Add atomic xchg and cmpxchg operation support for i8 and i16 types
with tests.

Note that since the current implementation of
Atomic::<{i8,i16}>::{load,store}() is READ_ONCE()/WRITE_ONCE()-based.
The atomicity between load/store and xchg/cmpxchg is only guaranteed if
the architecture has native RmW support, hence i8/i16 is currently
AtomicImpl only when CONFIG_ARCH_SUPPORTS_ATOMIC_RWM=y.

[boqun: Make i8/i16 AtomicImpl only when
CONFIG_ARCH_SUPPORTS_ATOMIC_RWM=y]

Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Link: https://patch.msgid.link/20251228120546.1602275-4-fujita.tomonori@gmail.com

Add atomic operation support for i8 and i16 types using volatile
read/write and smp_load_acquire/smp_store_release helpers.

[boqun: Adjust [1] to avoid introduction of
impl_atomic_only_load_and_store_ops!() in the middle]

Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Reviewed-by: Gary Guo <gary@garyguo.net>
Reviewed-by: Joel Fernandes <joelagnelf@nvidia.com>
Link: https://lore.kernel.org/all/20251228120546.1602275-1-fujita.tomonori@gmail.com/ [1]
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Link: https://patch.msgid.link/20251211113826.1299077-4-fujita.tomonori@gmail.com

Rework the internal AtomicOps macro plumbing to generate per-type
implementations from a mapping list.

Capture the trait definition once and reuse it for both declaration
and per-type impl expansion to reduce duplication and keep future
extensions simple.

This is a preparatory refactor for enabling i8/i16 atomics cleanly.

Signed-off-by: FUJITA Tomonori <fujita.tomonori@gmail.com>
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Link: https://patch.msgid.link/20251228120546.1602275-2-fujita.tomonori@gmail.com

Preparation for generic atomic implementation. To unify the
implementation of a generic method over `i32` and `i64`, the C side
atomic methods need to be grouped so that in a generic method, they can
be referred as <type>::<method>, otherwise their parameters and return
value are different between `i32` and `i64`, which would require using
`transmute()` to unify the type into a `T`.

Introduce `AtomicImpl` to represent a basic type in Rust that has the
direct mapping to an atomic implementation from C. Use a sealed trait to
restrict `AtomicImpl` to only support `i32` and `i64` for now.

Further, different methods are put into different `*Ops` trait groups,
and this is for the future when smaller types like `i8`/`i16` are
supported but only with a limited set of API (e.g. only set(), load(),
xchg() and cmpxchg(), no add() or sub() etc).

While the atomic mod is introduced, documentation is also added for
memory models and data races.

Also bump my role to the maintainer of ATOMIC INFRASTRUCTURE to reflect
my responsibility on the Rust atomic mod.

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/20250719030827.61357-3-boqun.feng@gmail.com/