Skip to contentSkip to conversation

perf lock: Random updates for the locking analysis (v2)

PATCH Applied Namhyung Kim 2 participants 8 messages Jan 27, 2022 – Feb 10, 2022 perf events

lore
Share & tools
lore

Namhyung Kim

author maintainer
#0 lore raw

cover perf lock: Random updates for the locking analysis (v2)

Hello,

I have some updates in the perf lock command (focused on 'report'). The main change is to add -c (or --combine-locks) option to aggregate results based on lock class name.

* changes from v1)
- rebased onto recent acme/perf/core
- add Jiri's Acked-by

Without this option, the result deals with lock addresses so instances in the same lock class will have separate entries like below:

  # perf lock report
                  Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns) 
         rcu_read_lock     251225          0               0               0               0               0 
   &(ei->i_block_re...       8731          0               0               0               0               0 
   &sb->s_type->i_l...       8731          0               0               0               0               0 
   hrtimer_bases.lo...       5261          0               0               0               0               0 
   hrtimer_bases.lo...       2626          0               0               0               0               0 
   hrtimer_bases.lo...       1953          0               0               0               0               0 
   hrtimer_bases.lo...       1382          0               0               0               0               0 
   cpu_hotplug_lock...       1350          0               0               0               0               0 
   hrtimer_bases.lo...       1273          0               0               0               0               0 
   hrtimer_bases.lo...       1269          0               0               0               0               0 
   hrtimer_bases.lo...       1198          0               0               0               0               0 
   hrtimer_bases.lo...       1116          0               0               0               0               0 
           &base->lock       1109          0               0               0               0               0 
   hrtimer_bases.lo...       1067          0               0               0               0               0 
   hrtimer_bases.lo...       1052          0               0               0               0               0 
   hrtimer_bases.lo...        957          0               0               0               0               0 
   hrtimer_bases.lo...        948          0               0               0               0               0 
          css_set_lock        791          0               0               0               0               0 
   hrtimer_bases.lo...        752          0               0               0               0               0 
   &lruvec->lru_loc...        747          5           11254           56272           18317            1412 
   hrtimer_bases.lo...        738          0               0               0               0               0 
   &newf->file_lock...        706         15            1025           15388            2279             618 
   hrtimer_bases.lo...        702          0               0               0               0               0 
   hrtimer_bases.lo...        694          0               0               0               0               0 

...

With -c option, the hrtimer_bases.lock would be combined into a single entry. Also note that the lock names are correctly displayed now.

  # perf lock report -c
                  Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns) 
         rcu_read_lock     251225          0               0               0               0               0 
    hrtimer_bases.lock      39449          0               0               0               0               0 
   &sb->s_type->i_l...      10301          1             662             662             662             662 
      ptlock_ptr(page)      10173          2             701            1402             760             642 
   &(ei->i_block_re...       8732          0               0               0               0               0 
           &base->lock       6705          0               0               0               0               0 
           &p->pi_lock       5549          0               0               0               0               0 
   &dentry->d_lockr...       5010          4            1274            5097            1844             789 
             &ep->lock       2750          0               0               0               0               0 
   &(__futex_data.q...       2331          0               0               0               0               0 
                (null)       1878          0               0               0               0               0 
      cpu_hotplug_lock       1350          0               0               0               0               0 
      &____s->seqcount       1349          0               0               0               0               0 
      &newf->file_lock       1001         15            1025           15388            2279             618 

...

Maybe we can make it default later (with a config and --no-combine-locks).

You can get it from 'perf/lock-combine-v2' branch at

git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git

Thanks,
Namhyung

Namhyung Kim (6):
perf lock: Convert lockhash_table to use hlist
perf lock: Change type of lock_stat->addr to u64 perf lock: Sort map info based on class name
perf lock: Fix lock name length check for printing perf lock: Add -c/--combine-locks option
perf lock: Carefully combine lock stats for discarded entries

 tools/perf/Documentation/perf-lock.txt |   4 +
 tools/perf/builtin-lock.c              | 155 +++++++++++++++++++------
 2 files changed, 124 insertions(+), 35 deletions(-)

base-commit: e783362eb54cd99b2cac8b3a9aeac942e6f6ac07

Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
author maintainer
#1 lore raw

PATCH 1/6 perf lock: Convert lockhash_table to use hlist

The hlist_head has a single entry so we can save some memory.

---
 tools/perf/builtin-lock.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)
tools/perf/builtin-lock.c +8 -8
index d70131b7b1b1..43139166f02e 100644
@@ -38,13 +38,13 @@ static struct perf_session *session;
 #define LOCKHASH_BITS		12
 #define LOCKHASH_SIZE		(1UL << LOCKHASH_BITS)
 
-static struct list_head lockhash_table[LOCKHASH_SIZE];
+static struct hlist_head lockhash_table[LOCKHASH_SIZE];
 
 #define __lockhashfn(key)	hash_long((unsigned long)key, LOCKHASH_BITS)
 #define lockhashentry(key)	(lockhash_table + __lockhashfn((key)))
 
 struct lock_stat {
-	struct list_head	hash_entry;
+	struct hlist_node	hash_entry;
 	struct rb_node		rb;		/* used for sorting */
 
 	/*
@@ -317,10 +317,10 @@ static struct lock_stat *pop_from_result(void)
 
 static struct lock_stat *lock_stat_findnew(void *addr, const char *name)
 {
-	struct list_head *entry = lockhashentry(addr);
+	struct hlist_head *entry = lockhashentry(addr);
 	struct lock_stat *ret, *new;
 
-	list_for_each_entry(ret, entry, hash_entry) {
+	hlist_for_each_entry(ret, entry, hash_entry) {
 		if (ret->addr == addr)
 			return ret;
 	}
@@ -339,7 +339,7 @@ static struct lock_stat *lock_stat_findnew(void *addr, const char *name)
 	strcpy(new->name, name);
 	new->wait_time_min = ULLONG_MAX;
 
-	list_add(&new->hash_entry, entry);
+	hlist_add_head(&new->hash_entry, entry);
 	return new;
 
 alloc_failed:
@@ -781,7 +781,7 @@ static void dump_map(void)
 
 	pr_info("Address of instance: name of class\n");
 	for (i = 0; i < LOCKHASH_SIZE; i++) {
-		list_for_each_entry(st, &lockhash_table[i], hash_entry) {
+		hlist_for_each_entry(st, &lockhash_table[i], hash_entry) {
 			pr_info(" %p: %s\n", st->addr, st->name);
 		}
 	}
@@ -838,7 +838,7 @@ static void sort_result(void)
 	struct lock_stat *st;
 
 	for (i = 0; i < LOCKHASH_SIZE; i++) {
-		list_for_each_entry(st, &lockhash_table[i], hash_entry) {
+		hlist_for_each_entry(st, &lockhash_table[i], hash_entry) {
 			insert_to_result(st, compare);
 		}
 	}
@@ -990,7 +990,7 @@ int cmd_lock(int argc, const char **argv)
 	int rc = 0;
 
 	for (i = 0; i < LOCKHASH_SIZE; i++)
-		INIT_LIST_HEAD(lockhash_table + i);
+		INIT_HLIST_HEAD(lockhash_table + i);
 
 	argc = parse_options_subcommand(argc, argv, lock_options, lock_subcommands,
 					lock_usage, PARSE_OPT_STOP_AT_NON_OPTION);
Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
author maintainer
#2 lore raw

PATCH 2/6 perf lock: Change type of lock_stat->addr to u64

In reply to Namhyung Kim

As evsel__intval() returns u64, we can just use it as is.

---
 tools/perf/builtin-lock.c | 35 +++++++++--------------------------
 1 file changed, 9 insertions(+), 26 deletions(-)
tools/perf/builtin-lock.c +9 -26
index 43139166f02e..c4b5c3d71ae3 100644
@@ -47,12 +47,7 @@ struct lock_stat {
 	struct hlist_node	hash_entry;
 	struct rb_node		rb;		/* used for sorting */
 
-	/*
-	 * FIXME: evsel__intval() returns u64,
-	 * so address of lockdep_map should be treated as 64bit.
-	 * Is there more better solution?
-	 */
-	void			*addr;		/* address of lockdep_map, used as ID */
+	u64			addr;		/* address of lockdep_map, used as ID */
 	char			*name;		/* for strcpy(), we cannot use const */
 
 	unsigned int		nr_acquire;
@@ -106,7 +101,7 @@ struct lock_seq_stat {
 	struct list_head        list;
 	int			state;
 	u64			prev_event_time;
-	void                    *addr;
+	u64                     addr;
 
 	int                     read_count;
 };
@@ -315,7 +310,7 @@ static struct lock_stat *pop_from_result(void)
 	return container_of(node, struct lock_stat, rb);
 }
 
-static struct lock_stat *lock_stat_findnew(void *addr, const char *name)
+static struct lock_stat *lock_stat_findnew(u64 addr, const char *name)
 {
 	struct hlist_head *entry = lockhashentry(addr);
 	struct lock_stat *ret, *new;
@@ -361,7 +356,7 @@ struct trace_lock_handler {
 			     struct perf_sample *sample);
 };
 
-static struct lock_seq_stat *get_seq(struct thread_stat *ts, void *addr)
+static struct lock_seq_stat *get_seq(struct thread_stat *ts, u64 addr)
 {
 	struct lock_seq_stat *seq;
 
@@ -400,16 +395,13 @@ enum acquire_flags {
 static int report_lock_acquire_event(struct evsel *evsel,
 				     struct perf_sample *sample)
 {
-	void *addr;
 	struct lock_stat *ls;
 	struct thread_stat *ts;
 	struct lock_seq_stat *seq;
 	const char *name = evsel__strval(evsel, sample, "name");
-	u64 tmp	 = evsel__intval(evsel, sample, "lockdep_addr");
+	u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
 	int flag = evsel__intval(evsel, sample, "flags");
 
-	memcpy(&addr, &tmp, sizeof(void *));
-
 	ls = lock_stat_findnew(addr, name);
 	if (!ls)
 		return -ENOMEM;
@@ -472,15 +464,12 @@ static int report_lock_acquire_event(struct evsel *evsel,
 static int report_lock_acquired_event(struct evsel *evsel,
 				      struct perf_sample *sample)
 {
-	void *addr;
 	struct lock_stat *ls;
 	struct thread_stat *ts;
 	struct lock_seq_stat *seq;
 	u64 contended_term;
 	const char *name = evsel__strval(evsel, sample, "name");
-	u64 tmp = evsel__intval(evsel, sample, "lockdep_addr");
-
-	memcpy(&addr, &tmp, sizeof(void *));
+	u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
 
 	ls = lock_stat_findnew(addr, name);
 	if (!ls)
@@ -535,14 +524,11 @@ static int report_lock_acquired_event(struct evsel *evsel,
 static int report_lock_contended_event(struct evsel *evsel,
 				       struct perf_sample *sample)
 {
-	void *addr;
 	struct lock_stat *ls;
 	struct thread_stat *ts;
 	struct lock_seq_stat *seq;
 	const char *name = evsel__strval(evsel, sample, "name");
-	u64 tmp = evsel__intval(evsel, sample, "lockdep_addr");
-
-	memcpy(&addr, &tmp, sizeof(void *));
+	u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
 
 	ls = lock_stat_findnew(addr, name);
 	if (!ls)
@@ -590,14 +576,11 @@ static int report_lock_contended_event(struct evsel *evsel,
 static int report_lock_release_event(struct evsel *evsel,
 				     struct perf_sample *sample)
 {
-	void *addr;
 	struct lock_stat *ls;
 	struct thread_stat *ts;
 	struct lock_seq_stat *seq;
 	const char *name = evsel__strval(evsel, sample, "name");
-	u64 tmp = evsel__intval(evsel, sample, "lockdep_addr");
-
-	memcpy(&addr, &tmp, sizeof(void *));
+	u64 addr = evsel__intval(evsel, sample, "lockdep_addr");
 
 	ls = lock_stat_findnew(addr, name);
 	if (!ls)
@@ -782,7 +765,7 @@ static void dump_map(void)
 	pr_info("Address of instance: name of class\n");
 	for (i = 0; i < LOCKHASH_SIZE; i++) {
 		hlist_for_each_entry(st, &lockhash_table[i], hash_entry) {
-			pr_info(" %p: %s\n", st->addr, st->name);
+			pr_info(" %#llx: %s\n", (unsigned long long)st->addr, st->name);
 		}
 	}
 }
Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
author maintainer
#3 lore raw

PATCH 3/6 perf lock: Sort map info based on class name

In reply to Namhyung Kim

Instead of the random order, sort it by lock class name.

Before:

  # perf lock info -m
  Address of instance: name of class
   0xffffa0d940ac5310: &dentry->d_lockref.lock
   0xffffa0c20b0e1cb0: &dentry->d_lockref.lock
   0xffffa0d8e051cc48: &base->lock
   0xffffa0d94f992110: &anon_vma->rwsem
   0xffffa0d947a4f278: (null)
   0xffffa0c208f6e108: &map->lock
   0xffffa0c213ad32c8: &cfs_rq->removed.lock
   0xffffa0c20d695888: &parent->list_lock
   0xffffa0c278775278: (null)
   0xffffa0c212ad4690: &dentry->d_lockref.lock

After:

  # perf lock info -m
  Address of instance: name of class
   0xffffa0c20d538800: &(&sig->stats_lock)->lock
   0xffffa0c216d4ec40: &(&sig->stats_lock)->lock
   0xffffa1fe4cb04610: &(__futex_data.queues)[i].lock
   0xffffa1fe4cb07750: &(__futex_data.queues)[i].lock
   0xffffa1fe4cb07b50: &(__futex_data.queues)[i].lock
   0xffffa1fe4cb0b850: &(__futex_data.queues)[i].lock
   0xffffa1fe4cb0bcd0: &(__futex_data.queues)[i].lock
   0xffffa1fe4cb0e5d0: &(__futex_data.queues)[i].lock
   0xffffa1fe4cb11ad0: &(__futex_data.queues)[i].lock
---
 tools/perf/builtin-lock.c | 20 +++++++++++++++++++-
 1 file changed, 19 insertions(+), 1 deletion(-)
tools/perf/builtin-lock.c +19 -1
index c4b5c3d71ae3..8078f7ca826d 100644
@@ -757,6 +757,21 @@ static void dump_threads(void)
 	}
 }
 
+static int compare_maps(struct lock_stat *a, struct lock_stat *b)
+{
+	int ret;
+
+	if (a->name && b->name)
+		ret = strcmp(a->name, b->name);
+	else
+		ret = !!a->name - !!b->name;
+
+	if (!ret)
+		return a->addr < b->addr;
+	else
+		return ret < 0;
+}
+
 static void dump_map(void)
 {
 	unsigned int i;
@@ -765,9 +780,12 @@ static void dump_map(void)
 	pr_info("Address of instance: name of class\n");
 	for (i = 0; i < LOCKHASH_SIZE; i++) {
 		hlist_for_each_entry(st, &lockhash_table[i], hash_entry) {
-			pr_info(" %#llx: %s\n", (unsigned long long)st->addr, st->name);
+			insert_to_result(st, compare_maps);
 		}
 	}
+
+	while ((st = pop_from_result()))
+		pr_info(" %#llx: %s\n", (unsigned long long)st->addr, st->name);
 }
 
 static int dump_info(void)
Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
author maintainer
#4 lore raw

PATCH 4/6 perf lock: Fix lock name length check for printing

In reply to Namhyung Kim

It has 20 character spaces for name so lock names shorter than 20 should be printed without ellipsis.

Before:

 # perf lock report
                Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns)
       rcu_read_lock     251225          0               0               0               0               0
 &(ei->i_block_re...       8731          0               0               0               0               0
 &sb->s_type->i_l...       8731          0               0               0               0               0
 hrtimer_bases.lo...       5261          0               0               0               0               0
 hrtimer_bases.lo...       2626          0               0               0               0               0
 hrtimer_bases.lo...       1953          0               0               0               0               0
 hrtimer_bases.lo...       1382          0               0               0               0               0
 cpu_hotplug_lock...       1350          0               0               0               0               0

After:

 # perf lock report
                Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns)
       rcu_read_lock     251225          0               0               0               0               0
 &(ei->i_block_re...       8731          0               0               0               0               0
 &sb->s_type->i_l...       8731          0               0               0               0               0
  hrtimer_bases.lock       5261          0               0               0               0               0
  hrtimer_bases.lock       2626          0               0               0               0               0
  hrtimer_bases.lock       1953          0               0               0               0               0
  hrtimer_bases.lock       1382          0               0               0               0               0
    cpu_hotplug_lock       1350          0               0               0               0               0
---
 tools/perf/builtin-lock.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/builtin-lock.c +1 -1
index 8078f7ca826d..f6adf3cdd1e2 100644
@@ -710,7 +710,7 @@ static void print_result(void)
 		}
 		bzero(cut_name, 20);
 
-		if (strlen(st->name) < 16) {
+		if (strlen(st->name) < 20) {
 			/* output raw name */
 			pr_info("%20s ", st->name);
 		} else {
Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
author maintainer
#5 lore raw

PATCH 5/6 perf lock: Add -c/--combine-locks option

In reply to Namhyung Kim

The -c or --combine-locks option is to merge lock instances in the same class into a single entry. It compares the name of the locks and marks duplicated entries using lock_stat->combined.

 # perf lock report
                Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns)
       rcu_read_lock     251225          0               0               0               0               0
 &(ei->i_block_re...       8731          0               0               0               0               0
 &sb->s_type->i_l...       8731          0               0               0               0               0
  hrtimer_bases.lock       5261          0               0               0               0               0
  hrtimer_bases.lock       2626          0               0               0               0               0
  hrtimer_bases.lock       1953          0               0               0               0               0
  hrtimer_bases.lock       1382          0               0               0               0               0
    cpu_hotplug_lock       1350          0               0               0               0               0
  hrtimer_bases.lock       1273          0               0               0               0               0
  hrtimer_bases.lock       1269          0               0               0               0               0
 # perf lock report -c
                Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns)
       rcu_read_lock     251225          0               0               0               0               0
  hrtimer_bases.lock      39450          0               0               0               0               0
 &sb->s_type->i_l...      10301          1             662             662             662             662
    ptlock_ptr(page)      10173          2             701            1402             760             642
 &(ei->i_block_re...       8732          0               0               0               0               0
        &xa->xa_lock       8088          0               0               0               0               0
         &base->lock       6705          0               0               0               0               0
         &p->pi_lock       5549          0               0               0               0               0
 &dentry->d_lockr...       5010          4            1274            5097            1844             789
           &ep->lock       3958          0               0               0               0               0
---
 tools/perf/Documentation/perf-lock.txt |  4 ++
 tools/perf/builtin-lock.c              | 68 ++++++++++++++++++++++++++
 2 files changed, 72 insertions(+)
tools/perf/Documentation/perf-lock.txt +4 -0
index 1b4d452923d7..f5eb95788969 100644
@@ -54,6 +54,10 @@ REPORT OPTIONS
         Sorting key. Possible values: acquired (default), contended,
 	avg_wait, wait_total, wait_max, wait_min.
 
+-c::
+--combine-locks::
+	Merge lock instances in the same class (based on name).
+
 INFO OPTIONS
 ------------
 
tools/perf/builtin-lock.c +68 -0

116 lines not on this page · raw message

Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
author maintainer
#6 lore raw

PATCH 6/6 perf lock: Carefully combine lock stats for discarded entries

In reply to Namhyung Kim

Stats from discarded entries should be omitted. But a lock class may have both good and bad entries. If the first entry was bad, we can zero-fill the stats and only add good stats if any. The entry can remove the discard state if it finds a good entry later.

---
 tools/perf/builtin-lock.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
tools/perf/builtin-lock.c +16 -0
index bbfeba79426a..57b9ebd7118a 100644
@@ -295,6 +295,9 @@ static void combine_lock_stats(struct lock_stat *st)
 			ret = !!st->name - !!p->name;
 
 		if (ret == 0) {
+			if (st->discard)
+				goto out;
+
 			p->nr_acquired += st->nr_acquired;
 			p->nr_contended += st->nr_contended;
 			p->wait_time_total += st->wait_time_total;
@@ -307,6 +310,10 @@ static void combine_lock_stats(struct lock_stat *st)
 			if (p->wait_time_max < st->wait_time_max)
 				p->wait_time_max = st->wait_time_max;
 
+			/* now it got a new !discard record */
+			p->discard = 0;
+
+out:
 			st->combined = 1;
 			return;
 		}
@@ -319,6 +326,15 @@ static void combine_lock_stats(struct lock_stat *st)
 
 	rb_link_node(&st->rb, parent, rb);
 	rb_insert_color(&st->rb, &sorted);
+
+	if (st->discard) {
+		st->nr_acquired = 0;
+		st->nr_contended = 0;
+		st->wait_time_total = 0;
+		st->avg_wait_time = 0;
+		st->wait_time_min = ULLONG_MAX;
+		st->wait_time_max = 0;
+	}
 }
 
 static void insert_to_result(struct lock_stat *st,
Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
maintainer Applied
#7 lore raw
In reply to Namhyung Kim

Em Wed, Jan 26, 2022 at 04:00:44PM -0800, Namhyung Kim escreveu:

Quoted 9 lines

Hello,

I have some updates in the perf lock command (focused on 'report'). The main change is to add -c (or --combine-locks) option to aggregate results based on lock class name.

* changes from v1)
- rebased onto recent acme/perf/core
- add Jiri's Acked-by

Thanks, applied.

- Arnaldo

Quoted 80 lines

Without this option, the result deals with lock addresses so instances in the same lock class will have separate entries like below:

  # perf lock report
                  Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns) 
         rcu_read_lock     251225          0               0               0               0               0 
   &(ei->i_block_re...       8731          0               0               0               0               0 
   &sb->s_type->i_l...       8731          0               0               0               0               0 
   hrtimer_bases.lo...       5261          0               0               0               0               0 
   hrtimer_bases.lo...       2626          0               0               0               0               0 
   hrtimer_bases.lo...       1953          0               0               0               0               0 
   hrtimer_bases.lo...       1382          0               0               0               0               0 
   cpu_hotplug_lock...       1350          0               0               0               0               0 
   hrtimer_bases.lo...       1273          0               0               0               0               0 
   hrtimer_bases.lo...       1269          0               0               0               0               0 
   hrtimer_bases.lo...       1198          0               0               0               0               0 
   hrtimer_bases.lo...       1116          0               0               0               0               0 
           &base->lock       1109          0               0               0               0               0 
   hrtimer_bases.lo...       1067          0               0               0               0               0 
   hrtimer_bases.lo...       1052          0               0               0               0               0 
   hrtimer_bases.lo...        957          0               0               0               0               0 
   hrtimer_bases.lo...        948          0               0               0               0               0 
          css_set_lock        791          0               0               0               0               0 
   hrtimer_bases.lo...        752          0               0               0               0               0 
   &lruvec->lru_loc...        747          5           11254           56272           18317            1412 
   hrtimer_bases.lo...        738          0               0               0               0               0 
   &newf->file_lock...        706         15            1025           15388            2279             618 
   hrtimer_bases.lo...        702          0               0               0               0               0 
   hrtimer_bases.lo...        694          0               0               0               0               0 

...

With -c option, the hrtimer_bases.lock would be combined into a single entry. Also note that the lock names are correctly displayed now.

  # perf lock report -c
                  Name   acquired  contended   avg wait (ns) total wait (ns)   max wait (ns)   min wait (ns) 
         rcu_read_lock     251225          0               0               0               0               0 
    hrtimer_bases.lock      39449          0               0               0               0               0 
   &sb->s_type->i_l...      10301          1             662             662             662             662 
      ptlock_ptr(page)      10173          2             701            1402             760             642 
   &(ei->i_block_re...       8732          0               0               0               0               0 
           &base->lock       6705          0               0               0               0               0 
           &p->pi_lock       5549          0               0               0               0               0 
   &dentry->d_lockr...       5010          4            1274            5097            1844             789 
             &ep->lock       2750          0               0               0               0               0 
   &(__futex_data.q...       2331          0               0               0               0               0 
                (null)       1878          0               0               0               0               0 
      cpu_hotplug_lock       1350          0               0               0               0               0 
      &____s->seqcount       1349          0               0               0               0               0 
      &newf->file_lock       1001         15            1025           15388            2279             618 

...

Maybe we can make it default later (with a config and --no-combine-locks).

You can get it from 'perf/lock-combine-v2' branch at

git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git

Thanks,
Namhyung

Namhyung Kim (6):
perf lock: Convert lockhash_table to use hlist
perf lock: Change type of lock_stat->addr to u64 perf lock: Sort map info based on class name
perf lock: Fix lock name length check for printing perf lock: Add -c/--combine-locks option
perf lock: Carefully combine lock stats for discarded entries

 tools/perf/Documentation/perf-lock.txt |   4 +
 tools/perf/builtin-lock.c              | 155 +++++++++++++++++++------
 2 files changed, 124 insertions(+), 35 deletions(-)

base-commit: e783362eb54cd99b2cac8b3a9aeac942e6f6ac07

Signature · 1 line
2.35.0.rc0.227.g00780c9af4-goog
Signature · 1 line
- Arnaldo