Subject: re^2: pmax interrupt problem solved
To: None <port-pmax@netbsd.org>
From: Terry R. Friedrichsen <terry@venus.sunquest.com>
List: port-pmax
Date: 03/08/2000 10:11:37
FWIW, I just finished rebuilding the MI SCSI kernel on my DS5000/200 with
the patch I posted earlier.

Let me make that more clear:  I am running a kernel with my patch, and
rebuilding the entire kernel from scratch while running the patched kernel.
Why?  Just as a sort-of torture test of the patch and benchmark of MI SCSI.

Under old-style SCSI, building a kernel from scratch took 2 hours and 36
minutes.

Under MI SCSI with the patched interrupt dispatch code, building a kernel
from scratch took 2 hours and 11 minutes, so this is a significant improve-
ment.



Michael Hitch wrote, regarding my patch:

> I had already commited a similar fix yesterday, but I'm going to change
> it a bit now that I understand what the code was attempting to do.

Hmmm - my sup run at 1 a. m. didn't pick that up; I'll be anxious to see
what you come up with.  I thought about getting fancier with the switch()
statements, but since what I wrote was already faster than the existing
code, I didn't bother.




Terry R. Friedrichsen

terry@venus.sunquest.com