fox@CS.McGill.CA (Colin BRADLEY) wrote:

>		What is it I should have done so that the new kernel could
>		execute things? I can't remember exactly what the error 
>		messages were as they came up on the con during 'rc', and
>		'netstart', but I think it was something to the effect
>		that 'cannot execute /usr/libexec/getty' and a bunch of 
>		other stuff... (inetd, ...)

Did you put "options COMPAT_10" into your kernel config?  You MUST use
COMPAT_10 until you've completed the upgrade process (and even then
it's a good idea to keep it.)

To head off another potential problem:

cd /usr/src/etc/etc.i386
install -c -m 544 MAKEDEV /dev
cd /dev
./MAKEDEV all

>	At this point I am back on my 1.0 kernel, and am more or less
>	manually going through the source tree seeing what will build 
>	and what will not...
>
>	Everything up till /usr/src/lib/libc seems to go. When it gets
>	to 

>install -c -o bin -g bin -m 444  libc.so.12.0 /usr/lib
>install -c -o bin -g bin -m 444 btree.cat3 /usr/share/man/cat3/btree.0
>Bus error - core dumped
>*** Error code 138
>
>Stop.
>*** Error code 1
>
>Stop.

Installing libc.so when you're not running an up-to-date kernel is
always a Bad Idea.  Every dynamically linked binary in your system
(usually everything but /bin/* and /sbin/*) will break.

>		Did I have to be running the new kernel for this to
>		work properly? If so, what was the problem with the 
>		kernel that didn't allow things to function? If not,
>		is the problem libc.so.12.0? Is it a dynamic linking 
>		problem? (I'm kinda fuzzy on this area...)

This will happen whenever your installed libc.so is newer than your
kernel.  Something changed in the syscall interface (which is why
COMPAT_10 exists), so the new libc won't work with the old kernel.

>                 -chewing the same piece of gum since '77
That's carrying economy a bit too far, I think. :-)

Back to work...
-- 
Mike Long <mike.long@analog.com>       PGP 2.6.2 public key signature:
VLSI Design Engineer                  CCBF225E7D3F7ECB2C8F7ABB15D9BE7B
Analog Devices, CPD Division
Norwood, MA 02062 USA                assert(*this!=opinionof(Analog));