Subject: port-sparc/8430: sparc 1.4.1 boot.fs still has some problems....
To: None <>
From: None <>
List: netbsd-bugs
Date: 09/17/1999 11:06:42
>Number:         8430
>Category:       port-sparc
>Synopsis:       sparc 1.4.1 boot.fs still has some problems....
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    port-sparc-maintainer (NetBSD/sparc Portmaster)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Fri Sep 17 11:05:01 1999
>Originator:     John Refling
>Release:        1.4.1
this is a sun CLASSIC

I have tried boot-141.fs 'I)nstall' only and these are some
of the problems I have discovered:

1.  There is a difference in the number of sectors reported
    at boot up (in the system probe) and that used in the
    disklabel.  Why? and which is correct?

2.  When asking for the DNS domain name, it is not clear how
    to enter none.  Typing 'none' gives hostnames which look
    like 'name.none', using 'done' gives similar 'name.done'
    and using a <cr> and then '.' gives 'name..'.  The other
    ports seem to use a '.'.  Don't know if that works here
    but it should and should be documented on the prompt line.

3.  When prompted by "Configure additional media type arguments
    (none) [le0]", and I press <cr> I get a line in my /etc/
    ifconfig.le0 like: "inet ipc netmask 0xffffff00 media le0".
    This syntax error prevents the network from working.
    I think the "(none) [le0]" probably should be "(le0) [none]"
    and, correspondingly, elsewhere.

4.  I'm lucky enough to have 2 ethernet cards in this machine,
    but it is not clear during the install which one it will be
    used for ftp'ing the distribution.  Perhaps this is something
    I need to do when it asks if you need to break out for further
    network configuration, but I'm not sure.

5.  What is 10base5?  It must be the AUI port.  One of the
    ethernet cards has a AUI and BNC, however, it still reports
    as 10base5 only.

6.  When the installation fails for any reason (perhaps I just
    control-c'ed since I made a typo error, when restarted, it will
    go thru most of it again then fail near the end with "exec:
    mount_ffs -o async /dev/sd0a /mnt: device busy; FATAL ERROR:
    Cannot mount filesystems. Double check your configuration and
    restart the initialization process."  Of course, if you restart,
    the same thing happens over and over!  The bottom line is that
    mount_ffs should not be considered to have failed if the device
    is already mounted properly.  Note that newfs which might have
    the same issue fails gracefully by just saying "newfs: /dev/sd0a
    is mounted on /mnt", but the install script continues.  I have
    seen the same problem in the i386 port, too.