[ ... ] >>> An option vector (or dt-cpu-ftrs) could be defined to disable msgsndp >>> to get KVM performance back. > > An option vector would require a PAPR change. Unless the architecture > reserves some bits for the implementation, but I don't think so. Same > for CAS. > >> Qemu/KVM populates /proc/device-tree/hypervisor, so we *could* look at >> that. Though adding PowerVM/KVM specific hacks is obviously a very >> slippery slope. > > QEMU could advertise a property "emulated-msgsndp", or something similar, > which would be interpreted by Linux as a CPU feature and taken into account > when doing the IPIs. This is really a PowerVM optimization. > The CPU setup for XIVE needs a cleanup also. There is no need to allocate > interrupts for IPIs anymore in that case. We need to keep these for the other cores. The XIVE layer is unchanged. C.
