*BSD News Article 6861


Return to BSD News archive

Path: sserve!manuel.anu.edu.au!munnari.oz.au!sgiblab!sgigate!olivea!uunet!news.univie.ac.at!alijku11!k390670
From: K390670@ALIJKU11.BITNET
Newsgroups: comp.unix.bsd
Subject: Re: Compiling kernel locks machine (was: takes a long time)
Message-ID: <92295.112907K390670@ALIJKU11.BITNET>
Date: 21 Oct 92 11:09:13 GMT
Organization: Johannes Kepler University Linz - Computing Center
Lines: 43

Similar to some other people who have posted about this problem, the
following happens on my machine (486-50 with 212 MB Conner, 16MB RAM):

When I install the sources, change into /usr/sys and do a "make all",
the machine happily compiles up to a certain point. Then hard disk
activity stops, but the machine isn't completely frozen either:
typing characters on the keyboard produces console echo.

Sometimes <CTRL-C> will break into the shell prompt under such
circumstances, and 1 (in words: one) more program (like pwd, ls)
can be run.  After that, just console echo works (and *nothing* else).

Sometimes <CTRL-C> doesn't work as soon as the hard disk activity is
stopped while the machine is running gcc.

Sometimes gcc stops with the message: Signal 6 received. Aborting.
(not literally, but it is signal 6). (Of course, I don't kill it, it
just happens.)

The 3 forms of the problem occur at random times during make all,
but can be reproduced for /usr/src/lib/libg++/libg++/Integer.cc
(which means that I don't ever get past Integer.cc.)

Integer.cc is a rather long program (47K source), but not abnormally
so.

As I just did "install" on the entire disk, I *don't* know if I
have any swapspace configured, but 16MB RAM should really suffice.

So far, I don't have any patches installed (this recompile was
supposed to be a test run...)

Any suggestions are highly welcome.

I start thinking of dumping 386BSD and trying LINUX instead. I mean,
I can fix bugs myself, but if a kernel recompile fails, I get into
some sort of "hen and egg" problem...

------------------------------------------------------------------------------
Gregor Glawitsch                         I  Everybody should believe in
Institute of Microcomputer Applications  I  something - I believe I should
Johannes Kepler Univ. Linz, Austria      I  have another beer...
------------------------------------------------------------------------------