Go Back   nV News Forums > Linux Support Forums > NVIDIA Linux

Newegg Daily Deals

Reply
 
Thread Tools
Old 12-06-05, 07:39 AM   #1
hklygre
Registered User
 
Join Date: Oct 2005
Posts: 16
Default Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

Hi,

I have never gotten X to run properly (i.e. longer than 3-5 minutes) using two Quadro 4500 PCI-E cards in a HP Workstation XW9300 (dual opteron dual-core 280, 16 GB RAM).

The system hangs after a few minutes - it stutters for a couple of seconds, then stops completely. At least once I got several NVRM-messages during the stuttering, and then an Oops - by then the system was completely dead.

Oops captured using serial console.

Option irqpoll (as suggested in error message) causes the computer to loop endlessly in nv_sata during boot.


- Håvard
Attached Files
File Type: txt hpws-oops.txt (3.7 KB, 117 views)
File Type: zip nvidia-bug-report.zip (21.1 KB, 132 views)
hklygre is offline   Reply With Quote
Old 12-06-05, 07:43 AM   #2
zander
NVIDIA Corporation
 
zander's Avatar
 
Join Date: Aug 2002
Posts: 3,740
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

Is this a regression with 1.0-8174 or a pre-existing problem?
zander is offline   Reply With Quote
Old 12-06-05, 08:11 AM   #3
hklygre
Registered User
 
Join Date: Oct 2005
Posts: 16
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

Yes and/or no :-)

I've had several problems with this system. The main problem turned out to be unrelated to nvidia (see http://www.nvnews.net/vbulletin/showthread.php?p=753861 - i needed to boot with pci=noacpi for both PCI-E bridges to be seen).

I have gotten a similar system (single core opterons, dual quadro 4300s, 8 GB ram, scsi-disks, but still a HP XW9300 so same chipset) to work with the 76.76 driver - but I didn't get stereo on both screens (see http://www.nvnews.net/vbulletin/showthread.php?t=58855 ). I don't have that system available for testing any more.

In other words - it _might_ work properly under 76.76, but as we don't get stereo it doesn't help us. If you're interested I could test it anyway. With 81.74 I don't need to run OpenGL apps for the system to crash - last time it crashed with only the gnome desktop and one gnome-terminal running less.

- Håvard
hklygre is offline   Reply With Quote
Old 12-06-05, 03:23 PM   #4
netllama
NVIDIA Corporation
 
Join Date: Dec 2004
Posts: 8,763
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

The 1.21 BIOS has alot of known problems that can result in insability. Do these problems reproduce if you update to the 1.29 BIOS for the xw9300?

Thanks,
Lonni
netllama is offline   Reply With Quote
Old 12-07-05, 03:52 AM   #5
hklygre
Registered User
 
Join Date: Oct 2005
Posts: 16
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

Hi,

I've upgraded to 1.29 rev a now, and while things are looking better they are still not OK. I can run the system for quite a while now - it doesn't crash in "regular desktop usage" anymore. The quadro cards are on separate IRQs now (the new bios allows me to change an ACPI-setting which helps).

I'm able to run ivview (from OpenInventor) with the provided sample models for as long as I like, in stereo, on both screens, for however long I like.

Trying to run our (in-house) performer-based application however, causes an instant crash. Sometimes without logging anything (serial console), sometimes with a trace. This application runs perfectly fine on other systems here - all Debian Sarge based, all nVidia (Quadro 3400, GeForce 6800 Ultra, GeForce 7800 GTX - only the first on two screens (TwinView) and in stereo).

Kernel log and nvidia bug report attached.
Attached Files
File Type: zip nvidia-bug-report.zip (21.2 KB, 111 views)
File Type: txt hpws-oops-2.txt (1.7 KB, 114 views)
hklygre is offline   Reply With Quote
Old 12-07-05, 12:22 PM   #6
netllama
NVIDIA Corporation
 
Join Date: Dec 2004
Posts: 8,763
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

Is there a reason why you're booting with pci=noacpi? Does the crash reproduce if you do not boot with that option?

If you can provide your performer-based application, along with reproduction instructions, I'll see if I can reproduce the problem on the xw9300 that I have here. If you're concerned about confidentiality of your application, please email linux-bugs@nvidia.com referencing this forum thread, and we discuss this further via email.

Thanks,
Lonni
netllama is offline   Reply With Quote
Old 12-12-05, 02:23 AM   #7
hklygre
Registered User
 
Join Date: Oct 2005
Posts: 16
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

Sorry about the late answer - I've been doing quite a bit of testing here (as well as the work I'm supposed to be doing).

Quote:
Originally Posted by netllama
Is there a reason why you're booting with pci=noacpi? Does the crash reproduce if you do not boot with that option?
I'm not booting with pci=noacpi anymore - after upgrading the BIOS, that option is not necessary. (I needed it before to find both graphics cards).

Quote:
Originally Posted by netllama
If you can provide your performer-based application, along with reproduction instructions, I'll see if I can reproduce the problem on the xw9300 that I have here.
One annoying problem that manifests in performer is that glXGetVisualFromFBConfigSGIX doesn't work. This problem is easily tested - when perfly (performer demo application) is started the DRAW process dies with a segmentation fault. This problem can be resolved by supplying a GLX Visual ID to performer (in perfly it's the -t parameter).

Another problem is that I cannot get performer (perfly) to operate properly in multihead mode (no Xinerama) - there doesn't seem to be any acceleration on the second card, and as performer synchronises all outputs that slows things to a crawl. Performance with Xinerama seems to be OK however, so that may be the solution. (We're used to SGI here, so multihead seems most natural).

There are indications that the problem with hard lockups may be hardware related - I'm going to do some more testing here to try and narrow the problem down. I installed windows on the computer to test graphics there, and get consistent crashes in some games. I'll do some swapping of cards and see what happens.

On a related note - does nVidia have an OpenGL capture tool that works in Linux? I'm trying to recreate the problem in a smaller application than the one we're developing, but I'm having no luck so far.

- Håvard
hklygre is offline   Reply With Quote
Old 12-12-05, 10:38 AM   #8
netllama
NVIDIA Corporation
 
Join Date: Dec 2004
Posts: 8,763
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

The perfly problems with 1.0-8174 and default visuals is known bug 194926.

Thanks,
Lonni
netllama is offline   Reply With Quote

Old 07-11-06, 07:37 AM   #9
macfrag
Registered User
 
Join Date: Jul 2006
Posts: 1
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

Quote:
Originally Posted by netllama
The perfly problems with 1.0-8174 and default visuals is known bug 194926.

Thanks,
Lonni
Hello,

could you please provide some more information about this bug?
I keep having perfly crashing with driver version>1.0-8178.
I was forced to downgrade to 1.0-7676 in order for perfly to work.

I looked around but I did not found much information

thank you for your time.
macfrag is offline   Reply With Quote
Old 07-11-06, 10:09 AM   #10
netllama
NVIDIA Corporation
 
Join Date: Dec 2004
Posts: 8,763
Default Re: Dual Quadro 4500 and driver 8174: Crash. (irq5: nobody cared)

This bug should be resolved in the next driver release.

Thanks,
Lonni
netllama is offline   Reply With Quote
Reply


Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 07:03 PM.


Powered by vBulletin® Version 3.7.1
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Copyright ©1998 - 2014, nV News.