> Suggested workaround. None.
>
> Resolution status. This erratum will be corrected in a future stepping
> of the AMD-K6 processor.
>
> ------------------------ Typo's are mine -------------------------
>
> (*) And B and C stepping which are K6-MMX.
>
> This is a serious problem for Linux right?
>
> Assuming SMM has something to do with system calls or interrupts, Or
> would Linux never allow me to load a NULL segment register? In 32bit
> mode, I don't expect to be able to load segment registers but WINE and
> DOSEMU would have to be able to do this right?
SMM is normally used only for power management and such, so I think this
would be only a minor problem for Linux. As a workaround it might be
possible to have the SMM handler check for null selectors (knowing that
they would be handled incorrectly), and it could then return to a stub
that would reload the zeroes from userland before returning to the
interrupted program.
Regards,
Bill
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu