Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754263Ab3JHMyN (ORCPT ); Tue, 8 Oct 2013 08:54:13 -0400 Received: from e23smtp07.au.ibm.com ([202.81.31.140]:34299 "EHLO e23smtp07.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751796Ab3JHMyM (ORCPT ); Tue, 8 Oct 2013 08:54:12 -0400 Message-ID: <5254005F.4010103@linux.vnet.ibm.com> Date: Tue, 08 Oct 2013 18:23:51 +0530 From: Hemant User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:17.0) Gecko/20130625 Thunderbird/17.0.7 MIME-Version: 1.0 To: Masami Hiramatsu CC: linux-kernel@vger.kernel.org, srikar@linux.vnet.ibm.com, peterz@infradead.org, oleg@redhat.com, hegdevasant@linux.vnet.ibm.com, mingo@redhat.com, anton@redhat.com, systemtap@sourceware.org, namhyung@kernel.org, aravinda@linux.vnet.ibm.com Subject: Re: [PATCH v2 0/3] Perf support to SDT markers References: <20131007063911.11693.33624.stgit@hemant-fedora> <5253A04E.4020809@hitachi.com> In-Reply-To: <5253A04E.4020809@hitachi.com> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit X-TM-AS-MML: No X-Content-Scanned: Fidelis XPS MAILER x-cbid: 13100812-0260-0000-0000-000003BF15F9 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2706 Lines: 91 On 10/08/2013 11:33 AM, Masami Hiramatsu wrote: > Hi, Hi Masami and thanks a lot for testing these patches. > > (2013/10/07 15:46), Hemant Kumar wrote: >> This patchset helps in probing dtrace style markers(SDT) present in user space >> applications through perf. Notes/markes are placed at important places by the >> developers. They have a negligible overhead when not enabled. We can enable >> them and probe at these places and find some important information like the >> arguments' values, etc. > Thank you for updating! > By the way, could you rebase this series to the latest -tip kernel next time? > It is more convenient for us to test & merge it. Yes, I will rebase this series next time. > > [...] >> - Now use perf to list the markers in the app: >> # perf probe --markers -x ./user_app >> >> %user_app:foo_start >> %user_app:fun_start > Here, what I got. > > $ ./perf probe --markers -x /lib64/librt.so > %librt:lll_futex_wake > > OK, it lists up markers. > > $ ./perf probe --markers -x ./perf > SDT markers not present! > Error : Failed to find SDT markers!(-1) > > OK, this shows an error, INHO, it'd better show which binary doesn't have markers. OK, will include the binary name too. > > $ ./perf probe --markers > Failed to open (null) > Error : Failed to find SDT markers!(-1) > > Hmm, it's a bit odd error. It should check if there is an elf binary is specified. Yeah, correct. > > $ ./perf probe --markers /lib64/librt.so > %librt:lll_futex_wake > > It seems to work without -x. However, that is only when the path is started with /. > And also the usage is not described by --help. > > $ ./perf probe --markers ./perf > Semantic error :File always requires line number or lazy pattern. > Error: Parse Error. (-22) > > So, to avoid confusion, --markers should abort without -x(--exec) option. Yes, we can restrict action of --markers for -x option only. I thought of keeping the same behavior as in case of function listing but I guess restricting it would be better since, --markers should only be available in user space. > >> - And then use perf probe to add a probe point : >> >> # perf probe -x ./user_app -a '%user_app:foo_start' > And also, without -x, -a also exits with an odd error message. > > $ ./perf probe -a %libm:slowexp_p6 > Failed to open (null) > Error: Failed to add events. (-1) Ok, will rectify this behavior. > > I'll review each patch soon. Thanks, looking forward to that. - Hemant -- 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/