[PLUG] Installing nvidia-legacy340 on Slackware 14.2

[Ben Koenig]

On Tuesday, January 15, 2019 7:22:04 PM PST Dick Steffens wrote:
> On 1/15/19 3:20 PM, Dick Steffens wrote:
> > I'm trying to downgrade from the nVidia legacy 390 driver to the
> > legacy 340 driver. To do so I first downloaded the
> > nvidia-legacy340-kernel.tar.gz, and unpacked it. Then I downloaded
> > NVIDIA-Linux-x86_64-340.104.run and moved it to the
> > nvidia-legacy340-kernel directory. Then I downloaded
> > NVIDIA-Linux-x86_64-390.87.run from the nVidia site. I ran
> > nvidia-legacy340-kernel.SlackBuild
> > 
> > Next I downloaded and unpacked nvidia-legacy340-driver.tar.gz. I
> > downloaded nvidia-modprobe-340.104.tar.bz2,
> > nvidia-persistenced-340.104.tar.bz2, nvidia-settings-340.104.tar.bz2,
> > nvidia-xconfig-340.104.tar.bz2, NVIDIA-Linux-x86_64-340.104.run and
> > moved them all into nvidia-legacy340-driver. Then I ran
> > nvidia-legacy340-driver.SlackBuild.
> > 
> > Before installing the packages I ran nvidia-switch --remove
> > 
> > Next I ran installpkg
> > /tmp/nvidia-legacy340-kernel-340.104_4.4.157-x86_64-1_SBo.tgz . It
> > reported Package
> > nvidia-legacy340-kernel-340.104_4.4.157-x86_64-1_SBo.tgz installed.
> > Then I ran installpkg
> > /tmp/nvidia-legacy340-driver-340.104-x86_64-1_SBo.tgz. It reported
> > Package nvidia-legacy340-driver-340.104-x86_64-1_SBo.tgz installed.
> > 
> > When I run startx I get the error:
> > 
> > (EE)
> > Fatal server error:
> > (EE) no screens found(EE)
> > (EE)
> > 
> > <...>
> > 
> > In dmesg I find:
> > 
> > [257997.749754] NVRM: API mismatch: the client has the version
> > 340.104, but
> >                 NVRM: this kernel module has the version 390.87. Please
> >                 NVRM: make sure that this kernel module and all NVIDIA
> > driver
> >                 NVRM: components have the same version.
> > 
> > I thought nvidia-switch --remove should have removed the old kernel.
> > What did I miss this time?
> 
> I neglected to restart. Once I did that I was able to startx. However, I
> had hoped the downgrade to version 340 would solve my problem of not
> getting the two monitors in the positions I want them.
> 
> I noticed some nvidia_modeset messages running by when I was restarting.
> Here's what I see in dmesg:
> 
> [   15.457662] nvidia_modeset: Unknown symbol nv_kthread_q_flush (err 0)
> [   15.457683] nvidia_modeset: Unknown symbol
> nv_kthread_q_schedule_q_item (err 0)
> [   15.457711] nvidia_modeset: Unknown symbol nvidia_get_rm_ops (err 0)
> [   15.457728] nvidia_modeset: Unknown symbol nv_kthread_q_item_init (err 0)
> [   15.457740] nvidia_modeset: Unknown symbol nv_kthread_q_stop (err 0) [  
> 15.457759] nvidia_modeset: Unknown symbol nv_kthread_q_init (err 0)
> 
> And near the end:
> 
> [   39.326881] NVRM: Your system is not currently configured to drive a
> VGA console
> [   39.326884] NVRM: on the primary VGA device. The NVIDIA Linux
> graphics driver
> [   39.326886] NVRM: requires the use of a text-mode VGA console. Use of
> other console
> [   39.326888] NVRM: drivers including, but not limited to, vesafb, may
> result in
> [   39.326889] NVRM: corruption and stability problems, and is not
> supported.
> [   67.447694] NVRM: API mismatch: the client has the version 390.87, but
>                 NVRM: this kernel module has the version 340.104. Please
>                 NVRM: make sure that this kernel module and all NVIDIA
> driver
>                 NVRM: components have the same version.
> [   67.447710] NVRM: nvidia_frontend_ioctl: minor 255, module->ioctl
> failed, error -22
> 
> Do those messages have anything to do with my positioning problem? And
> if not, are they significant for my regular operation?

You are trying to fix the problem without any knowledge of what the problem is. 
This is why I constantly tell people to slow down. The log you just posted 
hits on multiple topics in one single braindump. 

To break it down into 3 topics...

First, Go back to nvidia-390, there was never a reason to downgrade.

Second, boot errors about a VGA console are normal. The Nvidia module 
complains, but it is a meaningless error and can be ignored.

Third, I don't think you ever told us what exactly you are trying to do with 
your multi-monitor config. I remember a lot of confusion and chaos about 
getting X started and using xorg.config files in your home directory (which does 
not work, X doesn't load config files from $HOME..).

Since your driver packages are now FUBAR, run the following steps to get back 
to a working display:
- Uninstall any nvidia driver packages on your system
- reinstall the base distribution packages that were overwritten
slackpkg reinstall xorg-server libdrm mesa

- Make sure you DON'T have a file named /etc/X11/xorg.conf
- delete the nouveau blacklist file in /etc/modprobe.d/
- reboot

After you reset back to the sane defaults, you can explain to us what it is 
you are trying to do and maybe someone can advise a path forward.