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 * Copyright 2014 Igor Kozhukhov <ikozhukhov@gmail.com>.
25 * Copyright 2019 Nexenta Systems, Inc. All rights reserved.
26 * Copyright 2020 Joyent, Inc.
27 */
28
29 #ifndef _SYS_ZONE_H
30 #define _SYS_ZONE_H
31
32 #include <sys/types.h>
33 #include <sys/param.h>
34 #include <sys/tsol/label.h>
35 #include <sys/uadmin.h>
36 #include <netinet/in.h>
37
38 #ifdef _KERNEL
39 /*
40 * Many includes are kernel-only to reduce namespace pollution of
41 * userland applications.
42 */
43 #include <sys/mutex.h>
44 #include <sys/rctl.h>
45 #include <sys/ipc_rctl.h>
46 #include <sys/pset.h>
47 #include <sys/cred.h>
48 #include <sys/netstack.h>
49 #include <sys/ksynch.h>
50 #include <sys/socket_impl.h>
51 #include <sys/secflags.h>
52 #include <sys/cpu_uarray.h>
53 #include <sys/list.h>
54 #include <sys/loadavg.h>
55 #endif /* _KERNEL */
56
57 #ifdef __cplusplus
58 extern "C" {
59 #endif
60
61 /*
62 * NOTE
63 *
64 * The contents of this file are private to the implementation of
65 * Solaris and are subject to change at any time without notice.
66 * Applications and drivers using these interfaces may fail to
67 * run on future releases.
68 */
69
70 /* Available both in kernel and for user space */
71
72 /* zone id restrictions and special ids */
73 #define MAX_ZONEID 9999
74 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */
75 #define MIN_ZONEID 0 /* minimum zone ID on system */
76 #define GLOBAL_ZONEID 0
77 #define ZONEID_WIDTH 4 /* for printf */
78
79 /*
80 * Special zoneid_t token to refer to all zones.
81 */
82 #define ALL_ZONES (-1)
83
84 /* system call subcodes */
85 #define ZONE_CREATE 0
86 #define ZONE_DESTROY 1
87 #define ZONE_GETATTR 2
88 #define ZONE_ENTER 3
89 #define ZONE_LIST 4
90 #define ZONE_SHUTDOWN 5
91 #define ZONE_LOOKUP 6
92 #define ZONE_BOOT 7
93 #define ZONE_VERSION 8
94 #define ZONE_SETATTR 9
95 #define ZONE_ADD_DATALINK 10
96 #define ZONE_DEL_DATALINK 11
97 #define ZONE_CHECK_DATALINK 12
98 #define ZONE_LIST_DATALINK 13
99
100 /* zone attributes */
101 #define ZONE_ATTR_ROOT 1
102 #define ZONE_ATTR_NAME 2
103 #define ZONE_ATTR_STATUS 3
104 #define ZONE_ATTR_PRIVSET 4
105 #define ZONE_ATTR_UNIQID 5
106 #define ZONE_ATTR_POOLID 6
107 #define ZONE_ATTR_INITPID 7
108 #define ZONE_ATTR_SLBL 8
109 #define ZONE_ATTR_INITNAME 9
110 #define ZONE_ATTR_BOOTARGS 10
111 #define ZONE_ATTR_BRAND 11
112 #define ZONE_ATTR_PHYS_MCAP 12
113 #define ZONE_ATTR_SCHED_CLASS 13
114 #define ZONE_ATTR_FLAGS 14
115 #define ZONE_ATTR_HOSTID 15
116 #define ZONE_ATTR_FS_ALLOWED 16
117 #define ZONE_ATTR_NETWORK 17
118 #define ZONE_ATTR_INITNORESTART 20
119 #define ZONE_ATTR_SECFLAGS 21
120 #define ZONE_ATTR_DID 22
121
122 /* Start of the brand-specific attribute namespace */
123 #define ZONE_ATTR_BRAND_ATTRS 32768
124
125 #define ZONE_FS_ALLOWED_MAX 1024
126
127 #define ZONE_EVENT_CHANNEL "com.sun:zones:status"
128 #define ZONE_EVENT_STATUS_CLASS "status"
129 #define ZONE_EVENT_STATUS_SUBCLASS "change"
130
131 #define ZONE_EVENT_UNINITIALIZED "uninitialized"
132 #define ZONE_EVENT_INITIALIZED "initialized"
133 #define ZONE_EVENT_READY "ready"
134 #define ZONE_EVENT_RUNNING "running"
135 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down"
136
137 #define ZONE_CB_NAME "zonename"
138 #define ZONE_CB_NEWSTATE "newstate"
139 #define ZONE_CB_OLDSTATE "oldstate"
140 #define ZONE_CB_TIMESTAMP "when"
141 #define ZONE_CB_ZONEID "zoneid"
142
143 /*
144 * Exit values that may be returned by scripts or programs invoked by various
145 * zone commands.
146 *
147 * These are defined as:
148 *
149 * ZONE_SUBPROC_OK
150 * ===============
151 * The subprocess completed successfully.
152 *
153 * ZONE_SUBPROC_USAGE
154 * ==================
155 * The subprocess failed with a usage message, or a usage message should
156 * be output in its behalf.
157 *
158 * ZONE_SUBPROC_NOTCOMPLETE
159 * ========================
160 * The subprocess did not complete, but the actions performed by the
161 * subprocess require no recovery actions by the user.
162 *
163 * For example, if the subprocess were called by "zoneadm install," the
164 * installation of the zone did not succeed but the user need not perform
165 * a "zoneadm uninstall" before attempting another install.
166 *
167 * ZONE_SUBPROC_FATAL
168 * ==================
169 * The subprocess failed in a fatal manner, usually one that will require
170 * some type of recovery action by the user.
171 *
172 * For example, if the subprocess were called by "zoneadm install," the
173 * installation of the zone did not succeed and the user will need to
174 * perform a "zoneadm uninstall" before another install attempt is
175 * possible.
176 *
177 * The non-success exit values are large to avoid accidental collision
178 * with values used internally by some commands (e.g. "Z_ERR" and
179 * "Z_USAGE" as used by zoneadm.)
180 */
181 #define ZONE_SUBPROC_OK 0
182 #define ZONE_SUBPROC_USAGE 253
183 #define ZONE_SUBPROC_NOTCOMPLETE 254
184 #define ZONE_SUBPROC_FATAL 255
185
186 #ifdef _SYSCALL32
187 typedef struct {
188 caddr32_t zone_name;
189 caddr32_t zone_root;
190 caddr32_t zone_privs;
191 size32_t zone_privssz;
192 caddr32_t rctlbuf;
193 size32_t rctlbufsz;
194 caddr32_t extended_error;
195 caddr32_t zfsbuf;
196 size32_t zfsbufsz;
197 int match; /* match level */
198 uint32_t doi; /* DOI for label */
199 caddr32_t label; /* label associated with zone */
200 int flags;
201 zoneid_t zone_did; /* zone debug ID */
202 } zone_def32;
203 #endif
204 typedef struct {
205 const char *zone_name;
206 const char *zone_root;
207 const struct priv_set *zone_privs;
208 size_t zone_privssz;
209 const char *rctlbuf;
210 size_t rctlbufsz;
211 int *extended_error;
212 const char *zfsbuf;
213 size_t zfsbufsz;
214 int match; /* match level */
215 uint32_t doi; /* DOI for label */
216 const bslabel_t *label; /* label associated with zone */
217 int flags;
218 zoneid_t zone_did; /* zone debug ID */
219 } zone_def;
220
221 /* extended error information */
222 #define ZE_UNKNOWN 0 /* No extended error info */
223 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */
224 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */
225 #define ZE_LABELINUSE 3 /* label is already in use by some other zone */
226
227 /*
228 * zone_status values
229 *
230 * You must modify zone_status_names in mdb(1M)'s genunix module
231 * (genunix/zone.c) when you modify this enum.
232 */
233 typedef enum {
234 ZONE_IS_UNINITIALIZED = 0,
235 ZONE_IS_INITIALIZED,
236 ZONE_IS_READY,
237 ZONE_IS_BOOTING,
238 ZONE_IS_RUNNING,
239 ZONE_IS_SHUTTING_DOWN,
240 ZONE_IS_EMPTY,
241 ZONE_IS_DOWN,
242 ZONE_IS_DYING,
243 ZONE_IS_DEAD
244 } zone_status_t;
245 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED
246 #define ZONE_MAX_STATE ZONE_IS_DEAD
247
248 /*
249 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also
250 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT.
251 */
252 typedef enum zone_cmd {
253 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING,
254 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT, Z_SHUTDOWN
255 } zone_cmd_t;
256
257 /*
258 * The structure of a request to zoneadmd.
259 */
260 typedef struct zone_cmd_arg {
261 uint64_t uniqid; /* unique "generation number" */
262 zone_cmd_t cmd; /* requested action */
263 uint32_t _pad; /* need consistent 32/64 bit alignmt */
264 char locale[MAXPATHLEN]; /* locale in which to render messages */
265 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */
266 } zone_cmd_arg_t;
267
268 /*
269 * Structure of zoneadmd's response to a request. A NULL return value means
270 * the caller should attempt to restart zoneadmd and retry.
271 */
272 typedef struct zone_cmd_rval {
273 int rval; /* return value of request */
274 char errbuf[1]; /* variable-sized buffer containing error messages */
275 } zone_cmd_rval_t;
276
277 /*
278 * The zone support infrastructure uses the zone name as a component
279 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters
280 * in length, so ZONENAME_MAX is limited by that.
281 */
282 #define ZONENAME_MAX 64
283
284 #define GLOBAL_ZONENAME "global"
285
286 /*
287 * Extended Regular expression (see regex(5)) which matches all valid zone
288 * names.
289 */
290 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}"
291
292 /*
293 * Where the zones support infrastructure places temporary files.
294 */
295 #define ZONES_TMPDIR "/var/run/zones"
296
297 /*
298 * The path to the door used by clients to communicate with zoneadmd.
299 */
300 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door"
301
302
303 /* zone_flags */
304 /*
305 * Threads that read or write the following flag must hold zone_lock.
306 */
307 #define ZF_REFCOUNTS_LOGGED 0x1 /* a thread logged the zone's refs */
308
309 /*
310 * The following threads are set when the zone is created and never changed.
311 * Threads that test for these flags don't have to hold zone_lock.
312 */
313 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */
314 #define ZF_IS_SCRATCH 0x4 /* scratch zone */
315 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */
316
317
318 /* zone_create flags */
319 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */
320
321 /* zone network properties */
322 #define ZONE_NETWORK_ADDRESS 1
323 #define ZONE_NETWORK_DEFROUTER 2
324
325 #define ZONE_NET_ADDRNAME "address"
326 #define ZONE_NET_RTRNAME "route"
327
328 typedef struct zone_net_data {
329 int zn_type;
330 int zn_len;
331 datalink_id_t zn_linkid;
332 uint8_t zn_val[1];
333 } zone_net_data_t;
334
335
336 #ifdef _KERNEL
337
338 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */
339
340 /*
341 * Indexes into ->zone_ustate array, summing the micro state of all threads in a
342 * particular zone.
343 */
344 #define ZONE_USTATE_STIME (0)
345 #define ZONE_USTATE_UTIME (1)
346 #define ZONE_USTATE_WTIME (2)
347 #define ZONE_USTATE_MAX (3)
348
349 struct pool;
350 struct brand;
351
352 /*
353 * Each of these constants identifies a kernel subsystem that acquires and
354 * releases zone references. Each subsystem that invokes
355 * zone_hold_ref() and zone_rele_ref() should specify the
356 * zone_ref_subsys_t constant associated with the subsystem. Tracked holds
357 * help users and developers quickly identify subsystems that stall zone
358 * shutdowns indefinitely.
359 *
360 * NOTE: You must modify zone_ref_subsys_names in usr/src/uts/common/os/zone.c
361 * when you modify this enumeration.
362 */
363 typedef enum zone_ref_subsys {
364 ZONE_REF_NFS, /* NFS */
365 ZONE_REF_NFSV4, /* NFSv4 */
366 ZONE_REF_SMBFS, /* SMBFS */
367 ZONE_REF_MNTFS, /* MNTFS */
368 ZONE_REF_LOFI, /* LOFI devices */
369 ZONE_REF_VFS, /* VFS infrastructure */
370 ZONE_REF_IPC, /* IPC infrastructure */
371 ZONE_REF_NUM_SUBSYS /* This must be the last entry. */
372 } zone_ref_subsys_t;
373
374 /*
375 * zone_ref represents a general-purpose references to a zone. Each zone's
376 * references are linked into the zone's zone_t::zone_ref_list. This allows
377 * debuggers to walk zones' references.
378 */
379 typedef struct zone_ref {
380 struct zone *zref_zone; /* the zone to which the reference refers */
381 list_node_t zref_linkage; /* linkage for zone_t::zone_ref_list */
382 } zone_ref_t;
383
384 /*
385 * Structure to record list of ZFS datasets exported to a zone.
386 */
387 typedef struct zone_dataset {
388 char *zd_dataset;
389 list_node_t zd_linkage;
390 } zone_dataset_t;
391
392 /*
393 * structure for zone kstats
394 */
395 typedef struct zone_kstat {
396 kstat_named_t zk_zonename;
397 kstat_named_t zk_usage;
398 kstat_named_t zk_value;
399 } zone_kstat_t;
400
401 struct cpucap;
402
403 typedef struct {
404 kstat_named_t zm_zonename;
405 kstat_named_t zm_pgpgin;
406 kstat_named_t zm_anonpgin;
407 kstat_named_t zm_execpgin;
408 kstat_named_t zm_fspgin;
409 kstat_named_t zm_anon_alloc_fail;
410 } zone_mcap_kstat_t;
411
412 typedef struct {
413 kstat_named_t zm_zonename; /* full name, kstat truncates name */
414 kstat_named_t zm_utime;
415 kstat_named_t zm_stime;
416 kstat_named_t zm_wtime;
417 kstat_named_t zm_avenrun1;
418 kstat_named_t zm_avenrun5;
419 kstat_named_t zm_avenrun15;
420 kstat_named_t zm_ffcap;
421 kstat_named_t zm_ffnoproc;
422 kstat_named_t zm_ffnomem;
423 kstat_named_t zm_ffmisc;
424 kstat_named_t zm_nested_intp;
425 kstat_named_t zm_init_pid;
426 kstat_named_t zm_boot_time;
427 } zone_misc_kstat_t;
428
429 typedef struct zone {
430 /*
431 * zone_name is never modified once set.
432 */
433 char *zone_name; /* zone's configuration name */
434 /*
435 * zone_nodename and zone_domain are never freed once allocated.
436 */
437 char *zone_nodename; /* utsname.nodename equivalent */
438 char *zone_domain; /* srpc_domain equivalent */
439 /*
440 * zone_hostid is used for per-zone hostid emulation.
441 * Currently it isn't modified after it's set (so no locks protect
442 * accesses), but that might have to change when we allow
443 * administrators to change running zones' properties.
444 *
445 * The global zone's zone_hostid must always be HW_INVALID_HOSTID so
446 * that zone_get_hostid() will function correctly.
447 */
448 uint32_t zone_hostid; /* zone's hostid, HW_INVALID_HOSTID */
449 /* if not emulated */
450 /*
451 * zone_lock protects the following fields of a zone_t:
452 * zone_ref
453 * zone_cred_ref
454 * zone_subsys_ref
455 * zone_ref_list
456 * zone_ntasks
457 * zone_flags
458 * zone_zsd
459 * zone_pfexecd
460 */
461 kmutex_t zone_lock;
462 /*
463 * zone_linkage is the zone's linkage into the active or
464 * death-row list. The field is protected by zonehash_lock.
465 */
466 list_node_t zone_linkage;
467 zoneid_t zone_id; /* ID of zone */
468 zoneid_t zone_did; /* persistent debug ID of zone */
469 uint_t zone_ref; /* count of zone_hold()s on zone */
470 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */
471 /*
472 * Fixed-sized array of subsystem-specific reference counts
473 * The sum of all of the counts must be less than or equal to zone_ref.
474 * The array is indexed by the counts' subsystems' zone_ref_subsys_t
475 * constants.
476 */
477 uint_t zone_subsys_ref[ZONE_REF_NUM_SUBSYS];
478 list_t zone_ref_list; /* list of zone_ref_t structs */
479 /*
480 * zone_rootvp and zone_rootpath can never be modified once set.
481 */
482 struct vnode *zone_rootvp; /* zone's root vnode */
483 char *zone_rootpath; /* Path to zone's root + '/' */
484 ushort_t zone_flags; /* misc flags */
485 zone_status_t zone_status; /* protected by zone_status_lock */
486 uint_t zone_ntasks; /* number of tasks executing in zone */
487 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */
488 /* counters in projects and tasks */
489 /* that are within the zone */
490 rctl_qty_t zone_nlwps; /* number of lwps in zone */
491 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */
492 rctl_qty_t zone_shmmax; /* System V shared memory usage */
493 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */
494
495 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */
496 uint32_t zone_shares; /* FSS shares allocated to zone */
497 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */
498 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */
499 /* kpd_locked_mem for all */
500 /* projects in zone. */
501 /* Also protects zone_max_swap */
502 /* grab after p_lock, before rcs_lock */
503 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */
504 /* zone */
505 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */
506 /* limit. Protected by */
507 /* zone_rctls->rcs_lock */
508 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */
509 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */
510 /* Protected by */
511 /* zone_rctls->rcs_lock */
512 kmutex_t zone_rctl_lock; /* protects zone_max_lofi */
513 rctl_qty_t zone_max_lofi; /* lofi devs for zone */
514 rctl_qty_t zone_max_lofi_ctl; /* current lofi limit. */
515 /* Protected by */
516 /* zone_rctls->rcs_lock */
517 list_t zone_zsd; /* list of Zone-Specific Data values */
518 kcondvar_t zone_cv; /* used to signal state changes */
519 struct proc *zone_zsched; /* Dummy kernel "zsched" process */
520 pid_t zone_proc_initpid; /* pid of "init" for this zone */
521 char *zone_initname; /* fs path to 'init' */
522 int zone_boot_err; /* for zone_boot() if boot fails */
523 char *zone_bootargs; /* arguments passed via zone_boot() */
524 uint64_t zone_phys_mcap; /* physical memory cap */
525 /*
526 * zone_kthreads is protected by zone_status_lock.
527 */
528 kthread_t *zone_kthreads; /* kernel threads in zone */
529 struct priv_set *zone_privset; /* limit set for zone */
530 /*
531 * zone_vfslist is protected by vfs_list_lock().
532 */
533 struct vfs *zone_vfslist; /* list of FS's mounted in zone */
534 uint64_t zone_uniqid; /* unique zone generation number */
535 struct cred *zone_kcred; /* kcred-like, zone-limited cred */
536 /*
537 * zone_pool is protected by pool_lock().
538 */
539 struct pool *zone_pool; /* pool the zone is bound to */
540 hrtime_t zone_pool_mod; /* last pool bind modification time */
541 /* zone_psetid is protected by cpu_lock */
542 psetid_t zone_psetid; /* pset the zone is bound to */
543
544 time_t zone_boot_time; /* Similar to boot_time */
545
546 /*
547 * The following two can be read without holding any locks. They are
548 * updated under cpu_lock.
549 */
550 int zone_ncpus; /* zone's idea of ncpus */
551 int zone_ncpus_online; /* zone's idea of ncpus_online */
552 /*
553 * List of ZFS datasets exported to this zone.
554 */
555 list_t zone_datasets; /* list of datasets */
556
557 ts_label_t *zone_slabel; /* zone sensitivity label */
558 int zone_match; /* require label match for packets */
559 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */
560
561 boolean_t zone_restart_init; /* Restart init if it dies? */
562 struct brand *zone_brand; /* zone's brand */
563 void *zone_brand_data; /* store brand specific data */
564 id_t zone_defaultcid; /* dflt scheduling class id */
565 kstat_t *zone_swapresv_kstat;
566 kstat_t *zone_lockedmem_kstat;
567 /*
568 * zone_dl_list is protected by zone_lock
569 */
570 list_t zone_dl_list;
571 netstack_t *zone_netstack;
572 struct cpucap *zone_cpucap; /* CPU caps data */
573 /*
574 * Solaris Auditing per-zone audit context
575 */
576 struct au_kcontext *zone_audit_kctxt;
577 /*
578 * For private use by mntfs.
579 */
580 struct mntelem *zone_mntfs_db;
581 krwlock_t zone_mntfs_db_lock;
582
583 struct klpd_reg *zone_pfexecd;
584
585 char *zone_fs_allowed;
586 rctl_qty_t zone_nprocs; /* number of processes in the zone */
587 rctl_qty_t zone_nprocs_ctl; /* current limit protected by */
588 /* zone_rctls->rcs_lock */
589 kstat_t *zone_nprocs_kstat;
590
591 kmutex_t zone_mcap_lock; /* protects mcap statistics */
592 kstat_t *zone_mcap_ksp;
593 zone_mcap_kstat_t *zone_mcap_stats;
594 uint64_t zone_pgpgin; /* pages paged in */
595 uint64_t zone_anonpgin; /* anon pages paged in */
596 uint64_t zone_execpgin; /* exec pages paged in */
597 uint64_t zone_fspgin; /* fs pages paged in */
598 uint64_t zone_anon_alloc_fail; /* cnt of anon alloc fails */
599
600 psecflags_t zone_secflags; /* default zone security-flags */
601
602 /*
603 * Misc. kstats and counters for zone cpu-usage aggregation.
604 */
605 kmutex_t zone_misc_lock; /* protects misc statistics */
606 kstat_t *zone_misc_ksp;
607 zone_misc_kstat_t *zone_misc_stats;
608 /* Accumulated microstate for all threads in this zone. */
609 cpu_uarray_t *zone_ustate;
610 /* fork-fail kstat tracking */
611 uint32_t zone_ffcap; /* hit an rctl cap */
612 uint32_t zone_ffnoproc; /* get proc/lwp error */
613 uint32_t zone_ffnomem; /* as_dup/memory error */
614 uint32_t zone_ffmisc; /* misc. other error */
615
616 uint32_t zone_nested_intp; /* nested interp. kstat */
617
618 struct loadavg_s zone_loadavg; /* loadavg for this zone */
619 uint64_t zone_hp_avenrun[3]; /* high-precision avenrun */
620 int zone_avenrun[3]; /* FSCALED avg. run queue len */
621
622 /*
623 * FSS stats updated once per second by fss_decay_usage.
624 */
625 uint32_t zone_fss_gen; /* FSS generation cntr */
626 uint64_t zone_run_ticks; /* tot # of ticks running */
627
628 /*
629 * DTrace-private per-zone state
630 */
631 int zone_dtrace_getf; /* # of unprivileged getf()s */
632
633 /*
634 * Synchronization primitives used to synchronize between mounts and
635 * zone creation/destruction.
636 */
637 int zone_mounts_in_progress;
638 kcondvar_t zone_mount_cv;
639 kmutex_t zone_mount_lock;
640 } zone_t;
641
642 /*
643 * Special value of zone_psetid to indicate that pools are disabled.
644 */
645 #define ZONE_PS_INVAL PS_MYID
646
647 extern zone_t zone0;
648 extern zone_t *global_zone;
649 extern uint_t maxzones;
650 extern rctl_hndl_t rc_zone_nlwps;
651 extern rctl_hndl_t rc_zone_nprocs;
652
653 extern long zone(int, void *, void *, void *, void *);
654 extern void zone_zsd_init(void);
655 extern void zone_init(void);
656 extern void zone_hold(zone_t *);
657 extern void zone_rele(zone_t *);
658 extern void zone_init_ref(zone_ref_t *);
659 extern void zone_hold_ref(zone_t *, zone_ref_t *, zone_ref_subsys_t);
660 extern void zone_rele_ref(zone_ref_t *, zone_ref_subsys_t);
661 extern void zone_cred_hold(zone_t *);
662 extern void zone_cred_rele(zone_t *);
663 extern void zone_task_hold(zone_t *);
664 extern void zone_task_rele(zone_t *);
665 extern zone_t *zone_find_by_id(zoneid_t);
666 extern zone_t *zone_find_by_label(const ts_label_t *);
667 extern zone_t *zone_find_by_name(char *);
668 extern zone_t *zone_find_by_any_path(const char *, boolean_t);
669 extern zone_t *zone_find_by_path(const char *);
670 extern zoneid_t getzoneid(void);
671 extern zoneid_t getzonedid(void);
672 extern zone_t *zone_find_by_id_nolock(zoneid_t);
673 extern int zone_datalink_walk(zoneid_t, int (*)(datalink_id_t, void *), void *);
674 extern int zone_check_datalink(zoneid_t *, datalink_id_t);
675 extern void zone_loadavg_update();
676
677 /*
678 * Zone-specific data (ZSD) APIs
679 */
680 /*
681 * The following is what code should be initializing its zone_key_t to if it
682 * calls zone_getspecific() without necessarily knowing that zone_key_create()
683 * has been called on the key.
684 */
685 #define ZONE_KEY_UNINITIALIZED 0
686
687 typedef uint_t zone_key_t;
688
689 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t),
690 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *));
691 extern int zone_key_delete(zone_key_t);
692 extern void *zone_getspecific(zone_key_t, zone_t *);
693 extern int zone_setspecific(zone_key_t, zone_t *, const void *);
694
695 /*
696 * The definition of a zsd_entry is truly private to zone.c and is only
697 * placed here so it can be shared with mdb.
698 *
699 * State maintained for each zone times each registered key, which tracks
700 * the state of the create, shutdown and destroy callbacks.
701 *
702 * zsd_flags is used to keep track of pending actions to avoid holding locks
703 * when calling the create/shutdown/destroy callbacks, since doing so
704 * could lead to deadlocks.
705 */
706 struct zsd_entry {
707 zone_key_t zsd_key; /* Key used to lookup value */
708 void *zsd_data; /* Caller-managed value */
709 /*
710 * Callbacks to be executed when a zone is created, shutdown, and
711 * destroyed, respectively.
712 */
713 void *(*zsd_create)(zoneid_t);
714 void (*zsd_shutdown)(zoneid_t, void *);
715 void (*zsd_destroy)(zoneid_t, void *);
716 list_node_t zsd_linkage;
717 uint16_t zsd_flags; /* See below */
718 kcondvar_t zsd_cv;
719 };
720
721 /*
722 * zsd_flags
723 */
724 #define ZSD_CREATE_NEEDED 0x0001
725 #define ZSD_CREATE_INPROGRESS 0x0002
726 #define ZSD_CREATE_COMPLETED 0x0004
727 #define ZSD_SHUTDOWN_NEEDED 0x0010
728 #define ZSD_SHUTDOWN_INPROGRESS 0x0020
729 #define ZSD_SHUTDOWN_COMPLETED 0x0040
730 #define ZSD_DESTROY_NEEDED 0x0100
731 #define ZSD_DESTROY_INPROGRESS 0x0200
732 #define ZSD_DESTROY_COMPLETED 0x0400
733
734 #define ZSD_CREATE_ALL \
735 (ZSD_CREATE_NEEDED|ZSD_CREATE_INPROGRESS|ZSD_CREATE_COMPLETED)
736 #define ZSD_SHUTDOWN_ALL \
737 (ZSD_SHUTDOWN_NEEDED|ZSD_SHUTDOWN_INPROGRESS|ZSD_SHUTDOWN_COMPLETED)
738 #define ZSD_DESTROY_ALL \
739 (ZSD_DESTROY_NEEDED|ZSD_DESTROY_INPROGRESS|ZSD_DESTROY_COMPLETED)
740
741 #define ZSD_ALL_INPROGRESS \
742 (ZSD_CREATE_INPROGRESS|ZSD_SHUTDOWN_INPROGRESS|ZSD_DESTROY_INPROGRESS)
743
744 /*
745 * Macros to help with zone visibility restrictions.
746 */
747
748 /*
749 * Is process in the global zone?
750 */
751 #define INGLOBALZONE(p) \
752 ((p)->p_zone == global_zone)
753
754 /*
755 * Can process view objects in given zone?
756 */
757 #define HASZONEACCESS(p, zoneid) \
758 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p))
759
760 /*
761 * Convenience macro to see if a resolved path is visible from within a
762 * given zone.
763 *
764 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the
765 * two strings must be equal. Since the rootpathlen has a trailing '/',
766 * we want to skip everything in the path up to (but not including) the
767 * trailing '/'.
768 */
769 #define ZONE_PATH_VISIBLE(path, zone) \
770 (strncmp((path), (zone)->zone_rootpath, \
771 (zone)->zone_rootpathlen - 1) == 0)
772
773 /*
774 * Convenience macro to go from the global view of a path to that seen
775 * from within said zone. It is the responsibility of the caller to
776 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is
777 * in fact visible from within the zone.
778 */
779 #define ZONE_PATH_TRANSLATE(path, zone) \
780 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \
781 (path) + (zone)->zone_rootpathlen - 2)
782
783 /*
784 * Special processes visible in all zones.
785 */
786 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1)
787
788 /*
789 * A root vnode of the current zone.
790 *
791 * NOTE: It may be necessary (initialization time for file sharing where an
792 * NGZ loads a file-sharing kernel module that does zsd initialization) to NOT
793 * use this macro. One should ASSERT() that curzone == active ZSD (an
794 * ASSERTion that's not always true at ZSD initialization time) during regular
795 * use of this macro.
796 */
797 #define ZONE_ROOTVP() (curzone->zone_rootvp)
798
799 /*
800 * Since a zone's root isn't necessarily an actual filesystem boundary
801 * (i.e. VROOT may not be set on zone->zone_rootvp) we need to not assume it.
802 * This macro helps in checking if a vnode is the current zone's rootvp.
803 * NOTE: Using the VN_ prefix, even though it's defined here in zone.h.
804 * NOTE2: See above warning about ZONE_ROOTVP().
805 */
806 #define VN_IS_CURZONEROOT(vp) (VN_CMP(vp, ZONE_ROOTVP()))
807
808 /*
809 * Zone-safe version of thread_create() to be used when the caller wants to
810 * create a kernel thread to run within the current zone's context.
811 */
812 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t,
813 pri_t);
814 extern void zthread_exit(void);
815
816 /*
817 * Functions for an external observer to register interest in a zone's status
818 * change. Observers will be woken up when the zone status equals the status
819 * argument passed in (in the case of zone_status_timedwait, the function may
820 * also return because of a timeout; zone_status_wait_sig may return early due
821 * to a signal being delivered; zone_status_timedwait_sig may return for any of
822 * the above reasons).
823 *
824 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and
825 * cv_wait_sig() respectively.
826 */
827 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t);
828 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t);
829 extern void zone_status_wait(zone_t *, zone_status_t);
830 extern int zone_status_wait_sig(zone_t *, zone_status_t);
831
832 /*
833 * Get the status of the zone (at the time it was called). The state may
834 * have progressed by the time it is returned.
835 */
836 extern zone_status_t zone_status_get(zone_t *);
837
838 /*
839 * Safely get the hostid of the specified zone (defaults to machine's hostid
840 * if the specified zone doesn't emulate a hostid). Passing NULL retrieves
841 * the global zone's (i.e., physical system's) hostid.
842 */
843 extern uint32_t zone_get_hostid(zone_t *);
844
845 /*
846 * Get the "kcred" credentials corresponding to the given zone.
847 */
848 extern struct cred *zone_get_kcred(zoneid_t);
849
850 /*
851 * Get/set the pool the zone is currently bound to.
852 */
853 extern struct pool *zone_pool_get(zone_t *);
854 extern void zone_pool_set(zone_t *, struct pool *);
855
856 /*
857 * Get/set the pset the zone is currently using.
858 */
859 extern psetid_t zone_pset_get(zone_t *);
860 extern void zone_pset_set(zone_t *, psetid_t);
861
862 /*
863 * Get the number of cpus/online-cpus visible from the given zone.
864 */
865 extern int zone_ncpus_get(zone_t *);
866 extern int zone_ncpus_online_get(zone_t *);
867
868 /*
869 * Returns true if the named pool/dataset is visible in the current zone.
870 */
871 extern int zone_dataset_visible(const char *, int *);
872
873 /*
874 * zone version of kadmin()
875 */
876 extern int zone_kadmin(int, int, const char *, cred_t *);
877 extern void zone_shutdown_global(void);
878
879 extern void mount_in_progress(zone_t *);
880 extern void mount_completed(zone_t *);
881
882 extern int zone_walk(int (*)(zone_t *, void *), void *);
883
884 extern rctl_hndl_t rc_zone_locked_mem;
885 extern rctl_hndl_t rc_zone_max_swap;
886 extern rctl_hndl_t rc_zone_max_lofi;
887
888 #endif /* _KERNEL */
889
890 #ifdef __cplusplus
891 }
892 #endif
893
894 #endif /* _SYS_ZONE_H */