z820 e5-2600 v2 ivy bridge upgrade

[Z820]

Might have to manually input the link guys, sorry.

I edited the Part Surfer link in the earlier post, and it is now working.

Warren_Admin, HP Enterprise Community Manager
Any personal opinions expressed are mine, and should not be interpreted as official statements on behalf of Hewlett Packard

[SalSimp]

---

@Dan_in_WGBU wrote:

Installing a jumper on E14 overrides the system ROM boot block protection circuit. . . 

---

Dan - Does the E14 jumper override a complex "protection circuit" or does it merely 'pull up' (Vcc) the /WP pin of the BIOS flash rom (pin 9 of the flash SOIC)?

I could pull the board and do a continuity test but if you know the answer it would save from doing it. TIA

[Dan_WGBU]

salsimp, on the -003 Z820 motherboard, installing E14 pulls up a signal going to the Intel PCH, which is read by the BIOS.  It is not connected to the write protect pin 9 on the SPI BIOS ROM.  This write protect pin is connected to a circuit on the motherboard, which controls state of the WP pin.   

I have been *really* busy in my day job (forum responses are done on my own time), and did not have time to give a detailed reply to this thread.  Let me add few comments.    

Let's all face the elephant in the room, OK?  All this discussions are now about trying to find clever, round-about ways to flash the boot block so it can support Ivy Bridge.  (agreed?)  This is still a bad idea, not just because I am required as an HP employee to warn you not to try this, but also because installing a BIOS is far more complicated than putting an image on a chip.  I do not fully understand all of the issues, but here are some: 

   - Even if one removes the BIOS IC and flashes it with an external ROM burner, there are several problems, since part of the "CMOS" are stored in multiple places now, not just in the battery-backed registers in the PCH: 

        - The SAS controller address will be lost, causing it not to function.

        - One LAN port MAC address will be lost, causing it not to function.  (I think this is true, (??))

        - Tools to restore these addresses are not available outside of HP.

        - Just flashing part of the ROM, like the boot block, with then make the ROM checksum invalid, and it might not boot. 

        - After flashing the ROM with an external programmer, it is in "manufacturing mode" to streamline motherboard testing.  It then has to be "locked" by special SW tools before release. 

        - The TPM (Trusted Platform Module) IC on the motherboard has requirements for signed drivers, and possibly BIOS. 

HP BIOS flash tools and programs take care of these items.  Flashing it manually does not.   

HP is NOT locking the boot block to sell more boards (really, believe me), but there are requirements from NIST and others that will not allow changing the boot block.  If programs were allowed that could modify the boot block, it is a good way to use an altered version to maliciously corrupt the boot block and make the board inoperative.  A truly evil virus/trojan. 

        - I do not know what other manufacturers do.  I heard (second hand) that other first-tier computer companies also comply to these requirements, but cannot verify this. 

The participants on this thread are very intelligent, and knowledgeable about computer systems.  I understand about wanting to upgrade, I have done this on my personal systems for a very long time.  However, BIOS programming and updating is a very complex process that few understand fully, and there are no short cuts to modifying the BIOS. 

I am trying to prevent any unintended damage to your systems, OK?  Please don't hate me.   😉 

[SalSimp]

---

@Dan_in_WGBU wrote:

---

I have been *really* busy in my day job (forum responses are done on my own time), and did not have time to give a detailed reply to this thread. ..

I think a lot of us with day jobs and families can relate to that.  As 'enthusiasts' on this Board, we really appreciate your insights as limited as they may be due to your HP association.

---

@Dan_in_WGBU wrote:

... on the -003 Z820 motherboard, installing E14 pulls up a signal going to the Intel PCH, which is read by the BIOS.  It is not connected to the write protect pin 9 on the SPI BIOS ROM.  This write protect pin is connected to a circuit on the motherboard, which controls state of the WP pin.    

---

Thanks, that is useful information to have as we continue with this project.

[SalSimp]

---

@Dan_in_WGBU wrote:

Let's all face the elephant in the room, OK?  All this discussions are now about trying to find clever, round-about ways to flash the boot block so it can support Ivy Bridge.  (agreed?) 

---

No argument, agreed.

---

@Dan_in_WGBU wrote:

... installing a BIOS is far more complicated than putting an image on a chip.  I do not fully understand all of the issues, but here are some: 

   - Even if one removes the BIOS IC and flashes it with an external ROM burner, there are several problems, since part of the "CMOS" are stored in multiple places now, not just in the battery-backed registers in the PCH: 

        - The SAS controller address will be lost, causing it not to function.

        - One LAN port MAC address will be lost, causing it not to function.  (I think this is true, (??))

        - Tools to restore these addresses are not available outside of HP.

        - Just flashing part of the ROM, like the boot block, with then make the ROM checksum invalid, and it might not boot. 

        - After flashing the ROM with an external programmer, it is in "manufacturing mode" to streamline motherboard testing.  It then has to be "locked" by special SW tools before release. 

        - The TPM (Trusted Platform Module) IC on the motherboard has requirements for signed drivers, and possibly BIOS. 

HP BIOS flash tools and programs take care of these items.  Flashing it manually does not.   

---

While I understand the complexities you point out, I believe there has to be a workaround, just as it was possible on the XW9400. There one did not need to modify the BIOS bootblock but merely replace it. Let me explain.

As with the XW9400, the first required step is to read the entire BIOS (all 16MB of it). The BIOS contains the 'bootblock', the code for the ME/ATM engine and the rest of the normal BIOS code needed for the system to POST properly. Indeed these portions are signed (Verisign'd) but I do not believe there is a single signature for the entire BIOS flash (16MB), but signature for the various segments.

Once the BIOS is read, one needs to find the location of the bootblock. Typically the protected bootblock are athe beginning or end of the flash. With a binary editor one would 'lift' the bootblock from the newer Ivy Bridge supporting BIOS binary in the HP Softpack (like J63_0 352.bin) and substitute it for the one in the BIOS image read. Now you have a full BIOS image that has the Ivy Bridge bootblock. The ME/AMT code has not been altered as is the remainder of the BIOS. Their signatures do not change so during POST they will be verified as not being corrupted (if it were, a BIOS crisis recovery action would solve the problem).

Ok, assuming all that works (may need tweaking as more is learned) the BIG problem that exists is how to write the BIOS image back into the flash chip. All of the above is useless unless we can write the BIOS.

So the first task is to read the BIOS image. Once done we can work on identifying sections of the BIOS (bootblock in particular). These first two steps are tedious and do take perseverance but achievable. That takes us to the most difficult part, flashing back the image. That is the tough one, but one step at a time

Well, so much for now. Feedback or thoughts from others on this thread can be very helpful, so fire away.

[SalSimp]

Attached is a WinZip file with the binary contents (16MB file) of the BIOS flash ROM. This binary file was obtained by reading the BIOS flash chip from a Z820 workstation with an E5-2690 V2 XEON. The BIOS version it contains is 3.52

I have changed the serial number it contained (at 0x3B93A, 0x3BA6A) to keep the source private. Did the same with the MAC addresses (at 0x1000,0X2000,0x3470C, 0x3B84C AND 0x3B985).

These changes likely make the signature checks fail as Dan pointed out. I just do not want the image source workstation to end up being cloned if this image were to be somehow flashed on a Z820. The purpose of this posting is to advance the search for the bootblock so it can be replaced in your own individual BIOS ROM image (as previously discussed).

Just as an initial hypothesis, I believe the bootblock begins at 0xFE0000. I also think that the ME/AMT code block begins at 0x005000 and has a length of 0x77B000. But this needs to be verified.

A little help in the verification of this would help.

[MickTurner]

Hi

There are similar discussions going round with respect to the upcoming Thunderbolt 2 expansion card - F3F43AA. I think the key thing for making it work is a GPIO header (I think 5 pins). I have spoken to HP technical support and they told me it would work but it sounded like they were just reading off the card technical specifications rather then truly understanding the subtleties of different motherboard versions.

I don't suppose you could check to see if the 708610-001 board has such a GPIO header?

Thanks!

Mick

[#8]

Anybody have any success upgrading to V2 processor using latest firmware?

[MSLEEE]

HP does NOT support to upgrade the processor to Ivy-Bridge processor (v2) for sandy-bridge system offcially. As mentioned, the boot-block should be udapted to support the ivy-bridge processor for sandy-bridge syste board. Only HP factory is able to update it. Even though you call HP Support, you will not be able to get the right system board whcich support the ivy-bridge processor.

I want to emphasize that HP supports only to add the same processor in the second CPU slot that was from the system. Upgrading any different speed of processor from origianl processor is not supported offcially. There are mutiple reasons on this.

[SalSimp]

---

@JohnL wrote:

I want to emphasize that HP supports only to add the same processor in the second CPU slot that was from the system. Upgrading any different speed of processor from origianl processor is not supported offcially. There are mutiple reasons on this.

---

So are you saying that the Z820 Base Units that HP sells without processors, but with a warranty, are not supported when processors are added by the buyer? So then what good is the warranty?