Since we've exposed Lock::from_raw() and Guard::new() publically, we
want to be able to make sure that we assert that a lock is actually held
when constructing a Guard for it to handle instances of unsafe
Guard::new() calls outside of our lock module.
Hence add a new method assert_is_held() to Backend, which uses lockdep
to check whether or not a lock has been acquired. When lockdep is
disabled, this has no overhead.
[Boqun: Resolve the conflicts with exposing Guard::new(), reword the
commit log a bit and format "unsafe { <statement>; }" into "unsafe {
<statement> }" for the consistency. ]
Signed-off-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Link: https://lore.kernel.org/r/20241125204139.656801-1-lyude@redhat.com
Since we added a `Lock::from_raw()` function previously, it makes sense to also introduce an interface for creating a `Guard` from a reference to a `Lock` for instances where we've derived the `Lock` from a raw pointer and know that the lock is already acquired, there are such usages in KMS API. [Boqun: Add backquotes to type names, reformat the commit log, reword a bit on the usage of KMS API] Signed-off-by: Lyude Paul <lyude@redhat.com> Reviewed-by: Filipe Xavier <felipe_life@live.com> Reviewed-by: Alice Ryhl <aliceryhl@google.com> Signed-off-by: Boqun Feng <boqun.feng@gmail.com> Link: https://lore.kernel.org/r/20241119231146.2298971-3-lyude@redhat.com
The KMS bindings [1] have a few bindings that require manually acquiring specific locks before calling certain functions. At the moment though, the only way of acquiring these locks in bindings is to simply call the C locking functions directly - since said locks are not initialized on the Rust side of things. However - if we add `#[repr(C)]` to `Lock<(), B>`, then given `()` is a ZST - `Lock<(), B>` becomes equivalent in data layout to its inner `B::State` type. Since locks in C don't have data explicitly associated with them anyway, we can take advantage of this to add a `Lock::from_raw()` function that can translate a raw pointer to `B::State` into its proper `Lock<(), B>` equivalent. This lets us simply acquire a reference to the lock in question and work with it like it was initialized on the Rust side of things, allowing us to use less unsafe code to implement bindings with lock requirements. [Boqun: Use "Link:" instead of a URL and format the commit log] Signed-off-by: Lyude Paul <lyude@redhat.com> Reviewed-by: Alice Ryhl <aliceryhl@google.com> Link: https://patchwork.freedesktop.org/series/131522/ [1] Signed-off-by: Boqun Feng <boqun.feng@gmail.com> Link: https://lore.kernel.org/r/20241119231146.2298971-2-lyude@redhat.com
Pull rust updates from Miguel Ojeda:
"Toolchain and infrastructure:
- Enable a series of lints, including safety-related ones, e.g. the
compiler will now warn about missing safety comments, as well as
unnecessary ones. How safety documentation is organized is a
frequent source of review comments, thus having the compiler guide
new developers on where they are expected (and where not) is very
nice.
- Start using '#[expect]': an interesting feature in Rust (stabilized
in 1.81.0) that makes the compiler warn if an expected warning was
_not_ emitted. This is useful to avoid forgetting cleaning up
locally ignored diagnostics ('#[allow]'s).
- Introduce '.clippy.toml' configuration file for Clippy, the Rust
linter, which will allow us to tweak its behaviour. For instance,
our first use cases are declaring a disallowed macro and, more
importantly, enabling the checking of private items.
- Lints-related fixes and cleanups related to the items above.
- Migrate from 'receiver_trait' to 'arbitrary_self_types': to get the
kernel into stable Rust, one of the major pieces of the puzzle is
the support to write custom types that can be used as 'self', i.e.
as receivers, since the kernel needs to write types such as 'Arc'
that common userspace Rust would not. 'arbitrary_self_types' has
been accepted to become stable, and this is one of the steps
required to get there.
- Remove usage of the 'new_uninit' unstable feature.
- Use custom C FFI types. Includes a new 'ffi' crate to contain our
custom mapping, instead of using the standard library 'core::ffi'
one. The actual remapping will be introduced in a later cycle.
- Map '__kernel_{size_t,ssize_t,ptrdiff_t}' to 'usize'/'isize'
instead of 32/64-bit integers.
- Fix 'size_t' in bindgen generated prototypes of C builtins.
- Warn on bindgen < 0.69.5 and libclang >= 19.1 due to a double issue
in the projects, which we managed to trigger with the upcoming
tracepoint support. It includes a build test since some
distributions backported the fix (e.g. Debian -- thanks!). All
major distributions we list should be now OK except Ubuntu non-LTS.
'macros' crate:
- Adapt the build system to be able run the doctests there too; and
clean up and enable the corresponding doctests.
'kernel' crate:
- Add 'alloc' module with generic kernel allocator support and remove
the dependency on the Rust standard library 'alloc' and the
extension traits we used to provide fallible methods with flags.
Add the 'Allocator' trait and its implementations '{K,V,KV}malloc'.
Add the 'Box' type (a heap allocation for a single value of type
'T' that is also generic over an allocator and considers the
kernel's GFP flags) and its shorthand aliases '{K,V,KV}Box'. Add
'ArrayLayout' type. Add 'Vec' (a contiguous growable array type)
and its shorthand aliases '{K,V,KV}Vec', including iterator
support.
For instance, now we may write code such as:
let mut v = KVec::new();
v.push(1, GFP_KERNEL)?;
assert_eq!(&v, &[1]);
Treewide, move as well old users to these new types.
- 'sync' module: add global lock support, including the
'GlobalLockBackend' trait; the 'Global{Lock,Guard,LockedBy}' types
and the 'global_lock!' macro. Add the 'Lock::try_lock' method.
- 'error' module: optimize 'Error' type to use 'NonZeroI32' and make
conversion functions public.
- 'page' module: add 'page_align' function.
- Add 'transmute' module with the existing 'FromBytes' and 'AsBytes'
traits.
- 'block::mq::request' module: improve rendered documentation.
- 'types' module: extend 'Opaque' type documentation and add simple
examples for the 'Either' types.
drm/panic:
- Clean up a series of Clippy warnings.
Documentation:
- Add coding guidelines for lints and the '#[expect]' feature.
- Add Ubuntu to the list of distributions in the Quick Start guide.
MAINTAINERS:
- Add Danilo Krummrich as maintainer of the new 'alloc' module.
And a few other small cleanups and fixes"
* tag 'rust-6.13' of https://github.com/Rust-for-Linux/linux: (82 commits)
rust: alloc: Fix `ArrayLayout` allocations
docs: rust: remove spurious item in `expect` list
rust: allow `clippy::needless_lifetimes`
rust: warn on bindgen < 0.69.5 and libclang >= 19.1
rust: use custom FFI integer types
rust: map `__kernel_size_t` and friends also to usize/isize
rust: fix size_t in bindgen prototypes of C builtins
rust: sync: add global lock support
rust: macros: enable the rest of the tests
rust: macros: enable paste! use from macro_rules!
rust: enable macros::module! tests
rust: kbuild: expand rusttest target for macros
rust: types: extend `Opaque` documentation
rust: block: fix formatting of `kernel::block::mq::request` module
rust: macros: fix documentation of the paste! macro
rust: kernel: fix THIS_MODULE header path in ThisModule doc comment
rust: page: add Rust version of PAGE_ALIGN
rust: helpers: remove unnecessary header includes
rust: exports: improve grammar in commentary
drm/panic: allow verbose version check
...
Currently FFI integer types are defined in libcore. This commit creates the `ffi` crate and asks bindgen to use that crate for FFI integer types instead of `core::ffi`. This commit is preparatory and no type changes are made in this commit yet. Signed-off-by: Gary Guo <gary@garyguo.net> Link: https://lore.kernel.org/r/20240913213041.395655-4-gary@garyguo.net [ Added `rustdoc`, `rusttest` and KUnit tests support. Rebased on top of `rust-next` (e.g. migrated more `core::ffi` cases). Reworded crate docs slightly and formatted. - Miguel ] Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Add support for creating global variables that are wrapped in a mutex or spinlock. The implementation here is intended to replace the global mutex workaround found in the Rust Binder RFC [1]. In both cases, the global lock must be initialized before first use. The macro is unsafe to use for the same reason. The separate initialization step is required because it is tricky to access the value of __ARCH_SPIN_LOCK_UNLOCKED from Rust. Doing so will require changes to the C side. That change will happen as a follow-up to this patch. Link: https://lore.kernel.org/rust-for-linux/20231101-rust-binder-v1-2-08ba9197f637@google.com/#Z31drivers:android:context.rs [1] Signed-off-by: Alice Ryhl <aliceryhl@google.com> Reviewed-by: Boqun Feng <boqun.feng@gmail.com> Link: https://lore.kernel.org/r/20241023-static-mutex-v6-1-d7efdadcc84f@google.com [ Simplified a few intra-doc links. Formatted a few comments. Reworded title. - Miguel ] Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Add a non-blocking trylock method to lock backend interface, mutex and spinlock implementations. It includes a C helper for spin_trylock. Rust Binder will use this method together with the new shrinker abstractions to avoid deadlocks in the memory shrinker. Link: https://lore.kernel.org/all/20240912-shrinker-v1-1-18b7f1253553@google.com Signed-off-by: Filipe Xavier <felipe_life@live.com> Reviewed-by: Fiona Behrens <me@kloenk.dev> Reviewed-by: Alice Ryhl <aliceryhl@google.com> Reviewed-by: Boqun Feng <boqun.feng@gmail.com> Link: https://lore.kernel.org/r/BL0PR02MB4914579914884B5D7473B3D6E96A2@BL0PR02MB4914.namprd02.prod.outlook.com [ Slightly reworded. - Miguel ] Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
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 <aliceryhl@google.com> Reviewed-by: Trevor Gross <tmgross@umich.edu> Tested-by: Gary Guo <gary@garyguo.net> Reviewed-by: Gary Guo <gary@garyguo.net> Link: https://lore.kernel.org/r/20240904204347.168520-5-ojeda@kernel.org Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
This introduces a new marker type for types that shouldn't be thread safe. By adding a field of this type to a struct, it becomes non-Send and non-Sync, which means that it cannot be accessed in any way from threads other than the one it was created on. This is useful for APIs that require globals such as `current` to remain constant while the value exists. We update two existing users in the Kernel to use this helper: * `Task::current()` - moving the return type of this value to a different thread would not be safe as you can no longer be guaranteed that the `current` pointer remains valid. * Lock guards. Mutexes and spinlocks should be unlocked on the same thread as where they were locked, so we enforce this using the Send trait. There are also additional users in later patches of this patchset. See [1] and [2] for the discussion that led to the introduction of this patch. Link: https://lore.kernel.org/all/nFDPJFnzE9Q5cqY7FwSMByRH2OAn_BpI4H53NQfWIlN6I2qfmAqnkp2wRqn0XjMO65OyZY4h6P4K2nAGKJpAOSzksYXaiAK_FoH_8QbgBI4=@proton.me/ [1] Link: https://lore.kernel.org/all/nFDPJFnzE9Q5cqY7FwSMByRH2OAn_BpI4H53NQfWIlN6I2qfmAqnkp2wRqn0XjMO65OyZY4h6P4K2nAGKJpAOSzksYXaiAK_FoH_8QbgBI4=@proton.me/ [2] Suggested-by: Benno Lossin <benno.lossin@proton.me> Reviewed-by: Benno Lossin <benno.lossin@proton.me> Reviewed-by: Trevor Gross <tmgross@umich.edu> Reviewed-by: Martin Rodriguez Reboredo <yakoyoku@gmail.com> Reviewed-by: Björn Roy Baron <bjorn3_gh@protonmail.com> Reviewed-by: Gary Guo <gary@garyguo.net> Signed-off-by: Alice Ryhl <aliceryhl@google.com> Link: https://lore.kernel.org/r/20240915-alice-file-v10-1-88484f7a3dcf@google.com Signed-off-by: Christian Brauner <brauner@kernel.org>
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 <aliceryhl@google.com>
Link: https://lore.kernel.org/r/20240401212303.537355-3-ojeda@kernel.org
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>