Merge branch 'optimize-bpf_tail_call'

Daniel Borkmann says:

====================
This gets rid of indirect jumps for BPF tail calls whenever possible.
The series adds emission for *direct* jumps for tail call maps in order
to avoid the retpoline overhead from a493a87f38cf ("bpf, x64: implement
retpoline for tail call") for situations that allow for it, meaning,
for known constant keys at verification time which are used as index
into the tail call map. See patch 7/8 for more general details.

Thanks!

v1  -> v2:
  - added more test cases
  - u8 ip_stable -> bool (Andrii)
  - removed bpf_map_poke_{un,}lock and simplified the code (Andrii)
  - added break into prog_array_map_poke_untrack since there's just
    one prog (Andrii)
  - fixed typo: for for in commit msg (Andrii)
  - reworked __bpf_arch_text_poke (Andrii)
  - added subtests, and comment on tests themselves, NULL-NULL
    transistion (Andrii)
  - in constant map key tracking I've moved the map_poke_track callback
    to once we've finished creating the poke tab as otherwise concurrent
    access from tail call map would blow up (since we realloc the table)
rfc -> v1:
  - Applied Alexei's and Andrii's feeback from
    https://lore.kernel.org/bpf/cover.1573779287.git.daniel@iogearbox.net/T/#t
====================

Signed-off-by: Alexei Starovoitov <ast@kernel.org>

6 files changed, 698 insertions, 0 deletions

diff --git a/tools/testing/selftests/bpf/prog_tests/tailcalls.c b/tools/testing/selftests/bpf/prog_tests/tailcalls.c
new file mode 100644
index 000000000000..bb8fe646dd9f
--- /dev/null
+++ b/tools/testing/selftests/bpf/prog_tests/tailcalls.c
@@ -0,0 +1,487 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <test_progs.h>
+
+/* test_tailcall_1 checks basic functionality by patching multiple locations
+ * in a single program for a single tail call slot with nop->jmp, jmp->nop
+ * and jmp->jmp rewrites. Also checks for nop->nop.
+ */
+static void test_tailcall_1(void)
+{
+	int err, map_fd, prog_fd, main_fd, i, j;
+	struct bpf_map *prog_array;
+	struct bpf_program *prog;
+	struct bpf_object *obj;
+	__u32 retval, duration;
+	char prog_name[32];
+	char buff[128] = {};
+
+	err = bpf_prog_load("tailcall1.o", BPF_PROG_TYPE_SCHED_CLS, &obj,
+			    &prog_fd);
+	if (CHECK_FAIL(err))
+		return;
+
+	prog = bpf_object__find_program_by_title(obj, "classifier");
+	if (CHECK_FAIL(!prog))
+		goto out;
+
+	main_fd = bpf_program__fd(prog);
+	if (CHECK_FAIL(main_fd < 0))
+		goto out;
+
+	prog_array = bpf_object__find_map_by_name(obj, "jmp_table");
+	if (CHECK_FAIL(!prog_array))
+		goto out;
+
+	map_fd = bpf_map__fd(prog_array);
+	if (CHECK_FAIL(map_fd < 0))
+		goto out;
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		snprintf(prog_name, sizeof(prog_name), "classifier/%i", i);
+
+		prog = bpf_object__find_program_by_title(obj, prog_name);
+		if (CHECK_FAIL(!prog))
+			goto out;
+
+		prog_fd = bpf_program__fd(prog);
+		if (CHECK_FAIL(prog_fd < 0))
+			goto out;
+
+		err = bpf_map_update_elem(map_fd, &i, &prog_fd, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+					&duration, &retval, NULL);
+		CHECK(err || retval != i, "tailcall",
+		      "err %d errno %d retval %d\n", err, errno, retval);
+
+		err = bpf_map_delete_elem(map_fd, &i);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 3, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		snprintf(prog_name, sizeof(prog_name), "classifier/%i", i);
+
+		prog = bpf_object__find_program_by_title(obj, prog_name);
+		if (CHECK_FAIL(!prog))
+			goto out;
+
+		prog_fd = bpf_program__fd(prog);
+		if (CHECK_FAIL(prog_fd < 0))
+			goto out;
+
+		err = bpf_map_update_elem(map_fd, &i, &prog_fd, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 0, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		j = bpf_map__def(prog_array)->max_entries - 1 - i;
+		snprintf(prog_name, sizeof(prog_name), "classifier/%i", j);
+
+		prog = bpf_object__find_program_by_title(obj, prog_name);
+		if (CHECK_FAIL(!prog))
+			goto out;
+
+		prog_fd = bpf_program__fd(prog);
+		if (CHECK_FAIL(prog_fd < 0))
+			goto out;
+
+		err = bpf_map_update_elem(map_fd, &i, &prog_fd, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		j = bpf_map__def(prog_array)->max_entries - 1 - i;
+
+		err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+					&duration, &retval, NULL);
+		CHECK(err || retval != j, "tailcall",
+		      "err %d errno %d retval %d\n", err, errno, retval);
+
+		err = bpf_map_delete_elem(map_fd, &i);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 3, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		err = bpf_map_delete_elem(map_fd, &i);
+		if (CHECK_FAIL(err >= 0 || errno != ENOENT))
+			goto out;
+
+		err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+					&duration, &retval, NULL);
+		CHECK(err || retval != 3, "tailcall",
+		      "err %d errno %d retval %d\n", err, errno, retval);
+	}
+
+out:
+	bpf_object__close(obj);
+}
+
+/* test_tailcall_2 checks that patching multiple programs for a single
+ * tail call slot works. It also jumps through several programs and tests
+ * the tail call limit counter.
+ */
+static void test_tailcall_2(void)
+{
+	int err, map_fd, prog_fd, main_fd, i;
+	struct bpf_map *prog_array;
+	struct bpf_program *prog;
+	struct bpf_object *obj;
+	__u32 retval, duration;
+	char prog_name[32];
+	char buff[128] = {};
+
+	err = bpf_prog_load("tailcall2.o", BPF_PROG_TYPE_SCHED_CLS, &obj,
+			    &prog_fd);
+	if (CHECK_FAIL(err))
+		return;
+
+	prog = bpf_object__find_program_by_title(obj, "classifier");
+	if (CHECK_FAIL(!prog))
+		goto out;
+
+	main_fd = bpf_program__fd(prog);
+	if (CHECK_FAIL(main_fd < 0))
+		goto out;
+
+	prog_array = bpf_object__find_map_by_name(obj, "jmp_table");
+	if (CHECK_FAIL(!prog_array))
+		goto out;
+
+	map_fd = bpf_map__fd(prog_array);
+	if (CHECK_FAIL(map_fd < 0))
+		goto out;
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		snprintf(prog_name, sizeof(prog_name), "classifier/%i", i);
+
+		prog = bpf_object__find_program_by_title(obj, prog_name);
+		if (CHECK_FAIL(!prog))
+			goto out;
+
+		prog_fd = bpf_program__fd(prog);
+		if (CHECK_FAIL(prog_fd < 0))
+			goto out;
+
+		err = bpf_map_update_elem(map_fd, &i, &prog_fd, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 2, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+
+	i = 2;
+	err = bpf_map_delete_elem(map_fd, &i);
+	if (CHECK_FAIL(err))
+		goto out;
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 1, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+
+	i = 0;
+	err = bpf_map_delete_elem(map_fd, &i);
+	if (CHECK_FAIL(err))
+		goto out;
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 3, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+out:
+	bpf_object__close(obj);
+}
+
+/* test_tailcall_3 checks that the count value of the tail call limit
+ * enforcement matches with expectations.
+ */
+static void test_tailcall_3(void)
+{
+	int err, map_fd, prog_fd, main_fd, data_fd, i, val;
+	struct bpf_map *prog_array, *data_map;
+	struct bpf_program *prog;
+	struct bpf_object *obj;
+	__u32 retval, duration;
+	char buff[128] = {};
+
+	err = bpf_prog_load("tailcall3.o", BPF_PROG_TYPE_SCHED_CLS, &obj,
+			    &prog_fd);
+	if (CHECK_FAIL(err))
+		return;
+
+	prog = bpf_object__find_program_by_title(obj, "classifier");
+	if (CHECK_FAIL(!prog))
+		goto out;
+
+	main_fd = bpf_program__fd(prog);
+	if (CHECK_FAIL(main_fd < 0))
+		goto out;
+
+	prog_array = bpf_object__find_map_by_name(obj, "jmp_table");
+	if (CHECK_FAIL(!prog_array))
+		goto out;
+
+	map_fd = bpf_map__fd(prog_array);
+	if (CHECK_FAIL(map_fd < 0))
+		goto out;
+
+	prog = bpf_object__find_program_by_title(obj, "classifier/0");
+	if (CHECK_FAIL(!prog))
+		goto out;
+
+	prog_fd = bpf_program__fd(prog);
+	if (CHECK_FAIL(prog_fd < 0))
+		goto out;
+
+	i = 0;
+	err = bpf_map_update_elem(map_fd, &i, &prog_fd, BPF_ANY);
+	if (CHECK_FAIL(err))
+		goto out;
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 1, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+
+	data_map = bpf_object__find_map_by_name(obj, "tailcall.bss");
+	if (CHECK_FAIL(!data_map || !bpf_map__is_internal(data_map)))
+		return;
+
+	data_fd = bpf_map__fd(data_map);
+	if (CHECK_FAIL(map_fd < 0))
+		return;
+
+	i = 0;
+	err = bpf_map_lookup_elem(data_fd, &i, &val);
+	CHECK(err || val != 33, "tailcall count", "err %d errno %d count %d\n",
+	      err, errno, val);
+
+	i = 0;
+	err = bpf_map_delete_elem(map_fd, &i);
+	if (CHECK_FAIL(err))
+		goto out;
+
+	err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+				&duration, &retval, NULL);
+	CHECK(err || retval != 0, "tailcall", "err %d errno %d retval %d\n",
+	      err, errno, retval);
+out:
+	bpf_object__close(obj);
+}
+
+/* test_tailcall_4 checks that the kernel properly selects indirect jump
+ * for the case where the key is not known. Latter is passed via global
+ * data to select different targets we can compare return value of.
+ */
+static void test_tailcall_4(void)
+{
+	int err, map_fd, prog_fd, main_fd, data_fd, i;
+	struct bpf_map *prog_array, *data_map;
+	struct bpf_program *prog;
+	struct bpf_object *obj;
+	__u32 retval, duration;
+	static const int zero = 0;
+	char buff[128] = {};
+	char prog_name[32];
+
+	err = bpf_prog_load("tailcall4.o", BPF_PROG_TYPE_SCHED_CLS, &obj,
+			    &prog_fd);
+	if (CHECK_FAIL(err))
+		return;
+
+	prog = bpf_object__find_program_by_title(obj, "classifier");
+	if (CHECK_FAIL(!prog))
+		goto out;
+
+	main_fd = bpf_program__fd(prog);
+	if (CHECK_FAIL(main_fd < 0))
+		goto out;
+
+	prog_array = bpf_object__find_map_by_name(obj, "jmp_table");
+	if (CHECK_FAIL(!prog_array))
+		goto out;
+
+	map_fd = bpf_map__fd(prog_array);
+	if (CHECK_FAIL(map_fd < 0))
+		goto out;
+
+	data_map = bpf_object__find_map_by_name(obj, "tailcall.bss");
+	if (CHECK_FAIL(!data_map || !bpf_map__is_internal(data_map)))
+		return;
+
+	data_fd = bpf_map__fd(data_map);
+	if (CHECK_FAIL(map_fd < 0))
+		return;
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		snprintf(prog_name, sizeof(prog_name), "classifier/%i", i);
+
+		prog = bpf_object__find_program_by_title(obj, prog_name);
+		if (CHECK_FAIL(!prog))
+			goto out;
+
+		prog_fd = bpf_program__fd(prog);
+		if (CHECK_FAIL(prog_fd < 0))
+			goto out;
+
+		err = bpf_map_update_elem(map_fd, &i, &prog_fd, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		err = bpf_map_update_elem(data_fd, &zero, &i, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+
+		err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+					&duration, &retval, NULL);
+		CHECK(err || retval != i, "tailcall",
+		      "err %d errno %d retval %d\n", err, errno, retval);
+	}
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		err = bpf_map_update_elem(data_fd, &zero, &i, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+
+		err = bpf_map_delete_elem(map_fd, &i);
+		if (CHECK_FAIL(err))
+			goto out;
+
+		err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+					&duration, &retval, NULL);
+		CHECK(err || retval != 3, "tailcall",
+		      "err %d errno %d retval %d\n", err, errno, retval);
+	}
+out:
+	bpf_object__close(obj);
+}
+
+/* test_tailcall_5 probes similarly to test_tailcall_4 that the kernel generates
+ * an indirect jump when the keys are const but different from different branches.
+ */
+static void test_tailcall_5(void)
+{
+	int err, map_fd, prog_fd, main_fd, data_fd, i, key[] = { 1111, 1234, 5678 };
+	struct bpf_map *prog_array, *data_map;
+	struct bpf_program *prog;
+	struct bpf_object *obj;
+	__u32 retval, duration;
+	static const int zero = 0;
+	char buff[128] = {};
+	char prog_name[32];
+
+	err = bpf_prog_load("tailcall5.o", BPF_PROG_TYPE_SCHED_CLS, &obj,
+			    &prog_fd);
+	if (CHECK_FAIL(err))
+		return;
+
+	prog = bpf_object__find_program_by_title(obj, "classifier");
+	if (CHECK_FAIL(!prog))
+		goto out;
+
+	main_fd = bpf_program__fd(prog);
+	if (CHECK_FAIL(main_fd < 0))
+		goto out;
+
+	prog_array = bpf_object__find_map_by_name(obj, "jmp_table");
+	if (CHECK_FAIL(!prog_array))
+		goto out;
+
+	map_fd = bpf_map__fd(prog_array);
+	if (CHECK_FAIL(map_fd < 0))
+		goto out;
+
+	data_map = bpf_object__find_map_by_name(obj, "tailcall.bss");
+	if (CHECK_FAIL(!data_map || !bpf_map__is_internal(data_map)))
+		return;
+
+	data_fd = bpf_map__fd(data_map);
+	if (CHECK_FAIL(map_fd < 0))
+		return;
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		snprintf(prog_name, sizeof(prog_name), "classifier/%i", i);
+
+		prog = bpf_object__find_program_by_title(obj, prog_name);
+		if (CHECK_FAIL(!prog))
+			goto out;
+
+		prog_fd = bpf_program__fd(prog);
+		if (CHECK_FAIL(prog_fd < 0))
+			goto out;
+
+		err = bpf_map_update_elem(map_fd, &i, &prog_fd, BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+	}
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		err = bpf_map_update_elem(data_fd, &zero, &key[i], BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+
+		err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+					&duration, &retval, NULL);
+		CHECK(err || retval != i, "tailcall",
+		      "err %d errno %d retval %d\n", err, errno, retval);
+	}
+
+	for (i = 0; i < bpf_map__def(prog_array)->max_entries; i++) {
+		err = bpf_map_update_elem(data_fd, &zero, &key[i], BPF_ANY);
+		if (CHECK_FAIL(err))
+			goto out;
+
+		err = bpf_map_delete_elem(map_fd, &i);
+		if (CHECK_FAIL(err))
+			goto out;
+
+		err = bpf_prog_test_run(main_fd, 1, buff, sizeof(buff), 0,
+					&duration, &retval, NULL);
+		CHECK(err || retval != 3, "tailcall",
+		      "err %d errno %d retval %d\n", err, errno, retval);
+	}
+out:
+	bpf_object__close(obj);
+}
+
+void test_tailcalls(void)
+{
+	if (test__start_subtest("tailcall_1"))
+		test_tailcall_1();
+	if (test__start_subtest("tailcall_2"))
+		test_tailcall_2();
+	if (test__start_subtest("tailcall_3"))
+		test_tailcall_3();
+	if (test__start_subtest("tailcall_4"))
+		test_tailcall_4();
+	if (test__start_subtest("tailcall_5"))
+		test_tailcall_5();
+}
diff --git a/tools/testing/selftests/bpf/progs/tailcall1.c b/tools/testing/selftests/bpf/progs/tailcall1.c
new file mode 100644
index 000000000000..63531e1a9fa4
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall1.c
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/bpf.h>
+
+#include "bpf_helpers.h"
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 3);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} jmp_table SEC(".maps");
+
+#define TAIL_FUNC(x) 				\
+	SEC("classifier/" #x)			\
+	int bpf_func_##x(struct __sk_buff *skb)	\
+	{					\
+		return x;			\
+	}
+TAIL_FUNC(0)
+TAIL_FUNC(1)
+TAIL_FUNC(2)
+
+SEC("classifier")
+int entry(struct __sk_buff *skb)
+{
+	/* Multiple locations to make sure we patch
+	 * all of them.
+	 */
+	bpf_tail_call(skb, &jmp_table, 0);
+	bpf_tail_call(skb, &jmp_table, 0);
+	bpf_tail_call(skb, &jmp_table, 0);
+	bpf_tail_call(skb, &jmp_table, 0);
+
+	bpf_tail_call(skb, &jmp_table, 1);
+	bpf_tail_call(skb, &jmp_table, 1);
+	bpf_tail_call(skb, &jmp_table, 1);
+	bpf_tail_call(skb, &jmp_table, 1);
+
+	bpf_tail_call(skb, &jmp_table, 2);
+	bpf_tail_call(skb, &jmp_table, 2);
+	bpf_tail_call(skb, &jmp_table, 2);
+	bpf_tail_call(skb, &jmp_table, 2);
+
+	return 3;
+}
+
+char __license[] SEC("license") = "GPL";
+int _version SEC("version") = 1;
diff --git a/tools/testing/selftests/bpf/progs/tailcall2.c b/tools/testing/selftests/bpf/progs/tailcall2.c
new file mode 100644
index 000000000000..21c85c477210
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall2.c
@@ -0,0 +1,59 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/bpf.h>
+
+#include "bpf_helpers.h"
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 5);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} jmp_table SEC(".maps");
+
+SEC("classifier/0")
+int bpf_func_0(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &jmp_table, 1);
+	return 0;
+}
+
+SEC("classifier/1")
+int bpf_func_1(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &jmp_table, 2);
+	return 1;
+}
+
+SEC("classifier/2")
+int bpf_func_2(struct __sk_buff *skb)
+{
+	return 2;
+}
+
+SEC("classifier/3")
+int bpf_func_3(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &jmp_table, 4);
+	return 3;
+}
+
+SEC("classifier/4")
+int bpf_func_4(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &jmp_table, 3);
+	return 4;
+}
+
+SEC("classifier")
+int entry(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &jmp_table, 0);
+	/* Check multi-prog update. */
+	bpf_tail_call(skb, &jmp_table, 2);
+	/* Check tail call limit. */
+	bpf_tail_call(skb, &jmp_table, 3);
+	return 3;
+}
+
+char __license[] SEC("license") = "GPL";
+int _version SEC("version") = 1;
diff --git a/tools/testing/selftests/bpf/progs/tailcall3.c b/tools/testing/selftests/bpf/progs/tailcall3.c
new file mode 100644
index 000000000000..1ecae198b8c1
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall3.c
@@ -0,0 +1,31 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/bpf.h>
+
+#include "bpf_helpers.h"
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 1);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} jmp_table SEC(".maps");
+
+static volatile int count;
+
+SEC("classifier/0")
+int bpf_func_0(struct __sk_buff *skb)
+{
+	count++;
+	bpf_tail_call(skb, &jmp_table, 0);
+	return 1;
+}
+
+SEC("classifier")
+int entry(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &jmp_table, 0);
+	return 0;
+}
+
+char __license[] SEC("license") = "GPL";
+int _version SEC("version") = 1;
diff --git a/tools/testing/selftests/bpf/progs/tailcall4.c b/tools/testing/selftests/bpf/progs/tailcall4.c
new file mode 100644
index 000000000000..499388758119
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall4.c
@@ -0,0 +1,33 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/bpf.h>
+
+#include "bpf_helpers.h"
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 3);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} jmp_table SEC(".maps");
+
+static volatile int selector;
+
+#define TAIL_FUNC(x)				\
+	SEC("classifier/" #x)			\
+	int bpf_func_##x(struct __sk_buff *skb)	\
+	{					\
+		return x;			\
+	}
+TAIL_FUNC(0)
+TAIL_FUNC(1)
+TAIL_FUNC(2)
+
+SEC("classifier")
+int entry(struct __sk_buff *skb)
+{
+	bpf_tail_call(skb, &jmp_table, selector);
+	return 3;
+}
+
+char __license[] SEC("license") = "GPL";
+int _version SEC("version") = 1;
diff --git a/tools/testing/selftests/bpf/progs/tailcall5.c b/tools/testing/selftests/bpf/progs/tailcall5.c
new file mode 100644
index 000000000000..49c64eb53f19
--- /dev/null
+++ b/tools/testing/selftests/bpf/progs/tailcall5.c
@@ -0,0 +1,40 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/bpf.h>
+
+#include "bpf_helpers.h"
+
+struct {
+	__uint(type, BPF_MAP_TYPE_PROG_ARRAY);
+	__uint(max_entries, 3);
+	__uint(key_size, sizeof(__u32));
+	__uint(value_size, sizeof(__u32));
+} jmp_table SEC(".maps");
+
+static volatile int selector;
+
+#define TAIL_FUNC(x)				\
+	SEC("classifier/" #x)			\
+	int bpf_func_##x(struct __sk_buff *skb)	\
+	{					\
+		return x;			\
+	}
+TAIL_FUNC(0)
+TAIL_FUNC(1)
+TAIL_FUNC(2)
+
+SEC("classifier")
+int entry(struct __sk_buff *skb)
+{
+	int idx = 0;
+
+	if (selector == 1234)
+		idx = 1;
+	else if (selector == 5678)
+		idx = 2;
+
+	bpf_tail_call(skb, &jmp_table, idx);
+	return 3;
+}
+
+char __license[] SEC("license") = "GPL";
+int _version SEC("version") = 1;