problems running OpenGL code

[Michael Powell]

William A. Mahaffey III wrote:

> On 12/12/15 11:01, William A. Mahaffey III wrote:
>> On 12/12/15 10:41, Polytropon wrote:
>>> On Sat, 12 Dec 2015 09:36:23 -0553.75, William A. Mahaffey III wrote:
>>>>
>>>> I have an inhouse OpenGL code named PreBFCGL.TEST which I am trying to
>>>> run. I get the following when I tried about 5 min. ago:
>>>>
>>>>
>>>> [wam at kabini1, ~, 9:30:21am] 936 % PreBFCGL.TEST
>>>> Segmentation fault (core dumped)
>>>> [wam at kabini1, ~, 9:30:38am] 937 % cat BFC.LOG
>>>> [...]
>>>> Xlib:  extension "GLX" missing on display ":0.0".
>>>> Xlib:  extension "GLX" missing on display ":0.0".
>>>> Error: No RGB visual with depth buffer
>>>>
>>>>           Done, Bailing out ....
>>>>    Saving Model Command File to Model_0.in .... done.
>>> It seems to be an X error - the GL extension (GLX) is not
>>> loaded. If you're running X with xorg.conf, make sure
>>>
>>>     Section "Module"
>>>         [...]
>>>             Load            "glx"
>>>         [...]
>>>     EndSection
>>
>> Check.
>>
>>>
>>> is being defined. For OpenGL, other things like direct
>>> rendering (DRI, DRM) might be needed. You can always
>>> check this with tools like glxinfo, xdriinfo, glxgears,
>>> or other programs that make use of OpenGL (for example,
>>> "xlock -mode fire" or "xlock -mode lament").
>>>
>>> The top of "glxinfo" output should look like this:
>>>
>>>     name of display: :0.0
>>>     display: :0  screen: 0
>>>     direct rendering: Yes
>>>     server glx vendor string: NVIDIA Corporation
>>>     server glx version string: 1.4
>>>     server glx extensions:
>>>     [...]
>>>     client glx vendor string: NVIDIA Corporation
>>>     client glx version string: 1.4
>>>     client glx extensions:
>>>     [...]
>>>     GLX version: 1.4
>>
>> Hmmmm:
>>
>> [root at kabini1, /etc, 10:56:18am] 767 % glxinfo
>> glxinfo: Command not found.
>> [root at kabini1, /etc, 10:56:19am] 768 % uname -a
>> FreeBSD kabini1.local 9.3-RELEASE-p24 FreeBSD 9.3-RELEASE-p24 #0: Sat
>> Aug 22 01:54:44 UTC 2015
>> root at amd64-builder.daemonology.net:/usr/obj/usr/src/sys/GENERIC amd64
>> [root at kabini1, /etc, 10:56:29am] 769 %
>>
>>>
>>> Of course your graphics hardware should be reflected accordingly. :-)
>>>
>>> For this to work, I had to put
>>>
>>>     drm_load="YES"
>>>     nvidia_load="YES"
>>
>> Hmmmm .... OK, just did  that. Do have them installed:
>>
>> [root at kabini1, /etc, 10:56:29am] 769 % grep -i drm LIST.installed.txt
>> drm-kmod-9.280208              FreeBSD 9.3 Kernel DRM driver
>> libdrm-2.4.60,1                Userspace interface to kernel Direct
>> Rendering Module services
>> [root at kabini1, /etc, 10:57:09am] 770 % grep -i nvidia LIST.installed.txt
>> nvidia-driver-340-340.93       NVidia graphics card binary drivers for
>> hardware OpenGL rendering
>> [root at kabini1, /etc, 10:57:16am] 771 %
>>
>> I just 'kldload'-ed drm & nvidia, I'll try again momentarily ....
>>
>>>
>>> in /boot/loader.conf - in your case, something similar will probably
>>> be needed. ALso make sure to install relevant libraries as needed
>>> (libGL, libGLU, libdrm, Mesa, ...).
>>
>> Check.
>>
>>>
>>>
>>>
>>>> Am I hosed until 3D OpenGL works for AMD jaguars ? If so, how is
>>>> progress on that front ? These are fairly common APUs, widely used in
>>>> laptops, some low-end desktops & some tablets, it would be sweet to get
>>>> this going :-) ....
>>> As far as I know, this special kind of "graphics card" isn't that
>>> easy to use. It starts with the proper driver. Which one are you
>>> currently using?
>>
>> Uuuuuuuuuhhhh .... Not sure. VESA, I think, how do I check ?
>>
>>
> 
> It is indeed VESA. I just kldload-ed drm & nvidia & still nogo, same
> error ....
> 

It has been way too many years since I've used ARI/AMD video. I'm going to 
assume here that you are trying to utilize the on-chip AMD/ATI, and not a 
second discreet card. These may be of interest:

https://wiki.freebsd.org/AMD_GPU

https://wiki.freebsd.org/Graphics

On the Graphics page in the section "Video driver loaded at boot time" is 
example of the kernel modules to load. When you look in /boot/kernel you 
will see a radeon.ko and a radeonkms.ko followed by a bunch of other 
firmwares which are needed by various models. Many seem to have the model 
name embedded in them and I don't see a particular one with 'Jaguar' in it 
(this is a 10.1 machine I'm looking at here). I believe that even with these 
modules loaded you will still need to install the appropriate driver for 
Xorg under ports/x11-drivers (which still may not support Jaguar yet). It's 
also a possibility that the xorg driver may support, but the kernel modules 
and firmwares haven't caught up yet.

What I don't have any clue is the exact support for the specific chip you are 
trying to use; it may not exist yet or may be a WIP. Maybe there is some 
bleeding edge patch or code around somewhere you can try if nothing 
currently in Release is working.

-Mike