[PATCH 0/3] x86/hvm: follow-up to f7039ee41b3d

[Jan Beulich <jbeulich@xxxxxxxx>]

Much of what gets done here was discussed in the context of
"ioreq: handle pending emulation racing with ioreq server
destruction" already.

1: fold hvm_io_assist() into its only caller
2: re-work hvm_wait_for_io() a little
3: fold both instances of looking up a hvm_ioreq_vcpu with a request pending

Jan