Re: [Qemu-devel] Patch: Adding ability for qemu-img to create SCSI VMware disk images

[Anthony Liguori]

Kevin Wolf wrote:
> Anthony Liguori schrieb:
>   
> > Kevin Wolf wrote:
> >     
> > > Anthony Liguori schrieb:
> > >   
> > >       
> > > > Aaron Mason wrote:
> > > >     
> > > >         
> > > > > Who said the images had to be used for QEMU?  As I said, I use the 
> > > > > utility primarily for creating VMware images. 
> > > > >       
> > > > >           
> > > > If it's not useful for QEMU, then I don't see that it's worth 
> > > > maintaining within QEMU.
> > > >     
> > > >         
> > > Then I wonder why we have the compat6 option for VMDK. I would be rather
> > > surprised if it was useful for qemu itself.
> > >   
> > >       
> > Yup.  I don't like the compat6 patch either as it's pretty invasive.
> >     
>
> Hm, we're both talking about ec36ba1? Because I can't see how it is
> invasing, with the exception of the changes to qemu-img. But we don't
> need these any more, we have -o now.
>   

I find it invasive because it extends the block subsytem API with a 
non-generic interface.

> > Introducing BLOCK_FLAG_BUSLOGIC and BLOCK_FLAG_LSILOGIC seems very wrong 
> > to me.  You're happy with it?
> >     
>
> No, I'm not happy with this particular patch, especially given that it
> doesn't apply to master any more. ;-)
>
> You said this was "not useful", so I'm just talking of the
> functionality, independent of the code. And code-wise, it really should
> be only a few lines to vmdk - which makes the create function longer by
> maybe five lines, but hardly more complex.
>   

I wouldn't mind a few line patch to vmdk's create function.

Regards,

Anthony Liguori

> Kevin
>