X-Star DP2710 Monitors not recognized

Question

I recently purchased two Korean IPS monitors, however, on plugging them in, all I was able to see was a flashing coloured screen, I've read elsewhere that this is related to EDID:

$dmesg | grep edid 
[    2.700416] [drm:drm_edid_block_valid] *ERROR* EDID checksum is invalid, remainder is 128

After some shinanigans, I switched to the nouveau drivers using the following instructions:

1. sudo nvidia-settings --uninstall
2. sudo apt-get remove --purge nvidia*
3. sudo apt-get remove --purge xserver-xorg-video-nouveau xserver-xorg-video-nv
4. sudo apt-get install nvidia-common
5. sudo apt-get install xserver-xorg-video-nouveau
6. sudo apt-get install --reinstall libgl1-mesa-glx libgl1-mesa-dri xserver-xorg-core
7. sudo dpkg-reconfigure xserver-xorg

Source: Ubuntu Wiki

They both work flawlessly for about ten minutes up to about 39 minutes so far (and let me tell you how glorious 5120x1440 is...) But once that ten minutes is up:

1. they both get covered in squares of colour almost like pixelation, OR;
2. my system freezes entirely to the point that all I can do is REISUB (GPU lockup - switching to fbcom is what I saw once when it responded to ctrl + alt + f1 one time).

syslog

Jan  8 12:04:31 monster kernel: [ 6384.233289] nouveau E[  PGRAPH][0000:01:00.0] TRAP ch 4 [0x027f99f000 compiz[2422]]
Jan  8 12:04:31 monster kernel: [ 6384.233298] nouveau E[  PGRAPH][0000:01:00.0] GPC1/TPC0/TEX: 0x80000049
Jan  8 12:04:31 monster kernel: [ 6384.233302] nouveau E[  PGRAPH][0000:01:00.0] GPC1/TPC1/TEX: 0x80000049
Jan  8 12:04:31 monster kernel: [ 6384.233308] nouveau E[  PGRAPH][0000:01:00.0] GPC2/TPC0/TEX: 0x80000049
Jan  8 12:04:31 monster kernel: [ 6384.233312] nouveau E[  PGRAPH][0000:01:00.0] GPC2/TPC1/TEX: 0x80000049
Jan  8 12:04:31 monster kernel: [ 6384.233330] nouveau E[   PFIFO][0000:01:00.0] PFIFO: read fault at 0x0018ad9000 [PAGE_NOT_PRESENT] from (unknown enum 0x00000000)/GPC2/(unknown enum 0x00000004) on channel $
Jan  8 12:04:38 monster kernel: [ 6390.861216] nouveau E[  PGRAPH][0000:01:00.0] TRAP ch 4 [0x027f99f000 compiz[2422]]
Jan  8 12:04:38 monster kernel: [ 6390.861222] nouveau E[  PGRAPH][0000:01:00.0] ROP0 0x80000000 0x80000001
Jan  8 12:04:38 monster kernel: [ 6390.861226] nouveau E[  PGRAPH][0000:01:00.0] ROP1 0x80000000 0x80000001
Jan  8 12:04:38 monster kernel: [ 6390.861230] nouveau E[  PGRAPH][0000:01:00.0] ROP2 0x80000000 0x80000001
Jan  8 12:04:44 monster kernel: [ 6396.660693] SysRq : This sysrq operation is disabled.
Jan  8 12:04:45 monster kernel: [ 6397.660695] SysRq : This sysrq operation is disabled.
Jan  8 12:04:46 monster kernel: [ 6398.724695] SysRq : Emergency Sync
Jan  8 12:04:46 monster kernel: [ 6398.750943] Emergency Sync complete
Jan  8 12:04:46 monster kernel: [ 6399.156691] SysRq : Emergency Remount R/O

---

In an attempt to fix this issue, I've tried installing 304,319 proprietary drivers. However, when I install the proprietary drivers I get a black screen on load and a direct reduction to 640x480 with only one screen functional. This presents another problem. Typically, the following would work to reset Unity:

dconf reset -f /org/compiz/
unity --reset-icons &disown

However, this fails saying that "GlxQueryExtension string is null for screen 0". So, when this happened, I figured maybe I need to go and set things up in nvidia-settings. So, I try opening nvidia-settings but it tells me:

You do not appear to be using a Nvidia X driver you should run nvidia xconfig

So I try running sudo nvidia-xconfig, nothing happens and here I am.

I've been reading elsewhere and I believe I may need to write a custom xorg.conf file for these monitors. But I have no idea how to do that.

---

System Specs:

I built this PC from scratch so I have detailed information about every part I used:

http://ca.pcpartpicker.com/p/QbH8

i7-3930k
32GB RAM
250GB SSD
EVGA GTX 660
2 x X-Star DP2710 2560x1440 monsters.. I mean monitors.

Help me get back to this, permanently and win a lifetime of good karma:

Research:

Blank monitor, flicker, snow, or other random live image corruption

If you use dual-head (two video outputs in use) or dual-link (a very high resolution monitor), and the VBIOS has too low clocks by default, the video memory bandwitdh will be insufficient for scanout. This can lead to a blank image, or live image corruption, that changes on every vertical refresh, and likely does not show up in screencaptures. For the monitor, this is not a case of an unsupported video mode but random video data.

Raising the card performance mode might help. Ask on IRC, #nouveau channel, how to do that. Instructions are not given here, because in the worst case, it may destroy your card, because power management is still a work in progress.

Source: http://nouveau.freedesktop.org/wiki/TroubleShooting/

Related Bugs:

https://bugs.freedesktop.org/show_bug.cgi?id=72180

Answer 1

This seems to be a bug related to the GTX 660

See: Bug 72180 – [NVE6] Random GPU Lockups, works with blob PGRAPH fw

Steps

1. Download: firmware_vbios.zip
2. Extract the files to an otherwise empty folder
3. Ensure that you have NVE6 chipset, in a terminal, type:

dmesg | grep nouveau | grep Chipset

You should receive something like... verify NVE6 in brackets

[ 2.318701] nouveau [ DEVICE][0000:01:00.0] Chipset: GK106 (NVE6)

4. Copy Files

sudo mkdir /lib/firmware/nouveau
sudo cp /path/to/files/* /lib/firmware/nouveau

5. Update initrd

Replace your kernel with the kernel that you are presently using, push tab for a listing:

sudo update-initramfs -c -k <YOUR_KERNEL>

6. Update kernel boot options

sudo nano /etc/default/grub

Add nouveau.config=NvGrUseFW=1 to GRUB_CMDLINE_LINUX_DEFAULT so that it looks something like:

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash nouveau.config=NvGrUseFW=1"

7. Restart and verify

sudo shutdown now -r
# on restart...
dmesg | grep external

Look for a line that looks like

[    2.484773] nouveau  [  PGRAPH][0000:01:00.0] using external firmware