bpf.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework
@ 2019-12-12 10:22 Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 01/10] libbpf: Recognize SK_REUSEPORT programs from section name Jakub Sitnicki
                   ` (10 more replies)
  0 siblings, 11 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

This change has been suggested by Martin Lau [0] during a review of a
related patch set that extends reuseport tests [1].

Patches 1 & 2 address a warning due to unrecognized section name from
libbpf when running reuseport tests. We don't want to carry this warning
into test_progs.

Patches 3-8 massage the reuseport tests to ease the switch to test_progs
framework. The intention here is to show the work. Happy to squash these,
if needed.

Patches 9-10 do the actual move and conversion to test_progs.

Output from a test_progs run after changes pasted below.

Thanks,
Jakub

[0] https://lore.kernel.org/bpf/20191123110751.6729-1-jakub@cloudflare.com/T/#m607d822caeb1eb5db101172821a78cc3896ff1c3
[1] https://lore.kernel.org/bpf/20191123110751.6729-1-jakub@cloudflare.com/T/#m55881bae9fb6e34837d07a0c0a7ffbc138f8d06f

---8<---
bash-5.0# ./test_progs -t select_reuseport
#33/1 IPv4/TCP LOOPBACK test_err_inner_map:OK
#33/2 IPv4/TCP LOOPBACK test_err_skb_data:OK
#33/3 IPv4/TCP LOOPBACK test_err_sk_select_port:OK
#33/4 IPv4/TCP LOOPBACK test_pass:OK
#33/5 IPv4/TCP LOOPBACK test_syncookie:OK
#33/6 IPv4/TCP LOOPBACK test_pass_on_err:OK
#33/7 IPv4/TCP LOOPBACK test_detach_bpf:OK
#33/8 IPv4/TCP INANY test_err_inner_map:OK
#33/9 IPv4/TCP INANY test_err_skb_data:OK
#33/10 IPv4/TCP INANY test_err_sk_select_port:OK
#33/11 IPv4/TCP INANY test_pass:OK
#33/12 IPv4/TCP INANY test_syncookie:OK
#33/13 IPv4/TCP INANY test_pass_on_err:OK
#33/14 IPv4/TCP INANY test_detach_bpf:OK
#33/15 IPv6/TCP LOOPBACK test_err_inner_map:OK
#33/16 IPv6/TCP LOOPBACK test_err_skb_data:OK
#33/17 IPv6/TCP LOOPBACK test_err_sk_select_port:OK
#33/18 IPv6/TCP LOOPBACK test_pass:OK
#33/19 IPv6/TCP LOOPBACK test_syncookie:OK
#33/20 IPv6/TCP LOOPBACK test_pass_on_err:OK
#33/21 IPv6/TCP LOOPBACK test_detach_bpf:OK
#33/22 IPv6/TCP INANY test_err_inner_map:OK
#33/23 IPv6/TCP INANY test_err_skb_data:OK
#33/24 IPv6/TCP INANY test_err_sk_select_port:OK
#33/25 IPv6/TCP INANY test_pass:OK
#33/26 IPv6/TCP INANY test_syncookie:OK
#33/27 IPv6/TCP INANY test_pass_on_err:OK
#33/28 IPv6/TCP INANY test_detach_bpf:OK
#33/29 IPv4/UDP LOOPBACK test_err_inner_map:OK
#33/30 IPv4/UDP LOOPBACK test_err_skb_data:OK
#33/31 IPv4/UDP LOOPBACK test_err_sk_select_port:OK
#33/32 IPv4/UDP LOOPBACK test_pass:OK
#33/33 IPv4/UDP LOOPBACK test_syncookie:OK
#33/34 IPv4/UDP LOOPBACK test_pass_on_err:OK
#33/35 IPv4/UDP LOOPBACK test_detach_bpf:OK
#33/36 IPv6/UDP LOOPBACK test_err_inner_map:OK
#33/37 IPv6/UDP LOOPBACK test_err_skb_data:OK
#33/38 IPv6/UDP LOOPBACK test_err_sk_select_port:OK
#33/39 IPv6/UDP LOOPBACK test_pass:OK
#33/40 IPv6/UDP LOOPBACK test_syncookie:OK
#33/41 IPv6/UDP LOOPBACK test_pass_on_err:OK
#33/42 IPv6/UDP LOOPBACK test_detach_bpf:OK
#33 select_reuseport:OK
Summary: 1/42 PASSED, 0 SKIPPED, 0 FAILED
--->8---


Jakub Sitnicki (10):
  libbpf: Recognize SK_REUSEPORT programs from section name
  selftests/bpf: Let libbpf determine program type from section name
  selftests/bpf: Use sa_family_t everywhere in reuseport tests
  selftests/bpf: Add helpers for getting socket family & type name
  selftests/bpf: Unroll the main loop in reuseport test
  selftests/bpf: Run reuseport tests in a loop
  selftests/bpf: Propagate errors during setup for reuseport tests
  selftests/bpf: Pull up printing the test name into test runner
  selftests/bpf: Move reuseport tests under prog_tests/
  selftests/bpf: Switch reuseport tests for test_progs framework

 tools/lib/bpf/libbpf.c                        |   1 +
 tools/testing/selftests/bpf/Makefile          |   2 +-
 .../select_reuseport.c}                       | 514 ++++++++++--------
 .../bpf/progs/test_select_reuseport_kern.c    |   2 +-
 4 files changed, 292 insertions(+), 227 deletions(-)
 rename tools/testing/selftests/bpf/{test_select_reuseport.c => prog_tests/select_reuseport.c} (54%)

-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 01/10] libbpf: Recognize SK_REUSEPORT programs from section name
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 02/10] selftests/bpf: Let libbpf determine program type " Jakub Sitnicki
                   ` (9 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Allow loading BPF object files that contain SK_REUSEPORT programs without
having to manually set the program type before load if the the section name
is set to "sk_reuseport".

Makes user-space code needed to load SK_REUSEPORT BPF program more concise.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 tools/lib/bpf/libbpf.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
index 3f09772192f1..13f6db467cdd 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -4982,6 +4982,7 @@ static const struct {
 	enum bpf_attach_type attach_type;
 } section_names[] = {
 	BPF_PROG_SEC("socket",			BPF_PROG_TYPE_SOCKET_FILTER),
+	BPF_PROG_SEC("sk_reuseport",		BPF_PROG_TYPE_SK_REUSEPORT),
 	BPF_PROG_SEC("kprobe/",			BPF_PROG_TYPE_KPROBE),
 	BPF_PROG_SEC("uprobe/",			BPF_PROG_TYPE_KPROBE),
 	BPF_PROG_SEC("kretprobe/",		BPF_PROG_TYPE_KPROBE),
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 02/10] selftests/bpf: Let libbpf determine program type from section name
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 01/10] libbpf: Recognize SK_REUSEPORT programs from section name Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 03/10] selftests/bpf: Use sa_family_t everywhere in reuseport tests Jakub Sitnicki
                   ` (8 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Now that libbpf can recognize SK_REUSEPORT programs, we no longer have to
pass a prog_type hint before loading the object file.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 .../selftests/bpf/progs/test_select_reuseport_kern.c |  2 +-
 tools/testing/selftests/bpf/test_select_reuseport.c  | 12 +++---------
 2 files changed, 4 insertions(+), 10 deletions(-)

diff --git a/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c b/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c
index ea7d84f01235..b1f09f5bb1cf 100644
--- a/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c
+++ b/tools/testing/selftests/bpf/progs/test_select_reuseport_kern.c
@@ -62,7 +62,7 @@ struct {
 	goto done;				\
 })
 
-SEC("select_by_skb_data")
+SEC("sk_reuseport")
 int _select_by_skb_data(struct sk_reuseport_md *reuse_md)
 {
 	__u32 linum, index = 0, flags = 0, index_zero = 0;
diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
index 7566c13eb51a..1e3cfe1cb28a 100644
--- a/tools/testing/selftests/bpf/test_select_reuseport.c
+++ b/tools/testing/selftests/bpf/test_select_reuseport.c
@@ -87,19 +87,11 @@ static void prepare_bpf_obj(void)
 	struct bpf_program *prog;
 	struct bpf_map *map;
 	int err;
-	struct bpf_object_open_attr attr = {
-		.file = "test_select_reuseport_kern.o",
-		.prog_type = BPF_PROG_TYPE_SK_REUSEPORT,
-	};
 
-	obj = bpf_object__open_xattr(&attr);
+	obj = bpf_object__open("test_select_reuseport_kern.o");
 	CHECK(IS_ERR_OR_NULL(obj), "open test_select_reuseport_kern.o",
 	      "obj:%p PTR_ERR(obj):%ld\n", obj, PTR_ERR(obj));
 
-	prog = bpf_program__next(NULL, obj);
-	CHECK(!prog, "get first bpf_program", "!prog\n");
-	bpf_program__set_type(prog, attr.prog_type);
-
 	map = bpf_object__find_map_by_name(obj, "outer_map");
 	CHECK(!map, "find outer_map", "!map\n");
 	err = bpf_map__reuse_fd(map, outer_map);
@@ -108,6 +100,8 @@ static void prepare_bpf_obj(void)
 	err = bpf_object__load(obj);
 	CHECK(err, "load bpf_object", "err:%d\n", err);
 
+	prog = bpf_program__next(NULL, obj);
+	CHECK(!prog, "get first bpf_program", "!prog\n");
 	select_by_skb_data_prog = bpf_program__fd(prog);
 	CHECK(select_by_skb_data_prog == -1, "get prog fd",
 	      "select_by_skb_data_prog:%d\n", select_by_skb_data_prog);
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 03/10] selftests/bpf: Use sa_family_t everywhere in reuseport tests
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 01/10] libbpf: Recognize SK_REUSEPORT programs from section name Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 02/10] selftests/bpf: Let libbpf determine program type " Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 04/10] selftests/bpf: Add helpers for getting socket family & type name Jakub Sitnicki
                   ` (7 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Update the only function that is not using sa_family_t in this source file.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 tools/testing/selftests/bpf/test_select_reuseport.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
index 1e3cfe1cb28a..a295a087a026 100644
--- a/tools/testing/selftests/bpf/test_select_reuseport.c
+++ b/tools/testing/selftests/bpf/test_select_reuseport.c
@@ -643,7 +643,7 @@ static void prepare_sk_fds(int type, sa_family_t family, bool inany)
 	}
 }
 
-static void setup_per_test(int type, unsigned short family, bool inany)
+static void setup_per_test(int type, sa_family_t family, bool inany)
 {
 	int ovr = -1, err;
 
@@ -680,12 +680,12 @@ static void test_all(void)
 	const int types[] = { SOCK_STREAM, SOCK_DGRAM, SOCK_STREAM };
 	const char * const type_strings[] = { "TCP", "UDP", "TCP" };
 	const char * const family_strings[] = { "IPv6", "IPv4" };
-	const unsigned short families[] = { AF_INET6, AF_INET };
+	const sa_family_t families[] = { AF_INET6, AF_INET };
 	const bool bind_inany[] = { false, false, true };
 	int t, f, err;
 
 	for (f = 0; f < ARRAY_SIZE(families); f++) {
-		unsigned short family = families[f];
+		sa_family_t family = families[f];
 
 		for (t = 0; t < ARRAY_SIZE(types); t++) {
 			bool inany = bind_inany[t];
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 04/10] selftests/bpf: Add helpers for getting socket family & type name
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (2 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 03/10] selftests/bpf: Use sa_family_t everywhere in reuseport tests Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 05/10] selftests/bpf: Unroll the main loop in reuseport test Jakub Sitnicki
                   ` (6 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Having string arrays to map socket family & type to a name prevents us from
unrolling the test runner loop in the subsequent patch. Introduce helpers
that do the same thing.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 .../selftests/bpf/test_select_reuseport.c     | 28 +++++++++++++++++--
 1 file changed, 25 insertions(+), 3 deletions(-)

diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
index a295a087a026..ef98a7de6704 100644
--- a/tools/testing/selftests/bpf/test_select_reuseport.c
+++ b/tools/testing/selftests/bpf/test_select_reuseport.c
@@ -674,12 +674,34 @@ static void cleanup(void)
 	bpf_object__close(obj);
 }
 
+static const char *family_str(sa_family_t family)
+{
+	switch (family) {
+	case AF_INET:
+		return "IPv4";
+	case AF_INET6:
+		return "IPv6";
+	default:
+		return "unknown";
+	}
+}
+
+static const char *sotype_str(int sotype)
+{
+	switch (sotype) {
+	case SOCK_STREAM:
+		return "TCP";
+	case SOCK_DGRAM:
+		return "UDP";
+	default:
+		return "unknown";
+	}
+}
+
 static void test_all(void)
 {
 	/* Extra SOCK_STREAM to test bind_inany==true */
 	const int types[] = { SOCK_STREAM, SOCK_DGRAM, SOCK_STREAM };
-	const char * const type_strings[] = { "TCP", "UDP", "TCP" };
-	const char * const family_strings[] = { "IPv6", "IPv4" };
 	const sa_family_t families[] = { AF_INET6, AF_INET };
 	const bool bind_inany[] = { false, false, true };
 	int t, f, err;
@@ -692,7 +714,7 @@ static void test_all(void)
 			int type = types[t];
 
 			printf("######## %s/%s %s ########\n",
-			       family_strings[f], type_strings[t],
+			       family_str(family), sotype_str(type),
 				inany ? " INANY  " : "LOOPBACK");
 
 			setup_per_test(type, family, inany);
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 05/10] selftests/bpf: Unroll the main loop in reuseport test
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (3 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 04/10] selftests/bpf: Add helpers for getting socket family & type name Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 06/10] selftests/bpf: Run reuseport tests in a loop Jakub Sitnicki
                   ` (5 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Prepare for iterating over individual tests without introducing another
nested loop in the main test function.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 .../selftests/bpf/test_select_reuseport.c     | 73 +++++++++++--------
 1 file changed, 41 insertions(+), 32 deletions(-)

diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
index ef98a7de6704..63ce2e75e758 100644
--- a/tools/testing/selftests/bpf/test_select_reuseport.c
+++ b/tools/testing/selftests/bpf/test_select_reuseport.c
@@ -698,47 +698,56 @@ static const char *sotype_str(int sotype)
 	}
 }
 
-static void test_all(void)
+static void test_config(int type, sa_family_t family, bool inany)
 {
-	/* Extra SOCK_STREAM to test bind_inany==true */
-	const int types[] = { SOCK_STREAM, SOCK_DGRAM, SOCK_STREAM };
-	const sa_family_t families[] = { AF_INET6, AF_INET };
-	const bool bind_inany[] = { false, false, true };
-	int t, f, err;
+	int err;
 
-	for (f = 0; f < ARRAY_SIZE(families); f++) {
-		sa_family_t family = families[f];
+	printf("######## %s/%s %s ########\n",
+	       family_str(family), sotype_str(type),
+	       inany ? " INANY  " : "LOOPBACK");
 
-		for (t = 0; t < ARRAY_SIZE(types); t++) {
-			bool inany = bind_inany[t];
-			int type = types[t];
+	setup_per_test(type, family, inany);
 
-			printf("######## %s/%s %s ########\n",
-			       family_str(family), sotype_str(type),
-				inany ? " INANY  " : "LOOPBACK");
+	test_err_inner_map(type, family);
 
-			setup_per_test(type, family, inany);
+	/* Install reuseport_array to the outer_map */
+	err = bpf_map_update_elem(outer_map, &index_zero,
+				  &reuseport_array, BPF_ANY);
+	CHECK(err == -1, "update_elem(outer_map)",
+	      "err:%d errno:%d\n", err, errno);
 
-			test_err_inner_map(type, family);
+	test_err_skb_data(type, family);
+	test_err_sk_select_port(type, family);
+	test_pass(type, family);
+	test_syncookie(type, family);
+	test_pass_on_err(type, family);
+	/* Must be the last test */
+	test_detach_bpf(type, family);
 
-			/* Install reuseport_array to the outer_map */
-			err = bpf_map_update_elem(outer_map, &index_zero,
-						  &reuseport_array, BPF_ANY);
-			CHECK(err == -1, "update_elem(outer_map)",
-			      "err:%d errno:%d\n", err, errno);
+	cleanup_per_test();
+	printf("\n");
+}
 
-			test_err_skb_data(type, family);
-			test_err_sk_select_port(type, family);
-			test_pass(type, family);
-			test_syncookie(type, family);
-			test_pass_on_err(type, family);
-			/* Must be the last test */
-			test_detach_bpf(type, family);
+#define BIND_INANY true
 
-			cleanup_per_test();
-			printf("\n");
-		}
-	}
+static void test_all(void)
+{
+	const struct config {
+		int sotype;
+		sa_family_t family;
+		bool inany;
+	} configs[] = {
+		{ SOCK_STREAM, AF_INET },
+		{ SOCK_STREAM, AF_INET, BIND_INANY },
+		{ SOCK_STREAM, AF_INET6 },
+		{ SOCK_STREAM, AF_INET6, BIND_INANY },
+		{ SOCK_DGRAM, AF_INET },
+		{ SOCK_DGRAM, AF_INET6 },
+	};
+	const struct config *c;
+
+	for (c = configs; c < configs + ARRAY_SIZE(configs); c++)
+		test_config(c->sotype, c->family, c->inany);
 }
 
 int main(int argc, const char **argv)
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 06/10] selftests/bpf: Run reuseport tests in a loop
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (4 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 05/10] selftests/bpf: Unroll the main loop in reuseport test Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 07/10] selftests/bpf: Propagate errors during setup for reuseport tests Jakub Sitnicki
                   ` (4 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Prepare for switching reuseport tests to test_progs framework. Loop over
the tests and perform setup/cleanup for each test separately, remembering
that with test_progs we can select tests to run.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 .../selftests/bpf/test_select_reuseport.c     | 55 ++++++++++++-------
 1 file changed, 34 insertions(+), 21 deletions(-)

diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
index 63ce2e75e758..cfff958da570 100644
--- a/tools/testing/selftests/bpf/test_select_reuseport.c
+++ b/tools/testing/selftests/bpf/test_select_reuseport.c
@@ -643,7 +643,8 @@ static void prepare_sk_fds(int type, sa_family_t family, bool inany)
 	}
 }
 
-static void setup_per_test(int type, sa_family_t family, bool inany)
+static void setup_per_test(int type, sa_family_t family, bool inany,
+			   bool no_inner_map)
 {
 	int ovr = -1, err;
 
@@ -652,9 +653,18 @@ static void setup_per_test(int type, sa_family_t family, bool inany)
 				  BPF_ANY);
 	CHECK(err == -1, "update_elem(tmp_index_ovr_map, 0, -1)",
 	      "err:%d errno:%d\n", err, errno);
+
+	/* Install reuseport_array to outer_map? */
+	if (no_inner_map)
+		return;
+
+	err = bpf_map_update_elem(outer_map, &index_zero, &reuseport_array,
+				  BPF_ANY);
+	CHECK(err == -1, "update_elem(outer_map, 0, reuseport_array)",
+	      "err:%d errno:%d\n", err, errno);
 }
 
-static void cleanup_per_test(void)
+static void cleanup_per_test(bool no_inner_map)
 {
 	int i, err;
 
@@ -662,6 +672,10 @@ static void cleanup_per_test(void)
 		close(sk_fds[i]);
 	close(epfd);
 
+	/* Delete reuseport_array from outer_map? */
+	if (no_inner_map)
+		return;
+
 	err = bpf_map_delete_elem(outer_map, &index_zero);
 	CHECK(err == -1, "delete_elem(outer_map)",
 	      "err:%d errno:%d\n", err, errno);
@@ -700,31 +714,30 @@ static const char *sotype_str(int sotype)
 
 static void test_config(int type, sa_family_t family, bool inany)
 {
-	int err;
+	const struct test {
+		void (*fn)(int sotype, sa_family_t family);
+		bool no_inner_map;
+	} tests[] = {
+		{ test_err_inner_map, true /* no_inner_map */ },
+		{ test_err_skb_data },
+		{ test_err_sk_select_port },
+		{ test_pass },
+		{ test_syncookie },
+		{ test_pass_on_err },
+		{ test_detach_bpf },
+	};
+	const struct test *t;
 
 	printf("######## %s/%s %s ########\n",
 	       family_str(family), sotype_str(type),
 	       inany ? " INANY  " : "LOOPBACK");
 
-	setup_per_test(type, family, inany);
-
-	test_err_inner_map(type, family);
-
-	/* Install reuseport_array to the outer_map */
-	err = bpf_map_update_elem(outer_map, &index_zero,
-				  &reuseport_array, BPF_ANY);
-	CHECK(err == -1, "update_elem(outer_map)",
-	      "err:%d errno:%d\n", err, errno);
-
-	test_err_skb_data(type, family);
-	test_err_sk_select_port(type, family);
-	test_pass(type, family);
-	test_syncookie(type, family);
-	test_pass_on_err(type, family);
-	/* Must be the last test */
-	test_detach_bpf(type, family);
+	for (t = tests; t < tests + ARRAY_SIZE(tests); t++) {
+		setup_per_test(type, family, inany, t->no_inner_map);
+		t->fn(type, family);
+		cleanup_per_test(t->no_inner_map);
+	}
 
-	cleanup_per_test();
 	printf("\n");
 }
 
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 07/10] selftests/bpf: Propagate errors during setup for reuseport tests
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (5 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 06/10] selftests/bpf: Run reuseport tests in a loop Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 08/10] selftests/bpf: Pull up printing the test name into test runner Jakub Sitnicki
                   ` (3 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Prepare for switching reuseport tests to test_progs framework, where we
don't have the luxury to terminate the process on failure.

Modify setup helpers to signal failure via the return value with the help
of a macro similar to the one currently in use by the tests.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 .../selftests/bpf/test_select_reuseport.c     | 136 +++++++++++-------
 1 file changed, 85 insertions(+), 51 deletions(-)

diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
index cfff958da570..cc35816b7b34 100644
--- a/tools/testing/selftests/bpf/test_select_reuseport.c
+++ b/tools/testing/selftests/bpf/test_select_reuseport.c
@@ -32,7 +32,7 @@
 static int result_map, tmp_index_ovr_map, linum_map, data_check_map;
 static enum result expected_results[NR_RESULTS];
 static int sk_fds[REUSEPORT_ARRAY_SIZE];
-static int reuseport_array, outer_map;
+static int reuseport_array = -1, outer_map = -1;
 static int select_by_skb_data_prog;
 static int saved_tcp_syncookie;
 static struct bpf_object *obj;
@@ -55,7 +55,16 @@ static union sa46 {
 	}								\
 })
 
-static void create_maps(void)
+#define RET_ERR(condition, tag, format...) ({				\
+	int __ret = !!(condition);					\
+	if (__ret) {							\
+		printf("%s(%d):FAIL:%s ", __func__, __LINE__, tag);	\
+		printf(format);						\
+		return -1;						\
+	}								\
+})
+
+static int create_maps(void)
 {
 	struct bpf_create_map_attr attr = {};
 
@@ -67,8 +76,8 @@ static void create_maps(void)
 	attr.max_entries = REUSEPORT_ARRAY_SIZE;
 
 	reuseport_array = bpf_create_map_xattr(&attr);
-	CHECK(reuseport_array == -1, "creating reuseport_array",
-	      "reuseport_array:%d errno:%d\n", reuseport_array, errno);
+	RET_ERR(reuseport_array == -1, "creating reuseport_array",
+		"reuseport_array:%d errno:%d\n", reuseport_array, errno);
 
 	/* Creating outer_map */
 	attr.name = "outer_map";
@@ -78,57 +87,61 @@ static void create_maps(void)
 	attr.max_entries = 1;
 	attr.inner_map_fd = reuseport_array;
 	outer_map = bpf_create_map_xattr(&attr);
-	CHECK(outer_map == -1, "creating outer_map",
-	      "outer_map:%d errno:%d\n", outer_map, errno);
+	RET_ERR(outer_map == -1, "creating outer_map",
+		"outer_map:%d errno:%d\n", outer_map, errno);
+
+	return 0;
 }
 
-static void prepare_bpf_obj(void)
+static int prepare_bpf_obj(void)
 {
 	struct bpf_program *prog;
 	struct bpf_map *map;
 	int err;
 
 	obj = bpf_object__open("test_select_reuseport_kern.o");
-	CHECK(IS_ERR_OR_NULL(obj), "open test_select_reuseport_kern.o",
-	      "obj:%p PTR_ERR(obj):%ld\n", obj, PTR_ERR(obj));
+	RET_ERR(IS_ERR_OR_NULL(obj), "open test_select_reuseport_kern.o",
+		"obj:%p PTR_ERR(obj):%ld\n", obj, PTR_ERR(obj));
 
 	map = bpf_object__find_map_by_name(obj, "outer_map");
-	CHECK(!map, "find outer_map", "!map\n");
+	RET_ERR(!map, "find outer_map", "!map\n");
 	err = bpf_map__reuse_fd(map, outer_map);
-	CHECK(err, "reuse outer_map", "err:%d\n", err);
+	RET_ERR(err, "reuse outer_map", "err:%d\n", err);
 
 	err = bpf_object__load(obj);
-	CHECK(err, "load bpf_object", "err:%d\n", err);
+	RET_ERR(err, "load bpf_object", "err:%d\n", err);
 
 	prog = bpf_program__next(NULL, obj);
-	CHECK(!prog, "get first bpf_program", "!prog\n");
+	RET_ERR(!prog, "get first bpf_program", "!prog\n");
 	select_by_skb_data_prog = bpf_program__fd(prog);
-	CHECK(select_by_skb_data_prog == -1, "get prog fd",
-	      "select_by_skb_data_prog:%d\n", select_by_skb_data_prog);
+	RET_ERR(select_by_skb_data_prog == -1, "get prog fd",
+		"select_by_skb_data_prog:%d\n", select_by_skb_data_prog);
 
 	map = bpf_object__find_map_by_name(obj, "result_map");
-	CHECK(!map, "find result_map", "!map\n");
+	RET_ERR(!map, "find result_map", "!map\n");
 	result_map = bpf_map__fd(map);
-	CHECK(result_map == -1, "get result_map fd",
-	      "result_map:%d\n", result_map);
+	RET_ERR(result_map == -1, "get result_map fd",
+		"result_map:%d\n", result_map);
 
 	map = bpf_object__find_map_by_name(obj, "tmp_index_ovr_map");
-	CHECK(!map, "find tmp_index_ovr_map", "!map\n");
+	RET_ERR(!map, "find tmp_index_ovr_map\n", "!map");
 	tmp_index_ovr_map = bpf_map__fd(map);
-	CHECK(tmp_index_ovr_map == -1, "get tmp_index_ovr_map fd",
-	      "tmp_index_ovr_map:%d\n", tmp_index_ovr_map);
+	RET_ERR(tmp_index_ovr_map == -1, "get tmp_index_ovr_map fd",
+		"tmp_index_ovr_map:%d\n", tmp_index_ovr_map);
 
 	map = bpf_object__find_map_by_name(obj, "linum_map");
-	CHECK(!map, "find linum_map", "!map\n");
+	RET_ERR(!map, "find linum_map", "!map\n");
 	linum_map = bpf_map__fd(map);
-	CHECK(linum_map == -1, "get linum_map fd",
-	      "linum_map:%d\n", linum_map);
+	RET_ERR(linum_map == -1, "get linum_map fd",
+		"linum_map:%d\n", linum_map);
 
 	map = bpf_object__find_map_by_name(obj, "data_check_map");
-	CHECK(!map, "find data_check_map", "!map\n");
+	RET_ERR(!map, "find data_check_map", "!map\n");
 	data_check_map = bpf_map__fd(map);
-	CHECK(data_check_map == -1, "get data_check_map fd",
-	      "data_check_map:%d\n", data_check_map);
+	RET_ERR(data_check_map == -1, "get data_check_map fd",
+		"data_check_map:%d\n", data_check_map);
+
+	return 0;
 }
 
 static void sa46_init_loopback(union sa46 *sa, sa_family_t family)
@@ -157,31 +170,34 @@ static int read_int_sysctl(const char *sysctl)
 	int fd, ret;
 
 	fd = open(sysctl, 0);
-	CHECK(fd == -1, "open(sysctl)", "sysctl:%s fd:%d errno:%d\n",
-	      sysctl, fd, errno);
+	RET_ERR(fd == -1, "open(sysctl)",
+		"sysctl:%s fd:%d errno:%d\n", sysctl, fd, errno);
 
 	ret = read(fd, buf, sizeof(buf));
-	CHECK(ret <= 0, "read(sysctl)", "sysctl:%s ret:%d errno:%d\n",
-	      sysctl, ret, errno);
-	close(fd);
+	RET_ERR(ret <= 0, "read(sysctl)",
+		"sysctl:%s ret:%d errno:%d\n", sysctl, ret, errno);
 
+	close(fd);
 	return atoi(buf);
 }
 
-static void write_int_sysctl(const char *sysctl, int v)
+static int write_int_sysctl(const char *sysctl, int v)
 {
 	int fd, ret, size;
 	char buf[16];
 
 	fd = open(sysctl, O_RDWR);
-	CHECK(fd == -1, "open(sysctl)", "sysctl:%s fd:%d errno:%d\n",
-	      sysctl, fd, errno);
+	RET_ERR(fd == -1, "open(sysctl)",
+		"sysctl:%s fd:%d errno:%d\n", sysctl, fd, errno);
 
 	size = snprintf(buf, sizeof(buf), "%d", v);
 	ret = write(fd, buf, size);
-	CHECK(ret != size, "write(sysctl)",
-	      "sysctl:%s ret:%d size:%d errno:%d\n", sysctl, ret, size, errno);
+	RET_ERR(ret != size, "write(sysctl)",
+		"sysctl:%s ret:%d size:%d errno:%d\n",
+		sysctl, ret, size, errno);
+
 	close(fd);
+	return 0;
 }
 
 static void restore_sysctls(void)
@@ -190,22 +206,25 @@ static void restore_sysctls(void)
 	write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, saved_tcp_syncookie);
 }
 
-static void enable_fastopen(void)
+static int enable_fastopen(void)
 {
 	int fo;
 
 	fo = read_int_sysctl(TCP_FO_SYSCTL);
-	write_int_sysctl(TCP_FO_SYSCTL, fo | 7);
+	if (fo < 0)
+		return -1;
+
+	return write_int_sysctl(TCP_FO_SYSCTL, fo | 7);
 }
 
-static void enable_syncookie(void)
+static int enable_syncookie(void)
 {
-	write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, 2);
+	return write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, 2);
 }
 
-static void disable_syncookie(void)
+static int disable_syncookie(void)
 {
-	write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, 0);
+	return write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, 0);
 }
 
 static __u32 get_linum(void)
@@ -683,9 +702,12 @@ static void cleanup_per_test(bool no_inner_map)
 
 static void cleanup(void)
 {
-	close(outer_map);
-	close(reuseport_array);
-	bpf_object__close(obj);
+	if (outer_map != -1)
+		close(outer_map);
+	if (reuseport_array != -1)
+		close(reuseport_array);
+	if (obj)
+		bpf_object__close(obj);
 }
 
 static const char *family_str(sa_family_t family)
@@ -765,16 +787,28 @@ static void test_all(void)
 
 int main(int argc, const char **argv)
 {
-	create_maps();
-	prepare_bpf_obj();
+	int ret = EXIT_FAILURE;
+
+	if (create_maps())
+		goto out;
+	if (prepare_bpf_obj())
+		goto out;
+
 	saved_tcp_fo = read_int_sysctl(TCP_FO_SYSCTL);
 	saved_tcp_syncookie = read_int_sysctl(TCP_SYNCOOKIE_SYSCTL);
-	enable_fastopen();
-	disable_syncookie();
+	if (saved_tcp_syncookie < 0 || saved_tcp_syncookie < 0)
+		goto out;
 	atexit(restore_sysctls);
 
+	if (enable_fastopen())
+		goto out;
+	if (disable_syncookie())
+		goto out;
+
 	test_all();
 
+	ret = EXIT_SUCCESS;
+out:
 	cleanup();
-	return 0;
+	return ret;
 }
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 08/10] selftests/bpf: Pull up printing the test name into test runner
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (6 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 07/10] selftests/bpf: Propagate errors during setup for reuseport tests Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 09/10] selftests/bpf: Move reuseport tests under prog_tests/ Jakub Sitnicki
                   ` (2 subsequent siblings)
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Again, prepare for switching reuseport tests to test_progs framework.
test_progs framework will print the subtest name for us if we set it.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 .../selftests/bpf/test_select_reuseport.c     | 27 +++++++++----------
 1 file changed, 12 insertions(+), 15 deletions(-)

diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/test_select_reuseport.c
index cc35816b7b34..0d5687feb689 100644
--- a/tools/testing/selftests/bpf/test_select_reuseport.c
+++ b/tools/testing/selftests/bpf/test_select_reuseport.c
@@ -441,7 +441,6 @@ static void test_err_inner_map(int type, sa_family_t family)
 		.pass_on_failure = 0,
 	};
 
-	printf("%s: ", __func__);
 	expected_results[DROP_ERR_INNER_MAP]++;
 	do_test(type, family, &cmd, DROP_ERR_INNER_MAP);
 	printf("OK\n");
@@ -449,7 +448,6 @@ static void test_err_inner_map(int type, sa_family_t family)
 
 static void test_err_skb_data(int type, sa_family_t family)
 {
-	printf("%s: ", __func__);
 	expected_results[DROP_ERR_SKB_DATA]++;
 	do_test(type, family, NULL, DROP_ERR_SKB_DATA);
 	printf("OK\n");
@@ -462,7 +460,6 @@ static void test_err_sk_select_port(int type, sa_family_t family)
 		.pass_on_failure = 0,
 	};
 
-	printf("%s: ", __func__);
 	expected_results[DROP_ERR_SK_SELECT_REUSEPORT]++;
 	do_test(type, family, &cmd, DROP_ERR_SK_SELECT_REUSEPORT);
 	printf("OK\n");
@@ -473,7 +470,6 @@ static void test_pass(int type, sa_family_t family)
 	struct cmd cmd;
 	int i;
 
-	printf("%s: ", __func__);
 	cmd.pass_on_failure = 0;
 	for (i = 0; i < REUSEPORT_ARRAY_SIZE; i++) {
 		expected_results[PASS]++;
@@ -494,7 +490,6 @@ static void test_syncookie(int type, sa_family_t family)
 	if (type != SOCK_STREAM)
 		return;
 
-	printf("%s: ", __func__);
 	/*
 	 * +1 for TCP-SYN and
 	 * +1 for the TCP-ACK (ack the syncookie)
@@ -530,7 +525,6 @@ static void test_pass_on_err(int type, sa_family_t family)
 		.pass_on_failure = 1,
 	};
 
-	printf("%s: ", __func__);
 	expected_results[PASS_ERR_SK_SELECT_REUSEPORT] += 1;
 	do_test(type, family, &cmd, PASS_ERR_SK_SELECT_REUSEPORT);
 	printf("OK\n");
@@ -545,7 +539,6 @@ static void test_detach_bpf(int type, sa_family_t family)
 	struct cmd cmd = {};
 	int optvalue = 0;
 
-	printf("%s: ", __func__);
 	err = setsockopt(sk_fds[0], SOL_SOCKET, SO_DETACH_REUSEPORT_BPF,
 			 &optvalue, sizeof(optvalue));
 	CHECK(err == -1, "setsockopt(SO_DETACH_REUSEPORT_BPF)",
@@ -584,7 +577,7 @@ static void test_detach_bpf(int type, sa_family_t family)
 	printf("OK\n");
 	close(cli_fd);
 #else
-	printf("%s: SKIP\n", __func__);
+	printf("SKIP\n");
 #endif
 }
 
@@ -734,19 +727,22 @@ static const char *sotype_str(int sotype)
 	}
 }
 
+#define TEST_INIT(fn, ...) { fn, #fn, __VA_ARGS__ }
+
 static void test_config(int type, sa_family_t family, bool inany)
 {
 	const struct test {
 		void (*fn)(int sotype, sa_family_t family);
+		const char *name;
 		bool no_inner_map;
 	} tests[] = {
-		{ test_err_inner_map, true /* no_inner_map */ },
-		{ test_err_skb_data },
-		{ test_err_sk_select_port },
-		{ test_pass },
-		{ test_syncookie },
-		{ test_pass_on_err },
-		{ test_detach_bpf },
+		TEST_INIT(test_err_inner_map, true /* no_inner_map */),
+		TEST_INIT(test_err_skb_data),
+		TEST_INIT(test_err_sk_select_port),
+		TEST_INIT(test_pass),
+		TEST_INIT(test_syncookie),
+		TEST_INIT(test_pass_on_err),
+		TEST_INIT(test_detach_bpf),
 	};
 	const struct test *t;
 
@@ -756,6 +752,7 @@ static void test_config(int type, sa_family_t family, bool inany)
 
 	for (t = tests; t < tests + ARRAY_SIZE(tests); t++) {
 		setup_per_test(type, family, inany, t->no_inner_map);
+		printf("%s: ", t->name);
 		t->fn(type, family);
 		cleanup_per_test(t->no_inner_map);
 	}
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 09/10] selftests/bpf: Move reuseport tests under prog_tests/
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (7 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 08/10] selftests/bpf: Pull up printing the test name into test runner Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-12 10:22 ` [PATCH bpf-next 10/10] selftests/bpf: Switch reuseport tests for test_progs framework Jakub Sitnicki
  2019-12-13  5:36 ` [PATCH bpf-next 00/10] " Alexei Starovoitov
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

Do a pure move the show the actual work needed to adapt the tests in
subsequent patch at the cost of breaking test_progs build for the moment.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 tools/testing/selftests/bpf/Makefile                            | 2 +-
 .../{test_select_reuseport.c => prog_tests/select_reuseport.c}  | 0
 2 files changed, 1 insertion(+), 1 deletion(-)
 rename tools/testing/selftests/bpf/{test_select_reuseport.c => prog_tests/select_reuseport.c} (100%)

diff --git a/tools/testing/selftests/bpf/Makefile b/tools/testing/selftests/bpf/Makefile
index e0fe01d9ec33..90de7d75b5c6 100644
--- a/tools/testing/selftests/bpf/Makefile
+++ b/tools/testing/selftests/bpf/Makefile
@@ -28,7 +28,7 @@ LDLIBS += -lcap -lelf -lrt -lpthread
 TEST_GEN_PROGS = test_verifier test_tag test_maps test_lru_map test_lpm_map test_progs \
 	test_align test_verifier_log test_dev_cgroup test_tcpbpf_user \
 	test_sock test_btf test_sockmap get_cgroup_id_user test_socket_cookie \
-	test_cgroup_storage test_select_reuseport \
+	test_cgroup_storage \
 	test_netcnt test_tcpnotify_user test_sock_fields test_sysctl test_hashmap \
 	test_cgroup_attach test_progs-no_alu32
 
diff --git a/tools/testing/selftests/bpf/test_select_reuseport.c b/tools/testing/selftests/bpf/prog_tests/select_reuseport.c
similarity index 100%
rename from tools/testing/selftests/bpf/test_select_reuseport.c
rename to tools/testing/selftests/bpf/prog_tests/select_reuseport.c
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* [PATCH bpf-next 10/10] selftests/bpf: Switch reuseport tests for test_progs framework
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (8 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 09/10] selftests/bpf: Move reuseport tests under prog_tests/ Jakub Sitnicki
@ 2019-12-12 10:22 ` Jakub Sitnicki
  2019-12-13  5:36 ` [PATCH bpf-next 00/10] " Alexei Starovoitov
  10 siblings, 0 replies; 14+ messages in thread
From: Jakub Sitnicki @ 2019-12-12 10:22 UTC (permalink / raw)
  To: bpf; +Cc: Martin Lau, kernel-team

The tests were originally written in abort-on-error style. With the switch
to test_progs we can no longer do that. So at the risk of not cleaning up
some resource on failure, we now return to the caller on error.

That said, failure inside one test should not affect others because we run
setup/cleanup before/after every test.

Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
---
 .../bpf/prog_tests/select_reuseport.c         | 267 +++++++++---------
 1 file changed, 131 insertions(+), 136 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/select_reuseport.c b/tools/testing/selftests/bpf/prog_tests/select_reuseport.c
index 0d5687feb689..2c37ae7dc214 100644
--- a/tools/testing/selftests/bpf/prog_tests/select_reuseport.c
+++ b/tools/testing/selftests/bpf/prog_tests/select_reuseport.c
@@ -20,8 +20,11 @@
 #include <bpf/libbpf.h>
 #include "bpf_rlimit.h"
 #include "bpf_util.h"
+
+#include "test_progs.h"
 #include "test_select_reuseport_common.h"
 
+#define MAX_TEST_NAME 80
 #define MIN_TCPHDR_LEN 20
 #define UDPHDR_LEN 8
 
@@ -34,9 +37,9 @@ static enum result expected_results[NR_RESULTS];
 static int sk_fds[REUSEPORT_ARRAY_SIZE];
 static int reuseport_array = -1, outer_map = -1;
 static int select_by_skb_data_prog;
-static int saved_tcp_syncookie;
+static int saved_tcp_syncookie = -1;
 static struct bpf_object *obj;
-static int saved_tcp_fo;
+static int saved_tcp_fo = -1;
 static __u32 index_zero;
 static int epfd;
 
@@ -46,20 +49,16 @@ static union sa46 {
 	sa_family_t family;
 } srv_sa;
 
-#define CHECK(condition, tag, format...) ({				\
-	int __ret = !!(condition);					\
-	if (__ret) {							\
-		printf("%s(%d):FAIL:%s ", __func__, __LINE__, tag);	\
-		printf(format);						\
-		exit(-1);						\
+#define RET_IF(condition, tag, format...) ({				\
+	if (CHECK_FAIL(condition)) {					\
+		printf(tag " " format);					\
+		return;							\
 	}								\
 })
 
 #define RET_ERR(condition, tag, format...) ({				\
-	int __ret = !!(condition);					\
-	if (__ret) {							\
-		printf("%s(%d):FAIL:%s ", __func__, __LINE__, tag);	\
-		printf(format);						\
+	if (CHECK_FAIL(condition)) {					\
+		printf(tag " " format);					\
 		return -1;						\
 	}								\
 })
@@ -202,8 +201,10 @@ static int write_int_sysctl(const char *sysctl, int v)
 
 static void restore_sysctls(void)
 {
-	write_int_sysctl(TCP_FO_SYSCTL, saved_tcp_fo);
-	write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, saved_tcp_syncookie);
+	if (saved_tcp_fo != -1)
+		write_int_sysctl(TCP_FO_SYSCTL, saved_tcp_fo);
+	if (saved_tcp_syncookie != -1)
+		write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, saved_tcp_syncookie);
 }
 
 static int enable_fastopen(void)
@@ -227,14 +228,14 @@ static int disable_syncookie(void)
 	return write_int_sysctl(TCP_SYNCOOKIE_SYSCTL, 0);
 }
 
-static __u32 get_linum(void)
+static long get_linum(void)
 {
 	__u32 linum;
 	int err;
 
 	err = bpf_map_lookup_elem(linum_map, &index_zero, &linum);
-	CHECK(err == -1, "lookup_elem(linum_map)", "err:%d errno:%d\n",
-	      err, errno);
+	RET_ERR(err == -1, "lookup_elem(linum_map)", "err:%d errno:%d\n",
+		err, errno);
 
 	return linum;
 }
@@ -250,12 +251,12 @@ static void check_data(int type, sa_family_t family, const struct cmd *cmd,
 	addrlen = sizeof(cli_sa);
 	err = getsockname(cli_fd, (struct sockaddr *)&cli_sa,
 			  &addrlen);
-	CHECK(err == -1, "getsockname(cli_fd)", "err:%d errno:%d\n",
-	      err, errno);
+	RET_IF(err == -1, "getsockname(cli_fd)", "err:%d errno:%d\n",
+	       err, errno);
 
 	err = bpf_map_lookup_elem(data_check_map, &index_zero, &result);
-	CHECK(err == -1, "lookup_elem(data_check_map)", "err:%d errno:%d\n",
-	      err, errno);
+	RET_IF(err == -1, "lookup_elem(data_check_map)", "err:%d errno:%d\n",
+	       err, errno);
 
 	if (type == SOCK_STREAM) {
 		expected.len = MIN_TCPHDR_LEN;
@@ -297,22 +298,22 @@ static void check_data(int type, sa_family_t family, const struct cmd *cmd,
 		printf("expected: (0x%x, %u, %u)\n",
 		       expected.eth_protocol, expected.ip_protocol,
 		       expected.bind_inany);
-		CHECK(1, "data_check result != expected",
-		      "bpf_prog_linum:%u\n", get_linum());
+		RET_IF(1, "data_check result != expected",
+		       "bpf_prog_linum:%ld\n", get_linum());
 	}
 
-	CHECK(!result.hash, "data_check result.hash empty",
-	      "result.hash:%u", result.hash);
+	RET_IF(!result.hash, "data_check result.hash empty",
+	       "result.hash:%u", result.hash);
 
 	expected.len += cmd ? sizeof(*cmd) : 0;
 	if (type == SOCK_STREAM)
-		CHECK(expected.len > result.len, "expected.len > result.len",
-		      "expected.len:%u result.len:%u bpf_prog_linum:%u\n",
-		      expected.len, result.len, get_linum());
+		RET_IF(expected.len > result.len, "expected.len > result.len",
+		       "expected.len:%u result.len:%u bpf_prog_linum:%ld\n",
+		       expected.len, result.len, get_linum());
 	else
-		CHECK(expected.len != result.len, "expected.len != result.len",
-		      "expected.len:%u result.len:%u bpf_prog_linum:%u\n",
-		      expected.len, result.len, get_linum());
+		RET_IF(expected.len != result.len, "expected.len != result.len",
+		       "expected.len:%u result.len:%u bpf_prog_linum:%ld\n",
+		       expected.len, result.len, get_linum());
 }
 
 static void check_results(void)
@@ -323,8 +324,8 @@ static void check_results(void)
 
 	for (i = 0; i < NR_RESULTS; i++) {
 		err = bpf_map_lookup_elem(result_map, &i, &results[i]);
-		CHECK(err == -1, "lookup_elem(result_map)",
-		      "i:%u err:%d errno:%d\n", i, err, errno);
+		RET_IF(err == -1, "lookup_elem(result_map)",
+		       "i:%u err:%d errno:%d\n", i, err, errno);
 	}
 
 	for (i = 0; i < NR_RESULTS; i++) {
@@ -350,10 +351,10 @@ static void check_results(void)
 		printf(", %u", expected_results[i]);
 	printf("]\n");
 
-	CHECK(expected_results[broken] != results[broken],
-	      "unexpected result",
-	      "expected_results[%u] != results[%u] bpf_prog_linum:%u\n",
-	      broken, broken, get_linum());
+	RET_IF(expected_results[broken] != results[broken],
+	       "unexpected result",
+	       "expected_results[%u] != results[%u] bpf_prog_linum:%ld\n",
+	       broken, broken, get_linum());
 }
 
 static int send_data(int type, sa_family_t family, void *data, size_t len,
@@ -363,17 +364,17 @@ static int send_data(int type, sa_family_t family, void *data, size_t len,
 	int fd, err;
 
 	fd = socket(family, type, 0);
-	CHECK(fd == -1, "socket()", "fd:%d errno:%d\n", fd, errno);
+	RET_ERR(fd == -1, "socket()", "fd:%d errno:%d\n", fd, errno);
 
 	sa46_init_loopback(&cli_sa, family);
 	err = bind(fd, (struct sockaddr *)&cli_sa, sizeof(cli_sa));
-	CHECK(fd == -1, "bind(cli_sa)", "err:%d errno:%d\n", err, errno);
+	RET_ERR(fd == -1, "bind(cli_sa)", "err:%d errno:%d\n", err, errno);
 
 	err = sendto(fd, data, len, MSG_FASTOPEN, (struct sockaddr *)&srv_sa,
 		     sizeof(srv_sa));
-	CHECK(err != len && expected >= PASS,
-	      "sendto()", "family:%u err:%d errno:%d expected:%d\n",
-	      family, err, errno, expected);
+	RET_ERR(err != len && expected >= PASS,
+		"sendto()", "family:%u err:%d errno:%d expected:%d\n",
+		family, err, errno, expected);
 
 	return fd;
 }
@@ -388,47 +389,49 @@ static void do_test(int type, sa_family_t family, struct cmd *cmd,
 
 	cli_fd = send_data(type, family, cmd, cmd ? sizeof(*cmd) : 0,
 			   expected);
+	if (cli_fd < 0)
+		return;
 	nev = epoll_wait(epfd, &ev, 1, expected >= PASS ? 5 : 0);
-	CHECK((nev <= 0 && expected >= PASS) ||
-	      (nev > 0 && expected < PASS),
-	      "nev <> expected",
-	      "nev:%d expected:%d type:%d family:%d data:(%d, %d)\n",
-	      nev, expected, type, family,
-	      cmd ? cmd->reuseport_index : -1,
-	      cmd ? cmd->pass_on_failure : -1);
+	RET_IF((nev <= 0 && expected >= PASS) ||
+	       (nev > 0 && expected < PASS),
+	       "nev <> expected",
+	       "nev:%d expected:%d type:%d family:%d data:(%d, %d)\n",
+	       nev, expected, type, family,
+	       cmd ? cmd->reuseport_index : -1,
+	       cmd ? cmd->pass_on_failure : -1);
 	check_results();
 	check_data(type, family, cmd, cli_fd);
 
 	if (expected < PASS)
 		return;
 
-	CHECK(expected != PASS_ERR_SK_SELECT_REUSEPORT &&
-	      cmd->reuseport_index != ev.data.u32,
-	      "check cmd->reuseport_index",
-	      "cmd:(%u, %u) ev.data.u32:%u\n",
-	      cmd->pass_on_failure, cmd->reuseport_index, ev.data.u32);
+	RET_IF(expected != PASS_ERR_SK_SELECT_REUSEPORT &&
+	       cmd->reuseport_index != ev.data.u32,
+	       "check cmd->reuseport_index",
+	       "cmd:(%u, %u) ev.data.u32:%u\n",
+	       cmd->pass_on_failure, cmd->reuseport_index, ev.data.u32);
 
 	srv_fd = sk_fds[ev.data.u32];
 	if (type == SOCK_STREAM) {
 		int new_fd = accept(srv_fd, NULL, 0);
 
-		CHECK(new_fd == -1, "accept(srv_fd)",
-		      "ev.data.u32:%u new_fd:%d errno:%d\n",
-		      ev.data.u32, new_fd, errno);
+		RET_IF(new_fd == -1, "accept(srv_fd)",
+		       "ev.data.u32:%u new_fd:%d errno:%d\n",
+		       ev.data.u32, new_fd, errno);
 
 		nread = recv(new_fd, &rcv_cmd, sizeof(rcv_cmd), MSG_DONTWAIT);
-		CHECK(nread != sizeof(rcv_cmd),
-		      "recv(new_fd)",
-		      "ev.data.u32:%u nread:%zd sizeof(rcv_cmd):%zu errno:%d\n",
-		      ev.data.u32, nread, sizeof(rcv_cmd), errno);
+		RET_IF(nread != sizeof(rcv_cmd),
+		       "recv(new_fd)",
+		       "ev.data.u32:%u nread:%zd sizeof(rcv_cmd):%zu errno:%d\n",
+		       ev.data.u32, nread, sizeof(rcv_cmd), errno);
 
 		close(new_fd);
 	} else {
 		nread = recv(srv_fd, &rcv_cmd, sizeof(rcv_cmd), MSG_DONTWAIT);
-		CHECK(nread != sizeof(rcv_cmd),
-		      "recv(sk_fds)",
-		      "ev.data.u32:%u nread:%zd sizeof(rcv_cmd):%zu errno:%d\n",
-		      ev.data.u32, nread, sizeof(rcv_cmd), errno);
+		RET_IF(nread != sizeof(rcv_cmd),
+		       "recv(sk_fds)",
+		       "ev.data.u32:%u nread:%zd sizeof(rcv_cmd):%zu errno:%d\n",
+		       ev.data.u32, nread, sizeof(rcv_cmd), errno);
 	}
 
 	close(cli_fd);
@@ -443,14 +446,12 @@ static void test_err_inner_map(int type, sa_family_t family)
 
 	expected_results[DROP_ERR_INNER_MAP]++;
 	do_test(type, family, &cmd, DROP_ERR_INNER_MAP);
-	printf("OK\n");
 }
 
 static void test_err_skb_data(int type, sa_family_t family)
 {
 	expected_results[DROP_ERR_SKB_DATA]++;
 	do_test(type, family, NULL, DROP_ERR_SKB_DATA);
-	printf("OK\n");
 }
 
 static void test_err_sk_select_port(int type, sa_family_t family)
@@ -462,7 +463,6 @@ static void test_err_sk_select_port(int type, sa_family_t family)
 
 	expected_results[DROP_ERR_SK_SELECT_REUSEPORT]++;
 	do_test(type, family, &cmd, DROP_ERR_SK_SELECT_REUSEPORT);
-	printf("OK\n");
 }
 
 static void test_pass(int type, sa_family_t family)
@@ -476,7 +476,6 @@ static void test_pass(int type, sa_family_t family)
 		cmd.reuseport_index = i;
 		do_test(type, family, &cmd, PASS);
 	}
-	printf("OK\n");
 }
 
 static void test_syncookie(int type, sa_family_t family)
@@ -505,17 +504,16 @@ static void test_syncookie(int type, sa_family_t family)
 	 */
 	err = bpf_map_update_elem(tmp_index_ovr_map, &index_zero,
 				  &tmp_index, BPF_ANY);
-	CHECK(err == -1, "update_elem(tmp_index_ovr_map, 0, 1)",
-	      "err:%d errno:%d\n", err, errno);
+	RET_IF(err == -1, "update_elem(tmp_index_ovr_map, 0, 1)",
+	       "err:%d errno:%d\n", err, errno);
 	do_test(type, family, &cmd, PASS);
 	err = bpf_map_lookup_elem(tmp_index_ovr_map, &index_zero,
 				  &tmp_index);
-	CHECK(err == -1 || tmp_index != -1,
-	      "lookup_elem(tmp_index_ovr_map)",
-	      "err:%d errno:%d tmp_index:%d\n",
-	      err, errno, tmp_index);
+	RET_IF(err == -1 || tmp_index != -1,
+	       "lookup_elem(tmp_index_ovr_map)",
+	       "err:%d errno:%d tmp_index:%d\n",
+	       err, errno, tmp_index);
 	disable_syncookie();
-	printf("OK\n");
 }
 
 static void test_pass_on_err(int type, sa_family_t family)
@@ -527,7 +525,6 @@ static void test_pass_on_err(int type, sa_family_t family)
 
 	expected_results[PASS_ERR_SK_SELECT_REUSEPORT] += 1;
 	do_test(type, family, &cmd, PASS_ERR_SK_SELECT_REUSEPORT);
-	printf("OK\n");
 }
 
 static void test_detach_bpf(int type, sa_family_t family)
@@ -541,43 +538,45 @@ static void test_detach_bpf(int type, sa_family_t family)
 
 	err = setsockopt(sk_fds[0], SOL_SOCKET, SO_DETACH_REUSEPORT_BPF,
 			 &optvalue, sizeof(optvalue));
-	CHECK(err == -1, "setsockopt(SO_DETACH_REUSEPORT_BPF)",
-	      "err:%d errno:%d\n", err, errno);
+	RET_IF(err == -1, "setsockopt(SO_DETACH_REUSEPORT_BPF)",
+	       "err:%d errno:%d\n", err, errno);
 
 	err = setsockopt(sk_fds[1], SOL_SOCKET, SO_DETACH_REUSEPORT_BPF,
 			 &optvalue, sizeof(optvalue));
-	CHECK(err == 0 || errno != ENOENT, "setsockopt(SO_DETACH_REUSEPORT_BPF)",
-	      "err:%d errno:%d\n", err, errno);
+	RET_IF(err == 0 || errno != ENOENT,
+	       "setsockopt(SO_DETACH_REUSEPORT_BPF)",
+	       "err:%d errno:%d\n", err, errno);
 
 	for (i = 0; i < NR_RESULTS; i++) {
 		err = bpf_map_lookup_elem(result_map, &i, &tmp);
-		CHECK(err == -1, "lookup_elem(result_map)",
-		      "i:%u err:%d errno:%d\n", i, err, errno);
+		RET_IF(err == -1, "lookup_elem(result_map)",
+		       "i:%u err:%d errno:%d\n", i, err, errno);
 		nr_run_before += tmp;
 	}
 
 	cli_fd = send_data(type, family, &cmd, sizeof(cmd), PASS);
+	if (cli_fd < 0)
+		return;
 	nev = epoll_wait(epfd, &ev, 1, 5);
-	CHECK(nev <= 0, "nev <= 0",
-	      "nev:%d expected:1 type:%d family:%d data:(0, 0)\n",
-	      nev,  type, family);
+	RET_IF(nev <= 0, "nev <= 0",
+	       "nev:%d expected:1 type:%d family:%d data:(0, 0)\n",
+	       nev,  type, family);
 
 	for (i = 0; i < NR_RESULTS; i++) {
 		err = bpf_map_lookup_elem(result_map, &i, &tmp);
-		CHECK(err == -1, "lookup_elem(result_map)",
-		      "i:%u err:%d errno:%d\n", i, err, errno);
+		RET_IF(err == -1, "lookup_elem(result_map)",
+		       "i:%u err:%d errno:%d\n", i, err, errno);
 		nr_run_after += tmp;
 	}
 
-	CHECK(nr_run_before != nr_run_after,
-	      "nr_run_before != nr_run_after",
-	      "nr_run_before:%u nr_run_after:%u\n",
-	      nr_run_before, nr_run_after);
+	RET_IF(nr_run_before != nr_run_after,
+	       "nr_run_before != nr_run_after",
+	       "nr_run_before:%u nr_run_after:%u\n",
+	       nr_run_before, nr_run_after);
 
-	printf("OK\n");
 	close(cli_fd);
 #else
-	printf("SKIP\n");
+	test__skip();
 #endif
 }
 
@@ -600,58 +599,58 @@ static void prepare_sk_fds(int type, sa_family_t family, bool inany)
 	 */
 	for (i = first; i >= 0; i--) {
 		sk_fds[i] = socket(family, type, 0);
-		CHECK(sk_fds[i] == -1, "socket()", "sk_fds[%d]:%d errno:%d\n",
-		      i, sk_fds[i], errno);
+		RET_IF(sk_fds[i] == -1, "socket()", "sk_fds[%d]:%d errno:%d\n",
+		       i, sk_fds[i], errno);
 		err = setsockopt(sk_fds[i], SOL_SOCKET, SO_REUSEPORT,
 				 &optval, sizeof(optval));
-		CHECK(err == -1, "setsockopt(SO_REUSEPORT)",
-		      "sk_fds[%d] err:%d errno:%d\n",
-		      i, err, errno);
+		RET_IF(err == -1, "setsockopt(SO_REUSEPORT)",
+		       "sk_fds[%d] err:%d errno:%d\n",
+		       i, err, errno);
 
 		if (i == first) {
 			err = setsockopt(sk_fds[i], SOL_SOCKET,
 					 SO_ATTACH_REUSEPORT_EBPF,
 					 &select_by_skb_data_prog,
 					 sizeof(select_by_skb_data_prog));
-			CHECK(err == -1, "setsockopt(SO_ATTACH_REUEPORT_EBPF)",
-			      "err:%d errno:%d\n", err, errno);
+			RET_IF(err == -1, "setsockopt(SO_ATTACH_REUEPORT_EBPF)",
+			       "err:%d errno:%d\n", err, errno);
 		}
 
 		err = bind(sk_fds[i], (struct sockaddr *)&srv_sa, addrlen);
-		CHECK(err == -1, "bind()", "sk_fds[%d] err:%d errno:%d\n",
-		      i, err, errno);
+		RET_IF(err == -1, "bind()", "sk_fds[%d] err:%d errno:%d\n",
+		       i, err, errno);
 
 		if (type == SOCK_STREAM) {
 			err = listen(sk_fds[i], 10);
-			CHECK(err == -1, "listen()",
-			      "sk_fds[%d] err:%d errno:%d\n",
-			      i, err, errno);
+			RET_IF(err == -1, "listen()",
+			       "sk_fds[%d] err:%d errno:%d\n",
+			       i, err, errno);
 		}
 
 		err = bpf_map_update_elem(reuseport_array, &i, &sk_fds[i],
 					  BPF_NOEXIST);
-		CHECK(err == -1, "update_elem(reuseport_array)",
-		      "sk_fds[%d] err:%d errno:%d\n", i, err, errno);
+		RET_IF(err == -1, "update_elem(reuseport_array)",
+		       "sk_fds[%d] err:%d errno:%d\n", i, err, errno);
 
 		if (i == first) {
 			socklen_t addrlen = sizeof(srv_sa);
 
 			err = getsockname(sk_fds[i], (struct sockaddr *)&srv_sa,
 					  &addrlen);
-			CHECK(err == -1, "getsockname()",
-			      "sk_fds[%d] err:%d errno:%d\n", i, err, errno);
+			RET_IF(err == -1, "getsockname()",
+			       "sk_fds[%d] err:%d errno:%d\n", i, err, errno);
 		}
 	}
 
 	epfd = epoll_create(1);
-	CHECK(epfd == -1, "epoll_create(1)",
-	      "epfd:%d errno:%d\n", epfd, errno);
+	RET_IF(epfd == -1, "epoll_create(1)",
+	       "epfd:%d errno:%d\n", epfd, errno);
 
 	ev.events = EPOLLIN;
 	for (i = 0; i < REUSEPORT_ARRAY_SIZE; i++) {
 		ev.data.u32 = i;
 		err = epoll_ctl(epfd, EPOLL_CTL_ADD, sk_fds[i], &ev);
-		CHECK(err, "epoll_ctl(EPOLL_CTL_ADD)", "sk_fds[%d]\n", i);
+		RET_IF(err, "epoll_ctl(EPOLL_CTL_ADD)", "sk_fds[%d]\n", i);
 	}
 }
 
@@ -663,8 +662,8 @@ static void setup_per_test(int type, sa_family_t family, bool inany,
 	prepare_sk_fds(type, family, inany);
 	err = bpf_map_update_elem(tmp_index_ovr_map, &index_zero, &ovr,
 				  BPF_ANY);
-	CHECK(err == -1, "update_elem(tmp_index_ovr_map, 0, -1)",
-	      "err:%d errno:%d\n", err, errno);
+	RET_IF(err == -1, "update_elem(tmp_index_ovr_map, 0, -1)",
+	       "err:%d errno:%d\n", err, errno);
 
 	/* Install reuseport_array to outer_map? */
 	if (no_inner_map)
@@ -672,8 +671,8 @@ static void setup_per_test(int type, sa_family_t family, bool inany,
 
 	err = bpf_map_update_elem(outer_map, &index_zero, &reuseport_array,
 				  BPF_ANY);
-	CHECK(err == -1, "update_elem(outer_map, 0, reuseport_array)",
-	      "err:%d errno:%d\n", err, errno);
+	RET_IF(err == -1, "update_elem(outer_map, 0, reuseport_array)",
+	       "err:%d errno:%d\n", err, errno);
 }
 
 static void cleanup_per_test(bool no_inner_map)
@@ -689,8 +688,8 @@ static void cleanup_per_test(bool no_inner_map)
 		return;
 
 	err = bpf_map_delete_elem(outer_map, &index_zero);
-	CHECK(err == -1, "delete_elem(outer_map)",
-	      "err:%d errno:%d\n", err, errno);
+	RET_IF(err == -1, "delete_elem(outer_map)",
+	       "err:%d errno:%d\n", err, errno);
 }
 
 static void cleanup(void)
@@ -729,7 +728,7 @@ static const char *sotype_str(int sotype)
 
 #define TEST_INIT(fn, ...) { fn, #fn, __VA_ARGS__ }
 
-static void test_config(int type, sa_family_t family, bool inany)
+static void test_config(int sotype, sa_family_t family, bool inany)
 {
 	const struct test {
 		void (*fn)(int sotype, sa_family_t family);
@@ -744,20 +743,21 @@ static void test_config(int type, sa_family_t family, bool inany)
 		TEST_INIT(test_pass_on_err),
 		TEST_INIT(test_detach_bpf),
 	};
+	char s[MAX_TEST_NAME];
 	const struct test *t;
 
-	printf("######## %s/%s %s ########\n",
-	       family_str(family), sotype_str(type),
-	       inany ? " INANY  " : "LOOPBACK");
-
 	for (t = tests; t < tests + ARRAY_SIZE(tests); t++) {
-		setup_per_test(type, family, inany, t->no_inner_map);
-		printf("%s: ", t->name);
-		t->fn(type, family);
+		snprintf(s, sizeof(s), "%s/%s %s %s",
+			 family_str(family), sotype_str(sotype),
+			 inany ? "INANY" : "LOOPBACK", t->name);
+
+		if (!test__start_subtest(s))
+			continue;
+
+		setup_per_test(sotype, family, inany, t->no_inner_map);
+		t->fn(sotype, family);
 		cleanup_per_test(t->no_inner_map);
 	}
-
-	printf("\n");
 }
 
 #define BIND_INANY true
@@ -782,10 +782,8 @@ static void test_all(void)
 		test_config(c->sotype, c->family, c->inany);
 }
 
-int main(int argc, const char **argv)
+void test_select_reuseport(void)
 {
-	int ret = EXIT_FAILURE;
-
 	if (create_maps())
 		goto out;
 	if (prepare_bpf_obj())
@@ -795,7 +793,6 @@ int main(int argc, const char **argv)
 	saved_tcp_syncookie = read_int_sysctl(TCP_SYNCOOKIE_SYSCTL);
 	if (saved_tcp_syncookie < 0 || saved_tcp_syncookie < 0)
 		goto out;
-	atexit(restore_sysctls);
 
 	if (enable_fastopen())
 		goto out;
@@ -803,9 +800,7 @@ int main(int argc, const char **argv)
 		goto out;
 
 	test_all();
-
-	ret = EXIT_SUCCESS;
 out:
 	cleanup();
-	return ret;
+	restore_sysctls();
 }
-- 
2.23.0


^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework
  2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
                   ` (9 preceding siblings ...)
  2019-12-12 10:22 ` [PATCH bpf-next 10/10] selftests/bpf: Switch reuseport tests for test_progs framework Jakub Sitnicki
@ 2019-12-13  5:36 ` Alexei Starovoitov
  2019-12-13 16:30   ` Martin Lau
  10 siblings, 1 reply; 14+ messages in thread
From: Alexei Starovoitov @ 2019-12-13  5:36 UTC (permalink / raw)
  To: Jakub Sitnicki; +Cc: bpf, Martin Lau, kernel-team

On Thu, Dec 12, 2019 at 11:22:49AM +0100, Jakub Sitnicki wrote:
> This change has been suggested by Martin Lau [0] during a review of a
> related patch set that extends reuseport tests [1].
> 
> Patches 1 & 2 address a warning due to unrecognized section name from
> libbpf when running reuseport tests. We don't want to carry this warning
> into test_progs.
> 
> Patches 3-8 massage the reuseport tests to ease the switch to test_progs
> framework. The intention here is to show the work. Happy to squash these,
> if needed.
> 
> Patches 9-10 do the actual move and conversion to test_progs.
> 
> Output from a test_progs run after changes pasted below.

Thank you for doing this conversion.
Looks great to me.

Martin,
could you please review ?


^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework
  2019-12-13  5:36 ` [PATCH bpf-next 00/10] " Alexei Starovoitov
@ 2019-12-13 16:30   ` Martin Lau
  2019-12-13 20:41     ` Alexei Starovoitov
  0 siblings, 1 reply; 14+ messages in thread
From: Martin Lau @ 2019-12-13 16:30 UTC (permalink / raw)
  To: Alexei Starovoitov; +Cc: Jakub Sitnicki, bpf, kernel-team

On Thu, Dec 12, 2019 at 09:36:36PM -0800, Alexei Starovoitov wrote:
> On Thu, Dec 12, 2019 at 11:22:49AM +0100, Jakub Sitnicki wrote:
> > This change has been suggested by Martin Lau [0] during a review of a
> > related patch set that extends reuseport tests [1].
> > 
> > Patches 1 & 2 address a warning due to unrecognized section name from
> > libbpf when running reuseport tests. We don't want to carry this warning
> > into test_progs.
> > 
> > Patches 3-8 massage the reuseport tests to ease the switch to test_progs
> > framework. The intention here is to show the work. Happy to squash these,
> > if needed.
> > 
> > Patches 9-10 do the actual move and conversion to test_progs.
> > 
> > Output from a test_progs run after changes pasted below.
> 
> Thank you for doing this conversion.
> Looks great to me.
> 
> Martin,
> could you please review ?
Looks awesome.  Appreciate for moving this to test_progs!

Acked-by: Martin KaFai Lau <kafai@fb.com>

^ permalink raw reply	[flat|nested] 14+ messages in thread

* Re: [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework
  2019-12-13 16:30   ` Martin Lau
@ 2019-12-13 20:41     ` Alexei Starovoitov
  0 siblings, 0 replies; 14+ messages in thread
From: Alexei Starovoitov @ 2019-12-13 20:41 UTC (permalink / raw)
  To: Martin Lau; +Cc: Jakub Sitnicki, bpf, kernel-team

On Fri, Dec 13, 2019 at 8:31 AM Martin Lau <kafai@fb.com> wrote:
>
> On Thu, Dec 12, 2019 at 09:36:36PM -0800, Alexei Starovoitov wrote:
> > On Thu, Dec 12, 2019 at 11:22:49AM +0100, Jakub Sitnicki wrote:
> > > This change has been suggested by Martin Lau [0] during a review of a
> > > related patch set that extends reuseport tests [1].
> > >
> > > Patches 1 & 2 address a warning due to unrecognized section name from
> > > libbpf when running reuseport tests. We don't want to carry this warning
> > > into test_progs.
> > >
> > > Patches 3-8 massage the reuseport tests to ease the switch to test_progs
> > > framework. The intention here is to show the work. Happy to squash these,
> > > if needed.
> > >
> > > Patches 9-10 do the actual move and conversion to test_progs.
> > >
> > > Output from a test_progs run after changes pasted below.
> >
> > Thank you for doing this conversion.
> > Looks great to me.
> >
> > Martin,
> > could you please review ?
> Looks awesome.  Appreciate for moving this to test_progs!
>
> Acked-by: Martin KaFai Lau <kafai@fb.com>

Applied. Thanks

^ permalink raw reply	[flat|nested] 14+ messages in thread

end of thread, other threads:[~2019-12-13 20:41 UTC | newest]

Thread overview: 14+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-12-12 10:22 [PATCH bpf-next 00/10] Switch reuseport tests for test_progs framework Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 01/10] libbpf: Recognize SK_REUSEPORT programs from section name Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 02/10] selftests/bpf: Let libbpf determine program type " Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 03/10] selftests/bpf: Use sa_family_t everywhere in reuseport tests Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 04/10] selftests/bpf: Add helpers for getting socket family & type name Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 05/10] selftests/bpf: Unroll the main loop in reuseport test Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 06/10] selftests/bpf: Run reuseport tests in a loop Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 07/10] selftests/bpf: Propagate errors during setup for reuseport tests Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 08/10] selftests/bpf: Pull up printing the test name into test runner Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 09/10] selftests/bpf: Move reuseport tests under prog_tests/ Jakub Sitnicki
2019-12-12 10:22 ` [PATCH bpf-next 10/10] selftests/bpf: Switch reuseport tests for test_progs framework Jakub Sitnicki
2019-12-13  5:36 ` [PATCH bpf-next 00/10] " Alexei Starovoitov
2019-12-13 16:30   ` Martin Lau
2019-12-13 20:41     ` Alexei Starovoitov

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).