Print this page
7029 want per-process exploit mitigation features (secflags)
7030 want basic address space layout randomization (aslr)
7031 noexec_user_stack should be a secflag
7032 want a means to forbid mappings around NULL.
| 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 * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
23 23 * Copyright 2015 Joyent, Inc. All rights reserved.
24 24 * Copyright 2014 Nexenta Systems, Inc. All rights reserved.
25 25 * Copyright 2014 Igor Kozhukhov <ikozhukhov@gmail.com>.
26 26 */
27 27
28 28 #ifndef _SYS_ZONE_H
29 29 #define _SYS_ZONE_H
30 30
31 31 #include <sys/types.h>
32 32 #include <sys/mutex.h>
|
↓ open down ↓ |
32 lines elided |
↑ open up ↑ |
33 33 #include <sys/param.h>
34 34 #include <sys/rctl.h>
35 35 #include <sys/ipc_rctl.h>
36 36 #include <sys/pset.h>
37 37 #include <sys/tsol/label.h>
38 38 #include <sys/cred.h>
39 39 #include <sys/netstack.h>
40 40 #include <sys/uadmin.h>
41 41 #include <sys/ksynch.h>
42 42 #include <sys/socket_impl.h>
43 +#include <sys/secflags.h>
43 44 #include <netinet/in.h>
44 45
45 46 #ifdef __cplusplus
46 47 extern "C" {
47 48 #endif
48 49
49 50 /*
50 51 * NOTE
51 52 *
52 53 * The contents of this file are private to the implementation of
53 54 * Solaris and are subject to change at any time without notice.
54 55 * Applications and drivers using these interfaces may fail to
55 56 * run on future releases.
56 57 */
57 58
58 59 /* Available both in kernel and for user space */
59 60
60 61 /* zone id restrictions and special ids */
61 62 #define MAX_ZONEID 9999
62 63 #define MIN_USERZONEID 1 /* lowest user-creatable zone ID */
63 64 #define MIN_ZONEID 0 /* minimum zone ID on system */
64 65 #define GLOBAL_ZONEID 0
65 66 #define ZONEID_WIDTH 4 /* for printf */
66 67
67 68 /*
68 69 * Special zoneid_t token to refer to all zones.
69 70 */
70 71 #define ALL_ZONES (-1)
71 72
72 73 /* system call subcodes */
73 74 #define ZONE_CREATE 0
74 75 #define ZONE_DESTROY 1
75 76 #define ZONE_GETATTR 2
76 77 #define ZONE_ENTER 3
77 78 #define ZONE_LIST 4
78 79 #define ZONE_SHUTDOWN 5
79 80 #define ZONE_LOOKUP 6
80 81 #define ZONE_BOOT 7
81 82 #define ZONE_VERSION 8
82 83 #define ZONE_SETATTR 9
83 84 #define ZONE_ADD_DATALINK 10
84 85 #define ZONE_DEL_DATALINK 11
85 86 #define ZONE_CHECK_DATALINK 12
86 87 #define ZONE_LIST_DATALINK 13
87 88
88 89 /* zone attributes */
89 90 #define ZONE_ATTR_ROOT 1
90 91 #define ZONE_ATTR_NAME 2
91 92 #define ZONE_ATTR_STATUS 3
92 93 #define ZONE_ATTR_PRIVSET 4
93 94 #define ZONE_ATTR_UNIQID 5
94 95 #define ZONE_ATTR_POOLID 6
95 96 #define ZONE_ATTR_INITPID 7
96 97 #define ZONE_ATTR_SLBL 8
|
↓ open down ↓ |
44 lines elided |
↑ open up ↑ |
97 98 #define ZONE_ATTR_INITNAME 9
98 99 #define ZONE_ATTR_BOOTARGS 10
99 100 #define ZONE_ATTR_BRAND 11
100 101 #define ZONE_ATTR_PHYS_MCAP 12
101 102 #define ZONE_ATTR_SCHED_CLASS 13
102 103 #define ZONE_ATTR_FLAGS 14
103 104 #define ZONE_ATTR_HOSTID 15
104 105 #define ZONE_ATTR_FS_ALLOWED 16
105 106 #define ZONE_ATTR_NETWORK 17
106 107 #define ZONE_ATTR_INITNORESTART 20
108 +#define ZONE_ATTR_SECFLAGS 21
107 109
108 110 /* Start of the brand-specific attribute namespace */
109 111 #define ZONE_ATTR_BRAND_ATTRS 32768
110 112
111 113 #define ZONE_FS_ALLOWED_MAX 1024
112 114
113 115 #define ZONE_EVENT_CHANNEL "com.sun:zones:status"
114 116 #define ZONE_EVENT_STATUS_CLASS "status"
115 117 #define ZONE_EVENT_STATUS_SUBCLASS "change"
116 118
117 119 #define ZONE_EVENT_UNINITIALIZED "uninitialized"
118 120 #define ZONE_EVENT_INITIALIZED "initialized"
119 121 #define ZONE_EVENT_READY "ready"
120 122 #define ZONE_EVENT_RUNNING "running"
121 123 #define ZONE_EVENT_SHUTTING_DOWN "shutting_down"
122 124
123 125 #define ZONE_CB_NAME "zonename"
124 126 #define ZONE_CB_NEWSTATE "newstate"
125 127 #define ZONE_CB_OLDSTATE "oldstate"
126 128 #define ZONE_CB_TIMESTAMP "when"
127 129 #define ZONE_CB_ZONEID "zoneid"
128 130
129 131 /*
130 132 * Exit values that may be returned by scripts or programs invoked by various
131 133 * zone commands.
132 134 *
133 135 * These are defined as:
134 136 *
135 137 * ZONE_SUBPROC_OK
136 138 * ===============
137 139 * The subprocess completed successfully.
138 140 *
139 141 * ZONE_SUBPROC_USAGE
140 142 * ==================
141 143 * The subprocess failed with a usage message, or a usage message should
142 144 * be output in its behalf.
143 145 *
144 146 * ZONE_SUBPROC_NOTCOMPLETE
145 147 * ========================
146 148 * The subprocess did not complete, but the actions performed by the
147 149 * subprocess require no recovery actions by the user.
148 150 *
149 151 * For example, if the subprocess were called by "zoneadm install," the
150 152 * installation of the zone did not succeed but the user need not perform
151 153 * a "zoneadm uninstall" before attempting another install.
152 154 *
153 155 * ZONE_SUBPROC_FATAL
154 156 * ==================
155 157 * The subprocess failed in a fatal manner, usually one that will require
156 158 * some type of recovery action by the user.
157 159 *
158 160 * For example, if the subprocess were called by "zoneadm install," the
159 161 * installation of the zone did not succeed and the user will need to
160 162 * perform a "zoneadm uninstall" before another install attempt is
161 163 * possible.
162 164 *
163 165 * The non-success exit values are large to avoid accidental collision
164 166 * with values used internally by some commands (e.g. "Z_ERR" and
165 167 * "Z_USAGE" as used by zoneadm.)
166 168 */
167 169 #define ZONE_SUBPROC_OK 0
168 170 #define ZONE_SUBPROC_USAGE 253
169 171 #define ZONE_SUBPROC_NOTCOMPLETE 254
170 172 #define ZONE_SUBPROC_FATAL 255
171 173
172 174 #ifdef _SYSCALL32
173 175 typedef struct {
174 176 caddr32_t zone_name;
175 177 caddr32_t zone_root;
176 178 caddr32_t zone_privs;
177 179 size32_t zone_privssz;
178 180 caddr32_t rctlbuf;
179 181 size32_t rctlbufsz;
180 182 caddr32_t extended_error;
181 183 caddr32_t zfsbuf;
182 184 size32_t zfsbufsz;
183 185 int match; /* match level */
184 186 uint32_t doi; /* DOI for label */
185 187 caddr32_t label; /* label associated with zone */
186 188 int flags;
187 189 } zone_def32;
188 190 #endif
189 191 typedef struct {
190 192 const char *zone_name;
191 193 const char *zone_root;
192 194 const struct priv_set *zone_privs;
193 195 size_t zone_privssz;
194 196 const char *rctlbuf;
195 197 size_t rctlbufsz;
196 198 int *extended_error;
197 199 const char *zfsbuf;
198 200 size_t zfsbufsz;
199 201 int match; /* match level */
200 202 uint32_t doi; /* DOI for label */
201 203 const bslabel_t *label; /* label associated with zone */
202 204 int flags;
203 205 } zone_def;
204 206
205 207 /* extended error information */
206 208 #define ZE_UNKNOWN 0 /* No extended error info */
207 209 #define ZE_CHROOTED 1 /* tried to zone_create from chroot */
208 210 #define ZE_AREMOUNTS 2 /* there are mounts within the zone */
209 211 #define ZE_LABELINUSE 3 /* label is already in use by some other zone */
210 212
211 213 /*
212 214 * zone_status values
213 215 *
214 216 * You must modify zone_status_names in mdb(1M)'s genunix module
215 217 * (genunix/zone.c) when you modify this enum.
216 218 */
217 219 typedef enum {
218 220 ZONE_IS_UNINITIALIZED = 0,
219 221 ZONE_IS_INITIALIZED,
220 222 ZONE_IS_READY,
221 223 ZONE_IS_BOOTING,
222 224 ZONE_IS_RUNNING,
223 225 ZONE_IS_SHUTTING_DOWN,
224 226 ZONE_IS_EMPTY,
225 227 ZONE_IS_DOWN,
226 228 ZONE_IS_DYING,
227 229 ZONE_IS_DEAD
228 230 } zone_status_t;
229 231 #define ZONE_MIN_STATE ZONE_IS_UNINITIALIZED
230 232 #define ZONE_MAX_STATE ZONE_IS_DEAD
231 233
232 234 /*
233 235 * Valid commands which may be issued by zoneadm to zoneadmd. The kernel also
234 236 * communicates with zoneadmd, but only uses Z_REBOOT and Z_HALT.
235 237 */
236 238 typedef enum zone_cmd {
237 239 Z_READY, Z_BOOT, Z_FORCEBOOT, Z_REBOOT, Z_HALT, Z_NOTE_UNINSTALLING,
238 240 Z_MOUNT, Z_FORCEMOUNT, Z_UNMOUNT, Z_SHUTDOWN
239 241 } zone_cmd_t;
240 242
241 243 /*
242 244 * The structure of a request to zoneadmd.
243 245 */
244 246 typedef struct zone_cmd_arg {
245 247 uint64_t uniqid; /* unique "generation number" */
246 248 zone_cmd_t cmd; /* requested action */
247 249 uint32_t _pad; /* need consistent 32/64 bit alignmt */
248 250 char locale[MAXPATHLEN]; /* locale in which to render messages */
249 251 char bootbuf[BOOTARGS_MAX]; /* arguments passed to zone_boot() */
250 252 } zone_cmd_arg_t;
251 253
252 254 /*
253 255 * Structure of zoneadmd's response to a request. A NULL return value means
254 256 * the caller should attempt to restart zoneadmd and retry.
255 257 */
256 258 typedef struct zone_cmd_rval {
257 259 int rval; /* return value of request */
258 260 char errbuf[1]; /* variable-sized buffer containing error messages */
259 261 } zone_cmd_rval_t;
260 262
261 263 /*
262 264 * The zone support infrastructure uses the zone name as a component
263 265 * of unix domain (AF_UNIX) sockets, which are limited to 108 characters
264 266 * in length, so ZONENAME_MAX is limited by that.
265 267 */
266 268 #define ZONENAME_MAX 64
267 269
268 270 #define GLOBAL_ZONENAME "global"
269 271
270 272 /*
271 273 * Extended Regular expression (see regex(5)) which matches all valid zone
272 274 * names.
273 275 */
274 276 #define ZONENAME_REGEXP "[a-zA-Z0-9][-_.a-zA-Z0-9]{0,62}"
275 277
276 278 /*
277 279 * Where the zones support infrastructure places temporary files.
278 280 */
279 281 #define ZONES_TMPDIR "/var/run/zones"
280 282
281 283 /*
282 284 * The path to the door used by clients to communicate with zoneadmd.
283 285 */
284 286 #define ZONE_DOOR_PATH ZONES_TMPDIR "/%s.zoneadmd_door"
285 287
286 288
287 289 /* zone_flags */
288 290 /*
289 291 * Threads that read or write the following flag must hold zone_lock.
290 292 */
291 293 #define ZF_REFCOUNTS_LOGGED 0x1 /* a thread logged the zone's refs */
292 294
293 295 /*
294 296 * The following threads are set when the zone is created and never changed.
295 297 * Threads that test for these flags don't have to hold zone_lock.
296 298 */
297 299 #define ZF_HASHED_LABEL 0x2 /* zone has a unique label */
298 300 #define ZF_IS_SCRATCH 0x4 /* scratch zone */
299 301 #define ZF_NET_EXCL 0x8 /* Zone has an exclusive IP stack */
300 302
301 303
302 304 /* zone_create flags */
303 305 #define ZCF_NET_EXCL 0x1 /* Create a zone with exclusive IP */
304 306
305 307 /* zone network properties */
306 308 #define ZONE_NETWORK_ADDRESS 1
307 309 #define ZONE_NETWORK_DEFROUTER 2
308 310
309 311 #define ZONE_NET_ADDRNAME "address"
310 312 #define ZONE_NET_RTRNAME "route"
311 313
312 314 typedef struct zone_net_data {
313 315 int zn_type;
314 316 int zn_len;
315 317 datalink_id_t zn_linkid;
316 318 uint8_t zn_val[1];
317 319 } zone_net_data_t;
318 320
319 321
320 322 #ifdef _KERNEL
321 323
322 324 /*
323 325 * We need to protect the definition of 'list_t' from userland applications and
324 326 * libraries which may be defining ther own versions.
325 327 */
326 328 #include <sys/list.h>
327 329 #include <sys/loadavg.h>
328 330
329 331 #define GLOBAL_ZONEUNIQID 0 /* uniqid of the global zone */
330 332
331 333 struct pool;
332 334 struct brand;
333 335
334 336 /*
335 337 * Each of these constants identifies a kernel subsystem that acquires and
336 338 * releases zone references. Each subsystem that invokes
337 339 * zone_hold_ref() and zone_rele_ref() should specify the
338 340 * zone_ref_subsys_t constant associated with the subsystem. Tracked holds
339 341 * help users and developers quickly identify subsystems that stall zone
340 342 * shutdowns indefinitely.
341 343 *
342 344 * NOTE: You must modify zone_ref_subsys_names in usr/src/uts/common/os/zone.c
343 345 * when you modify this enumeration.
344 346 */
345 347 typedef enum zone_ref_subsys {
346 348 ZONE_REF_NFS, /* NFS */
347 349 ZONE_REF_NFSV4, /* NFSv4 */
348 350 ZONE_REF_SMBFS, /* SMBFS */
349 351 ZONE_REF_MNTFS, /* MNTFS */
350 352 ZONE_REF_LOFI, /* LOFI devices */
351 353 ZONE_REF_VFS, /* VFS infrastructure */
352 354 ZONE_REF_IPC, /* IPC infrastructure */
353 355 ZONE_REF_NUM_SUBSYS /* This must be the last entry. */
354 356 } zone_ref_subsys_t;
355 357
356 358 /*
357 359 * zone_ref represents a general-purpose references to a zone. Each zone's
358 360 * references are linked into the zone's zone_t::zone_ref_list. This allows
359 361 * debuggers to walk zones' references.
360 362 */
361 363 typedef struct zone_ref {
362 364 struct zone *zref_zone; /* the zone to which the reference refers */
363 365 list_node_t zref_linkage; /* linkage for zone_t::zone_ref_list */
364 366 } zone_ref_t;
365 367
366 368 /*
367 369 * Structure to record list of ZFS datasets exported to a zone.
368 370 */
369 371 typedef struct zone_dataset {
370 372 char *zd_dataset;
371 373 list_node_t zd_linkage;
372 374 } zone_dataset_t;
373 375
374 376 /*
375 377 * structure for zone kstats
376 378 */
377 379 typedef struct zone_kstat {
378 380 kstat_named_t zk_zonename;
379 381 kstat_named_t zk_usage;
380 382 kstat_named_t zk_value;
381 383 } zone_kstat_t;
382 384
383 385 struct cpucap;
384 386
385 387 typedef struct {
386 388 kstat_named_t zm_zonename;
387 389 kstat_named_t zm_pgpgin;
388 390 kstat_named_t zm_anonpgin;
389 391 kstat_named_t zm_execpgin;
390 392 kstat_named_t zm_fspgin;
391 393 kstat_named_t zm_anon_alloc_fail;
392 394 } zone_mcap_kstat_t;
393 395
394 396 typedef struct {
395 397 kstat_named_t zm_zonename; /* full name, kstat truncates name */
396 398 kstat_named_t zm_utime;
397 399 kstat_named_t zm_stime;
398 400 kstat_named_t zm_wtime;
399 401 kstat_named_t zm_avenrun1;
400 402 kstat_named_t zm_avenrun5;
401 403 kstat_named_t zm_avenrun15;
402 404 kstat_named_t zm_ffcap;
403 405 kstat_named_t zm_ffnoproc;
404 406 kstat_named_t zm_ffnomem;
405 407 kstat_named_t zm_ffmisc;
406 408 kstat_named_t zm_nested_intp;
407 409 kstat_named_t zm_init_pid;
408 410 kstat_named_t zm_boot_time;
409 411 } zone_misc_kstat_t;
410 412
411 413 typedef struct zone {
412 414 /*
413 415 * zone_name is never modified once set.
414 416 */
415 417 char *zone_name; /* zone's configuration name */
416 418 /*
417 419 * zone_nodename and zone_domain are never freed once allocated.
418 420 */
419 421 char *zone_nodename; /* utsname.nodename equivalent */
420 422 char *zone_domain; /* srpc_domain equivalent */
421 423 /*
422 424 * zone_hostid is used for per-zone hostid emulation.
423 425 * Currently it isn't modified after it's set (so no locks protect
424 426 * accesses), but that might have to change when we allow
425 427 * administrators to change running zones' properties.
426 428 *
427 429 * The global zone's zone_hostid must always be HW_INVALID_HOSTID so
428 430 * that zone_get_hostid() will function correctly.
429 431 */
430 432 uint32_t zone_hostid; /* zone's hostid, HW_INVALID_HOSTID */
431 433 /* if not emulated */
432 434 /*
433 435 * zone_lock protects the following fields of a zone_t:
434 436 * zone_ref
435 437 * zone_cred_ref
436 438 * zone_subsys_ref
437 439 * zone_ref_list
438 440 * zone_ntasks
439 441 * zone_flags
440 442 * zone_zsd
441 443 * zone_pfexecd
442 444 */
443 445 kmutex_t zone_lock;
444 446 /*
445 447 * zone_linkage is the zone's linkage into the active or
446 448 * death-row list. The field is protected by zonehash_lock.
447 449 */
448 450 list_node_t zone_linkage;
449 451 zoneid_t zone_id; /* ID of zone */
450 452 uint_t zone_ref; /* count of zone_hold()s on zone */
451 453 uint_t zone_cred_ref; /* count of zone_hold_cred()s on zone */
452 454 /*
453 455 * Fixed-sized array of subsystem-specific reference counts
454 456 * The sum of all of the counts must be less than or equal to zone_ref.
455 457 * The array is indexed by the counts' subsystems' zone_ref_subsys_t
456 458 * constants.
457 459 */
458 460 uint_t zone_subsys_ref[ZONE_REF_NUM_SUBSYS];
459 461 list_t zone_ref_list; /* list of zone_ref_t structs */
460 462 /*
461 463 * zone_rootvp and zone_rootpath can never be modified once set.
462 464 */
463 465 struct vnode *zone_rootvp; /* zone's root vnode */
464 466 char *zone_rootpath; /* Path to zone's root + '/' */
465 467 ushort_t zone_flags; /* misc flags */
466 468 zone_status_t zone_status; /* protected by zone_status_lock */
467 469 uint_t zone_ntasks; /* number of tasks executing in zone */
468 470 kmutex_t zone_nlwps_lock; /* protects zone_nlwps, and *_nlwps */
469 471 /* counters in projects and tasks */
470 472 /* that are within the zone */
471 473 rctl_qty_t zone_nlwps; /* number of lwps in zone */
472 474 rctl_qty_t zone_nlwps_ctl; /* protected by zone_rctls->rcs_lock */
473 475 rctl_qty_t zone_shmmax; /* System V shared memory usage */
474 476 ipc_rqty_t zone_ipc; /* System V IPC id resource usage */
475 477
476 478 uint_t zone_rootpathlen; /* strlen(zone_rootpath) + 1 */
477 479 uint32_t zone_shares; /* FSS shares allocated to zone */
478 480 rctl_set_t *zone_rctls; /* zone-wide (zone.*) rctls */
479 481 kmutex_t zone_mem_lock; /* protects zone_locked_mem and */
480 482 /* kpd_locked_mem for all */
481 483 /* projects in zone. */
482 484 /* Also protects zone_max_swap */
483 485 /* grab after p_lock, before rcs_lock */
484 486 rctl_qty_t zone_locked_mem; /* bytes of locked memory in */
485 487 /* zone */
486 488 rctl_qty_t zone_locked_mem_ctl; /* Current locked memory */
487 489 /* limit. Protected by */
488 490 /* zone_rctls->rcs_lock */
489 491 rctl_qty_t zone_max_swap; /* bytes of swap reserved by zone */
490 492 rctl_qty_t zone_max_swap_ctl; /* current swap limit. */
491 493 /* Protected by */
492 494 /* zone_rctls->rcs_lock */
493 495 kmutex_t zone_rctl_lock; /* protects zone_max_lofi */
494 496 rctl_qty_t zone_max_lofi; /* lofi devs for zone */
495 497 rctl_qty_t zone_max_lofi_ctl; /* current lofi limit. */
496 498 /* Protected by */
497 499 /* zone_rctls->rcs_lock */
498 500 list_t zone_zsd; /* list of Zone-Specific Data values */
499 501 kcondvar_t zone_cv; /* used to signal state changes */
500 502 struct proc *zone_zsched; /* Dummy kernel "zsched" process */
501 503 pid_t zone_proc_initpid; /* pid of "init" for this zone */
502 504 char *zone_initname; /* fs path to 'init' */
503 505 int zone_boot_err; /* for zone_boot() if boot fails */
504 506 char *zone_bootargs; /* arguments passed via zone_boot() */
505 507 uint64_t zone_phys_mcap; /* physical memory cap */
506 508 /*
507 509 * zone_kthreads is protected by zone_status_lock.
508 510 */
509 511 kthread_t *zone_kthreads; /* kernel threads in zone */
510 512 struct priv_set *zone_privset; /* limit set for zone */
511 513 /*
512 514 * zone_vfslist is protected by vfs_list_lock().
513 515 */
514 516 struct vfs *zone_vfslist; /* list of FS's mounted in zone */
515 517 uint64_t zone_uniqid; /* unique zone generation number */
516 518 struct cred *zone_kcred; /* kcred-like, zone-limited cred */
517 519 /*
518 520 * zone_pool is protected by pool_lock().
519 521 */
520 522 struct pool *zone_pool; /* pool the zone is bound to */
521 523 hrtime_t zone_pool_mod; /* last pool bind modification time */
522 524 /* zone_psetid is protected by cpu_lock */
523 525 psetid_t zone_psetid; /* pset the zone is bound to */
524 526
525 527 time_t zone_boot_time; /* Similar to boot_time */
526 528
527 529 /*
528 530 * The following two can be read without holding any locks. They are
529 531 * updated under cpu_lock.
530 532 */
531 533 int zone_ncpus; /* zone's idea of ncpus */
532 534 int zone_ncpus_online; /* zone's idea of ncpus_online */
533 535 /*
534 536 * List of ZFS datasets exported to this zone.
535 537 */
536 538 list_t zone_datasets; /* list of datasets */
537 539
538 540 ts_label_t *zone_slabel; /* zone sensitivity label */
539 541 int zone_match; /* require label match for packets */
540 542 tsol_mlp_list_t zone_mlps; /* MLPs on zone-private addresses */
541 543
542 544 boolean_t zone_restart_init; /* Restart init if it dies? */
543 545 struct brand *zone_brand; /* zone's brand */
544 546 void *zone_brand_data; /* store brand specific data */
545 547 id_t zone_defaultcid; /* dflt scheduling class id */
546 548 kstat_t *zone_swapresv_kstat;
547 549 kstat_t *zone_lockedmem_kstat;
548 550 /*
549 551 * zone_dl_list is protected by zone_lock
550 552 */
551 553 list_t zone_dl_list;
552 554 netstack_t *zone_netstack;
553 555 struct cpucap *zone_cpucap; /* CPU caps data */
554 556 /*
555 557 * Solaris Auditing per-zone audit context
556 558 */
557 559 struct au_kcontext *zone_audit_kctxt;
558 560 /*
559 561 * For private use by mntfs.
560 562 */
561 563 struct mntelem *zone_mntfs_db;
562 564 krwlock_t zone_mntfs_db_lock;
563 565
564 566 struct klpd_reg *zone_pfexecd;
565 567
566 568 char *zone_fs_allowed;
567 569 rctl_qty_t zone_nprocs; /* number of processes in the zone */
568 570 rctl_qty_t zone_nprocs_ctl; /* current limit protected by */
569 571 /* zone_rctls->rcs_lock */
570 572 kstat_t *zone_nprocs_kstat;
|
↓ open down ↓ |
454 lines elided |
↑ open up ↑ |
571 573
572 574 kmutex_t zone_mcap_lock; /* protects mcap statistics */
573 575 kstat_t *zone_mcap_ksp;
574 576 zone_mcap_kstat_t *zone_mcap_stats;
575 577 uint64_t zone_pgpgin; /* pages paged in */
576 578 uint64_t zone_anonpgin; /* anon pages paged in */
577 579 uint64_t zone_execpgin; /* exec pages paged in */
578 580 uint64_t zone_fspgin; /* fs pages paged in */
579 581 uint64_t zone_anon_alloc_fail; /* cnt of anon alloc fails */
580 582
583 + psecflags_t zone_secflags; /* default zone security-flags */
584 +
581 585 /*
582 586 * Misc. kstats and counters for zone cpu-usage aggregation.
583 587 * The zone_Xtime values are the sum of the micro-state accounting
584 588 * values for all threads that are running or have run in the zone.
585 589 * This is tracked in msacct.c as threads change state.
586 590 * The zone_stime is the sum of the LMS_SYSTEM times.
587 591 * The zone_utime is the sum of the LMS_USER times.
588 592 * The zone_wtime is the sum of the LMS_WAIT_CPU times.
589 593 * As with per-thread micro-state accounting values, these values are
590 594 * not scaled to nanosecs. The scaling is done by the
591 595 * zone_misc_kstat_update function when kstats are requested.
592 596 */
593 597 kmutex_t zone_misc_lock; /* protects misc statistics */
594 598 kstat_t *zone_misc_ksp;
595 599 zone_misc_kstat_t *zone_misc_stats;
596 600 uint64_t zone_stime; /* total system time */
597 601 uint64_t zone_utime; /* total user time */
598 602 uint64_t zone_wtime; /* total time waiting in runq */
599 603 /* fork-fail kstat tracking */
600 604 uint32_t zone_ffcap; /* hit an rctl cap */
601 605 uint32_t zone_ffnoproc; /* get proc/lwp error */
602 606 uint32_t zone_ffnomem; /* as_dup/memory error */
603 607 uint32_t zone_ffmisc; /* misc. other error */
604 608
605 609 uint32_t zone_nested_intp; /* nested interp. kstat */
606 610
607 611 struct loadavg_s zone_loadavg; /* loadavg for this zone */
608 612 uint64_t zone_hp_avenrun[3]; /* high-precision avenrun */
609 613 int zone_avenrun[3]; /* FSCALED avg. run queue len */
610 614
611 615 /*
612 616 * FSS stats updated once per second by fss_decay_usage.
613 617 */
614 618 uint32_t zone_fss_gen; /* FSS generation cntr */
615 619 uint64_t zone_run_ticks; /* tot # of ticks running */
616 620
617 621 /*
618 622 * DTrace-private per-zone state
619 623 */
620 624 int zone_dtrace_getf; /* # of unprivileged getf()s */
621 625
622 626 /*
623 627 * Synchronization primitives used to synchronize between mounts and
624 628 * zone creation/destruction.
625 629 */
626 630 int zone_mounts_in_progress;
627 631 kcondvar_t zone_mount_cv;
628 632 kmutex_t zone_mount_lock;
629 633 } zone_t;
630 634
631 635 /*
632 636 * Special value of zone_psetid to indicate that pools are disabled.
633 637 */
634 638 #define ZONE_PS_INVAL PS_MYID
635 639
636 640
637 641 extern zone_t zone0;
638 642 extern zone_t *global_zone;
639 643 extern uint_t maxzones;
640 644 extern rctl_hndl_t rc_zone_nlwps;
641 645 extern rctl_hndl_t rc_zone_nprocs;
642 646
643 647 extern long zone(int, void *, void *, void *, void *);
644 648 extern void zone_zsd_init(void);
645 649 extern void zone_init(void);
646 650 extern void zone_hold(zone_t *);
647 651 extern void zone_rele(zone_t *);
648 652 extern void zone_init_ref(zone_ref_t *);
649 653 extern void zone_hold_ref(zone_t *, zone_ref_t *, zone_ref_subsys_t);
650 654 extern void zone_rele_ref(zone_ref_t *, zone_ref_subsys_t);
651 655 extern void zone_cred_hold(zone_t *);
652 656 extern void zone_cred_rele(zone_t *);
653 657 extern void zone_task_hold(zone_t *);
654 658 extern void zone_task_rele(zone_t *);
655 659 extern zone_t *zone_find_by_id(zoneid_t);
656 660 extern zone_t *zone_find_by_label(const ts_label_t *);
657 661 extern zone_t *zone_find_by_name(char *);
658 662 extern zone_t *zone_find_by_any_path(const char *, boolean_t);
659 663 extern zone_t *zone_find_by_path(const char *);
660 664 extern zoneid_t getzoneid(void);
661 665 extern zone_t *zone_find_by_id_nolock(zoneid_t);
662 666 extern int zone_datalink_walk(zoneid_t, int (*)(datalink_id_t, void *), void *);
663 667 extern int zone_check_datalink(zoneid_t *, datalink_id_t);
664 668 extern void zone_loadavg_update();
665 669
666 670 /*
667 671 * Zone-specific data (ZSD) APIs
668 672 */
669 673 /*
670 674 * The following is what code should be initializing its zone_key_t to if it
671 675 * calls zone_getspecific() without necessarily knowing that zone_key_create()
672 676 * has been called on the key.
673 677 */
674 678 #define ZONE_KEY_UNINITIALIZED 0
675 679
676 680 typedef uint_t zone_key_t;
677 681
678 682 extern void zone_key_create(zone_key_t *, void *(*)(zoneid_t),
679 683 void (*)(zoneid_t, void *), void (*)(zoneid_t, void *));
680 684 extern int zone_key_delete(zone_key_t);
681 685 extern void *zone_getspecific(zone_key_t, zone_t *);
682 686 extern int zone_setspecific(zone_key_t, zone_t *, const void *);
683 687
684 688 /*
685 689 * The definition of a zsd_entry is truly private to zone.c and is only
686 690 * placed here so it can be shared with mdb.
687 691 *
688 692 * State maintained for each zone times each registered key, which tracks
689 693 * the state of the create, shutdown and destroy callbacks.
690 694 *
691 695 * zsd_flags is used to keep track of pending actions to avoid holding locks
692 696 * when calling the create/shutdown/destroy callbacks, since doing so
693 697 * could lead to deadlocks.
694 698 */
695 699 struct zsd_entry {
696 700 zone_key_t zsd_key; /* Key used to lookup value */
697 701 void *zsd_data; /* Caller-managed value */
698 702 /*
699 703 * Callbacks to be executed when a zone is created, shutdown, and
700 704 * destroyed, respectively.
701 705 */
702 706 void *(*zsd_create)(zoneid_t);
703 707 void (*zsd_shutdown)(zoneid_t, void *);
704 708 void (*zsd_destroy)(zoneid_t, void *);
705 709 list_node_t zsd_linkage;
706 710 uint16_t zsd_flags; /* See below */
707 711 kcondvar_t zsd_cv;
708 712 };
709 713
710 714 /*
711 715 * zsd_flags
712 716 */
713 717 #define ZSD_CREATE_NEEDED 0x0001
714 718 #define ZSD_CREATE_INPROGRESS 0x0002
715 719 #define ZSD_CREATE_COMPLETED 0x0004
716 720 #define ZSD_SHUTDOWN_NEEDED 0x0010
717 721 #define ZSD_SHUTDOWN_INPROGRESS 0x0020
718 722 #define ZSD_SHUTDOWN_COMPLETED 0x0040
719 723 #define ZSD_DESTROY_NEEDED 0x0100
720 724 #define ZSD_DESTROY_INPROGRESS 0x0200
721 725 #define ZSD_DESTROY_COMPLETED 0x0400
722 726
723 727 #define ZSD_CREATE_ALL \
724 728 (ZSD_CREATE_NEEDED|ZSD_CREATE_INPROGRESS|ZSD_CREATE_COMPLETED)
725 729 #define ZSD_SHUTDOWN_ALL \
726 730 (ZSD_SHUTDOWN_NEEDED|ZSD_SHUTDOWN_INPROGRESS|ZSD_SHUTDOWN_COMPLETED)
727 731 #define ZSD_DESTROY_ALL \
728 732 (ZSD_DESTROY_NEEDED|ZSD_DESTROY_INPROGRESS|ZSD_DESTROY_COMPLETED)
729 733
730 734 #define ZSD_ALL_INPROGRESS \
731 735 (ZSD_CREATE_INPROGRESS|ZSD_SHUTDOWN_INPROGRESS|ZSD_DESTROY_INPROGRESS)
732 736
733 737 /*
734 738 * Macros to help with zone visibility restrictions.
735 739 */
736 740
737 741 /*
738 742 * Is process in the global zone?
739 743 */
740 744 #define INGLOBALZONE(p) \
741 745 ((p)->p_zone == global_zone)
742 746
743 747 /*
744 748 * Can process view objects in given zone?
745 749 */
746 750 #define HASZONEACCESS(p, zoneid) \
747 751 ((p)->p_zone->zone_id == (zoneid) || INGLOBALZONE(p))
748 752
749 753 /*
750 754 * Convenience macro to see if a resolved path is visible from within a
751 755 * given zone.
752 756 *
753 757 * The basic idea is that the first (zone_rootpathlen - 1) bytes of the
754 758 * two strings must be equal. Since the rootpathlen has a trailing '/',
755 759 * we want to skip everything in the path up to (but not including) the
756 760 * trailing '/'.
757 761 */
758 762 #define ZONE_PATH_VISIBLE(path, zone) \
759 763 (strncmp((path), (zone)->zone_rootpath, \
760 764 (zone)->zone_rootpathlen - 1) == 0)
761 765
762 766 /*
763 767 * Convenience macro to go from the global view of a path to that seen
764 768 * from within said zone. It is the responsibility of the caller to
765 769 * ensure that the path is a resolved one (ie, no '..'s or '.'s), and is
766 770 * in fact visible from within the zone.
767 771 */
768 772 #define ZONE_PATH_TRANSLATE(path, zone) \
769 773 (ASSERT(ZONE_PATH_VISIBLE(path, zone)), \
770 774 (path) + (zone)->zone_rootpathlen - 2)
771 775
772 776 /*
773 777 * Special processes visible in all zones.
774 778 */
775 779 #define ZONE_SPECIALPID(x) ((x) == 0 || (x) == 1)
776 780
777 781 /*
778 782 * Zone-safe version of thread_create() to be used when the caller wants to
779 783 * create a kernel thread to run within the current zone's context.
780 784 */
781 785 extern kthread_t *zthread_create(caddr_t, size_t, void (*)(), void *, size_t,
782 786 pri_t);
783 787 extern void zthread_exit(void);
784 788
785 789 /*
786 790 * Functions for an external observer to register interest in a zone's status
787 791 * change. Observers will be woken up when the zone status equals the status
788 792 * argument passed in (in the case of zone_status_timedwait, the function may
789 793 * also return because of a timeout; zone_status_wait_sig may return early due
790 794 * to a signal being delivered; zone_status_timedwait_sig may return for any of
791 795 * the above reasons).
792 796 *
793 797 * Otherwise these behave identically to cv_timedwait(), cv_wait(), and
794 798 * cv_wait_sig() respectively.
795 799 */
796 800 extern clock_t zone_status_timedwait(zone_t *, clock_t, zone_status_t);
797 801 extern clock_t zone_status_timedwait_sig(zone_t *, clock_t, zone_status_t);
798 802 extern void zone_status_wait(zone_t *, zone_status_t);
799 803 extern int zone_status_wait_sig(zone_t *, zone_status_t);
800 804
801 805 /*
802 806 * Get the status of the zone (at the time it was called). The state may
803 807 * have progressed by the time it is returned.
804 808 */
805 809 extern zone_status_t zone_status_get(zone_t *);
806 810
807 811 /*
808 812 * Safely get the hostid of the specified zone (defaults to machine's hostid
809 813 * if the specified zone doesn't emulate a hostid). Passing NULL retrieves
810 814 * the global zone's (i.e., physical system's) hostid.
811 815 */
812 816 extern uint32_t zone_get_hostid(zone_t *);
813 817
814 818 /*
815 819 * Get the "kcred" credentials corresponding to the given zone.
816 820 */
817 821 extern struct cred *zone_get_kcred(zoneid_t);
818 822
819 823 /*
820 824 * Get/set the pool the zone is currently bound to.
821 825 */
822 826 extern struct pool *zone_pool_get(zone_t *);
823 827 extern void zone_pool_set(zone_t *, struct pool *);
824 828
825 829 /*
826 830 * Get/set the pset the zone is currently using.
827 831 */
828 832 extern psetid_t zone_pset_get(zone_t *);
829 833 extern void zone_pset_set(zone_t *, psetid_t);
830 834
831 835 /*
832 836 * Get the number of cpus/online-cpus visible from the given zone.
833 837 */
834 838 extern int zone_ncpus_get(zone_t *);
835 839 extern int zone_ncpus_online_get(zone_t *);
836 840
837 841 /*
838 842 * Returns true if the named pool/dataset is visible in the current zone.
839 843 */
840 844 extern int zone_dataset_visible(const char *, int *);
841 845
842 846 /*
843 847 * zone version of kadmin()
844 848 */
845 849 extern int zone_kadmin(int, int, const char *, cred_t *);
846 850 extern void zone_shutdown_global(void);
847 851
848 852 extern void mount_in_progress(zone_t *);
849 853 extern void mount_completed(zone_t *);
850 854
851 855 extern int zone_walk(int (*)(zone_t *, void *), void *);
852 856
853 857 extern rctl_hndl_t rc_zone_locked_mem;
854 858 extern rctl_hndl_t rc_zone_max_swap;
855 859 extern rctl_hndl_t rc_zone_max_lofi;
856 860
857 861 #endif /* _KERNEL */
858 862
859 863 #ifdef __cplusplus
860 864 }
861 865 #endif
862 866
863 867 #endif /* _SYS_ZONE_H */
|
↓ open down ↓ |
273 lines elided |
↑ open up ↑ |
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX