From: Yonghong Song <yonghong.song@linux.dev>
To: Dave Marchevsky <davemarchevsky@fb.com>, bpf@vger.kernel.org
Cc: Alexei Starovoitov <ast@kernel.org>,
Daniel Borkmann <daniel@iogearbox.net>,
Andrii Nakryiko <andrii@kernel.org>,
Martin KaFai Lau <martin.lau@kernel.org>,
Kernel Team <kernel-team@fb.com>
Subject: Re: [PATCH v1 bpf-next 4/4] selftests/bpf: Add tests exercising aggregate type BTF field search
Date: Mon, 30 Oct 2023 14:10:25 -0700 [thread overview]
Message-ID: <45000107-b119-46d5-aa01-c3f08d0a1921@linux.dev> (raw)
In-Reply-To: <20231023220030.2556229-5-davemarchevsky@fb.com>
On 10/23/23 3:00 PM, Dave Marchevsky wrote:
> The newly-added test file attempts to kptr_xchg a prog_test_ref_kfunc
> kptr into a kptr field in a variety of nested aggregate types. If the
> verifier recognizes that there's a kptr field where we're trying to
> kptr_xchg, then the aggregate type digging logic works as expected.
>
> Some of the refactoring changes in this series are tested as well.
> Specifically:
> * BTF_FIELDS_MAX is now higher and represents the max size of the
> growable array. Confirm that btf_parse_fields fails for a type which
> contains too many fields.
> * If we've already seen BTF_FIELDS_MAX fields, we should continue
> looking for fields and fail if we find another one, otherwise the
> search should succeed and return BTF_FIELDS_MAX btf_field_infos.
> Confirm that this edge case works as expected.
>
> Signed-off-by: Dave Marchevsky <davemarchevsky@fb.com>
> ---
> .../selftests/bpf/prog_tests/array_kptr.c | 12 ++
> .../testing/selftests/bpf/progs/array_kptr.c | 179 ++++++++++++++++++
> 2 files changed, 191 insertions(+)
> create mode 100644 tools/testing/selftests/bpf/prog_tests/array_kptr.c
> create mode 100644 tools/testing/selftests/bpf/progs/array_kptr.c
>
> diff --git a/tools/testing/selftests/bpf/prog_tests/array_kptr.c b/tools/testing/selftests/bpf/prog_tests/array_kptr.c
> new file mode 100644
> index 000000000000..9d088520bdfe
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/prog_tests/array_kptr.c
> @@ -0,0 +1,12 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/* Copyright (c) 2023 Meta Platforms, Inc. and affiliates. */
> +
> +#include <test_progs.h>
> +
> +#include "array_kptr.skel.h"
> +
> +void test_array_kptr(void)
> +{
> + if (env.has_testmod)
> + RUN_TESTS(array_kptr);
> +}
> diff --git a/tools/testing/selftests/bpf/progs/array_kptr.c b/tools/testing/selftests/bpf/progs/array_kptr.c
> new file mode 100644
> index 000000000000..f34872e74024
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/progs/array_kptr.c
> @@ -0,0 +1,179 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/* Copyright (c) 2023 Meta Platforms, Inc. and affiliates. */
> +
> +#include <vmlinux.h>
> +#include <bpf/bpf_tracing.h>
> +#include <bpf/bpf_helpers.h>
> +#include "../bpf_testmod/bpf_testmod_kfunc.h"
> +#include "bpf_misc.h"
> +
> +struct val {
> + int d;
> + struct prog_test_ref_kfunc __kptr *ref_ptr;
> +};
> +
> +struct val2 {
> + char c;
> + struct val v;
> +};
> +
> +struct val_holder {
> + int e;
> + struct val2 first[2];
> + int f;
> + struct val second[2];
> +};
> +
> +struct array_map {
> + __uint(type, BPF_MAP_TYPE_ARRAY);
> + __type(key, int);
> + __type(value, struct val);
> + __uint(max_entries, 10);
> +} array_map SEC(".maps");
> +
> +struct array_map2 {
> + __uint(type, BPF_MAP_TYPE_ARRAY);
> + __type(key, int);
> + __type(value, struct val2);
> + __uint(max_entries, 10);
> +} array_map2 SEC(".maps");
> +
> +__hidden struct val array[25];
> +__hidden struct val double_array[5][5];
> +__hidden struct val_holder double_holder_array[2][2];
> +
> +/* Some tests need their own section to force separate bss arraymap,
> + * otherwise above arrays wouldn't have btf_field_info either
> + */
> +#define private(name) SEC(".bss." #name) __hidden __attribute__((aligned(8)))
> +private(A) struct val array_too_big[300];
> +
> +private(B) struct val exactly_max_fields[256];
> +private(B) int ints[50];
> +
> +SEC("tc")
> +__success __retval(0)
> +int test_arraymap(void *ctx)
> +{
> + struct prog_test_ref_kfunc *p;
> + unsigned long dummy = 0;
> + struct val *v;
> + int idx = 0;
> +
> + v = bpf_map_lookup_elem(&array_map, &idx);
> + if (!v)
> + return 1;
> +
> + p = bpf_kfunc_call_test_acquire(&dummy);
> + if (!p)
> + return 2;
> +
> + p = bpf_kptr_xchg(&v->ref_ptr, p);
> + if (p) {
> + bpf_kfunc_call_test_release(p);
> + return 3;
> + }
> +
> + return 0;
> +}
> +
> ...
> +
> +SEC("tc")
> +__failure __msg("map '.bss.A' has no valid kptr")
The .bss.A might have valid kptr.
To reflect realiaty, maybe error message can be
'has too many special fields'?
> +int test_array_fail__too_big(void *ctx)
> +{
> + /* array_too_big's btf_record parsing will fail due to the
> + * number of btf_field_infos being > BTF_FIELDS_MAX
> + */
> + return test_array_xchg(&array_too_big[50]);
> +}
> +
> +char _license[] SEC("license") = "GPL";
next prev parent reply other threads:[~2023-10-30 21:10 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-23 22:00 [PATCH v1 bpf-next 0/4] Descend into struct, array types when searching for fields Dave Marchevsky
2023-10-23 22:00 ` [PATCH v1 bpf-next 1/4] bpf: Fix btf_get_field_type to fail for multiple bpf_refcount fields Dave Marchevsky
2023-10-30 17:56 ` Yonghong Song
2023-11-01 21:56 ` Andrii Nakryiko
2023-10-23 22:00 ` [PATCH v1 bpf-next 2/4] bpf: Refactor btf_find_field with btf_field_info_search Dave Marchevsky
2023-10-28 14:52 ` Jiri Olsa
2023-10-30 19:31 ` Yonghong Song
2023-10-30 19:56 ` Yonghong Song
2023-11-01 21:56 ` Andrii Nakryiko
2023-10-23 22:00 ` [PATCH v1 bpf-next 3/4] btf: Descend into structs and arrays during special field search Dave Marchevsky
2023-10-26 1:24 ` kernel test robot
2023-10-30 12:56 ` Jiri Olsa
2023-10-30 20:56 ` Yonghong Song
2023-11-01 21:56 ` Andrii Nakryiko
2023-10-23 22:00 ` [PATCH v1 bpf-next 4/4] selftests/bpf: Add tests exercising aggregate type BTF " Dave Marchevsky
2023-10-23 23:32 ` kernel test robot
2023-10-30 21:10 ` Yonghong Song [this message]
2023-11-01 21:56 ` Andrii Nakryiko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=45000107-b119-46d5-aa01-c3f08d0a1921@linux.dev \
--to=yonghong.song@linux.dev \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=davemarchevsky@fb.com \
--cc=kernel-team@fb.com \
--cc=martin.lau@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox