Closed VPaulV closed 1 year ago
same here. really confusing how things goes... i continuously testing which makes me tired.
UPD: 525.60.13 same problem
Is this only happening on Xorg or is it happening on Wayland too? I have a laptop GPU NVIDIA 3060 and I'm wondering whether I should update or not.
In my case it is xorg-server 21.1.4, would be nice if you could update and check if you have the same problem
I'm having similar issues with driver 525 as well.
Does this happen with the proprietary driver (of the same version) as well?
Yes
Operating System and Version
Arch latest
Kernel Release
Linux 6.0.12-arch1-1
Hardware: GPU
GPU 0: NVIDIA GeForce RTX 2060 Laptop GPU
Describe the bug
Problem happens when using an external display using hdmi, similar to the others who posted here, everything becomes super laggy and takes 3-4sec to render. For me this only happens when using only the external display with the laptop screen disabled, if i use both displays then it seems to work fine. But i dock my laptop when at home or during presentations so would prefer to be able to keep screen closed and only using external display or projector.
To Reproduce
Update to Nvidia Proprietary or Open 525
Connect external screen with HDMI
Bug Incidence
Always
Oh that is a good point, I totally forgot to write, for me the issue is only on laptops with intel+nvidia GPUs
I just updated, and I only faced a minor issue so far, which is my external monitor flickered a lot for almost a minute after boot. After that, I was even able to play my games (verified with Spider-Man Remastered and Witcher 3).
Operating System and Version
Manjaro (latest testing branch)
Kernel Release
Linux 6.1.0-1-MANJARO
Hardware: GPU
GPU 0: NVIDIA GeForce RTX 3060 Mobile
Driver version 525.60.11 (Installed through Manjaro repo, not sure if it is proprietary or open source one).
Using Wayland and connected monitor using HDMI.
SAMSUNG LU28R55 Resolution 3840x2160 Scale 115%
I have horrible flickering and ghosting / image retention (image burns in for 10+ minutes). Switching back to 470.161.03 (through mhdw) the problem is no longer persistent
Operating System and Version
Manjaro XFCE edition
Kernel Release
Linux 6.1 but also tested rollback to Linux 5.15
Hardware: GPU
GPU 0: NVIDIA GeForce GTX 1080 TI
Driver version
Using Xorg and Displayport on QHD KOORUI display (no clue what panel is inside this brand)
UPD: New driver - 525.78.01, same issue
I had time to check, it is broken from 525 version, 520 and below works just fine. Unfortunately, my expertise in GPU drivers is very limited, but if one could suggest in what part of driver this may happen I could try to partially revert patches and see if it helps
Is this only happening on Xorg or is it happening on Wayland too? I have a laptop GPU NVIDIA 3060 and I'm wondering whether I should update or not.
It's happening on Wayland for me. Somebody linked me this posthttps://www.linuxquestions.org/questions/showthread.php?p=6404174#post6404174, where there are some people indicating they've been able to get 525.78.01 working with kernel 6.1.5. Not sure if they're using Wayland, however.
I have similar issues. Archlinux, laptop, intel+nvidia GPU, using nvidia modeset kernel 6.1.1 + nvidia 525.60.11 works kernel 6.1.6 + nvidia 525.78.01 external display freeze/blackscreen
The only reproducible misbehavior I manage to check the log complains
[nvidia-modeset idling display engine timed out](nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000c57e:0:0:1128)
in the dmesg.
Same here, with errors relating to display idling:
[ 26.740169] nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000c67e:0:0:1128
I'm on Xorg, and it occours with both open and closed drivers
Same here
[ 22.898786] nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000c67e:0:0:1128
Info:
After update to 525 version, my external monitor gets to 1 frame per 10 seconds in case if only external monitor is enabled.
Driver: 525.85.12
Kernel: 5.15.0-58-generic
xserver-xorg-core: 2:21.1.3-2ubuntu2.5
Ubuntu 22.04.1 LTS
Feb 01 22:45:41 msi-u /usr/libexec/gdm-x-session[2336]: (II) NVIDIA(G0): Setting mode "HDMI-1-0: nvidia-auto-select @3840x2160 +0+0 {AllowGSYNC=Off, ViewPortIn=3840x2160, ViewPortOut=3840x2160+0+0}"
Feb 01 22:46:17 msi-u kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000c67e:0:0:1128
Feb 01 22:46:19 msi-u kernel: nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000c67e:0:0:1128
Feb 01 22:46:47 msi-u /usr/libexec/gdm-x-session[2336]: (II) NVIDIA(G0): Setting mode "HDMI-1-0: nvidia-auto-select @3840x2160 +0+0 {AllowGSYNC=Off, ViewPortIn=3840x2160, ViewPortOut=3840x2160+0+0}"
Are there any workaround besides downgrading the nvidia drivers and corresponding cuda version?
I have similar issues. Archlinux, laptop, intel+nvidia GPU, using nvidia modeset kernel 6.1.1 + nvidia 525.60.11 works kernel 6.1.6 + nvidia 525.78.01 external display freeze/blackscreen
The only reproducible misbehavior I manage to check the log complains
[nvidia-modeset idling display engine timed out](nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000c57e:0:0:1128)
in the dmesg.
Following up. Updated to Archlinux kernel 6.1.9 and nvidia 525.85.05. Still using modeset. External Monitor works fine for now. No more "timed out" errors in the dmesg neither.
Got a new update today.
One external screen continues black due to
[ 22.898786] nvidia-modeset: ERROR: GPU:0: Idling display engine timed out: 0x0000c67e:0:0:1128
the other one was working for 15 minutes and then died with the following info
[ 700.457227] pcieport 0000:00:01.1: pciehp: Slot(0): Link Down
[ 700.457231] pcieport 0000:00:01.1: pciehp: Slot(0): Card not present
[ 700.457240] snd_hda_intel 0000:01:00.1: can't change power state from D3cold to D0 (config space inaccessible)
[ 700.586286] NVRM: GPU at PCI:0000:01:00: GPU-d285da56-46b0-17be-5844-3892e0d2d716
[ 700.586294] NVRM: Xid (PCI:0000:01:00): 79, pid='<unknown>', name=<unknown>, GPU has fallen off the bus.
[ 700.586298] NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.
[ 700.586310] NVRM: A GPU crash dump has been created. If possible, please run
NVRM: nvidia-bug-report.sh as root to collect this data before
NVRM: the NVIDIA kernel module is unloaded.
[ 700.854261] snd_hda_intel 0000:01:00.1: can't change power state from D3cold to D0 (config space inaccessible)
[ 701.250270] snd_hda_codec_hdmi hdaudioC0D0: Unable to sync register 0x6f0100. -5
[ 701.250281] snd_hda_codec_hdmi hdaudioC0D0: HDMI: invalid ELD buf size -1
[ 701.250286] snd_hda_codec_hdmi hdaudioC0D0: HDMI: invalid ELD buf size -1
[ 701.250290] snd_hda_codec_hdmi hdaudioC0D0: HDMI: invalid ELD buf size -1
[ 701.250293] snd_hda_codec_hdmi hdaudioC0D0: HDMI: invalid ELD buf size -1
[ 701.450548] pci 0000:01:00.1: Removing from iommu group 8
[ 701.450566] NVRM: Attempting to remove device 0000:01:00.0 with non-zero usage count!
After the crash running nvidia-smi
produces the following message
Unable to determine the device handle for GPU0000:01:00.0: Unknown Error
Here's the bug report nvidia-bug-report.log.gz
I have observed that some people have been successful with the 6.* Linux kernel. As a result, I built 6.1.11 and tried every NVIDIA driver above version 515.86.01. Unfortunately, all of these drivers have the same issue with external displays, including the latest version 525.89.02.
I have observed that some people have been successful with the 6.* Linux kernel. As a result, I built 6.1.11 and tried every NVIDIA driver above version 515.86.01. Unfortunately, all of these drivers have the same issue with external displays, including the latest version 525.89.02.
I have the same problem as yours (but with an AMD+NVIDIA GPU laptop), tested every single combination, I'm currently using 520.56.06, if I upgrade the bug reappears again. I'm not sure if it's an issue with Nvidia or xorg, a fix was already provided on xorg for this issue https://gitlab.freedesktop.org/xorg/xserver/-/issues/1028 but it no longer works on latest versions of Nvidia drivers.
I have same issue, and can reproduce in different way. My setup is Ubuntu server with RTX 1660 Super (proprietary NVIDIA 525 driver ) with KVM, while I have power up KVM connected to HDMI port everything is working correctly. Issue is happening when I disconnecting power from KVM what is recognized by GPU as monitor disconnection, in syslog start to see log:
Feb 19 17:01:29 kernel: [27595.505195] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device HDMI-0
Frame rate drops to few fps and it is visible in nvenc stat x264 stream.
I'm having similar issues with driver 525 as well.
Does this happen with the proprietary driver (of the same version) as well?
Yes
Operating System and Version
Arch latest
Kernel Release
Linux 6.0.12-arch1-1
Hardware: GPU
GPU 0: NVIDIA GeForce RTX 2060 Laptop GPU
Describe the bug
Problem happens when using an external display using hdmi, similar to the others who posted here, everything becomes super laggy and takes 3-4sec to render. For me this only happens when using only the external display with the laptop screen disabled, if i use both displays then it seems to work fine. But i dock my laptop when at home or during presentations so would prefer to be able to keep screen closed and only using external display or projector.
To Reproduce
Update to Nvidia Proprietary or Open 525 Connect external screen with HDMI
Bug Incidence
Always
Just an update to my issue, till now I was using Nvidia 520.xx dkms with the latest stock kernel so I could use presentation mode but with the release of kernel 6.2, Nvidia 520.xx dkms modules stopped building into the kernel. With 520 no longer building in 6.2 and 525 still lagging when in presentation mode, I decided to try Nvidia-beta from the AUR. nvidia beta is currently using version 530.xx and so far it seems to have resolved my lag issue in multi-monitor presentation mode.
I'm using regular Nvidia beta but I do believe there is an Nvidia-open-beta in the AUR. Everyone's issue in this thread was similar but not exactly the same so I can't promise it will fix everyone's problems but if your on Arch it's worth a shot. For other distros, you may be able to get Nvidia beta source from their site. Anyway just wanted to update everyone here since till now there was no real solution to our problem. Hope it helps.
Same here, I have 3090 with 4 monitors on Arch Linux. Lagging a lot since the 525 driver update.
I'm having similar issues with driver 525 as well. Does this happen with the proprietary driver (of the same version) as well? Yes Operating System and Version Arch latest Kernel Release Linux 6.0.12-arch1-1 Hardware: GPU GPU 0: NVIDIA GeForce RTX 2060 Laptop GPU Describe the bug Problem happens when using an external display using hdmi, similar to the others who posted here, everything becomes super laggy and takes 3-4sec to render. For me this only happens when using only the external display with the laptop screen disabled, if i use both displays then it seems to work fine. But i dock my laptop when at home or during presentations so would prefer to be able to keep screen closed and only using external display or projector. To Reproduce
Update to Nvidia Proprietary or Open 525 Connect external screen with HDMI
Bug Incidence Always
Just an update to my issue, till now I was using Nvidia 520.xx dkms with the latest stock kernel so I could use presentation mode but with the release of kernel 6.2, Nvidia 520.xx dkms modules stopped building into the kernel. With 520 no longer building in 6.2 and 525 still lagging when in presentation mode, I decided to try Nvidia-beta from the AUR. nvidia beta is currently using version 530.xx and so far it seems to have resolved my lag issue in multi-monitor presentation mode.
I'm using regular Nvidia beta but I do believe there is an Nvidia-open-beta in the AUR. Everyone's issue in this thread was similar but not exactly the same so I can't promise it will fix everyone's problems but if your on Arch it's worth a shot. For other distros, you may be able to get Nvidia beta source from their site. Anyway just wanted to update everyone here since till now there was no real solution to our problem. Hope it helps.
Ubuntu 22.10 Nvidia RTX3060 Laptop AMD Ryzen 5 5600H with AMD graphics 530.30.02 driver (beta)
Steel shit. Now I hate NVIDIA. On my second exactly the same laptop, but with graphics from AMD - everything works.
@IvanGreen It's really unfortunate, they just don't care, it's been at least 2 years since this bug was mentioned (on other forums). It was somehow fixed with version 520.56.06 but next versions reintroduced the bug. The two options for now are to keep using the old version or to disable the AMD GPU and relies on Nvidia only.
Ubuntu 22.10 Nvidia RTX3060 Laptop AMD Ryzen 5 5600H with AMD graphics 530.30.02 driver (beta)
Steel shit. Now I hate NVIDIA. On my second exactly the same laptop, but with graphics from AMD - everything works.
Did you try to update kernel as well? (I didn't test myself, just curious about your results). Also, nowadays NVIDIA has way better Linux support than they had 5 years ago, so no reason to hate them. For now you can just use the 515/520 drivers, they are fine
Ubuntu 22.10 Nvidia RTX3060 Laptop AMD Ryzen 5 5600H with AMD graphics 530.30.02 driver (beta) Steel shit. Now I hate NVIDIA. On my second exactly the same laptop, but with graphics from AMD - everything works.
Did you try to update kernel as well? (I didn't test myself, just curious about your results). Also, nowadays NVIDIA has way better Linux support than they had 5 years ago, so no reason to hate them. For now you can just use the 515/520 drivers, they are fine
Yes, I tried to update the kernel - unfortunately it did not help. I have two additional monitors working (2.5k, 21:9 each, one hdmi, one type-c) only with version 420 drivers. And if I turn off programmatically the main display of the laptop, friezes begin. I haven't tried the version with 520.56.06 drivers yet, but I've tried it with the proprietary 515/520 drivers (I don't know the sub-version) - it still doesn't work.
Hey guys, great news. Just have tried last Linux LTS kernel with 530.41.03 driver and everything works as should. Please try and confirm. My setup:
Please note if you are using IBT you would want 6.2+ kernel.
This problem is brutal because I was running on 515 to avoid it on Ubuntu 22.10, but at least in the current beta the 515 totally hang the systemd-udevd
kernel task in the 6.2 kernel which ships with ubuntu 23.04 beta.
I can confirm that upgrade to v530 fixed the problem for me, though.
Can we probably close the issue since it seems to be resolved with v530+ and 6.1+ kernel?
NVIDIA Open GPU Kernel Modules Version
525.60.11
Does this happen with the proprietary driver (of the same version) as well?
Yes
Operating System and Version
Gentoo Linux
Kernel Release
Linux home 5.15.80-gentoo-x86_64
Hardware: GPU
GPU 0: NVIDIA GeForce RTX 3080 Laptop GPU
Describe the bug
Hi Guys,
After updating to the 525.60.11 (both opensource and proprietary) I have a problem with the external screen. When I connect (with hdmi) everything on the screen is extremely laggy, like rendering happens on the intel GPU instead of the NVIDIA card. By laggy I mean every change on the screen takes a few seconds to redraw. This happens only on the Intel+NVIDIA GPU laptops.
Please let me know if I can provide any additional logs that will help.
To Reproduce
Bug Incidence
Always
nvidia-bug-report.log.gz
nvidia-bug-report.log.gz (515.86.01 version that works as should) nvidia-bug-report.log.gz (525.60.11 with extremely laggy external display)
More Info
No response