Re: [pim] Implementation Question

[Stig Venaas <stig.venaas at uninett.no>]

Nikhil Ninan wrote:
> Hello there,
>                 I am researching into PIM-SM protocol behaviour. This 
> requires me to create PIM-SM messsages to be used in my test-bed. For 
> the life of me, I cant get the PIM-SM checksum value calculated 
> properly. In the specification it says, that checksum is the one's 
> complement of the PIM message. Does that include the IP header and if 
> so, what about the checksum field within the IP header. Any help would 
> be much appreaciated.

First of all, the IP header checksum should be taken care of by the
operating system. If this is IPv4 PIM, the PIM checksum is based on
just the PIM message itself (for IPv6 there is a pseudo header with
info from IP header as well), see 4.9 in RFC 4601.

I think the following code should work for the checksum. It's based
on some code from libnet and pim6sd. I cooked this together now and
have not tested it. You would call pim_cksum() with pointer to PIM
header where the PIM payload follows directly behind in the same
buffer.

Stig

/* thanks to libnet code */
uint32_t in_cksum(uint16_t *addr, size_t len) {
    uint32_t sum = 0;

    while (len > 1) {
        sum += *addr++;
        len -= 2;
    }
    if (len == 1)
        sum += *addr;

    return (sum);
}

void pim_cksum(struct pim *pim, size_t len) {
    uint32_t sum;

    pim->pim_cksum = 0;
    /* pad with 0 if odd len, assuming buffer large enough */
    if (len % 2)
        ((char *)pim)[len] = 0;
    sum = in_cksum((uint16_t *)pim, len);
    sum = (sum >> 16) + (sum & 0xffff);
    sum += sum >> 16;
    pim->pim_cksum = ~sum & 0xffff;
}

Stig

> Cheers
> Nikhil Ninan
>
> _______________________________________________
> pim mailing list
> pim at ietf.org
> https://www1.ietf.org/mailman/listinfo/pim


_______________________________________________
pim mailing list
pim at ietf.org
https://www1.ietf.org/mailman/listinfo/pim