[PATCH 2/2] ARM: dts: at91: sama7g5/sama7g5ek: align with Linux DT

[Eugen.Hristev at microchip.com]

On 3/7/22 1:05 PM, Tudor Ambarus - M18064 wrote:
> On 3/7/22 10:42, Eugen Hristev wrote:
>> Align the DT for sama7g5 SoC and sama7g5 EK board with Linux devicetree
>> in version 5.18.
> 
> cool
> 
>>
>> Some things remain still different, due to some things yet unimplemented in
>> certain drivers. These include in PMC, pinctrl, and others.
>>
>> Signed-off-by: Eugen Hristev <eugen.hristev at microchip.com>
>> ---
>>   arch/arm/dts/at91-sama7g5ek.dts | 727 ++++++++++++++++++++----
>>   arch/arm/dts/sama7g5.dtsi       | 942 +++++++++++++++++++++++++++-----
>>   2 files changed, 1419 insertions(+), 250 deletions(-)
>>
> 
> For arch/arm/dts/at91-sama7g5ek.dts:
> 1/ pinctrl_flx3_default: flx3_default  uses in linux bias-disable, you
> should update it too.

Linux should be updated.
If the bias is floating for these console pins, disconnecting the 
console cable stops the booting process because of stray characters 
detected on the console. The cable has an included pull-up which solves 
this.
(I will send a patch to fix Linux ), but it has to stay like this in 
Uboot until linux is fixed, otherwise boot fails

> 2/ I see linux's pinctrl_sdmmc0_default: sdmmc0_default splits the pins
> in cmd_data and ck_cd_rstn_vddsel. Can we do the same or it depends on
> something that's unsupported?
> 
We cannot do anything at the moment. It doesn't work because the pinctrl 
driver does not support subnodes like this:

sdmmc_pins : pins {
  subnode1 {
   some pin mux...
  };
  subnode2 {
   some other pin mux...
  };
}

It has to be fixed in the pinctrl driver. Until then, this has to stay 
like this otherwise pinctrl cannot probe the pins in uboot.

Could you double check the QSPI part and pins ? I have not tested this 
part with this patch, and I do not wish to break anything.
You can easily pull the patchset from here :

https://source.denx.de/u-boot/custodians/u-boot-at91/-/tree/testing

Thanks,
Eugen