Search code examples
debuggingchecksumcalculationfalse-positivehex-file

Fixing checksum issues on intel HEX files

I have made the following intel hex file snippet to test with a couple 8051 processor simulators:

:1000F5007002501F500CED2562FDEE3561FEEF35A7
:1001050060FFE56233F562E56133F561E56033F57E

One I used is located here: http://www.jroweb.de/8051/

From my research, it is to my understanding that a checksum of the intel hex file is calculated by summing all the pairs of hex digits (except last), then ANDing the result with 255 to get the 8-bit value, inversing the value, adding 1, and doing mod 256.

I basically followed the math from a respondent from these forums:

https://social.msdn.microsoft.com/Forums/en-US/a1736ae8-2db6-4657-a98c-e3e447ebfda3/calculate-intel-hex-file-format-checksum

When I performed the calculations, the checksum values for each line in the above snippet are correct, however in the 8051 processor simulator program I mentioned above, it shows me a checksum error on the last line. It thinks the value should be 7F and not 7E.

Is it a possibility that a false positive exists in the last hex file line (which confuses certain software to believe that 7F is the correct value)? If so, how should I arrange the last line in my hex file to fix it?

Solution

  • AFAICT your checksums are correct. Based on this and some other records I've fed to it, the simulator seems to have an off-by-one bug on some inputs.

    You have several options, including:

    • edit the simulator's t8051m.ini file and set IgnoreChecksum to 1

    • humour the simulator by editing your hex records to carry the checksums it expects instead of the correct checksums

    • disassemble the simulator's .exe file, find the bug, and fix it. (Before you go to those lengths, you could try reporting the problem to the author. It might be a known issue with a fix that the author just never got around to publishing on his website.)

    • use some other simulator