Avoid 64bit division in multilist index functions

The number of sublists in a multilist is relatively small. We dont need
64 bits to calculate an index. 32 bits is sufficient and makes the
code more efficient.

Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> 
Reviewed-by: Mark Maybee <mark.maybee@delphix.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes #12288

module/zfs/arc.c

@@ -7459,9 +7459,10 @@ arc_state_multilist_index_func(multilist_t *ml, void *obj)
 	 * Also, the low order bits of the hash value are thought to be
 	 * distributed evenly. Otherwise, in the case that the multilist
 	 * has a power of two number of sublists, each sublists' usage
-	 * would not be evenly distributed.
+	 * would not be evenly distributed. In this context full 64bit
+	 * division would be a waste of time, so limit it to 32 bits.
 	 */
-	return (buf_hash(hdr->b_spa, &hdr->b_dva, hdr->b_birth) %
+	return ((unsigned int)buf_hash(hdr->b_spa, &hdr->b_dva, hdr->b_birth) %
 	    multilist_get_num_sublists(ml));
 }
 

module/zfs/dbuf.c

@@ -622,9 +622,10 @@ dbuf_cache_multilist_index_func(multilist_t *ml, void *obj)
 	 * Also, the low order bits of the hash value are thought to be
 	 * distributed evenly. Otherwise, in the case that the multilist
 	 * has a power of two number of sublists, each sublists' usage
-	 * would not be evenly distributed.
+	 * would not be evenly distributed. In this context full 64bit
+	 * division would be a waste of time, so limit it to 32 bits.
 	 */
-	return (dbuf_hash(db->db_objset, db->db.db_object,
+	return ((unsigned int)dbuf_hash(db->db_objset, db->db.db_object,
 	    db->db_level, db->db_blkid) %
 	    multilist_get_num_sublists(ml));
 }

module/zfs/dmu_objset.c

@@ -399,7 +399,15 @@ static unsigned int
 dnode_multilist_index_func(multilist_t *ml, void *obj)
 {
 	dnode_t *dn = obj;
-	return (dnode_hash(dn->dn_objset, dn->dn_object) %
+
+	/*
+	 * The low order bits of the hash value are thought to be
+	 * distributed evenly. Otherwise, in the case that the multilist
+	 * has a power of two number of sublists, each sublists' usage
+	 * would not be evenly distributed. In this context full 64bit
+	 * division would be a waste of time, so limit it to 32 bits.
+	 */
+	return ((unsigned int)dnode_hash(dn->dn_objset, dn->dn_object) %
 	    multilist_get_num_sublists(ml));
 }
 

module/zfs/metaslab.c

@@ -1874,7 +1874,12 @@ static unsigned int
 metaslab_idx_func(multilist_t *ml, void *arg)
 {
 	metaslab_t *msp = arg;
-	return (msp->ms_id % multilist_get_num_sublists(ml));
+
+	/*
+	 * ms_id values are allocated sequentially, so full 64bit
+	 * division would be a waste of time, so limit it to 32 bits.
+	 */
+	return ((unsigned int)msp->ms_id % multilist_get_num_sublists(ml));
 }
 
 uint64_t