2 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
4 * Copyright (c) 1999 Poul-Henning Kamp.
5 * Copyright (c) 2009 James Gritton.
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution.
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
54 #define JAIL_API_VERSION 2
57 * For all xprison structs, always keep the pr_version an int and
58 * the first variable so userspace can easily distinguish them.
64 char pr_path[MAXPATHLEN];
65 char pr_host[MAXHOSTNAMELEN];
74 cpusetid_t pr_cpusetid;
75 char pr_path[MAXPATHLEN];
76 char pr_host[MAXHOSTNAMELEN];
77 char pr_name[MAXHOSTNAMELEN];
82 * sizeof(xprison) will be malloced + size needed for all
83 * IPv4 and IPv6 addesses. Offsets are based numbers of addresses.
85 struct in_addr pr_ip4[];
86 struct in6_addr pr_ip6[];
89 #define XPRISON_VERSION 3
92 PRISON_STATE_INVALID = 0, /* New prison, not ready to be seen */
93 PRISON_STATE_ALIVE, /* Current prison, visible to all */
94 PRISON_STATE_DYING /* Removed but holding resources, */
95 }; /* optionally visible. */
98 * Flags for jail_set and jail_get.
100 #define JAIL_CREATE 0x01 /* Create jail if it doesn't exist */
101 #define JAIL_UPDATE 0x02 /* Update parameters of existing jail */
102 #define JAIL_ATTACH 0x04 /* Attach to jail upon creation */
103 #define JAIL_DYING 0x08 /* Allow getting a dying jail */
104 #define JAIL_SET_MASK 0x0f
105 #define JAIL_GET_MASK 0x08
107 #define JAIL_SYS_DISABLE 0
108 #define JAIL_SYS_NEW 1
109 #define JAIL_SYS_INHERIT 2
116 int jail(struct jail *);
117 int jail_set(struct iovec *, unsigned int, int);
118 int jail_get(struct iovec *, unsigned int, int);
119 int jail_attach(int);
120 int jail_remove(int);
125 #include <sys/queue.h>
126 #include <sys/sysctl.h>
127 #include <sys/lock.h>
128 #include <sys/mutex.h>
129 #include <sys/_task.h>
131 #define JAIL_MAX 999999
133 #ifdef MALLOC_DECLARE
134 MALLOC_DECLARE(M_PRISON);
138 #if defined(_KERNEL) || defined(_WANT_PRISON)
142 #define HOSTUUIDLEN 64
143 #define OSRELEASELEN 32
149 * This structure describes a prison. It is pointed to by all struct
150 * ucreds's of the inmates. pr_ref keeps track of them and is used to
151 * delete the struture when the last inmate is dead.
155 * (c) set only during creation before the structure is shared, no mutex
157 * (m) locked by pr_mtx
158 * (p) locked by pr_mtx, and also at least shared allprison_lock required
160 * (q) locked by both pr_mtx and allprison_lock
161 * (r) atomic via refcount(9), pr_mtx and allprison_lock required to
165 TAILQ_ENTRY(prison) pr_list; /* (a) all prisons */
166 int pr_id; /* (c) prison id */
167 volatile u_int pr_ref; /* (r) refcount */
168 volatile u_int pr_uref; /* (r) user (alive) refcount */
169 unsigned pr_flags; /* (p) PR_* flags */
170 LIST_HEAD(, prison) pr_children; /* (a) list of child jails */
171 LIST_ENTRY(prison) pr_sibling; /* (a) next in parent's list */
172 struct prison *pr_parent; /* (c) containing jail */
174 struct task pr_task; /* (c) destroy task */
175 struct osd pr_osd; /* (p) additional data */
176 struct cpuset *pr_cpuset; /* (p) cpuset */
177 struct vnet *pr_vnet; /* (c) network stack */
178 struct vnode *pr_root; /* (c) vnode to rdir */
179 int pr_ip4s; /* (p) number of v4 IPs */
180 int pr_ip6s; /* (p) number of v6 IPs */
181 struct in_addr *pr_ip4; /* (p) v4 IPs of jail */
182 struct in6_addr *pr_ip6; /* (p) v6 IPs of jail */
183 struct prison_racct *pr_prison_racct; /* (c) racct jail proxy */
185 int pr_childcount; /* (a) number of child jails */
186 int pr_childmax; /* (p) maximum child jails */
187 unsigned pr_allow; /* (p) PR_ALLOW_* flags */
188 int pr_securelevel; /* (p) securelevel */
189 int pr_enforce_statfs; /* (p) statfs permission */
190 int pr_devfs_rsnum; /* (p) devfs ruleset */
191 enum prison_state pr_state; /* (q) state in life cycle */
193 int pr_osreldate; /* (c) kern.osreldate value */
194 unsigned long pr_hostid; /* (p) jail hostid */
195 char pr_name[MAXHOSTNAMELEN]; /* (p) admin jail name */
196 char pr_path[MAXPATHLEN]; /* (c) chroot path */
197 char pr_hostname[MAXHOSTNAMELEN]; /* (p) jail hostname */
198 char pr_domainname[MAXHOSTNAMELEN]; /* (p) jail domainname */
199 char pr_hostuuid[HOSTUUIDLEN]; /* (p) jail hostuuid */
200 char pr_osrelease[OSRELEASELEN]; /* (c) kern.osrelease value */
203 struct prison_racct {
204 LIST_ENTRY(prison_racct) prr_next;
205 char prr_name[MAXHOSTNAMELEN];
207 struct racct *prr_racct;
209 #endif /* _KERNEL || _WANT_PRISON */
212 /* Flag bits set via options */
213 #define PR_PERSIST 0x00000001 /* Can exist without processes */
214 #define PR_HOST 0x00000002 /* Virtualize hostname et al */
215 #define PR_IP4_USER 0x00000004 /* Restrict IPv4 addresses */
216 #define PR_IP6_USER 0x00000008 /* Restrict IPv6 addresses */
217 #define PR_VNET 0x00000010 /* Virtual network stack */
218 #define PR_IP4_SADDRSEL 0x00000080 /* Do IPv4 src addr sel. or use the */
219 /* primary jail address. */
220 #define PR_IP6_SADDRSEL 0x00000100 /* Do IPv6 src addr sel. or use the */
221 /* primary jail address. */
223 /* Internal flag bits */
224 #define PR_REMOVE 0x01000000 /* In process of being removed */
225 #define PR_IP4 0x02000000 /* IPv4 restricted or disabled */
226 /* by this jail or an ancestor */
227 #define PR_IP6 0x04000000 /* IPv6 restricted or disabled */
228 /* by this jail or an ancestor */
229 #define PR_COMPLETE_PROC 0x08000000 /* prison_complete called from */
230 /* prison_proc_free, releases uref */
234 * Bits not noted here may be used for dynamic allow.mount.xxxfs.
236 #define PR_ALLOW_SET_HOSTNAME 0x00000001
237 #define PR_ALLOW_SYSVIPC 0x00000002
238 #define PR_ALLOW_RAW_SOCKETS 0x00000004
239 #define PR_ALLOW_CHFLAGS 0x00000008
240 #define PR_ALLOW_MOUNT 0x00000010
241 #define PR_ALLOW_QUOTAS 0x00000020
242 #define PR_ALLOW_SOCKET_AF 0x00000040
243 #define PR_ALLOW_MLOCK 0x00000080
244 #define PR_ALLOW_READ_MSGBUF 0x00000100
245 #define PR_ALLOW_UNPRIV_DEBUG 0x00000200
246 #define PR_ALLOW_SUSER 0x00000400
247 #define PR_ALLOW_RESERVED_PORTS 0x00008000
248 #define PR_ALLOW_KMEM_ACCESS 0x00010000 /* reserved, not used yet */
249 #define PR_ALLOW_ALL_STATIC 0x000187ff
252 * PR_ALLOW_DIFFERENCES determines which flags are able to be
253 * different between the parent and child jail upon creation.
255 #define PR_ALLOW_DIFFERENCES (PR_ALLOW_UNPRIV_DEBUG)
260 #define PR_METHOD_CREATE 0
261 #define PR_METHOD_GET 1
262 #define PR_METHOD_SET 2
263 #define PR_METHOD_CHECK 3
264 #define PR_METHOD_ATTACH 4
265 #define PR_METHOD_REMOVE 5
266 #define PR_MAXMETHOD 6
269 * Lock/unlock a prison.
270 * XXX These exist not so much for general convenience, but to be useable in
271 * the FOREACH_PRISON_DESCENDANT_LOCKED macro which can't handle them in
272 * non-function form as currently defined.
275 prison_lock(struct prison *pr)
278 mtx_lock(&pr->pr_mtx);
282 prison_unlock(struct prison *pr)
285 mtx_unlock(&pr->pr_mtx);
288 /* Traverse a prison's immediate children. */
289 #define FOREACH_PRISON_CHILD(ppr, cpr) \
290 LIST_FOREACH(cpr, &(ppr)->pr_children, pr_sibling)
293 * Preorder traversal of all of a prison's descendants.
294 * This ugly loop allows the macro to be followed by a single block
295 * as expected in a looping primitive.
297 #define FOREACH_PRISON_DESCENDANT(ppr, cpr, descend) \
298 for ((cpr) = (ppr), (descend) = 1; \
299 ((cpr) = (((descend) && !LIST_EMPTY(&(cpr)->pr_children)) \
300 ? LIST_FIRST(&(cpr)->pr_children) \
303 : (((descend) = LIST_NEXT(cpr, pr_sibling) != NULL) \
304 ? LIST_NEXT(cpr, pr_sibling) \
305 : (cpr)->pr_parent))));) \
311 * As above, but lock descendants on the way down and unlock on the way up.
313 #define FOREACH_PRISON_DESCENDANT_LOCKED(ppr, cpr, descend) \
314 for ((cpr) = (ppr), (descend) = 1; \
315 ((cpr) = (((descend) && !LIST_EMPTY(&(cpr)->pr_children)) \
316 ? LIST_FIRST(&(cpr)->pr_children) \
319 : ((prison_unlock(cpr), \
320 (descend) = LIST_NEXT(cpr, pr_sibling) != NULL) \
321 ? LIST_NEXT(cpr, pr_sibling) \
322 : (cpr)->pr_parent))));) \
323 if ((descend) ? (prison_lock(cpr), 0) : 1) \
328 * As above, but also keep track of the level descended to.
330 #define FOREACH_PRISON_DESCENDANT_LOCKED_LEVEL(ppr, cpr, descend, level)\
331 for ((cpr) = (ppr), (descend) = 1, (level) = 0; \
332 ((cpr) = (((descend) && !LIST_EMPTY(&(cpr)->pr_children)) \
333 ? (level++, LIST_FIRST(&(cpr)->pr_children)) \
336 : ((prison_unlock(cpr), \
337 (descend) = LIST_NEXT(cpr, pr_sibling) != NULL) \
338 ? LIST_NEXT(cpr, pr_sibling) \
339 : (level--, (cpr)->pr_parent)))));) \
340 if ((descend) ? (prison_lock(cpr), 0) : 1) \
345 * Attributes of the physical system, and the root of the jail tree.
347 extern struct prison prison0;
349 TAILQ_HEAD(prisonlist, prison);
350 extern struct prisonlist allprison;
351 extern struct sx allprison_lock;
354 * Sysctls to describe jail parameters.
356 SYSCTL_DECL(_security_jail_param);
358 #define SYSCTL_JAIL_PARAM(module, param, type, fmt, descr) \
359 SYSCTL_PROC(_security_jail_param ## module, OID_AUTO, param, \
360 (type) | CTLFLAG_MPSAFE, NULL, 0, sysctl_jail_param, fmt, descr)
361 #define SYSCTL_JAIL_PARAM_STRING(module, param, access, len, descr) \
362 SYSCTL_PROC(_security_jail_param ## module, OID_AUTO, param, \
363 CTLTYPE_STRING | CTLFLAG_MPSAFE | (access), NULL, len, \
364 sysctl_jail_param, "A", descr)
365 #define SYSCTL_JAIL_PARAM_STRUCT(module, param, access, len, fmt, descr)\
366 SYSCTL_PROC(_security_jail_param ## module, OID_AUTO, param, \
367 CTLTYPE_STRUCT | CTLFLAG_MPSAFE | (access), NULL, len, \
368 sysctl_jail_param, fmt, descr)
369 #define SYSCTL_JAIL_PARAM_NODE(module, descr) \
370 SYSCTL_NODE(_security_jail_param, OID_AUTO, module, CTLFLAG_MPSAFE, \
372 #define SYSCTL_JAIL_PARAM_SUBNODE(parent, module, descr) \
373 SYSCTL_NODE(_security_jail_param_##parent, OID_AUTO, module, \
374 CTLFLAG_MPSAFE, 0, descr)
375 #define SYSCTL_JAIL_PARAM_SYS_NODE(module, access, descr) \
376 SYSCTL_JAIL_PARAM_NODE(module, descr); \
377 SYSCTL_JAIL_PARAM(_##module, , CTLTYPE_INT | (access), "E,jailsys", \
381 * Kernel support functions for jail().
390 * Return 1 if the passed credential is in a jail, otherwise 0.
392 #define jailed(cred) (cred->cr_prison != &prison0)
394 int jailed_without_vnet(struct ucred *);
395 void getcredhostname(struct ucred *, char *, size_t);
396 void getcreddomainname(struct ucred *, char *, size_t);
397 void getcredhostuuid(struct ucred *, char *, size_t);
398 void getcredhostid(struct ucred *, unsigned long *);
399 void getjailname(struct ucred *cred, char *name, size_t len);
400 void prison0_init(void);
401 int prison_allow(struct ucred *, unsigned);
402 int prison_check(struct ucred *cred1, struct ucred *cred2);
403 int prison_owns_vnet(struct ucred *);
404 int prison_canseemount(struct ucred *cred, struct mount *mp);
405 void prison_enforce_statfs(struct ucred *cred, struct mount *mp,
407 struct prison *prison_find(int prid);
408 struct prison *prison_find_child(struct prison *, int);
409 struct prison *prison_find_name(struct prison *, const char *);
410 int prison_flag(struct ucred *, unsigned);
411 void prison_free(struct prison *pr);
412 void prison_free_locked(struct prison *pr);
413 void prison_hold(struct prison *pr);
414 void prison_hold_locked(struct prison *pr);
415 void prison_proc_hold(struct prison *);
416 void prison_proc_free(struct prison *);
417 void prison_set_allow(struct ucred *cred, unsigned flag, int enable);
418 int prison_ischild(struct prison *, struct prison *);
419 bool prison_isalive(struct prison *);
420 bool prison_isvalid(struct prison *);
421 int prison_equal_ip4(struct prison *, struct prison *);
422 int prison_get_ip4(struct ucred *cred, struct in_addr *ia);
423 int prison_local_ip4(struct ucred *cred, struct in_addr *ia);
424 int prison_remote_ip4(struct ucred *cred, struct in_addr *ia);
425 int prison_check_ip4(const struct ucred *, const struct in_addr *);
426 int prison_check_ip4_locked(const struct prison *, const struct in_addr *);
427 int prison_saddrsel_ip4(struct ucred *, struct in_addr *);
428 int prison_restrict_ip4(struct prison *, struct in_addr *);
429 int prison_qcmp_v4(const void *, const void *);
431 int prison_equal_ip6(struct prison *, struct prison *);
432 int prison_get_ip6(struct ucred *, struct in6_addr *);
433 int prison_local_ip6(struct ucred *, struct in6_addr *, int);
434 int prison_remote_ip6(struct ucred *, struct in6_addr *);
435 int prison_check_ip6(const struct ucred *, const struct in6_addr *);
436 int prison_check_ip6_locked(const struct prison *, const struct in6_addr *);
437 int prison_saddrsel_ip6(struct ucred *, struct in6_addr *);
438 int prison_restrict_ip6(struct prison *, struct in6_addr *);
439 int prison_qcmp_v6(const void *, const void *);
441 int prison_check_af(struct ucred *cred, int af);
442 int prison_if(struct ucred *cred, const struct sockaddr *sa);
443 char *prison_name(struct prison *, struct prison *);
444 int prison_priv_check(struct ucred *cred, int priv);
445 int sysctl_jail_param(SYSCTL_HANDLER_ARGS);
446 unsigned prison_add_allow(const char *prefix, const char *name,
447 const char *prefix_descr, const char *descr);
448 void prison_add_vfs(struct vfsconf *vfsp);
449 void prison_racct_foreach(void (*callback)(struct racct *racct,
450 void *arg2, void *arg3), void (*pre)(void), void (*post)(void),
451 void *arg2, void *arg3);
452 struct prison_racct *prison_racct_find(const char *name);
453 void prison_racct_hold(struct prison_racct *prr);
454 void prison_racct_free(struct prison_racct *prr);
457 #endif /* !_SYS_JAIL_H_ */