spurious interrupts

[Jonathan Hue]

Why do all of our Suns spit out "spurious level X interrupt" kernel
printfs?  (Sun-3/[12]60, SunOS 3.5)

We see these whenever we rewind the 1/2" tape drive (Fujitsu on a Xylogics
472), which seem to be harmless, and sometimes when using a National
Instrments GPIB controller (connected to an Eikonix scanner, using their
driver).  These aren't harmless, as the machine dies soon afterwards,
usually due to a kernel bus error.

I think I know what a spurious interrupt is, it's when something tugs on
IRQ[1-7], and the CPU performs an interrupt acknowledge cycle and doesn't
see DSACK[01], so it gets a bus error.  I'm curious why they occur, it
seems like a hw bug on someone's part.

-Jonathan	uunet!jvc!jonathan