Skip to content

<atomic>: Workaround __iso_volatile_{load,store}64 being non-atomic with clang-cl on 32-bit x86 #5708

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Sep 10, 2025
Merged

<atomic>: Workaround __iso_volatile_{load,store}64 being non-atomic with clang-cl on 32-bit x86 #5708

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6cfdbde
Workaround non-atomic `__iso_volatile_{load,store}64` on x86 clang-cl
cpplearner Sep 3, 2025
7e01026
Point to LLVM-126516 instead
cpplearner Sep 3, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 18 additions & 0 deletions stl/inc/atomic
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1000,7 +1000,12 @@ struct _Atomic_storage<_Ty, 8> { // lock-free using 8-byte intrinsics
void store(const _TVal _Value) noexcept { // store with sequential consistency
const auto _Mem = _STD _Atomic_address_as<long long>(_Storage);
const long long _As_bytes = _STD _Atomic_reinterpret_as<long long>(_Value);
#if defined(__clang__) && defined(_M_IX86) // TRANSITION, LLVM-156228
static_assert(_M_IX86_FP != 0, "8 byte atomic store is not supported on clang-cl with /arch:IA32");
__atomic_store_n(_Mem, _As_bytes, __ATOMIC_SEQ_CST);
#else // ^^^ workaround / no workaround vvv
_ATOMIC_STORE_64_SEQ_CST(_Mem, _As_bytes);
#endif // ^^^ no workaround ^^^
}

void store(const _TVal _Value, const memory_order _Order) noexcept { // store with given memory order
Expand All @@ -1010,9 +1015,17 @@ struct _Atomic_storage<_Ty, 8> { // lock-free using 8-byte intrinsics
_Check_store_memory_order(_Order);

if (_Order == memory_order_relaxed) {
#if defined(__clang__) && defined(_M_IX86) // TRANSITION, LLVM-156228
__atomic_store_n(_Mem, _As_bytes, __ATOMIC_RELAXED);
#else // ^^^ workaround / no workaround vvv
__iso_volatile_store64(_Mem, _As_bytes);
#endif // ^^^ no workaround ^^^
} else if (_Order == memory_order_release) {
#if defined(__clang__) && defined(_M_IX86) // TRANSITION, LLVM-156228
__atomic_store_n(_Mem, _As_bytes, __ATOMIC_RELEASE);
#else // ^^^ workaround / no workaround vvv
__STORE_RELEASE(64, _Mem, _As_bytes);
#endif // ^^^ no workaround ^^^
} else {
store(_Value);
}
Expand All @@ -1025,7 +1038,12 @@ struct _Atomic_storage<_Ty, 8> { // lock-free using 8-byte intrinsics
#if _STD_ATOMIC_USE_ARM64_LDAR_STLR == 1
_ATOMIC_LOAD_ARM64(_As_bytes, 64, _Mem, _Order)
#else // ^^^ _STD_ATOMIC_USE_ARM64_LDAR_STLR == 1 / _STD_ATOMIC_USE_ARM64_LDAR_STLR != 1 vvv
#if defined(__clang__) && defined(_M_IX86) // TRANSITION, LLVM-156228
static_assert(_M_IX86_FP != 0, "8 byte atomic load is not supported on clang-cl with /arch:IA32");
_As_bytes = __atomic_load_n(_Mem, __ATOMIC_RELAXED);
#else // ^^^ workaround / no workaround vvv
_As_bytes = __iso_volatile_load64(_Mem);
#endif // ^^^ no workaround ^^^
_ATOMIC_POST_LOAD_BARRIER_AS_NEEDED(_Order)
#endif // ^^^ _STD_ATOMIC_USE_ARM64_LDAR_STLR != 1 ^^^
return reinterpret_cast<_TVal&>(_As_bytes);
Expand Down