Title: REG_WAIT timeout in optc3x_disable_crtc at first modeset on both
Raphael iGPU and RX 7600 (MSI PRO B650-P WIFI, HDMI) - display stays
black

Summary
With nomodeset removed, amdgpu initializes both GPUs but logs a REG_WAIT 
timeout in optc*_disable_crtc, and the display stays black while the system 
keeps running (commands run over SSH during the black screen). The same failure 
occurs on two different GPUs on the same board and monitor: the Ryzen's Raphael 
iGPU and a Navi 33 card. With nomodeset the system is stable on the firmware 
framebuffer.

Hardware
- Board: MSI PRO B650-P WIFI (MS-7D78), BIOS E7D78AMS.1I0 (03/12/2025)
- CPU: AMD Ryzen 9 7900X (Raphael iGPU, 0000:12:00.0)
- GPU: ASRock Radeon RX 7600 Challenger OC (Navi 33, 0000:03:00.0)
- Display: Viotek 2560x1440 monitor on HDMI only (EDID vendor VTK, product 
12800)
- BIOS: Re-Size BAR enabled, Above 4G enabled, SR-IOV disabled, Integrated 
Graphics UMA Auto, HybridGraphics disabled, Secure Boot enabled, Initiate 
Graphic Adapter [confirm: PEG or IGD for the iGPU boot]

Software
- Ubuntu 24.04.5 LTS, kernel 6.17.0-40-generic (HWE), X11 session (Wayland 
disabled in GDM)
- amdgpu-dkms 6.16.13 (out-of-tree, from AMD's amdgpu-install, 30.30.1)
- /etc/modprobe.d/amdgpu.conf: options amdgpu sg_display=0 gfx_off=0 runpm=0 
(gfx_off is reported as an unknown parameter and ignored)

Reproduction
1. Boot with kernel command line: ro quiet splash (no nomodeset).
2. amdgpu loads; VBIOS, PSP and DMUB firmware initialize without errors.
3. On the first modeset the timeout below is logged and the display goes black.

Logs, iGPU (monitor on the motherboard HDMI)
```
amdgpu 0000:12:00.0: amdgpu: [drm] DMUB hardware initialized: version=0x05002000
amdgpu 0000:12:00.0: amdgpu: runtime pm is manually disabled
amdgpu 0000:12:00.0: [drm] fb0: amdgpudrmfb frame buffer device
amdgpu 0000:12:00.0: amdgpu: [drm] REG_WAIT timeout 1us * 100000 tries - 
optc31_disable_crtc line:145
```
No further messages from 0000:12:00.0 follow.

Logs, RX 7600 (monitor on the card's HDMI)
```
amdgpu 0000:03:00.0: amdgpu: [drm] Loading DMUB firmware via PSP: 
version=0x07002D00
amdgpu 0000:03:00.0: amdgpu: [drm] REG_WAIT timeout 1us * 150000 tries - 
optc32_disable_crtc line:195   (logged twice)
```

Tried with no change
- amdgpu.aspm=0, runpm=0 and sg_display=0
- dc=0 and dpm=0 (in an earlier attempt)
- Ubuntu 26.04.1 live image (kernel 7.0, Mesa 26.0.x, in-tree amdgpu), RX 7600 
on HDMI: no desktop (cursor only by default; solid white screen with 
amdgpu.aspm=0 amdgpu.runpm=0). I captured no logs from those boots.

Other notes
- ASRock support told me they reproduced a similar failure on an MSI PRO 
B650M-A WIFI with a Ryzen 7 7800X3D, but not on their own Z790 Taichi, B650E 
Taichi or A620M-HDV/M.2 boards.
- Both GPUs fail at the same step on the same board and monitor. I have no 
second display to test, so I can't rule out the monitor or HDMI link.
- Launchpad bug (Ubuntu kernel): 
https://bugs.launchpad.net/ubuntu/+source/linux-hwe-6.17/+bug/2150485

Questions
- Is this a known issue with optc31/optc32 disable_crtc timeouts at first 
modeset?
- Which debug options (drm.debug bits, dcdebugmask) would be most useful? I can 
test patches or parameters and capture logs over SSH.

Full kernel logs (journalctl -b -k) of both failing boots are attached.

** Attachment added: "amdgpu-logs.txt"
   
https://bugs.launchpad.net/ubuntu/+source/linux-hwe-6.17/+bug/2150485/+attachment/6009485/+files/amdgpu-logs.txt

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2150485

Title:
  amdgpu: REG_WAIT timeout and display blackout on RX 7600 (Navi 33)
  with kernel 6.17

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux-hwe-6.17/+bug/2150485/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to