public inbox for linux-arm-kernel@lists.infradead.org 
 help / color / mirror / Atom feed
From: mtitinger@baylibre•com (Marc Titinger)
To: linux-arm-kernel@lists•infradead.org
Subject: [PATCH RFC 18/27] drivers: cpu-pd: Add PM Domain governor for CPUs
Date: Thu, 19 Nov 2015 09:50:30 +0100	[thread overview]
Message-ID: <564D8D56.1080206@baylibre.com> (raw)
In-Reply-To: <20151118184245.GA28755@red-moon>

On 18/11/2015 19:42, Lorenzo Pieralisi wrote:
> On Tue, Nov 17, 2015 at 03:37:42PM -0700, Lina Iyer wrote:
>> A PM domain comprising of CPUs may be powered off when all the CPUs in
>> the domain are powered down. Powering down a CPU domain is generally a
>> expensive operation and therefore the power performance trade offs
>> should be considered. The time between the last CPU powering down and
>> the first CPU powering up in a domain, is the time available for the
>> domain to sleep. Ideally, the sleep time of the domain should fulfill
>> the residency requirement of the domains' idle state.
>>
>> To do this effectively, read the time before the wakeup of the cluster's
>> CPUs and ensure that the domain's idle state sleep time guarantees the
>> QoS requirements of each of the CPU, the PM QoS CPU_DMA_LATENCY and the
>> state's residency.
>
> To me this information should be part of the CPUidle governor (it is
> already there), we should not split the decision into multiple layers.
>
> The problem you are facing is that the CPUidle governor(s) do not take
> cross cpus relationship into account, I do not think that adding another
> decision layer in the power domain subsystem helps, you are doing that
> just because adding it to the existing CPUidle governor(s) is invasive.
>
> Why can't we use the power domain work you put together to eg disable
> idle states that share multiple cpus and make them "visible" only
> when the power domain that encompass them is actually going down ?
>
> You could use the power domains information to detect states that
> are shared between cpus.
>
> It is just an idea, what I am saying is that having another governor in
> the power domain subsytem does not make much sense, you split the
> decision in two layers while there is actually one, the existing
> CPUidle governor and that's where the decision should be taken.
>
> Thoughts appreciated.

Maybe this is silly and not thought-through, but I wonder if the 
responsibilities could be split or instance with an outer control loop 
that has the heuristic to compute the next tick time, and the required 
cpu-power needed during that time slot, and an inner control loop 
(genpd) that has a per-domain QoS and can optimize power consumption.

Marc.

>
> Lorenzo
>
>> Signed-off-by: Lina Iyer <lina.iyer@linaro•org>
>> ---
>>   drivers/base/power/cpu-pd.c | 83 ++++++++++++++++++++++++++++++++++++++++++++-
>>   1 file changed, 82 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/base/power/cpu-pd.c b/drivers/base/power/cpu-pd.c
>> index 617ce54..a00abc1 100644
>> --- a/drivers/base/power/cpu-pd.c
>> +++ b/drivers/base/power/cpu-pd.c
>> @@ -21,6 +21,7 @@
>>   #include <linux/pm_qos.h>
>>   #include <linux/rculist.h>
>>   #include <linux/slab.h>
>> +#include <linux/tick.h>
>>
>>   #define CPU_PD_NAME_MAX 36
>>
>> @@ -66,6 +67,86 @@ static void get_cpus_in_domain(struct generic_pm_domain *genpd,
>>   	}
>>   }
>>
>> +static bool cpu_pd_down_ok(struct dev_pm_domain *pd)
>> +{
>> +	struct generic_pm_domain *genpd = pd_to_genpd(pd);
>> +	struct cpu_pm_domain *cpu_pd = to_cpu_pd(genpd);
>> +	int qos = pm_qos_request(PM_QOS_CPU_DMA_LATENCY);
>> +	u64 sleep_ns = ~0;
>> +	ktime_t earliest;
>> +	int cpu;
>> +	int i;
>> +
>> +	/* Reset the last set genpd state, default to index 0 */
>> +	genpd->state_idx = 0;
>> +
>> +	/* We dont want to power down, if QoS is 0 */
>> +	if (!qos)
>> +		return false;
>> +
>> +	/*
>> +	 * Find the sleep time for the cluster.
>> +	 * The time between now and the first wake up of any CPU that
>> +	 * are in this domain hierarchy is the time available for the
>> +	 * domain to be idle.
>> +	 */
>> +	earliest.tv64 = KTIME_MAX;
>> +	for_each_cpu_and(cpu, cpu_pd->cpus, cpu_online_mask) {
>> +		struct device *cpu_dev = get_cpu_device(cpu);
>> +		struct gpd_timing_data *td;
>> +
>> +		td = &dev_gpd_data(cpu_dev)->td;
>> +
>> +		if (earliest.tv64 < td->next_wakeup.tv64)
>> +			earliest = td->next_wakeup;
>> +	}
>> +
>> +	sleep_ns = ktime_to_ns(ktime_sub(earliest, ktime_get()));
>> +	if (sleep_ns <= 0)
>> +		return false;
>> +
>> +	/*
>> +	 * Find the deepest sleep state that satisfies the residency
>> +	 * requirement and the QoS constraint
>> +	 */
>> +	for (i = genpd->state_count - 1; i > 0; i--) {
>> +		u64 state_sleep_ns;
>> +
>> +		state_sleep_ns = genpd->states[i].power_off_latency_ns +
>> +			genpd->states[i].power_on_latency_ns +
>> +			genpd->states[i].residency_ns;
>> +
>> +		/*
>> +		 * If we cant sleep to save power in the state, move on
>> +		 * to the next lower idle state.
>> +		 */
>> +		if (state_sleep_ns > sleep_ns)
>> +		       continue;
>> +
>> +		/*
>> +		 * We also dont want to sleep more than we should to
>> +		 * gaurantee QoS.
>> +		 */
>> +		if (state_sleep_ns < (qos * NSEC_PER_USEC))
>> +			break;
>> +	}
>> +
>> +	if (i >= 0)
>> +		genpd->state_idx = i;
>> +
>> +	return  (i >= 0) ? true : false;
>> +}
>> +
>> +static bool cpu_stop_ok(struct device *dev)
>> +{
>> +	return true;
>> +}
>> +
>> +struct dev_power_governor cpu_pd_gov = {
>> +	.power_down_ok = cpu_pd_down_ok,
>> +	.stop_ok = cpu_stop_ok,
>> +};
>> +
>>   static int cpu_pd_power_off(struct generic_pm_domain *genpd)
>>   {
>>   	struct cpu_pm_domain *pd = to_cpu_pd(genpd);
>> @@ -183,7 +264,7 @@ int of_register_cpu_pm_domain(struct device_node *dn,
>>
>>   	/* Register the CPU genpd */
>>   	pr_debug("adding %s as CPU PM domain.\n", pd->genpd->name);
>> -	ret = of_pm_genpd_init(dn, pd->genpd, &simple_qos_governor, false);
>> +	ret = of_pm_genpd_init(dn, pd->genpd, &cpu_pd_gov, false);
>>   	if (ret) {
>>   		pr_err("Unable to initialize domain %s\n", dn->full_name);
>>   		return ret;
>> --
>> 2.1.4
>>

  reply	other threads:[~2015-11-19  8:50 UTC|newest]

Thread overview: 83+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-11-17 22:37 [PATCH RFC 00/27] PM/Domains: Cluster idle support for ARM SoCs Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 01/27] PM / Domains: core changes for multiple states Lina Iyer
2015-12-09 13:58   ` Ulf Hansson
2015-12-17 17:58     ` Axel Haslam
2015-12-17 21:19       ` Ulf Hansson
2015-11-17 22:37 ` [PATCH RFC 02/27] PM / Domains: Allow domain power states to be read from DT Lina Iyer
2015-12-10 16:53   ` Ulf Hansson
2015-12-15 10:07     ` Marc Titinger
2015-12-15 22:14       ` Lina Iyer
2015-12-16 21:36       ` Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 03/27] PM / Domain: Add additional state specific param Lina Iyer
2015-11-19 21:33   ` Kevin Hilman
2015-11-17 22:37 ` [PATCH RFC 04/27] PM / Domains: make governor select deepest state Lina Iyer
2015-12-11  9:13   ` Ulf Hansson
2015-11-17 22:37 ` [PATCH RFC 05/27] PM / Domains: remove old power on/off latencies Lina Iyer
2015-11-18 14:57   ` [PATCH] ARM: imx6: pm: declare pm domain latency on power_state struct Lina Iyer
2015-11-23 13:31     ` Lucas Stach
2015-11-23 13:42       ` Lucas Stach
2015-12-04 23:19         ` Lina Iyer
2015-12-11  9:16   ` [PATCH RFC 05/27] PM / Domains: remove old power on/off latencies Ulf Hansson
2015-11-17 22:37 ` [PATCH RFC 06/27] PM / Domains: add debugfs 'states' and 'timings' seq files Lina Iyer
2015-12-11 11:46   ` Ulf Hansson
2015-12-16 11:07     ` Marc Titinger
2015-12-16 12:48       ` Ulf Hansson
2015-12-16 14:12         ` Marc Titinger
2015-11-17 22:37 ` [PATCH RFC 07/27] PM / Domains: Read domain residency from DT Lina Iyer
2015-11-24 20:41   ` Stephen Boyd
2015-12-11 11:54   ` Ulf Hansson
2015-11-17 22:37 ` [PATCH RFC 08/27] PM / Domains: Support IRQ safe PM domains Lina Iyer
2016-01-14 14:42   ` Ulf Hansson
2016-01-14 18:33     ` Lina Iyer
2016-01-15  8:55       ` Ulf Hansson
2016-01-15 16:57         ` Lina Iyer
2016-01-15 22:08           ` Ulf Hansson
2016-01-18 16:58             ` Lina Iyer
2016-01-18 17:00               ` Lina Iyer
2016-01-19 10:01               ` Ulf Hansson
2015-11-17 22:37 ` [PATCH RFC 09/27] PM / Domains: Attempt runtime suspend of IRQ safe parent domain Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 10/27] drivers: power: Introduce PM domains for CPUs/clusters Lina Iyer
2015-11-24 20:52   ` Stephen Boyd
2015-11-17 22:37 ` [PATCH RFC 11/27] drivers: cpu: Define CPU devices as IRQ safe Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 12/27] ARM: cpuidle: remove cpu parameter from the cpuidle_ops suspend hook Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 13/27] ARM: cpuidle: Add runtime PM support for CPU idle Lina Iyer
2015-11-18  8:50   ` Zhaoyang Huang
2015-11-18 14:17     ` Lina Iyer
2015-11-19 22:10   ` Kevin Hilman
2015-11-17 22:37 ` [PATCH RFC 14/27] tick: get next wakeup event for the CPU Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 15/27] PM / Domains: Add next_wakeup to device's timing data Lina Iyer
2015-11-19 22:19   ` Kevin Hilman
2015-11-20 15:58     ` Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 16/27] ARM: cpuidle: Record the next wakeup event of the CPU Lina Iyer
2015-11-19 23:35   ` Kevin Hilman
2015-11-20 16:28     ` Lina Iyer
2015-11-24 18:29       ` Kevin Hilman
2015-11-17 22:37 ` [PATCH RFC 17/27] drivers: cpu-pd: Record CPUs that are part of the domain Lina Iyer
2015-11-24 21:00   ` Stephen Boyd
2015-11-25 14:13     ` Lina Iyer
2015-11-25 19:12       ` Stephen Boyd
2015-11-25 20:20         ` Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 18/27] drivers: cpu-pd: Add PM Domain governor for CPUs Lina Iyer
2015-11-18 18:42   ` Lorenzo Pieralisi
2015-11-19  8:50     ` Marc Titinger [this message]
2015-11-20 17:39       ` Lina Iyer
2015-11-19 23:52     ` Kevin Hilman
2015-11-20 16:21       ` Lorenzo Pieralisi
2015-11-20 16:42         ` Lina Iyer
2015-11-20  0:03   ` Kevin Hilman
2015-11-17 22:37 ` [PATCH RFC 19/27] drivers: cpu-pd: Invoke CPU PM runtime on hotplug Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 20/27] Documentation: ARM: topology: 'cluster' property for cluster nodes Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 21/27] drivers: cpu-pd: Parse topology to setup CPU PM domains Lina Iyer
2015-12-07 14:54   ` Lorenzo Pieralisi
2015-12-08 18:05     ` Lina Iyer
2015-12-10 18:11       ` Lorenzo Pieralisi
2015-12-11  9:04         ` Geert Uytterhoeven
2015-12-11 20:51           ` Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 22/27] drivers: firmware: PSCI: Export psci_has_ext_power_state() Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 23/27] ARM64: psci: Support cluster idle states for OS-Initated Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 24/27] arm64: dts: Add Qualcomm MSM8916, MTP8916, APQ8016, SBC8016 ids Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 25/27] devicetree: bindings: Document qcom, msm-id and qcom, board-id Lina Iyer
2015-11-19 14:36   ` Rob Herring
2015-11-19 15:36     ` [PATCH RFC 25/27] devicetree: bindings: Document qcom, msm-id and qcom,board-id Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 26/27] ARM64: dts: Add PSCI cpuidle support for MSM8916 Lina Iyer
2015-11-17 22:37 ` [PATCH RFC 27/27] ARM64: dts: Define CPU power domain " Lina Iyer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=564D8D56.1080206@baylibre.com \
    --to=mtitinger@baylibre$(echo .)com \
    --cc=linux-arm-kernel@lists$(echo .)infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox