[weblacky]

The L1 dallas holds a COPY of the serial number (not the primary source anyway, the small socketed ATMEL chip on the board has the serial...otherwise the system could never auto-configure a blank chip), while I don't think anyone has publicly tried it, if you could place ANY DALLAS DS1220 compatible RTC of the right density in these machines and they worked...that would be great, but there must be a reason we don't, right?

But the L1 RTC stores ALL NVRAM settings and I assume that means the logs, hw inventory, and hardware IO setup registers settings, etc. It doesn't just hold a serial number, is holds all the same info as a PC BIOS config would need to (including whatever the MIPS uses instead instead of ACPI tables).

Also, I don't trust the Irix and CDROM diags. Power-on diags are good enough for most things but I've never seen a machine pass those darn thing without some kind of freezing a test (one or the other), the onsite IDE never worked well whenever I tried it...

[jwhat]

Hi kokoboi,

I know that you are sure that this is not related to DS1742W battery, but to verify serial "stickiness" can you print out the L1 log, after powering down (removing from power completely) and then up.

If the machine is re-initialising the serial as part of L1 startup then it will be in the L1 log.

If it is resetting the serial you will see something like this:

>> M200XXXX-192.168.XXX.XXX-L2>l1 log
>> 001c07:
>> 11/01/20 23:41:34 checksum Error - common header initialized
>> 11/01/20 23:41:34 nvram checksum error - initializing core data.
>> 11/01/20 23:41:34 nvram checksum error - initializing extended data.
>> 11/01/20 23:41:34 nvram checksum error - log pointers invalid, log pointers reset
>> 11/01/20 23:41:34 L1 booting 1.44.0
>> 11/01/20 23:41:34 ChiServ IP59
>> 11/01/20 23:41:34 Checking for Type
>> 11/01/20 23:41:34 -- ChiServ Type set
>> 11/01/20 23:41:36 ** fixing invalid SSN value
>> 11/01/20 23:41:36 ** fixing BSN mismatch
>> 11/01/20 23:41:36 USB0: waiting on open
>> 11/01/20 23:41:41 USB0: opened
>> 11/01/20 23:41:41 USB0: registered for events
>> 11/01/20 23:41:42 INFO: System serial number reassigned (8000) to M200XXXX from attached L2.
>> 11/01/20 23:43:37 L1 booting 1.44.0
>> 11/01/20 23:43:37 ChiServ IP59
>> 11/01/20 23:43:37 Checking for Type
>> 11/01/20 23:43:37 -- ChiServ Type set
>> 11/01/20 23:43:40 USB0: waiting on open
>> 11/01/20 23:43:41 USB0: opened
>> 11/01/20 23:43:41 USB0: registered for events

I say "something like" as this was from O350 node, which reports as "ChiServ" and took on the serial from the NUMALINK connected machines.

Cheers from Oz,

jwhat/John.

[weblacky]

His log shows settings are incorrect or missing for SCSI controller and Ethernet (at times). This appears to be a "hardware not found issue". He likely has a serial but that doesn't prove the SRAM DS1200AB chip he got magically supports having the correct entries after initialization. Log indicates when accessing HW entry on BASEIO settings...nothing is found.

I think the accessing of the inventory settings or the writing of them isn't working due to wrong RTC. The serial issue is farther down the list but he's not complaining of it in his logs.

This: http://forums.irixnet.org/thread-2867-po...l#pid26531

Seems to say it's not finding hardware, then tries to reinit hardware inventory...still not finding hardware. I'm questioning is this RTC he used (which isn't a the correct one) is the reason for the missing hardware due to incorrect entries...which it's trying it fix. System Serial isn't a part of the base issue for now...finding basic IO hardware and initializing is.

[chulofiasco]

Initializing Base I/O Ethernet Interface...Failed. MII Status Register = 0x7949

That’s a normal error if the Ethernet on the IO9 is disconnected.

[weblacky]

Initializing Base I/O Ethernet Interface...Failed.  MII Status Register = 0x7949

That’s a normal error if the Ethernet on the IO9 is disconnected.

Oh, I'd not seen that...

[kokoboi]

As I previously explained with the last Dallas (from the photo) I was able to setup the serial number and the machine holds it just fine.
However if the dallas chip is fake (which is mostly likely) it probably doesn't function like it should be. The glitchwrks has been purchased, but I'm not so optimistic. The machine worked few years ago and even with dead nvram it should boot.
I can not run any diagnostics as I can't boot IRIX at the moment.

you can run diagnostics off of the cdrom.

Could you let me know how to run the diagnostics from the irix 6.5.30 install cd ?

[chulofiasco]

As I previously explained with the last Dallas (from the photo) I was able to setup the serial number and the machine holds it just fine.
However if the dallas chip is fake (which is mostly likely) it probably doesn't function like it should be. The glitchwrks has been purchased, but I'm not so optimistic. The machine worked few years ago and even with dead nvram it should boot.
I can not run any diagnostics as I can't boot IRIX at the moment.

you can run diagnostics off of the cdrom.

Could you let me know how to run the diagnostics from the irix 6.5.30 install cd ?

you should just have to put the CD-ROM in the drive and click the button on the prom menu for diagnostics.

[chulofiasco]

Actually, upon further review, you’re not going to find the diags in the CD.

Tezro isn’t using ide, the install CDs have ide for IP30 and IP32. (Octane and O2)

You’ll need to find a copy of smdk for this in order to run diags.

[indigofan]

I am having some odd issues trying to get 6.5.30 installed on a Terzo...

001c01-L1>debug 0x08
debug switches set to 0x0008
001c01-L1>pwr u
001c01-L1>

No POD mode... Is there any other way to get into it?

My machine is keeping the date, so guess the Dallas chip is fine.. Haven't done anything yet with the IO9.. Thoughts anyone?

I have gotten as far as loading installation tools, then on reset, it just kind of hangs..

[weblacky]

I am having some odd issues trying to get 6.5.30 installed on a Terzo...

001c01-L1>debug 0x08
debug switches set to 0x0008
001c01-L1>pwr u
001c01-L1>

No POD mode... Is there any other way to get into it?

My machine is keeping the date, so guess the Dallas chip is fine.. Haven't done anything yet with the IO9.. Thoughts anyone?

I have gotten as far as loading installation tools, then on reset, it just kind of hangs..

I normally used debug 0x010d for basic POD.