> Which seems to show that there's a design flaw or bug of some sort in
> the lpt driver.

Maybe not so much, I think Jason described it as the way Alpha deals with
interrupts, and infact you can find stay interrupts for all devices (I've
seen them usually for ethernet and SCSI controllers).  Its not really that
much of a problem, except that there is a counter that can cause your system
to lock up if a certain number of strays are generated in a particular amount
of time .. I believe a work around was just to bump up the count to a
much larger number ..

I've see strays with 1.6.2, but I've not had a machine lock up due to strays
since 1.5.x (or early 1.6.x)