question about PMU interrupt in binding

question about PMU interrupt in binding

[Stuart Yoder]

Stephen/Will,

Had a question about the definition of "interrupts" in the
PMU binding:

> commit d84c47837d800c0db1c9f2015a1565504417f812
> Author: Stephen Boyd <sboyd@codeaurora.org>
> Date:   Fri Feb 7 21:01:24 2014 +0000
> 
>     devicetree: bindings: Document Krait performance monitor units (PMU)
> 
>     Document the Krait PMU compatible string.
> 
>     Cc: <devicetree@vger.kernel.org>
>     Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
>     Signed-off-by: Will Deacon <will.deacon@arm.com>
> 
> diff --git a/Documentation/devicetree/bindings/arm/pmu.txt b/Documentation/devicetree/bindings/arm/pmu.txt
> index 3e1e498..ce73144 100644
> --- a/Documentation/devicetree/bindings/arm/pmu.txt
> +++ b/Documentation/devicetree/bindings/arm/pmu.txt
> @@ -16,7 +16,14 @@ Required properties:
>         "arm,arm11mpcore-pmu"
>         "arm,arm1176-pmu"
>         "arm,arm1136-pmu"
> -- interrupts : 1 combined interrupt or 1 per core.
> +       "qcom,krait-pmu"
> +- interrupts : 1 combined interrupt or 1 per core. If the interrupt is a per-cpu
> +               interrupt (PPI) then 1 interrupt should be specified.

If the interrupt is per-cpu do you mean 1 interrupt _per cpu_ should
be specified?  As written it seems to say _only_  "1 interrupt" should be
specificed.

Thanks,
Stuart

question about PMU interrupt in binding

[Mark Rutland]

On Tue, Feb 24, 2015 at 03:44:27PM +0000, Stuart Yoder wrote:
> Stephen/Will,
> 
> Had a question about the definition of "interrupts" in the
> PMU binding:
> 
> > commit d84c47837d800c0db1c9f2015a1565504417f812
> > Author: Stephen Boyd <sboyd@codeaurora.org>
> > Date:   Fri Feb 7 21:01:24 2014 +0000
> > 
> >     devicetree: bindings: Document Krait performance monitor units (PMU)
> > 
> >     Document the Krait PMU compatible string.
> > 
> >     Cc: <devicetree@vger.kernel.org>
> >     Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
> >     Signed-off-by: Will Deacon <will.deacon@arm.com>
> > 
> > diff --git a/Documentation/devicetree/bindings/arm/pmu.txt b/Documentation/devicetree/bindings/arm/pmu.txt
> > index 3e1e498..ce73144 100644
> > --- a/Documentation/devicetree/bindings/arm/pmu.txt
> > +++ b/Documentation/devicetree/bindings/arm/pmu.txt
> > @@ -16,7 +16,14 @@ Required properties:
> >         "arm,arm11mpcore-pmu"
> >         "arm,arm1176-pmu"
> >         "arm,arm1136-pmu"
> > -- interrupts : 1 combined interrupt or 1 per core.
> > +       "qcom,krait-pmu"
> > +- interrupts : 1 combined interrupt or 1 per core. If the interrupt is a per-cpu
> > +               interrupt (PPI) then 1 interrupt should be specified.
> 
> If the interrupt is per-cpu do you mean 1 interrupt _per cpu_ should
> be specified?  As written it seems to say _only_  "1 interrupt" should be
> specificed.

The specification as written is correct.

If it's a per-cpu interrupt (i.e. a PPI), we'll recognise this fact and
handle it appropriately in the backend, on the assumption that if the
PMU is connected via a PPI all CPUs share this PPI.

Thanks,
Mark.

question about PMU interrupt in binding

[Stuart Yoder]

> -----Original Message-----
> From: Mark Rutland [mailto:mark.rutland at arm.com]
> Sent: Tuesday, February 24, 2015 10:20 AM
> To: Yoder Stuart-B08248
> Cc: sboyd at codeaurora.org; Will Deacon; linux-arm-kernel at lists.infradead.org
> Subject: Re: question about PMU interrupt in binding
> 
> On Tue, Feb 24, 2015 at 03:44:27PM +0000, Stuart Yoder wrote:
> > Stephen/Will,
> >
> > Had a question about the definition of "interrupts" in the
> > PMU binding:
> >
> > > commit d84c47837d800c0db1c9f2015a1565504417f812
> > > Author: Stephen Boyd <sboyd@codeaurora.org>
> > > Date:   Fri Feb 7 21:01:24 2014 +0000
> > >
> > >     devicetree: bindings: Document Krait performance monitor units (PMU)
> > >
> > >     Document the Krait PMU compatible string.
> > >
> > >     Cc: <devicetree@vger.kernel.org>
> > >     Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
> > >     Signed-off-by: Will Deacon <will.deacon@arm.com>
> > >
> > > diff --git a/Documentation/devicetree/bindings/arm/pmu.txt
> b/Documentation/devicetree/bindings/arm/pmu.txt
> > > index 3e1e498..ce73144 100644
> > > --- a/Documentation/devicetree/bindings/arm/pmu.txt
> > > +++ b/Documentation/devicetree/bindings/arm/pmu.txt
> > > @@ -16,7 +16,14 @@ Required properties:
> > >         "arm,arm11mpcore-pmu"
> > >         "arm,arm1176-pmu"
> > >         "arm,arm1136-pmu"
> > > -- interrupts : 1 combined interrupt or 1 per core.
> > > +       "qcom,krait-pmu"
> > > +- interrupts : 1 combined interrupt or 1 per core. If the interrupt is a per-cpu
> > > +               interrupt (PPI) then 1 interrupt should be specified.
> >
> > If the interrupt is per-cpu do you mean 1 interrupt _per cpu_ should
> > be specified?  As written it seems to say _only_  "1 interrupt" should be
> > specificed.
> 
> The specification as written is correct.
> 
> If it's a per-cpu interrupt (i.e. a PPI), we'll recognise this fact and
> handle it appropriately in the backend, on the assumption that if the
> PMU is connected via a PPI all CPUs share this PPI.

Thanks.

Stuart