projects / linux.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3999e49) | patch
locks: give the blocked_hash its own spinlock
authorJeff Layton <jlayton@redhat.com>
Fri, 21 Jun 2013 12:58:20 +0000 (08:58 -0400)
committerAl Viro <viro@zeniv.linux.org.uk>
Sat, 29 Jun 2013 08:57:46 +0000 (12:57 +0400)
commit7b2296afb392bc21a50f42e7c7f4b19d3fea8c6d
tree360a8f35cf75f0bbcd1b984a6348f4c9e715e159tree | snapshot
parent3999e49364193f7dbbba66e2be655fe91ba1fcedcommit | diff
locks: give the blocked_hash its own spinlock

There's no reason we have to protect the blocked_hash and file_lock_list
with the same spinlock. With the tests I have, breaking it in two gives
a barely measurable performance benefit, but it seems reasonable to make
this locking as granular as possible.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Documentation/filesystems/Locking diff | blob | history
fs/locks.c diff | blob | history
Linux Kernel tree with some unpolished code for Colibri/Apalis modules