Re: [PATCH v1 3/3] PM / AVS: rockchip-cpu-avs: add driver handling Rockchip cpu avs

[Rob Herring]

On Tue, Aug 23, 2016 at 05:29:16PM +0800, Finley Xiao wrote:
> 
> 
> å 2016/8/23 17:03, Heiko StÃbner åé:
> >Hi Finley,
> >
> >Am Dienstag, 23. August 2016, 12:10:27 schrieb Finley Xiao:
> >>å 2016/8/19 21:36, Rob Herring åé:
> >>>On Tue, Aug 16, 2016 at 10:38:59AM +0800, Finlye Xiao wrote:
> >>>>From: Finley Xiao <finley.xiao@xxxxxxxxxxxxxx>
> >>>>
> >>>>This patch supports adjusting opp's voltage according to leakage
> >>>>
> >>>>Signed-off-by: Finley Xiao <finley.xiao@xxxxxxxxxxxxxx>
> >>>>---
> >>>>
> >>>>   .../devicetree/bindings/power/rockchip-cpu-avs.txt |  37 +++
> >>>>   drivers/power/avs/Kconfig                          |   8 +
> >>>>   drivers/power/avs/Makefile                         |   1 +
> >>>>   drivers/power/avs/rockchip-cpu-avs.c               | 314
> >>>>   +++++++++++++++++++++ 4 files changed, 360 insertions(+)
> >>>>   create mode 100644
> >>>>   Documentation/devicetree/bindings/power/rockchip-cpu-avs.txt create
> >>>>   mode 100644 drivers/power/avs/rockchip-cpu-avs.c
> >>>>
> >>>>diff --git a/Documentation/devicetree/bindings/power/rockchip-cpu-avs.txt
> >>>>b/Documentation/devicetree/bindings/power/rockchip-cpu-avs.txt new file
> >>>>mode 100644
> >>>>index 0000000..90f6b08
> >>>>--- /dev/null
> >>>>+++ b/Documentation/devicetree/bindings/power/rockchip-cpu-avs.txt
> >>>>@@ -0,0 +1,37 @@
> >>>>+Rockchip cpu avs device tree bindings
> >>>>+-------------------------------------
> >>>>+
> >>>>+Under the same frequency, the operating voltage tends to decrease with
> >>>>+increasing leakage. so it is necessary to adjust opp's voltage according
> >>>>+to leakage for power.
> >>>>+
> >>>>+
> >>>>+Required properties:
> >>>>+- compatible: Should be one of the following.
> >>>>+  - "rockchip,rk3399-cpu-avs" - for RK3399 SoCs.
> >>>>+- leakage-volt-<name>: Named leakage-volt property. At runtime, the
> >>>>+  platform can find a cpu's cluster_id according to it's cpu_id and
> >>>>match
> >>>>+  leakage-volt-<name> property. The property is an array of 3-tuples
> >>>>+  items, and each item consists of leakage and voltage like
> >>>>+  <min-leakage-mA max-leakage-mA vol-uV>.
> >>>>+	min-leakage: minimum leakage in mA.
> >>>>+	max-leakage: maximum leakage in mA.
> >>>>+	vol: voltage in microvolt.
> >>>How do you determine these values? When do they vary?
> >>How do you determine these values?
> >>
> >>run antutu-benchmark
> >>leakage     freq 	min_volt
> >>30mA 	    1608MHz 	1025mV
> >>40mA 	    1608MHz 	1000mV
> >>50mA 	    1608MHz 	975mV
> >>
> >>  From the table, we see the min_volt decrease with increasing leakage.
> >>So we can set the default opp-microvolt 1025mV for 1608MHz,
> >>and add an leakage_volt_cluster0 property as follows.
> >>leakage_volt_cluster0 = <
> >>		/*  min(mA)   max(mA)    volt(uV)*/
> >>		    0         40        0
> >>		    41        50        (-25000)
> >>		    51        254       (-50000)
> >>
> >>>;
> >>If the leakage is between 41mA and 50mAïthe opp-microvolt will subtract
> >>25mVã If the leakage is between 41mA and 50mAïthe opp-microvolt will
> >>subtract 50mVã
> >>
> >>When do they vary?
> >>
> >>  From the code,
> >>cpufreq_online
> >>--cpufreq_driver->init(policy);  /* add new opp table */
> >>--blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
> >>CPUFREQ_START, policy);
> >>--cpufreq_init_policy
> >>----cpufreq_set_policy
> >>--------__cpufreq_governor(policy, CPUFREQ_GOV_START);
> >>
> >>The cpu's opp table is added in the init function(cpufreq_init)ïand I will
> >>register a cpufreq notifier, once the first cpu of cluster is onlined,
> >>my notifer will be called, and if the event is CPUFREQ_STARTïit will modify
> >>the opp-microvolt according to leakage_volt_cluster0.
> >I think what Rob was asking with "when do they vary" is, are these leakage
> >values the same for each and every rk3399 soc or are they determined
> >separately for each board-type using the rk3399.
> 
> Oh, sorry, I misunderstood.
> They are the same for every rk3399 soc.

Then I think the values should go in the kernel at least until we define 
a common way to handle this.

Rob