diff options
| author | Ricardo Neri <ricardo.neri-calderon@linux.intel.com> | 2023-04-06 13:31:40 -0700 |
|---|---|---|
| committer | Peter Zijlstra <peterz@infradead.org> | 2023-05-08 10:58:35 +0200 |
| commit | 18ad34532755feb5b9f4284b07769b1bfec18ab3 (patch) | |
| tree | 40dce503a9ee80901b07c6895434250ca9efd96f /kernel/sched | |
| parent | ef7657d4d2d6a8456aa624010de456c32a135fe9 (diff) | |
| download | linux-next-18ad34532755feb5b9f4284b07769b1bfec18ab3.tar.gz | |
sched/fair: Let low-priority cores help high-priority busy SMT cores
Using asym_packing priorities within an SMT core is straightforward. Just
follow the priorities that hardware indicates.
When balancing load from an SMT core, also consider the idle state of its
siblings. Priorities do not reflect that an SMT core divides its throughput
among all its busy siblings. They only makes sense when exactly one sibling
is busy.
Indicate that active balance is needed if the destination CPU has lower
priority than the source CPU but the latter has busy SMT siblings.
Make find_busiest_queue() not skip higher-priority SMT cores with more than
busy sibling.
Suggested-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-5-ricardo.neri-calderon@linux.intel.com
Diffstat (limited to 'kernel/sched')
| -rw-r--r-- | kernel/sched/fair.c | 16 |
1 files changed, 14 insertions, 2 deletions
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index a8a02ae7d082..85ce2494234d 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -10551,8 +10551,15 @@ static struct rq *find_busiest_queue(struct lb_env *env, nr_running == 1) continue; - /* Make sure we only pull tasks from a CPU of lower priority */ + /* + * Make sure we only pull tasks from a CPU of lower priority + * when balancing between SMT siblings. + * + * If balancing between cores, let lower priority CPUs help + * SMT cores with more than one busy sibling. + */ if ((env->sd->flags & SD_ASYM_PACKING) && + sched_use_asym_prio(env->sd, i) && sched_asym_prefer(i, env->dst_cpu) && nr_running == 1) continue; @@ -10645,10 +10652,15 @@ asym_active_balance(struct lb_env *env) * priority CPUs in order to pack all tasks in the highest priority * CPUs. When done between cores, do it only if the whole core if the * whole core is idle. + * + * If @env::src_cpu is an SMT core with busy siblings, let + * the lower priority @env::dst_cpu help it. Do not follow + * CPU priority. */ return env->idle != CPU_NOT_IDLE && (env->sd->flags & SD_ASYM_PACKING) && sched_use_asym_prio(env->sd, env->dst_cpu) && - sched_asym_prefer(env->dst_cpu, env->src_cpu); + (sched_asym_prefer(env->dst_cpu, env->src_cpu) || + !sched_use_asym_prio(env->sd, env->src_cpu)); } static inline bool |