zone.c revision cf8f45c7690afabe63bdb8066b11db58d708ad09
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License, Version 1.0 only
* (the "License"). You may not use this file except in compliance
* with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or http://www.opensolaris.org/os/licensing.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright 2005 Sun Microsystems, Inc. All rights reserved.
* Use is subject to license terms.
*/
#pragma ident "%Z%%M% %I% %E% SMI"
/*
* Zones
*
* A zone is a named collection of processes, namespace constraints,
* and other system resources which comprise a secure and manageable
* application containment facility.
*
* Zones (represented by the reference counted zone_t) are tracked in
* the kernel in the zonehash. Elsewhere in the kernel, Zone IDs
* (zoneid_t) are used to track zone association. Zone IDs are
* dynamically generated when the zone is created; if a persistent
* identifier is needed (core files, accounting logs, audit trail,
* etc.), the zone name should be used.
*
*
* Global Zone:
*
* The global zone (zoneid 0) is automatically associated with all
* system resources that have not been bound to a user-created zone.
* This means that even systems where zones are not in active use
* have a global zone, and all processes, mounts, etc. are
* associated with that zone. The global zone is generally
* unconstrained in terms of privileges and access, though the usual
* credential and privilege based restrictions apply.
*
*
* Zone States:
*
* The states in which a zone may be in and the transitions are as
* follows:
*
* ZONE_IS_UNINITIALIZED: primordial state for a zone. The partially
* initialized zone is added to the list of active zones on the system but
* isn't accessible.
*
* ZONE_IS_READY: zsched (the kernel dummy process for a zone) is
* ready. The zone is made visible after the ZSD constructor callbacks are
* executed. A zone remains in this state until it transitions into
* the ZONE_IS_BOOTING state as a result of a call to zone_boot().
*
* ZONE_IS_BOOTING: in this shortlived-state, zsched attempts to start
* init. Should that fail, the zone proceeds to the ZONE_IS_SHUTTING_DOWN
* state.
*
* ZONE_IS_RUNNING: The zone is open for business: zsched has
* successfully started init. A zone remains in this state until
* zone_shutdown() is called.
*
* ZONE_IS_SHUTTING_DOWN: zone_shutdown() has been called, the system is
* killing all processes running in the zone. The zone remains
* in this state until there are no more user processes running in the zone.
* zone_create(), zone_enter(), and zone_destroy() on this zone will fail.
* Since zone_shutdown() is restartable, it may be called successfully
* multiple times for the same zone_t. Setting of the zone's state to
* ZONE_IS_SHUTTING_DOWN is synchronized with mounts, so VOP_MOUNT() may check
* the zone's status without worrying about it being a moving target.
*
* ZONE_IS_EMPTY: zone_shutdown() has been called, and there
* are no more user processes in the zone. The zone remains in this
* state until there are no more kernel threads associated with the
* zone. zone_create(), zone_enter(), and zone_destroy() on this zone will
* fail.
*
* ZONE_IS_DOWN: All kernel threads doing work on behalf of the zone
* have exited. zone_shutdown() returns. Henceforth it is not possible to
* join the zone or create kernel threads therein.
*
* ZONE_IS_DYING: zone_destroy() has been called on the zone; zone
* remains in this state until zsched exits. Calls to zone_find_by_*()
* return NULL from now on.
*
* ZONE_IS_DEAD: zsched has exited (zone_ntasks == 0). There are no
* processes or threads doing work on behalf of the zone. The zone is
* removed from the list of active zones. zone_destroy() returns, and
* the zone can be recreated.
*
* ZONE_IS_FREE (internal state): zone_ref goes to 0, ZSD destructor
* callbacks are executed, and all memory associated with the zone is
* freed.
*
* Threads can wait for the zone to enter a requested state by using
* zone_status_wait() or zone_status_timedwait() with the desired
* state passed in as an argument. Zone state transitions are
* uni-directional; it is not possible to move back to an earlier state.
*
*
* Zone-Specific Data:
*
* Subsystems needing to maintain zone-specific data can store that
* data using the ZSD mechanism. This provides a zone-specific data
* store, similar to thread-specific data (see pthread_getspecific(3C)
* or the TSD code in uts/common/disp/thread.c. Also, ZSD can be used
* to register callbacks to be invoked when a zone is created, shut
* down, or destroyed. This can be used to initialize zone-specific
* data for new zones and to clean up when zones go away.
*
*
* Data Structures:
*
* The per-zone structure (zone_t) is reference counted, and freed
* when all references are released. zone_hold and zone_rele can be
* used to adjust the reference count. In addition, reference counts
* associated with the cred_t structure are tracked separately using
* zone_cred_hold and zone_cred_rele.
*
* Pointers to active zone_t's are stored in two hash tables; one
* for searching by id, the other for searching by name. Lookups
* can be performed on either basis, using zone_find_by_id and
* zone_find_by_name. Both return zone_t pointers with the zone
* held, so zone_rele should be called when the pointer is no longer
* needed. Zones can also be searched by path; zone_find_by_path
* returns the zone with which a path name is associated (global
* zone if the path is not within some other zone's file system
* hierarchy). This currently requires iterating through each zone,
* so it is slower than an id or name search via a hash table.
*
*
* Locking:
*
* zonehash_lock: This is a top-level global lock used to protect the
* zone hash tables and lists. Zones cannot be created or destroyed
* while this lock is held.
* zone_status_lock: This is a global lock protecting zone state.
* Zones cannot change state while this lock is held. It also
* protects the list of kernel threads associated with a zone.
* zone_lock: This is a per-zone lock used to protect several fields of
* the zone_t (see <sys/zone.h> for details). In addition, holding
* this lock means that the zone cannot go away.
* zsd_key_lock: This is a global lock protecting the key state for ZSD.
* zone_deathrow_lock: This is a global lock protecting the "deathrow"
* list (a list of zones in the ZONE_IS_DEAD state).
*
* Ordering requirements:
* pool_lock --> cpu_lock --> zonehash_lock --> zone_status_lock -->
* zone_lock --> zsd_key_lock --> pidlock --> p_lock
*
* Blocking memory allocations are permitted while holding any of the
* zone locks.
*
*
* System Call Interface:
*
* The zone subsystem can be managed and queried from user level with
* the following system calls (all subcodes of the primary "zone"
* system call):
* - zone_create: creates a zone with selected attributes (name,
* root path, privileges, resource controls, ZFS datasets)
* - zone_enter: allows the current process to enter a zone
* - zone_getattr: reports attributes of a zone
* - zone_list: lists all zones active in the system
* - zone_lookup: looks up zone id based on name
* - zone_shutdown: initiates shutdown process (see states above)
* - zone_destroy: completes shutdown process (see states above)
*
*/
#include <sys/priv_impl.h>
#include <sys/cred.h>
#include <c2/audit.h>
#include <sys/ddi.h>
#include <sys/debug.h>
#include <sys/file.h>
#include <sys/kmem.h>
#include <sys/mutex.h>
#include <sys/pathname.h>
#include <sys/proc.h>
#include <sys/project.h>
#include <sys/sysevent.h>
#include <sys/task.h>
#include <sys/systm.h>
#include <sys/types.h>
#include <sys/utsname.h>
#include <sys/vnode.h>
#include <sys/vfs.h>
#include <sys/systeminfo.h>
#include <sys/policy.h>
#include <sys/cred_impl.h>
#include <sys/contract_impl.h>
#include <sys/contract/process_impl.h>
#include <sys/class.h>
#include <sys/pool.h>
#include <sys/pool_pset.h>
#include <sys/pset.h>
#include <sys/log.h>
#include <sys/sysmacros.h>
#include <sys/callb.h>
#include <sys/vmparam.h>
#include <sys/corectl.h>
#include <sys/door.h>
#include <sys/cpuvar.h>
#include <sys/fs/snode.h>
#include <sys/uadmin.h>
#include <sys/session.h>
#include <sys/cmn_err.h>
#include <sys/modhash.h>
#include <sys/nvpair.h>
#include <sys/rctl.h>
#include <sys/fss.h>
#include <sys/zone.h>
/*
* cv used to signal that all references to the zone have been released. This
* needs to be global since there may be multiple waiters, and the first to
* wake up will free the zone_t, hence we cannot use zone->zone_cv.
*/
static kcondvar_t zone_destroy_cv;
/*
* Lock used to serialize access to zone_cv. This could have been per-zone,
* but then we'd need another lock for zone_destroy_cv, and why bother?
*/
static kmutex_t zone_status_lock;
/*
* ZSD-related global variables.
*/
static kmutex_t zsd_key_lock; /* protects the following two */
/*
* The next caller of zone_key_create() will be assigned a key of ++zsd_keyval.
*/
static zone_key_t zsd_keyval = 0;
/*
* Global list of registered keys. We use this when a new zone is created.
*/
static list_t zsd_registered_keys;
int zone_hash_size = 256;
static mod_hash_t *zonehashbyname, *zonehashbyid;
static kmutex_t zonehash_lock;
static uint_t zonecount;
static id_space_t *zoneid_space;
/*
* The global zone (aka zone0) is the all-seeing, all-knowing zone in which the
* kernel proper runs, and which manages all other zones.
*
* Although not declared as static, the variable "zone0" should not be used
* except for by code that needs to reference the global zone early on in boot,
* before it is fully initialized. All other consumers should use
* 'global_zone'.
*/
zone_t zone0;
zone_t *global_zone = NULL; /* Set when the global zone is initialized */
/*
* List of active zones, protected by zonehash_lock.
*/
static list_t zone_active;
/*
* List of destroyed zones that still have outstanding cred references.
* Used for debugging. Uses a separate lock to avoid lock ordering
* problems in zone_free.
*/
static list_t zone_deathrow;
static kmutex_t zone_deathrow_lock;
/* number of zones is limited by virtual interface limit in IP */
uint_t maxzones = 8192;
/* Event channel to sent zone state change notifications */
evchan_t *zone_event_chan;
/*
* This table holds the mapping from kernel zone states to
* states visible in the state notification API.
* The idea is that we only expose "obvious" states and
* do not expose states which are just implementation details.
*/
const char *zone_status_table[] = {
ZONE_EVENT_UNINITIALIZED, /* uninitialized */
ZONE_EVENT_READY, /* ready */
ZONE_EVENT_READY, /* booting */
ZONE_EVENT_RUNNING, /* running */
ZONE_EVENT_SHUTTING_DOWN, /* shutting_down */
ZONE_EVENT_SHUTTING_DOWN, /* empty */
ZONE_EVENT_SHUTTING_DOWN, /* down */
ZONE_EVENT_SHUTTING_DOWN, /* dying */
ZONE_EVENT_UNINITIALIZED, /* dead */
};
/*
* This isn't static so lint doesn't complain.
*/
rctl_hndl_t rc_zone_cpu_shares;
rctl_hndl_t rc_zone_nlwps;
/*
* Synchronization primitives used to synchronize between mounts and zone
* creation/destruction.
*/
static int mounts_in_progress;
static kcondvar_t mount_cv;
static kmutex_t mount_lock;
const char * const zone_initname = "/sbin/init";
static int zone_shutdown(zoneid_t zoneid);
/*
* Bump this number when you alter the zone syscall interfaces; this is
* because we need to have support for previous API versions in libc
* to support patching; libc calls into the kernel to determine this number.
*
* Version 1 of the API is the version originally shipped with Solaris 10
* Version 2 alters the zone_create system call in order to support more
* arguments by moving the args into a structure; and to do better
* error reporting when zone_create() fails.
* Version 3 alters the zone_create system call in order to support the
* import of ZFS datasets to zones.
*/
static const int ZONE_SYSCALL_API_VERSION = 3;
/*
* Certain filesystems (such as NFS and autofs) need to know which zone
* the mount is being placed in. Because of this, we need to be able to
* ensure that a zone isn't in the process of being created such that
* nfs_mount() thinks it is in the global zone, while by the time it
* gets added the list of mounted zones, it ends up on zoneA's mount
* list.
*
* The following functions: block_mounts()/resume_mounts() and
* mount_in_progress()/mount_completed() are used by zones and the VFS
* layer (respectively) to synchronize zone creation and new mounts.
*
* The semantics are like a reader-reader lock such that there may
* either be multiple mounts (or zone creations, if that weren't
* serialized by zonehash_lock) in progress at the same time, but not
* both.
*
* We use cv's so the user can ctrl-C out of the operation if it's
* taking too long.
*
* The semantics are such that there is unfair bias towards the
* "current" operation. This means that zone creations may starve if
* there is a rapid succession of new mounts coming in to the system, or
* there is a remote possibility that zones will be created at such a
* rate that new mounts will not be able to proceed.
*/
/*
* Prevent new mounts from progressing to the point of calling
* VFS_MOUNT(). If there are already mounts in this "region", wait for
* them to complete.
*/
static int
block_mounts(void)
{
int retval = 0;
/*
* Since it may block for a long time, block_mounts() shouldn't be
* called with zonehash_lock held.
*/
ASSERT(MUTEX_NOT_HELD(&zonehash_lock));
mutex_enter(&mount_lock);
while (mounts_in_progress > 0) {
if (cv_wait_sig(&mount_cv, &mount_lock) == 0)
goto signaled;
}
/*
* A negative value of mounts_in_progress indicates that mounts
* have been blocked by (-mounts_in_progress) different callers.
*/
mounts_in_progress--;
retval = 1;
signaled:
mutex_exit(&mount_lock);
return (retval);
}
/*
* The VFS layer may progress with new mounts as far as we're concerned.
* Allow them to progress if we were the last obstacle.
*/
static void
resume_mounts(void)
{
mutex_enter(&mount_lock);
if (++mounts_in_progress == 0)
cv_broadcast(&mount_cv);
mutex_exit(&mount_lock);
}
/*
* The VFS layer is busy with a mount; zones should wait until all
* mounts are completed to progress.
*/
void
mount_in_progress(void)
{
mutex_enter(&mount_lock);
while (mounts_in_progress < 0)
cv_wait(&mount_cv, &mount_lock);
mounts_in_progress++;
mutex_exit(&mount_lock);
}
/*
* VFS is done with one mount; wake up any waiting block_mounts()
* callers if this is the last mount.
*/
void
mount_completed(void)
{
mutex_enter(&mount_lock);
if (--mounts_in_progress == 0)
cv_broadcast(&mount_cv);
mutex_exit(&mount_lock);
}
/*
* ZSD routines.
*
* Zone Specific Data (ZSD) is modeled after Thread Specific Data as
* defined by the pthread_key_create() and related interfaces.
*
* Kernel subsystems may register one or more data items and/or
* callbacks to be executed when a zone is created, shutdown, or
* destroyed.
*
* Unlike the thread counterpart, destructor callbacks will be executed
* even if the data pointer is NULL and/or there are no constructor
* callbacks, so it is the responsibility of such callbacks to check for
* NULL data values if necessary.
*
* The locking strategy and overall picture is as follows:
*
* When someone calls zone_key_create(), a template ZSD entry is added to the
* global list "zsd_registered_keys", protected by zsd_key_lock. The
* constructor callback is called immediately on all existing zones, and a
* copy of the ZSD entry added to the per-zone zone_zsd list (protected by
* zone_lock). As this operation requires the list of zones, the list of
* registered keys, and the per-zone list of ZSD entries to remain constant
* throughout the entire operation, it must grab zonehash_lock, zone_lock for
* all existing zones, and zsd_key_lock, in that order. Similar locking is
* needed when zone_key_delete() is called. It is thus sufficient to hold
* zsd_key_lock *or* zone_lock to prevent additions to or removals from the
* per-zone zone_zsd list.
*
* Note that this implementation does not make a copy of the ZSD entry if a
* constructor callback is not provided. A zone_getspecific() on such an
* uninitialized ZSD entry will return NULL.
*
* When new zones are created constructor callbacks for all registered ZSD
* entries will be called.
*
* The framework does not provide any locking around zone_getspecific() and
* zone_setspecific() apart from that needed for internal consistency, so
* callers interested in atomic "test-and-set" semantics will need to provide
* their own locking.
*/
void
zone_key_create(zone_key_t *keyp, void *(*create)(zoneid_t),
void (*shutdown)(zoneid_t, void *), void (*destroy)(zoneid_t, void *))
{
struct zsd_entry *zsdp;
struct zsd_entry *t;
struct zone *zone;
zsdp = kmem_alloc(sizeof (*zsdp), KM_SLEEP);
zsdp->zsd_data = NULL;
zsdp->zsd_create = create;
zsdp->zsd_shutdown = shutdown;
zsdp->zsd_destroy = destroy;
mutex_enter(&zonehash_lock); /* stop the world */
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone))
mutex_enter(&zone->zone_lock); /* lock all zones */
mutex_enter(&zsd_key_lock);
*keyp = zsdp->zsd_key = ++zsd_keyval;
ASSERT(zsd_keyval != 0);
list_insert_tail(&zsd_registered_keys, zsdp);
mutex_exit(&zsd_key_lock);
if (create != NULL) {
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone)) {
t = kmem_alloc(sizeof (*t), KM_SLEEP);
t->zsd_key = *keyp;
t->zsd_data = (*create)(zone->zone_id);
t->zsd_create = create;
t->zsd_shutdown = shutdown;
t->zsd_destroy = destroy;
list_insert_tail(&zone->zone_zsd, t);
}
}
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone))
mutex_exit(&zone->zone_lock);
mutex_exit(&zonehash_lock);
}
/*
* Helper function to find the zsd_entry associated with the key in the
* given list.
*/
static struct zsd_entry *
zsd_find(list_t *l, zone_key_t key)
{
struct zsd_entry *zsd;
for (zsd = list_head(l); zsd != NULL; zsd = list_next(l, zsd)) {
if (zsd->zsd_key == key) {
/*
* Move to head of list to keep list in MRU order.
*/
if (zsd != list_head(l)) {
list_remove(l, zsd);
list_insert_head(l, zsd);
}
return (zsd);
}
}
return (NULL);
}
/*
* Function called when a module is being unloaded, or otherwise wishes
* to unregister its ZSD key and callbacks.
*/
int
zone_key_delete(zone_key_t key)
{
struct zsd_entry *zsdp = NULL;
zone_t *zone;
mutex_enter(&zonehash_lock); /* Zone create/delete waits for us */
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone))
mutex_enter(&zone->zone_lock); /* lock all zones */
mutex_enter(&zsd_key_lock);
zsdp = zsd_find(&zsd_registered_keys, key);
if (zsdp == NULL)
goto notfound;
list_remove(&zsd_registered_keys, zsdp);
mutex_exit(&zsd_key_lock);
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone)) {
struct zsd_entry *del;
void *data;
if (!(zone->zone_flags & ZF_DESTROYED)) {
del = zsd_find(&zone->zone_zsd, key);
if (del != NULL) {
data = del->zsd_data;
ASSERT(del->zsd_shutdown == zsdp->zsd_shutdown);
ASSERT(del->zsd_destroy == zsdp->zsd_destroy);
list_remove(&zone->zone_zsd, del);
kmem_free(del, sizeof (*del));
} else {
data = NULL;
}
if (zsdp->zsd_shutdown)
zsdp->zsd_shutdown(zone->zone_id, data);
if (zsdp->zsd_destroy)
zsdp->zsd_destroy(zone->zone_id, data);
}
mutex_exit(&zone->zone_lock);
}
mutex_exit(&zonehash_lock);
kmem_free(zsdp, sizeof (*zsdp));
return (0);
notfound:
mutex_exit(&zsd_key_lock);
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone))
mutex_exit(&zone->zone_lock);
mutex_exit(&zonehash_lock);
return (-1);
}
/*
* ZSD counterpart of pthread_setspecific().
*/
int
zone_setspecific(zone_key_t key, zone_t *zone, const void *data)
{
struct zsd_entry *t;
struct zsd_entry *zsdp = NULL;
mutex_enter(&zone->zone_lock);
t = zsd_find(&zone->zone_zsd, key);
if (t != NULL) {
/*
* Replace old value with new
*/
t->zsd_data = (void *)data;
mutex_exit(&zone->zone_lock);
return (0);
}
/*
* If there was no previous value, go through the list of registered
* keys.
*
* We avoid grabbing zsd_key_lock until we are sure we need it; this is
* necessary for shutdown callbacks to be able to execute without fear
* of deadlock.
*/
mutex_enter(&zsd_key_lock);
zsdp = zsd_find(&zsd_registered_keys, key);
if (zsdp == NULL) { /* Key was not registered */
mutex_exit(&zsd_key_lock);
mutex_exit(&zone->zone_lock);
return (-1);
}
/*
* Add a zsd_entry to this zone, using the template we just retrieved
* to initialize the constructor and destructor(s).
*/
t = kmem_alloc(sizeof (*t), KM_SLEEP);
t->zsd_key = key;
t->zsd_data = (void *)data;
t->zsd_create = zsdp->zsd_create;
t->zsd_shutdown = zsdp->zsd_shutdown;
t->zsd_destroy = zsdp->zsd_destroy;
list_insert_tail(&zone->zone_zsd, t);
mutex_exit(&zsd_key_lock);
mutex_exit(&zone->zone_lock);
return (0);
}
/*
* ZSD counterpart of pthread_getspecific().
*/
void *
zone_getspecific(zone_key_t key, zone_t *zone)
{
struct zsd_entry *t;
void *data;
mutex_enter(&zone->zone_lock);
t = zsd_find(&zone->zone_zsd, key);
data = (t == NULL ? NULL : t->zsd_data);
mutex_exit(&zone->zone_lock);
return (data);
}
/*
* Function used to initialize a zone's list of ZSD callbacks and data
* when the zone is being created. The callbacks are initialized from
* the template list (zsd_registered_keys), and the constructor
* callback executed (if one exists).
*
* This is called before the zone is made publicly available, hence no
* need to grab zone_lock.
*
* Although we grab and release zsd_key_lock, new entries cannot be
* added to or removed from the zsd_registered_keys list until we
* release zonehash_lock, so there isn't a window for a
* zone_key_create() to come in after we've dropped zsd_key_lock but
* before the zone is added to the zone list, such that the constructor
* callbacks aren't executed for the new zone.
*/
static void
zone_zsd_configure(zone_t *zone)
{
struct zsd_entry *zsdp;
struct zsd_entry *t;
zoneid_t zoneid = zone->zone_id;
ASSERT(MUTEX_HELD(&zonehash_lock));
ASSERT(list_head(&zone->zone_zsd) == NULL);
mutex_enter(&zsd_key_lock);
for (zsdp = list_head(&zsd_registered_keys); zsdp != NULL;
zsdp = list_next(&zsd_registered_keys, zsdp)) {
if (zsdp->zsd_create != NULL) {
t = kmem_alloc(sizeof (*t), KM_SLEEP);
t->zsd_key = zsdp->zsd_key;
t->zsd_create = zsdp->zsd_create;
t->zsd_data = (*t->zsd_create)(zoneid);
t->zsd_shutdown = zsdp->zsd_shutdown;
t->zsd_destroy = zsdp->zsd_destroy;
list_insert_tail(&zone->zone_zsd, t);
}
}
mutex_exit(&zsd_key_lock);
}
enum zsd_callback_type { ZSD_CREATE, ZSD_SHUTDOWN, ZSD_DESTROY };
/*
* Helper function to execute shutdown or destructor callbacks.
*/
static void
zone_zsd_callbacks(zone_t *zone, enum zsd_callback_type ct)
{
struct zsd_entry *zsdp;
struct zsd_entry *t;
zoneid_t zoneid = zone->zone_id;
ASSERT(ct == ZSD_SHUTDOWN || ct == ZSD_DESTROY);
ASSERT(ct != ZSD_SHUTDOWN || zone_status_get(zone) >= ZONE_IS_EMPTY);
ASSERT(ct != ZSD_DESTROY || zone_status_get(zone) >= ZONE_IS_DOWN);
mutex_enter(&zone->zone_lock);
if (ct == ZSD_DESTROY) {
if (zone->zone_flags & ZF_DESTROYED) {
/*
* Make sure destructors are only called once.
*/
mutex_exit(&zone->zone_lock);
return;
}
zone->zone_flags |= ZF_DESTROYED;
}
mutex_exit(&zone->zone_lock);
/*
* Both zsd_key_lock and zone_lock need to be held in order to add or
* remove a ZSD key, (either globally as part of
* zone_key_create()/zone_key_delete(), or on a per-zone basis, as is
* possible through zone_setspecific()), so it's sufficient to hold
* zsd_key_lock here.
*
* This is a good thing, since we don't want to recursively try to grab
* zone_lock if a callback attempts to do something like a crfree() or
* zone_rele().
*/
mutex_enter(&zsd_key_lock);
for (zsdp = list_head(&zsd_registered_keys); zsdp != NULL;
zsdp = list_next(&zsd_registered_keys, zsdp)) {
zone_key_t key = zsdp->zsd_key;
/* Skip if no callbacks registered */
if (ct == ZSD_SHUTDOWN && zsdp->zsd_shutdown == NULL)
continue;
if (ct == ZSD_DESTROY && zsdp->zsd_destroy == NULL)
continue;
/*
* Call the callback with the zone-specific data if we can find
* any, otherwise with NULL.
*/
t = zsd_find(&zone->zone_zsd, key);
if (t != NULL) {
if (ct == ZSD_SHUTDOWN) {
t->zsd_shutdown(zoneid, t->zsd_data);
} else {
ASSERT(ct == ZSD_DESTROY);
t->zsd_destroy(zoneid, t->zsd_data);
}
} else {
if (ct == ZSD_SHUTDOWN) {
zsdp->zsd_shutdown(zoneid, NULL);
} else {
ASSERT(ct == ZSD_DESTROY);
zsdp->zsd_destroy(zoneid, NULL);
}
}
}
mutex_exit(&zsd_key_lock);
}
/*
* Called when the zone is going away; free ZSD-related memory, and
* destroy the zone_zsd list.
*/
static void
zone_free_zsd(zone_t *zone)
{
struct zsd_entry *t, *next;
/*
* Free all the zsd_entry's we had on this zone.
*/
for (t = list_head(&zone->zone_zsd); t != NULL; t = next) {
next = list_next(&zone->zone_zsd, t);
list_remove(&zone->zone_zsd, t);
kmem_free(t, sizeof (*t));
}
list_destroy(&zone->zone_zsd);
}
/*
* Frees memory associated with the zone dataset list.
*/
static void
zone_free_datasets(zone_t *zone)
{
zone_dataset_t *t, *next;
for (t = list_head(&zone->zone_datasets); t != NULL; t = next) {
next = list_next(&zone->zone_datasets, t);
list_remove(&zone->zone_datasets, t);
kmem_free(t->zd_dataset, strlen(t->zd_dataset) + 1);
kmem_free(t, sizeof (*t));
}
list_destroy(&zone->zone_datasets);
}
/*
* zone.cpu-shares resource control support.
*/
/*ARGSUSED*/
static rctl_qty_t
zone_cpu_shares_usage(rctl_t *rctl, struct proc *p)
{
ASSERT(MUTEX_HELD(&p->p_lock));
return (p->p_zone->zone_shares);
}
/*ARGSUSED*/
static int
zone_cpu_shares_set(rctl_t *rctl, struct proc *p, rctl_entity_p_t *e,
rctl_qty_t nv)
{
ASSERT(MUTEX_HELD(&p->p_lock));
ASSERT(e->rcep_t == RCENTITY_ZONE);
if (e->rcep_p.zone == NULL)
return (0);
e->rcep_p.zone->zone_shares = nv;
return (0);
}
static rctl_ops_t zone_cpu_shares_ops = {
rcop_no_action,
zone_cpu_shares_usage,
zone_cpu_shares_set,
rcop_no_test
};
/*ARGSUSED*/
static rctl_qty_t
zone_lwps_usage(rctl_t *r, proc_t *p)
{
rctl_qty_t nlwps;
zone_t *zone = p->p_zone;
ASSERT(MUTEX_HELD(&p->p_lock));
mutex_enter(&zone->zone_nlwps_lock);
nlwps = zone->zone_nlwps;
mutex_exit(&zone->zone_nlwps_lock);
return (nlwps);
}
/*ARGSUSED*/
static int
zone_lwps_test(rctl_t *r, proc_t *p, rctl_entity_p_t *e, rctl_val_t *rcntl,
rctl_qty_t incr, uint_t flags)
{
rctl_qty_t nlwps;
ASSERT(MUTEX_HELD(&p->p_lock));
ASSERT(e->rcep_t == RCENTITY_ZONE);
if (e->rcep_p.zone == NULL)
return (0);
ASSERT(MUTEX_HELD(&(e->rcep_p.zone->zone_nlwps_lock)));
nlwps = e->rcep_p.zone->zone_nlwps;
if (nlwps + incr > rcntl->rcv_value)
return (1);
return (0);
}
/*ARGSUSED*/
static int
zone_lwps_set(rctl_t *rctl, struct proc *p, rctl_entity_p_t *e, rctl_qty_t nv) {
ASSERT(MUTEX_HELD(&p->p_lock));
ASSERT(e->rcep_t == RCENTITY_ZONE);
if (e->rcep_p.zone == NULL)
return (0);
e->rcep_p.zone->zone_nlwps_ctl = nv;
return (0);
}
static rctl_ops_t zone_lwps_ops = {
rcop_no_action,
zone_lwps_usage,
zone_lwps_set,
zone_lwps_test,
};
/*
* Helper function to brand the zone with a unique ID.
*/
static void
zone_uniqid(zone_t *zone)
{
static uint64_t uniqid = 0;
ASSERT(MUTEX_HELD(&zonehash_lock));
zone->zone_uniqid = uniqid++;
}
/*
* Returns a held pointer to the "kcred" for the specified zone.
*/
struct cred *
zone_get_kcred(zoneid_t zoneid)
{
zone_t *zone;
cred_t *cr;
if ((zone = zone_find_by_id(zoneid)) == NULL)
return (NULL);
cr = zone->zone_kcred;
crhold(cr);
zone_rele(zone);
return (cr);
}
/*
* Called very early on in boot to initialize the ZSD list so that
* zone_key_create() can be called before zone_init(). It also initializes
* portions of zone0 which may be used before zone_init() is called. The
* variable "global_zone" will be set when zone0 is fully initialized by
* zone_init().
*/
void
zone_zsd_init(void)
{
mutex_init(&zonehash_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&zsd_key_lock, NULL, MUTEX_DEFAULT, NULL);
list_create(&zsd_registered_keys, sizeof (struct zsd_entry),
offsetof(struct zsd_entry, zsd_linkage));
list_create(&zone_active, sizeof (zone_t),
offsetof(zone_t, zone_linkage));
list_create(&zone_deathrow, sizeof (zone_t),
offsetof(zone_t, zone_linkage));
mutex_init(&zone0.zone_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&zone0.zone_nlwps_lock, NULL, MUTEX_DEFAULT, NULL);
zone0.zone_shares = 1;
zone0.zone_nlwps_ctl = INT_MAX;
zone0.zone_name = GLOBAL_ZONENAME;
zone0.zone_nodename = utsname.nodename;
zone0.zone_domain = srpc_domain;
zone0.zone_ref = 1;
zone0.zone_id = GLOBAL_ZONEID;
zone0.zone_status = ZONE_IS_RUNNING;
zone0.zone_rootpath = "/";
zone0.zone_rootpathlen = 2;
zone0.zone_psetid = ZONE_PS_INVAL;
zone0.zone_ncpus = 0;
zone0.zone_ncpus_online = 0;
zone0.zone_proc_initpid = 1;
list_create(&zone0.zone_zsd, sizeof (struct zsd_entry),
offsetof(struct zsd_entry, zsd_linkage));
list_insert_head(&zone_active, &zone0);
/*
* The root filesystem is not mounted yet, so zone_rootvp cannot be set
* to anything meaningful. It is assigned to be 'rootdir' in
* vfs_mountroot().
*/
zone0.zone_rootvp = NULL;
zone0.zone_vfslist = NULL;
zone0.zone_bootargs = NULL;
zone0.zone_privset = kmem_alloc(sizeof (priv_set_t), KM_SLEEP);
/*
* The global zone has all privileges
*/
priv_fillset(zone0.zone_privset);
/*
* Add p0 to the global zone
*/
zone0.zone_zsched = &p0;
p0.p_zone = &zone0;
}
/*
* Called by main() to initialize the zones framework.
*/
void
zone_init(void)
{
rctl_dict_entry_t *rde;
rctl_val_t *dval;
rctl_set_t *set;
rctl_alloc_gp_t *gp;
rctl_entity_p_t e;
int res;
ASSERT(curproc == &p0);
/*
* Create ID space for zone IDs. ID 0 is reserved for the
* global zone.
*/
zoneid_space = id_space_create("zoneid_space", 1, MAX_ZONEID);
/*
* Initialize generic zone resource controls, if any.
*/
rc_zone_cpu_shares = rctl_register("zone.cpu-shares",
RCENTITY_ZONE, RCTL_GLOBAL_SIGNAL_NEVER | RCTL_GLOBAL_DENY_NEVER |
RCTL_GLOBAL_NOBASIC |
RCTL_GLOBAL_COUNT, FSS_MAXSHARES, FSS_MAXSHARES,
&zone_cpu_shares_ops);
rc_zone_nlwps = rctl_register("zone.max-lwps", RCENTITY_ZONE,
RCTL_GLOBAL_NOACTION | RCTL_GLOBAL_NOBASIC | RCTL_GLOBAL_COUNT,
INT_MAX, INT_MAX, &zone_lwps_ops);
/*
* Create a rctl_val with PRIVILEGED, NOACTION, value = 1. Then attach
* this at the head of the rctl_dict_entry for ``zone.cpu-shares''.
*/
dval = kmem_cache_alloc(rctl_val_cache, KM_SLEEP);
bzero(dval, sizeof (rctl_val_t));
dval->rcv_value = 1;
dval->rcv_privilege = RCPRIV_PRIVILEGED;
dval->rcv_flagaction = RCTL_LOCAL_NOACTION;
dval->rcv_action_recip_pid = -1;
rde = rctl_dict_lookup("zone.cpu-shares");
(void) rctl_val_list_insert(&rde->rcd_default_value, dval);
/*
* Initialize the ``global zone''.
*/
set = rctl_set_create();
gp = rctl_set_init_prealloc(RCENTITY_ZONE);
mutex_enter(&p0.p_lock);
e.rcep_p.zone = &zone0;
e.rcep_t = RCENTITY_ZONE;
zone0.zone_rctls = rctl_set_init(RCENTITY_ZONE, &p0, &e, set,
gp);
zone0.zone_nlwps = p0.p_lwpcnt;
zone0.zone_ntasks = 1;
mutex_exit(&p0.p_lock);
rctl_prealloc_destroy(gp);
/*
* pool_default hasn't been initialized yet, so we let pool_init() take
* care of making the global zone is in the default pool.
*/
mutex_enter(&zonehash_lock);
zone_uniqid(&zone0);
ASSERT(zone0.zone_uniqid == GLOBAL_ZONEUNIQID);
mutex_exit(&zonehash_lock);
zonehashbyid = mod_hash_create_idhash("zone_by_id", zone_hash_size,
mod_hash_null_valdtor);
zonehashbyname = mod_hash_create_strhash("zone_by_name",
zone_hash_size, mod_hash_null_valdtor);
zonecount = 1;
(void) mod_hash_insert(zonehashbyid, (mod_hash_key_t)GLOBAL_ZONEID,
(mod_hash_val_t)&zone0);
(void) mod_hash_insert(zonehashbyname, (mod_hash_key_t)zone0.zone_name,
(mod_hash_val_t)&zone0);
/*
* We avoid setting zone_kcred until now, since kcred is initialized
* sometime after zone_zsd_init() and before zone_init().
*/
zone0.zone_kcred = kcred;
/*
* The global zone is fully initialized (except for zone_rootvp which
* will be set when the root filesystem is mounted).
*/
global_zone = &zone0;
/*
* Setup an event channel to send zone status change notifications on
*/
res = sysevent_evc_bind(ZONE_EVENT_CHANNEL, &zone_event_chan,
EVCH_CREAT);
if (res)
panic("Sysevent_evc_bind failed during zone setup.\n");
}
static void
zone_free(zone_t *zone)
{
ASSERT(zone != global_zone);
ASSERT(zone->zone_ntasks == 0);
ASSERT(zone->zone_nlwps == 0);
ASSERT(zone->zone_cred_ref == 0);
ASSERT(zone->zone_kcred == NULL);
ASSERT(zone_status_get(zone) == ZONE_IS_DEAD ||
zone_status_get(zone) == ZONE_IS_UNINITIALIZED);
/* remove from deathrow list */
if (zone_status_get(zone) == ZONE_IS_DEAD) {
ASSERT(zone->zone_ref == 0);
mutex_enter(&zone_deathrow_lock);
list_remove(&zone_deathrow, zone);
mutex_exit(&zone_deathrow_lock);
}
zone_free_zsd(zone);
zone_free_datasets(zone);
if (zone->zone_rootvp != NULL)
VN_RELE(zone->zone_rootvp);
if (zone->zone_rootpath)
kmem_free(zone->zone_rootpath, zone->zone_rootpathlen);
if (zone->zone_name != NULL)
kmem_free(zone->zone_name, ZONENAME_MAX);
if (zone->zone_nodename != NULL)
kmem_free(zone->zone_nodename, _SYS_NMLN);
if (zone->zone_domain != NULL)
kmem_free(zone->zone_domain, _SYS_NMLN);
if (zone->zone_privset != NULL)
kmem_free(zone->zone_privset, sizeof (priv_set_t));
if (zone->zone_rctls != NULL)
rctl_set_free(zone->zone_rctls);
if (zone->zone_bootargs != NULL)
kmem_free(zone->zone_bootargs, ZONEBOOTARGS_MAX);
id_free(zoneid_space, zone->zone_id);
mutex_destroy(&zone->zone_lock);
cv_destroy(&zone->zone_cv);
kmem_free(zone, sizeof (zone_t));
}
/*
* See block comment at the top of this file for information about zone
* status values.
*/
/*
* Convenience function for setting zone status.
*/
static void
zone_status_set(zone_t *zone, zone_status_t status)
{
nvlist_t *nvl = NULL;
ASSERT(MUTEX_HELD(&zone_status_lock));
ASSERT(status > ZONE_MIN_STATE && status <= ZONE_MAX_STATE &&
status >= zone_status_get(zone));
if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, KM_SLEEP) ||
nvlist_add_string(nvl, ZONE_CB_NAME, zone->zone_name) ||
nvlist_add_string(nvl, ZONE_CB_NEWSTATE,
zone_status_table[status]) ||
nvlist_add_string(nvl, ZONE_CB_OLDSTATE,
zone_status_table[zone->zone_status]) ||
nvlist_add_int32(nvl, ZONE_CB_ZONEID, zone->zone_id) ||
nvlist_add_uint64(nvl, ZONE_CB_TIMESTAMP, (uint64_t)gethrtime()) ||
sysevent_evc_publish(zone_event_chan, ZONE_EVENT_STATUS_CLASS,
ZONE_EVENT_STATUS_SUBCLASS,
"sun.com", "kernel", nvl, EVCH_SLEEP)) {
#ifdef DEBUG
(void) printf(
"Failed to allocate and send zone state change event.\n");
#endif
}
nvlist_free(nvl);
zone->zone_status = status;
cv_broadcast(&zone->zone_cv);
}
/*
* Public function to retrieve the zone status. The zone status may
* change after it is retrieved.
*/
zone_status_t
zone_status_get(zone_t *zone)
{
return (zone->zone_status);
}
static int
zone_set_bootargs(zone_t *zone, const char *zone_bootargs)
{
char *bootargs = kmem_zalloc(ZONEBOOTARGS_MAX, KM_SLEEP);
size_t len;
int err;
err = copyinstr(zone_bootargs, bootargs, ZONEBOOTARGS_MAX - 1, &len);
if (err != 0) {
kmem_free(bootargs, ZONEBOOTARGS_MAX);
return (err); /* EFAULT or ENAMETOOLONG */
}
bootargs[len] = '\0';
ASSERT(zone->zone_bootargs == NULL);
zone->zone_bootargs = bootargs;
return (0);
}
/*
* Block indefinitely waiting for (zone_status >= status)
*/
void
zone_status_wait(zone_t *zone, zone_status_t status)
{
ASSERT(status > ZONE_MIN_STATE && status <= ZONE_MAX_STATE);
mutex_enter(&zone_status_lock);
while (zone->zone_status < status) {
cv_wait(&zone->zone_cv, &zone_status_lock);
}
mutex_exit(&zone_status_lock);
}
/*
* Private CPR-safe version of zone_status_wait().
*/
static void
zone_status_wait_cpr(zone_t *zone, zone_status_t status, char *str)
{
callb_cpr_t cprinfo;
ASSERT(status > ZONE_MIN_STATE && status <= ZONE_MAX_STATE);
CALLB_CPR_INIT(&cprinfo, &zone_status_lock, callb_generic_cpr,
str);
mutex_enter(&zone_status_lock);
while (zone->zone_status < status) {
CALLB_CPR_SAFE_BEGIN(&cprinfo);
cv_wait(&zone->zone_cv, &zone_status_lock);
CALLB_CPR_SAFE_END(&cprinfo, &zone_status_lock);
}
/*
* zone_status_lock is implicitly released by the following.
*/
CALLB_CPR_EXIT(&cprinfo);
}
/*
* Block until zone enters requested state or signal is received. Return (0)
* if signaled, non-zero otherwise.
*/
int
zone_status_wait_sig(zone_t *zone, zone_status_t status)
{
ASSERT(status > ZONE_MIN_STATE && status <= ZONE_MAX_STATE);
mutex_enter(&zone_status_lock);
while (zone->zone_status < status) {
if (!cv_wait_sig(&zone->zone_cv, &zone_status_lock)) {
mutex_exit(&zone_status_lock);
return (0);
}
}
mutex_exit(&zone_status_lock);
return (1);
}
/*
* Block until the zone enters the requested state or the timeout expires,
* whichever happens first. Return (-1) if operation timed out, time remaining
* otherwise.
*/
clock_t
zone_status_timedwait(zone_t *zone, clock_t tim, zone_status_t status)
{
clock_t timeleft = 0;
ASSERT(status > ZONE_MIN_STATE && status <= ZONE_MAX_STATE);
mutex_enter(&zone_status_lock);
while (zone->zone_status < status && timeleft != -1) {
timeleft = cv_timedwait(&zone->zone_cv, &zone_status_lock, tim);
}
mutex_exit(&zone_status_lock);
return (timeleft);
}
/*
* Block until the zone enters the requested state, the current process is
* signaled, or the timeout expires, whichever happens first. Return (-1) if
* operation timed out, 0 if signaled, time remaining otherwise.
*/
clock_t
zone_status_timedwait_sig(zone_t *zone, clock_t tim, zone_status_t status)
{
clock_t timeleft = tim - lbolt;
ASSERT(status > ZONE_MIN_STATE && status <= ZONE_MAX_STATE);
mutex_enter(&zone_status_lock);
while (zone->zone_status < status) {
timeleft = cv_timedwait_sig(&zone->zone_cv, &zone_status_lock,
tim);
if (timeleft <= 0)
break;
}
mutex_exit(&zone_status_lock);
return (timeleft);
}
/*
* Zones have two reference counts: one for references from credential
* structures (zone_cred_ref), and one (zone_ref) for everything else.
* This is so we can allow a zone to be rebooted while there are still
* outstanding cred references, since certain drivers cache dblks (which
* implicitly results in cached creds). We wait for zone_ref to drop to
* 0 (actually 1), but not zone_cred_ref. The zone structure itself is
* later freed when the zone_cred_ref drops to 0, though nothing other
* than the zone id and privilege set should be accessed once the zone
* is "dead".
*
* A debugging flag, zone_wait_for_cred, can be set to a non-zero value
* to force halt/reboot to block waiting for the zone_cred_ref to drop
* to 0. This can be useful to flush out other sources of cached creds
* that may be less innocuous than the driver case.
*/
int zone_wait_for_cred = 0;
static void
zone_hold_locked(zone_t *z)
{
ASSERT(MUTEX_HELD(&z->zone_lock));
z->zone_ref++;
ASSERT(z->zone_ref != 0);
}
void
zone_hold(zone_t *z)
{
mutex_enter(&z->zone_lock);
zone_hold_locked(z);
mutex_exit(&z->zone_lock);
}
/*
* If the non-cred ref count drops to 1 and either the cred ref count
* is 0 or we aren't waiting for cred references, the zone is ready to
* be destroyed.
*/
#define ZONE_IS_UNREF(zone) ((zone)->zone_ref == 1 && \
(!zone_wait_for_cred || (zone)->zone_cred_ref == 0))
void
zone_rele(zone_t *z)
{
boolean_t wakeup;
mutex_enter(&z->zone_lock);
ASSERT(z->zone_ref != 0);
z->zone_ref--;
if (z->zone_ref == 0 && z->zone_cred_ref == 0) {
/* no more refs, free the structure */
mutex_exit(&z->zone_lock);
zone_free(z);
return;
}
/* signal zone_destroy so the zone can finish halting */
wakeup = (ZONE_IS_UNREF(z) && zone_status_get(z) >= ZONE_IS_DEAD);
mutex_exit(&z->zone_lock);
if (wakeup) {
/*
* Grabbing zonehash_lock here effectively synchronizes with
* zone_destroy() to avoid missed signals.
*/
mutex_enter(&zonehash_lock);
cv_broadcast(&zone_destroy_cv);
mutex_exit(&zonehash_lock);
}
}
void
zone_cred_hold(zone_t *z)
{
mutex_enter(&z->zone_lock);
z->zone_cred_ref++;
ASSERT(z->zone_cred_ref != 0);
mutex_exit(&z->zone_lock);
}
void
zone_cred_rele(zone_t *z)
{
boolean_t wakeup;
mutex_enter(&z->zone_lock);
ASSERT(z->zone_cred_ref != 0);
z->zone_cred_ref--;
if (z->zone_ref == 0 && z->zone_cred_ref == 0) {
/* no more refs, free the structure */
mutex_exit(&z->zone_lock);
zone_free(z);
return;
}
/*
* If zone_destroy is waiting for the cred references to drain
* out, and they have, signal it.
*/
wakeup = (zone_wait_for_cred && ZONE_IS_UNREF(z) &&
zone_status_get(z) >= ZONE_IS_DEAD);
mutex_exit(&z->zone_lock);
if (wakeup) {
/*
* Grabbing zonehash_lock here effectively synchronizes with
* zone_destroy() to avoid missed signals.
*/
mutex_enter(&zonehash_lock);
cv_broadcast(&zone_destroy_cv);
mutex_exit(&zonehash_lock);
}
}
void
zone_task_hold(zone_t *z)
{
mutex_enter(&z->zone_lock);
z->zone_ntasks++;
ASSERT(z->zone_ntasks != 0);
mutex_exit(&z->zone_lock);
}
void
zone_task_rele(zone_t *zone)
{
uint_t refcnt;
mutex_enter(&zone->zone_lock);
ASSERT(zone->zone_ntasks != 0);
refcnt = --zone->zone_ntasks;
if (refcnt > 1) { /* Common case */
mutex_exit(&zone->zone_lock);
return;
}
zone_hold_locked(zone); /* so we can use the zone_t later */
mutex_exit(&zone->zone_lock);
if (refcnt == 1) {
/*
* See if the zone is shutting down.
*/
mutex_enter(&zone_status_lock);
if (zone_status_get(zone) != ZONE_IS_SHUTTING_DOWN) {
goto out;
}
/*
* Make sure the ntasks didn't change since we
* dropped zone_lock.
*/
mutex_enter(&zone->zone_lock);
if (refcnt != zone->zone_ntasks) {
mutex_exit(&zone->zone_lock);
goto out;
}
mutex_exit(&zone->zone_lock);
/*
* No more user processes in the zone. The zone is empty.
*/
zone_status_set(zone, ZONE_IS_EMPTY);
goto out;
}
ASSERT(refcnt == 0);
/*
* zsched has exited; the zone is dead.
*/
zone->zone_zsched = NULL; /* paranoia */
mutex_enter(&zone_status_lock);
zone_status_set(zone, ZONE_IS_DEAD);
out:
mutex_exit(&zone_status_lock);
zone_rele(zone);
}
zoneid_t
getzoneid(void)
{
return (curproc->p_zone->zone_id);
}
/*
* Internal versions of zone_find_by_*(). These don't zone_hold() or
* check the validity of a zone's state.
*/
static zone_t *
zone_find_all_by_id(zoneid_t zoneid)
{
mod_hash_val_t hv;
zone_t *zone = NULL;
ASSERT(MUTEX_HELD(&zonehash_lock));
if (mod_hash_find(zonehashbyid,
(mod_hash_key_t)(uintptr_t)zoneid, &hv) == 0)
zone = (zone_t *)hv;
return (zone);
}
static zone_t *
zone_find_all_by_name(char *name)
{
mod_hash_val_t hv;
zone_t *zone = NULL;
ASSERT(MUTEX_HELD(&zonehash_lock));
if (mod_hash_find(zonehashbyname, (mod_hash_key_t)name, &hv) == 0)
zone = (zone_t *)hv;
return (zone);
}
/*
* Public interface for looking up a zone by zoneid. Only returns the zone if
* it is fully initialized, and has not yet begun the zone_destroy() sequence.
* Caller must call zone_rele() once it is done with the zone.
*
* The zone may begin the zone_destroy() sequence immediately after this
* function returns, but may be safely used until zone_rele() is called.
*/
zone_t *
zone_find_by_id(zoneid_t zoneid)
{
zone_t *zone;
zone_status_t status;
mutex_enter(&zonehash_lock);
if ((zone = zone_find_all_by_id(zoneid)) == NULL) {
mutex_exit(&zonehash_lock);
return (NULL);
}
status = zone_status_get(zone);
if (status < ZONE_IS_READY || status > ZONE_IS_DOWN) {
/*
* For all practical purposes the zone doesn't exist.
*/
mutex_exit(&zonehash_lock);
return (NULL);
}
zone_hold(zone);
mutex_exit(&zonehash_lock);
return (zone);
}
/*
* Similar to zone_find_by_id, but using zone name as the key.
*/
zone_t *
zone_find_by_name(char *name)
{
zone_t *zone;
zone_status_t status;
mutex_enter(&zonehash_lock);
if ((zone = zone_find_all_by_name(name)) == NULL) {
mutex_exit(&zonehash_lock);
return (NULL);
}
status = zone_status_get(zone);
if (status < ZONE_IS_READY || status > ZONE_IS_DOWN) {
/*
* For all practical purposes the zone doesn't exist.
*/
mutex_exit(&zonehash_lock);
return (NULL);
}
zone_hold(zone);
mutex_exit(&zonehash_lock);
return (zone);
}
/*
* Similar to zone_find_by_id(), using the path as a key. For instance,
* if there is a zone "foo" rooted at /foo/root, and the path argument
* is "/foo/root/proc", it will return the held zone_t corresponding to
* zone "foo".
*
* zone_find_by_path() always returns a non-NULL value, since at the
* very least every path will be contained in the global zone.
*
* As with the other zone_find_by_*() functions, the caller is
* responsible for zone_rele()ing the return value of this function.
*/
zone_t *
zone_find_by_path(const char *path)
{
zone_t *zone;
zone_t *zret = NULL;
zone_status_t status;
if (path == NULL) {
/*
* Call from rootconf().
*/
zone_hold(global_zone);
return (global_zone);
}
ASSERT(*path == '/');
mutex_enter(&zonehash_lock);
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone)) {
if (ZONE_PATH_VISIBLE(path, zone))
zret = zone;
}
ASSERT(zret != NULL);
status = zone_status_get(zret);
if (status < ZONE_IS_READY || status > ZONE_IS_DOWN) {
/*
* Zone practically doesn't exist.
*/
zret = global_zone;
}
zone_hold(zret);
mutex_exit(&zonehash_lock);
return (zret);
}
/*
* Get the number of cpus visible to this zone. The system-wide global
* 'ncpus' is returned if pools are disabled, the caller is in the
* global zone, or a NULL zone argument is passed in.
*/
int
zone_ncpus_get(zone_t *zone)
{
int myncpus = zone == NULL ? 0 : zone->zone_ncpus;
return (myncpus != 0 ? myncpus : ncpus);
}
/*
* Get the number of online cpus visible to this zone. The system-wide
* global 'ncpus_online' is returned if pools are disabled, the caller
* is in the global zone, or a NULL zone argument is passed in.
*/
int
zone_ncpus_online_get(zone_t *zone)
{
int myncpus_online = zone == NULL ? 0 : zone->zone_ncpus_online;
return (myncpus_online != 0 ? myncpus_online : ncpus_online);
}
/*
* Return the pool to which the zone is currently bound.
*/
pool_t *
zone_pool_get(zone_t *zone)
{
ASSERT(pool_lock_held());
return (zone->zone_pool);
}
/*
* Set the zone's pool pointer and update the zone's visibility to match
* the resources in the new pool.
*/
void
zone_pool_set(zone_t *zone, pool_t *pool)
{
ASSERT(pool_lock_held());
ASSERT(MUTEX_HELD(&cpu_lock));
zone->zone_pool = pool;
zone_pset_set(zone, pool->pool_pset->pset_id);
}
/*
* Return the cached value of the id of the processor set to which the
* zone is currently bound. The value will be ZONE_PS_INVAL if the pools
* facility is disabled.
*/
psetid_t
zone_pset_get(zone_t *zone)
{
ASSERT(MUTEX_HELD(&cpu_lock));
return (zone->zone_psetid);
}
/*
* Set the cached value of the id of the processor set to which the zone
* is currently bound. Also update the zone's visibility to match the
* resources in the new processor set.
*/
void
zone_pset_set(zone_t *zone, psetid_t newpsetid)
{
psetid_t oldpsetid;
ASSERT(MUTEX_HELD(&cpu_lock));
oldpsetid = zone_pset_get(zone);
if (oldpsetid == newpsetid)
return;
/*
* Global zone sees all.
*/
if (zone != global_zone) {
zone->zone_psetid = newpsetid;
if (newpsetid != ZONE_PS_INVAL)
pool_pset_visibility_add(newpsetid, zone);
if (oldpsetid != ZONE_PS_INVAL)
pool_pset_visibility_remove(oldpsetid, zone);
}
/*
* Disabling pools, so we should start using the global values
* for ncpus and ncpus_online.
*/
if (newpsetid == ZONE_PS_INVAL) {
zone->zone_ncpus = 0;
zone->zone_ncpus_online = 0;
}
}
/*
* Walk the list of active zones and issue the provided callback for
* each of them.
*
* Caller must not be holding any locks that may be acquired under
* zonehash_lock. See comment at the beginning of the file for a list of
* common locks and their interactions with zones.
*/
int
zone_walk(int (*cb)(zone_t *, void *), void *data)
{
zone_t *zone;
int ret = 0;
zone_status_t status;
mutex_enter(&zonehash_lock);
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone)) {
/*
* Skip zones that shouldn't be externally visible.
*/
status = zone_status_get(zone);
if (status < ZONE_IS_READY || status > ZONE_IS_DOWN)
continue;
/*
* Bail immediately if any callback invocation returns a
* non-zero value.
*/
ret = (*cb)(zone, data);
if (ret != 0)
break;
}
mutex_exit(&zonehash_lock);
return (ret);
}
static int
zone_set_root(zone_t *zone, const char *upath)
{
vnode_t *vp;
int trycount;
int error = 0;
char *path;
struct pathname upn, pn;
size_t pathlen;
if ((error = pn_get((char *)upath, UIO_USERSPACE, &upn)) != 0)
return (error);
pn_alloc(&pn);
/* prevent infinite loop */
trycount = 10;
for (;;) {
if (--trycount <= 0) {
error = ESTALE;
goto out;
}
if ((error = lookuppn(&upn, &pn, FOLLOW, NULLVPP, &vp)) == 0) {
/*
* VOP_ACCESS() may cover 'vp' with a new
* filesystem, if 'vp' is an autoFS vnode.
* Get the new 'vp' if so.
*/
if ((error = VOP_ACCESS(vp, VEXEC, 0, CRED())) == 0 &&
(vp->v_vfsmountedhere == NULL ||
(error = traverse(&vp)) == 0)) {
pathlen = pn.pn_pathlen + 2;
path = kmem_alloc(pathlen, KM_SLEEP);
(void) strncpy(path, pn.pn_path,
pn.pn_pathlen + 1);
path[pathlen - 2] = '/';
path[pathlen - 1] = '\0';
pn_free(&pn);
pn_free(&upn);
/* Success! */
break;
}
VN_RELE(vp);
}
if (error != ESTALE)
goto out;
}
ASSERT(error == 0);
zone->zone_rootvp = vp; /* we hold a reference to vp */
zone->zone_rootpath = path;
zone->zone_rootpathlen = pathlen;
return (0);
out:
pn_free(&pn);
pn_free(&upn);
return (error);
}
#define isalnum(c) (((c) >= '0' && (c) <= '9') || \
((c) >= 'a' && (c) <= 'z') || \
((c) >= 'A' && (c) <= 'Z'))
static int
zone_set_name(zone_t *zone, const char *uname)
{
char *kname = kmem_zalloc(ZONENAME_MAX, KM_SLEEP);
size_t len;
int i, err;
if ((err = copyinstr(uname, kname, ZONENAME_MAX, &len)) != 0) {
kmem_free(kname, ZONENAME_MAX);
return (err); /* EFAULT or ENAMETOOLONG */
}
/* must be less than ZONENAME_MAX */
if (len == ZONENAME_MAX && kname[ZONENAME_MAX - 1] != '\0') {
kmem_free(kname, ZONENAME_MAX);
return (EINVAL);
}
/*
* Name must start with an alphanumeric and must contain only
* alphanumerics, '-', '_' and '.'.
*/
if (!isalnum(kname[0])) {
kmem_free(kname, ZONENAME_MAX);
return (EINVAL);
}
for (i = 1; i < len - 1; i++) {
if (!isalnum(kname[i]) && kname[i] != '-' && kname[i] != '_' &&
kname[i] != '.') {
kmem_free(kname, ZONENAME_MAX);
return (EINVAL);
}
}
zone->zone_name = kname;
return (0);
}
/*
* Similar to thread_create(), but makes sure the thread is in the appropriate
* zone's zsched process (curproc->p_zone->zone_zsched) before returning.
*/
/*ARGSUSED*/
kthread_t *
zthread_create(
caddr_t stk,
size_t stksize,
void (*proc)(),
void *arg,
size_t len,
pri_t pri)
{
kthread_t *t;
zone_t *zone = curproc->p_zone;
proc_t *pp = zone->zone_zsched;
zone_hold(zone); /* Reference to be dropped when thread exits */
/*
* No-one should be trying to create threads if the zone is shutting
* down and there aren't any kernel threads around. See comment
* in zthread_exit().
*/
ASSERT(!(zone->zone_kthreads == NULL &&
zone_status_get(zone) >= ZONE_IS_EMPTY));
/*
* Create a thread, but don't let it run until we've finished setting
* things up.
*/
t = thread_create(stk, stksize, proc, arg, len, pp, TS_STOPPED, pri);
ASSERT(t->t_forw == NULL);
mutex_enter(&zone_status_lock);
if (zone->zone_kthreads == NULL) {
t->t_forw = t->t_back = t;
} else {
kthread_t *tx = zone->zone_kthreads;
t->t_forw = tx;
t->t_back = tx->t_back;
tx->t_back->t_forw = t;
tx->t_back = t;
}
zone->zone_kthreads = t;
mutex_exit(&zone_status_lock);
mutex_enter(&pp->p_lock);
t->t_proc_flag |= TP_ZTHREAD;
project_rele(t->t_proj);
t->t_proj = project_hold(pp->p_task->tk_proj);
/*
* Setup complete, let it run.
*/
thread_lock(t);
t->t_schedflag |= TS_ALLSTART;
setrun_locked(t);
thread_unlock(t);
mutex_exit(&pp->p_lock);
return (t);
}
/*
* Similar to thread_exit(). Must be called by threads created via
* zthread_exit().
*/
void
zthread_exit(void)
{
kthread_t *t = curthread;
proc_t *pp = curproc;
zone_t *zone = pp->p_zone;
mutex_enter(&zone_status_lock);
/*
* Reparent to p0
*/
kpreempt_disable();
mutex_enter(&pp->p_lock);
t->t_proc_flag &= ~TP_ZTHREAD;
t->t_procp = &p0;
hat_thread_exit(t);
mutex_exit(&pp->p_lock);
kpreempt_enable();
if (t->t_back == t) {
ASSERT(t->t_forw == t);
/*
* If the zone is empty, once the thread count
* goes to zero no further kernel threads can be
* created. This is because if the creator is a process
* in the zone, then it must have exited before the zone
* state could be set to ZONE_IS_EMPTY.
* Otherwise, if the creator is a kernel thread in the
* zone, the thread count is non-zero.
*
* This really means that non-zone kernel threads should
* not create zone kernel threads.
*/
zone->zone_kthreads = NULL;
if (zone_status_get(zone) == ZONE_IS_EMPTY) {
zone_status_set(zone, ZONE_IS_DOWN);
}
} else {
t->t_forw->t_back = t->t_back;
t->t_back->t_forw = t->t_forw;
if (zone->zone_kthreads == t)
zone->zone_kthreads = t->t_forw;
}
mutex_exit(&zone_status_lock);
zone_rele(zone);
thread_exit();
/* NOTREACHED */
}
static void
zone_chdir(vnode_t *vp, vnode_t **vpp, proc_t *pp)
{
vnode_t *oldvp;
/* we're going to hold a reference here to the directory */
VN_HOLD(vp);
#ifdef C2_AUDIT
if (audit_active) /* update abs cwd/root path see c2audit.c */
audit_chdirec(vp, vpp);
#endif
mutex_enter(&pp->p_lock);
oldvp = *vpp;
*vpp = vp;
mutex_exit(&pp->p_lock);
if (oldvp != NULL)
VN_RELE(oldvp);
}
/*
* Convert an rctl value represented by an nvlist_t into an rctl_val_t.
*/
static int
nvlist2rctlval(nvlist_t *nvl, rctl_val_t *rv)
{
nvpair_t *nvp = NULL;
boolean_t priv_set = B_FALSE;
boolean_t limit_set = B_FALSE;
boolean_t action_set = B_FALSE;
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
const char *name;
uint64_t ui64;
name = nvpair_name(nvp);
if (nvpair_type(nvp) != DATA_TYPE_UINT64)
return (EINVAL);
(void) nvpair_value_uint64(nvp, &ui64);
if (strcmp(name, "privilege") == 0) {
/*
* Currently only privileged values are allowed, but
* this may change in the future.
*/
if (ui64 != RCPRIV_PRIVILEGED)
return (EINVAL);
rv->rcv_privilege = ui64;
priv_set = B_TRUE;
} else if (strcmp(name, "limit") == 0) {
rv->rcv_value = ui64;
limit_set = B_TRUE;
} else if (strcmp(name, "action") == 0) {
if (ui64 != RCTL_LOCAL_NOACTION &&
ui64 != RCTL_LOCAL_DENY)
return (EINVAL);
rv->rcv_flagaction = ui64;
action_set = B_TRUE;
} else {
return (EINVAL);
}
}
if (!(priv_set && limit_set && action_set))
return (EINVAL);
rv->rcv_action_signal = 0;
rv->rcv_action_recipient = NULL;
rv->rcv_action_recip_pid = -1;
rv->rcv_firing_time = 0;
return (0);
}
void
zone_icode(void)
{
proc_t *p = ttoproc(curthread);
struct core_globals *cg;
/*
* For all purposes (ZONE_ATTR_INITPID and restart_init),
* storing just the pid of init is sufficient.
*/
p->p_zone->zone_proc_initpid = p->p_pid;
/*
* Allocate user address space and stack segment
*/
p->p_cstime = p->p_stime = p->p_cutime = p->p_utime = 0;
p->p_usrstack = (caddr_t)USRSTACK32;
p->p_model = DATAMODEL_ILP32;
p->p_stkprot = PROT_ZFOD & ~PROT_EXEC;
p->p_datprot = PROT_ZFOD & ~PROT_EXEC;
p->p_stk_ctl = INT32_MAX;
p->p_as = as_alloc();
p->p_as->a_userlimit = (caddr_t)USERLIMIT32;
(void) hat_setup(p->p_as->a_hat, HAT_INIT);
cg = zone_getspecific(core_zone_key, p->p_zone);
ASSERT(cg != NULL);
corectl_path_hold(cg->core_default_path);
corectl_content_hold(cg->core_default_content);
p->p_corefile = cg->core_default_path;
p->p_content = cg->core_default_content;
init_mstate(curthread, LMS_SYSTEM);
p->p_zone->zone_boot_err = exec_init(zone_initname, 0,
p->p_zone->zone_bootargs);
mutex_enter(&zone_status_lock);
if (p->p_zone->zone_boot_err != 0) {
/*
* Make sure we are still in the booting state-- we could have
* raced and already be shutting down, or even further along.
*/
if (zone_status_get(p->p_zone) == ZONE_IS_BOOTING)
zone_status_set(p->p_zone, ZONE_IS_SHUTTING_DOWN);
mutex_exit(&zone_status_lock);
/* It's gone bad, dispose of the process */
if (proc_exit(CLD_EXITED, p->p_zone->zone_boot_err) != 0) {
mutex_enter(&p->p_lock);
ASSERT(p->p_flag & SEXITLWPS);
lwp_exit();
}
} else {
if (zone_status_get(p->p_zone) == ZONE_IS_BOOTING)
zone_status_set(p->p_zone, ZONE_IS_RUNNING);
mutex_exit(&zone_status_lock);
/* cause the process to return to userland. */
lwp_rtt();
}
}
struct zsched_arg {
zone_t *zone;
nvlist_t *nvlist;
};
/*
* Per-zone "sched" workalike. The similarity to "sched" doesn't have
* anything to do with scheduling, but rather with the fact that
* per-zone kernel threads are parented to zsched, just like regular
* kernel threads are parented to sched (p0).
*
* zsched is also responsible for launching init for the zone.
*/
static void
zsched(void *arg)
{
struct zsched_arg *za = arg;
proc_t *pp = curproc;
proc_t *initp = proc_init;
zone_t *zone = za->zone;
cred_t *cr, *oldcred;
rctl_set_t *set;
rctl_alloc_gp_t *gp;
contract_t *ct = NULL;
task_t *tk, *oldtk;
rctl_entity_p_t e;
kproject_t *pj;
nvlist_t *nvl = za->nvlist;
nvpair_t *nvp = NULL;
bcopy("zsched", u.u_psargs, sizeof ("zsched"));
bcopy("zsched", u.u_comm, sizeof ("zsched"));
u.u_argc = 0;
u.u_argv = NULL;
u.u_envp = NULL;
closeall(P_FINFO(pp));
/*
* We are this zone's "zsched" process. As the zone isn't generally
* visible yet we don't need to grab any locks before initializing its
* zone_proc pointer.
*/
zone_hold(zone); /* this hold is released by zone_destroy() */
zone->zone_zsched = pp;
mutex_enter(&pp->p_lock);
pp->p_zone = zone;
mutex_exit(&pp->p_lock);
/*
* Disassociate process from its 'parent'; parent ourselves to init
* (pid 1) and change other values as needed.
*/
sess_create();
mutex_enter(&pidlock);
proc_detach(pp);
pp->p_ppid = 1;
pp->p_flag |= SZONETOP;
pp->p_ancpid = 1;
pp->p_parent = initp;
pp->p_psibling = NULL;
if (initp->p_child)
initp->p_child->p_psibling = pp;
pp->p_sibling = initp->p_child;
initp->p_child = pp;
/* Decrement what newproc() incremented. */
upcount_dec(crgetruid(CRED()), GLOBAL_ZONEID);
/*
* Our credentials are about to become kcred-like, so we don't care
* about the caller's ruid.
*/
upcount_inc(crgetruid(kcred), zone->zone_id);
mutex_exit(&pidlock);
/*
* getting out of global zone, so decrement lwp counts
*/
pj = pp->p_task->tk_proj;
mutex_enter(&global_zone->zone_nlwps_lock);
pj->kpj_nlwps -= pp->p_lwpcnt;
global_zone->zone_nlwps -= pp->p_lwpcnt;
mutex_exit(&global_zone->zone_nlwps_lock);
/*
* Create and join a new task in project '0' of this zone.
*
* We don't need to call holdlwps() since we know we're the only lwp in
* this process.
*
* task_join() returns with p_lock held.
*/
tk = task_create(0, zone);
mutex_enter(&cpu_lock);
oldtk = task_join(tk, 0);
mutex_exit(&curproc->p_lock);
mutex_exit(&cpu_lock);
task_rele(oldtk);
/*
* add lwp counts to zsched's zone, and increment project's task count
* due to the task created in the above tasksys_settaskid
*/
pj = pp->p_task->tk_proj;
mutex_enter(&zone->zone_nlwps_lock);
pj->kpj_nlwps += pp->p_lwpcnt;
pj->kpj_ntasks += 1;
zone->zone_nlwps += pp->p_lwpcnt;
mutex_exit(&zone->zone_nlwps_lock);
/*
* The process was created by a process in the global zone, hence the
* credentials are wrong. We might as well have kcred-ish credentials.
*/
cr = zone->zone_kcred;
crhold(cr);
mutex_enter(&pp->p_crlock);
oldcred = pp->p_cred;
pp->p_cred = cr;
mutex_exit(&pp->p_crlock);
crfree(oldcred);
/*
* Hold credentials again (for thread)
*/
crhold(cr);
/*
* p_lwpcnt can't change since this is a kernel process.
*/
crset(pp, cr);
/*
* Chroot
*/
zone_chdir(zone->zone_rootvp, &PTOU(pp)->u_cdir, pp);
zone_chdir(zone->zone_rootvp, &PTOU(pp)->u_rdir, pp);
/*
* Initialize zone's rctl set.
*/
set = rctl_set_create();
gp = rctl_set_init_prealloc(RCENTITY_ZONE);
mutex_enter(&pp->p_lock);
e.rcep_p.zone = zone;
e.rcep_t = RCENTITY_ZONE;
zone->zone_rctls = rctl_set_init(RCENTITY_ZONE, pp, &e, set, gp);
mutex_exit(&pp->p_lock);
rctl_prealloc_destroy(gp);
/*
* Apply the rctls passed in to zone_create(). This is basically a list
* assignment: all of the old values are removed and the new ones
* inserted. That is, if an empty list is passed in, all values are
* removed.
*/
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
rctl_dict_entry_t *rde;
rctl_hndl_t hndl;
char *name;
nvlist_t **nvlarray;
uint_t i, nelem;
int error; /* For ASSERT()s */
name = nvpair_name(nvp);
hndl = rctl_hndl_lookup(name);
ASSERT(hndl != -1);
rde = rctl_dict_lookup_hndl(hndl);
ASSERT(rde != NULL);
for (; /* ever */; ) {
rctl_val_t oval;
mutex_enter(&pp->p_lock);
error = rctl_local_get(hndl, NULL, &oval, pp);
mutex_exit(&pp->p_lock);
ASSERT(error == 0); /* Can't fail for RCTL_FIRST */
ASSERT(oval.rcv_privilege != RCPRIV_BASIC);
if (oval.rcv_privilege == RCPRIV_SYSTEM)
break;
mutex_enter(&pp->p_lock);
error = rctl_local_delete(hndl, &oval, pp);
mutex_exit(&pp->p_lock);
ASSERT(error == 0);
}
error = nvpair_value_nvlist_array(nvp, &nvlarray, &nelem);
ASSERT(error == 0);
for (i = 0; i < nelem; i++) {
rctl_val_t *nvalp;
nvalp = kmem_cache_alloc(rctl_val_cache, KM_SLEEP);
error = nvlist2rctlval(nvlarray[i], nvalp);
ASSERT(error == 0);
/*
* rctl_local_insert can fail if the value being
* inserted is a duplicate; this is OK.
*/
mutex_enter(&pp->p_lock);
if (rctl_local_insert(hndl, nvalp, pp) != 0)
kmem_cache_free(rctl_val_cache, nvalp);
mutex_exit(&pp->p_lock);
}
}
/*
* Tell the world that we're done setting up.
*
* At this point we want to set the zone status to ZONE_IS_READY
* and atomically set the zone's processor set visibility. Once
* we drop pool_lock() this zone will automatically get updated
* to reflect any future changes to the pools configuration.
*/
pool_lock();
mutex_enter(&cpu_lock);
mutex_enter(&zonehash_lock);
zone_uniqid(zone);
zone_zsd_configure(zone);
if (pool_state == POOL_ENABLED)
zone_pset_set(zone, pool_default->pool_pset->pset_id);
mutex_enter(&zone_status_lock);
ASSERT(zone_status_get(zone) == ZONE_IS_UNINITIALIZED);
zone_status_set(zone, ZONE_IS_READY);
mutex_exit(&zone_status_lock);
mutex_exit(&zonehash_lock);
mutex_exit(&cpu_lock);
pool_unlock();
/*
* Once we see the zone transition to the ZONE_IS_BOOTING state,
* we launch init, and set the state to running.
*/
zone_status_wait_cpr(zone, ZONE_IS_BOOTING, "zsched");
if (zone_status_get(zone) == ZONE_IS_BOOTING) {
id_t cid;
/*
* Ok, this is a little complicated. We need to grab the
* zone's pool's scheduling class ID; note that by now, we
* are already bound to a pool if we need to be (zoneadmd
* will have done that to us while we're in the READY
* state). *But* the scheduling class for the zone's 'init'
* must be explicitly passed to newproc, which doesn't
* respect pool bindings.
*
* We hold the pool_lock across the call to newproc() to
* close the obvious race: the pool's scheduling class
* could change before we manage to create the LWP with
* classid 'cid'.
*/
pool_lock();
cid = pool_get_class(zone->zone_pool);
if (cid == -1)
cid = defaultcid;
/*
* If this fails, zone_boot will ultimately fail. The
* state of the zone will be set to SHUTTING_DOWN-- userland
* will have to tear down the zone, and fail, or try again.
*/
if ((zone->zone_boot_err = newproc(zone_icode, NULL, cid,
minclsyspri - 1, &ct)) != 0) {
mutex_enter(&zone_status_lock);
zone_status_set(zone, ZONE_IS_SHUTTING_DOWN);
mutex_exit(&zone_status_lock);
}
pool_unlock();
}
/*
* Wait for zone_destroy() to be called. This is what we spend
* most of our life doing.
*/
zone_status_wait_cpr(zone, ZONE_IS_DYING, "zsched");
if (ct)
/*
* At this point the process contract should be empty.
* (Though if it isn't, it's not the end of the world.)
*/
VERIFY(contract_abandon(ct, curproc, B_TRUE) == 0);
/*
* Allow kcred to be freed when all referring processes
* (including this one) go away. We can't just do this in
* zone_free because we need to wait for the zone_cred_ref to
* drop to 0 before calling zone_free, and the existence of
* zone_kcred will prevent that. Thus, we call crfree here to
* balance the crdup in zone_create. The crhold calls earlier
* in zsched will be dropped when the thread and process exit.
*/
crfree(zone->zone_kcred);
zone->zone_kcred = NULL;
exit(CLD_EXITED, 0);
}
/*
* Helper function to determine if there are any submounts of the
* provided path. Used to make sure the zone doesn't "inherit" any
* mounts from before it is created.
*/
static uint_t
zone_mount_count(const char *rootpath)
{
vfs_t *vfsp;
uint_t count = 0;
size_t rootpathlen = strlen(rootpath);
/*
* Holding zonehash_lock prevents race conditions with
* vfs_list_add()/vfs_list_remove() since we serialize with
* zone_find_by_path().
*/
ASSERT(MUTEX_HELD(&zonehash_lock));
/*
* The rootpath must end with a '/'
*/
ASSERT(rootpath[rootpathlen - 1] == '/');
/*
* This intentionally does not count the rootpath itself if that
* happens to be a mount point.
*/
vfs_list_read_lock();
vfsp = rootvfs;
do {
if (strncmp(rootpath, refstr_value(vfsp->vfs_mntpt),
rootpathlen) == 0)
count++;
vfsp = vfsp->vfs_next;
} while (vfsp != rootvfs);
vfs_list_unlock();
return (count);
}
/*
* Helper function to make sure that a zone created on 'rootpath'
* wouldn't end up containing other zones' rootpaths.
*/
static boolean_t
zone_is_nested(const char *rootpath)
{
zone_t *zone;
size_t rootpathlen = strlen(rootpath);
size_t len;
ASSERT(MUTEX_HELD(&zonehash_lock));
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone)) {
if (zone == global_zone)
continue;
len = strlen(zone->zone_rootpath);
if (strncmp(rootpath, zone->zone_rootpath,
MIN(rootpathlen, len)) == 0)
return (B_TRUE);
}
return (B_FALSE);
}
static int
zone_set_privset(zone_t *zone, const priv_set_t *zone_privs,
size_t zone_privssz)
{
priv_set_t *privs = kmem_alloc(sizeof (priv_set_t), KM_SLEEP);
if (zone_privssz < sizeof (priv_set_t))
return (set_errno(ENOMEM));
if (copyin(zone_privs, privs, sizeof (priv_set_t))) {
kmem_free(privs, sizeof (priv_set_t));
return (EFAULT);
}
zone->zone_privset = privs;
return (0);
}
/*
* We make creative use of nvlists to pass in rctls from userland. The list is
* a list of the following structures:
*
* (name = rctl_name, value = nvpair_list_array)
*
* Where each element of the nvpair_list_array is of the form:
*
* [(name = "privilege", value = RCPRIV_PRIVILEGED),
* (name = "limit", value = uint64_t),
* (name = "action", value = (RCTL_LOCAL_NOACTION || RCTL_LOCAL_DENY))]
*/
static int
parse_rctls(caddr_t ubuf, size_t buflen, nvlist_t **nvlp)
{
nvpair_t *nvp = NULL;
nvlist_t *nvl = NULL;
char *kbuf;
int error;
rctl_val_t rv;
*nvlp = NULL;
if (buflen == 0)
return (0);
if ((kbuf = kmem_alloc(buflen, KM_NOSLEEP)) == NULL)
return (ENOMEM);
if (copyin(ubuf, kbuf, buflen)) {
error = EFAULT;
goto out;
}
if (nvlist_unpack(kbuf, buflen, &nvl, KM_SLEEP) != 0) {
/*
* nvl may have been allocated/free'd, but the value set to
* non-NULL, so we reset it here.
*/
nvl = NULL;
error = EINVAL;
goto out;
}
while ((nvp = nvlist_next_nvpair(nvl, nvp)) != NULL) {
rctl_dict_entry_t *rde;
rctl_hndl_t hndl;
nvlist_t **nvlarray;
uint_t i, nelem;
char *name;
error = EINVAL;
name = nvpair_name(nvp);
if (strncmp(nvpair_name(nvp), "zone.", sizeof ("zone.") - 1)
!= 0 || nvpair_type(nvp) != DATA_TYPE_NVLIST_ARRAY) {
goto out;
}
if ((hndl = rctl_hndl_lookup(name)) == -1) {
goto out;
}
rde = rctl_dict_lookup_hndl(hndl);
error = nvpair_value_nvlist_array(nvp, &nvlarray, &nelem);
ASSERT(error == 0);
for (i = 0; i < nelem; i++) {
if (error = nvlist2rctlval(nvlarray[i], &rv))
goto out;
}
if (rctl_invalid_value(rde, &rv)) {
error = EINVAL;
goto out;
}
}
error = 0;
*nvlp = nvl;
out:
kmem_free(kbuf, buflen);
if (error && nvl != NULL)
nvlist_free(nvl);
return (error);
}
int
zone_create_error(int er_error, int er_ext, int *er_out) {
if (er_out != NULL) {
if (copyout(&er_ext, er_out, sizeof (int))) {
return (set_errno(EFAULT));
}
}
return (set_errno(er_error));
}
/*
* Parses a comma-separated list of ZFS datasets into a per-zone dictionary.
*/
static int
parse_zfs(zone_t *zone, caddr_t ubuf, size_t buflen)
{
char *kbuf;
char *dataset, *next;
zone_dataset_t *zd;
size_t len;
if (ubuf == NULL || buflen == 0)
return (0);
if ((kbuf = kmem_alloc(buflen, KM_NOSLEEP)) == NULL)
return (ENOMEM);
if (copyin(ubuf, kbuf, buflen) != 0) {
kmem_free(kbuf, buflen);
return (EFAULT);
}
dataset = next = kbuf;
for (;;) {
zd = kmem_alloc(sizeof (zone_dataset_t), KM_SLEEP);
next = strchr(dataset, ',');
if (next == NULL)
len = strlen(dataset);
else
len = next - dataset;
zd->zd_dataset = kmem_alloc(len + 1, KM_SLEEP);
bcopy(dataset, zd->zd_dataset, len);
zd->zd_dataset[len] = '\0';
list_insert_head(&zone->zone_datasets, zd);
if (next == NULL)
break;
dataset = next + 1;
}
kmem_free(kbuf, buflen);
return (0);
}
/*
* System call to create/initialize a new zone named 'zone_name', rooted
* at 'zone_root', with a zone-wide privilege limit set of 'zone_privs',
* and initialized with the zone-wide rctls described in 'rctlbuf'.
*
* If extended error is non-null, we may use it to return more detailed
* error information.
*/
static zoneid_t
zone_create(const char *zone_name, const char *zone_root,
const priv_set_t *zone_privs, size_t zone_privssz,
caddr_t rctlbuf, size_t rctlbufsz,
caddr_t zfsbuf, size_t zfsbufsz, int *extended_error)
{
struct zsched_arg zarg;
nvlist_t *rctls = NULL;
proc_t *pp = curproc;
zone_t *zone, *ztmp;
zoneid_t zoneid;
int error;
int error2 = 0;
char *str;
cred_t *zkcr;
if (secpolicy_zone_config(CRED()) != 0)
return (set_errno(EPERM));
/* can't boot zone from within chroot environment */
if (PTOU(pp)->u_rdir != NULL && PTOU(pp)->u_rdir != rootdir)
return (zone_create_error(ENOTSUP, ZE_CHROOTED,
extended_error));
zone = kmem_zalloc(sizeof (zone_t), KM_SLEEP);
zoneid = zone->zone_id = id_alloc(zoneid_space);
zone->zone_status = ZONE_IS_UNINITIALIZED;
zone->zone_pool = pool_default;
zone->zone_pool_mod = gethrtime();
zone->zone_psetid = ZONE_PS_INVAL;
zone->zone_ncpus = 0;
zone->zone_ncpus_online = 0;
mutex_init(&zone->zone_lock, NULL, MUTEX_DEFAULT, NULL);
mutex_init(&zone->zone_nlwps_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&zone->zone_cv, NULL, CV_DEFAULT, NULL);
list_create(&zone->zone_zsd, sizeof (struct zsd_entry),
offsetof(struct zsd_entry, zsd_linkage));
list_create(&zone->zone_datasets, sizeof (zone_dataset_t),
offsetof(zone_dataset_t, zd_linkage));
if ((error = zone_set_name(zone, zone_name)) != 0) {
zone_free(zone);
return (zone_create_error(error, 0, extended_error));
}
if ((error = zone_set_root(zone, zone_root)) != 0) {
zone_free(zone);
return (zone_create_error(error, 0, extended_error));
}
if ((error = zone_set_privset(zone, zone_privs, zone_privssz)) != 0) {
zone_free(zone);
return (zone_create_error(error, 0, extended_error));
}
/* initialize node name to be the same as zone name */
zone->zone_nodename = kmem_alloc(_SYS_NMLN, KM_SLEEP);
(void) strncpy(zone->zone_nodename, zone->zone_name, _SYS_NMLN);
zone->zone_nodename[_SYS_NMLN - 1] = '\0';
zone->zone_domain = kmem_alloc(_SYS_NMLN, KM_SLEEP);
zone->zone_domain[0] = '\0';
zone->zone_shares = 1;
zone->zone_bootargs = NULL;
/*
* Zsched initializes the rctls.
*/
zone->zone_rctls = NULL;
if ((error = parse_rctls(rctlbuf, rctlbufsz, &rctls)) != 0) {
zone_free(zone);
return (zone_create_error(error, 0, extended_error));
}
if ((error = parse_zfs(zone, zfsbuf, zfsbufsz)) != 0) {
zone_free(zone);
return (set_errno(error));
}
/*
* Stop all lwps since that's what normally happens as part of fork().
* This needs to happen before we grab any locks to avoid deadlock
* (another lwp in the process could be waiting for the held lock).
*/
if (curthread != pp->p_agenttp && !holdlwps(SHOLDFORK)) {
zone_free(zone);
if (rctls)
nvlist_free(rctls);
return (zone_create_error(error, 0, extended_error));
}
if (block_mounts() == 0) {
mutex_enter(&pp->p_lock);
if (curthread != pp->p_agenttp)
continuelwps(pp);
mutex_exit(&pp->p_lock);
zone_free(zone);
if (rctls)
nvlist_free(rctls);
return (zone_create_error(error, 0, extended_error));
}
/*
* Set up credential for kernel access. After this, any errors
* should go through the dance in errout rather than calling
* zone_free directly.
*/
zone->zone_kcred = crdup(kcred);
crsetzone(zone->zone_kcred, zone);
priv_intersect(zone->zone_privset, &CR_PPRIV(zone->zone_kcred));
priv_intersect(zone->zone_privset, &CR_EPRIV(zone->zone_kcred));
priv_intersect(zone->zone_privset, &CR_IPRIV(zone->zone_kcred));
priv_intersect(zone->zone_privset, &CR_LPRIV(zone->zone_kcred));
mutex_enter(&zonehash_lock);
/*
* Make sure zone doesn't already exist.
*/
if ((ztmp = zone_find_all_by_name(zone->zone_name)) != NULL) {
zone_status_t status;
status = zone_status_get(ztmp);
if (status == ZONE_IS_READY || status == ZONE_IS_RUNNING)
error = EEXIST;
else
error = EBUSY;
goto errout;
}
/*
* Don't allow zone creations which would cause one zone's rootpath to
* be accessible from that of another (non-global) zone.
*/
if (zone_is_nested(zone->zone_rootpath)) {
error = EBUSY;
goto errout;
}
ASSERT(zonecount != 0); /* check for leaks */
if (zonecount + 1 > maxzones) {
error = ENOMEM;
goto errout;
}
if (zone_mount_count(zone->zone_rootpath) != 0) {
error = EBUSY;
error2 = ZE_AREMOUNTS;
goto errout;
}
/*
* Zone is still incomplete, but we need to drop all locks while
* zsched() initializes this zone's kernel process. We
* optimistically add the zone to the hashtable and associated
* lists so a parallel zone_create() doesn't try to create the
* same zone.
*/
zonecount++;
(void) mod_hash_insert(zonehashbyid,
(mod_hash_key_t)(uintptr_t)zone->zone_id,
(mod_hash_val_t)(uintptr_t)zone);
str = kmem_alloc(strlen(zone->zone_name) + 1, KM_SLEEP);
(void) strcpy(str, zone->zone_name);
(void) mod_hash_insert(zonehashbyname, (mod_hash_key_t)str,
(mod_hash_val_t)(uintptr_t)zone);
/*
* Insert into active list. At this point there are no 'hold's
* on the zone, but everyone else knows not to use it, so we can
* continue to use it. zsched() will do a zone_hold() if the
* newproc() is successful.
*/
list_insert_tail(&zone_active, zone);
mutex_exit(&zonehash_lock);
zarg.zone = zone;
zarg.nvlist = rctls;
/*
* The process, task, and project rctls are probably wrong;
* we need an interface to get the default values of all rctls,
* and initialize zsched appropriately. I'm not sure that that
* makes much of a difference, though.
*/
if (error = newproc(zsched, (void *)&zarg, syscid, minclsyspri, NULL)) {
/*
* We need to undo all globally visible state.
*/
mutex_enter(&zonehash_lock);
list_remove(&zone_active, zone);
(void) mod_hash_destroy(zonehashbyname,
(mod_hash_key_t)(uintptr_t)zone->zone_name);
(void) mod_hash_destroy(zonehashbyid,
(mod_hash_key_t)(uintptr_t)zone->zone_id);
ASSERT(zonecount > 1);
zonecount--;
goto errout;
}
/*
* Zone creation can't fail from now on.
*/
/*
* Let the other lwps continue.
*/
mutex_enter(&pp->p_lock);
if (curthread != pp->p_agenttp)
continuelwps(pp);
mutex_exit(&pp->p_lock);
/*
* Wait for zsched to finish initializing the zone.
*/
zone_status_wait(zone, ZONE_IS_READY);
/*
* The zone is fully visible, so we can let mounts progress.
*/
resume_mounts();
if (rctls)
nvlist_free(rctls);
return (zoneid);
errout:
mutex_exit(&zonehash_lock);
/*
* Let the other lwps continue.
*/
mutex_enter(&pp->p_lock);
if (curthread != pp->p_agenttp)
continuelwps(pp);
mutex_exit(&pp->p_lock);
resume_mounts();
if (rctls)
nvlist_free(rctls);
/*
* There is currently one reference to the zone, a cred_ref from
* zone_kcred. To free the zone, we call crfree, which will call
* zone_cred_rele, which will call zone_free.
*/
ASSERT(zone->zone_cred_ref == 1); /* for zone_kcred */
ASSERT(zone->zone_kcred->cr_ref == 1);
ASSERT(zone->zone_ref == 0);
zkcr = zone->zone_kcred;
zone->zone_kcred = NULL;
crfree(zkcr); /* triggers call to zone_free */
return (zone_create_error(error, error2, extended_error));
}
/*
* Cause the zone to boot. This is pretty simple, since we let zoneadmd do
* the heavy lifting.
*/
static int
zone_boot(zoneid_t zoneid, const char *bootargs)
{
int err;
zone_t *zone;
if (secpolicy_zone_config(CRED()) != 0)
return (set_errno(EPERM));
if (zoneid < MIN_USERZONEID || zoneid > MAX_ZONEID)
return (set_errno(EINVAL));
mutex_enter(&zonehash_lock);
/*
* Look for zone under hash lock to prevent races with calls to
* zone_shutdown, zone_destroy, etc.
*/
if ((zone = zone_find_all_by_id(zoneid)) == NULL) {
mutex_exit(&zonehash_lock);
return (set_errno(EINVAL));
}
if ((err = zone_set_bootargs(zone, bootargs)) != 0) {
mutex_exit(&zonehash_lock);
return (set_errno(err));
}
mutex_enter(&zone_status_lock);
if (zone_status_get(zone) != ZONE_IS_READY) {
mutex_exit(&zone_status_lock);
mutex_exit(&zonehash_lock);
return (set_errno(EINVAL));
}
zone_status_set(zone, ZONE_IS_BOOTING);
mutex_exit(&zone_status_lock);
zone_hold(zone); /* so we can use the zone_t later */
mutex_exit(&zonehash_lock);
if (zone_status_wait_sig(zone, ZONE_IS_RUNNING) == 0) {
zone_rele(zone);
return (set_errno(EINTR));
}
/*
* Boot (starting init) might have failed, in which case the zone
* will go to the SHUTTING_DOWN state; an appropriate errno will
* be placed in zone->zone_boot_err, and so we return that.
*/
err = zone->zone_boot_err;
zone_rele(zone);
return (err ? set_errno(err) : 0);
}
/*
* Kills all user processes in the zone, waiting for them all to exit
* before returning.
*/
static int
zone_empty(zone_t *zone)
{
int waitstatus;
/*
* We need to drop zonehash_lock before killing all
* processes, otherwise we'll deadlock with zone_find_*
* which can be called from the exit path.
*/
ASSERT(MUTEX_NOT_HELD(&zonehash_lock));
while ((waitstatus = zone_status_timedwait_sig(zone, lbolt + hz,
ZONE_IS_EMPTY)) == -1) {
killall(zone->zone_id);
}
/*
* return EINTR if we were signaled
*/
if (waitstatus == 0)
return (EINTR);
return (0);
}
/*
* Systemcall to start the zone's halt sequence. By the time this
* function successfully returns, all user processes and kernel threads
* executing in it will have exited, ZSD shutdown callbacks executed,
* and the zone status set to ZONE_IS_DOWN.
*
* It is possible that the call will interrupt itself if the caller is the
* parent of any process running in the zone, and doesn't have SIGCHLD blocked.
*/
static int
zone_shutdown(zoneid_t zoneid)
{
int error;
zone_t *zone;
zone_status_t status;
if (secpolicy_zone_config(CRED()) != 0)
return (set_errno(EPERM));
if (zoneid < MIN_USERZONEID || zoneid > MAX_ZONEID)
return (set_errno(EINVAL));
/*
* Block mounts so that VFS_MOUNT() can get an accurate view of
* the zone's status with regards to ZONE_IS_SHUTTING down.
*
* e.g. NFS can fail the mount if it determines that the zone
* has already begun the shutdown sequence.
*/
if (block_mounts() == 0)
return (set_errno(EINTR));
mutex_enter(&zonehash_lock);
/*
* Look for zone under hash lock to prevent races with other
* calls to zone_shutdown and zone_destroy.
*/
if ((zone = zone_find_all_by_id(zoneid)) == NULL) {
mutex_exit(&zonehash_lock);
resume_mounts();
return (set_errno(EINVAL));
}
mutex_enter(&zone_status_lock);
status = zone_status_get(zone);
/*
* Fail if the zone isn't fully initialized yet.
*/
if (status < ZONE_IS_READY) {
mutex_exit(&zone_status_lock);
mutex_exit(&zonehash_lock);
resume_mounts();
return (set_errno(EINVAL));
}
/*
* If conditions required for zone_shutdown() to return have been met,
* return success.
*/
if (status >= ZONE_IS_DOWN) {
mutex_exit(&zone_status_lock);
mutex_exit(&zonehash_lock);
resume_mounts();
return (0);
}
/*
* If zone_shutdown() hasn't been called before, go through the motions.
* If it has, there's nothing to do but wait for the kernel threads to
* drain.
*/
if (status < ZONE_IS_EMPTY) {
uint_t ntasks;
mutex_enter(&zone->zone_lock);
if ((ntasks = zone->zone_ntasks) != 1) {
/*
* There's still stuff running.
*/
zone_status_set(zone, ZONE_IS_SHUTTING_DOWN);
}
mutex_exit(&zone->zone_lock);
if (ntasks == 1) {
/*
* The only way to create another task is through
* zone_enter(), which will block until we drop
* zonehash_lock. The zone is empty.
*/
if (zone->zone_kthreads == NULL) {
/*
* Skip ahead to ZONE_IS_DOWN
*/
zone_status_set(zone, ZONE_IS_DOWN);
} else {
zone_status_set(zone, ZONE_IS_EMPTY);
}
}
}
zone_hold(zone); /* so we can use the zone_t later */
mutex_exit(&zone_status_lock);
mutex_exit(&zonehash_lock);
resume_mounts();
if (error = zone_empty(zone)) {
zone_rele(zone);
return (set_errno(error));
}
/*
* After the zone status goes to ZONE_IS_DOWN this zone will no
* longer be notified of changes to the pools configuration, so
* in order to not end up with a stale pool pointer, we point
* ourselves at the default pool and remove all resource
* visibility. This is especially important as the zone_t may
* languish on the deathrow for a very long time waiting for
* cred's to drain out.
*
* This rebinding of the zone can happen multiple times
* (presumably due to interrupted or parallel systemcalls)
* without any adverse effects.
*/
if (pool_lock_intr() != 0) {
zone_rele(zone);
return (set_errno(EINTR));
}
if (pool_state == POOL_ENABLED) {
mutex_enter(&cpu_lock);
zone_pool_set(zone, pool_default);
/*
* The zone no longer needs to be able to see any cpus.
*/
zone_pset_set(zone, ZONE_PS_INVAL);
mutex_exit(&cpu_lock);
}
pool_unlock();
/*
* ZSD shutdown callbacks can be executed multiple times, hence
* it is safe to not be holding any locks across this call.
*/
zone_zsd_callbacks(zone, ZSD_SHUTDOWN);
mutex_enter(&zone_status_lock);
if (zone->zone_kthreads == NULL && zone_status_get(zone) < ZONE_IS_DOWN)
zone_status_set(zone, ZONE_IS_DOWN);
mutex_exit(&zone_status_lock);
/*
* Wait for kernel threads to drain.
*/
if (!zone_status_wait_sig(zone, ZONE_IS_DOWN)) {
zone_rele(zone);
return (set_errno(EINTR));
}
zone_rele(zone);
return (0);
}
/*
* Systemcall entry point to finalize the zone halt process. The caller
* must have already successfully callefd zone_shutdown().
*
* Upon successful completion, the zone will have been fully destroyed:
* zsched will have exited, destructor callbacks executed, and the zone
* removed from the list of active zones.
*/
static int
zone_destroy(zoneid_t zoneid)
{
uint64_t uniqid;
zone_t *zone;
zone_status_t status;
if (secpolicy_zone_config(CRED()) != 0)
return (set_errno(EPERM));
if (zoneid < MIN_USERZONEID || zoneid > MAX_ZONEID)
return (set_errno(EINVAL));
mutex_enter(&zonehash_lock);
/*
* Look for zone under hash lock to prevent races with other
* calls to zone_destroy.
*/
if ((zone = zone_find_all_by_id(zoneid)) == NULL) {
mutex_exit(&zonehash_lock);
return (set_errno(EINVAL));
}
if (zone_mount_count(zone->zone_rootpath) != 0) {
mutex_exit(&zonehash_lock);
return (set_errno(EBUSY));
}
mutex_enter(&zone_status_lock);
status = zone_status_get(zone);
if (status < ZONE_IS_DOWN) {
mutex_exit(&zone_status_lock);
mutex_exit(&zonehash_lock);
return (set_errno(EBUSY));
} else if (status == ZONE_IS_DOWN) {
zone_status_set(zone, ZONE_IS_DYING); /* Tell zsched to exit */
}
mutex_exit(&zone_status_lock);
zone_hold(zone);
mutex_exit(&zonehash_lock);
/*
* wait for zsched to exit
*/
zone_status_wait(zone, ZONE_IS_DEAD);
zone_zsd_callbacks(zone, ZSD_DESTROY);
uniqid = zone->zone_uniqid;
zone_rele(zone);
zone = NULL; /* potentially free'd */
mutex_enter(&zonehash_lock);
for (; /* ever */; ) {
boolean_t unref;
if ((zone = zone_find_all_by_id(zoneid)) == NULL ||
zone->zone_uniqid != uniqid) {
/*
* The zone has gone away. Necessary conditions
* are met, so we return success.
*/
mutex_exit(&zonehash_lock);
return (0);
}
mutex_enter(&zone->zone_lock);
unref = ZONE_IS_UNREF(zone);
mutex_exit(&zone->zone_lock);
if (unref) {
/*
* There is only one reference to the zone -- that
* added when the zone was added to the hashtables --
* and things will remain this way until we drop
* zonehash_lock... we can go ahead and cleanup the
* zone.
*/
break;
}
if (cv_wait_sig(&zone_destroy_cv, &zonehash_lock) == 0) {
/* Signaled */
mutex_exit(&zonehash_lock);
return (set_errno(EINTR));
}
}
/*
* It is now safe to let the zone be recreated; remove it from the
* lists. The memory will not be freed until the last cred
* reference goes away.
*/
ASSERT(zonecount > 1); /* must be > 1; can't destroy global zone */
zonecount--;
/* remove from active list and hash tables */
list_remove(&zone_active, zone);
(void) mod_hash_destroy(zonehashbyname,
(mod_hash_key_t)zone->zone_name);
(void) mod_hash_destroy(zonehashbyid,
(mod_hash_key_t)(uintptr_t)zone->zone_id);
mutex_exit(&zonehash_lock);
/*
* Release the root vnode; we're not using it anymore. Nor should any
* other thread that might access it exist.
*/
if (zone->zone_rootvp != NULL) {
VN_RELE(zone->zone_rootvp);
zone->zone_rootvp = NULL;
}
/* add to deathrow list */
mutex_enter(&zone_deathrow_lock);
list_insert_tail(&zone_deathrow, zone);
mutex_exit(&zone_deathrow_lock);
/*
* Drop last reference (which was added by zsched()), this will
* free the zone unless there are outstanding cred references.
*/
zone_rele(zone);
return (0);
}
/*
* Systemcall entry point for zone_getattr(2).
*/
static ssize_t
zone_getattr(zoneid_t zoneid, int attr, void *buf, size_t bufsize)
{
size_t size;
int error = 0, err;
zone_t *zone;
char *zonepath;
zone_status_t zone_status;
pid_t initpid;
boolean_t global = (curproc->p_zone == global_zone);
mutex_enter(&zonehash_lock);
if ((zone = zone_find_all_by_id(zoneid)) == NULL) {
mutex_exit(&zonehash_lock);
return (set_errno(EINVAL));
}
zone_status = zone_status_get(zone);
if (zone_status < ZONE_IS_READY) {
mutex_exit(&zonehash_lock);
return (set_errno(EINVAL));
}
zone_hold(zone);
mutex_exit(&zonehash_lock);
/*
* If not in the global zone, don't show information about other zones.
*/
if (!global && curproc->p_zone != zone) {
zone_rele(zone);
return (set_errno(EINVAL));
}
switch (attr) {
case ZONE_ATTR_ROOT:
if (global) {
/*
* Copy the path to trim the trailing "/" (except for
* the global zone).
*/
if (zone != global_zone)
size = zone->zone_rootpathlen - 1;
else
size = zone->zone_rootpathlen;
zonepath = kmem_alloc(size, KM_SLEEP);
bcopy(zone->zone_rootpath, zonepath, size);
zonepath[size - 1] = '\0';
} else {
/*
* Caller is not in the global zone, just return
* faked-up path for current zone.
*/
zonepath = "/";
size = 2;
}
if (bufsize > size)
bufsize = size;
if (buf != NULL) {
err = copyoutstr(zonepath, buf, bufsize, NULL);
if (err != 0 && err != ENAMETOOLONG)
error = EFAULT;
}
if (global)
kmem_free(zonepath, size);
break;
case ZONE_ATTR_NAME:
size = strlen(zone->zone_name) + 1;
if (bufsize > size)
bufsize = size;
if (buf != NULL) {
err = copyoutstr(zone->zone_name, buf, bufsize, NULL);
if (err != 0 && err != ENAMETOOLONG)
error = EFAULT;
}
break;
case ZONE_ATTR_STATUS:
/*
* Since we're not holding zonehash_lock, the zone status
* may be anything; leave it up to userland to sort it out.
*/
size = sizeof (zone_status);
if (bufsize > size)
bufsize = size;
zone_status = zone_status_get(zone);
if (buf != NULL &&
copyout(&zone_status, buf, bufsize) != 0)
error = EFAULT;
break;
case ZONE_ATTR_PRIVSET:
size = sizeof (priv_set_t);
if (bufsize > size)
bufsize = size;
if (buf != NULL &&
copyout(zone->zone_privset, buf, bufsize) != 0)
error = EFAULT;
break;
case ZONE_ATTR_UNIQID:
size = sizeof (zone->zone_uniqid);
if (bufsize > size)
bufsize = size;
if (buf != NULL &&
copyout(&zone->zone_uniqid, buf, bufsize) != 0)
error = EFAULT;
break;
case ZONE_ATTR_POOLID:
{
pool_t *pool;
poolid_t poolid;
if (pool_lock_intr() != 0) {
error = EINTR;
break;
}
pool = zone_pool_get(zone);
poolid = pool->pool_id;
pool_unlock();
size = sizeof (poolid);
if (bufsize > size)
bufsize = size;
if (buf != NULL && copyout(&poolid, buf, size) != 0)
error = EFAULT;
}
break;
case ZONE_ATTR_INITPID:
size = sizeof (initpid);
if (bufsize > size)
bufsize = size;
initpid = zone->zone_proc_initpid;
if (initpid == -1) {
error = ESRCH;
break;
}
if (buf != NULL &&
copyout(&initpid, buf, bufsize) != 0)
error = EFAULT;
break;
default:
error = EINVAL;
}
zone_rele(zone);
if (error)
return (set_errno(error));
return ((ssize_t)size);
}
/*
* Return zero if the process has at least one vnode mapped in to its
* address space which shouldn't be allowed to change zones.
*/
static int
as_can_change_zones(void)
{
proc_t *pp = curproc;
struct seg *seg;
struct as *as = pp->p_as;
vnode_t *vp;
int allow = 1;
ASSERT(pp->p_as != &kas);
AS_LOCK_ENTER(&as, &as->a_lock, RW_READER);
for (seg = AS_SEGFIRST(as); seg != NULL; seg = AS_SEGNEXT(as, seg)) {
/*
* if we can't get a backing vnode for this segment then skip
* it.
*/
vp = NULL;
if (SEGOP_GETVP(seg, seg->s_base, &vp) != 0 || vp == NULL)
continue;
if (!vn_can_change_zones(vp)) { /* bail on first match */
allow = 0;
break;
}
}
AS_LOCK_EXIT(&as, &as->a_lock);
return (allow);
}
/*
* Systemcall entry point for zone_enter().
*
* The current process is injected into said zone. In the process
* it will change its project membership, privileges, rootdir/cwd,
* zone-wide rctls, and pool association to match those of the zone.
*
* The first zone_enter() called while the zone is in the ZONE_IS_READY
* state will transition it to ZONE_IS_RUNNING. Processes may only
* enter a zone that is "ready" or "running".
*/
static int
zone_enter(zoneid_t zoneid)
{
zone_t *zone;
vnode_t *vp;
proc_t *pp = curproc;
contract_t *ct;
cont_process_t *ctp;
task_t *tk, *oldtk;
kproject_t *zone_proj0;
cred_t *cr, *newcr;
pool_t *oldpool, *newpool;
sess_t *sp;
uid_t uid;
zone_status_t status;
int err = 0;
rctl_entity_p_t e;
if (secpolicy_zone_config(CRED()) != 0)
return (set_errno(EPERM));
if (zoneid < MIN_USERZONEID || zoneid > MAX_ZONEID)
return (set_errno(EINVAL));
/*
* Stop all lwps so we don't need to hold a lock to look at
* curproc->p_zone. This needs to happen before we grab any
* locks to avoid deadlock (another lwp in the process could
* be waiting for the held lock).
*/
if (curthread != pp->p_agenttp && !holdlwps(SHOLDFORK))
return (set_errno(EINTR));
/*
* Make sure we're not changing zones with files open or mapped in
* to our address space which shouldn't be changing zones.
*/
if (!files_can_change_zones()) {
err = EBADF;
goto out;
}
if (!as_can_change_zones()) {
err = EFAULT;
goto out;
}
mutex_enter(&zonehash_lock);
if (pp->p_zone != global_zone) {
mutex_exit(&zonehash_lock);
err = EINVAL;
goto out;
}
zone = zone_find_all_by_id(zoneid);
if (zone == NULL) {
mutex_exit(&zonehash_lock);
err = EINVAL;
goto out;
}
/*
* To prevent processes in a zone from holding contracts on
* extrazonal resources, and to avoid process contract
* memberships which span zones, contract holders and processes
* which aren't the sole members of their encapsulating process
* contracts are not allowed to zone_enter.
*/
ctp = pp->p_ct_process;
ct = &ctp->conp_contract;
mutex_enter(&ct->ct_lock);
mutex_enter(&pp->p_lock);
if ((avl_numnodes(&pp->p_ct_held) != 0) || (ctp->conp_nmembers != 1)) {
mutex_exit(&pp->p_lock);
mutex_exit(&ct->ct_lock);
mutex_exit(&zonehash_lock);
pool_unlock();
err = EINVAL;
goto out;
}
/*
* Moreover, we don't allow processes whose encapsulating
* process contracts have inherited extrazonal contracts.
* While it would be easier to eliminate all process contracts
* with inherited contracts, we need to be able to give a
* restarted init (or other zone-penetrating process) its
* predecessor's contracts.
*/
if (ctp->conp_ninherited != 0) {
contract_t *next;
for (next = list_head(&ctp->conp_inherited); next;
next = list_next(&ctp->conp_inherited, next)) {
if (contract_getzuniqid(next) != zone->zone_uniqid) {
mutex_exit(&pp->p_lock);
mutex_exit(&ct->ct_lock);
mutex_exit(&zonehash_lock);
pool_unlock();
err = EINVAL;
goto out;
}
}
}
mutex_exit(&pp->p_lock);
mutex_exit(&ct->ct_lock);
status = zone_status_get(zone);
if (status < ZONE_IS_READY || status >= ZONE_IS_SHUTTING_DOWN) {
/*
* Can't join
*/
mutex_exit(&zonehash_lock);
err = EINVAL;
goto out;
}
/*
* Make sure new priv set is within the permitted set for caller
*/
if (!priv_issubset(zone->zone_privset, &CR_OPPRIV(CRED()))) {
mutex_exit(&zonehash_lock);
err = EPERM;
goto out;
}
/*
* We want to momentarily drop zonehash_lock while we optimistically
* bind curproc to the pool it should be running in. This is safe
* since the zone can't disappear (we have a hold on it).
*/
zone_hold(zone);
mutex_exit(&zonehash_lock);
/*
* Grab pool_lock to keep the pools configuration from changing
* and to stop ourselves from getting rebound to another pool
* until we join the zone.
*/
if (pool_lock_intr() != 0) {
zone_rele(zone);
err = EINTR;
goto out;
}
ASSERT(secpolicy_pool(CRED()) == 0);
/*
* Bind ourselves to the pool currently associated with the zone.
*/
oldpool = curproc->p_pool;
newpool = zone_pool_get(zone);
if (pool_state == POOL_ENABLED && newpool != oldpool &&
(err = pool_do_bind(newpool, P_PID, P_MYID,
POOL_BIND_ALL)) != 0) {
pool_unlock();
zone_rele(zone);
goto out;
}
/*
* Grab cpu_lock now; we'll need it later when we call
* task_join().
*/
mutex_enter(&cpu_lock);
mutex_enter(&zonehash_lock);
/*
* Make sure the zone hasn't moved on since we dropped zonehash_lock.
*/
if (zone_status_get(zone) >= ZONE_IS_SHUTTING_DOWN) {
/*
* Can't join anymore.
*/
mutex_exit(&zonehash_lock);
mutex_exit(&cpu_lock);
if (pool_state == POOL_ENABLED &&
newpool != oldpool)
(void) pool_do_bind(oldpool, P_PID, P_MYID,
POOL_BIND_ALL);
pool_unlock();
zone_rele(zone);
err = EINVAL;
goto out;
}
mutex_enter(&pp->p_lock);
zone_proj0 = zone->zone_zsched->p_task->tk_proj;
/* verify that we do not exceed and task or lwp limits */
mutex_enter(&zone->zone_nlwps_lock);
/* add new lwps to zone and zone's proj0 */
zone_proj0->kpj_nlwps += pp->p_lwpcnt;
zone->zone_nlwps += pp->p_lwpcnt;
/* add 1 task to zone's proj0 */
zone_proj0->kpj_ntasks += 1;
mutex_exit(&pp->p_lock);
mutex_exit(&zone->zone_nlwps_lock);
/* remove lwps from proc's old zone and old project */
mutex_enter(&pp->p_zone->zone_nlwps_lock);
pp->p_zone->zone_nlwps -= pp->p_lwpcnt;
pp->p_task->tk_proj->kpj_nlwps -= pp->p_lwpcnt;
mutex_exit(&pp->p_zone->zone_nlwps_lock);
/*
* Joining the zone cannot fail from now on.
*
* This means that a lot of the following code can be commonized and
* shared with zsched().
*/
/*
* Reset the encapsulating process contract's zone.
*/
ASSERT(ct->ct_mzuniqid == GLOBAL_ZONEUNIQID);
contract_setzuniqid(ct, zone->zone_uniqid);
/*
* Create a new task and associate the process with the project keyed
* by (projid,zoneid).
*
* We might as well be in project 0; the global zone's projid doesn't
* make much sense in a zone anyhow.
*
* This also increments zone_ntasks, and returns with p_lock held.
*/
tk = task_create(0, zone);
oldtk = task_join(tk, 0);
mutex_exit(&cpu_lock);
pp->p_flag |= SZONETOP;
pp->p_zone = zone;
/*
* call RCTLOP_SET functions on this proc
*/
e.rcep_p.zone = zone;
e.rcep_t = RCENTITY_ZONE;
(void) rctl_set_dup(NULL, NULL, pp, &e, zone->zone_rctls, NULL,
RCD_CALLBACK);
mutex_exit(&pp->p_lock);
/*
* We don't need to hold any of zsched's locks here; not only do we know
* the process and zone aren't going away, we know its session isn't
* changing either.
*
* By joining zsched's session here, we mimic the behavior in the
* global zone of init's sid being the pid of sched. We extend this
* to all zlogin-like zone_enter()'ing processes as well.
*/
mutex_enter(&pidlock);
sp = zone->zone_zsched->p_sessp;
SESS_HOLD(sp);
mutex_enter(&pp->p_lock);
pgexit(pp);
SESS_RELE(pp->p_sessp);
pp->p_sessp = sp;
pgjoin(pp, zone->zone_zsched->p_pidp);
mutex_exit(&pp->p_lock);
mutex_exit(&pidlock);
mutex_exit(&zonehash_lock);
/*
* We're firmly in the zone; let pools progress.
*/
pool_unlock();
task_rele(oldtk);
/*
* We don't need to retain a hold on the zone since we already
* incremented zone_ntasks, so the zone isn't going anywhere.
*/
zone_rele(zone);
/*
* Chroot
*/
vp = zone->zone_rootvp;
zone_chdir(vp, &PTOU(pp)->u_cdir, pp);
zone_chdir(vp, &PTOU(pp)->u_rdir, pp);
/*
* Change process credentials
*/
newcr = cralloc();
mutex_enter(&pp->p_crlock);
cr = pp->p_cred;
crcopy_to(cr, newcr);
crsetzone(newcr, zone);
pp->p_cred = newcr;
/*
* Restrict all process privilege sets to zone limit
*/
priv_intersect(zone->zone_privset, &CR_PPRIV(newcr));
priv_intersect(zone->zone_privset, &CR_EPRIV(newcr));
priv_intersect(zone->zone_privset, &CR_IPRIV(newcr));
priv_intersect(zone->zone_privset, &CR_LPRIV(newcr));
mutex_exit(&pp->p_crlock);
crset(pp, newcr);
/*
* Adjust upcount to reflect zone entry.
*/
uid = crgetruid(newcr);
mutex_enter(&pidlock);
upcount_dec(uid, GLOBAL_ZONEID);
upcount_inc(uid, zoneid);
mutex_exit(&pidlock);
/*
* Set up core file path and content.
*/
set_core_defaults();
out:
/*
* Let the other lwps continue.
*/
mutex_enter(&pp->p_lock);
if (curthread != pp->p_agenttp)
continuelwps(pp);
mutex_exit(&pp->p_lock);
return (err != 0 ? set_errno(err) : 0);
}
/*
* Systemcall entry point for zone_list(2).
*
* Processes running in a (non-global) zone only see themselves.
*/
static int
zone_list(zoneid_t *zoneidlist, uint_t *numzones)
{
zoneid_t *zoneids;
zone_t *zone;
uint_t user_nzones, real_nzones;
int error = 0;
uint_t i;
if (copyin(numzones, &user_nzones, sizeof (uint_t)) != 0)
return (set_errno(EFAULT));
if (curproc->p_zone != global_zone) {
/* just return current zone */
real_nzones = 1;
zoneids = kmem_alloc(sizeof (zoneid_t), KM_SLEEP);
zoneids[0] = curproc->p_zone->zone_id;
} else {
mutex_enter(&zonehash_lock);
real_nzones = zonecount;
if (real_nzones) {
zoneids = kmem_alloc(real_nzones * sizeof (zoneid_t),
KM_SLEEP);
i = 0;
for (zone = list_head(&zone_active); zone != NULL;
zone = list_next(&zone_active, zone))
zoneids[i++] = zone->zone_id;
ASSERT(i == real_nzones);
}
mutex_exit(&zonehash_lock);
}
if (user_nzones > real_nzones)
user_nzones = real_nzones;
if (copyout(&real_nzones, numzones, sizeof (uint_t)) != 0)
error = EFAULT;
else if (zoneidlist != NULL && user_nzones != 0) {
if (copyout(zoneids, zoneidlist,
user_nzones * sizeof (zoneid_t)) != 0)
error = EFAULT;
}
if (real_nzones)
kmem_free(zoneids, real_nzones * sizeof (zoneid_t));
if (error)
return (set_errno(error));
else
return (0);
}
/*
* Systemcall entry point for zone_lookup(2).
*
* Non-global zones are only able to see themselves.
*/
static zoneid_t
zone_lookup(const char *zone_name)
{
char *kname;
zone_t *zone;
zoneid_t zoneid;
int err;
if (zone_name == NULL) {
/* return caller's zone id */
return (getzoneid());
}
kname = kmem_zalloc(ZONENAME_MAX, KM_SLEEP);
if ((err = copyinstr(zone_name, kname, ZONENAME_MAX, NULL)) != 0) {
kmem_free(kname, ZONENAME_MAX);
return (set_errno(err));
}
mutex_enter(&zonehash_lock);
zone = zone_find_all_by_name(kname);
kmem_free(kname, ZONENAME_MAX);
if (zone == NULL || zone_status_get(zone) < ZONE_IS_READY ||
(curproc->p_zone != global_zone && curproc->p_zone != zone)) {
/* in non-global zone, can only lookup own name */
mutex_exit(&zonehash_lock);
return (set_errno(EINVAL));
}
zoneid = zone->zone_id;
mutex_exit(&zonehash_lock);
return (zoneid);
}
static int
zone_version(int *version_arg)
{
int version = ZONE_SYSCALL_API_VERSION;
if (copyout(&version, version_arg, sizeof (int)) != 0)
return (set_errno(EFAULT));
return (0);
}
/* ARGSUSED */
long
zone(int cmd, void *arg1, void *arg2, void *arg3, void *arg4)
{
zone_def zs;
switch (cmd) {
case ZONE_CREATE:
if (get_udatamodel() == DATAMODEL_NATIVE) {
if (copyin(arg1, &zs, sizeof (zone_def))) {
return (set_errno(EFAULT));
}
} else {
#ifdef _SYSCALL32_IMPL
zone_def32 zs32;
if (copyin(arg1, &zs32, sizeof (zone_def32))) {
return (set_errno(EFAULT));
}
zs.zone_name =
(const char *)(unsigned long)zs32.zone_name;
zs.zone_root =
(const char *)(unsigned long)zs32.zone_root;
zs.zone_privs =
(const struct priv_set *)
(unsigned long)zs32.zone_privs;
zs.rctlbuf = (caddr_t)(unsigned long)zs32.rctlbuf;
zs.rctlbufsz = zs32.rctlbufsz;
zs.zfsbuf = (caddr_t)(unsigned long)zs32.zfsbuf;
zs.zfsbufsz = zs32.zfsbufsz;
zs.extended_error =
(int *)(unsigned long)zs32.extended_error;
#else
panic("get_udatamodel() returned bogus result\n");
#endif
}
return (zone_create(zs.zone_name, zs.zone_root,
zs.zone_privs, zs.zone_privssz,
(caddr_t)zs.rctlbuf, zs.rctlbufsz,
(caddr_t)zs.zfsbuf, zs.zfsbufsz,
zs.extended_error));
case ZONE_BOOT:
return (zone_boot((zoneid_t)(uintptr_t)arg1,
(const char *)arg2));
case ZONE_DESTROY:
return (zone_destroy((zoneid_t)(uintptr_t)arg1));
case ZONE_GETATTR:
return (zone_getattr((zoneid_t)(uintptr_t)arg1,
(int)(uintptr_t)arg2, arg3, (size_t)arg4));
case ZONE_ENTER:
return (zone_enter((zoneid_t)(uintptr_t)arg1));
case ZONE_LIST:
return (zone_list((zoneid_t *)arg1, (uint_t *)arg2));
case ZONE_SHUTDOWN:
return (zone_shutdown((zoneid_t)(uintptr_t)arg1));
case ZONE_LOOKUP:
return (zone_lookup((const char *)arg1));
case ZONE_VERSION:
return (zone_version((int *)arg1));
default:
return (set_errno(EINVAL));
}
}
struct zarg {
zone_t *zone;
zone_cmd_arg_t arg;
};
static int
zone_lookup_door(const char *zone_name, door_handle_t *doorp)
{
char *buf;
size_t buflen;
int error;
buflen = sizeof (ZONE_DOOR_PATH) + strlen(zone_name);
buf = kmem_alloc(buflen, KM_SLEEP);
(void) snprintf(buf, buflen, ZONE_DOOR_PATH, zone_name);
error = door_ki_open(buf, doorp);
kmem_free(buf, buflen);
return (error);
}
static void
zone_release_door(door_handle_t *doorp)
{
door_ki_rele(*doorp);
*doorp = NULL;
}
static void
zone_ki_call_zoneadmd(struct zarg *zargp)
{
door_handle_t door = NULL;
door_arg_t darg, save_arg;
char *zone_name;
size_t zone_namelen;
zoneid_t zoneid;
zone_t *zone;
zone_cmd_arg_t arg;
uint64_t uniqid;
size_t size;
int error;
int retry;
zone = zargp->zone;
arg = zargp->arg;
kmem_free(zargp, sizeof (*zargp));
zone_namelen = strlen(zone->zone_name) + 1;
zone_name = kmem_alloc(zone_namelen, KM_SLEEP);
bcopy(zone->zone_name, zone_name, zone_namelen);
zoneid = zone->zone_id;
uniqid = zone->zone_uniqid;
/*
* zoneadmd may be down, but at least we can empty out the zone.
* We can ignore the return value of zone_empty() since we're called
* from a kernel thread and know we won't be delivered any signals.
*/
ASSERT(curproc == &p0);
(void) zone_empty(zone);
ASSERT(zone_status_get(zone) >= ZONE_IS_EMPTY);
zone_rele(zone);
size = sizeof (arg);
darg.rbuf = (char *)&arg;
darg.data_ptr = (char *)&arg;
darg.rsize = size;
darg.data_size = size;
darg.desc_ptr = NULL;
darg.desc_num = 0;
save_arg = darg;
/*
* Since we're not holding a reference to the zone, any number of
* things can go wrong, including the zone disappearing before we get a
* chance to talk to zoneadmd.
*/
for (retry = 0; /* forever */; retry++) {
if (door == NULL &&
(error = zone_lookup_door(zone_name, &door)) != 0) {
goto next;
}
ASSERT(door != NULL);
if ((error = door_ki_upcall(door, &darg)) == 0) {
break;
}
switch (error) {
case EINTR:
/* FALLTHROUGH */
case EAGAIN: /* process may be forking */
/*
* Back off for a bit
*/
break;
case EBADF:
zone_release_door(&door);
if (zone_lookup_door(zone_name, &door) != 0) {
/*
* zoneadmd may be dead, but it may come back to
* life later.
*/
break;
}
break;
default:
cmn_err(CE_WARN,
"zone_ki_call_zoneadmd: door_ki_upcall error %d\n",
error);
goto out;
}
next:
/*
* If this isn't the same zone_t that we originally had in mind,
* then this is the same as if two kadmin requests come in at
* the same time: the first one wins. This means we lose, so we
* bail.
*/
if ((zone = zone_find_by_id(zoneid)) == NULL) {
/*
* Problem is solved.
*/
break;
}
if (zone->zone_uniqid != uniqid) {
/*
* zoneid recycled
*/
zone_rele(zone);
break;
}
/*
* We could zone_status_timedwait(), but there doesn't seem to
* be much point in doing that (plus, it would mean that
* zone_free() isn't called until this thread exits).
*/
zone_rele(zone);
delay(hz);
darg = save_arg;
}
out:
if (door != NULL) {
zone_release_door(&door);
}
kmem_free(zone_name, zone_namelen);
thread_exit();
}
/*
* Entry point for uadmin() to tell the zone to go away or reboot. The caller
* is a process in the zone to be modified.
*
* In order to shutdown the zone, we will hand off control to zoneadmd
* (running in the global zone) via a door. We do a half-hearted job at
* killing all processes in the zone, create a kernel thread to contact
* zoneadmd, and make note of the "uniqid" of the zone. The uniqid is
* a form of generation number used to let zoneadmd (as well as
* zone_destroy()) know exactly which zone they're re talking about.
*/
int
zone_uadmin(int cmd, int fcn, cred_t *credp)
{
struct zarg *zargp;
zone_cmd_t zcmd;
zone_t *zone;
zone = curproc->p_zone;
ASSERT(getzoneid() != GLOBAL_ZONEID);
switch (cmd) {
case A_SHUTDOWN:
switch (fcn) {
case AD_HALT:
case AD_POWEROFF:
zcmd = Z_HALT;
break;
case AD_BOOT:
zcmd = Z_REBOOT;
break;
case AD_IBOOT:
case AD_SBOOT:
case AD_SIBOOT:
case AD_NOSYNC:
return (ENOTSUP);
default:
return (EINVAL);
}
break;
case A_REBOOT:
zcmd = Z_REBOOT;
break;
case A_FTRACE:
case A_REMOUNT:
case A_FREEZE:
case A_DUMP:
return (ENOTSUP);
default:
ASSERT(cmd != A_SWAPCTL); /* handled by uadmin() */
return (EINVAL);
}
if (secpolicy_zone_admin(credp, B_FALSE))
return (EPERM);
mutex_enter(&zone_status_lock);
/*
* zone_status can't be ZONE_IS_EMPTY or higher since curproc
* is in the zone.
*/
ASSERT(zone_status_get(zone) < ZONE_IS_EMPTY);
if (zone_status_get(zone) > ZONE_IS_RUNNING) {
/*
* This zone is already on its way down.
*/
mutex_exit(&zone_status_lock);
return (0);
}
/*
* Prevent future zone_enter()s
*/
zone_status_set(zone, ZONE_IS_SHUTTING_DOWN);
mutex_exit(&zone_status_lock);
/*
* Kill everyone now and call zoneadmd later.
* zone_ki_call_zoneadmd() will do a more thorough job of this
* later.
*/
killall(zone->zone_id);
/*
* Now, create the thread to contact zoneadmd and do the rest of the
* work. This thread can't be created in our zone otherwise
* zone_destroy() would deadlock.
*/
zargp = kmem_alloc(sizeof (*zargp), KM_SLEEP);
zargp->arg.cmd = zcmd;
zargp->arg.uniqid = zone->zone_uniqid;
(void) strcpy(zargp->arg.locale, "C");
zone_hold(zargp->zone = zone);
(void) thread_create(NULL, 0, zone_ki_call_zoneadmd, zargp, 0, &p0,
TS_RUN, minclsyspri);
exit(CLD_EXITED, 0);
return (EINVAL);
}
/*
* Entry point so kadmin(A_SHUTDOWN, ...) can set the global zone's
* status to ZONE_IS_SHUTTING_DOWN.
*/
void
zone_shutdown_global(void)
{
ASSERT(curproc->p_zone == global_zone);
mutex_enter(&zone_status_lock);
ASSERT(zone_status_get(global_zone) == ZONE_IS_RUNNING);
zone_status_set(global_zone, ZONE_IS_SHUTTING_DOWN);
mutex_exit(&zone_status_lock);
}
/*
* Returns true if the named dataset is visible in the current zone.
* The 'write' parameter is set to 1 if the dataset is also writable.
*/
int
zone_dataset_visible(const char *dataset, int *write)
{
zone_dataset_t *zd;
size_t len;
zone_t *zone = curproc->p_zone;
if (dataset[0] == '\0')
return (0);
/*
* Walk the list once, looking for datasets which match exactly, or
* specify a dataset underneath an exported dataset. If found, return
* true and note that it is writable.
*/
for (zd = list_head(&zone->zone_datasets); zd != NULL;
zd = list_next(&zone->zone_datasets, zd)) {
len = strlen(zd->zd_dataset);
if (strlen(dataset) >= len &&
bcmp(dataset, zd->zd_dataset, len) == 0 &&
(dataset[len] == '\0' || dataset[len] == '/' ||
dataset[len] == '@')) {
if (write)
*write = 1;
return (1);
}
}
/*
* Walk the list a second time, searching for datasets which are parents
* of exported datasets. These should be visible, but read-only.
*
* Note that we also have to support forms such as 'pool/dataset/', with
* a trailing slash.
*/
for (zd = list_head(&zone->zone_datasets); zd != NULL;
zd = list_next(&zone->zone_datasets, zd)) {
len = strlen(dataset);
if (dataset[len - 1] == '/')
len--; /* Ignore trailing slash */
if (len < strlen(zd->zd_dataset) &&
bcmp(dataset, zd->zd_dataset, len) == 0 &&
zd->zd_dataset[len] == '/') {
if (write)
*write = 0;
return (1);
}
}
return (0);
}