linux/rust/kernel/workqueue.rs, branch v6.14

rust: workqueue: add missing newline to pr_info! examples

2025-03-06T19:44:06Z

The documentation examples in rust/kernel/workqueue.rs use pr_info! calls that lack a trailing newline. To maintain consistency with kernel logging practices, this patch adds the newline to all affected examples. Fixes: 15b286d1fd05 ("rust: workqueue: add examples") Reported-by: Miguel Ojeda Link: https://github.com/Rust-for-Linux/linux/issues/1139 Signed-off-by: Alban Kurti Reviewed-by: Alice Ryhl Link: https://lore.kernel.org/r/20250206-printing_fix-v3-5-a85273b501ae@invicto.ai [ Replaced Closes with Link since it fixes part of the issue. - Miguel ] Signed-off-by: Miguel Ojeda

Merge tag 'rust-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux

2025-01-22T01:48:03Z

Pull rust updates from Miguel Ojeda: "Toolchain and infrastructure: - Finish the move to custom FFI integer types started in the previous cycle and finally map 'long' to 'isize' and 'char' to 'u8'. Do a few cleanups on top thanks to that. - Start to use 'derive(CoercePointee)' on Rust >= 1.84.0. This is a major milestone on the path to build the kernel using only stable Rust features. In particular, previously we were using the unstable features 'coerce_unsized', 'dispatch_from_dyn' and 'unsize', and now we will use the new 'derive_coerce_pointee' one, which is on track to stabilization. This new feature is a macro that essentially expands into code that internally uses the unstable features that we were using before, without having to expose those. With it, stable Rust users, including the kernel, will be able to build custom smart pointers that work with trait objects, e.g.: fn f(p: &Arc) { pr_info!("{p}\n"); } let a: Arc = Arc::new(42i32, GFP_KERNEL)?; let b: Arc = Arc::new("hello there", GFP_KERNEL)?; f(&a); // Prints "42". f(&b); // Prints "hello there". Together with the 'arbitrary_self_types' feature that we started using in the previous cycle, using our custom smart pointers like 'Arc' will eventually only rely in stable Rust. - Introduce 'PROCMACROLDFLAGS' environment variable to allow to link Rust proc macros using different flags than those used for linking Rust host programs (e.g. when 'rustc' uses a different C library than the host programs' one), which Android needs. - Help kernel builds under macOS with Rust enabled by accomodating other naming conventions for dynamic libraries (i.e. '.so' vs. '.dylib') which are used for Rust procedural macros. The actual support for macOS (i.e. the rest of the pieces needed) is provided out-of-tree by others, following the policy used for other parts of the kernel by Kbuild. - Run Clippy for 'rusttest' code too and clean the bits it spotted. - Provide Clippy with the minimum supported Rust version to improve the suggestions it gives. - Document 'bindgen' 0.71.0 regression. 'kernel' crate: - 'build_error!': move users of the hidden function to the documented macro, prevent such uses in the future by moving the function elsewhere and add the macro to the prelude. - 'types' module: add improved version of 'ForeignOwnable::borrow_mut' (which was removed in the past since it was problematic); change 'ForeignOwnable' pointer type to '*mut'. - 'alloc' module: implement 'Display' for 'Box' and align the 'Debug' implementation to it; add example (doctest) for 'ArrayLayout::new()' - 'sync' module: document 'PhantomData' in 'Arc'; use 'NonNull::new_unchecked' in 'ForeignOwnable for Arc' impl. - 'uaccess' module: accept 'Vec's with different allocators in 'UserSliceReader::read_all'. - 'workqueue' module: enable run-testing a couple more doctests. - 'error' module: simplify 'from_errno()'. - 'block' module: fix formatting in code documentation (a lint to catch these is being implemented). - Avoid 'unwrap()'s in doctests, which also improves the examples by showing how kernel code is supposed to be written. - Avoid 'as' casts with 'cast{,_mut}' calls which are a bit safer. And a few other cleanups" * tag 'rust-6.14' of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux: (32 commits) kbuild: rust: add PROCMACROLDFLAGS rust: uaccess: generalize userSliceReader to support any Vec rust: kernel: add improved version of `ForeignOwnable::borrow_mut` rust: kernel: reorder `ForeignOwnable` items rust: kernel: change `ForeignOwnable` pointer to mut rust: arc: split unsafe block, add missing comment rust: types: avoid `as` casts rust: arc: use `NonNull::new_unchecked` rust: use derive(CoercePointee) on rustc >= 1.84.0 rust: alloc: add doctest for `ArrayLayout::new()` rust: init: update `stack_try_pin_init` examples rust: error: import `kernel`'s `LayoutError` instead of `core`'s rust: str: replace unwraps with question mark operators rust: page: remove unnecessary helper function from doctest rust: rbtree: remove unwrap in asserts rust: init: replace unwraps with question mark operators rust: use host dylib naming convention to support macOS rust: add `build_error!` to the prelude rust: kernel: move `build_error` hidden function to prevent mistakes rust: use the `build_error!` macro, not the hidden function ...

rust: workqueue: Enable execution of doctests

2024-12-17T23:37:21Z

Having the Rust doctests enabled these workqueue tests are built but not executed as the final callers of the print_*() functions are missing. Add them. The result is # rust_doctest_kernel_workqueue_rs_0.location: rust/kernel/workqueue.rs:35 rust_doctests_kernel: The value is: 42 ok 94 rust_doctest_kernel_workqueue_rs_0 # rust_doctest_kernel_workqueue_rs_3.location: rust/kernel/workqueue.rs:78 rust_doctests_kernel: The value is: 24 rust_doctests_kernel: The second value is: 42 ok 97 rust_doctest_kernel_workqueue_rs_3 Without this change the "The value ..." outputs are not there meaning that this test code is not run. Reviewed-by: Alice Ryhl Signed-off-by: Dirk Behme Reviewed-by: Boqun Feng Acked-by: Tejun Heo Link: https://lore.kernel.org/r/cb953202-0dbe-4127-8a8e-6a75258c2116@gmail.com [ Reworded slightly. - Miguel ] Signed-off-by: Miguel Ojeda

rust: add safety comment in workqueue traits

2024-12-03T20:47:58Z

Add missing safety comments for the implementation of the unsafe traits WorkItemPointer and RawWorkItem for Arc in workqueue.rs Link: https://github.com/Rust-for-Linux/linux/issues/351. Co-developed-by: Vangelis Mamalakis Signed-off-by: Vangelis Mamalakis Suggested-by: Miguel Ojeda Reviewed-by: Alice Ryhl Signed-off-by: Konstantin Andrikopoulos Signed-off-by: Tejun Heo

rust: treewide: switch to our kernel `Box` type

2024-10-15T20:56:59Z

Now that we got the kernel `Box` type in place, convert all existing `Box` users to make use of it. Reviewed-by: Alice Ryhl Reviewed-by: Benno Lossin Reviewed-by: Gary Guo Signed-off-by: Danilo Krummrich Link: https://lore.kernel.org/r/20241004154149.93856-13-dakr@kernel.org Signed-off-by: Miguel Ojeda

rust: enable `clippy::unnecessary_safety_comment` lint

2024-10-07T19:39:05Z

In Rust 1.67.0, Clippy added the `unnecessary_safety_comment` lint [1], which is the "inverse" of `undocumented_unsafe_blocks`: it finds places where safe code has a `// SAFETY` comment attached. The lint currently finds 3 places where we had such mistakes, thus it seems already quite useful. Thus clean those and enable it. Link: https://rust-lang.github.io/rust-clippy/master/index.html#/unnecessary_safety_comment [1] Reviewed-by: Alice Ryhl Reviewed-by: Trevor Gross Reviewed-by: Vincenzo Palazzo Tested-by: Gary Guo Reviewed-by: Gary Guo Link: https://lore.kernel.org/r/20240904204347.168520-6-ojeda@kernel.org Signed-off-by: Miguel Ojeda

rust: enable `clippy::undocumented_unsafe_blocks` lint

2024-10-07T19:39:05Z

Checking that we are not missing any `// SAFETY` comments in our `unsafe` blocks is something we have wanted to do for a long time, as well as cleaning up the remaining cases that were not documented [1]. Back when Rust for Linux started, this was something that could have been done via a script, like Rust's `tidy`. Soon after, in Rust 1.58.0, Clippy implemented the `undocumented_unsafe_blocks` lint [2]. Even though the lint has a few false positives, e.g. in some cases where attributes appear between the comment and the `unsafe` block [3], there are workarounds and the lint seems quite usable already. Thus enable the lint now. We still have a few cases to clean up, so just allow those for the moment by writing a `TODO` comment -- some of those may be good candidates for new contributors. Link: https://github.com/Rust-for-Linux/linux/issues/351 [1] Link: https://rust-lang.github.io/rust-clippy/master/#/undocumented_unsafe_blocks [2] Link: https://github.com/rust-lang/rust-clippy/issues/13189 [3] Reviewed-by: Alice Ryhl Reviewed-by: Trevor Gross Tested-by: Gary Guo Reviewed-by: Gary Guo Link: https://lore.kernel.org/r/20240904204347.168520-5-ojeda@kernel.org Signed-off-by: Miguel Ojeda

rust: workqueue: remove unneeded ``#[allow(clippy::new_ret_no_self)]`

2024-10-07T08:49:15Z

Perform the same clean commit b2516f7af9d2 ("rust: kernel: remove `#[allow(clippy::new_ret_no_self)]`") did for a case that appeared in workqueue in parallel in commit 7324b88975c5 ("rust: workqueue: add helper for defining work_struct fields"): Clippy triggered a false positive on its `new_ret_no_self` lint when using the `pin_init!` macro. Since Rust 1.67.0, that does not happen anymore, since Clippy learnt to not warn about `-> impl Trait` [1][2]. The kernel nowadays uses Rust 1.72.1, thus remove the `#[allow]`. Link: https://github.com/rust-lang/rust-clippy/issues/7344 [1] Link: https://github.com/rust-lang/rust-clippy/pull/9733 [2] Reviewed-by: Alice Ryhl Reviewed-by: Trevor Gross Tested-by: Gary Guo Reviewed-by: Gary Guo Link: https://lore.kernel.org/r/20240904204347.168520-2-ojeda@kernel.org Signed-off-by: Miguel Ojeda

rust: kernel: make impl_has_work compatible with more generics

2024-07-08T20:39:39Z

Make the impl_has_work macro compatible with more complex generics such as lifetimes and const generic arguments. Signed-off-by: Roland Xu Link: https://lore.kernel.org/r/ME0P282MB4890A180B99490CC65EF64FDCCEB2@ME0P282MB4890.AUSP282.PROD.OUTLOOK.COM Suggested-by: Benno Lossin Link: https://github.com/Rust-for-Linux/linux/issues/1077 [ Wrapped message to 72 columns. - Miguel ] Signed-off-by: Miguel Ojeda

rust: kernel: remove redundant imports

2024-05-05T17:22:25Z

Rust's `unused_imports` lint covers both unused and redundant imports. In the upcoming 1.78.0, the lint detects more cases of redundant imports [1], e.g.: error: the item `bindings` is imported redundantly --> rust/kernel/print.rs:38:9 | 38 | use crate::bindings; | ^^^^^^^^^^^^^^^ the item `bindings` is already defined by prelude Most cases are `use crate::bindings`, plus a few other items like `Box`. Thus clean them up. Note that, in the `bindings` case, the message "defined by prelude" above means the extern prelude, i.e. the `--extern` flags we pass. Link: https://github.com/rust-lang/rust/pull/117772 [1] Reviewed-by: Alice Ryhl Link: https://lore.kernel.org/r/20240401212303.537355-3-ojeda@kernel.org Signed-off-by: Miguel Ojeda