[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJ-ks9=UwOz=-NR0MJ3pjKZNGjb7+3R-shFqYKQF206s5Tp6ew@mail.gmail.com>
Date: Mon, 12 Jan 2026 12:48:52 -0500
From: Tamir Duberstein <tamird@...il.com>
To: Gary Guo <gary@...yguo.net>
Cc: Miguel Ojeda <ojeda@...nel.org>, Boqun Feng <boqun.feng@...il.com>,
Björn Roy Baron <bjorn3_gh@...tonmail.com>,
Benno Lossin <lossin@...nel.org>, Andreas Hindborg <a.hindborg@...nel.org>,
Alice Ryhl <aliceryhl@...gle.com>, Trevor Gross <tmgross@...ch.edu>,
Danilo Krummrich <dakr@...nel.org>, rust-for-linux@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 12/12] rust: macros: support `#[cfg]` properly in
`#[vtable]` macro.
On Mon, Jan 12, 2026 at 12:12 PM Gary Guo <gary@...nel.org> wrote:
>
> From: Gary Guo <gary@...yguo.net>
>
> Currently, we generate `HAS_` constants as long as the definition exists in
> the source, regardless if it is cfg-ed out or not.
>
> Currently, uses of `#[cfg]` present in both trait and impl, so it is not a
> problem; however if only the impl side uses `#[cfg]` then `HAS_` constants
> will incorrectly be true while it shouldnt't.
>
> With `syn` support, we can now implement `#[cfg]` handling properly by
> propagating the `#[cfg]` attributes to the constants.
>
> Signed-off-by: Gary Guo <gary@...yguo.net>
> ---
> rust/macros/helpers.rs | 9 +++++++++
> rust/macros/vtable.rs | 13 ++++---------
> 2 files changed, 13 insertions(+), 9 deletions(-)
>
> diff --git a/rust/macros/helpers.rs b/rust/macros/helpers.rs
> index adfa60d8f42d8..c17287da801dd 100644
> --- a/rust/macros/helpers.rs
> +++ b/rust/macros/helpers.rs
> @@ -7,6 +7,7 @@
> Parse,
> ParseStream, //
> },
> + Attribute,
> Error,
> LitStr,
> Result, //
> @@ -53,3 +54,11 @@ pub(crate) fn file() -> String {
> proc_macro::Span::call_site().file()
> }
> }
> +
> +/// Obtain all `#[cfg]` attributes.
> +pub(crate) fn gather_cfg_attrs(attr: &[Attribute]) -> Vec<Attribute> {
> + attr.iter()
> + .filter(|a| a.path().is_ident("cfg"))
> + .cloned()
> + .collect()
> +}
I believe there's no need to collect this into a vector, `quote`
supports iterators.
https://docs.rs/quote/latest/quote/macro.quote.html#interpolation
> diff --git a/rust/macros/vtable.rs b/rust/macros/vtable.rs
> index 72ae0a1816a04..c6510b0c4ea1d 100644
> --- a/rust/macros/vtable.rs
> +++ b/rust/macros/vtable.rs
> @@ -23,7 +23,6 @@
>
> fn handle_trait(mut item: ItemTrait) -> Result<ItemTrait> {
> let mut gen_items = Vec::new();
> - let mut gen_consts = HashSet::new();
>
> gen_items.push(parse_quote! {
> /// A marker to prevent implementors from forgetting to use [`#[vtable]`](vtable)
> @@ -38,22 +37,17 @@ fn handle_trait(mut item: ItemTrait) -> Result<ItemTrait> {
> &format!("HAS_{}", name.to_string().to_uppercase()),
> name.span(),
> );
> - // Skip if it's declared already -- this can happen if `#[cfg]` is used to selectively
> - // define functions.
> - // FIXME: `#[cfg]` should be copied and propagated to the generated consts.
> - if gen_consts.contains(&gen_const_name) {
> - continue;
> - }
>
> // We don't know on the implementation-site whether a method is required or provided
> // so we have to generate a const for all methods.
> + let cfg_attrs = crate::helpers::gather_cfg_attrs(&fn_item.attrs);
> let comment =
> format!("Indicates if the `{name}` method is overridden by the implementor.");
> gen_items.push(parse_quote! {
> + #(#cfg_attrs)*
> #[doc = #comment]
> const #gen_const_name: bool = false;
> });
> - gen_consts.insert(gen_const_name);
> }
> }
>
> @@ -87,10 +81,11 @@ fn handle_impl(mut item: ItemImpl) -> Result<ItemImpl> {
> if defined_consts.contains(&gen_const_name) {
> continue;
> }
> + let cfg_attrs = crate::helpers::gather_cfg_attrs(&fn_item.attrs);
> gen_items.push(parse_quote! {
> + #(#cfg_attrs)*
> const #gen_const_name: bool = true;
> });
> - defined_consts.insert(gen_const_name);
> }
> }
>
> --
> 2.51.2
>
Reviewed-by: Tamir Duberstein <tamird@...il.com>
Powered by blists - more mailing lists