Intel NUC DCCP847DYE Results

I have a Classic and Pro FlashcatUSB, the settings are correct: SPI NOR and 3.3v, it also verifies 100% match after programming (I have many years experience programming SPI so this is not the issue)

The bios version is the latest from Intel



Note: the updated bios with UBU works fine, only the uCodes padding is causing a problem. The manual modded virgin bios with only the uCodes updated for some reason does not boot.

Okay thanks for showing me. I hope, we’ll get no problems with this edited v66 (write protection included).

I’ve got news: I found some incoherences in alignment. Now it should be right.
But I only changed the old microcodes to new ones & subtracted the overcrowed space. I’ve done no checksum correction.

@Lost_N_BIOS please could you check my changes with your skills …

Edit: Rebuild version added. Should be fine for testing to flash … or, L’n’B?

1
2
3
4
5
 
Datei: Winbond_W25Q64_00-7FFFFF_Rebuild.bin
CRC-32: 79e03a2b
MD4: 1199275909ab4d581adac7a19f58404a
MD5: 8ce537dc8222714db19c6f53de759377
SHA-1: 086c6f18cf8b793ecfe1b1cbf09a53f9eb853998
 

Winbond_W25Q64_00-7FFFFF_New.zip (4.48 MB)

Winbond_W25Q64_00-7FFFFF_Rebuild.zip (4.48 MB)

Tested “Rebuild” first, this one did not boot…

Tested “New” and this one seems to work, the CPU info Bios also shows correctly and UBU shows all the uCodes as latest



Can you tell me the steps you did to replace the uCodes in case I need to in the future?

Thank you both for all your help!

Update: When I attempt to update the "New" bin with UBU I get the following

1
 
parseFile: invalid data checksum 56h, should be DEh
 


Edit: The UBU checksum error does not happen with "Rebuild," but this bin is bad as it will not boot.

@ricktendo - That checksum can be corrected, or ignored since it’s working now, but I would correct it (it’s microcode data checksum) @MiesMosel - it’s at 11h in the header, or you can correct and test ricktendo (@ 002804b9h when viewing the entire BIOS in hex)

@MiesMosel - thanks! Could you tell me what was wrong with the BIOS (microcodes) at post #8 or #10 (Should be same, just #8 have vBIOS update also, #10 did not)
Everything looks same as mine, except you put replacements in original locations, is that the key? Usually that doesn’t matter as long as rest is aligned/proper and FIT fixed if present.
I thought about doing that too, but wasn’t sure it would matter, and thought maybe I had wrong offset values or something, but looks like we used same (800h multiples)

I checked your “NEW” only, since he said other failed. Looks fine to me, except I would fix that checksum

Thanks Lost_N_BIOS for the informations about MCs checksum, that’s what I like to know more.
I’ll have a look @ 11h in the header (002804B9h in whole bios).

@ricktendo
I only exchanged the microcodes by hex-editor with beeing aware of the 800h alignment (800h+800h=1000h).
If a MC is x400h (p.e. 2400h/3400h) long, then it needs an additional 400h (with FFs) block before the next MC.
And this additional 400h-block you have to subtract/delete from the FFs after the whole microcode-section (below 206A7).

@Lost_N_BIOS
What I have seen by extraction through MMTool, there’s a header before the whole microcodes section/block (1818h length). If microcodes are changed, the (checksum in the) header may has to be changed, too. Edit: Done!

With UEFITool_NE and FIT option I have no experience, but maybe I could learn from you. Edit: Okay, found your Guide.

So I’m happy to understand now 8-bit checksum.

@ricktendo
Here’s the checksum-patched BIOS-file. Give it a try.
Edit: And the same one with updated (modded) OROMs by UBU.

Winbond_W25Q64_00-7FFFFF_patched.zip (4.48 MB)

mod_Winbond_W25Q64_00-7FFFFF_patched.zip (4.52 MB)

Thank you for all your help @MiesMosel and @Lost_N_BIOS , the patched bios with the uCode is now updated and working PERFECTLY!

I have updated the rest of the components with UBU (no more checksum error,) also the Intel MEI, and VBIOS… The CPU info is displaying bug is no more and info is displaying normally in the Intel Visual BIOS

Thanks @MiesMosel - I kept original microcode module overall size exact same, so nothing should have needed removed at the end, mine same size as yours, and alignment same too. So I think keeping original locations was the key why yours worked and mine failed this time, don’t you agree?
Please show mw the header you mean, that’s way to long to be a header, and only thing I see before microcode is a padding file of 18h and before that is a PEI module 53695C7A-BE6C-4724-A2D9-19B16E6FC202. Microcode volume itself has a header of 18h too, that’s the one I mentioned to change in my previous post.
UEFITool NE Alpha will always detect if checksum is wrong there, and you can either “rebuild” with UEFITool regular, or use the value UEFITool NE gives you and manually change. Is one of those 18h headers the one you mean? The microcode header was at 11h on the microcode volume itself, or at 2804b9h on entire file in hex

No FIT module in this one so nothing to fix there this time. Thanks for reminding me of my guide, I need to update that due to it’s now much easier to fix manually because of updates to UEFITool NE

@ricktendo - you’re welcome, but seems MiesMosel got this one all correct for you, thanks again @MiesMosel

@Lost_N_BIOS Can you please do me a favor and update the uCode of my bios again to preserve the padding, latest UBU still cannot upgrade them properly like you can

Download: https://mega.nz/#!e9RGjYBD!D5gJ2fPYnFw8z…E3agW8D7FFiKs4s

Here is the desired Before and After



Hey @ricktendo - yes, no problem. I am not sure why UBU is still doing funky stuff with ucode sometimes…?
I see 2-3 BIOS this week with bad edits, we discussed a few times too, but I’m not sure if he doesn’t quite understand what I mean, even with images showing, or if he thinks it’s OK/safe sometimes (And it may be), I’m not sure why it’s not correct every time like it used to be?
Did you test the BIOS with the UBU padding issue, if yes was it bricked, or did it boot but not properly load the ucodes or what?

You’ll have to learn to do this yourself manually, this BIOS is an easy one
http://s000.tinyupload.com/index.php?fil…772927349298646

@Lost_N_BIOS can you point me again to your tutorial?

Editing uCode with ubu does not brick the device, it only breaks the version display in bios

@Lost_N_BIOS I just tried your modded bios and it will not boot.

@Lost_N_BIOS I just tried your modded bios and it will not boot.

@ricktendo - Good to hear the UBU edit is not bricking the device at least! Not showing the ucode means it’s not loading it either

!!Ahh!! Very sorry about the non-boot BIOS! Of course, now that we have this again I remember I think we had that before too, and I had to edit another way, sorry I forgot to check back on this before making you a new BIOS!
Sorry about that, I’ll make a note in your folder on this to keep this in mind next time around!

Please test all of these, and let me know which boot and do not break microcode shown (all should boot I hope), then let me know, then I will update my notes with which method to use next time
I suspect M2M ones will work/boot, M2U may not boot, CC and D0 is test on which method needs to be done for the actual uCode edit and for those to be shown to you and not broken/not-loaded.
I noticed UEFITool is making changes inside next module (MemoryInit), that we’re not editing, and I didn’t notice last time we tried to figure out why MiesMosel’s edits were working and mine were not

So, if you want to be sure you can boot, and not have to go through non-boot hassles again, maybe only testing M2MD0 and M2MCC would be good enough to let us know with method onthe ucodes to use.
And then, if you wanted, you could confirm the M2U ones both fail to boot, but this not as important as the M2M tests, if both of those boot anyway
http://s000.tinyupload.com/index.php?fil…942404117023315

@Lost_N_BIOS as you suspected NONE of the M2U* images booted

M2MCC boot but the CPU/uCode in bios was not loaded
M2MD0 also boot and CPU/uCode in bios loaded correctly, so this appears to be the WINNER!



Thank you for your help!

@ricktendo - Great! And you’re welcome

So, here’s the breakdown of the issue, in case you wondered. M2U edits were done with UEFITool, this to preserve the padding initially (v25), I think we checked out V25/26 previously and both were the same edit via hex, so version didn’t matter.
Anyway, I found after saving, it’s making some changes in the module following the microcodes, which it shouldn’t be doing, there may be other edits below that done too but I didn’t look further after seeing that as that alone was plenty of non-desired edit, any single byte changed not desired since we didn’t edit that module.

M2M files were straight hex edits on the entire file, microcode body block removed and hex edited outside of BIOS but same/same, then pasted back in once work done. CC was part of starting alignment location of the last microcode in one file, D0 in the other.
CC aligned that ucode at a 400h alignment offset, and D0 used 800h alignment offset. So your BIOS requires 800h offset to be used for the starting location each ucode (800, 1000, 1800, 2000 etc).
The main issues here were UEFITool editing other modules, and then the alignment of ucodes. Microcodes were kept in there original order this time, but now that we know all this I don’t think that matters (we can test and confirm this if you want?)

@Lost_N_BIOS if you want me to test further images feel free to post them I will gladly test, its past my bedtime so I will do this tomorrow some time.

If you ever need help SPI flashing I am your guy, I will gladly help you out if needed

Thanks @ricktendo - here is mod BIOS with ucodes not in matching order of original BIOS, properly edited other than that, lets see if it works
This is how I normally put them in when editing, order being highest name value to least.

uCode-OrderChangeTest.png


http://s000.tinyupload.com/index.php?fil…589404201520048

And thank you, if I run into any issues with SPI flashing I can’t resolve I’ll remember to tag you to hop in and advise

Ahh, I just realized your CPU is 206A7, so in all test BIOS this is last ucode anyway Once you confirm above test is OK, I’ll make you new one with your ucode first

@Lost_N_BIOS sorry for the late reply, I got sick and I could not test your our of order uCode Bios updates, also Intel released a new version of the bios and of course they did not bother to update the uCodes, any chance if you have the time you could update them for me?

https://mega.nz/#!fxR11KiC!lGqKseBRy7xB-…TdFIbu6obV5F5vs

Thanks for everything!