From d797fdc5c5c245fbb05f553e68cb95d962fbdd01 Mon Sep 17 00:00:00 2001 From: Sam Liao Date: Tue, 7 Jun 2011 23:49:46 +0800 Subject: perf tools: Add inverted call graph report support. Add "caller/callee" option to support inverted butterfly report, in the inverted report (with caller option), the call graph start from the callee's ancestor. Users can use such view to catch system's performance bottleneck from a sysprof like view. Using this option with specified sort order like pid gives us high level view of call graph statistics. Also add "-G" alias for inverted call graph. Signed-off-by: Sam Liao Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Cc: David Ahern Signed-off-by: Frederic Weisbecker --- tools/perf/builtin-report.c | 33 +++++++++++++++++++++++++++------ 1 file changed, 27 insertions(+), 6 deletions(-) (limited to 'tools/perf/builtin-report.c') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 287a173523a..271e252dc65 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -45,7 +45,8 @@ static struct perf_read_values show_threads_values; static const char default_pretty_printing_style[] = "normal"; static const char *pretty_printing_style = default_pretty_printing_style; -static char callchain_default_opt[] = "fractal,0.5"; +static char callchain_default_opt[] = "fractal,0.5,callee"; +static bool inverted_callchain; static symbol_filter_t annotate_init; static int perf_session__add_hist_entry(struct perf_session *session, @@ -386,13 +387,29 @@ parse_callchain_opt(const struct option *opt __used, const char *arg, if (!tok) goto setup; - tok2 = strtok(NULL, ","); callchain_param.min_percent = strtod(tok, &endptr); if (tok == endptr) return -1; - if (tok2) + /* get the print limit */ + tok2 = strtok(NULL, ","); + if (!tok2) + goto setup; + + if (tok2[0] != 'c') { callchain_param.print_limit = strtod(tok2, &endptr); + tok2 = strtok(NULL, ","); + if (!tok2) + goto setup; + } + + /* get the call chain order */ + if (!strcmp(tok2, "caller")) + callchain_param.order = ORDER_CALLER; + else if (!strcmp(tok2, "callee")) + callchain_param.order = ORDER_CALLEE; + else + return -1; setup: if (callchain_register_param(&callchain_param) < 0) { fprintf(stderr, "Can't register callchain params\n"); @@ -436,9 +453,10 @@ static const struct option options[] = { "regex filter to identify parent, see: '--sort parent'"), OPT_BOOLEAN('x', "exclude-other", &symbol_conf.exclude_other, "Only display entries with parent-match"), - OPT_CALLBACK_DEFAULT('g', "call-graph", NULL, "output_type,min_percent", - "Display callchains using output_type (graph, flat, fractal, or none) and min percent threshold. " - "Default: fractal,0.5", &parse_callchain_opt, callchain_default_opt), + OPT_CALLBACK_DEFAULT('g', "call-graph", NULL, "output_type,min_percent, call_order", + "Display callchains using output_type (graph, flat, fractal, or none) , min percent threshold and callchain order. " + "Default: fractal,0.5,callee", &parse_callchain_opt, callchain_default_opt), + OPT_BOOLEAN('G', "inverted", &inverted_callchain, "alias for inverted call graph"), OPT_STRING('d', "dsos", &symbol_conf.dso_list_str, "dso[,dso...]", "only consider symbols in these dsos"), OPT_STRING('C', "comms", &symbol_conf.comm_list_str, "comm[,comm...]", @@ -467,6 +485,9 @@ int cmd_report(int argc, const char **argv, const char *prefix __used) else if (use_tui) use_browser = 1; + if (inverted_callchain) + callchain_param.order = ORDER_CALLER; + if (strcmp(input_name, "-") != 0) setup_browser(true); else -- cgit v1.2.3-70-g09d2 From cb1955b86c86782ff20037da42ef030057501c34 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 29 Jun 2011 23:52:52 +0200 Subject: perf tools: Only display parent field if explictly sorted We don't need to display the parent field if the parent sorting machinery is only used for parent filtering (as in "-p foo"). However if parent filtering is used in combination with explicit parent sorting ( -s parent), we want to display it. Result with: perf report -p kernel_thread -s parent Before: # Overhead Parent symbol # ........ ............. # 0.07% | --- ioread8 ata_sff_check_status ata_sff_tf_load ata_sff_qc_issue ata_bmdma_qc_issue ata_qc_issue ata_scsi_translate ata_scsi_queuecmd scsi_dispatch_cmd scsi_request_fn __blk_run_queue __make_request generic_make_request submit_bio submit_bh journal_submit_commit_record jbd2_journal_commit_transaction kjournald2 kthread kernel_thread_helpe After: # Overhead Parent symbol # ........ ............. # 0.07% kernel_thread_helper | --- ioread8 ata_sff_check_status ata_sff_tf_load ata_sff_qc_issue ata_bmdma_qc_issue ata_qc_issue ata_scsi_translate ata_scsi_queuecmd scsi_dispatch_cmd scsi_request_fn __blk_run_queue __make_request generic_make_request submit_bio submit_bh journal_submit_commit_record jbd2_journal_commit_transaction kjournald2 kthread kernel_thread_helper Signed-off-by: Frederic Weisbecker Cc: Ingo Molnar Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Stephane Eranian Cc: David Ahern Cc: Sam Liao --- tools/perf/builtin-report.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'tools/perf/builtin-report.c') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 271e252dc65..5d43d0181d6 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -525,7 +525,14 @@ int cmd_report(int argc, const char **argv, const char *prefix __used) if (parent_pattern != default_parent_pattern) { if (sort_dimension__add("parent") < 0) return -1; - sort_parent.elide = 1; + + /* + * Only show the parent fields if we explicitly + * sort that way. If we only use parent machinery + * for filtering, we don't want it. + */ + if (!strstr(sort_order, "parent")) + sort_parent.elide = 1; } else symbol_conf.exclude_other = false; -- cgit v1.2.3-70-g09d2 From 5d67be97f8903d05ce53597fb5f3bc25a45e8026 Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Mon, 4 Jul 2011 21:57:50 +1000 Subject: perf report/annotate/script: Add option to specify a CPU range Add an option to perf report/annotate/script to specify which CPUs to operate on. This enables us to take a single system wide profile and analyse each CPU (or group of CPUs) in isolation. This was useful when profiling a multiprocess workload where the bottleneck was on one CPU but this was hidden in the overall profile. Per process and per thread breakdowns didn't help because multiple processes were running on each CPU and no single process consumed an entire CPU. The patch converts the list of CPUs returned by cpu_map__new into a bitmap for fast lookup. I wanted to use -C to be consistent with perf top/record/stat, but unfortunately perf report already uses -C . v2: Incorporate suggestions from David Ahern: - Added -c to perf script - Check that SAMPLE_CPU is set when -c is used - Update documentation v3: Create perf_session__cpu_bitmap() Signed-off-by: Anton Blanchard Acked-by: David Ahern Cc: Arnaldo Carvalho de Melo Cc: Peter Zijlstra Cc: Paul Mackerras Link: http://lkml.kernel.org/r/20110704215750.11647eb9@kryten Signed-off-by: Ingo Molnar --- tools/perf/Documentation/perf-annotate.txt | 6 +++++ tools/perf/Documentation/perf-report.txt | 6 +++++ tools/perf/Documentation/perf-script.txt | 6 +++++ tools/perf/builtin-annotate.c | 15 ++++++++++++ tools/perf/builtin-report.c | 15 ++++++++++++ tools/perf/builtin-script.c | 13 ++++++++++ tools/perf/util/session.c | 38 ++++++++++++++++++++++++++++++ tools/perf/util/session.h | 3 +++ 8 files changed, 102 insertions(+) (limited to 'tools/perf/builtin-report.c') diff --git a/tools/perf/Documentation/perf-annotate.txt b/tools/perf/Documentation/perf-annotate.txt index 6f5a498608b..85c5f026930 100644 --- a/tools/perf/Documentation/perf-annotate.txt +++ b/tools/perf/Documentation/perf-annotate.txt @@ -66,6 +66,12 @@ OPTIONS used. This interfaces starts by centering on the line with more samples, TAB/UNTAB cycles through the lines with more samples. +-c:: +--cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can + be provided as a comma-separated list with no space: 0,1. Ranges of + CPUs are specified with -: 0-2. Default is to report samples on all + CPUs. + SEE ALSO -------- linkperf:perf-record[1], linkperf:perf-report[1] diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt index cfa8e513d0f..04253c07d19 100644 --- a/tools/perf/Documentation/perf-report.txt +++ b/tools/perf/Documentation/perf-report.txt @@ -128,6 +128,12 @@ OPTIONS --symfs=:: Look for files with symbols relative to this directory. +-c:: +--cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can + be provided as a comma-separated list with no space: 0,1. Ranges of + CPUs are specified with -: 0-2. Default is to report samples on all + CPUs. + SEE ALSO -------- linkperf:perf-stat[1] diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt index c6068cb43f5..db017867d9e 100644 --- a/tools/perf/Documentation/perf-script.txt +++ b/tools/perf/Documentation/perf-script.txt @@ -182,6 +182,12 @@ OPTIONS --hide-call-graph:: When printing symbols do not display call chain. +-c:: +--cpu:: Only report samples for the list of CPUs provided. Multiple CPUs can + be provided as a comma-separated list with no space: 0,1. Ranges of + CPUs are specified with -: 0-2. Default is to report samples on all + CPUs. + SEE ALSO -------- linkperf:perf-record[1], linkperf:perf-script-perl[1], diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 7b139e1e7e8..555aefd7fe0 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -28,6 +28,8 @@ #include "util/hist.h" #include "util/session.h" +#include + static char const *input_name = "perf.data"; static bool force, use_tui, use_stdio; @@ -38,6 +40,9 @@ static bool print_line; static const char *sym_hist_filter; +static const char *cpu_list; +static DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS); + static int perf_evlist__add_sample(struct perf_evlist *evlist, struct perf_sample *sample, struct perf_evsel *evsel, @@ -90,6 +95,9 @@ static int process_sample_event(union perf_event *event, return -1; } + if (cpu_list && !test_bit(sample->cpu, cpu_bitmap)) + return 0; + if (!al.filtered && perf_evlist__add_sample(session->evlist, sample, evsel, &al)) { pr_warning("problem incrementing symbol count, " @@ -177,6 +185,12 @@ static int __cmd_annotate(void) if (session == NULL) return -ENOMEM; + if (cpu_list) { + ret = perf_session__cpu_bitmap(session, cpu_list, cpu_bitmap); + if (ret) + goto out_delete; + } + ret = perf_session__process_events(session, &event_ops); if (ret) goto out_delete; @@ -252,6 +266,7 @@ static const struct option options[] = { "print matching source lines (may be slow)"), OPT_BOOLEAN('P', "full-paths", &full_paths, "Don't shorten the displayed pathnames"), + OPT_STRING('c', "cpu", &cpu_list, "cpu", "list of cpus to profile"), OPT_END() }; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 5d43d0181d6..f854efda768 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -33,6 +33,8 @@ #include "util/sort.h" #include "util/hist.h" +#include + static char const *input_name = "perf.data"; static bool force, use_tui, use_stdio; @@ -49,6 +51,9 @@ static char callchain_default_opt[] = "fractal,0.5,callee"; static bool inverted_callchain; static symbol_filter_t annotate_init; +static const char *cpu_list; +static DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS); + static int perf_session__add_hist_entry(struct perf_session *session, struct addr_location *al, struct perf_sample *sample, @@ -117,6 +122,9 @@ static int process_sample_event(union perf_event *event, if (al.filtered || (hide_unresolved && al.sym == NULL)) return 0; + if (cpu_list && !test_bit(sample->cpu, cpu_bitmap)) + return 0; + if (al.map != NULL) al.map->dso->hit = 1; @@ -263,6 +271,12 @@ static int __cmd_report(void) if (session == NULL) return -ENOMEM; + if (cpu_list) { + ret = perf_session__cpu_bitmap(session, cpu_list, cpu_bitmap); + if (ret) + goto out_delete; + } + if (show_threads) perf_read_values_init(&show_threads_values); @@ -473,6 +487,7 @@ static const struct option options[] = { "Only display entries resolved to a symbol"), OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory", "Look for files with symbols relative to this directory"), + OPT_STRING('c', "cpu", &cpu_list, "cpu", "list of cpus to profile"), OPT_END() }; diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 3056b45b3dd..09024ec2ab2 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -13,6 +13,7 @@ #include "util/util.h" #include "util/evlist.h" #include "util/evsel.h" +#include static char const *script_name; static char const *generate_script_lang; @@ -21,6 +22,8 @@ static u64 last_timestamp; static u64 nr_unordered; extern const struct option record_options[]; static bool no_callchain; +static const char *cpu_list; +static DECLARE_BITMAP(cpu_bitmap, MAX_NR_CPUS); enum perf_output_field { PERF_OUTPUT_COMM = 1U << 0, @@ -453,6 +456,10 @@ static int process_sample_event(union perf_event *event, last_timestamp = sample->time; return 0; } + + if (cpu_list && !test_bit(sample->cpu, cpu_bitmap)) + return 0; + scripting_ops->process_event(event, sample, evsel, session, thread); session->hists.stats.total_period += sample->period; @@ -1075,6 +1082,7 @@ static const struct option options[] = { OPT_CALLBACK('f', "fields", NULL, "str", "comma separated output fields prepend with 'type:'. Valid types: hw,sw,trace,raw. Fields: comm,tid,pid,time,cpu,event,trace,ip,sym,dso,addr", parse_output_fields), + OPT_STRING('c', "cpu", &cpu_list, "cpu", "list of cpus to profile"), OPT_END() }; @@ -1255,6 +1263,11 @@ int cmd_script(int argc, const char **argv, const char *prefix __used) if (session == NULL) return -ENOMEM; + if (cpu_list) { + if (perf_session__cpu_bitmap(session, cpu_list, cpu_bitmap)) + return -1; + } + if (!no_callchain) symbol_conf.use_callchain = true; else diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 558bcf99694..080e5336d89 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -12,6 +12,7 @@ #include "session.h" #include "sort.h" #include "util.h" +#include "cpumap.h" static int perf_session__open(struct perf_session *self, bool force) { @@ -1282,3 +1283,40 @@ void perf_session__print_ip(union perf_event *event, } } } + +int perf_session__cpu_bitmap(struct perf_session *session, + const char *cpu_list, unsigned long *cpu_bitmap) +{ + int i; + struct cpu_map *map; + + for (i = 0; i < PERF_TYPE_MAX; ++i) { + struct perf_evsel *evsel; + + evsel = perf_session__find_first_evtype(session, i); + if (!evsel) + continue; + + if (!(evsel->attr.sample_type & PERF_SAMPLE_CPU)) { + pr_err("File does not contain CPU events. " + "Remove -c option to proceed.\n"); + return -1; + } + } + + map = cpu_map__new(cpu_list); + + for (i = 0; i < map->nr; i++) { + int cpu = map->map[i]; + + if (cpu >= MAX_NR_CPUS) { + pr_err("Requested CPU %d too large. " + "Consider raising MAX_NR_CPUS\n", cpu); + return -1; + } + + set_bit(cpu, cpu_bitmap); + } + + return 0; +} diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index de4178d7bb7..5de754f4b7f 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -172,4 +172,7 @@ void perf_session__print_ip(union perf_event *event, struct perf_session *session, int print_sym, int print_dso); +int perf_session__cpu_bitmap(struct perf_session *session, + const char *cpu_list, unsigned long *cpu_bitmap); + #endif /* __PERF_SESSION_H */ -- cgit v1.2.3-70-g09d2