Re: [Qemu-devel] [PATCH] Make default invocation of block drivers safer

[Anthony Liguori <aliguori@linux.vnet.ibm.com>]

On 07/15/2010 04:20 AM, Daniel P. Berrange wrote:
> On Wed, Jul 14, 2010 at 01:50:02PM -0500, Anthony Liguori wrote:
>    
>> On 07/14/2010 01:43 PM, Christoph Hellwig wrote:
>>      
>>> Err, strong NACK.  Please don't start messing with the contents of the
>>> data plane, we're getting into real trouble there.  It's perfectly
>>> valid for a guest to create an image inside an image, and with hardware
>>> support for nested virtualization I guess this use case will become
>>> rather common, just as it already is on S/390 with VM.
>>>
>>>        
>> Then we have to remove block format probing.
>>
>> The two things are fundamentally incompatible.
>>      
> FWIW, the latest libvirt code will now always set a fmt=XXX arg even
> for raw, so probing will never be performed. We also always set a backing
> store format when creating qcow2 files.
>    

I assume libvirt probes once, then updates the guest's XML?

This is definitely the most user friendly option of all except for the 
fact that we don't have a writeable config file in QEMU today.

Regards,

Anthony Liguori

> The main remaining unsolved issues are backing stores with non-qcow2
> files for which QEMU will always probe, and pre-existing qcow2 files
> which have been created without backing store formats.  It would be
> nice to have an explicit arg to disable all backing store probing,
> forcing backing format to either be raw, or match the parent image
> format.
>
> Regards,
> Daniel
>