1 /* 2 * CDDL HEADER START 3 * 4 * The contents of this file are subject to the terms of the 5 * Common Development and Distribution License (the "License"). 6 * You may not use this file except in compliance with the License. 7 * 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 9 * or http://www.opensolaris.org/os/licensing. 10 * See the License for the specific language governing permissions 11 * and limitations under the License. 12 * 13 * When distributing Covered Code, include this CDDL HEADER in each 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 15 * If applicable, add the following below this CDDL HEADER, with the 16 * fields enclosed by brackets "[]" replaced with your own identifying 17 * information: Portions Copyright [yyyy] [name of copyright owner] 18 * 19 * CDDL HEADER END 20 */ 21 /* 22 * Copyright 2006 Sun Microsystems, Inc. All rights reserved. 23 * Use is subject to license terms. 24 */ 25 26 #ifndef _SYS_ZONE_H 27 #define _SYS_ZONE_H 28 29 #pragma ident "%Z%%M% %I% %E% SMI" 30 31 #include <sys/types.h> 32 #include <sys/mutex.h> 33 #include <sys/param.h> 34 #include <sys/rctl.h> 35 #include <sys/pset.h> 36 #include <sys/tsol/label.h> 37 #include <sys/uadmin.h> 38 39 #ifdef __cplusplus 40 extern "C" { 41 #endif 42 43 /* 44 * NOTE 45 * 46 * The contents of this file are private to the implementation of 47 * Solaris and are subject to change at any time without notice. 48 * Applications and drivers using these interfaces may fail to 49 * run on future releases. 50 */ 51 52 /* Available both in kernel and for user space */ 53 54 /* zone id restrictions and special ids */ 55 #define MAX_ZONEID 9999 56 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */ 57 #define MIN_ZONEID 0 /* minimum zone ID on system */ 58 #define GLOBAL_ZONEID 0 59 #define ZONEID_WIDTH 4 /* for printf */ 60 61 /* 62 * Special zoneid_t token to refer to all zones. 63 */ 64 #define ALL_ZONES (-1) 65 66 /* system call subcodes */ 67 #define ZONE_CREATE 0 68 #define ZONE_DESTROY 1 69 #define ZONE_GETATTR 2 70 #define ZONE_ENTER 3 71 #define ZONE_LIST 4 72 #define ZONE_SHUTDOWN 5 73 #define ZONE_LOOKUP 6 74 #define ZONE_BOOT 7 75 #define ZONE_VERSION 8 76 #define ZONE_SETATTR 9 77 78 /* zone attributes */ 79 #define ZONE_ATTR_ROOT 1 80 #define ZONE_ATTR_NAME 2 81 #define ZONE_ATTR_STATUS 3 82 #define ZONE_ATTR_PRIVSET 4 83 #define ZONE_ATTR_UNIQID 5 84 #define ZONE_ATTR_POOLID 6 85 #define ZONE_ATTR_INITPID 7 86 #define ZONE_ATTR_SLBL 8 87 #define ZONE_ATTR_INITNAME 9 88 #define ZONE_ATTR_BOOTARGS 10 89 90 #define ZONE_EVENT_CHANNEL "com.sun:zones:status" 91 #define ZONE_EVENT_STATUS_CLASS "status" 92 #define ZONE_EVENT_STATUS_SUBCLASS "change" 93 94 #define ZONE_EVENT_UNINITIALIZED "uninitialized" 95 #define ZONE_EVENT_READY "ready" 96 #define ZONE_EVENT_RUNNING "running" 97 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down" 98 99 #define ZONE_CB_NAME "zonename" 100 #define ZONE_CB_NEWSTATE "newstate" 101 #define ZONE_CB_OLDSTATE "oldstate" 102 #define ZONE_CB_TIMESTAMP "when" 103 #define ZONE_CB_ZONEID "zoneid" 104 105 #ifdef _SYSCALL32 106 typedef struct { 107 caddr32_t zone_name; 108 caddr32_t zone_root; 109 caddr32_t zone_privs; 110 size32_t zone_privssz; 111 caddr32_t rctlbuf; 112 size32_t rctlbufsz; 113 caddr32_t extended_error; 114 caddr32_t zfsbuf; 115 size32_t zfsbufsz; 116 int match; /* match level */ 117 uint32_t doi; /* DOI for label */ 118 caddr32_t label; /* label associated with zone */ 119 } zone_def32; 120 #endif 121 typedef struct { 122 const char *zone_name; 123 const char *zone_root; 124 const struct priv_set *zone_privs; 125 size_t zone_privssz; 126 const char *rctlbuf; 127 size_t rctlbufsz; 128 int *extended_error; 129 const char *zfsbuf; 130 size_t zfsbufsz; 131 int match; /* match level */ 132 uint32_t doi; /* DOI for label */ 133 const bslabel_t *label; /* label associated with zone */ 134 } zone_def; 135 136 /* extended error information */ 137 #define ZE_UNKNOWN 0 /* No extended error info */ 138 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */ 139 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */ 140 141 /* zone_status */ 142 typedef enum { 143 ZONE_IS_UNINITIALIZED = 0, 144 ZONE_IS_READY, 145 ZONE_IS_BOOTING, 146 ZONE_IS_RUNNING, 147 ZONE_IS_SHUTTING_DOWN, 148 ZONE_IS_EMPTY, 149 ZONE_IS_DOWN, 150 ZONE_IS_DYING, 151 ZONE_IS_DEAD 152 } zone_status_t; 153 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED 154 #define ZONE_MAX_STATE ZONE_IS_DEAD 155 156 /* 157 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also 158 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT. 159 */ 160 typedef enum zone_cmd { 161 Z_READY, Z_BOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING, 162 Z_MOUNT, Z_UNMOUNT 163 } zone_cmd_t; 164 165 /* 166 * The structure of a request to zoneadmd. 167 */ 168 typedef struct zone_cmd_arg { 169 uint64_t uniqid; /* unique "generation number" */ 170 zone_cmd_t cmd; /* requested action */ 171 uint32_t _pad; /* need consistent 32/64 bit alignmt */ 172 char locale[MAXPATHLEN]; /* locale in which to render messages */ 173 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */ 174 } zone_cmd_arg_t; 175 176 /* 177 * Structure of zoneadmd's response to a request. A NULL return value means 178 * the caller should attempt to restart zoneadmd and retry. 179 */ 180 typedef struct zone_cmd_rval { 181 int rval; /* return value of request */ 182 char errbuf[1]; /* variable-sized buffer containing error messages */ 183 } zone_cmd_rval_t; 184 185 /* 186 * The zone support infrastructure uses the zone name as a component 187 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters 188 * in length, so ZONENAME_MAX is limited by that. 189 */ 190 #define ZONENAME_MAX 64 191 192 #define GLOBAL_ZONENAME "global" 193 194 /* 195 * Extended Regular expression (see regex(5)) which matches all valid zone 196 * names. 197 */ 198 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}" 199 200 /* 201 * Where the zones support infrastructure places temporary files. 202 */ 203 #define ZONES_TMPDIR "/var/run/zones" 204 205 /* 206 * The path to the door used by clients to communicate with zoneadmd. 207 */ 208 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door" 209 210 #ifdef _KERNEL 211 /* 212 * We need to protect the definition of 'list_t' from userland applications and 213 * libraries which may be defining ther own versions. 214 */ 215 #include <sys/list.h> 216 217 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */ 218 219 /* zone_flags */ 220 #define ZF_DESTROYED 0x1 /* ZSD destructor callbacks run */ 221 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */ 222 #define ZF_IS_SCRATCH 0x4 /* scratch zone */ 223 224 struct pool; 225 226 /* 227 * Structure to record list of ZFS datasets exported to a zone. 228 */ 229 typedef struct zone_dataset { 230 char *zd_dataset; 231 list_node_t zd_linkage; 232 } zone_dataset_t; 233 234 typedef struct zone { 235 /* 236 * zone_name is never modified once set. 237 */ 238 char *zone_name; /* zone's configuration name */ 239 /* 240 * zone_nodename and zone_domain are never freed once allocated. 241 */ 242 char *zone_nodename; /* utsname.nodename equivalent */ 243 char *zone_domain; /* srpc_domain equivalent */ 244 /* 245 * zone_lock protects the following fields of a zone_t: 246 * zone_ref 247 * zone_cred_ref 248 * zone_ntasks 249 * zone_flags 250 * zone_zsd 251 */ 252 kmutex_t zone_lock; 253 /* 254 * zone_linkage is the zone's linkage into the active or 255 * death-row list. The field is protected by zonehash_lock. 256 */ 257 list_node_t zone_linkage; 258 zoneid_t zone_id; /* ID of zone */ 259 uint_t zone_ref; /* count of zone_hold()s on zone */ 260 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */ 261 /* 262 * zone_rootvp and zone_rootpath can never be modified once set. 263 */ 264 struct vnode *zone_rootvp; /* zone's root vnode */ 265 char *zone_rootpath; /* Path to zone's root + '/' */ 266 ushort_t zone_flags; /* misc flags */ 267 zone_status_t zone_status; /* protected by zone_status_lock */ 268 uint_t zone_ntasks; /* number of tasks executing in zone */ 269 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */ 270 /* counters in projects and tasks */ 271 /* that are within the zone */ 272 rctl_qty_t zone_nlwps; /* number of lwps in zone */ 273 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */ 274 275 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */ 276 uint32_t zone_shares; /* FSS shares allocated to zone */ 277 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */ 278 list_t zone_zsd; /* list of Zone-Specific Data values */ 279 kcondvar_t zone_cv; /* used to signal state changes */ 280 struct proc *zone_zsched; /* Dummy kernel "zsched" process */ 281 pid_t zone_proc_initpid; /* pid of "init" for this zone */ 282 char *zone_initname; /* fs path to 'init' */ 283 int zone_boot_err; /* for zone_boot() if boot fails */ 284 char *zone_bootargs; /* arguments passed via zone_boot() */ 285 /* 286 * zone_kthreads is protected by zone_status_lock. 287 */ 288 kthread_t *zone_kthreads; /* kernel threads in zone */ 289 struct priv_set *zone_privset; /* limit set for zone */ 290 /* 291 * zone_vfslist is protected by vfs_list_lock(). 292 */ 293 struct vfs *zone_vfslist; /* list of FS's mounted in zone */ 294 uint64_t zone_uniqid; /* unique zone generation number */ 295 struct cred *zone_kcred; /* kcred-like, zone-limited cred */ 296 /* 297 * zone_pool is protected by pool_lock(). 298 */ 299 struct pool *zone_pool; /* pool the zone is bound to */ 300 hrtime_t zone_pool_mod; /* last pool bind modification time */ 301 /* zone_psetid is protected by cpu_lock */ 302 psetid_t zone_psetid; /* pset the zone is bound to */ 303 /* 304 * The following two can be read without holding any locks. They are 305 * updated under cpu_lock. 306 */ 307 int zone_ncpus; /* zone's idea of ncpus */ 308 int zone_ncpus_online; /* zone's idea of ncpus_online */ 309 /* 310 * List of ZFS datasets exported to this zone. 311 */ 312 list_t zone_datasets; /* list of datasets */ 313 314 ts_label_t *zone_slabel; /* zone sensitivity label */ 315 int zone_match; /* require label match for packets */ 316 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */ 317 318 } zone_t; 319 320 /* 321 * Special value of zone_psetid to indicate that pools are disabled. 322 */ 323 #define ZONE_PS_INVAL PS_MYID 324 325 extern zone_t zone0; 326 extern zone_t *global_zone; 327 extern uint_t maxzones; 328 extern rctl_hndl_t rc_zone_nlwps; 329 330 extern const char * const zone_initname; 331 332 extern long zone(int, void *, void *, void *, void *); 333 extern void zone_zsd_init(void); 334 extern void zone_init(void); 335 extern void zone_hold(zone_t *); 336 extern void zone_rele(zone_t *); 337 extern void zone_cred_hold(zone_t *); 338 extern void zone_cred_rele(zone_t *); 339 extern void zone_task_hold(zone_t *); 340 extern void zone_task_rele(zone_t *); 341 extern zone_t *zone_find_by_id(zoneid_t); 342 extern zone_t *zone_find_by_label(const ts_label_t *); 343 extern zone_t *zone_find_by_name(char *); 344 extern zone_t *zone_find_by_any_path(const char *, boolean_t); 345 extern zone_t *zone_find_by_path(const char *); 346 extern zoneid_t getzoneid(void); 347 348 /* 349 * Zone-specific data (ZSD) APIs 350 */ 351 /* 352 * The following is what code should be initializing its zone_key_t to if it 353 * calls zone_getspecific() without necessarily knowing that zone_key_create() 354 * has been called on the key. 355 */ 356 #define ZONE_KEY_UNINITIALIZED 0 357 358 typedef uint_t zone_key_t; 359 360 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t), 361 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *)); 362 extern int zone_key_delete(zone_key_t); 363 extern void *zone_getspecific(zone_key_t, zone_t *); 364 extern int zone_setspecific(zone_key_t, zone_t *, const void *); 365 366 /* 367 * The definition of a zsd_entry is truly private to zone.c and is only 368 * placed here so it can be shared with mdb. 369 */ 370 struct zsd_entry { 371 zone_key_t zsd_key; /* Key used to lookup value */ 372 void *zsd_data; /* Caller-managed value */ 373 /* 374 * Callbacks to be executed when a zone is created, shutdown, and 375 * destroyed, respectively. 376 */ 377 void *(*zsd_create)(zoneid_t); 378 void (*zsd_shutdown)(zoneid_t, void *); 379 void (*zsd_destroy)(zoneid_t, void *); 380 list_node_t zsd_linkage; 381 }; 382 383 /* 384 * Macros to help with zone visibility restrictions. 385 */ 386 387 /* 388 * Is process in the global zone? 389 */ 390 #define INGLOBALZONE(p) \ 391 ((p)->p_zone == global_zone) 392 393 /* 394 * Can process view objects in given zone? 395 */ 396 #define HASZONEACCESS(p, zoneid) \ 397 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p)) 398 399 /* 400 * Convenience macro to see if a resolved path is visible from within a 401 * given zone. 402 * 403 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the 404 * two strings must be equal. Since the rootpathlen has a trailing '/', 405 * we want to skip everything in the path up to (but not including) the 406 * trailing '/'. 407 */ 408 #define ZONE_PATH_VISIBLE(path, zone) \ 409 (strncmp((path), (zone)->zone_rootpath, \ 410 (zone)->zone_rootpathlen - 1) == 0) 411 412 /* 413 * Convenience macro to go from the global view of a path to that seen 414 * from within said zone. It is the responsibility of the caller to 415 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is 416 * in fact visible from within the zone. 417 */ 418 #define ZONE_PATH_TRANSLATE(path, zone) \ 419 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \ 420 (path) + (zone)->zone_rootpathlen - 2) 421 422 /* 423 * Special processes visible in all zones. 424 */ 425 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1) 426 427 /* 428 * Zone-safe version of thread_create() to be used when the caller wants to 429 * create a kernel thread to run within the current zone's context. 430 */ 431 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t, 432 pri_t); 433 extern void zthread_exit(void); 434 435 /* 436 * Functions for an external observer to register interest in a zone's status 437 * change. Observers will be woken up when the zone status equals the status 438 * argument passed in (in the case of zone_status_timedwait, the function may 439 * also return because of a timeout; zone_status_wait_sig may return early due 440 * to a signal being delivered; zone_status_timedwait_sig may return for any of 441 * the above reasons). 442 * 443 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and 444 * cv_wait_sig() respectively. 445 */ 446 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t); 447 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t); 448 extern void zone_status_wait(zone_t *, zone_status_t); 449 extern int zone_status_wait_sig(zone_t *, zone_status_t); 450 451 /* 452 * Get the status of the zone (at the time it was called). The state may 453 * have progressed by the time it is returned. 454 */ 455 extern zone_status_t zone_status_get(zone_t *); 456 457 /* 458 * Get the "kcred" credentials corresponding to the given zone. 459 */ 460 extern struct cred *zone_get_kcred(zoneid_t); 461 462 /* 463 * Get/set the pool the zone is currently bound to. 464 */ 465 extern struct pool *zone_pool_get(zone_t *); 466 extern void zone_pool_set(zone_t *, struct pool *); 467 468 /* 469 * Get/set the pset the zone is currently using. 470 */ 471 extern psetid_t zone_pset_get(zone_t *); 472 extern void zone_pset_set(zone_t *, psetid_t); 473 474 /* 475 * Get the number of cpus/online-cpus visible from the given zone. 476 */ 477 extern int zone_ncpus_get(zone_t *); 478 extern int zone_ncpus_online_get(zone_t *); 479 480 /* 481 * Returns true if the named pool/dataset is visible in the current zone. 482 */ 483 extern int zone_dataset_visible(const char *, int *); 484 485 /* 486 * zone version of kadmin() 487 */ 488 extern int zone_kadmin(int, int, const char *, cred_t *); 489 extern void zone_shutdown_global(void); 490 491 extern void mount_in_progress(void); 492 extern void mount_completed(void); 493 494 extern int zone_walk(int (*)(zone_t *, void *), void *); 495 496 #endif /* _KERNEL */ 497 498 #ifdef __cplusplus 499 } 500 #endif 501 502 #endif /* _SYS_ZONE_H */ 503