Re: [PATCH] kvm: notify host when guest paniced

[Wen Congyang <wency@cn.fujitsu.com>]

At 02/29/2012 06:39 PM, Avi Kivity Wrote:
> On 02/29/2012 12:17 PM, Wen Congyang wrote:
>>>>>
>>>> Yes, crash can be so severe that it is not even detected by a kernel
>>>> itself, so not OOPS message even printed. But in most cases if kernel is
>>>> functional enough to print OOPS it is functional enough to call single
>>>> hypercall instruction.
>>>
>>> Why not print the oops to virtio-serial?  Or even just a regular serial
>>> port?  That's what bare metal does.
>>
>> If virtio-serial's driver has bug or the guest doesn't have such device...
> 
> We have the same issue with the hypercall; and virtio-serial is
> available on many deployed versions.

virtio-serial is available, but it is an optional device. If the guest does
not have this device, the guest cannot tell the host that is is paniced. So
I still prefer to touch the hypervisor.

Thanks
Wen Congyang

> 
>>>
>>>>>> Having special kdump
>>>>>> kernel that transfers dump to a host via virtio-serial channel though
>>>>>> sounds interesting. May be that's what you mean.
>>>>>
>>>>> Yes.  The "panic, starting dump" signal should be initiated by the
>>>>> panicking kernel though, in case the dump fails.
>>>>>
>>>> Then panic hypercall sounds like a reasonable solution.
>>>
>>> It is, but I'm trying to see if we can get away with doing nothing.
>>>
>>
>> If we have a reliable way with doing nothing, it is better. But I donot
>> find such way.
> 
> We won't have a 100% reliable way.  But I think a variant of the driver
> that doesn't use interrupts, or just using the ordinary serial driver,
> should be reliable enough.
>