On Sat, 25 Sep 2010, Bill McGonigle wrote:
> I'm trying these two under f13, and not getting past udev starting on the
> pv_ops kernel. (Xen oops) The non-xen kernel works fine on this hardware. A
> pv_ops version from back around January was working well for me, but I wanted
> to grab the x86_64 multilib-problem patched version. That brought me to
> upgrade my xen-* versions, and f12 to f13.
> The top of the stack has 'xen_force_evtchn_callback' on it. I see udev's
> [OK] message, and then it very reliably crashes (I can get it to tell me it's
> going down for reboot with a three-finger salute, but nothing really
> Is a serial console the only way to get a good Xen log? Looking at this
> board, I think I need to take the heatsink off the CPU to get at the serial
> header, so I haven't tried yet...
In your case it might be the only way to get the relevant information. I
also have a computer without a serial port, and have tried other two
things to get debugging information though they may not help in your case.
If the computer boots far enough you can get the xen serial output with
the xm dmesg or xl dmesg commands. Adding the console_to_ring and
conring_size=65536 options to the xen boot line may give more information.
I have also booted a customized xen dom0 live CD as a KVM guest where you
can redirect the console output to a file in the parent machine. That
does have the disadvantage that some of the hardware it sees is emulated
so it may behave differently.
If you haven't already seen them it is worth looking at
xen mailing list