Rebase master to b117

1 files changed, 37 insertions, 49 deletions

diff --git a/module/zfs/dnode.c b/module/zfs/dnode.c
index 538a141b0..cf49b97f1 100644
--- a/module/zfs/dnode.c
+++ b/module/zfs/dnode.c
@@ -156,7 +156,7 @@ dnode_verify(dnode_t *dn)
 	}
 	if (dn->dn_phys->dn_type != DMU_OT_NONE)
 		ASSERT3U(dn->dn_phys->dn_nlevels, <=, dn->dn_nlevels);
-	ASSERT(dn->dn_object == DMU_META_DNODE_OBJECT || dn->dn_dbuf != NULL);
+	ASSERT(DMU_OBJECT_IS_SPECIAL(dn->dn_object) || dn->dn_dbuf != NULL);
 	if (dn->dn_dbuf != NULL) {
 		ASSERT3P(dn->dn_phys, ==,
 		    (dnode_phys_t *)dn->dn_dbuf->db.db_data +
@@ -321,6 +321,7 @@ dnode_destroy(dnode_t *dn)
 	}
 	ASSERT(NULL == list_head(&dn->dn_dbufs));
 #endif
+	ASSERT(dn->dn_oldphys == NULL);
 
 	mutex_enter(&os->os_lock);
 	list_remove(&os->os_dnodes, dn);
@@ -417,8 +418,7 @@ void
 dnode_reallocate(dnode_t *dn, dmu_object_type_t ot, int blocksize,
     dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx)
 {
-	int i, nblkptr;
-	dmu_buf_impl_t *db = NULL;
+	int nblkptr;
 
 	ASSERT3U(blocksize, >=, SPA_MINBLOCKSIZE);
 	ASSERT3U(blocksize, <=, SPA_MAXBLOCKSIZE);
@@ -430,42 +430,25 @@ dnode_reallocate(dnode_t *dn, dmu_object_type_t ot, int blocksize,
 	ASSERT3U(bonustype, <, DMU_OT_NUMTYPES);
 	ASSERT3U(bonuslen, <=, DN_MAX_BONUSLEN);
 
-	for (i = 0; i < TXG_SIZE; i++)
-		ASSERT(!list_link_active(&dn->dn_dirty_link[i]));
-
 	/* clean up any unreferenced dbufs */
 	dnode_evict_dbufs(dn);
-	ASSERT3P(list_head(&dn->dn_dbufs), ==, NULL);
-
-	/*
-	 * XXX I should really have a generation number to tell if we
-	 * need to do this...
-	 */
-	if (blocksize != dn->dn_datablksz ||
-	    dn->dn_bonustype != bonustype || dn->dn_bonuslen != bonuslen) {
-		/* free all old data */
-		dnode_free_range(dn, 0, -1ULL, tx);
-	}
-
-	nblkptr = 1 + ((DN_MAX_BONUSLEN - bonuslen) >> SPA_BLKPTRSHIFT);
 
-	/* change blocksize */
 	rw_enter(&dn->dn_struct_rwlock, RW_WRITER);
-	if (blocksize != dn->dn_datablksz &&
-	    (!BP_IS_HOLE(&dn->dn_phys->dn_blkptr[0]) ||
-	    list_head(&dn->dn_dbufs) != NULL)) {
-		db = dbuf_hold(dn, 0, FTAG);
-		dbuf_new_size(db, blocksize, tx);
-	}
-	dnode_setdblksz(dn, blocksize);
 	dnode_setdirty(dn, tx);
-	dn->dn_next_bonuslen[tx->tx_txg&TXG_MASK] = bonuslen;
-	dn->dn_next_blksz[tx->tx_txg&TXG_MASK] = blocksize;
+	if (dn->dn_datablksz != blocksize) {
+		/* change blocksize */
+		ASSERT(dn->dn_maxblkid == 0 &&
+		    (BP_IS_HOLE(&dn->dn_phys->dn_blkptr[0]) ||
+		    dnode_block_freed(dn, 0)));
+		dnode_setdblksz(dn, blocksize);
+		dn->dn_next_blksz[tx->tx_txg&TXG_MASK] = blocksize;
+	}
+	if (dn->dn_bonuslen != bonuslen)
+		dn->dn_next_bonuslen[tx->tx_txg&TXG_MASK] = bonuslen;
+	nblkptr = 1 + ((DN_MAX_BONUSLEN - bonuslen) >> SPA_BLKPTRSHIFT);
 	if (dn->dn_nblkptr != nblkptr)
 		dn->dn_next_nblkptr[tx->tx_txg&TXG_MASK] = nblkptr;
 	rw_exit(&dn->dn_struct_rwlock);
-	if (db)
-		dbuf_rele(db, FTAG);
 
 	/* change type */
 	dn->dn_type = ot;
@@ -569,6 +552,22 @@ dnode_hold_impl(objset_impl_t *os, uint64_t object, int flag,
 	 */
 	ASSERT(spa_config_held(os->os_spa, SCL_ALL, RW_WRITER) == 0);
 
+	if (object == DMU_USERUSED_OBJECT || object == DMU_GROUPUSED_OBJECT) {
+		dn = (object == DMU_USERUSED_OBJECT) ?
+		    os->os_userused_dnode : os->os_groupused_dnode;
+		if (dn == NULL)
+			return (ENOENT);
+		type = dn->dn_type;
+		if ((flag & DNODE_MUST_BE_ALLOCATED) && type == DMU_OT_NONE)
+			return (ENOENT);
+		if ((flag & DNODE_MUST_BE_FREE) && type != DMU_OT_NONE)
+			return (EEXIST);
+		DNODE_VERIFY(dn);
+		(void) refcount_add(&dn->dn_holds, tag);
+		*dnp = dn;
+		return (0);
+	}
+
 	if (object == 0 || object >= DN_MAX_OBJECT)
 		return (EINVAL);
 
@@ -627,7 +626,8 @@ dnode_hold_impl(objset_impl_t *os, uint64_t object, int flag,
 	type = dn->dn_type;
 	if (dn->dn_free_txg ||
 	    ((flag & DNODE_MUST_BE_ALLOCATED) && type == DMU_OT_NONE) ||
-	    ((flag & DNODE_MUST_BE_FREE) && type != DMU_OT_NONE)) {
+	    ((flag & DNODE_MUST_BE_FREE) &&
+	    (type != DMU_OT_NONE || dn->dn_oldphys))) {
 		mutex_exit(&dn->dn_mtx);
 		dbuf_rele(db, FTAG);
 		return (type == DMU_OT_NONE ? ENOENT : EEXIST);
@@ -692,8 +692,10 @@ dnode_setdirty(dnode_t *dn, dmu_tx_t *tx)
 	objset_impl_t *os = dn->dn_objset;
 	uint64_t txg = tx->tx_txg;
 
-	if (dn->dn_object == DMU_META_DNODE_OBJECT)
+	if (DMU_OBJECT_IS_SPECIAL(dn->dn_object)) {
+		dsl_dataset_dirty(os->os_dsl_dataset, tx);
 		return;
+	}
 
 	DNODE_VERIFY(dn);
 
@@ -1189,11 +1191,6 @@ dnode_block_freed(dnode_t *dn, uint64_t blkid)
 	if (dn->dn_free_txg)
 		return (TRUE);
 
-	/*
-	 * If dn_datablkshift is not set, then there's only a single
-	 * block, in which case there will never be a free range so it
-	 * won't matter.
-	 */
 	range_tofind.fr_blkid = blkid;
 	mutex_enter(&dn->dn_mtx);
 	for (i = 0; i < TXG_SIZE; i++) {
@@ -1278,7 +1275,7 @@ dnode_next_offset_level(dnode_t *dn, int flags, uint64_t *offset,
 	dprintf("probing object %llu offset %llx level %d of %u\n",
 	    dn->dn_object, *offset, lvl, dn->dn_phys->dn_nlevels);
 
-	hole = flags & DNODE_FIND_HOLE;
+	hole = ((flags & DNODE_FIND_HOLE) != 0);
 	inc = (flags & DNODE_FIND_BACKWARDS) ? -1 : 1;
 	ASSERT(txg == 0 || !hole);
 
@@ -1325,16 +1322,7 @@ dnode_next_offset_level(dnode_t *dn, int flags, uint64_t *offset,
 
 		for (i = (*offset >> span) & (blkfill - 1);
 		    i >= 0 && i < blkfill; i += inc) {
-			boolean_t newcontents = B_TRUE;
-			if (txg) {
-				int j;
-				newcontents = B_FALSE;
-				for (j = 0; j < dnp[i].dn_nblkptr; j++) {
-					if (dnp[i].dn_blkptr[j].blk_birth > txg)
-						newcontents = B_TRUE;
-				}
-			}
-			if (!dnp[i].dn_type == hole && newcontents)
+			if ((dnp[i].dn_type == DMU_OT_NONE) == hole)
 				break;
 			*offset += (1ULL << span) * inc;
 		}