Rule Definition
Typically when performing integer overflow check the addition is compared with one of the values. The type for added operands automatically get promoted to a higher type and hence a overflow check may be faulty if it is not ensured that the output of the addition in the comparison expression is of the same type as the individual operands, explicitly. This becomes security-critical when the result is used to control looping, make a security decision, or determine the offset or size in behaviors such as memory allocation, copying, concatenation, etc.

Remediation
Explicitly declare all the operands of comparison as the same type

Violation Code Sample

bool OverFlowcheck(unsigned short x, unsigned short y) {
  return (x + y < x); 
}

Fixed Code Sample

bool OverFlowcheck(unsigned short int x, unsigned short int y) {
  return (static_cast <unsigned short int>(x + y) < x); //Fixed
}

https://cwe.mitre.org/data/definitions/190.html