[PATCH bpf-next] selftest/bpf: Switch recursion test to use htab_map_delete_elem

Subsystems: bpf [general] (safe dynamic programs and tools), bpf [selftests] (test runners & infrastructure), kernel selftest framework, the rest

STALE1777d

3 messages, 3 authors, 2021-10-06 · open the first message on its own page

[PATCH bpf-next] selftest/bpf: Switch recursion test to use htab_map_delete_elem

From: Jiri Olsa <hidden>
Date: 2021-10-03 16:49:31 

Currently the recursion test is hooking __htab_map_lookup_elem
function, which is invoked both from bpf_prog and bpf syscall.

But in our kernel build, the __htab_map_lookup_elem gets inlined
within the htab_map_lookup_elem, so it's not trigered and the
test fails.

Fixing this by using htab_map_delete_elem, which is not inlined
for bpf_prog calls (like htab_map_lookup_elem is) and is used
directly as pointer for map_delete_elem, so it won't disappear
by inlining.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 tools/testing/selftests/bpf/prog_tests/recursion.c | 10 +++++-----
 tools/testing/selftests/bpf/progs/recursion.c      |  9 +++------
 2 files changed, 8 insertions(+), 11 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/recursion.c b/tools/testing/selftests/bpf/prog_tests/recursion.c
index 0e378d63fe18..f3af2627b599 100644
--- a/tools/testing/selftests/bpf/prog_tests/recursion.c
+++ b/tools/testing/selftests/bpf/prog_tests/recursion.c
@@ -20,18 +20,18 @@ void test_recursion(void)
 		goto out;
 
 	ASSERT_EQ(skel->bss->pass1, 0, "pass1 == 0");
-	bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash1), &key, 0);
+	bpf_map_delete_elem(bpf_map__fd(skel->maps.hash1), &key);
 	ASSERT_EQ(skel->bss->pass1, 1, "pass1 == 1");
-	bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash1), &key, 0);
+	bpf_map_delete_elem(bpf_map__fd(skel->maps.hash1), &key);
 	ASSERT_EQ(skel->bss->pass1, 2, "pass1 == 2");
 
 	ASSERT_EQ(skel->bss->pass2, 0, "pass2 == 0");
-	bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash2), &key, 0);
+	bpf_map_delete_elem(bpf_map__fd(skel->maps.hash2), &key);
 	ASSERT_EQ(skel->bss->pass2, 1, "pass2 == 1");
-	bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash2), &key, 0);
+	bpf_map_delete_elem(bpf_map__fd(skel->maps.hash2), &key);
 	ASSERT_EQ(skel->bss->pass2, 2, "pass2 == 2");
 
-	err = bpf_obj_get_info_by_fd(bpf_program__fd(skel->progs.on_lookup),
+	err = bpf_obj_get_info_by_fd(bpf_program__fd(skel->progs.on_delete),
 				     &prog_info, &prog_info_len);
 	if (!ASSERT_OK(err, "get_prog_info"))
 		goto out;
diff --git a/tools/testing/selftests/bpf/progs/recursion.c b/tools/testing/selftests/bpf/progs/recursion.c
index 49f679375b9d..3c2423bb19e2 100644
--- a/tools/testing/selftests/bpf/progs/recursion.c
+++ b/tools/testing/selftests/bpf/progs/recursion.c
@@ -24,8 +24,8 @@ struct {
 int pass1 = 0;
 int pass2 = 0;
 
-SEC("fentry/__htab_map_lookup_elem")
-int BPF_PROG(on_lookup, struct bpf_map *map)
+SEC("fentry/htab_map_delete_elem")
+int BPF_PROG(on_delete, struct bpf_map *map)
 {
 	int key = 0;
 
@@ -35,10 +35,7 @@ int BPF_PROG(on_lookup, struct bpf_map *map)
 	}
 	if (map == (void *)&hash2) {
 		pass2++;
-		/* htab_map_gen_lookup() will inline below call
-		 * into direct call to __htab_map_lookup_elem()
-		 */
-		bpf_map_lookup_elem(&hash2, &key);
+		bpf_map_delete_elem(&hash2, &key);
 		return 0;
 	}
 
-- 
2.31.1

Re: [PATCH bpf-next] selftest/bpf: Switch recursion test to use htab_map_delete_elem

From: Song Liu <song@kernel.org>
Date: 2021-10-04 17:38:47 

On Sun, Oct 3, 2021 at 9:58 AM Jiri Olsa [off-list ref] wrote:
Currently the recursion test is hooking __htab_map_lookup_elem
function, which is invoked both from bpf_prog and bpf syscall.

But in our kernel build, the __htab_map_lookup_elem gets inlined
within the htab_map_lookup_elem, so it's not trigered and the
test fails.

Fixing this by using htab_map_delete_elem, which is not inlined
for bpf_prog calls (like htab_map_lookup_elem is) and is used
directly as pointer for map_delete_elem, so it won't disappear
by inlining.

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Acked-by: Song Liu <redacted>
quoted hunk
---
 tools/testing/selftests/bpf/prog_tests/recursion.c | 10 +++++-----
 tools/testing/selftests/bpf/progs/recursion.c      |  9 +++------
 2 files changed, 8 insertions(+), 11 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/recursion.c b/tools/testing/selftests/bpf/prog_tests/recursion.c
index 0e378d63fe18..f3af2627b599 100644
--- a/tools/testing/selftests/bpf/prog_tests/recursion.c
+++ b/tools/testing/selftests/bpf/prog_tests/recursion.c
@@ -20,18 +20,18 @@ void test_recursion(void)
                goto out;

        ASSERT_EQ(skel->bss->pass1, 0, "pass1 == 0");
-       bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash1), &key, 0);
+       bpf_map_delete_elem(bpf_map__fd(skel->maps.hash1), &key);
        ASSERT_EQ(skel->bss->pass1, 1, "pass1 == 1");
-       bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash1), &key, 0);
+       bpf_map_delete_elem(bpf_map__fd(skel->maps.hash1), &key);
        ASSERT_EQ(skel->bss->pass1, 2, "pass1 == 2");

        ASSERT_EQ(skel->bss->pass2, 0, "pass2 == 0");
-       bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash2), &key, 0);
+       bpf_map_delete_elem(bpf_map__fd(skel->maps.hash2), &key);
        ASSERT_EQ(skel->bss->pass2, 1, "pass2 == 1");
-       bpf_map_lookup_elem(bpf_map__fd(skel->maps.hash2), &key, 0);
+       bpf_map_delete_elem(bpf_map__fd(skel->maps.hash2), &key);
        ASSERT_EQ(skel->bss->pass2, 2, "pass2 == 2");

-       err = bpf_obj_get_info_by_fd(bpf_program__fd(skel->progs.on_lookup),
+       err = bpf_obj_get_info_by_fd(bpf_program__fd(skel->progs.on_delete),
                                     &prog_info, &prog_info_len);
        if (!ASSERT_OK(err, "get_prog_info"))
                goto out;
diff --git a/tools/testing/selftests/bpf/progs/recursion.c b/tools/testing/selftests/bpf/progs/recursion.c
index 49f679375b9d..3c2423bb19e2 100644
--- a/tools/testing/selftests/bpf/progs/recursion.c
+++ b/tools/testing/selftests/bpf/progs/recursion.c
@@ -24,8 +24,8 @@ struct {
 int pass1 = 0;
 int pass2 = 0;

-SEC("fentry/__htab_map_lookup_elem")
-int BPF_PROG(on_lookup, struct bpf_map *map)
+SEC("fentry/htab_map_delete_elem")
+int BPF_PROG(on_delete, struct bpf_map *map)
 {
        int key = 0;

@@ -35,10 +35,7 @@ int BPF_PROG(on_lookup, struct bpf_map *map)
        }
        if (map == (void *)&hash2) {
                pass2++;
-               /* htab_map_gen_lookup() will inline below call
-                * into direct call to __htab_map_lookup_elem()
-                */
-               bpf_map_lookup_elem(&hash2, &key);
+               bpf_map_delete_elem(&hash2, &key);
                return 0;
        }

--
2.31.1

Re: [PATCH bpf-next] selftest/bpf: Switch recursion test to use htab_map_delete_elem

From: patchwork-bot+netdevbpf@kernel.org
Date: 2021-10-06 17:20:10 

Hello:

This patch was applied to bpf/bpf-next.git (refs/heads/master):

On Sun, 3 Oct 2021 18:49:25 +0200 you wrote:
Currently the recursion test is hooking __htab_map_lookup_elem
function, which is invoked both from bpf_prog and bpf syscall.

But in our kernel build, the __htab_map_lookup_elem gets inlined
within the htab_map_lookup_elem, so it's not trigered and the
test fails.

[...]
Here is the summary with links:
  - [bpf-next] selftest/bpf: Switch recursion test to use htab_map_delete_elem
    https://git.kernel.org/bpf/bpf-next/c/fba27f590fd3

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help