No X with kernels 4.19.0 and 4.19.1 -- RESOLVED

Fred Krogh · Last edited by Fred Krogh on Wed Nov 14, 2018 11:22 pm; edited 1 time in total

Kernel 4.18.10 is working just fine, but I'm not able to get X working with 4.19. I've looked at a lot of places where somewhat similar problems are discussed, but have found nothing that works for me.
Some things that may help in pointing to the problem. In ~/.xsessions.errors

eccerr0r · Posted: Thu Nov 08, 2018 12:05 am Post subject:

If you change kernel versions, you must also rebuild the dynamically loadable kernel modules, of which nvidia-drivers is one.

Make sure your

# eselect kernel list

is pointing to the correct kernel source in /usr/src, and re-emerge nvidia-drivers.

However one problem that comes up frequently is that nvidia-drivers tends to lag latest and greatest kernels, and cannot be recompiled for the newer kernels without surgery to the code. It can be hard to fix due to the binary distribution nature of the driver. As I do not own an nvidia card that requires the use nvidia-drivers, I can't comment about this.

As a few asides:

1 - The error message you posted is a side effect of, or unrelated to, the real problem - not the direct cause.
2 - why do you want to get rid of pam, as long as all components use pam it seems like a good thing ensuring all authentication goes through the same codebase instead of DIY which can cause race conditions if it even applies to your own box... If you don't want authentication I'd look into other methods, you could simply tell pam not to require any credentials when asked. This is still a bad idea IMHO, but personally think pam is a good thing.
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

Fred Krogh · Posted: Thu Nov 08, 2018 12:16 am Post subject:

I compile the kernel with the following script

eccerr0r · Posted: Thu Nov 08, 2018 1:25 am Post subject:

The error messages printed aren't particularly helpful other than possibly pointing at a module version mismatch.

If there's nothing in dmesg, did the module even load? (lsmod after seeing the error?)
_________________
Intel Core i7 2700K/Radeon R7 250/24GB DDR3/256GB SSD
What am I supposed watching?

Ralphred · Guru Joined: 31 Dec 2013 Posts: 501

This may be completely unrelated, but in switching between 4.18 and 4.19 my gfx card changed address on the pci bus.

Fred Krogh · Posted: Fri Nov 09, 2018 11:20 pm Post subject:

Sorry, I've just been able to get back to this. The table below summarizes what I have found. I'm hoping this suggest something to somebody. I'm running xorg-server-1.20.3. Thanks,

NeddySeagoon · Posted: Sat Nov 10, 2018 10:20 am Post subject:

Fred Krogh,

Stand by to wait.

The evil binary blob nvidia-drivers doesn't build on 4.19.x yet.
nvidia need to fix it.

I understand that there may be some user patches around the web.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

Fred Krogh · Posted: Sat Nov 10, 2018 6:38 pm Post subject:

I'm not sure what to believe, but http://rglinuxtech.com/?p=2415 says the latest nvidia-driver work with kernel 4.19.x. If this is the case something else must be causing my problem.

Naib · Posted: Sat Nov 10, 2018 6:42 pm Post subject:

Fred Krogh · Posted: Wed Nov 14, 2018 11:22 pm Post subject:

I've just installed kernel 4.19.2, and X is working. I'll just call this mystery resolved.

Just realized it had to do with using genkernel on a aborted attempt to use systemd. Evidently that took priority, and I just happened to notice that it was always trying to use that kernel. Once I deleted that all works as I expected.