Re: [PATCH v3] cpuidle: psci: Do not suspend topology CPUs on PREEMPT_RT

[Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>]

On 24/01/2023 16:34, Sudeep Holla wrote:
> On Thu, Jan 19, 2023 at 07:42:28PM +0100, Krzysztof Kozlowski wrote:
>> The runtime Power Management of CPU topology is not compatible with
>> PREEMPT_RT:
>> 1. Core cpuidle path disables IRQs.
>> 2. Core cpuidle calls cpuidle-psci.
>> 3. cpuidle-psci in __psci_enter_domain_idle_state() calls
>>    pm_runtime_put_sync_suspend() and pm_runtime_get_sync() which use
>>    spinlocks (which are sleeping on PREEMPT_RT).
>>
>> Deep sleep modes are not a priority of Realtime kernels because the
>> latencies might become unpredictable.  On the other hand the PSCI CPU
>> idle power domain is a parent of other devices and power domain
>> controllers, thus it cannot be simply skipped (e.g. on Qualcomm SM8250).
>>
>> Disable the runtime PM calls from cpuidle-psci, which effectively stops
>> suspending the cpuidle PSCI domain.  This is a trade-off between making
>> PREEMPT_RT working and still having a proper power domain hierarchy in
>> the system.
>>
>> Cc: Adrien Thierry <athierry@redhat.com>
>> Cc: Brian Masney <bmasney@redhat.com>
>> Cc: linux-rt-users@vger.kernel.org
>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
>>
>> ---
>>
>> Changes since v1:
>> 1. Re-work commit msg.
>> 2. Add note to Kconfig.
>>
>> Several other patches were dropped, as this is the only one actually
>> needed.  It effectively stops PSCI cpuidle power domains from suspending
>> thus solving all other issues I experienced.
>> ---
>>  drivers/cpuidle/Kconfig.arm    | 3 +++
>>  drivers/cpuidle/cpuidle-psci.c | 4 ++--
>>  2 files changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/cpuidle/Kconfig.arm b/drivers/cpuidle/Kconfig.arm
>> index 747aa537389b..24429b5bfd1c 100644
>> --- a/drivers/cpuidle/Kconfig.arm
>> +++ b/drivers/cpuidle/Kconfig.arm
>> @@ -24,6 +24,9 @@ config ARM_PSCI_CPUIDLE
>>  	  It provides an idle driver that is capable of detecting and
>>  	  managing idle states through the PSCI firmware interface.
>>  
>> +	  The driver is not yet compatible with PREEMPT_RT: no idle states will
>> +	  be entered by CPUs on such kernel.
>> +
> 
> Any particular reason for even compiling this file in or allowing the
> ARM_PSCI_CPUIDLE when PREEMPT_RT=y ? If we can't enter idle states, we
> can as well compile this file out ?

It's the power domain sued for other devices, so we need it. Otherwise
other devices will keep waiting for this missing power domain provider.

Best regards,
Krzysztof