Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 955930 - media-libs/mesa-25.1.0: Nouveau fails to start after upgrading from 25.0.5
Summary: media-libs/mesa-25.1.0: Nouveau fails to start after upgrading from 25.0.5
Status: UNCONFIRMED
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: Current packages (show other bugs)
Hardware: All Linux
: Normal normal
Assignee: Gentoo X packagers
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2025-05-13 19:37 UTC by opensauce04
Modified: 2025-05-15 20:56 UTC (History)
1 user (show)

See Also:
Package list:
Runtime testing required: ---


Attachments
`dmesg` output from a boot after installing Mesa 25.1.0 (mesa-dmesg.txt,90.23 KB, text/plain)
2025-05-13 19:37 UTC, opensauce04
Details
Output of `emerge --info mesa` (mesa-emerge-info.txt,8.74 KB, text/plain)
2025-05-13 19:38 UTC, opensauce04
Details
`eglinfo` from Mesa 25.0.5 (eglinfo-25.0.5.txt,164.98 KB, text/plain)
2025-05-15 16:13 UTC, opensauce04
Details
Verbose `eglinfo` from Mesa 25.1.0 (eglinfo-25.1.0-verbose.txt,113.33 KB, text/plain)
2025-05-15 16:13 UTC, opensauce04
Details

Note You need to log in before you can comment on or make changes to this bug.
Description opensauce04 2025-05-13 19:37:41 UTC
Created attachment 928765 [details]
`dmesg` output from a boot after installing Mesa 25.1.0

After upgrading from Mesa 25.0.5 to 25.1.0, I rebooted my computer and was met with a black screen. I attempted to reboot again, and was met with a black screen again. I managed to successfully move to a TTY by using Ctrl+Alt+F4, and checked `dmesg` to find the following message:

```
[    7.830216] nouveau 0000:06:00.0: bios: version 94.03.1b.00.4a
[    7.830383] Loading firmware: nvidia/ga103/acr/ucode_ahesasc.bin
[    7.832836] Loading firmware: nvidia/ga103/acr/ucode_asb.bin
[    7.833319] Loading firmware: nvidia/ga103/acr/ucode_unload.bin
[    7.833851] Loading firmware: nvidia/ga103/gr/NET_img.bin
[    7.835262] Loading firmware: nvidia/ga103/gr/fecs_bl.bin
[    7.835533] Loading firmware: nvidia/ga103/gr/fecs_sig.bin
[    7.835732] Loading firmware: nvidia/ga103/gr/gpccs_bl.bin
[    7.835903] Loading firmware: nvidia/ga103/gr/gpccs_sig.bin
[    7.836108] Loading firmware: nvidia/ga103/sec2/sig.bin
[    7.836542] Loading firmware: nvidia/ga103/sec2/image.bin
[    7.837237] Loading firmware: nvidia/ga103/sec2/desc.bin
[    7.837395] Loading firmware: nvidia/ga103/sec2/hs_bl_sig.bin
[    7.837708] Console: switching to colour dummy device 80x25
[    7.837766] nouveau 0000:06:00.0: vgaarb: deactivate vga console
[    7.837868] nouveau 0000:06:00.0: fb: 8192 MiB GDDR6
[    7.837872] Loading firmware: nvidia/ga103/nvdec/scrubber.bin
[    7.867042] nouveau 0000:06:00.0: drm: VRAM: 8192 MiB
[    7.867045] nouveau 0000:06:00.0: drm: GART: 536870912 MiB
[    7.867047] nouveau 0000:06:00.0: drm: BIT table 'A' not found
[    7.867048] nouveau 0000:06:00.0: drm: BIT table 'L' not found
[    7.867049] nouveau 0000:06:00.0: drm: TMDS table version 2.0
[    7.867753] nouveau 0000:06:00.0: drm: MM: using COPY for buffer copies
[    7.868867] snd_hda_intel 0000:06:00.1: bound 0000:06:00.0 (ops disp50xx_modifiers [nouveau])
[    7.870568] nouveau 0000:06:00.0: [drm] Registered 4 planes with drm panic
[    7.870571] [drm] Initialized nouveau 1.4.0 for 0000:06:00.0 on minor 0
[    7.981369] fbcon: nouveaudrmfb (fb0) is primary device
[    8.128437] nouveau 0000:06:00.0: disp: outp 04:0006:0f82: training failed
[    8.128441] nouveau 0000:06:00.0: drm: [DRM/00000004:kmsOutp] [DP_TRAIN retrain:0 mst:0 lttprs:0 post_lt_adj:0 nr:4 bw:540000] (ret:-1)
[    8.223474] nouveau 0000:06:00.0: disp: outp 04:0006:0f82: training failed
[    8.223477] nouveau 0000:06:00.0: drm: [DRM/00000004:kmsOutp] [DP_TRAIN retrain:0 mst:0 lttprs:0 post_lt_adj:0 nr:4 bw:270000] (ret:-1)
[    8.318419] nouveau 0000:06:00.0: disp: outp 04:0006:0f82: training failed
[    8.318422] nouveau 0000:06:00.0: drm: [DRM/00000004:kmsOutp] [DP_TRAIN retrain:0 mst:0 lttprs:0 post_lt_adj:0 nr:4 bw:162000] (ret:-1)
[    8.441046] Console: switching to colour frame buffer device 240x67
[    8.442136] nouveau 0000:06:00.0: [drm] fb0: nouveaudrmfb frame buffer device
```

After downgrading to Mesa 25.0.5, I rebooted and managed to enter my desktop environment without any issue.

I have attached a full dmesg output as well as emerge information.
Comment 1 opensauce04 2025-05-13 19:38:21 UTC
Created attachment 928766 [details]
Output of `emerge --info mesa`
Comment 2 Matt Turner gentoo-dev 2025-05-14 15:54:22 UTC
Thanks for the report.

The errors in `dmesg` look like link training failures. That is, issues on the display-side of things, rather than Mesa.

Can you try 25.1.0 again and double check that it really is something related to Mesa causing this?
Comment 3 opensauce04 2025-05-14 21:08:16 UTC
Upgraded to Mesa 25.1.0 again, wit that being the only package touched by Portage, and experienced the same issue. Downgraded to 25.0.5 again, and there were no issues again. Is there anything else I can try to help diagnose the issue?
Comment 4 Matt Turner gentoo-dev 2025-05-15 12:05:11 UTC
Okay, thanks. That sounds conclusive.

Let's gather some information:

- `lspci -nn` to see what NVIDIA graphics card you have

From the firmware messages I think it's a 3060 Ampere.

- With mesa-25.0.5 installed, `eglinfo` output to show what driver is in use

- With mesa-25.1.0 installed, `LIBGL_DEBUG=verbose eglinfo` output to show some messages about what driver it's trying to load and perhaps why it's failing

You might also try without VIDEO_CARDS=zink. I believe in 25.0.5 the nouveau OpenGL driver is in use, and in 25.1.0 nvk+zink is the new default for OpenGL.
Comment 5 opensauce04 2025-05-15 16:13:02 UTC
Created attachment 928899 [details]
`eglinfo` from Mesa 25.0.5
Comment 6 opensauce04 2025-05-15 16:13:20 UTC
Created attachment 928900 [details]
Verbose `eglinfo` from Mesa 25.1.0
Comment 7 opensauce04 2025-05-15 16:13:34 UTC
I will try without Zink in a few hours
Comment 8 Matt Turner gentoo-dev 2025-05-15 17:50:58 UTC
Indeed, it switched from nouveau to zink with the update.
Comment 9 opensauce04 2025-05-15 20:18:13 UTC
Removing `zink` from `VIDEO_CARDS` made the issue go away, so this seems to be a Zink problem.
Comment 10 opensauce04 2025-05-15 20:56:12 UTC
By the way if it wasn't clear from the information I've provided, I'm using a 3060 Ti.