Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#40057 - [nvidia] Xorg Freeze, nvidia 337.12, EQ overflowing

Attached to Project: Arch Linux
Opened by Rodrigo M (RD777) - Wednesday, 23 April 2014, 19:59 GMT
Last edited by Sven-Hendrik Haase (Svenstaro) - Sunday, 29 June 2014, 15:17 GMT
Task Type Bug Report
Category Packages: Extra
Status Closed
Assigned To Ionut Biru (wonder)
Sven-Hendrik Haase (Svenstaro)
Architecture x86_64
Severity Medium
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:
Experiencing constant xorg freezes where only the mouse moves. Xorg log states that there is an ongoing mi(EQ) overflow. However, I have not been able to detect the actual source of the problem. Freezes can happen when moving windows around, opening smplayer, resizing windows on KDE. This happened also with the previous version of nvidia 334.12 but does not happen on 331.39 or 331.38. Freezes can happen with sli enabled or disabled, with desktop effects enabled or disabled. They can happen 20 seconds after Xorg has started or 1 hour.

System specs:
3.14.1-1 x86_64
nvidia 337.12
2x Nvidia GTX660 TI (sli on/off)
KDE
xorg 1.15.1-1


Additional info:
* package version(s)
linux 3.14.1-1 x86_64
nvidia 337.12
xorg 1.15.1-1
* config and/or log files etc.
Attached Xorg log and backtrace done with gdb
http://pastebin.com/KptEMPWG
http://pastebin.com/d8fXEXG7


Steps to reproduce:
Moving or resizing windows.
Launching windows

   Xorg.crash.log (8 KiB)
   gdb-Xorg.txt (7.3 KiB)
This task depends upon

Closed by  Sven-Hendrik Haase (Svenstaro)
Sunday, 29 June 2014, 15:17 GMT
Reason for closing:  Deferred
Comment by Rulet (rulet) - Thursday, 24 April 2014, 09:35 GMT
Rodrigo M, how did you downgrade to nvidia 331.49? Because nvidia 331.49 requares kernel<3.14
Comment by Rodrigo M (RD777) - Thursday, 24 April 2014, 16:16 GMT
I downgraded the kernel as well as mesa.

EDIT: I made a mistake, I overwrote the xorg.log with the output of gdb. Here is an old version of the xorg.log during a crash with nvidia-334.21 and kernel 3.13.7

http://pastebin.com/0CGDkK02

I will try to post a new xorg.log tonight.
Comment by Rulet (rulet) - Thursday, 24 April 2014, 16:26 GMT
Rodrigo M, and to what kernel you downgraded and what exactly packages you installed?
Have you used this commanf before downgrading?:
sudo pacman -Rdds nvidia nvidia-utils nvidia-libgl
And what nvidia card are you using?

Comment by Rodrigo M (RD777) - Thursday, 24 April 2014, 17:20 GMT
I am using two Nvidia 660ti, however the problem occurs with sli enabled/disabled.
I downgraded to the kernel 3.13.8
I downgraded to nvidia 331.38
nvidia-utils 331.38
nvidia-libl 331.38
mesa 10.0.3

As for the command I used
sudo pacman -U downgrading all packages thanks to /var/cache/pacman/pkg with the same -U
Comment by Rodrigo M (RD777) - Thursday, 24 April 2014, 17:32 GMT
I have also compiled the kernel and the nvidia-blob on my computer with the same version of gcc to see if the problem could be fixed easily. It didn't work. I have also changed from TSC to HPET which didn't seem to stop the crashes. I have not tried to change the size of miEQ and compile xorg, it's one of the things I'll try this weekend.
Comment by Rulet (rulet) - Thursday, 24 April 2014, 18:11 GMT
By the way, what DE are you using? I'm using Gnome which is in Arch linux now and there are glitches even in terminal(slow moving of
cursor). Can this be because of nvidia drivers?



Comment by Rodrigo M (RD777) - Friday, 25 April 2014, 01:12 GMT
I am currently using KDE. I am sorry but I could not help you to determine if it's the cause of your glitches, although I can tell you that 334.21 and 337.12 are really buggy drivers according to the nvidia forums.
Comment by Rulet (rulet) - Friday, 25 April 2014, 07:43 GMT
That's what I can't understand, why not to use latest stable nvidia driver 331.67? And by the way. it possible to use nvidia driver from official site somehow to repack current nvidia arch package?
Comment by Börje Holmberg (linfan) - Monday, 05 May 2014, 17:22 GMT
I downgraded nvidia, nvidia-utils, nvidia-libgl, lib32-nvidia-utils and lib32-nvidia-libgl to version 334 and put them in IgnorPkg in /etc/pacman.conf. Typing in gnome was delayed with 337. I would also prefer not to use beta drivers unless you have enabled testing.
Comment by Rodrigo M (RD777) - Wednesday, 14 May 2014, 00:52 GMT
Here is a new log, happened with kernel 3.14.3-2 and nvidia-blob 337.12. Same problem, happened after 15 min of uptime.

http://pastebin.com/pNbtDc0u
Comment by Sven-Hendrik Haase (Svenstaro) - Saturday, 31 May 2014, 16:10 GMT
Test 337.25 please.
Comment by Börje Holmberg (linfan) - Saturday, 31 May 2014, 16:23 GMT
All seems fine now for me once i ditched gnome 3 some days ago. Again enjoying linux :-) Running mate with gtk2. Guess being bad ass bleeding edge has its pro's and con's. But luckily there are choices with linux.
Comment by Rodrigo M (RD777) - Sunday, 08 June 2014, 14:01 GMT
I am currently testing 337.25. I have forced powermizer to the maximum level and so far it seems to help.
Without powermizer forced and with sli enabled, I am able to reproduce the crash after only 5 or 10 min of uptime. I will add an xorg log.
At the moment, I am testing powermizer forced with sli disabled to see how long it takes for my xorg to crash.
I will also test sli disabled without powermizer forced.

On another note, I appreciate the help. This problem is quite annoying since what is the point of using a rolling-release distro if I have to downgrade the kernel and nvidia for my system to remain stable or not crash every 20 min..... :(
Comment by Rodrigo M (RD777) - Sunday, 08 June 2014, 20:41 GMT
6 hours of uptime at the moment, however, not without a lot of user inputs.... Movies playing on a loop. (Sli disabled, powermized on adaptive mode) (latest kernel and 337.25)
Comment by Sven-Hendrik Haase (Svenstaro) - Sunday, 29 June 2014, 15:16 GMT
This might be fixed in one of the coming kernel releases as aplattner just pushed a relevant fix into mainline that might also hit this bug.

I'm not really sure what we can do here. The user worked around the bug for now and we don't seem to have other users who are hit by this. There are no further patches.

All in all, I'm going to close this because there doesn't appear to be anything we can do anyway.

Loading...