PROM Bugs

[Ludis Langens]

Roger Heflin wrote:
> 
> On Sun, 4 Jul 1999, Ludis Langens wrote:
> 
> > It's sad because these bugs could be easily found through "mechanical"
> > means.  In other words, a computer could have found them without human
> > intervention or being programmed with specific knowledge about the chip.
> >  Not finding these bugs is equivalent to building a high stakes race
> > motor from a junk yard heap without magnafluxing.
> 
> I take it these are the standard array overflow/underflow bugs that
> could be caught with something like purify?   If so, that is somewhat
> sad since they are pretty easy to catch.

They aren't really overflow/underflow type bugs - they're more like
random memory reads.  Some of them could be caught because the data
"type" doesn't match the implied "type" from other references to the
same location.  A few will show up as array limit violations.  Others
can be caught because the X register could have several values, each
pointing at a different data "type".

If GM's assembler had a psuedo-op which declared the programmer's
register content assumptions, then these bugs could have been easily caught.

-- 
Ludis Langens                               ludis (at) cruzers (dot) com
Mac, Fiero, & engine controller goodies:  http://www.cruzers.com/~ludis/