diff options
| author | Benjamin LaHaise <bcrl@linux.intel.com> | 2006-02-01 03:05:30 -0800 |
|---|---|---|
| committer | Linus Torvalds <torvalds@g5.osdl.org> | 2006-02-01 08:53:15 -0800 |
| commit | 9884fd8df195fe48d4e1be2279b419be96127cae (patch) | |
| tree | 1438571d206c1d488efb8454c64ae00c51438a2e /mm/vmscan.c | |
| parent | c84db23c6e587d3ab00a41c51fedf758e1f6ecd4 (diff) | |
| download | op-kernel-dev-9884fd8df195fe48d4e1be2279b419be96127cae.zip op-kernel-dev-9884fd8df195fe48d4e1be2279b419be96127cae.tar.gz | |
[PATCH] Use 32 bit division in slab_put_obj()
Improve the performance of slab_put_obj(). Without the cast, gcc considers
ptrdiff_t a 64 bit signed integer and ends up emitting code to use a full
signed 128 bit divide on EM64T, which is substantially slower than a 32 bit
unsigned divide.
I noticed this when looking at the profile of a case where the slab balance
is just on edge and thrashes back and forth freeing a block.
Signed-off-by: Benjamin LaHaise <benjamin.c.lahaise@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'mm/vmscan.c')
0 files changed, 0 insertions, 0 deletions
