On Thu, Nov 09, 2006 at 01:53:02AM -0500, Thor Lancelot Simon wrote:
> It is very
> hard for me to concieve of a plausible security model where one would
> want to treat this operation differently from a raw memory write (...)

It could be useful in i386_set_ioperm(). You could allow writes to I/O
ports that don't belong to DMA-capable hardware, and forbid all the
others, for example. (KAUTH_REQ_MACHDEP_X86_IOPERM would need an
additional argument for that.)

Pavel