Skip to content

Commit

Permalink
Increase ZFS_OBJ_MTX_SZ further
Browse files Browse the repository at this point in the history
Larger values have been associated with better concurrency and less
hanging of memory reclaim operations. So raise this number
further.

Too big and you get memory allocation failures. This number (767)
was chosen semi-arbitrarily but 3000+ has been shown to be too high.
  • Loading branch information
DeHackEd authored and DHE committed Feb 3, 2015
1 parent 56330f4 commit 359accf
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion include/sys/zfs_vfsops.h
Original file line number Diff line number Diff line change
Expand Up @@ -90,7 +90,7 @@ typedef struct zfs_sb {
uint64_t z_groupquota_obj;
uint64_t z_replay_eof; /* New end of file - replay only */
sa_attr_type_t *z_attr_table; /* SA attr mapping->id */
#define ZFS_OBJ_MTX_SZ 256
#define ZFS_OBJ_MTX_SZ 767
kmutex_t z_hold_mtx[ZFS_OBJ_MTX_SZ]; /* znode hold locks */
} zfs_sb_t;

Expand Down

0 comments on commit 359accf

Please sign in to comment.