Re: [PATCH 1/2] dt-bindings: mmc: Document Aspeed SD controller

["Andrew Jeffery" <andrew@aj.id.au>]

On Thu, 11 Jul 2019, at 01:20, Rob Herring wrote:
> On Wed, Jul 10, 2019 at 8:16 AM Andrew Jeffery <andrew@aj.id.au> wrote:
> >
> > The ASPEED SD/SDIO/eMMC controller exposes two slots implementing the
> > SDIO Host Specification v2.00, with 1 or 4 bit data buses, or an 8 bit
> > data bus if only a single slot is enabled.
> >
> > Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
> > ---
> >  .../bindings/mmc/sdhci-of-aspeed.yaml         | 91 +++++++++++++++++++
> 
> aspeed,sdhci.yaml

Ack. Previously I had separate documents for different compatibles, wasn't sure
how to tackle name one document covering multiple compatibles.

> 
> >  1 file changed, 91 insertions(+)
> >  create mode 100644 Documentation/devicetree/bindings/mmc/sdhci-of-aspeed.yaml
> >
> > diff --git a/Documentation/devicetree/bindings/mmc/sdhci-of-aspeed.yaml b/Documentation/devicetree/bindings/mmc/sdhci-of-aspeed.yaml
> > new file mode 100644
> > index 000000000000..e98a2ac4d46d
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/mmc/sdhci-of-aspeed.yaml
> > @@ -0,0 +1,91 @@
> > +# SPDX-License-Identifier: GPL-2.0-or-later
> 
> The preferred license is (GPL-2.0 OR BSD-2-Clause) if that is okay with you.

I should poke people internally again, I had asked, just haven't got a clear
answer. What was the justification for the preference (just so I can pass that
on)?

> 
> > +%YAML 1.2
> > +---
> > +$id: http://devicetree.org/schemas/mmc/sdhci-of-aspeed.yaml#
> > +$schema: http://devicetree.org/meta-schemas/core.yaml#
> > +
> > +title: ASPEED SD/SDIO/eMMC Controller
> > +
> > +maintainers:
> > +  - Andrew Jeffery <andrew@aj.id.au>
> > +  - Ryan Chen <ryanchen.aspeed@gmail.com>
> > +
> > +description: |+
> > +  The ASPEED SD/SDIO/eMMC controller exposes two slots implementing the SDIO
> > +  Host Specification v2.00, with 1 or 4 bit data buses, or an 8 bit data bus if
> > +  only a single slot is enabled.
> > +
> > +  The two slots are supported by a common configuration area. As the SDHCIs for
> > +  the slots are dependent on the common configuration area, they are described
> > +  as child nodes.
> > +
> > +properties:
> > +  compatible:
> > +    enum: [ aspeed,ast2400-sdc, aspeed,ast2500-sdc ]
> > +  reg:
> > +    description: Common configuration registers
> 
> This should have a 'maxItems: 1'. Same for the child reg.

Ack.

> 
> > +  ranges: true
> > +  clocks:
> > +    maxItems: 1
> > +    description: The SD/SDIO controller clock gate
> > +  sdhci:
> 
> This needs to be a pattern (under patternProperties) as these have
> unit-addresses.

Ah, I didn't think about that.

> 
> > +    type: object
> > +    properties:
> > +      compatible:
> > +        allOf:
> > +          - enum: [ aspeed,ast2400-sdhci, aspeed,ast2500-sdhci ]
> > +          - const: sdhci
> 
> This condition can never be true. What you need is s/allOf/items/.
> However, 'sdhci' is not really a useful compatible because every
> implementation has quirks, so I'd drop it.

Yeah, I was tossing up whether to include "sdhci". I'll drop it as you
suggest.

Pity the shorthand doesn't work how I expected. Might explain
some of the behaviour I was seeing with the bindings make targets
though.

> 
> > +      reg:
> > +        description: The SDHCI registers
> > +      clocks:
> > +        maxItems: 1
> > +        description: The SD bus clock
> > +      slot:
> > +        allOf:
> > +          - $ref: /schemas/types.yaml#/definitions/uint32
> > +          - enum: [0, 1]
> 
> Is this really needed? Offset 0x100 is slot 0 and offset 0x200 is slot
> 1. Does that ever change?

It doesn't in the SoCs at hand, but the downstream impact (driver
implementation) is that you need to derive the slot index from the
address and I was trying to avoid that. The slot index influences
some bit index calculations.

> 
> > +      interrupts:
> > +        maxItems: 1
> > +        description: The SD interrupt shared between both slots
> > +    required:
> > +      - compatible
> > +      - reg
> > +      - clocks
> > +      - slot
> > +      - interrupts
> > +
> > +required:
> > +  - compatible
> > +  - reg
> > +  - ranges
> > +  - clocks
> 
> #address-cells and #size-cells are required too.

Ack.

> 
> You should also add 'additionalProperties: false' here so other random
> properties can't be present.

Ah yes. Will fix.

> 
> > +
> > +examples:
> > +  - |
> > +    #include <dt-bindings/clock/aspeed-clock.h>
> > +    sdc@1e740000 {
> > +            compatible = "aspeed,ast2500-sdc";
> > +            reg = <0x1e740000 0x100>;
> > +            #address-cells = <1>;
> > +            #size-cells = <1>;
> > +            ranges;
> 
> It's preferred to limit the range here and then the child addresses
> are 0x100 and 0x200.

Okay. Was just trying to dodge mental arithmetic where I could, but
if that's the preference then I'll fix it up.

Thanks for the review!

Andrew

> 
> > +            clocks = <&syscon ASPEED_CLK_GATE_SDCLK>;
> > +
> > +            sdhci0: sdhci@1e740100 {
> > +                    compatible = "aspeed,ast2500-sdhci", "sdhci";
> > +                    reg = <0x1e740100 0x100>;
> > +                    slot = <0>;
> > +                    interrupts = <26>;
> > +                    sdhci,auto-cmd12;
> > +                    clocks = <&syscon ASPEED_CLK_SDIO>;
> > +            };
> > +
> > +            sdhci1: sdhci@1e740200 {
> > +                    compatible = "aspeed,ast2500-sdhci", "sdhci";
> > +                    reg = <0x1e740200 0x100>;
> > +                    slot = <1>;
> > +                    interrupts = <26>;
> > +                    sdhci,auto-cmd12;
> > +                    clocks = <&syscon ASPEED_CLK_SDIO>;
> > +            };
> > +    };
> > --
> > 2.20.1
> >
>