Sun

Ultra Enterprise 2 - boot issues

So I picked up an Ultra Enterprise 2 at the recycle center today, along with a couple SBus cards (they got an assload of Sun stuff in last Friday, and I missed getting a nice modern SPARC tower by about an hour - they were just finishing tearing it apart when I walked in.) It's a dual 300MHz UltraSPARC-II system with 1GB RAM and a 75MHz Creator 3D series 1 display adapter, and I supplemented it with a SunSwift 100Base-T/SCSI card and a Turbo GX accelerated framebuffer card. The only problem is, it's not quite working. There's no video output at all, on either card. The serial console comes up fine, which is encouraging, though it insists on going through the full POST every boot, I think because the NVRAM battery seems to be dead. The big problem is that I can't get it into OBP at all; the POST completes and it claims to be entering OBP, but it never goes to a prompt, and sending a break from the terminal doesn't get one either. Here's the logfile I took:

Code: Select all

Hardware Power ON
Master CPU online
Slave-Offline!
Slave Offline, disabled

Button Power ON
Master CPU online
Slave-Offline!
Slave Offline, disabled

Probing keyboard Done
%o0 = 0000.0000.0055.4001

Executing Power On SelfTest

0>@(#)  Sun Ultra Enterprise 2 FCS_POST, version SB3.3.8 11/16/1998 1:42 PM
0>   UltraSparc2 Version 2.0
0>Ecache Probe
0>   Ecache size 2048 Kb
0>Ecache Tag Test
0>Ecache RAM Test
0>Ecache Address Line Test
0>Initialize and Verify Ecache
0>SC Initialization
0>    SC_MP id=acf1, UPA Number=4, Impl=0, Ver=3
0>SC Dual Tag RAM Test
0>    Clearing DTAG's.
0>Initialize SC_MP memory control registers
0>BMX Test
0>    Checking BMX's
0>Probing Memory
0>   Found Memory Group #0      64Mb   64Mb   64Mb   64Mb
0>   Found Memory Group #1      64Mb   64Mb   64Mb   64Mb
0>   Found Memory Group #2      64Mb   64Mb   64Mb   64Mb
0>   Found Memory Group #3      64Mb   64Mb   64Mb   64Mb
0>            Found 1024 Megabytes of usable Main Memory
0>SIMM Group      Base Addr         Size    Group Status
0>   0         00000000.00000000   10000000     00
0>   1         00000000.20000000   10000000     00
0>   2         00000000.40000000   10000000     00
0>   3         00000000.60000000   10000000     00
0>Quick Memory Test
0>Clear and Test Stack Memory
0>
SelfTest Initializing
0>Basic CPU Test
0>    Instruction Cache Tag RAM Test
0>    Instruction Cache Instruction RAM Test
0>    Instruction Cache Next Field RAM Test
0>    Instruction Cache Pre-decode RAM Test
0>    Data Cache RAM Test
0>    Data Cache Tags Test
0>MMU Enable Test
0>    DMMU Registers Access Test
0>    DMMU TLB DATA RAM Access Test
0>    DMMU TLB TAGS Access Test
0>    IMMU Registers Access Test
0>    IMMU TLB DATA RAM Access Test
0>    IMMU TLB TAGS Access Test
0>    DMMU Init
0>    IMMU Init
0>    Mapping Selftest Enabling MMUs
0>FPU Register Test
0>    FPU Registers and Data Path Test
0>    FSR Read/Write Test
0>EPROMs Test
0>    PROM Datapath Test
0>Serial Ports Test
0>    Slavio Serial Ports Test
0>NVRAM TOD Test
0>    M48T59 (TOD) Init
0>    M48T59 (TOD) Functional Part 1 Test
0>Memory Test
0>    Memory Clear Test
0>   Test being relocated into Memory
0>    Memory RAM (blk) Test
0>   Test being relocated into Memory
0>    Memory Stress Test
0>   Test being relocated into Memory
0>    Memory Address Line Test
0>   Test being relocated into Memory
0>Forcing ECC Faults Test
0>    ECC CE Pattern Test
0>    ECC CE Check bit Test
0>    ECC UE Pattern Test
0>    ECC UE Check bit Test
0>SysIO Registers Test
0>    SysIO Regsiter Initialization
0>    IOMMU Registers and RAM Test
0>    Streaming Buffer Registers and RAM Test
0>    SBus Control and Config Registers Test
0>    SysIO RAM Initialization
0>SysIO Functional Test
0>    Mapping Selftest Enabling MMUs
0>    Clear Interrupt Map and State Registers
0>    SysIO Interrupts  Test
0>    SysIO Timers/Counters Test
0>    IOMMU Virtual Address TLB Tag Compare Test
0>    Streaming Buffer Flush Test
0>    DMA Merge Buffer Test
0>CPU Speed
0>    CPU 0 Running at 296 MHZ.
0>Ecache Stress Test
0>    Ecache Stress Test
0>APC Test
0>    APC Registers Tests Test
0>    APC DVMA Test
0>Data Cache Test
0>    Dcache Init
0>    Dcache Enable Test
0>    Dcache Functionality Test
0>FEPS Test
0>    Parallel Port Registers Test
0>       Parallel Port ID is: 0x2
0>    Parallel Port DVMA burst mode read/write Test
0>    FAS366 Registers Test
0>    ESP FAS366 DVMA burst mode read/write Test
0>    FEPS Internal Loopbacks Test
0>    Ethernet Tranceiver Internal Loopbacks Test
0>CPU Functional Test
0>    Mapping Selftest Enabling MMUs
0>    SPARC Atomic Instructions Test
0>    CPU Dispatch Control Register Test
0>    CPU Softint Registers and Interrupts Test
0>    CPU Tick and Tick Compare Registers Test
0>    Uni-Processor Cache Coherence Test
0>      Base_address = 00000000
0>    UltraSPARC-2 Prefetch Instructions Test
0>    SRAM Mode = 22, Clock Mode = 3:1, ELIM = 3, PCON = 0f7, MCAP = 13
0>    Ecache Size Limited to 00200000; 000006b6.3dc0803b
0>
<< POST COMPLETE >>
0>**Entering OBP (3b)


Power On Selftest Completed
Status  = 0000.0000.0000.0000 0000.01ff.f007.87f0 1f66.0000.01c1.0105


Software Power ON
Master CPU online
Slave-Offline!
Slave Offline, disabled

@(#) Sun Ultra 2 UPA/SBus 3.25 Version 0 created 1999/12/03 11:35
Clearing DTAGS  Done
Probing Memory Done
MEM BASE = 0000.0000.6000.0000
MEM SIZE = 0000.0000.1000.0000
MMUs ON
Copy Done
PC = 0000.01ff.f000.2c60
PC = 0000.0000.0000.2ca4
Decompressing into Memory Done
Size = 0000.0000.0006.70e0
ttya initialized
SC Control: EWP:0 IAP:0 FATAL:0 WAKEUP:0 BXIR:0 BPOR:0 SXIR:0 SPOR:1 POR:0
Incorrect configuration checksum;
Setting NVRAM parameters to default values.
Setting diag-switch? NVRAM parameter to true
Probing Memory Bank #0  64  64  64  64 : 256 Megabytes
Probing Memory Bank #1  64  64  64  64 : 256 Megabytes
Probing Memory Bank #2  64  64  64  64 : 256 Megabytes
Probing Memory Bank #3  64  64  64  64 : 256 Megabytes

And that's where it hangs. Any idea as to what's going wrong here?
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/MT-32/D-10, Oberheim Matrix-6, Yamaha DX7/FB-01, Korg MS-20 Mini, Ensoniq Mirage/SQ-80, Sequential Circuits Prophet-600, Hohner String Performer

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup
Update: figured it out. One of the CPUs is either bad or a mismatch. (Edit: it's bad and a mismatch.) When I remove it and boot it in single-processor mode, it works fine - except for the part where evidently a paranoid sysadmin sledgehammered both hard drives (but carefully reinstalled them!) before sending it off to the recycle center! Guess I'll have to rustle up a couple new hard drives (and maybe a replacement slave CPU...)

Still, not a bad haul at all. Even a single 300MHz UltraSPARC II is going to be a serious boost over my Ultra I, and having dual-head capability and 100Mbit Ethernet is just icing on the cake.
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/MT-32/D-10, Oberheim Matrix-6, Yamaha DX7/FB-01, Korg MS-20 Mini, Ensoniq Mirage/SQ-80, Sequential Circuits Prophet-600, Hohner String Performer

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup
Them boys is drive punchers alright. Whatcha gon' do with it once you've got it running to perfection?
Project:
Temporarily lost at sea...
Plan:
World domination! Or something...
Probably gonna throw OpenBSD on it and hook it up to a SANE-supported SCSI scanner I got from the recycle center a while back - should be a little nicer than my current scanning solution, a Win9x-only SCSI-over-parallel-port scanner hooked up to a PIII laptop.
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/MT-32/D-10, Oberheim Matrix-6, Yamaha DX7/FB-01, Korg MS-20 Mini, Ensoniq Mirage/SQ-80, Sequential Circuits Prophet-600, Hohner String Performer

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup
My scanner's hooked up to a circa-1990 Pentium 90 running WinNT 4.0 service pack 6, for the simple reason that I found HP's color correction software to be way better than anything in SANE or the Gimp, but then last time I checked was probably over 10 years ago, so they've probably caught up by now... :P
Project:
Temporarily lost at sea...
Plan:
World domination! Or something...
vishnu wrote: Them boys is drive punchers alright. Whatcha gon' do with it once you've got it running to perfection?


I've seen drive punchers and box bashers.. makes me want to smash em up good.. you can't blame the hardware
and sometimes the software is fine too, the real issue is PEBCAK . :mrgreen:
MAYA, nut-
:Octane2: :Octane2: Octane 2 R14k 600 V12 4GB, Octane2 R14K 600 V10 1GB ,
:Onyx2: :Onyx2: Onyx2 IR3 4GB Quad R14K 500 DIVO, Onyx2 IR Quad R12K 400 2GB,
:Indigo2: SGI Indigo 2 R8K75 TEAL Extreme 256MB,
:Indigo2IMP: SGI Indigo 2 R10K 195 Solid Impact 256MB, MAX Impact Pending
,
Apple G5 Quad, NV Quadro 4500 + 7800GT, 12GB RAM
Sun Blade 1000 Dual 900 XVR 1000 4GB
Sun Blade 2000 Dual 1200 XVR 1200 8GB
commodorejohn wrote: Probably gonna throw OpenBSD on it and hook it up to a SANE-supported SCSI scanner I got from the recycle center a while back - should be a little nicer than my current scanning solution, a Win9x-only SCSI-over-parallel-port scanner hooked up to a PIII laptop.


It might be a good idea to stay away from OpenBSD, especially if you plan on using any X Windows system. Even if hardware is supported
doesn't mean that its perfect at all. It displays , but its so dreadfully slow and unresponsive that you wished you stuck with the command line.
I tried running OpenBSD on sun blade 2K dual 1200 with xvr1200 graphics and 2GB ram and it sucked worse than running it with OpenSXCE.
So I reverted back to Solaris 10.u5
Though I think in your case you might be better off running an older version of Solaris, Say 5.8 or at most 5.9
you could compile, or grab what you need from OpenCSW or BlastWave ( http://ftp.ussg.iu.edu/solaris/blastwav ... sparc/5.8/ )
MAYA, nut-
:Octane2: :Octane2: Octane 2 R14k 600 V12 4GB, Octane2 R14K 600 V10 1GB ,
:Onyx2: :Onyx2: Onyx2 IR3 4GB Quad R14K 500 DIVO, Onyx2 IR Quad R12K 400 2GB,
:Indigo2: SGI Indigo 2 R8K75 TEAL Extreme 256MB,
:Indigo2IMP: SGI Indigo 2 R10K 195 Solid Impact 256MB, MAX Impact Pending
,
Apple G5 Quad, NV Quadro 4500 + 7800GT, 12GB RAM
Sun Blade 1000 Dual 900 XVR 1000 4GB
Sun Blade 2000 Dual 1200 XVR 1200 8GB
Seriously? I ran it on a 166MHz Ultra 1 with a measly 128MB RAM and it was perfectly usable. Faster than Solaris 8, even. Even did X fine, as long as I stuck with lightweight window managers like WindowMaker, which I was planning on doing anyway.
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/MT-32/D-10, Oberheim Matrix-6, Yamaha DX7/FB-01, Korg MS-20 Mini, Ensoniq Mirage/SQ-80, Sequential Circuits Prophet-600, Hohner String Performer

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup
Ryan Fox wrote: I tried running OpenBSD on sun blade 2K dual 1200 with xvr1200 graphics and 2GB ram and it sucked worse than running it with OpenSXCE.

The XVR-1200 being a completely undocumented beast, the X server for it is unaccelerated. It's no surprise display is slow, especially when using a wide resolution (such as 1920x1080).
:Indigo: R4000 :Indigo: R4000 :Indigo: R4000 :Indigo2: R4400 :Indigo2IMP: R4400 :Indigo2: R8000 :Indigo2IMP: R10000 :Indy: R4000PC :Indy: R4000SC :Indy: R4600 :Indy: R5000SC :O2: R5000 :O2: RM7000 :Octane: 2xR10000 :Octane: R12000 :O200: 2xR12000 :O200: - :O200: 2x2xR10000 :Fuel: R16000 :O3x0: 4xR16000 :A350:
among more than 150 machines : Apollo, Data General, Digital, HP, IBM, MIPS before SGI , Motorola, NeXT, SGI, Solbourne, Sun...
Yeah, that'd explain it. Luckily, both the Creator series and the GX series are supported by OpenBSD, so I should have (simple) acceleration on both displays.

I should be getting a replacement CPU and hard disk this week, so I should hopefully have it up and running before long :)
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/MT-32/D-10, Oberheim Matrix-6, Yamaha DX7/FB-01, Korg MS-20 Mini, Ensoniq Mirage/SQ-80, Sequential Circuits Prophet-600, Hohner String Performer

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup
Well, it took forever to actually get replacement hard drives, but it's now up and running with two 18GB drives and running OpenBSD just fine :) Now I just need to get X acceleration working...

Question, though: is there a way to force the Turbo GX framebuffer to display at a particular resolution/refresh rate? It doesn't generate a sync signal when I use it with the simple hard-wired 13w3-to-VGA adapter I use with the Creator framebuffers, and I suspect that it's just opting not to output a picture because it's not getting a valid resolution on the 13w3 sense lines. I have a KVM module with switches for the sense lines, but it doesn't work at all...
Computers: Amiga 1200, DEC VAXStation 4000/60, DEC MicroPDP-11/73
Synthesizers: Roland JX-10/MT-32/D-10, Oberheim Matrix-6, Yamaha DX7/FB-01, Korg MS-20 Mini, Ensoniq Mirage/SQ-80, Sequential Circuits Prophet-600, Hohner String Performer

"'Legacy code' often differs from its suggested alternative by actually working and scaling." - Bjarne Stroustrup
commodorejohn wrote: Question, though: is there a way to force the Turbo GX framebuffer to display at a particular resolution/refresh rate?

If it is the primary display, yes; see the cgsix(4) manpage for details. If it is not the primary display, this might be possible with a hairy nvramrc, refer to the Sun frame buffer FAQ for hints.
:Indigo: R4000 :Indigo: R4000 :Indigo: R4000 :Indigo2: R4400 :Indigo2IMP: R4400 :Indigo2: R8000 :Indigo2IMP: R10000 :Indy: R4000PC :Indy: R4000SC :Indy: R4600 :Indy: R5000SC :O2: R5000 :O2: RM7000 :Octane: 2xR10000 :Octane: R12000 :O200: 2xR12000 :O200: - :O200: 2x2xR10000 :Fuel: R16000 :O3x0: 4xR16000 :A350:
among more than 150 machines : Apollo, Data General, Digital, HP, IBM, MIPS before SGI , Motorola, NeXT, SGI, Solbourne, Sun...