Re: [RFC 2/6] roles: provide abstraction for the possible domain roles

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Tue, 8 Aug 2023, Daniel P. Smith wrote:
> > Thanks for the example, it is finally clear!
> > 
> > I don't think we should add hardware_dom=1 to the command line (I don't
> > know if it was just an example to make it easier to understand for me).
> > Instead it should a property on device tree. hardware_dom=1 is not great
> > because it is tied to the order of domain construction. Instead it
> > should be a device tree property under the specific domain. That way you
> > can clearly specify which one is tradition dom0, or which is
> > hardware_domain and which is control_domain.
> 
> I was describing with today's code, how you would launch a late hardware
> domain. The logic to trigger the hardware transfer from dom0 to the hardware
> domain is triggered by the command line parameter `hardware-dom`. With
> hyperlaunch, this can eventually be retired and separate control domain and
> hardware domain can be constructed at launch as we expose access to
> roles/capabilities through its DT interface.

+1


> > > > which I don't understand
> > > 
> > > Did the above help?
> > 
> > A lot, thanks!
> 
> Your welcome, glad I was able to articulate it better this time.

Thanks for the explanation, it all looks good to me.