Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit ce9027a6 authored by Stephen Boyd's avatar Stephen Boyd Committed by Mahesh Sivasubramanian
Browse files

drivers: cpuidle: lpm-levels: Remove hotplug checking



This check doesn't make sense anymore because we're not entering
this code from hotplug paths anymore. Hotplug goes directly
through PSCI paths to the firmware. It used to be that we wanted
to handle the case where hotplug was happening but the cluster
was idle so we wanted to override the idle flag to consider the
cluster as idling instead of completely going away. But that
isn't true anymore with PSCI.

Change-Id: I6003fcfc6fa430fdadd3a9b82b789953b7a13a5f
Signed-off-by: default avatarStephen Boyd <sboyd@codeaurora.org>
parent 14681b57
Loading
Loading
Loading
Loading
+0 −11
Original line number Diff line number Diff line
@@ -953,17 +953,6 @@ static int cluster_select(struct lpm_cluster *cluster, bool from_idle,
		latency_us = pm_qos_request_for_cpumask(PM_QOS_CPU_DMA_LATENCY,
							&mask);

	/*
	 * If atleast one of the core in the cluster is online, the cluster
	 * low power modes should be determined by the idle characteristics
	 * even if the last core enters the low power mode as a part of
	 * hotplug.
	 */

	if (!from_idle && num_online_cpus() > 1 &&
		cpumask_intersects(&cluster->child_cpus, cpu_online_mask))
		from_idle = true;

	for (i = 0; i < cluster->nlevels; i++) {
		struct lpm_cluster_level *level = &cluster->levels[i];
		struct power_params *pwr_params = &level->pwr;