| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CAP-5=fU6eUUJeoKTsO0nv=yFvG=Wu_z_1VVpyyWNx42h+BpSKQ@mail.gmail.com>
Date: Fri, 28 Jun 2024 11:34:51 -0700
From: Ian Rogers <irogers@...gle.com>
To: Adrian Hunter <adrian.hunter@...el.com>
Cc: Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>, Namhyung Kim <namhyung@...nel.org>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>, Jiri Olsa <jolsa@...nel.org>,
Kan Liang <kan.liang@...ux.intel.com>, John Garry <john.g.garry@...cle.com>,
Will Deacon <will@...nel.org>, James Clark <james.clark@....com>,
Mike Leach <mike.leach@...aro.org>, Leo Yan <leo.yan@...ux.dev>,
Suzuki K Poulose <suzuki.poulose@....com>, Yicong Yang <yangyicong@...ilicon.com>,
Jonathan Cameron <jonathan.cameron@...wei.com>, Nick Terrell <terrelln@...com>,
Nick Desaulniers <ndesaulniers@...gle.com>, Oliver Upton <oliver.upton@...ux.dev>,
Anshuman Khandual <anshuman.khandual@....com>, Song Liu <song@...nel.org>,
Ilkka Koskinen <ilkka@...amperecomputing.com>, Huacai Chen <chenhuacai@...nel.org>,
Yanteng Si <siyanteng@...ngson.cn>, Sun Haiyong <sunhaiyong@...ngson.cn>,
linux-kernel@...r.kernel.org, linux-perf-users@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH v2 01/27] perf auxevent: Zero size dummy tool
On Fri, Jun 28, 2024 at 10:45 AM Adrian Hunter <adrian.hunter@...el.com> wrote:
>
> In subject: "auxevent" -> "auxtrace"
Ack.
> On 26/06/24 23:36, Ian Rogers wrote:
> > The dummy tool is passed as a placeholder to allow a container_of to
> > get additional parameters. As the tool isn't used, make it a zero
> > sized array saving 320 bytes on an x86_64 build.
> >
> > s390-cpumsf's dummy tool struct was unused, so remove.
> >
> > Signed-off-by: Ian Rogers <irogers@...gle.com>
> > ---
> > tools/perf/util/arm-spe.c | 6 +++---
> > tools/perf/util/cs-etm.c | 6 +++---
> > tools/perf/util/intel-bts.c | 6 +++---
> > tools/perf/util/intel-pt.c | 7 +++----
> > tools/perf/util/s390-cpumsf.c | 5 -----
> > 5 files changed, 12 insertions(+), 18 deletions(-)
> >
> > diff --git a/tools/perf/util/arm-spe.c b/tools/perf/util/arm-spe.c
> > index afbd5869f6bf..ee0d5064ddd4 100644
> > --- a/tools/perf/util/arm-spe.c
> > +++ b/tools/perf/util/arm-spe.c
> > @@ -1074,8 +1074,8 @@ static void arm_spe_print_info(__u64 *arr)
> > }
> >
> > struct arm_spe_synth {
> > - struct perf_tool dummy_tool;
> > struct perf_session *session;
> > + struct perf_tool dummy_tool[0];
>
> [] is preferred to [0]
>
> > };
> >
> > static int arm_spe_event_synth(struct perf_tool *tool,
> > @@ -1084,7 +1084,7 @@ static int arm_spe_event_synth(struct perf_tool *tool,
> > struct machine *machine __maybe_unused)
> > {
> > struct arm_spe_synth *arm_spe_synth =
> > - container_of(tool, struct arm_spe_synth, dummy_tool);
> > + container_of(tool, struct arm_spe_synth, dummy_tool[0]);
> >
> > return perf_session__deliver_synth_event(arm_spe_synth->session,
> > event, NULL);
> > @@ -1098,7 +1098,7 @@ static int arm_spe_synth_event(struct perf_session *session,
> > memset(&arm_spe_synth, 0, sizeof(struct arm_spe_synth));
> > arm_spe_synth.session = session;
> >
> > - return perf_event__synthesize_attr(&arm_spe_synth.dummy_tool, attr, 1,
> > + return perf_event__synthesize_attr(arm_spe_synth.dummy_tool, attr, 1,
>
> Passing a pointer to an object that doesn't exist there is NAK IMO
Not disagreeing but I think it is at least more intention revealing
and efficient than passing a struct that isn't initialized because we
know the tool won't be used.
> It would be better to just write another version of
> perf_event__synthesize_attr().
I think we've gotten ourselves into this mess because we're doing poor
man's OO and we're trying to pass values via a tools/container_of
rather than just pass a value. I'm not keen on making this change
larger but I'll see if this can be done and we just explicitly pass
some "void* data" like value.
Thanks,
Ian
> > &id, arm_spe_event_synth);
> > }
> >
> > diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c
> > index 32818bd7cd17..9fd2967d4e3f 100644
> > --- a/tools/perf/util/cs-etm.c
> > +++ b/tools/perf/util/cs-etm.c
> > @@ -1596,8 +1596,8 @@ static int cs_etm__synth_branch_sample(struct cs_etm_queue *etmq,
> > }
> >
> > struct cs_etm_synth {
> > - struct perf_tool dummy_tool;
> > struct perf_session *session;
> > + struct perf_tool dummy_tool[0];
> > };
> >
> > static int cs_etm__event_synth(struct perf_tool *tool,
> > @@ -1606,7 +1606,7 @@ static int cs_etm__event_synth(struct perf_tool *tool,
> > struct machine *machine __maybe_unused)
> > {
> > struct cs_etm_synth *cs_etm_synth =
> > - container_of(tool, struct cs_etm_synth, dummy_tool);
> > + container_of(tool, struct cs_etm_synth, dummy_tool[0]);
> >
> > return perf_session__deliver_synth_event(cs_etm_synth->session,
> > event, NULL);
> > @@ -1620,7 +1620,7 @@ static int cs_etm__synth_event(struct perf_session *session,
> > memset(&cs_etm_synth, 0, sizeof(struct cs_etm_synth));
> > cs_etm_synth.session = session;
> >
> > - return perf_event__synthesize_attr(&cs_etm_synth.dummy_tool, attr, 1,
> > + return perf_event__synthesize_attr(cs_etm_synth.dummy_tool, attr, 1,
> > &id, cs_etm__event_synth);
> > }
> >
> > diff --git a/tools/perf/util/intel-bts.c b/tools/perf/util/intel-bts.c
> > index ec1b3bd9f530..fb8fec3a3c36 100644
> > --- a/tools/perf/util/intel-bts.c
> > +++ b/tools/perf/util/intel-bts.c
> > @@ -738,8 +738,8 @@ static bool intel_bts_evsel_is_auxtrace(struct perf_session *session,
> > }
> >
> > struct intel_bts_synth {
> > - struct perf_tool dummy_tool;
> > struct perf_session *session;
> > + struct perf_tool dummy_tool[0];
> > };
> >
> > static int intel_bts_event_synth(struct perf_tool *tool,
> > @@ -748,7 +748,7 @@ static int intel_bts_event_synth(struct perf_tool *tool,
> > struct machine *machine __maybe_unused)
> > {
> > struct intel_bts_synth *intel_bts_synth =
> > - container_of(tool, struct intel_bts_synth, dummy_tool);
> > + container_of(tool, struct intel_bts_synth, dummy_tool[0]);
> >
> > return perf_session__deliver_synth_event(intel_bts_synth->session,
> > event, NULL);
> > @@ -762,7 +762,7 @@ static int intel_bts_synth_event(struct perf_session *session,
> > memset(&intel_bts_synth, 0, sizeof(struct intel_bts_synth));
> > intel_bts_synth.session = session;
> >
> > - return perf_event__synthesize_attr(&intel_bts_synth.dummy_tool, attr, 1,
> > + return perf_event__synthesize_attr(intel_bts_synth.dummy_tool, attr, 1,
> > &id, intel_bts_event_synth);
> > }
> >
> > diff --git a/tools/perf/util/intel-pt.c b/tools/perf/util/intel-pt.c
> > index d6d7b7512505..b8b90773baa2 100644
> > --- a/tools/perf/util/intel-pt.c
> > +++ b/tools/perf/util/intel-pt.c
> > @@ -3660,8 +3660,8 @@ static int intel_pt_queue_data(struct perf_session *session,
> > }
> >
> > struct intel_pt_synth {
> > - struct perf_tool dummy_tool;
> > struct perf_session *session;
> > + struct perf_tool dummy_tool[0];
> > };
> >
> > static int intel_pt_event_synth(struct perf_tool *tool,
> > @@ -3670,7 +3670,7 @@ static int intel_pt_event_synth(struct perf_tool *tool,
> > struct machine *machine __maybe_unused)
> > {
> > struct intel_pt_synth *intel_pt_synth =
> > - container_of(tool, struct intel_pt_synth, dummy_tool);
> > + container_of(tool, struct intel_pt_synth, dummy_tool[0]);
> >
> > return perf_session__deliver_synth_event(intel_pt_synth->session, event,
> > NULL);
> > @@ -3687,8 +3687,7 @@ static int intel_pt_synth_event(struct perf_session *session, const char *name,
> >
> > memset(&intel_pt_synth, 0, sizeof(struct intel_pt_synth));
> > intel_pt_synth.session = session;
> > -
> > - err = perf_event__synthesize_attr(&intel_pt_synth.dummy_tool, attr, 1,
> > + err = perf_event__synthesize_attr(intel_pt_synth.dummy_tool, attr, 1,
> > &id, intel_pt_event_synth);
> > if (err)
> > pr_err("%s: failed to synthesize '%s' event type\n",
> > diff --git a/tools/perf/util/s390-cpumsf.c b/tools/perf/util/s390-cpumsf.c
> > index 6fe478b0b61b..4ec583e511af 100644
> > --- a/tools/perf/util/s390-cpumsf.c
> > +++ b/tools/perf/util/s390-cpumsf.c
> > @@ -952,11 +952,6 @@ s390_cpumsf_process_event(struct perf_session *session,
> > return err;
> > }
> >
> > -struct s390_cpumsf_synth {
> > - struct perf_tool cpumsf_tool;
> > - struct perf_session *session;
> > -};
>
> Should really be a separate patch
>
> > -
> > static int
> > s390_cpumsf_process_auxtrace_event(struct perf_session *session,
> > union perf_event *event __maybe_unused,
>
Powered by blists - more mailing lists