Re: [PATCH v2 4/4] docs/firmware: add feature flag for qemu variable store

["Daniel P. Berrangé" <berrange@redhat.com>]

On Wed, Mar 19, 2025 at 12:51:16PM +0100, Gerd Hoffmann wrote:
> On Wed, Mar 19, 2025 at 11:37:40AM +0000, Daniel P. Berrangé wrote:
> 
> > > > > +# @qemu-vars: The firmware expects qemu to provide an efi variable
> > > > > +#             store, via "uefi-vars-sysbus" or "uefi-vars-x64" device.
> > 
> > I wonder if 'qemu-vars' is the right name here ? It feels like the specification
> > for such device is effectively defined by UEFI, with any hypervisor providing a
> > impl. Perhaps just call it 'uefi-vars-dev' or some name that's relevant for
> > what EDK2 calls it ?
> 
> 'host-uefi-vars' maybe?  Or 'vmm-uefi-vars'?
> 
> > > There is 'stateless' already for 'firmware image in r/o flash'.
> > 
> > What's the behaviour of UEFI if build with JSON vars support, but without
> > QEMU providing any JSON vars backend ?
> 
> It will panic.

In that case, we must not reuse 'stateless' with such builds, as that's
quite different semantics & incompatible with current usage.

We would need a new 'uefi-vars' mode, or just declare that we must
always use 'memory'  not 'flash' for such builds.

> 
> > We would want to expand the 'stateless' docs to mention that this feature
> > flag indicates optional support for persistence in that case.
> 
> Optional is not possible do due to the way variable store support is
> organized in edk2.  Firmware can't switch between smm and non-smm
> configuration at runtime for similar reasons.
> 
> take care,
>   Gerd
> 

With regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|