sched/fair: Don't set LBF_ALL_PINNED unnecessarily
authorVincent Guittot <vincent.guittot@linaro.org>
Thu, 7 Jan 2021 10:33:24 +0000 (11:33 +0100)
committerPeter Zijlstra <peterz@infradead.org>
Thu, 14 Jan 2021 10:20:11 +0000 (11:20 +0100)
commit8a41dfcda7a32ed4435c00d98a9dc7156b08b671
tree5458b36c971aadd7c3a29e4a30678d88c5c5db97
parentfc488ffd4297f661b3e9d7450dcdb9089a53df7c
sched/fair: Don't set LBF_ALL_PINNED unnecessarily

Setting LBF_ALL_PINNED during active load balance is only valid when there
is only 1 running task on the rq otherwise this ends up increasing the
balance interval whereas other tasks could migrate after the next interval
once they become cache-cold as an example.

LBF_ALL_PINNED flag is now always set it by default. It is then cleared
when we find one task that can be pulled when calling detach_tasks() or
during active migration.

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Link: https://lkml.kernel.org/r/20210107103325.30851-3-vincent.guittot@linaro.org
kernel/sched/fair.c