2014-01-31 21:48:40

by Zoran Markovic

[permalink] [raw]
Subject: [RFC PATCH] power: move pm_qos update timeout handler to power-efficient workqueue

From: Shaibal Dutta <[email protected]>

To avoid waking up idle CPUs, allow the scheduler to select the best CPU
to handle pm_qos update timeouts. This extends idle residency times and
conserves power.

This functionality is enabled when CONFIG_WQ_POWER_EFFICIENT is selected.

Cc: Pavel Machek <[email protected]>
Cc: "Rafael J. Wysocki" <[email protected]>
Cc: Len Brown <[email protected]>
Signed-off-by: Shaibal Dutta <[email protected]>
[[email protected]: Rebased to latest kernel. Fixed code alignment.
Added commit message.]
Signed-off-by: Zoran Markovic <[email protected]>
---
kernel/power/qos.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/kernel/power/qos.c b/kernel/power/qos.c
index 8dff9b4..5e35a3a 100644
--- a/kernel/power/qos.c
+++ b/kernel/power/qos.c
@@ -405,7 +405,8 @@ void pm_qos_update_request_timeout(struct pm_qos_request *req, s32 new_value,
pm_qos_array[req->pm_qos_class]->constraints,
&req->node, PM_QOS_UPDATE_REQ, new_value);

- schedule_delayed_work(&req->work, usecs_to_jiffies(timeout_us));
+ queue_delayed_work(system_power_efficient_wq,
+ &req->work, usecs_to_jiffies(timeout_us));
}

/**
--
1.7.9.5