Few microoptimizations to dbuf layer

Move db_link into the same cache line as db_blkid and db_level.
It allows significantly reduce avl_add() time in dbuf_create() on
systems with large RAM and huge number of dbufs per dnode.

Avoid few accesses to dbuf_caches[].size, which is highly congested
under high IOPS and never stays in cache for a long time.  Use local
value we are receiving from zfs_refcount_add_many() any way.

Remove cache_size_bytes_max bump from dbuf_evict_one().  I don't see
a point to do it on dbuf eviction after we done it on insertion in
dbuf_rele_and_unlock().

Reviewed-by: Matt Ahrens <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: Alexander Motin <[email protected]>
Sponsored-By: iXsystems, Inc.
Closes #9931 

1 files changed, 7 insertions, 6 deletions

diff --git a/include/sys/dbuf.h b/include/sys/dbuf.h
index f6880cdb2..d04e08baa 100644
--- a/include/sys/dbuf.h
+++ b/include/sys/dbuf.h
@@ -206,6 +206,13 @@ typedef struct dmu_buf_impl {
 	 */
 	struct dmu_buf_impl *db_hash_next;
 
+	/*
+	 * Our link on the owner dnodes's dn_dbufs list.
+	 * Protected by its dn_dbufs_mtx.  Should be on the same cache line
+	 * as db_level and db_blkid for the best avl_add() performance.
+	 */
+	avl_node_t db_link;
+
 	/* our block number */
 	uint64_t db_blkid;
 
@@ -260,12 +267,6 @@ typedef struct dmu_buf_impl {
 	/* List of dirty records for the buffer sorted newest to oldest. */
 	list_t db_dirty_records;
 
-	/*
-	 * Our link on the owner dnodes's dn_dbufs list.
-	 * Protected by its dn_dbufs_mtx.
-	 */
-	avl_node_t db_link;
-
 	/* Link in dbuf_cache or dbuf_metadata_cache */
 	multilist_node_t db_cache_link;