diff options
| author | Andrew Morton <akpm@osdl.org> | 2004-05-09 23:29:07 -0700 |
|---|---|---|
| committer | Linus Torvalds <torvalds@ppc970.osdl.org> | 2004-05-09 23:29:07 -0700 |
| commit | a690c9b7ac0ddc28785c38526a69a7fe2e692500 (patch) | |
| tree | 34cb700c99bd558918a7735e39518f75f3dd5f4e /kernel | |
| parent | df65cdbf6bc82e79658c35e14221fef34f4db6d0 (diff) | |
| download | history-a690c9b7ac0ddc28785c38526a69a7fe2e692500.tar.gz | |
[PATCH] sched: cpu load management cleanup
From: Ingo Molnar <mingo@elte.hu>
This does the source/target cleanup. This is a no-functionality patch which
also adds more comments to explain these functions.
Diffstat (limited to 'kernel')
| -rw-r--r-- | kernel/sched.c | 26 |
1 files changed, 16 insertions, 10 deletions
diff --git a/kernel/sched.c b/kernel/sched.c index c6d7192eae5afa..90ee4fb0b60fd5 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -645,9 +645,12 @@ void kick_process(task_t *p) EXPORT_SYMBOL_GPL(kick_process); /* - * Return a low guess at the load of cpu. + * Return a low guess at the load of a migration-source cpu. + * + * We want to under-estimate the load of migration sources, to + * balance conservatively. */ -static inline unsigned long get_low_cpu_load(int cpu) +static inline unsigned long source_load(int cpu) { runqueue_t *rq = cpu_rq(cpu); unsigned long load_now = rq->nr_running * SCHED_LOAD_SCALE; @@ -655,7 +658,10 @@ static inline unsigned long get_low_cpu_load(int cpu) return min(rq->cpu_load, load_now); } -static inline unsigned long get_high_cpu_load(int cpu) +/* + * Return a high guess at the load of a migration-target cpu + */ +static inline unsigned long target_load(int cpu) { runqueue_t *rq = cpu_rq(cpu); unsigned long load_now = rq->nr_running * SCHED_LOAD_SCALE; @@ -751,8 +757,8 @@ static int try_to_wake_up(task_t * p, unsigned int state, int sync) if (cpu == this_cpu || unlikely(!cpu_isset(this_cpu, p->cpus_allowed))) goto out_set_cpu; - load = get_low_cpu_load(cpu); - this_load = get_high_cpu_load(this_cpu); + load = source_load(cpu); + this_load = target_load(this_cpu); /* Don't pull the task off an idle CPU to a busy one */ if (load < SCHED_LOAD_SCALE/2 && this_load > SCHED_LOAD_SCALE/2) @@ -1207,9 +1213,9 @@ static int sched_best_cpu(struct task_struct *p, struct sched_domain *sd) for_each_cpu_mask(i, tmp) { unsigned long load; if (i == this_cpu) - load = get_low_cpu_load(i); + load = source_load(i); else - load = get_high_cpu_load(i) + SCHED_LOAD_SCALE; + load = target_load(i) + SCHED_LOAD_SCALE; if (min_load > load) { best_cpu = i; @@ -1426,9 +1432,9 @@ find_busiest_group(struct sched_domain *sd, int this_cpu, for_each_cpu_mask(i, tmp) { /* Bias balancing toward cpus of our domain */ if (local_group) - load = get_high_cpu_load(i); + load = target_load(i); else - load = get_low_cpu_load(i); + load = source_load(i); nr_cpus++; avg_load += load; @@ -1548,7 +1554,7 @@ static runqueue_t *find_busiest_queue(struct sched_group *group) cpus_and(tmp, group->cpumask, cpu_online_map); for_each_cpu_mask(i, tmp) { - load = get_low_cpu_load(i); + load = source_load(i); if (load > max_load) { max_load = load; |