From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pavel Emelyanov Subject: [PATCH 0/2] inet_diag: Fix old/new inet_diag_req-s structures Date: Wed, 11 Jan 2012 12:35:32 +0400 Message-ID: <4F0D49D4.7040100@parallels.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit To: David Miller , Stephen Hemminger , Linux Netdev List Return-path: Received: from mailhub.sw.ru ([195.214.232.25]:28617 "EHLO relay.sw.ru" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752172Ab2AKIfn (ORCPT ); Wed, 11 Jan 2012 03:35:43 -0500 Sender: netdev-owner@vger.kernel.org List-ID: The recent changes to inet_diag broke the compilation of tools using kernel headers. Fix this by doing two sequential renames: inet_diag_req -> inet_diag_req_v2 inet_diag_req_compat -> inet_diag_req After this the changes in the user-space visible part of the inet_diag.h are * add the inet_diag_req_v2 * extend the INET_DIAG_ extensions list The set applies to current net tree. David, Stephen, I'm a little bit confused with the situation with this header. Can you please comment on the (quite ancient) commit 73c1f4a0 ([TCPDIAG]: Just rename everything to inet_diag) from acme@? This commit did the similar thing with inet_diag header -- flushed away the old struct and introduced a new one keeping just the binary compatibility. Why was the breakage of tools caused by this commit considered as acceptable? Thanks, Pavel