> NOTE TO AL COX: I will bring up the AMD halt issue AFTER I get the SMP
> info/contact(s) we need --> ONLY IF you think it is still necessary -->
> IF the newer AMD chips don't have the bug should we bother to press the
> issue?
Basically there are three things from a Linux point of view that are
important
1. Is this a CPU or Linux bug
2. If its a CPU bug is there a workaround
3. If its a CPU bug is there an easy way to find and report
affected chips
Its a bug, chips have bugs. In fact a lot of people don't realise quite
how buggy CPU's often are
CPU's that have had "crash as any user bugs" include:
Many 386's (POPAD)
Pentium (F00F)
Cyrix (COMA)
68040 (UNLK A7)
TurboSparc
Earlier UltraSparc in 64bit mode
etc..
So its really about knowing what the constraints are and how to best avoid
the problem being possible.
Alan
