1 /* $NetBSD: lfs_segment.c,v 1.259 2015/09/01 06:08:37 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@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_segment.c 8.10 (Berkeley) 6/10/95 60 */ 61 62 #include <sys/cdefs.h> 63 __KERNEL_RCSID(0, "$NetBSD: lfs_segment.c,v 1.259 2015/09/01 06:08:37 dholland Exp $"); 64 65 #ifdef DEBUG 66 # define vndebug(vp, str) do { \ 67 if (VTOI(vp)->i_flag & IN_CLEANING) \ 68 DLOG((DLOG_WVNODE, "not writing ino %d because %s (op %d)\n", \ 69 VTOI(vp)->i_number, (str), op)); \ 70 } while(0) 71 #else 72 # define vndebug(vp, str) 73 #endif 74 #define ivndebug(vp, str) \ 75 DLOG((DLOG_WVNODE, "ino %d: %s\n", VTOI(vp)->i_number, (str))) 76 77 #if defined(_KERNEL_OPT) 78 #include "opt_ddb.h" 79 #endif 80 81 #include <sys/param.h> 82 #include <sys/systm.h> 83 #include <sys/namei.h> 84 #include <sys/kernel.h> 85 #include <sys/resourcevar.h> 86 #include <sys/file.h> 87 #include <sys/stat.h> 88 #include <sys/buf.h> 89 #include <sys/proc.h> 90 #include <sys/vnode.h> 91 #include <sys/mount.h> 92 #include <sys/kauth.h> 93 #include <sys/syslog.h> 94 95 #include <miscfs/specfs/specdev.h> 96 #include <miscfs/fifofs/fifo.h> 97 98 #include <ufs/lfs/ulfs_inode.h> 99 #include <ufs/lfs/ulfsmount.h> 100 #include <ufs/lfs/ulfs_extern.h> 101 102 #include <ufs/lfs/lfs.h> 103 #include <ufs/lfs/lfs_accessors.h> 104 #include <ufs/lfs/lfs_kernel.h> 105 #include <ufs/lfs/lfs_extern.h> 106 107 #include <uvm/uvm.h> 108 #include <uvm/uvm_extern.h> 109 110 MALLOC_JUSTDEFINE(M_SEGMENT, "LFS segment", "Segment for LFS"); 111 112 static void lfs_generic_callback(struct buf *, void (*)(struct buf *)); 113 static void lfs_free_aiodone(struct buf *); 114 static void lfs_super_aiodone(struct buf *); 115 static void lfs_cluster_aiodone(struct buf *); 116 static void lfs_cluster_callback(struct buf *); 117 118 /* 119 * Determine if it's OK to start a partial in this segment, or if we need 120 * to go on to a new segment. 121 */ 122 #define LFS_PARTIAL_FITS(fs) \ 123 (lfs_sb_getfsbpseg(fs) - \ 124 (lfs_sb_getoffset(fs) - lfs_sb_getcurseg(fs)) > \ 125 lfs_sb_getfrag(fs)) 126 127 /* 128 * Figure out whether we should do a checkpoint write or go ahead with 129 * an ordinary write. 130 */ 131 #define LFS_SHOULD_CHECKPOINT(fs, flags) \ 132 ((flags & SEGM_CLEAN) == 0 && \ 133 ((fs->lfs_nactive > LFS_MAX_ACTIVE || \ 134 (flags & SEGM_CKP) || \ 135 lfs_sb_getnclean(fs) < LFS_MAX_ACTIVE))) 136 137 int lfs_match_fake(struct lfs *, struct buf *); 138 void lfs_newseg(struct lfs *); 139 void lfs_supercallback(struct buf *); 140 void lfs_updatemeta(struct segment *); 141 void lfs_writesuper(struct lfs *, daddr_t); 142 int lfs_writevnodes(struct lfs *fs, struct mount *mp, 143 struct segment *sp, int dirops); 144 145 static void lfs_shellsort(struct lfs *, struct buf **, union lfs_blocks *, 146 int, int); 147 148 int lfs_allclean_wakeup; /* Cleaner wakeup address. */ 149 int lfs_writeindir = 1; /* whether to flush indir on non-ckp */ 150 int lfs_clean_vnhead = 0; /* Allow freeing to head of vn list */ 151 int lfs_dirvcount = 0; /* # active dirops */ 152 153 /* Statistics Counters */ 154 int lfs_dostats = 1; 155 struct lfs_stats lfs_stats; 156 157 /* op values to lfs_writevnodes */ 158 #define VN_REG 0 159 #define VN_DIROP 1 160 #define VN_EMPTY 2 161 #define VN_CLEAN 3 162 163 /* 164 * XXX KS - Set modification time on the Ifile, so the cleaner can 165 * read the fs mod time off of it. We don't set IN_UPDATE here, 166 * since we don't really need this to be flushed to disk (and in any 167 * case that wouldn't happen to the Ifile until we checkpoint). 168 */ 169 void 170 lfs_imtime(struct lfs *fs) 171 { 172 struct timespec ts; 173 struct inode *ip; 174 175 ASSERT_MAYBE_SEGLOCK(fs); 176 vfs_timestamp(&ts); 177 ip = VTOI(fs->lfs_ivnode); 178 lfs_dino_setmtime(fs, ip->i_din, ts.tv_sec); 179 lfs_dino_setmtimensec(fs, ip->i_din, ts.tv_nsec); 180 } 181 182 /* 183 * Ifile and meta data blocks are not marked busy, so segment writes MUST be 184 * single threaded. Currently, there are two paths into lfs_segwrite, sync() 185 * and getnewbuf(). They both mark the file system busy. Lfs_vflush() 186 * explicitly marks the file system busy. So lfs_segwrite is safe. I think. 187 */ 188 189 #define IS_FLUSHING(fs,vp) ((fs)->lfs_flushvp == (vp)) 190 191 int 192 lfs_vflush(struct vnode *vp) 193 { 194 struct inode *ip; 195 struct lfs *fs; 196 struct segment *sp; 197 struct buf *bp, *nbp, *tbp, *tnbp; 198 int error; 199 int flushed; 200 int relock; 201 202 ip = VTOI(vp); 203 fs = VFSTOULFS(vp->v_mount)->um_lfs; 204 relock = 0; 205 206 top: 207 KASSERT(mutex_owned(vp->v_interlock) == false); 208 KASSERT(mutex_owned(&lfs_lock) == false); 209 KASSERT(mutex_owned(&bufcache_lock) == false); 210 ASSERT_NO_SEGLOCK(fs); 211 if (ip->i_flag & IN_CLEANING) { 212 ivndebug(vp,"vflush/in_cleaning"); 213 mutex_enter(&lfs_lock); 214 LFS_CLR_UINO(ip, IN_CLEANING); 215 LFS_SET_UINO(ip, IN_MODIFIED); 216 mutex_exit(&lfs_lock); 217 218 /* 219 * Toss any cleaning buffers that have real counterparts 220 * to avoid losing new data. 221 */ 222 mutex_enter(vp->v_interlock); 223 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) { 224 nbp = LIST_NEXT(bp, b_vnbufs); 225 if (!LFS_IS_MALLOC_BUF(bp)) 226 continue; 227 /* 228 * Look for pages matching the range covered 229 * by cleaning blocks. It's okay if more dirty 230 * pages appear, so long as none disappear out 231 * from under us. 232 */ 233 if (bp->b_lblkno > 0 && vp->v_type == VREG && 234 vp != fs->lfs_ivnode) { 235 struct vm_page *pg; 236 voff_t off; 237 238 for (off = lfs_lblktosize(fs, bp->b_lblkno); 239 off < lfs_lblktosize(fs, bp->b_lblkno + 1); 240 off += PAGE_SIZE) { 241 pg = uvm_pagelookup(&vp->v_uobj, off); 242 if (pg == NULL) 243 continue; 244 if ((pg->flags & PG_CLEAN) == 0 || 245 pmap_is_modified(pg)) { 246 lfs_sb_addavail(fs, 247 lfs_btofsb(fs, 248 bp->b_bcount)); 249 wakeup(&fs->lfs_availsleep); 250 mutex_exit(vp->v_interlock); 251 lfs_freebuf(fs, bp); 252 mutex_enter(vp->v_interlock); 253 bp = NULL; 254 break; 255 } 256 } 257 } 258 for (tbp = LIST_FIRST(&vp->v_dirtyblkhd); tbp; 259 tbp = tnbp) 260 { 261 tnbp = LIST_NEXT(tbp, b_vnbufs); 262 if (tbp->b_vp == bp->b_vp 263 && tbp->b_lblkno == bp->b_lblkno 264 && tbp != bp) 265 { 266 lfs_sb_addavail(fs, lfs_btofsb(fs, 267 bp->b_bcount)); 268 wakeup(&fs->lfs_availsleep); 269 mutex_exit(vp->v_interlock); 270 lfs_freebuf(fs, bp); 271 mutex_enter(vp->v_interlock); 272 bp = NULL; 273 break; 274 } 275 } 276 } 277 } else { 278 mutex_enter(vp->v_interlock); 279 } 280 281 /* If the node is being written, wait until that is done */ 282 while (WRITEINPROG(vp)) { 283 ivndebug(vp,"vflush/writeinprog"); 284 cv_wait(&vp->v_cv, vp->v_interlock); 285 } 286 error = vdead_check(vp, VDEAD_NOWAIT); 287 mutex_exit(vp->v_interlock); 288 289 /* Protect against deadlock in vinvalbuf() */ 290 lfs_seglock(fs, SEGM_SYNC | ((error != 0) ? SEGM_RECLAIM : 0)); 291 if (error != 0) { 292 fs->lfs_reclino = ip->i_number; 293 } 294 295 /* If we're supposed to flush a freed inode, just toss it */ 296 if (ip->i_lfs_iflags & LFSI_DELETED) { 297 DLOG((DLOG_VNODE, "lfs_vflush: ino %d freed, not flushing\n", 298 ip->i_number)); 299 /* Drain v_numoutput */ 300 mutex_enter(vp->v_interlock); 301 while (vp->v_numoutput > 0) { 302 cv_wait(&vp->v_cv, vp->v_interlock); 303 } 304 KASSERT(vp->v_numoutput == 0); 305 mutex_exit(vp->v_interlock); 306 307 mutex_enter(&bufcache_lock); 308 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) { 309 nbp = LIST_NEXT(bp, b_vnbufs); 310 311 KASSERT((bp->b_flags & B_GATHERED) == 0); 312 if (bp->b_oflags & BO_DELWRI) { /* XXX always true? */ 313 lfs_sb_addavail(fs, lfs_btofsb(fs, bp->b_bcount)); 314 wakeup(&fs->lfs_availsleep); 315 } 316 /* Copied from lfs_writeseg */ 317 if (bp->b_iodone != NULL) { 318 mutex_exit(&bufcache_lock); 319 biodone(bp); 320 mutex_enter(&bufcache_lock); 321 } else { 322 bremfree(bp); 323 LFS_UNLOCK_BUF(bp); 324 mutex_enter(vp->v_interlock); 325 bp->b_flags &= ~(B_READ | B_GATHERED); 326 bp->b_oflags = (bp->b_oflags & ~BO_DELWRI) | BO_DONE; 327 bp->b_error = 0; 328 reassignbuf(bp, vp); 329 mutex_exit(vp->v_interlock); 330 brelse(bp, 0); 331 } 332 } 333 mutex_exit(&bufcache_lock); 334 LFS_CLR_UINO(ip, IN_CLEANING); 335 LFS_CLR_UINO(ip, IN_MODIFIED | IN_ACCESSED); 336 ip->i_flag &= ~IN_ALLMOD; 337 DLOG((DLOG_VNODE, "lfs_vflush: done not flushing ino %d\n", 338 ip->i_number)); 339 lfs_segunlock(fs); 340 341 KASSERT(LIST_FIRST(&vp->v_dirtyblkhd) == NULL); 342 343 return 0; 344 } 345 346 fs->lfs_flushvp = vp; 347 if (LFS_SHOULD_CHECKPOINT(fs, fs->lfs_sp->seg_flags)) { 348 error = lfs_segwrite(vp->v_mount, SEGM_CKP | SEGM_SYNC); 349 fs->lfs_flushvp = NULL; 350 KASSERT(fs->lfs_flushvp_fakevref == 0); 351 lfs_segunlock(fs); 352 353 /* Make sure that any pending buffers get written */ 354 mutex_enter(vp->v_interlock); 355 while (vp->v_numoutput > 0) { 356 cv_wait(&vp->v_cv, vp->v_interlock); 357 } 358 KASSERT(LIST_FIRST(&vp->v_dirtyblkhd) == NULL); 359 KASSERT(vp->v_numoutput == 0); 360 mutex_exit(vp->v_interlock); 361 362 return error; 363 } 364 sp = fs->lfs_sp; 365 366 flushed = 0; 367 if (VPISEMPTY(vp)) { 368 lfs_writevnodes(fs, vp->v_mount, sp, VN_EMPTY); 369 ++flushed; 370 } else if ((ip->i_flag & IN_CLEANING) && 371 (fs->lfs_sp->seg_flags & SEGM_CLEAN)) { 372 ivndebug(vp,"vflush/clean"); 373 lfs_writevnodes(fs, vp->v_mount, sp, VN_CLEAN); 374 ++flushed; 375 } else if (lfs_dostats) { 376 if (!VPISEMPTY(vp) || (VTOI(vp)->i_flag & IN_ALLMOD)) 377 ++lfs_stats.vflush_invoked; 378 ivndebug(vp,"vflush"); 379 } 380 381 #ifdef DIAGNOSTIC 382 if (vp->v_uflag & VU_DIROP) { 383 DLOG((DLOG_VNODE, "lfs_vflush: flushing VU_DIROP\n")); 384 /* panic("lfs_vflush: VU_DIROP being flushed...this can\'t happen"); */ 385 } 386 #endif 387 388 do { 389 #ifdef DEBUG 390 int loopcount = 0; 391 #endif 392 do { 393 if (LIST_FIRST(&vp->v_dirtyblkhd) != NULL) { 394 relock = lfs_writefile(fs, sp, vp); 395 if (relock && vp != fs->lfs_ivnode) { 396 /* 397 * Might have to wait for the 398 * cleaner to run; but we're 399 * still not done with this vnode. 400 * XXX we can do better than this. 401 */ 402 KDASSERT(ip->i_number != LFS_IFILE_INUM); 403 lfs_writeinode(fs, sp, ip); 404 mutex_enter(&lfs_lock); 405 LFS_SET_UINO(ip, IN_MODIFIED); 406 mutex_exit(&lfs_lock); 407 lfs_writeseg(fs, sp); 408 lfs_segunlock(fs); 409 lfs_segunlock_relock(fs); 410 goto top; 411 } 412 } 413 /* 414 * If we begin a new segment in the middle of writing 415 * the Ifile, it creates an inconsistent checkpoint, 416 * since the Ifile information for the new segment 417 * is not up-to-date. Take care of this here by 418 * sending the Ifile through again in case there 419 * are newly dirtied blocks. But wait, there's more! 420 * This second Ifile write could *also* cross a segment 421 * boundary, if the first one was large. The second 422 * one is guaranteed to be no more than 8 blocks, 423 * though (two segment blocks and supporting indirects) 424 * so the third write *will not* cross the boundary. 425 */ 426 if (vp == fs->lfs_ivnode) { 427 lfs_writefile(fs, sp, vp); 428 lfs_writefile(fs, sp, vp); 429 } 430 #ifdef DEBUG 431 if (++loopcount > 2) 432 log(LOG_NOTICE, "lfs_vflush: looping count=%d\n", loopcount); 433 #endif 434 } while (lfs_writeinode(fs, sp, ip)); 435 } while (lfs_writeseg(fs, sp) && ip->i_number == LFS_IFILE_INUM); 436 437 if (lfs_dostats) { 438 ++lfs_stats.nwrites; 439 if (sp->seg_flags & SEGM_SYNC) 440 ++lfs_stats.nsync_writes; 441 if (sp->seg_flags & SEGM_CKP) 442 ++lfs_stats.ncheckpoints; 443 } 444 /* 445 * If we were called from somewhere that has already held the seglock 446 * (e.g., lfs_markv()), the lfs_segunlock will not wait for 447 * the write to complete because we are still locked. 448 * Since lfs_vflush() must return the vnode with no dirty buffers, 449 * we must explicitly wait, if that is the case. 450 * 451 * We compare the iocount against 1, not 0, because it is 452 * artificially incremented by lfs_seglock(). 453 */ 454 mutex_enter(&lfs_lock); 455 if (fs->lfs_seglock > 1) { 456 while (fs->lfs_iocount > 1) 457 (void)mtsleep(&fs->lfs_iocount, PRIBIO + 1, 458 "lfs_vflush", 0, &lfs_lock); 459 } 460 mutex_exit(&lfs_lock); 461 462 lfs_segunlock(fs); 463 464 /* Wait for these buffers to be recovered by aiodoned */ 465 mutex_enter(vp->v_interlock); 466 while (vp->v_numoutput > 0) { 467 cv_wait(&vp->v_cv, vp->v_interlock); 468 } 469 KASSERT(LIST_FIRST(&vp->v_dirtyblkhd) == NULL); 470 KASSERT(vp->v_numoutput == 0); 471 mutex_exit(vp->v_interlock); 472 473 fs->lfs_flushvp = NULL; 474 KASSERT(fs->lfs_flushvp_fakevref == 0); 475 476 return (0); 477 } 478 479 struct lfs_writevnodes_ctx { 480 int op; 481 struct lfs *fs; 482 }; 483 static bool 484 lfs_writevnodes_selector(void *cl, struct vnode *vp) 485 { 486 struct lfs_writevnodes_ctx *c = cl; 487 struct inode *ip = VTOI(vp); 488 int op = c->op; 489 490 if (ip == NULL || vp->v_type == VNON) 491 return false; 492 if ((op == VN_DIROP && !(vp->v_uflag & VU_DIROP)) || 493 (op != VN_DIROP && op != VN_CLEAN && (vp->v_uflag & VU_DIROP))) { 494 vndebug(vp, "dirop"); 495 return false; 496 } 497 if (op == VN_EMPTY && !VPISEMPTY(vp)) { 498 vndebug(vp,"empty"); 499 return false;; 500 } 501 if (op == VN_CLEAN && ip->i_number != LFS_IFILE_INUM && 502 vp != c->fs->lfs_flushvp && !(ip->i_flag & IN_CLEANING)) { 503 vndebug(vp,"cleaning"); 504 return false; 505 } 506 mutex_enter(&lfs_lock); 507 if (vp == c->fs->lfs_unlockvp) { 508 mutex_exit(&lfs_lock); 509 return false; 510 } 511 mutex_exit(&lfs_lock); 512 513 return true; 514 } 515 516 int 517 lfs_writevnodes(struct lfs *fs, struct mount *mp, struct segment *sp, int op) 518 { 519 struct inode *ip; 520 struct vnode *vp; 521 struct vnode_iterator *marker; 522 struct lfs_writevnodes_ctx ctx; 523 int inodes_written = 0; 524 int error = 0; 525 526 /* 527 * XXX This was TAILQ_FOREACH_REVERSE on &mp->mnt_vnodelist. 528 * XXX The rationale is unclear, the initial commit had no information. 529 * XXX If the order really matters we have to sort the vnodes first. 530 */ 531 532 ASSERT_SEGLOCK(fs); 533 vfs_vnode_iterator_init(mp, &marker); 534 ctx.op = op; 535 ctx.fs = fs; 536 while ((vp = vfs_vnode_iterator_next(marker, 537 lfs_writevnodes_selector, &ctx)) != NULL) { 538 ip = VTOI(vp); 539 540 /* 541 * Write the inode/file if dirty and it's not the IFILE. 542 */ 543 if (((ip->i_flag & IN_ALLMOD) || !VPISEMPTY(vp)) && 544 ip->i_number != LFS_IFILE_INUM) { 545 error = lfs_writefile(fs, sp, vp); 546 if (error) { 547 vrele(vp); 548 if (error == EAGAIN) { 549 /* 550 * This error from lfs_putpages 551 * indicates we need to drop 552 * the segment lock and start 553 * over after the cleaner has 554 * had a chance to run. 555 */ 556 lfs_writeinode(fs, sp, ip); 557 lfs_writeseg(fs, sp); 558 if (!VPISEMPTY(vp) && 559 !WRITEINPROG(vp) && 560 !(ip->i_flag & IN_ALLMOD)) { 561 mutex_enter(&lfs_lock); 562 LFS_SET_UINO(ip, IN_MODIFIED); 563 mutex_exit(&lfs_lock); 564 } 565 break; 566 } 567 error = 0; /* XXX not quite right */ 568 continue; 569 } 570 571 if (!VPISEMPTY(vp)) { 572 if (WRITEINPROG(vp)) { 573 ivndebug(vp,"writevnodes/write2"); 574 } else if (!(ip->i_flag & IN_ALLMOD)) { 575 mutex_enter(&lfs_lock); 576 LFS_SET_UINO(ip, IN_MODIFIED); 577 mutex_exit(&lfs_lock); 578 } 579 } 580 (void) lfs_writeinode(fs, sp, ip); 581 inodes_written++; 582 } 583 vrele(vp); 584 } 585 vfs_vnode_iterator_destroy(marker); 586 return error; 587 } 588 589 /* 590 * Do a checkpoint. 591 */ 592 int 593 lfs_segwrite(struct mount *mp, int flags) 594 { 595 struct buf *bp; 596 struct inode *ip; 597 struct lfs *fs; 598 struct segment *sp; 599 struct vnode *vp; 600 SEGUSE *segusep; 601 int do_ckp, did_ckp, error; 602 unsigned n, segleft, maxseg, sn, i, curseg; 603 int writer_set = 0; 604 int dirty; 605 int redo; 606 SEGSUM *ssp; 607 int um_error; 608 609 fs = VFSTOULFS(mp)->um_lfs; 610 ASSERT_MAYBE_SEGLOCK(fs); 611 612 if (fs->lfs_ronly) 613 return EROFS; 614 615 lfs_imtime(fs); 616 617 /* 618 * Allocate a segment structure and enough space to hold pointers to 619 * the maximum possible number of buffers which can be described in a 620 * single summary block. 621 */ 622 do_ckp = LFS_SHOULD_CHECKPOINT(fs, flags); 623 624 /* We can't do a partial write and checkpoint at the same time. */ 625 if (do_ckp) 626 flags &= ~SEGM_SINGLE; 627 628 lfs_seglock(fs, flags | (do_ckp ? SEGM_CKP : 0)); 629 sp = fs->lfs_sp; 630 if (sp->seg_flags & (SEGM_CLEAN | SEGM_CKP)) 631 do_ckp = 1; 632 633 /* 634 * If lfs_flushvp is non-NULL, we are called from lfs_vflush, 635 * in which case we have to flush *all* buffers off of this vnode. 636 * We don't care about other nodes, but write any non-dirop nodes 637 * anyway in anticipation of another getnewvnode(). 638 * 639 * If we're cleaning we only write cleaning and ifile blocks, and 640 * no dirops, since otherwise we'd risk corruption in a crash. 641 */ 642 if (sp->seg_flags & SEGM_CLEAN) 643 lfs_writevnodes(fs, mp, sp, VN_CLEAN); 644 else if (!(sp->seg_flags & SEGM_FORCE_CKP)) { 645 do { 646 um_error = lfs_writevnodes(fs, mp, sp, VN_REG); 647 if ((sp->seg_flags & SEGM_SINGLE) && 648 lfs_sb_getcurseg(fs) != fs->lfs_startseg) { 649 DLOG((DLOG_SEG, "lfs_segwrite: breaking out of segment write at daddr 0x%jx\n", (uintmax_t)lfs_sb_getoffset(fs))); 650 break; 651 } 652 653 if (do_ckp || fs->lfs_dirops == 0) { 654 if (!writer_set) { 655 lfs_writer_enter(fs, "lfs writer"); 656 writer_set = 1; 657 } 658 error = lfs_writevnodes(fs, mp, sp, VN_DIROP); 659 if (um_error == 0) 660 um_error = error; 661 /* In case writevnodes errored out */ 662 lfs_flush_dirops(fs); 663 ssp = (SEGSUM *)(sp->segsum); 664 lfs_ss_setflags(fs, ssp, 665 lfs_ss_getflags(fs, ssp) & ~(SS_CONT)); 666 lfs_finalize_fs_seguse(fs); 667 } 668 if (do_ckp && um_error) { 669 lfs_segunlock_relock(fs); 670 sp = fs->lfs_sp; 671 } 672 } while (do_ckp && um_error != 0); 673 } 674 675 /* 676 * If we are doing a checkpoint, mark everything since the 677 * last checkpoint as no longer ACTIVE. 678 */ 679 if (do_ckp || fs->lfs_doifile) { 680 segleft = lfs_sb_getnseg(fs); 681 curseg = 0; 682 for (n = 0; n < lfs_sb_getsegtabsz(fs); n++) { 683 dirty = 0; 684 if (bread(fs->lfs_ivnode, lfs_sb_getcleansz(fs) + n, 685 lfs_sb_getbsize(fs), B_MODIFY, &bp)) 686 panic("lfs_segwrite: ifile read"); 687 segusep = (SEGUSE *)bp->b_data; 688 maxseg = min(segleft, lfs_sb_getsepb(fs)); 689 for (i = 0; i < maxseg; i++) { 690 sn = curseg + i; 691 if (sn != lfs_dtosn(fs, lfs_sb_getcurseg(fs)) && 692 segusep->su_flags & SEGUSE_ACTIVE) { 693 segusep->su_flags &= ~SEGUSE_ACTIVE; 694 --fs->lfs_nactive; 695 ++dirty; 696 } 697 fs->lfs_suflags[fs->lfs_activesb][sn] = 698 segusep->su_flags; 699 if (lfs_sb_getversion(fs) > 1) 700 ++segusep; 701 else 702 segusep = (SEGUSE *) 703 ((SEGUSE_V1 *)segusep + 1); 704 } 705 706 if (dirty) 707 error = LFS_BWRITE_LOG(bp); /* Ifile */ 708 else 709 brelse(bp, 0); 710 segleft -= lfs_sb_getsepb(fs); 711 curseg += lfs_sb_getsepb(fs); 712 } 713 } 714 715 KASSERT(LFS_SEGLOCK_HELD(fs)); 716 717 did_ckp = 0; 718 if (do_ckp || fs->lfs_doifile) { 719 vp = fs->lfs_ivnode; 720 #ifdef DEBUG 721 int loopcount = 0; 722 #endif 723 do { 724 #ifdef DEBUG 725 LFS_ENTER_LOG("pretend", __FILE__, __LINE__, 0, 0, curproc->p_pid); 726 #endif 727 mutex_enter(&lfs_lock); 728 fs->lfs_flags &= ~LFS_IFDIRTY; 729 mutex_exit(&lfs_lock); 730 731 ip = VTOI(vp); 732 733 if (LIST_FIRST(&vp->v_dirtyblkhd) != NULL) { 734 /* 735 * Ifile has no pages, so we don't need 736 * to check error return here. 737 */ 738 lfs_writefile(fs, sp, vp); 739 /* 740 * Ensure the Ifile takes the current segment 741 * into account. See comment in lfs_vflush. 742 */ 743 lfs_writefile(fs, sp, vp); 744 lfs_writefile(fs, sp, vp); 745 } 746 747 if (ip->i_flag & IN_ALLMOD) 748 ++did_ckp; 749 #if 0 750 redo = (do_ckp ? lfs_writeinode(fs, sp, ip) : 0); 751 #else 752 redo = lfs_writeinode(fs, sp, ip); 753 #endif 754 redo += lfs_writeseg(fs, sp); 755 mutex_enter(&lfs_lock); 756 redo += (fs->lfs_flags & LFS_IFDIRTY); 757 mutex_exit(&lfs_lock); 758 #ifdef DEBUG 759 if (++loopcount > 2) 760 log(LOG_NOTICE, "lfs_segwrite: looping count=%d\n", 761 loopcount); 762 #endif 763 } while (redo && do_ckp); 764 765 /* 766 * Unless we are unmounting, the Ifile may continue to have 767 * dirty blocks even after a checkpoint, due to changes to 768 * inodes' atime. If we're checkpointing, it's "impossible" 769 * for other parts of the Ifile to be dirty after the loop 770 * above, since we hold the segment lock. 771 */ 772 mutex_enter(vp->v_interlock); 773 if (LIST_EMPTY(&vp->v_dirtyblkhd)) { 774 LFS_CLR_UINO(ip, IN_ALLMOD); 775 } 776 #ifdef DIAGNOSTIC 777 else if (do_ckp) { 778 int do_panic = 0; 779 LIST_FOREACH(bp, &vp->v_dirtyblkhd, b_vnbufs) { 780 if (bp->b_lblkno < lfs_sb_getcleansz(fs) + 781 lfs_sb_getsegtabsz(fs) && 782 !(bp->b_flags & B_GATHERED)) { 783 printf("ifile lbn %ld still dirty (flags %lx)\n", 784 (long)bp->b_lblkno, 785 (long)bp->b_flags); 786 ++do_panic; 787 } 788 } 789 if (do_panic) 790 panic("dirty blocks"); 791 } 792 #endif 793 mutex_exit(vp->v_interlock); 794 } else { 795 (void) lfs_writeseg(fs, sp); 796 } 797 798 /* Note Ifile no longer needs to be written */ 799 fs->lfs_doifile = 0; 800 if (writer_set) 801 lfs_writer_leave(fs); 802 803 /* 804 * If we didn't write the Ifile, we didn't really do anything. 805 * That means that (1) there is a checkpoint on disk and (2) 806 * nothing has changed since it was written. 807 * 808 * Take the flags off of the segment so that lfs_segunlock 809 * doesn't have to write the superblock either. 810 */ 811 if (do_ckp && !did_ckp) { 812 sp->seg_flags &= ~SEGM_CKP; 813 } 814 815 if (lfs_dostats) { 816 ++lfs_stats.nwrites; 817 if (sp->seg_flags & SEGM_SYNC) 818 ++lfs_stats.nsync_writes; 819 if (sp->seg_flags & SEGM_CKP) 820 ++lfs_stats.ncheckpoints; 821 } 822 lfs_segunlock(fs); 823 return (0); 824 } 825 826 /* 827 * Write the dirty blocks associated with a vnode. 828 */ 829 int 830 lfs_writefile(struct lfs *fs, struct segment *sp, struct vnode *vp) 831 { 832 struct inode *ip; 833 int i, frag; 834 SEGSUM *ssp; 835 int error; 836 837 ASSERT_SEGLOCK(fs); 838 error = 0; 839 ip = VTOI(vp); 840 841 lfs_acquire_finfo(fs, ip->i_number, ip->i_gen); 842 843 if (vp->v_uflag & VU_DIROP) { 844 ssp = (SEGSUM *)sp->segsum; 845 lfs_ss_setflags(fs, ssp, 846 lfs_ss_getflags(fs, ssp) | (SS_DIROP|SS_CONT)); 847 } 848 849 if (sp->seg_flags & SEGM_CLEAN) { 850 lfs_gather(fs, sp, vp, lfs_match_fake); 851 /* 852 * For a file being flushed, we need to write *all* blocks. 853 * This means writing the cleaning blocks first, and then 854 * immediately following with any non-cleaning blocks. 855 * The same is true of the Ifile since checkpoints assume 856 * that all valid Ifile blocks are written. 857 */ 858 if (IS_FLUSHING(fs, vp) || vp == fs->lfs_ivnode) { 859 lfs_gather(fs, sp, vp, lfs_match_data); 860 /* 861 * Don't call VOP_PUTPAGES: if we're flushing, 862 * we've already done it, and the Ifile doesn't 863 * use the page cache. 864 */ 865 } 866 } else { 867 lfs_gather(fs, sp, vp, lfs_match_data); 868 /* 869 * If we're flushing, we've already called VOP_PUTPAGES 870 * so don't do it again. Otherwise, we want to write 871 * everything we've got. 872 */ 873 if (!IS_FLUSHING(fs, vp)) { 874 mutex_enter(vp->v_interlock); 875 error = VOP_PUTPAGES(vp, 0, 0, 876 PGO_CLEANIT | PGO_ALLPAGES | PGO_LOCKED); 877 } 878 } 879 880 /* 881 * It may not be necessary to write the meta-data blocks at this point, 882 * as the roll-forward recovery code should be able to reconstruct the 883 * list. 884 * 885 * We have to write them anyway, though, under two conditions: (1) the 886 * vnode is being flushed (for reuse by vinvalbuf); or (2) we are 887 * checkpointing. 888 * 889 * BUT if we are cleaning, we might have indirect blocks that refer to 890 * new blocks not being written yet, in addition to fragments being 891 * moved out of a cleaned segment. If that is the case, don't 892 * write the indirect blocks, or the finfo will have a small block 893 * in the middle of it! 894 * XXX in this case isn't the inode size wrong too? 895 */ 896 frag = 0; 897 if (sp->seg_flags & SEGM_CLEAN) { 898 for (i = 0; i < ULFS_NDADDR; i++) 899 if (ip->i_lfs_fragsize[i] > 0 && 900 ip->i_lfs_fragsize[i] < lfs_sb_getbsize(fs)) 901 ++frag; 902 } 903 #ifdef DIAGNOSTIC 904 if (frag > 1) 905 panic("lfs_writefile: more than one fragment!"); 906 #endif 907 if (IS_FLUSHING(fs, vp) || 908 (frag == 0 && (lfs_writeindir || (sp->seg_flags & SEGM_CKP)))) { 909 lfs_gather(fs, sp, vp, lfs_match_indir); 910 lfs_gather(fs, sp, vp, lfs_match_dindir); 911 lfs_gather(fs, sp, vp, lfs_match_tindir); 912 } 913 lfs_release_finfo(fs); 914 915 return error; 916 } 917 918 /* 919 * Update segment accounting to reflect this inode's change of address. 920 */ 921 static int 922 lfs_update_iaddr(struct lfs *fs, struct segment *sp, struct inode *ip, daddr_t ndaddr) 923 { 924 struct buf *bp; 925 daddr_t daddr; 926 IFILE *ifp; 927 SEGUSE *sup; 928 ino_t ino; 929 int redo_ifile; 930 u_int32_t sn; 931 932 redo_ifile = 0; 933 934 /* 935 * If updating the ifile, update the super-block. Update the disk 936 * address and access times for this inode in the ifile. 937 */ 938 ino = ip->i_number; 939 if (ino == LFS_IFILE_INUM) { 940 daddr = lfs_sb_getidaddr(fs); 941 lfs_sb_setidaddr(fs, LFS_DBTOFSB(fs, ndaddr)); 942 } else { 943 LFS_IENTRY(ifp, fs, ino, bp); 944 daddr = lfs_if_getdaddr(fs, ifp); 945 lfs_if_setdaddr(fs, ifp, LFS_DBTOFSB(fs, ndaddr)); 946 (void)LFS_BWRITE_LOG(bp); /* Ifile */ 947 } 948 949 /* 950 * If this is the Ifile and lfs_offset is set to the first block 951 * in the segment, dirty the new segment's accounting block 952 * (XXX should already be dirty?) and tell the caller to do it again. 953 */ 954 if (ip->i_number == LFS_IFILE_INUM) { 955 sn = lfs_dtosn(fs, lfs_sb_getoffset(fs)); 956 if (lfs_sntod(fs, sn) + lfs_btofsb(fs, lfs_sb_getsumsize(fs)) == 957 lfs_sb_getoffset(fs)) { 958 LFS_SEGENTRY(sup, fs, sn, bp); 959 KASSERT(bp->b_oflags & BO_DELWRI); 960 LFS_WRITESEGENTRY(sup, fs, sn, bp); 961 /* fs->lfs_flags |= LFS_IFDIRTY; */ 962 redo_ifile |= 1; 963 } 964 } 965 966 /* 967 * The inode's last address should not be in the current partial 968 * segment, except under exceptional circumstances (lfs_writevnodes 969 * had to start over, and in the meantime more blocks were written 970 * to a vnode). Both inodes will be accounted to this segment 971 * in lfs_writeseg so we need to subtract the earlier version 972 * here anyway. The segment count can temporarily dip below 973 * zero here; keep track of how many duplicates we have in 974 * "dupino" so we don't panic below. 975 */ 976 if (daddr >= lfs_sb_getlastpseg(fs) && daddr <= lfs_sb_getoffset(fs)) { 977 ++sp->ndupino; 978 DLOG((DLOG_SEG, "lfs_writeinode: last inode addr in current pseg " 979 "(ino %d daddr 0x%llx) ndupino=%d\n", ino, 980 (long long)daddr, sp->ndupino)); 981 } 982 /* 983 * Account the inode: it no longer belongs to its former segment, 984 * though it will not belong to the new segment until that segment 985 * is actually written. 986 */ 987 if (daddr != LFS_UNUSED_DADDR) { 988 u_int32_t oldsn = lfs_dtosn(fs, daddr); 989 #ifdef DIAGNOSTIC 990 int ndupino = (sp->seg_number == oldsn) ? sp->ndupino : 0; 991 #endif 992 LFS_SEGENTRY(sup, fs, oldsn, bp); 993 #ifdef DIAGNOSTIC 994 if (sup->su_nbytes + DINOSIZE(fs) * ndupino < DINOSIZE(fs)) { 995 printf("lfs_writeinode: negative bytes " 996 "(segment %" PRIu32 " short by %d, " 997 "oldsn=%" PRIu32 ", cursn=%" PRIu32 998 ", daddr=%" PRId64 ", su_nbytes=%u, " 999 "ndupino=%d)\n", 1000 lfs_dtosn(fs, daddr), 1001 (int)DINOSIZE(fs) * 1002 (1 - sp->ndupino) - sup->su_nbytes, 1003 oldsn, sp->seg_number, daddr, 1004 (unsigned int)sup->su_nbytes, 1005 sp->ndupino); 1006 panic("lfs_writeinode: negative bytes"); 1007 sup->su_nbytes = DINOSIZE(fs); 1008 } 1009 #endif 1010 DLOG((DLOG_SU, "seg %d -= %d for ino %d inode\n", 1011 lfs_dtosn(fs, daddr), DINOSIZE(fs), ino)); 1012 sup->su_nbytes -= DINOSIZE(fs); 1013 redo_ifile |= 1014 (ino == LFS_IFILE_INUM && !(bp->b_flags & B_GATHERED)); 1015 if (redo_ifile) { 1016 mutex_enter(&lfs_lock); 1017 fs->lfs_flags |= LFS_IFDIRTY; 1018 mutex_exit(&lfs_lock); 1019 /* Don't double-account */ 1020 lfs_sb_setidaddr(fs, 0x0); 1021 } 1022 LFS_WRITESEGENTRY(sup, fs, oldsn, bp); /* Ifile */ 1023 } 1024 1025 return redo_ifile; 1026 } 1027 1028 int 1029 lfs_writeinode(struct lfs *fs, struct segment *sp, struct inode *ip) 1030 { 1031 struct buf *bp; 1032 union lfs_dinode *cdp; 1033 struct vnode *vp = ITOV(ip); 1034 daddr_t daddr; 1035 int i, ndx; 1036 int redo_ifile = 0; 1037 int gotblk = 0; 1038 int count; 1039 SEGSUM *ssp; 1040 1041 ASSERT_SEGLOCK(fs); 1042 if (!(ip->i_flag & IN_ALLMOD) && !(vp->v_uflag & VU_DIROP)) 1043 return (0); 1044 1045 /* Can't write ifile when writer is not set */ 1046 KASSERT(ip->i_number != LFS_IFILE_INUM || fs->lfs_writer > 0 || 1047 (sp->seg_flags & SEGM_CLEAN)); 1048 1049 /* 1050 * If this is the Ifile, see if writing it here will generate a 1051 * temporary misaccounting. If it will, do the accounting and write 1052 * the blocks, postponing the inode write until the accounting is 1053 * solid. 1054 */ 1055 count = 0; 1056 while (vp == fs->lfs_ivnode) { 1057 int redo = 0; 1058 1059 if (sp->idp == NULL && sp->ibp == NULL && 1060 (sp->seg_bytes_left < lfs_sb_getibsize(fs) || 1061 sp->sum_bytes_left < sizeof(int32_t))) { 1062 (void) lfs_writeseg(fs, sp); 1063 continue; 1064 } 1065 1066 /* Look for dirty Ifile blocks */ 1067 LIST_FOREACH(bp, &fs->lfs_ivnode->v_dirtyblkhd, b_vnbufs) { 1068 if (!(bp->b_flags & B_GATHERED)) { 1069 redo = 1; 1070 break; 1071 } 1072 } 1073 1074 if (redo == 0) 1075 redo = lfs_update_iaddr(fs, sp, ip, 0x0); 1076 if (redo == 0) 1077 break; 1078 1079 if (sp->idp) { 1080 lfs_dino_setinumber(fs, sp->idp, 0); 1081 sp->idp = NULL; 1082 } 1083 ++count; 1084 if (count > 2) 1085 log(LOG_NOTICE, "lfs_writeinode: looping count=%d\n", count); 1086 lfs_writefile(fs, sp, fs->lfs_ivnode); 1087 } 1088 1089 /* Allocate a new inode block if necessary. */ 1090 if ((ip->i_number != LFS_IFILE_INUM || sp->idp == NULL) && 1091 sp->ibp == NULL) { 1092 /* Allocate a new segment if necessary. */ 1093 if (sp->seg_bytes_left < lfs_sb_getibsize(fs) || 1094 sp->sum_bytes_left < sizeof(int32_t)) 1095 (void) lfs_writeseg(fs, sp); 1096 1097 /* Get next inode block. */ 1098 daddr = lfs_sb_getoffset(fs); 1099 lfs_sb_addoffset(fs, lfs_btofsb(fs, lfs_sb_getibsize(fs))); 1100 sp->ibp = *sp->cbpp++ = 1101 getblk(VTOI(fs->lfs_ivnode)->i_devvp, 1102 LFS_FSBTODB(fs, daddr), lfs_sb_getibsize(fs), 0, 0); 1103 gotblk++; 1104 1105 /* Zero out inode numbers */ 1106 for (i = 0; i < LFS_INOPB(fs); ++i) { 1107 union lfs_dinode *tmpdi; 1108 1109 tmpdi = (union lfs_dinode *)((char *)sp->ibp->b_data + 1110 DINOSIZE(fs) * i); 1111 lfs_dino_setinumber(fs, tmpdi, 0); 1112 } 1113 1114 ++sp->start_bpp; 1115 lfs_sb_subavail(fs, lfs_btofsb(fs, lfs_sb_getibsize(fs))); 1116 /* Set remaining space counters. */ 1117 sp->seg_bytes_left -= lfs_sb_getibsize(fs); 1118 sp->sum_bytes_left -= sizeof(int32_t); 1119 ndx = lfs_sb_getsumsize(fs) / sizeof(int32_t) - 1120 sp->ninodes / LFS_INOPB(fs) - 1; 1121 /* XXX ondisk32 */ 1122 ((int32_t *)(sp->segsum))[ndx] = daddr; 1123 } 1124 1125 /* Check VU_DIROP in case there is a new file with no data blocks */ 1126 if (vp->v_uflag & VU_DIROP) { 1127 ssp = (SEGSUM *)sp->segsum; 1128 lfs_ss_setflags(fs, ssp, 1129 lfs_ss_getflags(fs, ssp) | (SS_DIROP|SS_CONT)); 1130 } 1131 1132 /* Update the inode times and copy the inode onto the inode page. */ 1133 /* XXX kludge --- don't redirty the ifile just to put times on it */ 1134 if (ip->i_number != LFS_IFILE_INUM) 1135 LFS_ITIMES(ip, NULL, NULL, NULL); 1136 1137 /* 1138 * If this is the Ifile, and we've already written the Ifile in this 1139 * partial segment, just overwrite it (it's not on disk yet) and 1140 * continue. 1141 * 1142 * XXX we know that the bp that we get the second time around has 1143 * already been gathered. 1144 */ 1145 if (ip->i_number == LFS_IFILE_INUM && sp->idp) { 1146 lfs_copy_dinode(fs, sp->idp, ip->i_din); 1147 ip->i_lfs_osize = ip->i_size; 1148 return 0; 1149 } 1150 1151 bp = sp->ibp; 1152 cdp = DINO_IN_BLOCK(fs, bp->b_data, sp->ninodes % LFS_INOPB(fs)); 1153 lfs_copy_dinode(fs, cdp, ip->i_din); 1154 1155 /* 1156 * This inode is on its way to disk; clear its VU_DIROP status when 1157 * the write is complete. 1158 */ 1159 if (vp->v_uflag & VU_DIROP) { 1160 if (!(sp->seg_flags & SEGM_CLEAN)) 1161 ip->i_flag |= IN_CDIROP; 1162 else { 1163 DLOG((DLOG_DIROP, "lfs_writeinode: not clearing dirop for cleaned ino %d\n", (int)ip->i_number)); 1164 } 1165 } 1166 1167 /* 1168 * If cleaning, link counts and directory file sizes cannot change, 1169 * since those would be directory operations---even if the file 1170 * we are writing is marked VU_DIROP we should write the old values. 1171 * If we're not cleaning, of course, update the values so we get 1172 * current values the next time we clean. 1173 */ 1174 if (sp->seg_flags & SEGM_CLEAN) { 1175 if (vp->v_uflag & VU_DIROP) { 1176 lfs_dino_setnlink(fs, cdp, ip->i_lfs_odnlink); 1177 /* if (vp->v_type == VDIR) */ 1178 lfs_dino_setsize(fs, cdp, ip->i_lfs_osize); 1179 } 1180 } else { 1181 ip->i_lfs_odnlink = lfs_dino_getnlink(fs, cdp); 1182 ip->i_lfs_osize = ip->i_size; 1183 } 1184 1185 1186 /* We can finish the segment accounting for truncations now */ 1187 lfs_finalize_ino_seguse(fs, ip); 1188 1189 /* 1190 * If we are cleaning, ensure that we don't write UNWRITTEN disk 1191 * addresses to disk; possibly change the on-disk record of 1192 * the inode size, either by reverting to the previous size 1193 * (in the case of cleaning) or by verifying the inode's block 1194 * holdings (in the case of files being allocated as they are being 1195 * written). 1196 * XXX By not writing UNWRITTEN blocks, we are making the lfs_avail 1197 * XXX count on disk wrong by the same amount. We should be 1198 * XXX able to "borrow" from lfs_avail and return it after the 1199 * XXX Ifile is written. See also in lfs_writeseg. 1200 */ 1201 1202 /* Check file size based on highest allocated block */ 1203 if (((lfs_dino_getmode(fs, ip->i_din) & LFS_IFMT) == LFS_IFREG || 1204 (lfs_dino_getmode(fs, ip->i_din) & LFS_IFMT) == LFS_IFDIR) && 1205 ip->i_size > ((ip->i_lfs_hiblk + 1) << lfs_sb_getbshift(fs))) { 1206 lfs_dino_setsize(fs, cdp, (ip->i_lfs_hiblk + 1) << lfs_sb_getbshift(fs)); 1207 DLOG((DLOG_SEG, "lfs_writeinode: ino %d size %" PRId64 " -> %" 1208 PRId64 "\n", (int)ip->i_number, ip->i_size, lfs_dino_getsize(fs, cdp))); 1209 } 1210 if (ip->i_lfs_effnblks != lfs_dino_getblocks(fs, ip->i_din)) { 1211 DLOG((DLOG_SEG, "lfs_writeinode: cleansing ino %d eff %jd != nblk %d)" 1212 " at %jx\n", ip->i_number, (intmax_t)ip->i_lfs_effnblks, 1213 lfs_dino_getblocks(fs, ip->i_din), (uintmax_t)lfs_sb_getoffset(fs))); 1214 for (i=0; i<ULFS_NDADDR; i++) { 1215 if (lfs_dino_getdb(fs, cdp, i) == UNWRITTEN) { 1216 DLOG((DLOG_SEG, "lfs_writeinode: wiping UNWRITTEN\n")); 1217 lfs_dino_setdb(fs, cdp, i, 0); 1218 } 1219 } 1220 for (i=0; i<ULFS_NIADDR; i++) { 1221 if (lfs_dino_getib(fs, cdp, i) == UNWRITTEN) { 1222 DLOG((DLOG_SEG, "lfs_writeinode: wiping UNWRITTEN\n")); 1223 lfs_dino_setib(fs, cdp, i, 0); 1224 } 1225 } 1226 } 1227 1228 #ifdef DIAGNOSTIC 1229 /* 1230 * Check dinode held blocks against dinode size. 1231 * This should be identical to the check in lfs_vget(). 1232 */ 1233 for (i = (lfs_dino_getsize(fs, cdp) + lfs_sb_getbsize(fs) - 1) >> lfs_sb_getbshift(fs); 1234 i < ULFS_NDADDR; i++) { 1235 KASSERT(i >= 0); 1236 if ((lfs_dino_getmode(fs, cdp) & LFS_IFMT) == LFS_IFLNK) 1237 continue; 1238 if (((lfs_dino_getmode(fs, cdp) & LFS_IFMT) == LFS_IFBLK || 1239 (lfs_dino_getmode(fs, cdp) & LFS_IFMT) == LFS_IFCHR) && i == 0) 1240 continue; 1241 if (lfs_dino_getdb(fs, cdp, i) != 0) { 1242 # ifdef DEBUG 1243 lfs_dump_dinode(fs, cdp); 1244 # endif 1245 panic("writing inconsistent inode"); 1246 } 1247 } 1248 #endif /* DIAGNOSTIC */ 1249 1250 if (ip->i_flag & IN_CLEANING) 1251 LFS_CLR_UINO(ip, IN_CLEANING); 1252 else { 1253 /* XXX IN_ALLMOD */ 1254 LFS_CLR_UINO(ip, IN_ACCESSED | IN_ACCESS | IN_CHANGE | 1255 IN_UPDATE | IN_MODIFY); 1256 if (ip->i_lfs_effnblks == lfs_dino_getblocks(fs, ip->i_din)) 1257 LFS_CLR_UINO(ip, IN_MODIFIED); 1258 else { 1259 DLOG((DLOG_VNODE, "lfs_writeinode: ino %d: real " 1260 "blks=%d, eff=%jd\n", ip->i_number, 1261 lfs_dino_getblocks(fs, ip->i_din), (intmax_t)ip->i_lfs_effnblks)); 1262 } 1263 } 1264 1265 if (ip->i_number == LFS_IFILE_INUM) { 1266 /* We know sp->idp == NULL */ 1267 sp->idp = DINO_IN_BLOCK(fs, bp, sp->ninodes % LFS_INOPB(fs)); 1268 1269 /* Not dirty any more */ 1270 mutex_enter(&lfs_lock); 1271 fs->lfs_flags &= ~LFS_IFDIRTY; 1272 mutex_exit(&lfs_lock); 1273 } 1274 1275 if (gotblk) { 1276 mutex_enter(&bufcache_lock); 1277 LFS_LOCK_BUF(bp); 1278 brelsel(bp, 0); 1279 mutex_exit(&bufcache_lock); 1280 } 1281 1282 /* Increment inode count in segment summary block. */ 1283 1284 ssp = (SEGSUM *)sp->segsum; 1285 lfs_ss_setninos(fs, ssp, lfs_ss_getninos(fs, ssp) + 1); 1286 1287 /* If this page is full, set flag to allocate a new page. */ 1288 if (++sp->ninodes % LFS_INOPB(fs) == 0) 1289 sp->ibp = NULL; 1290 1291 redo_ifile = lfs_update_iaddr(fs, sp, ip, bp->b_blkno); 1292 1293 KASSERT(redo_ifile == 0); 1294 return (redo_ifile); 1295 } 1296 1297 int 1298 lfs_gatherblock(struct segment *sp, struct buf *bp, kmutex_t *mptr) 1299 { 1300 struct lfs *fs; 1301 int vers; 1302 int j, blksinblk; 1303 1304 ASSERT_SEGLOCK(sp->fs); 1305 /* 1306 * If full, finish this segment. We may be doing I/O, so 1307 * release and reacquire the splbio(). 1308 */ 1309 #ifdef DIAGNOSTIC 1310 if (sp->vp == NULL) 1311 panic ("lfs_gatherblock: Null vp in segment"); 1312 #endif 1313 fs = sp->fs; 1314 blksinblk = howmany(bp->b_bcount, lfs_sb_getbsize(fs)); 1315 if (sp->sum_bytes_left < sizeof(int32_t) * blksinblk || 1316 sp->seg_bytes_left < bp->b_bcount) { 1317 if (mptr) 1318 mutex_exit(mptr); 1319 lfs_updatemeta(sp); 1320 1321 vers = lfs_fi_getversion(fs, sp->fip); 1322 (void) lfs_writeseg(fs, sp); 1323 1324 /* Add the current file to the segment summary. */ 1325 lfs_acquire_finfo(fs, VTOI(sp->vp)->i_number, vers); 1326 1327 if (mptr) 1328 mutex_enter(mptr); 1329 return (1); 1330 } 1331 1332 if (bp->b_flags & B_GATHERED) { 1333 DLOG((DLOG_SEG, "lfs_gatherblock: already gathered! Ino %ju," 1334 " lbn %" PRId64 "\n", 1335 (uintmax_t)lfs_fi_getino(fs, sp->fip), bp->b_lblkno)); 1336 return (0); 1337 } 1338 1339 /* Insert into the buffer list, update the FINFO block. */ 1340 bp->b_flags |= B_GATHERED; 1341 1342 *sp->cbpp++ = bp; 1343 for (j = 0; j < blksinblk; j++) { 1344 unsigned bn; 1345 1346 bn = lfs_fi_getnblocks(fs, sp->fip); 1347 lfs_fi_setnblocks(fs, sp->fip, bn+1); 1348 lfs_fi_setblock(fs, sp->fip, bn, bp->b_lblkno + j); 1349 /* This block's accounting moves from lfs_favail to lfs_avail */ 1350 lfs_deregister_block(sp->vp, bp->b_lblkno + j); 1351 } 1352 1353 sp->sum_bytes_left -= sizeof(int32_t) * blksinblk; 1354 sp->seg_bytes_left -= bp->b_bcount; 1355 return (0); 1356 } 1357 1358 int 1359 lfs_gather(struct lfs *fs, struct segment *sp, struct vnode *vp, 1360 int (*match)(struct lfs *, struct buf *)) 1361 { 1362 struct buf *bp, *nbp; 1363 int count = 0; 1364 1365 ASSERT_SEGLOCK(fs); 1366 if (vp->v_type == VBLK) 1367 return 0; 1368 KASSERT(sp->vp == NULL); 1369 sp->vp = vp; 1370 mutex_enter(&bufcache_lock); 1371 1372 #ifndef LFS_NO_BACKBUF_HACK 1373 /* This is a hack to see if ordering the blocks in LFS makes a difference. */ 1374 # define BUF_OFFSET \ 1375 (((char *)&LIST_NEXT(bp, b_vnbufs)) - (char *)bp) 1376 # define BACK_BUF(BP) \ 1377 ((struct buf *)(((char *)(BP)->b_vnbufs.le_prev) - BUF_OFFSET)) 1378 # define BEG_OF_LIST \ 1379 ((struct buf *)(((char *)&LIST_FIRST(&vp->v_dirtyblkhd)) - BUF_OFFSET)) 1380 1381 loop: 1382 /* Find last buffer. */ 1383 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); 1384 bp && LIST_NEXT(bp, b_vnbufs) != NULL; 1385 bp = LIST_NEXT(bp, b_vnbufs)) 1386 /* nothing */; 1387 for (; bp && bp != BEG_OF_LIST; bp = nbp) { 1388 nbp = BACK_BUF(bp); 1389 #else /* LFS_NO_BACKBUF_HACK */ 1390 loop: 1391 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) { 1392 nbp = LIST_NEXT(bp, b_vnbufs); 1393 #endif /* LFS_NO_BACKBUF_HACK */ 1394 if ((bp->b_cflags & BC_BUSY) != 0 || 1395 (bp->b_flags & B_GATHERED) != 0 || !match(fs, bp)) { 1396 #ifdef DEBUG 1397 if (vp == fs->lfs_ivnode && 1398 (bp->b_cflags & BC_BUSY) != 0 && 1399 (bp->b_flags & B_GATHERED) == 0) 1400 log(LOG_NOTICE, "lfs_gather: ifile lbn %" 1401 PRId64 " busy (%x) at 0x%jx", 1402 bp->b_lblkno, bp->b_flags, 1403 (uintmax_t)lfs_sb_getoffset(fs)); 1404 #endif 1405 continue; 1406 } 1407 #ifdef DIAGNOSTIC 1408 # ifdef LFS_USE_B_INVAL 1409 if ((bp->b_flags & BC_INVAL) != 0 && bp->b_iodone == NULL) { 1410 DLOG((DLOG_SEG, "lfs_gather: lbn %" PRId64 1411 " is BC_INVAL\n", bp->b_lblkno)); 1412 VOP_PRINT(bp->b_vp); 1413 } 1414 # endif /* LFS_USE_B_INVAL */ 1415 if (!(bp->b_oflags & BO_DELWRI)) 1416 panic("lfs_gather: bp not BO_DELWRI"); 1417 if (!(bp->b_flags & B_LOCKED)) { 1418 DLOG((DLOG_SEG, "lfs_gather: lbn %" PRId64 1419 " blk %" PRId64 " not B_LOCKED\n", 1420 bp->b_lblkno, 1421 LFS_DBTOFSB(fs, bp->b_blkno))); 1422 VOP_PRINT(bp->b_vp); 1423 panic("lfs_gather: bp not B_LOCKED"); 1424 } 1425 #endif 1426 if (lfs_gatherblock(sp, bp, &bufcache_lock)) { 1427 goto loop; 1428 } 1429 count++; 1430 } 1431 mutex_exit(&bufcache_lock); 1432 lfs_updatemeta(sp); 1433 KASSERT(sp->vp == vp); 1434 sp->vp = NULL; 1435 return count; 1436 } 1437 1438 #if DEBUG 1439 # define DEBUG_OOFF(n) do { \ 1440 if (ooff == 0) { \ 1441 DLOG((DLOG_SEG, "lfs_updatemeta[%d]: warning: writing " \ 1442 "ino %d lbn %" PRId64 " at 0x%" PRIx32 \ 1443 ", was 0x0 (or %" PRId64 ")\n", \ 1444 (n), ip->i_number, lbn, ndaddr, daddr)); \ 1445 } \ 1446 } while (0) 1447 #else 1448 # define DEBUG_OOFF(n) 1449 #endif 1450 1451 /* 1452 * Change the given block's address to ndaddr, finding its previous 1453 * location using ulfs_bmaparray(). 1454 * 1455 * Account for this change in the segment table. 1456 * 1457 * called with sp == NULL by roll-forwarding code. 1458 */ 1459 void 1460 lfs_update_single(struct lfs *fs, struct segment *sp, 1461 struct vnode *vp, daddr_t lbn, daddr_t ndaddr, int size) 1462 { 1463 SEGUSE *sup; 1464 struct buf *bp; 1465 struct indir a[ULFS_NIADDR + 2], *ap; 1466 struct inode *ip; 1467 daddr_t daddr, ooff; 1468 int num, error; 1469 int bb, osize, obb; 1470 1471 ASSERT_SEGLOCK(fs); 1472 KASSERT(sp == NULL || sp->vp == vp); 1473 ip = VTOI(vp); 1474 1475 error = ulfs_bmaparray(vp, lbn, &daddr, a, &num, NULL, NULL); 1476 if (error) 1477 panic("lfs_updatemeta: ulfs_bmaparray returned %d", error); 1478 1479 KASSERT(daddr <= LFS_MAX_DADDR(fs)); 1480 if (daddr > 0) 1481 daddr = LFS_DBTOFSB(fs, daddr); 1482 1483 bb = lfs_numfrags(fs, size); 1484 switch (num) { 1485 case 0: 1486 ooff = lfs_dino_getdb(fs, ip->i_din, lbn); 1487 DEBUG_OOFF(0); 1488 if (ooff == UNWRITTEN) 1489 lfs_dino_setblocks(fs, ip->i_din, 1490 lfs_dino_getblocks(fs, ip->i_din) + bb); 1491 else { 1492 /* possible fragment truncation or extension */ 1493 obb = lfs_btofsb(fs, ip->i_lfs_fragsize[lbn]); 1494 lfs_dino_setblocks(fs, ip->i_din, 1495 lfs_dino_getblocks(fs, ip->i_din) + (bb-obb)); 1496 } 1497 lfs_dino_setdb(fs, ip->i_din, lbn, ndaddr); 1498 break; 1499 case 1: 1500 ooff = lfs_dino_getib(fs, ip->i_din, a[0].in_off); 1501 DEBUG_OOFF(1); 1502 if (ooff == UNWRITTEN) 1503 lfs_dino_setblocks(fs, ip->i_din, 1504 lfs_dino_getblocks(fs, ip->i_din) + bb); 1505 lfs_dino_setib(fs, ip->i_din, a[0].in_off, ndaddr); 1506 break; 1507 default: 1508 ap = &a[num - 1]; 1509 if (bread(vp, ap->in_lbn, lfs_sb_getbsize(fs), 1510 B_MODIFY, &bp)) 1511 panic("lfs_updatemeta: bread bno %" PRId64, 1512 ap->in_lbn); 1513 1514 ooff = lfs_iblock_get(fs, bp->b_data, ap->in_off); 1515 DEBUG_OOFF(num); 1516 if (ooff == UNWRITTEN) 1517 lfs_dino_setblocks(fs, ip->i_din, 1518 lfs_dino_getblocks(fs, ip->i_din) + bb); 1519 lfs_iblock_set(fs, bp->b_data, ap->in_off, ndaddr); 1520 (void) VOP_BWRITE(bp->b_vp, bp); 1521 } 1522 1523 KASSERT(ooff == 0 || ooff == UNWRITTEN || ooff == daddr); 1524 1525 /* Update hiblk when extending the file */ 1526 if (lbn > ip->i_lfs_hiblk) 1527 ip->i_lfs_hiblk = lbn; 1528 1529 /* 1530 * Though we'd rather it couldn't, this *can* happen right now 1531 * if cleaning blocks and regular blocks coexist. 1532 */ 1533 /* KASSERT(daddr < fs->lfs_lastpseg || daddr > ndaddr); */ 1534 1535 /* 1536 * Update segment usage information, based on old size 1537 * and location. 1538 */ 1539 if (daddr > 0) { 1540 u_int32_t oldsn = lfs_dtosn(fs, daddr); 1541 #ifdef DIAGNOSTIC 1542 int ndupino; 1543 1544 if (sp && sp->seg_number == oldsn) { 1545 ndupino = sp->ndupino; 1546 } else { 1547 ndupino = 0; 1548 } 1549 #endif 1550 KASSERT(oldsn < lfs_sb_getnseg(fs)); 1551 if (lbn >= 0 && lbn < ULFS_NDADDR) 1552 osize = ip->i_lfs_fragsize[lbn]; 1553 else 1554 osize = lfs_sb_getbsize(fs); 1555 LFS_SEGENTRY(sup, fs, oldsn, bp); 1556 #ifdef DIAGNOSTIC 1557 if (sup->su_nbytes + DINOSIZE(fs) * ndupino < osize) { 1558 printf("lfs_updatemeta: negative bytes " 1559 "(segment %" PRIu32 " short by %" PRId64 1560 ")\n", lfs_dtosn(fs, daddr), 1561 (int64_t)osize - 1562 (DINOSIZE(fs) * ndupino + sup->su_nbytes)); 1563 printf("lfs_updatemeta: ino %llu, lbn %" PRId64 1564 ", addr = 0x%" PRIx64 "\n", 1565 (unsigned long long)ip->i_number, lbn, daddr); 1566 printf("lfs_updatemeta: ndupino=%d\n", ndupino); 1567 panic("lfs_updatemeta: negative bytes"); 1568 sup->su_nbytes = osize - 1569 DINOSIZE(fs) * ndupino; 1570 } 1571 #endif 1572 DLOG((DLOG_SU, "seg %" PRIu32 " -= %d for ino %d lbn %" PRId64 1573 " db 0x%" PRIx64 "\n", 1574 lfs_dtosn(fs, daddr), osize, 1575 ip->i_number, lbn, daddr)); 1576 sup->su_nbytes -= osize; 1577 if (!(bp->b_flags & B_GATHERED)) { 1578 mutex_enter(&lfs_lock); 1579 fs->lfs_flags |= LFS_IFDIRTY; 1580 mutex_exit(&lfs_lock); 1581 } 1582 LFS_WRITESEGENTRY(sup, fs, oldsn, bp); 1583 } 1584 /* 1585 * Now that this block has a new address, and its old 1586 * segment no longer owns it, we can forget about its 1587 * old size. 1588 */ 1589 if (lbn >= 0 && lbn < ULFS_NDADDR) 1590 ip->i_lfs_fragsize[lbn] = size; 1591 } 1592 1593 /* 1594 * Update the metadata that points to the blocks listed in the FINFO 1595 * array. 1596 */ 1597 void 1598 lfs_updatemeta(struct segment *sp) 1599 { 1600 struct buf *sbp; 1601 struct lfs *fs; 1602 struct vnode *vp; 1603 daddr_t lbn; 1604 int i, nblocks, num; 1605 int __diagused nblocks_orig; 1606 int bb; 1607 int bytesleft, size; 1608 unsigned lastlength; 1609 union lfs_blocks tmpptr; 1610 1611 fs = sp->fs; 1612 vp = sp->vp; 1613 ASSERT_SEGLOCK(fs); 1614 1615 /* 1616 * This used to be: 1617 * 1618 * nblocks = &sp->fip->fi_blocks[sp->fip->fi_nblocks] - sp->start_lbp; 1619 * 1620 * that is, it allowed for the possibility that start_lbp did 1621 * not point to the beginning of the finfo block pointer area. 1622 * This particular formulation is six kinds of painful in the 1623 * lfs64 world where we have two sizes of block pointer, so 1624 * unless/until everything can be cleaned up to not move 1625 * start_lbp around but instead use an offset, we do the 1626 * following: 1627 * 1. Get NEXT_FINFO(sp->fip). This is the same pointer as 1628 * &sp->fip->fi_blocks[sp->fip->fi_nblocks], just the wrong 1629 * type. (Ugh.) 1630 * 2. Cast it to void *, then assign it to a temporary 1631 * union lfs_blocks. 1632 * 3. Subtract start_lbp from that. 1633 * 4. Save the value of nblocks in blocks_orig so we can 1634 * assert below that it hasn't changed without repeating this 1635 * rubbish. 1636 * 1637 * XXX. 1638 */ 1639 lfs_blocks_fromvoid(fs, &tmpptr, (void *)NEXT_FINFO(fs, sp->fip)); 1640 nblocks = lfs_blocks_sub(fs, &tmpptr, &sp->start_lbp); 1641 nblocks_orig = nblocks; 1642 1643 KASSERT(nblocks >= 0); 1644 KASSERT(vp != NULL); 1645 if (nblocks == 0) 1646 return; 1647 1648 /* 1649 * This count may be high due to oversize blocks from lfs_gop_write. 1650 * Correct for this. (XXX we should be able to keep track of these.) 1651 */ 1652 for (i = 0; i < nblocks; i++) { 1653 if (sp->start_bpp[i] == NULL) { 1654 DLOG((DLOG_SEG, "lfs_updatemeta: nblocks = %d, not %d\n", i, nblocks)); 1655 nblocks = i; 1656 break; 1657 } 1658 num = howmany(sp->start_bpp[i]->b_bcount, lfs_sb_getbsize(fs)); 1659 KASSERT(sp->start_bpp[i]->b_lblkno >= 0 || num == 1); 1660 nblocks -= num - 1; 1661 } 1662 1663 #if 0 1664 /* pre-lfs64 assertion */ 1665 KASSERT(vp->v_type == VREG || 1666 nblocks == &sp->fip->fi_blocks[sp->fip->fi_nblocks] - sp->start_lbp); 1667 #else 1668 KASSERT(vp->v_type == VREG || nblocks == nblocks_orig); 1669 #endif 1670 KASSERT(nblocks == sp->cbpp - sp->start_bpp); 1671 1672 /* 1673 * Sort the blocks. 1674 * 1675 * We have to sort even if the blocks come from the 1676 * cleaner, because there might be other pending blocks on the 1677 * same inode...and if we don't sort, and there are fragments 1678 * present, blocks may be written in the wrong place. 1679 */ 1680 lfs_shellsort(fs, sp->start_bpp, &sp->start_lbp, nblocks, lfs_sb_getbsize(fs)); 1681 1682 /* 1683 * Record the length of the last block in case it's a fragment. 1684 * If there are indirect blocks present, they sort last. An 1685 * indirect block will be lfs_bsize and its presence indicates 1686 * that you cannot have fragments. 1687 * 1688 * XXX This last is a lie. A cleaned fragment can coexist with 1689 * XXX a later indirect block. This will continue to be 1690 * XXX true until lfs_markv is fixed to do everything with 1691 * XXX fake blocks (including fake inodes and fake indirect blocks). 1692 */ 1693 lastlength = ((sp->start_bpp[nblocks - 1]->b_bcount - 1) & 1694 lfs_sb_getbmask(fs)) + 1; 1695 lfs_fi_setlastlength(fs, sp->fip, lastlength); 1696 1697 /* 1698 * Assign disk addresses, and update references to the logical 1699 * block and the segment usage information. 1700 */ 1701 for (i = nblocks; i--; ++sp->start_bpp) { 1702 sbp = *sp->start_bpp; 1703 lbn = lfs_blocks_get(fs, &sp->start_lbp, 0); 1704 KASSERT(sbp->b_lblkno == lbn); 1705 1706 sbp->b_blkno = LFS_FSBTODB(fs, lfs_sb_getoffset(fs)); 1707 1708 /* 1709 * If we write a frag in the wrong place, the cleaner won't 1710 * be able to correctly identify its size later, and the 1711 * segment will be uncleanable. (Even worse, it will assume 1712 * that the indirect block that actually ends the list 1713 * is of a smaller size!) 1714 */ 1715 if ((sbp->b_bcount & lfs_sb_getbmask(fs)) && i != 0) 1716 panic("lfs_updatemeta: fragment is not last block"); 1717 1718 /* 1719 * For each subblock in this possibly oversized block, 1720 * update its address on disk. 1721 */ 1722 KASSERT(lbn >= 0 || sbp->b_bcount == lfs_sb_getbsize(fs)); 1723 KASSERT(vp == sbp->b_vp); 1724 for (bytesleft = sbp->b_bcount; bytesleft > 0; 1725 bytesleft -= lfs_sb_getbsize(fs)) { 1726 size = MIN(bytesleft, lfs_sb_getbsize(fs)); 1727 bb = lfs_numfrags(fs, size); 1728 lbn = lfs_blocks_get(fs, &sp->start_lbp, 0); 1729 lfs_blocks_inc(fs, &sp->start_lbp); 1730 lfs_update_single(fs, sp, sp->vp, lbn, lfs_sb_getoffset(fs), 1731 size); 1732 lfs_sb_addoffset(fs, bb); 1733 } 1734 1735 } 1736 1737 /* This inode has been modified */ 1738 LFS_SET_UINO(VTOI(vp), IN_MODIFIED); 1739 } 1740 1741 /* 1742 * Move lfs_offset to a segment earlier than newsn. 1743 */ 1744 int 1745 lfs_rewind(struct lfs *fs, int newsn) 1746 { 1747 int sn, osn, isdirty; 1748 struct buf *bp; 1749 SEGUSE *sup; 1750 1751 ASSERT_SEGLOCK(fs); 1752 1753 osn = lfs_dtosn(fs, lfs_sb_getoffset(fs)); 1754 if (osn < newsn) 1755 return 0; 1756 1757 /* lfs_avail eats the remaining space in this segment */ 1758 lfs_sb_subavail(fs, lfs_sb_getfsbpseg(fs) - (lfs_sb_getoffset(fs) - lfs_sb_getcurseg(fs))); 1759 1760 /* Find a low-numbered segment */ 1761 for (sn = 0; sn < lfs_sb_getnseg(fs); ++sn) { 1762 LFS_SEGENTRY(sup, fs, sn, bp); 1763 isdirty = sup->su_flags & SEGUSE_DIRTY; 1764 brelse(bp, 0); 1765 1766 if (!isdirty) 1767 break; 1768 } 1769 if (sn == lfs_sb_getnseg(fs)) 1770 panic("lfs_rewind: no clean segments"); 1771 if (newsn >= 0 && sn >= newsn) 1772 return ENOENT; 1773 lfs_sb_setnextseg(fs, lfs_sntod(fs, sn)); 1774 lfs_newseg(fs); 1775 lfs_sb_setoffset(fs, lfs_sb_getcurseg(fs)); 1776 1777 return 0; 1778 } 1779 1780 /* 1781 * Start a new partial segment. 1782 * 1783 * Return 1 when we entered to a new segment. 1784 * Otherwise, return 0. 1785 */ 1786 int 1787 lfs_initseg(struct lfs *fs) 1788 { 1789 struct segment *sp = fs->lfs_sp; 1790 SEGSUM *ssp; 1791 struct buf *sbp; /* buffer for SEGSUM */ 1792 int repeat = 0; /* return value */ 1793 1794 ASSERT_SEGLOCK(fs); 1795 /* Advance to the next segment. */ 1796 if (!LFS_PARTIAL_FITS(fs)) { 1797 SEGUSE *sup; 1798 struct buf *bp; 1799 1800 /* lfs_avail eats the remaining space */ 1801 lfs_sb_subavail(fs, lfs_sb_getfsbpseg(fs) - (lfs_sb_getoffset(fs) - 1802 lfs_sb_getcurseg(fs))); 1803 /* Wake up any cleaning procs waiting on this file system. */ 1804 lfs_wakeup_cleaner(fs); 1805 lfs_newseg(fs); 1806 repeat = 1; 1807 lfs_sb_setoffset(fs, lfs_sb_getcurseg(fs)); 1808 1809 sp->seg_number = lfs_dtosn(fs, lfs_sb_getcurseg(fs)); 1810 sp->seg_bytes_left = lfs_fsbtob(fs, lfs_sb_getfsbpseg(fs)); 1811 1812 /* 1813 * If the segment contains a superblock, update the offset 1814 * and summary address to skip over it. 1815 */ 1816 LFS_SEGENTRY(sup, fs, sp->seg_number, bp); 1817 if (sup->su_flags & SEGUSE_SUPERBLOCK) { 1818 lfs_sb_addoffset(fs, lfs_btofsb(fs, LFS_SBPAD)); 1819 sp->seg_bytes_left -= LFS_SBPAD; 1820 } 1821 brelse(bp, 0); 1822 /* Segment zero could also contain the labelpad */ 1823 if (lfs_sb_getversion(fs) > 1 && sp->seg_number == 0 && 1824 lfs_sb_gets0addr(fs) < lfs_btofsb(fs, LFS_LABELPAD)) { 1825 lfs_sb_addoffset(fs, 1826 lfs_btofsb(fs, LFS_LABELPAD) - lfs_sb_gets0addr(fs)); 1827 sp->seg_bytes_left -= 1828 LFS_LABELPAD - lfs_fsbtob(fs, lfs_sb_gets0addr(fs)); 1829 } 1830 } else { 1831 sp->seg_number = lfs_dtosn(fs, lfs_sb_getcurseg(fs)); 1832 sp->seg_bytes_left = lfs_fsbtob(fs, lfs_sb_getfsbpseg(fs) - 1833 (lfs_sb_getoffset(fs) - lfs_sb_getcurseg(fs))); 1834 } 1835 lfs_sb_setlastpseg(fs, lfs_sb_getoffset(fs)); 1836 1837 /* Record first address of this partial segment */ 1838 if (sp->seg_flags & SEGM_CLEAN) { 1839 fs->lfs_cleanint[fs->lfs_cleanind] = lfs_sb_getoffset(fs); 1840 if (++fs->lfs_cleanind >= LFS_MAX_CLEANIND) { 1841 /* "1" is the artificial inc in lfs_seglock */ 1842 mutex_enter(&lfs_lock); 1843 while (fs->lfs_iocount > 1) { 1844 mtsleep(&fs->lfs_iocount, PRIBIO + 1, 1845 "lfs_initseg", 0, &lfs_lock); 1846 } 1847 mutex_exit(&lfs_lock); 1848 fs->lfs_cleanind = 0; 1849 } 1850 } 1851 1852 sp->fs = fs; 1853 sp->ibp = NULL; 1854 sp->idp = NULL; 1855 sp->ninodes = 0; 1856 sp->ndupino = 0; 1857 1858 sp->cbpp = sp->bpp; 1859 1860 /* Get a new buffer for SEGSUM */ 1861 sbp = lfs_newbuf(fs, VTOI(fs->lfs_ivnode)->i_devvp, 1862 LFS_FSBTODB(fs, lfs_sb_getoffset(fs)), lfs_sb_getsumsize(fs), LFS_NB_SUMMARY); 1863 1864 /* ... and enter it into the buffer list. */ 1865 *sp->cbpp = sbp; 1866 sp->cbpp++; 1867 lfs_sb_addoffset(fs, lfs_btofsb(fs, lfs_sb_getsumsize(fs))); 1868 1869 sp->start_bpp = sp->cbpp; 1870 1871 /* Set point to SEGSUM, initialize it. */ 1872 ssp = sp->segsum = sbp->b_data; 1873 memset(ssp, 0, lfs_sb_getsumsize(fs)); 1874 lfs_ss_setnext(fs, ssp, lfs_sb_getnextseg(fs)); 1875 lfs_ss_setnfinfo(fs, ssp, 0); 1876 lfs_ss_setninos(fs, ssp, 0); 1877 lfs_ss_setmagic(fs, ssp, SS_MAGIC); 1878 1879 /* Set pointer to first FINFO, initialize it. */ 1880 sp->fip = SEGSUM_FINFOBASE(fs, sp->segsum); 1881 lfs_fi_setnblocks(fs, sp->fip, 0); 1882 lfs_fi_setlastlength(fs, sp->fip, 0); 1883 lfs_blocks_fromfinfo(fs, &sp->start_lbp, sp->fip); 1884 1885 sp->seg_bytes_left -= lfs_sb_getsumsize(fs); 1886 sp->sum_bytes_left = lfs_sb_getsumsize(fs) - SEGSUM_SIZE(fs); 1887 1888 return (repeat); 1889 } 1890 1891 /* 1892 * Remove SEGUSE_INVAL from all segments. 1893 */ 1894 void 1895 lfs_unset_inval_all(struct lfs *fs) 1896 { 1897 SEGUSE *sup; 1898 struct buf *bp; 1899 int i; 1900 1901 for (i = 0; i < lfs_sb_getnseg(fs); i++) { 1902 LFS_SEGENTRY(sup, fs, i, bp); 1903 if (sup->su_flags & SEGUSE_INVAL) { 1904 sup->su_flags &= ~SEGUSE_INVAL; 1905 LFS_WRITESEGENTRY(sup, fs, i, bp); 1906 } else 1907 brelse(bp, 0); 1908 } 1909 } 1910 1911 /* 1912 * Return the next segment to write. 1913 */ 1914 void 1915 lfs_newseg(struct lfs *fs) 1916 { 1917 CLEANERINFO *cip; 1918 SEGUSE *sup; 1919 struct buf *bp; 1920 int curseg, isdirty, sn, skip_inval; 1921 1922 ASSERT_SEGLOCK(fs); 1923 1924 /* Honor LFCNWRAPSTOP */ 1925 mutex_enter(&lfs_lock); 1926 while (lfs_sb_getnextseg(fs) < lfs_sb_getcurseg(fs) && fs->lfs_nowrap) { 1927 if (fs->lfs_wrappass) { 1928 log(LOG_NOTICE, "%s: wrappass=%d\n", 1929 lfs_sb_getfsmnt(fs), fs->lfs_wrappass); 1930 fs->lfs_wrappass = 0; 1931 break; 1932 } 1933 fs->lfs_wrapstatus = LFS_WRAP_WAITING; 1934 wakeup(&fs->lfs_nowrap); 1935 log(LOG_NOTICE, "%s: waiting at log wrap\n", lfs_sb_getfsmnt(fs)); 1936 mtsleep(&fs->lfs_wrappass, PVFS, "newseg", 10 * hz, 1937 &lfs_lock); 1938 } 1939 fs->lfs_wrapstatus = LFS_WRAP_GOING; 1940 mutex_exit(&lfs_lock); 1941 1942 LFS_SEGENTRY(sup, fs, lfs_dtosn(fs, lfs_sb_getnextseg(fs)), bp); 1943 DLOG((DLOG_SU, "lfs_newseg: seg %d := 0 in newseg\n", 1944 lfs_dtosn(fs, lfs_sb_getnextseg(fs)))); 1945 sup->su_flags |= SEGUSE_DIRTY | SEGUSE_ACTIVE; 1946 sup->su_nbytes = 0; 1947 sup->su_nsums = 0; 1948 sup->su_ninos = 0; 1949 LFS_WRITESEGENTRY(sup, fs, lfs_dtosn(fs, lfs_sb_getnextseg(fs)), bp); 1950 1951 LFS_CLEANERINFO(cip, fs, bp); 1952 lfs_ci_shiftcleantodirty(fs, cip, 1); 1953 lfs_sb_setnclean(fs, lfs_ci_getclean(fs, cip)); 1954 LFS_SYNC_CLEANERINFO(cip, fs, bp, 1); 1955 1956 lfs_sb_setlastseg(fs, lfs_sb_getcurseg(fs)); 1957 lfs_sb_setcurseg(fs, lfs_sb_getnextseg(fs)); 1958 skip_inval = 1; 1959 for (sn = curseg = lfs_dtosn(fs, lfs_sb_getcurseg(fs)) + lfs_sb_getinterleave(fs);;) { 1960 sn = (sn + 1) % lfs_sb_getnseg(fs); 1961 1962 if (sn == curseg) { 1963 if (skip_inval) 1964 skip_inval = 0; 1965 else 1966 panic("lfs_nextseg: no clean segments"); 1967 } 1968 LFS_SEGENTRY(sup, fs, sn, bp); 1969 isdirty = sup->su_flags & (SEGUSE_DIRTY | (skip_inval ? SEGUSE_INVAL : 0)); 1970 /* Check SEGUSE_EMPTY as we go along */ 1971 if (isdirty && sup->su_nbytes == 0 && 1972 !(sup->su_flags & SEGUSE_EMPTY)) 1973 LFS_WRITESEGENTRY(sup, fs, sn, bp); 1974 else 1975 brelse(bp, 0); 1976 1977 if (!isdirty) 1978 break; 1979 } 1980 if (skip_inval == 0) 1981 lfs_unset_inval_all(fs); 1982 1983 ++fs->lfs_nactive; 1984 lfs_sb_setnextseg(fs, lfs_sntod(fs, sn)); 1985 if (lfs_dostats) { 1986 ++lfs_stats.segsused; 1987 } 1988 } 1989 1990 static struct buf * 1991 lfs_newclusterbuf(struct lfs *fs, struct vnode *vp, daddr_t addr, 1992 int n) 1993 { 1994 struct lfs_cluster *cl; 1995 struct buf **bpp, *bp; 1996 1997 ASSERT_SEGLOCK(fs); 1998 cl = (struct lfs_cluster *)pool_get(&fs->lfs_clpool, PR_WAITOK); 1999 bpp = (struct buf **)pool_get(&fs->lfs_bpppool, PR_WAITOK); 2000 memset(cl, 0, sizeof(*cl)); 2001 cl->fs = fs; 2002 cl->bpp = bpp; 2003 cl->bufcount = 0; 2004 cl->bufsize = 0; 2005 2006 /* If this segment is being written synchronously, note that */ 2007 if (fs->lfs_sp->seg_flags & SEGM_SYNC) { 2008 cl->flags |= LFS_CL_SYNC; 2009 cl->seg = fs->lfs_sp; 2010 ++cl->seg->seg_iocount; 2011 } 2012 2013 /* Get an empty buffer header, or maybe one with something on it */ 2014 bp = getiobuf(vp, true); 2015 bp->b_dev = NODEV; 2016 bp->b_blkno = bp->b_lblkno = addr; 2017 bp->b_iodone = lfs_cluster_callback; 2018 bp->b_private = cl; 2019 2020 return bp; 2021 } 2022 2023 int 2024 lfs_writeseg(struct lfs *fs, struct segment *sp) 2025 { 2026 struct buf **bpp, *bp, *cbp, *newbp, *unbusybp; 2027 SEGUSE *sup; 2028 SEGSUM *ssp; 2029 int i; 2030 int do_again, nblocks, byteoffset; 2031 size_t el_size; 2032 struct lfs_cluster *cl; 2033 u_short ninos; 2034 struct vnode *devvp; 2035 char *p = NULL; 2036 struct vnode *vp; 2037 int32_t *daddrp; /* XXX ondisk32 */ 2038 int changed; 2039 u_int32_t sum; 2040 size_t sumstart; 2041 #ifdef DEBUG 2042 FINFO *fip; 2043 int findex; 2044 #endif 2045 2046 ASSERT_SEGLOCK(fs); 2047 2048 ssp = (SEGSUM *)sp->segsum; 2049 2050 /* 2051 * If there are no buffers other than the segment summary to write, 2052 * don't do anything. If we are the end of a dirop sequence, however, 2053 * write the empty segment summary anyway, to help out the 2054 * roll-forward agent. 2055 */ 2056 if ((nblocks = sp->cbpp - sp->bpp) == 1) { 2057 if ((lfs_ss_getflags(fs, ssp) & (SS_DIROP | SS_CONT)) != SS_DIROP) 2058 return 0; 2059 } 2060 2061 /* Note if partial segment is being written by the cleaner */ 2062 if (sp->seg_flags & SEGM_CLEAN) 2063 lfs_ss_setflags(fs, ssp, lfs_ss_getflags(fs, ssp) | SS_CLEAN); 2064 2065 /* Note if we are writing to reclaim */ 2066 if (sp->seg_flags & SEGM_RECLAIM) { 2067 lfs_ss_setflags(fs, ssp, lfs_ss_getflags(fs, ssp) | SS_RECLAIM); 2068 lfs_ss_setreclino(fs, ssp, fs->lfs_reclino); 2069 } 2070 2071 devvp = VTOI(fs->lfs_ivnode)->i_devvp; 2072 2073 /* Update the segment usage information. */ 2074 LFS_SEGENTRY(sup, fs, sp->seg_number, bp); 2075 2076 /* Loop through all blocks, except the segment summary. */ 2077 for (bpp = sp->bpp; ++bpp < sp->cbpp; ) { 2078 if ((*bpp)->b_vp != devvp) { 2079 sup->su_nbytes += (*bpp)->b_bcount; 2080 DLOG((DLOG_SU, "seg %" PRIu32 " += %ld for ino %d" 2081 " lbn %" PRId64 " db 0x%" PRIx64 "\n", 2082 sp->seg_number, (*bpp)->b_bcount, 2083 VTOI((*bpp)->b_vp)->i_number, (*bpp)->b_lblkno, 2084 (*bpp)->b_blkno)); 2085 } 2086 } 2087 2088 #ifdef DEBUG 2089 /* Check for zero-length and zero-version FINFO entries. */ 2090 fip = SEGSUM_FINFOBASE(fs, ssp); 2091 for (findex = 0; findex < lfs_ss_getnfinfo(fs, ssp); findex++) { 2092 KDASSERT(lfs_fi_getnblocks(fs, fip) > 0); 2093 KDASSERT(lfs_fi_getversion(fs, fip) > 0); 2094 fip = NEXT_FINFO(fs, fip); 2095 } 2096 #endif /* DEBUG */ 2097 2098 ninos = (lfs_ss_getninos(fs, ssp) + LFS_INOPB(fs) - 1) / LFS_INOPB(fs); 2099 DLOG((DLOG_SU, "seg %d += %d for %d inodes\n", 2100 sp->seg_number, 2101 lfs_ss_getninos(fs, ssp) * DINOSIZE(fs), 2102 lfs_ss_getninos(fs, ssp))); 2103 sup->su_nbytes += lfs_ss_getninos(fs, ssp) * DINOSIZE(fs); 2104 /* sup->su_nbytes += lfs_sb_getsumsize(fs); */ 2105 if (lfs_sb_getversion(fs) == 1) 2106 sup->su_olastmod = time_second; 2107 else 2108 sup->su_lastmod = time_second; 2109 sup->su_ninos += ninos; 2110 ++sup->su_nsums; 2111 lfs_sb_subavail(fs, lfs_btofsb(fs, lfs_sb_getsumsize(fs))); 2112 2113 do_again = !(bp->b_flags & B_GATHERED); 2114 LFS_WRITESEGENTRY(sup, fs, sp->seg_number, bp); /* Ifile */ 2115 2116 /* 2117 * Mark blocks B_BUSY, to prevent then from being changed between 2118 * the checksum computation and the actual write. 2119 * 2120 * If we are cleaning, check indirect blocks for UNWRITTEN, and if 2121 * there are any, replace them with copies that have UNASSIGNED 2122 * instead. 2123 */ 2124 mutex_enter(&bufcache_lock); 2125 for (bpp = sp->bpp, i = nblocks - 1; i--;) { 2126 ++bpp; 2127 bp = *bpp; 2128 if (bp->b_iodone != NULL) { /* UBC or malloced buffer */ 2129 bp->b_cflags |= BC_BUSY; 2130 continue; 2131 } 2132 2133 while (bp->b_cflags & BC_BUSY) { 2134 DLOG((DLOG_SEG, "lfs_writeseg: avoiding potential" 2135 " data summary corruption for ino %d, lbn %" 2136 PRId64 "\n", 2137 VTOI(bp->b_vp)->i_number, bp->b_lblkno)); 2138 bp->b_cflags |= BC_WANTED; 2139 cv_wait(&bp->b_busy, &bufcache_lock); 2140 } 2141 bp->b_cflags |= BC_BUSY; 2142 mutex_exit(&bufcache_lock); 2143 unbusybp = NULL; 2144 2145 /* 2146 * Check and replace indirect block UNWRITTEN bogosity. 2147 * XXX See comment in lfs_writefile. 2148 */ 2149 if (bp->b_lblkno < 0 && bp->b_vp != devvp && bp->b_vp && 2150 lfs_dino_getblocks(fs, VTOI(bp->b_vp)->i_din) != 2151 VTOI(bp->b_vp)->i_lfs_effnblks) { 2152 DLOG((DLOG_VNODE, "lfs_writeseg: cleansing ino %d (%jd != %d)\n", 2153 VTOI(bp->b_vp)->i_number, 2154 (intmax_t)VTOI(bp->b_vp)->i_lfs_effnblks, 2155 lfs_dino_getblocks(fs, VTOI(bp->b_vp)->i_din))); 2156 /* Make a copy we'll make changes to */ 2157 newbp = lfs_newbuf(fs, bp->b_vp, bp->b_lblkno, 2158 bp->b_bcount, LFS_NB_IBLOCK); 2159 newbp->b_blkno = bp->b_blkno; 2160 memcpy(newbp->b_data, bp->b_data, 2161 newbp->b_bcount); 2162 2163 changed = 0; 2164 /* XXX ondisk32 */ 2165 for (daddrp = (int32_t *)(newbp->b_data); 2166 daddrp < (int32_t *)((char *)newbp->b_data + 2167 newbp->b_bcount); daddrp++) { 2168 if (*daddrp == UNWRITTEN) { 2169 ++changed; 2170 *daddrp = 0; 2171 } 2172 } 2173 /* 2174 * Get rid of the old buffer. Don't mark it clean, 2175 * though, if it still has dirty data on it. 2176 */ 2177 if (changed) { 2178 DLOG((DLOG_SEG, "lfs_writeseg: replacing UNWRITTEN(%d):" 2179 " bp = %p newbp = %p\n", changed, bp, 2180 newbp)); 2181 *bpp = newbp; 2182 bp->b_flags &= ~B_GATHERED; 2183 bp->b_error = 0; 2184 if (bp->b_iodone != NULL) { 2185 DLOG((DLOG_SEG, "lfs_writeseg: " 2186 "indir bp should not be B_CALL\n")); 2187 biodone(bp); 2188 bp = NULL; 2189 } else { 2190 /* Still on free list, leave it there */ 2191 unbusybp = bp; 2192 /* 2193 * We have to re-decrement lfs_avail 2194 * since this block is going to come 2195 * back around to us in the next 2196 * segment. 2197 */ 2198 lfs_sb_subavail(fs, 2199 lfs_btofsb(fs, bp->b_bcount)); 2200 } 2201 } else { 2202 lfs_freebuf(fs, newbp); 2203 } 2204 } 2205 mutex_enter(&bufcache_lock); 2206 if (unbusybp != NULL) { 2207 unbusybp->b_cflags &= ~BC_BUSY; 2208 if (unbusybp->b_cflags & BC_WANTED) 2209 cv_broadcast(&bp->b_busy); 2210 } 2211 } 2212 mutex_exit(&bufcache_lock); 2213 2214 /* 2215 * Compute checksum across data and then across summary; the first 2216 * block (the summary block) is skipped. Set the create time here 2217 * so that it's guaranteed to be later than the inode mod times. 2218 */ 2219 sum = 0; 2220 if (lfs_sb_getversion(fs) == 1) 2221 el_size = sizeof(u_long); 2222 else 2223 el_size = sizeof(u_int32_t); 2224 for (bpp = sp->bpp, i = nblocks - 1; i--; ) { 2225 ++bpp; 2226 /* Loop through gop_write cluster blocks */ 2227 for (byteoffset = 0; byteoffset < (*bpp)->b_bcount; 2228 byteoffset += lfs_sb_getbsize(fs)) { 2229 #ifdef LFS_USE_B_INVAL 2230 if (((*bpp)->b_cflags & BC_INVAL) != 0 && 2231 (*bpp)->b_iodone != NULL) { 2232 if (copyin((void *)(*bpp)->b_saveaddr + 2233 byteoffset, dp, el_size)) { 2234 panic("lfs_writeseg: copyin failed [1]:" 2235 " ino %d blk %" PRId64, 2236 VTOI((*bpp)->b_vp)->i_number, 2237 (*bpp)->b_lblkno); 2238 } 2239 } else 2240 #endif /* LFS_USE_B_INVAL */ 2241 { 2242 sum = lfs_cksum_part((char *) 2243 (*bpp)->b_data + byteoffset, el_size, sum); 2244 } 2245 } 2246 } 2247 if (lfs_sb_getversion(fs) == 1) 2248 lfs_ss_setocreate(fs, ssp, time_second); 2249 else { 2250 lfs_ss_setcreate(fs, ssp, time_second); 2251 lfs_sb_addserial(fs, 1); 2252 lfs_ss_setserial(fs, ssp, lfs_sb_getserial(fs)); 2253 lfs_ss_setident(fs, ssp, lfs_sb_getident(fs)); 2254 } 2255 lfs_ss_setdatasum(fs, ssp, lfs_cksum_fold(sum)); 2256 sumstart = lfs_ss_getsumstart(fs); 2257 lfs_ss_setsumsum(fs, ssp, cksum((char *)ssp + sumstart, 2258 lfs_sb_getsumsize(fs) - sumstart)); 2259 2260 mutex_enter(&lfs_lock); 2261 lfs_sb_subbfree(fs, (lfs_btofsb(fs, ninos * lfs_sb_getibsize(fs)) + 2262 lfs_btofsb(fs, lfs_sb_getsumsize(fs)))); 2263 lfs_sb_adddmeta(fs, (lfs_btofsb(fs, ninos * lfs_sb_getibsize(fs)) + 2264 lfs_btofsb(fs, lfs_sb_getsumsize(fs)))); 2265 mutex_exit(&lfs_lock); 2266 2267 /* 2268 * When we simply write the blocks we lose a rotation for every block 2269 * written. To avoid this problem, we cluster the buffers into a 2270 * chunk and write the chunk. MAXPHYS is the largest size I/O 2271 * devices can handle, use that for the size of the chunks. 2272 * 2273 * Blocks that are already clusters (from GOP_WRITE), however, we 2274 * don't bother to copy into other clusters. 2275 */ 2276 2277 #define CHUNKSIZE MAXPHYS 2278 2279 if (devvp == NULL) 2280 panic("devvp is NULL"); 2281 for (bpp = sp->bpp, i = nblocks; i;) { 2282 cbp = lfs_newclusterbuf(fs, devvp, (*bpp)->b_blkno, i); 2283 cl = cbp->b_private; 2284 2285 cbp->b_flags |= B_ASYNC; 2286 cbp->b_cflags |= BC_BUSY; 2287 cbp->b_bcount = 0; 2288 2289 #if defined(DEBUG) && defined(DIAGNOSTIC) 2290 if (bpp - sp->bpp > (lfs_sb_getsumsize(fs) - SEGSUM_SIZE(fs)) 2291 / sizeof(int32_t)) { 2292 panic("lfs_writeseg: real bpp overwrite"); 2293 } 2294 if (bpp - sp->bpp > lfs_segsize(fs) / lfs_sb_getfsize(fs)) { 2295 panic("lfs_writeseg: theoretical bpp overwrite"); 2296 } 2297 #endif 2298 2299 /* 2300 * Construct the cluster. 2301 */ 2302 mutex_enter(&lfs_lock); 2303 ++fs->lfs_iocount; 2304 mutex_exit(&lfs_lock); 2305 while (i && cbp->b_bcount < CHUNKSIZE) { 2306 bp = *bpp; 2307 2308 if (bp->b_bcount > (CHUNKSIZE - cbp->b_bcount)) 2309 break; 2310 if (cbp->b_bcount > 0 && !(cl->flags & LFS_CL_MALLOC)) 2311 break; 2312 2313 /* Clusters from GOP_WRITE are expedited */ 2314 if (bp->b_bcount > lfs_sb_getbsize(fs)) { 2315 if (cbp->b_bcount > 0) 2316 /* Put in its own buffer */ 2317 break; 2318 else { 2319 cbp->b_data = bp->b_data; 2320 } 2321 } else if (cbp->b_bcount == 0) { 2322 p = cbp->b_data = lfs_malloc(fs, CHUNKSIZE, 2323 LFS_NB_CLUSTER); 2324 cl->flags |= LFS_CL_MALLOC; 2325 } 2326 #ifdef DIAGNOSTIC 2327 if (lfs_dtosn(fs, LFS_DBTOFSB(fs, bp->b_blkno + 2328 btodb(bp->b_bcount - 1))) != 2329 sp->seg_number) { 2330 printf("blk size %d daddr %" PRIx64 2331 " not in seg %d\n", 2332 bp->b_bcount, bp->b_blkno, 2333 sp->seg_number); 2334 panic("segment overwrite"); 2335 } 2336 #endif 2337 2338 #ifdef LFS_USE_B_INVAL 2339 /* 2340 * Fake buffers from the cleaner are marked as B_INVAL. 2341 * We need to copy the data from user space rather than 2342 * from the buffer indicated. 2343 * XXX == what do I do on an error? 2344 */ 2345 if ((bp->b_cflags & BC_INVAL) != 0 && 2346 bp->b_iodone != NULL) { 2347 if (copyin(bp->b_saveaddr, p, bp->b_bcount)) 2348 panic("lfs_writeseg: " 2349 "copyin failed [2]"); 2350 } else 2351 #endif /* LFS_USE_B_INVAL */ 2352 if (cl->flags & LFS_CL_MALLOC) { 2353 /* copy data into our cluster. */ 2354 memcpy(p, bp->b_data, bp->b_bcount); 2355 p += bp->b_bcount; 2356 } 2357 2358 cbp->b_bcount += bp->b_bcount; 2359 cl->bufsize += bp->b_bcount; 2360 2361 bp->b_flags &= ~B_READ; 2362 bp->b_error = 0; 2363 cl->bpp[cl->bufcount++] = bp; 2364 2365 vp = bp->b_vp; 2366 mutex_enter(&bufcache_lock); 2367 mutex_enter(vp->v_interlock); 2368 bp->b_oflags &= ~(BO_DELWRI | BO_DONE); 2369 reassignbuf(bp, vp); 2370 vp->v_numoutput++; 2371 mutex_exit(vp->v_interlock); 2372 mutex_exit(&bufcache_lock); 2373 2374 bpp++; 2375 i--; 2376 } 2377 if (fs->lfs_sp->seg_flags & SEGM_SYNC) 2378 BIO_SETPRIO(cbp, BPRIO_TIMECRITICAL); 2379 else 2380 BIO_SETPRIO(cbp, BPRIO_TIMELIMITED); 2381 mutex_enter(devvp->v_interlock); 2382 devvp->v_numoutput++; 2383 mutex_exit(devvp->v_interlock); 2384 VOP_STRATEGY(devvp, cbp); 2385 curlwp->l_ru.ru_oublock++; 2386 } 2387 2388 if (lfs_dostats) { 2389 ++lfs_stats.psegwrites; 2390 lfs_stats.blocktot += nblocks - 1; 2391 if (fs->lfs_sp->seg_flags & SEGM_SYNC) 2392 ++lfs_stats.psyncwrites; 2393 if (fs->lfs_sp->seg_flags & SEGM_CLEAN) { 2394 ++lfs_stats.pcleanwrites; 2395 lfs_stats.cleanblocks += nblocks - 1; 2396 } 2397 } 2398 2399 return (lfs_initseg(fs) || do_again); 2400 } 2401 2402 void 2403 lfs_writesuper(struct lfs *fs, daddr_t daddr) 2404 { 2405 struct buf *bp; 2406 struct vnode *devvp = VTOI(fs->lfs_ivnode)->i_devvp; 2407 int s; 2408 2409 ASSERT_MAYBE_SEGLOCK(fs); 2410 #ifdef DIAGNOSTIC 2411 if (fs->lfs_is64) { 2412 KASSERT(fs->lfs_dlfs_u.u_64.dlfs_magic == LFS64_MAGIC); 2413 } else { 2414 KASSERT(fs->lfs_dlfs_u.u_32.dlfs_magic == LFS_MAGIC); 2415 } 2416 #endif 2417 /* 2418 * If we can write one superblock while another is in 2419 * progress, we risk not having a complete checkpoint if we crash. 2420 * So, block here if a superblock write is in progress. 2421 */ 2422 mutex_enter(&lfs_lock); 2423 s = splbio(); 2424 while (fs->lfs_sbactive) { 2425 mtsleep(&fs->lfs_sbactive, PRIBIO+1, "lfs sb", 0, 2426 &lfs_lock); 2427 } 2428 fs->lfs_sbactive = daddr; 2429 splx(s); 2430 mutex_exit(&lfs_lock); 2431 2432 /* Set timestamp of this version of the superblock */ 2433 if (lfs_sb_getversion(fs) == 1) 2434 lfs_sb_setotstamp(fs, time_second); 2435 lfs_sb_settstamp(fs, time_second); 2436 2437 /* The next chunk of code relies on this assumption */ 2438 CTASSERT(sizeof(struct dlfs) == sizeof(struct dlfs64)); 2439 2440 /* Checksum the superblock and copy it into a buffer. */ 2441 lfs_sb_setcksum(fs, lfs_sb_cksum(fs)); 2442 bp = lfs_newbuf(fs, devvp, 2443 LFS_FSBTODB(fs, daddr), LFS_SBPAD, LFS_NB_SBLOCK); 2444 memcpy(bp->b_data, &fs->lfs_dlfs_u, sizeof(struct dlfs)); 2445 memset((char *)bp->b_data + sizeof(struct dlfs), 0, 2446 LFS_SBPAD - sizeof(struct dlfs)); 2447 2448 bp->b_cflags |= BC_BUSY; 2449 bp->b_flags = (bp->b_flags & ~B_READ) | B_ASYNC; 2450 bp->b_oflags &= ~(BO_DONE | BO_DELWRI); 2451 bp->b_error = 0; 2452 bp->b_iodone = lfs_supercallback; 2453 2454 if (fs->lfs_sp != NULL && fs->lfs_sp->seg_flags & SEGM_SYNC) 2455 BIO_SETPRIO(bp, BPRIO_TIMECRITICAL); 2456 else 2457 BIO_SETPRIO(bp, BPRIO_TIMELIMITED); 2458 curlwp->l_ru.ru_oublock++; 2459 2460 mutex_enter(devvp->v_interlock); 2461 devvp->v_numoutput++; 2462 mutex_exit(devvp->v_interlock); 2463 2464 mutex_enter(&lfs_lock); 2465 ++fs->lfs_iocount; 2466 mutex_exit(&lfs_lock); 2467 VOP_STRATEGY(devvp, bp); 2468 } 2469 2470 /* 2471 * Logical block number match routines used when traversing the dirty block 2472 * chain. 2473 */ 2474 int 2475 lfs_match_fake(struct lfs *fs, struct buf *bp) 2476 { 2477 2478 ASSERT_SEGLOCK(fs); 2479 return LFS_IS_MALLOC_BUF(bp); 2480 } 2481 2482 #if 0 2483 int 2484 lfs_match_real(struct lfs *fs, struct buf *bp) 2485 { 2486 2487 ASSERT_SEGLOCK(fs); 2488 return (lfs_match_data(fs, bp) && !lfs_match_fake(fs, bp)); 2489 } 2490 #endif 2491 2492 int 2493 lfs_match_data(struct lfs *fs, struct buf *bp) 2494 { 2495 2496 ASSERT_SEGLOCK(fs); 2497 return (bp->b_lblkno >= 0); 2498 } 2499 2500 int 2501 lfs_match_indir(struct lfs *fs, struct buf *bp) 2502 { 2503 daddr_t lbn; 2504 2505 ASSERT_SEGLOCK(fs); 2506 lbn = bp->b_lblkno; 2507 return (lbn < 0 && (-lbn - ULFS_NDADDR) % LFS_NINDIR(fs) == 0); 2508 } 2509 2510 int 2511 lfs_match_dindir(struct lfs *fs, struct buf *bp) 2512 { 2513 daddr_t lbn; 2514 2515 ASSERT_SEGLOCK(fs); 2516 lbn = bp->b_lblkno; 2517 return (lbn < 0 && (-lbn - ULFS_NDADDR) % LFS_NINDIR(fs) == 1); 2518 } 2519 2520 int 2521 lfs_match_tindir(struct lfs *fs, struct buf *bp) 2522 { 2523 daddr_t lbn; 2524 2525 ASSERT_SEGLOCK(fs); 2526 lbn = bp->b_lblkno; 2527 return (lbn < 0 && (-lbn - ULFS_NDADDR) % LFS_NINDIR(fs) == 2); 2528 } 2529 2530 static void 2531 lfs_free_aiodone(struct buf *bp) 2532 { 2533 struct lfs *fs; 2534 2535 KERNEL_LOCK(1, curlwp); 2536 fs = bp->b_private; 2537 ASSERT_NO_SEGLOCK(fs); 2538 lfs_freebuf(fs, bp); 2539 KERNEL_UNLOCK_LAST(curlwp); 2540 } 2541 2542 static void 2543 lfs_super_aiodone(struct buf *bp) 2544 { 2545 struct lfs *fs; 2546 2547 KERNEL_LOCK(1, curlwp); 2548 fs = bp->b_private; 2549 ASSERT_NO_SEGLOCK(fs); 2550 mutex_enter(&lfs_lock); 2551 fs->lfs_sbactive = 0; 2552 if (--fs->lfs_iocount <= 1) 2553 wakeup(&fs->lfs_iocount); 2554 wakeup(&fs->lfs_sbactive); 2555 mutex_exit(&lfs_lock); 2556 lfs_freebuf(fs, bp); 2557 KERNEL_UNLOCK_LAST(curlwp); 2558 } 2559 2560 static void 2561 lfs_cluster_aiodone(struct buf *bp) 2562 { 2563 struct lfs_cluster *cl; 2564 struct lfs *fs; 2565 struct buf *tbp, *fbp; 2566 struct vnode *vp, *devvp, *ovp; 2567 struct inode *ip; 2568 int error; 2569 2570 KERNEL_LOCK(1, curlwp); 2571 2572 error = bp->b_error; 2573 cl = bp->b_private; 2574 fs = cl->fs; 2575 devvp = VTOI(fs->lfs_ivnode)->i_devvp; 2576 ASSERT_NO_SEGLOCK(fs); 2577 2578 /* Put the pages back, and release the buffer */ 2579 while (cl->bufcount--) { 2580 tbp = cl->bpp[cl->bufcount]; 2581 KASSERT(tbp->b_cflags & BC_BUSY); 2582 if (error) { 2583 tbp->b_error = error; 2584 } 2585 2586 /* 2587 * We're done with tbp. If it has not been re-dirtied since 2588 * the cluster was written, free it. Otherwise, keep it on 2589 * the locked list to be written again. 2590 */ 2591 vp = tbp->b_vp; 2592 2593 tbp->b_flags &= ~B_GATHERED; 2594 2595 LFS_BCLEAN_LOG(fs, tbp); 2596 2597 mutex_enter(&bufcache_lock); 2598 if (tbp->b_iodone == NULL) { 2599 KASSERT(tbp->b_flags & B_LOCKED); 2600 bremfree(tbp); 2601 if (vp) { 2602 mutex_enter(vp->v_interlock); 2603 reassignbuf(tbp, vp); 2604 mutex_exit(vp->v_interlock); 2605 } 2606 tbp->b_flags |= B_ASYNC; /* for biodone */ 2607 } 2608 2609 if (((tbp->b_flags | tbp->b_oflags) & 2610 (B_LOCKED | BO_DELWRI)) == B_LOCKED) 2611 LFS_UNLOCK_BUF(tbp); 2612 2613 if (tbp->b_oflags & BO_DONE) { 2614 DLOG((DLOG_SEG, "blk %d biodone already (flags %lx)\n", 2615 cl->bufcount, (long)tbp->b_flags)); 2616 } 2617 2618 if (tbp->b_iodone != NULL && !LFS_IS_MALLOC_BUF(tbp)) { 2619 /* 2620 * A buffer from the page daemon. 2621 * We use the same iodone as it does, 2622 * so we must manually disassociate its 2623 * buffers from the vp. 2624 */ 2625 if ((ovp = tbp->b_vp) != NULL) { 2626 /* This is just silly */ 2627 mutex_enter(ovp->v_interlock); 2628 brelvp(tbp); 2629 mutex_exit(ovp->v_interlock); 2630 tbp->b_vp = vp; 2631 tbp->b_objlock = vp->v_interlock; 2632 } 2633 /* Put it back the way it was */ 2634 tbp->b_flags |= B_ASYNC; 2635 /* Master buffers have BC_AGE */ 2636 if (tbp->b_private == tbp) 2637 tbp->b_cflags |= BC_AGE; 2638 } 2639 mutex_exit(&bufcache_lock); 2640 2641 biodone(tbp); 2642 2643 /* 2644 * If this is the last block for this vnode, but 2645 * there are other blocks on its dirty list, 2646 * set IN_MODIFIED/IN_CLEANING depending on what 2647 * sort of block. Only do this for our mount point, 2648 * not for, e.g., inode blocks that are attached to 2649 * the devvp. 2650 * XXX KS - Shouldn't we set *both* if both types 2651 * of blocks are present (traverse the dirty list?) 2652 */ 2653 mutex_enter(vp->v_interlock); 2654 mutex_enter(&lfs_lock); 2655 if (vp != devvp && vp->v_numoutput == 0 && 2656 (fbp = LIST_FIRST(&vp->v_dirtyblkhd)) != NULL) { 2657 ip = VTOI(vp); 2658 DLOG((DLOG_SEG, "lfs_cluster_aiodone: mark ino %d\n", 2659 ip->i_number)); 2660 if (LFS_IS_MALLOC_BUF(fbp)) 2661 LFS_SET_UINO(ip, IN_CLEANING); 2662 else 2663 LFS_SET_UINO(ip, IN_MODIFIED); 2664 } 2665 cv_broadcast(&vp->v_cv); 2666 mutex_exit(&lfs_lock); 2667 mutex_exit(vp->v_interlock); 2668 } 2669 2670 /* Fix up the cluster buffer, and release it */ 2671 if (cl->flags & LFS_CL_MALLOC) 2672 lfs_free(fs, bp->b_data, LFS_NB_CLUSTER); 2673 putiobuf(bp); 2674 2675 /* Note i/o done */ 2676 if (cl->flags & LFS_CL_SYNC) { 2677 if (--cl->seg->seg_iocount == 0) 2678 wakeup(&cl->seg->seg_iocount); 2679 } 2680 mutex_enter(&lfs_lock); 2681 #ifdef DIAGNOSTIC 2682 if (fs->lfs_iocount == 0) 2683 panic("lfs_cluster_aiodone: zero iocount"); 2684 #endif 2685 if (--fs->lfs_iocount <= 1) 2686 wakeup(&fs->lfs_iocount); 2687 mutex_exit(&lfs_lock); 2688 2689 KERNEL_UNLOCK_LAST(curlwp); 2690 2691 pool_put(&fs->lfs_bpppool, cl->bpp); 2692 cl->bpp = NULL; 2693 pool_put(&fs->lfs_clpool, cl); 2694 } 2695 2696 static void 2697 lfs_generic_callback(struct buf *bp, void (*aiodone)(struct buf *)) 2698 { 2699 /* reset b_iodone for when this is a single-buf i/o. */ 2700 bp->b_iodone = aiodone; 2701 2702 workqueue_enqueue(uvm.aiodone_queue, &bp->b_work, NULL); 2703 } 2704 2705 static void 2706 lfs_cluster_callback(struct buf *bp) 2707 { 2708 2709 lfs_generic_callback(bp, lfs_cluster_aiodone); 2710 } 2711 2712 void 2713 lfs_supercallback(struct buf *bp) 2714 { 2715 2716 lfs_generic_callback(bp, lfs_super_aiodone); 2717 } 2718 2719 /* 2720 * The only buffers that are going to hit these functions are the 2721 * segment write blocks, or the segment summaries, or the superblocks. 2722 * 2723 * All of the above are created by lfs_newbuf, and so do not need to be 2724 * released via brelse. 2725 */ 2726 void 2727 lfs_callback(struct buf *bp) 2728 { 2729 2730 lfs_generic_callback(bp, lfs_free_aiodone); 2731 } 2732 2733 /* 2734 * Shellsort (diminishing increment sort) from Data Structures and 2735 * Algorithms, Aho, Hopcraft and Ullman, 1983 Edition, page 290; 2736 * see also Knuth Vol. 3, page 84. The increments are selected from 2737 * formula (8), page 95. Roughly O(N^3/2). 2738 */ 2739 /* 2740 * This is our own private copy of shellsort because we want to sort 2741 * two parallel arrays (the array of buffer pointers and the array of 2742 * logical block numbers) simultaneously. Note that we cast the array 2743 * of logical block numbers to a unsigned in this routine so that the 2744 * negative block numbers (meta data blocks) sort AFTER the data blocks. 2745 */ 2746 2747 static void 2748 lfs_shellsort(struct lfs *fs, 2749 struct buf **bp_array, union lfs_blocks *lb_array, 2750 int nmemb, int size) 2751 { 2752 static int __rsshell_increments[] = { 4, 1, 0 }; 2753 int incr, *incrp, t1, t2; 2754 struct buf *bp_temp; 2755 2756 #ifdef DEBUG 2757 incr = 0; 2758 for (t1 = 0; t1 < nmemb; t1++) { 2759 for (t2 = 0; t2 * size < bp_array[t1]->b_bcount; t2++) { 2760 if (lfs_blocks_get(fs, lb_array, incr++) != bp_array[t1]->b_lblkno + t2) { 2761 /* dump before panic */ 2762 printf("lfs_shellsort: nmemb=%d, size=%d\n", 2763 nmemb, size); 2764 incr = 0; 2765 for (t1 = 0; t1 < nmemb; t1++) { 2766 const struct buf *bp = bp_array[t1]; 2767 2768 printf("bp[%d]: lbn=%" PRIu64 ", size=%" 2769 PRIu64 "\n", t1, 2770 (uint64_t)bp->b_bcount, 2771 (uint64_t)bp->b_lblkno); 2772 printf("lbns:"); 2773 for (t2 = 0; t2 * size < bp->b_bcount; 2774 t2++) { 2775 printf(" %jd", 2776 (intmax_t)lfs_blocks_get(fs, lb_array, incr++)); 2777 } 2778 printf("\n"); 2779 } 2780 panic("lfs_shellsort: inconsistent input"); 2781 } 2782 } 2783 } 2784 #endif 2785 2786 for (incrp = __rsshell_increments; (incr = *incrp++) != 0;) 2787 for (t1 = incr; t1 < nmemb; ++t1) 2788 for (t2 = t1 - incr; t2 >= 0;) 2789 if ((u_int64_t)bp_array[t2]->b_lblkno > 2790 (u_int64_t)bp_array[t2 + incr]->b_lblkno) { 2791 bp_temp = bp_array[t2]; 2792 bp_array[t2] = bp_array[t2 + incr]; 2793 bp_array[t2 + incr] = bp_temp; 2794 t2 -= incr; 2795 } else 2796 break; 2797 2798 /* Reform the list of logical blocks */ 2799 incr = 0; 2800 for (t1 = 0; t1 < nmemb; t1++) { 2801 for (t2 = 0; t2 * size < bp_array[t1]->b_bcount; t2++) { 2802 lfs_blocks_set(fs, lb_array, incr++, 2803 bp_array[t1]->b_lblkno + t2); 2804 } 2805 } 2806 } 2807 2808 /* 2809 * Set up an FINFO entry for a new file. The fip pointer is assumed to 2810 * point at uninitialized space. 2811 */ 2812 void 2813 lfs_acquire_finfo(struct lfs *fs, ino_t ino, int vers) 2814 { 2815 struct segment *sp = fs->lfs_sp; 2816 SEGSUM *ssp; 2817 2818 KASSERT(vers > 0); 2819 2820 if (sp->seg_bytes_left < lfs_sb_getbsize(fs) || 2821 sp->sum_bytes_left < FINFOSIZE(fs) + LFS_BLKPTRSIZE(fs)) 2822 (void) lfs_writeseg(fs, fs->lfs_sp); 2823 2824 sp->sum_bytes_left -= FINFOSIZE(fs); 2825 ssp = (SEGSUM *)sp->segsum; 2826 lfs_ss_setnfinfo(fs, ssp, lfs_ss_getnfinfo(fs, ssp) + 1); 2827 lfs_fi_setnblocks(fs, sp->fip, 0); 2828 lfs_fi_setino(fs, sp->fip, ino); 2829 lfs_fi_setversion(fs, sp->fip, vers); 2830 } 2831 2832 /* 2833 * Release the FINFO entry, either clearing out an unused entry or 2834 * advancing us to the next available entry. 2835 */ 2836 void 2837 lfs_release_finfo(struct lfs *fs) 2838 { 2839 struct segment *sp = fs->lfs_sp; 2840 SEGSUM *ssp; 2841 2842 if (lfs_fi_getnblocks(fs, sp->fip) != 0) { 2843 sp->fip = NEXT_FINFO(fs, sp->fip); 2844 lfs_blocks_fromfinfo(fs, &sp->start_lbp, sp->fip); 2845 } else { 2846 /* XXX shouldn't this update sp->fip? */ 2847 sp->sum_bytes_left += FINFOSIZE(fs); 2848 ssp = (SEGSUM *)sp->segsum; 2849 lfs_ss_setnfinfo(fs, ssp, lfs_ss_getnfinfo(fs, ssp) - 1); 2850 } 2851 } 2852