Skip to contentSkip to conversation

perf sched stats: Fix memory leaks

PATCH Superseded Namhyung Kim 3 participants 7 messages Jul 13 – Sep 10 perf events

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Namhyung Kim

author maintainer
#0 lore raw

cover perf sched stats: Fix memory leaks

Hello,

This is a small series to fix memory leaks in the perf sched stats. I've tested the changes with the address sanitizer.

Thanks,
Namhyung

Namhyung Kim (3):
perf sched: Add missing perf_session__delete() perf sched: Fix memory leaks in perf sched stats report perf sched: Free subcommand string after perf sched stats

 tools/perf/builtin-sched.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)
Signature · 1 line
2.55.0.795.g602f6c329a-goog
author maintainer
#1 lore raw

PATCH 1/3 perf sched: Add missing perf_session__delete()

The perf sched stats record missed to release the session and ASAN reported a leak.

Fixes c3030995f23b ("perf sched stats: Add record and rawdump support")
Cc Swapnil Sapkal <[email protected]>
---
 tools/perf/builtin-sched.c | 1 +
 1 file changed, 1 insertion(+)
tools/perf/builtin-sched.c +1 -0
index 6e0d8f4d270e2d88..68147a0be5c710ed 100644
@@ -4027,6 +4027,7 @@ static int perf_sched__schedstat_record(struct perf_sched *sched,
 	else
 		fprintf(stderr, "[ perf sched stats: Failed !! ]\n");
 
+	perf_session__delete(session);
 	evlist__put(evlist);
 	close(fd);
 	return err;
Signature · 1 line
2.55.0.795.g602f6c329a-goog
#2 lore raw

Re: [PATCH v1 1/3] perf sched: Add missing perf_session__delete()

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider: - [Medium] The patch introduces a double close of the file descriptor `fd` because…

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The patch introduces a double close of the file descriptor fd because perf_session__delete(session) inherently closes it before the explicit close(fd). --

--- Patch [1]: [PATCH v1 1/3] perf sched: Add missing perf_session__delete() --- commit 2da9bd6a7ce8e17e442d8be81b1b57837aaee6e1
Author: Namhyung Kim <[email protected]>

perf sched: Add missing perf_session__delete()

This patch addresses an ASAN-reported memory leak in the perf sched stats record command by explicitly deleting the session before returning.

tools/perf/builtin-sched.c static int perf_sched__schedstat_record(struct perf_sched *sched, · line 4032

Quoted 11 lines
tools/perf/builtin-sched.c +1 -0
index 6e0d8f4d270e2..68147a0be5c71 100644
@@ -4027,6 +4027,7 @@ static int perf_sched__schedstat_record(struct perf_sched *sched,
 	else
 		fprintf(stderr, "[ perf sched stats: Failed !! ]\n");
 
+	perf_session__delete(session);
 	evlist__put(evlist);
 	close(fd);

[Severity: Medium]
Does adding perf_session__delete() result in closing the file descriptor twice?

Since perf_session__delete() internally closes the file descriptor associated with the session data, the subsequent explicit close(fd) will attempt to close it again. This could return EBADF or erroneously close an unrelated file descriptor if another thread opens a file in the meantime.

Quoted 2 lines
tools/perf/builtin-sched.c +0 -0
 	return err;

}

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
author maintainer
#3 lore raw

Re: [PATCH v1 1/3] perf sched: Add missing perf_session__delete()

Quoted 28 lines · On Mon, Jul 13, 2026 at 08:59:07PM +0000, [email protected] wrote:

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The patch introduces a double close of the file descriptor fd because perf_session__delete(session) inherently closes it before the explicit close(fd). --

--- Patch [1]: [PATCH v1 1/3] perf sched: Add missing perf_session__delete() --- commit 2da9bd6a7ce8e17e442d8be81b1b57837aaee6e1
Author: Namhyung Kim <[email protected]>

perf sched: Add missing perf_session__delete()

This patch addresses an ASAN-reported memory leak in the perf sched stats record command by explicitly deleting the session before returning.

tools/perf/builtin-sched.c static int perf_sched__schedstat_record(struct perf_sched *sched, · line 4032

Quoted 11 lines
tools/perf/builtin-sched.c +1 -0
index 6e0d8f4d270e2..68147a0be5c71 100644
@@ -4027,6 +4027,7 @@ static int perf_sched__schedstat_record(struct perf_sched *sched,
 	else
 		fprintf(stderr, "[ perf sched stats: Failed !! ]\n");
 
+	perf_session__delete(session);
 	evlist__put(evlist);
 	close(fd);

[Severity: Medium]
Does adding perf_session__delete() result in closing the file descriptor twice?

Quoted 4 lines

Since perf_session__delete() internally closes the file descriptor associated with the session data, the subsequent explicit close(fd) will attempt to close it again. This could return EBADF or erroneously close an unrelated file descriptor if another thread opens a file in the meantime.

Hmm.. looks right. I think I can just remove the close().

Thanks,
Namhyung

#4 lore raw

Re: [PATCH v1 1/3] perf sched: Add missing perf_session__delete()

Quoted 14 lines · Namhyung Kim
Quoted 13 lines · On Mon, Jul 13, 2026 at 08:59:07PM +0000, [email protected] wrote:

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

tools/perf/builtin-sched.c static int perf_sched__schedstat_record(struct perf_sched *sched, · line 4032

Quoted 8 lines

+++ b/tools/perf/builtin-sched.c

tools/perf/builtin-sched.c +1 -0
@@ -4027,6 +4027,7 @@ static int perf_sched__schedstat_record(struct perf_sched *sched,
 	else
 		fprintf(stderr, "[ perf sched stats: Failed !! ]\n");
 
+	perf_session__delete(session);
 	evlist__put(evlist);
 	close(fd);

[Severity: Medium]
Does adding perf_session__delete() result in closing the file descriptor twice?

Quoted 6 lines
Quoted 4 lines

Since perf_session__delete() internally closes the file descriptor associated with the session data, the subsequent explicit close(fd) will attempt to close it again. This could return EBADF or erroneously close an unrelated file descriptor if another thread opens a file in the meantime.

Hmm.. looks right. I think I can just remove the close().

Ok, waiting for v2 then.

- Arnaldo

author maintainer
#5 lore raw

PATCH 2/3 perf sched: Fix memory leaks in perf sched stats report

In reply to Namhyung Kim

The second pass data is not saved in the list and only used to calculate delta from the first pass. Let's free the data after use.

Fixes 5a357ae6ad63 ("perf sched stats: Add support for report subcommand")
Cc Swapnil Sapkal <[email protected]>
---
 tools/perf/builtin-sched.c | 4 ++++
 1 file changed, 4 insertions(+)
tools/perf/builtin-sched.c +4 -0
index 68147a0be5c710ed..439e50d9a936f6c1 100644
@@ -4632,6 +4632,8 @@ static int perf_sched__process_schedstat(const struct perf_tool *tool __maybe_un
 			domain_second_pass = list_first_entry(&cpu_second_pass->domain_head,
 							      struct schedstat_domain, domain_list);
 			store_schedstat_cpu_diff(temp);
+			free(temp->cpu_data);
+			free(temp);
 		}
 	} else if (event->header.type == PERF_RECORD_SCHEDSTAT_DOMAIN) {
 		struct schedstat_cpu *cpu_tail;
@@ -4652,6 +4654,8 @@ static int perf_sched__process_schedstat(const struct perf_tool *tool __maybe_un
 		} else {
 			store_schedstat_domain_diff(temp);
 			domain_second_pass = list_next_entry(domain_second_pass, domain_list);
+			free(temp->domain_data);
+			free(temp);
 		}
 	}
 
Signature · 1 line
2.55.0.795.g602f6c329a-goog
author maintainer
#6 lore raw

PATCH 3/3 perf sched: Free subcommand string after perf sched stats

In reply to Namhyung Kim

The first entry of the sched_usage is dynamically allocated in parse_options_subcommand() so it should be released at the end. Do not return from a subcommand directly.

Fixes 064790a3d4a8 ("perf sched stats: Add support for diff subcommand")
Cc Swapnil Sapkal <[email protected]>
---
 tools/perf/builtin-sched.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
tools/perf/builtin-sched.c +5 -4
index 439e50d9a936f6c1..cc5b61cb9291fa02 100644
@@ -5260,19 +5260,20 @@ int cmd_sched(int argc, const char **argv)
 			if (argc)
 				argc = parse_options(argc, argv, stats_options,
 						     stats_usage, 0);
-			return perf_sched__schedstat_record(&sched, argc, argv);
+			ret = perf_sched__schedstat_record(&sched, argc, argv);
 		} else if (argv[0] && !strcmp(argv[0], "report")) {
 			if (argc)
 				argc = parse_options(argc, argv, stats_options,
 						     stats_usage, 0);
-			return perf_sched__schedstat_report(&sched);
+			ret = perf_sched__schedstat_report(&sched);
 		} else if (argv[0] && !strcmp(argv[0], "diff")) {
 			if (argc)
 				argc = parse_options(argc, argv, stats_options,
 						     stats_usage, 0);
-			return perf_sched__schedstat_diff(&sched, argc, argv);
+			ret = perf_sched__schedstat_diff(&sched, argc, argv);
+		} else {
+			ret = perf_sched__schedstat_live(&sched, argc, argv);
 		}
-		return perf_sched__schedstat_live(&sched, argc, argv);
 	} else {
 		usage_with_options(sched_usage, sched_options);
 	}
Signature · 1 line
2.55.0.795.g602f6c329a-goog