Return to BSD News archive
Path: sserve!manuel.anu.edu.au!munnari.oz.au!hp9000.csc.cuhk.hk!saimiri.primate.wisc.edu!zaphod.mps.ohio-state.edu!howland.reston.ans.net!wupost!uunet!pipex!bnr.co.uk!uknet!cf-cm!myrddin!paul
From: paul@isl.cf.ac.uk (Paul,,,)
Newsgroups: comp.os.386bsd.bugs
Subject: Error in patch 89
Message-ID: <1993Mar5.160055.17730@cm.cf.ac.uk>
Date: 5 Mar 93 16:00:54 GMT
Sender: news@cm.cf.ac.uk (Network News System)
Reply-To: paul@isl-gate.isl.cf.ac.uk
Organization: Intelligent Systems Lab., ELSYM, University of Wales, College of
Cardiff
Lines: 37
I found a mistake in patch89. The patch to the awk config doesn't contain anything except the header details !!
I assume it should set HAVESTROD to 1 instead of 0.
Incidentally, I've now got a full source tree, including kernel, compiled with
gcc-2.3.3 and -m486. There weren't that many changes needed maybe the next patchkit release should include them.
On a seperate note. Since updating was going to involve a lot of big compiles I decided
now would be a good time to investigate that "cache causing SIGBUS and SIGSEG" problem.
I tried compiling the kernel with external cache on and had to reboot several times
before it was finished. I tried a few times just to be sure it was a persistent
problem. Then I switched the external cache off and it compiled straight through
twice in succession. I then compiled the whole source tree without any problems twice
as well.
It seems pretty conclusive evidence that the cache is the cause of these SIG's, what I
want to know now is WHY?
I checked my cache chips and they're above spec. They're 20ns when the manual states
25ns are required. I really can't see how over spec chips could cause problems. Could
someone with more hardware experience shed some light on this?
In fact I can't see why the cache would be causing these problems in the first place.
Wouldn't the cache hardware deal with cache misses etc itself or is the cache hardware
just failing under the demands 386bsd is putting on it.
This is a really annoying problem that quite a number of people seem to be experiencing. I appreciate some discussion as to probable causes and any solutions we
may be able to adopt -- short of replacing my motherboard which is not an option for
me.
I'd consider replacing my cache chips but I don't want to go to my dealer and demand
SLOWER cache chips without some knowledge to back up my arguments (my machine's still
under guarantee).
---
Paul Richards, University of Wales, College Cardiff