]> www.pilppa.org Git - linux-2.6-omap-h63xx.git/commit
projects / linux-2.6-omap-h63xx.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 9f4cbec) | patch
[XFS] On machines with more than 8 cpus, when running parallel I/O
authorDavid Chinner <dgc@sgi.com>
Tue, 14 Mar 2006 02:13:09 +0000 (13:13 +1100)
committerNathan Scott <nathans@sgi.com>
Tue, 14 Mar 2006 02:13:09 +0000 (13:13 +1100)
commit8d280b98cfe3c0b69c37d355218975c1c0279bb0
tree2dc1deaec23a7da29b72152a4225c2600dacf1d4tree | snapshot
parent9f4cbecd7e5ee6390fecd6032dc04ca8c9805dc9commit | diff
[XFS] On machines with more than 8 cpus, when running parallel I/O
threads, the incore superblock lock becomes the limiting factor for
buffered write throughput. Make the contended fields in the incore
superblock use per-cpu counters so that there is no global lock to limit
scalability.

SGI-PV: 946630
SGI-Modid: xfs-linux-melb:xfs-kern:25106a

Signed-off-by: David Chinner <dgc@sgi.com>
Signed-off-by: Nathan Scott <nathans@sgi.com>
fs/xfs/linux-2.6/xfs_linux.h diff | blob | history
fs/xfs/xfs_fsops.c diff | blob | history
fs/xfs/xfs_mount.c diff | blob | history
fs/xfs/xfs_mount.h diff | blob | history
fs/xfs/xfs_vfsops.c diff | blob | history
Unnamed repository; edit this file 'description' to name the repository.