From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH v2 0/2] Tracepoint for tcp retransmission Date: Sun, 05 Feb 2012 16:48:12 -0500 (EST) Message-ID: <20120205.164812.383526470219039003.davem@davemloft.net> References: <20120205125325.GA31578@elastic.org> <20120205191708.GA5337@neilslaptop.think-freely.org> <20120205200428.GB2816@nuttenaction> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: nhorman@tuxdriver.com, fche@elastic.org, satoru.moriya@hds.com, netdev@vger.kernel.org, tgraf@infradead.org, stephen.hemminger@vyatta.com, eric.dumazet@gmail.com, seiji.aguchi@hds.com, fche@sourceware.org, mathieu.desnoyers@polymtl.ca, rostedt@goodmis.org To: hagen@jauu.net Return-path: Received: from shards.monkeyblade.net ([198.137.202.13]:43332 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752003Ab2BEVwQ (ORCPT ); Sun, 5 Feb 2012 16:52:16 -0500 In-Reply-To: <20120205200428.GB2816@nuttenaction> Sender: netdev-owner@vger.kernel.org List-ID: Let me put an end to this dicussion right now, I'm not applying any patches which just add tracepoints into the TCP stack. That's not the solution, and you can hoot and hollar all you want about it, I simply do not care. The solution is a way to manage socket events, and be able to choose which sockets emit those events, in a useful way, via the perf event infrastructure. This means we need a perf event socket layer that allows filtering on socket flow ID, which enables per-socket event and statistics gathering, and so forth.