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 (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved. 23 * Copyright 2014 Nexenta Systems, Inc. All rights reserved. 24 */ 25 26 #ifndef _SYS_ZONE_H 27 #define _SYS_ZONE_H 28 29 #include <sys/types.h> 30 #include <sys/mutex.h> 31 #include <sys/param.h> 32 #include <sys/rctl.h> 33 #include <sys/ipc_rctl.h> 34 #include <sys/pset.h> 35 #include <sys/tsol/label.h> 36 #include <sys/cred.h> 37 #include <sys/netstack.h> 38 #include <sys/uadmin.h> 39 #include <sys/ksynch.h> 40 #include <sys/socket_impl.h> 41 #include <netinet/in.h> 42 43 #ifdef __cplusplus 44 extern "C" { 45 #endif 46 47 /* 48 * NOTE 49 * 50 * The contents of this file are private to the implementation of 51 * Solaris and are subject to change at any time without notice. 52 * Applications and drivers using these interfaces may fail to 53 * run on future releases. 54 */ 55 56 /* Available both in kernel and for user space */ 57 58 /* zone id restrictions and special ids */ 59 #define MAX_ZONEID 9999 60 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */ 61 #define MIN_ZONEID 0 /* minimum zone ID on system */ 62 #define GLOBAL_ZONEID 0 63 #define ZONEID_WIDTH 4 /* for printf */ 64 65 /* 66 * Special zoneid_t token to refer to all zones. 67 */ 68 #define ALL_ZONES (-1) 69 70 /* system call subcodes */ 71 #define ZONE_CREATE 0 72 #define ZONE_DESTROY 1 73 #define ZONE_GETATTR 2 74 #define ZONE_ENTER 3 75 #define ZONE_LIST 4 76 #define ZONE_SHUTDOWN 5 77 #define ZONE_LOOKUP 6 78 #define ZONE_BOOT 7 79 #define ZONE_VERSION 8 80 #define ZONE_SETATTR 9 81 #define ZONE_ADD_DATALINK 10 82 #define ZONE_DEL_DATALINK 11 83 #define ZONE_CHECK_DATALINK 12 84 #define ZONE_LIST_DATALINK 13 85 86 /* zone attributes */ 87 #define ZONE_ATTR_ROOT 1 88 #define ZONE_ATTR_NAME 2 89 #define ZONE_ATTR_STATUS 3 90 #define ZONE_ATTR_PRIVSET 4 91 #define ZONE_ATTR_UNIQID 5 92 #define ZONE_ATTR_POOLID 6 93 #define ZONE_ATTR_INITPID 7 94 #define ZONE_ATTR_SLBL 8 95 #define ZONE_ATTR_INITNAME 9 96 #define ZONE_ATTR_BOOTARGS 10 97 #define ZONE_ATTR_BRAND 11 98 #define ZONE_ATTR_PHYS_MCAP 12 99 #define ZONE_ATTR_SCHED_CLASS 13 100 #define ZONE_ATTR_FLAGS 14 101 #define ZONE_ATTR_HOSTID 15 102 #define ZONE_ATTR_FS_ALLOWED 16 103 #define ZONE_ATTR_NETWORK 17 104 105 /* Start of the brand-specific attribute namespace */ 106 #define ZONE_ATTR_BRAND_ATTRS 32768 107 108 #define ZONE_FS_ALLOWED_MAX 1024 109 110 #define ZONE_EVENT_CHANNEL "com.sun:zones:status" 111 #define ZONE_EVENT_STATUS_CLASS "status" 112 #define ZONE_EVENT_STATUS_SUBCLASS "change" 113 114 #define ZONE_EVENT_UNINITIALIZED "uninitialized" 115 #define ZONE_EVENT_INITIALIZED "initialized" 116 #define ZONE_EVENT_READY "ready" 117 #define ZONE_EVENT_RUNNING "running" 118 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down" 119 120 #define ZONE_CB_NAME "zonename" 121 #define ZONE_CB_NEWSTATE "newstate" 122 #define ZONE_CB_OLDSTATE "oldstate" 123 #define ZONE_CB_TIMESTAMP "when" 124 #define ZONE_CB_ZONEID "zoneid" 125 126 /* 127 * Exit values that may be returned by scripts or programs invoked by various 128 * zone commands. 129 * 130 * These are defined as: 131 * 132 * ZONE_SUBPROC_OK 133 * =============== 134 * The subprocess completed successfully. 135 * 136 * ZONE_SUBPROC_USAGE 137 * ================== 138 * The subprocess failed with a usage message, or a usage message should 139 * be output in its behalf. 140 * 141 * ZONE_SUBPROC_NOTCOMPLETE 142 * ======================== 143 * The subprocess did not complete, but the actions performed by the 144 * subprocess require no recovery actions by the user. 145 * 146 * For example, if the subprocess were called by "zoneadm install," the 147 * installation of the zone did not succeed but the user need not perform 148 * a "zoneadm uninstall" before attempting another install. 149 * 150 * ZONE_SUBPROC_FATAL 151 * ================== 152 * The subprocess failed in a fatal manner, usually one that will require 153 * some type of recovery action by the user. 154 * 155 * For example, if the subprocess were called by "zoneadm install," the 156 * installation of the zone did not succeed and the user will need to 157 * perform a "zoneadm uninstall" before another install attempt is 158 * possible. 159 * 160 * The non-success exit values are large to avoid accidental collision 161 * with values used internally by some commands (e.g. "Z_ERR" and 162 * "Z_USAGE" as used by zoneadm.) 163 */ 164 #define ZONE_SUBPROC_OK 0 165 #define ZONE_SUBPROC_USAGE 253 166 #define ZONE_SUBPROC_NOTCOMPLETE 254 167 #define ZONE_SUBPROC_FATAL 255 168 169 #ifdef _SYSCALL32 170 typedef struct { 171 caddr32_t zone_name; 172 caddr32_t zone_root; 173 caddr32_t zone_privs; 174 size32_t zone_privssz; 175 caddr32_t rctlbuf; 176 size32_t rctlbufsz; 177 caddr32_t extended_error; 178 caddr32_t zfsbuf; 179 size32_t zfsbufsz; 180 int match; /* match level */ 181 uint32_t doi; /* DOI for label */ 182 caddr32_t label; /* label associated with zone */ 183 int flags; 184 } zone_def32; 185 #endif 186 typedef struct { 187 const char *zone_name; 188 const char *zone_root; 189 const struct priv_set *zone_privs; 190 size_t zone_privssz; 191 const char *rctlbuf; 192 size_t rctlbufsz; 193 int *extended_error; 194 const char *zfsbuf; 195 size_t zfsbufsz; 196 int match; /* match level */ 197 uint32_t doi; /* DOI for label */ 198 const bslabel_t *label; /* label associated with zone */ 199 int flags; 200 } zone_def; 201 202 /* extended error information */ 203 #define ZE_UNKNOWN 0 /* No extended error info */ 204 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */ 205 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */ 206 #define ZE_LABELINUSE 3 /* label is already in use by some other zone */ 207 208 /* 209 * zone_status values 210 * 211 * You must modify zone_status_names in mdb(1M)'s genunix module 212 * (genunix/zone.c) when you modify this enum. 213 */ 214 typedef enum { 215 ZONE_IS_UNINITIALIZED = 0, 216 ZONE_IS_INITIALIZED, 217 ZONE_IS_READY, 218 ZONE_IS_BOOTING, 219 ZONE_IS_RUNNING, 220 ZONE_IS_SHUTTING_DOWN, 221 ZONE_IS_EMPTY, 222 ZONE_IS_DOWN, 223 ZONE_IS_DYING, 224 ZONE_IS_DEAD 225 } zone_status_t; 226 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED 227 #define ZONE_MAX_STATE ZONE_IS_DEAD 228 229 /* 230 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also 231 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT. 232 */ 233 typedef enum zone_cmd { 234 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING, 235 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT, Z_SHUTDOWN 236 } zone_cmd_t; 237 238 /* 239 * The structure of a request to zoneadmd. 240 */ 241 typedef struct zone_cmd_arg { 242 uint64_t uniqid; /* unique "generation number" */ 243 zone_cmd_t cmd; /* requested action */ 244 uint32_t _pad; /* need consistent 32/64 bit alignmt */ 245 char locale[MAXPATHLEN]; /* locale in which to render messages */ 246 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */ 247 } zone_cmd_arg_t; 248 249 /* 250 * Structure of zoneadmd's response to a request. A NULL return value means 251 * the caller should attempt to restart zoneadmd and retry. 252 */ 253 typedef struct zone_cmd_rval { 254 int rval; /* return value of request */ 255 char errbuf[1]; /* variable-sized buffer containing error messages */ 256 } zone_cmd_rval_t; 257 258 /* 259 * The zone support infrastructure uses the zone name as a component 260 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters 261 * in length, so ZONENAME_MAX is limited by that. 262 */ 263 #define ZONENAME_MAX 64 264 265 #define GLOBAL_ZONENAME "global" 266 267 /* 268 * Extended Regular expression (see regex(5)) which matches all valid zone 269 * names. 270 */ 271 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}" 272 273 /* 274 * Where the zones support infrastructure places temporary files. 275 */ 276 #define ZONES_TMPDIR "/var/run/zones" 277 278 /* 279 * The path to the door used by clients to communicate with zoneadmd. 280 */ 281 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door" 282 283 284 /* zone_flags */ 285 /* 286 * Threads that read or write the following flag must hold zone_lock. 287 */ 288 #define ZF_REFCOUNTS_LOGGED 0x1 /* a thread logged the zone's refs */ 289 290 /* 291 * The following threads are set when the zone is created and never changed. 292 * Threads that test for these flags don't have to hold zone_lock. 293 */ 294 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */ 295 #define ZF_IS_SCRATCH 0x4 /* scratch zone */ 296 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */ 297 298 299 /* zone_create flags */ 300 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */ 301 302 /* zone network properties */ 303 #define ZONE_NETWORK_ADDRESS 1 304 #define ZONE_NETWORK_DEFROUTER 2 305 306 #define ZONE_NET_ADDRNAME "address" 307 #define ZONE_NET_RTRNAME "route" 308 309 typedef struct zone_net_data { 310 int zn_type; 311 int zn_len; 312 datalink_id_t zn_linkid; 313 uint8_t zn_val[1]; 314 } zone_net_data_t; 315 316 317 #ifdef _KERNEL 318 319 /* 320 * We need to protect the definition of 'list_t' from userland applications and 321 * libraries which may be defining ther own versions. 322 */ 323 #include <sys/list.h> 324 325 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */ 326 327 struct pool; 328 struct brand; 329 330 /* 331 * Each of these constants identifies a kernel subsystem that acquires and 332 * releases zone references. Each subsystem that invokes 333 * zone_hold_ref() and zone_rele_ref() should specify the 334 * zone_ref_subsys_t constant associated with the subsystem. Tracked holds 335 * help users and developers quickly identify subsystems that stall zone 336 * shutdowns indefinitely. 337 * 338 * NOTE: You must modify zone_ref_subsys_names in usr/src/uts/common/os/zone.c 339 * when you modify this enumeration. 340 */ 341 typedef enum zone_ref_subsys { 342 ZONE_REF_NFS, /* NFS */ 343 ZONE_REF_NFSV4, /* NFSv4 */ 344 ZONE_REF_SMBFS, /* SMBFS */ 345 ZONE_REF_MNTFS, /* MNTFS */ 346 ZONE_REF_LOFI, /* LOFI devices */ 347 ZONE_REF_VFS, /* VFS infrastructure */ 348 ZONE_REF_IPC, /* IPC infrastructure */ 349 ZONE_REF_NUM_SUBSYS /* This must be the last entry. */ 350 } zone_ref_subsys_t; 351 352 /* 353 * zone_ref represents a general-purpose references to a zone. Each zone's 354 * references are linked into the zone's zone_t::zone_ref_list. This allows 355 * debuggers to walk zones' references. 356 */ 357 typedef struct zone_ref { 358 struct zone *zref_zone; /* the zone to which the reference refers */ 359 list_node_t zref_linkage; /* linkage for zone_t::zone_ref_list */ 360 } zone_ref_t; 361 362 /* 363 * Structure to record list of ZFS datasets exported to a zone. 364 */ 365 typedef struct zone_dataset { 366 char *zd_dataset; 367 list_node_t zd_linkage; 368 } zone_dataset_t; 369 370 /* 371 * structure for zone kstats 372 */ 373 typedef struct zone_kstat { 374 kstat_named_t zk_zonename; 375 kstat_named_t zk_usage; 376 kstat_named_t zk_value; 377 } zone_kstat_t; 378 379 struct cpucap; 380 381 typedef struct zone { 382 /* 383 * zone_name is never modified once set. 384 */ 385 char *zone_name; /* zone's configuration name */ 386 /* 387 * zone_nodename and zone_domain are never freed once allocated. 388 */ 389 char *zone_nodename; /* utsname.nodename equivalent */ 390 char *zone_domain; /* srpc_domain equivalent */ 391 /* 392 * zone_hostid is used for per-zone hostid emulation. 393 * Currently it isn't modified after it's set (so no locks protect 394 * accesses), but that might have to change when we allow 395 * administrators to change running zones' properties. 396 * 397 * The global zone's zone_hostid must always be HW_INVALID_HOSTID so 398 * that zone_get_hostid() will function correctly. 399 */ 400 uint32_t zone_hostid; /* zone's hostid, HW_INVALID_HOSTID */ 401 /* if not emulated */ 402 /* 403 * zone_lock protects the following fields of a zone_t: 404 * zone_ref 405 * zone_cred_ref 406 * zone_subsys_ref 407 * zone_ref_list 408 * zone_ntasks 409 * zone_flags 410 * zone_zsd 411 * zone_pfexecd 412 */ 413 kmutex_t zone_lock; 414 /* 415 * zone_linkage is the zone's linkage into the active or 416 * death-row list. The field is protected by zonehash_lock. 417 */ 418 list_node_t zone_linkage; 419 zoneid_t zone_id; /* ID of zone */ 420 uint_t zone_ref; /* count of zone_hold()s on zone */ 421 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */ 422 /* 423 * Fixed-sized array of subsystem-specific reference counts 424 * The sum of all of the counts must be less than or equal to zone_ref. 425 * The array is indexed by the counts' subsystems' zone_ref_subsys_t 426 * constants. 427 */ 428 uint_t zone_subsys_ref[ZONE_REF_NUM_SUBSYS]; 429 list_t zone_ref_list; /* list of zone_ref_t structs */ 430 /* 431 * zone_rootvp and zone_rootpath can never be modified once set. 432 */ 433 struct vnode *zone_rootvp; /* zone's root vnode */ 434 char *zone_rootpath; /* Path to zone's root + '/' */ 435 ushort_t zone_flags; /* misc flags */ 436 zone_status_t zone_status; /* protected by zone_status_lock */ 437 uint_t zone_ntasks; /* number of tasks executing in zone */ 438 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */ 439 /* counters in projects and tasks */ 440 /* that are within the zone */ 441 rctl_qty_t zone_nlwps; /* number of lwps in zone */ 442 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */ 443 rctl_qty_t zone_shmmax; /* System V shared memory usage */ 444 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */ 445 446 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */ 447 uint32_t zone_shares; /* FSS shares allocated to zone */ 448 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */ 449 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */ 450 /* kpd_locked_mem for all */ 451 /* projects in zone. */ 452 /* Also protects zone_max_swap */ 453 /* grab after p_lock, before rcs_lock */ 454 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */ 455 /* zone */ 456 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */ 457 /* limit. Protected by */ 458 /* zone_rctls->rcs_lock */ 459 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */ 460 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */ 461 /* Protected by */ 462 /* zone_rctls->rcs_lock */ 463 kmutex_t zone_rctl_lock; /* protects zone_max_lofi */ 464 rctl_qty_t zone_max_lofi; /* lofi devs for zone */ 465 rctl_qty_t zone_max_lofi_ctl; /* current lofi limit. */ 466 /* Protected by */ 467 /* zone_rctls->rcs_lock */ 468 list_t zone_zsd; /* list of Zone-Specific Data values */ 469 kcondvar_t zone_cv; /* used to signal state changes */ 470 struct proc *zone_zsched; /* Dummy kernel "zsched" process */ 471 pid_t zone_proc_initpid; /* pid of "init" for this zone */ 472 char *zone_initname; /* fs path to 'init' */ 473 int zone_boot_err; /* for zone_boot() if boot fails */ 474 char *zone_bootargs; /* arguments passed via zone_boot() */ 475 uint64_t zone_phys_mcap; /* physical memory cap */ 476 /* 477 * zone_kthreads is protected by zone_status_lock. 478 */ 479 kthread_t *zone_kthreads; /* kernel threads in zone */ 480 struct priv_set *zone_privset; /* limit set for zone */ 481 /* 482 * zone_vfslist is protected by vfs_list_lock(). 483 */ 484 struct vfs *zone_vfslist; /* list of FS's mounted in zone */ 485 uint64_t zone_uniqid; /* unique zone generation number */ 486 struct cred *zone_kcred; /* kcred-like, zone-limited cred */ 487 /* 488 * zone_pool is protected by pool_lock(). 489 */ 490 struct pool *zone_pool; /* pool the zone is bound to */ 491 hrtime_t zone_pool_mod; /* last pool bind modification time */ 492 /* zone_psetid is protected by cpu_lock */ 493 psetid_t zone_psetid; /* pset the zone is bound to */ 494 495 time_t zone_boot_time; /* Similar to boot_time */ 496 497 /* 498 * The following two can be read without holding any locks. They are 499 * updated under cpu_lock. 500 */ 501 int zone_ncpus; /* zone's idea of ncpus */ 502 int zone_ncpus_online; /* zone's idea of ncpus_online */ 503 /* 504 * List of ZFS datasets exported to this zone. 505 */ 506 list_t zone_datasets; /* list of datasets */ 507 508 ts_label_t *zone_slabel; /* zone sensitivity label */ 509 int zone_match; /* require label match for packets */ 510 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */ 511 512 boolean_t zone_restart_init; /* Restart init if it dies? */ 513 struct brand *zone_brand; /* zone's brand */ 514 void *zone_brand_data; /* store brand specific data */ 515 id_t zone_defaultcid; /* dflt scheduling class id */ 516 kstat_t *zone_swapresv_kstat; 517 kstat_t *zone_lockedmem_kstat; 518 /* 519 * zone_dl_list is protected by zone_lock 520 */ 521 list_t zone_dl_list; 522 netstack_t *zone_netstack; 523 struct cpucap *zone_cpucap; /* CPU caps data */ 524 /* 525 * Solaris Auditing per-zone audit context 526 */ 527 struct au_kcontext *zone_audit_kctxt; 528 /* 529 * For private use by mntfs. 530 */ 531 struct mntelem *zone_mntfs_db; 532 krwlock_t zone_mntfs_db_lock; 533 534 struct klpd_reg *zone_pfexecd; 535 536 char *zone_fs_allowed; 537 rctl_qty_t zone_nprocs; /* number of processes in the zone */ 538 rctl_qty_t zone_nprocs_ctl; /* current limit protected by */ 539 /* zone_rctls->rcs_lock */ 540 kstat_t *zone_nprocs_kstat; 541 542 /* 543 * DTrace-private per-zone state 544 */ 545 int zone_dtrace_getf; /* # of unprivileged getf()s */ 546 } zone_t; 547 548 /* 549 * Special value of zone_psetid to indicate that pools are disabled. 550 */ 551 #define ZONE_PS_INVAL PS_MYID 552 553 554 extern zone_t zone0; 555 extern zone_t *global_zone; 556 extern uint_t maxzones; 557 extern rctl_hndl_t rc_zone_nlwps; 558 extern rctl_hndl_t rc_zone_nprocs; 559 560 extern long zone(int, void *, void *, void *, void *); 561 extern void zone_zsd_init(void); 562 extern void zone_init(void); 563 extern void zone_hold(zone_t *); 564 extern void zone_rele(zone_t *); 565 extern void zone_init_ref(zone_ref_t *); 566 extern void zone_hold_ref(zone_t *, zone_ref_t *, zone_ref_subsys_t); 567 extern void zone_rele_ref(zone_ref_t *, zone_ref_subsys_t); 568 extern void zone_cred_hold(zone_t *); 569 extern void zone_cred_rele(zone_t *); 570 extern void zone_task_hold(zone_t *); 571 extern void zone_task_rele(zone_t *); 572 extern zone_t *zone_find_by_id(zoneid_t); 573 extern zone_t *zone_find_by_label(const ts_label_t *); 574 extern zone_t *zone_find_by_name(char *); 575 extern zone_t *zone_find_by_any_path(const char *, boolean_t); 576 extern zone_t *zone_find_by_path(const char *); 577 extern zoneid_t getzoneid(void); 578 extern zone_t *zone_find_by_id_nolock(zoneid_t); 579 extern int zone_datalink_walk(zoneid_t, int (*)(datalink_id_t, void *), void *); 580 extern int zone_check_datalink(zoneid_t *, datalink_id_t); 581 582 /* 583 * Zone-specific data (ZSD) APIs 584 */ 585 /* 586 * The following is what code should be initializing its zone_key_t to if it 587 * calls zone_getspecific() without necessarily knowing that zone_key_create() 588 * has been called on the key. 589 */ 590 #define ZONE_KEY_UNINITIALIZED 0 591 592 typedef uint_t zone_key_t; 593 594 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t), 595 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *)); 596 extern int zone_key_delete(zone_key_t); 597 extern void *zone_getspecific(zone_key_t, zone_t *); 598 extern int zone_setspecific(zone_key_t, zone_t *, const void *); 599 600 /* 601 * The definition of a zsd_entry is truly private to zone.c and is only 602 * placed here so it can be shared with mdb. 603 * 604 * State maintained for each zone times each registered key, which tracks 605 * the state of the create, shutdown and destroy callbacks. 606 * 607 * zsd_flags is used to keep track of pending actions to avoid holding locks 608 * when calling the create/shutdown/destroy callbacks, since doing so 609 * could lead to deadlocks. 610 */ 611 struct zsd_entry { 612 zone_key_t zsd_key; /* Key used to lookup value */ 613 void *zsd_data; /* Caller-managed value */ 614 /* 615 * Callbacks to be executed when a zone is created, shutdown, and 616 * destroyed, respectively. 617 */ 618 void *(*zsd_create)(zoneid_t); 619 void (*zsd_shutdown)(zoneid_t, void *); 620 void (*zsd_destroy)(zoneid_t, void *); 621 list_node_t zsd_linkage; 622 uint16_t zsd_flags; /* See below */ 623 kcondvar_t zsd_cv; 624 }; 625 626 /* 627 * zsd_flags 628 */ 629 #define ZSD_CREATE_NEEDED 0x0001 630 #define ZSD_CREATE_INPROGRESS 0x0002 631 #define ZSD_CREATE_COMPLETED 0x0004 632 #define ZSD_SHUTDOWN_NEEDED 0x0010 633 #define ZSD_SHUTDOWN_INPROGRESS 0x0020 634 #define ZSD_SHUTDOWN_COMPLETED 0x0040 635 #define ZSD_DESTROY_NEEDED 0x0100 636 #define ZSD_DESTROY_INPROGRESS 0x0200 637 #define ZSD_DESTROY_COMPLETED 0x0400 638 639 #define ZSD_CREATE_ALL \ 640 (ZSD_CREATE_NEEDED|ZSD_CREATE_INPROGRESS|ZSD_CREATE_COMPLETED) 641 #define ZSD_SHUTDOWN_ALL \ 642 (ZSD_SHUTDOWN_NEEDED|ZSD_SHUTDOWN_INPROGRESS|ZSD_SHUTDOWN_COMPLETED) 643 #define ZSD_DESTROY_ALL \ 644 (ZSD_DESTROY_NEEDED|ZSD_DESTROY_INPROGRESS|ZSD_DESTROY_COMPLETED) 645 646 #define ZSD_ALL_INPROGRESS \ 647 (ZSD_CREATE_INPROGRESS|ZSD_SHUTDOWN_INPROGRESS|ZSD_DESTROY_INPROGRESS) 648 649 /* 650 * Macros to help with zone visibility restrictions. 651 */ 652 653 /* 654 * Is process in the global zone? 655 */ 656 #define INGLOBALZONE(p) \ 657 ((p)->p_zone == global_zone) 658 659 /* 660 * Can process view objects in given zone? 661 */ 662 #define HASZONEACCESS(p, zoneid) \ 663 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p)) 664 665 /* 666 * Convenience macro to see if a resolved path is visible from within a 667 * given zone. 668 * 669 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the 670 * two strings must be equal. Since the rootpathlen has a trailing '/', 671 * we want to skip everything in the path up to (but not including) the 672 * trailing '/'. 673 */ 674 #define ZONE_PATH_VISIBLE(path, zone) \ 675 (strncmp((path), (zone)->zone_rootpath, \ 676 (zone)->zone_rootpathlen - 1) == 0) 677 678 /* 679 * Convenience macro to go from the global view of a path to that seen 680 * from within said zone. It is the responsibility of the caller to 681 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is 682 * in fact visible from within the zone. 683 */ 684 #define ZONE_PATH_TRANSLATE(path, zone) \ 685 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \ 686 (path) + (zone)->zone_rootpathlen - 2) 687 688 /* 689 * Special processes visible in all zones. 690 */ 691 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1) 692 693 /* 694 * Zone-safe version of thread_create() to be used when the caller wants to 695 * create a kernel thread to run within the current zone's context. 696 */ 697 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t, 698 pri_t); 699 extern void zthread_exit(void); 700 701 /* 702 * Functions for an external observer to register interest in a zone's status 703 * change. Observers will be woken up when the zone status equals the status 704 * argument passed in (in the case of zone_status_timedwait, the function may 705 * also return because of a timeout; zone_status_wait_sig may return early due 706 * to a signal being delivered; zone_status_timedwait_sig may return for any of 707 * the above reasons). 708 * 709 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and 710 * cv_wait_sig() respectively. 711 */ 712 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t); 713 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t); 714 extern void zone_status_wait(zone_t *, zone_status_t); 715 extern int zone_status_wait_sig(zone_t *, zone_status_t); 716 717 /* 718 * Get the status of the zone (at the time it was called). The state may 719 * have progressed by the time it is returned. 720 */ 721 extern zone_status_t zone_status_get(zone_t *); 722 723 /* 724 * Safely get the hostid of the specified zone (defaults to machine's hostid 725 * if the specified zone doesn't emulate a hostid). Passing NULL retrieves 726 * the global zone's (i.e., physical system's) hostid. 727 */ 728 extern uint32_t zone_get_hostid(zone_t *); 729 730 /* 731 * Get the "kcred" credentials corresponding to the given zone. 732 */ 733 extern struct cred *zone_get_kcred(zoneid_t); 734 735 /* 736 * Get/set the pool the zone is currently bound to. 737 */ 738 extern struct pool *zone_pool_get(zone_t *); 739 extern void zone_pool_set(zone_t *, struct pool *); 740 741 /* 742 * Get/set the pset the zone is currently using. 743 */ 744 extern psetid_t zone_pset_get(zone_t *); 745 extern void zone_pset_set(zone_t *, psetid_t); 746 747 /* 748 * Get the number of cpus/online-cpus visible from the given zone. 749 */ 750 extern int zone_ncpus_get(zone_t *); 751 extern int zone_ncpus_online_get(zone_t *); 752 753 /* 754 * Returns true if the named pool/dataset is visible in the current zone. 755 */ 756 extern int zone_dataset_visible(const char *, int *); 757 758 /* 759 * zone version of kadmin() 760 */ 761 extern int zone_kadmin(int, int, const char *, cred_t *); 762 extern void zone_shutdown_global(void); 763 764 extern void mount_in_progress(void); 765 extern void mount_completed(void); 766 767 extern int zone_walk(int (*)(zone_t *, void *), void *); 768 769 extern rctl_hndl_t rc_zone_locked_mem; 770 extern rctl_hndl_t rc_zone_max_swap; 771 extern rctl_hndl_t rc_zone_max_lofi; 772 773 #endif /* _KERNEL */ 774 775 #ifdef __cplusplus 776 } 777 #endif 778 779 #endif /* _SYS_ZONE_H */ 780