sched/numa: Add SD_PERFER_SIBLING to CPU domain
authorAlex Shi <alex.shi@intel.com>
Fri, 20 Jul 2012 06:19:50 +0000 (14:19 +0800)
committerIngo Molnar <mingo@kernel.org>
Thu, 26 Jul 2012 09:46:58 +0000 (11:46 +0200)
commit6956dc568f34107f1d02b24f87efe7250803fc87
tree6372cf6bbb73c0afd0f4b449344915b846956d67
parent8323f26ce3425460769605a6aece7a174edaa7d1
sched/numa: Add SD_PERFER_SIBLING to CPU domain

Commit 8e7fbcbc22c ("sched: Remove stale power aware scheduling remnants
and dysfunctional knobs") removed SD_PERFER_SIBLING from the CPU domain.

On NUMA machines this causes that load_balance() doesn't perfer LCPU in
 same physical CPU package.

It causes some actual performance regressions on our NUMA machines from
Core2 to NHM and SNB.

Adding this domain flag again recovers the performance drop.

This change doesn't have any bad impact on any of my benchmarks:
 specjbb, kbuild, fio, hackbench .. etc, on all my machines.

Signed-off-by: Alex Shi <alex.shi@intel.com>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1342765190-21540-1-git-send-email-alex.shi@intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
include/linux/topology.h