Re: [PATCH 0/7] dra7: Fixes for MMC devicetree node

[Kishon Vijay Abraham I]

Hi,

On Tuesday 06 June 2017 04:21 PM, Tony Lindgren wrote:
> * Kishon Vijay Abraham I <kishon@xxxxxx> [170606 03:40]:
>> omap3-overo-base.dtsi has the following dt node for mmc.
>> /* optional on board WiFi */
>> &mmc2 {
>>         pinctrl-names = "default";
>>         pinctrl-0 = <&mmc2_pins>;
>>         vmmc-supply = <&w3cbw003c_npoweron>;
>>         vqmmc-supply = <&w3cbw003c_bt_nreset>;
>>         vmmc_aux-supply = <&w3cbw003c_wifi_nreset>;
>>         bus-width = <4>;
>>         cap-sdio-irq;
>>         non-removable;
>> };
>>
>> It has both "vqmmc-supply" and "vmmc_aux-supply". However omap-hsmmc driver has
>> never parsed vqmmc so far. So I'd assume bt was never functional and it's used
>> only for wifi.
> 
> Hmm yeah the vqmmc-supply probably got copied from some
> other SoC for w3cbw003c.

Okay. So I think we can remove vqmmc-supply here without breaking any existing
functionality.
> 
>> So for the change suggested by Ulf, Can I do something like below, so that we
>> can keep wifi functional (ignoring bt since that was never functional)?
>> -       vqmmc-supply = <&w3cbw003c_bt_nreset>;
>> -       vmmc_aux-supply = <&w3cbw003c_wifi_nreset>;
>> +       vqmmc-supply = <&w3cbw003c_wifi_nreset>;
> 
> Optional sdio pins 4 - 8 need separate power and that's
> why we have the vmmc_aux supply to start with. But here
> there are only four pins, so my guess is that only one
> regulator is needed and the the other two are gpio pins
> for wland and bt that nowadays should be handled by the
> mmc power sequence driver.

Yeah. Just now checked the pwrseq driver (pwrseq_simple), but it has it's own
binding. So even if we move to the pwrseq, we still have to support
"vqmmc-supply = <&w3cbw003c_wifi_nreset>" to avoid breaking old dt compatibiltiy.

Thanks
Kishon