X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=blobdiff_plain;f=module%2Fzfs%2Fvdev.c;h=205a1d1aabbceb7a92088d873ae5211463ccc2f8;hb=af26c4d4ab545767456d8c21ed48e9e01ce6a3e7;hp=604a673f954c5c5a2d4475bd83ccd07651cb41ab;hpb=d6320ddb78fa89c4d0fc2af00ae53c7c70992f96;p=zfs.git

diff --git a/module/zfs/vdev.c b/module/zfs/vdev.c
index 604a673..205a1d1 100644
--- a/module/zfs/vdev.c
+++ b/module/zfs/vdev.c
@@ -21,6 +21,8 @@
 
 /*
  * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Copyright 2011 Nexenta Systems, Inc.  All rights reserved.
+ * Copyright (c) 2012 by Delphix. All rights reserved.
  */
 
 #include <sys/zfs_context.h>
@@ -107,7 +109,7 @@ vdev_get_min_asize(vdev_t *vd)
 	vdev_t *pvd = vd->vdev_parent;
 
 	/*
-	 * The our parent is NULL (inactive spare or cache) or is the root,
+	 * If our parent is NULL (inactive spare or cache) or is the root,
 	 * just return our own asize.
 	 */
 	if (pvd == NULL)
@@ -193,7 +195,7 @@ vdev_add_child(vdev_t *pvd, vdev_t *cvd)
 	pvd->vdev_children = MAX(pvd->vdev_children, id + 1);
 	newsize = pvd->vdev_children * sizeof (vdev_t *);
 
-	newchild = kmem_zalloc(newsize, KM_SLEEP);
+	newchild = kmem_zalloc(newsize, KM_PUSHPAGE);
 	if (pvd->vdev_child != NULL) {
 		bcopy(pvd->vdev_child, newchild, oldsize);
 		kmem_free(pvd->vdev_child, oldsize);
@@ -263,7 +265,7 @@ vdev_compact_children(vdev_t *pvd)
 		if (pvd->vdev_child[c])
 			newc++;
 
-	newchild = kmem_alloc(newc * sizeof (vdev_t *), KM_SLEEP);
+	newchild = kmem_alloc(newc * sizeof (vdev_t *), KM_PUSHPAGE);
 
 	for (c = newc = 0; c < oldc; c++) {
 		if ((cvd = pvd->vdev_child[c]) != NULL) {
@@ -286,11 +288,12 @@ vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops)
 	vdev_t *vd;
 	int t;
 
-	vd = kmem_zalloc(sizeof (vdev_t), KM_SLEEP);
+	vd = kmem_zalloc(sizeof (vdev_t), KM_PUSHPAGE);
 
 	if (spa->spa_root_vdev == NULL) {
 		ASSERT(ops == &vdev_root_ops);
 		spa->spa_root_vdev = vd;
+		spa->spa_load_guid = spa_generate_guid(NULL);
 	}
 
 	if (guid == 0 && ops != &vdev_hole_ops) {
@@ -317,6 +320,8 @@ vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops)
 	vd->vdev_state = VDEV_STATE_CLOSED;
 	vd->vdev_ishole = (ops == &vdev_hole_ops);
 
+	list_link_init(&vd->vdev_config_dirty_node);
+	list_link_init(&vd->vdev_state_dirty_node);
 	mutex_init(&vd->vdev_dtl_lock, NULL, MUTEX_DEFAULT, NULL);
 	mutex_init(&vd->vdev_stat_lock, NULL, MUTEX_DEFAULT, NULL);
 	mutex_init(&vd->vdev_probe_lock, NULL, MUTEX_DEFAULT, NULL);
@@ -490,7 +495,7 @@ vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id,
 		    &vd->vdev_removing);
 	}
 
-	if (parent && !parent->vdev_parent) {
+	if (parent && !parent->vdev_parent && alloctype != VDEV_ALLOC_ATTACH) {
 		ASSERT(alloctype == VDEV_ALLOC_LOAD ||
 		    alloctype == VDEV_ALLOC_ADD ||
 		    alloctype == VDEV_ALLOC_SPLIT ||
@@ -667,6 +672,8 @@ vdev_top_transfer(vdev_t *svd, vdev_t *tvd)
 	svd->vdev_ms_shift = 0;
 	svd->vdev_ms_count = 0;
 
+	if (tvd->vdev_mg)
+		ASSERT3P(tvd->vdev_mg, ==, svd->vdev_mg);
 	tvd->vdev_mg = svd->vdev_mg;
 	tvd->vdev_ms = svd->vdev_ms;
 
@@ -740,6 +747,7 @@ vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops)
 
 	mvd->vdev_asize = cvd->vdev_asize;
 	mvd->vdev_min_asize = cvd->vdev_min_asize;
+	mvd->vdev_max_asize = cvd->vdev_max_asize;
 	mvd->vdev_ashift = cvd->vdev_ashift;
 	mvd->vdev_state = cvd->vdev_state;
 	mvd->vdev_crtxg = cvd->vdev_crtxg;
@@ -831,7 +839,7 @@ vdev_metaslab_init(vdev_t *vd, uint64_t txg)
 
 	ASSERT(oldc <= newc);
 
-	mspp = kmem_zalloc(newc * sizeof (*mspp), KM_SLEEP);
+	mspp = kmem_zalloc(newc * sizeof (*mspp), KM_PUSHPAGE | KM_NODEBUG);
 
 	if (oldc != 0) {
 		bcopy(vd->vdev_ms, mspp, oldc * sizeof (*mspp));
@@ -986,7 +994,7 @@ vdev_probe(vdev_t *vd, zio_t *zio)
 	mutex_enter(&vd->vdev_probe_lock);
 
 	if ((pio = vd->vdev_probe_zio) == NULL) {
-		vps = kmem_zalloc(sizeof (*vps), KM_SLEEP);
+		vps = kmem_zalloc(sizeof (*vps), KM_PUSHPAGE);
 
 		vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE |
 		    ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_AGGREGATE |
@@ -1067,6 +1075,15 @@ vdev_open_child(void *arg)
 boolean_t
 vdev_uses_zvols(vdev_t *vd)
 {
+/*
+ * Stacking zpools on top of zvols is unsupported until we implement a method
+ * for determining if an arbitrary block device is a zvol without using the
+ * path.  Solaris would check the 'zvol' path component but this does not
+ * exist in the Linux port, so we really should do something like stat the
+ * file and check the major number.  This is complicated by the fact that
+ * we need to do this portably in user or kernel space.
+ */
+#if 0
 	int c;
 
 	if (vd->vdev_path && strncmp(vd->vdev_path, ZVOL_DIR,
@@ -1075,6 +1092,7 @@ vdev_uses_zvols(vdev_t *vd)
 	for (c = 0; c < vd->vdev_children; c++)
 		if (vdev_uses_zvols(vd->vdev_child[c]))
 			return (B_TRUE);
+#endif
 	return (B_FALSE);
 }
 
@@ -1115,7 +1133,8 @@ vdev_open(vdev_t *vd)
 	spa_t *spa = vd->vdev_spa;
 	int error;
 	uint64_t osize = 0;
-	uint64_t asize, psize;
+	uint64_t max_osize = 0;
+	uint64_t asize, max_asize, psize;
 	uint64_t ashift = 0;
 	int c;
 
@@ -1147,7 +1166,7 @@ vdev_open(vdev_t *vd)
 		return (ENXIO);
 	}
 
-	error = vd->vdev_ops->vdev_op_open(vd, &osize, &ashift);
+	error = vd->vdev_ops->vdev_op_open(vd, &osize, &max_osize, &ashift);
 
 	/*
 	 * Reset the vdev_reopening flag so that we actually close
@@ -1205,6 +1224,7 @@ vdev_open(vdev_t *vd)
 	}
 
 	osize = P2ALIGN(osize, (uint64_t)sizeof (vdev_label_t));
+	max_osize = P2ALIGN(max_osize, (uint64_t)sizeof (vdev_label_t));
 
 	if (vd->vdev_children == 0) {
 		if (osize < SPA_MINDEVSIZE) {
@@ -1214,6 +1234,8 @@ vdev_open(vdev_t *vd)
 		}
 		psize = osize;
 		asize = osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE);
+		max_asize = max_osize - (VDEV_LABEL_START_SIZE +
+		    VDEV_LABEL_END_SIZE);
 	} else {
 		if (vd->vdev_parent != NULL && osize < SPA_MINDEVSIZE -
 		    (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE)) {
@@ -1223,6 +1245,7 @@ vdev_open(vdev_t *vd)
 		}
 		psize = 0;
 		asize = osize;
+		max_asize = max_osize;
 	}
 
 	vd->vdev_psize = psize;
@@ -1242,6 +1265,7 @@ vdev_open(vdev_t *vd)
 		 * For testing purposes, a higher ashift can be requested.
 		 */
 		vd->vdev_asize = asize;
+		vd->vdev_max_asize = max_asize;
 		vd->vdev_ashift = MAX(ashift, vd->vdev_ashift);
 	} else {
 		/*
@@ -1252,6 +1276,7 @@ vdev_open(vdev_t *vd)
 			    VDEV_AUX_BAD_LABEL);
 			return (EINVAL);
 		}
+		vd->vdev_max_asize = max_asize;
 	}
 
 	/*
@@ -1293,13 +1318,18 @@ vdev_open(vdev_t *vd)
  * contents.  This needs to be done before vdev_load() so that we don't
  * inadvertently do repair I/Os to the wrong device.
  *
+ * If 'strict' is false ignore the spa guid check. This is necessary because
+ * if the machine crashed during a re-guid the new guid might have been written
+ * to all of the vdev labels, but not the cached config. The strict check
+ * will be performed when the pool is opened again using the mos config.
+ *
  * This function will only return failure if one of the vdevs indicates that it
  * has since been destroyed or exported.  This is only possible if
  * /etc/zfs/zpool.cache was readonly at the time.  Otherwise, the vdev state
  * will be updated but the function will return 0.
  */
 int
-vdev_validate(vdev_t *vd)
+vdev_validate(vdev_t *vd, boolean_t strict)
 {
 	spa_t *spa = vd->vdev_spa;
 	nvlist_t *label;
@@ -1308,7 +1338,7 @@ vdev_validate(vdev_t *vd)
 	int c;
 
 	for (c = 0; c < vd->vdev_children; c++)
-		if (vdev_validate(vd->vdev_child[c]) != 0)
+		if (vdev_validate(vd->vdev_child[c], strict) != 0)
 			return (EBADF);
 
 	/*
@@ -1338,8 +1368,9 @@ vdev_validate(vdev_t *vd)
 			return (0);
 		}
 
-		if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID,
-		    &guid) != 0 || guid != spa_guid(spa)) {
+		if (strict && (nvlist_lookup_uint64(label,
+		    ZPOOL_CONFIG_POOL_GUID, &guid) != 0 ||
+		    guid != spa_guid(spa))) {
 			vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN,
 			    VDEV_AUX_CORRUPT_DATA);
 			nvlist_free(label);
@@ -1412,8 +1443,8 @@ vdev_validate(vdev_t *vd)
 void
 vdev_close(vdev_t *vd)
 {
-	spa_t *spa = vd->vdev_spa;
 	vdev_t *pvd = vd->vdev_parent;
+	ASSERTV(spa_t *spa = vd->vdev_spa);
 
 	ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
 
@@ -1502,7 +1533,7 @@ vdev_reopen(vdev_t *vd)
 		    !l2arc_vdev_present(vd))
 			l2arc_add_vdev(spa, vd);
 	} else {
-		(void) vdev_validate(vd);
+		(void) vdev_validate(vd, B_TRUE);
 	}
 
 	/*
@@ -1795,8 +1826,7 @@ vdev_dtl_sync(vdev_t *vd, uint64_t txg)
 
 	if (vd->vdev_detached) {
 		if (smo->smo_object != 0) {
-			int err = dmu_object_free(mos, smo->smo_object, tx);
-			ASSERT3U(err, ==, 0);
+			VERIFY(0 == dmu_object_free(mos, smo->smo_object, tx));
 			smo->smo_object = 0;
 		}
 		dmu_tx_commit(tx);
@@ -2034,7 +2064,7 @@ vdev_sync_done(vdev_t *vd, uint64_t txg)
 
 	ASSERT(!vd->vdev_ishole);
 
-	while (msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg)))
+	while ((msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg))))
 		metaslab_sync_done(msp, txg);
 
 	if (reassess)
@@ -2387,7 +2417,7 @@ vdev_clear(spa_t *spa, vdev_t *vd)
 		if (vd->vdev_aux == NULL && !vdev_is_dead(vd))
 			spa_async_request(spa, SPA_ASYNC_RESILVER);
 
-		spa_event_notify(spa, vd, ESC_ZFS_VDEV_CLEAR);
+		spa_event_notify(spa, vd, FM_EREPORT_ZFS_DEVICE_CLEAR);
 	}
 
 	/*
@@ -2477,6 +2507,7 @@ vdev_get_stats(vdev_t *vd, vdev_stat_t *vs)
 	vs->vs_rsize = vdev_get_min_asize(vd);
 	if (vd->vdev_ops->vdev_op_leaf)
 		vs->vs_rsize += VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE;
+	vs->vs_esize = vd->vdev_max_asize - vd->vdev_asize;
 	mutex_exit(&vd->vdev_stat_lock);
 
 	/*
@@ -3042,13 +3073,17 @@ vdev_set_state(vdev_t *vd, boolean_t isopen, vdev_state_t state, vdev_aux_t aux)
 
 /*
  * Check the vdev configuration to ensure that it's capable of supporting
- * a root pool. Currently, we do not support RAID-Z or partial configuration.
- * In addition, only a single top-level vdev is allowed and none of the leaves
- * can be wholedisks.
+ * a root pool.
  */
 boolean_t
 vdev_is_bootable(vdev_t *vd)
 {
+#if defined(__sun__) || defined(__sun)
+	/*
+	 * Currently, we do not support RAID-Z or partial configuration.
+	 * In addition, only a single top-level vdev is allowed and none of the
+	 * leaves can be wholedisks.
+	 */
 	int c;
 
 	if (!vd->vdev_ops->vdev_op_leaf) {
@@ -3069,6 +3104,7 @@ vdev_is_bootable(vdev_t *vd)
 		if (!vdev_is_bootable(vd->vdev_child[c]))
 			return (B_FALSE);
 	}
+#endif /* __sun__ || __sun */
 	return (B_TRUE);
 }
 
@@ -3084,7 +3120,8 @@ vdev_load_log_state(vdev_t *nvd, vdev_t *ovd)
 	int c;
 
 	ASSERT(nvd->vdev_top->vdev_islog);
-	ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
+	ASSERT(spa_config_held(nvd->vdev_spa,
+	    SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL);
 	ASSERT3U(nvd->vdev_guid, ==, ovd->vdev_guid);
 
 	for (c = 0; c < nvd->vdev_children; c++)
@@ -3155,3 +3192,14 @@ vdev_split(vdev_t *vd)
 	}
 	vdev_propagate_state(cvd);
 }
+
+#if defined(_KERNEL) && defined(HAVE_SPL)
+EXPORT_SYMBOL(vdev_fault);
+EXPORT_SYMBOL(vdev_degrade);
+EXPORT_SYMBOL(vdev_online);
+EXPORT_SYMBOL(vdev_offline);
+EXPORT_SYMBOL(vdev_clear);
+
+module_param(zfs_scrub_limit, int, 0644);
+MODULE_PARM_DESC(zfs_scrub_limit, "Max scrub/resilver I/O per leaf vdev");
+#endif