Trying to change the poly in this crc algorithm, need help

[iso9001]

I have an application that requires a CRC. I looked/asked around for an efficient algorithm and found this one from Dallas's One Wire protocol.

Dallas uses the poly 0x99. If it were up to me, I'de use that poly too, but its not up to me and I've been stuck with 0x11D.

My problem is that I have no idea how the number they are using have been calculated. I've searched and searched, but found nothing.

My CRC code right now works fine except that it for loops for every bit in every bite. An entire 8 byte message takes over 300uS @ 40Mhz,

I know that they picked 0x99 for a reason and that its possible that I can't take theirs and modify it, so if anyone has any other code it would really help me out a lot,

Can't wait till the day each pic has a tiny FPGA built-in... Then I can continue on being a crappy programmer


//-----------------------------------------------------------------------------
// Calculate crc8 checksum
// This is the checksum as used in the Dallas One Wire protocol.
// Advantage of this checksum over other crc's is that because of a pecularity
// of this specific polynomial it allows bytewise processing without bit
// shifting. It's very fast, small and uses no RAM.
//
// Based on: Maxim Application Note 27,
// http://www.maxim-ic.com/appnotes.cfm/appnote_number/542
// Original C code: T. Scott Dattalo
// http://www.dattalo.com/technical/software/pic/crc_8bit.c
// Optimized by: Dirk Van den Berge (dvdb)
// donated to the CCS-forum 12-jan-2004
// http://www.ccsinfo.com/forum/viewtopic.php?t=17170
//-----------------------------------------------------------------------------
int8 Calc_Crc8(int8 *Buffer, int8 byte_cntr)
{
int8 crc=0;

#ASM
//copy pointer to pointer register (CCS doesn't support lfsr assembly instruction)
movff Buffer,FSR0L
movff &Buffer+1,FSR0H

loop:
movf POSTINC0,w // load w with next databyte
xorwf crc,f // xor with accumulated crc (accumulated crc is no longer valid now)
movlw 0 // w will accumulate the new crc
btfsc crc,0
xorlw 0x5e // could also be iorlw
btfsc crc,1
xorlw 0xbc
btfsc crc,2
xorlw 0x61
btfsc crc,3
xorlw 0xc2
btfsc crc,4
xorlw 0x9d
btfsc crc,5
xorlw 0x23
btfsc crc,6
xorlw 0x46
btfsc crc,7
xorlw 0x8c
movwf crc // store accumulated crc
decfsz byte_cntr,f
goto loop // next databyte
// done, crc is in w

movwf crc // store in result
#ENDASM

return crc;
}

[iso9001]

No one knows a better CRC Algorithm or any advice on changing this one to work with my poly ????

Maybe even a table driven appliction ?

[PCM programmer]

The problem is in how you posed your question. It left me uncertain
as to exactly what you want. To exlain:

[Ttelmah]

Also, as a further comment, 0x11D, is a very unusual CRC (9bits). Modifying an 8bit CRC, is not going to get very far for this...

Best Wishes

[treitmey]

I thought the CRC.C and ex_crc in the samples would work.
The poly is defined. ... Don't you just refined the new one?

[iso9001]

I think there is plenty of information on the problem, I have a poly definded by outside sources as 11D and need an efficient CRC algorithm. My original question was if it was possible to modify the posted alg, or if anyone had an different one.

treitmey: Nah, ccs's is a 16bit CRC. Thanks for looking tho.

Ttelmah: Yea, I know its wierd. I have a feeling that 0x1D might also work but have not verified this.



What I don't understand is how Dallas has a poly of 0x99 but these are the number used in the code:

0x5e, 0xbc, 0x61, 0xc2, 0x9d, 0x23, 0x46, 0x8c


I'm still pretty confused how a I could set it up in a table driven application.

Anyone have experience with this ?

[jds-pic]

[jds-pic]

you will also want to read the entirety of this:
http://www.riccibitti.com/crcguide.htm

jds-pic