Re: [PATCH v2 3/3] spi: spi_amd: Update Kconfig dependencies

["Rangoju, Raju" <raju.rangoju@amd.com>]

On 5/28/2025 1:02 PM, Geert Uytterhoeven wrote:
> Hi Raju,
> 
> Thanks for your patch, which is now commit dbb79974193a2932 ("spi:
> spi_amd: Update Kconfig dependencies") upstream.
> 
> On Fri, 16 May 2025 at 12:09, Raju Rangoju <Raju.Rangoju@amd.com> wrote:
>> Add X86 and SPI_MEM as dependencies for the spi_amd driver to ensure it is
>> built only on relevant platforms and with the required SPI memory
>> framework.
> 
> This sounds like you are restricting the driver further (BTW, what are
> "relevant platforms"?)...

AMD x86 platforms.

> 
>> Co-developed-by: Akshata MukundShetty <akshata.mukundshetty@amd.com>
>> Signed-off-by: Akshata MukundShetty <akshata.mukundshetty@amd.com>
>> Signed-off-by: Raju Rangoju <Raju.Rangoju@amd.com>
> 
>> --- a/drivers/spi/Kconfig
>> +++ b/drivers/spi/Kconfig
>> @@ -1267,7 +1267,8 @@ config SPI_ZYNQMP_GQSPI
>>   config SPI_AMD
>>          tristate "AMD SPI controller"
>>          depends on PCI
>> -       depends on SPI_MASTER || COMPILE_TEST
>> +       depends on SPI_MASTER || X86 || COMPILE_TEST
> 
> ... but this broadens it instead?
> And how can it be used on X86 if SPI_MASTER=n?
> 
> Seeing an ACPI match table in drivers/spi/spi-amd.c, perhaps you wanted
> to have a dependency on ACPI (or X86 && ACPI) somewhere?
> 
> BTW, as you now have a single Kconfig symbol gating both spi-amd.c
> and spi-amd-pci.c, this means you can no longer build spi-amd.c if
> CONFIG_PCI=n.

Sure Greert. Thanks for the suggestion.
We are planning to have separate Kconfig for both ACPI and PCI based 
drivers in the upcoming series.

> 
>> +       depends on SPI_MEM
>>          help
>>            Enables SPI controller driver for AMD SoC.
> 
> Gr{oetje,eeting}s,
> 
>                          Geert
>