perf tools: Add support for PERF_SAMPLE_IDENTIFIER

Enable parsing of samples with sample format bit PERF_SAMPLE_IDENTIFIER.
In addition, if the kernel supports it, prefer it to selecting
PERF_SAMPLE_ID thereby allowing non-matching sample types.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/1377591794-30553-8-git-send-email-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This commit is contained in:
Adrian Hunter 2013-08-27 11:23:09 +03:00 committed by Arnaldo Carvalho de Melo
parent faf967068e
commit 75562573ba
9 changed files with 310 additions and 22 deletions

View File

@ -365,7 +365,7 @@ static int process_read_event(struct perf_tool *tool,
static int perf_report__setup_sample_type(struct perf_report *rep)
{
struct perf_session *self = rep->session;
u64 sample_type = perf_evlist__sample_type(self->evlist);
u64 sample_type = perf_evlist__combined_sample_type(self->evlist);
if (!self->fd_pipe && !(sample_type & PERF_SAMPLE_CALLCHAIN)) {
if (sort__has_parent) {

View File

@ -72,7 +72,7 @@ int test__basic_mmap(void)
}
evsels[i]->attr.wakeup_events = 1;
perf_evsel__set_sample_id(evsels[i]);
perf_evsel__set_sample_id(evsels[i], false);
perf_evlist__add(evlist, evsels[i]);

View File

@ -53,7 +53,8 @@ struct read_event {
(PERF_SAMPLE_IP | PERF_SAMPLE_TID | \
PERF_SAMPLE_TIME | PERF_SAMPLE_ADDR | \
PERF_SAMPLE_ID | PERF_SAMPLE_STREAM_ID | \
PERF_SAMPLE_CPU | PERF_SAMPLE_PERIOD)
PERF_SAMPLE_CPU | PERF_SAMPLE_PERIOD | \
PERF_SAMPLE_IDENTIFIER)
struct sample_event {
struct perf_event_header header;

View File

@ -49,6 +49,21 @@ struct perf_evlist *perf_evlist__new(void)
return evlist;
}
/**
* perf_evlist__set_id_pos - set the positions of event ids.
* @evlist: selected event list
*
* Events with compatible sample types all have the same id_pos
* and is_pos. For convenience, put a copy on evlist.
*/
void perf_evlist__set_id_pos(struct perf_evlist *evlist)
{
struct perf_evsel *first = perf_evlist__first(evlist);
evlist->id_pos = first->id_pos;
evlist->is_pos = first->is_pos;
}
static void perf_evlist__purge(struct perf_evlist *evlist)
{
struct perf_evsel *pos, *n;
@ -79,15 +94,20 @@ void perf_evlist__delete(struct perf_evlist *evlist)
void perf_evlist__add(struct perf_evlist *evlist, struct perf_evsel *entry)
{
list_add_tail(&entry->node, &evlist->entries);
++evlist->nr_entries;
if (!evlist->nr_entries++)
perf_evlist__set_id_pos(evlist);
}
void perf_evlist__splice_list_tail(struct perf_evlist *evlist,
struct list_head *list,
int nr_entries)
{
bool set_id_pos = !evlist->nr_entries;
list_splice_tail(list, &evlist->entries);
evlist->nr_entries += nr_entries;
if (set_id_pos)
perf_evlist__set_id_pos(evlist);
}
void __perf_evlist__set_leader(struct list_head *list)
@ -349,6 +369,55 @@ struct perf_evsel *perf_evlist__id2evsel(struct perf_evlist *evlist, u64 id)
return NULL;
}
static int perf_evlist__event2id(struct perf_evlist *evlist,
union perf_event *event, u64 *id)
{
const u64 *array = event->sample.array;
ssize_t n;
n = (event->header.size - sizeof(event->header)) >> 3;
if (event->header.type == PERF_RECORD_SAMPLE) {
if (evlist->id_pos >= n)
return -1;
*id = array[evlist->id_pos];
} else {
if (evlist->is_pos > n)
return -1;
n -= evlist->is_pos;
*id = array[n];
}
return 0;
}
static struct perf_evsel *perf_evlist__event2evsel(struct perf_evlist *evlist,
union perf_event *event)
{
struct hlist_head *head;
struct perf_sample_id *sid;
int hash;
u64 id;
if (evlist->nr_entries == 1)
return perf_evlist__first(evlist);
if (perf_evlist__event2id(evlist, event, &id))
return NULL;
/* Synthesized events have an id of zero */
if (!id)
return perf_evlist__first(evlist);
hash = hash_64(id, PERF_EVLIST__HLIST_BITS);
head = &evlist->heads[hash];
hlist_for_each_entry(sid, head, node) {
if (sid->id == id)
return sid->evsel;
}
return NULL;
}
union perf_event *perf_evlist__mmap_read(struct perf_evlist *evlist, int idx)
{
struct perf_mmap *md = &evlist->mmap[idx];
@ -659,20 +728,40 @@ int perf_evlist__set_filter(struct perf_evlist *evlist, const char *filter)
bool perf_evlist__valid_sample_type(struct perf_evlist *evlist)
{
struct perf_evsel *first = perf_evlist__first(evlist), *pos = first;
struct perf_evsel *pos;
list_for_each_entry_continue(pos, &evlist->entries, node) {
if (first->attr.sample_type != pos->attr.sample_type)
if (evlist->nr_entries == 1)
return true;
if (evlist->id_pos < 0 || evlist->is_pos < 0)
return false;
list_for_each_entry(pos, &evlist->entries, node) {
if (pos->id_pos != evlist->id_pos ||
pos->is_pos != evlist->is_pos)
return false;
}
return true;
}
u64 perf_evlist__sample_type(struct perf_evlist *evlist)
u64 __perf_evlist__combined_sample_type(struct perf_evlist *evlist)
{
struct perf_evsel *first = perf_evlist__first(evlist);
return first->attr.sample_type;
struct perf_evsel *evsel;
if (evlist->combined_sample_type)
return evlist->combined_sample_type;
list_for_each_entry(evsel, &evlist->entries, node)
evlist->combined_sample_type |= evsel->attr.sample_type;
return evlist->combined_sample_type;
}
u64 perf_evlist__combined_sample_type(struct perf_evlist *evlist)
{
evlist->combined_sample_type = 0;
return __perf_evlist__combined_sample_type(evlist);
}
bool perf_evlist__valid_read_format(struct perf_evlist *evlist)
@ -727,6 +816,9 @@ u16 perf_evlist__id_hdr_size(struct perf_evlist *evlist)
if (sample_type & PERF_SAMPLE_CPU)
size += sizeof(data->cpu) * 2;
if (sample_type & PERF_SAMPLE_IDENTIFIER)
size += sizeof(data->id);
out:
return size;
}
@ -885,7 +977,10 @@ int perf_evlist__start_workload(struct perf_evlist *evlist)
int perf_evlist__parse_sample(struct perf_evlist *evlist, union perf_event *event,
struct perf_sample *sample)
{
struct perf_evsel *evsel = perf_evlist__first(evlist);
struct perf_evsel *evsel = perf_evlist__event2evsel(evlist, event);
if (!evsel)
return -EFAULT;
return perf_evsel__parse_sample(evsel, event, sample);
}

View File

@ -32,6 +32,9 @@ struct perf_evlist {
int nr_fds;
int nr_mmaps;
int mmap_len;
int id_pos;
int is_pos;
u64 combined_sample_type;
struct {
int cork_fd;
pid_t pid;
@ -85,6 +88,8 @@ union perf_event *perf_evlist__mmap_read(struct perf_evlist *self, int idx);
int perf_evlist__open(struct perf_evlist *evlist);
void perf_evlist__close(struct perf_evlist *evlist);
void perf_evlist__set_id_pos(struct perf_evlist *evlist);
bool perf_can_sample_identifier(void);
void perf_evlist__config(struct perf_evlist *evlist,
struct perf_record_opts *opts);
@ -121,7 +126,8 @@ void __perf_evlist__set_leader(struct list_head *list);
void perf_evlist__set_leader(struct perf_evlist *evlist);
u64 perf_evlist__read_format(struct perf_evlist *evlist);
u64 perf_evlist__sample_type(struct perf_evlist *evlist);
u64 __perf_evlist__combined_sample_type(struct perf_evlist *evlist);
u64 perf_evlist__combined_sample_type(struct perf_evlist *evlist);
bool perf_evlist__sample_id_all(struct perf_evlist *evlist);
u16 perf_evlist__id_hdr_size(struct perf_evlist *evlist);

View File

@ -31,7 +31,7 @@ static struct {
#define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))
static int __perf_evsel__sample_size(u64 sample_type)
int __perf_evsel__sample_size(u64 sample_type)
{
u64 mask = sample_type & PERF_SAMPLE_MASK;
int size = 0;
@ -47,6 +47,72 @@ static int __perf_evsel__sample_size(u64 sample_type)
return size;
}
/**
* __perf_evsel__calc_id_pos - calculate id_pos.
* @sample_type: sample type
*
* This function returns the position of the event id (PERF_SAMPLE_ID or
* PERF_SAMPLE_IDENTIFIER) in a sample event i.e. in the array of struct
* sample_event.
*/
static int __perf_evsel__calc_id_pos(u64 sample_type)
{
int idx = 0;
if (sample_type & PERF_SAMPLE_IDENTIFIER)
return 0;
if (!(sample_type & PERF_SAMPLE_ID))
return -1;
if (sample_type & PERF_SAMPLE_IP)
idx += 1;
if (sample_type & PERF_SAMPLE_TID)
idx += 1;
if (sample_type & PERF_SAMPLE_TIME)
idx += 1;
if (sample_type & PERF_SAMPLE_ADDR)
idx += 1;
return idx;
}
/**
* __perf_evsel__calc_is_pos - calculate is_pos.
* @sample_type: sample type
*
* This function returns the position (counting backwards) of the event id
* (PERF_SAMPLE_ID or PERF_SAMPLE_IDENTIFIER) in a non-sample event i.e. if
* sample_id_all is used there is an id sample appended to non-sample events.
*/
static int __perf_evsel__calc_is_pos(u64 sample_type)
{
int idx = 1;
if (sample_type & PERF_SAMPLE_IDENTIFIER)
return 1;
if (!(sample_type & PERF_SAMPLE_ID))
return -1;
if (sample_type & PERF_SAMPLE_CPU)
idx += 1;
if (sample_type & PERF_SAMPLE_STREAM_ID)
idx += 1;
return idx;
}
void perf_evsel__calc_id_pos(struct perf_evsel *evsel)
{
evsel->id_pos = __perf_evsel__calc_id_pos(evsel->attr.sample_type);
evsel->is_pos = __perf_evsel__calc_is_pos(evsel->attr.sample_type);
}
void hists__init(struct hists *hists)
{
memset(hists, 0, sizeof(*hists));
@ -63,6 +129,7 @@ void __perf_evsel__set_sample_bit(struct perf_evsel *evsel,
if (!(evsel->attr.sample_type & bit)) {
evsel->attr.sample_type |= bit;
evsel->sample_size += sizeof(u64);
perf_evsel__calc_id_pos(evsel);
}
}
@ -72,12 +139,19 @@ void __perf_evsel__reset_sample_bit(struct perf_evsel *evsel,
if (evsel->attr.sample_type & bit) {
evsel->attr.sample_type &= ~bit;
evsel->sample_size -= sizeof(u64);
perf_evsel__calc_id_pos(evsel);
}
}
void perf_evsel__set_sample_id(struct perf_evsel *evsel)
void perf_evsel__set_sample_id(struct perf_evsel *evsel,
bool can_sample_identifier)
{
perf_evsel__set_sample_bit(evsel, ID);
if (can_sample_identifier) {
perf_evsel__reset_sample_bit(evsel, ID);
perf_evsel__set_sample_bit(evsel, IDENTIFIER);
} else {
perf_evsel__set_sample_bit(evsel, ID);
}
evsel->attr.read_format |= PERF_FORMAT_ID;
}
@ -90,6 +164,7 @@ void perf_evsel__init(struct perf_evsel *evsel,
INIT_LIST_HEAD(&evsel->node);
hists__init(&evsel->hists);
evsel->sample_size = __perf_evsel__sample_size(attr->sample_type);
perf_evsel__calc_id_pos(evsel);
}
struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr, int idx)
@ -509,7 +584,7 @@ void perf_evsel__config(struct perf_evsel *evsel,
* We need ID even in case of single event, because
* PERF_SAMPLE_READ process ID specific data.
*/
perf_evsel__set_sample_id(evsel);
perf_evsel__set_sample_id(evsel, false);
/*
* Apply group format only if we belong to group
@ -1088,6 +1163,11 @@ static int perf_evsel__parse_id_sample(const struct perf_evsel *evsel,
array += ((event->header.size -
sizeof(event->header)) / sizeof(u64)) - 1;
if (type & PERF_SAMPLE_IDENTIFIER) {
sample->id = *array;
array--;
}
if (type & PERF_SAMPLE_CPU) {
u.val64 = *array;
if (swapped) {
@ -1184,6 +1264,12 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
if (evsel->sample_size + sizeof(event->header) > event->header.size)
return -EFAULT;
data->id = -1ULL;
if (type & PERF_SAMPLE_IDENTIFIER) {
data->id = *array;
array++;
}
if (type & PERF_SAMPLE_IP) {
data->ip = *array;
array++;
@ -1214,7 +1300,6 @@ int perf_evsel__parse_sample(struct perf_evsel *evsel, union perf_event *event,
array++;
}
data->id = -1ULL;
if (type & PERF_SAMPLE_ID) {
data->id = *array;
array++;
@ -1396,6 +1481,11 @@ int perf_event__synthesize_sample(union perf_event *event, u64 type,
array = event->sample.array;
if (type & PERF_SAMPLE_IDENTIFIER) {
*array = sample->id;
array++;
}
if (type & PERF_SAMPLE_IP) {
*array = sample->ip;
array++;
@ -1584,6 +1674,7 @@ static int sample_type__fprintf(FILE *fp, bool *first, u64 value)
bit_name(READ), bit_name(CALLCHAIN), bit_name(ID), bit_name(CPU),
bit_name(PERIOD), bit_name(STREAM_ID), bit_name(RAW),
bit_name(BRANCH_STACK), bit_name(REGS_USER), bit_name(STACK_USER),
bit_name(IDENTIFIER),
{ .name = NULL, }
};
#undef bit_name

View File

@ -48,6 +48,12 @@ struct perf_sample_id {
* @name - Can be set to retain the original event name passed by the user,
* so that when showing results in tools such as 'perf stat', we
* show the name used, not some alias.
* @id_pos: the position of the event id (PERF_SAMPLE_ID or
* PERF_SAMPLE_IDENTIFIER) in a sample event i.e. in the array of
* struct sample_event
* @is_pos: the position (counting backwards) of the event id (PERF_SAMPLE_ID or
* PERF_SAMPLE_IDENTIFIER) in a non-sample event i.e. if sample_id_all
* is used there is an id sample appended to non-sample events
*/
struct perf_evsel {
struct list_head node;
@ -74,6 +80,8 @@ struct perf_evsel {
} handler;
struct cpu_map *cpus;
unsigned int sample_size;
int id_pos;
int is_pos;
bool supported;
bool needs_swap;
/* parse modifier helper */
@ -104,6 +112,9 @@ void perf_evsel__delete(struct perf_evsel *evsel);
void perf_evsel__config(struct perf_evsel *evsel,
struct perf_record_opts *opts);
int __perf_evsel__sample_size(u64 sample_type);
void perf_evsel__calc_id_pos(struct perf_evsel *evsel);
bool perf_evsel__is_cache_op_valid(u8 type, u8 op);
#define PERF_EVSEL__MAX_ALIASES 8
@ -142,7 +153,8 @@ void __perf_evsel__reset_sample_bit(struct perf_evsel *evsel,
#define perf_evsel__reset_sample_bit(evsel, bit) \
__perf_evsel__reset_sample_bit(evsel, PERF_SAMPLE_##bit)
void perf_evsel__set_sample_id(struct perf_evsel *evsel);
void perf_evsel__set_sample_id(struct perf_evsel *evsel,
bool use_sample_identifier);
int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, int nthreads,
const char *filter);

View File

@ -1,11 +1,83 @@
#include "evlist.h"
#include "evsel.h"
#include "cpumap.h"
#include "parse-events.h"
typedef void (*setup_probe_fn_t)(struct perf_evsel *evsel);
static int perf_do_probe_api(setup_probe_fn_t fn, int cpu, const char *str)
{
struct perf_evlist *evlist;
struct perf_evsel *evsel;
int err = -EAGAIN, fd;
evlist = perf_evlist__new();
if (!evlist)
return -ENOMEM;
if (parse_events(evlist, str))
goto out_delete;
evsel = perf_evlist__first(evlist);
fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, 0);
if (fd < 0)
goto out_delete;
close(fd);
fn(evsel);
fd = sys_perf_event_open(&evsel->attr, -1, cpu, -1, 0);
if (fd < 0) {
if (errno == EINVAL)
err = -EINVAL;
goto out_delete;
}
close(fd);
err = 0;
out_delete:
perf_evlist__delete(evlist);
return err;
}
static bool perf_probe_api(setup_probe_fn_t fn)
{
const char *try[] = {"cycles:u", "instructions:u", "cpu-clock", NULL};
struct cpu_map *cpus;
int cpu, ret, i = 0;
cpus = cpu_map__new(NULL);
if (!cpus)
return false;
cpu = cpus->map[0];
cpu_map__delete(cpus);
do {
ret = perf_do_probe_api(fn, cpu, try[i++]);
if (!ret)
return true;
} while (ret == -EAGAIN && try[i]);
return false;
}
static void perf_probe_sample_identifier(struct perf_evsel *evsel)
{
evsel->attr.sample_type |= PERF_SAMPLE_IDENTIFIER;
}
bool perf_can_sample_identifier(void)
{
return perf_probe_api(perf_probe_sample_identifier);
}
void perf_evlist__config(struct perf_evlist *evlist,
struct perf_record_opts *opts)
{
struct perf_evsel *evsel;
bool use_sample_identifier = false;
/*
* Set the evsel leader links before we configure attributes,
* since some might depend on this info.
@ -16,10 +88,21 @@ void perf_evlist__config(struct perf_evlist *evlist,
if (evlist->cpus->map[0] < 0)
opts->no_inherit = true;
list_for_each_entry(evsel, &evlist->entries, node) {
list_for_each_entry(evsel, &evlist->entries, node)
perf_evsel__config(evsel, opts);
if (evlist->nr_entries > 1)
perf_evsel__set_sample_id(evsel);
if (evlist->nr_entries > 1) {
struct perf_evsel *first = perf_evlist__first(evlist);
list_for_each_entry(evsel, &evlist->entries, node) {
if (evsel->attr.sample_type == first->attr.sample_type)
continue;
use_sample_identifier = perf_can_sample_identifier();
break;
}
list_for_each_entry(evsel, &evlist->entries, node)
perf_evsel__set_sample_id(evsel, use_sample_identifier);
}
perf_evlist__set_id_pos(evlist);
}

View File

@ -739,7 +739,7 @@ static void perf_session__print_tstamp(struct perf_session *session,
union perf_event *event,
struct perf_sample *sample)
{
u64 sample_type = perf_evlist__sample_type(session->evlist);
u64 sample_type = __perf_evlist__combined_sample_type(session->evlist);
if (event->header.type != PERF_RECORD_SAMPLE &&
!perf_evlist__sample_id_all(session->evlist)) {