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