Can confirm. 

But I'm a bit worried - there is a part before the "core_link_enable_stream"
that seems wrong to me in initializing the device, which seems to occure in all
dmesg outputs.

I've attached line numbers via less -N - maybe it helps.

837 [    8.686299] [drm] BIOS signature incorrect 0 0
838 [    8.686326] ATOM BIOS: 113-RAVEN-110
839 [    8.686357] [drm] vm size is 262144 GB, 4 levels, block size is 9-bit,
fragment size is 9-bit
...
854 [    8.688113] [drm] Found VCN firmware Version: 1.73 Family ID: 18
...
 857 [    8.797714] [drm] DM_PPLIB: values for Invalid clock
    858 [    8.797715] [drm] DM_PPLIB:   0 in kHz
    859 [    8.797715] [drm] DM_PPLIB:   400000 in kHz
    860 [    8.797716] [drm] DM_PPLIB:   933000 in kHz
    861 [    8.797716] [drm] DM_PPLIB:   1067000 in kHz
    862 [    8.797809] WARNING: CPU: 6 PID: 725 at
drivers/gpu/drm/amd/amdgpu/../display/dc/calcs/dcn_calcs.c:1380
dcn_bw_update_from_pplib+0x16b/0x280 [amdgpu]

This seems plain wrong to me - shouldn't it mention here the engine clock?

This WARNING stems from not getting the clock values right - seems intended.

The same warning occurs again, I wildly guess for the memory (?!).

Line 862-1017 - only afterwards the mentioned "core_link_enable_stream" bug
occurs. It seems to me - as a total layman - that the initialization fails and
afterwards bogus happens.