LinuxQuestions.org - Radeon 9000 3D Acceleration

- Slackware (https://www.linuxquestions.org/questions/slackware-14/)

- - Radeon 9000 3D Acceleration (https://www.linuxquestions.org/questions/slackware-14/radeon-9000-3d-acceleration-113860/)

Radeon 9000 3D Acceleration

I am somewhat of a noob to linux. I had Mandrake 9.1 but decided to switch to a somewhat more complicated distro Slackware 9.1(learning purposes). I have a Radeon 9000 Pro 128mb and Slackware is using the standard vesa driver for the card so I don't have an 3D Acceleration, but it is strange as I can use some OpenGL things! I don't really care about some of them as I just want to be able to play Wolfenstein Enemy Territory. ATI Linux drivers yeah right install those and your X server doesnt like to start I know I tried with Mandrake. I can't even install the rpm with the drivers in the first place with Slack someone wanna help me?

rpm -ivh --force <dirver's name> would be the best way

please dont use rpms they just suck ... there is a driver built by xfree86 for you to use.

just edit /etc/X11/XF86Config

Section "Device"
Identifier "VESA Framebuffer"
Driver "vesa"
#VideoRam 4096
# Insert Clocks lines here if appropriate
EndSection

change the line that says Driver "vesa"
so it looks like Driver "radeon"

This is a generic radeon driver but you can get a better one from the ati web site ... if you are trying to use dri you should read the howto.

you will probably also want GLcore and glx loading ... i think they are at the top of the XF86Config file by default in the modules section.

Yeah i'm kind having the same problem, am finding the xfree86 one a bit sluggish

Has any1 had luck with installing the ati driver for Radeon 9000 in slack?

Definitely. I've done it 10 times over.
The advice above is all wrong for the 9000 Pro if you want 3d acceleration.(same card I've got.)

If you had searched, you'll find a lot of info about this and a lot of it posted by me.. heheh

Alright, if you've got slack 9.1, it installs the kernel source be default. This is good, you'll need it. For slack 9.0 you'll need to get the kernel source first.
Here are the general steps
As root:

Code:

rpm2tgz fglrx<whatever>.rpm

installpkg fglrx<whatever>.tgz

cd /lib/modules/fglrx/build_mod/

./make.sh

cd ..

./make_install.sh

Bam. Drivers installed.

Now, change the driver line listed above from vesa (or radeon) to fglrx, make sure the default color depth is set to 24, and make sure this section exists (it's at the end in mine) :

Code:

Section "DRI"

        Group 0

        Mode 0666

EndSection

That should be it buddy :)

-Shade

Edit to add:
See you in Enemy Territory ;) My handle is Shade there, too.

forgot to mention that i have a nforce2 mobo which conflict with an Ati based video card (windoz no prob's but in linux not unless its an a nvidia based vid) wich i think is the major prob

Have seen a few good posts and am looking into it

will let ya know how i get on

Shade post you configuration file as my Enemy Territory won't run now. It runs then goes to black screen then back to the desktop and hour glasses forever.

My /etc/X11/XF86Config file:

Code:

# File generated by xf86config.

# Hacked to bits by Mike Shade

 

# **********************************************************************

# Module section -- this  section  is used to specify

# which dynamically loadable modules to load.

# **********************************************************************

 

Section "Module"

 

# This loads the DBE extension module.

 

    Load        "dbe"   # Double buffer extension

 

# This loads the miscellaneous extensions module, and disables

# initialisation of the XFree86-DGA extension within that module.

    SubSection  "extmod"

      Option    "omit xfree86-dga"   # don't initialise the DGA extension

    EndSubSection

 

# This loads the font modules

    Load        "type1"

    Load        "freetype"

    Load        "speedo"

 

# This loads the GLX module

    Load       "glx"

    Load       "dri"

EndSection

 

# **********************************************************************

# Files section.  This allows default font and rgb paths to be set

# **********************************************************************

 

Section "Files"

 

# The location of the RGB database.  Note, this is the name of the

# file minus the extension (like ".txt" or ".db").  There is normally

# no need to change the default.

 

    RgbPath     "/usr/X11R6/lib/X11/rgb"

 

# Multiple FontPath entries are allowed (which are concatenated together),

# as well as specifying multiple comma-separated entries in one FontPath

# command (or a combination of both methods)

#

# If you don't have a floating point coprocessor and emacs, Mosaic or other

# programs take long to start up, try moving the Type1 and Speedo directory

# to the end of this list (or comment them out).

#

 

    FontPath   "/usr/X11R6/lib/X11/fonts/local/"

    FontPath   "/usr/X11R6/lib/X11/fonts/misc/"

    FontPath   "/usr/X11R6/lib/X11/fonts/75dpi/:unscaled"

    FontPath   "/usr/X11R6/lib/X11/fonts/100dpi/:unscaled"

    FontPath   "/usr/X11R6/lib/X11/fonts/Type1/"

    FontPath   "/usr/X11R6/lib/X11/fonts/CID/"

    FontPath   "/usr/X11R6/lib/X11/fonts/Speedo/"

    FontPath   "/usr/X11R6/lib/X11/fonts/75dpi/"

    FontPath   "/usr/X11R6/lib/X11/fonts/100dpi/"

    FontPath   "/usr/X11R6/lib/X11/fonts/cyrillic/"

 

# ModulePath can be used to set a search path for the X server modules.

# The default path is shown here.

 

#    ModulePath "/usr/X11R6/lib/modules"

 

EndSection

 

# **********************************************************************

# Server flags section.

# **********************************************************************

 

Section "ServerFlags"

 

# Uncomment this to cause a core dump at the spot where a signal is

# received.  This may leave the console in an unusable state, but may

# provide a better stack trace in the core dump to aid in debugging

 

#    Option     "NoTrapSignals"

 

# Uncomment this to disable the <Crtl><Alt><Fn> VT switch sequence

# (where n is 1 through 12).  This allows clients to receive these key

# events.

 

#    Option     "DontVTSwitch"

 

# Uncomment this to disable the <Crtl><Alt><BS> server abort sequence

# This allows clients to receive this key event.

 

#    Option     "DontZap"

 

# Uncomment this to disable the <Crtl><Alt><KP_+>/<KP_-> mode switching

# sequences.  This allows clients to receive these key events.

 

#    Option     "DontZoom"

 

# Uncomment this to disable tuning with the xvidtune client. With

# it the client can still run and fetch card and monitor attributes,

# but it will not be allowed to change them. If it tries it will

# receive a protocol error.

 

#    Option     "DisableVidModeExtension"

 

# Uncomment this to enable the use of a non-local xvidtune client.

 

#    Option     "AllowNonLocalXvidtune"

 

# Uncomment this to disable dynamically modifying the input device

# (mouse and keyboard) settings.

 

#    Option     "DisableModInDev"

 

# Uncomment this to enable the use of a non-local client to

# change the keyboard or mouse settings (currently only xset).

 

#    Option     "AllowNonLocalModInDev"

 

# Set the basic blanking screen saver timeout.

 

#    Option      "blank time"    "10"    # 10 minutes

 

# Set the DPMS timeouts.  These are set here because they are global

# rather than screen-specific.  These settings alone don't enable DPMS.

# It is enabled per-screen (or per-monitor), and even then only when

# the driver supports it.

 

#    Option      "standby time"  "20"

#    Option      "suspend time"  "30"

#    Option      "off time"      "60"

 

# On some platform the server needs to estimate the sizes of PCI

# memory and pio ranges. This is done by assuming that PCI ranges

# don't overlap. Some broken BIOSes tend to set ranges of inactive

# devices wrong. Here one can adjust how aggressive the assumptions

# should be. Default is 0.

 

# Option   "EstimateSizesAggresively" "0"

 

EndSection

 

# **********************************************************************

# Input devices

# **********************************************************************

 

# **********************************************************************

# Core keyboard's InputDevice section

# **********************************************************************

 

Section "InputDevice"

 

    Identifier  "Keyboard1"

    Driver      "Keyboard"

# For most OSs the protocol can be omitted (it defaults to "Standard").

# When using XQUEUE (only for SVR3 and SVR4, but not Solaris),

# uncomment the following line.

 

#    Option     "Protocol"      "Xqueue"

 

# Set the keyboard auto repeat parameters.  Not all platforms implement

# this.

 

#    Option     "AutoRepeat"    "500 5"

 

# Specifiy which keyboard LEDs can be user-controlled (eg, with xset(1)).

 

#    Option     "Xleds" "1 2 3"

 

# To disable the XKEYBOARD extension, uncomment XkbDisable.

 

#    Option     "XkbDisable"

 

# To customise the XKB settings to suit your keyboard, modify the

# lines below (which are the defaults).  For example, for a European

# keyboard, you will probably want to use one of:

#

#    Option     "XkbModel"      "pc102"

#    Option     "XkbModel"      "pc105"

#

# If you have a US "windows" keyboard you will want:

#

    Option     "XkbModel"      "pc104"

#

# If you'd like to switch the positions of your capslock and

# control keys, use:

#

#    Option     "XkbOptions"    "ctrl:swapcaps"

 

 

# These are the default XKB settings for XFree86

#

#    Option     "XkbRules"      "xfree86"

#    Option     "XkbModel"      "pc101"

#    Option     "XkbLayout"     "us"

#    Option     "XkbVariant"    ""

#    Option     "XkbOptions"    ""

 

EndSection

 

 

# **********************************************************************

# Core Pointer's InputDevice section

# **********************************************************************

 

Section "InputDevice"

 

# Identifier and driver

 

    Identifier  "Mouse1"

    Driver      "mouse"

 

# On platforms where PnP mouse detection is supported the following

# protocol setting can be used when using a newer PnP mouse:

 

    Option     "Protocol"      "IMPS/2"

    Option "Device"      "/dev/mouse"

    Option "ZAxisMapping" "4 5"

 

EndSection

 

# **********************************************************************

# Monitor section

# **********************************************************************

 

# Any number of monitor sections may be present

 

Section "Monitor"

 

    Identifier  "My Monitor"

    HorizSync   31.5 - 50.0

 

#    HorizSync  30-64         # multisync

#    HorizSync  31.5, 35.2    # multiple fixed sync frequencies

#    HorizSync  15-25, 30-50  # multiple ranges of sync frequencies

 

# VertRefresh is in Hz unless units are specified.

# VertRefresh may be a comma separated list of discrete values, or a

# comma separated list of ranges of values.

# NOTE: THE VALUES HERE ARE EXAMPLES ONLY.  REFER TO YOUR MONITOR'S

# USER MANUAL FOR THE CORRECT NUMBERS.

 

    VertRefresh 40-90

 

EndSection

 

 

# **********************************************************************

# Graphics device section

# **********************************************************************

 

# Any number of graphics device sections may be present

 

Section "Device"

    Identifier  "fglrx"

    Driver      "fglrx"

EndSection

 

# **********************************************************************

# Screen sections

# **********************************************************************

 

# Any number of screen sections may be present.  Each describes

# the configuration of a single screen.  A single specific screen section

# may be specified from the X server command line with the "-screen"

# option.

Section "Screen"

    Identifier  "Screen 1"

    Device      "fglrx"

    Monitor     "My Monitor"

 

# If your card can handle it, a higher default color depth (like 24 or 32)

# is highly recommended.

 

#   DefaultDepth 8

#   DefaultDepth 16

   DefaultDepth 24

#   DefaultDepth 32

 

# "1024x768" is also a conservative usable default resolution.  If you

# have a better monitor, feel free to try resolutions such as

# "1152x864", "1280x1024", "1600x1200", and "1800x1400" (or whatever your

# card/monitor can produce)

 

    Subsection "Display"

        Depth       8

        Modes "1280x1024" "800x600" "640x480"

    EndSubsection

    Subsection "Display"

        Depth       16

        Modes "1280x1024" "800x600" "640x480"

    EndSubsection

    Subsection "Display"

        Depth       24

        Modes "1280x1024" "800x600" "640x480"

    EndSubsection

    Subsection "Display"

        Depth       32

        Modes "1280x1024" "800x600" "640x480"

    EndSubsection

 

EndSection

 

# **********************************************************************

# ServerLayout sections.

# **********************************************************************

 

# Any number of ServerLayout sections may be present.  Each describes

# the way multiple screens are organised.  A specific ServerLayout

# section may be specified from the X server command line with the

# "-layout" option.  In the absence of this, the first section is used.

# When now ServerLayout section is present, the first Screen section

# is used alone.

 

Section "ServerLayout"

 

# The Identifier line must be present

    Identifier  "Simple Layout"

 

# Each Screen line specifies a Screen section name, and optionally

# the relative position of other screens.  The four names after

# primary screen name are the screens to the top, bottom, left and right

# of the primary screen.  In this example, screen 2 is located to the

# right of screen 1.

 

    Screen "Screen 1"

 

# Each InputDevice line specifies an InputDevice section name and

# optionally some options to specify the way the device is to be

# used.  Those options include "CorePointer", "CoreKeyboard" and

# "SendCoreEvents".

 

    InputDevice "Mouse1" "CorePointer"

    InputDevice "Keyboard1" "CoreKeyboard"

 

EndSection

 

Section "DRI"

        Group 0

        Mode 0666

EndSection

I just got mine working on the nforce2 chipset FINALLY!!! Not 30 mins ago matter o fact. Already played UT2003 for the true test.

I followed Shade's post elsewhere on this forum... Here it is.

The only things I had to make sure of, was I didn't use the nvgart module from nvidia's site. I recompiled my kernel and made sure the nvidia agpgart support was enabled under 'character devices' (it was already btw) in kernel 2.4.22.

Then, I had to change the line in XF86Config-4 that tells it to look for the card @ PCI:3:0:0, and change it to read PCI:3:0:1. It also worked when I just commented that line out.

" Then, I had to change the line in XF86Config-4 that tells it to look for the card @ PCI:3:0:0, and change it to read PCI:3:0:1. It also worked when I just commented that line out."

Could you describe it more specific?
I am a little confused on what are u trying to say.
Or You could just port the part before the changes was made and the part that's after the changes.

Also, what driver are u using? 3.2.8?

I fixed it with the help of what Shade said, sometimes you do have to tell it the location so it will load mine just had a few problems it works fine now but thanks

Also, do if you are using 3.2.8
Do you have any idea how to get the DIR initialized?
My wouldn't never. it always say bad driver module or something!
What's wrong with ATI's driver?

Running fglrxconfig generates a new file /etc/X11/XF86Config-4. The PCI Bus line I was referring to is shown below. It should be somewhere around line 480.

# === Misc Options ===
Option "UseFastTLS" "0"
Option "BlockSignalsOnLock" "on"
Option "UseInternalAGPGART" "no"
Option "ForceGenericCPU" "no"
BusID "PCI:3:0:0" # vendor=1002, device=496e
Screen 0
EndSection

I didn't use the ATI driver, I downloaded the FireGL driver from the schneider-de website for XFree4.3.0. Be sure you change the driver entry to fglrx as shown below:

# === ATI device section ===

Section "Device"
Identifier "ATI Graphics Adapter"
Driver "fglrx"

Keep in mind, that I am running an nForce2 chipset mobo when you set the location for the BusID. You should be able to find your vid card address by typing the command 'dmesg' and hunting through its output. Unless of course, you are using an ASUS A7N8X Deluxe as I am, in which case it will probably be the same.

I have the same mobo as yours and i have ati radeon 9600 pro.
Can gave me a link of that file?

Sure. http://www.schneider-digital.de/html/download_ati.html
Be sure to download the correct driver for your XFree version. Any newer distro will be 4.3.0.

i had just download the driver from http://www.schneider-digital.de/html/download_ati.html and had used
rpm -i --force fglrx-glc22-4.3.0-3.2.8.rpm.
After installation changed XF86Config manually, then i got

Fire GL built-in AGP-support
Based on agpgart interface v0.99 (c) Jeff Hartmann
agpgart: Maximum main memory to use for agp memory: 816M
agpgart: Detected nVidia nForce2 chipset
agpgart: Bridge device is in AGP v1/v2 mode
agpgart: AGP aperture is 128M @ 0xe0000000
Power management callback for AGP chipset installed
[fglrx] AGP detected, AgpState = 0x00000000 (hardware caps of chipset)
[fglrx:firegl_unlock] *ERROR* Process 9780 using kernel context 0
[fglrx:drm_ioremapfree] *ERROR* [mappings] Attempt to free NULL pointer

from dmesg

and

(EE) fglrx(0): [agp] unable to acquire AGP, error "xf86_ENOMEM"
(EE) fglrx(0): cannot init AGP

(WW) fglrx(0): ***********************************************
(WW) fglrx(0): * DRI initialization failed! *
(WW) fglrx(0): * (maybe driver kernel module missing or bad) *
(WW) fglrx(0): * 2D acceleraton available (MMIO) *
(WW) fglrx(0): * no 3D acceleration available *
(WW) fglrx(0): ********************************************* *

again from XF86 log

Coud you figure out what's going on here
I had kernel2.4.22, redhatlinux9 with xfree86-4.3.0

What distro are you running? You should update your profile to include that info ;) (edit: just noticed it's redhat :) ) What does lsmod show?

You may have to recompile your kernel to include DRI support, but I may be wrong. Hopefully someone more knowlegable comes along to help.

Also, post your monitor and graphic device sections from xf86config-4. You did run 'fglrxconfig' to create this file, right?

I can't really speak for an rpm install. I changed it from an rpm to a .tgz (slack package) and still had to cd around to a few different directories to make/make install a few things.

What you mean by "had to cd around to a few different directories to make/make install a few things."
lsmod does show fglrx
Also, i do have DRI support
Can you post your kernel configuration file, so i could compare what's wrong with mine?

I mean it took several steps to get the driver installed as shown in Shade's instructions. I posted the link on the 1st page, but here it is again: http://www.linuxquestions.org/questi...ghlight=radeon

Here's a link to my kernel config. Better I upload it to my old site than everyone have to scroll 2 miles right here ;)

The kernel config is pretty well stock. I haven't gotten around to doing a proper recompile since I re-installed last week (I did a new install to make installing a new Audigy painless). All I really made sure of during this one, is that the nforce support was selected under character devices and that AGPGART was compiled in as well so I could quickly see if this was going to work. I swore I had at least changed the processor from a 486 to an Athlon, but it doesnt appear that I had :scratch: .....

Anyway, give it a look.

Quote:

Originally posted by Scruff
Also, post your monitor and graphic device sections from xf86config-4. You did run 'fglrxconfig' to create this file, right?

I just realized you still hadn't answered these questions.

Quote:

Originally posted by Shade
Definitely. I've done it 10 times over.
The advice above is all wrong for the 9000 Pro if you want 3d acceleration.(same card I've got.)

Here are the general steps

<<snip>>

That should be it buddy :)

-Shade

Shade; you're a legend!

Was struggling with those damn ATi drivers for hours, but I came and found your steps and I'm now enjoying full resolution, hardware accelerated, Linux goodness. :D

/pats on back

You know, I changed back to slack (used it back in the first release) after getting fed up with RedHat and the none standard builds, fedora etc. I had less problems setting up hardware acceleration - actually, was a snap after reading this post - on my Radeon 9000Pro than I ever had with RH9.
Thanks a bunch guys. Great stuff Slack.

Wow... Glad to see that a lot of people have had success with my instructions. :)
I may put up a more thorough how-to.

I've gone through a few more installs since that post, just trying things out differently and acouple of other things I noticed --

agpgart must be compiled as module -- not in the kernel. If compiled in kernel, you will get an odd message from dmesg about the fglrx module having to be inserted prior to any other DRI module -- which I deduced to be agpgart compiled in.

To help with the 9600 --
Ati's newest driver released supports it. I believe you may follow the same install procedure as stated before for Slack 9.1 --
rpm2tgz fglrx-2.5.8<whatever>.rpm
installpkg fglrx-2.5.8<whatever>.tgz

edit XF86Config-4 to use fglrx driver.

And there's the difference for the 9600 -- this is just a hunch based on the errors I saw listed in a previous post of this thread:

Option "hw_cursor"

Try that under the Device section for the vid card.

-Shade

Hi!

Today I installed slackware 9.1 and compiled a 2.4.23 kernel.

I've downloaded the latest driver from ATI, and followed Shade's instructions, but got a very strange error:

root@darkstar:/lib/modules/fglrx/build_mod# ./make
-bash: ./make: No such file or directory
root@darkstar:/lib/modules/fglrx/build_mod# ./make.sh
ATI module generator V 2.0
==========================
initializing...
probing for VMA API version...
cleaning...
patching 'highmem.h'...
skipping patch for 'drmP.h', not needed
patching 'drm_os_linux.h'...
doing script based build for kernel 2.4.x and similar
compiling 'agpgart_be.c'...
compiling 'agp3.c'...
compiling 'i7505-agp.c'...
compiling 'nvidia-agp.c'...
compiling 'firegl_public.c'...
linking of fglrx kernel module...
duplicating results into driver repository...
done.
==============================
You must change your working directory to /lib/modules/fglrx
and then call ./make_install.sh in order to install the built module.
==============================
root@darkstar:/lib/modules/fglrx/build_mod# cd ..
root@darkstar:/lib/modules/fglrx# ./make_install.sh
- creating symlink
- recreating module dependency list
- trying a sample load of the kernel module
Warning: loading /lib/modules/2.4.23/kernel/drivers/char/drm/fglrx.o will taint the kernel: non-GPL license - Proprietary. (C) 2002 - ATI Technologies, Starnberg, GERMANY
See http://www.tux.org/lkml/#export-tainted for information about tainted modules
/lib/modules/2.4.23/kernel/drivers/char/drm/fglrx.o: init_module: Operation not permitted
/lib/modules/2.4.23/kernel/drivers/char/drm/fglrx.o: Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters.
You may find more information in syslog or the output from dmesg
/lib/modules/2.4.23/kernel/drivers/char/drm/fglrx.o: insmod /lib/modules/2.4.23/kernel/drivers/char/drm/fglrx.o failed
/lib/modules/2.4.23/kernel/drivers/char/drm/fglrx.o: insmod fglrx failed
failed.
root@darkstar:/lib/modules/fglrx#

??? What the hell is this? :) Unfortunately XFREE 4.3's generic ati and radeon drivers don't work for my system, so the problem is critical, cause vesa suxx

T-Cat --

Check to see if you maybe compiled agpgart into the kernel.

Use dmesg and see what the last line is after you try a modprobe fglrx.
If it says something like, fglrx must be loaded prior to any other kernel DRI manager.. then that's your problem ;)

Recompile 2.4.23 with agpgart as module.

-Shade

Thx Shade!! :)

Everything went OK after I compiled agpgart and the radeon support from the DRI section as a module

Sorry to bring back this old thread, but I was wondering if anyone could help me out. I followed the steps listed, and checked my kernel config and am positive that AGP support is builtin. I have an nForce2, and was thinking that maybe that PCI Bus ID thing might be the problem. My mobo is an NF7-S by Abit. Also, this line at the bottom of dmesg caught my eye: [fglrx:firegl_unlock] *ERROR* Process 16747 using kernel context 0. Help is much appreciated.

it does on my machine too. it doesn't says that when my system boots up. but after i play tux racer, there is several line which was about it.
maybe upgrade kernel to 2.6.1?
i will stay with 2.4.23. just want a kernel compiling break!!