I’m running Fedora 36 with NVIDIA driver 515.65.01 via xorg-x11-drv-nvidia-515.65.01-1.fc36.x86_64 from rpmfusion. I’m running X11, not wayland. I am see 4-5 second freezes when a new GUI application (that uses the GPU?) is launched. I have found out that I can replicate the lag simply by calling glxinfo:
[ts@spark ]$ time glxinfo
[...]
real 0m4,501s
user 0m0,038s
sys 0m4,051s
Now to the really curious part - I understand if the following sounds crazy, but I have now reproduced it a few times: I can temporarily cure the problem by repeatedly running glxinfo. The first few times the command takes +4 seconds, but then it runs in a tenth of the time:
[ts@spark easyviz-3]$ time glxinfo
[...]
real 0m0,367s
user 0m0,027s
sys 0m0,037s
At the time the switch to fast happens, I see the following in dmesg output:
[Oct12 13:11] NVRM: GPU at PCI:0000:01:00: GPU-f1f1b432-dc15-a2ec-5790-de38a0615d5c
[ +0,000003] NVRM: Xid (PCI:0000:01:00): 61, pid='<unknown>', name=<unknown>, 0d20(321c) 00000000 00000000
There is nothing else of particular interest in the dmesg output related to nvidia or GPU, that I can spot:
$ dmesg -H|grep -iE 'gpu|nvidia'
[ +0,000000] Command line: BOOT_IMAGE=(hd0,gpt2)/vmlinuz-5.19.13-200.fc36.x86_64 root=UUID=c29c5306-bc62-4ffa-a82a-88334dd1ac74 ro rootflags=subvol=root rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau nvidia-drm.modeset=1 initcall_blacklist=simpledrm_platform_driver_init
[ +0,000001] Kernel command line: BOOT_IMAGE=(hd0,gpt2)/vmlinuz-5.19.13-200.fc36.x86_64 root=UUID=c29c5306-bc62-4ffa-a82a-88334dd1ac74 ro rootflags=subvol=root rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau nvidia-drm.modeset=1 initcall_blacklist=simpledrm_platform_driver_init
[ +0,017421] input: HDA NVidia HDMI/DP,pcm=3 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card2/input6
[ +0,000031] input: HDA NVidia HDMI/DP,pcm=7 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card2/input7
[ +0,000021] input: HDA NVidia HDMI/DP,pcm=8 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card2/input8
[ +0,000028] input: HDA NVidia HDMI/DP,pcm=9 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card2/input9
[ +0,000015] input: HDA NVidia HDMI/DP,pcm=10 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card2/input10
[ +0,000017] input: HDA NVidia HDMI/DP,pcm=11 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card2/input11
[ +0,000018] input: HDA NVidia HDMI/DP,pcm=12 as /devices/pci0000:00/0000:00:01.0/0000:01:00.1/sound/card2/input12
[ +0,181858] nvidia: loading out-of-tree module taints kernel.
[ +0,000006] nvidia: module license 'NVIDIA' taints kernel.
[ +0,006110] nvidia: module verification failed: signature and/or required key missing - tainting kernel
[ +0,044151] nvidia-nvlink: Nvlink Core is being initialized, major device number 511
[ +0,000493] nvidia 0000:01:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=io+mem
[ +0,043519] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 515.65.01 Wed Jul 20 14:00:58 UTC 2022
[ +0,060257] nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint.
[ +0,008284] nvidia-uvm: Loaded the UVM driver, major device number 509.
[ +0,037297] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 515.65.01 Wed Jul 20 13:43:59 UTC 2022
[ +0,002627] [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
[ +0,059976] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 0
[Oct12 12:46] nvidia-modeset: WARNING: GPU:0: Lost display notification (0:0x00000000); continuing.
[Oct12 12:49] nvidia-modeset: WARNING: GPU:0: Lost display notification (0:0x00000000); continuing.
[Oct12 13:11] NVRM: GPU at PCI:0000:01:00: GPU-f1f1b432-dc15-a2ec-5790-de38a0615d5c
nvidia-bug-report.log.gz (437.2 KB)