On Mon, 3 Nov 1997, Jason Thorpe wrote:
> Ping round trip times are often affected by interrupt latency (ping a
> system running Windows NT sometime :-).  Macs are notorious for having
> braindamaged interrupt layouts... this could be biting you.

I haven't gotten around to looking at this yet, but I think the problem is
that the level 4 interrupt handler is special-cased (for better serial
performance maybe?) and doesn't process soft interrupts afterwards like
the other interrupt handlers do. In (most?) other Macs, only the serial
chip is at level 4, but on the AVs, the ethernet is there too. So... it
looks like after ether_input() posts its softnet interrupt, it doesn't get
handled until the next clock tick, which might be up to 17ms later.

Anybody (Bill Studenmund?) know why lev4intr just does an rte instead of
jra rei like the other handlers?

> Someone might want to take a look at the interrupt code I wrote for
> the hp300 port.

I think Allen Briggs is looking into exactly that :) (or at least
something very similar :)

Name: Dave Huang     |   Mammal, mammal / their names are called /
INet: khym@bga.com   |   they raise a paw / the bat, the cat /
FurryMUCK: Dahan     |   dolphin and dog / koala bear and hog -- TMBG
Dahan: Hani G Y+C 22 Y++ L+++ W- C++ T++ A+ E+ S++ V++ F- Q+++ P+ B+ PA+ PL++