vfs_init.c revision 1.48.18.1
1 /* $NetBSD: vfs_init.c,v 1.48.18.1 2019/06/10 22:09:04 christos Exp $ */ 2 3 /*- 4 * Copyright (c) 1998, 2000, 2008 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 21 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 22 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 23 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 24 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33 /* 34 * Copyright (c) 1989, 1993 35 * The Regents of the University of California. All rights reserved. 36 * 37 * This code is derived from software contributed 38 * to Berkeley by John Heidemann of the UCLA Ficus project. 39 * 40 * Source: * @(#)i405_init.c 2.10 92/04/27 UCLA Ficus project 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 * @(#)vfs_init.c 8.5 (Berkeley) 5/11/95 67 */ 68 69 #include <sys/cdefs.h> 70 __KERNEL_RCSID(0, "$NetBSD: vfs_init.c,v 1.48.18.1 2019/06/10 22:09:04 christos Exp $"); 71 72 #include <sys/param.h> 73 #include <sys/mount.h> 74 #include <sys/time.h> 75 #include <sys/vnode.h> 76 #include <sys/stat.h> 77 #include <sys/namei.h> 78 #include <sys/ucred.h> 79 #include <sys/buf.h> 80 #include <sys/errno.h> 81 #include <sys/kmem.h> 82 #include <sys/systm.h> 83 #include <sys/module.h> 84 #include <sys/dirhash.h> 85 #include <sys/sysctl.h> 86 #include <sys/kauth.h> 87 88 /* 89 * Sigh, such primitive tools are these... 90 */ 91 #if 0 92 #define DODEBUG(A) A 93 #else 94 #define DODEBUG(A) 95 #endif 96 97 pool_cache_t pnbuf_cache; 98 99 /* 100 * The global list of vnode operations. 101 */ 102 extern const struct vnodeop_desc * const vfs_op_descs[]; 103 104 /* 105 * These vnodeopv_descs are listed here because they are not 106 * associated with any particular file system, and thus cannot 107 * be initialized by vfs_attach(). 108 */ 109 extern const struct vnodeopv_desc dead_vnodeop_opv_desc; 110 extern const struct vnodeopv_desc fifo_vnodeop_opv_desc; 111 extern const struct vnodeopv_desc spec_vnodeop_opv_desc; 112 113 const struct vnodeopv_desc * const vfs_special_vnodeopv_descs[] = { 114 &dead_vnodeop_opv_desc, 115 &fifo_vnodeop_opv_desc, 116 &spec_vnodeop_opv_desc, 117 NULL, 118 }; 119 120 struct vfs_list_head vfs_list = /* vfs list */ 121 LIST_HEAD_INITIALIZER(vfs_list); 122 123 static kauth_listener_t mount_listener; 124 125 /* 126 * This code doesn't work if the defn is **vnodop_defns with cc. 127 * The problem is because of the compiler sometimes putting in an 128 * extra level of indirection for arrays. It's an interesting 129 * "feature" of C. 130 */ 131 typedef int (*PFI)(void *); 132 133 /* 134 * A miscellaneous routine. 135 * A generic "default" routine that just returns an error. 136 */ 137 /*ARGSUSED*/ 138 int 139 vn_default_error(void *v) 140 { 141 142 return (EOPNOTSUPP); 143 } 144 145 static struct sysctllog *vfs_sysctllog; 146 147 /* 148 * Top level filesystem related information gathering. 149 */ 150 static void 151 sysctl_vfs_setup(void) 152 { 153 extern int vfs_magiclinks; 154 155 sysctl_createv(&vfs_sysctllog, 0, NULL, NULL, 156 CTLFLAG_PERMANENT, 157 CTLTYPE_NODE, "generic", 158 SYSCTL_DESCR("Non-specific vfs related information"), 159 NULL, 0, NULL, 0, 160 CTL_VFS, VFS_GENERIC, CTL_EOL); 161 sysctl_createv(&vfs_sysctllog, 0, NULL, NULL, 162 CTLFLAG_PERMANENT, 163 CTLTYPE_STRING, "fstypes", 164 SYSCTL_DESCR("List of file systems present"), 165 sysctl_vfs_generic_fstypes, 0, NULL, 0, 166 CTL_VFS, VFS_GENERIC, CTL_CREATE, CTL_EOL); 167 sysctl_createv(&vfs_sysctllog, 0, NULL, NULL, 168 CTLFLAG_PERMANENT|CTLFLAG_READWRITE, 169 CTLTYPE_INT, "magiclinks", 170 SYSCTL_DESCR("Whether \"magic\" symlinks are expanded"), 171 NULL, 0, &vfs_magiclinks, 0, 172 CTL_VFS, VFS_GENERIC, VFS_MAGICLINKS, CTL_EOL); 173 } 174 175 176 /* 177 * vfs_init.c 178 * 179 * Allocate and fill in operations vectors. 180 * 181 * An undocumented feature of this approach to defining operations is that 182 * there can be multiple entries in vfs_opv_descs for the same operations 183 * vector. This allows third parties to extend the set of operations 184 * supported by another layer in a binary compatibile way. For example, 185 * assume that NFS needed to be modified to support Ficus. NFS has an entry 186 * (probably nfs_vnopdeop_decls) declaring all the operations NFS supports by 187 * default. Ficus could add another entry (ficus_nfs_vnodeop_decl_entensions) 188 * listing those new operations Ficus adds to NFS, all without modifying the 189 * NFS code. (Of couse, the OTW NFS protocol still needs to be munged, but 190 * that is a(whole)nother story.) This is a feature. 191 */ 192 193 /* 194 * Init the vector, if it needs it. 195 * Also handle backwards compatibility. 196 */ 197 static void 198 vfs_opv_init_explicit(const struct vnodeopv_desc *vfs_opv_desc) 199 { 200 int (**opv_desc_vector)(void *); 201 const struct vnodeopv_entry_desc *opve_descp; 202 203 opv_desc_vector = *(vfs_opv_desc->opv_desc_vector_p); 204 205 for (opve_descp = vfs_opv_desc->opv_desc_ops; 206 opve_descp->opve_op; 207 opve_descp++) { 208 /* 209 * Sanity check: is this operation listed 210 * in the list of operations? We check this 211 * by seeing if its offset is zero. Since 212 * the default routine should always be listed 213 * first, it should be the only one with a zero 214 * offset. Any other operation with a zero 215 * offset is probably not listed in 216 * vfs_op_descs, and so is probably an error. 217 * 218 * A panic here means the layer programmer 219 * has committed the all-too common bug 220 * of adding a new operation to the layer's 221 * list of vnode operations but 222 * not adding the operation to the system-wide 223 * list of supported operations. 224 */ 225 if (opve_descp->opve_op->vdesc_offset == 0 && 226 opve_descp->opve_op->vdesc_offset != VOFFSET(vop_default)) { 227 printf("operation %s not listed in %s.\n", 228 opve_descp->opve_op->vdesc_name, "vfs_op_descs"); 229 panic ("vfs_opv_init: bad operation"); 230 } 231 232 /* 233 * Fill in this entry. 234 */ 235 opv_desc_vector[opve_descp->opve_op->vdesc_offset] = 236 opve_descp->opve_impl; 237 } 238 } 239 240 static void 241 vfs_opv_init_default(const struct vnodeopv_desc *vfs_opv_desc) 242 { 243 int j; 244 int (**opv_desc_vector)(void *); 245 246 opv_desc_vector = *(vfs_opv_desc->opv_desc_vector_p); 247 248 /* 249 * Force every operations vector to have a default routine. 250 */ 251 if (opv_desc_vector[VOFFSET(vop_default)] == NULL) 252 panic("vfs_opv_init: operation vector without default routine."); 253 254 for (j = 0; j < VNODE_OPS_COUNT; j++) 255 if (opv_desc_vector[j] == NULL) 256 opv_desc_vector[j] = 257 opv_desc_vector[VOFFSET(vop_default)]; 258 } 259 260 void 261 vfs_opv_init(const struct vnodeopv_desc * const *vopvdpp) 262 { 263 int (**opv_desc_vector)(void *); 264 int i; 265 266 /* 267 * Allocate the vectors. 268 */ 269 for (i = 0; vopvdpp[i] != NULL; i++) { 270 opv_desc_vector = 271 kmem_alloc(VNODE_OPS_COUNT * sizeof(PFI), KM_SLEEP); 272 memset(opv_desc_vector, 0, VNODE_OPS_COUNT * sizeof(PFI)); 273 *(vopvdpp[i]->opv_desc_vector_p) = opv_desc_vector; 274 DODEBUG(printf("vector at %p allocated\n", 275 opv_desc_vector_p)); 276 } 277 278 /* 279 * ...and fill them in. 280 */ 281 for (i = 0; vopvdpp[i] != NULL; i++) 282 vfs_opv_init_explicit(vopvdpp[i]); 283 284 /* 285 * Finally, go back and replace unfilled routines 286 * with their default. 287 */ 288 for (i = 0; vopvdpp[i] != NULL; i++) 289 vfs_opv_init_default(vopvdpp[i]); 290 } 291 292 void 293 vfs_opv_free(const struct vnodeopv_desc * const *vopvdpp) 294 { 295 int i; 296 297 /* 298 * Free the vectors allocated in vfs_opv_init(). 299 */ 300 for (i = 0; vopvdpp[i] != NULL; i++) { 301 kmem_free(*(vopvdpp[i]->opv_desc_vector_p), 302 VNODE_OPS_COUNT * sizeof(PFI)); 303 *(vopvdpp[i]->opv_desc_vector_p) = NULL; 304 } 305 } 306 307 #ifdef DEBUG 308 static void 309 vfs_op_check(void) 310 { 311 int i; 312 313 DODEBUG(printf("Vnode_interface_init.\n")); 314 315 /* 316 * Check offset of each op. 317 */ 318 for (i = 0; vfs_op_descs[i]; i++) { 319 if (vfs_op_descs[i]->vdesc_offset != i) 320 panic("vfs_op_check: vfs_op_desc[] offset mismatch"); 321 } 322 323 if (i != VNODE_OPS_COUNT) { 324 panic("vfs_op_check: vnode ops count mismatch (%d != %d)", 325 i, VNODE_OPS_COUNT); 326 } 327 328 DODEBUG(printf ("vfs_opv_numops=%d\n", VNODE_OPS_COUNT)); 329 } 330 #endif /* DEBUG */ 331 332 /* 333 * Common routine to check if an unprivileged mount is allowed. 334 * 335 * We export just this part (i.e., without the access control) so that if a 336 * secmodel wants to implement finer grained user mounts it can do so without 337 * copying too much code. More elaborate policies (i.e., specific users allowed 338 * to also create devices and/or introduce set-id binaries, or export 339 * file-systems) will require a different implementation. 340 * 341 * This routine is intended to be called from listener context, and as such 342 * does not take credentials as an argument. 343 */ 344 int 345 usermount_common_policy(struct mount *mp, u_long flags) 346 { 347 348 /* No exporting if unprivileged. */ 349 if (flags & MNT_EXPORTED) 350 return EPERM; 351 352 /* Must have 'nosuid' and 'nodev'. */ 353 if ((flags & MNT_NODEV) == 0 || (flags & MNT_NOSUID) == 0) 354 return EPERM; 355 356 /* Retain 'noexec'. */ 357 if ((mp->mnt_flag & MNT_NOEXEC) && (flags & MNT_NOEXEC) == 0) 358 return EPERM; 359 360 return 0; 361 } 362 363 static int 364 mount_listener_cb(kauth_cred_t cred, kauth_action_t action, void *cookie, 365 void *arg0, void *arg1, void *arg2, void *arg3) 366 { 367 int result; 368 enum kauth_system_req req; 369 370 result = KAUTH_RESULT_DEFER; 371 req = (enum kauth_system_req)arg0; 372 373 if (action != KAUTH_SYSTEM_MOUNT) 374 return result; 375 376 if (req == KAUTH_REQ_SYSTEM_MOUNT_GET) 377 result = KAUTH_RESULT_ALLOW; 378 else if (req == KAUTH_REQ_SYSTEM_MOUNT_DEVICE) { 379 vnode_t *devvp = arg2; 380 mode_t access_mode = (mode_t)(unsigned long)arg3; 381 int error; 382 383 error = VOP_ACCESS(devvp, access_mode, cred); 384 if (!error) 385 result = KAUTH_RESULT_ALLOW; 386 } 387 388 return result; 389 } 390 391 /* 392 * Initialize the vnode structures and initialize each file system type. 393 */ 394 void 395 vfsinit(void) 396 { 397 398 /* 399 * Attach sysctl nodes 400 */ 401 sysctl_vfs_setup(); 402 403 /* 404 * Initialize the namei pathname buffer pool and cache. 405 */ 406 pnbuf_cache = pool_cache_init(MAXPATHLEN, 0, 0, 0, "pnbufpl", 407 NULL, IPL_NONE, NULL, NULL, NULL); 408 KASSERT(pnbuf_cache != NULL); 409 410 /* 411 * Initialize the vnode table 412 */ 413 vntblinit(); 414 415 /* 416 * Initialize the vnode name cache 417 */ 418 nchinit(); 419 420 #ifdef DEBUG 421 /* 422 * Check the list of vnode operations. 423 */ 424 vfs_op_check(); 425 #endif 426 427 /* 428 * Initialize the special vnode operations. 429 */ 430 vfs_opv_init(vfs_special_vnodeopv_descs); 431 432 /* 433 * Initialise generic dirhash. 434 */ 435 dirhash_init(); 436 437 /* 438 * Initialise VFS hooks. 439 */ 440 vfs_hooks_init(); 441 442 mount_listener = kauth_listen_scope(KAUTH_SCOPE_SYSTEM, 443 mount_listener_cb, NULL); 444 445 /* 446 * Establish each file system which was statically 447 * included in the kernel. 448 */ 449 module_init_class(MODULE_CLASS_VFS); 450 } 451 452 /* 453 * Drop a reference to a file system type. 454 */ 455 void 456 vfs_delref(struct vfsops *vfs) 457 { 458 459 mutex_enter(&vfs_list_lock); 460 vfs->vfs_refcount--; 461 mutex_exit(&vfs_list_lock); 462 } 463 464 /* 465 * Establish a file system and initialize it. 466 */ 467 int 468 vfs_attach(struct vfsops *vfs) 469 { 470 struct vfsops *v; 471 int error = 0; 472 473 mutex_enter(&vfs_list_lock); 474 475 /* 476 * Make sure this file system doesn't already exist. 477 */ 478 LIST_FOREACH(v, &vfs_list, vfs_list) { 479 if (strcmp(vfs->vfs_name, v->vfs_name) == 0) { 480 error = EEXIST; 481 goto out; 482 } 483 } 484 485 /* 486 * Initialize the vnode operations for this file system. 487 */ 488 vfs_opv_init(vfs->vfs_opv_descs); 489 490 /* 491 * Now initialize the file system itself. 492 */ 493 (*vfs->vfs_init)(); 494 495 /* 496 * ...and link it into the kernel's list. 497 */ 498 LIST_INSERT_HEAD(&vfs_list, vfs, vfs_list); 499 500 /* 501 * Sanity: make sure the reference count is 0. 502 */ 503 vfs->vfs_refcount = 0; 504 out: 505 mutex_exit(&vfs_list_lock); 506 return (error); 507 } 508 509 /* 510 * Remove a file system from the kernel. 511 */ 512 int 513 vfs_detach(struct vfsops *vfs) 514 { 515 struct vfsops *v; 516 int error = 0; 517 518 mutex_enter(&vfs_list_lock); 519 520 /* 521 * Make sure no one is using the filesystem. 522 */ 523 if (vfs->vfs_refcount != 0) { 524 error = EBUSY; 525 goto out; 526 } 527 528 /* 529 * ...and remove it from the kernel's list. 530 */ 531 LIST_FOREACH(v, &vfs_list, vfs_list) { 532 if (v == vfs) { 533 LIST_REMOVE(v, vfs_list); 534 break; 535 } 536 } 537 538 if (v == NULL) { 539 error = ESRCH; 540 goto out; 541 } 542 543 /* 544 * Now run the file system-specific cleanups. 545 */ 546 (*vfs->vfs_done)(); 547 548 /* 549 * Free the vnode operations vector. 550 */ 551 vfs_opv_free(vfs->vfs_opv_descs); 552 out: 553 mutex_exit(&vfs_list_lock); 554 return (error); 555 } 556 557 void 558 vfs_reinit(void) 559 { 560 struct vfsops *vfs; 561 562 mutex_enter(&vfs_list_lock); 563 LIST_FOREACH(vfs, &vfs_list, vfs_list) { 564 if (vfs->vfs_reinit) { 565 vfs->vfs_refcount++; 566 mutex_exit(&vfs_list_lock); 567 (*vfs->vfs_reinit)(); 568 mutex_enter(&vfs_list_lock); 569 vfs->vfs_refcount--; 570 } 571 } 572 mutex_exit(&vfs_list_lock); 573 } 574
Indexes created Sat Sep 20 22:09:52 GMT 2025