sched/fair: Remove SIS_AVG_CPU
authorMel Gorman <mgorman@techsingularity.net>
Mon, 25 Jan 2021 08:59:06 +0000 (08:59 +0000)
committerPeter Zijlstra <peterz@infradead.org>
Wed, 27 Jan 2021 16:26:43 +0000 (17:26 +0100)
commite6e0dc2d5497f7f3ed970052917e2923c6f453f4
treeed7ddaf907a849aa8e16e5c932b6d97f79cbf6af
parent1875dc5b8ff4690547c446ef222083e28e2d9463
sched/fair: Remove SIS_AVG_CPU

SIS_AVG_CPU was introduced as a means of avoiding a search when the
average search cost indicated that the search would likely fail. It was
a blunt instrument and disabled by commit 4c77b18cf8b7 ("sched/fair: Make
select_idle_cpu() more aggressive") and later replaced with a proportional
search depth by commit 1ad3aaf3fcd2 ("sched/core: Implement new approach
to scale select_idle_cpu()").

While there are corner cases where SIS_AVG_CPU is better, it has now been
disabled for almost three years. As the intent of SIS_PROP is to reduce
the time complexity of select_idle_cpu(), lets drop SIS_AVG_CPU and focus
on SIS_PROP as a throttling mechanism.

Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/20210125085909.4600-2-mgorman@techsingularity.net
kernel/sched/fair.c
kernel/sched/features.h