Always provide atomic CAS for MSP430 and AVR

taiki-e · Aug 12, 2022 · 668c029 · 668c029
1 parent a4aec1e
commit 668c029
Show file tree

Hide file tree

Showing 7 changed files with 567 additions and 174 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -10,6 +10,10 @@ Note: In this file, do not use the hard wrap in the middle of a sentence for com
 
 ## [Unreleased]
 
+- Always provide atomic CAS for MSP430 and AVR.
+
+  This previously required unsafe cfg `portable_atomic_unsafe_assume_single_core`, but since all MSP430 and AVR are single-core, we can safely provide atomic CAS based on disabling interrupts.
+
 - Support `fence` and `compiler_fence` on MSP430.
 
 - Update safety requirements for unsafe cfg `portable_atomic_unsafe_assume_single_core` to mention use of privileged instructions to disable interrupts.

diff --git a/README.md b/README.md
@@ -14,7 +14,7 @@ Portable atomic types including support for 128-bit atomics, atomic float, etc.
 - Provide `AtomicF32` and `AtomicF64`. (optional)
 <!-- - Provide generic `Atomic<T>` type. (optional) -->
 - Provide atomic load/store for targets where atomic is not available at all in the standard library. (RISC-V without A-extension, MSP430, AVR)
-- Provide atomic CAS for targets where atomic CAS is not available in the standard library. (thumbv6m, RISC-V without A-extension, MSP430, AVR) (optional, [single-core only](#optional-cfg))
+- Provide atomic CAS for targets where atomic CAS is not available in the standard library. (thumbv6m, RISC-V without A-extension, MSP430, AVR) (optional and [single-core only](#optional-cfg) for ARM and RISC-V, always enabled for MSP430 and AVR)
 - Provide equivalents on the target that the standard library's atomic-related APIs cause LLVM errors. (fence/compiler_fence on MSP430)
 - Provide stable equivalents of the standard library atomic types' unstable APIs, such as [`AtomicPtr::fetch_*`](https://github.com/rust-lang/rust/issues/99108), [`AtomicBool::fetch_not`](https://github.com/rust-lang/rust/issues/98485).
 - Make features that require newer compilers, such as [fetch_max](https://doc.rust-lang.org/std/sync/atomic/struct.AtomicUsize.html#method.fetch_max), [fetch_min](https://doc.rust-lang.org/std/sync/atomic/struct.AtomicUsize.html#method.fetch_min), [fetch_update](https://doc.rust-lang.org/std/sync/atomic/struct.AtomicPtr.html#method.fetch_update), and [stronger CAS failure ordering](https://github.com/rust-lang/rust/pull/98383) available on Rust 1.34+.
@@ -79,7 +79,9 @@ See [this list](https://github.com/taiki-e/portable-atomic/issues/10#issuecommen
 
   Enabling this cfg for targets that have atomic CAS will result in a compile error.
 
-  ARMv6-M (thumbv6m), RISC-V without A-extension, MSP430, and AVR are currently supported. See [#26] for support of no-std pre-v6 ARM and multi-core systems.
+  ARMv6-M (thumbv6m), RISC-V without A-extension are currently supported. See [#26] for support of no-std pre-v6 ARM and multi-core systems.
+
+  Since all MSP430 and AVR are single-core, we always provide atomic CAS for them without this cfg.
 
   Feel free to submit an issue if your target is not supported yet.
 

diff --git a/src/imp/interrupt/mod.rs b/src/imp/interrupt/mod.rs
@@ -15,6 +15,7 @@
 //
 // AVR, which is single core[^avr1] and LLVM also generates code that disables
 // interrupts [^avr2] in atomic ops by default, is considered the latter.
+// MSP430 as well.
 //
 // [^avr1]: https://github.com/llvm/llvm-project/blob/llvmorg-15.0.0-rc1/llvm/lib/Target/AVR/AVRExpandPseudoInsts.cpp#L1008
 // [^avr2]: https://github.com/llvm/llvm-project/blob/llvmorg-15.0.0-rc1/llvm/test/CodeGen/AVR/atomics/load16.ll#L5
@@ -138,7 +139,6 @@ impl AtomicBool {
         }
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn swap(&self, val: bool, _order: Ordering) -> bool {
         // SAFETY: any data races are prevented by disabling interrupts (see
@@ -147,7 +147,6 @@ impl AtomicBool {
         with(|| unsafe { self.v.get().replace(val as u8) != 0 })
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn compare_exchange(
         &self,
@@ -171,7 +170,6 @@ impl AtomicBool {
         })
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn compare_exchange_weak(
         &self,
@@ -183,7 +181,6 @@ impl AtomicBool {
         self.compare_exchange(current, new, success, failure)
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn fetch_and(&self, val: bool, _order: Ordering) -> bool {
         // SAFETY: any data races are prevented by disabling interrupts (see
@@ -196,7 +193,6 @@ impl AtomicBool {
         })
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn fetch_nand(&self, val: bool, order: Ordering) -> bool {
         if val {
@@ -210,7 +206,6 @@ impl AtomicBool {
         }
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn fetch_or(&self, val: bool, _order: Ordering) -> bool {
         // SAFETY: any data races are prevented by disabling interrupts (see
@@ -223,7 +218,6 @@ impl AtomicBool {
         })
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn fetch_xor(&self, val: bool, _order: Ordering) -> bool {
         // SAFETY: any data races are prevented by disabling interrupts (see
@@ -317,7 +311,6 @@ impl<T> AtomicPtr<T> {
         }
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn swap(&self, ptr: *mut T, _order: Ordering) -> *mut T {
         // SAFETY: any data races are prevented by disabling interrupts (see
@@ -326,7 +319,6 @@ impl<T> AtomicPtr<T> {
         with(|| unsafe { self.p.get().replace(ptr) })
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn compare_exchange(
         &self,
@@ -350,7 +342,6 @@ impl<T> AtomicPtr<T> {
         })
     }
 
-    #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
     #[inline]
     pub(crate) fn compare_exchange_weak(
         &self,
@@ -472,7 +463,6 @@ macro_rules! atomic_int {
     };
     (cas, $atomic_type:ident, $int_type:ident, $align:expr) => {
         impl $atomic_type {
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn swap(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -481,7 +471,6 @@ macro_rules! atomic_int {
                 with(|| unsafe { self.v.get().replace(val) })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn compare_exchange(
                 &self,
@@ -505,7 +494,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn compare_exchange_weak(
                 &self,
@@ -517,7 +505,6 @@ macro_rules! atomic_int {
                 self.compare_exchange(current, new, success, failure)
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_add(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -530,7 +517,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_sub(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -543,7 +529,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_and(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -556,7 +541,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_nand(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -569,7 +553,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_or(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -582,7 +565,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_xor(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -595,7 +577,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_max(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -608,7 +589,6 @@ macro_rules! atomic_int {
                 })
             }
 
-            #[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
             #[inline]
             pub(crate) fn fetch_min(&self, val: $int_type, _order: Ordering) -> $int_type {
                 // SAFETY: any data races are prevented by disabling interrupts (see
@@ -647,34 +627,30 @@ atomic_int!(load_store_atomic, AtomicI16, i16, 2);
 atomic_int!(load_store_atomic, AtomicU16, u16, 2);
 
 #[cfg(not(target_pointer_width = "16"))]
-#[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
 atomic_int!(load_store_atomic, AtomicI32, i32, 4);
 #[cfg(not(target_pointer_width = "16"))]
-#[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
 atomic_int!(load_store_atomic, AtomicU32, u32, 4);
 #[cfg(target_pointer_width = "16")]
-#[cfg(any(test, all(feature = "fallback", portable_atomic_unsafe_assume_single_core)))]
+#[cfg(any(test, feature = "fallback"))]
 atomic_int!(load_store_critical_session, AtomicI32, i32, 4);
 #[cfg(target_pointer_width = "16")]
-#[cfg(any(test, all(feature = "fallback", portable_atomic_unsafe_assume_single_core)))]
+#[cfg(any(test, feature = "fallback"))]
 atomic_int!(load_store_critical_session, AtomicU32, u32, 4);
 
 #[cfg(not(any(target_pointer_width = "16", target_pointer_width = "32")))]
-#[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
 atomic_int!(load_store_atomic, AtomicI64, i64, 8);
 #[cfg(not(any(target_pointer_width = "16", target_pointer_width = "32")))]
-#[cfg(any(test, portable_atomic_unsafe_assume_single_core))]
 atomic_int!(load_store_atomic, AtomicU64, u64, 8);
 #[cfg(any(target_pointer_width = "16", target_pointer_width = "32"))]
-#[cfg(any(test, all(feature = "fallback", portable_atomic_unsafe_assume_single_core)))]
+#[cfg(any(test, feature = "fallback"))]
 atomic_int!(load_store_critical_session, AtomicI64, i64, 8);
 #[cfg(any(target_pointer_width = "16", target_pointer_width = "32"))]
-#[cfg(any(test, all(feature = "fallback", portable_atomic_unsafe_assume_single_core)))]
+#[cfg(any(test, feature = "fallback"))]
 atomic_int!(load_store_critical_session, AtomicU64, u64, 8);
 
-#[cfg(any(test, all(feature = "fallback", portable_atomic_unsafe_assume_single_core)))]
+#[cfg(any(test, feature = "fallback"))]
 atomic_int!(load_store_critical_session, AtomicI128, i128, 16);
-#[cfg(any(test, all(feature = "fallback", portable_atomic_unsafe_assume_single_core)))]
+#[cfg(any(test, feature = "fallback"))]
 atomic_int!(load_store_critical_session, AtomicU128, u128, 16);
 
 #[cfg(test)]

diff --git a/src/imp/mod.rs b/src/imp/mod.rs
@@ -98,10 +98,12 @@ mod fallback;
 // On AVR, we always use critical section based fallback implementation.
 // AVR can be safely assumed to be single-core, so this is sound.
 // https://github.com/llvm/llvm-project/blob/llvmorg-15.0.0-rc1/llvm/lib/Target/AVR/AVRExpandPseudoInsts.cpp#L1008
+// MSP430 as well.
 #[cfg(any(
     all(test, target_os = "none"),
     portable_atomic_unsafe_assume_single_core,
-    target_arch = "avr"
+    target_arch = "avr",
+    target_arch = "msp430",
 ))]
 #[cfg_attr(portable_atomic_no_cfg_target_has_atomic, cfg(any(test, portable_atomic_no_atomic_cas)))]
 #[cfg_attr(
@@ -147,12 +149,6 @@ mod interrupt;
 pub(crate) use self::core_atomic::{
     AtomicBool, AtomicI16, AtomicI8, AtomicIsize, AtomicPtr, AtomicU16, AtomicU8, AtomicUsize,
 };
-// MSP430
-#[cfg(not(portable_atomic_unsafe_assume_single_core))]
-#[cfg(target_arch = "msp430")]
-pub(crate) use self::msp430::{
-    AtomicBool, AtomicI16, AtomicI8, AtomicIsize, AtomicPtr, AtomicU16, AtomicU8, AtomicUsize,
-};
 // RISC-V without A-extension
 #[cfg(not(portable_atomic_unsafe_assume_single_core))]
 #[cfg(any(target_arch = "riscv32", target_arch = "riscv64"))]
@@ -162,15 +158,13 @@ pub(crate) use self::riscv::{
     AtomicBool, AtomicI16, AtomicI8, AtomicIsize, AtomicPtr, AtomicU16, AtomicU8, AtomicUsize,
 };
 // no core Atomic{Isize,Usize,Bool,Ptr}/Atomic{I,U}{8,16} & assume single core => critical section based fallback
-#[cfg(any(portable_atomic_unsafe_assume_single_core, target_arch = "avr"))]
-#[cfg_attr(
-    portable_atomic_no_cfg_target_has_atomic,
-    cfg(any(portable_atomic_no_atomic_cas, target_arch = "avr"))
-)]
-#[cfg_attr(
-    not(portable_atomic_no_cfg_target_has_atomic),
-    cfg(any(not(target_has_atomic = "ptr"), target_arch = "avr"))
-)]
+#[cfg(any(
+    portable_atomic_unsafe_assume_single_core,
+    target_arch = "avr",
+    target_arch = "msp430"
+))]
+#[cfg_attr(portable_atomic_no_cfg_target_has_atomic, cfg(portable_atomic_no_atomic_cas))]
+#[cfg_attr(not(portable_atomic_no_cfg_target_has_atomic), cfg(not(target_has_atomic = "ptr")))]
 pub(crate) use self::interrupt::{
     AtomicBool, AtomicI16, AtomicI8, AtomicIsize, AtomicPtr, AtomicU16, AtomicU8, AtomicUsize,
 };
@@ -210,7 +204,11 @@ pub(crate) use self::core_atomic::{AtomicI32, AtomicU32};
 pub(crate) use self::riscv::{AtomicI32, AtomicU32};
 // no core Atomic{I,U}32 & no CAS & assume single core => critical section based fallback
 #[cfg(any(not(target_pointer_width = "16"), feature = "fallback"))]
-#[cfg(portable_atomic_unsafe_assume_single_core)]
+#[cfg(any(
+    portable_atomic_unsafe_assume_single_core,
+    target_arch = "avr",
+    target_arch = "msp430"
+))]
 #[cfg_attr(portable_atomic_no_cfg_target_has_atomic, cfg(portable_atomic_no_atomic_cas))]
 #[cfg_attr(not(portable_atomic_no_cfg_target_has_atomic), cfg(not(target_has_atomic = "ptr")))]
 pub(crate) use self::interrupt::{AtomicI32, AtomicU32};
@@ -251,7 +249,11 @@ pub(crate) use self::fallback::{AtomicI64, AtomicU64};
     not(any(target_pointer_width = "16", target_pointer_width = "32")),
     feature = "fallback"
 ))]
-#[cfg(portable_atomic_unsafe_assume_single_core)]
+#[cfg(any(
+    portable_atomic_unsafe_assume_single_core,
+    target_arch = "avr",
+    target_arch = "msp430"
+))]
 #[cfg_attr(portable_atomic_no_cfg_target_has_atomic, cfg(portable_atomic_no_atomic_cas))]
 #[cfg_attr(not(portable_atomic_no_cfg_target_has_atomic), cfg(not(target_has_atomic = "ptr")))]
 pub(crate) use self::interrupt::{AtomicI64, AtomicU64};
@@ -316,7 +318,11 @@ pub(crate) use self::s390x::{AtomicI128, AtomicU128};
 pub(crate) use self::fallback::{AtomicI128, AtomicU128};
 // no core Atomic{I,U}128 & no CAS & assume_single_core => critical section based fallback
 #[cfg(feature = "fallback")]
-#[cfg(portable_atomic_unsafe_assume_single_core)]
+#[cfg(any(
+    portable_atomic_unsafe_assume_single_core,
+    target_arch = "avr",
+    target_arch = "msp430"
+))]
 #[cfg_attr(portable_atomic_no_cfg_target_has_atomic, cfg(portable_atomic_no_atomic_cas))]
 #[cfg_attr(not(portable_atomic_no_cfg_target_has_atomic), cfg(not(target_has_atomic = "ptr")))]
 pub(crate) use self::interrupt::{AtomicI128, AtomicU128};