Re: [PATCH 1/5] ARM: bcm2835: Define standard pinctrl groups in the gpio node.

[Stephen Warren <swarren@wwwdotorg.org>]

On 03/03/2016 03:23 PM, Eric Anholt wrote:
> Stephen Warren <swarren@wwwdotorg.org> writes:
>
>> On 02/26/2016 11:19 AM, Eric Anholt wrote:
>>> The BCM2835-ARM-Peripherals.pdf documentation specifies what the
>>> function selects do for the pins, and there are a bunch of obvious
>>> groupings to be made.  With these created, we'll be able to replace
>>> bcm2835-rpi.dtsi's main "set all of these pins to alt0" with
>>> references to specific groups we want enabled.
>>
>>> diff --git a/arch/arm/boot/dts/bcm283x.dtsi b/arch/arm/boot/dts/bcm283x.dtsi
>>
>>> +			spi0_gpio7: spi0_gpio7 {
>>> +				brcm,pins = <7 8 9 10 11>;
>>> +				brcm,function = <BCM2835_FSEL_ALT0>;
>>> +			};
>>
>> This is too many pins.
>>
>> - It includes both MOSI and MISO, although a particular use-case may
>> only use 1 of those.
>>
>> - It includes both chip-select signals, whereas a particular use-case
>> may use 0, 1, or 2 of those. This is especially true since IIRC the
>> mainline bcm283x SPI driver wants to only use GPIOs for chip-selects,
>> not SPI-controller-generated chip-select signals, to avoid some issues
>> with the HW generation of these signals.
>>
>>
>> I believe a similar comment applies to other SPI nodes too.
>>
>>> +			pcm_gpio18: pcm_gpio18 {
>>> +				brcm,pins = <18 19 20 21>;
>>> +				brcm,function = <BCM2835_FSEL_ALT0>;
>>> +			};
>>
>> Here too, I wonder if some people might want only one of DIN/DOUT and
>> not both?
>>
>>> +			uart1_gpio36: uart1_gpio36 {
>>> +				brcm,pins = <36 37 38 39>;
>>> +				brcm,function = <BCM2835_FSEL_ALT2>;
>>> +			};
>>
>> Similarly, I think for UARTS, TX/RX and RTS/CTS should always be in
>> different nodes so people can choose 2- or 4-wire mode. Most of the UART
>> nodes are already split like this, but this one isn't.
>>
>>> +			emmc_gpio22: emmc_gpio22 {
>>> +				brcm,pins = <22 23 24 25 26 27>;
>>> +				brcm,function = <BCM2835_FSEL_ALT3>;
>>> +			};
>>
>> 1-wire (1 data wire, plus CLK/CMD) eMMC is possible in theory, although
>> I don't know whether it makes sense to support this?
>
> Nothing here precludes making alternative pin groups for special
> situations like you're bringing up here.  I'm just trying to bring
> sanity to the giant lists of pins we have currently, that happen to
> correspond to these.
>
> Of your suggestions, making uart1_gpio36 split out cts/rts like the rest
> makes a lot of sense to me.  Of the others, they seem like speculation
> more than "we should fix this because it's not what people want."  Can
> you provide specific feedback of what you'd like changed to get an Ack?

All of the points I raised should be fixed. I don't believe any of the 
groups that affect more than minimal sets of pins are useful. Indeed, 
using groups at all is rather tenuous; it'd be far better to list the 
precise sets of pins only as and when they're used.