Re: [PATCH 08/13] clk: qcom: hfpll: CLK_IGNORE_UNUSED

[Jorge Ramirez]

On 1/22/19 19:47, Stephen Boyd wrote:
> Quoting Jorge Ramirez (2019-01-17 02:46:21)
>> On 1/17/19 11:08, Viresh Kumar wrote:
>>> On 17-01-19, 09:38, Jorge Ramirez wrote:
>>>> COMMON_CLK_DISABLED_UNUSED relies on the enable_count reference counter
>>>> to disable the clocks that were enabled by the firwmare and not by the
>>>> drivers.
>>>>
>>>> the cpufreq driver does not enable the cpu clock.
>>>>
>>>> so when clk_change_rate is called, the enable_count counter is not
>>>> incremented and therefore it just remains null since this was enabled by
>>>> the firmware.
>>>>
>>>> I tried doing:
>>>>
>>>> diff --git a/drivers/cpufreq/cpufreq-dt.c b/drivers/cpufreq/cpufreq-dt.c
>>>> index e58bfcb..5a9f83e 100644
>>>> --- a/drivers/cpufreq/cpufreq-dt.c
>>>> +++ b/drivers/cpufreq/cpufreq-dt.c
>>>> @@ -124,6 +124,10 @@ static int resources_available(void)
>>>>                 return ret;
>>>>         }
>>>>
>>>> + ret = clk_prepare_enable(cpu_clk);
>>>> + if (ret)
>>>> +         return ret;
>>>> +
>>>>         clk_put(cpu_clk);
>>>>
>>>>         name = find_supply_name(cpu_dev);
>>>>
>>>>
>>>> and that removed the need for CLK_IGNORE_UNUSED. But I am not sure of
>>>> the system wide consequences of that change to cpufreq.
>>>
>>> If the cpufreq driver enables it then it should disable it on exit as
>>> well, right ? And in that case if you unload your driver's module, you
>>> will hang the system as the clock will get disabled :)
>>
>> ah, of course, sorry was over-thinking this thing :)
>>
>>>
>>> Every other platform must either be marking it with CLK_IGNORE_UNUSED
>>> or they must be doing clk_enable from somewhere, maybe the CPU online
>>> path, not sure though.
>>>
>>
>> since this clock is enabled by the firmware, it seems to me that using
>> CLK_IGNORE_UNUSED remains the best option.
>>
> 
> What do you do about CPUs being offlined? Presumably when the CPU is
> gone the system doesn't need to keep the clk enabled anymore.
> 
> 

the kernel does not take any action - it is under firmware control; and
since the clock is shared between all cores I there will have to be some
involvement not only at TF-A/firmware level but also at the TrustedOS
level for when the last core is offlined.

I dont have visibility on either of those projects though.