Support C++17 over-aligned allocation

Summary:
Add support for C++17 over-aligned allocation:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2016/p0035r4.html

Supporting all 10 operators means we avoid thunking thru libstdc++-v3/libsupc++ and just call jemalloc directly.

It's also worth noting that there is now an aligned *and sized* operator delete:
```
void operator delete(void* ptr, std::size_t size, std::align_val_t al) noexcept;
```

If JeMalloc did not provide this, the default implementation would ignore the size parameter entirely:
https://github.com/gcc-mirror/gcc/blob/master/libstdc%2B%2B-v3/libsupc%2B%2B/del_opsa.cc#L30-L33

(I must also update ax_cxx_compile_stdcxx.m4 to a newer version with C++17 support.)

Test Plan:
Wrote a simple test that allocates and then deletes an over-aligned type:
```
struct alignas(32) Foo {};
Foo *f;

int main()
{
  f = new Foo;
  delete f;
}
```

Before this change, both new and delete go thru PLT, and we end up calling regular old free:
```
(gdb) disassemble
Dump of assembler code for function main():
...
   0x00000000004029b7 <+55>:    call   0x4022d0 <_ZnwmSt11align_val_t@plt>
...
   0x00000000004029d5 <+85>:    call   0x4022e0 <_ZdlPvmSt11align_val_t@plt>
...
(gdb) s
free (ptr=0x7ffff6408020) at /home/engshare/third-party2/jemalloc/master/src/jemalloc.git-trunk/src/jemalloc.c:2842
2842            if (!free_fastpath(ptr, 0, false)) {
```

After this change, we directly call new/delete and ultimately call sdallocx:
```
(gdb) disassemble
Dump of assembler code for function main():
...
   0x0000000000402b77 <+55>:    call   0x496ca0 <operator new(unsigned long, std::align_val_t)>
...
   0x0000000000402b95 <+85>:    call   0x496e60 <operator delete(void*, unsigned long, std::align_val_t)>
...
(gdb) s
116             je_sdallocx_noflags(ptr, size);
```

configure.ac

@@ -290,8 +290,11 @@ if test "x$enable_cxx" = "x1" ; then
   dnl Require at least c++14, which is the first version to support sized
   dnl deallocation.  C++ support is not compiled otherwise.
   m4_include([m4/ax_cxx_compile_stdcxx.m4])
-  AX_CXX_COMPILE_STDCXX([14], [noext], [optional])
+  AX_CXX_COMPILE_STDCXX([17], [noext], [optional])
-  if test "x${HAVE_CXX14}" = "x1" ; then
+  if test "x${HAVE_CXX17}" != "x1"; then
+    AX_CXX_COMPILE_STDCXX([14], [noext], [optional])
+  fi
+  if test "x${HAVE_CXX14}" = "x1" -o "x${HAVE_CXX17}" = "x1"; then
     JE_CXXFLAGS_ADD([-Wall])
     JE_CXXFLAGS_ADD([-Wextra])
     JE_CXXFLAGS_ADD([-g3])

include/jemalloc/internal/jemalloc_internal_decls.h

@@ -5,6 +5,7 @@
 #ifdef _WIN32
 #  include <windows.h>
 #  include "msvc_compat/windows_extra.h"
+#  include "msvc_compat/strings.h"
 #  ifdef _WIN64
 #    if LG_VADDR <= 32
 #      error Generate the headers using x64 vcargs

m4/ax_cxx_compile_stdcxx.m4

@@ -1,5 +1,5 @@
 # ===========================================================================
-#   http://www.gnu.org/software/autoconf-archive/ax_cxx_compile_stdcxx.html
+#  https://www.gnu.org/software/autoconf-archive/ax_cxx_compile_stdcxx.html
 # ===========================================================================
 #
 # SYNOPSIS
@@ -33,21 +33,23 @@
 #   Copyright (c) 2014, 2015 Google Inc.; contributed by Alexey Sokolov <sokolov@google.com>
 #   Copyright (c) 2015 Paul Norman <penorman@mac.com>
 #   Copyright (c) 2015 Moritz Klammler <moritz@klammler.eu>
+#   Copyright (c) 2016, 2018 Krzesimir Nowak <qdlacz@gmail.com>
+#   Copyright (c) 2019 Enji Cooper <yaneurabeya@gmail.com>
 #
 #   Copying and distribution of this file, with or without modification, are
 #   permitted in any medium without royalty provided the copyright notice
 #   and this notice are preserved.  This file is offered as-is, without any
 #   warranty.
 
-#serial 4
+#serial 11
 
 dnl  This macro is based on the code from the AX_CXX_COMPILE_STDCXX_11 macro
 dnl  (serial version number 13).
 
 AC_DEFUN([AX_CXX_COMPILE_STDCXX], [dnl
-  m4_if([$1], [11], [],
+  m4_if([$1], [11], [ax_cxx_compile_alternatives="11 0x"],
-        [$1], [14], [],
+        [$1], [14], [ax_cxx_compile_alternatives="14 1y"],
-        [$1], [17], [m4_fatal([support for C++17 not yet implemented in AX_CXX_COMPILE_STDCXX])],
+        [$1], [17], [ax_cxx_compile_alternatives="17 1z"],
         [m4_fatal([invalid first argument `$1' to AX_CXX_COMPILE_STDCXX])])dnl
   m4_if([$2], [], [],
         [$2], [ext], [],
@@ -59,18 +61,11 @@ AC_DEFUN([AX_CXX_COMPILE_STDCXX], [dnl
         [m4_fatal([invalid third argument `$3' to AX_CXX_COMPILE_STDCXX])])
   AC_LANG_PUSH([C++])dnl
   ac_success=no
-  AC_CACHE_CHECK(whether $CXX supports C++$1 features by default,
-  ax_cv_cxx_compile_cxx$1,
-  [AC_COMPILE_IFELSE([AC_LANG_SOURCE([_AX_CXX_COMPILE_STDCXX_testbody_$1])],
-    [ax_cv_cxx_compile_cxx$1=yes],
-    [ax_cv_cxx_compile_cxx$1=no])])
-  if test x$ax_cv_cxx_compile_cxx$1 = xyes; then
-    ac_success=yes
-  fi
 
   m4_if([$2], [noext], [], [dnl
   if test x$ac_success = xno; then
-    for switch in -std=gnu++$1 -std=gnu++0x; do
+    for alternative in ${ax_cxx_compile_alternatives}; do
+      switch="-std=gnu++${alternative}"
       cachevar=AS_TR_SH([ax_cv_cxx_compile_cxx$1_$switch])
       AC_CACHE_CHECK(whether $CXX supports C++$1 features with $switch,
                      $cachevar,
@@ -96,22 +91,27 @@ AC_DEFUN([AX_CXX_COMPILE_STDCXX], [dnl
     dnl HP's aCC needs +std=c++11 according to:
     dnl http://h21007.www2.hp.com/portal/download/files/unprot/aCxx/PDF_Release_Notes/769149-001.pdf
     dnl Cray's crayCC needs "-h std=c++11"
-    for switch in -std=c++$1 -std=c++0x +std=c++$1 "-h std=c++$1"; do
+    for alternative in ${ax_cxx_compile_alternatives}; do
-      cachevar=AS_TR_SH([ax_cv_cxx_compile_cxx$1_$switch])
+      for switch in -std=c++${alternative} +std=c++${alternative} "-h std=c++${alternative}"; do
-      AC_CACHE_CHECK(whether $CXX supports C++$1 features with $switch,
+        cachevar=AS_TR_SH([ax_cv_cxx_compile_cxx$1_$switch])
-                     $cachevar,
+        AC_CACHE_CHECK(whether $CXX supports C++$1 features with $switch,
-        [ac_save_CXX="$CXX"
+                       $cachevar,
-         CXX="$CXX $switch"
+          [ac_save_CXX="$CXX"
-         AC_COMPILE_IFELSE([AC_LANG_SOURCE([_AX_CXX_COMPILE_STDCXX_testbody_$1])],
+           CXX="$CXX $switch"
-          [eval $cachevar=yes],
+           AC_COMPILE_IFELSE([AC_LANG_SOURCE([_AX_CXX_COMPILE_STDCXX_testbody_$1])],
-          [eval $cachevar=no])
+            [eval $cachevar=yes],
-         CXX="$ac_save_CXX"])
+            [eval $cachevar=no])
-      if eval test x\$$cachevar = xyes; then
+           CXX="$ac_save_CXX"])
-        CXX="$CXX $switch"
+        if eval test x\$$cachevar = xyes; then
-        if test -n "$CXXCPP" ; then
+          CXX="$CXX $switch"
-          CXXCPP="$CXXCPP $switch"
+          if test -n "$CXXCPP" ; then
+            CXXCPP="$CXXCPP $switch"
+          fi
+          ac_success=yes
+          break
         fi
-        ac_success=yes
+      done
+      if test x$ac_success = xyes; then
         break
       fi
     done
@@ -148,6 +148,11 @@ m4_define([_AX_CXX_COMPILE_STDCXX_testbody_14],
   _AX_CXX_COMPILE_STDCXX_testbody_new_in_14
 )
 
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_17],
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_11
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_14
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_17
+)
 
 dnl  Tests for new features in C++11
 
@@ -185,11 +190,13 @@ namespace cxx11
 
     struct Base
     {
+      virtual ~Base() {}
       virtual void f() {}
     };
 
     struct Derived : public Base
     {
+      virtual ~Derived() override {}
       virtual void f() override {}
     };
 
@@ -518,7 +525,7 @@ namespace cxx14
 
   }
 
-  namespace test_digit_seperators
+  namespace test_digit_separators
   {
 
     constexpr auto ten_million = 100'000'000;
@@ -560,3 +567,385 @@ namespace cxx14
 #endif  // __cplusplus >= 201402L
 
 ]])
+
+
+dnl  Tests for new features in C++17
+
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_new_in_17], [[
+
+// If the compiler admits that it is not ready for C++17, why torture it?
+// Hopefully, this will speed up the test.
+
+#ifndef __cplusplus
+
+#error "This is not a C++ compiler"
+
+#elif __cplusplus < 201703L
+
+#error "This is not a C++17 compiler"
+
+#else
+
+#include <initializer_list>
+#include <utility>
+#include <type_traits>
+
+namespace cxx17
+{
+
+  namespace test_constexpr_lambdas
+  {
+
+    constexpr int foo = [](){return 42;}();
+
+  }
+
+  namespace test::nested_namespace::definitions
+  {
+
+  }
+
+  namespace test_fold_expression
+  {
+
+    template<typename... Args>
+    int multiply(Args... args)
+    {
+      return (args * ... * 1);
+    }
+
+    template<typename... Args>
+    bool all(Args... args)
+    {
+      return (args && ...);
+    }
+
+  }
+
+  namespace test_extended_static_assert
+  {
+
+    static_assert (true);
+
+  }
+
+  namespace test_auto_brace_init_list
+  {
+
+    auto foo = {5};
+    auto bar {5};
+
+    static_assert(std::is_same<std::initializer_list<int>, decltype(foo)>::value);
+    static_assert(std::is_same<int, decltype(bar)>::value);
+  }
+
+  namespace test_typename_in_template_template_parameter
+  {
+
+    template<template<typename> typename X> struct D;
+
+  }
+
+  namespace test_fallthrough_nodiscard_maybe_unused_attributes
+  {
+
+    int f1()
+    {
+      return 42;
+    }
+
+    [[nodiscard]] int f2()
+    {
+      [[maybe_unused]] auto unused = f1();
+
+      switch (f1())
+      {
+      case 17:
+        f1();
+        [[fallthrough]];
+      case 42:
+        f1();
+      }
+      return f1();
+    }
+
+  }
+
+  namespace test_extended_aggregate_initialization
+  {
+
+    struct base1
+    {
+      int b1, b2 = 42;
+    };
+
+    struct base2
+    {
+      base2() {
+        b3 = 42;
+      }
+      int b3;
+    };
+
+    struct derived : base1, base2
+    {
+        int d;
+    };
+
+    derived d1 {{1, 2}, {}, 4};  // full initialization
+    derived d2 {{}, {}, 4};      // value-initialized bases
+
+  }
+
+  namespace test_general_range_based_for_loop
+  {
+
+    struct iter
+    {
+      int i;
+
+      int& operator* ()
+      {
+        return i;
+      }
+
+      const int& operator* () const
+      {
+        return i;
+      }
+
+      iter& operator++()
+      {
+        ++i;
+        return *this;
+      }
+    };
+
+    struct sentinel
+    {
+      int i;
+    };
+
+    bool operator== (const iter& i, const sentinel& s)
+    {
+      return i.i == s.i;
+    }
+
+    bool operator!= (const iter& i, const sentinel& s)
+    {
+      return !(i == s);
+    }
+
+    struct range
+    {
+      iter begin() const
+      {
+        return {0};
+      }
+
+      sentinel end() const
+      {
+        return {5};
+      }
+    };
+
+    void f()
+    {
+      range r {};
+
+      for (auto i : r)
+      {
+        [[maybe_unused]] auto v = i;
+      }
+    }
+
+  }
+
+  namespace test_lambda_capture_asterisk_this_by_value
+  {
+
+    struct t
+    {
+      int i;
+      int foo()
+      {
+        return [*this]()
+        {
+          return i;
+        }();
+      }
+    };
+
+  }
+
+  namespace test_enum_class_construction
+  {
+
+    enum class byte : unsigned char
+    {};
+
+    byte foo {42};
+
+  }
+
+  namespace test_constexpr_if
+  {
+
+    template <bool cond>
+    int f ()
+    {
+      if constexpr(cond)
+      {
+        return 13;
+      }
+      else
+      {
+        return 42;
+      }
+    }
+
+  }
+
+  namespace test_selection_statement_with_initializer
+  {
+
+    int f()
+    {
+      return 13;
+    }
+
+    int f2()
+    {
+      if (auto i = f(); i > 0)
+      {
+        return 3;
+      }
+
+      switch (auto i = f(); i + 4)
+      {
+      case 17:
+        return 2;
+
+      default:
+        return 1;
+      }
+    }
+
+  }
+
+  namespace test_template_argument_deduction_for_class_templates
+  {
+
+    template <typename T1, typename T2>
+    struct pair
+    {
+      pair (T1 p1, T2 p2)
+        : m1 {p1},
+          m2 {p2}
+      {}
+
+      T1 m1;
+      T2 m2;
+    };
+
+    void f()
+    {
+      [[maybe_unused]] auto p = pair{13, 42u};
+    }
+
+  }
+
+  namespace test_non_type_auto_template_parameters
+  {
+
+    template <auto n>
+    struct B
+    {};
+
+    B<5> b1;
+    B<'a'> b2;
+
+  }
+
+  namespace test_structured_bindings
+  {
+
+    int arr[2] = { 1, 2 };
+    std::pair<int, int> pr = { 1, 2 };
+
+    auto f1() -> int(&)[2]
+    {
+      return arr;
+    }
+
+    auto f2() -> std::pair<int, int>&
+    {
+      return pr;
+    }
+
+    struct S
+    {
+      int x1 : 2;
+      volatile double y1;
+    };
+
+    S f3()
+    {
+      return {};
+    }
+
+    auto [ x1, y1 ] = f1();
+    auto& [ xr1, yr1 ] = f1();
+    auto [ x2, y2 ] = f2();
+    auto& [ xr2, yr2 ] = f2();
+    const auto [ x3, y3 ] = f3();
+
+  }
+
+  namespace test_exception_spec_type_system
+  {
+
+    struct Good {};
+    struct Bad {};
+
+    void g1() noexcept;
+    void g2();
+
+    template<typename T>
+    Bad
+    f(T*, T*);
+
+    template<typename T1, typename T2>
+    Good
+    f(T1*, T2*);
+
+    static_assert (std::is_same_v<Good, decltype(f(g1, g2))>);
+
+  }
+
+  namespace test_inline_variables
+  {
+
+    template<class T> void f(T)
+    {}
+
+    template<class T> inline T g(T)
+    {
+      return T{};
+    }
+
+    template<> inline void f<>(int)
+    {}
+
+    template<> int g<>(int)
+    {
+      return 5;
+    }
+
+  }
+
+}  // namespace cxx17
+
+#endif  // __cplusplus < 201703L
+
+]])

src/jemalloc_cpp.cpp

@@ -39,6 +39,20 @@ void	operator delete(void *ptr, std::size_t size) noexcept;
 void	operator delete[](void *ptr, std::size_t size) noexcept;
 #endif
 
+#if __cpp_aligned_new >= 201606
+/* C++17's over-aligned operators. */
+void	*operator new(std::size_t size, std::align_val_t);
+void	*operator new(std::size_t size, std::align_val_t, const std::nothrow_t &) noexcept;
+void	*operator new[](std::size_t size, std::align_val_t);
+void	*operator new[](std::size_t size, std::align_val_t, const std::nothrow_t &) noexcept;
+void	operator delete(void* ptr, std::align_val_t) noexcept;
+void	operator delete(void* ptr, std::align_val_t, const std::nothrow_t &) noexcept;
+void	operator delete(void* ptr, std::size_t size, std::align_val_t al) noexcept;
+void	operator delete[](void* ptr, std::align_val_t) noexcept;
+void	operator delete[](void* ptr, std::align_val_t, const std::nothrow_t &) noexcept;
+void	operator delete[](void* ptr, std::size_t size, std::align_val_t al) noexcept;
+#endif
+
 JEMALLOC_NOINLINE
 static void *
 handleOOM(std::size_t size, bool nothrow) {
@@ -76,12 +90,46 @@ JEMALLOC_ALWAYS_INLINE
 void *
 newImpl(std::size_t size) noexcept(IsNoExcept) {
 	void *ptr = je_malloc(size);
-	if (likely(ptr != nullptr))
+	if (likely(ptr != nullptr)) {
 		return ptr;
+	}
 
 	return handleOOM(size, IsNoExcept);
 }
 
+template <bool IsNoExcept>
+JEMALLOC_ALWAYS_INLINE
+void *
+alignedNewImpl(std::size_t size, std::align_val_t alignment) noexcept(IsNoExcept) {
+	void *ptr = je_aligned_alloc(static_cast<std::size_t>(alignment), size);
+	if (likely(ptr != nullptr)) {
+		return ptr;
+	}
+
+	return handleOOM(size, IsNoExcept);
+}
+
+JEMALLOC_ALWAYS_INLINE
+void
+sizedDeleteImpl(void* ptr, std::size_t size) noexcept {
+	if (unlikely(ptr == nullptr)) {
+		return;
+	}
+	je_sdallocx_noflags(ptr, size);
+}
+
+JEMALLOC_ALWAYS_INLINE
+void
+alignedSizedDeleteImpl(void* ptr, std::size_t size, std::align_val_t alignment) noexcept {
+	if (config_debug) {
+		assert(((size_t)alignment & ((size_t)alignment - 1)) == 0);
+	}
+	if (unlikely(ptr == nullptr)) {
+		return;
+	}
+	je_sdallocx(ptr, size, MALLOCX_ALIGN(alignment));
+}
+
 void *
 operator new(std::size_t size) {
 	return newImpl<false>(size);
@@ -102,6 +150,30 @@ operator new[](std::size_t size, const std::nothrow_t &) noexcept {
 	return newImpl<true>(size);
 }
 
+#if __cpp_aligned_new >= 201606
+
+void *
+operator new(std::size_t size, std::align_val_t alignment) {
+	return alignedNewImpl<false>(size, alignment);
+}
+
+void *
+operator new(std::size_t size, std::align_val_t alignment, const std::nothrow_t &) noexcept {
+	return alignedNewImpl<true>(size, alignment);
+}
+
+void *
+operator new[](std::size_t size, std::align_val_t alignment) {
+	return alignedNewImpl<false>(size, alignment);
+}
+
+void *
+operator new[](std::size_t size, std::align_val_t alignment, const std::nothrow_t &) noexcept {
+	return alignedNewImpl<true>(size, alignment);
+}
+
+#endif  // __cpp_aligned_new
+
 void
 operator delete(void *ptr) noexcept {
 	je_free(ptr);
@@ -125,17 +197,46 @@ void operator delete[](void *ptr, const std::nothrow_t &) noexcept {
 
 void
 operator delete(void *ptr, std::size_t size) noexcept {
-	if (unlikely(ptr == nullptr)) {
+	sizedDeleteImpl(ptr, size);
-		return;
-	}
-	je_sdallocx_noflags(ptr, size);
 }
 
-void operator delete[](void *ptr, std::size_t size) noexcept {
+void
-	if (unlikely(ptr == nullptr)) {
+operator delete[](void *ptr, std::size_t size) noexcept {
-		return;
+	sizedDeleteImpl(ptr, size);
-	}
-	je_sdallocx_noflags(ptr, size);
 }
 
 #endif  // __cpp_sized_deallocation
+
+#if __cpp_aligned_new >= 201606
+
+void
+operator delete(void* ptr, std::align_val_t) noexcept {
+	je_free(ptr);
+}
+
+void
+operator delete(void* ptr, std::align_val_t, const std::nothrow_t&) noexcept {
+	je_free(ptr);
+}
+
+void
+operator delete[](void* ptr, std::align_val_t) noexcept {
+	je_free(ptr);
+}
+
+void
+operator delete[](void* ptr, std::align_val_t, const std::nothrow_t&) noexcept {
+	je_free(ptr);
+}
+
+void
+operator delete(void* ptr, std::size_t size, std::align_val_t alignment) noexcept {
+	alignedSizedDeleteImpl(ptr, size, alignment);
+}
+
+void
+operator delete[](void* ptr, std::size_t size, std::align_val_t alignment) noexcept {
+	alignedSizedDeleteImpl(ptr, size, alignment);
+}
+
+#endif  // __cpp_aligned_new