portable-atomic

• Work around core::sync::atomic bug which may cause data races with non-relaxed load/CAS on pre-v6 ARM Linux/Android due to missing memory barrier. (#112)

• Make {8,16,32}-bit non-relaxed load compatible with read-only memory on pre-v6 ARM Linux/Android. Previously only relaxed load was compatible with read-only memory. (#112)

• Make from_ptr const fn on Rust 1.58+. Previously it was const fn only on Rust 1.83+. (c09d0e6)

• Optimize non-relaxed load/store on pre-v6 ARM Linux/Android. (#112)

• Optimize fence on pre-v6 ARM Linux/Android. (#112)

• Optimize 128-bit CAS/swap/min/max on s390x. This also includes removing the code complexity that caused the bug fixed in 1.14.0. (33cd7e5, 6c8f764)

• Documentation improvements.

• Work around core::sync::atomic bug which may cause data races with non-relaxed load/CAS on pre-v6 ARM Linux/Android due to missing memory barrier. (#112)

• Make {8,16,32}-bit non-relaxed load compatible with read-only memory on pre-v6 ARM Linux/Android. Previously only relaxed load was compatible with read-only memory. (#112)

• Make from_ptr const fn on Rust 1.58+. Previously it was const fn only on Rust 1.83+. (c09d0e6)

• Optimize non-relaxed load/store on pre-v6 ARM Linux/Android. (#112)

• Optimize fence on pre-v6 ARM Linux/Android. (#112)

• Optimize 128-bit CAS/swap/min/max on s390x. This also includes removing the code complexity that caused the bug fixed in 1.14.0. (33cd7e5, 6c8f764)

• Documentation improvements.

• Guarantee that is_lock_free to always return the same result. (e74f384)

• Fix bug in 128-bit fetch_{min,max} on pre-z196 s390x. (a6c6c4b)

• Fix bug in atomic float fetch_{add,sub,max,min} on AArch64 when FEAT_LSFE is enabled. (There are currently no known CPUs that support FEAT_LSFE.) (b105f9bc)

• Fix panic in build script when custom target with non-standard name on non-nightly rustc. (02ca737)

• Improve robustness of interrupt disable/restore on RISC-V and Xtensa. (88cb3d5)

• Optimize SeqCst fence on x86. (9127b31)

• Optimize 128-bit CAS when FEAT_LSE is not enabled. (671b9a89)

• Optimize SeqCst atomics on AArch64 Windows MSVC. (671b9a89)

• Optimize 8-bit fetch_and/fetch_or/fetch_xor/fetch_not on AVR when RMW instructions available. (6f2f11f2)

• Optimize 128-bit RMW on s390x. (b8c502a9)

• Optimize run-time CPU feature detection. (cae7032)

• Enable run-time detection by default on AArch64 Apple targets. (6c92015)

• Enable run-time detection by default on AArch64 illumos. (59e3636b)

• aarch64: Support run-time detection of FEAT_LSE2 on Windows. (16bc44f0)

• Make portable_atomic_no_outline_atomics cfg no-op instead of compile error on unsupported architectures. (6b8e3eb)

• Improve compile-time detection of AVR target features. (2b5a9a4f)

• Documentation improvements.

• Guarantee that is_lock_free to always return the same result. (e74f384)

• Fix bug in 128-bit fetch_{min,max} on pre-z196 s390x. (a6c6c4b)

• Fix bug in atomic float fetch_{add,sub,max,min} on AArch64 when FEAT_LSFE is enabled. (There are currently no known CPUs that support FEAT_LSFE.) (b105f9bc)

• Fix panic in build script when custom target with non-standard name on non-nightly rustc. (02ca737)

• Improve robustness of interrupt disable/restore on RISC-V and Xtensa. (88cb3d5)

• Optimize SeqCst fence on x86. (9127b31)

• Optimize 128-bit CAS when FEAT_LSE is not enabled. (671b9a89)

• Optimize SeqCst atomics on AArch64 Windows MSVC. (671b9a89)

• Optimize 8-bit fetch_and/fetch_or/fetch_xor/fetch_not on AVR when RMW instructions available. (6f2f11f2)

• Optimize 128-bit RMW on s390x. (b8c502a9)

• Optimize run-time CPU feature detection. (cae7032)

• Enable run-time detection by default on AArch64 Apple targets. (6c92015)

• Enable run-time detection by default on AArch64 illumos. (59e3636b)

• aarch64: Support run-time detection of FEAT_LSE2 on Windows. (16bc44f0)

• Make portable_atomic_no_outline_atomics cfg no-op instead of compile error on unsupported architectures. (6b8e3eb)

• Improve compile-time detection of AVR target features. (2b5a9a4f)

• Documentation improvements.

• Update to stabilized PowerPC64 inline assembly. (92b02f8a)

• Work around rustc_codegen_gcc bugs on x86_64. (ae4c501)

• Optimize x86_64 128-bit atomics. (a9d61eb, 90a17ca4)

• Improve compile-time detection of RISC-V target features. (535fced)

• Enable release immutability.

• Update to stabilized PowerPC64 inline assembly. (92b02f8a)

• Work around rustc_codegen_gcc bugs on x86_64. (ae4c501)

• Optimize x86_64 128-bit atomics. (a9d61eb, 90a17ca4)

• Improve compile-time detection of RISC-V target features. (535fced)

• Enable release immutability.

• Add unsafe-assume-privileged feature / portable_atomic_unsafe_assume_privileged cfg for safer lock-based fallback on multi-core privileged environments. (b084ee1)

• Support unsafe-assume-single-core/unsafe-assume-privileged on all 32-bit Arm targets. Previously it was only M-profile and pre-v6 targets. (7e07f5e)

• Make AtomicPtr::fetch_* strict-provenance compatible on all environments. Previously it was only strict-provenance compatible on cfg(miri) and otherwise permissive-provenance compatible. (4306943)

• Ensure sequential consistency in lock-based fallback when SeqCst is used. (7e80742)

• Support compile-time detection for x86_64 VMOVDQA. (f7bb1aa)

• Improve compile-time detection of s390x miscellaneous-extensions-3. (11045fe)

• Optimize AVR 8-bit swap when RMW instructions available. (8cedb34)

• Optimize interrupt restore on RISC-V. (9b97a2a)

• Add unsafe-assume-privileged feature / portable_atomic_unsafe_assume_privileged cfg for safer lock-based fallback on multi-core privileged environments. (b084ee1)

• Support unsafe-assume-single-core/unsafe-assume-privileged on all 32-bit Arm targets. Previously it was only M-profile and pre-v6 targets. (7e07f5e)

• Make AtomicPtr::fetch_* strict-provenance compatible on all environments. Previously it was only strict-provenance compatible on cfg(miri) and otherwise permissive-provenance compatible. (4306943)

• Ensure sequential consistency in lock-based fallback when SeqCst is used. (7e80742)

• Support compile-time detection for x86_64 VMOVDQA. (f7bb1aa)

• Improve compile-time detection of s390x miscellaneous-extensions-3. (11045fe)

• Optimize AVR 8-bit swap when RMW instructions available. (8cedb34)

• Optimize interrupt restore on RISC-V. (9b97a2a)

• Fix build error on no-std pre-v6 Arm targets due to the recent upstream change. (83f6f3e)

• Support unsafe-assume-single-core feature / portable_atomic_unsafe_assume_single_core cfg for targets with CAS. (38e9572)

• Improve compile-time detection of s390x target feature. (5ae0ef5)

• Documentation improvements. (c84f720)

• Fix build error on no-std pre-v6 Arm targets due to the recent upstream change. (83f6f3e)

• Support unsafe-assume-single-core feature / portable_atomic_unsafe_assume_single_core cfg for targets with CAS. (38e9572)

• Improve compile-time detection of s390x target feature. (5ae0ef5)

• Documentation improvements. (c84f720)

• Fix build error when building aarch64/arm64ec/powerpc64/s390x targets for Miri or ThreadSanitizer since nightly-2025-05-31.

• aarch64: Optimize atomic floats when FEAT_LSFE is enabled. (#201)

• Improve compile-time detection of RISC-V Zacas extension. (b7634e2)

• Improve run-time detection on linux-musl. (7fdad7f)

• Fix build error when building aarch64/arm64ec/powerpc64/s390x targets for Miri or ThreadSanitizer since nightly-2025-05-31.

• aarch64: Optimize atomic floats when FEAT_LSFE is enabled. (#201)

• Improve compile-time detection of RISC-V Zacas extension. (b7634e2)

• Improve run-time detection on linux-musl. (7fdad7f)

• Optimize interrupt restore on thumbv6m. (dd2004a)

• Work around nightly-2025-02-24 rustc regression causing "cannot use value of type *mut T for inline assembly" error on RISC-V without A extension, MSP430, and pre-v6 no-std Arm targets. (eeb0235)

• Support AtomicF16 and AtomicF128 for unstable f16 and f128 under unstable cfgs. (#200)

• RISC-V Zacas extension support is no longer experimental. (#206)

• Improve support of run-time detection and outline-atomics:

  • riscv: Enable run-time detection of Zacas extension by default on Linux/Android. (#207)
  • aarch64: Support run-time detection of FEAT_LRCPC3/FEAT_LSE128 on FreeBSD. (6a5075d)
  • powerpc64: Support run-time detection of quadword-atomics on AIX (currently disabled by default because detection support for AIX is experimental). (#102)

• Work around nightly-2025-02-24 rustc regression causing "cannot use value of type *mut T for inline assembly" error on RISC-V without A extension, MSP430, and pre-v6 no-std Arm targets. (eeb0235)

• Support AtomicF16 and AtomicF128 for unstable f16 and f128 under unstable cfgs. (#200)

• RISC-V Zacas extension support is no longer experimental. (#206)

• Improve support of run-time detection and outline-atomics:

  • riscv: Enable run-time detection of Zacas extension by default on Linux/Android. (#207)
  • aarch64: Support run-time detection of FEAT_LRCPC3/FEAT_LSE128 on FreeBSD. (6a5075d)
  • powerpc64: Support run-time detection of quadword-atomics on AIX (currently disabled by default because detection support for AIX is experimental). (#102)

• Update to stabilized s390x and Arm64EC inline assembly. (97645c1, e1d1a97)

• Make get_mut const fn on Rust 1.83+. (0dea68c)

• Make from_ptr const fn on Rust 1.83+. (align to the std atomic change in Rust 1.84) (50532d8)

• Various optimizations:

  • RISC-V without A-extension: Optimize 16-bit fetch_not when Zabha enabled. (a487a09)
  • s390x: Optimize 128-bit CAS/RMW. (fba028d, 33ab2c1)
  • PowerPC64: Optimize 128-bit Acquire/AcqRel/SeqCst CAS/RMW and 128-bit CAS with Relaxed failure ordering. (33ab2c1)
  • AVR: Optimize 8-bit load/store. (33ab2c1)

• Improve support of run-time detection and outline-atomics:

  • Enable run-time detection by default on powerpc64 and aarch64 linux-uclibc. (#193)
  • Improve run-time detection of powerpc64 quadword-atomics. (1e3bfda)
  • Improve run-time detection of Zhaoxin CPU. (f283d2a)

• Support RISC-V Zacas extension on pre-1.82 rustc. (#194)

• Improve compile-time detection of RISC-V Zaamo/Zabha extensions. (673137a)

• Respect RUSTC_BOOTSTRAP=-1 recently added in nightly in rustc version detection. (5b2847a)

• Update to stabilized s390x and Arm64EC inline assembly. (97645c1, e1d1a97)

• Make get_mut const fn on Rust 1.83+. (0dea68c)

• Make from_ptr const fn on Rust 1.83+. (align to the std atomic change in Rust 1.84) (50532d8)

• Various optimizations:

  • RISC-V without A-extension: Optimize 16-bit fetch_not when Zabha enabled. (a487a09)
  • s390x: Optimize 128-bit CAS/RMW. (fba028d, 33ab2c1)
  • PowerPC64: Optimize 128-bit Acquire/AcqRel/SeqCst CAS/RMW and 128-bit CAS with Relaxed failure ordering. (33ab2c1)
  • AVR: Optimize 8-bit load/store. (33ab2c1)

• Improve support of run-time detection and outline-atomics:

  • Enable run-time detection by default on powerpc64 and aarch64 linux-uclibc. (#193)
  • Improve run-time detection of powerpc64 quadword-atomics. (1e3bfda)
  • Improve run-time detection of Zhaoxin CPU. (f283d2a)

• Support RISC-V Zacas extension on pre-1.82 rustc. (#194)

• Improve compile-time detection of RISC-V Zaamo/Zabha extensions. (673137a)

• Respect RUSTC_BOOTSTRAP=-1 recently added in nightly in rustc version detection. (5b2847a)

• RISC-V without A-extension: Support RMW when Zaamo extension enabled (even when unsafe-assume-single-core disabled). (#185, 9983a8b) See "operations don't require disabling interrupts" list in interrupt module's readme for the operations provided.

• Support run-time detection of RISC-V Zacas extension (currently disabled by default). (#183)

• Support 128-bit atomics on Arm64EC (currently nightly-only) (#184)

• Improve compile-time detection of powerpc64 quadword-atomics. (3eb8507)

• Improve diagnostics when method that requires CAS is unavailable. (#181)

  Before:

  error[E0599]: no method named `compare_exchange` found for struct `portable_atomic::AtomicUsize` in the current scope
    --> src/race.rs:60:24
     |
  60 |             self.inner.compare_exchange(0, value.get(), Ordering::AcqRel, Ordering::Acquire);
     |                        ^^^^^^^^^^^^^^^^ method not found in `AtomicUsize`
  

  After:

  error[E0277]: `compare_exchange` requires atomic CAS but not available on this target by default
      --> src/race.rs:60:24
       |
  60   |             self.inner.compare_exchange(0, value.get(), Ordering::AcqRel, Ordering::Acquire);
       |                        ^^^^^^^^^^^^^^^^ this associated function is not available on this target by default
       |
       = help: the trait `HasCompareExchange` is not implemented for `&portable_atomic::AtomicUsize`
       = note: consider enabling one of the `unsafe-assume-single-core` or `critical-section` Cargo features
       = note: see <https://docs.rs/portable-atomic/latest/portable_atomic/#optional-features> for more.
  

• Improve compile error messages for some other cases (19716ac, 61dcaaa)

• Various improvements to RISC-V.

  • riscv64: Support 128-bit atomics when Zacas extension enabled. (173) This is currently marked as experimental because LLVM marking the corresponding target feature as experimental.
  • riscv32: Support 64-bit atomics when Zacas extension enabled. (173) This is currently marked as experimental because LLVM marking the corresponding target feature as experimental.
  • Improvements for RISC-V without A-extension:
    • Support zaamo target feature. When building for single-core RISC-V without A-extension, this is equivalent to force-amo feature (8abba4b)
    • Support zabha target feature. (694364a)
    • Strengthen SeqCst store to improve compatibility with code that uses atomic instruction mapping that differs from LLVM and GCC. (5b10b15)

• Improve support of run-time detection and outline-atomics:

  • aarch64: Support run-time detection of FEAT_LRCPC3/FEAT_LSE128 for load/store. (#174)
  • aarch64: Support run-time detection of FEAT_LSE2 on OpenBSD. (4f8c735)
  • aarch64: Support run-time detection of FEAT_LSE/FEAT_LSE2 on illumos (currently disabled by default because illumos AArch64 port is experimental). (#175)
  • powerpc64: Support run-time detection on OpenBSD 7.6+ (currently disabled by default for compatibility with old versions). (09a967b)

• Support AArch64 FEAT_LRCPC3/FEAT_LSE128 with pre-16 LLVM. (#178)

• Improve compile-time detection of AArch64 FEAT_LSE2/FEAT_LRCPC3/FEAT_LSE128. (10d47de)

• Relax minimal version of serde (supported via optional feature) to 1.0.60.

• Support run-time detection for cmpxchg16b on x86_64 on pre-1.69 rustc. (#154)

• Make into_inner const fn on Rust 1.56+. (align to the std atomic change in Rust 1.79) (dee1f89)

• Work around rustc_codegen_gcc bug on x86_64. (d938f77)

• Optimize x86_64 atomics.

  • Optimize 128-bit load/store on Zhaoxin CPU with AVX. (86cee8f)
  • Optimize 128-bit SeqCst store on Intel/AMD/Zhaoxin CPU with AVX. (#156, 0483042)
  • Remove needless test in CAS. (573e025)

• Make rustc version detection robust for custom toolchains. (f8ea85e)

• Respect RUSTC_WRAPPER in rustc version detection.

• Our build script is now less likely to be re-run unnecessarily in versions where the cargo bug fix is available (cargo 1.79+). (52c277b)

• Add cfg_{has,no}_atomic_{8,16,32,64,128,ptr} macros to enable code when the corresponding atomic implementation is available/unavailable.

• Add cfg_{has,no}_atomic_cas macros to enable code when atomic CAS/RMW implementation is available/unavailable.

• Improve support for RISC-V targets without atomic CAS.

• Fix bug in i{8,16} fetch_{or,xor} on RISC-V without A-extension where unsafe-assume-single-core and force-amo are enabled.

• Optimize swap for targets that do not have native atomic CAS instructions.

Note: This release has been yanked due to a bug fixed in 1.5.1.

• Add from_ptr.

• Add force-amo feature (portable_atomic_force_amo cfg) for single-core RISC-V without A-extension. (#124)

• Support run-time detection on AArch64 on pre-1.61 rustc. (#98)

  This also solves a compatibility issue with rustc_codegen_cranelift.

• Support run-time detection of FEAT_LSE2. (#126)

• Support run-time detection of FEAT_LSE on AArch64 NetBSD. (#66)

• Acknowledge ESP-IDF targets' 64-bit atomics are not lock-free. See #122 for more.

• Optimize 128-bit weak CAS on powerpc64.

• Optimize interrupt disable on no-std pre-v6 Arm where unsafe-assume-single-core and disable-fiq are enabled. (771c45d)

• Improve detection of Apple hardware. (5c3a43b)

• Improve compatibility with the future version of Miri.

• Optimize AArch64 128-bit atomic store/swap/fetch_and/fetch_or when the lse128 target feature is enabled at compile-time. (#68)

• Optimize AArch64 128-bit atomic load/store when the rcpc3 target feature is enabled at compile-time. (#68)

• Optimize inline assemblies on Arm, AArch64, and MSP430.

• Optimize AtomicBool on RISC-V/LoongArch64. This is the same as rust-lang/rust#114034, but is available for all rustc versions.

• Improve compatibility with the future version of Miri.

• Allow using embedded-related cfgs as Cargo features. (#94, thanks @Dirbaio)

  Originally, we were providing these as cfgs instead of features, but based on a strong request from the embedded ecosystem, we have agreed to provide them as features as well. See #94 for more.

  cfgs are kept and can be used as aliases for features.

• Acknowledge all x86_64 Apple targets support 128-bit atomics.

  Our code already recognizes this via cfg(target_feature), so this only affects docs and users using pre-1.69 stable rustc.

  See also rust-lang/rust#112150.

• Optimize 128-bit atomics on AArch64/s390x.

• Fix build error on AArch64 ILP32 ABI targets (tier 3).

• Optimize 128-bit atomics on s390x.

• Fix bug in powerpc64/s390x 128-bit atomic RMWs on old nightly.

• Optimize 128-bit atomics on powerpc64/s390x.

• Documentation improvements.

• Add require-cas feature. (#100)

  If your crate supports no-std environment and requires atomic CAS, enabling this feature will allow the portable-atomic to display helpful error messages to users on targets requiring additional action on the user side to provide atomic CAS.

  [dependencies]
  portable-atomic = { version = "1.3", default-features = false, features = ["require-cas"] }
  

  See #100 for more.

• Support portable_atomic_unsafe_assume_single_core cfg on Xtensa targets without atomic CAS. (#86)

• Fix bug in AArch64 128-bit SeqCst load when FEAT_LSE2 is enabled at compile-time. This is the same bug that was fixed in the recently released GCC 13.1. LLVM also has the same bug, which had not yet been fixed when the patch was created; I will open a bug report if necessary after looking into the situation in LLVM. (a29154b)

• Fix compile error on bpf{eb,el}-unknown-none (tier 3) and mipsel-sony-psx (tier 3) when critical-section feature is disabled.

• Various optimizations

  • Optimize x86_64 128-bit outline-atomics. This improves performance by up to 15% in concurrent RMW/store for cases where the cmpxchg16b target feature is not available at compile-time. (40c4cd4)
  • Optimize x86_64 128-bit load that uses cmpxchg16b. (40c4cd4)
  • Optimize AArch64 128-bit load that uses FEAT_LSE. (40c4cd4)
  • Optimize pre-Armv6 Linux/Android 64-bit atomics. (efacc89)
  • Support outline-atomics for powerpc64 128-bit atomics. This is currently disabled by default, and can be enabled by --cfg portable_atomic_outline_atomics. (#90)
  • Optimize AArch64 outline-atomics on linux-musl. On linux-musl, outline-atomics is enabled by default only when dynamic linking is enabled. When static linking is enabled, this can be enabled by --cfg portable_atomic_outline_atomics. See the atomic128 module's readme for more. (8418235, 31d0862)

• Make 64-bit atomics lock-free on Arm Linux/Android targets that do not have 64-bit atomics (e.g., armv5te-unknown-linux-gnueabi, arm-linux-androideabi, etc.) when the kernel version is 3.1 or later. (#82)

• Fix AArch64 128-bit atomics performance regression on Apple hardware. (#89)

• Optimize 128-bit atomics on AArch64, x86_64, powerpc64, and s390x.

• Add Atomic{I,U}*::bit_{set,clear,toggle} and AtomicPtr::bit_{set,clear,toggle}. (#72)

  They correspond to x86's lock bt{s,r,c}, and the implementation calls them on x86/x86_64.

• Add AtomicU*::{fetch_neg,neg} methods. Previously it was only available on AtomicI* and AtomicF*.

• Add as_ptr method to all atomic types. (#79)

• Make AtomicF{32,64}::as_bits const fn on Rust 1.58+. (#79)

• Relax ordering in Serialize impl to reflect the upstream change.

• Optimize x86_64 outline-atomics for 128-bit atomics.

  • Support outline-atomics for cmpxchg16b on Rust 1.69+ (i.e., on Rust 1.69+, x86_64 128-bit atomics is lock-free on all Intel chips and almost all AMD chips, even if cmpxchg16b is not available at compile-time.). Previously it was only nightly. (#80)
  • portable-atomic no longer enables outline-atomics on target where run-time CPU feature detection is not available. (#80)

• Optimize AArch64 outline-atomics for 128-bit atomics.

  • Support more targets and improve performance. (#63, #64, #67, #69, #75, #76, #77) See the atomic128 module's readme for a list of platforms that support outline-atomics. Most of these improvements have already been submitted and accepted in rust-lang/stdarch and will soon be available in std::arch::is_aarch64_feature_detected.
  • portable-atomic no longer enables outline-atomics on target where run-time CPU feature detection is not available.

• Performance improvements. (#70, #81, 6c189ae, 13c92b0, etc.)

• Improve support for old nightly. (#73, 872feb9)

• Documentation improvements.

• Optimize Atomic{I,U}*::{fetch_not,not} methods. (#62)

• Add critical-section feature to use critical-section on targets where atomic CAS is not natively available. (#51, thanks @Dirbaio)

  This is useful to get atomic CAS when --cfg portable_atomic_unsafe_assume_single_core can't be used, such as multi-core targets, unprivileged code running under some RTOS, or environments where disabling interrupts needs extra care due to e.g. real-time requirements.

  See documentation for more.

• Remove outline-atomics feature. This was no-op since 0.3.19.

• Documentation improvements.

The latest version of portable-atomic is 1.x. This release makes portable-atomic 0.3 is built on top of portable-atomic 1.x to make bug fixes and improvements such as support for new targets in 1.x available to the ecosystem that depends on older portable-atomic. portable-atomic 0.3 is still maintained passively, but upgrading to portable-atomic 1.x is recommended. (There are no breaking changes from 0.3, except that a deprecated no-op outline-atomics Cargo feature has been removed.) (#99)

• Add AtomicI*::{fetch_neg,neg} and AtomicF*::fetch_neg methods. (#54)

  AtomicI*::neg are equivalent to the corresponding fetch_* methods, but do not return the previous value. They are intended for optimization on platforms that have atomic instructions for the corresponding operation, such as x86's lock neg.

  Currently, optimizations by these methods (neg) are only guaranteed for x86/x86_64.

• Add Atomic{I,U}*::{fetch_not,not} methods. (#54)

  Atomic{I,U}*::not are equivalent to the corresponding fetch_* methods, but do not return the previous value. They are intended for optimization on platforms that have atomic instructions for the corresponding operation, such as x86's lock not, MSP430's inv.

  Currently, optimizations by these methods (not) are only guaranteed for x86/x86_64 and MSP430.

  (Note: AtomicBool already has fetch_not and not methods.)

• Enable outline-atomics for 128-bit atomics by default. (#57) See #57 for more.

• Improve support for old nightly compilers.

• Fix build error when not using portable_atomic_unsafe_assume_single_core cfg on AVR and MSP430 custom targets. (#50)

  Since 0.3.11, atomic CAS was supported without the cfg on AVR and MSP430 builtin targets, but that change was not applied to custom targets.

• Optimize x86_64 128-bit atomic load/store on AMD CPU with AVX. (#49)

• Improve support for custom targets on old rustc.

• Add Atomic{I,U}*::{add,sub,and,or,xor} and AtomicBool::{and,or,xor} methods. (#47)

  They are equivalent to the corresponding fetch_* methods, but do not return the previous value. They are intended for optimization on platforms that implement atomics using inline assembly, such as the MSP430.

  Currently, optimizations by these methods (add,sub,and,or,xor) are only guaranteed for MSP430; on x86/x86_64, LLVM can optimize in most cases, so cases, where this would improve things, should be rare.

• Various improvements to portable_atomic_unsafe_assume_single_core cfg. (#44, #40)

  • Support disabling FIQs on pre-v6 Arm under portable_atomic_disable_fiq cfg.
  • Support RISC-V supervisor mode under portable_atomic_s_mode cfg.
  • Optimize interrupt restore on AVR and MSP430. (#40)
  • Documentation improvements.

  See #44 for more.

• Implement workaround for std cpuid bug due to LLVM bug (rust-lang/rust#101346, llvm/llvm-project#57550).

  • Our use case is likely not affected, but we implement this just in case.
  • We've confirmed that the uses of inline assembly in this crate are not affected by this LLVM bug.

• Optimize atomic load/store on no-std pre-v6 Arm when portable_atomic_unsafe_assume_single_core cfg is used. (#36)

• Support pre-power8 powerpc64le. powerpc64le's default cpu version is power8, but you can technically compile it for the old cpu using the unsafe -C target-cpu rustc flag.

• Use track_caller when debug assertions are enabled on Rust 1.46+.

• Make powerpc64 128-bit atomics compatible with Miri and ThreadSanitizer on LLVM 15+.

• Document that 128-bit atomics are compatible with Miri and ThreadSanitizer on recent nightly.

• Support atomic CAS on no-std pre-v6 Arm targets (e.g., thumbv4t-none-eabi) under unsafe cfg portable_atomic_unsafe_assume_single_core. (#28)

• Always provide atomic CAS for MSP430 and AVR. (#31)

  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. (On MSP430, the standard library's fences are currently unavailable due to LLVM errors.)

• Update safety requirements for unsafe cfg portable_atomic_unsafe_assume_single_core to mention use of privileged instructions to disable interrupts.

• Atomic operations based on disabling interrupts on single-core systems are now considered lock-free.

  The previous behavior was inconsistent because we consider the pre-v6 Arm Linux's atomic operations provided in a similar way by the Linux kernel to be lock-free.

• Respect -Z allow-features.

• Optimize AArch64 128-bit atomic load when the lse target feature is enabled at compile-time. (#20)

• Fix build error on old Miri.

• Documentation improvements.

• Make AArch64 and s390x 128-bit atomics compatible with Miri and ThreadSanitizer.

• Provide stable equivalent of #![feature(strict_provenance_atomic_ptr)]. (#23)

  • AtomicPtr::fetch_ptr_{add,sub}
  • AtomicPtr::fetch_byte_{add,sub}
  • AtomicPtr::fetch_{or,and,xor}

  These APIs are compatible with strict-provenance on cfg(miri). Otherwise, they are compatible with permissive-provenance. Once #![feature(strict_provenance_atomic_ptr)] is stabilized, these APIs will be strict-provenance compatible in all cases from the version in which it is stabilized.

• Provide stable equivalent of #![feature(atomic_bool_fetch_not)]. (#24)

  • AtomicBool::fetch_not

• Optimize x86_64 128-bit RMWs. (#22)

• Optimize x86_64 outline-atomics.

• Optimize inline assemblies on Arm and AArch64.

• Revert thumbv6m atomic load/store changes made in 0.3.5. This is because rust-lang/rust#99595 has been reverted, so this is no longer needed.

• Fix build failure due to the existence of the specs directory.

• Documentation improvements.

• Optimize inline assemblies on x86_64, RISC-V, and MSP430.

Note: This release has been yanked due to a bug fixed in 0.3.6.

• Provide thumbv6m atomic load/store which is planned to be removed from the standard library in rust-lang/rust#99595. (#18)

• Optimize inline assemblies on AArch64, RISC-V, and powerpc64.

• Optimize x86_64 128-bit atomic store.

• Allow CAS failure ordering stronger than success ordering. (#17)

• Optimize x86_64 128-bit atomic load/store on Intel CPU with AVX. (#16)

• Support native 128-bit atomic operations for powerpc64 (le or pwr8+, currently nightly-only).

• Fix behavior differences between stable and nightly. (#15)

• Optimize AArch64 128-bit atomic load/store when the lse2 target feature is enabled at compile-time. (#11)

• Relax ordering in Debug impl to reflect std changes. (#12)

• Support native 128-bit atomic operations for s390x (currently nightly-only).

• Add AtomicF{32,64}::fetch_abs.

• Add #[must_use] to constructors.

• Use 128-bit atomic operation mappings same as LLVM on AArch64.

• Remove parking_lot optional feature to allow the use of this crate within global allocators.

• Implement AArch64 outline-atomics.

• Remove i128 feature. Atomic{I,U}128 are now always enabled.

• Add outline-atomics feature. Currently, this is the same as the 0.1's i128-dynamic, except that fallback feature is not implicitly enabled.

• Remove i128-dynamic feature in favor of outline-atomics feature.

• Add AtomicF{32,64}::as_bits.

• Support native 128-bit atomic operations for AArch64 at Rust 1.59+. This was previously supported only on nightly. (#6)

• Fix inline assembly for RISC-V without A-extension.

Note: This release has been yanked due to a bug fixed in 0.1.3.

• Add parking_lot feature to use parking_lot in global locks of fallback implementation.

• Fix bug in cmpxchg16b support. (#5)

Note: This release has been yanked due to a bug fixed in 0.1.3.

• Fix doc cfg on Atomic{I,U}128.

Note: This release has been yanked due to a bug fixed in 0.1.3.

Initial release