[PATCH bpf-next 1/2] libbpf: handle nulled-out program in struct_ops correctly

[PATCH bpf-next 1/2] libbpf: handle nulled-out program in struct_ops correctly

[Andrii Nakryiko]

If struct_ops has one of program callbacks set declaratively and host
kernel is old and doesn't support this callback, libbpf will allow to
load such struct_ops as long as that callback was explicitly nulled-out
(presumably through skeleton). This is all working correctly, except we
won't reset corresponding program slot to NULL before bailing out, which
will lead to libbpf not detecting that BPF program has to be not
auto-loaded. Fix this by unconditionally resetting corresponding program
slot to NULL.

Fixes: c911fc61a7ce ("libbpf: Skip zeroed or null fields if not found in the kernel type.")
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
---
 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 97eb6e5dd7c8..898d5d34ecea 100644
--- a/tools/lib/bpf/libbpf.c
+++ b/tools/lib/bpf/libbpf.c
@@ -1148,6 +1148,7 @@ static int bpf_map__init_kern_struct_ops(struct bpf_map *map)
 			 * presented in the kernel BTF.
 			 */
 			if (libbpf_is_mem_zeroed(mdata, msize)) {
+				st_ops->progs[i] = NULL;
 				pr_info("struct_ops %s: member %s not found in kernel, skipping it as it's set to zero\n",
 					map->name, mname);
 				continue;
-- 
2.43.0

[PATCH bpf-next 2/2] selftests/bpf: validate nulled-out struct_ops program is handled properly

[Andrii Nakryiko]

Add a selftests validating that it's possible to have some struct_ops
callback set declaratively, then disable it (by setting to NULL)
programmatically. Libbpf should detect that such program should be
loaded, even if host kernel doesn't have type information for it.

Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
---
 .../bpf/prog_tests/test_struct_ops_module.c    | 18 ++++++++++++++++--
 .../selftests/bpf/progs/struct_ops_module.c    |  7 +++++++
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/bpf/prog_tests/test_struct_ops_module.c b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_module.c
index 7cf2b9ddd3e1..bd39586abd5a 100644
--- a/tools/testing/selftests/bpf/prog_tests/test_struct_ops_module.c
+++ b/tools/testing/selftests/bpf/prog_tests/test_struct_ops_module.c
@@ -66,6 +66,7 @@ static void test_struct_ops_load(void)
 	 * auto-loading, or it will fail to load.
 	 */
 	bpf_program__set_autoload(skel->progs.test_2, false);
+	bpf_map__set_autocreate(skel->maps.testmod_zeroed, false);
 
 	err = struct_ops_module__load(skel);
 	if (!ASSERT_OK(err, "struct_ops_module_load"))
@@ -103,6 +104,10 @@ static void test_struct_ops_not_zeroed(void)
 	if (!ASSERT_OK_PTR(skel, "struct_ops_module_open"))
 		return;
 
+	skel->struct_ops.testmod_zeroed->zeroed = 0;
+	/* zeroed_op prog should be not loaded automatically now */
+	skel->struct_ops.testmod_zeroed->zeroed_op = NULL;
+
 	err = struct_ops_module__load(skel);
 	ASSERT_OK(err, "struct_ops_module_load");
 
@@ -118,6 +123,7 @@ static void test_struct_ops_not_zeroed(void)
 	 * value of "zeroed" is non-zero.
 	 */
 	skel->struct_ops.testmod_zeroed->zeroed = 0xdeadbeef;
+	skel->struct_ops.testmod_zeroed->zeroed_op = NULL;
 	err = struct_ops_module__load(skel);
 	ASSERT_ERR(err, "struct_ops_module_load_not_zeroed");
 
@@ -148,15 +154,23 @@ static void test_struct_ops_incompatible(void)
 {
 	struct struct_ops_module *skel;
 	struct bpf_link *link;
+	int err;
 
-	skel = struct_ops_module__open_and_load();
-	if (!ASSERT_OK_PTR(skel, "open_and_load"))
+	skel = struct_ops_module__open();
+	if (!ASSERT_OK_PTR(skel, "struct_ops_module_open"))
 		return;
 
+	bpf_map__set_autocreate(skel->maps.testmod_zeroed, false);
+
+	err = struct_ops_module__load(skel);
+	if (!ASSERT_OK(err, "skel_load"))
+		goto cleanup;
+
 	link = bpf_map__attach_struct_ops(skel->maps.testmod_incompatible);
 	if (ASSERT_OK_PTR(link, "attach_struct_ops"))
 		bpf_link__destroy(link);
 
+cleanup:
 	struct_ops_module__destroy(skel);
 }
 
diff --git a/tools/testing/selftests/bpf/progs/struct_ops_module.c b/tools/testing/selftests/bpf/progs/struct_ops_module.c
index 63b065dae002..40109be2b3ae 100644
--- a/tools/testing/selftests/bpf/progs/struct_ops_module.c
+++ b/tools/testing/selftests/bpf/progs/struct_ops_module.c
@@ -63,10 +63,17 @@ struct bpf_testmod_ops___zeroed {
 	int zeroed;
 };
 
+SEC("?struct_ops/test_3")
+int BPF_PROG(zeroed_op)
+{
+	return 1;
+}
+
 SEC(".struct_ops.link")
 struct bpf_testmod_ops___zeroed testmod_zeroed = {
 	.test_1 = (void *)test_1,
 	.test_2 = (void *)test_2_v2,
+	.zeroed_op = (void *)zeroed_op,
 };
 
 struct bpf_testmod_ops___incompatible {
-- 
2.43.0

Re: [PATCH bpf-next 2/2] selftests/bpf: validate nulled-out struct_ops program is handled properly

[Martin KaFai Lau]

On 4/27/24 8:09 PM, Andrii Nakryiko wrote:
> Add a selftests validating that it's possible to have some struct_ops
> callback set declaratively, then disable it (by setting to NULL)
> programmatically. Libbpf should detect that such program should be

such program should be /not/ loaded ?

> loaded, even if host kernel doesn't have type information for it.
> 

> @@ -103,6 +104,10 @@ static void test_struct_ops_not_zeroed(void)
>   	if (!ASSERT_OK_PTR(skel, "struct_ops_module_open"))
>   		return;
>   
> +	skel->struct_ops.testmod_zeroed->zeroed = 0;
> +	/* zeroed_op prog should be not loaded automatically now */
> +	skel->struct_ops.testmod_zeroed->zeroed_op = NULL;
> +
>   	err = struct_ops_module__load(skel);
>   	ASSERT_OK(err, "struct_ops_module_load");
>   
> @@ -118,6 +123,7 @@ static void test_struct_ops_not_zeroed(void)
>   	 * value of "zeroed" is non-zero.
>   	 */
>   	skel->struct_ops.testmod_zeroed->zeroed = 0xdeadbeef;
> +	skel->struct_ops.testmod_zeroed->zeroed_op = NULL;
>   	err = struct_ops_module__load(skel);
>   	ASSERT_ERR(err, "struct_ops_module_load_not_zeroed");
>   

Re: [PATCH bpf-next 2/2] selftests/bpf: validate nulled-out struct_ops program is handled properly

[Andrii Nakryiko]

On Mon, Apr 29, 2024 at 2:29 PM Martin KaFai Lau <martin.lau@linux.dev> wrote:
>
> On 4/27/24 8:09 PM, Andrii Nakryiko wrote:
> > Add a selftests validating that it's possible to have some struct_ops
> > callback set declaratively, then disable it (by setting to NULL)
> > programmatically. Libbpf should detect that such program should be
>
> such program should be /not/ loaded ?

yep, can you fix it up while applying or should I send a new revision?

>
> > loaded, even if host kernel doesn't have type information for it.
> >
>
> > @@ -103,6 +104,10 @@ static void test_struct_ops_not_zeroed(void)
> >       if (!ASSERT_OK_PTR(skel, "struct_ops_module_open"))
> >               return;
> >
> > +     skel->struct_ops.testmod_zeroed->zeroed = 0;
> > +     /* zeroed_op prog should be not loaded automatically now */
> > +     skel->struct_ops.testmod_zeroed->zeroed_op = NULL;
> > +
> >       err = struct_ops_module__load(skel);
> >       ASSERT_OK(err, "struct_ops_module_load");
> >
> > @@ -118,6 +123,7 @@ static void test_struct_ops_not_zeroed(void)
> >        * value of "zeroed" is non-zero.
> >        */
> >       skel->struct_ops.testmod_zeroed->zeroed = 0xdeadbeef;
> > +     skel->struct_ops.testmod_zeroed->zeroed_op = NULL;
> >       err = struct_ops_module__load(skel);
> >       ASSERT_ERR(err, "struct_ops_module_load_not_zeroed");
> >
>

Re: [PATCH bpf-next 2/2] selftests/bpf: validate nulled-out struct_ops program is handled properly

[Martin KaFai Lau]

On 4/29/24 3:35 PM, Andrii Nakryiko wrote:
> On Mon, Apr 29, 2024 at 2:29 PM Martin KaFai Lau <martin.lau@linux.dev> wrote:
>>
>> On 4/27/24 8:09 PM, Andrii Nakryiko wrote:
>>> Add a selftests validating that it's possible to have some struct_ops
>>> callback set declaratively, then disable it (by setting to NULL)
>>> programmatically. Libbpf should detect that such program should be
>>
>> such program should be /not/ loaded ?
> 
> yep, can you fix it up while applying or should I send a new revision?

I will take care of it. No need to respin.

Re: [PATCH bpf-next 1/2] libbpf: handle nulled-out program in struct_ops correctly

[patchwork-bot+netdevbpf]

Hello:

This series was applied to bpf/bpf-next.git (master)
by Martin KaFai Lau <martin.lau@kernel.org>:

On Sat, 27 Apr 2024 20:09:53 -0700 you wrote:
> If struct_ops has one of program callbacks set declaratively and host
> kernel is old and doesn't support this callback, libbpf will allow to
> load such struct_ops as long as that callback was explicitly nulled-out
> (presumably through skeleton). This is all working correctly, except we
> won't reset corresponding program slot to NULL before bailing out, which
> will lead to libbpf not detecting that BPF program has to be not
> auto-loaded. Fix this by unconditionally resetting corresponding program
> slot to NULL.
> 
> [...]

Here is the summary with links:
  - [bpf-next,1/2] libbpf: handle nulled-out program in struct_ops correctly
    https://git.kernel.org/bpf/bpf-next/c/f973fccd43d3
  - [bpf-next,2/2] selftests/bpf: validate nulled-out struct_ops program is handled properly
    https://git.kernel.org/bpf/bpf-next/c/1bba3b3d373d

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html