selftests/cgroup: Don't require synchronous populated update on task exit

[ Upstream commit 6680c162b4850976ee52b57372eddc4450c1d074 ]

test_cgcore_populated (test_core) and test_cgkill_{simple,tree,forkbomb}
(test_kill) check cgroup.events "populated 0" immediately after reaping
child tasks with waitpid(). This used to work because cgroup_task_exit() in
do_exit() unlinked tasks from css_sets before exit_notify() woke up
waitpid().

d245698d727a ("cgroup: Defer task cgroup unlink until after the task is done
switching out") moved the unlink to cgroup_task_dead() in
finish_task_switch(), which runs after exit_notify(). The populated counter
is now decremented after the parent's waitpid() can return, so there is no
longer a synchronous ordering guarantee. On PREEMPT_RT, where
cgroup_task_dead() is further deferred through lazy irq_work, the race
window is even larger.

The synchronous populated transition was never part of the cgroup interface
contract - it was an implementation artifact. Use cg_read_strcmp_wait() which
retries for up to 1 second, matching what these tests actually need to
verify: that the cgroup eventually becomes unpopulated after all tasks exit.

Fixes: d245698d727a ("cgroup: Defer task cgroup unlink until after the task is done switching out")
Reported-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Tejun Heo <tj@kernel.org>
Tested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Christian Brauner <brauner@kernel.org>
Cc: cgroups@vger.kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

4 files changed, 23 insertions, 4 deletions

diff --git a/tools/testing/selftests/cgroup/lib/cgroup_util.c b/tools/testing/selftests/cgroup/lib/cgroup_util.c
index 44c52f620fda..4b0f2c46d432 100644
--- a/tools/testing/selftests/cgroup/lib/cgroup_util.c
+++ b/tools/testing/selftests/cgroup/lib/cgroup_util.c
@@ -123,6 +123,21 @@ int cg_read_strcmp(const char *cgroup, const char *control,
 	return ret;
 }
 
+int cg_read_strcmp_wait(const char *cgroup, const char *control,
+			    const char *expected)
+{
+	int i, ret;
+
+	for (i = 0; i < 100; i++) {
+		ret = cg_read_strcmp(cgroup, control, expected);
+		if (!ret)
+			return ret;
+		usleep(10000);
+	}
+
+	return ret;
+}
+
 int cg_read_strstr(const char *cgroup, const char *control, const char *needle)
 {
 	char buf[PAGE_SIZE];
diff --git a/tools/testing/selftests/cgroup/lib/include/cgroup_util.h b/tools/testing/selftests/cgroup/lib/include/cgroup_util.h
index 7ab2824ed7b5..1cbe3b0ac6f7 100644
--- a/tools/testing/selftests/cgroup/lib/include/cgroup_util.h
+++ b/tools/testing/selftests/cgroup/lib/include/cgroup_util.h
@@ -59,6 +59,8 @@ extern int cg_read(const char *cgroup, const char *control,
 		   char *buf, size_t len);
 extern int cg_read_strcmp(const char *cgroup, const char *control,
 			  const char *expected);
+extern int cg_read_strcmp_wait(const char *cgroup, const char *control,
+				   const char *expected);
 extern int cg_read_strstr(const char *cgroup, const char *control,
 			  const char *needle);
 extern long cg_read_long(const char *cgroup, const char *control);
diff --git a/tools/testing/selftests/cgroup/test_core.c b/tools/testing/selftests/cgroup/test_core.c
index 102262555a59..7b83c7e7c9d4 100644
--- a/tools/testing/selftests/cgroup/test_core.c
+++ b/tools/testing/selftests/cgroup/test_core.c
@@ -233,7 +233,8 @@ static int test_cgcore_populated(const char *root)
 	if (err)
 		goto cleanup;
 
-	if (cg_read_strcmp(cg_test_d, "cgroup.events", "populated 0\n"))
+	if (cg_read_strcmp_wait(cg_test_d, "cgroup.events",
+				   "populated 0\n"))
 		goto cleanup;
 
 	/* Remove cgroup. */
diff --git a/tools/testing/selftests/cgroup/test_kill.c b/tools/testing/selftests/cgroup/test_kill.c
index c8c9d306925b..f6cd23a8ecc7 100644
--- a/tools/testing/selftests/cgroup/test_kill.c
+++ b/tools/testing/selftests/cgroup/test_kill.c
@@ -86,7 +86,7 @@ cleanup:
 		wait_for_pid(pids[i]);
 
 	if (ret == KSFT_PASS &&
-	    cg_read_strcmp(cgroup, "cgroup.events", "populated 0\n"))
+	    cg_read_strcmp_wait(cgroup, "cgroup.events", "populated 0\n"))
 		ret = KSFT_FAIL;
 
 	if (cgroup)
@@ -190,7 +190,8 @@ cleanup:
 		wait_for_pid(pids[i]);
 
 	if (ret == KSFT_PASS &&
-	    cg_read_strcmp(cgroup[0], "cgroup.events", "populated 0\n"))
+	    cg_read_strcmp_wait(cgroup[0], "cgroup.events",
+				   "populated 0\n"))
 		ret = KSFT_FAIL;
 
 	for (i = 9; i >= 0 && cgroup[i]; i--) {
@@ -251,7 +252,7 @@ cleanup:
 		wait_for_pid(pid);
 
 	if (ret == KSFT_PASS &&
-	    cg_read_strcmp(cgroup, "cgroup.events", "populated 0\n"))
+	    cg_read_strcmp_wait(cgroup, "cgroup.events", "populated 0\n"))
 		ret = KSFT_FAIL;
 
 	if (cgroup)