169.07 BUG: The 7 series multi core bug still exsists.

[alexhavy]

Originally Posted by Rainmaker87

It almost seems like the issue are primarily with the Go cards (at this point), anyone else notice that?

I have an issue with nvidia driver. Have HP Xeon with a single hiperthreaded CPU and NVIDIA Quadro NVS 285 card. The driver causes reboot of my system (Red Hat EL 2.9.67) seconds (up to a minute or two) after the X starts. No errors in logs. Tried 3 driver versions up to 169.07, nothing works. Submitted a bug report to NVIDIA - no response. Will eventually try 169.09. Doubt it will work though... This is some kind of conflict with the kernel or stability problem, no one wants to deal with it.

My original post is here:
http://www.nvnews.net/vbulletin/showthread.php?t=106302

[sk8glad]

I can confirm that I have the same bug (flickers and hardfreezes) with a Go7400 and on another laptop with a Go7300. The first one is an HP laptop, the other is a Dell. I've had this bug for quite some time too (I think more than six months now). The only way to resolve it for me is to restart X.

[clint]

Me too. Go7400 hard freezes but only when DPMS kicks in. Only way to resolve is reboot.

[99vadim]

Read this - http://www.nvnews.net/vbulletin/show...t=96673&page=3. Adding options nvidia (or nvidia_new depending on your distro) NVreg_RegistryDwords="PerfLevelSrc=0x2222" to /etc/modprobe.d/nvidia (or nvidia-kernel-nkc, once again it depends on the distro) resolves the freezing issues.

[pferraro]

Originally Posted by 99vadim

Read this - http://www.nvnews.net/vbulletin/show...t=96673&page=3. Adding options nvidia (or nvidia_new depending on your distro) NVreg_RegistryDwords="PerfLevelSrc=0x2222" to /etc/modprobe.d/nvidia (or nvidia-kernel-nkc, once again it depends on the distro) resolves the freezing issues.

This is NOT an acceptable solution for most laptop users (i.e. users with Go cards) who have battery and heat dissipation limitations since this registry value will disable frequency scaling on the GPU.

[99vadim]

On my GeForce Go 7200 laptop the difference in temperature is NONE. Thermal monitor shows the same value with and without the PerfLevelSrc=0x2222 option. The only difference on my HP laptop is lack of freezes & flashes.

[clint]

But setting the PerfLevelSrc to 0x2222 is a hack without statement from nvidia so far. I guess we can expect an official way of solution.

[pferraro]

Originally Posted by 99vadim

On my GeForce Go 7200 laptop the difference in temperature is NONE. Thermal monitor shows the same value with and without the PerfLevelSrc=0x2222 option. The only difference on my HP laptop is lack of freezes & flashes.

On higher end cards (e.g. my GeForce Go 7950 GTX) heat indeed becomes a problem at these higher clock speeds. I start to see red pixel artifacts when the temperature gets too high.
http://en.wikipedia.org/wiki/Compari...ce_Go_7_series

[SilentLexx]

Xorg freeze again, even on 6600 + 169.09 ((

But system work more stable with geforce 6600.

[hilbert70]

I had also freezing problems with my desktop. But I seem to have fixed it with the kernel parameter "pci=routeirq" this because I saw non carched irq in my messages and this seems to fix my freezing problem.

[pferraro]

Originally Posted by hilbert70

I had also freezing problems with my desktop. But I seem to have fixed it with the kernel parameter "pci=routeirq" this because I saw non carched irq in my messages and this seems to fix my freezing problem.

I tried booting with pci=routeirq. So far, I got the same flicker (aka black flash) and the following message in dmesg:
[ 1952.352866] NVRM: Xid (0001:00): 6, PE0000 0300 0491077b 0000fdcc 00eeeeec 04b00000

Usually, I see a second L1->L0 Xid message, but not this time. Curious. I will continue with compiz enabled and see if X still locks up.

[pferraro]

Originally Posted by pferraro

I tried booting with pci=routeirq. So far, I got the same flicker (aka black flash) and the following message in dmesg:
[ 1952.352866] NVRM: Xid (0001:00): 6, PE0000 0300 0491077b 0000fdcc 00eeeeec 04b00000

Usually, I see a second L1->L0 Xid message, but not this time. Curious. I will continue with compiz enabled and see if X still locks up.

No luck. X froze after a couple hours of use. Again, this is with a GeForce Go 7950 GTX. Anyone at nvidia listening...?