Single board computer GPIO HAL?

I saw that too. I think a LinuxCNC hal driver needs to access the registers directly, going through the Kernel driver might not be fast enough. Maybe someone can chime in.

I have no idea what the hell I'm doing tho. Once I get enough info I'll figure out who I can bribe to hook me up with a LinuxCNC driver.

Before now I hadn't powered up the board, I was mostly going by what I've seen on the Seeed forum and such. I got Manjaro installed, apparently the /sys/class/gpio thing is the old Linux way of accessing GPIO so it doesn't just work in Arch based distros. I installed libgpiod and it's dependencies. There are tools to detect the GPIO but it just shows some obscure info relating to 4 "gpiochip"s. I came across this thread on the Seeed forum where someone Traced out all of the GPIO physical pins and mapped them out.
forum.seeedstudio.com/t/gpio-pins-not-re...ing-in-code/252187/3

I have no idea if libgpiod respects RT-preempt or how any of that works

here is another manufacturer for SBC boards

www.impulse-embedded.co.uk/systems/singl...rd-computer-kits.htm

they do quite a few options if you browse the site, the costs of what I was looking at 12-18m ago was not too bad.

it's just another option.....

blazini36 wrote:

Before now I hadn't powered up the board, I was mostly going by what I've seen on the Seeed forum and such. I got Manjaro installed, apparently the /sys/class/gpio thing is the old Linux way of accessing GPIO so it doesn't just work in Arch based distros. I installed libgpiod and it's dependencies. There are tools to detect the GPIO but it just shows some obscure info relating to 4 "gpiochip"s. I came across this thread on the Seeed forum where someone Traced out all of the GPIO physical pins and mapped them out.
forum.seeedstudio.com/t/gpio-pins-not-re...ing-in-code/252187/3

I have no idea if libgpiod respects RT-preempt or how any of that works

The last time I played with ports at the hardware level it was in DOS before Linuxcnc and Mach1 existed.
My understanding is that hardware ports were mapped to a memory address so if you wrote data to that address, it was transmitted to the hardware.

But this is low level stuff so the kernel has to be involved.

Given the kernel includes some procedures and even header files to support GPIO ports, I would use them. Heck, its in the kernel so if you patch the kernel to support  PREEMPT_RT, surely it will work!