| View previous topic :: View next topic |
| Author |
Message |
drumgod
n00b
Joined: 12 Dec 2003
Posts: 61
|
 Posted: Thu Sep 25, 2014 8:15 pm Post subject: Corrupted and missing screen/fonts/letters? (with screencap) |
 |
|
Hi guys,
My X screen and frame buffer consoles have varying degrees of corruption and or fuzzy/missing letters. (see screencap)
The card is a PCI version of a GeForce 6200 (NV40 according to nouveau) with 512M
The machine is a dual Xenon 3gig Dell PowerEdge SC1430.
Things I've tried in various combinations:
emerge -e world
Kernels 3.14, 3.16, 3.16.2, 3.16.3 built from kernel.org sources
Nouvea driver with KMS (3.16* kernels)
Nvidia driver with vesa framebuffer (3.14, 3.16.2 and 3.16.3 kernels)
Downgraded mesa, xorg-server, eudev
Moved video card to different slot
Replaced video card with the same model
Things I've noticed:
Some combos are more stable but none work completely reliably. (e.g. Everything works fine but capital Cs are not displayed)
I know things are going to go bad quickly when my fonts are slightly larger in claws mail.
Nouveau often reports cache read errors and reports GPU lockup. (but not always)
Old vesa framebuffer (passing vga=795 IIRC) seems pretty stable but scrolls extremely slowly.
On startup, between grub and nouveau/kms framebuffer, the screen goes static-y with patterns that look like a stereogram.
Often, when nouveau/kms framebuffer comes up, the "static" remains at the top and bottom of the screen until I change consoles.
Screencap is here <-click
The large solid black lines across the screencap were not visible at the time the screencap was taken.
The more static-y lines in the background were visible.
Any ideas on what's causing this? TIA |
|
| Back to top |
|
 |
kernelOfTruth
Watchman
Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)
|
|
| Back to top |
|
|
drumgod
n00b
Joined: 12 Dec 2003
Posts: 61
|
| Posted: Fri Sep 26, 2014 1:06 am Post subject: |
|
|
| kernelOfTruth wrote: | to rule out a hardware issue with broken memory chips on the gpu: ...
|
I replaced the entire, actual video card with another of the same manufacturer and model. I also tried them each in both of the two available slots, so I don't think it's a hardware issue... Well, at least not the hardware of the vid card anyway. (Hmmm. The PCI vid card is inserted a PCI-X slot if that helps any...)
I have not actually "exercised" either of the vid cards or the full system too much though. I kinda' dropped the hardware line when I threw the other card in and that didn't fix it. I did leave memtest86 running for several hours without issue and completed an error free "emerge -e world".
Is there any test in particular you can think of that might yield more helpful info? |
|
| Back to top |
|
|
kernelOfTruth
Watchman
Joined: 20 Dec 2005
Posts: 6111
Location: Vienna, Austria; Germany; hello world :)
|
| Posted: Fri Sep 26, 2014 4:44 pm Post subject: |
|
|
so if I understood it correctly:
No matter what different card (replacement, same card type, nvidia/nouveau) you're operating on that system (mainboard)
it's showing that behavior ?
Any bios settings you could tinker with related to PCI (timings, etc.) ?
A kernel configuration "issue" comes to mind - but for this we have to know since when this is happening ... (looks more like a hardware or driver issue though)
If I remember correctly I had slightly similar patterns with a GT7600 card when it was overheating or even before that card,
is the PSU ok ?
Well, if vga or vesa works it has to be an issue with KMS (most likely regression if it wasn't there right from the start with the system - but nvidia-drivers practically to 100% rules it out),
so trying e.g. SystemRescueCD with a 3.14 based (alternative) kernel version would be interesting to try out - if that also shows this pattern:
booting up -> keyboard layout -> wizard (X setup)
Any chance to run a somewhat older kernel with prior nvidia driver versions ? Perhaps they changed something, too and it got adopted into nouveau ?
This doesn't appear in VT/console, right ?
Could you get your hands on a different model ? (collegues, friends, ... ?)
_________________
https://github.com/kernelOfTruth/ZFS-for-SystemRescueCD/tree/ZFS-for-SysRescCD-4.9.0
https://github.com/kernelOfTruth/pulseaudio-equalizer-ladspa
Hardcore Gentoo Linux user since 2004  |
|
| Back to top |
|
|
drumgod
n00b
Joined: 12 Dec 2003
Posts: 61
|
| Posted: Mon Sep 29, 2014 9:21 pm Post subject: |
|
|
| kernelOfTruth wrote: | so if I understood it correctly:
No matter what different card (replacement, same card type, nvidia/nouveau) you're operating on that system (mainboard)
it's showing that behavior ?
|
That is correct.
| kernelOfTruth wrote: | Any bios settings you could tinker with related to PCI (timings, etc.) ?
|
I can change the IRQs used by each PCI device somewhat but nothing more than that.
| kernelOfTruth wrote: | A kernel configuration "issue" comes to mind - but for this we have to know since when this is happening ... (looks more like a hardware or driver issue though)
If I remember correctly I had slightly similar patterns with a GT7600 card when it was overheating or even before that card,
is the PSU ok ?
|
As far as I can tell. I haven't put a multi-meter on it though.
| kernelOfTruth wrote: | Well, if vga or vesa works it has to be an issue with KMS (most likely regression if it wasn't there right from the start with the system - but nvidia-drivers practically to 100% rules it out),
so trying e.g. SystemRescueCD with a 3.14 based (alternative) kernel version would be interesting to try out - if that also shows this pattern:
booting up -> keyboard layout -> wizard (X setup)
|
I've not used SystemRescueCD before but I grabbed it and booted the system.
I'm not quite sure I understand your -> notation above but these some of my boot results:
Alternative 64bit kernel + default options = (try one) total meltdown with completely garbled screen
Alternative 64bit kernel + default graphical environment = (try one) total meltdown with completely garbled screen
Alternative 64bit kernel + high resolution console (1024x768, lol) = comes up fine. Then "startx" works okay but X also in a painfully low resolution.
Alternative 64bit kernel + default graphical environment = same behavior as "high resolution" but with and even more painfully low resolution.
Alternative 64bit kernel + default graphical environment = (try two) frambuffer terminal came up okay. "startx" takes me to X with similar problems as my normal system.
Alternative 64bit kernel + default options = (try two) frambuffer terminal came up okay. "startx" takes me to X with similar problems as my normal system.
Ugh. Two different results with the same boot options. yuck. It would seem that I am having the same issues when booting from the rescue CD. Is there anything in particular you would like me to try while booting from the disk?
| kernelOfTruth wrote: | Any chance to run a somewhat older kernel with prior nvidia driver versions ? Perhaps they changed something, too and it got adopted into nouveau ?
|
I will see if I can downgrade kernel/nvidia to something compatable with each other...
| kernelOfTruth wrote: | This doesn't appear in VT/console, right ?
|
Standard console does not seem to be effected.
| kernelOfTruth wrote: | | Could you get your hands on a different model ? (collegues, friends, ... ?) |
Not easily. The problem being it's in a PCI-X slot. It would seem there aren't many PCI video cards around that are compatible with PCI-X slots.
Thanks again for your help... |
|
| Back to top |
|
|
drumgod
n00b
Joined: 12 Dec 2003
Posts: 61
|
| Posted: Wed Oct 01, 2014 2:16 pm Post subject: |
|
|
| Hmm... Well, everything has been working fine with kernel 3.12.28 and nvidia-drivers 173.14.39 |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Desktop Environments
