Hi team

This is a huge long shot but does anybody have access to a dump of an HP ZX2000 itanium workstation? I have a board with a damaged bios chip and a damaged transistor, and would like to attempt a rescue.

It's an early EFI bios. Apparently, it's been dumped before: Shoehorning a dual core Itanium into a ZX2000

Alternatively, if anyone knows whether its possible to extract valid binaries from HP Itanium update files [Example EFI type update] please let me know.

These come in Linux, EFI, OpenVMS formats... but nothing looks like an appropriate binary that could be programmed with a hardware flasher. Any help appreciated!


Edit by Fernando: Thread title customized

@IntelModder - You'll have to dump your current bricked BIOS, and if it's not empty, try to find location of serial, UUID, LAN MAC ID so you can transfer over to this dump

If you can dump the BIOS, upload a copy here and maybe I can tell from that how to rebuild from stock package. Which system do you have? Your title says ZX2000, but you linked rx2600 BIOS

@Lost_N_BIOS Apologies, I've updated that link to the relevant file.

Getting a dump is not possible. The board is damaged and the bios chip is not salvageable. If I could source a binary dump I was going to buy and solder on a replacement Intel 28F640J chip.

@IntelModder - How badly is the original chip damaged? If the board is damaged, how do you think a new BIOS chip will help, or is it only the BIOS chip damaged?

rx2600 link is still the only link above, both link me to rx2600 system downloads
https://support.hpe.com/hpsc/swd/public/...nvOid=4184#tab2 << 2003 files
https://support.hpe.com/hpsc/swd/public/...1&swEnvOid=4184 << 2009 files

@Lost_N_BIOS the link is for the rx2600, zx2000 and zx6000. They're the same family and compatible systems are listed in the release notes.

Too badly to get a dump... but because replacements for the damaged transistor and bios chip are only a few dollars total I thought I'd give repairing it a go before I give up on it. If I can't source the files the board is probably going to electronic recycling.

So that download is for your system too? If yes, sorry, I missed that. I will check the files now. I checked the release notes, and don't see zx2000 listed as a compatible system, only the rx2600 - I do see zx2000 mentioned, but not any capacity that would mean "This BIOS also for zx2000 system"

You never mentioned, how bad is the BIOS damaged? I may be able to link you to an adapter you can get it into to try and dump, once you remove it, unless it's busted in half or crushed etc.

As far as I can tell, yes it is:

Zitat

---

This package (hp_fwzre_108.efi) is an EFI application to flash the System and BMC Firmware of hp workstation zx2000, hp workstation zx6000 and hp server rx2600, and contains the firmware images.

---

I have one of these with a tsop56 adapter lying around from another project: Flashcat USB xPort. So I'm confident that if I can source a binary I can program a replacement chip.

The old chip is busted, it's almost split in half. I don't have a photo available, unfortunately. But i can try and get one sometime within the next week if you think it might help.

Well busted in half explains enough, we can't use that to look at. I see in the release notes now how this may be applied to this system, but it certainly doesn't list it in a way that would be normal for a multi-platform BIOS. I'd expect a list at top or bottom of compatible devices or systems it can be used with etc.
I'm looking at the update_SFW.efi now to see if I can figure out how to get a file out of it for you.

What you mentioned, I don't see in the release notes? This is all I see mentioning this system

System FW 01.82
Intel has notified HP and other computer manufacturers of an issue with the Itanium processor. Multiprocessor (MP) systems running IA-32 applications that use the "cmpxchg8b" instruction could experience hangs, halts, or data integrity issues on your system while running Windows or Linux; HP-UX is not susceptible. HP multiprocessor systems that may be affected are: the HP Server rx2600, HP Server rx5670, and the HP Workstation zx6000. The issue does not occur on the first generation Itanium processor-based systems or on recently announced Itanium based single processor systems such as the HP Workstation zx2000. It also does not occur on any other Intel processors or PA-RISC or Alpha systems. This issue is not unique to HP and is a published Intel erratum on the Itanium processor (Erratum 53).

The problem has not been reported by any Itanium processor customers. However, because Intel is strongly committed to delivering high quality products, it has developed a new PAL (Processor Abstraction Layer, a part of system firmware) that addresses the issue. HP has incorporated this new PAL into firmware revision 01.82 for the HP Server rx2600, HP Workstation zx6000 and zx2000 products.


Maybe we get different end page from the same links? I also don't even find hp_fwzre_108.efi on that page?

At release note tab, on top I see TITLE: rx2600 System, BMC and iLO MP Firmware - and the only mentions of zx2000 are the three above
And in description tab - SFW and BMC for rx2600
In certain configurations, zx2000, zx6000 and rx2600 PCI bus speed may be restricted to 66 MHz

Ohh! I see, I was looking at the 2009 download page, sorry for my confusion! I'll check the 2003 one now too.

* Edit

I found the _FIT_ table in a 4396kb file from the hp.fwsrre.108.efi - but this file, and the main one + other, do not look like full BIOS files, only partial updates of certain modules in each one.
I think this only applies the changes and leaves the base code in place. So I don't think we can get full BIOS from these .efi files

Damn. Thanks anyway @Lost_N_BIOS! Your expertise is much appreciated

You're welcome, and thank you too @IntelModder - sorry I couldn't be of more help. I also tried to find a dump for a while in google, no luck

If you have second system you can dump from, we can use that to get the other going, then just need to change serial and LAN ID probably.

If anyone in future ever needs this firmware, or is interested in IA64 firmware/early EFI 1.10 generally, it can be found attached.

This was dumped with hardware programmer from the flash chip. There seems to be some PE files, but nothing resembling a firmware volume.

Hello!
Dear @IntelModder and other contributors, that's great that I've found this topic. Let me describe what happened during the reanimation of my old zx2000 machine. Really need advice on what to do next ))

I have 3 zx2000 system boards at hand:
- the original one, worked without problems for many years and eventually died (it still can run EFI and fails with machine checks when loading OS; HP diagnostic utility finds a number of damages);
- 2 ebayed replacements, both are of older hardware revisions and with outdated firmware. Every attempt to boot ends with a swf hang. I tried several single core Itanium chips, including a 1.0Ghz Deerfield (the original system was equipped with it), 1.4Ghz and 1.5Ghz Madisons (not specified in the zx2000 docs however the original machine was OK with both), and finally an old 900Mhz McKinley - however, with minor differences the result is the same: ia64 firmware hangs and after the timeout BMC reports a firmware error.

I assume that the probability of having 2 physically damaged replacements from different sellers is close to 0 and what I see is really a firmware hanging. What could have caused this? Besides a newer Itanium chip, it can be a scsi (excluded trying different versions of Lsi adapter and even by unplugging it at all), VGA (not fully excluded bacause I have a limited number of AGP VGAs at hand, but the old system is OK even with newer Radeons), and a priori anything else (it seems paranoid, but I have changed the pw supply, fans and the front panel card).

I guess there's something I don't know. Why I was so successful with the original system board and everything ends in failure with new ones?

There is a surgical solution: solder off the AMI chip from the old board and attach it to the new one. A softer version of that is to use the ISP interface, of course, undocumented. Which way would you recommend? Perhaps there are some nuances? Is there any chance to use ISP? Thank you in advance.

Regards,
Valery

Hi @valeryz2001

It's good to see someone else with one of these machines. Mine is currently in storage and I cannot test anything for you, but I may be able to share some extra information.

1. Firmware hanging was incredibly common on mine. They do not appear to like a range of USB devices - mine locked and threw an MCA on a certain USB drive and a certain Bluetooth module.

2. I wouldn't worry about your limited AGP cards. Despite a lack of drivers for most cards, I only found one which would hard lock the system - a 3dLabs Wildcat VP880 Pro. Best bet I found was ATI cards (including PCI versions) and I tested up to an HD2600 AGP.

You don't say which firmware versions are on the boards. If you can get a board to drop into EFI I'd try updating to the latest firmware (still available on the HP site last I checked).

It seems you've got it covered, but I'd only recommend using the 900mhz McKinley and clearing the CMOS by removing the battery for a period prior to a boot attempt. Just make sure the time is set before trying the update - the update gets unhappy if it's not set.

Hi @IntelModder, thanks for a lot of useful and new info!

I saw the first MC about 8 years ago and immediately purchased a used motherboard, then facing everything described for the first time. All available zx2000 motherboards seem to be taken from the early deployed rack systems, thus all have very old firmware. But my problem turned out to be unrelated to the board: a scsi adapter began to die. So I was given a delay until now. But this time it's the board itself ((

Zitat

---

You don't say which firmware versions are on the boards

---

BMC info command says about it's own firmware only, it's 1.12 on one board and a bit newer on another. So i'm not aware of what is inside the AMI chip.

Zitat

---

I'd only recommend using the 900mhz McKinley

---

The only therapeutic option is to test McKinley on the most recent board. As to one I bought 8 years ago, it was tested several times with McKinley and showed something different in behavior than did with newer chips, I cannot recall exactly what. For the purity of experimenting I forgot to test the McKinley itself: though my original board is _dead_, it still can run efi, so this should be done on McKinley too.

Zitat

---

Just make sure the time is set before trying the update - the update gets unhappy if it's not set

---

Thank you, I missed it.

V.