Re: Dom0 lockup

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Thu, Jan 06, 2011 at 01:02:27PM +0530, Daniel Mathew wrote:
> Hi,
> 
> I have been doing some development on HVM. I am passing data between dom0
> and an HVM domU in chunks of 256K in a ring buffer, similar to what the
> console does. A few seconds after starting domU, dom0 locks up; even the

Can you try to limit the guest and the domU to only use one CPU each?
And pin them? It might be that you are missing a couple of memory barriers
and you get your producer/consumer indexes out of sync?

> SysRq keys don't work. For debugging purposes, I have inserted some printk's
> in the hvmemul_read_io function that handles I/O in HVM mode. Prior to

Does the hypervisor serial console work? If so, you can make it print where
the guest or the Dom0 are stuck (it will only give you the eip, and a jumble
of stack, but you should be able to find where what those addresses correspond
to).

> starting domU, I hook up a dom0 file to capture the printk output. I found
> that the logs in this file are ok in the first half; after that, the printk
> outputs are jumbled. After a reset, I found that /var/log/xen/xend.log was
> also corrupted. This has been happening a few times, and I'm not sure how to
> debug the issue. Request some suggestions.
> 
> My development machine is a 4GB Core 2 Duo box. dom0: Debian Lenny (kernel
> version 2.6.32-25, 64bit).
> 
> Thanks,
> Daniel.

> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel