[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: 40r6 locks up during boot

On Mon, 2009-01-12 at 11:04 +0900, John wrote:
> John wrote:
> > Frans Pop wrote:
> > Ah, so. It's the DVD image.
> > 
> It's different, but not satisfactory.
> I'm copying by hand atm, I'll look at a serial cable (I think I can do 
> that) later, so I will give a bare minimum for the moment.
> Security framework initialized
> SELinux diasbled at boot
> Capability LSM initialized
> Initializing xgroup subsys ns
>                             cpuacct
>                             devices
> CPU: L1 cache 64K etc
> CPU: L2 cache 512K etc
> CPU: Amd Athlon (tm) Dual Core Processor 4450B stepping 02
> checking HLT instruction ... ok
> general protection fault: 3edc [#1]
> Modules linked in:
> Pid 0, comm: swapper Not tainted (2.6.26-1-486 #1)
> EIP 0060:[<c03633f0d>] EFLAGS 00000286 CPU: 0
> EIP is at 0xc0363f0d
> I will leave off now, lots of hex numbers which I'm likely to mistranscribe.
> It ends with kernel panic - not syncing -: Attempted to kill the idle task!
> Oh, the top item on the call trace:
> [<c0107a71>] apply_paravirt+0x81/0x91.

Hmm, this suggests that adding "noreplace-paravirt" to your command line
might help.

It'd be interesting to see all the hex numbers once you've got serial
connected up though, especially since the crashing EIP is at 0xc03xxxxxx
which implies it is in a module. 

I vaguely recall some traffic on LKML about dynamic patching of stuff in
modules, need to try and remember more to dig it out.

> and it's called from vgacon_set_cursor_size.

Called from? Did you see this address on the stack? It might be that it
is attempting to replace some code in this function so it appears on the
stack spuriously.

Ian Campbell

A woman was in love with fourteen soldiers.  It was clearly platoonic.

Attachment: signature.asc
Description: This is a digitally signed message part

Reply to: