> >On Tue, Nov 08, 2011 at 12:29:53AM +0100, Hans Schillstrom wrote: >> >We prefer skb_header_pointer instead. If conntrack is enabled, we can >> >benefit from defragmention. >> >> In our case conntrack will not be there > >Yes, but if conntrack is there, we benefit from fragment reassembly if >you use skb_header_pointer. > >> >Please, replace all pskb_may_pull by skb_header_pointer in this code. >> > >> >We can assume that the IP header is linear (not fragmented). >> >> I ran in to this issue in IPv6 testing so I got a little bit "paranoid". >> Are you sure that the embedded IP and L4 header in the ICMP msg also is unfragmented. >> Is this true for both IPv6 & IPv4 ? > >No sorry. I was refering to normal IP header in one packet. > >> From what I remember when I was testing IPv6 icmp and digged into the original header (on a 2.6.32 kernel) >> pskb_may_pull was needed. > >Yes, it is indeed needed. > >> [snip] [snip] > >Welcome, let's see if we can get this into 3.3 since we cannot make it >for 3.2. > >BTW, do you have some number of this running with and without >conntrack? It would be interesting to have. I didn't save them, but I can make a new benchmark later on. Regards Hans -- To unsubscribe from this list: send the line "unsubscribe netfilter-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
