3.3.0-6-liquorix crashed on first boot.

3.3.6 vanilla has not crashed for several boots now.
I used your '3.3.1-zen Kernel Configuration' and left out some AVM ISDN driver modules, because they failed to compile for god knows why.

Everything in x86_64 of course.

For now I will keep on testing vanilla 3.3.6.
Back to top

Interesting, then I feel it's likely that BFS is causing your hardware problems. Can you confirm by merging Con's patch from here? -> ck.kolivas.org/patches/3.0/3.3/3.3-ck1/patches/3.3-sched-bfs-420.patch

I'm not aware of any other invasive patches that would cause problems with slab, so that's your best bet.
Back to top

OK, here is what I found out:

- Don't ever use your old notebook to compile if you're used to a 6 times faster monster at home ;-)

- BFS patch is to blame, no matter if using BFQ or CFQ

- After a boot with BFS patch kworker/0:2 and ksoftirqd/0 take up 50% each of CPU core 0 for roughly one minute, core 1 remains unused

- The kernel either panics within that period or it will remain stable


I feel that it's easier to crash it without PSU, coincidence?
Back to top

Well, I showed Con this thread and he recognized it's a real bug. At first he suspected I was running slab, but that's not the case for you, since I explicitly set slub in the configuration.

If you want to get more help, you should talk directly with Con. Join #ck @ irc.oftc.net, and ask if there's anything you can do to help with your kernel panics. He might come up with some rough patches you can try that modify parts of BFS he suspects would cause the oops you're running into.
Back to top

If anyone hits the same problem here is how it stands:
- the panics started with 3.3.4
- conman can't reproduce it himself
- anyway he has no time to hunt it down and probably won't for a long while
- there is no other maintainer

So i'm back to stock for now.
Or is there a boot parameter for chosing the CPU scheduler?
Back to top