Arch Linux

Snostorm

Member

Registered: 2009-04-12

Posts: 18

Nvidia 195.36.15-2 and -1 both brake system

My arch 64 worked fien before this update some one please help me out I am not completely new to linux, but I can't even seem to find the right packages for the NVidia installer to work with the latest kernel so I am either stuck with my previous kernel and workinf nvidia drivers the 190.53 or no 3d video drivers.  when the 195.36.15 drivers are installed system hangs on waiting for dbus events.

Snostorm

Member

Registered: 2009-04-12

Posts: 18

Re: Nvidia 195.36.15-2 and -1 both brake system

Okay kids went away so i looked a little more,
I ditched the 195.36.15 drivers and got the system to boot then re installed the drivers and tried loadign the module and the system hangs, I found in a newbie thread whcih I knew about before thath setting vga= on the boot will get the system past the processing dbus events but it realy seems to just delay the loading of the drivers because as soon as the drivers try to load the system hangs.

Also went over to the NVIDAI site and checked that the diver I need is still supoorted whcih it is I have a 2 GFORCE 8600 GT cards

I might try removing one as others that have had this issue also had dual video cards.

there wer eno obvious reasons for the failure in the nvidia forums

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

It sounds like I have the same issue. When you say dbus events do you mean UDev events?

I tried recompiling the kernel, the drivers, option nomodeset, and seemingly 50 other things.

My post is here:
http://bbs.archlinux.org/viewtopic.php?id=96011

biplust

Member

Registered: 2009-05-03

Posts: 7

Re: Nvidia 195.36.15-2 and -1 both brake system

I can't offer help but I can confirm the same on kernel 2.6.32-ARCH and 2.6.33-zen.
Right now I downgraded both my kernel and my driver.

What kind of mainboard do you have ?, Does it have a nForce chipset ?

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

I've got an EVGA x58 3-way SLI motherboard with an Intel chipset. The previous driver (190.56 I believe) works but would require patching for 2.6.33.

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

If I pass vga=0x0369 to the kernel it gets past the "Waiting for Udev" events pretty quickly. Logging out doesn't cause a freeze either, but attempting to enter a console with Ctrl-Alt-F1 does.

biplust

Member

Registered: 2009-05-03

Posts: 7

Re: Nvidia 195.36.15-2 and -1 both brake system

SLI you say?, Well I also have a twoway SLI, maybe that's the cause I'll try pulling out one of my cards when I get home and see whats happens

Snostorm

Member

Registered: 2009-04-12

Posts: 18

Re: Nvidia 195.36.15-2 and -1 both brake system

sorry been busy with other issues I went back to nouveau driver for the time being, yes I have the EVGA nforce 790i extreame mainboard chipset, it is also SLI I have not taken off the SLI bridge yet and removed tmy second graphics card, but that seems to be a common point for this problem is the sli main board,, weather or not you actually have sli enabled in the driver.  and I did mean uevents no udev. I also found the settign vga=xxx that it got past the hang but that initializing x would hang the ssytem imeadiatly,  I did see another thread that indicated it would startx on his system but hang when switching back to a console from X11

sekenre

Member

Registered: 2009-04-14

Posts: 16

Re: Nvidia 195.36.15-2 and -1 both brake system

Hi,

I can confirm this problem also, GeForce 6200, no sli or other fancy configuration. The screen turns bright green when X tries to start and the machine hangs completely.

Only solution was to downgrade kernel to 2.6.32 and nvidia to 190.53

Bleah.

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

I don't think SLI is the issue. I've got another machine with dual 260's that's running Arch64, kernel 2.6.33, and latest nvidia drivers with no problems.

Upgrading to 2.6.33.3 didn't help. Now the system freezes on UDev uevents and the vga option doesn't help anymore :-(

Tried reverting to nvidia-190.56 patched for 2.6.33 and that fails as well. I'm now wondering if it's another issue. Any way to determine what event is causing the hang?

Edit: 2.6.33.3, not 2

Last edited by markz79 (2010-04-28 04:18:29)

abefar

Member

Registered: 2010-04-28

Posts: 14

Re: Nvidia 195.36.15-2 and -1 both brake system

The exact same thing happened for me (hanging at udev events) when I added an additional 7900 GTX to my machine, so I was pretty sure that it was a SLI issue.

However, 2.6.33.3 didn't change anything for me - booting using vga=XYZ still works, but the default 60Hz is annoying as hell to look at, and my monitor goes out of sync when exiting X.

I'm using a X58 chipset (ASUS P6T Deluxe V2) with a Core i7 920 CPU and two 7900GTX's - do any of you have similar hardware?

EDIT:
I tried adding 'nvidia' to the MODULES array in /etc/rc.conf, but then the boot sequence just hangs at "Loading modules..." instead (the step just before udev uevents).
Then I deleted it from the list and added it to MODULES in /etc/mkinitcpio.conf and built a new initcpio (using mkinitcpio -p kernel26). Now, it hangs after "loading udevd...done" - even before getting to the blue and white coloured text.
Adding the vga= parameter would probably enable me to boot again, but building the new initcpio didn't solve anything at all, so don't try it.

Again, I didn't have any problems at all with the exact same setup before adding a second video card, so I think this has something to do with SLI in one way or another. Maybe removing the SLI bridge would work, but that would be a pretty useless solution, wouldn't it?

I have even tried reinstalling Arch twice to different harddisks, without any results...

Last edited by abefar (2010-04-28 14:47:11)

biplust

Member

Registered: 2009-05-03

Posts: 7

Re: Nvidia 195.36.15-2 and -1 both brake system

Another confirmation from some ubuntu users:

https://bugs.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/548362

It would seem that that kernel > 2.6.32 and nvidia 195.36.xx
don't play well in some cases, maybe a problem to do with the i2c_core module or one of the other i2c modules ? , Does any one know if it was changed in the update from 2.6.31 to 2.6.32 ?

I'll continue looking for solutions

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

I'm going to try removing the SLI bridge. That sounds promising. My work machine with dual 260s isn't using a bridge and it's fine.

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

Removing the SLI bridge didn't help. Still hangs on waiting for UDev events. I think it actually segfaults as a call trace is displayed after a while.

Tried updating the BIOS as well but that didn't help either.

biplust

Member

Registered: 2009-05-03

Posts: 7

Re: Nvidia 195.36.15-2 and -1 both brake system

Well I came across something interesting if I ran

#nvidia-xconfig -a

I get a machine check error, can you see if you guys also get one ?, mine seems to be related to hyper-threading (maybe?), I'm posting the decoded version.
BTW, might be totally unrelated to any of your errors

CPU 1: Machine Check Exception: 5 Bank 5: b200220024080400
TSC f9fa8431fd
CPU 1 BANK 5 TSC f9fa8431fd 
TIME 1272689794 Sat May  1 07:56:34 2010
MCG status:RIPV MCIP 
MCi status:
Uncorrected error
Error enabled
Processor context corrupt
MCA: Internal Timer error
STATUS b200220024080400 MCGSTATUS 5
(Fields were incomplete)
PROCESSOR 0:10676 TIME 1272689794 SOCKET 0 APIC 2
Mo human readable MCE decoding support on this CPU type.
Run the message through 'mcelog --ascii' to decode.
CPU 1: Machine Check Exception: 5 Bank 0: b200004000000B00
TSC f9fa8131fd
HARDWARE ERROR. This is *NOT* a software problem!
Please contact your hardware vendor
CPU 1 BANK 0 TSC f9fa8131fd 
TIME 1272689794 Sat May  1 07:56:34 2010
MCG status:RIPV MCIP 
MCi status:
Uncorrected error
Error enabled
Processor context corrupt
MCA: BUS Level-0 Responed-to-request Generic Memory-access Request-no-timeout Error
BQ_DCU_READ_TYPE BQ_ERR_HARD_TYPE BQ_ERR_HARD_TYPE
timeout BINIT (ROB timeout). No micro-instruction retired for some time
STATUS b200004000000b00 MCGSTATUS 5
(Fields were incomplete)
PROCESSOR 0:10676 TIME 1272689794 SOCKET 0 APIC 2
No human readable MCE decoding support on this CPU type.
Run the message through 'mcelog --ascii' to decode.
Some CPUs didn't answer in synchronization
This is not a software prrblem!
Machine check! Processor context corrupt
kernel panic - not syncing: Fatal machine check on current CPU
Pid: 3192. comm: hald-addon-stor Tainted: P M 2.6_33-zenZ-ARCH-20100428-stable
Call Trace:
<#MC> [<ffffffff812de7b2>] ? panic+0x89/0x153
[<ffffffff812de8bc>] ? prlntk+0x40/0x4c
[<ffffffff81014260>] ? mce_panic+0x200/0x250
[<ffffffff81014bff>] ? do_machine_check+0x68f/0x940
[<ffffffff812e22ff>] ? _lock_kernel+0xbf/0x110
[<ffffffff8110bab0>] ? blkdev_open+0x0/0xc0
[<ffffffff812e275c>] ? machine_check+0x1c/0x30
[<ffffffff8110bab0)] ? blkdev_open+0x0/0xc0
[<ffffffff812e22ff>] ? _lock_kernel+0xbf/0x110
<<EOE>> [<ffffffffa0007000>] ? scsi_execute+0x130/0x170 [scsi_mod]
[<ffffffff8110b6e3>] ? __blkdev_get+0x33/0x3f0
[<ffffffff8110bab0>] ? blkdev_open+0x0/0xc0
[<ffffffff8110bb19>] ? blkdev_open+0x69/0xc0
[<ffffffff810db2c1>] ? __dentry_open+0x131/0x360
[<ffffffff810eaf98>] ? do_flip_open+0xa7B/0xdb0
[<ffffffffa0001fc2>] ? scsi_set_medium_remobal+0x82/0xb0 [scsi_mod]
[<ffffffff810666aa>] ? module_refcount+0x4a/0x60
[<ffffffffa00001c8>] ? scsi_device_put+0x58/0x60 [scsi_mod]
[<ffffffff810f3273>] ? iput+0x23/0x70
[<ffffffff8100ae15>] ? read_tsc+0X5/0x20
[<ffffffff810f61d4>] ? alloc_fd+0xe4/0x140
[<ffffffff810dca8a>] ? do_sys_open+0x6a/0x140
[<ffffffff81002ec2>] ? system_call_fastpath+0x16/Ox1b
Rebooting in 30 seconds..

correction : Has nothing to do with hyper-threading, faulty CPU ? (I doubt it but still)

Last edited by biplust (2010-05-01 07:37:08)

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

@biplust:

I downgraded to kernel26-2.6.32.10 and nvidia-190.53-1 to get my machine booting again.
Running nvidia-xconfig -a successfully produces a config file in this environment.

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

The issue could be with the VGA arbiter. Sounsd like a kernel config option resolved the issue for a couple folks:
http://www.nvnews.net/vbulletin/showthread.php?t=150212

I haven't tried it yet and don't know that I'll have the time.

biplust

Member

Registered: 2009-05-03

Posts: 7

Re: Nvidia 195.36.15-2 and -1 both brake system

markz79: I have a  Ton of papers to grade and a test so i think I'll let this go until I have some more time. maybe by then some other distro will offer a solution.

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

nVidia has posted a patch:
http://www.nvnews.net/vbulletin/showthr … ?p=2241328

It did not work for me -- still hanging at "Waiting..."

biplust

Member

Registered: 2009-05-03

Posts: 7

Re: Nvidia 195.36.15-2 and -1 both brake system

markz79, I can confirm it DOES work for me.:) (on kernle26-zen 2.6.33), so 10x.

I should say that my framebuffer is initialized through grub2 and nvidia module is loaded after login (blacklisted module).

markz79

Member

Registered: 2009-05-17

Posts: 74

Re: Nvidia 195.36.15-2 and -1 both brake system

Works for me too now. Had to disable my onboard sound, though. I still have to use the vga boot option.