Checksum Algorithm mismatch

[MatrixEditor]

Checksum Algorithm

Although, the checksum algorithm is discovered and decompiled here, the implemented method in python always returns a wrong result.

Description of the algorithm

The call which is done in BuildSADPPacket() is the following:

 uVar7 = Checksum(
    *(ushort **)(buffer + 0x14), 
    (uint)*(byte *)((int)*(ushort **)(buffer + 0x14) + 3)
 );

Note, that the conversion to unsigned short * will produce a pointer that combine two bytes each position. Therefore, +3 at the end has the following effect:

  ubyte *buf = [0x10, 0x20, 0x30, 0x40];
  ushort *dest = (ushort *)buf;
  //           = [0x1020, 0x3040] 

The next step is to convert the returned unsigned short value into a single byte value:

  ushort *buf = [0x1020, 0x3040]
  byte dest = *((byte *)buf)
  //        = 0x10

The input buffer is the following:

  ┌───────────────────────────────────────────────┐
  │Buffer Structure                               │
  ├─────────┬─────────┬─────┬─────┬─────────┬─────┤
  │0x21 0x02│0x01 0x42│.. ..│.. ..│0x06 0x04│.. ..│
  └─────────┴─────────┴─────┴─────┴─────────┴─────┘

Basically, the algorithm adds the uint16 values starting from position 0 until a index variable is reaching 0. Even if the values are copied from a packed which was sent by the SADP Tool there is a checksum mismatch.

[MatrixEditor]

This issue has been fixed since patch version 1.1.0. The main algorithm implementation was correct, only the input parameters were interpreted wrong. See checksum.h for details on the arguments.