[Request] Complete BIOS Dump for Dell 5820 Workstation

BIOS/UEFI Modding BIOS Modding Requests
1

I am quite new here. Had an issue with my BIOS. It’s flashing 3 amber, 3 white, but it can’t access any IO, and no keyboard is lit up.

I just extracted the BIOS from the chip 25Q256JVF0 16 pin 32 MB, but it contains only a UEFI Image, which is just a BIOS region. I am banging my head as it doesn’t have the other descriptors, GBE, and Me. During one of the dumps earlier when I couldnt get the dump consistently(minor variations) it did generate one with all since it wasnt consistent I largely ignored. Hence, the confusion: does this motherboard contain dual BIOS or just one with all regions?

Board is 002KVM and Intel Xeon W 2155

Couldn’t locate the right category. Please move if it isn’t appropriate.

Thank you so much for the help
Thank you for moving!

Appreciate it.

2

@CodeNinja
Welcome to the Win-Raid Forum!
What you extracted was just the content of the “BIOS Region”, but for the flash of a new BIOS you need is a complete Dump (SPI) of the BIOS chip.
Good luck!

3

Yes Thank you for confirming. I managed to prepare it. But was wondering

Here is the uefi tool ne’s screen capture.

image
image518×419 6.63 KB

4

It didn’t go well. looks like my ME region is wrong. So getting constant 4 amber lights followed by another amber light and restarts in a loop.

Want to learn more about how to verify this before flashing again. Any help is appreciated.

thank you

5

Not the best learning example/model for starters…
Besides the EC fw IC (SPI_2) and excluding here the TPM, there’s 2 main ICs or shadow/toggle dual-cpu (SPI_1/3)

image
image1141×694 125 KB

Anyway… FPT dump by Service jumper and both readings of the ICs, IS A MUST before proceeding with any erase/write operations, next was trying to re-initialize/Clean the CSME FW…(CS SPS 4 most likely…)
Clean/blank IC programming is always last resource.
And before any of this…INFORMATION is the priority to research, especially the board (MATIRA 3) schematics and a possible boardview file (dont ask for it), that’s all, good luck.

6

Thanks for that. I always wondered what the use of the SPI_DEBUG is.

I already flashed it, thinking I have a valid image. But it looks like it’s not. I was about to get a replacement motherboard. Thought of just trying on this motherboard to see if I can understand how this works.

So the only way out of this situation is to get a flashable bios and flash it?

Sorry, this is my first attempt. So questions can reflect the same :slight_smile:

7

By your post on Dell forums… seems you have 3 of those… what does this tell you to try?
Anyway, these are server class boards, requires a bit more dedication and for sure a more experienced user, over_n-out.

8

Your image helped me a lot to see whats going on Thank you so much

Yes, I like this system. Wanted to keep it for a while. When this one died, I was sad and got another one from the used market. I was about to order the motherboard, but thought it’s worth trying it, as worst I will lose this one.

Actually, I have two 5820s and another one is 3620(just an i7 variant), once I fix that, I will probably stash it somewhere.

Ok, here is the recap

Day 1: I installed the Dell updates; I didn’t know it included BIOS updates. It didn’t force me to reboot or allow me to reboot later

Day 2: Due to my dev drive resize i needed to install hyper v as part of that forced me to reboot. And the bios was a mess due to two party conflict

Day 10(a lot later, but let’s just pretend its 10th day): started extracting bios using esp32 serprog, and independent clips and soldered wire on pins (some of them like 16, 7), among 10 dumps, all are having some variance in size, which was accurate, but there are bytes different. But one of those backups is structurally correct. Let’s call it day 10 backup

Day 11: swithced to solder that to esp32 directly and cut short cables it started generating identical dumps and then i moved to ch341a and only one change but before that change all were idneticaly and after change all are identical. So, mostly 11-day dumps are pretty solid with a caveat. It only generated 32 MB with BIOS region ‘none’ of the others present there. But it has the service tag inside
Day12: today used the day 10 backup as a base to replace the me region from the bios exe and the bios region from the exe as well. Then extracted two nvram guids with the Uefiedit tool 0.28. Verified in Uefi NE alpha 73, no errors in there.

If I am going to attempt again. Will use spi_debug to connect to the chip.

If I had located the service_mode jumper earlier, I would have bridged it to extract.
Not sure how I got one dump that had everything(used serprog with esp32 if that makes any difference), as in my previous post. Not sure that’s the correct one.