NVDA(OpenGL): Channel exception!

Were you on a Dell, HP, or Compaq? The NVIDIA settlement includes Windows machines, as well: http://www.nvidiasettlement.com/affectedmodels.html

Just to follow up on my previous posts:

I found that when this problem occurred it consistently happened on the same GT 120. I called Apple and they replaced that card. I haven't had any errors since, but it's only been a couple of weeks.

Nope, this is a machine I built, not covered by the settlement 😟
Still trying to stress 10.6.3 drivers, if it fails again I'm using Windows 7 until Lion is released, hopefully it doesnt have pathetic excuse for GPU acceleration.
Awesome how they're moving computing off CPU to GPU, but can't even code proper GPU drivers, well done.

Will see if I get any details from kernel when crashing, I'm using debug=0x144
Mar 14 13:38:50 macpro kernel[0]: IOKernelDebugger: registering debugger

Hi guys,

I think I might be on the same boat here.
I have a Mac Mini with the following graphics card:
+Chipset Model:NVIDIA GeForce 320M+
Type:GPU
Bus:PCI
+VRAM (Total):128 MB+
+Vendor:NVIDIA (0x10de)+

After 3 months without any problem, I started to meet some completely random freezes where the display would show some weird artifacts.
At first the inputs are still responsive, and after a while everything freezes.
Sometime when I reboot it's back to normal, sometimes the display is still corrupted (usually if I wait a little before rebooting, which could confirm my recent findings).
I first checked the RAM, swapping the sticks, then the disk, then re-installed the system, PMU, SMC, you name it...no luck.
As the crash are pretty random it's hard to tell if the issue is solved or not, and I have to wait and hope for the best.

And yesterday, after 2 months of random crashes, sometimes everyday, sometimes not for a whole week, I found a 100% reproduction method.
I get the same freeze every time I try to import the same 10 videos from my camera, basically copying the files from the camera to the internal HD.
I tried 10 times, deleting and importing the files again, and again and again I got the freeze (between each crash I tried swapping the RAM, etc...)

After checking the crash logs I found those, for every instance of the crash:
+3/14/11 6:52:57 PM kernel NVDA(OpenGL): Channel exception! exception type = 0x8 = Fifo: Watchdog Timeout Error+
+3/14/11 6:52:57 PM kernel 0000006e+
+3/14/11 6:52:57 PM kernel 00000010 00008697 00000474 00000000+
+3/14/11 6:52:57 PM kernel 0000047e 00001528 00000000 0000001b+
+3/14/11 6:52:57 PM kernel 00000000 00000000 01900501+
+3/14/11 6:52:57 PM kernel 00000000 00000000+
+3/14/11 6:52:57 PM kernel NVDA(OpenGL): Channel exception! exception type = 0xd = GR: SW Notify Error+
+3/14/11 6:52:57 PM kernel 0000006e+
+3/14/11 6:52:57 PM kernel 00200000 00008697 00000474 00000010+
+3/14/11 6:52:57 PM kernel 0000047e 00001408 00000001 0000000a+
+3/14/11 6:52:57 PM kernel 00000000 00000000 00000403+
+3/14/11 6:52:57 PM kernel 00000010 00000000+
+3/14/11 6:52:58 PM kernel NVDA(OpenGL): Channel exception! exception type = 0xd = GR: SW Notify Error+
+3/14/11 6:52:58 PM kernel 0000006e+
+3/14/11 6:52:58 PM kernel 00200000 00008697 00000474 00000010+
+3/14/11 6:52:58 PM kernel 0000047e 00001408 00000001 0000000a+
+3/14/11 6:52:58 PM kernel 00000000 00000000 00000413+
+3/14/11 6:52:58 PM kernel 00000010 00000000+

And finally found this thread!!

*Anyway, here is the good news (well sort of):*
Now that I had a 100% reproduction method I was able to test workarounds or fixes.
I first tried to bump up the fan RPM from 1700 to 3000, once the CPU temperature reaches 60°C, after a few seconds the freeze occurs.
Only now, instead of copying successfully 3 files I get 6 before the crash.
I then tried Fan Control and setup a curve to avoid reaching 60°C, and guess what, with a max temperature of 56°C, all 10 files copy without a glitch.
I have only used the Mini for 1 day since the workaround is in place, but no freeze so far.

I am pretty sure the graphics card should not have any problem operating at 60°C.
Now I am not sure what to do, my Mac is still covered by the warranty, but it might be difficult to have Apple admit that there is a problem (even with my logs, and workaround explanation).
On the other hand I am not so thrilled about waiting for the graphics card to deteriorate and die.

Something else I am wondering is, if it's normal for the internal temperature to go from a 40°C to a 60°C in a matter of seconds, just by copying files around?

Anyway sorry for the lengthy post, thought I told my story in case it could help others, any comment or advice is welcome!!

pep.

Welcome aboard, lol, do you mean CPU or GPU? CPU temp should not shoot up 20C if youre just copying/moving files, as for the GPU 60C does not sound too high for me, I don't know everyones max thermal spec for their GPU, but mine (8800 GTX G80) has been tested beyond 100C, and it's fine, as for OS X, GPU temps sit around 55-70C, there may be more than a single cause for all these errors.

Still no luck crashing the vanilla 10.6.3 kexts here, been running 2 x world of warcrafts running at once, on Ultra settings, still too early to determine stability.

null

It was the CPU temperature. I have tried several monitoring apps (smcFanControl, Temp Monitor, etc...) and the CPU temperature was showing a jump from 40°C to 60°C every time.

I'll check again tonight what's the GPU temperature like (I could not see it with the apps I tried). Maybe something (HDD?) is overheating and causing the graphics card to fail...

Temp it's fine in my system (35ºC) and my vga it's at 40ºC (Geforce 250GTS, with Thermaltake DuOrb), max (50ºC), running latest 10.6.6 drivers from NVIDIA.

17-03-11 14:57:47 kernel NVDA(OpenGL): Channel exception! exception type = 0x6 = Fifo: Parse Error
17-03-11 14:57:47 kernel 00000069

It's a software problem, because of the "abandon" of nvidia drivers.

(I'm from Chile xd)

10.6.7 j869 partially solved this problem, the error occurs for 3 seconds, but then you can keep playing or etc.

Now the log shows

19-03-11 16:37:28 kernel NVDA(OpenGL): Channel exception! exception type = 0xd = GR: SW Notify Error
19-03-11 16:37:28 kernel 0000006e
19-03-11 16:37:28 kernel 00200000 00008297 00000474 00000010
19-03-11 16:37:28 kernel 0000047e 00001538 00000001 0000000c
19-03-11 16:37:28 kernel 00000000 00000000 00000401
19-03-11 16:37:28 kernel 00000010 00000000

But I haven't to restart my computer.

Nope, 10.6.7 its worse for nvidia, after a restart the error occurs more frequently 😟

Mac OS X Snow Leopard, the best OS??? NO

After a week of use with the Fan Control active, I can report I did not meet another freeze.

Ah, I came here to see if 10.6.7 would be an improvement for this ongoing problem. I already had a sinking feeling because the changelog doesn't mention anything to that effect.

It is truly pathetic for a company of this size and financial and technological clout to ignore this problem affecting so many (and so many different) users and configurations.

Will be sticking with 10.6.5 then, because it is more or less manageable. I know which programs trigger it on my system (Unison for instance) and minimize using those or stopped use them at all - like with the new iPhoto - 50 euros wasted there.

There used to be a time when Apple took pride in their products, it seems pleasing customers after the purchase is no longer of any importance.

Hooray! I have just upgraded to 10.6.7 and it appears to have fixed the problem. Previously iPhoto 11 was unusable and I would get the freeze (and this error) just by clicking on a photo to view it after a few seconds. Have been running iPhoto solid for last 1/2 hour and no freezes or errors!

Hopefully this is the end of this very annoying problem!

Hi there!

I know I'm just adding to the bucket, but I have the MacBook Pro 15" i7 2.66 (2010 model), with the NVIDIA GeForce GT 300M chipset running 10.6.6 (Snow Leopard). I've already reported this to Apple through the Feedback link supplied earlier, but didn't even get an acknowledgment.

@Jasx86, can you please let us know if that resolves the issue? Am also getting these annoying NVDA(OpenGL): Channel exception! exception type = 0xd = GR: SW Notify Error and the machine crashes.

Skype crashes the machine a lot - presumably it uses some form of OpenGL processing for the eye candy and I can generally get the machine to crash quite quickly using the latest version of Skype v5.0.

Cheers!

I wasn't able to fix the problem entirely with any versions of the graphics kexts, from nvidia or apple, so for the past week I've just been using Windows 7 x64, been gaming hard for a week and not one single issue from the graphics card.

In other news 10.6.7 is now released, someone should try installing it and see if it's finally fixed (not counting on it myself)

Just updated to 10.6.7 less than 2 hours ago, and already have produced hard lockup


Mar 22 17:23:20 macpro kernel[0]: (default pager): [KERNEL]: Switching ON Emergency paging segment
Mar 22 17:23:27 macpro kernel[0]: (default pager): [KERNEL]: Recovered emergency paging segment
Mar 22 19:04:10 macpro kernel[0]: (default pager): [KERNEL]: Switching ON Emergency paging segment
Mar 22 19:04:11 macpro kernel[0]: (default pager): [KERNEL]: Recovered emergency paging segment
Mar 22 19:09:57 macpro kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0xd = GR: SW Notify Error
Mar 22 19:09:57 macpro kernel[0]: 0000006e
Mar 22 19:09:57 macpro kernel[0]: 00200000 00005097 00000654 00000100
Mar 22 19:09:57 macpro kernel[0]: 0000065e 00001458 00000006 0000000e
Mar 22 19:09:57 macpro kernel[0]: 00000000 00000000 00000002
Mar 22 19:09:57 macpro kernel[0]: 00000100 00000000
Mar 22 19:09:57 macpro kernel[0]: NVDA(OpenGL): Channel exception! exception type = 0xd = GR: SW Notify Error
Mar 22 19:09:57 macpro kernel[0]: 0000006e
Mar 22 19:09:57 macpro kernel[0]: 00200000 00005097 00000654 00000010
Mar 22 19:09:57 macpro kernel[0]: 0000065e 00001458 00024c01 0000000e
Mar 22 19:09:57 macpro kernel[0]: 00000000 00000000 00000013
Mar 22 19:09:57 macpro kernel[0]: 00000010 00000000
Mar 22 19:10:18 macpro kernel[0]: NVDA(OpenGL): Channel timeout!

retrieved kernel log via ssh, unsure what the messages above timeout are.
you guys may aswell just grab a copy of Windows 7 x64 and run it, at least its not going to crap on you every few hours, apple/nvidia are never going to fix this.