LinuxQuestions.org - dual-head support, radeon 300, fglrx

- Linux - Hardware (https://www.linuxquestions.org/questions/linux-hardware-18/)

- - dual-head support, radeon 300, fglrx (https://www.linuxquestions.org/questions/linux-hardware-18/dual-head-support-radeon-300-fglrx-718433/)

dual-head support, radeon 300, fglrx

Hello,

I'm trying to follow these instructions for dual-head setup with ATI configuration:
http://www.yolinux.com/TUTORIALS/Lin...lMonitors.html

I'm running Slackware 12.2, kernel 2.6.27.7-smp, with Radeon X300. I have one integrated video connector, and and add-in graphics card (2VGA ports).

(1) I downloaded the ATI driver for Radeon X300, and installed it.

Code:

#sh ati-driver-installer-9-3-x86.x86_64.run

I received an error and was pointed to look at the install log
here is the tailing bit of /usr/share/ati/fglrx-install.log:

Code:

WARNING: "pci_enable_msi" [/lib/modules/fglrx/build_mod/2.6.x/fglrx.ko] undefined!

WARNING: "pci_disable_msi" [/lib/modules/fglrx/build_mod/2.6.x/fglrx.ko] undefined!

  CC      /lib/modules/fglrx/build_mod/2.6.x/fglrx.mod.o

  LD [M]  /lib/modules/fglrx/build_mod/2.6.x/fglrx.ko

make[1]: Leaving directory `/usr/src/linux-2.6.27.7'

build succeeded with return value 0

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.

- recreating module dependency list

- trying a sample load of the kernel modules

failed.

[Error] Kernel Module : Failed to install compiled kernel module - please consult readme.

I am unsure which readme to consult here. I tried looking in /lib/modules/fglrx, but this appears to be the incorrect. I'm also unclear as to what I am supposed to be consulting in the readme.

(2) I tried running 'Xorg -configure' , and I receive the error:

Code:

fglrx: No matching Device section for instance (BusID PCI:1:0:1) found

(3) relevant lspci info

Code:

01:00.0 VGA compatible controller: ATI Technologies Inc RV370 5B60 [Radeon X300 (PCIE)] (prog-if 00 [VGA controller])

        Subsystem: ATI Technologies Inc Unknown device 0302

        Flags: fast devsel, IRQ 11

        Memory at d0000000 (32-bit, prefetchable) [size=128M]

        I/O ports at dc00 [size=256]

        Memory at dfde0000 (32-bit, non-prefetchable) [size=64K]

        Expansion ROM at dfe00000 [disabled] [size=128K]

        Capabilities: [50] Power Management version 2

        Capabilities: [58] Express Endpoint, MSI 00

        Capabilities: [80] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-

        Capabilities: [100] Advanced Error Reporting <?>

        Kernel modules: fglrx, radeonfb



01:00.1 Display controller: ATI Technologies Inc RV370 [Radeon X300SE]

        Subsystem: ATI Technologies Inc Unknown device 0303

        Flags: bus master, fast devsel, latency 0

        Memory at dfdf0000 (32-bit, non-prefetchable) [size=64K]

        Capabilities: [50] Power Management version 2

        Capabilities: [58] Express Endpoint, MSI 00

Thanks in advance for any help or suggestions. Let me know if I need to provide any other information.

Have you gotten past the

$ glxinfo | grep -i "direct"
direct rendering: Yes

part? If not you may need to move the ati version of libGL over the Mesa version of the same. And maybe even compile the fglrx kernel module by hand. Depending on how you installed the driver. Then you can worry about dual head.

HTH

Code:

# glxinfo | grep -i "direct"

Error: unable to open display 



# glxinfo

Error: unable to open display

Quote:

If not you may need to move the ati version of libGL over the Mesa version of the same.

I'm not sure I understand this sentence, or how to implement it. Does this mean there are two libGL versions on my machine, and I should replace the Mesa version with the ati version? How do I do that?
Searching for libGL*, I see:

Code:

/usr/lib/libGL.so.1.2

/usr/lib/libGLU.so

/usr/lib/libGL.so

/usr/lib/xorg/modules/extensions/libGLcore.so

/usr/lib/xorg/modules/extensions/libGLcore.la

/usr/lib/libGLw.so.1

/usr/lib/libGL.so.1

/usr/lib/libGLU.so.1

/usr/lib/libGLU.so.1.3.070003

/usr/lib/libGLw.so

/usr/lib/libGLw.so.1.0.0

/usr/share/ati/libGLdir.txt

but I imagine something else is going on.

I noticed this post:
http://www.linuxquestions.org/questi...6/#post3448699
and that in the error log from the fglrx install:

Code:

WARNING: "pci_enable_msi" [/lib/modules/fglrx/build_mod/2.6.x/fglrx.ko] undefined!

WARNING: "pci_disable_msi" [/lib/modules/fglrx/build_mod/2.6.x/fglrx.ko] undefined!

Ostensibly those are related. But you are suggesting recompiling a different aspect of the kernel.

Quote:

Depending on how you installed the driver.

I just downloaded the appropriate driver and did:
#sh ati-driver-installer-9-3-x86.x86_64.run

Another bit of information:
If I try to plug a monitor into the VGA for the integrated graphics card and boot, I get a screen telling me to unplug it, use the add-in graphics card instead, then reboot. I don't know if that reveals anything.

Thanks for your help.

With a distro specific install I just use dpkg-divert in debian. And it tells me of the conflict and it's location(s).

/usr/lib/libGL.so.* (mesa's default location / what X actually uses at boot)
/usr/share/ati/* (and other locations depending on version / distro)

$ find / -iname '*libGL.so*'

$ cat /usr/share/ati/libGLdir.txt

I'm running slackware 12.2, here are the outputs for those commands:

# find / -iname '*libGL.so*'
/usr/lib/libGL.so.1.2
/usr/lib/FGL.renamed.libGL.so.1.2
/usr/lib/libGL.so
/usr/lib/libGL.so.1

# cat /usr/share/ati/libGLdir.txt
/usr/lib

Is direct rendering enabled?

$ glxinfo | grep -i "direct"

Too many parts making up the whole IMO. It looks like your libGL is ATI's version. Double check timestamps (ls -l) to see if it makes sense. You also need that section DRI stuff in your xorg.conf. And various other parts to make the whole. Also note that for ATI, 3D acceleration may not be enabled when using a dual-head configuration. Or when using compiz/beryl or AIGLX.

I had previously ran this remotely while not having started X, so I received an Error. Here it is properly:

Code:

# glxinfo | grep -i "direct" 

direct rendering: No (If you want to find out why, try setting LIBGL_DEBUG=verbose)

OpenGL renderer string: Mesa GLX Indirect

all the dates from 2009-04-11 is when I installed the ati driver, so it makes sense that libGL is ati, yes.

Code:

ls -l /usr/lib | grep 'libGL'

-rwxr-xr-x  1 root root  354508 2008-07-16 13:27 FGL.renamed.libGL.so.1.2

lrwxrwxrwx  1 root root      21 2009-04-11 11:06 libGL.so -> /usr/lib/libGL.so.1.2

lrwxrwxrwx  1 root root      21 2009-04-11 11:06 libGL.so.1 -> /usr/lib/libGL.so.1.2

-rw-r--r--  1 root root  564480 2009-04-11 11:06 libGL.so.1.2

lrwxrwxrwx  1 root root      11 2009-04-11 02:50 libGLU.so -> libGLU.so.1

lrwxrwxrwx  1 root root      20 2009-04-11 02:50 libGLU.so.1 -> libGLU.so.1.3.070003

-rwxr-xr-x  1 root root  485836 2008-07-16 13:27 libGLU.so.1.3.070003

lrwxrwxrwx  1 root root      11 2009-04-11 02:50 libGLw.so -> libGLw.so.1

lrwxrwxrwx  1 root root      15 2009-04-11 02:50 libGLw.so.1 -> libGLw.so.1.0.0

-rwxr-xr-x  1 root root    10956 2008-07-16 13:27 libGLw.so.1.0.0

Quote:

You also need that section DRI stuff in your xorg.conf

Which "stuff" is this? Would you be able to give me more detailed info about what to add to xorg.conf? Thanks again for your help until now.

Quote:

Originally Posted by noviciado (Post 3515231)

Code:

# glxinfo | grep -i "direct" 

direct rendering: No (If you want to find out why, try setting LIBGL_DEBUG=verbose)

OpenGL renderer string: Mesa GLX Indirect

This pretty much says that you're NOT using ATI's libGL.

server glx vendor string: SGI
OpenGL renderer string: ATI Radeon HD 4550

You'd have those two if you were. Specific to your card of course.

In your xorg.conf you might need stuff like this:

Section "Extensions"
Option "Composite" "false"
EndSection

Section "DRI"
Group "video"
Mode 0660
EndSection

(although I seem to be missing my section DRI and with DRI for all users. Generally without it, only root has acceleration.)

Once you swap in ATI's libGL, you'll need to restart X to have it take affect. And depending on how you install it, you may need to compile, copy, and modprobe the fglrx module independently. /usr/src/modules.tar.bz or some other non-descript location/name from what I recall. nVidia drivers may not be as feature-full, but they're a little easier to install IMO. Although just as hard to configure.

HTH

Hmm, you're right. I forgot my first post; I had errors installing the driver (and libGL with it I imagine).
I guess I have to build the fglrx module 'by hand'?

I get errors installing the ati driver, which point me to installing fglrx's built module...but that too fails.

Code:

WARNING: "pci_enable_msi" [/lib/modules/fglrx/build_mod/2.6.x/fglrx.ko] undefined!

WARNING: "pci_disable_msi" [/lib/modules/fglrx/build_mod/2.6.x/fglrx.ko] undefined!

  CC      /lib/modules/fglrx/build_mod/2.6.x/fglrx.mod.o

  LD [M]  /lib/modules/fglrx/build_mod/2.6.x/fglrx.ko

make[1]: Leaving directory `/usr/src/linux-2.6.27.7'

build succeeded with return value 0

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.

- recreating module dependency list

- trying a sample load of the kernel modules

failed.

[Error] Kernel Module : Failed to install compiled kernel module - please consult readme.

Code:

# ./make_install.sh 

- recreating module dependency list

- trying a sample load of the kernel modules

FATAL: Error inserting fglrx (/lib/modules/2.6.27.7-smp/kernel/drivers/char/drm/fglrx.ko): Unknown symbol in module, or unknown parameter (see dmesg)

failed.

I've never recompiled or built anything for the kernel, so I'd have to read up quite a bit before trying to go anywhere further with this--if, indeed, that is what it appears I must do next.
Thanks for you help until now.