[libcxx-commits] [libcxx] [libc++] Add an ABI setting to harden unique_ptr<T[]>::operator[] (PR #91798)

[Louis Dionne via libcxx-commits]

================
@@ -0,0 +1,55 @@
+// -*- 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___MEMORY_ARRAY_COOKIE_H
+#define _LIBCPP___MEMORY_ARRAY_COOKIE_H
+
+#include <__config>
+
+#include <__configuration/abi.h>
+#include <__type_traits/integral_constant.h>
+#include <__type_traits/is_trivially_destructible.h>
+#include <__type_traits/negation.h>
+#include <cstddef>
+
+#if !defined(_LIBCPP_HAS_NO_PRAGMA_SYSTEM_HEADER)
+#  pragma GCC system_header
+#endif
+
+_LIBCPP_BEGIN_NAMESPACE_STD
+
+// Trait representing whether a type requires an array cookie at the start of its allocation when
+// allocated as `new T[n]` and deallocated as `delete array`.
+//
+// Under the Itanium C++ ABI [1], we know that an array cookie is available unless `T` is trivially
+// destructible and the call to `operator delete[]` is not a sized operator delete. Under ABIs other
+// than the Itanium ABI, we assume there are no array cookies.
+//
+// [1]: https://itanium-cxx-abi.github.io/cxx-abi/abi.html#array-cookies
+#ifdef _LIBCPP_ABI_ITANIUM
+// TODO: Use a builtin instead
+// TODO: We should factor in the choice of the usual deallocation function in this determination.
+template <class _Tp>
+struct __has_array_cookie : _Not<is_trivially_destructible<_Tp> > {};
----------------
ldionne wrote:

@efriedma-quic What do you think about this? It is a bit more conservative than we could be, but I don't really want to start trying to implement the resolution of `operator delete[]` in the library. It looks like a compiler builtin should be fairly easy to provide for both getting the array cookie and telling us whether an array cookie is available.

I'm curious to know whether you think this is a reasonable first implementation.

https://github.com/llvm/llvm-project/pull/91798