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 * Copyright 2014 Igor Kozhukhov <ikozhukhov@gmail.com>. 25 */ 26 27 #ifndef _SYS_ZONE_H 28 #define _SYS_ZONE_H 29 30 #include <sys/types.h> 31 #include <sys/mutex.h> 32 #include <sys/param.h> 33 #include <sys/rctl.h> 34 #include <sys/ipc_rctl.h> 35 #include <sys/pset.h> 36 #include <sys/tsol/label.h> 37 #include <sys/cred.h> 38 #include <sys/netstack.h> 39 #include <sys/uadmin.h> 40 #include <sys/ksynch.h> 41 #include <sys/socket_impl.h> 42 #include <netinet/in.h> 43 44 #ifdef __cplusplus 45 extern "C" { 46 #endif 47 48 /* 49 * NOTE 50 * 51 * The contents of this file are private to the implementation of 52 * Solaris and are subject to change at any time without notice. 53 * Applications and drivers using these interfaces may fail to 54 * run on future releases. 55 */ 56 57 /* Available both in kernel and for user space */ 58 59 /* zone id restrictions and special ids */ 60 #define MAX_ZONEID 9999 61 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */ 62 #define MIN_ZONEID 0 /* minimum zone ID on system */ 63 #define GLOBAL_ZONEID 0 64 #define ZONEID_WIDTH 4 /* for printf */ 65 66 /* 67 * Special zoneid_t token to refer to all zones. 68 */ 69 #define ALL_ZONES (-1) 70 71 /* system call subcodes */ 72 #define ZONE_CREATE 0 73 #define ZONE_DESTROY 1 74 #define ZONE_GETATTR 2 75 #define ZONE_ENTER 3 76 #define ZONE_LIST 4 77 #define ZONE_SHUTDOWN 5 78 #define ZONE_LOOKUP 6 79 #define ZONE_BOOT 7 80 #define ZONE_VERSION 8 81 #define ZONE_SETATTR 9 82 #define ZONE_ADD_DATALINK 10 83 #define ZONE_DEL_DATALINK 11 84 #define ZONE_CHECK_DATALINK 12 85 #define ZONE_LIST_DATALINK 13 86 87 /* zone attributes */ 88 #define ZONE_ATTR_ROOT 1 89 #define ZONE_ATTR_NAME 2 90 #define ZONE_ATTR_STATUS 3 91 #define ZONE_ATTR_PRIVSET 4 92 #define ZONE_ATTR_UNIQID 5 93 #define ZONE_ATTR_POOLID 6 94 #define ZONE_ATTR_INITPID 7 95 #define ZONE_ATTR_SLBL 8 96 #define ZONE_ATTR_INITNAME 9 97 #define ZONE_ATTR_BOOTARGS 10 98 #define ZONE_ATTR_BRAND 11 99 #define ZONE_ATTR_PHYS_MCAP 12 100 #define ZONE_ATTR_SCHED_CLASS 13 101 #define ZONE_ATTR_FLAGS 14 102 #define ZONE_ATTR_HOSTID 15 103 #define ZONE_ATTR_FS_ALLOWED 16 104 #define ZONE_ATTR_NETWORK 17 105 106 /* Start of the brand-specific attribute namespace */ 107 #define ZONE_ATTR_BRAND_ATTRS 32768 108 109 #define ZONE_FS_ALLOWED_MAX 1024 110 111 #define ZONE_EVENT_CHANNEL "com.sun:zones:status" 112 #define ZONE_EVENT_STATUS_CLASS "status" 113 #define ZONE_EVENT_STATUS_SUBCLASS "change" 114 115 #define ZONE_EVENT_UNINITIALIZED "uninitialized" 116 #define ZONE_EVENT_INITIALIZED "initialized" 117 #define ZONE_EVENT_READY "ready" 118 #define ZONE_EVENT_RUNNING "running" 119 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down" 120 121 #define ZONE_CB_NAME "zonename" 122 #define ZONE_CB_NEWSTATE "newstate" 123 #define ZONE_CB_OLDSTATE "oldstate" 124 #define ZONE_CB_TIMESTAMP "when" 125 #define ZONE_CB_ZONEID "zoneid" 126 127 /* 128 * Exit values that may be returned by scripts or programs invoked by various 129 * zone commands. 130 * 131 * These are defined as: 132 * 133 * ZONE_SUBPROC_OK 134 * =============== 135 * The subprocess completed successfully. 136 * 137 * ZONE_SUBPROC_USAGE 138 * ================== 139 * The subprocess failed with a usage message, or a usage message should 140 * be output in its behalf. 141 * 142 * ZONE_SUBPROC_NOTCOMPLETE 143 * ======================== 144 * The subprocess did not complete, but the actions performed by the 145 * subprocess require no recovery actions by the user. 146 * 147 * For example, if the subprocess were called by "zoneadm install," the 148 * installation of the zone did not succeed but the user need not perform 149 * a "zoneadm uninstall" before attempting another install. 150 * 151 * ZONE_SUBPROC_FATAL 152 * ================== 153 * The subprocess failed in a fatal manner, usually one that will require 154 * some type of recovery action by the user. 155 * 156 * For example, if the subprocess were called by "zoneadm install," the 157 * installation of the zone did not succeed and the user will need to 158 * perform a "zoneadm uninstall" before another install attempt is 159 * possible. 160 * 161 * The non-success exit values are large to avoid accidental collision 162 * with values used internally by some commands (e.g. "Z_ERR" and 163 * "Z_USAGE" as used by zoneadm.) 164 */ 165 #define ZONE_SUBPROC_OK 0 166 #define ZONE_SUBPROC_USAGE 253 167 #define ZONE_SUBPROC_NOTCOMPLETE 254 168 #define ZONE_SUBPROC_FATAL 255 169 170 #ifdef _SYSCALL32 171 typedef struct { 172 caddr32_t zone_name; 173 caddr32_t zone_root; 174 caddr32_t zone_privs; 175 size32_t zone_privssz; 176 caddr32_t rctlbuf; 177 size32_t rctlbufsz; 178 caddr32_t extended_error; 179 caddr32_t zfsbuf; 180 size32_t zfsbufsz; 181 int match; /* match level */ 182 uint32_t doi; /* DOI for label */ 183 caddr32_t label; /* label associated with zone */ 184 int flags; 185 } zone_def32; 186 #endif 187 typedef struct { 188 const char *zone_name; 189 const char *zone_root; 190 const struct priv_set *zone_privs; 191 size_t zone_privssz; 192 const char *rctlbuf; 193 size_t rctlbufsz; 194 int *extended_error; 195 const char *zfsbuf; 196 size_t zfsbufsz; 197 int match; /* match level */ 198 uint32_t doi; /* DOI for label */ 199 const bslabel_t *label; /* label associated with zone */ 200 int flags; 201 } zone_def; 202 203 /* extended error information */ 204 #define ZE_UNKNOWN 0 /* No extended error info */ 205 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */ 206 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */ 207 #define ZE_LABELINUSE 3 /* label is already in use by some other zone */ 208 209 /* 210 * zone_status values 211 * 212 * You must modify zone_status_names in mdb(1M)'s genunix module 213 * (genunix/zone.c) when you modify this enum. 214 */ 215 typedef enum { 216 ZONE_IS_UNINITIALIZED = 0, 217 ZONE_IS_INITIALIZED, 218 ZONE_IS_READY, 219 ZONE_IS_BOOTING, 220 ZONE_IS_RUNNING, 221 ZONE_IS_SHUTTING_DOWN, 222 ZONE_IS_EMPTY, 223 ZONE_IS_DOWN, 224 ZONE_IS_DYING, 225 ZONE_IS_DEAD 226 } zone_status_t; 227 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED 228 #define ZONE_MAX_STATE ZONE_IS_DEAD 229 230 /* 231 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also 232 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT. 233 */ 234 typedef enum zone_cmd { 235 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING, 236 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT, Z_SHUTDOWN 237 } zone_cmd_t; 238 239 /* 240 * The structure of a request to zoneadmd. 241 */ 242 typedef struct zone_cmd_arg { 243 uint64_t uniqid; /* unique "generation number" */ 244 zone_cmd_t cmd; /* requested action */ 245 uint32_t _pad; /* need consistent 32/64 bit alignmt */ 246 char locale[MAXPATHLEN]; /* locale in which to render messages */ 247 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */ 248 } zone_cmd_arg_t; 249 250 /* 251 * Structure of zoneadmd's response to a request. A NULL return value means 252 * the caller should attempt to restart zoneadmd and retry. 253 */ 254 typedef struct zone_cmd_rval { 255 int rval; /* return value of request */ 256 char errbuf[1]; /* variable-sized buffer containing error messages */ 257 } zone_cmd_rval_t; 258 259 /* 260 * The zone support infrastructure uses the zone name as a component 261 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters 262 * in length, so ZONENAME_MAX is limited by that. 263 */ 264 #define ZONENAME_MAX 64 265 266 #define GLOBAL_ZONENAME "global" 267 268 /* 269 * Extended Regular expression (see regex(5)) which matches all valid zone 270 * names. 271 */ 272 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}" 273 274 /* 275 * Where the zones support infrastructure places temporary files. 276 */ 277 #define ZONES_TMPDIR "/var/run/zones" 278 279 /* 280 * The path to the door used by clients to communicate with zoneadmd. 281 */ 282 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door" 283 284 285 /* zone_flags */ 286 /* 287 * Threads that read or write the following flag must hold zone_lock. 288 */ 289 #define ZF_REFCOUNTS_LOGGED 0x1 /* a thread logged the zone's refs */ 290 291 /* 292 * The following threads are set when the zone is created and never changed. 293 * Threads that test for these flags don't have to hold zone_lock. 294 */ 295 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */ 296 #define ZF_IS_SCRATCH 0x4 /* scratch zone */ 297 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */ 298 299 300 /* zone_create flags */ 301 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */ 302 303 /* zone network properties */ 304 #define ZONE_NETWORK_ADDRESS 1 305 #define ZONE_NETWORK_DEFROUTER 2 306 307 #define ZONE_NET_ADDRNAME "address" 308 #define ZONE_NET_RTRNAME "route" 309 310 typedef struct zone_net_data { 311 int zn_type; 312 int zn_len; 313 datalink_id_t zn_linkid; 314 uint8_t zn_val[1]; 315 } zone_net_data_t; 316 317 318 #ifdef _KERNEL 319 320 /* 321 * We need to protect the definition of 'list_t' from userland applications and 322 * libraries which may be defining ther own versions. 323 */ 324 #include <sys/list.h> 325 #include <sys/loadavg.h> 326 327 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */ 328 329 struct pool; 330 struct brand; 331 332 /* 333 * Each of these constants identifies a kernel subsystem that acquires and 334 * releases zone references. Each subsystem that invokes 335 * zone_hold_ref() and zone_rele_ref() should specify the 336 * zone_ref_subsys_t constant associated with the subsystem. Tracked holds 337 * help users and developers quickly identify subsystems that stall zone 338 * shutdowns indefinitely. 339 * 340 * NOTE: You must modify zone_ref_subsys_names in usr/src/uts/common/os/zone.c 341 * when you modify this enumeration. 342 */ 343 typedef enum zone_ref_subsys { 344 ZONE_REF_NFS, /* NFS */ 345 ZONE_REF_NFSV4, /* NFSv4 */ 346 ZONE_REF_SMBFS, /* SMBFS */ 347 ZONE_REF_MNTFS, /* MNTFS */ 348 ZONE_REF_LOFI, /* LOFI devices */ 349 ZONE_REF_VFS, /* VFS infrastructure */ 350 ZONE_REF_IPC, /* IPC infrastructure */ 351 ZONE_REF_NUM_SUBSYS /* This must be the last entry. */ 352 } zone_ref_subsys_t; 353 354 /* 355 * zone_ref represents a general-purpose references to a zone. Each zone's 356 * references are linked into the zone's zone_t::zone_ref_list. This allows 357 * debuggers to walk zones' references. 358 */ 359 typedef struct zone_ref { 360 struct zone *zref_zone; /* the zone to which the reference refers */ 361 list_node_t zref_linkage; /* linkage for zone_t::zone_ref_list */ 362 } zone_ref_t; 363 364 /* 365 * Structure to record list of ZFS datasets exported to a zone. 366 */ 367 typedef struct zone_dataset { 368 char *zd_dataset; 369 list_node_t zd_linkage; 370 } zone_dataset_t; 371 372 /* 373 * structure for zone kstats 374 */ 375 typedef struct zone_kstat { 376 kstat_named_t zk_zonename; 377 kstat_named_t zk_usage; 378 kstat_named_t zk_value; 379 } zone_kstat_t; 380 381 struct cpucap; 382 383 typedef struct { 384 kstat_named_t zm_zonename; /* full name, kstat truncates name */ 385 kstat_named_t zm_utime; 386 kstat_named_t zm_stime; 387 kstat_named_t zm_wtime; 388 kstat_named_t zm_avenrun1; 389 kstat_named_t zm_avenrun5; 390 kstat_named_t zm_avenrun15; 391 kstat_named_t zm_run_ticks; 392 kstat_named_t zm_run_wait; 393 kstat_named_t zm_fss_shr_pct; 394 kstat_named_t zm_fss_pri_hi; 395 kstat_named_t zm_fss_pri_avg; 396 } zone_misc_kstat_t; 397 398 typedef struct zone { 399 /* 400 * zone_name is never modified once set. 401 */ 402 char *zone_name; /* zone's configuration name */ 403 /* 404 * zone_nodename and zone_domain are never freed once allocated. 405 */ 406 char *zone_nodename; /* utsname.nodename equivalent */ 407 char *zone_domain; /* srpc_domain equivalent */ 408 /* 409 * zone_hostid is used for per-zone hostid emulation. 410 * Currently it isn't modified after it's set (so no locks protect 411 * accesses), but that might have to change when we allow 412 * administrators to change running zones' properties. 413 * 414 * The global zone's zone_hostid must always be HW_INVALID_HOSTID so 415 * that zone_get_hostid() will function correctly. 416 */ 417 uint32_t zone_hostid; /* zone's hostid, HW_INVALID_HOSTID */ 418 /* if not emulated */ 419 /* 420 * zone_lock protects the following fields of a zone_t: 421 * zone_ref 422 * zone_cred_ref 423 * zone_subsys_ref 424 * zone_ref_list 425 * zone_ntasks 426 * zone_flags 427 * zone_zsd 428 * zone_pfexecd 429 */ 430 kmutex_t zone_lock; 431 /* 432 * zone_linkage is the zone's linkage into the active or 433 * death-row list. The field is protected by zonehash_lock. 434 */ 435 list_node_t zone_linkage; 436 zoneid_t zone_id; /* ID of zone */ 437 uint_t zone_ref; /* count of zone_hold()s on zone */ 438 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */ 439 /* 440 * Fixed-sized array of subsystem-specific reference counts 441 * The sum of all of the counts must be less than or equal to zone_ref. 442 * The array is indexed by the counts' subsystems' zone_ref_subsys_t 443 * constants. 444 */ 445 uint_t zone_subsys_ref[ZONE_REF_NUM_SUBSYS]; 446 list_t zone_ref_list; /* list of zone_ref_t structs */ 447 /* 448 * zone_rootvp and zone_rootpath can never be modified once set. 449 */ 450 struct vnode *zone_rootvp; /* zone's root vnode */ 451 char *zone_rootpath; /* Path to zone's root + '/' */ 452 ushort_t zone_flags; /* misc flags */ 453 zone_status_t zone_status; /* protected by zone_status_lock */ 454 uint_t zone_ntasks; /* number of tasks executing in zone */ 455 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */ 456 /* counters in projects and tasks */ 457 /* that are within the zone */ 458 rctl_qty_t zone_nlwps; /* number of lwps in zone */ 459 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */ 460 rctl_qty_t zone_shmmax; /* System V shared memory usage */ 461 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */ 462 463 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */ 464 uint32_t zone_shares; /* FSS shares allocated to zone */ 465 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */ 466 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */ 467 /* kpd_locked_mem for all */ 468 /* projects in zone. */ 469 /* Also protects zone_max_swap */ 470 /* grab after p_lock, before rcs_lock */ 471 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */ 472 /* zone */ 473 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */ 474 /* limit. Protected by */ 475 /* zone_rctls->rcs_lock */ 476 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */ 477 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */ 478 /* Protected by */ 479 /* zone_rctls->rcs_lock */ 480 kmutex_t zone_rctl_lock; /* protects zone_max_lofi */ 481 rctl_qty_t zone_max_lofi; /* lofi devs for zone */ 482 rctl_qty_t zone_max_lofi_ctl; /* current lofi limit. */ 483 /* Protected by */ 484 /* zone_rctls->rcs_lock */ 485 list_t zone_zsd; /* list of Zone-Specific Data values */ 486 kcondvar_t zone_cv; /* used to signal state changes */ 487 struct proc *zone_zsched; /* Dummy kernel "zsched" process */ 488 pid_t zone_proc_initpid; /* pid of "init" for this zone */ 489 char *zone_initname; /* fs path to 'init' */ 490 int zone_boot_err; /* for zone_boot() if boot fails */ 491 char *zone_bootargs; /* arguments passed via zone_boot() */ 492 uint64_t zone_phys_mcap; /* physical memory cap */ 493 /* 494 * zone_kthreads is protected by zone_status_lock. 495 */ 496 kthread_t *zone_kthreads; /* kernel threads in zone */ 497 struct priv_set *zone_privset; /* limit set for zone */ 498 /* 499 * zone_vfslist is protected by vfs_list_lock(). 500 */ 501 struct vfs *zone_vfslist; /* list of FS's mounted in zone */ 502 uint64_t zone_uniqid; /* unique zone generation number */ 503 struct cred *zone_kcred; /* kcred-like, zone-limited cred */ 504 /* 505 * zone_pool is protected by pool_lock(). 506 */ 507 struct pool *zone_pool; /* pool the zone is bound to */ 508 hrtime_t zone_pool_mod; /* last pool bind modification time */ 509 /* zone_psetid is protected by cpu_lock */ 510 psetid_t zone_psetid; /* pset the zone is bound to */ 511 512 time_t zone_boot_time; /* Similar to boot_time */ 513 514 /* 515 * The following two can be read without holding any locks. They are 516 * updated under cpu_lock. 517 */ 518 int zone_ncpus; /* zone's idea of ncpus */ 519 int zone_ncpus_online; /* zone's idea of ncpus_online */ 520 /* 521 * List of ZFS datasets exported to this zone. 522 */ 523 list_t zone_datasets; /* list of datasets */ 524 525 ts_label_t *zone_slabel; /* zone sensitivity label */ 526 int zone_match; /* require label match for packets */ 527 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */ 528 529 boolean_t zone_restart_init; /* Restart init if it dies? */ 530 struct brand *zone_brand; /* zone's brand */ 531 void *zone_brand_data; /* store brand specific data */ 532 id_t zone_defaultcid; /* dflt scheduling class id */ 533 kstat_t *zone_swapresv_kstat; 534 kstat_t *zone_lockedmem_kstat; 535 /* 536 * zone_dl_list is protected by zone_lock 537 */ 538 list_t zone_dl_list; 539 netstack_t *zone_netstack; 540 struct cpucap *zone_cpucap; /* CPU caps data */ 541 /* 542 * Solaris Auditing per-zone audit context 543 */ 544 struct au_kcontext *zone_audit_kctxt; 545 /* 546 * For private use by mntfs. 547 */ 548 struct mntelem *zone_mntfs_db; 549 krwlock_t zone_mntfs_db_lock; 550 551 struct klpd_reg *zone_pfexecd; 552 553 char *zone_fs_allowed; 554 rctl_qty_t zone_nprocs; /* number of processes in the zone */ 555 rctl_qty_t zone_nprocs_ctl; /* current limit protected by */ 556 /* zone_rctls->rcs_lock */ 557 kstat_t *zone_nprocs_kstat; 558 559 /* 560 * Misc. kstats and counters for zone cpu-usage aggregation. 561 * The zone_Xtime values are the sum of the micro-state accounting 562 * values for all threads that are running or have run in the zone. 563 * This is tracked in msacct.c as threads change state. 564 * The zone_stime is the sum of the LMS_SYSTEM times. 565 * The zone_utime is the sum of the LMS_USER times. 566 * The zone_wtime is the sum of the LMS_WAIT_CPU times. 567 * As with per-thread micro-state accounting values, these values are 568 * not scaled to nanosecs. The scaling is done by the 569 * zone_misc_kstat_update function when kstats are requested. 570 */ 571 kmutex_t zone_misc_lock; /* protects misc statistics */ 572 kstat_t *zone_misc_ksp; 573 zone_misc_kstat_t *zone_misc_stats; 574 uint64_t zone_stime; /* total system time */ 575 uint64_t zone_utime; /* total user time */ 576 uint64_t zone_wtime; /* total time waiting in runq */ 577 578 struct loadavg_s zone_loadavg; /* loadavg for this zone */ 579 uint64_t zone_hp_avenrun[3]; /* high-precision avenrun */ 580 int zone_avenrun[3]; /* FSCALED avg. run queue len */ 581 582 /* 583 * FSS stats updated once per second by fss_decay_usage. 584 */ 585 uint32_t zone_fss_gen; /* FSS generation cntr */ 586 uint64_t zone_run_ticks; /* tot # of ticks running */ 587 588 /* 589 * DTrace-private per-zone state 590 */ 591 int zone_dtrace_getf; /* # of unprivileged getf()s */ 592 } zone_t; 593 594 /* 595 * Special value of zone_psetid to indicate that pools are disabled. 596 */ 597 #define ZONE_PS_INVAL PS_MYID 598 599 600 extern zone_t zone0; 601 extern zone_t *global_zone; 602 extern uint_t maxzones; 603 extern rctl_hndl_t rc_zone_nlwps; 604 extern rctl_hndl_t rc_zone_nprocs; 605 606 extern long zone(int, void *, void *, void *, void *); 607 extern void zone_zsd_init(void); 608 extern void zone_init(void); 609 extern void zone_hold(zone_t *); 610 extern void zone_rele(zone_t *); 611 extern void zone_init_ref(zone_ref_t *); 612 extern void zone_hold_ref(zone_t *, zone_ref_t *, zone_ref_subsys_t); 613 extern void zone_rele_ref(zone_ref_t *, zone_ref_subsys_t); 614 extern void zone_cred_hold(zone_t *); 615 extern void zone_cred_rele(zone_t *); 616 extern void zone_task_hold(zone_t *); 617 extern void zone_task_rele(zone_t *); 618 extern zone_t *zone_find_by_id(zoneid_t); 619 extern zone_t *zone_find_by_label(const ts_label_t *); 620 extern zone_t *zone_find_by_name(char *); 621 extern zone_t *zone_find_by_any_path(const char *, boolean_t); 622 extern zone_t *zone_find_by_path(const char *); 623 extern zoneid_t getzoneid(void); 624 extern zone_t *zone_find_by_id_nolock(zoneid_t); 625 extern int zone_datalink_walk(zoneid_t, int (*)(datalink_id_t, void *), void *); 626 extern int zone_check_datalink(zoneid_t *, datalink_id_t); 627 extern void zone_loadavg_update(); 628 629 /* 630 * Zone-specific data (ZSD) APIs 631 */ 632 /* 633 * The following is what code should be initializing its zone_key_t to if it 634 * calls zone_getspecific() without necessarily knowing that zone_key_create() 635 * has been called on the key. 636 */ 637 #define ZONE_KEY_UNINITIALIZED 0 638 639 typedef uint_t zone_key_t; 640 641 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t), 642 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *)); 643 extern int zone_key_delete(zone_key_t); 644 extern void *zone_getspecific(zone_key_t, zone_t *); 645 extern int zone_setspecific(zone_key_t, zone_t *, const void *); 646 647 /* 648 * The definition of a zsd_entry is truly private to zone.c and is only 649 * placed here so it can be shared with mdb. 650 * 651 * State maintained for each zone times each registered key, which tracks 652 * the state of the create, shutdown and destroy callbacks. 653 * 654 * zsd_flags is used to keep track of pending actions to avoid holding locks 655 * when calling the create/shutdown/destroy callbacks, since doing so 656 * could lead to deadlocks. 657 */ 658 struct zsd_entry { 659 zone_key_t zsd_key; /* Key used to lookup value */ 660 void *zsd_data; /* Caller-managed value */ 661 /* 662 * Callbacks to be executed when a zone is created, shutdown, and 663 * destroyed, respectively. 664 */ 665 void *(*zsd_create)(zoneid_t); 666 void (*zsd_shutdown)(zoneid_t, void *); 667 void (*zsd_destroy)(zoneid_t, void *); 668 list_node_t zsd_linkage; 669 uint16_t zsd_flags; /* See below */ 670 kcondvar_t zsd_cv; 671 }; 672 673 /* 674 * zsd_flags 675 */ 676 #define ZSD_CREATE_NEEDED 0x0001 677 #define ZSD_CREATE_INPROGRESS 0x0002 678 #define ZSD_CREATE_COMPLETED 0x0004 679 #define ZSD_SHUTDOWN_NEEDED 0x0010 680 #define ZSD_SHUTDOWN_INPROGRESS 0x0020 681 #define ZSD_SHUTDOWN_COMPLETED 0x0040 682 #define ZSD_DESTROY_NEEDED 0x0100 683 #define ZSD_DESTROY_INPROGRESS 0x0200 684 #define ZSD_DESTROY_COMPLETED 0x0400 685 686 #define ZSD_CREATE_ALL \ 687 (ZSD_CREATE_NEEDED|ZSD_CREATE_INPROGRESS|ZSD_CREATE_COMPLETED) 688 #define ZSD_SHUTDOWN_ALL \ 689 (ZSD_SHUTDOWN_NEEDED|ZSD_SHUTDOWN_INPROGRESS|ZSD_SHUTDOWN_COMPLETED) 690 #define ZSD_DESTROY_ALL \ 691 (ZSD_DESTROY_NEEDED|ZSD_DESTROY_INPROGRESS|ZSD_DESTROY_COMPLETED) 692 693 #define ZSD_ALL_INPROGRESS \ 694 (ZSD_CREATE_INPROGRESS|ZSD_SHUTDOWN_INPROGRESS|ZSD_DESTROY_INPROGRESS) 695 696 /* 697 * Macros to help with zone visibility restrictions. 698 */ 699 700 /* 701 * Is process in the global zone? 702 */ 703 #define INGLOBALZONE(p) \ 704 ((p)->p_zone == global_zone) 705 706 /* 707 * Can process view objects in given zone? 708 */ 709 #define HASZONEACCESS(p, zoneid) \ 710 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p)) 711 712 /* 713 * Convenience macro to see if a resolved path is visible from within a 714 * given zone. 715 * 716 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the 717 * two strings must be equal. Since the rootpathlen has a trailing '/', 718 * we want to skip everything in the path up to (but not including) the 719 * trailing '/'. 720 */ 721 #define ZONE_PATH_VISIBLE(path, zone) \ 722 (strncmp((path), (zone)->zone_rootpath, \ 723 (zone)->zone_rootpathlen - 1) == 0) 724 725 /* 726 * Convenience macro to go from the global view of a path to that seen 727 * from within said zone. It is the responsibility of the caller to 728 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is 729 * in fact visible from within the zone. 730 */ 731 #define ZONE_PATH_TRANSLATE(path, zone) \ 732 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \ 733 (path) + (zone)->zone_rootpathlen - 2) 734 735 /* 736 * Special processes visible in all zones. 737 */ 738 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1) 739 740 /* 741 * Zone-safe version of thread_create() to be used when the caller wants to 742 * create a kernel thread to run within the current zone's context. 743 */ 744 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t, 745 pri_t); 746 extern void zthread_exit(void); 747 748 /* 749 * Functions for an external observer to register interest in a zone's status 750 * change. Observers will be woken up when the zone status equals the status 751 * argument passed in (in the case of zone_status_timedwait, the function may 752 * also return because of a timeout; zone_status_wait_sig may return early due 753 * to a signal being delivered; zone_status_timedwait_sig may return for any of 754 * the above reasons). 755 * 756 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and 757 * cv_wait_sig() respectively. 758 */ 759 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t); 760 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t); 761 extern void zone_status_wait(zone_t *, zone_status_t); 762 extern int zone_status_wait_sig(zone_t *, zone_status_t); 763 764 /* 765 * Get the status of the zone (at the time it was called). The state may 766 * have progressed by the time it is returned. 767 */ 768 extern zone_status_t zone_status_get(zone_t *); 769 770 /* 771 * Safely get the hostid of the specified zone (defaults to machine's hostid 772 * if the specified zone doesn't emulate a hostid). Passing NULL retrieves 773 * the global zone's (i.e., physical system's) hostid. 774 */ 775 extern uint32_t zone_get_hostid(zone_t *); 776 777 /* 778 * Get the "kcred" credentials corresponding to the given zone. 779 */ 780 extern struct cred *zone_get_kcred(zoneid_t); 781 782 /* 783 * Get/set the pool the zone is currently bound to. 784 */ 785 extern struct pool *zone_pool_get(zone_t *); 786 extern void zone_pool_set(zone_t *, struct pool *); 787 788 /* 789 * Get/set the pset the zone is currently using. 790 */ 791 extern psetid_t zone_pset_get(zone_t *); 792 extern void zone_pset_set(zone_t *, psetid_t); 793 794 /* 795 * Get the number of cpus/online-cpus visible from the given zone. 796 */ 797 extern int zone_ncpus_get(zone_t *); 798 extern int zone_ncpus_online_get(zone_t *); 799 800 /* 801 * Returns true if the named pool/dataset is visible in the current zone. 802 */ 803 extern int zone_dataset_visible(const char *, int *); 804 805 /* 806 * zone version of kadmin() 807 */ 808 extern int zone_kadmin(int, int, const char *, cred_t *); 809 extern void zone_shutdown_global(void); 810 811 extern void mount_in_progress(void); 812 extern void mount_completed(void); 813 814 extern int zone_walk(int (*)(zone_t *, void *), void *); 815 816 extern rctl_hndl_t rc_zone_locked_mem; 817 extern rctl_hndl_t rc_zone_max_swap; 818 extern rctl_hndl_t rc_zone_max_lofi; 819 820 #endif /* _KERNEL */ 821 822 #ifdef __cplusplus 823 } 824 #endif 825 826 #endif /* _SYS_ZONE_H */ 827