GPIO pins not responding in code

[zerrofour4]

Hello.

I’ve been stuck all day trying to get any of the GPIO pins to change state from a small python script.

I followed the example in the wiki, and the IRQTest.py just said false with every loop (ran as root).

I used this library https://github.com/vsergeev/python-periphery that is also in the IRQtest.py with this code:

import time
from periphery import GPIO

gpio = GPIO("/dev/gpiochip0", 40, "out")
gpio.write(False)
time.sleep(1)
gpio.write(True)
time.sleep(2)

gpio.close()

Basically, I am trying to activate a relay.

The pin stays at 1.38 volts while the script runs (as root). I’ve tried /dev/gpiochip1 as well, no change.

I can change the pin state in the bios, and when I set it to low, the pin reads at .03 volts. So I think the GPIO pins themselves are fine. I am stumped as to why doing anything from linux just doesn’t work.

I tried Debian 10 and Fedora 32

Thanks

[zerrofour4]

Some more digging on my end…
I am trying to use libgpiod, which accesses the gpio pins through a character device. The sysfs method is deprecated and removed in newer kernels, so everybody is going to have to adjust anyway.

See https://git.kernel.org/pub/scm/libs/libgpiod/libgpiod.git/about/

So, what I am now essentially asking is…What is the correlation between the physical pin number and the line number? 4 gpiochips show up, and 0 and 1 both have 80 lines.

# gpiodetect
gpiochip0 [INT3453:00] (80 lines)
gpiochip1 [INT3453:01] (80 lines)
gpiochip2 [INT3453:02] (20 lines)
gpiochip3 [INT3453:03] (35 lines)

Which gpiochip is the 40 pin setup on the board?

gpioinfo
gpiochip0 - 80 lines:
	line   0:      unnamed       unused   input  active-high 
	line   1:      unnamed       unused   input  active-high 
	line   2:      unnamed       unused   input  active-high 
	line   3:      unnamed       unused   input  active-high 
	line   4:      unnamed       unused   input  active-high 
	line   5:      unnamed       unused   input  active-high 
	line   6:      unnamed       unused   input  active-high 
	line   7:      unnamed       unused   input  active-high 
	line   8:      unnamed       unused   input  active-high 
	line   9:      unnamed       unused   input  active-high 
	line  10:      unnamed       unused  output  active-high 
	line  11:      unnamed       unused  output  active-high 
	line  12:      unnamed       unused  output  active-high 
	line  13:      unnamed       unused   input  active-high 

From other pages I’ve seed, the broadcom chips have labels. And there is no correlation between pin number and line number. I’ve already caused a crash and reboot by setting line 47 to high.

[zerrofour4]

I did some trial and error and nailed down the chip and line number of a few pins using a relay and patience.

I then stared at the numbers in the bios for the pins and realized there is a pattern.

Here are my results, most of them checked by hand using the gpioset utility.

bios_number	physical_pin	chip_number	line_number
60	10	0	60
61	8	0	61
79	23	1	3
80	24	1	4
81	26	1	5
82	21	1	6
83	19	1	7
88	11	1	12
110	3	1	34
111	5	1	35
112	27	1	36
113	28	1	37
114	22	1	38
115	32	1	39
134	36	1	58
136	13	1	60
137	15	1	61
139	29	1	63
140	31	1	64
141	33	1	65
143	37	1	67
145	16	1	69
146	18	1	70
161	7	2	5
162	12	2	6
163	35	2	7
164	38	2	8
165	40	2	9

[zerrofour4]

@zerrofour4

import time
from periphery import GPIO

gpio = GPIO("/dev/gpiochip2", 9, "out")
gpio.write(False)
time.sleep(1)
gpio.write(True)
time.sleep(2)

gpio.close()

Now, going back to my original snippet, I see that physical pin 40 (off in the far corner ) is on /dev/gpiochip2 on line 9. And now it works as expected!

If at somepoint Ubuntu or Debian finally drops GPIO sysfs, I hope someone googling around sees this post. Or if you are using Fedora IoT right now.

I’ll probably make a fork of python-periphery to integrate this mapping more easily, so you only have to worry about the physical pin layout number.

[ansonhe97]

Thanks @zerrofour4 for the good work!

This is very useful for other users to see.

[James_Hawthorne]

Thanks so much for this @zerrofour4, saved me a lot of time and frustration

[blazini36]

Thank you for the fine detective work. I just started looking into the GPIO thing on the Odyssey definitely saved me some work by posting your findings

[Odess]

GPIO sysfs stil exist, just the numbers changed, just add 512 to every number.

[Lode]

Confirmed that GPIO sysfs works with 512 added to every number.
e.g.
pin on connect 28 → BCM1_ID_SCL which is 389 in your program, now works as 901.
see seeed odyssey pin overview for the other numbers

#define IO_PIN_ENABLE_3point3V 901 //389 // connector pin 28 for EN258 pcb

[PJ_Glasso]

Hi there,
I think you need some delay’s in there.

gpio = GPIO("/dev/gpiochip0", 40, "out")
Delay (500);
gpio.write(False)
Delay (500);
time.sleep(1)

HTH
GL PJ