Received-SPF: pass (google.com: domain of linux-kernel+bounces-110312-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: <20240315-hid-bpf-sleepable-v4-0-5658f2540564@kernel.org>
 <20240315-hid-bpf-sleepable-v4-4-5658f2540564@kernel.org> <bee98dce6c5bf59fb1ad06855f0a1740b8195d45.camel@gmail.com>
In-Reply-To: <bee98dce6c5bf59fb1ad06855f0a1740b8195d45.camel@gmail.com>
From: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Date: Thu, 21 Mar 2024 17:09:32 +0100
Message-ID: <CAO-hwJLh00--9AbS=7z4+W7FgA+hLA8hViMmaMMfz2mERcMwhQ@mail.gmail.com>
Subject: Re: [PATCH bpf-next v4 4/6] bpf/helpers: mark the callback of
 bpf_timer_set_sleepable_cb() as sleepable
To: Eduard Zingerman <eddyz87@gmail.com>
Cc: Benjamin Tissoires <bentiss@kernel.org>, Alexei Starovoitov <ast@kernel.org>, 
	Daniel Borkmann <daniel@iogearbox.net>, Andrii Nakryiko <andrii@kernel.org>, 
	Martin KaFai Lau <martin.lau@linux.dev>, Song Liu <song@kernel.org>, 
	Yonghong Song <yonghong.song@linux.dev>, John Fastabend <john.fastabend@gmail.com>, 
	KP Singh <kpsingh@kernel.org>, Stanislav Fomichev <sdf@google.com>, Hao Luo <haoluo@google.com>, 
	Jiri Olsa <jolsa@kernel.org>, Mykola Lysenko <mykolal@fb.com>, Shuah Khan <shuah@kernel.org>, 
	bpf@vger.kernel.org, linux-kernel@vger.kernel.org, 
	linux-kselftest@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tue, Mar 19, 2024 at 12:54=E2=80=AFAM Eduard Zingerman <eddyz87@gmail.co=
m> wrote:
>
> On Fri, 2024-03-15 at 15:29 +0100, Benjamin Tissoires wrote:
> [...]
>
> > @@ -5279,7 +5281,8 @@ static int map_kptr_match_type(struct bpf_verifie=
r_env *env,
> >
> >  static bool in_sleepable(struct bpf_verifier_env *env)
> >  {
> > -     return env->prog->sleepable;
> > +     return env->prog->sleepable ||
> > +            (env->cur_state && env->cur_state->in_sleepable);
> >  }
>
> I was curious why 'env->cur_state &&' check was needed and found that
> removing it caused an error in the following fragment:
>
> static int do_misc_fixups(struct bpf_verifier_env *env)
> {
>                 ...
>                 if (is_storage_get_function(insn->imm)) {
>                         if (!in_sleepable(env) ||
>                             env->insn_aux_data[i + delta].storage_get_fun=
c_atomic)
>                                 insn_buf[0] =3D BPF_MOV64_IMM(BPF_REG_5, =
(__force __s32)GFP_ATOMIC);
>                         else
>                                 insn_buf[0] =3D BPF_MOV64_IMM(BPF_REG_5, =
(__force __s32)GFP_KERNEL);
>                         ...
>                 }
>                 ...
> }
>
> When do_misc_fixups() is done env->cur_state is NULL.
> Current implementation would use GFP_ATOMIC allocation even for
> sleepable callbacks, where GFP_KERNEL is sufficient.
> Is this is something we want to address?

I honestly have no idea of the impact there.

AFAICT, if env->cur_state is not set, we don't even know if the
callback will be sleepable or not, so if there is a small penalty,
then it's the safest option, no?

Cheers,
Benjamin

>
> >
> >  /* The non-sleepable programs and sleepable programs with explicit bpf=
_rcu_read_lock()
>