Re: [Qemu-devel] [PATCH v2 5/6] x86: fix up 32 bit phys_bits case

[Daniel P. Berrange]

On Tue, Jul 05, 2016 at 12:29:30PM +0100, Dr. David Alan Gilbert wrote:
> * Daniel P. Berrange (address@hidden) wrote:
> > On Mon, Jul 04, 2016 at 08:16:08PM +0100, Dr. David Alan Gilbert (git) 
> > wrote:
> > > From: "Dr. David Alan Gilbert" <address@hidden>
> > > 
> > > On 32 bit systems fix up phys_bits to be consistent with what
> > > we tell the guest; don't ever bother with using the phys_bits
> > > property.
> > 
> > > @@ -2990,6 +2986,15 @@ static void x86_cpu_realizefn(DeviceState *dev, 
> > > Error **errp)
> > >              cpu->phys_bits = host_phys_bits;
> > >  
> > >          }
> > > +    } else {
> > > +        /* For 32 bit systems don't use the user set value, but keep
> > > +         * phys_bits consistent with what we tell the guest.
> > > +         */
> > > +        if (env->features[FEAT_1_EDX] & CPUID_PSE36) {
> > > +            cpu->phys_bits = 36;
> > > +        } else {
> > > +            cpu->phys_bits = 32;
> > > +        }
> > 
> > I kind of feel like we should report an error and exit if the
> > user/app has provided a phys_bits property value, rather than
> > silently ignoring their provided value, on the basis that this
> > is a user/app configuration error.
> 
> Do we have an easy way to tell that the user has set the parameter
> as opposed to it being the default?

Not sure if there's an official way, but you could perhaps default
phys_bits to -1, and treat -1 as being equivalent to 0 if set by
the user.


Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|