From 59b4caeb797494043f5f3b98a610f5d9b75eefa3 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Sun, 6 Dec 2009 10:16:30 +0100 Subject: perf tools: Correct size computation in tracepoint_id_to_path() The size argument to zalloc should be the size of desired structure, not the pointer to it. The semantic patch that makes this change is as follows: (http://coccinelle.lip6.fr/) // @expression@ expression *x; @@ x = <+... -sizeof(x) +sizeof(*x) ...+>// Signed-off-by: Julia Lawall Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: Signed-off-by: Ingo Molnar --- tools/perf/util/parse-events.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 9e5dbd66d34..448a13b5201 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -197,7 +197,7 @@ struct tracepoint_path *tracepoint_id_to_path(u64 config) if (id == config) { closedir(evt_dir); closedir(sys_dir); - path = zalloc(sizeof(path)); + path = zalloc(sizeof(*path)); path->system = malloc(MAX_EVENT_LENGTH); if (!path->system) { free(path); -- cgit From 028c515253761084c6594bf9ac9b194b51d87065 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Sun, 6 Dec 2009 20:07:29 +0900 Subject: perf timechart: Fix header handling Update "struct trace_entry" to match with current one. And remove "size" field from it. If it has "size", it become cause of alignment mismatch of structure with kernel. Signed-off-by: OGAWA Hirofumi Acked-by: Arjan van de Ven Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <87ljhg8ioe.fsf@devron.myhome.or.jp> Signed-off-by: Ingo Molnar --- tools/perf/builtin-timechart.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index cb58b6605fc..c0f29ed0996 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -302,12 +302,11 @@ process_exit_event(event_t *event) } struct trace_entry { - u32 size; unsigned short type; unsigned char flags; unsigned char preempt_count; int pid; - int tgid; + int lock_depth; }; struct power_entry { @@ -489,6 +488,7 @@ process_sample_event(event_t *event) u64 stamp = 0; u32 cpu = 0; u32 pid = 0; + u32 size, *size_ptr; struct trace_entry *te; if (sample_type & PERF_SAMPLE_IP) @@ -518,9 +518,13 @@ process_sample_event(event_t *event) if (sample_type & PERF_SAMPLE_PERIOD) cursor++; - te = (void *)&event->sample.array[cursor]; + size_ptr = (void *)&event->sample.array[cursor]; - if (sample_type & PERF_SAMPLE_RAW && te->size > 0) { + size = *size_ptr; + size_ptr++; + + te = (void *)size_ptr; + if (sample_type & PERF_SAMPLE_RAW && size > 0) { char *event_str; struct power_entry *pe; -- cgit From 180f95e29aa8782c019caa64ede2a28d8ab62564 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Sun, 6 Dec 2009 20:08:24 +0900 Subject: perf: Make common SAMPLE_EVENT parser Currently, sample event data is parsed for each commands, and it is assuming that the data is not including other data. (E.g. timechart, trace, etc. can't parse the event if it has PERF_SAMPLE_CALLCHAIN) So, even if we record the superset data for multiple commands at a time, commands can't parse. etc. To fix it, this makes common sample event parser, and use it to parse sample event correctly. (PERF_SAMPLE_READ is unsupported for now though, it seems to be not using.) Signed-off-by: OGAWA Hirofumi Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <87hbs48imv.fsf@devron.myhome.or.jp> Signed-off-by: Ingo Molnar --- tools/perf/builtin-kmem.c | 36 ++++++++--------------- tools/perf/builtin-report.c | 39 ++++++++++++------------ tools/perf/builtin-sched.c | 38 ++++++++---------------- tools/perf/builtin-timechart.c | 56 ++++++++++------------------------- tools/perf/builtin-trace.c | 48 ++++++++++-------------------- tools/perf/util/event.c | 67 ++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/event.h | 17 ++++++++++- 7 files changed, 155 insertions(+), 146 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 047fef74bd5..f218990de0c 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -320,35 +320,23 @@ process_raw_event(event_t *raw_event __used, void *more_data, static int process_sample_event(event_t *event) { - u64 ip = event->ip.ip; - u64 timestamp = -1; - u32 cpu = -1; - u64 period = 1; - void *more_data = event->ip.__more_data; - struct thread *thread = threads__findnew(event->ip.pid); - - if (sample_type & PERF_SAMPLE_TIME) { - timestamp = *(u64 *)more_data; - more_data += sizeof(u64); - } + struct sample_data data; + struct thread *thread; - if (sample_type & PERF_SAMPLE_CPU) { - cpu = *(u32 *)more_data; - more_data += sizeof(u32); - more_data += sizeof(u32); /* reserved */ - } + memset(&data, 0, sizeof(data)); + data.time = -1; + data.cpu = -1; + data.period = 1; - if (sample_type & PERF_SAMPLE_PERIOD) { - period = *(u64 *)more_data; - more_data += sizeof(u64); - } + event__parse_sample(event, sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, - event->ip.pid, event->ip.tid, - (void *)(long)ip, - (long long)period); + data.pid, data.tid, + (void *)(long)data.ip, + (long long)data.period); + thread = threads__findnew(event->ip.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); @@ -357,7 +345,7 @@ static int process_sample_event(event_t *event) dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - process_raw_event(event, more_data, cpu, timestamp, thread); + process_raw_event(event, data.raw_data, data.cpu, data.time, thread); return 0; } diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 383c4ab4f9a..2b9eb3a553e 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -605,44 +605,41 @@ static int validate_chain(struct ip_callchain *chain, event_t *event) static int process_sample_event(event_t *event) { - u64 ip = event->ip.ip; - u64 period = 1; - void *more_data = event->ip.__more_data; - struct ip_callchain *chain = NULL; + struct sample_data data; int cpumode; struct addr_location al; - struct thread *thread = threads__findnew(event->ip.pid); + struct thread *thread; - if (sample_type & PERF_SAMPLE_PERIOD) { - period = *(u64 *)more_data; - more_data += sizeof(u64); - } + memset(&data, 0, sizeof(data)); + data.period = 1; + + event__parse_sample(event, sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, - event->ip.pid, event->ip.tid, - (void *)(long)ip, - (long long)period); + data.pid, data.tid, + (void *)(long)data.ip, + (long long)data.period); if (sample_type & PERF_SAMPLE_CALLCHAIN) { unsigned int i; - chain = (void *)more_data; - - dump_printf("... chain: nr:%Lu\n", chain->nr); + dump_printf("... chain: nr:%Lu\n", data.callchain->nr); - if (validate_chain(chain, event) < 0) { + if (validate_chain(data.callchain, event) < 0) { pr_debug("call-chain problem with event, " "skipping it.\n"); return 0; } if (dump_trace) { - for (i = 0; i < chain->nr; i++) - dump_printf("..... %2d: %016Lx\n", i, chain->ips[i]); + for (i = 0; i < data.callchain->nr; i++) + dump_printf("..... %2d: %016Lx\n", + i, data.callchain->ips[i]); } } + thread = threads__findnew(data.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); @@ -657,7 +654,7 @@ static int process_sample_event(event_t *event) cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; thread__find_addr_location(thread, cpumode, - MAP__FUNCTION, ip, &al, NULL); + MAP__FUNCTION, data.ip, &al, NULL); /* * We have to do this here as we may have a dso with no symbol hit that * has a name longer than the ones with symbols sampled. @@ -675,12 +672,12 @@ static int process_sample_event(event_t *event) if (sym_list && al.sym && !strlist__has_entry(sym_list, al.sym->name)) return 0; - if (hist_entry__add(&al, chain, period)) { + if (hist_entry__add(&al, data.callchain, data.period)) { pr_debug("problem incrementing symbol count, skipping event\n"); return -1; } - event__stats.total += period; + event__stats.total += data.period; return 0; } diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 26b782f26ee..45c46c79049 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1598,40 +1598,26 @@ process_raw_event(event_t *raw_event __used, void *more_data, static int process_sample_event(event_t *event) { + struct sample_data data; struct thread *thread; - u64 ip = event->ip.ip; - u64 timestamp = -1; - u32 cpu = -1; - u64 period = 1; - void *more_data = event->ip.__more_data; if (!(sample_type & PERF_SAMPLE_RAW)) return 0; - thread = threads__findnew(event->ip.pid); + memset(&data, 0, sizeof(data)); + data.time = -1; + data.cpu = -1; + data.period = -1; - if (sample_type & PERF_SAMPLE_TIME) { - timestamp = *(u64 *)more_data; - more_data += sizeof(u64); - } - - if (sample_type & PERF_SAMPLE_CPU) { - cpu = *(u32 *)more_data; - more_data += sizeof(u32); - more_data += sizeof(u32); /* reserved */ - } - - if (sample_type & PERF_SAMPLE_PERIOD) { - period = *(u64 *)more_data; - more_data += sizeof(u64); - } + event__parse_sample(event, sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, - event->ip.pid, event->ip.tid, - (void *)(long)ip, - (long long)period); + data.pid, data.tid, + (void *)(long)data.ip, + (long long)data.period); + thread = threads__findnew(data.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); @@ -1640,10 +1626,10 @@ static int process_sample_event(event_t *event) dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - if (profile_cpu != -1 && profile_cpu != (int) cpu) + if (profile_cpu != -1 && profile_cpu != (int)data.cpu) return 0; - process_raw_event(event, more_data, cpu, timestamp, thread); + process_raw_event(event, data.raw_data, data.cpu, data.time, thread); return 0; } diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index c0f29ed0996..f472df9561e 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -483,48 +483,22 @@ static void sched_switch(int cpu, u64 timestamp, struct trace_entry *te) static int process_sample_event(event_t *event) { - int cursor = 0; - u64 addr = 0; - u64 stamp = 0; - u32 cpu = 0; - u32 pid = 0; - u32 size, *size_ptr; + struct sample_data data; struct trace_entry *te; - if (sample_type & PERF_SAMPLE_IP) - cursor++; - - if (sample_type & PERF_SAMPLE_TID) { - pid = event->sample.array[cursor]>>32; - cursor++; - } - if (sample_type & PERF_SAMPLE_TIME) { - stamp = event->sample.array[cursor++]; + memset(&data, 0, sizeof(data)); - if (!first_time || first_time > stamp) - first_time = stamp; - if (last_time < stamp) - last_time = stamp; + event__parse_sample(event, sample_type, &data); + if (sample_type & PERF_SAMPLE_TIME) { + if (!first_time || first_time > data.time) + first_time = data.time; + if (last_time < data.time) + last_time = data.time; } - if (sample_type & PERF_SAMPLE_ADDR) - addr = event->sample.array[cursor++]; - if (sample_type & PERF_SAMPLE_ID) - cursor++; - if (sample_type & PERF_SAMPLE_STREAM_ID) - cursor++; - if (sample_type & PERF_SAMPLE_CPU) - cpu = event->sample.array[cursor++] & 0xFFFFFFFF; - if (sample_type & PERF_SAMPLE_PERIOD) - cursor++; - - size_ptr = (void *)&event->sample.array[cursor]; - - size = *size_ptr; - size_ptr++; - te = (void *)size_ptr; - if (sample_type & PERF_SAMPLE_RAW && size > 0) { + te = (void *)data.raw_data; + if (sample_type & PERF_SAMPLE_RAW && data.raw_size > 0) { char *event_str; struct power_entry *pe; @@ -536,19 +510,19 @@ process_sample_event(event_t *event) return 0; if (strcmp(event_str, "power:power_start") == 0) - c_state_start(cpu, stamp, pe->value); + c_state_start(data.cpu, data.time, pe->value); if (strcmp(event_str, "power:power_end") == 0) - c_state_end(cpu, stamp); + c_state_end(data.cpu, data.time); if (strcmp(event_str, "power:power_frequency") == 0) - p_state_change(cpu, stamp, pe->value); + p_state_change(data.cpu, data.time, pe->value); if (strcmp(event_str, "sched:sched_wakeup") == 0) - sched_wakeup(cpu, stamp, pid, te); + sched_wakeup(data.cpu, data.time, data.pid, te); if (strcmp(event_str, "sched:sched_switch") == 0) - sched_switch(cpu, stamp, te); + sched_switch(data.cpu, data.time, te); } return 0; } diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index abb914aa7be..c2fcc34486f 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -66,58 +66,40 @@ static u64 sample_type; static int process_sample_event(event_t *event) { - u64 ip = event->ip.ip; - u64 timestamp = -1; - u32 cpu = -1; - u64 period = 1; - void *more_data = event->ip.__more_data; - struct thread *thread = threads__findnew(event->ip.pid); - - if (sample_type & PERF_SAMPLE_TIME) { - timestamp = *(u64 *)more_data; - more_data += sizeof(u64); - } + struct sample_data data; + struct thread *thread; - if (sample_type & PERF_SAMPLE_CPU) { - cpu = *(u32 *)more_data; - more_data += sizeof(u32); - more_data += sizeof(u32); /* reserved */ - } + memset(&data, 0, sizeof(data)); + data.time = -1; + data.cpu = -1; + data.period = 1; - if (sample_type & PERF_SAMPLE_PERIOD) { - period = *(u64 *)more_data; - more_data += sizeof(u64); - } + event__parse_sample(event, sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, - event->ip.pid, event->ip.tid, - (void *)(long)ip, - (long long)period); + data.pid, data.tid, + (void *)(long)data.ip, + (long long)data.period); + thread = threads__findnew(event->ip.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); return -1; } - dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - if (sample_type & PERF_SAMPLE_RAW) { - struct { - u32 size; - char data[0]; - } *raw = more_data; - /* * FIXME: better resolve from pid from the struct trace_entry * field, although it should be the same than this perf * event pid */ - scripting_ops->process_event(cpu, raw->data, raw->size, - timestamp, thread->comm); + scripting_ops->process_event(data.cpu, data.raw_data, + data.raw_size, + data.time, thread->comm); } - event__stats.total += period; + event__stats.total += data.period; return 0; } diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 414b89d1bde..4dcecafa85d 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -310,3 +310,70 @@ int event__preprocess_sample(const event_t *self, struct addr_location *al, al->level == 'H' ? "[hypervisor]" : ""); return 0; } + +int event__parse_sample(event_t *event, u64 type, struct sample_data *data) +{ + u64 *array = event->sample.array; + + if (type & PERF_SAMPLE_IP) { + data->ip = event->ip.ip; + array++; + } + + if (type & PERF_SAMPLE_TID) { + u32 *p = (u32 *)array; + data->pid = p[0]; + data->tid = p[1]; + array++; + } + + if (type & PERF_SAMPLE_TIME) { + data->time = *array; + array++; + } + + if (type & PERF_SAMPLE_ADDR) { + data->addr = *array; + array++; + } + + if (type & PERF_SAMPLE_ID) { + data->id = *array; + array++; + } + + if (type & PERF_SAMPLE_STREAM_ID) { + data->stream_id = *array; + array++; + } + + if (type & PERF_SAMPLE_CPU) { + u32 *p = (u32 *)array; + data->cpu = *p; + array++; + } + + if (type & PERF_SAMPLE_PERIOD) { + data->period = *array; + array++; + } + + if (type & PERF_SAMPLE_READ) { + pr_debug("PERF_SAMPLE_READ is unsuported for now\n"); + return -1; + } + + if (type & PERF_SAMPLE_CALLCHAIN) { + data->callchain = (struct ip_callchain *)array; + array += 1 + data->callchain->nr; + } + + if (type & PERF_SAMPLE_RAW) { + u32 *p = (u32 *)array; + data->raw_size = *p; + p++; + data->raw_data = p; + } + + return 0; +} diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index a4cc8105cf6..c7a78eef8e5 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -56,11 +56,25 @@ struct read_event { u64 id; }; -struct sample_event{ +struct sample_event { struct perf_event_header header; u64 array[]; }; +struct sample_data { + u64 ip; + u32 pid, tid; + u64 time; + u64 addr; + u64 id; + u64 stream_id; + u32 cpu; + u64 period; + struct ip_callchain *callchain; + u32 raw_size; + void *raw_data; +}; + #define BUILD_ID_SIZE 20 struct build_id_event { @@ -155,5 +169,6 @@ int event__process_task(event_t *self); struct addr_location; int event__preprocess_sample(const event_t *self, struct addr_location *al, symbol_filter_t filter); +int event__parse_sample(event_t *event, u64 type, struct sample_data *data); #endif /* __PERF_RECORD_H */ -- cgit From 7691b1ec2e4a8d4bd88dcf88b29792399ebe1c91 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Sun, 6 Dec 2009 20:10:49 +0900 Subject: perf tools: Misc small fixes - util/header.c "len" is aligned to 64. So, it tries to write the out of long_name buffer. So, this use "zero_buf" to write aligned area. - util/trace-event-read.c "size" is not including nul byte. So, this allocates it, and set '\0'. - util/trace-event-parse.c It needs parens to calc correct size. Signed-off-by: OGAWA Hirofumi Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <87d42s8iiu.fsf_-_@devron.myhome.or.jp> Signed-off-by: Ingo Molnar --- tools/perf/util/header.c | 9 +++++++-- tools/perf/util/trace-event-parse.c | 2 +- tools/perf/util/trace-event-read.c | 3 ++- 3 files changed, 10 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 4805e6dfd23..08b6759287f 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -187,7 +187,9 @@ static int do_write(int fd, const void *buf, size_t size) static int __dsos__write_buildid_table(struct list_head *head, int fd) { +#define NAME_ALIGN 64 struct dso *pos; + static const char zero_buf[NAME_ALIGN]; list_for_each_entry(pos, head, node) { int err; @@ -197,14 +199,17 @@ static int __dsos__write_buildid_table(struct list_head *head, int fd) if (!pos->has_build_id) continue; len = pos->long_name_len + 1; - len = ALIGN(len, 64); + len = ALIGN(len, NAME_ALIGN); memset(&b, 0, sizeof(b)); memcpy(&b.build_id, pos->build_id, sizeof(pos->build_id)); b.header.size = sizeof(b) + len; err = do_write(fd, &b, sizeof(b)); if (err < 0) return err; - err = do_write(fd, pos->long_name, len); + err = do_write(fd, pos->long_name, pos->long_name_len + 1); + if (err < 0) + return err; + err = do_write(fd, zero_buf, len - pos->long_name_len + 1); if (err < 0) return err; } diff --git a/tools/perf/util/trace-event-parse.c b/tools/perf/util/trace-event-parse.c index 0302405aa2c..6ffe9d63d85 100644 --- a/tools/perf/util/trace-event-parse.c +++ b/tools/perf/util/trace-event-parse.c @@ -177,7 +177,7 @@ void parse_proc_kallsyms(char *file, unsigned int size __unused) func_count++; } - func_list = malloc_or_die(sizeof(*func_list) * func_count + 1); + func_list = malloc_or_die(sizeof(*func_list) * (func_count + 1)); i = 0; while (list) { diff --git a/tools/perf/util/trace-event-read.c b/tools/perf/util/trace-event-read.c index 342dfdd43f8..1744422cafc 100644 --- a/tools/perf/util/trace-event-read.c +++ b/tools/perf/util/trace-event-read.c @@ -145,8 +145,9 @@ static void read_proc_kallsyms(void) if (!size) return; - buf = malloc_or_die(size); + buf = malloc_or_die(size + 1); read_or_die(buf, size); + buf[size] = '\0'; parse_proc_kallsyms(buf, size); -- cgit From b9b1e1c71a9481c0c34ed5bed42f1bfa730fd39e Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 6 Dec 2009 18:03:10 -0200 Subject: perf buildid-list: Fix copy'n'paste help message MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260129790-11520-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-buildid-list.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index 7dee9d19ab7..dcb6143a000 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -19,7 +19,7 @@ static char const *input_name = "perf.data"; static int force; static const char *const buildid_list_usage[] = { - "perf report []", + "perf buildid-list []", NULL }; -- cgit From c0777c5aa835a97ccc77d82e55388940f0140a61 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Mon, 7 Dec 2009 12:04:49 +0800 Subject: perf/sched: Fix 'perf sched trace' If we use 'perf sched trace', it will call symbol__init() again, and can lead to a perf tool crash: [root@localhost perf]# ./perf sched trace *** glibc detected *** ./perf: free(): invalid next size (normal): 0x094c1898 *** ======= Backtrace: ========= /lib/libc.so.6[0xb7602404] /lib/libc.so.6(cfree+0x96)[0xb76043b6] ./perf[0x80730fe] ./perf[0x8074c97] ./perf[0x805eb59] ./perf[0x80536fd] ./perf[0x804b618] ./perf[0x804bdc3] /lib/libc.so.6(__libc_start_main+0xe5)[0xb75a9735] ./perf[0x804af81] ======= Memory map: ======== 08048000-08158000 r-xp 00000000 fe:00 556831 /home/eric/.... 08158000-08168000 rw-p 0010f000 fe:00 556831 /home/eric/... 08168000-085fe000 rw-p 00000000 00:00 0 094ab000-094cc000 rw-p 00000000 00:00 0 [heap] Signed-off-by: Xiao Guangrong LKML-Reference: <4B1C7EE1.8030906@cn.fujitsu.com> Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker Signed-off-by: Ingo Molnar --- tools/perf/builtin-sched.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 45c46c79049..7481ebdb17e 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1888,13 +1888,18 @@ static int __cmd_record(int argc, const char **argv) int cmd_sched(int argc, const char **argv, const char *prefix __used) { - symbol__init(0); - argc = parse_options(argc, argv, sched_options, sched_usage, PARSE_OPT_STOP_AT_NON_OPTION); if (!argc) usage_with_options(sched_usage, sched_options); + /* + * Aliased to 'perf trace' for now: + */ + if (!strcmp(argv[0], "trace")) + return cmd_trace(argc, argv, prefix); + + symbol__init(0); if (!strncmp(argv[0], "rec", 3)) { return __cmd_record(argc, argv); } else if (!strncmp(argv[0], "lat", 3)) { @@ -1918,11 +1923,6 @@ int cmd_sched(int argc, const char **argv, const char *prefix __used) usage_with_options(replay_usage, replay_options); } __cmd_replay(); - } else if (!strcmp(argv[0], "trace")) { - /* - * Aliased to 'perf trace' for now: - */ - return cmd_trace(argc, argv, prefix); } else { usage_with_options(sched_usage, sched_options); } -- cgit From d8bd9e0aedabcb47887712497bc386a06ddcbd12 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Mon, 7 Dec 2009 12:06:29 +0800 Subject: perf_event: Fix raw event processing We use 'data.raw_data' parameter to call process_raw_event(), but data.raw_data buffer not include data size. it can make perf tool crash. This bug was introduced by commit 180f95e29a ("perf: Make common SAMPLE_EVENT parser"). Signed-off-by: Xiao Guangrong Cc: Pekka Enberg Cc: Eduard - Gabriel Munteanu Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: OGAWA Hirofumi Cc: Peter Zijlstra Cc: Li Zefan LKML-Reference: <4B1C7F45.5080105@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-kmem.c | 11 ++++++++--- tools/perf/builtin-sched.c | 11 ++++++++--- 2 files changed, 16 insertions(+), 6 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index f218990de0c..f84d7a3db68 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -289,13 +289,17 @@ static void process_free_event(struct raw_event_sample *raw, } static void -process_raw_event(event_t *raw_event __used, void *more_data, +process_raw_event(event_t *raw_event __used, u32 size, void *data, int cpu, u64 timestamp, struct thread *thread) { - struct raw_event_sample *raw = more_data; + struct raw_event_sample *raw; struct event *event; int type; + raw = malloc_or_die(sizeof(*raw)+size); + raw->size = size; + memcpy(raw->data, data, size); + type = trace_parse_common_type(raw->data); event = trace_find_event(type); @@ -345,7 +349,8 @@ static int process_sample_event(event_t *event) dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - process_raw_event(event, data.raw_data, data.cpu, data.time, thread); + process_raw_event(event, data.raw_size, data.raw_data, data.cpu, + data.time, thread); return 0; } diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 7481ebdb17e..4655e16b929 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1570,13 +1570,17 @@ process_sched_migrate_task_event(struct raw_event_sample *raw, } static void -process_raw_event(event_t *raw_event __used, void *more_data, +process_raw_event(event_t *raw_event __used, u32 size, void *data, int cpu, u64 timestamp, struct thread *thread) { - struct raw_event_sample *raw = more_data; + struct raw_event_sample *raw; struct event *event; int type; + raw = malloc_or_die(sizeof(*raw)+size); + raw->size = size; + memcpy(raw->data, data, size); + type = trace_parse_common_type(raw->data); event = trace_find_event(type); @@ -1629,7 +1633,8 @@ static int process_sample_event(event_t *event) if (profile_cpu != -1 && profile_cpu != (int)data.cpu) return 0; - process_raw_event(event, data.raw_data, data.cpu, data.time, thread); + process_raw_event(event, data.raw_size, data.raw_data, data.cpu, + data.time, thread); return 0; } -- cgit From d9541ed3241bb6c2b805d3ea0e87563cf2a0c5c3 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Mon, 7 Dec 2009 12:07:15 +0800 Subject: perf_event: Fix __dsos__write_buildid_table() The remain buff size is 'len - pos->long_name_len - 1', not 'len - pos->long_name_len + 1' This bug was introduced by commit 7691b1e ("perf tools: Misc small fixes"). Signed-off-by: Xiao Guangrong Acked-by: OGAWA Hirofumi Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: Peter Zijlstra Cc: Li Zefan LKML-Reference: <4B1C7F73.80707@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- tools/perf/util/header.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 08b6759287f..59a9c0b3033 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -209,7 +209,7 @@ static int __dsos__write_buildid_table(struct list_head *head, int fd) err = do_write(fd, pos->long_name, pos->long_name_len + 1); if (err < 0) return err; - err = do_write(fd, zero_buf, len - pos->long_name_len + 1); + err = do_write(fd, zero_buf, len - pos->long_name_len - 1); if (err < 0) return err; } -- cgit From f48f669d42e133db839af16656fd720107ef6742 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Mon, 7 Dec 2009 13:04:04 +0800 Subject: perf_event: Eliminate raw->size raw->size is not used, this patch just cleans it up. Signed-off-by: Xiao Guangrong Cc: Frederic Weisbecker Cc: Paul Mackerras Cc: OGAWA Hirofumi Cc: Peter Zijlstra Cc: Li Zefan LKML-Reference: <4B1C8CC4.4050007@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-kmem.c | 38 +++++++------------ tools/perf/builtin-sched.c | 94 +++++++++++++++++++++------------------------- 2 files changed, 56 insertions(+), 76 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index f84d7a3db68..7551a5f834b 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -57,11 +57,6 @@ static struct rb_root root_caller_sorted; static unsigned long total_requested, total_allocated; static unsigned long nr_allocs, nr_cross_allocs; -struct raw_event_sample { - u32 size; - char data[0]; -}; - #define PATH_SYS_NODE "/sys/devices/system/node" static void init_cpunode_map(void) @@ -201,7 +196,7 @@ static void insert_caller_stat(unsigned long call_site, } } -static void process_alloc_event(struct raw_event_sample *raw, +static void process_alloc_event(void *data, struct event *event, int cpu, u64 timestamp __used, @@ -214,10 +209,10 @@ static void process_alloc_event(struct raw_event_sample *raw, int bytes_alloc; int node1, node2; - ptr = raw_field_value(event, "ptr", raw->data); - call_site = raw_field_value(event, "call_site", raw->data); - bytes_req = raw_field_value(event, "bytes_req", raw->data); - bytes_alloc = raw_field_value(event, "bytes_alloc", raw->data); + ptr = raw_field_value(event, "ptr", data); + call_site = raw_field_value(event, "call_site", data); + bytes_req = raw_field_value(event, "bytes_req", data); + bytes_alloc = raw_field_value(event, "bytes_alloc", data); insert_alloc_stat(call_site, ptr, bytes_req, bytes_alloc, cpu); insert_caller_stat(call_site, bytes_req, bytes_alloc); @@ -227,7 +222,7 @@ static void process_alloc_event(struct raw_event_sample *raw, if (node) { node1 = cpunode_map[cpu]; - node2 = raw_field_value(event, "node", raw->data); + node2 = raw_field_value(event, "node", data); if (node1 != node2) nr_cross_allocs++; } @@ -262,7 +257,7 @@ static struct alloc_stat *search_alloc_stat(unsigned long ptr, return NULL; } -static void process_free_event(struct raw_event_sample *raw, +static void process_free_event(void *data, struct event *event, int cpu, u64 timestamp __used, @@ -271,7 +266,7 @@ static void process_free_event(struct raw_event_sample *raw, unsigned long ptr; struct alloc_stat *s_alloc, *s_caller; - ptr = raw_field_value(event, "ptr", raw->data); + ptr = raw_field_value(event, "ptr", data); s_alloc = search_alloc_stat(ptr, 0, &root_alloc_stat, ptr_cmp); if (!s_alloc) @@ -289,35 +284,30 @@ static void process_free_event(struct raw_event_sample *raw, } static void -process_raw_event(event_t *raw_event __used, u32 size, void *data, +process_raw_event(event_t *raw_event __used, void *data, int cpu, u64 timestamp, struct thread *thread) { - struct raw_event_sample *raw; struct event *event; int type; - raw = malloc_or_die(sizeof(*raw)+size); - raw->size = size; - memcpy(raw->data, data, size); - - type = trace_parse_common_type(raw->data); + type = trace_parse_common_type(data); event = trace_find_event(type); if (!strcmp(event->name, "kmalloc") || !strcmp(event->name, "kmem_cache_alloc")) { - process_alloc_event(raw, event, cpu, timestamp, thread, 0); + process_alloc_event(data, event, cpu, timestamp, thread, 0); return; } if (!strcmp(event->name, "kmalloc_node") || !strcmp(event->name, "kmem_cache_alloc_node")) { - process_alloc_event(raw, event, cpu, timestamp, thread, 1); + process_alloc_event(data, event, cpu, timestamp, thread, 1); return; } if (!strcmp(event->name, "kfree") || !strcmp(event->name, "kmem_cache_free")) { - process_free_event(raw, event, cpu, timestamp, thread); + process_free_event(data, event, cpu, timestamp, thread); return; } } @@ -349,7 +339,7 @@ static int process_sample_event(event_t *event) dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - process_raw_event(event, data.raw_size, data.raw_data, data.cpu, + process_raw_event(event, data.raw_data, data.cpu, data.time, thread); return 0; diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 4655e16b929..19f43faa9f8 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -628,11 +628,6 @@ static void test_calibrations(void) printf("the sleep test took %Ld nsecs\n", T1-T0); } -struct raw_event_sample { - u32 size; - char data[0]; -}; - #define FILL_FIELD(ptr, field, event, data) \ ptr.field = (typeof(ptr.field)) raw_field_value(event, #field, data) @@ -1356,7 +1351,7 @@ static void sort_lat(void) static struct trace_sched_handler *trace_handler; static void -process_sched_wakeup_event(struct raw_event_sample *raw, +process_sched_wakeup_event(void *data, struct event *event, int cpu __used, u64 timestamp __used, @@ -1364,13 +1359,13 @@ process_sched_wakeup_event(struct raw_event_sample *raw, { struct trace_wakeup_event wakeup_event; - FILL_COMMON_FIELDS(wakeup_event, event, raw->data); + FILL_COMMON_FIELDS(wakeup_event, event, data); - FILL_ARRAY(wakeup_event, comm, event, raw->data); - FILL_FIELD(wakeup_event, pid, event, raw->data); - FILL_FIELD(wakeup_event, prio, event, raw->data); - FILL_FIELD(wakeup_event, success, event, raw->data); - FILL_FIELD(wakeup_event, cpu, event, raw->data); + FILL_ARRAY(wakeup_event, comm, event, data); + FILL_FIELD(wakeup_event, pid, event, data); + FILL_FIELD(wakeup_event, prio, event, data); + FILL_FIELD(wakeup_event, success, event, data); + FILL_FIELD(wakeup_event, cpu, event, data); if (trace_handler->wakeup_event) trace_handler->wakeup_event(&wakeup_event, event, cpu, timestamp, thread); @@ -1469,7 +1464,7 @@ map_switch_event(struct trace_switch_event *switch_event, static void -process_sched_switch_event(struct raw_event_sample *raw, +process_sched_switch_event(void *data, struct event *event, int this_cpu, u64 timestamp __used, @@ -1477,15 +1472,15 @@ process_sched_switch_event(struct raw_event_sample *raw, { struct trace_switch_event switch_event; - FILL_COMMON_FIELDS(switch_event, event, raw->data); + FILL_COMMON_FIELDS(switch_event, event, data); - FILL_ARRAY(switch_event, prev_comm, event, raw->data); - FILL_FIELD(switch_event, prev_pid, event, raw->data); - FILL_FIELD(switch_event, prev_prio, event, raw->data); - FILL_FIELD(switch_event, prev_state, event, raw->data); - FILL_ARRAY(switch_event, next_comm, event, raw->data); - FILL_FIELD(switch_event, next_pid, event, raw->data); - FILL_FIELD(switch_event, next_prio, event, raw->data); + FILL_ARRAY(switch_event, prev_comm, event, data); + FILL_FIELD(switch_event, prev_pid, event, data); + FILL_FIELD(switch_event, prev_prio, event, data); + FILL_FIELD(switch_event, prev_state, event, data); + FILL_ARRAY(switch_event, next_comm, event, data); + FILL_FIELD(switch_event, next_pid, event, data); + FILL_FIELD(switch_event, next_prio, event, data); if (curr_pid[this_cpu] != (u32)-1) { /* @@ -1502,7 +1497,7 @@ process_sched_switch_event(struct raw_event_sample *raw, } static void -process_sched_runtime_event(struct raw_event_sample *raw, +process_sched_runtime_event(void *data, struct event *event, int cpu __used, u64 timestamp __used, @@ -1510,17 +1505,17 @@ process_sched_runtime_event(struct raw_event_sample *raw, { struct trace_runtime_event runtime_event; - FILL_ARRAY(runtime_event, comm, event, raw->data); - FILL_FIELD(runtime_event, pid, event, raw->data); - FILL_FIELD(runtime_event, runtime, event, raw->data); - FILL_FIELD(runtime_event, vruntime, event, raw->data); + FILL_ARRAY(runtime_event, comm, event, data); + FILL_FIELD(runtime_event, pid, event, data); + FILL_FIELD(runtime_event, runtime, event, data); + FILL_FIELD(runtime_event, vruntime, event, data); if (trace_handler->runtime_event) trace_handler->runtime_event(&runtime_event, event, cpu, timestamp, thread); } static void -process_sched_fork_event(struct raw_event_sample *raw, +process_sched_fork_event(void *data, struct event *event, int cpu __used, u64 timestamp __used, @@ -1528,12 +1523,12 @@ process_sched_fork_event(struct raw_event_sample *raw, { struct trace_fork_event fork_event; - FILL_COMMON_FIELDS(fork_event, event, raw->data); + FILL_COMMON_FIELDS(fork_event, event, data); - FILL_ARRAY(fork_event, parent_comm, event, raw->data); - FILL_FIELD(fork_event, parent_pid, event, raw->data); - FILL_ARRAY(fork_event, child_comm, event, raw->data); - FILL_FIELD(fork_event, child_pid, event, raw->data); + FILL_ARRAY(fork_event, parent_comm, event, data); + FILL_FIELD(fork_event, parent_pid, event, data); + FILL_ARRAY(fork_event, child_comm, event, data); + FILL_FIELD(fork_event, child_pid, event, data); if (trace_handler->fork_event) trace_handler->fork_event(&fork_event, event, cpu, timestamp, thread); @@ -1550,7 +1545,7 @@ process_sched_exit_event(struct event *event, } static void -process_sched_migrate_task_event(struct raw_event_sample *raw, +process_sched_migrate_task_event(void *data, struct event *event, int cpu __used, u64 timestamp __used, @@ -1558,46 +1553,42 @@ process_sched_migrate_task_event(struct raw_event_sample *raw, { struct trace_migrate_task_event migrate_task_event; - FILL_COMMON_FIELDS(migrate_task_event, event, raw->data); + FILL_COMMON_FIELDS(migrate_task_event, event, data); - FILL_ARRAY(migrate_task_event, comm, event, raw->data); - FILL_FIELD(migrate_task_event, pid, event, raw->data); - FILL_FIELD(migrate_task_event, prio, event, raw->data); - FILL_FIELD(migrate_task_event, cpu, event, raw->data); + FILL_ARRAY(migrate_task_event, comm, event, data); + FILL_FIELD(migrate_task_event, pid, event, data); + FILL_FIELD(migrate_task_event, prio, event, data); + FILL_FIELD(migrate_task_event, cpu, event, data); if (trace_handler->migrate_task_event) trace_handler->migrate_task_event(&migrate_task_event, event, cpu, timestamp, thread); } static void -process_raw_event(event_t *raw_event __used, u32 size, void *data, +process_raw_event(event_t *raw_event __used, void *data, int cpu, u64 timestamp, struct thread *thread) { - struct raw_event_sample *raw; struct event *event; int type; - raw = malloc_or_die(sizeof(*raw)+size); - raw->size = size; - memcpy(raw->data, data, size); - type = trace_parse_common_type(raw->data); + type = trace_parse_common_type(data); event = trace_find_event(type); if (!strcmp(event->name, "sched_switch")) - process_sched_switch_event(raw, event, cpu, timestamp, thread); + process_sched_switch_event(data, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_stat_runtime")) - process_sched_runtime_event(raw, event, cpu, timestamp, thread); + process_sched_runtime_event(data, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_wakeup")) - process_sched_wakeup_event(raw, event, cpu, timestamp, thread); + process_sched_wakeup_event(data, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_wakeup_new")) - process_sched_wakeup_event(raw, event, cpu, timestamp, thread); + process_sched_wakeup_event(data, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_process_fork")) - process_sched_fork_event(raw, event, cpu, timestamp, thread); + process_sched_fork_event(data, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_process_exit")) process_sched_exit_event(event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_migrate_task")) - process_sched_migrate_task_event(raw, event, cpu, timestamp, thread); + process_sched_migrate_task_event(data, event, cpu, timestamp, thread); } static int process_sample_event(event_t *event) @@ -1633,8 +1624,7 @@ static int process_sample_event(event_t *event) if (profile_cpu != -1 && profile_cpu != (int)data.cpu) return 0; - process_raw_event(event, data.raw_size, data.raw_data, data.cpu, - data.time, thread); + process_raw_event(event, data.raw_data, data.cpu, data.time, thread); return 0; } -- cgit From 3a9a0beba2913edaae39ff8b4645fee10c3acf37 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sun, 6 Dec 2009 20:41:52 -0600 Subject: perf trace/scripting: Fix compile error when libperl not installed When I added the xs callbacks into perf, I forgot to re-check the no-libperl case. This patch fixes the undefined reference error for that. Reported-by: Arnaldo Carvalho de Melo Signed-off-by: Tom Zanussi Cc: Frederic Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260153712.6564.4.camel@tropicana> Signed-off-by: Ingo Molnar --- tools/perf/util/trace-event-perl.c | 3 --- tools/perf/util/trace-event-perl.h | 4 ++++ 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/trace-event-perl.c b/tools/perf/util/trace-event-perl.c index 51e833fd58c..59564b22d9c 100644 --- a/tools/perf/util/trace-event-perl.c +++ b/tools/perf/util/trace-event-perl.c @@ -32,9 +32,6 @@ void xs_init(pTHX); -void boot_Perf__Trace__Context(pTHX_ CV *cv); -void boot_DynaLoader(pTHX_ CV *cv); - void xs_init(pTHX) { const char *file = __FILE__; diff --git a/tools/perf/util/trace-event-perl.h b/tools/perf/util/trace-event-perl.h index 8fe0d866fe1..e88fb26137b 100644 --- a/tools/perf/util/trace-event-perl.h +++ b/tools/perf/util/trace-event-perl.h @@ -34,9 +34,13 @@ typedef int INTERP; #define dXSUB_SYS #define pTHX_ static inline void newXS(const char *a, void *b, const char *c) {} +static void boot_Perf__Trace__Context(pTHX_ CV *cv) {} +static void boot_DynaLoader(pTHX_ CV *cv) {} #else #include #include +void boot_Perf__Trace__Context(pTHX_ CV *cv); +void boot_DynaLoader(pTHX_ CV *cv); typedef PerlInterpreter * INTERP; #endif -- cgit From 67a6259ec97b8408f86f2fe8459d2233f0b0987d Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Sun, 6 Dec 2009 23:31:59 -0600 Subject: perf trace/scripting: Don't display 'scripting unsupported' msg unnecessarily The 'scripting unsupported' message should only be displayed when the -s or -g options are used, and not when they aren't, as the current code does. Signed-off-by: Tom Zanussi Cc: rostedt@goodmis.org Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <1260163919-6679-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/util/trace-event-perl.c | 64 ++++++++++++++++++++++++++++++++------ 1 file changed, 55 insertions(+), 9 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/trace-event-perl.c b/tools/perf/util/trace-event-perl.c index 59564b22d9c..a5ffe60db5d 100644 --- a/tools/perf/util/trace-event-perl.c +++ b/tools/perf/util/trace-event-perl.c @@ -570,26 +570,72 @@ struct scripting_ops perl_scripting_ops = { .generate_script = perl_generate_script, }; -#ifdef NO_LIBPERL -void setup_perl_scripting(void) +static void print_unsupported_msg(void) { fprintf(stderr, "Perl scripting not supported." - " Install libperl and rebuild perf to enable it. e.g. " - "apt-get install libperl-dev (ubuntu), yum install " - "perl-ExtUtils-Embed (Fedora), etc.\n"); + " Install libperl and rebuild perf to enable it.\n" + "For example:\n # apt-get install libperl-dev (ubuntu)" + "\n # yum install perl-ExtUtils-Embed (Fedora)" + "\n etc.\n"); } -#else -void setup_perl_scripting(void) + +static int perl_start_script_unsupported(const char *script __unused) +{ + print_unsupported_msg(); + + return -1; +} + +static int perl_stop_script_unsupported(void) +{ + return 0; +} + +static void perl_process_event_unsupported(int cpu __unused, + void *data __unused, + int size __unused, + unsigned long long nsecs __unused, + char *comm __unused) +{ +} + +static int perl_generate_script_unsupported(const char *outfile __unused) +{ + print_unsupported_msg(); + + return -1; +} + +struct scripting_ops perl_scripting_unsupported_ops = { + .name = "Perl", + .start_script = perl_start_script_unsupported, + .stop_script = perl_stop_script_unsupported, + .process_event = perl_process_event_unsupported, + .generate_script = perl_generate_script_unsupported, +}; + +static void register_perl_scripting(struct scripting_ops *scripting_ops) { int err; - err = script_spec_register("Perl", &perl_scripting_ops); + err = script_spec_register("Perl", scripting_ops); if (err) die("error registering Perl script extension"); - err = script_spec_register("pl", &perl_scripting_ops); + err = script_spec_register("pl", scripting_ops); if (err) die("error registering pl script extension"); scripting_context = malloc(sizeof(struct scripting_context)); } + +#ifdef NO_LIBPERL +void setup_perl_scripting(void) +{ + register_perl_scripting(&perl_scripting_unsupported_ops); +} +#else +void setup_perl_scripting(void) +{ + register_perl_scripting(&perl_scripting_ops); +} #endif -- cgit From 180570fdb7a3c404b599f0a318c2ccf86e4827ed Mon Sep 17 00:00:00 2001 From: Ulrich Drepper Date: Sun, 6 Dec 2009 13:25:30 -0500 Subject: perf tools: Optimize parse_subsystem_tracepoint_event() Uses of strcat are almost always signs that someone is too lazy to think about the code a bit more carefully. One always has to know about the lengths of the strings involved to avoid buffer overflows. This is one case where the size of the object code for me is reduced by 38 bytes. The code should also be faster, especially if flags is non-NULL. Signed-off-by: Ulrich Drepper Cc: a.p.zijlstra@chello.nl Cc: fweisbec@gmail.com Cc: jaswinderrajput@gmail.com Cc: paulus@samba.org LKML-Reference: <200912061825.nB6IPUa1023306@hs20-bc2-1.build.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/util/parse-events.c | 15 +++------------ 1 file changed, 3 insertions(+), 12 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 448a13b5201..e5bc0fb016b 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -467,7 +467,6 @@ parse_subsystem_tracepoint_event(char *sys_name, char *flags) while ((evt_ent = readdir(evt_dir))) { char event_opt[MAX_EVOPT_LEN + 1]; int len; - unsigned int rem = MAX_EVOPT_LEN; if (!strcmp(evt_ent->d_name, ".") || !strcmp(evt_ent->d_name, "..") @@ -475,20 +474,12 @@ parse_subsystem_tracepoint_event(char *sys_name, char *flags) || !strcmp(evt_ent->d_name, "filter")) continue; - len = snprintf(event_opt, MAX_EVOPT_LEN, "%s:%s", sys_name, - evt_ent->d_name); + len = snprintf(event_opt, MAX_EVOPT_LEN, "%s:%s%s%s", sys_name, + evt_ent->d_name, flags ? ":" : "", + flags ?: ""); if (len < 0) return EVT_FAILED; - rem -= len; - if (flags) { - if (rem < strlen(flags) + 1) - return EVT_FAILED; - - strcat(event_opt, ":"); - strcat(event_opt, flags); - } - if (parse_events(NULL, event_opt, 0)) return EVT_FAILED; } -- cgit From d56728b8d7fb3e1e5e5f97b88fdf6b43a35b4f5e Mon Sep 17 00:00:00 2001 From: Juha Leppanen Date: Mon, 7 Dec 2009 12:00:40 -0500 Subject: perf probe: Fix strtailcmp() to compare s1and s2[0] Fix strtailcmp() to compare s1[0] and s2[0]. strtailcmp() returns 0 if "a" and "b" or "a" and "ab", it's a wrong behavior. This patch fixes it. Signed-off-by: "Juha Leppanen" Acked-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Juha Leppanen Cc: Frederic Weisbecker LKML-Reference: <20091207170040.19230.37464.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/util/probe-finder.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c index 293cdfc1b8c..4585f1d8679 100644 --- a/tools/perf/util/probe-finder.c +++ b/tools/perf/util/probe-finder.c @@ -106,7 +106,7 @@ static int strtailcmp(const char *s1, const char *s2) { int i1 = strlen(s1); int i2 = strlen(s2); - while (--i1 > 0 && --i2 > 0) { + while (--i1 >= 0 && --i2 >= 0) { if (s1[i1] != s2[i2]) return s1[i1] - s2[i2]; } -- cgit From e1d2017b24fb31602f1128e6a8b2afc54c9283cd Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Mon, 7 Dec 2009 12:00:46 -0500 Subject: perf probe: Fix event namelist to duplicate string Fix event namelist to duplicate string. Without duplicating, adding multiple probes causes stack overwrite bug, because it reuses a buffer on stack while the buffer is already added in the namelist. String duplication solves this bug because only contents of the buffer is copied to the namelist. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <20091207170046.19230.55557.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/util/probe-event.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index cd7fbda5e2a..de0d91385c9 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -413,12 +413,13 @@ static struct strlist *get_perf_event_names(int fd) rawlist = get_trace_kprobe_event_rawlist(fd); - sl = strlist__new(false, NULL); + sl = strlist__new(true, NULL); for (i = 0; i < strlist__nr_entries(rawlist); i++) { ent = strlist__entry(rawlist, i); parse_trace_kprobe_event(ent->s, &group, &event, NULL); strlist__add(sl, event); free(group); + free(event); } strlist__delete(rawlist); @@ -480,5 +481,6 @@ void add_trace_kprobe_events(struct probe_point *probes, int nr_probes) strlist__add(namelist, event); } } + strlist__delete(namelist); close(fd); } -- cgit From 849884508ecbe2d220131840e4cc7c32ca24ebe3 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Mon, 7 Dec 2009 12:00:53 -0500 Subject: perf probe: Check e_snprintf() format string Check e_snprintf() format string by gcc, and fix a bug of e_snprintf() caller. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <20091207170053.19230.7690.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/util/probe-event.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index de0d91385c9..88e18044993 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -47,6 +47,9 @@ #define semantic_error(msg ...) die("Semantic error :" msg) /* If there is no space to write, returns -E2BIG. */ +static int e_snprintf(char *str, size_t size, const char *format, ...) + __attribute__((format(printf, 3, 4))); + static int e_snprintf(char *str, size_t size, const char *format, ...) { int ret; @@ -258,7 +261,7 @@ int synthesize_perf_probe_event(struct probe_point *pp) ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s", pp->function, offs, pp->retprobe ? "%return" : "", line); else - ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s", pp->file, line); + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", pp->file, line); if (ret <= 0) goto error; len = ret; -- cgit From d3a2dbf844d81b4b9c9ad6044563c294e7a48cac Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Mon, 7 Dec 2009 12:00:59 -0500 Subject: perf probe: Use pr_debug for debug message Use pr_debug() for "missing vmlinux" debugging message. Signed-off-by: Masami Hiramatsu Cc: systemtap Cc: DLE Cc: Frederic Weisbecker Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <20091207170059.19230.51459.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-probe.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index a58e11b7ea8..8993a1f4e1c 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -194,8 +194,8 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) if (session.need_dwarf) die("Could not open vmlinux/module file."); - pr_warning("Could not open vmlinux/module file." - " Try to use symbols.\n"); + pr_debug("Could not open vmlinux/module file." + " Try to use symbols.\n"); goto end_dwarf; } -- cgit From 278498d438781426d8f315b65f7bca023a26fcc0 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 8 Dec 2009 17:02:40 -0500 Subject: perf probe: Change event list format Change event list format for user readability. perf probe --list shows event list in "[GROUP:EVENT] EVENT-DEFINITION" format, but this format is different from the output of perf-list, and EVENT-DEFINITION is a bit blunt. This patch changes the format to more user friendly one. Before: [probe:schedule_0] schedule+10 prev cpu After: probe:schedule_0 (on schedule+10 with prev cpu) Signed-off-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Jim Keniston Cc: Ananth N Mavinakayanahalli Cc: Christoph Hellwig Cc: Frank Ch. Eigler Cc: Frederic Weisbecker Cc: Jason Baron Cc: K.Prasad Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE LKML-Reference: <20091208220240.10142.42916.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/util/probe-event.c | 27 +++++++++++++++++++++++++-- 1 file changed, 25 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 88e18044993..a20e3827324 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -379,11 +379,29 @@ static void clear_probe_point(struct probe_point *pp) memset(pp, 0, sizeof(pp)); } +/* Show an event */ +static void show_perf_probe_event(const char *group, const char *event, + const char *place, struct probe_point *pp) +{ + int i; + char buf[128]; + + e_snprintf(buf, 128, "%s:%s", group, event); + printf(" %-40s (on %s", buf, place); + + if (pp->nr_args > 0) { + printf(" with"); + for (i = 0; i < pp->nr_args; i++) + printf(" %s", pp->args[i]); + } + printf(")\n"); +} + /* List up current perf-probe events */ void show_perf_probe_events(void) { unsigned int i; - int fd; + int fd, nr; char *group, *event; struct probe_point pp; struct strlist *rawlist; @@ -396,8 +414,13 @@ void show_perf_probe_events(void) for (i = 0; i < strlist__nr_entries(rawlist); i++) { ent = strlist__entry(rawlist, i); parse_trace_kprobe_event(ent->s, &group, &event, &pp); + /* Synthesize only event probe point */ + nr = pp.nr_args; + pp.nr_args = 0; synthesize_perf_probe_event(&pp); - printf("[%s:%s]\t%s\n", group, event, pp.probes[0]); + pp.nr_args = nr; + /* Show an event */ + show_perf_probe_event(group, event, pp.probes[0], &pp); free(group); free(event); clear_probe_point(&pp); -- cgit From a9b495b0d35859971d6896293f6d0a0d880c7dfb Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 8 Dec 2009 17:02:47 -0500 Subject: perf probe: Change probe-added message more user-friendly Change probe-added message more user-friendly expression and show usage of new events. Before: Added new event: p:probe/schedule_0 schedule+10 prev=%ax cpu=%bx After: Added new event: probe:schedule_1 (on schedule+1 with prev cpu) You can now use it on all perf tools, such as: perf record -e probe:schedule_1 -a sleep 1 Signed-off-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Jim Keniston Cc: Ananth N Mavinakayanahalli Cc: Christoph Hellwig Cc: Frank Ch. Eigler Cc: Frederic Weisbecker Cc: Jason Baron Cc: K.Prasad Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE LKML-Reference: <20091208220247.10142.91642.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/util/probe-event.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index a20e3827324..2c4d3017441 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -453,17 +453,13 @@ static struct strlist *get_perf_event_names(int fd) return sl; } -static int write_trace_kprobe_event(int fd, const char *buf) +static void write_trace_kprobe_event(int fd, const char *buf) { int ret; ret = write(fd, buf, strlen(buf)); if (ret <= 0) die("Failed to create event."); - else - printf("Added new event: %s\n", buf); - - return ret; } static void get_new_event_name(char *buf, size_t len, const char *base, @@ -503,10 +499,19 @@ void add_trace_kprobe_events(struct probe_point *probes, int nr_probes) PERFPROBE_GROUP, event, pp->probes[i]); write_trace_kprobe_event(fd, buf); + printf("Added new event:\n"); + /* Get the first parameter (probe-point) */ + sscanf(pp->probes[i], "%s", buf); + show_perf_probe_event(PERFPROBE_GROUP, event, + buf, pp); /* Add added event name to namelist */ strlist__add(namelist, event); } } + /* Show how to use the event. */ + printf("\nYou can now use it on all perf tools, such as:\n\n"); + printf("\tperf record -e %s:%s -a sleep 1\n\n", PERFPROBE_GROUP, event); + strlist__delete(namelist); close(fd); } -- cgit From d1bde3f755e8652faad59e264c466c4baab68fa8 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 8 Dec 2009 17:02:54 -0500 Subject: perf probe: Fix add-probe command syntax without --add option Fix add-probe command syntax without --add option. perf-probe supports add-probe command without --add option. But it treats each argument as an event definition. e.g. perf probe func arg1 arg2 is interpreted as perf probe --add func --add arg1 --add arg2 But it may be useless in many cases. This patch fixes this syntax to fold those arguments into one event definition if there is no --add option. With this change, above command is interpreted as below; perf probe --add "func arg1 arg2" Signed-off-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Jim Keniston Cc: Ananth N Mavinakayanahalli Cc: Christoph Hellwig Cc: Frank Ch. Eigler Cc: Frederic Weisbecker Cc: Jason Baron Cc: K.Prasad Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE LKML-Reference: <20091208220254.10142.73767.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-probe.c | 33 ++++++++++++++++++++++++++------- 1 file changed, 26 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index 8993a1f4e1c..1347fdf5337 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -79,6 +79,25 @@ static void parse_probe_event(const char *str) pr_debug("%d arguments\n", pp->nr_args); } +static void parse_probe_event_argv(int argc, const char **argv) +{ + int i, len; + char *buf; + + /* Bind up rest arguments */ + len = 0; + for (i = 0; i < argc; i++) + len += strlen(argv[i]) + 1; + buf = zalloc(len + 1); + if (!buf) + die("Failed to allocate memory for binding arguments."); + len = 0; + for (i = 0; i < argc; i++) + len += sprintf(&buf[len], "%s ", argv[i]); + parse_probe_event(buf); + free(buf); +} + static int opt_add_probe_event(const struct option *opt __used, const char *str, int unset __used) { @@ -160,7 +179,7 @@ static const struct option options[] = { int cmd_probe(int argc, const char **argv, const char *prefix __used) { - int i, j, ret; + int i, ret; #ifndef NO_LIBDWARF int fd; #endif @@ -168,8 +187,8 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) argc = parse_options(argc, argv, options, probe_usage, PARSE_OPT_STOP_AT_NON_OPTION); - for (i = 0; i < argc; i++) - parse_probe_event(argv[i]); + if (argc > 0) + parse_probe_event_argv(argc, argv); if ((session.nr_probe == 0 && !listing) || (session.nr_probe != 0 && listing)) @@ -200,8 +219,8 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) } /* Searching probe points */ - for (j = 0; j < session.nr_probe; j++) { - pp = &session.probes[j]; + for (i = 0; i < session.nr_probe; i++) { + pp = &session.probes[i]; if (pp->found) continue; @@ -223,8 +242,8 @@ end_dwarf: #endif /* !NO_LIBDWARF */ /* Synthesize probes without dwarf */ - for (j = 0; j < session.nr_probe; j++) { - pp = &session.probes[j]; + for (i = 0; i < session.nr_probe; i++) { + pp = &session.probes[i]; if (pp->found) /* This probe is already found. */ continue; -- cgit From 17f88fcd667a914b6f4dca146c9a09492fcd57b8 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 8 Dec 2009 17:03:02 -0500 Subject: perf probe: Remove event suffix number _0 Remove event suffix number _0 if it is the first. The first event has no suffix, and from the second, each event has suffix number counted from _1. This reduces typing cost :-). Signed-off-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Jim Keniston Cc: Ananth N Mavinakayanahalli Cc: Christoph Hellwig Cc: Frank Ch. Eigler Cc: Frederic Weisbecker Cc: Jason Baron Cc: K.Prasad Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE LKML-Reference: <20091208220301.10142.50031.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/util/probe-event.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 2c4d3017441..31beedcf61c 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -466,7 +466,16 @@ static void get_new_event_name(char *buf, size_t len, const char *base, struct strlist *namelist) { int i, ret; - for (i = 0; i < MAX_EVENT_INDEX; i++) { + + /* Try no suffix */ + ret = e_snprintf(buf, len, "%s", base); + if (ret < 0) + die("snprintf() failed: %s", strerror(-ret)); + if (!strlist__has_entry(namelist, buf)) + return; + + /* Try to add suffix */ + for (i = 1; i < MAX_EVENT_INDEX; i++) { ret = e_snprintf(buf, len, "%s_%d", base, i); if (ret < 0) die("snprintf() failed: %s", strerror(-ret)); -- cgit From f984f03da35357b23d53e9cad29e909810857451 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 8 Dec 2009 17:03:09 -0500 Subject: perf probe: Support vmlinux on cwd by default Support vmlinux on current working direcotry by default and also update file-open messages. Now perf probe searches ./vmlinux too. Signed-off-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Jim Keniston Cc: Ananth N Mavinakayanahalli Cc: Christoph Hellwig Cc: Frank Ch. Eigler Cc: Frederic Weisbecker Cc: Jason Baron Cc: K.Prasad Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE LKML-Reference: <20091208220309.10142.33040.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-probe.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index 1347fdf5337..1c97e133a3f 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -43,11 +43,12 @@ #include "util/probe-event.h" /* Default vmlinux search paths */ -#define NR_SEARCH_PATH 3 +#define NR_SEARCH_PATH 4 const char *default_search_path[NR_SEARCH_PATH] = { "/lib/modules/%s/build/vmlinux", /* Custom build kernel */ "/usr/lib/debug/lib/modules/%s/vmlinux", /* Red Hat debuginfo */ "/boot/vmlinux-debug-%s", /* Ubuntu */ +"./vmlinux", /* CWD */ }; #define MAX_PATH_LEN 256 @@ -205,13 +206,14 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) #else /* !NO_LIBDWARF */ pr_debug("Some probes require debuginfo.\n"); - if (session.vmlinux) + if (session.vmlinux) { + pr_debug("Try to open %s.", session.vmlinux); fd = open(session.vmlinux, O_RDONLY); - else + } else fd = open_default_vmlinux(); if (fd < 0) { if (session.need_dwarf) - die("Could not open vmlinux/module file."); + die("Could not open debuginfo file."); pr_debug("Could not open vmlinux/module file." " Try to use symbols.\n"); -- cgit From fa28244d12337eebcc620b23852ec3cf29582ff9 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 8 Dec 2009 17:03:23 -0500 Subject: perf probe: Support --del option Support perf probe --del option. Currently, perf probe can have only one event for each --del option. If you'd like to delete several probe events, you need to specify --del for each events. Signed-off-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Jim Keniston Cc: Ananth N Mavinakayanahalli Cc: Christoph Hellwig Cc: Frank Ch. Eigler Cc: Frederic Weisbecker Cc: Jason Baron Cc: K.Prasad Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE LKML-Reference: <20091208220323.10142.62079.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-probe.c | 33 +++++++++++++++++++-- tools/perf/util/probe-event.c | 69 ++++++++++++++++++++++++++++++++++++++++--- tools/perf/util/probe-event.h | 1 + 3 files changed, 96 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index 1c97e133a3f..5a47c1e11f7 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -35,6 +35,7 @@ #include "perf.h" #include "builtin.h" #include "util/util.h" +#include "util/strlist.h" #include "util/event.h" #include "util/debug.h" #include "util/parse-options.h" @@ -61,6 +62,7 @@ static struct { int need_dwarf; int nr_probe; struct probe_point probes[MAX_PROBES]; + struct strlist *dellist; } session; static bool listing; @@ -107,6 +109,17 @@ static int opt_add_probe_event(const struct option *opt __used, return 0; } +static int opt_del_probe_event(const struct option *opt __used, + const char *str, int unset __used) +{ + if (str) { + if (!session.dellist) + session.dellist = strlist__new(true, NULL); + strlist__add(session.dellist, str); + } + return 0; +} + #ifndef NO_LIBDWARF static int open_default_vmlinux(void) { @@ -141,6 +154,7 @@ static int open_default_vmlinux(void) static const char * const probe_usage[] = { "perf probe [] 'PROBEDEF' ['PROBEDEF' ...]", "perf probe [] --add 'PROBEDEF' [--add 'PROBEDEF' ...]", + "perf probe [] --del '[GROUP:]EVENT' ...", "perf probe --list", NULL }; @@ -152,7 +166,9 @@ static const struct option options[] = { OPT_STRING('k', "vmlinux", &session.vmlinux, "file", "vmlinux/module pathname"), #endif - OPT_BOOLEAN('l', "list", &listing, "list up current probes"), + OPT_BOOLEAN('l', "list", &listing, "list up current probe events"), + OPT_CALLBACK('d', "del", NULL, "[GROUP:]EVENT", "delete a probe event.", + opt_del_probe_event), OPT_CALLBACK('a', "add", NULL, #ifdef NO_LIBDWARF "FUNC[+OFFS|%return] [ARG ...]", @@ -191,15 +207,26 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) if (argc > 0) parse_probe_event_argv(argc, argv); - if ((session.nr_probe == 0 && !listing) || - (session.nr_probe != 0 && listing)) + if ((session.nr_probe == 0 && !session.dellist && !listing)) usage_with_options(probe_usage, options); if (listing) { + if (session.nr_probe != 0 || session.dellist) { + pr_warning(" Error: Don't use --list with" + " --add/--del.\n"); + usage_with_options(probe_usage, options); + } show_perf_probe_events(); return 0; } + if (session.dellist) { + del_trace_kprobe_events(session.dellist); + strlist__delete(session.dellist); + if (session.nr_probe == 0) + return 0; + } + if (session.need_dwarf) #ifdef NO_LIBDWARF die("Debuginfo-analysis is not supported"); diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 31beedcf61c..9480d9941cc 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -430,10 +430,11 @@ void show_perf_probe_events(void) } /* Get current perf-probe event names */ -static struct strlist *get_perf_event_names(int fd) +static struct strlist *get_perf_event_names(int fd, bool include_group) { unsigned int i; char *group, *event; + char buf[128]; struct strlist *sl, *rawlist; struct str_node *ent; @@ -443,7 +444,12 @@ static struct strlist *get_perf_event_names(int fd) for (i = 0; i < strlist__nr_entries(rawlist); i++) { ent = strlist__entry(rawlist, i); parse_trace_kprobe_event(ent->s, &group, &event, NULL); - strlist__add(sl, event); + if (include_group) { + if (e_snprintf(buf, 128, "%s:%s", group, event) < 0) + die("Failed to copy group:event name."); + strlist__add(sl, buf); + } else + strlist__add(sl, event); free(group); free(event); } @@ -457,9 +463,10 @@ static void write_trace_kprobe_event(int fd, const char *buf) { int ret; + pr_debug("Writing event: %s\n", buf); ret = write(fd, buf, strlen(buf)); if (ret <= 0) - die("Failed to create event."); + die("Failed to write event: %s", strerror(errno)); } static void get_new_event_name(char *buf, size_t len, const char *base, @@ -496,7 +503,7 @@ void add_trace_kprobe_events(struct probe_point *probes, int nr_probes) fd = open_kprobe_events(O_RDWR, O_APPEND); /* Get current event names */ - namelist = get_perf_event_names(fd); + namelist = get_perf_event_names(fd, false); for (j = 0; j < nr_probes; j++) { pp = probes + j; @@ -524,3 +531,57 @@ void add_trace_kprobe_events(struct probe_point *probes, int nr_probes) strlist__delete(namelist); close(fd); } + +static void del_trace_kprobe_event(int fd, const char *group, + const char *event, struct strlist *namelist) +{ + char buf[128]; + + if (e_snprintf(buf, 128, "%s:%s", group, event) < 0) + die("Failed to copy event."); + if (!strlist__has_entry(namelist, buf)) { + pr_warning("Warning: event \"%s\" is not found.\n", buf); + return; + } + /* Convert from perf-probe event to trace-kprobe event */ + if (e_snprintf(buf, 128, "-:%s/%s", group, event) < 0) + die("Failed to copy event."); + + write_trace_kprobe_event(fd, buf); + printf("Remove event: %s:%s\n", group, event); +} + +void del_trace_kprobe_events(struct strlist *dellist) +{ + int fd; + unsigned int i; + const char *group, *event; + char *p, *str; + struct str_node *ent; + struct strlist *namelist; + + fd = open_kprobe_events(O_RDWR, O_APPEND); + /* Get current event names */ + namelist = get_perf_event_names(fd, true); + + for (i = 0; i < strlist__nr_entries(dellist); i++) { + ent = strlist__entry(dellist, i); + str = strdup(ent->s); + if (!str) + die("Failed to copy event."); + p = strchr(str, ':'); + if (p) { + group = str; + *p = '\0'; + event = p + 1; + } else { + group = PERFPROBE_GROUP; + event = str; + } + del_trace_kprobe_event(fd, group, event, namelist); + free(str); + } + strlist__delete(namelist); + close(fd); +} + diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h index 0c6fe56fe38..f752159124a 100644 --- a/tools/perf/util/probe-event.h +++ b/tools/perf/util/probe-event.h @@ -10,6 +10,7 @@ extern void parse_trace_kprobe_event(const char *str, char **group, char **event, struct probe_point *pp); extern int synthesize_trace_kprobe_event(struct probe_point *pp); extern void add_trace_kprobe_events(struct probe_point *probes, int nr_probes); +extern void del_trace_kprobe_events(struct strlist *dellist); extern void show_perf_probe_events(void); /* Maximum index number of event-name postfix */ -- cgit From c937fe20cb6d9e24c6ad5f9f0c64d64c78411057 Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Tue, 8 Dec 2009 17:03:30 -0500 Subject: perf probe: Update perf-probe document Add --list and --del option descriptions to perf-probe.txt. Signed-off-by: Masami Hiramatsu Cc: Steven Rostedt Cc: Jim Keniston Cc: Ananth N Mavinakayanahalli Cc: Christoph Hellwig Cc: Frank Ch. Eigler Cc: Frederic Weisbecker Cc: Jason Baron Cc: K.Prasad Cc: Peter Zijlstra Cc: Srikar Dronamraju Cc: Arnaldo Carvalho de Melo Cc: systemtap Cc: DLE Cc: Frederic Weisbecker LKML-Reference: <20091208220330.10142.73296.stgit@dhcp-100-2-132.bos.redhat.com> Signed-off-by: Ingo Molnar --- tools/perf/Documentation/perf-probe.txt | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-probe.txt b/tools/perf/Documentation/perf-probe.txt index 9270594e6df..8fa6bf99fcb 100644 --- a/tools/perf/Documentation/perf-probe.txt +++ b/tools/perf/Documentation/perf-probe.txt @@ -8,10 +8,13 @@ perf-probe - Define new dynamic tracepoints SYNOPSIS -------- [verse] -'perf probe' [options] --add 'PROBE' [--add 'PROBE' ...] +'perf probe' [options] --add='PROBE' [...] or -'perf probe' [options] 'PROBE' ['PROBE' ...] - +'perf probe' [options] PROBE +or +'perf probe' [options] --del='[GROUP:]EVENT' [...] +or +'perf probe' --list DESCRIPTION ----------- @@ -31,8 +34,16 @@ OPTIONS Be more verbose (show parsed arguments, etc). -a:: ---add:: - Define a probe point (see PROBE SYNTAX for detail) +--add=:: + Define a probe event (see PROBE SYNTAX for detail). + +-d:: +--del=:: + Delete a probe event. + +-l:: +--list:: + List up current probe events. PROBE SYNTAX ------------ -- cgit From c0c9e72150c07b4a6766cd24a6f685ec2dc9c343 Mon Sep 17 00:00:00 2001 From: Xiao Guangrong Date: Wed, 9 Dec 2009 17:51:30 +0800 Subject: perf sched: Fix for getting task's execution time In current code, task's execute time is got by reading '/proc//sched' file, it's wrong if the task is created by pthread_create(), because every thread task has same pid. This way also has two demerits: 1: 'perf sched replay' can't work if the kernel is not compiled with the 'CONFIG_SCHED_DEBUG' option 2: perf tool should depend on proc file system So, this patch uses PERF_COUNT_SW_TASK_CLOCK to get task's execution time instead of reading /proc file. Changelog v2 -> v3: use PERF_COUNT_SW_TASK_CLOCK instead of rusage() as Ingo's suggestion Reported-by: Torok Edwin Signed-off-by: Xiao Guangrong Cc: Xiao Guangrong Cc: Peter Zijlstra Cc: Frederic Weisbecker Cc: Paul Mackerras LKML-Reference: <4B1F7322.80103@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-sched.c | 55 +++++++++++++++++++++++----------------------- 1 file changed, 27 insertions(+), 28 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 19f43faa9f8..b12b23ac06f 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -13,7 +13,6 @@ #include "util/debug.h" #include "util/data_map.h" -#include #include #include @@ -414,34 +413,33 @@ static u64 get_cpu_usage_nsec_parent(void) return sum; } -static u64 get_cpu_usage_nsec_self(void) +static int self_open_counters(void) { - char filename [] = "/proc/1234567890/sched"; - unsigned long msecs, nsecs; - char *line = NULL; - u64 total = 0; - size_t len = 0; - ssize_t chars; - FILE *file; - int ret; + struct perf_event_attr attr; + int fd; - sprintf(filename, "/proc/%d/sched", getpid()); - file = fopen(filename, "r"); - BUG_ON(!file); + memset(&attr, 0, sizeof(attr)); - while ((chars = getline(&line, &len, file)) != -1) { - ret = sscanf(line, "se.sum_exec_runtime : %ld.%06ld\n", - &msecs, &nsecs); - if (ret == 2) { - total = msecs*1e6 + nsecs; - break; - } - } - if (line) - free(line); - fclose(file); + attr.type = PERF_TYPE_SOFTWARE; + attr.config = PERF_COUNT_SW_TASK_CLOCK; - return total; + fd = sys_perf_event_open(&attr, 0, -1, -1, 0); + + if (fd < 0) + die("Error: sys_perf_event_open() syscall returned" + "with %d (%s)\n", fd, strerror(errno)); + return fd; +} + +static u64 get_cpu_usage_nsec_self(int fd) +{ + u64 runtime; + int ret; + + ret = read(fd, &runtime, sizeof(runtime)); + BUG_ON(ret != sizeof(runtime)); + + return runtime; } static void *thread_func(void *ctx) @@ -450,9 +448,11 @@ static void *thread_func(void *ctx) u64 cpu_usage_0, cpu_usage_1; unsigned long i, ret; char comm2[22]; + int fd; sprintf(comm2, ":%s", this_task->comm); prctl(PR_SET_NAME, comm2); + fd = self_open_counters(); again: ret = sem_post(&this_task->ready_for_work); @@ -462,16 +462,15 @@ again: ret = pthread_mutex_unlock(&start_work_mutex); BUG_ON(ret); - cpu_usage_0 = get_cpu_usage_nsec_self(); + cpu_usage_0 = get_cpu_usage_nsec_self(fd); for (i = 0; i < this_task->nr_events; i++) { this_task->curr_event = i; process_sched_event(this_task, this_task->atoms[i]); } - cpu_usage_1 = get_cpu_usage_nsec_self(); + cpu_usage_1 = get_cpu_usage_nsec_self(fd); this_task->cpu_usage = cpu_usage_1 - cpu_usage_0; - ret = sem_post(&this_task->work_done_sem); BUG_ON(ret); -- cgit From 5660ce34241ab204bf78fbcaa5e09318c2748d37 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Wed, 9 Dec 2009 20:26:18 +0100 Subject: perf tools: Correct size given to memset Memset should be given the size of the structure, not the size of the pointer. The semantic patch that makes this change is as follows: (http://coccinelle.lip6.fr/) // @@ type T; T *x; expression E; @@ memset(x, E, sizeof( + * x)) // Signed-off-by: Julia Lawall Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: Signed-off-by: Ingo Molnar --- tools/perf/util/probe-event.c | 2 +- tools/perf/util/trace-event-parse.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 9480d9941cc..d14a4585bca 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -376,7 +376,7 @@ static void clear_probe_point(struct probe_point *pp) free(pp->args); for (i = 0; i < pp->found; i++) free(pp->probes[i]); - memset(pp, 0, sizeof(pp)); + memset(pp, 0, sizeof(*pp)); } /* Show an event */ diff --git a/tools/perf/util/trace-event-parse.c b/tools/perf/util/trace-event-parse.c index 6ffe9d63d85..c5c32be040b 100644 --- a/tools/perf/util/trace-event-parse.c +++ b/tools/perf/util/trace-event-parse.c @@ -1477,7 +1477,7 @@ process_fields(struct event *event, struct print_flag_sym **list, char **tok) goto out_free; field = malloc_or_die(sizeof(*field)); - memset(field, 0, sizeof(field)); + memset(field, 0, sizeof(*field)); value = arg_eval(arg); field->value = strdup(value); -- cgit From 3786310afe738070be31c439b8deeaeb69b9735d Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Wed, 9 Dec 2009 21:40:08 +0100 Subject: perf sched: Add max delay time snapshot When we have a maximum latency reported for a task, we need a convenient way to find the matching location to the raw traces or to perf sched map that shows where the task has been eventually scheduled in. This gives a pointer to retrieve the events that occured during this max latency. Signed-off-by: Frederic Weisbecker Reviewed-by: Xiao Guangrong Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras LKML-Reference: <1260391208-6808-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-sched.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index b12b23ac06f..7cca7c15b40 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -140,6 +140,7 @@ struct work_atoms { struct thread *thread; struct rb_node node; u64 max_lat; + u64 max_lat_at; u64 total_lat; u64 nb_atoms; u64 total_runtime; @@ -1013,8 +1014,10 @@ add_sched_in_event(struct work_atoms *atoms, u64 timestamp) delta = atom->sched_in_time - atom->wake_up_time; atoms->total_lat += delta; - if (delta > atoms->max_lat) + if (delta > atoms->max_lat) { atoms->max_lat = delta; + atoms->max_lat_at = timestamp; + } atoms->nb_atoms++; } @@ -1210,10 +1213,11 @@ static void output_lat_thread(struct work_atoms *work_list) avg = work_list->total_lat / work_list->nb_atoms; - printf("|%11.3f ms |%9llu | avg:%9.3f ms | max:%9.3f ms |\n", + printf("|%11.3f ms |%9llu | avg:%9.3f ms | max:%9.3f ms | max at: %9.6f s\n", (double)work_list->total_runtime / 1e6, work_list->nb_atoms, (double)avg / 1e6, - (double)work_list->max_lat / 1e6); + (double)work_list->max_lat / 1e6, + (double)work_list->max_lat_at / 1e9); } static int pid_cmp(struct work_atoms *l, struct work_atoms *r) @@ -1704,9 +1708,9 @@ static void __cmd_lat(void) read_events(); sort_lat(); - printf("\n -----------------------------------------------------------------------------------------\n"); - printf(" Task | Runtime ms | Switches | Average delay ms | Maximum delay ms |\n"); - printf(" -----------------------------------------------------------------------------------------\n"); + printf("\n ---------------------------------------------------------------------------------------------------------------\n"); + printf(" Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | Maximum delay at |\n"); + printf(" ---------------------------------------------------------------------------------------------------------------\n"); next = rb_first(&sorted_atom_root); -- cgit From 90b86a9f7dc22e7ff8e8c79ed553860454ff8dd9 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 10 Dec 2009 15:21:57 +0800 Subject: perf kmem: Show usage if no option is specified As Ingo suggested, make "perf kmem" show help information. "perf kmem stat [--caller] [--alloc] .." will show memory statistics. Signed-off-by: Li Zefan Acked-by: Pekka Enberg LKML-Reference: <4B20A195.8030106@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- tools/perf/Documentation/perf-kmem.txt | 13 +++++---- tools/perf/builtin-kmem.c | 52 +++++++++++++++++++--------------- 2 files changed, 37 insertions(+), 28 deletions(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-kmem.txt b/tools/perf/Documentation/perf-kmem.txt index 44b0ce35c28..eac4d852e7c 100644 --- a/tools/perf/Documentation/perf-kmem.txt +++ b/tools/perf/Documentation/perf-kmem.txt @@ -8,16 +8,16 @@ perf-kmem - Tool to trace/measure kernel memory(slab) properties SYNOPSIS -------- [verse] -'perf kmem' {record} [] +'perf kmem' {record|stat} [] DESCRIPTION ----------- -There's two variants of perf kmem: +There are two variants of perf kmem: 'perf kmem record ' to record the kmem events of an arbitrary workload. - 'perf kmem' to report kernel memory statistics. + 'perf kmem stat' to report kernel memory statistics. OPTIONS ------- @@ -25,8 +25,11 @@ OPTIONS --input=:: Select the input file (default: perf.data) ---stat=:: - Select per callsite or per allocation statistics +--caller:: + Show per-callsite statistics + +--alloc:: + Show per-allocation statistics -s :: --sort=:: diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 7551a5f834b..1b04787ed90 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -526,7 +526,7 @@ static int __cmd_kmem(void) } static const char * const kmem_usage[] = { - "perf kmem [] {record}", + "perf kmem [] {record|stat}", NULL }; @@ -686,18 +686,17 @@ static int parse_sort_opt(const struct option *opt __used, return 0; } -static int parse_stat_opt(const struct option *opt __used, +static int parse_caller_opt(const struct option *opt __used, const char *arg, int unset __used) { - if (!arg) - return -1; + caller_flag = (alloc_flag + 1); + return 0; +} - if (strcmp(arg, "alloc") == 0) - alloc_flag = (caller_flag + 1); - else if (strcmp(arg, "caller") == 0) - caller_flag = (alloc_flag + 1); - else - return -1; +static int parse_alloc_opt(const struct option *opt __used, + const char *arg, int unset __used) +{ + alloc_flag = (caller_flag + 1); return 0; } @@ -722,14 +721,17 @@ static int parse_line_opt(const struct option *opt __used, static const struct option kmem_options[] = { OPT_STRING('i', "input", &input_name, "file", "input file name"), - OPT_CALLBACK(0, "stat", NULL, "|", - "stat selector, Pass 'alloc' or 'caller'.", - parse_stat_opt), + OPT_CALLBACK_NOOPT(0, "caller", NULL, NULL, + "show per-callsite statistics", + parse_caller_opt), + OPT_CALLBACK_NOOPT(0, "alloc", NULL, NULL, + "show per-allocation statistics", + parse_alloc_opt), OPT_CALLBACK('s', "sort", NULL, "key[,key2...]", "sort by keys: ptr, call_site, bytes, hit, pingpong, frag", parse_sort_opt), OPT_CALLBACK('l', "line", NULL, "num", - "show n lins", + "show n lines", parse_line_opt), OPT_BOOLEAN(0, "raw-ip", &raw_ip, "show raw ip instead of symbol"), OPT_END() @@ -773,18 +775,22 @@ int cmd_kmem(int argc, const char **argv, const char *prefix __used) argc = parse_options(argc, argv, kmem_options, kmem_usage, 0); - if (argc && !strncmp(argv[0], "rec", 3)) - return __cmd_record(argc, argv); - else if (argc) + if (!argc) usage_with_options(kmem_usage, kmem_options); - if (list_empty(&caller_sort)) - setup_sorting(&caller_sort, default_sort_order); - if (list_empty(&alloc_sort)) - setup_sorting(&alloc_sort, default_sort_order); + if (!strncmp(argv[0], "rec", 3)) { + return __cmd_record(argc, argv); + } else if (!strcmp(argv[0], "stat")) { + setup_cpunode_map(); + + if (list_empty(&caller_sort)) + setup_sorting(&caller_sort, default_sort_order); + if (list_empty(&alloc_sort)) + setup_sorting(&alloc_sort, default_sort_order); - setup_cpunode_map(); + return __cmd_kmem(); + } - return __cmd_kmem(); + return 0; } -- cgit From bc3abfb1b50964ffbbd0fc4e1ffe598b1b63a8c7 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 10 Dec 2009 15:22:17 +0800 Subject: perf tools: Align long options which have no short forms Before: $ ./perf kmem ... -l, --line show n lines --raw-ip show raw ip instead of symbol After: $ ./perf kmem ... -l, --line show n lines --raw-ip show raw ip instead of symbol Signed-off-by: Li Zefan Cc: Pekka Enberg LKML-Reference: <4B20A1A9.3040104@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- tools/perf/util/parse-options.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'tools') diff --git a/tools/perf/util/parse-options.c b/tools/perf/util/parse-options.c index 6d8af48c925..efebd5b476b 100644 --- a/tools/perf/util/parse-options.c +++ b/tools/perf/util/parse-options.c @@ -430,6 +430,9 @@ int usage_with_options_internal(const char * const *usagestr, pos = fprintf(stderr, " "); if (opts->short_name) pos += fprintf(stderr, "-%c", opts->short_name); + else + pos += fprintf(stderr, " "); + if (opts->long_name && opts->short_name) pos += fprintf(stderr, ", "); if (opts->long_name) -- cgit From 8b4825bf8da5c07e80496b749e9a50d675df4119 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 9 Dec 2009 20:09:37 -0200 Subject: perf symbols: dsos__read_build_ids() should read both user and kernel buildids MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260396578-19116-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/symbol.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index fffcb937cdc..e7508ad3450 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -938,8 +938,9 @@ static bool __dsos__read_build_ids(struct list_head *head) bool dsos__read_build_ids(void) { - return __dsos__read_build_ids(&dsos__kernel) || - __dsos__read_build_ids(&dsos__user); + bool kbuildids = __dsos__read_build_ids(&dsos__kernel), + ubuildids = __dsos__read_build_ids(&dsos__user); + return kbuildids || ubuildids; } /* -- cgit From 716d69e4fda0563ef67d62ee44baa17b377b9b23 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Wed, 9 Dec 2009 20:09:38 -0200 Subject: perf symbols: perf_header__read_build_ids() offset'n'size should be u64 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As off_t is a long, so breaking things on 32-bit land. Now buildids work on 32-bit land. [root@ana ~]# uname -a Linux ana.ghostprotocols.net 2.6.31.6-162.fc12.i686 #1 SMP Fri Dec 4 01:09:09 EST 2009 i686 i686 i386 GNU/Linux [root@ana ~]# perf buildid-list | tail -5 136ee6792ba2ae57870ecd87369f4ae3194d5b27 /lib/libreadline.so.6.0 d202dcb1ad48d140065783657d37ae3f2d9ab83f /usr/bin/gdb 0a56c0c00dcc2e9e581ae9997f31957c9c4671df /usr/lib/libdwarf.so.0.0 5f9e6ac95241cbb3227608e0ff2a2e0cbbe72439 /home/acme/bin/perf 925d19eccc2ddb1c9d74dd178a011426f1b124a8 /bin/sleep [root@ana ~]# Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260396578-19116-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/data_map.c | 4 ++-- tools/perf/util/data_map.h | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index ca0bedf637c..59b65d0bd7c 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -100,11 +100,11 @@ process_event(event_t *event, unsigned long offset, unsigned long head) } } -int perf_header__read_build_ids(int input, off_t offset, off_t size) +int perf_header__read_build_ids(int input, u64 offset, u64 size) { struct build_id_event bev; char filename[PATH_MAX]; - off_t limit = offset + size; + u64 limit = offset + size; int err = -1; while (offset < limit) { diff --git a/tools/perf/util/data_map.h b/tools/perf/util/data_map.h index 3180ff7e363..258a87bcc4f 100644 --- a/tools/perf/util/data_map.h +++ b/tools/perf/util/data_map.h @@ -27,6 +27,6 @@ int mmap_dispatch_perf_file(struct perf_header **pheader, int full_paths, int *cwdlen, char **cwd); -int perf_header__read_build_ids(int input, off_t offset, off_t file_size); +int perf_header__read_build_ids(int input, u64 offset, u64 file_size); #endif -- cgit From 7931241694b25589658b1ceb02218d2750540ae0 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Thu, 10 Dec 2009 08:43:34 +0100 Subject: perf kmem: Fix unused argument build warning Fix: builtin-kmem.c: In function 'parse_caller_opt': builtin-kmem.c:690: error: unused parameter 'arg' builtin-kmem.c: In function 'parse_alloc_opt': builtin-kmem.c:697: error: unused parameter 'arg' Cc: Li Zefan Cc: Pekka Enberg LKML-Reference: <4B20A195.8030106@cn.fujitsu.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-kmem.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 1b04787ed90..5f209514f65 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -687,14 +687,14 @@ static int parse_sort_opt(const struct option *opt __used, } static int parse_caller_opt(const struct option *opt __used, - const char *arg, int unset __used) + const char *arg __used, int unset __used) { caller_flag = (alloc_flag + 1); return 0; } static int parse_alloc_opt(const struct option *opt __used, - const char *arg, int unset __used) + const char *arg __used, int unset __used) { alloc_flag = (caller_flag + 1); return 0; -- cgit From cc835752ae3634acd2d487fdf5152f6075f45aef Mon Sep 17 00:00:00 2001 From: Jamie Iles Date: Fri, 11 Dec 2009 09:21:00 +0000 Subject: perf tools: Allow cross compiling For embedded platforms, we want to be able to build the perf tools on a build machine to run on a different arch. This patch allows $CROSS_COMPILE to set the cross compiler. Additionally, if NO_LIBPERL is set, then don't use perl include paths as they will be for the host arch. Signed-off-by: Jamie Iles Cc: Peter Zijlstra LKML-Reference: <1260523260-15694-2-git-send-email-jamie.iles@picochip.com> Signed-off-by: Ingo Molnar --- tools/perf/Makefile | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 23ec66098bd..e2ee3b589af 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -237,8 +237,8 @@ lib = lib export prefix bindir sharedir sysconfdir -CC = gcc -AR = ar +CC = $(CROSS_COMPILE)gcc +AR = $(CROSS_COMPILE)ar RM = rm -f TAR = tar FIND = find @@ -492,8 +492,10 @@ else LIB_OBJS += util/probe-finder.o endif +ifndef NO_LIBPERL PERL_EMBED_LDOPTS = `perl -MExtUtils::Embed -e ldopts 2>/dev/null` PERL_EMBED_CCOPTS = `perl -MExtUtils::Embed -e ccopts 2>/dev/null` +endif ifneq ($(shell sh -c "(echo '\#include '; echo '\#include '; echo 'int main(void) { perl_alloc(); return 0; }') | $(CC) -x c - $(PERL_EMBED_CCOPTS) -o /dev/null $(PERL_EMBED_LDOPTS) > /dev/null 2>&1 && echo y"), y) BASIC_CFLAGS += -DNO_LIBPERL -- cgit From 58e9f94138c1d9c47f6a63632ca7a78fc6dcc15f Mon Sep 17 00:00:00 2001 From: Jamie Iles Date: Fri, 11 Dec 2009 12:20:09 +0000 Subject: perf tools: Allow building for ARM Add definitions of rmb() and cpu_relax() and include the ARM unistd.h header. The __kuser_memory_barrier helper in the helper page is used to provide the correct memory barrier depending on the CPU type. [ The rmb() will work on v6 and v7, segfault on v5. Dynamic detection to add v5 support will be added later. ] Signed-off-by: Jamie Iles Cc: Russell King Cc: Peter Zijlstra Cc: Mikael Pettersson LKML-Reference: <1260534009-5394-1-git-send-email-jamie.iles@picochip.com> Signed-off-by: Ingo Molnar --- tools/perf/perf.h | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'tools') diff --git a/tools/perf/perf.h b/tools/perf/perf.h index 454d5d55f32..75f941bfba9 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -59,6 +59,18 @@ #define cpu_relax() asm volatile ("hint @pause" ::: "memory") #endif +#ifdef __arm__ +#include "../../arch/arm/include/asm/unistd.h" +/* + * Use the __kuser_memory_barrier helper in the CPU helper page. See + * arch/arm/kernel/entry-armv.S in the kernel source for details. + */ +#define rmb() asm volatile("mov r0, #0xffff0fff; mov lr, pc;" \ + "sub pc, r0, #95" ::: "r0", "lr", "cc", \ + "memory") +#define cpu_relax() asm volatile("":::"memory") +#endif + #include #include #include -- cgit From 9958e1f0aee632c3665162c9c93cf8fde8006a94 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 14:50:36 -0200 Subject: perf symbols: Rename kthreads to kmaps, using another abstraction for it MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Using a struct thread instance just to hold the kernel space maps (vmlinux + modules) is overkill and confuses people trying to understand the perf symbols abstractions. The kernel maps are really present in all threads, i.e. the kernel is a library, not a separate thread. So introduce the 'map_groups' abstraction and use it for the kernel maps, now in the kmaps global variable. It, in turn, will move, together with the threads list to the perf_file abstraction, so that we can support multiple perf_file instances, needed by perf diff. Brainstormed-with: Eduardo Habkost Signed-off-by: Arnaldo Carvalho de Melo Cc: Eduardo Habkost Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260550239-5372-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-kmem.c | 2 +- tools/perf/util/event.c | 11 +++---- tools/perf/util/symbol.c | 73 ++++++++++++++++++++++++----------------------- tools/perf/util/symbol.h | 4 +-- tools/perf/util/thread.c | 62 ++++++++++++++++++++++------------------ tools/perf/util/thread.h | 39 ++++++++++++++++--------- 6 files changed, 106 insertions(+), 85 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 5f209514f65..fe73435192b 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -403,7 +403,7 @@ static void __print_result(struct rb_root *root, int n_lines, int is_caller) if (is_caller) { addr = data->call_site; if (!raw_ip) - sym = thread__find_function(kthread, addr, NULL); + sym = map_groups__find_function(kmaps, addr, NULL); } else addr = data->ptr; diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 4dcecafa85d..ba0de90cd3d 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -254,13 +254,14 @@ void thread__find_addr_location(struct thread *self, u8 cpumode, struct addr_location *al, symbol_filter_t filter) { - struct thread *thread = al->thread = self; + struct map_groups *mg = &self->mg; + al->thread = self; al->addr = addr; if (cpumode & PERF_RECORD_MISC_KERNEL) { al->level = 'k'; - thread = kthread; + mg = kmaps; } else if (cpumode & PERF_RECORD_MISC_USER) al->level = '.'; else { @@ -270,7 +271,7 @@ void thread__find_addr_location(struct thread *self, u8 cpumode, return; } try_again: - al->map = thread__find_map(thread, type, al->addr); + al->map = map_groups__find(mg, type, al->addr); if (al->map == NULL) { /* * If this is outside of all known maps, and is a negative @@ -281,8 +282,8 @@ try_again: * "[vdso]" dso, but for now lets use the old trick of looking * in the whole kernel symbol list. */ - if ((long long)al->addr < 0 && thread != kthread) { - thread = kthread; + if ((long long)al->addr < 0 && mg != kmaps) { + mg = kmaps; goto try_again; } al->sym = NULL; diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index e7508ad3450..93620234294 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -29,11 +29,11 @@ enum dso_origin { }; static void dsos__add(struct list_head *head, struct dso *dso); -static struct map *thread__find_map_by_name(struct thread *self, char *name); +static struct map *map_groups__find_by_name(struct map_groups *self, char *name); static struct map *map__new2(u64 start, struct dso *dso, enum map_type type); struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr); static int dso__load_kernel_sym(struct dso *self, struct map *map, - struct thread *thread, symbol_filter_t filter); + struct map_groups *mg, symbol_filter_t filter); unsigned int symbol__priv_size; static int vmlinux_path__nr_entries; static char **vmlinux_path; @@ -43,8 +43,8 @@ static struct symbol_conf symbol_conf__defaults = { .try_vmlinux_path = true, }; -static struct thread kthread_mem; -struct thread *kthread = &kthread_mem; +static struct map_groups kmaps_mem; +struct map_groups *kmaps = &kmaps_mem; bool dso__loaded(const struct dso *self, enum map_type type) { @@ -79,7 +79,7 @@ static void symbols__fixup_end(struct rb_root *self) curr->end = roundup(curr->start, 4096); } -static void __thread__fixup_maps_end(struct thread *self, enum map_type type) +static void __map_groups__fixup_end(struct map_groups *self, enum map_type type) { struct map *prev, *curr; struct rb_node *nd, *prevnd = rb_first(&self->maps[type]); @@ -102,11 +102,11 @@ static void __thread__fixup_maps_end(struct thread *self, enum map_type type) curr->end = ~0UL; } -static void thread__fixup_maps_end(struct thread *self) +static void map_groups__fixup_end(struct map_groups *self) { int i; for (i = 0; i < MAP__NR_TYPES; ++i) - __thread__fixup_maps_end(self, i); + __map_groups__fixup_end(self, i); } static struct symbol *symbol__new(u64 start, u64 len, const char *name) @@ -364,8 +364,8 @@ out_failure: * kernel range is broken in several maps, named [kernel].N, as we don't have * the original ELF section names vmlinux have. */ -static int dso__split_kallsyms(struct dso *self, struct map *map, struct thread *thread, - symbol_filter_t filter) +static int dso__split_kallsyms(struct dso *self, struct map *map, + struct map_groups *mg, symbol_filter_t filter) { struct map *curr_map = map; struct symbol *pos; @@ -382,13 +382,13 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, struct thread module = strchr(pos->name, '\t'); if (module) { - if (!thread->use_modules) + if (!mg->use_modules) goto discard_symbol; *module++ = '\0'; if (strcmp(self->name, module)) { - curr_map = thread__find_map_by_name(thread, module); + curr_map = map_groups__find_by_name(mg, module); if (curr_map == NULL) { pr_debug("/proc/{kallsyms,modules} " "inconsistency!\n"); @@ -419,7 +419,7 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, struct thread } curr_map->map_ip = curr_map->unmap_ip = identity__map_ip; - __thread__insert_map(thread, curr_map); + map_groups__insert(mg, curr_map); ++kernel_range; } @@ -440,7 +440,7 @@ discard_symbol: rb_erase(&pos->rb_node, root); static int dso__load_kallsyms(struct dso *self, struct map *map, - struct thread *thread, symbol_filter_t filter) + struct map_groups *mg, symbol_filter_t filter) { if (dso__load_all_kallsyms(self, map) < 0) return -1; @@ -448,13 +448,13 @@ static int dso__load_kallsyms(struct dso *self, struct map *map, symbols__fixup_end(&self->symbols[map->type]); self->origin = DSO__ORIG_KERNEL; - return dso__split_kallsyms(self, map, thread, filter); + return dso__split_kallsyms(self, map, mg, filter); } size_t kernel_maps__fprintf(FILE *fp) { size_t printed = fprintf(fp, "Kernel maps:\n"); - printed += thread__fprintf_maps(kthread, fp); + printed += map_groups__fprintf_maps(kmaps, fp); return printed + fprintf(fp, "END kernel maps\n"); } @@ -745,7 +745,7 @@ out: } static int dso__load_sym(struct dso *self, struct map *map, - struct thread *thread, const char *name, int fd, + struct map_groups *mg, const char *name, int fd, symbol_filter_t filter, int kernel, int kmodule) { struct map *curr_map = map; @@ -849,7 +849,7 @@ static int dso__load_sym(struct dso *self, struct map *map, snprintf(dso_name, sizeof(dso_name), "%s%s", self->short_name, section_name); - curr_map = thread__find_map_by_name(thread, dso_name); + curr_map = map_groups__find_by_name(mg, dso_name); if (curr_map == NULL) { u64 start = sym.st_value; @@ -868,7 +868,7 @@ static int dso__load_sym(struct dso *self, struct map *map, curr_map->map_ip = identity__map_ip; curr_map->unmap_ip = identity__map_ip; curr_dso->origin = DSO__ORIG_KERNEL; - __thread__insert_map(kthread, curr_map); + map_groups__insert(kmaps, curr_map); dsos__add(&dsos__kernel, curr_dso); } else curr_dso = curr_map->dso; @@ -1094,7 +1094,7 @@ int dso__load(struct dso *self, struct map *map, symbol_filter_t filter) dso__set_loaded(self, map->type); if (self->kernel) - return dso__load_kernel_sym(self, map, kthread, filter); + return dso__load_kernel_sym(self, map, kmaps, filter); name = malloc(size); if (!name) @@ -1180,7 +1180,7 @@ out: return ret; } -static struct map *thread__find_map_by_name(struct thread *self, char *name) +static struct map *map_groups__find_by_name(struct map_groups *self, char *name) { struct rb_node *nd; @@ -1228,7 +1228,7 @@ static int dsos__set_modules_path_dir(char *dirname) (int)(dot - dent->d_name), dent->d_name); strxfrchar(dso_name, '-', '_'); - map = thread__find_map_by_name(kthread, dso_name); + map = map_groups__find_by_name(kmaps, dso_name); if (map == NULL) continue; @@ -1281,7 +1281,7 @@ static struct map *map__new2(u64 start, struct dso *dso, enum map_type type) return self; } -static int thread__create_module_maps(struct thread *self) +static int map_groups__create_module_maps(struct map_groups *self) { char *line = NULL; size_t n; @@ -1338,7 +1338,7 @@ static int thread__create_module_maps(struct thread *self) dso->has_build_id = true; dso->origin = DSO__ORIG_KMODULE; - __thread__insert_map(self, map); + map_groups__insert(self, map); dsos__add(&dsos__kernel, dso); } @@ -1353,7 +1353,8 @@ out_failure: return -1; } -static int dso__load_vmlinux(struct dso *self, struct map *map, struct thread *thread, +static int dso__load_vmlinux(struct dso *self, struct map *map, + struct map_groups *mg, const char *vmlinux, symbol_filter_t filter) { int err = -1, fd; @@ -1387,14 +1388,14 @@ static int dso__load_vmlinux(struct dso *self, struct map *map, struct thread *t return -1; dso__set_loaded(self, map->type); - err = dso__load_sym(self, map, thread, self->long_name, fd, filter, 1, 0); + err = dso__load_sym(self, map, mg, self->long_name, fd, filter, 1, 0); close(fd); return err; } static int dso__load_kernel_sym(struct dso *self, struct map *map, - struct thread *thread, symbol_filter_t filter) + struct map_groups *mg, symbol_filter_t filter) { int err; bool is_kallsyms; @@ -1404,7 +1405,7 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, pr_debug("Looking at the vmlinux_path (%d entries long)\n", vmlinux_path__nr_entries); for (i = 0; i < vmlinux_path__nr_entries; ++i) { - err = dso__load_vmlinux(self, map, thread, + err = dso__load_vmlinux(self, map, mg, vmlinux_path[i], filter); if (err > 0) { pr_debug("Using %s for symbols\n", @@ -1420,12 +1421,12 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, if (is_kallsyms) goto do_kallsyms; - err = dso__load_vmlinux(self, map, thread, self->long_name, filter); + err = dso__load_vmlinux(self, map, mg, self->long_name, filter); if (err <= 0) { pr_info("The file %s cannot be used, " "trying to use /proc/kallsyms...", self->long_name); do_kallsyms: - err = dso__load_kallsyms(self, map, thread, filter); + err = dso__load_kallsyms(self, map, mg, filter); if (err > 0 && !is_kallsyms) dso__set_long_name(self, strdup("[kernel.kallsyms]")); } @@ -1508,7 +1509,7 @@ size_t dsos__fprintf_buildid(FILE *fp) __dsos__fprintf_buildid(&dsos__user, fp)); } -static int thread__create_kernel_map(struct thread *self, const char *vmlinux) +static int map_groups__create_kernel_map(struct map_groups *self, const char *vmlinux) { struct map *kmap; struct dso *kernel = dso__new(vmlinux ?: "[kernel.kallsyms]"); @@ -1533,7 +1534,7 @@ static int thread__create_kernel_map(struct thread *self, const char *vmlinux) sizeof(kernel->build_id)) == 0) kernel->has_build_id = true; - __thread__insert_map(self, kmap); + map_groups__insert(self, kmap); dsos__add(&dsos__kernel, kernel); dsos__add(&dsos__user, vdso); @@ -1607,23 +1608,23 @@ int symbol__init(struct symbol_conf *conf) elf_version(EV_CURRENT); symbol__priv_size = pconf->priv_size; - thread__init(kthread, 0); + map_groups__init(kmaps); if (pconf->try_vmlinux_path && vmlinux_path__init() < 0) return -1; - if (thread__create_kernel_map(kthread, pconf->vmlinux_name) < 0) { + if (map_groups__create_kernel_map(kmaps, pconf->vmlinux_name) < 0) { vmlinux_path__exit(); return -1; } - kthread->use_modules = pconf->use_modules; - if (pconf->use_modules && thread__create_module_maps(kthread) < 0) + kmaps->use_modules = pconf->use_modules; + if (pconf->use_modules && map_groups__create_module_maps(kmaps) < 0) pr_debug("Failed to load list of modules in use, " "continuing...\n"); /* * Now that we have all the maps created, just set the ->end of them: */ - thread__fixup_maps_end(kthread); + map_groups__fixup_end(kmaps); return 0; } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 17003efa0b3..6e1da1ea631 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -113,8 +113,8 @@ size_t kernel_maps__fprintf(FILE *fp); int symbol__init(struct symbol_conf *conf); -struct thread; -struct thread *kthread; +struct map_groups; +struct map_groups *kmaps; extern struct list_head dsos__user, dsos__kernel; extern struct dso *vdso; #endif /* __PERF_SYMBOL */ diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 603f5610861..a1285129c83 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -9,11 +9,9 @@ static struct rb_root threads; static struct thread *last_match; -void thread__init(struct thread *self, pid_t pid) +void map_groups__init(struct map_groups *self) { int i; - self->pid = pid; - self->comm = NULL; for (i = 0; i < MAP__NR_TYPES; ++i) { self->maps[i] = RB_ROOT; INIT_LIST_HEAD(&self->removed_maps[i]); @@ -25,7 +23,8 @@ static struct thread *thread__new(pid_t pid) struct thread *self = zalloc(sizeof(*self)); if (self != NULL) { - thread__init(self, pid); + map_groups__init(&self->mg); + self->pid = pid; self->comm = malloc(32); if (self->comm) snprintf(self->comm, 32, ":%d", self->pid); @@ -57,8 +56,8 @@ static const char *map_type__name[MAP__NR_TYPES] = { [MAP__FUNCTION] = "Functions", }; -static size_t __thread__fprintf_maps(struct thread *self, - enum map_type type, FILE *fp) +static size_t __map_groups__fprintf_maps(struct map_groups *self, + enum map_type type, FILE *fp) { size_t printed = fprintf(fp, "%s:\n", map_type__name[type]); struct rb_node *nd; @@ -76,16 +75,16 @@ static size_t __thread__fprintf_maps(struct thread *self, return printed; } -size_t thread__fprintf_maps(struct thread *self, FILE *fp) +size_t map_groups__fprintf_maps(struct map_groups *self, FILE *fp) { size_t printed = 0, i; for (i = 0; i < MAP__NR_TYPES; ++i) - printed += __thread__fprintf_maps(self, i, fp); + printed += __map_groups__fprintf_maps(self, i, fp); return printed; } -static size_t __thread__fprintf_removed_maps(struct thread *self, - enum map_type type, FILE *fp) +static size_t __map_groups__fprintf_removed_maps(struct map_groups *self, + enum map_type type, FILE *fp) { struct map *pos; size_t printed = 0; @@ -101,20 +100,25 @@ static size_t __thread__fprintf_removed_maps(struct thread *self, return printed; } -static size_t thread__fprintf_removed_maps(struct thread *self, FILE *fp) +static size_t map_groups__fprintf_removed_maps(struct map_groups *self, FILE *fp) { size_t printed = 0, i; for (i = 0; i < MAP__NR_TYPES; ++i) - printed += __thread__fprintf_removed_maps(self, i, fp); + printed += __map_groups__fprintf_removed_maps(self, i, fp); return printed; } -static size_t thread__fprintf(struct thread *self, FILE *fp) +static size_t map_groups__fprintf(struct map_groups *self, FILE *fp) { - size_t printed = fprintf(fp, "Thread %d %s\n", self->pid, self->comm); - printed += thread__fprintf_removed_maps(self, fp); + size_t printed = map_groups__fprintf_maps(self, fp); printed += fprintf(fp, "Removed maps:\n"); - return printed + thread__fprintf_removed_maps(self, fp); + return printed + map_groups__fprintf_removed_maps(self, fp); +} + +static size_t thread__fprintf(struct thread *self, FILE *fp) +{ + return fprintf(fp, "Thread %d %s\n", self->pid, self->comm) + + map_groups__fprintf(&self->mg, fp); } struct thread *threads__findnew(pid_t pid) @@ -168,7 +172,8 @@ struct thread *register_idle_thread(void) return thread; } -static void thread__remove_overlappings(struct thread *self, struct map *map) +static void map_groups__remove_overlappings(struct map_groups *self, + struct map *map) { struct rb_root *root = &self->maps[map->type]; struct rb_node *next = rb_first(root); @@ -238,12 +243,15 @@ struct map *maps__find(struct rb_root *maps, u64 ip) void thread__insert_map(struct thread *self, struct map *map) { - thread__remove_overlappings(self, map); - maps__insert(&self->maps[map->type], map); + map_groups__remove_overlappings(&self->mg, map); + map_groups__insert(&self->mg, map); } -static int thread__clone_maps(struct thread *self, struct thread *parent, - enum map_type type) +/* + * XXX This should not really _copy_ te maps, but refcount them. + */ +static int map_groups__clone(struct map_groups *self, + struct map_groups *parent, enum map_type type) { struct rb_node *nd; for (nd = rb_first(&parent->maps[type]); nd; nd = rb_next(nd)) { @@ -251,7 +259,7 @@ static int thread__clone_maps(struct thread *self, struct thread *parent, struct map *new = map__clone(map); if (new == NULL) return -ENOMEM; - thread__insert_map(self, new); + map_groups__insert(self, new); } return 0; } @@ -267,7 +275,7 @@ int thread__fork(struct thread *self, struct thread *parent) return -ENOMEM; for (i = 0; i < MAP__NR_TYPES; ++i) - if (thread__clone_maps(self, parent, i) < 0) + if (map_groups__clone(&self->mg, &parent->mg, i) < 0) return -ENOMEM; return 0; } @@ -286,11 +294,11 @@ size_t threads__fprintf(FILE *fp) return ret; } -struct symbol *thread__find_symbol(struct thread *self, - enum map_type type, u64 addr, - symbol_filter_t filter) +struct symbol *map_groups__find_symbol(struct map_groups *self, + enum map_type type, u64 addr, + symbol_filter_t filter) { - struct map *map = thread__find_map(self, type, addr); + struct map *map = map_groups__find(self, type, addr); if (map != NULL) return map__find_symbol(map, map->map_ip(map, addr), filter); diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index 686d6e914d9..a6333f3716a 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -5,52 +5,63 @@ #include #include "symbol.h" -struct thread { - struct rb_node rb_node; +struct map_groups { struct rb_root maps[MAP__NR_TYPES]; struct list_head removed_maps[MAP__NR_TYPES]; - pid_t pid; bool use_modules; +}; + +struct thread { + struct rb_node rb_node; + struct map_groups mg; + pid_t pid; char shortname[3]; char *comm; int comm_len; }; -void thread__init(struct thread *self, pid_t pid); +void map_groups__init(struct map_groups *self); int thread__set_comm(struct thread *self, const char *comm); int thread__comm_len(struct thread *self); struct thread *threads__findnew(pid_t pid); struct thread *register_idle_thread(void); void thread__insert_map(struct thread *self, struct map *map); int thread__fork(struct thread *self, struct thread *parent); -size_t thread__fprintf_maps(struct thread *self, FILE *fp); +size_t map_groups__fprintf_maps(struct map_groups *self, FILE *fp); size_t threads__fprintf(FILE *fp); void maps__insert(struct rb_root *maps, struct map *map); struct map *maps__find(struct rb_root *maps, u64 addr); -static inline struct map *thread__find_map(struct thread *self, +static inline void map_groups__insert(struct map_groups *self, struct map *map) +{ + maps__insert(&self->maps[map->type], map); +} + +static inline struct map *map_groups__find(struct map_groups *self, enum map_type type, u64 addr) { - return self ? maps__find(&self->maps[type], addr) : NULL; + return maps__find(&self->maps[type], addr); } -static inline void __thread__insert_map(struct thread *self, struct map *map) +static inline struct map *thread__find_map(struct thread *self, + enum map_type type, u64 addr) { - maps__insert(&self->maps[map->type], map); + return self ? map_groups__find(&self->mg, type, addr) : NULL; } void thread__find_addr_location(struct thread *self, u8 cpumode, enum map_type type, u64 addr, struct addr_location *al, symbol_filter_t filter); -struct symbol *thread__find_symbol(struct thread *self, - enum map_type type, u64 addr, - symbol_filter_t filter); +struct symbol *map_groups__find_symbol(struct map_groups *self, + enum map_type type, u64 addr, + symbol_filter_t filter); static inline struct symbol * -thread__find_function(struct thread *self, u64 addr, symbol_filter_t filter) +map_groups__find_function(struct map_groups *self, u64 addr, + symbol_filter_t filter) { - return thread__find_symbol(self, MAP__FUNCTION, addr, filter); + return map_groups__find_symbol(self, MAP__FUNCTION, addr, filter); } #endif /* __PERF_THREAD_H */ -- cgit From 6893d4ee67cc100348088328cac73d56f7186aa3 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 14:50:37 -0200 Subject: perf symbols: Introduce symbol_type__is_a MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For selecting the right types of symbols in /proc/kallsyms, will be followed by elf_symbol_type__is_a, for the same purpose on ELF symtabs. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260550239-5372-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/event.h | 4 ++-- tools/perf/util/symbol.c | 15 +++++++++++---- 2 files changed, 13 insertions(+), 6 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index c7a78eef8e5..d9a65d91e92 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -103,10 +103,10 @@ void event__print_totals(void); enum map_type { MAP__FUNCTION = 0, - - MAP__NR_TYPES, }; +#define MAP__NR_TYPES (MAP__FUNCTION + 1) + struct map { union { struct rb_node rb_node; diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 93620234294..bb21c96e8e1 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -56,6 +56,16 @@ static void dso__set_loaded(struct dso *self, enum map_type type) self->loaded |= (1 << type); } +static bool symbol_type__is_a(char symbol_type, enum map_type map_type) +{ + switch (map_type) { + case MAP__FUNCTION: + return symbol_type == 'T' || symbol_type == 'W'; + default: + return false; + } +} + static void symbols__fixup_end(struct rb_root *self) { struct rb_node *nd, *prevnd = rb_first(self); @@ -327,10 +337,7 @@ static int dso__load_all_kallsyms(struct dso *self, struct map *map) continue; symbol_type = toupper(line[len]); - /* - * We're interested only in code ('T'ext) - */ - if (symbol_type != 'T' && symbol_type != 'W') + if (!symbol_type__is_a(symbol_type, map->type)) continue; symbol_name = line + len + 2; -- cgit From d45868d38c1d08d50abf3e884710a938d19fa93c Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 14:50:38 -0200 Subject: perf symbols: Introduce ELF counterparts to symbol_type__is_a MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit For selecting the right types of symbols in ELF symtabs. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260550239-5372-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/symbol.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index bb21c96e8e1..db8dc97b548 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -751,6 +751,26 @@ out: return 0; } +static bool elf_sym__is_a(GElf_Sym *self, enum map_type type) +{ + switch (type) { + case MAP__FUNCTION: + return elf_sym__is_function(self); + default: + return false; + } +} + +static bool elf_sec__is_a(GElf_Shdr *self, Elf_Data *secstrs, enum map_type type) +{ + switch (type) { + case MAP__FUNCTION: + return elf_sec__is_text(self, secstrs); + default: + return false; + } +} + static int dso__load_sym(struct dso *self, struct map *map, struct map_groups *mg, const char *name, int fd, symbol_filter_t filter, int kernel, int kmodule) @@ -825,7 +845,7 @@ static int dso__load_sym(struct dso *self, struct map *map, int is_label = elf_sym__is_label(&sym); const char *section_name; - if (!is_label && !elf_sym__is_function(&sym)) + if (!is_label && !elf_sym__is_a(&sym, map->type)) continue; sec = elf_getscn(elf, sym.st_shndx); @@ -834,7 +854,7 @@ static int dso__load_sym(struct dso *self, struct map *map, gelf_getshdr(sec, &shdr); - if (is_label && !elf_sec__is_text(&shdr, secstrs)) + if (is_label && !elf_sec__is_a(&shdr, secstrs, map->type)) continue; elf_name = elf_sym__name(&sym, symstrs); -- cgit From f1dfa0b1c1a90d4d3bf515ab04a6b6222e086293 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 14:50:39 -0200 Subject: perf symbols: Add support for 'variable' symtabs MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Example: { u64 addr = strtoull(sym_filter, NULL, 16); struct map *map = map_groups__find(kmaps, MAP__VARIABLE, addr); if (map == NULL) pr_err("couldn't find map!\n"); else { struct symbol *sym = map__find_symbol(map, addr, NULL); if (sym == NULL) pr_err("couldn't find addr!\n"); else pr_info("addr %#Lx is in %s global var\n", addr, sym->name); } exit(0); } Added just after symbol__init() call in 'perf top', then: { u64 addr = strtoull(sym_filter, NULL, 16); struct map *map = map_groups__find(kmaps, MAP__VARIABLE, addr); if (map == NULL) pr_err("couldn't find map!\n"); else { struct symbol *sym = map__find_symbol(map, addr, NULL); if (sym == NULL) pr_err("couldn't find addr!\n"); else pr_info("addr %#Lx is in %s global var\n", addr, sym->name); } exit(0); } [root@doppio linux-2.6-tip]# grep ' [dD] ' /proc/kallsyms | grep ' sched' ffffffff817827d8 d sched_nr_latency ffffffff81782ce0 d sched_domains_mutex ffffffff8178c070 d schedstr.22423 ffffffff817909a0 d sched_register_mutex ffffffff81823490 d sched_feat_names ffffffff81823558 d scheduler_running ffffffff818235b8 d sched_clock_running ffffffff818235bc D sched_clock_stable ffffffff81824f00 d sched_switch_trace [root@doppio linux-2.6-tip]# perf top -s 0xffffffff817827d9 addr 0xffffffff817827d9 is in sched_nr_latency global var [root@doppio linux-2.6-tip]# perf top -s ffffffff81782ce0 addr 0xffffffff81782ce0 is in sched_domains_mutex global var [root@doppio linux-2.6-tip]# [root@doppio linux-2.6-tip]# perf top -s ffffffff81782ce0 --vmlinux OFF The file OFF cannot be used, trying to use /proc/kallsyms...addr 0xffffffff81782ce0 is in sched_domains_mutex global var [root@doppio linux-2.6-tip]# perf top -s ffffffff818235bc --vmlinux OFF The file OFF cannot be used, trying to use /proc/kallsyms...addr 0xffffffff818235bc is in sched_clock_stable global var [root@doppio linux-2.6-tip]# So it works with both /proc/kallsyms and with ELF symtabs, either the one on the vmlinux explicitely passed via --vmlinux or in one in the vmlinux_path that matches the buildid for the running kernel or the one found in the buildid header section in a perf.data file. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260550239-5372-4-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/event.h | 3 ++- tools/perf/util/symbol.c | 66 +++++++++++++++++++++++++++++++++++++----------- 2 files changed, 53 insertions(+), 16 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index d9a65d91e92..56640946b5a 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -103,9 +103,10 @@ void event__print_totals(void); enum map_type { MAP__FUNCTION = 0, + MAP__VARIABLE, }; -#define MAP__NR_TYPES (MAP__FUNCTION + 1) +#define MAP__NR_TYPES (MAP__VARIABLE + 1) struct map { union { diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index db8dc97b548..e63ddb469de 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -61,6 +61,8 @@ static bool symbol_type__is_a(char symbol_type, enum map_type map_type) switch (map_type) { case MAP__FUNCTION: return symbol_type == 'T' || symbol_type == 'W'; + case MAP__VARIABLE: + return symbol_type == 'D' || symbol_type == 'd'; default: return false; } @@ -551,6 +553,13 @@ static inline int elf_sym__is_function(const GElf_Sym *sym) sym->st_shndx != SHN_UNDEF; } +static inline bool elf_sym__is_object(const GElf_Sym *sym) +{ + return elf_sym__type(sym) == STT_OBJECT && + sym->st_name != 0 && + sym->st_shndx != SHN_UNDEF; +} + static inline int elf_sym__is_label(const GElf_Sym *sym) { return elf_sym__type(sym) == STT_NOTYPE && @@ -571,6 +580,12 @@ static inline int elf_sec__is_text(const GElf_Shdr *shdr, return strstr(elf_sec__name(shdr, secstrs), "text") != NULL; } +static inline bool elf_sec__is_data(const GElf_Shdr *shdr, + const Elf_Data *secstrs) +{ + return strstr(elf_sec__name(shdr, secstrs), "data") != NULL; +} + static inline const char *elf_sym__name(const GElf_Sym *sym, const Elf_Data *symstrs) { @@ -756,6 +771,8 @@ static bool elf_sym__is_a(GElf_Sym *self, enum map_type type) switch (type) { case MAP__FUNCTION: return elf_sym__is_function(self); + case MAP__VARIABLE: + return elf_sym__is_object(self); default: return false; } @@ -766,6 +783,8 @@ static bool elf_sec__is_a(GElf_Shdr *self, Elf_Data *secstrs, enum map_type type switch (type) { case MAP__FUNCTION: return elf_sec__is_text(self, secstrs); + case MAP__VARIABLE: + return elf_sec__is_data(self, secstrs); default: return false; } @@ -1536,42 +1555,59 @@ size_t dsos__fprintf_buildid(FILE *fp) __dsos__fprintf_buildid(&dsos__user, fp)); } -static int map_groups__create_kernel_map(struct map_groups *self, const char *vmlinux) +static struct dso *dsos__create_kernel( const char *vmlinux) { - struct map *kmap; struct dso *kernel = dso__new(vmlinux ?: "[kernel.kallsyms]"); if (kernel == NULL) - return -1; - - kmap = map__new2(0, kernel, MAP__FUNCTION); - if (kmap == NULL) - goto out_delete_kernel_dso; + return NULL; - kmap->map_ip = kmap->unmap_ip = identity__map_ip; kernel->short_name = "[kernel]"; kernel->kernel = 1; vdso = dso__new("[vdso]"); if (vdso == NULL) - goto out_delete_kernel_map; + goto out_delete_kernel_dso; dso__set_loaded(vdso, MAP__FUNCTION); if (sysfs__read_build_id("/sys/kernel/notes", kernel->build_id, sizeof(kernel->build_id)) == 0) kernel->has_build_id = true; - map_groups__insert(self, kmap); dsos__add(&dsos__kernel, kernel); dsos__add(&dsos__user, vdso); - return 0; + return kernel; -out_delete_kernel_map: - map__delete(kmap); out_delete_kernel_dso: dso__delete(kernel); - return -1; + return NULL; +} + +static int map_groups__create_kernel_maps(struct map_groups *self, const char *vmlinux) +{ + struct map *functions, *variables; + struct dso *kernel = dsos__create_kernel(vmlinux); + + if (kernel == NULL) + return -1; + + functions = map__new2(0, kernel, MAP__FUNCTION); + if (functions == NULL) + return -1; + + variables = map__new2(0, kernel, MAP__VARIABLE); + if (variables == NULL) { + map__delete(functions); + return -1; + } + + functions->map_ip = functions->unmap_ip = + variables->map_ip = variables->unmap_ip = identity__map_ip; + map_groups__insert(self, functions); + map_groups__insert(self, variables); + + return 0; } static void vmlinux_path__exit(void) @@ -1640,7 +1676,7 @@ int symbol__init(struct symbol_conf *conf) if (pconf->try_vmlinux_path && vmlinux_path__init() < 0) return -1; - if (map_groups__create_kernel_map(kmaps, pconf->vmlinux_name) < 0) { + if (map_groups__create_kernel_maps(kmaps, pconf->vmlinux_name) < 0) { vmlinux_path__exit(); return -1; } -- cgit From 22ccec57f8732de22fd87bce43e8edcb71453c72 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 15:23:28 -0200 Subject: perf symbols: Add missing "Variables" entry to map_type__name MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260552208-6824-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/thread.c | 1 + 1 file changed, 1 insertion(+) (limited to 'tools') diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index a1285129c83..b68a00ea412 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -54,6 +54,7 @@ int thread__comm_len(struct thread *self) static const char *map_type__name[MAP__NR_TYPES] = { [MAP__FUNCTION] = "Functions", + [MAP__VARIABLE] = "Variables", }; static size_t __map_groups__fprintf_maps(struct map_groups *self, -- cgit From 79406cd789f745ac6aa9d597895f904a98a14007 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 18:50:22 -0200 Subject: perf symbols: Allow lookups by symbol name too MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Configurable via symbol_conf.sort_by_name, so that the cost of an extra rb_node on all 'struct symbol' instances is not paid by tools that only want to decode addresses. How to use it: symbol_conf.sort_by_name = true; symbol_init(&symbol_conf); struct map *map = map_groups__find_by_name(kmaps, MAP__VARIABLE, "[kernel.kallsyms]"); if (map == NULL) { pr_err("couldn't find map!\n"); kernel_maps__fprintf(stdout); } else { struct symbol *sym = map__find_symbol_by_name(map, sym_filter, NULL); if (sym == NULL) pr_err("couldn't find symbol %s!\n", sym_filter); else pr_info("symbol %s: %#Lx-%#Lx \n", sym_filter, sym->start, sym->end); } Looking over the vmlinux/kallsyms is common enough that I'll add a variable to the upcoming struct perf_session to avoid the need to use map_groups__find_by_name to get the main vmlinux/kallsyms map. The above example looks on the 'variable' symtab, but it is just like that for the functions one. Also the sort operation is done when we first use map__find_symbol_by_name, in a lazy way. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Masami Hiramatsu Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260564622-12392-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/event.h | 2 + tools/perf/util/map.c | 85 +++++++++++++++++++++++--------------- tools/perf/util/symbol.c | 104 +++++++++++++++++++++++++++++++++++++++++++---- tools/perf/util/symbol.h | 10 ++++- tools/perf/util/thread.h | 3 ++ 5 files changed, 163 insertions(+), 41 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 56640946b5a..51a96c2effd 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -151,6 +151,8 @@ int map__overlap(struct map *l, struct map *r); size_t map__fprintf(struct map *self, FILE *fp); struct symbol *map__find_symbol(struct map *self, u64 addr, symbol_filter_t filter); +struct symbol *map__find_symbol_by_name(struct map *self, const char *name, + symbol_filter_t filter); void map__fixup_start(struct map *self); void map__fixup_end(struct map *self); diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 69f94fe9db2..175f1f6b691 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -104,45 +104,66 @@ void map__fixup_end(struct map *self) #define DSO__DELETED "(deleted)" -struct symbol *map__find_symbol(struct map *self, u64 addr, - symbol_filter_t filter) +static int map__load(struct map *self, symbol_filter_t filter) { - if (!dso__loaded(self->dso, self->type)) { - int nr = dso__load(self->dso, self, filter); - - if (nr < 0) { - if (self->dso->has_build_id) { - char sbuild_id[BUILD_ID_SIZE * 2 + 1]; - - build_id__sprintf(self->dso->build_id, - sizeof(self->dso->build_id), - sbuild_id); - pr_warning("%s with build id %s not found", - self->dso->long_name, sbuild_id); - } else - pr_warning("Failed to open %s", - self->dso->long_name); - pr_warning(", continuing without symbols\n"); - return NULL; - } else if (nr == 0) { - const char *name = self->dso->long_name; - const size_t len = strlen(name); - const size_t real_len = len - sizeof(DSO__DELETED); - - if (len > sizeof(DSO__DELETED) && - strcmp(name + real_len + 1, DSO__DELETED) == 0) { - pr_warning("%.*s was updated, restart the long running apps that use it!\n", - (int)real_len, name); - } else { - pr_warning("no symbols found in %s, maybe install a debug package?\n", name); - } - return NULL; + const char *name = self->dso->long_name; + int nr = dso__load(self->dso, self, filter); + + if (nr < 0) { + if (self->dso->has_build_id) { + char sbuild_id[BUILD_ID_SIZE * 2 + 1]; + + build_id__sprintf(self->dso->build_id, + sizeof(self->dso->build_id), + sbuild_id); + pr_warning("%s with build id %s not found", + name, sbuild_id); + } else + pr_warning("Failed to open %s", name); + + pr_warning(", continuing without symbols\n"); + return -1; + } else if (nr == 0) { + const size_t len = strlen(name); + const size_t real_len = len - sizeof(DSO__DELETED); + + if (len > sizeof(DSO__DELETED) && + strcmp(name + real_len + 1, DSO__DELETED) == 0) { + pr_warning("%.*s was updated, restart the long " + "running apps that use it!\n", + (int)real_len, name); + } else { + pr_warning("no symbols found in %s, maybe install " + "a debug package?\n", name); } + + return -1; } + return 0; +} + +struct symbol *map__find_symbol(struct map *self, u64 addr, + symbol_filter_t filter) +{ + if (!dso__loaded(self->dso, self->type) && map__load(self, filter) < 0) + return NULL; + return self->dso->find_symbol(self->dso, self->type, addr); } +struct symbol *map__find_symbol_by_name(struct map *self, const char *name, + symbol_filter_t filter) +{ + if (!dso__loaded(self->dso, self->type) && map__load(self, filter) < 0) + return NULL; + + if (!dso__sorted_by_name(self->dso, self->type)) + dso__sort_by_name(self->dso, self->type); + + return dso__find_symbol_by_name(self->dso, self->type, name); +} + struct map *map__clone(struct map *self) { struct map *map = malloc(sizeof(*self)); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index e63ddb469de..8134c49deae 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -29,7 +29,6 @@ enum dso_origin { }; static void dsos__add(struct list_head *head, struct dso *dso); -static struct map *map_groups__find_by_name(struct map_groups *self, char *name); static struct map *map__new2(u64 start, struct dso *dso, enum map_type type); struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr); static int dso__load_kernel_sym(struct dso *self, struct map *map, @@ -51,11 +50,21 @@ bool dso__loaded(const struct dso *self, enum map_type type) return self->loaded & (1 << type); } +bool dso__sorted_by_name(const struct dso *self, enum map_type type) +{ + return self->sorted_by_name & (1 << type); +} + static void dso__set_loaded(struct dso *self, enum map_type type) { self->loaded |= (1 << type); } +static void dso__set_sorted_by_name(struct dso *self, enum map_type type) +{ + self->sorted_by_name |= (1 << type); +} + static bool symbol_type__is_a(char symbol_type, enum map_type map_type) { switch (map_type) { @@ -176,11 +185,12 @@ struct dso *dso__new(const char *name) dso__set_long_name(self, self->name); self->short_name = self->name; for (i = 0; i < MAP__NR_TYPES; ++i) - self->symbols[i] = RB_ROOT; + self->symbols[i] = self->symbol_names[i] = RB_ROOT; self->find_symbol = dso__find_symbol; self->slen_calculated = 0; self->origin = DSO__ORIG_NOT_FOUND; self->loaded = 0; + self->sorted_by_name = 0; self->has_build_id = 0; } @@ -258,11 +268,85 @@ static struct symbol *symbols__find(struct rb_root *self, u64 ip) return NULL; } -struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr) +struct symbol_name_rb_node { + struct rb_node rb_node; + struct symbol sym; +}; + +static void symbols__insert_by_name(struct rb_root *self, struct symbol *sym) +{ + struct rb_node **p = &self->rb_node; + struct rb_node *parent = NULL; + struct symbol_name_rb_node *symn = ((void *)sym) - sizeof(*parent), *s; + + while (*p != NULL) { + parent = *p; + s = rb_entry(parent, struct symbol_name_rb_node, rb_node); + if (strcmp(sym->name, s->sym.name) < 0) + p = &(*p)->rb_left; + else + p = &(*p)->rb_right; + } + rb_link_node(&symn->rb_node, parent, p); + rb_insert_color(&symn->rb_node, self); +} + +static void symbols__sort_by_name(struct rb_root *self, struct rb_root *source) +{ + struct rb_node *nd; + + for (nd = rb_first(source); nd; nd = rb_next(nd)) { + struct symbol *pos = rb_entry(nd, struct symbol, rb_node); + symbols__insert_by_name(self, pos); + } +} + +static struct symbol *symbols__find_by_name(struct rb_root *self, const char *name) +{ + struct rb_node *n; + + if (self == NULL) + return NULL; + + n = self->rb_node; + + while (n) { + struct symbol_name_rb_node *s; + int cmp; + + s = rb_entry(n, struct symbol_name_rb_node, rb_node); + cmp = strcmp(name, s->sym.name); + + if (cmp < 0) + n = n->rb_left; + else if (cmp > 0) + n = n->rb_right; + else + return &s->sym; + } + + return NULL; +} + +struct symbol *dso__find_symbol(struct dso *self, + enum map_type type, u64 addr) { return symbols__find(&self->symbols[type], addr); } +struct symbol *dso__find_symbol_by_name(struct dso *self, enum map_type type, + const char *name) +{ + return symbols__find_by_name(&self->symbol_names[type], name); +} + +void dso__sort_by_name(struct dso *self, enum map_type type) +{ + dso__set_sorted_by_name(self, type); + return symbols__sort_by_name(&self->symbol_names[type], + &self->symbols[type]); +} + int build_id__sprintf(u8 *self, int len, char *bf) { char *bid = bf; @@ -397,7 +481,7 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, *module++ = '\0'; if (strcmp(self->name, module)) { - curr_map = map_groups__find_by_name(mg, module); + curr_map = map_groups__find_by_name(mg, map->type, module); if (curr_map == NULL) { pr_debug("/proc/{kallsyms,modules} " "inconsistency!\n"); @@ -895,7 +979,7 @@ static int dso__load_sym(struct dso *self, struct map *map, snprintf(dso_name, sizeof(dso_name), "%s%s", self->short_name, section_name); - curr_map = map_groups__find_by_name(mg, dso_name); + curr_map = map_groups__find_by_name(mg, map->type, dso_name); if (curr_map == NULL) { u64 start = sym.st_value; @@ -1226,11 +1310,12 @@ out: return ret; } -static struct map *map_groups__find_by_name(struct map_groups *self, char *name) +struct map *map_groups__find_by_name(struct map_groups *self, + enum map_type type, const char *name) { struct rb_node *nd; - for (nd = rb_first(&self->maps[MAP__FUNCTION]); nd; nd = rb_next(nd)) { + for (nd = rb_first(&self->maps[type]); nd; nd = rb_next(nd)) { struct map *map = rb_entry(nd, struct map, rb_node); if (map->dso && strcmp(map->dso->name, name) == 0) @@ -1274,7 +1359,7 @@ static int dsos__set_modules_path_dir(char *dirname) (int)(dot - dent->d_name), dent->d_name); strxfrchar(dso_name, '-', '_'); - map = map_groups__find_by_name(kmaps, dso_name); + map = map_groups__find_by_name(kmaps, MAP__FUNCTION, dso_name); if (map == NULL) continue; @@ -1671,6 +1756,9 @@ int symbol__init(struct symbol_conf *conf) elf_version(EV_CURRENT); symbol__priv_size = pconf->priv_size; + if (pconf->sort_by_name) + symbol__priv_size += (sizeof(struct symbol_name_rb_node) - + sizeof(struct symbol)); map_groups__init(kmaps); if (pconf->try_vmlinux_path && vmlinux_path__init() < 0) diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 6e1da1ea631..51c401307bf 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -52,7 +52,8 @@ struct symbol { struct symbol_conf { unsigned short priv_size; bool try_vmlinux_path, - use_modules; + use_modules, + sort_by_name; const char *vmlinux_name; }; @@ -74,6 +75,7 @@ struct addr_location { struct dso { struct list_head node; struct rb_root symbols[MAP__NR_TYPES]; + struct rb_root symbol_names[MAP__NR_TYPES]; struct symbol *(*find_symbol)(struct dso *self, enum map_type type, u64 addr); u8 adjust_symbols:1; @@ -81,6 +83,7 @@ struct dso { u8 has_build_id:1; u8 kernel:1; unsigned char origin; + u8 sorted_by_name; u8 loaded; u8 build_id[BUILD_ID_SIZE]; u16 long_name_len; @@ -93,6 +96,9 @@ struct dso *dso__new(const char *name); void dso__delete(struct dso *self); bool dso__loaded(const struct dso *self, enum map_type type); +bool dso__sorted_by_name(const struct dso *self, enum map_type type); + +void dso__sort_by_name(struct dso *self, enum map_type type); struct dso *dsos__findnew(const char *name); int dso__load(struct dso *self, struct map *map, symbol_filter_t filter); @@ -103,6 +109,8 @@ size_t dso__fprintf_buildid(struct dso *self, FILE *fp); size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); char dso__symtab_origin(const struct dso *self); void dso__set_build_id(struct dso *self, void *build_id); +struct symbol *dso__find_symbol_by_name(struct dso *self, enum map_type type, + const char *name); int filename__read_build_id(const char *filename, void *bf, size_t size); int sysfs__read_build_id(const char *filename, void *bf, size_t size); diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index a6333f3716a..1751802a09b 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -64,4 +64,7 @@ map_groups__find_function(struct map_groups *self, u64 addr, { return map_groups__find_symbol(self, MAP__FUNCTION, addr, filter); } + +struct map *map_groups__find_by_name(struct map_groups *self, + enum map_type type, const char *name); #endif /* __PERF_THREAD_H */ -- cgit From ea08d8cbd162fe3756e3e2298efbe0b8b12f92d1 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 18:56:39 -0200 Subject: perf symbols: Ditch dso->find_symbol MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It is always wired to dso__find_symbol. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260564999-13371-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/util/map.c | 2 +- tools/perf/util/symbol.c | 2 -- tools/perf/util/symbol.h | 3 +-- 3 files changed, 2 insertions(+), 5 deletions(-) (limited to 'tools') diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 175f1f6b691..76bdca640a9 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -149,7 +149,7 @@ struct symbol *map__find_symbol(struct map *self, u64 addr, if (!dso__loaded(self->dso, self->type) && map__load(self, filter) < 0) return NULL; - return self->dso->find_symbol(self->dso, self->type, addr); + return dso__find_symbol(self->dso, self->type, addr); } struct symbol *map__find_symbol_by_name(struct map *self, const char *name, diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index 8134c49deae..d3d9fed74f1 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -30,7 +30,6 @@ enum dso_origin { static void dsos__add(struct list_head *head, struct dso *dso); static struct map *map__new2(u64 start, struct dso *dso, enum map_type type); -struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr); static int dso__load_kernel_sym(struct dso *self, struct map *map, struct map_groups *mg, symbol_filter_t filter); unsigned int symbol__priv_size; @@ -186,7 +185,6 @@ struct dso *dso__new(const char *name) self->short_name = self->name; for (i = 0; i < MAP__NR_TYPES; ++i) self->symbols[i] = self->symbol_names[i] = RB_ROOT; - self->find_symbol = dso__find_symbol; self->slen_calculated = 0; self->origin = DSO__ORIG_NOT_FOUND; self->loaded = 0; diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index 51c401307bf..cf99f88adf3 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -76,8 +76,6 @@ struct dso { struct list_head node; struct rb_root symbols[MAP__NR_TYPES]; struct rb_root symbol_names[MAP__NR_TYPES]; - struct symbol *(*find_symbol)(struct dso *self, - enum map_type type, u64 addr); u8 adjust_symbols:1; u8 slen_calculated:1; u8 has_build_id:1; @@ -109,6 +107,7 @@ size_t dso__fprintf_buildid(struct dso *self, FILE *fp); size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); char dso__symtab_origin(const struct dso *self); void dso__set_build_id(struct dso *self, void *build_id); +struct symbol *dso__find_symbol(struct dso *self, enum map_type type, u64 addr); struct symbol *dso__find_symbol_by_name(struct dso *self, enum map_type type, const char *name); -- cgit From 94c744b6c0c6c5802a85ebfebbec429ac5851f2b Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Fri, 11 Dec 2009 21:24:02 -0200 Subject: perf tools: Introduce perf_session class MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit That does all the initialization boilerplate, opening the file, reading the header, checking if it is valid, etc. And that will as well have the threads list, kmap (now) global variable, etc, so that we can handle two (or more) perf.data files describing sessions to compare. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260573842-19720-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/Makefile | 3 ++ tools/perf/builtin-annotate.c | 15 +++++--- tools/perf/builtin-buildid-list.c | 55 ++++----------------------- tools/perf/builtin-kmem.c | 13 +++++-- tools/perf/builtin-record.c | 25 ++++++------ tools/perf/builtin-report.c | 20 ++++++---- tools/perf/builtin-sched.c | 13 +++++-- tools/perf/builtin-timechart.c | 15 +++++--- tools/perf/builtin-trace.c | 20 ++++++---- tools/perf/util/data_map.c | 71 ++++++---------------------------- tools/perf/util/data_map.h | 9 ++--- tools/perf/util/header.c | 28 +++----------- tools/perf/util/header.h | 4 +- tools/perf/util/session.c | 80 +++++++++++++++++++++++++++++++++++++++ tools/perf/util/session.h | 16 ++++++++ 15 files changed, 206 insertions(+), 181 deletions(-) create mode 100644 tools/perf/util/session.c create mode 100644 tools/perf/util/session.h (limited to 'tools') diff --git a/tools/perf/Makefile b/tools/perf/Makefile index e2ee3b589af..406999668ca 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -356,7 +356,9 @@ LIB_H += util/parse-options.h LIB_H += util/parse-events.h LIB_H += util/quote.h LIB_H += util/util.h +LIB_H += util/header.h LIB_H += util/help.h +LIB_H += util/session.h LIB_H += util/strbuf.h LIB_H += util/string.h LIB_H += util/strlist.h @@ -405,6 +407,7 @@ LIB_OBJS += util/callchain.o LIB_OBJS += util/values.o LIB_OBJS += util/debug.o LIB_OBJS += util/map.o +LIB_OBJS += util/session.o LIB_OBJS += util/thread.o LIB_OBJS += util/trace-event-parse.o LIB_OBJS += util/trace-event-read.o diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 0bf2e8f9af5..21a78d30d53 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -25,6 +25,7 @@ #include "util/thread.h" #include "util/sort.h" #include "util/hist.h" +#include "util/session.h" #include "util/data_map.h" static char const *input_name = "perf.data"; @@ -462,21 +463,23 @@ static struct perf_file_handler file_handler = { static int __cmd_annotate(void) { - struct perf_header *header; + struct perf_session *session = perf_session__new(input_name, O_RDONLY, force); struct thread *idle; int ret; + if (session == NULL) + return -ENOMEM; + idle = register_idle_thread(); register_perf_file_handler(&file_handler); - ret = mmap_dispatch_perf_file(&header, input_name, 0, 0, - &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); if (ret) - return ret; + goto out_delete; if (dump_trace) { event__print_totals(); - return 0; + goto out_delete; } if (verbose > 3) @@ -489,6 +492,8 @@ static int __cmd_annotate(void) output__resort(event__total[0]); find_annotations(); +out_delete: + perf_session__delete(session); return ret; } diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index dcb6143a000..bfd16a1594e 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -11,8 +11,8 @@ #include "util/cache.h" #include "util/data_map.h" #include "util/debug.h" -#include "util/header.h" #include "util/parse-options.h" +#include "util/session.h" #include "util/symbol.h" static char const *input_name = "perf.data"; @@ -55,56 +55,17 @@ static int perf_file_section__process_buildids(struct perf_file_section *self, static int __cmd_buildid_list(void) { int err = -1; - struct perf_header *header; - struct perf_file_header f_header; - struct stat input_stat; - int input = open(input_name, O_RDONLY); + struct perf_session *session = perf_session__new(input_name, O_RDONLY, force); - if (input < 0) { - pr_err("failed to open file: %s", input_name); - if (!strcmp(input_name, "perf.data")) - pr_err(" (try 'perf record' first)"); - pr_err("\n"); - goto out; - } - - err = fstat(input, &input_stat); - if (err < 0) { - perror("failed to stat file"); - goto out_close; - } - - if (!force && input_stat.st_uid && (input_stat.st_uid != geteuid())) { - pr_err("file %s not owned by current user or root\n", - input_name); - goto out_close; - } - - if (!input_stat.st_size) { - pr_info("zero-sized file, nothing to do!\n"); - goto out_close; - } - - err = -1; - header = perf_header__new(); - if (header == NULL) - goto out_close; - - if (perf_file_header__read(&f_header, header, input) < 0) { - pr_warning("incompatible file format"); - goto out_close; - } + if (session == NULL) + return -1; - err = perf_header__process_sections(header, input, + err = perf_header__process_sections(&session->header, session->fd, perf_file_section__process_buildids); + if (err >= 0) + dsos__fprintf_buildid(stdout); - if (err < 0) - goto out_close; - - dsos__fprintf_buildid(stdout); -out_close: - close(input); -out: + perf_session__delete(session); return err; } diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index fe73435192b..2071d248591 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -6,6 +6,7 @@ #include "util/symbol.h" #include "util/thread.h" #include "util/header.h" +#include "util/session.h" #include "util/parse-options.h" #include "util/trace-event.h" @@ -20,7 +21,6 @@ typedef int (*sort_fn_t)(struct alloc_stat *, struct alloc_stat *); static char const *input_name = "perf.data"; -static struct perf_header *header; static u64 sample_type; static int alloc_flag; @@ -367,11 +367,18 @@ static struct perf_file_handler file_handler = { static int read_events(void) { + int err; + struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); + + if (session == NULL) + return -ENOMEM; + register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, 0, 0, - &event__cwdlen, &event__cwd); + err = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + perf_session__delete(session); + return err; } static double fragmentation(unsigned long n_req, unsigned long n_alloc) diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 0e519c667e3..4decbd14eae 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -17,6 +17,7 @@ #include "util/header.h" #include "util/event.h" #include "util/debug.h" +#include "util/session.h" #include "util/symbol.h" #include @@ -62,7 +63,7 @@ static int nr_cpu = 0; static int file_new = 1; -struct perf_header *header = NULL; +static struct perf_session *session; struct mmap_data { int counter; @@ -216,12 +217,12 @@ static struct perf_header_attr *get_header_attr(struct perf_event_attr *a, int n { struct perf_header_attr *h_attr; - if (nr < header->attrs) { - h_attr = header->attr[nr]; + if (nr < session->header.attrs) { + h_attr = session->header.attr[nr]; } else { h_attr = perf_header_attr__new(a); if (h_attr != NULL) - if (perf_header__add_attr(header, h_attr) < 0) { + if (perf_header__add_attr(&session->header, h_attr) < 0) { perf_header_attr__delete(h_attr); h_attr = NULL; } @@ -395,9 +396,9 @@ static void open_counters(int cpu, pid_t pid) static void atexit_header(void) { - header->data_size += bytes_written; + session->header.data_size += bytes_written; - perf_header__write(header, output, true); + perf_header__write(&session->header, output, true); } static int __cmd_record(int argc, const char **argv) @@ -440,24 +441,24 @@ static int __cmd_record(int argc, const char **argv) exit(-1); } - header = perf_header__new(); - if (header == NULL) { + session = perf_session__new(output_name, O_WRONLY, force); + if (session == NULL) { pr_err("Not enough memory for reading perf file header\n"); return -1; } if (!file_new) { - err = perf_header__read(header, output); + err = perf_header__read(&session->header, output); if (err < 0) return err; } if (raw_samples) { - perf_header__set_feat(header, HEADER_TRACE_INFO); + perf_header__set_feat(&session->header, HEADER_TRACE_INFO); } else { for (i = 0; i < nr_counters; i++) { if (attrs[i].sample_type & PERF_SAMPLE_RAW) { - perf_header__set_feat(header, HEADER_TRACE_INFO); + perf_header__set_feat(&session->header, HEADER_TRACE_INFO); break; } } @@ -481,7 +482,7 @@ static int __cmd_record(int argc, const char **argv) } if (file_new) { - err = perf_header__write(header, output, false); + err = perf_header__write(&session->header, output, false); if (err < 0) return err; } diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 2b9eb3a553e..e2ec49a9b73 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -22,6 +22,7 @@ #include "perf.h" #include "util/debug.h" #include "util/header.h" +#include "util/session.h" #include "util/parse-options.h" #include "util/parse-events.h" @@ -52,7 +53,7 @@ static int exclude_other = 1; static char callchain_default_opt[] = "fractal,0.5"; -static struct perf_header *header; +static struct perf_session *session; static u64 sample_type; @@ -701,7 +702,7 @@ static int process_read_event(event_t *event) { struct perf_event_attr *attr; - attr = perf_header__find_attr(event->read.id, header); + attr = perf_header__find_attr(event->read.id, &session->header); if (show_threads) { const char *name = attr ? __event_name(attr->type, attr->config) @@ -766,6 +767,10 @@ static int __cmd_report(void) struct thread *idle; int ret; + session = perf_session__new(input_name, O_RDONLY, force); + if (session == NULL) + return -ENOMEM; + idle = register_idle_thread(); thread__comm_adjust(idle); @@ -774,14 +779,14 @@ static int __cmd_report(void) register_perf_file_handler(&file_handler); - ret = mmap_dispatch_perf_file(&header, input_name, force, - full_paths, &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, full_paths, + &event__cwdlen, &event__cwd); if (ret) - return ret; + goto out_delete; if (dump_trace) { event__print_totals(); - return 0; + goto out_delete; } if (verbose > 3) @@ -796,7 +801,8 @@ static int __cmd_report(void) if (show_threads) perf_read_values_destroy(&show_threads_values); - +out_delete: + perf_session__delete(session); return ret; } diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 7cca7c15b40..65021fe1361 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -6,6 +6,7 @@ #include "util/symbol.h" #include "util/thread.h" #include "util/header.h" +#include "util/session.h" #include "util/parse-options.h" #include "util/trace-event.h" @@ -21,7 +22,6 @@ static char const *input_name = "perf.data"; -static struct perf_header *header; static u64 sample_type; static char default_sort_order[] = "avg, max, switch, runtime"; @@ -1663,11 +1663,18 @@ static struct perf_file_handler file_handler = { static int read_events(void) { + int err; + struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); + + if (session == NULL) + return -ENOMEM; + register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, 0, 0, - &event__cwdlen, &event__cwd); + err = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + perf_session__delete(session); + return err; } static void print_bad_events(void) diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index f472df9561e..759dd2b35fd 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1059,15 +1059,17 @@ static struct perf_file_handler file_handler = { static int __cmd_timechart(void) { - struct perf_header *header; + struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); int ret; + if (session == NULL) + return -ENOMEM; + register_perf_file_handler(&file_handler); - ret = mmap_dispatch_perf_file(&header, input_name, 0, 0, - &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); if (ret) - return EXIT_FAILURE; + goto out_delete; process_samples(); @@ -1079,8 +1081,9 @@ static int __cmd_timechart(void) pr_info("Written %2.1f seconds of trace to %s.\n", (last_time - first_time) / 1000000000.0, output_name); - - return EXIT_SUCCESS; +out_delete: + perf_session__delete(session); + return ret; } static const char * const timechart_usage[] = { diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index c2fcc34486f..0756664666f 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -7,6 +7,7 @@ #include "util/header.h" #include "util/exec_cmd.h" #include "util/trace-event.h" +#include "util/session.h" static char const *script_name; static char const *generate_script_lang; @@ -61,7 +62,7 @@ static int cleanup_scripting(void) static char const *input_name = "perf.data"; -static struct perf_header *header; +static struct perf_session *session; static u64 sample_type; static int process_sample_event(event_t *event) @@ -126,11 +127,18 @@ static struct perf_file_handler file_handler = { static int __cmd_trace(void) { + int err; + + session = perf_session__new(input_name, O_RDONLY, 0); + if (session == NULL) + return -ENOMEM; + register_idle_thread(); register_perf_file_handler(&file_handler); - return mmap_dispatch_perf_file(&header, input_name, - 0, 0, &event__cwdlen, &event__cwd); + err = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + perf_session__delete(session); + return err; } struct script_spec { @@ -348,11 +356,7 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) return -1; } - header = perf_header__new(); - if (header == NULL) - return -1; - - perf_header__read(header, input); + perf_header__read(&session->header, input); err = scripting_ops->generate_script("perf-trace"); goto out; } diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index 59b65d0bd7c..6d46dda53a2 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -129,23 +129,16 @@ out: return err; } -int mmap_dispatch_perf_file(struct perf_header **pheader, - const char *input_name, - int force, - int full_paths, - int *cwdlen, - char **cwd) +int perf_session__process_events(struct perf_session *self, + int full_paths, int *cwdlen, char **cwd) { int err; - struct perf_header *header; unsigned long head, shift; unsigned long offset = 0; - struct stat input_stat; size_t page_size; u64 sample_type; event_t *event; uint32_t size; - int input; char *buf; if (curr_handler == NULL) { @@ -155,56 +148,19 @@ int mmap_dispatch_perf_file(struct perf_header **pheader, page_size = getpagesize(); - input = open(input_name, O_RDONLY); - if (input < 0) { - pr_err("Failed to open file: %s", input_name); - if (!strcmp(input_name, "perf.data")) - pr_err(" (try 'perf record' first)"); - pr_err("\n"); - return -errno; - } - - if (fstat(input, &input_stat) < 0) { - pr_err("failed to stat file"); - err = -errno; - goto out_close; - } - - err = -EACCES; - if (!force && input_stat.st_uid && (input_stat.st_uid != geteuid())) { - pr_err("file: %s not owned by current user or root\n", - input_name); - goto out_close; - } - - if (input_stat.st_size == 0) { - pr_info("zero-sized file, nothing to do!\n"); - goto done; - } - - err = -ENOMEM; - header = perf_header__new(); - if (header == NULL) - goto out_close; - - err = perf_header__read(header, input); - if (err < 0) - goto out_delete; - *pheader = header; - head = header->data_offset; - - sample_type = perf_header__sample_type(header); + head = self->header.data_offset; + sample_type = perf_header__sample_type(&self->header); err = -EINVAL; if (curr_handler->sample_type_check && curr_handler->sample_type_check(sample_type) < 0) - goto out_delete; + goto out_err; if (!full_paths) { if (getcwd(__cwd, sizeof(__cwd)) == NULL) { pr_err("failed to get the current directory\n"); err = -errno; - goto out_delete; + goto out_err; } *cwd = __cwd; *cwdlen = strlen(*cwd); @@ -219,11 +175,11 @@ int mmap_dispatch_perf_file(struct perf_header **pheader, remap: buf = mmap(NULL, page_size * mmap_window, PROT_READ, - MAP_SHARED, input, offset); + MAP_SHARED, self->fd, offset); if (buf == MAP_FAILED) { pr_err("failed to mmap file\n"); err = -errno; - goto out_delete; + goto out_err; } more: @@ -273,19 +229,14 @@ more: head += size; - if (offset + head >= header->data_offset + header->data_size) + if (offset + head >= self->header.data_offset + self->header.data_size) goto done; - if (offset + head < (unsigned long)input_stat.st_size) + if (offset + head < self->size) goto more; done: err = 0; -out_close: - close(input); - +out_err: return err; -out_delete: - perf_header__delete(header); - goto out_close; } diff --git a/tools/perf/util/data_map.h b/tools/perf/util/data_map.h index 258a87bcc4f..98c5b823388 100644 --- a/tools/perf/util/data_map.h +++ b/tools/perf/util/data_map.h @@ -3,6 +3,7 @@ #include "event.h" #include "header.h" +#include "session.h" typedef int (*event_type_handler_t)(event_t *); @@ -21,12 +22,8 @@ struct perf_file_handler { }; void register_perf_file_handler(struct perf_file_handler *handler); -int mmap_dispatch_perf_file(struct perf_header **pheader, - const char *input_name, - int force, - int full_paths, - int *cwdlen, - char **cwd); +int perf_session__process_events(struct perf_session *self, + int full_paths, int *cwdlen, char **cwd); int perf_header__read_build_ids(int input, u64 offset, u64 file_size); #endif diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index 59a9c0b3033..f2e8d871511 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -58,35 +58,19 @@ int perf_header_attr__add_id(struct perf_header_attr *self, u64 id) return 0; } -/* - * Create new perf.data header: - */ -struct perf_header *perf_header__new(void) +int perf_header__init(struct perf_header *self) { - struct perf_header *self = zalloc(sizeof(*self)); - - if (self != NULL) { - self->size = 1; - self->attr = malloc(sizeof(void *)); - - if (self->attr == NULL) { - free(self); - self = NULL; - } - } - - return self; + self->size = 1; + self->attr = malloc(sizeof(void *)); + return self->attr == NULL ? -ENOMEM : 0; } -void perf_header__delete(struct perf_header *self) +void perf_header__exit(struct perf_header *self) { int i; - for (i = 0; i < self->attrs; ++i) - perf_header_attr__delete(self->attr[i]); - + perf_header_attr__delete(self->attr[i]); free(self->attr); - free(self); } int perf_header__add_attr(struct perf_header *self, diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h index d1dbe2b79c4..d118d05d3ab 100644 --- a/tools/perf/util/header.h +++ b/tools/perf/util/header.h @@ -55,8 +55,8 @@ struct perf_header { DECLARE_BITMAP(adds_features, HEADER_FEAT_BITS); }; -struct perf_header *perf_header__new(void); -void perf_header__delete(struct perf_header *self); +int perf_header__init(struct perf_header *self); +void perf_header__exit(struct perf_header *self); int perf_header__read(struct perf_header *self, int fd); int perf_header__write(struct perf_header *self, int fd, bool at_exit); diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c new file mode 100644 index 00000000000..707ce1cb162 --- /dev/null +++ b/tools/perf/util/session.c @@ -0,0 +1,80 @@ +#include + +#include +#include + +#include "session.h" +#include "util.h" + +static int perf_session__open(struct perf_session *self, bool force) +{ + struct stat input_stat; + + self->fd = open(self->filename, O_RDONLY); + if (self->fd < 0) { + pr_err("failed to open file: %s", self->filename); + if (!strcmp(self->filename, "perf.data")) + pr_err(" (try 'perf record' first)"); + pr_err("\n"); + return -errno; + } + + if (fstat(self->fd, &input_stat) < 0) + goto out_close; + + if (!force && input_stat.st_uid && (input_stat.st_uid != geteuid())) { + pr_err("file %s not owned by current user or root\n", + self->filename); + goto out_close; + } + + if (!input_stat.st_size) { + pr_info("zero-sized file (%s), nothing to do!\n", + self->filename); + goto out_close; + } + + if (perf_header__read(&self->header, self->fd) < 0) { + pr_err("incompatible file format"); + goto out_close; + } + + self->size = input_stat.st_size; + return 0; + +out_close: + close(self->fd); + self->fd = -1; + return -1; +} + +struct perf_session *perf_session__new(const char *filename, int mode, bool force) +{ + size_t len = strlen(filename) + 1; + struct perf_session *self = zalloc(sizeof(*self) + len); + + if (self == NULL) + goto out; + + if (perf_header__init(&self->header) < 0) + goto out_delete; + + memcpy(self->filename, filename, len); + + if (mode == O_RDONLY && perf_session__open(self, force) < 0) { + perf_session__delete(self); + self = NULL; + } +out: + return self; +out_delete: + free(self); + return NULL; +} + +void perf_session__delete(struct perf_session *self) +{ + perf_header__exit(&self->header); + close(self->fd); + free(self); +} diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h new file mode 100644 index 00000000000..f3699c8c8ed --- /dev/null +++ b/tools/perf/util/session.h @@ -0,0 +1,16 @@ +#ifndef __PERF_SESSION_H +#define __PERF_SESSION_H + +#include "header.h" + +struct perf_session { + struct perf_header header; + unsigned long size; + int fd; + char filename[0]; +}; + +struct perf_session *perf_session__new(const char *filename, int mode, bool force); +void perf_session__delete(struct perf_session *self); + +#endif /* __PERF_SESSION_H */ -- cgit From 2044279d1e07c90edd64324f09c7adf1acfc42e5 Mon Sep 17 00:00:00 2001 From: Hitoshi Mitake Date: Sun, 13 Dec 2009 17:01:59 +0900 Subject: perf bench: Add "all" pseudo subsystem and "all" pseudo suite This patch adds a new "all" pseudo subsystem and an "all" pseudo suite. These are for testing all subsystem and its all suite, or all suite of one subsystem. (This patch also contains a few trivial comment fixes for bench/* and output style fixes. I judged that there are no necessity to make them into individual patch.) Example of use: | % ./perf bench sched all # Test all suites of sched subsystem | # Running sched/messaging benchmark... | # 20 sender and receiver processes per group | # 10 groups == 400 processes run | | Total time: 0.414 [sec] | | # Running sched/pipe benchmark... | # Extecuted 1000000 pipe operations between two tasks | | Total time: 10.999 [sec] | | 10.999317 usecs/op | 90914 ops/sec | | % ./perf bench all # Test all suites of all subsystems | # Running sched/messaging benchmark... | # 20 sender and receiver processes per group | # 10 groups == 400 processes run | | Total time: 0.420 [sec] | | # Running sched/pipe benchmark... | # Extecuted 1000000 pipe operations between two tasks | | Total time: 11.741 [sec] | | 11.741346 usecs/op | 85169 ops/sec | | # Running mem/memcpy benchmark... | # Copying 1MB Bytes from 0x7ff33e920010 to 0x7ff3401ae010 ... | | 808.407437 MB/Sec Signed-off-by: Hitoshi Mitake Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Frederic Weisbecker LKML-Reference: <1260691319-4683-1-git-send-email-mitake@dcl.info.waseda.ac.jp> Signed-off-by: Ingo Molnar --- tools/perf/bench/sched-messaging.c | 2 +- tools/perf/bench/sched-pipe.c | 5 ++-- tools/perf/builtin-bench.c | 57 +++++++++++++++++++++++++++++++++++--- 3 files changed, 57 insertions(+), 7 deletions(-) (limited to 'tools') diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c index 605a2a959aa..e249eb22b43 100644 --- a/tools/perf/bench/sched-messaging.c +++ b/tools/perf/bench/sched-messaging.c @@ -1,6 +1,6 @@ /* * - * builtin-bench-messaging.c + * sched-messaging.c * * messaging: Benchmark for scheduler and IPC mechanisms * diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c index 238185f9797..2354f3830b1 100644 --- a/tools/perf/bench/sched-pipe.c +++ b/tools/perf/bench/sched-pipe.c @@ -1,6 +1,6 @@ /* * - * builtin-bench-pipe.c + * sched-pipe.c * * pipe: Benchmark for pipe() * @@ -87,7 +87,8 @@ int bench_sched_pipe(int argc, const char **argv, if (pid) { retpid = waitpid(pid, &wait_stat, 0); assert((retpid == pid) && WIFEXITED(wait_stat)); - return 0; + } else { + exit(0); } switch (bench_format) { diff --git a/tools/perf/builtin-bench.c b/tools/perf/builtin-bench.c index e043eb83092..46996774e55 100644 --- a/tools/perf/builtin-bench.c +++ b/tools/perf/builtin-bench.c @@ -31,6 +31,9 @@ struct bench_suite { const char *summary; int (*fn)(int, const char **, const char *); }; + \ +/* sentinel: easy for help */ +#define suite_all { "all", "test all suite (pseudo suite)", NULL } static struct bench_suite sched_suites[] = { { "messaging", @@ -39,6 +42,7 @@ static struct bench_suite sched_suites[] = { { "pipe", "Flood of communication over pipe() between two processes", bench_sched_pipe }, + suite_all, { NULL, NULL, NULL } @@ -48,6 +52,7 @@ static struct bench_suite mem_suites[] = { { "memcpy", "Simple memory copy in various ways", bench_mem_memcpy }, + suite_all, { NULL, NULL, NULL } @@ -66,6 +71,9 @@ static struct bench_subsys subsystems[] = { { "mem", "memory access performance", mem_suites }, + { "all", /* sentinel: easy for help */ + "test all subsystem (pseudo subsystem)", + NULL }, { NULL, NULL, NULL } @@ -75,11 +83,11 @@ static void dump_suites(int subsys_index) { int i; - printf("List of available suites for %s...\n\n", + printf("# List of available suites for %s...\n\n", subsystems[subsys_index].name); for (i = 0; subsystems[subsys_index].suites[i].name; i++) - printf("\t%s: %s\n", + printf("%14s: %s\n", subsystems[subsys_index].suites[i].name, subsystems[subsys_index].suites[i].summary); @@ -110,10 +118,10 @@ static void print_usage(void) printf("\t%s\n", bench_usage[i]); printf("\n"); - printf("List of available subsystems...\n\n"); + printf("# List of available subsystems...\n\n"); for (i = 0; subsystems[i].name; i++) - printf("\t%s: %s\n", + printf("%14s: %s\n", subsystems[i].name, subsystems[i].summary); printf("\n"); } @@ -131,6 +139,37 @@ static int bench_str2int(char *str) return BENCH_FORMAT_UNKNOWN; } +static void all_suite(struct bench_subsys *subsys) /* FROM HERE */ +{ + int i; + const char *argv[2]; + struct bench_suite *suites = subsys->suites; + + argv[1] = NULL; + /* + * TODO: + * preparing preset parameters for + * embedded, ordinary PC, HPC, etc... + * will be helpful + */ + for (i = 0; suites[i].fn; i++) { + printf("# Running %s/%s benchmark...\n", + subsys->name, + suites[i].name); + + argv[1] = suites[i].name; + suites[i].fn(1, argv, NULL); + printf("\n"); + } +} + +static void all_subsystem(void) +{ + int i; + for (i = 0; subsystems[i].suites; i++) + all_suite(&subsystems[i]); +} + int cmd_bench(int argc, const char **argv, const char *prefix __used) { int i, j, status = 0; @@ -155,6 +194,11 @@ int cmd_bench(int argc, const char **argv, const char *prefix __used) goto end; } + if (!strcmp(argv[0], "all")) { + all_subsystem(); + goto end; + } + for (i = 0; subsystems[i].name; i++) { if (strcmp(subsystems[i].name, argv[0])) continue; @@ -165,6 +209,11 @@ int cmd_bench(int argc, const char **argv, const char *prefix __used) goto end; } + if (!strcmp(argv[1], "all")) { + all_suite(&subsystems[i]); + goto end; + } + for (j = 0; subsystems[i].suites[j].name; j++) { if (strcmp(subsystems[i].suites[j].name, argv[1])) continue; -- cgit From 2cd9046cc53dd2625e2cf5854d6cbb1ba61de914 Mon Sep 17 00:00:00 2001 From: David Miller Date: Sun, 13 Dec 2009 23:56:22 -0800 Subject: perf sched: Fix build failure on sparc Here, tvec->tv_usec is "unsigned int" not "unsigned long". Since the type is different on every platform, it's probably best to just use long printf formats and cast. Signed-off-by: David S. Miller Cc: Peter Zijlstra Cc: Mike Galbraith Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Frederic Weisbecker LKML-Reference: <20091213.235622.53363059.davem@davemloft.net> Signed-off-by: Ingo Molnar --- tools/perf/bench/sched-messaging.c | 6 ++++-- tools/perf/bench/sched-pipe.c | 6 ++++-- 2 files changed, 8 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c index e249eb22b43..81cee78181f 100644 --- a/tools/perf/bench/sched-messaging.c +++ b/tools/perf/bench/sched-messaging.c @@ -320,10 +320,12 @@ int bench_sched_messaging(int argc, const char **argv, num_groups, num_groups * 2 * num_fds, thread_mode ? "threads" : "processes"); printf(" %14s: %lu.%03lu [sec]\n", "Total time", - diff.tv_sec, diff.tv_usec/1000); + diff.tv_sec, + (unsigned long) (diff.tv_usec/1000)); break; case BENCH_FORMAT_SIMPLE: - printf("%lu.%03lu\n", diff.tv_sec, diff.tv_usec/1000); + printf("%lu.%03lu\n", diff.tv_sec, + (unsigned long) (diff.tv_usec/1000)); break; default: /* reaching here is something disaster */ diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c index 2354f3830b1..4f77c7c2764 100644 --- a/tools/perf/bench/sched-pipe.c +++ b/tools/perf/bench/sched-pipe.c @@ -100,7 +100,8 @@ int bench_sched_pipe(int argc, const char **argv, result_usec += diff.tv_usec; printf(" %14s: %lu.%03lu [sec]\n\n", "Total time", - diff.tv_sec, diff.tv_usec/1000); + diff.tv_sec, + (unsigned long) (diff.tv_usec/1000)); printf(" %14lf usecs/op\n", (double)result_usec / (double)loops); @@ -111,7 +112,8 @@ int bench_sched_pipe(int argc, const char **argv, case BENCH_FORMAT_SIMPLE: printf("%lu.%03lu\n", - diff.tv_sec, diff.tv_usec / 1000); + diff.tv_sec, + (unsigned long) (diff.tv_usec / 1000)); break; default: -- cgit From d8f66248d6f25f7c935cc5307c43bf394db07272 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 13 Dec 2009 19:50:24 -0200 Subject: perf session: Pass the perf_session to the event handling operations MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit They will need it to get the right threads list, etc. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260741029-4430-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 2 +- tools/perf/builtin-kmem.c | 2 +- tools/perf/builtin-record.c | 11 +++++++---- tools/perf/builtin-report.c | 9 ++++----- tools/perf/builtin-sched.c | 6 ++++-- tools/perf/builtin-timechart.c | 13 +++++-------- tools/perf/builtin-top.c | 28 +++++++++++++++++----------- tools/perf/builtin-trace.c | 23 ++++++++++------------- tools/perf/util/data_map.c | 27 ++++++++++++++------------- tools/perf/util/data_map.h | 5 ++++- tools/perf/util/event.c | 37 +++++++++++++++++++++++-------------- tools/perf/util/event.h | 19 +++++++++++++------ 12 files changed, 103 insertions(+), 79 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 21a78d30d53..2e418b9fc1c 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -132,7 +132,7 @@ static int hist_entry__add(struct addr_location *al, u64 count) return 0; } -static int process_sample_event(event_t *event) +static int process_sample_event(event_t *event, struct perf_session *session __used) { struct addr_location al; diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 2071d248591..101b2682b4d 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -312,7 +312,7 @@ process_raw_event(event_t *raw_event __used, void *data, } } -static int process_sample_event(event_t *event) +static int process_sample_event(event_t *event, struct perf_session *session __used) { struct sample_data data; struct thread *thread; diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 4decbd14eae..b7e15a1b1ec 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -123,7 +123,8 @@ static void write_event(event_t *buf, size_t size) write_output(buf, size); } -static int process_synthesized_event(event_t *event) +static int process_synthesized_event(event_t *event, + struct perf_session *self __used) { write_event(event, event->header.size); return 0; @@ -488,9 +489,10 @@ static int __cmd_record(int argc, const char **argv) } if (!system_wide) - event__synthesize_thread(pid, process_synthesized_event); + event__synthesize_thread(pid, process_synthesized_event, + session); else - event__synthesize_threads(process_synthesized_event); + event__synthesize_threads(process_synthesized_event, session); if (target_pid == -1 && argc) { pid = fork(); @@ -510,7 +512,8 @@ static int __cmd_record(int argc, const char **argv) */ usleep(1000); event__synthesize_thread(pid, - process_synthesized_event); + process_synthesized_event, + session); } child_pid = pid; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index e2ec49a9b73..dcd8fedc298 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -53,8 +53,6 @@ static int exclude_other = 1; static char callchain_default_opt[] = "fractal,0.5"; -static struct perf_session *session; - static u64 sample_type; struct symbol_conf symbol_conf; @@ -604,7 +602,7 @@ static int validate_chain(struct ip_callchain *chain, event_t *event) return 0; } -static int process_sample_event(event_t *event) +static int process_sample_event(event_t *event, struct perf_session *session __used) { struct sample_data data; int cpumode; @@ -683,7 +681,7 @@ static int process_sample_event(event_t *event) return 0; } -static int process_comm_event(event_t *event) +static int process_comm_event(event_t *event, struct perf_session *session __used) { struct thread *thread = threads__findnew(event->comm.pid); @@ -698,7 +696,7 @@ static int process_comm_event(event_t *event) return 0; } -static int process_read_event(event_t *event) +static int process_read_event(event_t *event, struct perf_session *session __used) { struct perf_event_attr *attr; @@ -766,6 +764,7 @@ static int __cmd_report(void) { struct thread *idle; int ret; + struct perf_session *session; session = perf_session__new(input_name, O_RDONLY, force); if (session == NULL) diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 65021fe1361..48ab283ed86 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1594,7 +1594,8 @@ process_raw_event(event_t *raw_event __used, void *data, process_sched_migrate_task_event(data, event, cpu, timestamp, thread); } -static int process_sample_event(event_t *event) +static int process_sample_event(event_t *event, + struct perf_session *session __used) { struct sample_data data; struct thread *thread; @@ -1632,7 +1633,8 @@ static int process_sample_event(event_t *event) return 0; } -static int process_lost_event(event_t *event __used) +static int process_lost_event(event_t *event __used, + struct perf_session *session __used) { nr_lost_chunks++; nr_lost_events += event->lost.lost; diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 759dd2b35fd..db6caae1a40 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -281,21 +281,19 @@ static int cpus_cstate_state[MAX_CPUS]; static u64 cpus_pstate_start_times[MAX_CPUS]; static u64 cpus_pstate_state[MAX_CPUS]; -static int -process_comm_event(event_t *event) +static int process_comm_event(event_t *event, struct perf_session *session __used) { pid_set_comm(event->comm.pid, event->comm.comm); return 0; } -static int -process_fork_event(event_t *event) + +static int process_fork_event(event_t *event, struct perf_session *session __used) { pid_fork(event->fork.pid, event->fork.ppid, event->fork.time); return 0; } -static int -process_exit_event(event_t *event) +static int process_exit_event(event_t *event, struct perf_session *session __used) { pid_exit(event->fork.pid, event->fork.time); return 0; @@ -594,8 +592,7 @@ static u64 sample_time(event_t *event) * We first queue all events, sorted backwards by insertion. * The order will get flipped later. */ -static int -queue_sample_event(event_t *event) +static int queue_sample_event(event_t *event, struct perf_session *session __used) { struct sample_wrapper *copy, *prev; int size; diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index e0a374d0e43..b13f4262554 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -965,14 +965,14 @@ static void event__process_sample(const event_t *self, int counter) } } -static int event__process(event_t *event) +static int event__process(event_t *event, struct perf_session *session) { switch (event->header.type) { case PERF_RECORD_COMM: - event__process_comm(event); + event__process_comm(event, session); break; case PERF_RECORD_MMAP: - event__process_mmap(event); + event__process_mmap(event, session); break; default: break; @@ -999,7 +999,8 @@ static unsigned int mmap_read_head(struct mmap_data *md) return head; } -static void mmap_read_counter(struct mmap_data *md) +static void perf_session__mmap_read_counter(struct perf_session *self, + struct mmap_data *md) { unsigned int head = mmap_read_head(md); unsigned int old = md->prev; @@ -1054,7 +1055,7 @@ static void mmap_read_counter(struct mmap_data *md) if (event->header.type == PERF_RECORD_SAMPLE) event__process_sample(event, md->counter); else - event__process(event); + event__process(event, self); old += size; } @@ -1064,13 +1065,13 @@ static void mmap_read_counter(struct mmap_data *md) static struct pollfd event_array[MAX_NR_CPUS * MAX_COUNTERS]; static struct mmap_data mmap_array[MAX_NR_CPUS][MAX_COUNTERS]; -static void mmap_read(void) +static void perf_session__mmap_read(struct perf_session *self) { int i, counter; for (i = 0; i < nr_cpus; i++) { for (counter = 0; counter < nr_counters; counter++) - mmap_read_counter(&mmap_array[i][counter]); + perf_session__mmap_read_counter(self, &mmap_array[i][counter]); } } @@ -1155,11 +1156,16 @@ static int __cmd_top(void) pthread_t thread; int i, counter; int ret; + /* + * XXX perf_session__new should allow passing a O_MMAP, so that all this + * mmap reading, etc is encapsulated in it. + */ + struct perf_session *session = NULL; if (target_pid != -1) - event__synthesize_thread(target_pid, event__process); + event__synthesize_thread(target_pid, event__process, session); else - event__synthesize_threads(event__process); + event__synthesize_threads(event__process, session); for (i = 0; i < nr_cpus; i++) { group_fd = -1; @@ -1170,7 +1176,7 @@ static int __cmd_top(void) /* Wait for a minimal set of events before starting the snapshot */ poll(event_array, nr_poll, 100); - mmap_read(); + perf_session__mmap_read(session); if (pthread_create(&thread, NULL, display_thread, NULL)) { printf("Could not create display thread.\n"); @@ -1190,7 +1196,7 @@ static int __cmd_top(void) while (1) { int hits = samples; - mmap_read(); + perf_session__mmap_read(session); if (hits == samples) ret = poll(event_array, nr_poll, 100); diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 0756664666f..3057e1d387b 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -62,10 +62,9 @@ static int cleanup_scripting(void) static char const *input_name = "perf.data"; -static struct perf_session *session; static u64 sample_type; -static int process_sample_event(event_t *event) +static int process_sample_event(event_t *event, struct perf_session *session __used) { struct sample_data data; struct thread *thread; @@ -125,20 +124,12 @@ static struct perf_file_handler file_handler = { .sample_type_check = sample_type_check, }; -static int __cmd_trace(void) +static int __cmd_trace(struct perf_session *session) { - int err; - - session = perf_session__new(input_name, O_RDONLY, 0); - if (session == NULL) - return -ENOMEM; - register_idle_thread(); register_perf_file_handler(&file_handler); - err = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); - perf_session__delete(session); - return err; + return perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); } struct script_spec { @@ -313,6 +304,7 @@ static const struct option options[] = { int cmd_trace(int argc, const char **argv, const char *prefix __used) { int err; + struct perf_session *session; symbol__init(0); @@ -330,6 +322,10 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) setup_pager(); + session = perf_session__new(input_name, O_RDONLY, 0); + if (session == NULL) + return -ENOMEM; + if (generate_script_lang) { struct stat perf_stat; @@ -367,8 +363,9 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) goto out; } - err = __cmd_trace(); + err = __cmd_trace(session); + perf_session__delete(session); cleanup_scripting(); out: return err; diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index 6d46dda53a2..22bcdfeff55 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -8,7 +8,8 @@ static struct perf_file_handler *curr_handler; static unsigned long mmap_window = 32; static char __cwd[PATH_MAX]; -static int process_event_stub(event_t *event __used) +static int process_event_stub(event_t *event __used, + struct perf_session *session __used) { dump_printf(": unhandled!\n"); return 0; @@ -61,8 +62,8 @@ void event__print_totals(void) event__name[i], event__total[i]); } -static int -process_event(event_t *event, unsigned long offset, unsigned long head) +static int process_event(event_t *event, struct perf_session *session, + unsigned long offset, unsigned long head) { trace_event(event); @@ -77,23 +78,23 @@ process_event(event_t *event, unsigned long offset, unsigned long head) switch (event->header.type) { case PERF_RECORD_SAMPLE: - return curr_handler->process_sample_event(event); + return curr_handler->process_sample_event(event, session); case PERF_RECORD_MMAP: - return curr_handler->process_mmap_event(event); + return curr_handler->process_mmap_event(event, session); case PERF_RECORD_COMM: - return curr_handler->process_comm_event(event); + return curr_handler->process_comm_event(event, session); case PERF_RECORD_FORK: - return curr_handler->process_fork_event(event); + return curr_handler->process_fork_event(event, session); case PERF_RECORD_EXIT: - return curr_handler->process_exit_event(event); + return curr_handler->process_exit_event(event, session); case PERF_RECORD_LOST: - return curr_handler->process_lost_event(event); + return curr_handler->process_lost_event(event, session); case PERF_RECORD_READ: - return curr_handler->process_read_event(event); + return curr_handler->process_read_event(event, session); case PERF_RECORD_THROTTLE: - return curr_handler->process_throttle_event(event); + return curr_handler->process_throttle_event(event, session); case PERF_RECORD_UNTHROTTLE: - return curr_handler->process_unthrottle_event(event); + return curr_handler->process_unthrottle_event(event, session); default: curr_handler->total_unknown++; return -1; @@ -209,7 +210,7 @@ more: (void *)(long)event->header.size, event->header.type); - if (!size || process_event(event, offset, head) < 0) { + if (!size || process_event(event, self, offset, head) < 0) { dump_printf("%p [%p]: skipping unknown header type: %d\n", (void *)(offset + head), diff --git a/tools/perf/util/data_map.h b/tools/perf/util/data_map.h index 98c5b823388..6d4ae52bc07 100644 --- a/tools/perf/util/data_map.h +++ b/tools/perf/util/data_map.h @@ -5,7 +5,10 @@ #include "header.h" #include "session.h" -typedef int (*event_type_handler_t)(event_t *); +struct perf_session; + +typedef int (*event_type_handler_t)(event_t *self, + struct perf_session *session); struct perf_file_handler { event_type_handler_t process_sample_event; diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index ba0de90cd3d..e2c489533c6 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -5,7 +5,9 @@ #include "thread.h" static pid_t event__synthesize_comm(pid_t pid, int full, - int (*process)(event_t *event)) + int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session) { event_t ev; char filename[PATH_MAX]; @@ -54,7 +56,7 @@ out_race: if (!full) { ev.comm.tid = pid; - process(&ev); + process(&ev, session); goto out_fclose; } @@ -72,7 +74,7 @@ out_race: ev.comm.tid = pid; - process(&ev); + process(&ev, session); } closedir(tasks); @@ -86,7 +88,9 @@ out_failure: } static int event__synthesize_mmap_events(pid_t pid, pid_t tgid, - int (*process)(event_t *event)) + int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session) { char filename[PATH_MAX]; FILE *fp; @@ -141,7 +145,7 @@ static int event__synthesize_mmap_events(pid_t pid, pid_t tgid, ev.mmap.pid = tgid; ev.mmap.tid = pid; - process(&ev); + process(&ev, session); } } @@ -149,15 +153,20 @@ static int event__synthesize_mmap_events(pid_t pid, pid_t tgid, return 0; } -int event__synthesize_thread(pid_t pid, int (*process)(event_t *event)) +int event__synthesize_thread(pid_t pid, + int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session) { - pid_t tgid = event__synthesize_comm(pid, 1, process); + pid_t tgid = event__synthesize_comm(pid, 1, process, session); if (tgid == -1) return -1; - return event__synthesize_mmap_events(pid, tgid, process); + return event__synthesize_mmap_events(pid, tgid, process, session); } -void event__synthesize_threads(int (*process)(event_t *event)) +void event__synthesize_threads(int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session) { DIR *proc; struct dirent dirent, *next; @@ -171,7 +180,7 @@ void event__synthesize_threads(int (*process)(event_t *event)) if (*end) /* only interested in proper numerical dirents */ continue; - event__synthesize_thread(pid, process); + event__synthesize_thread(pid, process, session); } closedir(proc); @@ -182,7 +191,7 @@ int event__cwdlen; struct events_stats event__stats; -int event__process_comm(event_t *self) +int event__process_comm(event_t *self, struct perf_session *session __used) { struct thread *thread = threads__findnew(self->comm.pid); @@ -196,14 +205,14 @@ int event__process_comm(event_t *self) return 0; } -int event__process_lost(event_t *self) +int event__process_lost(event_t *self, struct perf_session *session __used) { dump_printf(": id:%Ld: lost:%Ld\n", self->lost.id, self->lost.lost); event__stats.lost += self->lost.lost; return 0; } -int event__process_mmap(event_t *self) +int event__process_mmap(event_t *self, struct perf_session *session __used) { struct thread *thread = threads__findnew(self->mmap.pid); struct map *map = map__new(&self->mmap, MAP__FUNCTION, @@ -224,7 +233,7 @@ int event__process_mmap(event_t *self) return 0; } -int event__process_task(event_t *self) +int event__process_task(event_t *self, struct perf_session *session __used) { struct thread *thread = threads__findnew(self->fork.pid); struct thread *parent = threads__findnew(self->fork.ppid); diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 51a96c2effd..6b6429b63da 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -156,18 +156,25 @@ struct symbol *map__find_symbol_by_name(struct map *self, const char *name, void map__fixup_start(struct map *self); void map__fixup_end(struct map *self); -int event__synthesize_thread(pid_t pid, int (*process)(event_t *event)); -void event__synthesize_threads(int (*process)(event_t *event)); +struct perf_session; + +int event__synthesize_thread(pid_t pid, + int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session); +void event__synthesize_threads(int (*process)(event_t *event, + struct perf_session *session), + struct perf_session *session); extern char *event__cwd; extern int event__cwdlen; extern struct events_stats event__stats; extern unsigned long event__total[PERF_RECORD_MAX]; -int event__process_comm(event_t *self); -int event__process_lost(event_t *self); -int event__process_mmap(event_t *self); -int event__process_task(event_t *self); +int event__process_comm(event_t *self, struct perf_session *session); +int event__process_lost(event_t *self, struct perf_session *session); +int event__process_mmap(event_t *self, struct perf_session *session); +int event__process_task(event_t *self, struct perf_session *session); struct addr_location; int event__preprocess_sample(const event_t *self, struct addr_location *al, -- cgit From 301a0b020210360c6e441c7765521248bc87d58e Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 13 Dec 2009 19:50:25 -0200 Subject: perf session: Ditch register_perf_file_handler MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Pass the event_ops to perf_session__process_events instead. Also move the event_ops definition to session.h, starting to move things around to their right place, trimming the many unneeded headers we have. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260741029-4430-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/Makefile | 1 - tools/perf/builtin-annotate.c | 10 +++++----- tools/perf/builtin-buildid-list.c | 1 - tools/perf/builtin-kmem.c | 8 +++----- tools/perf/builtin-report.c | 7 ++----- tools/perf/builtin-sched.c | 7 +++---- tools/perf/builtin-timechart.c | 9 ++++----- tools/perf/builtin-trace.c | 8 +++----- tools/perf/util/data_map.c | 41 +++++++++++++++++---------------------- tools/perf/util/data_map.h | 32 ------------------------------ tools/perf/util/header.c | 2 +- tools/perf/util/session.c | 3 ++- tools/perf/util/session.h | 26 ++++++++++++++++++++++++- 13 files changed, 66 insertions(+), 89 deletions(-) delete mode 100644 tools/perf/util/data_map.h (limited to 'tools') diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 406999668ca..a4cb7925538 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -370,7 +370,6 @@ LIB_H += util/values.h LIB_H += util/sort.h LIB_H += util/hist.h LIB_H += util/thread.h -LIB_H += util/data_map.h LIB_H += util/probe-finder.h LIB_H += util/probe-event.h diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 2e418b9fc1c..43e3bb35444 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -26,7 +26,6 @@ #include "util/sort.h" #include "util/hist.h" #include "util/session.h" -#include "util/data_map.h" static char const *input_name = "perf.data"; @@ -454,7 +453,7 @@ static void find_annotations(void) } } -static struct perf_file_handler file_handler = { +static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_mmap_event = event__process_mmap, .process_comm_event = event__process_comm, @@ -463,7 +462,8 @@ static struct perf_file_handler file_handler = { static int __cmd_annotate(void) { - struct perf_session *session = perf_session__new(input_name, O_RDONLY, force); + struct perf_session *session = perf_session__new(input_name, O_RDONLY, + force); struct thread *idle; int ret; @@ -471,9 +471,9 @@ static int __cmd_annotate(void) return -ENOMEM; idle = register_idle_thread(); - register_perf_file_handler(&file_handler); - ret = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, &event_ops, 0, + &event__cwdlen, &event__cwd); if (ret) goto out_delete; diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index bfd16a1594e..2629f76d95f 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -9,7 +9,6 @@ #include "builtin.h" #include "perf.h" #include "util/cache.h" -#include "util/data_map.h" #include "util/debug.h" #include "util/parse-options.h" #include "util/session.h" diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 101b2682b4d..6f74bd8d20e 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -12,7 +12,6 @@ #include "util/trace-event.h" #include "util/debug.h" -#include "util/data_map.h" #include @@ -359,7 +358,7 @@ static int sample_type_check(u64 type) return 0; } -static struct perf_file_handler file_handler = { +static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, .sample_type_check = sample_type_check, @@ -374,9 +373,8 @@ static int read_events(void) return -ENOMEM; register_idle_thread(); - register_perf_file_handler(&file_handler); - - err = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + err = perf_session__process_events(session, &event_ops, 0, + &event__cwdlen, &event__cwd); perf_session__delete(session); return err; } diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index dcd8fedc298..c203eaf73a2 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -27,7 +27,6 @@ #include "util/parse-options.h" #include "util/parse-events.h" -#include "util/data_map.h" #include "util/thread.h" #include "util/sort.h" #include "util/hist.h" @@ -748,7 +747,7 @@ static int sample_type_check(u64 type) return 0; } -static struct perf_file_handler file_handler = { +static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_mmap_event = event__process_mmap, .process_comm_event = process_comm_event, @@ -776,9 +775,7 @@ static int __cmd_report(void) if (show_threads) perf_read_values_init(&show_threads_values); - register_perf_file_handler(&file_handler); - - ret = perf_session__process_events(session, full_paths, + ret = perf_session__process_events(session, &event_ops, full_paths, &event__cwdlen, &event__cwd); if (ret) goto out_delete; diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 48ab283ed86..b5b44723626 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -12,7 +12,6 @@ #include "util/trace-event.h" #include "util/debug.h" -#include "util/data_map.h" #include @@ -1656,7 +1655,7 @@ static int sample_type_check(u64 type) return 0; } -static struct perf_file_handler file_handler = { +static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, .process_lost_event = process_lost_event, @@ -1672,9 +1671,9 @@ static int read_events(void) return -ENOMEM; register_idle_thread(); - register_perf_file_handler(&file_handler); - err = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + err = perf_session__process_events(session, &event_ops, 0, + &event__cwdlen, &event__cwd); perf_session__delete(session); return err; } diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index db6caae1a40..4b95cec6b4c 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -30,7 +30,7 @@ #include "util/parse-options.h" #include "util/parse-events.h" #include "util/event.h" -#include "util/data_map.h" +#include "util/session.h" #include "util/svghelper.h" static char const *input_name = "perf.data"; @@ -1046,7 +1046,7 @@ static int sample_type_check(u64 type) return 0; } -static struct perf_file_handler file_handler = { +static struct perf_event_ops event_ops = { .process_comm_event = process_comm_event, .process_fork_event = process_fork_event, .process_exit_event = process_exit_event, @@ -1062,9 +1062,8 @@ static int __cmd_timechart(void) if (session == NULL) return -ENOMEM; - register_perf_file_handler(&file_handler); - - ret = perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, &event_ops, 0, + &event__cwdlen, &event__cwd); if (ret) goto out_delete; diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 3057e1d387b..c404dece950 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -57,7 +57,6 @@ static int cleanup_scripting(void) #include "util/debug.h" #include "util/trace-event.h" -#include "util/data_map.h" #include "util/exec_cmd.h" static char const *input_name = "perf.data"; @@ -118,7 +117,7 @@ static int sample_type_check(u64 type) return 0; } -static struct perf_file_handler file_handler = { +static struct perf_event_ops event_ops = { .process_sample_event = process_sample_event, .process_comm_event = event__process_comm, .sample_type_check = sample_type_check, @@ -127,9 +126,8 @@ static struct perf_file_handler file_handler = { static int __cmd_trace(struct perf_session *session) { register_idle_thread(); - register_perf_file_handler(&file_handler); - - return perf_session__process_events(session, 0, &event__cwdlen, &event__cwd); + return perf_session__process_events(session, &event_ops, 0, + &event__cwdlen, &event__cwd); } struct script_spec { diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index 22bcdfeff55..ba5bcfa1f90 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -1,10 +1,8 @@ -#include "data_map.h" #include "symbol.h" #include "util.h" #include "debug.h" +#include "session.h" - -static struct perf_file_handler *curr_handler; static unsigned long mmap_window = 32; static char __cwd[PATH_MAX]; @@ -15,7 +13,7 @@ static int process_event_stub(event_t *event __used, return 0; } -void register_perf_file_handler(struct perf_file_handler *handler) +static void perf_event_ops__fill_defaults(struct perf_event_ops *handler) { if (!handler->process_sample_event) handler->process_sample_event = process_event_stub; @@ -35,8 +33,6 @@ void register_perf_file_handler(struct perf_file_handler *handler) handler->process_throttle_event = process_event_stub; if (!handler->process_unthrottle_event) handler->process_unthrottle_event = process_event_stub; - - curr_handler = handler; } static const char *event__name[] = { @@ -63,6 +59,7 @@ void event__print_totals(void) } static int process_event(event_t *event, struct perf_session *session, + struct perf_event_ops *ops, unsigned long offset, unsigned long head) { trace_event(event); @@ -78,25 +75,25 @@ static int process_event(event_t *event, struct perf_session *session, switch (event->header.type) { case PERF_RECORD_SAMPLE: - return curr_handler->process_sample_event(event, session); + return ops->process_sample_event(event, session); case PERF_RECORD_MMAP: - return curr_handler->process_mmap_event(event, session); + return ops->process_mmap_event(event, session); case PERF_RECORD_COMM: - return curr_handler->process_comm_event(event, session); + return ops->process_comm_event(event, session); case PERF_RECORD_FORK: - return curr_handler->process_fork_event(event, session); + return ops->process_fork_event(event, session); case PERF_RECORD_EXIT: - return curr_handler->process_exit_event(event, session); + return ops->process_exit_event(event, session); case PERF_RECORD_LOST: - return curr_handler->process_lost_event(event, session); + return ops->process_lost_event(event, session); case PERF_RECORD_READ: - return curr_handler->process_read_event(event, session); + return ops->process_read_event(event, session); case PERF_RECORD_THROTTLE: - return curr_handler->process_throttle_event(event, session); + return ops->process_throttle_event(event, session); case PERF_RECORD_UNTHROTTLE: - return curr_handler->process_unthrottle_event(event, session); + return ops->process_unthrottle_event(event, session); default: - curr_handler->total_unknown++; + ops->total_unknown++; return -1; } } @@ -131,6 +128,7 @@ out: } int perf_session__process_events(struct perf_session *self, + struct perf_event_ops *ops, int full_paths, int *cwdlen, char **cwd) { int err; @@ -142,10 +140,7 @@ int perf_session__process_events(struct perf_session *self, uint32_t size; char *buf; - if (curr_handler == NULL) { - pr_debug("Forgot to register perf file handler\n"); - return -EINVAL; - } + perf_event_ops__fill_defaults(ops); page_size = getpagesize(); @@ -153,8 +148,8 @@ int perf_session__process_events(struct perf_session *self, sample_type = perf_header__sample_type(&self->header); err = -EINVAL; - if (curr_handler->sample_type_check && - curr_handler->sample_type_check(sample_type) < 0) + if (ops->sample_type_check && + ops->sample_type_check(sample_type) < 0) goto out_err; if (!full_paths) { @@ -210,7 +205,7 @@ more: (void *)(long)event->header.size, event->header.type); - if (!size || process_event(event, self, offset, head) < 0) { + if (!size || process_event(event, self, ops, offset, head) < 0) { dump_printf("%p [%p]: skipping unknown header type: %d\n", (void *)(offset + head), diff --git a/tools/perf/util/data_map.h b/tools/perf/util/data_map.h deleted file mode 100644 index 6d4ae52bc07..00000000000 --- a/tools/perf/util/data_map.h +++ /dev/null @@ -1,32 +0,0 @@ -#ifndef __PERF_DATAMAP_H -#define __PERF_DATAMAP_H - -#include "event.h" -#include "header.h" -#include "session.h" - -struct perf_session; - -typedef int (*event_type_handler_t)(event_t *self, - struct perf_session *session); - -struct perf_file_handler { - event_type_handler_t process_sample_event; - event_type_handler_t process_mmap_event; - event_type_handler_t process_comm_event; - event_type_handler_t process_fork_event; - event_type_handler_t process_exit_event; - event_type_handler_t process_lost_event; - event_type_handler_t process_read_event; - event_type_handler_t process_throttle_event; - event_type_handler_t process_unthrottle_event; - int (*sample_type_check)(u64 sample_type); - unsigned long total_unknown; -}; - -void register_perf_file_handler(struct perf_file_handler *handler); -int perf_session__process_events(struct perf_session *self, - int full_paths, int *cwdlen, char **cwd); -int perf_header__read_build_ids(int input, u64 offset, u64 file_size); - -#endif diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c index f2e8d871511..8a0bca55106 100644 --- a/tools/perf/util/header.c +++ b/tools/perf/util/header.c @@ -8,8 +8,8 @@ #include "header.h" #include "../perf.h" #include "trace-event.h" +#include "session.h" #include "symbol.h" -#include "data_map.h" #include "debug.h" /* diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 707ce1cb162..39766868d43 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -48,7 +48,8 @@ out_close: return -1; } -struct perf_session *perf_session__new(const char *filename, int mode, bool force) +struct perf_session *perf_session__new(const char *filename, int mode, + bool force) { size_t len = strlen(filename) + 1; struct perf_session *self = zalloc(sizeof(*self) + len); diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index f3699c8c8ed..7a4c32c545f 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -1,6 +1,7 @@ #ifndef __PERF_SESSION_H #define __PERF_SESSION_H +#include "event.h" #include "header.h" struct perf_session { @@ -10,7 +11,30 @@ struct perf_session { char filename[0]; }; -struct perf_session *perf_session__new(const char *filename, int mode, bool force); +typedef int (*event_op)(event_t *self, struct perf_session *session); + +struct perf_event_ops { + event_op process_sample_event; + event_op process_mmap_event; + event_op process_comm_event; + event_op process_fork_event; + event_op process_exit_event; + event_op process_lost_event; + event_op process_read_event; + event_op process_throttle_event; + event_op process_unthrottle_event; + int (*sample_type_check)(u64 sample_type); + unsigned long total_unknown; +}; + +struct perf_session *perf_session__new(const char *filename, int mode, + bool force); void perf_session__delete(struct perf_session *self); +int perf_session__process_events(struct perf_session *self, + struct perf_event_ops *event_ops, + int full_paths, int *cwdlen, char **cwd); + +int perf_header__read_build_ids(int input, u64 offset, u64 file_size); + #endif /* __PERF_SESSION_H */ -- cgit From 13df45ca1c9a430d5f53862854070fcc324e015c Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 13 Dec 2009 19:50:26 -0200 Subject: perf session: Register the idle thread in perf_session__process_events MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit No need for all tools to register it and then immediately call perf_session__process_events. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260741029-4430-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 3 --- tools/perf/builtin-kmem.c | 1 - tools/perf/builtin-report.c | 4 ---- tools/perf/builtin-sched.c | 2 -- tools/perf/builtin-trace.c | 1 - tools/perf/util/data_map.c | 16 ++++++++++++++++ tools/perf/util/thread.c | 12 ------------ tools/perf/util/thread.h | 1 - 8 files changed, 16 insertions(+), 24 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 43e3bb35444..93d765a746f 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -464,14 +464,11 @@ static int __cmd_annotate(void) { struct perf_session *session = perf_session__new(input_name, O_RDONLY, force); - struct thread *idle; int ret; if (session == NULL) return -ENOMEM; - idle = register_idle_thread(); - ret = perf_session__process_events(session, &event_ops, 0, &event__cwdlen, &event__cwd); if (ret) diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 6f74bd8d20e..37a849906af 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -372,7 +372,6 @@ static int read_events(void) if (session == NULL) return -ENOMEM; - register_idle_thread(); err = perf_session__process_events(session, &event_ops, 0, &event__cwdlen, &event__cwd); perf_session__delete(session); diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index c203eaf73a2..4b37ac4f128 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -761,7 +761,6 @@ static struct perf_event_ops event_ops = { static int __cmd_report(void) { - struct thread *idle; int ret; struct perf_session *session; @@ -769,9 +768,6 @@ static int __cmd_report(void) if (session == NULL) return -ENOMEM; - idle = register_idle_thread(); - thread__comm_adjust(idle); - if (show_threads) perf_read_values_init(&show_threads_values); diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index b5b44723626..847ed51248b 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1670,8 +1670,6 @@ static int read_events(void) if (session == NULL) return -ENOMEM; - register_idle_thread(); - err = perf_session__process_events(session, &event_ops, 0, &event__cwdlen, &event__cwd); perf_session__delete(session); diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index c404dece950..40cb896581d 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -125,7 +125,6 @@ static struct perf_event_ops event_ops = { static int __cmd_trace(struct perf_session *session) { - register_idle_thread(); return perf_session__process_events(session, &event_ops, 0, &event__cwdlen, &event__cwd); } diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index ba5bcfa1f90..36e3bfe7319 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -1,6 +1,7 @@ #include "symbol.h" #include "util.h" #include "debug.h" +#include "thread.h" #include "session.h" static unsigned long mmap_window = 32; @@ -127,6 +128,18 @@ out: return err; } +static struct thread *perf_session__register_idle_thread(struct perf_session *self __used) +{ + struct thread *thread = threads__findnew(0); + + if (!thread || thread__set_comm(thread, "swapper")) { + pr_err("problem inserting idle task.\n"); + thread = NULL; + } + + return thread; +} + int perf_session__process_events(struct perf_session *self, struct perf_event_ops *ops, int full_paths, int *cwdlen, char **cwd) @@ -140,6 +153,9 @@ int perf_session__process_events(struct perf_session *self, uint32_t size; char *buf; + if (perf_session__register_idle_thread(self) == NULL) + return -ENOMEM; + perf_event_ops__fill_defaults(ops); page_size = getpagesize(); diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index b68a00ea412..5c0ab14f3db 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -161,18 +161,6 @@ struct thread *threads__findnew(pid_t pid) return th; } -struct thread *register_idle_thread(void) -{ - struct thread *thread = threads__findnew(0); - - if (!thread || thread__set_comm(thread, "swapper")) { - fprintf(stderr, "problem inserting idle task.\n"); - exit(-1); - } - - return thread; -} - static void map_groups__remove_overlappings(struct map_groups *self, struct map *map) { diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index 1751802a09b..2e35e1f6bb4 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -24,7 +24,6 @@ void map_groups__init(struct map_groups *self); int thread__set_comm(struct thread *self, const char *comm); int thread__comm_len(struct thread *self); struct thread *threads__findnew(pid_t pid); -struct thread *register_idle_thread(void); void thread__insert_map(struct thread *self, struct map *map); int thread__fork(struct thread *self, struct thread *parent); size_t map_groups__fprintf_maps(struct map_groups *self, FILE *fp); -- cgit From ec913369733923dbfd6bdff5953a918107059701 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 13 Dec 2009 19:50:27 -0200 Subject: perf session: Reduce the number of parms to perf_session__process_events MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit By having the cwd/cwdlen in the perf_session struct and full_paths in perf_event_ops. Now its just a matter of passing the ops. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260741029-4430-4-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 3 +-- tools/perf/builtin-kmem.c | 3 +-- tools/perf/builtin-report.c | 6 ++---- tools/perf/builtin-sched.c | 3 +-- tools/perf/builtin-timechart.c | 3 +-- tools/perf/builtin-trace.c | 3 +-- tools/perf/util/data_map.c | 32 ++++++++++++++++---------------- tools/perf/util/event.c | 8 +++----- tools/perf/util/session.c | 4 ++++ tools/perf/util/session.h | 7 +++++-- 10 files changed, 35 insertions(+), 37 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 93d765a746f..a931b133f3a 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -469,8 +469,7 @@ static int __cmd_annotate(void) if (session == NULL) return -ENOMEM; - ret = perf_session__process_events(session, &event_ops, 0, - &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, &event_ops); if (ret) goto out_delete; diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 37a849906af..237155fa756 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -372,8 +372,7 @@ static int read_events(void) if (session == NULL) return -ENOMEM; - err = perf_session__process_events(session, &event_ops, 0, - &event__cwdlen, &event__cwd); + err = perf_session__process_events(session, &event_ops); perf_session__delete(session); return err; } diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 4b37ac4f128..26b94786094 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -39,7 +39,6 @@ static struct strlist *dso_list, *comm_list, *sym_list; static int force; -static int full_paths; static int show_nr_samples; static int show_threads; @@ -771,8 +770,7 @@ static int __cmd_report(void) if (show_threads) perf_read_values_init(&show_threads_values); - ret = perf_session__process_events(session, &event_ops, full_paths, - &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, &event_ops); if (ret) goto out_delete; @@ -877,7 +875,7 @@ static const struct option options[] = { "pretty printing style key: normal raw"), OPT_STRING('s', "sort", &sort_order, "key[,key2...]", "sort by key(s): pid, comm, dso, symbol, parent"), - OPT_BOOLEAN('P', "full-paths", &full_paths, + OPT_BOOLEAN('P', "full-paths", &event_ops.full_paths, "Don't shorten the pathnames taking into account the cwd"), OPT_STRING('p', "parent", &parent_pattern, "regex", "regex filter to identify parent, see: '--sort parent'"), diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 847ed51248b..1e4e508339a 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1670,8 +1670,7 @@ static int read_events(void) if (session == NULL) return -ENOMEM; - err = perf_session__process_events(session, &event_ops, 0, - &event__cwdlen, &event__cwd); + err = perf_session__process_events(session, &event_ops); perf_session__delete(session); return err; } diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 4b95cec6b4c..b5211facddc 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1062,8 +1062,7 @@ static int __cmd_timechart(void) if (session == NULL) return -ENOMEM; - ret = perf_session__process_events(session, &event_ops, 0, - &event__cwdlen, &event__cwd); + ret = perf_session__process_events(session, &event_ops); if (ret) goto out_delete; diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 40cb896581d..b7eb3fcc224 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -125,8 +125,7 @@ static struct perf_event_ops event_ops = { static int __cmd_trace(struct perf_session *session) { - return perf_session__process_events(session, &event_ops, 0, - &event__cwdlen, &event__cwd); + return perf_session__process_events(session, &event_ops); } struct script_spec { diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index 36e3bfe7319..ba2eb2ce018 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -4,9 +4,6 @@ #include "thread.h" #include "session.h" -static unsigned long mmap_window = 32; -static char __cwd[PATH_MAX]; - static int process_event_stub(event_t *event __used, struct perf_session *session __used) { @@ -141,8 +138,7 @@ static struct thread *perf_session__register_idle_thread(struct perf_session *se } int perf_session__process_events(struct perf_session *self, - struct perf_event_ops *ops, - int full_paths, int *cwdlen, char **cwd) + struct perf_event_ops *ops) { int err; unsigned long head, shift; @@ -168,17 +164,21 @@ int perf_session__process_events(struct perf_session *self, ops->sample_type_check(sample_type) < 0) goto out_err; - if (!full_paths) { - if (getcwd(__cwd, sizeof(__cwd)) == NULL) { - pr_err("failed to get the current directory\n"); + if (!ops->full_paths) { + char bf[PATH_MAX]; + + if (getcwd(bf, sizeof(bf)) == NULL) { err = -errno; +out_getcwd_err: + pr_err("failed to get the current directory\n"); goto out_err; } - *cwd = __cwd; - *cwdlen = strlen(*cwd); - } else { - *cwd = NULL; - *cwdlen = 0; + self->cwd = strdup(bf); + if (self->cwd == NULL) { + err = -ENOMEM; + goto out_getcwd_err; + } + self->cwdlen = strlen(self->cwd); } shift = page_size * (head / page_size); @@ -186,7 +186,7 @@ int perf_session__process_events(struct perf_session *self, head -= shift; remap: - buf = mmap(NULL, page_size * mmap_window, PROT_READ, + buf = mmap(NULL, page_size * self->mmap_window, PROT_READ, MAP_SHARED, self->fd, offset); if (buf == MAP_FAILED) { pr_err("failed to mmap file\n"); @@ -201,12 +201,12 @@ more: if (!size) size = 8; - if (head + event->header.size >= page_size * mmap_window) { + if (head + event->header.size >= page_size * self->mmap_window) { int munmap_ret; shift = page_size * (head / page_size); - munmap_ret = munmap(buf, page_size * mmap_window); + munmap_ret = munmap(buf, page_size * self->mmap_window); assert(munmap_ret == 0); offset += shift; diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index e2c489533c6..40d8d842a21 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -1,6 +1,7 @@ #include #include "event.h" #include "debug.h" +#include "session.h" #include "string.h" #include "thread.h" @@ -186,9 +187,6 @@ void event__synthesize_threads(int (*process)(event_t *event, closedir(proc); } -char *event__cwd; -int event__cwdlen; - struct events_stats event__stats; int event__process_comm(event_t *self, struct perf_session *session __used) @@ -212,11 +210,11 @@ int event__process_lost(event_t *self, struct perf_session *session __used) return 0; } -int event__process_mmap(event_t *self, struct perf_session *session __used) +int event__process_mmap(event_t *self, struct perf_session *session) { struct thread *thread = threads__findnew(self->mmap.pid); struct map *map = map__new(&self->mmap, MAP__FUNCTION, - event__cwd, event__cwdlen); + session->cwd, session->cwdlen); dump_printf(" %d/%d: [%p(%p) @ %p]: %s\n", self->mmap.pid, self->mmap.tid, diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 39766868d43..534a8770ee7 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -61,6 +61,9 @@ struct perf_session *perf_session__new(const char *filename, int mode, goto out_delete; memcpy(self->filename, filename, len); + self->mmap_window = 32; + self->cwd = NULL; + self->cwdlen = 0; if (mode == O_RDONLY && perf_session__open(self, force) < 0) { perf_session__delete(self); @@ -77,5 +80,6 @@ void perf_session__delete(struct perf_session *self) { perf_header__exit(&self->header); close(self->fd); + free(self->cwd); free(self); } diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 7a4c32c545f..1e0da9ca31a 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -7,7 +7,10 @@ struct perf_session { struct perf_header header; unsigned long size; + unsigned long mmap_window; int fd; + int cwdlen; + char *cwd; char filename[0]; }; @@ -25,6 +28,7 @@ struct perf_event_ops { event_op process_unthrottle_event; int (*sample_type_check)(u64 sample_type); unsigned long total_unknown; + bool full_paths; }; struct perf_session *perf_session__new(const char *filename, int mode, @@ -32,8 +36,7 @@ struct perf_session *perf_session__new(const char *filename, int mode, void perf_session__delete(struct perf_session *self); int perf_session__process_events(struct perf_session *self, - struct perf_event_ops *event_ops, - int full_paths, int *cwdlen, char **cwd); + struct perf_event_ops *event_ops); int perf_header__read_build_ids(int input, u64 offset, u64 file_size); -- cgit From b3165f414416a717f72a376720564012af5a2e01 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 13 Dec 2009 19:50:28 -0200 Subject: perf session: Move the global threads list to perf_session MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit So that we can process two perf.data files. We still need to add a O_MMAP mode for perf_session so that we can do all the mmap stuff in it. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260741029-4430-5-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 6 ++-- tools/perf/builtin-kmem.c | 4 +-- tools/perf/builtin-report.c | 10 +++---- tools/perf/builtin-sched.c | 68 ++++++++++++++++++++++++++----------------- tools/perf/builtin-top.c | 19 +++++++----- tools/perf/builtin-trace.c | 4 +-- tools/perf/util/data_map.c | 4 +-- tools/perf/util/event.c | 18 ++++++------ tools/perf/util/event.h | 4 +-- tools/perf/util/session.c | 4 ++- tools/perf/util/session.h | 5 ++++ tools/perf/util/thread.c | 22 +++++++------- tools/perf/util/thread.h | 4 +-- 13 files changed, 98 insertions(+), 74 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index a931b133f3a..795f865c136 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -131,14 +131,14 @@ static int hist_entry__add(struct addr_location *al, u64 count) return 0; } -static int process_sample_event(event_t *event, struct perf_session *session __used) +static int process_sample_event(event_t *event, struct perf_session *session) { struct addr_location al; dump_printf("(IP, %d): %d: %p\n", event->header.misc, event->ip.pid, (void *)(long)event->ip.ip); - if (event__preprocess_sample(event, &al, symbol_filter) < 0) { + if (event__preprocess_sample(event, session, &al, symbol_filter) < 0) { fprintf(stderr, "problem processing %d event, skipping it.\n", event->header.type); return -1; @@ -479,7 +479,7 @@ static int __cmd_annotate(void) } if (verbose > 3) - threads__fprintf(stdout); + perf_session__fprintf(session, stdout); if (verbose > 2) dsos__fprintf(stdout); diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 237155fa756..de194958fe6 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -311,7 +311,7 @@ process_raw_event(event_t *raw_event __used, void *data, } } -static int process_sample_event(event_t *event, struct perf_session *session __used) +static int process_sample_event(event_t *event, struct perf_session *session) { struct sample_data data; struct thread *thread; @@ -329,7 +329,7 @@ static int process_sample_event(event_t *event, struct perf_session *session __u (void *)(long)data.ip, (long long)data.period); - thread = threads__findnew(event->ip.pid); + thread = perf_session__findnew(session, event->ip.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 26b94786094..efa8147b899 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -600,7 +600,7 @@ static int validate_chain(struct ip_callchain *chain, event_t *event) return 0; } -static int process_sample_event(event_t *event, struct perf_session *session __used) +static int process_sample_event(event_t *event, struct perf_session *session) { struct sample_data data; int cpumode; @@ -636,7 +636,7 @@ static int process_sample_event(event_t *event, struct perf_session *session __u } } - thread = threads__findnew(data.pid); + thread = perf_session__findnew(session, data.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); @@ -679,9 +679,9 @@ static int process_sample_event(event_t *event, struct perf_session *session __u return 0; } -static int process_comm_event(event_t *event, struct perf_session *session __used) +static int process_comm_event(event_t *event, struct perf_session *session) { - struct thread *thread = threads__findnew(event->comm.pid); + struct thread *thread = perf_session__findnew(session, event->comm.pid); dump_printf(": %s:%d\n", event->comm.comm, event->comm.pid); @@ -780,7 +780,7 @@ static int __cmd_report(void) } if (verbose > 3) - threads__fprintf(stdout); + perf_session__fprintf(session, stdout); if (verbose > 2) dsos__fprintf(stdout); diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 1e4e508339a..8d58d9e07a7 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -730,18 +730,21 @@ struct trace_migrate_task_event { struct trace_sched_handler { void (*switch_event)(struct trace_switch_event *, + struct perf_session *, struct event *, int cpu, u64 timestamp, struct thread *thread); void (*runtime_event)(struct trace_runtime_event *, + struct perf_session *, struct event *, int cpu, u64 timestamp, struct thread *thread); void (*wakeup_event)(struct trace_wakeup_event *, + struct perf_session *, struct event *, int cpu, u64 timestamp, @@ -754,6 +757,7 @@ struct trace_sched_handler { struct thread *thread); void (*migrate_task_event)(struct trace_migrate_task_event *, + struct perf_session *session, struct event *, int cpu, u64 timestamp, @@ -763,6 +767,7 @@ struct trace_sched_handler { static void replay_wakeup_event(struct trace_wakeup_event *wakeup_event, + struct perf_session *session __used, struct event *event, int cpu __used, u64 timestamp __used, @@ -789,6 +794,7 @@ static u64 cpu_last_switched[MAX_CPUS]; static void replay_switch_event(struct trace_switch_event *switch_event, + struct perf_session *session __used, struct event *event, int cpu, u64 timestamp, @@ -1022,6 +1028,7 @@ add_sched_in_event(struct work_atoms *atoms, u64 timestamp) static void latency_switch_event(struct trace_switch_event *switch_event, + struct perf_session *session, struct event *event __used, int cpu, u64 timestamp, @@ -1045,8 +1052,8 @@ latency_switch_event(struct trace_switch_event *switch_event, die("hm, delta: %Ld < 0 ?\n", delta); - sched_out = threads__findnew(switch_event->prev_pid); - sched_in = threads__findnew(switch_event->next_pid); + sched_out = perf_session__findnew(session, switch_event->prev_pid); + sched_in = perf_session__findnew(session, switch_event->next_pid); out_events = thread_atoms_search(&atom_root, sched_out, &cmp_pid); if (!out_events) { @@ -1074,12 +1081,13 @@ latency_switch_event(struct trace_switch_event *switch_event, static void latency_runtime_event(struct trace_runtime_event *runtime_event, + struct perf_session *session, struct event *event __used, int cpu, u64 timestamp, struct thread *this_thread __used) { - struct thread *thread = threads__findnew(runtime_event->pid); + struct thread *thread = perf_session__findnew(session, runtime_event->pid); struct work_atoms *atoms = thread_atoms_search(&atom_root, thread, &cmp_pid); BUG_ON(cpu >= MAX_CPUS || cpu < 0); @@ -1096,6 +1104,7 @@ latency_runtime_event(struct trace_runtime_event *runtime_event, static void latency_wakeup_event(struct trace_wakeup_event *wakeup_event, + struct perf_session *session, struct event *__event __used, int cpu __used, u64 timestamp, @@ -1109,7 +1118,7 @@ latency_wakeup_event(struct trace_wakeup_event *wakeup_event, if (!wakeup_event->success) return; - wakee = threads__findnew(wakeup_event->pid); + wakee = perf_session__findnew(session, wakeup_event->pid); atoms = thread_atoms_search(&atom_root, wakee, &cmp_pid); if (!atoms) { thread_atoms_insert(wakee); @@ -1143,6 +1152,7 @@ latency_wakeup_event(struct trace_wakeup_event *wakeup_event, static void latency_migrate_task_event(struct trace_migrate_task_event *migrate_task_event, + struct perf_session *session, struct event *__event __used, int cpu __used, u64 timestamp, @@ -1158,7 +1168,7 @@ latency_migrate_task_event(struct trace_migrate_task_event *migrate_task_event, if (profile_cpu == -1) return; - migrant = threads__findnew(migrate_task_event->pid); + migrant = perf_session__findnew(session, migrate_task_event->pid); atoms = thread_atoms_search(&atom_root, migrant, &cmp_pid); if (!atoms) { thread_atoms_insert(migrant); @@ -1353,7 +1363,7 @@ static void sort_lat(void) static struct trace_sched_handler *trace_handler; static void -process_sched_wakeup_event(void *data, +process_sched_wakeup_event(void *data, struct perf_session *session, struct event *event, int cpu __used, u64 timestamp __used, @@ -1370,7 +1380,8 @@ process_sched_wakeup_event(void *data, FILL_FIELD(wakeup_event, cpu, event, data); if (trace_handler->wakeup_event) - trace_handler->wakeup_event(&wakeup_event, event, cpu, timestamp, thread); + trace_handler->wakeup_event(&wakeup_event, session, event, + cpu, timestamp, thread); } /* @@ -1388,6 +1399,7 @@ static char next_shortname2 = '0'; static void map_switch_event(struct trace_switch_event *switch_event, + struct perf_session *session, struct event *event __used, int this_cpu, u64 timestamp, @@ -1415,8 +1427,8 @@ map_switch_event(struct trace_switch_event *switch_event, die("hm, delta: %Ld < 0 ?\n", delta); - sched_out = threads__findnew(switch_event->prev_pid); - sched_in = threads__findnew(switch_event->next_pid); + sched_out = perf_session__findnew(session, switch_event->prev_pid); + sched_in = perf_session__findnew(session, switch_event->next_pid); curr_thread[this_cpu] = sched_in; @@ -1466,7 +1478,7 @@ map_switch_event(struct trace_switch_event *switch_event, static void -process_sched_switch_event(void *data, +process_sched_switch_event(void *data, struct perf_session *session, struct event *event, int this_cpu, u64 timestamp __used, @@ -1493,13 +1505,14 @@ process_sched_switch_event(void *data, nr_context_switch_bugs++; } if (trace_handler->switch_event) - trace_handler->switch_event(&switch_event, event, this_cpu, timestamp, thread); + trace_handler->switch_event(&switch_event, session, event, + this_cpu, timestamp, thread); curr_pid[this_cpu] = switch_event.next_pid; } static void -process_sched_runtime_event(void *data, +process_sched_runtime_event(void *data, struct perf_session *session, struct event *event, int cpu __used, u64 timestamp __used, @@ -1513,7 +1526,7 @@ process_sched_runtime_event(void *data, FILL_FIELD(runtime_event, vruntime, event, data); if (trace_handler->runtime_event) - trace_handler->runtime_event(&runtime_event, event, cpu, timestamp, thread); + trace_handler->runtime_event(&runtime_event, session, event, cpu, timestamp, thread); } static void @@ -1533,7 +1546,8 @@ process_sched_fork_event(void *data, FILL_FIELD(fork_event, child_pid, event, data); if (trace_handler->fork_event) - trace_handler->fork_event(&fork_event, event, cpu, timestamp, thread); + trace_handler->fork_event(&fork_event, event, + cpu, timestamp, thread); } static void @@ -1547,7 +1561,7 @@ process_sched_exit_event(struct event *event, } static void -process_sched_migrate_task_event(void *data, +process_sched_migrate_task_event(void *data, struct perf_session *session, struct event *event, int cpu __used, u64 timestamp __used, @@ -1563,12 +1577,13 @@ process_sched_migrate_task_event(void *data, FILL_FIELD(migrate_task_event, cpu, event, data); if (trace_handler->migrate_task_event) - trace_handler->migrate_task_event(&migrate_task_event, event, cpu, timestamp, thread); + trace_handler->migrate_task_event(&migrate_task_event, session, + event, cpu, timestamp, thread); } static void -process_raw_event(event_t *raw_event __used, void *data, - int cpu, u64 timestamp, struct thread *thread) +process_raw_event(event_t *raw_event __used, struct perf_session *session, + void *data, int cpu, u64 timestamp, struct thread *thread) { struct event *event; int type; @@ -1578,23 +1593,22 @@ process_raw_event(event_t *raw_event __used, void *data, event = trace_find_event(type); if (!strcmp(event->name, "sched_switch")) - process_sched_switch_event(data, event, cpu, timestamp, thread); + process_sched_switch_event(data, session, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_stat_runtime")) - process_sched_runtime_event(data, event, cpu, timestamp, thread); + process_sched_runtime_event(data, session, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_wakeup")) - process_sched_wakeup_event(data, event, cpu, timestamp, thread); + process_sched_wakeup_event(data, session, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_wakeup_new")) - process_sched_wakeup_event(data, event, cpu, timestamp, thread); + process_sched_wakeup_event(data, session, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_process_fork")) process_sched_fork_event(data, event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_process_exit")) process_sched_exit_event(event, cpu, timestamp, thread); if (!strcmp(event->name, "sched_migrate_task")) - process_sched_migrate_task_event(data, event, cpu, timestamp, thread); + process_sched_migrate_task_event(data, session, event, cpu, timestamp, thread); } -static int process_sample_event(event_t *event, - struct perf_session *session __used) +static int process_sample_event(event_t *event, struct perf_session *session) { struct sample_data data; struct thread *thread; @@ -1615,7 +1629,7 @@ static int process_sample_event(event_t *event, (void *)(long)data.ip, (long long)data.period); - thread = threads__findnew(data.pid); + thread = perf_session__findnew(session, data.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); @@ -1627,7 +1641,7 @@ static int process_sample_event(event_t *event, if (profile_cpu != -1 && profile_cpu != (int)data.cpu) return 0; - process_raw_event(event, data.raw_data, data.cpu, data.time, thread); + process_raw_event(event, session, data.raw_data, data.cpu, data.time, thread); return 0; } diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index b13f4262554..0f7a4da2924 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -20,8 +20,9 @@ #include "perf.h" -#include "util/symbol.h" #include "util/color.h" +#include "util/session.h" +#include "util/symbol.h" #include "util/thread.h" #include "util/util.h" #include @@ -926,7 +927,8 @@ static int symbol_filter(struct map *map, struct symbol *sym) return 0; } -static void event__process_sample(const event_t *self, int counter) +static void event__process_sample(const event_t *self, + struct perf_session *session, int counter) { u64 ip = self->ip.ip; struct sym_entry *syme; @@ -946,7 +948,7 @@ static void event__process_sample(const event_t *self, int counter) return; } - if (event__preprocess_sample(self, &al, symbol_filter) < 0 || + if (event__preprocess_sample(self, session, &al, symbol_filter) < 0 || al.sym == NULL) return; @@ -1053,7 +1055,7 @@ static void perf_session__mmap_read_counter(struct perf_session *self, } if (event->header.type == PERF_RECORD_SAMPLE) - event__process_sample(event, md->counter); + event__process_sample(event, self, md->counter); else event__process(event, self); old += size; @@ -1157,10 +1159,13 @@ static int __cmd_top(void) int i, counter; int ret; /* - * XXX perf_session__new should allow passing a O_MMAP, so that all this - * mmap reading, etc is encapsulated in it. + * FIXME: perf_session__new should allow passing a O_MMAP, so that all this + * mmap reading, etc is encapsulated in it. Use O_WRONLY for now. */ - struct perf_session *session = NULL; + struct perf_session *session = perf_session__new(NULL, O_WRONLY, false); + + if (session == NULL) + return -ENOMEM; if (target_pid != -1) event__synthesize_thread(target_pid, event__process, session); diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index b7eb3fcc224..d7653237505 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -63,7 +63,7 @@ static char const *input_name = "perf.data"; static u64 sample_type; -static int process_sample_event(event_t *event, struct perf_session *session __used) +static int process_sample_event(event_t *event, struct perf_session *session) { struct sample_data data; struct thread *thread; @@ -81,7 +81,7 @@ static int process_sample_event(event_t *event, struct perf_session *session __u (void *)(long)data.ip, (long long)data.period); - thread = threads__findnew(event->ip.pid); + thread = perf_session__findnew(session, event->ip.pid); if (thread == NULL) { pr_debug("problem processing %d event, skipping it.\n", event->header.type); diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index ba2eb2ce018..44dea211cc6 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -125,9 +125,9 @@ out: return err; } -static struct thread *perf_session__register_idle_thread(struct perf_session *self __used) +static struct thread *perf_session__register_idle_thread(struct perf_session *self) { - struct thread *thread = threads__findnew(0); + struct thread *thread = perf_session__findnew(self, 0); if (!thread || thread__set_comm(thread, "swapper")) { pr_err("problem inserting idle task.\n"); diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 40d8d842a21..2d09c29b3a6 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -189,9 +189,9 @@ void event__synthesize_threads(int (*process)(event_t *event, struct events_stats event__stats; -int event__process_comm(event_t *self, struct perf_session *session __used) +int event__process_comm(event_t *self, struct perf_session *session) { - struct thread *thread = threads__findnew(self->comm.pid); + struct thread *thread = perf_session__findnew(session, self->comm.pid); dump_printf(": %s:%d\n", self->comm.comm, self->comm.pid); @@ -212,7 +212,7 @@ int event__process_lost(event_t *self, struct perf_session *session __used) int event__process_mmap(event_t *self, struct perf_session *session) { - struct thread *thread = threads__findnew(self->mmap.pid); + struct thread *thread = perf_session__findnew(session, self->mmap.pid); struct map *map = map__new(&self->mmap, MAP__FUNCTION, session->cwd, session->cwdlen); @@ -231,10 +231,10 @@ int event__process_mmap(event_t *self, struct perf_session *session) return 0; } -int event__process_task(event_t *self, struct perf_session *session __used) +int event__process_task(event_t *self, struct perf_session *session) { - struct thread *thread = threads__findnew(self->fork.pid); - struct thread *parent = threads__findnew(self->fork.ppid); + struct thread *thread = perf_session__findnew(session, self->fork.pid); + struct thread *parent = perf_session__findnew(session, self->fork.ppid); dump_printf("(%d:%d):(%d:%d)\n", self->fork.pid, self->fork.tid, self->fork.ppid, self->fork.ptid); @@ -300,11 +300,11 @@ try_again: } } -int event__preprocess_sample(const event_t *self, struct addr_location *al, - symbol_filter_t filter) +int event__preprocess_sample(const event_t *self, struct perf_session *session, + struct addr_location *al, symbol_filter_t filter) { u8 cpumode = self->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; - struct thread *thread = threads__findnew(self->ip.pid); + struct thread *thread = perf_session__findnew(session, self->ip.pid); if (thread == NULL) return -1; diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 6b6429b63da..bb090257570 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -177,8 +177,8 @@ int event__process_mmap(event_t *self, struct perf_session *session); int event__process_task(event_t *self, struct perf_session *session); struct addr_location; -int event__preprocess_sample(const event_t *self, struct addr_location *al, - symbol_filter_t filter); +int event__preprocess_sample(const event_t *self, struct perf_session *session, + struct addr_location *al, symbol_filter_t filter); int event__parse_sample(event_t *event, u64 type, struct sample_data *data); #endif /* __PERF_RECORD_H */ diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 534a8770ee7..09836a537fc 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -51,7 +51,7 @@ out_close: struct perf_session *perf_session__new(const char *filename, int mode, bool force) { - size_t len = strlen(filename) + 1; + size_t len = filename ? strlen(filename) + 1 : 0; struct perf_session *self = zalloc(sizeof(*self) + len); if (self == NULL) @@ -61,6 +61,8 @@ struct perf_session *perf_session__new(const char *filename, int mode, goto out_delete; memcpy(self->filename, filename, len); + self->threads = RB_ROOT; + self->last_match = NULL; self->mmap_window = 32; self->cwd = NULL; self->cwdlen = 0; diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 1e0da9ca31a..1dbef7cdd48 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -3,11 +3,16 @@ #include "event.h" #include "header.h" +#include + +struct thread; struct perf_session { struct perf_header header; unsigned long size; unsigned long mmap_window; + struct rb_root threads; + struct thread *last_match; int fd; int cwdlen; char *cwd; diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 5c0ab14f3db..634b7f7140d 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -2,13 +2,11 @@ #include #include #include +#include "session.h" #include "thread.h" #include "util.h" #include "debug.h" -static struct rb_root threads; -static struct thread *last_match; - void map_groups__init(struct map_groups *self) { int i; @@ -122,9 +120,9 @@ static size_t thread__fprintf(struct thread *self, FILE *fp) map_groups__fprintf(&self->mg, fp); } -struct thread *threads__findnew(pid_t pid) +struct thread *perf_session__findnew(struct perf_session *self, pid_t pid) { - struct rb_node **p = &threads.rb_node; + struct rb_node **p = &self->threads.rb_node; struct rb_node *parent = NULL; struct thread *th; @@ -133,15 +131,15 @@ struct thread *threads__findnew(pid_t pid) * so most of the time we dont have to look up * the full rbtree: */ - if (last_match && last_match->pid == pid) - return last_match; + if (self->last_match && self->last_match->pid == pid) + return self->last_match; while (*p != NULL) { parent = *p; th = rb_entry(parent, struct thread, rb_node); if (th->pid == pid) { - last_match = th; + self->last_match = th; return th; } @@ -154,8 +152,8 @@ struct thread *threads__findnew(pid_t pid) th = thread__new(pid); if (th != NULL) { rb_link_node(&th->rb_node, parent, p); - rb_insert_color(&th->rb_node, &threads); - last_match = th; + rb_insert_color(&th->rb_node, &self->threads); + self->last_match = th; } return th; @@ -269,12 +267,12 @@ int thread__fork(struct thread *self, struct thread *parent) return 0; } -size_t threads__fprintf(FILE *fp) +size_t perf_session__fprintf(struct perf_session *self, FILE *fp) { size_t ret = 0; struct rb_node *nd; - for (nd = rb_first(&threads); nd; nd = rb_next(nd)) { + for (nd = rb_first(&self->threads); nd; nd = rb_next(nd)) { struct thread *pos = rb_entry(nd, struct thread, rb_node); ret += thread__fprintf(pos, fp); diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index 2e35e1f6bb4..e93abf2d9cb 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -23,11 +23,11 @@ struct thread { void map_groups__init(struct map_groups *self); int thread__set_comm(struct thread *self, const char *comm); int thread__comm_len(struct thread *self); -struct thread *threads__findnew(pid_t pid); +struct thread *perf_session__findnew(struct perf_session *self, pid_t pid); void thread__insert_map(struct thread *self, struct map *map); int thread__fork(struct thread *self, struct thread *parent); size_t map_groups__fprintf_maps(struct map_groups *self, FILE *fp); -size_t threads__fprintf(FILE *fp); +size_t perf_session__fprintf(struct perf_session *self, FILE *fp); void maps__insert(struct rb_root *maps, struct map *map); struct map *maps__find(struct rb_root *maps, u64 addr); -- cgit From 4aa65636411ccb12f006a6ad593930655c445ff6 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Sun, 13 Dec 2009 19:50:29 -0200 Subject: perf session: Move kmaps to perf_session MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There is still some more work to do to disentangle map creation from DSO loading, but this happens only for the kernel, and for the early adopters of perf diff, where this disentanglement matters most, we'll be testing different kernels, so no problem here. Further clarification: right now we create the kernel maps for the various modules and discontiguous kernel text maps when loading the DSO, we should do it as a two step process, first creating the maps, for multiple mappings with the same DSO store, then doing the dso load just once, for the first hit on one of the maps sharing this DSO backing store. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260741029-4430-6-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 2 +- tools/perf/builtin-buildid-list.c | 4 +- tools/perf/builtin-kmem.c | 41 +++++++++--------- tools/perf/builtin-record.c | 2 +- tools/perf/builtin-report.c | 14 +++--- tools/perf/builtin-sched.c | 4 +- tools/perf/builtin-timechart.c | 3 +- tools/perf/builtin-top.c | 4 +- tools/perf/builtin-trace.c | 2 +- tools/perf/util/event.c | 13 +++--- tools/perf/util/event.h | 10 +++-- tools/perf/util/map.c | 14 +++--- tools/perf/util/session.c | 19 ++++++--- tools/perf/util/session.h | 6 ++- tools/perf/util/symbol.c | 89 +++++++++++++++++++-------------------- tools/perf/util/symbol.h | 11 ++--- tools/perf/util/thread.c | 3 +- tools/perf/util/thread.h | 11 ++--- 18 files changed, 134 insertions(+), 118 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 795f865c136..e44c54c79be 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -463,7 +463,7 @@ static struct perf_event_ops event_ops = { static int __cmd_annotate(void) { struct perf_session *session = perf_session__new(input_name, O_RDONLY, - force); + force, &symbol_conf); int ret; if (session == NULL) diff --git a/tools/perf/builtin-buildid-list.c b/tools/perf/builtin-buildid-list.c index 2629f76d95f..7c36e4b2ecc 100644 --- a/tools/perf/builtin-buildid-list.c +++ b/tools/perf/builtin-buildid-list.c @@ -54,8 +54,8 @@ static int perf_file_section__process_buildids(struct perf_file_section *self, static int __cmd_buildid_list(void) { int err = -1; - struct perf_session *session = perf_session__new(input_name, O_RDONLY, force); - + struct perf_session *session = perf_session__new(input_name, O_RDONLY, + force, NULL); if (session == NULL) return -1; diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index de194958fe6..e79ecbc1718 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -364,19 +364,6 @@ static struct perf_event_ops event_ops = { .sample_type_check = sample_type_check, }; -static int read_events(void) -{ - int err; - struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); - - if (session == NULL) - return -ENOMEM; - - err = perf_session__process_events(session, &event_ops); - perf_session__delete(session); - return err; -} - static double fragmentation(unsigned long n_req, unsigned long n_alloc) { if (n_alloc == 0) @@ -385,7 +372,8 @@ static double fragmentation(unsigned long n_req, unsigned long n_alloc) return 100.0 - (100.0 * n_req / n_alloc); } -static void __print_result(struct rb_root *root, int n_lines, int is_caller) +static void __print_result(struct rb_root *root, struct perf_session *session, + int n_lines, int is_caller) { struct rb_node *next; @@ -406,7 +394,7 @@ static void __print_result(struct rb_root *root, int n_lines, int is_caller) if (is_caller) { addr = data->call_site; if (!raw_ip) - sym = map_groups__find_function(kmaps, addr, NULL); + sym = map_groups__find_function(&session->kmaps, session, addr, NULL); } else addr = data->ptr; @@ -447,12 +435,12 @@ static void print_summary(void) printf("Cross CPU allocations: %lu/%lu\n", nr_cross_allocs, nr_allocs); } -static void print_result(void) +static void print_result(struct perf_session *session) { if (caller_flag) - __print_result(&root_caller_sorted, caller_lines, 1); + __print_result(&root_caller_sorted, session, caller_lines, 1); if (alloc_flag) - __print_result(&root_alloc_sorted, alloc_lines, 0); + __print_result(&root_alloc_sorted, session, alloc_lines, 0); print_summary(); } @@ -520,12 +508,21 @@ static void sort_result(void) static int __cmd_kmem(void) { + int err; + struct perf_session *session = perf_session__new(input_name, O_RDONLY, + 0, NULL); + if (session == NULL) + return -ENOMEM; + setup_pager(); - read_events(); + err = perf_session__process_events(session, &event_ops); + if (err != 0) + goto out_delete; sort_result(); - print_result(); - - return 0; + print_result(session); +out_delete: + perf_session__delete(session); + return err; } static const char * const kmem_usage[] = { diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index b7e15a1b1ec..a66a58d5281 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -442,7 +442,7 @@ static int __cmd_record(int argc, const char **argv) exit(-1); } - session = perf_session__new(output_name, O_WRONLY, force); + session = perf_session__new(output_name, O_WRONLY, force, NULL); if (session == NULL) { pr_err("Not enough memory for reading perf file header\n"); return -1; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index efa8147b899..3487224670a 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -414,6 +414,7 @@ static int call__match(struct symbol *sym) } static struct symbol **resolve_callchain(struct thread *thread, + struct perf_session *session, struct ip_callchain *chain, struct symbol **parent) { @@ -447,8 +448,8 @@ static struct symbol **resolve_callchain(struct thread *thread, continue; } - thread__find_addr_location(thread, cpumode, MAP__FUNCTION, - ip, &al, NULL); + thread__find_addr_location(thread, session, cpumode, + MAP__FUNCTION, ip, &al, NULL); if (al.sym != NULL) { if (sort__has_parent && !*parent && call__match(al.sym)) @@ -467,6 +468,7 @@ static struct symbol **resolve_callchain(struct thread *thread, */ static int hist_entry__add(struct addr_location *al, + struct perf_session *session, struct ip_callchain *chain, u64 count) { struct symbol **syms = NULL, *parent = NULL; @@ -474,7 +476,7 @@ static int hist_entry__add(struct addr_location *al, struct hist_entry *he; if ((sort__has_parent || callchain) && chain) - syms = resolve_callchain(al->thread, chain, &parent); + syms = resolve_callchain(al->thread, session, chain, &parent); he = __hist_entry__add(al, parent, count, &hit); if (he == NULL) @@ -650,7 +652,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) cpumode = event->header.misc & PERF_RECORD_MISC_CPUMODE_MASK; - thread__find_addr_location(thread, cpumode, + thread__find_addr_location(thread, session, cpumode, MAP__FUNCTION, data.ip, &al, NULL); /* * We have to do this here as we may have a dso with no symbol hit that @@ -669,7 +671,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) if (sym_list && al.sym && !strlist__has_entry(sym_list, al.sym->name)) return 0; - if (hist_entry__add(&al, data.callchain, data.period)) { + if (hist_entry__add(&al, session, data.callchain, data.period)) { pr_debug("problem incrementing symbol count, skipping event\n"); return -1; } @@ -763,7 +765,7 @@ static int __cmd_report(void) int ret; struct perf_session *session; - session = perf_session__new(input_name, O_RDONLY, force); + session = perf_session__new(input_name, O_RDONLY, force, &symbol_conf); if (session == NULL) return -ENOMEM; diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 8d58d9e07a7..bce05dff6df 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1679,8 +1679,8 @@ static struct perf_event_ops event_ops = { static int read_events(void) { int err; - struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); - + struct perf_session *session = perf_session__new(input_name, O_RDONLY, + 0, NULL); if (session == NULL) return -ENOMEM; diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index b5211facddc..27018531404 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1056,7 +1056,8 @@ static struct perf_event_ops event_ops = { static int __cmd_timechart(void) { - struct perf_session *session = perf_session__new(input_name, O_RDONLY, 0); + struct perf_session *session = perf_session__new(input_name, O_RDONLY, + 0, NULL); int ret; if (session == NULL) diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 0f7a4da2924..3b212bb2e4d 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -1162,8 +1162,8 @@ static int __cmd_top(void) * FIXME: perf_session__new should allow passing a O_MMAP, so that all this * mmap reading, etc is encapsulated in it. Use O_WRONLY for now. */ - struct perf_session *session = perf_session__new(NULL, O_WRONLY, false); - + struct perf_session *session = perf_session__new(NULL, O_WRONLY, false, + &symbol_conf); if (session == NULL) return -ENOMEM; diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index d7653237505..9d89ae423ca 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -318,7 +318,7 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) setup_pager(); - session = perf_session__new(input_name, O_RDONLY, 0); + session = perf_session__new(input_name, O_RDONLY, 0, NULL); if (session == NULL) return -ENOMEM; diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 2d09c29b3a6..222efb1fc3b 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -256,7 +256,8 @@ int event__process_task(event_t *self, struct perf_session *session) return 0; } -void thread__find_addr_location(struct thread *self, u8 cpumode, +void thread__find_addr_location(struct thread *self, + struct perf_session *session, u8 cpumode, enum map_type type, u64 addr, struct addr_location *al, symbol_filter_t filter) @@ -268,7 +269,7 @@ void thread__find_addr_location(struct thread *self, u8 cpumode, if (cpumode & PERF_RECORD_MISC_KERNEL) { al->level = 'k'; - mg = kmaps; + mg = &session->kmaps; } else if (cpumode & PERF_RECORD_MISC_USER) al->level = '.'; else { @@ -289,14 +290,14 @@ try_again: * "[vdso]" dso, but for now lets use the old trick of looking * in the whole kernel symbol list. */ - if ((long long)al->addr < 0 && mg != kmaps) { - mg = kmaps; + if ((long long)al->addr < 0 && mg != &session->kmaps) { + mg = &session->kmaps; goto try_again; } al->sym = NULL; } else { al->addr = al->map->map_ip(al->map, al->addr); - al->sym = map__find_symbol(al->map, al->addr, filter); + al->sym = map__find_symbol(al->map, session, al->addr, filter); } } @@ -311,7 +312,7 @@ int event__preprocess_sample(const event_t *self, struct perf_session *session, dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - thread__find_addr_location(thread, cpumode, MAP__FUNCTION, + thread__find_addr_location(thread, session, cpumode, MAP__FUNCTION, self->ip.ip, al, filter); dump_printf(" ...... dso: %s\n", al->map ? al->map->dso->long_name : diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index bb090257570..035ecf3c25c 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -149,15 +149,17 @@ void map__delete(struct map *self); struct map *map__clone(struct map *self); int map__overlap(struct map *l, struct map *r); size_t map__fprintf(struct map *self, FILE *fp); -struct symbol *map__find_symbol(struct map *self, u64 addr, - symbol_filter_t filter); + +struct perf_session; + +struct symbol *map__find_symbol(struct map *self, struct perf_session *session, + u64 addr, symbol_filter_t filter); struct symbol *map__find_symbol_by_name(struct map *self, const char *name, + struct perf_session *session, symbol_filter_t filter); void map__fixup_start(struct map *self); void map__fixup_end(struct map *self); -struct perf_session; - int event__synthesize_thread(pid_t pid, int (*process)(event_t *event, struct perf_session *session), diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index 76bdca640a9..8b3dd467adb 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -104,10 +104,11 @@ void map__fixup_end(struct map *self) #define DSO__DELETED "(deleted)" -static int map__load(struct map *self, symbol_filter_t filter) +static int map__load(struct map *self, struct perf_session *session, + symbol_filter_t filter) { const char *name = self->dso->long_name; - int nr = dso__load(self->dso, self, filter); + int nr = dso__load(self->dso, self, session, filter); if (nr < 0) { if (self->dso->has_build_id) { @@ -143,19 +144,20 @@ static int map__load(struct map *self, symbol_filter_t filter) return 0; } -struct symbol *map__find_symbol(struct map *self, u64 addr, - symbol_filter_t filter) +struct symbol *map__find_symbol(struct map *self, struct perf_session *session, + u64 addr, symbol_filter_t filter) { - if (!dso__loaded(self->dso, self->type) && map__load(self, filter) < 0) + if (!dso__loaded(self->dso, self->type) && map__load(self, session, filter) < 0) return NULL; return dso__find_symbol(self->dso, self->type, addr); } struct symbol *map__find_symbol_by_name(struct map *self, const char *name, + struct perf_session *session, symbol_filter_t filter) { - if (!dso__loaded(self->dso, self->type) && map__load(self, filter) < 0) + if (!dso__loaded(self->dso, self->type) && map__load(self, session, filter) < 0) return NULL; if (!dso__sorted_by_name(self->dso, self->type)) diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index 09836a537fc..fe87a2f2e5a 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -49,7 +49,7 @@ out_close: } struct perf_session *perf_session__new(const char *filename, int mode, - bool force) + bool force, struct symbol_conf *conf) { size_t len = filename ? strlen(filename) + 1 : 0; struct perf_session *self = zalloc(sizeof(*self) + len); @@ -58,7 +58,7 @@ struct perf_session *perf_session__new(const char *filename, int mode, goto out; if (perf_header__init(&self->header) < 0) - goto out_delete; + goto out_free; memcpy(self->filename, filename, len); self->threads = RB_ROOT; @@ -66,16 +66,21 @@ struct perf_session *perf_session__new(const char *filename, int mode, self->mmap_window = 32; self->cwd = NULL; self->cwdlen = 0; + map_groups__init(&self->kmaps); - if (mode == O_RDONLY && perf_session__open(self, force) < 0) { - perf_session__delete(self); - self = NULL; - } + if (perf_session__create_kernel_maps(self, conf) < 0) + goto out_delete; + + if (mode == O_RDONLY && perf_session__open(self, force) < 0) + goto out_delete; out: return self; -out_delete: +out_free: free(self); return NULL; +out_delete: + perf_session__delete(self); + return NULL; } void perf_session__delete(struct perf_session *self) diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 1dbef7cdd48..20b2c9cc834 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -3,19 +3,23 @@ #include "event.h" #include "header.h" +#include "thread.h" #include struct thread; +struct symbol_conf; struct perf_session { struct perf_header header; unsigned long size; unsigned long mmap_window; + struct map_groups kmaps; struct rb_root threads; struct thread *last_match; int fd; int cwdlen; char *cwd; + bool use_modules; char filename[0]; }; @@ -37,7 +41,7 @@ struct perf_event_ops { }; struct perf_session *perf_session__new(const char *filename, int mode, - bool force); + bool force, struct symbol_conf *conf); void perf_session__delete(struct perf_session *self); int perf_session__process_events(struct perf_session *self, diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index d3d9fed74f1..185b9eec192 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1,5 +1,6 @@ #include "util.h" #include "../perf.h" +#include "session.h" #include "string.h" #include "symbol.h" #include "thread.h" @@ -31,7 +32,7 @@ enum dso_origin { static void dsos__add(struct list_head *head, struct dso *dso); static struct map *map__new2(u64 start, struct dso *dso, enum map_type type); static int dso__load_kernel_sym(struct dso *self, struct map *map, - struct map_groups *mg, symbol_filter_t filter); + struct perf_session *session, symbol_filter_t filter); unsigned int symbol__priv_size; static int vmlinux_path__nr_entries; static char **vmlinux_path; @@ -41,9 +42,6 @@ static struct symbol_conf symbol_conf__defaults = { .try_vmlinux_path = true, }; -static struct map_groups kmaps_mem; -struct map_groups *kmaps = &kmaps_mem; - bool dso__loaded(const struct dso *self, enum map_type type) { return self->loaded & (1 << type); @@ -456,7 +454,7 @@ out_failure: * the original ELF section names vmlinux have. */ static int dso__split_kallsyms(struct dso *self, struct map *map, - struct map_groups *mg, symbol_filter_t filter) + struct perf_session *session, symbol_filter_t filter) { struct map *curr_map = map; struct symbol *pos; @@ -473,13 +471,13 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, module = strchr(pos->name, '\t'); if (module) { - if (!mg->use_modules) + if (!session->use_modules) goto discard_symbol; *module++ = '\0'; if (strcmp(self->name, module)) { - curr_map = map_groups__find_by_name(mg, map->type, module); + curr_map = map_groups__find_by_name(&session->kmaps, map->type, module); if (curr_map == NULL) { pr_debug("/proc/{kallsyms,modules} " "inconsistency!\n"); @@ -510,7 +508,7 @@ static int dso__split_kallsyms(struct dso *self, struct map *map, } curr_map->map_ip = curr_map->unmap_ip = identity__map_ip; - map_groups__insert(mg, curr_map); + map_groups__insert(&session->kmaps, curr_map); ++kernel_range; } @@ -531,7 +529,7 @@ discard_symbol: rb_erase(&pos->rb_node, root); static int dso__load_kallsyms(struct dso *self, struct map *map, - struct map_groups *mg, symbol_filter_t filter) + struct perf_session *session, symbol_filter_t filter) { if (dso__load_all_kallsyms(self, map) < 0) return -1; @@ -539,14 +537,7 @@ static int dso__load_kallsyms(struct dso *self, struct map *map, symbols__fixup_end(&self->symbols[map->type]); self->origin = DSO__ORIG_KERNEL; - return dso__split_kallsyms(self, map, mg, filter); -} - -size_t kernel_maps__fprintf(FILE *fp) -{ - size_t printed = fprintf(fp, "Kernel maps:\n"); - printed += map_groups__fprintf_maps(kmaps, fp); - return printed + fprintf(fp, "END kernel maps\n"); + return dso__split_kallsyms(self, map, session, filter); } static int dso__load_perf_map(struct dso *self, struct map *map, @@ -873,7 +864,7 @@ static bool elf_sec__is_a(GElf_Shdr *self, Elf_Data *secstrs, enum map_type type } static int dso__load_sym(struct dso *self, struct map *map, - struct map_groups *mg, const char *name, int fd, + struct perf_session *session, const char *name, int fd, symbol_filter_t filter, int kernel, int kmodule) { struct map *curr_map = map; @@ -977,7 +968,7 @@ static int dso__load_sym(struct dso *self, struct map *map, snprintf(dso_name, sizeof(dso_name), "%s%s", self->short_name, section_name); - curr_map = map_groups__find_by_name(mg, map->type, dso_name); + curr_map = map_groups__find_by_name(&session->kmaps, map->type, dso_name); if (curr_map == NULL) { u64 start = sym.st_value; @@ -996,7 +987,7 @@ static int dso__load_sym(struct dso *self, struct map *map, curr_map->map_ip = identity__map_ip; curr_map->unmap_ip = identity__map_ip; curr_dso->origin = DSO__ORIG_KERNEL; - map_groups__insert(kmaps, curr_map); + map_groups__insert(&session->kmaps, curr_map); dsos__add(&dsos__kernel, curr_dso); } else curr_dso = curr_map->dso; @@ -1211,7 +1202,8 @@ char dso__symtab_origin(const struct dso *self) return origin[self->origin]; } -int dso__load(struct dso *self, struct map *map, symbol_filter_t filter) +int dso__load(struct dso *self, struct map *map, struct perf_session *session, + symbol_filter_t filter) { int size = PATH_MAX; char *name; @@ -1222,7 +1214,7 @@ int dso__load(struct dso *self, struct map *map, symbol_filter_t filter) dso__set_loaded(self, map->type); if (self->kernel) - return dso__load_kernel_sym(self, map, kmaps, filter); + return dso__load_kernel_sym(self, map, session, filter); name = malloc(size); if (!name) @@ -1323,7 +1315,7 @@ struct map *map_groups__find_by_name(struct map_groups *self, return NULL; } -static int dsos__set_modules_path_dir(char *dirname) +static int perf_session__set_modules_path_dir(struct perf_session *self, char *dirname) { struct dirent *dent; DIR *dir = opendir(dirname); @@ -1343,7 +1335,7 @@ static int dsos__set_modules_path_dir(char *dirname) snprintf(path, sizeof(path), "%s/%s", dirname, dent->d_name); - if (dsos__set_modules_path_dir(path) < 0) + if (perf_session__set_modules_path_dir(self, path) < 0) goto failure; } else { char *dot = strrchr(dent->d_name, '.'), @@ -1357,7 +1349,7 @@ static int dsos__set_modules_path_dir(char *dirname) (int)(dot - dent->d_name), dent->d_name); strxfrchar(dso_name, '-', '_'); - map = map_groups__find_by_name(kmaps, MAP__FUNCTION, dso_name); + map = map_groups__find_by_name(&self->kmaps, MAP__FUNCTION, dso_name); if (map == NULL) continue; @@ -1377,7 +1369,7 @@ failure: return -1; } -static int dsos__set_modules_path(void) +static int perf_session__set_modules_path(struct perf_session *self) { struct utsname uts; char modules_path[PATH_MAX]; @@ -1388,7 +1380,7 @@ static int dsos__set_modules_path(void) snprintf(modules_path, sizeof(modules_path), "/lib/modules/%s/kernel", uts.release); - return dsos__set_modules_path_dir(modules_path); + return perf_session__set_modules_path_dir(self, modules_path); } /* @@ -1410,7 +1402,7 @@ static struct map *map__new2(u64 start, struct dso *dso, enum map_type type) return self; } -static int map_groups__create_module_maps(struct map_groups *self) +static int perf_session__create_module_maps(struct perf_session *self) { char *line = NULL; size_t n; @@ -1467,14 +1459,14 @@ static int map_groups__create_module_maps(struct map_groups *self) dso->has_build_id = true; dso->origin = DSO__ORIG_KMODULE; - map_groups__insert(self, map); + map_groups__insert(&self->kmaps, map); dsos__add(&dsos__kernel, dso); } free(line); fclose(file); - return dsos__set_modules_path(); + return perf_session__set_modules_path(self); out_delete_line: free(line); @@ -1483,7 +1475,7 @@ out_failure: } static int dso__load_vmlinux(struct dso *self, struct map *map, - struct map_groups *mg, + struct perf_session *session, const char *vmlinux, symbol_filter_t filter) { int err = -1, fd; @@ -1517,14 +1509,14 @@ static int dso__load_vmlinux(struct dso *self, struct map *map, return -1; dso__set_loaded(self, map->type); - err = dso__load_sym(self, map, mg, self->long_name, fd, filter, 1, 0); + err = dso__load_sym(self, map, session, self->long_name, fd, filter, 1, 0); close(fd); return err; } static int dso__load_kernel_sym(struct dso *self, struct map *map, - struct map_groups *mg, symbol_filter_t filter) + struct perf_session *session, symbol_filter_t filter) { int err; bool is_kallsyms; @@ -1534,7 +1526,7 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, pr_debug("Looking at the vmlinux_path (%d entries long)\n", vmlinux_path__nr_entries); for (i = 0; i < vmlinux_path__nr_entries; ++i) { - err = dso__load_vmlinux(self, map, mg, + err = dso__load_vmlinux(self, map, session, vmlinux_path[i], filter); if (err > 0) { pr_debug("Using %s for symbols\n", @@ -1550,12 +1542,12 @@ static int dso__load_kernel_sym(struct dso *self, struct map *map, if (is_kallsyms) goto do_kallsyms; - err = dso__load_vmlinux(self, map, mg, self->long_name, filter); + err = dso__load_vmlinux(self, map, session, self->long_name, filter); if (err <= 0) { pr_info("The file %s cannot be used, " "trying to use /proc/kallsyms...", self->long_name); do_kallsyms: - err = dso__load_kallsyms(self, map, mg, filter); + err = dso__load_kallsyms(self, map, session, filter); if (err > 0 && !is_kallsyms) dso__set_long_name(self, strdup("[kernel.kallsyms]")); } @@ -1757,23 +1749,30 @@ int symbol__init(struct symbol_conf *conf) if (pconf->sort_by_name) symbol__priv_size += (sizeof(struct symbol_name_rb_node) - sizeof(struct symbol)); - map_groups__init(kmaps); if (pconf->try_vmlinux_path && vmlinux_path__init() < 0) return -1; - if (map_groups__create_kernel_maps(kmaps, pconf->vmlinux_name) < 0) { - vmlinux_path__exit(); + return 0; +} + +int perf_session__create_kernel_maps(struct perf_session *self, + struct symbol_conf *conf) +{ + const struct symbol_conf *pconf = conf ?: &symbol_conf__defaults; + + if (map_groups__create_kernel_maps(&self->kmaps, + pconf->vmlinux_name) < 0) return -1; - } - kmaps->use_modules = pconf->use_modules; - if (pconf->use_modules && map_groups__create_module_maps(kmaps) < 0) - pr_debug("Failed to load list of modules in use, " - "continuing...\n"); + self->use_modules = pconf->use_modules; + + if (pconf->use_modules && perf_session__create_module_maps(self) < 0) + pr_debug("Failed to load list of modules for session %s, " + "continuing...\n", self->filename); /* * Now that we have all the maps created, just set the ->end of them: */ - map_groups__fixup_end(kmaps); + map_groups__fixup_end(&self->kmaps); return 0; } diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index cf99f88adf3..941ef331790 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -98,8 +98,11 @@ bool dso__sorted_by_name(const struct dso *self, enum map_type type); void dso__sort_by_name(struct dso *self, enum map_type type); +struct perf_session; + struct dso *dsos__findnew(const char *name); -int dso__load(struct dso *self, struct map *map, symbol_filter_t filter); +int dso__load(struct dso *self, struct map *map, struct perf_session *session, + symbol_filter_t filter); void dsos__fprintf(FILE *fp); size_t dsos__fprintf_buildid(FILE *fp); @@ -116,12 +119,10 @@ int sysfs__read_build_id(const char *filename, void *bf, size_t size); bool dsos__read_build_ids(void); int build_id__sprintf(u8 *self, int len, char *bf); -size_t kernel_maps__fprintf(FILE *fp); - int symbol__init(struct symbol_conf *conf); +int perf_session__create_kernel_maps(struct perf_session *self, + struct symbol_conf *conf); -struct map_groups; -struct map_groups *kmaps; extern struct list_head dsos__user, dsos__kernel; extern struct dso *vdso; #endif /* __PERF_SYMBOL */ diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index 634b7f7140d..4a08dcf50b6 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -282,13 +282,14 @@ size_t perf_session__fprintf(struct perf_session *self, FILE *fp) } struct symbol *map_groups__find_symbol(struct map_groups *self, + struct perf_session *session, enum map_type type, u64 addr, symbol_filter_t filter) { struct map *map = map_groups__find(self, type, addr); if (map != NULL) - return map__find_symbol(map, map->map_ip(map, addr), filter); + return map__find_symbol(map, session, map->map_ip(map, addr), filter); return NULL; } diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index e93abf2d9cb..c206f72c888 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -8,7 +8,6 @@ struct map_groups { struct rb_root maps[MAP__NR_TYPES]; struct list_head removed_maps[MAP__NR_TYPES]; - bool use_modules; }; struct thread { @@ -49,19 +48,21 @@ static inline struct map *thread__find_map(struct thread *self, return self ? map_groups__find(&self->mg, type, addr) : NULL; } -void thread__find_addr_location(struct thread *self, u8 cpumode, +void thread__find_addr_location(struct thread *self, + struct perf_session *session, u8 cpumode, enum map_type type, u64 addr, struct addr_location *al, symbol_filter_t filter); struct symbol *map_groups__find_symbol(struct map_groups *self, + struct perf_session *session, enum map_type type, u64 addr, symbol_filter_t filter); static inline struct symbol * -map_groups__find_function(struct map_groups *self, u64 addr, - symbol_filter_t filter) +map_groups__find_function(struct map_groups *self, struct perf_session *session, + u64 addr, symbol_filter_t filter) { - return map_groups__find_symbol(self, MAP__FUNCTION, addr, filter); + return map_groups__find_symbol(self, session, MAP__FUNCTION, addr, filter); } struct map *map_groups__find_by_name(struct map_groups *self, -- cgit From b9bf089212d95746ce66482bcdbc7e77a0651088 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 11:37:11 -0200 Subject: perf tools: No need for three rb_trees for sorting hist entries MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit All hist entries are in only one of them, so use just one and a temporary rb_root while sorting/collapsing. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260797831-11220-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 2 +- tools/perf/builtin-report.c | 2 +- tools/perf/util/hist.c | 36 ++++++++++++++++++++---------------- tools/perf/util/hist.h | 10 ---------- 4 files changed, 22 insertions(+), 28 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index e44c54c79be..f25e89e9c9b 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -432,7 +432,7 @@ static void find_annotations(void) { struct rb_node *nd; - for (nd = rb_first(&output_hists); nd; nd = rb_next(nd)) { + for (nd = rb_first(&hist); nd; nd = rb_next(nd)) { struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node); struct sym_priv *priv; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 3487224670a..9cbdcbc4cd5 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -567,7 +567,7 @@ static size_t output__fprintf(FILE *fp, u64 total_samples) fprintf(fp, "#\n"); print_entries: - for (nd = rb_first(&output_hists); nd; nd = rb_next(nd)) { + for (nd = rb_first(&hist); nd; nd = rb_next(nd)) { pos = rb_entry(nd, struct hist_entry, rb_node); ret += hist_entry__fprintf(fp, pos, total_samples); } diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 0ebf6ee16ca..b40e37ded4b 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -1,8 +1,6 @@ #include "hist.h" struct rb_root hist; -struct rb_root collapse_hists; -struct rb_root output_hists; int callchain; struct callchain_param callchain_param = { @@ -102,9 +100,9 @@ void hist_entry__free(struct hist_entry *he) * collapse the histogram */ -void collapse__insert_entry(struct hist_entry *he) +static void collapse__insert_entry(struct rb_root *root, struct hist_entry *he) { - struct rb_node **p = &collapse_hists.rb_node; + struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; struct hist_entry *iter; int64_t cmp; @@ -128,34 +126,40 @@ void collapse__insert_entry(struct hist_entry *he) } rb_link_node(&he->rb_node, parent, p); - rb_insert_color(&he->rb_node, &collapse_hists); + rb_insert_color(&he->rb_node, root); } void collapse__resort(void) { + struct rb_root tmp; struct rb_node *next; struct hist_entry *n; if (!sort__need_collapse) return; + tmp = RB_ROOT; next = rb_first(&hist); + while (next) { n = rb_entry(next, struct hist_entry, rb_node); next = rb_next(&n->rb_node); rb_erase(&n->rb_node, &hist); - collapse__insert_entry(n); + collapse__insert_entry(&tmp, n); } + + hist = tmp; } /* * reverse the map, sort on count. */ -void output__insert_entry(struct hist_entry *he, u64 min_callchain_hits) +static void output__insert_entry(struct rb_root *root, struct hist_entry *he, + u64 min_callchain_hits) { - struct rb_node **p = &output_hists.rb_node; + struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; struct hist_entry *iter; @@ -174,29 +178,29 @@ void output__insert_entry(struct hist_entry *he, u64 min_callchain_hits) } rb_link_node(&he->rb_node, parent, p); - rb_insert_color(&he->rb_node, &output_hists); + rb_insert_color(&he->rb_node, root); } void output__resort(u64 total_samples) { + struct rb_root tmp; struct rb_node *next; struct hist_entry *n; - struct rb_root *tree = &hist; u64 min_callchain_hits; min_callchain_hits = total_samples * (callchain_param.min_percent / 100); - if (sort__need_collapse) - tree = &collapse_hists; - - next = rb_first(tree); + tmp = RB_ROOT; + next = rb_first(&hist); while (next) { n = rb_entry(next, struct hist_entry, rb_node); next = rb_next(&n->rb_node); - rb_erase(&n->rb_node, tree); - output__insert_entry(n, min_callchain_hits); + rb_erase(&n->rb_node, &hist); + output__insert_entry(&tmp, n, min_callchain_hits); } + + hist = tmp; } diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index 3020db0c929..a6cb1485e3b 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -25,16 +25,8 @@ #include "sort.h" extern struct rb_root hist; -extern struct rb_root collapse_hists; -extern struct rb_root output_hists; extern int callchain; extern struct callchain_param callchain_param; -extern unsigned long total; -extern unsigned long total_mmap; -extern unsigned long total_comm; -extern unsigned long total_fork; -extern unsigned long total_unknown; -extern unsigned long total_lost; struct hist_entry *__hist_entry__add(struct addr_location *al, struct symbol *parent, @@ -42,9 +34,7 @@ struct hist_entry *__hist_entry__add(struct addr_location *al, extern int64_t hist_entry__cmp(struct hist_entry *, struct hist_entry *); extern int64_t hist_entry__collapse(struct hist_entry *, struct hist_entry *); extern void hist_entry__free(struct hist_entry *); -extern void collapse__insert_entry(struct hist_entry *); extern void collapse__resort(void); -extern void output__insert_entry(struct hist_entry *, u64); extern void output__resort(u64); #endif /* __PERF_HIST_H */ -- cgit From 4e4f06e4c8f17ea96f7dd76251cab99511026401 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 13:10:39 -0200 Subject: perf session: Move the hist_entries rb tree to perf_session MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As we'll need to sort multiple times for multiple perf sessions, so that we can then do a diff. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260803439-16783-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 19 ++++++++------- tools/perf/builtin-kmem.c | 2 +- tools/perf/builtin-report.c | 55 +++++++++++++++++++++++------------------- tools/perf/builtin-sched.c | 2 +- tools/perf/builtin-timechart.c | 2 +- tools/perf/builtin-trace.c | 2 +- tools/perf/util/data_map.c | 2 +- tools/perf/util/hist.c | 43 ++++++++++++++++++--------------- tools/perf/util/hist.h | 43 +++++++++++---------------------- tools/perf/util/session.h | 5 +++- 10 files changed, 86 insertions(+), 89 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index f25e89e9c9b..1f1341f1dd8 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -121,10 +121,12 @@ static void hist_hit(struct hist_entry *he, u64 ip) h->ip[offset]); } -static int hist_entry__add(struct addr_location *al, u64 count) +static int perf_session__add_hist_entry(struct perf_session *self, + struct addr_location *al, u64 count) { bool hit; - struct hist_entry *he = __hist_entry__add(al, NULL, count, &hit); + struct hist_entry *he = __perf_session__add_hist_entry(self, al, NULL, + count, &hit); if (he == NULL) return -ENOMEM; hist_hit(he, al->addr); @@ -144,7 +146,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) return -1; } - if (hist_entry__add(&al, 1)) { + if (perf_session__add_hist_entry(session, &al, 1)) { fprintf(stderr, "problem incrementing symbol count, " "skipping event\n"); return -1; @@ -428,11 +430,11 @@ static void annotate_sym(struct hist_entry *he) free_source_line(he, len); } -static void find_annotations(void) +static void perf_session__find_annotations(struct perf_session *self) { struct rb_node *nd; - for (nd = rb_first(&hist); nd; nd = rb_next(nd)) { + for (nd = rb_first(&self->hists); nd; nd = rb_next(nd)) { struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node); struct sym_priv *priv; @@ -484,10 +486,9 @@ static int __cmd_annotate(void) if (verbose > 2) dsos__fprintf(stdout); - collapse__resort(); - output__resort(event__total[0]); - - find_annotations(); + perf_session__collapse_resort(session); + perf_session__output_resort(session, event__total[0]); + perf_session__find_annotations(session); out_delete: perf_session__delete(session); diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index e79ecbc1718..3b329c66b50 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -344,7 +344,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) return 0; } -static int sample_type_check(u64 type) +static int sample_type_check(u64 type, struct perf_session *session __used) { sample_type = type; diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 9cbdcbc4cd5..854427f0e57 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -38,6 +38,7 @@ static char *dso_list_str, *comm_list_str, *sym_list_str, static struct strlist *dso_list, *comm_list, *sym_list; static int force; +static bool use_callchain; static int show_nr_samples; @@ -312,8 +313,9 @@ hist_entry_callchain__fprintf(FILE *fp, struct hist_entry *self, return ret; } -static size_t -hist_entry__fprintf(FILE *fp, struct hist_entry *self, u64 total_samples) +static size_t hist_entry__fprintf(FILE *fp, struct hist_entry *self, + struct perf_session *session, + u64 total_samples) { struct sort_entry *se; size_t ret; @@ -345,7 +347,7 @@ hist_entry__fprintf(FILE *fp, struct hist_entry *self, u64 total_samples) ret += fprintf(fp, "\n"); - if (callchain) { + if (session->use_callchain) { int left_margin = 0; if (sort__first_dimension == SORT_COMM) { @@ -422,7 +424,7 @@ static struct symbol **resolve_callchain(struct thread *thread, struct symbol **syms = NULL; unsigned int i; - if (callchain) { + if (session->use_callchain) { syms = calloc(chain->nr, sizeof(*syms)); if (!syms) { fprintf(stderr, "Can't allocate memory for symbols\n"); @@ -454,7 +456,7 @@ static struct symbol **resolve_callchain(struct thread *thread, if (sort__has_parent && !*parent && call__match(al.sym)) *parent = al.sym; - if (!callchain) + if (!session->use_callchain) break; syms[i] = al.sym; } @@ -467,25 +469,25 @@ static struct symbol **resolve_callchain(struct thread *thread, * collect histogram counts */ -static int hist_entry__add(struct addr_location *al, - struct perf_session *session, - struct ip_callchain *chain, u64 count) +static int perf_session__add_hist_entry(struct perf_session *self, + struct addr_location *al, + struct ip_callchain *chain, u64 count) { struct symbol **syms = NULL, *parent = NULL; bool hit; struct hist_entry *he; - if ((sort__has_parent || callchain) && chain) - syms = resolve_callchain(al->thread, session, chain, &parent); + if ((sort__has_parent || self->use_callchain) && chain) + syms = resolve_callchain(al->thread, self, chain, &parent); - he = __hist_entry__add(al, parent, count, &hit); + he = __perf_session__add_hist_entry(self, al, parent, count, &hit); if (he == NULL) return -ENOMEM; if (hit) he->count += count; - if (callchain) { + if (self->use_callchain) { if (!hit) callchain_init(&he->callchain); append_chain(&he->callchain, chain, syms); @@ -495,7 +497,8 @@ static int hist_entry__add(struct addr_location *al, return 0; } -static size_t output__fprintf(FILE *fp, u64 total_samples) +static size_t perf_session__fprintf_hist_entries(struct perf_session *self, + u64 total_samples, FILE *fp) { struct hist_entry *pos; struct sort_entry *se; @@ -567,9 +570,9 @@ static size_t output__fprintf(FILE *fp, u64 total_samples) fprintf(fp, "#\n"); print_entries: - for (nd = rb_first(&hist); nd; nd = rb_next(nd)) { + for (nd = rb_first(&self->hists); nd; nd = rb_next(nd)) { pos = rb_entry(nd, struct hist_entry, rb_node); - ret += hist_entry__fprintf(fp, pos, total_samples); + ret += hist_entry__fprintf(fp, pos, self, total_samples); } if (sort_order == default_sort_order && @@ -671,7 +674,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) if (sym_list && al.sym && !strlist__has_entry(sym_list, al.sym->name)) return 0; - if (hist_entry__add(&al, session, data.callchain, data.period)) { + if (perf_session__add_hist_entry(session, &al, data.callchain, data.period)) { pr_debug("problem incrementing symbol count, skipping event\n"); return -1; } @@ -719,7 +722,7 @@ static int process_read_event(event_t *event, struct perf_session *session __use return 0; } -static int sample_type_check(u64 type) +static int sample_type_check(u64 type, struct perf_session *session) { sample_type = type; @@ -730,14 +733,14 @@ static int sample_type_check(u64 type) " perf record without -g?\n"); return -1; } - if (callchain) { + if (session->use_callchain) { fprintf(stderr, "selected -g but no callchain data." " Did you call perf record without" " -g?\n"); return -1; } - } else if (callchain_param.mode != CHAIN_NONE && !callchain) { - callchain = 1; + } else if (callchain_param.mode != CHAIN_NONE && !session->use_callchain) { + session->use_callchain = true; if (register_callchain_param(&callchain_param) < 0) { fprintf(stderr, "Can't register callchain" " params\n"); @@ -769,6 +772,8 @@ static int __cmd_report(void) if (session == NULL) return -ENOMEM; + session->use_callchain = use_callchain; + if (show_threads) perf_read_values_init(&show_threads_values); @@ -787,9 +792,9 @@ static int __cmd_report(void) if (verbose > 2) dsos__fprintf(stdout); - collapse__resort(); - output__resort(event__stats.total); - output__fprintf(stdout, event__stats.total); + perf_session__collapse_resort(session); + perf_session__output_resort(session, event__stats.total); + perf_session__fprintf_hist_entries(session, event__stats.total, stdout); if (show_threads) perf_read_values_destroy(&show_threads_values); @@ -805,7 +810,7 @@ parse_callchain_opt(const struct option *opt __used, const char *arg, char *tok; char *endptr; - callchain = 1; + use_callchain = true; if (!arg) return 0; @@ -826,7 +831,7 @@ parse_callchain_opt(const struct option *opt __used, const char *arg, else if (!strncmp(tok, "none", strlen(arg))) { callchain_param.mode = CHAIN_NONE; - callchain = 0; + use_callchain = true; return 0; } diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index bce05dff6df..412ae924640 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -1655,7 +1655,7 @@ static int process_lost_event(event_t *event __used, return 0; } -static int sample_type_check(u64 type) +static int sample_type_check(u64 type, struct perf_session *session __used) { sample_type = type; diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 27018531404..3a0a89e4152 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -1033,7 +1033,7 @@ static void process_samples(void) } } -static int sample_type_check(u64 type) +static int sample_type_check(u64 type, struct perf_session *session __used) { sample_type = type; diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 9d89ae423ca..adce442dd60 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -103,7 +103,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) return 0; } -static int sample_type_check(u64 type) +static int sample_type_check(u64 type, struct perf_session *session __used) { sample_type = type; diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index 44dea211cc6..08c4cf5e66b 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -161,7 +161,7 @@ int perf_session__process_events(struct perf_session *self, err = -EINVAL; if (ops->sample_type_check && - ops->sample_type_check(sample_type) < 0) + ops->sample_type_check(sample_type, self) < 0) goto out_err; if (!ops->full_paths) { diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index b40e37ded4b..b9828fce7bf 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -1,7 +1,6 @@ #include "hist.h" - -struct rb_root hist; -int callchain; +#include "session.h" +#include "sort.h" struct callchain_param callchain_param = { .mode = CHAIN_GRAPH_REL, @@ -12,11 +11,12 @@ struct callchain_param callchain_param = { * histogram, sorted on item, collects counts */ -struct hist_entry *__hist_entry__add(struct addr_location *al, - struct symbol *sym_parent, - u64 count, bool *hit) +struct hist_entry *__perf_session__add_hist_entry(struct perf_session *self, + struct addr_location *al, + struct symbol *sym_parent, + u64 count, bool *hit) { - struct rb_node **p = &hist.rb_node; + struct rb_node **p = &self->hists.rb_node; struct rb_node *parent = NULL; struct hist_entry *he; struct hist_entry entry = { @@ -52,7 +52,7 @@ struct hist_entry *__hist_entry__add(struct addr_location *al, return NULL; *he = entry; rb_link_node(&he->rb_node, parent, p); - rb_insert_color(&he->rb_node, &hist); + rb_insert_color(&he->rb_node, &self->hists); *hit = false; return he; } @@ -129,7 +129,7 @@ static void collapse__insert_entry(struct rb_root *root, struct hist_entry *he) rb_insert_color(&he->rb_node, root); } -void collapse__resort(void) +void perf_session__collapse_resort(struct perf_session *self) { struct rb_root tmp; struct rb_node *next; @@ -139,31 +139,33 @@ void collapse__resort(void) return; tmp = RB_ROOT; - next = rb_first(&hist); + next = rb_first(&self->hists); while (next) { n = rb_entry(next, struct hist_entry, rb_node); next = rb_next(&n->rb_node); - rb_erase(&n->rb_node, &hist); + rb_erase(&n->rb_node, &self->hists); collapse__insert_entry(&tmp, n); } - hist = tmp; + self->hists = tmp; } /* * reverse the map, sort on count. */ -static void output__insert_entry(struct rb_root *root, struct hist_entry *he, - u64 min_callchain_hits) +static void perf_session__insert_output_hist_entry(struct perf_session *self, + struct rb_root *root, + struct hist_entry *he, + u64 min_callchain_hits) { struct rb_node **p = &root->rb_node; struct rb_node *parent = NULL; struct hist_entry *iter; - if (callchain) + if (self->use_callchain) callchain_param.sort(&he->sorted_chain, &he->callchain, min_callchain_hits, &callchain_param); @@ -181,7 +183,7 @@ static void output__insert_entry(struct rb_root *root, struct hist_entry *he, rb_insert_color(&he->rb_node, root); } -void output__resort(u64 total_samples) +void perf_session__output_resort(struct perf_session *self, u64 total_samples) { struct rb_root tmp; struct rb_node *next; @@ -192,15 +194,16 @@ void output__resort(u64 total_samples) total_samples * (callchain_param.min_percent / 100); tmp = RB_ROOT; - next = rb_first(&hist); + next = rb_first(&self->hists); while (next) { n = rb_entry(next, struct hist_entry, rb_node); next = rb_next(&n->rb_node); - rb_erase(&n->rb_node, &hist); - output__insert_entry(&tmp, n, min_callchain_hits); + rb_erase(&n->rb_node, &self->hists); + perf_session__insert_output_hist_entry(self, &tmp, n, + min_callchain_hits); } - hist = tmp; + self->hists = tmp; } diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index a6cb1485e3b..7efdb1b6d8c 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -1,40 +1,25 @@ #ifndef __PERF_HIST_H #define __PERF_HIST_H -#include "../builtin.h" -#include "util.h" - -#include "color.h" -#include -#include "cache.h" -#include -#include "symbol.h" -#include "string.h" +#include #include "callchain.h" -#include "strlist.h" -#include "values.h" - -#include "../perf.h" -#include "debug.h" -#include "header.h" - -#include "parse-options.h" -#include "parse-events.h" -#include "thread.h" -#include "sort.h" - -extern struct rb_root hist; -extern int callchain; extern struct callchain_param callchain_param; -struct hist_entry *__hist_entry__add(struct addr_location *al, - struct symbol *parent, - u64 count, bool *hit); +struct perf_session; +struct hist_entry; +struct addr_location; +struct symbol; + +struct hist_entry *__perf_session__add_hist_entry(struct perf_session *self, + struct addr_location *al, + struct symbol *parent, + u64 count, bool *hit); extern int64_t hist_entry__cmp(struct hist_entry *, struct hist_entry *); extern int64_t hist_entry__collapse(struct hist_entry *, struct hist_entry *); -extern void hist_entry__free(struct hist_entry *); -extern void collapse__resort(void); -extern void output__resort(u64); +void hist_entry__free(struct hist_entry *); + +void perf_session__output_resort(struct perf_session *self, u64 total_samples); +void perf_session__collapse_resort(struct perf_session *self); #endif /* __PERF_HIST_H */ diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 20b2c9cc834..759d96022a3 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -16,10 +16,12 @@ struct perf_session { struct map_groups kmaps; struct rb_root threads; struct thread *last_match; + struct rb_root hists; int fd; int cwdlen; char *cwd; bool use_modules; + bool use_callchain; char filename[0]; }; @@ -35,7 +37,8 @@ struct perf_event_ops { event_op process_read_event; event_op process_throttle_event; event_op process_unthrottle_event; - int (*sample_type_check)(u64 sample_type); + int (*sample_type_check)(u64 sample_type, + struct perf_session *session); unsigned long total_unknown; bool full_paths; }; -- cgit From a328626b61aeda1a7d00a80c475c76ca1b815e0d Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 14:22:59 -0200 Subject: perf session: Adopt resolve_callchain MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is really a generic library routine, so declutter builtin-report.c a bit by moving it to the library. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260807780-19377-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-report.c | 62 ++------------------------------------------- tools/perf/util/session.c | 59 ++++++++++++++++++++++++++++++++++++++++++ tools/perf/util/session.h | 7 +++++ 3 files changed, 68 insertions(+), 60 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 854427f0e57..5141cdccbb6 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -407,64 +407,6 @@ static int thread__set_comm_adjust(struct thread *self, const char *comm) return 0; } -static int call__match(struct symbol *sym) -{ - if (sym->name && !regexec(&parent_regex, sym->name, 0, NULL, 0)) - return 1; - - return 0; -} - -static struct symbol **resolve_callchain(struct thread *thread, - struct perf_session *session, - struct ip_callchain *chain, - struct symbol **parent) -{ - u8 cpumode = PERF_RECORD_MISC_USER; - struct symbol **syms = NULL; - unsigned int i; - - if (session->use_callchain) { - syms = calloc(chain->nr, sizeof(*syms)); - if (!syms) { - fprintf(stderr, "Can't allocate memory for symbols\n"); - exit(-1); - } - } - - for (i = 0; i < chain->nr; i++) { - u64 ip = chain->ips[i]; - struct addr_location al; - - if (ip >= PERF_CONTEXT_MAX) { - switch (ip) { - case PERF_CONTEXT_HV: - cpumode = PERF_RECORD_MISC_HYPERVISOR; break; - case PERF_CONTEXT_KERNEL: - cpumode = PERF_RECORD_MISC_KERNEL; break; - case PERF_CONTEXT_USER: - cpumode = PERF_RECORD_MISC_USER; break; - default: - break; - } - continue; - } - - thread__find_addr_location(thread, session, cpumode, - MAP__FUNCTION, ip, &al, NULL); - if (al.sym != NULL) { - if (sort__has_parent && !*parent && - call__match(al.sym)) - *parent = al.sym; - if (!session->use_callchain) - break; - syms[i] = al.sym; - } - } - - return syms; -} - /* * collect histogram counts */ @@ -478,8 +420,8 @@ static int perf_session__add_hist_entry(struct perf_session *self, struct hist_entry *he; if ((sort__has_parent || self->use_callchain) && chain) - syms = resolve_callchain(al->thread, self, chain, &parent); - + syms = perf_session__resolve_callchain(self, al->thread, + chain, &parent); he = __perf_session__add_hist_entry(self, al, parent, count, &hit); if (he == NULL) return -ENOMEM; diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index fe87a2f2e5a..ecd54bedfb1 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -4,6 +4,7 @@ #include #include "session.h" +#include "sort.h" #include "util.h" static int perf_session__open(struct perf_session *self, bool force) @@ -90,3 +91,61 @@ void perf_session__delete(struct perf_session *self) free(self->cwd); free(self); } + +static bool symbol__match_parent_regex(struct symbol *sym) +{ + if (sym->name && !regexec(&parent_regex, sym->name, 0, NULL, 0)) + return 1; + + return 0; +} + +struct symbol **perf_session__resolve_callchain(struct perf_session *self, + struct thread *thread, + struct ip_callchain *chain, + struct symbol **parent) +{ + u8 cpumode = PERF_RECORD_MISC_USER; + struct symbol **syms = NULL; + unsigned int i; + + if (self->use_callchain) { + syms = calloc(chain->nr, sizeof(*syms)); + if (!syms) { + fprintf(stderr, "Can't allocate memory for symbols\n"); + exit(-1); + } + } + + for (i = 0; i < chain->nr; i++) { + u64 ip = chain->ips[i]; + struct addr_location al; + + if (ip >= PERF_CONTEXT_MAX) { + switch (ip) { + case PERF_CONTEXT_HV: + cpumode = PERF_RECORD_MISC_HYPERVISOR; break; + case PERF_CONTEXT_KERNEL: + cpumode = PERF_RECORD_MISC_KERNEL; break; + case PERF_CONTEXT_USER: + cpumode = PERF_RECORD_MISC_USER; break; + default: + break; + } + continue; + } + + thread__find_addr_location(thread, self, cpumode, + MAP__FUNCTION, ip, &al, NULL); + if (al.sym != NULL) { + if (sort__has_parent && !*parent && + symbol__match_parent_regex(al.sym)) + *parent = al.sym; + if (!self->use_callchain) + break; + syms[i] = al.sym; + } + } + + return syms; +} diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 759d96022a3..a8f3a49ca43 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -6,7 +6,9 @@ #include "thread.h" #include +struct ip_callchain; struct thread; +struct symbol; struct symbol_conf; struct perf_session { @@ -50,6 +52,11 @@ void perf_session__delete(struct perf_session *self); int perf_session__process_events(struct perf_session *self, struct perf_event_ops *event_ops); +struct symbol **perf_session__resolve_callchain(struct perf_session *self, + struct thread *thread, + struct ip_callchain *chain, + struct symbol **parent); + int perf_header__read_build_ids(int input, u64 offset, u64 file_size); #endif /* __PERF_SESSION_H */ -- cgit From c019879bcc5692ec9267c1cedad91f1794d0b693 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 14:23:00 -0200 Subject: perf session: Adopt the sample_type variable MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit All tools had copies, and perf diff would have to specify a sample_type_check method just for copying it. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260807780-19377-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-kmem.c | 10 +++------- tools/perf/builtin-report.c | 12 ++++-------- tools/perf/builtin-sched.c | 12 ++++-------- tools/perf/builtin-timechart.c | 36 +++++++++++++++--------------------- tools/perf/builtin-trace.c | 12 ++++-------- tools/perf/util/data_map.c | 6 ++---- tools/perf/util/session.h | 4 ++-- 7 files changed, 34 insertions(+), 58 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c index 3b329c66b50..dda60869faa 100644 --- a/tools/perf/builtin-kmem.c +++ b/tools/perf/builtin-kmem.c @@ -20,8 +20,6 @@ typedef int (*sort_fn_t)(struct alloc_stat *, struct alloc_stat *); static char const *input_name = "perf.data"; -static u64 sample_type; - static int alloc_flag; static int caller_flag; @@ -321,7 +319,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) data.cpu = -1; data.period = 1; - event__parse_sample(event, sample_type, &data); + event__parse_sample(event, session->sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, @@ -344,11 +342,9 @@ static int process_sample_event(event_t *event, struct perf_session *session) return 0; } -static int sample_type_check(u64 type, struct perf_session *session __used) +static int sample_type_check(struct perf_session *session) { - sample_type = type; - - if (!(sample_type & PERF_SAMPLE_RAW)) { + if (!(session->sample_type & PERF_SAMPLE_RAW)) { fprintf(stderr, "No trace sample to read. Did you call perf record " "without -R?"); diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 5141cdccbb6..142c475f991 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -52,8 +52,6 @@ static int exclude_other = 1; static char callchain_default_opt[] = "fractal,0.5"; -static u64 sample_type; - struct symbol_conf symbol_conf; @@ -557,7 +555,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) memset(&data, 0, sizeof(data)); data.period = 1; - event__parse_sample(event, sample_type, &data); + event__parse_sample(event, session->sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, @@ -565,7 +563,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) (void *)(long)data.ip, (long long)data.period); - if (sample_type & PERF_SAMPLE_CALLCHAIN) { + if (session->sample_type & PERF_SAMPLE_CALLCHAIN) { unsigned int i; dump_printf("... chain: nr:%Lu\n", data.callchain->nr); @@ -664,11 +662,9 @@ static int process_read_event(event_t *event, struct perf_session *session __use return 0; } -static int sample_type_check(u64 type, struct perf_session *session) +static int sample_type_check(struct perf_session *session) { - sample_type = type; - - if (!(sample_type & PERF_SAMPLE_CALLCHAIN)) { + if (!(session->sample_type & PERF_SAMPLE_CALLCHAIN)) { if (sort__has_parent) { fprintf(stderr, "selected --sort parent, but no" " callchain data. Did you call" diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c index 412ae924640..d67f274adba 100644 --- a/tools/perf/builtin-sched.c +++ b/tools/perf/builtin-sched.c @@ -21,8 +21,6 @@ static char const *input_name = "perf.data"; -static u64 sample_type; - static char default_sort_order[] = "avg, max, switch, runtime"; static char *sort_order = default_sort_order; @@ -1613,7 +1611,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) struct sample_data data; struct thread *thread; - if (!(sample_type & PERF_SAMPLE_RAW)) + if (!(session->sample_type & PERF_SAMPLE_RAW)) return 0; memset(&data, 0, sizeof(data)); @@ -1621,7 +1619,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) data.cpu = -1; data.period = -1; - event__parse_sample(event, sample_type, &data); + event__parse_sample(event, session->sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, @@ -1655,11 +1653,9 @@ static int process_lost_event(event_t *event __used, return 0; } -static int sample_type_check(u64 type, struct perf_session *session __used) +static int sample_type_check(struct perf_session *session __used) { - sample_type = type; - - if (!(sample_type & PERF_SAMPLE_RAW)) { + if (!(session->sample_type & PERF_SAMPLE_RAW)) { fprintf(stderr, "No trace sample to read. Did you call perf record " "without -R?"); diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c index 3a0a89e4152..ffd81e87ce6 100644 --- a/tools/perf/builtin-timechart.c +++ b/tools/perf/builtin-timechart.c @@ -36,9 +36,6 @@ static char const *input_name = "perf.data"; static char const *output_name = "output.svg"; - -static u64 sample_type; - static unsigned int numcpus; static u64 min_freq; /* Lowest CPU frequency seen */ static u64 max_freq; /* Highest CPU frequency seen */ @@ -478,17 +475,16 @@ static void sched_switch(int cpu, u64 timestamp, struct trace_entry *te) } -static int -process_sample_event(event_t *event) +static int process_sample_event(event_t *event, struct perf_session *session) { struct sample_data data; struct trace_entry *te; memset(&data, 0, sizeof(data)); - event__parse_sample(event, sample_type, &data); + event__parse_sample(event, session->sample_type, &data); - if (sample_type & PERF_SAMPLE_TIME) { + if (session->sample_type & PERF_SAMPLE_TIME) { if (!first_time || first_time > data.time) first_time = data.time; if (last_time < data.time) @@ -496,7 +492,7 @@ process_sample_event(event_t *event) } te = (void *)data.raw_data; - if (sample_type & PERF_SAMPLE_RAW && data.raw_size > 0) { + if (session->sample_type & PERF_SAMPLE_RAW && data.raw_size > 0) { char *event_str; struct power_entry *pe; @@ -573,16 +569,16 @@ static void end_sample_processing(void) } } -static u64 sample_time(event_t *event) +static u64 sample_time(event_t *event, const struct perf_session *session) { int cursor; cursor = 0; - if (sample_type & PERF_SAMPLE_IP) + if (session->sample_type & PERF_SAMPLE_IP) cursor++; - if (sample_type & PERF_SAMPLE_TID) + if (session->sample_type & PERF_SAMPLE_TID) cursor++; - if (sample_type & PERF_SAMPLE_TIME) + if (session->sample_type & PERF_SAMPLE_TIME) return event->sample.array[cursor]; return 0; } @@ -592,7 +588,7 @@ static u64 sample_time(event_t *event) * We first queue all events, sorted backwards by insertion. * The order will get flipped later. */ -static int queue_sample_event(event_t *event, struct perf_session *session __used) +static int queue_sample_event(event_t *event, struct perf_session *session) { struct sample_wrapper *copy, *prev; int size; @@ -606,7 +602,7 @@ static int queue_sample_event(event_t *event, struct perf_session *session __use memset(copy, 0, size); copy->next = NULL; - copy->timestamp = sample_time(event); + copy->timestamp = sample_time(event, session); memcpy(©->data, event, event->sample.header.size); @@ -1018,7 +1014,7 @@ static void write_svg_file(const char *filename) svg_close(); } -static void process_samples(void) +static void process_samples(struct perf_session *session) { struct sample_wrapper *cursor; event_t *event; @@ -1029,15 +1025,13 @@ static void process_samples(void) while (cursor) { event = (void *)&cursor->data; cursor = cursor->next; - process_sample_event(event); + process_sample_event(event, session); } } -static int sample_type_check(u64 type, struct perf_session *session __used) +static int sample_type_check(struct perf_session *session) { - sample_type = type; - - if (!(sample_type & PERF_SAMPLE_RAW)) { + if (!(session->sample_type & PERF_SAMPLE_RAW)) { fprintf(stderr, "No trace samples found in the file.\n" "Have you used 'perf timechart record' to record it?\n"); return -1; @@ -1067,7 +1061,7 @@ static int __cmd_timechart(void) if (ret) goto out_delete; - process_samples(); + process_samples(session); end_sample_processing(); diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index adce442dd60..9ee976dc395 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -61,8 +61,6 @@ static int cleanup_scripting(void) static char const *input_name = "perf.data"; -static u64 sample_type; - static int process_sample_event(event_t *event, struct perf_session *session) { struct sample_data data; @@ -73,7 +71,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) data.cpu = -1; data.period = 1; - event__parse_sample(event, sample_type, &data); + event__parse_sample(event, session->sample_type, &data); dump_printf("(IP, %d): %d/%d: %p period: %Ld\n", event->header.misc, @@ -88,7 +86,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) return -1; } - if (sample_type & PERF_SAMPLE_RAW) { + if (session->sample_type & PERF_SAMPLE_RAW) { /* * FIXME: better resolve from pid from the struct trace_entry * field, although it should be the same than this perf @@ -103,11 +101,9 @@ static int process_sample_event(event_t *event, struct perf_session *session) return 0; } -static int sample_type_check(u64 type, struct perf_session *session __used) +static int sample_type_check(struct perf_session *session) { - sample_type = type; - - if (!(sample_type & PERF_SAMPLE_RAW)) { + if (!(session->sample_type & PERF_SAMPLE_RAW)) { fprintf(stderr, "No trace sample to read. Did you call perf record " "without -R?"); diff --git a/tools/perf/util/data_map.c b/tools/perf/util/data_map.c index 08c4cf5e66b..b557b836de3 100644 --- a/tools/perf/util/data_map.c +++ b/tools/perf/util/data_map.c @@ -144,7 +144,6 @@ int perf_session__process_events(struct perf_session *self, unsigned long head, shift; unsigned long offset = 0; size_t page_size; - u64 sample_type; event_t *event; uint32_t size; char *buf; @@ -157,11 +156,10 @@ int perf_session__process_events(struct perf_session *self, page_size = getpagesize(); head = self->header.data_offset; - sample_type = perf_header__sample_type(&self->header); + self->sample_type = perf_header__sample_type(&self->header); err = -EINVAL; - if (ops->sample_type_check && - ops->sample_type_check(sample_type, self) < 0) + if (ops->sample_type_check && ops->sample_type_check(self) < 0) goto out_err; if (!ops->full_paths) { diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index a8f3a49ca43..4e8a21c5304 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -19,6 +19,7 @@ struct perf_session { struct rb_root threads; struct thread *last_match; struct rb_root hists; + u64 sample_type; int fd; int cwdlen; char *cwd; @@ -39,8 +40,7 @@ struct perf_event_ops { event_op process_read_event; event_op process_throttle_event; event_op process_unthrottle_event; - int (*sample_type_check)(u64 sample_type, - struct perf_session *session); + int (*sample_type_check)(struct perf_session *session); unsigned long total_unknown; bool full_paths; }; -- cgit From f823e441ab4dfaeaf17832fa1931e0dc0fde304d Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 15:06:01 -0200 Subject: perf session: Event statistics also are per session MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260810361-22828-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 2 +- tools/perf/builtin-report.c | 7 +++---- tools/perf/builtin-trace.c | 2 +- tools/perf/util/event.c | 6 ++---- tools/perf/util/event.h | 5 ----- tools/perf/util/session.h | 3 +++ 6 files changed, 10 insertions(+), 15 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 1f1341f1dd8..fa833f50763 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -487,7 +487,7 @@ static int __cmd_annotate(void) dsos__fprintf(stdout); perf_session__collapse_resort(session); - perf_session__output_resort(session, event__total[0]); + perf_session__output_resort(session, session->event_total[0]); perf_session__find_annotations(session); out_delete: perf_session__delete(session); diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index 142c475f991..d12ea4ab0d2 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -619,8 +619,7 @@ static int process_sample_event(event_t *event, struct perf_session *session) return -1; } - event__stats.total += data.period; - + session->events_stats.total += data.period; return 0; } @@ -731,8 +730,8 @@ static int __cmd_report(void) dsos__fprintf(stdout); perf_session__collapse_resort(session); - perf_session__output_resort(session, event__stats.total); - perf_session__fprintf_hist_entries(session, event__stats.total, stdout); + perf_session__output_resort(session, session->events_stats.total); + perf_session__fprintf_hist_entries(session, session->events_stats.total, stdout); if (show_threads) perf_read_values_destroy(&show_threads_values); diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 9ee976dc395..07da66511bd 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -96,8 +96,8 @@ static int process_sample_event(event_t *event, struct perf_session *session) data.raw_size, data.time, thread->comm); } - event__stats.total += data.period; + session->events_stats.total += data.period; return 0; } diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 222efb1fc3b..375fb6dca1c 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -187,8 +187,6 @@ void event__synthesize_threads(int (*process)(event_t *event, closedir(proc); } -struct events_stats event__stats; - int event__process_comm(event_t *self, struct perf_session *session) { struct thread *thread = perf_session__findnew(session, self->comm.pid); @@ -203,10 +201,10 @@ int event__process_comm(event_t *self, struct perf_session *session) return 0; } -int event__process_lost(event_t *self, struct perf_session *session __used) +int event__process_lost(event_t *self, struct perf_session *session) { dump_printf(": id:%Ld: lost:%Ld\n", self->lost.id, self->lost.lost); - event__stats.lost += self->lost.lost; + session->events_stats.lost += self->lost.lost; return 0; } diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h index 035ecf3c25c..a92e0b039a6 100644 --- a/tools/perf/util/event.h +++ b/tools/perf/util/event.h @@ -168,11 +168,6 @@ void event__synthesize_threads(int (*process)(event_t *event, struct perf_session *session), struct perf_session *session); -extern char *event__cwd; -extern int event__cwdlen; -extern struct events_stats event__stats; -extern unsigned long event__total[PERF_RECORD_MAX]; - int event__process_comm(event_t *self, struct perf_session *session); int event__process_lost(event_t *self, struct perf_session *session); int event__process_mmap(event_t *self, struct perf_session *session); diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 4e8a21c5304..bdfc4b8eee7 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -5,6 +5,7 @@ #include "header.h" #include "thread.h" #include +#include "../../../include/linux/perf_event.h" struct ip_callchain; struct thread; @@ -18,6 +19,8 @@ struct perf_session { struct map_groups kmaps; struct rb_root threads; struct thread *last_match; + struct events_stats events_stats; + unsigned long event_total[PERF_RECORD_MAX]; struct rb_root hists; u64 sample_type; int fd; -- cgit From c8829c7a31c7e0156b230fa8d5a14be9881d7677 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 20:09:29 -0200 Subject: perf util: Remove setup_sorting dups MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit And it is also needed by 'perf diff'. Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260828571-3613-1-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-annotate.c | 17 +---------------- tools/perf/builtin-report.c | 17 +---------------- tools/perf/util/sort.c | 15 +++++++++++++++ tools/perf/util/sort.h | 2 ++ 4 files changed, 19 insertions(+), 32 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index fa833f50763..2e2855a685c 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -521,21 +521,6 @@ static const struct option options[] = { OPT_END() }; -static void setup_sorting(void) -{ - char *tmp, *tok, *str = strdup(sort_order); - - for (tok = strtok_r(str, ", ", &tmp); - tok; tok = strtok_r(NULL, ", ", &tmp)) { - if (sort_dimension__add(tok) < 0) { - error("Unknown --sort key: `%s'", tok); - usage_with_options(annotate_usage, options); - } - } - - free(str); -} - int cmd_annotate(int argc, const char **argv, const char *prefix __used) { if (symbol__init(&symbol_conf) < 0) @@ -543,7 +528,7 @@ int cmd_annotate(int argc, const char **argv, const char *prefix __used) argc = parse_options(argc, argv, options, annotate_usage, 0); - setup_sorting(); + setup_sorting(annotate_usage, options); if (argc) { /* diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index d12ea4ab0d2..beff7c090d2 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -843,21 +843,6 @@ static const struct option options[] = { OPT_END() }; -static void setup_sorting(void) -{ - char *tmp, *tok, *str = strdup(sort_order); - - for (tok = strtok_r(str, ", ", &tmp); - tok; tok = strtok_r(NULL, ", ", &tmp)) { - if (sort_dimension__add(tok) < 0) { - error("Unknown --sort key: `%s'", tok); - usage_with_options(report_usage, options); - } - } - - free(str); -} - static void setup_list(struct strlist **list, const char *list_str, struct sort_entry *se, const char *list_name, FILE *fp) @@ -884,7 +869,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __used) argc = parse_options(argc, argv, options, report_usage, 0); - setup_sorting(); + setup_sorting(report_usage, options); if (parent_pattern != default_parent_pattern) { sort_dimension__add("parent"); diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c index b490354d1b2..cff1c316fa9 100644 --- a/tools/perf/util/sort.c +++ b/tools/perf/util/sort.c @@ -288,3 +288,18 @@ int sort_dimension__add(const char *tok) return -ESRCH; } + +void setup_sorting(const char * const usagestr[], const struct option *opts) +{ + char *tmp, *tok, *str = strdup(sort_order); + + for (tok = strtok_r(str, ", ", &tmp); + tok; tok = strtok_r(NULL, ", ", &tmp)) { + if (sort_dimension__add(tok) < 0) { + error("Unknown --sort key: `%s'", tok); + usage_with_options(usagestr, opts); + } + } + + free(str); +} diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h index 333e664ff45..cb6151c026f 100644 --- a/tools/perf/util/sort.h +++ b/tools/perf/util/sort.h @@ -81,6 +81,8 @@ struct sort_entry { extern struct sort_entry sort_thread; extern struct list_head hist_entry__sort_list; +void setup_sorting(const char * const usagestr[], const struct option *opts); + extern int repsep_fprintf(FILE *fp, const char *fmt, ...); extern size_t sort__thread_print(FILE *, struct hist_entry *, unsigned int); extern size_t sort__comm_print(FILE *, struct hist_entry *, unsigned int); -- cgit From b38d34645cc52136b6c99623fef7ded26742e573 Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 20:09:30 -0200 Subject: perf record: Rename perf.data to perf.data.old if --force/-f is used MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Suggested-by: Steven Rostedt Signed-off-by: Arnaldo Carvalho de Melo Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras LKML-Reference: <1260828571-3613-2-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/builtin-record.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index a66a58d5281..66979a5553b 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -421,10 +421,19 @@ static int __cmd_record(int argc, const char **argv) signal(SIGINT, sig_handler); if (!stat(output_name, &st) && st.st_size) { - if (!force && !append_file) { - fprintf(stderr, "Error, output file %s exists, use -A to append or -f to overwrite.\n", - output_name); - exit(-1); + if (!force) { + if (!append_file) { + pr_err("Error, output file %s exists, use -A " + "to append or -f to overwrite.\n", + output_name); + exit(-1); + } + } else { + char oldname[PATH_MAX]; + snprintf(oldname, sizeof(oldname), "%s.old", + output_name); + unlink(oldname); + rename(output_name, oldname); } } else { append_file = 0; -- cgit From 86a9eee047ba09a714c3b8e27c9df2bbf715393a Mon Sep 17 00:00:00 2001 From: Arnaldo Carvalho de Melo Date: Mon, 14 Dec 2009 20:09:31 -0200 Subject: perf diff: Introduce tool to show performance difference MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit I guess it is enough to show some examples: [root@doppio linux-2.6-tip]# rm -f perf.data* [root@doppio linux-2.6-tip]# ls -la perf.data* ls: cannot access perf.data*: No such file or directory [root@doppio linux-2.6-tip]# perf record -f find / > /dev/null [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.062 MB perf.data (~2699 samples) ] [root@doppio linux-2.6-tip]# ls -la perf.data* -rw------- 1 root root 74440 2009-12-14 20:03 perf.data [root@doppio linux-2.6-tip]# perf record -f find / > /dev/null [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.062 MB perf.data (~2692 samples) ] [root@doppio linux-2.6-tip]# ls -la perf.data* -rw------- 1 root root 74280 2009-12-14 20:03 perf.data -rw------- 1 root root 74440 2009-12-14 20:03 perf.data.old [root@doppio linux-2.6-tip]# perf diff | head -5 1 -34994580 /lib64/libc-2.10.1.so _IO_vfprintf_internal 2 -15307806 [kernel.kallsyms] __kmalloc 3 +1 +3665941 /lib64/libc-2.10.1.so __GI_memmove 4 +4 +23508995 /lib64/libc-2.10.1.so _int_malloc 5 +7 +38538813 [kernel.kallsyms] __d_lookup [root@doppio linux-2.6-tip]# perf diff -p | head -5 1 +1.00% /lib64/libc-2.10.1.so _IO_vfprintf_internal 2 [kernel.kallsyms] __kmalloc 3 +1 /lib64/libc-2.10.1.so __GI_memmove 4 +4 /lib64/libc-2.10.1.so _int_malloc 5 +7 -1.00% [kernel.kallsyms] __d_lookup [root@doppio linux-2.6-tip]# perf diff -v | head -5 1 361449551 326454971 -34994580 /lib64/libc-2.10.1.so _IO_vfprintf_internal 2 151009241 135701435 -15307806 [kernel.kallsyms] __kmalloc 3 +1 101805328 105471269 +3665941 /lib64/libc-2.10.1.so __GI_memmove 4 +4 78041440 101550435 +23508995 /lib64/libc-2.10.1.so _int_malloc 5 +7 59536172 98074985 +38538813 [kernel.kallsyms] __d_lookup [root@doppio linux-2.6-tip]# perf diff -vp | head -5 1 9.00% 8.00% +1.00% /lib64/libc-2.10.1.so _IO_vfprintf_internal 2 3.00% 3.00% [kernel.kallsyms] __kmalloc 3 +1 2.00% 2.00% /lib64/libc-2.10.1.so __GI_memmove 4 +4 2.00% 2.00% /lib64/libc-2.10.1.so _int_malloc 5 +7 1.00% 2.00% -1.00% [kernel.kallsyms] __d_lookup [root@doppio linux-2.6-tip]# This should be enough for diffs where the system is non volatile, i.e. when one doesn't updates binaries. For volatile environments, stay tuned for the next perf tool feature: a buildid cache populated by 'perf record', managed by 'perf buildid-cache' a-la ccache, and used by all the report tools. Signed-off-by: Arnaldo Carvalho de Melo Cc: "Paul E. McKenney" Cc: Stephen Hemminger Cc: Frédéric Weisbecker Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Paul Mackerras Cc: Paul E. McKenney LKML-Reference: <1260828571-3613-3-git-send-email-acme@infradead.org> Signed-off-by: Ingo Molnar --- tools/perf/Documentation/perf-diff.txt | 31 ++++ tools/perf/Makefile | 1 + tools/perf/builtin-diff.c | 288 +++++++++++++++++++++++++++++++++ tools/perf/builtin.h | 1 + tools/perf/command-list.txt | 1 + tools/perf/perf.c | 1 + tools/perf/util/sort.h | 8 +- 7 files changed, 329 insertions(+), 2 deletions(-) create mode 100644 tools/perf/Documentation/perf-diff.txt create mode 100644 tools/perf/builtin-diff.c (limited to 'tools') diff --git a/tools/perf/Documentation/perf-diff.txt b/tools/perf/Documentation/perf-diff.txt new file mode 100644 index 00000000000..bd1ee55cef6 --- /dev/null +++ b/tools/perf/Documentation/perf-diff.txt @@ -0,0 +1,31 @@ +perf-diff(1) +============== + +NAME +---- +perf-diff - Read perf.data (created by perf record) and display the profile + +SYNOPSIS +-------- +[verse] +'perf diff' [oldfile] [newfile] + +DESCRIPTION +----------- +This command displays the performance difference among two perf.data files +captured via perf record. + +If no parameters are passed it will assume perf.data.old and perf.data. + +OPTIONS +------- +-p:: +--percentage:: + Show percentages instead of raw counters +-v:: +--verbose:: + Be verbose, for instance, show the raw counters in addition to the + diff. +SEE ALSO +-------- +linkperf:perf-record[1] diff --git a/tools/perf/Makefile b/tools/perf/Makefile index a4cb7925538..87a424e4cda 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -427,6 +427,7 @@ BUILTIN_OBJS += bench/sched-messaging.o BUILTIN_OBJS += bench/sched-pipe.o BUILTIN_OBJS += bench/mem-memcpy.o +BUILTIN_OBJS += builtin-diff.o BUILTIN_OBJS += builtin-help.o BUILTIN_OBJS += builtin-sched.o BUILTIN_OBJS += builtin-buildid-list.o diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c new file mode 100644 index 00000000000..0d528018ffb --- /dev/null +++ b/tools/perf/builtin-diff.c @@ -0,0 +1,288 @@ +/* + * builtin-diff.c + * + * Builtin diff command: Analyze two perf.data input files, look up and read + * DSOs and symbol information, sort them and produce a diff. + */ +#include "builtin.h" + +#include "util/debug.h" +#include "util/event.h" +#include "util/hist.h" +#include "util/session.h" +#include "util/sort.h" +#include "util/symbol.h" +#include "util/util.h" + +#include + +static char const *input_old = "perf.data.old", + *input_new = "perf.data"; +static int force; +static bool show_percent; + +struct symbol_conf symbol_conf; + +static int perf_session__add_hist_entry(struct perf_session *self, + struct addr_location *al, u64 count) +{ + bool hit; + struct hist_entry *he = __perf_session__add_hist_entry(self, al, NULL, + count, &hit); + if (he == NULL) + return -ENOMEM; + + if (hit) + he->count += count; + + return 0; +} + +static int diff__process_sample_event(event_t *event, struct perf_session *session) +{ + struct addr_location al; + struct sample_data data = { .period = 1, }; + + dump_printf("(IP, %d): %d: %p\n", event->header.misc, + event->ip.pid, (void *)(long)event->ip.ip); + + if (event__preprocess_sample(event, session, &al, NULL) < 0) { + pr_warning("problem processing %d event, skipping it.\n", + event->header.type); + return -1; + } + + event__parse_sample(event, session->sample_type, &data); + + if (al.sym && perf_session__add_hist_entry(session, &al, data.period)) { + pr_warning("problem incrementing symbol count, skipping event\n"); + return -1; + } + + session->events_stats.total += data.period; + return 0; +} + +static struct perf_event_ops event_ops = { + .process_sample_event = diff__process_sample_event, + .process_mmap_event = event__process_mmap, + .process_comm_event = event__process_comm, + .process_exit_event = event__process_task, + .process_fork_event = event__process_task, + .process_lost_event = event__process_lost, +}; + +static void perf_session__insert_hist_entry_by_name(struct rb_root *root, + struct hist_entry *he) +{ + struct rb_node **p = &root->rb_node; + struct rb_node *parent = NULL; + struct hist_entry *iter; + + while (*p != NULL) { + int cmp; + parent = *p; + iter = rb_entry(parent, struct hist_entry, rb_node); + + cmp = strcmp(he->map->dso->name, iter->map->dso->name); + if (cmp > 0) + p = &(*p)->rb_left; + else if (cmp < 0) + p = &(*p)->rb_right; + else { + cmp = strcmp(he->sym->name, iter->sym->name); + if (cmp > 0) + p = &(*p)->rb_left; + else + p = &(*p)->rb_right; + } + } + + rb_link_node(&he->rb_node, parent, p); + rb_insert_color(&he->rb_node, root); +} + +static void perf_session__resort_by_name(struct perf_session *self) +{ + unsigned long position = 1; + struct rb_root tmp = RB_ROOT; + struct rb_node *next = rb_first(&self->hists); + + while (next != NULL) { + struct hist_entry *n = rb_entry(next, struct hist_entry, rb_node); + + next = rb_next(&n->rb_node); + rb_erase(&n->rb_node, &self->hists); + n->position = position++; + perf_session__insert_hist_entry_by_name(&tmp, n); + } + + self->hists = tmp; +} + +static struct hist_entry * +perf_session__find_hist_entry_by_name(struct perf_session *self, + struct hist_entry *he) +{ + struct rb_node *n = self->hists.rb_node; + + while (n) { + struct hist_entry *iter = rb_entry(n, struct hist_entry, rb_node); + int cmp = strcmp(he->map->dso->name, iter->map->dso->name); + + if (cmp > 0) + n = n->rb_left; + else if (cmp < 0) + n = n->rb_right; + else { + cmp = strcmp(he->sym->name, iter->sym->name); + if (cmp > 0) + n = n->rb_left; + else if (cmp < 0) + n = n->rb_right; + else + return iter; + } + } + + return NULL; +} + +static void perf_session__match_hists(struct perf_session *old_session, + struct perf_session *new_session) +{ + struct rb_node *nd; + + perf_session__resort_by_name(old_session); + + for (nd = rb_first(&new_session->hists); nd; nd = rb_next(nd)) { + struct hist_entry *pos = rb_entry(nd, struct hist_entry, rb_node); + pos->pair = perf_session__find_hist_entry_by_name(old_session, pos); + } +} + +static size_t hist_entry__fprintf_matched(struct hist_entry *self, + unsigned long pos, + struct perf_session *session, + struct perf_session *pair_session, + FILE *fp) +{ + u64 old_count = 0; + char displacement[16]; + size_t printed; + + if (self->pair != NULL) { + long pdiff = (long)self->pair->position - (long)pos; + old_count = self->pair->count; + if (pdiff == 0) + goto blank; + snprintf(displacement, sizeof(displacement), "%+4ld", pdiff); + } else { +blank: memset(displacement, ' ', sizeof(displacement)); + } + + printed = fprintf(fp, "%4lu %5.5s ", pos, displacement); + + if (show_percent) { + double old_percent = (old_count * 100) / pair_session->events_stats.total, + new_percent = (self->count * 100) / session->events_stats.total; + double diff = old_percent - new_percent; + + if (verbose) + printed += fprintf(fp, " %3.2f%% %3.2f%%", old_percent, new_percent); + + if ((u64)diff != 0) + printed += fprintf(fp, " %+4.2F%%", diff); + else + printed += fprintf(fp, " "); + } else { + if (verbose) + printed += fprintf(fp, " %9Lu %9Lu", old_count, self->count); + printed += fprintf(fp, " %+9Ld", (s64)self->count - (s64)old_count); + } + + return printed + fprintf(fp, " %25.25s %s\n", + self->map->dso->name, self->sym->name); +} + +static size_t perf_session__fprintf_matched_hists(struct perf_session *self, + struct perf_session *pair, + FILE *fp) +{ + struct rb_node *nd; + size_t printed = 0; + unsigned long pos = 1; + + for (nd = rb_first(&self->hists); nd; nd = rb_next(nd)) { + struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node); + printed += hist_entry__fprintf_matched(he, pos++, self, pair, fp); + } + + return printed; +} + +static int __cmd_diff(void) +{ + int ret, i; + struct perf_session *session[2]; + + session[0] = perf_session__new(input_old, O_RDONLY, force, &symbol_conf); + session[1] = perf_session__new(input_new, O_RDONLY, force, &symbol_conf); + if (session[0] == NULL || session[1] == NULL) + return -ENOMEM; + + for (i = 0; i < 2; ++i) { + ret = perf_session__process_events(session[i], &event_ops); + if (ret) + goto out_delete; + perf_session__output_resort(session[i], session[i]->events_stats.total); + } + + perf_session__match_hists(session[0], session[1]); + perf_session__fprintf_matched_hists(session[1], session[0], stdout); +out_delete: + for (i = 0; i < 2; ++i) + perf_session__delete(session[i]); + return ret; +} + +static const char *const diff_usage[] = { + "perf diff [] [old_file] [new_file]", +}; + +static const struct option options[] = { + OPT_BOOLEAN('v', "verbose", &verbose, + "be more verbose (show symbol address, etc)"), + OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace, + "dump raw trace in ASCII"), + OPT_BOOLEAN('f', "force", &force, "don't complain, do it"), + OPT_BOOLEAN('m', "modules", &symbol_conf.use_modules, + "load module symbols - WARNING: use only with -k and LIVE kernel"), + OPT_BOOLEAN('p', "percentages", &show_percent, + "Don't shorten the pathnames taking into account the cwd"), + OPT_BOOLEAN('P', "full-paths", &event_ops.full_paths, + "Don't shorten the pathnames taking into account the cwd"), + OPT_END() +}; + +int cmd_diff(int argc, const char **argv, const char *prefix __used) +{ + if (symbol__init(&symbol_conf) < 0) + return -1; + + setup_sorting(diff_usage, options); + + argc = parse_options(argc, argv, options, diff_usage, 0); + if (argc) { + if (argc > 2) + usage_with_options(diff_usage, options); + if (argc == 2) { + input_old = argv[0]; + input_new = argv[1]; + } else + input_new = argv[0]; + } + + setup_pager(); + return __cmd_diff(); +} diff --git a/tools/perf/builtin.h b/tools/perf/builtin.h index a3d8bf65f26..18035b1f16c 100644 --- a/tools/perf/builtin.h +++ b/tools/perf/builtin.h @@ -17,6 +17,7 @@ extern int check_pager_config(const char *cmd); extern int cmd_annotate(int argc, const char **argv, const char *prefix); extern int cmd_bench(int argc, const char **argv, const char *prefix); extern int cmd_buildid_list(int argc, const char **argv, const char *prefix); +extern int cmd_diff(int argc, const char **argv, const char *prefix); extern int cmd_help(int argc, const char **argv, const char *prefix); extern int cmd_sched(int argc, const char **argv, const char *prefix); extern int cmd_list(int argc, const char **argv, const char *prefix); diff --git a/tools/perf/command-list.txt b/tools/perf/command-list.txt index 02b09ea17a3..71dc7c3fe7b 100644 --- a/tools/perf/command-list.txt +++ b/tools/perf/command-list.txt @@ -5,6 +5,7 @@ perf-annotate mainporcelain common perf-bench mainporcelain common perf-buildid-list mainporcelain common +perf-diff mainporcelain common perf-list mainporcelain common perf-sched mainporcelain common perf-record mainporcelain common diff --git a/tools/perf/perf.c b/tools/perf/perf.c index cf64049bc9b..873e55fab37 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -286,6 +286,7 @@ static void handle_internal_command(int argc, const char **argv) const char *cmd = argv[0]; static struct cmd_struct commands[] = { { "buildid-list", cmd_buildid_list, 0 }, + { "diff", cmd_diff, 0 }, { "help", cmd_help, 0 }, { "list", cmd_list, 0 }, { "record", cmd_record, 0 }, diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h index cb6151c026f..925f083e1ee 100644 --- a/tools/perf/util/sort.h +++ b/tools/perf/util/sort.h @@ -49,9 +49,13 @@ struct hist_entry { struct symbol *sym; u64 ip; char level; - struct symbol *parent; + struct symbol *parent; struct callchain_node callchain; - struct rb_root sorted_chain; + union { + unsigned long position; + struct hist_entry *pair; + struct rb_root sorted_chain; + }; }; enum sort_type { -- cgit From c249a4ce796b30b742bb4854bf3039ced12ef8e5 Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Tue, 15 Dec 2009 03:15:22 +0100 Subject: perf tools: Make symbol_conf static perf top, report and annotate all define their own symbol_conf, it should be static. Signed-off-by: Frederic Weisbecker Cc: Peter Zijlstra Cc: Arnaldo Carvalho de Melo Cc: Paul Mackerras LKML-Reference: <1260843322-6602-1-git-send-regression-fweisbec@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-report.c | 2 +- tools/perf/builtin-top.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index beff7c090d2..40389c0e38c 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -52,7 +52,7 @@ static int exclude_other = 1; static char callchain_default_opt[] = "fractal,0.5"; -struct symbol_conf symbol_conf; +static struct symbol_conf symbol_conf; static size_t diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 3b212bb2e4d..296e809c253 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -80,7 +80,7 @@ static int dump_symtab = 0; static bool hide_kernel_symbols = false; static bool hide_user_symbols = false; static struct winsize winsize; -struct symbol_conf symbol_conf; +static struct symbol_conf symbol_conf; /* * Source -- cgit From 586bc5cce88be993dad584c3936c49f945368551 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 15 Dec 2009 02:53:35 -0600 Subject: perf trace/scripting: Add support for script args One oversight of the original scripting_ops patch was a lack of support for passing args to handler scripts. This adds argc/argv to the start_script() scripting_op, and changes the rw-by-file script to take 'comm' arg rather than the 'perf' value currently hard-coded. It also takes the opportunity to do some related minor cleanup. Signed-off-by: Tom Zanussi Cc: fweisbec@gmail.com Cc: rostedt@goodmis.org LKML-Reference: <1260867220-15699-2-git-send-email-tzanussi@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-trace.c | 19 ++++++-------- tools/perf/scripts/perl/bin/rw-by-file-report | 2 +- tools/perf/scripts/perl/rw-by-file.pl | 5 ++-- tools/perf/util/trace-event-perl.c | 36 +++++++++++++++++++-------- tools/perf/util/trace-event.h | 2 +- 5 files changed, 38 insertions(+), 26 deletions(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 07da66511bd..88b0353d401 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -12,7 +12,9 @@ static char const *script_name; static char const *generate_script_lang; -static int default_start_script(const char *script __attribute((unused))) +static int default_start_script(const char *script __unused, + int argc __unused, + const char **argv __unused) { return 0; } @@ -22,7 +24,7 @@ static int default_stop_script(void) return 0; } -static int default_generate_script(const char *outfile __attribute ((unused))) +static int default_generate_script(const char *outfile __unused) { return 0; } @@ -302,15 +304,8 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) setup_scripting(); - argc = parse_options(argc, argv, options, annotate_usage, 0); - if (argc) { - /* - * Special case: if there's an argument left then assume tha - * it's a symbol filter: - */ - if (argc > 1) - usage_with_options(annotate_usage, options); - } + argc = parse_options(argc, argv, options, annotate_usage, + PARSE_OPT_STOP_AT_NON_OPTION); setup_pager(); @@ -350,7 +345,7 @@ int cmd_trace(int argc, const char **argv, const char *prefix __used) } if (script_name) { - err = scripting_ops->start_script(script_name); + err = scripting_ops->start_script(script_name, argc, argv); if (err) goto out; } diff --git a/tools/perf/scripts/perl/bin/rw-by-file-report b/tools/perf/scripts/perl/bin/rw-by-file-report index f5dcf9cb5bd..1c0567516ab 100644 --- a/tools/perf/scripts/perl/bin/rw-by-file-report +++ b/tools/perf/scripts/perl/bin/rw-by-file-report @@ -1,5 +1,5 @@ #!/bin/bash -perf trace -s ~/libexec/perf-core/scripts/perl/rw-by-file.pl +perf trace -s ~/libexec/perf-core/scripts/perl/rw-by-file.pl $1 diff --git a/tools/perf/scripts/perl/rw-by-file.pl b/tools/perf/scripts/perl/rw-by-file.pl index 61f91561d84..2a39097687b 100644 --- a/tools/perf/scripts/perl/rw-by-file.pl +++ b/tools/perf/scripts/perl/rw-by-file.pl @@ -18,8 +18,9 @@ use lib "./Perf-Trace-Util/lib"; use Perf::Trace::Core; use Perf::Trace::Util; -# change this to the comm of the program you're interested in -my $for_comm = "perf"; +my $usage = "perf trace -s rw-by-file.pl \n"; + +my $for_comm = shift or die $usage; my %reads; my %writes; diff --git a/tools/perf/util/trace-event-perl.c b/tools/perf/util/trace-event-perl.c index a5ffe60db5d..6f10e760245 100644 --- a/tools/perf/util/trace-event-perl.c +++ b/tools/perf/util/trace-event-perl.c @@ -267,7 +267,7 @@ int common_lock_depth(struct scripting_context *context) } static void perl_process_event(int cpu, void *data, - int size __attribute((unused)), + int size __unused, unsigned long long nsecs, char *comm) { struct format_field *field; @@ -359,28 +359,42 @@ static void run_start_sub(void) /* * Start trace script */ -static int perl_start_script(const char *script) +static int perl_start_script(const char *script, int argc, const char **argv) { - const char *command_line[2] = { "", NULL }; + const char **command_line; + int i, err = 0; + command_line = malloc((argc + 2) * sizeof(const char *)); + command_line[0] = ""; command_line[1] = script; + for (i = 2; i < argc + 2; i++) + command_line[i] = argv[i - 2]; my_perl = perl_alloc(); perl_construct(my_perl); - if (perl_parse(my_perl, xs_init, 2, (char **)command_line, - (char **)NULL)) - return -1; + if (perl_parse(my_perl, xs_init, argc + 2, (char **)command_line, + (char **)NULL)) { + err = -1; + goto error; + } perl_run(my_perl); - if (SvTRUE(ERRSV)) - return -1; + if (SvTRUE(ERRSV)) { + err = -1; + goto error; + } run_start_sub(); + free(command_line); fprintf(stderr, "perf trace started with Perl script %s\n\n", script); - return 0; +error: + perl_free(my_perl); + free(command_line); + + return err; } /* @@ -579,7 +593,9 @@ static void print_unsupported_msg(void) "\n etc.\n"); } -static int perl_start_script_unsupported(const char *script __unused) +static int perl_start_script_unsupported(const char *script __unused, + int argc __unused, + const char **argv __unused) { print_unsupported_msg(); diff --git a/tools/perf/util/trace-event.h b/tools/perf/util/trace-event.h index 81698d5e650..6ad405620c9 100644 --- a/tools/perf/util/trace-event.h +++ b/tools/perf/util/trace-event.h @@ -270,7 +270,7 @@ enum trace_flag_type { struct scripting_ops { const char *name; - int (*start_script) (const char *); + int (*start_script) (const char *script, int argc, const char **argv); int (*stop_script) (void); void (*process_event) (int cpu, void *data, int size, unsigned long long nsecs, char *comm); -- cgit From a3a7cb7bb1d7bd989982314cf6f90ec392890006 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 15 Dec 2009 02:53:36 -0600 Subject: perf trace/scripting: Don't install unneeded files README and Makefile.PL don't need to be installed for Perl run-time support. Signed-off-by: Tom Zanussi Cc: fweisbec@gmail.com Cc: rostedt@goodmis.org LKML-Reference: <1260867220-15699-3-git-send-email-tzanussi@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/Makefile | 2 -- 1 file changed, 2 deletions(-) (limited to 'tools') diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 87a424e4cda..7814dbbd401 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -996,8 +996,6 @@ install: all $(INSTALL) scripts/perl/Perf-Trace-Util/lib/Perf/Trace/* -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util/lib/Perf/Trace' $(INSTALL) scripts/perl/*.pl -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl' $(INSTALL) scripts/perl/bin/* -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/bin' - $(INSTALL) scripts/perl/Perf-Trace-Util/Makefile.PL -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util' - $(INSTALL) scripts/perl/Perf-Trace-Util/README -t '$(DESTDIR_SQ)$(perfexec_instdir_SQ)/scripts/perl/Perf-Trace-Util' ifdef BUILT_INS $(INSTALL) -d -m 755 '$(DESTDIR_SQ)$(perfexec_instdir_SQ)' $(INSTALL) $(BUILT_INS) '$(DESTDIR_SQ)$(perfexec_instdir_SQ)' -- cgit From 8f11d85a0e7e9025acea7493e6864089c8b52f42 Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 15 Dec 2009 02:53:37 -0600 Subject: perf trace/scripting: Check return val of perl_run() The return value from perl_run() is currently ignored, but it should be checked and used to exit perf if there are problems loading the script. Signed-off-by: Tom Zanussi Cc: fweisbec@gmail.com Cc: rostedt@goodmis.org LKML-Reference: <1260867220-15699-4-git-send-email-tzanussi@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/util/trace-event-perl.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/util/trace-event-perl.c b/tools/perf/util/trace-event-perl.c index 6f10e760245..6d6d76b8a21 100644 --- a/tools/perf/util/trace-event-perl.c +++ b/tools/perf/util/trace-event-perl.c @@ -379,7 +379,11 @@ static int perl_start_script(const char *script, int argc, const char **argv) goto error; } - perl_run(my_perl); + if (perl_run(my_perl)) { + err = -1; + goto error; + } + if (SvTRUE(ERRSV)) { err = -1; goto error; -- cgit From 4b9c0c596ea826ef784eb83f663c5351ed01ba6d Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 15 Dec 2009 02:53:38 -0600 Subject: perf trace/scripting: List available scripts Lists the available perf trace scripts, one per line e.g.: root@tropicana:~# perf trace -l List of available trace scripts: workqueue-stats workqueue stats (ins/exe/create/destroy) wakeup-latency system-wide min/max/avg wakeup latency rw-by-file r/w activity for a program, by file check-perf-trace useless but exhaustive test script rw-by-pid system-wide r/w activity To be consistent with the other listing options in perf, the current latency trace option was changed to '-L', and '-l' is now used to access the script listing as: To create the list, it searches each scripts/*/bin directory for files ending with "-report" and reads information found in certain comment lines contained in those shell scripts: - if the comment line starts with "description:", the rest of the line is used as a 'half-line' description. To keep each line in the list to a single line, the description should be limited to 40 characters (the rest of the line contains the script name and args) - if the comment line starts with "args:", the rest of the line names the args the script supports. Required args should be surrounded by <> brackets, optional args by [] brackets. The current scripts in scripts/perl/bin have also been updated with description: and args: comments. Signed-off-by: Tom Zanussi Cc: fweisbec@gmail.com Cc: rostedt@goodmis.org LKML-Reference: <1260867220-15699-5-git-send-email-tzanussi@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-trace.c | 199 ++++++++++++++++++++- .../perf/scripts/perl/bin/check-perf-trace-report | 1 + tools/perf/scripts/perl/bin/rw-by-file-report | 2 + tools/perf/scripts/perl/bin/rw-by-pid-report | 1 + tools/perf/scripts/perl/bin/wakeup-latency-report | 1 + tools/perf/scripts/perl/bin/workqueue-stats-report | 1 + 6 files changed, 204 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 88b0353d401..7674153c4bb 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -274,6 +274,201 @@ static int parse_scriptname(const struct option *opt __used, return 0; } +#define for_each_lang(scripts_dir, lang_dirent, lang_next) \ + while (!readdir_r(scripts_dir, &lang_dirent, &lang_next) && \ + lang_next) \ + if (lang_dirent.d_type == DT_DIR && \ + (strcmp(lang_dirent.d_name, ".")) && \ + (strcmp(lang_dirent.d_name, ".."))) + +#define for_each_script(lang_dir, script_dirent, script_next) \ + while (!readdir_r(lang_dir, &script_dirent, &script_next) && \ + script_next) \ + if (script_dirent.d_type != DT_DIR) + + +#define RECORD_SUFFIX "-record" +#define REPORT_SUFFIX "-report" + +struct script_desc { + struct list_head node; + char *name; + char *half_liner; + char *args; +}; + +LIST_HEAD(script_descs); + +static struct script_desc *script_desc__new(const char *name) +{ + struct script_desc *s = zalloc(sizeof(*s)); + + if (s != NULL) + s->name = strdup(name); + + return s; +} + +static void script_desc__delete(struct script_desc *s) +{ + free(s->name); + free(s); +} + +static void script_desc__add(struct script_desc *s) +{ + list_add_tail(&s->node, &script_descs); +} + +static struct script_desc *script_desc__find(const char *name) +{ + struct script_desc *s; + + list_for_each_entry(s, &script_descs, node) + if (strcasecmp(s->name, name) == 0) + return s; + return NULL; +} + +static struct script_desc *script_desc__findnew(const char *name) +{ + struct script_desc *s = script_desc__find(name); + + if (s) + return s; + + s = script_desc__new(name); + if (!s) + goto out_delete_desc; + + script_desc__add(s); + + return s; + +out_delete_desc: + script_desc__delete(s); + + return NULL; +} + +static char *ends_with(char *str, const char *suffix) +{ + size_t suffix_len = strlen(suffix); + char *p = str; + + if (strlen(str) > suffix_len) { + p = str + strlen(str) - suffix_len; + if (!strncmp(p, suffix, suffix_len)) + return p; + } + + return NULL; +} + +static char *ltrim(char *str) +{ + int len = strlen(str); + + while (len && isspace(*str)) { + len--; + str++; + } + + return str; +} + +static int read_script_info(struct script_desc *desc, const char *filename) +{ + char line[BUFSIZ], *p; + FILE *fp; + + fp = fopen(filename, "r"); + if (!fp) + return -1; + + while (fgets(line, sizeof(line), fp)) { + p = ltrim(line); + if (strlen(p) == 0) + continue; + if (*p != '#') + continue; + p++; + if (strlen(p) && *p == '!') + continue; + + p = ltrim(p); + if (strlen(p) && p[strlen(p) - 1] == '\n') + p[strlen(p) - 1] = '\0'; + + if (!strncmp(p, "description:", strlen("description:"))) { + p += strlen("description:"); + desc->half_liner = strdup(ltrim(p)); + continue; + } + + if (!strncmp(p, "args:", strlen("args:"))) { + p += strlen("args:"); + desc->args = strdup(ltrim(p)); + continue; + } + } + + fclose(fp); + + return 0; +} + +static int list_available_scripts(const struct option *opt __used, + const char *s __used, int unset __used) +{ + struct dirent *script_next, *lang_next, script_dirent, lang_dirent; + char scripts_path[MAXPATHLEN]; + DIR *scripts_dir, *lang_dir; + char script_path[MAXPATHLEN]; + char lang_path[MAXPATHLEN]; + struct script_desc *desc; + char first_half[BUFSIZ]; + char *script_root; + char *str; + + snprintf(scripts_path, MAXPATHLEN, "%s/scripts", perf_exec_path()); + + scripts_dir = opendir(scripts_path); + if (!scripts_dir) + return -1; + + for_each_lang(scripts_dir, lang_dirent, lang_next) { + snprintf(lang_path, MAXPATHLEN, "%s/%s/bin", scripts_path, + lang_dirent.d_name); + lang_dir = opendir(lang_path); + if (!lang_dir) + continue; + + for_each_script(lang_dir, script_dirent, script_next) { + script_root = strdup(script_dirent.d_name); + str = ends_with(script_root, REPORT_SUFFIX); + if (str) { + *str = '\0'; + desc = script_desc__findnew(script_root); + snprintf(script_path, MAXPATHLEN, "%s/%s", + lang_path, script_dirent.d_name); + read_script_info(desc, script_path); + } + free(script_root); + } + } + + fprintf(stdout, "List of available trace scripts:\n"); + list_for_each_entry(desc, &script_descs, node) { + sprintf(first_half, "%s %s", desc->name, + desc->args ? desc->args : ""); + fprintf(stdout, " %-36s %s\n", first_half, + desc->half_liner ? desc->half_liner : ""); + } + + exit(0); +} + static const char * const annotate_usage[] = { "perf trace [] ", NULL @@ -284,8 +479,10 @@ static const struct option options[] = { "dump raw trace in ASCII"), OPT_BOOLEAN('v', "verbose", &verbose, "be more verbose (show symbol address, etc)"), - OPT_BOOLEAN('l', "latency", &latency_format, + OPT_BOOLEAN('L', "Latency", &latency_format, "show latency attributes (irqs/preemption disabled, etc)"), + OPT_CALLBACK_NOOPT('l', "list", NULL, NULL, "list available scripts", + list_available_scripts), OPT_CALLBACK('s', "script", NULL, "name", "script file name (lang:script name, script name, or *)", parse_scriptname), diff --git a/tools/perf/scripts/perl/bin/check-perf-trace-report b/tools/perf/scripts/perl/bin/check-perf-trace-report index 89948b01502..7fc4a033dd4 100644 --- a/tools/perf/scripts/perl/bin/check-perf-trace-report +++ b/tools/perf/scripts/perl/bin/check-perf-trace-report @@ -1,4 +1,5 @@ #!/bin/bash +# description: useless but exhaustive test script perf trace -s ~/libexec/perf-core/scripts/perl/check-perf-trace.pl diff --git a/tools/perf/scripts/perl/bin/rw-by-file-report b/tools/perf/scripts/perl/bin/rw-by-file-report index 1c0567516ab..eddb9ccce6a 100644 --- a/tools/perf/scripts/perl/bin/rw-by-file-report +++ b/tools/perf/scripts/perl/bin/rw-by-file-report @@ -1,4 +1,6 @@ #!/bin/bash +# description: r/w activity for a program, by file +# args: perf trace -s ~/libexec/perf-core/scripts/perl/rw-by-file.pl $1 diff --git a/tools/perf/scripts/perl/bin/rw-by-pid-report b/tools/perf/scripts/perl/bin/rw-by-pid-report index cea16f78a3a..7f44c25cc85 100644 --- a/tools/perf/scripts/perl/bin/rw-by-pid-report +++ b/tools/perf/scripts/perl/bin/rw-by-pid-report @@ -1,4 +1,5 @@ #!/bin/bash +# description: system-wide r/w activity perf trace -s ~/libexec/perf-core/scripts/perl/rw-by-pid.pl diff --git a/tools/perf/scripts/perl/bin/wakeup-latency-report b/tools/perf/scripts/perl/bin/wakeup-latency-report index 85769dc456e..fce3adcb324 100644 --- a/tools/perf/scripts/perl/bin/wakeup-latency-report +++ b/tools/perf/scripts/perl/bin/wakeup-latency-report @@ -1,4 +1,5 @@ #!/bin/bash +# description: system-wide min/max/avg wakeup latency perf trace -s ~/libexec/perf-core/scripts/perl/wakeup-latency.pl diff --git a/tools/perf/scripts/perl/bin/workqueue-stats-report b/tools/perf/scripts/perl/bin/workqueue-stats-report index aa68435be92..71cfbd182fb 100644 --- a/tools/perf/scripts/perl/bin/workqueue-stats-report +++ b/tools/perf/scripts/perl/bin/workqueue-stats-report @@ -1,4 +1,5 @@ #!/bin/bash +# description: workqueue stats (ins/exe/create/destroy) perf trace -s ~/libexec/perf-core/scripts/perl/workqueue-stats.pl -- cgit From 3875294f5c0d7b9ef96ffc373d8a956ebd7c0c7f Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 15 Dec 2009 02:53:39 -0600 Subject: perf trace/scripting: Add 'record' and 'report' options Allow scripts to be recorded/executed by simply specifying the script root name (the script name minus extension) along with 'record' or 'report' to 'perf trace'. The script names shown by 'perf trace -l' can be directly used to run the command-line contained within the corresponding '-record' and '-report' versions of scripts in the scripts/*/bin directories. For example, to record the trace data needed to run the wakeup-latency.pl script, the user can easily find the name of the corresponding script from the script list and invoke it using 'perf trace record', without having to remember the details of how to do the same thing using the lower-level perf trace command-line options: root@tropicana:~# perf trace -l List of available trace scripts: workqueue-stats workqueue stats (ins/exe/create/destroy) wakeup-latency system-wide min/max/avg wakeup latency rw-by-file r/w activity for a program, by file check-perf-trace useless but exhaustive test script rw-by-pid system-wide r/w activity root@tropicana:~# perf trace record wakeup-latency ^C[ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.296 MB perf.data (~12931 samples) ] To run the wakeup-latency.pl script using the captured data, change 'record' to 'report' in the command-line: root@tropicana:~# perf trace report wakeup-latency wakeup_latency stats: total_wakeups: 65 avg_wakeup_latency (ns): 22417 min_wakeup_latency (ns): 3470 max_wakeup_latency (ns): 223311 perf trace Perl script stopped If the script takes options, thay can be simply added to the end of the 'report' invocation: root@tropicana:~# perf trace record rw-by-file ^C[ perf record: Woken up 2 times to write data ] [ perf record: Captured and wrote 0.782 MB perf.data (~34171 samples) ] root@tropicana:~# perf trace report rw-by-file perf file read counts for perf: fd # reads bytes_requested ------ ---------- ----------- 122 1934 1980416 120 1 32 file write counts for perf: fd # writes bytes_written ------ ---------- ----------- 3 4006 280568 perf trace Perl script stopped Signed-off-by: Tom Zanussi Cc: fweisbec@gmail.com Cc: rostedt@goodmis.org LKML-Reference: <1260867220-15699-6-git-send-email-tzanussi@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/builtin-trace.c | 84 +++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 83 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c index 7674153c4bb..7e744f77404 100644 --- a/tools/perf/builtin-trace.c +++ b/tools/perf/builtin-trace.c @@ -469,6 +469,49 @@ static int list_available_scripts(const struct option *opt __used, exit(0); } +static char *get_script_path(const char *script_root, const char *suffix) +{ + struct dirent *script_next, *lang_next, script_dirent, lang_dirent; + char scripts_path[MAXPATHLEN]; + char script_path[MAXPATHLEN]; + DIR *scripts_dir, *lang_dir; + char lang_path[MAXPATHLEN]; + char *str, *__script_root; + char *path = NULL; + + snprintf(scripts_path, MAXPATHLEN, "%s/scripts", perf_exec_path()); + + scripts_dir = opendir(scripts_path); + if (!scripts_dir) + return NULL; + + for_each_lang(scripts_dir, lang_dirent, lang_next) { + snprintf(lang_path, MAXPATHLEN, "%s/%s/bin", scripts_path, + lang_dirent.d_name); + lang_dir = opendir(lang_path); + if (!lang_dir) + continue; + + for_each_script(lang_dir, script_dirent, script_next) { + __script_root = strdup(script_dirent.d_name); + str = ends_with(__script_root, suffix); + if (str) { + *str = '\0'; + if (strcmp(__script_root, script_root)) + continue; + snprintf(script_path, MAXPATHLEN, "%s/%s", + lang_path, script_dirent.d_name); + path = strdup(script_path); + free(__script_root); + break; + } + free(__script_root); + } + } + + return path; +} + static const char * const annotate_usage[] = { "perf trace [] ", NULL @@ -494,8 +537,47 @@ static const struct option options[] = { int cmd_trace(int argc, const char **argv, const char *prefix __used) { - int err; struct perf_session *session; + const char *suffix = NULL; + const char **__argv; + char *script_path; + int i, err; + + if (argc >= 2 && strncmp(argv[1], "rec", strlen("rec")) == 0) { + if (argc < 3) { + fprintf(stderr, + "Please specify a record script\n"); + return -1; + } + suffix = RECORD_SUFFIX; + } + + if (argc >= 2 && strncmp(argv[1], "rep", strlen("rep")) == 0) { + if (argc < 3) { + fprintf(stderr, + "Please specify a report script\n"); + return -1; + } + suffix = REPORT_SUFFIX; + } + + if (suffix) { + script_path = get_script_path(argv[2], suffix); + if (!script_path) { + fprintf(stderr, "script not found\n"); + return -1; + } + + __argv = malloc((argc + 1) * sizeof(const char *)); + __argv[0] = "/bin/sh"; + __argv[1] = script_path; + for (i = 3; i < argc; i++) + __argv[i - 1] = argv[i]; + __argv[argc - 1] = NULL; + + execvp("/bin/sh", (char **)__argv); + exit(-1); + } symbol__init(0); -- cgit From a6005123ce22770dbd91bc3cb637ce0807ab959b Mon Sep 17 00:00:00 2001 From: Tom Zanussi Date: Tue, 15 Dec 2009 02:53:40 -0600 Subject: perf trace/scripting: Update Documentation Update the perf-trace page with new and missing options and remove some unused ones. Signed-off-by: Tom Zanussi Cc: fweisbec@gmail.com Cc: rostedt@goodmis.org LKML-Reference: <1260867220-15699-7-git-send-email-tzanussi@gmail.com> Signed-off-by: Ingo Molnar --- tools/perf/Documentation/perf-trace.txt | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) (limited to 'tools') diff --git a/tools/perf/Documentation/perf-trace.txt b/tools/perf/Documentation/perf-trace.txt index 07065efa60e..60e5900da48 100644 --- a/tools/perf/Documentation/perf-trace.txt +++ b/tools/perf/Documentation/perf-trace.txt @@ -8,18 +8,43 @@ perf-trace - Read perf.data (created by perf record) and display trace output SYNOPSIS -------- [verse] -'perf trace' [-i | --input=file] symbol_name +'perf trace' {record