From 787bef174f055343c69a9639e6e05a564980ed4c Mon Sep 17 00:00:00 2001 From: David Ahern Date: Fri, 27 May 2011 14:28:43 -0600 Subject: perf script: "sym" field really means show IP data Currently the "sym" output field is used to dump instruction pointers and callchain stack. Sample addresses can also be converted to symbols, so the meaning of "sym" needs to be fixed. This patch adds an "ip" option and if it is selected the user can also opt to dump symbols for them. If the user opts to dump IP without syms only the address is shown. Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1306528124-25861-2-git-send-email-dsahern@gmail.com Signed-off-by: David Ahern Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 31 +++++++++++++++++++++---------- 1 file changed, 21 insertions(+), 10 deletions(-) (limited to 'tools/perf/builtin-script.c') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 22747de7234b..0852db2ea155 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -30,7 +30,8 @@ enum perf_output_field { PERF_OUTPUT_CPU = 1U << 4, PERF_OUTPUT_EVNAME = 1U << 5, PERF_OUTPUT_TRACE = 1U << 6, - PERF_OUTPUT_SYM = 1U << 7, + PERF_OUTPUT_IP = 1U << 7, + PERF_OUTPUT_SYM = 1U << 8, }; struct output_option { @@ -44,6 +45,7 @@ struct output_option { {.str = "cpu", .field = PERF_OUTPUT_CPU}, {.str = "event", .field = PERF_OUTPUT_EVNAME}, {.str = "trace", .field = PERF_OUTPUT_TRACE}, + {.str = "ip", .field = PERF_OUTPUT_IP}, {.str = "sym", .field = PERF_OUTPUT_SYM}, }; @@ -60,7 +62,8 @@ static struct { .fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | - PERF_OUTPUT_EVNAME | PERF_OUTPUT_SYM, + PERF_OUTPUT_EVNAME | PERF_OUTPUT_IP | + PERF_OUTPUT_SYM, .invalid_fields = PERF_OUTPUT_TRACE, }, @@ -70,7 +73,8 @@ static struct { .fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | - PERF_OUTPUT_EVNAME | PERF_OUTPUT_SYM, + PERF_OUTPUT_EVNAME | PERF_OUTPUT_IP | + PERF_OUTPUT_SYM, .invalid_fields = PERF_OUTPUT_TRACE, }, @@ -88,7 +92,8 @@ static struct { .fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | - PERF_OUTPUT_EVNAME | PERF_OUTPUT_SYM, + PERF_OUTPUT_EVNAME | PERF_OUTPUT_IP | + PERF_OUTPUT_SYM, .invalid_fields = PERF_OUTPUT_TRACE, }, @@ -157,15 +162,20 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel, !perf_session__has_traces(session, "record -R")) return -EINVAL; - if (PRINT_FIELD(SYM)) { + if (PRINT_FIELD(IP)) { if (perf_event_attr__check_stype(attr, PERF_SAMPLE_IP, "IP", - PERF_OUTPUT_SYM)) + PERF_OUTPUT_IP)) return -EINVAL; if (!no_callchain && !(attr->sample_type & PERF_SAMPLE_CALLCHAIN)) symbol_conf.use_callchain = false; } + if (PRINT_FIELD(SYM) && !PRINT_FIELD(IP)) { + pr_err("Display of symbols requested but IP is not selected.\n" + "No addresses to convert to symbols.\n"); + return -EINVAL; + } if ((PRINT_FIELD(PID) || PRINT_FIELD(TID)) && perf_event_attr__check_stype(attr, PERF_SAMPLE_TID, "TID", @@ -230,7 +240,7 @@ static void print_sample_start(struct perf_sample *sample, if (PRINT_FIELD(COMM)) { if (latency_format) printf("%8.8s ", thread->comm); - else if (PRINT_FIELD(SYM) && symbol_conf.use_callchain) + else if (PRINT_FIELD(IP) && symbol_conf.use_callchain) printf("%s ", thread->comm); else printf("%16s ", thread->comm); @@ -288,12 +298,13 @@ static void process_event(union perf_event *event __unused, print_trace_event(sample->cpu, sample->raw_data, sample->raw_size); - if (PRINT_FIELD(SYM)) { + if (PRINT_FIELD(IP)) { if (!symbol_conf.use_callchain) printf(" "); else printf("\n"); - perf_session__print_symbols(event, sample, session); + perf_session__print_ip(event, sample, session, + PRINT_FIELD(SYM)); } printf("\n"); @@ -985,7 +996,7 @@ static const struct option options[] = { OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory", "Look for files with symbols relative to this directory"), 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,sym", + "comma separated output fields prepend with 'type:'. Valid types: hw,sw,trace,raw. Fields: comm,tid,pid,time,cpu,event,trace,ip,sym", parse_output_fields), OPT_END() -- cgit v1.2.3 From 610723f24eeb842025178a6722fa9108c4e157b6 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Fri, 27 May 2011 14:28:44 -0600 Subject: perf script: Make printing of dso a separate field option The 'sym' option displays both the function name and the DSO it comes from. Split the display of the dso into a separate option. This allows display of the ip address and symbol without the dso, thus shortening line lengths - and decluttering the output a bit. Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1306528124-25861-3-git-send-email-dsahern@gmail.com Signed-off-by: David Ahern Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 17 ++++++++++++----- 1 file changed, 12 insertions(+), 5 deletions(-) (limited to 'tools/perf/builtin-script.c') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 0852db2ea155..a8bd00f2e557 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -32,6 +32,7 @@ enum perf_output_field { PERF_OUTPUT_TRACE = 1U << 6, PERF_OUTPUT_IP = 1U << 7, PERF_OUTPUT_SYM = 1U << 8, + PERF_OUTPUT_DSO = 1U << 9, }; struct output_option { @@ -47,6 +48,7 @@ struct output_option { {.str = "trace", .field = PERF_OUTPUT_TRACE}, {.str = "ip", .field = PERF_OUTPUT_IP}, {.str = "sym", .field = PERF_OUTPUT_SYM}, + {.str = "dso", .field = PERF_OUTPUT_DSO}, }; /* default set to maintain compatibility with current format */ @@ -63,7 +65,7 @@ static struct { .fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | PERF_OUTPUT_EVNAME | PERF_OUTPUT_IP | - PERF_OUTPUT_SYM, + PERF_OUTPUT_SYM | PERF_OUTPUT_DSO, .invalid_fields = PERF_OUTPUT_TRACE, }, @@ -74,7 +76,7 @@ static struct { .fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | PERF_OUTPUT_EVNAME | PERF_OUTPUT_IP | - PERF_OUTPUT_SYM, + PERF_OUTPUT_SYM | PERF_OUTPUT_DSO, .invalid_fields = PERF_OUTPUT_TRACE, }, @@ -93,7 +95,7 @@ static struct { .fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | PERF_OUTPUT_EVNAME | PERF_OUTPUT_IP | - PERF_OUTPUT_SYM, + PERF_OUTPUT_SYM | PERF_OUTPUT_DSO, .invalid_fields = PERF_OUTPUT_TRACE, }, @@ -176,6 +178,11 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel, "No addresses to convert to symbols.\n"); return -EINVAL; } + if (PRINT_FIELD(DSO) && !PRINT_FIELD(IP)) { + pr_err("Display of DSO requested but IP is not selected.\n" + "No addresses to convert to dso.\n"); + return -EINVAL; + } if ((PRINT_FIELD(PID) || PRINT_FIELD(TID)) && perf_event_attr__check_stype(attr, PERF_SAMPLE_TID, "TID", @@ -304,7 +311,7 @@ static void process_event(union perf_event *event __unused, else printf("\n"); perf_session__print_ip(event, sample, session, - PRINT_FIELD(SYM)); + PRINT_FIELD(SYM), PRINT_FIELD(DSO)); } printf("\n"); @@ -996,7 +1003,7 @@ static const struct option options[] = { OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory", "Look for files with symbols relative to this directory"), 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", + "comma separated output fields prepend with 'type:'. Valid types: hw,sw,trace,raw. Fields: comm,tid,pid,time,cpu,event,trace,ip,sym,dso", parse_output_fields), OPT_END() -- cgit v1.2.3 From 7cec0922389e080d11ec43dd23aa778e136bd1e1 Mon Sep 17 00:00:00 2001 From: David Ahern Date: Mon, 30 May 2011 13:08:23 -0600 Subject: perf script: Add printing of sample address Resolve to a function or variable if possible and if the sym option is enabled. Cc: Frederic Weisbecker Cc: Ingo Molnar Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/1306782503-22002-1-git-send-email-dsahern@gmail.com Signed-off-by: David Ahern Signed-off-by: Arnaldo Carvalho de Melo --- tools/perf/builtin-script.c | 84 +++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 77 insertions(+), 7 deletions(-) (limited to 'tools/perf/builtin-script.c') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index a8bd00f2e557..3056b45b3dd6 100644 --- a/tools/perf/builtin-script.c +++ b/tools/perf/builtin-script.c @@ -33,6 +33,7 @@ enum perf_output_field { PERF_OUTPUT_IP = 1U << 7, PERF_OUTPUT_SYM = 1U << 8, PERF_OUTPUT_DSO = 1U << 9, + PERF_OUTPUT_ADDR = 1U << 10, }; struct output_option { @@ -49,6 +50,7 @@ struct output_option { {.str = "ip", .field = PERF_OUTPUT_IP}, {.str = "sym", .field = PERF_OUTPUT_SYM}, {.str = "dso", .field = PERF_OUTPUT_DSO}, + {.str = "addr", .field = PERF_OUTPUT_ADDR}, }; /* default set to maintain compatibility with current format */ @@ -173,14 +175,22 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel, !(attr->sample_type & PERF_SAMPLE_CALLCHAIN)) symbol_conf.use_callchain = false; } - if (PRINT_FIELD(SYM) && !PRINT_FIELD(IP)) { - pr_err("Display of symbols requested but IP is not selected.\n" - "No addresses to convert to symbols.\n"); + + if (PRINT_FIELD(ADDR) && + perf_event_attr__check_stype(attr, PERF_SAMPLE_ADDR, "ADDR", + PERF_OUTPUT_ADDR)) + return -EINVAL; + + if (PRINT_FIELD(SYM) && !PRINT_FIELD(IP) && !PRINT_FIELD(ADDR)) { + pr_err("Display of symbols requested but neither sample IP nor " + "sample address\nis selected. Hence, no addresses to convert " + "to symbols.\n"); return -EINVAL; } - if (PRINT_FIELD(DSO) && !PRINT_FIELD(IP)) { - pr_err("Display of DSO requested but IP is not selected.\n" - "No addresses to convert to dso.\n"); + if (PRINT_FIELD(DSO) && !PRINT_FIELD(IP) && !PRINT_FIELD(ADDR)) { + pr_err("Display of DSO requested but neither sample IP nor " + "sample address\nis selected. Hence, no addresses to convert " + "to DSO.\n"); return -EINVAL; } @@ -288,6 +298,63 @@ static void print_sample_start(struct perf_sample *sample, } } +static bool sample_addr_correlates_sym(struct perf_event_attr *attr) +{ + if ((attr->type == PERF_TYPE_SOFTWARE) && + ((attr->config == PERF_COUNT_SW_PAGE_FAULTS) || + (attr->config == PERF_COUNT_SW_PAGE_FAULTS_MIN) || + (attr->config == PERF_COUNT_SW_PAGE_FAULTS_MAJ))) + return true; + + return false; +} + +static void print_sample_addr(union perf_event *event, + struct perf_sample *sample, + struct perf_session *session, + struct thread *thread, + struct perf_event_attr *attr) +{ + struct addr_location al; + u8 cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; + const char *symname, *dsoname; + + printf("%16" PRIx64, sample->addr); + + if (!sample_addr_correlates_sym(attr)) + return; + + thread__find_addr_map(thread, session, cpumode, MAP__FUNCTION, + event->ip.pid, sample->addr, &al); + if (!al.map) + thread__find_addr_map(thread, session, cpumode, MAP__VARIABLE, + event->ip.pid, sample->addr, &al); + + al.cpu = sample->cpu; + al.sym = NULL; + + if (al.map) + al.sym = map__find_symbol(al.map, al.addr, NULL); + + if (PRINT_FIELD(SYM)) { + if (al.sym && al.sym->name) + symname = al.sym->name; + else + symname = ""; + + printf(" %16s", symname); + } + + if (PRINT_FIELD(DSO)) { + if (al.map && al.map->dso && al.map->dso->name) + dsoname = al.map->dso->name; + else + dsoname = ""; + + printf(" (%s)", dsoname); + } +} + static void process_event(union perf_event *event __unused, struct perf_sample *sample, struct perf_evsel *evsel, @@ -305,6 +372,9 @@ static void process_event(union perf_event *event __unused, print_trace_event(sample->cpu, sample->raw_data, sample->raw_size); + if (PRINT_FIELD(ADDR)) + print_sample_addr(event, sample, session, thread, attr); + if (PRINT_FIELD(IP)) { if (!symbol_conf.use_callchain) printf(" "); @@ -1003,7 +1073,7 @@ static const struct option options[] = { OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory", "Look for files with symbols relative to this directory"), 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", + "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_END() -- cgit v1.2.3 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/builtin-script.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'tools/perf/builtin-script.c') diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c index 3056b45b3dd6..09024ec2ab2e 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 -- cgit v1.2.3