Skip to contentSkip to conversation

selftests/bpf: wq: fix skel leak in serial_test_wq()

PATCH In review Kery Qi 3 participants 3 messages Jan 11 – Jan 13

Subsystems (3)
lore
Share & tools
lore

Kery Qi

author
#0 lore raw

PATCH 1/1 selftests/bpf: wq: fix skel leak in serial_test_wq()

serial_test_wq() returns early when ASSERT_OK_PTR(wq_skel, "wq_skel_load") fails. In that case wq__open_and_load() may still have returned a non-NULL skeleton, and the early return skips wq__destroy(), leaking resources and triggering ASAN leak reports in selftests runs.

Jump to the common clean_up label instead, so wq__destroy() is executed on all exit paths. Also fix the missing semicolon after 'goto clean_up'.

Fixes 8290dba51910 ("selftests/bpf: wq: add bpf_wq_start() checks")
---
 tools/testing/selftests/bpf/prog_tests/wq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/wq.c +1 -1
index 1dcdeda84853..328c393e7167 100644
@@ -17,7 +17,7 @@ void serial_test_wq(void)
 
 	wq_skel = wq__open_and_load();
 	if (!ASSERT_OK_PTR(wq_skel, "wq_skel_load"))
-		return;
+		goto clean_up;
 
 	err = wq__attach(wq_skel);
 	if (!ASSERT_OK(err, "wq_attach"))
Signature · 1 line
2.34.1
#1 lore raw
Is there a clean_up label defined in serial_test_wq()? Looking at the function, I only see wq__destroy(wq_skel) at the end without any label. The only label in the file is out: in…

tools/testing/selftests/bpf/prog_tests/wq.c void serial_test_wq(void) · line 20

Quoted 10 lines
tools/testing/selftests/bpf/prog_tests/wq.c +1 -1
index 15c67d23128b..d000d9dbdf46 100644
@@ -17,7 +17,7 @@ void serial_test_wq(void)
 
 	wq_skel = wq__open_and_load();
 	if (!ASSERT_OK_PTR(wq_skel, "wq_skel_load"))
-		return;
+		goto clean_up;

Is there a clean_up label defined in serial_test_wq()? Looking at the function, I only see wq__destroy(wq_skel) at the end without any label. The only label in the file is out: in test_failure_map_no_btf().

This appears to cause a build failure since the goto target does not exist.

---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md

CI run summary: https://github.com/kernel-patches/bpf/actions/runs/20899649668

#2 lore raw
In reply to Kery Qi
Quoted 19 lines · Kery Qi

serial_test_wq() returns early when ASSERT_OK_PTR(wq_skel, "wq_skel_load") fails. In that case wq__open_and_load() may still have returned a non-NULL skeleton, and the early return skips wq__destroy(), leaking resources and triggering ASAN leak reports in selftests runs.

Jump to the common clean_up label instead, so wq__destroy() is executed on all exit paths. Also fix the missing semicolon after 'goto clean_up'.

Fixes 8290dba51910 ("selftests/bpf: wq: add bpf_wq_start() checks")
---
  tools/testing/selftests/bpf/prog_tests/wq.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
tools/testing/selftests/bpf/prog_tests/wq.c +0 -0
index 1dcdeda84853..328c393e7167 100644

tools/testing/selftests/bpf/prog_tests/wq.c void serial_test_wq(void) · line 23

Quoted 8 lines
tools/testing/selftests/bpf/prog_tests/wq.c +1 -1
@@ -17,7 +17,7 @@ void serial_test_wq(void)
 
  	wq_skel = wq__open_and_load();
  	if (!ASSERT_OK_PTR(wq_skel, "wq_skel_load"))
-		return;
+		goto clean_up;
 
  	err = wq__attach(wq_skel);
  	if (!ASSERT_OK(err, "wq_attach"))

I think the code is correct as is, open_and_load() either returns NULL or valid loaded skeleton.
There is nothing to cleanup if NULL is returned. You can check generated wq.skel.h to see that on load error path wq__destory() is called.