search

pcengines / apu2-documentation

Documentation and scripts for building and adjusting PC Engines APU2 firmware
https://pcengines.github.io/apu2-documentation/
208 stars 45 forks source link

Update APU2C BIOS #258

Closed KyleSanderson closed 2 years ago

KyleSanderson commented 7 years ago

Hello,

I'm still seeing duplicate printouts over the RS232 connection and it's rather painful to do anything with ncurses... Any tips as to when we can see a sync for the mainline bios shipped by PCEngines? This bug was fixed quite some time ago as per the git tree.

pietrushnic commented 7 years ago

@KyleSanderson thanks for contributing to PC Engines repository.

We are in process of mainlining code to coreboot. Please track this gerrit item to get more information. Still a lot of testing is needed, so if you can help any feedback is valuable.

Other issues are SeaBIOS and sortbootorder itself.

In case of double prints I don't see those for long time. Please describe your configuration in more detail, especially:

We can try reproduce issue and help you improve situation.

KyleSanderson commented 7 years ago

@pietrushnic

Thanks for following up. The binaries are from PCEngines.ch, latest being from March 11th 2016 just adding iPXE support. It's been painful, as even TinyCore Linux (shell) doesn't work with these bins, however I've gotten IPFire (bad), Gentoo (took a bit), and more recently PFSense going on my APU2C4.

Any updates would be appreciated, I have the API module however I'm still terrified of bricking the thing.

pietrushnic commented 7 years ago

@KyleSanderson I just posted information about status here. Unfortunately support in OS is out of scope - we can try to help but mostly you are on your own. I will think about posting installation procedures for various OSes in documenation, but there are higher priority issues now.

Binary from PC Engines site is not recent one but use stable code that works and I agree that trying custom built binaries requires recovery procedure and development knowledge. I hope recent code will mature fast, but as I wrote testing is needed to make it stable.

KyleSanderson commented 7 years ago

I know. The OS references were it's difficult to install an OS if the RS232 connection is near completely wrecked. Is there no later version that could be pinned and shipped with at least the console connection stable prior to mainlining the code in coreboot?

pietrushnic commented 7 years ago

@KyleSanderson I have no control on PC Engines releases. There were beta versions but I assume those were with quality no good enough to release. I will ask and see what we can do with this problem. Best would be to reproduce that on my side and fix. Can you provide instructions how to reproduce your issue ?

pietrushnic commented 7 years ago

@KyleSanderson please post serial console output of what you see, it may be that you just see OS output incorrect and firmware output correct. In that case you have to attach OS people not firmware part.

KyleSanderson commented 7 years ago

The device is in production now as my gateway. I can't source out the git tree anymore however there was a one-line change apparently fixing the double printout (pretty sure I found it on the board). If it doesn't ring a bell that's totally fine. I'll watch the gerrit and hope PCEngines decides to one day pin a new stable version :-).

pietrushnic commented 7 years ago

@KyleSanderson sorry for delayed reply. The key problem with seeing output twice is sgabios, which we need because lack of VGA output. Sgabios redirects vga output to console. If given distro have serial enabled you will see logs printed twice. Not much we can do with that. The solution is to disable console logging in installer if it already publish to VGA, if not then serial should look fine.

Let me know, if there is anything I can help.

mkopec commented 2 years ago

Looks like the issue was resolved in https://pcengines.github.io/#mr-4

updated SeaBIOS 1.11.0.1 (removed sgabios)

Closing this issue.