Blair Sadewitz wrote:
> There is an [undocumented] 'x86-64' architecture type in gcc 4.1 which
> _apparently_ generates instructions common to AMD64 and EMT64.  I'm
> looking into this further to make sure that this is indeed the case.
> If it is, then it is my position that:
>
> -- we should patch the gcc(1) manual page to reflect this
>
> -- NetBSD/amd64 builds should use this by default.
>
> Comments?
EMT64 was cloned from amd64, and meant to be compatible. Now, in a few 
corner cases, Intel didn't make it quite compatible, but these cases 
were fixed, and the differences left are pretty obscure and should not 
appear in any kernel code, especially not generated code.

Can you give an example on how incompatible code cold be generated?

Also, the Linux compat issues you were seeing: I very strongly doubt 
that they were due to any EMT64/AMD64 differences. Unless you stumbled 
on an undocumented difference..

- Frank