From 1953287bfe8afcbbd235bd6c42c9df06d52438dc Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Fri, 7 Aug 2009 07:11:05 +0200 Subject: perf tools: Fix call-chain cumul hit based sub-total (fractal mode) The callchain fractal mode builds each new total hits in a new branch of profiling by using the parent's hits of the current branch plus the hits of the children. This is wrong, the total hits of a branch should be made of the sum of every children hits, we must ignore the parent hits in this scope. This patch also fixes another mistake with the hit counting. Now the rates are correct. Signed-off-by: Frederic Weisbecker Cc: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Mike Galbraith Cc: Pekka Enberg Signed-off-by: Ingo Molnar --- tools/perf/builtin-report.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools/perf/builtin-report.c') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 8cb58d68a00..da402e18656 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -901,7 +901,7 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, int i; if (callchain_param.mode == CHAIN_GRAPH_REL) - new_total = self->cumul_hit; + new_total = self->children_hit; else new_total = total_samples; @@ -930,7 +930,7 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, ret += ipchain__fprintf_graph(fp, chain, depth, new_depth_mask, i++, new_total, - child->cumul_hit); + cumul_hits(child)); } ret += callchain__fprintf_graph(fp, child, new_total, depth + 1, -- cgit v1.2.3-70-g09d2 From 8f18aec535b5ca513dd13b531730177d35175ffa Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 6 Aug 2009 19:40:28 +0200 Subject: perf report: Fix per task mult-counter stat reporting Brice Goglin reported: > I can easily sort them by thread id, but I don't know how to match > my 4 events with each group of 4 lines. Also report the counter id and the time running/enabled stats (in case the counter got time-shared). Reported-by: Brice Goglin Signed-off-by: Peter Zijlstra Tested-by: Brice Goglin Signed-off-by: Ingo Molnar --- tools/perf/builtin-report.c | 31 +++++++++++++++++++++++++++---- tools/perf/util/parse-events.c | 8 +++++++- tools/perf/util/parse-events.h | 1 + 3 files changed, 35 insertions(+), 5 deletions(-) (limited to 'tools/perf/builtin-report.c') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index da402e18656..84205462e07 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -112,7 +112,9 @@ struct read_event { struct perf_event_header header; u32 pid,tid; u64 value; - u64 format[3]; + u64 time_enabled; + u64 time_running; + u64 id; }; typedef union event_union { @@ -1690,14 +1692,37 @@ static void trace_event(event_t *event) dprintf(".\n"); } +static struct perf_header *header; + +static struct perf_counter_attr *perf_header__find_attr(u64 id) +{ + int i; + + for (i = 0; i < header->attrs; i++) { + struct perf_header_attr *attr = header->attr[i]; + int j; + + for (j = 0; j < attr->ids; j++) { + if (attr->id[j] == id) + return &attr->attr; + } + } + + return NULL; +} + static int process_read_event(event_t *event, unsigned long offset, unsigned long head) { - dprintf("%p [%p]: PERF_EVENT_READ: %d %d %Lu\n", + struct perf_counter_attr *attr = perf_header__find_attr(event->read.id); + + dprintf("%p [%p]: PERF_EVENT_READ: %d %d %s %Lu\n", (void *)(offset + head), (void *)(long)(event->header.size), event->read.pid, event->read.tid, + attr ? __event_name(attr->type, attr->config) + : "FAIL", event->read.value); return 0; @@ -1743,8 +1768,6 @@ process_event(event_t *event, unsigned long offset, unsigned long head) return 0; } -static struct perf_header *header; - static u64 perf_header__sample_type(void) { u64 sample_type = 0; diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 7bdad8df22a..f77407b5832 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -223,9 +223,15 @@ char *event_name(int counter) { u64 config = attrs[counter].config; int type = attrs[counter].type; + + return __event_name(type, config); +} + +char *__event_name(int type, u64 config) +{ static char buf[32]; - if (attrs[counter].type == PERF_TYPE_RAW) { + if (type == PERF_TYPE_RAW) { sprintf(buf, "raw 0x%llx", config); return buf; } diff --git a/tools/perf/util/parse-events.h b/tools/perf/util/parse-events.h index 1ea5d09b6eb..192a962e3a0 100644 --- a/tools/perf/util/parse-events.h +++ b/tools/perf/util/parse-events.h @@ -10,6 +10,7 @@ extern int nr_counters; extern struct perf_counter_attr attrs[MAX_COUNTERS]; extern char *event_name(int ctr); +extern char *__event_name(int type, u64 config); extern int parse_events(const struct option *opt, const char *str, int unset); -- cgit v1.2.3-70-g09d2 From 94cb9e385d5b4d55a5ae389baa10ad2835ea39bb Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Thu, 6 Aug 2009 14:43:17 -0300 Subject: perf report: Add debug help for the finding of symbol bugs - show the symtab origin (DSO, build-id, kernel, etc) Used with perf report --verbose: [acme@doppio linux-2.6-tip]$ perf report -v | head -16 5.17% firefox /usr/lib64/xulrunner-1.9.1/libxul.so 0x00000000005d8eee f [.] imgContainer::DrawFrameTo(gfxIImageFrame*, gfxIImageFrame*, nsRect&) 2.56% firefox /lib64/libpthread-2.10.1.so 0x0000000000008e02 d [.] __pthread_mutex_lock_internal 1.94% firefox /usr/lib64/xulrunner-1.9.1/libxul.so 0x0000000000d0af8f f [.] SearchTable 1.75% firefox [kernel] 0xffffffffff60013b k [.] vread_hpet 1.63% firefox /lib64/libpthread-2.10.1.so 0x000000000000a404 d [.] __pthread_mutex_unlock 1.47% firefox /usr/lib64/xulrunner-1.9.1/libmozjs.so 0x00000000000482ea f [.] js_Interpret 1.42% firefox /usr/lib64/xulrunner-1.9.1/libmozjs.so 0x000000000003eda3 f [.] JS_CallTracer 1.24% firefox [kernel] 0xffffffff8102ca4a k [k] read_hpet 1.16% firefox [kernel] 0xffffffff810f3dd4 k [k] fget_light 1.11% firefox /usr/lib64/xulrunner-1.9.1/libmozjs.so 0x00000000000567ff f [.] js_TraceObject 0.98% firefox /usr/lib64/firefox-3.5.2/firefox 0x000000000000dd23 b [.] arena_ralloc [acme@doppio linux-2.6-tip]$ The new field is just after the symbol address. To help in figuring out symbol resolution bugs. Signed-off-by: Arnaldo Carvalho de Melo Acked-by: Peter Zijlstra Signed-off-by: Ingo Molnar --- tools/perf/builtin-report.c | 3 ++- tools/perf/util/symbol.c | 57 ++++++++++++++++++++++++++++++++++++--------- tools/perf/util/symbol.h | 2 ++ 3 files changed, 50 insertions(+), 12 deletions(-) (limited to 'tools/perf/builtin-report.c') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 84205462e07..a5e2f8df411 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -700,7 +700,8 @@ sort__sym_print(FILE *fp, struct hist_entry *self, unsigned int width __used) size_t ret = 0; if (verbose) - ret += repsep_fprintf(fp, "%#018llx ", (u64)self->ip); + ret += repsep_fprintf(fp, "%#018llx %c ", (u64)self->ip, + dso__symtab_origin(self->dso)); ret += repsep_fprintf(fp, "[%c] ", self->level); if (self->sym) { diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 16ddca20294..f1dcede1430 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -24,6 +24,16 @@ const char *sym_hist_filter; #define DMGL_ANSI (1 << 1) /* Include const, volatile, etc */ #endif +enum dso_origin { + DSO__ORIG_KERNEL = 0, + DSO__ORIG_JAVA_JIT, + DSO__ORIG_FEDORA, + DSO__ORIG_UBUNTU, + DSO__ORIG_BUILDID, + DSO__ORIG_DSO, + DSO__ORIG_NOT_FOUND, +}; + static struct symbol *symbol__new(u64 start, u64 len, const char *name, unsigned int priv_size, u64 obj_start, int verbose) @@ -81,6 +91,7 @@ struct dso *dso__new(const char *name, unsigned int sym_priv_size) self->sym_priv_size = sym_priv_size; self->find_symbol = dso__find_symbol; self->slen_calculated = 0; + self->origin = DSO__ORIG_NOT_FOUND; } return self; @@ -710,7 +721,7 @@ static char *dso__read_build_id(struct dso *self, int verbose) ++raw; bid += 2; } - if (verbose) + if (verbose >= 2) printf("%s(%s): %s\n", __func__, self->name, build_id); out_elf_end: elf_end(elf); @@ -720,11 +731,26 @@ out: return build_id; } +char dso__symtab_origin(const struct dso *self) +{ + static const char origin[] = { + [DSO__ORIG_KERNEL] = 'k', + [DSO__ORIG_JAVA_JIT] = 'j', + [DSO__ORIG_FEDORA] = 'f', + [DSO__ORIG_UBUNTU] = 'u', + [DSO__ORIG_BUILDID] = 'b', + [DSO__ORIG_DSO] = 'd', + }; + + if (self == NULL || self->origin == DSO__ORIG_NOT_FOUND) + return '!'; + return origin[self->origin]; +} + int dso__load(struct dso *self, symbol_filter_t filter, int verbose) { int size = PATH_MAX; char *name = malloc(size), *build_id = NULL; - int variant = 0; int ret = -1; int fd; @@ -733,19 +759,26 @@ int dso__load(struct dso *self, symbol_filter_t filter, int verbose) self->adjust_symbols = 0; - if (strncmp(self->name, "/tmp/perf-", 10) == 0) - return dso__load_perf_map(self, filter, verbose); + if (strncmp(self->name, "/tmp/perf-", 10) == 0) { + ret = dso__load_perf_map(self, filter, verbose); + self->origin = ret > 0 ? DSO__ORIG_JAVA_JIT : + DSO__ORIG_NOT_FOUND; + return ret; + } + + self->origin = DSO__ORIG_FEDORA - 1; more: do { - switch (variant) { - case 0: /* Fedora */ + self->origin++; + switch (self->origin) { + case DSO__ORIG_FEDORA: snprintf(name, size, "/usr/lib/debug%s.debug", self->name); break; - case 1: /* Ubuntu */ + case DSO__ORIG_UBUNTU: snprintf(name, size, "/usr/lib/debug%s", self->name); break; - case 2: + case DSO__ORIG_BUILDID: build_id = dso__read_build_id(self, verbose); if (build_id != NULL) { snprintf(name, size, @@ -754,16 +787,15 @@ more: free(build_id); break; } - variant++; + self->origin++; /* Fall thru */ - case 3: /* Sane people */ + case DSO__ORIG_DSO: snprintf(name, size, "%s", self->name); break; default: goto out; } - variant++; fd = open(name, O_RDONLY); } while (fd < 0); @@ -899,6 +931,9 @@ int dso__load_kernel(struct dso *self, const char *vmlinux, if (err <= 0) err = dso__load_kallsyms(self, filter, verbose); + if (err > 0) + self->origin = DSO__ORIG_KERNEL; + return err; } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 2f92b21c712..1e003ec2f4b 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -26,6 +26,7 @@ struct dso { unsigned int sym_priv_size; unsigned char adjust_symbols; unsigned char slen_calculated; + unsigned char origin; char name[0]; }; @@ -49,6 +50,7 @@ int dso__load_modules(struct dso *self, symbol_filter_t filter, int verbose); int dso__load(struct dso *self, symbol_filter_t filter, int verbose); size_t dso__fprintf(struct dso *self, FILE *fp); +char dso__symtab_origin(const struct dso *self); void symbol__init(void); #endif /* _PERF_SYMBOL_ */ -- cgit v1.2.3-70-g09d2 From b1a88349c37624755b28ac3b3152b48f52c1f487 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 8 Aug 2009 02:16:24 +0200 Subject: perf tools: callchain: Fix 'perf report' display to be callchain by default If we recorded with -g option to record the callchain, right now we require a -g option to perf report as well - and people reported this as unnecessary complication: the user already specified -g once, no need to require it a second time. So if the recording includes call-chains, display the callchain by default from perf report. ( The user can override this default using "-g none" option from perf report. ) Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Mike Galbraith LKML-Reference: <1249690585-9145-3-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-report.c | 16 +++++++++++++++- tools/perf/util/callchain.h | 1 + 2 files changed, 16 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 a5e2f8df411..c4a8e108e52 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -68,7 +68,7 @@ static int callchain; static struct callchain_param callchain_param = { - .mode = CHAIN_GRAPH_ABS, + .mode = CHAIN_GRAPH_REL, .min_percent = 0.5 }; @@ -1836,6 +1836,13 @@ static int __cmd_report(void) " -g?\n"); exit(-1); } + } else if (callchain_param.mode != CHAIN_NONE && !callchain) { + callchain = 1; + if (register_callchain_param(&callchain_param) < 0) { + fprintf(stderr, "Can't register callchain" + " params\n"); + exit(-1); + } } if (load_kernel() < 0) { @@ -1974,6 +1981,13 @@ parse_callchain_opt(const struct option *opt __used, const char *arg, else if (!strncmp(tok, "fractal", strlen(arg))) callchain_param.mode = CHAIN_GRAPH_REL; + else if (!strncmp(tok, "none", strlen(arg))) { + callchain_param.mode = CHAIN_NONE; + callchain = 0; + + return 0; + } + else return -1; diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index b2d128e07c8..a926ae4f5a1 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -7,6 +7,7 @@ #include "symbol.h" enum chain_mode { + CHAIN_NONE, CHAIN_FLAT, CHAIN_GRAPH_ABS, CHAIN_GRAPH_REL -- cgit v1.2.3-70-g09d2 From 25446036cbfc2c89faacdb4fb4603943d2197dc6 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Sat, 8 Aug 2009 02:16:25 +0200 Subject: perf tools: callchain: Fix sum of percentages to be 100% by displaying amount of ignored chains in fractal mode When we filter the callchains below a given percentage, we ignore them and the end result only shows entries that have an upper percentage than the filter threshold. It seems to users then that we have an imbalance in the percentage, as if the sum inside a profiled branch doesn't reach 100%. Since in the past there have been real perf report bugs that showed the same sypmtom, it would be nice to assure the user that the data is perfect and trustable and it all sums up to 100.00%. So fix this by displaying the remaining hits that have been filtered but without more detail than their amount in each branches. Example while filtering below 50%: 7.73% [k] delay_tsc | |--98.22%-- __const_udelay | | | |--86.37%-- ath5k_hw_register_timeout | | ath5k_hw_noise_floor_calibration | | ath5k_hw_reset | | ath5k_reset | | ath5k_config | | ieee80211_hw_config | | | | | |--88.53%-- ieee80211_scan_work | | | worker_thread | | | kthread | | | child_rip | | --11.47%-- [...] | --13.63%-- [...] --1.78%-- [...] Reported-by: Ingo Molnar Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Mike Galbraith LKML-Reference: <1249690585-9145-4-git-send-email-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-report.c | 47 ++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 44 insertions(+), 3 deletions(-) (limited to 'tools/perf/builtin-report.c') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index c4a8e108e52..99274cec0ad 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -891,6 +891,21 @@ ipchain__fprintf_graph(FILE *fp, struct callchain_list *chain, int depth, return ret; } +static struct symbol *rem_sq_bracket; +static struct callchain_list rem_hits; + +static void init_rem_hits(void) +{ + rem_sq_bracket = malloc(sizeof(*rem_sq_bracket) + 6); + if (!rem_sq_bracket) { + fprintf(stderr, "Not enough memory to display remaining hits\n"); + return; + } + + strcpy(rem_sq_bracket->name, "[...]"); + rem_hits.sym = rem_sq_bracket; +} + static size_t callchain__fprintf_graph(FILE *fp, struct callchain_node *self, u64 total_samples, int depth, int depth_mask) @@ -900,6 +915,7 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, struct callchain_list *chain; int new_depth_mask = depth_mask; u64 new_total; + u64 remaining; size_t ret = 0; int i; @@ -908,17 +924,25 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, else new_total = total_samples; + remaining = new_total; + node = rb_first(&self->rb_root); while (node) { + u64 cumul; + child = rb_entry(node, struct callchain_node, rb_node); + cumul = cumul_hits(child); + remaining -= cumul; /* * The depth mask manages the output of pipes that show * the depth. We don't want to keep the pipes of the current - * level for the last child of this depth + * level for the last child of this depth. + * Except if we have remaining filtered hits. They will + * supersede the last child */ next = rb_next(node); - if (!next) + if (!next && (callchain_param.mode != CHAIN_GRAPH_REL || !remaining)) new_depth_mask &= ~(1 << (depth - 1)); /* @@ -933,7 +957,7 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, ret += ipchain__fprintf_graph(fp, chain, depth, new_depth_mask, i++, new_total, - cumul_hits(child)); + cumul); } ret += callchain__fprintf_graph(fp, child, new_total, depth + 1, @@ -941,6 +965,19 @@ callchain__fprintf_graph(FILE *fp, struct callchain_node *self, node = next; } + if (callchain_param.mode == CHAIN_GRAPH_REL && + remaining && remaining != new_total) { + + if (!rem_sq_bracket) + return ret; + + new_depth_mask &= ~(1 << (depth - 1)); + + ret += ipchain__fprintf_graph(fp, &rem_hits, depth, + new_depth_mask, 0, new_total, + remaining); + } + return ret; } @@ -1361,6 +1398,8 @@ static size_t output__fprintf(FILE *fp, u64 total_samples) unsigned int width; char *col_width = col_width_list_str; + init_rem_hits(); + fprintf(fp, "# Samples: %Ld\n", (u64)total_samples); fprintf(fp, "#\n"); @@ -1432,6 +1471,8 @@ print_entries: } fprintf(fp, "\n"); + free(rem_sq_bracket); + return ret; } -- cgit v1.2.3-70-g09d2