Re: [PATCH bpf-next v11 2/4] selftests/bpf: Add test for recursive attachment of tracing progs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Dec 22, 2023 at 7:12 AM Dmitrii Dolgov <9erthalion6@xxxxxxxxx> wrote:
>
> Verify the fact that only one fentry prog could be attached to another
> fentry, building up an attachment chain of limited size. Use existing
> bpf_testmod as a start of the chain.
>
> Acked-by: Jiri Olsa <olsajiri@xxxxxxxxx>
> Signed-off-by: Dmitrii Dolgov <9erthalion6@xxxxxxxxx>

Acked-by: Song Liu <song@xxxxxxxxxx>

With a few nits below.

>
> diff --git a/tools/testing/selftests/bpf/prog_tests/recursive_attach.c b/tools/testing/selftests/bpf/prog_tests/recursive_attach.c
> new file mode 100644
> index 000000000000..e9e576de6723
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/prog_tests/recursive_attach.c
> @@ -0,0 +1,111 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/* Copyright (c) 2023 Red Hat, Inc. */
> +#include <test_progs.h>
> +#include "fentry_recursive.skel.h"
> +#include "fentry_recursive_target.skel.h"
> +#include <bpf/btf.h>
> +#include "bpf/libbpf_internal.h"
> +
> +/*
> + * Test recursive attachment of tracing progs with more than one nesting level
> + * is not possible. Create a chain of attachment, verify that the last prog
> + * will fail. Depending on the arguments, following cases are tested:
> + *
> + * - Recursive loading of tracing progs, without attaching (attach = false,
> + *   detach = false). The chain looks like this:
> + *       load target
> + *       load fentry1 -> target
> + *       load fentry2 -> fentry1 (fail)
> + *
> + * - Recursive attach of tracing progs (attach = true, detach = false). The
> + *   chain looks like this:
> + *       load target
> + *       load fentry1 -> target
> + *       attach fentry1 -> target
> + *       load fentry2 -> fentry1 (fail)
> + *
> + * - Recursive attach and detach of tracing progs (attach = true, detach =
> + *   true). This validates that attach_tracing_prog flag will be set throughout
> + *   the whole lifecycle of an fentry prog, independently from whether it's
> + *   detached. The chain looks like this:
> + *       load target
> + *       load fentry1 -> target
> + *       attach fentry1 -> target
> + *       detach fentry1
> + *       load fentry2 -> fentry1 (fail)
> + */
> +static void test_recursive_fentry_chain(bool attach, bool detach)
> +{
> +       struct fentry_recursive_target *target_skel = NULL;
> +       struct fentry_recursive *tracing_chain[2] = {};
> +       struct bpf_program *prog;
> +       int prev_fd, err;
> +
> +       target_skel = fentry_recursive_target__open_and_load();
> +       if (!ASSERT_OK_PTR(target_skel, "fentry_recursive_target__open_and_load"))
> +               goto close_prog;

nit: This is not wrong, but we can just return here.

> +
> +       /* Create an attachment chain with two fentry progs */
> +       for (int i = 0; i < 2; i++) {
> +               tracing_chain[i] = fentry_recursive__open();
> +               if (!ASSERT_OK_PTR(tracing_chain[i], "fentry_recursive__open"))
> +                       goto close_prog;
> +
> +               /*
> +                * The first prog in the chain is going to be attached to the target
> +                * fentry program, the second one to the previous in the chain.
> +                */
> +               prog = tracing_chain[i]->progs.recursive_attach;
> +               if (i == 0) {
> +                       prev_fd = bpf_program__fd(target_skel->progs.test1);
> +                       err = bpf_program__set_attach_target(prog, prev_fd, "test1");
> +               } else {
> +                       prev_fd = bpf_program__fd(tracing_chain[i-1]->progs.recursive_attach);
> +                       err = bpf_program__set_attach_target(prog, prev_fd, "recursive_attach");
> +               }

nit: I don't really like these "if (i == 0)" cases. But it is not too bad.

> +
> +               if (!ASSERT_OK(err, "bpf_program__set_attach_target"))
> +                       goto close_prog;
> +
> +               err = fentry_recursive__load(tracing_chain[i]);
> +               /* The first attach should succeed, the second fail */
> +               if (i == 0) {
> +                       if (!ASSERT_OK(err, "fentry_recursive__load"))
> +                               goto close_prog;
> +
> +                       if (attach) {
> +                               err = fentry_recursive__attach(tracing_chain[i]);
> +                               if (!ASSERT_OK(err, "fentry_recursive__attach"))
> +                                       goto close_prog;
> +                       }
> +
> +                       if (detach) {
> +                               /*
> +                                * Flag attach_tracing_prog should still be set, preventing
> +                                * attachment of the following prog.
> +                                */
> +                               fentry_recursive__detach(tracing_chain[i]);
> +                       }
> +               } else {
> +                       if (!ASSERT_ERR(err, "fentry_recursive__load"))
> +                               goto close_prog;
> +               }
> +       }
> +
> +close_prog:
> +       fentry_recursive_target__destroy(target_skel);
> +       for (int i = 0; i < 2; i++) {
> +               if (tracing_chain[i])

nit: This NULL check is not necessary.

> +                       fentry_recursive__destroy(tracing_chain[i]);
> +       }
> +}
> +
> +void test_recursive_fentry(void)
> +{
> +       if (test__start_subtest("attach"))
> +               test_recursive_fentry_chain(true, false);
> +       if (test__start_subtest("load"))
> +               test_recursive_fentry_chain(false, false);
> +       if (test__start_subtest("detach"))
> +               test_recursive_fentry_chain(true, true);
> +}
[...]





[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux