diff mbox

[RFD,02/10] cpuidle: Checking the zero latency inside the governors does not make sense.

Message ID 1413986273-28522-3-git-send-email-daniel.lezcano@linaro.org
State New
Headers show

Commit Message

Daniel Lezcano Oct. 22, 2014, 1:57 p.m. UTC 
If the zero latency is required, we don't want to invoke any cpuidle code at
all.

Move the check within the governors and do the check before selecting the
state in order to fallback to the default idle function.

Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
---
 drivers/cpuidle/governors/ladder.c | 6 ------
 drivers/cpuidle/governors/menu.c   | 4 ----
 kernel/sched/idle.c                | 8 ++++++++
 3 files changed, 8 insertions(+), 10 deletions(-)

Comments

Len Brown Oct. 28, 2014, 3:11 a.m. UTC | #1 
after the patch '[V2,1/5] sched: idle: cpuidle: Check the latency req
before idle' is applied,
this one becomes a no-op.

indeed, they'd conflict both logically and physically.
physically the check for latency_req == 0 was already removed from the
governors.
logically, the upper level, cpu_idle_loop() already invokes
cpu_idle_poll() directly
when latency_req is 0.  Indeed, that would prevent the code in this
patch from running
at all, since cpu_idle_call() is no longer invoked in that case.
Further, why would
you propose invoking "use_default" instead of cpu_idle_poll() in the
latency_req == 0 case?

thanks,
Len Brown, Intel Open Source Technology Center
diff mbox

Patch

diff --git a/drivers/cpuidle/governors/ladder.c b/drivers/cpuidle/governors/ladder.c
index 044ee0d..a7f18e7 100644
--- a/drivers/cpuidle/governors/ladder.c
+++ b/drivers/cpuidle/governors/ladder.c
@@ -71,12 +71,6 @@  static int ladder_select_state(struct cpuidle_driver *drv,
 	int last_residency, last_idx = ldev->last_state_idx;
 	int latency_req = pm_qos_request(PM_QOS_CPU_DMA_LATENCY);
 
-	/* Special case when user has set very strict latency requirement */
-	if (unlikely(latency_req == 0)) {
-		ladder_do_selection(ldev, last_idx, 0);
-		return 0;
-	}
-
 	last_state = &ldev->states[last_idx];
 
 	if (drv->states[last_idx].flags & CPUIDLE_FLAG_TIME_VALID) {
diff --git a/drivers/cpuidle/governors/menu.c b/drivers/cpuidle/governors/menu.c
index 34db2fb..4ed3915 100644
--- a/drivers/cpuidle/governors/menu.c
+++ b/drivers/cpuidle/governors/menu.c
@@ -302,10 +302,6 @@  static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev)
 
 	data->last_state_idx = CPUIDLE_DRIVER_STATE_START - 1;
 
-	/* Special case when user has set very strict latency requirement */
-	if (unlikely(latency_req == 0))
-		return 0;
-
 	/* determine the expected residency time, round up */
 	data->next_timer_us = ktime_to_us(tick_nohz_get_sleep_length());
 
diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c
index 11e7bc4..3042b924 100644
--- a/kernel/sched/idle.c
+++ b/kernel/sched/idle.c
@@ -78,6 +78,7 @@  static void cpuidle_idle_call(void)
 {
 	struct cpuidle_device *dev = __this_cpu_read(cpuidle_devices);
 	struct cpuidle_driver *drv = cpuidle_get_cpu_driver(dev);
+	int latency_req = pm_qos_request(PM_QOS_CPU_DMA_LATENCY);
 	int next_state, entered_state;
 	unsigned int broadcast;
 
@@ -104,6 +105,13 @@  static void cpuidle_idle_call(void)
 	rcu_idle_enter();
 
 	/*
+	 * The latency requirement does not allow any latency, jump to
+	 * the default idle function
+	 */
+	if (latency_req == 0)
+		goto use_default;
+
+	/*
 	 * Ask the cpuidle framework to choose a convenient idle state.
 	 * Fall back to the default arch idle method on errors.
 	 */