So we still have got an apparent contradiction between the Pi4 specs for UART2 with ALT4 and the statement on the schematic: There's no contradiction - the GPIO capabilities are a feature of the 2711 SoC, while the pin usage is a feature of the Pi 4 board. The 2711 has potentially wider application...
I'm not aware of a kernel feature like that - it sounds more like a screen saver configuration option, i.e. up in userland at the desktop or application level.
After reading around a bit, I think this might work (but I've not tried): [Unit] Description=This service sets the indicator LED into the heartbeat mode on shutdown. DefaultDependencies=no After=final.target [Service] Type=oneshot ExecStart=/webroot/gStation/sh/heartbeat_indicator.sh [Install] Wante...
Have a look in /etc/systemd/logind.conf for HandlePowerKey. The default value (if nothing is specified) is "poweroff", but it might have been changed by something. This page explains the options: https://www.freedesktop.org/software/sy ... .conf.html
You almost got it. What you didn't figure out is that the "leds" node replaces your rp41_pins node, i.e.: // Overlay for rp41 gpio line names /dts-v1/; /plugin/; /{ compatible = "brcm,bcm2835"; fragment@0 { target = <&leds>; __overlay__ { pinctrl-names = "default"; pinctrl-0 = <&rp41_gpio>; }; }; fr...
We need an active driver to hang pinctrl requests on, preferably one which doesn't already use pinctrl. In the absence of a driver specific to your application, the gpio-leds driver is a suitable candidate (but be aware that other overlays may try the same trick). The vga666 overlay is an example of...
Your new DT node (rp41_pins) has no compatible string - without one it will be ignored. pinctrl claims and configures pins on behalf of a driver, but releases them if the driver doesn't initialise (probe) correctly.
One other thing I just noticed: something does happen when I assert GPIO17 - the red LED goes off. But that's it. That could be an indication that the voltage is dropping too low, briefly. Do you have a too-small resistor in there somewhere? I don't have a Pi to hand, so this is a bit vague, but th...
Re: the default pull, the overlay overrides the default behaviour so it pulls up (which wouldn't work well if active_low=0, but it isn't so you are ok). Does xev show any activity when you ground the pin? Is the gpio-key module loaded? Does gpioinfo show the GPIO as being claimed? (You may need to "...
Another factor to consider is that gpio 17 pulls low by default. Run "raspi-gpio get 17" to confirm that the input is low all the time, which will cause the gpio-key driver to ignore that input as a stuck key.
raspi-gpio gives a more accurate picture of the current GPIO state. gpioinfo shows what Linux thinks it is set to, but since it doesn't read back the start-of-day state anything changed via dt-blob or config.txt won't show up.
I was about to write that nomdediot is probably a new user, but I see he's been around longer than you, Dougie, albeit with slightly fewer posts, so who knows. Although not an employee of Raspberry Pi themself ("they" as a singular pronoun was the Merriam Webster word of 2019), aBUGSworstnightmare (...
Have you read the CM schematics? They're as official as they come, and they show that both CSI and DSI have dedicated pins on the SoC, hence no chance to reuse them for something else.
Apart from a missing semicolon after the "line-name" property, your overlay looks good. The problem is caused by the sequence of operations in the BCM2835's combined pinctrl and gpio driver. Currently the driver creates and registers the GPIO half of the driver before the pinctrl side, but this fail...
On all Pis, UART0 is a PL011 that appear to Linux as ttyAMA<n>, and UART1 is an 8250 clone that appears as ttyS0. On a Pi4, UART2-5 are additional PL011s that also appear as ttyAMA<n>. The number starts at 0 for the first enabled PL011 and counts up through all the enabled interfaces. The numbering ...