Hi,
	Just thought I'd let everyone know what happened with my trials and
tribulations vis a vis Silverlining and "Mkfs".

Jesse Byler wrote:
> Secondly, why can't you use APS to format your NetBSD (AUX) partitions, and
> use whatever other formatting utility on the MacOS partition?  For that
> matter, you could use APS for now, and if it turns out the driver isn't
> compatible with the new SCSI manager, you can replace the driver with a new
> one.  (You don't have to reformat the partition just to install a new driver).

I ended up using APS 2.7.3 and reformatted the disk, then partitioned it
with APS and installed the driver that came with APS.  My only complaint with
this APS version is that it insisted on creating a 3 Mb (minimum allowed)
Eschatology partition and a 0-length Free partition, but no big deal.
The important thing was that it allowed me to create the A/UX 2.0 Root&Usr
partition and the A/UX Swap and then the HFS Mac partition last (on the disk).

I then rebooted off of the CD-ROM and did a bare-bones 7.5.3 install with
no customizations/additions (no sense doing that when you're getting blown
away by "Mkfs" everytime (-: ).

Having done this, I got curious: what if I tried to boot it?  I thought it
would die because APS 2.7.3 is so old; I was sure it wouldn't live happily
with SCSI Manager 4.3.  Well, it booted anyway, and I've had no problems
since then, so I can't really explain it.  Anyway, my disk drive is now
Silverlining-free  (-:

Once I was up and running, it was time for the Big Test: Mkfs.  Well, I ran
it, it ran, it said no errors, I did a "Restart" and crossed my fingers.
Damned if it didn't come up!  No corruption!  Hallelujah!

Bob Nestor tried to explain my earlier "Mkfs" corruption:
> Mkfs has a problem addressing any disk block above a certain point.  It 
> only uses a 6-byte CMD for talking to the SCSI Manager, which limits it to 
> 21 bits for block address.  When it exceeds this, it happily truncates the 
> numbers causing them to wrap down to lower numbered blocks.  So if in you 
> example you tried to run Mkfs on the 900Meg Usr Partition, it would have 
> clobbered the MacOS Partition and maybe the Driver and Partition Map as 
> well.  Since you got a "?" in the reboot, it's safe to say you luckily 
> didn't zap the Driver Partition.

While this is a perfectly cogent and reasonable observation, I don't think
it was the case in my situation.  I used the latest Mkfs from the stuff in
the NetBSD-1.2/mac68k directory on ftp.NetBSD.ORG, and as you can see from
the above, after having partitioned with APS I was able to successfully
Mkfs the partition.  And at this point in time, surely there are people in
NetBSD/mac68k land who are using Mkfs'd NetBSD partitions larger than 1 Gb?
(Also, in my case, the partition size is actually something like 900-some-odd
 Mb, i.e. < 2 ^ 21 512-byte blocks.)

Anyway, the bottom line is that from there I was able to get NetBSD 1.2
installed and the system is up and running:

netbsd4me:1:38 [/home/earle] % rup macbsd4me
macbsd4me                up            23:51,  load average: 2.20 1.60 1.05

At this point, I'm having two problems of note with the system.  One minor,
one serious:

(1) I'm using "dt" (1.1.4) and I change the font size to be larger (is there
    any way to set the default to be larger without having to recompile it?).
    But when console messages come through, they blast through "dt" in the
    default small console font.  Messes up the screen a bit.  Any way to avoid
    this?

(2) Unfortunately, I'm having *serious* problems with NFS mounting.  I'm using
    NIS and "amd", and using the same AMD maps as I do on my NetBSD/SPARC 1.2
    system.  I've found that I could NFS mount my home directory from the
    NetBSD/SPARC system just fine (I think ... it worked on Friday, but right
    now I'm having problems).  But *any* attempt to mount other users' home
    directories from a Solaris 2.4 SPARCserver 1000 results in hangs.

    The amd process spawns a sub-process to do the mount, which promptly gets
    stuck in "D" state (WCHAN of "netio").  I watched the network traffic and
    found that the Mac makes the initial portmap request, then when the
    Solaris server says "AUTH_NULL success" in return (back to port 1021 on
    the Mac - using priviledged ports), the Mac sends back an ICMP Destination
    Port Unreachable to the SPARCserver!!!   Aieee!  This doesn't happen on
    the NetBSD/SPARC machine, which is happy to either amd automount the very
    same things from the SPARCserver, or do a manual "mount" which works just
    fine as well.  But on the Mac it hangs.  )-:

    I'd give more details but right now I can't get onto the Mac for the same
    reason (apparently), but now it's affected my own home directory mount :-(
    (And I forgot to allow root logins/ftp over the network, thinking that
     since my own login was working, I didn't need to do that yet.  Sigh.)

netbsd4me:1:39 [/home/earle] % telnet macbsd4me
Trying []...
Connected to macbsd4me.
Escape character is '^]'.

NetBSD/mac68k (macbsd4me) (ttyp1)

login: earle
Password:

(Hangs here)

Anyway, thanks to everyone for their advice, I'm up and running.  Now if only
I could solve this NFS hang problem ... (I will followup with more details
when I can get back onto the machine  )-: )

	- Greg

P.S. This IIci only has the internal video, which I'm running in 1-bit mode
     for dt's sake.  Can anyone tell me what NuBus video cards I could get
     that would allow me to run the Xmacbsd "colorkit" that's available in
     ftp://ftp.MacBSD.COM/pub/NetBSD/X/colorkit/ ???  (The README.colorkit
     in that directory implies that the Xmacbsd found there won't work on
     the IIci's internal video; of course, that document is 5 months old now)