Not sure if this is related, but I swapped out the 1660 for a GTX 1050 Ti. The system ran fine since my last post (10/29/2019) until today: There was a short freeze (1-2 secs) from which the x server recovered with the following journal:
Nov 07 12:07:06 . kernel: NVRM: GPU at PCI:0000:08:00: GPU-4bff5633-7dea-ca21-dbba-febd483902bc
Nov 07 12:07:06 . kernel: NVRM: GPU Board Serial Number:
Nov 07 12:07:06 . kernel: NVRM: Xid (PCI:0000:08:00): 68, pid=1569, CCMDs 00000027 0000c2b0
Nov 07 12:08:15 . /usr/lib/gdm-x-session[1565]: nvLock: client timed out, taking the lock
apart from the different GPU, the system is unchanged.
kaputt0nsqv: yours looks like a regular gnome bug of overqueueing frames. Try setting
__GL_MaxFramesAllowed=1
in your system environment to work around it.
The rest of you: you all seem to be running 3rd gen Ryzen, since this seems to only happen while (mostly) idle, maybe somthing wrong with pcie power management, try setting kernel parameters
pcie_aspm=off
and/or
processor.max_cstate=1
and check if that helps with the situation.
I also have the same problem, where my screen suddenly starts freezing and only mouse pointer works, while the rest of my applications that require GPU starts lagging. Iâm using a Ryzen 3600 on a MSI x570, with a RTX 2070 SUPER. I have tried applying @generixâs proposed solutions, setting pcie_asm=off didnât work â so iâm trying if processor.max_cstate=1 will help. Any other suggestions?
I got a similar problem. It occurs randomly, generally after a couple of hours
Nov 14 02:31:01 archdesktop kernel: NVRM: GPU at PCI:0000:05:00: GPU-2be4b6b5-9e1f-b8d5-6304-36fe416969c8
Nov 14 02:31:01 archdesktop kernel: NVRM: GPU Board Serial Number:
Nov 14 02:31:01 archdesktop kernel: NVRM: Xid (PCI:0000:05:00): 61, pid=441, 0cec(3098) 00000000 00000000
Nov 14 02:31:05 archdesktop /usr/lib/gdm-x-session[1457]: (II) event8 - Logitech MX Master 2S: SYN_DROPPED event - some input events have been lost.
Nov 14 02:31:05 archdesktop /usr/lib/gdm-x-session[1457]: (EE) client bug: timer event8 debounce: offset negative (-304ms)
Nov 14 02:31:05 archdesktop /usr/lib/gdm-x-session[1457]: (EE) client bug: timer event8 debounce short: offset negative (-317ms)
Nov 14 02:31:10 archdesktop /usr/lib/gdm-x-session[1457]: (II) event8 - Logitech MX Master 2S: SYN_DROPPED event - some input events have been lost.
Nov 14 02:31:23 archdesktop kernel: NVRM: Xid (PCI:0000:05:00): 8, pid=441, Channel 00000018
Nov 14 02:31:23 archdesktop /usr/lib/gdm-x-session[1457]: (EE) NVIDIA(0): The NVIDIA X driver has encountered an error; attempting to
Nov 14 02:31:23 archdesktop /usr/lib/gdm-x-session[1457]: (EE) NVIDIA(0): recover...
Nov 14 02:31:29 archdesktop /usr/lib/gdm-x-session[1457]: (WW) NVIDIA: Wait for channel idle timed out.
Nov 14 02:31:39 archdesktop /usr/lib/gdm-x-session[1457]: (EE) NVIDIA(GPU-0): Failed to initialize DMA.
CPU : AMD Ryzen 7 3700X Graphic card : Asus GeForce RTX 2080 SUPER Dual OC EVO Motherboard : ASUS TUF Gaming X570-Plus OS : Arch Linux 5.3.10-arch1-1 Driver version : 440.31-1 (dkms)
striffly, looks you already found the correct thread, didnât look at your hw specs.
Would be nice if the thread starter could edit the thread title to reflect that Ryzen 3rd gen systems are affected by this.
Did any of you have positive/negative results with limiting cstates, i.e. kernel parameter processor.max_cstate=1 ?
You have faced same issue recently, please help to provide applications or anything specific running which is causing issue. It will help me replicate issue locally.
RTX GeForce 2080 Ti
Ryzen 3900X
Ubuntu 19.10
nvidia-driver-435/eoan,now
Every couple days Iâll get the Xid 61 error and need to do a hard reboot. Iâve disabled all of the gnome extensions and have also set __GL_MaxFramesAllowed=1 as was mentioned, but to no avail.
MaxFramesAllowed was for a different issue.
On a another thread about a different issue on the same Ryzen platform, users reported that was fixed by a recently released bios update. Did you check if a bios update is available for your mainboard?
Hmm, I also use the same mobo as @striffly with a Ryzen 3600x and RTX 2070 SUPER, and get the same problem. Maybe the mobo and/or CPU is the source of the problem?
Itâs definitely the combination that triggers this bug
Ryzen 3rd gen CPU
ASUS x570 + B450 mainboards
Nvidia Turing GPU
Anyone not falling into that category? Otherwise my guess would be a bios bug since so far only one mainboard vendor seems affected.
Edit : Day 7 without any bugs. I reset the BIOS from the dedicated interface, but I suspect another one of my changes caused this bug resolution.
For those who have this Xid 61 bug: do you have your GC placed on the PCI slot at the bottom of your motherboard? Did you try to place it on the top one?
I have been trying to replicate issue by running applications like chrome; firefox; spotify and glmark2 benchmark on a system which is running since 3 weeks but no luck so far.
striffly, looking at the ASUS manual for their x570/B450 mainboards, running a single graphics card in any but the first slot (next to the cpu) is an unsupported configuration.
The slots are all for gpus but have to be populated in a specific order depending on number of gpus/Ryzen gen.
Iâm here to chime in, I also experience the same bug, RTX 2060. Happens around once a week and does not require any heavy load to occur. 3 times it happened to me so far:
Opera with Video playing (so HW is being used) + FTL
Opera with Video playing + image rendering in another opera window on 2nd screen
Opera with Video playing + image rendering in another opera window on 2nd screen
do you have your GC placed on the PCI slot at the bottom of your motherboard? Did you try to place it on the top one?
striffly, looking at the ASUS manual for their x570/B450 mainboards, running a single graphics card in any but the first slot (next to the cpu) is an unsupported configuration.
I have B450 MB. GC is in the top one next to the CPU.