matthieu castet wrote:
> I don't know how is encapsulated ip in dvb, but this patch seem quite
> ugly : why we can't read the protocol in the llc header ?
>
> For the atm it is done like that :
> if (!clip_vcc->encap || skb->len < RFC1483LLC_LEN ||
> memcmp(skb->data,
> llc_oui,sizeof(llc_oui))) skb->protocol = htons(ETH_P_IP);
> else {
> skb->protocol = ((u16 *) skb->data)[3];
> skb_pull(skb,RFC1483LLC_LEN);
> }
>
> where RFC1483LLC_LEN=8 and
> static const unsigned char llc_oui[] = {
> 0xaa, /* DSAP: non-ISO */
> 0xaa, /* SSAP: non-ISO */
> 0x03, /* Ctrl: Unnumbered Information Command PDU */
> 0x00, /* OUI: EtherType */
> 0x00,
> 0x00 };
>
> In sumary if there no llc header we are in vcmux mode and we assume IPV4
> and otherwise we read the protocol in the LLC header.
Yeah, we should do it that way. Thanks for the information!
> Also instead of removing the first 8 Bytes by hand which could be
> expensive, couldn't use the kernel network utils like skb_pull ?
I'll simply change the skb_put() call to skip the LLC/SNAP header.
Thanks,
Johannes
