From mboxrd@z Thu Jan 1 00:00:00 1970 From: Parav Pandit Subject: ip6 route output() and ip_route_output_key() by drivers Date: Wed, 30 Jun 2010 00:24:04 -0700 (PDT) Message-ID: <750794.80583.qm@web30108.mail.mud.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii To: netdev@vger.kernel.org Return-path: Received: from web30108.mail.mud.yahoo.com ([209.191.69.40]:46698 "HELO web30108.mail.mud.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1752769Ab0F3Haq (ORCPT ); Wed, 30 Jun 2010 03:30:46 -0400 Sender: netdev-owner@vger.kernel.org List-ID: Hi, Our device driver uses ip6 route output() and typecast the return value of dst_entry to rt6_info (similar to (net/ip6/icmp.c) Driver uses (a) rt6i_dst.addr, (b) rt6i_dst.plen and (c) rt6i_gateway to get the route entry (ip/mask, gw, netdev). Can anyone please confirm that drivers and/or netfilter kernel modules are allowed to use them? netdev drivers like : cnic.c, cxgb3, nes are using the ipv4 API (but not IPv6 API). Also ip_route_output_key() rtable and dst_entry doesn't seem to provide route entry's subnet mask/prefix value. how drivers/netfilter kernel module can get a ip/mask gateway value while given a destination IP address as input? I really appreciate your inputs. I hope I am asking it on the relevant mailing list. Regards, Parav Pandit