From mboxrd@z Thu Jan 1 00:00:00 1970 From: Buster Holzbauer Subject: Checksum Question about new sk_buffs Date: Wed, 22 Jul 2009 12:42:11 -0700 (PDT) Message-ID: <362475.96996.qm@web30508.mail.mud.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE To: netdev@vger.kernel.org Return-path: Received: from web30508.mail.mud.yahoo.com ([68.142.200.121]:36076 "HELO web30508.mail.mud.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1752263AbZGVTmL convert rfc822-to-8bit (ORCPT ); Wed, 22 Jul 2009 15:42:11 -0400 Sender: netdev-owner@vger.kernel.org List-ID: Hello, I am a researcher at a university working on a new protocol based off o= f TCP, using the 2.6.26.5 kernel on a pair of gentoo machines. I need t= o send additional sk_buffs containing redundancy information as well as= the sk_buffs TCP would normally send containing data.=A0 The redundancy sk_buffs are constructed by a function called by tcp_pus= h (). I create these sk_buffs by calling sk_stream_alloc_skb(), then up= dating a few fields in the header (such as sequence number), but I'm cl= early doing something wrong or missing a step. I believe this because a= fter adding a printk to the csum_error label in tcp_rcv_established, th= e receiver-side logs show that these redundancy sk_buffs are triggering= this label. The normal data sk_buffs, as expected, do not trigger this= printk. Does anyone have any information about the specific steps to create a n= ew sk_buff, update header fields, and populate it with data so that the= receiver processes the sk_buff correctly?=A0 Thanks in advance for any help you can provide, Buster Holzbauer =20