dmu_objset: replace dnode_hash impl with cityhash4

As mentioned in PR openzfs#16131, replacing CRC-based hash with cityhash4 could slightly improve the performance by eliminating memory access. Replacing algorightm is safe since the hash result is not persisted. Reviewed by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: Tino Reichardt <[email protected]> Signed-off-by: Shengqi Chen <[email protected]> Closes openzfs#16131 Closes openzfs#16483
openzfsonwindows · Jan 22, 2025 · f92819b · f92819b
1 parent 9e1e8a2
commit f92819b
Showing 1 changed file with 3 additions and 16 deletions.
diff --git a/module/zfs/dmu_objset.c b/module/zfs/dmu_objset.c
@@ -66,6 +66,7 @@
 #include "zfs_namecheck.h"
 #include <sys/vdev_impl.h>
 #include <sys/arc.h>
+#include <cityhash.h>
 
 /*
  * Needed to close a window in dnode_move() that allows the objset to be freed
@@ -404,27 +405,13 @@ dmu_objset_byteswap(void *buf, size_t size)
 }
 
 /*
- * The hash is a CRC-based hash of the objset_t pointer and the object number.
+ * Runs cityhash4 on the objset_t pointer and the object number.
  */
 static uint64_t
 dnode_hash(const objset_t *os, uint64_t obj)
 {
 	uintptr_t osv = (uintptr_t)os;
-	uint64_t crc = -1ULL;
-
-	ASSERT(zfs_crc64_table[128] == ZFS_CRC64_POLY);
-	/*
-	 * The lower 11 bits of the pointer don't have much entropy, because
-	 * the objset_t is more than 1KB long and so likely aligned to 2KB.
-	 */
-	crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (osv >> 11)) & 0xFF];
-	crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (obj >> 0)) & 0xFF];
-	crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (obj >> 8)) & 0xFF];
-	crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (obj >> 16)) & 0xFF];
-
-	crc ^= (osv>>14) ^ (obj>>24);
-
-	return (crc);
+	return (cityhash4((uint64_t)osv, obj, 0, 0));
 }
 
 static unsigned int