Re: [PATCH v2] clocksource/drivers/sh_cmt: Always leave device running after probe

[Daniel Lezcano]

On 11/5/25 17:06, Niklas Söderlund wrote:
> On 2025-11-05 16:36:15 +0100, Daniel Lezcano wrote:
> > On 10/16/25 20:20, Niklas Söderlund wrote:
> > > The CMT device can be used as both a clocksource and a clockevent
> > > provider. The driver tries to be smart and power itself on and off, as
> > > well as enabling and disabling its clock when it's not in operation.
> > > This behavior is slightly altered if the CMT is used as an early
> > > platform device in which case the device is left powered on after probe,
> > > but the clock is still enabled and disabled at runtime.
> > >
> > > This has worked for a long time, but recent improvements in PREEMPT_RT
> > > and PROVE_LOCKING have highlighted an issue. As the CMT registers itself
> > > as a clockevent provider, clockevents_register_device(), it needs to use
> > > raw spinlocks internally as this is the context of which the clockevent
> > > framework interacts with the CMT driver. However in the context of
> > > holding a raw spinlock the CMT driver can't really manage its power
> > > state or clock with calls to pm_runtime_*() and clk_*() as these calls
> > > end up in other platform drivers using regular spinlocks to control
> > > power and clocks.
> >
> > So the fix is to remove PM management in the driver ?
>
> Yes. As I understand it we can't do runtime pm in these drivers as the
> core calls into the functions with the raw spinlock held. I hope we can
> improve this in future.


IIUC, the changes done for PREEMPT_RT prevent to use pm_runtime by 
functions running in atomic context because the spinlocks are actually 
mutexes.

But if PREEMPT_RT is not set, then everything is running as usual.

This change drops the PM while it should be working for kernel compiled 
without PREEMPT_RT.

I suggest to handle the case with/out PREEMPT_RT.

Hopefully pm_runtime will be fixed with PREEMPT_RT and you won't have to 
reintroduce pm_runtime in this driver but just remove the PREEMPT_RT case.


--
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog