[PATCH v1 00/11] KVM: x86: break the xAPIC barrier

[Radim KrÄmÃÅ]

RFC: http://www.spinics.net/lists/kvm/msg132036.html

This series allows userspace to create and send interrupts to VCPUs with
APIC ID > 255.

v1: 32 bit API extensions now use a toggleable capability called
KVM_CAP_X2APIC_API.

* new [1/11], loosely related and should have been posted long ago
* [2/11] improved comment for kvm_apic_map_get_dest_lapic() [Peter]
* [3/11] use round_up [Peter]
* [6/11] drop big endian format in lapic get/set ioctl [Paolo]
* [6/11] exclude APICv optimization of APIC ID register
* [6/11] rename __kvm_apic_state_fixup() parameter
* new [7/11], split from the previous patch
* [8/11] added r-b [Peter]
* [10/11] rewritten with a toggleable capability [Paolo]
* [10/11] dropped MSI_ADDR_EXT_DEST_ID to enforce reserved bits

Latest Linux, FreeBSD, and Windows run fine, but I haven't tested VFIO.
Linux works even with APIC ID over 255.  (QEMU and seabios patches to
achieve that are not yet posted.)


Radim KrÄmÃÅ (11):
  KVM: x86: bump KVM_SOFT_MAX_VCPUS to 240
  KVM: x86: add kvm_apic_map_get_dest_lapic
  KVM: x86: dynamic kvm_apic_map
  KVM: x86: use u16 for logical VCPU mask in lapic
  KVM: x86: use generic function for MSI parsing
  KVM: x86: use hardware-compatible format for APIC ID register
  KVM: VMX: optimize APIC ID read with APICv
  KVM: x86: directly call recalculate_apic_map on lapic restore
  KVM: x86: reset lapic base in kvm_lapic_reset
  KVM: x86: add KVM_CAP_X2APIC_API
  KVM: x86: bump MAX_VCPUS to 288

 Documentation/virtual/kvm/api.txt |  26 +++
 arch/x86/include/asm/kvm_host.h   |  17 +-
 arch/x86/kvm/irq_comm.c           |  29 ++--
 arch/x86/kvm/lapic.c              | 357 ++++++++++++++++++--------------------
 arch/x86/kvm/lapic.h              |   9 +-
 arch/x86/kvm/vmx.c                |   5 +-
 arch/x86/kvm/x86.c                |  14 ++
 include/uapi/linux/kvm.h          |   1 +
 8 files changed, 248 insertions(+), 210 deletions(-)

-- 
2.9.0