Re: [PATCH v2 bpf-next 6/7] libbpf: support kernel module ksym externs

From: Andrii Nakryiko <andrii.nakryiko@gmail.com>
To: Hao Luo <haoluo@google.com>
Cc: Andrii Nakryiko <andrii@kernel.org>, bpf <bpf@vger.kernel.org>,
	Networking <netdev@vger.kernel.org>,
	Alexei Starovoitov <ast@fb.com>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Kernel Team <kernel-team@fb.com>
Subject: Re: [PATCH v2 bpf-next 6/7] libbpf: support kernel module ksym externs
Date: Mon, 11 Jan 2021 13:39:25 -0800	[thread overview]
Message-ID: <CAEf4Bza7XcXMGb-85SE+2ymtRsxvEooWg18dvjh6xztJ6k3akQ@mail.gmail.com> (raw)
In-Reply-To: <CA+khW7jWiTRe36Uc5zKzk_bHmC+R_QZ43EBRo0gmPGhZHiOrqA@mail.gmail.com>

On Mon, Jan 11, 2021 at 11:00 AM Hao Luo <haoluo@google.com> wrote:
>
> Acked-by: Hao Luo <haoluo@google.com>, with a couple of nits.
>
> On Fri, Jan 8, 2021 at 2:09 PM Andrii Nakryiko <andrii@kernel.org> wrote:
> >
> > Add support for searching for ksym externs not just in vmlinux BTF, but across
> > all module BTFs, similarly to how it's done for CO-RE relocations. Kernels
> > that expose module BTFs through sysfs are assumed to support new ldimm64
> > instruction extension with BTF FD provided in insn[1].imm field, so no extra
> > feature detection is performed.
> >
> > Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
> > ---
> >  tools/lib/bpf/libbpf.c | 47 +++++++++++++++++++++++++++---------------
> >  1 file changed, 30 insertions(+), 17 deletions(-)
> >
> > diff --git a/tools/lib/bpf/libbpf.c b/tools/lib/bpf/libbpf.c
> > index 6ae748f6ea11..57559a71e4de 100644
> > --- a/tools/lib/bpf/libbpf.c
> > +++ b/tools/lib/bpf/libbpf.c
> [...]
> > @@ -7319,7 +7321,8 @@ static int bpf_object__read_kallsyms_file(struct bpf_object *obj)
> >  static int bpf_object__resolve_ksyms_btf_id(struct bpf_object *obj)
> >  {
> >         struct extern_desc *ext;
> > -       int i, id;
> > +       struct btf *btf;
> > +       int i, j, id, btf_fd, err;
> >
> >         for (i = 0; i < obj->nr_extern; i++) {
> >                 const struct btf_type *targ_var, *targ_type;
> > @@ -7331,8 +7334,22 @@ static int bpf_object__resolve_ksyms_btf_id(struct bpf_object *obj)
> >                 if (ext->type != EXT_KSYM || !ext->ksym.type_id)
> >                         continue;
> >
> > -               id = btf__find_by_name_kind(obj->btf_vmlinux, ext->name,
> > -                                           BTF_KIND_VAR);
> > +               btf = obj->btf_vmlinux;
> > +               btf_fd = 0;
> > +               id = btf__find_by_name_kind(btf, ext->name, BTF_KIND_VAR);
>
> Is "if (id <= 0)" better? Just in case, more error code is introduced in future.

There is id <= 0 right below after special-casing -ENOENT, so all
works as you want, no?

>
> > +               if (id == -ENOENT) {
> > +                       err = load_module_btfs(obj);
> > +                       if (err)
> > +                               return err;
> > +
> > +                       for (j = 0; j < obj->btf_module_cnt; j++) {
> > +                               btf = obj->btf_modules[j].btf;
> > +                               btf_fd = obj->btf_modules[j].fd;
> > +                               id = btf__find_by_name_kind(btf, ext->name, BTF_KIND_VAR);
> > +                               if (id != -ENOENT)
> > +                                       break;
> > +                       }
> > +               }
> >                 if (id <= 0) {
>
> Nit: the warning message isn't accurate any more, right? We also
> searched kernel modules' BTF.

Right, how about just "failed to find BTF ID in kernel BTF"? Where
"kernel BTF" is "vmlinux BTF or any of kernel modules' BTFs"?

>
> >                         pr_warn("extern (ksym) '%s': failed to find BTF ID in vmlinux BTF.\n",
> >                                 ext->name);
> > @@ -7343,24 +7360,19 @@ static int bpf_object__resolve_ksyms_btf_id(struct bpf_object *obj)
> [...]
>
> > --
> > 2.24.1
> >