[RESEND PATCH v5 2/2] sched/fair: Scan cluster before scanning LLC in wake-up path

[Yicong Yang]

From: Barry Song <song.bao.hua at hisilicon.com>

For platforms having clusters like Kunpeng920, CPUs within the same cluster
have lower latency when synchronizing and accessing shared resources like
cache. Thus, this patch tries to find an idle cpu within the cluster of the
target CPU before scanning the whole LLC to gain lower latency.

Note neither Kunpeng920 nor x86 Jacobsville supports SMT, so this patch
doesn't consider SMT for this moment.

Testing has been done on Kunpeng920 by pinning tasks to one numa and two
numa. On Kunpeng920, Each numa has 8 clusters and each cluster has 4 CPUs.

With this patch, We noticed enhancement on tbench within one numa or cross
two numa.

On numa 0:
                           tip/core                 patched
Hmean     1        345.89 (   0.00%)      393.96 *  13.90%*
Hmean     2        697.77 (   0.00%)      786.04 *  12.65%*
Hmean     4       1392.51 (   0.00%)     1570.26 *  12.76%*
Hmean     8       2800.61 (   0.00%)     3083.98 *  10.12%*
Hmean     16      5514.27 (   0.00%)     6116.00 *  10.91%*
Hmean     32     10869.81 (   0.00%)    10782.98 *  -0.80%*
Hmean     64      8315.22 (   0.00%)     8519.84 *   2.46%*
Hmean     128     6324.47 (   0.00%)     7159.35 *  13.20%*

On numa 0-1:
                           tip/core                 patched
Hmean     1        348.68 (   0.00%)      387.91 *  11.25%*
Hmean     2        693.57 (   0.00%)      774.91 *  11.73%*
Hmean     4       1369.26 (   0.00%)     1475.48 *   7.76%*
Hmean     8       2772.99 (   0.00%)     2984.61 *   7.63%*
Hmean     16      4825.83 (   0.00%)     5873.13 *  21.70%*
Hmean     32     10250.32 (   0.00%)    11688.06 *  14.03%*
Hmean     64     16309.51 (   0.00%)    19889.48 *  21.95%*
Hmean     128    13022.32 (   0.00%)    16005.64 *  22.91%*
Hmean     256    11335.79 (   0.00%)    13821.74 *  21.93%*

Tested-by: Yicong Yang <yangyicong at hisilicon.com>
Signed-off-by: Barry Song <song.bao.hua at hisilicon.com>
Signed-off-by: Yicong Yang <yangyicong at hisilicon.com>
Reviewed-by: Tim Chen <tim.c.chen at linux.intel.com>
---
 kernel/sched/fair.c | 44 +++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 41 insertions(+), 3 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 914096c5b1ae..25d2900ae221 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -6377,6 +6377,40 @@ static inline int select_idle_smt(struct task_struct *p, struct sched_domain *sd
 
 #endif /* CONFIG_SCHED_SMT */
 
+#ifdef CONFIG_SCHED_CLUSTER
+/*
+ * Scan the cluster domain for idle CPUs and clear cluster cpumask after scanning
+ */
+static inline int scan_cluster(struct task_struct *p, struct cpumask *cpus,
+			       int target, int *nr)
+{
+	struct sched_domain *sd = rcu_dereference(per_cpu(sd_cluster, target));
+	int cpu, idle_cpu;
+
+	/* TODO: Support SMT system with cluster topology */
+	if (!sched_smt_active() && sd) {
+		for_each_cpu_and(cpu, cpus, sched_domain_span(sd)) {
+			if (!--*nr)
+				return -1;
+
+			idle_cpu = __select_idle_cpu(cpu, p);
+			if ((unsigned int)idle_cpu < nr_cpumask_bits)
+				return idle_cpu;
+		}
+
+		cpumask_andnot(cpus, cpus, sched_domain_span(sd));
+	}
+
+	return -1;
+}
+#else
+static inline int scan_cluster(struct task_struct *p, struct cpumask *cpus,
+			       int target, int *nr)
+{
+	return -1;
+}
+#endif
+
 /*
  * Scan the LLC domain for idle CPUs; this is dynamically regulated by
  * comparing the average scan cost (tracked in sd->avg_scan_cost) against the
@@ -6437,6 +6471,10 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, bool
 		}
 	}
 
+	idle_cpu = scan_cluster(p, cpus, target, &nr);
+	if ((unsigned int)idle_cpu < nr_cpumask_bits)
+		return idle_cpu;
+
 	for_each_cpu_wrap(cpu, cpus, target + 1) {
 		if (has_idle_core) {
 			i = select_idle_core(p, cpu, cpus, &idle_cpu);
@@ -6444,7 +6482,7 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, bool
 				return i;
 
 		} else {
-			if (!--nr)
+			if (--nr <= 0)
 				return -1;
 			idle_cpu = __select_idle_cpu(cpu, p);
 			if ((unsigned int)idle_cpu < nr_cpumask_bits)
@@ -6543,7 +6581,7 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 	/*
 	 * If the previous CPU is cache affine and idle, don't be stupid:
 	 */
-	if (prev != target && cpus_share_cache(prev, target) &&
+	if (prev != target && cpus_share_lowest_cache(prev, target) &&
 	    (available_idle_cpu(prev) || sched_idle_cpu(prev)) &&
 	    asym_fits_capacity(task_util, prev))
 		return prev;
@@ -6569,7 +6607,7 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target)
 	p->recent_used_cpu = prev;
 	if (recent_used_cpu != prev &&
 	    recent_used_cpu != target &&
-	    cpus_share_cache(recent_used_cpu, target) &&
+	    cpus_share_lowest_cache(recent_used_cpu, target) &&
 	    (available_idle_cpu(recent_used_cpu) || sched_idle_cpu(recent_used_cpu)) &&
 	    cpumask_test_cpu(p->recent_used_cpu, p->cpus_ptr) &&
 	    asym_fits_capacity(task_util, recent_used_cpu)) {
-- 
2.24.0