Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756104AbZLJVWQ (ORCPT ); Thu, 10 Dec 2009 16:22:16 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753448AbZLJVWO (ORCPT ); Thu, 10 Dec 2009 16:22:14 -0500 Received: from va3ehsobe002.messaging.microsoft.com ([216.32.180.12]:26210 "EHLO VA3EHSOBE002.bigfish.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750963AbZLJVWN (ORCPT ); Thu, 10 Dec 2009 16:22:13 -0500 X-SpamScore: -21 X-BigFish: VPS-21(zz1432R98dN936eMzz1202hz4fhzz2fh6bh61h) X-Spam-TCS-SCL: 0:0 Message-ID: <4B216670.60807@am.sony.com> Date: Thu, 10 Dec 2009 13:21:52 -0800 From: Tim Bird User-Agent: Thunderbird 2.0.0.14 (X11/20080501) MIME-Version: 1.0 To: "rostedt@goodmis.org" CC: "Bird, Tim" , linux kernel , Ingo Molnar , Frederic Weisbecker Subject: Re: [PATCH 4/4] ftrace - function_duration Documentation References: <4B2027C7.9010509@am.sony.com> <1260416840.2146.134.camel@gandalf.stny.rr.com> In-Reply-To: <1260416840.2146.134.camel@gandalf.stny.rr.com> Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 10 Dec 2009 21:21:53.0756 (UTC) FILETIME=[CB8349C0:01CA79DE] X-SEL-encryption-scan: scanned X-Reverse-DNS: mail8.fw-sd.sony.com Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1611 Lines: 38 Steven Rostedt wrote: > On Wed, 2009-12-09 at 14:42 -0800, Tim Bird wrote: >> Documentation for function_duration tracer. > >> +The purpose of the function duration tracer is to help find the >> +longest-running, most time-consuming functions in the Linux kernel. >> +This is especially important for early initialization, to reduce >> +kernel boot time. >> + >> +Function duration tracing uses the ftrace ability to hook both the >> +entry and exit of a function, to record duration information for >> +kernel functions. This is similar to the function_graph tracer, but >> +the function_duration tracer is specially written to support filtering >> +by duration and decreased tracer overhead while filtering. >> + >> +This extends the amount of time that a trace can cover, and reduces >> +interference with the timing of the traced activity. > > BTW, have you taken a look at the function profiler? It also keeps track > of all durations without using the ring buffer. It just adds them up and > gives an average time. I haven't, but I should. However, my tool has some other options (like subtracting time spent in interrupts) that I don't think the profile approach currently supports. -- Tim ============================= Tim Bird Architecture Group Chair, CE Linux Forum Senior Staff Engineer, Sony Corporation of America ============================= -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/