Re: [PATCH v2 2/5] cpuidle: psci: Mark as PREEMPT_RT safe

[Krzysztof Kozlowski]

On 12/01/2023 12:00, Sebastian Andrzej Siewior wrote:
> On 2022-12-19 16:15:00 [+0100], Krzysztof Kozlowski wrote:
>> The PSCI cpuidle power domain in power_off callback uses
>> __this_cpu_write() so it is PREEMPT_RT safe.  This allows to use it in
> 
> Why does __this_cpu_write() matter here?

I'll reword to "not using sleeping primitives nor spinlock_t"

> 
>> Realtime kernels and solves errors like:
>>
>>   BUG: scheduling while atomic: swapper/2/0/0x00000002
>>   Hardware name: Qualcomm Technologies, Inc. Robotics RB5 (DT)
>>   Call trace:
>>    dump_backtrace.part.0+0xe0/0xf0
>>    show_stack+0x18/0x40
>>    dump_stack_lvl+0x68/0x84
>>    dump_stack+0x18/0x34
>>    __schedule_bug+0x60/0x80
>>    __schedule+0x628/0x800
>>    schedule_rtlock+0x28/0x5c
>>    rtlock_slowlock_locked+0x360/0xd30
>>    rt_spin_lock+0x88/0xb0
>>    genpd_lock_nested_spin+0x1c/0x30
>>    genpd_power_off.part.0.isra.0+0x20c/0x2a0
>>    genpd_runtime_suspend+0x150/0x2bc
>>    __rpm_callback+0x48/0x170
>>    rpm_callback+0x6c/0x7c
>>    rpm_suspend+0x108/0x660
>>    __pm_runtime_suspend+0x4c/0x8c
>>    __psci_enter_domain_idle_state.constprop.0+0x54/0xe0
>>    psci_enter_domain_idle_state+0x18/0x2c
>>    cpuidle_enter_state+0x8c/0x4e0
>>    cpuidle_enter+0x38/0x50
>>    do_idle+0x248/0x2f0
>>    cpu_startup_entry+0x24/0x30
>>    secondary_start_kernel+0x130/0x154
>>    __secondary_switched+0xb0/0xb4
> 
> This is to a sleeping lock (spinlock_t) in an IRQ-off region which
> starts in do_idle(). You could describe the problem and to solution you
> aim for instead pasting a backtrace into the commit description and
> adding a flow in the code.

I'll extend the description.

> 
> I don't see how your commit description matches your change in code. One
> might think "Oh. If I see this pattern, I need an irqsafe lock to fix
> it".


Best regards,
Krzysztof