[Linux-stm32] [PATCH v7 10/10] ARM: dts: stm32: add STM32MP1-based Phytec board

[Alexandre TORGUE]

Hello Ahmad

On 4/3/23 11:28, Ahmad Fatoum wrote:
> Hello Alexandre,
> 
> On 03.04.23 11:15, Alexandre TORGUE wrote:
>> On 3/30/23 07:04, Steffen Trumtrar wrote:
>>> Add the Phytec STM32MP1-3 Dev board. The devboard uses a Phytec
>>> stm32m157c-som.
>>>
>>> Signed-off-by: Steffen Trumtrar <s.trumtrar at pengutronix.de>
>>> ---
> 
> [snip]
> 
>>> +/ {
>>> +    model = "PHYTEC phyCORE-STM32MP1-3 Dev Board";
>>> +    compatible = "phytec,phycore-stm32mp1-3",
>>> +             "phytec,phycore-stm32mp157c-som", "st,stm32mp157";
>>> +
>>> +    aliases {
>>> +        mmc0 = &sdmmc1;
>>> +        mmc1 = &sdmmc2;
>>> +        mmc2 = &sdmmc3;
>>
>> mmc aliases are still used in linux?
> 
> Since kernel commit fa2d0aa96941 ("mmc: core: Allow setting slot
> index via device tree alias") added with v5.10-rc1, aliases
> dictate kernel numbering of /dev/mmcblk devices, so it's good
> to have these aliases for consistency as the alternative of
> using PARTUUID breaks down once the same image is flashed to
> both SD and eMMC and use of an initrd is not always practical.

Oh ok. I thought the opposite. Sorry for this mistake.

Cheers
Alex

>>> +&sdmmc1 {
>>> +    secure-status = "disabled";
>>> +};
>>> +
>>> +&sdmmc2 {
>>> +    status = "okay";
>>> +    secure-status = "disabled";
>>> +};
>>
>> What is the need to put the secure status disabled for SDMMC nodes ?
> 
> TF-A parses them, but for SDMMC, it should have no effect, so I agree
> they can be removed.
> 
> Cheers,
> Ahmad
> 
>