Re: [Qemu-devel] [Patch v1 01/29] qmp: details about CPU definitions in query-cpu-definitions

[David Hildenbrand <dahi@linux.vnet.ibm.com>]

> On Tue, Aug 02, 2016 at 03:23:30PM +0200, David Hildenbrand wrote:
> [...]
> > > > +#
> > > > +# @static: #optional whether a CPU definition is static and will not change
> > > > +#          between QEMU versions / QEMU machines. A static model is always
> > > > +#          migration-safe. If not provided, information is not available.    
> > > 
> > > I assume static models don't change depending on the
> > > machine-type, either. If that's case, we should document that.  
> > 
> > That's what I meant with "QEMU machines", should that be "QEMU machine types"
> > instead?  
> 
> Sorry, I was reading it too quickly. You did document it. :)
> 
> We could extend it to mention the other guarantees (just like my
> previous suggestion for CpuModelExpansionType::static): "will not
> change depending on QEMU version, machine type, machine options,
> and accelerator options".

Sounds good to me!

> 
> >   
> > > 
> > > I believe in this case we don't need to make it optional: just
> > > make the field always present and set it to "false" by default.  
> > 
> > That is true for x86, do you know about the other architectures (arm, ppc)?
> > I'd like to avoid returning false information here for other architectures.  
> 
> As being "static" is not a fact about the existing code, but just
> a guarantee about what the developers are going to do in the
> future, static=false just means that the developers didn't make
> any promises yet (so I don't think it would ever be false
> information).
> 
> In other words, I believe we can safely assume a CPU model is not
> guaranteed to be static unless the maintainers decided to
> explicitly document it as static (and change the data returned by
> query-cpu-definitions).
> 
> (I am assuming that changing it from "false" to "true" in a new
> QEMU version won't be a problem for anybody.)
> 

Hmm, if "static" means, the model will never be changed, but it was changed in
the past, this sounds somewhat strange. I would rather say then "information
is not available" == no guarantee.

But if nobody else sees a problem with that, I can just set it to stable=false
on all other architectures.

Thanks!

David