From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754963AbYIWCh2 (ORCPT ); Mon, 22 Sep 2008 22:37:28 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754175AbYIWChQ (ORCPT ); Mon, 22 Sep 2008 22:37:16 -0400 Received: from bc.sympatico.ca ([209.226.175.184]:40851 "EHLO tomts22-srv.bellnexxia.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754074AbYIWChP (ORCPT ); Mon, 22 Sep 2008 22:37:15 -0400 X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: AqEEAFTt10hMQWq+/2dsb2JhbACBXbUcgWY Date: Mon, 22 Sep 2008 22:36:54 -0400 From: Mathieu Desnoyers To: Peter Zijlstra Cc: Steven Rostedt , Martin Bligh , Linux Kernel Mailing List , Linus Torvalds , Thomas Gleixner , od@novell.com, "Frank Ch. Eigler" Subject: Re: Unified tracing buffer Message-ID: <20080923023654.GF24937@Krystal> References: <33307c790809191433w246c0283l55a57c196664ce77@mail.gmail.com> <1222077437.7470.5.camel@lappy.programming.kicks-ass.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Disposition: inline In-Reply-To: <1222077437.7470.5.camel@lappy.programming.kicks-ass.net> X-Editor: vi X-Info: http://krystal.dyndns.org:8080 X-Operating-System: Linux/2.6.21.3-grsec (i686) X-Uptime: 22:31:40 up 110 days, 7:12, 7 users, load average: 1.04, 0.45, 0.34 User-Agent: Mutt/1.5.16 (2007-06-11) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org * Peter Zijlstra (a.p.zijlstra@chello.nl) wrote: > On Sat, 2008-09-20 at 05:03 -0400, Steven Rostedt wrote: > > > Oh, and all commands should start with the namespace. > > > > ring_buffer_alloc() > > ring_buffer_free() > > ring_buffer_record_event() > > I really think we should separate the ringbuffer management from the > event stuff. > Sure, I am strongly in favor of separating those two, given they represent two different things. However, the requirement I have heard at KS2008 was to provide - Unified buffering mechanism - Timestamps synchronized across all buffers - Unified event IDs management, so events from various sources could be shared between tools. - As of my understanding, unified event structure, which can be exported to userspace and be shared across different tools. - Unified buffer control/management mechanism. These all represent different infrastructure parts, but are all needed if we want tools to be able to share the data exported through those buffers. Relay is a good example of having only a _single_ of these layers in common : there is currently no way the different relay users can share the data they collect because they have simply no idea how others structure their data. Mathieu -- Mathieu Desnoyers OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F BA06 3F25 A8FE 3BAE 9A68