGSUM_V1) : sizeof(SEGSUM)) 635 636 637 /* 638 * On-disk super block. 639 */ 640 struct dlfs { 641 #define LFS_MAGIC 0x070162 642 u_int32_t dlfs_magic; /* 0: magic number */ 643 #define LFS_VERSION 2 644 u_int32_t dlfs_version; /* 4: version number */ 645 646 u_int32_t dlfs_size; /* 8: number of blocks in fs (v1) */ 647 /* number of frags in fs (v2) */ 648 u_int32_t dlfs_ssize; /* 12: number of blocks per segment (v1) */ 649 /* number of bytes per segment (v2) */ 650 u_int32_t dlfs_dsize; /* 16: number of disk blocks in fs */ 651 u_int32_t dlfs_bsize; /* 20: file system block size */ 652 u_int32_t dlfs_fsize; /* 24: size of frag blocks in fs */ 653 u_int32_t dlfs_frag; /* 28: number of frags in a block in fs */ 654 655 /* Checkpoint region. */ 656 u_int32_t dlfs_freehd; /* 32: start of the free list */ 657 int32_t dlfs_bfree; /* 36: number of free disk blocks */ 658 u_int32_t dlfs_nfiles; /* 40: number of allocated inodes */ 659 int32_t dlfs_avail; /* 44: blocks available for writing */ 660 int32_t dlfs_uinodes; /* 48: inodes in cache not yet on disk */ 661 int32_t dlfs_idaddr; /* 52: inode file disk address */ 662 u_int32_t dlfs_ifile; /* 56: inode file inode number */ 663 int32_t dlfs_lastseg; /* 60: address of last segment written */ 664 int32_t dlfs_nextseg; /* 64: address of next segment to write */ 665 int32_t dlfs_curseg; /* 68: current segment being written */ 666 int32_t dlfs_offset; /* 72: offset in curseg for next partial */ 667 int32_t dlfs_lastpseg; /* 76: address of last partial written */ 668 u_int32_t dlfs_inopf; /* 80: v1: time stamp; v2: inodes per frag */ 669 #define dlfs_otstamp dlfs_inopf 670 671 /* These are configuration parameters. */ 672 u_int32_t dlfs_minfree; /* 84: minimum percentage of free blocks */ 673 674 /* These fields can be computed from the others. */ 675 u_int64_t dlfs_maxfilesize; /* 88: maximum representable file size */ 676 u_int32_t dlfs_fsbpseg; /* 96: fsb per segment */ 677 u_int32_t dlfs_inopb; /* 100: inodes per block */ 678 u_int32_t dlfs_ifpb; /* 104: IFILE entries per block */ 679 u_int32_t dlfs_sepb; /* 108: SEGUSE entries per block */ 680 u_int32_t dlfs_nindir; /* 112: indirect pointers per block */ 681 u_int32_t dlfs_nseg; /* 116: number of segments */ 682 u_int32_t dlfs_nspf; /* 120: number of sectors per fragment */ 683 u_int32_t dlfs_cleansz; /* 124: cleaner info size in blocks */ 684 u_int32_t dlfs_segtabsz; /* 128: segment table size in blocks */ 685 u_int32_t dlfs_segmask; /* 132: calculate offset within a segment */ 686 u_int32_t dlfs_segshift; /* 136: fast mult/div for segments */ 687 u_int32_t dlfs_bshift; /* 140: calc block number from file offset */ 688 u_int32_t dlfs_ffshift; /* 144: fast mult/div for frag from file */ 689 u_int32_t dlfs_fbshift; /* 148: fast mult/div for frag from block */ 690 u_int64_t dlfs_bmask; /* 152: calc block offset from file offset */ 691 u_int64_t dlfs_ffmask; /* 160: calc frag offset from file offset */ 692 u_int64_t dlfs_fbmask; /* 168: calc frag offset from block offset */ 693 u_int32_t dlfs_blktodb; /* 176: blktodb and dbtoblk shift constant */ 694 u_int32_t dlfs_sushift; /* 180: fast mult/div for segusage table */ 695 696 int32_t dlfs_maxsymlinklen; /* 184: max length of an internal symlink */ 697 #define LFS_MIN_SBINTERVAL 5 /* minimum superblock segment spacing */ 698 #define LFS_MAXNUMSB 10 /* 188: superblock disk offsets */ 699 int32_t dlfs_sboffs[LFS_MAXNUMSB]; 700 701 u_int32_t dlfs_nclean; /* 228: Number of clean segments */ 702 u_char dlfs_fsmnt[MNAMELEN]; /* 232: name mounted on */ 703 #define LFS_PF_CLEAN 0x1 704 u_int16_t dlfs_pflags; /* 322: file system persistent flags */ 705 int32_t dlfs_dmeta; /* 324: total number of dirty summaries */ 706 u_int32_t dlfs_minfreeseg; /* 328: segments not counted in bfree */ 707 u_int32_t dlfs_sumsize; /* 332: size of summary blocks */ 708 u_int64_t dlfs_serial; /* 336: serial number */ 709 u_int32_t dlfs_ibsize; /* 344: size of inode blocks */ 710 int32_t dlfs_start; /* 348: start of segment 0 */ 711 u_int64_t dlfs_tstamp; /* 352: time stamp */ 712 #define LFS_44INODEFMT 0 713 #define LFS_MAXINODEFMT 0 714 u_int32_t dlfs_inodefmt; /* 360: inode format version */ 715 u_int32_t dlfs_interleave; /* 364: segment interleave */ 716 u_int32_t dlfs_ident; /* 368: per-fs identifier */ 717 u_int32_t dlfs_fsbtodb; /* 372: fsbtodb abd dbtodsb shift constant */ 718 u_int32_t dlfs_resvseg; /* 376: segments reserved for the cleaner */ 719 int8_t dlfs_pad[128]; /* 380: round to 512 bytes */ 720 /* Checksum -- last valid disk field. */ 721 u_int32_t dlfs_cksum; /* 508: checksum for superblock checking */ 722 }; 723 724 /* Type used for the inode bitmap */ 725 typedef u_int32_t lfs_bm_t; 726 727 /* 728 * Linked list of segments whose byte count needs updating following a 729 * file truncation. 730 */ 731 struct segdelta { 732 long segnum; 733 size_t num; 734 LIST_ENTRY(segdelta) list; 735 }; 736 737 /* 738 * In-memory super block. 739 */ 740 struct lfs { 741 struct dlfs lfs_dlfs; /* on-disk parameters */ 742 #define lfs_magic lfs_dlfs.dlfs_magic 743 #define lfs_version lfs_dlfs.dlfs_version 744 #define lfs_size lfs_dlfs.dlfs_size 745 #define lfs_ssize lfs_dlfs.dlfs_ssize 746 #define lfs_dsize lfs_dlfs.dlfs_dsize 747 #define lfs_bsize lfs_dlfs.dlfs_bsize 748 #define lfs_fsize lfs_dlfs.dlfs_fsize 749 #define lfs_frag lfs_dlfs.dlfs_frag 750 #define lfs_freehd lfs_dlfs.dlfs_freehd 751 #define lfs_bfree lfs_dlfs.dlfs_bfree 752 #define lfs_nfiles lfs_dlfs.dlfs_nfiles 753 #define lfs_avail lfs_dlfs.dlfs_avail 754 #define lfs_uinodes lfs_dlfs.dlfs_uinodes 755 #define lfs_idaddr lfs_dlfs.dlfs_idaddr 756 #define lfs_ifile lfs_dlfs.dlfs_ifile 757 #define lfs_lastseg lfs_dlfs.dlfs_lastseg 758 #define lfs_nextseg lfs_dlfs.dlfs_nextseg 759 #define lfs_curseg lfs_dlfs.dlfs_curseg 760 #define lfs_offset lfs_dlfs.dlfs_offset 761 #define lfs_lastpseg lfs_dlfs.dlfs_lastpseg 762 #define lfs_otstamp lfs_dlfs.dlfs_inopf 763 #define lfs_inopf lfs_dlfs.dlfs_inopf 764 #define lfs_minfree lfs_dlfs.dlfs_minfree 765 #define lfs_maxfilesize lfs_dlfs.dlfs_maxfilesize 766 #define lfs_fsbpseg lfs_dlfs.dlfs_fsbpseg 767 #define lfs_inopb lfs_dlfs.dlfs_inopb 768 #define lfs_ifpb lfs_dlfs.dlfs_ifpb 769 #define lfs_sepb lfs_dlfs.dlfs_sepb 770 #define lfs_nindir lfs_dlfs.dlfs_nindir 771 #define lfs_nseg lfs_dlfs.dlfs_nseg 772 #define lfs_nspf lfs_dlfs.dlfs_nspf 773 #define lfs_cleansz lfs_dlfs.dlfs_cleansz 774 #define lfs_segtabsz lfs_dlfs.dlfs_segtabsz 775 #define lfs_segmask lfs_dlfs.dlfs_segmask 776 #define lfs_segshift lfs_dlfs.dlfs_segshift 777 #define lfs_bmask lfs_dlfs.dlfs_bmask 778 #define lfs_bshift lfs_dlfs.dlfs_bshift 779 #define lfs_ffmask lfs_dlfs.dlfs_ffmask 780 #define lfs_ffshift lfs_dlfs.dlfs_ffshift 781 #define lfs_fbmask lfs_dlfs.dlfs_fbmask 782 #define lfs_fbshift lfs_dlfs.dlfs_fbshift 783 #define lfs_blktodb lfs_dlfs.dlfs_blktodb 784 #define lfs_fsbtodb lfs_dlfs.dlfs_fsbtodb 785 #define lfs_sushift lfs_dlfs.dlfs_sushift 786 #define lfs_maxsymlinklen lfs_dlfs.dlfs_maxsymlinklen 787 #define lfs_sboffs lfs_dlfs.dlfs_sboffs 788 #define lfs_cksum lfs_dlfs.dlfs_cksum 789 #define lfs_pflags lfs_dlfs.dlfs_pflags 790 #define lfs_fsmnt lfs_dlfs.dlfs_fsmnt 791 #define lfs_nclean lfs_dlfs.dlfs_nclean 792 #define lfs_dmeta lfs_dlfs.dlfs_dmeta 793 #define lfs_minfreeseg lfs_dlfs.dlfs_minfreeseg 794 #define lfs_sumsize lfs_dlfs.dlfs_sumsize 795 #define lfs_serial lfs_dlfs.dlfs_serial 796 #define lfs_ibsize lfs_dlfs.dlfs_ibsize 797 #define lfs_start lfs_dlfs.dlfs_start 798 #define lfs_tstamp lfs_dlfs.dlfs_tstamp 799 #define lfs_inodefmt lfs_dlfs.dlfs_inodefmt 800 #define lfs_interleave lfs_dlfs.dlfs_interleave 801 #define lfs_ident lfs_dlfs.dlfs_ident 802 #define lfs_resvseg lfs_dlfs.dlfs_resvseg 803 804 /* These fields are set at mount time and are meaningless on disk. */ 805 struct segment *lfs_sp; /* current segment being written */ 806 struct vnode *lfs_ivnode; /* vnode for the ifile */ 807 u_int32_t lfs_seglock; /* single-thread the segment writer */ 808 pid_t lfs_lockpid; /* pid of lock holder */ 809 lwpid_t lfs_locklwp; /* lwp of lock holder */ 810 u_int32_t lfs_iocount; /* number of ios pending */ 811 u_int32_t lfs_writer; /* don't allow any dirops to start */ 812 u_int32_t lfs_dirops; /* count of active directory ops */ 813 u_int32_t lfs_dirvcount; /* count of VDIROP nodes in this fs */ 814 u_int32_t lfs_doifile; /* Write ifile blocks on next write */ 815 u_int32_t lfs_nactive; /* Number of segments since last ckp */ 816 int8_t lfs_fmod; /* super block modified flag */ 817 int8_t lfs_ronly; /* mounted read-only flag */ 818 #define LFS_NOTYET 0x01 819 #define LFS_IFDIRTY 0x02 820 #define LFS_WARNED 0x04 821 #define LFS_UNDIROP 0x08 822 int8_t lfs_flags; /* currently unused flag */ 823 u_int16_t lfs_activesb; /* toggle between superblocks */ 824 daddr_t lfs_sbactive; /* disk address of current sb write */ 825 struct vnode *lfs_flushvp; /* vnode being flushed */ 826 int lfs_flushvp_fakevref; /* fake vref count for flushvp */ 827 struct vnode *lfs_unlockvp; /* being inactivated in lfs_segunlock */ 828 u_int32_t lfs_diropwait; /* # procs waiting on dirop flush */ 829 size_t lfs_devbsize; /* Device block size */ 830 size_t lfs_devbshift; /* Device block shift */ 831 krwlock_t lfs_fraglock; 832 krwlock_t lfs_iflock; /* Ifile lock */ 833 kcondvar_t lfs_stopcv; /* Wrap lock */ 834 struct lwp *lfs_stoplwp; 835 pid_t lfs_rfpid; /* Process ID of roll-forward agent */ 836 int lfs_nadirop; /* number of active dirop nodes */ 837 long lfs_ravail; /* blocks pre-reserved for writing */ 838 long lfs_favail; /* blocks pre-reserved for writing */ 839 res_t *lfs_resblk; /* Reserved memory for pageout */ 840 TAILQ_HEAD(, inode) lfs_dchainhd; /* dirop vnodes */ 841 TAILQ_HEAD(, inode) lfs_pchainhd; /* paging vnodes */ 842 #define LFS_RESHASH_WIDTH 17 843 LIST_HEAD(, lfs_res_blk) lfs_reshash[LFS_RESHASH_WIDTH]; 844 int lfs_pdflush; /* pagedaemon wants us to flush */ 845 u_int32_t **lfs_suflags; /* Segment use flags */ 846 #ifdef _KERNEL 847 struct pool lfs_clpool; /* Pool for struct lfs_cluster */ 848 struct pool lfs_bpppool; /* Pool for bpp */ 849 struct pool lfs_segpool; /* Pool for struct segment */ 850 #endif /* _KERNEL */ 851 #define LFS_MAX_CLEANIND 64 852 int32_t lfs_cleanint[LFS_MAX_CLEANIND]; /* Active cleaning intervals */ 853 int lfs_cleanind; /* Index into intervals */ 854 int lfs_sleepers; /* # procs sleeping this fs */ 855 int lfs_pages; /* dirty pages blaming this fs */ 856 lfs_bm_t *lfs_ino_bitmap; /* Inuse inodes bitmap */ 857 int lfs_nowrap; /* Suspend log wrap */ 858 int lfs_wrappass; /* Allow first log wrap requester to pass */ 859 int lfs_wrapstatus; /* Wrap status */ 860 int lfs_reclino; /* Inode being reclaimed */ 861 int lfs_startseg; /* Segment we started writing at */ 862 LIST_HEAD(, segdelta) lfs_segdhd; /* List of pending trunc accounting events */ 863 }; 864 865 /* NINDIR is the number of indirects in a file system block. */ 866 #define NINDIR(fs) ((fs)->lfs_nindir) 867 868 /* INOPB is the number of inodes in a secondary storage block. */ 869 #define INOPB(fs) ((fs)->lfs_inopb) 870 /* INOPF is the number of inodes in a fragment. */ 871 #define INOPF(fs) ((fs)->lfs_inopf) 872 873 #define blksize(fs, ip, lbn) \ 874 (((lbn) >= ULFS_NDADDR || (ip)->i_ffs1_size >= ((lbn) + 1) << (fs)->lfs_bshift) \ 875 ? (fs)->lfs_bsize \ 876 : (fragroundup(fs, blkoff(fs, (ip)->i_ffs1_size)))) 877 #define blkoff(fs, loc) ((int)((loc) & (fs)->lfs_bmask)) 878 #define fragoff(fs, loc) /* calculates (loc % fs->lfs_fsize) */ \ 879 ((int)((loc) & (fs)->lfs_ffmask)) 880 881 #if defined (_KERNEL) 882 #define fsbtodb(fs, b) ((b) << ((fs)->lfs_ffshift - DEV_BSHIFT)) 883 #define dbtofsb(fs, b) ((b) >> ((fs)->lfs_ffshift - DEV_BSHIFT)) 884 #else 885 #define fsbtodb(fs, b) ((b) << (fs)->lfs_fsbtodb) 886 #define dbtofsb(fs, b) ((b) >> (fs)->lfs_fsbtodb) 887 #endif 888 889 #define lblkno(fs, loc) ((loc) >> (fs)->lfs_bshift) 890 #define lblktosize(fs, blk) ((blk) << (fs)->lfs_bshift) 891 892 #define fsbtob(fs, b) ((b) << (fs)->lfs_ffshift) 893 #define btofsb(fs, b) ((b) >> (fs)->lfs_ffshift) 894 895 #define numfrags(fs, loc) /* calculates (loc / fs->lfs_fsize) */ \ 896 ((loc) >> (fs)->lfs_ffshift) 897 #define blkroundup(fs, size) /* calculates roundup(size, fs->lfs_bsize) */ \ 898 ((off_t)(((size) + (fs)->lfs_bmask) & (~(fs)->lfs_bmask))) 899 #define fragroundup(fs, size) /* calculates roundup(size, fs->lfs_fsize) */ \ 900 ((off_t)(((size) + (fs)->lfs_ffmask) & (~(fs)->lfs_ffmask))) 901 #define fragstoblks(fs, frags)/* calculates (frags / fs->fs_frag) */ \ 902 ((frags) >> (fs)->lfs_fbshift) 903 #define blkstofrags(fs, blks) /* calculates (blks * fs->fs_frag) */ \ 904 ((blks) << (fs)->lfs_fbshift) 905 #define fragnum(fs, fsb) /* calculates (fsb % fs->lfs_frag) */ \ 906 ((fsb) & ((fs)->lfs_frag - 1)) 907 #define blknum(fs, fsb) /* calculates rounddown(fsb, fs->lfs_frag) */ \ 908 ((fsb) &~ ((fs)->lfs_frag - 1)) 909 #define dblksize(fs, dp, lbn) \ 910 (((lbn) >= ULFS_NDADDR || (dp)->di_size >= ((lbn) + 1) << (fs)->lfs_bshift)\ 911 ? (fs)->lfs_bsize \ 912 : (fragroundup(fs, blkoff(fs, (dp)->di_size)))) 913 914 #define segsize(fs) ((fs)->lfs_version == 1 ? \ 915 lblktosize((fs), (fs)->lfs_ssize) : \ 916 (fs)->lfs_ssize) 917 #define segtod(fs, seg) (((fs)->lfs_version == 1 ? \ 918 (fs)->lfs_ssize << (fs)->lfs_blktodb : \ 919 btofsb((fs), (fs)->lfs_ssize)) * (seg)) 920 #define dtosn(fs, daddr) /* block address to segment number */ \ 921 ((uint32_t)(((daddr) - (fs)->lfs_start) / segtod((fs), 1))) 922 #define sntod(fs, sn) /* segment number to disk address */ \ 923 ((daddr_t)(segtod((fs), (sn)) + (fs)->lfs_start)) 924 925 /* 926 * Structures used by lfs_bmapv and lfs_markv to communicate information 927 * about inodes and data blocks. 928 */ 929 typedef struct block_info { 930 u_int32_t bi_inode; /* inode # */ 931 int32_t bi_lbn; /* logical block w/in file */ 932 int32_t bi_daddr; /* disk address of block */ 933 u_int64_t bi_segcreate; /* origin segment create time */ 934 int bi_version; /* file version number */ 935 void *bi_bp; /* data buffer */ 936 int bi_size; /* size of the block (if fragment) */ 937 } BLOCK_INFO; 938 939 /* Compatibility for 1.5 binaries */ 940 typedef struct block_info_15 { 941 u_int32_t bi_inode; /* inode # */ 942 int32_t bi_lbn; /* logical block w/in file */ 943 int32_t bi_daddr; /* disk address of block */ 944 u_int32_t bi_segcreate; /* origin segment create time */ 945 int bi_version; /* file version number */ 946 void *bi_bp; /* data buffer */ 947 int bi_size; /* size of the block (if fragment) */ 948 } BLOCK_INFO_15; 949 950 /* In-memory description of a segment about to be written. */ 951 struct segment { 952 struct lfs *fs; /* file system pointer */ 953 struct buf **bpp; /* pointer to buffer array */ 954 struct buf **cbpp; /* pointer to next available bp */ 955 struct buf **start_bpp; /* pointer to first bp in this set */ 956 struct buf *ibp; /* buffer pointer to inode page */ 957 struct ulfs1_dinode *idp; /* pointer to ifile dinode */ 958 struct finfo *fip; /* current fileinfo pointer */ 959 struct vnode *vp; /* vnode being gathered */ 960 void *segsum; /* segment summary info */ 961 u_int32_t ninodes; /* number of inodes in this segment */ 962 int32_t seg_bytes_left; /* bytes left in segment */ 963 int32_t sum_bytes_left; /* bytes left in summary block */ 964 u_int32_t seg_number; /* number of this segment */ 965 int32_t *start_lbp; /* beginning lbn for this set */ 966 967 #define SEGM_CKP 0x0001 /* doing a checkpoint */ 968 #define SEGM_CLEAN 0x0002 /* cleaner call; don't sort */ 969 #define SEGM_SYNC 0x0004 /* wait for segment */ 970 #define SEGM_PROT 0x0008 /* don't inactivate at segunlock */ 971 #define SEGM_PAGEDAEMON 0x0010 /* pagedaemon called us */ 972 #define SEGM_WRITERD 0x0020 /* LFS writed called us */ 973 #define SEGM_FORCE_CKP 0x0040 /* Force checkpoint right away */ 974 #define SEGM_RECLAIM 0x0080 /* Writing to reclaim vnode */ 975 #define SEGM_SINGLE 0x0100 /* Opportunistic writevnodes */ 976 u_int16_t seg_flags; /* run-time flags for this segment */ 977 u_int32_t seg_iocount; /* number of ios pending */ 978 int ndupino; /* number of duplicate inodes */ 979 }; 980 981 #ifdef _KERNEL 982 struct lfs_cluster { 983 size_t bufsize; /* Size of kept data */ 984 struct buf **bpp; /* Array of kept buffers */ 985 int bufcount; /* Number of kept buffers */ 986 #define LFS_CL_MALLOC 0x00000001 987 #define LFS_CL_SHIFT 0x00000002 988 #define LFS_CL_SYNC 0x00000004 989 u_int32_t flags; /* Flags */ 990 struct lfs *fs; /* LFS that this belongs to */ 991 struct segment *seg; /* Segment structure, for LFS_CL_SYNC */ 992 }; 993 994 /* 995 * Splay tree containing block numbers allocated through lfs_balloc. 996 */ 997 struct lbnentry { 998 SPLAY_ENTRY(lbnentry) entry; 999 daddr_t lbn; 1000 }; 1001 #endif /* _KERNEL */ 1002 1003 /* 1004 * LFS inode extensions. 1005 */ 1006 struct lfs_inode_ext { 1007 off_t lfs_osize; /* size of file on disk */ 1008 u_int32_t lfs_effnblocks; /* number of blocks when i/o completes */ 1009 size_t lfs_fragsize[ULFS_NDADDR]; /* size of on-disk direct blocks */ 1010 TAILQ_ENTRY(inode) lfs_dchain; /* Dirop chain. */ 1011 TAILQ_ENTRY(inode) lfs_pchain; /* Paging chain. */ 1012 #define LFSI_NO_GOP_WRITE 0x01 1013 #define LFSI_DELETED 0x02 1014 #define LFSI_WRAPBLOCK 0x04 1015 #define LFSI_WRAPWAIT 0x08 1016 #define LFSI_BMAP 0x10 1017 u_int32_t lfs_iflags; /* Inode flags */ 1018 daddr_t lfs_hiblk; /* Highest lbn held by inode */ 1019 #ifdef _KERNEL 1020 SPLAY_HEAD(lfs_splay, lbnentry) lfs_lbtree; /* Tree of balloc'd lbns */ 1021 int lfs_nbtree; /* Size of tree */ 1022 LIST_HEAD(, segdelta) lfs_segdhd; 1023 #endif 1024 int16_t lfs_odnlink; /* on-disk nlink count for cleaner */ 1025 }; 1026 #define i_lfs_osize inode_ext.lfs->lfs_osize 1027 #define i_lfs_effnblks inode_ext.lfs->lfs_effnblocks 1028 #define i_lfs_fragsize inode_ext.lfs->lfs_fragsize 1029 #define i_lfs_dchain inode_ext.lfs->lfs_dchain 1030 #define i_lfs_pchain inode_ext.lfs->lfs_pchain 1031 #define i_lfs_iflags inode_ext.lfs->lfs_iflags 1032 #define i_lfs_hiblk inode_ext.lfs->lfs_hiblk 1033 #define i_lfs_lbtree inode_ext.lfs->lfs_lbtree 1034 #define i_lfs_nbtree inode_ext.lfs->lfs_nbtree 1035 #define i_lfs_segdhd inode_ext.lfs->lfs_segdhd 1036 #define i_lfs_odnlink inode_ext.lfs->lfs_odnlink 1037 1038 /* 1039 * Macros for determining free space on the disk, with the variable metadata 1040 * of segment summaries and inode blocks taken into account. 1041 */ 1042 /* 1043 * Estimate number of clean blocks not available for writing because 1044 * they will contain metadata or overhead. This is calculated as 1045 * 1046 * E = ((C * M / D) * D + (0) * (T - D)) / T 1047 * or more simply 1048 * E = (C * M) / T 1049 * 1050 * where 1051 * C is the clean space, 1052 * D is the dirty space, 1053 * M is the dirty metadata, and 1054 * T = C + D is the total space on disk. 1055 * 1056 * This approximates the old formula of E = C * M / D when D is close to T, 1057 * but avoids falsely reporting "disk full" when the sample size (D) is small. 1058 */ 1059 #define LFS_EST_CMETA(F) (int32_t)(( \ 1060 ((F)->lfs_dmeta * (int64_t)(F)->lfs_nclean) / \ 1061 ((F)->lfs_nseg))) 1062 1063 /* Estimate total size of the disk not including metadata */ 1064 #define LFS_EST_NONMETA(F) ((F)->lfs_dsize - (F)->lfs_dmeta - LFS_EST_CMETA(F)) 1065 1066 /* Estimate number of blocks actually available for writing */ 1067 #define LFS_EST_BFREE(F) ((F)->lfs_bfree > LFS_EST_CMETA(F) ? \ 1068 (F)->lfs_bfree - LFS_EST_CMETA(F) : 0) 1069 1070 /* Amount of non-meta space not available to mortal man */ 1071 #define LFS_EST_RSVD(F) (int32_t)((LFS_EST_NONMETA(F) * \ 1072 (u_int64_t)(F)->lfs_minfree) / \ 1073 100) 1074 1075 /* Can credential C write BB blocks */ 1076 #define ISSPACE(F, BB, C) \ 1077 ((((C) == NOCRED || kauth_cred_geteuid(C) == 0) && \ 1078 LFS_EST_BFREE(F) >= (BB)) || \ 1079 (kauth_cred_geteuid(C) != 0 && IS_FREESPACE(F, BB))) 1080 1081 /* Can an ordinary user write BB blocks */ 1082 #define IS_FREESPACE(F, BB) \ 1083 (LFS_EST_BFREE(F) >= (BB) + LFS_EST_RSVD(F)) 1084 1085 /* 1086 * The minimum number of blocks to create a new inode. This is: 1087 * directory direct block (1) + ULFS_NIADDR indirect blocks + inode block (1) + 1088 * ifile direct block (1) + ULFS_NIADDR indirect blocks = 3 + 2 * ULFS_NIADDR blocks. 1089 */ 1090 #define LFS_NRESERVE(F) (btofsb((F), (2 * ULFS_NIADDR + 3) << (F)->lfs_bshift)) 1091 1092 /* Statistics Counters */ 1093 struct lfs_stats { /* Must match sysctl list in lfs_vfsops.h ! */ 1094 u_int segsused; 1095 u_int psegwrites; 1096 u_int psyncwrites; 1097 u_int pcleanwrites; 1098 u_int blocktot; 1099 u_int cleanblocks; 1100 u_int ncheckpoints; 1101 u_int nwrites; 1102 u_int nsync_writes; 1103 u_int wait_exceeded; 1104 u_int write_exceeded; 1105 u_int flush_invoked; 1106 u_int vflush_invoked; 1107 u_int clean_inlocked; 1108 u_int clean_vnlocked; 1109 u_int segs_reclaimed; 1110 }; 1111 #ifdef _KERNEL 1112 extern struct lfs_stats lfs_stats; 1113 #endif 1114 1115 /* Fcntls to take the place of the lfs syscalls */ 1116 struct lfs_fcntl_markv { 1117 BLOCK_INFO *blkiov; /* blocks to relocate */ 1118 int blkcnt; /* number of blocks */ 1119 }; 1120 1121 #define LFCNSEGWAITALL _FCNR_FSPRIV('L', 14, struct timeval) 1122 #define LFCNSEGWAIT _FCNR_FSPRIV('L', 15, struct timeval) 1123 #define LFCNBMAPV _FCNRW_FSPRIV('L', 2, struct lfs_fcntl_markv) 1124 #define LFCNMARKV _FCNRW_FSPRIV('L', 3, struct lfs_fcntl_markv) 1125 #define LFCNRECLAIM _FCNO_FSPRIV('L', 4) 1126 1127 struct lfs_fhandle { 1128 char space[28]; /* FHANDLE_SIZE_COMPAT (but used from userland too) */ 1129 }; 1130 #define LFCNREWIND _FCNR_FSPRIV('L', 6, int) 1131 #define LFCNINVAL _FCNR_FSPRIV('L', 7, int) 1132 #define LFCNRESIZE _FCNR_FSPRIV('L', 8, int) 1133 #define LFCNWRAPSTOP _FCNR_FSPRIV('L', 9, int) 1134 #define LFCNWRAPGO _FCNR_FSPRIV('L', 10, int) 1135 #define LFCNIFILEFH _FCNW_FSPRIV('L', 11, struct lfs_fhandle) 1136 #define LFCNWRAPPASS _FCNR_FSPRIV('L', 12, int) 1137 # define LFS_WRAP_GOING 0x0 1138 # define LFS_WRAP_WAITING 0x1 1139 #define LFCNWRAPSTATUS _FCNW_FSPRIV('L', 13, int) 1140 1141 /* 1142 * Compat. Defined for kernel only. Userland always uses 1143 * "the one true version". 1144 */ 1145 #ifdef _KERNEL 1146 #include <compat/sys/time_types.h> 1147 1148 #define LFCNSEGWAITALL_COMPAT _FCNW_FSPRIV('L', 0, struct timeval50) 1149 #define LFCNSEGWAIT_COMPAT _FCNW_FSPRIV('L', 1, struct timeval50) 1150 #define LFCNIFILEFH_COMPAT _FCNW_FSPRIV('L', 5, struct lfs_fhandle) 1151 #define LFCNIFILEFH_COMPAT2 _FCN_FSPRIV(F_FSOUT, 'L', 11, 32) 1152 #define LFCNWRAPSTOP_COMPAT _FCNO_FSPRIV('L', 9) 1153 #define LFCNWRAPGO_COMPAT _FCNO_FSPRIV('L', 10) 1154 #define LFCNSEGWAITALL_COMPAT_50 _FCNR_FSPRIV('L', 0, struct timeval50) 1155 #define LFCNSEGWAIT_COMPAT_50 _FCNR_FSPRIV('L', 1, struct timeval50) 1156 #endif 1157 1158 #ifdef _KERNEL 1159 /* XXX MP */ 1160 #define LFS_SEGLOCK_HELD(fs) \ 1161 ((fs)->lfs_seglock != 0 && \ 1162 (fs)->lfs_lockpid == curproc->p_pid && \ 1163 (fs)->lfs_locklwp == curlwp->l_lid) 1164 #endif /* _KERNEL */ 1165 1166 /* Debug segment lock */ 1167 #ifdef notyet 1168 # define ASSERT_SEGLOCK(fs) KASSERT(LFS_SEGLOCK_HELD(fs)) 1169 # define ASSERT_NO_SEGLOCK(fs) KASSERT(!LFS_SEGLOCK_HELD(fs)) 1170 # define ASSERT_DUNNO_SEGLOCK(fs) 1171 # define ASSERT_MAYBE_SEGLOCK(fs) 1172 #else /* !notyet */ 1173 # define ASSERT_DUNNO_SEGLOCK(fs) \ 1174 DLOG((DLOG_SEG, "lfs func %s seglock wrong (%d)\n", __func__, \ 1175 LFS_SEGLOCK_HELD(fs))) 1176 # define ASSERT_SEGLOCK(fs) do { \ 1177 if (!LFS_SEGLOCK_HELD(fs)) { \ 1178 DLOG((DLOG_SEG, "lfs func %s seglock wrong (0)\n", __func__)); \ 1179 } \ 1180 } while(0) 1181 # define ASSERT_NO_SEGLOCK(fs) do { \ 1182 if (LFS_SEGLOCK_HELD(fs)) { \ 1183 DLOG((DLOG_SEG, "lfs func %s seglock wrong (1)\n", __func__)); \ 1184 } \ 1185 } while(0) 1186 # define ASSERT_MAYBE_SEGLOCK(x) 1187 #endif /* !notyet */ 1188 1189 __BEGIN_DECLS 1190 void lfs_itimes(struct inode *, const struct timespec *, 1191 const struct timespec *, const struct timespec *); 1192 __END_DECLS 1193 1194 #endif /* !_UFS_LFS_LFS_H_ */ 1195
Indexes created Tue Sep 30 17:09:57 GMT 2025