On Mon, 18 Feb 2019 14:49:55 -0500
Tony Krowiak <address@hidden> wrote:
This patch series introduces hot plug/unplug of a vfio-ap device.
To hot plug a vfio-ap device, the QEMU device_add function may be used:
(qemu) device_add vfio-ap,sysfsdev=$path-to-mdev
Where $path-to-mdev is the absolute path to the mediated matrix device
to which AP resources to be used by the guest have been assigned.
A vfio-ap device can be hot plugged only if:
1. A vfio-ap device has not been attached to the ap-bus (only one is
allowed per guest) either via the QEMU command line or a prior hot
plug action.
2. The guest was started with the CPU model feature for AP enabled
(e.g., -cpu host,ap=on).
To hot unplug a vfio-ap device, the QEMU device-del function may be used:
(qemu) device_del vfio-ap,sysfsdev=$path-to-mdev
Where $path-to-mdev is the absolute path to the mediated matrix device
specified when the vfio-ap device was attached to the virtual machine's
ap-bus.
A vfio-ap device can be hot unplugged only if:
1. A vfio-ap device has been attached to the virtual machine's ap-abus
either via the QEMU command line or a prior hot plug action.
2. The guest was started with the CPU model feature for AP enabled
(e.g., -cpu host,ap=on).
Tony Krowiak (2):
s390x/vfio-ap: Implement hot plug/unplug of vfio-ap device
s390x/vfio-ap: document hot plug/unplug of vfio-ap device
docs/vfio-ap.txt | 58 +++++++++++++++++++++++++++++++++++++++++++++++-----
hw/s390x/ap-bridge.c | 12 ++++++++++-
hw/vfio/ap.c | 2 +-
3 files changed, 65 insertions(+), 7 deletions(-)
In the meantime, the signature of qbus_set_hotplug_handler() has
changed to take an Object instead of a DeviceState, but I just went
ahead and did that trivial change myself.
Thanks, applied.