I’m currently working on a project involving a custom SBC based off the Pocket Beagle. I’m in the process of building a image for it and I am trying to utilize a blank board ID mentioned here: https://octavosystems.com/app_notes/osd335x-eeprom-during-boot/. However upon testing, I found out that the image for some reason does not boot on the original Pocket Beagle. Why is this the case? As far as I know, utilizing a blank board ID is not the recommended way to boot however I don’t see why it shouldn’t boot with the original Pocket Beagle. Thanks.
It should work on an original pocket beagle. Like I told Sam Park in a separate thread, you basically must have a serial debug adapter connected to the device to debug u-boot issues.
I can say that I have followed rcn’s instructions for building and installing u-boot and followed octavo’s directions for modifying a board with a blank eeprom, and things worked perfectly, for me.
Were you using u-boot version 2018.03rc1 when you made it work? That is the only thing that I had to change when following pocket beagle image tutorial since the 0002-NFM-Production-eeprom-assume-device-is-BeagleBone-Bl.patch provided in the blank board ID method is written for that version. In case I made a mistake somewhere I will go back and try again.
I’m off work today so I can’t say exactly what versions I was using, but I recall following the instructions verbatim and everything worked perfectly.
I suppose if your modifications to the u-boot source were bad enough, you could cause boot to fail. If I recall correctly, I started by allowing the blank eeprom case to follow the black boot path. That seemed easiest to get started and I think it just added an OR to an if statement.
Figuring out what’s happening without a serial debug port is feeling around blind. If you’re not getting paid for it, it’s not so bad, but it wastes a lot of time. No way could I do that at work.