VMX check_for_null_selector()

VMX check_for_null_selector()

[Jan Beulich]

What is the hidden background for using this function when the guest
is in protected mode? When the selector used is a null one (and not
in 64-bit mode), the instruction should not cause an
EXIT_REASON_IO_INSTRUCTION exit at all, but rather should cause
a GP fault in the guest. In 64-bit mode, a null selector isn't invalid,
and hence doesn't need checking for.
What I'm trying to determine is whether the function must be fixed
(to deal with (a) non-zero CS base addresses and (b) the fact that
on INS there shouldn't be checking for segment overrides, but ES
should be used in all cases) or can be removed.

Thanks, Jan

Re: VMX check_for_null_selector()

[Keir Fraser]

Good question, it's something I've been meaning to ask the Intel guys
about...

 K.

On 22/11/06 08:46, "Jan Beulich" <jbeulich@novell.com> wrote:

> What is the hidden background for using this function when the guest
> is in protected mode? When the selector used is a null one (and not
> in 64-bit mode), the instruction should not cause an
> EXIT_REASON_IO_INSTRUCTION exit at all, but rather should cause
> a GP fault in the guest. In 64-bit mode, a null selector isn't invalid,
> and hence doesn't need checking for.
> What I'm trying to determine is whether the function must be fixed
> (to deal with (a) non-zero CS base addresses and (b) the fact that
> on INS there shouldn't be checking for segment overrides, but ES
> should be used in all cases) or can be removed.
> 
> Thanks, Jan
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel