PCF files do not match the actual pinout for GPIOs

[jeremyherbert]

It appears that the GPIO pins in the pcf file here: https://github.com/gregdavill/OrangeCrab-examples/blob/main/verilog/orangecrab_r0.2.pcf are not correct.

They do not match the pinout here: https://gregdavill.github.io/OrangeCrab/r0.2/docs/pinout/ and at least from my quick look at the schematic, the picture is the correct reference.

[jeremyherbert]

I had a closer look at this, it seems that the pinout numbers on the silkscreen do not match the GPIO numbering; is this intended?

[umarcor]

@jeremyherbert, my perception is that the constraints file is correct, in the sense that it responds to some criteria; however, the criteria is not coherent with the silkscreen. I would propose the following modifications to the names in the constraints file:

constraint	silkscreen	pin	proposal
GPIO0[0]	0	N17	GPIO_0
GPIO0[1]	1	M18	GPIO_1
GPIO0[2]	5	B10	GPIO_5
GPIO0[3]	6	B9	GPIO_6
GPIO0[4]	9	C8	GPIO_9
GPIO0[5]	10	B8	GPIO_10
GPIO0[6]	11	A8	GPIO_11
GPIO0[7]	12	H2	GPIO_12
GPIO0[8]	13	J2	GPIO_13
GPIO0[9]	A0	L4	GPIO_A0
GPIO0[10]	A1	N3	GPIO_A1
GPIO0[11]	A2	N4	GPIO_A2
GPIO0[12]	A3	H4	GPIO_A3

[gregdavill]

The pinouts on the silkscreen are following the standard 'feather' numbering, which I think has some historical reasons for the numbering.

As @umarcor has pointed out, the GPIO numbers line up, but they have gaps.
I think we should update the pcf to cover the numbers actually present on the silkscreen.

[jeremyherbert]

Thanks!