Re: [PATCH 12/17] ARM: OMAP2+: remove misuse of IRQF_NO_SUSPEND flag

[Tony Lindgren <tony@atomide.com>]

* Sudeep Holla <sudeep.holla@arm.com> [151013 03:46]:
> 
> 
> On 12/10/15 21:28, Tony Lindgren wrote:
> >* Tony Lindgren <tony@atomide.com> [151012 13:27]:
> >>* Sudeep Holla <sudeep.holla@arm.com> [150921 08:52]:
> >>>The IRQF_NO_SUSPEND flag is used to identify the interrupts that should
> >>>be left enabled so as to allow them to work as expected during the
> >>>suspend-resume cycle, but doesn't guarantee that it will wake the system
> >>>from a suspended state, enable_irq_wake is recommended to be used for
> >>>the wakeup.
> >>>
> >>>This patch removes the use of IRQF_NO_SUSPEND flags replacing it with
> >>>enable_irq_wake instead.
> >>
> >>Applying into omap-for-v4.4/cleanup thanks.
> >
> >Actually I don't think this does the right thing. The interrupts
> >in the $subject patch are in the always on powerdomain, and we really
> 
> Agreed
> 
> >want them to be excluded from the suspend.
> >
> 
> OK but what's wrong with this patch. At-least the name suggest it's a
> wakeup interrupt. And using IRQF_NO_SUSPEND for the wakeup interrupt is
> simply wrong.

Hmm so if we have a separate always on irq controller for the wake-up events
and we want to keep it always on and exclude it from any suspend related
things.. Why would we not use IRQF_NO_SUSPEND on it?

Above you say "The IRQF_NO_SUSPEND flag is used to identify the interrupts
that should be left enabled so as to allow them to work as expected during
the suspend-resume cycle..." and that's exactly what we want to do here :)

For the dedicated wake-up interrupts, we have separate registers to enable
and disable them. The $subject irq is the shared interrupt that allows
making use of the pin specific wake-up interrupts, and for those yes we
are using enable_irq_wake().

> >So not applying without further explanations.
> >
> 
> But I don't understand the real need for IRQF_NO_SUSPEND over wakeup APIs ?

Because in the $subject case we just want to always keep it on and
never suspend it. It's unrelated to the wakeup APIs at least for the
omap related SoCs.

Regards,

Tony