]> git.saurik.com Git - apple/xnu.git/blob - bsd/vfs/vfs_init.c
441d9269fea3d7344c43c97c50ca19f791aac10b
[apple/xnu.git] / bsd / vfs / vfs_init.c
1 /*
2 * Copyright (c) 2000-2010 Apple Inc. All rights reserved.
3 *
4 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
5 *
6 * This file contains Original Code and/or Modifications of Original Code
7 * as defined in and that are subject to the Apple Public Source License
8 * Version 2.0 (the 'License'). You may not use this file except in
9 * compliance with the License. The rights granted to you under the License
10 * may not be used to create, or enable the creation or redistribution of,
11 * unlawful or unlicensed copies of an Apple operating system, or to
12 * circumvent, violate, or enable the circumvention or violation of, any
13 * terms of an Apple operating system software license agreement.
14 *
15 * Please obtain a copy of the License at
16 * http://www.opensource.apple.com/apsl/ and read it before using this file.
17 *
18 * The Original Code and all software distributed under the License are
19 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
20 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
21 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
22 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
23 * Please see the License for the specific language governing rights and
24 * limitations under the License.
25 *
26 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
27 */
28 /* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */
29 /*
30 * Copyright (c) 1989, 1993
31 * The Regents of the University of California. All rights reserved.
32 *
33 * This code is derived from software contributed
34 * to Berkeley by John Heidemann of the UCLA Ficus project.
35 *
36 * Source: * @(#)i405_init.c 2.10 92/04/27 UCLA Ficus project
37 *
38 * Redistribution and use in source and binary forms, with or without
39 * modification, are permitted provided that the following conditions
40 * are met:
41 * 1. Redistributions of source code must retain the above copyright
42 * notice, this list of conditions and the following disclaimer.
43 * 2. Redistributions in binary form must reproduce the above copyright
44 * notice, this list of conditions and the following disclaimer in the
45 * documentation and/or other materials provided with the distribution.
46 * 3. All advertising materials mentioning features or use of this software
47 * must display the following acknowledgement:
48 * This product includes software developed by the University of
49 * California, Berkeley and its contributors.
50 * 4. 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 * NOTICE: This file was modified by SPARTA, Inc. in 2005 to introduce
70 * support for mandatory and extensible security protections. This notice
71 * is included in support of clause 2.2 (b) of the Apple Public License,
72 * Version 2.0.
73 */
74
75
76 #include <sys/param.h>
77 #include <sys/mount_internal.h>
78 #include <sys/time.h>
79 #include <sys/vm.h>
80 #include <sys/vnode_internal.h>
81 #include <sys/stat.h>
82 #include <sys/namei.h>
83 #include <sys/ucred.h>
84 #include <sys/errno.h>
85 #include <kern/kalloc.h>
86 #include <sys/decmpfs.h>
87
88 #if CONFIG_MACF
89 #include <security/mac_framework.h>
90 #include <sys/kauth.h>
91 #endif
92 #if QUOTA
93 #include <sys/quota.h>
94 #endif
95
96 /*
97 * Sigh, such primitive tools are these...
98 */
99 #if 0
100 #define DODEBUG(A) A
101 #else
102 #define DODEBUG(A)
103 #endif
104
105 ZONE_DECLARE(mount_zone, "mount", sizeof(struct mount), ZC_ZFREE_CLEARMEM);
106
107 __private_extern__ void vntblinit(void);
108
109 extern const struct vnodeopv_desc *vfs_opv_descs[];
110 /* a list of lists of vnodeops defns */
111 extern struct vnodeop_desc *vfs_op_descs[];
112 /* and the operations they perform */
113 /*
114 * This code doesn't work if the defn is **vnodop_defns with cc.
115 * The problem is because of the compiler sometimes putting in an
116 * extra level of indirection for arrays. It's an interesting
117 * "feature" of C.
118 */
119 int vfs_opv_numops;
120
121 typedef int (*PFIvp)(void *);
122
123 /*
124 * A miscellaneous routine.
125 * A generic "default" routine that just returns an error.
126 */
127 int
128 vn_default_error(void)
129 {
130 return ENOTSUP;
131 }
132
133 /*
134 * vfs_init.c
135 *
136 * Allocate and fill in operations vectors.
137 *
138 * An undocumented feature of this approach to defining operations is that
139 * there can be multiple entries in vfs_opv_descs for the same operations
140 * vector. This allows third parties to extend the set of operations
141 * supported by another layer in a binary compatibile way. For example,
142 * assume that NFS needed to be modified to support Ficus. NFS has an entry
143 * (probably nfs_vnopdeop_decls) declaring all the operations NFS supports by
144 * default. Ficus could add another entry (ficus_nfs_vnodeop_decl_entensions)
145 * listing those new operations Ficus adds to NFS, all without modifying the
146 * NFS code. (Of couse, the OTW NFS protocol still needs to be munged, but
147 * that is a(whole)nother story.) This is a feature.
148 */
149 void
150 vfs_opv_init(void)
151 {
152 int i, j, k;
153 int(***opv_desc_vector_p)(void *);
154 int(**opv_desc_vector)(void *);
155 const struct vnodeopv_entry_desc *opve_descp;
156
157 /*
158 * Allocate the dynamic vectors and fill them in.
159 */
160 for (i = 0; vfs_opv_descs[i]; i++) {
161 opv_desc_vector_p = vfs_opv_descs[i]->opv_desc_vector_p;
162 /*
163 * Allocate and init the vector, if it needs it.
164 * Also handle backwards compatibility.
165 */
166 if (*opv_desc_vector_p == NULL) {
167 *opv_desc_vector_p = kheap_alloc(KHEAP_DEFAULT,
168 vfs_opv_numops * sizeof(PFIvp), Z_WAITOK | Z_ZERO);
169 DODEBUG(printf("vector at %x allocated\n",
170 opv_desc_vector_p));
171 }
172 opv_desc_vector = *opv_desc_vector_p;
173 for (j = 0; vfs_opv_descs[i]->opv_desc_ops[j].opve_op; j++) {
174 opve_descp = &(vfs_opv_descs[i]->opv_desc_ops[j]);
175
176 /* Silently skip known-disabled operations */
177 if (opve_descp->opve_op->vdesc_flags & VDESC_DISABLED) {
178 printf("vfs_fsadd: Ignoring reference in %p to disabled operation %s.\n",
179 vfs_opv_descs[i], opve_descp->opve_op->vdesc_name);
180 continue;
181 }
182
183 /*
184 * Sanity check: is this operation listed
185 * in the list of operations? We check this
186 * by seeing if its offest is zero. Since
187 * the default routine should always be listed
188 * first, it should be the only one with a zero
189 * offset. Any other operation with a zero
190 * offset is probably not listed in
191 * vfs_op_descs, and so is probably an error.
192 *
193 * A panic here means the layer programmer
194 * has committed the all-too common bug
195 * of adding a new operation to the layer's
196 * list of vnode operations but
197 * not adding the operation to the system-wide
198 * list of supported operations.
199 */
200 if (opve_descp->opve_op->vdesc_offset == 0 &&
201 opve_descp->opve_op !=
202 VDESC(vnop_default)) {
203 printf("operation %s not listed in %s.\n",
204 opve_descp->opve_op->vdesc_name,
205 "vfs_op_descs");
206 panic("vfs_opv_init: bad operation");
207 }
208 /*
209 * Fill in this entry.
210 */
211 opv_desc_vector[opve_descp->opve_op->vdesc_offset] =
212 opve_descp->opve_impl;
213 }
214 }
215 /*
216 * Finally, go back and replace unfilled routines
217 * with their default. (Sigh, an O(n^3) algorithm. I
218 * could make it better, but that'd be work, and n is small.)
219 */
220 for (i = 0; vfs_opv_descs[i]; i++) {
221 opv_desc_vector = *(vfs_opv_descs[i]->opv_desc_vector_p);
222 /*
223 * Force every operations vector to have a default routine.
224 */
225 if (opv_desc_vector[VOFFSET(vnop_default)] == NULL) {
226 panic("vfs_opv_init: operation vector without default routine.");
227 }
228 for (k = 0; k < vfs_opv_numops; k++) {
229 if (opv_desc_vector[k] == NULL) {
230 opv_desc_vector[k] =
231 opv_desc_vector[VOFFSET(vnop_default)];
232 }
233 }
234 }
235 }
236
237 /*
238 * Initialize known vnode operations vectors.
239 */
240 void
241 vfs_op_init(void)
242 {
243 int i;
244
245 DODEBUG(printf("Vnode_interface_init.\n"));
246 /*
247 * Set all vnode vectors to a well known value.
248 */
249 for (i = 0; vfs_opv_descs[i]; i++) {
250 *(vfs_opv_descs[i]->opv_desc_vector_p) = NULL;
251 }
252 /*
253 * Figure out how many ops there are by counting the table,
254 * and assign each its offset.
255 */
256 for (vfs_opv_numops = 0, i = 0; vfs_op_descs[i]; i++) {
257 /* Silently skip known-disabled operations */
258 if (vfs_op_descs[i]->vdesc_flags & VDESC_DISABLED) {
259 continue;
260 }
261 vfs_op_descs[i]->vdesc_offset = vfs_opv_numops;
262 vfs_opv_numops++;
263 }
264 DODEBUG(printf("vfs_opv_numops=%d\n", vfs_opv_numops));
265 }
266
267 /*
268 * Routines having to do with the management of the vnode table.
269 */
270 extern struct vnodeops dead_vnodeops;
271 extern struct vnodeops spec_vnodeops;
272
273 /* vars for vnode lock */
274 lck_grp_t * vnode_lck_grp;
275 lck_grp_attr_t * vnode_lck_grp_attr;
276 lck_attr_t * vnode_lck_attr;
277
278 #if CONFIG_TRIGGERS
279 /* vars for vnode trigger resolver */
280 lck_grp_t * trigger_vnode_lck_grp;
281 lck_grp_attr_t * trigger_vnode_lck_grp_attr;
282 lck_attr_t * trigger_vnode_lck_attr;
283 #endif
284
285 lck_grp_t * fd_vn_lck_grp;
286 lck_grp_attr_t * fd_vn_lck_grp_attr;
287 lck_attr_t * fd_vn_lck_attr;
288
289 /* vars for vnode list lock */
290 lck_grp_t * vnode_list_lck_grp;
291 lck_grp_attr_t * vnode_list_lck_grp_attr;
292 lck_attr_t * vnode_list_lck_attr;
293 lck_spin_t * vnode_list_spin_lock;
294 lck_mtx_t * spechash_mtx_lock;
295
296 /* vars for vfsconf lock */
297 lck_grp_t * fsconf_lck_grp;
298 lck_grp_attr_t * fsconf_lck_grp_attr;
299 lck_attr_t * fsconf_lck_attr;
300
301
302 /* vars for mount lock */
303 lck_grp_t * mnt_lck_grp;
304 lck_grp_attr_t * mnt_lck_grp_attr;
305 lck_attr_t * mnt_lck_attr;
306
307 /* vars for mount list lock */
308 lck_grp_t * mnt_list_lck_grp;
309 lck_grp_attr_t * mnt_list_lck_grp_attr;
310 lck_attr_t * mnt_list_lck_attr;
311 lck_mtx_t * mnt_list_mtx_lock;
312
313 /* vars for sync mutex */
314 lck_grp_t * sync_mtx_lck_grp;
315 lck_grp_attr_t * sync_mtx_lck_grp_attr;
316 lck_attr_t * sync_mtx_lck_attr;
317 lck_mtx_t * sync_mtx_lck;
318
319 lck_mtx_t *pkg_extensions_lck;
320
321 struct mount * dead_mountp;
322
323 /*
324 * Initialize the vnode structures and initialize each file system type.
325 */
326 void
327 vfsinit(void)
328 {
329 struct vfstable *vfsp;
330 int i, maxtypenum;
331 struct mount * mp;
332
333 /* Allocate vnode list lock group attribute and group */
334 vnode_list_lck_grp_attr = lck_grp_attr_alloc_init();
335
336 vnode_list_lck_grp = lck_grp_alloc_init("vnode list", vnode_list_lck_grp_attr);
337
338 /* Allocate vnode list lock attribute */
339 vnode_list_lck_attr = lck_attr_alloc_init();
340
341 /* Allocate vnode list lock */
342 vnode_list_spin_lock = lck_spin_alloc_init(vnode_list_lck_grp, vnode_list_lck_attr);
343
344 /* Allocate spec hash list lock */
345 spechash_mtx_lock = lck_mtx_alloc_init(vnode_list_lck_grp, vnode_list_lck_attr);
346
347 /* Allocate the package extensions table lock */
348 pkg_extensions_lck = lck_mtx_alloc_init(vnode_list_lck_grp, vnode_list_lck_attr);
349
350 /* allocate vnode lock group attribute and group */
351 vnode_lck_grp_attr = lck_grp_attr_alloc_init();
352
353 vnode_lck_grp = lck_grp_alloc_init("vnode", vnode_lck_grp_attr);
354
355 /* Allocate vnode lock attribute */
356 vnode_lck_attr = lck_attr_alloc_init();
357
358 #if CONFIG_TRIGGERS
359 trigger_vnode_lck_grp_attr = lck_grp_attr_alloc_init();
360 trigger_vnode_lck_grp = lck_grp_alloc_init("trigger_vnode", trigger_vnode_lck_grp_attr);
361 trigger_vnode_lck_attr = lck_attr_alloc_init();
362 #endif
363 /* Allocate per fd vnode data lock attribute and group */
364 fd_vn_lck_grp_attr = lck_grp_attr_alloc_init();
365 fd_vn_lck_grp = lck_grp_alloc_init("fd_vnode_data", fd_vn_lck_grp_attr);
366 fd_vn_lck_attr = lck_attr_alloc_init();
367
368 /* Allocate fs config lock group attribute and group */
369 fsconf_lck_grp_attr = lck_grp_attr_alloc_init();
370
371 fsconf_lck_grp = lck_grp_alloc_init("fs conf", fsconf_lck_grp_attr);
372
373 /* Allocate fs config lock attribute */
374 fsconf_lck_attr = lck_attr_alloc_init();
375
376 /* Allocate mount point related lock structures */
377
378 /* Allocate mount list lock group attribute and group */
379 mnt_list_lck_grp_attr = lck_grp_attr_alloc_init();
380
381 mnt_list_lck_grp = lck_grp_alloc_init("mount list", mnt_list_lck_grp_attr);
382
383 /* Allocate mount list lock attribute */
384 mnt_list_lck_attr = lck_attr_alloc_init();
385
386 /* Allocate mount list lock */
387 mnt_list_mtx_lock = lck_mtx_alloc_init(mnt_list_lck_grp, mnt_list_lck_attr);
388
389
390 /* allocate mount lock group attribute and group */
391 mnt_lck_grp_attr = lck_grp_attr_alloc_init();
392
393 mnt_lck_grp = lck_grp_alloc_init("mount", mnt_lck_grp_attr);
394
395 /* Allocate mount lock attribute */
396 mnt_lck_attr = lck_attr_alloc_init();
397
398 /* Allocate sync lock */
399 sync_mtx_lck_grp_attr = lck_grp_attr_alloc_init();
400 sync_mtx_lck_grp = lck_grp_alloc_init("sync thread", sync_mtx_lck_grp_attr);
401 sync_mtx_lck_attr = lck_attr_alloc_init();
402 sync_mtx_lck = lck_mtx_alloc_init(sync_mtx_lck_grp, sync_mtx_lck_attr);
403
404 /*
405 * Initialize the vnode table
406 */
407 vntblinit();
408 /*
409 * Initialize the filesystem event mechanism.
410 */
411 vfs_event_init();
412 /*
413 * Initialize the vnode name cache
414 */
415 nchinit();
416
417 /*
418 * Build vnode operation vectors.
419 */
420 vfs_op_init();
421 vfs_opv_init(); /* finish the job */
422 /*
423 * Initialize each file system type in the static list,
424 * until the first NULL ->vfs_vfsops is encountered.
425 */
426 maxtypenum = VT_NON;
427 for (vfsp = vfsconf, i = 0; i < maxvfsslots; i++, vfsp++) {
428 struct vfsconf vfsc;
429 if (vfsp->vfc_vfsops == (struct vfsops *)0) {
430 break;
431 }
432 if (i) {
433 vfsconf[i - 1].vfc_next = vfsp;
434 }
435 if (maxtypenum <= vfsp->vfc_typenum) {
436 maxtypenum = vfsp->vfc_typenum + 1;
437 }
438
439 bzero(&vfsc, sizeof(struct vfsconf));
440 vfsc.vfc_reserved1 = 0;
441 bcopy(vfsp->vfc_name, vfsc.vfc_name, sizeof(vfsc.vfc_name));
442 vfsc.vfc_typenum = vfsp->vfc_typenum;
443 vfsc.vfc_refcount = vfsp->vfc_refcount;
444 vfsc.vfc_flags = vfsp->vfc_flags;
445 vfsc.vfc_reserved2 = 0;
446 vfsc.vfc_reserved3 = 0;
447
448 if (vfsp->vfc_vfsops->vfs_sysctl) {
449 struct sysctl_oid *oidp = NULL;
450 struct sysctl_oid oid = SYSCTL_STRUCT_INIT(_vfs, vfsp->vfc_typenum, , CTLTYPE_NODE | CTLFLAG_KERN | CTLFLAG_RW | CTLFLAG_LOCKED, NULL, 0, vfs_sysctl_node, "-", "");
451
452 oidp = kheap_alloc(KHEAP_DEFAULT, sizeof(struct sysctl_oid), Z_WAITOK);
453 *oidp = oid;
454
455 /* Memory for VFS oid held by vfsentry forever */
456 vfsp->vfc_sysctl = oidp;
457 oidp->oid_name = vfsp->vfc_name;
458 sysctl_register_oid(vfsp->vfc_sysctl);
459 }
460
461 (*vfsp->vfc_vfsops->vfs_init)(&vfsc);
462
463 numused_vfsslots++;
464 numregistered_fses++;
465 }
466 /* next vfc_typenum to be used */
467 maxvfstypenum = maxtypenum;
468
469 /*
470 * Initialize the vnop authorization scope.
471 */
472 vnode_authorize_init();
473
474 /*
475 * Initialiize the quota system.
476 */
477 #if QUOTA
478 dqinit();
479 #endif
480
481 /*
482 * create a mount point for dead vnodes
483 */
484 mp = zalloc_flags(mount_zone, Z_WAITOK | Z_ZERO);
485 /* Initialize the default IO constraints */
486 mp->mnt_maxreadcnt = mp->mnt_maxwritecnt = MAXPHYS;
487 mp->mnt_segreadcnt = mp->mnt_segwritecnt = 32;
488 mp->mnt_maxsegreadsize = mp->mnt_maxreadcnt;
489 mp->mnt_maxsegwritesize = mp->mnt_maxwritecnt;
490 mp->mnt_devblocksize = DEV_BSIZE;
491 mp->mnt_alignmentmask = PAGE_MASK;
492 mp->mnt_ioqueue_depth = MNT_DEFAULT_IOQUEUE_DEPTH;
493 mp->mnt_ioscale = 1;
494 mp->mnt_ioflags = 0;
495 mp->mnt_realrootvp = NULLVP;
496 mp->mnt_authcache_ttl = CACHED_LOOKUP_RIGHT_TTL;
497
498 TAILQ_INIT(&mp->mnt_vnodelist);
499 TAILQ_INIT(&mp->mnt_workerqueue);
500 TAILQ_INIT(&mp->mnt_newvnodes);
501 mp->mnt_flag = MNT_LOCAL;
502 mp->mnt_lflag = MNT_LDEAD;
503 mount_lock_init(mp);
504
505 #if CONFIG_MACF
506 mac_mount_label_init(mp);
507 mac_mount_label_associate(vfs_context_kernel(), mp);
508 #endif
509 dead_mountp = mp;
510
511 #if FS_COMPRESSION
512 decmpfs_init();
513 #endif
514
515 nspace_resolver_init();
516 }
517
518 void
519 vnode_list_lock(void)
520 {
521 lck_spin_lock_grp(vnode_list_spin_lock, vnode_list_lck_grp);
522 }
523
524 void
525 vnode_list_unlock(void)
526 {
527 lck_spin_unlock(vnode_list_spin_lock);
528 }
529
530 void
531 mount_list_lock(void)
532 {
533 lck_mtx_lock(mnt_list_mtx_lock);
534 }
535
536 void
537 mount_list_unlock(void)
538 {
539 lck_mtx_unlock(mnt_list_mtx_lock);
540 }
541
542 void
543 mount_lock_init(mount_t mp)
544 {
545 lck_mtx_init(&mp->mnt_mlock, mnt_lck_grp, mnt_lck_attr);
546 lck_mtx_init(&mp->mnt_iter_lock, mnt_lck_grp, mnt_lck_attr);
547 lck_mtx_init(&mp->mnt_renamelock, mnt_lck_grp, mnt_lck_attr);
548 lck_rw_init(&mp->mnt_rwlock, mnt_lck_grp, mnt_lck_attr);
549 }
550
551 void
552 mount_lock_destroy(mount_t mp)
553 {
554 lck_mtx_destroy(&mp->mnt_mlock, mnt_lck_grp);
555 lck_mtx_destroy(&mp->mnt_iter_lock, mnt_lck_grp);
556 lck_mtx_destroy(&mp->mnt_renamelock, mnt_lck_grp);
557 lck_rw_destroy(&mp->mnt_rwlock, mnt_lck_grp);
558 }
559
560
561 /*
562 * Name: vfstable_add
563 *
564 * Description: Add a filesystem to the vfsconf list at the first
565 * unused slot. If no slots are available, return an
566 * error.
567 *
568 * Parameter: nvfsp vfsconf for VFS to add
569 *
570 * Returns: 0 Success
571 * -1 Failure
572 *
573 * Notes: The vfsconf should be treated as a linked list by
574 * all external references, as the implementation is
575 * expected to change in the future. The linkage is
576 * through ->vfc_next, and the list is NULL terminated.
577 *
578 * Warning: This code assumes that vfsconf[0] is non-empty.
579 */
580 struct vfstable *
581 vfstable_add(struct vfstable *nvfsp)
582 {
583 int slot;
584 struct vfstable *slotp, *allocated = NULL;
585 struct sysctl_oid *oidp = NULL;
586
587
588 if (nvfsp->vfc_vfsops->vfs_sysctl) {
589 struct sysctl_oid oid = SYSCTL_STRUCT_INIT(_vfs, nvfsp->vfc_typenum, , CTLTYPE_NODE | CTLFLAG_KERN | CTLFLAG_RW | CTLFLAG_LOCKED, NULL, 0, vfs_sysctl_node, "-", "");
590
591 oidp = kheap_alloc(KHEAP_DEFAULT, sizeof(struct sysctl_oid), Z_WAITOK);
592 *oidp = oid;
593 }
594
595 /*
596 * Find the next empty slot; we recognize an empty slot by a
597 * NULL-valued ->vfc_vfsops, so if we delete a VFS, we must
598 * ensure we set the entry back to NULL.
599 */
600 findslot:
601 mount_list_lock();
602 for (slot = 0; slot < maxvfsslots; slot++) {
603 if (vfsconf[slot].vfc_vfsops == NULL) {
604 break;
605 }
606 }
607 if (slot == maxvfsslots) {
608 if (allocated == NULL) {
609 mount_list_unlock();
610 /* out of static slots; allocate one instead */
611 allocated = kheap_alloc(KHEAP_DEFAULT, sizeof(struct vfstable),
612 Z_WAITOK);
613 goto findslot;
614 } else {
615 slotp = allocated;
616 }
617 } else {
618 slotp = &vfsconf[slot];
619 }
620
621 /*
622 * Replace the contents of the next empty slot with the contents
623 * of the provided nvfsp.
624 *
625 * Note; Takes advantage of the fact that 'slot' was left
626 * with the value of 'maxvfslots' in the allocation case.
627 */
628 bcopy(nvfsp, slotp, sizeof(struct vfstable));
629 if (slot != 0) {
630 slotp->vfc_next = vfsconf[slot - 1].vfc_next;
631 vfsconf[slot - 1].vfc_next = slotp;
632 } else {
633 slotp->vfc_next = NULL;
634 }
635
636 if (slotp != allocated) {
637 /* used a statically allocated slot */
638 numused_vfsslots++;
639 }
640 numregistered_fses++;
641
642 if (oidp) {
643 /* Memory freed in vfstable_del after unregistration */
644 slotp->vfc_sysctl = oidp;
645 oidp->oid_name = slotp->vfc_name;
646 sysctl_register_oid(slotp->vfc_sysctl);
647 }
648
649 mount_list_unlock();
650
651 if (allocated && allocated != slotp) {
652 /* did allocation, but ended up using static slot */
653 kheap_free(KHEAP_DEFAULT, allocated, sizeof(struct vfstable));
654 }
655
656 return slotp;
657 }
658
659 /*
660 * Name: vfstable_del
661 *
662 * Description: Remove a filesystem from the vfsconf list by name.
663 * If no such filesystem exists, return an error.
664 *
665 * Parameter: fs_name name of VFS to remove
666 *
667 * Returns: 0 Success
668 * -1 Failure
669 *
670 * Notes: Hopefully all filesystems have unique names.
671 */
672 int
673 vfstable_del(struct vfstable * vtbl)
674 {
675 struct vfstable **vcpp;
676 struct vfstable *vcdelp;
677
678 #if DEBUG
679 lck_mtx_assert(mnt_list_mtx_lock, LCK_MTX_ASSERT_OWNED);
680 #endif /* DEBUG */
681
682 /*
683 * Traverse the list looking for vtbl; if found, *vcpp
684 * will contain the address of the pointer to the entry to
685 * be removed.
686 */
687 for (vcpp = &vfsconf; *vcpp; vcpp = &(*vcpp)->vfc_next) {
688 if (*vcpp == vtbl) {
689 break;
690 }
691 }
692
693 if (*vcpp == NULL) {
694 return ESRCH; /* vtbl not on vfsconf list */
695 }
696 if ((*vcpp)->vfc_sysctl) {
697 sysctl_unregister_oid((*vcpp)->vfc_sysctl);
698 (*vcpp)->vfc_sysctl->oid_name = NULL;
699 kheap_free(KHEAP_DEFAULT, (*vcpp)->vfc_sysctl, sizeof(struct sysctl_oid));
700 }
701
702 /* Unlink entry */
703 vcdelp = *vcpp;
704 *vcpp = (*vcpp)->vfc_next;
705
706 /*
707 * Is this an entry from our static table? We find out by
708 * seeing if the pointer to the object to be deleted places
709 * the object in the address space containing the table (or not).
710 */
711 if (vcdelp >= vfsconf && vcdelp < (vfsconf + maxvfsslots)) { /* Y */
712 /* Mark as empty for vfscon_add() */
713 bzero(vcdelp, sizeof(struct vfstable));
714 numregistered_fses--;
715 numused_vfsslots--;
716 } else { /* N */
717 /*
718 * This entry was dynamically allocated; we must free it;
719 * we would prefer to have just linked the caller's
720 * vfsconf onto our list, but it may not be persistent
721 * because of the previous (copying) implementation.
722 */
723 numregistered_fses--;
724 mount_list_unlock();
725 kheap_free(KHEAP_DEFAULT, vcdelp, sizeof(struct vfstable));
726 mount_list_lock();
727 }
728
729 #if DEBUG
730 lck_mtx_assert(mnt_list_mtx_lock, LCK_MTX_ASSERT_OWNED);
731 #endif /* DEBUG */
732
733 return 0;
734 }
735
736 void
737 SPECHASH_LOCK(void)
738 {
739 lck_mtx_lock(spechash_mtx_lock);
740 }
741
742 void
743 SPECHASH_UNLOCK(void)
744 {
745 lck_mtx_unlock(spechash_mtx_lock);
746 }