Response to recent bug report

[utzoo!decvax!cca!sri-unix!hplabs!menlo70!nsc!bill]

>From bill Mon Apr 12 14:21:43 1982
To: menlo70!ucbvax!sdcsvax!phonlab!donn
In-Response-to: <phonlab.210>
Subject: Your 2.8bsd questions
Cc: bill menlo70!ucbvax!carl menlo70!ucbvax!sdcsvax!sdchema!fred!donn menlo70!virus!mike

In answer to your netnews queries:

	1) MFPS,MTPS. Yes this has been done before on V6 systems. The
reason we haven't used it yet is because of two bits of errata that I have
not yet determined:
		a) How "portable" is this to other 11's ? I heard a rumor
		   that this breaks on 11/60's, but don't have one to try it
		   with.
		b) On some V6 systems where I saw this done they first set
		   the spl to one lower, then to the target spl. Is this 
		   getting around some obscure 11 glitch or what?

	2) Clock mess. There are two reasons for spl6 in clock: Idiotic
management of timeout routines, and various obscure drivers diddling things in
KDSA5(and in strange cases, KDSA2). There is quite a jumble of ideas here so
no attempt will be made to fix 2.8. Instead, 2.81, the new release, has been
rewritten to aviod the previously mentioned problems. Until 2.81 becomes
available, you can turn the spl6 into a 5 if: you have no one using KDSA5 
except mapped buffers and no dz's being used with interrupt silo code. In
many cases, the first condition is optional, but...

	3) Awk. 2.81 has Awk recompilable on 11/40's. This is neccessary cause
the system uses it for reconfiguration now.


	Bill Jolitz.
	(ucbvax!william)