I am a Archlinux user, and update my kernel and nvidia driver all the time, and with every combination I tried in at least the last 6 months I have the same issue with nvidia drivers.
To reproduce, keep scrolling(page up and page down multiple times, fast) a text file in sublime text editor 2 or transfer a large number of files on filezilla, then this start to appear on error log, and after a time the system X freeze, music keeps playing and mouse moving, but I can’t change to terminal mode, and need to do a hard reset.
Ago 21 17:25:45 kernel: NVRM: Xid (0000:01:00): 6, PE0001
Ago 21 17:25:48 kernel: NVRM: Xid (0000:01:00): 6, PE0001
Ago 21 17:25:49 kernel: NVRM: Xid (0000:01:00): 6, PE0001
Ago 21 17:25:50 kernel: NVRM: Xid (0000:01:00): 6, PE0001
[This file was removed because it was flagged as potentially malicious] (76.5 KB) nvidia-bug-report.log.gz (94 KB)
Updated log with the error, was listening to music with audacious and browsing when the crash happen, I could remotely login with ssh and execute the bug-report
This error generally indicates that data was corrupted on its way from the CPU to the GPU, either in the GPU itself or because some other system component corrupted it. Do you have a different GPU you could try maybe? If your system has multiple PCIe slots, try moving the GPU from one slot to another. Please also make sure that the graphics board is firmly seated in the socket and that it is not overheating and that the heatsink and fan (if it has one) are clear of dust and debris.
It’s a computer from my work, I can’t open to change or clean, but I had the same problem with my previous computer here, with a 7300gs, I can play games(Counter strike 1.6, Stealth Bastard Deluxe, browse and watch videos for days without rebooting), I only have this problem with this two programs, I do not have other platform with nvidia where I can install archlinux to try now.
i’ve also had this problem with an nvidia quadro nvs 295 and fedora 20. i stay up-to-date against the fedora updates-testing repo and use the rpmfusion repo for my drivers (currently 304.119). i use four screens and xinerama (no twinview) to bind them into one display. i can’t reproduce this problem in a deterministic way. i can’t identify any pattern of use that triggers it. but it happens frequently (sometimes a few times during the day, sometimes every few days). it’s been plaguing me for about a month. it started after a fedup upgrade from fedora 19 to fedora 20. did anyone every find a solution to this? i found this post because i have similar message in /var/log/messages.
I have the exact same video device (Quadro nvs 295), and have the same errors. It only started happening after upgrading to Fedora 20 (Fedora 18 worked perfectly) and am using rpmfusion’s 331.89.
In my case, it usually occurs around the time the screensaver (just displaying graphics emulating monochrome text display) has run for hours, sometime between late night and early morning. The PC is still running, as I can ssh to it from another PC.
It also almost always occurs after suspend/wakeup, where the video locks up with either blank or noisy and hung display (not pixel glitches.)
I tried dusting off the video card, but hasn’t solved it.