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/ipc_rctl.h> 36 #include <sys/pset.h> 37 #include <sys/tsol/label.h> 38 #include <sys/uadmin.h> 39 40 #ifdef __cplusplus 41 extern "C" { 42 #endif 43 44 /* 45 * NOTE 46 * 47 * The contents of this file are private to the implementation of 48 * Solaris and are subject to change at any time without notice. 49 * Applications and drivers using these interfaces may fail to 50 * run on future releases. 51 */ 52 53 /* Available both in kernel and for user space */ 54 55 /* zone id restrictions and special ids */ 56 #define MAX_ZONEID 9999 57 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */ 58 #define MIN_ZONEID 0 /* minimum zone ID on system */ 59 #define GLOBAL_ZONEID 0 60 #define ZONEID_WIDTH 4 /* for printf */ 61 62 /* 63 * Special zoneid_t token to refer to all zones. 64 */ 65 #define ALL_ZONES (-1) 66 67 /* system call subcodes */ 68 #define ZONE_CREATE 0 69 #define ZONE_DESTROY 1 70 #define ZONE_GETATTR 2 71 #define ZONE_ENTER 3 72 #define ZONE_LIST 4 73 #define ZONE_SHUTDOWN 5 74 #define ZONE_LOOKUP 6 75 #define ZONE_BOOT 7 76 #define ZONE_VERSION 8 77 #define ZONE_SETATTR 9 78 79 /* zone attributes */ 80 #define ZONE_ATTR_ROOT 1 81 #define ZONE_ATTR_NAME 2 82 #define ZONE_ATTR_STATUS 3 83 #define ZONE_ATTR_PRIVSET 4 84 #define ZONE_ATTR_UNIQID 5 85 #define ZONE_ATTR_POOLID 6 86 #define ZONE_ATTR_INITPID 7 87 #define ZONE_ATTR_SLBL 8 88 #define ZONE_ATTR_INITNAME 9 89 #define ZONE_ATTR_BOOTARGS 10 90 91 #define ZONE_EVENT_CHANNEL "com.sun:zones:status" 92 #define ZONE_EVENT_STATUS_CLASS "status" 93 #define ZONE_EVENT_STATUS_SUBCLASS "change" 94 95 #define ZONE_EVENT_UNINITIALIZED "uninitialized" 96 #define ZONE_EVENT_READY "ready" 97 #define ZONE_EVENT_RUNNING "running" 98 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down" 99 100 #define ZONE_CB_NAME "zonename" 101 #define ZONE_CB_NEWSTATE "newstate" 102 #define ZONE_CB_OLDSTATE "oldstate" 103 #define ZONE_CB_TIMESTAMP "when" 104 #define ZONE_CB_ZONEID "zoneid" 105 106 #ifdef _SYSCALL32 107 typedef struct { 108 caddr32_t zone_name; 109 caddr32_t zone_root; 110 caddr32_t zone_privs; 111 size32_t zone_privssz; 112 caddr32_t rctlbuf; 113 size32_t rctlbufsz; 114 caddr32_t extended_error; 115 caddr32_t zfsbuf; 116 size32_t zfsbufsz; 117 int match; /* match level */ 118 uint32_t doi; /* DOI for label */ 119 caddr32_t label; /* label associated with zone */ 120 } zone_def32; 121 #endif 122 typedef struct { 123 const char *zone_name; 124 const char *zone_root; 125 const struct priv_set *zone_privs; 126 size_t zone_privssz; 127 const char *rctlbuf; 128 size_t rctlbufsz; 129 int *extended_error; 130 const char *zfsbuf; 131 size_t zfsbufsz; 132 int match; /* match level */ 133 uint32_t doi; /* DOI for label */ 134 const bslabel_t *label; /* label associated with zone */ 135 } zone_def; 136 137 /* extended error information */ 138 #define ZE_UNKNOWN 0 /* No extended error info */ 139 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */ 140 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */ 141 142 /* zone_status */ 143 typedef enum { 144 ZONE_IS_UNINITIALIZED = 0, 145 ZONE_IS_READY, 146 ZONE_IS_BOOTING, 147 ZONE_IS_RUNNING, 148 ZONE_IS_SHUTTING_DOWN, 149 ZONE_IS_EMPTY, 150 ZONE_IS_DOWN, 151 ZONE_IS_DYING, 152 ZONE_IS_DEAD 153 } zone_status_t; 154 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED 155 #define ZONE_MAX_STATE ZONE_IS_DEAD 156 157 /* 158 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also 159 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT. 160 */ 161 typedef enum zone_cmd { 162 Z_READY, Z_BOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING, 163 Z_MOUNT, Z_UNMOUNT 164 } zone_cmd_t; 165 166 /* 167 * The structure of a request to zoneadmd. 168 */ 169 typedef struct zone_cmd_arg { 170 uint64_t uniqid; /* unique "generation number" */ 171 zone_cmd_t cmd; /* requested action */ 172 uint32_t _pad; /* need consistent 32/64 bit alignmt */ 173 char locale[MAXPATHLEN]; /* locale in which to render messages */ 174 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */ 175 } zone_cmd_arg_t; 176 177 /* 178 * Structure of zoneadmd's response to a request. A NULL return value means 179 * the caller should attempt to restart zoneadmd and retry. 180 */ 181 typedef struct zone_cmd_rval { 182 int rval; /* return value of request */ 183 char errbuf[1]; /* variable-sized buffer containing error messages */ 184 } zone_cmd_rval_t; 185 186 /* 187 * The zone support infrastructure uses the zone name as a component 188 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters 189 * in length, so ZONENAME_MAX is limited by that. 190 */ 191 #define ZONENAME_MAX 64 192 193 #define GLOBAL_ZONENAME "global" 194 195 /* 196 * Extended Regular expression (see regex(5)) which matches all valid zone 197 * names. 198 */ 199 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}" 200 201 /* 202 * Where the zones support infrastructure places temporary files. 203 */ 204 #define ZONES_TMPDIR "/var/run/zones" 205 206 /* 207 * The path to the door used by clients to communicate with zoneadmd. 208 */ 209 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door" 210 211 #ifdef _KERNEL 212 /* 213 * We need to protect the definition of 'list_t' from userland applications and 214 * libraries which may be defining ther own versions. 215 */ 216 #include <sys/list.h> 217 218 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */ 219 220 /* zone_flags */ 221 #define ZF_DESTROYED 0x1 /* ZSD destructor callbacks run */ 222 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */ 223 #define ZF_IS_SCRATCH 0x4 /* scratch zone */ 224 225 struct pool; 226 227 /* 228 * Structure to record list of ZFS datasets exported to a zone. 229 */ 230 typedef struct zone_dataset { 231 char *zd_dataset; 232 list_node_t zd_linkage; 233 } zone_dataset_t; 234 235 typedef struct zone { 236 /* 237 * zone_name is never modified once set. 238 */ 239 char *zone_name; /* zone's configuration name */ 240 /* 241 * zone_nodename and zone_domain are never freed once allocated. 242 */ 243 char *zone_nodename; /* utsname.nodename equivalent */ 244 char *zone_domain; /* srpc_domain equivalent */ 245 /* 246 * zone_lock protects the following fields of a zone_t: 247 * zone_ref 248 * zone_cred_ref 249 * zone_ntasks 250 * zone_flags 251 * zone_zsd 252 */ 253 kmutex_t zone_lock; 254 /* 255 * zone_linkage is the zone's linkage into the active or 256 * death-row list. The field is protected by zonehash_lock. 257 */ 258 list_node_t zone_linkage; 259 zoneid_t zone_id; /* ID of zone */ 260 uint_t zone_ref; /* count of zone_hold()s on zone */ 261 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */ 262 /* 263 * zone_rootvp and zone_rootpath can never be modified once set. 264 */ 265 struct vnode *zone_rootvp; /* zone's root vnode */ 266 char *zone_rootpath; /* Path to zone's root + '/' */ 267 ushort_t zone_flags; /* misc flags */ 268 zone_status_t zone_status; /* protected by zone_status_lock */ 269 uint_t zone_ntasks; /* number of tasks executing in zone */ 270 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */ 271 /* counters in projects and tasks */ 272 /* that are within the zone */ 273 rctl_qty_t zone_nlwps; /* number of lwps in zone */ 274 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */ 275 rctl_qty_t zone_shmmax; /* System V shared memory usage */ 276 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */ 277 278 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */ 279 uint32_t zone_shares; /* FSS shares allocated to zone */ 280 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */ 281 list_t zone_zsd; /* list of Zone-Specific Data values */ 282 kcondvar_t zone_cv; /* used to signal state changes */ 283 struct proc *zone_zsched; /* Dummy kernel "zsched" process */ 284 pid_t zone_proc_initpid; /* pid of "init" for this zone */ 285 char *zone_initname; /* fs path to 'init' */ 286 int zone_boot_err; /* for zone_boot() if boot fails */ 287 char *zone_bootargs; /* arguments passed via zone_boot() */ 288 /* 289 * zone_kthreads is protected by zone_status_lock. 290 */ 291 kthread_t *zone_kthreads; /* kernel threads in zone */ 292 struct priv_set *zone_privset; /* limit set for zone */ 293 /* 294 * zone_vfslist is protected by vfs_list_lock(). 295 */ 296 struct vfs *zone_vfslist; /* list of FS's mounted in zone */ 297 uint64_t zone_uniqid; /* unique zone generation number */ 298 struct cred *zone_kcred; /* kcred-like, zone-limited cred */ 299 /* 300 * zone_pool is protected by pool_lock(). 301 */ 302 struct pool *zone_pool; /* pool the zone is bound to */ 303 hrtime_t zone_pool_mod; /* last pool bind modification time */ 304 /* zone_psetid is protected by cpu_lock */ 305 psetid_t zone_psetid; /* pset the zone is bound to */ 306 /* 307 * The following two can be read without holding any locks. They are 308 * updated under cpu_lock. 309 */ 310 int zone_ncpus; /* zone's idea of ncpus */ 311 int zone_ncpus_online; /* zone's idea of ncpus_online */ 312 /* 313 * List of ZFS datasets exported to this zone. 314 */ 315 list_t zone_datasets; /* list of datasets */ 316 317 ts_label_t *zone_slabel; /* zone sensitivity label */ 318 int zone_match; /* require label match for packets */ 319 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */ 320 321 } zone_t; 322 323 /* 324 * Special value of zone_psetid to indicate that pools are disabled. 325 */ 326 #define ZONE_PS_INVAL PS_MYID 327 328 extern zone_t zone0; 329 extern zone_t *global_zone; 330 extern uint_t maxzones; 331 extern rctl_hndl_t rc_zone_nlwps; 332 333 extern const char * const zone_initname; 334 335 extern long zone(int, void *, void *, void *, void *); 336 extern void zone_zsd_init(void); 337 extern void zone_init(void); 338 extern void zone_hold(zone_t *); 339 extern void zone_rele(zone_t *); 340 extern void zone_cred_hold(zone_t *); 341 extern void zone_cred_rele(zone_t *); 342 extern void zone_task_hold(zone_t *); 343 extern void zone_task_rele(zone_t *); 344 extern zone_t *zone_find_by_id(zoneid_t); 345 extern zone_t *zone_find_by_label(const ts_label_t *); 346 extern zone_t *zone_find_by_name(char *); 347 extern zone_t *zone_find_by_any_path(const char *, boolean_t); 348 extern zone_t *zone_find_by_path(const char *); 349 extern zoneid_t getzoneid(void); 350 351 /* 352 * Zone-specific data (ZSD) APIs 353 */ 354 /* 355 * The following is what code should be initializing its zone_key_t to if it 356 * calls zone_getspecific() without necessarily knowing that zone_key_create() 357 * has been called on the key. 358 */ 359 #define ZONE_KEY_UNINITIALIZED 0 360 361 typedef uint_t zone_key_t; 362 363 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t), 364 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *)); 365 extern int zone_key_delete(zone_key_t); 366 extern void *zone_getspecific(zone_key_t, zone_t *); 367 extern int zone_setspecific(zone_key_t, zone_t *, const void *); 368 369 /* 370 * The definition of a zsd_entry is truly private to zone.c and is only 371 * placed here so it can be shared with mdb. 372 */ 373 struct zsd_entry { 374 zone_key_t zsd_key; /* Key used to lookup value */ 375 void *zsd_data; /* Caller-managed value */ 376 /* 377 * Callbacks to be executed when a zone is created, shutdown, and 378 * destroyed, respectively. 379 */ 380 void *(*zsd_create)(zoneid_t); 381 void (*zsd_shutdown)(zoneid_t, void *); 382 void (*zsd_destroy)(zoneid_t, void *); 383 list_node_t zsd_linkage; 384 }; 385 386 /* 387 * Macros to help with zone visibility restrictions. 388 */ 389 390 /* 391 * Is process in the global zone? 392 */ 393 #define INGLOBALZONE(p) \ 394 ((p)->p_zone == global_zone) 395 396 /* 397 * Can process view objects in given zone? 398 */ 399 #define HASZONEACCESS(p, zoneid) \ 400 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p)) 401 402 /* 403 * Convenience macro to see if a resolved path is visible from within a 404 * given zone. 405 * 406 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the 407 * two strings must be equal. Since the rootpathlen has a trailing '/', 408 * we want to skip everything in the path up to (but not including) the 409 * trailing '/'. 410 */ 411 #define ZONE_PATH_VISIBLE(path, zone) \ 412 (strncmp((path), (zone)->zone_rootpath, \ 413 (zone)->zone_rootpathlen - 1) == 0) 414 415 /* 416 * Convenience macro to go from the global view of a path to that seen 417 * from within said zone. It is the responsibility of the caller to 418 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is 419 * in fact visible from within the zone. 420 */ 421 #define ZONE_PATH_TRANSLATE(path, zone) \ 422 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \ 423 (path) + (zone)->zone_rootpathlen - 2) 424 425 /* 426 * Special processes visible in all zones. 427 */ 428 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1) 429 430 /* 431 * Zone-safe version of thread_create() to be used when the caller wants to 432 * create a kernel thread to run within the current zone's context. 433 */ 434 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t, 435 pri_t); 436 extern void zthread_exit(void); 437 438 /* 439 * Functions for an external observer to register interest in a zone's status 440 * change. Observers will be woken up when the zone status equals the status 441 * argument passed in (in the case of zone_status_timedwait, the function may 442 * also return because of a timeout; zone_status_wait_sig may return early due 443 * to a signal being delivered; zone_status_timedwait_sig may return for any of 444 * the above reasons). 445 * 446 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and 447 * cv_wait_sig() respectively. 448 */ 449 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t); 450 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t); 451 extern void zone_status_wait(zone_t *, zone_status_t); 452 extern int zone_status_wait_sig(zone_t *, zone_status_t); 453 454 /* 455 * Get the status of the zone (at the time it was called). The state may 456 * have progressed by the time it is returned. 457 */ 458 extern zone_status_t zone_status_get(zone_t *); 459 460 /* 461 * Get the "kcred" credentials corresponding to the given zone. 462 */ 463 extern struct cred *zone_get_kcred(zoneid_t); 464 465 /* 466 * Get/set the pool the zone is currently bound to. 467 */ 468 extern struct pool *zone_pool_get(zone_t *); 469 extern void zone_pool_set(zone_t *, struct pool *); 470 471 /* 472 * Get/set the pset the zone is currently using. 473 */ 474 extern psetid_t zone_pset_get(zone_t *); 475 extern void zone_pset_set(zone_t *, psetid_t); 476 477 /* 478 * Get the number of cpus/online-cpus visible from the given zone. 479 */ 480 extern int zone_ncpus_get(zone_t *); 481 extern int zone_ncpus_online_get(zone_t *); 482 483 /* 484 * Returns true if the named pool/dataset is visible in the current zone. 485 */ 486 extern int zone_dataset_visible(const char *, int *); 487 488 /* 489 * zone version of kadmin() 490 */ 491 extern int zone_kadmin(int, int, const char *, cred_t *); 492 extern void zone_shutdown_global(void); 493 494 extern void mount_in_progress(void); 495 extern void mount_completed(void); 496 497 extern int zone_walk(int (*)(zone_t *, void *), void *); 498 499 #endif /* _KERNEL */ 500 501 #ifdef __cplusplus 502 } 503 #endif 504 505 #endif /* _SYS_ZONE_H */ 506