All virtual terminals have become corrupt

Davidian1024 · 02-21-2005, 12:24 AM

I have run into an interesting problem. An X application caused my X server to crash. This left the video totally frozen. I was able to partially get out of this situation by logging into my machine from another through ssh. I killed the offending process but things did not return to normal. The video was still frozen on my machine. So I tried starting X from the remote terminal. X came up and ran normally.

This is where I am stuck. Now when I try to switch back to a virtual console it does not switch to a text mode. All that changes is a black area appears in the upper left corner of the screen. I have discovered that that black area is showing what is on the virtual console. It does not show text. I believe it is the graphical representation of the virtual terminals text buffer in memory.

I have blindly tried logging in on a virtual terminal and entering 'reset', and 'echo [ctrl+v][esc]c'. These have no effect. I am sure a reboot would fix this but I am interested in a more elegant solution. Mainly because I am trying to see how long I can keep the system running without rebooting it.

For now I will try to make due with a "graphics mode" only environment. Though, I don't know how long I can hold out. If anyone has any tricks, please let me know.

Davidian1024 · 02-21-2005, 08:54 PM

Anyone?

Artanicus · 02-22-2005, 04:09 AM

Im frequently suffering from the same problem.. I fix it the same way you do, and sometimes starting a new X and shutting it down restores the VT:s back in order.. What kinda video hardware and X versions are you running? It would be nice to be able to narrow this problem down and some day find a cure.. d:

Davidian1024 · 02-22-2005, 10:41 AM

Here's the info:

Radeon 9800 SE video card
nForce 2 chipset
Xorg 6.8.1
Linux kernel 2.6.10
ATI fglrx driver 8.10.19

Note: When this happened I was using version 8.8.25 of ATI's fglrx driver. I thought maybe if I upgraded that it might fix it. No such luck.

Artanicus · 02-22-2005, 10:45 AM

Im running:

Nvidia GeForce FX 5200
nForce 2 chipset
Xorg 6.8.2
Linux kernel 2.4.29
Nvidia driver 1.0-6629

The thing we have in common is the nForce 2 chipset.. Do you have the drivers installed for that? I dont.. hmm.

Davidian1024 · 02-22-2005, 10:48 AM

I wasn't aware of any nVidia driver for it. I've just been using the chipset driver that came with the kernel. 3D acceleration seems to work fine this way.

Artanicus · 02-22-2005, 10:51 AM

yeah, I was a tad suprised too to find out it actually has a driver.. it claims to be mainly for the onboard soundcard and ethernet, but since I dont use the sound and the ethernet works fine out of the box, I havnt installed the driver..

Could be just that its a tad buggy.. Getting weird errors from it upon boot.. *sigh*
Guess itll never be solved..

Davidian1024 · 02-22-2005, 10:56 AM

OK, I knew about the sound and NIC drivers for the nForce but I've never needed them so I left them alone. I've never gotten any errors from the chipset driver. What kind of errors are you getting.

Artanicus · 02-22-2005, 11:05 AM

a grab from the syslog from the time of my last boot, points of interest in bold:

Code:

Feb  6 12:15:05 indigo kernel: Detected 2171.623 MHz processor.
Feb  6 12:15:05 indigo kernel: Console: colour dummy device 80x25
Feb  6 12:15:05 indigo kernel: Calibrating delay loop... 4325.37 BogoMIPS
Feb  6 12:15:05 indigo kernel: Page-cache hash table entries: 262144 (order: 8, 1048576 bytes)
Feb  6 12:15:05 indigo kernel: CPU: AMD Athlon(tm) XP 3000+ stepping 00
Feb  6 12:15:05 indigo kernel: POSIX conformance testing by UNIFIX
Feb  6 12:15:05 indigo kernel: mtrr: v1.40 (20010327) Richard Gooch (rgooch@atnf.csiro.au)
Feb  6 12:15:05 indigo kernel: mtrr: detected mtrr type: Intel
Feb  6 12:15:05 indigo kernel: PCI: Probing PCI hardware (bus 00)
Feb  6 12:15:05 indigo kernel: PCI: nForce2 C1 Halt Disconnect fixup
Feb  6 12:15:05 indigo kernel: Initializing RT netlink socket
Feb  6 12:15:05 indigo kernel: Starting kswapd
Feb  6 12:15:05 indigo kernel: Console: switching to colour frame buffer device 128x48
Feb  6 12:15:05 indigo kernel: pty: 512 Unix98 ptys configured
Feb  6 12:15:05 indigo kernel: RAMDISK driver initialized: 16 RAM disks of 7777K size 1024 blocksize
Feb  6 12:15:05 indigo kernel: NFORCE2: BIOS didn't set cable bits correctly. Enabling workaround.
Feb  6 12:15:05 indigo kernel: NFORCE2: BIOS didn't set cable bits correctly. Enabling workaround.

Davidian1024 · 02-22-2005, 11:19 AM

Apparently I am getting an error from the chipset driver at startup

Code:

Feb  6 23:02:37 localhost kernel: NFORCE2: IDE controller at PCI slot 0000:00:09.0
Feb  6 23:02:37 localhost kernel: NFORCE2: chipset revision 162
Feb  6 23:02:37 localhost kernel: NFORCE2: not 100%% native mode: will probe irqs later
Feb  6 23:02:37 localhost kernel: NFORCE2: BIOS didn't set cable bits correctly. Enabling workaround.
Feb  6 23:02:37 localhost kernel: NFORCE2: 0000:00:09.0 (rev a2) UDMA133 controller