Removing options MIPS3 cause DS5000 to loop during boot

[Erik Bertelsen <bertelsen.erik%gmail.com@localhost>]

I have a DS5000/133 where I try to keep NetBSD running and self-buildable.

This has not been possible for the last year or so, but after recent
work on the mips architecture, the situation has improved somewhat.

I used to have a trimmed-down custom kernel that just supports the
hardware present in the machine, but this kernel does not work. A
current GENERIC kernel does, however, boot and then fails when going
multi-user, but that's another story about updating userland.

Now I have started from scratch and done several reductions from
GENERIC to see when it stops working.

If I remove support for non 3MIN machines like this:

 diff S2 S4
34,36d33
< options       DEC_3100        # DECstation 2100, 3100 (kn01)
< options       DEC_5100        # DECsystem 5100 (kn230)
< options       DEC_3MAX        # DECstation 5000/200 (kn02)
38,43d34
< options       DEC_MAXINE      # Personal DECstation 5000/xx (kn02ca)
< options       DEC_3MAXPLUS    # DECstation 5000/240 (kn03), 5000/260 (kn05)
<
< #options      DEC_5400        # DECsystem 5400 (kn210)
< #options      DEC_5500        # DECsystem 5500 (kn220)
< #options      DEC_5800        # DECsystem 58x0 (kn5800)
211,220d201
< # 3MAX has TURBOchannel but no IOASIC
< ibus0 at      tc? slot ? offset ?     # 3MAX system slot
<
< # PMAX, 3MAX and MIPSMATE common devices
< mcclock* at   ibus0 addr ?            # DS1287 RTC
< dz*   at      ibus0 addr ?            # DC7085 quad line serial device
<
< lkkbd*        at      dz? line 0              # Keyboard
< vsms* at      dz? line 1              # Mouse
<
236,241d216
< # PMAX and MIPSMATE
< ibus0 at      mainbus0
< pm*   at      ibus0 addr ?            # 3100 onboard fb
< le*   at      ibus0 addr ?            # 3100, 5100 onboard LANCE
< sii*  at      ibus0 addr ?            # DC7061 onboard SCSI
<
257d231
< wsdisplay*    at      pm?
264d237
< scsibus* at sii?

the machine can still boot to single user and give me a prompt where I can work.

If I further remove MIPS3 support:

 diff S4 S5
30d29
< options       MIPS3           # R4000/R4400 CPUs

then the machine boots to the point of starting init and then hangs
solidly -- the only recovery seems to power cycle the machine.

The full console output is:

KN02-BA V5.7j
>>boot 3/rz4/netbsd.s5

NetBSD/pmax 1.6ZC FFS Primary Bootstrap

NetBSD/pmax 5.99.40 Secondary Bootstrap, Revision 1.5
(root%sockdev.h.erikb.net@localhost, Sat Oct 10 23:31:25 CEST 2009)

Boot: 3/rz4/netbsd.s5
3056528+110568 [189424+184633]=0x360db8
Starting at 0x80030000

segment  0 start 00000000 size 02000000
phys segment: 0x2000000 @ 0
adding 0x46d000 @ 0x393000 to freelist 1
adding 0x1800000 @ 0x800000 to freelist 0
pmap_steal_memory: seg 0: 0x393 0x393 0x800 0x800
pmap_steal_memory: seg 0: 0x3b9 0x3b9 0x800 0x800
Copyright (c) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005,
    2006, 2007, 2008, 2009, 2010, 2011
    The NetBSD Foundation, Inc.  All rights reserved.
Copyright (c) 1982, 1986, 1989, 1991, 1993
    The Regents of the University of California.  All rights reserved.

NetBSD 5.99.49 (S5) #0: Thu Apr 21 17:21:51 CEST 2011
        
erik%x8-nb.h.recerik.dk@localhost:/home/NetBSD/crossbuild/obj.pmax/sys/arch/pmax/compile/S5
DECstation 5000/133 (3MIN)
total memory = 32768 KB
avail memory = 28304 KB
mainbus0 (root)
cpu0 at mainbus0: MIPS R3000A CPU (0x230) Rev. 3.0 with MIPS R3010 FPC Rev. 4.0
cpu0: 64KB/4B direct-mapped Instruction cache, 64 TLB entries
cpu0: 128KB/4B direct-mapped write-through Data cache
tc0 at mainbus0: 12.5 MHz clock
ioasic0 at tc0 slot 3 offset 0x0
le0 at ioasic0 offset 0xc0000: address 08:00:2b:32:56:6f
le0: 32 receive buffers, 8 transmit buffers
zsc0 at ioasic0 offset 0x100000
vsms0 at zsc0 channel 0
wsmouse0 at vsms0 mux 0
zstty0 at zsc0 channel 1
zsc1 at ioasic0 offset 0x180000
lkkbd0 at zsc1 channel 0
wskbd0 at lkkbd0 mux 1
zstty1 at zsc1 channel 1 (console i/o)
mcclock0 at ioasic0 offset 0x200000: mc146818 or compatible
asc0 at ioasic0 offset 0x300000: NCR53C94, 25MHz, SCSI ID 7
scsibus0 at asc0: 8 targets, 8 luns per target
scsibus0: waiting 2 seconds for devices to settle...
sd0 at scsibus0 target 4 lun 0: <QUANTUM, QM39100TD-S, N1B0> disk fixed
sd0: 8683 MB, 8057 cyl, 10 head, 220 sec, 512 bytes/sect x 17783249 sectors
sd0: sync (200.00ns offset 15), 8-bit (5.000MB/s) transfers, tagged queueing
boot device: sd0
root on sd0a dumps on sd0b
root file system type: ffs
pid 1(init): ABI set to O32 (e_flags=0x1007)


There is not difference in the console messages between the S4 and S5
kernels except that the 'avail memory' has increased slightly (because
of a smaller kernel?)

Can anyone explain this? I'll be happy to test patches to resolve this.

kind regards
- Erik