Stuck in a reboot loop when CM4 is installed on the official IO board running Retropie v4.7.1.

[gregmoudy]

I've seen this happen twice now with two different CM4 modules. I'm working with Retropie v.4.7.1 on a CM4001000 and a CM4102032.

CM4001000
While on the official IO board, copied and installed the patch, installed the safe shutdown script, installed the CM4 into the GPiCase2 and everything worked great. Removed the CM4 from the case, removed the SD card and mounted it to turn off the patch to switch it back to HDMI output, put it back on the official IO board, started the CM4 and then was stuck in an endless reboot loop. Mounted the SD card on another machine and examined the boot partition and saw the config.txt was missing.

CM4102032
While on the official IO board, copied the patch over but did not run the install batch file yet as I was still working on some stuff connected to a TV, installed the safe shutdown script, did a reboot and got stuck in an endless reboot loop before I could put it in the GPiCase2. Was able to put the the CM4 in usb boot mode, looked at the boot partition and again the config.txt was missing. Restored the config.txt from a backup, ran the install batch for the patch, put it in the GPiCase2 and it is working again.

Am I doing something wrong here or out of order? I've followed the steps outlined in README.md:

"Make sure that the GPiCase2 patch is installed.
If the patch file is not installed and you install the script first, you will lose the “config.txt” file, which will cause the CM4 to fail to boot."

Should I be able to move the CM4 back and forth between the 2 devices as long as I'm installing and uninstalling the patch which sets up the display output? I'm not blocked by this but I'm basically in a spot where I can't throw the CM4 back onto the official IO board without this happening. Luckily I don't plan on doing this now in the future since I'm using a CM4102032 with wifi, but anyone with a non wifi compute module 4 I worry will run into this when they need to go back to an IO board to do some maintenance or file management.

I haven't had a chance to dig into it and try to debug it but clearly something about this safe shutdown script or the way I'm using it and booting up on an IO board is causing something to fail and wipe out the config.txt after having it happen twice on 2 different modules.

I'd be interested to hear if anyone else has experienced this. I'm willing to provide any additional info needed and I still have the CM4001000 and SD setup in a way were I can reproduce this again fairly easily.

[EctoOne]

I've ordered the GPI Case 2 a few days ago and prepped an SD card on a normal Pi4. And after installing the Safe Shutdown Script, I also was stuck in a boot loop.

I'm not that familiar with Linux/Coding but I assume that the script needs to have a way to know if it is docked or not and it probably does that by checking if there is anything connected. And since this is probably not the case for a normal Pi4 or the IO board, it just fails to set the correct config.

My Case hasn't arrived yet but I'm a little bit disappointed that I probably can't use the same card in both devices (especially since I couldn't get the version with the dock). I actually wanted to disable Bluetooth and Wifi on the GPI Case to save power. And just move the card to a Pi4 if I want to update or whatever. Well, I guess I still can do that after I've set up WiFi. It's still not great. Or I might just do a clean installation and don't install the save shutdown script.