c - CRC mismatch using calculation with CRC RevEng results - Stack Overflow

admin管理员组

文章数量:1277406

Hello and thanks in advance,

Background: I have to communicate with a device for which I have to write a 10-bit CRC with initial value 0000010000 (0x10) and poly:

x¹⁰ + x⁷ + x³ + x² + x + 1 --> 0x08F

Problem: If I look for the CRC algorithm using the messages I receive and "CRC RevEng":

reveng -w 10 -i 10 -s 010000FF0300005A 00F87F000000033D 00000000000003E3 0000FFFFFFFF012D

I get the actual algorithm:

width=10  poly=0x08f  init=0x010  refin=false  refout=false  xorout=0x000  check=0x06c  residue=0x000  name=(none)

BUT, if I reverse the operation, and use the CRC algorithm parameters to calculate the CRC, I do not get the CRCs I entered in the first place (0x005A, 0x033D, 0x03E3, 0x012D).

reveng -w 10 -i 10 -p 08f -c 010000FF0300 00F87F000000 000000000000 0000FFFFFFFF
00fc
0316
0331
007a

What am I missing?

Thank you,

Hello and thanks in advance,

Background: I have to communicate with a device for which I have to write a 10-bit CRC with initial value 0000010000 (0x10) and poly:

x¹⁰ + x⁷ + x³ + x² + x + 1 --> 0x08F

Problem: If I look for the CRC algorithm using the messages I receive and "CRC RevEng":

reveng -w 10 -i 10 -s 010000FF0300005A 00F87F000000033D 00000000000003E3 0000FFFFFFFF012D

I get the actual algorithm:

width=10  poly=0x08f  init=0x010  refin=false  refout=false  xorout=0x000  check=0x06c  residue=0x000  name=(none)

BUT, if I reverse the operation, and use the CRC algorithm parameters to calculate the CRC, I do not get the CRCs I entered in the first place (0x005A, 0x033D, 0x03E3, 0x012D).

reveng -w 10 -i 10 -p 08f -c 010000FF0300 00F87F000000 000000000000 0000FFFFFFFF
00fc
0316
0331
007a

What am I missing?

Thank you,

• c
• reverse-engineering
• crc

Share Improve this question Follow asked Feb 25 at 9:56 Eloi SanchezEloi Sanchez 6544 bronze badges 1

• What with the init=0x010? Is that the initial value of the checksum? That's odd in that case, usually it is either all zeroes or all ones. – Lundin Commented Feb 25 at 10:57

Add a comment  | 

1 Answer 1

Sorted by: Reset to default 5

The input length in bits is no longer a multiple of 8 after chopping off the 10-bit CRC at the end, so the input character width cannot be set to 8. It cannot be set to 4 either to specify the input as nybbles, so the only sensible alternative is to set the input character width to 1 (-a 1) and specify the input in binary. The output width can be set to 4 (-A 4) to show the CRC in hexadecimal for convenience.

$ reveng -w 10 -i 10 -p 8f -a 1 -A 4 -c 000000010000000000000000111111110000001100000000000000
05a

If you really wanted to go to the effort, you could specify the 54 input bits in octal since it is a multiple of 3 (input width -a 3), here with the output width set to 4 (-A 4):

$ reveng -w 10 -i 10 -p 8f -a 3 -A 4 -c 002000007760140000
05a

It is also possible to check the original hexadecimal strings including the CRC on the end produce a CRC of 0:

$ reveng -w 10 -i 10 -p 8f -c 010000FF0300005A 00F87F000000033D 00000000000003E3 0000FFFFFFFF012D
0000
0000
0000
0000

本文标签： cCRC mismatch using calculation with CRC RevEng resultsStack Overflow