>I was wondering if there is a known bug with Pent systems and 256k pipelined
>cache (external of course).  I cannot successfully compile a kernel with the
>external cache turned on.

If turning the external cache off makes it work, then it can't be a gcc bug.
I don't know of any "known bugs" with Pentium systems and external caches,
but it's certainly common enough for 486 system chipsets to mishandle their
external caches, typically failing to do bus snooping or not handling the
invalidate commands correctly.  You'd think that chipset manufacturers would
learn from previous mistakes and not continually reimplement them; you'd
apparently be sadly mistaken...

(I have such a 486.  *Someday* I have to get around to getting a new
motherboard.)