On Thu, May 30, 2019 at 03:38:23PM +1000, David Gibson wrote:
> PAPR has its own PCI hotplug mechanism which we implemented quite a
> while ago. However, we weren't able to hotplug devices under PCI
> bridges, because the pci-bridge code would try to handle the hotplug
> using SHPC which won't work with PAPR.
>
> We now have the infrastructure in the hotplug core to have the machine
> take control of the hotplug before the bus tries to claim it, which
> lets us address this. While we're about it, make a bunch of cleanups
> and also allow hot plug of P2P bridges themselves (this turns out to
> be quite straightforward once the rest is in place).
>
> For now, we don't support hot unplug of bridges - this seems to cause
> problems, at least for a Linux guest. I suspect the problem is on the
> guest side, but for now just prohibit until we have time to
> investigate properly.
I've fixed the problem Alexey pointed out, and I've merged this
tentatively to the ppc-for-4.1 branch.
>
> Changes since v1:
> * Added proper cover letter
> * Addressed some minor comments
>
> David Gibson (8):
> spapr: Clean up device node name generation for PCI devices
> spapr: Clean up device tree construction for PCI devices
> spapr: Clean up dt creation for PCI buses
> spapr: Clean up spapr_drc_populate_dt()
> spapr: Clean up DRC index construction
> spapr: Don't use bus number for building DRC ids
> spapr: Direct all PCI hotplug to host bridge, rather than P2P bridge
> spapr: Allow hot plug/unplug of PCI bridges and devices under PCI
> bridges
>
> hw/ppc/spapr.c | 25 +-
> hw/ppc/spapr_drc.c | 13 +-
> hw/ppc/spapr_pci.c | 486 ++++++++++++++++++++++--------------
> include/hw/pci-host/spapr.h | 4 +-
> include/hw/ppc/spapr_drc.h | 3 +-
> 5 files changed, 327 insertions(+), 204 deletions(-)
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson