Re: [PATCH 1/6] soc: qcom: Move some socinfo defines to the header, expand them

[Elliot Berman]

On Fri, Apr 05, 2024 at 10:41:29AM +0200, Konrad Dybcio wrote:
> In preparation for parsing the chip "feature code" (FC) and "product
> code" (PC) (essentially the parameters that let us conclusively
> characterize the sillicon we're running on, including various speed
> bins), move the socinfo version defines to the public header and
> include some more FC/PC defines.
> 
> Signed-off-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxx>
> ---
>  drivers/soc/qcom/socinfo.c       |  8 --------
>  include/linux/soc/qcom/socinfo.h | 36 ++++++++++++++++++++++++++++++++++++
>  2 files changed, 36 insertions(+), 8 deletions(-)
> 
..
> diff --git a/include/linux/soc/qcom/socinfo.h b/include/linux/soc/qcom/socinfo.h
..
> @@ -74,4 +84,30 @@ struct socinfo {
>  	__le32 boot_core;
>  };
>  
> +/* Internal feature codes */
> +enum feature_code {
> +	/* External feature codes */
> +	SOCINFO_FC_UNKNOWN = 0x0,
> +	SOCINFO_FC_AA,
> +	SOCINFO_FC_AB,
> +	SOCINFO_FC_AC,
> +	SOCINFO_FC_AD,
> +	SOCINFO_FC_AE,
> +	SOCINFO_FC_AF,
> +	SOCINFO_FC_AG,
> +	SOCINFO_FC_AH,
> +	SOCINFO_FC_EXT_RESERVE,
> +};

SOCINFO_FC_EXT_RESERVE was a convenient limit since we mapped
SOCINFO_FC_AA -> string "AA" via an array, and we've only needed the 8
feature codes so far.

We should remove the EXT_RESERVE and test for the Y0-YF (internal
feature code) values instead.

> +
> +/* Internal feature codes */
> +/* Valid values: 0 <= n <= 0xf */
> +#define SOCINFO_FC_Yn(n)		(0xf1 + n)
> +#define SOCINFO_FC_INT_RESERVE		SOCINFO_FC_Yn(0x10)

We probably should've named this SOCINFO_FC_INT_MAX. Reserve implies
it's reserved for some future use, but it's really the max value it
could be.

> +
> +/* Product codes */
> +#define SOCINFO_PC_UNKNOWN		0
> +/* Valid values: 0 <= n <= 8, the rest is reserved */
> +#define SOCINFO_PCn(n)			(n + 1)
> +#define SOCINFO_PC_RESERVE		(BIT(31) - 1)

Similar comments here as the SOCINFO_FC_EXT_*. It's more like known
values are [0,8], but more values could come in future chipsets.

> +
>  #endif
>