selftests/bpf: Add more tests for kptrs in maps

Firstly, ensure programs successfully load when using all of the
supported maps. Then, extend existing tests to test more cases at
runtime. We are currently testing both the synchronous freeing of items
and asynchronous destruction when map is freed, but the code needs to be
adjusted a bit to be able to also accomodate support for percpu maps.

We now do a delete on the item (and update for array maps which has a
similar effect for kptrs) to perform a synchronous free of the kptr, and
test destruction both for the synchronous and asynchronous deletion.
Next time the program runs, it should observe the refcount as 1 since
all existing references should have been released by then. By running
the program after both possible paths freeing kptrs, we establish that
they correctly release resources. Next, we augment the existing test to
also test the same code path shared by all local storage maps using a
task local storage map.

Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20230225154010.391965-4-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>

3 files changed, 451 insertions, 65 deletions

diff --git a/tools/testing/selftests/bpf/prog_tests/map_kptr.c b/tools/testing/selftests/bpf/prog_tests/map_kptr.c
index 3533a4ecad01..8743df599567 100644
--- a/tools/testing/selftests/bpf/prog_tests/map_kptr.c
+++ b/tools/testing/selftests/bpf/prog_tests/map_kptr.c
@@ -4,70 +4,160 @@
 
 #include "map_kptr.skel.h"
 #include "map_kptr_fail.skel.h"
+#include "rcu_tasks_trace_gp.skel.h"
 
 static void test_map_kptr_success(bool test_run)
 {
+	LIBBPF_OPTS(bpf_test_run_opts, lopts);
 	LIBBPF_OPTS(bpf_test_run_opts, opts,
 		.data_in = &pkt_v4,
 		.data_size_in = sizeof(pkt_v4),
 		.repeat = 1,
 	);
+	int key = 0, ret, cpu;
 	struct map_kptr *skel;
-	int key = 0, ret;
-	char buf[16];
+	char buf[16], *pbuf;
 
 	skel = map_kptr__open_and_load();
 	if (!ASSERT_OK_PTR(skel, "map_kptr__open_and_load"))
 		return;
 
-	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref), &opts);
-	ASSERT_OK(ret, "test_map_kptr_ref refcount");
-	ASSERT_OK(opts.retval, "test_map_kptr_ref retval");
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref1), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref1 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref1 retval");
 	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref2), &opts);
 	ASSERT_OK(ret, "test_map_kptr_ref2 refcount");
 	ASSERT_OK(opts.retval, "test_map_kptr_ref2 retval");
 
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_ls_map_kptr_ref1), &lopts);
+	ASSERT_OK(ret, "test_ls_map_kptr_ref1 refcount");
+	ASSERT_OK(lopts.retval, "test_ls_map_kptr_ref1 retval");
+
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_ls_map_kptr_ref2), &lopts);
+	ASSERT_OK(ret, "test_ls_map_kptr_ref2 refcount");
+	ASSERT_OK(lopts.retval, "test_ls_map_kptr_ref2 retval");
+
 	if (test_run)
 		goto exit;
 
+	cpu = libbpf_num_possible_cpus();
+	if (!ASSERT_GT(cpu, 0, "libbpf_num_possible_cpus"))
+		goto exit;
+
+	pbuf = calloc(cpu, sizeof(buf));
+	if (!ASSERT_OK_PTR(pbuf, "calloc(pbuf)"))
+		goto exit;
+
 	ret = bpf_map__update_elem(skel->maps.array_map,
 				   &key, sizeof(key), buf, sizeof(buf), 0);
 	ASSERT_OK(ret, "array_map update");
-	ret = bpf_map__update_elem(skel->maps.array_map,
-				   &key, sizeof(key), buf, sizeof(buf), 0);
-	ASSERT_OK(ret, "array_map update2");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
+
+	ret = bpf_map__update_elem(skel->maps.pcpu_array_map,
+				   &key, sizeof(key), pbuf, cpu * sizeof(buf), 0);
+	ASSERT_OK(ret, "pcpu_array_map update");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
 
-	ret = bpf_map__update_elem(skel->maps.hash_map,
-				   &key, sizeof(key), buf, sizeof(buf), 0);
-	ASSERT_OK(ret, "hash_map update");
 	ret = bpf_map__delete_elem(skel->maps.hash_map, &key, sizeof(key), 0);
 	ASSERT_OK(ret, "hash_map delete");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
+
+	ret = bpf_map__delete_elem(skel->maps.pcpu_hash_map, &key, sizeof(key), 0);
+	ASSERT_OK(ret, "pcpu_hash_map delete");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
 
-	ret = bpf_map__update_elem(skel->maps.hash_malloc_map,
-				   &key, sizeof(key), buf, sizeof(buf), 0);
-	ASSERT_OK(ret, "hash_malloc_map update");
 	ret = bpf_map__delete_elem(skel->maps.hash_malloc_map, &key, sizeof(key), 0);
 	ASSERT_OK(ret, "hash_malloc_map delete");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
+
+	ret = bpf_map__delete_elem(skel->maps.pcpu_hash_malloc_map, &key, sizeof(key), 0);
+	ASSERT_OK(ret, "pcpu_hash_malloc_map delete");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
 
-	ret = bpf_map__update_elem(skel->maps.lru_hash_map,
-				   &key, sizeof(key), buf, sizeof(buf), 0);
-	ASSERT_OK(ret, "lru_hash_map update");
 	ret = bpf_map__delete_elem(skel->maps.lru_hash_map, &key, sizeof(key), 0);
 	ASSERT_OK(ret, "lru_hash_map delete");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
+
+	ret = bpf_map__delete_elem(skel->maps.lru_pcpu_hash_map, &key, sizeof(key), 0);
+	ASSERT_OK(ret, "lru_pcpu_hash_map delete");
+	skel->data->ref--;
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_map_kptr_ref3), &opts);
+	ASSERT_OK(ret, "test_map_kptr_ref3 refcount");
+	ASSERT_OK(opts.retval, "test_map_kptr_ref3 retval");
 
+	ret = bpf_prog_test_run_opts(bpf_program__fd(skel->progs.test_ls_map_kptr_ref_del), &lopts);
+	ASSERT_OK(ret, "test_ls_map_kptr_ref_del delete");
+	skel->data->ref--;
+	ASSERT_OK(lopts.retval, "test_ls_map_kptr_ref_del retval");
+
+	free(pbuf);
 exit:
 	map_kptr__destroy(skel);
 }
 
-void test_map_kptr(void)
+static int kern_sync_rcu_tasks_trace(struct rcu_tasks_trace_gp *rcu)
 {
-	if (test__start_subtest("success")) {
+	long gp_seq = READ_ONCE(rcu->bss->gp_seq);
+	LIBBPF_OPTS(bpf_test_run_opts, opts);
+
+	if (!ASSERT_OK(bpf_prog_test_run_opts(bpf_program__fd(rcu->progs.do_call_rcu_tasks_trace),
+					      &opts), "do_call_rcu_tasks_trace"))
+		return -EFAULT;
+	if (!ASSERT_OK(opts.retval, "opts.retval == 0"))
+		return -EFAULT;
+	while (gp_seq == READ_ONCE(rcu->bss->gp_seq))
+		sched_yield();
+	return 0;
+}
+
+void serial_test_map_kptr(void)
+{
+	struct rcu_tasks_trace_gp *skel;
+
+	RUN_TESTS(map_kptr_fail);
+
+	skel = rcu_tasks_trace_gp__open_and_load();
+	if (!ASSERT_OK_PTR(skel, "rcu_tasks_trace_gp__open_and_load"))
+		return;
+	if (!ASSERT_OK(rcu_tasks_trace_gp__attach(skel), "rcu_tasks_trace_gp__attach"))
+		goto end;
+
+	if (test__start_subtest("success-map")) {
+		test_map_kptr_success(true);
+
+		ASSERT_OK(kern_sync_rcu_tasks_trace(skel), "sync rcu_tasks_trace");
+		ASSERT_OK(kern_sync_rcu(), "sync rcu");
+		/* Observe refcount dropping to 1 on bpf_map_free_deferred */
 		test_map_kptr_success(false);
-		/* Do test_run twice, so that we see refcount going back to 1
-		 * after we leave it in map from first iteration.
-		 */
+
+		ASSERT_OK(kern_sync_rcu_tasks_trace(skel), "sync rcu_tasks_trace");
+		ASSERT_OK(kern_sync_rcu(), "sync rcu");
+		/* Observe refcount dropping to 1 on synchronous delete elem */
 		test_map_kptr_success(true);
 	}
 
-	RUN_TESTS(map_kptr_fail);
+end:
+	rcu_tasks_trace_gp__destroy(skel);
+	return;
 }
diff --git a/tools/testing/selftests/bpf/progs/map_kptr.c b/tools/testing/selftests/bpf/progs/map_kptr.c
index 228ec45365a8..a24d17bc17eb 100644
--- a/tools/testing/selftests/bpf/progs/map_kptr.c
+++ b/tools/testing/selftests/bpf/progs/map_kptr.c
@@ -15,6 +15,13 @@ struct array_map {
 	__uint(max_entries, 1);
 } array_map SEC(".maps");
 
+struct pcpu_array_map {
+	__uint(type, BPF_MAP_TYPE_PERCPU_ARRAY);
+	__type(key, int);
+	__type(value, struct map_value);
+	__uint(max_entries, 1);
+} pcpu_array_map SEC(".maps");
+
 struct hash_map {
 	__uint(type, BPF_MAP_TYPE_HASH);
 	__type(key, int);
@@ -22,6 +29,13 @@ struct hash_map {
 	__uint(max_entries, 1);
 } hash_map SEC(".maps");
 
+struct pcpu_hash_map {
+	__uint(type, BPF_MAP_TYPE_PERCPU_HASH);
+	__type(key, int);
+	__type(value, struct map_value);
+	__uint(max_entries, 1);
+} pcpu_hash_map SEC(".maps");
+
 struct hash_malloc_map {
 	__uint(type, BPF_MAP_TYPE_HASH);
 	__type(key, int);
@@ -30,6 +44,14 @@ struct hash_malloc_map {
 	__uint(map_flags, BPF_F_NO_PREALLOC);
 } hash_malloc_map SEC(".maps");
 
+struct pcpu_hash_malloc_map {
+	__uint(type, BPF_MAP_TYPE_PERCPU_HASH);
+	__type(key, int);
+	__type(value, struct map_value);
+	__uint(max_entries, 1);
+	__uint(map_flags, BPF_F_NO_PREALLOC);
+} pcpu_hash_malloc_map SEC(".maps");
+
 struct lru_hash_map {
 	__uint(type, BPF_MAP_TYPE_LRU_HASH);
 	__type(key, int);
@@ -37,6 +59,41 @@ struct lru_hash_map {
 	__uint(max_entries, 1);
 } lru_hash_map SEC(".maps");
 
+struct lru_pcpu_hash_map {
+	__uint(type, BPF_MAP_TYPE_LRU_PERCPU_HASH);
+	__type(key, int);
+	__type(value, struct map_value);
+	__uint(max_entries, 1);
+} lru_pcpu_hash_map SEC(".maps");
+
+struct cgrp_ls_map {
+	__uint(type, BPF_MAP_TYPE_CGRP_STORAGE);
+	__uint(map_flags, BPF_F_NO_PREALLOC);
+	__type(key, int);
+	__type(value, struct map_value);
+} cgrp_ls_map SEC(".maps");
+
+struct task_ls_map {
+	__uint(type, BPF_MAP_TYPE_TASK_STORAGE);
+	__uint(map_flags, BPF_F_NO_PREALLOC);
+	__type(key, int);
+	__type(value, struct map_value);
+} task_ls_map SEC(".maps");
+
+struct inode_ls_map {
+	__uint(type, BPF_MAP_TYPE_INODE_STORAGE);
+	__uint(map_flags, BPF_F_NO_PREALLOC);
+	__type(key, int);
+	__type(value, struct map_value);
+} inode_ls_map SEC(".maps");
+
+struct sk_ls_map {
+	__uint(type, BPF_MAP_TYPE_SK_STORAGE);
+	__uint(map_flags, BPF_F_NO_PREALLOC);
+	__type(key, int);
+	__type(value, struct map_value);
+} sk_ls_map SEC(".maps");
+
 #define DEFINE_MAP_OF_MAP(map_type, inner_map_type, name)       \
 	struct {                                                \
 		__uint(type, map_type);                         \
@@ -160,6 +217,58 @@ int test_map_kptr(struct __sk_buff *ctx)
 	return 0;
 }
 
+SEC("tp_btf/cgroup_mkdir")
+int BPF_PROG(test_cgrp_map_kptr, struct cgroup *cgrp, const char *path)
+{
+	struct map_value *v;
+
+	v = bpf_cgrp_storage_get(&cgrp_ls_map, cgrp, NULL, BPF_LOCAL_STORAGE_GET_F_CREATE);
+	if (v)
+		test_kptr(v);
+	return 0;
+}
+
+SEC("lsm/inode_unlink")
+int BPF_PROG(test_task_map_kptr, struct inode *inode, struct dentry *victim)
+{
+	struct task_struct *task;
+	struct map_value *v;
+
+	task = bpf_get_current_task_btf();
+	if (!task)
+		return 0;
+	v = bpf_task_storage_get(&task_ls_map, task, NULL, BPF_LOCAL_STORAGE_GET_F_CREATE);
+	if (v)
+		test_kptr(v);
+	return 0;
+}
+
+SEC("lsm/inode_unlink")
+int BPF_PROG(test_inode_map_kptr, struct inode *inode, struct dentry *victim)
+{
+	struct map_value *v;
+
+	v = bpf_inode_storage_get(&inode_ls_map, inode, NULL, BPF_LOCAL_STORAGE_GET_F_CREATE);
+	if (v)
+		test_kptr(v);
+	return 0;
+}
+
+SEC("tc")
+int test_sk_map_kptr(struct __sk_buff *ctx)
+{
+	struct map_value *v;
+	struct bpf_sock *sk;
+
+	sk = ctx->sk;
+	if (!sk)
+		return 0;
+	v = bpf_sk_storage_get(&sk_ls_map, sk, NULL, BPF_LOCAL_STORAGE_GET_F_CREATE);
+	if (v)
+		test_kptr(v);
+	return 0;
+}
+
 SEC("tc")
 int test_map_in_map_kptr(struct __sk_buff *ctx)
 {
@@ -189,106 +298,257 @@ int test_map_in_map_kptr(struct __sk_buff *ctx)
 	return 0;
 }
 
-SEC("tc")
-int test_map_kptr_ref(struct __sk_buff *ctx)
+int ref = 1;
+
+static __always_inline
+int test_map_kptr_ref_pre(struct map_value *v)
 {
 	struct prog_test_ref_kfunc *p, *p_st;
 	unsigned long arg = 0;
-	struct map_value *v;
-	int key = 0, ret;
+	int ret;
 
 	p = bpf_kfunc_call_test_acquire(&arg);
 	if (!p)
 		return 1;
+	ref++;
 
 	p_st = p->next;
-	if (p_st->cnt.refs.counter != 2) {
+	if (p_st->cnt.refs.counter != ref) {
 		ret = 2;
 		goto end;
 	}
 
-	v = bpf_map_lookup_elem(&array_map, &key);
-	if (!v) {
-		ret = 3;
-		goto end;
-	}
-
 	p = bpf_kptr_xchg(&v->ref_ptr, p);
 	if (p) {
-		ret = 4;
+		ret = 3;
 		goto end;
 	}
-	if (p_st->cnt.refs.counter != 2)
-		return 5;
+	if (p_st->cnt.refs.counter != ref)
+		return 4;
 
 	p = bpf_kfunc_call_test_kptr_get(&v->ref_ptr, 0, 0);
 	if (!p)
-		return 6;
-	if (p_st->cnt.refs.counter != 3) {
-		ret = 7;
+		return 5;
+	ref++;
+	if (p_st->cnt.refs.counter != ref) {
+		ret = 6;
 		goto end;
 	}
 	bpf_kfunc_call_test_release(p);
-	if (p_st->cnt.refs.counter != 2)
-		return 8;
+	ref--;
+	if (p_st->cnt.refs.counter != ref)
+		return 7;
 
 	p = bpf_kptr_xchg(&v->ref_ptr, NULL);
 	if (!p)
-		return 9;
+		return 8;
 	bpf_kfunc_call_test_release(p);
-	if (p_st->cnt.refs.counter != 1)
-		return 10;
+	ref--;
+	if (p_st->cnt.refs.counter != ref)
+		return 9;
 
 	p = bpf_kfunc_call_test_acquire(&arg);
 	if (!p)
-		return 11;
+		return 10;
+	ref++;
 	p = bpf_kptr_xchg(&v->ref_ptr, p);
 	if (p) {
-		ret = 12;
+		ret = 11;
 		goto end;
 	}
-	if (p_st->cnt.refs.counter != 2)
-		return 13;
+	if (p_st->cnt.refs.counter != ref)
+		return 12;
 	/* Leave in map */
 
 	return 0;
 end:
+	ref--;
 	bpf_kfunc_call_test_release(p);
 	return ret;
 }
 
-SEC("tc")
-int test_map_kptr_ref2(struct __sk_buff *ctx)
+static __always_inline
+int test_map_kptr_ref_post(struct map_value *v)
 {
 	struct prog_test_ref_kfunc *p, *p_st;
-	struct map_value *v;
-	int key = 0;
-
-	v = bpf_map_lookup_elem(&array_map, &key);
-	if (!v)
-		return 1;
 
 	p_st = v->ref_ptr;
-	if (!p_st || p_st->cnt.refs.counter != 2)
-		return 2;
+	if (!p_st || p_st->cnt.refs.counter != ref)
+		return 1;
 
 	p = bpf_kptr_xchg(&v->ref_ptr, NULL);
 	if (!p)
-		return 3;
-	if (p_st->cnt.refs.counter != 2) {
+		return 2;
+	if (p_st->cnt.refs.counter != ref) {
 		bpf_kfunc_call_test_release(p);
-		return 4;
+		return 3;
 	}
 
 	p = bpf_kptr_xchg(&v->ref_ptr, p);
 	if (p) {
 		bpf_kfunc_call_test_release(p);
-		return 5;
+		return 4;
 	}
-	if (p_st->cnt.refs.counter != 2)
-		return 6;
+	if (p_st->cnt.refs.counter != ref)
+		return 5;
+
+	return 0;
+}
+
+#define TEST(map)                            \
+	v = bpf_map_lookup_elem(&map, &key); \
+	if (!v)                              \
+		return -1;                   \
+	ret = test_map_kptr_ref_pre(v);      \
+	if (ret)                             \
+		return ret;
+
+#define TEST_PCPU(map)                                 \
+	v = bpf_map_lookup_percpu_elem(&map, &key, 0); \
+	if (!v)                                        \
+		return -1;                             \
+	ret = test_map_kptr_ref_pre(v);                \
+	if (ret)                                       \
+		return ret;
+
+SEC("tc")
+int test_map_kptr_ref1(struct __sk_buff *ctx)
+{
+	struct map_value *v, val = {};
+	int key = 0, ret;
+
+	bpf_map_update_elem(&hash_map, &key, &val, 0);
+	bpf_map_update_elem(&hash_malloc_map, &key, &val, 0);
+	bpf_map_update_elem(&lru_hash_map, &key, &val, 0);
+
+	bpf_map_update_elem(&pcpu_hash_map, &key, &val, 0);
+	bpf_map_update_elem(&pcpu_hash_malloc_map, &key, &val, 0);
+	bpf_map_update_elem(&lru_pcpu_hash_map, &key, &val, 0);
+
+	TEST(array_map);
+	TEST(hash_map);
+	TEST(hash_malloc_map);
+	TEST(lru_hash_map);
+
+	TEST_PCPU(pcpu_array_map);
+	TEST_PCPU(pcpu_hash_map);
+	TEST_PCPU(pcpu_hash_malloc_map);
+	TEST_PCPU(lru_pcpu_hash_map);
+
+	return 0;
+}
+
+#undef TEST
+#undef TEST_PCPU
+
+#define TEST(map)                            \
+	v = bpf_map_lookup_elem(&map, &key); \
+	if (!v)                              \
+		return -1;                   \
+	ret = test_map_kptr_ref_post(v);     \
+	if (ret)                             \
+		return ret;
+
+#define TEST_PCPU(map)                                 \
+	v = bpf_map_lookup_percpu_elem(&map, &key, 0); \
+	if (!v)                                        \
+		return -1;                             \
+	ret = test_map_kptr_ref_post(v);               \
+	if (ret)                                       \
+		return ret;
+
+SEC("tc")
+int test_map_kptr_ref2(struct __sk_buff *ctx)
+{
+	struct map_value *v;
+	int key = 0, ret;
+
+	TEST(array_map);
+	TEST(hash_map);
+	TEST(hash_malloc_map);
+	TEST(lru_hash_map);
+
+	TEST_PCPU(pcpu_array_map);
+	TEST_PCPU(pcpu_hash_map);
+	TEST_PCPU(pcpu_hash_malloc_map);
+	TEST_PCPU(lru_pcpu_hash_map);
 
 	return 0;
 }
 
+#undef TEST
+#undef TEST_PCPU
+
+SEC("tc")
+int test_map_kptr_ref3(struct __sk_buff *ctx)
+{
+	struct prog_test_ref_kfunc *p;
+	unsigned long sp = 0;
+
+	p = bpf_kfunc_call_test_acquire(&sp);
+	if (!p)
+		return 1;
+	ref++;
+	if (p->cnt.refs.counter != ref) {
+		bpf_kfunc_call_test_release(p);
+		return 2;
+	}
+	bpf_kfunc_call_test_release(p);
+	ref--;
+	return 0;
+}
+
+SEC("syscall")
+int test_ls_map_kptr_ref1(void *ctx)
+{
+	struct task_struct *current;
+	struct map_value *v;
+	int ret;
+
+	current = bpf_get_current_task_btf();
+	if (!current)
+		return 100;
+	v = bpf_task_storage_get(&task_ls_map, current, NULL, 0);
+	if (v)
+		return 150;
+	v = bpf_task_storage_get(&task_ls_map, current, NULL, BPF_LOCAL_STORAGE_GET_F_CREATE);
+	if (!v)
+		return 200;
+	return test_map_kptr_ref_pre(v);
+}
+
+SEC("syscall")
+int test_ls_map_kptr_ref2(void *ctx)
+{
+	struct task_struct *current;
+	struct map_value *v;
+	int ret;
+
+	current = bpf_get_current_task_btf();
+	if (!current)
+		return 100;
+	v = bpf_task_storage_get(&task_ls_map, current, NULL, 0);
+	if (!v)
+		return 200;
+	return test_map_kptr_ref_post(v);
+}
+
+SEC("syscall")
+int test_ls_map_kptr_ref_del(void *ctx)
+{
+	struct task_struct *current;
+	struct map_value *v;
+	int ret;
+
+	current = bpf_get_current_task_btf();
+	if (!current)
+		return 100;
+	v = bpf_task_storage_get(&task_ls_map, current, NULL, 0);
+	if (!v)
+		return 200;
+	if (!v->ref_ptr)
+		return 300;
+	return bpf_task_storage_delete(&task_ls_map, current);
+}
+
 char _license[] SEC("license") = "GPL";
diff --git a/tools/testing/selftests/bpf/progs/rcu_tasks_trace_gp.c b/tools/testing/selftests/bpf/progs/rcu_tasks_trace_gp.c
new file mode 100644
index 000000000000..df4873558634
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/rcu_tasks_trace_gp.c
@@ -0,0 +1,36 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <vmlinux.h>
+#include <bpf/bpf_tracing.h>
+#include <bpf/bpf_helpers.h>
+
+struct task_ls_map {
+	__uint(type, BPF_MAP_TYPE_TASK_STORAGE);
+	__uint(map_flags, BPF_F_NO_PREALLOC);
+	__type(key, int);
+	__type(value, int);
+} task_ls_map SEC(".maps");
+
+long gp_seq;
+
+SEC("syscall")
+int do_call_rcu_tasks_trace(void *ctx)
+{
+    struct task_struct *current;
+    int *v;
+
+    current = bpf_get_current_task_btf();
+    v = bpf_task_storage_get(&task_ls_map, current, NULL, BPF_LOCAL_STORAGE_GET_F_CREATE);
+    if (!v)
+        return 1;
+    /* Invoke call_rcu_tasks_trace */
+    return bpf_task_storage_delete(&task_ls_map, current);
+}
+
+SEC("kprobe/rcu_tasks_trace_postgp")
+int rcu_tasks_trace_postgp(void *ctx)
+{
+    __sync_add_and_fetch(&gp_seq, 1);
+    return 0;
+}
+
+char _license[] SEC("license") = "GPL";