Re: [PATCH v1 3/9] ACPI: processor: idle: Return failure when get lpi_state->arch_flags failed

[lihuisong (C)]

在 2025/10/23 18:09, Rafael J. Wysocki 写道:
> On Thu, Oct 23, 2025 at 11:59 AM lihuisong (C) <lihuisong@xxxxxxxxxx> wrote:
> > 在 2025/10/22 3:36, Rafael J. Wysocki 写道:
> > > On Mon, Sep 29, 2025 at 11:38 AM Huisong Li <lihuisong@xxxxxxxxxx> wrote:
> > > > The architecture specific context loss flags is important for ARM.
> > > > And this flag is used to control the execution of different code
> > > > flows in acpi_processor_ffh_lpi_enter().
> > > >
> > > > So it is better to return failure when get lpi_state->arch_flags
> > > > failed.
> > > A failure means no idle states at all.
> > Actually, I didn't know why driver should continue to do cpu idle
> > scaling if the idle state doesn't meet the developer's expectations.🙂
> There may be a firmware bug in one state description, but it doesn't
> mean that the other states are unusable, does it?
Ack
> > > Wouldn't it be better to skip the state with invalid arch flags?
> > This arch flags is important.  And acpi_processor_ffh_lpi_enter will use it.
> > There is no other place to verify its validity. so here do it.
> > This check is just to prevent potential issues in cpuidle scaling later.
> I'm saying to treat this particular state as invalid and skip it
> without rejecting all of the other states that may be valid.

Ok, handle this case the same as illegal entry method in patch 2/9 if it 
is ok to you.