I've never found that additive sum somewhere. I'm starting to think there is no such sum to be verified. mutante I think the best is to disassemble either HPQflash or rompaq to find the verification algorithm.... we need to know what's checked...what kind of sum and what address range.....at bios it's 8 bit, 16bit or crc, sometimes only the odd bytes are taken and at second sum the even bytes for calculation...sometimes different address ranges....
Is there any possibility to re-tattoo motherboard like DELL,maybe that "trigger" for fullflash the BIOS by original HPQflash is on the motherboard?In some older HP models there is written on motherboard if the computer is freeDOS or linux or not.Or we just have to wait BIOS update with unusable SLIC 2.1 and correct it with WoW?
I think we should play with bytes....at the second marker change one byte +1 to 46h and the next byte -1 to 4Ch, try to flash with original flasher..error? If error then correct offset 2 to 52h only If error then correct offset 3 to 4Bh only If error then correct offset 4 to 48h only Somebody try please? Edit there, not the marker header: Code: Offset 0 1 2 3 4 5 6 7 8 9 A B C D E F 000F6210 45 4D 53 4C 49 43 2D 4D 50 43 57 49 4E 44 4F 57 EMSLIC-MPCWINDOW We can fiddle with bytes and try to flash with original, we can go step by step to see how sensitive the integrity check will be....change some bytes try to flash.. again and again....keeping the sum....
Hi, Yen! I tried all this with original HPQFlash.exe - always error. Moreover, I drew attention that 68MAD.BAK (can be created and loaded to ROM with original rompaq.exe and patched rompaq_.exe) more than the 68MAD.BIN (original) to 6h (first byte is FFh instead of F9h). I tried to restore original 68MAD.BIN (renamed to 68MAD.BAK) with the original rompaq.exe -> succesfull. So, I tried your method to modify 68MAD.BAK (first byte F9h and other six bytes in your string higher by one) CRC32 are equal -> error. And yet, I drew attention, when using rompaq_.exe (patched) in command line I can see restoring (/R) and updating (/U) proceed banks 1...14 (always with ORIGINAL and modded 68MAD.BIN). But backuping (/B) proceed banks 1...15. It seems, rompaq_.exe (and may be original rompaq.exe for 68MAD) cann't to overwrite 15s bank (it's may be addresses F0000-FFFFF).
So It seems the checksum problem is already solved at the patched flasher. The question: How to make the flasher to flash the last bank 15? Either patch the flasher or..hmm the biosfile itself has a flag somewhere to tell the flasher that bank 15 has to be updated! If HP wants to make this be updated as well, the biosfile 'has to tell' the flasher, just do it! maybe remove the 14 banks limit? Difficult? mutante? What do you think about? Or what about these special hP tools? Edit: there is 'uniflash' a universal flasher, but quite obsolete......