Re: [PATCH] Re: SSE related security hole

[H. Peter Anvin]

>
> I get this feeling that Intel screwed up on specifying how to
> initialize this whole state.
>

Indeed.  Logically, FNINIT should have been extended to initialize it all -
- it is a security hole that it doesn't initialize MMX properly.
Alternatively, for SSE only, an INITP instruction could have been added
that an SSE-enabled OS can use at the time OSXFSR or whatever that flag is
called is set.