Re: [PATCH v5 02/10] dt-bindings: aspeed-sgpio: Add ast2600 sgpio compatibles.

[Andrew Jeffery]

On Fri, 11 Jun 2021, at 01:53, Rob Herring wrote:
> On Tue, Jun 08, 2021 at 06:25:37PM +0800, Steven Lee wrote:
> > AST2600 SoC has 2 SGPIO master interfaces one with 128 pins another one
> > with 80 pins. Add ast2600-sgpiom0-80 and ast2600-sgpiom-128 compatibles
> > and update descriptions to introduce the max number of available gpio
> > pins that AST2600 supported.
> > 
> > Signed-off-by: Steven Lee <steven_lee@xxxxxxxxxxxxxx>
> > Reviewed-by: Andrew Jeffery <andrew@xxxxxxxx>
> > ---
> >  Documentation/devicetree/bindings/gpio/aspeed,sgpio.yaml | 9 ++++++---
> >  1 file changed, 6 insertions(+), 3 deletions(-)
> > 
> > diff --git a/Documentation/devicetree/bindings/gpio/aspeed,sgpio.yaml b/Documentation/devicetree/bindings/gpio/aspeed,sgpio.yaml
> > index b2ae211411ff..0e42eded3c1e 100644
> > --- a/Documentation/devicetree/bindings/gpio/aspeed,sgpio.yaml
> > +++ b/Documentation/devicetree/bindings/gpio/aspeed,sgpio.yaml
> > @@ -10,9 +10,10 @@ maintainers:
> >    - Andrew Jeffery <andrew@xxxxxxxx>
> >  
> >  description:
> > -  This SGPIO controller is for ASPEED AST2500 SoC, it supports up to 80 full
> > -  featured Serial GPIOs. Each of the Serial GPIO pins can be programmed to
> > -  support the following options
> > +  This SGPIO controller is for ASPEED AST2400, AST2500 and AST2600 SoC,
> > +  AST2600 have two sgpio master one with 128 pins another one with 80 pins,
> > +  AST2500/AST2400 have one sgpio master with 80 pins. Each of the Serial
> > +  GPIO pins can be programmed to support the following options
> >    - Support interrupt option for each input port and various interrupt
> >      sensitivity option (level-high, level-low, edge-high, edge-low)
> >    - Support reset tolerance option for each output port
> > @@ -25,6 +26,8 @@ properties:
> >      enum:
> >        - aspeed,ast2400-sgpio
> >        - aspeed,ast2500-sgpio
> > +      - aspeed,ast2600-sgpiom-80
> > +      - aspeed,ast2600-sgpiom-128
> 
> If the number of GPIOs is the only difference, then I don't think you 
> should get rid of ngpios. It's one thing if it varies from one SoC to 
> the next, but if something is per instance we should have a property.
> 

There are two issues:

1. The maximum number of GPIOs supported by the controller
2. The maximum number of GPIOs supported by the platform

These are different because of what the controller does - here's some previous discussion on the topic:

https://lore.kernel.org/linux-gpio/f2875111-9ba9-43b7-b2a4-d00c8725f5a0@xxxxxxxxxxxxxxxx/

We've used ngpios to describe 2; this decision was made prior to the 2600 design - the SGPIO controller for both the 2400 and 2500 supported a maximum of 80 GPIOs. With the 2600 we have to differentiate between the two SGPIO controllers because they support a different maximum number of GPIOs. The proposed approach of different compatibles keeps the behaviour of ngpios the same across all controller implementations.

Cheers,

Andrew