[geeks] Ultra 2: what's U1401?

[Ethan Hawke]

I would try memory first, while the test is in regards to the APC, the 
error is generated by the CPU's memory controller (ECC error).

There could also be problems with the E$ on the CPU modules, so you can 
try either swapping them around or trying a different one.

The APC "Aurora Personality Chip", was part of the chipset for the SS5, 
although it was reused just for sound duties in the U1/U2 (and perhaps 
others).
It's one of the few ASICs made by Sun with no (known) available User 
Manual, just that brief datasheet you mentioned.

As the APC is just for sound (in the Ultras), if it was faulty I can't 
see it causing these problems.

I may be wrong, but try memory and CPUs first, at least the CPUs are 
replaceable as opposed to the failing E$ on my U1...

Regards,
Ethan

On 9/1/21 3:30 AM, Mouse wrote:
> I've got an Ultra 2 that fails POST.  But the error is cryptic enough I
> can't tell whether it's anything reasonably fixable.  I'm wondering if
> anyone can tell me, or point me towards doc that can tell me.
>
> The error I'm seeing is
>
> 0>APC Test
> 0>    APC Registers Tests Test
> 0>    APC DVMA Test
> 0>ERROR: TEST=APC,SUBTEST=APC DVMA ID=16.2
> 0>Component under test: APC U1401
> 0>    Unexpected event occured
> 0>    Data Access Error from address 00000000.001319c0
> 0> tl  tt         tstate                 tpc               tnpc
> 0> 01  63  00000011.80001603  000001ff.f004bd24  000001ff.f004bd28
> 0>    AFSR 00000000.00100000
> 0>    (CE) Correctable ECC Error
> 0>   SDBH[143:72] Syndrome is c8. CE Error on bit 67
> 0>    SDBH = 00000000.000000c8 SDBL = 00000000.00000060
>
> I found 802-2561-11.pdf, but it's little help.  The block diagram on
> page C-2 seems to indicate that the APC is purely for audio.  C-29 has
> another diagram, but all that does is tell me it's not a RAM stick (the
> "Correectable ECC Error" part made me hope it was RAM).  So far I
> haven't found anything in there that actually helps me.
>
> .idprom output is, I suspect, characteristic of a dead battery:
>
> ok .idprom
> Format/Type: ff ff Ethernet: ff ff ff ff ff ff Date: ff ff ff ff
> Serial: ff ef ff Checksum: ff
> ok
>
> It occurs to me that a dead battery could cause POST failure, but if so
> it's unexpected to the POST author; it complains about U1401, while the
> NVRAM is U2006.
>
> I've tried to reset the MAC, but whenever I try mkp or mkpl it wedges,
> needing a hard power-cycle.  The machine does have a disk in it, but
> when I try to "boot disk -s" I get
>
> Boot device: /sbus/SUNW,fas at e,8800000/sd at 0,0  File and args: -s
>
> and it wedges there.
>
> Any thoughts?  I'd hate to send this off to the scrapper if it doesn't
> deserve it, but I also don't want to beat my head against the wall of
> trying to make broken hardware work.
>
> /~\ The ASCII				  Mouse
> \ / Ribbon Campaign
>   X  Against HTML		mouse at rodents-montreal.org
> / \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B
> _______________________________________________
> GEEKS:  http://www.sunhelp.org/mailman/listinfo/geeks