2 * Copyright (c) 1999 Poul-Henning Kamp.
3 * Copyright (c) 2008 Bjoern A. Zeeb.
4 * Copyright (c) 2009 James Gritton.
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 #include <sys/cdefs.h>
30 __FBSDID("$FreeBSD$");
34 #include "opt_inet6.h"
36 #include <sys/param.h>
37 #include <sys/types.h>
38 #include <sys/kernel.h>
39 #include <sys/systm.h>
40 #include <sys/errno.h>
41 #include <sys/sysproto.h>
42 #include <sys/malloc.h>
46 #include <sys/taskqueue.h>
47 #include <sys/fcntl.h>
50 #include <sys/mutex.h>
51 #include <sys/racct.h>
52 #include <sys/refcount.h>
54 #include <sys/sysent.h>
55 #include <sys/namei.h>
56 #include <sys/mount.h>
57 #include <sys/queue.h>
58 #include <sys/socket.h>
59 #include <sys/syscallsubr.h>
60 #include <sys/sysctl.h>
61 #include <sys/vnode.h>
66 #include <netinet/in.h>
69 prison_qcmp_v4(const void *ip1, const void *ip2)
74 * We need to compare in HBO here to get the list sorted as expected
75 * by the result of the code. Sorting NBO addresses gives you
76 * interesting results. If you do not understand, do not try.
78 iaa = ntohl(((const struct in_addr *)ip1)->s_addr);
79 iab = ntohl(((const struct in_addr *)ip2)->s_addr);
82 * Do not simply return the difference of the two numbers, the int is
94 * Restrict a prison's IP address list with its parent's, possibly replacing
95 * it. Return true if the replacement buffer was used (or would have been).
98 prison_restrict_ip4(struct prison *pr, struct in_addr *newip4)
104 if (!(pr->pr_flags & PR_IP4_USER)) {
105 /* This has no user settings, so just copy the parent's list. */
106 if (pr->pr_ip4s < ppr->pr_ip4s) {
108 * There's no room for the parent's list. Use the
109 * new list buffer, which is assumed to be big enough
110 * (if it was passed). If there's no buffer, try to
114 if (newip4 == NULL) {
115 newip4 = malloc(ppr->pr_ip4s * sizeof(*newip4),
120 if (newip4 != NULL) {
121 bcopy(ppr->pr_ip4, newip4,
122 ppr->pr_ip4s * sizeof(*newip4));
123 free(pr->pr_ip4, M_PRISON);
125 pr->pr_ip4s = ppr->pr_ip4s;
129 pr->pr_ip4s = ppr->pr_ip4s;
131 bcopy(ppr->pr_ip4, pr->pr_ip4,
132 pr->pr_ip4s * sizeof(*newip4));
133 else if (pr->pr_ip4 != NULL) {
134 free(pr->pr_ip4, M_PRISON);
137 } else if (pr->pr_ip4s > 0) {
138 /* Remove addresses that aren't in the parent. */
139 for (ij = 0; ij < ppr->pr_ip4s; ij++)
140 if (pr->pr_ip4[0].s_addr == ppr->pr_ip4[ij].s_addr)
142 if (ij < ppr->pr_ip4s)
145 bcopy(pr->pr_ip4 + 1, pr->pr_ip4,
146 --pr->pr_ip4s * sizeof(*pr->pr_ip4));
149 for (ij = 1; ii < pr->pr_ip4s; ) {
150 if (pr->pr_ip4[ii].s_addr == ppr->pr_ip4[0].s_addr) {
154 switch (ij >= ppr->pr_ip4s ? -1 :
155 prison_qcmp_v4(&pr->pr_ip4[ii], &ppr->pr_ip4[ij])) {
157 bcopy(pr->pr_ip4 + ii + 1, pr->pr_ip4 + ii,
158 (--pr->pr_ip4s - ii) * sizeof(*pr->pr_ip4));
169 if (pr->pr_ip4s == 0) {
170 free(pr->pr_ip4, M_PRISON);
178 * Pass back primary IPv4 address of this jail.
180 * If not restricted return success but do not alter the address. Caller has
181 * to make sure to initialize it correctly (e.g. INADDR_ANY).
183 * Returns 0 on success, EAFNOSUPPORT if the jail doesn't allow IPv4.
184 * Address returned in NBO.
187 prison_get_ip4(struct ucred *cred, struct in_addr *ia)
191 KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
192 KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
194 pr = cred->cr_prison;
195 if (!(pr->pr_flags & PR_IP4))
197 mtx_lock(&pr->pr_mtx);
198 if (!(pr->pr_flags & PR_IP4)) {
199 mtx_unlock(&pr->pr_mtx);
202 if (pr->pr_ip4 == NULL) {
203 mtx_unlock(&pr->pr_mtx);
204 return (EAFNOSUPPORT);
207 ia->s_addr = pr->pr_ip4[0].s_addr;
208 mtx_unlock(&pr->pr_mtx);
213 * Return 1 if we should do proper source address selection or are not jailed.
214 * We will return 0 if we should bypass source address selection in favour
215 * of the primary jail IPv4 address. Only in this case *ia will be updated and
217 * Return EAFNOSUPPORT, in case this jail does not allow IPv4.
220 prison_saddrsel_ip4(struct ucred *cred, struct in_addr *ia)
226 KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
227 KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
232 pr = cred->cr_prison;
233 if (pr->pr_flags & PR_IP4_SADDRSEL)
236 lia.s_addr = INADDR_ANY;
237 error = prison_get_ip4(cred, &lia);
240 if (lia.s_addr == INADDR_ANY)
243 ia->s_addr = lia.s_addr;
248 * Return true if pr1 and pr2 have the same IPv4 address restrictions.
251 prison_equal_ip4(struct prison *pr1, struct prison *pr2)
258 * No need to lock since the PR_IP4_USER flag can't be altered for
261 while (pr1 != &prison0 &&
263 !(pr1->pr_flags & PR_VNET) &&
265 !(pr1->pr_flags & PR_IP4_USER))
266 pr1 = pr1->pr_parent;
267 while (pr2 != &prison0 &&
269 !(pr2->pr_flags & PR_VNET) &&
271 !(pr2->pr_flags & PR_IP4_USER))
272 pr2 = pr2->pr_parent;
277 * Make sure our (source) address is set to something meaningful to this
280 * Returns 0 if jail doesn't restrict IPv4 or if address belongs to jail,
281 * EADDRNOTAVAIL if the address doesn't belong, or EAFNOSUPPORT if the jail
282 * doesn't allow IPv4. Address passed in in NBO and returned in NBO.
285 prison_local_ip4(struct ucred *cred, struct in_addr *ia)
291 KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
292 KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
294 pr = cred->cr_prison;
295 if (!(pr->pr_flags & PR_IP4))
297 mtx_lock(&pr->pr_mtx);
298 if (!(pr->pr_flags & PR_IP4)) {
299 mtx_unlock(&pr->pr_mtx);
302 if (pr->pr_ip4 == NULL) {
303 mtx_unlock(&pr->pr_mtx);
304 return (EAFNOSUPPORT);
307 ia0.s_addr = ntohl(ia->s_addr);
309 if (ia0.s_addr == INADDR_ANY) {
311 * In case there is only 1 IPv4 address, bind directly.
313 if (pr->pr_ip4s == 1)
314 ia->s_addr = pr->pr_ip4[0].s_addr;
315 mtx_unlock(&pr->pr_mtx);
319 error = prison_check_ip4_locked(pr, ia);
320 if (error == EADDRNOTAVAIL && ia0.s_addr == INADDR_LOOPBACK) {
321 ia->s_addr = pr->pr_ip4[0].s_addr;
325 mtx_unlock(&pr->pr_mtx);
330 * Rewrite destination address in case we will connect to loopback address.
332 * Returns 0 on success, EAFNOSUPPORT if the jail doesn't allow IPv4.
333 * Address passed in in NBO and returned in NBO.
336 prison_remote_ip4(struct ucred *cred, struct in_addr *ia)
340 KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
341 KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
343 pr = cred->cr_prison;
344 if (!(pr->pr_flags & PR_IP4))
346 mtx_lock(&pr->pr_mtx);
347 if (!(pr->pr_flags & PR_IP4)) {
348 mtx_unlock(&pr->pr_mtx);
351 if (pr->pr_ip4 == NULL) {
352 mtx_unlock(&pr->pr_mtx);
353 return (EAFNOSUPPORT);
356 if (ntohl(ia->s_addr) == INADDR_LOOPBACK &&
357 prison_check_ip4_locked(pr, ia) == EADDRNOTAVAIL) {
358 ia->s_addr = pr->pr_ip4[0].s_addr;
359 mtx_unlock(&pr->pr_mtx);
364 * Return success because nothing had to be changed.
366 mtx_unlock(&pr->pr_mtx);
371 * Check if given address belongs to the jail referenced by cred/prison.
373 * Returns 0 if address belongs to jail,
374 * EADDRNOTAVAIL if the address doesn't belong to the jail.
377 prison_check_ip4_locked(const struct prison *pr, const struct in_addr *ia)
382 * Check the primary IP.
384 if (pr->pr_ip4[0].s_addr == ia->s_addr)
388 * All the other IPs are sorted so we can do a binary search.
394 d = prison_qcmp_v4(&pr->pr_ip4[i+1], ia);
403 return (EADDRNOTAVAIL);
407 prison_check_ip4(const struct ucred *cred, const struct in_addr *ia)
412 KASSERT(cred != NULL, ("%s: cred is NULL", __func__));
413 KASSERT(ia != NULL, ("%s: ia is NULL", __func__));
415 pr = cred->cr_prison;
416 if (!(pr->pr_flags & PR_IP4))
418 mtx_lock(&pr->pr_mtx);
419 if (!(pr->pr_flags & PR_IP4)) {
420 mtx_unlock(&pr->pr_mtx);
423 if (pr->pr_ip4 == NULL) {
424 mtx_unlock(&pr->pr_mtx);
425 return (EAFNOSUPPORT);
428 error = prison_check_ip4_locked(pr, ia);
429 mtx_unlock(&pr->pr_mtx);