*BSD News Article 19365


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!elroy.jpl.nasa.gov!decwrl!usenet.coe.montana.edu!bsd.coe.montana.edu!nate
From: nate@bsd.coe.montana.edu (Nate Williams)
Newsgroups: comp.os.386bsd.questions
Subject: Re: 386bsd 0.1 Patchkit 0.2.4 - Problems, Questions
Date: 10 Aug 1993 22:31:30 GMT
Organization: Montana Stateu University, Bozeman  MT
Lines: 112
Message-ID: <2497k2$qq2@pdq.coe.montana.edu>
References: <CBIpzw.6BM@csi.compuserve.com>
NNTP-Posting-Host: bsd.coe.montana.edu

In article <CBIpzw.6BM@csi.compuserve.com>,
M Foley <mfoley@csi.compuserve.com> wrote:
In article <CBIpzw.6BM@csi.compuserve.com>,
M Foley <mfoley@csi.compuserve.com> wrote:
>I just installed patchkit 0.2.4 to my 386BSD system. It seems to be
>working OK, but I have the following problems:
>
>1.  When I built a new kernel, the vers.o was still missing. I had to
>    run "/sys/conf/newvers.sh ; cc -c vers.c", by hand from
>    /sys/compile/FOLEY (my compile directory). I could then finish 'make'.
>    I also get the feeling that it is still incrementing by 2 since my
>    rev level now says 0.1.2. However, since I ran it by hand once, and I
>    saw is run again during the final make, this may not be true. Also, 
>    I do not see the machine name inside the '()' on the version title
>    line when the system boots. The parenthesis are empty. Why? Is this 
>    function still broken, or did I do something wrong?

Did you read the install notes?  It said to run buildworld.sh, and then
that script recommended that you do a make clean in all of your kernel
directories, (or simply remove them)

$ cd /sys/compile
$ rm -rf FOLEY
$ cd /sys/i386/conf
$ configure FOLEY 
(make sure that configure was re-compiled via the buildworld.sh script)
$ cd /sys/compile/FOLEY
$ make depend
$ make all
...

>3.  BIGGEST PROBLEM: When running buildworld.sh (for hours and hours), 
>    my system has stopped working twice now on 
>    "nroff -mandoc /usr/src/lib/libc/sys/intro.2 > intro.0"
>    It also hung once doing "... >csh.0". The system is not completely
>    dead since I can ^C and terminate the script. I notice that my

Your system is swapping itself to death, because nroff(groff) uses up lots
of memory due to a bug in g++1 that didn't free any memory.

>    "... /lib/libc" directory now has lots of '.0' files that probably
>    didn't get moved to the right place. Why is this procedure not finishing?

That's because a make install hasn't been run.

>    How can I get it to finish? Is there a way to restart without having
>    to do the whole thing over? 

Make will pick up where it's left off, so you can continue (leave it for
a week or so. :-) ) and it will keep going.

>    I am running a 80386-25 w/8087, 8MB of
>    memory and whatever default swap space tiny bsd put on my drive - I
>    believe 10000 blocks. I am using a 240MB partition on my 431MB Seagate
>    (15530 ?) hard drive. I am not running out of disk space. I did not get
>    an error in my afterinstall.OUT.

the 5 MB swap partition is not adequate for any machine with any version of
386BSD available publically.  

>4.  With the patchkit, I can now access my 2nd hard drive. BUT, I have a 
>    problem. I create a disklabel using "disklabel -R /dev/rwd1a 41mb" where 
>    '41mb' is a disklabel format file describing my 43MB Seagate 157A IDE 
>    drive. I then do "newfs -m 0 wd1" to create a file system - which works 
>    fine. I mount it, use it, etc. However, when I reboot and try to mount it 
>    again, I get a "BOGUS BOOT BLOCK" message and have to redo the label and 
>    recreate the file system all over. Why isn't the label/file system
>    information 'sticking'? Any ideas?

Check to make sure that the d partition has an offset of 0.

>5.  When I boot, I now get many "strayintr ff" messages on the console.
>    Fairly quickly, it says there are too may of these type of error and it
>    will stop logging them (I forget the exact wording). What are these?
>    should I be worried?

These are from the lp driver, it may be a bug or bad hardware.  (I'm
not sure which one it is)

>6.  I really want access to my new com2 (formerly com3), but I don't think
>    I can set my board to IRQ 5. Can this IRQ be changed in the 
>    /sys/i386/conf/MACHINE file? Can I set it to IRQ 4 like my com0 or
>    will this screw up com0 (i.e., must the devices be on separate IRQ's
>    or is the driver smart enough to daisy-chain)?

You must have a separate IRQ for every com port with the standard multi-
I/O cards in PC's.

>7.  There is now a directory and patch info for 'pcfs'. What is it? Is it
>    access to DOS partitions? Remote pcnfs-like stuff? What? Patch 101
>    claims to contain the man pages for pcfs, but I cannot "man pcfs". 
>    Perhaps my problem from #3 above has caused these pages to not be
>    installed.

Right on the money.

>
>8.  Question not related to the patch-kit: What should my '/etc/ttys' line 
>    say to specify a default com port baud rate of 13400? Do I need to make 
>    changes to my gettytab? Sticking in std.38400 as the getty argument didn't
>    work.

Read the man page on ttys, after you get it installed. :-)



Nate
-- 
nate@bsd.coe.montana.edu     |  In the middle of it ........ again. 
nate@cs.montana.edu          |  Running/supporting one of many freely available 
work #: (406) 994-4836       |  Operating Systems for [34]86 machines.
home #: (406) 586-0579       |  (based on Net/2, name changes all the time :-)