xfs: optimize log flushing in xfs_fsync
authorChristoph Hellwig <hch@infradead.org>
Fri, 5 Feb 2010 09:57:55 +0000 (09:57 +0000)
committerAlex Elder <aelder@sgi.com>
Fri, 12 Feb 2010 19:45:14 +0000 (13:45 -0600)
commit180040b89ee2aed88c0a0b1fcf7ada9a512b12e3
treeddc3f6aead139aad5d5c5e0e786fb0b91090505a
parent87185517de81101da5afbc82cefdeed6eeaa38fb
xfs: optimize log flushing in xfs_fsync

If we have a pinned inode it must have a log item attached to it.
Usually that log item will have ili_last_lsn already set, in which
case we only need to flush the log up to that LSN instead of doing a
full log force.  This gives speedups of about 5% in some fsync heavy
workloads.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Alex Elder <aelder@sgi.com>
fs/xfs/xfs_vnodeops.c