CM5 uses a completely different GPIO controller, so it's not inconceivable that there may be a difference. However, the CM5 hardware implementation is more flexible so ultimately it shouldn't be a step backwards.
The pinctrl utility will give you a more complete picture of the GPIOs and what state they are in. See what you get from "pinctrl 26". You can also use "pinctrl poll 26" to watch for level changes on a pin (it doesn't use interrupts so you might miss very short pulses).
I do notice that the GT928 driver explicitly changes the interrupt GPIO to an output in some situations - reset and waking from sleep - so that makes it unusual. GPIOs enabled for interrupts will appear in /proc/interrupts, so that gives you something else to watch.
Have a play, and report back with your findings.
The pinctrl utility will give you a more complete picture of the GPIOs and what state they are in. See what you get from "pinctrl 26". You can also use "pinctrl poll 26" to watch for level changes on a pin (it doesn't use interrupts so you might miss very short pulses).
I do notice that the GT928 driver explicitly changes the interrupt GPIO to an output in some situations - reset and waking from sleep - so that makes it unusual. GPIOs enabled for interrupts will appear in /proc/interrupts, so that gives you something else to watch.
Have a play, and report back with your findings.
Statistics: Posted by PhilE — Thu May 29, 2025 8:22 am