On Tue, 11 Aug 1998, Charles M. Hannum wrote:

> 
> >When I tried enabling interrupts in these functions (which one _should_
> >be able to), the machine fell over running lmbench (which it shouldn't).
> 
> Well, I just `#if 0'd all the code that turns off interrupts during
> TLB and cache flushes, and ran lmbench, and it didn't even hiccup.
> It's conceivable that this had something to do with the
> splimp-splclock inversion that I fixed when I added splaudio, but I'm
> not sure.  I definitely don't see any bug here, though.
I think Neils problem is the state of his old hacked source tree.
Certainly with the most recent versions of cpufunc_asm.S things should
work if the irq disabled are #if 0'ed out.
However attention needs to be paid to the suswintr() and fuswintr()
functions i.e. if they are called from an interrupt occurring during a
cache clean/purge operation this can result in a failed cache clean as you
cannot guarentee that all user land addresses will be correctly purged
from the cache.

Cheers,
				Mark