Bug 1180442 - amdgpu not detected on boot with eDP connector
amdgpu not detected on boot with eDP connector
Status: RESOLVED FIXED
Classification: openSUSE
Product: openSUSE Tumbleweed
Classification: openSUSE
Component: Kernel
Current
Other Other
: P5 - None : Normal (vote)
: ---
Assigned To: openSUSE Kernel Bugs
E-mail List
https://gitlab.freedesktop.org/drm/am...
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2020-12-29 21:09 UTC by John Shand
Modified: 2021-01-12 08:34 UTC (History)
3 users (show)

See Also:
Found By: ---
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---


Attachments
Hardware Profile (10.79 KB, text/plain)
2020-12-29 21:09 UTC, John Shand
Details
dmidecode (9.41 KB, text/plain)
2021-01-01 23:09 UTC, John Shand
Details
lscpu (2.31 KB, text/plain)
2021-01-01 23:09 UTC, John Shand
Details
lspci (2.54 KB, text/plain)
2021-01-01 23:10 UTC, John Shand
Details
hardware info detailed (865.17 KB, text/plain)
2021-01-04 09:03 UTC, John Shand
Details
inxi-amdgpu-working-hardware (10.79 KB, text/plain)
2021-01-12 07:55 UTC, John Shand
Details

Note You need to log in before you can comment on or make changes to this bug.
Description John Shand 2020-12-29 21:09:17 UTC
Created attachment 844743 [details]
Hardware Profile

Hi there, just tried kernel version 5.10.3 and unfortunately I still need to specify nomodeset in the grub command line, not graphics driver listed.  must be llvmpipe instead.
Comment 1 John Shand 2020-12-29 21:14:41 UTC
upon boot i get these errors after I put in nomodeset:

kernel: [drm:amdgpu_init [amdgpu]] *ERROR* VGACON disables amdgpu kernel modesetting.
kernel: [drm:amdgpu_init [amdgpu]] *ERROR* VGACON disables amdgpu kernel modesetting.
Comment 2 Takashi Iwai 2020-12-30 07:43:05 UTC
Could you report it to the upstream bug tracker, e.g. gitlab.freedesktop.org Issues?

The error messages with nomodeset are expected, it just indicates that the KMS isn't used.
Comment 3 John Shand 2021-01-01 23:09:00 UTC
Created attachment 844775 [details]
dmidecode
Comment 4 John Shand 2021-01-01 23:09:46 UTC
Created attachment 844776 [details]
lscpu
Comment 5 John Shand 2021-01-01 23:10:13 UTC
Created attachment 844777 [details]
lspci
Comment 6 John Shand 2021-01-01 23:27:26 UTC
link to git lab https://gitlab.freedesktop.org/drm/amd/-/issues/1413
Comment 7 John Shand 2021-01-02 07:45:15 UTC
if the needinfo flag was selected, could you please help me out by requesting what information you need?  I'm very new to this
Comment 8 Takashi Iwai 2021-01-02 07:49:39 UTC
It was meant for asking to report to the upstream tracker, i.e. the URL to the upstream tracker entry.
Comment 9 John Shand 2021-01-02 07:53:12 UTC
oh, I see.  I did it this morning sometime.  sorry for the misunderstanding
Comment 10 Takashi Iwai 2021-01-02 16:23:41 UTC
Admittedly the meaning of NEEDINFO flag is to bloat, so of course no problem you asking about it.

In anyway, let's see how the upstream devs handle this issue.  Thanks.
Comment 11 John Shand 2021-01-04 09:03:59 UTC
Created attachment 844793 [details]
hardware info detailed
Comment 12 John Shand 2021-01-05 07:44:30 UTC
i just tried booting the amdgpu driver with the new 5.10.4 kernel.  same result.  come on guys.  this is getting old
Comment 13 Takashi Iwai 2021-01-05 08:47:20 UTC
Just to make sure: could you check whether the latest kernel-vanilla package shows the same problem?  There is one amdgpu patch that was added to kernel-default for addressing a kernel Oops.  I don't think this is the cause of your problem, but we can verify that by testing kernel-vanilla.

If the problem is still seen with kernel-vanilla, and Arch Linux with the same kernel version works, it must be either the different kernel configuration and/or the user-space stuff.
Comment 14 John Shand 2021-01-05 09:52:45 UTC
I have tried the vanilla kernel and the problem is not there.
Comment 15 Takashi Iwai 2021-01-05 10:16:30 UTC
(In reply to John Shand from comment #14)
> I have tried the vanilla kernel and the problem is not there.

OK, and which kernel version was it?

I noticed that kernel-vanilla-5.10.4 also contains the same fix (that was merged in Linus tree), while kerne-vanilla-5.10.3 doesn't.
Comment 16 John Shand 2021-01-05 10:44:39 UTC
ineresting.  i am on kernel 5.10.4.  i would have thought that vanilla means no patches
Comment 17 John Shand 2021-01-05 10:45:58 UTC
i can't believe the kernel guys let that one out of the bag.  that for me is a show stopper
Comment 18 Takashi Iwai 2021-01-05 10:48:21 UTC
(In reply to John Shand from comment #17)
> i can't believe the kernel guys let that one out of the bag.

You're expecting them having always a magic crystal...

Please double-check that kernel-default-5.10.4 breaks and kernel-vanilla-5.10.4 works.  If this is confirmed in 100%, it's a good hint for tracking further.
Comment 19 John Shand 2021-01-05 20:54:22 UTC
(In reply to Takashi Iwai from comment #18)
> (In reply to John Shand from comment #17)
> > i can't believe the kernel guys let that one out of the bag.
> 
> You're expecting them having always a magic crystal...
> 
> Please double-check that kernel-default-5.10.4 breaks and
> kernel-vanilla-5.10.4 works.  If this is confirmed in 100%, it's a good hint
> for tracking further.

I was mistaken with the kernel 5.10.4 vanilla it doesn't work.  arch linux kernel 5.10.4 does work
Comment 20 John Shand 2021-01-12 07:54:55 UTC
amdgpu load issue disappeared with inxi-amdgpu-working-hardware.txt file
Comment 21 John Shand 2021-01-12 07:55:29 UTC
Created attachment 845019 [details]
inxi-amdgpu-working-hardware
Comment 22 John Shand 2021-01-12 07:56:21 UTC
i have information to give you now.  this issue is not present now.  thanks for your hard work.
Comment 23 Takashi Iwai 2021-01-12 08:34:57 UTC
Thanks for updates.  Let's close now.