ACPI FIXVW2 4-20-2011 a regression?
 

/APIC doesn't work at all without /SMP on same line. It is just
ignored without it.

Are you sure of this statement? I thought that /APIC vs /PIC had to do
with how interrupts are routed and that /APIC only makes sense on
mother boards that support the more up-to-date interrupt handling.

At its root, it's a design choice that one has to make about what are
the default and non-default configurations for an operating system. If
you look at this from a wider perspective than merely OS/2 platform
drivers, you'll find that although the various combinations are *in
theory* orthogonal, in practice the choice of one generally dictates the
choice of another. In Windows NT, for example, not all of the possible
combinations of ACPI, local APICs, and SMP exist as HALs (e.g.
HALACPI.DLL v. HALAPIC.DLL v. HALAAPIC.DLL v. HAPMACPI.DLL v.
HALMPS.DLL). It's essentially a design decision what combinations to
support, with what defaults, and when to say "If you don't have X then
you're just going to have to do without Y as well.". What Allan is
saying is that the design choice made here is that if you, the system
administrator, choose the ACPI platform driver, then you get APIC and
SMP as a job lot: either you have both or you get neither.