Hi,

On Thu, Jun 26, 2025 at 02:02:51AM -0700, Mike Larkin wrote:
> > >...
> > > Doesn't this break with XO kernels?
> >
> > Good point! I have not considered XO, yet.  Mh, worst case this would
> > mean we can not use SEV-(ES|SNP) together PKU support?  Or would there
> > be a way to temporarily remap code readable and then remap XO?
> >
> 
> It was just a thought. XO kernel will require much more to be done, this would
> just add one more thing to the list. No action needed.

thanks!  I see.  And I will put X0 kernel on my list of things to
remember :)

> > >...
> > > I don't understand this assumption. next byte 0xe4 does not imply
> > > "in $0x71, %al"; what is meant by this comment? (also the next one).
> >
> > I am looking at the current byte at RIP, not the next one.  Code like this:
> >
> > ffffffff81903a76:       e4 71                   in     $0x71,%al
> >
> > Does this make sense?
> 
> My point was "0xe4" just means "in  $xxx, %al", it doesn't mean "$0x71".
> Look at the previous examples (0xef, 0xed); I'd recommend making the
> comment in this case be more generic.

I see, will do!

Thanks and take care,
Hans-Joerg

• smime.p7s (5K)