lfs_segment.c revision 1.167
1 /* $NetBSD: lfs_segment.c,v 1.167 2005/09/26 13:52:20 yamt Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 2000, 2001, 2002, 2003 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Konrad E. Schroder <perseant (at) hhhh.org>. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. All advertising materials mentioning features or use of this software 19 * must display the following acknowledgement: 20 * This product includes software developed by the NetBSD 21 * Foundation, Inc. and its contributors. 22 * 4. Neither the name of The NetBSD Foundation nor the names of its 23 * contributors may be used to endorse or promote products derived 24 * from this software without specific prior written permission. 25 * 26 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 27 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 28 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 29 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 30 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 31 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 32 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 33 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 34 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 35 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 36 * POSSIBILITY OF SUCH DAMAGE. 37 */ 38 /* 39 * Copyright (c) 1991, 1993 40 * The Regents of the University of California. All rights reserved. 41 * 42 * Redistribution and use in source and binary forms, with or without 43 * modification, are permitted provided that the following conditions 44 * are met: 45 * 1. Redistributions of source code must retain the above copyright 46 * notice, this list of conditions and the following disclaimer. 47 * 2. Redistributions in binary form must reproduce the above copyright 48 * notice, this list of conditions and the following disclaimer in the 49 * documentation and/or other materials provided with the distribution. 50 * 3. Neither the name of the University nor the names of its contributors 51 * may be used to endorse or promote products derived from this software 52 * without specific prior written permission. 53 * 54 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 55 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 56 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 57 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 58 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 59 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 60 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 61 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 62 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 63 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 64 * SUCH DAMAGE. 65 * 66 * @(#)lfs_segment.c 8.10 (Berkeley) 6/10/95 67 */ 68 69 #include <sys/cdefs.h> 70 __KERNEL_RCSID(0, "$NetBSD: lfs_segment.c,v 1.167 2005/09/26 13:52:20 yamt Exp $"); 71 72 #ifdef DEBUG 73 # define vndebug(vp, str) do { \ 74 if (VTOI(vp)->i_flag & IN_CLEANING) \ 75 DLOG((DLOG_WVNODE, "not writing ino %d because %s (op %d)\n", \ 76 VTOI(vp)->i_number, (str), op)); \ 77 } while(0) 78 #else 79 # define vndebug(vp, str) 80 #endif 81 #define ivndebug(vp, str) \ 82 DLOG((DLOG_WVNODE, "ino %d: %s\n", VTOI(vp)->i_number, (str))) 83 84 #if defined(_KERNEL_OPT) 85 #include "opt_ddb.h" 86 #endif 87 88 #include <sys/param.h> 89 #include <sys/systm.h> 90 #include <sys/namei.h> 91 #include <sys/kernel.h> 92 #include <sys/resourcevar.h> 93 #include <sys/file.h> 94 #include <sys/stat.h> 95 #include <sys/buf.h> 96 #include <sys/proc.h> 97 #include <sys/vnode.h> 98 #include <sys/mount.h> 99 100 #include <miscfs/specfs/specdev.h> 101 #include <miscfs/fifofs/fifo.h> 102 103 #include <ufs/ufs/inode.h> 104 #include <ufs/ufs/dir.h> 105 #include <ufs/ufs/ufsmount.h> 106 #include <ufs/ufs/ufs_extern.h> 107 108 #include <ufs/lfs/lfs.h> 109 #include <ufs/lfs/lfs_extern.h> 110 111 #include <uvm/uvm.h> 112 #include <uvm/uvm_extern.h> 113 114 MALLOC_DEFINE(M_SEGMENT, "LFS segment", "Segment for LFS"); 115 116 extern int count_lock_queue(void); 117 extern struct simplelock vnode_free_list_slock; /* XXX */ 118 extern struct simplelock bqueue_slock; /* XXX */ 119 120 static void lfs_generic_callback(struct buf *, void (*)(struct buf *)); 121 static void lfs_super_aiodone(struct buf *); 122 static void lfs_cluster_aiodone(struct buf *); 123 static void lfs_cluster_callback(struct buf *); 124 125 /* 126 * Determine if it's OK to start a partial in this segment, or if we need 127 * to go on to a new segment. 128 */ 129 #define LFS_PARTIAL_FITS(fs) \ 130 ((fs)->lfs_fsbpseg - ((fs)->lfs_offset - (fs)->lfs_curseg) > \ 131 fragstofsb((fs), (fs)->lfs_frag)) 132 133 int lfs_match_fake(struct lfs *, struct buf *); 134 void lfs_newseg(struct lfs *); 135 /* XXX ondisk32 */ 136 void lfs_shellsort(struct buf **, int32_t *, int, int); 137 void lfs_supercallback(struct buf *); 138 void lfs_updatemeta(struct segment *); 139 void lfs_writesuper(struct lfs *, daddr_t); 140 int lfs_writevnodes(struct lfs *fs, struct mount *mp, 141 struct segment *sp, int dirops); 142 143 int lfs_allclean_wakeup; /* Cleaner wakeup address. */ 144 int lfs_writeindir = 1; /* whether to flush indir on non-ckp */ 145 int lfs_clean_vnhead = 0; /* Allow freeing to head of vn list */ 146 int lfs_dirvcount = 0; /* # active dirops */ 147 148 /* Statistics Counters */ 149 int lfs_dostats = 1; 150 struct lfs_stats lfs_stats; 151 152 /* op values to lfs_writevnodes */ 153 #define VN_REG 0 154 #define VN_DIROP 1 155 #define VN_EMPTY 2 156 #define VN_CLEAN 3 157 158 /* 159 * XXX KS - Set modification time on the Ifile, so the cleaner can 160 * read the fs mod time off of it. We don't set IN_UPDATE here, 161 * since we don't really need this to be flushed to disk (and in any 162 * case that wouldn't happen to the Ifile until we checkpoint). 163 */ 164 void 165 lfs_imtime(struct lfs *fs) 166 { 167 struct timespec ts; 168 struct inode *ip; 169 170 ASSERT_MAYBE_SEGLOCK(fs); 171 nanotime(&ts); 172 ip = VTOI(fs->lfs_ivnode); 173 ip->i_ffs1_mtime = ts.tv_sec; 174 ip->i_ffs1_mtimensec = ts.tv_nsec; 175 } 176 177 /* 178 * Ifile and meta data blocks are not marked busy, so segment writes MUST be 179 * single threaded. Currently, there are two paths into lfs_segwrite, sync() 180 * and getnewbuf(). They both mark the file system busy. Lfs_vflush() 181 * explicitly marks the file system busy. So lfs_segwrite is safe. I think. 182 */ 183 184 #define SET_FLUSHING(fs,vp) (fs)->lfs_flushvp = (vp) 185 #define IS_FLUSHING(fs,vp) ((fs)->lfs_flushvp == (vp)) 186 #define CLR_FLUSHING(fs,vp) (fs)->lfs_flushvp = NULL 187 188 int 189 lfs_vflush(struct vnode *vp) 190 { 191 struct inode *ip; 192 struct lfs *fs; 193 struct segment *sp; 194 struct buf *bp, *nbp, *tbp, *tnbp; 195 int error, s; 196 int flushed; 197 #if 0 198 int redo; 199 #endif 200 201 ip = VTOI(vp); 202 fs = VFSTOUFS(vp->v_mount)->um_lfs; 203 204 ASSERT_NO_SEGLOCK(fs); 205 if (ip->i_flag & IN_CLEANING) { 206 ivndebug(vp,"vflush/in_cleaning"); 207 LFS_CLR_UINO(ip, IN_CLEANING); 208 LFS_SET_UINO(ip, IN_MODIFIED); 209 210 /* 211 * Toss any cleaning buffers that have real counterparts 212 * to avoid losing new data. 213 */ 214 s = splbio(); 215 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) { 216 nbp = LIST_NEXT(bp, b_vnbufs); 217 if (!LFS_IS_MALLOC_BUF(bp)) 218 continue; 219 /* 220 * Look for pages matching the range covered 221 * by cleaning blocks. It's okay if more dirty 222 * pages appear, so long as none disappear out 223 * from under us. 224 */ 225 if (bp->b_lblkno > 0 && vp->v_type == VREG && 226 vp != fs->lfs_ivnode) { 227 struct vm_page *pg; 228 voff_t off; 229 230 simple_lock(&vp->v_interlock); 231 for (off = lblktosize(fs, bp->b_lblkno); 232 off < lblktosize(fs, bp->b_lblkno + 1); 233 off += PAGE_SIZE) { 234 pg = uvm_pagelookup(&vp->v_uobj, off); 235 if (pg == NULL) 236 continue; 237 if ((pg->flags & PG_CLEAN) == 0 || 238 pmap_is_modified(pg)) { 239 fs->lfs_avail += btofsb(fs, 240 bp->b_bcount); 241 wakeup(&fs->lfs_avail); 242 lfs_freebuf(fs, bp); 243 bp = NULL; 244 goto nextbp; 245 } 246 } 247 simple_unlock(&vp->v_interlock); 248 } 249 for (tbp = LIST_FIRST(&vp->v_dirtyblkhd); tbp; 250 tbp = tnbp) 251 { 252 tnbp = LIST_NEXT(tbp, b_vnbufs); 253 if (tbp->b_vp == bp->b_vp 254 && tbp->b_lblkno == bp->b_lblkno 255 && tbp != bp) 256 { 257 fs->lfs_avail += btofsb(fs, 258 bp->b_bcount); 259 wakeup(&fs->lfs_avail); 260 lfs_freebuf(fs, bp); 261 bp = NULL; 262 break; 263 } 264 } 265 nextbp: 266 ; 267 } 268 splx(s); 269 } 270 271 /* If the node is being written, wait until that is done */ 272 simple_lock(&vp->v_interlock); 273 s = splbio(); 274 if (WRITEINPROG(vp)) { 275 ivndebug(vp,"vflush/writeinprog"); 276 ltsleep(vp, (PRIBIO+1), "lfs_vw", 0, &vp->v_interlock); 277 } 278 splx(s); 279 simple_unlock(&vp->v_interlock); 280 281 /* Protect against VXLOCK deadlock in vinvalbuf() */ 282 lfs_seglock(fs, SEGM_SYNC); 283 284 /* If we're supposed to flush a freed inode, just toss it */ 285 /* XXX - seglock, so these buffers can't be gathered, right? */ 286 if (ip->i_mode == 0) { 287 DLOG((DLOG_VNODE, "lfs_vflush: ino %d freed, not flushing\n", 288 ip->i_number)); 289 s = splbio(); 290 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) { 291 nbp = LIST_NEXT(bp, b_vnbufs); 292 if (bp->b_flags & B_DELWRI) { /* XXX always true? */ 293 fs->lfs_avail += btofsb(fs, bp->b_bcount); 294 wakeup(&fs->lfs_avail); 295 } 296 /* Copied from lfs_writeseg */ 297 if (bp->b_flags & B_CALL) { 298 biodone(bp); 299 } else { 300 bremfree(bp); 301 LFS_UNLOCK_BUF(bp); 302 bp->b_flags &= ~(B_ERROR | B_READ | B_DELWRI | 303 B_GATHERED); 304 bp->b_flags |= B_DONE; 305 reassignbuf(bp, vp); 306 brelse(bp); 307 } 308 } 309 splx(s); 310 LFS_CLR_UINO(ip, IN_CLEANING); 311 LFS_CLR_UINO(ip, IN_MODIFIED | IN_ACCESSED); 312 ip->i_flag &= ~IN_ALLMOD; 313 DLOG((DLOG_VNODE, "lfs_vflush: done not flushing ino %d\n", 314 ip->i_number)); 315 lfs_segunlock(fs); 316 return 0; 317 } 318 319 SET_FLUSHING(fs,vp); 320 if (fs->lfs_nactive > LFS_MAX_ACTIVE || 321 (fs->lfs_sp->seg_flags & SEGM_CKP)) { 322 error = lfs_segwrite(vp->v_mount, SEGM_CKP | SEGM_SYNC); 323 CLR_FLUSHING(fs,vp); 324 lfs_segunlock(fs); 325 return error; 326 } 327 sp = fs->lfs_sp; 328 329 flushed = 0; 330 if (VPISEMPTY(vp)) { 331 lfs_writevnodes(fs, vp->v_mount, sp, VN_EMPTY); 332 ++flushed; 333 } else if ((ip->i_flag & IN_CLEANING) && 334 (fs->lfs_sp->seg_flags & SEGM_CLEAN)) { 335 ivndebug(vp,"vflush/clean"); 336 lfs_writevnodes(fs, vp->v_mount, sp, VN_CLEAN); 337 ++flushed; 338 } else if (lfs_dostats) { 339 if (!VPISEMPTY(vp) || (VTOI(vp)->i_flag & IN_ALLMOD)) 340 ++lfs_stats.vflush_invoked; 341 ivndebug(vp,"vflush"); 342 } 343 344 #ifdef DIAGNOSTIC 345 if (vp->v_flag & VDIROP) { 346 DLOG((DLOG_VNODE, "lfs_vflush: flushing VDIROP\n")); 347 /* panic("lfs_vflush: VDIROP being flushed...this can\'t happen"); */ 348 } 349 if (vp->v_usecount < 0) { 350 printf("usecount=%ld\n", (long)vp->v_usecount); 351 panic("lfs_vflush: usecount<0"); 352 } 353 #endif 354 355 #if 1 356 do { 357 do { 358 if (LIST_FIRST(&vp->v_dirtyblkhd) != NULL) 359 lfs_writefile(fs, sp, vp); 360 } while (lfs_writeinode(fs, sp, ip)); 361 } while (lfs_writeseg(fs, sp) && ip->i_number == LFS_IFILE_INUM); 362 #else 363 if (flushed && vp != fs->lfs_ivnode) 364 lfs_writeseg(fs, sp); 365 else do { 366 simple_lock(&fs->lfs_interlock); 367 fs->lfs_flags &= ~LFS_IFDIRTY; 368 simple_unlock(&fs->lfs_interlock); 369 lfs_writefile(fs, sp, vp); 370 redo = lfs_writeinode(fs, sp, ip); 371 redo += lfs_writeseg(fs, sp); 372 simple_lock(&fs->lfs_interlock); 373 redo += (fs->lfs_flags & LFS_IFDIRTY); 374 simple_unlock(&fs->lfs_interlock); 375 } while (redo && vp == fs->lfs_ivnode); 376 #endif 377 if (lfs_dostats) { 378 ++lfs_stats.nwrites; 379 if (sp->seg_flags & SEGM_SYNC) 380 ++lfs_stats.nsync_writes; 381 if (sp->seg_flags & SEGM_CKP) 382 ++lfs_stats.ncheckpoints; 383 } 384 /* 385 * If we were called from somewhere that has already held the seglock 386 * (e.g., lfs_markv()), the lfs_segunlock will not wait for 387 * the write to complete because we are still locked. 388 * Since lfs_vflush() must return the vnode with no dirty buffers, 389 * we must explicitly wait, if that is the case. 390 * 391 * We compare the iocount against 1, not 0, because it is 392 * artificially incremented by lfs_seglock(). 393 */ 394 simple_lock(&fs->lfs_interlock); 395 if (fs->lfs_seglock > 1) { 396 while (fs->lfs_iocount > 1) 397 (void)ltsleep(&fs->lfs_iocount, PRIBIO + 1, 398 "lfs_vflush", 0, &fs->lfs_interlock); 399 } 400 simple_unlock(&fs->lfs_interlock); 401 402 lfs_segunlock(fs); 403 404 /* Wait for these buffers to be recovered by aiodoned */ 405 s = splbio(); 406 simple_lock(&global_v_numoutput_slock); 407 while (vp->v_numoutput > 0) { 408 vp->v_flag |= VBWAIT; 409 ltsleep(&vp->v_numoutput, PRIBIO + 1, "lfs_vf2", 0, 410 &global_v_numoutput_slock); 411 } 412 simple_unlock(&global_v_numoutput_slock); 413 splx(s); 414 415 CLR_FLUSHING(fs,vp); 416 return (0); 417 } 418 419 int 420 lfs_writevnodes(struct lfs *fs, struct mount *mp, struct segment *sp, int op) 421 { 422 struct inode *ip; 423 struct vnode *vp, *nvp; 424 int inodes_written = 0, only_cleaning; 425 426 ASSERT_SEGLOCK(fs); 427 #ifndef LFS_NO_BACKVP_HACK 428 /* BEGIN HACK */ 429 #define VN_OFFSET \ 430 (((caddr_t)&LIST_NEXT(vp, v_mntvnodes)) - (caddr_t)vp) 431 #define BACK_VP(VP) \ 432 ((struct vnode *)(((caddr_t)(VP)->v_mntvnodes.le_prev) - VN_OFFSET)) 433 #define BEG_OF_VLIST \ 434 ((struct vnode *)(((caddr_t)&LIST_FIRST(&mp->mnt_vnodelist)) \ 435 - VN_OFFSET)) 436 437 /* Find last vnode. */ 438 loop: for (vp = LIST_FIRST(&mp->mnt_vnodelist); 439 vp && LIST_NEXT(vp, v_mntvnodes) != NULL; 440 vp = LIST_NEXT(vp, v_mntvnodes)); 441 for (; vp && vp != BEG_OF_VLIST; vp = nvp) { 442 nvp = BACK_VP(vp); 443 #else 444 loop: 445 for (vp = LIST_FIRST(&mp->mnt_vnodelist); vp; vp = nvp) { 446 nvp = LIST_NEXT(vp, v_mntvnodes); 447 #endif 448 /* 449 * If the vnode that we are about to sync is no longer 450 * associated with this mount point, start over. 451 */ 452 if (vp->v_mount != mp) { 453 DLOG((DLOG_VNODE, "lfs_writevnodes: starting over\n")); 454 /* 455 * After this, pages might be busy 456 * due to our own previous putpages. 457 * Start actual segment write here to avoid deadlock. 458 */ 459 (void)lfs_writeseg(fs, sp); 460 goto loop; 461 } 462 463 if (vp->v_type == VNON) { 464 continue; 465 } 466 467 ip = VTOI(vp); 468 if ((op == VN_DIROP && !(vp->v_flag & VDIROP)) || 469 (op != VN_DIROP && op != VN_CLEAN && 470 (vp->v_flag & VDIROP))) { 471 vndebug(vp,"dirop"); 472 continue; 473 } 474 475 if (op == VN_EMPTY && !VPISEMPTY(vp)) { 476 vndebug(vp,"empty"); 477 continue; 478 } 479 480 if (op == VN_CLEAN && ip->i_number != LFS_IFILE_INUM 481 && vp != fs->lfs_flushvp 482 && !(ip->i_flag & IN_CLEANING)) { 483 vndebug(vp,"cleaning"); 484 continue; 485 } 486 487 if (lfs_vref(vp)) { 488 vndebug(vp,"vref"); 489 continue; 490 } 491 492 only_cleaning = 0; 493 /* 494 * Write the inode/file if dirty and it's not the IFILE. 495 */ 496 if ((ip->i_flag & IN_ALLMOD) || !VPISEMPTY(vp)) { 497 only_cleaning = 498 ((ip->i_flag & IN_ALLMOD) == IN_CLEANING); 499 500 if (ip->i_number != LFS_IFILE_INUM) { 501 lfs_writefile(fs, sp, vp); 502 if (!VPISEMPTY(vp)) { 503 if (WRITEINPROG(vp)) { 504 ivndebug(vp,"writevnodes/write2"); 505 } else if (!(ip->i_flag & IN_ALLMOD)) { 506 LFS_SET_UINO(ip, IN_MODIFIED); 507 } 508 } 509 (void) lfs_writeinode(fs, sp, ip); 510 inodes_written++; 511 } 512 } 513 514 if (lfs_clean_vnhead && only_cleaning) 515 lfs_vunref_head(vp); 516 else 517 lfs_vunref(vp); 518 } 519 return inodes_written; 520 } 521 522 /* 523 * Do a checkpoint. 524 */ 525 int 526 lfs_segwrite(struct mount *mp, int flags) 527 { 528 struct buf *bp; 529 struct inode *ip; 530 struct lfs *fs; 531 struct segment *sp; 532 struct vnode *vp; 533 SEGUSE *segusep; 534 int do_ckp, did_ckp, error, s; 535 unsigned n, segleft, maxseg, sn, i, curseg; 536 int writer_set = 0; 537 int dirty; 538 int redo; 539 540 fs = VFSTOUFS(mp)->um_lfs; 541 ASSERT_MAYBE_SEGLOCK(fs); 542 543 if (fs->lfs_ronly) 544 return EROFS; 545 546 lfs_imtime(fs); 547 548 /* 549 * Allocate a segment structure and enough space to hold pointers to 550 * the maximum possible number of buffers which can be described in a 551 * single summary block. 552 */ 553 do_ckp = (flags & SEGM_CKP) || fs->lfs_nactive > LFS_MAX_ACTIVE; 554 lfs_seglock(fs, flags | (do_ckp ? SEGM_CKP : 0)); 555 sp = fs->lfs_sp; 556 557 /* 558 * If lfs_flushvp is non-NULL, we are called from lfs_vflush, 559 * in which case we have to flush *all* buffers off of this vnode. 560 * We don't care about other nodes, but write any non-dirop nodes 561 * anyway in anticipation of another getnewvnode(). 562 * 563 * If we're cleaning we only write cleaning and ifile blocks, and 564 * no dirops, since otherwise we'd risk corruption in a crash. 565 */ 566 if (sp->seg_flags & SEGM_CLEAN) 567 lfs_writevnodes(fs, mp, sp, VN_CLEAN); 568 else if (!(sp->seg_flags & SEGM_FORCE_CKP)) { 569 lfs_writevnodes(fs, mp, sp, VN_REG); 570 if (!fs->lfs_dirops || !fs->lfs_flushvp) { 571 error = lfs_writer_enter(fs, "lfs writer"); 572 if (error) { 573 DLOG((DLOG_SEG, "segwrite mysterious error\n")); 574 /* XXX why not segunlock? */ 575 pool_put(&fs->lfs_bpppool, sp->bpp); 576 sp->bpp = NULL; 577 pool_put(&fs->lfs_segpool, sp); 578 sp = fs->lfs_sp = NULL; 579 return (error); 580 } 581 writer_set = 1; 582 lfs_writevnodes(fs, mp, sp, VN_DIROP); 583 ((SEGSUM *)(sp->segsum))->ss_flags &= ~(SS_CONT); 584 } 585 } 586 587 /* 588 * If we are doing a checkpoint, mark everything since the 589 * last checkpoint as no longer ACTIVE. 590 */ 591 if (do_ckp) { 592 segleft = fs->lfs_nseg; 593 curseg = 0; 594 for (n = 0; n < fs->lfs_segtabsz; n++) { 595 dirty = 0; 596 if (bread(fs->lfs_ivnode, 597 fs->lfs_cleansz + n, fs->lfs_bsize, NOCRED, &bp)) 598 panic("lfs_segwrite: ifile read"); 599 segusep = (SEGUSE *)bp->b_data; 600 maxseg = min(segleft, fs->lfs_sepb); 601 for (i = 0; i < maxseg; i++) { 602 sn = curseg + i; 603 if (sn != dtosn(fs, fs->lfs_curseg) && 604 segusep->su_flags & SEGUSE_ACTIVE) { 605 segusep->su_flags &= ~SEGUSE_ACTIVE; 606 --fs->lfs_nactive; 607 ++dirty; 608 } 609 fs->lfs_suflags[fs->lfs_activesb][sn] = 610 segusep->su_flags; 611 if (fs->lfs_version > 1) 612 ++segusep; 613 else 614 segusep = (SEGUSE *) 615 ((SEGUSE_V1 *)segusep + 1); 616 } 617 618 if (dirty) 619 error = LFS_BWRITE_LOG(bp); /* Ifile */ 620 else 621 brelse(bp); 622 segleft -= fs->lfs_sepb; 623 curseg += fs->lfs_sepb; 624 } 625 } 626 627 LOCK_ASSERT(LFS_SEGLOCK_HELD(fs)); 628 629 did_ckp = 0; 630 if (do_ckp || fs->lfs_doifile) { 631 vp = fs->lfs_ivnode; 632 vn_lock(vp, LK_EXCLUSIVE); 633 do { 634 #ifdef DEBUG 635 LFS_ENTER_LOG("pretend", __FILE__, __LINE__, 0, 0, curproc->p_pid); 636 #endif 637 simple_lock(&fs->lfs_interlock); 638 fs->lfs_flags &= ~LFS_IFDIRTY; 639 simple_unlock(&fs->lfs_interlock); 640 641 ip = VTOI(vp); 642 643 if (LIST_FIRST(&vp->v_dirtyblkhd) != NULL) 644 lfs_writefile(fs, sp, vp); 645 646 if (ip->i_flag & IN_ALLMOD) 647 ++did_ckp; 648 redo = lfs_writeinode(fs, sp, ip); 649 redo += lfs_writeseg(fs, sp); 650 simple_lock(&fs->lfs_interlock); 651 redo += (fs->lfs_flags & LFS_IFDIRTY); 652 simple_unlock(&fs->lfs_interlock); 653 } while (redo && do_ckp); 654 655 /* 656 * Unless we are unmounting, the Ifile may continue to have 657 * dirty blocks even after a checkpoint, due to changes to 658 * inodes' atime. If we're checkpointing, it's "impossible" 659 * for other parts of the Ifile to be dirty after the loop 660 * above, since we hold the segment lock. 661 */ 662 s = splbio(); 663 if (LIST_EMPTY(&vp->v_dirtyblkhd)) { 664 LFS_CLR_UINO(ip, IN_ALLMOD); 665 } 666 #ifdef DIAGNOSTIC 667 else if (do_ckp) { 668 int do_panic = 0; 669 LIST_FOREACH(bp, &vp->v_dirtyblkhd, b_vnbufs) { 670 if (bp->b_lblkno < fs->lfs_cleansz + 671 fs->lfs_segtabsz && 672 !(bp->b_flags & B_GATHERED)) { 673 printf("ifile lbn %ld still dirty (flags %lx)\n", 674 (long)bp->b_lblkno, 675 (long)bp->b_flags); 676 ++do_panic; 677 } 678 } 679 if (do_panic) 680 panic("dirty blocks"); 681 } 682 #endif 683 splx(s); 684 VOP_UNLOCK(vp, 0); 685 } else { 686 (void) lfs_writeseg(fs, sp); 687 } 688 689 /* Note Ifile no longer needs to be written */ 690 fs->lfs_doifile = 0; 691 if (writer_set) 692 lfs_writer_leave(fs); 693 694 /* 695 * If we didn't write the Ifile, we didn't really do anything. 696 * That means that (1) there is a checkpoint on disk and (2) 697 * nothing has changed since it was written. 698 * 699 * Take the flags off of the segment so that lfs_segunlock 700 * doesn't have to write the superblock either. 701 */ 702 if (do_ckp && !did_ckp) { 703 sp->seg_flags &= ~SEGM_CKP; 704 } 705 706 if (lfs_dostats) { 707 ++lfs_stats.nwrites; 708 if (sp->seg_flags & SEGM_SYNC) 709 ++lfs_stats.nsync_writes; 710 if (sp->seg_flags & SEGM_CKP) 711 ++lfs_stats.ncheckpoints; 712 } 713 lfs_segunlock(fs); 714 return (0); 715 } 716 717 /* 718 * Write the dirty blocks associated with a vnode. 719 */ 720 void 721 lfs_writefile(struct lfs *fs, struct segment *sp, struct vnode *vp) 722 { 723 struct buf *bp; 724 struct finfo *fip; 725 struct inode *ip; 726 IFILE *ifp; 727 int i, frag; 728 729 ASSERT_SEGLOCK(fs); 730 ip = VTOI(vp); 731 732 if (sp->seg_bytes_left < fs->lfs_bsize || 733 sp->sum_bytes_left < sizeof(struct finfo)) 734 (void) lfs_writeseg(fs, sp); 735 736 sp->sum_bytes_left -= FINFOSIZE; 737 ++((SEGSUM *)(sp->segsum))->ss_nfinfo; 738 739 if (vp->v_flag & VDIROP) 740 ((SEGSUM *)(sp->segsum))->ss_flags |= (SS_DIROP|SS_CONT); 741 742 fip = sp->fip; 743 fip->fi_nblocks = 0; 744 fip->fi_ino = ip->i_number; 745 LFS_IENTRY(ifp, fs, fip->fi_ino, bp); 746 fip->fi_version = ifp->if_version; 747 brelse(bp); 748 749 if (sp->seg_flags & SEGM_CLEAN) { 750 lfs_gather(fs, sp, vp, lfs_match_fake); 751 /* 752 * For a file being flushed, we need to write *all* blocks. 753 * This means writing the cleaning blocks first, and then 754 * immediately following with any non-cleaning blocks. 755 * The same is true of the Ifile since checkpoints assume 756 * that all valid Ifile blocks are written. 757 */ 758 if (IS_FLUSHING(fs,vp) || vp == fs->lfs_ivnode) { 759 lfs_gather(fs, sp, vp, lfs_match_data); 760 /* 761 * Don't call VOP_PUTPAGES: if we're flushing, 762 * we've already done it, and the Ifile doesn't 763 * use the page cache. 764 */ 765 } 766 } else { 767 lfs_gather(fs, sp, vp, lfs_match_data); 768 /* 769 * If we're flushing, we've already called VOP_PUTPAGES 770 * so don't do it again. Otherwise, we want to write 771 * everything we've got. 772 */ 773 if (!IS_FLUSHING(fs, vp)) { 774 simple_lock(&vp->v_interlock); 775 VOP_PUTPAGES(vp, 0, 0, 776 PGO_CLEANIT | PGO_ALLPAGES | PGO_LOCKED); 777 } 778 } 779 780 /* 781 * It may not be necessary to write the meta-data blocks at this point, 782 * as the roll-forward recovery code should be able to reconstruct the 783 * list. 784 * 785 * We have to write them anyway, though, under two conditions: (1) the 786 * vnode is being flushed (for reuse by vinvalbuf); or (2) we are 787 * checkpointing. 788 * 789 * BUT if we are cleaning, we might have indirect blocks that refer to 790 * new blocks not being written yet, in addition to fragments being 791 * moved out of a cleaned segment. If that is the case, don't 792 * write the indirect blocks, or the finfo will have a small block 793 * in the middle of it! 794 * XXX in this case isn't the inode size wrong too? 795 */ 796 frag = 0; 797 if (sp->seg_flags & SEGM_CLEAN) { 798 for (i = 0; i < NDADDR; i++) 799 if (ip->i_lfs_fragsize[i] > 0 && 800 ip->i_lfs_fragsize[i] < fs->lfs_bsize) 801 ++frag; 802 } 803 #ifdef DIAGNOSTIC 804 if (frag > 1) 805 panic("lfs_writefile: more than one fragment!"); 806 #endif 807 if (IS_FLUSHING(fs, vp) || 808 (frag == 0 && (lfs_writeindir || (sp->seg_flags & SEGM_CKP)))) { 809 lfs_gather(fs, sp, vp, lfs_match_indir); 810 lfs_gather(fs, sp, vp, lfs_match_dindir); 811 lfs_gather(fs, sp, vp, lfs_match_tindir); 812 } 813 fip = sp->fip; 814 if (fip->fi_nblocks != 0) { 815 sp->fip = (FINFO*)((caddr_t)fip + FINFOSIZE + 816 sizeof(int32_t) * (fip->fi_nblocks)); 817 sp->start_lbp = &sp->fip->fi_blocks[0]; 818 } else { 819 sp->sum_bytes_left += FINFOSIZE; 820 --((SEGSUM *)(sp->segsum))->ss_nfinfo; 821 } 822 } 823 824 int 825 lfs_writeinode(struct lfs *fs, struct segment *sp, struct inode *ip) 826 { 827 struct buf *bp, *ibp; 828 struct ufs1_dinode *cdp; 829 IFILE *ifp; 830 SEGUSE *sup; 831 daddr_t daddr; 832 int32_t *daddrp; /* XXX ondisk32 */ 833 ino_t ino; 834 int error, i, ndx, fsb = 0; 835 int redo_ifile = 0; 836 int gotblk = 0; 837 838 ASSERT_SEGLOCK(fs); 839 if (!(ip->i_flag & IN_ALLMOD)) 840 return (0); 841 842 /* Allocate a new inode block if necessary. */ 843 if ((ip->i_number != LFS_IFILE_INUM || sp->idp == NULL) && 844 sp->ibp == NULL) { 845 /* Allocate a new segment if necessary. */ 846 if (sp->seg_bytes_left < fs->lfs_ibsize || 847 sp->sum_bytes_left < sizeof(int32_t)) 848 (void) lfs_writeseg(fs, sp); 849 850 /* Get next inode block. */ 851 daddr = fs->lfs_offset; 852 fs->lfs_offset += btofsb(fs, fs->lfs_ibsize); 853 sp->ibp = *sp->cbpp++ = 854 getblk(VTOI(fs->lfs_ivnode)->i_devvp, 855 fsbtodb(fs, daddr), fs->lfs_ibsize, 0, 0); 856 gotblk++; 857 858 /* Zero out inode numbers */ 859 for (i = 0; i < INOPB(fs); ++i) 860 ((struct ufs1_dinode *)sp->ibp->b_data)[i].di_inumber = 861 0; 862 863 ++sp->start_bpp; 864 fs->lfs_avail -= btofsb(fs, fs->lfs_ibsize); 865 /* Set remaining space counters. */ 866 sp->seg_bytes_left -= fs->lfs_ibsize; 867 sp->sum_bytes_left -= sizeof(int32_t); 868 ndx = fs->lfs_sumsize / sizeof(int32_t) - 869 sp->ninodes / INOPB(fs) - 1; 870 ((int32_t *)(sp->segsum))[ndx] = daddr; 871 } 872 873 /* Update the inode times and copy the inode onto the inode page. */ 874 /* XXX kludge --- don't redirty the ifile just to put times on it */ 875 if (ip->i_number != LFS_IFILE_INUM) 876 LFS_ITIMES(ip, NULL, NULL, NULL); 877 878 /* 879 * If this is the Ifile, and we've already written the Ifile in this 880 * partial segment, just overwrite it (it's not on disk yet) and 881 * continue. 882 * 883 * XXX we know that the bp that we get the second time around has 884 * already been gathered. 885 */ 886 if (ip->i_number == LFS_IFILE_INUM && sp->idp) { 887 *(sp->idp) = *ip->i_din.ffs1_din; 888 ip->i_lfs_osize = ip->i_size; 889 return 0; 890 } 891 892 bp = sp->ibp; 893 cdp = ((struct ufs1_dinode *)bp->b_data) + (sp->ninodes % INOPB(fs)); 894 *cdp = *ip->i_din.ffs1_din; 895 896 /* 897 * If we are cleaning, ensure that we don't write UNWRITTEN disk 898 * addresses to disk; possibly change the on-disk record of 899 * the inode size, either by reverting to the previous size 900 * (in the case of cleaning) or by verifying the inode's block 901 * holdings (in the case of files being allocated as they are being 902 * written). 903 * XXX By not writing UNWRITTEN blocks, we are making the lfs_avail 904 * XXX count on disk wrong by the same amount. We should be 905 * XXX able to "borrow" from lfs_avail and return it after the 906 * XXX Ifile is written. See also in lfs_writeseg. 907 */ 908 909 /* Check file size based on highest allocated block */ 910 if (((ip->i_ffs1_mode & IFMT) == IFREG || 911 (ip->i_ffs1_mode & IFMT) == IFDIR) && 912 ip->i_size > ((ip->i_lfs_hiblk + 1) << fs->lfs_bshift)) { 913 cdp->di_size = (ip->i_lfs_hiblk + 1) << fs->lfs_bshift; 914 DLOG((DLOG_SEG, "lfs_writeinode: ino %d size %" PRId64 " -> %" 915 PRId64 "\n", (int)ip->i_number, ip->i_size, cdp->di_size)); 916 } 917 if (ip->i_lfs_effnblks != ip->i_ffs1_blocks) { 918 if (ip->i_flags & IN_CLEANING) 919 cdp->di_size = ip->i_lfs_osize; 920 DLOG((DLOG_SEG, "lfs_writeinode: cleansing ino %d eff %d != nblk %d)" 921 " at %x\n", ip->i_number, ip->i_lfs_effnblks, 922 ip->i_ffs1_blocks, fs->lfs_offset)); 923 for (daddrp = cdp->di_db; daddrp < cdp->di_ib + NIADDR; 924 daddrp++) { 925 if (*daddrp == UNWRITTEN) { 926 DLOG((DLOG_SEG, "lfs_writeinode: wiping UNWRITTEN\n")); 927 *daddrp = 0; 928 } 929 } 930 } else { 931 /* If all blocks are going to disk, update "size on disk" */ 932 ip->i_lfs_osize = ip->i_size; 933 } 934 935 #ifdef DIAGNOSTIC 936 /* 937 * Check dinode held blocks against dinode size. 938 * This should be identical to the check in lfs_vget(). 939 */ 940 for (i = (cdp->di_size + fs->lfs_bsize - 1) >> fs->lfs_bshift; 941 i < NDADDR; i++) { 942 KASSERT(i >= 0); 943 if ((cdp->di_mode & IFMT) == IFLNK) 944 continue; 945 if (((cdp->di_mode & IFMT) == IFBLK || 946 (cdp->di_mode & IFMT) == IFCHR) && i == 0) 947 continue; 948 if (cdp->di_db[i] != 0) { 949 # ifdef DEBUG 950 lfs_dump_dinode(cdp); 951 # endif 952 panic("writing inconsistent inode"); 953 } 954 } 955 #endif /* DIAGNOSTIC */ 956 957 if (ip->i_flag & IN_CLEANING) 958 LFS_CLR_UINO(ip, IN_CLEANING); 959 else { 960 /* XXX IN_ALLMOD */ 961 LFS_CLR_UINO(ip, IN_ACCESSED | IN_ACCESS | IN_CHANGE | 962 IN_UPDATE | IN_MODIFY); 963 if (ip->i_lfs_effnblks == ip->i_ffs1_blocks) 964 LFS_CLR_UINO(ip, IN_MODIFIED); 965 else 966 DLOG((DLOG_VNODE, "lfs_writeinode: ino %d: real blks=%d, " 967 "eff=%d\n", ip->i_number, ip->i_ffs1_blocks, 968 ip->i_lfs_effnblks)); 969 } 970 971 if (ip->i_number == LFS_IFILE_INUM) /* We know sp->idp == NULL */ 972 sp->idp = ((struct ufs1_dinode *)bp->b_data) + 973 (sp->ninodes % INOPB(fs)); 974 if (gotblk) { 975 LFS_LOCK_BUF(bp); 976 brelse(bp); 977 } 978 979 /* Increment inode count in segment summary block. */ 980 ++((SEGSUM *)(sp->segsum))->ss_ninos; 981 982 /* If this page is full, set flag to allocate a new page. */ 983 if (++sp->ninodes % INOPB(fs) == 0) 984 sp->ibp = NULL; 985 986 /* 987 * If updating the ifile, update the super-block. Update the disk 988 * address and access times for this inode in the ifile. 989 */ 990 ino = ip->i_number; 991 if (ino == LFS_IFILE_INUM) { 992 daddr = fs->lfs_idaddr; 993 fs->lfs_idaddr = dbtofsb(fs, bp->b_blkno); 994 } else { 995 LFS_IENTRY(ifp, fs, ino, ibp); 996 daddr = ifp->if_daddr; 997 ifp->if_daddr = dbtofsb(fs, bp->b_blkno) + fsb; 998 error = LFS_BWRITE_LOG(ibp); /* Ifile */ 999 } 1000 1001 /* 1002 * The inode's last address should not be in the current partial 1003 * segment, except under exceptional circumstances (lfs_writevnodes 1004 * had to start over, and in the meantime more blocks were written 1005 * to a vnode). Both inodes will be accounted to this segment 1006 * in lfs_writeseg so we need to subtract the earlier version 1007 * here anyway. The segment count can temporarily dip below 1008 * zero here; keep track of how many duplicates we have in 1009 * "dupino" so we don't panic below. 1010 */ 1011 if (daddr >= fs->lfs_lastpseg && daddr <= dbtofsb(fs, bp->b_blkno)) { 1012 ++sp->ndupino; 1013 DLOG((DLOG_SEG, "lfs_writeinode: last inode addr in current pseg " 1014 "(ino %d daddr 0x%llx) ndupino=%d\n", ino, 1015 (long long)daddr, sp->ndupino)); 1016 } 1017 /* 1018 * Account the inode: it no longer belongs to its former segment, 1019 * though it will not belong to the new segment until that segment 1020 * is actually written. 1021 */ 1022 if (daddr != LFS_UNUSED_DADDR) { 1023 u_int32_t oldsn = dtosn(fs, daddr); 1024 #ifdef DIAGNOSTIC 1025 int ndupino = (sp->seg_number == oldsn) ? sp->ndupino : 0; 1026 #endif 1027 LFS_SEGENTRY(sup, fs, oldsn, bp); 1028 #ifdef DIAGNOSTIC 1029 if (sup->su_nbytes + 1030 sizeof (struct ufs1_dinode) * ndupino 1031 < sizeof (struct ufs1_dinode)) { 1032 printf("lfs_writeinode: negative bytes " 1033 "(segment %" PRIu32 " short by %d, " 1034 "oldsn=%" PRIu32 ", cursn=%" PRIu32 1035 ", daddr=%" PRId64 ", su_nbytes=%u, " 1036 "ndupino=%d)\n", 1037 dtosn(fs, daddr), 1038 (int)sizeof (struct ufs1_dinode) * 1039 (1 - sp->ndupino) - sup->su_nbytes, 1040 oldsn, sp->seg_number, daddr, 1041 (unsigned int)sup->su_nbytes, 1042 sp->ndupino); 1043 panic("lfs_writeinode: negative bytes"); 1044 sup->su_nbytes = sizeof (struct ufs1_dinode); 1045 } 1046 #endif 1047 DLOG((DLOG_SU, "seg %d -= %d for ino %d inode\n", 1048 dtosn(fs, daddr), sizeof (struct ufs1_dinode), ino)); 1049 sup->su_nbytes -= sizeof (struct ufs1_dinode); 1050 redo_ifile = 1051 (ino == LFS_IFILE_INUM && !(bp->b_flags & B_GATHERED)); 1052 if (redo_ifile) { 1053 simple_lock(&fs->lfs_interlock); 1054 fs->lfs_flags |= LFS_IFDIRTY; 1055 simple_unlock(&fs->lfs_interlock); 1056 } 1057 LFS_WRITESEGENTRY(sup, fs, oldsn, bp); /* Ifile */ 1058 } 1059 return (redo_ifile); 1060 } 1061 1062 int 1063 lfs_gatherblock(struct segment *sp, struct buf *bp, int *sptr) 1064 { 1065 struct lfs *fs; 1066 int vers; 1067 int j, blksinblk; 1068 1069 ASSERT_SEGLOCK(sp->fs); 1070 /* 1071 * If full, finish this segment. We may be doing I/O, so 1072 * release and reacquire the splbio(). 1073 */ 1074 #ifdef DIAGNOSTIC 1075 if (sp->vp == NULL) 1076 panic ("lfs_gatherblock: Null vp in segment"); 1077 #endif 1078 fs = sp->fs; 1079 blksinblk = howmany(bp->b_bcount, fs->lfs_bsize); 1080 if (sp->sum_bytes_left < sizeof(int32_t) * blksinblk || 1081 sp->seg_bytes_left < bp->b_bcount) { 1082 if (sptr) 1083 splx(*sptr); 1084 lfs_updatemeta(sp); 1085 1086 vers = sp->fip->fi_version; 1087 (void) lfs_writeseg(fs, sp); 1088 1089 sp->fip->fi_version = vers; 1090 sp->fip->fi_ino = VTOI(sp->vp)->i_number; 1091 /* Add the current file to the segment summary. */ 1092 ++((SEGSUM *)(sp->segsum))->ss_nfinfo; 1093 sp->sum_bytes_left -= FINFOSIZE; 1094 1095 if (sptr) 1096 *sptr = splbio(); 1097 return (1); 1098 } 1099 1100 if (bp->b_flags & B_GATHERED) { 1101 DLOG((DLOG_SEG, "lfs_gatherblock: already gathered! Ino %d," 1102 " lbn %" PRId64 "\n", 1103 sp->fip->fi_ino, bp->b_lblkno)); 1104 return (0); 1105 } 1106 1107 /* Insert into the buffer list, update the FINFO block. */ 1108 bp->b_flags |= B_GATHERED; 1109 1110 *sp->cbpp++ = bp; 1111 for (j = 0; j < blksinblk; j++) { 1112 sp->fip->fi_blocks[sp->fip->fi_nblocks++] = bp->b_lblkno + j; 1113 /* This block's accounting moves from lfs_favail to lfs_avail */ 1114 lfs_deregister_block(sp->vp, bp->b_lblkno + j); 1115 } 1116 1117 sp->sum_bytes_left -= sizeof(int32_t) * blksinblk; 1118 sp->seg_bytes_left -= bp->b_bcount; 1119 return (0); 1120 } 1121 1122 int 1123 lfs_gather(struct lfs *fs, struct segment *sp, struct vnode *vp, 1124 int (*match)(struct lfs *, struct buf *)) 1125 { 1126 struct buf *bp, *nbp; 1127 int s, count = 0; 1128 1129 ASSERT_SEGLOCK(fs); 1130 KASSERT(sp->vp == NULL); 1131 sp->vp = vp; 1132 s = splbio(); 1133 1134 #ifndef LFS_NO_BACKBUF_HACK 1135 /* This is a hack to see if ordering the blocks in LFS makes a difference. */ 1136 # define BUF_OFFSET \ 1137 (((caddr_t)&LIST_NEXT(bp, b_vnbufs)) - (caddr_t)bp) 1138 # define BACK_BUF(BP) \ 1139 ((struct buf *)(((caddr_t)(BP)->b_vnbufs.le_prev) - BUF_OFFSET)) 1140 # define BEG_OF_LIST \ 1141 ((struct buf *)(((caddr_t)&LIST_FIRST(&vp->v_dirtyblkhd)) - BUF_OFFSET)) 1142 1143 loop: 1144 /* Find last buffer. */ 1145 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); 1146 bp && LIST_NEXT(bp, b_vnbufs) != NULL; 1147 bp = LIST_NEXT(bp, b_vnbufs)) 1148 /* nothing */; 1149 for (; bp && bp != BEG_OF_LIST; bp = nbp) { 1150 nbp = BACK_BUF(bp); 1151 #else /* LFS_NO_BACKBUF_HACK */ 1152 loop: 1153 for (bp = LIST_FIRST(&vp->v_dirtyblkhd); bp; bp = nbp) { 1154 nbp = LIST_NEXT(bp, b_vnbufs); 1155 #endif /* LFS_NO_BACKBUF_HACK */ 1156 if ((bp->b_flags & (B_BUSY|B_GATHERED)) || !match(fs, bp)) { 1157 #ifdef DEBUG 1158 if (vp == fs->lfs_ivnode && 1159 (bp->b_flags & (B_BUSY|B_GATHERED)) == B_BUSY) 1160 DLOG((DLOG_SEG, "lfs_gather: ifile lbn %" 1161 PRId64 " busy (%x)", 1162 bp->b_lblkno, bp->b_flags)); 1163 #endif 1164 continue; 1165 } 1166 if (vp->v_type == VBLK) { 1167 /* For block devices, just write the blocks. */ 1168 /* XXX Do we even need to do this? */ 1169 /* 1170 * Get the block before bwrite, 1171 * so we don't corrupt the free list 1172 */ 1173 bp->b_flags |= B_BUSY; 1174 bremfree(bp); 1175 bwrite(bp); 1176 } else { 1177 #ifdef DIAGNOSTIC 1178 # ifdef LFS_USE_B_INVAL 1179 if ((bp->b_flags & (B_CALL|B_INVAL)) == B_INVAL) { 1180 DLOG((DLOG_SEG, "lfs_gather: lbn %" PRId64 1181 " is B_INVAL\n", bp->b_lblkno)); 1182 VOP_PRINT(bp->b_vp); 1183 } 1184 # endif /* LFS_USE_B_INVAL */ 1185 if (!(bp->b_flags & B_DELWRI)) 1186 panic("lfs_gather: bp not B_DELWRI"); 1187 if (!(bp->b_flags & B_LOCKED)) { 1188 DLOG((DLOG_SEG, "lfs_gather: lbn %" PRId64 1189 " blk %" PRId64 " not B_LOCKED\n", 1190 bp->b_lblkno, 1191 dbtofsb(fs, bp->b_blkno))); 1192 VOP_PRINT(bp->b_vp); 1193 panic("lfs_gather: bp not B_LOCKED"); 1194 } 1195 #endif 1196 if (lfs_gatherblock(sp, bp, &s)) { 1197 goto loop; 1198 } 1199 } 1200 count++; 1201 } 1202 splx(s); 1203 lfs_updatemeta(sp); 1204 KASSERT(sp->vp == vp); 1205 sp->vp = NULL; 1206 return count; 1207 } 1208 1209 #if DEBUG 1210 # define DEBUG_OOFF(n) do { \ 1211 if (ooff == 0) { \ 1212 DLOG((DLOG_SEG, "lfs_updatemeta[%d]: warning: writing " \ 1213 "ino %d lbn %" PRId64 " at 0x%" PRIx32 \ 1214 ", was 0x0 (or %" PRId64 ")\n", \ 1215 (n), ip->i_number, lbn, ndaddr, daddr)); \ 1216 } \ 1217 } while (0) 1218 #else 1219 # define DEBUG_OOFF(n) 1220 #endif 1221 1222 /* 1223 * Change the given block's address to ndaddr, finding its previous 1224 * location using ufs_bmaparray(). 1225 * 1226 * Account for this change in the segment table. 1227 * 1228 * called with sp == NULL by roll-forwarding code. 1229 */ 1230 void 1231 lfs_update_single(struct lfs *fs, struct segment *sp, struct vnode *vp, 1232 daddr_t lbn, int32_t ndaddr, int size) 1233 { 1234 SEGUSE *sup; 1235 struct buf *bp; 1236 struct indir a[NIADDR + 2], *ap; 1237 struct inode *ip; 1238 daddr_t daddr, ooff; 1239 int num, error; 1240 int bb, osize, obb; 1241 1242 ASSERT_SEGLOCK(fs); 1243 KASSERT(sp == NULL || sp->vp == vp); 1244 ip = VTOI(vp); 1245 1246 error = ufs_bmaparray(vp, lbn, &daddr, a, &num, NULL, NULL); 1247 if (error) 1248 panic("lfs_updatemeta: ufs_bmaparray returned %d", error); 1249 1250 daddr = (daddr_t)((int32_t)daddr); /* XXX ondisk32 */ 1251 KASSERT(daddr <= LFS_MAX_DADDR); 1252 if (daddr > 0) 1253 daddr = dbtofsb(fs, daddr); 1254 1255 bb = fragstofsb(fs, numfrags(fs, size)); 1256 switch (num) { 1257 case 0: 1258 ooff = ip->i_ffs1_db[lbn]; 1259 DEBUG_OOFF(0); 1260 if (ooff == UNWRITTEN) 1261 ip->i_ffs1_blocks += bb; 1262 else { 1263 /* possible fragment truncation or extension */ 1264 obb = btofsb(fs, ip->i_lfs_fragsize[lbn]); 1265 ip->i_ffs1_blocks += (bb - obb); 1266 } 1267 ip->i_ffs1_db[lbn] = ndaddr; 1268 break; 1269 case 1: 1270 ooff = ip->i_ffs1_ib[a[0].in_off]; 1271 DEBUG_OOFF(1); 1272 if (ooff == UNWRITTEN) 1273 ip->i_ffs1_blocks += bb; 1274 ip->i_ffs1_ib[a[0].in_off] = ndaddr; 1275 break; 1276 default: 1277 ap = &a[num - 1]; 1278 if (bread(vp, ap->in_lbn, fs->lfs_bsize, NOCRED, &bp)) 1279 panic("lfs_updatemeta: bread bno %" PRId64, 1280 ap->in_lbn); 1281 1282 /* XXX ondisk32 */ 1283 ooff = ((int32_t *)bp->b_data)[ap->in_off]; 1284 DEBUG_OOFF(num); 1285 if (ooff == UNWRITTEN) 1286 ip->i_ffs1_blocks += bb; 1287 /* XXX ondisk32 */ 1288 ((int32_t *)bp->b_data)[ap->in_off] = ndaddr; 1289 (void) VOP_BWRITE(bp); 1290 } 1291 1292 KASSERT(ooff == 0 || ooff == UNWRITTEN || ooff == daddr); 1293 1294 /* Update hiblk when extending the file */ 1295 if (lbn > ip->i_lfs_hiblk) 1296 ip->i_lfs_hiblk = lbn; 1297 1298 /* 1299 * Though we'd rather it couldn't, this *can* happen right now 1300 * if cleaning blocks and regular blocks coexist. 1301 */ 1302 /* KASSERT(daddr < fs->lfs_lastpseg || daddr > ndaddr); */ 1303 1304 /* 1305 * Update segment usage information, based on old size 1306 * and location. 1307 */ 1308 if (daddr > 0) { 1309 u_int32_t oldsn = dtosn(fs, daddr); 1310 #ifdef DIAGNOSTIC 1311 int ndupino; 1312 1313 if (sp && sp->seg_number == oldsn) { 1314 ndupino = sp->ndupino; 1315 } else { 1316 ndupino = 0; 1317 } 1318 #endif 1319 KASSERT(oldsn >= 0 && oldsn < fs->lfs_nseg); 1320 if (lbn >= 0 && lbn < NDADDR) 1321 osize = ip->i_lfs_fragsize[lbn]; 1322 else 1323 osize = fs->lfs_bsize; 1324 LFS_SEGENTRY(sup, fs, oldsn, bp); 1325 #ifdef DIAGNOSTIC 1326 if (sup->su_nbytes + sizeof (struct ufs1_dinode) * ndupino 1327 < osize) { 1328 printf("lfs_updatemeta: negative bytes " 1329 "(segment %" PRIu32 " short by %" PRId64 1330 ")\n", dtosn(fs, daddr), 1331 (int64_t)osize - 1332 (sizeof (struct ufs1_dinode) * ndupino + 1333 sup->su_nbytes)); 1334 printf("lfs_updatemeta: ino %llu, lbn %" PRId64 1335 ", addr = 0x%" PRIx64 "\n", 1336 (unsigned long long)ip->i_number, lbn, daddr); 1337 printf("lfs_updatemeta: ndupino=%d\n", ndupino); 1338 panic("lfs_updatemeta: negative bytes"); 1339 sup->su_nbytes = osize - 1340 sizeof (struct ufs1_dinode) * ndupino; 1341 } 1342 #endif 1343 DLOG((DLOG_SU, "seg %" PRIu32 " -= %d for ino %d lbn %" PRId64 1344 " db 0x%" PRIx64 "\n", 1345 dtosn(fs, daddr), osize, 1346 ip->i_number, lbn, daddr)); 1347 sup->su_nbytes -= osize; 1348 if (!(bp->b_flags & B_GATHERED)) { 1349 simple_lock(&fs->lfs_interlock); 1350 fs->lfs_flags |= LFS_IFDIRTY; 1351 simple_unlock(&fs->lfs_interlock); 1352 } 1353 LFS_WRITESEGENTRY(sup, fs, oldsn, bp); 1354 } 1355 /* 1356 * Now that this block has a new address, and its old 1357 * segment no longer owns it, we can forget about its 1358 * old size. 1359 */ 1360 if (lbn >= 0 && lbn < NDADDR) 1361 ip->i_lfs_fragsize[lbn] = size; 1362 } 1363 1364 /* 1365 * Update the metadata that points to the blocks listed in the FINFO 1366 * array. 1367 */ 1368 void 1369 lfs_updatemeta(struct segment *sp) 1370 { 1371 struct buf *sbp; 1372 struct lfs *fs; 1373 struct vnode *vp; 1374 daddr_t lbn; 1375 int i, nblocks, num; 1376 int bb; 1377 int bytesleft, size; 1378 1379 ASSERT_SEGLOCK(sp->fs); 1380 vp = sp->vp; 1381 nblocks = &sp->fip->fi_blocks[sp->fip->fi_nblocks] - sp->start_lbp; 1382 KASSERT(nblocks >= 0); 1383 KASSERT(vp != NULL); 1384 if (nblocks == 0) 1385 return; 1386 1387 /* 1388 * This count may be high due to oversize blocks from lfs_gop_write. 1389 * Correct for this. (XXX we should be able to keep track of these.) 1390 */ 1391 fs = sp->fs; 1392 for (i = 0; i < nblocks; i++) { 1393 if (sp->start_bpp[i] == NULL) { 1394 DLOG((DLOG_SEG, "lfs_updatemeta: nblocks = %d, not %d\n", i, nblocks)); 1395 nblocks = i; 1396 break; 1397 } 1398 num = howmany(sp->start_bpp[i]->b_bcount, fs->lfs_bsize); 1399 KASSERT(sp->start_bpp[i]->b_lblkno >= 0 || num == 1); 1400 nblocks -= num - 1; 1401 } 1402 1403 KASSERT(vp->v_type == VREG || 1404 nblocks == &sp->fip->fi_blocks[sp->fip->fi_nblocks] - sp->start_lbp); 1405 KASSERT(nblocks == sp->cbpp - sp->start_bpp); 1406 1407 /* 1408 * Sort the blocks. 1409 * 1410 * We have to sort even if the blocks come from the 1411 * cleaner, because there might be other pending blocks on the 1412 * same inode...and if we don't sort, and there are fragments 1413 * present, blocks may be written in the wrong place. 1414 */ 1415 lfs_shellsort(sp->start_bpp, sp->start_lbp, nblocks, fs->lfs_bsize); 1416 1417 /* 1418 * Record the length of the last block in case it's a fragment. 1419 * If there are indirect blocks present, they sort last. An 1420 * indirect block will be lfs_bsize and its presence indicates 1421 * that you cannot have fragments. 1422 * 1423 * XXX This last is a lie. A cleaned fragment can coexist with 1424 * XXX a later indirect block. This will continue to be 1425 * XXX true until lfs_markv is fixed to do everything with 1426 * XXX fake blocks (including fake inodes and fake indirect blocks). 1427 */ 1428 sp->fip->fi_lastlength = ((sp->start_bpp[nblocks - 1]->b_bcount - 1) & 1429 fs->lfs_bmask) + 1; 1430 1431 /* 1432 * Assign disk addresses, and update references to the logical 1433 * block and the segment usage information. 1434 */ 1435 for (i = nblocks; i--; ++sp->start_bpp) { 1436 sbp = *sp->start_bpp; 1437 lbn = *sp->start_lbp; 1438 KASSERT(sbp->b_lblkno == lbn); 1439 1440 sbp->b_blkno = fsbtodb(fs, fs->lfs_offset); 1441 1442 /* 1443 * If we write a frag in the wrong place, the cleaner won't 1444 * be able to correctly identify its size later, and the 1445 * segment will be uncleanable. (Even worse, it will assume 1446 * that the indirect block that actually ends the list 1447 * is of a smaller size!) 1448 */ 1449 if ((sbp->b_bcount & fs->lfs_bmask) && i != 0) 1450 panic("lfs_updatemeta: fragment is not last block"); 1451 1452 /* 1453 * For each subblock in this possibly oversized block, 1454 * update its address on disk. 1455 */ 1456 KASSERT(lbn >= 0 || sbp->b_bcount == fs->lfs_bsize); 1457 KASSERT(vp == sbp->b_vp); 1458 for (bytesleft = sbp->b_bcount; bytesleft > 0; 1459 bytesleft -= fs->lfs_bsize) { 1460 size = MIN(bytesleft, fs->lfs_bsize); 1461 bb = fragstofsb(fs, numfrags(fs, size)); 1462 lbn = *sp->start_lbp++; 1463 lfs_update_single(fs, sp, sp->vp, lbn, fs->lfs_offset, 1464 size); 1465 fs->lfs_offset += bb; 1466 } 1467 1468 } 1469 } 1470 1471 /* 1472 * Move lfs_offset to a segment earlier than sn. 1473 */ 1474 int 1475 lfs_rewind(struct lfs *fs, int newsn) 1476 { 1477 int sn, osn, isdirty; 1478 struct buf *bp; 1479 SEGUSE *sup; 1480 1481 ASSERT_SEGLOCK(fs); 1482 1483 osn = dtosn(fs, fs->lfs_offset); 1484 if (osn < newsn) 1485 return 0; 1486 1487 /* lfs_avail eats the remaining space in this segment */ 1488 fs->lfs_avail -= fs->lfs_fsbpseg - (fs->lfs_offset - fs->lfs_curseg); 1489 1490 /* Find a low-numbered segment */ 1491 for (sn = 0; sn < fs->lfs_nseg; ++sn) { 1492 LFS_SEGENTRY(sup, fs, sn, bp); 1493 isdirty = sup->su_flags & SEGUSE_DIRTY; 1494 brelse(bp); 1495 1496 if (!isdirty) 1497 break; 1498 } 1499 if (sn == fs->lfs_nseg) 1500 panic("lfs_rewind: no clean segments"); 1501 if (sn >= newsn) 1502 return ENOENT; 1503 fs->lfs_nextseg = sn; 1504 lfs_newseg(fs); 1505 fs->lfs_offset = fs->lfs_curseg; 1506 1507 return 0; 1508 } 1509 1510 /* 1511 * Start a new partial segment. 1512 * 1513 * Return 1 when we entered to a new segment. 1514 * Otherwise, return 0. 1515 */ 1516 int 1517 lfs_initseg(struct lfs *fs) 1518 { 1519 struct segment *sp = fs->lfs_sp; 1520 SEGSUM *ssp; 1521 struct buf *sbp; /* buffer for SEGSUM */ 1522 int repeat = 0; /* return value */ 1523 1524 ASSERT_SEGLOCK(fs); 1525 /* Advance to the next segment. */ 1526 if (!LFS_PARTIAL_FITS(fs)) { 1527 SEGUSE *sup; 1528 struct buf *bp; 1529 1530 /* lfs_avail eats the remaining space */ 1531 fs->lfs_avail -= fs->lfs_fsbpseg - (fs->lfs_offset - 1532 fs->lfs_curseg); 1533 /* Wake up any cleaning procs waiting on this file system. */ 1534 wakeup(&lfs_allclean_wakeup); 1535 wakeup(&fs->lfs_nextseg); 1536 lfs_newseg(fs); 1537 repeat = 1; 1538 fs->lfs_offset = fs->lfs_curseg; 1539 1540 sp->seg_number = dtosn(fs, fs->lfs_curseg); 1541 sp->seg_bytes_left = fsbtob(fs, fs->lfs_fsbpseg); 1542 1543 /* 1544 * If the segment contains a superblock, update the offset 1545 * and summary address to skip over it. 1546 */ 1547 LFS_SEGENTRY(sup, fs, sp->seg_number, bp); 1548 if (sup->su_flags & SEGUSE_SUPERBLOCK) { 1549 fs->lfs_offset += btofsb(fs, LFS_SBPAD); 1550 sp->seg_bytes_left -= LFS_SBPAD; 1551 } 1552 brelse(bp); 1553 /* Segment zero could also contain the labelpad */ 1554 if (fs->lfs_version > 1 && sp->seg_number == 0 && 1555 fs->lfs_start < btofsb(fs, LFS_LABELPAD)) { 1556 fs->lfs_offset += 1557 btofsb(fs, LFS_LABELPAD) - fs->lfs_start; 1558 sp->seg_bytes_left -= 1559 LFS_LABELPAD - fsbtob(fs, fs->lfs_start); 1560 } 1561 } else { 1562 sp->seg_number = dtosn(fs, fs->lfs_curseg); 1563 sp->seg_bytes_left = fsbtob(fs, fs->lfs_fsbpseg - 1564 (fs->lfs_offset - fs->lfs_curseg)); 1565 } 1566 fs->lfs_lastpseg = fs->lfs_offset; 1567 1568 /* Record first address of this partial segment */ 1569 if (sp->seg_flags & SEGM_CLEAN) { 1570 fs->lfs_cleanint[fs->lfs_cleanind] = fs->lfs_offset; 1571 if (++fs->lfs_cleanind >= LFS_MAX_CLEANIND) { 1572 /* "1" is the artificial inc in lfs_seglock */ 1573 simple_lock(&fs->lfs_interlock); 1574 while (fs->lfs_iocount > 1) { 1575 ltsleep(&fs->lfs_iocount, PRIBIO + 1, 1576 "lfs_initseg", 0, &fs->lfs_interlock); 1577 } 1578 simple_unlock(&fs->lfs_interlock); 1579 fs->lfs_cleanind = 0; 1580 } 1581 } 1582 1583 sp->fs = fs; 1584 sp->ibp = NULL; 1585 sp->idp = NULL; 1586 sp->ninodes = 0; 1587 sp->ndupino = 0; 1588 1589 sp->cbpp = sp->bpp; 1590 1591 /* Get a new buffer for SEGSUM */ 1592 sbp = lfs_newbuf(fs, VTOI(fs->lfs_ivnode)->i_devvp, 1593 fsbtodb(fs, fs->lfs_offset), fs->lfs_sumsize, LFS_NB_SUMMARY); 1594 1595 /* ... and enter it into the buffer list. */ 1596 *sp->cbpp = sbp; 1597 sp->cbpp++; 1598 fs->lfs_offset += btofsb(fs, fs->lfs_sumsize); 1599 1600 sp->start_bpp = sp->cbpp; 1601 1602 /* Set point to SEGSUM, initialize it. */ 1603 ssp = sp->segsum = sbp->b_data; 1604 memset(ssp, 0, fs->lfs_sumsize); 1605 ssp->ss_next = fs->lfs_nextseg; 1606 ssp->ss_nfinfo = ssp->ss_ninos = 0; 1607 ssp->ss_magic = SS_MAGIC; 1608 1609 /* Set pointer to first FINFO, initialize it. */ 1610 sp->fip = (struct finfo *)((caddr_t)sp->segsum + SEGSUM_SIZE(fs)); 1611 sp->fip->fi_nblocks = 0; 1612 sp->start_lbp = &sp->fip->fi_blocks[0]; 1613 sp->fip->fi_lastlength = 0; 1614 1615 sp->seg_bytes_left -= fs->lfs_sumsize; 1616 sp->sum_bytes_left = fs->lfs_sumsize - SEGSUM_SIZE(fs); 1617 1618 return (repeat); 1619 } 1620 1621 /* 1622 * Remove SEGUSE_INVAL from all segments. 1623 */ 1624 void 1625 lfs_unset_inval_all(struct lfs *fs) 1626 { 1627 SEGUSE *sup; 1628 struct buf *bp; 1629 int i; 1630 1631 for (i = 0; i < fs->lfs_nseg; i++) { 1632 LFS_SEGENTRY(sup, fs, i, bp); 1633 if (sup->su_flags & SEGUSE_INVAL) { 1634 sup->su_flags &= ~SEGUSE_INVAL; 1635 VOP_BWRITE(bp); 1636 } else 1637 brelse(bp); 1638 } 1639 } 1640 1641 /* 1642 * Return the next segment to write. 1643 */ 1644 void 1645 lfs_newseg(struct lfs *fs) 1646 { 1647 CLEANERINFO *cip; 1648 SEGUSE *sup; 1649 struct buf *bp; 1650 int curseg, isdirty, sn, skip_inval; 1651 1652 ASSERT_SEGLOCK(fs); 1653 LFS_SEGENTRY(sup, fs, dtosn(fs, fs->lfs_nextseg), bp); 1654 DLOG((DLOG_SU, "lfs_newseg: seg %d := 0 in newseg\n", 1655 dtosn(fs, fs->lfs_nextseg))); 1656 sup->su_flags |= SEGUSE_DIRTY | SEGUSE_ACTIVE; 1657 sup->su_nbytes = 0; 1658 sup->su_nsums = 0; 1659 sup->su_ninos = 0; 1660 LFS_WRITESEGENTRY(sup, fs, dtosn(fs, fs->lfs_nextseg), bp); 1661 1662 LFS_CLEANERINFO(cip, fs, bp); 1663 --cip->clean; 1664 ++cip->dirty; 1665 fs->lfs_nclean = cip->clean; 1666 LFS_SYNC_CLEANERINFO(cip, fs, bp, 1); 1667 1668 fs->lfs_lastseg = fs->lfs_curseg; 1669 fs->lfs_curseg = fs->lfs_nextseg; 1670 skip_inval = 1; 1671 for (sn = curseg = dtosn(fs, fs->lfs_curseg) + fs->lfs_interleave;;) { 1672 sn = (sn + 1) % fs->lfs_nseg; 1673 if (sn == curseg) { 1674 if (skip_inval) 1675 skip_inval = 0; 1676 else 1677 panic("lfs_nextseg: no clean segments"); 1678 } 1679 LFS_SEGENTRY(sup, fs, sn, bp); 1680 isdirty = sup->su_flags & (SEGUSE_DIRTY | (skip_inval ? SEGUSE_INVAL : 0)); 1681 /* Check SEGUSE_EMPTY as we go along */ 1682 if (isdirty && sup->su_nbytes == 0 && 1683 !(sup->su_flags & SEGUSE_EMPTY)) 1684 LFS_WRITESEGENTRY(sup, fs, sn, bp); 1685 else 1686 brelse(bp); 1687 1688 if (!isdirty) 1689 break; 1690 } 1691 if (skip_inval == 0) 1692 lfs_unset_inval_all(fs); 1693 1694 ++fs->lfs_nactive; 1695 fs->lfs_nextseg = sntod(fs, sn); 1696 if (lfs_dostats) { 1697 ++lfs_stats.segsused; 1698 } 1699 } 1700 1701 static struct buf * 1702 lfs_newclusterbuf(struct lfs *fs, struct vnode *vp, daddr_t addr, int n) 1703 { 1704 struct lfs_cluster *cl; 1705 struct buf **bpp, *bp; 1706 int s; 1707 1708 ASSERT_SEGLOCK(fs); 1709 cl = (struct lfs_cluster *)pool_get(&fs->lfs_clpool, PR_WAITOK); 1710 bpp = (struct buf **)pool_get(&fs->lfs_bpppool, PR_WAITOK); 1711 memset(cl, 0, sizeof(*cl)); 1712 cl->fs = fs; 1713 cl->bpp = bpp; 1714 cl->bufcount = 0; 1715 cl->bufsize = 0; 1716 1717 /* If this segment is being written synchronously, note that */ 1718 if (fs->lfs_sp->seg_flags & SEGM_SYNC) { 1719 cl->flags |= LFS_CL_SYNC; 1720 cl->seg = fs->lfs_sp; 1721 ++cl->seg->seg_iocount; 1722 } 1723 1724 /* Get an empty buffer header, or maybe one with something on it */ 1725 s = splbio(); 1726 bp = pool_get(&bufpool, PR_WAITOK); /* XXX should use lfs_malloc? */ 1727 splx(s); 1728 memset(bp, 0, sizeof(*bp)); 1729 BUF_INIT(bp); 1730 1731 bp->b_flags = B_BUSY | B_CALL; 1732 bp->b_dev = NODEV; 1733 bp->b_blkno = bp->b_lblkno = addr; 1734 bp->b_iodone = lfs_cluster_callback; 1735 bp->b_private = cl; 1736 bp->b_vp = vp; 1737 1738 return bp; 1739 } 1740 1741 int 1742 lfs_writeseg(struct lfs *fs, struct segment *sp) 1743 { 1744 struct buf **bpp, *bp, *cbp, *newbp; 1745 SEGUSE *sup; 1746 SEGSUM *ssp; 1747 int i, s; 1748 int do_again, nblocks, byteoffset; 1749 size_t el_size; 1750 struct lfs_cluster *cl; 1751 u_short ninos; 1752 struct vnode *devvp; 1753 char *p = NULL; 1754 struct vnode *vp; 1755 int32_t *daddrp; /* XXX ondisk32 */ 1756 int changed; 1757 u_int32_t sum; 1758 1759 ASSERT_SEGLOCK(fs); 1760 /* 1761 * If there are no buffers other than the segment summary to write 1762 * and it is not a checkpoint, don't do anything. On a checkpoint, 1763 * even if there aren't any buffers, you need to write the superblock. 1764 */ 1765 if ((nblocks = sp->cbpp - sp->bpp) == 1) 1766 return (0); 1767 1768 devvp = VTOI(fs->lfs_ivnode)->i_devvp; 1769 1770 /* Update the segment usage information. */ 1771 LFS_SEGENTRY(sup, fs, sp->seg_number, bp); 1772 1773 /* Loop through all blocks, except the segment summary. */ 1774 for (bpp = sp->bpp; ++bpp < sp->cbpp; ) { 1775 if ((*bpp)->b_vp != devvp) { 1776 sup->su_nbytes += (*bpp)->b_bcount; 1777 DLOG((DLOG_SU, "seg %" PRIu32 " += %ld for ino %d" 1778 " lbn %" PRId64 " db 0x%" PRIx64 "\n", 1779 sp->seg_number, (*bpp)->b_bcount, 1780 VTOI((*bpp)->b_vp)->i_number, (*bpp)->b_lblkno, 1781 (*bpp)->b_blkno)); 1782 } 1783 } 1784 1785 ssp = (SEGSUM *)sp->segsum; 1786 1787 ninos = (ssp->ss_ninos + INOPB(fs) - 1) / INOPB(fs); 1788 DLOG((DLOG_SU, "seg %d += %d for %d inodes\n", 1789 sp->seg_number, ssp->ss_ninos * sizeof (struct ufs1_dinode), 1790 ssp->ss_ninos)); 1791 sup->su_nbytes += ssp->ss_ninos * sizeof (struct ufs1_dinode); 1792 /* sup->su_nbytes += fs->lfs_sumsize; */ 1793 if (fs->lfs_version == 1) 1794 sup->su_olastmod = time.tv_sec; 1795 else 1796 sup->su_lastmod = time.tv_sec; 1797 sup->su_ninos += ninos; 1798 ++sup->su_nsums; 1799 fs->lfs_dmeta += (btofsb(fs, fs->lfs_sumsize) + btofsb(fs, ninos * 1800 fs->lfs_ibsize)); 1801 fs->lfs_avail -= btofsb(fs, fs->lfs_sumsize); 1802 1803 do_again = !(bp->b_flags & B_GATHERED); 1804 LFS_WRITESEGENTRY(sup, fs, sp->seg_number, bp); /* Ifile */ 1805 1806 /* 1807 * Mark blocks B_BUSY, to prevent then from being changed between 1808 * the checksum computation and the actual write. 1809 * 1810 * If we are cleaning, check indirect blocks for UNWRITTEN, and if 1811 * there are any, replace them with copies that have UNASSIGNED 1812 * instead. 1813 */ 1814 for (bpp = sp->bpp, i = nblocks - 1; i--;) { 1815 ++bpp; 1816 bp = *bpp; 1817 if (bp->b_flags & B_CALL) { /* UBC or malloced buffer */ 1818 bp->b_flags |= B_BUSY; 1819 continue; 1820 } 1821 1822 simple_lock(&bp->b_interlock); 1823 s = splbio(); 1824 while (bp->b_flags & B_BUSY) { 1825 DLOG((DLOG_SEG, "lfs_writeseg: avoiding potential" 1826 " data summary corruption for ino %d, lbn %" 1827 PRId64 "\n", 1828 VTOI(bp->b_vp)->i_number, bp->b_lblkno)); 1829 bp->b_flags |= B_WANTED; 1830 ltsleep(bp, (PRIBIO + 1), "lfs_writeseg", 0, 1831 &bp->b_interlock); 1832 splx(s); 1833 s = splbio(); 1834 } 1835 bp->b_flags |= B_BUSY; 1836 splx(s); 1837 simple_unlock(&bp->b_interlock); 1838 1839 /* 1840 * Check and replace indirect block UNWRITTEN bogosity. 1841 * XXX See comment in lfs_writefile. 1842 */ 1843 if (bp->b_lblkno < 0 && bp->b_vp != devvp && bp->b_vp && 1844 VTOI(bp->b_vp)->i_ffs1_blocks != 1845 VTOI(bp->b_vp)->i_lfs_effnblks) { 1846 DLOG((DLOG_VNODE, "lfs_writeseg: cleansing ino %d (%d != %d)\n", 1847 VTOI(bp->b_vp)->i_number, 1848 VTOI(bp->b_vp)->i_lfs_effnblks, 1849 VTOI(bp->b_vp)->i_ffs1_blocks)); 1850 /* Make a copy we'll make changes to */ 1851 newbp = lfs_newbuf(fs, bp->b_vp, bp->b_lblkno, 1852 bp->b_bcount, LFS_NB_IBLOCK); 1853 newbp->b_blkno = bp->b_blkno; 1854 memcpy(newbp->b_data, bp->b_data, 1855 newbp->b_bcount); 1856 1857 changed = 0; 1858 /* XXX ondisk32 */ 1859 for (daddrp = (int32_t *)(newbp->b_data); 1860 daddrp < (int32_t *)(newbp->b_data + 1861 newbp->b_bcount); daddrp++) { 1862 if (*daddrp == UNWRITTEN) { 1863 ++changed; 1864 *daddrp = 0; 1865 } 1866 } 1867 /* 1868 * Get rid of the old buffer. Don't mark it clean, 1869 * though, if it still has dirty data on it. 1870 */ 1871 if (changed) { 1872 DLOG((DLOG_SEG, "lfs_writeseg: replacing UNWRITTEN(%d):" 1873 " bp = %p newbp = %p\n", changed, bp, 1874 newbp)); 1875 *bpp = newbp; 1876 bp->b_flags &= ~(B_ERROR | B_GATHERED); 1877 if (bp->b_flags & B_CALL) { 1878 DLOG((DLOG_SEG, "lfs_writeseg: " 1879 "indir bp should not be B_CALL\n")); 1880 s = splbio(); 1881 biodone(bp); 1882 splx(s); 1883 bp = NULL; 1884 } else { 1885 /* Still on free list, leave it there */ 1886 s = splbio(); 1887 bp->b_flags &= ~B_BUSY; 1888 if (bp->b_flags & B_WANTED) 1889 wakeup(bp); 1890 splx(s); 1891 /* 1892 * We have to re-decrement lfs_avail 1893 * since this block is going to come 1894 * back around to us in the next 1895 * segment. 1896 */ 1897 fs->lfs_avail -= 1898 btofsb(fs, bp->b_bcount); 1899 } 1900 } else { 1901 lfs_freebuf(fs, newbp); 1902 } 1903 } 1904 } 1905 /* 1906 * Compute checksum across data and then across summary; the first 1907 * block (the summary block) is skipped. Set the create time here 1908 * so that it's guaranteed to be later than the inode mod times. 1909 */ 1910 sum = 0; 1911 if (fs->lfs_version == 1) 1912 el_size = sizeof(u_long); 1913 else 1914 el_size = sizeof(u_int32_t); 1915 for (bpp = sp->bpp, i = nblocks - 1; i--; ) { 1916 ++bpp; 1917 /* Loop through gop_write cluster blocks */ 1918 for (byteoffset = 0; byteoffset < (*bpp)->b_bcount; 1919 byteoffset += fs->lfs_bsize) { 1920 #ifdef LFS_USE_B_INVAL 1921 if (((*bpp)->b_flags & (B_CALL | B_INVAL)) == 1922 (B_CALL | B_INVAL)) { 1923 if (copyin((caddr_t)(*bpp)->b_saveaddr + 1924 byteoffset, dp, el_size)) { 1925 panic("lfs_writeseg: copyin failed [1]:" 1926 " ino %d blk %" PRId64, 1927 VTOI((*bpp)->b_vp)->i_number, 1928 (*bpp)->b_lblkno); 1929 } 1930 } else 1931 #endif /* LFS_USE_B_INVAL */ 1932 { 1933 sum = lfs_cksum_part( 1934 (*bpp)->b_data + byteoffset, el_size, sum); 1935 } 1936 } 1937 } 1938 if (fs->lfs_version == 1) 1939 ssp->ss_ocreate = time.tv_sec; 1940 else { 1941 ssp->ss_create = time.tv_sec; 1942 ssp->ss_serial = ++fs->lfs_serial; 1943 ssp->ss_ident = fs->lfs_ident; 1944 } 1945 ssp->ss_datasum = lfs_cksum_fold(sum); 1946 ssp->ss_sumsum = cksum(&ssp->ss_datasum, 1947 fs->lfs_sumsize - sizeof(ssp->ss_sumsum)); 1948 1949 simple_lock(&fs->lfs_interlock); 1950 fs->lfs_bfree -= (btofsb(fs, ninos * fs->lfs_ibsize) + 1951 btofsb(fs, fs->lfs_sumsize)); 1952 simple_unlock(&fs->lfs_interlock); 1953 1954 /* 1955 * When we simply write the blocks we lose a rotation for every block 1956 * written. To avoid this problem, we cluster the buffers into a 1957 * chunk and write the chunk. MAXPHYS is the largest size I/O 1958 * devices can handle, use that for the size of the chunks. 1959 * 1960 * Blocks that are already clusters (from GOP_WRITE), however, we 1961 * don't bother to copy into other clusters. 1962 */ 1963 1964 #define CHUNKSIZE MAXPHYS 1965 1966 if (devvp == NULL) 1967 panic("devvp is NULL"); 1968 for (bpp = sp->bpp, i = nblocks; i;) { 1969 cbp = lfs_newclusterbuf(fs, devvp, (*bpp)->b_blkno, i); 1970 cl = cbp->b_private; 1971 1972 cbp->b_flags |= B_ASYNC | B_BUSY; 1973 cbp->b_bcount = 0; 1974 1975 #if defined(DEBUG) && defined(DIAGNOSTIC) 1976 if (bpp - sp->bpp > (fs->lfs_sumsize - SEGSUM_SIZE(fs)) 1977 / sizeof(int32_t)) { 1978 panic("lfs_writeseg: real bpp overwrite"); 1979 } 1980 if (bpp - sp->bpp > segsize(fs) / fs->lfs_fsize) { 1981 panic("lfs_writeseg: theoretical bpp overwrite"); 1982 } 1983 #endif 1984 1985 /* 1986 * Construct the cluster. 1987 */ 1988 simple_lock(&fs->lfs_interlock); 1989 ++fs->lfs_iocount; 1990 simple_unlock(&fs->lfs_interlock); 1991 while (i && cbp->b_bcount < CHUNKSIZE) { 1992 bp = *bpp; 1993 1994 if (bp->b_bcount > (CHUNKSIZE - cbp->b_bcount)) 1995 break; 1996 if (cbp->b_bcount > 0 && !(cl->flags & LFS_CL_MALLOC)) 1997 break; 1998 1999 /* Clusters from GOP_WRITE are expedited */ 2000 if (bp->b_bcount > fs->lfs_bsize) { 2001 if (cbp->b_bcount > 0) 2002 /* Put in its own buffer */ 2003 break; 2004 else { 2005 cbp->b_data = bp->b_data; 2006 } 2007 } else if (cbp->b_bcount == 0) { 2008 p = cbp->b_data = lfs_malloc(fs, CHUNKSIZE, 2009 LFS_NB_CLUSTER); 2010 cl->flags |= LFS_CL_MALLOC; 2011 } 2012 #ifdef DIAGNOSTIC 2013 if (dtosn(fs, dbtofsb(fs, bp->b_blkno + 2014 btodb(bp->b_bcount - 1))) != 2015 sp->seg_number) { 2016 printf("blk size %d daddr %" PRIx64 2017 " not in seg %d\n", 2018 bp->b_bcount, bp->b_blkno, 2019 sp->seg_number); 2020 panic("segment overwrite"); 2021 } 2022 #endif 2023 2024 #ifdef LFS_USE_B_INVAL 2025 /* 2026 * Fake buffers from the cleaner are marked as B_INVAL. 2027 * We need to copy the data from user space rather than 2028 * from the buffer indicated. 2029 * XXX == what do I do on an error? 2030 */ 2031 if ((bp->b_flags & (B_CALL|B_INVAL)) == 2032 (B_CALL|B_INVAL)) { 2033 if (copyin(bp->b_saveaddr, p, bp->b_bcount)) 2034 panic("lfs_writeseg: " 2035 "copyin failed [2]"); 2036 } else 2037 #endif /* LFS_USE_B_INVAL */ 2038 if (cl->flags & LFS_CL_MALLOC) { 2039 /* copy data into our cluster. */ 2040 memcpy(p, bp->b_data, bp->b_bcount); 2041 p += bp->b_bcount; 2042 } 2043 2044 cbp->b_bcount += bp->b_bcount; 2045 cl->bufsize += bp->b_bcount; 2046 2047 bp->b_flags &= ~(B_ERROR | B_READ | B_DELWRI | B_DONE); 2048 cl->bpp[cl->bufcount++] = bp; 2049 vp = bp->b_vp; 2050 s = splbio(); 2051 reassignbuf(bp, vp); 2052 V_INCR_NUMOUTPUT(vp); 2053 splx(s); 2054 2055 bpp++; 2056 i--; 2057 } 2058 if (fs->lfs_sp->seg_flags & SEGM_SYNC) 2059 BIO_SETPRIO(cbp, BPRIO_TIMECRITICAL); 2060 else 2061 BIO_SETPRIO(cbp, BPRIO_TIMELIMITED); 2062 s = splbio(); 2063 V_INCR_NUMOUTPUT(devvp); 2064 splx(s); 2065 VOP_STRATEGY(devvp, cbp); 2066 curproc->p_stats->p_ru.ru_oublock++; 2067 } 2068 2069 if (lfs_dostats) { 2070 ++lfs_stats.psegwrites; 2071 lfs_stats.blocktot += nblocks - 1; 2072 if (fs->lfs_sp->seg_flags & SEGM_SYNC) 2073 ++lfs_stats.psyncwrites; 2074 if (fs->lfs_sp->seg_flags & SEGM_CLEAN) { 2075 ++lfs_stats.pcleanwrites; 2076 lfs_stats.cleanblocks += nblocks - 1; 2077 } 2078 } 2079 return (lfs_initseg(fs) || do_again); 2080 } 2081 2082 void 2083 lfs_writesuper(struct lfs *fs, daddr_t daddr) 2084 { 2085 struct buf *bp; 2086 int s; 2087 struct vnode *devvp = VTOI(fs->lfs_ivnode)->i_devvp; 2088 2089 ASSERT_MAYBE_SEGLOCK(fs); 2090 #ifdef DIAGNOSTIC 2091 KASSERT(fs->lfs_magic == LFS_MAGIC); 2092 #endif 2093 /* 2094 * If we can write one superblock while another is in 2095 * progress, we risk not having a complete checkpoint if we crash. 2096 * So, block here if a superblock write is in progress. 2097 */ 2098 simple_lock(&fs->lfs_interlock); 2099 s = splbio(); 2100 while (fs->lfs_sbactive) { 2101 ltsleep(&fs->lfs_sbactive, PRIBIO+1, "lfs sb", 0, 2102 &fs->lfs_interlock); 2103 } 2104 fs->lfs_sbactive = daddr; 2105 splx(s); 2106 simple_unlock(&fs->lfs_interlock); 2107 2108 /* Set timestamp of this version of the superblock */ 2109 if (fs->lfs_version == 1) 2110 fs->lfs_otstamp = time.tv_sec; 2111 fs->lfs_tstamp = time.tv_sec; 2112 2113 /* Checksum the superblock and copy it into a buffer. */ 2114 fs->lfs_cksum = lfs_sb_cksum(&(fs->lfs_dlfs)); 2115 bp = lfs_newbuf(fs, devvp, 2116 fsbtodb(fs, daddr), LFS_SBPAD, LFS_NB_SBLOCK); 2117 memset(bp->b_data + sizeof(struct dlfs), 0, 2118 LFS_SBPAD - sizeof(struct dlfs)); 2119 *(struct dlfs *)bp->b_data = fs->lfs_dlfs; 2120 2121 bp->b_flags |= B_BUSY | B_CALL | B_ASYNC; 2122 bp->b_flags &= ~(B_DONE | B_ERROR | B_READ | B_DELWRI); 2123 bp->b_iodone = lfs_supercallback; 2124 2125 if (fs->lfs_sp != NULL && fs->lfs_sp->seg_flags & SEGM_SYNC) 2126 BIO_SETPRIO(bp, BPRIO_TIMECRITICAL); 2127 else 2128 BIO_SETPRIO(bp, BPRIO_TIMELIMITED); 2129 curproc->p_stats->p_ru.ru_oublock++; 2130 s = splbio(); 2131 V_INCR_NUMOUTPUT(bp->b_vp); 2132 splx(s); 2133 simple_lock(&fs->lfs_interlock); 2134 ++fs->lfs_iocount; 2135 simple_unlock(&fs->lfs_interlock); 2136 VOP_STRATEGY(devvp, bp); 2137 } 2138 2139 /* 2140 * Logical block number match routines used when traversing the dirty block 2141 * chain. 2142 */ 2143 int 2144 lfs_match_fake(struct lfs *fs, struct buf *bp) 2145 { 2146 2147 ASSERT_SEGLOCK(fs); 2148 return LFS_IS_MALLOC_BUF(bp); 2149 } 2150 2151 #if 0 2152 int 2153 lfs_match_real(struct lfs *fs, struct buf *bp) 2154 { 2155 2156 ASSERT_SEGLOCK(fs); 2157 return (lfs_match_data(fs, bp) && !lfs_match_fake(fs, bp)); 2158 } 2159 #endif 2160 2161 int 2162 lfs_match_data(struct lfs *fs, struct buf *bp) 2163 { 2164 2165 ASSERT_SEGLOCK(fs); 2166 return (bp->b_lblkno >= 0); 2167 } 2168 2169 int 2170 lfs_match_indir(struct lfs *fs, struct buf *bp) 2171 { 2172 daddr_t lbn; 2173 2174 ASSERT_SEGLOCK(fs); 2175 lbn = bp->b_lblkno; 2176 return (lbn < 0 && (-lbn - NDADDR) % NINDIR(fs) == 0); 2177 } 2178 2179 int 2180 lfs_match_dindir(struct lfs *fs, struct buf *bp) 2181 { 2182 daddr_t lbn; 2183 2184 ASSERT_SEGLOCK(fs); 2185 lbn = bp->b_lblkno; 2186 return (lbn < 0 && (-lbn - NDADDR) % NINDIR(fs) == 1); 2187 } 2188 2189 int 2190 lfs_match_tindir(struct lfs *fs, struct buf *bp) 2191 { 2192 daddr_t lbn; 2193 2194 ASSERT_SEGLOCK(fs); 2195 lbn = bp->b_lblkno; 2196 return (lbn < 0 && (-lbn - NDADDR) % NINDIR(fs) == 2); 2197 } 2198 2199 /* 2200 * XXX - The only buffers that are going to hit these functions are the 2201 * segment write blocks, or the segment summaries, or the superblocks. 2202 * 2203 * All of the above are created by lfs_newbuf, and so do not need to be 2204 * released via brelse. 2205 */ 2206 void 2207 lfs_callback(struct buf *bp) 2208 { 2209 struct lfs *fs; 2210 2211 fs = bp->b_private; 2212 ASSERT_NO_SEGLOCK(fs); 2213 lfs_freebuf(fs, bp); 2214 } 2215 2216 static void 2217 lfs_super_aiodone(struct buf *bp) 2218 { 2219 struct lfs *fs; 2220 2221 fs = bp->b_private; 2222 ASSERT_NO_SEGLOCK(fs); 2223 simple_lock(&fs->lfs_interlock); 2224 fs->lfs_sbactive = 0; 2225 if (--fs->lfs_iocount <= 1) 2226 wakeup(&fs->lfs_iocount); 2227 simple_unlock(&fs->lfs_interlock); 2228 wakeup(&fs->lfs_sbactive); 2229 lfs_freebuf(fs, bp); 2230 } 2231 2232 static void 2233 lfs_cluster_aiodone(struct buf *bp) 2234 { 2235 struct lfs_cluster *cl; 2236 struct lfs *fs; 2237 struct buf *tbp, *fbp; 2238 struct vnode *vp, *devvp; 2239 struct inode *ip; 2240 int s, error=0; 2241 2242 if (bp->b_flags & B_ERROR) 2243 error = bp->b_error; 2244 2245 cl = bp->b_private; 2246 fs = cl->fs; 2247 devvp = VTOI(fs->lfs_ivnode)->i_devvp; 2248 ASSERT_NO_SEGLOCK(fs); 2249 2250 /* Put the pages back, and release the buffer */ 2251 while (cl->bufcount--) { 2252 tbp = cl->bpp[cl->bufcount]; 2253 KASSERT(tbp->b_flags & B_BUSY); 2254 if (error) { 2255 tbp->b_flags |= B_ERROR; 2256 tbp->b_error = error; 2257 } 2258 2259 /* 2260 * We're done with tbp. If it has not been re-dirtied since 2261 * the cluster was written, free it. Otherwise, keep it on 2262 * the locked list to be written again. 2263 */ 2264 vp = tbp->b_vp; 2265 2266 tbp->b_flags &= ~B_GATHERED; 2267 2268 LFS_BCLEAN_LOG(fs, tbp); 2269 2270 if (!(tbp->b_flags & B_CALL)) { 2271 KASSERT(tbp->b_flags & B_LOCKED); 2272 s = splbio(); 2273 simple_lock(&bqueue_slock); 2274 bremfree(tbp); 2275 simple_unlock(&bqueue_slock); 2276 if (vp) 2277 reassignbuf(tbp, vp); 2278 splx(s); 2279 tbp->b_flags |= B_ASYNC; /* for biodone */ 2280 } 2281 2282 if ((tbp->b_flags & (B_LOCKED | B_DELWRI)) == B_LOCKED) 2283 LFS_UNLOCK_BUF(tbp); 2284 2285 if (tbp->b_flags & B_DONE) { 2286 DLOG((DLOG_SEG, "blk %d biodone already (flags %lx)\n", 2287 cl->bufcount, (long)tbp->b_flags)); 2288 } 2289 2290 if ((tbp->b_flags & B_CALL) && !LFS_IS_MALLOC_BUF(tbp)) { 2291 /* 2292 * A buffer from the page daemon. 2293 * We use the same iodone as it does, 2294 * so we must manually disassociate its 2295 * buffers from the vp. 2296 */ 2297 if (tbp->b_vp) { 2298 /* This is just silly */ 2299 s = splbio(); 2300 brelvp(tbp); 2301 tbp->b_vp = vp; 2302 splx(s); 2303 } 2304 /* Put it back the way it was */ 2305 tbp->b_flags |= B_ASYNC; 2306 /* Master buffers have B_AGE */ 2307 if (tbp->b_private == tbp) 2308 tbp->b_flags |= B_AGE; 2309 } 2310 s = splbio(); 2311 biodone(tbp); 2312 2313 /* 2314 * If this is the last block for this vnode, but 2315 * there are other blocks on its dirty list, 2316 * set IN_MODIFIED/IN_CLEANING depending on what 2317 * sort of block. Only do this for our mount point, 2318 * not for, e.g., inode blocks that are attached to 2319 * the devvp. 2320 * XXX KS - Shouldn't we set *both* if both types 2321 * of blocks are present (traverse the dirty list?) 2322 */ 2323 simple_lock(&global_v_numoutput_slock); 2324 if (vp != devvp && vp->v_numoutput == 0 && 2325 (fbp = LIST_FIRST(&vp->v_dirtyblkhd)) != NULL) { 2326 ip = VTOI(vp); 2327 DLOG((DLOG_SEG, "lfs_cluster_aiodone: mark ino %d\n", 2328 ip->i_number)); 2329 if (LFS_IS_MALLOC_BUF(fbp)) 2330 LFS_SET_UINO(ip, IN_CLEANING); 2331 else 2332 LFS_SET_UINO(ip, IN_MODIFIED); 2333 } 2334 simple_unlock(&global_v_numoutput_slock); 2335 splx(s); 2336 wakeup(vp); 2337 } 2338 2339 /* Fix up the cluster buffer, and release it */ 2340 if (cl->flags & LFS_CL_MALLOC) 2341 lfs_free(fs, bp->b_data, LFS_NB_CLUSTER); 2342 s = splbio(); 2343 pool_put(&bufpool, bp); /* XXX should use lfs_free? */ 2344 splx(s); 2345 2346 /* Note i/o done */ 2347 if (cl->flags & LFS_CL_SYNC) { 2348 if (--cl->seg->seg_iocount == 0) 2349 wakeup(&cl->seg->seg_iocount); 2350 } 2351 simple_lock(&fs->lfs_interlock); 2352 #ifdef DIAGNOSTIC 2353 if (fs->lfs_iocount == 0) 2354 panic("lfs_cluster_aiodone: zero iocount"); 2355 #endif 2356 if (--fs->lfs_iocount <= 1) 2357 wakeup(&fs->lfs_iocount); 2358 simple_unlock(&fs->lfs_interlock); 2359 2360 pool_put(&fs->lfs_bpppool, cl->bpp); 2361 cl->bpp = NULL; 2362 pool_put(&fs->lfs_clpool, cl); 2363 } 2364 2365 static void 2366 lfs_generic_callback(struct buf *bp, void (*aiodone)(struct buf *)) 2367 { 2368 /* reset b_iodone for when this is a single-buf i/o. */ 2369 bp->b_iodone = aiodone; 2370 2371 simple_lock(&uvm.aiodoned_lock); /* locks uvm.aio_done */ 2372 TAILQ_INSERT_TAIL(&uvm.aio_done, bp, b_freelist); 2373 wakeup(&uvm.aiodoned); 2374 simple_unlock(&uvm.aiodoned_lock); 2375 } 2376 2377 static void 2378 lfs_cluster_callback(struct buf *bp) 2379 { 2380 2381 lfs_generic_callback(bp, lfs_cluster_aiodone); 2382 } 2383 2384 void 2385 lfs_supercallback(struct buf *bp) 2386 { 2387 2388 lfs_generic_callback(bp, lfs_super_aiodone); 2389 } 2390 2391 /* 2392 * Shellsort (diminishing increment sort) from Data Structures and 2393 * Algorithms, Aho, Hopcraft and Ullman, 1983 Edition, page 290; 2394 * see also Knuth Vol. 3, page 84. The increments are selected from 2395 * formula (8), page 95. Roughly O(N^3/2). 2396 */ 2397 /* 2398 * This is our own private copy of shellsort because we want to sort 2399 * two parallel arrays (the array of buffer pointers and the array of 2400 * logical block numbers) simultaneously. Note that we cast the array 2401 * of logical block numbers to a unsigned in this routine so that the 2402 * negative block numbers (meta data blocks) sort AFTER the data blocks. 2403 */ 2404 2405 void 2406 lfs_shellsort(struct buf **bp_array, int32_t *lb_array, int nmemb, int size) 2407 { 2408 static int __rsshell_increments[] = { 4, 1, 0 }; 2409 int incr, *incrp, t1, t2; 2410 struct buf *bp_temp; 2411 2412 #ifdef DEBUG 2413 incr = 0; 2414 for (t1 = 0; t1 < nmemb; t1++) { 2415 for (t2 = 0; t2 * size < bp_array[t1]->b_bcount; t2++) { 2416 if (lb_array[incr++] != bp_array[t1]->b_lblkno + t2) { 2417 /* dump before panic */ 2418 printf("lfs_shellsort: nmemb=%d, size=%d\n", 2419 nmemb, size); 2420 incr = 0; 2421 for (t1 = 0; t1 < nmemb; t1++) { 2422 const struct buf *bp = bp_array[t1]; 2423 2424 printf("bp[%d]: lbn=%" PRIu64 ", size=%" 2425 PRIu64 "\n", t1, 2426 (uint64_t)bp->b_bcount, 2427 (uint64_t)bp->b_lblkno); 2428 printf("lbns:"); 2429 for (t2 = 0; t2 * size < bp->b_bcount; 2430 t2++) { 2431 printf(" %" PRId32, 2432 lb_array[incr++]); 2433 } 2434 printf("\n"); 2435 } 2436 panic("lfs_shellsort: inconsistent input"); 2437 } 2438 } 2439 } 2440 #endif 2441 2442 for (incrp = __rsshell_increments; (incr = *incrp++) != 0;) 2443 for (t1 = incr; t1 < nmemb; ++t1) 2444 for (t2 = t1 - incr; t2 >= 0;) 2445 if ((u_int32_t)bp_array[t2]->b_lblkno > 2446 (u_int32_t)bp_array[t2 + incr]->b_lblkno) { 2447 bp_temp = bp_array[t2]; 2448 bp_array[t2] = bp_array[t2 + incr]; 2449 bp_array[t2 + incr] = bp_temp; 2450 t2 -= incr; 2451 } else 2452 break; 2453 2454 /* Reform the list of logical blocks */ 2455 incr = 0; 2456 for (t1 = 0; t1 < nmemb; t1++) { 2457 for (t2 = 0; t2 * size < bp_array[t1]->b_bcount; t2++) { 2458 lb_array[incr++] = bp_array[t1]->b_lblkno + t2; 2459 } 2460 } 2461 } 2462 2463 /* 2464 * Check VXLOCK. Return 1 if the vnode is locked. Otherwise, vget it. 2465 */ 2466 int 2467 lfs_vref(struct vnode *vp) 2468 { 2469 ASSERT_MAYBE_SEGLOCK(VTOI(vp)->i_lfs); 2470 /* 2471 * If we return 1 here during a flush, we risk vinvalbuf() not 2472 * being able to flush all of the pages from this vnode, which 2473 * will cause it to panic. So, return 0 if a flush is in progress. 2474 */ 2475 if (vp->v_flag & VXLOCK) { 2476 if (IS_FLUSHING(VTOI(vp)->i_lfs, vp)) { 2477 return 0; 2478 } 2479 return (1); 2480 } 2481 return (vget(vp, 0)); 2482 } 2483 2484 /* 2485 * This is vrele except that we do not want to VOP_INACTIVE this vnode. We 2486 * inline vrele here to avoid the vn_lock and VOP_INACTIVE call at the end. 2487 */ 2488 void 2489 lfs_vunref(struct vnode *vp) 2490 { 2491 ASSERT_MAYBE_SEGLOCK(VTOI(vp)->i_lfs); 2492 /* 2493 * Analogous to lfs_vref, if the node is flushing, fake it. 2494 */ 2495 if ((vp->v_flag & VXLOCK) && IS_FLUSHING(VTOI(vp)->i_lfs, vp)) { 2496 return; 2497 } 2498 2499 simple_lock(&vp->v_interlock); 2500 #ifdef DIAGNOSTIC 2501 if (vp->v_usecount <= 0) { 2502 printf("lfs_vunref: inum is %llu\n", (unsigned long long) 2503 VTOI(vp)->i_number); 2504 printf("lfs_vunref: flags are 0x%lx\n", (u_long)vp->v_flag); 2505 printf("lfs_vunref: usecount = %ld\n", (long)vp->v_usecount); 2506 panic("lfs_vunref: v_usecount < 0"); 2507 } 2508 #endif 2509 vp->v_usecount--; 2510 if (vp->v_usecount > 0) { 2511 simple_unlock(&vp->v_interlock); 2512 return; 2513 } 2514 /* 2515 * insert at tail of LRU list 2516 */ 2517 simple_lock(&vnode_free_list_slock); 2518 if (vp->v_holdcnt > 0) 2519 TAILQ_INSERT_TAIL(&vnode_hold_list, vp, v_freelist); 2520 else 2521 TAILQ_INSERT_TAIL(&vnode_free_list, vp, v_freelist); 2522 simple_unlock(&vnode_free_list_slock); 2523 simple_unlock(&vp->v_interlock); 2524 } 2525 2526 /* 2527 * We use this when we have vnodes that were loaded in solely for cleaning. 2528 * There is no reason to believe that these vnodes will be referenced again 2529 * soon, since the cleaning process is unrelated to normal filesystem 2530 * activity. Putting cleaned vnodes at the tail of the list has the effect 2531 * of flushing the vnode LRU. So, put vnodes that were loaded only for 2532 * cleaning at the head of the list, instead. 2533 */ 2534 void 2535 lfs_vunref_head(struct vnode *vp) 2536 { 2537 2538 ASSERT_SEGLOCK(VTOI(vp)->i_lfs); 2539 simple_lock(&vp->v_interlock); 2540 #ifdef DIAGNOSTIC 2541 if (vp->v_usecount == 0) { 2542 panic("lfs_vunref: v_usecount<0"); 2543 } 2544 #endif 2545 vp->v_usecount--; 2546 if (vp->v_usecount > 0) { 2547 simple_unlock(&vp->v_interlock); 2548 return; 2549 } 2550 /* 2551 * insert at head of LRU list 2552 */ 2553 simple_lock(&vnode_free_list_slock); 2554 if (vp->v_holdcnt > 0) 2555 TAILQ_INSERT_TAIL(&vnode_hold_list, vp, v_freelist); 2556 else 2557 TAILQ_INSERT_HEAD(&vnode_free_list, vp, v_freelist); 2558 simple_unlock(&vnode_free_list_slock); 2559 simple_unlock(&vp->v_interlock); 2560 } 2561 2562
Indexes created Tue Sep 23 07:09:52 GMT 2025