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

Newegg Daily Deals

Reply
 
Thread Tools
Old 10-31-02, 12:59 PM   #109
esrever_otua
Registered User
 
Join Date: Oct 2002
Posts: 2
Default APM Breaks drivers...

Interesting development; I had noticed that I was suffering from the same fate as several other people here; having finally started X successfull, when I tried to shutdown X the screen would blank and the machine locked up hard, and if I tried to switch to one of the virtual consoles the screen would go funky (lots of multicoloured vertical bars) and would lock up hard. Since then I have made only one change, which was to recompile my 2.4.19 kernel with ACPI instead of APM (completely disabled APM, completely enabled ACPI) and now, whilst X still takes about a minute to start, I no longer have either of the two problems mentioned above....

See my previous post for machine spec's

cheers,
esrever_otua
esrever_otua is offline   Reply With Quote
Old 10-31-02, 06:27 PM   #110
debrules
Guest
 
Posts: n/a
Unhappy mem leak?

Has anyone else noticed a memory leak in NVdriver? I have compiled and installed the driver without any problems/errors. When I leave my system up and running for one day my free mem is reduced significantly. I've tried this without running X and the NVdriver module loaded and it still happens, so it's not X. I have also tried this without loading the NVdriver module and the free mem does not reduce at all....

Config:
Dell 8100
Debian Kernel 2.4.18
P4, 1.5ghz
256mb RAM
XFree86 4.2.1

Anyone know of any patches?
  Reply With Quote
Old 10-31-02, 06:31 PM   #111
gigsvoo
Registered User
 
Join Date: Oct 2002
Posts: 4
Default

I am not sure about the memory but something things just go slow, pause then resume... so I put VeryNice to prioritize the CPU thread priority.
gigsvoo is offline   Reply With Quote
Old 11-03-02, 06:55 PM   #112
kappax
Registered User
 
Join Date: Oct 2002
Posts: 56
Default Re: mem leak?

Quote:
Originally posted by debrules
Has anyone else noticed a memory leak in NVdriver? I have compiled and installed the driver without any problems/errors. When I leave my system up and running for one day my free mem is reduced significantly. I've tried this without running X and the NVdriver module loaded and it still happens, so it's not X. I have also tried this without loading the NVdriver module and the free mem does not reduce at all....

Config:
Dell 8100
Debian Kernel 2.4.18
P4, 1.5ghz
256mb RAM
XFree86 4.2.1

Anyone know of any patches?
I may be wrong, but it is not that free mem is gone, or leaked away, but my bet is that it is cached, type top at a console and see if the number by "cached" is high, if so there is no problem. it saves stufff in ram and reuses it if needed, but if the ram is needed for someting new that is not in the ram already it will free the ram by removing some of the cached stuff out. On a normal day i have over 90% to 99% of my ram used and NOT FREE, but that does not mean that i am missing my ram. it is cached, stuff like quake3 and ut2003 and mozillla that i had played a day a go will load faster because it is in cached area of the ram. I would not worrie about it till you start getting errors about programs not running due to the lac of ram : )
kappax is offline   Reply With Quote
Old 11-04-02, 04:23 AM   #113
hageshii
Guest
 
Posts: n/a
Default Instability after a few startx's

I've noticed that if you run startx from the command-line, X starts and runs with nVidia drivers with no problems (for me). But if I kill the X session and restart X with a 'startx', then the screen flickers (like it's about to switch resolutions) and the AGP bus locks up (along with the rest of the system). While locked, thin, vertical, unevenly-spaced, different-color bars appear across the screen. I'd tried tweaking my BIOS's "AGP drive strength" settings with little luck. It usually happened after two or three startx's.

But I may have found a solution. It seems that if I do a "rmmod NVdriver" after killing the X session, X comes up with no problems. I've performed about 10 iterations of this with no crashes. So it seems like the drivers leave the AGP bus in some stale, funky state or something. Here are some stats:

AMD Athlon 900MHz
VIA VT82Cxx chipset on an MSI mainboard
geForce2 something or other from LeadTek
Driver: 1.0-3123
Kernel: 2.4.19
Dist: LFS 4.0
XFree86 4.2.0
Using Xinerama / TwinView
AGP: 4X
Manual Drive Strength: 0xA8

-dannyd
  Reply With Quote
Old 11-12-02, 09:50 AM   #114
ceddy
Registered User
 
Join Date: Nov 2002
Posts: 5
Send a message via ICQ to ceddy
Default 30 sec delay to start X too

Ok, you can add me to the 30sec delay to start X as well, but I'm NOT running on AMD.

My config is:

NV driver: 2960
Gfx card: Dimand Viper 550 PCI (TNT based)
Distro: Debian Sarge
XFree: 4.2.1
Proc: Intel PII 450

Had the same problem with some previous NV drivers, with XFree 4.0 and Debian Woody 3.0

I've not upgraded to the very latest NV driver as it usually doesn't bring anything new on those old cards. Btw, that would be nice to release the source of those old cards

Thanks
ceddy is offline   Reply With Quote
Old 11-16-02, 03:48 PM   #115
capt_nemo
Registered User
 
Join Date: Nov 2002
Posts: 6
Default 30 sec delay

I have had a ~30 sec delay since I can remember, and I am running a p3. I am useing mandrake 8.2,p3 500, 1.0-3123 from src rpms, kernel 2.4.19 compiled myself, a gf4mx-440, soyo 6ba+ mobo, xfree 4.2. I had the same problem with 2960.
capt_nemo is offline   Reply With Quote
Old 11-17-02, 10:25 PM   #116
etr
Registered User
 
Join Date: Aug 2002
Location: NC, US
Posts: 55
Default AGP and nForce

I have been unable to get the 3123 drivers to use AGP on my nForce board. I started a thread on this, but have since found other similar threads without a recorded resolution. I've banged on it a fair amount, and do not think the problem is with my configuration. (If you think otherwise, please take a look at "AGP and nForce". If you can see anything I might have missed, I am definitely game to try it.) I might also add that I have tried the 2960 drivers to resolve this problem--to no avail.

If anyone HAS gotten AGP to work with the nForce, I would love to hear about it (even if you don't have any suggestions). In particular, I would be interested in knowing whether you are using the on-board audio or an add-in card. (I am using an add-in GF4 Ti 4400.) I am wondering how many others are having problems similar to mine--at least a few, judging by other posts.
etr is offline   Reply With Quote

Old 11-22-02, 12:32 AM   #117
whammy
Guest
 
Posts: n/a
Unhappy apps segfault when using 3123 drivers

I have been working on a graphics engine (Ogre) and have been trying to debug a segfault problem that has appeared with the 3123 drivers.

The problem occurs when Ogre opens the rendering screen using SDL_Init which in turn calls on functions in the libX11 library to create a window. The segfault occurs when one of the functions tries to clear the errno return code. The interesting thing about this is that if you launch the ogre using gdb, then probe the offending line that produced the segfault, gdb reports that the memory address is perfectly valid (i.e. no access violations). I set a breakpoint at the line were the crash takes place and tried to manually set the memory location. Again, no access violations. But if you allow the code to run. It crashes every time.

Now this is were is gets weird. If you change the memory footprint by adding some NOPs to the code, the code either runs, or the problem moves somewhere else.

Since X runs just fine, I'm reasonably confident that this isn't a problem with the X11 lib. Also, if you switch back to the default nv drivers, the problems stop.

In an attempt to get more info on the problem, I put the nvidia 3123 drivers back in place and launched the app using a memory management sniffer to see if some memory was getting stomped on. It reported an unusual function call to modify_ldt() inside libGL.so.1.0.3123 which can modify a processe's memory management. So my thoughts are that there is something not kosher going on here with the memory management that's causing these phantom segfaults to occur.

Anyone got any ideas on this?

RH8.0
2.4.18-18
GeForce2 MX200 (PCI)
Xfree 4.2
  Reply With Quote
Old 11-26-02, 03:38 AM   #118
doc-holliday
Guest
 
Posts: n/a
Default Black Screen under X

Hi there,

I have sometimes a black Screen under my X, if i change from console to X.

Greetings
  Reply With Quote
Old 11-26-02, 04:33 AM   #119
McCoy
Registered User
 
Join Date: Nov 2002
Posts: 1
Default

Hi,

I was using my GeForce3 without problems (aparently) with 2880 and 2960 drivers on my box with debian woody (I use the tar.gz files in the nVIDIA homepage).

When I changed to the new 3123 everything seemed to be allright, and I followed using them normally.

However, yesterday, I was writing some code in C with OpenGL (I had been some time without doing such a thing) and problems arised!

I'm using glut library for X-Windows interaction, and the problem appeard when using the Depth test.

With a simple:
glClear (GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);

the window appears ok, but the redisplay is not good and the windows seems to "get ruined" (sorry about my english, I don't know how to explain accurately the state the window finishes in) and I got the computer completely lock up (I have to do a "hard" reset). It seems some problem with video mem or something...

If a change the line:

glClear (GL_COLOR_BUFFER_BIT);

Everything is OK (excepts the depth test, of course :-)

All the symlinks and the files in /usr/lib and /usr/X11R6/lib seems to be OK, and no errors are getting from the system.

I tryed some glut examples from the glut bundle and the problem is the same when treating with the GL_DEPTH_BUFFER_BIT.

I also tried changing the AGP options (NVAGP, AGPGART, no AGP at all, but it seems not to have got to do with it).

Finally I "downgraded" to 2960 and everything is ok with the same codes.

Could it be a bug in 3123 version?
Any ideas?

I'm sorry about my english, I hope the problem can be understood.

Cheers.
McCoy is offline   Reply With Quote
Old 11-29-02, 03:53 PM   #120
LinuxSam
Registered User
 
LinuxSam's Avatar
 
Join Date: Nov 2002
Location: Kiruna, Sweden
Posts: 5
Send a message via ICQ to LinuxSam
Default nVIdia driver 3123 and kernel 2.4.20

It seems the 3123 driver and the 2.4.20 kernel doesn't mix...

when I try to compile the nVidia kernel module I get a lot of "unresolved symbols".

/LinuxSam
LinuxSam 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
173.14.35 Linux Driver annb NVIDIA Legacy Graphics Cards 0 06-13-12 06:46 PM
Best Linux driver for Quadro FX2700M dclement NVIDIA Linux 0 05-07-12 06:58 AM
302.07 (beta) for Linux x86/x86_64 released AaronP NVIDIA Linux 0 05-02-12 09:55 AM
_which_ quadro cards have stereo support under driver 3123 codingfrenzy NVIDIA Linux 3 12-29-02 03:24 PM
download problem with linux NVIDIA driver packages Timo NVIDIA Linux 7 07-31-02 09:22 AM

All times are GMT -5. The time now is 10:32 AM.


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