575 release feedback & discussion

@dinosaur is it in essence the same as you posted like 2 months ago or were there any important updates?

Thanks! :)

Linux 6.18 changes some irq-related API calls. The new patch accounts for them.

1 Like

Thanks a lot! I’ve recently forked CachyOS’ Pkgbuilds to build the 575 dkms packages with your previous patch, I’ll integrate your new one for kernel 6.18. If that can help anyone here feel free to use it.

Thanks for the update, For issue external Monitors not working we have different bug 5525444 where it got reproduced locally and engineering team is working on it. For current issue where CPU is reaching 100% along with display hang(5366590) is not reproducible in house. I am archiving 5366590 from our end, We request you to check on latest 580 driver reply back to this thread if you hit this issue again on latest driver along with nvidia-bug-report.

Hi @chris.bainbridge1

I have tried multiple times to reproduce issue on below handy notebook but could not duplicate issue.

Notebook rebooted successfully without any such error message.

Lenovo Legion R7000P2021H + Driver 570.124.04 + NVIDIA GeForce RTX 3060 + Dock W19TDS + 2 displays connected via DP and 1 display connected by Type C

Could you please help to test on latest 580 / 590 released driver and share feedback.

Hi @chris.bainbridge1

I was able to repro black screen issue on one of displays after numerous attempts and indeed the repro frequency is very low.

I am wandering to know if you upgraded to latest 580 branch or 590 drivers and still sees this issue.

@amrits Have you tried with “hybrid graphics” enabled in the BIOS? I just upgraded to 590.44.01 and saw the black screen issue almost immediately when opening/closing laptop lid with “hybrid graphics” enabled in the BIOS. I have not seen it yet with discrete graphics enabled in the BIOS, but this may just be because the probability is lower.

@amrits I just upgraded to 590.44.01 and booted with the dock connected and nvidia discrete graphics enabled in the BIOS (not hybrid). The errors were still present. nvidia-bug-report log attached.

nvidia-bug-report.log.gz (549.7 KB)

$ zcat nvidia-bug-report.log.gz |grep ERROR
2025-12-08T11:09:47.467745+00:00 ubuntu2404 kernel: [drm:nv_drm_dev_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add connector for NvKmsKapiDisplay 0x00000400
2025-12-08T11:09:57.553753+00:00 ubuntu2404 kernel: [drm:nv_drm_dev_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to get dynamic displays
2025-12-08T11:10:07.561713+00:00 ubuntu2404 kernel: [drm:nv_drm_handle_dynamic_display_connected [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add encoder for NvKmsKapiDisplay 0x00000001
2025-12-08T11:10:17.657740+00:00 ubuntu2404 kernel: [drm:nv_drm_handle_dynamic_display_connected [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add encoder for NvKmsKapiDisplay 0x00000002
2025-12-08T11:48:22.466698+00:00 ubuntu2404 kernel: [drm:nv_drm_dev_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add connector for NvKmsKapiDisplay 0x00000400
2025-12-08T11:48:32.555742+00:00 ubuntu2404 kernel: [drm:nv_drm_dev_load [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to get dynamic displays
2025-12-08T11:48:36.625746+00:00 ubuntu2404 kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
2025-12-08T11:48:39.697769+00:00 ubuntu2404 kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
2025-12-08T11:48:42.641753+00:00 ubuntu2404 kernel: [drm:nv_drm_handle_dynamic_display_connected [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add encoder for NvKmsKapiDisplay 0x00000001
2025-12-08T11:48:42.769758+00:00 ubuntu2404 kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
2025-12-08T11:48:47.953783+00:00 ubuntu2404 kernel: [drm:nv_drm_atomic_commit [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Flip event timeout on head 0
2025-12-08T11:48:52.641748+00:00 ubuntu2404 kernel: [drm:nv_drm_handle_dynamic_display_connected [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to add encoder for NvKmsKapiDisplay 0x00000002
...

I almost noticed many errors about PRIME failing:

[    44.261] (EE) NVIDIA(G0): Failed to configure Reverse PRIME on output DP-1-2.1.
[    46.898] (EE) NVIDIA(G0): Failed to configure Reverse PRIME on output DP-1-2.2.
[    46.910] (EE) NVIDIA(G0): Failed to configure Reverse PRIME on output DP-1-2.3.
[    46.913] (EE) NVIDIA(G0): Failed to configure Reverse PRIME on output DP-1-2.1.
[    60.579] (EE) NVIDIA(G0): Failed to configure Reverse PRIME on output DP-1-2.1.

Hi @chris.bainbridge1

I enabled Hybrid Graphics in BIOS and installed latest driver 590.44.01 and able to repro black screen once after trying more than 50 cycles of closing/opening lid.

I will share further updates once there is feedback available from engineering team.

Hi @chris.bainbridge1

I wanted to check if you observe any display freeze or any other issue on your system when you see such errors.

@amrits 2 of the external monitors don’t come up and just have black screen. But the actual laptop display seems ok.

@amrits That is low, it happens more often for me. I just did the same test - open or close the lid 50 times - and got a black screen 10 times. So on my setup the failure rate around 20%.

Dinosaur’s patch does not work on Debian’s 6.17 kernels, so I’ve put together relevant Debian patches for v550 and added one file from v580 and now it seems to work fine:
nvidia575-linux6.17-debian.patch.txt (9.1 KB)

I know this driver version is soon to be deprecated, but I’d like to report a bug with Monster Hunter Wilds.

As a reminder and as it has been already reported, the game crashes on RTX 50 with drivers 580+ (or has an absurd number of vertex explosions in the latest 590, but at least it doesn’t crash anymore). As such I cannot guarantee it has been resolved on the newer versions.

It happens that sometimes the game will crash when looking at specific equipments, in particular the “Numinous Crown”, either from the blacksmith or the inventory. Starting a hunt of Zoh Shia, the monter related to this equipment will also make the game crash. If I switch Proton version, it triggers a shader recompilation and then the problem disappears, which tends to indicate a shader issue (hence my suspicions about a driver issue). I don’t know if this is reproducible on older cards and with the latest drivers.

It tends to come back every other day, and it seems that it can be triggered running the game with an empty shader cache. By the way, the shader compilation is rather quick (3 minutes max), contrary to what I experienced on Windows (20+ minutes), I cannot tell if this is normal or another nail in the coffin of the shader compilation pipeline.

I cannot test it right now with the latest drivers, as I’m locked out of the game for 24h thanks to Denuvo, as switching Proton’s version is considered as a new installation of the game. I’ve attached two report logs from yesterday and today.

nvidia-bug-report.log.gz (920.7 KB)

nvidia-bug-report.log.gz (888.1 KB)

Just for the record, technically it is not even deprecated but already dead. It was a temporary “new feature branch” (non-production/stable drivers), and hasn’t received updates despite being affected by security vulnerabilities which is the easiest way to tell if a branch is no longer supported.

570.xx on the other hand is still supported last I know of.

Imagine NVIDIA would still be interested in knowing about regressions going from 570 to 575 that are still affecting 580+ though.

2 Likes

Thanks for pointing this! I think this wasn’t communicated loud enough anywhere here, so for everyone coming here in the future:

VERSION 575 HAS SECURITY ISSUES


and is probably unlikely to receive any fixes ;-(
I believe this is the relevant CVE list:

Hi @vrachatte i just want to let you know, that the issue seems to be solved on the latest 590 build.

1 Like

Dear @BlueGoliath

I wanted to check if you are still observing random black screen bug with latest 590 released drivers.

Thank you in advance for all the help and inputs.

I have not had any black screens with my 5060 TI on 590. It only happened with my 4060.

Thanks @BlueGoliath I spent few hours playing multiple steam titles on NVIDIA GeForce RTX 4080 but could not duplicate issue.

Would you be able to share again video / graphics settings for game title, steam launch options (if any) along with new bug report captured as soon as you see black screen.