On Tue, 21 Sep 2021, Rahul Singh wrote:
> Hi Stefano,
> 
> > On 16 Sep 2021, at 9:26 pm, Stefano Stabellini <sstabellini@kernel.org> wrote:
> > 
> > On Thu, 16 Sep 2021, Rahul Singh wrote:
> >> Hi Stefano,
> >> 
> >>> On 10 Sep 2021, at 1:26 am, Stefano Stabellini <sstabellini@kernel.org> wrote:
> >>> 
> >>> On Thu, 19 Aug 2021, Rahul Singh wrote:
> >>>> The existing VPCI support available for X86 is adapted for Arm.
> >>>> When the device is added to XEN via the hyper call
> >>>> “PHYSDEVOP_pci_device_add”, VPCI handler for the config space
> >>>> access is added to the Xen to emulate the PCI devices config space.
> >>> 
> >>> This is done just for device discovery, right?
> >>> 
> >>> Although it is currently not implemented (and I am not asking to
> >>> implement it now, I am only trying to understand the architecture), it
> >>> would be possible to discover all PCI devices just by walking down the
> >>> PCI hierarchy by ourselves in Xen (no Dom0 interactions) given that we
> >>> have an ECAM driver.
> >>> 
> >>> I take that would be the way to implement PCI support for Dom0less?
> >> 
> >> It is not possible to discover PCI devices in XEN if enumeration is not done before XEN boot.
> >> If boot firmware did the enumeration,  XEN will discover the PCI device.
> > 
> > OK, but if the boot firmware does the enumeration, how will Xen discover
> > the PCI device exactly? Will Xen discover it because the PCI device will
> > be present in device tree explicitly (will have its own device tree
> > node)? Or will Xen discover it by walking the PCI bus?
> 
> Yes Xen discover it by walking the PCI bus. There is already a function in XEN scan_pci_devices() that we can use 
> to discover the PCI device for dom0less system.

OK. Please add a statement to the commit message on how dom0less support
could be implemented. It doesn't have to be implemented now of course
but it would be good to keep a note for future reference.