Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:7 as permitted sender) client-ip=2620:137:e000::3:7;
MIME-Version: 1.0
References: <20231011120857.251943-1-zhouchuyi@bytedance.com> <20231011120857.251943-4-zhouchuyi@bytedance.com>
In-Reply-To: <20231011120857.251943-4-zhouchuyi@bytedance.com>
From:   Andrii Nakryiko <andrii.nakryiko@gmail.com>
Date:   Fri, 13 Oct 2023 14:27:15 -0700
Message-ID: <CAEf4BzYGZiTUHPkjuF81vWZWPH-x4rxz1s9+T0rh-dsrO5ZwDg@mail.gmail.com>
Subject: Re: [PATCH bpf-next v5 3/8] bpf: Introduce task open coded iterator kfuncs
To:     Chuyi Zhou <zhouchuyi@bytedance.com>
Cc:     bpf@vger.kernel.org, ast@kernel.org, daniel@iogearbox.net,
        andrii@kernel.org, martin.lau@kernel.org, tj@kernel.org,
        linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Oct 11, 2023 at 5:09=E2=80=AFAM Chuyi Zhou <zhouchuyi@bytedance.com=
> wrote:
>
> This patch adds kfuncs bpf_iter_task_{new,next,destroy} which allow
> creation and manipulation of struct bpf_iter_task in open-coded iterator
> style. BPF programs can use these kfuncs or through bpf_for_each macro to
> iterate all processes in the system.
>
> The API design keep consistent with SEC("iter/task"). bpf_iter_task_new()
> accepts a specific task and iterating type which allows:
>
> 1. iterating all process in the system(BPF_TASK_ITER_ALL_PROCS)
>
> 2. iterating all threads in the system(BPF_TASK_ITER_ALL_THREADS)
>
> 3. iterating all threads of a specific task(BPF_TASK_ITER_PROC_THREADS)
>
> Signed-off-by: Chuyi Zhou <zhouchuyi@bytedance.com>
> ---
>  kernel/bpf/helpers.c                          |  3 +
>  kernel/bpf/task_iter.c                        | 82 +++++++++++++++++++
>  .../testing/selftests/bpf/bpf_experimental.h  |  5 ++
>  3 files changed, 90 insertions(+)
>
> diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
> index cb24c4a916df..690763751f6e 100644
> --- a/kernel/bpf/helpers.c
> +++ b/kernel/bpf/helpers.c
> @@ -2555,6 +2555,9 @@ BTF_ID_FLAGS(func, bpf_iter_num_destroy, KF_ITER_DE=
STROY)
>  BTF_ID_FLAGS(func, bpf_iter_css_task_new, KF_ITER_NEW | KF_TRUSTED_ARGS)
>  BTF_ID_FLAGS(func, bpf_iter_css_task_next, KF_ITER_NEXT | KF_RET_NULL)
>  BTF_ID_FLAGS(func, bpf_iter_css_task_destroy, KF_ITER_DESTROY)
> +BTF_ID_FLAGS(func, bpf_iter_task_new, KF_ITER_NEW | KF_TRUSTED_ARGS)
> +BTF_ID_FLAGS(func, bpf_iter_task_next, KF_ITER_NEXT | KF_RET_NULL)
> +BTF_ID_FLAGS(func, bpf_iter_task_destroy, KF_ITER_DESTROY)
>  BTF_ID_FLAGS(func, bpf_dynptr_adjust)
>  BTF_ID_FLAGS(func, bpf_dynptr_is_null)
>  BTF_ID_FLAGS(func, bpf_dynptr_is_rdonly)
> diff --git a/kernel/bpf/task_iter.c b/kernel/bpf/task_iter.c
> index 2cfcb4dd8a37..caeddad3d2f1 100644
> --- a/kernel/bpf/task_iter.c
> +++ b/kernel/bpf/task_iter.c
> @@ -856,6 +856,88 @@ __bpf_kfunc void bpf_iter_css_task_destroy(struct bp=
f_iter_css_task *it)
>         bpf_mem_free(&bpf_global_ma, kit->css_it);
>  }
>
> +struct bpf_iter_task {
> +       __u64 __opaque[3];
> +} __attribute__((aligned(8)));
> +
> +struct bpf_iter_task_kern {
> +       struct task_struct *task;
> +       struct task_struct *pos;
> +       unsigned int flags;
> +} __attribute__((aligned(8)));
> +
> +enum {
> +       BPF_TASK_ITER_ALL_PROCS,
> +       BPF_TASK_ITER_ALL_THREADS,
> +       BPF_TASK_ITER_PROC_THREADS
> +};
> +
> +__bpf_kfunc int bpf_iter_task_new(struct bpf_iter_task *it,
> +               struct task_struct *task, unsigned int flags)
> +{
> +       struct bpf_iter_task_kern *kit =3D (void *)it;
> +
> +       BUILD_BUG_ON(sizeof(struct bpf_iter_task_kern) > sizeof(struct bp=
f_iter_task));
> +       BUILD_BUG_ON(__alignof__(struct bpf_iter_task_kern) !=3D
> +                                       __alignof__(struct bpf_iter_task)=
);
> +
> +       kit->task =3D kit->pos =3D NULL;
> +       switch (flags) {
> +       case BPF_TASK_ITER_ALL_THREADS:
> +       case BPF_TASK_ITER_ALL_PROCS:
> +       case BPF_TASK_ITER_PROC_THREADS:
> +               break;
> +       default:
> +               return -EINVAL;
> +       }
> +
> +       if (flags =3D=3D BPF_TASK_ITER_PROC_THREADS)
> +               kit->task =3D task;
> +       else
> +               kit->task =3D &init_task;
> +       kit->pos =3D kit->task;
> +       kit->flags =3D flags;
> +       return 0;
> +}
> +
> +__bpf_kfunc struct task_struct *bpf_iter_task_next(struct bpf_iter_task =
*it)
> +{
> +       struct bpf_iter_task_kern *kit =3D (void *)it;
> +       struct task_struct *pos;
> +       unsigned int flags;
> +
> +       flags =3D kit->flags;
> +       pos =3D kit->pos;
> +
> +       if (!pos)
> +               goto out;
> +
> +       if (flags =3D=3D BPF_TASK_ITER_ALL_PROCS)
> +               goto get_next_task;
> +
> +       kit->pos =3D next_thread(kit->pos);
> +       if (kit->pos =3D=3D kit->task) {
> +               if (flags =3D=3D BPF_TASK_ITER_PROC_THREADS) {
> +                       kit->pos =3D NULL;
> +                       goto out;
> +               }
> +       } else
> +               goto out;

nit: this should have {} around it to match the other if branch

but actually, why goto out instead of return pos? same above, return
pos instead of goto out?


> +
> +get_next_task:
> +       kit->pos =3D next_task(kit->pos);
> +       kit->task =3D kit->pos;
> +       if (kit->pos =3D=3D &init_task)
> +               kit->pos =3D NULL;
> +
> +out:
> +       return pos;
> +}
> +
> +__bpf_kfunc void bpf_iter_task_destroy(struct bpf_iter_task *it)
> +{
> +}
> +
>  DEFINE_PER_CPU(struct mmap_unlock_irq_work, mmap_unlock_work);
>
>  static void do_mmap_read_unlock(struct irq_work *entry)
> diff --git a/tools/testing/selftests/bpf/bpf_experimental.h b/tools/testi=
ng/selftests/bpf/bpf_experimental.h
> index 8b53537e0f27..1ec82997cce7 100644
> --- a/tools/testing/selftests/bpf/bpf_experimental.h
> +++ b/tools/testing/selftests/bpf/bpf_experimental.h
> @@ -457,5 +457,10 @@ extern int bpf_iter_css_task_new(struct bpf_iter_css=
_task *it,
>  extern struct task_struct *bpf_iter_css_task_next(struct bpf_iter_css_ta=
sk *it) __weak __ksym;
>  extern void bpf_iter_css_task_destroy(struct bpf_iter_css_task *it) __we=
ak __ksym;
>
> +struct bpf_iter_task;
> +extern int bpf_iter_task_new(struct bpf_iter_task *it,
> +               struct task_struct *task, unsigned int flags) __weak __ks=
ym;
> +extern struct task_struct *bpf_iter_task_next(struct bpf_iter_task *it) =
__weak __ksym;
> +extern void bpf_iter_task_destroy(struct bpf_iter_task *it) __weak __ksy=
m;
>
>  #endif
> --
> 2.20.1
>