Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753811AbbHRRgp (ORCPT ); Tue, 18 Aug 2015 13:36:45 -0400 Received: from mail-wi0-f171.google.com ([209.85.212.171]:33890 "EHLO mail-wi0-f171.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752974AbbHRRgn (ORCPT ); Tue, 18 Aug 2015 13:36:43 -0400 MIME-Version: 1.0 In-Reply-To: <1439889946-28986-4-git-send-email-kan.liang@intel.com> References: <1439889946-28986-1-git-send-email-kan.liang@intel.com> <1439889946-28986-4-git-send-email-kan.liang@intel.com> Date: Tue, 18 Aug 2015 10:36:41 -0700 Message-ID: Subject: Re: [PATCH RFC 03/10] perf,tools: support option --socket From: Stephane Eranian To: "Liang, Kan" Cc: Arnaldo Carvalho de Melo , Peter Zijlstra , "mingo@redhat.com" , Jiri Olsa , Namhyung Kim , "ak@linux.intel.com" , LKML Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 11412 Lines: 295 On Tue, Aug 18, 2015 at 2:25 AM, wrote: > From: Kan Liang > > Introduce a new option for perf report to show the event information in > the same socket together. > When this option is set, perf report will force to sort by socket. > I would say processor socket, otherwise it is a bit confusing. At first, I thought you were talking about network sockets. > $ perf report --stdio --socket > # To display the perf.data header info, please use > --header/--header-only options. > # > # > # Total Lost Samples: 0 > # > # Samples: 686 of event 'cycles' > # Event count (approx.): 349215462 > # > # > # Socket: 0 > # Processor Socket: > # Overhead Command Shared Object Symbol > # ........ ......... ................ > ................................. > # > 97.05% test test [.] plusB_c > 0.98% test test [.] plusA_c > 0.16% test [kernel.vmlinux] [k] add_mm_counter_fast > 0.15% swapper [kernel.vmlinux] [k] note_gp_changes > 0.15% perf [kernel.vmlinux] [k] unmap_single_vma > 0.06% perf [kernel.vmlinux] [k] run_timer_softirq > 0.00% swapper [kernel.vmlinux] [k] native_write_msr > # > # Socket: 1 > # > # Overhead Command Shared Object Symbol > # ........ ......... ................ > ................................. > # > 0.93% perf [kernel.vmlinux] [k] smp_call_function_single > 0.19% perf [kernel.vmlinux] [k] page_fault > 0.19% swapper [kernel.vmlinux] [k] pm_qos_request > 0.12% rcu_sched [kernel.vmlinux] [k] > dyntick_save_progress_counter > 0.00% swapper [kernel.vmlinux] [k] wake_up_process > 0.00% swapper [kernel.vmlinux] [k] __do_softirq > 0.00% swapper [kernel.vmlinux] [k] run_timer_softirq > 0.00% swapper [kernel.vmlinux] [k] native_write_msr > 0.00% perf [kernel.vmlinux] [k] native_write_msr > > Signed-off-by: Kan Liang > --- > tools/perf/Documentation/perf-report.txt | 3 +++ > tools/perf/builtin-diff.c | 2 +- > tools/perf/builtin-report.c | 17 ++++++++++++++++- > tools/perf/builtin-top.c | 2 +- > tools/perf/ui/stdio/hist.c | 14 ++++++++++---- > tools/perf/util/cpumap.c | 14 ++++++++++++++ > tools/perf/util/cpumap.h | 2 ++ > tools/perf/util/hist.h | 2 +- > tools/perf/util/sort.c | 6 ++++++ > tools/perf/util/symbol.h | 1 + > 10 files changed, 55 insertions(+), 8 deletions(-) > > diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt > index ca76b0d..49a42e4 100644 > --- a/tools/perf/Documentation/perf-report.txt > +++ b/tools/perf/Documentation/perf-report.txt > @@ -293,6 +293,9 @@ OPTIONS > --group:: > Show event group information together. > > +--socket:: > + Show event information in the same socket together. > + > --demangle:: > Demangle symbol names to human readable form. It's enabled by default, > disable with --no-demangle. > diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c > index 0b180a8..0dfd91f 100644 > --- a/tools/perf/builtin-diff.c > +++ b/tools/perf/builtin-diff.c > @@ -667,7 +667,7 @@ static void hists__process(struct hists *hists) > hists__precompute(hists); > hists__output_resort(hists, NULL); > > - hists__fprintf(hists, true, 0, 0, 0, stdout); > + hists__fprintf(hists, true, 0, 0, 0, stdout, -1); > } > > static void data__fprintf(void) > diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c > index 62b285e..6fdf9f4 100644 > --- a/tools/perf/builtin-report.c > +++ b/tools/perf/builtin-report.c > @@ -345,7 +345,17 @@ static int perf_evlist__tty_browse_hists(struct perf_evlist *evlist, > continue; > > hists__fprintf_nr_sample_events(hists, rep, evname, stdout); > - hists__fprintf(hists, true, 0, 0, rep->min_percent, stdout); > + > + if (symbol_conf.socket) { > + int i; > + > + for (i = 0; i < max_socket_num; i++) { > + fprintf(stdout, "#\n# Socket: %d\n#\n", i); > + hists__fprintf(hists, true, 0, 0, rep->min_percent, stdout, i); > + } > + } else > + hists__fprintf(hists, true, 0, 0, rep->min_percent, stdout, -1); > + > fprintf(stdout, "\n\n"); > } > > @@ -724,6 +734,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused) > "Show a column with the sum of periods"), > OPT_BOOLEAN(0, "group", &symbol_conf.event_group, > "Show event group information together"), > + OPT_BOOLEAN(0, "socket", &symbol_conf.socket, > + "Show event information in the same socket together"), > OPT_CALLBACK_NOOPT('b', "branch-stack", &branch_mode, "", > "use branch records for per branch histogram filling", > parse_branch_mode), > @@ -909,6 +921,9 @@ repeat: > > sort__setup_elide(stdout); > > + if (symbol_conf.socket) > + set_max_socket_num(); > + > ret = __cmd_report(&report); > if (ret == K_SWITCH_INPUT_DATA) { > perf_session__delete(session); > diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c > index bfe24f1..deffe44 100644 > --- a/tools/perf/builtin-top.c > +++ b/tools/perf/builtin-top.c > @@ -295,7 +295,7 @@ static void perf_top__print_sym_table(struct perf_top *top) > hists__output_recalc_col_len(hists, top->print_entries - printed); > putchar('\n'); > hists__fprintf(hists, false, top->print_entries - printed, win_width, > - top->min_percent, stdout); > + top->min_percent, stdout, -1); > } > > static void prompt_integer(int *target, const char *msg) > diff --git a/tools/perf/ui/stdio/hist.c b/tools/perf/ui/stdio/hist.c > index dfcbc90..2f512b8 100644 > --- a/tools/perf/ui/stdio/hist.c > +++ b/tools/perf/ui/stdio/hist.c > @@ -323,7 +323,8 @@ static int hist_entry__snprintf(struct hist_entry *he, struct perf_hpp *hpp) > return 0; > > perf_hpp__for_each_format(fmt) { > - if (perf_hpp__should_skip(fmt)) > + if (perf_hpp__should_skip(fmt) || > + !strcmp(fmt->name, "SOCKET")) > continue; > > /* > @@ -371,7 +372,7 @@ static int hist_entry__fprintf(struct hist_entry *he, size_t size, > } > > size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows, > - int max_cols, float min_pcnt, FILE *fp) > + int max_cols, float min_pcnt, FILE *fp, int socket) > { > struct perf_hpp_fmt *fmt; > struct rb_node *nd; > @@ -402,7 +403,8 @@ size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows, > fprintf(fp, "# "); > > perf_hpp__for_each_format(fmt) { > - if (perf_hpp__should_skip(fmt)) > + if (perf_hpp__should_skip(fmt) || > + !strcmp(fmt->name, "SOCKET")) > continue; > > if (!first) > @@ -428,7 +430,8 @@ size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows, > perf_hpp__for_each_format(fmt) { > unsigned int i; > > - if (perf_hpp__should_skip(fmt)) > + if (perf_hpp__should_skip(fmt) || > + !strcmp(fmt->name, "SOCKET")) > continue; > > if (!first) > @@ -465,6 +468,9 @@ print_entries: > if (h->filtered) > continue; > > + if ((socket >= 0) && (cpu__get_socket(h->cpu) != socket)) > + continue; > + > percent = hist_entry__get_percent_limit(h); > if (percent < min_pcnt) > continue; > diff --git a/tools/perf/util/cpumap.c b/tools/perf/util/cpumap.c > index 7f25e6c..ae03426 100644 > --- a/tools/perf/util/cpumap.c > +++ b/tools/perf/util/cpumap.c > @@ -407,6 +407,20 @@ out: > pr_err("Failed to read max cpus, using default of %d\n", max_cpu_num); > } > > +void set_max_socket_num(void) > +{ > + int cpu, socket; > + > + set_max_cpu_num(); > + > + max_socket_num = 1; > + for (cpu = 0; cpu < max_cpu_num; cpu++) { > + socket = cpu__get_socket(cpu); > + if (max_socket_num < (socket + 1)) > + max_socket_num = socket + 1; > + } > +} > + > /* Determine highest possible node in the system for sparse allocation */ > static void set_max_node_num(void) > { > diff --git a/tools/perf/util/cpumap.h b/tools/perf/util/cpumap.h > index effb56e..094edd9 100644 > --- a/tools/perf/util/cpumap.h > +++ b/tools/perf/util/cpumap.h > @@ -56,9 +56,11 @@ static inline bool cpu_map__empty(const struct cpu_map *map) > > int max_cpu_num; > int max_node_num; > +int max_socket_num; > int *cpunode_map; > > int cpu__setup_cpunode_map(void); > +void set_max_socket_num(void); > > static inline int cpu__max_node(void) > { > diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h > index af80fb5..b188a62 100644 > --- a/tools/perf/util/hist.h > +++ b/tools/perf/util/hist.h > @@ -139,7 +139,7 @@ void events_stats__inc(struct events_stats *stats, u32 type); > size_t events_stats__fprintf(struct events_stats *stats, FILE *fp); > > size_t hists__fprintf(struct hists *hists, bool show_header, int max_rows, > - int max_cols, float min_pcnt, FILE *fp); > + int max_cols, float min_pcnt, FILE *fp, int socket); > size_t perf_evlist__fprintf_nr_events(struct perf_evlist *evlist, FILE *fp); > > void hists__filter_by_dso(struct hists *hists); > diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c > index 245e254..70f0526 100644 > --- a/tools/perf/util/sort.c > +++ b/tools/perf/util/sort.c > @@ -1930,6 +1930,12 @@ int setup_sorting(void) > return err; > } > > + if (symbol_conf.socket) { > + err = sort_dimension__add("socket"); > + if (err < 0) > + return err; > + } > + > reset_dimensions(); > > /* > diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h > index a4cde92..7c153c7 100644 > --- a/tools/perf/util/symbol.h > +++ b/tools/perf/util/symbol.h > @@ -101,6 +101,7 @@ struct symbol_conf { > annotate_asm_raw, > annotate_src, > event_group, > + socket, > demangle, > demangle_kernel, > filter_relative, > -- > 1.8.3.1 > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/