From mboxrd@z Thu Jan 1 00:00:00 1970 From: Song Liu Subject: Re: [BUG] perf: intel_pt won't display kernel function Date: Wed, 3 Apr 2019 18:59:51 +0000 Message-ID: <425141B8-8908-4D48-90FD-8806BA9E9E12@fb.com> References: <20190403143738.GB32001@krava> <20190403145353.GE32553@kernel.org> <20190403151548.GF32553@kernel.org> <20190403185053.GH32553@kernel.org> <124332A9-146A-446F-A82A-910C8319170E@fb.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <124332A9-146A-446F-A82A-910C8319170E@fb.com> Content-Language: en-US Content-ID: <548D2E056D721F49BDB858F1DD886E8E@namprd15.prod.outlook.com> Sender: linux-kernel-owner@vger.kernel.org To: Arnaldo Carvalho de Melo Cc: Jiri Olsa , Adrian Hunter , Andi Kleen , "jolsa@kernel.org" , "namhyung@kernel.org" , "linux-kernel@vger.kernel.org" , "linux-perf-users@vger.kernel.org" , Andi Kleen List-Id: linux-perf-users.vger.kernel.org > On Apr 3, 2019, at 11:55 AM, Song Liu wrote: >=20 >=20 >=20 >> On Apr 3, 2019, at 11:50 AM, Arnaldo Carvalho de Melo wrote: >>=20 >> Em Wed, Apr 03, 2019 at 04:27:56PM +0000, Song Liu escreveu: >>>=20 >>>=20 >>>> On Apr 3, 2019, at 8:15 AM, Arnaldo Carvalho de Melo wrote: >>>>=20 >>>> Em Wed, Apr 03, 2019 at 11:53:53AM -0300, Arnaldo Carvalho de Melo esc= reveu: >>>>> Em Wed, Apr 03, 2019 at 04:37:38PM +0200, Jiri Olsa escreveu: >>>>>> hi, >>>>>> perf script --call-trace stop working for me recently, >>>>>> and displays only user space functions >>>>>>=20 >>>>>> I bisected that to: >>>>>> 7b612e291a5a perf tools: Synthesize PERF_RECORD_* for loaded BPF pro= grams >>>>>>=20 >>>>>> data from following comands will display user space functions only: >>>>>> # perf-with-kcore record pt -e intel_pt// -- ls >>>>>> # perf-with-kcore script pt --call-trace >>>>>>=20 >>>>>> when I disable the bpf synthesizing (patch below), kernel functions = are back >>>>>>=20 >>>>>> I guess the new events mess up with intel_pt decoder somehow >>>>>=20 >>>>> I.e. I'm adding the patch below to my perf/urgent branch. >>>>=20 >>>> Song, that is what I have, can I have your Acked-by and please conside= r >>>> taking a look at the bug Jiri reported, >>>>=20 >>>> Thanks, >>>>=20 >>>> - Arnaldo >>>=20 >>> Current logic with --no-bpf-event is to generate PERF_RECORD_KSYMBOL, b= ut not >>> PERF_RECORD_BPF_EVENT: >>=20 >> I see... The opts is done later, after querying the kernel for existing >> BPF programs so that at least the KSYMBOL ones can be generated. So I'll >> keep it as is, no need for this patch. >>=20 >>> https://urldefense.proofpoint.com/v2/url?u=3Dhttps-3A__elixir.bootlin.c= om_linux_v5.1-2Drc3_source_tools_perf_util_bpf-2Devent.c-23L254&d=3DDwIBAg&= c=3D5VD0RTtNlTh3ycd41b3MUw&r=3DdR8692q0_uaizy0jkrBJQM5k2hfm4CiFxYT8KaysFrg&= m=3D8ITYgIrqFBIy1uQFQPHrsX5kBX5RlmhwcknKYJa2eGs&s=3DMdB2zA9JstwC93qKEonsfTQ= eQYiw-yeOLbPxm3QMRo8&e=3D >>>=20 >>> I will look into the intel_pt problem.=20 >>=20 >> Ok. >>=20 >>> In the meanwhile, let's fix it for now.=20 >>=20 >> I'll wait a bit. >=20 > I guess something like this might be a better fix? (Sorry for missing=20 > commit message): >=20 > diff --git i/tools/perf/util/map.c w/tools/perf/util/map.c > index e32628cd20a7..741430a35dca 100644 > --- i/tools/perf/util/map.c > +++ w/tools/perf/util/map.c > @@ -261,6 +261,12 @@ bool __map__is_extra_kernel_map(const struct map *ma= p) > return kmap && kmap->name[0]; > } >=20 > +bool __map__is_bpf_prog(const struct map *map) > +{ > + const char *name =3D map->dso->short_name; > + return name && (strstr(name, "bpf_prog_") =3D=3D name); > +} > + > bool map__has_symbols(const struct map *map) > { > return dso__has_symbols(map->dso); > diff --git i/tools/perf/util/map.h w/tools/perf/util/map.h > index 0e20749f2c55..01079f1f4375 100644 > --- i/tools/perf/util/map.h > +++ w/tools/perf/util/map.h > @@ -159,10 +159,12 @@ int map__set_kallsyms_ref_reloc_sym(struct map *map= , const char *symbol_name, >=20 > bool __map__is_kernel(const struct map *map); > bool __map__is_extra_kernel_map(const struct map *map); > +bool __map__is_bpf_prog(const struct map *map); >=20 > static inline bool __map__is_kmodule(const struct map *map) > { > - return !__map__is_kernel(map) && !__map__is_extra_kernel_map(map)= ; > + return !__map__is_kernel(map) && !__map__is_extra_kernel_map(map)= && > + !__map__is_bpf_prog(map); > } >=20 > bool map__has_symbols(const struct map *map); >=20 > Thanks, > Song Actually, it should be something like this: diff --git i/tools/perf/util/map.c w/tools/perf/util/map.c index e32628cd20a7..31cd23612529 100644 --- i/tools/perf/util/map.c +++ w/tools/perf/util/map.c @@ -261,6 +261,11 @@ bool __map__is_extra_kernel_map(const struct map *map) return kmap && kmap->name[0]; } +bool __map__is_bpf_prog(const struct map *map) +{ + return map->dso->binary_type =3D=3D DSO_BINARY_TYPE__BPF_PROG_INFO; +} + bool map__has_symbols(const struct map *map) { return dso__has_symbols(map->dso); diff --git i/tools/perf/util/map.h w/tools/perf/util/map.h index 0e20749f2c55..01079f1f4375 100644 --- i/tools/perf/util/map.h +++ w/tools/perf/util/map.h @@ -159,10 +159,12 @@ int map__set_kallsyms_ref_reloc_sym(struct map *map, = const char *symbol_name, bool __map__is_kernel(const struct map *map); bool __map__is_extra_kernel_map(const struct map *map); +bool __map__is_bpf_prog(const struct map *map); static inline bool __map__is_kmodule(const struct map *map) { - return !__map__is_kernel(map) && !__map__is_extra_kernel_map(map); + return !__map__is_kernel(map) && !__map__is_extra_kernel_map(map) &= & + !__map__is_bpf_prog(map); } bool map__has_symbols(const struct map *map); >=20 >>=20 >>> Acked-by: Song Liu >>>=20 >>> Thanks,=20 >>> Song >>>=20 >>>=20 >>>> commit 011318ccc2024ba03e96c32a06f74ca5d6ab5503 >>>> Author: Arnaldo Carvalho de Melo >>>> Date: Wed Apr 3 12:05:15 2019 -0300 >>>>=20 >>>> perf record: Do not synthesize BPF records when --no-bpf-event is use= d >>>>=20 >>>> By default we synthesize and ask the kernel for BPF events, having a >>>> --no-bpf-event option to disable that, which can be useful, for >>>> instance, if there are still bugs in that code, which seems to be the >>>> case as reported by Jiri Olsa in: >>>>=20 >>>> "[BUG] perf: intel_pt won't display kernel function" >>>> https://lore.kernel.org/lkml/20190403143738.GB32001@krava >>>>=20 >>>> So add the check for record_opts.no_bpf_event when considering >>>> synthesizing BPF events for pre-existing BPF programs in 'perf record= '. >>>>=20 >>>> The reported bug needs further analysis and is a separate matter. >>>>=20 >>>> Cc: Adrian Hunter >>>> Cc: Alexei Starovoitov >>>> Cc: Andrii Nakryiko >>>> Cc: Daniel Borkmann >>>> Cc: Jiri Olsa >>>> Cc: Martin KaFai Lau >>>> Cc: Namhyung Kim >>>> Cc: Peter Zijlstra >>>> Cc: Song Liu >>>> Cc: Yonghong Song >>>> Fixes: 7b612e291a5a ("perf tools: Synthesize PERF_RECORD_* for loaded= BPF programs") >>>> Link: https://lore.kernel.org/lkml/20190403145353.GE32553@kernel.org >>>> Signed-off-by: Arnaldo Carvalho de Melo >>>>=20 >>>> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c >>>> index 4e2d953d4bc5..17d772f192ad 100644 >>>> --- a/tools/perf/builtin-record.c >>>> +++ b/tools/perf/builtin-record.c >>>> @@ -1114,10 +1114,11 @@ static int record__synthesize(struct record *r= ec, bool tail) >>>> return err; >>>> } >>>>=20 >>>> - err =3D perf_event__synthesize_bpf_events(session, process_synthesiz= ed_event, >>>> - machine, opts); >>>> - if (err < 0) >>>> - pr_warning("Couldn't synthesize bpf events.\n"); >>>> + if (!opts->no_bpf_event) { >>>> + err =3D perf_event__synthesize_bpf_events(session, process_synthesi= zed_event, machine, opts); >>>> + if (err < 0) >>>> + pr_warning("Couldn't synthesize bpf events.\n"); >>>> + } >>>>=20 >>>> err =3D __machine__synthesize_threads(machine, tool, &opts->target, r= ec->evlist->threads, >>>> process_synthesized_event, opts->sample_address, >>=20 >> --=20 >>=20 >> - Arnaldo