[Qemu-devel] [RFC PATCH v3 0/4] AMBA platform device passthrough

[Alvise Rigo]

These patches are the next iteration of "[Qemu-devel] [RFC v2 0/4] AMBA
platform device passthrough" and are based on the latest version of VFIO
for platform devices "[RFC PATCH v6 00/20] VFIO support for platform
devices on ARM" and the initial patch series "[Qemu-devel] [RFC v3 0/10] KVM
platform device passthrough".

- [PATCH 1/4] Possibility to bind a wider class of devices. In
  particular the patch proposes a solution to enhance a bit the device
  tree nodes generation, allowing to specify more than one compatibility
  property (handy for AMBA devices). This was required by the DMA330
  that needs "arm,pl330","arm,primecell" as compatibility string,
  together with a clock source defined in the device tree. In the future
  VFIO will be able to tell if we are dealing with an AMBA device;
  before that, we have to rely on the compatibility string to state
  that.

- [PATCH 3/4] The last version of VFIO allows to pass an eventfd file descriptor
  to unmask a host interrupt. This can be used in conjunction with IRQFD to
  benefit of the resampler feature; in this way QEMU no longer has to unmask
  the interrupt because the kernel will take care of it.

- [PATCH 4/4] If event_notifier_init fails to create a new eventfd, it
  will fallback using pipe/pipe2 and we end up passing to the kernel a
  wrong file descriptor. This patch force to use eventfd.

  Changes since v2:

- The EXEC_FLAG now is always used by VFIO if it is supported. QEMU does not
  need to query the kernel any more to know whether using this flag or not.
  For this reason, the patch [1/4] of the previous version has been removed.

- Included patch [2/4] from Eric Auger, required by patch [3/4] and [4/4].

- The patch [3/4] of the previous version has been also removed since the
  introduction of the new mechanism of handling IRQs and EOIs through IRQFD.
  The patch can nevertheless be used to improve the standard, timer based,
  mechanism.

- [PATCH 4/4] Instead of checking whether eventfd has been used by
  event_notifier_init, a new method (eventfd_notifier_init) is added to fulfill
  this purpose.

Alvise Rigo (3):
  Add AMBA devices support to VFIO
  Force eventfd as notifying mechanism for VFIO
  Let VFIO handle the unmasking of interrupts at EOI

Eric Auger (1):
  vfio: Add irqfd support in platform device

 hw/arm/virt.c                 |  59 +++++++++++---
 hw/intc/arm_gic_kvm.c         |   1 +
 hw/vfio/pci.c                 |  18 ++---
 hw/vfio/platform.c            | 179 ++++++++++++++++++++++++++++++++++++++----
 include/qemu/event_notifier.h |   1 +
 util/event_notifier-posix.c   |  24 +++++-
 6 files changed, 244 insertions(+), 38 deletions(-)

-- 
1.9.1