Ultra 30 booting problems

Rob Craig

unread,

Feb 15, 2021, 8:19:51 AM2/15/21

to

Hi guys,
I have an ultra 30 and I can't get it to boot.

Keyboard LEDS indicate on startup: CAPS LOCK: ON COMPOSITE: OFF SCROLL LOCK: ON NUM LOCK: ON.
This shows in the manual: "reserved" (I've got no idea what that means for the POST)

If I connect a serial cable to see the post I get:

@(#) Sun Ultra 30 UPA/PCI POST 1.1.1 03/04/97

CPU: UltraSPARC 2 (MHz: 296 MID: 0 Ecache Size: 2048KB)
Init System BSS
STATUS =FAILED
Unexpected event occurred - Trap
tl tt tstate tpc tnpc
01 34 00000044.80001404 ffffffff.f0088910 ffffffff.f0088238
DMMU SFSR 00000000.0015007b
DMMU SFAR ffffffff.f00e03f2
(FV) Fault Valid Bit Set
(OW) Overwrite Bit Set
(ASI) 0x15
(E) Side effect Bit Set
(CT) Context 3
(PR) Privilege Bit Set
IMMU SFSR 00000000.0015007b
(FV) Fault Valid Bit Set
(ASI) 0x84
(FT) Privilege Violation
(CT) Context 2
AFSR 00000001.a45800fd
AFAR 000001ff.f0000210
(ME) Multiple Errors
(PRIV) Privileged Code
(ETP) Ecache Tag Parity Error, (ETS) Syndrome = 8
(BERR) Bus Error
(EDP) Ecache Parity Error
(CE) Correctable ECC Error
(P_SYND) Ecache Parity Syndrome = 00fd
SDBH = 00000000.00000000 SDBL = 00000000.000001f2
DEMON Level 1
Select one of the following functions
0 Run POST and return to OBP
1 Demon Level 2
2 Initiate Soft Reset
3 Peek/Poke device
4 CPU Basic Menus
5 FPU Basic Menus
6 MMU Basic Menus
7 Cache Basic Menus
8 Ecache Basic Menus
9 Psycho Basic Menus
a SC Basic Menus
b NVRAM/TOD Basic Menus
c UPA Freq Margin
d Toggle Test Flags

Any ideas? I've replaced the CPU, I've tried moving and reseating the RAM and video card.
Thanks,

chris

unread,

Feb 15, 2021, 2:44:23 PM2/15/21

to

Sounds like it could be a ram fault, quite common on older machines,
especially if they have been stored in less than ideal conditions. Try
swapping the ram around, starting with a minimum count supported.

Also try removoing and reseating cpu modules and if two, try one at
a time...

Chris

Rob Craig

unread,

Feb 17, 2021, 5:48:14 AM2/17/21

to

Thanks for the reply Chris. I've been moving the RAM around and testing the banks. If I hold down STOP on the power on to bypass the POST check where it errors out straight away I see the RAM is detected in each back I put it into:

@(#) Sun Ultra 30 UPA/PCI 3.9 Version 5 created 1997/04/11 10:03
Probing keyboard Done

Skipping POST because of L1 keyboard command.

Clearing E$ Tags Done
Clearing I/D TLBs Done
Probing Memory
SIMM population : 0000.0000.0000.0002
SIMM esize info : 0000.0000.0000.0002
SIMM msize info : 0000.0000.0000.0002
MEM BASE = 0000.0000.0000.0000
MEM SIZE = 0000.0000.0400.0000
Memory interleave: Disabled
MMUs ON
Copy Done
PC = 0000.01ff.f000.262c
PC = 0000.0000.0000.2670
Decompressing into Memory Done
Size = 0000.0000.0007.3950
ttya initialized
SC Control: EWP:0 IAP:0 FATAL:0 WAKEUP:0 BXIR:0 BPOR:1 SXIR:0 SPOR:0 POR:0
Starting real time clock...
Incorrect configuration checksum;
Setting NVRAM parameters to default values.
Setting diag-switch? NVRAM parameter to true
Probing Memory Bank #0 32 + 32 : 64 Megabytes
Probing Memory Bank #1 0 + 0 : 0 Megabytes
Probing Memory Bank #2 0 + 0 : 0 Megabytes
Probing Memory Bank #3 0 + 0 : 0 Megabytes
Probing Memory Bank #4 0 + 0 : 0 Megabytes
Probing Memory Bank #5 0 + 0 : 0 Megabytes
Probing Memory Bank #6 0 + 0 : 0 Megabytes
Probing Memory Bank #7 0 + 0 : 0 Megabytes
Probing Floppy: drive detected on ID0
Probing EBUS SUNW,CS4231
Probing UPA Slot at 1e,0 SUNW,ffb
Probing UPA Slot at 1d,0 Nothing there
Probing /pci@1f,4000 at Device 3 scsi disk tape
Probing /pci@1f,4000 at Device 2 Nothing there
Probing /pci@1f,4000 at Device 4 Nothing there
Probing /pci@1f,4000 at Device 5 Nothing there
Probing /pci@1f,2000 at Device 1 Nothing there
Probing /pci@1f,2000 at Device 2 Nothing there

So I've found two of the SIMMS give a bad parity 'l'' error on the boot up so I removed those and try booting with 2 simms only. The above screen is where it gets too. But this is obviously holding down STOP to by pass the POST test. IF I boot normally I still get the same POST error as above even with the "bad parity" ram removed.

chris

unread,

Feb 18, 2021, 8:16:21 AM2/18/21

to

Not sure about the Ultra 30, but if the idprom battery is duff, the
machine won't have a valid mac address and it won't pass post. The fix
is not too difficult, but needs a Dremel and soldering skills to fix
it. Usually takes me about an hour to do that. Loads of info on the
web, youtube etc, but alternatively, there's an Ebay US seller
offering replacements. Try a search on sun idprom.

The Ultra 30 was a fairly hot machine back in the day, well worth
fixing up, so don't give up...

Chris

Rob Craig

unread,

Feb 18, 2021, 9:08:36 AM2/18/21

to

Yeah thanks for the support. I definitely want to restore it. I have done the dremel job on a Dallas RTC for a PC. A bit of a hassle but worked. I do think its a motherboard issue as I purchased a IDPROM from the US a few years back as the original ones battery was dead. I remember that even having a dead battery I could still boot the machine and at least type commands in the open manage screen. I changed the MAC address via the commands to match the new IDPROM and I managed to get one or two boots then the problems started. It would not boot. I think stuck it into storage for a few years and pulled it out the other day to troubleshoot it. Hence where we are now.

Both IDROMS (the old and “new”) report the same “Contents are invalid” error. I guess the “new” one’s battery has died too?

chris

unread,

Feb 18, 2021, 12:13:36 PM2/18/21

to

The only way to verify the idprom is to reload the relevant bytes,
incl the checksum, then cycle power, reset and see if the settings
are still there.

I've been working through some of my Sun collection over the past
few months and probably done a dozen or more iproms. Started by
epoxying batteries on top the idprom, but current method is to epoxy
a coin cell battery holder onto the end of the package, so that the
battery is replaceable, even with power on. On some sbus machines,
height is limited, so fit onto the end is the only solution and
even that varies with machine type. Using CR2032 cells now, much
higher capacity than original, which should give > 10 years life,
probably more than the remaining life of the machine.

Good luck with it anyway. If you've got that far, should be able
to finish it. The sunhelp.org and the classiccmp.org mailing lists
are a good place to keep an eye on as well...

Chris

Rob Craig

unread,

Feb 19, 2021, 6:55:07 AM2/19/21

to

Ok thanks Chris. I've just dug out my other Sun workstations too for cleaning and fixing. I've got a SunBlade 1000 Dual Sparc III+ and another Sunblade 1000. It's going to be a Sun SPARC month. LOL