Yet Another LockUp!

[shalcrow]

Hello Andy,

Here's the info you may need.
First off it freezes anywhere in X but not at any specific time its always random, yet always gauranteed to occur.

Videocard:
Model: Aladdin TNT2
IRQ: 11
Video BIOS: 02.05.20.02.00
Card Type: AGP

Driver Version:
NVIDIA Accelerated Graphics Driver for Linux-x86' (1.0-4363)

Distro:
RedHat9.0
kernel2.4.20-9

Videocard BIOS:
Model: Aladdin TNT2
IRQ: 11
Video BIOS: 02.05.20.02.00
Card Type: AGP

Processor:
Pentium 3 - 866Mhz

RAM:
256MB SDRAM

Motherboard:
PC133 Systemboard
Socket370
M754 series

Attached is a copy of the XF86Config file. I will post a copy of the Xfree86.0.log in a new thread as it appears I can only attach one file at a time.
Any further questions, let me know.

Regards,
Simon Halcrow.

[shalcrow]

Andy,

Please find attached a copy of the X log.

Regards,
Simon Halcrow.

[shalcrow]

hey guys. its been a while since ive posted this message. and i havent received any feedback. Are any of you able ot help me in any way?
any information would be helpfull.

Thanks
Simon Halcrow.

[CooL_SPoT]

:(
I have the same problem with aladdin tnt2....
i have the problem also in MORPHIX game edition and in MANDRAKE 9.1
my card is not AGP, it's onboard...
so i guess it's a problem with the drivers...

[mmaster]

is IRQ 11 being shared w/anything else. I had strange lockups till I resolved an IRQ sharing issue... my video card and my promise ide controller were sharing an IRQ...

i'm *NOT* sure, but i think you can
cat /proc/interrupts

or

lspci -v

to list what IRQs are being shared...

you may find dmesgs to be enlightening...

dmesg |more
or
dmesg |grep Sharing

report back on your results

[shalcrow]

'dmesg |grep Sharing' returned nothing.

this is my interupts:
$ cat /proc/interrupts
CPU0
0: 1008723 XT-PIC timer
1: 6154 XT-PIC keyboard
2: 0 XT-PIC cascade
5: 1224186 XT-PIC cmpci
8: 1 XT-PIC rtc
10: 550645 XT-PIC usb-ohci
11: 772914 XT-PIC nvidia
12: 195569 XT-PIC PS/2 Mouse
14: 36954 XT-PIC ide0
15: 154620 XT-PIC ide1
NMI: 0
ERR: 0

and attached are my lspci -v results

thankyou for any further help my can provide.
Simon Halcrow

[mmaster]

it doesn't look like there is any irq sharing going on. that was my best suggestion...

you're not overclocking the card or the system are you? some systems are very flaky when overclocked.... i know i have to run mine default.


edit: more questions I needed to ask...

what version of the drivers are you using?

why use:
Option "NvAGP" "2"

i'd try changing to "3" which is load nvagp first, then agpgart needed...

when the lockups happen, what are you doing? playing games? system is just idle and becomes unresponsive?

perhaps you should disable any power management in the bios... sometimes "waking up" doesn't work as well as you'd hope...

[DrOzz]

here is my Device section in my config file....
Section "Device"
# no known options
Identifier "NVIDIA GeForce 2 MX"
#Driver "nv"
Driver "nvidia" VendorName "NVIDIA GeForce 2 MX" BoardName "NVIDIA GeForce 2 MX" Option "NoLogo" "1"
Option "RenderAccel" "0" Option "NvAgp" "1"
#BusID
EndSection
as you can see i used three different options, but that didn't fix the random lockups that i was experiencing either...i don't know if it is some freak thing that happened to me, or if it will work for others, but by default, my computer is a 1.2Ghz, when i declocked it to 900mhz i have been recieving no lockups whatsoever...i just got my uptime during my typing of this post, and i know its only 18 hours, but just to think, i was getting a lockup within 30mins of the computer being on before..
1:18pm up 18:26, 1 user, load average: 0.04, 0.18, 0.17

[mmaster]

underclocking helped you? my *wild* guess is that maybe yer having a voltage issue. can you check in the bios to see what voltage is being supplied and then check your cpu specs to see what voltage it requires? also check the memory specs. for example, you wouldn't wanna run pc100 on a pc133 bus...

also, did you buy your cpu retail or oem? if oem, i'd find some way to make ABSOLUTELY certain you didn't get a remarked chip (a 900mhz chip "remarked" and sold as if it were a 1ghz for example).

[DrOzz]

well voltages are fine, and my cpu is a retail 1.2ghz AMD T-Bird....that stuff is fine...and as of memory i won't even go there because i totally disagree with what your saying and what your getting at with it...and just for spite i use pc133 ram...

[mmaster]

wasn't trying to pick a fight or anything..

my understanding is that when you run pc100 in a 133 bus, you are, in effect, overclocking the memory and that can cause issues.

I run pc133 in my 100 mhz bus... (i.e. the reverse of the above situation - "underclocking" my ram, if you will)

as I said, it was a wild guess. I was only tryin' to help, don't get all upset on me here :)

[DrOzz]

heh i wasn't trying to pick a fight, i was just stating that i disagree...when you use pc100 on a board that supports pc133, it plain and simple runs the memory at 100...i just wanted to state that...it dont' matter...and i appreciated your commenting on my situation...but as of now i have no explanation on why my machine stablely works when i de-clock...i am sure the answer will arise sometime, and i am sure your right that it has something to do with voltages, and i am convinced it does, but thats neither here nor there, i am just happy that i am up and running, even if i had to sacrifice 300mhz...

[shalcrow]

mmaster,

I checked my bios, and totally forgot i was overclocking my cpu. anyway, i reduced it back to default. so i guess i'll wait for the inevidable lockup. (had my hopes up in other efforts! always got let down)
anyway, i will check that power management section in the bios too. from memory that should be fine.
ive changed the NvAgp to "0"
even tho in the bios it is using onboard agp?!

no doubt, i'll be writing here again this afternoon.

Simon "sceptic" Halcrow.

ps. i love nvidia right now~~~

[mmaster]

I'm off for the weekend starting friday evening. I'll try to check the board, but no promises.

It may not be your vid card causing all this. Some systems (windows or linux or whatever) just don't respond well to overclocking. It might be, though, cuz sometimes when you overclock your cpu, you overclock your agp and pci slots as well and that could cause some goofiness...

why'd you set the NvAGP to 0? 3 is the "universal" setting.

[shalcrow]

I was just trying it for a while.
ended up changing to three a short time ago.
Am going to try a movie soon to see if it will freeze.

does andy know that this could factor into all these problems? you should let him know.

Simon Halcrow

[shalcrow]

nope didnt stop the lockups.
i was just running the configure script for wine and it locked up on me.

[feeyo]

Hi,

The lock up could have something to do with the tempature of your processor but i doubt its that. what temp. is it?

And how did you install the nvidia drivers?
Did you uninstall the older or dummy drivers first??

Because i had that problem 2 when installing the nvidia drivers without removing the dummy drivers.

[shalcrow]

No, i just installed the new ones over the old.
can you tell me how to remove both?

I will reinstall the latest driver fresh.

hopefully this may solve this issue.

Simon Halcrow.

[feeyo]

Hi,
I will tell you how to uninstall both and reinstall the new one's only: (but this is on SuSE 8.1)

what you do is download the new ones to a know directory.

Reboot your machine into linux safe settings. (runlevel 3)

then login with root and do a yast.
install/remove software, search for nvidia and remove them all.
Then get out of yast, cd to the official nvidia drivers and type:

rpm -Uhv NVIDIA_GLX-1.0-4363.suse81.i386.rpm

rpm -Uhv NVIDIA_kernel-1.0-4363.suse81.i386.rpm

reboot and ready.
Dont forget this is the SuSE way what i know of. i hope you understand it a bit and do it the RH way.

[mmaster]

or try

nivdia-installer -f --update

if you are using newer drivers...

[shalcrow]

that didnt work either. i am already using the latest driver.
do you think it could be gdm causing this problem?

[mmaster]

ok then...

do this:

init 3
(changes you to cli mode - no gui)

su - (supply root password)

nvidia-installer --update -f

and see if the drivers reinstall

when yer done,
exit
(to cancel the su mode)

init 5
( to go back to gui mode)

other than that, i'm stumped...

and, for the record, I'm not convinced it's the drivers...

and, for giggles, do this:
rpm -qa |grep nvidia
to see if you installed the rpm versions...

if they are installed, you can
rpm -e RPMPACKAGENAME
where RPMPACKAGENAME are the nvidia rpms listed with the -qa line...

just be careful, and make sure to read the man pages and be sure of what you want to accomplish... we don't wanna muck up your system for you...

[feeyo]

Hi maybe try this.
this is your XF86Config:

Section "Module"

# Option "NvAGP" "3"
# Option "NoLogo" "true"
Load "dbe"
Load "extmod"
# Load "fbdevhw"
Load "glx"
Load "record"
Load "freetype"
Load "type1"
# Load "dri"
EndSection

Try to change it to this:
Section "Module"
Load "freetype"
Load "speedo"
Load "glx"
Load "dbe"
Load "extmod"
Load "type1"
EndSection

And at the "device" add this options:


BusID "1:0:0"
Driver "nvidia"
Option "NoRenderAccel"
VendorName "NVidia"
Option "NvAGP" "3"
Option "DigitalVibrance" "255"

Before you this make a backup ;)
try it out.