RFC 1331 contains the C code (and the required fcs look-up
table) used to generate the CCITT 'CRC' A/K/A the 16-bit
FCS (Frame Check Sequence) as described in RFC1331
appendix B:

"The HDLC polynomial: x**0 + x**5 + x**12 + x**16 (0x8408)."

as used in the PPP protocol:

http://www.armware.dk/RFC/rfc/rfc1331.html

Here are the simple test cases I sent through their code to
verify it's operation:

#define PPPINITFCS      0xffff  /* Initial FCS value  */

// Test, 1 of 3:
     Process_Buffer[0] = 0x00;
     Process_Buffer[1] = 0x00;
     PPP_TempUInt = pppfcs(PPPINITFCS, Process_Buffer, 2 );
//   Result: 0xf0b8

// Test, 2 of 3:
//   Process_Buffer[0] = 0xFF;
//   Process_Buffer[1] = 0xFF;
//   PPP_TempUInt = pppfcs(        0, Process_Buffer, 2 );
//   Result: 0xf0b8

// Test, 3 of 3:
//   Process_Buffer[0] = 0xb8;    // LSByte
//   Process_Buffer[1] = 0xf0;    // MSByte
//   PPP_TempUInt = pppfcs(   0xf0B8, Process_Buffer, 2 );
//   Result: 0x0000

One thing that is not mentioned (at least in RFC 1331) is
that this fcs must be COMPLEMENTED before inclusion
(or comparison with) the fcs value sent in the PPP packet. I
spent a day, maybe two chasing that rabbit to find out why
PPP packets I was receiving (from commercial equipment) didn't
agree with the FCS I was calculatating ...

Also be sure to respect ENDIAN-ness when generating and
comparing these values with values picked out of a byte stream.

Feeding your ASCII test string "12...89"in Process_Buffer into the
FCS code results in:

       PPP_TempUInt = pppfcs(PPPINITFCS, Process_Buffer, 9 );

Result in PPP_TempUInt:   0x6F91
Complementing this gives: 0x906E

Hope this helps - and sorry, I have nothing on the "CRC-16" code.

Jim  (seemingly ex-RF guru now doing S/W and hopelessly
         addicted to PICLIST)

PS. I coded this up in Borland C ver. 4.5 and ran it on a PC
under DOS. I've got a bit more code (including the Borland
C version of the FCS C code) if you're interested. Let me know.

----- Original Message -----
From: "bernard wood" <woodbj@CLEAR.NET.NZ>
To: <PICLIST@MITVMA.MIT.EDU>
Sent: Saturday, July 28, 2001 6:02 AM
Subject: [EE]: verify CRC-16 on "123456789" is BB3Dh??


Hi all,
I've coded routines but need to verify that they do indeed produce the
correct CRC.

To do this, I need 2 known CRC check values.  These are obtained by running
thru
the "standard" check string "123456789" of your algorithm and obtaining the
result.

I think that the check with CRC-16 on "123456789" is BB3Dh (can someone
verify this?)

Can someone (with known good code) please tell me
what the 2 check CRC's are for both
CRC-CCITT (x^16+x^12+x^5+1, or POLY=1021h)
and
CRC-16 (x^16+x^15+x^2+1, or POLY=8005h)

or alternatively, send me known good code in C or VB6?

Thanks in advance,
Bernard Wood
New Zealand

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.