From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751454AbbIFHsP (ORCPT ); Sun, 6 Sep 2015 03:48:15 -0400 Received: from szxga03-in.huawei.com ([119.145.14.66]:12262 "EHLO szxga03-in.huawei.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751095AbbIFHsI (ORCPT ); Sun, 6 Sep 2015 03:48:08 -0400 Message-ID: <55EBEF99.8010506@huawei.com> Date: Sun, 6 Sep 2015 15:47:37 +0800 From: "Wangnan (F)" User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0 MIME-Version: 1.0 To: Namhyung Kim CC: Arnaldo Carvalho de Melo , Ingo Molnar , Peter Zijlstra , Jiri Olsa , LKML , Masami Hiramatsu , Subject: Re: [PATCH v2 1/5] perf probe: Split add_perf_probe_events() References: <1441368963-11565-1-git-send-email-namhyung@kernel.org> In-Reply-To: <1441368963-11565-1-git-send-email-namhyung@kernel.org> Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: 7bit X-Originating-IP: [10.111.66.109] X-CFilter-Loop: Reflected X-Mirapoint-Virus-RAPID-Raw: score=unknown(0), refid=str=0001.0A020205.55EBEF9C.0123,ss=1,re=0.000,recu=0.000,reip=0.000,cl=1,cld=1,fgs=0, ip=0.0.0.0, so=2013-05-26 15:14:31, dmn=2013-03-21 17:37:32 X-Mirapoint-Loop-Id: 0239eb82dcf045dec5f127bd655f6492 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Namhyung, Thanks for this patchset. Could you plase have a look at patch 5/27 and 6/27 in my newest pull request? These 2 patches utilize new probing API to create probe point and collect probe_trace_events. I'm not very sure I fully understand your design principle, especially the cleanup part, because I can see different functions dealing with cleanup: cleanup_perf_probe_events del_perf_probe_events clear_perf_probe_event clear_probe_trace_event But non of them works perfectly for me. In bpf_prog_priv__clear() function of 6/27, I copied some code from cleanup_perf_probe_events(), because I think when destroying bpf programs, the probe_trace_events should also be cleanuped, but we don't need call exit_symbol_maps() many times, because we are in 'perf record', and not sure whether other parts of perf need symbol maps. Otherwise I think directly calling cleanup_perf_probe_events() sould be better. You can find patch from: http://lkml.kernel.org/n/1441523623-152703-6-git-send-email-wangnan0@huawei.com http://lkml.kernel.org/n/1441523623-152703-7-git-send-email-wangnan0@huawei.com Thank you. On 2015/9/4 20:15, Namhyung Kim wrote: > The add_perf_probe_events() does 3 things: > > 1. convert all perf events to trace events > 2. add all trace events to kernel > 3. cleanup all trace events > > But sometimes we need to do something with the trace events. So split > the funtion into three, so that it can access intermediate trace events > via struct __event_package if needed. > > Acked-by: Masami Hiramatsu > Signed-off-by: Namhyung Kim > --- > tools/perf/util/probe-event.c | 39 +++++++++++++++++++++++++++++++++++---- > 1 file changed, 35 insertions(+), 4 deletions(-) > > diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c > index eb5f18b75402..2c762f41e7a5 100644 > --- a/tools/perf/util/probe-event.c > +++ b/tools/perf/util/probe-event.c > @@ -2765,9 +2765,10 @@ struct __event_package { > int ntevs; > }; > > -int add_perf_probe_events(struct perf_probe_event *pevs, int npevs) > +static int convert_perf_probe_events(struct perf_probe_event *pevs, int npevs, > + struct __event_package **ppkgs) > { > - int i, j, ret; > + int i, ret; > struct __event_package *pkgs; > > ret = 0; > @@ -2792,12 +2793,21 @@ int add_perf_probe_events(struct perf_probe_event *pevs, int npevs) > ret = convert_to_probe_trace_events(pkgs[i].pev, > &pkgs[i].tevs); > if (ret < 0) > - goto end; > + return ret; > pkgs[i].ntevs = ret; > } > /* This just release blacklist only if allocated */ > kprobe_blacklist__release(); > > + *ppkgs = pkgs; > + > + return 0; > +} > + > +static int apply_perf_probe_events(struct __event_package *pkgs, int npevs) > +{ > + int i, ret = 0; > + > /* Loop 2: add all events */ > for (i = 0; i < npevs; i++) { > ret = __add_probe_trace_events(pkgs[i].pev, pkgs[i].tevs, > @@ -2806,7 +2816,16 @@ int add_perf_probe_events(struct perf_probe_event *pevs, int npevs) > if (ret < 0) > break; > } > -end: > + return ret; > +} > + > +static void cleanup_perf_probe_events(struct __event_package *pkgs, int npevs) > +{ > + int i, j; > + > + if (pkgs == NULL) > + return; > + > /* Loop 3: cleanup and free trace events */ > for (i = 0; i < npevs; i++) { > for (j = 0; j < pkgs[i].ntevs; j++) > @@ -2815,6 +2834,18 @@ end: > } > free(pkgs); > exit_symbol_maps(); > +} > + > +int add_perf_probe_events(struct perf_probe_event *pevs, int npevs) > +{ > + int ret; > + struct __event_package *pkgs = NULL; > + > + ret = convert_perf_probe_events(pevs, npevs, &pkgs); > + if (ret == 0) > + ret = apply_perf_probe_events(pkgs, npevs); > + > + cleanup_perf_probe_events(pkgs, npevs); > > return ret; > }