Print this page
11083 support NFS server in zone
Portions contributed by: Dan Kruchinin <dan.kruchinin@nexenta.com>
Portions contributed by: Stepan Zastupov <stepan.zastupov@gmail.com>
Portions contributed by: Joyce McIntosh <joyce.mcintosh@nexenta.com>
Portions contributed by: Mike Zeller <mike@mikezeller.net>
Portions contributed by: Dan McDonald <danmcd@joyent.com>
Portions contributed by: Gordon Ross <gordon.w.ross@gmail.com>
Portions contributed by: Vitaliy Gusev <gusev.vitaliy@gmail.com>
Reviewed by: Rick McNeal <rick.mcneal@nexenta.com>
Reviewed by: Rob Gittins <rob.gittins@nexenta.com>
Reviewed by: Sanjay Nadkarni <sanjay.nadkarni@nexenta.com>
Reviewed by: Jason King <jbk@joyent.com>
Reviewed by: C Fraire <cfraire@me.com>
Change-Id: I22f289d357503f9b48a0bc2482cc4328a6d43d16
| 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
|
↓ open down ↓ |
10 lines elided |
↑ open up ↑ |
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 * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
23 - * Copyright 2018 Joyent, Inc.
24 - * Copyright 2019 Nexenta Systems, Inc. All rights reserved.
25 24 * Copyright 2014 Igor Kozhukhov <ikozhukhov@gmail.com>.
25 + * Copyright 2019 Nexenta Systems, Inc. All rights reserved.
26 + * Copyright 2020 Joyent, Inc.
26 27 */
27 28
28 29 #ifndef _SYS_ZONE_H
29 30 #define _SYS_ZONE_H
30 31
31 32 #include <sys/types.h>
32 33 #include <sys/param.h>
33 34 #include <sys/tsol/label.h>
34 35 #include <sys/uadmin.h>
35 36 #include <netinet/in.h>
36 37
37 38 #ifdef _KERNEL
38 39 /*
39 40 * Many includes are kernel-only to reduce namespace pollution of
40 41 * userland applications.
41 42 */
42 43 #include <sys/mutex.h>
43 44 #include <sys/rctl.h>
44 45 #include <sys/ipc_rctl.h>
45 46 #include <sys/pset.h>
46 47 #include <sys/cred.h>
47 48 #include <sys/netstack.h>
48 49 #include <sys/ksynch.h>
49 50 #include <sys/socket_impl.h>
50 51 #include <sys/secflags.h>
51 52 #include <sys/cpu_uarray.h>
52 53 #include <sys/list.h>
53 54 #include <sys/loadavg.h>
54 55 #endif /* _KERNEL */
55 56
56 57 #ifdef __cplusplus
57 58 extern "C" {
58 59 #endif
59 60
60 61 /*
61 62 * NOTE
62 63 *
63 64 * The contents of this file are private to the implementation of
64 65 * Solaris and are subject to change at any time without notice.
65 66 * Applications and drivers using these interfaces may fail to
66 67 * run on future releases.
67 68 */
68 69
69 70 /* Available both in kernel and for user space */
70 71
71 72 /* zone id restrictions and special ids */
72 73 #define MAX_ZONEID 9999
73 74 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */
74 75 #define MIN_ZONEID 0 /* minimum zone ID on system */
75 76 #define GLOBAL_ZONEID 0
76 77 #define ZONEID_WIDTH 4 /* for printf */
77 78
78 79 /*
79 80 * Special zoneid_t token to refer to all zones.
80 81 */
81 82 #define ALL_ZONES (-1)
82 83
83 84 /* system call subcodes */
84 85 #define ZONE_CREATE 0
85 86 #define ZONE_DESTROY 1
86 87 #define ZONE_GETATTR 2
87 88 #define ZONE_ENTER 3
88 89 #define ZONE_LIST 4
89 90 #define ZONE_SHUTDOWN 5
90 91 #define ZONE_LOOKUP 6
91 92 #define ZONE_BOOT 7
92 93 #define ZONE_VERSION 8
93 94 #define ZONE_SETATTR 9
94 95 #define ZONE_ADD_DATALINK 10
95 96 #define ZONE_DEL_DATALINK 11
96 97 #define ZONE_CHECK_DATALINK 12
97 98 #define ZONE_LIST_DATALINK 13
98 99
99 100 /* zone attributes */
100 101 #define ZONE_ATTR_ROOT 1
101 102 #define ZONE_ATTR_NAME 2
102 103 #define ZONE_ATTR_STATUS 3
103 104 #define ZONE_ATTR_PRIVSET 4
104 105 #define ZONE_ATTR_UNIQID 5
105 106 #define ZONE_ATTR_POOLID 6
106 107 #define ZONE_ATTR_INITPID 7
107 108 #define ZONE_ATTR_SLBL 8
108 109 #define ZONE_ATTR_INITNAME 9
109 110 #define ZONE_ATTR_BOOTARGS 10
110 111 #define ZONE_ATTR_BRAND 11
111 112 #define ZONE_ATTR_PHYS_MCAP 12
112 113 #define ZONE_ATTR_SCHED_CLASS 13
113 114 #define ZONE_ATTR_FLAGS 14
114 115 #define ZONE_ATTR_HOSTID 15
115 116 #define ZONE_ATTR_FS_ALLOWED 16
116 117 #define ZONE_ATTR_NETWORK 17
117 118 #define ZONE_ATTR_INITNORESTART 20
118 119 #define ZONE_ATTR_SECFLAGS 21
119 120
120 121 /* Start of the brand-specific attribute namespace */
121 122 #define ZONE_ATTR_BRAND_ATTRS 32768
122 123
123 124 #define ZONE_FS_ALLOWED_MAX 1024
124 125
125 126 #define ZONE_EVENT_CHANNEL "com.sun:zones:status"
126 127 #define ZONE_EVENT_STATUS_CLASS "status"
127 128 #define ZONE_EVENT_STATUS_SUBCLASS "change"
128 129
129 130 #define ZONE_EVENT_UNINITIALIZED "uninitialized"
130 131 #define ZONE_EVENT_INITIALIZED "initialized"
131 132 #define ZONE_EVENT_READY "ready"
132 133 #define ZONE_EVENT_RUNNING "running"
133 134 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down"
134 135
135 136 #define ZONE_CB_NAME "zonename"
136 137 #define ZONE_CB_NEWSTATE "newstate"
137 138 #define ZONE_CB_OLDSTATE "oldstate"
138 139 #define ZONE_CB_TIMESTAMP "when"
139 140 #define ZONE_CB_ZONEID "zoneid"
140 141
141 142 /*
142 143 * Exit values that may be returned by scripts or programs invoked by various
143 144 * zone commands.
144 145 *
145 146 * These are defined as:
146 147 *
147 148 * ZONE_SUBPROC_OK
148 149 * ===============
149 150 * The subprocess completed successfully.
150 151 *
151 152 * ZONE_SUBPROC_USAGE
152 153 * ==================
153 154 * The subprocess failed with a usage message, or a usage message should
154 155 * be output in its behalf.
155 156 *
156 157 * ZONE_SUBPROC_NOTCOMPLETE
157 158 * ========================
158 159 * The subprocess did not complete, but the actions performed by the
159 160 * subprocess require no recovery actions by the user.
160 161 *
161 162 * For example, if the subprocess were called by "zoneadm install," the
162 163 * installation of the zone did not succeed but the user need not perform
163 164 * a "zoneadm uninstall" before attempting another install.
164 165 *
165 166 * ZONE_SUBPROC_FATAL
166 167 * ==================
167 168 * The subprocess failed in a fatal manner, usually one that will require
168 169 * some type of recovery action by the user.
169 170 *
170 171 * For example, if the subprocess were called by "zoneadm install," the
171 172 * installation of the zone did not succeed and the user will need to
172 173 * perform a "zoneadm uninstall" before another install attempt is
173 174 * possible.
174 175 *
175 176 * The non-success exit values are large to avoid accidental collision
176 177 * with values used internally by some commands (e.g. "Z_ERR" and
177 178 * "Z_USAGE" as used by zoneadm.)
178 179 */
179 180 #define ZONE_SUBPROC_OK 0
180 181 #define ZONE_SUBPROC_USAGE 253
181 182 #define ZONE_SUBPROC_NOTCOMPLETE 254
182 183 #define ZONE_SUBPROC_FATAL 255
183 184
184 185 #ifdef _SYSCALL32
185 186 typedef struct {
186 187 caddr32_t zone_name;
187 188 caddr32_t zone_root;
188 189 caddr32_t zone_privs;
189 190 size32_t zone_privssz;
190 191 caddr32_t rctlbuf;
191 192 size32_t rctlbufsz;
192 193 caddr32_t extended_error;
193 194 caddr32_t zfsbuf;
194 195 size32_t zfsbufsz;
195 196 int match; /* match level */
196 197 uint32_t doi; /* DOI for label */
197 198 caddr32_t label; /* label associated with zone */
198 199 int flags;
199 200 } zone_def32;
200 201 #endif
201 202 typedef struct {
202 203 const char *zone_name;
203 204 const char *zone_root;
204 205 const struct priv_set *zone_privs;
205 206 size_t zone_privssz;
206 207 const char *rctlbuf;
207 208 size_t rctlbufsz;
208 209 int *extended_error;
209 210 const char *zfsbuf;
210 211 size_t zfsbufsz;
211 212 int match; /* match level */
212 213 uint32_t doi; /* DOI for label */
213 214 const bslabel_t *label; /* label associated with zone */
214 215 int flags;
215 216 } zone_def;
216 217
217 218 /* extended error information */
218 219 #define ZE_UNKNOWN 0 /* No extended error info */
219 220 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */
220 221 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */
221 222 #define ZE_LABELINUSE 3 /* label is already in use by some other zone */
222 223
223 224 /*
224 225 * zone_status values
225 226 *
226 227 * You must modify zone_status_names in mdb(1M)'s genunix module
227 228 * (genunix/zone.c) when you modify this enum.
228 229 */
229 230 typedef enum {
230 231 ZONE_IS_UNINITIALIZED = 0,
231 232 ZONE_IS_INITIALIZED,
232 233 ZONE_IS_READY,
233 234 ZONE_IS_BOOTING,
234 235 ZONE_IS_RUNNING,
235 236 ZONE_IS_SHUTTING_DOWN,
236 237 ZONE_IS_EMPTY,
237 238 ZONE_IS_DOWN,
238 239 ZONE_IS_DYING,
239 240 ZONE_IS_DEAD
240 241 } zone_status_t;
241 242 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED
242 243 #define ZONE_MAX_STATE ZONE_IS_DEAD
243 244
244 245 /*
245 246 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also
246 247 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT.
247 248 */
248 249 typedef enum zone_cmd {
249 250 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING,
250 251 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT, Z_SHUTDOWN
251 252 } zone_cmd_t;
252 253
253 254 /*
254 255 * The structure of a request to zoneadmd.
255 256 */
256 257 typedef struct zone_cmd_arg {
257 258 uint64_t uniqid; /* unique "generation number" */
258 259 zone_cmd_t cmd; /* requested action */
259 260 uint32_t _pad; /* need consistent 32/64 bit alignmt */
260 261 char locale[MAXPATHLEN]; /* locale in which to render messages */
261 262 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */
262 263 } zone_cmd_arg_t;
263 264
264 265 /*
265 266 * Structure of zoneadmd's response to a request. A NULL return value means
266 267 * the caller should attempt to restart zoneadmd and retry.
267 268 */
268 269 typedef struct zone_cmd_rval {
269 270 int rval; /* return value of request */
270 271 char errbuf[1]; /* variable-sized buffer containing error messages */
271 272 } zone_cmd_rval_t;
272 273
273 274 /*
274 275 * The zone support infrastructure uses the zone name as a component
275 276 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters
276 277 * in length, so ZONENAME_MAX is limited by that.
277 278 */
278 279 #define ZONENAME_MAX 64
279 280
280 281 #define GLOBAL_ZONENAME "global"
281 282
282 283 /*
283 284 * Extended Regular expression (see regex(5)) which matches all valid zone
284 285 * names.
285 286 */
286 287 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}"
287 288
288 289 /*
289 290 * Where the zones support infrastructure places temporary files.
290 291 */
291 292 #define ZONES_TMPDIR "/var/run/zones"
292 293
293 294 /*
294 295 * The path to the door used by clients to communicate with zoneadmd.
295 296 */
296 297 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door"
297 298
298 299
299 300 /* zone_flags */
300 301 /*
301 302 * Threads that read or write the following flag must hold zone_lock.
302 303 */
303 304 #define ZF_REFCOUNTS_LOGGED 0x1 /* a thread logged the zone's refs */
304 305
305 306 /*
306 307 * The following threads are set when the zone is created and never changed.
307 308 * Threads that test for these flags don't have to hold zone_lock.
308 309 */
309 310 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */
310 311 #define ZF_IS_SCRATCH 0x4 /* scratch zone */
311 312 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */
312 313
313 314
314 315 /* zone_create flags */
315 316 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */
316 317
317 318 /* zone network properties */
318 319 #define ZONE_NETWORK_ADDRESS 1
319 320 #define ZONE_NETWORK_DEFROUTER 2
320 321
321 322 #define ZONE_NET_ADDRNAME "address"
322 323 #define ZONE_NET_RTRNAME "route"
323 324
324 325 typedef struct zone_net_data {
325 326 int zn_type;
326 327 int zn_len;
327 328 datalink_id_t zn_linkid;
328 329 uint8_t zn_val[1];
329 330 } zone_net_data_t;
330 331
331 332
332 333 #ifdef _KERNEL
333 334
334 335 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */
335 336
336 337 /*
337 338 * Indexes into ->zone_ustate array, summing the micro state of all threads in a
338 339 * particular zone.
339 340 */
340 341 #define ZONE_USTATE_STIME (0)
341 342 #define ZONE_USTATE_UTIME (1)
342 343 #define ZONE_USTATE_WTIME (2)
343 344 #define ZONE_USTATE_MAX (3)
344 345
345 346 struct pool;
346 347 struct brand;
347 348
348 349 /*
349 350 * Each of these constants identifies a kernel subsystem that acquires and
350 351 * releases zone references. Each subsystem that invokes
351 352 * zone_hold_ref() and zone_rele_ref() should specify the
352 353 * zone_ref_subsys_t constant associated with the subsystem. Tracked holds
353 354 * help users and developers quickly identify subsystems that stall zone
354 355 * shutdowns indefinitely.
355 356 *
356 357 * NOTE: You must modify zone_ref_subsys_names in usr/src/uts/common/os/zone.c
357 358 * when you modify this enumeration.
358 359 */
359 360 typedef enum zone_ref_subsys {
360 361 ZONE_REF_NFS, /* NFS */
361 362 ZONE_REF_NFSV4, /* NFSv4 */
362 363 ZONE_REF_SMBFS, /* SMBFS */
363 364 ZONE_REF_MNTFS, /* MNTFS */
364 365 ZONE_REF_LOFI, /* LOFI devices */
365 366 ZONE_REF_VFS, /* VFS infrastructure */
366 367 ZONE_REF_IPC, /* IPC infrastructure */
367 368 ZONE_REF_NUM_SUBSYS /* This must be the last entry. */
368 369 } zone_ref_subsys_t;
369 370
370 371 /*
371 372 * zone_ref represents a general-purpose references to a zone. Each zone's
372 373 * references are linked into the zone's zone_t::zone_ref_list. This allows
373 374 * debuggers to walk zones' references.
374 375 */
375 376 typedef struct zone_ref {
376 377 struct zone *zref_zone; /* the zone to which the reference refers */
377 378 list_node_t zref_linkage; /* linkage for zone_t::zone_ref_list */
378 379 } zone_ref_t;
379 380
380 381 /*
381 382 * Structure to record list of ZFS datasets exported to a zone.
382 383 */
383 384 typedef struct zone_dataset {
384 385 char *zd_dataset;
385 386 list_node_t zd_linkage;
386 387 } zone_dataset_t;
387 388
388 389 /*
389 390 * structure for zone kstats
390 391 */
391 392 typedef struct zone_kstat {
392 393 kstat_named_t zk_zonename;
393 394 kstat_named_t zk_usage;
394 395 kstat_named_t zk_value;
395 396 } zone_kstat_t;
396 397
397 398 struct cpucap;
398 399
399 400 typedef struct {
400 401 kstat_named_t zm_zonename;
401 402 kstat_named_t zm_pgpgin;
402 403 kstat_named_t zm_anonpgin;
403 404 kstat_named_t zm_execpgin;
404 405 kstat_named_t zm_fspgin;
405 406 kstat_named_t zm_anon_alloc_fail;
406 407 } zone_mcap_kstat_t;
407 408
408 409 typedef struct {
409 410 kstat_named_t zm_zonename; /* full name, kstat truncates name */
410 411 kstat_named_t zm_utime;
411 412 kstat_named_t zm_stime;
412 413 kstat_named_t zm_wtime;
413 414 kstat_named_t zm_avenrun1;
414 415 kstat_named_t zm_avenrun5;
415 416 kstat_named_t zm_avenrun15;
416 417 kstat_named_t zm_ffcap;
417 418 kstat_named_t zm_ffnoproc;
418 419 kstat_named_t zm_ffnomem;
419 420 kstat_named_t zm_ffmisc;
420 421 kstat_named_t zm_nested_intp;
421 422 kstat_named_t zm_init_pid;
422 423 kstat_named_t zm_boot_time;
423 424 } zone_misc_kstat_t;
424 425
425 426 typedef struct zone {
426 427 /*
427 428 * zone_name is never modified once set.
428 429 */
429 430 char *zone_name; /* zone's configuration name */
430 431 /*
431 432 * zone_nodename and zone_domain are never freed once allocated.
432 433 */
433 434 char *zone_nodename; /* utsname.nodename equivalent */
434 435 char *zone_domain; /* srpc_domain equivalent */
435 436 /*
436 437 * zone_hostid is used for per-zone hostid emulation.
437 438 * Currently it isn't modified after it's set (so no locks protect
438 439 * accesses), but that might have to change when we allow
439 440 * administrators to change running zones' properties.
440 441 *
441 442 * The global zone's zone_hostid must always be HW_INVALID_HOSTID so
442 443 * that zone_get_hostid() will function correctly.
443 444 */
444 445 uint32_t zone_hostid; /* zone's hostid, HW_INVALID_HOSTID */
445 446 /* if not emulated */
446 447 /*
447 448 * zone_lock protects the following fields of a zone_t:
448 449 * zone_ref
449 450 * zone_cred_ref
450 451 * zone_subsys_ref
451 452 * zone_ref_list
452 453 * zone_ntasks
453 454 * zone_flags
454 455 * zone_zsd
455 456 * zone_pfexecd
456 457 */
457 458 kmutex_t zone_lock;
458 459 /*
459 460 * zone_linkage is the zone's linkage into the active or
460 461 * death-row list. The field is protected by zonehash_lock.
461 462 */
462 463 list_node_t zone_linkage;
463 464 zoneid_t zone_id; /* ID of zone */
464 465 uint_t zone_ref; /* count of zone_hold()s on zone */
465 466 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */
466 467 /*
467 468 * Fixed-sized array of subsystem-specific reference counts
468 469 * The sum of all of the counts must be less than or equal to zone_ref.
469 470 * The array is indexed by the counts' subsystems' zone_ref_subsys_t
470 471 * constants.
471 472 */
472 473 uint_t zone_subsys_ref[ZONE_REF_NUM_SUBSYS];
473 474 list_t zone_ref_list; /* list of zone_ref_t structs */
474 475 /*
475 476 * zone_rootvp and zone_rootpath can never be modified once set.
476 477 */
477 478 struct vnode *zone_rootvp; /* zone's root vnode */
478 479 char *zone_rootpath; /* Path to zone's root + '/' */
479 480 ushort_t zone_flags; /* misc flags */
480 481 zone_status_t zone_status; /* protected by zone_status_lock */
481 482 uint_t zone_ntasks; /* number of tasks executing in zone */
482 483 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */
483 484 /* counters in projects and tasks */
484 485 /* that are within the zone */
485 486 rctl_qty_t zone_nlwps; /* number of lwps in zone */
486 487 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */
487 488 rctl_qty_t zone_shmmax; /* System V shared memory usage */
488 489 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */
489 490
490 491 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */
491 492 uint32_t zone_shares; /* FSS shares allocated to zone */
492 493 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */
493 494 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */
494 495 /* kpd_locked_mem for all */
495 496 /* projects in zone. */
496 497 /* Also protects zone_max_swap */
497 498 /* grab after p_lock, before rcs_lock */
498 499 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */
499 500 /* zone */
500 501 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */
501 502 /* limit. Protected by */
502 503 /* zone_rctls->rcs_lock */
503 504 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */
504 505 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */
505 506 /* Protected by */
506 507 /* zone_rctls->rcs_lock */
507 508 kmutex_t zone_rctl_lock; /* protects zone_max_lofi */
508 509 rctl_qty_t zone_max_lofi; /* lofi devs for zone */
509 510 rctl_qty_t zone_max_lofi_ctl; /* current lofi limit. */
510 511 /* Protected by */
511 512 /* zone_rctls->rcs_lock */
512 513 list_t zone_zsd; /* list of Zone-Specific Data values */
513 514 kcondvar_t zone_cv; /* used to signal state changes */
514 515 struct proc *zone_zsched; /* Dummy kernel "zsched" process */
515 516 pid_t zone_proc_initpid; /* pid of "init" for this zone */
516 517 char *zone_initname; /* fs path to 'init' */
517 518 int zone_boot_err; /* for zone_boot() if boot fails */
518 519 char *zone_bootargs; /* arguments passed via zone_boot() */
519 520 uint64_t zone_phys_mcap; /* physical memory cap */
520 521 /*
521 522 * zone_kthreads is protected by zone_status_lock.
522 523 */
523 524 kthread_t *zone_kthreads; /* kernel threads in zone */
524 525 struct priv_set *zone_privset; /* limit set for zone */
525 526 /*
526 527 * zone_vfslist is protected by vfs_list_lock().
527 528 */
528 529 struct vfs *zone_vfslist; /* list of FS's mounted in zone */
529 530 uint64_t zone_uniqid; /* unique zone generation number */
530 531 struct cred *zone_kcred; /* kcred-like, zone-limited cred */
531 532 /*
532 533 * zone_pool is protected by pool_lock().
533 534 */
534 535 struct pool *zone_pool; /* pool the zone is bound to */
535 536 hrtime_t zone_pool_mod; /* last pool bind modification time */
536 537 /* zone_psetid is protected by cpu_lock */
537 538 psetid_t zone_psetid; /* pset the zone is bound to */
538 539
539 540 time_t zone_boot_time; /* Similar to boot_time */
540 541
541 542 /*
542 543 * The following two can be read without holding any locks. They are
543 544 * updated under cpu_lock.
544 545 */
545 546 int zone_ncpus; /* zone's idea of ncpus */
546 547 int zone_ncpus_online; /* zone's idea of ncpus_online */
547 548 /*
548 549 * List of ZFS datasets exported to this zone.
549 550 */
550 551 list_t zone_datasets; /* list of datasets */
551 552
552 553 ts_label_t *zone_slabel; /* zone sensitivity label */
553 554 int zone_match; /* require label match for packets */
554 555 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */
555 556
556 557 boolean_t zone_restart_init; /* Restart init if it dies? */
557 558 struct brand *zone_brand; /* zone's brand */
558 559 void *zone_brand_data; /* store brand specific data */
559 560 id_t zone_defaultcid; /* dflt scheduling class id */
560 561 kstat_t *zone_swapresv_kstat;
561 562 kstat_t *zone_lockedmem_kstat;
562 563 /*
563 564 * zone_dl_list is protected by zone_lock
564 565 */
565 566 list_t zone_dl_list;
566 567 netstack_t *zone_netstack;
567 568 struct cpucap *zone_cpucap; /* CPU caps data */
568 569 /*
569 570 * Solaris Auditing per-zone audit context
570 571 */
571 572 struct au_kcontext *zone_audit_kctxt;
572 573 /*
573 574 * For private use by mntfs.
574 575 */
575 576 struct mntelem *zone_mntfs_db;
576 577 krwlock_t zone_mntfs_db_lock;
577 578
578 579 struct klpd_reg *zone_pfexecd;
579 580
580 581 char *zone_fs_allowed;
581 582 rctl_qty_t zone_nprocs; /* number of processes in the zone */
582 583 rctl_qty_t zone_nprocs_ctl; /* current limit protected by */
583 584 /* zone_rctls->rcs_lock */
584 585 kstat_t *zone_nprocs_kstat;
585 586
586 587 kmutex_t zone_mcap_lock; /* protects mcap statistics */
587 588 kstat_t *zone_mcap_ksp;
588 589 zone_mcap_kstat_t *zone_mcap_stats;
589 590 uint64_t zone_pgpgin; /* pages paged in */
590 591 uint64_t zone_anonpgin; /* anon pages paged in */
591 592 uint64_t zone_execpgin; /* exec pages paged in */
592 593 uint64_t zone_fspgin; /* fs pages paged in */
593 594 uint64_t zone_anon_alloc_fail; /* cnt of anon alloc fails */
594 595
595 596 psecflags_t zone_secflags; /* default zone security-flags */
596 597
597 598 /*
598 599 * Misc. kstats and counters for zone cpu-usage aggregation.
599 600 */
600 601 kmutex_t zone_misc_lock; /* protects misc statistics */
601 602 kstat_t *zone_misc_ksp;
602 603 zone_misc_kstat_t *zone_misc_stats;
603 604 /* Accumulated microstate for all threads in this zone. */
604 605 cpu_uarray_t *zone_ustate;
605 606 /* fork-fail kstat tracking */
606 607 uint32_t zone_ffcap; /* hit an rctl cap */
607 608 uint32_t zone_ffnoproc; /* get proc/lwp error */
608 609 uint32_t zone_ffnomem; /* as_dup/memory error */
609 610 uint32_t zone_ffmisc; /* misc. other error */
610 611
611 612 uint32_t zone_nested_intp; /* nested interp. kstat */
612 613
613 614 struct loadavg_s zone_loadavg; /* loadavg for this zone */
614 615 uint64_t zone_hp_avenrun[3]; /* high-precision avenrun */
615 616 int zone_avenrun[3]; /* FSCALED avg. run queue len */
616 617
617 618 /*
618 619 * FSS stats updated once per second by fss_decay_usage.
619 620 */
620 621 uint32_t zone_fss_gen; /* FSS generation cntr */
621 622 uint64_t zone_run_ticks; /* tot # of ticks running */
622 623
623 624 /*
624 625 * DTrace-private per-zone state
625 626 */
626 627 int zone_dtrace_getf; /* # of unprivileged getf()s */
627 628
628 629 /*
629 630 * Synchronization primitives used to synchronize between mounts and
630 631 * zone creation/destruction.
631 632 */
|
↓ open down ↓ |
596 lines elided |
↑ open up ↑ |
632 633 int zone_mounts_in_progress;
633 634 kcondvar_t zone_mount_cv;
634 635 kmutex_t zone_mount_lock;
635 636 } zone_t;
636 637
637 638 /*
638 639 * Special value of zone_psetid to indicate that pools are disabled.
639 640 */
640 641 #define ZONE_PS_INVAL PS_MYID
641 642
642 -
643 643 extern zone_t zone0;
644 644 extern zone_t *global_zone;
645 645 extern uint_t maxzones;
646 646 extern rctl_hndl_t rc_zone_nlwps;
647 647 extern rctl_hndl_t rc_zone_nprocs;
648 648
649 649 extern long zone(int, void *, void *, void *, void *);
650 650 extern void zone_zsd_init(void);
651 651 extern void zone_init(void);
652 652 extern void zone_hold(zone_t *);
653 653 extern void zone_rele(zone_t *);
654 654 extern void zone_init_ref(zone_ref_t *);
655 655 extern void zone_hold_ref(zone_t *, zone_ref_t *, zone_ref_subsys_t);
656 656 extern void zone_rele_ref(zone_ref_t *, zone_ref_subsys_t);
657 657 extern void zone_cred_hold(zone_t *);
658 658 extern void zone_cred_rele(zone_t *);
659 659 extern void zone_task_hold(zone_t *);
660 660 extern void zone_task_rele(zone_t *);
661 661 extern zone_t *zone_find_by_id(zoneid_t);
662 662 extern zone_t *zone_find_by_label(const ts_label_t *);
663 663 extern zone_t *zone_find_by_name(char *);
664 664 extern zone_t *zone_find_by_any_path(const char *, boolean_t);
665 665 extern zone_t *zone_find_by_path(const char *);
666 666 extern zoneid_t getzoneid(void);
667 667 extern zone_t *zone_find_by_id_nolock(zoneid_t);
668 668 extern int zone_datalink_walk(zoneid_t, int (*)(datalink_id_t, void *), void *);
669 669 extern int zone_check_datalink(zoneid_t *, datalink_id_t);
670 670 extern void zone_loadavg_update();
671 671
672 672 /*
673 673 * Zone-specific data (ZSD) APIs
674 674 */
675 675 /*
676 676 * The following is what code should be initializing its zone_key_t to if it
677 677 * calls zone_getspecific() without necessarily knowing that zone_key_create()
678 678 * has been called on the key.
679 679 */
680 680 #define ZONE_KEY_UNINITIALIZED 0
681 681
682 682 typedef uint_t zone_key_t;
683 683
684 684 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t),
685 685 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *));
686 686 extern int zone_key_delete(zone_key_t);
687 687 extern void *zone_getspecific(zone_key_t, zone_t *);
688 688 extern int zone_setspecific(zone_key_t, zone_t *, const void *);
689 689
690 690 /*
691 691 * The definition of a zsd_entry is truly private to zone.c and is only
692 692 * placed here so it can be shared with mdb.
693 693 *
694 694 * State maintained for each zone times each registered key, which tracks
695 695 * the state of the create, shutdown and destroy callbacks.
696 696 *
697 697 * zsd_flags is used to keep track of pending actions to avoid holding locks
698 698 * when calling the create/shutdown/destroy callbacks, since doing so
699 699 * could lead to deadlocks.
700 700 */
701 701 struct zsd_entry {
702 702 zone_key_t zsd_key; /* Key used to lookup value */
703 703 void *zsd_data; /* Caller-managed value */
704 704 /*
705 705 * Callbacks to be executed when a zone is created, shutdown, and
706 706 * destroyed, respectively.
707 707 */
708 708 void *(*zsd_create)(zoneid_t);
709 709 void (*zsd_shutdown)(zoneid_t, void *);
710 710 void (*zsd_destroy)(zoneid_t, void *);
711 711 list_node_t zsd_linkage;
712 712 uint16_t zsd_flags; /* See below */
713 713 kcondvar_t zsd_cv;
714 714 };
715 715
716 716 /*
717 717 * zsd_flags
718 718 */
719 719 #define ZSD_CREATE_NEEDED 0x0001
720 720 #define ZSD_CREATE_INPROGRESS 0x0002
721 721 #define ZSD_CREATE_COMPLETED 0x0004
722 722 #define ZSD_SHUTDOWN_NEEDED 0x0010
723 723 #define ZSD_SHUTDOWN_INPROGRESS 0x0020
724 724 #define ZSD_SHUTDOWN_COMPLETED 0x0040
725 725 #define ZSD_DESTROY_NEEDED 0x0100
726 726 #define ZSD_DESTROY_INPROGRESS 0x0200
727 727 #define ZSD_DESTROY_COMPLETED 0x0400
728 728
729 729 #define ZSD_CREATE_ALL \
730 730 (ZSD_CREATE_NEEDED|ZSD_CREATE_INPROGRESS|ZSD_CREATE_COMPLETED)
731 731 #define ZSD_SHUTDOWN_ALL \
732 732 (ZSD_SHUTDOWN_NEEDED|ZSD_SHUTDOWN_INPROGRESS|ZSD_SHUTDOWN_COMPLETED)
733 733 #define ZSD_DESTROY_ALL \
734 734 (ZSD_DESTROY_NEEDED|ZSD_DESTROY_INPROGRESS|ZSD_DESTROY_COMPLETED)
735 735
736 736 #define ZSD_ALL_INPROGRESS \
737 737 (ZSD_CREATE_INPROGRESS|ZSD_SHUTDOWN_INPROGRESS|ZSD_DESTROY_INPROGRESS)
738 738
739 739 /*
740 740 * Macros to help with zone visibility restrictions.
741 741 */
742 742
743 743 /*
744 744 * Is process in the global zone?
745 745 */
746 746 #define INGLOBALZONE(p) \
747 747 ((p)->p_zone == global_zone)
748 748
749 749 /*
750 750 * Can process view objects in given zone?
751 751 */
752 752 #define HASZONEACCESS(p, zoneid) \
753 753 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p))
754 754
755 755 /*
756 756 * Convenience macro to see if a resolved path is visible from within a
757 757 * given zone.
758 758 *
759 759 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the
760 760 * two strings must be equal. Since the rootpathlen has a trailing '/',
761 761 * we want to skip everything in the path up to (but not including) the
762 762 * trailing '/'.
763 763 */
764 764 #define ZONE_PATH_VISIBLE(path, zone) \
765 765 (strncmp((path), (zone)->zone_rootpath, \
766 766 (zone)->zone_rootpathlen - 1) == 0)
767 767
768 768 /*
769 769 * Convenience macro to go from the global view of a path to that seen
770 770 * from within said zone. It is the responsibility of the caller to
771 771 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is
772 772 * in fact visible from within the zone.
773 773 */
|
↓ open down ↓ |
121 lines elided |
↑ open up ↑ |
774 774 #define ZONE_PATH_TRANSLATE(path, zone) \
775 775 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \
776 776 (path) + (zone)->zone_rootpathlen - 2)
777 777
778 778 /*
779 779 * Special processes visible in all zones.
780 780 */
781 781 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1)
782 782
783 783 /*
784 + * A root vnode of the current zone.
785 + *
786 + * NOTE: It may be necessary (initialization time for file sharing where an
787 + * NGZ loads a file-sharing kernel module that does zsd initialization) to NOT
788 + * use this macro. One should ASSERT() that curzone == active ZSD (an
789 + * ASSERTion that's not always true at ZSD initialization time) during regular
790 + * use of this macro.
791 + */
792 +#define ZONE_ROOTVP() (curzone->zone_rootvp)
793 +
794 +/*
795 + * Since a zone's root isn't necessarily an actual filesystem boundary
796 + * (i.e. VROOT may not be set on zone->zone_rootvp) we need to not assume it.
797 + * This macro helps in checking if a vnode is the current zone's rootvp.
798 + * NOTE: Using the VN_ prefix, even though it's defined here in zone.h.
799 + * NOTE2: See above warning about ZONE_ROOTVP().
800 + */
801 +#define VN_IS_CURZONEROOT(vp) (VN_CMP(vp, ZONE_ROOTVP()))
802 +
803 +/*
784 804 * Zone-safe version of thread_create() to be used when the caller wants to
785 805 * create a kernel thread to run within the current zone's context.
786 806 */
787 807 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t,
788 808 pri_t);
789 809 extern void zthread_exit(void);
790 810
791 811 /*
792 812 * Functions for an external observer to register interest in a zone's status
793 813 * change. Observers will be woken up when the zone status equals the status
794 814 * argument passed in (in the case of zone_status_timedwait, the function may
795 815 * also return because of a timeout; zone_status_wait_sig may return early due
796 816 * to a signal being delivered; zone_status_timedwait_sig may return for any of
797 817 * the above reasons).
798 818 *
799 819 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and
800 820 * cv_wait_sig() respectively.
801 821 */
802 822 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t);
803 823 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t);
804 824 extern void zone_status_wait(zone_t *, zone_status_t);
805 825 extern int zone_status_wait_sig(zone_t *, zone_status_t);
806 826
807 827 /*
808 828 * Get the status of the zone (at the time it was called). The state may
809 829 * have progressed by the time it is returned.
810 830 */
811 831 extern zone_status_t zone_status_get(zone_t *);
812 832
813 833 /*
814 834 * Safely get the hostid of the specified zone (defaults to machine's hostid
815 835 * if the specified zone doesn't emulate a hostid). Passing NULL retrieves
816 836 * the global zone's (i.e., physical system's) hostid.
817 837 */
818 838 extern uint32_t zone_get_hostid(zone_t *);
819 839
820 840 /*
821 841 * Get the "kcred" credentials corresponding to the given zone.
822 842 */
823 843 extern struct cred *zone_get_kcred(zoneid_t);
824 844
825 845 /*
826 846 * Get/set the pool the zone is currently bound to.
827 847 */
828 848 extern struct pool *zone_pool_get(zone_t *);
829 849 extern void zone_pool_set(zone_t *, struct pool *);
830 850
831 851 /*
832 852 * Get/set the pset the zone is currently using.
833 853 */
834 854 extern psetid_t zone_pset_get(zone_t *);
835 855 extern void zone_pset_set(zone_t *, psetid_t);
836 856
837 857 /*
838 858 * Get the number of cpus/online-cpus visible from the given zone.
839 859 */
840 860 extern int zone_ncpus_get(zone_t *);
841 861 extern int zone_ncpus_online_get(zone_t *);
842 862
843 863 /*
844 864 * Returns true if the named pool/dataset is visible in the current zone.
845 865 */
846 866 extern int zone_dataset_visible(const char *, int *);
847 867
848 868 /*
849 869 * zone version of kadmin()
850 870 */
851 871 extern int zone_kadmin(int, int, const char *, cred_t *);
852 872 extern void zone_shutdown_global(void);
853 873
854 874 extern void mount_in_progress(zone_t *);
855 875 extern void mount_completed(zone_t *);
856 876
857 877 extern int zone_walk(int (*)(zone_t *, void *), void *);
858 878
859 879 extern rctl_hndl_t rc_zone_locked_mem;
860 880 extern rctl_hndl_t rc_zone_max_swap;
861 881 extern rctl_hndl_t rc_zone_max_lofi;
862 882
863 883 #endif /* _KERNEL */
864 884
865 885 #ifdef __cplusplus
866 886 }
867 887 #endif
868 888
869 889 #endif /* _SYS_ZONE_H */
|
↓ open down ↓ |
76 lines elided |
↑ open up ↑ |
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX