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 2007 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/cred.h> 39 #include <sys/netstack.h> 40 #include <sys/uadmin.h> 41 42 #ifdef __cplusplus 43 extern "C" { 44 #endif 45 46 /* 47 * NOTE 48 * 49 * The contents of this file are private to the implementation of 50 * Solaris and are subject to change at any time without notice. 51 * Applications and drivers using these interfaces may fail to 52 * run on future releases. 53 */ 54 55 /* Available both in kernel and for user space */ 56 57 /* zone id restrictions and special ids */ 58 #define MAX_ZONEID 9999 59 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */ 60 #define MIN_ZONEID 0 /* minimum zone ID on system */ 61 #define GLOBAL_ZONEID 0 62 #define ZONEID_WIDTH 4 /* for printf */ 63 64 /* 65 * Special zoneid_t token to refer to all zones. 66 */ 67 #define ALL_ZONES (-1) 68 69 /* system call subcodes */ 70 #define ZONE_CREATE 0 71 #define ZONE_DESTROY 1 72 #define ZONE_GETATTR 2 73 #define ZONE_ENTER 3 74 #define ZONE_LIST 4 75 #define ZONE_SHUTDOWN 5 76 #define ZONE_LOOKUP 6 77 #define ZONE_BOOT 7 78 #define ZONE_VERSION 8 79 #define ZONE_SETATTR 9 80 #define ZONE_ADD_DATALINK 10 81 #define ZONE_DEL_DATALINK 11 82 #define ZONE_CHECK_DATALINK 12 83 #define ZONE_LIST_DATALINK 13 84 85 /* zone attributes */ 86 #define ZONE_ATTR_ROOT 1 87 #define ZONE_ATTR_NAME 2 88 #define ZONE_ATTR_STATUS 3 89 #define ZONE_ATTR_PRIVSET 4 90 #define ZONE_ATTR_UNIQID 5 91 #define ZONE_ATTR_POOLID 6 92 #define ZONE_ATTR_INITPID 7 93 #define ZONE_ATTR_SLBL 8 94 #define ZONE_ATTR_INITNAME 9 95 #define ZONE_ATTR_BOOTARGS 10 96 #define ZONE_ATTR_BRAND 11 97 #define ZONE_ATTR_PHYS_MCAP 12 98 #define ZONE_ATTR_SCHED_CLASS 13 99 #define ZONE_ATTR_FLAGS 14 100 101 /* Start of the brand-specific attribute namespace */ 102 #define ZONE_ATTR_BRAND_ATTRS 32768 103 104 #define ZONE_EVENT_CHANNEL "com.sun:zones:status" 105 #define ZONE_EVENT_STATUS_CLASS "status" 106 #define ZONE_EVENT_STATUS_SUBCLASS "change" 107 108 #define ZONE_EVENT_UNINITIALIZED "uninitialized" 109 #define ZONE_EVENT_READY "ready" 110 #define ZONE_EVENT_RUNNING "running" 111 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down" 112 113 #define ZONE_CB_NAME "zonename" 114 #define ZONE_CB_NEWSTATE "newstate" 115 #define ZONE_CB_OLDSTATE "oldstate" 116 #define ZONE_CB_TIMESTAMP "when" 117 #define ZONE_CB_ZONEID "zoneid" 118 119 /* 120 * Exit values that may be returned by scripts or programs invoked by various 121 * zone commands. 122 * 123 * These are defined as: 124 * 125 * ZONE_SUBPROC_OK 126 * =============== 127 * The subprocess completed successfully. 128 * 129 * ZONE_SUBPROC_USAGE 130 * ================== 131 * The subprocess failed with a usage message, or a usage message should 132 * be output in its behalf. 133 * 134 * ZONE_SUBPROC_NOTCOMPLETE 135 * ======================== 136 * The subprocess did not complete, but the actions performed by the 137 * subprocess require no recovery actions by the user. 138 * 139 * For example, if the subprocess were called by "zoneadm install," the 140 * installation of the zone did not succeed but the user need not perform 141 * a "zoneadm uninstall" before attempting another install. 142 * 143 * ZONE_SUBPROC_FATAL 144 * ================== 145 * The subprocess failed in a fatal manner, usually one that will require 146 * some type of recovery action by the user. 147 * 148 * For example, if the subprocess were called by "zoneadm install," the 149 * installation of the zone did not succeed and the user will need to 150 * perform a "zoneadm uninstall" before another install attempt is 151 * possible. 152 * 153 * The non-success exit values are large to avoid accidental collision 154 * with values used internally by some commands (e.g. "Z_ERR" and 155 * "Z_USAGE" as used by zoneadm.) 156 */ 157 #define ZONE_SUBPROC_OK 0 158 #define ZONE_SUBPROC_USAGE 253 159 #define ZONE_SUBPROC_NOTCOMPLETE 254 160 #define ZONE_SUBPROC_FATAL 255 161 162 #ifdef _SYSCALL32 163 typedef struct { 164 caddr32_t zone_name; 165 caddr32_t zone_root; 166 caddr32_t zone_privs; 167 size32_t zone_privssz; 168 caddr32_t rctlbuf; 169 size32_t rctlbufsz; 170 caddr32_t extended_error; 171 caddr32_t zfsbuf; 172 size32_t zfsbufsz; 173 int match; /* match level */ 174 uint32_t doi; /* DOI for label */ 175 caddr32_t label; /* label associated with zone */ 176 int flags; 177 } zone_def32; 178 #endif 179 typedef struct { 180 const char *zone_name; 181 const char *zone_root; 182 const struct priv_set *zone_privs; 183 size_t zone_privssz; 184 const char *rctlbuf; 185 size_t rctlbufsz; 186 int *extended_error; 187 const char *zfsbuf; 188 size_t zfsbufsz; 189 int match; /* match level */ 190 uint32_t doi; /* DOI for label */ 191 const bslabel_t *label; /* label associated with zone */ 192 int flags; 193 } zone_def; 194 195 /* extended error information */ 196 #define ZE_UNKNOWN 0 /* No extended error info */ 197 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */ 198 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */ 199 200 /* zone_status */ 201 typedef enum { 202 ZONE_IS_UNINITIALIZED = 0, 203 ZONE_IS_READY, 204 ZONE_IS_BOOTING, 205 ZONE_IS_RUNNING, 206 ZONE_IS_SHUTTING_DOWN, 207 ZONE_IS_EMPTY, 208 ZONE_IS_DOWN, 209 ZONE_IS_DYING, 210 ZONE_IS_DEAD 211 } zone_status_t; 212 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED 213 #define ZONE_MAX_STATE ZONE_IS_DEAD 214 215 /* 216 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also 217 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT. 218 */ 219 typedef enum zone_cmd { 220 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING, 221 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT 222 } zone_cmd_t; 223 224 /* 225 * The structure of a request to zoneadmd. 226 */ 227 typedef struct zone_cmd_arg { 228 uint64_t uniqid; /* unique "generation number" */ 229 zone_cmd_t cmd; /* requested action */ 230 uint32_t _pad; /* need consistent 32/64 bit alignmt */ 231 char locale[MAXPATHLEN]; /* locale in which to render messages */ 232 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */ 233 } zone_cmd_arg_t; 234 235 /* 236 * Structure of zoneadmd's response to a request. A NULL return value means 237 * the caller should attempt to restart zoneadmd and retry. 238 */ 239 typedef struct zone_cmd_rval { 240 int rval; /* return value of request */ 241 char errbuf[1]; /* variable-sized buffer containing error messages */ 242 } zone_cmd_rval_t; 243 244 /* 245 * The zone support infrastructure uses the zone name as a component 246 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters 247 * in length, so ZONENAME_MAX is limited by that. 248 */ 249 #define ZONENAME_MAX 64 250 251 #define GLOBAL_ZONENAME "global" 252 253 /* 254 * Extended Regular expression (see regex(5)) which matches all valid zone 255 * names. 256 */ 257 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}" 258 259 /* 260 * Where the zones support infrastructure places temporary files. 261 */ 262 #define ZONES_TMPDIR "/var/run/zones" 263 264 /* 265 * The path to the door used by clients to communicate with zoneadmd. 266 */ 267 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door" 268 269 /* zone_flags */ 270 #define ZF_DESTROYED 0x1 /* ZSD destructor callbacks run */ 271 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */ 272 #define ZF_IS_SCRATCH 0x4 /* scratch zone */ 273 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */ 274 275 /* zone_create flags */ 276 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */ 277 278 #ifdef _KERNEL 279 /* 280 * We need to protect the definition of 'list_t' from userland applications and 281 * libraries which may be defining ther own versions. 282 */ 283 #include <sys/list.h> 284 285 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */ 286 287 struct pool; 288 struct brand; 289 struct dlnamelist; 290 291 /* 292 * Structure to record list of ZFS datasets exported to a zone. 293 */ 294 typedef struct zone_dataset { 295 char *zd_dataset; 296 list_node_t zd_linkage; 297 } zone_dataset_t; 298 299 /* 300 * structure for zone kstats 301 */ 302 typedef struct zone_kstat { 303 kstat_named_t zk_zonename; 304 kstat_named_t zk_usage; 305 kstat_named_t zk_value; 306 } zone_kstat_t; 307 308 typedef struct zone { 309 /* 310 * zone_name is never modified once set. 311 */ 312 char *zone_name; /* zone's configuration name */ 313 /* 314 * zone_nodename and zone_domain are never freed once allocated. 315 */ 316 char *zone_nodename; /* utsname.nodename equivalent */ 317 char *zone_domain; /* srpc_domain equivalent */ 318 /* 319 * zone_lock protects the following fields of a zone_t: 320 * zone_ref 321 * zone_cred_ref 322 * zone_ntasks 323 * zone_flags 324 * zone_zsd 325 */ 326 kmutex_t zone_lock; 327 /* 328 * zone_linkage is the zone's linkage into the active or 329 * death-row list. The field is protected by zonehash_lock. 330 */ 331 list_node_t zone_linkage; 332 zoneid_t zone_id; /* ID of zone */ 333 uint_t zone_ref; /* count of zone_hold()s on zone */ 334 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */ 335 /* 336 * zone_rootvp and zone_rootpath can never be modified once set. 337 */ 338 struct vnode *zone_rootvp; /* zone's root vnode */ 339 char *zone_rootpath; /* Path to zone's root + '/' */ 340 ushort_t zone_flags; /* misc flags */ 341 zone_status_t zone_status; /* protected by zone_status_lock */ 342 uint_t zone_ntasks; /* number of tasks executing in zone */ 343 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */ 344 /* counters in projects and tasks */ 345 /* that are within the zone */ 346 rctl_qty_t zone_nlwps; /* number of lwps in zone */ 347 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */ 348 rctl_qty_t zone_shmmax; /* System V shared memory usage */ 349 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */ 350 351 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */ 352 uint32_t zone_shares; /* FSS shares allocated to zone */ 353 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */ 354 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */ 355 /* kpd_locked_mem for all */ 356 /* projects in zone. */ 357 /* Also protects zone_max_swap */ 358 /* grab after p_lock, before rcs_lock */ 359 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */ 360 /* zone */ 361 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */ 362 /* limit. Protected by */ 363 /* zone_rctls->rcs_lock */ 364 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */ 365 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */ 366 /* Protected by */ 367 /* zone_rctls->rcs_lock */ 368 list_t zone_zsd; /* list of Zone-Specific Data values */ 369 kcondvar_t zone_cv; /* used to signal state changes */ 370 struct proc *zone_zsched; /* Dummy kernel "zsched" process */ 371 pid_t zone_proc_initpid; /* pid of "init" for this zone */ 372 char *zone_initname; /* fs path to 'init' */ 373 int zone_boot_err; /* for zone_boot() if boot fails */ 374 char *zone_bootargs; /* arguments passed via zone_boot() */ 375 uint64_t zone_phys_mcap; /* physical memory cap */ 376 /* 377 * zone_kthreads is protected by zone_status_lock. 378 */ 379 kthread_t *zone_kthreads; /* kernel threads in zone */ 380 struct priv_set *zone_privset; /* limit set for zone */ 381 /* 382 * zone_vfslist is protected by vfs_list_lock(). 383 */ 384 struct vfs *zone_vfslist; /* list of FS's mounted in zone */ 385 uint64_t zone_uniqid; /* unique zone generation number */ 386 struct cred *zone_kcred; /* kcred-like, zone-limited cred */ 387 /* 388 * zone_pool is protected by pool_lock(). 389 */ 390 struct pool *zone_pool; /* pool the zone is bound to */ 391 hrtime_t zone_pool_mod; /* last pool bind modification time */ 392 /* zone_psetid is protected by cpu_lock */ 393 psetid_t zone_psetid; /* pset the zone is bound to */ 394 /* 395 * The following two can be read without holding any locks. They are 396 * updated under cpu_lock. 397 */ 398 int zone_ncpus; /* zone's idea of ncpus */ 399 int zone_ncpus_online; /* zone's idea of ncpus_online */ 400 /* 401 * List of ZFS datasets exported to this zone. 402 */ 403 list_t zone_datasets; /* list of datasets */ 404 405 ts_label_t *zone_slabel; /* zone sensitivity label */ 406 int zone_match; /* require label match for packets */ 407 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */ 408 409 boolean_t zone_restart_init; /* Restart init if it dies? */ 410 struct brand *zone_brand; /* zone's brand */ 411 id_t zone_defaultcid; /* dflt scheduling class id */ 412 kstat_t *zone_swapresv_kstat; 413 kstat_t *zone_lockedmem_kstat; 414 /* 415 * zone_dl_list is protected by zone_lock 416 */ 417 struct dlnamelist *zone_dl_list; 418 netstack_t *zone_netstack; 419 } zone_t; 420 421 /* 422 * Special value of zone_psetid to indicate that pools are disabled. 423 */ 424 #define ZONE_PS_INVAL PS_MYID 425 426 427 extern zone_t zone0; 428 extern zone_t *global_zone; 429 extern uint_t maxzones; 430 extern rctl_hndl_t rc_zone_nlwps; 431 432 extern long zone(int, void *, void *, void *, void *); 433 extern void zone_zsd_init(void); 434 extern void zone_init(void); 435 extern void zone_hold(zone_t *); 436 extern void zone_rele(zone_t *); 437 extern void zone_cred_hold(zone_t *); 438 extern void zone_cred_rele(zone_t *); 439 extern void zone_task_hold(zone_t *); 440 extern void zone_task_rele(zone_t *); 441 extern zone_t *zone_find_by_id(zoneid_t); 442 extern zone_t *zone_find_by_label(const ts_label_t *); 443 extern zone_t *zone_find_by_name(char *); 444 extern zone_t *zone_find_by_any_path(const char *, boolean_t); 445 extern zone_t *zone_find_by_path(const char *); 446 extern zoneid_t getzoneid(void); 447 extern zone_t *zone_find_by_id_nolock(zoneid_t); 448 449 /* 450 * Zone-specific data (ZSD) APIs 451 */ 452 /* 453 * The following is what code should be initializing its zone_key_t to if it 454 * calls zone_getspecific() without necessarily knowing that zone_key_create() 455 * has been called on the key. 456 */ 457 #define ZONE_KEY_UNINITIALIZED 0 458 459 typedef uint_t zone_key_t; 460 461 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t), 462 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *)); 463 extern int zone_key_delete(zone_key_t); 464 extern void *zone_getspecific(zone_key_t, zone_t *); 465 extern int zone_setspecific(zone_key_t, zone_t *, const void *); 466 467 /* 468 * The definition of a zsd_entry is truly private to zone.c and is only 469 * placed here so it can be shared with mdb. 470 */ 471 struct zsd_entry { 472 zone_key_t zsd_key; /* Key used to lookup value */ 473 void *zsd_data; /* Caller-managed value */ 474 /* 475 * Callbacks to be executed when a zone is created, shutdown, and 476 * destroyed, respectively. 477 */ 478 void *(*zsd_create)(zoneid_t); 479 void (*zsd_shutdown)(zoneid_t, void *); 480 void (*zsd_destroy)(zoneid_t, void *); 481 list_node_t zsd_linkage; 482 }; 483 484 /* 485 * Macros to help with zone visibility restrictions. 486 */ 487 488 /* 489 * Is process in the global zone? 490 */ 491 #define INGLOBALZONE(p) \ 492 ((p)->p_zone == global_zone) 493 494 /* 495 * Can process view objects in given zone? 496 */ 497 #define HASZONEACCESS(p, zoneid) \ 498 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p)) 499 500 /* 501 * Convenience macro to see if a resolved path is visible from within a 502 * given zone. 503 * 504 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the 505 * two strings must be equal. Since the rootpathlen has a trailing '/', 506 * we want to skip everything in the path up to (but not including) the 507 * trailing '/'. 508 */ 509 #define ZONE_PATH_VISIBLE(path, zone) \ 510 (strncmp((path), (zone)->zone_rootpath, \ 511 (zone)->zone_rootpathlen - 1) == 0) 512 513 /* 514 * Convenience macro to go from the global view of a path to that seen 515 * from within said zone. It is the responsibility of the caller to 516 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is 517 * in fact visible from within the zone. 518 */ 519 #define ZONE_PATH_TRANSLATE(path, zone) \ 520 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \ 521 (path) + (zone)->zone_rootpathlen - 2) 522 523 /* 524 * Special processes visible in all zones. 525 */ 526 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1) 527 528 /* 529 * Zone-safe version of thread_create() to be used when the caller wants to 530 * create a kernel thread to run within the current zone's context. 531 */ 532 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t, 533 pri_t); 534 extern void zthread_exit(void); 535 536 /* 537 * Functions for an external observer to register interest in a zone's status 538 * change. Observers will be woken up when the zone status equals the status 539 * argument passed in (in the case of zone_status_timedwait, the function may 540 * also return because of a timeout; zone_status_wait_sig may return early due 541 * to a signal being delivered; zone_status_timedwait_sig may return for any of 542 * the above reasons). 543 * 544 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and 545 * cv_wait_sig() respectively. 546 */ 547 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t); 548 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t); 549 extern void zone_status_wait(zone_t *, zone_status_t); 550 extern int zone_status_wait_sig(zone_t *, zone_status_t); 551 552 /* 553 * Get the status of the zone (at the time it was called). The state may 554 * have progressed by the time it is returned. 555 */ 556 extern zone_status_t zone_status_get(zone_t *); 557 558 /* 559 * Get the "kcred" credentials corresponding to the given zone. 560 */ 561 extern struct cred *zone_get_kcred(zoneid_t); 562 563 /* 564 * Get/set the pool the zone is currently bound to. 565 */ 566 extern struct pool *zone_pool_get(zone_t *); 567 extern void zone_pool_set(zone_t *, struct pool *); 568 569 /* 570 * Get/set the pset the zone is currently using. 571 */ 572 extern psetid_t zone_pset_get(zone_t *); 573 extern void zone_pset_set(zone_t *, psetid_t); 574 575 /* 576 * Get the number of cpus/online-cpus visible from the given zone. 577 */ 578 extern int zone_ncpus_get(zone_t *); 579 extern int zone_ncpus_online_get(zone_t *); 580 581 /* 582 * Returns true if the named pool/dataset is visible in the current zone. 583 */ 584 extern int zone_dataset_visible(const char *, int *); 585 586 /* 587 * zone version of kadmin() 588 */ 589 extern int zone_kadmin(int, int, const char *, cred_t *); 590 extern void zone_shutdown_global(void); 591 592 extern void mount_in_progress(void); 593 extern void mount_completed(void); 594 595 extern int zone_walk(int (*)(zone_t *, void *), void *); 596 597 extern rctl_hndl_t rc_zone_locked_mem; 598 extern rctl_hndl_t rc_zone_max_swap; 599 600 #endif /* _KERNEL */ 601 602 #ifdef __cplusplus 603 } 604 #endif 605 606 #endif /* _SYS_ZONE_H */ 607