diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2015-07-04 08:17:29 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2015-07-04 08:17:29 -0700 |
commit | c1776a18e3b5a3559f3dff5df0ecce570abd3a9f (patch) | |
tree | 5b97defe199f434f5d9829bf7d8df6124e6ce5f1 | |
parent | 91cca0f0ffa3e485ad9d5b44744f23318c8f99ab (diff) | |
parent | b9df84fd7c05cc300d6d14f022b8a00773ebcf8c (diff) |
Merge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf updates from Ingo Molnar:
"This tree includes an x86 PMU scheduling fix, but most changes are
late breaking tooling fixes and updates:
User visible fixes:
- Create config.detected into OUTPUT directory, fixing parallel
builds sharing the same source directory (Aaro Kiskinen)
- Allow to specify custom linker command, fixing some MIPS64 builds.
(Aaro Kiskinen)
- Fix to show proper convergence stats in 'perf bench numa' (Srikar
Dronamraju)
User visible changes:
- Validate syscall list passed via -e argument to 'perf trace'.
(Arnaldo Carvalho de Melo)
- Introduce 'perf stat --per-thread' (Jiri Olsa)
- Check access permission for --kallsyms and --vmlinux (Li Zhang)
- Move toggling event logic from 'perf top' and into hists browser,
allowing freeze/unfreeze with event lists with more than one entry
(Namhyung Kim)
- Add missing newlines when dumping PERF_RECORD_FINISHED_ROUND and
showing the Aggregated stats in 'perf report -D' (Adrian Hunter)
Infrastructure fixes:
- Add missing break for PERF_RECORD_ITRACE_START, which caused those
events samples to be parsed as well as PERF_RECORD_LOST_SAMPLES.
ITRACE_START only appears when Intel PT or BTS are present, so..
(Jiri Olsa)
- Call the perf_session destructor when bailing out in the inject,
kmem, report, kvm and mem tools (Taeung Song)
Infrastructure changes:
- Move stuff out of 'perf stat' and into the lib for further use
(Jiri Olsa)
- Reference count the cpu_map and thread_map classes (Jiri Olsa)
- Set evsel->{cpus,threads} from the evlist, if not set, allowing the
generalization of some 'perf stat' functions that previously were
accessing private static evlist variable (Jiri Olsa)
- Delete an unnecessary check before the calling free_event_desc()
(Markus Elfring)
- Allow auxtrace data alignment (Adrian Hunter)
- Allow events with dot (Andi Kleen)
- Fix failure to 'perf probe' events on arm (He Kuang)
- Add testing for Makefile.perf (Jiri Olsa)
- Add test for make install with prefix (Jiri Olsa)
- Fix single target build dependency check (Jiri Olsa)
- Access thread_map entries via accessors, prep patch to hold more
info per entry, for ongoing 'perf stat --per-thread' work (Jiri
Olsa)
- Use __weak definition from compiler.h (Sukadev Bhattiprolu)
- Split perf_pmu__new_alias() (Sukadev Bhattiprolu)"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (54 commits)
perf tools: Allow to specify custom linker command
perf tools: Create config.detected into OUTPUT directory
perf mem: Fill in the missing session freeing after an error occurs
perf kvm: Fill in the missing session freeing after an error occurs
perf report: Fill in the missing session freeing after an error occurs
perf kmem: Fill in the missing session freeing after an error occurs
perf inject: Fill in the missing session freeing after an error occurs
perf tools: Add missing break for PERF_RECORD_ITRACE_START
perf/x86: Fix 'active_events' imbalance
perf symbols: Check access permission when reading symbol files
perf stat: Introduce --per-thread option
perf stat: Introduce print_counters function
perf stat: Using init_stats instead of memset
perf stat: Rename print_interval to process_interval
perf stat: Remove perf_evsel__read_cb function
perf stat: Move perf_stat initialization counter process code
perf stat: Move zero_per_pkg into counter process code
perf stat: Separate counters reading and processing
perf stat: Introduce read_counters function
perf stat: Introduce perf_evsel__read function
...
54 files changed, 885 insertions, 422 deletions
diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 5801a14f7524..3658de47900f 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -357,34 +357,24 @@ void x86_release_hardware(void) */ int x86_add_exclusive(unsigned int what) { - int ret = -EBUSY, i; - - if (atomic_inc_not_zero(&x86_pmu.lbr_exclusive[what])) - return 0; + int i; - mutex_lock(&pmc_reserve_mutex); - for (i = 0; i < ARRAY_SIZE(x86_pmu.lbr_exclusive); i++) { - if (i != what && atomic_read(&x86_pmu.lbr_exclusive[i])) - goto out; + if (!atomic_inc_not_zero(&x86_pmu.lbr_exclusive[what])) { + mutex_lock(&pmc_reserve_mutex); + for (i = 0; i < ARRAY_SIZE(x86_pmu.lbr_exclusive); i++) { + if (i != what && atomic_read(&x86_pmu.lbr_exclusive[i])) + goto fail_unlock; + } + atomic_inc(&x86_pmu.lbr_exclusive[what]); + mutex_unlock(&pmc_reserve_mutex); } - atomic_inc(&x86_pmu.lbr_exclusive[what]); - ret = 0; + atomic_inc(&active_events); + return 0; -out: +fail_unlock: mutex_unlock(&pmc_reserve_mutex); - - /* - * Assuming that all exclusive events will share the PMI handler - * (which checks active_events for whether there is work to do), - * we can bump active_events counter right here, except for - * x86_lbr_exclusive_lbr events that go through x86_pmu_event_init() - * path, which already bumps active_events for them. - */ - if (!ret && what != x86_lbr_exclusive_lbr) - atomic_inc(&active_events); - - return ret; + return -EBUSY; } void x86_del_exclusive(unsigned int what) diff --git a/tools/build/Makefile.build b/tools/build/Makefile.build index a51244a8022f..faca2bf6a430 100644 --- a/tools/build/Makefile.build +++ b/tools/build/Makefile.build @@ -25,7 +25,7 @@ build-dir := $(srctree)/tools/build include $(build-dir)/Build.include # do not force detected configuration --include .config-detected +-include $(OUTPUT).config-detected # Init all relevant variables used in build files so # 1) they have correct type diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt index 04e150d83e7d..47469abdcc1c 100644 --- a/tools/perf/Documentation/perf-stat.txt +++ b/tools/perf/Documentation/perf-stat.txt @@ -144,6 +144,10 @@ is a useful mode to detect imbalance between physical cores. To enable this mod use --per-core in addition to -a. (system-wide). The output includes the core number and the number of online logical processors on that physical processor. +--per-thread:: +Aggregate counts per monitored threads, when monitoring threads (-t option) +or processes (-p option). + -D msecs:: --delay msecs:: After starting the program, wait msecs before measuring. This is useful to diff --git a/tools/perf/Makefile b/tools/perf/Makefile index d31a7bbd7cee..480546d5f13b 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -83,8 +83,8 @@ build-test: # # All other targets get passed through: # -%: +%: FORCE $(print_msg) $(make) -.PHONY: tags TAGS +.PHONY: tags TAGS FORCE Makefile diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf index 1af0cfeb7a57..7a4b549214e3 100644 --- a/tools/perf/Makefile.perf +++ b/tools/perf/Makefile.perf @@ -110,7 +110,7 @@ $(OUTPUT)PERF-VERSION-FILE: ../../.git/HEAD $(Q)touch $(OUTPUT)PERF-VERSION-FILE CC = $(CROSS_COMPILE)gcc -LD = $(CROSS_COMPILE)ld +LD ?= $(CROSS_COMPILE)ld AR = $(CROSS_COMPILE)ar PKG_CONFIG = $(CROSS_COMPILE)pkg-config @@ -545,7 +545,7 @@ config-clean: clean: $(LIBTRACEEVENT)-clean $(LIBAPI)-clean config-clean $(call QUIET_CLEAN, core-objs) $(RM) $(LIB_FILE) $(OUTPUT)perf-archive $(OUTPUT)perf-with-kcore $(LANG_BINDINGS) $(Q)find . -name '*.o' -delete -o -name '\.*.cmd' -delete -o -name '\.*.d' -delete - $(Q)$(RM) .config-detected + $(Q)$(RM) $(OUTPUT).config-detected $(call QUIET_CLEAN, core-progs) $(RM) $(ALL_PROGRAMS) perf perf-read-vdso32 perf-read-vdsox32 $(call QUIET_CLEAN, core-gen) $(RM) *.spec *.pyc *.pyo */*.pyc */*.pyo $(OUTPUT)common-cmds.h TAGS tags cscope* $(OUTPUT)PERF-VERSION-FILE $(OUTPUT)FEATURE-DUMP $(OUTPUT)util/*-bison* $(OUTPUT)util/*-flex* $(QUIET_SUBDIR0)Documentation $(QUIET_SUBDIR1) clean diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c index 52ec66b23607..01b06492bd6a 100644 --- a/tools/perf/builtin-inject.c +++ b/tools/perf/builtin-inject.c @@ -630,12 +630,13 @@ int cmd_inject(int argc, const char **argv, const char *prefix __maybe_unused) if (inject.session == NULL) return -1; - if (symbol__init(&inject.session->header.env) < 0) - return -1; + ret = symbol__init(&inject.session->header.env); + if (ret < 0) + goto out_delete; ret = __cmd_inject(&inject); +out_delete: perf_session__delete(inject.session); - return ret; } diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 950f296dfcf7..23b1faaaa4cc 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -1916,7 +1916,7 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused) if (!perf_evlist__find_tracepoint_by_name(session->evlist, "kmem:kmalloc")) { pr_err(errmsg, "slab", "slab"); - return -1; + goto out_delete; } } @@ -1927,7 +1927,7 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __maybe_unused) "kmem:mm_page_alloc"); if (evsel == NULL) { pr_err(errmsg, "page", "page"); - return -1; + goto out_delete; } kmem_page_size = pevent_get_page_size(evsel->tp_format->pevent); diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c index 74878cd75078..fc1cffb1b7a2 100644 --- a/tools/perf/builtin-kvm.c +++ b/tools/perf/builtin-kvm.c @@ -1061,8 +1061,10 @@ static int read_events(struct perf_kvm_stat *kvm) symbol__init(&kvm->session->header.env); - if (!perf_session__has_traces(kvm->session, "kvm record")) - return -EINVAL; + if (!perf_session__has_traces(kvm->session, "kvm record")) { + ret = -EINVAL; + goto out_delete; + } /* * Do not use 'isa' recorded in kvm_exit tracepoint since it is not @@ -1070,9 +1072,13 @@ static int read_events(struct perf_kvm_stat *kvm) */ ret = cpu_isa_config(kvm); if (ret < 0) - return ret; + goto out_delete; - return perf_session__process_events(kvm->session); + ret = perf_session__process_events(kvm->session); + +out_delete: + perf_session__delete(kvm->session); + return ret; } static int parse_target_str(struct perf_kvm_stat *kvm) diff --git a/tools/perf/builtin-mem.c b/tools/perf/builtin-mem.c index da2ec06f0742..80170aace5d4 100644 --- a/tools/perf/builtin-mem.c +++ b/tools/perf/builtin-mem.c @@ -124,7 +124,6 @@ static int report_raw_events(struct perf_mem *mem) .mode = PERF_DATA_MODE_READ, .force = mem->force, }; - int err = -EINVAL; int ret; struct perf_session *session = perf_session__new(&file, false, &mem->tool); @@ -135,24 +134,21 @@ static int report_raw_events(struct perf_mem *mem) if (mem->cpu_list) { ret = perf_session__cpu_bitmap(session, mem->cpu_list, mem->cpu_bitmap); - if (ret) + if (ret < 0) goto out_delete; } - if (symbol__init(&session->header.env) < 0) - return -1; + ret = symbol__init(&session->header.env); + if (ret < 0) + goto out_delete; printf("# PID, TID, IP, ADDR, LOCAL WEIGHT, DSRC, SYMBOL\n"); - err = perf_session__process_events(session); - if (err) - return err; - - return 0; + ret = perf_session__process_events(session); out_delete: perf_session__delete(session); - return err; + return ret; } static int report_events(int argc, const char **argv, struct perf_mem *mem) diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 32626ea3e227..95a47719aec3 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -742,6 +742,17 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused) argc = parse_options(argc, argv, options, report_usage, 0); + if (symbol_conf.vmlinux_name && + access(symbol_conf.vmlinux_name, R_OK)) { + pr_err("Invalid file: %s\n", symbol_conf.vmlinux_name); + return -EINVAL; + } + if (symbol_conf.kallsyms_name && + access(symbol_conf.kallsyms_name, R_OK)) { + pr_err("Invalid file: %s\n", symbol_conf.kallsyms_name); + return -EINVAL; + } + if (report.use_stdio) use_browser = 0; else if (report.use_tui) @@ -828,8 +839,10 @@ repeat: if (report.header || report.header_only) { perf_session__fprintf_info(session, stdout, report.show_full_info); - if (report.header_only) - return 0; + if (report.header_only) { + ret = 0; + goto error; + } } else if (use_browser == 0) { fputs("# To display the perf.data header info, please use --header/--header-only options.\n#\n", stdout); diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index fcf99bdeb19e..37e301a32f43 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -67,10 +67,7 @@ #define CNTR_NOT_SUPPORTED "<not supported>" #define CNTR_NOT_COUNTED "<not counted>" -static void print_stat(int argc, const char **argv); -static void print_counter_aggr(struct perf_evsel *counter, char *prefix); -static void print_counter(struct perf_evsel *counter, char *prefix); -static void print_aggr(char *prefix); +static void print_counters(struct timespec *ts, int argc, const char **argv); /* Default events used for perf stat -T */ static const char *transaction_attrs = { @@ -141,96 +138,9 @@ static inline void diff_timespec(struct timespec *r, struct timespec *a, } } -static inline struct cpu_map *perf_evsel__cpus(struct perf_evsel *evsel) +static void perf_stat__reset_stats(void) { - return (evsel->cpus && !target.cpu_list) ? evsel->cpus : evsel_list->cpus; -} - -static inline int perf_evsel__nr_cpus(struct perf_evsel *evsel) -{ - return perf_evsel__cpus(evsel)->nr; -} - -static void perf_evsel__reset_stat_priv(struct perf_evsel *evsel) -{ - int i; - struct perf_stat *ps = evsel->priv; - - for (i = 0; i < 3; i++) - init_stats(&ps->res_stats[i]); - - perf_stat_evsel_id_init(evsel); -} - -static int perf_evsel__alloc_stat_priv(struct perf_evsel *evsel) -{ - evsel->priv = zalloc(sizeof(struct perf_stat)); - if (evsel->priv == NULL) - return -ENOMEM; - perf_evsel__reset_stat_priv(evsel); - return 0; -} - -static void perf_evsel__free_stat_priv(struct perf_evsel *evsel) -{ - zfree(&evsel->priv); -} - -static int perf_evsel__alloc_prev_raw_counts(struct perf_evsel *evsel) -{ - struct perf_counts *counts; - - counts = perf_counts__new(perf_evsel__nr_cpus(evsel)); - if (counts) - evsel->prev_raw_counts = counts; - - return counts ? 0 : -ENOMEM; -} - -static void perf_evsel__free_prev_raw_counts(struct perf_evsel *evsel) -{ - perf_counts__delete(evsel->prev_raw_counts); - evsel->prev_raw_counts = NULL; -} - -static void perf_evlist__free_stats(struct perf_evlist *evlist) -{ - struct perf_evsel *evsel; - - evlist__for_each(evlist, evsel) { - perf_evsel__free_stat_priv(evsel); - perf_evsel__free_counts(evsel); - perf_evsel__free_prev_raw_counts(evsel); - } -} - -static int perf_evlist__alloc_stats(struct perf_evlist *evlist, bool alloc_raw) -{ - struct perf_evsel *evsel; - - evlist__for_each(evlist, evsel) { - if (perf_evsel__alloc_stat_priv(evsel) < 0 || - perf_evsel__alloc_counts(evsel, perf_evsel__nr_cpus(evsel)) < 0 || - (alloc_raw && perf_evsel__alloc_prev_raw_counts(evsel) < 0)) - goto out_free; - } - - return 0; - -out_free: - perf_evlist__free_stats(evlist); - return -1; -} - -static void perf_stat__reset_stats(struct perf_evlist *evlist) -{ - struct perf_evsel *evsel; - - evlist__for_each(evlist, evsel) { - perf_evsel__reset_stat_priv(evsel); - perf_evsel__reset_counts(evsel, perf_evsel__nr_cpus(evsel)); - } - + perf_evlist__reset_stats(evsel_list); perf_stat__reset_shadow_stats(); } @@ -304,8 +214,9 @@ static int check_per_pkg(struct perf_evsel *counter, int cpu, bool *skip) return 0; } -static int read_cb(struct perf_evsel *evsel, int cpu, int thread __maybe_unused, - struct perf_counts_values *count) +static int +process_counter_values(struct perf_evsel *evsel, int cpu, int thread, + struct perf_counts_values *count) { struct perf_counts_values *aggr = &evsel->counts->aggr; static struct perf_counts_values zero; @@ -320,13 +231,13 @@ static int read_cb(struct perf_evsel *evsel, int cpu, int thread __maybe_unused, count = &zero; switch (aggr_mode) { + case AGGR_THREAD: case AGGR_CORE: case AGGR_SOCKET: case AGGR_NONE: if (!evsel->snapshot) - perf_evsel__compute_deltas(evsel, cpu, count); + perf_evsel__compute_deltas(evsel, cpu, thread, count); perf_counts_values__scale(count, scale, NULL); - evsel->counts->cpu[cpu] = *count; if (aggr_mode == AGGR_NONE) perf_stat__update_shadow_stats(evsel, count->values, cpu); break; @@ -343,26 +254,48 @@ static int read_cb(struct perf_evsel *evsel, int cpu, int thread __maybe_unused, return 0; } -static int read_counter(struct perf_evsel *counter); +static int process_counter_maps(struct perf_evsel *counter) +{ + int nthreads = thread_map__nr(counter->threads); + int ncpus = perf_evsel__nr_cpus(counter); + int cpu, thread; -/* - * Read out the results of a single counter: - * aggregate counts across CPUs in system-wide mode - */ -static int read_counter_aggr(struct perf_evsel *counter) + if (counter->system_wide) + nthreads = 1; + + for (thread = 0; thread < nthreads; thread++) { + for (cpu = 0; cpu < ncpus; cpu++) { + if (process_counter_values(counter, cpu, thread, + perf_counts(counter->counts, cpu, thread))) + return -1; + } + } + + return 0; +} + +static int process_counter(struct perf_evsel *counter) { struct perf_counts_values *aggr = &counter->counts->aggr; struct perf_stat *ps = counter->priv; u64 *count = counter->counts->aggr.values; - int i; + int i, ret; aggr->val = aggr->ena = aggr->run = 0; + init_stats(ps->res_stats); - if (read_counter(counter)) - return -1; + if (counter->per_pkg) + zero_per_pkg(counter); + + ret = process_counter_maps(counter); + if (ret) + return ret; + + if (aggr_mode != AGGR_GLOBAL) + return 0; if (!counter->snapshot) - perf_evsel__compute_deltas(counter, -1, aggr); + perf_evsel__compute_deltas(counter, -1, -1, aggr); perf_counts_values__scale(aggr, scale, &counter->counts->scaled); for (i = 0; i < 3; i++) @@ -397,12 +330,12 @@ static int read_counter(struct perf_evsel *counter) if (counter->system_wide) nthreads = 1; - if (counter->per_pkg) - zero_per_pkg(counter); - for (thread = 0; thread < nthreads; thread++) { for (cpu = 0; cpu < ncpus; cpu++) { - if (perf_evsel__read_cb(counter, cpu, thread, read_cb)) + struct perf_counts_values *count; + + count = perf_counts(counter->counts, cpu, thread); + if (perf_evsel__read(counter, cpu, thread, count)) return -1; } } @@ -410,68 +343,34 @@ static int read_counter(struct perf_evsel *counter) return 0; } -static void print_interval(void) +static void read_counters(bool close) { - static int num_print_interval; struct perf_evsel *counter; - struct perf_stat *ps; - struct timespec ts, rs; - char prefix[64]; - if (aggr_mode == AGGR_GLOBAL) { - evlist__for_each(evsel_list, counter) { - ps = counter->priv; - memset(ps->res_stats, 0, sizeof(ps->res_stats)); - read_counter_aggr(counter); - } - } else { - evlist__for_each(evsel_list, counter) { - ps = counter->priv; - memset(ps->res_stats, 0, sizeof(ps->res_stats)); - read_counter(counter); - } - } + evlist__for_each(evsel_list, counter) { + if (read_counter(counter)) + pr_warning("failed to read counter %s\n", counter->name); - clock_gettime(CLOCK_MONOTONIC, &ts); - diff_timespec(&rs, &ts, &ref_time); - sprintf(prefix, "%6lu.%09lu%s", rs.tv_sec, rs.tv_nsec, csv_sep); + if (process_counter(counter)) + pr_warning("failed to process counter %s\n", counter->name); - if (num_print_interval == 0 && !csv_output) { - switch (aggr_mode) { - case AGGR_SOCKET: - fprintf(output, "# time socket cpus counts %*s events\n", unit_width, "unit"); - break; - case AGGR_CORE: - fprintf(output, "# time core cpus counts %*s events\n", unit_width, "unit"); - break; - case AGGR_NONE: - fprintf(output, "# time CPU counts %*s events\n", unit_width, "unit"); - break; - case AGGR_GLOBAL: - default: - fprintf(output, "# time counts %*s events\n", unit_width, "unit"); + if (close) { + perf_evsel__close_fd(counter, perf_evsel__nr_cpus(counter), + thread_map__nr(evsel_list->threads)); } } +} - if (++num_print_interval == 25) - num_print_interval = 0; +static void process_interval(void) +{ + struct timespec ts, rs; - switch (aggr_mode) { - case AGGR_CORE: - case AGGR_SOCKET: - print_aggr(prefix); - break; - case AGGR_NONE: - evlist__for_each(evsel_list, counter) - print_counter(counter, prefix); - break; - case AGGR_GLOBAL: - default: - evlist__for_each(evsel_list, counter) - print_counter_aggr(counter, prefix); - } + read_counters(false); - fflush(output); + clock_gettime(CLOCK_MONOTONIC, &ts |