CRC Polynomial

The same 32 bit poly is used for all Ethernet rates.

Regards, Allan

Yes is always the same:

x(32)+x(26)+x(23)+x(22)+x(16)+x(12)+x(11)+x(10)+x(8)+x(7)+x(5)+x(4)+x(2)+x+1

Only take care that it is transmitted MSB first (while all the other fields are LSB): does anyone know why ?

Regards

Luc

A common CRC implementation (especially in 10 Mb/s systems) uses a

32-bit shift register with feedback elements for the various terms in the polynomial above. The output of the shift register becomes the serial bit sequence transmitted onto the LAN. Sending the CRC from MSB to LSB means that, once the actual data is complete, the implementation can simply keeps shifting bits out onto the LAN for another 32 bit times; i.e., the hardware implementation is easier if the CRC is sent in the specified bit order.

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

"Rich Seifert" ha scritto nel messaggio news: snipped-for-privacy@news.active-news.com...

Thanks a lot Mr. Seifert ! We definitely need a new version of your great book.

Thanks again

Luc

Thanks Srini

This reminds me of a discussion of bit order for MAC addresses.

A CRC doesn't really have a MSB and LSB, as it is not a number, but a series of coefficients for a polynomial written as a binary bit pattern.

Now, it might make some sense for the X**0 coefficient to match the

2**0 bit. Note, though, that bit numbering is not always with the LSB as bit 0. IBM mainframes, consistent with big endian byte addressing, and even though they supply no bit addressing instructions, numbers bits with the MSB as bit 0. It seems, then, not so out of place that the bits shift the other way.

-- glen