@@ -46,7 +46,7 @@
void build_cpu_to_node_map(void);
-#define SD_CPU_INIT (struct sched_domain) { \
+#define SD_CPU_INIT(cpu) (struct sched_domain) { \
.parent = NULL, \
.child = NULL, \
.groups = NULL, \
@@ -65,6 +65,7 @@ void build_cpu_to_node_map(void);
| SD_BALANCE_EXEC \
| SD_BALANCE_FORK \
| SD_WAKE_AFFINE, \
+ | arch_sd_local_flags(0, cpu)\
.last_balance = jiffies, \
.balance_interval = 1, \
.nr_balance_failed = 0, \
@@ -51,7 +51,7 @@ static inline const struct cpumask *cpumask_of_node(int node)
*/
/* sched_domains SD_CPU_INIT for TILE architecture */
-#define SD_CPU_INIT (struct sched_domain) { \
+#define SD_CPU_INIT(cpu) (struct sched_domain) { \
.min_interval = 4, \
.max_interval = 128, \
.busy_factor = 64, \
@@ -71,6 +71,7 @@ static inline const struct cpumask *cpumask_of_node(int node)
| 0*SD_WAKE_AFFINE \
| 0*SD_SHARE_CPUPOWER \
| 0*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_local_flags(0, cpu) \
| 0*SD_SERIALIZE \
, \
.last_balance = jiffies, \
@@ -763,6 +763,7 @@ enum cpu_idle_type {
#define SD_BALANCE_WAKE 0x0010 /* Balance on wakeup */
#define SD_WAKE_AFFINE 0x0020 /* Wake task to waking CPU */
#define SD_SHARE_CPUPOWER 0x0080 /* Domain members share cpu power */
+#define SD_SHARE_POWERDOMAIN 0x0100 /* Domain members share power domain */
#define SD_SHARE_PKG_RESOURCES 0x0200 /* Domain members share cpu pkg resources */
#define SD_SERIALIZE 0x0400 /* Only a single load balancing instance */
#define SD_ASYM_PACKING 0x0800 /* Place busy groups earlier in the domain */
@@ -85,7 +85,7 @@ int arch_update_cpu_topology(void);
#define ARCH_HAS_SCHED_WAKE_IDLE
/* Common values for SMT siblings */
#ifndef SD_SIBLING_INIT
-#define SD_SIBLING_INIT (struct sched_domain) { \
+#define SD_SIBLING_INIT(cpu) (struct sched_domain) { \
.min_interval = 1, \
.max_interval = 2, \
.busy_factor = 64, \
@@ -99,6 +99,8 @@ int arch_update_cpu_topology(void);
| 1*SD_WAKE_AFFINE \
| 1*SD_SHARE_CPUPOWER \
| 1*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_local_flags(SD_SHARE_CPUPOWER|\
+ SD_SHARE_PKG_RESOURCES, cpu) \
| 0*SD_SERIALIZE \
| 0*SD_PREFER_SIBLING \
| arch_sd_sibling_asym_packing() \
@@ -113,7 +115,7 @@ int arch_update_cpu_topology(void);
#ifdef CONFIG_SCHED_MC
/* Common values for MC siblings. for now mostly derived from SD_CPU_INIT */
#ifndef SD_MC_INIT
-#define SD_MC_INIT (struct sched_domain) { \
+#define SD_MC_INIT(cpu) (struct sched_domain) { \
.min_interval = 1, \
.max_interval = 4, \
.busy_factor = 64, \
@@ -131,6 +133,8 @@ int arch_update_cpu_topology(void);
| 1*SD_WAKE_AFFINE \
| 0*SD_SHARE_CPUPOWER \
| 1*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_local_flags( \
+ SD_SHARE_PKG_RESOURCES, cpu) \
| 0*SD_SERIALIZE \
, \
.last_balance = jiffies, \
@@ -141,7 +145,7 @@ int arch_update_cpu_topology(void);
/* Common values for CPUs */
#ifndef SD_CPU_INIT
-#define SD_CPU_INIT (struct sched_domain) { \
+#define SD_CPU_INIT(cpu) (struct sched_domain) { \
.min_interval = 1, \
.max_interval = 4, \
.busy_factor = 64, \
@@ -161,6 +165,7 @@ int arch_update_cpu_topology(void);
| 1*SD_WAKE_AFFINE \
| 0*SD_SHARE_CPUPOWER \
| 0*SD_SHARE_PKG_RESOURCES \
+ | arch_sd_local_flags(0, cpu) \
| 0*SD_SERIALIZE \
| 1*SD_PREFER_SIBLING \
, \
@@ -5460,6 +5460,11 @@ int __weak arch_sd_sibling_asym_packing(void)
return 0*SD_ASYM_PACKING;
}
+int __weak arch_sd_local_flags(int level, int cpu)
+{
+ return 1*SD_SHARE_POWERDOMAIN;
+}
+
/*
* Initializers for schedule domains
* Non-inlined to reduce accumulated stack pressure in build_sched_domains()
@@ -5473,10 +5478,10 @@ int __weak arch_sd_sibling_asym_packing(void)
#define SD_INIT_FUNC(type) \
static noinline struct sched_domain * \
-sd_init_##type(struct sched_domain_topology_level *tl, int cpu) \
+sd_init_##type(struct sched_domain_topology_level *tl, int cpu) \
{ \
struct sched_domain *sd = *per_cpu_ptr(tl->data.sd, cpu); \
- *sd = SD_##type##_INIT; \
+ *sd = SD_##type##_INIT(cpu); \
SD_INIT_NAME(sd, type); \
sd->private = &tl->data; \
return sd; \
@@ -5652,6 +5657,7 @@ sd_numa_init(struct sched_domain_topology_level *tl, int cpu)
| 0*SD_WAKE_AFFINE
| 0*SD_SHARE_CPUPOWER
| 0*SD_SHARE_PKG_RESOURCES
+ | 1*SD_SHARE_POWERDOMAIN
| 1*SD_SERIALIZE
| 0*SD_PREFER_SIBLING
| sd_local_flags(level)
The function arch_sd_local_flags is used to set flags in sched_domains according to the platform architecture. A new flag SD_SHARE_POWERDOMAIN is also created to reflect whether groups of CPUs in a sched_domain level can or not reach different power state. As an example, the flag should be cleared at CPU level if groups of cores can be power gated independently. This information is used to decide if it's worth packing some tasks in a group of CPUs in order to power gate the other groups instead of spreading the tasks. The default behavior of the scheduler is to spread tasks across CPUs and groups of CPUs so the flag is set into all sched_domains. The cpu parameter of arch_sd_local_flags can be used by architecture to fine tune the scheduler domain flags. As an example SD_SHARE_POWERDOMAIN flag can be set differently for groups of CPUs according to DT information Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org> --- arch/ia64/include/asm/topology.h | 3 ++- arch/tile/include/asm/topology.h | 3 ++- include/linux/sched.h | 1 + include/linux/topology.h | 11 ++++++++--- kernel/sched/core.c | 10 ++++++++-- 5 files changed, 21 insertions(+), 7 deletions(-)