Received-SPF: pass (google.com: domain of linux-kernel+bounces-208734-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Precedence: bulk
MIME-Version: 1.0
References: <cover.1717881178.git.dxu@dxuuu.xyz> <e172bf47f32c6e716322bc85bb84d78b1398bd7c.1717881178.git.dxu@dxuuu.xyz>
In-Reply-To: <e172bf47f32c6e716322bc85bb84d78b1398bd7c.1717881178.git.dxu@dxuuu.xyz>
From: Alexei Starovoitov <alexei.starovoitov@gmail.com>
Date: Mon, 10 Jun 2024 11:30:31 -0700
Message-ID: <CAADnVQLE=XcpZ4SnW=NARG0D5Ya6iU1-1CayTVmArnxpSzWSFA@mail.gmail.com>
Subject: Re: [PATCH bpf-next v4 08/12] bpf: verifier: Relax caller
 requirements for kfunc projection type args
To: Daniel Xu <dxu@dxuuu.xyz>
Cc: Alexei Starovoitov <ast@kernel.org>, Daniel Borkmann <daniel@iogearbox.net>, 
	Andrii Nakryiko <andrii@kernel.org>, Jiri Olsa <olsajiri@gmail.com>, 
	Quentin Monnet <quentin@isovalent.com>, Alan Maguire <alan.maguire@oracle.com>, 
	Arnaldo Carvalho de Melo <acme@kernel.org>, Eddy Z <eddyz87@gmail.com>, 
	John Fastabend <john.fastabend@gmail.com>, Martin KaFai Lau <martin.lau@linux.dev>, 
	Song Liu <song@kernel.org>, Yonghong Song <yonghong.song@linux.dev>, KP Singh <kpsingh@kernel.org>, 
	Stanislav Fomichev <sdf@google.com>, Hao Luo <haoluo@google.com>, Jiri Olsa <jolsa@kernel.org>, 
	bpf <bpf@vger.kernel.org>, LKML <linux-kernel@vger.kernel.org>, 
	Kernel Team <kernel-team@meta.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Sat, Jun 8, 2024 at 2:16=E2=80=AFPM Daniel Xu <dxu@dxuuu.xyz> wrote:
>
> Currently, if a kfunc accepts a projection type as an argument (eg
> struct __sk_buff *), the caller must exactly provide exactly the same
> type with provable provenance.
>
> However in practice, kfuncs that accept projection types _must_ cast to
> the underlying type before use b/c projection type layouts are
> completely made up. Thus, it is ok to relax the verifier rules around
> implicit conversions.
>
> We will use this functionality in the next commit when we align kfuncs
> to user-facing types.
>
> Signed-off-by: Daniel Xu <dxu@dxuuu.xyz>
> ---
>  kernel/bpf/verifier.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
> index 81a3d2ced78d..0808beca3837 100644
> --- a/kernel/bpf/verifier.c
> +++ b/kernel/bpf/verifier.c
> @@ -11257,6 +11257,8 @@ static int process_kf_arg_ptr_to_btf_id(struct bp=
f_verifier_env *env,
>         bool strict_type_match =3D false;
>         const struct btf *reg_btf;
>         const char *reg_ref_tname;
> +       bool taking_projection;
> +       bool struct_same;
>         u32 reg_ref_id;
>
>         if (base_type(reg->type) =3D=3D PTR_TO_BTF_ID) {
> @@ -11300,7 +11302,13 @@ static int process_kf_arg_ptr_to_btf_id(struct b=
pf_verifier_env *env,
>
>         reg_ref_t =3D btf_type_skip_modifiers(reg_btf, reg_ref_id, &reg_r=
ef_id);
>         reg_ref_tname =3D btf_name_by_offset(reg_btf, reg_ref_t->name_off=
);
> -       if (!btf_struct_ids_match(&env->log, reg_btf, reg_ref_id, reg->of=
f, meta->btf, ref_id, strict_type_match)) {
> +       struct_same =3D btf_struct_ids_match(&env->log, reg_btf, reg_ref_=
id, reg->off, meta->btf, ref_id, strict_type_match);
> +       /* If kfunc is accepting a projection type (ie. __sk_buff), it ca=
nnot
> +        * actually use it -- it must cast to the underlying type. So we =
allow
> +        * caller to pass in the underlying type.
> +        */
> +       taking_projection =3D !strcmp(ref_tname, "__sk_buff") && !strcmp(=
reg_ref_tname, "sk_buff");

xdp_md/buff probably as well?

And with that share the code with btf_is_prog_ctx_type() ?

> +       if (!taking_projection && !struct_same) {
>                 verbose(env, "kernel function %s args#%d expected pointer=
 to %s %s but R%d has a pointer to %s %s\n",
>                         meta->func_name, argno, btf_type_str(ref_t), ref_=
tname, argno + 1,
>                         btf_type_str(reg_ref_t), reg_ref_tname);
> --
> 2.44.0
>