Re: [PATCH 1/2] Increment virtio-net savevm version to avoid conflict with upstream QEMU.

[Anthony Liguori <aliguori@us.ibm.com>]

Alex Williamson wrote:
> On Wed, 2009-04-29 at 15:53 -0500, Anthony Liguori wrote: 
>   
>> -#define VIRTIO_NET_VM_VERSION    6
>> +/* Version 7 has TAP_VNET_HDR support.  This is reserved in upstream QEMU to
>> + * avoid future conflict.
>> + * We can't assume verisons > 7 have TAP_VNET_HDR support until this is merged
>> + * in upstream QEMU.
>> + */
>> +#define VIRTIO_NET_VM_VERSION    7
>>     
>
> It seems like you're saying you're only going to reserve version number
> 7, and not the 4 bytes of savevm we're using for version 7 here.
> Couldn't we fix this by adding a dummy patch to qemu to bump to version
> 7, and push/pop a 4 byte zero from the savevm?  Then we could change the
> code below to >= 7.  Qemu should probably puke on a savevm image with
> non-zero in this location until the kvm code gets merged.  Looks like
> one byte would be more than sufficient if we wanted to make that change
> now too.  Thanks,
>   

I'd rather just merge vnet into upstream QEMU as quickly as possible.  
All I have to do to reserve a field is just hope noone submits a patch 
incrementing version id until we submit vnet support :-)

--
Regards,

Anthony Liguori