[SOLVED] X freeze at system boot after update (not system)

NuLL3rr0r · Last edited by NuLL3rr0r on Sat May 11, 2019 10:16 am; edited 1 time in total

I am going crazy over this since it has been almost a month. I mostly do emerge --sync && emerge -auDN @world daily and since I do not turn off this laptop most of the time even when I carry it around I did not notice what exactly ruined my Gentoo installation.

Now, every time I turn on this laptop when X starts it freezes with a blank screen. I see the HDD LED blinking even after X freezing, so I know the system is not freezing (I haven't tried ssh-ing to the system since at the moment this is the only system I have). The funny thing is when I remove xdm service from the default runlevel the system boots fine and when I start xdm service (sddm is my login manager) it sometimes successfully pass me through to the login screen and sometimes it won't and it freezes. So, I have to hold the power button down since the keyboard is also frozen and I cannot go to another tty.

When I go to my rescue shell after the freeze I cannot find anything unusual in Xorg.log no errors, no warnings, nothing (Xorg log when it freezes https://pastebin.com/2505vG1e - Xorg log when it won't freeze https://pastebin.com/qJycJvYP). Though I always have some complaints from thermald and newcpufreq in my system logs even before this happened (https://pastebin.com/RJReZVCX). The kernel config is also the same.

So, any help on how to track down this issue where I have no clue is really appreciated.

jburns · Posted: Thu May 02, 2019 9:44 pm Post subject:

There was a problem with some systems that would hang with some versions of linux and nvidia-drivers, you could try a newer version of linux.

Hu · Administrator Joined: 06 Mar 2007 Posts: 21867

Can you reproduce this problem with an untainted kernel? If so, it may be easier to debug.

NuLL3rr0r · Posted: Fri May 03, 2019 3:03 pm Post subject:

@jburns I tried the latest v5.1-rc7 and it's still the same.

NuLL3rr0r · Posted: Fri May 03, 2019 3:06 pm Post subject:

@Hu I am using the vanilla sources and I started from scratch just adding the required laptop and vmware drivers/settings to the kernel. If I remove the .config and rebuild the kernel would that suffice for building an untainted kernel? Or, the kernel picks up the previous config?

Hu · Administrator Joined: 06 Mar 2007 Posts: 21867

A kernel is tainted by loading certain types of modules. You don't need to rebuild. Just blacklist those modules, then reboot and do not load them. Your dmesg will tell you when loading a module taints the system. The proprietary nVidia module is known to taint the kernel.

jburns · Posted: Sat May 04, 2019 5:05 am Post subject:

Some kernel frame buffer drivers can cause nvidia failures. Pastebin the kernel config and dmesg.

NuLL3rr0r · Posted: Sat May 04, 2019 10:45 am Post subject:

Thank you! Here is the kernel .config https://pastebin.com/XquwqzVj and the dmesg https://pastebin.com/W517P1Q9

jburns · Posted: Sat May 04, 2019 7:13 pm Post subject:

I don't see any problems in .config or dmesg. The only thing I noticed is that the nvidia drivers load earlier on your system than on mine.
In .config you have CONFIG_AGP=y when there is no AGP on your system and CONFIG_VGA_ARB=y which you probably do not need.

NuLL3rr0r · Posted: Sat May 04, 2019 10:08 pm Post subject:

@jburns, thanks for figuring that out. At the moment upgrading the system's GCC and rebuilding. I will try it and report back. Thank you!

NuLL3rr0r · Posted: Sat May 11, 2019 10:14 am Post subject:

Hmm, it seems the issue is gone for me. I did not even touch the kernel config. Just switched the gcc profile to 9.1.0 and:

NuLL3rr0r · Posted: Mon Aug 19, 2019 1:18 am Post subject:

Sorry, to bring up this dead thread. But, I finally figured out the causes and found a workaround for this issue. For anyone coming across this, blacklisting NVIDIA drivers is the workaround. Read more about it here.