On Mon, Feb 24, 2025 at 9:47 PM Tao Chen <chen.dylane@xxxxxxxxx> wrote:
>
> 在 2025/2/25 09:15, Andrii Nakryiko 写道:
> > On Mon, Feb 24, 2025 at 9:02 AM Tao Chen <chen.dylane@xxxxxxxxx> 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@xxxxxxxxxxxxxx>
> >> Reviewed-by: Jiri Olsa <jolsa@xxxxxxxxxx>
> >> Reviewed-by: Eduard Zingerman <eddyz87@xxxxxxxxx>
> >> Signed-off-by: Tao Chen <chen.dylane@xxxxxxxxx>
> >> ---
> >> tools/lib/bpf/libbpf.h | 19 ++++++++++++-
> >> tools/lib/bpf/libbpf.map | 1 +
> >> tools/lib/bpf/libbpf_probes.c | 51 +++++++++++++++++++++++++++++++++++
> >> 3 files changed, 70 insertions(+), 1 deletion(-)
> >>
> >
> > [...]
> >
> >> + 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 (ret > 0)
> >> + return 1; /* assume supported */
> >> +
> >> + /* If BPF verifier recognizes BPF kfunc but it's not supported for
> >> + * given BPF program type, it will emit "calling kernel function
> >> + * <name> is not allowed". If the kfunc id is invalid,
> >> + * it will emit "kernel btf_id <id> is not a function". If BTF fd
> >> + * invalid in module BTF, it will emit "invalid module BTF fd specified" or
> >> + * "negative offset disallowed for kernel module function call". If
> >> + * kfunc prog not dev buound, it will emit "metadata kfuncs require
> >> + * device-bound program".
> >> + */
> >> + if (strstr(buf, "not allowed") || strstr(buf, "not a function") ||
> >> + strstr(buf, "invalid module BTF fd") ||
> >
> > why is invalid module BTF FD not an error (negative return)?
> >
> >> + strstr(buf, "negative offset disallowed") ||
> >> + strstr(buf, "device-bound program"))
> >> + return 0;
> >> +
> >> + return 1;
> >> +}
> >> +
> >> int libbpf_probe_bpf_helper(enum bpf_prog_type prog_type, enum bpf_func_id helper_id,
> >> const void *opts)
> >> {
> >> --
> >> 2.43.0
> >>
>
> In probe_prog_load, err will be checked and converted into either 0 or 1.
I guess what I was getting at is that providing invalid module BTF FD
is not a "not supported" case, it's an error case (and so should
result in negative return)
>
> --
> Best Regards
> Tao Chen
