[mirror_zfs.git] / module / zfs / spa_config.c

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (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 2008 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 */

#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/nvpair.h>
#include <sys/uio.h>
#include <sys/fs/zfs.h>
#include <sys/vdev_impl.h>
#include <sys/zfs_ioctl.h>
#include <sys/utsname.h>
#include <sys/systeminfo.h>
#include <sys/sunddi.h>
#ifdef _KERNEL
#include <sys/kobj.h>
#endif

/*
 * Pool configuration repository.
 *
 * Pool configuration is stored as a packed nvlist on the filesystem.  By
 * default, all pools are stored in /etc/zfs/zpool.cache and loaded on boot
 * (when the ZFS module is loaded).  Pools can also have the 'cachefile'
 * property set that allows them to be stored in an alternate location until
 * the control of external software.
 *
 * For each cache file, we have a single nvlist which holds all the
 * configuration information.  When the module loads, we read this information
 * from /etc/zfs/zpool.cache and populate the SPA namespace.  This namespace is
 * maintained independently in spa.c.  Whenever the namespace is modified, or
 * the configuration of a pool is changed, we call spa_config_sync(), which
 * walks through all the active pools and writes the configuration to disk.
 */

static uint64_t spa_config_generation = 1;

/*
 * This can be overridden in userland to preserve an alternate namespace for
 * userland pools when doing testing.
 */
const char *spa_config_path = ZPOOL_CACHE;

/*
 * Called when the module is first loaded, this routine loads the configuration
 * file into the SPA namespace.  It does not actually open or load the pools; it
 * only populates the namespace.
 */
void
spa_config_load(void)
{
	void *buf = NULL;
	nvlist_t *nvlist, *child;
	nvpair_t *nvpair;
	spa_t *spa;
	char *pathname;
	struct _buf *file;
	uint64_t fsize;

	/*
	 * Open the configuration file.
	 */
	pathname = kmem_alloc(MAXPATHLEN, KM_SLEEP);

	(void) snprintf(pathname, MAXPATHLEN, "%s%s",
	    (rootdir != NULL) ? "./" : "", spa_config_path);

	file = kobj_open_file(pathname);

	kmem_free(pathname, MAXPATHLEN);

	if (file == (struct _buf *)-1)
		return;

	if (kobj_get_filesize(file, &fsize) != 0)
		goto out;

	buf = kmem_alloc(fsize, KM_SLEEP);

	/*
	 * Read the nvlist from the file.
	 */
	if (kobj_read_file(file, buf, fsize, 0) < 0)
		goto out;

	/*
	 * Unpack the nvlist.
	 */
	if (nvlist_unpack(buf, fsize, &nvlist, KM_SLEEP) != 0)
		goto out;

	/*
	 * Iterate over all elements in the nvlist, creating a new spa_t for
	 * each one with the specified configuration.
	 */
	mutex_enter(&spa_namespace_lock);
	nvpair = NULL;
	while ((nvpair = nvlist_next_nvpair(nvlist, nvpair)) != NULL) {

		if (nvpair_type(nvpair) != DATA_TYPE_NVLIST)
			continue;

		VERIFY(nvpair_value_nvlist(nvpair, &child) == 0);

		if (spa_lookup(nvpair_name(nvpair)) != NULL)
			continue;
		spa = spa_add(nvpair_name(nvpair), NULL);

		/*
		 * We blindly duplicate the configuration here.  If it's
		 * invalid, we will catch it when the pool is first opened.
		 */
		VERIFY(nvlist_dup(child, &spa->spa_config, 0) == 0);
	}
	mutex_exit(&spa_namespace_lock);

	nvlist_free(nvlist);

out:
	if (buf != NULL)
		kmem_free(buf, fsize);

	kobj_close_file(file);
}

static void
spa_config_write(spa_config_dirent_t *dp, nvlist_t *nvl)
{
	size_t buflen;
	char *buf;
	vnode_t *vp;
	int oflags = FWRITE | FTRUNC | FCREAT | FOFFMAX;
	char *temp;

	/*
	 * If the nvlist is empty (NULL), then remove the old cachefile.
	 */
	if (nvl == NULL) {
		(void) vn_remove(dp->scd_path, UIO_SYSSPACE, RMFILE);
		return;
	}

	/*
	 * Pack the configuration into a buffer.
	 */
	VERIFY(nvlist_size(nvl, &buflen, NV_ENCODE_XDR) == 0);

	buf = kmem_alloc(buflen, KM_SLEEP);
	temp = kmem_zalloc(MAXPATHLEN, KM_SLEEP);

	VERIFY(nvlist_pack(nvl, &buf, &buflen, NV_ENCODE_XDR,
	    KM_SLEEP) == 0);

	/*
	 * Write the configuration to disk.  We need to do the traditional
	 * 'write to temporary file, sync, move over original' to make sure we
	 * always have a consistent view of the data.
	 */
	(void) snprintf(temp, MAXPATHLEN, "%s.tmp", dp->scd_path);

	if (vn_open(temp, UIO_SYSSPACE, oflags, 0644, &vp, CRCREAT, 0) == 0) {
		if (vn_rdwr(UIO_WRITE, vp, buf, buflen, 0, UIO_SYSSPACE,
		    0, RLIM64_INFINITY, kcred, NULL) == 0 &&
		    VOP_FSYNC(vp, FSYNC, kcred, NULL) == 0) {
			(void) vn_rename(temp, dp->scd_path, UIO_SYSSPACE);
		}
		(void) VOP_CLOSE(vp, oflags, 1, 0, kcred, NULL);
		VN_RELE(vp);
	}

	(void) vn_remove(temp, UIO_SYSSPACE, RMFILE);

	kmem_free(buf, buflen);
	kmem_free(temp, MAXPATHLEN);
}

/*
 * Synchronize pool configuration to disk.  This must be called with the
 * namespace lock held.
 */
void
spa_config_sync(spa_t *target, boolean_t removing, boolean_t postsysevent)
{
	spa_config_dirent_t *dp, *tdp;
	nvlist_t *nvl;

	ASSERT(MUTEX_HELD(&spa_namespace_lock));

	if (rootdir == NULL)
		return;

	/*
	 * Iterate over all cachefiles for the pool, past or present.  When the
	 * cachefile is changed, the new one is pushed onto this list, allowing
	 * us to update previous cachefiles that no longer contain this pool.
	 */
	for (dp = list_head(&target->spa_config_list); dp != NULL;
	    dp = list_next(&target->spa_config_list, dp)) {
		spa_t *spa = NULL;
		if (dp->scd_path == NULL)
			continue;

		/*
		 * Iterate over all pools, adding any matching pools to 'nvl'.
		 */
		nvl = NULL;
		while ((spa = spa_next(spa)) != NULL) {
			if (spa == target && removing)
				continue;

			mutex_enter(&spa->spa_props_lock);
			tdp = list_head(&spa->spa_config_list);
			if (spa->spa_config == NULL ||
			    tdp->scd_path == NULL ||
			    strcmp(tdp->scd_path, dp->scd_path) != 0) {
				mutex_exit(&spa->spa_props_lock);
				continue;
			}

			if (nvl == NULL)
				VERIFY(nvlist_alloc(&nvl, NV_UNIQUE_NAME,
				    KM_SLEEP) == 0);

			VERIFY(nvlist_add_nvlist(nvl, spa->spa_name,
			    spa->spa_config) == 0);
			mutex_exit(&spa->spa_props_lock);
		}

		spa_config_write(dp, nvl);
		nvlist_free(nvl);
	}

	/*
	 * Remove any config entries older than the current one.
	 */
	dp = list_head(&target->spa_config_list);
	while ((tdp = list_next(&target->spa_config_list, dp)) != NULL) {
		list_remove(&target->spa_config_list, tdp);
		if (tdp->scd_path != NULL)
			spa_strfree(tdp->scd_path);
		kmem_free(tdp, sizeof (spa_config_dirent_t));
	}

	spa_config_generation++;

	if (postsysevent)
		spa_event_notify(target, NULL, ESC_ZFS_CONFIG_SYNC);
}

/*
 * Sigh.  Inside a local zone, we don't have access to /etc/zfs/zpool.cache,
 * and we don't want to allow the local zone to see all the pools anyway.
 * So we have to invent the ZFS_IOC_CONFIG ioctl to grab the configuration
 * information for all pool visible within the zone.
 */
nvlist_t *
spa_all_configs(uint64_t *generation)
{
	nvlist_t *pools;
	spa_t *spa = NULL;

	if (*generation == spa_config_generation)
		return (NULL);

	VERIFY(nvlist_alloc(&pools, NV_UNIQUE_NAME, KM_SLEEP) == 0);

	mutex_enter(&spa_namespace_lock);
	while ((spa = spa_next(spa)) != NULL) {
		if (INGLOBALZONE(curproc) ||
		    zone_dataset_visible(spa_name(spa), NULL)) {
			mutex_enter(&spa->spa_props_lock);
			VERIFY(nvlist_add_nvlist(pools, spa_name(spa),
			    spa->spa_config) == 0);
			mutex_exit(&spa->spa_props_lock);
		}
	}
	*generation = spa_config_generation;
	mutex_exit(&spa_namespace_lock);

	return (pools);
}

void
spa_config_set(spa_t *spa, nvlist_t *config)
{
	mutex_enter(&spa->spa_props_lock);
	if (spa->spa_config != NULL)
		nvlist_free(spa->spa_config);
	spa->spa_config = config;
	mutex_exit(&spa->spa_props_lock);
}

/*
 * Generate the pool's configuration based on the current in-core state.
 * We infer whether to generate a complete config or just one top-level config
 * based on whether vd is the root vdev.
 */
nvlist_t *
spa_config_generate(spa_t *spa, vdev_t *vd, uint64_t txg, int getstats)
{
	nvlist_t *config, *nvroot;
	vdev_t *rvd = spa->spa_root_vdev;
	unsigned long hostid = 0;
	boolean_t locked = B_FALSE;

	if (vd == NULL) {
		vd = rvd;
		locked = B_TRUE;
		spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER);
	}

	ASSERT(spa_config_held(spa, SCL_CONFIG | SCL_STATE, RW_READER) ==
	    (SCL_CONFIG | SCL_STATE));

	/*
	 * If txg is -1, report the current value of spa->spa_config_txg.
	 */
	if (txg == -1ULL)
		txg = spa->spa_config_txg;

	VERIFY(nvlist_alloc(&config, NV_UNIQUE_NAME, KM_SLEEP) == 0);

	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_VERSION,
	    spa_version(spa)) == 0);
	VERIFY(nvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME,
	    spa_name(spa)) == 0);
	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE,
	    spa_state(spa)) == 0);
	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_TXG,
	    txg) == 0);
	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_GUID,
	    spa_guid(spa)) == 0);
	(void) ddi_strtoul(hw_serial, NULL, 10, &hostid);
	if (hostid != 0) {
		VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_HOSTID,
		    hostid) == 0);
	}
	VERIFY(nvlist_add_string(config, ZPOOL_CONFIG_HOSTNAME,
	    utsname.nodename) == 0);

	if (vd != rvd) {
		VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_TOP_GUID,
		    vd->vdev_top->vdev_guid) == 0);
		VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_GUID,
		    vd->vdev_guid) == 0);
		if (vd->vdev_isspare)
			VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_IS_SPARE,
			    1ULL) == 0);
		if (vd->vdev_islog)
			VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_IS_LOG,
			    1ULL) == 0);
		vd = vd->vdev_top;		/* label contains top config */
	}

	nvroot = vdev_config_generate(spa, vd, getstats, B_FALSE, B_FALSE);
	VERIFY(nvlist_add_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, nvroot) == 0);
	nvlist_free(nvroot);

	if (locked)
		spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG);

	return (config);
}

/*
 * For a pool that's not currently a booting rootpool, update all disk labels,
 * generate a fresh config based on the current in-core state, and sync the
 * global config cache.
 */
void
spa_config_update(spa_t *spa, int what)
{
	spa_config_update_common(spa, what, FALSE);
}

/*
 * Update all disk labels, generate a fresh config based on the current
 * in-core state, and sync the global config cache (do not sync the config
 * cache if this is a booting rootpool).
 */
void
spa_config_update_common(spa_t *spa, int what, boolean_t isroot)
{
	vdev_t *rvd = spa->spa_root_vdev;
	uint64_t txg;
	int c;

	ASSERT(MUTEX_HELD(&spa_namespace_lock));

	spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
	txg = spa_last_synced_txg(spa) + 1;
	if (what == SPA_CONFIG_UPDATE_POOL) {
		vdev_config_dirty(rvd);
	} else {
		/*
		 * If we have top-level vdevs that were added but have
		 * not yet been prepared for allocation, do that now.
		 * (It's safe now because the config cache is up to date,
		 * so it will be able to translate the new DVAs.)
		 * See comments in spa_vdev_add() for full details.
		 */
		for (c = 0; c < rvd->vdev_children; c++) {
			vdev_t *tvd = rvd->vdev_child[c];
			if (tvd->vdev_ms_array == 0) {
				vdev_init(tvd, txg);
				vdev_config_dirty(tvd);
			}
		}
	}
	spa_config_exit(spa, SCL_ALL, FTAG);

	/*
	 * Wait for the mosconfig to be regenerated and synced.
	 */
	txg_wait_synced(spa->spa_dsl_pool, txg);

	/*
	 * Update the global config cache to reflect the new mosconfig.
	 */
	if (!isroot)
		spa_config_sync(spa, B_FALSE, what != SPA_CONFIG_UPDATE_POOL);

	if (what == SPA_CONFIG_UPDATE_POOL)
		spa_config_update_common(spa, SPA_CONFIG_UPDATE_VDEVS, isroot);
}
Commit	Line	Data
34dc7c2f BB	1	/*
	2	* CDDL HEADER START
	3	*
	4	* The contents of this file are subject to the terms of the
	5	* Common Development and Distribution License (the "License").
	6	* You may not use this file except in compliance with the License.
	7	*
	8	* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
	9	* or http://www.opensolaris.org/os/licensing.
	10	* See the License for the specific language governing permissions
	11	* and limitations under the License.
	12	*
	13	* When distributing Covered Code, include this CDDL HEADER in each
	14	* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
	15	* If applicable, add the following below this CDDL HEADER, with the
	16	* fields enclosed by brackets "[]" replaced with your own identifying
	17	* information: Portions Copyright [yyyy] [name of copyright owner]
	18	*
	19	* CDDL HEADER END
	20	*/
	21
	22	/*
	23	* Copyright 2008 Sun Microsystems, Inc. All rights reserved.
	24	* Use is subject to license terms.
	25	*/
	26
34dc7c2f BB	27	#include <sys/spa.h>
	28	#include <sys/spa_impl.h>
	29	#include <sys/nvpair.h>
	30	#include <sys/uio.h>
	31	#include <sys/fs/zfs.h>
	32	#include <sys/vdev_impl.h>
	33	#include <sys/zfs_ioctl.h>
	34	#include <sys/utsname.h>
	35	#include <sys/systeminfo.h>
	36	#include <sys/sunddi.h>
	37	#ifdef _KERNEL
	38	#include <sys/kobj.h>
	39	#endif
	40
	41	/*
	42	* Pool configuration repository.
	43	*
	44	* Pool configuration is stored as a packed nvlist on the filesystem. By
	45	* default, all pools are stored in /etc/zfs/zpool.cache and loaded on boot
	46	* (when the ZFS module is loaded). Pools can also have the 'cachefile'
	47	* property set that allows them to be stored in an alternate location until
	48	* the control of external software.
	49	*
	50	* For each cache file, we have a single nvlist which holds all the
	51	* configuration information. When the module loads, we read this information
	52	* from /etc/zfs/zpool.cache and populate the SPA namespace. This namespace is
	53	* maintained independently in spa.c. Whenever the namespace is modified, or
	54	* the configuration of a pool is changed, we call spa_config_sync(), which
	55	* walks through all the active pools and writes the configuration to disk.
	56	*/
	57
	58	static uint64_t spa_config_generation = 1;
	59
	60	/*
	61	* This can be overridden in userland to preserve an alternate namespace for
	62	* userland pools when doing testing.
	63	*/
b128c09f	64	const char *spa_config_path = ZPOOL_CACHE;
34dc7c2f BB	65
	66	/*
	67	* Called when the module is first loaded, this routine loads the configuration
	68	* file into the SPA namespace. It does not actually open or load the pools; it
	69	* only populates the namespace.
	70	*/
	71	void
	72	spa_config_load(void)
	73	{
	74	void *buf = NULL;
	75	nvlist_t nvlist, child;
	76	nvpair_t *nvpair;
	77	spa_t *spa;
b128c09f	78	char *pathname;
34dc7c2f BB	79	struct _buf *file;
	80	uint64_t fsize;
	81
	82	/*
	83	* Open the configuration file.
	84	*/
b128c09f BB	85	pathname = kmem_alloc(MAXPATHLEN, KM_SLEEP);
	86
	87	(void) snprintf(pathname, MAXPATHLEN, "%s%s",
	88	(rootdir != NULL) ? "./" : "", spa_config_path);
34dc7c2f BB	89
34dc7c2f BB	90	file = kobj_open_file(pathname);
b128c09f BB	91
	92	kmem_free(pathname, MAXPATHLEN);
	93
34dc7c2f BB	94	if (file == (struct _buf *)-1)
	95	return;
	96
	97	if (kobj_get_filesize(file, &fsize) != 0)
	98	goto out;
	99
	100	buf = kmem_alloc(fsize, KM_SLEEP);
	101
	102	/*
	103	* Read the nvlist from the file.
	104	*/
	105	if (kobj_read_file(file, buf, fsize, 0) < 0)
	106	goto out;
	107
	108	/*
	109	* Unpack the nvlist.
	110	*/
	111	if (nvlist_unpack(buf, fsize, &nvlist, KM_SLEEP) != 0)
	112	goto out;
	113
	114	/*
	115	* Iterate over all elements in the nvlist, creating a new spa_t for
	116	* each one with the specified configuration.
	117	*/
	118	mutex_enter(&spa_namespace_lock);
	119	nvpair = NULL;
	120	while ((nvpair = nvlist_next_nvpair(nvlist, nvpair)) != NULL) {
	121
	122	if (nvpair_type(nvpair) != DATA_TYPE_NVLIST)
	123	continue;
	124
	125	VERIFY(nvpair_value_nvlist(nvpair, &child) == 0);
	126
	127	if (spa_lookup(nvpair_name(nvpair)) != NULL)
	128	continue;
	129	spa = spa_add(nvpair_name(nvpair), NULL);
	130
	131	/*
	132	* We blindly duplicate the configuration here. If it's
	133	* invalid, we will catch it when the pool is first opened.
	134	*/
	135	VERIFY(nvlist_dup(child, &spa->spa_config, 0) == 0);
	136	}
	137	mutex_exit(&spa_namespace_lock);
	138
	139	nvlist_free(nvlist);
	140
	141	out:
	142	if (buf != NULL)
	143	kmem_free(buf, fsize);
	144
	145	kobj_close_file(file);
	146	}
	147
34dc7c2f	148	static void
b128c09f	149	spa_config_write(spa_config_dirent_t dp, nvlist_t nvl)
34dc7c2f	150	{
34dc7c2f BB	151	size_t buflen;
	152	char *buf;
	153	vnode_t *vp;
	154	int oflags = FWRITE \| FTRUNC \| FCREAT \| FOFFMAX;
b128c09f BB	155	char *temp;
	156
	157	/*
	158	* If the nvlist is empty (NULL), then remove the old cachefile.
	159	*/
	160	if (nvl == NULL) {
	161	(void) vn_remove(dp->scd_path, UIO_SYSSPACE, RMFILE);
	162	return;
	163	}
34dc7c2f BB	164
	165	/*
	166	* Pack the configuration into a buffer.
	167	*/
b128c09f	168	VERIFY(nvlist_size(nvl, &buflen, NV_ENCODE_XDR) == 0);
34dc7c2f BB	169
34dc7c2f BB	170	buf = kmem_alloc(buflen, KM_SLEEP);
b128c09f	171	temp = kmem_zalloc(MAXPATHLEN, KM_SLEEP);
34dc7c2f	172
b128c09f	173	VERIFY(nvlist_pack(nvl, &buf, &buflen, NV_ENCODE_XDR,
34dc7c2f BB	174	KM_SLEEP) == 0);
	175
	176	/*
	177	* Write the configuration to disk. We need to do the traditional
	178	* 'write to temporary file, sync, move over original' to make sure we
	179	* always have a consistent view of the data.
	180	*/
b128c09f	181	(void) snprintf(temp, MAXPATHLEN, "%s.tmp", dp->scd_path);
34dc7c2f	182
b128c09f BB	183	if (vn_open(temp, UIO_SYSSPACE, oflags, 0644, &vp, CRCREAT, 0) == 0) {
	184	if (vn_rdwr(UIO_WRITE, vp, buf, buflen, 0, UIO_SYSSPACE,
	185	0, RLIM64_INFINITY, kcred, NULL) == 0 &&
	186	VOP_FSYNC(vp, FSYNC, kcred, NULL) == 0) {
	187	(void) vn_rename(temp, dp->scd_path, UIO_SYSSPACE);
	188	}
	189	(void) VOP_CLOSE(vp, oflags, 1, 0, kcred, NULL);
	190	VN_RELE(vp);
34dc7c2f BB	191	}
34dc7c2f BB	192
b128c09f	193	(void) vn_remove(temp, UIO_SYSSPACE, RMFILE);
34dc7c2f	194
34dc7c2f	195	kmem_free(buf, buflen);
b128c09f	196	kmem_free(temp, MAXPATHLEN);
34dc7c2f BB	197	}
	198
	199	/*
b128c09f BB	200	* Synchronize pool configuration to disk. This must be called with the
b128c09f BB	201	* namespace lock held.
34dc7c2f BB	202	*/
34dc7c2f BB	203	void
b128c09f	204	spa_config_sync(spa_t *target, boolean_t removing, boolean_t postsysevent)
34dc7c2f	205	{
b128c09f BB	206	spa_config_dirent_t dp, tdp;
b128c09f BB	207	nvlist_t *nvl;
34dc7c2f BB	208
	209	ASSERT(MUTEX_HELD(&spa_namespace_lock));
	210
fb5f0bc8 BB	211	if (rootdir == NULL)
	212	return;
	213
b128c09f BB	214	/*
	215	* Iterate over all cachefiles for the pool, past or present. When the
	216	* cachefile is changed, the new one is pushed onto this list, allowing
	217	* us to update previous cachefiles that no longer contain this pool.
	218	*/
	219	for (dp = list_head(&target->spa_config_list); dp != NULL;
	220	dp = list_next(&target->spa_config_list, dp)) {
	221	spa_t *spa = NULL;
	222	if (dp->scd_path == NULL)
	223	continue;
	224
	225	/*
	226	* Iterate over all pools, adding any matching pools to 'nvl'.
	227	*/
	228	nvl = NULL;
	229	while ((spa = spa_next(spa)) != NULL) {
	230	if (spa == target && removing)
	231	continue;
	232
	233	mutex_enter(&spa->spa_props_lock);
	234	tdp = list_head(&spa->spa_config_list);
	235	if (spa->spa_config == NULL \|\|
	236	tdp->scd_path == NULL \|\|
	237	strcmp(tdp->scd_path, dp->scd_path) != 0) {
	238	mutex_exit(&spa->spa_props_lock);
	239	continue;
	240	}
	241
	242	if (nvl == NULL)
	243	VERIFY(nvlist_alloc(&nvl, NV_UNIQUE_NAME,
	244	KM_SLEEP) == 0);
	245
	246	VERIFY(nvlist_add_nvlist(nvl, spa->spa_name,
	247	spa->spa_config) == 0);
	248	mutex_exit(&spa->spa_props_lock);
	249	}
	250
	251	spa_config_write(dp, nvl);
	252	nvlist_free(nvl);
	253	}
34dc7c2f BB	254
34dc7c2f BB	255	/*
b128c09f	256	* Remove any config entries older than the current one.
34dc7c2f	257	*/
b128c09f BB	258	dp = list_head(&target->spa_config_list);
	259	while ((tdp = list_next(&target->spa_config_list, dp)) != NULL) {
	260	list_remove(&target->spa_config_list, tdp);
	261	if (tdp->scd_path != NULL)
	262	spa_strfree(tdp->scd_path);
	263	kmem_free(tdp, sizeof (spa_config_dirent_t));
34dc7c2f BB	264	}
	265
	266	spa_config_generation++;
b128c09f BB	267
	268	if (postsysevent)
	269	spa_event_notify(target, NULL, ESC_ZFS_CONFIG_SYNC);
34dc7c2f BB	270	}
	271
	272	/*
	273	* Sigh. Inside a local zone, we don't have access to /etc/zfs/zpool.cache,
	274	* and we don't want to allow the local zone to see all the pools anyway.
	275	* So we have to invent the ZFS_IOC_CONFIG ioctl to grab the configuration
	276	* information for all pool visible within the zone.
	277	*/
	278	nvlist_t *
	279	spa_all_configs(uint64_t *generation)
	280	{
	281	nvlist_t *pools;
b128c09f	282	spa_t *spa = NULL;
34dc7c2f BB	283
	284	if (*generation == spa_config_generation)
	285	return (NULL);
	286
	287	VERIFY(nvlist_alloc(&pools, NV_UNIQUE_NAME, KM_SLEEP) == 0);
	288
34dc7c2f BB	289	mutex_enter(&spa_namespace_lock);
	290	while ((spa = spa_next(spa)) != NULL) {
	291	if (INGLOBALZONE(curproc) \|\|
	292	zone_dataset_visible(spa_name(spa), NULL)) {
b128c09f	293	mutex_enter(&spa->spa_props_lock);
34dc7c2f BB	294	VERIFY(nvlist_add_nvlist(pools, spa_name(spa),
34dc7c2f BB	295	spa->spa_config) == 0);
b128c09f	296	mutex_exit(&spa->spa_props_lock);
34dc7c2f BB	297	}
34dc7c2f BB	298	}
34dc7c2f	299	*generation = spa_config_generation;
b128c09f	300	mutex_exit(&spa_namespace_lock);
34dc7c2f BB	301
	302	return (pools);
	303	}
	304
	305	void
	306	spa_config_set(spa_t spa, nvlist_t config)
	307	{
b128c09f	308	mutex_enter(&spa->spa_props_lock);
34dc7c2f BB	309	if (spa->spa_config != NULL)
	310	nvlist_free(spa->spa_config);
	311	spa->spa_config = config;
b128c09f	312	mutex_exit(&spa->spa_props_lock);
34dc7c2f BB	313	}
	314
	315	/*
	316	* Generate the pool's configuration based on the current in-core state.
	317	* We infer whether to generate a complete config or just one top-level config
	318	* based on whether vd is the root vdev.
	319	*/
	320	nvlist_t *
	321	spa_config_generate(spa_t spa, vdev_t vd, uint64_t txg, int getstats)
	322	{
	323	nvlist_t config, nvroot;
	324	vdev_t *rvd = spa->spa_root_vdev;
	325	unsigned long hostid = 0;
b128c09f	326	boolean_t locked = B_FALSE;
34dc7c2f	327
b128c09f	328	if (vd == NULL) {
34dc7c2f	329	vd = rvd;
b128c09f BB	330	locked = B_TRUE;
	331	spa_config_enter(spa, SCL_CONFIG \| SCL_STATE, FTAG, RW_READER);
	332	}
	333
	334	ASSERT(spa_config_held(spa, SCL_CONFIG \| SCL_STATE, RW_READER) ==
	335	(SCL_CONFIG \| SCL_STATE));
34dc7c2f BB	336
	337	/*
	338	* If txg is -1, report the current value of spa->spa_config_txg.
	339	*/
	340	if (txg == -1ULL)
	341	txg = spa->spa_config_txg;
	342
	343	VERIFY(nvlist_alloc(&config, NV_UNIQUE_NAME, KM_SLEEP) == 0);
	344
	345	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_VERSION,
	346	spa_version(spa)) == 0);
	347	VERIFY(nvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME,
	348	spa_name(spa)) == 0);
	349	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE,
	350	spa_state(spa)) == 0);
	351	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_TXG,
	352	txg) == 0);
	353	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_POOL_GUID,
	354	spa_guid(spa)) == 0);
	355	(void) ddi_strtoul(hw_serial, NULL, 10, &hostid);
	356	if (hostid != 0) {
	357	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_HOSTID,
	358	hostid) == 0);
	359	}
	360	VERIFY(nvlist_add_string(config, ZPOOL_CONFIG_HOSTNAME,
	361	utsname.nodename) == 0);
	362
	363	if (vd != rvd) {
	364	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_TOP_GUID,
	365	vd->vdev_top->vdev_guid) == 0);
	366	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_GUID,
	367	vd->vdev_guid) == 0);
	368	if (vd->vdev_isspare)
	369	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_IS_SPARE,
	370	1ULL) == 0);
	371	if (vd->vdev_islog)
	372	VERIFY(nvlist_add_uint64(config, ZPOOL_CONFIG_IS_LOG,
	373	1ULL) == 0);
	374	vd = vd->vdev_top; /* label contains top config */
	375	}
	376
	377	nvroot = vdev_config_generate(spa, vd, getstats, B_FALSE, B_FALSE);
	378	VERIFY(nvlist_add_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, nvroot) == 0);
	379	nvlist_free(nvroot);
	380
b128c09f BB	381	if (locked)
	382	spa_config_exit(spa, SCL_CONFIG \| SCL_STATE, FTAG);
	383
34dc7c2f BB	384	return (config);
	385	}
	386
	387	/*
	388	* For a pool that's not currently a booting rootpool, update all disk labels,
	389	* generate a fresh config based on the current in-core state, and sync the
	390	* global config cache.
	391	*/
	392	void
	393	spa_config_update(spa_t *spa, int what)
	394	{
	395	spa_config_update_common(spa, what, FALSE);
	396	}
	397
	398	/*
	399	* Update all disk labels, generate a fresh config based on the current
	400	* in-core state, and sync the global config cache (do not sync the config
	401	* cache if this is a booting rootpool).
	402	*/
	403	void
	404	spa_config_update_common(spa_t *spa, int what, boolean_t isroot)
	405	{
	406	vdev_t *rvd = spa->spa_root_vdev;
	407	uint64_t txg;
	408	int c;
	409
	410	ASSERT(MUTEX_HELD(&spa_namespace_lock));
	411
b128c09f	412	spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
34dc7c2f BB	413	txg = spa_last_synced_txg(spa) + 1;
	414	if (what == SPA_CONFIG_UPDATE_POOL) {
	415	vdev_config_dirty(rvd);
	416	} else {
	417	/*
	418	* If we have top-level vdevs that were added but have
	419	* not yet been prepared for allocation, do that now.
	420	* (It's safe now because the config cache is up to date,
	421	* so it will be able to translate the new DVAs.)
	422	* See comments in spa_vdev_add() for full details.
	423	*/
	424	for (c = 0; c < rvd->vdev_children; c++) {
	425	vdev_t *tvd = rvd->vdev_child[c];
	426	if (tvd->vdev_ms_array == 0) {
	427	vdev_init(tvd, txg);
	428	vdev_config_dirty(tvd);
	429	}
	430	}
	431	}
b128c09f	432	spa_config_exit(spa, SCL_ALL, FTAG);
34dc7c2f BB	433
	434	/*
	435	* Wait for the mosconfig to be regenerated and synced.
	436	*/
	437	txg_wait_synced(spa->spa_dsl_pool, txg);
	438
	439	/*
	440	* Update the global config cache to reflect the new mosconfig.
	441	*/
	442	if (!isroot)
b128c09f	443	spa_config_sync(spa, B_FALSE, what != SPA_CONFIG_UPDATE_POOL);
34dc7c2f BB	444
	445	if (what == SPA_CONFIG_UPDATE_POOL)
	446	spa_config_update_common(spa, SPA_CONFIG_UPDATE_VDEVS, isroot);
	447	}