[RFC PATCH 06/12 v2] Attach CPU topology to specify each sched_domain's workload consolidation

[Yuyang Du <yuyang.du@intel.com>]

Defined SD_WORKLOAD_CONSOLIDATION flag in sched_domain. When this flag is
set, the workload consolidation applies to that domain. In addition, a
consolidating_coeff is defined in sched_domain to specify the degree of
consolidation in that domain.

Signed-off-by: Yuyang Du <yuyang.du@intel.com>
---
 include/linux/sched.h    |   13 +++++++++++++
 include/linux/topology.h |   16 ++++++++++++++++
 kernel/sched/core.c      |   41 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 70 insertions(+)

diff --git a/include/linux/sched.h b/include/linux/sched.h
index 25f54c7..f3f7d4a 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -877,6 +877,12 @@ enum cpu_idle_type {
 #define SD_OVERLAP		0x2000	/* sched_domains of this level overlap */
 #define SD_NUMA			0x4000	/* cross-node balancing */
 
+#ifdef CONFIG_WORKLOAD_CONSOLIDATION
+#define SD_WORKLOAD_CONSOLIDATION	0x8000	/* Higher concurrency in front */
+#else
+#define SD_WORKLOAD_CONSOLIDATION	0
+#endif
+
 extern int __weak arch_sd_sibiling_asym_packing(void);
 
 struct sched_domain_attr {
@@ -960,6 +966,13 @@ struct sched_domain {
 		struct rcu_head rcu;	/* used during destruction */
 	};
 
+#ifdef CONFIG_WORKLOAD_CONSOLIDATION
+	unsigned int total_groups;			/* total groups number */
+	unsigned int group_number;			/* this CPU's group sequence */
+	unsigned int consolidating_coeff;	/* consolidating coefficient */
+	struct sched_group *first_group;	/* ordered by CPU number */
+#endif
+
 	unsigned int span_weight;
 	/*
 	 * Span of all CPUs in this domain.
diff --git a/include/linux/topology.h b/include/linux/topology.h
index 7062330..334f83e 100644
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -66,6 +66,16 @@ int arch_update_cpu_topology(void);
 #define PENALTY_FOR_NODE_WITH_CPUS	(1)
 #endif
 
+#ifdef CONFIG_WORKLOAD_CONSOLIDATION
+#ifndef WORKLOAD_CONSOLIDATION_INIT
+#define WORKLOAD_CONSOLIDATION_INIT(n) .consolidating_coeff = (n),
+#endif
+#else
+#ifndef WORKLOAD_CONSOLIDATION_INIT
+#define WORKLOAD_CONSOLIDATION_INIT(n)
+#endif
+#endif
+
 /*
  * Below are the 3 major initializers used in building sched_domains:
  * SD_SIBLING_INIT, for SMT domains
@@ -102,12 +112,14 @@ int arch_update_cpu_topology(void);
 				| 0*SD_SERIALIZE			\
 				| 0*SD_PREFER_SIBLING			\
 				| arch_sd_sibling_asym_packing()	\
+				| 0*SD_WORKLOAD_CONSOLIDATION	\
 				,					\
 	.last_balance		= jiffies,				\
 	.balance_interval	= 1,					\
 	.smt_gain		= 1178,	/* 15% */			\
 	.max_newidle_lb_cost	= 0,					\
 	.next_decay_max_lb_cost	= jiffies,				\
+	WORKLOAD_CONSOLIDATION_INIT(0)				\
 }
 #endif
 #endif /* CONFIG_SCHED_SMT */
@@ -134,11 +146,13 @@ int arch_update_cpu_topology(void);
 				| 0*SD_SHARE_CPUPOWER			\
 				| 1*SD_SHARE_PKG_RESOURCES		\
 				| 0*SD_SERIALIZE			\
+				| 1*SD_WORKLOAD_CONSOLIDATION	\
 				,					\
 	.last_balance		= jiffies,				\
 	.balance_interval	= 1,					\
 	.max_newidle_lb_cost	= 0,					\
 	.next_decay_max_lb_cost	= jiffies,				\
+	WORKLOAD_CONSOLIDATION_INIT(180)			\
 }
 #endif
 #endif /* CONFIG_SCHED_MC */
@@ -167,11 +181,13 @@ int arch_update_cpu_topology(void);
 				| 0*SD_SHARE_PKG_RESOURCES		\
 				| 0*SD_SERIALIZE			\
 				| 1*SD_PREFER_SIBLING			\
+				| 1*SD_WORKLOAD_CONSOLIDATION	\
 				,					\
 	.last_balance		= jiffies,				\
 	.balance_interval	= 1,					\
 	.max_newidle_lb_cost	= 0,					\
 	.next_decay_max_lb_cost	= jiffies,				\
+	WORKLOAD_CONSOLIDATION_INIT(180)			\
 }
 #endif
 
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 0236455..cd92f2d 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4843,7 +4843,11 @@ set_table_entry(struct ctl_table *entry,
 static struct ctl_table *
 sd_alloc_ctl_domain_table(struct sched_domain *sd)
 {
+#ifdef CONFIG_WORKLOAD_CONSOLIDATION
+	struct ctl_table *table = sd_alloc_ctl_entry(15);
+#else
 	struct ctl_table *table = sd_alloc_ctl_entry(14);
+#endif
 
 	if (table == NULL)
 		return NULL;
@@ -4876,7 +4880,13 @@ sd_alloc_ctl_domain_table(struct sched_domain *sd)
 		sizeof(long), 0644, proc_doulongvec_minmax, false);
 	set_table_entry(&table[12], "name", sd->name,
 		CORENAME_MAX_SIZE, 0444, proc_dostring, false);
+#ifdef CONFIG_WORKLOAD_CONSOLIDATION
+	set_table_entry(&table[13], "consolidating_coeff", &sd->consolidating_coeff,
+		sizeof(int), 0644, proc_dointvec, false);
+	/* &table[14] is terminator */
+#else
 	/* &table[13] is terminator */
+#endif
 
 	return table;
 }
@@ -5497,6 +5507,33 @@ static void update_top_cache_domain(int cpu)
 	rcu_assign_pointer(per_cpu(sd_asym, cpu), sd);
 }
 
+#ifdef CONFIG_WORKLOAD_CONSOLIDATION
+static void update_domain_extra_info(struct sched_domain *sd)
+{
+	while (sd) {
+		int i = 0, j = 0, first, min = INT_MAX;
+		struct sched_group *group;
+
+		group = sd->groups;
+		first = group_first_cpu(group);
+		do {
+			int k = group_first_cpu(group);
+			i += 1;
+			if (k < first)
+				j += 1;
+			if (k < min) {
+				sd->first_group = group;
+				min = k;
+			}
+		} while (group = group->next, group != sd->groups);
+
+		sd->total_groups = i;
+		sd->group_number = j;
+		sd = sd->parent;
+	}
+}
+#endif
+
 /*
  * Attach the domain 'sd' to 'cpu' as its base domain. Callers must
  * hold the hotplug lock.
@@ -5545,6 +5582,10 @@ cpu_attach_domain(struct sched_domain *sd, struct root_domain *rd, int cpu)
 	destroy_sched_domains(tmp, cpu);
 
 	update_top_cache_domain(cpu);
+
+#ifdef CONFIG_WORKLOAD_CONSOLIDATION
+	update_domain_extra_info(sd);
+#endif
 }
 
 /* cpus with isolated domains */
-- 
1.7.9.5