IP packet checksum algorithm C # (formerly)

For the sake of pseudo IP, we specially made an IP package and used C # to rewrite the IP header checksum algorithm.
Document 1:
IP Header Format:
Version Number (4 digits)
IP header length (4 digits)
Service type (8 bits)
Packet Length (16 bits)
Id segment (16 bits)
Flag (16 bits)
Survival time (8 digits)
Transmission Protocol (8 bits)
Header checksum (16 bits)
Sending address (16 digits)
Destination Address (16 bits)
Option
Fill

Document 2:
The IP protocol uses a unified verification algorithm. The calculation is relatively simple: Set the checksum initial value to 0, and then calculate the checksum for every 16 bits or if the result is reversed. During verification, the data (including checksum) is summed according to the same algorithm. If the result is 0, the data is correct. If the value is not 0, a communication error occurs and the packet must be discarded.

Algorithm source code:

Public static UInt16 checksum (UInt16 [] buffer, int size)
{
Int32 cksum = 0;
Int counter;
Counter = 0;
While (size> 0)
{
UInt16 val = buffer [counter];
Cksum + = Convert. ToInt32 (buffer [counter]);
Counter + = 1;
Size =-1;
}
Cksum = (cksum> 16) + (cksum & 0 xffff );
Cksum + = (cksum> 16 );
Return (UInt16 )(~ Cksum );
}

Note: The buffer array is the entire IP package array and needs to be converted to UInt16 []; size is the length of the buffer array.
The method for converting byte [] to UInt16 [] is relatively simple. We will not describe it here.