Screen of death - How I solved it

Erdie · Last edited by Erdie on Tue Sep 12, 2006 11:08 am; edited 1 time in total

Sometimes, this means every few weeks, i get a strange error during bootup. Because i cannot make a Screenshot, I made a "Screenshot" with my digital camera:

http://www.erdie.de/screen-of-death.jpg

The initial message is: "unable to handle kernel NULL pointer reference"

My guess is, this is an issue with the preemtible kernel option. Does somebody know more?

thanks
Erdie
_________________
Desktop AMD Ryzen 9 5900X 32GB RAM, Asus GF GTX 1060.
Notebook Tuxedo Pulse 15 Gen1 AMD Ryzen 7 4800H mit Radeon Vega 7
Raspberry Pi 1 + 2 + 3B+ + Zero W

r00t440 · Tux's lil' helper Joined: 14 Sep 2005 Posts: 93

this is most likely a "kernel compile" issue. you could try the following:

1. If your are using the standard kernel from portage (e.g., gentoo-sources), try using non-standard kernel with new and latest patches (i.e., no-sources).

2. If you are using bleeding edge kernel try switching back to standard kernels.

3. If you suspect that a certain configuration is the culprit try turning it off.

Recompile, recompile, recompile...

Erdie · Posted: Thu Jun 15, 2006 12:32 pm Post subject:

This will be hard to realize because it takes sometimes serveral weeks before the error comes again
_________________
Desktop AMD Ryzen 9 5900X 32GB RAM, Asus GF GTX 1060.
Notebook Tuxedo Pulse 15 Gen1 AMD Ryzen 7 4800H mit Radeon Vega 7
Raspberry Pi 1 + 2 + 3B+ + Zero W

davidgurvich · Veteran Joined: 23 Apr 2004 Posts: 1063

Are you running an SMP kernel? This seems to be similar to amd64 powernow feature using SMP kernel.

Erdie · Posted: Thu Jun 15, 2006 1:10 pm Post subject:

No, i do not use SMP. But I have the preemtible Kernel option enabled and I guess the reason could be there. But I am using my machine for Digital Audio Workstation purposes so this option is very useful
_________________
Desktop AMD Ryzen 9 5900X 32GB RAM, Asus GF GTX 1060.
Notebook Tuxedo Pulse 15 Gen1 AMD Ryzen 7 4800H mit Radeon Vega 7
Raspberry Pi 1 + 2 + 3B+ + Zero W

Erdie · Posted: Sat Aug 05, 2006 4:48 am Post subject:

Now I changed my configuration. I just load my ehci_hcd and ohci_hcd with modules.autoload/kernel2.6 and i disabled couldplug in my system.
The results is that somtimes!! I get the same error in my boot screen but the system continues booting without any USB driver. Any ideas to avoid this?

Erdie
_________________
Desktop AMD Ryzen 9 5900X 32GB RAM, Asus GF GTX 1060.
Notebook Tuxedo Pulse 15 Gen1 AMD Ryzen 7 4800H mit Radeon Vega 7
Raspberry Pi 1 + 2 + 3B+ + Zero W

Erdie · Posted: Tue Sep 12, 2006 11:08 am Post subject:

How I fixed this error:

I found out that the system bootup does not stop if I include uhci_hcd and ehci_hcd in modules_autoload directly without coldplug.
But sometimes the error came up anyway and it was caused by loeading ehci_hcd. In this case the system does not lockup completely but the USB subsystem was corrupted.

My 3rd try was to assume that this could be a timing problem because it occours just sometimes non-systematiclly. Therefore I created a script: