Message-ID: <CAEf4BzbQv4D65kuYRr+i8aqGqUY+YT7oKGJNNBxSUUBsj+Zhrw@mail.gmail.com>
Date: Tue, 11 Feb 2025 14:24:01 -0800
From: Andrii Nakryiko <andrii.nakryiko@...il.com>
To: Tao Chen <chen.dylane@...il.com>
Cc: ast@...nel.org, daniel@...earbox.net, andrii@...nel.org, eddyz87@...il.com, 
	haoluo@...gle.com, jolsa@...nel.org, qmo@...nel.org, bpf@...r.kernel.org, 
	linux-kernel@...r.kernel.org, Tao Chen <dylane.chen@...iglobal.com>
Subject: Re: [PATCH bpf-next v6 3/4] libbpf: Add libbpf_probe_bpf_kfunc API

On Tue, Feb 11, 2025 at 3:19 AM Tao Chen <chen.dylane@...il.com> wrote:
>
> Similarly to libbpf_probe_bpf_helper, the libbpf_probe_bpf_kfunc
> used to test the availability of the different eBPF kfuncs on the
> current system.
>
> Cc: Tao Chen <dylane.chen@...iglobal.com>
> Signed-off-by: Tao Chen <chen.dylane@...il.com>
> ---
>  tools/lib/bpf/libbpf.h        | 19 +++++++++++++-
>  tools/lib/bpf/libbpf.map      |  1 +
>  tools/lib/bpf/libbpf_probes.c | 48 +++++++++++++++++++++++++++++++++++
>  3 files changed, 67 insertions(+), 1 deletion(-)
>
> diff --git a/tools/lib/bpf/libbpf.h b/tools/lib/bpf/libbpf.h
> index 3020ee45303a..e796e38cf255 100644
> --- a/tools/lib/bpf/libbpf.h
> +++ b/tools/lib/bpf/libbpf.h
> @@ -1680,7 +1680,24 @@ LIBBPF_API int libbpf_probe_bpf_map_type(enum bpf_map_type map_type, const void
>   */
>  LIBBPF_API int libbpf_probe_bpf_helper(enum bpf_prog_type prog_type,
>                                        enum bpf_func_id helper_id, const void *opts);
> -
> +/**
> + * @brief **libbpf_probe_bpf_kfunc()** detects if host kernel supports the
> + * use of a given BPF kfunc from specified BPF program type.
> + * @param prog_type BPF program type used to check the support of BPF kfunc
> + * @param kfunc_id The btf ID of BPF kfunc to check support for
> + * @param btf_fd The module BTF FD, if kfunc is defined in kernel module,
> + * btf_fd is used to point to module's BTF, which is >= 0, and -1 means kfunc
> + * defined in vmlinux.
> + * @param opts reserved for future extensibility, should be NULL
> + * @return 1, if given combination of program type and kfunc is supported; 0,
> + * if the combination is not supported; negative error code if feature
> + * detection for provided input arguments failed or can't be performed
> + *
> + * Make sure the process has required set of CAP_* permissions (or runs as
> + * root) when performing feature checking.
> + */
> +LIBBPF_API int libbpf_probe_bpf_kfunc(enum bpf_prog_type prog_type,
> +                                     int kfunc_id, int btf_fd, const void *opts);
>  /**
>   * @brief **libbpf_num_possible_cpus()** is a helper function to get the
>   * number of possible CPUs that the host kernel supports and expects.
> diff --git a/tools/lib/bpf/libbpf.map b/tools/lib/bpf/libbpf.map
> index b5a838de6f47..3bbfe13aeb6a 100644
> --- a/tools/lib/bpf/libbpf.map
> +++ b/tools/lib/bpf/libbpf.map
> @@ -438,4 +438,5 @@ LIBBPF_1.6.0 {
>                 bpf_linker__new_fd;
>                 btf__add_decl_attr;
>                 btf__add_type_attr;
> +               libbpf_probe_bpf_kfunc;
>  } LIBBPF_1.5.0;
> diff --git a/tools/lib/bpf/libbpf_probes.c b/tools/lib/bpf/libbpf_probes.c
> index 8ed92ea922b3..ab5591c385de 100644
> --- a/tools/lib/bpf/libbpf_probes.c
> +++ b/tools/lib/bpf/libbpf_probes.c
> @@ -431,6 +431,54 @@ static bool can_probe_prog_type(enum bpf_prog_type prog_type)
>         return true;
>  }
>
> +int libbpf_probe_bpf_kfunc(enum bpf_prog_type prog_type, int kfunc_id, int btf_fd,
> +                          const void *opts)
> +{
> +       struct bpf_insn insns[] = {
> +               BPF_RAW_INSN(BPF_JMP | BPF_CALL, 0, BPF_PSEUDO_KFUNC_CALL, 1, kfunc_id),
> +               BPF_EXIT_INSN(),
> +       };
> +       const size_t insn_cnt = ARRAY_SIZE(insns);
> +       char buf[4096];
> +       int fd_array[2] = {-1};
> +       int ret;
> +
> +       if (opts)
> +               return libbpf_err(-EINVAL);
> +
> +       if (!can_probe_prog_type(prog_type))
> +               return -EOPNOTSUPP;

libbpf_err() here

pw-bot: cr

> +
> +       if (btf_fd >= 0) {
> +               fd_array[1] = btf_fd;
> +       } else if (btf_fd == -1) {

let's not hard-code the equality, use < 0 (though I'd follow
verifier's offset == 0 convention for vmlinux BTF here as well to stay
conceptually consistent)

> +               /* insn.off = 0, means vmlinux btf */
> +               insns[0].off = 0;
> +       } else {
> +               return libbpf_err(-EINVAL);
> +       }
> +
> +       buf[0] = '\0';
> +       ret = probe_prog_load(prog_type, insns, insn_cnt, btf_fd >= 0 ? fd_array : NULL,
> +                             buf, sizeof(buf));
> +       if (ret < 0)
> +               return libbpf_err(ret);
> +
> +       /* If BPF verifier recognizes BPF kfunc but it's not supported for
> +        * given BPF program type, it will emit "calling kernel function
> +        * bpf_cpumask_create is not allowed", if the kfunc id is invalid,

bpf_cpumask_create -> <name> to keep comments generic?

> +        * it will emit "kernel btf_id 4294967295 is not a function". If btf fd

same as above, use <id> placeholder instead of specific number?

and keep BTF (all caps) use consistent, please

> +        * invalid in module btf, it will emit "invalid module BTF fd specified" or

ditto, btf -> BTF

> +        * "negative offset disallowed for kernel module function call"
> +        */
> +       if (ret == 0 && (strstr(buf, "not allowed") || strstr(buf, "not a function") ||
> +                       (strstr(buf, "invalid module BTF fd")) ||
> +                       (strstr(buf, "negative offset disallowed"))))

stylistically, given amount of checks, I'd probably go with the
following structure

if (ret > 0)
    return 1;

if (strstr(buf, "not allowed") ||
    strstr(buf, "not a function") ||
...)
    return 0;

> +               return 0;
> +
> +       return 1; /* assume supported */
> +}
> +
>  int libbpf_probe_bpf_helper(enum bpf_prog_type prog_type, enum bpf_func_id helper_id,
>                             const void *opts)
>  {
> --
> 2.43.0
>

Powered by blists - more mailing lists