[OpenCL] Document legacy atomics with generic address space

[Anastasia]

Summary

There were multiple requests from the developers to allow this functionality in OpenCL
https://github.com/KhronosGroup/OpenCL-Docs/issues/66

and this can be supported already in some tool flows i.e. using SPIRV-LLVM Translator.
However, there hasn't been enough progress on the core spec side.

Therefore, I suggest documenting this as a clang extension for now.

[svenvh]

Nitpick to make the sentence read more smoothly; otherwise LGTM!

clang/docs/LanguageExtensions.rst

1819

"from earlier than OpenCL 2.0 standard" -> "from the OpenCL 1.x standards"

clang/docs/LanguageExtensions.rst

 .. code-block:: c++
 
   #pragma OPENCL EXTENSION __cl_clang_variadic_functions : enable
   void foo(int a, ...); // compiled - no diagnostic generated
 
   #pragma OPENCL EXTENSION __cl_clang_variadic_functions : disable
   void bar(int a, ...); // error - variadic prototype is not allowed
 
+Legacy 1.x atomics with generic address space
+---------------------------------------------
+
+Clang allows use of atomic functions from the OpenCL 1.x standards
+with the generic address space pointer in C++ for OpenCL mode.
+
+This is a non-portable feature and might not be supported by all
+targets.
+
+**Example of Use**:
+
+.. code-block:: c++
+
+  void foo(__generic volatile unsigned int* a) {
+    atomic_add(a, 1);
+  }
+
 Builtin Functions
 =================
 
 Clang supports a number of builtin library functions with the same syntax as
 GCC, including things like ``__builtin_nan``, ``__builtin_constant_p``,
 ``__builtin_choose_expr``, ``__builtin_types_compatible_p``,
 ``__builtin_assume_aligned``, ``__sync_fetch_and_add``, etc.  In addition to
 the GCC builtins, Clang supports a number of builtins that GCC does not, which