rust-for-linux.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Alexandre Courbot" <acourbot@nvidia.com>
To: "Alice Ryhl" <aliceryhl@google.com>,
	"Alexandre Courbot" <acourbot@nvidia.com>
Cc: "Miguel Ojeda" <ojeda@kernel.org>,
	"Boqun Feng" <boqun.feng@gmail.com>,
	"Gary Guo" <gary@garyguo.net>,
	"Björn Roy Baron" <bjorn3_gh@protonmail.com>,
	"Benno Lossin" <lossin@kernel.org>,
	"Andreas Hindborg" <a.hindborg@kernel.org>,
	"Trevor Gross" <tmgross@umich.edu>,
	"Danilo Krummrich" <dakr@kernel.org>,
	"Nathan Chancellor" <nathan@kernel.org>,
	"Nicolas Schier" <nicolas.schier@linux.dev>,
	linux-kernel@vger.kernel.org, rust-for-linux@vger.kernel.org,
	linux-kbuild@vger.kernel.org
Subject: Re: [PATCH RESEND] rust: enable slice_flatten feature and abstract it through an extension trait
Date: Tue, 04 Nov 2025 19:57:02 +0900	[thread overview]
Message-ID: <DDZUUDRMJNC9.36PS0YTZCSFO6@nvidia.com> (raw)
In-Reply-To: <aQm7LNwPwxm8HZ_I@google.com>

On Tue Nov 4, 2025 at 5:37 PM JST, Alice Ryhl wrote:
> On Tue, Nov 04, 2025 at 12:06:54AM +0900, Alexandre Courbot wrote:
>> On Mon Nov 3, 2025 at 7:22 PM JST, Alice Ryhl wrote:
>> > On Sat, Nov 01, 2025 at 10:31:38PM +0900, Alexandre Courbot wrote:
>> >> In Rust 1.80, the previously unstable `slice::flatten` family of methods
>> >> have been stabilized and renamed to `as_flattened`.
>> >> 
>> >> This creates an issue as we want to use `as_flattened`, but need to
>> >> support the MSRV (which at the moment is Rust 1.78) where it is named
>> >> `flatten`.
>> >> 
>> >> Solve this by enabling the `slice_flatten` feature, and abstracting
>> >> `as_flatten` behind an extension trait that calls the right method
>> >> depending on the Rust version.
>> >> 
>> >> This extension trait can be removed once the MSRV passes 1.80.
>> >> 
>> >> Suggested-by: Miguel Ojeda <ojeda@kernel.org>
>> >> Link: https://lore.kernel.org/all/CANiq72kK4pG=O35NwxPNoTO17oRcg1yfGcvr3==Fi4edr+sfmw@mail.gmail.com/
>> >> Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
>> >> ---
>> >> This patch was part of the Nova GSP boot series [1], but since it
>> >> requires attention from the core Rust team (and possibly the build
>> >> maintainers?) and is otherwise buried under Nova patches, I am taking
>> >> the freedom to send it separately for visibility.
>> >> 
>> >> Hopefully it captures Miguel's suggestion [2] accurately, but please let
>> >> me know if I missed something.
>> >> 
>> >> Since the Nova GSP boot series makes use of this, I hope to eventually
>> >> merge it alongside the series, through the DRM tree.
>> >> 
>> >> Thanks!
>> >> 
>> >> [1] https://lore.kernel.org/all/20251029-gsp_boot-v7-0-34227afad347@nvidia.com/
>> >> [2] https://lore.kernel.org/all/CANiq72kK4pG=O35NwxPNoTO17oRcg1yfGcvr3==Fi4edr+sfmw@mail.gmail.com/
>> >> ---
>> >>  init/Kconfig           |  3 +++
>> >>  rust/kernel/lib.rs     |  4 ++++
>> >>  rust/kernel/slice.rs   | 63 ++++++++++++++++++++++++++++++++++++++++++++++++++
>> >>  scripts/Makefile.build |  3 ++-
>> >>  4 files changed, 72 insertions(+), 1 deletion(-)
>> >> 
>> >> diff --git a/init/Kconfig b/init/Kconfig
>> >> index cab3ad28ca49..7da93c9cccc3 100644
>> >> --- a/init/Kconfig
>> >> +++ b/init/Kconfig
>> >> @@ -147,6 +147,9 @@ config LD_CAN_USE_KEEP_IN_OVERLAY
>> >>  	# https://github.com/llvm/llvm-project/pull/130661
>> >>  	def_bool LD_IS_BFD || LLD_VERSION >= 210000
>> >>  
>> >> +config RUSTC_HAS_SLICE_AS_FLATTENED
>> >> +	def_bool RUSTC_VERSION >= 108000
>> >> +
>> >>  config RUSTC_HAS_COERCE_POINTEE
>> >>  	def_bool RUSTC_VERSION >= 108400
>> >>  
>> >> diff --git a/rust/kernel/lib.rs b/rust/kernel/lib.rs
>> >> index 3dd7bebe7888..2581a356d114 100644
>> >> --- a/rust/kernel/lib.rs
>> >> +++ b/rust/kernel/lib.rs
>> >> @@ -21,6 +21,9 @@
>> >>  #![feature(inline_const)]
>> >>  #![feature(pointer_is_aligned)]
>> >>  //
>> >> +// Stable since Rust 1.80.0.
>> >> +#![feature(slice_flatten)]
>> >> +//
>> >>  // Stable since Rust 1.81.0.
>> >>  #![feature(lint_reasons)]
>> >>  //
>> >> @@ -128,6 +131,7 @@
>> >>  pub mod security;
>> >>  pub mod seq_file;
>> >>  pub mod sizes;
>> >> +pub mod slice;
>> >>  mod static_assert;
>> >>  #[doc(hidden)]
>> >>  pub mod std_vendor;
>> >> diff --git a/rust/kernel/slice.rs b/rust/kernel/slice.rs
>> >> new file mode 100644
>> >> index 000000000000..7e837bec4bed
>> >> --- /dev/null
>> >> +++ b/rust/kernel/slice.rs
>> >> @@ -0,0 +1,63 @@
>> >> +// SPDX-License-Identifier: GPL-2.0
>> >> +
>> >> +//! Additional (and temporary) slice helpers.
>> >> +
>> >> +/// Extension trait providing a portable version of [`as_flattened`] and
>> >> +/// [`as_flattened_mut`].
>> >> +///
>> >> +/// In Rust 1.80, the previously unstable `slice::flatten` family of methods
>> >> +/// have been stabilized and renamed from `flatten` to `as_flattened`.
>> >> +///
>> >> +/// This creates an issue for as long as the MSRV is < 1.80, as the same functionality is provided
>> >> +/// by different methods depending on the compiler version.
>> >> +///
>> >> +/// This extension trait solves this by abstracting `as_flatten` and calling the correct  method
>> >> +/// depending on the Rust version.
>> >> +///
>> >> +/// This trait can be removed once the MSRV passes 1.80.
>> >> +///
>> >> +/// [`as_flattened`]: slice::as_flattened
>> >> +/// [`as_flattened_mut`]: slice::as_flattened_mut
>> >> +pub trait AsFlattened<T> {
>> >> +    /// Takes an `&[[T; N]]` and flattens it to a `&[T]`.
>> >> +    ///
>> >> +    /// This is an portable layer on top of [`as_flattened`]; see its documentation for details.
>> >> +    ///
>> >> +    /// [`as_flattened`]: slice::as_flattened
>> >> +    fn as_flattened_slice(&self) -> &[T];
>> >> +
>> >> +    /// Takes an `&mut [[T; N]]` and flattens it to a `&mut [T]`.
>> >> +    ///
>> >> +    /// This is an portable layer on top of [`as_flattened_mut`]; see its documentation for details.
>> >> +    ///
>> >> +    /// [`as_flattened_mut`]: slice::as_flattened_mut
>> >> +    fn as_flattened_slice_mut(&mut self) -> &mut [T];
>> >> +}
>> >> +
>> >> +impl<T, const N: usize> AsFlattened<T> for [[T; N]] {
>> >> +    #[allow(clippy::incompatible_msrv)]
>> >> +    fn as_flattened_slice(&self) -> &[T] {
>> >> +        #[cfg(not(CONFIG_RUSTC_HAS_SLICE_AS_FLATTENED))]
>> >> +        {
>> >> +            self.flatten()
>> >> +        }
>> >> +
>> >> +        #[cfg(CONFIG_RUSTC_HAS_SLICE_AS_FLATTENED)]
>> >> +        {
>> >> +            self.as_flattened()
>> >> +        }
>> >> +    }
>> >> +
>> >> +    #[allow(clippy::incompatible_msrv)]
>> >> +    fn as_flattened_slice_mut(&mut self) -> &mut [T] {
>> >> +        #[cfg(not(CONFIG_RUSTC_HAS_SLICE_AS_FLATTENED))]
>> >> +        {
>> >> +            self.flatten_mut()
>> >> +        }
>> >> +
>> >> +        #[cfg(CONFIG_RUSTC_HAS_SLICE_AS_FLATTENED)]
>> >> +        {
>> >> +            self.as_flattened_mut()
>> >> +        }
>> >
>> > Hmm. Why not have this match the name that this was stabilized under?
>> > That way, when bumping the MSRV, we can just remove this trait without
>> > changing the callers.
>> 
>> I expected that doing so would clash with the methods of the same name
>> in `slice` on Rust >= 1.80, but instead I just got a warning that my
>> `AsFlattened` import was ignored - it looks like Rust did the right
>> thing and picked the non-trait method by default.
>> 
>> So all we need to do to make your proposal work is to make the import
>> conditional on Rust being < 1.80. And we can also make the whole trait
>> and its impl block that way.
>> 
>> That would actually be much cleaner! Thanks a lot.
>
> One option is to put the trait in the prelude on relevant rustc
> versions. That way, users of the method don't need any conditional
> logic.

And with that last touch I think we have reached perfection! :) Thanks
again.

      reply	other threads:[~2025-11-04 10:57 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-01 13:31 [PATCH RESEND] rust: enable slice_flatten feature and abstract it through an extension trait Alexandre Courbot
2025-11-01 15:36 ` Miguel Ojeda
2025-11-02  2:25   ` Alexandre Courbot
2025-11-02 12:00     ` Miguel Ojeda
2025-11-02 14:28       ` Alexandre Courbot
2025-11-01 15:52 ` Danilo Krummrich
2025-11-03 10:22 ` Alice Ryhl
2025-11-03 15:06   ` Alexandre Courbot
2025-11-04  8:37     ` Alice Ryhl
2025-11-04 10:57       ` Alexandre Courbot [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=DDZUUDRMJNC9.36PS0YTZCSFO6@nvidia.com \
    --to=acourbot@nvidia.com \
    --cc=a.hindborg@kernel.org \
    --cc=aliceryhl@google.com \
    --cc=bjorn3_gh@protonmail.com \
    --cc=boqun.feng@gmail.com \
    --cc=dakr@kernel.org \
    --cc=gary@garyguo.net \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lossin@kernel.org \
    --cc=nathan@kernel.org \
    --cc=nicolas.schier@linux.dev \
    --cc=ojeda@kernel.org \
    --cc=rust-for-linux@vger.kernel.org \
    --cc=tmgross@umich.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).