]> git.saurik.com Git - apple/xnu.git/blob - bsd/kern/uipc_domain.c
xnu-2050.24.15.tar.gz
[apple/xnu.git] / bsd / kern / uipc_domain.c
1 /*
2 * Copyright (c) 1998-2012 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) 1982, 1986, 1993
31 * The Regents of the University of California. All rights reserved.
32 *
33 * Redistribution and use in source and binary forms, with or without
34 * modification, are permitted provided that the following conditions
35 * are met:
36 * 1. Redistributions of source code must retain the above copyright
37 * notice, this list of conditions and the following disclaimer.
38 * 2. Redistributions in binary form must reproduce the above copyright
39 * notice, this list of conditions and the following disclaimer in the
40 * documentation and/or other materials provided with the distribution.
41 * 3. All advertising materials mentioning features or use of this software
42 * must display the following acknowledgement:
43 * This product includes software developed by the University of
44 * California, Berkeley and its contributors.
45 * 4. Neither the name of the University nor the names of its contributors
46 * may be used to endorse or promote products derived from this software
47 * without specific prior written permission.
48 *
49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
59 * SUCH DAMAGE.
60 *
61 * @(#)uipc_domain.c 8.3 (Berkeley) 2/14/95
62 */
63
64 #include <sys/param.h>
65 #include <sys/socket.h>
66 #include <sys/protosw.h>
67 #include <sys/domain.h>
68 #include <sys/mcache.h>
69 #include <sys/mbuf.h>
70 #include <sys/time.h>
71 #include <sys/kernel.h>
72 #include <sys/systm.h>
73 #include <sys/proc_internal.h>
74 #include <sys/sysctl.h>
75 #include <sys/syslog.h>
76 #include <sys/queue.h>
77
78 #include <net/dlil.h>
79
80 #include <pexpert/pexpert.h>
81
82 void init_domain(struct domain *dp) __attribute__((section("__TEXT, initcode")));
83 void prepend_domain(struct domain *dp) __attribute__((section("__TEXT, initcode")));
84
85 void pfslowtimo(void *);
86
87 struct protosw *pffindprotonotype(int, int);
88 struct protosw *pffindprotonotype_locked(int , int , int);
89 struct domain *pffinddomain(int);
90 static void net_update_uptime(void);
91
92 /*
93 * Add/delete 'domain': Link structure into system list,
94 * invoke the domain init, and then the proto inits.
95 * To delete, just remove from the list (dom_refs must be zero)
96 */
97
98 lck_grp_t *domain_proto_mtx_grp;
99 lck_attr_t *domain_proto_mtx_attr;
100 static lck_grp_attr_t *domain_proto_mtx_grp_attr;
101 decl_lck_mtx_data(static, domain_proto_mtx);
102 extern int do_reclaim;
103
104 extern sysctlfn net_sysctl;
105
106 static u_int64_t _net_uptime;
107
108 static void
109 init_proto(struct protosw *pr)
110 {
111 TAILQ_INIT(&pr->pr_filter_head);
112 if (pr->pr_init)
113 (*pr->pr_init)();
114
115 /* Make sure pr_init isn't called again!! */
116 pr->pr_init = 0;
117 }
118
119 void
120 init_domain(struct domain *dp)
121 {
122 struct protosw *pr;
123
124 if ((dp->dom_mtx = lck_mtx_alloc_init(domain_proto_mtx_grp, domain_proto_mtx_attr)) == NULL) {
125 printf("init_domain: can't init domain mtx for domain=%s\n", dp->dom_name);
126 return; /* we have a problem... */
127 }
128
129 if (dp->dom_init)
130 (*dp->dom_init)();
131
132 /* and then init the currently installed protos in this domain */
133
134 for (pr = dp->dom_protosw; pr; pr = pr->pr_next) {
135 if (pr->pr_usrreqs == 0)
136 panic("domaininit: %ssw[%d] has no usrreqs!",
137 dp->dom_name,
138 (int)(pr - dp->dom_protosw));
139
140 #if __APPLE__
141 /*
142 * Warn that pr_fasttimo (now pr_unused) is deprecated since rdar://7617868
143 */
144 if (pr->pr_unused != NULL) {
145 printf("init_domain: warning %s, proto %d: pr_fasttimo is deprecated and won't be called\n",
146 dp->dom_name, pr->pr_protocol);
147 }
148 #endif
149
150 init_proto(pr);
151
152 }
153
154 /* Recompute for new protocol */
155 if (_max_linkhdr < 16) /* XXX - Sheesh; everything's ether? */
156 _max_linkhdr = 16;
157 _max_linkhdr = max_linkhdr; /* round it up */
158
159 if (dp->dom_protohdrlen > _max_protohdr)
160 _max_protohdr = dp->dom_protohdrlen;
161 _max_protohdr = max_protohdr; /* round it up */
162
163 max_hdr = max_linkhdr + max_protohdr;
164 max_datalen = MHLEN - max_hdr;
165 }
166
167 void
168 prepend_domain(struct domain *dp)
169 {
170 lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_OWNED);
171 dp->dom_next = domains;
172 domains = dp;
173 }
174
175 void
176 net_add_domain(struct domain *dp)
177 {
178 int do_unlock;
179
180 kprintf("Adding domain %s (family %d)\n", dp->dom_name,
181 dp->dom_family);
182 /* First, link in the domain */
183
184 do_unlock = domain_proto_mtx_lock();
185 prepend_domain(dp);
186
187 init_domain(dp);
188 domain_proto_mtx_unlock(do_unlock);
189
190 }
191
192 int
193 net_del_domain(struct domain *dp)
194 { register struct domain *dp1, *dp2;
195 register int retval = 0;
196 int do_unlock;
197
198 do_unlock = domain_proto_mtx_lock();
199
200 if (dp->dom_refs) {
201 domain_proto_mtx_unlock(do_unlock);
202 return(EBUSY);
203 }
204
205 for (dp2 = NULL, dp1 = domains; dp1; dp2 = dp1, dp1 = dp1->dom_next)
206 { if (dp == dp1)
207 break;
208 }
209 if (dp1)
210 { if (dp2)
211 dp2->dom_next = dp1->dom_next;
212 else
213 domains = dp1->dom_next;
214 } else
215 retval = EPFNOSUPPORT;
216 domain_proto_mtx_unlock(do_unlock);
217
218 return(retval);
219 }
220
221 /*
222 * net_add_proto - link a protosw into a domain's protosw chain
223 *
224 * note: protocols must use their own domain lock before calling net_add_proto
225 */
226 int
227 net_add_proto(struct protosw *pp, struct domain *dp)
228 { register struct protosw *pp1, *pp2;
229
230 for (pp2 = NULL, pp1 = dp->dom_protosw; pp1; pp1 = pp1->pr_next)
231 { if (pp1->pr_type == pp->pr_type &&
232 pp1->pr_protocol == pp->pr_protocol) {
233 return(EEXIST);
234 }
235 pp2 = pp1;
236 }
237 if (pp2 == NULL)
238 dp->dom_protosw = pp;
239 else
240 pp2->pr_next = pp;
241
242 init_proto(pp);
243
244 return(0);
245 }
246
247 /*
248 * net_del_proto - remove a protosw from a domain's protosw chain.
249 * Search the protosw chain for the element with matching data.
250 * Then unlink and return.
251 *
252 * note: protocols must use their own domain lock before calling net_del_proto
253 */
254 int
255 net_del_proto(int type, int protocol, struct domain *dp)
256 {
257 register struct protosw *pp1, *pp2;
258
259 for (pp2 = NULL, pp1 = dp->dom_protosw; pp1; pp1 = pp1->pr_next)
260 { if (pp1->pr_type == type &&
261 pp1->pr_protocol == protocol)
262 break;
263 pp2 = pp1;
264 }
265 if (pp1 == NULL) {
266 return(ENXIO);
267 }
268 if (pp2)
269 pp2->pr_next = pp1->pr_next;
270 else
271 dp->dom_protosw = pp1->pr_next;
272 return(0);
273 }
274
275
276 #if NS
277 extern struct domain nsdomain;
278 #endif
279 #if ISO
280 extern struct domain isodomain;
281 #endif
282 #if CCITT
283 extern struct domain ccittdomain;
284 #endif
285
286 #if NETAT
287 extern struct domain atalkdomain;
288 #endif
289 #if INET6
290 extern struct domain inet6domain;
291 #endif
292 #if IPSEC
293 extern struct domain keydomain;
294 #endif
295
296 extern struct domain routedomain, ndrvdomain, inetdomain;
297 extern struct domain systemdomain;
298
299 void
300 domaininit(void)
301 {
302 register struct domain *dp;
303 int do_unlock;
304
305 /*
306 * allocate lock group attribute and group for domain mutexes
307 */
308 domain_proto_mtx_grp_attr = lck_grp_attr_alloc_init();
309
310 domain_proto_mtx_grp = lck_grp_alloc_init("domain", domain_proto_mtx_grp_attr);
311
312 /*
313 * allocate the lock attribute for per domain mutexes
314 */
315 domain_proto_mtx_attr = lck_attr_alloc_init();
316
317 lck_mtx_init(&domain_proto_mtx, domain_proto_mtx_grp,
318 domain_proto_mtx_attr);
319 /*
320 * Add all the static domains to the domains list
321 */
322
323 do_unlock = domain_proto_mtx_lock();
324
325 prepend_domain(&localdomain);
326 prepend_domain(&inetdomain);
327 #if NETAT
328 prepend_domain(&atalkdomain);
329 #endif
330 #if INET6
331 prepend_domain(&inet6domain);
332 #endif
333 prepend_domain(&routedomain);
334
335 #if IPSEC
336 prepend_domain(&keydomain);
337 #endif
338
339 #if NS
340 prepend_domain(&nsdomain);
341 #endif
342 #if ISO
343 prepend_domain(&isodomain);
344 #endif
345 #if CCITT
346 prepend_domain(&ccittdomain);
347 #endif
348 prepend_domain(&ndrvdomain);
349
350 prepend_domain(&systemdomain);
351
352 /*
353 * Now ask them all to init (XXX including the routing domain,
354 * see above)
355 */
356 for (dp = domains; dp; dp = dp->dom_next)
357 init_domain(dp);
358
359 domain_proto_mtx_unlock(do_unlock);
360 timeout(pfslowtimo, NULL, 1);
361 }
362
363 static __inline__ struct domain *
364 pffinddomain_locked(int pf)
365 {
366 struct domain *dp;
367
368 dp = domains;
369 while (dp != NULL)
370 { if (dp->dom_family == pf) {
371 break;
372 }
373 dp = dp->dom_next;
374 }
375 return (dp);
376 }
377
378 struct protosw *
379 pffindtype(int family, int type)
380 {
381 register struct domain *dp;
382 register struct protosw *pr;
383 int do_unlock;
384
385 do_unlock = domain_proto_mtx_lock();
386 dp = pffinddomain_locked(family);
387 if (dp == NULL) {
388 domain_proto_mtx_unlock(do_unlock);
389 return (NULL);
390 }
391 for (pr = dp->dom_protosw; pr; pr = pr->pr_next)
392 if (pr->pr_type && pr->pr_type == type) {
393 domain_proto_mtx_unlock(do_unlock);
394 return (pr);
395 }
396 domain_proto_mtx_unlock(do_unlock);
397 return (0);
398 }
399
400 struct domain *
401 pffinddomain(int pf)
402 {
403 struct domain *dp;
404 int do_unlock;
405
406 do_unlock = domain_proto_mtx_lock();
407 dp = pffinddomain_locked(pf);
408 domain_proto_mtx_unlock(do_unlock);
409 return(dp);
410 }
411
412 struct protosw *
413 pffindproto(int family, int protocol, int type)
414 {
415 register struct protosw *pr;
416 int do_unlock;
417 do_unlock = domain_proto_mtx_lock();
418 pr = pffindproto_locked(family, protocol, type);
419 domain_proto_mtx_unlock(do_unlock);
420 return (pr);
421 }
422
423 struct protosw *
424 pffindproto_locked(int family, int protocol, int type)
425 {
426 register struct domain *dp;
427 register struct protosw *pr;
428 struct protosw *maybe = 0;
429
430 if (family == 0)
431 return (0);
432 dp = pffinddomain_locked(family);
433 if (dp == NULL) {
434 return (NULL);
435 }
436 for (pr = dp->dom_protosw; pr; pr = pr->pr_next) {
437 if ((pr->pr_protocol == protocol) && (pr->pr_type == type))
438 return (pr);
439
440 if (type == SOCK_RAW && pr->pr_type == SOCK_RAW &&
441 pr->pr_protocol == 0 && maybe == (struct protosw *)0)
442 maybe = pr;
443 }
444 return (maybe);
445 }
446
447 struct protosw *
448 pffindprotonotype_locked(int family, int protocol, __unused int type)
449 {
450 register struct domain *dp;
451 register struct protosw *pr;
452
453 if (family == 0)
454 return (0);
455 dp = pffinddomain_locked(family);
456 if (dp == NULL) {
457 return (NULL);
458 }
459 for (pr = dp->dom_protosw; pr; pr = pr->pr_next) {
460 if (pr->pr_protocol == protocol) {
461 return (pr);
462 }
463 }
464 return (NULL);
465 }
466
467 struct protosw *
468 pffindprotonotype(int family, int protocol)
469 {
470 register struct protosw *pr;
471 int do_unlock;
472 if (protocol == 0) {
473 return (NULL);
474 }
475 do_unlock = domain_proto_mtx_lock();
476 pr = pffindprotonotype_locked(family, protocol, 0);
477 domain_proto_mtx_unlock(do_unlock);
478 return (pr);
479 }
480
481 int
482 net_sysctl(int *name, u_int namelen, user_addr_t oldp, size_t *oldlenp,
483 user_addr_t newp, size_t newlen, __unused struct proc *p)
484 {
485 register struct domain *dp;
486 register struct protosw *pr;
487 int family, protocol, error;
488 int do_unlock;
489
490 /*
491 * All sysctl names at this level are nonterminal;
492 * next two components are protocol family and protocol number,
493 * then at least one addition component.
494 */
495 if (namelen < 3)
496 return (EISDIR); /* overloaded */
497 family = name[0];
498 protocol = name[1];
499
500 if (family == 0)
501 return (0);
502 do_unlock = domain_proto_mtx_lock();
503 for (dp = domains; dp; dp = dp->dom_next)
504 if (dp->dom_family == family)
505 goto found;
506 domain_proto_mtx_unlock(do_unlock);
507 return (ENOPROTOOPT);
508 found:
509 for (pr = dp->dom_protosw; pr; pr = pr->pr_next)
510 if (pr->pr_protocol == protocol && pr->pr_sysctl) {
511 error = (*pr->pr_sysctl)(name + 2, namelen - 2,
512 (void *)(uintptr_t)oldp, oldlenp, (void *)(uintptr_t)newp, newlen);
513 domain_proto_mtx_unlock(do_unlock);
514 return (error);
515 }
516 domain_proto_mtx_unlock(do_unlock);
517 return (ENOPROTOOPT);
518 }
519
520 void
521 pfctlinput(int cmd, struct sockaddr *sa)
522 {
523 pfctlinput2(cmd, sa, (void*)0);
524 }
525
526 void
527 pfctlinput2(int cmd, struct sockaddr *sa, void *ctlparam)
528 {
529 struct domain *dp;
530 struct protosw *pr;
531 int do_unlock;
532
533 if (!sa)
534 return;
535
536 do_unlock = domain_proto_mtx_lock();
537 for (dp = domains; dp; dp = dp->dom_next)
538 for (pr = dp->dom_protosw; pr; pr = pr->pr_next)
539 if (pr->pr_ctlinput)
540 (*pr->pr_ctlinput)(cmd, sa, ctlparam);
541 domain_proto_mtx_unlock(do_unlock);
542 }
543
544 void
545 pfslowtimo(__unused void *arg)
546 {
547 register struct domain *dp;
548 register struct protosw *pr;
549 int do_unlock;
550
551 /*
552 * Update coarse-grained networking timestamp (in sec.); the idea
553 * is to piggy-back on the periodic slow timeout callout to update
554 * the counter returnable via net_uptime().
555 */
556 net_update_uptime();
557
558 do_unlock = domain_proto_mtx_lock();
559 for (dp = domains; dp; dp = dp->dom_next)
560 for (pr = dp->dom_protosw; pr; pr = pr->pr_next) {
561 if (pr->pr_slowtimo)
562 (*pr->pr_slowtimo)();
563 if ((do_reclaim || (pr->pr_flags & PR_AGGDRAIN)) &&
564 pr->pr_drain)
565 (*pr->pr_drain)();
566 }
567 do_reclaim = 0;
568 domain_proto_mtx_unlock(do_unlock);
569 timeout(pfslowtimo, NULL, hz/PR_SLOWHZ);
570 }
571
572 static void
573 net_update_uptime(void)
574 {
575 struct timeval tv;
576
577 microuptime(&tv);
578 _net_uptime = tv.tv_sec;
579 }
580
581 /*
582 * An alternative way to obtain the coarse-grained uptime (in seconds)
583 * for networking code which do not require high-precision timestamp,
584 * as this is significantly cheaper than microuptime().
585 */
586 u_int64_t
587 net_uptime(void)
588 {
589 /* If we get here before pfslowtimo() fires for the first time */
590 if (_net_uptime == 0)
591 net_update_uptime();
592
593 return (_net_uptime);
594 }
595
596 int
597 domain_proto_mtx_lock(void)
598 {
599 int held = net_thread_check_lock(NET_THREAD_HELD_DOMAIN);
600 if (!held) {
601 lck_mtx_lock(&domain_proto_mtx);
602 net_thread_set_lock(NET_THREAD_HELD_DOMAIN);
603 }
604 lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_OWNED);
605 return !held;
606 }
607
608 void
609 domain_proto_mtx_unlock(int do_unlock)
610 {
611 if (do_unlock) {
612 net_thread_unset_lock(NET_THREAD_HELD_DOMAIN);
613 lck_mtx_unlock(&domain_proto_mtx);
614 lck_mtx_assert(&domain_proto_mtx, LCK_MTX_ASSERT_NOTOWNED);
615 }
616 }