Re: Re: pv_ops dom0 USB fixed

[Jeremy Fitzhardinge <jeremy@goop.org>]

Todd Deshane wrote:
> On Thu, Dec 11, 2008 at 1:17 PM, Jeremy Fitzhardinge <jeremy@goop.org> wrote:
>   
>> Todd Deshane wrote:
>>     
>>>> Well, for a start, disable MSI for now (pci=nomsi on the kernel command
>>>> line).  I haven't done any MSI support yet.
>>>>
>>>> In the nosmp boot, something seems amiss with interrupt routing.  nomsi
>>>> may
>>>> help there as well, but I'll need to have a closer look at the code to
>>>> work
>>>> out what's really going wrong here.  Hm, its probably:
>>>> [    0.000000] ACPI: Skipping IOAPIC probe due to 'noapic' option.
>>>>
>>>> I guess nosmp implies noapic?  Anyway, we don't do well without ioapics
>>>> at
>>>> present...
>>>>
>>>>  J
>>>>
>>>>
>>>>         
>>> The pci=nomsi helped a lot.
>>>
>>> I have booting (at least almost)
>>>
>>> Again two logs attached. This time the without nosmp does better.
>>>
>>> Gets to a checking battery state line then it seems to stop.
>>>
>>>       
>> What normally happens next?  Looks like its pretty close to the end of the
>> boot sequence; is that where it would normally print the login prompt?  Is
>> the machine on the network at that point?  Can you log in remotely?
>>
>>     
>
> Correct it is fully booted. This is the same thing shown on a normal
> linux boot. I have attached both logs for reference. The linux one
> is not exactly the same kernel, but it is close (both are rc7)
> The vanilla linux one doesn't have any xen dom0 patches.
>   

You can boot the same kernel.  The dom0 kernel will just boot native if 
you don't start it under Xen.

> On the normal linux one, even when putting the console on
> the serial port, gdm loads up on the monitor, but this
> doesn't happen for the dom0 kernel.
>   

I wouldn't expect X to work at this point.

> However I can SSH into the dom0 kernel fine.
>   

That's good; we can add a new card to known working ethernet controllers.

    J