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

Newegg Daily Deals

Reply
 
Thread Tools
Old 09-13-06, 08:52 AM   #1
rbw
Registered User
 
Join Date: Sep 2006
Posts: 6
Default Kernel/Module Crash gentoo

Hi,

From a while now, (unfortunately I cannot precisely identify when this started) I can't get the nvidia drivers packaged for gentoo working.
I tried *A LOT* of kenrnels (gentoo-sources, vanilla-sources, hardened-sources, mm-sources), and they all fail more or less with the
error shown in the dmesg section attached.
I'm now running:

Linux m2 2.6.18-rc6-mm2 #2 SMP PREEMPT Tue Sep 12 17:46:34
GMT+3 2006 i686 AMD Athlon(tm) Processor GNU/Linux

But just because this is the last one I tried.

The kernel command line is:

kernel (hd0,1)/boot/kernel-genkernel-x86-2.6.18-rc6-mm2 rw root=/dev/ram0 real_root=/dev/evms/ROOT init=/linuxrc ramdisk_size=8192 doevms2 dodmraid mttr video=vesafb:ywrap acpi=noirq pci=usepirqmask

I tried *MANY* variations of the kernel configuration, none works.

Is there anything else I can do to track down the problem? Honestly,
I'm beginning to think this could be a hardware problem.
Attached Files
File Type: txt msg.txt (3.3 KB, 134 views)
rbw is offline   Reply With Quote
Old 09-13-06, 11:58 AM   #2
netllama
NVIDIA Corporation
 
Join Date: Dec 2004
Posts: 8,763
Default Re: Kernel/Module Crash gentoo

How can this problem be reproduced?

If this also reproduces with 1.0-8774, please generate and attach an nvidia-bug-report.log.

Thanks,
Lonni
netllama is offline   Reply With Quote
Old 09-13-06, 01:06 PM   #3
energyman76b
Registered User
 
Join Date: Dec 2002
Location: Clausthal/Germany
Posts: 1,104
Default Re: Kernel/Module Crash gentoo

Hi,

maybe it helps if you stop using genkernel.

make menuconfig - and only build into the kernel what you really need. Same is true for modules.
Oh, and don't build any kernels with nvidia-framebuffer stuff enabled.
energyman76b is offline   Reply With Quote
Old 09-13-06, 02:13 PM   #4
jcat
Registered User
 
Join Date: May 2006
Posts: 8
Default Re: Kernel/Module Crash gentoo

Quote:
Originally Posted by energyman76b
Hi,

maybe it helps if you stop using genkernel.

make menuconfig - and only build into the kernel what you really need. Same is true for modules.
Oh, and don't build any kernels with nvidia-framebuffer stuff enabled.
You can still use genkernel, just use:

genkernel --menuconfig all

Then at least it still automates the proceedure. I use a modified config from the gentoo live CD with no probs.

I can post my config if it helps..



Cheers,
jcat
jcat is offline   Reply With Quote
Old 09-13-06, 02:20 PM   #5
rbw
Registered User
 
Join Date: Sep 2006
Posts: 6
Default Re: Kernel/Module Crash gentoo

netllama:
It's very easy to reproduce:

startx

I cannot run X *at all* with nvidia-drivers. It always crashes before
displaying anything, and I end up with a black screen. At least recently
I discovered that Alt-sysreq-K allowed me to get back to the console,
thus avoiding a hard reset.

Attaching the output of nvidia-bug-report.sh is easier said than done
It freezes, ie, becomes stuck waiting on "disk" (the process appears
with a "D" on the "STAT" column of "top") when it tries to cat the contents of:

/proc/driver/nvidia/agp/status

The same thing happens when I try to cat this file by "hand", maybe
this is an indication of something? (I will try to do this again, without
trying to run X, perhaps the driver is already wedged beyond hope,
but only after finishing this post

I edited the script (great! I bet things were hard enough without
smart asses editing *bug report scripts*, but there you go) to change the line that used append_glob on that directory to append the files in the directory individually, skipping "status", but then the script got stuck again when reading

/proc/driver/nvidia/cards/0

Then I gave up

I will try again after a reboot and post the result.

Thanks for the interest!
rbw is offline   Reply With Quote
Old 09-13-06, 02:30 PM   #6
rbw
Registered User
 
Join Date: Sep 2006
Posts: 6
Default Re: Kernel/Module Crash gentoo

Quote:
Originally Posted by energyman76b
Hi,

maybe it helps if you stop using genkernel.

make menuconfig - and only build into the kernel what you really need. Same is true for modules.
Oh, and don't build any kernels with nvidia-framebuffer stuff enabled.
Hi,

I almost always use

genkernel --oldconfig --loads --of --other --stuff

and run menuconfig by hand or make oldconfig with the .config
of a previous version.

I also run pretty lean kernels, with just about the least drivers I can,
so I don't think this is it.

Unfortunately, I just checked and verified I'm (trying to use) using
1.0.8762 drivers, which are not the latest version it seems.

I'll try again with 1.0.8774 to see what happens.
Thanks for the interest!
rbw is offline   Reply With Quote
Old 09-13-06, 03:34 PM   #7
rbw
Registered User
 
Join Date: Sep 2006
Posts: 6
Default Re: Kernel/Module Crash gentoo

Ok, now I tried with the latest version available on gentoo
(1.0.8774) and had the same results, only this time harder to get
back to the console (had to type *a lot* of sysreqs.) In a perverted
way this is good news since then I was not moronicaly complaining
about resolved problems

I had a similar oops, attached, and also could not run
nvidia-bur-report.sh without a reboot. I did reboot
and run it, and the result is also attached. I perused
this file and it reports a failed driver instalation with the
curses interface, do ignore it, I'm installing the driver
emerging the gentoo package (I tried some time
ago to install "by hand" but had similar results).

Is it possible for the card to fail in the "3d part" and still
work in 2d? I'm *really* suspecting this is a hardware
problem, but it's working fine in with the nv driver (non-3D-accelerated, obviously).
Attached Files
File Type: txt oops.txt (17.7 KB, 119 views)
File Type: log nvidia-bug-report.log (62.9 KB, 103 views)
rbw is offline   Reply With Quote
Old 09-13-06, 03:41 PM   #8
netllama
NVIDIA Corporation
 
Join Date: Dec 2004
Posts: 8,763
Default Re: Kernel/Module Crash gentoo

I have a few questions:
0) If you boot with the noapic and/or acpi=off kernel parameters (instead of acpi=noirq) does that have any impact on this problem?
1) Does this reproduce if you have no frambuffer console graphics support in your kernel (remove vesafb,etc)?
2) Does the problem persist if you set RenderAccel to false and NvAGP to 0 in xorg.conf ?
3) Have you verified that you're using the latest BIOS for the motherboard?

Thanks,
Lonni
netllama is offline   Reply With Quote

Old 09-13-06, 05:37 PM   #9
rbw
Registered User
 
Join Date: Sep 2006
Posts: 6
Default Re: Kernel/Module Crash gentoo

Quote:
Originally Posted by netllama
I have a few questions:
0) If you boot with the noapic and/or acpi=off kernel parameters (instead of acpi=noirq) does that have any impact on this problem?
1) Does this reproduce if you have no frambuffer console graphics support in your kernel (remove vesafb,etc)?
2) Does the problem persist if you set RenderAccel to false and NvAGP to 0 in xorg.conf ?
3) Have you verified that you're using the latest BIOS for the motherboard?

Thanks,
Lonni
0) No, same(ish) crash (I didn't check the stack trace line by line though.)
1) Yes.
2) Yes.
3) Boy, that was a big one. This is an old clunker (Asus A7V-133), but still I managed to find a new bios, (from 2002 , flashed it (not that simple when you don't have windows or dos on your machine), rebooted, the machine came back up (much to my surprise and happiness,) however, the problem persisted, it still crashes.

Considering this is such an old machine, perhaps it's not worth investing too much of your time on this. If it was something stoopid I was doing, ok, but this is becoming a research problem
It's not that I don't mind my machine not havind 3D graphics, I do mind, particularly since this had worked for years, but I don't like to waste precious resources (smart knowledgeable people's time) in something that in the end will help just one guy with an old machine.
If you don't see something obvious in the stack traces and the bug report, perhaps it's best to give this rest, and maybe I'll finally be motivated to upgrade my machine. It *has* to be some kind of hardware *thing*, otherwise someone else on the whole wide world would have similar issues, but they don't.
rbw is offline   Reply With Quote
Old 09-13-06, 05:41 PM   #10
netllama
NVIDIA Corporation
 
Join Date: Dec 2004
Posts: 8,763
Default Re: Kernel/Module Crash gentoo

I'm not convinced that this is a hardware problem (although it might be). Its just that it seems to be a problem isolated to your environment. The fact that you're running Gentoo just adds more variables to the equation since replicating your environment here would be difficult. If you've got the time & resources, it might not be a bad idea to install some other 'off the shelf' distribution (Debian, SuSE, Redhat, etc) to see whether the problem persists. At least that would help to determine whether its something in your Gentoo environment that is involved.

This could still be a SBIOS or VBIOS bug.

Thanks,
Lonni
netllama is offline   Reply With Quote
Old 09-15-06, 01:34 PM   #11
rbw
Registered User
 
Join Date: Sep 2006
Posts: 6
Default Re: Kernel/Module Crash gentoo

It's such a sad thing to be stoopid...
Just now, reading your last messsage (I was offline for a couple of days) I realized that I can check very easily whether this is a hardware problem or related to my particular environment by booting a livecd distro such as KNOPPIX, some of these have 3D drivers. If it works, than it is software! I could have done this ages ago...
Actually nstalling another system is not something I'm looking forward to, since this is my main workhorse computer, and I'd have to mess about with it more than I think is safe (it terms of time needed to get it back in working order in case something went wrong) at my current schedule.

You are quite right about gentoo, each system is quite unique (that's why I love it ). That anything at all works *at all* is a flipping miracle

So, that's why I'm not complaining, you pays your money...

And finally, no, thank YOU, for the effort!
rbw is offline   Reply With Quote
Old 09-16-06, 08:25 AM   #12
zarq
Registered User
 
Join Date: Mar 2003
Posts: 1
Default Re: Kernel/Module Crash gentoo

I have a similar problem. Also running Gentoo, and have tried with various combinations of kernels and driver versions. Attempting to start X gives a black screen, and Xorg.0.log ends right before the line that typically identifies the chipset, ie. the last lines I see are:

Code:
(**) NVIDIA(0): Depth 24, (--) framebuffer bpp 32
(==) NVIDIA(0): RGB weight 888
(==) NVIDIA(0): Default visual is TrueColor
(==) NVIDIA(0): Using gamma correction (1.0, 1.0, 1.0)
(**) NVIDIA(0): Enabling RENDER acceleration
Disabling render acceleration makes no difference, neither does disabling the render extension completely. If I remove the glx module from the xorg.conf, then I will also get a few more lines in the log complaining about that, which leads me to think it's not related to the render acceleration. From a bit of searching it looks like the next line in the Xorg.0.log should identify the chipset.

By the way, sshing to the box and doing a ps at this point shows the X process in brackets, [X], and supposedly using no memory at all. It can't be killed even with kill -9.

My hardware is quite different. It's a brand new Asus M2N-MX motherboard with an onboard GeForce 6100. It has an nForce 430 chipset and an AMD 64 3200+ installed, and I'm using 64bit Gentoo (amd64 platform) and the 64bit Nvidia drivers.

Works fine in XP, by the way.

My next step is to give up and buy an external card instead, disabling the onboard one. Kinda hoping not to have to do that, though.

Ru
zarq 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


Similar Threads
Thread Thread Starter Forum Replies Last Post
UEFI + nvidia blob = nvidia blob crash puchuu NVIDIA Linux 11 05-12-12 03:14 PM
gf 6150se crash! oddhornedant NVIDIA Linux 2 05-05-12 06:30 AM
xvideo crash elanthis NVIDIA Linux 5 10-29-02 08:12 PM
Crash when logout from X TGL NVIDIA Linux 10 09-13-02 09:22 PM
Gentoo and agp trouble john36 NVIDIA Linux 2 09-12-02 11:36 PM

All times are GMT -5. The time now is 10:27 AM.


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