NVidia driver/connection problems

[alepulver]

Hello.

I have an NVidia 6600 on FreeBSD 5.4-RELEASE, Xorg 6.9 and nvidia-driver 1.0.8178 and I am having ocassional problems when running some OpenGL problems.

Sometimes in the middle of the execution the system freezes for a while (about 30 seconds) and then the application crashes (not always, sometimes it continues normally) with messages like this:

Received signal 11, exiting...
X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 135 (XFree86-VidModeExtension)
Minor opcode of failed request: 10 (XF86VidModeSwitchToMode)
Value in failed request: 0x1200002
Serial number of failed request: 35768
Current serial number in output stream: 35771

And the NVidia kernel module outputs errors like the following:

NVRM: Xid: 8, Channel 00000002
NVRM: Xid: 27, L1 -> L0

The card was previously without the AGP slot lock set, and was outputting a lot of errors (because it was randomly being disconnected and reconnected by the vibration of the cooler I think), but now the lock is on and this is the only error I notice.

The errors when the AGP lock was off were like the previous one (but much more):

NVRM: Xid: 8, Channel 00000002
NVRM: Xid: 27, L1 -> L0
NVRM: Xid: 8, Channel 00000002
NVRM: Xid: 27, L1 -> L0
pid 99360 (conftest), uid 0: exited on signal 12 (core dumped)
pid 19323 (conftest), uid 0: exited on signal 11 (core dumped)
pid 57511 (conftest), uid 0: exited on signal 12 (core dumped)
pid 68673 (conftest), uid 0: exited on signal 11 (core dumped)
nvidia0: detached
nvidia0: <GeForce 6600> mem 0xcd000000-0xcdffffff,0xb0000000-0xbfffffff,0xce000000-0xceffffff irq 16 at device 0.0 on pci1
NVRM: Xid: 8, Channel 00000002
NVRM: Xid: 27, L1 -> L0

Here you can see it was disconnected for a moment.

Do you think these errors are still about connection problems? What can I do?

If more information is needed just ask.

Thanks and Best Regards,
Ale

[zander]

Does lowering the AGP rate help?

[bentman78]

I am kind of a newbie to FBSD, how do I lower the AGP Rate? How do I get an AGP Rate? My Bus Type says PCI, am I doing something wrong?

[alepulver]

zander:
Thank you for your reply.

I have tried lowering it to 4x, but the problem persists. It only happens with the same OpenGL application: a Quake II engine modification called QuDos.

The screen/system freezes for about 20 seconds, and then the kernel module outputs this:
NVRM: Xid: 8, Channel 00000002
NVRM: Xid: 27, L1 -> L0

The application crashes with the following message:
Received signal 11, exiting...
X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 135 (XFree86-VidModeExtension)
Minor opcode of failed request: 10 (XF86VidModeSwitchToMode)
Value in failed request: 0x1400002
Serial number of failed request: 855
Current serial number in output stream: 858

What do you think about this?

bentman78:
You can view it with the sysctl command:

% sysctl hw.nvidia.agp.card.rates
hw.nvidia.agp.card.rates: 8x 4x

But you have to change it from the BIOS (you can also view it from here).

[zander]

@alepulver: the vidmode extension problem doesn't look like it's caused by the NVIDIA driver; are you loading the extmod module? As to the Xid problem, does disabling AGP alltogether help? Also, which AGP GART driver are you using?

[alepulver]

I am loading the "extmod" module (omitting the "DGA" extension).

I have disabled AGP in the BIOS, but the problem persists (I changed "Primary Graphics Adapter" from AGP to PCI),

I am using the NVidia's AGP GART driver (I am not loading agp.ko and I have no NvAGP option in xorg.conf).

Are you sure these two problems are not related, because they always happen at the same time and under the same circumstances?

[alepulver]

Sorry for the bad description I did before. Here is a clarification:

The system freezes about 20 seconds when I get that NVRM message, during that time if I try to switch to a terminal it makes the executable crash, otherwise it continues normally.