Subject: Re: sysinst
To: Frederick Bruckman <fredb@immanent.net>
From: Bob Nestor <rnestor@augustmail.com>
List: port-mac68k
Date: 06/26/2002 18:49:37
On Tuesday, June 25, 2002, at 09:49 PM, Frederick Bruckman wrote:

> On Sat, 22 Jun 2002, Bob Nestor wrote:
>
>> Patches and Installaton Kernels are available at:
>>
>>     ftp://murphy.dyndns.org/pub/sysinst_1.6
>>
>> Please let me know if you have problems with the patches or the Install
>> Kernels.
>
> I built a 1.6B snapshot with your patches, and did three installs to
> an IDE drive after installing drivers with either FWB HD Toolkit or
> Apple Drive Setup, including a final setup with a full-house (a - h
> filled). All the previously reported problems seem to be gone! Bravo!
>
> Also tested nfs and ftp. Both fine.
>
> There are a couple of minor problems with the fstab generation:
>
> 1) The fs_passno field is '2' for root and '1' for all others, when it
> should be the other way 'round.

The handling for this is in the MI portion of the code and it appears to 
be correct.  Is it possible you were seeing this in a configuration that 
had other problems?  You might want to look at the code in 
"src/distrib/utils/sysinst/disks.c", the routine fsck_num() which is 
used to generate the fs_passno field while it is building the fstab.  If 
you see an error in there I'll fix it for the next round of Installation 
kernels, but it looks correct to me and it worked OK on my system 
installation tests.

> 2) I made one partition of type "Usr" but didn't mount it. This led
> to an fstab entry with a missing field, which led to dire warnings at
> first boot. "sysinstall" should have just omitted that line.

This appears to be an oversight in the MI code that builds the fstab.  
I've modified the code so that it generates a commented line in 
/etc/fstab for partitions that don't have a mount point specified.  If 
could be that other ports force the user to specify a mount point for 
all partitions where the mac68k port doesn't though.

I'm testing the changes right now and will start an Installation Kernel 
build shortly with the DHCP change Scott asked for.  The new Kernels 
will be posted in the morning along with the patch files.

-bob