Return to Level1Techs.com

2200G and 2400G non-bootable under Linux


#41

You are not wrong about it being forced but then that’s true for most things. Choice is mostly between one company that wants your money and another one that is willing to go further to get it. Companies sabotage one another and manipulate people constantly, no argument there. That’s the system we live in. On a brighter note Ubuntu can be a better replacement for Windows for basic computing. I hear only good things from the few people for whom I installed it for. In my use case it’s not usable though my issue here is much more with AMD for miss-advertisement. than with anything else.
I am comparing one OS to another, you have to compare to choose one regardless of why or how it is what it is. I am not comparing Linux development ecosystem to that of Windows. That would indeed be unfair.


#42

perhaps this is a good time to get started with ostree and atomic workstation?

https://fedoraproject.org/wiki/Workstation/AtomicWorkstation (warning work in progress but I imagine you’re not faint of heart since you switch kernels)


#43

I am not entirely sure how this would help with hardware support. Or am I am missing something. Also the system in question is a Server (with standalone GUI app) so Workstations OS might not be ideal.


#44

work station and server are the exact same. One is just tooled to have a more desktop like experience OOB while the other is not.


#45

Yes but that’s the point, server is headless, workstation comes with a desktop by default. I know I could remove the desktop but I don’t really know how that particular Fedora build would help with hardware support anyways.


#46

I decided to just use the “boot kit” legacy apu provided by AMD until the 2400G was usable under Linux. That did not work out either. With new BIOS the old apu no longer works - stuck at BIOS boot screen. Even jsut disabling SMT caused the system to have no display until bios reset. After downgrading BIOS with 2400G it is now stuck in reboot loop with either the apu or cpu and CLEAR CMOS has no effect.

There is definitely a defect in the BIOS as it doesn’t work properly even before Linux comes into play. I would certainly point the finger at Asrock at this point. Who knows maybe it’s the BIOS and not Linux that’s causing the freezes. For now I have returned new hardware and went back to the old fm2 apu which is stable. Hopefully Wendell does a review of this board at some point and the cause of the issue becomes apparent.


#47

It won’t help you but you’ll be able to “commit” your exact system configuration as far as I understand and if things work for you on one machine, they can work everywhere.

It is just my cursory understanding of the goals of the project. I don’t think we’re there yet.


#48

Yes, it does look a bit experimental, not ideal for a server. It feels great to have access to NAS files reliably again, so I will play a bit safer from now on with the server at least.


#49

I thought this was a sandbox server (:


#50

Between all the bickering, I wasn’t clear, did the original poster of the thread try the latest 4.16RC? Only those work anything approaching stable with Ryzen 2000 APUs, 4.15 has significant bugs and you can forget about older kernels. You won’t find the 4.16 kernel on any distro, it’s not released, you’ll need to grab it from git and compile it.


#51

You need to have the NEWEST kernel (4.15) to use raven ridge.


#52

I didn’t try 4.16rc before I switched back to fm2 hardware. I did read that it did not fix the issues I had though. It would make sense as there are many issues I experienced not just one. I was planning on using the boot kit apu until 4.16 was released but it got bricked so had to return it.


#53

Well, some people are managing to do OK with 4.16RC. It’d be worth a try, and certainly will have less bugs on Ryzen 2000 than 4.15 which had driver support that wasn’t ready. 4.16 won’t be released until April, by the way, I’m not sure if you want to wait that long…


#54

I would have, and planned on waiting but my hand was forced by defective BIOS bricking the motherboard. Will follow the developments and will probably order it again when things settle down.


#55

Some bugs (not yours?) are waiting to be fixed with 4.17 I just read on phroenix because the fixes are frozen out of the mainline development branch right now, that’s not good news, since 4.17 isn’t out until mid-June / July. I’m going to try it with 4.16RC though once I have the system, it’ll be a great upgrade for this AMD FX 8320e.


#56

Yes it’s a great upgrade, pretty much 50%+ performance at lower power not to mention modern video decoders. It does look like 4.16 won’t be quite stable, 4.17 should be stable though even that might no be as stable as I would like for a home server. Will have to keep an eye out for motherboard reviews to see if the issue have been fixed with BIOS.


#57

Yep. https://www.3dmark.com/compare/spy/3517777/spy/3200692


#58

Nice. Did you get a chance to test it on Linux?


#59

Sorry for the late reply. I tested it in Linux, 4.16 was unstable using the APU’s graphics… I ended up later bricking my ASRock x370 Pro4 BIOS trying to upgrade it, so I’m back to using an AMD FX until I can arrange for repairs or replacement of the motherboard. :frowning: Though at least my Nvidia cards are working better than AMD’s APU driver in Linux…


#60

How’d this happen? Corrupt BIOS or something?