linux/rust/kernel/sync/atomic.rs
Boqun Feng d132054360 rust: sync: atomic: Add the framework of arithmetic operations
One important set of atomic operations is the arithmetic operations,
i.e. add(), sub(), fetch_add(), add_return(), etc. However it may not
make senses for all the types that `AtomicType` to have arithmetic
operations, for example a `Foo(u32)` may not have a reasonable add() or
sub(), plus subword types (`u8` and `u16`) currently don't have
atomic arithmetic operations even on C side and might not have them in
the future in Rust (because they are usually suboptimal on a few
architecures). Therefore the plan is to add a few subtraits of
`AtomicType` describing which types have and can do atomic arithemtic
operations.

One trait `AtomicAdd` is added, and only add() and fetch_add() are
added. The rest will be added in the future.

Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Reviewed-by: Benno Lossin <lossin@kernel.org>
Reviewed-by: Elle Rhumsaa <elle@weathered-steel.dev>
Link: https://lore.kernel.org/all/20250719030827.61357-7-boqun.feng@gmail.com/
2025-09-15 09:38:34 +02:00

551 lines
19 KiB
Rust

// SPDX-License-Identifier: GPL-2.0
//! Atomic primitives.
//!
//! These primitives have the same semantics as their C counterparts: and the precise definitions of
//! semantics can be found at [`LKMM`]. Note that Linux Kernel Memory (Consistency) Model is the
//! only model for Rust code in kernel, and Rust's own atomics should be avoided.
//!
//! # Data races
//!
//! [`LKMM`] atomics have different rules regarding data races:
//!
//! - A normal write from C side is treated as an atomic write if
//! CONFIG_KCSAN_ASSUME_PLAIN_WRITES_ATOMIC=y.
//! - Mixed-size atomic accesses don't cause data races.
//!
//! [`LKMM`]: srctree/tools/memory-model/
mod internal;
pub mod ordering;
mod predefine;
pub use internal::AtomicImpl;
pub use ordering::{Acquire, Full, Relaxed, Release};
use crate::build_error;
use internal::{AtomicArithmeticOps, AtomicBasicOps, AtomicExchangeOps, AtomicRepr};
use ordering::OrderingType;
/// A memory location which can be safely modified from multiple execution contexts.
///
/// This has the same size, alignment and bit validity as the underlying type `T`. And it disables
/// niche optimization for the same reason as [`UnsafeCell`].
///
/// The atomic operations are implemented in a way that is fully compatible with the [Linux Kernel
/// Memory (Consistency) Model][LKMM], hence they should be modeled as the corresponding
/// [`LKMM`][LKMM] atomic primitives. With the help of [`Atomic::from_ptr()`] and
/// [`Atomic::as_ptr()`], this provides a way to interact with [C-side atomic operations]
/// (including those without the `atomic` prefix, e.g. `READ_ONCE()`, `WRITE_ONCE()`,
/// `smp_load_acquire()` and `smp_store_release()`).
///
/// # Invariants
///
/// `self.0` is a valid `T`.
///
/// [`UnsafeCell`]: core::cell::UnsafeCell
/// [LKMM]: srctree/tools/memory-model/
/// [C-side atomic operations]: srctree/Documentation/atomic_t.txt
#[repr(transparent)]
pub struct Atomic<T: AtomicType>(AtomicRepr<T::Repr>);
// SAFETY: `Atomic<T>` is safe to share among execution contexts because all accesses are atomic.
unsafe impl<T: AtomicType> Sync for Atomic<T> {}
/// Types that support basic atomic operations.
///
/// # Round-trip transmutability
///
/// `T` is round-trip transmutable to `U` if and only if both of these properties hold:
///
/// - Any valid bit pattern for `T` is also a valid bit pattern for `U`.
/// - Transmuting (e.g. using [`transmute()`]) a value of type `T` to `U` and then to `T` again
/// yields a value that is in all aspects equivalent to the original value.
///
/// # Safety
///
/// - [`Self`] must have the same size and alignment as [`Self::Repr`].
/// - [`Self`] must be [round-trip transmutable] to [`Self::Repr`].
///
/// Note that this is more relaxed than requiring the bi-directional transmutability (i.e.
/// [`transmute()`] is always sound between `U` and `T`) because of the support for atomic
/// variables over unit-only enums, see [Examples].
///
/// # Limitations
///
/// Because C primitives are used to implement the atomic operations, and a C function requires a
/// valid object of a type to operate on (i.e. no `MaybeUninit<_>`), hence at the Rust <-> C
/// surface, only types with all the bits initialized can be passed. As a result, types like `(u8,
/// u16)` (padding bytes are uninitialized) are currently not supported.
///
/// # Examples
///
/// A unit-only enum that implements [`AtomicType`]:
///
/// ```
/// use kernel::sync::atomic::{AtomicType, Atomic, Relaxed};
///
/// #[derive(Clone, Copy, PartialEq, Eq)]
/// #[repr(i32)]
/// enum State {
/// Uninit = 0,
/// Working = 1,
/// Done = 2,
/// };
///
/// // SAFETY: `State` and `i32` has the same size and alignment, and it's round-trip
/// // transmutable to `i32`.
/// unsafe impl AtomicType for State {
/// type Repr = i32;
/// }
///
/// let s = Atomic::new(State::Uninit);
///
/// assert_eq!(State::Uninit, s.load(Relaxed));
/// ```
/// [`transmute()`]: core::mem::transmute
/// [round-trip transmutable]: AtomicType#round-trip-transmutability
/// [Examples]: AtomicType#examples
pub unsafe trait AtomicType: Sized + Send + Copy {
/// The backing atomic implementation type.
type Repr: AtomicImpl;
}
/// Types that support atomic add operations.
///
/// # Safety
///
// TODO: Properly defines `wrapping_add` in the following comment.
/// `wrapping_add` any value of type `Self::Repr::Delta` obtained by [`Self::rhs_into_delta()`] to
/// any value of type `Self::Repr` obtained through transmuting a value of type `Self` to must
/// yield a value with a bit pattern also valid for `Self`.
pub unsafe trait AtomicAdd<Rhs = Self>: AtomicType {
/// Converts `Rhs` into the `Delta` type of the atomic implementation.
fn rhs_into_delta(rhs: Rhs) -> <Self::Repr as AtomicImpl>::Delta;
}
#[inline(always)]
const fn into_repr<T: AtomicType>(v: T) -> T::Repr {
// SAFETY: Per the safety requirement of `AtomicType`, `T` is round-trip transmutable to
// `T::Repr`, therefore the transmute operation is sound.
unsafe { core::mem::transmute_copy(&v) }
}
/// # Safety
///
/// `r` must be a valid bit pattern of `T`.
#[inline(always)]
const unsafe fn from_repr<T: AtomicType>(r: T::Repr) -> T {
// SAFETY: Per the safety requirement of the function, the transmute operation is sound.
unsafe { core::mem::transmute_copy(&r) }
}
impl<T: AtomicType> Atomic<T> {
/// Creates a new atomic `T`.
pub const fn new(v: T) -> Self {
// INVARIANT: Per the safety requirement of `AtomicType`, `into_repr(v)` is a valid `T`.
Self(AtomicRepr::new(into_repr(v)))
}
/// Creates a reference to an atomic `T` from a pointer of `T`.
///
/// This usually is used when communicating with C side or manipulating a C struct, see
/// examples below.
///
/// # Safety
///
/// - `ptr` is aligned to `align_of::<T>()`.
/// - `ptr` is valid for reads and writes for `'a`.
/// - For the duration of `'a`, other accesses to `*ptr` must not cause data races (defined
/// by [`LKMM`]) against atomic operations on the returned reference. Note that if all other
/// accesses are atomic, then this safety requirement is trivially fulfilled.
///
/// [`LKMM`]: srctree/tools/memory-model
///
/// # Examples
///
/// Using [`Atomic::from_ptr()`] combined with [`Atomic::load()`] or [`Atomic::store()`] can
/// achieve the same functionality as `READ_ONCE()`/`smp_load_acquire()` or
/// `WRITE_ONCE()`/`smp_store_release()` in C side:
///
/// ```
/// # use kernel::types::Opaque;
/// use kernel::sync::atomic::{Atomic, Relaxed, Release};
///
/// // Assume there is a C struct `foo`.
/// mod cbindings {
/// #[repr(C)]
/// pub(crate) struct foo {
/// pub(crate) a: i32,
/// pub(crate) b: i32
/// }
/// }
///
/// let tmp = Opaque::new(cbindings::foo { a: 1, b: 2 });
///
/// // struct foo *foo_ptr = ..;
/// let foo_ptr = tmp.get();
///
/// // SAFETY: `foo_ptr` is valid, and `.a` is in bounds.
/// let foo_a_ptr = unsafe { &raw mut (*foo_ptr).a };
///
/// // a = READ_ONCE(foo_ptr->a);
/// //
/// // SAFETY: `foo_a_ptr` is valid for read, and all other accesses on it is atomic, so no
/// // data race.
/// let a = unsafe { Atomic::from_ptr(foo_a_ptr) }.load(Relaxed);
/// # assert_eq!(a, 1);
///
/// // smp_store_release(&foo_ptr->a, 2);
/// //
/// // SAFETY: `foo_a_ptr` is valid for writes, and all other accesses on it is atomic, so
/// // no data race.
/// unsafe { Atomic::from_ptr(foo_a_ptr) }.store(2, Release);
/// ```
pub unsafe fn from_ptr<'a>(ptr: *mut T) -> &'a Self
where
T: Sync,
{
// CAST: `T` and `Atomic<T>` have the same size, alignment and bit validity.
// SAFETY: Per function safety requirement, `ptr` is a valid pointer and the object will
// live long enough. It's safe to return a `&Atomic<T>` because function safety requirement
// guarantees other accesses won't cause data races.
unsafe { &*ptr.cast::<Self>() }
}
/// Returns a pointer to the underlying atomic `T`.
///
/// Note that use of the return pointer must not cause data races defined by [`LKMM`].
///
/// # Guarantees
///
/// The returned pointer is valid and properly aligned (i.e. aligned to [`align_of::<T>()`]).
///
/// [`LKMM`]: srctree/tools/memory-model
/// [`align_of::<T>()`]: core::mem::align_of
pub const fn as_ptr(&self) -> *mut T {
// GUARANTEE: Per the function guarantee of `AtomicRepr::as_ptr()`, the `self.0.as_ptr()`
// must be a valid and properly aligned pointer for `T::Repr`, and per the safety guarantee
// of `AtomicType`, it's a valid and properly aligned pointer of `T`.
self.0.as_ptr().cast()
}
/// Returns a mutable reference to the underlying atomic `T`.
///
/// This is safe because the mutable reference of the atomic `T` guarantees exclusive access.
pub fn get_mut(&mut self) -> &mut T {
// CAST: `T` and `T::Repr` has the same size and alignment per the safety requirement of
// `AtomicType`, and per the type invariants `self.0` is a valid `T`, therefore the casting
// result is a valid pointer of `T`.
// SAFETY: The pointer is valid per the CAST comment above, and the mutable reference
// guarantees exclusive access.
unsafe { &mut *self.0.as_ptr().cast() }
}
}
impl<T: AtomicType> Atomic<T>
where
T::Repr: AtomicBasicOps,
{
/// Loads the value from the atomic `T`.
///
/// # Examples
///
/// ```
/// use kernel::sync::atomic::{Atomic, Relaxed};
///
/// let x = Atomic::new(42i32);
///
/// assert_eq!(42, x.load(Relaxed));
///
/// let x = Atomic::new(42i64);
///
/// assert_eq!(42, x.load(Relaxed));
/// ```
#[doc(alias("atomic_read", "atomic64_read"))]
#[inline(always)]
pub fn load<Ordering: ordering::AcquireOrRelaxed>(&self, _: Ordering) -> T {
let v = {
match Ordering::TYPE {
OrderingType::Relaxed => T::Repr::atomic_read(&self.0),
OrderingType::Acquire => T::Repr::atomic_read_acquire(&self.0),
_ => build_error!("Wrong ordering"),
}
};
// SAFETY: `v` comes from reading `self.0`, which is a valid `T` per the type invariants.
unsafe { from_repr(v) }
}
/// Stores a value to the atomic `T`.
///
/// # Examples
///
/// ```
/// use kernel::sync::atomic::{Atomic, Relaxed};
///
/// let x = Atomic::new(42i32);
///
/// assert_eq!(42, x.load(Relaxed));
///
/// x.store(43, Relaxed);
///
/// assert_eq!(43, x.load(Relaxed));
/// ```
#[doc(alias("atomic_set", "atomic64_set"))]
#[inline(always)]
pub fn store<Ordering: ordering::ReleaseOrRelaxed>(&self, v: T, _: Ordering) {
let v = into_repr(v);
// INVARIANT: `v` is a valid `T`, and is stored to `self.0` by `atomic_set*()`.
match Ordering::TYPE {
OrderingType::Relaxed => T::Repr::atomic_set(&self.0, v),
OrderingType::Release => T::Repr::atomic_set_release(&self.0, v),
_ => build_error!("Wrong ordering"),
}
}
}
impl<T: AtomicType> Atomic<T>
where
T::Repr: AtomicExchangeOps,
{
/// Atomic exchange.
///
/// Atomically updates `*self` to `v` and returns the old value of `*self`.
///
/// # Examples
///
/// ```
/// use kernel::sync::atomic::{Atomic, Acquire, Relaxed};
///
/// let x = Atomic::new(42);
///
/// assert_eq!(42, x.xchg(52, Acquire));
/// assert_eq!(52, x.load(Relaxed));
/// ```
#[doc(alias("atomic_xchg", "atomic64_xchg", "swap"))]
#[inline(always)]
pub fn xchg<Ordering: ordering::Ordering>(&self, v: T, _: Ordering) -> T {
let v = into_repr(v);
// INVARIANT: `self.0` is a valid `T` after `atomic_xchg*()` because `v` is transmutable to
// `T`.
let ret = {
match Ordering::TYPE {
OrderingType::Full => T::Repr::atomic_xchg(&self.0, v),
OrderingType::Acquire => T::Repr::atomic_xchg_acquire(&self.0, v),
OrderingType::Release => T::Repr::atomic_xchg_release(&self.0, v),
OrderingType::Relaxed => T::Repr::atomic_xchg_relaxed(&self.0, v),
}
};
// SAFETY: `ret` comes from reading `*self`, which is a valid `T` per type invariants.
unsafe { from_repr(ret) }
}
/// Atomic compare and exchange.
///
/// If `*self` == `old`, atomically updates `*self` to `new`. Otherwise, `*self` is not
/// modified.
///
/// Compare: The comparison is done via the byte level comparison between `*self` and `old`.
///
/// Ordering: When succeeds, provides the corresponding ordering as the `Ordering` type
/// parameter indicates, and a failed one doesn't provide any ordering, the load part of a
/// failed cmpxchg is a [`Relaxed`] load.
///
/// Returns `Ok(value)` if cmpxchg succeeds, and `value` is guaranteed to be equal to `old`,
/// otherwise returns `Err(value)`, and `value` is the current value of `*self`.
///
/// # Examples
///
/// ```
/// use kernel::sync::atomic::{Atomic, Full, Relaxed};
///
/// let x = Atomic::new(42);
///
/// // Checks whether cmpxchg succeeded.
/// let success = x.cmpxchg(52, 64, Relaxed).is_ok();
/// # assert!(!success);
///
/// // Checks whether cmpxchg failed.
/// let failure = x.cmpxchg(52, 64, Relaxed).is_err();
/// # assert!(failure);
///
/// // Uses the old value if failed, probably re-try cmpxchg.
/// match x.cmpxchg(52, 64, Relaxed) {
/// Ok(_) => { },
/// Err(old) => {
/// // do something with `old`.
/// # assert_eq!(old, 42);
/// }
/// }
///
/// // Uses the latest value regardlessly, same as atomic_cmpxchg() in C.
/// let latest = x.cmpxchg(42, 64, Full).unwrap_or_else(|old| old);
/// # assert_eq!(42, latest);
/// assert_eq!(64, x.load(Relaxed));
/// ```
///
/// [`Relaxed`]: ordering::Relaxed
#[doc(alias(
"atomic_cmpxchg",
"atomic64_cmpxchg",
"atomic_try_cmpxchg",
"atomic64_try_cmpxchg",
"compare_exchange"
))]
#[inline(always)]
pub fn cmpxchg<Ordering: ordering::Ordering>(
&self,
mut old: T,
new: T,
o: Ordering,
) -> Result<T, T> {
// Note on code generation:
//
// try_cmpxchg() is used to implement cmpxchg(), and if the helper functions are inlined,
// the compiler is able to figure out that branch is not needed if the users don't care
// about whether the operation succeeds or not. One exception is on x86, due to commit
// 44fe84459faf ("locking/atomic: Fix atomic_try_cmpxchg() semantics"), the
// atomic_try_cmpxchg() on x86 has a branch even if the caller doesn't care about the
// success of cmpxchg and only wants to use the old value. For example, for code like:
//
// let latest = x.cmpxchg(42, 64, Full).unwrap_or_else(|old| old);
//
// It will still generate code:
//
// movl $0x40, %ecx
// movl $0x34, %eax
// lock
// cmpxchgl %ecx, 0x4(%rsp)
// jne 1f
// 2:
// ...
// 1: movl %eax, %ecx
// jmp 2b
//
// This might be "fixed" by introducing a try_cmpxchg_exclusive() that knows the "*old"
// location in the C function is always safe to write.
if self.try_cmpxchg(&mut old, new, o) {
Ok(old)
} else {
Err(old)
}
}
/// Atomic compare and exchange and returns whether the operation succeeds.
///
/// If `*self` == `old`, atomically updates `*self` to `new`. Otherwise, `*self` is not
/// modified, `*old` is updated to the current value of `*self`.
///
/// "Compare" and "Ordering" part are the same as [`Atomic::cmpxchg()`].
///
/// Returns `true` means the cmpxchg succeeds otherwise returns `false`.
#[inline(always)]
fn try_cmpxchg<Ordering: ordering::Ordering>(&self, old: &mut T, new: T, _: Ordering) -> bool {
let mut tmp = into_repr(*old);
let new = into_repr(new);
// INVARIANT: `self.0` is a valid `T` after `atomic_try_cmpxchg*()` because `new` is
// transmutable to `T`.
let ret = {
match Ordering::TYPE {
OrderingType::Full => T::Repr::atomic_try_cmpxchg(&self.0, &mut tmp, new),
OrderingType::Acquire => {
T::Repr::atomic_try_cmpxchg_acquire(&self.0, &mut tmp, new)
}
OrderingType::Release => {
T::Repr::atomic_try_cmpxchg_release(&self.0, &mut tmp, new)
}
OrderingType::Relaxed => {
T::Repr::atomic_try_cmpxchg_relaxed(&self.0, &mut tmp, new)
}
}
};
// SAFETY: `tmp` comes from reading `*self`, which is a valid `T` per type invariants.
*old = unsafe { from_repr(tmp) };
ret
}
}
impl<T: AtomicType> Atomic<T>
where
T::Repr: AtomicArithmeticOps,
{
/// Atomic add.
///
/// Atomically updates `*self` to `(*self).wrapping_add(v)`.
///
/// # Examples
///
/// ```
/// use kernel::sync::atomic::{Atomic, Relaxed};
///
/// let x = Atomic::new(42);
///
/// assert_eq!(42, x.load(Relaxed));
///
/// x.add(12, Relaxed);
///
/// assert_eq!(54, x.load(Relaxed));
/// ```
#[inline(always)]
pub fn add<Rhs>(&self, v: Rhs, _: ordering::Relaxed)
where
T: AtomicAdd<Rhs>,
{
let v = T::rhs_into_delta(v);
// INVARIANT: `self.0` is a valid `T` after `atomic_add()` due to safety requirement of
// `AtomicAdd`.
T::Repr::atomic_add(&self.0, v);
}
/// Atomic fetch and add.
///
/// Atomically updates `*self` to `(*self).wrapping_add(v)`, and returns the value of `*self`
/// before the update.
///
/// # Examples
///
/// ```
/// use kernel::sync::atomic::{Atomic, Acquire, Full, Relaxed};
///
/// let x = Atomic::new(42);
///
/// assert_eq!(42, x.load(Relaxed));
///
/// assert_eq!(54, { x.fetch_add(12, Acquire); x.load(Relaxed) });
///
/// let x = Atomic::new(42);
///
/// assert_eq!(42, x.load(Relaxed));
///
/// assert_eq!(54, { x.fetch_add(12, Full); x.load(Relaxed) } );
/// ```
#[inline(always)]
pub fn fetch_add<Rhs, Ordering: ordering::Ordering>(&self, v: Rhs, _: Ordering) -> T
where
T: AtomicAdd<Rhs>,
{
let v = T::rhs_into_delta(v);
// INVARIANT: `self.0` is a valid `T` after `atomic_fetch_add*()` due to safety requirement
// of `AtomicAdd`.
let ret = {
match Ordering::TYPE {
OrderingType::Full => T::Repr::atomic_fetch_add(&self.0, v),
OrderingType::Acquire => T::Repr::atomic_fetch_add_acquire(&self.0, v),
OrderingType::Release => T::Repr::atomic_fetch_add_release(&self.0, v),
OrderingType::Relaxed => T::Repr::atomic_fetch_add_relaxed(&self.0, v),
}
};
// SAFETY: `ret` comes from reading `self.0`, which is a valid `T` per type invariants.
unsafe { from_repr(ret) }
}
}