* [PATCH v6 00/27] Constify tool pointers
@ 2024-07-18 0:59 Ian Rogers
2024-07-18 0:59 ` [PATCH v6 01/27] perf auxtrace: Remove dummy tools Ian Rogers
` (27 more replies)
0 siblings, 28 replies; 40+ messages in thread
From: Ian Rogers @ 2024-07-18 0:59 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa,
Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon,
James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang,
Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton,
Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev,
Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel,
linux-perf-users, linux-arm-kernel
struct perf_tool provides a set of function pointers that are called
through when processing perf data. To make filling the pointers less
cumbersome, if they are NULL perf_tools__fill_defaults will add
default do nothing implementations.
This change refactors struct perf_tool to have an init function that
provides the default implementation. The special use of NULL and
perf_tools__fill_defaults are removed. As a consequence the tool
pointers can then all be made const, which better reflects the
behavior a particular perf command would expect of the tool and to
some extent can reduce the cognitive load on someone working on a
command.
v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by.
v5: Rebase dropping asan fix merged by Namhyung.
v4: Simplify perf_session__deliver_synth_attr_event following Adrian's
suggestions.
v3: Just remove auxtrace dummy tools [Adrian] and make s390-cpumsf
struct removal its own patch [Adrian].
v2: Remove dummy tool initialization [Adrian] and make zero sized. Add
cs-etm fix for address sanitizer build, found necessary when
testing dummy tool change.
Ian Rogers (27):
perf auxtrace: Remove dummy tools
perf s390-cpumsf: Remove unused struct
perf tool: Constify tool pointers
perf tool: Move fill defaults into tool.c
perf tool: Add perf_tool__init
perf kmem: Use perf_tool__init
perf buildid-list: Use perf_tool__init
perf kvm: Use perf_tool__init
perf lock: Use perf_tool__init
perf evlist: Use perf_tool__init
perf record: Use perf_tool__init
perf c2c: Use perf_tool__init
perf script: Use perf_tool__init
perf inject: Use perf_tool__init
perf report: Use perf_tool__init
perf stat: Use perf_tool__init
perf annotate: Use perf_tool__init
perf sched: Use perf_tool__init
perf mem: Use perf_tool__init
perf timechart: Use perf_tool__init
perf diff: Use perf_tool__init
perf data convert json: Use perf_tool__init
perf data convert ctf: Use perf_tool__init
perf test event_update: Ensure tools is initialized
perf kwork: Use perf_tool__init
perf tool: Remove perf_tool__fill_defaults
perf session: Constify tool
tools/perf/arch/x86/util/event.c | 4 +-
tools/perf/bench/synthesize.c | 2 +-
tools/perf/builtin-annotate.c | 44 ++--
tools/perf/builtin-buildid-list.c | 10 +
tools/perf/builtin-c2c.c | 33 ++-
tools/perf/builtin-diff.c | 30 ++-
tools/perf/builtin-evlist.c | 10 +-
tools/perf/builtin-inject.c | 159 ++++++------
tools/perf/builtin-kmem.c | 20 +-
tools/perf/builtin-kvm.c | 19 +-
tools/perf/builtin-kwork.c | 33 ++-
tools/perf/builtin-lock.c | 41 ++--
tools/perf/builtin-mem.c | 37 +--
tools/perf/builtin-record.c | 47 ++--
tools/perf/builtin-report.c | 67 +++--
tools/perf/builtin-sched.c | 50 ++--
tools/perf/builtin-script.c | 106 ++++----
tools/perf/builtin-stat.c | 26 +-
tools/perf/builtin-timechart.c | 25 +-
tools/perf/builtin-top.c | 2 +-
tools/perf/builtin-trace.c | 4 +-
tools/perf/tests/cpumap.c | 6 +-
tools/perf/tests/dlfilter-test.c | 2 +-
tools/perf/tests/dwarf-unwind.c | 2 +-
tools/perf/tests/event_update.c | 9 +-
tools/perf/tests/stat.c | 6 +-
tools/perf/tests/thread-map.c | 2 +-
tools/perf/util/Build | 1 +
tools/perf/util/arm-spe.c | 55 +----
tools/perf/util/auxtrace.c | 12 +-
tools/perf/util/auxtrace.h | 20 +-
tools/perf/util/bpf-event.c | 4 +-
tools/perf/util/build-id.c | 34 +--
tools/perf/util/build-id.h | 8 +-
tools/perf/util/cs-etm.c | 39 +--
tools/perf/util/data-convert-bt.c | 34 ++-
tools/perf/util/data-convert-json.c | 47 ++--
tools/perf/util/event.c | 54 ++--
tools/perf/util/event.h | 38 +--
tools/perf/util/header.c | 6 +-
tools/perf/util/header.h | 4 +-
tools/perf/util/hisi-ptt.c | 6 +-
tools/perf/util/intel-bts.c | 37 +--
tools/perf/util/intel-pt.c | 30 +--
tools/perf/util/jitdump.c | 4 +-
tools/perf/util/s390-cpumsf.c | 11 +-
tools/perf/util/session.c | 366 +++-------------------------
tools/perf/util/session.h | 9 +-
tools/perf/util/synthetic-events.c | 80 +++---
tools/perf/util/synthetic-events.h | 70 +++---
tools/perf/util/tool.c | 294 ++++++++++++++++++++++
tools/perf/util/tool.h | 18 +-
tools/perf/util/tsc.c | 2 +-
53 files changed, 977 insertions(+), 1102 deletions(-)
create mode 100644 tools/perf/util/tool.c
--
2.45.2.1089.g2a221341d9-goog
^ permalink raw reply [flat|nested] 40+ messages in thread* [PATCH v6 01/27] perf auxtrace: Remove dummy tools 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers @ 2024-07-18 0:59 ` Ian Rogers 2024-07-18 0:59 ` [PATCH v6 02/27] perf s390-cpumsf: Remove unused struct Ian Rogers ` (26 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 0:59 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Cc: Leo Yan Add perf_session__deliver_synth_attr_event that synthesizes a perf_record_header_attr event with one id. Remove use of perf_event__synthesize_attr that necessitates the use of the dummy tool in order to pass the session. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> Tested-by: Adrian Hunter <adrian.hunter@intel.com> Tested-by: Leo Yan <leo.yan@arm.com> --- tools/perf/util/arm-spe.c | 49 ++++++++----------------------------- tools/perf/util/cs-etm.c | 33 ++----------------------- tools/perf/util/intel-bts.c | 31 +---------------------- tools/perf/util/intel-pt.c | 24 +----------------- tools/perf/util/session.c | 24 ++++++++++++++++++ tools/perf/util/session.h | 3 +++ 6 files changed, 41 insertions(+), 123 deletions(-) diff --git a/tools/perf/util/arm-spe.c b/tools/perf/util/arm-spe.c index afbd5869f6bf..fa40f3cb6266 100644 --- a/tools/perf/util/arm-spe.c +++ b/tools/perf/util/arm-spe.c @@ -1073,35 +1073,6 @@ static void arm_spe_print_info(__u64 *arr) fprintf(stdout, arm_spe_info_fmts[ARM_SPE_PMU_TYPE], arr[ARM_SPE_PMU_TYPE]); } -struct arm_spe_synth { - struct perf_tool dummy_tool; - struct perf_session *session; -}; - -static int arm_spe_event_synth(struct perf_tool *tool, - union perf_event *event, - struct perf_sample *sample __maybe_unused, - struct machine *machine __maybe_unused) -{ - struct arm_spe_synth *arm_spe_synth = - container_of(tool, struct arm_spe_synth, dummy_tool); - - return perf_session__deliver_synth_event(arm_spe_synth->session, - event, NULL); -} - -static int arm_spe_synth_event(struct perf_session *session, - struct perf_event_attr *attr, u64 id) -{ - struct arm_spe_synth arm_spe_synth; - - 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, - &id, arm_spe_event_synth); -} - static void arm_spe_set_event_name(struct evlist *evlist, u64 id, const char *name) { @@ -1172,7 +1143,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) spe->sample_flc = true; /* Level 1 data cache miss */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->l1d_miss_id = id; @@ -1180,7 +1151,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) id += 1; /* Level 1 data cache access */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->l1d_access_id = id; @@ -1192,7 +1163,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) spe->sample_llc = true; /* Last level cache miss */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->llc_miss_id = id; @@ -1200,7 +1171,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) id += 1; /* Last level cache access */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->llc_access_id = id; @@ -1212,7 +1183,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) spe->sample_tlb = true; /* TLB miss */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->tlb_miss_id = id; @@ -1220,7 +1191,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) id += 1; /* TLB access */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->tlb_access_id = id; @@ -1232,7 +1203,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) spe->sample_branch = true; /* Branch miss */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->branch_miss_id = id; @@ -1244,7 +1215,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) spe->sample_remote_access = true; /* Remote access */ - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->remote_access_id = id; @@ -1255,7 +1226,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) if (spe->synth_opts.mem) { spe->sample_memory = true; - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->memory_id = id; @@ -1276,7 +1247,7 @@ arm_spe_synth_events(struct arm_spe *spe, struct perf_session *session) attr.config = PERF_COUNT_HW_INSTRUCTIONS; attr.sample_period = spe->synth_opts.period; spe->instructions_sample_period = attr.sample_period; - err = arm_spe_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; spe->instructions_id = id; diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c index 5e9fbcfad7d4..7fd976ea954b 100644 --- a/tools/perf/util/cs-etm.c +++ b/tools/perf/util/cs-etm.c @@ -1595,35 +1595,6 @@ static int cs_etm__synth_branch_sample(struct cs_etm_queue *etmq, return ret; } -struct cs_etm_synth { - struct perf_tool dummy_tool; - struct perf_session *session; -}; - -static int cs_etm__event_synth(struct perf_tool *tool, - union perf_event *event, - struct perf_sample *sample __maybe_unused, - struct machine *machine __maybe_unused) -{ - struct cs_etm_synth *cs_etm_synth = - container_of(tool, struct cs_etm_synth, dummy_tool); - - return perf_session__deliver_synth_event(cs_etm_synth->session, - event, NULL); -} - -static int cs_etm__synth_event(struct perf_session *session, - struct perf_event_attr *attr, u64 id) -{ - struct cs_etm_synth cs_etm_synth; - - 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, - &id, cs_etm__event_synth); -} - static int cs_etm__synth_events(struct cs_etm_auxtrace *etm, struct perf_session *session) { @@ -1675,7 +1646,7 @@ static int cs_etm__synth_events(struct cs_etm_auxtrace *etm, attr.config = PERF_COUNT_HW_BRANCH_INSTRUCTIONS; attr.sample_period = 1; attr.sample_type |= PERF_SAMPLE_ADDR; - err = cs_etm__synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; etm->branches_sample_type = attr.sample_type; @@ -1698,7 +1669,7 @@ static int cs_etm__synth_events(struct cs_etm_auxtrace *etm, attr.config = PERF_COUNT_HW_INSTRUCTIONS; attr.sample_period = etm->synth_opts.period; etm->instructions_sample_period = attr.sample_period; - err = cs_etm__synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) return err; etm->instructions_sample_type = attr.sample_type; diff --git a/tools/perf/util/intel-bts.c b/tools/perf/util/intel-bts.c index ec1b3bd9f530..160d346cb450 100644 --- a/tools/perf/util/intel-bts.c +++ b/tools/perf/util/intel-bts.c @@ -737,35 +737,6 @@ static bool intel_bts_evsel_is_auxtrace(struct perf_session *session, return evsel->core.attr.type == bts->pmu_type; } -struct intel_bts_synth { - struct perf_tool dummy_tool; - struct perf_session *session; -}; - -static int intel_bts_event_synth(struct perf_tool *tool, - union perf_event *event, - struct perf_sample *sample __maybe_unused, - struct machine *machine __maybe_unused) -{ - struct intel_bts_synth *intel_bts_synth = - container_of(tool, struct intel_bts_synth, dummy_tool); - - return perf_session__deliver_synth_event(intel_bts_synth->session, - event, NULL); -} - -static int intel_bts_synth_event(struct perf_session *session, - struct perf_event_attr *attr, u64 id) -{ - struct intel_bts_synth intel_bts_synth; - - 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, - &id, intel_bts_event_synth); -} - static int intel_bts_synth_events(struct intel_bts *bts, struct perf_session *session) { @@ -814,7 +785,7 @@ static int intel_bts_synth_events(struct intel_bts *bts, attr.sample_type |= PERF_SAMPLE_ADDR; pr_debug("Synthesizing 'branches' event with id %" PRIu64 " sample type %#" PRIx64 "\n", id, (u64)attr.sample_type); - err = intel_bts_synth_event(session, &attr, id); + err = perf_session__deliver_synth_attr_event(session, &attr, id); if (err) { pr_err("%s: failed to synthesize 'branches' event type\n", __func__); diff --git a/tools/perf/util/intel-pt.c b/tools/perf/util/intel-pt.c index d6d7b7512505..1fc03bfd14c3 100644 --- a/tools/perf/util/intel-pt.c +++ b/tools/perf/util/intel-pt.c @@ -3659,37 +3659,15 @@ static int intel_pt_queue_data(struct perf_session *session, data_offset, timestamp); } -struct intel_pt_synth { - struct perf_tool dummy_tool; - struct perf_session *session; -}; - -static int intel_pt_event_synth(struct perf_tool *tool, - union perf_event *event, - struct perf_sample *sample __maybe_unused, - struct machine *machine __maybe_unused) -{ - struct intel_pt_synth *intel_pt_synth = - container_of(tool, struct intel_pt_synth, dummy_tool); - - return perf_session__deliver_synth_event(intel_pt_synth->session, event, - NULL); -} - static int intel_pt_synth_event(struct perf_session *session, const char *name, struct perf_event_attr *attr, u64 id) { - struct intel_pt_synth intel_pt_synth; int err; pr_debug("Synthesizing '%s' event with id %" PRIu64 " sample type %#" PRIx64 "\n", name, id, (u64)attr->sample_type); - 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, - &id, intel_pt_event_synth); + err = perf_session__deliver_synth_attr_event(session, attr, id); if (err) pr_err("%s: failed to synthesize '%s' event type\n", __func__, name); diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 5596bed1b8c8..18a8f4a84045 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1764,6 +1764,30 @@ int perf_session__deliver_synth_event(struct perf_session *session, return machines__deliver_event(&session->machines, evlist, event, sample, tool, 0, NULL); } +int perf_session__deliver_synth_attr_event(struct perf_session *session, + const struct perf_event_attr *attr, + u64 id) +{ + union { + struct { + struct perf_record_header_attr attr; + u64 ids[1]; + } attr_id; + union perf_event ev; + } ev = { + .attr_id.attr.header.type = PERF_RECORD_HEADER_ATTR, + .attr_id.attr.header.size = sizeof(ev.attr_id), + .attr_id.ids[0] = id, + }; + + if (attr->size != sizeof(ev.attr_id.attr.attr)) { + pr_debug("Unexpected perf_event_attr size\n"); + return -EINVAL; + } + ev.attr_id.attr.attr = *attr; + return perf_session__deliver_synth_event(session, &ev.ev, NULL); +} + static void event_swap(union perf_event *event, bool sample_id_all) { perf_event__swap_op swap; diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 4c29dc86956f..cc28976bb00e 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -154,6 +154,9 @@ extern volatile int session_done; int perf_session__deliver_synth_event(struct perf_session *session, union perf_event *event, struct perf_sample *sample); +int perf_session__deliver_synth_attr_event(struct perf_session *session, + const struct perf_event_attr *attr, + u64 id); int perf_session__dsos_hit_all(struct perf_session *session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 02/27] perf s390-cpumsf: Remove unused struct 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers 2024-07-18 0:59 ` [PATCH v6 01/27] perf auxtrace: Remove dummy tools Ian Rogers @ 2024-07-18 0:59 ` Ian Rogers 2024-07-18 0:59 ` [PATCH v6 03/27] perf tool: Constify tool pointers Ian Rogers ` (25 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 0:59 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel struct s390_cpumsf_synth was likely cargo culted from other auxtrace examples. It has no users, so remove. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/util/s390-cpumsf.c | 5 ----- 1 file changed, 5 deletions(-) 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; -}; - static int s390_cpumsf_process_auxtrace_event(struct perf_session *session, union perf_event *event __maybe_unused, -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 03/27] perf tool: Constify tool pointers 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers 2024-07-18 0:59 ` [PATCH v6 01/27] perf auxtrace: Remove dummy tools Ian Rogers 2024-07-18 0:59 ` [PATCH v6 02/27] perf s390-cpumsf: Remove unused struct Ian Rogers @ 2024-07-18 0:59 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 04/27] perf tool: Move fill defaults into tool.c Ian Rogers ` (24 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 0:59 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Cc: Leo Yan The tool pointer (to a struct largely of function pointers) is passed around but is unchanged except at initialization. Change parameter and variable types to be const to lower the possibilities of what could happen with a tool. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> Tested-by: Adrian Hunter <adrian.hunter@intel.com> Tested-by: Leo Yan <leo.yan@arm.com> --- tools/perf/arch/x86/util/event.c | 4 +- tools/perf/bench/synthesize.c | 2 +- tools/perf/builtin-annotate.c | 2 +- tools/perf/builtin-c2c.c | 2 +- tools/perf/builtin-diff.c | 2 +- tools/perf/builtin-inject.c | 70 ++++++++++++------------- tools/perf/builtin-kmem.c | 3 +- tools/perf/builtin-kvm.c | 4 +- tools/perf/builtin-kwork.c | 22 ++++---- tools/perf/builtin-lock.c | 4 +- tools/perf/builtin-mem.c | 4 +- tools/perf/builtin-record.c | 14 ++--- tools/perf/builtin-report.c | 12 ++--- tools/perf/builtin-sched.c | 34 ++++++------ tools/perf/builtin-script.c | 41 ++++++++------- tools/perf/builtin-stat.c | 8 +-- tools/perf/builtin-timechart.c | 8 +-- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 4 +- tools/perf/tests/cpumap.c | 6 +-- tools/perf/tests/dlfilter-test.c | 2 +- tools/perf/tests/dwarf-unwind.c | 2 +- tools/perf/tests/event_update.c | 8 +-- tools/perf/tests/stat.c | 6 +-- tools/perf/tests/thread-map.c | 2 +- tools/perf/util/arm-spe.c | 6 +-- tools/perf/util/auxtrace.c | 12 ++--- tools/perf/util/auxtrace.h | 20 ++++---- tools/perf/util/bpf-event.c | 4 +- tools/perf/util/build-id.c | 4 +- tools/perf/util/build-id.h | 4 +- tools/perf/util/cs-etm.c | 6 +-- tools/perf/util/data-convert-bt.c | 4 +- tools/perf/util/data-convert-json.c | 4 +- tools/perf/util/event.c | 34 ++++++------ tools/perf/util/event.h | 34 ++++++------ tools/perf/util/header.c | 6 +-- tools/perf/util/header.h | 4 +- tools/perf/util/hisi-ptt.c | 6 +-- tools/perf/util/intel-bts.c | 6 +-- tools/perf/util/intel-pt.c | 6 +-- tools/perf/util/jitdump.c | 4 +- tools/perf/util/s390-cpumsf.c | 6 +-- tools/perf/util/session.c | 30 +++++------ tools/perf/util/session.h | 2 +- tools/perf/util/synthetic-events.c | 80 ++++++++++++++--------------- tools/perf/util/synthetic-events.h | 70 ++++++++++++------------- tools/perf/util/tool.h | 8 +-- tools/perf/util/tsc.c | 2 +- 49 files changed, 314 insertions(+), 316 deletions(-) diff --git a/tools/perf/arch/x86/util/event.c b/tools/perf/arch/x86/util/event.c index e65b7dbe27fb..a0400707180c 100644 --- a/tools/perf/arch/x86/util/event.c +++ b/tools/perf/arch/x86/util/event.c @@ -15,7 +15,7 @@ #if defined(__x86_64__) struct perf_event__synthesize_extra_kmaps_cb_args { - struct perf_tool *tool; + const struct perf_tool *tool; perf_event__handler_t process; struct machine *machine; union perf_event *event; @@ -65,7 +65,7 @@ static int perf_event__synthesize_extra_kmaps_cb(struct map *map, void *data) return 0; } -int perf_event__synthesize_extra_kmaps(struct perf_tool *tool, +int perf_event__synthesize_extra_kmaps(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine) { diff --git a/tools/perf/bench/synthesize.c b/tools/perf/bench/synthesize.c index 7401ebbac100..9b333276cbdb 100644 --- a/tools/perf/bench/synthesize.c +++ b/tools/perf/bench/synthesize.c @@ -49,7 +49,7 @@ static const char *const bench_usage[] = { static atomic_t event_count; -static int process_synthesized_event(struct perf_tool *tool __maybe_unused, +static int process_synthesized_event(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index b10b7f005658..c0685d2c8de1 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -279,7 +279,7 @@ static int evsel__add_sample(struct evsel *evsel, struct perf_sample *sample, return ret; } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c index c157bd31f2e5..88c131d05186 100644 --- a/tools/perf/builtin-c2c.c +++ b/tools/perf/builtin-c2c.c @@ -273,7 +273,7 @@ static void compute_stats(struct c2c_hist_entry *c2c_he, update_stats(&cstats->load, weight); } -static int process_sample_event(struct perf_tool *tool __maybe_unused, +static int process_sample_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c index 57d300d8e570..4c0567882a7a 100644 --- a/tools/perf/builtin-diff.c +++ b/tools/perf/builtin-diff.c @@ -388,7 +388,7 @@ struct hist_entry_ops block_hist_ops = { .free = block_hist_free, }; -static int diff__process_sample_event(struct perf_tool *tool, +static int diff__process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index a212678d47be..be74e2bf1f00 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -134,7 +134,7 @@ struct event_entry { union perf_event event[]; }; -static int dso__inject_build_id(struct dso *dso, struct perf_tool *tool, +static int dso__inject_build_id(struct dso *dso, const struct perf_tool *tool, struct machine *machine, u8 cpumode, u32 flags); static int output_bytes(struct perf_inject *inject, void *buf, size_t sz) @@ -149,7 +149,7 @@ static int output_bytes(struct perf_inject *inject, void *buf, size_t sz) return 0; } -static int perf_event__repipe_synth(struct perf_tool *tool, +static int perf_event__repipe_synth(const struct perf_tool *tool, union perf_event *event) { struct perf_inject *inject = container_of(tool, struct perf_inject, @@ -158,7 +158,7 @@ static int perf_event__repipe_synth(struct perf_tool *tool, return output_bytes(inject, event, event->header.size); } -static int perf_event__repipe_oe_synth(struct perf_tool *tool, +static int perf_event__repipe_oe_synth(const struct perf_tool *tool, union perf_event *event, struct ordered_events *oe __maybe_unused) { @@ -166,7 +166,7 @@ static int perf_event__repipe_oe_synth(struct perf_tool *tool, } #ifdef HAVE_JITDUMP -static int perf_event__drop_oe(struct perf_tool *tool __maybe_unused, +static int perf_event__drop_oe(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct ordered_events *oe __maybe_unused) { @@ -188,7 +188,7 @@ static int perf_event__repipe_op4_synth(struct perf_session *session, return perf_event__repipe_synth(session->tool, event); } -static int perf_event__repipe_attr(struct perf_tool *tool, +static int perf_event__repipe_attr(const struct perf_tool *tool, union perf_event *event, struct evlist **pevlist) { @@ -206,7 +206,7 @@ static int perf_event__repipe_attr(struct perf_tool *tool, return perf_event__repipe_synth(tool, event); } -static int perf_event__repipe_event_update(struct perf_tool *tool, +static int perf_event__repipe_event_update(const struct perf_tool *tool, union perf_event *event, struct evlist **pevlist __maybe_unused) { @@ -237,7 +237,7 @@ static int copy_bytes(struct perf_inject *inject, struct perf_data *data, off_t static s64 perf_event__repipe_auxtrace(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct perf_inject *inject = container_of(tool, struct perf_inject, tool); int ret; @@ -284,7 +284,7 @@ perf_event__repipe_auxtrace(struct perf_session *session __maybe_unused, #endif -static int perf_event__repipe(struct perf_tool *tool, +static int perf_event__repipe(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -292,7 +292,7 @@ static int perf_event__repipe(struct perf_tool *tool, return perf_event__repipe_synth(tool, event); } -static int perf_event__drop(struct perf_tool *tool __maybe_unused, +static int perf_event__drop(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -300,7 +300,7 @@ static int perf_event__drop(struct perf_tool *tool __maybe_unused, return 0; } -static int perf_event__drop_aux(struct perf_tool *tool, +static int perf_event__drop_aux(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct perf_sample *sample, struct machine *machine __maybe_unused) @@ -341,13 +341,13 @@ perf_inject__cut_auxtrace_sample(struct perf_inject *inject, return ev; } -typedef int (*inject_handler)(struct perf_tool *tool, +typedef int (*inject_handler)(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine); -static int perf_event__repipe_sample(struct perf_tool *tool, +static int perf_event__repipe_sample(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, @@ -372,7 +372,7 @@ static int perf_event__repipe_sample(struct perf_tool *tool, return perf_event__repipe_synth(tool, event); } -static int perf_event__repipe_mmap(struct perf_tool *tool, +static int perf_event__repipe_mmap(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -386,7 +386,7 @@ static int perf_event__repipe_mmap(struct perf_tool *tool, } #ifdef HAVE_JITDUMP -static int perf_event__jit_repipe_mmap(struct perf_tool *tool, +static int perf_event__jit_repipe_mmap(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -455,7 +455,7 @@ static struct dso *findnew_dso(int pid, int tid, const char *filename, return dso; } -static int perf_event__repipe_buildid_mmap(struct perf_tool *tool, +static int perf_event__repipe_buildid_mmap(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -474,7 +474,7 @@ static int perf_event__repipe_buildid_mmap(struct perf_tool *tool, return perf_event__repipe(tool, event, sample, machine); } -static int perf_event__repipe_mmap2(struct perf_tool *tool, +static int perf_event__repipe_mmap2(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -500,7 +500,7 @@ static int perf_event__repipe_mmap2(struct perf_tool *tool, } #ifdef HAVE_JITDUMP -static int perf_event__jit_repipe_mmap2(struct perf_tool *tool, +static int perf_event__jit_repipe_mmap2(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -524,7 +524,7 @@ static int perf_event__jit_repipe_mmap2(struct perf_tool *tool, } #endif -static int perf_event__repipe_buildid_mmap2(struct perf_tool *tool, +static int perf_event__repipe_buildid_mmap2(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -565,7 +565,7 @@ static int perf_event__repipe_buildid_mmap2(struct perf_tool *tool, return 0; } -static int perf_event__repipe_fork(struct perf_tool *tool, +static int perf_event__repipe_fork(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -578,7 +578,7 @@ static int perf_event__repipe_fork(struct perf_tool *tool, return err; } -static int perf_event__repipe_comm(struct perf_tool *tool, +static int perf_event__repipe_comm(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -591,7 +591,7 @@ static int perf_event__repipe_comm(struct perf_tool *tool, return err; } -static int perf_event__repipe_namespaces(struct perf_tool *tool, +static int perf_event__repipe_namespaces(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -603,7 +603,7 @@ static int perf_event__repipe_namespaces(struct perf_tool *tool, return err; } -static int perf_event__repipe_exit(struct perf_tool *tool, +static int perf_event__repipe_exit(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -712,7 +712,7 @@ static bool perf_inject__lookup_known_build_id(struct perf_inject *inject, return false; } -static int dso__inject_build_id(struct dso *dso, struct perf_tool *tool, +static int dso__inject_build_id(struct dso *dso, const struct perf_tool *tool, struct machine *machine, u8 cpumode, u32 flags) { struct perf_inject *inject = container_of(tool, struct perf_inject, @@ -743,7 +743,7 @@ static int dso__inject_build_id(struct dso *dso, struct perf_tool *tool, return 0; } -int perf_event__inject_buildid(struct perf_tool *tool, union perf_event *event, +int perf_event__inject_buildid(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel __maybe_unused, struct machine *machine) @@ -776,7 +776,7 @@ int perf_event__inject_buildid(struct perf_tool *tool, union perf_event *event, return 0; } -static int perf_inject__sched_process_exit(struct perf_tool *tool, +static int perf_inject__sched_process_exit(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct perf_sample *sample, struct evsel *evsel __maybe_unused, @@ -796,7 +796,7 @@ static int perf_inject__sched_process_exit(struct perf_tool *tool, return 0; } -static int perf_inject__sched_switch(struct perf_tool *tool, +static int perf_inject__sched_switch(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, @@ -821,7 +821,7 @@ static int perf_inject__sched_switch(struct perf_tool *tool, } #ifdef HAVE_LIBTRACEEVENT -static int perf_inject__sched_stat(struct perf_tool *tool, +static int perf_inject__sched_stat(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct perf_sample *sample, struct evsel *evsel, @@ -866,7 +866,7 @@ static int guest_session__output_bytes(struct guest_session *gs, void *buf, size return ret < 0 ? ret : 0; } -static int guest_session__repipe(struct perf_tool *tool, +static int guest_session__repipe(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -1032,7 +1032,7 @@ static struct guest_id *guest_session__lookup_id(struct guest_session *gs, u64 i return NULL; } -static int process_attr(struct perf_tool *tool, union perf_event *event, +static int process_attr(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) { @@ -1160,7 +1160,7 @@ static u64 evlist__first_id(struct evlist *evlist) return 0; } -static int process_build_id(struct perf_tool *tool, +static int process_build_id(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -1210,7 +1210,7 @@ static int guest_session__add_build_ids(struct guest_session *gs) gs); } -static int guest_session__ksymbol_event(struct perf_tool *tool, +static int guest_session__ksymbol_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -1574,7 +1574,7 @@ static int guest_session__flush_events(struct guest_session *gs) return guest_session__inject_events(gs, -1); } -static int host__repipe(struct perf_tool *tool, +static int host__repipe(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -1647,7 +1647,7 @@ static int host__finished_init(struct perf_session *session, union perf_event *e * guest events up to the same time. Finally write out the FINISHED_ROUND event * itself. */ -static int host__finished_round(struct perf_tool *tool, +static int host__finished_round(const struct perf_tool *tool, union perf_event *event, struct ordered_events *oe) { @@ -1665,7 +1665,7 @@ static int host__finished_round(struct perf_tool *tool, return perf_event__repipe_oe_synth(tool, event, oe); } -static int host__context_switch(struct perf_tool *tool, +static int host__context_switch(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -1719,7 +1719,7 @@ static int evsel__check_stype(struct evsel *evsel, u64 sample_type, const char * return 0; } -static int drop_sample(struct perf_tool *tool __maybe_unused, +static int drop_sample(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, struct evsel *evsel __maybe_unused, diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 6fd95be5032b..859ff018eace 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -955,7 +955,7 @@ static bool perf_kmem__skip_sample(struct perf_sample *sample) typedef int (*tracepoint_handler)(struct evsel *evsel, struct perf_sample *sample); -static int process_sample_event(struct perf_tool *tool __maybe_unused, +static int process_sample_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, @@ -2061,4 +2061,3 @@ int cmd_kmem(int argc, const char **argv) return ret; } - diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 71165036e4ca..a3b903cf4311 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -1166,7 +1166,7 @@ static void print_result(struct perf_kvm_stat *kvm) } #if defined(HAVE_TIMERFD_SUPPORT) && defined(HAVE_LIBTRACEEVENT) -static int process_lost_event(struct perf_tool *tool, +static int process_lost_event(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -1187,7 +1187,7 @@ static bool skip_sample(struct perf_kvm_stat *kvm, return false; } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c index 56e3f3a5e03a..8ffaa80a2d1d 100644 --- a/tools/perf/builtin-kwork.c +++ b/tools/perf/builtin-kwork.c @@ -958,7 +958,7 @@ static int top_sched_switch_event(struct perf_kwork *kwork, } static struct kwork_class kwork_irq; -static int process_irq_handler_entry_event(struct perf_tool *tool, +static int process_irq_handler_entry_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -971,7 +971,7 @@ static int process_irq_handler_entry_event(struct perf_tool *tool, return 0; } -static int process_irq_handler_exit_event(struct perf_tool *tool, +static int process_irq_handler_exit_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1037,7 +1037,7 @@ static struct kwork_class kwork_irq = { }; static struct kwork_class kwork_softirq; -static int process_softirq_raise_event(struct perf_tool *tool, +static int process_softirq_raise_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1051,7 +1051,7 @@ static int process_softirq_raise_event(struct perf_tool *tool, return 0; } -static int process_softirq_entry_event(struct perf_tool *tool, +static int process_softirq_entry_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1065,7 +1065,7 @@ static int process_softirq_entry_event(struct perf_tool *tool, return 0; } -static int process_softirq_exit_event(struct perf_tool *tool, +static int process_softirq_exit_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1167,7 +1167,7 @@ static struct kwork_class kwork_softirq = { }; static struct kwork_class kwork_workqueue; -static int process_workqueue_activate_work_event(struct perf_tool *tool, +static int process_workqueue_activate_work_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1181,7 +1181,7 @@ static int process_workqueue_activate_work_event(struct perf_tool *tool, return 0; } -static int process_workqueue_execute_start_event(struct perf_tool *tool, +static int process_workqueue_execute_start_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1195,7 +1195,7 @@ static int process_workqueue_execute_start_event(struct perf_tool *tool, return 0; } -static int process_workqueue_execute_end_event(struct perf_tool *tool, +static int process_workqueue_execute_end_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1266,7 +1266,7 @@ static struct kwork_class kwork_workqueue = { }; static struct kwork_class kwork_sched; -static int process_sched_switch_event(struct perf_tool *tool, +static int process_sched_switch_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1945,12 +1945,12 @@ static int perf_kwork__report(struct perf_kwork *kwork) return 0; } -typedef int (*tracepoint_handler)(struct perf_tool *tool, +typedef int (*tracepoint_handler)(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine); -static int perf_kwork__process_tracepoint_sample(struct perf_tool *tool, +static int perf_kwork__process_tracepoint_sample(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c index 0253184b3b58..6efa9d646637 100644 --- a/tools/perf/builtin-lock.c +++ b/tools/perf/builtin-lock.c @@ -1501,7 +1501,7 @@ static const struct evsel_str_handler contention_tracepoints[] = { { "lock:contention_end", evsel__process_contention_end, }, }; -static int process_event_update(struct perf_tool *tool, +static int process_event_update(const struct perf_tool *tool, union perf_event *event, struct evlist **pevlist) { @@ -1520,7 +1520,7 @@ static int process_event_update(struct perf_tool *tool, typedef int (*tracepoint_handler)(struct evsel *evsel, struct perf_sample *sample); -static int process_sample_event(struct perf_tool *tool __maybe_unused, +static int process_sample_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-mem.c b/tools/perf/builtin-mem.c index 863fcd735dae..c2038a32543c 100644 --- a/tools/perf/builtin-mem.c +++ b/tools/perf/builtin-mem.c @@ -188,7 +188,7 @@ static int __cmd_record(int argc, const char **argv, struct perf_mem *mem) } static int -dump_raw_samples(struct perf_tool *tool, +dump_raw_samples(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -262,7 +262,7 @@ dump_raw_samples(struct perf_tool *tool, return 0; } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel __maybe_unused, diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index a94516e8c522..f27f725913d7 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -608,7 +608,7 @@ static int record__comp_enabled(struct record *rec) return rec->opts.comp_level > 0; } -static int process_synthesized_event(struct perf_tool *tool, +static int process_synthesized_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -619,7 +619,7 @@ static int process_synthesized_event(struct perf_tool *tool, static struct mutex synth_lock; -static int process_locked_synthesized_event(struct perf_tool *tool, +static int process_locked_synthesized_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -704,7 +704,7 @@ static void record__sig_exit(void) #ifdef HAVE_AUXTRACE_SUPPORT -static int record__process_auxtrace(struct perf_tool *tool, +static int record__process_auxtrace(const struct perf_tool *tool, struct mmap *map, union perf_event *event, void *data1, size_t len1, void *data2, size_t len2) @@ -1416,7 +1416,7 @@ static void set_timestamp_boundary(struct record *rec, u64 sample_time) rec->evlist->last_sample_time = sample_time; } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, @@ -3243,7 +3243,7 @@ static const char * const __record_usage[] = { }; const char * const *record_usage = __record_usage; -static int build_id__process_mmap(struct perf_tool *tool, union perf_event *event, +static int build_id__process_mmap(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) { /* @@ -3255,7 +3255,7 @@ static int build_id__process_mmap(struct perf_tool *tool, union perf_event *even return perf_event__process_mmap(tool, event, sample, machine); } -static int build_id__process_mmap2(struct perf_tool *tool, union perf_event *event, +static int build_id__process_mmap2(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) { /* @@ -3268,7 +3268,7 @@ static int build_id__process_mmap2(struct perf_tool *tool, union perf_event *eve return perf_event__process_mmap2(tool, event, sample, machine); } -static int process_timestamp_boundary(struct perf_tool *tool, +static int process_timestamp_boundary(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct perf_sample *sample, struct machine *machine __maybe_unused) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 6edc0d4ce6fb..c402d9b3dca6 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -263,7 +263,7 @@ static int process_feature_event(struct perf_session *session, return 0; } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, @@ -339,7 +339,7 @@ static int process_sample_event(struct perf_tool *tool, return ret; } -static int process_read_event(struct perf_tool *tool, +static int process_read_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct evsel *evsel, @@ -765,7 +765,7 @@ static void report__output_resort(struct report *rep) ui_progress__finish(); } -static int count_sample_event(struct perf_tool *tool __maybe_unused, +static int count_sample_event(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, struct evsel *evsel, @@ -777,7 +777,7 @@ static int count_sample_event(struct perf_tool *tool __maybe_unused, return 0; } -static int count_lost_samples_event(struct perf_tool *tool, +static int count_lost_samples_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine __maybe_unused) @@ -793,7 +793,7 @@ static int count_lost_samples_event(struct perf_tool *tool, return 0; } -static int process_attr(struct perf_tool *tool __maybe_unused, +static int process_attr(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct evlist **pevlist); @@ -1233,7 +1233,7 @@ parse_percent_limit(const struct option *opt, const char *str, return 0; } -static int process_attr(struct perf_tool *tool __maybe_unused, +static int process_attr(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct evlist **pevlist) { diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 8750b5f2d49b..2c60bd3a8149 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1489,7 +1489,7 @@ static void perf_sched__sort_lat(struct perf_sched *sched) } } -static int process_sched_wakeup_event(struct perf_tool *tool, +static int process_sched_wakeup_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1502,7 +1502,7 @@ static int process_sched_wakeup_event(struct perf_tool *tool, return 0; } -static int process_sched_wakeup_ignore(struct perf_tool *tool __maybe_unused, +static int process_sched_wakeup_ignore(const struct perf_tool *tool __maybe_unused, struct evsel *evsel __maybe_unused, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -1770,7 +1770,7 @@ static int map_switch_event(struct perf_sched *sched, struct evsel *evsel, return 0; } -static int process_sched_switch_event(struct perf_tool *tool, +static int process_sched_switch_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1796,7 +1796,7 @@ static int process_sched_switch_event(struct perf_tool *tool, return err; } -static int process_sched_runtime_event(struct perf_tool *tool, +static int process_sched_runtime_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1809,7 +1809,7 @@ static int process_sched_runtime_event(struct perf_tool *tool, return 0; } -static int perf_sched__process_fork_event(struct perf_tool *tool, +static int perf_sched__process_fork_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -1826,7 +1826,7 @@ static int perf_sched__process_fork_event(struct perf_tool *tool, return 0; } -static int process_sched_migrate_task_event(struct perf_tool *tool, +static int process_sched_migrate_task_event(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine) @@ -1839,12 +1839,12 @@ static int process_sched_migrate_task_event(struct perf_tool *tool, return 0; } -typedef int (*tracepoint_handler)(struct perf_tool *tool, +typedef int (*tracepoint_handler)(const struct perf_tool *tool, struct evsel *evsel, struct perf_sample *sample, struct machine *machine); -static int perf_sched__process_tracepoint_sample(struct perf_tool *tool __maybe_unused, +static int perf_sched__process_tracepoint_sample(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample, struct evsel *evsel, @@ -1860,7 +1860,7 @@ static int perf_sched__process_tracepoint_sample(struct perf_tool *tool __maybe_ return err; } -static int perf_sched__process_comm(struct perf_tool *tool __maybe_unused, +static int perf_sched__process_comm(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2506,7 +2506,7 @@ static void timehist_print_wakeup_event(struct perf_sched *sched, printf("\n"); } -static int timehist_sched_wakeup_ignore(struct perf_tool *tool __maybe_unused, +static int timehist_sched_wakeup_ignore(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct evsel *evsel __maybe_unused, struct perf_sample *sample __maybe_unused, @@ -2515,7 +2515,7 @@ static int timehist_sched_wakeup_ignore(struct perf_tool *tool __maybe_unused, return 0; } -static int timehist_sched_wakeup_event(struct perf_tool *tool, +static int timehist_sched_wakeup_event(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct evsel *evsel, struct perf_sample *sample, @@ -2599,7 +2599,7 @@ static void timehist_print_migration_event(struct perf_sched *sched, printf("\n"); } -static int timehist_migrate_task_event(struct perf_tool *tool, +static int timehist_migrate_task_event(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct evsel *evsel, struct perf_sample *sample, @@ -2627,7 +2627,7 @@ static int timehist_migrate_task_event(struct perf_tool *tool, return 0; } -static int timehist_sched_change_event(struct perf_tool *tool, +static int timehist_sched_change_event(const struct perf_tool *tool, union perf_event *event, struct evsel *evsel, struct perf_sample *sample, @@ -2758,7 +2758,7 @@ static int timehist_sched_change_event(struct perf_tool *tool, return rc; } -static int timehist_sched_switch_event(struct perf_tool *tool, +static int timehist_sched_switch_event(const struct perf_tool *tool, union perf_event *event, struct evsel *evsel, struct perf_sample *sample, @@ -2767,7 +2767,7 @@ static int timehist_sched_switch_event(struct perf_tool *tool, return timehist_sched_change_event(tool, event, evsel, sample, machine); } -static int process_lost(struct perf_tool *tool __maybe_unused, +static int process_lost(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine __maybe_unused) @@ -3010,13 +3010,13 @@ static void timehist_print_summary(struct perf_sched *sched, printf(" (x %d)\n", sched->max_cpu.cpu); } -typedef int (*sched_handler)(struct perf_tool *tool, +typedef int (*sched_handler)(const struct perf_tool *tool, union perf_event *event, struct evsel *evsel, struct perf_sample *sample, struct machine *machine); -static int perf_timehist__process_sample(struct perf_tool *tool, +static int perf_timehist__process_sample(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index c16224b1fef3..394bce9f5338 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -2399,7 +2399,7 @@ static bool filter_cpu(struct perf_sample *sample) return false; } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, @@ -2486,7 +2486,7 @@ static int process_sample_event(struct perf_tool *tool, // Used when scr->per_event_dump is not set static struct evsel_script es_stdout; -static int process_attr(struct perf_tool *tool, union perf_event *event, +static int process_attr(const struct perf_tool *tool, union perf_event *event, struct evlist **pevlist) { struct perf_script *scr = container_of(tool, struct perf_script, tool); @@ -2552,7 +2552,7 @@ static int process_attr(struct perf_tool *tool, union perf_event *event, return 0; } -static int print_event_with_time(struct perf_tool *tool, +static int print_event_with_time(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine, @@ -2588,14 +2588,14 @@ static int print_event_with_time(struct perf_tool *tool, return 0; } -static int print_event(struct perf_tool *tool, union perf_event *event, +static int print_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine, pid_t pid, pid_t tid) { return print_event_with_time(tool, event, sample, machine, pid, tid, 0); } -static int process_comm_event(struct perf_tool *tool, +static int process_comm_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2607,7 +2607,7 @@ static int process_comm_event(struct perf_tool *tool, event->comm.tid); } -static int process_namespaces_event(struct perf_tool *tool, +static int process_namespaces_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2619,7 +2619,7 @@ static int process_namespaces_event(struct perf_tool *tool, event->namespaces.tid); } -static int process_cgroup_event(struct perf_tool *tool, +static int process_cgroup_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2631,7 +2631,7 @@ static int process_cgroup_event(struct perf_tool *tool, sample->tid); } -static int process_fork_event(struct perf_tool *tool, +static int process_fork_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2643,7 +2643,7 @@ static int process_fork_event(struct perf_tool *tool, event->fork.pid, event->fork.tid, event->fork.time); } -static int process_exit_event(struct perf_tool *tool, +static int process_exit_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2656,7 +2656,7 @@ static int process_exit_event(struct perf_tool *tool, return perf_event__process_exit(tool, event, sample, machine); } -static int process_mmap_event(struct perf_tool *tool, +static int process_mmap_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2668,7 +2668,7 @@ static int process_mmap_event(struct perf_tool *tool, event->mmap.tid); } -static int process_mmap2_event(struct perf_tool *tool, +static int process_mmap2_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2680,7 +2680,7 @@ static int process_mmap2_event(struct perf_tool *tool, event->mmap2.tid); } -static int process_switch_event(struct perf_tool *tool, +static int process_switch_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2712,7 +2712,7 @@ static int process_auxtrace_error(struct perf_session *session, } static int -process_lost_event(struct perf_tool *tool, +process_lost_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2722,7 +2722,7 @@ process_lost_event(struct perf_tool *tool, } static int -process_throttle_event(struct perf_tool *tool __maybe_unused, +process_throttle_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2733,7 +2733,7 @@ process_throttle_event(struct perf_tool *tool __maybe_unused, } static int -process_finished_round_event(struct perf_tool *tool __maybe_unused, +process_finished_round_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct ordered_events *oe __maybe_unused) @@ -2743,7 +2743,7 @@ process_finished_round_event(struct perf_tool *tool __maybe_unused, } static int -process_bpf_events(struct perf_tool *tool __maybe_unused, +process_bpf_events(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -2755,7 +2755,7 @@ process_bpf_events(struct perf_tool *tool __maybe_unused, sample->tid); } -static int process_text_poke_events(struct perf_tool *tool, +static int process_text_poke_events(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -3757,7 +3757,7 @@ static int process_thread_map_event(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct perf_script *script = container_of(tool, struct perf_script, tool); if (dump_trace) @@ -3779,7 +3779,7 @@ static int process_cpu_map_event(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct perf_script *script = container_of(tool, struct perf_script, tool); if (dump_trace) @@ -3809,11 +3809,10 @@ static int process_feature_event(struct perf_session *session, static int perf_script__process_auxtrace_info(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; - int ret = perf_event__process_auxtrace_info(session, event); if (ret == 0) { + const struct perf_tool *tool = session->tool; struct perf_script *script = container_of(tool, struct perf_script, tool); ret = perf_script__setup_per_event_dump(script); diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 661832756a24..d460c46f15fd 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -248,7 +248,7 @@ static void perf_stat__reset_stats(void) perf_stat__reset_shadow_stats(); } -static int process_synthesized_event(struct perf_tool *tool __maybe_unused, +static int process_synthesized_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -2180,7 +2180,7 @@ static int process_stat_config_event(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct perf_stat *st = container_of(tool, struct perf_stat, tool); perf_event__read_stat_config(&stat_config, &event->stat_config); @@ -2229,7 +2229,7 @@ static int process_thread_map_event(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct perf_stat *st = container_of(tool, struct perf_stat, tool); if (st->threads) { @@ -2248,7 +2248,7 @@ static int process_cpu_map_event(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct perf_stat *st = container_of(tool, struct perf_stat, tool); struct perf_cpu_map *cpus; diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 19d4542ea18a..5bf818baa662 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -320,7 +320,7 @@ static int *cpus_cstate_state; static u64 *cpus_pstate_start_times; static u64 *cpus_pstate_state; -static int process_comm_event(struct perf_tool *tool, +static int process_comm_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -330,7 +330,7 @@ static int process_comm_event(struct perf_tool *tool, return 0; } -static int process_fork_event(struct perf_tool *tool, +static int process_fork_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -340,7 +340,7 @@ static int process_fork_event(struct perf_tool *tool, return 0; } -static int process_exit_event(struct perf_tool *tool, +static int process_exit_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -571,7 +571,7 @@ typedef int (*tracepoint_handler)(struct timechart *tchart, struct perf_sample *sample, const char *backtrace); -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index e8cbbf10d361..d6fcc6c122da 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -740,7 +740,7 @@ static int hist_iter__top_callback(struct hist_entry_iter *iter, return 0; } -static void perf_event__process_sample(struct perf_tool *tool, +static void perf_event__process_sample(const struct perf_tool *tool, const union perf_event *event, struct evsel *evsel, struct perf_sample *sample, diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 8449f2beb54d..3b9a0278d4a4 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -1637,7 +1637,7 @@ static int trace__process_event(struct trace *trace, struct machine *machine, return ret; } -static int trace__tool_process(struct perf_tool *tool, +static int trace__tool_process(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -3009,7 +3009,7 @@ static void trace__set_base_time(struct trace *trace, trace->base_time = sample->time; } -static int trace__process_sample(struct perf_tool *tool, +static int trace__process_sample(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, diff --git a/tools/perf/tests/cpumap.c b/tools/perf/tests/cpumap.c index bd8e396f3e57..2f0168b2a5a9 100644 --- a/tools/perf/tests/cpumap.c +++ b/tools/perf/tests/cpumap.c @@ -11,7 +11,7 @@ struct machine; -static int process_event_mask(struct perf_tool *tool __maybe_unused, +static int process_event_mask(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -47,7 +47,7 @@ static int process_event_mask(struct perf_tool *tool __maybe_unused, return 0; } -static int process_event_cpus(struct perf_tool *tool __maybe_unused, +static int process_event_cpus(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -73,7 +73,7 @@ static int process_event_cpus(struct perf_tool *tool __maybe_unused, return 0; } -static int process_event_range_cpus(struct perf_tool *tool __maybe_unused, +static int process_event_range_cpus(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) diff --git a/tools/perf/tests/dlfilter-test.c b/tools/perf/tests/dlfilter-test.c index da3a9b50b1b1..54f59d1246bc 100644 --- a/tools/perf/tests/dlfilter-test.c +++ b/tools/perf/tests/dlfilter-test.c @@ -62,7 +62,7 @@ static int test_result(const char *msg, int ret) return ret; } -static int process(struct perf_tool *tool, union perf_event *event, +static int process(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) { diff --git a/tools/perf/tests/dwarf-unwind.c b/tools/perf/tests/dwarf-unwind.c index d01aa931fe81..f85d391ced98 100644 --- a/tools/perf/tests/dwarf-unwind.c +++ b/tools/perf/tests/dwarf-unwind.c @@ -37,7 +37,7 @@ #define NO_TAIL_CALL_BARRIER __asm__ __volatile__("" : : : "memory"); #endif -static int mmap_handler(struct perf_tool *tool __maybe_unused, +static int mmap_handler(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) diff --git a/tools/perf/tests/event_update.c b/tools/perf/tests/event_update.c index d093a9b878d1..fdecad920f59 100644 --- a/tools/perf/tests/event_update.c +++ b/tools/perf/tests/event_update.c @@ -12,7 +12,7 @@ #include "tests.h" #include "debug.h" -static int process_event_unit(struct perf_tool *tool __maybe_unused, +static int process_event_unit(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -25,7 +25,7 @@ static int process_event_unit(struct perf_tool *tool __maybe_unused, return 0; } -static int process_event_scale(struct perf_tool *tool __maybe_unused, +static int process_event_scale(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -43,7 +43,7 @@ struct event_name { const char *name; }; -static int process_event_name(struct perf_tool *tool, +static int process_event_name(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -57,7 +57,7 @@ static int process_event_name(struct perf_tool *tool, return 0; } -static int process_event_cpus(struct perf_tool *tool __maybe_unused, +static int process_event_cpus(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) diff --git a/tools/perf/tests/stat.c b/tools/perf/tests/stat.c index 706780fb5695..6468cc0d0204 100644 --- a/tools/perf/tests/stat.c +++ b/tools/perf/tests/stat.c @@ -21,7 +21,7 @@ static bool has_term(struct perf_record_stat_config *config, return false; } -static int process_stat_config_event(struct perf_tool *tool __maybe_unused, +static int process_stat_config_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -62,7 +62,7 @@ static int test__synthesize_stat_config(struct test_suite *test __maybe_unused, return 0; } -static int process_stat_event(struct perf_tool *tool __maybe_unused, +static int process_stat_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -93,7 +93,7 @@ static int test__synthesize_stat(struct test_suite *test __maybe_unused, int sub return 0; } -static int process_stat_round_event(struct perf_tool *tool __maybe_unused, +static int process_stat_round_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) diff --git a/tools/perf/tests/thread-map.c b/tools/perf/tests/thread-map.c index 74308c1368fe..1fe521466bf4 100644 --- a/tools/perf/tests/thread-map.c +++ b/tools/perf/tests/thread-map.c @@ -60,7 +60,7 @@ static int test__thread_map(struct test_suite *test __maybe_unused, int subtest return 0; } -static int process_event(struct perf_tool *tool __maybe_unused, +static int process_event(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) diff --git a/tools/perf/util/arm-spe.c b/tools/perf/util/arm-spe.c index fa40f3cb6266..138ffc71b32d 100644 --- a/tools/perf/util/arm-spe.c +++ b/tools/perf/util/arm-spe.c @@ -899,7 +899,7 @@ static int arm_spe_context_switch(struct arm_spe *spe, union perf_event *event, static int arm_spe_process_event(struct perf_session *session, union perf_event *event, struct perf_sample *sample, - struct perf_tool *tool) + const struct perf_tool *tool) { int err = 0; u64 timestamp; @@ -947,7 +947,7 @@ static int arm_spe_process_event(struct perf_session *session, static int arm_spe_process_auxtrace_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct arm_spe *spe = container_of(session->auxtrace, struct arm_spe, auxtrace); @@ -985,7 +985,7 @@ static int arm_spe_process_auxtrace_event(struct perf_session *session, } static int arm_spe_flush(struct perf_session *session __maybe_unused, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct arm_spe *spe = container_of(session->auxtrace, struct arm_spe, auxtrace); diff --git a/tools/perf/util/auxtrace.c b/tools/perf/util/auxtrace.c index e2f317063eec..cbb773ed6f1a 100644 --- a/tools/perf/util/auxtrace.c +++ b/tools/perf/util/auxtrace.c @@ -1240,7 +1240,7 @@ void auxtrace_synth_error(struct perf_record_auxtrace_error *auxtrace_error, int } int perf_event__synthesize_auxtrace_info(struct auxtrace_record *itr, - struct perf_tool *tool, + const struct perf_tool *tool, struct perf_session *session, perf_event__handler_t process) { @@ -1831,7 +1831,7 @@ int __weak compat_auxtrace_mmap__write_tail(struct auxtrace_mmap *mm, u64 tail) static int __auxtrace_mmap__read(struct mmap *map, struct auxtrace_record *itr, - struct perf_tool *tool, process_auxtrace_t fn, + const struct perf_tool *tool, process_auxtrace_t fn, bool snapshot, size_t snapshot_size) { struct auxtrace_mmap *mm = &map->auxtrace_mmap; @@ -1942,14 +1942,14 @@ static int __auxtrace_mmap__read(struct mmap *map, } int auxtrace_mmap__read(struct mmap *map, struct auxtrace_record *itr, - struct perf_tool *tool, process_auxtrace_t fn) + const struct perf_tool *tool, process_auxtrace_t fn) { return __auxtrace_mmap__read(map, itr, tool, fn, false, 0); } int auxtrace_mmap__read_snapshot(struct mmap *map, struct auxtrace_record *itr, - struct perf_tool *tool, process_auxtrace_t fn, + const struct perf_tool *tool, process_auxtrace_t fn, size_t snapshot_size) { return __auxtrace_mmap__read(map, itr, tool, fn, true, snapshot_size); @@ -2829,7 +2829,7 @@ int auxtrace_parse_filters(struct evlist *evlist) } int auxtrace__process_event(struct perf_session *session, union perf_event *event, - struct perf_sample *sample, struct perf_tool *tool) + struct perf_sample *sample, const struct perf_tool *tool) { if (!session->auxtrace) return 0; @@ -2847,7 +2847,7 @@ void auxtrace__dump_auxtrace_sample(struct perf_session *session, session->auxtrace->dump_auxtrace_sample(session, sample); } -int auxtrace__flush_events(struct perf_session *session, struct perf_tool *tool) +int auxtrace__flush_events(struct perf_session *session, const struct perf_tool *tool) { if (!session->auxtrace) return 0; diff --git a/tools/perf/util/auxtrace.h b/tools/perf/util/auxtrace.h index 8a6ec9565835..d405efcd8708 100644 --- a/tools/perf/util/auxtrace.h +++ b/tools/perf/util/auxtrace.h @@ -208,17 +208,17 @@ struct auxtrace { int (*process_event)(struct perf_session *session, union perf_event *event, struct perf_sample *sample, - struct perf_tool *tool); + const struct perf_tool *tool); int (*process_auxtrace_event)(struct perf_session *session, union perf_event *event, - struct perf_tool *tool); + const struct perf_tool *tool); int (*queue_data)(struct perf_session *session, struct perf_sample *sample, union perf_event *event, u64 data_offset); void (*dump_auxtrace_sample)(struct perf_session *session, struct perf_sample *sample); int (*flush_events)(struct perf_session *session, - struct perf_tool *tool); + const struct perf_tool *tool); void (*free_events)(struct perf_session *session); void (*free)(struct perf_session *session); bool (*evsel_is_auxtrace)(struct perf_session *session, @@ -508,17 +508,17 @@ void auxtrace_mmap_params__set_idx(struct auxtrace_mmap_params *mp, struct evlist *evlist, struct evsel *evsel, int idx); -typedef int (*process_auxtrace_t)(struct perf_tool *tool, +typedef int (*process_auxtrace_t)(const struct perf_tool *tool, struct mmap *map, union perf_event *event, void *data1, size_t len1, void *data2, size_t len2); int auxtrace_mmap__read(struct mmap *map, struct auxtrace_record *itr, - struct perf_tool *tool, process_auxtrace_t fn); + const struct perf_tool *tool, process_auxtrace_t fn); int auxtrace_mmap__read_snapshot(struct mmap *map, struct auxtrace_record *itr, - struct perf_tool *tool, process_auxtrace_t fn, + const struct perf_tool *tool, process_auxtrace_t fn, size_t snapshot_size); int auxtrace_queues__init_nr(struct auxtrace_queues *queues, int nr_queues); @@ -639,10 +639,10 @@ int addr_filters__parse_bare_filter(struct addr_filters *filts, int auxtrace_parse_filters(struct evlist *evlist); int auxtrace__process_event(struct perf_session *session, union perf_event *event, - struct perf_sample *sample, struct perf_tool *tool); + struct perf_sample *sample, const struct perf_tool *tool); void auxtrace__dump_auxtrace_sample(struct perf_session *session, struct perf_sample *sample); -int auxtrace__flush_events(struct perf_session *session, struct perf_tool *tool); +int auxtrace__flush_events(struct perf_session *session, const struct perf_tool *tool); void auxtrace__free_events(struct perf_session *session); void auxtrace__free(struct perf_session *session); bool auxtrace__evsel_is_auxtrace(struct perf_session *session, @@ -809,7 +809,7 @@ static inline int auxtrace__process_event(struct perf_session *session __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { return 0; } @@ -822,7 +822,7 @@ void auxtrace__dump_auxtrace_sample(struct perf_session *session __maybe_unused, static inline int auxtrace__flush_events(struct perf_session *session __maybe_unused, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { return 0; } diff --git a/tools/perf/util/bpf-event.c b/tools/perf/util/bpf-event.c index 827695cd0408..13608237c50e 100644 --- a/tools/perf/util/bpf-event.c +++ b/tools/perf/util/bpf-event.c @@ -170,7 +170,7 @@ static int perf_event__synthesize_one_bpf_prog(struct perf_session *session, { struct perf_record_ksymbol *ksymbol_event = &event->ksymbol; struct perf_record_bpf_event *bpf_event = &event->bpf; - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct bpf_prog_info_node *info_node; struct perf_bpil *info_linear; struct bpf_prog_info *info; @@ -310,7 +310,7 @@ struct kallsyms_parse { union perf_event *event; perf_event__handler_t process; struct machine *machine; - struct perf_tool *tool; + const struct perf_tool *tool; }; static int diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c index 83a1581e8cf1..098fcc625d91 100644 --- a/tools/perf/util/build-id.c +++ b/tools/perf/util/build-id.c @@ -42,7 +42,7 @@ static bool no_buildid_cache; -int build_id__mark_dso_hit(struct perf_tool *tool __maybe_unused, +int build_id__mark_dso_hit(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct evsel *evsel __maybe_unused, @@ -67,7 +67,7 @@ int build_id__mark_dso_hit(struct perf_tool *tool __maybe_unused, return 0; } -static int perf_event__exit_del_thread(struct perf_tool *tool __maybe_unused, +static int perf_event__exit_del_thread(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, diff --git a/tools/perf/util/build-id.h b/tools/perf/util/build-id.h index 3fa8bffb07ca..ae87c4c58d5b 100644 --- a/tools/perf/util/build-id.h +++ b/tools/perf/util/build-id.h @@ -35,11 +35,11 @@ char *dso__build_id_filename(const struct dso *dso, char *bf, size_t size, char *__dso__build_id_filename(const struct dso *dso, char *bf, size_t size, bool is_debug, bool is_kallsyms); -int build_id__mark_dso_hit(struct perf_tool *tool, union perf_event *event, +int build_id__mark_dso_hit(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine); -int perf_event__inject_buildid(struct perf_tool *tool, union perf_event *event, +int perf_event__inject_buildid(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine); diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c index 7fd976ea954b..21023d2a58ee 100644 --- a/tools/perf/util/cs-etm.c +++ b/tools/perf/util/cs-etm.c @@ -788,7 +788,7 @@ static void cs_etm__dump_event(struct cs_etm_queue *etmq, } static int cs_etm__flush_events(struct perf_session *session, - struct perf_tool *tool) + const struct perf_tool *tool) { struct cs_etm_auxtrace *etm = container_of(session->auxtrace, struct cs_etm_auxtrace, @@ -2711,7 +2711,7 @@ static int cs_etm__process_switch_cpu_wide(struct cs_etm_auxtrace *etm, static int cs_etm__process_event(struct perf_session *session, union perf_event *event, struct perf_sample *sample, - struct perf_tool *tool) + const struct perf_tool *tool) { struct cs_etm_auxtrace *etm = container_of(session->auxtrace, struct cs_etm_auxtrace, @@ -2781,7 +2781,7 @@ static void dump_queued_data(struct cs_etm_auxtrace *etm, static int cs_etm__process_auxtrace_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct cs_etm_auxtrace *etm = container_of(session->auxtrace, struct cs_etm_auxtrace, diff --git a/tools/perf/util/data-convert-bt.c b/tools/perf/util/data-convert-bt.c index 2b732bccabad..9e2170604b66 100644 --- a/tools/perf/util/data-convert-bt.c +++ b/tools/perf/util/data-convert-bt.c @@ -792,7 +792,7 @@ static bool is_flush_needed(struct ctf_stream *cs) return cs->count >= STREAM_FLUSH_COUNT; } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *_event, struct perf_sample *sample, struct evsel *evsel, @@ -871,7 +871,7 @@ do { \ } while(0) #define __FUNC_PROCESS_NON_SAMPLE(_name, body) \ -static int process_##_name##_event(struct perf_tool *tool, \ +static int process_##_name##_event(const struct perf_tool *tool, \ union perf_event *_event, \ struct perf_sample *sample, \ struct machine *machine) \ diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c index 3cf64f5b23ee..905ea9823f9d 100644 --- a/tools/perf/util/data-convert-json.c +++ b/tools/perf/util/data-convert-json.c @@ -118,7 +118,7 @@ static void output_json_key_format(FILE *out, bool comma, int depth, va_end(args); } -static void output_sample_callchain_entry(struct perf_tool *tool, +static void output_sample_callchain_entry(const struct perf_tool *tool, u64 ip, struct addr_location *al) { struct convert_json *c = container_of(tool, struct convert_json, tool); @@ -146,7 +146,7 @@ static void output_sample_callchain_entry(struct perf_tool *tool, output_json_format(out, false, 4, "}"); } -static int process_sample_event(struct perf_tool *tool, +static int process_sample_event(const struct perf_tool *tool, union perf_event *event __maybe_unused, struct perf_sample *sample, struct evsel *evsel __maybe_unused, diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index f32f9abf6344..c2f0e7f40ad5 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -216,7 +216,7 @@ size_t perf_event__fprintf_cgroup(union perf_event *event, FILE *fp) event->cgroup.id, event->cgroup.path); } -int perf_event__process_comm(struct perf_tool *tool __maybe_unused, +int perf_event__process_comm(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -224,7 +224,7 @@ int perf_event__process_comm(struct perf_tool *tool __maybe_unused, return machine__process_comm_event(machine, event, sample); } -int perf_event__process_namespaces(struct perf_tool *tool __maybe_unused, +int perf_event__process_namespaces(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -232,7 +232,7 @@ int perf_event__process_namespaces(struct perf_tool *tool __maybe_unused, return machine__process_namespaces_event(machine, event, sample); } -int perf_event__process_cgroup(struct perf_tool *tool __maybe_unused, +int perf_event__process_cgroup(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -240,7 +240,7 @@ int perf_event__process_cgroup(struct perf_tool *tool __maybe_unused, return machine__process_cgroup_event(machine, event, sample); } -int perf_event__process_lost(struct perf_tool *tool __maybe_unused, +int perf_event__process_lost(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -248,7 +248,7 @@ int perf_event__process_lost(struct perf_tool *tool __maybe_unused, return machine__process_lost_event(machine, event, sample); } -int perf_event__process_aux(struct perf_tool *tool __maybe_unused, +int perf_event__process_aux(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine) @@ -256,7 +256,7 @@ int perf_event__process_aux(struct perf_tool *tool __maybe_unused, return machine__process_aux_event(machine, event); } -int perf_event__process_itrace_start(struct perf_tool *tool __maybe_unused, +int perf_event__process_itrace_start(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine) @@ -264,7 +264,7 @@ int perf_event__process_itrace_start(struct perf_tool *tool __maybe_unused, return machine__process_itrace_start_event(machine, event); } -int perf_event__process_aux_output_hw_id(struct perf_tool *tool __maybe_unused, +int perf_event__process_aux_output_hw_id(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine) @@ -272,7 +272,7 @@ int perf_event__process_aux_output_hw_id(struct perf_tool *tool __maybe_unused, return machine__process_aux_output_hw_id_event(machine, event); } -int perf_event__process_lost_samples(struct perf_tool *tool __maybe_unused, +int perf_event__process_lost_samples(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -280,7 +280,7 @@ int perf_event__process_lost_samples(struct perf_tool *tool __maybe_unused, return machine__process_lost_samples_event(machine, event, sample); } -int perf_event__process_switch(struct perf_tool *tool __maybe_unused, +int perf_event__process_switch(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine) @@ -288,7 +288,7 @@ int perf_event__process_switch(struct perf_tool *tool __maybe_unused, return machine__process_switch_event(machine, event); } -int perf_event__process_ksymbol(struct perf_tool *tool __maybe_unused, +int perf_event__process_ksymbol(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample __maybe_unused, struct machine *machine) @@ -296,7 +296,7 @@ int perf_event__process_ksymbol(struct perf_tool *tool __maybe_unused, return machine__process_ksymbol(machine, event, sample); } -int perf_event__process_bpf(struct perf_tool *tool __maybe_unused, +int perf_event__process_bpf(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -304,7 +304,7 @@ int perf_event__process_bpf(struct perf_tool *tool __maybe_unused, return machine__process_bpf(machine, event, sample); } -int perf_event__process_text_poke(struct perf_tool *tool __maybe_unused, +int perf_event__process_text_poke(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -387,7 +387,7 @@ size_t perf_event__fprintf_cpu_map(union perf_event *event, FILE *fp) return ret; } -int perf_event__process_mmap(struct perf_tool *tool __maybe_unused, +int perf_event__process_mmap(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -395,7 +395,7 @@ int perf_event__process_mmap(struct perf_tool *tool __maybe_unused, return machine__process_mmap_event(machine, event, sample); } -int perf_event__process_mmap2(struct perf_tool *tool __maybe_unused, +int perf_event__process_mmap2(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -410,7 +410,7 @@ size_t perf_event__fprintf_task(union perf_event *event, FILE *fp) event->fork.ppid, event->fork.ptid); } -int perf_event__process_fork(struct perf_tool *tool __maybe_unused, +int perf_event__process_fork(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -418,7 +418,7 @@ int perf_event__process_fork(struct perf_tool *tool __maybe_unused, return machine__process_fork_event(machine, event, sample); } -int perf_event__process_exit(struct perf_tool *tool __maybe_unused, +int perf_event__process_exit(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) @@ -587,7 +587,7 @@ size_t perf_event__fprintf(union perf_event *event, struct machine *machine, FIL return ret; } -int perf_event__process(struct perf_tool *tool __maybe_unused, +int perf_event__process(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct perf_sample *sample, struct machine *machine) diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index d8bcee2e9b93..4b24f1c580fd 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -267,71 +267,71 @@ struct perf_tool; void perf_event__read_stat_config(struct perf_stat_config *config, struct perf_record_stat_config *event); -int perf_event__process_comm(struct perf_tool *tool, +int perf_event__process_comm(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_lost(struct perf_tool *tool, +int perf_event__process_lost(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_lost_samples(struct perf_tool *tool, +int perf_event__process_lost_samples(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_aux(struct perf_tool *tool, +int perf_event__process_aux(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_itrace_start(struct perf_tool *tool, +int perf_event__process_itrace_start(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_aux_output_hw_id(struct perf_tool *tool, +int perf_event__process_aux_output_hw_id(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_switch(struct perf_tool *tool, +int perf_event__process_switch(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_namespaces(struct perf_tool *tool, +int perf_event__process_namespaces(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_cgroup(struct perf_tool *tool, +int perf_event__process_cgroup(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_mmap(struct perf_tool *tool, +int perf_event__process_mmap(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_mmap2(struct perf_tool *tool, +int perf_event__process_mmap2(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_fork(struct perf_tool *tool, +int perf_event__process_fork(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_exit(struct perf_tool *tool, +int perf_event__process_exit(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_ksymbol(struct perf_tool *tool, +int perf_event__process_ksymbol(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_bpf(struct perf_tool *tool, +int perf_event__process_bpf(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process_text_poke(struct perf_tool *tool, +int perf_event__process_text_poke(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__process(struct perf_tool *tool, +int perf_event__process(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 55e9553861d0..3309fe7f1d12 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -4326,7 +4326,7 @@ int perf_session__read_header(struct perf_session *session, int repipe_fd) int perf_event__process_feature(struct perf_session *session, union perf_event *event) { - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct feat_fd ff = { .fd = 0 }; struct perf_record_header_feature *fe = (struct perf_record_header_feature *)event; int type = fe->header.type; @@ -4405,7 +4405,7 @@ size_t perf_event__fprintf_event_update(union perf_event *event, FILE *fp) return ret; } -int perf_event__process_attr(struct perf_tool *tool __maybe_unused, +int perf_event__process_attr(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct evlist **pevlist) { @@ -4444,7 +4444,7 @@ int perf_event__process_attr(struct perf_tool *tool __maybe_unused, return 0; } -int perf_event__process_event_update(struct perf_tool *tool __maybe_unused, +int perf_event__process_event_update(const struct perf_tool *tool __maybe_unused, union perf_event *event, struct evlist **pevlist) { diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index 7c16a250e738..07ff647197ff 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -156,9 +156,9 @@ int perf_header__fprintf_info(struct perf_session *s, FILE *fp, bool full); int perf_event__process_feature(struct perf_session *session, union perf_event *event); -int perf_event__process_attr(struct perf_tool *tool, union perf_event *event, +int perf_event__process_attr(const struct perf_tool *tool, union perf_event *event, struct evlist **pevlist); -int perf_event__process_event_update(struct perf_tool *tool, +int perf_event__process_event_update(const struct perf_tool *tool, union perf_event *event, struct evlist **pevlist); size_t perf_event__fprintf_event_update(union perf_event *event, FILE *fp); diff --git a/tools/perf/util/hisi-ptt.c b/tools/perf/util/hisi-ptt.c index 37ea987017f6..e4cc4785f744 100644 --- a/tools/perf/util/hisi-ptt.c +++ b/tools/perf/util/hisi-ptt.c @@ -79,14 +79,14 @@ static void hisi_ptt_dump_event(struct hisi_ptt *ptt, unsigned char *buf, static int hisi_ptt_process_event(struct perf_session *session __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { return 0; } static int hisi_ptt_process_auxtrace_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct hisi_ptt *ptt = container_of(session->auxtrace, struct hisi_ptt, auxtrace); @@ -123,7 +123,7 @@ static int hisi_ptt_process_auxtrace_event(struct perf_session *session, } static int hisi_ptt_flush(struct perf_session *session __maybe_unused, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { return 0; } diff --git a/tools/perf/util/intel-bts.c b/tools/perf/util/intel-bts.c index 160d346cb450..27d9b5c9fec8 100644 --- a/tools/perf/util/intel-bts.c +++ b/tools/perf/util/intel-bts.c @@ -591,7 +591,7 @@ static int intel_bts_process_queues(struct intel_bts *bts, u64 timestamp) static int intel_bts_process_event(struct perf_session *session, union perf_event *event, struct perf_sample *sample, - struct perf_tool *tool) + const struct perf_tool *tool) { struct intel_bts *bts = container_of(session->auxtrace, struct intel_bts, auxtrace); @@ -634,7 +634,7 @@ static int intel_bts_process_event(struct perf_session *session, static int intel_bts_process_auxtrace_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct intel_bts *bts = container_of(session->auxtrace, struct intel_bts, auxtrace); @@ -675,7 +675,7 @@ static int intel_bts_process_auxtrace_event(struct perf_session *session, } static int intel_bts_flush(struct perf_session *session, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct intel_bts *bts = container_of(session->auxtrace, struct intel_bts, auxtrace); diff --git a/tools/perf/util/intel-pt.c b/tools/perf/util/intel-pt.c index 1fc03bfd14c3..fd2597613f3d 100644 --- a/tools/perf/util/intel-pt.c +++ b/tools/perf/util/intel-pt.c @@ -3449,7 +3449,7 @@ static int intel_pt_text_poke(struct intel_pt *pt, union perf_event *event) static int intel_pt_process_event(struct perf_session *session, union perf_event *event, struct perf_sample *sample, - struct perf_tool *tool) + const struct perf_tool *tool) { struct intel_pt *pt = container_of(session->auxtrace, struct intel_pt, auxtrace); @@ -3533,7 +3533,7 @@ static int intel_pt_process_event(struct perf_session *session, return err; } -static int intel_pt_flush(struct perf_session *session, struct perf_tool *tool) +static int intel_pt_flush(struct perf_session *session, const struct perf_tool *tool) { struct intel_pt *pt = container_of(session->auxtrace, struct intel_pt, auxtrace); @@ -3600,7 +3600,7 @@ static bool intel_pt_evsel_is_auxtrace(struct perf_session *session, static int intel_pt_process_auxtrace_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct intel_pt *pt = container_of(session->auxtrace, struct intel_pt, auxtrace); diff --git a/tools/perf/util/jitdump.c b/tools/perf/util/jitdump.c index 1f657ef8975f..5ce13653512b 100644 --- a/tools/perf/util/jitdump.c +++ b/tools/perf/util/jitdump.c @@ -424,7 +424,7 @@ static int jit_repipe_code_load(struct jit_buf_desc *jd, union jr_entry *jr) { struct perf_sample sample; union perf_event *event; - struct perf_tool *tool = jd->session->tool; + const struct perf_tool *tool = jd->session->tool; uint64_t code, addr; uintptr_t uaddr; char *filename; @@ -543,7 +543,7 @@ static int jit_repipe_code_move(struct jit_buf_desc *jd, union jr_entry *jr) { struct perf_sample sample; union perf_event *event; - struct perf_tool *tool = jd->session->tool; + const struct perf_tool *tool = jd->session->tool; char *filename; size_t size; struct stat st; diff --git a/tools/perf/util/s390-cpumsf.c b/tools/perf/util/s390-cpumsf.c index 4ec583e511af..73846b73d0cf 100644 --- a/tools/perf/util/s390-cpumsf.c +++ b/tools/perf/util/s390-cpumsf.c @@ -912,7 +912,7 @@ static int s390_cpumsf_process_event(struct perf_session *session, union perf_event *event, struct perf_sample *sample, - struct perf_tool *tool) + const struct perf_tool *tool) { struct s390_cpumsf *sf = container_of(session->auxtrace, struct s390_cpumsf, @@ -955,7 +955,7 @@ s390_cpumsf_process_event(struct perf_session *session, static int s390_cpumsf_process_auxtrace_event(struct perf_session *session, union perf_event *event __maybe_unused, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { struct s390_cpumsf *sf = container_of(session->auxtrace, struct s390_cpumsf, @@ -998,7 +998,7 @@ static void s390_cpumsf_free_events(struct perf_session *session __maybe_unused) } static int s390_cpumsf_flush(struct perf_session *session __maybe_unused, - struct perf_tool *tool __maybe_unused) + const struct perf_tool *tool __maybe_unused) { return 0; } diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 18a8f4a84045..7e9827e61d1b 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -102,7 +102,7 @@ static int perf_session__process_compressed_event(struct perf_session *session, static int perf_session__deliver_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool, + const struct perf_tool *tool, u64 file_offset, const char *file_path); @@ -328,7 +328,7 @@ static int process_event_synth_tracing_data_stub(struct perf_session *session return 0; } -static int process_event_synth_attr_stub(struct perf_tool *tool __maybe_unused, +static int process_event_synth_attr_stub(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct evlist **pevlist __maybe_unused) @@ -337,7 +337,7 @@ static int process_event_synth_attr_stub(struct perf_tool *tool __maybe_unused, return 0; } -static int process_event_synth_event_update_stub(struct perf_tool *tool __maybe_unused, +static int process_event_synth_event_update_stub(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct evlist **pevlist __maybe_unused) @@ -349,7 +349,7 @@ static int process_event_synth_event_update_stub(struct perf_tool *tool __maybe_ return 0; } -static int process_event_sample_stub(struct perf_tool *tool __maybe_unused, +static int process_event_sample_stub(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, struct evsel *evsel __maybe_unused, @@ -359,7 +359,7 @@ static int process_event_sample_stub(struct perf_tool *tool __maybe_unused, return 0; } -static int process_event_stub(struct perf_tool *tool __maybe_unused, +static int process_event_stub(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct perf_sample *sample __maybe_unused, struct machine *machine __maybe_unused) @@ -368,7 +368,7 @@ static int process_event_stub(struct perf_tool *tool __maybe_unused, return 0; } -static int process_finished_round_stub(struct perf_tool *tool __maybe_unused, +static int process_finished_round_stub(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct ordered_events *oe __maybe_unused) { @@ -1076,7 +1076,7 @@ static perf_event__swap_op perf_event__swap_ops[] = { * Flush every events below timestamp 7 * etc... */ -int perf_event__process_finished_round(struct perf_tool *tool __maybe_unused, +int perf_event__process_finished_round(const struct perf_tool *tool __maybe_unused, union perf_event *event __maybe_unused, struct ordered_events *oe) { @@ -1470,7 +1470,7 @@ static struct machine *machines__find_for_cpumode(struct machines *machines, } static int deliver_sample_value(struct evlist *evlist, - struct perf_tool *tool, + const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct sample_read_value *v, @@ -1502,7 +1502,7 @@ static int deliver_sample_value(struct evlist *evlist, } static int deliver_sample_group(struct evlist *evlist, - struct perf_tool *tool, + const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine, @@ -1521,7 +1521,7 @@ static int deliver_sample_group(struct evlist *evlist, return ret; } -static int evlist__deliver_sample(struct evlist *evlist, struct perf_tool *tool, +static int evlist__deliver_sample(struct evlist *evlist, const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine) { @@ -1546,7 +1546,7 @@ static int machines__deliver_event(struct machines *machines, struct evlist *evlist, union perf_event *event, struct perf_sample *sample, - struct perf_tool *tool, u64 file_offset, + const struct perf_tool *tool, u64 file_offset, const char *file_path) { struct evsel *evsel; @@ -1634,7 +1634,7 @@ static int machines__deliver_event(struct machines *machines, static int perf_session__deliver_event(struct perf_session *session, union perf_event *event, - struct perf_tool *tool, + const struct perf_tool *tool, u64 file_offset, const char *file_path) { @@ -1667,7 +1667,7 @@ static s64 perf_session__process_user_event(struct perf_session *session, const char *file_path) { struct ordered_events *oe = &session->ordered_events; - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct perf_sample sample = { .time = 0, }; int fd = perf_data__fd(session->data); int err; @@ -1754,7 +1754,7 @@ int perf_session__deliver_synth_event(struct perf_session *session, struct perf_sample *sample) { struct evlist *evlist = session->evlist; - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; events_stats__inc(&evlist->stats, event->header.type); @@ -1886,7 +1886,7 @@ static s64 perf_session__process_event(struct perf_session *session, const char *file_path) { struct evlist *evlist = session->evlist; - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; int ret; if (session->header.needs_swap) diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index cc28976bb00e..6886cc85600f 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -163,7 +163,7 @@ int perf_session__dsos_hit_all(struct perf_session *session); int perf_event__process_id_index(struct perf_session *session, union perf_event *event); -int perf_event__process_finished_round(struct perf_tool *tool, +int perf_event__process_finished_round(const struct perf_tool *tool, union perf_event *event, struct ordered_events *oe); diff --git a/tools/perf/util/synthetic-events.c b/tools/perf/util/synthetic-events.c index 5498048f56ea..7f884d70de81 100644 --- a/tools/perf/util/synthetic-events.c +++ b/tools/perf/util/synthetic-events.c @@ -1,4 +1,4 @@ -// SPDX-License-Identifier: GPL-2.0-only +// SPDX-License-Identifier: GPL-2.0-only #include "util/cgroup.h" #include "util/data.h" @@ -47,7 +47,7 @@ unsigned int proc_map_timeout = DEFAULT_PROC_MAP_PARSE_TIMEOUT; -int perf_tool__process_synth_event(struct perf_tool *tool, +int perf_tool__process_synth_event(const struct perf_tool *tool, union perf_event *event, struct machine *machine, perf_event__handler_t process) @@ -187,7 +187,7 @@ static int perf_event__prepare_comm(union perf_event *event, pid_t pid, pid_t ti return 0; } -pid_t perf_event__synthesize_comm(struct perf_tool *tool, +pid_t perf_event__synthesize_comm(const struct perf_tool *tool, union perf_event *event, pid_t pid, perf_event__handler_t process, struct machine *machine) @@ -218,7 +218,7 @@ static void perf_event__get_ns_link_info(pid_t pid, const char *ns, } } -int perf_event__synthesize_namespaces(struct perf_tool *tool, +int perf_event__synthesize_namespaces(const struct perf_tool *tool, union perf_event *event, pid_t pid, pid_t tgid, perf_event__handler_t process, @@ -257,7 +257,7 @@ int perf_event__synthesize_namespaces(struct perf_tool *tool, return 0; } -static int perf_event__synthesize_fork(struct perf_tool *tool, +static int perf_event__synthesize_fork(const struct perf_tool *tool, union perf_event *event, pid_t pid, pid_t tgid, pid_t ppid, perf_event__handler_t process, @@ -418,7 +418,7 @@ static void perf_record_mmap2__read_build_id(struct perf_record_mmap2 *event, dso__put(dso); } -int perf_event__synthesize_mmap_events(struct perf_tool *tool, +int perf_event__synthesize_mmap_events(const struct perf_tool *tool, union perf_event *event, pid_t pid, pid_t tgid, perf_event__handler_t process, @@ -542,7 +542,7 @@ int perf_event__synthesize_mmap_events(struct perf_tool *tool, } #ifdef HAVE_FILE_HANDLE -static int perf_event__synthesize_cgroup(struct perf_tool *tool, +static int perf_event__synthesize_cgroup(const struct perf_tool *tool, union perf_event *event, char *path, size_t mount_len, perf_event__handler_t process, @@ -582,7 +582,7 @@ static int perf_event__synthesize_cgroup(struct perf_tool *tool, return 0; } -static int perf_event__walk_cgroup_tree(struct perf_tool *tool, +static int perf_event__walk_cgroup_tree(const struct perf_tool *tool, union perf_event *event, char *path, size_t mount_len, perf_event__handler_t process, @@ -630,7 +630,7 @@ static int perf_event__walk_cgroup_tree(struct perf_tool *tool, return ret; } -int perf_event__synthesize_cgroups(struct perf_tool *tool, +int perf_event__synthesize_cgroups(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine) { @@ -657,7 +657,7 @@ int perf_event__synthesize_cgroups(struct perf_tool *tool, return 0; } #else -int perf_event__synthesize_cgroups(struct perf_tool *tool __maybe_unused, +int perf_event__synthesize_cgroups(const struct perf_tool *tool __maybe_unused, perf_event__handler_t process __maybe_unused, struct machine *machine __maybe_unused) { @@ -666,7 +666,7 @@ int perf_event__synthesize_cgroups(struct perf_tool *tool __maybe_unused, #endif struct perf_event__synthesize_modules_maps_cb_args { - struct perf_tool *tool; + const struct perf_tool *tool; perf_event__handler_t process; struct machine *machine; union perf_event *event; @@ -717,7 +717,7 @@ static int perf_event__synthesize_modules_maps_cb(struct map *map, void *data) return 0; } -int perf_event__synthesize_modules(struct perf_tool *tool, perf_event__handler_t process, +int perf_event__synthesize_modules(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine) { int rc; @@ -763,7 +763,7 @@ static int __event__synthesize_thread(union perf_event *comm_event, union perf_event *fork_event, union perf_event *namespaces_event, pid_t pid, int full, perf_event__handler_t process, - struct perf_tool *tool, struct machine *machine, + const struct perf_tool *tool, struct machine *machine, bool needs_mmap, bool mmap_data) { char filename[PATH_MAX]; @@ -852,7 +852,7 @@ static int __event__synthesize_thread(union perf_event *comm_event, return rc; } -int perf_event__synthesize_thread_map(struct perf_tool *tool, +int perf_event__synthesize_thread_map(const struct perf_tool *tool, struct perf_thread_map *threads, perf_event__handler_t process, struct machine *machine, @@ -929,7 +929,7 @@ int perf_event__synthesize_thread_map(struct perf_tool *tool, return err; } -static int __perf_event__synthesize_threads(struct perf_tool *tool, +static int __perf_event__synthesize_threads(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine, bool needs_mmap, @@ -993,7 +993,7 @@ static int __perf_event__synthesize_threads(struct perf_tool *tool, } struct synthesize_threads_arg { - struct perf_tool *tool; + const struct perf_tool *tool; perf_event__handler_t process; struct machine *machine; bool needs_mmap; @@ -1015,7 +1015,7 @@ static void *synthesize_threads_worker(void *arg) return NULL; } -int perf_event__synthesize_threads(struct perf_tool *tool, +int perf_event__synthesize_threads(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine, bool needs_mmap, bool mmap_data, @@ -1104,14 +1104,14 @@ int perf_event__synthesize_threads(struct perf_tool *tool, return err; } -int __weak perf_event__synthesize_extra_kmaps(struct perf_tool *tool __maybe_unused, +int __weak perf_event__synthesize_extra_kmaps(const struct perf_tool *tool __maybe_unused, perf_event__handler_t process __maybe_unused, struct machine *machine __maybe_unused) { return 0; } -static int __perf_event__synthesize_kernel_mmap(struct perf_tool *tool, +static int __perf_event__synthesize_kernel_mmap(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine) { @@ -1183,7 +1183,7 @@ static int __perf_event__synthesize_kernel_mmap(struct perf_tool *tool, return err; } -int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, +int perf_event__synthesize_kernel_mmap(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine) { @@ -1196,7 +1196,7 @@ int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, return perf_event__synthesize_extra_kmaps(tool, process, machine); } -int perf_event__synthesize_thread_map2(struct perf_tool *tool, +int perf_event__synthesize_thread_map2(const struct perf_tool *tool, struct perf_thread_map *threads, perf_event__handler_t process, struct machine *machine) @@ -1346,7 +1346,7 @@ static struct perf_record_cpu_map *cpu_map_event__new(const struct perf_cpu_map } -int perf_event__synthesize_cpu_map(struct perf_tool *tool, +int perf_event__synthesize_cpu_map(const struct perf_tool *tool, const struct perf_cpu_map *map, perf_event__handler_t process, struct machine *machine) @@ -1364,7 +1364,7 @@ int perf_event__synthesize_cpu_map(struct perf_tool *tool, return err; } -int perf_event__synthesize_stat_config(struct perf_tool *tool, +int perf_event__synthesize_stat_config(const struct perf_tool *tool, struct perf_stat_config *config, perf_event__handler_t process, struct machine *machine) @@ -1403,7 +1403,7 @@ int perf_event__synthesize_stat_config(struct perf_tool *tool, return err; } -int perf_event__synthesize_stat(struct perf_tool *tool, +int perf_event__synthesize_stat(const struct perf_tool *tool, struct perf_cpu cpu, u32 thread, u64 id, struct perf_counts_values *count, perf_event__handler_t process, @@ -1425,7 +1425,7 @@ int perf_event__synthesize_stat(struct perf_tool *tool, return process(tool, (union perf_event *) &event, NULL, machine); } -int perf_event__synthesize_stat_round(struct perf_tool *tool, +int perf_event__synthesize_stat_round(const struct perf_tool *tool, u64 evtime, u64 type, perf_event__handler_t process, struct machine *machine) @@ -1826,7 +1826,7 @@ int perf_event__synthesize_id_sample(__u64 *array, u64 type, const struct perf_s return (void *)array - (void *)start; } -int __perf_event__synthesize_id_index(struct perf_tool *tool, perf_event__handler_t process, +int __perf_event__synthesize_id_index(const struct perf_tool *tool, perf_event__handler_t process, struct evlist *evlist, struct machine *machine, size_t from) { union perf_event *ev; @@ -1918,13 +1918,13 @@ int __perf_event__synthesize_id_index(struct perf_tool *tool, perf_event__handle return err; } -int perf_event__synthesize_id_index(struct perf_tool *tool, perf_event__handler_t process, +int perf_event__synthesize_id_index(const struct perf_tool *tool, perf_event__handler_t process, struct evlist *evlist, struct machine *machine) { return __perf_event__synthesize_id_index(tool, process, evlist, machine, 0); } -int __machine__synthesize_threads(struct machine *machine, struct perf_tool *tool, +int __machine__synthesize_threads(struct machine *machine, const struct perf_tool *tool, struct target *target, struct perf_thread_map *threads, perf_event__handler_t process, bool needs_mmap, bool data_mmap, unsigned int nr_threads_synthesize) @@ -1985,7 +1985,7 @@ static struct perf_record_event_update *event_update_event__new(size_t size, u64 return ev; } -int perf_event__synthesize_event_update_unit(struct perf_tool *tool, struct evsel *evsel, +int perf_event__synthesize_event_update_unit(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process) { size_t size = strlen(evsel->unit); @@ -2002,7 +2002,7 @@ int perf_event__synthesize_event_update_unit(struct perf_tool *tool, struct evse return err; } -int perf_event__synthesize_event_update_scale(struct perf_tool *tool, struct evsel *evsel, +int perf_event__synthesize_event_update_scale(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process) { struct perf_record_event_update *ev; @@ -2019,7 +2019,7 @@ int perf_event__synthesize_event_update_scale(struct perf_tool *tool, struct evs return err; } -int perf_event__synthesize_event_update_name(struct perf_tool *tool, struct evsel *evsel, +int perf_event__synthesize_event_update_name(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process) { struct perf_record_event_update *ev; @@ -2036,7 +2036,7 @@ int perf_event__synthesize_event_update_name(struct perf_tool *tool, struct evse return err; } -int perf_event__synthesize_event_update_cpus(struct perf_tool *tool, struct evsel *evsel, +int perf_event__synthesize_event_update_cpus(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process) { struct synthesize_cpu_map_data syn_data = { .map = evsel->core.own_cpus }; @@ -2059,7 +2059,7 @@ int perf_event__synthesize_event_update_cpus(struct perf_tool *tool, struct evse return err; } -int perf_event__synthesize_attrs(struct perf_tool *tool, struct evlist *evlist, +int perf_event__synthesize_attrs(const struct perf_tool *tool, struct evlist *evlist, perf_event__handler_t process) { struct evsel *evsel; @@ -2087,7 +2087,7 @@ static bool has_scale(struct evsel *evsel) return evsel->scale != 1; } -int perf_event__synthesize_extra_attr(struct perf_tool *tool, struct evlist *evsel_list, +int perf_event__synthesize_extra_attr(const struct perf_tool *tool, struct evlist *evsel_list, perf_event__handler_t process, bool is_pipe) { struct evsel *evsel; @@ -2143,7 +2143,7 @@ int perf_event__synthesize_extra_attr(struct perf_tool *tool, struct evlist *evs return 0; } -int perf_event__synthesize_attr(struct perf_tool *tool, struct perf_event_attr *attr, +int perf_event__synthesize_attr(const struct perf_tool *tool, struct perf_event_attr *attr, u32 ids, u64 *id, perf_event__handler_t process) { union perf_event *ev; @@ -2177,7 +2177,7 @@ int perf_event__synthesize_attr(struct perf_tool *tool, struct perf_event_attr * } #ifdef HAVE_LIBTRACEEVENT -int perf_event__synthesize_tracing_data(struct perf_tool *tool, int fd, struct evlist *evlist, +int perf_event__synthesize_tracing_data(const struct perf_tool *tool, int fd, struct evlist *evlist, perf_event__handler_t process) { union perf_event ev; @@ -2225,7 +2225,7 @@ int perf_event__synthesize_tracing_data(struct perf_tool *tool, int fd, struct e } #endif -int perf_event__synthesize_build_id(struct perf_tool *tool, struct dso *pos, u16 misc, +int perf_event__synthesize_build_id(const struct perf_tool *tool, struct dso *pos, u16 misc, perf_event__handler_t process, struct machine *machine) { union perf_event ev; @@ -2249,7 +2249,7 @@ int perf_event__synthesize_build_id(struct perf_tool *tool, struct dso *pos, u16 return process(tool, &ev, NULL, machine); } -int perf_event__synthesize_stat_events(struct perf_stat_config *config, struct perf_tool *tool, +int perf_event__synthesize_stat_events(struct perf_stat_config *config, const struct perf_tool *tool, struct evlist *evlist, perf_event__handler_t process, bool attrs) { int err; @@ -2286,7 +2286,7 @@ int perf_event__synthesize_stat_events(struct perf_stat_config *config, struct p extern const struct perf_header_feature_ops feat_ops[HEADER_LAST_FEATURE]; -int perf_event__synthesize_features(struct perf_tool *tool, struct perf_session *session, +int perf_event__synthesize_features(const struct perf_tool *tool, struct perf_session *session, struct evlist *evlist, perf_event__handler_t process) { struct perf_header *header = &session->header; @@ -2349,7 +2349,7 @@ int perf_event__synthesize_features(struct perf_tool *tool, struct perf_session return ret; } -int perf_event__synthesize_for_pipe(struct perf_tool *tool, +int perf_event__synthesize_for_pipe(const struct perf_tool *tool, struct perf_session *session, struct perf_data *data, perf_event__handler_t process) diff --git a/tools/perf/util/synthetic-events.h b/tools/perf/util/synthetic-events.h index 53737d1619a4..31df7653677f 100644 --- a/tools/perf/util/synthetic-events.h +++ b/tools/perf/util/synthetic-events.h @@ -40,45 +40,45 @@ enum perf_record_synth { int parse_synth_opt(char *str); -typedef int (*perf_event__handler_t)(struct perf_tool *tool, union perf_event *event, +typedef int (*perf_event__handler_t)(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -int perf_event__synthesize_attrs(struct perf_tool *tool, struct evlist *evlist, perf_event__handler_t process); -int perf_event__synthesize_attr(struct perf_tool *tool, struct perf_event_attr *attr, u32 ids, u64 *id, perf_event__handler_t process); -int perf_event__synthesize_build_id(struct perf_tool *tool, struct dso *pos, u16 misc, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_cpu_map(struct perf_tool *tool, const struct perf_cpu_map *cpus, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_event_update_cpus(struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); -int perf_event__synthesize_event_update_name(struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); -int perf_event__synthesize_event_update_scale(struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); -int perf_event__synthesize_event_update_unit(struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); -int perf_event__synthesize_extra_attr(struct perf_tool *tool, struct evlist *evsel_list, perf_event__handler_t process, bool is_pipe); -int perf_event__synthesize_extra_kmaps(struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_features(struct perf_tool *tool, struct perf_session *session, struct evlist *evlist, perf_event__handler_t process); -int perf_event__synthesize_id_index(struct perf_tool *tool, perf_event__handler_t process, struct evlist *evlist, struct machine *machine); -int __perf_event__synthesize_id_index(struct perf_tool *tool, perf_event__handler_t process, struct evlist *evlist, struct machine *machine, size_t from); +int perf_event__synthesize_attrs(const struct perf_tool *tool, struct evlist *evlist, perf_event__handler_t process); +int perf_event__synthesize_attr(const struct perf_tool *tool, struct perf_event_attr *attr, u32 ids, u64 *id, perf_event__handler_t process); +int perf_event__synthesize_build_id(const struct perf_tool *tool, struct dso *pos, u16 misc, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_cpu_map(const struct perf_tool *tool, const struct perf_cpu_map *cpus, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_event_update_cpus(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); +int perf_event__synthesize_event_update_name(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); +int perf_event__synthesize_event_update_scale(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); +int perf_event__synthesize_event_update_unit(const struct perf_tool *tool, struct evsel *evsel, perf_event__handler_t process); +int perf_event__synthesize_extra_attr(const struct perf_tool *tool, struct evlist *evsel_list, perf_event__handler_t process, bool is_pipe); +int perf_event__synthesize_extra_kmaps(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_features(const struct perf_tool *tool, struct perf_session *session, struct evlist *evlist, perf_event__handler_t process); +int perf_event__synthesize_id_index(const struct perf_tool *tool, perf_event__handler_t process, struct evlist *evlist, struct machine *machine); +int __perf_event__synthesize_id_index(const struct perf_tool *tool, perf_event__handler_t process, struct evlist *evlist, struct machine *machine, size_t from); int perf_event__synthesize_id_sample(__u64 *array, u64 type, const struct perf_sample *sample); -int perf_event__synthesize_kernel_mmap(struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_mmap_events(struct perf_tool *tool, union perf_event *event, pid_t pid, pid_t tgid, perf_event__handler_t process, struct machine *machine, bool mmap_data); -int perf_event__synthesize_modules(struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_namespaces(struct perf_tool *tool, union perf_event *event, pid_t pid, pid_t tgid, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_cgroups(struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_kernel_mmap(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_mmap_events(const struct perf_tool *tool, union perf_event *event, pid_t pid, pid_t tgid, perf_event__handler_t process, struct machine *machine, bool mmap_data); +int perf_event__synthesize_modules(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_namespaces(const struct perf_tool *tool, union perf_event *event, pid_t pid, pid_t tgid, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_cgroups(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); int perf_event__synthesize_sample(union perf_event *event, u64 type, u64 read_format, const struct perf_sample *sample); -int perf_event__synthesize_stat_config(struct perf_tool *tool, struct perf_stat_config *config, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_stat_events(struct perf_stat_config *config, struct perf_tool *tool, struct evlist *evlist, perf_event__handler_t process, bool attrs); -int perf_event__synthesize_stat_round(struct perf_tool *tool, u64 time, u64 type, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_stat(struct perf_tool *tool, struct perf_cpu cpu, u32 thread, u64 id, struct perf_counts_values *count, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_thread_map2(struct perf_tool *tool, struct perf_thread_map *threads, perf_event__handler_t process, struct machine *machine); -int perf_event__synthesize_thread_map(struct perf_tool *tool, struct perf_thread_map *threads, perf_event__handler_t process, struct machine *machine, bool needs_mmap, bool mmap_data); -int perf_event__synthesize_threads(struct perf_tool *tool, perf_event__handler_t process, struct machine *machine, bool needs_mmap, bool mmap_data, unsigned int nr_threads_synthesize); -int perf_event__synthesize_tracing_data(struct perf_tool *tool, int fd, struct evlist *evlist, perf_event__handler_t process); -int perf_event__synth_time_conv(const struct perf_event_mmap_page *pc, struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); -pid_t perf_event__synthesize_comm(struct perf_tool *tool, union perf_event *event, pid_t pid, perf_event__handler_t process, struct machine *machine); - -int perf_tool__process_synth_event(struct perf_tool *tool, union perf_event *event, struct machine *machine, perf_event__handler_t process); +int perf_event__synthesize_stat_config(const struct perf_tool *tool, struct perf_stat_config *config, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_stat_events(struct perf_stat_config *config, const struct perf_tool *tool, struct evlist *evlist, perf_event__handler_t process, bool attrs); +int perf_event__synthesize_stat_round(const struct perf_tool *tool, u64 time, u64 type, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_stat(const struct perf_tool *tool, struct perf_cpu cpu, u32 thread, u64 id, struct perf_counts_values *count, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_thread_map2(const struct perf_tool *tool, struct perf_thread_map *threads, perf_event__handler_t process, struct machine *machine); +int perf_event__synthesize_thread_map(const struct perf_tool *tool, struct perf_thread_map *threads, perf_event__handler_t process, struct machine *machine, bool needs_mmap, bool mmap_data); +int perf_event__synthesize_threads(const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine, bool needs_mmap, bool mmap_data, unsigned int nr_threads_synthesize); +int perf_event__synthesize_tracing_data(const struct perf_tool *tool, int fd, struct evlist *evlist, perf_event__handler_t process); +int perf_event__synth_time_conv(const struct perf_event_mmap_page *pc, const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine); +pid_t perf_event__synthesize_comm(const struct perf_tool *tool, union perf_event *event, pid_t pid, perf_event__handler_t process, struct machine *machine); + +int perf_tool__process_synth_event(const struct perf_tool *tool, union perf_event *event, struct machine *machine, perf_event__handler_t process); size_t perf_event__sample_event_size(const struct perf_sample *sample, u64 type, u64 read_format); -int __machine__synthesize_threads(struct machine *machine, struct perf_tool *tool, +int __machine__synthesize_threads(struct machine *machine, const struct perf_tool *tool, struct target *target, struct perf_thread_map *threads, perf_event__handler_t process, bool needs_mmap, bool data_mmap, unsigned int nr_threads_synthesize); @@ -87,7 +87,7 @@ int machine__synthesize_threads(struct machine *machine, struct target *target, unsigned int nr_threads_synthesize); #ifdef HAVE_AUXTRACE_SUPPORT -int perf_event__synthesize_auxtrace_info(struct auxtrace_record *itr, struct perf_tool *tool, +int perf_event__synthesize_auxtrace_info(struct auxtrace_record *itr, const struct perf_tool *tool, struct perf_session *session, perf_event__handler_t process); #else // HAVE_AUXTRACE_SUPPORT @@ -96,7 +96,7 @@ int perf_event__synthesize_auxtrace_info(struct auxtrace_record *itr, struct per static inline int perf_event__synthesize_auxtrace_info(struct auxtrace_record *itr __maybe_unused, - struct perf_tool *tool __maybe_unused, + const struct perf_tool *tool __maybe_unused, struct perf_session *session __maybe_unused, perf_event__handler_t process __maybe_unused) { @@ -117,7 +117,7 @@ static inline int perf_event__synthesize_bpf_events(struct perf_session *session } #endif // HAVE_LIBBPF_SUPPORT -int perf_event__synthesize_for_pipe(struct perf_tool *tool, +int perf_event__synthesize_for_pipe(const struct perf_tool *tool, struct perf_session *session, struct perf_data *data, perf_event__handler_t process); diff --git a/tools/perf/util/tool.h b/tools/perf/util/tool.h index c957fb849ac6..b192d44fe91f 100644 --- a/tools/perf/util/tool.h +++ b/tools/perf/util/tool.h @@ -15,14 +15,14 @@ struct perf_tool; struct machine; struct ordered_events; -typedef int (*event_sample)(struct perf_tool *tool, union perf_event *event, +typedef int (*event_sample)(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine); -typedef int (*event_op)(struct perf_tool *tool, union perf_event *event, +typedef int (*event_op)(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); -typedef int (*event_attr_op)(struct perf_tool *tool, +typedef int (*event_attr_op)(const struct perf_tool *tool, union perf_event *event, struct evlist **pevlist); @@ -31,7 +31,7 @@ typedef s64 (*event_op3)(struct perf_session *session, union perf_event *event); typedef int (*event_op4)(struct perf_session *session, union perf_event *event, u64 data, const char *str); -typedef int (*event_oe)(struct perf_tool *tool, union perf_event *event, +typedef int (*event_oe)(const struct perf_tool *tool, union perf_event *event, struct ordered_events *oe); enum show_feature_header { diff --git a/tools/perf/util/tsc.c b/tools/perf/util/tsc.c index f19791d46e99..2e33a20e1e1b 100644 --- a/tools/perf/util/tsc.c +++ b/tools/perf/util/tsc.c @@ -72,7 +72,7 @@ int perf_read_tsc_conversion(const struct perf_event_mmap_page *pc, } int perf_event__synth_time_conv(const struct perf_event_mmap_page *pc, - struct perf_tool *tool, + const struct perf_tool *tool, perf_event__handler_t process, struct machine *machine) { -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 04/27] perf tool: Move fill defaults into tool.c 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (2 preceding siblings ...) 2024-07-18 0:59 ` [PATCH v6 03/27] perf tool: Constify tool pointers Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 05/27] perf tool: Add perf_tool__init Ian Rogers ` (23 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel The aim here is to eventually make perf_tool__fill_defaults an init function so that the tools struct is more const. Create a tool.c to go along with tool.h. Move perf_tool__fill_defaults out of session.c into tool.c along with the default stub values. Add perf_tool__compressed_is_stub for a test in perf_session__process_user_event. perf_session__process_compressed_event is only used from being default initialized so migrate into tool.c. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/util/Build | 1 + tools/perf/util/session.c | 310 +----------------------------------- tools/perf/util/session.h | 2 - tools/perf/util/tool.c | 325 ++++++++++++++++++++++++++++++++++++++ tools/perf/util/tool.h | 4 + 5 files changed, 331 insertions(+), 311 deletions(-) create mode 100644 tools/perf/util/tool.c diff --git a/tools/perf/util/Build b/tools/perf/util/Build index 0f18fe81ef0b..f3633342374a 100644 --- a/tools/perf/util/Build +++ b/tools/perf/util/Build @@ -65,6 +65,7 @@ perf-util-y += map.o perf-util-y += maps.o perf-util-y += pstack.o perf-util-y += session.o +perf-util-y += tool.o perf-util-y += sample-raw.o perf-util-y += s390-sample-raw.o perf-util-y += amd-sample-raw.o diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 7e9827e61d1b..5359ee055c85 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -38,68 +38,6 @@ #include "units.h" #include <internal/lib.h> -#ifdef HAVE_ZSTD_SUPPORT -static int perf_session__process_compressed_event(struct perf_session *session, - union perf_event *event, u64 file_offset, - const char *file_path) -{ - void *src; - size_t decomp_size, src_size; - u64 decomp_last_rem = 0; - size_t mmap_len, decomp_len = session->header.env.comp_mmap_len; - struct decomp *decomp, *decomp_last = session->active_decomp->decomp_last; - - if (decomp_last) { - decomp_last_rem = decomp_last->size - decomp_last->head; - decomp_len += decomp_last_rem; - } - - mmap_len = sizeof(struct decomp) + decomp_len; - decomp = mmap(NULL, mmap_len, PROT_READ|PROT_WRITE, - MAP_ANONYMOUS|MAP_PRIVATE, -1, 0); - if (decomp == MAP_FAILED) { - pr_err("Couldn't allocate memory for decompression\n"); - return -1; - } - - decomp->file_pos = file_offset; - decomp->file_path = file_path; - decomp->mmap_len = mmap_len; - decomp->head = 0; - - if (decomp_last_rem) { - memcpy(decomp->data, &(decomp_last->data[decomp_last->head]), decomp_last_rem); - decomp->size = decomp_last_rem; - } - - src = (void *)event + sizeof(struct perf_record_compressed); - src_size = event->pack.header.size - sizeof(struct perf_record_compressed); - - decomp_size = zstd_decompress_stream(session->active_decomp->zstd_decomp, src, src_size, - &(decomp->data[decomp_last_rem]), decomp_len - decomp_last_rem); - if (!decomp_size) { - munmap(decomp, mmap_len); - pr_err("Couldn't decompress data\n"); - return -1; - } - - decomp->size += decomp_size; - - if (session->active_decomp->decomp == NULL) - session->active_decomp->decomp = decomp; - else - session->active_decomp->decomp_last->next = decomp; - - session->active_decomp->decomp_last = decomp; - - pr_debug("decomp (B): %zd to %zd\n", src_size, decomp_size); - - return 0; -} -#else /* !HAVE_ZSTD_SUPPORT */ -#define perf_session__process_compressed_event perf_session__process_compressed_event_stub -#endif - static int perf_session__deliver_event(struct perf_session *session, union perf_event *event, const struct perf_tool *tool, @@ -319,251 +257,6 @@ void perf_session__delete(struct perf_session *session) free(session); } -static int process_event_synth_tracing_data_stub(struct perf_session *session - __maybe_unused, - union perf_event *event - __maybe_unused) -{ - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_event_synth_attr_stub(const struct perf_tool *tool __maybe_unused, - union perf_event *event __maybe_unused, - struct evlist **pevlist - __maybe_unused) -{ - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_event_synth_event_update_stub(const struct perf_tool *tool __maybe_unused, - union perf_event *event __maybe_unused, - struct evlist **pevlist - __maybe_unused) -{ - if (dump_trace) - perf_event__fprintf_event_update(event, stdout); - - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_event_sample_stub(const struct perf_tool *tool __maybe_unused, - union perf_event *event __maybe_unused, - struct perf_sample *sample __maybe_unused, - struct evsel *evsel __maybe_unused, - struct machine *machine __maybe_unused) -{ - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_event_stub(const struct perf_tool *tool __maybe_unused, - union perf_event *event __maybe_unused, - struct perf_sample *sample __maybe_unused, - struct machine *machine __maybe_unused) -{ - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_finished_round_stub(const struct perf_tool *tool __maybe_unused, - union perf_event *event __maybe_unused, - struct ordered_events *oe __maybe_unused) -{ - dump_printf(": unhandled!\n"); - return 0; -} - -static int skipn(int fd, off_t n) -{ - char buf[4096]; - ssize_t ret; - - while (n > 0) { - ret = read(fd, buf, min(n, (off_t)sizeof(buf))); - if (ret <= 0) - return ret; - n -= ret; - } - - return 0; -} - -static s64 process_event_auxtrace_stub(struct perf_session *session __maybe_unused, - union perf_event *event) -{ - dump_printf(": unhandled!\n"); - if (perf_data__is_pipe(session->data)) - skipn(perf_data__fd(session->data), event->auxtrace.size); - return event->auxtrace.size; -} - -static int process_event_op2_stub(struct perf_session *session __maybe_unused, - union perf_event *event __maybe_unused) -{ - dump_printf(": unhandled!\n"); - return 0; -} - - -static -int process_event_thread_map_stub(struct perf_session *session __maybe_unused, - union perf_event *event __maybe_unused) -{ - if (dump_trace) - perf_event__fprintf_thread_map(event, stdout); - - dump_printf(": unhandled!\n"); - return 0; -} - -static -int process_event_cpu_map_stub(struct perf_session *session __maybe_unused, - union perf_event *event __maybe_unused) -{ - if (dump_trace) - perf_event__fprintf_cpu_map(event, stdout); - - dump_printf(": unhandled!\n"); - return 0; -} - -static -int process_event_stat_config_stub(struct perf_session *session __maybe_unused, - union perf_event *event __maybe_unused) -{ - if (dump_trace) - perf_event__fprintf_stat_config(event, stdout); - - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_stat_stub(struct perf_session *perf_session __maybe_unused, - union perf_event *event) -{ - if (dump_trace) - perf_event__fprintf_stat(event, stdout); - - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_stat_round_stub(struct perf_session *perf_session __maybe_unused, - union perf_event *event) -{ - if (dump_trace) - perf_event__fprintf_stat_round(event, stdout); - - dump_printf(": unhandled!\n"); - return 0; -} - -static int process_event_time_conv_stub(struct perf_session *perf_session __maybe_unused, - union perf_event *event) -{ - if (dump_trace) - perf_event__fprintf_time_conv(event, stdout); - - dump_printf(": unhandled!\n"); - return 0; -} - -static int perf_session__process_compressed_event_stub(struct perf_session *session __maybe_unused, - union perf_event *event __maybe_unused, - u64 file_offset __maybe_unused, - const char *file_path __maybe_unused) -{ - dump_printf(": unhandled!\n"); - return 0; -} - -void perf_tool__fill_defaults(struct perf_tool *tool) -{ - if (tool->sample == NULL) - tool->sample = process_event_sample_stub; - if (tool->mmap == NULL) - tool->mmap = process_event_stub; - if (tool->mmap2 == NULL) - tool->mmap2 = process_event_stub; - if (tool->comm == NULL) - tool->comm = process_event_stub; - if (tool->namespaces == NULL) - tool->namespaces = process_event_stub; - if (tool->cgroup == NULL) - tool->cgroup = process_event_stub; - if (tool->fork == NULL) - tool->fork = process_event_stub; - if (tool->exit == NULL) - tool->exit = process_event_stub; - if (tool->lost == NULL) - tool->lost = perf_event__process_lost; - if (tool->lost_samples == NULL) - tool->lost_samples = perf_event__process_lost_samples; - if (tool->aux == NULL) - tool->aux = perf_event__process_aux; - if (tool->itrace_start == NULL) - tool->itrace_start = perf_event__process_itrace_start; - if (tool->context_switch == NULL) - tool->context_switch = perf_event__process_switch; - if (tool->ksymbol == NULL) - tool->ksymbol = perf_event__process_ksymbol; - if (tool->bpf == NULL) - tool->bpf = perf_event__process_bpf; - if (tool->text_poke == NULL) - tool->text_poke = perf_event__process_text_poke; - if (tool->aux_output_hw_id == NULL) - tool->aux_output_hw_id = perf_event__process_aux_output_hw_id; - if (tool->read == NULL) - tool->read = process_event_sample_stub; - if (tool->throttle == NULL) - tool->throttle = process_event_stub; - if (tool->unthrottle == NULL) - tool->unthrottle = process_event_stub; - if (tool->attr == NULL) - tool->attr = process_event_synth_attr_stub; - if (tool->event_update == NULL) - tool->event_update = process_event_synth_event_update_stub; - if (tool->tracing_data == NULL) - tool->tracing_data = process_event_synth_tracing_data_stub; - if (tool->build_id == NULL) - tool->build_id = process_event_op2_stub; - if (tool->finished_round == NULL) { - if (tool->ordered_events) - tool->finished_round = perf_event__process_finished_round; - else - tool->finished_round = process_finished_round_stub; - } - if (tool->id_index == NULL) - tool->id_index = process_event_op2_stub; - if (tool->auxtrace_info == NULL) - tool->auxtrace_info = process_event_op2_stub; - if (tool->auxtrace == NULL) - tool->auxtrace = process_event_auxtrace_stub; - if (tool->auxtrace_error == NULL) - tool->auxtrace_error = process_event_op2_stub; - if (tool->thread_map == NULL) - tool->thread_map = process_event_thread_map_stub; - if (tool->cpu_map == NULL) - tool->cpu_map = process_event_cpu_map_stub; - if (tool->stat_config == NULL) - tool->stat_config = process_event_stat_config_stub; - if (tool->stat == NULL) - tool->stat = process_stat_stub; - if (tool->stat_round == NULL) - tool->stat_round = process_stat_round_stub; - if (tool->time_conv == NULL) - tool->time_conv = process_event_time_conv_stub; - if (tool->feature == NULL) - tool->feature = process_event_op2_stub; - if (tool->compressed == NULL) - tool->compressed = perf_session__process_compressed_event; - if (tool->finished_init == NULL) - tool->finished_init = process_event_op2_stub; -} - static void swap_sample_id_all(union perf_event *event, void *data) { void *end = (void *) event + event->header.size; @@ -1672,8 +1365,7 @@ static s64 perf_session__process_user_event(struct perf_session *session, int fd = perf_data__fd(session->data); int err; - if (event->header.type != PERF_RECORD_COMPRESSED || - tool->compressed == perf_session__process_compressed_event_stub) + if (event->header.type != PERF_RECORD_COMPRESSED || perf_tool__compressed_is_stub(tool)) dump_event(session->evlist, event, file_offset, &sample, file_path); /* These events are processed right away */ diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 6886cc85600f..7f69baeae7fb 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -92,8 +92,6 @@ int perf_session__process_events(struct perf_session *session); int perf_session__queue_event(struct perf_session *s, union perf_event *event, u64 timestamp, u64 file_offset, const char *file_path); -void perf_tool__fill_defaults(struct perf_tool *tool); - int perf_session__resolve_callchain(struct perf_session *session, struct evsel *evsel, struct thread *thread, diff --git a/tools/perf/util/tool.c b/tools/perf/util/tool.c new file mode 100644 index 000000000000..17219ecb8fa6 --- /dev/null +++ b/tools/perf/util/tool.c @@ -0,0 +1,325 @@ +// SPDX-License-Identifier: GPL-2.0 +#include "data.h" +#include "debug.h" +#include "header.h" +#include "session.h" +#include "stat.h" +#include "tool.h" +#include "tsc.h" +#include <sys/mman.h> +#include <unistd.h> + +#ifdef HAVE_ZSTD_SUPPORT +static int perf_session__process_compressed_event(struct perf_session *session, + union perf_event *event, u64 file_offset, + const char *file_path) +{ + void *src; + size_t decomp_size, src_size; + u64 decomp_last_rem = 0; + size_t mmap_len, decomp_len = session->header.env.comp_mmap_len; + struct decomp *decomp, *decomp_last = session->active_decomp->decomp_last; + + if (decomp_last) { + decomp_last_rem = decomp_last->size - decomp_last->head; + decomp_len += decomp_last_rem; + } + + mmap_len = sizeof(struct decomp) + decomp_len; + decomp = mmap(NULL, mmap_len, PROT_READ|PROT_WRITE, + MAP_ANONYMOUS|MAP_PRIVATE, -1, 0); + if (decomp == MAP_FAILED) { + pr_err("Couldn't allocate memory for decompression\n"); + return -1; + } + + decomp->file_pos = file_offset; + decomp->file_path = file_path; + decomp->mmap_len = mmap_len; + decomp->head = 0; + + if (decomp_last_rem) { + memcpy(decomp->data, &(decomp_last->data[decomp_last->head]), decomp_last_rem); + decomp->size = decomp_last_rem; + } + + src = (void *)event + sizeof(struct perf_record_compressed); + src_size = event->pack.header.size - sizeof(struct perf_record_compressed); + + decomp_size = zstd_decompress_stream(session->active_decomp->zstd_decomp, src, src_size, + &(decomp->data[decomp_last_rem]), decomp_len - decomp_last_rem); + if (!decomp_size) { + munmap(decomp, mmap_len); + pr_err("Couldn't decompress data\n"); + return -1; + } + + decomp->size += decomp_size; + + if (session->active_decomp->decomp == NULL) + session->active_decomp->decomp = decomp; + else + session->active_decomp->decomp_last->next = decomp; + + session->active_decomp->decomp_last = decomp; + + pr_debug("decomp (B): %zd to %zd\n", src_size, decomp_size); + + return 0; +} +#endif + +static int process_event_synth_tracing_data_stub(struct perf_session *session + __maybe_unused, + union perf_event *event + __maybe_unused) +{ + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_event_synth_attr_stub(const struct perf_tool *tool __maybe_unused, + union perf_event *event __maybe_unused, + struct evlist **pevlist + __maybe_unused) +{ + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_event_synth_event_update_stub(const struct perf_tool *tool __maybe_unused, + union perf_event *event __maybe_unused, + struct evlist **pevlist + __maybe_unused) +{ + if (dump_trace) + perf_event__fprintf_event_update(event, stdout); + + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_event_sample_stub(const struct perf_tool *tool __maybe_unused, + union perf_event *event __maybe_unused, + struct perf_sample *sample __maybe_unused, + struct evsel *evsel __maybe_unused, + struct machine *machine __maybe_unused) +{ + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_event_stub(const struct perf_tool *tool __maybe_unused, + union perf_event *event __maybe_unused, + struct perf_sample *sample __maybe_unused, + struct machine *machine __maybe_unused) +{ + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_finished_round_stub(const struct perf_tool *tool __maybe_unused, + union perf_event *event __maybe_unused, + struct ordered_events *oe __maybe_unused) +{ + dump_printf(": unhandled!\n"); + return 0; +} + +static int skipn(int fd, off_t n) +{ + char buf[4096]; + ssize_t ret; + + while (n > 0) { + ret = read(fd, buf, min(n, (off_t)sizeof(buf))); + if (ret <= 0) + return ret; + n -= ret; + } + + return 0; +} + +static s64 process_event_auxtrace_stub(struct perf_session *session __maybe_unused, + union perf_event *event) +{ + dump_printf(": unhandled!\n"); + if (perf_data__is_pipe(session->data)) + skipn(perf_data__fd(session->data), event->auxtrace.size); + return event->auxtrace.size; +} + +static int process_event_op2_stub(struct perf_session *session __maybe_unused, + union perf_event *event __maybe_unused) +{ + dump_printf(": unhandled!\n"); + return 0; +} + + +static +int process_event_thread_map_stub(struct perf_session *session __maybe_unused, + union perf_event *event __maybe_unused) +{ + if (dump_trace) + perf_event__fprintf_thread_map(event, stdout); + + dump_printf(": unhandled!\n"); + return 0; +} + +static +int process_event_cpu_map_stub(struct perf_session *session __maybe_unused, + union perf_event *event __maybe_unused) +{ + if (dump_trace) + perf_event__fprintf_cpu_map(event, stdout); + + dump_printf(": unhandled!\n"); + return 0; +} + +static +int process_event_stat_config_stub(struct perf_session *session __maybe_unused, + union perf_event *event __maybe_unused) +{ + if (dump_trace) + perf_event__fprintf_stat_config(event, stdout); + + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_stat_stub(struct perf_session *perf_session __maybe_unused, + union perf_event *event) +{ + if (dump_trace) + perf_event__fprintf_stat(event, stdout); + + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_stat_round_stub(struct perf_session *perf_session __maybe_unused, + union perf_event *event) +{ + if (dump_trace) + perf_event__fprintf_stat_round(event, stdout); + + dump_printf(": unhandled!\n"); + return 0; +} + +static int process_event_time_conv_stub(struct perf_session *perf_session __maybe_unused, + union perf_event *event) +{ + if (dump_trace) + perf_event__fprintf_time_conv(event, stdout); + + dump_printf(": unhandled!\n"); + return 0; +} + +static int perf_session__process_compressed_event_stub(struct perf_session *session __maybe_unused, + union perf_event *event __maybe_unused, + u64 file_offset __maybe_unused, + const char *file_path __maybe_unused) +{ + dump_printf(": unhandled!\n"); + return 0; +} + +void perf_tool__fill_defaults(struct perf_tool *tool) +{ + if (tool->sample == NULL) + tool->sample = process_event_sample_stub; + if (tool->mmap == NULL) + tool->mmap = process_event_stub; + if (tool->mmap2 == NULL) + tool->mmap2 = process_event_stub; + if (tool->comm == NULL) + tool->comm = process_event_stub; + if (tool->namespaces == NULL) + tool->namespaces = process_event_stub; + if (tool->cgroup == NULL) + tool->cgroup = process_event_stub; + if (tool->fork == NULL) + tool->fork = process_event_stub; + if (tool->exit == NULL) + tool->exit = process_event_stub; + if (tool->lost == NULL) + tool->lost = perf_event__process_lost; + if (tool->lost_samples == NULL) + tool->lost_samples = perf_event__process_lost_samples; + if (tool->aux == NULL) + tool->aux = perf_event__process_aux; + if (tool->itrace_start == NULL) + tool->itrace_start = perf_event__process_itrace_start; + if (tool->context_switch == NULL) + tool->context_switch = perf_event__process_switch; + if (tool->ksymbol == NULL) + tool->ksymbol = perf_event__process_ksymbol; + if (tool->bpf == NULL) + tool->bpf = perf_event__process_bpf; + if (tool->text_poke == NULL) + tool->text_poke = perf_event__process_text_poke; + if (tool->aux_output_hw_id == NULL) + tool->aux_output_hw_id = perf_event__process_aux_output_hw_id; + if (tool->read == NULL) + tool->read = process_event_sample_stub; + if (tool->throttle == NULL) + tool->throttle = process_event_stub; + if (tool->unthrottle == NULL) + tool->unthrottle = process_event_stub; + if (tool->attr == NULL) + tool->attr = process_event_synth_attr_stub; + if (tool->event_update == NULL) + tool->event_update = process_event_synth_event_update_stub; + if (tool->tracing_data == NULL) + tool->tracing_data = process_event_synth_tracing_data_stub; + if (tool->build_id == NULL) + tool->build_id = process_event_op2_stub; + if (tool->finished_round == NULL) { + if (tool->ordered_events) + tool->finished_round = perf_event__process_finished_round; + else + tool->finished_round = process_finished_round_stub; + } + if (tool->id_index == NULL) + tool->id_index = process_event_op2_stub; + if (tool->auxtrace_info == NULL) + tool->auxtrace_info = process_event_op2_stub; + if (tool->auxtrace == NULL) + tool->auxtrace = process_event_auxtrace_stub; + if (tool->auxtrace_error == NULL) + tool->auxtrace_error = process_event_op2_stub; + if (tool->thread_map == NULL) + tool->thread_map = process_event_thread_map_stub; + if (tool->cpu_map == NULL) + tool->cpu_map = process_event_cpu_map_stub; + if (tool->stat_config == NULL) + tool->stat_config = process_event_stat_config_stub; + if (tool->stat == NULL) + tool->stat = process_stat_stub; + if (tool->stat_round == NULL) + tool->stat_round = process_stat_round_stub; + if (tool->time_conv == NULL) + tool->time_conv = process_event_time_conv_stub; + if (tool->feature == NULL) + tool->feature = process_event_op2_stub; + if (tool->compressed == NULL) { +#ifdef HAVE_ZSTD_SUPPORT + tool->compressed = perf_session__process_compressed_event; +#else + tool->compressed = perf_session__process_compressed_event_stub; +#endif + } + if (tool->finished_init == NULL) + tool->finished_init = process_event_op2_stub; +} + +bool perf_tool__compressed_is_stub(const struct perf_tool *tool) +{ + return tool->compressed == perf_session__process_compressed_event_stub; +} diff --git a/tools/perf/util/tool.h b/tools/perf/util/tool.h index b192d44fe91f..7913b698033a 100644 --- a/tools/perf/util/tool.h +++ b/tools/perf/util/tool.h @@ -88,4 +88,8 @@ struct perf_tool { enum show_feature_header show_feat_hdr; }; +void perf_tool__fill_defaults(struct perf_tool *tool); + +bool perf_tool__compressed_is_stub(const struct perf_tool *tool); + #endif /* __PERF_TOOL_H */ -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 05/27] perf tool: Add perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (3 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 04/27] perf tool: Move fill defaults into tool.c Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 06/27] perf kmem: Use perf_tool__init Ian Rogers ` (22 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Add init function that behaves like perf_tool__fill_defaults but assumes all values haven't been initialized. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/util/tool.c | 58 ++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/tool.h | 1 + 2 files changed, 59 insertions(+) diff --git a/tools/perf/util/tool.c b/tools/perf/util/tool.c index 17219ecb8fa6..cbd9b888bd73 100644 --- a/tools/perf/util/tool.c +++ b/tools/perf/util/tool.c @@ -230,6 +230,64 @@ static int perf_session__process_compressed_event_stub(struct perf_session *sess return 0; } +void perf_tool__init(struct perf_tool *tool, bool ordered_events) +{ + tool->ordered_events = ordered_events; + tool->ordering_requires_timestamps = false; + tool->namespace_events = false; + tool->cgroup_events = false; + tool->no_warn = false; + tool->show_feat_hdr = SHOW_FEAT_NO_HEADER; + + tool->sample = process_event_sample_stub; + tool->mmap = process_event_stub; + tool->mmap2 = process_event_stub; + tool->comm = process_event_stub; + tool->namespaces = process_event_stub; + tool->cgroup = process_event_stub; + tool->fork = process_event_stub; + tool->exit = process_event_stub; + tool->lost = perf_event__process_lost; + tool->lost_samples = perf_event__process_lost_samples; + tool->aux = perf_event__process_aux; + tool->itrace_start = perf_event__process_itrace_start; + tool->context_switch = perf_event__process_switch; + tool->ksymbol = perf_event__process_ksymbol; + tool->bpf = perf_event__process_bpf; + tool->text_poke = perf_event__process_text_poke; + tool->aux_output_hw_id = perf_event__process_aux_output_hw_id; + tool->read = process_event_sample_stub; + tool->throttle = process_event_stub; + tool->unthrottle = process_event_stub; + tool->attr = process_event_synth_attr_stub; + tool->event_update = process_event_synth_event_update_stub; + tool->tracing_data = process_event_synth_tracing_data_stub; + tool->build_id = process_event_op2_stub; + + if (ordered_events) + tool->finished_round = perf_event__process_finished_round; + else + tool->finished_round = process_finished_round_stub; + + tool->id_index = process_event_op2_stub; + tool->auxtrace_info = process_event_op2_stub; + tool->auxtrace = process_event_auxtrace_stub; + tool->auxtrace_error = process_event_op2_stub; + tool->thread_map = process_event_thread_map_stub; + tool->cpu_map = process_event_cpu_map_stub; + tool->stat_config = process_event_stat_config_stub; + tool->stat = process_stat_stub; + tool->stat_round = process_stat_round_stub; + tool->time_conv = process_event_time_conv_stub; + tool->feature = process_event_op2_stub; +#ifdef HAVE_ZSTD_SUPPORT + tool->compressed = perf_session__process_compressed_event; +#else + tool->compressed = perf_session__process_compressed_event_stub; +#endif + tool->finished_init = process_event_op2_stub; +} + void perf_tool__fill_defaults(struct perf_tool *tool) { if (tool->sample == NULL) diff --git a/tools/perf/util/tool.h b/tools/perf/util/tool.h index 7913b698033a..897c6c44b6b2 100644 --- a/tools/perf/util/tool.h +++ b/tools/perf/util/tool.h @@ -88,6 +88,7 @@ struct perf_tool { enum show_feature_header show_feat_hdr; }; +void perf_tool__init(struct perf_tool *tool, bool ordered_events); void perf_tool__fill_defaults(struct perf_tool *tool); bool perf_tool__compressed_is_stub(const struct perf_tool *tool); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 06/27] perf kmem: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (4 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 05/27] perf tool: Add perf_tool__init Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 07/27] perf buildid-list: " Ian Rogers ` (21 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Reduce the scope of the tool from global/static to just that of the cmd_kmem function where the session is scoped. Use the perf_tool__init to initialize default values. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-kmem.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 859ff018eace..b3cbac40b8c7 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -986,15 +986,6 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused, return err; } -static struct perf_tool perf_kmem = { - .sample = process_sample_event, - .comm = perf_event__process_comm, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .namespaces = perf_event__process_namespaces, - .ordered_events = true, -}; - static double fragmentation(unsigned long n_req, unsigned long n_alloc) { if (n_alloc == 0) @@ -1971,6 +1962,7 @@ int cmd_kmem(int argc, const char **argv) NULL }; struct perf_session *session; + struct perf_tool perf_kmem; static const char errmsg[] = "No %s allocation events found. Have you run 'perf kmem record --%s'?\n"; int ret = perf_config(kmem_config, NULL); @@ -1998,6 +1990,13 @@ int cmd_kmem(int argc, const char **argv) data.path = input_name; + perf_tool__init(&perf_kmem, /*ordered_events=*/true); + perf_kmem.sample = process_sample_event; + perf_kmem.comm = perf_event__process_comm; + perf_kmem.mmap = perf_event__process_mmap; + perf_kmem.mmap2 = perf_event__process_mmap2; + perf_kmem.namespaces = perf_event__process_namespaces; + kmem_session = session = perf_session__new(&data, &perf_kmem); if (IS_ERR(session)) return PTR_ERR(session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 07/27] perf buildid-list: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (5 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 06/27] perf kmem: Use perf_tool__init Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 08/27] perf kvm: " Ian Rogers ` (20 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Reduce scope of build_id__mark_dso_hit_ops to the scope of function perf_session__list_build_ids, its only use, and use perf_tool__init for the default values. Move perf_event__exit_del_thread to event.[ch] so it can be used in builtin-buildid-list.c. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-buildid-list.c | 10 ++++++++++ tools/perf/util/build-id.c | 32 ------------------------------- tools/perf/util/build-id.h | 4 +--- tools/perf/util/event.c | 20 +++++++++++++++++++ tools/perf/util/event.h | 4 ++++ 5 files changed, 35 insertions(+), 35 deletions(-) diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index 383d5de36ce4..52dfacaff8e3 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -89,6 +89,7 @@ static int perf_session__list_build_ids(bool force, bool with_hits) .mode = PERF_DATA_MODE_READ, .force = force, }; + struct perf_tool build_id__mark_dso_hit_ops; symbol__elf_init(); /* @@ -97,6 +98,15 @@ static int perf_session__list_build_ids(bool force, bool with_hits) if (filename__fprintf_build_id(input_name, stdout) > 0) goto out; + perf_tool__init(&build_id__mark_dso_hit_ops, /*ordered_events=*/true); + build_id__mark_dso_hit_ops.sample = build_id__mark_dso_hit; + build_id__mark_dso_hit_ops.mmap = perf_event__process_mmap; + build_id__mark_dso_hit_ops.mmap2 = perf_event__process_mmap2; + build_id__mark_dso_hit_ops.fork = perf_event__process_fork; + build_id__mark_dso_hit_ops.exit = perf_event__exit_del_thread; + build_id__mark_dso_hit_ops.attr = perf_event__process_attr; + build_id__mark_dso_hit_ops.build_id = perf_event__process_build_id; + session = perf_session__new(&data, &build_id__mark_dso_hit_ops); if (IS_ERR(session)) return PTR_ERR(session); diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c index 098fcc625d91..451d145fa4ed 100644 --- a/tools/perf/util/build-id.c +++ b/tools/perf/util/build-id.c @@ -67,38 +67,6 @@ int build_id__mark_dso_hit(const struct perf_tool *tool __maybe_unused, return 0; } -static int perf_event__exit_del_thread(const struct perf_tool *tool __maybe_unused, - union perf_event *event, - struct perf_sample *sample - __maybe_unused, - struct machine *machine) -{ - struct thread *thread = machine__findnew_thread(machine, - event->fork.pid, - event->fork.tid); - - dump_printf("(%d:%d):(%d:%d)\n", event->fork.pid, event->fork.tid, - event->fork.ppid, event->fork.ptid); - - if (thread) { - machine__remove_thread(machine, thread); - thread__put(thread); - } - - return 0; -} - -struct perf_tool build_id__mark_dso_hit_ops = { - .sample = build_id__mark_dso_hit, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .fork = perf_event__process_fork, - .exit = perf_event__exit_del_thread, - .attr = perf_event__process_attr, - .build_id = perf_event__process_build_id, - .ordered_events = true, -}; - int build_id__sprintf(const struct build_id *build_id, char *bf) { char *bid = bf; diff --git a/tools/perf/util/build-id.h b/tools/perf/util/build-id.h index ae87c4c58d5b..a212497bfdb0 100644 --- a/tools/perf/util/build-id.h +++ b/tools/perf/util/build-id.h @@ -16,11 +16,9 @@ struct build_id { size_t size; }; -struct nsinfo; - -extern struct perf_tool build_id__mark_dso_hit_ops; struct dso; struct feat_fd; +struct nsinfo; void build_id__init(struct build_id *bid, const u8 *data, size_t size); int build_id__sprintf(const struct build_id *build_id, char *bf); diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index c2f0e7f40ad5..aac96d5d1917 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -426,6 +426,26 @@ int perf_event__process_exit(const struct perf_tool *tool __maybe_unused, return machine__process_exit_event(machine, event, sample); } +int perf_event__exit_del_thread(const struct perf_tool *tool __maybe_unused, + union perf_event *event, + struct perf_sample *sample __maybe_unused, + struct machine *machine) +{ + struct thread *thread = machine__findnew_thread(machine, + event->fork.pid, + event->fork.tid); + + dump_printf("(%d:%d):(%d:%d)\n", event->fork.pid, event->fork.tid, + event->fork.ppid, event->fork.ptid); + + if (thread) { + machine__remove_thread(machine, thread); + thread__put(thread); + } + + return 0; +} + size_t perf_event__fprintf_aux(union perf_event *event, FILE *fp) { return fprintf(fp, " offset: %#"PRI_lx64" size: %#"PRI_lx64" flags: %#"PRI_lx64" [%s%s%s]\n", diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 4b24f1c580fd..f8742e6230a5 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -319,6 +319,10 @@ int perf_event__process_exit(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct machine *machine); +int perf_event__exit_del_thread(const struct perf_tool *tool, + union perf_event *event, + struct perf_sample *sample, + struct machine *machine); int perf_event__process_ksymbol(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 08/27] perf kvm: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (6 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 07/27] perf buildid-list: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 09/27] perf lock: " Ian Rogers ` (19 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-kvm.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index a3b903cf4311..692267b1b7e8 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -1603,19 +1603,17 @@ static int read_events(struct perf_kvm_stat *kvm) { int ret; - struct perf_tool eops = { - .sample = process_sample_event, - .comm = perf_event__process_comm, - .namespaces = perf_event__process_namespaces, - .ordered_events = true, - }; struct perf_data file = { .path = kvm->file_name, .mode = PERF_DATA_MODE_READ, .force = kvm->force, }; - kvm->tool = eops; + perf_tool__init(&kvm->tool, /*ordered_events=*/true); + kvm->tool.sample = process_sample_event; + kvm->tool.comm = perf_event__process_comm; + kvm->tool.namespaces = perf_event__process_namespaces; + kvm->session = perf_session__new(&file, &kvm->tool); if (IS_ERR(kvm->session)) { pr_err("Initializing perf session failed\n"); @@ -1919,14 +1917,13 @@ static int kvm_events_live(struct perf_kvm_stat *kvm, /* event handling */ + perf_tool__init(&kvm->tool, /*ordered_events=*/true); kvm->tool.sample = process_sample_event; kvm->tool.comm = perf_event__process_comm; kvm->tool.exit = perf_event__process_exit; kvm->tool.fork = perf_event__process_fork; kvm->tool.lost = process_lost_event; kvm->tool.namespaces = perf_event__process_namespaces; - kvm->tool.ordered_events = true; - perf_tool__fill_defaults(&kvm->tool); /* set defaults */ kvm->display_time = 1; -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 09/27] perf lock: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (7 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 08/27] perf kvm: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 10/27] perf evlist: " Ian Rogers ` (18 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-lock.c | 37 ++++++++++++++++++------------------- 1 file changed, 18 insertions(+), 19 deletions(-) diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c index 6efa9d646637..2c216427e929 100644 --- a/tools/perf/builtin-lock.c +++ b/tools/perf/builtin-lock.c @@ -1933,22 +1933,21 @@ static bool force; static int __cmd_report(bool display_info) { int err = -EINVAL; - struct perf_tool eops = { - .attr = perf_event__process_attr, - .event_update = process_event_update, - .sample = process_sample_event, - .comm = perf_event__process_comm, - .mmap = perf_event__process_mmap, - .namespaces = perf_event__process_namespaces, - .tracing_data = perf_event__process_tracing_data, - .ordered_events = true, - }; + struct perf_tool eops; struct perf_data data = { .path = input_name, .mode = PERF_DATA_MODE_READ, .force = force, }; + perf_tool__init(&eops, /*ordered_events=*/true); + eops.attr = perf_event__process_attr; + eops.event_update = process_event_update; + eops.sample = process_sample_event; + eops.comm = perf_event__process_comm; + eops.mmap = perf_event__process_mmap; + eops.namespaces = perf_event__process_namespaces; + eops.tracing_data = perf_event__process_tracing_data; session = perf_session__new(&data, &eops); if (IS_ERR(session)) { pr_err("Initializing perf session failed\n"); @@ -2069,15 +2068,7 @@ static int check_lock_contention_options(const struct option *options, static int __cmd_contention(int argc, const char **argv) { int err = -EINVAL; - struct perf_tool eops = { - .attr = perf_event__process_attr, - .event_update = process_event_update, - .sample = process_sample_event, - .comm = perf_event__process_comm, - .mmap = perf_event__process_mmap, - .tracing_data = perf_event__process_tracing_data, - .ordered_events = true, - }; + struct perf_tool eops; struct perf_data data = { .path = input_name, .mode = PERF_DATA_MODE_READ, @@ -2100,6 +2091,14 @@ static int __cmd_contention(int argc, const char **argv) con.result = &lockhash_table[0]; + perf_tool__init(&eops, /*ordered_events=*/true); + eops.attr = perf_event__process_attr; + eops.event_update = process_event_update; + eops.sample = process_sample_event; + eops.comm = perf_event__process_comm; + eops.mmap = perf_event__process_mmap; + eops.tracing_data = perf_event__process_tracing_data; + session = perf_session__new(use_bpf ? NULL : &data, &eops); if (IS_ERR(session)) { pr_err("Initializing perf session failed\n"); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 10/27] perf evlist: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (8 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 09/27] perf lock: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 11/27] perf record: " Ian Rogers ` (17 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-evlist.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/tools/perf/builtin-evlist.c b/tools/perf/builtin-evlist.c index 7117656939e7..818ab21c3f73 100644 --- a/tools/perf/builtin-evlist.c +++ b/tools/perf/builtin-evlist.c @@ -35,13 +35,13 @@ static int __cmd_evlist(const char *file_name, struct perf_attr_details *details .mode = PERF_DATA_MODE_READ, .force = details->force, }; - struct perf_tool tool = { - /* only needed for pipe mode */ - .attr = perf_event__process_attr, - .feature = process_header_feature, - }; + struct perf_tool tool; bool has_tracepoint = false; + perf_tool__init(&tool, /*ordered_events=*/false); + /* only needed for pipe mode */ + tool.attr = perf_event__process_attr; + tool.feature = process_header_feature; session = perf_session__new(&data, &tool); if (IS_ERR(session)) return PTR_ERR(session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 11/27] perf record: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (9 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 10/27] perf evlist: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 12/27] perf c2c: " Ian Rogers ` (16 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-record.c | 33 ++++++++++++++++++++------------- tools/perf/util/tool.c | 10 +++++----- tools/perf/util/tool.h | 6 ++++++ 3 files changed, 31 insertions(+), 18 deletions(-) diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index f27f725913d7..12a5170d9193 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -193,6 +193,15 @@ static const char *affinity_tags[PERF_AFFINITY_MAX] = { "SYS", "NODE", "CPU" }; +static int build_id__process_mmap(const struct perf_tool *tool, union perf_event *event, + struct perf_sample *sample, struct machine *machine); +static int build_id__process_mmap2(const struct perf_tool *tool, union perf_event *event, + struct perf_sample *sample, struct machine *machine); +static int process_timestamp_boundary(const struct perf_tool *tool, + union perf_event *event, + struct perf_sample *sample, + struct machine *machine); + #ifndef HAVE_GETTID static inline pid_t gettid(void) { @@ -1458,7 +1467,7 @@ static int process_buildids(struct record *rec) * first/last samples. */ if (rec->buildid_all && !rec->timestamp_boundary) - rec->tool.sample = NULL; + rec->tool.sample = process_event_sample_stub; return perf_session__process_events(session); } @@ -2386,6 +2395,16 @@ static int __cmd_record(struct record *rec, int argc, const char **argv) signal(SIGUSR2, SIG_IGN); } + perf_tool__init(tool, /*ordered_events=*/true); + tool->sample = process_sample_event; + tool->fork = perf_event__process_fork; + tool->exit = perf_event__process_exit; + tool->comm = perf_event__process_comm; + tool->namespaces = perf_event__process_namespaces; + tool->mmap = build_id__process_mmap; + tool->mmap2 = build_id__process_mmap2; + tool->itrace_start = process_timestamp_boundary; + tool->aux = process_timestamp_boundary; session = perf_session__new(data, tool); if (IS_ERR(session)) { pr_err("Perf session creation failed.\n"); @@ -3326,18 +3345,6 @@ static struct record record = { .ctl_fd_ack = -1, .synth = PERF_SYNTH_ALL, }, - .tool = { - .sample = process_sample_event, - .fork = perf_event__process_fork, - .exit = perf_event__process_exit, - .comm = perf_event__process_comm, - .namespaces = perf_event__process_namespaces, - .mmap = build_id__process_mmap, - .mmap2 = build_id__process_mmap2, - .itrace_start = process_timestamp_boundary, - .aux = process_timestamp_boundary, - .ordered_events = true, - }, }; const char record_callchain_help[] = CALLCHAIN_RECORD_HELP diff --git a/tools/perf/util/tool.c b/tools/perf/util/tool.c index cbd9b888bd73..a4ca16e5eefe 100644 --- a/tools/perf/util/tool.c +++ b/tools/perf/util/tool.c @@ -99,11 +99,11 @@ static int process_event_synth_event_update_stub(const struct perf_tool *tool __ return 0; } -static int process_event_sample_stub(const struct perf_tool *tool __maybe_unused, - union perf_event *event __maybe_unused, - struct perf_sample *sample __maybe_unused, - struct evsel *evsel __maybe_unused, - struct machine *machine __maybe_unused) +int process_event_sample_stub(const struct perf_tool *tool __maybe_unused, + union perf_event *event __maybe_unused, + struct perf_sample *sample __maybe_unused, + struct evsel *evsel __maybe_unused, + struct machine *machine __maybe_unused) { dump_printf(": unhandled!\n"); return 0; diff --git a/tools/perf/util/tool.h b/tools/perf/util/tool.h index 897c6c44b6b2..fb7e32d98dda 100644 --- a/tools/perf/util/tool.h +++ b/tools/perf/util/tool.h @@ -93,4 +93,10 @@ void perf_tool__fill_defaults(struct perf_tool *tool); bool perf_tool__compressed_is_stub(const struct perf_tool *tool); +int process_event_sample_stub(const struct perf_tool *tool, + union perf_event *event, + struct perf_sample *sample, + struct evsel *evsel, + struct machine *machine); + #endif /* __PERF_TOOL_H */ -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 12/27] perf c2c: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (10 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 11/27] perf record: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 13/27] perf script: " Ian Rogers ` (15 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-c2c.c | 31 +++++++++++++------------------ 1 file changed, 13 insertions(+), 18 deletions(-) diff --git a/tools/perf/builtin-c2c.c b/tools/perf/builtin-c2c.c index 88c131d05186..cd2bd573bfc3 100644 --- a/tools/perf/builtin-c2c.c +++ b/tools/perf/builtin-c2c.c @@ -385,24 +385,6 @@ static int process_sample_event(const struct perf_tool *tool __maybe_unused, goto out; } -static struct perf_c2c c2c = { - .tool = { - .sample = process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .exit = perf_event__process_exit, - .fork = perf_event__process_fork, - .lost = perf_event__process_lost, - .attr = perf_event__process_attr, - .auxtrace_info = perf_event__process_auxtrace_info, - .auxtrace = perf_event__process_auxtrace, - .auxtrace_error = perf_event__process_auxtrace_error, - .ordered_events = true, - .ordering_requires_timestamps = true, - }, -}; - static const char * const c2c_usage[] = { "perf c2c {record|report}", NULL @@ -3070,6 +3052,19 @@ static int perf_c2c__report(int argc, const char **argv) data.path = input_name; data.force = symbol_conf.force; + perf_tool__init(&c2c.tool, /*ordered_events=*/true); + c2c.tool.sample = process_sample_event; + c2c.tool.mmap = perf_event__process_mmap; + c2c.tool.mmap2 = perf_event__process_mmap2; + c2c.tool.comm = perf_event__process_comm; + c2c.tool.exit = perf_event__process_exit; + c2c.tool.fork = perf_event__process_fork; + c2c.tool.lost = perf_event__process_lost; + c2c.tool.attr = perf_event__process_attr; + c2c.tool.auxtrace_info = perf_event__process_auxtrace_info; + c2c.tool.auxtrace = perf_event__process_auxtrace; + c2c.tool.auxtrace_error = perf_event__process_auxtrace_error; + c2c.tool.ordering_requires_timestamps = true; session = perf_session__new(&data, &c2c.tool); if (IS_ERR(session)) { err = PTR_ERR(session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 13/27] perf script: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (11 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 12/27] perf c2c: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 14/27] perf inject: " Ian Rogers ` (14 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-script.c | 65 +++++++++++++++++-------------------- 1 file changed, 30 insertions(+), 35 deletions(-) diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 394bce9f5338..b4fc2971335b 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -3899,38 +3899,7 @@ int cmd_script(int argc, const char **argv) const char *dlfilter_file = NULL; const char **__argv; int i, j, err = 0; - struct perf_script script = { - .tool = { - .sample = process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .namespaces = perf_event__process_namespaces, - .cgroup = perf_event__process_cgroup, - .exit = perf_event__process_exit, - .fork = perf_event__process_fork, - .attr = process_attr, - .event_update = perf_event__process_event_update, -#ifdef HAVE_LIBTRACEEVENT - .tracing_data = perf_event__process_tracing_data, -#endif - .feature = process_feature_event, - .build_id = perf_event__process_build_id, - .id_index = perf_event__process_id_index, - .auxtrace_info = perf_script__process_auxtrace_info, - .auxtrace = perf_event__process_auxtrace, - .auxtrace_error = perf_event__process_auxtrace_error, - .stat = perf_event__process_stat_event, - .stat_round = process_stat_round_event, - .stat_config = process_stat_config_event, - .thread_map = process_thread_map_event, - .cpu_map = process_cpu_map_event, - .throttle = process_throttle_event, - .unthrottle = process_throttle_event, - .ordered_events = true, - .ordering_requires_timestamps = true, - }, - }; + struct perf_script script = {}; struct perf_data data = { .mode = PERF_DATA_MODE_READ, }; @@ -4102,10 +4071,8 @@ int cmd_script(int argc, const char **argv) data.path = input_name; data.force = symbol_conf.force; - if (unsorted_dump) { + if (unsorted_dump) dump_trace = true; - script.tool.ordered_events = false; - } if (symbol__validate_sym_arguments()) return -1; @@ -4296,6 +4263,34 @@ int cmd_script(int argc, const char **argv) use_browser = 0; } + perf_tool__init(&script.tool, !unsorted_dump); + script.tool.sample = process_sample_event; + script.tool.mmap = perf_event__process_mmap; + script.tool.mmap2 = perf_event__process_mmap2; + script.tool.comm = perf_event__process_comm; + script.tool.namespaces = perf_event__process_namespaces; + script.tool.cgroup = perf_event__process_cgroup; + script.tool.exit = perf_event__process_exit; + script.tool.fork = perf_event__process_fork; + script.tool.attr = process_attr; + script.tool.event_update = perf_event__process_event_update; +#ifdef HAVE_LIBTRACEEVENT + script.tool.tracing_data = perf_event__process_tracing_data; +#endif + script.tool.feature = process_feature_event; + script.tool.build_id = perf_event__process_build_id; + script.tool.id_index = perf_event__process_id_index; + script.tool.auxtrace_info = perf_script__process_auxtrace_info; + script.tool.auxtrace = perf_event__process_auxtrace; + script.tool.auxtrace_error = perf_event__process_auxtrace_error; + script.tool.stat = perf_event__process_stat_event; + script.tool.stat_round = process_stat_round_event; + script.tool.stat_config = process_stat_config_event; + script.tool.thread_map = process_thread_map_event; + script.tool.cpu_map = process_cpu_map_event; + script.tool.throttle = process_throttle_event; + script.tool.unthrottle = process_throttle_event; + script.tool.ordering_requires_timestamps = true; session = perf_session__new(&data, &script.tool); if (IS_ERR(session)) return PTR_ERR(session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 14/27] perf inject: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (12 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 13/27] perf script: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 15/27] perf report: " Ian Rogers ` (13 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-inject.c | 89 +++++++++++++++++-------------------- 1 file changed, 42 insertions(+), 47 deletions(-) diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index be74e2bf1f00..2866756d5060 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -2165,46 +2165,6 @@ static int __cmd_inject(struct perf_inject *inject) int cmd_inject(int argc, const char **argv) { struct perf_inject inject = { - .tool = { - .sample = perf_event__repipe_sample, - .read = perf_event__repipe_sample, - .mmap = perf_event__repipe, - .mmap2 = perf_event__repipe, - .comm = perf_event__repipe, - .namespaces = perf_event__repipe, - .cgroup = perf_event__repipe, - .fork = perf_event__repipe, - .exit = perf_event__repipe, - .lost = perf_event__repipe, - .lost_samples = perf_event__repipe, - .aux = perf_event__repipe, - .itrace_start = perf_event__repipe, - .aux_output_hw_id = perf_event__repipe, - .context_switch = perf_event__repipe, - .throttle = perf_event__repipe, - .unthrottle = perf_event__repipe, - .ksymbol = perf_event__repipe, - .bpf = perf_event__repipe, - .text_poke = perf_event__repipe, - .attr = perf_event__repipe_attr, - .event_update = perf_event__repipe_event_update, - .tracing_data = perf_event__repipe_op2_synth, - .finished_round = perf_event__repipe_oe_synth, - .build_id = perf_event__repipe_op2_synth, - .id_index = perf_event__repipe_op2_synth, - .auxtrace_info = perf_event__repipe_op2_synth, - .auxtrace_error = perf_event__repipe_op2_synth, - .time_conv = perf_event__repipe_op2_synth, - .thread_map = perf_event__repipe_op2_synth, - .cpu_map = perf_event__repipe_op2_synth, - .stat_config = perf_event__repipe_op2_synth, - .stat = perf_event__repipe_op2_synth, - .stat_round = perf_event__repipe_op2_synth, - .feature = perf_event__repipe_op2_synth, - .finished_init = perf_event__repipe_op2_synth, - .compressed = perf_event__repipe_op4_synth, - .auxtrace = perf_event__repipe_auxtrace, - }, .input_name = "-", .samples = LIST_HEAD_INIT(inject.samples), .output = { @@ -2269,6 +2229,7 @@ int cmd_inject(int argc, const char **argv) "perf inject [<options>]", NULL }; + bool ordered_events; if (!inject.itrace_synth_opts.set) { /* Disable eager loading of kernel symbols that adds overhead to perf inject. */ @@ -2333,7 +2294,47 @@ int cmd_inject(int argc, const char **argv) if (strcmp(inject.input_name, "-")) repipe = false; } - + ordered_events = inject.jit_mode || inject.sched_stat || + (inject.build_ids && !inject.build_id_all); + perf_tool__init(&inject.tool, ordered_events); + inject.tool.sample = perf_event__repipe_sample; + inject.tool.read = perf_event__repipe_sample; + inject.tool.mmap = perf_event__repipe; + inject.tool.mmap2 = perf_event__repipe; + inject.tool.comm = perf_event__repipe; + inject.tool.namespaces = perf_event__repipe; + inject.tool.cgroup = perf_event__repipe; + inject.tool.fork = perf_event__repipe; + inject.tool.exit = perf_event__repipe; + inject.tool.lost = perf_event__repipe; + inject.tool.lost_samples = perf_event__repipe; + inject.tool.aux = perf_event__repipe; + inject.tool.itrace_start = perf_event__repipe; + inject.tool.aux_output_hw_id = perf_event__repipe; + inject.tool.context_switch = perf_event__repipe; + inject.tool.throttle = perf_event__repipe; + inject.tool.unthrottle = perf_event__repipe; + inject.tool.ksymbol = perf_event__repipe; + inject.tool.bpf = perf_event__repipe; + inject.tool.text_poke = perf_event__repipe; + inject.tool.attr = perf_event__repipe_attr; + inject.tool.event_update = perf_event__repipe_event_update; + inject.tool.tracing_data = perf_event__repipe_op2_synth; + inject.tool.finished_round = perf_event__repipe_oe_synth; + inject.tool.build_id = perf_event__repipe_op2_synth; + inject.tool.id_index = perf_event__repipe_op2_synth; + inject.tool.auxtrace_info = perf_event__repipe_op2_synth; + inject.tool.auxtrace_error = perf_event__repipe_op2_synth; + inject.tool.time_conv = perf_event__repipe_op2_synth; + inject.tool.thread_map = perf_event__repipe_op2_synth; + inject.tool.cpu_map = perf_event__repipe_op2_synth; + inject.tool.stat_config = perf_event__repipe_op2_synth; + inject.tool.stat = perf_event__repipe_op2_synth; + inject.tool.stat_round = perf_event__repipe_op2_synth; + inject.tool.feature = perf_event__repipe_op2_synth; + inject.tool.finished_init = perf_event__repipe_op2_synth; + inject.tool.compressed = perf_event__repipe_op4_synth; + inject.tool.auxtrace = perf_event__repipe_auxtrace; inject.session = __perf_session__new(&data, repipe, output_fd(&inject), &inject.tool); @@ -2372,7 +2373,6 @@ int cmd_inject(int argc, const char **argv) * mmaps. We cannot generate the buildid hit list and * inject the jit mmaps at the same time for now. */ - inject.tool.ordered_events = true; inject.tool.ordering_requires_timestamps = true; if (known_build_ids != NULL) { inject.known_build_ids = @@ -2385,15 +2385,10 @@ int cmd_inject(int argc, const char **argv) } } - if (inject.sched_stat) { - inject.tool.ordered_events = true; - } - #ifdef HAVE_JITDUMP if (inject.jit_mode) { inject.tool.mmap2 = perf_event__jit_repipe_mmap2; inject.tool.mmap = perf_event__jit_repipe_mmap; - inject.tool.ordered_events = true; inject.tool.ordering_requires_timestamps = true; /* * JIT MMAP injection injects all MMAP events in one go, so it -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 15/27] perf report: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (13 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 14/27] perf inject: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 16/27] perf stat: " Ian Rogers ` (12 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-report.c | 55 ++++++++++++++++++------------------- 1 file changed, 27 insertions(+), 28 deletions(-) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index c402d9b3dca6..cd5caf1c974c 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -799,7 +799,7 @@ static int process_attr(const struct perf_tool *tool __maybe_unused, static void stats_setup(struct report *rep) { - memset(&rep->tool, 0, sizeof(rep->tool)); + perf_tool__init(&rep->tool, /*ordered_events=*/false); rep->tool.attr = process_attr; rep->tool.sample = count_sample_event; rep->tool.lost_samples = count_lost_samples_event; @@ -817,8 +817,7 @@ static int stats_print(struct report *rep) static void tasks_setup(struct report *rep) { - memset(&rep->tool, 0, sizeof(rep->tool)); - rep->tool.ordered_events = true; + perf_tool__init(&rep->tool, /*ordered_events=*/true); if (rep->mmaps_mode) { rep->tool.mmap = perf_event__process_mmap; rep->tool.mmap2 = perf_event__process_mmap2; @@ -1272,30 +1271,6 @@ int cmd_report(int argc, const char **argv) NULL }; struct report report = { - .tool = { - .sample = process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .namespaces = perf_event__process_namespaces, - .cgroup = perf_event__process_cgroup, - .exit = perf_event__process_exit, - .fork = perf_event__process_fork, - .lost = perf_event__process_lost, - .read = process_read_event, - .attr = process_attr, -#ifdef HAVE_LIBTRACEEVENT - .tracing_data = perf_event__process_tracing_data, -#endif - .build_id = perf_event__process_build_id, - .id_index = perf_event__process_id_index, - .auxtrace_info = perf_event__process_auxtrace_info, - .auxtrace = perf_event__process_auxtrace, - .event_update = perf_event__process_event_update, - .feature = process_feature_event, - .ordered_events = true, - .ordering_requires_timestamps = true, - }, .max_stack = PERF_MAX_STACK_DEPTH, .pretty_printing_style = "normal", .socket_filter = -1, @@ -1477,6 +1452,7 @@ int cmd_report(int argc, const char **argv) }; int ret = hists__init(); char sort_tmp[128]; + bool ordered_events = true; if (ret < 0) goto exit; @@ -1531,7 +1507,7 @@ int cmd_report(int argc, const char **argv) report.tasks_mode = true; if (dump_trace && report.disable_order) - report.tool.ordered_events = false; + ordered_events = false; if (quiet) perf_quiet_option(); @@ -1562,6 +1538,29 @@ int cmd_report(int argc, const char **argv) symbol_conf.skip_empty = report.skip_empty; repeat: + perf_tool__init(&report.tool, ordered_events); + report.tool.sample = process_sample_event; + report.tool.mmap = perf_event__process_mmap; + report.tool.mmap2 = perf_event__process_mmap2; + report.tool.comm = perf_event__process_comm; + report.tool.namespaces = perf_event__process_namespaces; + report.tool.cgroup = perf_event__process_cgroup; + report.tool.exit = perf_event__process_exit; + report.tool.fork = perf_event__process_fork; + report.tool.lost = perf_event__process_lost; + report.tool.read = process_read_event; + report.tool.attr = process_attr; +#ifdef HAVE_LIBTRACEEVENT + report.tool.tracing_data = perf_event__process_tracing_data; +#endif + report.tool.build_id = perf_event__process_build_id; + report.tool.id_index = perf_event__process_id_index; + report.tool.auxtrace_info = perf_event__process_auxtrace_info; + report.tool.auxtrace = perf_event__process_auxtrace; + report.tool.event_update = perf_event__process_event_update; + report.tool.feature = process_feature_event; + report.tool.ordering_requires_timestamps = true; + session = perf_session__new(&data, &report.tool); if (IS_ERR(session)) { ret = PTR_ERR(session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 16/27] perf stat: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (14 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 15/27] perf report: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 17/27] perf annotate: " Ian Rogers ` (11 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-stat.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index d460c46f15fd..a3d77a55d17f 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -2271,15 +2271,6 @@ static const char * const stat_report_usage[] = { }; static struct perf_stat perf_stat = { - .tool = { - .attr = perf_event__process_attr, - .event_update = perf_event__process_event_update, - .thread_map = process_thread_map_event, - .cpu_map = process_cpu_map_event, - .stat_config = process_stat_config_event, - .stat = perf_event__process_stat_event, - .stat_round = process_stat_round_event, - }, .aggr_mode = AGGR_UNSET, .aggr_level = 0, }; @@ -2322,6 +2313,15 @@ static int __cmd_report(int argc, const char **argv) perf_stat.data.path = input_name; perf_stat.data.mode = PERF_DATA_MODE_READ; + perf_tool__init(&perf_stat.tool, /*ordered_events=*/false); + perf_stat.tool.attr = perf_event__process_attr; + perf_stat.tool.event_update = perf_event__process_event_update; + perf_stat.tool.thread_map = process_thread_map_event; + perf_stat.tool.cpu_map = process_cpu_map_event; + perf_stat.tool.stat_config = process_stat_config_event; + perf_stat.tool.stat = perf_event__process_stat_event; + perf_stat.tool.stat_round = process_stat_round_event; + session = perf_session__new(&perf_stat.data, &perf_stat.tool); if (IS_ERR(session)) return PTR_ERR(session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 17/27] perf annotate: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (15 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 16/27] perf stat: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 18/27] perf sched: " Ian Rogers ` (10 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-annotate.c | 42 +++++++++++++++++------------------ 1 file changed, 20 insertions(+), 22 deletions(-) diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index c0685d2c8de1..598ab854b9f3 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -686,28 +686,7 @@ static const char * const annotate_usage[] = { int cmd_annotate(int argc, const char **argv) { - struct perf_annotate annotate = { - .tool = { - .sample = process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .exit = perf_event__process_exit, - .fork = perf_event__process_fork, - .namespaces = perf_event__process_namespaces, - .attr = perf_event__process_attr, - .build_id = perf_event__process_build_id, -#ifdef HAVE_LIBTRACEEVENT - .tracing_data = perf_event__process_tracing_data, -#endif - .id_index = perf_event__process_id_index, - .auxtrace_info = perf_event__process_auxtrace_info, - .auxtrace = perf_event__process_auxtrace, - .feature = process_feature_event, - .ordered_events = true, - .ordering_requires_timestamps = true, - }, - }; + struct perf_annotate annotate = {}; struct perf_data data = { .mode = PERF_DATA_MODE_READ, }; @@ -864,6 +843,25 @@ int cmd_annotate(int argc, const char **argv) data.path = input_name; + perf_tool__init(&annotate.tool, /*ordered_events=*/true); + annotate.tool.sample = process_sample_event; + annotate.tool.mmap = perf_event__process_mmap; + annotate.tool.mmap2 = perf_event__process_mmap2; + annotate.tool.comm = perf_event__process_comm; + annotate.tool.exit = perf_event__process_exit; + annotate.tool.fork = perf_event__process_fork; + annotate.tool.namespaces = perf_event__process_namespaces; + annotate.tool.attr = perf_event__process_attr; + annotate.tool.build_id = perf_event__process_build_id; +#ifdef HAVE_LIBTRACEEVENT + annotate.tool.tracing_data = perf_event__process_tracing_data; +#endif + annotate.tool.id_index = perf_event__process_id_index; + annotate.tool.auxtrace_info = perf_event__process_auxtrace_info; + annotate.tool.auxtrace = perf_event__process_auxtrace; + annotate.tool.feature = process_feature_event; + annotate.tool.ordering_requires_timestamps = true; + annotate.session = perf_session__new(&data, &annotate.tool); if (IS_ERR(annotate.session)) return PTR_ERR(annotate.session); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 18/27] perf sched: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (16 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 17/27] perf annotate: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 19/27] perf mem: " Ian Rogers ` (9 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-sched.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-) diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 2c60bd3a8149..34fe8e540c43 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -3100,7 +3100,6 @@ static int perf_sched__timehist(struct perf_sched *sched) sched->tool.tracing_data = perf_event__process_tracing_data; sched->tool.build_id = perf_event__process_build_id; - sched->tool.ordered_events = true; sched->tool.ordering_requires_timestamps = true; symbol_conf.use_callchain = sched->show_callchain; @@ -3605,14 +3604,6 @@ int cmd_sched(int argc, const char **argv) { static const char default_sort_order[] = "avg, max, switch, runtime"; struct perf_sched sched = { - .tool = { - .sample = perf_sched__process_tracepoint_sample, - .comm = perf_sched__process_comm, - .namespaces = perf_event__process_namespaces, - .lost = perf_event__process_lost, - .fork = perf_sched__process_fork_event, - .ordered_events = true, - }, .cmp_pid = LIST_HEAD_INIT(sched.cmp_pid), .sort_list = LIST_HEAD_INIT(sched.sort_list), .sort_order = default_sort_order, @@ -3733,6 +3724,13 @@ int cmd_sched(int argc, const char **argv) }; int ret; + perf_tool__init(&sched.tool, /*ordered_events=*/true); + sched.tool.sample = perf_sched__process_tracepoint_sample; + sched.tool.comm = perf_sched__process_comm; + sched.tool.namespaces = perf_event__process_namespaces; + sched.tool.lost = perf_event__process_lost; + sched.tool.fork = perf_sched__process_fork_event; + argc = parse_options_subcommand(argc, argv, sched_options, sched_subcommands, sched_usage, PARSE_OPT_STOP_AT_NON_OPTION); if (!argc) -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 19/27] perf mem: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (17 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 18/27] perf sched: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 20/27] perf timechart: " Ian Rogers ` (8 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-mem.c | 33 +++++++++++++++++---------------- 1 file changed, 17 insertions(+), 16 deletions(-) diff --git a/tools/perf/builtin-mem.c b/tools/perf/builtin-mem.c index c2038a32543c..c12ca0ec1237 100644 --- a/tools/perf/builtin-mem.c +++ b/tools/perf/builtin-mem.c @@ -285,7 +285,23 @@ static int report_raw_events(struct perf_mem *mem) .force = mem->force, }; int ret; - struct perf_session *session = perf_session__new(&data, &mem->tool); + struct perf_session *session; + + perf_tool__init(&mem->tool, /*ordered_events=*/true); + mem->tool.sample = process_sample_event; + mem->tool.mmap = perf_event__process_mmap; + mem->tool.mmap2 = perf_event__process_mmap2; + mem->tool.comm = perf_event__process_comm; + mem->tool.lost = perf_event__process_lost; + mem->tool.fork = perf_event__process_fork; + mem->tool.attr = perf_event__process_attr; + mem->tool.build_id = perf_event__process_build_id; + mem->tool.namespaces = perf_event__process_namespaces; + mem->tool.auxtrace_info = perf_event__process_auxtrace_info; + mem->tool.auxtrace = perf_event__process_auxtrace; + mem->tool.auxtrace_error = perf_event__process_auxtrace_error; + + session = perf_session__new(&data, &mem->tool); if (IS_ERR(session)) return PTR_ERR(session); @@ -449,21 +465,6 @@ int cmd_mem(int argc, const char **argv) { struct stat st; struct perf_mem mem = { - .tool = { - .sample = process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .lost = perf_event__process_lost, - .fork = perf_event__process_fork, - .attr = perf_event__process_attr, - .build_id = perf_event__process_build_id, - .namespaces = perf_event__process_namespaces, - .auxtrace_info = perf_event__process_auxtrace_info, - .auxtrace = perf_event__process_auxtrace, - .auxtrace_error = perf_event__process_auxtrace_error, - .ordered_events = true, - }, .input_name = "perf.data", /* * default to both load an store sampling -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 20/27] perf timechart: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (18 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 19/27] perf mem: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 21/27] perf diff: " Ian Rogers ` (7 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-timechart.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 5bf818baa662..218c8b44d7be 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1606,10 +1606,16 @@ static int __cmd_timechart(struct timechart *tchart, const char *output_name) .mode = PERF_DATA_MODE_READ, .force = tchart->force, }; - - struct perf_session *session = perf_session__new(&data, &tchart->tool); + struct perf_session *session; int ret = -EINVAL; + perf_tool__init(&tchart->tool, /*ordered_events=*/true); + tchart->tool.comm = process_comm_event; + tchart->tool.fork = process_fork_event; + tchart->tool.exit = process_exit_event; + tchart->tool.sample = process_sample_event; + + session = perf_session__new(&data, &tchart->tool); if (IS_ERR(session)) return PTR_ERR(session); @@ -1924,13 +1930,6 @@ parse_time(const struct option *opt, const char *arg, int __maybe_unused unset) int cmd_timechart(int argc, const char **argv) { struct timechart tchart = { - .tool = { - .comm = process_comm_event, - .fork = process_fork_event, - .exit = process_exit_event, - .sample = process_sample_event, - .ordered_events = true, - }, .proc_num = 15, .min_time = NSEC_PER_MSEC, .merge_dist = 1000, -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 21/27] perf diff: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (19 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 20/27] perf timechart: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 22/27] perf data convert json: " Ian Rogers ` (6 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-diff.c | 28 +++++++++++++--------------- 1 file changed, 13 insertions(+), 15 deletions(-) diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c index 4c0567882a7a..28c5208fcdc9 100644 --- a/tools/perf/builtin-diff.c +++ b/tools/perf/builtin-diff.c @@ -467,21 +467,7 @@ static int diff__process_sample_event(const struct perf_tool *tool, return ret; } -static struct perf_diff pdiff = { - .tool = { - .sample = diff__process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .exit = perf_event__process_exit, - .fork = perf_event__process_fork, - .lost = perf_event__process_lost, - .namespaces = perf_event__process_namespaces, - .cgroup = perf_event__process_cgroup, - .ordered_events = true, - .ordering_requires_timestamps = true, - }, -}; +static struct perf_diff pdiff; static struct evsel *evsel_match(struct evsel *evsel, struct evlist *evlist) @@ -1959,6 +1945,18 @@ int cmd_diff(int argc, const char **argv) if (ret < 0) return ret; + perf_tool__init(&pdiff.tool, /*ordered_events=*/true); + pdiff.tool.sample = diff__process_sample_event; + pdiff.tool.mmap = perf_event__process_mmap; + pdiff.tool.mmap2 = perf_event__process_mmap2; + pdiff.tool.comm = perf_event__process_comm; + pdiff.tool.exit = perf_event__process_exit; + pdiff.tool.fork = perf_event__process_fork; + pdiff.tool.lost = perf_event__process_lost; + pdiff.tool.namespaces = perf_event__process_namespaces; + pdiff.tool.cgroup = perf_event__process_cgroup; + pdiff.tool.ordering_requires_timestamps = true; + perf_config(diff__config, NULL); argc = parse_options(argc, argv, options, diff_usage, 0); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 22/27] perf data convert json: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (20 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 21/27] perf diff: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 23/27] perf data convert ctf: " Ian Rogers ` (5 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/util/data-convert-json.c | 43 ++++++++++++++--------------- 1 file changed, 20 insertions(+), 23 deletions(-) diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c index 905ea9823f9d..20bfb0884e9e 100644 --- a/tools/perf/util/data-convert-json.c +++ b/tools/perf/util/data-convert-json.c @@ -316,39 +316,36 @@ int bt_convert__perf2json(const char *input_name, const char *output_name, struct perf_session *session; int fd; int ret = -1; - struct convert_json c = { - .tool = { - .sample = process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .namespaces = perf_event__process_namespaces, - .cgroup = perf_event__process_cgroup, - .exit = perf_event__process_exit, - .fork = perf_event__process_fork, - .lost = perf_event__process_lost, -#ifdef HAVE_LIBTRACEEVENT - .tracing_data = perf_event__process_tracing_data, -#endif - .build_id = perf_event__process_build_id, - .id_index = perf_event__process_id_index, - .auxtrace_info = perf_event__process_auxtrace_info, - .auxtrace = perf_event__process_auxtrace, - .event_update = perf_event__process_event_update, - .ordered_events = true, - .ordering_requires_timestamps = true, - }, .first = true, .events_count = 0, }; - struct perf_data data = { .mode = PERF_DATA_MODE_READ, .path = input_name, .force = opts->force, }; + perf_tool__init(&c.tool, /*ordered_events=*/true); + c.tool.sample = process_sample_event; + c.tool.mmap = perf_event__process_mmap; + c.tool.mmap2 = perf_event__process_mmap2; + c.tool.comm = perf_event__process_comm; + c.tool.namespaces = perf_event__process_namespaces; + c.tool.cgroup = perf_event__process_cgroup; + c.tool.exit = perf_event__process_exit; + c.tool.fork = perf_event__process_fork; + c.tool.lost = perf_event__process_lost; +#ifdef HAVE_LIBTRACEEVENT + c.tool.tracing_data = perf_event__process_tracing_data; +#endif + c.tool.build_id = perf_event__process_build_id; + c.tool.id_index = perf_event__process_id_index; + c.tool.auxtrace_info = perf_event__process_auxtrace_info; + c.tool.auxtrace = perf_event__process_auxtrace; + c.tool.event_update = perf_event__process_event_update; + c.tool.ordering_requires_timestamps = true; + if (opts->all) { pr_err("--all is currently unsupported for JSON output.\n"); goto err; -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 23/27] perf data convert ctf: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (21 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 22/27] perf data convert json: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 24/27] perf test event_update: Ensure tools is initialized Ian Rogers ` (4 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/util/data-convert-bt.c | 30 ++++++++++++++---------------- 1 file changed, 14 insertions(+), 16 deletions(-) diff --git a/tools/perf/util/data-convert-bt.c b/tools/perf/util/data-convert-bt.c index 9e2170604b66..021e9b1d5cc5 100644 --- a/tools/perf/util/data-convert-bt.c +++ b/tools/perf/util/data-convert-bt.c @@ -1607,25 +1607,23 @@ int bt_convert__perf2ctf(const char *input, const char *path, .mode = PERF_DATA_MODE_READ, .force = opts->force, }; - struct convert c = { - .tool = { - .sample = process_sample_event, - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .comm = perf_event__process_comm, - .exit = perf_event__process_exit, - .fork = perf_event__process_fork, - .lost = perf_event__process_lost, - .tracing_data = perf_event__process_tracing_data, - .build_id = perf_event__process_build_id, - .namespaces = perf_event__process_namespaces, - .ordered_events = true, - .ordering_requires_timestamps = true, - }, - }; + struct convert c = {}; struct ctf_writer *cw = &c.writer; int err; + perf_tool__init(&c.tool, /*ordered_events=*/true); + c.tool.sample = process_sample_event; + c.tool.mmap = perf_event__process_mmap; + c.tool.mmap2 = perf_event__process_mmap2; + c.tool.comm = perf_event__process_comm; + c.tool.exit = perf_event__process_exit; + c.tool.fork = perf_event__process_fork; + c.tool.lost = perf_event__process_lost; + c.tool.tracing_data = perf_event__process_tracing_data; + c.tool.build_id = perf_event__process_build_id; + c.tool.namespaces = perf_event__process_namespaces; + c.tool.ordering_requires_timestamps = true; + if (opts->all) { c.tool.comm = process_comm_event; c.tool.exit = process_exit_event; -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 24/27] perf test event_update: Ensure tools is initialized 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (22 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 23/27] perf data convert ctf: " Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 25/27] perf kwork: Use perf_tool__init Ian Rogers ` (3 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Ensure tool is initialized to avoid lazy initialization pattern so that more uses of struct perf_tool can be made const. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/tests/event_update.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/perf/tests/event_update.c b/tools/perf/tests/event_update.c index fdecad920f59..d6b4ce3ef4ee 100644 --- a/tools/perf/tests/event_update.c +++ b/tools/perf/tests/event_update.c @@ -103,6 +103,7 @@ static int test__event_update(struct test_suite *test __maybe_unused, int subtes TEST_ASSERT_VAL("failed to synthesize attr update scale", !perf_event__synthesize_event_update_scale(NULL, evsel, process_event_scale)); + perf_tool__init(&tmp.tool, /*ordered_events=*/false); tmp.name = evsel__name(evsel); TEST_ASSERT_VAL("failed to synthesize attr update name", -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 25/27] perf kwork: Use perf_tool__init 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (23 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 24/27] perf test event_update: Ensure tools is initialized Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 26/27] perf tool: Remove perf_tool__fill_defaults Ian Rogers ` (2 subsequent siblings) 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Use perf_tool__init so that more uses of struct perf_tool can be const and not relying on perf_tool__fill_defaults. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/builtin-kwork.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/tools/perf/builtin-kwork.c b/tools/perf/builtin-kwork.c index 8ffaa80a2d1d..6a4281b8fd10 100644 --- a/tools/perf/builtin-kwork.c +++ b/tools/perf/builtin-kwork.c @@ -2322,12 +2322,6 @@ int cmd_kwork(int argc, const char **argv) { static struct perf_kwork kwork = { .class_list = LIST_HEAD_INIT(kwork.class_list), - .tool = { - .mmap = perf_event__process_mmap, - .mmap2 = perf_event__process_mmap2, - .sample = perf_kwork__process_tracepoint_sample, - .ordered_events = true, - }, .atom_page_list = LIST_HEAD_INIT(kwork.atom_page_list), .sort_list = LIST_HEAD_INIT(kwork.sort_list), .cmp_id = LIST_HEAD_INIT(kwork.cmp_id), @@ -2462,6 +2456,11 @@ int cmd_kwork(int argc, const char **argv) "record", "report", "latency", "timehist", "top", NULL }; + perf_tool__init(&kwork.tool, /*ordered_events=*/true); + kwork.tool.mmap = perf_event__process_mmap; + kwork.tool.mmap2 = perf_event__process_mmap2; + kwork.tool.sample = perf_kwork__process_tracepoint_sample; + argc = parse_options_subcommand(argc, argv, kwork_options, kwork_subcommands, kwork_usage, PARSE_OPT_STOP_AT_NON_OPTION); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 26/27] perf tool: Remove perf_tool__fill_defaults 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (24 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 25/27] perf kwork: Use perf_tool__init Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-18 1:00 ` [PATCH v6 27/27] perf session: Constify tool Ian Rogers 2024-07-19 8:50 ` [PATCH v6 00/27] Constify tool pointers Adrian Hunter 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Now all tools are fully initialized prior to use it has no use so remove. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/util/session.c | 6 --- tools/perf/util/tool.c | 89 --------------------------------------- tools/perf/util/tool.h | 1 - 3 files changed, 96 deletions(-) diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 5359ee055c85..ad5da61d2330 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1776,8 +1776,6 @@ static int __perf_session__process_pipe_events(struct perf_session *session) void *p; bool update_prog = false; - perf_tool__fill_defaults(tool); - /* * If it's from a file saving pipe data (by redirection), it would have * a file name other than "-". Then we can get the total size and show @@ -2201,8 +2199,6 @@ static int __perf_session__process_events(struct perf_session *session) struct ui_progress prog; int err; - perf_tool__fill_defaults(tool); - if (rd.data_size == 0) return -1; @@ -2255,8 +2251,6 @@ static int __perf_session__process_dir_events(struct perf_session *session) u64 total_size = perf_data__size(session->data); struct reader *rd; - perf_tool__fill_defaults(tool); - ui_progress__init_size(&prog, total_size, "Processing events..."); nr_readers = 1; diff --git a/tools/perf/util/tool.c b/tools/perf/util/tool.c index a4ca16e5eefe..3b7f390f26eb 100644 --- a/tools/perf/util/tool.c +++ b/tools/perf/util/tool.c @@ -288,95 +288,6 @@ void perf_tool__init(struct perf_tool *tool, bool ordered_events) tool->finished_init = process_event_op2_stub; } -void perf_tool__fill_defaults(struct perf_tool *tool) -{ - if (tool->sample == NULL) - tool->sample = process_event_sample_stub; - if (tool->mmap == NULL) - tool->mmap = process_event_stub; - if (tool->mmap2 == NULL) - tool->mmap2 = process_event_stub; - if (tool->comm == NULL) - tool->comm = process_event_stub; - if (tool->namespaces == NULL) - tool->namespaces = process_event_stub; - if (tool->cgroup == NULL) - tool->cgroup = process_event_stub; - if (tool->fork == NULL) - tool->fork = process_event_stub; - if (tool->exit == NULL) - tool->exit = process_event_stub; - if (tool->lost == NULL) - tool->lost = perf_event__process_lost; - if (tool->lost_samples == NULL) - tool->lost_samples = perf_event__process_lost_samples; - if (tool->aux == NULL) - tool->aux = perf_event__process_aux; - if (tool->itrace_start == NULL) - tool->itrace_start = perf_event__process_itrace_start; - if (tool->context_switch == NULL) - tool->context_switch = perf_event__process_switch; - if (tool->ksymbol == NULL) - tool->ksymbol = perf_event__process_ksymbol; - if (tool->bpf == NULL) - tool->bpf = perf_event__process_bpf; - if (tool->text_poke == NULL) - tool->text_poke = perf_event__process_text_poke; - if (tool->aux_output_hw_id == NULL) - tool->aux_output_hw_id = perf_event__process_aux_output_hw_id; - if (tool->read == NULL) - tool->read = process_event_sample_stub; - if (tool->throttle == NULL) - tool->throttle = process_event_stub; - if (tool->unthrottle == NULL) - tool->unthrottle = process_event_stub; - if (tool->attr == NULL) - tool->attr = process_event_synth_attr_stub; - if (tool->event_update == NULL) - tool->event_update = process_event_synth_event_update_stub; - if (tool->tracing_data == NULL) - tool->tracing_data = process_event_synth_tracing_data_stub; - if (tool->build_id == NULL) - tool->build_id = process_event_op2_stub; - if (tool->finished_round == NULL) { - if (tool->ordered_events) - tool->finished_round = perf_event__process_finished_round; - else - tool->finished_round = process_finished_round_stub; - } - if (tool->id_index == NULL) - tool->id_index = process_event_op2_stub; - if (tool->auxtrace_info == NULL) - tool->auxtrace_info = process_event_op2_stub; - if (tool->auxtrace == NULL) - tool->auxtrace = process_event_auxtrace_stub; - if (tool->auxtrace_error == NULL) - tool->auxtrace_error = process_event_op2_stub; - if (tool->thread_map == NULL) - tool->thread_map = process_event_thread_map_stub; - if (tool->cpu_map == NULL) - tool->cpu_map = process_event_cpu_map_stub; - if (tool->stat_config == NULL) - tool->stat_config = process_event_stat_config_stub; - if (tool->stat == NULL) - tool->stat = process_stat_stub; - if (tool->stat_round == NULL) - tool->stat_round = process_stat_round_stub; - if (tool->time_conv == NULL) - tool->time_conv = process_event_time_conv_stub; - if (tool->feature == NULL) - tool->feature = process_event_op2_stub; - if (tool->compressed == NULL) { -#ifdef HAVE_ZSTD_SUPPORT - tool->compressed = perf_session__process_compressed_event; -#else - tool->compressed = perf_session__process_compressed_event_stub; -#endif - } - if (tool->finished_init == NULL) - tool->finished_init = process_event_op2_stub; -} - bool perf_tool__compressed_is_stub(const struct perf_tool *tool) { return tool->compressed == perf_session__process_compressed_event_stub; diff --git a/tools/perf/util/tool.h b/tools/perf/util/tool.h index fb7e32d98dda..0874e7924687 100644 --- a/tools/perf/util/tool.h +++ b/tools/perf/util/tool.h @@ -89,7 +89,6 @@ struct perf_tool { }; void perf_tool__init(struct perf_tool *tool, bool ordered_events); -void perf_tool__fill_defaults(struct perf_tool *tool); bool perf_tool__compressed_is_stub(const struct perf_tool *tool); -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* [PATCH v6 27/27] perf session: Constify tool 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (25 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 26/27] perf tool: Remove perf_tool__fill_defaults Ian Rogers @ 2024-07-18 1:00 ` Ian Rogers 2024-07-19 8:50 ` [PATCH v6 00/27] Constify tool pointers Adrian Hunter 27 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-18 1:00 UTC (permalink / raw) To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Make tool const now that all uses are const and perf_tool__fill_defaults won't be used. The aim is to better capture that sessions don't mutate tools. Signed-off-by: Ian Rogers <irogers@google.com> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> --- tools/perf/util/session.c | 6 +++--- tools/perf/util/session.h | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index ad5da61d2330..0c7cc092887d 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -1765,7 +1765,7 @@ static int __perf_session__process_decomp_events(struct perf_session *session); static int __perf_session__process_pipe_events(struct perf_session *session) { struct ordered_events *oe = &session->ordered_events; - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct ui_progress prog; union perf_event *event; uint32_t size, cur_size = 0; @@ -2195,7 +2195,7 @@ static int __perf_session__process_events(struct perf_session *session) .in_place_update = session->data->in_place_update, }; struct ordered_events *oe = &session->ordered_events; - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; struct ui_progress prog; int err; @@ -2245,7 +2245,7 @@ static int __perf_session__process_events(struct perf_session *session) static int __perf_session__process_dir_events(struct perf_session *session) { struct perf_data *data = session->data; - struct perf_tool *tool = session->tool; + const struct perf_tool *tool = session->tool; int i, ret, readers, nr_readers; struct ui_progress prog; u64 total_size = perf_data__size(session->data); diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 7f69baeae7fb..7c8dd6956330 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -43,7 +43,7 @@ struct perf_session { u64 one_mmap_offset; struct ordered_events ordered_events; struct perf_data *data; - struct perf_tool *tool; + const struct perf_tool *tool; u64 bytes_transferred; u64 bytes_compressed; struct zstd_data zstd_data; -- 2.45.2.1089.g2a221341d9-goog ^ permalink raw reply related [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers ` (26 preceding siblings ...) 2024-07-18 1:00 ` [PATCH v6 27/27] perf session: Constify tool Ian Rogers @ 2024-07-19 8:50 ` Adrian Hunter 2024-07-19 16:26 ` Ian Rogers 27 siblings, 1 reply; 40+ messages in thread From: Adrian Hunter @ 2024-07-19 8:50 UTC (permalink / raw) To: Ian Rogers, Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On 18/07/24 03:59, Ian Rogers wrote: > struct perf_tool provides a set of function pointers that are called > through when processing perf data. To make filling the pointers less > cumbersome, if they are NULL perf_tools__fill_defaults will add > default do nothing implementations. > > This change refactors struct perf_tool to have an init function that > provides the default implementation. The special use of NULL and > perf_tools__fill_defaults are removed. As a consequence the tool > pointers can then all be made const, which better reflects the > behavior a particular perf command would expect of the tool and to > some extent can reduce the cognitive load on someone working on a > command. > > v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. The tags were really meant only for patch 1, the email that was replied to. But now for patches 2 and 3: Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> Looking at patches 4 to 25, they do not seem to offer any benefit. Instead of patch 26, presumably perf_tool__fill_defaults() could be moved to __perf_session__new(), which perhaps would allow patch 27 as it is. > v5: Rebase dropping asan fix merged by Namhyung. > v4: Simplify perf_session__deliver_synth_attr_event following Adrian's > suggestions. > v3: Just remove auxtrace dummy tools [Adrian] and make s390-cpumsf > struct removal its own patch [Adrian]. > v2: Remove dummy tool initialization [Adrian] and make zero sized. Add > cs-etm fix for address sanitizer build, found necessary when > testing dummy tool change. > > Ian Rogers (27): > perf auxtrace: Remove dummy tools > perf s390-cpumsf: Remove unused struct > perf tool: Constify tool pointers > perf tool: Move fill defaults into tool.c > perf tool: Add perf_tool__init > perf kmem: Use perf_tool__init > perf buildid-list: Use perf_tool__init > perf kvm: Use perf_tool__init > perf lock: Use perf_tool__init > perf evlist: Use perf_tool__init > perf record: Use perf_tool__init > perf c2c: Use perf_tool__init > perf script: Use perf_tool__init > perf inject: Use perf_tool__init > perf report: Use perf_tool__init > perf stat: Use perf_tool__init > perf annotate: Use perf_tool__init > perf sched: Use perf_tool__init > perf mem: Use perf_tool__init > perf timechart: Use perf_tool__init > perf diff: Use perf_tool__init > perf data convert json: Use perf_tool__init > perf data convert ctf: Use perf_tool__init > perf test event_update: Ensure tools is initialized > perf kwork: Use perf_tool__init > perf tool: Remove perf_tool__fill_defaults > perf session: Constify tool > > tools/perf/arch/x86/util/event.c | 4 +- > tools/perf/bench/synthesize.c | 2 +- > tools/perf/builtin-annotate.c | 44 ++-- > tools/perf/builtin-buildid-list.c | 10 + > tools/perf/builtin-c2c.c | 33 ++- > tools/perf/builtin-diff.c | 30 ++- > tools/perf/builtin-evlist.c | 10 +- > tools/perf/builtin-inject.c | 159 ++++++------ > tools/perf/builtin-kmem.c | 20 +- > tools/perf/builtin-kvm.c | 19 +- > tools/perf/builtin-kwork.c | 33 ++- > tools/perf/builtin-lock.c | 41 ++-- > tools/perf/builtin-mem.c | 37 +-- > tools/perf/builtin-record.c | 47 ++-- > tools/perf/builtin-report.c | 67 +++-- > tools/perf/builtin-sched.c | 50 ++-- > tools/perf/builtin-script.c | 106 ++++---- > tools/perf/builtin-stat.c | 26 +- > tools/perf/builtin-timechart.c | 25 +- > tools/perf/builtin-top.c | 2 +- > tools/perf/builtin-trace.c | 4 +- > tools/perf/tests/cpumap.c | 6 +- > tools/perf/tests/dlfilter-test.c | 2 +- > tools/perf/tests/dwarf-unwind.c | 2 +- > tools/perf/tests/event_update.c | 9 +- > tools/perf/tests/stat.c | 6 +- > tools/perf/tests/thread-map.c | 2 +- > tools/perf/util/Build | 1 + > tools/perf/util/arm-spe.c | 55 +---- > tools/perf/util/auxtrace.c | 12 +- > tools/perf/util/auxtrace.h | 20 +- > tools/perf/util/bpf-event.c | 4 +- > tools/perf/util/build-id.c | 34 +-- > tools/perf/util/build-id.h | 8 +- > tools/perf/util/cs-etm.c | 39 +-- > tools/perf/util/data-convert-bt.c | 34 ++- > tools/perf/util/data-convert-json.c | 47 ++-- > tools/perf/util/event.c | 54 ++-- > tools/perf/util/event.h | 38 +-- > tools/perf/util/header.c | 6 +- > tools/perf/util/header.h | 4 +- > tools/perf/util/hisi-ptt.c | 6 +- > tools/perf/util/intel-bts.c | 37 +-- > tools/perf/util/intel-pt.c | 30 +-- > tools/perf/util/jitdump.c | 4 +- > tools/perf/util/s390-cpumsf.c | 11 +- > tools/perf/util/session.c | 366 +++------------------------- > tools/perf/util/session.h | 9 +- > tools/perf/util/synthetic-events.c | 80 +++--- > tools/perf/util/synthetic-events.h | 70 +++--- > tools/perf/util/tool.c | 294 ++++++++++++++++++++++ > tools/perf/util/tool.h | 18 +- > tools/perf/util/tsc.c | 2 +- > 53 files changed, 977 insertions(+), 1102 deletions(-) > create mode 100644 tools/perf/util/tool.c > ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-19 8:50 ` [PATCH v6 00/27] Constify tool pointers Adrian Hunter @ 2024-07-19 16:26 ` Ian Rogers 2024-07-22 8:28 ` Adrian Hunter 2024-08-12 13:53 ` Arnaldo Carvalho de Melo 0 siblings, 2 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-19 16:26 UTC (permalink / raw) To: Adrian Hunter Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > On 18/07/24 03:59, Ian Rogers wrote: > > struct perf_tool provides a set of function pointers that are called > > through when processing perf data. To make filling the pointers less > > cumbersome, if they are NULL perf_tools__fill_defaults will add > > default do nothing implementations. > > > > This change refactors struct perf_tool to have an init function that > > provides the default implementation. The special use of NULL and > > perf_tools__fill_defaults are removed. As a consequence the tool > > pointers can then all be made const, which better reflects the > > behavior a particular perf command would expect of the tool and to > > some extent can reduce the cognitive load on someone working on a > > command. > > > > v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > > The tags were really meant only for patch 1, the email that was replied to. > > But now for patches 2 and 3: > > Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> Sorry for that, you'd mentioned that pt and bts testing which is impacted by more than just patch 1. > Looking at patches 4 to 25, they do not seem to offer any benefit. > > Instead of patch 26, presumably perf_tool__fill_defaults() could > be moved to __perf_session__new(), which perhaps would allow > patch 27 as it is. What I'm trying to do in the series is make it so that the tool isn't mutated during its use by session. Ideally we'd be passing a const tool to session_new, that's not possible because there's a hack to fix ordered events and pipe mode in session__new: https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 Imo, it isn't great to pass a tool to session__new where you say you want ordered events and then session just goes to change that for you. Altering that behavior was beyond the scope of this clean up, so tool is only const after session__new. The reason for doing this is to make it so that when I have a tool I can reason that nobody is doing things to change it under my feet. My builtin_cmd is in charge of what the tool is rather than some code buried in util that thought it was going to do me a favor. The code is a refactor and so the benefit is intended to be for the developer and how they reason about the use of tool. We generally use _init functions rather than having _fill_defaults, so there is a consistency argument. I don't expect any impact in terms of performance... Moving perf_tool__fill_defaults to __perf_session__new had issues with the existing code where NULL would be written over a function pointer expecting the later fill_defaults to fix it up, doesn't address coding consistency where _init is the norm, and adds another reason the tool passed to session__new can't be const. Thanks, Ian > > v5: Rebase dropping asan fix merged by Namhyung. > > v4: Simplify perf_session__deliver_synth_attr_event following Adrian's > > suggestions. > > v3: Just remove auxtrace dummy tools [Adrian] and make s390-cpumsf > > struct removal its own patch [Adrian]. > > v2: Remove dummy tool initialization [Adrian] and make zero sized. Add > > cs-etm fix for address sanitizer build, found necessary when > > testing dummy tool change. > > > > Ian Rogers (27): > > perf auxtrace: Remove dummy tools > > perf s390-cpumsf: Remove unused struct > > perf tool: Constify tool pointers > > perf tool: Move fill defaults into tool.c > > perf tool: Add perf_tool__init > > perf kmem: Use perf_tool__init > > perf buildid-list: Use perf_tool__init > > perf kvm: Use perf_tool__init > > perf lock: Use perf_tool__init > > perf evlist: Use perf_tool__init > > perf record: Use perf_tool__init > > perf c2c: Use perf_tool__init > > perf script: Use perf_tool__init > > perf inject: Use perf_tool__init > > perf report: Use perf_tool__init > > perf stat: Use perf_tool__init > > perf annotate: Use perf_tool__init > > perf sched: Use perf_tool__init > > perf mem: Use perf_tool__init > > perf timechart: Use perf_tool__init > > perf diff: Use perf_tool__init > > perf data convert json: Use perf_tool__init > > perf data convert ctf: Use perf_tool__init > > perf test event_update: Ensure tools is initialized > > perf kwork: Use perf_tool__init > > perf tool: Remove perf_tool__fill_defaults > > perf session: Constify tool > > > > tools/perf/arch/x86/util/event.c | 4 +- > > tools/perf/bench/synthesize.c | 2 +- > > tools/perf/builtin-annotate.c | 44 ++-- > > tools/perf/builtin-buildid-list.c | 10 + > > tools/perf/builtin-c2c.c | 33 ++- > > tools/perf/builtin-diff.c | 30 ++- > > tools/perf/builtin-evlist.c | 10 +- > > tools/perf/builtin-inject.c | 159 ++++++------ > > tools/perf/builtin-kmem.c | 20 +- > > tools/perf/builtin-kvm.c | 19 +- > > tools/perf/builtin-kwork.c | 33 ++- > > tools/perf/builtin-lock.c | 41 ++-- > > tools/perf/builtin-mem.c | 37 +-- > > tools/perf/builtin-record.c | 47 ++-- > > tools/perf/builtin-report.c | 67 +++-- > > tools/perf/builtin-sched.c | 50 ++-- > > tools/perf/builtin-script.c | 106 ++++---- > > tools/perf/builtin-stat.c | 26 +- > > tools/perf/builtin-timechart.c | 25 +- > > tools/perf/builtin-top.c | 2 +- > > tools/perf/builtin-trace.c | 4 +- > > tools/perf/tests/cpumap.c | 6 +- > > tools/perf/tests/dlfilter-test.c | 2 +- > > tools/perf/tests/dwarf-unwind.c | 2 +- > > tools/perf/tests/event_update.c | 9 +- > > tools/perf/tests/stat.c | 6 +- > > tools/perf/tests/thread-map.c | 2 +- > > tools/perf/util/Build | 1 + > > tools/perf/util/arm-spe.c | 55 +---- > > tools/perf/util/auxtrace.c | 12 +- > > tools/perf/util/auxtrace.h | 20 +- > > tools/perf/util/bpf-event.c | 4 +- > > tools/perf/util/build-id.c | 34 +-- > > tools/perf/util/build-id.h | 8 +- > > tools/perf/util/cs-etm.c | 39 +-- > > tools/perf/util/data-convert-bt.c | 34 ++- > > tools/perf/util/data-convert-json.c | 47 ++-- > > tools/perf/util/event.c | 54 ++-- > > tools/perf/util/event.h | 38 +-- > > tools/perf/util/header.c | 6 +- > > tools/perf/util/header.h | 4 +- > > tools/perf/util/hisi-ptt.c | 6 +- > > tools/perf/util/intel-bts.c | 37 +-- > > tools/perf/util/intel-pt.c | 30 +-- > > tools/perf/util/jitdump.c | 4 +- > > tools/perf/util/s390-cpumsf.c | 11 +- > > tools/perf/util/session.c | 366 +++------------------------- > > tools/perf/util/session.h | 9 +- > > tools/perf/util/synthetic-events.c | 80 +++--- > > tools/perf/util/synthetic-events.h | 70 +++--- > > tools/perf/util/tool.c | 294 ++++++++++++++++++++++ > > tools/perf/util/tool.h | 18 +- > > tools/perf/util/tsc.c | 2 +- > > 53 files changed, 977 insertions(+), 1102 deletions(-) > > create mode 100644 tools/perf/util/tool.c > > > ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-19 16:26 ` Ian Rogers @ 2024-07-22 8:28 ` Adrian Hunter 2024-07-22 16:06 ` Ian Rogers 2024-08-12 13:53 ` Arnaldo Carvalho de Melo 1 sibling, 1 reply; 40+ messages in thread From: Adrian Hunter @ 2024-07-22 8:28 UTC (permalink / raw) To: Ian Rogers Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On 19/07/24 19:26, Ian Rogers wrote: > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: >> >> On 18/07/24 03:59, Ian Rogers wrote: >>> struct perf_tool provides a set of function pointers that are called >>> through when processing perf data. To make filling the pointers less >>> cumbersome, if they are NULL perf_tools__fill_defaults will add >>> default do nothing implementations. >>> >>> This change refactors struct perf_tool to have an init function that >>> provides the default implementation. The special use of NULL and >>> perf_tools__fill_defaults are removed. As a consequence the tool >>> pointers can then all be made const, which better reflects the >>> behavior a particular perf command would expect of the tool and to >>> some extent can reduce the cognitive load on someone working on a >>> command. >>> >>> v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. >> >> The tags were really meant only for patch 1, the email that was replied to. >> >> But now for patches 2 and 3: >> >> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > > Sorry for that, you'd mentioned that pt and bts testing which is > impacted by more than just patch 1. > >> Looking at patches 4 to 25, they do not seem to offer any benefit. >> >> Instead of patch 26, presumably perf_tool__fill_defaults() could >> be moved to __perf_session__new(), which perhaps would allow >> patch 27 as it is. > > What I'm trying to do in the series is make it so that the tool isn't > mutated during its use by session. Ideally we'd be passing a const > tool to session_new, that's not possible because there's a hack to fix > ordered events and pipe mode in session__new: > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 > Imo, it isn't great to pass a tool to session__new where you say you > want ordered events and then session just goes to change that for you. > Altering that behavior was beyond the scope of this clean up, so tool > is only const after session__new. Seems like a separate issue. Since the session is created by __perf_session__new(), session->tool will always be a pointer to a const tool once there is: diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 7f69baeae7fb..7c8dd6956330 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -43,7 +43,7 @@ struct perf_session { u64 one_mmap_offset; struct ordered_events ordered_events; struct perf_data *data; - struct perf_tool *tool; + const struct perf_tool *tool; u64 bytes_transferred; u64 bytes_compressed; struct zstd_data zstd_data; > > The reason for doing this is to make it so that when I have a tool I > can reason that nobody is doing things to change it under my feet. It still can be changed by the caller of __perf_session__new(), since the tool itself is not const. Anything using container_of() like: static int process_sample_event(const struct perf_tool *tool, union perf_event *event, struct perf_sample *sample, struct evsel *evsel, struct machine *machine) { struct perf_script *scr = container_of(tool, struct perf_script, tool); can then change scr->tool without even having to cast away const. Really, 'tool' needs to be defined as const in the first place. > My > builtin_cmd is in charge of what the tool is rather than some code > buried in util that thought it was going to do me a favor. The code is > a refactor and so the benefit is intended to be for the developer and > how they reason about the use of tool. It creates another question though: since there is a lot of code before perf_tool__init() is called, does the caller mistakenly change tool before calling perf_tool__init() > how they reason about the use of tool. We generally use _init > functions rather than having _fill_defaults, so there is a consistency > argument. The caller does not need the "defaults", so why would it set them up. The session could just as easily do: if (tool->cb) tool->cb(...); else cb_stub(...); > I don't expect any impact in terms of performance... Moving > perf_tool__fill_defaults to __perf_session__new had issues with the > existing code where NULL would be written over a function pointer > expecting the later fill_defaults to fix it up, doesn't address coding > consistency where _init is the norm, and adds another reason the tool > passed to session__new can't be const. perf_tool__init() is not a steeping stone to making 'tool' a const in the first place. ^ permalink raw reply related [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-22 8:28 ` Adrian Hunter @ 2024-07-22 16:06 ` Ian Rogers 2024-07-22 17:45 ` Namhyung Kim 0 siblings, 1 reply; 40+ messages in thread From: Ian Rogers @ 2024-07-22 16:06 UTC (permalink / raw) To: Adrian Hunter Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Mon, Jul 22, 2024 at 1:29 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > On 19/07/24 19:26, Ian Rogers wrote: > > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > >> > >> On 18/07/24 03:59, Ian Rogers wrote: > >>> struct perf_tool provides a set of function pointers that are called > >>> through when processing perf data. To make filling the pointers less > >>> cumbersome, if they are NULL perf_tools__fill_defaults will add > >>> default do nothing implementations. > >>> > >>> This change refactors struct perf_tool to have an init function that > >>> provides the default implementation. The special use of NULL and > >>> perf_tools__fill_defaults are removed. As a consequence the tool > >>> pointers can then all be made const, which better reflects the > >>> behavior a particular perf command would expect of the tool and to > >>> some extent can reduce the cognitive load on someone working on a > >>> command. > >>> > >>> v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > >> > >> The tags were really meant only for patch 1, the email that was replied to. > >> > >> But now for patches 2 and 3: > >> > >> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > > > > Sorry for that, you'd mentioned that pt and bts testing which is > > impacted by more than just patch 1. > > > >> Looking at patches 4 to 25, they do not seem to offer any benefit. > >> > >> Instead of patch 26, presumably perf_tool__fill_defaults() could > >> be moved to __perf_session__new(), which perhaps would allow > >> patch 27 as it is. > > > > What I'm trying to do in the series is make it so that the tool isn't > > mutated during its use by session. Ideally we'd be passing a const > > tool to session_new, that's not possible because there's a hack to fix > > ordered events and pipe mode in session__new: > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 > > Imo, it isn't great to pass a tool to session__new where you say you > > want ordered events and then session just goes to change that for you. > > Altering that behavior was beyond the scope of this clean up, so tool > > is only const after session__new. > > Seems like a separate issue. Since the session is created > by __perf_session__new(), session->tool will always be a pointer > to a const tool once there is: > > diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h > index 7f69baeae7fb..7c8dd6956330 100644 > --- a/tools/perf/util/session.h > +++ b/tools/perf/util/session.h > @@ -43,7 +43,7 @@ struct perf_session { > u64 one_mmap_offset; > struct ordered_events ordered_events; > struct perf_data *data; > - struct perf_tool *tool; > + const struct perf_tool *tool; > u64 bytes_transferred; > u64 bytes_compressed; > struct zstd_data zstd_data; That's the case after these changes, but not before. > > > > The reason for doing this is to make it so that when I have a tool I > > can reason that nobody is doing things to change it under my feet. > > It still can be changed by the caller of __perf_session__new(), since > the tool itself is not const. > > Anything using container_of() like: > > static int process_sample_event(const struct perf_tool *tool, > union perf_event *event, > struct perf_sample *sample, > struct evsel *evsel, > struct machine *machine) > { > struct perf_script *scr = container_of(tool, struct perf_script, tool); > > can then change scr->tool without even having to cast away const. Agreed, but such things happen in builtin_cmd where the tool is defined and presumably they know what they are doing. My objection is to code in util mutating the tool as I want the tool to have predictable behavior. As callers that take a tool can call fill in defaults (not all) then the tool has to be mutable and I don't want this to be the case. > Really, 'tool' needs to be defined as const in the first place. I'd like this. The problem is initializing all the function pointers and making such initialization robust to extra functions being added to the tool API. It can be done in a long winded way but I couldn't devise macro magic to do it. The other problem is around variables like ordered_events that can't currently be const. The patches move us closer to this being a possibility. > > My > > builtin_cmd is in charge of what the tool is rather than some code > > buried in util that thought it was going to do me a favor. The code is > > a refactor and so the benefit is intended to be for the developer and > > how they reason about the use of tool. > > It creates another question though: since there is a lot of code > before perf_tool__init() is called, does the caller mistakenly > change tool before calling perf_tool__init() If they do this their function pointers will be clobbered and their code won't behave as expected, which I'd expect to be easy to observe. In C++ if you were to initialize memory and then use the memory for a placement new to create an object which would call the constructor, the expected behavior would be that the initialized memory's values would get overridden. I see the use of _init and _exit in the code as being our poor man replacements of constructors and destructors. > > how they reason about the use of tool. We generally use _init > > functions rather than having _fill_defaults, so there is a consistency > > argument. > > The caller does not need the "defaults", so why would it set them up. > The session could just as easily do: > > if (tool->cb) > tool->cb(...); > else > cb_stub(...); Multiplied by every stub, we'd probably need a helper function, how to handle argument passing. There's nothing wrong with this as an idea but I think of this code as trying to create a visitor pattern and this is a visitor pattern with a hard time for the caller. > > I don't expect any impact in terms of performance... Moving > > perf_tool__fill_defaults to __perf_session__new had issues with the > > existing code where NULL would be written over a function pointer > > expecting the later fill_defaults to fix it up, doesn't address coding > > consistency where _init is the norm, and adds another reason the tool > > passed to session__new can't be const. > > perf_tool__init() is not a steeping stone to making 'tool' a > const in the first place. It is because the patch series gets rid of fill in defaults which is why we have a mutable tool passed around. I don't think this is up for debate as the patch series clearly goes from a non-const tool to a const tool at the end. Changing perf_tool__init to make all the function pointers NULL and then making every caller have to do a: if (tool->cb) tool->cb(...); else cb_stub(...); I think it is a less elegant solution at the end, it is also a large and more invasive change. The various refactorings to make tool const, changing the aux use of tool, etc. wouldn't be impacted by such a change but I think it is out of scope for this patch series. Thanks, Ian ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-22 16:06 ` Ian Rogers @ 2024-07-22 17:45 ` Namhyung Kim 2024-07-22 17:50 ` Ian Rogers 0 siblings, 1 reply; 40+ messages in thread From: Namhyung Kim @ 2024-07-22 17:45 UTC (permalink / raw) To: Ian Rogers Cc: Adrian Hunter, Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel Hi, On Mon, Jul 22, 2024 at 9:06 AM Ian Rogers <irogers@google.com> wrote: > > On Mon, Jul 22, 2024 at 1:29 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > > On 19/07/24 19:26, Ian Rogers wrote: > > > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > >> > > >> On 18/07/24 03:59, Ian Rogers wrote: > > >>> struct perf_tool provides a set of function pointers that are called > > >>> through when processing perf data. To make filling the pointers less > > >>> cumbersome, if they are NULL perf_tools__fill_defaults will add > > >>> default do nothing implementations. > > >>> > > >>> This change refactors struct perf_tool to have an init function that > > >>> provides the default implementation. The special use of NULL and > > >>> perf_tools__fill_defaults are removed. As a consequence the tool > > >>> pointers can then all be made const, which better reflects the > > >>> behavior a particular perf command would expect of the tool and to > > >>> some extent can reduce the cognitive load on someone working on a > > >>> command. > > >>> > > >>> v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > > >> > > >> The tags were really meant only for patch 1, the email that was replied to. > > >> > > >> But now for patches 2 and 3: > > >> > > >> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > > > > > > Sorry for that, you'd mentioned that pt and bts testing which is > > > impacted by more than just patch 1. > > > > > >> Looking at patches 4 to 25, they do not seem to offer any benefit. > > >> > > >> Instead of patch 26, presumably perf_tool__fill_defaults() could > > >> be moved to __perf_session__new(), which perhaps would allow > > >> patch 27 as it is. > > > > > > What I'm trying to do in the series is make it so that the tool isn't > > > mutated during its use by session. Ideally we'd be passing a const > > > tool to session_new, that's not possible because there's a hack to fix > > > ordered events and pipe mode in session__new: > > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 > > > Imo, it isn't great to pass a tool to session__new where you say you > > > want ordered events and then session just goes to change that for you. > > > Altering that behavior was beyond the scope of this clean up, so tool > > > is only const after session__new. > > > > Seems like a separate issue. Since the session is created > > by __perf_session__new(), session->tool will always be a pointer > > to a const tool once there is: > > > > diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h > > index 7f69baeae7fb..7c8dd6956330 100644 > > --- a/tools/perf/util/session.h > > +++ b/tools/perf/util/session.h > > @@ -43,7 +43,7 @@ struct perf_session { > > u64 one_mmap_offset; > > struct ordered_events ordered_events; > > struct perf_data *data; > > - struct perf_tool *tool; > > + const struct perf_tool *tool; > > u64 bytes_transferred; > > u64 bytes_compressed; > > struct zstd_data zstd_data; > > That's the case after these changes, but not before. > > > > > > > The reason for doing this is to make it so that when I have a tool I > > > can reason that nobody is doing things to change it under my feet. > > > > It still can be changed by the caller of __perf_session__new(), since > > the tool itself is not const. > > > > Anything using container_of() like: > > > > static int process_sample_event(const struct perf_tool *tool, > > union perf_event *event, > > struct perf_sample *sample, > > struct evsel *evsel, > > struct machine *machine) > > { > > struct perf_script *scr = container_of(tool, struct perf_script, tool); > > > > can then change scr->tool without even having to cast away const. > > Agreed, but such things happen in builtin_cmd where the tool is > defined and presumably they know what they are doing. My objection is > to code in util mutating the tool as I want the tool to have > predictable behavior. As callers that take a tool can call fill in > defaults (not all) then the tool has to be mutable and I don't want > this to be the case. > > > Really, 'tool' needs to be defined as const in the first place. > > I'd like this. The problem is initializing all the function pointers > and making such initialization robust to extra functions being added > to the tool API. It can be done in a long winded way but I couldn't > devise macro magic to do it. The other problem is around variables > like ordered_events that can't currently be const. The patches move us > closer to this being a possibility. > > > > My > > > builtin_cmd is in charge of what the tool is rather than some code > > > buried in util that thought it was going to do me a favor. The code is > > > a refactor and so the benefit is intended to be for the developer and > > > how they reason about the use of tool. > > > > It creates another question though: since there is a lot of code > > before perf_tool__init() is called, does the caller mistakenly > > change tool before calling perf_tool__init() > > If they do this their function pointers will be clobbered and their > code won't behave as expected, which I'd expect to be easy to observe. > In C++ if you were to initialize memory and then use the memory for a > placement new to create an object which would call the constructor, > the expected behavior would be that the initialized memory's values > would get overridden. I see the use of _init and _exit in the code as > being our poor man replacements of constructors and destructors. > > > > how they reason about the use of tool. We generally use _init > > > functions rather than having _fill_defaults, so there is a consistency > > > argument. > > > > The caller does not need the "defaults", so why would it set them up. > > The session could just as easily do: > > > > if (tool->cb) > > tool->cb(...); > > else > > cb_stub(...); > > Multiplied by every stub, we'd probably need a helper function, how to > handle argument passing. There's nothing wrong with this as an idea > but I think of this code as trying to create a visitor pattern and > this is a visitor pattern with a hard time for the caller. > > > > I don't expect any impact in terms of performance... Moving > > > perf_tool__fill_defaults to __perf_session__new had issues with the > > > existing code where NULL would be written over a function pointer > > > expecting the later fill_defaults to fix it up, doesn't address coding > > > consistency where _init is the norm, and adds another reason the tool > > > passed to session__new can't be const. > > > > perf_tool__init() is not a steeping stone to making 'tool' a > > const in the first place. > > It is because the patch series gets rid of fill in defaults which is > why we have a mutable tool passed around. I don't think this is up for > debate as the patch series clearly goes from a non-const > tool to a const tool at the end. Changing perf_tool__init to make all > the function pointers NULL and then making every caller have to do a: > > if (tool->cb) > tool->cb(...); > else > cb_stub(...); > > I think it is a less elegant solution at the end, it is also a large > and more invasive change. The various refactorings to make tool const, > changing the aux use of tool, etc. wouldn't be impacted by such a > change but I think it is out of scope for this patch series. I don't think it's a large and invasive change. The tools are mostly zero-initialized so we don't need to reset to NULL. And tool->cb is called mostly from two functions: machines__deliver_event() and perf_session__process_user_event(). Can we change them to check NULL and get rid of perf_tool__fill_defaults() to keep it const? Thanks, Namhyung ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-22 17:45 ` Namhyung Kim @ 2024-07-22 17:50 ` Ian Rogers 2024-07-22 18:04 ` Ian Rogers 0 siblings, 1 reply; 40+ messages in thread From: Ian Rogers @ 2024-07-22 17:50 UTC (permalink / raw) To: Namhyung Kim Cc: Adrian Hunter, Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Mon, Jul 22, 2024 at 10:45 AM Namhyung Kim <namhyung@kernel.org> wrote: > > Hi, > > On Mon, Jul 22, 2024 at 9:06 AM Ian Rogers <irogers@google.com> wrote: > > > > On Mon, Jul 22, 2024 at 1:29 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > > > > On 19/07/24 19:26, Ian Rogers wrote: > > > > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > >> > > > >> On 18/07/24 03:59, Ian Rogers wrote: > > > >>> struct perf_tool provides a set of function pointers that are called > > > >>> through when processing perf data. To make filling the pointers less > > > >>> cumbersome, if they are NULL perf_tools__fill_defaults will add > > > >>> default do nothing implementations. > > > >>> > > > >>> This change refactors struct perf_tool to have an init function that > > > >>> provides the default implementation. The special use of NULL and > > > >>> perf_tools__fill_defaults are removed. As a consequence the tool > > > >>> pointers can then all be made const, which better reflects the > > > >>> behavior a particular perf command would expect of the tool and to > > > >>> some extent can reduce the cognitive load on someone working on a > > > >>> command. > > > >>> > > > >>> v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > > > >> > > > >> The tags were really meant only for patch 1, the email that was replied to. > > > >> > > > >> But now for patches 2 and 3: > > > >> > > > >> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > > > > > > > > Sorry for that, you'd mentioned that pt and bts testing which is > > > > impacted by more than just patch 1. > > > > > > > >> Looking at patches 4 to 25, they do not seem to offer any benefit. > > > >> > > > >> Instead of patch 26, presumably perf_tool__fill_defaults() could > > > >> be moved to __perf_session__new(), which perhaps would allow > > > >> patch 27 as it is. > > > > > > > > What I'm trying to do in the series is make it so that the tool isn't > > > > mutated during its use by session. Ideally we'd be passing a const > > > > tool to session_new, that's not possible because there's a hack to fix > > > > ordered events and pipe mode in session__new: > > > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 > > > > Imo, it isn't great to pass a tool to session__new where you say you > > > > want ordered events and then session just goes to change that for you. > > > > Altering that behavior was beyond the scope of this clean up, so tool > > > > is only const after session__new. > > > > > > Seems like a separate issue. Since the session is created > > > by __perf_session__new(), session->tool will always be a pointer > > > to a const tool once there is: > > > > > > diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h > > > index 7f69baeae7fb..7c8dd6956330 100644 > > > --- a/tools/perf/util/session.h > > > +++ b/tools/perf/util/session.h > > > @@ -43,7 +43,7 @@ struct perf_session { > > > u64 one_mmap_offset; > > > struct ordered_events ordered_events; > > > struct perf_data *data; > > > - struct perf_tool *tool; > > > + const struct perf_tool *tool; > > > u64 bytes_transferred; > > > u64 bytes_compressed; > > > struct zstd_data zstd_data; > > > > That's the case after these changes, but not before. > > > > > > > > > > The reason for doing this is to make it so that when I have a tool I > > > > can reason that nobody is doing things to change it under my feet. > > > > > > It still can be changed by the caller of __perf_session__new(), since > > > the tool itself is not const. > > > > > > Anything using container_of() like: > > > > > > static int process_sample_event(const struct perf_tool *tool, > > > union perf_event *event, > > > struct perf_sample *sample, > > > struct evsel *evsel, > > > struct machine *machine) > > > { > > > struct perf_script *scr = container_of(tool, struct perf_script, tool); > > > > > > can then change scr->tool without even having to cast away const. > > > > Agreed, but such things happen in builtin_cmd where the tool is > > defined and presumably they know what they are doing. My objection is > > to code in util mutating the tool as I want the tool to have > > predictable behavior. As callers that take a tool can call fill in > > defaults (not all) then the tool has to be mutable and I don't want > > this to be the case. > > > > > Really, 'tool' needs to be defined as const in the first place. > > > > I'd like this. The problem is initializing all the function pointers > > and making such initialization robust to extra functions being added > > to the tool API. It can be done in a long winded way but I couldn't > > devise macro magic to do it. The other problem is around variables > > like ordered_events that can't currently be const. The patches move us > > closer to this being a possibility. > > > > > > My > > > > builtin_cmd is in charge of what the tool is rather than some code > > > > buried in util that thought it was going to do me a favor. The code is > > > > a refactor and so the benefit is intended to be for the developer and > > > > how they reason about the use of tool. > > > > > > It creates another question though: since there is a lot of code > > > before perf_tool__init() is called, does the caller mistakenly > > > change tool before calling perf_tool__init() > > > > If they do this their function pointers will be clobbered and their > > code won't behave as expected, which I'd expect to be easy to observe. > > In C++ if you were to initialize memory and then use the memory for a > > placement new to create an object which would call the constructor, > > the expected behavior would be that the initialized memory's values > > would get overridden. I see the use of _init and _exit in the code as > > being our poor man replacements of constructors and destructors. > > > > > > how they reason about the use of tool. We generally use _init > > > > functions rather than having _fill_defaults, so there is a consistency > > > > argument. > > > > > > The caller does not need the "defaults", so why would it set them up. > > > The session could just as easily do: > > > > > > if (tool->cb) > > > tool->cb(...); > > > else > > > cb_stub(...); > > > > Multiplied by every stub, we'd probably need a helper function, how to > > handle argument passing. There's nothing wrong with this as an idea > > but I think of this code as trying to create a visitor pattern and > > this is a visitor pattern with a hard time for the caller. > > > > > > I don't expect any impact in terms of performance... Moving > > > > perf_tool__fill_defaults to __perf_session__new had issues with the > > > > existing code where NULL would be written over a function pointer > > > > expecting the later fill_defaults to fix it up, doesn't address coding > > > > consistency where _init is the norm, and adds another reason the tool > > > > passed to session__new can't be const. > > > > > > perf_tool__init() is not a steeping stone to making 'tool' a > > > const in the first place. > > > > It is because the patch series gets rid of fill in defaults which is > > why we have a mutable tool passed around. I don't think this is up for > > debate as the patch series clearly goes from a non-const > > tool to a const tool at the end. Changing perf_tool__init to make all > > the function pointers NULL and then making every caller have to do a: > > > > if (tool->cb) > > tool->cb(...); > > else > > cb_stub(...); > > > > I think it is a less elegant solution at the end, it is also a large > > and more invasive change. The various refactorings to make tool const, > > changing the aux use of tool, etc. wouldn't be impacted by such a > > change but I think it is out of scope for this patch series. > > I don't think it's a large and invasive change. The tools are mostly > zero-initialized so we don't need to reset to NULL. And tool->cb is > called mostly from two functions: machines__deliver_event() and > perf_session__process_user_event(). Can we change them to check > NULL and get rid of perf_tool__fill_defaults() to keep it const? As I said above, I don't think that is good style and is out of scope here. It clearly can be done as follow up, but I don't see how that fixes the style issue. Thanks, Ian ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-22 17:50 ` Ian Rogers @ 2024-07-22 18:04 ` Ian Rogers 2024-07-23 9:38 ` Adrian Hunter 0 siblings, 1 reply; 40+ messages in thread From: Ian Rogers @ 2024-07-22 18:04 UTC (permalink / raw) To: Namhyung Kim Cc: Adrian Hunter, Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Mon, Jul 22, 2024 at 10:50 AM Ian Rogers <irogers@google.com> wrote: > > On Mon, Jul 22, 2024 at 10:45 AM Namhyung Kim <namhyung@kernel.org> wrote: > > > > Hi, > > > > On Mon, Jul 22, 2024 at 9:06 AM Ian Rogers <irogers@google.com> wrote: > > > > > > On Mon, Jul 22, 2024 at 1:29 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > > > > > > On 19/07/24 19:26, Ian Rogers wrote: > > > > > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > > >> > > > > >> On 18/07/24 03:59, Ian Rogers wrote: > > > > >>> struct perf_tool provides a set of function pointers that are called > > > > >>> through when processing perf data. To make filling the pointers less > > > > >>> cumbersome, if they are NULL perf_tools__fill_defaults will add > > > > >>> default do nothing implementations. > > > > >>> > > > > >>> This change refactors struct perf_tool to have an init function that > > > > >>> provides the default implementation. The special use of NULL and > > > > >>> perf_tools__fill_defaults are removed. As a consequence the tool > > > > >>> pointers can then all be made const, which better reflects the > > > > >>> behavior a particular perf command would expect of the tool and to > > > > >>> some extent can reduce the cognitive load on someone working on a > > > > >>> command. > > > > >>> > > > > >>> v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > > > > >> > > > > >> The tags were really meant only for patch 1, the email that was replied to. > > > > >> > > > > >> But now for patches 2 and 3: > > > > >> > > > > >> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > > > > > > > > > > Sorry for that, you'd mentioned that pt and bts testing which is > > > > > impacted by more than just patch 1. > > > > > > > > > >> Looking at patches 4 to 25, they do not seem to offer any benefit. > > > > >> > > > > >> Instead of patch 26, presumably perf_tool__fill_defaults() could > > > > >> be moved to __perf_session__new(), which perhaps would allow > > > > >> patch 27 as it is. > > > > > > > > > > What I'm trying to do in the series is make it so that the tool isn't > > > > > mutated during its use by session. Ideally we'd be passing a const > > > > > tool to session_new, that's not possible because there's a hack to fix > > > > > ordered events and pipe mode in session__new: > > > > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 > > > > > Imo, it isn't great to pass a tool to session__new where you say you > > > > > want ordered events and then session just goes to change that for you. > > > > > Altering that behavior was beyond the scope of this clean up, so tool > > > > > is only const after session__new. > > > > > > > > Seems like a separate issue. Since the session is created > > > > by __perf_session__new(), session->tool will always be a pointer > > > > to a const tool once there is: > > > > > > > > diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h > > > > index 7f69baeae7fb..7c8dd6956330 100644 > > > > --- a/tools/perf/util/session.h > > > > +++ b/tools/perf/util/session.h > > > > @@ -43,7 +43,7 @@ struct perf_session { > > > > u64 one_mmap_offset; > > > > struct ordered_events ordered_events; > > > > struct perf_data *data; > > > > - struct perf_tool *tool; > > > > + const struct perf_tool *tool; > > > > u64 bytes_transferred; > > > > u64 bytes_compressed; > > > > struct zstd_data zstd_data; > > > > > > That's the case after these changes, but not before. > > > > > > > > > > > > > The reason for doing this is to make it so that when I have a tool I > > > > > can reason that nobody is doing things to change it under my feet. > > > > > > > > It still can be changed by the caller of __perf_session__new(), since > > > > the tool itself is not const. > > > > > > > > Anything using container_of() like: > > > > > > > > static int process_sample_event(const struct perf_tool *tool, > > > > union perf_event *event, > > > > struct perf_sample *sample, > > > > struct evsel *evsel, > > > > struct machine *machine) > > > > { > > > > struct perf_script *scr = container_of(tool, struct perf_script, tool); > > > > > > > > can then change scr->tool without even having to cast away const. > > > > > > Agreed, but such things happen in builtin_cmd where the tool is > > > defined and presumably they know what they are doing. My objection is > > > to code in util mutating the tool as I want the tool to have > > > predictable behavior. As callers that take a tool can call fill in > > > defaults (not all) then the tool has to be mutable and I don't want > > > this to be the case. > > > > > > > Really, 'tool' needs to be defined as const in the first place. > > > > > > I'd like this. The problem is initializing all the function pointers > > > and making such initialization robust to extra functions being added > > > to the tool API. It can be done in a long winded way but I couldn't > > > devise macro magic to do it. The other problem is around variables > > > like ordered_events that can't currently be const. The patches move us > > > closer to this being a possibility. > > > > > > > > My > > > > > builtin_cmd is in charge of what the tool is rather than some code > > > > > buried in util that thought it was going to do me a favor. The code is > > > > > a refactor and so the benefit is intended to be for the developer and > > > > > how they reason about the use of tool. > > > > > > > > It creates another question though: since there is a lot of code > > > > before perf_tool__init() is called, does the caller mistakenly > > > > change tool before calling perf_tool__init() > > > > > > If they do this their function pointers will be clobbered and their > > > code won't behave as expected, which I'd expect to be easy to observe. > > > In C++ if you were to initialize memory and then use the memory for a > > > placement new to create an object which would call the constructor, > > > the expected behavior would be that the initialized memory's values > > > would get overridden. I see the use of _init and _exit in the code as > > > being our poor man replacements of constructors and destructors. > > > > > > > > how they reason about the use of tool. We generally use _init > > > > > functions rather than having _fill_defaults, so there is a consistency > > > > > argument. > > > > > > > > The caller does not need the "defaults", so why would it set them up. > > > > The session could just as easily do: > > > > > > > > if (tool->cb) > > > > tool->cb(...); > > > > else > > > > cb_stub(...); > > > > > > Multiplied by every stub, we'd probably need a helper function, how to > > > handle argument passing. There's nothing wrong with this as an idea > > > but I think of this code as trying to create a visitor pattern and > > > this is a visitor pattern with a hard time for the caller. > > > > > > > > I don't expect any impact in terms of performance... Moving > > > > > perf_tool__fill_defaults to __perf_session__new had issues with the > > > > > existing code where NULL would be written over a function pointer > > > > > expecting the later fill_defaults to fix it up, doesn't address coding > > > > > consistency where _init is the norm, and adds another reason the tool > > > > > passed to session__new can't be const. > > > > > > > > perf_tool__init() is not a steeping stone to making 'tool' a > > > > const in the first place. > > > > > > It is because the patch series gets rid of fill in defaults which is > > > why we have a mutable tool passed around. I don't think this is up for > > > debate as the patch series clearly goes from a non-const > > > tool to a const tool at the end. Changing perf_tool__init to make all > > > the function pointers NULL and then making every caller have to do a: > > > > > > if (tool->cb) > > > tool->cb(...); > > > else > > > cb_stub(...); > > > > > > I think it is a less elegant solution at the end, it is also a large > > > and more invasive change. The various refactorings to make tool const, > > > changing the aux use of tool, etc. wouldn't be impacted by such a > > > change but I think it is out of scope for this patch series. > > > > I don't think it's a large and invasive change. The tools are mostly > > zero-initialized so we don't need to reset to NULL. And tool->cb is > > called mostly from two functions: machines__deliver_event() and > > perf_session__process_user_event(). Can we change them to check > > NULL and get rid of perf_tool__fill_defaults() to keep it const? > > As I said above, I don't think that is good style and is out of scope > here. It clearly can be done as follow up, but I don't see how that > fixes the style issue. Just to be clear on what the style issue is. We already have code: https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/builtin-record.c?h=perf-tools-next#n1461 ``` if (rec->buildid_all && !rec->timestamp_boundary) rec->tool.sample = NULL; ``` that relies on the special behavior of NULL in a function pointer being changed at dispatch time - a simple reading of that code would be anyone calling the function pointer would get a segv. I'm trying to make it so that NULL isn't magic in the context of tool and you can simply look at the tool to understand what its behavior is, much as a virtual method table would work if we could do proper object-oriented development. Thanks, Ian > Thanks, > Ian ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-22 18:04 ` Ian Rogers @ 2024-07-23 9:38 ` Adrian Hunter 2024-07-23 12:50 ` Ian Rogers 0 siblings, 1 reply; 40+ messages in thread From: Adrian Hunter @ 2024-07-23 9:38 UTC (permalink / raw) To: Ian Rogers, Namhyung Kim Cc: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On 22/07/24 21:04, Ian Rogers wrote: > On Mon, Jul 22, 2024 at 10:50 AM Ian Rogers <irogers@google.com> wrote: >> >> On Mon, Jul 22, 2024 at 10:45 AM Namhyung Kim <namhyung@kernel.org> wrote: >>> >>> Hi, >>> >>> On Mon, Jul 22, 2024 at 9:06 AM Ian Rogers <irogers@google.com> wrote: >>>> >>>> On Mon, Jul 22, 2024 at 1:29 AM Adrian Hunter <adrian.hunter@intel.com> wrote: >>>>> >>>>> On 19/07/24 19:26, Ian Rogers wrote: >>>>>> On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: >>>>>>> >>>>>>> On 18/07/24 03:59, Ian Rogers wrote: >>>>>>>> struct perf_tool provides a set of function pointers that are called >>>>>>>> through when processing perf data. To make filling the pointers less >>>>>>>> cumbersome, if they are NULL perf_tools__fill_defaults will add >>>>>>>> default do nothing implementations. >>>>>>>> >>>>>>>> This change refactors struct perf_tool to have an init function that >>>>>>>> provides the default implementation. The special use of NULL and >>>>>>>> perf_tools__fill_defaults are removed. As a consequence the tool >>>>>>>> pointers can then all be made const, which better reflects the >>>>>>>> behavior a particular perf command would expect of the tool and to >>>>>>>> some extent can reduce the cognitive load on someone working on a >>>>>>>> command. >>>>>>>> >>>>>>>> v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. >>>>>>> >>>>>>> The tags were really meant only for patch 1, the email that was replied to. >>>>>>> >>>>>>> But now for patches 2 and 3: >>>>>>> >>>>>>> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> >>>>>> >>>>>> Sorry for that, you'd mentioned that pt and bts testing which is >>>>>> impacted by more than just patch 1. >>>>>> >>>>>>> Looking at patches 4 to 25, they do not seem to offer any benefit. >>>>>>> >>>>>>> Instead of patch 26, presumably perf_tool__fill_defaults() could >>>>>>> be moved to __perf_session__new(), which perhaps would allow >>>>>>> patch 27 as it is. >>>>>> >>>>>> What I'm trying to do in the series is make it so that the tool isn't >>>>>> mutated during its use by session. Ideally we'd be passing a const >>>>>> tool to session_new, that's not possible because there's a hack to fix >>>>>> ordered events and pipe mode in session__new: >>>>>> https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 >>>>>> Imo, it isn't great to pass a tool to session__new where you say you >>>>>> want ordered events and then session just goes to change that for you. >>>>>> Altering that behavior was beyond the scope of this clean up, so tool >>>>>> is only const after session__new. >>>>> >>>>> Seems like a separate issue. Since the session is created >>>>> by __perf_session__new(), session->tool will always be a pointer >>>>> to a const tool once there is: >>>>> >>>>> diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h >>>>> index 7f69baeae7fb..7c8dd6956330 100644 >>>>> --- a/tools/perf/util/session.h >>>>> +++ b/tools/perf/util/session.h >>>>> @@ -43,7 +43,7 @@ struct perf_session { >>>>> u64 one_mmap_offset; >>>>> struct ordered_events ordered_events; >>>>> struct perf_data *data; >>>>> - struct perf_tool *tool; >>>>> + const struct perf_tool *tool; >>>>> u64 bytes_transferred; >>>>> u64 bytes_compressed; >>>>> struct zstd_data zstd_data; >>>> >>>> That's the case after these changes, but not before. >>>> >>>>>> >>>>>> The reason for doing this is to make it so that when I have a tool I >>>>>> can reason that nobody is doing things to change it under my feet. >>>>> >>>>> It still can be changed by the caller of __perf_session__new(), since >>>>> the tool itself is not const. >>>>> >>>>> Anything using container_of() like: >>>>> >>>>> static int process_sample_event(const struct perf_tool *tool, >>>>> union perf_event *event, >>>>> struct perf_sample *sample, >>>>> struct evsel *evsel, >>>>> struct machine *machine) >>>>> { >>>>> struct perf_script *scr = container_of(tool, struct perf_script, tool); >>>>> >>>>> can then change scr->tool without even having to cast away const. >>>> >>>> Agreed, but such things happen in builtin_cmd where the tool is >>>> defined and presumably they know what they are doing. My objection is >>>> to code in util mutating the tool as I want the tool to have >>>> predictable behavior. As callers that take a tool can call fill in >>>> defaults (not all) then the tool has to be mutable and I don't want >>>> this to be the case. >>>> >>>>> Really, 'tool' needs to be defined as const in the first place. >>>> >>>> I'd like this. The problem is initializing all the function pointers >>>> and making such initialization robust to extra functions being added >>>> to the tool API. It can be done in a long winded way but I couldn't >>>> devise macro magic to do it. The other problem is around variables >>>> like ordered_events that can't currently be const. The patches move us >>>> closer to this being a possibility. >>>> >>>>>> My >>>>>> builtin_cmd is in charge of what the tool is rather than some code >>>>>> buried in util that thought it was going to do me a favor. The code is >>>>>> a refactor and so the benefit is intended to be for the developer and >>>>>> how they reason about the use of tool. >>>>> >>>>> It creates another question though: since there is a lot of code >>>>> before perf_tool__init() is called, does the caller mistakenly >>>>> change tool before calling perf_tool__init() >>>> >>>> If they do this their function pointers will be clobbered and their >>>> code won't behave as expected, which I'd expect to be easy to observe. >>>> In C++ if you were to initialize memory and then use the memory for a >>>> placement new to create an object which would call the constructor, >>>> the expected behavior would be that the initialized memory's values >>>> would get overridden. I see the use of _init and _exit in the code as >>>> being our poor man replacements of constructors and destructors. >>>> >>>>>> how they reason about the use of tool. We generally use _init >>>>>> functions rather than having _fill_defaults, so there is a consistency >>>>>> argument. >>>>> >>>>> The caller does not need the "defaults", so why would it set them up. >>>>> The session could just as easily do: >>>>> >>>>> if (tool->cb) >>>>> tool->cb(...); >>>>> else >>>>> cb_stub(...); >>>> >>>> Multiplied by every stub, we'd probably need a helper function, how to >>>> handle argument passing. There's nothing wrong with this as an idea >>>> but I think of this code as trying to create a visitor pattern and >>>> this is a visitor pattern with a hard time for the caller. >>>> >>>>>> I don't expect any impact in terms of performance... Moving >>>>>> perf_tool__fill_defaults to __perf_session__new had issues with the >>>>>> existing code where NULL would be written over a function pointer >>>>>> expecting the later fill_defaults to fix it up, doesn't address coding >>>>>> consistency where _init is the norm, and adds another reason the tool >>>>>> passed to session__new can't be const. >>>>> >>>>> perf_tool__init() is not a steeping stone to making 'tool' a >>>>> const in the first place. >>>> >>>> It is because the patch series gets rid of fill in defaults which is >>>> why we have a mutable tool passed around. I don't think this is up for >>>> debate as the patch series clearly goes from a non-const >>>> tool to a const tool at the end. Changing perf_tool__init to make all >>>> the function pointers NULL and then making every caller have to do a: >>>> >>>> if (tool->cb) >>>> tool->cb(...); >>>> else >>>> cb_stub(...); >>>> >>>> I think it is a less elegant solution at the end, it is also a large >>>> and more invasive change. The various refactorings to make tool const, >>>> changing the aux use of tool, etc. wouldn't be impacted by such a >>>> change but I think it is out of scope for this patch series. >>> >>> I don't think it's a large and invasive change. The tools are mostly >>> zero-initialized so we don't need to reset to NULL. And tool->cb is >>> called mostly from two functions: machines__deliver_event() and >>> perf_session__process_user_event(). Can we change them to check >>> NULL and get rid of perf_tool__fill_defaults() to keep it const? >> >> As I said above, I don't think that is good style and is out of scope >> here. It clearly can be done as follow up, but I don't see how that >> fixes the style issue. > > Just to be clear on what the style issue is. We already have code: > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/builtin-record.c?h=perf-tools-next#n1461 > ``` > if (rec->buildid_all && !rec->timestamp_boundary) > rec->tool.sample = NULL; > ``` > that relies on the special behavior of NULL in a function pointer > being changed at dispatch time - a simple reading of that code would > be anyone calling the function pointer would get a segv. I'm trying to > make it so that NULL isn't magic in the context of tool and you can > simply look at the tool to understand what its behavior is, much as a > virtual method table would work if we could do proper object-oriented > development. In C, NULL or zero is often used as a special value to mean no-value. Optional callbacks that are NULL is also not remarkable. ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-23 9:38 ` Adrian Hunter @ 2024-07-23 12:50 ` Ian Rogers 0 siblings, 0 replies; 40+ messages in thread From: Ian Rogers @ 2024-07-23 12:50 UTC (permalink / raw) To: Adrian Hunter Cc: Namhyung Kim, Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Tue, Jul 23, 2024 at 2:38 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > On 22/07/24 21:04, Ian Rogers wrote: > > On Mon, Jul 22, 2024 at 10:50 AM Ian Rogers <irogers@google.com> wrote: > >> > >> On Mon, Jul 22, 2024 at 10:45 AM Namhyung Kim <namhyung@kernel.org> wrote: > >>> > >>> Hi, > >>> > >>> On Mon, Jul 22, 2024 at 9:06 AM Ian Rogers <irogers@google.com> wrote: > >>>> > >>>> On Mon, Jul 22, 2024 at 1:29 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > >>>>> > >>>>> On 19/07/24 19:26, Ian Rogers wrote: > >>>>>> On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > >>>>>>> > >>>>>>> On 18/07/24 03:59, Ian Rogers wrote: > >>>>>>>> struct perf_tool provides a set of function pointers that are called > >>>>>>>> through when processing perf data. To make filling the pointers less > >>>>>>>> cumbersome, if they are NULL perf_tools__fill_defaults will add > >>>>>>>> default do nothing implementations. > >>>>>>>> > >>>>>>>> This change refactors struct perf_tool to have an init function that > >>>>>>>> provides the default implementation. The special use of NULL and > >>>>>>>> perf_tools__fill_defaults are removed. As a consequence the tool > >>>>>>>> pointers can then all be made const, which better reflects the > >>>>>>>> behavior a particular perf command would expect of the tool and to > >>>>>>>> some extent can reduce the cognitive load on someone working on a > >>>>>>>> command. > >>>>>>>> > >>>>>>>> v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > >>>>>>> > >>>>>>> The tags were really meant only for patch 1, the email that was replied to. > >>>>>>> > >>>>>>> But now for patches 2 and 3: > >>>>>>> > >>>>>>> Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > >>>>>> > >>>>>> Sorry for that, you'd mentioned that pt and bts testing which is > >>>>>> impacted by more than just patch 1. > >>>>>> > >>>>>>> Looking at patches 4 to 25, they do not seem to offer any benefit. > >>>>>>> > >>>>>>> Instead of patch 26, presumably perf_tool__fill_defaults() could > >>>>>>> be moved to __perf_session__new(), which perhaps would allow > >>>>>>> patch 27 as it is. > >>>>>> > >>>>>> What I'm trying to do in the series is make it so that the tool isn't > >>>>>> mutated during its use by session. Ideally we'd be passing a const > >>>>>> tool to session_new, that's not possible because there's a hack to fix > >>>>>> ordered events and pipe mode in session__new: > >>>>>> https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 > >>>>>> Imo, it isn't great to pass a tool to session__new where you say you > >>>>>> want ordered events and then session just goes to change that for you. > >>>>>> Altering that behavior was beyond the scope of this clean up, so tool > >>>>>> is only const after session__new. > >>>>> > >>>>> Seems like a separate issue. Since the session is created > >>>>> by __perf_session__new(), session->tool will always be a pointer > >>>>> to a const tool once there is: > >>>>> > >>>>> diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h > >>>>> index 7f69baeae7fb..7c8dd6956330 100644 > >>>>> --- a/tools/perf/util/session.h > >>>>> +++ b/tools/perf/util/session.h > >>>>> @@ -43,7 +43,7 @@ struct perf_session { > >>>>> u64 one_mmap_offset; > >>>>> struct ordered_events ordered_events; > >>>>> struct perf_data *data; > >>>>> - struct perf_tool *tool; > >>>>> + const struct perf_tool *tool; > >>>>> u64 bytes_transferred; > >>>>> u64 bytes_compressed; > >>>>> struct zstd_data zstd_data; > >>>> > >>>> That's the case after these changes, but not before. > >>>> > >>>>>> > >>>>>> The reason for doing this is to make it so that when I have a tool I > >>>>>> can reason that nobody is doing things to change it under my feet. > >>>>> > >>>>> It still can be changed by the caller of __perf_session__new(), since > >>>>> the tool itself is not const. > >>>>> > >>>>> Anything using container_of() like: > >>>>> > >>>>> static int process_sample_event(const struct perf_tool *tool, > >>>>> union perf_event *event, > >>>>> struct perf_sample *sample, > >>>>> struct evsel *evsel, > >>>>> struct machine *machine) > >>>>> { > >>>>> struct perf_script *scr = container_of(tool, struct perf_script, tool); > >>>>> > >>>>> can then change scr->tool without even having to cast away const. > >>>> > >>>> Agreed, but such things happen in builtin_cmd where the tool is > >>>> defined and presumably they know what they are doing. My objection is > >>>> to code in util mutating the tool as I want the tool to have > >>>> predictable behavior. As callers that take a tool can call fill in > >>>> defaults (not all) then the tool has to be mutable and I don't want > >>>> this to be the case. > >>>> > >>>>> Really, 'tool' needs to be defined as const in the first place. > >>>> > >>>> I'd like this. The problem is initializing all the function pointers > >>>> and making such initialization robust to extra functions being added > >>>> to the tool API. It can be done in a long winded way but I couldn't > >>>> devise macro magic to do it. The other problem is around variables > >>>> like ordered_events that can't currently be const. The patches move us > >>>> closer to this being a possibility. > >>>> > >>>>>> My > >>>>>> builtin_cmd is in charge of what the tool is rather than some code > >>>>>> buried in util that thought it was going to do me a favor. The code is > >>>>>> a refactor and so the benefit is intended to be for the developer and > >>>>>> how they reason about the use of tool. > >>>>> > >>>>> It creates another question though: since there is a lot of code > >>>>> before perf_tool__init() is called, does the caller mistakenly > >>>>> change tool before calling perf_tool__init() > >>>> > >>>> If they do this their function pointers will be clobbered and their > >>>> code won't behave as expected, which I'd expect to be easy to observe. > >>>> In C++ if you were to initialize memory and then use the memory for a > >>>> placement new to create an object which would call the constructor, > >>>> the expected behavior would be that the initialized memory's values > >>>> would get overridden. I see the use of _init and _exit in the code as > >>>> being our poor man replacements of constructors and destructors. > >>>> > >>>>>> how they reason about the use of tool. We generally use _init > >>>>>> functions rather than having _fill_defaults, so there is a consistency > >>>>>> argument. > >>>>> > >>>>> The caller does not need the "defaults", so why would it set them up. > >>>>> The session could just as easily do: > >>>>> > >>>>> if (tool->cb) > >>>>> tool->cb(...); > >>>>> else > >>>>> cb_stub(...); > >>>> > >>>> Multiplied by every stub, we'd probably need a helper function, how to > >>>> handle argument passing. There's nothing wrong with this as an idea > >>>> but I think of this code as trying to create a visitor pattern and > >>>> this is a visitor pattern with a hard time for the caller. > >>>> > >>>>>> I don't expect any impact in terms of performance... Moving > >>>>>> perf_tool__fill_defaults to __perf_session__new had issues with the > >>>>>> existing code where NULL would be written over a function pointer > >>>>>> expecting the later fill_defaults to fix it up, doesn't address coding > >>>>>> consistency where _init is the norm, and adds another reason the tool > >>>>>> passed to session__new can't be const. > >>>>> > >>>>> perf_tool__init() is not a steeping stone to making 'tool' a > >>>>> const in the first place. > >>>> > >>>> It is because the patch series gets rid of fill in defaults which is > >>>> why we have a mutable tool passed around. I don't think this is up for > >>>> debate as the patch series clearly goes from a non-const > >>>> tool to a const tool at the end. Changing perf_tool__init to make all > >>>> the function pointers NULL and then making every caller have to do a: > >>>> > >>>> if (tool->cb) > >>>> tool->cb(...); > >>>> else > >>>> cb_stub(...); > >>>> > >>>> I think it is a less elegant solution at the end, it is also a large > >>>> and more invasive change. The various refactorings to make tool const, > >>>> changing the aux use of tool, etc. wouldn't be impacted by such a > >>>> change but I think it is out of scope for this patch series. > >>> > >>> I don't think it's a large and invasive change. The tools are mostly > >>> zero-initialized so we don't need to reset to NULL. And tool->cb is > >>> called mostly from two functions: machines__deliver_event() and > >>> perf_session__process_user_event(). Can we change them to check > >>> NULL and get rid of perf_tool__fill_defaults() to keep it const? > >> > >> As I said above, I don't think that is good style and is out of scope > >> here. It clearly can be done as follow up, but I don't see how that > >> fixes the style issue. > > > > Just to be clear on what the style issue is. We already have code: > > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/builtin-record.c?h=perf-tools-next#n1461 > > ``` > > if (rec->buildid_all && !rec->timestamp_boundary) > > rec->tool.sample = NULL; > > ``` > > that relies on the special behavior of NULL in a function pointer > > being changed at dispatch time - a simple reading of that code would > > be anyone calling the function pointer would get a segv. I'm trying to > > make it so that NULL isn't magic in the context of tool and you can > > simply look at the tool to understand what its behavior is, much as a > > virtual method table would work if we could do proper object-oriented > > development. > > In C, NULL or zero is often used as a special value to mean > no-value. Optional callbacks that are NULL is also not remarkable. Sure NULL means optional and has since the dawn of time, I don't see this adding anything to the conversation. What is remarkable here is passing around a collection of function pointers, then the caller of the function pointer is supposed to check whether what you want to be an optional value (I don't, nor does the existing code) function pointer is present, and if not present the caller needs to know to call another special function. This is complicated enough that the function calls could use helper macros, e.g.: #define CALL_TOOL_SAMPLE(tool, event, sample, evsel, machine) \ if (tool->sample) \ tool->sample(tool, event, sample, machine); \ else \ process_event_sample_stub(tool, event, sample, machine); you need about 38 of them. Now if I'm a buildtin_cmd writer wanting to know what gets called on tool, I need to check, do all callers of the tool use the macro? Perhaps the user decided to do their own NULL test and call a different function. Cases like tool->finished_round already need to go to one of two different stubs. There is clearly more cognitive load on the tool API user as now they must check all uses of 38 function pointers across the code base to understand the behavior of the tool, they can't just rely on the function pointer in the struct being the place the code will go. A simple function pointer call has imo become spaghetti. There is also the overhead checking the optional value - I've never seen a C programmer wanting to do this in 30 years of doing paid C development, the use of the function pointer was to avoid checking, etc. But maybe the kernel does this and it's an existing kernel style? I can't find it. Thanks, Ian ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-07-19 16:26 ` Ian Rogers 2024-07-22 8:28 ` Adrian Hunter @ 2024-08-12 13:53 ` Arnaldo Carvalho de Melo 2024-08-12 18:10 ` Ian Rogers 1 sibling, 1 reply; 40+ messages in thread From: Arnaldo Carvalho de Melo @ 2024-08-12 13:53 UTC (permalink / raw) To: Ian Rogers Cc: Adrian Hunter, Peter Zijlstra, Ingo Molnar, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Fri, Jul 19, 2024 at 09:26:57AM -0700, Ian Rogers wrote: > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > > On 18/07/24 03:59, Ian Rogers wrote: > > > struct perf_tool provides a set of function pointers that are called > > > through when processing perf data. To make filling the pointers less > > > cumbersome, if they are NULL perf_tools__fill_defaults will add > > > default do nothing implementations. > > > > > > This change refactors struct perf_tool to have an init function that > > > provides the default implementation. The special use of NULL and > > > perf_tools__fill_defaults are removed. As a consequence the tool > > > pointers can then all be made const, which better reflects the > > > behavior a particular perf command would expect of the tool and to > > > some extent can reduce the cognitive load on someone working on a > > > command. > > > > > > v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > > > > The tags were really meant only for patch 1, the email that was replied to. > > > > But now for patches 2 and 3: > > > > Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> Applied 1-3, 4 is not applying, I'll look at it later. - Arnaldo > > Sorry for that, you'd mentioned that pt and bts testing which is > impacted by more than just patch 1. > > > Looking at patches 4 to 25, they do not seem to offer any benefit. > > > > Instead of patch 26, presumably perf_tool__fill_defaults() could > > be moved to __perf_session__new(), which perhaps would allow > > patch 27 as it is. > > What I'm trying to do in the series is make it so that the tool isn't > mutated during its use by session. Ideally we'd be passing a const > tool to session_new, that's not possible because there's a hack to fix > ordered events and pipe mode in session__new: > https://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next.git/tree/tools/perf/util/session.c?h=perf-tools-next#n275 > Imo, it isn't great to pass a tool to session__new where you say you > want ordered events and then session just goes to change that for you. > Altering that behavior was beyond the scope of this clean up, so tool > is only const after session__new. > > The reason for doing this is to make it so that when I have a tool I > can reason that nobody is doing things to change it under my feet. My > builtin_cmd is in charge of what the tool is rather than some code > buried in util that thought it was going to do me a favor. The code is > a refactor and so the benefit is intended to be for the developer and > how they reason about the use of tool. We generally use _init > functions rather than having _fill_defaults, so there is a consistency > argument. I don't expect any impact in terms of performance... Moving > perf_tool__fill_defaults to __perf_session__new had issues with the > existing code where NULL would be written over a function pointer > expecting the later fill_defaults to fix it up, doesn't address coding > consistency where _init is the norm, and adds another reason the tool > passed to session__new can't be const. > > Thanks, > Ian > > > > v5: Rebase dropping asan fix merged by Namhyung. > > > v4: Simplify perf_session__deliver_synth_attr_event following Adrian's > > > suggestions. > > > v3: Just remove auxtrace dummy tools [Adrian] and make s390-cpumsf > > > struct removal its own patch [Adrian]. > > > v2: Remove dummy tool initialization [Adrian] and make zero sized. Add > > > cs-etm fix for address sanitizer build, found necessary when > > > testing dummy tool change. > > > > > > Ian Rogers (27): > > > perf auxtrace: Remove dummy tools > > > perf s390-cpumsf: Remove unused struct > > > perf tool: Constify tool pointers > > > perf tool: Move fill defaults into tool.c > > > perf tool: Add perf_tool__init > > > perf kmem: Use perf_tool__init > > > perf buildid-list: Use perf_tool__init > > > perf kvm: Use perf_tool__init > > > perf lock: Use perf_tool__init > > > perf evlist: Use perf_tool__init > > > perf record: Use perf_tool__init > > > perf c2c: Use perf_tool__init > > > perf script: Use perf_tool__init > > > perf inject: Use perf_tool__init > > > perf report: Use perf_tool__init > > > perf stat: Use perf_tool__init > > > perf annotate: Use perf_tool__init > > > perf sched: Use perf_tool__init > > > perf mem: Use perf_tool__init > > > perf timechart: Use perf_tool__init > > > perf diff: Use perf_tool__init > > > perf data convert json: Use perf_tool__init > > > perf data convert ctf: Use perf_tool__init > > > perf test event_update: Ensure tools is initialized > > > perf kwork: Use perf_tool__init > > > perf tool: Remove perf_tool__fill_defaults > > > perf session: Constify tool > > > > > > tools/perf/arch/x86/util/event.c | 4 +- > > > tools/perf/bench/synthesize.c | 2 +- > > > tools/perf/builtin-annotate.c | 44 ++-- > > > tools/perf/builtin-buildid-list.c | 10 + > > > tools/perf/builtin-c2c.c | 33 ++- > > > tools/perf/builtin-diff.c | 30 ++- > > > tools/perf/builtin-evlist.c | 10 +- > > > tools/perf/builtin-inject.c | 159 ++++++------ > > > tools/perf/builtin-kmem.c | 20 +- > > > tools/perf/builtin-kvm.c | 19 +- > > > tools/perf/builtin-kwork.c | 33 ++- > > > tools/perf/builtin-lock.c | 41 ++-- > > > tools/perf/builtin-mem.c | 37 +-- > > > tools/perf/builtin-record.c | 47 ++-- > > > tools/perf/builtin-report.c | 67 +++-- > > > tools/perf/builtin-sched.c | 50 ++-- > > > tools/perf/builtin-script.c | 106 ++++---- > > > tools/perf/builtin-stat.c | 26 +- > > > tools/perf/builtin-timechart.c | 25 +- > > > tools/perf/builtin-top.c | 2 +- > > > tools/perf/builtin-trace.c | 4 +- > > > tools/perf/tests/cpumap.c | 6 +- > > > tools/perf/tests/dlfilter-test.c | 2 +- > > > tools/perf/tests/dwarf-unwind.c | 2 +- > > > tools/perf/tests/event_update.c | 9 +- > > > tools/perf/tests/stat.c | 6 +- > > > tools/perf/tests/thread-map.c | 2 +- > > > tools/perf/util/Build | 1 + > > > tools/perf/util/arm-spe.c | 55 +---- > > > tools/perf/util/auxtrace.c | 12 +- > > > tools/perf/util/auxtrace.h | 20 +- > > > tools/perf/util/bpf-event.c | 4 +- > > > tools/perf/util/build-id.c | 34 +-- > > > tools/perf/util/build-id.h | 8 +- > > > tools/perf/util/cs-etm.c | 39 +-- > > > tools/perf/util/data-convert-bt.c | 34 ++- > > > tools/perf/util/data-convert-json.c | 47 ++-- > > > tools/perf/util/event.c | 54 ++-- > > > tools/perf/util/event.h | 38 +-- > > > tools/perf/util/header.c | 6 +- > > > tools/perf/util/header.h | 4 +- > > > tools/perf/util/hisi-ptt.c | 6 +- > > > tools/perf/util/intel-bts.c | 37 +-- > > > tools/perf/util/intel-pt.c | 30 +-- > > > tools/perf/util/jitdump.c | 4 +- > > > tools/perf/util/s390-cpumsf.c | 11 +- > > > tools/perf/util/session.c | 366 +++------------------------- > > > tools/perf/util/session.h | 9 +- > > > tools/perf/util/synthetic-events.c | 80 +++--- > > > tools/perf/util/synthetic-events.h | 70 +++--- > > > tools/perf/util/tool.c | 294 ++++++++++++++++++++++ > > > tools/perf/util/tool.h | 18 +- > > > tools/perf/util/tsc.c | 2 +- > > > 53 files changed, 977 insertions(+), 1102 deletions(-) > > > create mode 100644 tools/perf/util/tool.c > > > > > ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-08-12 13:53 ` Arnaldo Carvalho de Melo @ 2024-08-12 18:10 ` Ian Rogers 2024-08-12 19:54 ` Arnaldo Carvalho de Melo 0 siblings, 1 reply; 40+ messages in thread From: Ian Rogers @ 2024-08-12 18:10 UTC (permalink / raw) To: Arnaldo Carvalho de Melo Cc: Adrian Hunter, Peter Zijlstra, Ingo Molnar, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Mon, Aug 12, 2024 at 6:53 AM Arnaldo Carvalho de Melo <acme@kernel.org> wrote: > > On Fri, Jul 19, 2024 at 09:26:57AM -0700, Ian Rogers wrote: > > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > > > > On 18/07/24 03:59, Ian Rogers wrote: > > > > struct perf_tool provides a set of function pointers that are called > > > > through when processing perf data. To make filling the pointers less > > > > cumbersome, if they are NULL perf_tools__fill_defaults will add > > > > default do nothing implementations. > > > > > > > > This change refactors struct perf_tool to have an init function that > > > > provides the default implementation. The special use of NULL and > > > > perf_tools__fill_defaults are removed. As a consequence the tool > > > > pointers can then all be made const, which better reflects the > > > > behavior a particular perf command would expect of the tool and to > > > > some extent can reduce the cognitive load on someone working on a > > > > command. > > > > > > > > v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > > > > > > The tags were really meant only for patch 1, the email that was replied to. > > > > > > But now for patches 2 and 3: > > > > > > Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > > Applied 1-3, 4 is not applying, I'll look at it later. I have a rebase and can resend. I haven't addressed Adrian's feedback as I prefer callers of the tool function pointers not to have to do NULL tests, I'm trying to minimize spaghetti. Not sure if you're applying the whole series here or just the beginning. Thanks, Ian ^ permalink raw reply [flat|nested] 40+ messages in thread
* Re: [PATCH v6 00/27] Constify tool pointers 2024-08-12 18:10 ` Ian Rogers @ 2024-08-12 19:54 ` Arnaldo Carvalho de Melo 0 siblings, 0 replies; 40+ messages in thread From: Arnaldo Carvalho de Melo @ 2024-08-12 19:54 UTC (permalink / raw) To: Ian Rogers Cc: Adrian Hunter, Peter Zijlstra, Ingo Molnar, Namhyung Kim, Mark Rutland, Alexander Shishkin, Jiri Olsa, Kan Liang, John Garry, Will Deacon, James Clark, Mike Leach, Leo Yan, Suzuki K Poulose, Yicong Yang, Jonathan Cameron, Nick Terrell, Nick Desaulniers, Oliver Upton, Anshuman Khandual, Song Liu, Ilkka Koskinen, Athira Rajeev, Huacai Chen, Yanteng Si, Sun Haiyong, linux-kernel, linux-perf-users, linux-arm-kernel On Mon, Aug 12, 2024 at 11:10:02AM -0700, Ian Rogers wrote: > On Mon, Aug 12, 2024 at 6:53 AM Arnaldo Carvalho de Melo > <acme@kernel.org> wrote: > > > > On Fri, Jul 19, 2024 at 09:26:57AM -0700, Ian Rogers wrote: > > > On Fri, Jul 19, 2024 at 1:51 AM Adrian Hunter <adrian.hunter@intel.com> wrote: > > > > > > > > On 18/07/24 03:59, Ian Rogers wrote: > > > > > struct perf_tool provides a set of function pointers that are called > > > > > through when processing perf data. To make filling the pointers less > > > > > cumbersome, if they are NULL perf_tools__fill_defaults will add > > > > > default do nothing implementations. > > > > > > > > > > This change refactors struct perf_tool to have an init function that > > > > > provides the default implementation. The special use of NULL and > > > > > perf_tools__fill_defaults are removed. As a consequence the tool > > > > > pointers can then all be made const, which better reflects the > > > > > behavior a particular perf command would expect of the tool and to > > > > > some extent can reduce the cognitive load on someone working on a > > > > > command. > > > > > > > > > > v6: Rebase adding Adrian's reviewed-by/tested-by and Leo's tested-by. > > > > > > > > The tags were really meant only for patch 1, the email that was replied to. > > > > > > > > But now for patches 2 and 3: > > > > > > > > Reviewed-by: Adrian Hunter <adrian.hunter@intel.com> > > > > Applied 1-3, 4 is not applying, I'll look at it later. > > I have a rebase and can resend. I haven't addressed Adrian's feedback Please resend then. - Arnaldo > as I prefer callers of the tool function pointers not to have to do > NULL tests, I'm trying to minimize spaghetti. Not sure if you're > applying the whole series here or just the beginning. > > Thanks, > Ian ^ permalink raw reply [flat|nested] 40+ messages in thread
end of thread, other threads:[~2024-08-12 19:54 UTC | newest] Thread overview: 40+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2024-07-18 0:59 [PATCH v6 00/27] Constify tool pointers Ian Rogers 2024-07-18 0:59 ` [PATCH v6 01/27] perf auxtrace: Remove dummy tools Ian Rogers 2024-07-18 0:59 ` [PATCH v6 02/27] perf s390-cpumsf: Remove unused struct Ian Rogers 2024-07-18 0:59 ` [PATCH v6 03/27] perf tool: Constify tool pointers Ian Rogers 2024-07-18 1:00 ` [PATCH v6 04/27] perf tool: Move fill defaults into tool.c Ian Rogers 2024-07-18 1:00 ` [PATCH v6 05/27] perf tool: Add perf_tool__init Ian Rogers 2024-07-18 1:00 ` [PATCH v6 06/27] perf kmem: Use perf_tool__init Ian Rogers 2024-07-18 1:00 ` [PATCH v6 07/27] perf buildid-list: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 08/27] perf kvm: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 09/27] perf lock: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 10/27] perf evlist: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 11/27] perf record: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 12/27] perf c2c: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 13/27] perf script: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 14/27] perf inject: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 15/27] perf report: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 16/27] perf stat: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 17/27] perf annotate: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 18/27] perf sched: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 19/27] perf mem: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 20/27] perf timechart: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 21/27] perf diff: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 22/27] perf data convert json: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 23/27] perf data convert ctf: " Ian Rogers 2024-07-18 1:00 ` [PATCH v6 24/27] perf test event_update: Ensure tools is initialized Ian Rogers 2024-07-18 1:00 ` [PATCH v6 25/27] perf kwork: Use perf_tool__init Ian Rogers 2024-07-18 1:00 ` [PATCH v6 26/27] perf tool: Remove perf_tool__fill_defaults Ian Rogers 2024-07-18 1:00 ` [PATCH v6 27/27] perf session: Constify tool Ian Rogers 2024-07-19 8:50 ` [PATCH v6 00/27] Constify tool pointers Adrian Hunter 2024-07-19 16:26 ` Ian Rogers 2024-07-22 8:28 ` Adrian Hunter 2024-07-22 16:06 ` Ian Rogers 2024-07-22 17:45 ` Namhyung Kim 2024-07-22 17:50 ` Ian Rogers 2024-07-22 18:04 ` Ian Rogers 2024-07-23 9:38 ` Adrian Hunter 2024-07-23 12:50 ` Ian Rogers 2024-08-12 13:53 ` Arnaldo Carvalho de Melo 2024-08-12 18:10 ` Ian Rogers 2024-08-12 19:54 ` Arnaldo Carvalho de Melo
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).