perf tools: Allow to control synthesize during record
Namhyung Kim
maintainerPATCH 1/2 perf tools: Allow to control synthesize during record
Depending on the use case, it might require some kind of synthesize and some not. Make it controllable to turn off heavy operations like MMAP for all tasks.
Currently all users are converted to enable all the synthesis by default. It'll be updated in the later patch.
--- tools/perf/bench/synthesize.c | 4 +-- tools/perf/builtin-kvm.c | 2 +- tools/perf/builtin-record.c | 6 ++-- tools/perf/builtin-top.c | 2 +- tools/perf/builtin-trace.c | 4 +-- tools/perf/tests/code-reading.c | 3 +- tools/perf/tests/mmap-thread-lookup.c | 4 +-- tools/perf/util/synthetic-events.c | 45 ++++++++++++++++----------- tools/perf/util/synthetic-events.h | 8 ++--- 9 files changed, 44 insertions(+), 34 deletions(-)
tools/perf/bench/synthesize.c +2 -2
@@ -80,7 +80,7 @@ static int do_run_single_threaded(struct perf_session *session,NULL , target, threads, process_synthesized_event, - data_mmap, + true, data_mmap, nr_threads_synthesize);if (err)return err;
@@ -171,7 +171,7 @@ static int do_run_multi_threaded(struct target *target,NULL , target,NULL , process_synthesized_event, - false, + true, false, nr_threads_synthesize);if (err) {perf_session__delete (session);
tools/perf/builtin-kvm.c +1 -1
@@ -1456,7 +1456,7 @@ static int kvm_events_live(struct perf_kvm_stat *kvm,perf_session__set_id_hdr_size (kvm->session);ordered_events__set_copy_on_queue (&kvm->session->ordered_events, true);machine__synthesize_threads (&kvm->session->machines.host, &kvm->opts.target, - kvm->evlist->core.threads, false,1 ); + kvm->evlist->core.threads, true, false,1 ); err =kvm_live_open_events (kvm);if (err) goto out;
tools/perf/builtin-record.c +4 -2
@@ -1265,6 +1265,7 @@ static int record__synthesize_workload(struct record *rec, bool tail) err =perf_event__synthesize_thread_map (&rec->tool, thread_map, process_synthesized_event, &rec->session->machines.host, + true, rec->opts.sample_address);perf_thread_map__put (thread_map);return err;
@@ -1509,8 +1510,9 @@ static int record__synthesize(struct record *rec, bool tail) f = process_locked_synthesized_event; } - err =__machine__synthesize_threads (machine, tool, &opts->target, rec->evlist->core.threads, - f, opts->sample_address, + err =__machine__synthesize_threads (machine, tool, &opts->target, + rec->evlist->core.threads, + f, true, opts->sample_address, rec->opts.nr_threads_synthesize);if (rec->opts.nr_threads_synthesize >1 )
tools/perf/builtin-top.c +1 -1
@@ -1271,7 +1271,7 @@ static int __cmd_top(struct perf_top *top)pr_debug ("Couldn't synthesize cgroup events.\n" );machine__synthesize_threads (&top->session->machines.host, &opts->target, - top->evlist->core.threads, false, + top->evlist->core.threads, true, false, top->nr_threads_synthesize);if (top->nr_threads_synthesize >1 )
tools/perf/builtin-trace.c +2 -2
@@ -1620,8 +1620,8 @@ static int trace__symbols_init(struct trace *trace, struct evlist *evlist) goto out; err =__machine__synthesize_threads (trace->host, &trace->tool, &trace->opts.target, - evlist->core.threads, trace__tool_process, false, -1 ); + evlist->core.threads, trace__tool_process, + true, false,1 ); out:if (err)symbol__exit ();
tools/perf/tests/code-reading.c +2 -1
@@ -606,7 +606,8 @@ static int do_test_code_reading(bool try_kcore) } ret =perf_event__synthesize_thread_map (NULL , threads, - perf_event__process, machine, false); + perf_event__process, machine, + true, false);if (ret <0 ) {pr_debug ("perf_event__synthesize_thread_map failed\n" ); goto out_err;
tools/perf/tests/mmap-thread-lookup.c +2 -2
@@ -135,7 +135,7 @@ static int synth_all(struct machine *machine) {return perf_event__synthesize_threads (NULL , perf_event__process, - machine,0 ,1 ); + machine,1 ,0 ,1 ); }static int synth_process (struct machine *machine)
@@ -147,7 +147,7 @@ static int synth_process(struct machine *machine) err =perf_event__synthesize_thread_map (NULL , map, perf_event__process, - machine,0 ); + machine,1 ,0 );perf_thread_map__put (map);return err;
tools/perf/util/synthetic-events.c +26 -19
140 lines not on this page · raw message
tools/perf/util/synthetic-events.h +4 -4
@@ -52,8 +52,8 @@ int perf_event__synthesize_stat_events(struct perf_stat_config *config, struct pint 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,u32 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 mmap_data); -int perf_event__synthesize_threads (struct perf_tool *tool,perf_event__handler_t process,struct machine *machine,bool mmap_data,unsigned int nr_threads_synthesize); +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);
@@ -64,10 +64,10 @@ size_t perf_event__sample_event_size(const struct perf_sample *sample, u64 type,int __machine__synthesize_threads (struct machine *machine,struct perf_tool *tool,struct target *target,struct perf_thread_map *threads, -perf_event__handler_t process,bool data_mmap, +perf_event__handler_t process,bool needs_mmap,bool data_mmap,unsigned int nr_threads_synthesize);int machine__synthesize_threads (struct machine *machine,struct target *target, -struct perf_thread_map *threads,bool data_mmap, +struct perf_thread_map *threads,bool needs_mmap,bool data_mmap,unsigned int nr_threads_synthesize);#ifdef HAVE_AUXTRACE_SUPPORT
Signature · 1 line
2.32.0.432.gabb21c7263-goog
Namhyung Kim
maintainerPATCH 2/2 perf record: Add --synth option
Add an option to control synthesize behavior.
--synth <no|all|task|mmap|cgroup>
Fine-tune event synthesis: default=all
This can be useful when we know it doesn't need some synthesis like in a specific usecase and/or when using pipe:
$ perf record -a --all-cgroups --synth cgroup -o- sleep 1 | \
Quoted 1 line
perf report -i- -s cgroup
--- tools/perf/Documentation/perf-record.txt | 4 ++ tools/perf/builtin-record.c | 48 +++++++++++++++++++----- tools/perf/util/record.h | 1 + tools/perf/util/synthetic-events.c | 28 ++++++++++++++ tools/perf/util/synthetic-events.h | 12 ++++++ 5 files changed, 84 insertions(+), 9 deletions(-)
tools/perf/Documentation/perf-record.txt +4 -0
@@ -596,6 +596,10 @@ options. 'perf record --dry-run -e' can act as a BPF script compiler if llvm.dump-obj in config file is set to true. +--synth=TYPE:: +Collect and synthesize given type of events. Available types are: +task, mmap, cgroup, all, no + --tail-synthesize:: Instead of collecting non-sample events (for example, fork, comm, mmap) at the beginning of record, collect them during finalizing an output file.
tools/perf/builtin-record.c +39 -9
90 lines not on this page · raw message
tools/perf/util/record.h +1 -0
@@ -77,6 +77,7 @@ struct record_opts {int ctl_fd;int ctl_fd_ack;bool ctl_fd_close; +int synth; };extern const char *const *record_usage;
tools/perf/util/synthetic-events.c +28 -0
@@ -2186,3 +2186,31 @@ int perf_event__synthesize_features(struct perf_tool *tool, struct perf_sessionfree (ff.buf);return ret; } + +int parse_synth_opt (char *synth) +{ +char *p, *q; +int ret =0 ; + +if (synth ==NULL ) +return -1 ; + +for (q = synth; (p =strsep (&q,"," )); p = q) { +if (!strcasecmp (p,"no" ) || !strcasecmp (p,"none" )) +return 0 ; + +if (!strcasecmp (p,"all" )) +return PERF_SYNTH_ALL ; + +if (!strcasecmp (p,"task" )) + ret |=PERF_SYNTH_TASK ; +else if (!strcasecmp (p,"mmap" )) + ret |=PERF_SYNTH_TASK |PERF_SYNTH_MMAP ; +else if (!strcasecmp (p,"cgroup" )) + ret |=PERF_SYNTH_CGROUP ; +else +return -1 ; + } + +return ret; +}
tools/perf/util/synthetic-events.h +12 -0
@@ -26,6 +26,18 @@ struct target;union perf_event ; +enum perf_record_synth { +PERF_SYNTH_TASK =1 <<0 , +PERF_SYNTH_MMAP =1 <<1 , +PERF_SYNTH_CGROUP =1 <<2 , + +/* last element */ +PERF_SYNTH_MAX =1 <<3 , +}; +#define PERF_SYNTH_ALL (PERF_SYNTH_MAX - 1) + +int parse_synth_opt (char *str); +typedef int (*perf_event__handler_t )(struct perf_tool *tool,union perf_event *event,struct perf_sample *sample,struct machine *machine);
Signature · 1 line
2.32.0.432.gabb21c7263-goog
Re: [PATCH 2/2] perf record: Add --synth option
Quoted 26 lines · Namhyung Kim
Add an option to control synthesize behavior.
--synth <no|all|task|mmap|cgroup>
Fine-tune event synthesis: default=allThis can be useful when we know it doesn't need some synthesis like in a specific usecase and/or when using pipe:
$ perf record -a --all-cgroups --synth cgroup -o- sleep 1 | \Quoted 1 line
perf report -i- -s cgroup
Signed-off-by Namhyung Kim <[email protected]>--- tools/perf/Documentation/perf-record.txt | 4 ++ tools/perf/builtin-record.c | 48 +++++++++++++++++++----- tools/perf/util/record.h | 1 + tools/perf/util/synthetic-events.c | 28 ++++++++++++++ tools/perf/util/synthetic-events.h | 12 ++++++ 5 files changed, 84 insertions(+), 9 deletions(-)tools/perf/Documentation/perf-record.txt +0
-0
tools/perf/Documentation/perf-record.txt options. · line 602
Quoted 7 lines
tools/perf/Documentation/perf-record.txt +4
-0@@ -596,6 +596,10 @@ options. 'perf record --dry-run -e' can act as a BPF script compiler if llvm.dump-obj in config file is set to true. +--synth=TYPE:: +Collect and synthesize given type of events. Available types are: +task, mmap, cgroup, all, no +
I would explain what each option disables in the synthesis. It is not clear in your description
or your explain in the commit msg.
Quoted 178 lines
tools/perf/Documentation/perf-record.txt +0
-0--tail-synthesize:: Instead of collecting non-sample events (for example, fork, comm, mmap) at the beginning of record, collect them during finalizing an output file.tools/perf/builtin-record.c +39
-9@@ -1254,6 +1254,7 @@ static int record__synthesize_workload(struct record *rec, bool tail) {int err;struct perf_thread_map *thread_map; +bool needs_mmap = rec->opts.synth &PERF_SYNTH_MMAP ;if (rec->opts.tail_synthesize != tail)return 0 ;@@ -1265,7 +1266,7 @@ static int record__synthesize_workload(struct record *rec, bool tail) err =perf_event__synthesize_thread_map (&rec->tool, thread_map, process_synthesized_event, &rec->session->machines.host, - true, + needs_mmap, rec->opts.sample_address);perf_thread_map__put (thread_map);return err;@@ -1500,20 +1501,26 @@ static int record__synthesize(struct record *rec, bool tail)if (err <0 )pr_warning ("Couldn't synthesize bpf events.\n" ); - err =perf_event__synthesize_cgroups (tool, process_synthesized_event, - machine); -if (err <0 ) -pr_warning ("Couldn't synthesize cgroup events.\n" ); +if (rec->opts.synth &PERF_SYNTH_CGROUP ) { + err =perf_event__synthesize_cgroups (tool, process_synthesized_event, + machine); +if (err <0 ) +pr_warning ("Couldn't synthesize cgroup events.\n" ); + }if (rec->opts.nr_threads_synthesize >1 ) {perf_set_multithreaded (); f = process_locked_synthesized_event; } - err =__machine__synthesize_threads (machine, tool, &opts->target, - rec->evlist->core.threads, - f, true, opts->sample_address, - rec->opts.nr_threads_synthesize); +if (rec->opts.synth &PERF_SYNTH_TASK ) { +bool needs_mmap = rec->opts.synth &PERF_SYNTH_MMAP ; + + err =__machine__synthesize_threads (machine, tool, &opts->target, + rec->evlist->core.threads, + f, needs_mmap, opts->sample_address, + rec->opts.nr_threads_synthesize); + }if (rec->opts.nr_threads_synthesize >1 )perf_set_singlethreaded ();@@ -2422,6 +2429,26 @@ static int process_timestamp_boundary(struct perf_tool *tool,return 0 ; } +static int parse_record_synth_option (const struct option *opt, +const char *str, +int unset__maybe_unused ) +{ +struct record_opts *opts = opt->value; +char *p =strdup (str); + +if (p ==NULL ) +return -1 ; + + opts->synth =parse_synth_opt (p); +free (p); + +if (opts->synth <0 ) { +pr_err ("Invalid synth option: %s\n" , str); +return -1 ; + } +return 0 ; +} +/* * XXX Ideally would be local to cmd_record() and passed to a record__new * because we need to have access to it in record__exit, that is called @@ -2447,6 +2474,7 @@ static struct record record = { .nr_threads_synthesize = 1, .ctl_fd = -1, .ctl_fd_ack = -1, + .synth = PERF_SYNTH_ALL, }, .tool = { .sample = process_sample_event,@@ -2662,6 +2690,8 @@ static struct option __record_options[] = { "\t\t\t Optionally send control command completion ('ack\\n') to ack-fd descriptor.\n" "\t\t\t Alternatively, ctl-fifo / ack-fifo will be opened and used as ctl-fd / ack-fd.", parse_control_option), + OPT_CALLBACK(0, "synth", &record.opts, "no|all|task|mmap|cgroup", + "Fine-tune event synthesis: default=all", parse_record_synth_option), OPT_END() };tools/perf/util/record.h +1
-0@@ -77,6 +77,7 @@ struct record_opts {int ctl_fd;int ctl_fd_ack;bool ctl_fd_close; +int synth; };extern const char *const *record_usage;tools/perf/util/synthetic-events.c +28
-0@@ -2186,3 +2186,31 @@ int perf_event__synthesize_features(struct perf_tool *tool, struct perf_sessionfree (ff.buf);return ret; } + +int parse_synth_opt (char *synth) +{ +char *p, *q; +int ret =0 ; + +if (synth ==NULL ) +return -1 ; + +for (q = synth; (p =strsep (&q,"," )); p = q) { +if (!strcasecmp (p,"no" ) || !strcasecmp (p,"none" )) +return 0 ; + +if (!strcasecmp (p,"all" )) +return PERF_SYNTH_ALL ; + +if (!strcasecmp (p,"task" )) + ret |=PERF_SYNTH_TASK ; +else if (!strcasecmp (p,"mmap" )) + ret |=PERF_SYNTH_TASK |PERF_SYNTH_MMAP ; +else if (!strcasecmp (p,"cgroup" )) + ret |=PERF_SYNTH_CGROUP ; +else +return -1 ; + } + +return ret; +}tools/perf/util/synthetic-events.h +12
-1@@ -26,6 +26,18 @@ struct target;union perf_event ; +enum perf_record_synth { +PERF_SYNTH_TASK =1 <<0 , +PERF_SYNTH_MMAP =1 <<1 , +PERF_SYNTH_CGROUP =1 <<2 , + +/* last element */ +PERF_SYNTH_MAX =1 <<3 , +}; +#define PERF_SYNTH_ALL (PERF_SYNTH_MAX - 1) + +int parse_synth_opt (char *str); +typedef int (*perf_event__handler_t )(struct perf_tool *tool,union perf_event *event,struct perf_sample *sample,struct machine *machine); --2.32.0.432.gabb21c7263-goog
Namhyung Kim
maintainerRe: [PATCH 2/2] perf record: Add --synth option
Quoted 31 lines · Stephane Eranian
Quoted 29 lines · Namhyung Kim
Add an option to control synthesize behavior.
--synth <no|all|task|mmap|cgroup>
Fine-tune event synthesis: default=allThis can be useful when we know it doesn't need some synthesis like in a specific usecase and/or when using pipe:
$ perf record -a --all-cgroups --synth cgroup -o- sleep 1 | \Quoted 1 line
perf report -i- -s cgroup
Signed-off-by Namhyung Kim <[email protected]>--- tools/perf/Documentation/perf-record.txt | 4 ++ tools/perf/builtin-record.c | 48 +++++++++++++++++++----- tools/perf/util/record.h | 1 + tools/perf/util/synthetic-events.c | 28 ++++++++++++++ tools/perf/util/synthetic-events.h | 12 ++++++ 5 files changed, 84 insertions(+), 9 deletions(-)tools/perf/Documentation/perf-record.txt +0
-0@@ -596,6 +596,10 @@ options. 'perf record --dry-run -e' can act as a BPF script compiler if llvm.dump-obj in config file is set to true.
Quoted 8 lines
tools/perf/Documentation/perf-record.txt options. · line 602
Quoted 4 lines
tools/perf/Documentation/perf-record.txt +4
-0+--synth=TYPE:: +Collect and synthesize given type of events. Available types are: +task, mmap, cgroup, all, no +I would explain what each option disables in the synthesis. It is not clear in your description
or your explain in the commit msg.
Ok, will update.
Thanks,
Namhyung
Quoted 180 lines
Quoted 180 lines
tools/perf/Documentation/perf-record.txt +0
-0--tail-synthesize:: Instead of collecting non-sample events (for example, fork, comm, mmap) at the beginning of record, collect them during finalizing an output file.tools/perf/builtin-record.c +39
-9@@ -1254,6 +1254,7 @@ static int record__synthesize_workload(struct record *rec, bool tail) {int err;struct perf_thread_map *thread_map; +bool needs_mmap = rec->opts.synth &PERF_SYNTH_MMAP ;if (rec->opts.tail_synthesize != tail)return 0 ;@@ -1265,7 +1266,7 @@ static int record__synthesize_workload(struct record *rec, bool tail) err =perf_event__synthesize_thread_map (&rec->tool, thread_map, process_synthesized_event, &rec->session->machines.host, - true, + needs_mmap, rec->opts.sample_address);perf_thread_map__put (thread_map);return err;@@ -1500,20 +1501,26 @@ static int record__synthesize(struct record *rec, bool tail)if (err <0 )pr_warning ("Couldn't synthesize bpf events.\n" ); - err =perf_event__synthesize_cgroups (tool, process_synthesized_event, - machine); -if (err <0 ) -pr_warning ("Couldn't synthesize cgroup events.\n" ); +if (rec->opts.synth &PERF_SYNTH_CGROUP ) { + err =perf_event__synthesize_cgroups (tool, process_synthesized_event, + machine); +if (err <0 ) +pr_warning ("Couldn't synthesize cgroup events.\n" ); + }if (rec->opts.nr_threads_synthesize >1 ) {perf_set_multithreaded (); f = process_locked_synthesized_event; } - err =__machine__synthesize_threads (machine, tool, &opts->target, - rec->evlist->core.threads, - f, true, opts->sample_address, - rec->opts.nr_threads_synthesize); +if (rec->opts.synth &PERF_SYNTH_TASK ) { +bool needs_mmap = rec->opts.synth &PERF_SYNTH_MMAP ; + + err =__machine__synthesize_threads (machine, tool, &opts->target, + rec->evlist->core.threads, + f, needs_mmap, opts->sample_address, + rec->opts.nr_threads_synthesize); + }if (rec->opts.nr_threads_synthesize >1 )perf_set_singlethreaded ();@@ -2422,6 +2429,26 @@ static int process_timestamp_boundary(struct perf_tool *tool,return 0 ; } +static int parse_record_synth_option (const struct option *opt, +const char *str, +int unset__maybe_unused ) +{ +struct record_opts *opts = opt->value; +char *p =strdup (str); + +if (p ==NULL ) +return -1 ; + + opts->synth =parse_synth_opt (p); +free (p); + +if (opts->synth <0 ) { +pr_err ("Invalid synth option: %s\n" , str); +return -1 ; + } +return 0 ; +} +/* * XXX Ideally would be local to cmd_record() and passed to a record__new * because we need to have access to it in record__exit, that is called @@ -2447,6 +2474,7 @@ static struct record record = { .nr_threads_synthesize = 1, .ctl_fd = -1, .ctl_fd_ack = -1, + .synth = PERF_SYNTH_ALL, }, .tool = { .sample = process_sample_event,@@ -2662,6 +2690,8 @@ static struct option __record_options[] = { "\t\t\t Optionally send control command completion ('ack\\n') to ack-fd descriptor.\n" "\t\t\t Alternatively, ctl-fifo / ack-fifo will be opened and used as ctl-fd / ack-fd.", parse_control_option), + OPT_CALLBACK(0, "synth", &record.opts, "no|all|task|mmap|cgroup", + "Fine-tune event synthesis: default=all", parse_record_synth_option), OPT_END() };tools/perf/util/record.h +1
-0@@ -77,6 +77,7 @@ struct record_opts {int ctl_fd;int ctl_fd_ack;bool ctl_fd_close; +int synth; };extern const char *const *record_usage;tools/perf/util/synthetic-events.c +28
-0@@ -2186,3 +2186,31 @@ int perf_event__synthesize_features(struct perf_tool *tool, struct perf_sessionfree (ff.buf);return ret; } + +int parse_synth_opt (char *synth) +{ +char *p, *q; +int ret =0 ; + +if (synth ==NULL ) +return -1 ; + +for (q = synth; (p =strsep (&q,"," )); p = q) { +if (!strcasecmp (p,"no" ) || !strcasecmp (p,"none" )) +return 0 ; + +if (!strcasecmp (p,"all" )) +return PERF_SYNTH_ALL ; + +if (!strcasecmp (p,"task" )) + ret |=PERF_SYNTH_TASK ; +else if (!strcasecmp (p,"mmap" )) + ret |=PERF_SYNTH_TASK |PERF_SYNTH_MMAP ; +else if (!strcasecmp (p,"cgroup" )) + ret |=PERF_SYNTH_CGROUP ; +else +return -1 ; + } + +return ret; +}tools/perf/util/synthetic-events.h +12
-1@@ -26,6 +26,18 @@ struct target;union perf_event ; +enum perf_record_synth { +PERF_SYNTH_TASK =1 <<0 , +PERF_SYNTH_MMAP =1 <<1 , +PERF_SYNTH_CGROUP =1 <<2 , + +/* last element */ +PERF_SYNTH_MAX =1 <<3 , +}; +#define PERF_SYNTH_ALL (PERF_SYNTH_MAX - 1) + +int parse_synth_opt (char *str); +typedef int (*perf_event__handler_t )(struct perf_tool *tool,union perf_event *event,struct perf_sample *sample,struct machine *machine); --2.32.0.432.gabb21c7263-goog