Ars Technica article

[Grzegorz Junka]

On 13/04/2020 10:34, Niclas Zeising wrote:
> On 2020-04-13 12:27, Grzegorz Junka wrote:
>>
>>> Oh, I'm not blaming him at all.  And more to it, while Andrea Venturoli
>>> is currently helps to debug xf86-video-ati-legacy issues, I'm not that
>>> concerned about that driver, but rather that all other DRM ports except
>>> legacy locking up my laptop upon "kldload radeonkms".  This isn't 
>>> right;
>>> it should be debugged and fixed, and only then drm-legacy-kmod port can
>>> be removed.
>>>
>>
>> For what's is worth, since I moved from 12.0 to 12.1 none of the new 
>> DRM ports worked for me. The only working configuration was 
>> kernel/world from CURRENT (13) and ports compiled at least on 
>> 12.1-RELEASE. Are you trying binary distributions or compiling on 
>> your own?
>>
>
> When FreeBSD 12.1 was first released, there were issues with the 
> binary package of drm-fbsd12.0-kmod, since all packages still were 
> built on 12.0, and a kmod built on 12.0 didn't work on 12.1. Since 
> then, the package build cluster has changed to build all packages for 
> 12 on 12.1, so if you're running 12.1 the binary package should work.  
> If you are running 12-stable you need to compile the driver though.


They should but they don't. However, I didn't try the official packages. 
I am building my own with poudriere. drm-fbsd12.0-kmod compiled with 
poudriere on 12.1-RELEASE-p3 and installed on the same system (tried 
both, the official kernel/base and one compiled on my own) - didn't 
work, and same for 12-STABLE. However, Hans Petter Selasky sent me a 
13-CURRENT kernel/world to try and the drm-fbsd12.0-kmod I compiled on 
12-STABLE worked with that without issues.

Either the port is very fragile and I did something incorrectly or the 
port became dependent on some changes in the latest kernel ABI, at least 
for amdgpu (vega10 firmware modules in my case).

GrzegorzJ