RE: [PATCH v5 1/4] domctl: introduce a new domain create flag, XEN_DOMCTL_CDF_evtchn_fifo, ...

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Fri, 4 Dec 2020, Durrant, Paul wrote:
> > -----Original Message-----
> > From: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
> > Sent: 04 December 2020 17:45
> > To: Stefano Stabellini <sstabellini@xxxxxxxxxx>
> > Cc: Julien Grall <julien@xxxxxxx>; Jan Beulich <jbeulich@xxxxxxxx>; 
> > paul@xxxxxxx; Durrant, Paul
> > <pdurrant@xxxxxxxxxxxx>; Elnikety, Eslam <elnikety@xxxxxxxxxx>; 'Ian 
> > Jackson' <iwj@xxxxxxxxxxxxxx>;
> > 'Wei Liu' <wl@xxxxxxx>; 'Anthony PERARD' <anthony.perard@xxxxxxxxxx>; 
> > 'George Dunlap'
> > <george.dunlap@xxxxxxxxxx>; 'Christian Lindig' 
> > <christian.lindig@xxxxxxxxxx>; 'David Scott'
> > <dave@xxxxxxxxxx>; 'Volodymyr Babchuk' <Volodymyr_Babchuk@xxxxxxxx>; 'Roger 
> > Pau Monné'
> > <roger.pau@xxxxxxxxxx>; xen-devel@xxxxxxxxxxxxxxxxxxxx
> > Subject: RE: [EXTERNAL] [PATCH v5 1/4] domctl: introduce a new domain 
> > create flag,
> > XEN_DOMCTL_CDF_evtchn_fifo, ...
> > 
> > CAUTION: This email originated from outside of the organization. Do not 
> > click links or open
> > attachments unless you can confirm the sender and know the content is safe.
> > 
> > 
> > 
> > On 04/12/2020 17:41, Stefano Stabellini wrote:
> > >>> FAOD, I am sure there might be other features that need to be
> > >>> disabled. But we have to start somewhere :).
> > >> Absolutely top of the list, importance wise, is so we can test different
> > >> configurations, without needing to rebuild the hypervisor (and to a
> > >> lesser extent, without having to reboot).
> > >>
> > >> It is a mistake that events/grants/etc were ever available unilaterally
> > >> in HVM guests.  This is definitely a step in the right direction (but I
> > >> thought it would be too rude to ask Paul to make all of those CDF flags
> > >> at once).
> > > +1
> > >
> > > For FuSa we'll need to be able to disable them at some point soon.
> > 
> > FWIW, I have a proper plan for this stuff, which start alongside the
> > fixed toolstack ABI, and will cover all aspects of optional
> > functionality in a domain.
> > 
> 
> OK. Can we live with this series as it stands until that point? There is some 
> urgency to get at least these two things fixed.

I am happy to take things one step at a time, and this is a good step
forward.