[PATCH] sched: cleanup sched_group cpu_power setup
authorSiddha, Suresh B <suresh.b.siddha@intel.com>
Tue, 3 Oct 2006 08:14:09 +0000 (01:14 -0700)
committerLinus Torvalds <torvalds@g5.osdl.org>
Tue, 3 Oct 2006 15:04:06 +0000 (08:04 -0700)
Up to now sched group's cpu_power for each sched domain is initialized
independently.  This made the setup code ugly as the new sched domains are
getting added.

Make the sched group cpu_power setup code generic, by using domain child
field and new domain flag in sched_domain.  For most of the sched
domains(except NUMA), sched group's cpu_power is now computed generically
using the domain properties of itself and of the child domain.

sched groups in NUMA domains are setup little differently and hence they
don't use this generic mechanism.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Nick Piggin <nickpiggin@yahoo.com.au>
Cc: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
include/linux/sched.h
include/linux/topology.h
kernel/sched.c

index 8e26c90..331f450 100644 (file)
@@ -625,9 +625,17 @@ enum idle_type
 #define SD_WAKE_BALANCE                64      /* Perform balancing at task wakeup */
 #define SD_SHARE_CPUPOWER      128     /* Domain members share cpu power */
 #define SD_POWERSAVINGS_BALANCE        256     /* Balance for power savings */
+#define SD_SHARE_PKG_RESOURCES 512     /* Domain members share cpu pkg resources */
 
-#define BALANCE_FOR_POWER      ((sched_mc_power_savings || sched_smt_power_savings) \
-                                ? SD_POWERSAVINGS_BALANCE : 0)
+#define BALANCE_FOR_MC_POWER   \
+       (sched_smt_power_savings ? SD_POWERSAVINGS_BALANCE : 0)
+
+#define BALANCE_FOR_PKG_POWER  \
+       ((sched_mc_power_savings || sched_smt_power_savings) ?  \
+        SD_POWERSAVINGS_BALANCE : 0)
+
+#define test_sd_parent(sd, flag)       ((sd->parent &&         \
+                                        (sd->parent->flags & flag)) ? 1 : 0)
 
 
 struct sched_group {
index 486bec2..da508d1 100644 (file)
 #endif
 #endif /* CONFIG_SCHED_SMT */
 
+#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) {             \
+       .span                   = CPU_MASK_NONE,        \
+       .parent                 = NULL,                 \
+       .child                  = NULL,                 \
+       .groups                 = NULL,                 \
+       .min_interval           = 1,                    \
+       .max_interval           = 4,                    \
+       .busy_factor            = 64,                   \
+       .imbalance_pct          = 125,                  \
+       .cache_nice_tries       = 1,                    \
+       .per_cpu_gain           = 100,                  \
+       .busy_idx               = 2,                    \
+       .idle_idx               = 1,                    \
+       .newidle_idx            = 2,                    \
+       .wake_idx               = 1,                    \
+       .forkexec_idx           = 1,                    \
+       .flags                  = SD_LOAD_BALANCE       \
+                               | SD_BALANCE_NEWIDLE    \
+                               | SD_BALANCE_EXEC       \
+                               | SD_WAKE_AFFINE        \
+                               | SD_SHARE_PKG_RESOURCES\
+                               | BALANCE_FOR_MC_POWER, \
+       .last_balance           = jiffies,              \
+       .balance_interval       = 1,                    \
+       .nr_balance_failed      = 0,                    \
+}
+#endif
+#endif /* CONFIG_SCHED_MC */
+
 /* Common values for CPUs */
 #ifndef SD_CPU_INIT
 #define SD_CPU_INIT (struct sched_domain) {            \
                                | SD_BALANCE_NEWIDLE    \
                                | SD_BALANCE_EXEC       \
                                | SD_WAKE_AFFINE        \
-                               | BALANCE_FOR_POWER,    \
+                               | BALANCE_FOR_PKG_POWER,\
        .last_balance           = jiffies,              \
        .balance_interval       = 1,                    \
        .nr_balance_failed      = 0,                    \
        .nr_balance_failed      = 0,                    \
 }
 
-#ifdef CONFIG_SCHED_MC
-#ifndef SD_MC_INIT
-/* for now its same as SD_CPU_INIT.
- * TBD: Tune Domain parameters!
- */
-#define SD_MC_INIT   SD_CPU_INIT
-#endif
-#endif
-
 #ifdef CONFIG_NUMA
 #ifndef SD_NODE_INIT
 #error Please define an appropriate SD_NODE_INIT in include/asm/topology.h!!!
index 0feeacb..0a5e814 100644 (file)
@@ -2541,8 +2541,14 @@ static int load_balance(int this_cpu, struct rq *this_rq,
        struct rq *busiest;
        cpumask_t cpus = CPU_MASK_ALL;
 
+       /*
+        * When power savings policy is enabled for the parent domain, idle
+        * sibling can pick up load irrespective of busy siblings. In this case,
+        * let the state of idle sibling percolate up as IDLE, instead of
+        * portraying it as NOT_IDLE.
+        */
        if (idle != NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER &&
-           !sched_smt_power_savings)
+           !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
                sd_idle = 1;
 
        schedstat_inc(sd, lb_cnt[idle]);
@@ -2638,7 +2644,7 @@ redo:
        }
 
        if (!nr_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER &&
-           !sched_smt_power_savings)
+           !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
                return -1;
        return nr_moved;
 
@@ -2654,7 +2660,7 @@ out_one_pinned:
                sd->balance_interval *= 2;
 
        if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER &&
-                       !sched_smt_power_savings)
+           !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
                return -1;
        return 0;
 }
@@ -2676,7 +2682,14 @@ load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd)
        int sd_idle = 0;
        cpumask_t cpus = CPU_MASK_ALL;
 
-       if (sd->flags & SD_SHARE_CPUPOWER && !sched_smt_power_savings)
+       /*
+        * When power savings policy is enabled for the parent domain, idle
+        * sibling can pick up load irrespective of busy siblings. In this case,
+        * let the state of idle sibling percolate up as IDLE, instead of
+        * portraying it as NOT_IDLE.
+        */
+       if (sd->flags & SD_SHARE_CPUPOWER &&
+           !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
                sd_idle = 1;
 
        schedstat_inc(sd, lb_cnt[NEWLY_IDLE]);
@@ -2717,7 +2730,8 @@ redo:
 
        if (!nr_moved) {
                schedstat_inc(sd, lb_failed[NEWLY_IDLE]);
-               if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER)
+               if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER &&
+                   !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
                        return -1;
        } else
                sd->nr_balance_failed = 0;
@@ -2727,7 +2741,7 @@ redo:
 out_balanced:
        schedstat_inc(sd, lb_balanced[NEWLY_IDLE]);
        if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER &&
-                                       !sched_smt_power_savings)
+           !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
                return -1;
        sd->nr_balance_failed = 0;
 
@@ -5400,7 +5414,9 @@ static int sd_degenerate(struct sched_domain *sd)
        if (sd->flags & (SD_LOAD_BALANCE |
                         SD_BALANCE_NEWIDLE |
                         SD_BALANCE_FORK |
-                        SD_BALANCE_EXEC)) {
+                        SD_BALANCE_EXEC |
+                        SD_SHARE_CPUPOWER |
+                        SD_SHARE_PKG_RESOURCES)) {
                if (sd->groups != sd->groups->next)
                        return 0;
        }
@@ -5434,7 +5450,9 @@ sd_parent_degenerate(struct sched_domain *sd, struct sched_domain *parent)
                pflags &= ~(SD_LOAD_BALANCE |
                                SD_BALANCE_NEWIDLE |
                                SD_BALANCE_FORK |
-                               SD_BALANCE_EXEC);
+                               SD_BALANCE_EXEC |
+                               SD_SHARE_CPUPOWER |
+                               SD_SHARE_PKG_RESOURCES);
        }
        if (~cflags & pflags)
                return 0;
@@ -6241,12 +6259,65 @@ static void free_sched_groups(const cpumask_t *cpu_map)
 #endif
 
 /*
+ * Initialize sched groups cpu_power.
+ *
+ * cpu_power indicates the capacity of sched group, which is used while
+ * distributing the load between different sched groups in a sched domain.
+ * Typically cpu_power for all the groups in a sched domain will be same unless
+ * there are asymmetries in the topology. If there are asymmetries, group
+ * having more cpu_power will pickup more load compared to the group having
+ * less cpu_power.
+ *
+ * cpu_power will be a multiple of SCHED_LOAD_SCALE. This multiple represents
+ * the maximum number of tasks a group can handle in the presence of other idle
+ * or lightly loaded groups in the same sched domain.
+ */
+static void init_sched_groups_power(int cpu, struct sched_domain *sd)
+{
+       struct sched_domain *child;
+       struct sched_group *group;
+
+       WARN_ON(!sd || !sd->groups);
+
+       if (cpu != first_cpu(sd->groups->cpumask))
+               return;
+
+       child = sd->child;
+
+       /*
+        * For perf policy, if the groups in child domain share resources
+        * (for example cores sharing some portions of the cache hierarchy
+        * or SMT), then set this domain groups cpu_power such that each group
+        * can handle only one task, when there are other idle groups in the
+        * same sched domain.
+        */
+       if (!child || (!(sd->flags & SD_POWERSAVINGS_BALANCE) &&
+                      (child->flags &
+                       (SD_SHARE_CPUPOWER | SD_SHARE_PKG_RESOURCES)))) {
+               sd->groups->cpu_power = SCHED_LOAD_SCALE;
+               return;
+       }
+
+       sd->groups->cpu_power = 0;
+
+       /*
+        * add cpu_power of each child group to this groups cpu_power
+        */
+       group = child->groups;
+       do {
+               sd->groups->cpu_power += group->cpu_power;
+               group = group->next;
+       } while (group != child->groups);
+}
+
+/*
  * Build sched domains for a given set of cpus and attach the sched domains
  * to the individual cpus
  */
 static int build_sched_domains(const cpumask_t *cpu_map)
 {
        int i;
+       struct sched_domain *sd;
 #ifdef CONFIG_NUMA
        struct sched_group **sched_group_nodes = NULL;
        struct sched_group *sched_group_allnodes = NULL;
@@ -6456,72 +6527,20 @@ static int build_sched_domains(const cpumask_t *cpu_map)
        /* Calculate CPU power for physical packages and nodes */
 #ifdef CONFIG_SCHED_SMT
        for_each_cpu_mask(i, *cpu_map) {
-               struct sched_domain *sd;
                sd = &per_cpu(cpu_domains, i);
-               sd->groups->cpu_power = SCHED_LOAD_SCALE;
+               init_sched_groups_power(i, sd);
        }
 #endif
 #ifdef CONFIG_SCHED_MC
        for_each_cpu_mask(i, *cpu_map) {
-               int power;
-               struct sched_domain *sd;
                sd = &per_cpu(core_domains, i);
-               if (sched_smt_power_savings)
-                       power = SCHED_LOAD_SCALE * cpus_weight(sd->groups->cpumask);
-               else
-                       power = SCHED_LOAD_SCALE + (cpus_weight(sd->groups->cpumask)-1)
-                                           * SCHED_LOAD_SCALE / 10;
-               sd->groups->cpu_power = power;
+               init_sched_groups_power(i, sd);
        }
 #endif
 
        for_each_cpu_mask(i, *cpu_map) {
-               struct sched_domain *sd;
-#ifdef CONFIG_SCHED_MC
-               sd = &per_cpu(phys_domains, i);
-               if (i != first_cpu(sd->groups->cpumask))
-                       continue;
-
-               sd->groups->cpu_power = 0;
-               if (sched_mc_power_savings || sched_smt_power_savings) {
-                       int j;
-
-                       for_each_cpu_mask(j, sd->groups->cpumask) {
-                               struct sched_domain *sd1;
-                               sd1 = &per_cpu(core_domains, j);
-                               /*
-                                * for each core we will add once
-                                * to the group in physical domain
-                                */
-                               if (j != first_cpu(sd1->groups->cpumask))
-                                       continue;
-
-                               if (sched_smt_power_savings)
-                                       sd->groups->cpu_power += sd1->groups->cpu_power;
-                               else
-                                       sd->groups->cpu_power += SCHED_LOAD_SCALE;
-                       }
-               } else
-                       /*
-                        * This has to be < 2 * SCHED_LOAD_SCALE
-                        * Lets keep it SCHED_LOAD_SCALE, so that
-                        * while calculating NUMA group's cpu_power
-                        * we can simply do
-                        *  numa_group->cpu_power += phys_group->cpu_power;
-                        *
-                        * See "only add power once for each physical pkg"
-                        * comment below
-                        */
-                       sd->groups->cpu_power = SCHED_LOAD_SCALE;
-#else
-               int power;
                sd = &per_cpu(phys_domains, i);
-               if (sched_smt_power_savings)
-                       power = SCHED_LOAD_SCALE * cpus_weight(sd->groups->cpumask);
-               else
-                       power = SCHED_LOAD_SCALE;
-               sd->groups->cpu_power = power;
-#endif
+               init_sched_groups_power(i, sd);
        }
 
 #ifdef CONFIG_NUMA