Re: [OOPS 2.4.0-test2] Reproducable, good looking, probably apmd (is the same as oops while starting with debian)

[Russell King]

Soeren Sonnenburg writes:
> I reported this before but not as detailed, so maybe someone knows to
> deal with it...

> >>EIP; c0148bf4 <proc_permission+64/f4>   <=====

This is *exactly* what I was chasing down last night on the ARM
architecture.  In this case, it was caused by:

	kernel_thread(xxxx, NULL, 0);
				  ^ - should be
				CLONE_FS | CLONE_FILES | CLONE_SIGHAND

This was in the ARM-specific code, so the problem I found is not
exactly the same as yours.

Basically, there is a process running in your kernel which has no
knowledge of a root filesystem.  I don't see anything common between
kernel_thread() and your kernel messages to suspect any others
in the generic kernel tree - maybe its a result of a patch you applied?
   _____
  |_____| ------------------------------------------------- ---+---+-
  |   |         Russell King        [e-mail]      --- ---
  | | | |   http://www.arm.linux.org.uk/~rmk/aboutme.html    /  /  |
  | +-+-+                                                     --- -+-
  /   |               THE developer of ARM Linux              |+| /|\
 /  | | |                                                     ---  |
    +-+-+ -------------------------------------------------  /\\\  |