lfs_vnops.c revision 1.109.2.7
1 /* $NetBSD: lfs_vnops.c,v 1.109.2.7 2005/02/04 11:48:27 skrll 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) 1986, 1989, 1991, 1993, 1995 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_vnops.c 8.13 (Berkeley) 6/10/95 67 */ 68 69 #include <sys/cdefs.h> 70 __KERNEL_RCSID(0, "$NetBSD: lfs_vnops.c,v 1.109.2.7 2005/02/04 11:48:27 skrll Exp $"); 71 72 #include <sys/param.h> 73 #include <sys/systm.h> 74 #include <sys/namei.h> 75 #include <sys/resourcevar.h> 76 #include <sys/kernel.h> 77 #include <sys/file.h> 78 #include <sys/stat.h> 79 #include <sys/buf.h> 80 #include <sys/proc.h> 81 #include <sys/mount.h> 82 #include <sys/vnode.h> 83 #include <sys/malloc.h> 84 #include <sys/pool.h> 85 #include <sys/signalvar.h> 86 87 #include <miscfs/fifofs/fifo.h> 88 #include <miscfs/genfs/genfs.h> 89 #include <miscfs/specfs/specdev.h> 90 91 #include <ufs/ufs/inode.h> 92 #include <ufs/ufs/dir.h> 93 #include <ufs/ufs/ufsmount.h> 94 #include <ufs/ufs/ufs_extern.h> 95 96 #include <uvm/uvm.h> 97 #include <uvm/uvm_pmap.h> 98 #include <uvm/uvm_stat.h> 99 #include <uvm/uvm_pager.h> 100 101 #include <ufs/lfs/lfs.h> 102 #include <ufs/lfs/lfs_extern.h> 103 104 extern pid_t lfs_writer_daemon; 105 106 /* Global vfs data structures for lfs. */ 107 int (**lfs_vnodeop_p)(void *); 108 const struct vnodeopv_entry_desc lfs_vnodeop_entries[] = { 109 { &vop_default_desc, vn_default_error }, 110 { &vop_lookup_desc, ufs_lookup }, /* lookup */ 111 { &vop_create_desc, lfs_create }, /* create */ 112 { &vop_whiteout_desc, ufs_whiteout }, /* whiteout */ 113 { &vop_mknod_desc, lfs_mknod }, /* mknod */ 114 { &vop_open_desc, ufs_open }, /* open */ 115 { &vop_close_desc, lfs_close }, /* close */ 116 { &vop_access_desc, ufs_access }, /* access */ 117 { &vop_getattr_desc, lfs_getattr }, /* getattr */ 118 { &vop_setattr_desc, lfs_setattr }, /* setattr */ 119 { &vop_read_desc, lfs_read }, /* read */ 120 { &vop_write_desc, lfs_write }, /* write */ 121 { &vop_lease_desc, ufs_lease_check }, /* lease */ 122 { &vop_ioctl_desc, ufs_ioctl }, /* ioctl */ 123 { &vop_fcntl_desc, lfs_fcntl }, /* fcntl */ 124 { &vop_poll_desc, ufs_poll }, /* poll */ 125 { &vop_kqfilter_desc, genfs_kqfilter }, /* kqfilter */ 126 { &vop_revoke_desc, ufs_revoke }, /* revoke */ 127 { &vop_mmap_desc, lfs_mmap }, /* mmap */ 128 { &vop_fsync_desc, lfs_fsync }, /* fsync */ 129 { &vop_seek_desc, ufs_seek }, /* seek */ 130 { &vop_remove_desc, lfs_remove }, /* remove */ 131 { &vop_link_desc, lfs_link }, /* link */ 132 { &vop_rename_desc, lfs_rename }, /* rename */ 133 { &vop_mkdir_desc, lfs_mkdir }, /* mkdir */ 134 { &vop_rmdir_desc, lfs_rmdir }, /* rmdir */ 135 { &vop_symlink_desc, lfs_symlink }, /* symlink */ 136 { &vop_readdir_desc, ufs_readdir }, /* readdir */ 137 { &vop_readlink_desc, ufs_readlink }, /* readlink */ 138 { &vop_abortop_desc, ufs_abortop }, /* abortop */ 139 { &vop_inactive_desc, lfs_inactive }, /* inactive */ 140 { &vop_reclaim_desc, lfs_reclaim }, /* reclaim */ 141 { &vop_lock_desc, ufs_lock }, /* lock */ 142 { &vop_unlock_desc, ufs_unlock }, /* unlock */ 143 { &vop_bmap_desc, ufs_bmap }, /* bmap */ 144 { &vop_strategy_desc, lfs_strategy }, /* strategy */ 145 { &vop_print_desc, ufs_print }, /* print */ 146 { &vop_islocked_desc, ufs_islocked }, /* islocked */ 147 { &vop_pathconf_desc, ufs_pathconf }, /* pathconf */ 148 { &vop_advlock_desc, ufs_advlock }, /* advlock */ 149 { &vop_blkatoff_desc, lfs_blkatoff }, /* blkatoff */ 150 { &vop_valloc_desc, lfs_valloc }, /* valloc */ 151 { &vop_balloc_desc, lfs_balloc }, /* balloc */ 152 { &vop_vfree_desc, lfs_vfree }, /* vfree */ 153 { &vop_truncate_desc, lfs_truncate }, /* truncate */ 154 { &vop_update_desc, lfs_update }, /* update */ 155 { &vop_bwrite_desc, lfs_bwrite }, /* bwrite */ 156 { &vop_getpages_desc, lfs_getpages }, /* getpages */ 157 { &vop_putpages_desc, lfs_putpages }, /* putpages */ 158 { NULL, NULL } 159 }; 160 const struct vnodeopv_desc lfs_vnodeop_opv_desc = 161 { &lfs_vnodeop_p, lfs_vnodeop_entries }; 162 163 int (**lfs_specop_p)(void *); 164 const struct vnodeopv_entry_desc lfs_specop_entries[] = { 165 { &vop_default_desc, vn_default_error }, 166 { &vop_lookup_desc, spec_lookup }, /* lookup */ 167 { &vop_create_desc, spec_create }, /* create */ 168 { &vop_mknod_desc, spec_mknod }, /* mknod */ 169 { &vop_open_desc, spec_open }, /* open */ 170 { &vop_close_desc, lfsspec_close }, /* close */ 171 { &vop_access_desc, ufs_access }, /* access */ 172 { &vop_getattr_desc, lfs_getattr }, /* getattr */ 173 { &vop_setattr_desc, lfs_setattr }, /* setattr */ 174 { &vop_read_desc, ufsspec_read }, /* read */ 175 { &vop_write_desc, ufsspec_write }, /* write */ 176 { &vop_lease_desc, spec_lease_check }, /* lease */ 177 { &vop_ioctl_desc, spec_ioctl }, /* ioctl */ 178 { &vop_fcntl_desc, ufs_fcntl }, /* fcntl */ 179 { &vop_poll_desc, spec_poll }, /* poll */ 180 { &vop_kqfilter_desc, spec_kqfilter }, /* kqfilter */ 181 { &vop_revoke_desc, spec_revoke }, /* revoke */ 182 { &vop_mmap_desc, spec_mmap }, /* mmap */ 183 { &vop_fsync_desc, spec_fsync }, /* fsync */ 184 { &vop_seek_desc, spec_seek }, /* seek */ 185 { &vop_remove_desc, spec_remove }, /* remove */ 186 { &vop_link_desc, spec_link }, /* link */ 187 { &vop_rename_desc, spec_rename }, /* rename */ 188 { &vop_mkdir_desc, spec_mkdir }, /* mkdir */ 189 { &vop_rmdir_desc, spec_rmdir }, /* rmdir */ 190 { &vop_symlink_desc, spec_symlink }, /* symlink */ 191 { &vop_readdir_desc, spec_readdir }, /* readdir */ 192 { &vop_readlink_desc, spec_readlink }, /* readlink */ 193 { &vop_abortop_desc, spec_abortop }, /* abortop */ 194 { &vop_inactive_desc, lfs_inactive }, /* inactive */ 195 { &vop_reclaim_desc, lfs_reclaim }, /* reclaim */ 196 { &vop_lock_desc, ufs_lock }, /* lock */ 197 { &vop_unlock_desc, ufs_unlock }, /* unlock */ 198 { &vop_bmap_desc, spec_bmap }, /* bmap */ 199 { &vop_strategy_desc, spec_strategy }, /* strategy */ 200 { &vop_print_desc, ufs_print }, /* print */ 201 { &vop_islocked_desc, ufs_islocked }, /* islocked */ 202 { &vop_pathconf_desc, spec_pathconf }, /* pathconf */ 203 { &vop_advlock_desc, spec_advlock }, /* advlock */ 204 { &vop_blkatoff_desc, spec_blkatoff }, /* blkatoff */ 205 { &vop_valloc_desc, spec_valloc }, /* valloc */ 206 { &vop_vfree_desc, lfs_vfree }, /* vfree */ 207 { &vop_truncate_desc, spec_truncate }, /* truncate */ 208 { &vop_update_desc, lfs_update }, /* update */ 209 { &vop_bwrite_desc, vn_bwrite }, /* bwrite */ 210 { &vop_getpages_desc, spec_getpages }, /* getpages */ 211 { &vop_putpages_desc, spec_putpages }, /* putpages */ 212 { NULL, NULL } 213 }; 214 const struct vnodeopv_desc lfs_specop_opv_desc = 215 { &lfs_specop_p, lfs_specop_entries }; 216 217 int (**lfs_fifoop_p)(void *); 218 const struct vnodeopv_entry_desc lfs_fifoop_entries[] = { 219 { &vop_default_desc, vn_default_error }, 220 { &vop_lookup_desc, fifo_lookup }, /* lookup */ 221 { &vop_create_desc, fifo_create }, /* create */ 222 { &vop_mknod_desc, fifo_mknod }, /* mknod */ 223 { &vop_open_desc, fifo_open }, /* open */ 224 { &vop_close_desc, lfsfifo_close }, /* close */ 225 { &vop_access_desc, ufs_access }, /* access */ 226 { &vop_getattr_desc, lfs_getattr }, /* getattr */ 227 { &vop_setattr_desc, lfs_setattr }, /* setattr */ 228 { &vop_read_desc, ufsfifo_read }, /* read */ 229 { &vop_write_desc, ufsfifo_write }, /* write */ 230 { &vop_lease_desc, fifo_lease_check }, /* lease */ 231 { &vop_ioctl_desc, fifo_ioctl }, /* ioctl */ 232 { &vop_fcntl_desc, ufs_fcntl }, /* fcntl */ 233 { &vop_poll_desc, fifo_poll }, /* poll */ 234 { &vop_kqfilter_desc, fifo_kqfilter }, /* kqfilter */ 235 { &vop_revoke_desc, fifo_revoke }, /* revoke */ 236 { &vop_mmap_desc, fifo_mmap }, /* mmap */ 237 { &vop_fsync_desc, fifo_fsync }, /* fsync */ 238 { &vop_seek_desc, fifo_seek }, /* seek */ 239 { &vop_remove_desc, fifo_remove }, /* remove */ 240 { &vop_link_desc, fifo_link }, /* link */ 241 { &vop_rename_desc, fifo_rename }, /* rename */ 242 { &vop_mkdir_desc, fifo_mkdir }, /* mkdir */ 243 { &vop_rmdir_desc, fifo_rmdir }, /* rmdir */ 244 { &vop_symlink_desc, fifo_symlink }, /* symlink */ 245 { &vop_readdir_desc, fifo_readdir }, /* readdir */ 246 { &vop_readlink_desc, fifo_readlink }, /* readlink */ 247 { &vop_abortop_desc, fifo_abortop }, /* abortop */ 248 { &vop_inactive_desc, lfs_inactive }, /* inactive */ 249 { &vop_reclaim_desc, lfs_reclaim }, /* reclaim */ 250 { &vop_lock_desc, ufs_lock }, /* lock */ 251 { &vop_unlock_desc, ufs_unlock }, /* unlock */ 252 { &vop_bmap_desc, fifo_bmap }, /* bmap */ 253 { &vop_strategy_desc, fifo_strategy }, /* strategy */ 254 { &vop_print_desc, ufs_print }, /* print */ 255 { &vop_islocked_desc, ufs_islocked }, /* islocked */ 256 { &vop_pathconf_desc, fifo_pathconf }, /* pathconf */ 257 { &vop_advlock_desc, fifo_advlock }, /* advlock */ 258 { &vop_blkatoff_desc, fifo_blkatoff }, /* blkatoff */ 259 { &vop_valloc_desc, fifo_valloc }, /* valloc */ 260 { &vop_vfree_desc, lfs_vfree }, /* vfree */ 261 { &vop_truncate_desc, fifo_truncate }, /* truncate */ 262 { &vop_update_desc, lfs_update }, /* update */ 263 { &vop_bwrite_desc, lfs_bwrite }, /* bwrite */ 264 { &vop_putpages_desc, fifo_putpages }, /* putpages */ 265 { NULL, NULL } 266 }; 267 const struct vnodeopv_desc lfs_fifoop_opv_desc = 268 { &lfs_fifoop_p, lfs_fifoop_entries }; 269 270 /* 271 * A function version of LFS_ITIMES, for the UFS functions which call ITIMES 272 */ 273 void 274 lfs_itimes(struct inode *ip, struct timespec *acc, struct timespec *mod, struct timespec *cre) 275 { 276 LFS_ITIMES(ip, acc, mod, cre); 277 } 278 279 #define LFS_READWRITE 280 #include <ufs/ufs/ufs_readwrite.c> 281 #undef LFS_READWRITE 282 283 /* 284 * Synch an open file. 285 */ 286 /* ARGSUSED */ 287 int 288 lfs_fsync(void *v) 289 { 290 struct vop_fsync_args /* { 291 struct vnode *a_vp; 292 struct ucred *a_cred; 293 int a_flags; 294 off_t offlo; 295 off_t offhi; 296 struct lwp *a_l; 297 } */ *ap = v; 298 struct vnode *vp = ap->a_vp; 299 int error, wait; 300 301 /* 302 * Trickle sync checks for need to do a checkpoint after possible 303 * activity from the pagedaemon. 304 */ 305 if (ap->a_flags & FSYNC_LAZY) { 306 simple_lock(&lfs_subsys_lock); 307 wakeup(&lfs_writer_daemon); 308 simple_unlock(&lfs_subsys_lock); 309 return 0; 310 } 311 312 wait = (ap->a_flags & FSYNC_WAIT); 313 simple_lock(&vp->v_interlock); 314 error = VOP_PUTPAGES(vp, trunc_page(ap->a_offlo), 315 round_page(ap->a_offhi), 316 PGO_CLEANIT | (wait ? PGO_SYNCIO : 0)); 317 if (error) 318 return error; 319 error = VOP_UPDATE(vp, NULL, NULL, wait ? UPDATE_WAIT : 0); 320 if (error == 0 && ap->a_flags & FSYNC_CACHE) { 321 int l = 0; 322 error = VOP_IOCTL(VTOI(vp)->i_devvp, DIOCCACHESYNC, &l, FWRITE, 323 ap->a_l->l_proc->p_ucred, ap->a_l); 324 } 325 if (wait && !VPISEMPTY(vp)) 326 LFS_SET_UINO(VTOI(vp), IN_MODIFIED); 327 328 return error; 329 } 330 331 /* 332 * Take IN_ADIROP off, then call ufs_inactive. 333 */ 334 int 335 lfs_inactive(void *v) 336 { 337 struct vop_inactive_args /* { 338 struct vnode *a_vp; 339 struct lwp *a_l; 340 } */ *ap = v; 341 342 KASSERT(VTOI(ap->a_vp)->i_nlink == VTOI(ap->a_vp)->i_ffs_effnlink); 343 344 lfs_unmark_vnode(ap->a_vp); 345 346 /* 347 * The Ifile is only ever inactivated on unmount. 348 * Streamline this process by not giving it more dirty blocks. 349 */ 350 if (VTOI(ap->a_vp)->i_number == LFS_IFILE_INUM) { 351 LFS_CLR_UINO(VTOI(ap->a_vp), IN_ALLMOD); 352 VOP_UNLOCK(ap->a_vp, 0); 353 return 0; 354 } 355 356 return ufs_inactive(v); 357 } 358 359 /* 360 * These macros are used to bracket UFS directory ops, so that we can 361 * identify all the pages touched during directory ops which need to 362 * be ordered and flushed atomically, so that they may be recovered. 363 */ 364 /* 365 * XXX KS - Because we have to mark nodes VDIROP in order to prevent 366 * the cache from reclaiming them while a dirop is in progress, we must 367 * also manage the number of nodes so marked (otherwise we can run out). 368 * We do this by setting lfs_dirvcount to the number of marked vnodes; it 369 * is decremented during segment write, when VDIROP is taken off. 370 */ 371 #define SET_DIROP(vp) SET_DIROP2((vp), NULL) 372 #define SET_DIROP2(vp, vp2) lfs_set_dirop((vp), (vp2)) 373 static int lfs_set_dirop(struct vnode *, struct vnode *); 374 375 #define NRESERVE(fs) (btofsb(fs, (NIADDR + 3 + (2 * NIADDR + 3)) << fs->lfs_bshift)) 376 377 static int 378 lfs_set_dirop(struct vnode *vp, struct vnode *vp2) 379 { 380 struct lfs *fs; 381 int error; 382 383 KASSERT(VOP_ISLOCKED(vp)); 384 KASSERT(vp2 == NULL || VOP_ISLOCKED(vp2)); 385 386 fs = VTOI(vp)->i_lfs; 387 /* 388 * We might need one directory block plus supporting indirect blocks, 389 * plus an inode block and ifile page for the new vnode. 390 */ 391 if ((error = lfs_reserve(fs, vp, vp2, NRESERVE(fs))) != 0) 392 return (error); 393 394 if (fs->lfs_dirops == 0) 395 lfs_check(vp, LFS_UNUSED_LBN, 0); 396 restart: 397 simple_lock(&fs->lfs_interlock); 398 if (fs->lfs_writer) { 399 ltsleep(&fs->lfs_dirops, (PRIBIO + 1) | PNORELOCK, 400 "lfs_sdirop", 0, &fs->lfs_interlock); 401 goto restart; 402 } 403 simple_lock(&lfs_subsys_lock); 404 if (lfs_dirvcount > LFS_MAX_DIROP && fs->lfs_dirops == 0) { 405 wakeup(&lfs_writer_daemon); 406 simple_unlock(&lfs_subsys_lock); 407 simple_unlock(&fs->lfs_interlock); 408 preempt(1); 409 goto restart; 410 } 411 412 if (lfs_dirvcount > LFS_MAX_DIROP) { 413 simple_unlock(&fs->lfs_interlock); 414 #ifdef DEBUG_LFS 415 printf("lfs_set_dirop: sleeping with dirops=%d, " 416 "dirvcount=%d\n", fs->lfs_dirops, lfs_dirvcount); 417 #endif 418 if ((error = ltsleep(&lfs_dirvcount, 419 PCATCH | PUSER | PNORELOCK, "lfs_maxdirop", 0, 420 &lfs_subsys_lock)) != 0) { 421 goto unreserve; 422 } 423 goto restart; 424 } 425 simple_unlock(&lfs_subsys_lock); 426 427 ++fs->lfs_dirops; 428 fs->lfs_doifile = 1; 429 simple_unlock(&fs->lfs_interlock); 430 431 /* Hold a reference so SET_ENDOP will be happy */ 432 vref(vp); 433 if (vp2) 434 vref(vp2); 435 436 return 0; 437 438 unreserve: 439 lfs_reserve(fs, vp, vp2, -NRESERVE(fs)); 440 return error; 441 } 442 443 #define SET_ENDOP(fs, vp, str) SET_ENDOP2((fs), (vp), NULL, (str)) 444 #define SET_ENDOP2(fs, vp, vp2, str) { \ 445 --(fs)->lfs_dirops; \ 446 if (!(fs)->lfs_dirops) { \ 447 if ((fs)->lfs_nadirop) { \ 448 panic("SET_ENDOP: %s: no dirops but nadirop=%d", \ 449 (str), (fs)->lfs_nadirop); \ 450 } \ 451 wakeup(&(fs)->lfs_writer); \ 452 lfs_check((vp),LFS_UNUSED_LBN,0); \ 453 } \ 454 lfs_reserve((fs), vp, vp2, -NRESERVE(fs)); /* XXX */ \ 455 vrele(vp); \ 456 if (vp2) \ 457 vrele(vp2); \ 458 } 459 460 #define MARK_VNODE(vp) lfs_mark_vnode(vp) 461 #define UNMARK_VNODE(vp) lfs_unmark_vnode(vp) 462 463 void 464 lfs_mark_vnode(struct vnode *vp) 465 { 466 struct inode *ip = VTOI(vp); 467 struct lfs *fs = ip->i_lfs; 468 469 if (!(ip->i_flag & IN_ADIROP)) { 470 if (!(vp->v_flag & VDIROP)) { 471 (void)lfs_vref(vp); 472 ++lfs_dirvcount; 473 TAILQ_INSERT_TAIL(&fs->lfs_dchainhd, ip, i_lfs_dchain); 474 vp->v_flag |= VDIROP; 475 } 476 ++fs->lfs_nadirop; 477 ip->i_flag |= IN_ADIROP; 478 } else 479 KASSERT(vp->v_flag & VDIROP); 480 } 481 482 void 483 lfs_unmark_vnode(struct vnode *vp) 484 { 485 struct inode *ip = VTOI(vp); 486 487 if (ip->i_flag & IN_ADIROP) { 488 KASSERT(vp->v_flag & VDIROP); 489 --ip->i_lfs->lfs_nadirop; 490 ip->i_flag &= ~IN_ADIROP; 491 } 492 } 493 494 int 495 lfs_symlink(void *v) 496 { 497 struct vop_symlink_args /* { 498 struct vnode *a_dvp; 499 struct vnode **a_vpp; 500 struct componentname *a_cnp; 501 struct vattr *a_vap; 502 char *a_target; 503 } */ *ap = v; 504 int error; 505 506 if ((error = SET_DIROP(ap->a_dvp)) != 0) { 507 vput(ap->a_dvp); 508 return error; 509 } 510 MARK_VNODE(ap->a_dvp); 511 error = ufs_symlink(ap); 512 UNMARK_VNODE(ap->a_dvp); 513 if (*(ap->a_vpp)) 514 UNMARK_VNODE(*(ap->a_vpp)); 515 SET_ENDOP(VTOI(ap->a_dvp)->i_lfs,ap->a_dvp,"symlink"); 516 return (error); 517 } 518 519 int 520 lfs_mknod(void *v) 521 { 522 struct vop_mknod_args /* { 523 struct vnode *a_dvp; 524 struct vnode **a_vpp; 525 struct componentname *a_cnp; 526 struct vattr *a_vap; 527 } */ *ap = v; 528 struct vattr *vap = ap->a_vap; 529 struct vnode **vpp = ap->a_vpp; 530 struct componentname *cnp = ap->a_cnp; 531 struct inode *ip; 532 int error; 533 struct mount *mp; 534 ino_t ino; 535 536 if ((error = SET_DIROP(ap->a_dvp)) != 0) { 537 vput(ap->a_dvp); 538 return error; 539 } 540 MARK_VNODE(ap->a_dvp); 541 error = ufs_makeinode(MAKEIMODE(vap->va_type, vap->va_mode), 542 ap->a_dvp, vpp, cnp); 543 UNMARK_VNODE(ap->a_dvp); 544 if (*(ap->a_vpp)) 545 UNMARK_VNODE(*(ap->a_vpp)); 546 547 /* Either way we're done with the dirop at this point */ 548 SET_ENDOP(VTOI(ap->a_dvp)->i_lfs,ap->a_dvp,"mknod"); 549 550 if (error) 551 return (error); 552 553 ip = VTOI(*vpp); 554 mp = (*vpp)->v_mount; 555 ino = ip->i_number; 556 ip->i_flag |= IN_ACCESS | IN_CHANGE | IN_UPDATE; 557 if (vap->va_rdev != VNOVAL) { 558 /* 559 * Want to be able to use this to make badblock 560 * inodes, so don't truncate the dev number. 561 */ 562 #if 0 563 ip->i_ffs1_rdev = ufs_rw32(vap->va_rdev, 564 UFS_MPNEEDSWAP((*vpp)->v_mount)); 565 #else 566 ip->i_ffs1_rdev = vap->va_rdev; 567 #endif 568 } 569 /* 570 * Call fsync to write the vnode so that we don't have to deal with 571 * flushing it when it's marked VDIROP|VXLOCK. 572 * 573 * XXX KS - If we can't flush we also can't call vgone(), so must 574 * return. But, that leaves this vnode in limbo, also not good. 575 * Can this ever happen (barring hardware failure)? 576 */ 577 if ((error = VOP_FSYNC(*vpp, NOCRED, FSYNC_WAIT, 0, 0, 578 cnp->cn_lwp)) != 0) { 579 printf("Couldn't fsync in mknod (ino %d)---what do I do?\n", 580 VTOI(*vpp)->i_number); 581 return (error); 582 } 583 /* 584 * Remove vnode so that it will be reloaded by VFS_VGET and 585 * checked to see if it is an alias of an existing entry in 586 * the inode cache. 587 */ 588 /* Used to be vput, but that causes us to call VOP_INACTIVE twice. */ 589 VOP_UNLOCK(*vpp, 0); 590 lfs_vunref(*vpp); 591 (*vpp)->v_type = VNON; 592 vgone(*vpp); 593 error = VFS_VGET(mp, ino, vpp); 594 if (error != 0) { 595 *vpp = NULL; 596 return (error); 597 } 598 return (0); 599 } 600 601 int 602 lfs_create(void *v) 603 { 604 struct vop_create_args /* { 605 struct vnode *a_dvp; 606 struct vnode **a_vpp; 607 struct componentname *a_cnp; 608 struct vattr *a_vap; 609 } */ *ap = v; 610 int error; 611 612 if ((error = SET_DIROP(ap->a_dvp)) != 0) { 613 vput(ap->a_dvp); 614 return error; 615 } 616 MARK_VNODE(ap->a_dvp); 617 error = ufs_create(ap); 618 UNMARK_VNODE(ap->a_dvp); 619 if (*(ap->a_vpp)) 620 UNMARK_VNODE(*(ap->a_vpp)); 621 SET_ENDOP(VTOI(ap->a_dvp)->i_lfs,ap->a_dvp,"create"); 622 return (error); 623 } 624 625 int 626 lfs_mkdir(void *v) 627 { 628 struct vop_mkdir_args /* { 629 struct vnode *a_dvp; 630 struct vnode **a_vpp; 631 struct componentname *a_cnp; 632 struct vattr *a_vap; 633 } */ *ap = v; 634 int error; 635 636 if ((error = SET_DIROP(ap->a_dvp)) != 0) { 637 vput(ap->a_dvp); 638 return error; 639 } 640 MARK_VNODE(ap->a_dvp); 641 error = ufs_mkdir(ap); 642 UNMARK_VNODE(ap->a_dvp); 643 if (*(ap->a_vpp)) 644 UNMARK_VNODE(*(ap->a_vpp)); 645 SET_ENDOP(VTOI(ap->a_dvp)->i_lfs,ap->a_dvp,"mkdir"); 646 return (error); 647 } 648 649 int 650 lfs_remove(void *v) 651 { 652 struct vop_remove_args /* { 653 struct vnode *a_dvp; 654 struct vnode *a_vp; 655 struct componentname *a_cnp; 656 } */ *ap = v; 657 struct vnode *dvp, *vp; 658 int error; 659 660 dvp = ap->a_dvp; 661 vp = ap->a_vp; 662 if ((error = SET_DIROP2(dvp, vp)) != 0) { 663 if (dvp == vp) 664 vrele(vp); 665 else 666 vput(vp); 667 vput(dvp); 668 return error; 669 } 670 MARK_VNODE(dvp); 671 MARK_VNODE(vp); 672 error = ufs_remove(ap); 673 UNMARK_VNODE(dvp); 674 UNMARK_VNODE(vp); 675 676 SET_ENDOP2(VTOI(dvp)->i_lfs, dvp, vp, "remove"); 677 return (error); 678 } 679 680 int 681 lfs_rmdir(void *v) 682 { 683 struct vop_rmdir_args /* { 684 struct vnodeop_desc *a_desc; 685 struct vnode *a_dvp; 686 struct vnode *a_vp; 687 struct componentname *a_cnp; 688 } */ *ap = v; 689 struct vnode *vp; 690 int error; 691 692 vp = ap->a_vp; 693 if ((error = SET_DIROP2(ap->a_dvp, ap->a_vp)) != 0) { 694 vrele(ap->a_dvp); 695 if (ap->a_vp != ap->a_dvp) 696 VOP_UNLOCK(ap->a_dvp, 0); 697 vput(vp); 698 return error; 699 } 700 MARK_VNODE(ap->a_dvp); 701 MARK_VNODE(vp); 702 error = ufs_rmdir(ap); 703 UNMARK_VNODE(ap->a_dvp); 704 UNMARK_VNODE(vp); 705 706 SET_ENDOP2(VTOI(ap->a_dvp)->i_lfs, ap->a_dvp, vp, "rmdir"); 707 return (error); 708 } 709 710 int 711 lfs_link(void *v) 712 { 713 struct vop_link_args /* { 714 struct vnode *a_dvp; 715 struct vnode *a_vp; 716 struct componentname *a_cnp; 717 } */ *ap = v; 718 int error; 719 720 if ((error = SET_DIROP(ap->a_dvp)) != 0) { 721 vput(ap->a_dvp); 722 return error; 723 } 724 MARK_VNODE(ap->a_dvp); 725 error = ufs_link(ap); 726 UNMARK_VNODE(ap->a_dvp); 727 SET_ENDOP(VTOI(ap->a_dvp)->i_lfs,ap->a_dvp,"link"); 728 return (error); 729 } 730 731 int 732 lfs_rename(void *v) 733 { 734 struct vop_rename_args /* { 735 struct vnode *a_fdvp; 736 struct vnode *a_fvp; 737 struct componentname *a_fcnp; 738 struct vnode *a_tdvp; 739 struct vnode *a_tvp; 740 struct componentname *a_tcnp; 741 } */ *ap = v; 742 struct vnode *tvp, *fvp, *tdvp, *fdvp; 743 struct componentname *tcnp, *fcnp; 744 int error; 745 struct lfs *fs; 746 747 fs = VTOI(ap->a_fdvp)->i_lfs; 748 tvp = ap->a_tvp; 749 tdvp = ap->a_tdvp; 750 tcnp = ap->a_tcnp; 751 fvp = ap->a_fvp; 752 fdvp = ap->a_fdvp; 753 fcnp = ap->a_fcnp; 754 755 /* 756 * Check for cross-device rename. 757 * If it is, we don't want to set dirops, just error out. 758 * (In particular note that MARK_VNODE(tdvp) will DTWT on 759 * a cross-device rename.) 760 * 761 * Copied from ufs_rename. 762 */ 763 if ((fvp->v_mount != tdvp->v_mount) || 764 (tvp && (fvp->v_mount != tvp->v_mount))) { 765 error = EXDEV; 766 goto errout; 767 } 768 769 /* 770 * Check to make sure we're not renaming a vnode onto itself 771 * (deleting a hard link by renaming one name onto another); 772 * if we are we can't recursively call VOP_REMOVE since that 773 * would leave us with an unaccounted-for number of live dirops. 774 * 775 * Inline the relevant section of ufs_rename here, *before* 776 * calling SET_DIROP2. 777 */ 778 if (tvp && ((VTOI(tvp)->i_flags & (IMMUTABLE | APPEND)) || 779 (VTOI(tdvp)->i_flags & APPEND))) { 780 error = EPERM; 781 goto errout; 782 } 783 if (fvp == tvp) { 784 if (fvp->v_type == VDIR) { 785 error = EINVAL; 786 goto errout; 787 } 788 789 /* Release destination completely. */ 790 VOP_ABORTOP(tdvp, tcnp); 791 vput(tdvp); 792 vput(tvp); 793 794 /* Delete source. */ 795 vrele(fvp); 796 fcnp->cn_flags &= ~(MODMASK | SAVESTART); 797 fcnp->cn_flags |= LOCKPARENT | LOCKLEAF; 798 fcnp->cn_nameiop = DELETE; 799 if ((error = relookup(fdvp, &fvp, fcnp))){ 800 /* relookup blew away fdvp */ 801 return (error); 802 } 803 return (VOP_REMOVE(fdvp, fvp, fcnp)); 804 } 805 806 if ((error = SET_DIROP2(tdvp, tvp)) != 0) 807 goto errout; 808 MARK_VNODE(fdvp); 809 MARK_VNODE(tdvp); 810 MARK_VNODE(fvp); 811 if (tvp) { 812 MARK_VNODE(tvp); 813 } 814 815 error = ufs_rename(ap); 816 UNMARK_VNODE(fdvp); 817 UNMARK_VNODE(tdvp); 818 UNMARK_VNODE(fvp); 819 if (tvp) { 820 UNMARK_VNODE(tvp); 821 } 822 SET_ENDOP2(fs, tdvp, tvp, "rename"); 823 return (error); 824 825 errout: 826 VOP_ABORTOP(tdvp, ap->a_tcnp); /* XXX, why not in NFS? */ 827 if (tdvp == tvp) 828 vrele(tdvp); 829 else 830 vput(tdvp); 831 if (tvp) 832 vput(tvp); 833 VOP_ABORTOP(fdvp, ap->a_fcnp); /* XXX, why not in NFS? */ 834 vrele(fdvp); 835 vrele(fvp); 836 return (error); 837 } 838 839 /* XXX hack to avoid calling ITIMES in getattr */ 840 int 841 lfs_getattr(void *v) 842 { 843 struct vop_getattr_args /* { 844 struct vnode *a_vp; 845 struct vattr *a_vap; 846 struct ucred *a_cred; 847 struct lwp *a_l; 848 } */ *ap = v; 849 struct vnode *vp = ap->a_vp; 850 struct inode *ip = VTOI(vp); 851 struct vattr *vap = ap->a_vap; 852 struct lfs *fs = ip->i_lfs; 853 /* 854 * Copy from inode table 855 */ 856 vap->va_fsid = ip->i_dev; 857 vap->va_fileid = ip->i_number; 858 vap->va_mode = ip->i_mode & ~IFMT; 859 vap->va_nlink = ip->i_nlink; 860 vap->va_uid = ip->i_uid; 861 vap->va_gid = ip->i_gid; 862 vap->va_rdev = (dev_t)ip->i_ffs1_rdev; 863 vap->va_size = vp->v_size; 864 vap->va_atime.tv_sec = ip->i_ffs1_atime; 865 vap->va_atime.tv_nsec = ip->i_ffs1_atimensec; 866 vap->va_mtime.tv_sec = ip->i_ffs1_mtime; 867 vap->va_mtime.tv_nsec = ip->i_ffs1_mtimensec; 868 vap->va_ctime.tv_sec = ip->i_ffs1_ctime; 869 vap->va_ctime.tv_nsec = ip->i_ffs1_ctimensec; 870 vap->va_flags = ip->i_flags; 871 vap->va_gen = ip->i_gen; 872 /* this doesn't belong here */ 873 if (vp->v_type == VBLK) 874 vap->va_blocksize = BLKDEV_IOSIZE; 875 else if (vp->v_type == VCHR) 876 vap->va_blocksize = MAXBSIZE; 877 else 878 vap->va_blocksize = vp->v_mount->mnt_stat.f_iosize; 879 vap->va_bytes = fsbtob(fs, (u_quad_t)ip->i_lfs_effnblks); 880 vap->va_type = vp->v_type; 881 vap->va_filerev = ip->i_modrev; 882 return (0); 883 } 884 885 /* 886 * Check to make sure the inode blocks won't choke the buffer 887 * cache, then call ufs_setattr as usual. 888 */ 889 int 890 lfs_setattr(void *v) 891 { 892 struct vop_getattr_args /* { 893 struct vnode *a_vp; 894 struct vattr *a_vap; 895 struct ucred *a_cred; 896 struct lwp *a_l; 897 } */ *ap = v; 898 struct vnode *vp = ap->a_vp; 899 900 lfs_check(vp, LFS_UNUSED_LBN, 0); 901 return ufs_setattr(v); 902 } 903 904 /* 905 * Close called 906 * 907 * XXX -- we were using ufs_close, but since it updates the 908 * times on the inode, we might need to bump the uinodes 909 * count. 910 */ 911 /* ARGSUSED */ 912 int 913 lfs_close(void *v) 914 { 915 struct vop_close_args /* { 916 struct vnode *a_vp; 917 int a_fflag; 918 struct ucred *a_cred; 919 struct lwp *a_l; 920 } */ *ap = v; 921 struct vnode *vp = ap->a_vp; 922 struct inode *ip = VTOI(vp); 923 struct timespec ts; 924 925 if (vp == ip->i_lfs->lfs_ivnode && 926 vp->v_mount->mnt_iflag & IMNT_UNMOUNT) 927 return 0; 928 929 if (vp->v_usecount > 1 && vp != ip->i_lfs->lfs_ivnode) { 930 TIMEVAL_TO_TIMESPEC(&time, &ts); 931 LFS_ITIMES(ip, &ts, &ts, &ts); 932 } 933 return (0); 934 } 935 936 /* 937 * Close wrapper for special devices. 938 * 939 * Update the times on the inode then do device close. 940 */ 941 int 942 lfsspec_close(void *v) 943 { 944 struct vop_close_args /* { 945 struct vnode *a_vp; 946 int a_fflag; 947 struct ucred *a_cred; 948 struct lwp *a_l; 949 } */ *ap = v; 950 struct vnode *vp; 951 struct inode *ip; 952 struct timespec ts; 953 954 vp = ap->a_vp; 955 ip = VTOI(vp); 956 if (vp->v_usecount > 1) { 957 TIMEVAL_TO_TIMESPEC(&time, &ts); 958 LFS_ITIMES(ip, &ts, &ts, &ts); 959 } 960 return (VOCALL (spec_vnodeop_p, VOFFSET(vop_close), ap)); 961 } 962 963 /* 964 * Close wrapper for fifo's. 965 * 966 * Update the times on the inode then do device close. 967 */ 968 int 969 lfsfifo_close(void *v) 970 { 971 struct vop_close_args /* { 972 struct vnode *a_vp; 973 int a_fflag; 974 struct ucred *a_cred; 975 struct lwp *a_l; 976 } */ *ap = v; 977 struct vnode *vp; 978 struct inode *ip; 979 struct timespec ts; 980 981 vp = ap->a_vp; 982 ip = VTOI(vp); 983 if (ap->a_vp->v_usecount > 1) { 984 TIMEVAL_TO_TIMESPEC(&time, &ts); 985 LFS_ITIMES(ip, &ts, &ts, &ts); 986 } 987 return (VOCALL (fifo_vnodeop_p, VOFFSET(vop_close), ap)); 988 } 989 990 /* 991 * Reclaim an inode so that it can be used for other purposes. 992 */ 993 994 int 995 lfs_reclaim(void *v) 996 { 997 struct vop_reclaim_args /* { 998 struct vnode *a_vp; 999 struct lwp *a_l; 1000 } */ *ap = v; 1001 struct vnode *vp = ap->a_vp; 1002 struct inode *ip = VTOI(vp); 1003 int error; 1004 1005 KASSERT(ip->i_nlink == ip->i_ffs_effnlink); 1006 1007 LFS_CLR_UINO(ip, IN_ALLMOD); 1008 if ((error = ufs_reclaim(vp, ap->a_l))) 1009 return (error); 1010 pool_put(&lfs_dinode_pool, VTOI(vp)->i_din.ffs1_din); 1011 pool_put(&lfs_inoext_pool, ip->inode_ext.lfs); 1012 ip->inode_ext.lfs = NULL; 1013 pool_put(&lfs_inode_pool, vp->v_data); 1014 vp->v_data = NULL; 1015 return (0); 1016 } 1017 1018 /* 1019 * Read a block from a storage device. 1020 * In order to avoid reading blocks that are in the process of being 1021 * written by the cleaner---and hence are not mutexed by the normal 1022 * buffer cache / page cache mechanisms---check for collisions before 1023 * reading. 1024 * 1025 * We inline ufs_strategy to make sure that the VOP_BMAP occurs *before* 1026 * the active cleaner test. 1027 * 1028 * XXX This code assumes that lfs_markv makes synchronous checkpoints. 1029 */ 1030 int 1031 lfs_strategy(void *v) 1032 { 1033 struct vop_strategy_args /* { 1034 struct vnode *a_vp; 1035 struct buf *a_bp; 1036 } */ *ap = v; 1037 struct buf *bp; 1038 struct lfs *fs; 1039 struct vnode *vp; 1040 struct inode *ip; 1041 daddr_t tbn; 1042 int i, sn, error, slept; 1043 1044 bp = ap->a_bp; 1045 vp = ap->a_vp; 1046 ip = VTOI(vp); 1047 fs = ip->i_lfs; 1048 1049 /* lfs uses its strategy routine only for read */ 1050 KASSERT(bp->b_flags & B_READ); 1051 1052 if (vp->v_type == VBLK || vp->v_type == VCHR) 1053 panic("lfs_strategy: spec"); 1054 KASSERT(bp->b_bcount != 0); 1055 if (bp->b_blkno == bp->b_lblkno) { 1056 error = VOP_BMAP(vp, bp->b_lblkno, NULL, &bp->b_blkno, 1057 NULL); 1058 if (error) { 1059 bp->b_error = error; 1060 bp->b_flags |= B_ERROR; 1061 biodone(bp); 1062 return (error); 1063 } 1064 if ((long)bp->b_blkno == -1) /* no valid data */ 1065 clrbuf(bp); 1066 } 1067 if ((long)bp->b_blkno < 0) { /* block is not on disk */ 1068 biodone(bp); 1069 return (0); 1070 } 1071 1072 slept = 1; 1073 simple_lock(&fs->lfs_interlock); 1074 while (slept && fs->lfs_seglock) { 1075 simple_unlock(&fs->lfs_interlock); 1076 /* 1077 * Look through list of intervals. 1078 * There will only be intervals to look through 1079 * if the cleaner holds the seglock. 1080 * Since the cleaner is synchronous, we can trust 1081 * the list of intervals to be current. 1082 */ 1083 tbn = dbtofsb(fs, bp->b_blkno); 1084 sn = dtosn(fs, tbn); 1085 slept = 0; 1086 for (i = 0; i < fs->lfs_cleanind; i++) { 1087 if (sn == dtosn(fs, fs->lfs_cleanint[i]) && 1088 tbn >= fs->lfs_cleanint[i]) { 1089 #ifdef DEBUG_LFS 1090 printf("lfs_strategy: ino %d lbn %" PRId64 1091 " ind %d sn %d fsb %" PRIx32 1092 " given sn %d fsb %" PRIx64 "\n", 1093 ip->i_number, bp->b_lblkno, i, 1094 dtosn(fs, fs->lfs_cleanint[i]), 1095 fs->lfs_cleanint[i], sn, tbn); 1096 printf("lfs_strategy: sleeping on ino %d lbn %" 1097 PRId64 "\n", ip->i_number, bp->b_lblkno); 1098 #endif 1099 tsleep(&fs->lfs_seglock, PRIBIO+1, 1100 "lfs_strategy", 0); 1101 /* Things may be different now; start over. */ 1102 slept = 1; 1103 break; 1104 } 1105 } 1106 simple_lock(&fs->lfs_interlock); 1107 } 1108 simple_unlock(&fs->lfs_interlock); 1109 1110 vp = ip->i_devvp; 1111 VOP_STRATEGY(vp, bp); 1112 return (0); 1113 } 1114 1115 static void 1116 lfs_flush_dirops(struct lfs *fs) 1117 { 1118 struct inode *ip, *nip; 1119 struct vnode *vp; 1120 extern int lfs_dostats; 1121 struct segment *sp; 1122 int needunlock; 1123 1124 if (fs->lfs_ronly) 1125 return; 1126 1127 if (TAILQ_FIRST(&fs->lfs_dchainhd) == NULL) 1128 return; 1129 1130 if (lfs_dostats) 1131 ++lfs_stats.flush_invoked; 1132 1133 /* 1134 * Inline lfs_segwrite/lfs_writevnodes, but just for dirops. 1135 * Technically this is a checkpoint (the on-disk state is valid) 1136 * even though we are leaving out all the file data. 1137 */ 1138 lfs_imtime(fs); 1139 lfs_seglock(fs, SEGM_CKP); 1140 sp = fs->lfs_sp; 1141 1142 /* 1143 * lfs_writevnodes, optimized to get dirops out of the way. 1144 * Only write dirops, and don't flush files' pages, only 1145 * blocks from the directories. 1146 * 1147 * We don't need to vref these files because they are 1148 * dirops and so hold an extra reference until the 1149 * segunlock clears them of that status. 1150 * 1151 * We don't need to check for IN_ADIROP because we know that 1152 * no dirops are active. 1153 * 1154 */ 1155 for (ip = TAILQ_FIRST(&fs->lfs_dchainhd); ip != NULL; ip = nip) { 1156 nip = TAILQ_NEXT(ip, i_lfs_dchain); 1157 vp = ITOV(ip); 1158 1159 /* 1160 * All writes to directories come from dirops; all 1161 * writes to files' direct blocks go through the page 1162 * cache, which we're not touching. Reads to files 1163 * and/or directories will not be affected by writing 1164 * directory blocks inodes and file inodes. So we don't 1165 * really need to lock. If we don't lock, though, 1166 * make sure that we don't clear IN_MODIFIED 1167 * unnecessarily. 1168 */ 1169 if (vp->v_flag & VXLOCK) 1170 continue; 1171 if (vn_lock(vp, LK_EXCLUSIVE | LK_CANRECURSE | 1172 LK_NOWAIT) == 0) { 1173 needunlock = 1; 1174 } else { 1175 printf("lfs_flush_dirops: flushing locked ino %d\n", 1176 VTOI(vp)->i_number); 1177 needunlock = 0; 1178 } 1179 if (vp->v_type != VREG && 1180 ((ip->i_flag & IN_ALLMOD) || !VPISEMPTY(vp))) { 1181 lfs_writefile(fs, sp, vp); 1182 if (!VPISEMPTY(vp) && !WRITEINPROG(vp) && 1183 !(ip->i_flag & IN_ALLMOD)) { 1184 LFS_SET_UINO(ip, IN_MODIFIED); 1185 } 1186 } 1187 (void) lfs_writeinode(fs, sp, ip); 1188 if (needunlock) 1189 VOP_UNLOCK(vp, 0); 1190 else 1191 LFS_SET_UINO(ip, IN_MODIFIED); 1192 } 1193 /* We've written all the dirops there are */ 1194 ((SEGSUM *)(sp->segsum))->ss_flags &= ~(SS_CONT); 1195 (void) lfs_writeseg(fs, sp); 1196 lfs_segunlock(fs); 1197 } 1198 1199 /* 1200 * Provide a fcntl interface to sys_lfs_{segwait,bmapv,markv}. 1201 */ 1202 int 1203 lfs_fcntl(void *v) 1204 { 1205 struct vop_fcntl_args /* { 1206 struct vnode *a_vp; 1207 u_long a_command; 1208 caddr_t a_data; 1209 int a_fflag; 1210 struct ucred *a_cred; 1211 struct lwp *a_l; 1212 } */ *ap = v; 1213 struct timeval *tvp; 1214 BLOCK_INFO *blkiov; 1215 CLEANERINFO *cip; 1216 int blkcnt, error, oclean; 1217 struct lfs_fcntl_markv blkvp; 1218 struct proc *p; 1219 fsid_t *fsidp; 1220 struct lfs *fs; 1221 struct buf *bp; 1222 daddr_t off; 1223 1224 /* Only respect LFS fcntls on fs root or Ifile */ 1225 if (VTOI(ap->a_vp)->i_number != ROOTINO && 1226 VTOI(ap->a_vp)->i_number != LFS_IFILE_INUM) { 1227 return ufs_fcntl(v); 1228 } 1229 1230 /* Avoid locking a draining lock */ 1231 if (ap->a_vp->v_mount->mnt_iflag & IMNT_UNMOUNT) { 1232 return ESHUTDOWN; 1233 } 1234 1235 p = ap->a_l->l_proc; 1236 fs = VTOI(ap->a_vp)->i_lfs; 1237 fsidp = &ap->a_vp->v_mount->mnt_stat.f_fsidx; 1238 1239 switch (ap->a_command) { 1240 case LFCNSEGWAITALL: 1241 fsidp = NULL; 1242 /* FALLSTHROUGH */ 1243 case LFCNSEGWAIT: 1244 tvp = (struct timeval *)ap->a_data; 1245 simple_lock(&fs->lfs_interlock); 1246 ++fs->lfs_sleepers; 1247 simple_unlock(&fs->lfs_interlock); 1248 VOP_UNLOCK(ap->a_vp, 0); 1249 1250 error = lfs_segwait(fsidp, tvp); 1251 1252 VOP_LOCK(ap->a_vp, LK_EXCLUSIVE); 1253 simple_lock(&fs->lfs_interlock); 1254 if (--fs->lfs_sleepers == 0) 1255 wakeup(&fs->lfs_sleepers); 1256 simple_unlock(&fs->lfs_interlock); 1257 return error; 1258 1259 case LFCNBMAPV: 1260 case LFCNMARKV: 1261 if ((error = suser(p->p_ucred, &p->p_acflag)) != 0) 1262 return (error); 1263 blkvp = *(struct lfs_fcntl_markv *)ap->a_data; 1264 1265 blkcnt = blkvp.blkcnt; 1266 if ((u_int) blkcnt > LFS_MARKV_MAXBLKCNT) 1267 return (EINVAL); 1268 blkiov = malloc(blkcnt * sizeof(BLOCK_INFO), M_SEGMENT, M_WAITOK); 1269 if ((error = copyin(blkvp.blkiov, blkiov, 1270 blkcnt * sizeof(BLOCK_INFO))) != 0) { 1271 free(blkiov, M_SEGMENT); 1272 return error; 1273 } 1274 1275 simple_lock(&fs->lfs_interlock); 1276 ++fs->lfs_sleepers; 1277 simple_unlock(&fs->lfs_interlock); 1278 VOP_UNLOCK(ap->a_vp, 0); 1279 if (ap->a_command == LFCNBMAPV) 1280 error = lfs_bmapv(p, fsidp, blkiov, blkcnt); 1281 else /* LFCNMARKV */ 1282 error = lfs_markv(p, fsidp, blkiov, blkcnt); 1283 if (error == 0) 1284 error = copyout(blkiov, blkvp.blkiov, 1285 blkcnt * sizeof(BLOCK_INFO)); 1286 VOP_LOCK(ap->a_vp, LK_EXCLUSIVE); 1287 simple_lock(&fs->lfs_interlock); 1288 if (--fs->lfs_sleepers == 0) 1289 wakeup(&fs->lfs_sleepers); 1290 simple_unlock(&fs->lfs_interlock); 1291 free(blkiov, M_SEGMENT); 1292 return error; 1293 1294 case LFCNRECLAIM: 1295 /* 1296 * Flush dirops and write Ifile, allowing empty segments 1297 * to be immediately reclaimed. 1298 */ 1299 lfs_writer_enter(fs, "pndirop"); 1300 off = fs->lfs_offset; 1301 lfs_seglock(fs, SEGM_FORCE_CKP | SEGM_CKP); 1302 lfs_flush_dirops(fs); 1303 LFS_CLEANERINFO(cip, fs, bp); 1304 oclean = cip->clean; 1305 LFS_SYNC_CLEANERINFO(cip, fs, bp, 1); 1306 lfs_segwrite(ap->a_vp->v_mount, SEGM_FORCE_CKP); 1307 lfs_segunlock(fs); 1308 lfs_writer_leave(fs); 1309 1310 #ifdef DEBUG_LFS 1311 LFS_CLEANERINFO(cip, fs, bp); 1312 oclean = cip->clean; 1313 printf("lfs_fcntl: reclaim wrote %" PRId64 " blocks, cleaned " 1314 "%" PRId32 " segments (activesb %d)\n", 1315 fs->lfs_offset - off, cip->clean - oclean, 1316 fs->lfs_activesb); 1317 LFS_SYNC_CLEANERINFO(cip, fs, bp, 0); 1318 #endif 1319 1320 return 0; 1321 1322 default: 1323 return ufs_fcntl(v); 1324 } 1325 return 0; 1326 } 1327 1328 int 1329 lfs_getpages(void *v) 1330 { 1331 struct vop_getpages_args /* { 1332 struct vnode *a_vp; 1333 voff_t a_offset; 1334 struct vm_page **a_m; 1335 int *a_count; 1336 int a_centeridx; 1337 vm_prot_t a_access_type; 1338 int a_advice; 1339 int a_flags; 1340 } */ *ap = v; 1341 1342 if (VTOI(ap->a_vp)->i_number == LFS_IFILE_INUM && 1343 (ap->a_access_type & VM_PROT_WRITE) != 0) { 1344 return EPERM; 1345 } 1346 if ((ap->a_access_type & VM_PROT_WRITE) != 0) { 1347 LFS_SET_UINO(VTOI(ap->a_vp), IN_MODIFIED); 1348 } 1349 1350 /* 1351 * we're relying on the fact that genfs_getpages() always read in 1352 * entire filesystem blocks. 1353 */ 1354 return genfs_getpages(v); 1355 } 1356 1357 /* 1358 * Make sure that for all pages in every block in the given range, 1359 * either all are dirty or all are clean. If any of the pages 1360 * we've seen so far are dirty, put the vnode on the paging chain, 1361 * and mark it IN_PAGING. 1362 * 1363 * If checkfirst != 0, don't check all the pages but return at the 1364 * first dirty page. 1365 */ 1366 static int 1367 check_dirty(struct lfs *fs, struct vnode *vp, 1368 off_t startoffset, off_t endoffset, off_t blkeof, 1369 int flags, int checkfirst) 1370 { 1371 int by_list; 1372 struct vm_page *curpg = NULL; /* XXX: gcc */ 1373 struct vm_page *pgs[MAXBSIZE / PAGE_SIZE], *pg; 1374 off_t soff = 0; /* XXX: gcc */ 1375 voff_t off; 1376 int i; 1377 int nonexistent; 1378 int any_dirty; /* number of dirty pages */ 1379 int dirty; /* number of dirty pages in a block */ 1380 int tdirty; 1381 int pages_per_block = fs->lfs_bsize >> PAGE_SHIFT; 1382 1383 top: 1384 by_list = (vp->v_uobj.uo_npages <= 1385 ((endoffset - startoffset) >> PAGE_SHIFT) * 1386 UVM_PAGE_HASH_PENALTY); 1387 any_dirty = 0; 1388 1389 if (by_list) { 1390 curpg = TAILQ_FIRST(&vp->v_uobj.memq); 1391 } else { 1392 soff = startoffset; 1393 } 1394 while (by_list || soff < MIN(blkeof, endoffset)) { 1395 if (by_list) { 1396 /* 1397 * find the first page in a block. 1398 */ 1399 if (pages_per_block > 1) { 1400 while (curpg && (curpg->offset & fs->lfs_bmask)) 1401 curpg = TAILQ_NEXT(curpg, listq); 1402 } 1403 if (curpg == NULL) 1404 break; 1405 soff = curpg->offset; 1406 } 1407 1408 /* 1409 * Mark all pages in extended range busy; find out if any 1410 * of them are dirty. 1411 */ 1412 nonexistent = dirty = 0; 1413 for (i = 0; i == 0 || i < pages_per_block; i++) { 1414 if (by_list && pages_per_block <= 1) { 1415 pgs[i] = pg = curpg; 1416 } else { 1417 off = soff + (i << PAGE_SHIFT); 1418 pgs[i] = pg = uvm_pagelookup(&vp->v_uobj, off); 1419 if (pg == NULL) { 1420 ++nonexistent; 1421 continue; 1422 } 1423 } 1424 KASSERT(pg != NULL); 1425 while (pg->flags & PG_BUSY) { 1426 pg->flags |= PG_WANTED; 1427 UVM_UNLOCK_AND_WAIT(pg, &vp->v_interlock, 0, 1428 "lfsput", 0); 1429 simple_lock(&vp->v_interlock); 1430 if (by_list) { 1431 if (i > 0) 1432 uvm_page_unbusy(pgs, i); 1433 goto top; 1434 } 1435 } 1436 pg->flags |= PG_BUSY; 1437 UVM_PAGE_OWN(pg, "lfs_putpages"); 1438 1439 pmap_page_protect(pg, VM_PROT_NONE); 1440 tdirty = (pmap_clear_modify(pg) || 1441 (pg->flags & PG_CLEAN) == 0); 1442 dirty += tdirty; 1443 } 1444 if (pages_per_block > 0 && nonexistent >= pages_per_block) { 1445 if (by_list) { 1446 curpg = TAILQ_NEXT(curpg, listq); 1447 } else { 1448 soff += fs->lfs_bsize; 1449 } 1450 continue; 1451 } 1452 1453 any_dirty += dirty; 1454 KASSERT(nonexistent == 0); 1455 1456 /* 1457 * If any are dirty make all dirty; unbusy them, 1458 * but if we were asked to clean, wire them so that 1459 * the pagedaemon doesn't bother us about them while 1460 * they're on their way to disk. 1461 */ 1462 for (i = 0; i == 0 || i < pages_per_block; i++) { 1463 pg = pgs[i]; 1464 KASSERT(!((pg->flags & PG_CLEAN) && (pg->flags & PG_DELWRI))); 1465 if (dirty) { 1466 pg->flags &= ~PG_CLEAN; 1467 if (flags & PGO_FREE) { 1468 /* XXXUBC need better way to update */ 1469 simple_lock(&lfs_subsys_lock); 1470 lfs_subsys_pages += MIN(1, pages_per_block); 1471 simple_unlock(&lfs_subsys_lock); 1472 /* 1473 * Wire the page so that 1474 * pdaemon doesn't see it again. 1475 */ 1476 uvm_lock_pageq(); 1477 uvm_pagewire(pg); 1478 uvm_unlock_pageq(); 1479 1480 /* Suspended write flag */ 1481 pg->flags |= PG_DELWRI; 1482 } 1483 } 1484 if (pg->flags & PG_WANTED) 1485 wakeup(pg); 1486 pg->flags &= ~(PG_WANTED|PG_BUSY); 1487 UVM_PAGE_OWN(pg, NULL); 1488 } 1489 1490 if (checkfirst && any_dirty) 1491 break; 1492 1493 if (by_list) { 1494 curpg = TAILQ_NEXT(curpg, listq); 1495 } else { 1496 soff += MAX(PAGE_SIZE, fs->lfs_bsize); 1497 } 1498 } 1499 1500 /* 1501 * If any pages were dirty, mark this inode as "pageout requested", 1502 * and put it on the paging queue. 1503 * XXXUBC locking (check locking on dchainhd too) 1504 */ 1505 #ifdef notyet 1506 if (any_dirty) { 1507 if (!(ip->i_flags & IN_PAGING)) { 1508 ip->i_flags |= IN_PAGING; 1509 TAILQ_INSERT_TAIL(&fs->lfs_pchainhd, ip, i_lfs_pchain); 1510 } 1511 } 1512 #endif 1513 return any_dirty; 1514 } 1515 1516 /* 1517 * lfs_putpages functions like genfs_putpages except that 1518 * 1519 * (1) It needs to bounds-check the incoming requests to ensure that 1520 * they are block-aligned; if they are not, expand the range and 1521 * do the right thing in case, e.g., the requested range is clean 1522 * but the expanded range is dirty. 1523 * (2) It needs to explicitly send blocks to be written when it is done. 1524 * VOP_PUTPAGES is not ever called with the seglock held, so 1525 * we simply take the seglock and let lfs_segunlock wait for us. 1526 * XXX Actually we can be called with the seglock held, if we have 1527 * XXX to flush a vnode while lfs_markv is in operation. As of this 1528 * XXX writing we panic in this case. 1529 * 1530 * Assumptions: 1531 * 1532 * (1) The caller does not hold any pages in this vnode busy. If it does, 1533 * there is a danger that when we expand the page range and busy the 1534 * pages we will deadlock. 1535 * (2) We are called with vp->v_interlock held; we must return with it 1536 * released. 1537 * (3) We don't absolutely have to free pages right away, provided that 1538 * the request does not have PGO_SYNCIO. When the pagedaemon gives 1539 * us a request with PGO_FREE, we take the pages out of the paging 1540 * queue and wake up the writer, which will handle freeing them for us. 1541 * 1542 * We ensure that for any filesystem block, all pages for that 1543 * block are either resident or not, even if those pages are higher 1544 * than EOF; that means that we will be getting requests to free 1545 * "unused" pages above EOF all the time, and should ignore them. 1546 * 1547 * XXX note that we're (ab)using PGO_LOCKED as "seglock held". 1548 */ 1549 1550 int 1551 lfs_putpages(void *v) 1552 { 1553 int error; 1554 struct vop_putpages_args /* { 1555 struct vnode *a_vp; 1556 voff_t a_offlo; 1557 voff_t a_offhi; 1558 int a_flags; 1559 } */ *ap = v; 1560 struct vnode *vp; 1561 struct inode *ip; 1562 struct lfs *fs; 1563 struct segment *sp; 1564 off_t origoffset, startoffset, endoffset, origendoffset, blkeof; 1565 off_t off, max_endoffset; 1566 int s; 1567 boolean_t seglocked, sync, pagedaemon; 1568 struct vm_page *pg; 1569 UVMHIST_FUNC("lfs_putpages"); UVMHIST_CALLED(ubchist); 1570 1571 vp = ap->a_vp; 1572 ip = VTOI(vp); 1573 fs = ip->i_lfs; 1574 sync = (ap->a_flags & PGO_SYNCIO) != 0; 1575 pagedaemon = (curproc == uvm.pagedaemon_proc); 1576 1577 /* Putpages does nothing for metadata. */ 1578 if (vp == fs->lfs_ivnode || vp->v_type != VREG) { 1579 simple_unlock(&vp->v_interlock); 1580 return 0; 1581 } 1582 1583 /* 1584 * If there are no pages, don't do anything. 1585 */ 1586 if (vp->v_uobj.uo_npages == 0) { 1587 s = splbio(); 1588 if (LIST_FIRST(&vp->v_dirtyblkhd) == NULL && 1589 (vp->v_flag & VONWORKLST)) { 1590 vp->v_flag &= ~VONWORKLST; 1591 LIST_REMOVE(vp, v_synclist); 1592 } 1593 splx(s); 1594 simple_unlock(&vp->v_interlock); 1595 return 0; 1596 } 1597 1598 blkeof = blkroundup(fs, ip->i_size); 1599 1600 /* 1601 * Ignore requests to free pages past EOF but in the same block 1602 * as EOF, unless the request is synchronous. (XXX why sync?) 1603 * XXXUBC Make these pages look "active" so the pagedaemon won't 1604 * XXXUBC bother us with them again. 1605 */ 1606 if (!sync && ap->a_offlo >= ip->i_size && ap->a_offlo < blkeof) { 1607 origoffset = ap->a_offlo; 1608 for (off = origoffset; off < blkeof; off += fs->lfs_bsize) { 1609 pg = uvm_pagelookup(&vp->v_uobj, off); 1610 KASSERT(pg != NULL); 1611 while (pg->flags & PG_BUSY) { 1612 pg->flags |= PG_WANTED; 1613 UVM_UNLOCK_AND_WAIT(pg, &vp->v_interlock, 0, 1614 "lfsput2", 0); 1615 simple_lock(&vp->v_interlock); 1616 } 1617 uvm_lock_pageq(); 1618 uvm_pageactivate(pg); 1619 uvm_unlock_pageq(); 1620 } 1621 ap->a_offlo = blkeof; 1622 if (ap->a_offhi > 0 && ap->a_offhi <= ap->a_offlo) { 1623 simple_unlock(&vp->v_interlock); 1624 return 0; 1625 } 1626 } 1627 1628 /* 1629 * Extend page range to start and end at block boundaries. 1630 * (For the purposes of VOP_PUTPAGES, fragments don't exist.) 1631 */ 1632 origoffset = ap->a_offlo; 1633 origendoffset = ap->a_offhi; 1634 startoffset = origoffset & ~(fs->lfs_bmask); 1635 max_endoffset = (trunc_page(LLONG_MAX) >> fs->lfs_bshift) 1636 << fs->lfs_bshift; 1637 1638 if (origendoffset == 0 || ap->a_flags & PGO_ALLPAGES) { 1639 endoffset = max_endoffset; 1640 origendoffset = endoffset; 1641 } else { 1642 origendoffset = round_page(ap->a_offhi); 1643 endoffset = round_page(blkroundup(fs, origendoffset)); 1644 } 1645 1646 KASSERT(startoffset > 0 || endoffset >= startoffset); 1647 if (startoffset == endoffset) { 1648 /* Nothing to do, why were we called? */ 1649 simple_unlock(&vp->v_interlock); 1650 #ifdef DEBUG 1651 printf("lfs_putpages: startoffset = endoffset = %" PRId64 "\n", 1652 startoffset); 1653 #endif 1654 return 0; 1655 } 1656 1657 ap->a_offlo = startoffset; 1658 ap->a_offhi = endoffset; 1659 1660 if (!(ap->a_flags & PGO_CLEANIT)) 1661 return genfs_putpages(v); 1662 1663 /* 1664 * If there are more than one page per block, we don't want 1665 * to get caught locking them backwards; so set PGO_BUSYFAIL 1666 * to avoid deadlocks. 1667 */ 1668 ap->a_flags |= PGO_BUSYFAIL; 1669 1670 do { 1671 int r; 1672 1673 /* If no pages are dirty, we can just use genfs_putpages. */ 1674 if (check_dirty(fs, vp, startoffset, endoffset, blkeof, 1675 ap->a_flags, 1) != 0) 1676 break; 1677 1678 if ((r = genfs_putpages(v)) != EDEADLK) 1679 return r; 1680 1681 /* Start over. */ 1682 preempt(1); 1683 simple_lock(&vp->v_interlock); 1684 } while(1); 1685 1686 /* 1687 * Dirty and asked to clean. 1688 * 1689 * Pagedaemon can't actually write LFS pages; wake up 1690 * the writer to take care of that. The writer will 1691 * notice the pager inode queue and act on that. 1692 */ 1693 if (pagedaemon) { 1694 ++fs->lfs_pdflush; 1695 wakeup(&lfs_writer_daemon); 1696 simple_unlock(&vp->v_interlock); 1697 return EWOULDBLOCK; 1698 } 1699 1700 /* 1701 * If this is a file created in a recent dirop, we can't flush its 1702 * inode until the dirop is complete. Drain dirops, then flush the 1703 * filesystem (taking care of any other pending dirops while we're 1704 * at it). 1705 */ 1706 if ((ap->a_flags & (PGO_CLEANIT|PGO_LOCKED)) == PGO_CLEANIT && 1707 (vp->v_flag & VDIROP)) { 1708 int locked; 1709 1710 /* printf("putpages to clean VDIROP, flushing\n"); */ 1711 lfs_writer_enter(fs, "ppdirop"); 1712 locked = VOP_ISLOCKED(vp) && /* XXX */ 1713 vp->v_lock.lk_lockholder == curproc->p_pid; 1714 if (locked) 1715 VOP_UNLOCK(vp, 0); 1716 simple_unlock(&vp->v_interlock); 1717 1718 lfs_flush_fs(fs, sync ? SEGM_SYNC : 0); 1719 1720 simple_lock(&vp->v_interlock); 1721 if (locked) 1722 VOP_LOCK(vp, LK_EXCLUSIVE); 1723 lfs_writer_leave(fs); 1724 1725 /* XXX the flush should have taken care of this one too! */ 1726 } 1727 1728 /* 1729 * This is it. We are going to write some pages. From here on 1730 * down it's all just mechanics. 1731 * 1732 * Don't let genfs_putpages wait; lfs_segunlock will wait for us. 1733 */ 1734 ap->a_flags &= ~PGO_SYNCIO; 1735 1736 /* 1737 * If we've already got the seglock, flush the node and return. 1738 * The FIP has already been set up for us by lfs_writefile, 1739 * and FIP cleanup and lfs_updatemeta will also be done there, 1740 * unless genfs_putpages returns EDEADLK; then we must flush 1741 * what we have, and correct FIP and segment header accounting. 1742 */ 1743 1744 seglocked = (ap->a_flags & PGO_LOCKED) != 0; 1745 if (!seglocked) { 1746 simple_unlock(&vp->v_interlock); 1747 /* 1748 * Take the seglock, because we are going to be writing pages. 1749 */ 1750 error = lfs_seglock(fs, SEGM_PROT | (sync ? SEGM_SYNC : 0)); 1751 if (error != 0) 1752 return error; 1753 simple_lock(&vp->v_interlock); 1754 } 1755 1756 /* 1757 * VOP_PUTPAGES should not be called while holding the seglock. 1758 * XXXUBC fix lfs_markv, or do this properly. 1759 */ 1760 /* KASSERT(fs->lfs_seglock == 1); */ 1761 1762 /* 1763 * We assume we're being called with sp->fip pointing at blank space. 1764 * Account for a new FIP in the segment header, and set sp->vp. 1765 * (This should duplicate the setup at the top of lfs_writefile().) 1766 */ 1767 sp = fs->lfs_sp; 1768 if (!seglocked) { 1769 if (sp->seg_bytes_left < fs->lfs_bsize || 1770 sp->sum_bytes_left < sizeof(struct finfo)) 1771 (void) lfs_writeseg(fs, fs->lfs_sp); 1772 1773 sp->sum_bytes_left -= FINFOSIZE; 1774 ++((SEGSUM *)(sp->segsum))->ss_nfinfo; 1775 } 1776 KASSERT(sp->vp == NULL); 1777 sp->vp = vp; 1778 1779 if (!seglocked) { 1780 if (vp->v_flag & VDIROP) 1781 ((SEGSUM *)(sp->segsum))->ss_flags |= (SS_DIROP|SS_CONT); 1782 } 1783 1784 sp->fip->fi_nblocks = 0; 1785 sp->fip->fi_ino = ip->i_number; 1786 sp->fip->fi_version = ip->i_gen; 1787 1788 /* 1789 * Loop through genfs_putpages until all pages are gathered. 1790 * genfs_putpages() drops the interlock, so reacquire it if necessary. 1791 * Whenever we lose the interlock we have to rerun check_dirty, as 1792 * well. 1793 */ 1794 again: 1795 check_dirty(fs, vp, startoffset, endoffset, blkeof, ap->a_flags, 0); 1796 1797 if ((error = genfs_putpages(v)) == EDEADLK) { 1798 #ifdef DEBUG_LFS 1799 printf("lfs_putpages: genfs_putpages returned EDEADLK [2]" 1800 " ino %d off %x (seg %d)\n", 1801 ip->i_number, fs->lfs_offset, 1802 dtosn(fs, fs->lfs_offset)); 1803 #endif 1804 /* If nothing to write, short-circuit */ 1805 if (sp->cbpp - sp->bpp > 1) { 1806 /* Write gathered pages */ 1807 lfs_updatemeta(sp); 1808 (void) lfs_writeseg(fs, sp); 1809 1810 /* 1811 * Reinitialize brand new FIP and add us to it. 1812 * (This should duplicate the fixup in 1813 * lfs_gatherpages().) 1814 */ 1815 KASSERT(sp->vp == vp); 1816 sp->fip->fi_version = ip->i_gen; 1817 sp->fip->fi_ino = ip->i_number; 1818 /* Add us to the new segment summary. */ 1819 ++((SEGSUM *)(sp->segsum))->ss_nfinfo; 1820 sp->sum_bytes_left -= FINFOSIZE; 1821 } 1822 1823 /* Give the write a chance to complete */ 1824 preempt(1); 1825 1826 /* We've lost the interlock. Start over. */ 1827 simple_lock(&vp->v_interlock); 1828 goto again; 1829 } 1830 1831 KASSERT(sp->vp == vp); 1832 if (!seglocked) { 1833 sp->vp = NULL; /* XXX lfs_gather below will set this */ 1834 1835 /* Write indirect blocks as well */ 1836 lfs_gather(fs, fs->lfs_sp, vp, lfs_match_indir); 1837 lfs_gather(fs, fs->lfs_sp, vp, lfs_match_dindir); 1838 lfs_gather(fs, fs->lfs_sp, vp, lfs_match_tindir); 1839 1840 KASSERT(sp->vp == NULL); 1841 sp->vp = vp; 1842 } 1843 1844 /* 1845 * Blocks are now gathered into a segment waiting to be written. 1846 * All that's left to do is update metadata, and write them. 1847 */ 1848 lfs_updatemeta(sp); 1849 KASSERT(sp->vp == vp); 1850 sp->vp = NULL; 1851 1852 if (seglocked) { 1853 /* we're called by lfs_writefile. */ 1854 return error; 1855 } 1856 1857 /* 1858 * Clean up FIP, since we're done writing this file. 1859 * This should duplicate cleanup at the end of lfs_writefile(). 1860 */ 1861 if (sp->fip->fi_nblocks != 0) { 1862 sp->fip = (FINFO*)((caddr_t)sp->fip + FINFOSIZE + 1863 sizeof(int32_t) * sp->fip->fi_nblocks); 1864 sp->start_lbp = &sp->fip->fi_blocks[0]; 1865 } else { 1866 sp->sum_bytes_left += FINFOSIZE; 1867 --((SEGSUM *)(sp->segsum))->ss_nfinfo; 1868 } 1869 lfs_writeseg(fs, fs->lfs_sp); 1870 1871 /* 1872 * XXX - with the malloc/copy writeseg, the pages are freed by now 1873 * even if we don't wait (e.g. if we hold a nested lock). This 1874 * will not be true if we stop using malloc/copy. 1875 */ 1876 KASSERT(fs->lfs_sp->seg_flags & SEGM_PROT); 1877 lfs_segunlock(fs); 1878 1879 /* 1880 * Wait for v_numoutput to drop to zero. The seglock should 1881 * take care of this, but there is a slight possibility that 1882 * aiodoned might not have got around to our buffers yet. 1883 */ 1884 if (sync) { 1885 int s; 1886 1887 s = splbio(); 1888 simple_lock(&global_v_numoutput_slock); 1889 while (vp->v_numoutput > 0) { 1890 #ifdef DEBUG 1891 printf("ino %d sleeping on num %d\n", 1892 ip->i_number, vp->v_numoutput); 1893 #endif 1894 vp->v_flag |= VBWAIT; 1895 ltsleep(&vp->v_numoutput, PRIBIO + 1, "lfs_vn", 0, 1896 &global_v_numoutput_slock); 1897 } 1898 simple_unlock(&global_v_numoutput_slock); 1899 splx(s); 1900 } 1901 return error; 1902 } 1903 1904 /* 1905 * Return the last logical file offset that should be written for this file 1906 * if we're doing a write that ends at "size". If writing, we need to know 1907 * about sizes on disk, i.e. fragments if there are any; if reading, we need 1908 * to know about entire blocks. 1909 */ 1910 void 1911 lfs_gop_size(struct vnode *vp, off_t size, off_t *eobp, int flags) 1912 { 1913 struct inode *ip = VTOI(vp); 1914 struct lfs *fs = ip->i_lfs; 1915 daddr_t olbn, nlbn; 1916 1917 KASSERT(flags & (GOP_SIZE_READ | GOP_SIZE_WRITE)); 1918 KASSERT((flags & (GOP_SIZE_READ | GOP_SIZE_WRITE)) 1919 != (GOP_SIZE_READ | GOP_SIZE_WRITE)); 1920 1921 olbn = lblkno(fs, ip->i_size); 1922 nlbn = lblkno(fs, size); 1923 if (!(flags & GOP_SIZE_MEM) && nlbn < NDADDR && olbn <= nlbn) { 1924 *eobp = fragroundup(fs, size); 1925 } else { 1926 *eobp = blkroundup(fs, size); 1927 } 1928 } 1929 1930 #ifdef DEBUG 1931 void lfs_dump_vop(void *); 1932 1933 void 1934 lfs_dump_vop(void *v) 1935 { 1936 struct vop_putpages_args /* { 1937 struct vnode *a_vp; 1938 voff_t a_offlo; 1939 voff_t a_offhi; 1940 int a_flags; 1941 } */ *ap = v; 1942 1943 #ifdef DDB 1944 vfs_vnode_print(ap->a_vp, 0, printf); 1945 #endif 1946 lfs_dump_dinode(VTOI(ap->a_vp)->i_din.ffs1_din); 1947 } 1948 #endif 1949 1950 int 1951 lfs_mmap(void *v) 1952 { 1953 struct vop_mmap_args /* { 1954 const struct vnodeop_desc *a_desc; 1955 struct vnode *a_vp; 1956 int a_fflags; 1957 struct ucred *a_cred; 1958 struct lwp *a_l; 1959 } */ *ap = v; 1960 1961 if (VTOI(ap->a_vp)->i_number == LFS_IFILE_INUM) 1962 return EOPNOTSUPP; 1963 return ufs_mmap(v); 1964 } 1965
Indexes created Sun Oct 19 02:09:48 GMT 2025