FreeBSD NVIDIA support vs TrueOS NVIDIA support

mikepwagner

mikepwagner

I recently decided to move from Ubuntu back to BSD on an old Dell latitude 6510 laptop.

I had run PC-BSD on the laptop a year ago, and everything worked pretty well.

PC-BSD has apparently become TrueOS Desktop, and when I tried to install TrueOS, it warned me that my legacy NVIDIA card is not longer supported.

Since this all worked a year or so ago, am I likely to be successful if I try to install freeBSD directly, instead of using TrueOS desktop?

I think - though I am not sure - that NVIDIA provided freeBSD drivers in the past, and that I may be able to get a driver that works with my GPU directly from NVIDIA. Is that going to be a headache?
 
Go to nvidia site, check what is the driver to your model. Than locate it in /usr/ports/x11/nvidia* .

NVIDIA still proving fresh drivers for a bunch of new products.
 
Nvidia regularly drops support for legacy hardware from the mainline drivers. The last version before removing support for several chipsets are kept as legacy versions:
https://www.nvidia.com/object/unix.html

A list of what cards fall into "legacy" is also provided:
https://www.nvidia.com/object/IO_32667.html

For very old GPUs you might need older legacy versions; just work your way back from the latest one (340) until your card works. You can also read the supported hardware list in the release notes, but don't blindly trust them - I was running a legacy GeForce 4 till last year, which *should* have worked with the 340 driver, but only ran stable with the 304 branch.
 
I think that I will try this. I am a little confused by the manual. It seems as though there are many steps in Section 5.4 "Xorg Configuration", but when I look at videos on YouTube about installing a GUI on freeBSD, it seems like more people skip most of the steps.

This is what I see in videos:
  1. Install the xorg pkg
  2. Install some desktop environment
  3. If the desktop environment doesn't include a login manager, install slim
  4. Add hald_enable="YES" and dbus_enable="YES" (and maybe slim_enable="YES") to /etc/rc.conf
  5. Add an exec line to the user's .xinitrc to start the desktop environment

If I need to add NVIDIA 340 driver to work with my video card, can I just install that package between steps 1 and 2, or do I need to modify some conf file to tell xorg to nvidia-driver-340?

Thanks.
 
mikepwagner said:
If I need to add NVIDIA 340 driver to work with my video card, can I just install that package between steps 1 and 2, or do I need to modify some conf file to tell xorg to nvidia-driver-340?
Click to expand...
After installing the nvidia-driver 340, you may need to run nvidia-xconfig, that will generate an xorg.conf file. That surely worked for me anyway.

This is what I have installed:
nvidia-driver-340-340.96_1 NVidia graphics card binary drivers for hardware OpenGL rendering
nvidia-settings-355.11_3 Display Control Panel for X NVidia driver
nvidia-xconfig-367.35 Tool to manipulate X configuration files for the NVidia driver
Click to expand...
nvidia-settings may be useful if you need to tweak some parameter, like brightness, contrast and more.
 
mikepwagner said:
3. If the desktop environment doesn't include a login manager, install slim
4. Add hald_enable="YES" and dbus_enable="YES" (and maybe slim_enable="YES") to /etc/rc.conf
Click to expand...

You may skip that by now. I prefer to do things progressively. Just run startx if you setup .xinitrc as you said and run nvidia-xconfig as fellow ASX mentioned.

hald and dbus can be good and evil, at the same time! Test one step after another, is my humble suggestion.

[off topic on] For historical reasons with a notebook using the openchrome driver, I quit with login managers, so I'm used to type startx. As time passed, I put this into my .bashrc :

Code: 
if [ `tty` == "/dev/ttyv6" ]; then
  cp -p ~/.xinitrc-fluxbox ~/.xinitrc
else
  cp -p ~/.xinitrc-lumina ~/.xinitrc
fi
  startx &
  exit

May be stupid, may not... that's what I love *nix ... allowing people to do clever or stupid customized useful things :D

[/off topic on]
 
Was Nvidia the reason to switch back to FreeBSD? For some reason, I have 3x more trouble with Nvidia on Ubuntu than on FreeBSD. On FreeBSD, it seems always to work just fine. Maybe it's MIR related ...
 
ronaldlees said:
Was Nvidia the reason to switch back to FreeBSD? For some reason, I have 3x more trouble with Nvidia on Ubuntu than on FreeBSD. On FreeBSD, it seems always to work just fine. Maybe it's MIR related ...
Click to expand...

Not really. I went to ubuntu from PC-BSD mostly because I wanted to use gnucash with a PostGres backend, and at that the time, the PC-BSD gnucash package didn't include PostGres support. I tried to build it, and that didn't work very well, and I didn't feel like allocating enough time debug the issue.

I am not as interested in gnucash as I was, and I decided to move back to BSD.
 
I think "vanilla" FreeBSD nVidia support might be a bit ahead of TrueOS/PC-BSD. That being said, I know that for your (mikepwagner) graphics card, the 340xx line is the correct one. I use it whenever in doubt regarding the mainline nVidia driver. As ASX mentioned, running nvidia-xconfig is highly recommended. Bare in mind that this will generate an xorg.conf file that takes precedence over the xorg.conf.d directory where individual Section .conf files are stored. Therefore, if you do any adjustments to X11 configuration, those need to go to xorg.conf.
 
wblock@ said:
This is unlikely. TrueOS, which is the replacement for PC-BSD, is based on FreeBSD-CURRENT and includes the latest KMS stuff. It's cutting edge stuff.
Click to expand...

"TrueOS says "Legacy Nvidia cards are not working properly at this time. We are working to resolve this issue."
 
The older NVIDIA drivers (340.x, 304.x) are working perfectly fine in TrueOS now. The installer just does not include them (only has the latest driver), so you will need to install them manually afterwards before turning on the "nvidia" driver for X11.
 
You must log in or register to reply here.
Back
Top