Lines Matching refs:balance

1713     /* Check if run-queue part of active NUMA balance. */
1963      * balance improves then stop the search. While a better swap
2079      * balance domains, some of which do not cross NUMA boundaries.
4937  * load-balance operations.
5972     /* balance early to pull high priority tasks */
7157             sd = NULL; /* Prefer wake_affine over balance flags */
7714  * To achieve this balance we define a measure of imbalance which follows
7735  * of load-balance at each level inv. proportional to the number of CPUs in
7744  *         |         |     `- number of CPUs doing load-balance
7748  * Coupled with a limit on how many tasks we can migrate every balance pass,
8024          * meet load balance goals by pulling other tasks on src_cpu.
8064      * 3) too many balance attempts have failed.
8784  * If we were to balance group-wise we'd place two tasks in the first group and
8789  * by noticing the lower domain failed to reach balance and had difficulty
8794  * find_busiest_group() avoid some of the usual balance conditions to allow it
8817  * any benefit for the load balance.
9483                  * a real need of migration, periodic load balance will
9607  *             groups of a given sched_domain during load balance.
9608  * @env: load balance environment
9639          * the imbalance. The next load balance will take care of
9658              * amount of load to migrate in order to balance the
9769  * to restore balance.
9809     /* ASYM feature bypasses nice load balance check */
9939          * task, the next balance pass can still reduce the busiest
10126      * to do the newly idle load balance.
10252          * This changes load balance semantics a bit on who can move
10261          * moreover subsequent load balance cycles should correct the
10282          * We failed to reach balance because of affinity.
10315          * Increment the failure counter only on periodic balance.
10316          * We do not want newidle balance, which can be very
10343              * only after active load balance is finished.
10383      * We reach balance although we may have faced some affinity
10397      * We reach balance because all tasks are pinned at this level so
10409      * newidle_balance() disregards balance intervals, so we could
10456     /* used by idle balance, so cpu_busy = 0 */
10609  * This trades load-balance latency on larger machines for less cross talk.
10660          * Stop the load balance at this level. There is another
10722          * balance. Other idle CPUs have already rebalanced with
10725          * balance for itself and we need to update the
10800      * is idle. And the softirq performing nohz idle load balance
10901          * For asymmetric systems, we do not want to nicely balance
10918          * the others are - so just get a nohz balance going if it looks
10998          * the nohz idle balance, which should be avoided.
11057  * Internal function that runs load balance for all idle cpus. The load balance
11058  * can be a simple update of blocked load or a complete load balance with
11121          * If time for next balance is due,
11122          * do the balance.
11163     /* The full idle balance loop has been done */
11176  * In CONFIG_NO_HZ_COMMON case, the idle balance kickee will do the
11285      * for load-balance and preemption/IRQs are still disabled avoiding
11359     /* Move the next balance forward */
11400      * give the idle CPUs a chance to load balance. Else we may
11401      * load balance only within the local sched_domain hierarchy
11408     /* normal load balance */
11457     /* Invoke active balance to force migrate currently running task */
12079     .balance = balance_fair,