Print this page
OS-208 DTrace needs to use zone_did to match zone-limited enablings
OS-192 zone_create() warning on headnode
| Split |
Close |
| Expand all |
| Collapse all |
--- old/usr/src/uts/common/sys/zone.h
+++ new/usr/src/uts/common/sys/zone.h
1 1 /*
2 2 * CDDL HEADER START
3 3 *
4 4 * The contents of this file are subject to the terms of the
5 5 * Common Development and Distribution License (the "License").
6 6 * You may not use this file except in compliance with the License.
7 7 *
8 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 9 * or http://www.opensolaris.org/os/licensing.
10 10 * See the License for the specific language governing permissions
11 11 * and limitations under the License.
12 12 *
13 13 * When distributing Covered Code, include this CDDL HEADER in each
14 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 15 * If applicable, add the following below this CDDL HEADER, with the
16 16 * fields enclosed by brackets "[]" replaced with your own identifying
17 17 * information: Portions Copyright [yyyy] [name of copyright owner]
18 18 *
19 19 * CDDL HEADER END
20 20 */
21 21
22 22 /*
23 23 * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
24 24 * Copyright 2014 Igor Kozhukhov <ikozhukhov@gmail.com>.
25 25 * Copyright 2019 Nexenta Systems, Inc. All rights reserved.
26 26 * Copyright 2020 Joyent, Inc.
27 27 */
28 28
29 29 #ifndef _SYS_ZONE_H
30 30 #define _SYS_ZONE_H
31 31
32 32 #include <sys/types.h>
33 33 #include <sys/param.h>
34 34 #include <sys/tsol/label.h>
35 35 #include <sys/uadmin.h>
36 36 #include <netinet/in.h>
37 37
38 38 #ifdef _KERNEL
39 39 /*
40 40 * Many includes are kernel-only to reduce namespace pollution of
41 41 * userland applications.
42 42 */
43 43 #include <sys/mutex.h>
44 44 #include <sys/rctl.h>
45 45 #include <sys/ipc_rctl.h>
46 46 #include <sys/pset.h>
47 47 #include <sys/cred.h>
48 48 #include <sys/netstack.h>
49 49 #include <sys/ksynch.h>
50 50 #include <sys/socket_impl.h>
51 51 #include <sys/secflags.h>
52 52 #include <sys/cpu_uarray.h>
53 53 #include <sys/list.h>
54 54 #include <sys/loadavg.h>
55 55 #endif /* _KERNEL */
56 56
57 57 #ifdef __cplusplus
58 58 extern "C" {
59 59 #endif
60 60
61 61 /*
62 62 * NOTE
63 63 *
64 64 * The contents of this file are private to the implementation of
65 65 * Solaris and are subject to change at any time without notice.
66 66 * Applications and drivers using these interfaces may fail to
67 67 * run on future releases.
68 68 */
69 69
70 70 /* Available both in kernel and for user space */
71 71
72 72 /* zone id restrictions and special ids */
73 73 #define MAX_ZONEID 9999
74 74 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */
75 75 #define MIN_ZONEID 0 /* minimum zone ID on system */
76 76 #define GLOBAL_ZONEID 0
77 77 #define ZONEID_WIDTH 4 /* for printf */
78 78
79 79 /*
80 80 * Special zoneid_t token to refer to all zones.
81 81 */
82 82 #define ALL_ZONES (-1)
83 83
84 84 /* system call subcodes */
85 85 #define ZONE_CREATE 0
86 86 #define ZONE_DESTROY 1
87 87 #define ZONE_GETATTR 2
88 88 #define ZONE_ENTER 3
89 89 #define ZONE_LIST 4
90 90 #define ZONE_SHUTDOWN 5
91 91 #define ZONE_LOOKUP 6
92 92 #define ZONE_BOOT 7
93 93 #define ZONE_VERSION 8
94 94 #define ZONE_SETATTR 9
95 95 #define ZONE_ADD_DATALINK 10
96 96 #define ZONE_DEL_DATALINK 11
97 97 #define ZONE_CHECK_DATALINK 12
98 98 #define ZONE_LIST_DATALINK 13
99 99
100 100 /* zone attributes */
101 101 #define ZONE_ATTR_ROOT 1
102 102 #define ZONE_ATTR_NAME 2
103 103 #define ZONE_ATTR_STATUS 3
104 104 #define ZONE_ATTR_PRIVSET 4
105 105 #define ZONE_ATTR_UNIQID 5
106 106 #define ZONE_ATTR_POOLID 6
107 107 #define ZONE_ATTR_INITPID 7
108 108 #define ZONE_ATTR_SLBL 8
109 109 #define ZONE_ATTR_INITNAME 9
|
↓ open down ↓ |
109 lines elided |
↑ open up ↑ |
110 110 #define ZONE_ATTR_BOOTARGS 10
111 111 #define ZONE_ATTR_BRAND 11
112 112 #define ZONE_ATTR_PHYS_MCAP 12
113 113 #define ZONE_ATTR_SCHED_CLASS 13
114 114 #define ZONE_ATTR_FLAGS 14
115 115 #define ZONE_ATTR_HOSTID 15
116 116 #define ZONE_ATTR_FS_ALLOWED 16
117 117 #define ZONE_ATTR_NETWORK 17
118 118 #define ZONE_ATTR_INITNORESTART 20
119 119 #define ZONE_ATTR_SECFLAGS 21
120 +#define ZONE_ATTR_DID 22
120 121
121 122 /* Start of the brand-specific attribute namespace */
122 123 #define ZONE_ATTR_BRAND_ATTRS 32768
123 124
124 125 #define ZONE_FS_ALLOWED_MAX 1024
125 126
126 127 #define ZONE_EVENT_CHANNEL "com.sun:zones:status"
127 128 #define ZONE_EVENT_STATUS_CLASS "status"
128 129 #define ZONE_EVENT_STATUS_SUBCLASS "change"
129 130
130 131 #define ZONE_EVENT_UNINITIALIZED "uninitialized"
131 132 #define ZONE_EVENT_INITIALIZED "initialized"
132 133 #define ZONE_EVENT_READY "ready"
133 134 #define ZONE_EVENT_RUNNING "running"
134 135 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down"
135 136
136 137 #define ZONE_CB_NAME "zonename"
137 138 #define ZONE_CB_NEWSTATE "newstate"
138 139 #define ZONE_CB_OLDSTATE "oldstate"
139 140 #define ZONE_CB_TIMESTAMP "when"
140 141 #define ZONE_CB_ZONEID "zoneid"
141 142
142 143 /*
143 144 * Exit values that may be returned by scripts or programs invoked by various
144 145 * zone commands.
145 146 *
146 147 * These are defined as:
147 148 *
148 149 * ZONE_SUBPROC_OK
149 150 * ===============
150 151 * The subprocess completed successfully.
151 152 *
152 153 * ZONE_SUBPROC_USAGE
153 154 * ==================
154 155 * The subprocess failed with a usage message, or a usage message should
155 156 * be output in its behalf.
156 157 *
157 158 * ZONE_SUBPROC_NOTCOMPLETE
158 159 * ========================
159 160 * The subprocess did not complete, but the actions performed by the
160 161 * subprocess require no recovery actions by the user.
161 162 *
162 163 * For example, if the subprocess were called by "zoneadm install," the
163 164 * installation of the zone did not succeed but the user need not perform
164 165 * a "zoneadm uninstall" before attempting another install.
165 166 *
166 167 * ZONE_SUBPROC_FATAL
167 168 * ==================
168 169 * The subprocess failed in a fatal manner, usually one that will require
169 170 * some type of recovery action by the user.
170 171 *
171 172 * For example, if the subprocess were called by "zoneadm install," the
172 173 * installation of the zone did not succeed and the user will need to
173 174 * perform a "zoneadm uninstall" before another install attempt is
174 175 * possible.
175 176 *
176 177 * The non-success exit values are large to avoid accidental collision
177 178 * with values used internally by some commands (e.g. "Z_ERR" and
178 179 * "Z_USAGE" as used by zoneadm.)
179 180 */
180 181 #define ZONE_SUBPROC_OK 0
181 182 #define ZONE_SUBPROC_USAGE 253
182 183 #define ZONE_SUBPROC_NOTCOMPLETE 254
183 184 #define ZONE_SUBPROC_FATAL 255
184 185
185 186 #ifdef _SYSCALL32
186 187 typedef struct {
187 188 caddr32_t zone_name;
188 189 caddr32_t zone_root;
189 190 caddr32_t zone_privs;
|
↓ open down ↓ |
60 lines elided |
↑ open up ↑ |
190 191 size32_t zone_privssz;
191 192 caddr32_t rctlbuf;
192 193 size32_t rctlbufsz;
193 194 caddr32_t extended_error;
194 195 caddr32_t zfsbuf;
195 196 size32_t zfsbufsz;
196 197 int match; /* match level */
197 198 uint32_t doi; /* DOI for label */
198 199 caddr32_t label; /* label associated with zone */
199 200 int flags;
201 + zoneid_t zone_did; /* zone debug ID */
200 202 } zone_def32;
201 203 #endif
202 204 typedef struct {
203 205 const char *zone_name;
204 206 const char *zone_root;
205 207 const struct priv_set *zone_privs;
206 208 size_t zone_privssz;
207 209 const char *rctlbuf;
208 210 size_t rctlbufsz;
209 211 int *extended_error;
210 212 const char *zfsbuf;
211 213 size_t zfsbufsz;
212 214 int match; /* match level */
213 215 uint32_t doi; /* DOI for label */
214 216 const bslabel_t *label; /* label associated with zone */
215 217 int flags;
218 + zoneid_t zone_did; /* zone debug ID */
216 219 } zone_def;
217 220
218 221 /* extended error information */
219 222 #define ZE_UNKNOWN 0 /* No extended error info */
220 223 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */
221 224 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */
222 225 #define ZE_LABELINUSE 3 /* label is already in use by some other zone */
223 226
224 227 /*
225 228 * zone_status values
226 229 *
227 230 * You must modify zone_status_names in mdb(1M)'s genunix module
228 231 * (genunix/zone.c) when you modify this enum.
229 232 */
230 233 typedef enum {
231 234 ZONE_IS_UNINITIALIZED = 0,
232 235 ZONE_IS_INITIALIZED,
233 236 ZONE_IS_READY,
234 237 ZONE_IS_BOOTING,
235 238 ZONE_IS_RUNNING,
236 239 ZONE_IS_SHUTTING_DOWN,
237 240 ZONE_IS_EMPTY,
238 241 ZONE_IS_DOWN,
239 242 ZONE_IS_DYING,
240 243 ZONE_IS_DEAD
241 244 } zone_status_t;
242 245 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED
243 246 #define ZONE_MAX_STATE ZONE_IS_DEAD
244 247
245 248 /*
246 249 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also
247 250 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT.
248 251 */
249 252 typedef enum zone_cmd {
250 253 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING,
251 254 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT, Z_SHUTDOWN
252 255 } zone_cmd_t;
253 256
254 257 /*
255 258 * The structure of a request to zoneadmd.
256 259 */
257 260 typedef struct zone_cmd_arg {
258 261 uint64_t uniqid; /* unique "generation number" */
259 262 zone_cmd_t cmd; /* requested action */
260 263 uint32_t _pad; /* need consistent 32/64 bit alignmt */
261 264 char locale[MAXPATHLEN]; /* locale in which to render messages */
262 265 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */
263 266 } zone_cmd_arg_t;
264 267
265 268 /*
266 269 * Structure of zoneadmd's response to a request. A NULL return value means
267 270 * the caller should attempt to restart zoneadmd and retry.
268 271 */
269 272 typedef struct zone_cmd_rval {
270 273 int rval; /* return value of request */
271 274 char errbuf[1]; /* variable-sized buffer containing error messages */
272 275 } zone_cmd_rval_t;
273 276
274 277 /*
275 278 * The zone support infrastructure uses the zone name as a component
276 279 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters
277 280 * in length, so ZONENAME_MAX is limited by that.
278 281 */
279 282 #define ZONENAME_MAX 64
280 283
281 284 #define GLOBAL_ZONENAME "global"
282 285
283 286 /*
284 287 * Extended Regular expression (see regex(5)) which matches all valid zone
285 288 * names.
286 289 */
287 290 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}"
288 291
289 292 /*
290 293 * Where the zones support infrastructure places temporary files.
291 294 */
292 295 #define ZONES_TMPDIR "/var/run/zones"
293 296
294 297 /*
295 298 * The path to the door used by clients to communicate with zoneadmd.
296 299 */
297 300 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door"
298 301
299 302
300 303 /* zone_flags */
301 304 /*
302 305 * Threads that read or write the following flag must hold zone_lock.
303 306 */
304 307 #define ZF_REFCOUNTS_LOGGED 0x1 /* a thread logged the zone's refs */
305 308
306 309 /*
307 310 * The following threads are set when the zone is created and never changed.
308 311 * Threads that test for these flags don't have to hold zone_lock.
309 312 */
310 313 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */
311 314 #define ZF_IS_SCRATCH 0x4 /* scratch zone */
312 315 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */
313 316
314 317
315 318 /* zone_create flags */
316 319 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */
317 320
318 321 /* zone network properties */
319 322 #define ZONE_NETWORK_ADDRESS 1
320 323 #define ZONE_NETWORK_DEFROUTER 2
321 324
322 325 #define ZONE_NET_ADDRNAME "address"
323 326 #define ZONE_NET_RTRNAME "route"
324 327
325 328 typedef struct zone_net_data {
326 329 int zn_type;
327 330 int zn_len;
328 331 datalink_id_t zn_linkid;
329 332 uint8_t zn_val[1];
330 333 } zone_net_data_t;
331 334
332 335
333 336 #ifdef _KERNEL
334 337
335 338 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */
336 339
337 340 /*
338 341 * Indexes into ->zone_ustate array, summing the micro state of all threads in a
339 342 * particular zone.
340 343 */
341 344 #define ZONE_USTATE_STIME (0)
342 345 #define ZONE_USTATE_UTIME (1)
343 346 #define ZONE_USTATE_WTIME (2)
344 347 #define ZONE_USTATE_MAX (3)
345 348
346 349 struct pool;
347 350 struct brand;
348 351
349 352 /*
350 353 * Each of these constants identifies a kernel subsystem that acquires and
351 354 * releases zone references. Each subsystem that invokes
352 355 * zone_hold_ref() and zone_rele_ref() should specify the
353 356 * zone_ref_subsys_t constant associated with the subsystem. Tracked holds
354 357 * help users and developers quickly identify subsystems that stall zone
355 358 * shutdowns indefinitely.
356 359 *
357 360 * NOTE: You must modify zone_ref_subsys_names in usr/src/uts/common/os/zone.c
358 361 * when you modify this enumeration.
359 362 */
360 363 typedef enum zone_ref_subsys {
361 364 ZONE_REF_NFS, /* NFS */
362 365 ZONE_REF_NFSV4, /* NFSv4 */
363 366 ZONE_REF_SMBFS, /* SMBFS */
364 367 ZONE_REF_MNTFS, /* MNTFS */
365 368 ZONE_REF_LOFI, /* LOFI devices */
366 369 ZONE_REF_VFS, /* VFS infrastructure */
367 370 ZONE_REF_IPC, /* IPC infrastructure */
368 371 ZONE_REF_NUM_SUBSYS /* This must be the last entry. */
369 372 } zone_ref_subsys_t;
370 373
371 374 /*
372 375 * zone_ref represents a general-purpose references to a zone. Each zone's
373 376 * references are linked into the zone's zone_t::zone_ref_list. This allows
374 377 * debuggers to walk zones' references.
375 378 */
376 379 typedef struct zone_ref {
377 380 struct zone *zref_zone; /* the zone to which the reference refers */
378 381 list_node_t zref_linkage; /* linkage for zone_t::zone_ref_list */
379 382 } zone_ref_t;
380 383
381 384 /*
382 385 * Structure to record list of ZFS datasets exported to a zone.
383 386 */
384 387 typedef struct zone_dataset {
385 388 char *zd_dataset;
386 389 list_node_t zd_linkage;
387 390 } zone_dataset_t;
388 391
389 392 /*
390 393 * structure for zone kstats
391 394 */
392 395 typedef struct zone_kstat {
393 396 kstat_named_t zk_zonename;
394 397 kstat_named_t zk_usage;
395 398 kstat_named_t zk_value;
396 399 } zone_kstat_t;
397 400
398 401 struct cpucap;
399 402
400 403 typedef struct {
401 404 kstat_named_t zm_zonename;
402 405 kstat_named_t zm_pgpgin;
403 406 kstat_named_t zm_anonpgin;
404 407 kstat_named_t zm_execpgin;
405 408 kstat_named_t zm_fspgin;
406 409 kstat_named_t zm_anon_alloc_fail;
407 410 } zone_mcap_kstat_t;
408 411
409 412 typedef struct {
410 413 kstat_named_t zm_zonename; /* full name, kstat truncates name */
411 414 kstat_named_t zm_utime;
412 415 kstat_named_t zm_stime;
413 416 kstat_named_t zm_wtime;
414 417 kstat_named_t zm_avenrun1;
415 418 kstat_named_t zm_avenrun5;
416 419 kstat_named_t zm_avenrun15;
417 420 kstat_named_t zm_ffcap;
418 421 kstat_named_t zm_ffnoproc;
419 422 kstat_named_t zm_ffnomem;
420 423 kstat_named_t zm_ffmisc;
421 424 kstat_named_t zm_nested_intp;
422 425 kstat_named_t zm_init_pid;
423 426 kstat_named_t zm_boot_time;
424 427 } zone_misc_kstat_t;
425 428
426 429 typedef struct zone {
427 430 /*
428 431 * zone_name is never modified once set.
429 432 */
430 433 char *zone_name; /* zone's configuration name */
431 434 /*
432 435 * zone_nodename and zone_domain are never freed once allocated.
433 436 */
434 437 char *zone_nodename; /* utsname.nodename equivalent */
435 438 char *zone_domain; /* srpc_domain equivalent */
436 439 /*
437 440 * zone_hostid is used for per-zone hostid emulation.
438 441 * Currently it isn't modified after it's set (so no locks protect
439 442 * accesses), but that might have to change when we allow
440 443 * administrators to change running zones' properties.
441 444 *
442 445 * The global zone's zone_hostid must always be HW_INVALID_HOSTID so
443 446 * that zone_get_hostid() will function correctly.
444 447 */
445 448 uint32_t zone_hostid; /* zone's hostid, HW_INVALID_HOSTID */
446 449 /* if not emulated */
447 450 /*
448 451 * zone_lock protects the following fields of a zone_t:
449 452 * zone_ref
450 453 * zone_cred_ref
451 454 * zone_subsys_ref
452 455 * zone_ref_list
453 456 * zone_ntasks
454 457 * zone_flags
|
↓ open down ↓ |
229 lines elided |
↑ open up ↑ |
455 458 * zone_zsd
456 459 * zone_pfexecd
457 460 */
458 461 kmutex_t zone_lock;
459 462 /*
460 463 * zone_linkage is the zone's linkage into the active or
461 464 * death-row list. The field is protected by zonehash_lock.
462 465 */
463 466 list_node_t zone_linkage;
464 467 zoneid_t zone_id; /* ID of zone */
468 + zoneid_t zone_did; /* persistent debug ID of zone */
465 469 uint_t zone_ref; /* count of zone_hold()s on zone */
466 470 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */
467 471 /*
468 472 * Fixed-sized array of subsystem-specific reference counts
469 473 * The sum of all of the counts must be less than or equal to zone_ref.
470 474 * The array is indexed by the counts' subsystems' zone_ref_subsys_t
471 475 * constants.
472 476 */
473 477 uint_t zone_subsys_ref[ZONE_REF_NUM_SUBSYS];
474 478 list_t zone_ref_list; /* list of zone_ref_t structs */
475 479 /*
476 480 * zone_rootvp and zone_rootpath can never be modified once set.
477 481 */
478 482 struct vnode *zone_rootvp; /* zone's root vnode */
479 483 char *zone_rootpath; /* Path to zone's root + '/' */
480 484 ushort_t zone_flags; /* misc flags */
481 485 zone_status_t zone_status; /* protected by zone_status_lock */
482 486 uint_t zone_ntasks; /* number of tasks executing in zone */
483 487 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */
484 488 /* counters in projects and tasks */
485 489 /* that are within the zone */
486 490 rctl_qty_t zone_nlwps; /* number of lwps in zone */
487 491 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */
488 492 rctl_qty_t zone_shmmax; /* System V shared memory usage */
489 493 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */
490 494
491 495 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */
492 496 uint32_t zone_shares; /* FSS shares allocated to zone */
493 497 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */
494 498 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */
495 499 /* kpd_locked_mem for all */
496 500 /* projects in zone. */
497 501 /* Also protects zone_max_swap */
498 502 /* grab after p_lock, before rcs_lock */
499 503 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */
500 504 /* zone */
501 505 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */
502 506 /* limit. Protected by */
503 507 /* zone_rctls->rcs_lock */
504 508 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */
505 509 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */
506 510 /* Protected by */
507 511 /* zone_rctls->rcs_lock */
508 512 kmutex_t zone_rctl_lock; /* protects zone_max_lofi */
509 513 rctl_qty_t zone_max_lofi; /* lofi devs for zone */
510 514 rctl_qty_t zone_max_lofi_ctl; /* current lofi limit. */
511 515 /* Protected by */
512 516 /* zone_rctls->rcs_lock */
513 517 list_t zone_zsd; /* list of Zone-Specific Data values */
514 518 kcondvar_t zone_cv; /* used to signal state changes */
515 519 struct proc *zone_zsched; /* Dummy kernel "zsched" process */
516 520 pid_t zone_proc_initpid; /* pid of "init" for this zone */
517 521 char *zone_initname; /* fs path to 'init' */
518 522 int zone_boot_err; /* for zone_boot() if boot fails */
519 523 char *zone_bootargs; /* arguments passed via zone_boot() */
520 524 uint64_t zone_phys_mcap; /* physical memory cap */
521 525 /*
522 526 * zone_kthreads is protected by zone_status_lock.
523 527 */
524 528 kthread_t *zone_kthreads; /* kernel threads in zone */
525 529 struct priv_set *zone_privset; /* limit set for zone */
526 530 /*
527 531 * zone_vfslist is protected by vfs_list_lock().
528 532 */
529 533 struct vfs *zone_vfslist; /* list of FS's mounted in zone */
530 534 uint64_t zone_uniqid; /* unique zone generation number */
531 535 struct cred *zone_kcred; /* kcred-like, zone-limited cred */
532 536 /*
533 537 * zone_pool is protected by pool_lock().
534 538 */
535 539 struct pool *zone_pool; /* pool the zone is bound to */
536 540 hrtime_t zone_pool_mod; /* last pool bind modification time */
537 541 /* zone_psetid is protected by cpu_lock */
538 542 psetid_t zone_psetid; /* pset the zone is bound to */
539 543
540 544 time_t zone_boot_time; /* Similar to boot_time */
541 545
542 546 /*
543 547 * The following two can be read without holding any locks. They are
544 548 * updated under cpu_lock.
545 549 */
546 550 int zone_ncpus; /* zone's idea of ncpus */
547 551 int zone_ncpus_online; /* zone's idea of ncpus_online */
548 552 /*
549 553 * List of ZFS datasets exported to this zone.
550 554 */
551 555 list_t zone_datasets; /* list of datasets */
552 556
553 557 ts_label_t *zone_slabel; /* zone sensitivity label */
554 558 int zone_match; /* require label match for packets */
555 559 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */
556 560
557 561 boolean_t zone_restart_init; /* Restart init if it dies? */
558 562 struct brand *zone_brand; /* zone's brand */
559 563 void *zone_brand_data; /* store brand specific data */
560 564 id_t zone_defaultcid; /* dflt scheduling class id */
561 565 kstat_t *zone_swapresv_kstat;
562 566 kstat_t *zone_lockedmem_kstat;
563 567 /*
564 568 * zone_dl_list is protected by zone_lock
565 569 */
566 570 list_t zone_dl_list;
567 571 netstack_t *zone_netstack;
568 572 struct cpucap *zone_cpucap; /* CPU caps data */
569 573 /*
570 574 * Solaris Auditing per-zone audit context
571 575 */
572 576 struct au_kcontext *zone_audit_kctxt;
573 577 /*
574 578 * For private use by mntfs.
575 579 */
576 580 struct mntelem *zone_mntfs_db;
577 581 krwlock_t zone_mntfs_db_lock;
578 582
579 583 struct klpd_reg *zone_pfexecd;
580 584
581 585 char *zone_fs_allowed;
582 586 rctl_qty_t zone_nprocs; /* number of processes in the zone */
583 587 rctl_qty_t zone_nprocs_ctl; /* current limit protected by */
584 588 /* zone_rctls->rcs_lock */
585 589 kstat_t *zone_nprocs_kstat;
586 590
587 591 kmutex_t zone_mcap_lock; /* protects mcap statistics */
588 592 kstat_t *zone_mcap_ksp;
589 593 zone_mcap_kstat_t *zone_mcap_stats;
590 594 uint64_t zone_pgpgin; /* pages paged in */
591 595 uint64_t zone_anonpgin; /* anon pages paged in */
592 596 uint64_t zone_execpgin; /* exec pages paged in */
593 597 uint64_t zone_fspgin; /* fs pages paged in */
594 598 uint64_t zone_anon_alloc_fail; /* cnt of anon alloc fails */
595 599
596 600 psecflags_t zone_secflags; /* default zone security-flags */
597 601
598 602 /*
599 603 * Misc. kstats and counters for zone cpu-usage aggregation.
600 604 */
601 605 kmutex_t zone_misc_lock; /* protects misc statistics */
602 606 kstat_t *zone_misc_ksp;
603 607 zone_misc_kstat_t *zone_misc_stats;
604 608 /* Accumulated microstate for all threads in this zone. */
605 609 cpu_uarray_t *zone_ustate;
606 610 /* fork-fail kstat tracking */
607 611 uint32_t zone_ffcap; /* hit an rctl cap */
608 612 uint32_t zone_ffnoproc; /* get proc/lwp error */
609 613 uint32_t zone_ffnomem; /* as_dup/memory error */
610 614 uint32_t zone_ffmisc; /* misc. other error */
611 615
612 616 uint32_t zone_nested_intp; /* nested interp. kstat */
613 617
614 618 struct loadavg_s zone_loadavg; /* loadavg for this zone */
615 619 uint64_t zone_hp_avenrun[3]; /* high-precision avenrun */
616 620 int zone_avenrun[3]; /* FSCALED avg. run queue len */
617 621
618 622 /*
619 623 * FSS stats updated once per second by fss_decay_usage.
620 624 */
621 625 uint32_t zone_fss_gen; /* FSS generation cntr */
622 626 uint64_t zone_run_ticks; /* tot # of ticks running */
623 627
624 628 /*
625 629 * DTrace-private per-zone state
626 630 */
627 631 int zone_dtrace_getf; /* # of unprivileged getf()s */
628 632
629 633 /*
630 634 * Synchronization primitives used to synchronize between mounts and
631 635 * zone creation/destruction.
632 636 */
633 637 int zone_mounts_in_progress;
634 638 kcondvar_t zone_mount_cv;
635 639 kmutex_t zone_mount_lock;
636 640 } zone_t;
637 641
638 642 /*
639 643 * Special value of zone_psetid to indicate that pools are disabled.
640 644 */
641 645 #define ZONE_PS_INVAL PS_MYID
642 646
643 647 extern zone_t zone0;
644 648 extern zone_t *global_zone;
645 649 extern uint_t maxzones;
646 650 extern rctl_hndl_t rc_zone_nlwps;
647 651 extern rctl_hndl_t rc_zone_nprocs;
648 652
649 653 extern long zone(int, void *, void *, void *, void *);
650 654 extern void zone_zsd_init(void);
651 655 extern void zone_init(void);
652 656 extern void zone_hold(zone_t *);
653 657 extern void zone_rele(zone_t *);
654 658 extern void zone_init_ref(zone_ref_t *);
655 659 extern void zone_hold_ref(zone_t *, zone_ref_t *, zone_ref_subsys_t);
656 660 extern void zone_rele_ref(zone_ref_t *, zone_ref_subsys_t);
|
↓ open down ↓ |
182 lines elided |
↑ open up ↑ |
657 661 extern void zone_cred_hold(zone_t *);
658 662 extern void zone_cred_rele(zone_t *);
659 663 extern void zone_task_hold(zone_t *);
660 664 extern void zone_task_rele(zone_t *);
661 665 extern zone_t *zone_find_by_id(zoneid_t);
662 666 extern zone_t *zone_find_by_label(const ts_label_t *);
663 667 extern zone_t *zone_find_by_name(char *);
664 668 extern zone_t *zone_find_by_any_path(const char *, boolean_t);
665 669 extern zone_t *zone_find_by_path(const char *);
666 670 extern zoneid_t getzoneid(void);
671 +extern zoneid_t getzonedid(void);
667 672 extern zone_t *zone_find_by_id_nolock(zoneid_t);
668 673 extern int zone_datalink_walk(zoneid_t, int (*)(datalink_id_t, void *), void *);
669 674 extern int zone_check_datalink(zoneid_t *, datalink_id_t);
670 675 extern void zone_loadavg_update();
671 676
672 677 /*
673 678 * Zone-specific data (ZSD) APIs
674 679 */
675 680 /*
676 681 * The following is what code should be initializing its zone_key_t to if it
677 682 * calls zone_getspecific() without necessarily knowing that zone_key_create()
678 683 * has been called on the key.
679 684 */
680 685 #define ZONE_KEY_UNINITIALIZED 0
681 686
682 687 typedef uint_t zone_key_t;
683 688
684 689 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t),
685 690 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *));
686 691 extern int zone_key_delete(zone_key_t);
687 692 extern void *zone_getspecific(zone_key_t, zone_t *);
688 693 extern int zone_setspecific(zone_key_t, zone_t *, const void *);
689 694
690 695 /*
691 696 * The definition of a zsd_entry is truly private to zone.c and is only
692 697 * placed here so it can be shared with mdb.
693 698 *
694 699 * State maintained for each zone times each registered key, which tracks
695 700 * the state of the create, shutdown and destroy callbacks.
696 701 *
697 702 * zsd_flags is used to keep track of pending actions to avoid holding locks
698 703 * when calling the create/shutdown/destroy callbacks, since doing so
699 704 * could lead to deadlocks.
700 705 */
701 706 struct zsd_entry {
702 707 zone_key_t zsd_key; /* Key used to lookup value */
703 708 void *zsd_data; /* Caller-managed value */
704 709 /*
705 710 * Callbacks to be executed when a zone is created, shutdown, and
706 711 * destroyed, respectively.
707 712 */
708 713 void *(*zsd_create)(zoneid_t);
709 714 void (*zsd_shutdown)(zoneid_t, void *);
710 715 void (*zsd_destroy)(zoneid_t, void *);
711 716 list_node_t zsd_linkage;
712 717 uint16_t zsd_flags; /* See below */
713 718 kcondvar_t zsd_cv;
714 719 };
715 720
716 721 /*
717 722 * zsd_flags
718 723 */
719 724 #define ZSD_CREATE_NEEDED 0x0001
720 725 #define ZSD_CREATE_INPROGRESS 0x0002
721 726 #define ZSD_CREATE_COMPLETED 0x0004
722 727 #define ZSD_SHUTDOWN_NEEDED 0x0010
723 728 #define ZSD_SHUTDOWN_INPROGRESS 0x0020
724 729 #define ZSD_SHUTDOWN_COMPLETED 0x0040
725 730 #define ZSD_DESTROY_NEEDED 0x0100
726 731 #define ZSD_DESTROY_INPROGRESS 0x0200
727 732 #define ZSD_DESTROY_COMPLETED 0x0400
728 733
729 734 #define ZSD_CREATE_ALL \
730 735 (ZSD_CREATE_NEEDED|ZSD_CREATE_INPROGRESS|ZSD_CREATE_COMPLETED)
731 736 #define ZSD_SHUTDOWN_ALL \
732 737 (ZSD_SHUTDOWN_NEEDED|ZSD_SHUTDOWN_INPROGRESS|ZSD_SHUTDOWN_COMPLETED)
733 738 #define ZSD_DESTROY_ALL \
734 739 (ZSD_DESTROY_NEEDED|ZSD_DESTROY_INPROGRESS|ZSD_DESTROY_COMPLETED)
735 740
736 741 #define ZSD_ALL_INPROGRESS \
737 742 (ZSD_CREATE_INPROGRESS|ZSD_SHUTDOWN_INPROGRESS|ZSD_DESTROY_INPROGRESS)
738 743
739 744 /*
740 745 * Macros to help with zone visibility restrictions.
741 746 */
742 747
743 748 /*
744 749 * Is process in the global zone?
745 750 */
746 751 #define INGLOBALZONE(p) \
747 752 ((p)->p_zone == global_zone)
748 753
749 754 /*
750 755 * Can process view objects in given zone?
751 756 */
752 757 #define HASZONEACCESS(p, zoneid) \
753 758 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p))
754 759
755 760 /*
756 761 * Convenience macro to see if a resolved path is visible from within a
757 762 * given zone.
758 763 *
759 764 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the
760 765 * two strings must be equal. Since the rootpathlen has a trailing '/',
761 766 * we want to skip everything in the path up to (but not including) the
762 767 * trailing '/'.
763 768 */
764 769 #define ZONE_PATH_VISIBLE(path, zone) \
765 770 (strncmp((path), (zone)->zone_rootpath, \
766 771 (zone)->zone_rootpathlen - 1) == 0)
767 772
768 773 /*
769 774 * Convenience macro to go from the global view of a path to that seen
770 775 * from within said zone. It is the responsibility of the caller to
771 776 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is
772 777 * in fact visible from within the zone.
773 778 */
774 779 #define ZONE_PATH_TRANSLATE(path, zone) \
775 780 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \
776 781 (path) + (zone)->zone_rootpathlen - 2)
777 782
778 783 /*
779 784 * Special processes visible in all zones.
780 785 */
781 786 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1)
782 787
783 788 /*
784 789 * A root vnode of the current zone.
785 790 *
786 791 * NOTE: It may be necessary (initialization time for file sharing where an
787 792 * NGZ loads a file-sharing kernel module that does zsd initialization) to NOT
788 793 * use this macro. One should ASSERT() that curzone == active ZSD (an
789 794 * ASSERTion that's not always true at ZSD initialization time) during regular
790 795 * use of this macro.
791 796 */
792 797 #define ZONE_ROOTVP() (curzone->zone_rootvp)
793 798
794 799 /*
795 800 * Since a zone's root isn't necessarily an actual filesystem boundary
796 801 * (i.e. VROOT may not be set on zone->zone_rootvp) we need to not assume it.
797 802 * This macro helps in checking if a vnode is the current zone's rootvp.
798 803 * NOTE: Using the VN_ prefix, even though it's defined here in zone.h.
799 804 * NOTE2: See above warning about ZONE_ROOTVP().
800 805 */
801 806 #define VN_IS_CURZONEROOT(vp) (VN_CMP(vp, ZONE_ROOTVP()))
802 807
803 808 /*
804 809 * Zone-safe version of thread_create() to be used when the caller wants to
805 810 * create a kernel thread to run within the current zone's context.
806 811 */
807 812 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t,
808 813 pri_t);
809 814 extern void zthread_exit(void);
810 815
811 816 /*
812 817 * Functions for an external observer to register interest in a zone's status
813 818 * change. Observers will be woken up when the zone status equals the status
814 819 * argument passed in (in the case of zone_status_timedwait, the function may
815 820 * also return because of a timeout; zone_status_wait_sig may return early due
816 821 * to a signal being delivered; zone_status_timedwait_sig may return for any of
817 822 * the above reasons).
818 823 *
819 824 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and
820 825 * cv_wait_sig() respectively.
821 826 */
822 827 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t);
823 828 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t);
824 829 extern void zone_status_wait(zone_t *, zone_status_t);
825 830 extern int zone_status_wait_sig(zone_t *, zone_status_t);
826 831
827 832 /*
828 833 * Get the status of the zone (at the time it was called). The state may
829 834 * have progressed by the time it is returned.
830 835 */
831 836 extern zone_status_t zone_status_get(zone_t *);
832 837
833 838 /*
834 839 * Safely get the hostid of the specified zone (defaults to machine's hostid
835 840 * if the specified zone doesn't emulate a hostid). Passing NULL retrieves
836 841 * the global zone's (i.e., physical system's) hostid.
837 842 */
838 843 extern uint32_t zone_get_hostid(zone_t *);
839 844
840 845 /*
841 846 * Get the "kcred" credentials corresponding to the given zone.
842 847 */
843 848 extern struct cred *zone_get_kcred(zoneid_t);
844 849
845 850 /*
846 851 * Get/set the pool the zone is currently bound to.
847 852 */
848 853 extern struct pool *zone_pool_get(zone_t *);
849 854 extern void zone_pool_set(zone_t *, struct pool *);
850 855
851 856 /*
852 857 * Get/set the pset the zone is currently using.
853 858 */
854 859 extern psetid_t zone_pset_get(zone_t *);
855 860 extern void zone_pset_set(zone_t *, psetid_t);
856 861
857 862 /*
858 863 * Get the number of cpus/online-cpus visible from the given zone.
859 864 */
860 865 extern int zone_ncpus_get(zone_t *);
861 866 extern int zone_ncpus_online_get(zone_t *);
862 867
863 868 /*
864 869 * Returns true if the named pool/dataset is visible in the current zone.
865 870 */
866 871 extern int zone_dataset_visible(const char *, int *);
867 872
868 873 /*
869 874 * zone version of kadmin()
870 875 */
871 876 extern int zone_kadmin(int, int, const char *, cred_t *);
872 877 extern void zone_shutdown_global(void);
873 878
874 879 extern void mount_in_progress(zone_t *);
875 880 extern void mount_completed(zone_t *);
876 881
877 882 extern int zone_walk(int (*)(zone_t *, void *), void *);
878 883
879 884 extern rctl_hndl_t rc_zone_locked_mem;
880 885 extern rctl_hndl_t rc_zone_max_swap;
881 886 extern rctl_hndl_t rc_zone_max_lofi;
882 887
883 888 #endif /* _KERNEL */
884 889
885 890 #ifdef __cplusplus
886 891 }
887 892 #endif
888 893
889 894 #endif /* _SYS_ZONE_H */
|
↓ open down ↓ |
213 lines elided |
↑ open up ↑ |
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX