[request] Modded bios for Tyan S5533 MoBo

Wow, you did try to flash your bios via remote KVM? Do you have physical access to this machine or ist ‘really’ remote? (Did you remote boot, too? And no, I don’t think that’s the reason for the error, but in case something doesn’t work as expected…)

@wildgunman - Sorry! There is many things wrong here, hopefully it’s just with your method used. Is this an ES AFUDOS? If not, you cannot use /GAN. If this is the stock AFU Flash app, you also cannot use /GAN.
I assume this may be the stock tool, as I know I suggested you try (only with the batch file I edited/included), and if this is that failed then it’s to be expected since this is mod BIOS. I only wanted you to try it, as sent, to see if it allowed the flash or not.

Since it hasn’t, proceed as you normally did before with /GAN I suppose. I would remove all those switches though, unless you also used those previously without issue.

However, I do prefer and suggest FPT flashing, but you’d have to send me FPT Dump, and I’d have to reapply the edits to that, wouldn’t take but a few minutes, then you’d reflash with that.
Since this is Intel system, that should be the preferred method over all others, unless that also gets blocked that we can’t get around but usually there’s enough tricks to get around it.
If you want to try this, which I suggest, please wait for further instructions and do not try further AFU Flashes.

@lfb6 - do you know if the 03.00.05.402 package FPT works on these systems?

@lfb6

no, the USB stick is physically on the motherboard, but I have no monitor and use the image output of the KVM.

@Lost_N_BIOS


new situation - the motherboard is dead. I did not pay attention when updating the BMC firmware and used the image of another board, which works in one of my other servers. The end of the song, the board turns on for a moment and then goes out again … and in continuous loop. Any efforts on my part have brought nothing. I have now ordered a SOIC16 test cable for my CH341A Flash Writer, I do not have much hope of being able to flash the chip. The BMC firmware is only available in .IMA format. Inside is a mini-Linux. I will report if it worked. Nevertheless many thanks to all involved for their time and effort.

p.s: the afudos is an older Version, that works with /gan command.

Regards Christian

Sorry to hear, hope you can revive it so we can continue!!

AFU w/ /GAN would not give you any security errors >> /GAN means flash no, no checks, even if wrong board/BIOS etc. So that is why I assumed it was not ES
no other switches are usually used with /GAN, so that may be why you are getting error, likely the /ME one, I’ve never heard or seen any error with use of /GAN, it just bricks the board if wrong or bad BIOS, without any error, flashes as soon as you tell it too, without any checking, in any instance I’ve ever seen.
Additionally, I have all ES versions of AFUDOS I believe, and none are 3.05.03 (3.06.01 is the only v3), seeing that security/capsule check appears to be proving this is not an ES version as suspected. Your use of /GAN would be ignored on non-ES versions, so if that’s what you used before then it was flashing just because, not due to use of /GAN

@Lost_N_BIOS There’s a chance as spsInfo and SPSManuf from this package do, but as I told you the spsFPTW* is not too cooperative.

I’ll have a look at the Dos or/ and Efi version(s) tomorrow.


@wildgunman On a P9D-I there’s an undocumented jumper BMC_SW1 which disables the BMC, maybe your board has this possibility, too?

I will look tomorrow. Thanks for the info.

@lfb6 - what does not too cooperative mean?? It either functions or not, and should always be ran from an admin command prompt. Sorry if you mentioned before, if so, I forgot, what issue did you see with it?

@Lost_N_BIOS The original fparts.txt doesn’t have an entry for my Winbond 25Q64FV, so without diving any further in it I tried a newer fparts.txt from ME9.1 tools, but the information is a litttle different and spsFPTW now complains of course that the chip erase timeout is not correct…

ME 9.1;Description of fields:
; 1) Display name
; 2) Device ID (2 or 3 bytes)
; 3) Device Size (in bits)
; 4) Block Erase Size (in bytes - 256, 4K, 64K)
; 5) Block Erase Command
; 6) Write Granularity (1 or 64)
; 7) Enable Write Status Register Command (1- True, 0- False)

SPS 3
;Description of fields:

; 8) Chip Erase Command
; 9) Chip Erase Timeout (in milliseconds)

Found the missing information, but now the error is ‘no supported SPI flash devices installed’. Winbond W25Q64FV is named in the VSCC table- see picture in spoiler (picture taken from the chips I normally use, there ‘s a (minimal) chance that there is another chip on the board, but I’ have to open the case to check).

But taking the output of spsManuf it should be OK:

Checking ME Hardware and Firmware Status…passed
vsccommn.bin was created on 18:47:14 06/03/2013 GMT
SPI Flash ID #1 ME VSCC value is 0x2025, could not determine if device supports SFDP capability
Checking ME VSCC status…passed
Checking default spsManuf tests…passed

spsManuf Test Passed

And since these tools do (and did) work just once after every reboot, it’s a little time consuming (or in other words these tools are not too cooperative). But as you can see in the first picture: spsFPTW does at least read the Flash Descriptor…

spsftpw.jpg


sps3_3.jpg


Sorry, this has become quite off topic now!

Two things here @lfb6 - sorry I didn’t mention before, FPT isn’t going to work in windows unless you have the ME drivers installed.

As for fparts.txt, add entry line for W25Q64FV that matches your chip info, but that’s still not going to help unless/until you have ME drivers installed.
0xEF4017, 0x4000000, 0x1000, 0x20, 64, 0, 0xc7, 80000W25Q64FV
0xEF6018, 0x8000000, 0x1000, 0x20, 64, 0, 0xc7, 80000W25Q64FVSSIQ
0xEF4017, 0x4000000, 0x1000, 0x20, 64, 0, 0xc7, 80000W25Q64FVSSIQ

I assume the main issue is no drivers installed, there is always random error in windows without drivers installed.
Also, we may need specific DE version of FPT? You’ll know for sure if you get some error in DOS/EFI too

Both entries for 25Q64FV are there (third line (EF6018) is a W25Q128*)

W25Q64FV, 0xEF4017, 0x4000000, 0x1000, 0x20, 64, 0, 0xc7, 80000
W25Q64FV, 0xEF6017, 0x4000000, 0x1000, 0x20, 64, 0, 0xc7, 80000


What would a DE version be?

I meant DE specific version of SPSFPT, like the specific FITc you sent me

I found all four of these in a V10 FPT fparts, similar to what I listed above (can’t remember which I grabbed that set from)
W25Q64FV, 0xEF4017, 0x4000000, 0x1000, 0x20, 64, 0
W25Q64FV, 0xEF6017, 0x4000000, 0x1000, 0x20, 64, 0
W25Q64FVSSIQ, 0xEF4017, 0x4000000, 0x1000, 0x20, 64, 0
W25Q64FVSSIQ, 0xEF6017, 0x4000000, 0x1000, 0x20, 64, 0

Hurray, back in the game. The SOIC16 test clip arrived today and after a little soldering I could use it with the ch341a. Although I could not flash the .ima file, but at least delete the chip. After that, the board started again. In Windows, then I have flashed the correct firmware with socflash.

Nice to hear it! So, can you now flash the mod BIOS I sent you in #19, or do you need to fix some other stuff before you can get into BIOS again?

I have flashed the Bios with afudos es and /gan option - so far so good - no issues, board starts without problems :slight_smile:

flash3.PNG

Looks good then, great you recovered and mod BIOS is OK

@Lost_N_BIOS
Can you send me the Bios with updated ME?

Regards Christian