amdgpu panics

[Grzegorz Junka]

On 04/04/2020 10:27, Matthias Andree wrote:
> Thank you John for the comprehensive explanation. It took me a while to
> go through all the details, then again to recompile the ports and try to
> reinstall all packages.
>> What i discovered in the meantime is that it's not an isolated problem:
>>
>>
>> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=241787
>>
>> https://forums.freebsd.org/threads/upgrading-to-freebsd-12-1-release-resolving-an-issue-with-drm-fbsd12-0-kmod.72895/
>>
>>
>>
>> On my system I indeed had the jail at a different patch level than the
>> host system, although they were all running 12.1-RELEASE. I updated
>> the host and the jail to 12.1-RELEASE-p3. Poudriere noticed the
>> updated jail and deleted and recompiled all 2000+ packages. Then I
>> upgraded the system on which I wanted to install the packages to
>> 12.1-RELEASE-p3 too. Then I deleted drm-fbsd12.0-kmod and installed
>> drm-kmod. It reinstalled drm-fbsd12.0-kmod.
>>
>> The result? Blank screen!!!
>>
>> I start as single or normal user then do:
>>
>> kldload amdgpu
>>
>> I see the driver is loading various graphics kernel modules then the
>> screen goes blank and the whole system hangs. No panic is shown, no
>> restart, just hungs. Any SSH sessions to the system become stale. Only
>> hard reset is able to restart it.
>>
>> This is really frustrating and a really bad user experience. I
>> wouldn't be surprised if the remained desktop users moved to Linux or
>> other FreeBSD forks if they haven't already.
>>
>> The only option left I see is to also compile the kernel myself from
>> sources.
> Compared to 2,000 packages that seams a reasonable approach, and then IN
> THAT SAME LIVE SYSTEM also rebuild the graphics modules.
>
> I understand that the poudriere/pkg proponents have aggressively lobbied
> users to use pkg and poudriere for clean-room builds, but I wonder if it
> isn't easier to forgo poudriere for drivers and instead:
>
> obtain/update to 12.1-RELEASE-p3 sources in /usr/src (with svn, for
> instance)
> make buildworld buildkernel
> make installkernel
> edit your loader.conf[.local] so it doesn't load b0rked graphics modules,
> reboot into single-user
> mergemaster -Fp
> make installworld
> mergemaster -Fi
> make delete-old # important - there may be 12.0 parts that need removal,
> 12.1 for instance updated LLVM
> rebuild your kmods and drivers IN THIS LIVE SYSTEM RIGHT FROM PORTS (not
> poudriere)
> install kmods and drivers
> reboot and then gradually manually load kernel drivers such as amdgpu
> one by one so you know which work (enable them in the loader) and which
> won't.
>
> I am not sure if it helps for amdgpu, since I am using nvidia- which
> sort-of works (but GNOME frequently flakes out for my user but not other
> users)... but I'd think this approach forgoes any potential difference
> between the build jail and live system kernel sources
>
> Of course this rules out freebsd-update for kernel/system patching then,
> you'd update /usr/src and then make -DNOCLEAN buildworld buildkernel and
> install again once -p4 or newer come out.
>

I reinstalled the whole system from a newly compiled kernel and world. 
It didn't help. When I do "kldload amdgpu" the screen goes blank after 
loading one of the driver modules.

I have compiled the base on another system and firstly just unpacked the 
kernel files. When that didn't work I used FreeBSD-base packages to 
reinstall everything from the build server. It worked pretty well but 
didn't change a thing.

The build system has an NVidia card and an AMD (Phenom) process. The 
system on which I install has AMD Vega 64 card and another AMD (Ryzen) 
processor. I don't use any configuration to build for a specific 
architecture and I hope "drm-fbsd12.0-kmod" doesn't do any optimizations 
based on the architecture on which it's compiled.

How can I debug what's wrong?

Grzegorz J