Implement large_dnode pool feature

Justification
-------------

This feature adds support for variable length dnodes. Our motivation is
to eliminate the overhead associated with using spill blocks.  Spill
blocks are used to store system attribute data (i.e. file metadata) that
does not fit in the dnode's bonus buffer. By allowing a larger bonus
buffer area the use of a spill block can be avoided.  Spill blocks
potentially incur an additional read I/O for every dnode in a dnode
block. As a worst case example, reading 32 dnodes from a 16k dnode block
and all of the spill blocks could issue 33 separate reads. Now suppose
those dnodes have size 1024 and therefore don't need spill blocks.  Then
the worst case number of blocks read is reduced to from 33 to two--one
per dnode block. In practice spill blocks may tend to be co-located on
disk with the dnode blocks so the reduction in I/O would not be this
drastic. In a badly fragmented pool, however, the improvement could be
significant.

ZFS-on-Linux systems that make heavy use of extended attributes would
benefit from this feature. In particular, ZFS-on-Linux supports the
xattr=sa dataset property which allows file extended attribute data
to be stored in the dnode bonus buffer as an alternative to the
traditional directory-based format. Workloads such as SELinux and the
Lustre distributed filesystem often store enough xattr data to force
spill bocks when xattr=sa is in effect. Large dnodes may therefore
provide a performance benefit to such systems.

Other use cases that may benefit from this feature include files with
large ACLs and symbolic links with long target names. Furthermore,
this feature may be desirable on other platforms in case future
applications or features are developed that could make use of a
larger bonus buffer area.

Implementation
--------------

The size of a dnode may be a multiple of 512 bytes up to the size of
a dnode block (currently 16384 bytes). A dn_extra_slots field was
added to the current on-disk dnode_phys_t structure to describe the
size of the physical dnode on disk. The 8 bits for this field were
taken from the zero filled dn_pad2 field. The field represents how
many "extra" dnode_phys_t slots a dnode consumes in its dnode block.
This convention results in a value of 0 for 512 byte dnodes which
preserves on-disk format compatibility with older software.

Similarly, the in-memory dnode_t structure has a new dn_num_slots field
to represent the total number of dnode_phys_t slots consumed on disk.
Thus dn->dn_num_slots is 1 greater than the corresponding
dnp->dn_extra_slots. This difference in convention was adopted
because, unlike on-disk structures, backward compatibility is not a
concern for in-memory objects, so we used a more natural way to
represent size for a dnode_t.

The default size for newly created dnodes is determined by the value of
a new "dnodesize" dataset property. By default the property is set to
"legacy" which is compatible with older software. Setting the property
to "auto" will allow the filesystem to choose the most suitable dnode
size. Currently this just sets the default dnode size to 1k, but future
code improvements could dynamically choose a size based on observed
workload patterns. Dnodes of varying sizes can coexist within the same
dataset and even within the same dnode block. For example, to enable
automatically-sized dnodes, run

 # zfs set dnodesize=auto tank/fish

The user can also specify literal values for the dnodesize property.
These are currently limited to powers of two from 1k to 16k. The
power-of-2 limitation is only for simplicity of the user interface.
Internally the implementation can handle any multiple of 512 up to 16k,
and consumers of the DMU API can specify any legal dnode value.

The size of a new dnode is determined at object allocation time and
stored as a new field in the znode in-memory structure. New DMU
interfaces are added to allow the consumer to specify the dnode size
that a newly allocated object should use. Existing interfaces are
unchanged to avoid having to update every call site and to preserve
compatibility with external consumers such as Lustre. The new
interfaces names are given below. The versions of these functions that
don't take a dnodesize parameter now just call the _dnsize() versions
with a dnodesize of 0, which means use the legacy dnode size.

New DMU interfaces:
  dmu_object_alloc_dnsize()
  dmu_object_claim_dnsize()
  dmu_object_reclaim_dnsize()

New ZAP interfaces:
  zap_create_dnsize()
  zap_create_norm_dnsize()
  zap_create_flags_dnsize()
  zap_create_claim_norm_dnsize()
  zap_create_link_dnsize()

The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The
spa_maxdnodesize() function should be used to determine the maximum
bonus length for a pool.

These are a few noteworthy changes to key functions:

* The prototype for dnode_hold_impl() now takes a "slots" parameter.
  When the DNODE_MUST_BE_FREE flag is set, this parameter is used to
  ensure the hole at the specified object offset is large enough to
  hold the dnode being created. The slots parameter is also used
  to ensure a dnode does not span multiple dnode blocks. In both of
  these cases, if a failure occurs, ENOSPC is returned. Keep in mind,
  these failure cases are only possible when using DNODE_MUST_BE_FREE.

  If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0.
  dnode_hold_impl() will check if the requested dnode is already
  consumed as an extra dnode slot by an large dnode, in which case
  it returns ENOENT.

* The function dmu_object_alloc() advances to the next dnode block
  if dnode_hold_impl() returns an error for a requested object.
  This is because the beginning of the next dnode block is the only
  location it can safely assume to either be a hole or a valid
  starting point for a dnode.

* dnode_next_offset_level() and other functions that iterate
  through dnode blocks may no longer use a simple array indexing
  scheme. These now use the current dnode's dn_num_slots field to
  advance to the next dnode in the block. This is to ensure we
  properly skip the current dnode's bonus area and don't interpret it
  as a valid dnode.

zdb
---
The zdb command was updated to display a dnode's size under the
"dnsize" column when the object is dumped.

For ZIL create log records, zdb will now display the slot count for
the object.

ztest
-----
Ztest chooses a random dnodesize for every newly created object. The
random distribution is more heavily weighted toward small dnodes to
better simulate real-world datasets.

Unused bonus buffer space is filled with non-zero values computed from
the object number, dataset id, offset, and generation number.  This
helps ensure that the dnode traversal code properly skips the interior
regions of large dnodes, and that these interior regions are not
overwritten by data belonging to other dnodes. A new test visits each
object in a dataset. It verifies that the actual dnode size matches what
was stored in the ztest block tag when it was created. It also verifies
that the unused bonus buffer space is filled with the expected data
patterns.

ZFS Test Suite
--------------
Added six new large dnode-specific tests, and integrated the dnodesize
property into existing tests for zfs allow and send/recv.

Send/Receive
------------
ZFS send streams for datasets containing large dnodes cannot be received
on pools that don't support the large_dnode feature. A send stream with
large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be
unrecognized by an incompatible receiving pool so that the zfs receive
will fail gracefully.

While not implemented here, it may be possible to generate a
backward-compatible send stream from a dataset containing large
dnodes. The implementation may be tricky, however, because the send
object record for a large dnode would need to be resized to a 512
byte dnode, possibly kicking in a spill block in the process. This
means we would need to construct a new SA layout and possibly
register it in the SA layout object. The SA layout is normally just
sent as an ordinary object record. But if we are constructing new
layouts while generating the send stream we'd have to build the SA
layout object dynamically and send it at the end of the stream.

For sending and receiving between pools that do support large dnodes,
the drr_object send record type is extended with a new field to store
the dnode slot count. This field was repurposed from unused padding
in the structure.

ZIL Replay
----------
The dnode slot count is stored in the uppermost 8 bits of the lr_foid
field. The bits were unused as the object id is currently capped at
48 bits.

Resizing Dnodes
---------------
It should be possible to resize a dnode when it is dirtied if the
current dnodesize dataset property differs from the dnode's size, but
this functionality is not currently implemented. Clearly a dnode can
only grow if there are sufficient contiguous unused slots in the
dnode block, but it should always be possible to shrink a dnode.
Growing dnodes may be useful to reduce fragmentation in a pool with
many spill blocks in use. Shrinking dnodes may be useful to allow
sending a dataset to a pool that doesn't support the large_dnode
feature.

Feature Reference Counting
--------------------------
The reference count for the large_dnode pool feature tracks the
number of datasets that have ever contained a dnode of size larger
than 512 bytes. The first time a large dnode is created in a dataset
the dataset is converted to an extensible dataset. This is a one-way
operation and the only way to decrement the feature count is to
destroy the dataset, even if the dataset no longer contains any large
dnodes. The complexity of reference counting on a per-dnode basis was
too high, so we chose to track it on a per-dataset basis similarly to
the large_block feature.

Signed-off-by: Ned Bass <[email protected]>
Signed-off-by: Brian Behlendorf <[email protected]>
Closes #3542

11 files changed, 102 insertions, 17 deletions

diff --git a/include/sys/dmu.h b/include/sys/dmu.h
index 64d28334e..74001d8f3 100644
--- a/include/sys/dmu.h
+++ b/include/sys/dmu.h
@@ -334,10 +334,19 @@ typedef struct dmu_buf {
  */
 uint64_t dmu_object_alloc(objset_t *os, dmu_object_type_t ot,
     int blocksize, dmu_object_type_t bonus_type, int bonus_len, dmu_tx_t *tx);
+uint64_t dmu_object_alloc_dnsize(objset_t *os, dmu_object_type_t ot,
+    int blocksize, dmu_object_type_t bonus_type, int bonus_len,
+    int dnodesize, dmu_tx_t *tx);
 int dmu_object_claim(objset_t *os, uint64_t object, dmu_object_type_t ot,
     int blocksize, dmu_object_type_t bonus_type, int bonus_len, dmu_tx_t *tx);
+int dmu_object_claim_dnsize(objset_t *os, uint64_t object, dmu_object_type_t ot,
+    int blocksize, dmu_object_type_t bonus_type, int bonus_len,
+    int dnodesize, dmu_tx_t *tx);
 int dmu_object_reclaim(objset_t *os, uint64_t object, dmu_object_type_t ot,
     int blocksize, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *txp);
+int dmu_object_reclaim_dnsize(objset_t *os, uint64_t object,
+    dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype,
+    int bonuslen, int dnodesize, dmu_tx_t *txp);
 
 /*
  * Free an object from this objset.
@@ -756,6 +765,7 @@ typedef struct dmu_object_info {
 	uint8_t doi_compress;
 	uint8_t doi_nblkptr;
 	uint8_t doi_pad[4];
+	uint64_t doi_dnodesize;
 	uint64_t doi_physical_blocks_512;	/* data + metadata, 512b blks */
 	uint64_t doi_max_offset;
 	uint64_t doi_fill_count;		/* number of non-empty blocks */
@@ -797,6 +807,8 @@ void dmu_object_info_from_db(dmu_buf_t *db, dmu_object_info_t *doi);
 void dmu_object_size_from_db(dmu_buf_t *db, uint32_t *blksize,
     u_longlong_t *nblk512);
 
+void dmu_object_dnsize_from_db(dmu_buf_t *db, int *dnsize);
+
 typedef struct dmu_objset_stats {
 	uint64_t dds_num_clones; /* number of clones of this */
 	uint64_t dds_creation_txg;
@@ -854,6 +866,7 @@ extern struct dsl_dataset *dmu_objset_ds(objset_t *os);
 extern void dmu_objset_name(objset_t *os, char *buf);
 extern dmu_objset_type_t dmu_objset_type(objset_t *os);
 extern uint64_t dmu_objset_id(objset_t *os);
+extern uint64_t dmu_objset_dnodesize(objset_t *os);
 extern zfs_sync_type_t dmu_objset_syncprop(objset_t *os);
 extern zfs_logbias_op_t dmu_objset_logbias(objset_t *os);
 extern int dmu_snapshot_list_next(objset_t *os, int namelen, char *name,
diff --git a/include/sys/dmu_objset.h b/include/sys/dmu_objset.h
index 0f03699f1..81bb89bfc 100644
--- a/include/sys/dmu_objset.h
+++ b/include/sys/dmu_objset.h
@@ -88,6 +88,7 @@ struct objset {
 	list_node_t os_evicting_node;
 
 	/* can change, under dsl_dir's locks: */
+	uint64_t os_dnodesize; /* default dnode size for new objects */
 	enum zio_checksum os_checksum;
 	enum zio_compress os_compress;
 	uint8_t os_copies;
diff --git a/include/sys/dnode.h b/include/sys/dnode.h
index c5250d51a..cee4ea783 100644
--- a/include/sys/dnode.h
+++ b/include/sys/dnode.h
@@ -79,11 +79,18 @@ extern "C" {
 /*
  * Derived constants.
  */
-#define	DNODE_SIZE	(1 << DNODE_SHIFT)
-#define	DN_MAX_NBLKPTR	((DNODE_SIZE - DNODE_CORE_SIZE) >> SPA_BLKPTRSHIFT)
-#define	DN_MAX_BONUSLEN	(DNODE_SIZE - DNODE_CORE_SIZE - (1 << SPA_BLKPTRSHIFT))
+#define	DNODE_MIN_SIZE		(1 << DNODE_SHIFT)
+#define	DNODE_MAX_SIZE		(1 << DNODE_BLOCK_SHIFT)
+#define	DNODE_BLOCK_SIZE	(1 << DNODE_BLOCK_SHIFT)
+#define	DNODE_MIN_SLOTS		(DNODE_MIN_SIZE >> DNODE_SHIFT)
+#define	DNODE_MAX_SLOTS		(DNODE_MAX_SIZE >> DNODE_SHIFT)
+#define	DN_BONUS_SIZE(dnsize)	((dnsize) - DNODE_CORE_SIZE - \
+	(1 << SPA_BLKPTRSHIFT))
+#define	DN_SLOTS_TO_BONUSLEN(slots)	DN_BONUS_SIZE((slots) << DNODE_SHIFT)
+#define	DN_OLD_MAX_BONUSLEN	(DN_BONUS_SIZE(DNODE_MIN_SIZE))
+#define	DN_MAX_NBLKPTR	((DNODE_MIN_SIZE - DNODE_CORE_SIZE) >> SPA_BLKPTRSHIFT)
 #define	DN_MAX_OBJECT	(1ULL << DN_MAX_OBJECT_SHIFT)
-#define	DN_ZERO_BONUSLEN	(DN_MAX_BONUSLEN + 1)
+#define	DN_ZERO_BONUSLEN	(DN_BONUS_SIZE(DNODE_MAX_SIZE) + 1)
 #define	DN_KILL_SPILLBLK (1)
 
 #define	DNODES_PER_BLOCK_SHIFT	(DNODE_BLOCK_SHIFT - DNODE_SHIFT)
@@ -131,7 +138,8 @@ typedef struct dnode_phys {
 	uint8_t dn_flags;		/* DNODE_FLAG_* */
 	uint16_t dn_datablkszsec;	/* data block size in 512b sectors */
 	uint16_t dn_bonuslen;		/* length of dn_bonus */
-	uint8_t dn_pad2[4];
+	uint8_t dn_extra_slots;		/* # of subsequent slots consumed */
+	uint8_t dn_pad2[3];
 
 	/* accounting is protected by dn_dirty_mtx */
 	uint64_t dn_maxblkid;		/* largest allocated block ID */
@@ -140,8 +148,11 @@ typedef struct dnode_phys {
 	uint64_t dn_pad3[4];
 
 	/*
-	 * The tail region is 448 bytes, and there are three ways to
-	 * look at it.
+	 * The tail region is 448 bytes for a 512 byte dnode, and
+	 * correspondingly larger for larger dnode sizes. The spill
+	 * block pointer, when present, is always at the end of the tail
+	 * region. There are three ways this space may be used, using
+	 * a 512 byte dnode for this diagram:
 	 *
 	 * 0       64      128     192     256     320     384     448 (offset)
 	 * +---------------+---------------+---------------+-------+
@@ -149,23 +160,27 @@ typedef struct dnode_phys {
 	 * +---------------+---------------+---------------+-------+
 	 * | dn_blkptr[0]  | dn_bonus[0..319]                      |
 	 * +---------------+-----------------------+---------------+
-	 * | dn_blkptr[0]  | /                     | dn_spill      |
+	 * | dn_blkptr[0]  | dn_bonus[0..191]      | dn_spill      |
 	 * +---------------+-----------------------+---------------+
 	 */
 	union {
-		blkptr_t dn_blkptr[1+DN_MAX_BONUSLEN/sizeof (blkptr_t)];
+		blkptr_t dn_blkptr[1+DN_OLD_MAX_BONUSLEN/sizeof (blkptr_t)];
 		struct {
 			blkptr_t __dn_ignore1;
-			uint8_t dn_bonus[DN_MAX_BONUSLEN];
+			uint8_t dn_bonus[DN_OLD_MAX_BONUSLEN];
 		};
 		struct {
 			blkptr_t __dn_ignore2;
-			uint8_t __dn_ignore3[DN_MAX_BONUSLEN-sizeof (blkptr_t)];
+			uint8_t __dn_ignore3[DN_OLD_MAX_BONUSLEN -
+			    sizeof (blkptr_t)];
 			blkptr_t dn_spill;
 		};
 	};
 } dnode_phys_t;
 
+#define	DN_SPILL_BLKPTR(dnp)	(blkptr_t *)((char *)(dnp) + \
+	(((dnp)->dn_extra_slots + 1) << DNODE_SHIFT) - (1 << SPA_BLKPTRSHIFT))
+
 typedef struct dnode {
 	/*
 	 * Protects the structure of the dnode, including the number of levels
@@ -202,6 +217,7 @@ typedef struct dnode {
 	uint32_t dn_datablksz;		/* in bytes */
 	uint64_t dn_maxblkid;
 	uint8_t dn_next_type[TXG_SIZE];
+	uint8_t dn_num_slots;		/* metadnode slots consumed on disk */
 	uint8_t dn_next_nblkptr[TXG_SIZE];
 	uint8_t dn_next_nlevels[TXG_SIZE];
 	uint8_t dn_next_indblkshift[TXG_SIZE];
@@ -299,7 +315,7 @@ void dnode_rm_spill(dnode_t *dn, dmu_tx_t *tx);
 
 int dnode_hold(struct objset *dd, uint64_t object,
     void *ref, dnode_t **dnp);
-int dnode_hold_impl(struct objset *dd, uint64_t object, int flag,
+int dnode_hold_impl(struct objset *dd, uint64_t object, int flag, int dn_slots,
     void *ref, dnode_t **dnp);
 boolean_t dnode_add_ref(dnode_t *dn, void *ref);
 void dnode_rele(dnode_t *dn, void *ref);
@@ -307,9 +323,9 @@ void dnode_rele_and_unlock(dnode_t *dn, void *tag);
 void dnode_setdirty(dnode_t *dn, dmu_tx_t *tx);
 void dnode_sync(dnode_t *dn, dmu_tx_t *tx);
 void dnode_allocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, int ibs,
-    dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx);
+    dmu_object_type_t bonustype, int bonuslen, int dn_slots, dmu_tx_t *tx);
 void dnode_reallocate(dnode_t *dn, dmu_object_type_t ot, int blocksize,
-    dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx);
+    dmu_object_type_t bonustype, int bonuslen, int dn_slots, dmu_tx_t *tx);
 void dnode_free(dnode_t *dn, dmu_tx_t *tx);
 void dnode_byteswap(dnode_phys_t *dnp);
 void dnode_buf_byteswap(void *buf, size_t size);
diff --git a/include/sys/dsl_dataset.h b/include/sys/dsl_dataset.h
index a596642e3..f3e64a772 100644
--- a/include/sys/dsl_dataset.h
+++ b/include/sys/dsl_dataset.h
@@ -92,6 +92,13 @@ struct dsl_pool;
 #define	DS_FIELD_LARGE_BLOCKS "org.open-zfs:large_blocks"
 
 /*
+ * This field is present (with value=0) if this dataset may contain large
+ * dnodes (>512B).  If it is present, then this dataset is counted in the
+ * refcount of the SPA_FEATURE_LARGE_DNODE feature.
+ */
+#define	DS_FIELD_LARGE_DNODE "org.zfsonlinux:large_dnode"
+
+/*
  * DS_FLAG_CI_DATASET is set if the dataset contains a file system whose
  * name lookups should be performed case-insensitively.
  */
diff --git a/include/sys/fs/zfs.h b/include/sys/fs/zfs.h
index 9de50e9a2..0935dca77 100644
--- a/include/sys/fs/zfs.h
+++ b/include/sys/fs/zfs.h
@@ -137,6 +137,7 @@ typedef enum {
 	ZFS_PROP_DEDUP,
 	ZFS_PROP_MLSLABEL,
 	ZFS_PROP_SYNC,
+	ZFS_PROP_DNODESIZE,
 	ZFS_PROP_REFRATIO,
 	ZFS_PROP_WRITTEN,
 	ZFS_PROP_CLONES,
@@ -204,6 +205,7 @@ typedef enum {
 	ZPOOL_PROP_LEAKED,
 	ZPOOL_PROP_MAXBLOCKSIZE,
 	ZPOOL_PROP_TNAME,
+	ZPOOL_PROP_MAXDNODESIZE,
 	ZPOOL_NUM_PROPS
 } zpool_prop_t;
 
@@ -362,6 +364,16 @@ typedef enum {
 } zfs_xattr_type_t;
 
 typedef enum {
+	ZFS_DNSIZE_LEGACY = 0,
+	ZFS_DNSIZE_AUTO = 1,
+	ZFS_DNSIZE_1K = 1024,
+	ZFS_DNSIZE_2K = 2048,
+	ZFS_DNSIZE_4K = 4096,
+	ZFS_DNSIZE_8K = 8192,
+	ZFS_DNSIZE_16K = 16384
+} zfs_dnsize_type_t;
+
+typedef enum {
 	ZFS_REDUNDANT_METADATA_ALL,
 	ZFS_REDUNDANT_METADATA_MOST
 } zfs_redundant_metadata_type_t;
diff --git a/include/sys/sa_impl.h b/include/sys/sa_impl.h
index 6f2f1db6d..b68b7610b 100644
--- a/include/sys/sa_impl.h
+++ b/include/sys/sa_impl.h
@@ -235,7 +235,7 @@ struct sa_handle {
 #define	SA_BONUSTYPE_FROM_DB(db) \
 	(dmu_get_bonustype((dmu_buf_t *)db))
 
-#define	SA_BLKPTR_SPACE	(DN_MAX_BONUSLEN - sizeof (blkptr_t))
+#define	SA_BLKPTR_SPACE	(DN_OLD_MAX_BONUSLEN - sizeof (blkptr_t))
 
 #define	SA_LAYOUT_NUM(x, type) \
 	((!IS_SA_BONUSTYPE(type) ? 0 : (((IS_SA_BONUSTYPE(type)) && \
diff --git a/include/sys/spa.h b/include/sys/spa.h
index 7c09caf6c..17bf76de8 100644
--- a/include/sys/spa.h
+++ b/include/sys/spa.h
@@ -822,6 +822,7 @@ extern boolean_t spa_is_root(spa_t *spa);
 extern boolean_t spa_writeable(spa_t *spa);
 extern boolean_t spa_has_pending_synctask(spa_t *spa);
 extern int spa_maxblocksize(spa_t *spa);
+extern int spa_maxdnodesize(spa_t *spa);
 extern void zfs_blkptr_verify(spa_t *spa, const blkptr_t *bp);
 
 extern int spa_mode(spa_t *spa);
diff --git a/include/sys/zap.h b/include/sys/zap.h
index ed60b86db..70cbaae97 100644
--- a/include/sys/zap.h
+++ b/include/sys/zap.h
@@ -129,16 +129,30 @@ typedef enum zap_flags {
  *     MT_FIRST/MT_BEST matching will find entries that match without
  *     regard to case (eg. looking for "foo" can find an entry "Foo").
  * Eventually, other flags will permit unicode normalization as well.
+ *
+ * dnodesize specifies the on-disk size of the dnode for the new zapobj.
+ * Valid values are multiples of 512 up to DNODE_MAX_SIZE.
  */
 uint64_t zap_create(objset_t *ds, dmu_object_type_t ot,
     dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx);
+uint64_t zap_create_dnsize(objset_t *ds, dmu_object_type_t ot,
+    dmu_object_type_t bonustype, int bonuslen, int dnodesize, dmu_tx_t *tx);
 uint64_t zap_create_norm(objset_t *ds, int normflags, dmu_object_type_t ot,
     dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx);
+uint64_t zap_create_norm_dnsize(objset_t *ds, int normflags,
+    dmu_object_type_t ot, dmu_object_type_t bonustype, int bonuslen,
+    int dnodesize, dmu_tx_t *tx);
 uint64_t zap_create_flags(objset_t *os, int normflags, zap_flags_t flags,
     dmu_object_type_t ot, int leaf_blockshift, int indirect_blockshift,
     dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx);
+uint64_t zap_create_flags_dnsize(objset_t *os, int normflags,
+    zap_flags_t flags, dmu_object_type_t ot, int leaf_blockshift,
+    int indirect_blockshift, dmu_object_type_t bonustype, int bonuslen,
+    int dnodesize, dmu_tx_t *tx);
 uint64_t zap_create_link(objset_t *os, dmu_object_type_t ot,
     uint64_t parent_obj, const char *name, dmu_tx_t *tx);
+uint64_t zap_create_link_dnsize(objset_t *os, dmu_object_type_t ot,
+    uint64_t parent_obj, const char *name, int dnodesize, dmu_tx_t *tx);
 
 /*
  * Initialize an already-allocated object.
@@ -152,9 +166,14 @@ void mzap_create_impl(objset_t *os, uint64_t obj, int normflags,
  */
 int zap_create_claim(objset_t *ds, uint64_t obj, dmu_object_type_t ot,
     dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx);
+int zap_create_claim_dnsize(objset_t *ds, uint64_t obj, dmu_object_type_t ot,
+    dmu_object_type_t bonustype, int bonuslen, int dnodesize, dmu_tx_t *tx);
 int zap_create_claim_norm(objset_t *ds, uint64_t obj,
     int normflags, dmu_object_type_t ot,
     dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx);
+int zap_create_claim_norm_dnsize(objset_t *ds, uint64_t obj,
+    int normflags, dmu_object_type_t ot,
+    dmu_object_type_t bonustype, int bonuslen, int dnodesize, dmu_tx_t *tx);
 
 /*
  * The zapobj passed in must be a valid ZAP object for all of the
diff --git a/include/sys/zfs_ioctl.h b/include/sys/zfs_ioctl.h
index ac70f690e..301f83232 100644
--- a/include/sys/zfs_ioctl.h
+++ b/include/sys/zfs_ioctl.h
@@ -98,6 +98,7 @@ typedef enum drr_headertype {
 #define	DMU_BACKUP_FEATURE_EMBED_DATA_LZ4	(1<<17)
 /* flag #18 is reserved for a Delphix feature */
 #define	DMU_BACKUP_FEATURE_LARGE_BLOCKS		(1<<19)
+#define	DMU_BACKUP_FEATURE_LARGE_DNODE		(1<<20)
 
 /*
  * Mask of all supported backup features
@@ -105,7 +106,7 @@ typedef enum drr_headertype {
 #define	DMU_BACKUP_FEATURE_MASK	(DMU_BACKUP_FEATURE_DEDUP | \
     DMU_BACKUP_FEATURE_DEDUPPROPS | DMU_BACKUP_FEATURE_SA_SPILL | \
     DMU_BACKUP_FEATURE_EMBED_DATA | DMU_BACKUP_FEATURE_EMBED_DATA_LZ4 | \
-    DMU_BACKUP_FEATURE_LARGE_BLOCKS)
+    DMU_BACKUP_FEATURE_LARGE_BLOCKS | DMU_BACKUP_FEATURE_LARGE_DNODE)
 
 /* Are all features in the given flag word currently supported? */
 #define	DMU_STREAM_SUPPORTED(x)	(!((x) & ~DMU_BACKUP_FEATURE_MASK))
@@ -173,7 +174,8 @@ typedef struct dmu_replay_record {
 			uint32_t drr_bonuslen;
 			uint8_t drr_checksumtype;
 			uint8_t drr_compress;
-			uint8_t drr_pad[6];
+			uint8_t drr_dn_slots;
+			uint8_t drr_pad[5];
 			uint64_t drr_toguid;
 			/* bonus content follows */
 		} drr_object;
diff --git a/include/sys/zfs_znode.h b/include/sys/zfs_znode.h
index 43b591915..2c09feee1 100644
--- a/include/sys/zfs_znode.h
+++ b/include/sys/zfs_znode.h
@@ -196,6 +196,7 @@ typedef struct znode {
 	uint_t		z_blksz;	/* block size in bytes */
 	uint_t		z_seq;		/* modification sequence number */
 	uint64_t	z_mapcnt;	/* number of pages mapped to file */
+	uint64_t	z_dnodesize;	/* dnode size */
 	uint64_t	z_size;		/* file size (cached) */
 	uint64_t	z_links;	/* file links (cached) */
 	uint64_t	z_pflags;	/* pflags (cached) */
diff --git a/include/sys/zil.h b/include/sys/zil.h
index 65b14f1cd..1f04d1833 100644
--- a/include/sys/zil.h
+++ b/include/sys/zil.h
@@ -173,6 +173,19 @@ typedef enum zil_create {
 	(txtype) == TX_WRITE2)
 
 /*
+ * The number of dnode slots consumed by the object is stored in the 8
+ * unused upper bits of the object ID. We subtract 1 from the value
+ * stored on disk for compatibility with implementations that don't
+ * support large dnodes. The slot count for a single-slot dnode will
+ * contain 0 for those bits to preserve the log record format for
+ * "small" dnodes.
+ */
+#define	LR_FOID_GET_SLOTS(oid) (BF64_GET((oid), 56, 8) + 1)
+#define	LR_FOID_SET_SLOTS(oid, x) BF64_SET((oid), 56, 8, (x) - 1)
+#define	LR_FOID_GET_OBJ(oid) BF64_GET((oid), 0, DN_MAX_OBJECT_SHIFT)
+#define	LR_FOID_SET_OBJ(oid, x) BF64_SET((oid), 0, DN_MAX_OBJECT_SHIFT, (x))
+
+/*
  * Format of log records.
  * The fields are carefully defined to allow them to be aligned
  * and sized the same on sparc & intel architectures.