[PATCH 00/24] XSM: follow-on to XSAs 492 and 499

Working on those XSAs made pretty apparent that there's a lot of redundancy,
requiring changes in too many separate places if e.g. adding / altering /
removing a hook. Obviously while dealing with that, some other, smaller
tidying opportunities turned up as well.

01: XSM: reduce redundancy in hook machinery
02: XSM: make .grant_*() hooks dependent upon GRANT_TABLE=y
03: XSM: make .{,un}map_domain_pirq() hooks dependent upon HAS_PIRQ=y
04: XSM: make PCI hooks dependent upon HAS_PCI=y
05: XSM: make .iomem_mapping_vcpi() hook dependent upon HAS_VPCI=y
06: XSM: make .kexec() hook dependent upon KEXEC=y
07: XSM: make .hypfs() hook dependent upon HYPFS=y
08: XSM: make .hvm_altp2mhvm_op() hook dependent upon ALTP2M=y
09: XSM: make .hvm_param*() hooks dependent upon HVM=y
10: XSM: make .dm_op() hook dependent upon IORE_SERVER=y
11: XSM: make certain x86-specific hooks dependent upon PV=y
12: x86/mm: get_page_from_l1e() is PV-or-shadow-only
13: x86: restrict PHYSDEVOP_* when PV=n
14: XSM/dummy: fold cf_check into XSM_INLINE
15: XSM/dummy: drop redundant return statements
16: XSM: suppress hypercall when XSM=n
17: XSM: make Argo hooks well-formed ones
18: XSM: make XSM hooks well-formed ones
19: XSM: convert "allow" (Flask: "access") parameters to bool
20: XSM: fold xsm_{,un}map_domain_pirq() hooks
21: x86: type-correct last parameter of map_domain_pirq()
22: XSM: pass just SBDF to xsm_{,un}map_domain_irq()
23: XSM: fold xsm_{,un}map_domain_irq() hooks
24: XSM: fold xsm_{,un}bind_pt_irq() hooks

Jan