Which is valid, upper or lower of CRC_CRC in 16-bit CRC mode?

[yasuhikokoumoto]

Hello experts,

I have a question of the crc_demo sample in KINETIS_120MHZ_SC provided by the freescale.

According to Kinetis K70 Reference Manual "35.3.2.1 16-bit CRC", the usage of CRC_CRC is described as below.

6. Clear the CTRL[WAS] bit to start writing data values.
7. Write data values into CRC[HU:HL:LU:LL]. A CRC is computed on every data
value write, and the intermediate CRC result is stored back into CRC[LU:LL].
8. When all values have been written, read the final CRC result from CRC[LU:LL].

I think it would be inconsistent with the sample code.

The original code is as follows.

  sizeWords = sizeBytes>>1;
  j = 0;
  for(i=0;i<sizeWords;i++){ 
      data_in = (msg[j] << 8) | (msg[j+1]);
      j += 2;
      CRC_CRCH=data_in;                 //DES was CRC_CRC_LOW=data_in; 
  }
  if(j<sizeBytes)
  { 
     pCRCBytes = (uint8_t*)&CRC_CRC;
     *pCRCBytes++ = msg[j];
  }
  data_out=CRC_CRCH;                    //DES was CRC_CRC;

I guess it might be as follows.

  sizeWords = sizeBytes>>1;
  j = 0;
  for(i=0;i<sizeWords;i++){ 
      data_in = (msg[j+1] << 8) | (msg[j]); // CHANGE
      j += 2;
      CRC_CRCL=data_in;        　           // CHANGE 
  }
  if(j<sizeBytes)
  { 
     pCRCBytes = (uint8_t*)&CRC_CRC;
     *pCRCBytes++ = msg[j];
  }
  data_out=CRC_CRCL;                        //CHANGE

Am I not correct?

Can anyone teach me the truth?

Best regards,

Yasuhiko Koumoto.

[Paul_Tian]

Hi, Koumoto-san

Please pay more patience. I will check your issue. Thanks.

Best Regards

Paul

[yasuhikokoumoto]

Hi Paul,

thank you. I am interested in the result.

Best regards,
Yasuhiko Koumoto.

[PatriciaTeran]

Hi, Yasuhiko

Are you still having problems?

Regards

Patricia

[yasuhikokoumoto]

Hello Patricia,

I have still problem because I have not gotten any answer from Paul.

Best regards,

Yasuhiko Koumoto.

[Jorge_Gonzalez · Accepted solution]

Hello Yasuhiko Koumoto:

I think the sample code is correct and your confusion is about endianness. In the sample project, the message buffer is filled from a UART channel in sequence, so you can consider that the bytes are stored in Big Endian mode:

printf("\r\nPlease enter an ASCII Message:");
bByteCount = Input_Message(&strMsg[0]);
...
uint32_t  Input_Message(uint8_t *MsgBuf)
{
  uint8_t ch;
  uint32_t count;
  count = 0;
  do
  {
    ch = in_char();
    out_char(ch);
    if(ch == '\r')
    {     
      break;
    }
    MsgBuf[count++] = ch; 
    if(count >= MAX_CHARS)
    {
      break;
    }
  }while(1);
  return (count);
}

But the CRC registers in K70 are Little Endian, so the correct way to calculate CRC16 for this buffer is with the original code:

 sizeWords = sizeBytes>>1;
  j = 0;
  for(i=0;i<sizeWords;i++){ 
      data_in = (msg[j] << 8) | (msg[j+1]);
      j += 2;
      CRC_CRCL=data_in; 
  }
  if(j<sizeBytes)
  { 
     pCRCBytes = (uint8_t*)&CRC_CRC;
     *pCRCBytes++ = msg[j];
  }
  data_out=CRC_CRC;  // CRC_CRCL is also correct
 return(data_out);

Regards!,
Jorge Gonzalez

-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------