Received-SPF: pass (google.com: domain of linux-kernel+bounces-134312-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Date: Sun, 07 Apr 2024 09:54:19 +0000
To: Viresh Kumar <viresh.kumar@linaro.org>, "Rafael J. Wysocki" <rafael@kernel.org>, Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>, Miguel Ojeda <ojeda@kernel.org>, Alex Gaynor <alex.gaynor@gmail.com>, Wedson Almeida Filho <wedsonaf@gmail.com>, Boqun Feng <boqun.feng@gmail.com>, Gary Guo <gary@garyguo.net>, =?utf-8?Q?Bj=C3=B6rn_Roy_Baron?= <bjorn3_gh@protonmail.com>, Andreas Hindborg <a.hindborg@samsung.com>, Alice Ryhl <aliceryhl@google.com>
From: Benno Lossin <benno.lossin@proton.me>
Cc: linux-pm@vger.kernel.org, Vincent Guittot <vincent.guittot@linaro.org>, Stephen Boyd <sboyd@kernel.org>, Nishanth Menon <nm@ti.com>, rust-for-linux@vger.kernel.org, Manos Pitsidianakis <manos.pitsidianakis@linaro.org>, Erik Schilling <erik.schilling@linaro.org>, =?utf-8?Q?Alex_Benn=C3=A9e?= <alex.bennee@linaro.org>, Joakim Bech <joakim.bech@linaro.org>, linux-kernel@vger.kernel.org
Subject: Re: [RFC PATCH 1/3] rust: Add bindings for OPP framework
Message-ID: <ff6c7d5e-d6e9-4331-b8cc-eab139160e59@proton.me>
In-Reply-To: <06bb914eae00671a69b393bf86bb01ddec86c16f.1712314032.git.viresh.kumar@linaro.org>
References: <cover.1712314032.git.viresh.kumar@linaro.org> <06bb914eae00671a69b393bf86bb01ddec86c16f.1712314032.git.viresh.kumar@linaro.org>
Feedback-ID: 71780778:user:proton
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

Hi,

I took a quick look and left some comments from the Rust side of view.

On 05.04.24 13:09, Viresh Kumar wrote:
> +/// Equivalent to `struct dev_pm_opp_config` in the C Code.
> +pub struct Config<T: ConfigOps> {
> +    token: Option<i32>,
> +    clk_names: Option<Pin<Vec<CString>>>,

Why are you using `Pin<Vec<_>>`? The vector may reallocate the backing
storage at any point in time.

> +    prop_name: Option<Pin<CString>>,
> +    regulator_names: Option<Pin<Vec<CString>>>,
> +    genpd_names: Option<Pin<Vec<CString>>>,
> +    supported_hw: Option<Pin<Vec<u32>>>,
> +    required_devs: Option<Pin<Vec<Device>>>,
> +    _data: PhantomData<T>,
> +}

[...]

> +    /// Sets the configuration with the OPP core.
> +    pub fn set(&mut self, dev: &Device) -> Result<()> {
> +        // Already configured.
> +        if self.token.is_some() {

Why does the config hold onto this token? Would it make sense to consume
the config and return a `Handle` or `Token` abstraction? Then you don't
need to check if the config has been "used" before.

> +            return Err(EBUSY);
> +        }
> +
> +        let (_clk_list, clk_names) =3D match &self.clk_names {
> +            Some(x) =3D> {
> +                let list =3D to_c_str_array(x)?;
> +                let ptr =3D list.as_ptr();
> +                (Some(list), ptr)
> +            }
> +            None =3D> (None, ptr::null()),
> +        };

[...]

> +/// Operating performance point (OPP).
> +///
> +/// # Invariants
> +///
> +/// `ptr` is valid, non-null, and has a non-zero reference count. One of=
 the references is owned by
> +/// `self`, and will be decremented when `self` is dropped.
> +#[repr(transparent)]
> +pub struct OPP(*mut bindings::dev_pm_opp);

I think you should use the `ARef` pattern instead:

     #[repr(transparent)]
     pub struct OPP(Opaque<bindings::dev_pm_opp>);

     unsafe impl AlwaysRefCounted for OPP {
         // ...
     }

Then you can use `ARef<OPP>` everywhere you use `OPP` currently.

--=20
Cheers,
Benno