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 /*
23 * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
24 */
25
26 /*
27 * Copyright 2015 Joyent, Inc. All rights reserved.
28 * Copyright 2018 Nexenta Systems, Inc.
29 * Copyright 2014 Igor Kozhukhov <ikozhukhov@gmail.com>.
30 */
31
32 #ifndef _SYS_ZONE_H
33 #define _SYS_ZONE_H
34
35 #include <sys/types.h>
36 #include <sys/mutex.h>
37 #include <sys/param.h>
38 #include <sys/rctl.h>
39 #include <sys/ipc_rctl.h>
40 #include <sys/pset.h>
41 #include <sys/tsol/label.h>
42 #include <sys/cred.h>
43 #include <sys/netstack.h>
44 #include <sys/uadmin.h>
45 #include <sys/ksynch.h>
46 #include <sys/socket_impl.h>
47 #include <sys/secflags.h>
48 #include <netinet/in.h>
49
50 #ifdef __cplusplus
51 extern "C" {
52 #endif
53
54 /*
55 * NOTE
56 *
57 * The contents of this file are private to the implementation of
58 * Solaris and are subject to change at any time without notice.
59 * Applications and drivers using these interfaces may fail to
60 * run on future releases.
61 */
62
63 /* Available both in kernel and for user space */
64
65 /* zone id restrictions and special ids */
66 #define MAX_ZONEID 9999
67 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */
68 #define MIN_ZONEID 0 /* minimum zone ID on system */
69 #define GLOBAL_ZONEID 0
70 #define ZONEID_WIDTH 4 /* for printf */
71
72 /*
73 * Special zoneid_t token to refer to all zones.
74 */
75 #define ALL_ZONES (-1)
76
77 /* system call subcodes */
78 #define ZONE_CREATE 0
79 #define ZONE_DESTROY 1
80 #define ZONE_GETATTR 2
81 #define ZONE_ENTER 3
82 #define ZONE_LIST 4
83 #define ZONE_SHUTDOWN 5
84 #define ZONE_LOOKUP 6
85 #define ZONE_BOOT 7
86 #define ZONE_VERSION 8
87 #define ZONE_SETATTR 9
88 #define ZONE_ADD_DATALINK 10
89 #define ZONE_DEL_DATALINK 11
90 #define ZONE_CHECK_DATALINK 12
91 #define ZONE_LIST_DATALINK 13
92
93 /* zone attributes */
94 #define ZONE_ATTR_ROOT 1
95 #define ZONE_ATTR_NAME 2
96 #define ZONE_ATTR_STATUS 3
97 #define ZONE_ATTR_PRIVSET 4
98 #define ZONE_ATTR_UNIQID 5
99 #define ZONE_ATTR_POOLID 6
100 #define ZONE_ATTR_INITPID 7
101 #define ZONE_ATTR_SLBL 8
102 #define ZONE_ATTR_INITNAME 9
103 #define ZONE_ATTR_BOOTARGS 10
104 #define ZONE_ATTR_BRAND 11
105 #define ZONE_ATTR_PHYS_MCAP 12
106 #define ZONE_ATTR_SCHED_CLASS 13
107 #define ZONE_ATTR_FLAGS 14
108 #define ZONE_ATTR_HOSTID 15
109 #define ZONE_ATTR_FS_ALLOWED 16
110 #define ZONE_ATTR_NETWORK 17
111 #define ZONE_ATTR_INITNORESTART 20
112 #define ZONE_ATTR_SECFLAGS 21
113
114 /* Start of the brand-specific attribute namespace */
115 #define ZONE_ATTR_BRAND_ATTRS 32768
116
117 #define ZONE_FS_ALLOWED_MAX 1024
118
119 #define ZONE_EVENT_CHANNEL "com.sun:zones:status"
120 #define ZONE_EVENT_STATUS_CLASS "status"
121 #define ZONE_EVENT_STATUS_SUBCLASS "change"
122
123 #define ZONE_EVENT_UNINITIALIZED "uninitialized"
124 #define ZONE_EVENT_INITIALIZED "initialized"
125 #define ZONE_EVENT_READY "ready"
126 #define ZONE_EVENT_RUNNING "running"
127 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down"
128
129 #define ZONE_CB_NAME "zonename"
130 #define ZONE_CB_NEWSTATE "newstate"
131 #define ZONE_CB_OLDSTATE "oldstate"
132 #define ZONE_CB_TIMESTAMP "when"
133 #define ZONE_CB_ZONEID "zoneid"
134
135 /*
136 * Exit values that may be returned by scripts or programs invoked by various
137 * zone commands.
138 *
139 * These are defined as:
140 *
141 * ZONE_SUBPROC_OK
142 * ===============
143 * The subprocess completed successfully.
144 *
145 * ZONE_SUBPROC_USAGE
146 * ==================
147 * The subprocess failed with a usage message, or a usage message should
148 * be output in its behalf.
149 *
150 * ZONE_SUBPROC_NOTCOMPLETE
151 * ========================
152 * The subprocess did not complete, but the actions performed by the
153 * subprocess require no recovery actions by the user.
154 *
155 * For example, if the subprocess were called by "zoneadm install," the
156 * installation of the zone did not succeed but the user need not perform
157 * a "zoneadm uninstall" before attempting another install.
158 *
159 * ZONE_SUBPROC_FATAL
160 * ==================
161 * The subprocess failed in a fatal manner, usually one that will require
162 * some type of recovery action by the user.
163 *
164 * For example, if the subprocess were called by "zoneadm install," the
165 * installation of the zone did not succeed and the user will need to
166 * perform a "zoneadm uninstall" before another install attempt is
167 * possible.
168 *
169 * The non-success exit values are large to avoid accidental collision
170 * with values used internally by some commands (e.g. "Z_ERR" and
171 * "Z_USAGE" as used by zoneadm.)
172 */
173 #define ZONE_SUBPROC_OK 0
174 #define ZONE_SUBPROC_USAGE 253
175 #define ZONE_SUBPROC_NOTCOMPLETE 254
176 #define ZONE_SUBPROC_FATAL 255
177
178 #ifdef _SYSCALL32
179 typedef struct {
180 caddr32_t zone_name;
181 caddr32_t zone_root;
182 caddr32_t zone_privs;
183 size32_t zone_privssz;
184 caddr32_t rctlbuf;
185 size32_t rctlbufsz;
186 caddr32_t extended_error;
187 caddr32_t zfsbuf;
188 size32_t zfsbufsz;
189 int match; /* match level */
190 uint32_t doi; /* DOI for label */
191 caddr32_t label; /* label associated with zone */
192 int flags;
193 } zone_def32;
194 #endif
195 typedef struct {
196 const char *zone_name;
197 const char *zone_root;
198 const struct priv_set *zone_privs;
199 size_t zone_privssz;
200 const char *rctlbuf;
201 size_t rctlbufsz;
202 int *extended_error;
203 const char *zfsbuf;
204 size_t zfsbufsz;
205 int match; /* match level */
206 uint32_t doi; /* DOI for label */
207 const bslabel_t *label; /* label associated with zone */
208 int flags;
209 } zone_def;
210
211 /* extended error information */
212 #define ZE_UNKNOWN 0 /* No extended error info */
213 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */
214 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */
215 #define ZE_LABELINUSE 3 /* label is already in use by some other zone */
216
217 /*
218 * zone_status values
219 *
220 * You must modify zone_status_names in mdb(1M)'s genunix module
221 * (genunix/zone.c) when you modify this enum.
222 */
223 typedef enum {
224 ZONE_IS_UNINITIALIZED = 0,
225 ZONE_IS_INITIALIZED,
226 ZONE_IS_READY,
227 ZONE_IS_BOOTING,
228 ZONE_IS_RUNNING,
229 ZONE_IS_SHUTTING_DOWN,
230 ZONE_IS_EMPTY,
231 ZONE_IS_DOWN,
232 ZONE_IS_DYING,
233 ZONE_IS_DEAD
234 } zone_status_t;
235 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED
236 #define ZONE_MAX_STATE ZONE_IS_DEAD
237
238 /*
239 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also
240 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT.
241 */
242 typedef enum zone_cmd {
243 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING,
244 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT, Z_SHUTDOWN
245 } zone_cmd_t;
246
247 /*
248 * The structure of a request to zoneadmd.
249 */
250 typedef struct zone_cmd_arg {
251 uint64_t uniqid; /* unique "generation number" */
252 zone_cmd_t cmd; /* requested action */
253 uint32_t _pad; /* need consistent 32/64 bit alignmt */
254 char locale[MAXPATHLEN]; /* locale in which to render messages */
255 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */
256 } zone_cmd_arg_t;
257
258 /*
259 * Structure of zoneadmd's response to a request. A NULL return value means
260 * the caller should attempt to restart zoneadmd and retry.
261 */
262 typedef struct zone_cmd_rval {
263 int rval; /* return value of request */
264 char errbuf[1]; /* variable-sized buffer containing error messages */
265 } zone_cmd_rval_t;
266
267 /*
268 * The zone support infrastructure uses the zone name as a component
269 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters
270 * in length, so ZONENAME_MAX is limited by that.
271 */
272 #define ZONENAME_MAX 64
273
274 #define GLOBAL_ZONENAME "global"
275
276 /*
277 * Extended Regular expression (see regex(5)) which matches all valid zone
278 * names.
279 */
280 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}"
281
282 /*
283 * Where the zones support infrastructure places temporary files.
284 */
285 #define ZONES_TMPDIR "/var/run/zones"
286
287 /*
288 * The path to the door used by clients to communicate with zoneadmd.
289 */
290 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door"
291
292
293 /* zone_flags */
294 /*
295 * Threads that read or write the following flag must hold zone_lock.
296 */
297 #define ZF_REFCOUNTS_LOGGED 0x1 /* a thread logged the zone's refs */
298
299 /*
300 * The following threads are set when the zone is created and never changed.
301 * Threads that test for these flags don't have to hold zone_lock.
302 */
303 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */
304 #define ZF_IS_SCRATCH 0x4 /* scratch zone */
305 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */
306
307
308 /* zone_create flags */
309 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */
310
311 /* zone network properties */
312 #define ZONE_NETWORK_ADDRESS 1
313 #define ZONE_NETWORK_DEFROUTER 2
314
315 #define ZONE_NET_ADDRNAME "address"
316 #define ZONE_NET_RTRNAME "route"
317
318 typedef struct zone_net_data {
319 int zn_type;
320 int zn_len;
321 datalink_id_t zn_linkid;
322 uint8_t zn_val[1];
323 } zone_net_data_t;
324
325
326 #ifdef _KERNEL
327
328 /*
329 * We need to protect the definition of 'list_t' from userland applications and
330 * libraries which may be defining ther own versions.
331 */
332 #include <sys/list.h>
333 #include <sys/loadavg.h>
334
335 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */
336
337 struct pool;
338 struct brand;
339
340 /*
341 * Each of these constants identifies a kernel subsystem that acquires and
342 * releases zone references. Each subsystem that invokes
343 * zone_hold_ref() and zone_rele_ref() should specify the
344 * zone_ref_subsys_t constant associated with the subsystem. Tracked holds
345 * help users and developers quickly identify subsystems that stall zone
346 * shutdowns indefinitely.
347 *
348 * NOTE: You must modify zone_ref_subsys_names in usr/src/uts/common/os/zone.c
349 * when you modify this enumeration.
350 */
351 typedef enum zone_ref_subsys {
352 ZONE_REF_NFS, /* NFS */
353 ZONE_REF_NFSV4, /* NFSv4 */
354 ZONE_REF_SMBFS, /* SMBFS */
355 ZONE_REF_MNTFS, /* MNTFS */
356 ZONE_REF_LOFI, /* LOFI devices */
357 ZONE_REF_VFS, /* VFS infrastructure */
358 ZONE_REF_IPC, /* IPC infrastructure */
359 ZONE_REF_NUM_SUBSYS /* This must be the last entry. */
360 } zone_ref_subsys_t;
361
362 /*
363 * zone_ref represents a general-purpose references to a zone. Each zone's
364 * references are linked into the zone's zone_t::zone_ref_list. This allows
365 * debuggers to walk zones' references.
366 */
367 typedef struct zone_ref {
368 struct zone *zref_zone; /* the zone to which the reference refers */
369 list_node_t zref_linkage; /* linkage for zone_t::zone_ref_list */
370 } zone_ref_t;
371
372 /*
373 * Structure to record list of ZFS datasets exported to a zone.
374 */
375 typedef struct zone_dataset {
376 char *zd_dataset;
377 list_node_t zd_linkage;
378 } zone_dataset_t;
379
380 /*
381 * structure for zone kstats
382 */
383 typedef struct zone_kstat {
384 kstat_named_t zk_zonename;
385 kstat_named_t zk_usage;
386 kstat_named_t zk_value;
387 } zone_kstat_t;
388
389 struct cpucap;
390
391 typedef struct {
392 kstat_named_t zm_zonename;
393 kstat_named_t zm_pgpgin;
394 kstat_named_t zm_anonpgin;
395 kstat_named_t zm_execpgin;
396 kstat_named_t zm_fspgin;
397 kstat_named_t zm_anon_alloc_fail;
398 } zone_mcap_kstat_t;
399
400 typedef struct {
401 kstat_named_t zm_zonename; /* full name, kstat truncates name */
402 kstat_named_t zm_utime;
403 kstat_named_t zm_stime;
404 kstat_named_t zm_wtime;
405 kstat_named_t zm_avenrun1;
406 kstat_named_t zm_avenrun5;
407 kstat_named_t zm_avenrun15;
408 kstat_named_t zm_ffcap;
409 kstat_named_t zm_ffnoproc;
410 kstat_named_t zm_ffnomem;
411 kstat_named_t zm_ffmisc;
412 kstat_named_t zm_nested_intp;
413 kstat_named_t zm_init_pid;
414 kstat_named_t zm_boot_time;
415 } zone_misc_kstat_t;
416
417 typedef struct zone {
418 /*
419 * zone_name is never modified once set.
420 */
421 char *zone_name; /* zone's configuration name */
422 /*
423 * zone_nodename and zone_domain are never freed once allocated.
424 */
425 char *zone_nodename; /* utsname.nodename equivalent */
426 char *zone_domain; /* srpc_domain equivalent */
427 /*
428 * zone_hostid is used for per-zone hostid emulation.
429 * Currently it isn't modified after it's set (so no locks protect
430 * accesses), but that might have to change when we allow
431 * administrators to change running zones' properties.
432 *
433 * The global zone's zone_hostid must always be HW_INVALID_HOSTID so
434 * that zone_get_hostid() will function correctly.
435 */
436 uint32_t zone_hostid; /* zone's hostid, HW_INVALID_HOSTID */
437 /* if not emulated */
438 /*
439 * zone_lock protects the following fields of a zone_t:
440 * zone_ref
441 * zone_cred_ref
442 * zone_subsys_ref
443 * zone_ref_list
444 * zone_ntasks
445 * zone_flags
446 * zone_zsd
447 * zone_pfexecd
448 */
449 kmutex_t zone_lock;
450 /*
451 * zone_linkage is the zone's linkage into the active or
452 * death-row list. The field is protected by zonehash_lock.
453 */
454 list_node_t zone_linkage;
455 zoneid_t zone_id; /* ID of zone */
456 uint_t zone_ref; /* count of zone_hold()s on zone */
457 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */
458 /*
459 * Fixed-sized array of subsystem-specific reference counts
460 * The sum of all of the counts must be less than or equal to zone_ref.
461 * The array is indexed by the counts' subsystems' zone_ref_subsys_t
462 * constants.
463 */
464 uint_t zone_subsys_ref[ZONE_REF_NUM_SUBSYS];
465 list_t zone_ref_list; /* list of zone_ref_t structs */
466 /*
467 * zone_rootvp and zone_rootpath can never be modified once set.
468 */
469 struct vnode *zone_rootvp; /* zone's root vnode */
470 char *zone_rootpath; /* Path to zone's root + '/' */
471 ushort_t zone_flags; /* misc flags */
472 zone_status_t zone_status; /* protected by zone_status_lock */
473 uint_t zone_ntasks; /* number of tasks executing in zone */
474 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */
475 /* counters in projects and tasks */
476 /* that are within the zone */
477 rctl_qty_t zone_nlwps; /* number of lwps in zone */
478 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */
479 rctl_qty_t zone_shmmax; /* System V shared memory usage */
480 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */
481
482 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */
483 uint32_t zone_shares; /* FSS shares allocated to zone */
484 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */
485 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */
486 /* kpd_locked_mem for all */
487 /* projects in zone. */
488 /* Also protects zone_max_swap */
489 /* grab after p_lock, before rcs_lock */
490 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */
491 /* zone */
492 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */
493 /* limit. Protected by */
494 /* zone_rctls->rcs_lock */
495 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */
496 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */
497 /* Protected by */
498 /* zone_rctls->rcs_lock */
499 kmutex_t zone_rctl_lock; /* protects zone_max_lofi */
500 rctl_qty_t zone_max_lofi; /* lofi devs for zone */
501 rctl_qty_t zone_max_lofi_ctl; /* current lofi limit. */
502 /* Protected by */
503 /* zone_rctls->rcs_lock */
504 list_t zone_zsd; /* list of Zone-Specific Data values */
505 kcondvar_t zone_cv; /* used to signal state changes */
506 struct proc *zone_zsched; /* Dummy kernel "zsched" process */
507 pid_t zone_proc_initpid; /* pid of "init" for this zone */
508 char *zone_initname; /* fs path to 'init' */
509 int zone_boot_err; /* for zone_boot() if boot fails */
510 char *zone_bootargs; /* arguments passed via zone_boot() */
511 uint64_t zone_phys_mcap; /* physical memory cap */
512 /*
513 * zone_kthreads is protected by zone_status_lock.
514 */
515 kthread_t *zone_kthreads; /* kernel threads in zone */
516 struct priv_set *zone_privset; /* limit set for zone */
517 /*
518 * zone_vfslist is protected by vfs_list_lock().
519 */
520 struct vfs *zone_vfslist; /* list of FS's mounted in zone */
521 uint64_t zone_uniqid; /* unique zone generation number */
522 struct cred *zone_kcred; /* kcred-like, zone-limited cred */
523 /*
524 * zone_pool is protected by pool_lock().
525 */
526 struct pool *zone_pool; /* pool the zone is bound to */
527 hrtime_t zone_pool_mod; /* last pool bind modification time */
528 /* zone_psetid is protected by cpu_lock */
529 psetid_t zone_psetid; /* pset the zone is bound to */
530
531 time_t zone_boot_time; /* Similar to boot_time */
532
533 /*
534 * The following two can be read without holding any locks. They are
535 * updated under cpu_lock.
536 */
537 int zone_ncpus; /* zone's idea of ncpus */
538 int zone_ncpus_online; /* zone's idea of ncpus_online */
539 /*
540 * List of ZFS datasets exported to this zone.
541 */
542 list_t zone_datasets; /* list of datasets */
543
544 ts_label_t *zone_slabel; /* zone sensitivity label */
545 int zone_match; /* require label match for packets */
546 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */
547
548 boolean_t zone_restart_init; /* Restart init if it dies? */
549 struct brand *zone_brand; /* zone's brand */
550 void *zone_brand_data; /* store brand specific data */
551 id_t zone_defaultcid; /* dflt scheduling class id */
552 kstat_t *zone_swapresv_kstat;
553 kstat_t *zone_lockedmem_kstat;
554 /*
555 * zone_dl_list is protected by zone_lock
556 */
557 list_t zone_dl_list;
558 netstack_t *zone_netstack;
559 struct cpucap *zone_cpucap; /* CPU caps data */
560 /*
561 * Solaris Auditing per-zone audit context
562 */
563 struct au_kcontext *zone_audit_kctxt;
564 /*
565 * For private use by mntfs.
566 */
567 struct mntelem *zone_mntfs_db;
568 krwlock_t zone_mntfs_db_lock;
569
570 struct klpd_reg *zone_pfexecd;
571
572 char *zone_fs_allowed;
573 rctl_qty_t zone_nprocs; /* number of processes in the zone */
574 rctl_qty_t zone_nprocs_ctl; /* current limit protected by */
575 /* zone_rctls->rcs_lock */
576 kstat_t *zone_nprocs_kstat;
577
578 kmutex_t zone_mcap_lock; /* protects mcap statistics */
579 kstat_t *zone_mcap_ksp;
580 zone_mcap_kstat_t *zone_mcap_stats;
581 uint64_t zone_pgpgin; /* pages paged in */
582 uint64_t zone_anonpgin; /* anon pages paged in */
583 uint64_t zone_execpgin; /* exec pages paged in */
584 uint64_t zone_fspgin; /* fs pages paged in */
585 uint64_t zone_anon_alloc_fail; /* cnt of anon alloc fails */
586
587 psecflags_t zone_secflags; /* default zone security-flags */
588
589 /*
590 * Misc. kstats and counters for zone cpu-usage aggregation.
591 * The zone_Xtime values are the sum of the micro-state accounting
592 * values for all threads that are running or have run in the zone.
593 * This is tracked in msacct.c as threads change state.
594 * The zone_stime is the sum of the LMS_SYSTEM times.
595 * The zone_utime is the sum of the LMS_USER times.
596 * The zone_wtime is the sum of the LMS_WAIT_CPU times.
597 * As with per-thread micro-state accounting values, these values are
598 * not scaled to nanosecs. The scaling is done by the
599 * zone_misc_kstat_update function when kstats are requested.
600 */
601 kmutex_t zone_misc_lock; /* protects misc statistics */
602 kstat_t *zone_misc_ksp;
603 zone_misc_kstat_t *zone_misc_stats;
604 uint64_t zone_stime; /* total system time */
605 uint64_t zone_utime; /* total user time */
606 uint64_t zone_wtime; /* total time waiting in runq */
607 /* fork-fail kstat tracking */
608 uint32_t zone_ffcap; /* hit an rctl cap */
609 uint32_t zone_ffnoproc; /* get proc/lwp error */
610 uint32_t zone_ffnomem; /* as_dup/memory error */
611 uint32_t zone_ffmisc; /* misc. other error */
612
613 uint32_t zone_nested_intp; /* nested interp. kstat */
614
615 struct loadavg_s zone_loadavg; /* loadavg for this zone */
616 uint64_t zone_hp_avenrun[3]; /* high-precision avenrun */
617 int zone_avenrun[3]; /* FSCALED avg. run queue len */
618
619 /*
620 * FSS stats updated once per second by fss_decay_usage.
621 */
622 uint32_t zone_fss_gen; /* FSS generation cntr */
623 uint64_t zone_run_ticks; /* tot # of ticks running */
624
625 /*
626 * DTrace-private per-zone state
627 */
628 int zone_dtrace_getf; /* # of unprivileged getf()s */
629
630 /*
631 * Synchronization primitives used to synchronize between mounts and
632 * zone creation/destruction.
633 */
634 int zone_mounts_in_progress;
635 kcondvar_t zone_mount_cv;
636 kmutex_t zone_mount_lock;
637 } zone_t;
638
639 /*
640 * Special value of zone_psetid to indicate that pools are disabled.
641 */
642 #define ZONE_PS_INVAL PS_MYID
643
644 extern zone_t zone0;
645 extern zone_t *global_zone;
646 extern uint_t maxzones;
647 extern rctl_hndl_t rc_zone_nlwps;
648 extern rctl_hndl_t rc_zone_nprocs;
649
650 extern long zone(int, void *, void *, void *, void *);
651 extern void zone_zsd_init(void);
652 extern void zone_init(void);
653 extern void zone_hold(zone_t *);
654 extern void zone_rele(zone_t *);
655 extern void zone_init_ref(zone_ref_t *);
656 extern void zone_hold_ref(zone_t *, zone_ref_t *, zone_ref_subsys_t);
657 extern void zone_rele_ref(zone_ref_t *, zone_ref_subsys_t);
658 extern void zone_cred_hold(zone_t *);
659 extern void zone_cred_rele(zone_t *);
660 extern void zone_task_hold(zone_t *);
661 extern void zone_task_rele(zone_t *);
662 extern zone_t *zone_find_by_id(zoneid_t);
663 extern zone_t *zone_find_by_label(const ts_label_t *);
664 extern zone_t *zone_find_by_name(char *);
665 extern zone_t *zone_find_by_any_path(const char *, boolean_t);
666 extern zone_t *zone_find_by_path(const char *);
667 extern zoneid_t getzoneid(void);
668 extern zone_t *zone_find_by_id_nolock(zoneid_t);
669 extern int zone_datalink_walk(zoneid_t, int (*)(datalink_id_t, void *), void *);
670 extern int zone_check_datalink(zoneid_t *, datalink_id_t);
671 extern void zone_loadavg_update();
672
673 /*
674 * Zone-specific data (ZSD) APIs
675 */
676 /*
677 * The following is what code should be initializing its zone_key_t to if it
678 * calls zone_getspecific() without necessarily knowing that zone_key_create()
679 * has been called on the key.
680 */
681 #define ZONE_KEY_UNINITIALIZED 0
682
683 typedef uint_t zone_key_t;
684
685 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t),
686 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *));
687 extern int zone_key_delete(zone_key_t);
688 extern void *zone_getspecific(zone_key_t, zone_t *);
689 extern int zone_setspecific(zone_key_t, zone_t *, const void *);
690
691 /*
692 * The definition of a zsd_entry is truly private to zone.c and is only
693 * placed here so it can be shared with mdb.
694 *
695 * State maintained for each zone times each registered key, which tracks
696 * the state of the create, shutdown and destroy callbacks.
697 *
698 * zsd_flags is used to keep track of pending actions to avoid holding locks
699 * when calling the create/shutdown/destroy callbacks, since doing so
700 * could lead to deadlocks.
701 */
702 struct zsd_entry {
703 zone_key_t zsd_key; /* Key used to lookup value */
704 void *zsd_data; /* Caller-managed value */
705 /*
706 * Callbacks to be executed when a zone is created, shutdown, and
707 * destroyed, respectively.
708 */
709 void *(*zsd_create)(zoneid_t);
710 void (*zsd_shutdown)(zoneid_t, void *);
711 void (*zsd_destroy)(zoneid_t, void *);
712 list_node_t zsd_linkage;
713 uint16_t zsd_flags; /* See below */
714 kcondvar_t zsd_cv;
715 };
716
717 /*
718 * zsd_flags
719 */
720 #define ZSD_CREATE_NEEDED 0x0001
721 #define ZSD_CREATE_INPROGRESS 0x0002
722 #define ZSD_CREATE_COMPLETED 0x0004
723 #define ZSD_SHUTDOWN_NEEDED 0x0010
724 #define ZSD_SHUTDOWN_INPROGRESS 0x0020
725 #define ZSD_SHUTDOWN_COMPLETED 0x0040
726 #define ZSD_DESTROY_NEEDED 0x0100
727 #define ZSD_DESTROY_INPROGRESS 0x0200
728 #define ZSD_DESTROY_COMPLETED 0x0400
729
730 #define ZSD_CREATE_ALL \
731 (ZSD_CREATE_NEEDED|ZSD_CREATE_INPROGRESS|ZSD_CREATE_COMPLETED)
732 #define ZSD_SHUTDOWN_ALL \
733 (ZSD_SHUTDOWN_NEEDED|ZSD_SHUTDOWN_INPROGRESS|ZSD_SHUTDOWN_COMPLETED)
734 #define ZSD_DESTROY_ALL \
735 (ZSD_DESTROY_NEEDED|ZSD_DESTROY_INPROGRESS|ZSD_DESTROY_COMPLETED)
736
737 #define ZSD_ALL_INPROGRESS \
738 (ZSD_CREATE_INPROGRESS|ZSD_SHUTDOWN_INPROGRESS|ZSD_DESTROY_INPROGRESS)
739
740 /*
741 * Macros to help with zone visibility restrictions.
742 */
743
744 /*
745 * Is process in the global zone?
746 */
747 #define INGLOBALZONE(p) \
748 ((p)->p_zone == global_zone)
749
750 /*
751 * Can process view objects in given zone?
752 */
753 #define HASZONEACCESS(p, zoneid) \
754 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p))
755
756 /*
757 * Convenience macro to see if a resolved path is visible from within a
758 * given zone.
759 *
760 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the
761 * two strings must be equal. Since the rootpathlen has a trailing '/',
762 * we want to skip everything in the path up to (but not including) the
763 * trailing '/'.
764 */
765 #define ZONE_PATH_VISIBLE(path, zone) \
766 (strncmp((path), (zone)->zone_rootpath, \
767 (zone)->zone_rootpathlen - 1) == 0)
768
769 /*
770 * Convenience macro to go from the global view of a path to that seen
771 * from within said zone. It is the responsibility of the caller to
772 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is
773 * in fact visible from within the zone.
774 */
775 #define ZONE_PATH_TRANSLATE(path, zone) \
776 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \
777 (path) + (zone)->zone_rootpathlen - 2)
778
779 /*
780 * Special processes visible in all zones.
781 */
782 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1)
783
784 /*
785 * A root vnode of the current zone.
786 */
787 #define ZONE_ROOTVP() (curproc->p_zone->zone_rootvp)
788
789 /*
790 * Zone-safe version of thread_create() to be used when the caller wants to
791 * create a kernel thread to run within the current zone's context.
792 */
793 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t,
794 pri_t);
795 extern void zthread_exit(void);
796
797 /*
798 * Functions for an external observer to register interest in a zone's status
799 * change. Observers will be woken up when the zone status equals the status
800 * argument passed in (in the case of zone_status_timedwait, the function may
801 * also return because of a timeout; zone_status_wait_sig may return early due
802 * to a signal being delivered; zone_status_timedwait_sig may return for any of
803 * the above reasons).
804 *
805 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and
806 * cv_wait_sig() respectively.
807 */
808 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t);
809 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t);
810 extern void zone_status_wait(zone_t *, zone_status_t);
811 extern int zone_status_wait_sig(zone_t *, zone_status_t);
812
813 /*
814 * Get the status of the zone (at the time it was called). The state may
815 * have progressed by the time it is returned.
816 */
817 extern zone_status_t zone_status_get(zone_t *);
818
819 /*
820 * Safely get the hostid of the specified zone (defaults to machine's hostid
821 * if the specified zone doesn't emulate a hostid). Passing NULL retrieves
822 * the global zone's (i.e., physical system's) hostid.
823 */
824 extern uint32_t zone_get_hostid(zone_t *);
825
826 /*
827 * Get the "kcred" credentials corresponding to the given zone.
828 */
829 extern struct cred *zone_get_kcred(zoneid_t);
830
831 /*
832 * Get/set the pool the zone is currently bound to.
833 */
834 extern struct pool *zone_pool_get(zone_t *);
835 extern void zone_pool_set(zone_t *, struct pool *);
836
837 /*
838 * Get/set the pset the zone is currently using.
839 */
840 extern psetid_t zone_pset_get(zone_t *);
841 extern void zone_pset_set(zone_t *, psetid_t);
842
843 /*
844 * Get the number of cpus/online-cpus visible from the given zone.
845 */
846 extern int zone_ncpus_get(zone_t *);
847 extern int zone_ncpus_online_get(zone_t *);
848
849 /*
850 * Returns true if the named pool/dataset is visible in the current zone.
851 */
852 extern int zone_dataset_visible(const char *, int *);
853
854 /*
855 * zone version of kadmin()
856 */
857 extern int zone_kadmin(int, int, const char *, cred_t *);
858 extern void zone_shutdown_global(void);
859
860 extern void mount_in_progress(zone_t *);
861 extern void mount_completed(zone_t *);
862
863 extern int zone_walk(int (*)(zone_t *, void *), void *);
864
865 extern rctl_hndl_t rc_zone_locked_mem;
866 extern rctl_hndl_t rc_zone_max_swap;
867 extern rctl_hndl_t rc_zone_max_lofi;
868
869 #endif /* _KERNEL */
870
871 #ifdef __cplusplus
872 }
873 #endif
874
875 #endif /* _SYS_ZONE_H */