Based on the [PATCH 3/5], this patch changed the related interfaces in evlist &
evsel to support the operations to thread_map's bitmap. Then, we can use these
interfaces to insert a new forked thread into or remove a exited trhead from
thread_map and other related data structures.
Cc: David Ahern <dsah...@gmail.com>
Cc: Arjan van de Ven <ar...@linux.intel.com>
Cc: Namhyung Kim <namhy...@gmail.com>
Cc: Yanmin Zhang <yanmin.zh...@intel.com>
Cc: Wu Fengguang <fengguang...@intel.com>
Cc: Mike Galbraith <efa...@gmx.de>
Cc: Paul Mackerras <pau...@samba.org>
Cc: Peter Zijlstra <pet...@infradead.org>
Cc: Ingo Molnar <mi...@redhat.com>
Cc: Arnaldo Carvalho de Melo <a...@ghostprotocols.net>
Cc: Andrew Morton <a...@linux-foundation.org>
Signed-off-by: Chenggang Qin <chenggang....@taobao.com>
---
tools/perf/builtin-record.c | 25 ++-
tools/perf/builtin-stat.c | 7 +-
tools/perf/builtin-top.c | 14 +-
tools/perf/tests/mmap-basic.c | 4 +-
tools/perf/tests/open-syscall-all-cpus.c | 2 +-
tools/perf/tests/open-syscall-tp-fields.c | 3 +-
tools/perf/tests/open-syscall.c | 3 +-
tools/perf/tests/perf-record.c | 2 +-
tools/perf/util/evlist.c | 236 +++++++++++++++++++++++------
tools/perf/util/evlist.h | 39 +++--
tools/perf/util/evsel.c | 147 +++++++++++++++---
tools/perf/util/evsel.h | 38 +++--
tools/perf/util/python.c | 3 +-
13 files changed, 408 insertions(+), 115 deletions(-)
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index f3151d3..277303f 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -359,7 +359,7 @@ try_again:
goto out;
}
- if (perf_evlist__mmap(evlist, opts->mmap_pages, false) < 0) {
+ if (perf_evlist__mmap(evlist, opts->mmap_pages, false, -1, false) < 0) {
if (errno == EPERM) {
pr_err("Permission error mapping pages.\n"
"Consider increasing "
@@ -472,12 +472,21 @@ static int perf_record__mmap_read_all(struct perf_record
*rec)
int i;
int rc = 0;
- for (i = 0; i < rec->evlist->nr_mmaps; i++) {
- if (rec->evlist->mmap[i].base) {
- if (perf_record__mmap_read(rec, &rec->evlist->mmap[i])
!= 0) {
- rc = -1;
- goto out;
- }
+ if (cpu_map__all(rec->evlist->cpus)) {
+ for_each_set_bit(i, rec->evlist->threads->bitmap,
+ PID_MAX_DEFAULT) {
+ if (rec->evlist->mmap[i].base)
+ if (perf_record__mmap_read(rec,
+ &rec->evlist->mmap[i]) != 0){
+ rc = -1;
+ goto out;
+ }
+ }
+ } else {
+ for (i = 0; i < rec->evlist->nr_mmaps; i++) {
+ if (rec->evlist->mmap[i].base)
+ if (perf_record__mmap_read(rec,
+ &rec->evlist->mmap[i]) != 0) {
+ rc = -1;
+ goto out;
+ }
}
}
@@ -1161,7 +1170,7 @@ int cmd_record(int argc, const char **argv, const char
*prefix __maybe_unused)
err = -EINVAL;
goto out_free_fd;
}
-
+
err = __cmd_record(&record, argc, argv);
out_free_fd:
perf_evlist__delete_maps(evsel_list);
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index c247fac..74d5311 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -229,7 +229,7 @@ static int read_counter_aggr(struct perf_evsel *counter)
int i;
if (__perf_evsel__read(counter, perf_evsel__nr_cpus(counter),
- evsel_list->threads->nr, scale) < 0)
+ evsel_list->threads->bitmap, scale) < 0)
return -1;
for (i = 0; i < 3; i++)
@@ -394,13 +394,14 @@ static int __run_perf_stat(int argc __maybe_unused, const
char **argv)
if (no_aggr) {
list_for_each_entry(counter, &evsel_list->entries, node) {
read_counter(counter);
- perf_evsel__close_fd(counter,
perf_evsel__nr_cpus(counter), 1);
+ perf_evsel__close_fd(counter,
+ perf_evsel__nr_cpus(counter),
+ evsel_list->threads->bitmap);
}
} else {
list_for_each_entry(counter, &evsel_list->entries, node) {
read_counter_aggr(counter);
perf_evsel__close_fd(counter,
perf_evsel__nr_cpus(counter),
- evsel_list->threads->nr);
+ evsel_list->threads->bitmap);
}
}
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index c9ff395..b3650e3 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -68,6 +68,8 @@
#include <linux/unistd.h>
#include <linux/types.h>
+#include "asm/bug.h"
+
void get_term_dimensions(struct winsize *ws)
{
char *s = getenv("LINES");
@@ -823,7 +825,7 @@ static void perf_top__mmap_read_idx(struct perf_top *top,
int idx)
struct perf_evsel *evsel;
struct perf_session *session = top->session;
union perf_event *event;
- struct machine *machine;
+ struct machine *machine = NULL;
u8 origin;
int ret;
@@ -886,8 +888,12 @@ static void perf_top__mmap_read(struct perf_top *top)
{
int i;
- for (i = 0; i < top->evlist->nr_mmaps; i++)
- perf_top__mmap_read_idx(top, i);
+ if (cpu_map__all(top->evlist->cpus)) {
+ for_each_set_bit(i, top->evlist->threads->bitmap,
+ PID_MAX_DEFAULT)
+ perf_top__mmap_read_idx(top, i);
+ } else
+ for (i = 0; i < top->evlist->nr_mmaps; i++)
+ perf_top__mmap_read_idx(top, i);
}
static void perf_top__start_counters(struct perf_top *top)
@@ -996,7 +1002,7 @@ try_again:
}
}
- if (perf_evlist__mmap(evlist, top->mmap_pages, false) < 0) {
+ if (perf_evlist__mmap(evlist, top->mmap_pages, false, -1, false) < 0) {
ui__error("Failed to mmap with %d (%s)\n",
errno, strerror(errno));
goto out_err;
diff --git a/tools/perf/tests/mmap-basic.c b/tools/perf/tests/mmap-basic.c
index e174681..fac0316 100644
--- a/tools/perf/tests/mmap-basic.c
+++ b/tools/perf/tests/mmap-basic.c
@@ -101,7 +101,7 @@ int test__basic_mmap(void)
}
}
- if (perf_evlist__mmap(evlist, 128, true) < 0) {
+ if (perf_evlist__mmap(evlist, 128, true, -1, false) < 0) {
pr_debug("failed to mmap events: %d (%s)\n", errno,
strerror(errno));
goto out_close_fd;
@@ -151,7 +151,7 @@ out_munmap:
perf_evlist__munmap(evlist);
out_close_fd:
for (i = 0; i < nsyscalls; ++i)
- perf_evsel__close_fd(evsels[i], 1, threads->nr);
+ perf_evsel__close_fd(evsels[i], 1, threads->bitmap);
out_free_evlist:
perf_evlist__delete(evlist);
out_free_cpus:
diff --git a/tools/perf/tests/open-syscall-all-cpus.c
b/tools/perf/tests/open-syscall-all-cpus.c
index 31072ab..4d6f8ed 100644
--- a/tools/perf/tests/open-syscall-all-cpus.c
+++ b/tools/perf/tests/open-syscall-all-cpus.c
@@ -111,7 +111,7 @@ int test__open_syscall_event_on_all_cpus(void)
}
out_close_fd:
- perf_evsel__close_fd(evsel, 1, threads->nr);
+ perf_evsel__close_fd(evsel, 1, threads->bitmap);
out_evsel_delete:
perf_evsel__delete(evsel);
out_thread_map_delete:
diff --git a/tools/perf/tests/open-syscall-tp-fields.c
b/tools/perf/tests/open-syscall-tp-fields.c
index 1c52fdc..5613863 100644
--- a/tools/perf/tests/open-syscall-tp-fields.c
+++ b/tools/perf/tests/open-syscall-tp-fields.c
@@ -44,6 +44,7 @@ int test__syscall_open_tp_fields(void)
perf_evsel__config(evsel, &opts);
evlist->threads->map[0] = getpid();
+ set_bit(0, evlist->threads->bitmap);
err = perf_evlist__open(evlist);
if (err < 0) {
@@ -51,7 +52,7 @@ int test__syscall_open_tp_fields(void)
goto out_delete_evlist;
}
- err = perf_evlist__mmap(evlist, UINT_MAX, false);
+ err = perf_evlist__mmap(evlist, UINT_MAX, false, -1, false);
if (err < 0) {
pr_debug("perf_evlist__mmap: %s\n", strerror(errno));
goto out_delete_evlist;
diff --git a/tools/perf/tests/open-syscall.c b/tools/perf/tests/open-syscall.c
index 98be8b5..6b0b1da 100644
--- a/tools/perf/tests/open-syscall.c
+++ b/tools/perf/tests/open-syscall.c
@@ -2,6 +2,7 @@
#include "evsel.h"
#include "debug.h"
#include "tests.h"
+#include <linux/bitops.h>
int test__open_syscall_event(void)
{
@@ -57,7 +58,7 @@ int test__open_syscall_event(void)
err = 0;
out_close_fd:
- perf_evsel__close_fd(evsel, 1, threads->nr);
+ perf_evsel__close_fd(evsel, 1, threads->bitmap);
out_evsel_delete:
perf_evsel__delete(evsel);
out_thread_map_delete:
diff --git a/tools/perf/tests/perf-record.c b/tools/perf/tests/perf-record.c
index 70e0d44..9678d7b 100644
--- a/tools/perf/tests/perf-record.c
+++ b/tools/perf/tests/perf-record.c
@@ -139,7 +139,7 @@ int test__PERF_RECORD(void)
* fds in the same CPU to be injected in the same mmap ring buffer
* (using ioctl(PERF_EVENT_IOC_SET_OUTPUT)).
*/
- err = perf_evlist__mmap(evlist, opts.mmap_pages, false);
+ err = perf_evlist__mmap(evlist, opts.mmap_pages, false, -1, false);
if (err < 0) {
pr_debug("perf_evlist__mmap: %s\n", strerror(errno));
goto out_delete_evlist;
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 7052934..75907cb 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -23,9 +23,6 @@
#include <linux/bitops.h>
#include <linux/hash.h>
-#define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))
-#define SID(e, x, y) xyarray__entry(e->sample_id, x, y)
-
void perf_evlist__init(struct perf_evlist *evlist, struct cpu_map *cpus,
struct thread_map *threads)
{
@@ -224,7 +221,9 @@ void perf_evlist__disable(struct perf_evlist *evlist)
list_for_each_entry(pos, &evlist->entries, node) {
if (perf_evsel__is_group_member(pos))
continue;
- for (thread = 0; thread < evlist->threads->nr; thread++)
+
+ for_each_set_bit(thread, evlist->threads->bitmap,
+ PID_MAX_DEFAULT)
ioctl(FD(pos, cpu, thread),
PERF_EVENT_IOC_DISABLE, 0);
}
@@ -240,20 +239,52 @@ void perf_evlist__enable(struct perf_evlist *evlist)
list_for_each_entry(pos, &evlist->entries, node) {
if (perf_evsel__is_group_member(pos))
continue;
- for (thread = 0; thread < evlist->threads->nr; thread++)
+ for_each_set_bit(thread, evlist->threads->bitmap,
+ PID_MAX_DEFAULT)
ioctl(FD(pos, cpu, thread),
PERF_EVENT_IOC_ENABLE, 0);
}
}
}
+static int perf_evlist__realloc_pollfd(struct perf_evlist *evlist)
+{
+ int nfds = cpu_map__nr(evlist->cpus) * evlist->threads->max_nr *
+ evlist->nr_entries;
+ struct pollfd *pollfd;
+
+ pollfd = realloc(evlist->pollfd, sizeof(struct pollfd) * nfds);
+
+ if (pollfd == NULL)
+ goto out;
+
+ evlist->pollfd = pollfd;
+
+ return 0;
+out:
+ return -ENOMEM;
+}
+
static int perf_evlist__alloc_pollfd(struct perf_evlist *evlist)
{
- int nfds = cpu_map__nr(evlist->cpus) * evlist->threads->nr *
evlist->nr_entries;
+ int nfds = cpu_map__nr(evlist->cpus) * evlist->threads->max_nr *
+ evlist->nr_entries;
evlist->pollfd = malloc(sizeof(struct pollfd) * nfds);
return evlist->pollfd != NULL ? 0 : -ENOMEM;
}
+void perf_evlist__remove_pollfd(struct perf_evlist *evlist, int nr_thread)
+{
+ int cpu;
+ int entry;
+ int row_size = evlist->threads->max_nr * evlist->nr_entries;
+
+ for (cpu = 0; cpu < cpu_map__nr(evlist->cpus); cpu++) {
+ for (entry = 0; entry < evlist->nr_entries; entry++) {
+ evlist->pollfd[cpu * row_size + nr_thread + entry].fd =
-1;
+ evlist->nr_fds--;
+ }
+ }
+}
+
void perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd)
{
fcntl(fd, F_SETFL, O_NONBLOCK);
@@ -275,11 +306,30 @@ static void perf_evlist__id_hash(struct perf_evlist
*evlist,
hlist_add_head(&sid->node, &evlist->heads[hash]);
}
+static void perf_evlist__id_hash_del(struct perf_evsel *evsel,
+ int cpu, int thread)
+{
+ struct perf_sample_id *sid = SID(evsel, cpu, thread);
+
+ hlist_del(&sid->node);
+ sid->id = 0;
+ sid->evsel = NULL;
+}
+
void perf_evlist__id_add(struct perf_evlist *evlist, struct perf_evsel *evsel,
int cpu, int thread, u64 id)
{
perf_evlist__id_hash(evlist, evsel, cpu, thread, id);
- evsel->id[evsel->ids++] = id;
+ set_bit(ID_BITMAP_POS(cpu, thread), evsel->id_bitmap);
+ evsel->id[ID_BITMAP_POS(cpu, thread)] = id;
+}
+
+void perf_evlist__id_remove(struct perf_evsel *evsel,
+ int cpu, int thread)
+{
+ perf_evlist__id_hash_del(evsel, cpu, thread);
+ clear_bit(ID_BITMAP_POS(cpu, thread), evsel->id_bitmap);
+ evsel->id[ID_BITMAP_POS(cpu, thread)] = -1;
}
static int perf_evlist__id_add_fd(struct perf_evlist *evlist,
@@ -304,7 +354,7 @@ static int perf_evlist__id_add_fd(struct perf_evlist
*evlist,
struct perf_evsel *perf_evlist__id2evsel(struct perf_evlist *evlist, u64 id)
{
- struct hlist_head *head;
+ struct hlist_head *head = NULL;
struct hlist_node *pos;
struct perf_sample_id *sid;
int hash;
@@ -407,13 +457,41 @@ void perf_evlist__munmap(struct perf_evlist *evlist)
static int perf_evlist__alloc_mmap(struct perf_evlist *evlist)
{
+ int max_nr_mmaps;
+
evlist->nr_mmaps = cpu_map__nr(evlist->cpus);
- if (cpu_map__all(evlist->cpus))
- evlist->nr_mmaps = evlist->threads->nr;
- evlist->mmap = zalloc(evlist->nr_mmaps * sizeof(struct perf_mmap));
+ max_nr_mmaps = evlist->nr_mmaps;
+
+ if (cpu_map__all(evlist->cpus)) {
+ evlist->nr_mmaps = 0;
+ max_nr_mmaps = evlist->threads->max_nr;
+ }
+
+ evlist->mmap = zalloc(max_nr_mmaps * sizeof(struct perf_mmap));
return evlist->mmap != NULL ? 0 : -ENOMEM;
}
+static int perf_evlist__realloc_mmap(struct perf_evlist *evlist)
+{
+ struct perf_mmap *mt;
+
+ if (!cpu_map__all(evlist->cpus))
+ return 0;
+
+ mt = realloc(evlist->mmap, evlist->threads->max_nr *
+ sizeof(struct perf_mmap));
+
+ if (mt == NULL) {
+ printf("mmap realloc failed\n");
+ goto out;
+ }
+
+ evlist->mmap = mt;
+
+ return 0;
+out:
+ return -1;
+}
+
static int __perf_evlist__mmap(struct perf_evlist *evlist,
int idx, int prot, int mask, int fd)
{
@@ -426,6 +504,9 @@ static int __perf_evlist__mmap(struct perf_evlist *evlist,
return -1;
}
+ if (cpu_map__all(evlist->cpus))
+ evlist->nr_mmaps++;
+
perf_evlist__add_pollfd(evlist, fd);
return 0;
}
@@ -438,7 +519,8 @@ static int perf_evlist__mmap_per_cpu(struct perf_evlist
*evlist, int prot, int m
for (cpu = 0; cpu < evlist->cpus->nr; cpu++) {
int output = -1;
- for (thread = 0; thread < evlist->threads->nr; thread++) {
+ for_each_set_bit(thread, evlist->threads->bitmap,
+ PID_MAX_DEFAULT) {
list_for_each_entry(evsel, &evlist->entries, node) {
int fd = FD(evsel, cpu, thread);
@@ -471,37 +553,55 @@ out_unmap:
return -1;
}
-static int perf_evlist__mmap_per_thread(struct perf_evlist *evlist, int prot,
int mask)
+static int perf_evlist__mmap_a_thread(struct perf_evlist *evlist, int prot,
+ int mask, int thread_id)
{
struct perf_evsel *evsel;
- int thread;
+ int output = -1;
- for (thread = 0; thread < evlist->threads->nr; thread++) {
- int output = -1;
+ list_for_each_entry(evsel, &evlist->entries, node) {
+ int fd = FD(evsel, 0, thread_id);
- list_for_each_entry(evsel, &evlist->entries, node) {
- int fd = FD(evsel, 0, thread);
+ if (fd <= 0)
+ continue;
- if (output == -1) {
- output = fd;
- if (__perf_evlist__mmap(evlist, thread,
- prot, mask, output) < 0)
- goto out_unmap;
- } else {
- if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT,
output) != 0)
- goto out_unmap;
- }
+ if (output == -1) {
+ output = fd;
+ if (__perf_evlist__mmap(evlist, thread_id,
+ prot, mask, output) < 0)
+ return -1;
+ } else
+ if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT, output) != 0)
+ return -1;
- if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
- perf_evlist__id_add_fd(evlist, evsel, 0, thread,
fd) < 0)
- goto out_unmap;
- }
+ if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
+ perf_evlist__id_add_fd(evlist, evsel, 0, thread_id, fd) < 0)
+ return -1;
+ }
+ return 0;
+}
+
+static int perf_evlist__mmap_per_thread(struct perf_evlist *evlist, int prot,
+ int mask, int nr_append)
+{
+ int thread;
+
+ if (nr_append >= 0) {
+ if (perf_evlist__mmap_a_thread(evlist, prot, mask,
+ nr_append) < 0)
+ goto out_unmap;
+
+ return 0;
+ }
+
+ for_each_set_bit(thread, evlist->threads->bitmap, PID_MAX_DEFAULT) {
+ if (perf_evlist__mmap_a_thread(evlist, prot, mask, thread) < 0)
+ goto out_unmap;
}
return 0;
out_unmap:
- for (thread = 0; thread < evlist->threads->nr; thread++) {
+ for_each_set_bit(thread, evlist->threads->bitmap, PID_MAX_DEFAULT) {
if (evlist->mmap[thread].base != NULL) {
munmap(evlist->mmap[thread].base, evlist->mmap_len);
evlist->mmap[thread].base = NULL;
@@ -510,6 +610,35 @@ out_unmap:
return -1;
}
+static void perf_evlist__reset_heads(struct perf_evlist *evlist, int old_cpu)
+{
+ struct perf_evsel *evsel;
+ int i;
+ int cpu, thread;
+ int hash;
+
+ for (i = 0; i < PERF_EVLIST__HLIST_SIZE; i++)
+ INIT_HLIST_HEAD(&evlist->heads[i]);
+
+ list_for_each_entry(evsel, &evlist->entries, node) {
+ if (evsel->attr.read_format & PERF_FORMAT_ID) {
+ for (cpu = 0; cpu < old_cpu; cpu++)
+ for_each_set_bit(thread,
+ evlist->threads->bitmap,
+ PID_MAX_DEFAULT) {
+ struct perf_sample_id *sid;
+ sid = SID(evsel, cpu, thread);
+
+ if (sid->id != 0) {
+ hash = hash_64(sid->id,
+
PERF_EVLIST__HLIST_BITS);
+ hlist_add_head(&(sid->node),
+
&(evlist->heads[hash]));
+ }
+ }
+ }
+ }
+}
+
/** perf_evlist__mmap - Create per cpu maps to receive events
*
* @evlist - list of events
@@ -526,13 +655,16 @@ out_unmap:
* Using perf_evlist__read_on_cpu does this automatically.
*/
int perf_evlist__mmap(struct perf_evlist *evlist, unsigned int pages,
- bool overwrite)
+ bool overwrite, int nr_append, bool re_alloc)
{
struct perf_evsel *evsel;
const struct cpu_map *cpus = evlist->cpus;
const struct thread_map *threads = evlist->threads;
int prot = PROT_READ | (overwrite ? 0 : PROT_WRITE), mask;
+ if (evlist->mmap && nr_append < 0)
+ return -1;
+
/* 512 kiB: default amount of unprivileged mlocked memory */
if (pages == UINT_MAX)
pages = (512 * 1024) / page_size;
@@ -544,21 +676,35 @@ int perf_evlist__mmap(struct perf_evlist *evlist,
unsigned int pages,
if (evlist->mmap == NULL && perf_evlist__alloc_mmap(evlist) < 0)
return -ENOMEM;
+ if (evlist->mmap && re_alloc && perf_evlist__realloc_mmap(evlist) < 0)
+ return -ENOMEM;
+
if (evlist->pollfd == NULL && perf_evlist__alloc_pollfd(evlist) < 0)
return -ENOMEM;
+ if (evlist->pollfd && re_alloc &&
+ perf_evlist__realloc_pollfd(evlist) < 0)
+ return -ENOMEM;
+
evlist->overwrite = overwrite;
evlist->mmap_len = (pages + 1) * page_size;
list_for_each_entry(evsel, &evlist->entries, node) {
if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
evsel->sample_id == NULL &&
- perf_evsel__alloc_id(evsel, cpu_map__nr(cpus), threads->nr)
< 0)
+ perf_evsel__alloc_id(evsel, cpu_map__nr(cpus),
+ threads->max_nr) < 0)
+ return -ENOMEM;
+
+ if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
+ evsel->sample_id && re_alloc &&
+ perf_evsel__realloc_id(evsel, cpu_map__nr(cpus),
+ threads->max_nr) < 0)
return -ENOMEM;
}
+ if (re_alloc)
+ perf_evlist__reset_heads(evlist, cpu_map__nr(cpus));
+
if (cpu_map__all(cpus))
- return perf_evlist__mmap_per_thread(evlist, prot, mask);
+ return perf_evlist__mmap_per_thread(evlist, prot, mask,
nr_append);
return perf_evlist__mmap_per_cpu(evlist, prot, mask);
}
@@ -572,6 +718,9 @@ int perf_evlist__create_maps(struct perf_evlist *evlist,
if (evlist->threads == NULL)
return -1;
+ if (evlist->threads->map[0] == -1)
+ set_bit(0, evlist->threads->bitmap);
+
if (perf_target__has_task(target))
evlist->cpus = cpu_map__dummy_new();
else if (!perf_target__has_cpu(target) && !target->uses_mmap)
@@ -601,14 +750,14 @@ int perf_evlist__apply_filters(struct perf_evlist *evlist)
{
struct perf_evsel *evsel;
int err = 0;
- const int ncpus = cpu_map__nr(evlist->cpus),
- nthreads = evlist->threads->nr;
+ const int ncpus = cpu_map__nr(evlist->cpus);
+ BITMAP *thread_bitmap = evlist->threads->bitmap;
list_for_each_entry(evsel, &evlist->entries, node) {
if (evsel->filter == NULL)
continue;
- err = perf_evsel__set_filter(evsel, ncpus, nthreads,
evsel->filter);
+ err = perf_evsel__set_filter(evsel, ncpus, thread_bitmap,
evsel->filter);
if (err)
break;
}
@@ -620,11 +769,11 @@ int perf_evlist__set_filter(struct perf_evlist *evlist,
const char *filter)
{
struct perf_evsel *evsel;
int err = 0;
- const int ncpus = cpu_map__nr(evlist->cpus),
- nthreads = evlist->threads->nr;
+ const int ncpus = cpu_map__nr(evlist->cpus);
+ BITMAP *thread_bitmap = evlist->threads->bitmap;
list_for_each_entry(evsel, &evlist->entries, node) {
- err = perf_evsel__set_filter(evsel, ncpus, nthreads, filter);
+ err = perf_evsel__set_filter(evsel, ncpus, thread_bitmap,
filter);
if (err)
break;
}
@@ -707,7 +856,7 @@ void perf_evlist__set_selected(struct perf_evlist *evlist,
int perf_evlist__open(struct perf_evlist *evlist)
{
struct perf_evsel *evsel;
- int err, ncpus, nthreads;
+ int err, ncpus;
list_for_each_entry(evsel, &evlist->entries, node) {
err = perf_evsel__open(evsel, evlist->cpus, evlist->threads);
@@ -718,10 +867,9 @@ int perf_evlist__open(struct perf_evlist *evlist)
return 0;
out_err:
ncpus = evlist->cpus ? evlist->cpus->nr : 1;
- nthreads = evlist->threads ? evlist->threads->nr : 1;
list_for_each_entry_reverse(evsel, &evlist->entries, node)
- perf_evsel__close(evsel, ncpus, nthreads);
+ perf_evsel__close(evsel, ncpus, evlist->threads->bitmap);
errno = -err;
return err;
diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index 56003f7..5df4f2b 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -2,12 +2,14 @@
#define __PERF_EVLIST_H 1
#include <linux/list.h>
+#include <linux/bitops.h>
#include <stdio.h>
#include "../perf.h"
#include "event.h"
#include "evsel.h"
#include "util.h"
#include <unistd.h>
+#include "thread_map.h"
struct pollfd;
struct thread_map;
@@ -18,23 +20,23 @@ struct perf_record_opts;
#define PERF_EVLIST__HLIST_SIZE (1 << PERF_EVLIST__HLIST_BITS)
struct perf_evlist {
- struct list_head entries;
- struct hlist_head heads[PERF_EVLIST__HLIST_SIZE];
- int nr_entries;
- int nr_fds;
- int nr_mmaps;
- int mmap_len;
+ struct list_head entries;
+ struct hlist_head heads[PERF_EVLIST__HLIST_SIZE];
+ int nr_entries;
+ int nr_fds;
+ int nr_mmaps;
+ int mmap_len;
struct {
- int cork_fd;
- pid_t pid;
+ int cork_fd;
+ pid_t pid;
} workload;
- bool overwrite;
- union perf_event event_copy;
- struct perf_mmap *mmap;
- struct pollfd *pollfd;
- struct thread_map *threads;
- struct cpu_map *cpus;
- struct perf_evsel *selected;
+ bool overwrite;
+ union perf_event event_copy;
+ struct perf_mmap *mmap;
+ struct pollfd *pollfd;
+ struct thread_map *threads;
+ struct cpu_map *cpus;
+ struct perf_evsel *selected;
};
struct perf_evsel_str_handler {
@@ -68,8 +70,13 @@ perf_evlist__find_tracepoint_by_id(struct perf_evlist
*evlist, int id);
void perf_evlist__id_add(struct perf_evlist *evlist, struct perf_evsel *evsel,
int cpu, int thread, u64 id);
+void perf_evlist__id_remove(struct perf_evsel *evsel,
+ int cpu, int thread);
+
void perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd);
+void perf_evlist__remove_pollfd(struct perf_evlist *evlist, int nr_thread);
+
struct perf_evsel *perf_evlist__id2evsel(struct perf_evlist *evlist, u64 id);
union perf_event *perf_evlist__mmap_read(struct perf_evlist *self, int idx);
@@ -85,7 +92,7 @@ int perf_evlist__prepare_workload(struct perf_evlist *evlist,
int perf_evlist__start_workload(struct perf_evlist *evlist);
int perf_evlist__mmap(struct perf_evlist *evlist, unsigned int pages,
- bool overwrite);
+ bool overwrite, int nr_append, bool re_alloc);
void perf_evlist__munmap(struct perf_evlist *evlist);
void perf_evlist__disable(struct perf_evlist *evlist);
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index a34167f..6dd366e 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -8,7 +8,7 @@
*/
#include <byteswap.h>
-#include <linux/bitops.h>
+#include <linux/bitmap.h>
#include "asm/bug.h"
#include "debugfs.h"
#include "event-parse.h"
@@ -21,8 +21,7 @@
#include <linux/hw_breakpoint.h>
#include <linux/perf_event.h>
#include "perf_regs.h"
-
-#define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))
+#include "debug.h"
static int __perf_evsel__sample_size(u64 sample_type)
{
@@ -546,13 +545,29 @@ int perf_evsel__alloc_fd(struct perf_evsel *evsel, int
ncpus, int nthreads)
return evsel->fd != NULL ? 0 : -ENOMEM;
}
-int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, int nthreads,
+int perf_evsel__realloc_fd(struct perf_evsel *evsel,
+ int ncpus, int max_nthreads)
+{
+ int old_nthreads = evsel->fd->row_size;
+ int cpu, thread;
+
+ if (xyarray__realloc(&(evsel->fd), ncpus, ncpus, max_nthreads) < 0)
+ return -1;
+
+ for (cpu = 0; cpu < ncpus; cpu++)
+ for (thread = old_nthreads; thread < max_nthreads; thread++)
+ FD(evsel, cpu, thread) = -1;
+
+ return 0;
+}
+
+int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus,
+ BITMAP *thread_bitmap,
const char *filter)
{
int cpu, thread;
for (cpu = 0; cpu < ncpus; cpu++) {
- for (thread = 0; thread < nthreads; thread++) {
+ for_each_set_bit(thread, thread_bitmap, PID_MAX_DEFAULT) {
int fd = FD(evsel, cpu, thread),
err = ioctl(fd, PERF_EVENT_IOC_SET_FILTER, filter);
@@ -570,12 +585,33 @@ int perf_evsel__alloc_id(struct perf_evsel *evsel, int
ncpus, int nthreads)
if (evsel->sample_id == NULL)
return -ENOMEM;
- evsel->id = zalloc(ncpus * nthreads * sizeof(u64));
+ evsel->id = zalloc(ID_MAX_DEFAULT * sizeof(u64));
if (evsel->id == NULL) {
xyarray__delete(evsel->sample_id);
evsel->sample_id = NULL;
return -ENOMEM;
}
+ evsel->ids = ncpus * nthreads;
+
+ return 0;
+}
+
+int perf_evsel__realloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
+{
+ u64 *id = NULL;
+ size_t old_nthreads = evsel->sample_id->row_size /
+ sizeof(struct perf_sample_id);
+
+ if (xyarray__realloc(&(evsel->sample_id), ncpus, ncpus, nthreads) < 0)
+ return -ENOMEM;
+
+ id = realloc(evsel->id, ncpus * nthreads * sizeof(u64));
+ if (id == NULL) {
+ xyarray__realloc(&(evsel->sample_id), ncpus, ncpus,
+ old_nthreads);
+ return -ENOMEM;
+ }
+
+ evsel->id = id;
+ evsel->ids = ncpus * nthreads;
return 0;
}
@@ -601,14 +637,17 @@ void perf_evsel__free_id(struct perf_evsel *evsel)
evsel->id = NULL;
}
-void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
+void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus,
+ BITMAP *thread_bitmap)
{
int cpu, thread;
for (cpu = 0; cpu < ncpus; cpu++)
- for (thread = 0; thread < nthreads; ++thread) {
- close(FD(evsel, cpu, thread));
- FD(evsel, cpu, thread) = -1;
+ for_each_set_bit(thread, thread_bitmap, PID_MAX_DEFAULT) {
+ if (FD(evsel, cpu, thread)) {
+ close(FD(evsel, cpu, thread));
+ FD(evsel, cpu, thread) = -1;
+ }
}
}
@@ -659,7 +698,7 @@ int __perf_evsel__read_on_cpu(struct perf_evsel *evsel,
}
int __perf_evsel__read(struct perf_evsel *evsel,
- int ncpus, int nthreads, bool scale)
+ int ncpus, BITMAP *thread_bitmap, bool scale)
{
size_t nv = scale ? 3 : 1;
int cpu, thread;
@@ -668,7 +707,7 @@ int __perf_evsel__read(struct perf_evsel *evsel,
aggr->val = aggr->ena = aggr->run = 0;
for (cpu = 0; cpu < ncpus; cpu++) {
- for (thread = 0; thread < nthreads; thread++) {
+ for_each_set_bit(thread, thread_bitmap, PID_MAX_DEFAULT) {
if (FD(evsel, cpu, thread) < 0)
continue;
@@ -722,6 +761,68 @@ static int get_group_fd(struct perf_evsel *evsel, int cpu,
int thread)
return fd;
}
+int perf_evsel__append_open(struct perf_evsel *evsel,
+ struct cpu_map *cpus, struct thread_map *threads,
+ int append_nr, bool ralloc_need)
+{
+ int cpu;
+ unsigned long flags = 0;
+ int pid = -1, err;
+
+ if (ralloc_need)
+ if (perf_evsel__realloc_fd(evsel, cpus->nr, threads->max_nr) <
0)
+ return -ENOMEM;
+
+ if (evsel->cgrp) {
+ flags = PERF_FLAG_PID_CGROUP;
+ pid = evsel->cgrp->fd;
+ }
+
+ for (cpu = 0; cpu < cpus->nr; cpu++) {
+ int group_fd;
+
+ if (!evsel->cgrp)
+ pid = threads->map[append_nr];
+
+ group_fd = get_group_fd(evsel, cpu, append_nr);
+
+ FD(evsel, cpu, append_nr) = sys_perf_event_open(&evsel->attr,
+ pid,
cpus->map[cpu],
+ group_fd,
flags);
+
+ if (FD(evsel, cpu, append_nr) < 0) {
+ err = errno;
+ FD(evsel, cpu, append_nr) = -1;
+
+ if (err == ESRCH) {
+ int tid = threads->map[append_nr];
+
+ ui__error("A ESRCH error is got. May be the "
+ "target task [%d] exited.\n",
+ tid);
+ return err;
+ } else if (err == EMFILE) {
+ ui__error("Too many events (threads) are
opened.\n"
+ "Try again after reducing the number
of events\n");
+ goto out_err;
+ }
+
+ ui__error("The sys_perf_event_open() syscall "
+ "returned with %d (%s). /bin/dmesg "
+ "may provide additional information.\n"
+ "No CONFIG_PERF_EVENTS=y kernel support "
+ "configured?\n", err, strerror(err));
+
+ goto out_err;
+ }
+ }
+
+ return 0;
+out_err:
+ exit_browser(0);
+ exit(0);
+}
+
static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
struct thread_map *threads)
{
@@ -730,7 +831,7 @@ static int __perf_evsel__open(struct perf_evsel *evsel,
struct cpu_map *cpus,
int pid = -1, err;
if (evsel->fd == NULL &&
- perf_evsel__alloc_fd(evsel, cpus->nr, threads->nr) < 0)
+ perf_evsel__alloc_fd(evsel, cpus->nr, threads->max_nr) < 0)
return -ENOMEM;
if (evsel->cgrp) {
@@ -739,8 +840,7 @@ static int __perf_evsel__open(struct perf_evsel *evsel,
struct cpu_map *cpus,
}
for (cpu = 0; cpu < cpus->nr; cpu++) {
-
- for (thread = 0; thread < threads->nr; thread++) {
+ for_each_set_bit(thread, threads->bitmap, PID_MAX_DEFAULT) {
int group_fd;
if (!evsel->cgrp)
@@ -763,21 +863,21 @@ static int __perf_evsel__open(struct perf_evsel *evsel,
struct cpu_map *cpus,
out_close:
do {
- while (--thread >= 0) {
- close(FD(evsel, cpu, thread));
- FD(evsel, cpu, thread) = -1;
- }
- thread = threads->nr;
+ for_each_set_bit(thread, threads->bitmap, PID_MAX_DEFAULT)
+ if (FD(evsel, cpu, thread) != -1) {
+ close(FD(evsel, cpu, thread));
+ FD(evsel, cpu, thread) = -1;
+ }
} while (--cpu >= 0);
return err;
}
-void perf_evsel__close(struct perf_evsel *evsel, int ncpus, int nthreads)
+void perf_evsel__close(struct perf_evsel *evsel, int ncpus,
+ BITMAP *thread_bitmap)
{
if (evsel->fd == NULL)
return;
- perf_evsel__close_fd(evsel, ncpus, nthreads);
+ perf_evsel__close_fd(evsel, ncpus, thread_bitmap);
perf_evsel__free_fd(evsel);
evsel->fd = NULL;
}
@@ -803,7 +903,7 @@ static struct {
.map = {
.max_nr = MAX_THREADS_NR_DEFAULT,
.nr = 1,
- .bitmap = empty_thread_bitmap,
+ .bitmap = empty_thread_bitmap,
},
.threads = { -1, },
};
@@ -830,6 +930,7 @@ int perf_evsel__open_per_cpu(struct perf_evsel *evsel,
{
bitmap_zero(empty_thread_map.map.bitmap, PID_MAX_DEFAULT);
set_bit(0, empty_thread_map.map.bitmap);
+
return __perf_evsel__open(evsel, cpus, &empty_thread_map.map);
}
diff --git a/tools/perf/util/evsel.h b/tools/perf/util/evsel.h
index 3d2b801..065d27b 100644
--- a/tools/perf/util/evsel.h
+++ b/tools/perf/util/evsel.h
@@ -2,6 +2,7 @@
#define __PERF_EVSEL_H 1
#include <linux/list.h>
+#include <linux/bitops.h>
#include <stdbool.h>
#include <stddef.h>
#include <linux/perf_event.h>
@@ -9,6 +10,17 @@
#include "xyarray.h"
#include "cgroup.h"
#include "hist.h"
+
+#define ID_MAX_DEFAULT (CPU_MAX_DEFAULT * PID_MAX_DEFAULT)
+
+/*
+ * find the bit position of (cpu, thread)
+ */
+#define ID_BITMAP_POS(cpu, thread) \
+ (cpu * PID_MAX_DEFAULT + thread)
+
+#define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))
+#define SID(e, x, y) xyarray__entry(e->sample_id, x, y)
struct perf_counts_values {
union {
@@ -52,6 +64,7 @@ struct perf_evsel {
struct xyarray *fd;
struct xyarray *sample_id;
u64 *id;
+ DECLARE_BITMAP(id_bitmap, ID_MAX_DEFAULT);
struct perf_counts *counts;
int idx;
u32 ids;
@@ -112,13 +125,15 @@ int __perf_evsel__hw_cache_type_op_res_name(u8 type, u8
op, u8 result,
const char *perf_evsel__name(struct perf_evsel *evsel);
int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
+int perf_evsel__realloc_fd(struct perf_evsel *evsel, int ncpus, int
max_nthreads);
int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads);
+int perf_evsel__realloc_id(struct perf_evsel *evsel, int ncpus, int
max_nthreads);
int perf_evsel__alloc_counts(struct perf_evsel *evsel, int ncpus);
void perf_evsel__free_fd(struct perf_evsel *evsel);
void perf_evsel__free_id(struct perf_evsel *evsel);
-void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
+void perf_evsel__close_fd(struct perf_evsel *evsel, int ncpus, BITMAP
*thread_bitmap);
-int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, int nthreads,
+int perf_evsel__set_filter(struct perf_evsel *evsel, int ncpus, BITMAP
*thread_bitmap,
const char *filter);
int perf_evsel__open_per_cpu(struct perf_evsel *evsel,
@@ -127,7 +142,10 @@ int perf_evsel__open_per_thread(struct perf_evsel *evsel,
struct thread_map *threads);
int perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
struct thread_map *threads);
-void perf_evsel__close(struct perf_evsel *evsel, int ncpus, int nthreads);
+int perf_evsel__append_open(struct perf_evsel *evsel, struct cpu_map *cpus,
+ struct thread_map *threads, int append_nr,
+ bool need_realloc);
+void perf_evsel__close(struct perf_evsel *evsel, int ncpus, BITMAP
*thread_bitmap);
struct perf_sample;
@@ -187,7 +205,7 @@ static inline int perf_evsel__read_on_cpu_scaled(struct
perf_evsel *evsel,
return __perf_evsel__read_on_cpu(evsel, cpu, thread, true);
}
-int __perf_evsel__read(struct perf_evsel *evsel, int ncpus, int nthreads,
+int __perf_evsel__read(struct perf_evsel *evsel, int ncpus, BITMAP
*thread_bitmap,
bool scale);
/**
@@ -195,12 +213,12 @@ int __perf_evsel__read(struct perf_evsel *evsel, int
ncpus, int nthreads,
*
* @evsel - event selector to read value
* @ncpus - Number of cpus affected, from zero
- * @nthreads - Number of threads affected, from zero
+ * @thread_bitmap - Bitmap of threads map affected.
*/
static inline int perf_evsel__read(struct perf_evsel *evsel,
- int ncpus, int nthreads)
+ int ncpus, BITMAP *thread_bitmap)
{
- return __perf_evsel__read(evsel, ncpus, nthreads, false);
+ return __perf_evsel__read(evsel, ncpus, thread_bitmap, false);
}
/**
@@ -208,12 +226,12 @@ static inline int perf_evsel__read(struct perf_evsel
*evsel,
*
* @evsel - event selector to read value
* @ncpus - Number of cpus affected, from zero
- * @nthreads - Number of threads affected, from zero
+ * @thread_bitmap - Bitmap of threads map affected.
*/
static inline int perf_evsel__read_scaled(struct perf_evsel *evsel,
- int ncpus, int nthreads)
+ int ncpus, BITMAP *thread_bitmap)
{
- return __perf_evsel__read(evsel, ncpus, nthreads, true);
+ return __perf_evsel__read(evsel, ncpus, thread_bitmap, true);
}
void hists__init(struct hists *hists);
diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c
index a2657fd..f54b362 100644
--- a/tools/perf/util/python.c
+++ b/tools/perf/util/python.c
@@ -420,6 +420,7 @@ struct pyrf_thread_map {
PyObject_HEAD
struct thread_map *threads;
+ DECLARE_BITMAP(thread_bitmap, PID_MAX_DEFAULT);
};
static int pyrf_thread_map__init(struct pyrf_thread_map *pthreads,
@@ -704,7 +705,7 @@ static PyObject *pyrf_evlist__mmap(struct pyrf_evlist
*pevlist,
&pages, &overwrite))
return NULL;
- if (perf_evlist__mmap(evlist, pages, overwrite) < 0) {
+ if (perf_evlist__mmap(evlist, pages, overwrite, -1, false) < 0) {
PyErr_SetFromErrno(PyExc_OSError);
return NULL;
}
--
1.7.9.5
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
