I have been using NetBSD/SPARC for nearly a year now, but have never had a
chance to try out the "mothership" version before (-:

I have a SPARCstation and a Mac at home, and at work, workstations rule the
roost.  PCs are a very scarce commodity around the places I haunt.

Anyway, there is one PC around - a 486DX/33 - and recently it was upgraded
with hardware that is now very NetBSD-friendly.  I found out from the user
that it now has a Seagate 1 Gb SCSI disk, and gee, he was only using the
first 210 Mb for DOS/Windows/etc.  And I was more than welcome to swallow the
remaining 790 Mb for my evil and nefarious purposes.  Bwah hah hah!

So, I tried to start my first-ever NetBSD/i386 install, with the 1.0 dist.
Not being a PC type, I have a few questions about things I ran into.

Here is the hardware that I know about:
	486DX/33		Adaptec 1542CF SCSI controller
	AMI BIOS 06/06/92	Seagate ST31200N 1 Gb (1006 Mb) SCSI disk
	16 Mb RAM		Toshiba XM-3401TA internal CD-ROM drive
	1.44 Mb 3.5" floppy	ATI Mach 32 video card
	1.2 Mb 5.25" floppy	Novell NE2000 Ethernet card

I have three basic questions:

(1) I ran "pfdisk" and it said this:

    pfdisk> L
    #Partition table on device: 0
    geometry 1006 64 32 (cyls heads sectors)
    #	 ID	First (cyl)	Last (cyl)	Name	# start, length
    1	  1	    0		    9		DOS12	# 32, 20448
    2	  5	   10		  209		DOSex	# 20480, 409600
    3	  0	    0		    0		empty	# 0, 0
    4	165	  210		 1004		unkno	# 430080, 1628160

    When I booted off the INST-10.FS floppy, it said

    aha0 targ 0 lun 0 : <SEAGATE ST31200N	8630>  SCSI 2 direct fixed
    sd0 at scsibus0: 1006 MB, 2700 cyl, 9 head, 84 sec, 512 bytes/sec

    The "pfdisk" man page (.doc) says that it gives cylinders, heads and
    sectors, but this "1006" above is the disk size in Mb, not the # of
    cylinders.  Is the man page (doc) wrong?  Is "pfdisk" really supposed to
    report in Mb units instead?

    If the NetBSD driver is right, the NetBSD partition starts on a head
    boundary (430080 = 512 * 84 * 10), but not on a cylinder boundary
    (512 * 84 * 9 * N).  Will this be a problem?

(2) When I tried to boot off of INST-10.FS for the first time, I got

------------------------------------------------------------------------------
NetBSD 1.0 (GENERICAHA)  #3: Sun Oct 23 20:58:04 PDT 1994
CPU: i486DX (486-class CPU)
real mem: = 16646144
avail mem = 14589952
using 228 buffers containing 933888 bytes of memory
pc0 at isa0 port 0x60-0x6f irq 1 : color
fdc0 at isa0 port 0x3f0-0x3f7 irq 6 drq 2
fd0 at fdc0 drive 0:  1.44 MB 80 cyl, 2 head, 18 sec
fd1 at fdc0 drive 1:  1.2  MB 80 cyl, 2 head, 15 sec
aha0: 1542C/CF detected, unlocking mailbox
aha0 at isa0 port 0x330-0x333 irq 11 drq 5
scsibus0 at aha0
aha0 targ 0 lun 0: <SEAGATE  ST31200N	8630>  SCSI 2  direct fixed
sd0 at scsibus0: 1006 MB, 2700 cyl, 9 head, 84 sec, 512 bytes/sec
aha0 targ 4 lun 0: <TOSHIBA CD-ROM XM-3401TA1094>  SCSI 2  read only removable
cd0 at scsibus0: drive empty
------------------------------------------------------------------------------

    At this point, the boot hard hangs.

    For some reason, I wondered if it was unhappy that there wasn't any CD-ROM
    in the drive.  So I tossed an ISO 9660 CD-ROM into a caddy and shoved it
    in, and power-cycled the machine.  This time it worked:

------------------------------------------------------------------------------
NetBSD 1.0 (GENERICAHA)  #3: Sun Oct 23 20:58:04 PDT 1994
[ ... same as before ... ]
aha0 at isa0 port 0x330-0x333 irq 11 drq 5
scsibus0 at aha0
aha0 targ 0 lun 0: <SEAGATE  ST31200N	8630>  SCSI 2  direct fixed
sd0 at scsibus0: 1006 MB, 2700 cyl, 9 head, 84 sec, 512 bytes/sec
aha0 targ 4 lun 0: <TOSHIBA CD-ROM XM-3401TA1094>  SCSI 2  read only removable
cd0 at scsibus0: cd present, 165987 x 2048 byte records
ie0: Unknown AT&T board type code 15
npx0 at isa0  port 0xf0 - 0xff : using exception 16
biomask 840 netmask 2 ttymask 2
changing root device to fd0a
WARNING: no swap space found
/etc/rc: Can't open /etc/rc

Enter pathname of shell or RETURN for sh:
------------------------------------------------------------------------------

    Should I have expected the boot to hang if no CD-ROM is in the CD drive?

(3) The INSTALL says that the NE2000 is supported; why does the boot probe
    identify it as an "Unknown AT&T board type"?  Looking at if_ie.c and
    if_ed.c, it appears that the system thinks that the Ethernet board is
    Intel i82586-based, not NSC DS8390 or WD83C690-based?  That's odd, since
    when DOS boots it says:

Novell NE2000 Ethernet MLID v1.34 (910603)
Int 5, Port 340, Node Address 1B3A114C
Max Frame 1514 Bytes, Line Speed 10 Mbps
Board 1, Frame ETHERNET_II
Board 2, Frame ETHERNET_802.3

    So I would have thought it would have shown up as an "ed0", not an "ie0".

    Why did the kernel see an ie0?  Or is the kernel seeing something *else*
    living at port 0x360 and irq 7 and decide it was an alien bogon
    StarLAN card?  (-:  I am now worried that I won't be able to load the main
    distribution from a Sun over the net until/unless I have a custom kernel
    with "device ed0 at isa? port 0x340 irq 5 iomem 0x?????" in it.  Yes?

    I also note looking at the various config files that different machines
    use different ports and iomem addresses for these "ed0" boards; is there
    any easy way to tell what values one should use (perhaps from DOS, given
    where I'm at at the moment) under NetBSD?  I assume the above tells me that
    it's at port 0x340 and irq 5, but I have no idea what the iomem address
    should be (nor the iosiz parameter).

I apologize for the length; I am trying to get all of my questions answered
at once so I can get this install rolling ... Thanks for any answers.

(Please at least Cc: me, as I'm not on "i386-users" yet - at least, not 'til
 I get it installed and booted (-: )

	- Greg