On Thu, Dec 20, 2018 at 11:06 AM Mika Westerberg <mika.westerberg@xxxxxxxxxxxxxxx> wrote: > > On Wed, Dec 19, 2018 at 06:09:15PM +0100, Lukas Wunner wrote: > > > I think better example where this fails is normal Thunderbolt device > > > (not host) which includes PCIe switch and there is an PCIe endpoint, say > > > network interface connected to one of the downstream ports. That > > > downstream port has "Slot implemented" set but is not hotplug capable. > > > > > > So the device would work correctly but if you take the recent "Runtime > > > D3, RTD3" system such as Lenovo Carbon X1 6th gen it keeps the whole > > > PCIe hierarchy from entering D3cold. I would rather not to break that ;-) > > > > Yeah but as you say, those are Downstream Ports. What if you constrain > > it to Root Ports? > > Yes, that could work. I did not test it yet, though. Thinking this bit > further maybe we can contstrain it to ports that have slot implemented > set and have an ACPI companion instead of just root ports? Point being > that ports without ACPI companion could not possibly get ACPI Notify() > either. I like this idea. We can basically assume, at least to begin with, that ACPI companions of PCIe ports are there for a reason. That reason very well may be to provide a way to handle notifications.
