[QEMU PATCH 0/2] cxl: fixes for hot-removal of devices

[Joshua Lant <joshualant@gmail.com>]

As requested from the discussion in:
https://lore.kernel.org/linux-cxl/20260429135717.3048713-1-joshualant@gmail.com/T/#t

Two bugs have been found when testing cxl topologies for managed
hot-remove of a device:

Patch 1 properly unmaps the backend memory device, allowing it to be
remapped if the removed device is re-added.

Patch 2 sets the power controller present capability in the downstream
port. This is required to notify the device from the guest of removal.

Modifications from linked RFC (patches 3 & 4):

- Originally patch 3 (1 here) unmaps only on exit. It should also unmap in the
error goto in the realize function.
- "Fixes:" tags added to commit messages.

Joshua Lant (2):
  cxl-type3: Properly unmap the memory-backend on device exit
  cxl_downstream: enable power controller present capability.

 hw/mem/cxl_type3.c             | 6 ++++++
 hw/pci-bridge/cxl_downstream.c | 2 ++
 2 files changed, 8 insertions(+)

-- 
2.47.3