[lldb] [llvm] [libcxx] [mlir] [clang-tools-extra] [compiler-rt] [lld] [flang] [clang] [libc++][numeric] P0543R3: Saturation arithmetic (PR #77967)

[Mark de Wever via cfe-commits]

================
@@ -0,0 +1,119 @@
+// -*- C++ -*-
+//===----------------------------------------------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef _LIBCPP___NUMERIC_SATURATION_ARITHMETIC_H
+#define _LIBCPP___NUMERIC_SATURATION_ARITHMETIC_H
+
+#include <__concepts/arithmetic.h>
+#include <__config>
+#include <__type_traits/decay.h>
+#include <__utility/cmp.h>
+#include <limits>
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+#if _LIBCPP_STD_VER >= 26
+
+template <typename _Tp>
+concept __libcpp_standard_integer = __libcpp_unsigned_integer<decay_t<_Tp>> || __libcpp_signed_integer<decay_t<_Tp>>;
+
+template <__libcpp_standard_integer _Tp>
+_LIBCPP_HIDE_FROM_ABI constexpr _Tp add_sat(_Tp __x, _Tp __y) noexcept {
+  if (_Tp __sum; !__builtin_add_overflow(__x, __y, &__sum))
----------------
mordante wrote:

I asked both, extend `__builtin_add_overflow` for 128 bit, but typically there is a set like
`__builtin_[su]addl{0,2}_overflow` too, I think it would be good to keep that pattern from Clang's PoV. Having a new builtin also makes it easy for us to query whether `__builtin_add_overflow` has 128-bit support.

https://github.com/llvm/llvm-project/pull/77967