[PATCH 3/3] perf script: dump software events and samples from hardware-based profiling
David Ahern
comm-pid [cpu] secs.usecs: event: IP symbol dso
Example:
perf record -v -ga -e cs -c 1 -- sleep 5
perf script
(Line lengths wrapped):
sshd-794 [000] 2572.863440: context-switches: ffffffff810355de \
perf_event_task_sched_out ([kernel.kallsyms])
sshd-794 [000] 2572.863440: context-switches: ffffffff81382b01 \
schedule ([kernel.kallsyms])
sshd-794 [000] 2572.863440: context-switches: ffffffff8138380a \
schedule_hrtimeout_range_clock ([kernel.kallsyms])
sshd-794 [000] 2572.863440: context-switches: ffffffff813838e6 \
schedule_hrtimeout_range ([kernel.kallsyms])
sshd-794 [000] 2572.863440: context-switches: ffffffff8110c55d \
poll_schedule_timeout ([kernel.kallsyms])
sshd-794 [000] 2572.863440: context-switches: ffffffff8110cd84 \
do_select ([kernel.kallsyms])
or 'perf script -G'
swapper-0 [001] 2572.863188: context-switches: ffffffff810355de ...
sshd-794 [000] 2572.863440: context-switches: ffffffff810355de ...
kworker/0:1-10 [000] 2572.863451: context-switches: ffffffff810355de ...
Signed-off-by: David Ahern <daa...@cisco.com>
---
tools/perf/Documentation/perf-script.txt | 19 +++++
tools/perf/builtin-script.c | 32 +++++++++
tools/perf/util/session.c | 113 ++++++++++++++++++++++++++++++
tools/perf/util/session.h | 6 ++
4 files changed, 170 insertions(+), 0 deletions(-)
diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt
index 29ad942..99db652 100644
--- a/tools/perf/Documentation/perf-script.txt
+++ b/tools/perf/Documentation/perf-script.txt
@@ -112,6 +112,25 @@ OPTIONS
--debug-mode::
Do various checks like samples ordering and lost events.
+-k::
+--vmlinux=<file>::
+ vmlinux pathname
+
+--kallsyms=<file>::
+ kallsyms pathname
+
+--symfs=<directory>::
+ Look for files with symbols relative to this directory.
+
+-U::
+--show-unresolved::
+ Display all addresses including unresolved to a symbol.
+
+-G::
+--hide-call-graph::
+ Do not display call chain.
+
+--
SEE ALSO
--------
linkperf:perf-record[1], linkperf:perf-script-perl[1],
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index f59d482..f1cad74b 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -19,6 +19,8 @@ static bool debug_mode;
static u64 last_timestamp;
static u64 nr_unordered;
extern const struct option record_options[];
+static bool show_unresolved;
+static bool no_callchain;
static void process_event(union perf_event *event,
struct perf_sample *sample,
@@ -57,6 +59,11 @@ static void process_event(union perf_event *event,
print_tracepoint_event(sample->cpu, sample->raw_data,
sample->raw_size, sample->time,
thread->comm);
+ } else if ((attr->type == PERF_TYPE_SOFTWARE) ||
+ ((attr->type == PERF_TYPE_HARDWARE) &&
+ (attr->config == PERF_COUNT_HW_CPU_CYCLES))) {
+ perf_session__print_sample(event, sample, session, attr,
+ show_unresolved);
} else {
evname = __event_name(attr->type, attr->config);
if (verbose)
@@ -130,7 +137,10 @@ static int process_sample_event(union perf_event *event,
static struct perf_event_ops event_ops = {
.sample = process_sample_event,
+ .mmap = perf_event__process_mmap,
.comm = perf_event__process_comm,
+ .exit = perf_event__process_task,
+ .fork = perf_event__process_task,
.attr = perf_event__process_attr,
.event_type = perf_event__process_event_type,
.tracing_data = perf_event__process_tracing_data,
@@ -620,6 +630,16 @@ static const struct option options[] = {
"input file name"),
OPT_BOOLEAN('d', "debug-mode", &debug_mode,
"do various checks like samples ordering and lost events"),
+ OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
+ "file", "vmlinux pathname"),
+ OPT_STRING(0, "kallsyms", &symbol_conf.kallsyms_name,
+ "file", "kallsyms pathname"),
+ OPT_BOOLEAN('G', "hide-call-graph", &no_callchain,
+ "Do not display call chain"),
+ OPT_BOOLEAN('U', "show-unresolved", &show_unresolved,
+ "Display all entries including unresolved to a symbol"),
+ OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory",
+ "Look for files with symbols relative to this directory"),
OPT_END()
};
@@ -763,6 +783,18 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
exit(-1);
}
+ if (no_callchain)
+ symbol_conf.use_callchain = false;
+
+ else {
+ symbol_conf.use_callchain = true;
+ if (callchain_register_param(&callchain_param) < 0) {
+ error("Can't register callchain params\n");
+ err = -EINVAL;
+ goto out;
+ }
+ }
+
if (rec_script_path)
script_path = rec_script_path;
if (rep_script_path)
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index a3a871f..a6c3a56 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -10,6 +10,7 @@
#include "session.h"
#include "sort.h"
#include "util.h"
+#include "trace-event.h"
static int perf_session__open(struct perf_session *self, bool force)
{
@@ -1137,3 +1138,115 @@ size_t perf_session__fprintf_dsos_buildid(struct perf_session *self, FILE *fp,
size_t ret = machine__fprintf_dsos_buildid(&self->host_machine, fp, with_hits);
return ret + machines__fprintf_dsos_buildid(&self->machines, fp, with_hits);
}
+
+static inline void print_one_symbol(const char *comm, pid_t pid,
+ u32 cpu, u64 secs, u64 usecs, const char *evname,
+ u64 addr, const char *symname, const char *dsoname)
+{
+ printf("%16s-%-5d ", comm, pid);
+
+ if (cpu != (u32) -1)
+ printf("[%03d]", cpu);
+
+ printf(" %5lu.%06lu: %s: ", secs, usecs, evname);
+
+ printf("%16" PRIx64 " %s (%s)\n",
+ addr, symname, dsoname);
+
+ return;
+}
+
+void perf_session__print_sample(union perf_event *event,
+ struct perf_sample *sample,
+ struct perf_session *session,
+ struct perf_event_attr *attr,
+ bool show_unresolved)
+{
+ struct callchain_cursor_node *node, *prev;
+ struct addr_location al;
+ const char *evname = NULL;
+ const char *comm;
+ const char *symname, *dsoname;
+ u32 cpu = -1;
+ u64 secs = 0, usecs = 0;
+
+ if (perf_event__preprocess_sample(event, session, &al, sample,
+ NULL) < 0) {
+ error("problem processing %d event, skipping it.\n",
+ event->header.type);
+ return;
+ }
+
+ if (session->sample_type & PERF_SAMPLE_TIME) {
+ u64 nsecs = sample->time;
+ secs = nsecs / NSECS_PER_SEC;
+ nsecs -= secs * NSECS_PER_SEC;
+ usecs = nsecs / NSECS_PER_USEC;
+ }
+
+ evname = __event_name(attr->type, attr->config);
+ if (!evname)
+ evname = "(unknown)";
+
+ comm = al.thread->comm_set ? al.thread->comm : "-";
+
+ if (attr->sample_type & PERF_SAMPLE_CPU)
+ cpu = sample->cpu;
+
+ if (symbol_conf.use_callchain && sample->callchain) {
+
+ if (perf_session__resolve_callchain(session, al.thread,
+ sample->callchain, NULL) != 0) {
+ if (verbose)
+ error("Failed to resolve callchain. Skipping\n");
+ return;
+ }
+
+ node = session->callchain_cursor.first;
+ if (!node)
+ return;
+
+ while (node) {
+ if (node->sym && node->sym->name)
+ symname = node->sym->name;
+ else if (show_unresolved)
+ symname = "";
+ else
+ goto next;
+
+ if (node->map && node->map->dso && node->map->dso->name)
+ dsoname = node->map->dso->name;
+ else if (show_unresolved)
+ dsoname = "";
+ else
+ goto next;
+
+ print_one_symbol(comm, al.thread->pid, cpu, secs, usecs,
+ evname, node->ip, symname, dsoname);
+
+next:
+ prev = node;
+ node = node->next;
+ }
+ /* put a spacer between samples when callchains are dumped */
+ printf("\n");
+
+ } else {
+ if (al.sym && al.sym->name)
+ symname = al.sym->name;
+ else if (show_unresolved)
+ symname = "";
+ else
+ return;
+
+ if (al.map && al.map->dso && al.map->dso->name)
+ dsoname = al.map->dso->name;
+ else if (show_unresolved)
+ dsoname = "";
+ else
+ return;
+
+ print_one_symbol(comm, al.thread->pid, cpu, secs, usecs,
+ evname, al.addr, symname, dsoname);
+ }
+}
diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h
index 977b3a1..3827048 100644
--- a/tools/perf/util/session.h
+++ b/tools/perf/util/session.h
@@ -165,4 +165,10 @@ static inline int perf_session__parse_sample(struct perf_session *session,
session->sample_id_all, sample);
}
+void perf_session__print_sample(union perf_event *event,
+ struct perf_sample *sample,
+ struct perf_session *session,
+ struct perf_event_attr *attr,
+ bool show_unresolved);
+
#endif /* __PERF_SESSION_H */
--
1.7.4
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majo...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
David Ahern
independent.
David Ahern (3):
perf script: change process_event prototype
perf script: prepare to handle more than tracepoint events
perf script: dump software events and samples from hardware-based
profiling
tools/perf/Documentation/perf-script.txt | 19 +++
tools/perf/builtin-script.c | 118 ++++++++++++++-----
.../util/scripting-engines/trace-event-python.c | 20 +++-
tools/perf/util/session.c | 113 +++++++++++++++++++
tools/perf/util/session.h | 6 +
tools/perf/util/trace-event-parse.c | 4 +-
tools/perf/util/trace-event-scripting.c | 8 +-
tools/perf/util/trace-event.h | 10 +-
8 files changed, 253 insertions(+), 45 deletions(-)
David Ahern
Signed-off-by: David Ahern <daa...@cisco.com>
---
tools/perf/builtin-script.c | 40 +++++++++++--------
.../util/scripting-engines/trace-event-python.c | 20 ++++++++-
tools/perf/util/trace-event-scripting.c | 8 +--
tools/perf/util/trace-event.h | 6 ++-
4 files changed, 47 insertions(+), 27 deletions(-)
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 5f40df6..0bee150 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -20,6 +20,27 @@ static u64 last_timestamp;
static u64 nr_unordered;
extern const struct option record_options[];
+static void process_event(union perf_event *event,
+ struct perf_sample *sample,
+ struct perf_session *session)
+{
+ struct thread *thread = perf_session__findnew(session, event->ip.pid);
+
+ if (thread == NULL) {
+ pr_debug("problem processing %d event, skipping it.\n",
+ event->header.type);
+ return;
+ }
+
+ /*
+ * FIXME: better resolve from pid from the struct trace_entry
+ * field, although it should be the same than this perf
+ * event pid
+ */
+ print_event(sample->cpu, sample->raw_data, sample->raw_size,
+ sample->time, thread->comm);
+}
+
static int default_start_script(const char *script __unused,
int argc __unused,
const char **argv __unused)
@@ -40,7 +61,7 @@ static int default_generate_script(const char *outfile __unused)
static struct scripting_ops default_scripting_ops = {
.start_script = default_start_script,
.stop_script = default_stop_script,
- .process_event = print_event,
+ .process_event = process_event,
.generate_script = default_generate_script,
};
@@ -67,14 +88,6 @@ static int process_sample_event(union perf_event *event,
struct perf_sample *sample,
struct perf_session *session)
{
- struct thread *thread = perf_session__findnew(session, event->ip.pid);
-
- if (thread == NULL) {
- pr_debug("problem processing %d event, skipping it.\n",
- event->header.type);
- return -1;
- }
-
if (session->sample_type & PERF_SAMPLE_RAW) {
if (debug_mode) {
if (sample->time < last_timestamp) {
@@ -86,14 +99,7 @@ static int process_sample_event(union perf_event *event,
last_timestamp = sample->time;
return 0;
}
- /*
- * 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(sample->cpu, sample->raw_data,
- sample->raw_size,
- sample->time, thread->comm);
+ scripting_ops->process_event(event, sample, session);
}
session->hists.stats.total_period += sample->period;
diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
index 2040b85..5b03fb6 100644
--- a/tools/perf/util/scripting-engines/trace-event-python.c
+++ b/tools/perf/util/scripting-engines/trace-event-python.c
@@ -204,9 +204,9 @@ static inline struct event *find_cache_event(int type)
return event;
}
-static void python_process_event(int cpu, void *data,
- int size __unused,
- unsigned long long nsecs, char *comm)
+static void python_process_event(union perf_event *pevent,
+ struct perf_sample *sample,
+ struct perf_session *session)
{
PyObject *handler, *retval, *context, *t, *obj, *dict = NULL;
static char handler_name[256];
@@ -218,6 +218,20 @@ static void python_process_event(int cpu, void *data,
int type;
int pid;
+ int cpu = sample->cpu;
+ void *data = sample->raw_data;
+ unsigned long long nsecs = sample->time;
+ char *comm;
+ struct thread *thread;
+
+ thread = perf_session__findnew(session, pevent->ip.pid);
+ if (thread == NULL) {
+ pr_debug("problem processing %d event, skipping it.\n",
+ pevent->header.type);
+ return;
+ }
+ comm = thread->comm;
+
t = PyTuple_New(MAX_FIELDS);
if (!t)
Py_FatalError("couldn't create Python tuple");
diff --git a/tools/perf/util/trace-event-scripting.c b/tools/perf/util/trace-event-scripting.c
index f7af2fc..91f1b33 100644
--- a/tools/perf/util/trace-event-scripting.c
+++ b/tools/perf/util/trace-event-scripting.c
@@ -36,11 +36,9 @@ static int stop_script_unsupported(void)
return 0;
}
-static void process_event_unsupported(int cpu __unused,
- void *data __unused,
- int size __unused,
- unsigned long long nsecs __unused,
- char *comm __unused)
+static void process_event_unsupported(union perf_event *event __unused,
+ struct perf_sample *sample __unused,
+ struct perf_session *session __unused)
{
}
diff --git a/tools/perf/util/trace-event.h b/tools/perf/util/trace-event.h
index b5f12ca..632b87b1 100644
--- a/tools/perf/util/trace-event.h
+++ b/tools/perf/util/trace-event.h
@@ -3,6 +3,7 @@
#include <stdbool.h>
#include "parse-events.h"
+#include "session.h"
#define __unused __attribute__((unused))
@@ -278,8 +279,9 @@ struct scripting_ops {
const char *name;
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);
+ void (*process_event) (union perf_event *event,
+ struct perf_sample *sample,
+ struct perf_session *session);
int (*generate_script) (const char *outfile);
};
David Ahern
tools/perf/util/trace-event-parse.c | 4 +-
tools/perf/util/trace-event.h | 4 +-
3 files changed, 50 insertions(+), 32 deletions(-)
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 0bee150..f59d482 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -24,21 +24,45 @@ static void process_event(union perf_event *event,
struct perf_sample *sample,
struct perf_session *session)
{
- struct thread *thread = perf_session__findnew(session, event->ip.pid);
-
- if (thread == NULL) {
- pr_debug("problem processing %d event, skipping it.\n",
- event->header.type);
- return;
+ struct perf_event_attr *attr;
+ struct thread *thread;
+ const char *evname = NULL;
+ static bool check_raw = true;
+
+ attr = perf_header__find_attr(sample->id, &session->header);
+ if (!attr) {
+ pr_debug("No attributes found for sample id %" PRId64 ". "
+ "skipping it.\n", sample->id);
}
- /*
- * FIXME: better resolve from pid from the struct trace_entry
- * field, although it should be the same than this perf
- * event pid
- */
- print_event(sample->cpu, sample->raw_data, sample->raw_size,
- sample->time, thread->comm);
+ if (attr->type == PERF_TYPE_TRACEPOINT) {
+ if (check_raw) {
+ if (!perf_session__has_traces(session, "record -R"))
+ return;
+ check_raw = false;
+ }
+
+ thread = perf_session__findnew(session, event->ip.pid);
+ if (thread == NULL) {
+ pr_debug("problem processing %d event, skipping it.\n",
+ event->header.type);
+ return;
+ }
+
+ /*
+ * FIXME: better resolve from pid from the struct trace_entry
+ * field, although it should be the same than this perf
+ * event pid
+ */
+ print_tracepoint_event(sample->cpu, sample->raw_data,
+ sample->raw_size, sample->time,
+ thread->comm);
+ } else {
+ evname = __event_name(attr->type, attr->config);
+ if (verbose)
+ warning("support for event %s not implemented. skipping\n",
+ evname ? evname : "(unknown)");
+ }
}
static int default_start_script(const char *script __unused,
@@ -88,19 +112,17 @@ static int process_sample_event(union perf_event *event,
struct perf_sample *sample,
struct perf_session *session)
{
- if (session->sample_type & PERF_SAMPLE_RAW) {
- if (debug_mode) {
- if (sample->time < last_timestamp) {
- pr_err("Samples misordered, previous: %" PRIu64
- " this: %" PRIu64 "\n", last_timestamp,
- sample->time);
- nr_unordered++;
- }
- last_timestamp = sample->time;
- return 0;
+ if (debug_mode) {
+ if (sample->time < last_timestamp) {
+ pr_err("Samples misordered, previous: %" PRIu64
+ " this: %" PRIu64 "\n", last_timestamp,
+ sample->time);
+ nr_unordered++;
}
- scripting_ops->process_event(event, sample, session);
+ last_timestamp = sample->time;
+ return 0;
}
+ scripting_ops->process_event(event, sample, session);
session->hists.stats.total_period += sample->period;
return 0;
@@ -778,10 +800,6 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
if (session == NULL)
return -ENOMEM;
- if (strcmp(input_name, "-") &&
- !perf_session__has_traces(session, "record -R"))
- return -EINVAL;
-
if (generate_script_lang) {
struct stat perf_stat;
diff --git a/tools/perf/util/trace-event-parse.c b/tools/perf/util/trace-event-parse.c
index d8e622d..efe1628 100644
--- a/tools/perf/util/trace-event-parse.c
+++ b/tools/perf/util/trace-event-parse.c
@@ -2988,8 +2988,8 @@ pretty_print_func_graph(void *data, int size, struct event *event,
printf("\n");
}
-void print_event(int cpu, void *data, int size, unsigned long long nsecs,
- char *comm)
+void print_tracepoint_event(int cpu, void *data, int size,
+ unsigned long long nsecs, char *comm)
{
struct event *event;
unsigned long secs;
diff --git a/tools/perf/util/trace-event.h b/tools/perf/util/trace-event.h
index 632b87b1..cafabf3 100644
--- a/tools/perf/util/trace-event.h
+++ b/tools/perf/util/trace-event.h
@@ -177,8 +177,8 @@ void print_printk(void);
int parse_ftrace_file(char *buf, unsigned long size);
int parse_event_file(char *buf, unsigned long size, char *sys);
-void print_event(int cpu, void *data, int size, unsigned long long nsecs,
- char *comm);
+void print_tracepoint_event(int cpu, void *data, int size,
+ unsigned long long nsecs, char *comm);
extern int file_bigendian;
extern int host_bigendian;
Frederic Weisbecker
> Prepare for handling of samples for any event type.
>
> Signed-off-by: David Ahern <daa...@cisco.com>
> ---
> tools/perf/builtin-script.c | 40 +++++++++++--------
> .../util/scripting-engines/trace-event-python.c | 20 ++++++++-
What about Perl?
> tools/perf/util/trace-event-scripting.c | 8 +--
> tools/perf/util/trace-event.h | 6 ++-
> 4 files changed, 47 insertions(+), 27 deletions(-)
>
> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> index 5f40df6..0bee150 100644
> --- a/tools/perf/builtin-script.c
> +++ b/tools/perf/builtin-script.c
> @@ -20,6 +20,27 @@ static u64 last_timestamp;
> static u64 nr_unordered;
> extern const struct option record_options[];
>
> +static void process_event(union perf_event *event,
> + struct perf_sample *sample,
> + struct perf_session *session)
> +{
> + struct thread *thread = perf_session__findnew(session, event->ip.pid);
> +
> + if (thread == NULL) {
> + pr_debug("problem processing %d event, skipping it.\n",
> + event->header.type);
> + return;
> + }
Seems the thread is needed by any endpoints. It would be better to resolve
it from process_sample_event and pass it to the process_event() handler.
> +
> + /*
> + * FIXME: better resolve from pid from the struct trace_entry
> + * field, although it should be the same than this perf
> + * event pid
> + */
> + print_event(sample->cpu, sample->raw_data, sample->raw_size,
> + sample->time, thread->comm);
> +}
[...]
> diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
> index 2040b85..5b03fb6 100644
> --- a/tools/perf/util/scripting-engines/trace-event-python.c
> +++ b/tools/perf/util/scripting-engines/trace-event-python.c
> @@ -204,9 +204,9 @@ static inline struct event *find_cache_event(int type)
> return event;
> }
>
> -static void python_process_event(int cpu, void *data,
> - int size __unused,
> - unsigned long long nsecs, char *comm)
> +static void python_process_event(union perf_event *pevent,
> + struct perf_sample *sample,
> + struct perf_session *session)
> {
> PyObject *handler, *retval, *context, *t, *obj, *dict = NULL;
> static char handler_name[256];
> @@ -218,6 +218,20 @@ static void python_process_event(int cpu, void *data,
> int type;
> int pid;
Please avoid such blank line in the middle of local vars declaration.
> + int cpu = sample->cpu;
> + void *data = sample->raw_data;
> + unsigned long long nsecs = sample->time;
> + char *comm;
> + struct thread *thread;
> +
> + thread = perf_session__findnew(session, pevent->ip.pid);
> + if (thread == NULL) {
> + pr_debug("problem processing %d event, skipping it.\n",
> + pevent->header.type);
> + return;
> + }
> + comm = thread->comm;
> +
Frederic Weisbecker
> @@ -763,6 +783,18 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
> exit(-1);
> }
>
> + if (no_callchain)
> + symbol_conf.use_callchain = false;
> +
> + else {
> + symbol_conf.use_callchain = true;
> + if (callchain_register_param(&callchain_param) < 0) {
That call doesn't seem needed. Register callchain params is only useful
for later callchain sorting.
> + error("Can't register callchain params\n");
> + err = -EINVAL;
> + goto out;
> + }
> + }
> +
> if (rec_script_path)
> script_path = rec_script_path;
> if (rep_script_path)
[...]
Hmm, that's a wrong way of walking through callchains. In fact it's not
a classical list. node->next can be a ghost entry from a previous callchain
that we kept cached in order to optimize allocations.
You need the accessors callchain_cursor_current() and callchain_cursor_advance().
David Ahern
On 03/02/2011 07:40 PM, Frederic Weisbecker wrote:
> On Wed, Mar 02, 2011 at 10:29:18AM -0700, David Ahern wrote:
>> Prepare for handling of samples for any event type.
>>
>> Signed-off-by: David Ahern <daa...@cisco.com>
>> ---
>> tools/perf/builtin-script.c | 40 +++++++++++--------
>> .../util/scripting-engines/trace-event-python.c | 20 ++++++++-
>
> What about Perl?
Oversight. Stumbled onto it yesterday afternoon. I have the perl module
installed now, so builds will get it too from now on.
>
>> tools/perf/util/trace-event-scripting.c | 8 +--
>> tools/perf/util/trace-event.h | 6 ++-
>> 4 files changed, 47 insertions(+), 27 deletions(-)
>>
>> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
>> index 5f40df6..0bee150 100644
>> --- a/tools/perf/builtin-script.c
>> +++ b/tools/perf/builtin-script.c
>> @@ -20,6 +20,27 @@ static u64 last_timestamp;
>> static u64 nr_unordered;
>> extern const struct option record_options[];
>>
>> +static void process_event(union perf_event *event,
>> + struct perf_sample *sample,
>> + struct perf_session *session)
>> +{
>> + struct thread *thread = perf_session__findnew(session, event->ip.pid);
>> +
>> + if (thread == NULL) {
>> + pr_debug("problem processing %d event, skipping it.\n",
>> + event->header.type);
>> + return;
>> + }
>
> Seems the thread is needed by any endpoints. It would be better to resolve
> it from process_sample_event and pass it to the process_event() handler.
tracepoints yes; S/W samples no. If you want it I'll add it.
ok.
David
David Ahern
On 03/02/2011 08:05 PM, Frederic Weisbecker wrote:
> On Wed, Mar 02, 2011 at 10:29:20AM -0700, David Ahern wrote:
>> @@ -763,6 +783,18 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
>> exit(-1);
>> }
>>
>> + if (no_callchain)
>> + symbol_conf.use_callchain = false;
>> +
>> + else {
>> + symbol_conf.use_callchain = true;
>> + if (callchain_register_param(&callchain_param) < 0) {
>
> That call doesn't seem needed. Register callchain params is only useful
> for later callchain sorting.
ok. I see that now. I'll remove.
Ok. I'll look at perf-report's callchain_append again.
David
Frederic Weisbecker
>
>
> On 03/02/2011 07:40 PM, Frederic Weisbecker wrote:
> > Seems the thread is needed by any endpoints. It would be better to resolve
> > it from process_sample_event and pass it to the process_event() handler.
>
> tracepoints yes; S/W samples no. If you want it I'll add it.
Hmm, yeah that looks more sensible, the majority of users seem to need it.
Thanks.
Frederic Weisbecker
Yeah callchain_append() is too much a generic name for something actually
rather specific. In fact callchain_append() adds a callchain, already
resolved, to a histogram in order to produce that statistical tree in the end
that you can have with perf report.
But you don't need those statistical tree of callchains, it's only used
by perf report for now. Instead you rather need to treat every callchains
individually and print each of them.
So you don't even need callchain_append(). All you need in the end is
to use perf_session__resolve_callchain() that resolves the raw struct ip_callchain
(only made of raw ips) into a cursor (list of ips resolved into symbols and so) and
walk through the cursor with the two accessors.
Ah I forgot, you first need to use callchain_cursor_commit() in order to initialize
the position in the cursor.
So:
1) Resolve with perf_session__resolve_callchain()
2) commit with callchain_cursor_commit()
3) iterate with callchain_cursor_current(), callchain_cursor_advance()
Thanks.
David Ahern
On 03/03/11 10:19, Frederic Weisbecker wrote:
> On Thu, Mar 03, 2011 at 07:20:10AM -0700, David Ahern wrote:
>> On 03/02/2011 08:05 PM, Frederic Weisbecker wrote:
>>> Hmm, that's a wrong way of walking through callchains. In fact it's not
>>> a classical list. node->next can be a ghost entry from a previous callchain
>>> that we kept cached in order to optimize allocations.
>>>
>>> You need the accessors callchain_cursor_current() and callchain_cursor_advance().
>>
>> Ok. I'll look at perf-report's callchain_append again.
>
> Yeah callchain_append() is too much a generic name for something actually
> rather specific. In fact callchain_append() adds a callchain, already
> resolved, to a histogram in order to produce that statistical tree in the end
> that you can have with perf report.
>
> But you don't need those statistical tree of callchains, it's only used
> by perf report for now. Instead you rather need to treat every callchains
> individually and print each of them.
>
> So you don't even need callchain_append(). All you need in the end is
Right, just using that as an example of how callchains are handled.
> to use perf_session__resolve_callchain() that resolves the raw struct ip_callchain
> (only made of raw ips) into a cursor (list of ips resolved into symbols and so) and
> walk through the cursor with the two accessors.
>
> Ah I forgot, you first need to use callchain_cursor_commit() in order to initialize
> the position in the cursor.
>
> So:
>
> 1) Resolve with perf_session__resolve_callchain()
> 2) commit with callchain_cursor_commit()
> 3) iterate with callchain_cursor_current(), callchain_cursor_advance()
yes, I figured out the missing commit, and I changed the loop to:
while (1) {
node = callchain_cursor_current(cursor);
if (!node)
break;
... (print chain)
callchain_cursor_advance(cursor);
}
Thanks for the comments.
What about the python and perl engines? Right now they are tracepoint
specific. I do not have a sufficient background in either to expand to
other sample types.
David
Frederic Weisbecker
> On 03/03/11 10:19, Frederic Weisbecker wrote:
> > to use perf_session__resolve_callchain() that resolves the raw struct ip_callchain
> > (only made of raw ips) into a cursor (list of ips resolved into symbols and so) and
> > walk through the cursor with the two accessors.
> >
> > Ah I forgot, you first need to use callchain_cursor_commit() in order to initialize
> > the position in the cursor.
> >
> > So:
> >
> > 1) Resolve with perf_session__resolve_callchain()
> > 2) commit with callchain_cursor_commit()
> > 3) iterate with callchain_cursor_current(), callchain_cursor_advance()
>
> yes, I figured out the missing commit, and I changed the loop to:
> while (1) {
> node = callchain_cursor_current(cursor);
> if (!node)
> break;
>
> ... (print chain)
>
> callchain_cursor_advance(cursor);
> }
Yep.
>
> Thanks for the comments.
>
> What about the python and perl engines? Right now they are tracepoint
> specific. I do not have a sufficient background in either to expand to
> other sample types.
Not a problem, the new parameters layout paves the way to make it later feasible
if someone is interested.
David Ahern
Signed-off-by: David Ahern <daa...@cisco.com>
---
tools/perf/builtin-script.c | 35 ++++++++++++++-----
.../perf/util/scripting-engines/trace-event-perl.c | 12 +++++--
.../util/scripting-engines/trace-event-python.c | 12 +++++--
tools/perf/util/trace-event-scripting.c | 10 +++---
tools/perf/util/trace-event.h | 8 +++-
5 files changed, 55 insertions(+), 22 deletions(-)
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 5f40df6..48bdf94 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -20,6 +20,21 @@ static u64 last_timestamp;
static u64 nr_unordered;
extern const struct option record_options[];
+static void process_event(union perf_event *event __unused,
+ struct perf_sample *sample,
+ struct perf_session *session __unused,
+ struct perf_event_attr *attr __unused,
+ struct thread *thread)
+{
+ /*
+ * FIXME: better resolve from pid from the struct trace_entry
+ * field, although it should be the same than this perf
+ * event pid
+ */
+ print_event(sample->cpu, sample->raw_data, sample->raw_size,
+ sample->time, thread->comm);
+}
+
static int default_start_script(const char *script __unused,
int argc __unused,
const char **argv __unused)
@@ -40,7 +55,7 @@ static int default_generate_script(const char *outfile __unused)
static struct scripting_ops default_scripting_ops = {
.start_script = default_start_script,
.stop_script = default_stop_script,
- .process_event = print_event,
+ .process_event = process_event,
.generate_script = default_generate_script,
};
@@ -67,6 +82,7 @@ static int process_sample_event(union perf_event *event,
struct perf_sample *sample,
struct perf_session *session)
{
+ struct perf_event_attr *attr;
struct thread *thread = perf_session__findnew(session, event->ip.pid);
if (thread == NULL) {
@@ -75,6 +91,13 @@ static int process_sample_event(union perf_event *event,
return -1;
}
+ attr = perf_header__find_attr(sample->id, &session->header);
+ if (!attr) {
+ pr_debug("No attributes found for sample id %" PRId64 ". "
+ "skipping it.\n", sample->id);
+ return -1;
+ }
+
if (session->sample_type & PERF_SAMPLE_RAW) {
if (debug_mode) {
if (sample->time < last_timestamp) {
@@ -86,14 +109,8 @@ static int process_sample_event(union perf_event *event,
last_timestamp = sample->time;
return 0;
}
- /*
- * 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(sample->cpu, sample->raw_data,
- sample->raw_size,
- sample->time, thread->comm);
+ scripting_ops->process_event(event, sample, session,
+ attr, thread);
}
session->hists.stats.total_period += sample->period;
diff --git a/tools/perf/util/scripting-engines/trace-event-perl.c b/tools/perf/util/scripting-engines/trace-event-perl.c
index 9368081..ed69d4c 100644
--- a/tools/perf/util/scripting-engines/trace-event-perl.c
+++ b/tools/perf/util/scripting-engines/trace-event-perl.c
@@ -245,9 +245,11 @@ static inline struct event *find_cache_event(int type)
return event;
}
-static void perl_process_event(int cpu, void *data,
- int size __unused,
- unsigned long long nsecs, char *comm)
+static void perl_process_event(union perf_event *pevent __unused,
+ struct perf_sample *sample,
+ struct perf_session *session __unused,
+ struct perf_event_attr *attr __unused,
+ struct thread *thread)
{
struct format_field *field;
static char handler[256];
@@ -256,6 +258,10 @@ static void perl_process_event(int cpu, void *data,
struct event *event;
int type;
int pid;
+ int cpu = sample->cpu;
+ void *data = sample->raw_data;
+ unsigned long long nsecs = sample->time;
+ char *comm = thread->comm;
dSP;
diff --git a/tools/perf/util/scripting-engines/trace-event-python.c b/tools/perf/util/scripting-engines/trace-event-python.c
index 2040b85..9fded61 100644
--- a/tools/perf/util/scripting-engines/trace-event-python.c
+++ b/tools/perf/util/scripting-engines/trace-event-python.c
@@ -204,9 +204,11 @@ static inline struct event *find_cache_event(int type)
return event;
}
-static void python_process_event(int cpu, void *data,
- int size __unused,
- unsigned long long nsecs, char *comm)
+static void python_process_event(union perf_event *pevent __unused,
+ struct perf_sample *sample,
+ struct perf_session *session __unused,
+ struct perf_event_attr *attr __unused,
+ struct thread *thread)
{
PyObject *handler, *retval, *context, *t, *obj, *dict = NULL;
static char handler_name[256];
@@ -217,6 +219,10 @@ static void python_process_event(int cpu, void *data,
unsigned n = 0;
int type;
int pid;
+ int cpu = sample->cpu;
+ void *data = sample->raw_data;
+ unsigned long long nsecs = sample->time;
+ char *comm = thread->comm;
t = PyTuple_New(MAX_FIELDS);
if (!t)
diff --git a/tools/perf/util/trace-event-scripting.c b/tools/perf/util/trace-event-scripting.c
index f7af2fc..6f91d3c 100644
--- a/tools/perf/util/trace-event-scripting.c
+++ b/tools/perf/util/trace-event-scripting.c
@@ -36,11 +36,11 @@ static int stop_script_unsupported(void)
return 0;
}
-static void process_event_unsupported(int cpu __unused,
- void *data __unused,
- int size __unused,
- unsigned long long nsecs __unused,
- char *comm __unused)
+static void process_event_unsupported(union perf_event *event __unused,
+ struct perf_sample *sample __unused,
+ struct perf_session *session __unused,
+ struct perf_event_attr *attr __unused,
+ struct thread *thread __unused)
{
}
diff --git a/tools/perf/util/trace-event.h b/tools/perf/util/trace-event.h
index b5f12ca..b4b939d 100644
--- a/tools/perf/util/trace-event.h
+++ b/tools/perf/util/trace-event.h
@@ -3,6 +3,7 @@
#include <stdbool.h>
#include "parse-events.h"
+#include "session.h"
#define __unused __attribute__((unused))
@@ -278,8 +279,11 @@ struct scripting_ops {
const char *name;
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);
+ void (*process_event) (union perf_event *event,
+ struct perf_sample *sample,
+ struct perf_session *session,
+ struct perf_event_attr *attr,
+ struct thread *thread);
int (*generate_script) (const char *outfile);
};
--
1.7.4
--
David Ahern
- fixed process_event handler in trace-event-perl
- removed unneeded call to callchain_register_param
- corrected walking of callchains in perf_session__print_symbols
- complete rewrite for handling non-tracepoint events
David Ahern (2):
perf script: change process_event prototype
perf script: support custom field selection for output
tools/perf/Documentation/perf-script.txt | 24 ++
tools/perf/builtin-script.c | 234 ++++++++++++++++++--
tools/perf/util/output.h | 18 ++
.../perf/util/scripting-engines/trace-event-perl.c | 12 +-
.../util/scripting-engines/trace-event-python.c | 12 +-
tools/perf/util/session.c | 75 +++++++
tools/perf/util/session.h | 5 +
tools/perf/util/trace-event-parse.c | 71 +++++--
tools/perf/util/trace-event-scripting.c | 10 +-
tools/perf/util/trace-event.h | 8 +-
10 files changed, 420 insertions(+), 49 deletions(-)
create mode 100644 tools/perf/util/output.h
David Ahern
Options include comm (command name), tid (thread id), pid (process id),
time (perf timestamp), cpu, event (for event name), trace (for trace
data), and sym (for IP, symbol name and DSO name).
This options allows dumping software events with symbols -- my primary
goal -- and also symbols for trace events if a so desired.
Default is set to maintain compatibility with current output.
Thanks to Frederic Weisbecker for detailed suggestions on this approach.
Examples (output compressed)
1. trace, default format
perf record -ga -e sched:sched_switch
perf script
swapper-0 [000] 534.955086: sched_switch: prev_comm=swapper prev_pid=0 ...
sshd-1678 [000] 534.955093: sched_switch: prev_comm=sshd prev_pid=1678 ...
sshd-1675 [001] 534.960353: sched_switch: prev_comm=sshd prev_pid=1675 ...
2. trace, custom format
perf record -ga -e sched:sched_switch
perf script -f comm,pid,time,trace,sym
swapper-0 534.955086: prev_comm=swapper prev_pid=0 prev_prio=120 ...
ffffffff81030350 perf_trace_sched_switch ([kernel.kallsyms])
ffffffff81382ac5 schedule ([kernel.kallsyms])
ffffffff8100134a cpu_idle ([kernel.kallsyms])
ffffffff81370b39 rest_init ([kernel.kallsyms])
ffffffff81696c23 start_kernel ([kernel.kallsyms].init.text)
ffffffff816962af x86_64_start_reservations ([kernel.kallsyms].init.text)
ffffffff816963b9 x86_64_start_kernel ([kernel.kallsyms].init.text)
sshd-1678 534.955093: prev_comm=sshd prev_pid=1678 prev_prio=120 ...
ffffffff81030350 perf_trace_sched_switch ([kernel.kallsyms])
ffffffff81382ac5 schedule ([kernel.kallsyms])
ffffffff81382fa0 schedule_timeout ([kernel.kallsyms])
ffffffff81363f16 unix_stream_recvmsg ([kernel.kallsyms])
ffffffff812e35b4 __sock_recvmsg ([kernel.kallsyms])
ffffffff812e368b sock_aio_read ([kernel.kallsyms])
ffffffff810fd2f2 do_sync_read ([kernel.kallsyms])
ffffffff810fd978 vfs_read ([kernel.kallsyms])
ffffffff810fda02 sys_read ([kernel.kallsyms])
ffffffff81002bc2 system_call ([kernel.kallsyms])
7f1647e51210 __GI___libc_read (/lib64/libc-2.12.90.so)
sshd-1675 534.960353: prev_comm=sshd prev_pid=1675 prev_prio=120 ...
ffffffff81030350 perf_trace_sched_switch ([kernel.kallsyms])
ffffffff81382ac5 schedule ([kernel.kallsyms])
ffffffff81382fa0 schedule_timeout ([kernel.kallsyms])
ffffffff81363f16 unix_stream_recvmsg ([kernel.kallsyms])
ffffffff812e35b4 __sock_recvmsg ([kernel.kallsyms])
ffffffff812e368b sock_aio_read ([kernel.kallsyms])
ffffffff810fd2f2 do_sync_read ([kernel.kallsyms])
ffffffff810fd978 vfs_read ([kernel.kallsyms])
ffffffff810fda02 sys_read ([kernel.kallsyms])
ffffffff81002bc2 system_call ([kernel.kallsyms])
7f1647e51210 __GI___libc_read (/lib64/libc-2.12.90.so)
3. Software events
perf record -e cs -c 1 -gaT
perf script -f comm,tid,pid,time,sym
sshd 1015/1015 328.654782
ffffffff810355de perf_event_task_sched_out ([kernel.kallsyms])
ffffffff81382aa1 schedule ([kernel.kallsyms])
ffffffff813837aa schedule_hrtimeout_range_clock ([kernel.kallsyms])
ffffffff81383886 schedule_hrtimeout_range ([kernel.kallsyms])
ffffffff8110c4f9 poll_schedule_timeout ([kernel.kallsyms])
ffffffff8110cd20 do_select ([kernel.kallsyms])
ffffffff8110ced8 core_sys_select ([kernel.kallsyms])
ffffffff8110d00d sys_select ([kernel.kallsyms])
ffffffff81002bc2 system_call ([kernel.kallsyms])
7fd685d30e93 __GI_select (/lib64/libc-2.12.90.so)
sshd 1629/1629 328.655460
ffffffff810355de perf_event_task_sched_out ([kernel.kallsyms])
ffffffff81382aa1 schedule ([kernel.kallsyms])
ffffffff8103d4d7 __cond_resched ([kernel.kallsyms])
ffffffff81382cc7 _cond_resched ([kernel.kallsyms])
ffffffff81382d0a wait_for_common ([kernel.kallsyms])
ffffffff81382e90 wait_for_completion ([kernel.kallsyms])
ffffffff8108092e stop_one_cpu ([kernel.kallsyms])
ffffffff8103b41d sched_exec ([kernel.kallsyms])
ffffffff811035e0 do_execve ([kernel.kallsyms])
ffffffff8100a5bd sys_execve ([kernel.kallsyms])
ffffffff8100301c stub_execve ([kernel.kallsyms])
7fd685d01f67 __execve (/lib64/libc-2.12.90.so)
4. generate a process list from a recorded session:
perf script -f comm,pid
Signed-off-by: David Ahern <daa...@cisco.com>
---
tools/perf/Documentation/perf-script.txt | 24 ++++
tools/perf/builtin-script.c | 207 +++++++++++++++++++++++++++---
tools/perf/util/output.h | 18 +++
tools/perf/util/session.c | 75 +++++++++++
tools/perf/util/session.h | 5 +
tools/perf/util/trace-event-parse.c | 71 ++++++++--
6 files changed, 369 insertions(+), 31 deletions(-)
create mode 100644 tools/perf/util/output.h
diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt
index 29ad942..f527287 100644
--- a/tools/perf/Documentation/perf-script.txt
+++ b/tools/perf/Documentation/perf-script.txt
@@ -112,6 +112,30 @@ OPTIONS
--debug-mode::
Do various checks like samples ordering and lost events.
+-k::
+--vmlinux=<file>::
+ vmlinux pathname
+
+--kallsyms=<file>::
+ kallsyms pathname
+
+--symfs=<directory>::
+ Look for files with symbols relative to this directory.
+
+-U::
+--show-unresolved::
+ Display all addresses including unresolved to a symbol.
+
+-G::
+--hide-call-graph::
+ When printing symbols do not display call chain.
+
+-f::
+--fields
+ Comma separated list of fields to print. Options are:
+ comm, tid, pid, time, cpu, sym, trace, event
+
+--
SEE ALSO
--------
linkperf:perf-record[1], linkperf:perf-script-perl[1],
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 48bdf94..c6b9db1 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -12,6 +12,7 @@
#include "util/trace-event.h"
#include "util/parse-options.h"
#include "util/util.h"
+#include "util/output.h"
static char const *script_name;
static char const *generate_script_lang;
@@ -19,6 +20,81 @@ static bool debug_mode;
static u64 last_timestamp;
static u64 nr_unordered;
extern const struct option record_options[];
+static bool show_unresolved;
+static bool no_callchain;
+
+struct output_option {
+ const char *str;
+ enum perf_output_field field;
+} all_output_options [] = {
+ {.str = "comm", .field = PERF_OUTPUT_COMM},
+ {.str = "tid", .field = PERF_OUTPUT_TID},
+ {.str = "pid", .field = PERF_OUTPUT_PID},
+ {.str = "time", .field = PERF_OUTPUT_TIME},
+ {.str = "cpu", .field = PERF_OUTPUT_CPU},
+ {.str = "sym", .field = PERF_OUTPUT_SYM},
+ {.str = "trace", .field = PERF_OUTPUT_TRACE},
+ {.str = "event", .field = PERF_OUTPUT_EVNAME},
+};
+
+/* default set to maintain compatibility with current format */
+#define output_fields_default (PERF_OUTPUT_COMM | PERF_OUTPUT_PID | \
+ PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | \
+ PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE)
+
+u64 output_fields = output_fields_default;
+
+static void print_sample_oneline(struct perf_sample *sample,
+ struct perf_event_attr *attr,
+ struct thread *thread)
+{
+ const char *evname;
+
+ if (PRINT_FIELD(COMM)) {
+ int n = 16;
+ if (PRINT_FIELD(SYM) && !no_callchain)
+ n = 1;
+ printf("%*s ", n, thread->comm);
+ }
+
+ if (PRINT_FIELD(PID) && PRINT_FIELD(TID))
+ printf("%5d/%-5d ", sample->pid, sample->tid);
+ else if (PRINT_FIELD(PID))
+ printf("%5d ", sample->pid);
+ else if (PRINT_FIELD(TID))
+ printf("%5d ", sample->tid);
+
+ if (PRINT_FIELD(CPU))
+ printf("[%03d] ", sample->cpu);
+
+ if (PRINT_FIELD(TIME)) {
+ unsigned long secs, usecs, nsecs;
+
+ nsecs = sample->time;
+ secs = nsecs / NSECS_PER_SEC;
+ nsecs -= secs * NSECS_PER_SEC;
+ usecs = nsecs / NSECS_PER_USEC;
+ printf("%8lu.%06lu ", secs, usecs);
+ }
+
+ if (PRINT_FIELD(EVNAME)) {
+ if (attr->type == PERF_TYPE_TRACEPOINT) {
+ struct event *event;
+ int type;
+
+ type = trace_parse_common_type(sample->raw_data);
+ event = trace_find_event(type);
+ if (event)
+ evname = event->name;
+ } else
+ evname = __event_name(attr->type, attr->config);
+
+ printf("%s ", evname ? evname : "(unknown)");
+ }
+
+ if (!no_callchain)
+ printf("\n");
+}
static void process_event(union perf_event *event __unused,
struct perf_sample *sample,
@@ -31,8 +107,15 @@ static void process_event(union perf_event *event __unused,
* field, although it should be the same than this perf
* event pid
*/
- print_event(sample->cpu, sample->raw_data, sample->raw_size,
- sample->time, thread->comm);
+ if (PRINT_FIELD(TRACE))
+ print_event(sample->cpu, sample->raw_data,
+ sample->raw_size, sample->time,
+ thread->comm);
+ else
+ print_sample_oneline(sample, attr, thread);
+
+ if (PRINT_FIELD(SYM))
+ perf_session__print_symbols(event, sample, session, show_unresolved);
}
static int default_start_script(const char *script __unused,
@@ -98,20 +181,17 @@ static int process_sample_event(union perf_event *event,
return -1;
}
- if (session->sample_type & PERF_SAMPLE_RAW) {
- if (debug_mode) {
- if (sample->time < last_timestamp) {
- pr_err("Samples misordered, previous: %" PRIu64
- " this: %" PRIu64 "\n", last_timestamp,
- sample->time);
- nr_unordered++;
- }
- last_timestamp = sample->time;
- return 0;
+ if (debug_mode) {
+ if (sample->time < last_timestamp) {
+ pr_err("Samples misordered, previous: %" PRIu64
+ " this: %" PRIu64 "\n", last_timestamp,
+ sample->time);
+ nr_unordered++;
}
- scripting_ops->process_event(event, sample, session,
- attr, thread);
+ last_timestamp = sample->time;
+ return 0;
}
+ scripting_ops->process_event(event, sample, session, attr, thread);
session->hists.stats.total_period += sample->period;
return 0;
@@ -119,7 +199,10 @@ static int process_sample_event(union perf_event *event,
static struct perf_event_ops event_ops = {
.sample = process_sample_event,
+ .mmap = perf_event__process_mmap,
.comm = perf_event__process_comm,
+ .exit = perf_event__process_task,
+ .fork = perf_event__process_task,
.attr = perf_event__process_attr,
.event_type = perf_event__process_event_type,
.tracing_data = perf_event__process_tracing_data,
@@ -149,6 +232,39 @@ static int __cmd_script(struct perf_session *session)
return ret;
}
+static int perf_session__check_attr(struct perf_session *session)
+{
+ if ((output_fields & PERF_OUTPUT_TRACE) &&
+ !perf_session__has_traces(session, "record -R"))
+ return -EINVAL;
+
+ if ((output_fields & PERF_OUTPUT_SYM) &&
+ !(session->sample_type & PERF_SAMPLE_IP)) {
+ pr_err("Samples do not contain IP data.\n");
+ return -EINVAL;
+ }
+
+ if ((output_fields & (PERF_OUTPUT_TID | PERF_OUTPUT_PID)) &&
+ !(session->sample_type & PERF_SAMPLE_TID)) {
+ pr_err("Samples do not contain TID/PID data.\n");
+ return -EINVAL;
+ }
+
+ if ((output_fields & PERF_OUTPUT_TIME) &&
+ !(session->sample_type & PERF_SAMPLE_TIME)) {
+ pr_err("Samples do not contain timestamps.\n");
+ return -EINVAL;
+ }
+
+ if ((output_fields & PERF_OUTPUT_CPU) &&
+ !(session->sample_type & PERF_SAMPLE_CPU)) {
+ pr_err("Samples do not contain cpu.\n");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
struct script_spec {
struct list_head node;
struct scripting_ops *ops;
@@ -297,6 +413,43 @@ static int parse_scriptname(const struct option *opt __used,
return 0;
}
+static int parse_output_fields(const struct option *opt __used,
+ const char *arg, int unset __used)
+{
+ char *tok;
+ int i, imax = sizeof(all_output_options) / sizeof(struct output_option);
+ int rc = 0;
+ char *str = strdup(arg);
+
+ if (!str)
+ return -ENOMEM;
+
+ /* remove default */
+ output_fields = 0;
+
+ tok = strtok(str, ",");
+ while (1) {
+ for (i = 0; i < imax; ++i) {
+ if (strcmp(tok, all_output_options[i].str) == 0) {
+ output_fields |= all_output_options[i].field;
+ break;
+ }
+ }
+ if (i == imax) {
+ fprintf(stderr, "Invalid field requested.");
+ rc = -EINVAL;
+ break;
+ }
+
+ tok = strtok(NULL, ",");
+ if (!tok)
+ break;
+ }
+
+ free(str);
+ return rc;
+}
+
/* Helper function for filesystems that return a dent->d_type DT_UNKNOWN */
static int is_directory(const char *base_path, const struct dirent *dent)
{
@@ -609,6 +762,19 @@ static const struct option options[] = {
"input file name"),
OPT_BOOLEAN('d', "debug-mode", &debug_mode,
"do various checks like samples ordering and lost events"),
+ OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
+ "file", "vmlinux pathname"),
+ OPT_STRING(0, "kallsyms", &symbol_conf.kallsyms_name,
+ "file", "kallsyms pathname"),
+ OPT_BOOLEAN('G', "hide-call-graph", &no_callchain,
+ "When printing symbols do not display call chain"),
+ OPT_BOOLEAN('U', "show-unresolved", &show_unresolved,
+ "Display all entries including unresolved to a symbol"),
+ OPT_STRING(0, "symfs", &symbol_conf.symfs, "directory",
+ "Look for files with symbols relative to this directory"),
+ OPT_CALLBACK('f', "fields", NULL, "str",
+ "comma separated output fields. Options: comm,tid,pid,time,cpu,sym,trace,event",
+ parse_output_fields),
OPT_END()
};
@@ -752,6 +918,8 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
exit(-1);
}
+ symbol_conf.use_callchain = no_callchain ? false : true;
+
if (rec_script_path)
script_path = rec_script_path;
if (rep_script_path)
@@ -790,13 +958,20 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
return -ENOMEM;
if (strcmp(input_name, "-") &&
- !perf_session__has_traces(session, "record -R"))
+ (perf_session__check_attr(session) < 0))
return -EINVAL;
if (generate_script_lang) {
struct stat perf_stat;
+ int input;
+
+ if (output_fields != output_fields_default) {
+ fprintf(stderr,
+ "custom fields not supported for generated scripts");
+ return -1;
+ }
- int input = open(input_name, O_RDONLY);
+ input = open(input_name, O_RDONLY);
if (input < 0) {
perror("failed to open file");
exit(-1);
diff --git a/tools/perf/util/output.h b/tools/perf/util/output.h
new file mode 100644
index 0000000..9d79852
--- /dev/null
+++ b/tools/perf/util/output.h
@@ -0,0 +1,18 @@
+#ifndef _PERF_OUTPUT_H
+#define _PERF_OUTPUT_H
+
+enum perf_output_field {
+ PERF_OUTPUT_COMM = 1U << 0,
+ PERF_OUTPUT_TID = 1U << 1,
+ PERF_OUTPUT_PID = 1U << 2,
+ PERF_OUTPUT_TIME = 1U << 3,
+ PERF_OUTPUT_CPU = 1U << 4,
+ PERF_OUTPUT_SYM = 1U << 5,
+ PERF_OUTPUT_TRACE = 1U << 6,
+ PERF_OUTPUT_EVNAME = 1U << 7,
+};
+
+extern u64 output_fields;
+#define PRINT_FIELD(x) (output_fields & PERF_OUTPUT_##x)
+
+#endif
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index a3a871f..c862a71 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -10,6 +10,7 @@
#include "session.h"
#include "sort.h"
#include "util.h"
+#include "trace-event.h"
static int perf_session__open(struct perf_session *self, bool force)
{
@@ -1137,3 +1138,77 @@ size_t perf_session__fprintf_dsos_buildid(struct perf_session *self, FILE *fp,
size_t ret = machine__fprintf_dsos_buildid(&self->host_machine, fp, with_hits);
return ret + machines__fprintf_dsos_buildid(&self->machines, fp, with_hits);
}
+
+void perf_session__print_symbols(union perf_event *event,
+ struct perf_sample *sample,
+ struct perf_session *session,
+ bool show_unresolved)
+{
+ struct addr_location al;
+ const char *symname, *dsoname;
+ struct callchain_cursor *cursor = &session->callchain_cursor;
+ struct callchain_cursor_node *node;
+
+ if (perf_event__preprocess_sample(event, session, &al, sample,
+ NULL) < 0) {
+ error("problem processing %d event, skipping it.\n",
+ event->header.type);
+ return;
+ }
+
+ if (symbol_conf.use_callchain && sample->callchain) {
+
+ if (perf_session__resolve_callchain(session, al.thread,
+ sample->callchain, NULL) != 0) {
+ if (verbose)
+ error("Failed to resolve callchain. Skipping\n");
+ return;
+ }
+ callchain_cursor_commit(cursor);
+ if (cursor->nr == 0)
+ return;
+
+ while (1) {
+ node = callchain_cursor_current(cursor);
+ if (!node)
+ break;
+
+ if (node->sym && node->sym->name)
+ symname = node->sym->name;
+ else if (show_unresolved)
+ symname = "";
+ else
+ goto next;
+
+ if (node->map && node->map->dso && node->map->dso->name)
+ dsoname = node->map->dso->name;
+ else if (show_unresolved)
+ dsoname = "";
+ else
+ goto next;
+
+ printf("\t%16" PRIx64 " %s (%s)\n", node->ip, symname, dsoname);
+
+next:
+ callchain_cursor_advance(cursor);
+ }
+ printf("\n");
+
+ } else {
+ if (al.sym && al.sym->name)
+ symname = al.sym->name;
+ else if (show_unresolved)
+ symname = "";
+ else
+ return;
+
+ if (al.map && al.map->dso && al.map->dso->name)
+ dsoname = al.map->dso->name;
+ else if (show_unresolved)
+ dsoname = "";
+ else
+ return;
+
+ printf("%16" PRIx64 " %s (%s)\n", al.addr, symname, dsoname);
+ }
+}
diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h
index 977b3a1..d77a2fa 100644
--- a/tools/perf/util/session.h
+++ b/tools/perf/util/session.h
@@ -165,4 +165,9 @@ static inline int perf_session__parse_sample(struct perf_session *session,
session->sample_id_all, sample);
}
+void perf_session__print_symbols(union perf_event *event,
+ struct perf_sample *sample,
+ struct perf_session *session,
+ bool show_unresolved);
+
#endif /* __PERF_SESSION_H */
diff --git a/tools/perf/util/trace-event-parse.c b/tools/perf/util/trace-event-parse.c
index d8e622d..9ac6524 100644
--- a/tools/perf/util/trace-event-parse.c
+++ b/tools/perf/util/trace-event-parse.c
@@ -32,6 +32,7 @@
#include "../perf.h"
#include "util.h"
#include "trace-event.h"
+#include "output.h"
int header_page_ts_offset;
int header_page_ts_size;
@@ -2683,13 +2684,18 @@ static void print_graph_proc(int pid, const char *comm)
/* sign + log10(MAX_INT) + '\0' */
char pid_str[11];
int spaces = 0;
- int len;
+ int len = 0;
int i;
sprintf(pid_str, "%d", pid);
/* 1 stands for the "-" character */
- len = strlen(comm) + strlen(pid_str) + 1;
+ if (PRINT_FIELD(COMM) && PRINT_FIELD(PID))
+ len = strlen(comm) + strlen(pid_str) + 1;
+ else if (PRINT_FIELD(COMM))
+ len = strlen(comm);
+ else if (PRINT_FIELD(PID))
+ len = strlen(pid_str);
if (len < TRACE_GRAPH_PROCINFO_LENGTH)
spaces = TRACE_GRAPH_PROCINFO_LENGTH - len;
@@ -2698,7 +2704,12 @@ static void print_graph_proc(int pid, const char *comm)
for (i = 0; i < spaces / 2; i++)
printf(" ");
- printf("%s-%s", comm, pid_str);
+ if (PRINT_FIELD(COMM) && PRINT_FIELD(PID))
+ printf("%s-%s", comm, pid_str);
+ else if (PRINT_FIELD(COMM))
+ printf("%s", comm);
+ else if (PRINT_FIELD(PID))
+ printf("%s", pid_str);
/* Last spaces to align center */
for (i = 0; i < spaces - (spaces / 2); i++)
@@ -2884,10 +2895,14 @@ pretty_print_func_ent(void *data, int size, struct event *event,
void *copy_data;
unsigned long val;
- printf("%5lu.%06lu | ", secs, usecs);
+ if (PRINT_FIELD(TIME))
+ printf("%5lu.%06lu | ", secs, usecs);
+
+ if (PRINT_FIELD(CPU))
+ print_graph_cpu(cpu);
- print_graph_cpu(cpu);
- print_graph_proc(pid, comm);
+ if (PRINT_FIELD(COMM) || PRINT_FIELD(PID))
+ print_graph_proc(pid, comm);
printf(" | ");
@@ -2932,10 +2947,14 @@ pretty_print_func_ret(void *data, int size __unused, struct event *event,
struct format_field *field;
int i;
- printf("%5lu.%06lu | ", secs, usecs);
+ if (PRINT_FIELD(TIME))
+ printf("%5lu.%06lu | ", secs, usecs);
+
+ if (PRINT_FIELD(CPU))
+ print_graph_cpu(cpu);
- print_graph_cpu(cpu);
- print_graph_proc(pid, comm);
+ if (PRINT_FIELD(COMM) || PRINT_FIELD(PID))
+ print_graph_proc(pid, comm);
printf(" | ");
@@ -2989,7 +3008,7 @@ pretty_print_func_graph(void *data, int size, struct event *event,
}
void print_event(int cpu, void *data, int size, unsigned long long nsecs,
- char *comm)
+ char *comm)
{
struct event *event;
unsigned long secs;
@@ -3016,13 +3035,35 @@ void print_event(int cpu, void *data, int size, unsigned long long nsecs,
pid, comm, secs, usecs);
if (latency_format) {
- printf("%8.8s-%-5d %3d",
- comm, pid, cpu);
+ if (PRINT_FIELD(COMM) && PRINT_FIELD(PID))
+ printf("%8.8s-%-5d", comm, pid);
+ else if (PRINT_FIELD(COMM))
+ printf("%8.8s", comm);
+ else if (PRINT_FIELD(PID))
+ printf("%5d", pid);
+
+ if (PRINT_FIELD(CPU))
+ printf(" %3d", cpu);
+
print_lat_fmt(data, size);
- } else
- printf("%16s-%-5d [%03d]", comm, pid, cpu);
+ } else {
+ if (PRINT_FIELD(COMM) && PRINT_FIELD(PID))
+ printf("%16s-%-5d", comm, pid);
+ else if (PRINT_FIELD(COMM))
+ printf("%16s", comm);
+ else if (PRINT_FIELD(PID))
+ printf("%5d", pid);
+
+ if (PRINT_FIELD(CPU))
+ printf(" [%03d]", cpu);
+ }
- printf(" %5lu.%06lu: %s: ", secs, usecs, event->name);
+ if (PRINT_FIELD(TIME))
+ printf(" %5lu.%06lu:", secs, usecs);
+
+ if (PRINT_FIELD(EVNAME))
+ printf(" %s:", event->name);
+ printf(" ");
if (event->flags & EVENT_FL_FAILED) {
printf("EVENT '%s' FAILED TO PARSE\n",
Frederic Weisbecker
> Allow a user to select which fields to print to stdout for event data.
> Options include comm (command name), tid (thread id), pid (process id),
> time (perf timestamp), cpu, event (for event name), trace (for trace
> data), and sym (for IP, symbol name and DSO name).
>
> This options allows dumping software events with symbols -- my primary
> goal -- and also symbols for trace events if a so desired.
>
> Default is set to maintain compatibility with current output.
>
> Thanks to Frederic Weisbecker for detailed suggestions on this approach.
>
>
> Examples (output compressed)
>
> 1. trace, default format
>
> perf record -ga -e sched:sched_switch
> perf script
>
> swapper-0 [000] 534.955086: sched_switch: prev_comm=swapper prev_pid=0 ...
> sshd-1678 [000] 534.955093: sched_switch: prev_comm=sshd prev_pid=1678 ...
> sshd-1675 [001] 534.960353: sched_switch: prev_comm=sshd prev_pid=1675 ...
>
> 2. trace, custom format
>
> perf record -ga -e sched:sched_switch
> perf script -f comm,pid,time,trace,sym
Nice.
> diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt
> index 29ad942..f527287 100644
> --- a/tools/perf/Documentation/perf-script.txt
> +++ b/tools/perf/Documentation/perf-script.txt
> @@ -112,6 +112,30 @@ OPTIONS
> --debug-mode::
> Do various checks like samples ordering and lost events.
>
> +-k::
> +--vmlinux=<file>::
> + vmlinux pathname
> +
> +--kallsyms=<file>::
> + kallsyms pathname
> +
> +--symfs=<directory>::
> + Look for files with symbols relative to this directory.
> +
> +-U::
> +--show-unresolved::
> + Display all addresses including unresolved to a symbol.
We should always do it I think. As long as the sym format is asked, try
to resolve it, otherwise print the raw address.
Hm, we should have one default for tracepoint events and one
for others. For example dso and sym make sense for hardware,
breakpoint and software event, but it makes few sense for tracepoints.
The print_event() thing we have in trace-event-parse.c should really handle only
the raw trace itself. More on that below when I reach that file.
> + else
> + print_sample_oneline(sample, attr, thread);
> +
> + if (PRINT_FIELD(SYM))
> + perf_session__print_symbols(event, sample, session, show_unresolved);
> }
>
> struct script_spec {
> struct list_head node;
> struct scripting_ops *ops;
> @@ -297,6 +413,43 @@ static int parse_scriptname(const struct option *opt __used,
> return 0;
> }
[...]
I guess you don't need the above check.
For now we don't use that function because the function graph tracer
is not yet supported by perf.
Just forget about that, we'll take care of that later. Consider
pretty_print_func_graph() can't be called.
trace-event-parse.c and its print_event() pretty printing should not care
about the output format. It only needs to print the raw trace itself.
All the comm, time, etc... things should be handled from perf script
core.
In fact, print_event() should really be renamed print_trace() or something.
Other than that, looks like a right direction.
Thanks!
David Ahern
On 03/07/11 09:50, Frederic Weisbecker wrote:
>> +-U::
>> +--show-unresolved::
>> + Display all addresses including unresolved to a symbol.
>
> We should always do it I think. As long as the sym format is asked, try
> to resolve it, otherwise print the raw address.
Ok. I'll remove.
>> +/* default set to maintain compatibility with current format */
>> +#define output_fields_default (PERF_OUTPUT_COMM | PERF_OUTPUT_PID | \
>> + PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | \
>> + PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE)
>> +
>> +u64 output_fields = output_fields_default;
>
> Hm, we should have one default for tracepoint events and one
> for others. For example dso and sym make sense for hardware,
> breakpoint and software event, but it makes few sense for tracepoints.
That default was strictly to maintain backwards compatibility with
existing output. So you would like to see:
out_fields_def_trace = <the above default>
out_fields_def_sw = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | \
PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | \
PERF_OUTPUT_EVNAM | PERF_OUTPUT_SYM
If user does not specify custom option, set output_fields to default
based on event type - which conceptually can change sample to sample
(though perf currently can't handle a S/W and a trace event in the same
file).
As for H/W events, the cycles format should be the same as the S/W
format. After that perf-record does not mix well with H/W events.
>> static void process_event(union perf_event *event __unused,
>> struct perf_sample *sample,
>> @@ -31,8 +107,15 @@ static void process_event(union perf_event *event __unused,
>> * field, although it should be the same than this perf
>> * event pid
>> */
>> - print_event(sample->cpu, sample->raw_data, sample->raw_size,
>> - sample->time, thread->comm);
>> + if (PRINT_FIELD(TRACE))
>> + print_event(sample->cpu, sample->raw_data,
>> + sample->raw_size, sample->time,
>> + thread->comm);
>
> The print_event() thing we have in trace-event-parse.c should really handle only
> the raw trace itself. More on that below when I reach that file.
Ok
>> + callchain_cursor_commit(cursor);
>> + if (cursor->nr == 0)
>> + return;
>
> I guess you don't need the above check.
Ok. I'll remove.
That certainly helps make it simpler. I left the printing of those
fields in that file b/c of that function.
>
> trace-event-parse.c and its print_event() pretty printing should not care
> about the output format. It only needs to print the raw trace itself.
> All the comm, time, etc... things should be handled from perf script
> core.
>
> In fact, print_event() should really be renamed print_trace() or something.
Ok. I'll pull the comm, pid, time options into builtin-script's version
and rename the function.
David
Frederic Weisbecker
> On 03/07/11 09:50, Frederic Weisbecker wrote:
> >> +/* default set to maintain compatibility with current format */
> >> +#define output_fields_default (PERF_OUTPUT_COMM | PERF_OUTPUT_PID | \
> >> + PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | \
> >> + PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE)
> >> +
> >> +u64 output_fields = output_fields_default;
> >
> > Hm, we should have one default for tracepoint events and one
> > for others. For example dso and sym make sense for hardware,
> > breakpoint and software event, but it makes few sense for tracepoints.
>
> That default was strictly to maintain backwards compatibility with
> existing output.
Well, we don't necessarily need to keep a backward compatibility. If
one day we realize the comm is not needed on the traces output, we can
just remove it right away. We only want to keep the exisiting default
because it (seems to) make sense. Now the default necessary headers
that make sense for tracepoint may not be the same than for hardware
events.
> So you would like to see:
>
> out_fields_def_trace = <the above default>
>
> out_fields_def_sw = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | \
> PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | \
> PERF_OUTPUT_EVNAM | PERF_OUTPUT_SYM
Yep, I guess that would be sensible.
>
> If user does not specify custom option, set output_fields to default
> based on event type - which conceptually can change sample to sample
Exactly. Which means that a global output_fields wouldn't make
much sense anymore.
> (though perf currently can't handle a S/W and a trace event in the same
> file).
Yeah, but let's prepare the ground for that.
> As for H/W events, the cycles format should be the same as the S/W
> format.
Probably. Unless we find later that H/W event need some special treatment
over software ones.
> After that perf-record does not mix well with H/W events.
Hm?
Thanks!
David Ahern
- removed show-unresolved option
- setup different output defaults based on event type
- moved printing of comm, pid/tid, time and event name from print_event to
builtin-script and renamed print_event to print_trace_event
- removed unneeded cursor depth check
- updated to latest perf-core to get Arnaldo's recent makeovers
- created smaller, focused patches
v1 -> v2:
- fixed process_event handler in trace-event-perl
- removed unneeded call to callchain_register_param
- corrected walking of callchains in perf_session__print_symbols
- complete rewrite for handling non-tracepoint events
David Ahern (5):
perf script: change process_event prototype
perf script: move printing of 'common' data from print_event and
rename
perf script: support custom field selection for output
perf script: add support for dumping symbols
perf script: add support for H/W and S/W events
tools/perf/Documentation/perf-script.txt | 22 ++
tools/perf/builtin-script.c | 296 ++++++++++++++++++--
tools/perf/util/parse-events.c | 22 ++
tools/perf/util/parse-events.h | 1 +
.../perf/util/scripting-engines/trace-event-perl.c | 11 +-
.../util/scripting-engines/trace-event-python.c | 11 +-
tools/perf/util/session.c | 61 ++++
tools/perf/util/session.h | 4 +
tools/perf/util/trace-event-parse.c | 106 +-------
tools/perf/util/trace-event-scripting.c | 9 +-
tools/perf/util/trace-event.h | 10 +-
11 files changed, 416 insertions(+), 137 deletions(-)
--
1.7.4
David Ahern
Options include comm (command name), tid (thread id), pid (process id),
trace data).
Default is set to maintain compatibility with current output; this
feature does alter output format slightly -- no '-' between command
and pid/tid.
Thanks to Frederic Weisbecker for detailed suggestions on this approach.
Examples (output compressed)
1. trace, default format
perf record -ga -e sched:sched_switch
perf script
swapper 0 [000] 537.037184: sched_switch: prev_comm=swapper prev_pid=0...
sshd 1675 [000] 537.037309: sched_switch: prev_comm=sshd prev_pid=1675...
netstat 1692 [001] 537.038664: sched_switch: prev_comm=netstat prev_pid=1692...
2. trace, custom format
perf record -ga -e sched:sched_switch
perf script -f comm,pid,time,trace <--- omitting cpu and event name
swapper 0 537.037184: prev_comm=swapper prev_pid=0 prev_prio=120 ...
sshd 1675 537.037309: prev_comm=sshd prev_pid=1675 prev_prio=120 ...
netstat 1692 537.038664: prev_comm=netstat prev_pid=1692 prev_prio=120 ...
Signed-off-by: David Ahern <daa...@cisco.com>
---
tools/perf/Documentation/perf-script.txt | 5 +
tools/perf/builtin-script.c | 147 ++++++++++++++++++++++++++----
2 files changed, 133 insertions(+), 19 deletions(-)
diff --git a/tools/perf/Documentation/perf-script.txt b/tools/perf/Documentation/perf-script.txt
index 29ad942..b73cf58 100644
--- a/tools/perf/Documentation/perf-script.txt
+++ b/tools/perf/Documentation/perf-script.txt
@@ -112,6 +112,11 @@ OPTIONS
--debug-mode::
Do various checks like samples ordering and lost events.
+-f::
+--fields
+ Comma separated list of fields to print. Options are:
+ comm, tid, pid, time, cpu, event, trace
+
SEE ALSO
--------
linkperf:perf-record[1], linkperf:perf-script-perl[1],
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 0a79da2..e541059 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -20,6 +20,38 @@ static u64 last_timestamp;
static u64 nr_unordered;
extern const struct option record_options[];
+enum perf_output_field {
+ PERF_OUTPUT_COMM = 1U << 0,
+ PERF_OUTPUT_TID = 1U << 1,
+ PERF_OUTPUT_PID = 1U << 2,
+ PERF_OUTPUT_TIME = 1U << 3,
+ PERF_OUTPUT_CPU = 1U << 4,
+ PERF_OUTPUT_EVNAME = 1U << 5,
+ PERF_OUTPUT_TRACE = 1U << 6,
+};
+
+struct output_option {
+ const char *str;
+ enum perf_output_field field;
+} all_output_options [] = {
+ {.str = "comm", .field = PERF_OUTPUT_COMM},
+ {.str = "tid", .field = PERF_OUTPUT_TID},
+ {.str = "pid", .field = PERF_OUTPUT_PID},
+ {.str = "time", .field = PERF_OUTPUT_TIME},
+ {.str = "cpu", .field = PERF_OUTPUT_CPU},
+ {.str = "event", .field = PERF_OUTPUT_EVNAME},
+ {.str = "trace", .field = PERF_OUTPUT_TRACE},
+};
+
+/* default set to maintain compatibility with current format */
+static u64 output_fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID | \
+ PERF_OUTPUT_CPU | PERF_OUTPUT_TIME | \
+ PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE;
+
+static bool output_set_by_user;
+
+#define PRINT_FIELD(x) (output_fields & PERF_OUTPUT_##x)
+
static void print_sample_start(struct perf_sample *sample,
struct thread *thread)
{
@@ -28,24 +60,45 @@ static void print_sample_start(struct perf_sample *sample,
const char *evname = NULL;
unsigned long secs;
unsigned long usecs;
- unsigned long long nsecs = sample->time;
-
- if (latency_format)
- printf("%8.8s-%-5d %3d", thread->comm, sample->tid, sample->cpu);
- else
- printf("%16s-%-5d [%03d]", thread->comm, sample->tid, sample->cpu);
-
- secs = nsecs / NSECS_PER_SEC;
- nsecs -= secs * NSECS_PER_SEC;
- usecs = nsecs / NSECS_PER_USEC;
- printf(" %5lu.%06lu: ", secs, usecs);
+ unsigned long long nsecs;
- type = trace_parse_common_type(sample->raw_data);
- event = trace_find_event(type);
- if (event)
- evname = event->name;
-
- printf("%s: ", evname ? evname : "(unknown)");
+ if (PRINT_FIELD(COMM)) {
+ if (latency_format)
+ printf("%8.8s ", thread->comm);
+ else
+ printf("%16s ", thread->comm);
+ }
+
+ if (PRINT_FIELD(PID) && PRINT_FIELD(TID))
+ printf("%5d/%-5d ", sample->pid, sample->tid);
+ else if (PRINT_FIELD(PID))
+ printf("%5d ", sample->pid);
+ else if (PRINT_FIELD(TID))
+ printf("%5d ", sample->tid);
+
+ if (PRINT_FIELD(CPU)) {
+ if (latency_format)
+ printf("%3d ", sample->cpu);
+ else
+ printf("[%03d] ", sample->cpu);
+ }
+
+ if (PRINT_FIELD(TIME)) {
+ nsecs = sample->time;
+ secs = nsecs / NSECS_PER_SEC;
+ nsecs -= secs * NSECS_PER_SEC;
+ usecs = nsecs / NSECS_PER_USEC;
+ printf("%5lu.%06lu: ", secs, usecs);
+ }
+
+ if (PRINT_FIELD(EVNAME)) {
+ type = trace_parse_common_type(sample->raw_data);
+ event = trace_find_event(type);
+ if (event)
+ evname = event->name;
+
+ printf("%s: ", evname ? evname : "(unknown)");
+ }
}
static void process_event(union perf_event *event __unused,
@@ -54,7 +107,11 @@ static void process_event(union perf_event *event __unused,
struct thread *thread)
{
print_sample_start(sample, thread);
- print_trace_event(sample->cpu, sample->raw_data, sample->raw_size);
+
+ if (PRINT_FIELD(TRACE))
+ print_trace_event(sample->cpu, sample->raw_data,
+ sample->raw_size);
+
printf("\n");
}
@@ -311,6 +368,48 @@ static int parse_scriptname(const struct option *opt __used,
return 0;
}
+static int parse_output_fields(const struct option *opt __used,
+ const char *arg, int unset __used)
+{
+ char *tok;
+ int i, imax = sizeof(all_output_options) / sizeof(struct output_option);
+ int rc = 0;
+ char *str = strdup(arg);
+
+ if (!str)
+ return -ENOMEM;
+
+ tok = strtok(str, ",");
+ if (!tok) {
+ fprintf(stderr, "Invalid field string.");
+ return -EINVAL;
+ }
+
+ output_fields = 0;
+ while (1) {
+ for (i = 0; i < imax; ++i) {
+ if (strcmp(tok, all_output_options[i].str) == 0) {
+ output_fields |= all_output_options[i].field;
+ break;
+ }
+ }
+ if (i == imax) {
+ fprintf(stderr, "Invalid field requested.");
+ rc = -EINVAL;
+ break;
+ }
+
+ tok = strtok(NULL, ",");
+ if (!tok)
+ break;
+ }
+
+ output_set_by_user = true;
+
+ free(str);
+ return rc;
+}
+
/* Helper function for filesystems that return a dent->d_type DT_UNKNOWN */
static int is_directory(const char *base_path, const struct dirent *dent)
{
@@ -623,6 +722,9 @@ static const struct option options[] = {
"input file name"),
OPT_BOOLEAN('d', "debug-mode", &debug_mode,
"do various checks like samples ordering and lost events"),
+ OPT_CALLBACK('f', "fields", NULL, "str",
+ "comma separated output fields. Options: comm,tid,pid,time,cpu,event,trace",
+ parse_output_fields),
OPT_END()
};
@@ -809,8 +911,15 @@ int cmd_script(int argc, const char **argv, const char *prefix __used)
if (generate_script_lang) {
struct stat perf_stat;
+ int input;
+
+ if (output_set_by_user) {
+ fprintf(stderr,
+ "custom fields not supported for generated scripts");
+ return -1;
+ }
- int input = open(input_name, O_RDONLY);
+ input = open(input_name, O_RDONLY);
if (input < 0) {
perror("failed to open file");
exit(-1);