Richard B. Johnson wrote:
> I ran my test program on the 2.1.63 kernel. The fix works! I also
> have a 64-bit random-number generator running on another machine
> to check for other problems on the Pentium. Without the prefix, one
> has to be careful because a loop-forever is possible. I fix this with
> an alarm-clock timeout of the child. Basically, I have been executing
> 64-bits of random garbage (with the known sequences excluded). This
> has been running since 3-AM. I will run this for several weeks or
> until another bad sequence is found. The machine it's running on is
> presently not in use otherwise.
> <SNIP>
J.P. Racine
Osiris Technologies Inc.
racinejp@osiris.net