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

Newegg Daily Deals

Reply
 
Thread Tools
Old 07-02-03, 09:22 AM   #1
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default ali agp still unstable

hello,

has this (ali aladdin V agp unstable) known problem been fixed yet? i've tried the 4363 drivers and a couple of previous ones too but my system is still unstable (mandrake 9.1, also tried slackware8.1). opengl games exit shortly after loading them(i've checked GLX & NV-GLX & NVIDIA-GLX are ok), back to a desktop with the virtual desktop enabled for some reason (i wasnt using it before-it uses a 800x600 screen to looks around a 1024x768 desktop(probably since game 800x600)). but also even not using an opengl application-mozilla causes my sytem to properly hang (can still move mouse but screen/desktop is locked up). i've tried the different agp modes, but no difference. does anyone know if or at what stage the resolution of this problem is at?win98 also needs a ali agp patch, so will a new kernel be needed or agpgart or what? my current kernel is 2.4.21-0.13mdk. any help/advice would be much appreciated.

system: AMD k6-2 500mhz, 192mb ram, MSI MS-5159 motherboard(probably still got the latest BIOS as no updates in years) with the horrible ali aladdin 5 M1531/M1543 chipset , geforce2mx with tv-out, dual boot with win98.

thanks.
Mr Marmalade is offline   Reply With Quote
Old 07-04-03, 08:57 AM   #2
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default

hello,
i'm trying to build the 4363 driver for my system with the ali agp bits
enabled to try and fix my problems of system (+opengl) instability. i changed the bits in the os-registry.c
file as follows:
static int NVreg_EnableALiAGP = 1; (was 0)
static int NVreg_ReqAGPRate = 1; (was 7; bit unsure of this one as it says "option expects a bitmask"?)
static int NVreg_UpdateKernelAGP = 0; (was 1)

once i save the os-registry.c file and try to 'make' it as root it spews out a load of errors(2+1/2 pages worth).
the error includes this(actually says this block twice):

/usr/include/linux/version.h:2:2: #error "================================================= ======"
/usr/include/linux/version.h:3:2: #error "You should not include /usr/include/{linux,asm}/ header"
/usr/include/linux/version.h:4:2: #error "files directly for the compilation of kernel modules."
/usr/include/linux/version.h:5:2: #error ""
/usr/include/linux/version.h:6:2: #error "glibc now uses kernel header files from a well-defined"
/usr/include/linux/version.h:7:2: #error "working kernel version (as recommended by Linus Torvalds)"
/usr/include/linux/version.h:8:2: #error "These files are glibc internal and may not match the"
/usr/include/linux/version.h:9:2: #error "currently running kernel. They should only be"
/usr/include/linux/version.h:10:2: #error "included via other system header files - user space"
/usr/include/linux/version.h:11:2: #error "programs should not directly include <linux/*.h> or"
/usr/include/linux/version.h:12:2: #error "<asm/*.h> as well."
/usr/include/linux/version.h:13:2: #error ""
/usr/include/linux/version.h:14:2: #error "To build kernel modules please do the following:"
/usr/include/linux/version.h:15:2: #error ""
/usr/include/linux/version.h:16:2: #error " o Have the kernel sources installed"
/usr/include/linux/version.h:17:2: #error ""
/usr/include/linux/version.h:18:2: #error " o Make sure that the symbolic link"
/usr/include/linux/version.h:19:2: #error " /lib/modules/`uname -r`/build exists and points to"
/usr/include/linux/version.h:20:2: #error " the matching kernel source directory"
/usr/include/linux/version.h:21:2: #error ""
/usr/include/linux/version.h:22:2: #error " o Now copy /boot/vmlinuz.version.h to"
/usr/include/linux/version.h:23:2: #error " /lib/modules/`uname -r`/build/include/linux/version.h"
/usr/include/linux/version.h:24:2: #error ""
/usr/include/linux/version.h:25:2: #error " o When compiling, make sure to use the following"
/usr/include/linux/version.h:26:2: #error " compiler option to use the correct include files:"
/usr/include/linux/version.h:27:2: #error ""
/usr/include/linux/version.h:28:2: #error " -I/lib/modules/`uname -r`/build/include"
/usr/include/linux/version.h:29:2: #error ""
/usr/include/linux/version.h:30:2: #error " instead of"
/usr/include/linux/version.h:31:2: #error ""
/usr/include/linux/version.h:32:2: #error " -I/usr/include/linux"
/usr/include/linux/version.h:33:2: #error ""
/usr/include/linux/version.h:34:2: #error " Please adjust the Makefile accordingly."
/usr/include/linux/version.h:35:2: #error "================================================= ======"

reading the following FAQ from the nvidia readme:

Q: I can't build the NVIDIA kernel module, or I can build the NVIDIA
kernel module, but modprobe/insmod fails to load the module into
my kernel. What's wrong?

A: These problems are generally caused by the build using the wrong kernel
header files (ie header files for a different kernel version than
the one you are running). The convention used to be that kernel
header files should be stored in "/usr/include/linux/", but that
is deprecated in favor of "/lib/modules/`uname -r`/build/include".
The nvidia-installer should be able to determine the location on your
system; however, if you encounter a problem you can force the build
to use certain header files by using the --kernel-include-dir option.
Obviously, for this to work, you need the appropriate kernel header
files installed on your system. Consult the documentation that came
with your distribution; some distributions don't install the kernel
header files by default, or they install headers that don't coincide
properly with the kernel you are running.

it sounds as though its looking for the headers in the wrong place, i've checked and the headers exist
in the "/usr/include/linux" directory (lots of "*.h" files that konkerer says are headers,
so i'm assuming they are headers!) and not the "/lib/modules/....." one.
there's a bit in the Makefile that mentions:

KERNDIR=/lib/modules/$(shell uname -r)
should i change this to
KERNDIR=/usr/incluse/linux

???

please someone respond/help as this ali agp (aladdin5) chipset is driving me nuts.

(for reference i'm trying to go along with the following ali help page : http://www.blutsuppe.de/freesoft.html
, although its a bit dated)
Mr Marmalade is offline   Reply With Quote
Old 07-08-03, 06:11 AM   #3
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default

ok, i've fixed the problem so i'm writing it here in case anyone else needs to know how i did it:
do as per
http://www.blutsuppe.de/freesoft.html
but works with latest 4363 drivers (os-registry is located in usr/src/nv within folder you extracted driver to). once edited the 3 bits as per the link above, save & exit. typing make then does some stuff but not enought to install the drivers. then go back to the main extracted folder and you'll see a file called 'nvidia-installer*', which won't run in the same way as the original *.run file; instead run it by typing "./nvidia-installer" without the quotation marks. and then do as it says. make sure to edit the xf86config file as per the nvidia readme (it was inside the driver, you extracted it earlier). now it all works!
note: i never fixed the problem of compiling not working in mandrake 9.1, probably due to missing kernel modules (?), so i went back to slackware 8.1 and it all went beautifully! bless slackware's sturdy cotton hiking socks.
horrah! and now quake2 works in openGL... now only got to re-sort out my network and sound system and the cd-roms and the...
__________________
My Photos: http://photobucket.com/albums/v643/visual_etc/
PC: 486sx 25MHz, 4MB RAM, 1MB VLB Gfx.
Mr Marmalade is offline   Reply With Quote
Old 12-11-03, 04:10 PM   #4
DerKostka
Registered User
 
Join Date: Dec 2003
Posts: 14
Send a message via ICQ to DerKostka
Default

Williot run under mandrake9.2 with my ti4200 and the newest nvidia driver ?
I really hope so!!!

p.s the link is dead , could you give me a new ?



EDIT:

I am sorry i have currently installed Mandrake 9.2 not 9.1!

Last edited by DerKostka; 12-11-03 at 04:59 PM.
DerKostka is offline   Reply With Quote
Old 12-11-03, 04:48 PM   #5
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default

Feels like ages since i made that post and i can't remember much about it. now i'm using slackware9.1. its got kernel 2.4.22 i think which i think mandrake9.1 may use.
i'm not sure if i ever overcame the ali agp problem since if you look in the readme for the latest nvidia drivers (which don't appear to have been updated in ages!!!) it still says that this chipset is mince:

__________________________________________________ ________________________

(app-g) APPENDIX G: ALI SPECIFIC ISSUES
__________________________________________________ ________________________

The following tips may help stabilize problematic ALI systems:

o Disable TURBO AGP MODE in the BIOS.

o When using a P5A upgrade to BIOS Revision 1002 BETA 2.

o When using 1007, 1007A or 1009 adjust the IO Recovery Time to
4 cycles.

o AGP is disabled by default on some ALi chipsets (ALi1541, ALi1647)
to work around severe system stability problems with these chipsets.
See the comments for NVreg_EnableALiAGP in os-registry.c to force
AGP on anyway.

I think i eventually gave up on this problem - i just use linux now for everything except games and scanning. maybe once the new 2.6 kernel comes along it will help this issue. then again, probably not.

But, just for you, I'll have a wee shot again. I'll try and backtrack and get back to you.

Just call me Santa.
__________________
My Photos: http://photobucket.com/albums/v643/visual_etc/
PC: 486sx 25MHz, 4MB RAM, 1MB VLB Gfx.
Mr Marmalade is offline   Reply With Quote
Old 12-11-03, 04:54 PM   #6
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default

i think i found that guide, or at least some sort of guide:

Nvidia and ALi woes

The ALi 1541 chipset is a major pain in the * with AGP cards. Apparently this problem is neither restricted to Linux or to Nvidia cards. I own a ASUS P5A-B board which has this blessed chipset. It took me over a year of frustrated crashes to get it to cooperate. Here is an explanation of the steps that were necessary for me to finally get it to a working stage. I wrote this with the absolute Linux newbie in mind, so don't be offended if you already know what ls and cd are.


To start with, I strongly encourage you to upgrade your system to:

The newest stable BIOS version. Upgrading your BIOS isn't really bad, and it doesn't involve that many risks, either. I would advise against BETA BIOS versions. In my case that didn't help at all. A detailed explanation of how to upgrade your BIOS and were to download the newest files can usually be found on the homepage of your mainboard producer. Generally it involves making a DOS boot disk and copying some files onto it. If you don't own a DOS boot disk, you might want to try FreeDOS, which can be downloaded at http://www.ibiblio.org/pub/micro/pc-...nstdisk/1.44/. There are also some instructions how to put these files on a disc. Then you'll need to make some space on the disc - just delete the docs and util directories with your favorite file manager or rm -r docs util them. I have not tested this, though. It's probably much easier to take a MSDOS or Windows 95/98/ME (the DOS-based Windows versions) boot disk/CD and press F5 when they start to boot, so it won't load with memory managers.
The latest of the stable kernel releases. If you are unhappy compiling your own, check out the FTP server of your favourite distribution, and install the newest kernel or upgrade to a newer version of your distro. Newer 2.4 kernels have bugfixes in the AGPGART code, which you will want to use.
Upgrade to the latest Nvidia drivers. Get them from www.nvidia.com. You need the GLX package for your distribution, and the source for the kernel driver (the tar file, not the RPM).There are GLX RPM packages for all major distributions. If you are RPM-based (i.e. Red Hat, SuSE, Mandrake...), just download them, change to root with the command su, and install them with rpm -Uvh NAME_OF_DOWNLOADED_RPM_FILE. Kernel driver installation is below.


Unfortunately, we cannot use the standard RPM kernel driver file. The first showstopper would be that many of you will have a custom kernel at this point. But anyway we need to tune some parameters in the registry of the kernel driver, so we must do this from source. I have uploaded my changed os-registry.c file here, but I strongly advise you to do the changes yourself. My file is against version 1.0-2960 of the NVidia kernel drivers, don't use with any other version!!


Extract the kernel driver tar file with tar -x -f NAME_OF_DOWNLOADED_TAR_FILE -z. This creates a new directory (look with ls, its something like NVIDIA_kernel-1.0-2960). Change into it with cd NAME_OF_DIRECTORY. Now we must edit the files to set the AGP speed to 1 and to enable the ALi chipset fixes. Use your favourite editor to edit the file os-registry.c. This could be the KDE or Gnome editor. I use joe, a text-based editor. If you want help on an editor, remember that you can type man PROGRAM_NAME in a shell window. At about line 100, there begins an explanation of the problems found. We definitely want AGP, so we change the parameter

U032 NVreg_EnableALiAGP = 0;
to
U032 NVreg_EnableALiAGP = 1;
Directly afterwards comes the part where they talk about the bit-mask of acceptable AGP rates. My board crashes if I use more than 1x AGP, so we'll change this. Change the preceding
#if 0
to
#if 1
or the code won't get compiled at all. Now take the
U032 NVreg_ReqAGPRate = 0x7;
and change it to
U032 NVreg_ReqAGPRate = 0x1;
The last change in this file is the authors "cover-my-tail flag". It seems the Nvidia coders did spend quite some time trying to get this stable. Once again change the preceding #if 0 to #if 1. Ok, this was it, folks. Now save your changes. Then type make to compile the driver. The resulting mode must be copied to a location were the loader will find it automatically. To find out your kernel version, type uname -r.. Now copy the file NVdriver to your module directory by typing cp NVdriver /lib/modules/KERNEL_VERSION/kernel/drivers/video.


The next step is to adapt the X configuration. Change to the /etc/X11 directory. On some systems this directory might have a slightly different name. Just type cd /etc/X11 and then press TAB (the key with those two arrows pointing in different directions on the left hand side of the keyboard). This is completion, and will complete names you type for you. Now edit the file XF86Config or XF86Config-4.x (type XF86C and use TAB to complete the name - its easy). This file is divided in sections. We are looking for the "Device" section. Make sure that Driver is set to "nvidia" in this section. Now beneath the Driver line add the following three lines:

Option "NvAGP" "3"
Option "PageFlip" "false"
Option "NoRenderAccel" "True"
This should do the trick. Make sure that all these lines are still in the "Device" section. I'm not quite sure if those two options are really necessary. Pageflipping only works on GeForce cards anyway, and I have a TNT2 Ultra. I'd like to hear from you if it works with different parameters. Make sure that all these lines are still in the "Device" section. Now save the file. If you are in the graphical environment (called X in Unix), close all programs and log out. Unfortunately, you must reboot now, because the Nvidia card must be reset to the new values. Unloading the drivers and loading the new ones might cause problems (rebooting is not common in Unix). Voila - it should work now.


I'm very interested in success stories. If you have any improvements or find errors in this document, please don't hesitate to tell me. My email address is obsidian at obsidian dot de.



On another note: I overclocked my K6-2/500 to 550 MHz two days ago by setting the front side bus to 110 MHz. Now I know this is pretty lame from an evil gamerz point of view and pretty braindamaged from a chip designers point of view
Anyway its rock stable and I really get those 10% more in Q3A. The next thing I'm going to do is experiment with NVclock and overclock my NVidia card. Maybe I can get RTCW to run somewhat smooth.
__________________
My Photos: http://photobucket.com/albums/v643/visual_etc/
PC: 486sx 25MHz, 4MB RAM, 1MB VLB Gfx.
Mr Marmalade is offline   Reply With Quote
Old 12-11-03, 05:09 PM   #7
DerKostka
Registered User
 
Join Date: Dec 2003
Posts: 14
Send a message via ICQ to DerKostka
Default

Wow, what a nice Howto ! Cool,


But i cannot try it because there is one Problem :

Quote:
Extract the kernel driver tar file with tar -x -f NAME_OF_DOWNLOADED_TAR_FILE -z. This creates a new directory (look with ls, its something like NVIDIA_kernel-1.0-2960).
There is the Beta 46.20 Version installed on my Mandrake9.2 System and i did it via the +++.run File. Now, how can i extract the *.run file ?
DerKostka is offline   Reply With Quote
Old 12-11-03, 05:11 PM   #8
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default

im doing stuff justnow with this but heres how to do that bit:

sh NVIDIA-Linux-x86-1.0-4496-pkg2.run --extract-only
__________________
My Photos: http://photobucket.com/albums/v643/visual_etc/
PC: 486sx 25MHz, 4MB RAM, 1MB VLB Gfx.
Mr Marmalade is offline   Reply With Quote

Old 12-11-03, 05:14 PM   #9
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default

what on earth is this:
"There is the Beta 46.20 Version installed on my Mandrake9.2 System"
new nvidia drivers or a kernel thing, you've lost me? im following the guide vaguely but for the 4496 drivers.
__________________
My Photos: http://photobucket.com/albums/v643/visual_etc/
PC: 486sx 25MHz, 4MB RAM, 1MB VLB Gfx.
Mr Marmalade is offline   Reply With Quote
Old 12-11-03, 05:15 PM   #10
DerKostka
Registered User
 
Join Date: Dec 2003
Posts: 14
Send a message via ICQ to DerKostka
Default

Oh, yes ! I`m sorry i just have forgotten that --extract-only. I have heard about it before, but i just startet off with Linux some Weeks ago, so now I'll try to follow the instructions....


+++
Thanks for your fast Reply !
DerKostka is offline   Reply With Quote
Old 12-11-03, 05:19 PM   #11
DerKostka
Registered User
 
Join Date: Dec 2003
Posts: 14
Send a message via ICQ to DerKostka
Default

Sorry but i don`t really get you !

I mean, i've installed that version because i thought it would fix the issue with the Ali-Mainboard, but in fact it did not, as i read in another Forum. So i'll try to follow this Howto using the actual beta driver.
DerKostka is offline   Reply With Quote
Old 12-11-03, 06:27 PM   #12
Mr Marmalade
Refrigerated User
 
Mr Marmalade's Avatar
 
Join Date: Jul 2003
Location: Scotland.
Posts: 40
Default

i think i may have got OpenGL working. here's how i did it, in 19 easy steps:

first: go into the directory where you've got the NVIDIA driver .run file.

1.
$ sh NVIDIA-Linux-x86-1.0-4496-pkg2.run --extract-only

2.
$ cd NVIDIA-Linux-x86-1.0-4496-pkg2

3.
$ cd usr

4.
$ cd src

5.
$ cd nv

6.
$ pico os-registry.c <<< or use whatever editor you like.

7.
find the following line:
static int NVreg_EnableALiAGP = 0;
and change the "0" to a "1".

8.
find the following line:
static int NVreg_ReqAGPRate = 7;
and change the "7" to a "1".
here's my understanding of what a bit-mask might be:
7=1+2+4
3=1+2
1=1
i.e. 15=1+2+4+8.
sounds plausible.

9.
find the following line:
static int NVreg_UpdateKernelAGP = 1;
and change the "1" to a "0".

10.
save & exit editor.

11.
become root user. type 'make'. type 'make install'.
bit foggy about what i did here, but i think that's it ^.

12.
$ pico /etc/inittab
such that runlevel is 3, i.e. non-X.
reboot

13.
as root enter the directiry;
NVIDIA-Linux-x86-1.0-4496-pkg2

13.
$ ./nvidia-installer

NOW JUST GO ON TO EDIT XF86CONFIG AS STATED IN README. WICH GOES SOMETHING LIKE THIS:

14.
$ pico /etc/X11/XF86Config

15.
find the following line:
Load "glx"
uncomment it to enable it (remove the # sign).

16.
find the following line:
Driver "nv"
change nv to nvidia.

17.
save & exit.

18.
load up X by typing
$ startx

19.
reedit inittab to change runlevel back to 4.
__________________
My Photos: http://photobucket.com/albums/v643/visual_etc/
PC: 486sx 25MHz, 4MB RAM, 1MB VLB Gfx.
Mr Marmalade 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
Gentoo and agp trouble john36 NVIDIA Linux 2 09-12-02 11:36 PM
I give up: WinXP + AGP + ALi eL_PuSHeR Microsoft Windows XP And Vista 4 09-07-02 06:35 PM
How much heat can a GeForce4 Ti 4600 handle? Velkator NVIDIA GeForce 7, 8, And 9 Series 4 09-03-02 05:41 PM
Ali M1541 chipset problem perun NVIDIA Linux 3 08-12-02 06:35 PM
X freezes when agp 4x is enabled jinksed NVIDIA Linux 1 08-07-02 05:56 PM

All times are GMT -5. The time now is 10:46 AM.


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