sched: Remove load_balance_newidle()
Peter Zijlstra [Wed, 23 Dec 2009 14:29:42 +0000 (15:29 +0100)]
The two functions: load_balance{,_newidle}() are very similar, with the
following differences:

 - rq->lock usage
 - sb->balance_interval updates
 - *balance check

So remove the load_balance_newidle() call with load_balance(.idle =
CPU_NEWLY_IDLE), explicitly unlock the rq->lock before calling (would be
done by double_lock_balance() anyway), and ignore the other differences
for now.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

kernel/sched_fair.c

index 1040832..de5ab12 100644 (file)
@@ -3010,125 +3010,6 @@ out:
 }
 
 /*
- * Check this_cpu to ensure it is balanced within domain. Attempt to move
- * tasks if there is an imbalance.
- *
- * Called from schedule when this_rq is about to become idle (CPU_NEWLY_IDLE).
- * this_rq is locked.
- */
-static int
-load_balance_newidle(int this_cpu, struct rq *this_rq, struct sched_domain *sd)
-{
-       struct sched_group *group;
-       struct rq *busiest = NULL;
-       unsigned long imbalance;
-       int ld_moved = 0;
-       int sd_idle = 0;
-       int all_pinned = 0;
-       struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask);
-
-       cpumask_copy(cpus, cpu_active_mask);
-
-       /*
-        * 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 CPU_NOT_IDLE.
-        */
-       if (sd->flags & SD_SHARE_CPUPOWER &&
-           !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
-               sd_idle = 1;
-
-       schedstat_inc(sd, lb_count[CPU_NEWLY_IDLE]);
-redo:
-       update_shares_locked(this_rq, sd);
-       group = find_busiest_group(sd, this_cpu, &imbalance, CPU_NEWLY_IDLE,
-                                  &sd_idle, cpus, NULL);
-       if (!group) {
-               schedstat_inc(sd, lb_nobusyg[CPU_NEWLY_IDLE]);
-               goto out_balanced;
-       }
-
-       busiest = find_busiest_queue(group, CPU_NEWLY_IDLE, imbalance, cpus);
-       if (!busiest) {
-               schedstat_inc(sd, lb_nobusyq[CPU_NEWLY_IDLE]);
-               goto out_balanced;
-       }
-
-       BUG_ON(busiest == this_rq);
-
-       schedstat_add(sd, lb_imbalance[CPU_NEWLY_IDLE], imbalance);
-
-       ld_moved = 0;
-       if (busiest->nr_running > 1) {
-               /* Attempt to move tasks */
-               double_lock_balance(this_rq, busiest);
-               /* this_rq->clock is already updated */
-               update_rq_clock(busiest);
-               ld_moved = move_tasks(this_rq, this_cpu, busiest,
-                                       imbalance, sd, CPU_NEWLY_IDLE,
-                                       &all_pinned);
-               double_unlock_balance(this_rq, busiest);
-
-               if (unlikely(all_pinned)) {
-                       cpumask_clear_cpu(cpu_of(busiest), cpus);
-                       if (!cpumask_empty(cpus))
-                               goto redo;
-               }
-       }
-
-       if (!ld_moved) {
-               int active_balance = 0;
-
-               schedstat_inc(sd, lb_failed[CPU_NEWLY_IDLE]);
-               sd->nr_balance_failed++;
-
-               if (need_active_balance(sd, sd_idle, CPU_NEWLY_IDLE)) {
-                       double_lock_balance(this_rq, busiest);
-
-                       /*
-                        * don't kick the migration_thread, if the curr
-                        * task on busiest cpu can't be moved to this_cpu
-                        */
-                       if (!cpumask_test_cpu(this_cpu,
-                                             &busiest->curr->cpus_allowed)) {
-                               double_unlock_balance(this_rq, busiest);
-                               all_pinned = 1;
-                               return ld_moved;
-                       }
-
-                       if (!busiest->active_balance) {
-                               busiest->active_balance = 1;
-                               busiest->push_cpu = this_cpu;
-                               active_balance = 1;
-                       }
-
-                       double_unlock_balance(this_rq, busiest);
-                       /*
-                        * Should not call ttwu while holding a rq->lock
-                        */
-                       raw_spin_unlock(&this_rq->lock);
-                       if (active_balance)
-                               wake_up_process(busiest->migration_thread);
-                       raw_spin_lock(&this_rq->lock);
-               }
-       } else
-               sd->nr_balance_failed = 0;
-
-       update_shares_locked(this_rq, sd);
-       return ld_moved;
-
-out_balanced:
-       schedstat_inc(sd, lb_balanced[CPU_NEWLY_IDLE]);
-       if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER &&
-           !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE))
-               return -1;
-       sd->nr_balance_failed = 0;
-
-       return 0;
-}
-
-/*
  * idle_balance is called by schedule() if this_cpu is about to become
  * idle. Attempts to pull tasks from other CPUs.
  */
@@ -3143,16 +3024,23 @@ static void idle_balance(int this_cpu, struct rq *this_rq)
        if (this_rq->avg_idle < sysctl_sched_migration_cost)
                return;
 
+       /*
+        * Drop the rq->lock, but keep IRQ/preempt disabled.
+        */
+       raw_spin_unlock(&this_rq->lock);
+
        for_each_domain(this_cpu, sd) {
                unsigned long interval;
+               int balance = 1;
 
                if (!(sd->flags & SD_LOAD_BALANCE))
                        continue;
 
-               if (sd->flags & SD_BALANCE_NEWIDLE)
+               if (sd->flags & SD_BALANCE_NEWIDLE) {
                        /* If we've pulled tasks over stop searching: */
-                       pulled_task = load_balance_newidle(this_cpu, this_rq,
-                                                          sd);
+                       pulled_task = load_balance(this_cpu, this_rq,
+                                                  sd, CPU_NEWLY_IDLE, &balance);
+               }
 
                interval = msecs_to_jiffies(sd->balance_interval);
                if (time_after(next_balance, sd->last_balance + interval))
@@ -3162,6 +3050,9 @@ static void idle_balance(int this_cpu, struct rq *this_rq)
                        break;
                }
        }
+
+       raw_spin_lock(&this_rq->lock);
+
        if (pulled_task || time_after(jiffies, this_rq->next_balance)) {
                /*
                 * We are going idle. next_balance may be set based on