Re: [PATCH v3 5/5] dt-bindings: Update the riscv,isa string description

[Rob Herring]

On Wed, Jul 31, 2019 at 6:58 PM Atish Patra <atish.patra@xxxxxxx> wrote:
>
> Since the RISC-V specification states that ISA description strings are
> case-insensitive, there's no functional difference between mixed-case,
> upper-case, and lower-case ISA strings. Thus, to simplify parsing,
> specify that the letters present in "riscv,isa" must be all lowercase.
>
> Suggested-by: Paul Walmsley <paul.walmsley@xxxxxxxxxx>
> Signed-off-by: Atish Patra <atish.patra@xxxxxxx>
> ---
>  Documentation/devicetree/bindings/riscv/cpus.yaml | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/Documentation/devicetree/bindings/riscv/cpus.yaml b/Documentation/devicetree/bindings/riscv/cpus.yaml
> index c899111aa5e3..4f0acb00185a 100644
> --- a/Documentation/devicetree/bindings/riscv/cpus.yaml
> +++ b/Documentation/devicetree/bindings/riscv/cpus.yaml
> @@ -46,10 +46,12 @@ properties:
>            - rv64imafdc
>      description:
>        Identifies the specific RISC-V instruction set architecture
> -      supported by the hart.  These are documented in the RISC-V
> +      supported by the hart. These are documented in the RISC-V
>        User-Level ISA document, available from
>        https://riscv.org/specifications/
>
> +      Letters in the riscv,isa string must be all lowercase.
> +

The schemas are case sensitive this looks pretty pointless without the
context of the commit msg. Can you prefix with 'While the
specification is case insensitive, "

For some background, FDT generally always has been case sensitive too
(dtc won't merge/override nodes/properties with differing case). It's
really only some older true OF systems that were case insensitive. The
kernel had a mixture of case sensitive and insensitive comparisons
somewhat depending on the arch and whether of_prop_cmp/of_node_cmp or
str*cmp functions were used. There's been a lot of clean-up and now
most comparisons are case sensitive with only Sparc having some
deviation.

Rob