> Come on, Intel is stressing like hell to get this thing fixed (I
> estimate the cost of replacing the defective CPUs at around $5e9), but
> they have certain restrictions like "please don't release this yet,
> until we're sure it works". So they asked BSDI to "keep the lid on the
> source for the fix for another -=*DAY*=- .
>
> BSDI has signed an NDA with Intel, such that Intel has some control
> over what gets said in public. Have you signed such an NDA? If Red Hat
> would have had such an agreement in place with Intel, then the fix
> would have been released for Linux at the same moment, under the same
> restrictions.
Point taken, please excuse my reactionary responce.
> > I dont know about anyone else here, but losing my internal cache is not at
> > all in option, the performance is absolutly positivly not an option.
>
> Neither for Intel would this be acceptable. You can disable your cache
> from the BIOS.
I dont know about that.. Well have to wait and see.. And disableing it
from cmos wouldn't do it.. It would have to be populated by executing a
'proper' invalid opcode catching the exception then locking it down.