[www-releases] r297634 - Check in the 4.0.0 release

Hans Wennborg via llvm-commits llvm-commits at lists.llvm.org
Mon Mar 13 09:30:21 PDT 2017


Added: www-releases/trunk/4.0.0/docs/_sources/ProgrammersManual.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/ProgrammersManual.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/ProgrammersManual.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/ProgrammersManual.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,4017 @@
+========================
+LLVM Programmer's Manual
+========================
+
+.. contents::
+   :local:
+
+.. warning::
+   This is always a work in progress.
+
+.. _introduction:
+
+Introduction
+============
+
+This document is meant to highlight some of the important classes and interfaces
+available in the LLVM source-base.  This manual is not intended to explain what
+LLVM is, how it works, and what LLVM code looks like.  It assumes that you know
+the basics of LLVM and are interested in writing transformations or otherwise
+analyzing or manipulating the code.
+
+This document should get you oriented so that you can find your way in the
+continuously growing source code that makes up the LLVM infrastructure.  Note
+that this manual is not intended to serve as a replacement for reading the
+source code, so if you think there should be a method in one of these classes to
+do something, but it's not listed, check the source.  Links to the `doxygen
+<http://llvm.org/doxygen/>`__ sources are provided to make this as easy as
+possible.
+
+The first section of this document describes general information that is useful
+to know when working in the LLVM infrastructure, and the second describes the
+Core LLVM classes.  In the future this manual will be extended with information
+describing how to use extension libraries, such as dominator information, CFG
+traversal routines, and useful utilities like the ``InstVisitor`` (`doxygen
+<http://llvm.org/doxygen/InstVisitor_8h-source.html>`__) template.
+
+.. _general:
+
+General Information
+===================
+
+This section contains general information that is useful if you are working in
+the LLVM source-base, but that isn't specific to any particular API.
+
+.. _stl:
+
+The C++ Standard Template Library
+---------------------------------
+
+LLVM makes heavy use of the C++ Standard Template Library (STL), perhaps much
+more than you are used to, or have seen before.  Because of this, you might want
+to do a little background reading in the techniques used and capabilities of the
+library.  There are many good pages that discuss the STL, and several books on
+the subject that you can get, so it will not be discussed in this document.
+
+Here are some useful links:
+
+#. `cppreference.com
+   <http://en.cppreference.com/w/>`_ - an excellent
+   reference for the STL and other parts of the standard C++ library.
+
+#. `C++ In a Nutshell <http://www.tempest-sw.com/cpp/>`_ - This is an O'Reilly
+   book in the making.  It has a decent Standard Library Reference that rivals
+   Dinkumware's, and is unfortunately no longer free since the book has been
+   published.
+
+#. `C++ Frequently Asked Questions <http://www.parashift.com/c++-faq-lite/>`_.
+
+#. `SGI's STL Programmer's Guide <http://www.sgi.com/tech/stl/>`_ - Contains a
+   useful `Introduction to the STL
+   <http://www.sgi.com/tech/stl/stl_introduction.html>`_.
+
+#. `Bjarne Stroustrup's C++ Page
+   <http://www.research.att.com/%7Ebs/C++.html>`_.
+
+#. `Bruce Eckel's Thinking in C++, 2nd ed. Volume 2 Revision 4.0
+   (even better, get the book)
+   <http://www.mindview.net/Books/TICPP/ThinkingInCPP2e.html>`_.
+
+You are also encouraged to take a look at the :doc:`LLVM Coding Standards
+<CodingStandards>` guide which focuses on how to write maintainable code more
+than where to put your curly braces.
+
+.. _resources:
+
+Other useful references
+-----------------------
+
+#. `Using static and shared libraries across platforms
+   <http://www.fortran-2000.com/ArnaudRecipes/sharedlib.html>`_
+
+.. _apis:
+
+Important and useful LLVM APIs
+==============================
+
+Here we highlight some LLVM APIs that are generally useful and good to know
+about when writing transformations.
+
+.. _isa:
+
+The ``isa<>``, ``cast<>`` and ``dyn_cast<>`` templates
+------------------------------------------------------
+
+The LLVM source-base makes extensive use of a custom form of RTTI.  These
+templates have many similarities to the C++ ``dynamic_cast<>`` operator, but
+they don't have some drawbacks (primarily stemming from the fact that
+``dynamic_cast<>`` only works on classes that have a v-table).  Because they are
+used so often, you must know what they do and how they work.  All of these
+templates are defined in the ``llvm/Support/Casting.h`` (`doxygen
+<http://llvm.org/doxygen/Casting_8h-source.html>`__) file (note that you very
+rarely have to include this file directly).
+
+``isa<>``:
+  The ``isa<>`` operator works exactly like the Java "``instanceof``" operator.
+  It returns true or false depending on whether a reference or pointer points to
+  an instance of the specified class.  This can be very useful for constraint
+  checking of various sorts (example below).
+
+``cast<>``:
+  The ``cast<>`` operator is a "checked cast" operation.  It converts a pointer
+  or reference from a base class to a derived class, causing an assertion
+  failure if it is not really an instance of the right type.  This should be
+  used in cases where you have some information that makes you believe that
+  something is of the right type.  An example of the ``isa<>`` and ``cast<>``
+  template is:
+
+  .. code-block:: c++
+
+    static bool isLoopInvariant(const Value *V, const Loop *L) {
+      if (isa<Constant>(V) || isa<Argument>(V) || isa<GlobalValue>(V))
+        return true;
+
+      // Otherwise, it must be an instruction...
+      return !L->contains(cast<Instruction>(V)->getParent());
+    }
+
+  Note that you should **not** use an ``isa<>`` test followed by a ``cast<>``,
+  for that use the ``dyn_cast<>`` operator.
+
+``dyn_cast<>``:
+  The ``dyn_cast<>`` operator is a "checking cast" operation.  It checks to see
+  if the operand is of the specified type, and if so, returns a pointer to it
+  (this operator does not work with references).  If the operand is not of the
+  correct type, a null pointer is returned.  Thus, this works very much like
+  the ``dynamic_cast<>`` operator in C++, and should be used in the same
+  circumstances.  Typically, the ``dyn_cast<>`` operator is used in an ``if``
+  statement or some other flow control statement like this:
+
+  .. code-block:: c++
+
+    if (auto *AI = dyn_cast<AllocationInst>(Val)) {
+      // ...
+    }
+
+  This form of the ``if`` statement effectively combines together a call to
+  ``isa<>`` and a call to ``cast<>`` into one statement, which is very
+  convenient.
+
+  Note that the ``dyn_cast<>`` operator, like C++'s ``dynamic_cast<>`` or Java's
+  ``instanceof`` operator, can be abused.  In particular, you should not use big
+  chained ``if/then/else`` blocks to check for lots of different variants of
+  classes.  If you find yourself wanting to do this, it is much cleaner and more
+  efficient to use the ``InstVisitor`` class to dispatch over the instruction
+  type directly.
+
+``cast_or_null<>``:
+  The ``cast_or_null<>`` operator works just like the ``cast<>`` operator,
+  except that it allows for a null pointer as an argument (which it then
+  propagates).  This can sometimes be useful, allowing you to combine several
+  null checks into one.
+
+``dyn_cast_or_null<>``:
+  The ``dyn_cast_or_null<>`` operator works just like the ``dyn_cast<>``
+  operator, except that it allows for a null pointer as an argument (which it
+  then propagates).  This can sometimes be useful, allowing you to combine
+  several null checks into one.
+
+These five templates can be used with any classes, whether they have a v-table
+or not.  If you want to add support for these templates, see the document
+:doc:`How to set up LLVM-style RTTI for your class hierarchy
+<HowToSetUpLLVMStyleRTTI>`
+
+.. _string_apis:
+
+Passing strings (the ``StringRef`` and ``Twine`` classes)
+---------------------------------------------------------
+
+Although LLVM generally does not do much string manipulation, we do have several
+important APIs which take strings.  Two important examples are the Value class
+-- which has names for instructions, functions, etc. -- and the ``StringMap``
+class which is used extensively in LLVM and Clang.
+
+These are generic classes, and they need to be able to accept strings which may
+have embedded null characters.  Therefore, they cannot simply take a ``const
+char *``, and taking a ``const std::string&`` requires clients to perform a heap
+allocation which is usually unnecessary.  Instead, many LLVM APIs use a
+``StringRef`` or a ``const Twine&`` for passing strings efficiently.
+
+.. _StringRef:
+
+The ``StringRef`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``StringRef`` data type represents a reference to a constant string (a
+character array and a length) and supports the common operations available on
+``std::string``, but does not require heap allocation.
+
+It can be implicitly constructed using a C style null-terminated string, an
+``std::string``, or explicitly with a character pointer and length.  For
+example, the ``StringRef`` find function is declared as:
+
+.. code-block:: c++
+
+  iterator find(StringRef Key);
+
+and clients can call it using any one of:
+
+.. code-block:: c++
+
+  Map.find("foo");                 // Lookup "foo"
+  Map.find(std::string("bar"));    // Lookup "bar"
+  Map.find(StringRef("\0baz", 4)); // Lookup "\0baz"
+
+Similarly, APIs which need to return a string may return a ``StringRef``
+instance, which can be used directly or converted to an ``std::string`` using
+the ``str`` member function.  See ``llvm/ADT/StringRef.h`` (`doxygen
+<http://llvm.org/doxygen/classllvm_1_1StringRef_8h-source.html>`__) for more
+information.
+
+You should rarely use the ``StringRef`` class directly, because it contains
+pointers to external memory it is not generally safe to store an instance of the
+class (unless you know that the external storage will not be freed).
+``StringRef`` is small and pervasive enough in LLVM that it should always be
+passed by value.
+
+The ``Twine`` class
+^^^^^^^^^^^^^^^^^^^
+
+The ``Twine`` (`doxygen <http://llvm.org/doxygen/classllvm_1_1Twine.html>`__)
+class is an efficient way for APIs to accept concatenated strings.  For example,
+a common LLVM paradigm is to name one instruction based on the name of another
+instruction with a suffix, for example:
+
+.. code-block:: c++
+
+    New = CmpInst::Create(..., SO->getName() + ".cmp");
+
+The ``Twine`` class is effectively a lightweight `rope
+<http://en.wikipedia.org/wiki/Rope_(computer_science)>`_ which points to
+temporary (stack allocated) objects.  Twines can be implicitly constructed as
+the result of the plus operator applied to strings (i.e., a C strings, an
+``std::string``, or a ``StringRef``).  The twine delays the actual concatenation
+of strings until it is actually required, at which point it can be efficiently
+rendered directly into a character array.  This avoids unnecessary heap
+allocation involved in constructing the temporary results of string
+concatenation.  See ``llvm/ADT/Twine.h`` (`doxygen
+<http://llvm.org/doxygen/Twine_8h_source.html>`__) and :ref:`here <dss_twine>`
+for more information.
+
+As with a ``StringRef``, ``Twine`` objects point to external memory and should
+almost never be stored or mentioned directly.  They are intended solely for use
+when defining a function which should be able to efficiently accept concatenated
+strings.
+
+.. _formatting_strings:
+
+Formatting strings (the ``formatv`` function)
+---------------------------------------------
+While LLVM doesn't necessarily do a lot of string manipulation and parsing, it
+does do a lot of string formatting.  From diagnostic messages, to llvm tool
+outputs such as ``llvm-readobj`` to printing verbose disassembly listings and
+LLDB runtime logging, the need for string formatting is pervasive.
+
+The ``formatv`` is similar in spirit to ``printf``, but uses a different syntax
+which borrows heavily from Python and C#.  Unlike ``printf`` it deduces the type
+to be formatted at compile time, so it does not need a format specifier such as
+``%d``.  This reduces the mental overhead of trying to construct portable format
+strings, especially for platform-specific types like ``size_t`` or pointer types.
+Unlike both ``printf`` and Python, it additionally fails to compile if LLVM does
+not know how to format the type.  These two properties ensure that the function
+is both safer and simpler to use than traditional formatting methods such as 
+the ``printf`` family of functions.
+
+Simple formatting
+^^^^^^^^^^^^^^^^^
+
+A call to ``formatv`` involves a single **format string** consisting of 0 or more
+**replacement sequences**, followed by a variable length list of **replacement values**.
+A replacement sequence is a string of the form ``{N[[,align]:style]}``.
+
+``N`` refers to the 0-based index of the argument from the list of replacement
+values.  Note that this means it is possible to reference the same parameter
+multiple times, possibly with different style and/or alignment options, in any order.
+
+``align`` is an optional string specifying the width of the field to format
+the value into, and the alignment of the value within the field.  It is specified as
+an optional **alignment style** followed by a positive integral **field width**.  The
+alignment style can be one of the characters ``-`` (left align), ``=`` (center align),
+or ``+`` (right align).  The default is right aligned.  
+
+``style`` is an optional string consisting of a type specific that controls the
+formatting of the value.  For example, to format a floating point value as a percentage,
+you can use the style option ``P``.
+
+Custom formatting
+^^^^^^^^^^^^^^^^^
+
+There are two ways to customize the formatting behavior for a type.
+
+1. Provide a template specialization of ``llvm::format_provider<T>`` for your
+   type ``T`` with the appropriate static format method.
+
+  .. code-block:: c++
+  
+    namespace llvm {
+      template<>
+      struct format_provider<MyFooBar> {
+        static void format(const MyFooBar &V, raw_ostream &Stream, StringRef Style) {
+          // Do whatever is necessary to format `V` into `Stream`
+        }
+      };
+      void foo() {
+        MyFooBar X;
+        std::string S = formatv("{0}", X);
+      }
+    }
+    
+  This is a useful extensibility mechanism for adding support for formatting your own
+  custom types with your own custom Style options.  But it does not help when you want
+  to extend the mechanism for formatting a type that the library already knows how to
+  format.  For that, we need something else.
+    
+2. Provide a **format adapter** inheriting from ``llvm::FormatAdapter<T>``.
+
+  .. code-block:: c++
+  
+    namespace anything {
+      struct format_int_custom : public llvm::FormatAdapter<int> {
+        explicit format_int_custom(int N) : llvm::FormatAdapter<int>(N) {}
+        void format(llvm::raw_ostream &Stream, StringRef Style) override {
+          // Do whatever is necessary to format ``this->Item`` into ``Stream``
+        }
+      };
+    }
+    namespace llvm {
+      void foo() {
+        std::string S = formatv("{0}", anything::format_int_custom(42));
+      }
+    }
+    
+  If the type is detected to be derived from ``FormatAdapter<T>``, ``formatv``
+  will call the
+  ``format`` method on the argument passing in the specified style.  This allows
+  one to provide custom formatting of any type, including one which already has
+  a builtin format provider.
+
+``formatv`` Examples
+^^^^^^^^^^^^^^^^^^^^
+Below is intended to provide an incomplete set of examples demonstrating
+the usage of ``formatv``.  More information can be found by reading the
+doxygen documentation or by looking at the unit test suite.
+
+
+.. code-block:: c++
+  
+  std::string S;
+  // Simple formatting of basic types and implicit string conversion.
+  S = formatv("{0} ({1:P})", 7, 0.35);  // S == "7 (35.00%)"
+  
+  // Out-of-order referencing and multi-referencing
+  outs() << formatv("{0} {2} {1} {0}", 1, "test", 3); // prints "1 3 test 1"
+  
+  // Left, right, and center alignment
+  S = formatv("{0,7}",  'a');  // S == "      a";
+  S = formatv("{0,-7}", 'a');  // S == "a      ";
+  S = formatv("{0,=7}", 'a');  // S == "   a   ";
+  S = formatv("{0,+7}", 'a');  // S == "      a";
+  
+  // Custom styles
+  S = formatv("{0:N} - {0:x} - {1:E}", 12345, 123908342); // S == "12,345 - 0x3039 - 1.24E8"
+  
+  // Adapters
+  S = formatv("{0}", fmt_align(42, AlignStyle::Center, 7));  // S == "  42   "
+  S = formatv("{0}", fmt_repeat("hi", 3)); // S == "hihihi"
+  S = formatv("{0}", fmt_pad("hi", 2, 6)); // S == "  hi      "
+  
+  // Ranges
+  std::vector<int> V = {8, 9, 10};
+  S = formatv("{0}", make_range(V.begin(), V.end())); // S == "8, 9, 10"
+  S = formatv("{0:$[+]}", make_range(V.begin(), V.end())); // S == "8+9+10"
+  S = formatv("{0:$[ + ]@[x]}", make_range(V.begin(), V.end())); // S == "0x8 + 0x9 + 0xA"
+
+.. _error_apis:
+
+Error handling
+--------------
+
+Proper error handling helps us identify bugs in our code, and helps end-users
+understand errors in their tool usage. Errors fall into two broad categories:
+*programmatic* and *recoverable*, with different strategies for handling and
+reporting.
+
+Programmatic Errors
+^^^^^^^^^^^^^^^^^^^
+
+Programmatic errors are violations of program invariants or API contracts, and
+represent bugs within the program itself. Our aim is to document invariants, and
+to abort quickly at the point of failure (providing some basic diagnostic) when
+invariants are broken at runtime.
+
+The fundamental tools for handling programmatic errors are assertions and the
+llvm_unreachable function. Assertions are used to express invariant conditions,
+and should include a message describing the invariant:
+
+.. code-block:: c++
+
+  assert(isPhysReg(R) && "All virt regs should have been allocated already.");
+
+The llvm_unreachable function can be used to document areas of control flow
+that should never be entered if the program invariants hold:
+
+.. code-block:: c++
+
+  enum { Foo, Bar, Baz } X = foo();
+
+  switch (X) {
+    case Foo: /* Handle Foo */; break;
+    case Bar: /* Handle Bar */; break;
+    default:
+      llvm_unreachable("X should be Foo or Bar here");
+  }
+
+Recoverable Errors
+^^^^^^^^^^^^^^^^^^
+
+Recoverable errors represent an error in the program's environment, for example
+a resource failure (a missing file, a dropped network connection, etc.), or
+malformed input. These errors should be detected and communicated to a level of
+the program where they can be handled appropriately. Handling the error may be
+as simple as reporting the issue to the user, or it may involve attempts at
+recovery.
+
+Recoverable errors are modeled using LLVM's ``Error`` scheme. This scheme
+represents errors using function return values, similar to classic C integer
+error codes, or C++'s ``std::error_code``. However, the ``Error`` class is
+actually a lightweight wrapper for user-defined error types, allowing arbitrary
+information to be attached to describe the error. This is similar to the way C++
+exceptions allow throwing of user-defined types.
+
+Success values are created by calling ``Error::success()``, E.g.:
+
+.. code-block:: c++
+
+  Error foo() {
+    // Do something.
+    // Return success.
+    return Error::success();
+  }
+
+Success values are very cheap to construct and return - they have minimal
+impact on program performance.
+
+Failure values are constructed using ``make_error<T>``, where ``T`` is any class
+that inherits from the ErrorInfo utility, E.g.:
+
+.. code-block:: c++
+
+  class BadFileFormat : public ErrorInfo<BadFileFormat> {
+  public:
+    static char ID;
+    std::string Path;
+
+    BadFileFormat(StringRef Path) : Path(Path.str()) {}
+
+    void log(raw_ostream &OS) const override {
+      OS << Path << " is malformed";
+    }
+
+    std::error_code convertToErrorCode() const override {
+      return make_error_code(object_error::parse_failed);
+    }
+  };
+
+  char FileExists::ID; // This should be declared in the C++ file.
+
+  Error printFormattedFile(StringRef Path) {
+    if (<check for valid format>)
+      return make_error<InvalidObjectFile>(Path);
+    // print file contents.
+    return Error::success();
+  }
+
+Error values can be implicitly converted to bool: true for error, false for
+success, enabling the following idiom:
+
+.. code-block:: c++
+
+  Error mayFail();
+
+  Error foo() {
+    if (auto Err = mayFail())
+      return Err;
+    // Success! We can proceed.
+    ...
+
+For functions that can fail but need to return a value the ``Expected<T>``
+utility can be used. Values of this type can be constructed with either a
+``T``, or an ``Error``. Expected<T> values are also implicitly convertible to
+boolean, but with the opposite convention to ``Error``: true for success, false
+for error. If success, the ``T`` value can be accessed via the dereference
+operator. If failure, the ``Error`` value can be extracted using the
+``takeError()`` method. Idiomatic usage looks like:
+
+.. code-block:: c++
+
+  Expected<FormattedFile> openFormattedFile(StringRef Path) {
+    // If badly formatted, return an error.
+    if (auto Err = checkFormat(Path))
+      return std::move(Err);
+    // Otherwise return a FormattedFile instance.
+    return FormattedFile(Path);
+  }
+
+  Error processFormattedFile(StringRef Path) {
+    // Try to open a formatted file
+    if (auto FileOrErr = openFormattedFile(Path)) {
+      // On success, grab a reference to the file and continue.
+      auto &File = *FileOrErr;
+      ...
+    } else
+      // On error, extract the Error value and return it.
+      return FileOrErr.takeError();
+  }
+
+If an ``Expected<T>`` value is in success mode then the ``takeError()`` method
+will return a success value. Using this fact, the above function can be
+rewritten as:
+
+.. code-block:: c++
+
+  Error processFormattedFile(StringRef Path) {
+    // Try to open a formatted file
+    auto FileOrErr = openFormattedFile(Path);
+    if (auto Err = FileOrErr.takeError())
+      // On error, extract the Error value and return it.
+      return Err;
+    // On success, grab a reference to the file and continue.
+    auto &File = *FileOrErr;
+    ...
+  }
+
+This second form is often more readable for functions that involve multiple
+``Expected<T>`` values as it limits the indentation required.
+
+All ``Error`` instances, whether success or failure, must be either checked or
+moved from (via ``std::move`` or a return) before they are destructed.
+Accidentally discarding an unchecked error will cause a program abort at the
+point where the unchecked value's destructor is run, making it easy to identify
+and fix violations of this rule.
+
+Success values are considered checked once they have been tested (by invoking
+the boolean conversion operator):
+
+.. code-block:: c++
+
+  if (auto Err = canFail(...))
+    return Err; // Failure value - move error to caller.
+
+  // Safe to continue: Err was checked.
+
+In contrast, the following code will always cause an abort, even if ``canFail``
+returns a success value:
+
+.. code-block:: c++
+
+    canFail();
+    // Program will always abort here, even if canFail() returns Success, since
+    // the value is not checked.
+
+Failure values are considered checked once a handler for the error type has
+been activated:
+
+.. code-block:: c++
+
+  handleErrors(
+    processFormattedFile(...),
+    [](const BadFileFormat &BFF) {
+      report("Unable to process " + BFF.Path + ": bad format");
+    },
+    [](const FileNotFound &FNF) {
+      report("File not found " + FNF.Path);
+    });
+
+The ``handleErrors`` function takes an error as its first argument, followed by
+a variadic list of "handlers", each of which must be a callable type (a
+function, lambda, or class with a call operator) with one argument. The
+``handleErrors`` function will visit each handler in the sequence and check its
+argument type against the dynamic type of the error, running the first handler
+that matches. This is the same decision process that is used decide which catch
+clause to run for a C++ exception.
+
+Since the list of handlers passed to ``handleErrors`` may not cover every error
+type that can occur, the ``handleErrors`` function also returns an Error value
+that must be checked or propagated. If the error value that is passed to
+``handleErrors`` does not match any of the handlers it will be returned from
+handleErrors. Idiomatic use of ``handleErrors`` thus looks like:
+
+.. code-block:: c++
+
+  if (auto Err =
+        handleErrors(
+          processFormattedFile(...),
+          [](const BadFileFormat &BFF) {
+            report("Unable to process " + BFF.Path + ": bad format");
+          },
+          [](const FileNotFound &FNF) {
+            report("File not found " + FNF.Path);
+          }))
+    return Err;
+
+In cases where you truly know that the handler list is exhaustive the
+``handleAllErrors`` function can be used instead. This is identical to
+``handleErrors`` except that it will terminate the program if an unhandled
+error is passed in, and can therefore return void. The ``handleAllErrors``
+function should generally be avoided: the introduction of a new error type
+elsewhere in the program can easily turn a formerly exhaustive list of errors
+into a non-exhaustive list, risking unexpected program termination. Where
+possible, use handleErrors and propagate unknown errors up the stack instead.
+
+For tool code, where errors can be handled by printing an error message then
+exiting with an error code, the :ref:`ExitOnError <err_exitonerr>` utility
+may be a better choice than handleErrors, as it simplifies control flow when
+calling fallible functions.
+
+StringError
+"""""""""""
+
+Many kinds of errors have no recovery strategy, the only action that can be
+taken is to report them to the user so that the user can attempt to fix the
+environment. In this case representing the error as a string makes perfect
+sense. LLVM provides the ``StringError`` class for this purpose. It takes two
+arguments: A string error message, and an equivalent ``std::error_code`` for
+interoperability:
+
+.. code-block:: c++
+
+  make_error<StringError>("Bad executable",
+                          make_error_code(errc::executable_format_error"));
+
+If you're certain that the error you're building will never need to be converted
+to a ``std::error_code`` you can use the ``inconvertibleErrorCode()`` function:
+
+.. code-block:: c++
+
+  make_error<StringError>("Bad executable", inconvertibleErrorCode());
+
+This should be done only after careful consideration. If any attempt is made to
+convert this error to a ``std::error_code`` it will trigger immediate program
+termination. Unless you are certain that your errors will not need
+interoperability you should look for an existing ``std::error_code`` that you
+can convert to, and even (as painful as it is) consider introducing a new one as
+a stopgap measure.
+
+Interoperability with std::error_code and ErrorOr
+"""""""""""""""""""""""""""""""""""""""""""""""""
+
+Many existing LLVM APIs use ``std::error_code`` and its partner ``ErrorOr<T>``
+(which plays the same role as ``Expected<T>``, but wraps a ``std::error_code``
+rather than an ``Error``). The infectious nature of error types means that an
+attempt to change one of these functions to return ``Error`` or ``Expected<T>``
+instead often results in an avalanche of changes to callers, callers of callers,
+and so on. (The first such attempt, returning an ``Error`` from
+MachOObjectFile's constructor, was abandoned after the diff reached 3000 lines,
+impacted half a dozen libraries, and was still growing).
+
+To solve this problem, the ``Error``/``std::error_code`` interoperability requirement was
+introduced. Two pairs of functions allow any ``Error`` value to be converted to a
+``std::error_code``, any ``Expected<T>`` to be converted to an ``ErrorOr<T>``, and vice
+versa:
+
+.. code-block:: c++
+
+  std::error_code errorToErrorCode(Error Err);
+  Error errorCodeToError(std::error_code EC);
+
+  template <typename T> ErrorOr<T> expectedToErrorOr(Expected<T> TOrErr);
+  template <typename T> Expected<T> errorOrToExpected(ErrorOr<T> TOrEC);
+
+
+Using these APIs it is easy to make surgical patches that update individual
+functions from ``std::error_code`` to ``Error``, and from ``ErrorOr<T>`` to
+``Expected<T>``.
+
+Returning Errors from error handlers
+""""""""""""""""""""""""""""""""""""
+
+Error recovery attempts may themselves fail. For that reason, ``handleErrors``
+actually recognises three different forms of handler signature:
+
+.. code-block:: c++
+
+  // Error must be handled, no new errors produced:
+  void(UserDefinedError &E);
+
+  // Error must be handled, new errors can be produced:
+  Error(UserDefinedError &E);
+
+  // Original error can be inspected, then re-wrapped and returned (or a new
+  // error can be produced):
+  Error(std::unique_ptr<UserDefinedError> E);
+
+Any error returned from a handler will be returned from the ``handleErrors``
+function so that it can be handled itself, or propagated up the stack.
+
+.. _err_exitonerr:
+
+Using ExitOnError to simplify tool code
+"""""""""""""""""""""""""""""""""""""""
+
+Library code should never call ``exit`` for a recoverable error, however in tool
+code (especially command line tools) this can be a reasonable approach. Calling
+``exit`` upon encountering an error dramatically simplifies control flow as the
+error no longer needs to be propagated up the stack. This allows code to be
+written in straight-line style, as long as each fallible call is wrapped in a
+check and call to exit. The ``ExitOnError`` class supports this pattern by
+providing call operators that inspect ``Error`` values, stripping the error away
+in the success case and logging to ``stderr`` then exiting in the failure case.
+
+To use this class, declare a global ``ExitOnError`` variable in your program:
+
+.. code-block:: c++
+
+  ExitOnError ExitOnErr;
+
+Calls to fallible functions can then be wrapped with a call to ``ExitOnErr``,
+turning them into non-failing calls:
+
+.. code-block:: c++
+
+  Error mayFail();
+  Expected<int> mayFail2();
+
+  void foo() {
+    ExitOnErr(mayFail());
+    int X = ExitOnErr(mayFail2());
+  }
+
+On failure, the error's log message will be written to ``stderr``, optionally
+preceded by a string "banner" that can be set by calling the setBanner method. A
+mapping can also be supplied from ``Error`` values to exit codes using the
+``setExitCodeMapper`` method:
+
+.. code-block:: c++
+
+  int main(int argc, char *argv[]) {
+    ExitOnErr.setBanner(std::string(argv[0]) + " error:");
+    ExitOnErr.setExitCodeMapper(
+      [](const Error &Err) {
+        if (Err.isA<BadFileFormat>())
+          return 2;
+        return 1;
+      });
+
+Use ``ExitOnError`` in your tool code where possible as it can greatly improve
+readability.
+
+Fallible constructors
+"""""""""""""""""""""
+
+Some classes require resource acquisition or other complex initialization that
+can fail during construction. Unfortunately constructors can't return errors,
+and having clients test objects after they're constructed to ensure that they're
+valid is error prone as it's all too easy to forget the test. To work around
+this, use the named constructor idiom and return an ``Expected<T>``:
+
+.. code-block:: c++
+
+  class Foo {
+  public:
+
+    static Expected<Foo> Create(Resource R1, Resource R2) {
+      Error Err;
+      Foo F(R1, R2, Err);
+      if (Err)
+        return std::move(Err);
+      return std::move(F);
+    }
+
+  private:
+
+    Foo(Resource R1, Resource R2, Error &Err) {
+      ErrorAsOutParameter EAO(&Err);
+      if (auto Err2 = R1.acquire()) {
+        Err = std::move(Err2);
+        return;
+      }
+      Err = R2.acquire();
+    }
+  };
+
+
+Here, the named constructor passes an ``Error`` by reference into the actual
+constructor, which the constructor can then use to return errors. The
+``ErrorAsOutParameter`` utility sets the ``Error`` value's checked flag on entry
+to the constructor so that the error can be assigned to, then resets it on exit
+to force the client (the named constructor) to check the error.
+
+By using this idiom, clients attempting to construct a Foo receive either a
+well-formed Foo or an Error, never an object in an invalid state.
+
+Propagating and consuming errors based on types
+"""""""""""""""""""""""""""""""""""""""""""""""
+
+In some contexts, certain types of error are known to be benign. For example,
+when walking an archive, some clients may be happy to skip over badly formatted
+object files rather than terminating the walk immediately. Skipping badly
+formatted objects could be achieved using an elaborate handler method, but the
+Error.h header provides two utilities that make this idiom much cleaner: the
+type inspection method, ``isA``, and the ``consumeError`` function:
+
+.. code-block:: c++
+
+  Error walkArchive(Archive A) {
+    for (unsigned I = 0; I != A.numMembers(); ++I) {
+      auto ChildOrErr = A.getMember(I);
+      if (auto Err = ChildOrErr.takeError()) {
+        if (Err.isA<BadFileFormat>())
+          consumeError(std::move(Err))
+        else
+          return Err;
+      }
+      auto &Child = *ChildOrErr;
+      // Use Child
+      ...
+    }
+    return Error::success();
+  }
+
+Concatenating Errors with joinErrors
+""""""""""""""""""""""""""""""""""""
+
+In the archive walking example above ``BadFileFormat`` errors are simply
+consumed and ignored. If the client had wanted report these errors after
+completing the walk over the archive they could use the ``joinErrors`` utility:
+
+.. code-block:: c++
+
+  Error walkArchive(Archive A) {
+    Error DeferredErrs = Error::success();
+    for (unsigned I = 0; I != A.numMembers(); ++I) {
+      auto ChildOrErr = A.getMember(I);
+      if (auto Err = ChildOrErr.takeError())
+        if (Err.isA<BadFileFormat>())
+          DeferredErrs = joinErrors(std::move(DeferredErrs), std::move(Err));
+        else
+          return Err;
+      auto &Child = *ChildOrErr;
+      // Use Child
+      ...
+    }
+    return DeferredErrs;
+  }
+
+The ``joinErrors`` routine builds a special error type called ``ErrorList``,
+which holds a list of user defined errors. The ``handleErrors`` routine
+recognizes this type and will attempt to handle each of the contained erorrs in
+order. If all contained errors can be handled, ``handleErrors`` will return
+``Error::success()``, otherwise ``handleErrors`` will concatenate the remaining
+errors and return the resulting ``ErrorList``.
+
+Building fallible iterators and iterator ranges
+"""""""""""""""""""""""""""""""""""""""""""""""
+
+The archive walking examples above retrieve archive members by index, however
+this requires considerable boiler-plate for iteration and error checking. We can
+clean this up by using ``Error`` with the "fallible iterator" pattern. The usual
+C++ iterator patterns do not allow for failure on increment, but we can
+incorporate support for it by having iterators hold an Error reference through
+which they can report failure. In this pattern, if an increment operation fails
+the failure is recorded via the Error reference and the iterator value is set to
+the end of the range in order to terminate the loop. This ensures that the
+dereference operation is safe anywhere that an ordinary iterator dereference
+would be safe (i.e. when the iterator is not equal to end). Where this pattern
+is followed (as in the ``llvm::object::Archive`` class) the result is much
+cleaner iteration idiom:
+
+.. code-block:: c++
+
+  Error Err;
+  for (auto &Child : Ar->children(Err)) {
+    // Use Child - we only enter the loop when it's valid
+    ...
+  }
+  // Check Err after the loop to ensure it didn't break due to an error.
+  if (Err)
+    return Err;
+
+.. _function_apis:
+
+More information on Error and its related utilities can be found in the
+Error.h header file.
+
+Passing functions and other callable objects
+--------------------------------------------
+
+Sometimes you may want a function to be passed a callback object. In order to
+support lambda expressions and other function objects, you should not use the
+traditional C approach of taking a function pointer and an opaque cookie:
+
+.. code-block:: c++
+
+    void takeCallback(bool (*Callback)(Function *, void *), void *Cookie);
+
+Instead, use one of the following approaches:
+
+Function template
+^^^^^^^^^^^^^^^^^
+
+If you don't mind putting the definition of your function into a header file,
+make it a function template that is templated on the callable type.
+
+.. code-block:: c++
+
+    template<typename Callable>
+    void takeCallback(Callable Callback) {
+      Callback(1, 2, 3);
+    }
+
+The ``function_ref`` class template
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``function_ref``
+(`doxygen <http://llvm.org/docs/doxygen/html/classllvm_1_1function__ref_3_01Ret_07Params_8_8_8_08_4.html>`__) class
+template represents a reference to a callable object, templated over the type
+of the callable. This is a good choice for passing a callback to a function,
+if you don't need to hold onto the callback after the function returns. In this
+way, ``function_ref`` is to ``std::function`` as ``StringRef`` is to
+``std::string``.
+
+``function_ref<Ret(Param1, Param2, ...)>`` can be implicitly constructed from
+any callable object that can be called with arguments of type ``Param1``,
+``Param2``, ..., and returns a value that can be converted to type ``Ret``.
+For example:
+
+.. code-block:: c++
+
+    void visitBasicBlocks(Function *F, function_ref<bool (BasicBlock*)> Callback) {
+      for (BasicBlock &BB : *F)
+        if (Callback(&BB))
+          return;
+    }
+
+can be called using:
+
+.. code-block:: c++
+
+    visitBasicBlocks(F, [&](BasicBlock *BB) {
+      if (process(BB))
+        return isEmpty(BB);
+      return false;
+    });
+
+Note that a ``function_ref`` object contains pointers to external memory, so it
+is not generally safe to store an instance of the class (unless you know that
+the external storage will not be freed). If you need this ability, consider
+using ``std::function``. ``function_ref`` is small enough that it should always
+be passed by value.
+
+.. _DEBUG:
+
+The ``DEBUG()`` macro and ``-debug`` option
+-------------------------------------------
+
+Often when working on your pass you will put a bunch of debugging printouts and
+other code into your pass.  After you get it working, you want to remove it, but
+you may need it again in the future (to work out new bugs that you run across).
+
+Naturally, because of this, you don't want to delete the debug printouts, but
+you don't want them to always be noisy.  A standard compromise is to comment
+them out, allowing you to enable them if you need them in the future.
+
+The ``llvm/Support/Debug.h`` (`doxygen
+<http://llvm.org/doxygen/Debug_8h-source.html>`__) file provides a macro named
+``DEBUG()`` that is a much nicer solution to this problem.  Basically, you can
+put arbitrary code into the argument of the ``DEBUG`` macro, and it is only
+executed if '``opt``' (or any other tool) is run with the '``-debug``' command
+line argument:
+
+.. code-block:: c++
+
+  DEBUG(errs() << "I am here!\n");
+
+Then you can run your pass like this:
+
+.. code-block:: none
+
+  $ opt < a.bc > /dev/null -mypass
+  <no output>
+  $ opt < a.bc > /dev/null -mypass -debug
+  I am here!
+
+Using the ``DEBUG()`` macro instead of a home-brewed solution allows you to not
+have to create "yet another" command line option for the debug output for your
+pass.  Note that ``DEBUG()`` macros are disabled for non-asserts builds, so they
+do not cause a performance impact at all (for the same reason, they should also
+not contain side-effects!).
+
+One additional nice thing about the ``DEBUG()`` macro is that you can enable or
+disable it directly in gdb.  Just use "``set DebugFlag=0``" or "``set
+DebugFlag=1``" from the gdb if the program is running.  If the program hasn't
+been started yet, you can always just run it with ``-debug``.
+
+.. _DEBUG_TYPE:
+
+Fine grained debug info with ``DEBUG_TYPE`` and the ``-debug-only`` option
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Sometimes you may find yourself in a situation where enabling ``-debug`` just
+turns on **too much** information (such as when working on the code generator).
+If you want to enable debug information with more fine-grained control, you
+should define the ``DEBUG_TYPE`` macro and use the ``-debug-only`` option as
+follows:
+
+.. code-block:: c++
+
+  #define DEBUG_TYPE "foo"
+  DEBUG(errs() << "'foo' debug type\n");
+  #undef  DEBUG_TYPE
+  #define DEBUG_TYPE "bar"
+  DEBUG(errs() << "'bar' debug type\n"));
+  #undef  DEBUG_TYPE
+
+Then you can run your pass like this:
+
+.. code-block:: none
+
+  $ opt < a.bc > /dev/null -mypass
+  <no output>
+  $ opt < a.bc > /dev/null -mypass -debug
+  'foo' debug type
+  'bar' debug type
+  $ opt < a.bc > /dev/null -mypass -debug-only=foo
+  'foo' debug type
+  $ opt < a.bc > /dev/null -mypass -debug-only=bar
+  'bar' debug type
+  $ opt < a.bc > /dev/null -mypass -debug-only=foo,bar
+  'foo' debug type
+  'bar' debug type
+
+Of course, in practice, you should only set ``DEBUG_TYPE`` at the top of a file,
+to specify the debug type for the entire module. Be careful that you only do
+this after including Debug.h and not around any #include of headers. Also, you
+should use names more meaningful than "foo" and "bar", because there is no
+system in place to ensure that names do not conflict. If two different modules
+use the same string, they will all be turned on when the name is specified.
+This allows, for example, all debug information for instruction scheduling to be
+enabled with ``-debug-only=InstrSched``, even if the source lives in multiple
+files. The name must not include a comma (,) as that is used to separate the
+arguments of the ``-debug-only`` option.
+
+For performance reasons, -debug-only is not available in optimized build
+(``--enable-optimized``) of LLVM.
+
+The ``DEBUG_WITH_TYPE`` macro is also available for situations where you would
+like to set ``DEBUG_TYPE``, but only for one specific ``DEBUG`` statement.  It
+takes an additional first parameter, which is the type to use.  For example, the
+preceding example could be written as:
+
+.. code-block:: c++
+
+  DEBUG_WITH_TYPE("foo", errs() << "'foo' debug type\n");
+  DEBUG_WITH_TYPE("bar", errs() << "'bar' debug type\n"));
+
+.. _Statistic:
+
+The ``Statistic`` class & ``-stats`` option
+-------------------------------------------
+
+The ``llvm/ADT/Statistic.h`` (`doxygen
+<http://llvm.org/doxygen/Statistic_8h-source.html>`__) file provides a class
+named ``Statistic`` that is used as a unified way to keep track of what the LLVM
+compiler is doing and how effective various optimizations are.  It is useful to
+see what optimizations are contributing to making a particular program run
+faster.
+
+Often you may run your pass on some big program, and you're interested to see
+how many times it makes a certain transformation.  Although you can do this with
+hand inspection, or some ad-hoc method, this is a real pain and not very useful
+for big programs.  Using the ``Statistic`` class makes it very easy to keep
+track of this information, and the calculated information is presented in a
+uniform manner with the rest of the passes being executed.
+
+There are many examples of ``Statistic`` uses, but the basics of using it are as
+follows:
+
+#. Define your statistic like this:
+
+  .. code-block:: c++
+
+    #define DEBUG_TYPE "mypassname"   // This goes before any #includes.
+    STATISTIC(NumXForms, "The # of times I did stuff");
+
+  The ``STATISTIC`` macro defines a static variable, whose name is specified by
+  the first argument.  The pass name is taken from the ``DEBUG_TYPE`` macro, and
+  the description is taken from the second argument.  The variable defined
+  ("NumXForms" in this case) acts like an unsigned integer.
+
+#. Whenever you make a transformation, bump the counter:
+
+  .. code-block:: c++
+
+    ++NumXForms;   // I did stuff!
+
+That's all you have to do.  To get '``opt``' to print out the statistics
+gathered, use the '``-stats``' option:
+
+.. code-block:: none
+
+  $ opt -stats -mypassname < program.bc > /dev/null
+  ... statistics output ...
+
+Note that in order to use the '``-stats``' option, LLVM must be
+compiled with assertions enabled.
+
+When running ``opt`` on a C file from the SPEC benchmark suite, it gives a
+report that looks like this:
+
+.. code-block:: none
+
+   7646 bitcodewriter   - Number of normal instructions
+    725 bitcodewriter   - Number of oversized instructions
+ 129996 bitcodewriter   - Number of bitcode bytes written
+   2817 raise           - Number of insts DCEd or constprop'd
+   3213 raise           - Number of cast-of-self removed
+   5046 raise           - Number of expression trees converted
+     75 raise           - Number of other getelementptr's formed
+    138 raise           - Number of load/store peepholes
+     42 deadtypeelim    - Number of unused typenames removed from symtab
+    392 funcresolve     - Number of varargs functions resolved
+     27 globaldce       - Number of global variables removed
+      2 adce            - Number of basic blocks removed
+    134 cee             - Number of branches revectored
+     49 cee             - Number of setcc instruction eliminated
+    532 gcse            - Number of loads removed
+   2919 gcse            - Number of instructions removed
+     86 indvars         - Number of canonical indvars added
+     87 indvars         - Number of aux indvars removed
+     25 instcombine     - Number of dead inst eliminate
+    434 instcombine     - Number of insts combined
+    248 licm            - Number of load insts hoisted
+   1298 licm            - Number of insts hoisted to a loop pre-header
+      3 licm            - Number of insts hoisted to multiple loop preds (bad, no loop pre-header)
+     75 mem2reg         - Number of alloca's promoted
+   1444 cfgsimplify     - Number of blocks simplified
+
+Obviously, with so many optimizations, having a unified framework for this stuff
+is very nice.  Making your pass fit well into the framework makes it more
+maintainable and useful.
+
+.. _ViewGraph:
+
+Viewing graphs while debugging code
+-----------------------------------
+
+Several of the important data structures in LLVM are graphs: for example CFGs
+made out of LLVM :ref:`BasicBlocks <BasicBlock>`, CFGs made out of LLVM
+:ref:`MachineBasicBlocks <MachineBasicBlock>`, and :ref:`Instruction Selection
+DAGs <SelectionDAG>`.  In many cases, while debugging various parts of the
+compiler, it is nice to instantly visualize these graphs.
+
+LLVM provides several callbacks that are available in a debug build to do
+exactly that.  If you call the ``Function::viewCFG()`` method, for example, the
+current LLVM tool will pop up a window containing the CFG for the function where
+each basic block is a node in the graph, and each node contains the instructions
+in the block.  Similarly, there also exists ``Function::viewCFGOnly()`` (does
+not include the instructions), the ``MachineFunction::viewCFG()`` and
+``MachineFunction::viewCFGOnly()``, and the ``SelectionDAG::viewGraph()``
+methods.  Within GDB, for example, you can usually use something like ``call
+DAG.viewGraph()`` to pop up a window.  Alternatively, you can sprinkle calls to
+these functions in your code in places you want to debug.
+
+Getting this to work requires a small amount of setup.  On Unix systems
+with X11, install the `graphviz <http://www.graphviz.org>`_ toolkit, and make
+sure 'dot' and 'gv' are in your path.  If you are running on Mac OS X, download
+and install the Mac OS X `Graphviz program
+<http://www.pixelglow.com/graphviz/>`_ and add
+``/Applications/Graphviz.app/Contents/MacOS/`` (or wherever you install it) to
+your path. The programs need not be present when configuring, building or
+running LLVM and can simply be installed when needed during an active debug
+session.
+
+``SelectionDAG`` has been extended to make it easier to locate *interesting*
+nodes in large complex graphs.  From gdb, if you ``call DAG.setGraphColor(node,
+"color")``, then the next ``call DAG.viewGraph()`` would highlight the node in
+the specified color (choices of colors can be found at `colors
+<http://www.graphviz.org/doc/info/colors.html>`_.) More complex node attributes
+can be provided with ``call DAG.setGraphAttrs(node, "attributes")`` (choices can
+be found at `Graph attributes <http://www.graphviz.org/doc/info/attrs.html>`_.)
+If you want to restart and clear all the current graph attributes, then you can
+``call DAG.clearGraphAttrs()``.
+
+Note that graph visualization features are compiled out of Release builds to
+reduce file size.  This means that you need a Debug+Asserts or Release+Asserts
+build to use these features.
+
+.. _datastructure:
+
+Picking the Right Data Structure for a Task
+===========================================
+
+LLVM has a plethora of data structures in the ``llvm/ADT/`` directory, and we
+commonly use STL data structures.  This section describes the trade-offs you
+should consider when you pick one.
+
+The first step is a choose your own adventure: do you want a sequential
+container, a set-like container, or a map-like container?  The most important
+thing when choosing a container is the algorithmic properties of how you plan to
+access the container.  Based on that, you should use:
+
+
+* a :ref:`map-like <ds_map>` container if you need efficient look-up of a
+  value based on another value.  Map-like containers also support efficient
+  queries for containment (whether a key is in the map).  Map-like containers
+  generally do not support efficient reverse mapping (values to keys).  If you
+  need that, use two maps.  Some map-like containers also support efficient
+  iteration through the keys in sorted order.  Map-like containers are the most
+  expensive sort, only use them if you need one of these capabilities.
+
+* a :ref:`set-like <ds_set>` container if you need to put a bunch of stuff into
+  a container that automatically eliminates duplicates.  Some set-like
+  containers support efficient iteration through the elements in sorted order.
+  Set-like containers are more expensive than sequential containers.
+
+* a :ref:`sequential <ds_sequential>` container provides the most efficient way
+  to add elements and keeps track of the order they are added to the collection.
+  They permit duplicates and support efficient iteration, but do not support
+  efficient look-up based on a key.
+
+* a :ref:`string <ds_string>` container is a specialized sequential container or
+  reference structure that is used for character or byte arrays.
+
+* a :ref:`bit <ds_bit>` container provides an efficient way to store and
+  perform set operations on sets of numeric id's, while automatically
+  eliminating duplicates.  Bit containers require a maximum of 1 bit for each
+  identifier you want to store.
+
+Once the proper category of container is determined, you can fine tune the
+memory use, constant factors, and cache behaviors of access by intelligently
+picking a member of the category.  Note that constant factors and cache behavior
+can be a big deal.  If you have a vector that usually only contains a few
+elements (but could contain many), for example, it's much better to use
+:ref:`SmallVector <dss_smallvector>` than :ref:`vector <dss_vector>`.  Doing so
+avoids (relatively) expensive malloc/free calls, which dwarf the cost of adding
+the elements to the container.
+
+.. _ds_sequential:
+
+Sequential Containers (std::vector, std::list, etc)
+---------------------------------------------------
+
+There are a variety of sequential containers available for you, based on your
+needs.  Pick the first in this section that will do what you want.
+
+.. _dss_arrayref:
+
+llvm/ADT/ArrayRef.h
+^^^^^^^^^^^^^^^^^^^
+
+The ``llvm::ArrayRef`` class is the preferred class to use in an interface that
+accepts a sequential list of elements in memory and just reads from them.  By
+taking an ``ArrayRef``, the API can be passed a fixed size array, an
+``std::vector``, an ``llvm::SmallVector`` and anything else that is contiguous
+in memory.
+
+.. _dss_fixedarrays:
+
+Fixed Size Arrays
+^^^^^^^^^^^^^^^^^
+
+Fixed size arrays are very simple and very fast.  They are good if you know
+exactly how many elements you have, or you have a (low) upper bound on how many
+you have.
+
+.. _dss_heaparrays:
+
+Heap Allocated Arrays
+^^^^^^^^^^^^^^^^^^^^^
+
+Heap allocated arrays (``new[]`` + ``delete[]``) are also simple.  They are good
+if the number of elements is variable, if you know how many elements you will
+need before the array is allocated, and if the array is usually large (if not,
+consider a :ref:`SmallVector <dss_smallvector>`).  The cost of a heap allocated
+array is the cost of the new/delete (aka malloc/free).  Also note that if you
+are allocating an array of a type with a constructor, the constructor and
+destructors will be run for every element in the array (re-sizable vectors only
+construct those elements actually used).
+
+.. _dss_tinyptrvector:
+
+llvm/ADT/TinyPtrVector.h
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+``TinyPtrVector<Type>`` is a highly specialized collection class that is
+optimized to avoid allocation in the case when a vector has zero or one
+elements.  It has two major restrictions: 1) it can only hold values of pointer
+type, and 2) it cannot hold a null pointer.
+
+Since this container is highly specialized, it is rarely used.
+
+.. _dss_smallvector:
+
+llvm/ADT/SmallVector.h
+^^^^^^^^^^^^^^^^^^^^^^
+
+``SmallVector<Type, N>`` is a simple class that looks and smells just like
+``vector<Type>``: it supports efficient iteration, lays out elements in memory
+order (so you can do pointer arithmetic between elements), supports efficient
+push_back/pop_back operations, supports efficient random access to its elements,
+etc.
+
+The advantage of SmallVector is that it allocates space for some number of
+elements (N) **in the object itself**.  Because of this, if the SmallVector is
+dynamically smaller than N, no malloc is performed.  This can be a big win in
+cases where the malloc/free call is far more expensive than the code that
+fiddles around with the elements.
+
+This is good for vectors that are "usually small" (e.g. the number of
+predecessors/successors of a block is usually less than 8).  On the other hand,
+this makes the size of the SmallVector itself large, so you don't want to
+allocate lots of them (doing so will waste a lot of space).  As such,
+SmallVectors are most useful when on the stack.
+
+SmallVector also provides a nice portable and efficient replacement for
+``alloca``.
+
+.. note::
+
+   Prefer to use ``SmallVectorImpl<T>`` as a parameter type.
+
+   In APIs that don't care about the "small size" (most?), prefer to use
+   the ``SmallVectorImpl<T>`` class, which is basically just the "vector
+   header" (and methods) without the elements allocated after it. Note that
+   ``SmallVector<T, N>`` inherits from ``SmallVectorImpl<T>`` so the
+   conversion is implicit and costs nothing. E.g.
+
+   .. code-block:: c++
+
+      // BAD: Clients cannot pass e.g. SmallVector<Foo, 4>.
+      hardcodedSmallSize(SmallVector<Foo, 2> &Out);
+      // GOOD: Clients can pass any SmallVector<Foo, N>.
+      allowsAnySmallSize(SmallVectorImpl<Foo> &Out);
+
+      void someFunc() {
+        SmallVector<Foo, 8> Vec;
+        hardcodedSmallSize(Vec); // Error.
+        allowsAnySmallSize(Vec); // Works.
+      }
+
+   Even though it has "``Impl``" in the name, this is so widely used that
+   it really isn't "private to the implementation" anymore. A name like
+   ``SmallVectorHeader`` would be more appropriate.
+
+.. _dss_vector:
+
+<vector>
+^^^^^^^^
+
+``std::vector`` is well loved and respected.  It is useful when SmallVector
+isn't: when the size of the vector is often large (thus the small optimization
+will rarely be a benefit) or if you will be allocating many instances of the
+vector itself (which would waste space for elements that aren't in the
+container).  vector is also useful when interfacing with code that expects
+vectors :).
+
+One worthwhile note about std::vector: avoid code like this:
+
+.. code-block:: c++
+
+  for ( ... ) {
+     std::vector<foo> V;
+     // make use of V.
+  }
+
+Instead, write this as:
+
+.. code-block:: c++
+
+  std::vector<foo> V;
+  for ( ... ) {
+     // make use of V.
+     V.clear();
+  }
+
+Doing so will save (at least) one heap allocation and free per iteration of the
+loop.
+
+.. _dss_deque:
+
+<deque>
+^^^^^^^
+
+``std::deque`` is, in some senses, a generalized version of ``std::vector``.
+Like ``std::vector``, it provides constant time random access and other similar
+properties, but it also provides efficient access to the front of the list.  It
+does not guarantee continuity of elements within memory.
+
+In exchange for this extra flexibility, ``std::deque`` has significantly higher
+constant factor costs than ``std::vector``.  If possible, use ``std::vector`` or
+something cheaper.
+
+.. _dss_list:
+
+<list>
+^^^^^^
+
+``std::list`` is an extremely inefficient class that is rarely useful.  It
+performs a heap allocation for every element inserted into it, thus having an
+extremely high constant factor, particularly for small data types.
+``std::list`` also only supports bidirectional iteration, not random access
+iteration.
+
+In exchange for this high cost, std::list supports efficient access to both ends
+of the list (like ``std::deque``, but unlike ``std::vector`` or
+``SmallVector``).  In addition, the iterator invalidation characteristics of
+std::list are stronger than that of a vector class: inserting or removing an
+element into the list does not invalidate iterator or pointers to other elements
+in the list.
+
+.. _dss_ilist:
+
+llvm/ADT/ilist.h
+^^^^^^^^^^^^^^^^
+
+``ilist<T>`` implements an 'intrusive' doubly-linked list.  It is intrusive,
+because it requires the element to store and provide access to the prev/next
+pointers for the list.
+
+``ilist`` has the same drawbacks as ``std::list``, and additionally requires an
+``ilist_traits`` implementation for the element type, but it provides some novel
+characteristics.  In particular, it can efficiently store polymorphic objects,
+the traits class is informed when an element is inserted or removed from the
+list, and ``ilist``\ s are guaranteed to support a constant-time splice
+operation.
+
+These properties are exactly what we want for things like ``Instruction``\ s and
+basic blocks, which is why these are implemented with ``ilist``\ s.
+
+Related classes of interest are explained in the following subsections:
+
+* :ref:`ilist_traits <dss_ilist_traits>`
+
+* :ref:`iplist <dss_iplist>`
+
+* :ref:`llvm/ADT/ilist_node.h <dss_ilist_node>`
+
+* :ref:`Sentinels <dss_ilist_sentinel>`
+
+.. _dss_packedvector:
+
+llvm/ADT/PackedVector.h
+^^^^^^^^^^^^^^^^^^^^^^^
+
+Useful for storing a vector of values using only a few number of bits for each
+value.  Apart from the standard operations of a vector-like container, it can
+also perform an 'or' set operation.
+
+For example:
+
+.. code-block:: c++
+
+  enum State {
+      None = 0x0,
+      FirstCondition = 0x1,
+      SecondCondition = 0x2,
+      Both = 0x3
+  };
+
+  State get() {
+      PackedVector<State, 2> Vec1;
+      Vec1.push_back(FirstCondition);
+
+      PackedVector<State, 2> Vec2;
+      Vec2.push_back(SecondCondition);
+
+      Vec1 |= Vec2;
+      return Vec1[0]; // returns 'Both'.
+  }
+
+.. _dss_ilist_traits:
+
+ilist_traits
+^^^^^^^^^^^^
+
+``ilist_traits<T>`` is ``ilist<T>``'s customization mechanism. ``iplist<T>``
+(and consequently ``ilist<T>``) publicly derive from this traits class.
+
+.. _dss_iplist:
+
+iplist
+^^^^^^
+
+``iplist<T>`` is ``ilist<T>``'s base and as such supports a slightly narrower
+interface.  Notably, inserters from ``T&`` are absent.
+
+``ilist_traits<T>`` is a public base of this class and can be used for a wide
+variety of customizations.
+
+.. _dss_ilist_node:
+
+llvm/ADT/ilist_node.h
+^^^^^^^^^^^^^^^^^^^^^
+
+``ilist_node<T>`` implements the forward and backward links that are expected
+by the ``ilist<T>`` (and analogous containers) in the default manner.
+
+``ilist_node<T>``\ s are meant to be embedded in the node type ``T``, usually
+``T`` publicly derives from ``ilist_node<T>``.
+
+.. _dss_ilist_sentinel:
+
+Sentinels
+^^^^^^^^^
+
+``ilist``\ s have another specialty that must be considered.  To be a good
+citizen in the C++ ecosystem, it needs to support the standard container
+operations, such as ``begin`` and ``end`` iterators, etc.  Also, the
+``operator--`` must work correctly on the ``end`` iterator in the case of
+non-empty ``ilist``\ s.
+
+The only sensible solution to this problem is to allocate a so-called *sentinel*
+along with the intrusive list, which serves as the ``end`` iterator, providing
+the back-link to the last element.  However conforming to the C++ convention it
+is illegal to ``operator++`` beyond the sentinel and it also must not be
+dereferenced.
+
+These constraints allow for some implementation freedom to the ``ilist`` how to
+allocate and store the sentinel.  The corresponding policy is dictated by
+``ilist_traits<T>``.  By default a ``T`` gets heap-allocated whenever the need
+for a sentinel arises.
+
+While the default policy is sufficient in most cases, it may break down when
+``T`` does not provide a default constructor.  Also, in the case of many
+instances of ``ilist``\ s, the memory overhead of the associated sentinels is
+wasted.  To alleviate the situation with numerous and voluminous
+``T``-sentinels, sometimes a trick is employed, leading to *ghostly sentinels*.
+
+Ghostly sentinels are obtained by specially-crafted ``ilist_traits<T>`` which
+superpose the sentinel with the ``ilist`` instance in memory.  Pointer
+arithmetic is used to obtain the sentinel, which is relative to the ``ilist``'s
+``this`` pointer.  The ``ilist`` is augmented by an extra pointer, which serves
+as the back-link of the sentinel.  This is the only field in the ghostly
+sentinel which can be legally accessed.
+
+.. _dss_other:
+
+Other Sequential Container options
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Other STL containers are available, such as ``std::string``.
+
+There are also various STL adapter classes such as ``std::queue``,
+``std::priority_queue``, ``std::stack``, etc.  These provide simplified access
+to an underlying container but don't affect the cost of the container itself.
+
+.. _ds_string:
+
+String-like containers
+----------------------
+
+There are a variety of ways to pass around and use strings in C and C++, and
+LLVM adds a few new options to choose from.  Pick the first option on this list
+that will do what you need, they are ordered according to their relative cost.
+
+Note that it is generally preferred to *not* pass strings around as ``const
+char*``'s.  These have a number of problems, including the fact that they
+cannot represent embedded nul ("\0") characters, and do not have a length
+available efficiently.  The general replacement for '``const char*``' is
+StringRef.
+
+For more information on choosing string containers for APIs, please see
+:ref:`Passing Strings <string_apis>`.
+
+.. _dss_stringref:
+
+llvm/ADT/StringRef.h
+^^^^^^^^^^^^^^^^^^^^
+
+The StringRef class is a simple value class that contains a pointer to a
+character and a length, and is quite related to the :ref:`ArrayRef
+<dss_arrayref>` class (but specialized for arrays of characters).  Because
+StringRef carries a length with it, it safely handles strings with embedded nul
+characters in it, getting the length does not require a strlen call, and it even
+has very convenient APIs for slicing and dicing the character range that it
+represents.
+
+StringRef is ideal for passing simple strings around that are known to be live,
+either because they are C string literals, std::string, a C array, or a
+SmallVector.  Each of these cases has an efficient implicit conversion to
+StringRef, which doesn't result in a dynamic strlen being executed.
+
+StringRef has a few major limitations which make more powerful string containers
+useful:
+
+#. You cannot directly convert a StringRef to a 'const char*' because there is
+   no way to add a trailing nul (unlike the .c_str() method on various stronger
+   classes).
+
+#. StringRef doesn't own or keep alive the underlying string bytes.
+   As such it can easily lead to dangling pointers, and is not suitable for
+   embedding in datastructures in most cases (instead, use an std::string or
+   something like that).
+
+#. For the same reason, StringRef cannot be used as the return value of a
+   method if the method "computes" the result string.  Instead, use std::string.
+
+#. StringRef's do not allow you to mutate the pointed-to string bytes and it
+   doesn't allow you to insert or remove bytes from the range.  For editing
+   operations like this, it interoperates with the :ref:`Twine <dss_twine>`
+   class.
+
+Because of its strengths and limitations, it is very common for a function to
+take a StringRef and for a method on an object to return a StringRef that points
+into some string that it owns.
+
+.. _dss_twine:
+
+llvm/ADT/Twine.h
+^^^^^^^^^^^^^^^^
+
+The Twine class is used as an intermediary datatype for APIs that want to take a
+string that can be constructed inline with a series of concatenations.  Twine
+works by forming recursive instances of the Twine datatype (a simple value
+object) on the stack as temporary objects, linking them together into a tree
+which is then linearized when the Twine is consumed.  Twine is only safe to use
+as the argument to a function, and should always be a const reference, e.g.:
+
+.. code-block:: c++
+
+  void foo(const Twine &T);
+  ...
+  StringRef X = ...
+  unsigned i = ...
+  foo(X + "." + Twine(i));
+
+This example forms a string like "blarg.42" by concatenating the values
+together, and does not form intermediate strings containing "blarg" or "blarg.".
+
+Because Twine is constructed with temporary objects on the stack, and because
+these instances are destroyed at the end of the current statement, it is an
+inherently dangerous API.  For example, this simple variant contains undefined
+behavior and will probably crash:
+
+.. code-block:: c++
+
+  void foo(const Twine &T);
+  ...
+  StringRef X = ...
+  unsigned i = ...
+  const Twine &Tmp = X + "." + Twine(i);
+  foo(Tmp);
+
+... because the temporaries are destroyed before the call.  That said, Twine's
+are much more efficient than intermediate std::string temporaries, and they work
+really well with StringRef.  Just be aware of their limitations.
+
+.. _dss_smallstring:
+
+llvm/ADT/SmallString.h
+^^^^^^^^^^^^^^^^^^^^^^
+
+SmallString is a subclass of :ref:`SmallVector <dss_smallvector>` that adds some
+convenience APIs like += that takes StringRef's.  SmallString avoids allocating
+memory in the case when the preallocated space is enough to hold its data, and
+it calls back to general heap allocation when required.  Since it owns its data,
+it is very safe to use and supports full mutation of the string.
+
+Like SmallVector's, the big downside to SmallString is their sizeof.  While they
+are optimized for small strings, they themselves are not particularly small.
+This means that they work great for temporary scratch buffers on the stack, but
+should not generally be put into the heap: it is very rare to see a SmallString
+as the member of a frequently-allocated heap data structure or returned
+by-value.
+
+.. _dss_stdstring:
+
+std::string
+^^^^^^^^^^^
+
+The standard C++ std::string class is a very general class that (like
+SmallString) owns its underlying data.  sizeof(std::string) is very reasonable
+so it can be embedded into heap data structures and returned by-value.  On the
+other hand, std::string is highly inefficient for inline editing (e.g.
+concatenating a bunch of stuff together) and because it is provided by the
+standard library, its performance characteristics depend a lot of the host
+standard library (e.g. libc++ and MSVC provide a highly optimized string class,
+GCC contains a really slow implementation).
+
+The major disadvantage of std::string is that almost every operation that makes
+them larger can allocate memory, which is slow.  As such, it is better to use
+SmallVector or Twine as a scratch buffer, but then use std::string to persist
+the result.
+
+.. _ds_set:
+
+Set-Like Containers (std::set, SmallSet, SetVector, etc)
+--------------------------------------------------------
+
+Set-like containers are useful when you need to canonicalize multiple values
+into a single representation.  There are several different choices for how to do
+this, providing various trade-offs.
+
+.. _dss_sortedvectorset:
+
+A sorted 'vector'
+^^^^^^^^^^^^^^^^^
+
+If you intend to insert a lot of elements, then do a lot of queries, a great
+approach is to use a vector (or other sequential container) with
+std::sort+std::unique to remove duplicates.  This approach works really well if
+your usage pattern has these two distinct phases (insert then query), and can be
+coupled with a good choice of :ref:`sequential container <ds_sequential>`.
+
+This combination provides the several nice properties: the result data is
+contiguous in memory (good for cache locality), has few allocations, is easy to
+address (iterators in the final vector are just indices or pointers), and can be
+efficiently queried with a standard binary search (e.g.
+``std::lower_bound``; if you want the whole range of elements comparing
+equal, use ``std::equal_range``).
+
+.. _dss_smallset:
+
+llvm/ADT/SmallSet.h
+^^^^^^^^^^^^^^^^^^^
+
+If you have a set-like data structure that is usually small and whose elements
+are reasonably small, a ``SmallSet<Type, N>`` is a good choice.  This set has
+space for N elements in place (thus, if the set is dynamically smaller than N,
+no malloc traffic is required) and accesses them with a simple linear search.
+When the set grows beyond N elements, it allocates a more expensive
+representation that guarantees efficient access (for most types, it falls back
+to :ref:`std::set <dss_set>`, but for pointers it uses something far better,
+:ref:`SmallPtrSet <dss_smallptrset>`.
+
+The magic of this class is that it handles small sets extremely efficiently, but
+gracefully handles extremely large sets without loss of efficiency.  The
+drawback is that the interface is quite small: it supports insertion, queries
+and erasing, but does not support iteration.
+
+.. _dss_smallptrset:
+
+llvm/ADT/SmallPtrSet.h
+^^^^^^^^^^^^^^^^^^^^^^
+
+``SmallPtrSet`` has all the advantages of ``SmallSet`` (and a ``SmallSet`` of
+pointers is transparently implemented with a ``SmallPtrSet``), but also supports
+iterators.  If more than N insertions are performed, a single quadratically
+probed hash table is allocated and grows as needed, providing extremely
+efficient access (constant time insertion/deleting/queries with low constant
+factors) and is very stingy with malloc traffic.
+
+Note that, unlike :ref:`std::set <dss_set>`, the iterators of ``SmallPtrSet``
+are invalidated whenever an insertion occurs.  Also, the values visited by the
+iterators are not visited in sorted order.
+
+.. _dss_stringset:
+
+llvm/ADT/StringSet.h
+^^^^^^^^^^^^^^^^^^^^
+
+``StringSet`` is a thin wrapper around :ref:`StringMap\<char\> <dss_stringmap>`,
+and it allows efficient storage and retrieval of unique strings.
+
+Functionally analogous to ``SmallSet<StringRef>``, ``StringSet`` also supports
+iteration. (The iterator dereferences to a ``StringMapEntry<char>``, so you
+need to call ``i->getKey()`` to access the item of the StringSet.)  On the
+other hand, ``StringSet`` doesn't support range-insertion and
+copy-construction, which :ref:`SmallSet <dss_smallset>` and :ref:`SmallPtrSet
+<dss_smallptrset>` do support.
+
+.. _dss_denseset:
+
+llvm/ADT/DenseSet.h
+^^^^^^^^^^^^^^^^^^^
+
+DenseSet is a simple quadratically probed hash table.  It excels at supporting
+small values: it uses a single allocation to hold all of the pairs that are
+currently inserted in the set.  DenseSet is a great way to unique small values
+that are not simple pointers (use :ref:`SmallPtrSet <dss_smallptrset>` for
+pointers).  Note that DenseSet has the same requirements for the value type that
+:ref:`DenseMap <dss_densemap>` has.
+
+.. _dss_sparseset:
+
+llvm/ADT/SparseSet.h
+^^^^^^^^^^^^^^^^^^^^
+
+SparseSet holds a small number of objects identified by unsigned keys of
+moderate size.  It uses a lot of memory, but provides operations that are almost
+as fast as a vector.  Typical keys are physical registers, virtual registers, or
+numbered basic blocks.
+
+SparseSet is useful for algorithms that need very fast clear/find/insert/erase
+and fast iteration over small sets.  It is not intended for building composite
+data structures.
+
+.. _dss_sparsemultiset:
+
+llvm/ADT/SparseMultiSet.h
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+SparseMultiSet adds multiset behavior to SparseSet, while retaining SparseSet's
+desirable attributes. Like SparseSet, it typically uses a lot of memory, but
+provides operations that are almost as fast as a vector.  Typical keys are
+physical registers, virtual registers, or numbered basic blocks.
+
+SparseMultiSet is useful for algorithms that need very fast
+clear/find/insert/erase of the entire collection, and iteration over sets of
+elements sharing a key. It is often a more efficient choice than using composite
+data structures (e.g. vector-of-vectors, map-of-vectors). It is not intended for
+building composite data structures.
+
+.. _dss_FoldingSet:
+
+llvm/ADT/FoldingSet.h
+^^^^^^^^^^^^^^^^^^^^^
+
+FoldingSet is an aggregate class that is really good at uniquing
+expensive-to-create or polymorphic objects.  It is a combination of a chained
+hash table with intrusive links (uniqued objects are required to inherit from
+FoldingSetNode) that uses :ref:`SmallVector <dss_smallvector>` as part of its ID
+process.
+
+Consider a case where you want to implement a "getOrCreateFoo" method for a
+complex object (for example, a node in the code generator).  The client has a
+description of **what** it wants to generate (it knows the opcode and all the
+operands), but we don't want to 'new' a node, then try inserting it into a set
+only to find out it already exists, at which point we would have to delete it
+and return the node that already exists.
+
+To support this style of client, FoldingSet perform a query with a
+FoldingSetNodeID (which wraps SmallVector) that can be used to describe the
+element that we want to query for.  The query either returns the element
+matching the ID or it returns an opaque ID that indicates where insertion should
+take place.  Construction of the ID usually does not require heap traffic.
+
+Because FoldingSet uses intrusive links, it can support polymorphic objects in
+the set (for example, you can have SDNode instances mixed with LoadSDNodes).
+Because the elements are individually allocated, pointers to the elements are
+stable: inserting or removing elements does not invalidate any pointers to other
+elements.
+
+.. _dss_set:
+
+<set>
+^^^^^
+
+``std::set`` is a reasonable all-around set class, which is decent at many
+things but great at nothing.  std::set allocates memory for each element
+inserted (thus it is very malloc intensive) and typically stores three pointers
+per element in the set (thus adding a large amount of per-element space
+overhead).  It offers guaranteed log(n) performance, which is not particularly
+fast from a complexity standpoint (particularly if the elements of the set are
+expensive to compare, like strings), and has extremely high constant factors for
+lookup, insertion and removal.
+
+The advantages of std::set are that its iterators are stable (deleting or
+inserting an element from the set does not affect iterators or pointers to other
+elements) and that iteration over the set is guaranteed to be in sorted order.
+If the elements in the set are large, then the relative overhead of the pointers
+and malloc traffic is not a big deal, but if the elements of the set are small,
+std::set is almost never a good choice.
+
+.. _dss_setvector:
+
+llvm/ADT/SetVector.h
+^^^^^^^^^^^^^^^^^^^^
+
+LLVM's ``SetVector<Type>`` is an adapter class that combines your choice of a
+set-like container along with a :ref:`Sequential Container <ds_sequential>` The
+important property that this provides is efficient insertion with uniquing
+(duplicate elements are ignored) with iteration support.  It implements this by
+inserting elements into both a set-like container and the sequential container,
+using the set-like container for uniquing and the sequential container for
+iteration.
+
+The difference between SetVector and other sets is that the order of iteration
+is guaranteed to match the order of insertion into the SetVector.  This property
+is really important for things like sets of pointers.  Because pointer values
+are non-deterministic (e.g. vary across runs of the program on different
+machines), iterating over the pointers in the set will not be in a well-defined
+order.
+
+The drawback of SetVector is that it requires twice as much space as a normal
+set and has the sum of constant factors from the set-like container and the
+sequential container that it uses.  Use it **only** if you need to iterate over
+the elements in a deterministic order.  SetVector is also expensive to delete
+elements out of (linear time), unless you use its "pop_back" method, which is
+faster.
+
+``SetVector`` is an adapter class that defaults to using ``std::vector`` and a
+size 16 ``SmallSet`` for the underlying containers, so it is quite expensive.
+However, ``"llvm/ADT/SetVector.h"`` also provides a ``SmallSetVector`` class,
+which defaults to using a ``SmallVector`` and ``SmallSet`` of a specified size.
+If you use this, and if your sets are dynamically smaller than ``N``, you will
+save a lot of heap traffic.
+
+.. _dss_uniquevector:
+
+llvm/ADT/UniqueVector.h
+^^^^^^^^^^^^^^^^^^^^^^^
+
+UniqueVector is similar to :ref:`SetVector <dss_setvector>` but it retains a
+unique ID for each element inserted into the set.  It internally contains a map
+and a vector, and it assigns a unique ID for each value inserted into the set.
+
+UniqueVector is very expensive: its cost is the sum of the cost of maintaining
+both the map and vector, it has high complexity, high constant factors, and
+produces a lot of malloc traffic.  It should be avoided.
+
+.. _dss_immutableset:
+
+llvm/ADT/ImmutableSet.h
+^^^^^^^^^^^^^^^^^^^^^^^
+
+ImmutableSet is an immutable (functional) set implementation based on an AVL
+tree.  Adding or removing elements is done through a Factory object and results
+in the creation of a new ImmutableSet object.  If an ImmutableSet already exists
+with the given contents, then the existing one is returned; equality is compared
+with a FoldingSetNodeID.  The time and space complexity of add or remove
+operations is logarithmic in the size of the original set.
+
+There is no method for returning an element of the set, you can only check for
+membership.
+
+.. _dss_otherset:
+
+Other Set-Like Container Options
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The STL provides several other options, such as std::multiset and the various
+"hash_set" like containers (whether from C++ TR1 or from the SGI library).  We
+never use hash_set and unordered_set because they are generally very expensive
+(each insertion requires a malloc) and very non-portable.
+
+std::multiset is useful if you're not interested in elimination of duplicates,
+but has all the drawbacks of :ref:`std::set <dss_set>`.  A sorted vector
+(where you don't delete duplicate entries) or some other approach is almost
+always better.
+
+.. _ds_map:
+
+Map-Like Containers (std::map, DenseMap, etc)
+---------------------------------------------
+
+Map-like containers are useful when you want to associate data to a key.  As
+usual, there are a lot of different ways to do this. :)
+
+.. _dss_sortedvectormap:
+
+A sorted 'vector'
+^^^^^^^^^^^^^^^^^
+
+If your usage pattern follows a strict insert-then-query approach, you can
+trivially use the same approach as :ref:`sorted vectors for set-like containers
+<dss_sortedvectorset>`.  The only difference is that your query function (which
+uses std::lower_bound to get efficient log(n) lookup) should only compare the
+key, not both the key and value.  This yields the same advantages as sorted
+vectors for sets.
+
+.. _dss_stringmap:
+
+llvm/ADT/StringMap.h
+^^^^^^^^^^^^^^^^^^^^
+
+Strings are commonly used as keys in maps, and they are difficult to support
+efficiently: they are variable length, inefficient to hash and compare when
+long, expensive to copy, etc.  StringMap is a specialized container designed to
+cope with these issues.  It supports mapping an arbitrary range of bytes to an
+arbitrary other object.
+
+The StringMap implementation uses a quadratically-probed hash table, where the
+buckets store a pointer to the heap allocated entries (and some other stuff).
+The entries in the map must be heap allocated because the strings are variable
+length.  The string data (key) and the element object (value) are stored in the
+same allocation with the string data immediately after the element object.
+This container guarantees the "``(char*)(&Value+1)``" points to the key string
+for a value.
+
+The StringMap is very fast for several reasons: quadratic probing is very cache
+efficient for lookups, the hash value of strings in buckets is not recomputed
+when looking up an element, StringMap rarely has to touch the memory for
+unrelated objects when looking up a value (even when hash collisions happen),
+hash table growth does not recompute the hash values for strings already in the
+table, and each pair in the map is store in a single allocation (the string data
+is stored in the same allocation as the Value of a pair).
+
+StringMap also provides query methods that take byte ranges, so it only ever
+copies a string if a value is inserted into the table.
+
+StringMap iteratation order, however, is not guaranteed to be deterministic, so
+any uses which require that should instead use a std::map.
+
+.. _dss_indexmap:
+
+llvm/ADT/IndexedMap.h
+^^^^^^^^^^^^^^^^^^^^^
+
+IndexedMap is a specialized container for mapping small dense integers (or
+values that can be mapped to small dense integers) to some other type.  It is
+internally implemented as a vector with a mapping function that maps the keys
+to the dense integer range.
+
+This is useful for cases like virtual registers in the LLVM code generator: they
+have a dense mapping that is offset by a compile-time constant (the first
+virtual register ID).
+
+.. _dss_densemap:
+
+llvm/ADT/DenseMap.h
+^^^^^^^^^^^^^^^^^^^
+
+DenseMap is a simple quadratically probed hash table.  It excels at supporting
+small keys and values: it uses a single allocation to hold all of the pairs
+that are currently inserted in the map.  DenseMap is a great way to map
+pointers to pointers, or map other small types to each other.
+
+There are several aspects of DenseMap that you should be aware of, however.
+The iterators in a DenseMap are invalidated whenever an insertion occurs,
+unlike map.  Also, because DenseMap allocates space for a large number of
+key/value pairs (it starts with 64 by default), it will waste a lot of space if
+your keys or values are large.  Finally, you must implement a partial
+specialization of DenseMapInfo for the key that you want, if it isn't already
+supported.  This is required to tell DenseMap about two special marker values
+(which can never be inserted into the map) that it needs internally.
+
+DenseMap's find_as() method supports lookup operations using an alternate key
+type.  This is useful in cases where the normal key type is expensive to
+construct, but cheap to compare against.  The DenseMapInfo is responsible for
+defining the appropriate comparison and hashing methods for each alternate key
+type used.
+
+.. _dss_valuemap:
+
+llvm/IR/ValueMap.h
+^^^^^^^^^^^^^^^^^^^
+
+ValueMap is a wrapper around a :ref:`DenseMap <dss_densemap>` mapping
+``Value*``\ s (or subclasses) to another type.  When a Value is deleted or
+RAUW'ed, ValueMap will update itself so the new version of the key is mapped to
+the same value, just as if the key were a WeakVH.  You can configure exactly how
+this happens, and what else happens on these two events, by passing a ``Config``
+parameter to the ValueMap template.
+
+.. _dss_intervalmap:
+
+llvm/ADT/IntervalMap.h
+^^^^^^^^^^^^^^^^^^^^^^
+
+IntervalMap is a compact map for small keys and values.  It maps key intervals
+instead of single keys, and it will automatically coalesce adjacent intervals.
+When the map only contains a few intervals, they are stored in the map object
+itself to avoid allocations.
+
+The IntervalMap iterators are quite big, so they should not be passed around as
+STL iterators.  The heavyweight iterators allow a smaller data structure.
+
+.. _dss_map:
+
+<map>
+^^^^^
+
+std::map has similar characteristics to :ref:`std::set <dss_set>`: it uses a
+single allocation per pair inserted into the map, it offers log(n) lookup with
+an extremely large constant factor, imposes a space penalty of 3 pointers per
+pair in the map, etc.
+
+std::map is most useful when your keys or values are very large, if you need to
+iterate over the collection in sorted order, or if you need stable iterators
+into the map (i.e. they don't get invalidated if an insertion or deletion of
+another element takes place).
+
+.. _dss_mapvector:
+
+llvm/ADT/MapVector.h
+^^^^^^^^^^^^^^^^^^^^
+
+``MapVector<KeyT,ValueT>`` provides a subset of the DenseMap interface.  The
+main difference is that the iteration order is guaranteed to be the insertion
+order, making it an easy (but somewhat expensive) solution for non-deterministic
+iteration over maps of pointers.
+
+It is implemented by mapping from key to an index in a vector of key,value
+pairs.  This provides fast lookup and iteration, but has two main drawbacks:
+the key is stored twice and removing elements takes linear time.  If it is
+necessary to remove elements, it's best to remove them in bulk using
+``remove_if()``.
+
+.. _dss_inteqclasses:
+
+llvm/ADT/IntEqClasses.h
+^^^^^^^^^^^^^^^^^^^^^^^
+
+IntEqClasses provides a compact representation of equivalence classes of small
+integers.  Initially, each integer in the range 0..n-1 has its own equivalence
+class.  Classes can be joined by passing two class representatives to the
+join(a, b) method.  Two integers are in the same class when findLeader() returns
+the same representative.
+
+Once all equivalence classes are formed, the map can be compressed so each
+integer 0..n-1 maps to an equivalence class number in the range 0..m-1, where m
+is the total number of equivalence classes.  The map must be uncompressed before
+it can be edited again.
+
+.. _dss_immutablemap:
+
+llvm/ADT/ImmutableMap.h
+^^^^^^^^^^^^^^^^^^^^^^^
+
+ImmutableMap is an immutable (functional) map implementation based on an AVL
+tree.  Adding or removing elements is done through a Factory object and results
+in the creation of a new ImmutableMap object.  If an ImmutableMap already exists
+with the given key set, then the existing one is returned; equality is compared
+with a FoldingSetNodeID.  The time and space complexity of add or remove
+operations is logarithmic in the size of the original map.
+
+.. _dss_othermap:
+
+Other Map-Like Container Options
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The STL provides several other options, such as std::multimap and the various
+"hash_map" like containers (whether from C++ TR1 or from the SGI library).  We
+never use hash_set and unordered_set because they are generally very expensive
+(each insertion requires a malloc) and very non-portable.
+
+std::multimap is useful if you want to map a key to multiple values, but has all
+the drawbacks of std::map.  A sorted vector or some other approach is almost
+always better.
+
+.. _ds_bit:
+
+Bit storage containers (BitVector, SparseBitVector)
+---------------------------------------------------
+
+Unlike the other containers, there are only two bit storage containers, and
+choosing when to use each is relatively straightforward.
+
+One additional option is ``std::vector<bool>``: we discourage its use for two
+reasons 1) the implementation in many common compilers (e.g.  commonly
+available versions of GCC) is extremely inefficient and 2) the C++ standards
+committee is likely to deprecate this container and/or change it significantly
+somehow.  In any case, please don't use it.
+
+.. _dss_bitvector:
+
+BitVector
+^^^^^^^^^
+
+The BitVector container provides a dynamic size set of bits for manipulation.
+It supports individual bit setting/testing, as well as set operations.  The set
+operations take time O(size of bitvector), but operations are performed one word
+at a time, instead of one bit at a time.  This makes the BitVector very fast for
+set operations compared to other containers.  Use the BitVector when you expect
+the number of set bits to be high (i.e. a dense set).
+
+.. _dss_smallbitvector:
+
+SmallBitVector
+^^^^^^^^^^^^^^
+
+The SmallBitVector container provides the same interface as BitVector, but it is
+optimized for the case where only a small number of bits, less than 25 or so,
+are needed.  It also transparently supports larger bit counts, but slightly less
+efficiently than a plain BitVector, so SmallBitVector should only be used when
+larger counts are rare.
+
+At this time, SmallBitVector does not support set operations (and, or, xor), and
+its operator[] does not provide an assignable lvalue.
+
+.. _dss_sparsebitvector:
+
+SparseBitVector
+^^^^^^^^^^^^^^^
+
+The SparseBitVector container is much like BitVector, with one major difference:
+Only the bits that are set, are stored.  This makes the SparseBitVector much
+more space efficient than BitVector when the set is sparse, as well as making
+set operations O(number of set bits) instead of O(size of universe).  The
+downside to the SparseBitVector is that setting and testing of random bits is
+O(N), and on large SparseBitVectors, this can be slower than BitVector.  In our
+implementation, setting or testing bits in sorted order (either forwards or
+reverse) is O(1) worst case.  Testing and setting bits within 128 bits (depends
+on size) of the current bit is also O(1).  As a general statement,
+testing/setting bits in a SparseBitVector is O(distance away from last set bit).
+
+.. _debugging:
+
+Debugging
+=========
+
+A handful of `GDB pretty printers
+<https://sourceware.org/gdb/onlinedocs/gdb/Pretty-Printing.html>`__ are
+provided for some of the core LLVM libraries. To use them, execute the
+following (or add it to your ``~/.gdbinit``)::
+
+  source /path/to/llvm/src/utils/gdb-scripts/prettyprinters.py
+
+It also might be handy to enable the `print pretty
+<http://ftp.gnu.org/old-gnu/Manuals/gdb/html_node/gdb_57.html>`__ option to
+avoid data structures being printed as a big block of text.
+
+.. _common:
+
+Helpful Hints for Common Operations
+===================================
+
+This section describes how to perform some very simple transformations of LLVM
+code.  This is meant to give examples of common idioms used, showing the
+practical side of LLVM transformations.
+
+Because this is a "how-to" section, you should also read about the main classes
+that you will be working with.  The :ref:`Core LLVM Class Hierarchy Reference
+<coreclasses>` contains details and descriptions of the main classes that you
+should know about.
+
+.. _inspection:
+
+Basic Inspection and Traversal Routines
+---------------------------------------
+
+The LLVM compiler infrastructure have many different data structures that may be
+traversed.  Following the example of the C++ standard template library, the
+techniques used to traverse these various data structures are all basically the
+same.  For a enumerable sequence of values, the ``XXXbegin()`` function (or
+method) returns an iterator to the start of the sequence, the ``XXXend()``
+function returns an iterator pointing to one past the last valid element of the
+sequence, and there is some ``XXXiterator`` data type that is common between the
+two operations.
+
+Because the pattern for iteration is common across many different aspects of the
+program representation, the standard template library algorithms may be used on
+them, and it is easier to remember how to iterate.  First we show a few common
+examples of the data structures that need to be traversed.  Other data
+structures are traversed in very similar ways.
+
+.. _iterate_function:
+
+Iterating over the ``BasicBlock`` in a ``Function``
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+It's quite common to have a ``Function`` instance that you'd like to transform
+in some way; in particular, you'd like to manipulate its ``BasicBlock``\ s.  To
+facilitate this, you'll need to iterate over all of the ``BasicBlock``\ s that
+constitute the ``Function``.  The following is an example that prints the name
+of a ``BasicBlock`` and the number of ``Instruction``\ s it contains:
+
+.. code-block:: c++
+
+  // func is a pointer to a Function instance
+  for (Function::iterator i = func->begin(), e = func->end(); i != e; ++i)
+    // Print out the name of the basic block if it has one, and then the
+    // number of instructions that it contains
+    errs() << "Basic block (name=" << i->getName() << ") has "
+               << i->size() << " instructions.\n";
+
+Note that i can be used as if it were a pointer for the purposes of invoking
+member functions of the ``Instruction`` class.  This is because the indirection
+operator is overloaded for the iterator classes.  In the above code, the
+expression ``i->size()`` is exactly equivalent to ``(*i).size()`` just like
+you'd expect.
+
+.. _iterate_basicblock:
+
+Iterating over the ``Instruction`` in a ``BasicBlock``
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Just like when dealing with ``BasicBlock``\ s in ``Function``\ s, it's easy to
+iterate over the individual instructions that make up ``BasicBlock``\ s.  Here's
+a code snippet that prints out each instruction in a ``BasicBlock``:
+
+.. code-block:: c++
+
+  // blk is a pointer to a BasicBlock instance
+  for (BasicBlock::iterator i = blk->begin(), e = blk->end(); i != e; ++i)
+     // The next statement works since operator<<(ostream&,...)
+     // is overloaded for Instruction&
+     errs() << *i << "\n";
+
+
+However, this isn't really the best way to print out the contents of a
+``BasicBlock``!  Since the ostream operators are overloaded for virtually
+anything you'll care about, you could have just invoked the print routine on the
+basic block itself: ``errs() << *blk << "\n";``.
+
+.. _iterate_insiter:
+
+Iterating over the ``Instruction`` in a ``Function``
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+If you're finding that you commonly iterate over a ``Function``'s
+``BasicBlock``\ s and then that ``BasicBlock``'s ``Instruction``\ s,
+``InstIterator`` should be used instead.  You'll need to include
+``llvm/IR/InstIterator.h`` (`doxygen
+<http://llvm.org/doxygen/InstIterator_8h.html>`__) and then instantiate
+``InstIterator``\ s explicitly in your code.  Here's a small example that shows
+how to dump all instructions in a function to the standard error stream:
+
+.. code-block:: c++
+
+  #include "llvm/IR/InstIterator.h"
+
+  // F is a pointer to a Function instance
+  for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
+    errs() << *I << "\n";
+
+Easy, isn't it?  You can also use ``InstIterator``\ s to fill a work list with
+its initial contents.  For example, if you wanted to initialize a work list to
+contain all instructions in a ``Function`` F, all you would need to do is
+something like:
+
+.. code-block:: c++
+
+  std::set<Instruction*> worklist;
+  // or better yet, SmallPtrSet<Instruction*, 64> worklist;
+
+  for (inst_iterator I = inst_begin(F), E = inst_end(F); I != E; ++I)
+    worklist.insert(&*I);
+
+The STL set ``worklist`` would now contain all instructions in the ``Function``
+pointed to by F.
+
+.. _iterate_convert:
+
+Turning an iterator into a class pointer (and vice-versa)
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Sometimes, it'll be useful to grab a reference (or pointer) to a class instance
+when all you've got at hand is an iterator.  Well, extracting a reference or a
+pointer from an iterator is very straight-forward.  Assuming that ``i`` is a
+``BasicBlock::iterator`` and ``j`` is a ``BasicBlock::const_iterator``:
+
+.. code-block:: c++
+
+  Instruction& inst = *i;   // Grab reference to instruction reference
+  Instruction* pinst = &*i; // Grab pointer to instruction reference
+  const Instruction& inst = *j;
+
+However, the iterators you'll be working with in the LLVM framework are special:
+they will automatically convert to a ptr-to-instance type whenever they need to.
+Instead of dereferencing the iterator and then taking the address of the result,
+you can simply assign the iterator to the proper pointer type and you get the
+dereference and address-of operation as a result of the assignment (behind the
+scenes, this is a result of overloading casting mechanisms).  Thus the second
+line of the last example,
+
+.. code-block:: c++
+
+  Instruction *pinst = &*i;
+
+is semantically equivalent to
+
+.. code-block:: c++
+
+  Instruction *pinst = i;
+
+It's also possible to turn a class pointer into the corresponding iterator, and
+this is a constant time operation (very efficient).  The following code snippet
+illustrates use of the conversion constructors provided by LLVM iterators.  By
+using these, you can explicitly grab the iterator of something without actually
+obtaining it via iteration over some structure:
+
+.. code-block:: c++
+
+  void printNextInstruction(Instruction* inst) {
+    BasicBlock::iterator it(inst);
+    ++it; // After this line, it refers to the instruction after *inst
+    if (it != inst->getParent()->end()) errs() << *it << "\n";
+  }
+
+Unfortunately, these implicit conversions come at a cost; they prevent these
+iterators from conforming to standard iterator conventions, and thus from being
+usable with standard algorithms and containers.  For example, they prevent the
+following code, where ``B`` is a ``BasicBlock``, from compiling:
+
+.. code-block:: c++
+
+  llvm::SmallVector<llvm::Instruction *, 16>(B->begin(), B->end());
+
+Because of this, these implicit conversions may be removed some day, and
+``operator*`` changed to return a pointer instead of a reference.
+
+.. _iterate_complex:
+
+Finding call sites: a slightly more complex example
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Say that you're writing a FunctionPass and would like to count all the locations
+in the entire module (that is, across every ``Function``) where a certain
+function (i.e., some ``Function *``) is already in scope.  As you'll learn
+later, you may want to use an ``InstVisitor`` to accomplish this in a much more
+straight-forward manner, but this example will allow us to explore how you'd do
+it if you didn't have ``InstVisitor`` around.  In pseudo-code, this is what we
+want to do:
+
+.. code-block:: none
+
+  initialize callCounter to zero
+  for each Function f in the Module
+    for each BasicBlock b in f
+      for each Instruction i in b
+        if (i is a CallInst and calls the given function)
+          increment callCounter
+
+And the actual code is (remember, because we're writing a ``FunctionPass``, our
+``FunctionPass``-derived class simply has to override the ``runOnFunction``
+method):
+
+.. code-block:: c++
+
+  Function* targetFunc = ...;
+
+  class OurFunctionPass : public FunctionPass {
+    public:
+      OurFunctionPass(): callCounter(0) { }
+
+      virtual runOnFunction(Function& F) {
+        for (Function::iterator b = F.begin(), be = F.end(); b != be; ++b) {
+          for (BasicBlock::iterator i = b->begin(), ie = b->end(); i != ie; ++i) {
+            if (CallInst* callInst = dyn_cast<CallInst>(&*i)) {
+              // We know we've encountered a call instruction, so we
+              // need to determine if it's a call to the
+              // function pointed to by m_func or not.
+              if (callInst->getCalledFunction() == targetFunc)
+                ++callCounter;
+            }
+          }
+        }
+      }
+
+    private:
+      unsigned callCounter;
+  };
+
+.. _calls_and_invokes:
+
+Treating calls and invokes the same way
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+You may have noticed that the previous example was a bit oversimplified in that
+it did not deal with call sites generated by 'invoke' instructions.  In this,
+and in other situations, you may find that you want to treat ``CallInst``\ s and
+``InvokeInst``\ s the same way, even though their most-specific common base
+class is ``Instruction``, which includes lots of less closely-related things.
+For these cases, LLVM provides a handy wrapper class called ``CallSite``
+(`doxygen <http://llvm.org/doxygen/classllvm_1_1CallSite.html>`__) It is
+essentially a wrapper around an ``Instruction`` pointer, with some methods that
+provide functionality common to ``CallInst``\ s and ``InvokeInst``\ s.
+
+This class has "value semantics": it should be passed by value, not by reference
+and it should not be dynamically allocated or deallocated using ``operator new``
+or ``operator delete``.  It is efficiently copyable, assignable and
+constructable, with costs equivalents to that of a bare pointer.  If you look at
+its definition, it has only a single pointer member.
+
+.. _iterate_chains:
+
+Iterating over def-use & use-def chains
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Frequently, we might have an instance of the ``Value`` class (`doxygen
+<http://llvm.org/doxygen/classllvm_1_1Value.html>`__) and we want to determine
+which ``User`` s use the ``Value``.  The list of all ``User``\ s of a particular
+``Value`` is called a *def-use* chain.  For example, let's say we have a
+``Function*`` named ``F`` to a particular function ``foo``.  Finding all of the
+instructions that *use* ``foo`` is as simple as iterating over the *def-use*
+chain of ``F``:
+
+.. code-block:: c++
+
+  Function *F = ...;
+
+  for (User *U : F->users()) {
+    if (Instruction *Inst = dyn_cast<Instruction>(U)) {
+      errs() << "F is used in instruction:\n";
+      errs() << *Inst << "\n";
+    }
+
+Alternatively, it's common to have an instance of the ``User`` Class (`doxygen
+<http://llvm.org/doxygen/classllvm_1_1User.html>`__) and need to know what
+``Value``\ s are used by it.  The list of all ``Value``\ s used by a ``User`` is
+known as a *use-def* chain.  Instances of class ``Instruction`` are common
+``User`` s, so we might want to iterate over all of the values that a particular
+instruction uses (that is, the operands of the particular ``Instruction``):
+
+.. code-block:: c++
+
+  Instruction *pi = ...;
+
+  for (Use &U : pi->operands()) {
+    Value *v = U.get();
+    // ...
+  }
+
+Declaring objects as ``const`` is an important tool of enforcing mutation free
+algorithms (such as analyses, etc.).  For this purpose above iterators come in
+constant flavors as ``Value::const_use_iterator`` and
+``Value::const_op_iterator``.  They automatically arise when calling
+``use/op_begin()`` on ``const Value*``\ s or ``const User*``\ s respectively.
+Upon dereferencing, they return ``const Use*``\ s.  Otherwise the above patterns
+remain unchanged.
+
+.. _iterate_preds:
+
+Iterating over predecessors & successors of blocks
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Iterating over the predecessors and successors of a block is quite easy with the
+routines defined in ``"llvm/IR/CFG.h"``.  Just use code like this to
+iterate over all predecessors of BB:
+
+.. code-block:: c++
+
+  #include "llvm/IR/CFG.h"
+  BasicBlock *BB = ...;
+
+  for (pred_iterator PI = pred_begin(BB), E = pred_end(BB); PI != E; ++PI) {
+    BasicBlock *Pred = *PI;
+    // ...
+  }
+
+Similarly, to iterate over successors use ``succ_iterator/succ_begin/succ_end``.
+
+.. _simplechanges:
+
+Making simple changes
+---------------------
+
+There are some primitive transformation operations present in the LLVM
+infrastructure that are worth knowing about.  When performing transformations,
+it's fairly common to manipulate the contents of basic blocks.  This section
+describes some of the common methods for doing so and gives example code.
+
+.. _schanges_creating:
+
+Creating and inserting new ``Instruction``\ s
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+*Instantiating Instructions*
+
+Creation of ``Instruction``\ s is straight-forward: simply call the constructor
+for the kind of instruction to instantiate and provide the necessary parameters.
+For example, an ``AllocaInst`` only *requires* a (const-ptr-to) ``Type``.  Thus:
+
+.. code-block:: c++
+
+  AllocaInst* ai = new AllocaInst(Type::Int32Ty);
+
+will create an ``AllocaInst`` instance that represents the allocation of one
+integer in the current stack frame, at run time.  Each ``Instruction`` subclass
+is likely to have varying default parameters which change the semantics of the
+instruction, so refer to the `doxygen documentation for the subclass of
+Instruction <http://llvm.org/doxygen/classllvm_1_1Instruction.html>`_ that
+you're interested in instantiating.
+
+*Naming values*
+
+It is very useful to name the values of instructions when you're able to, as
+this facilitates the debugging of your transformations.  If you end up looking
+at generated LLVM machine code, you definitely want to have logical names
+associated with the results of instructions!  By supplying a value for the
+``Name`` (default) parameter of the ``Instruction`` constructor, you associate a
+logical name with the result of the instruction's execution at run time.  For
+example, say that I'm writing a transformation that dynamically allocates space
+for an integer on the stack, and that integer is going to be used as some kind
+of index by some other code.  To accomplish this, I place an ``AllocaInst`` at
+the first point in the first ``BasicBlock`` of some ``Function``, and I'm
+intending to use it within the same ``Function``.  I might do:
+
+.. code-block:: c++
+
+  AllocaInst* pa = new AllocaInst(Type::Int32Ty, 0, "indexLoc");
+
+where ``indexLoc`` is now the logical name of the instruction's execution value,
+which is a pointer to an integer on the run time stack.
+
+*Inserting instructions*
+
+There are essentially three ways to insert an ``Instruction`` into an existing
+sequence of instructions that form a ``BasicBlock``:
+
+* Insertion into an explicit instruction list
+
+  Given a ``BasicBlock* pb``, an ``Instruction* pi`` within that ``BasicBlock``,
+  and a newly-created instruction we wish to insert before ``*pi``, we do the
+  following:
+
+  .. code-block:: c++
+
+      BasicBlock *pb = ...;
+      Instruction *pi = ...;
+      Instruction *newInst = new Instruction(...);
+
+      pb->getInstList().insert(pi, newInst); // Inserts newInst before pi in pb
+
+  Appending to the end of a ``BasicBlock`` is so common that the ``Instruction``
+  class and ``Instruction``-derived classes provide constructors which take a
+  pointer to a ``BasicBlock`` to be appended to.  For example code that looked
+  like:
+
+  .. code-block:: c++
+
+    BasicBlock *pb = ...;
+    Instruction *newInst = new Instruction(...);
+
+    pb->getInstList().push_back(newInst); // Appends newInst to pb
+
+  becomes:
+
+  .. code-block:: c++
+
+    BasicBlock *pb = ...;
+    Instruction *newInst = new Instruction(..., pb);
+
+  which is much cleaner, especially if you are creating long instruction
+  streams.
+
+* Insertion into an implicit instruction list
+
+  ``Instruction`` instances that are already in ``BasicBlock``\ s are implicitly
+  associated with an existing instruction list: the instruction list of the
+  enclosing basic block.  Thus, we could have accomplished the same thing as the
+  above code without being given a ``BasicBlock`` by doing:
+
+  .. code-block:: c++
+
+    Instruction *pi = ...;
+    Instruction *newInst = new Instruction(...);
+
+    pi->getParent()->getInstList().insert(pi, newInst);
+
+  In fact, this sequence of steps occurs so frequently that the ``Instruction``
+  class and ``Instruction``-derived classes provide constructors which take (as
+  a default parameter) a pointer to an ``Instruction`` which the newly-created
+  ``Instruction`` should precede.  That is, ``Instruction`` constructors are
+  capable of inserting the newly-created instance into the ``BasicBlock`` of a
+  provided instruction, immediately before that instruction.  Using an
+  ``Instruction`` constructor with a ``insertBefore`` (default) parameter, the
+  above code becomes:
+
+  .. code-block:: c++
+
+    Instruction* pi = ...;
+    Instruction* newInst = new Instruction(..., pi);
+
+  which is much cleaner, especially if you're creating a lot of instructions and
+  adding them to ``BasicBlock``\ s.
+
+* Insertion using an instance of ``IRBuilder``
+
+  Inserting several ``Instruction``\ s can be quite laborious using the previous
+  methods. The ``IRBuilder`` is a convenience class that can be used to add
+  several instructions to the end of a ``BasicBlock`` or before a particular
+  ``Instruction``. It also supports constant folding and renaming named
+  registers (see ``IRBuilder``'s template arguments).
+
+  The example below demonstrates a very simple use of the ``IRBuilder`` where
+  three instructions are inserted before the instruction ``pi``. The first two
+  instructions are Call instructions and third instruction multiplies the return
+  value of the two calls.
+
+  .. code-block:: c++
+
+    Instruction *pi = ...;
+    IRBuilder<> Builder(pi);
+    CallInst* callOne = Builder.CreateCall(...);
+    CallInst* callTwo = Builder.CreateCall(...);
+    Value* result = Builder.CreateMul(callOne, callTwo);
+
+  The example below is similar to the above example except that the created
+  ``IRBuilder`` inserts instructions at the end of the ``BasicBlock`` ``pb``.
+
+  .. code-block:: c++
+
+    BasicBlock *pb = ...;
+    IRBuilder<> Builder(pb);
+    CallInst* callOne = Builder.CreateCall(...);
+    CallInst* callTwo = Builder.CreateCall(...);
+    Value* result = Builder.CreateMul(callOne, callTwo);
+
+  See :doc:`tutorial/LangImpl03` for a practical use of the ``IRBuilder``.
+
+
+.. _schanges_deleting:
+
+Deleting Instructions
+^^^^^^^^^^^^^^^^^^^^^
+
+Deleting an instruction from an existing sequence of instructions that form a
+BasicBlock_ is very straight-forward: just call the instruction's
+``eraseFromParent()`` method.  For example:
+
+.. code-block:: c++
+
+  Instruction *I = .. ;
+  I->eraseFromParent();
+
+This unlinks the instruction from its containing basic block and deletes it.  If
+you'd just like to unlink the instruction from its containing basic block but
+not delete it, you can use the ``removeFromParent()`` method.
+
+.. _schanges_replacing:
+
+Replacing an Instruction with another Value
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Replacing individual instructions
+"""""""""""""""""""""""""""""""""
+
+Including "`llvm/Transforms/Utils/BasicBlockUtils.h
+<http://llvm.org/doxygen/BasicBlockUtils_8h-source.html>`_" permits use of two
+very useful replace functions: ``ReplaceInstWithValue`` and
+``ReplaceInstWithInst``.
+
+.. _schanges_deleting_sub:
+
+Deleting Instructions
+"""""""""""""""""""""
+
+* ``ReplaceInstWithValue``
+
+  This function replaces all uses of a given instruction with a value, and then
+  removes the original instruction.  The following example illustrates the
+  replacement of the result of a particular ``AllocaInst`` that allocates memory
+  for a single integer with a null pointer to an integer.
+
+  .. code-block:: c++
+
+    AllocaInst* instToReplace = ...;
+    BasicBlock::iterator ii(instToReplace);
+
+    ReplaceInstWithValue(instToReplace->getParent()->getInstList(), ii,
+                         Constant::getNullValue(PointerType::getUnqual(Type::Int32Ty)));
+
+* ``ReplaceInstWithInst``
+
+  This function replaces a particular instruction with another instruction,
+  inserting the new instruction into the basic block at the location where the
+  old instruction was, and replacing any uses of the old instruction with the
+  new instruction.  The following example illustrates the replacement of one
+  ``AllocaInst`` with another.
+
+  .. code-block:: c++
+
+    AllocaInst* instToReplace = ...;
+    BasicBlock::iterator ii(instToReplace);
+
+    ReplaceInstWithInst(instToReplace->getParent()->getInstList(), ii,
+                        new AllocaInst(Type::Int32Ty, 0, "ptrToReplacedInt"));
+
+
+Replacing multiple uses of Users and Values
+"""""""""""""""""""""""""""""""""""""""""""
+
+You can use ``Value::replaceAllUsesWith`` and ``User::replaceUsesOfWith`` to
+change more than one use at a time.  See the doxygen documentation for the
+`Value Class <http://llvm.org/doxygen/classllvm_1_1Value.html>`_ and `User Class
+<http://llvm.org/doxygen/classllvm_1_1User.html>`_, respectively, for more
+information.
+
+.. _schanges_deletingGV:
+
+Deleting GlobalVariables
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+Deleting a global variable from a module is just as easy as deleting an
+Instruction.  First, you must have a pointer to the global variable that you
+wish to delete.  You use this pointer to erase it from its parent, the module.
+For example:
+
+.. code-block:: c++
+
+  GlobalVariable *GV = .. ;
+
+  GV->eraseFromParent();
+
+
+.. _create_types:
+
+How to Create Types
+-------------------
+
+In generating IR, you may need some complex types.  If you know these types
+statically, you can use ``TypeBuilder<...>::get()``, defined in
+``llvm/Support/TypeBuilder.h``, to retrieve them.  ``TypeBuilder`` has two forms
+depending on whether you're building types for cross-compilation or native
+library use.  ``TypeBuilder<T, true>`` requires that ``T`` be independent of the
+host environment, meaning that it's built out of types from the ``llvm::types``
+(`doxygen <http://llvm.org/doxygen/namespacellvm_1_1types.html>`__) namespace
+and pointers, functions, arrays, etc. built of those.  ``TypeBuilder<T, false>``
+additionally allows native C types whose size may depend on the host compiler.
+For example,
+
+.. code-block:: c++
+
+  FunctionType *ft = TypeBuilder<types::i<8>(types::i<32>*), true>::get();
+
+is easier to read and write than the equivalent
+
+.. code-block:: c++
+
+  std::vector<const Type*> params;
+  params.push_back(PointerType::getUnqual(Type::Int32Ty));
+  FunctionType *ft = FunctionType::get(Type::Int8Ty, params, false);
+
+See the `class comment
+<http://llvm.org/doxygen/TypeBuilder_8h-source.html#l00001>`_ for more details.
+
+.. _threading:
+
+Threads and LLVM
+================
+
+This section describes the interaction of the LLVM APIs with multithreading,
+both on the part of client applications, and in the JIT, in the hosted
+application.
+
+Note that LLVM's support for multithreading is still relatively young.  Up
+through version 2.5, the execution of threaded hosted applications was
+supported, but not threaded client access to the APIs.  While this use case is
+now supported, clients *must* adhere to the guidelines specified below to ensure
+proper operation in multithreaded mode.
+
+Note that, on Unix-like platforms, LLVM requires the presence of GCC's atomic
+intrinsics in order to support threaded operation.  If you need a
+multhreading-capable LLVM on a platform without a suitably modern system
+compiler, consider compiling LLVM and LLVM-GCC in single-threaded mode, and
+using the resultant compiler to build a copy of LLVM with multithreading
+support.
+
+.. _shutdown:
+
+Ending Execution with ``llvm_shutdown()``
+-----------------------------------------
+
+When you are done using the LLVM APIs, you should call ``llvm_shutdown()`` to
+deallocate memory used for internal structures.
+
+.. _managedstatic:
+
+Lazy Initialization with ``ManagedStatic``
+------------------------------------------
+
+``ManagedStatic`` is a utility class in LLVM used to implement static
+initialization of static resources, such as the global type tables.  In a
+single-threaded environment, it implements a simple lazy initialization scheme.
+When LLVM is compiled with support for multi-threading, however, it uses
+double-checked locking to implement thread-safe lazy initialization.
+
+.. _llvmcontext:
+
+Achieving Isolation with ``LLVMContext``
+----------------------------------------
+
+``LLVMContext`` is an opaque class in the LLVM API which clients can use to
+operate multiple, isolated instances of LLVM concurrently within the same
+address space.  For instance, in a hypothetical compile-server, the compilation
+of an individual translation unit is conceptually independent from all the
+others, and it would be desirable to be able to compile incoming translation
+units concurrently on independent server threads.  Fortunately, ``LLVMContext``
+exists to enable just this kind of scenario!
+
+Conceptually, ``LLVMContext`` provides isolation.  Every LLVM entity
+(``Module``\ s, ``Value``\ s, ``Type``\ s, ``Constant``\ s, etc.) in LLVM's
+in-memory IR belongs to an ``LLVMContext``.  Entities in different contexts
+*cannot* interact with each other: ``Module``\ s in different contexts cannot be
+linked together, ``Function``\ s cannot be added to ``Module``\ s in different
+contexts, etc.  What this means is that is is safe to compile on multiple
+threads simultaneously, as long as no two threads operate on entities within the
+same context.
+
+In practice, very few places in the API require the explicit specification of a
+``LLVMContext``, other than the ``Type`` creation/lookup APIs.  Because every
+``Type`` carries a reference to its owning context, most other entities can
+determine what context they belong to by looking at their own ``Type``.  If you
+are adding new entities to LLVM IR, please try to maintain this interface
+design.
+
+.. _jitthreading:
+
+Threads and the JIT
+-------------------
+
+LLVM's "eager" JIT compiler is safe to use in threaded programs.  Multiple
+threads can call ``ExecutionEngine::getPointerToFunction()`` or
+``ExecutionEngine::runFunction()`` concurrently, and multiple threads can run
+code output by the JIT concurrently.  The user must still ensure that only one
+thread accesses IR in a given ``LLVMContext`` while another thread might be
+modifying it.  One way to do that is to always hold the JIT lock while accessing
+IR outside the JIT (the JIT *modifies* the IR by adding ``CallbackVH``\ s).
+Another way is to only call ``getPointerToFunction()`` from the
+``LLVMContext``'s thread.
+
+When the JIT is configured to compile lazily (using
+``ExecutionEngine::DisableLazyCompilation(false)``), there is currently a `race
+condition <http://llvm.org/bugs/show_bug.cgi?id=5184>`_ in updating call sites
+after a function is lazily-jitted.  It's still possible to use the lazy JIT in a
+threaded program if you ensure that only one thread at a time can call any
+particular lazy stub and that the JIT lock guards any IR access, but we suggest
+using only the eager JIT in threaded programs.
+
+.. _advanced:
+
+Advanced Topics
+===============
+
+This section describes some of the advanced or obscure API's that most clients
+do not need to be aware of.  These API's tend manage the inner workings of the
+LLVM system, and only need to be accessed in unusual circumstances.
+
+.. _SymbolTable:
+
+The ``ValueSymbolTable`` class
+------------------------------
+
+The ``ValueSymbolTable`` (`doxygen
+<http://llvm.org/doxygen/classllvm_1_1ValueSymbolTable.html>`__) class provides
+a symbol table that the :ref:`Function <c_Function>` and Module_ classes use for
+naming value definitions.  The symbol table can provide a name for any Value_.
+
+Note that the ``SymbolTable`` class should not be directly accessed by most
+clients.  It should only be used when iteration over the symbol table names
+themselves are required, which is very special purpose.  Note that not all LLVM
+Value_\ s have names, and those without names (i.e. they have an empty name) do
+not exist in the symbol table.
+
+Symbol tables support iteration over the values in the symbol table with
+``begin/end/iterator`` and supports querying to see if a specific name is in the
+symbol table (with ``lookup``).  The ``ValueSymbolTable`` class exposes no
+public mutator methods, instead, simply call ``setName`` on a value, which will
+autoinsert it into the appropriate symbol table.
+
+.. _UserLayout:
+
+The ``User`` and owned ``Use`` classes' memory layout
+-----------------------------------------------------
+
+The ``User`` (`doxygen <http://llvm.org/doxygen/classllvm_1_1User.html>`__)
+class provides a basis for expressing the ownership of ``User`` towards other
+`Value instance <http://llvm.org/doxygen/classllvm_1_1Value.html>`_\ s.  The
+``Use`` (`doxygen <http://llvm.org/doxygen/classllvm_1_1Use.html>`__) helper
+class is employed to do the bookkeeping and to facilitate *O(1)* addition and
+removal.
+
+.. _Use2User:
+
+Interaction and relationship between ``User`` and ``Use`` objects
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+A subclass of ``User`` can choose between incorporating its ``Use`` objects or
+refer to them out-of-line by means of a pointer.  A mixed variant (some ``Use``
+s inline others hung off) is impractical and breaks the invariant that the
+``Use`` objects belonging to the same ``User`` form a contiguous array.
+
+We have 2 different layouts in the ``User`` (sub)classes:
+
+* Layout a)
+
+  The ``Use`` object(s) are inside (resp. at fixed offset) of the ``User``
+  object and there are a fixed number of them.
+
+* Layout b)
+
+  The ``Use`` object(s) are referenced by a pointer to an array from the
+  ``User`` object and there may be a variable number of them.
+
+As of v2.4 each layout still possesses a direct pointer to the start of the
+array of ``Use``\ s.  Though not mandatory for layout a), we stick to this
+redundancy for the sake of simplicity.  The ``User`` object also stores the
+number of ``Use`` objects it has. (Theoretically this information can also be
+calculated given the scheme presented below.)
+
+Special forms of allocation operators (``operator new``) enforce the following
+memory layouts:
+
+* Layout a) is modelled by prepending the ``User`` object by the ``Use[]``
+  array.
+
+  .. code-block:: none
+
+    ...---.---.---.---.-------...
+      | P | P | P | P | User
+    '''---'---'---'---'-------'''
+
+* Layout b) is modelled by pointing at the ``Use[]`` array.
+
+  .. code-block:: none
+
+    .-------...
+    | User
+    '-------'''
+        |
+        v
+        .---.---.---.---...
+        | P | P | P | P |
+        '---'---'---'---'''
+
+*(In the above figures* '``P``' *stands for the* ``Use**`` *that is stored in
+each* ``Use`` *object in the member* ``Use::Prev`` *)*
+
+.. _Waymarking:
+
+The waymarking algorithm
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+Since the ``Use`` objects are deprived of the direct (back)pointer to their
+``User`` objects, there must be a fast and exact method to recover it.  This is
+accomplished by the following scheme:
+
+A bit-encoding in the 2 LSBits (least significant bits) of the ``Use::Prev``
+allows to find the start of the ``User`` object:
+
+* ``00`` --- binary digit 0
+
+* ``01`` --- binary digit 1
+
+* ``10`` --- stop and calculate (``s``)
+
+* ``11`` --- full stop (``S``)
+
+Given a ``Use*``, all we have to do is to walk till we get a stop and we either
+have a ``User`` immediately behind or we have to walk to the next stop picking
+up digits and calculating the offset:
+
+.. code-block:: none
+
+  .---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.---.----------------
+  | 1 | s | 1 | 0 | 1 | 0 | s | 1 | 1 | 0 | s | 1 | 1 | s | 1 | S | User (or User*)
+  '---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'---'----------------
+      |+15                |+10            |+6         |+3     |+1
+      |                   |               |           |       | __>
+      |                   |               |           | __________>
+      |                   |               | ______________________>
+      |                   | ______________________________________>
+      | __________________________________________________________>
+
+Only the significant number of bits need to be stored between the stops, so that
+the *worst case is 20 memory accesses* when there are 1000 ``Use`` objects
+associated with a ``User``.
+
+.. _ReferenceImpl:
+
+Reference implementation
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+The following literate Haskell fragment demonstrates the concept:
+
+.. code-block:: haskell
+
+  > import Test.QuickCheck
+  >
+  > digits :: Int -> [Char] -> [Char]
+  > digits 0 acc = '0' : acc
+  > digits 1 acc = '1' : acc
+  > digits n acc = digits (n `div` 2) $ digits (n `mod` 2) acc
+  >
+  > dist :: Int -> [Char] -> [Char]
+  > dist 0 [] = ['S']
+  > dist 0 acc = acc
+  > dist 1 acc = let r = dist 0 acc in 's' : digits (length r) r
+  > dist n acc = dist (n - 1) $ dist 1 acc
+  >
+  > takeLast n ss = reverse $ take n $ reverse ss
+  >
+  > test = takeLast 40 $ dist 20 []
+  >
+
+Printing <test> gives: ``"1s100000s11010s10100s1111s1010s110s11s1S"``
+
+The reverse algorithm computes the length of the string just by examining a
+certain prefix:
+
+.. code-block:: haskell
+
+  > pref :: [Char] -> Int
+  > pref "S" = 1
+  > pref ('s':'1':rest) = decode 2 1 rest
+  > pref (_:rest) = 1 + pref rest
+  >
+  > decode walk acc ('0':rest) = decode (walk + 1) (acc * 2) rest
+  > decode walk acc ('1':rest) = decode (walk + 1) (acc * 2 + 1) rest
+  > decode walk acc _ = walk + acc
+  >
+
+Now, as expected, printing <pref test> gives ``40``.
+
+We can *quickCheck* this with following property:
+
+.. code-block:: haskell
+
+  > testcase = dist 2000 []
+  > testcaseLength = length testcase
+  >
+  > identityProp n = n > 0 && n <= testcaseLength ==> length arr == pref arr
+  >     where arr = takeLast n testcase
+  >
+
+As expected <quickCheck identityProp> gives:
+
+::
+
+  *Main> quickCheck identityProp
+  OK, passed 100 tests.
+
+Let's be a bit more exhaustive:
+
+.. code-block:: haskell
+
+  >
+  > deepCheck p = check (defaultConfig { configMaxTest = 500 }) p
+  >
+
+And here is the result of <deepCheck identityProp>:
+
+::
+
+  *Main> deepCheck identityProp
+  OK, passed 500 tests.
+
+.. _Tagging:
+
+Tagging considerations
+^^^^^^^^^^^^^^^^^^^^^^
+
+To maintain the invariant that the 2 LSBits of each ``Use**`` in ``Use`` never
+change after being set up, setters of ``Use::Prev`` must re-tag the new
+``Use**`` on every modification.  Accordingly getters must strip the tag bits.
+
+For layout b) instead of the ``User`` we find a pointer (``User*`` with LSBit
+set).  Following this pointer brings us to the ``User``.  A portable trick
+ensures that the first bytes of ``User`` (if interpreted as a pointer) never has
+the LSBit set. (Portability is relying on the fact that all known compilers
+place the ``vptr`` in the first word of the instances.)
+
+.. _polymorphism:
+
+Designing Type Hiercharies and Polymorphic Interfaces
+-----------------------------------------------------
+
+There are two different design patterns that tend to result in the use of
+virtual dispatch for methods in a type hierarchy in C++ programs. The first is
+a genuine type hierarchy where different types in the hierarchy model
+a specific subset of the functionality and semantics, and these types nest
+strictly within each other. Good examples of this can be seen in the ``Value``
+or ``Type`` type hierarchies.
+
+A second is the desire to dispatch dynamically across a collection of
+polymorphic interface implementations. This latter use case can be modeled with
+virtual dispatch and inheritance by defining an abstract interface base class
+which all implementations derive from and override. However, this
+implementation strategy forces an **"is-a"** relationship to exist that is not
+actually meaningful. There is often not some nested hierarchy of useful
+generalizations which code might interact with and move up and down. Instead,
+there is a singular interface which is dispatched across a range of
+implementations.
+
+The preferred implementation strategy for the second use case is that of
+generic programming (sometimes called "compile-time duck typing" or "static
+polymorphism"). For example, a template over some type parameter ``T`` can be
+instantiated across any particular implementation that conforms to the
+interface or *concept*. A good example here is the highly generic properties of
+any type which models a node in a directed graph. LLVM models these primarily
+through templates and generic programming. Such templates include the
+``LoopInfoBase`` and ``DominatorTreeBase``. When this type of polymorphism
+truly needs **dynamic** dispatch you can generalize it using a technique
+called *concept-based polymorphism*. This pattern emulates the interfaces and
+behaviors of templates using a very limited form of virtual dispatch for type
+erasure inside its implementation. You can find examples of this technique in
+the ``PassManager.h`` system, and there is a more detailed introduction to it
+by Sean Parent in several of his talks and papers:
+
+#. `Inheritance Is The Base Class of Evil
+   <http://channel9.msdn.com/Events/GoingNative/2013/Inheritance-Is-The-Base-Class-of-Evil>`_
+   - The GoingNative 2013 talk describing this technique, and probably the best
+   place to start.
+#. `Value Semantics and Concepts-based Polymorphism
+   <http://www.youtube.com/watch?v=_BpMYeUFXv8>`_ - The C++Now! 2012 talk
+   describing this technique in more detail.
+#. `Sean Parent's Papers and Presentations
+   <http://github.com/sean-parent/sean-parent.github.com/wiki/Papers-and-Presentations>`_
+   - A Github project full of links to slides, video, and sometimes code.
+
+When deciding between creating a type hierarchy (with either tagged or virtual
+dispatch) and using templates or concepts-based polymorphism, consider whether
+there is some refinement of an abstract base class which is a semantically
+meaningful type on an interface boundary. If anything more refined than the
+root abstract interface is meaningless to talk about as a partial extension of
+the semantic model, then your use case likely fits better with polymorphism and
+you should avoid using virtual dispatch. However, there may be some exigent
+circumstances that require one technique or the other to be used.
+
+If you do need to introduce a type hierarchy, we prefer to use explicitly
+closed type hierarchies with manual tagged dispatch and/or RTTI rather than the
+open inheritance model and virtual dispatch that is more common in C++ code.
+This is because LLVM rarely encourages library consumers to extend its core
+types, and leverages the closed and tag-dispatched nature of its hierarchies to
+generate significantly more efficient code. We have also found that a large
+amount of our usage of type hierarchies fits better with tag-based pattern
+matching rather than dynamic dispatch across a common interface. Within LLVM we
+have built custom helpers to facilitate this design. See this document's
+section on :ref:`isa and dyn_cast <isa>` and our :doc:`detailed document
+<HowToSetUpLLVMStyleRTTI>` which describes how you can implement this
+pattern for use with the LLVM helpers.
+
+.. _abi_breaking_checks:
+
+ABI Breaking Checks
+-------------------
+
+Checks and asserts that alter the LLVM C++ ABI are predicated on the
+preprocessor symbol `LLVM_ENABLE_ABI_BREAKING_CHECKS` -- LLVM
+libraries built with `LLVM_ENABLE_ABI_BREAKING_CHECKS` are not ABI
+compatible LLVM libraries built without it defined.  By default,
+turning on assertions also turns on `LLVM_ENABLE_ABI_BREAKING_CHECKS`
+so a default +Asserts build is not ABI compatible with a
+default -Asserts build.  Clients that want ABI compatibility
+between +Asserts and -Asserts builds should use the CMake or autoconf
+build systems to set `LLVM_ENABLE_ABI_BREAKING_CHECKS` independently
+of `LLVM_ENABLE_ASSERTIONS`.
+
+.. _coreclasses:
+
+The Core LLVM Class Hierarchy Reference
+=======================================
+
+``#include "llvm/IR/Type.h"``
+
+header source: `Type.h <http://llvm.org/doxygen/Type_8h-source.html>`_
+
+doxygen info: `Type Clases <http://llvm.org/doxygen/classllvm_1_1Type.html>`_
+
+The Core LLVM classes are the primary means of representing the program being
+inspected or transformed.  The core LLVM classes are defined in header files in
+the ``include/llvm/IR`` directory, and implemented in the ``lib/IR``
+directory. It's worth noting that, for historical reasons, this library is
+called ``libLLVMCore.so``, not ``libLLVMIR.so`` as you might expect.
+
+.. _Type:
+
+The Type class and Derived Types
+--------------------------------
+
+``Type`` is a superclass of all type classes.  Every ``Value`` has a ``Type``.
+``Type`` cannot be instantiated directly but only through its subclasses.
+Certain primitive types (``VoidType``, ``LabelType``, ``FloatType`` and
+``DoubleType``) have hidden subclasses.  They are hidden because they offer no
+useful functionality beyond what the ``Type`` class offers except to distinguish
+themselves from other subclasses of ``Type``.
+
+All other types are subclasses of ``DerivedType``.  Types can be named, but this
+is not a requirement.  There exists exactly one instance of a given shape at any
+one time.  This allows type equality to be performed with address equality of
+the Type Instance.  That is, given two ``Type*`` values, the types are identical
+if the pointers are identical.
+
+.. _m_Type:
+
+Important Public Methods
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+* ``bool isIntegerTy() const``: Returns true for any integer type.
+
+* ``bool isFloatingPointTy()``: Return true if this is one of the five
+  floating point types.
+
+* ``bool isSized()``: Return true if the type has known size.  Things
+  that don't have a size are abstract types, labels and void.
+
+.. _derivedtypes:
+
+Important Derived Types
+^^^^^^^^^^^^^^^^^^^^^^^
+
+``IntegerType``
+  Subclass of DerivedType that represents integer types of any bit width.  Any
+  bit width between ``IntegerType::MIN_INT_BITS`` (1) and
+  ``IntegerType::MAX_INT_BITS`` (~8 million) can be represented.
+
+  * ``static const IntegerType* get(unsigned NumBits)``: get an integer
+    type of a specific bit width.
+
+  * ``unsigned getBitWidth() const``: Get the bit width of an integer type.
+
+``SequentialType``
+  This is subclassed by ArrayType and VectorType.
+
+  * ``const Type * getElementType() const``: Returns the type of each
+    of the elements in the sequential type.
+
+  * ``uint64_t getNumElements() const``: Returns the number of elements
+    in the sequential type.
+
+``ArrayType``
+  This is a subclass of SequentialType and defines the interface for array
+  types.
+
+``PointerType``
+  Subclass of Type for pointer types.
+
+``VectorType``
+  Subclass of SequentialType for vector types.  A vector type is similar to an
+  ArrayType but is distinguished because it is a first class type whereas
+  ArrayType is not.  Vector types are used for vector operations and are usually
+  small vectors of an integer or floating point type.
+
+``StructType``
+  Subclass of DerivedTypes for struct types.
+
+.. _FunctionType:
+
+``FunctionType``
+  Subclass of DerivedTypes for function types.
+
+  * ``bool isVarArg() const``: Returns true if it's a vararg function.
+
+  * ``const Type * getReturnType() const``: Returns the return type of the
+    function.
+
+  * ``const Type * getParamType (unsigned i)``: Returns the type of the ith
+    parameter.
+
+  * ``const unsigned getNumParams() const``: Returns the number of formal
+    parameters.
+
+.. _Module:
+
+The ``Module`` class
+--------------------
+
+``#include "llvm/IR/Module.h"``
+
+header source: `Module.h <http://llvm.org/doxygen/Module_8h-source.html>`_
+
+doxygen info: `Module Class <http://llvm.org/doxygen/classllvm_1_1Module.html>`_
+
+The ``Module`` class represents the top level structure present in LLVM
+programs.  An LLVM module is effectively either a translation unit of the
+original program or a combination of several translation units merged by the
+linker.  The ``Module`` class keeps track of a list of :ref:`Function
+<c_Function>`\ s, a list of GlobalVariable_\ s, and a SymbolTable_.
+Additionally, it contains a few helpful member functions that try to make common
+operations easy.
+
+.. _m_Module:
+
+Important Public Members of the ``Module`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* ``Module::Module(std::string name = "")``
+
+  Constructing a Module_ is easy.  You can optionally provide a name for it
+  (probably based on the name of the translation unit).
+
+* | ``Module::iterator`` - Typedef for function list iterator
+  | ``Module::const_iterator`` - Typedef for const_iterator.
+  | ``begin()``, ``end()``, ``size()``, ``empty()``
+
+  These are forwarding methods that make it easy to access the contents of a
+  ``Module`` object's :ref:`Function <c_Function>` list.
+
+* ``Module::FunctionListType &getFunctionList()``
+
+  Returns the list of :ref:`Function <c_Function>`\ s.  This is necessary to use
+  when you need to update the list or perform a complex action that doesn't have
+  a forwarding method.
+
+----------------
+
+* | ``Module::global_iterator`` - Typedef for global variable list iterator
+  | ``Module::const_global_iterator`` - Typedef for const_iterator.
+  | ``global_begin()``, ``global_end()``, ``global_size()``, ``global_empty()``
+
+  These are forwarding methods that make it easy to access the contents of a
+  ``Module`` object's GlobalVariable_ list.
+
+* ``Module::GlobalListType &getGlobalList()``
+
+  Returns the list of GlobalVariable_\ s.  This is necessary to use when you
+  need to update the list or perform a complex action that doesn't have a
+  forwarding method.
+
+----------------
+
+* ``SymbolTable *getSymbolTable()``
+
+  Return a reference to the SymbolTable_ for this ``Module``.
+
+----------------
+
+* ``Function *getFunction(StringRef Name) const``
+
+  Look up the specified function in the ``Module`` SymbolTable_.  If it does not
+  exist, return ``null``.
+
+* ``Function *getOrInsertFunction(const std::string &Name, const FunctionType
+  *T)``
+
+  Look up the specified function in the ``Module`` SymbolTable_.  If it does not
+  exist, add an external declaration for the function and return it.
+
+* ``std::string getTypeName(const Type *Ty)``
+
+  If there is at least one entry in the SymbolTable_ for the specified Type_,
+  return it.  Otherwise return the empty string.
+
+* ``bool addTypeName(const std::string &Name, const Type *Ty)``
+
+  Insert an entry in the SymbolTable_ mapping ``Name`` to ``Ty``.  If there is
+  already an entry for this name, true is returned and the SymbolTable_ is not
+  modified.
+
+.. _Value:
+
+The ``Value`` class
+-------------------
+
+``#include "llvm/IR/Value.h"``
+
+header source: `Value.h <http://llvm.org/doxygen/Value_8h-source.html>`_
+
+doxygen info: `Value Class <http://llvm.org/doxygen/classllvm_1_1Value.html>`_
+
+The ``Value`` class is the most important class in the LLVM Source base.  It
+represents a typed value that may be used (among other things) as an operand to
+an instruction.  There are many different types of ``Value``\ s, such as
+Constant_\ s, Argument_\ s.  Even Instruction_\ s and :ref:`Function
+<c_Function>`\ s are ``Value``\ s.
+
+A particular ``Value`` may be used many times in the LLVM representation for a
+program.  For example, an incoming argument to a function (represented with an
+instance of the Argument_ class) is "used" by every instruction in the function
+that references the argument.  To keep track of this relationship, the ``Value``
+class keeps a list of all of the ``User``\ s that is using it (the User_ class
+is a base class for all nodes in the LLVM graph that can refer to ``Value``\ s).
+This use list is how LLVM represents def-use information in the program, and is
+accessible through the ``use_*`` methods, shown below.
+
+Because LLVM is a typed representation, every LLVM ``Value`` is typed, and this
+Type_ is available through the ``getType()`` method.  In addition, all LLVM
+values can be named.  The "name" of the ``Value`` is a symbolic string printed
+in the LLVM code:
+
+.. code-block:: llvm
+
+  %foo = add i32 1, 2
+
+.. _nameWarning:
+
+The name of this instruction is "foo". **NOTE** that the name of any value may
+be missing (an empty string), so names should **ONLY** be used for debugging
+(making the source code easier to read, debugging printouts), they should not be
+used to keep track of values or map between them.  For this purpose, use a
+``std::map`` of pointers to the ``Value`` itself instead.
+
+One important aspect of LLVM is that there is no distinction between an SSA
+variable and the operation that produces it.  Because of this, any reference to
+the value produced by an instruction (or the value available as an incoming
+argument, for example) is represented as a direct pointer to the instance of the
+class that represents this value.  Although this may take some getting used to,
+it simplifies the representation and makes it easier to manipulate.
+
+.. _m_Value:
+
+Important Public Members of the ``Value`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* | ``Value::use_iterator`` - Typedef for iterator over the use-list
+  | ``Value::const_use_iterator`` - Typedef for const_iterator over the
+    use-list
+  | ``unsigned use_size()`` - Returns the number of users of the value.
+  | ``bool use_empty()`` - Returns true if there are no users.
+  | ``use_iterator use_begin()`` - Get an iterator to the start of the
+    use-list.
+  | ``use_iterator use_end()`` - Get an iterator to the end of the use-list.
+  | ``User *use_back()`` - Returns the last element in the list.
+
+  These methods are the interface to access the def-use information in LLVM.
+  As with all other iterators in LLVM, the naming conventions follow the
+  conventions defined by the STL_.
+
+* ``Type *getType() const``
+  This method returns the Type of the Value.
+
+* | ``bool hasName() const``
+  | ``std::string getName() const``
+  | ``void setName(const std::string &Name)``
+
+  This family of methods is used to access and assign a name to a ``Value``, be
+  aware of the :ref:`precaution above <nameWarning>`.
+
+* ``void replaceAllUsesWith(Value *V)``
+
+  This method traverses the use list of a ``Value`` changing all User_\ s of the
+  current value to refer to "``V``" instead.  For example, if you detect that an
+  instruction always produces a constant value (for example through constant
+  folding), you can replace all uses of the instruction with the constant like
+  this:
+
+  .. code-block:: c++
+
+    Inst->replaceAllUsesWith(ConstVal);
+
+.. _User:
+
+The ``User`` class
+------------------
+
+``#include "llvm/IR/User.h"``
+
+header source: `User.h <http://llvm.org/doxygen/User_8h-source.html>`_
+
+doxygen info: `User Class <http://llvm.org/doxygen/classllvm_1_1User.html>`_
+
+Superclass: Value_
+
+The ``User`` class is the common base class of all LLVM nodes that may refer to
+``Value``\ s.  It exposes a list of "Operands" that are all of the ``Value``\ s
+that the User is referring to.  The ``User`` class itself is a subclass of
+``Value``.
+
+The operands of a ``User`` point directly to the LLVM ``Value`` that it refers
+to.  Because LLVM uses Static Single Assignment (SSA) form, there can only be
+one definition referred to, allowing this direct connection.  This connection
+provides the use-def information in LLVM.
+
+.. _m_User:
+
+Important Public Members of the ``User`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``User`` class exposes the operand list in two ways: through an index access
+interface and through an iterator based interface.
+
+* | ``Value *getOperand(unsigned i)``
+  | ``unsigned getNumOperands()``
+
+  These two methods expose the operands of the ``User`` in a convenient form for
+  direct access.
+
+* | ``User::op_iterator`` - Typedef for iterator over the operand list
+  | ``op_iterator op_begin()`` - Get an iterator to the start of the operand
+    list.
+  | ``op_iterator op_end()`` - Get an iterator to the end of the operand list.
+
+  Together, these methods make up the iterator based interface to the operands
+  of a ``User``.
+
+
+.. _Instruction:
+
+The ``Instruction`` class
+-------------------------
+
+``#include "llvm/IR/Instruction.h"``
+
+header source: `Instruction.h
+<http://llvm.org/doxygen/Instruction_8h-source.html>`_
+
+doxygen info: `Instruction Class
+<http://llvm.org/doxygen/classllvm_1_1Instruction.html>`_
+
+Superclasses: User_, Value_
+
+The ``Instruction`` class is the common base class for all LLVM instructions.
+It provides only a few methods, but is a very commonly used class.  The primary
+data tracked by the ``Instruction`` class itself is the opcode (instruction
+type) and the parent BasicBlock_ the ``Instruction`` is embedded into.  To
+represent a specific type of instruction, one of many subclasses of
+``Instruction`` are used.
+
+Because the ``Instruction`` class subclasses the User_ class, its operands can
+be accessed in the same way as for other ``User``\ s (with the
+``getOperand()``/``getNumOperands()`` and ``op_begin()``/``op_end()`` methods).
+An important file for the ``Instruction`` class is the ``llvm/Instruction.def``
+file.  This file contains some meta-data about the various different types of
+instructions in LLVM.  It describes the enum values that are used as opcodes
+(for example ``Instruction::Add`` and ``Instruction::ICmp``), as well as the
+concrete sub-classes of ``Instruction`` that implement the instruction (for
+example BinaryOperator_ and CmpInst_).  Unfortunately, the use of macros in this
+file confuses doxygen, so these enum values don't show up correctly in the
+`doxygen output <http://llvm.org/doxygen/classllvm_1_1Instruction.html>`_.
+
+.. _s_Instruction:
+
+Important Subclasses of the ``Instruction`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. _BinaryOperator:
+
+* ``BinaryOperator``
+
+  This subclasses represents all two operand instructions whose operands must be
+  the same type, except for the comparison instructions.
+
+.. _CastInst:
+
+* ``CastInst``
+  This subclass is the parent of the 12 casting instructions.  It provides
+  common operations on cast instructions.
+
+.. _CmpInst:
+
+* ``CmpInst``
+
+  This subclass respresents the two comparison instructions,
+  `ICmpInst <LangRef.html#i_icmp>`_ (integer opreands), and
+  `FCmpInst <LangRef.html#i_fcmp>`_ (floating point operands).
+
+.. _TerminatorInst:
+
+* ``TerminatorInst``
+
+  This subclass is the parent of all terminator instructions (those which can
+  terminate a block).
+
+.. _m_Instruction:
+
+Important Public Members of the ``Instruction`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* ``BasicBlock *getParent()``
+
+  Returns the BasicBlock_ that this
+  ``Instruction`` is embedded into.
+
+* ``bool mayWriteToMemory()``
+
+  Returns true if the instruction writes to memory, i.e. it is a ``call``,
+  ``free``, ``invoke``, or ``store``.
+
+* ``unsigned getOpcode()``
+
+  Returns the opcode for the ``Instruction``.
+
+* ``Instruction *clone() const``
+
+  Returns another instance of the specified instruction, identical in all ways
+  to the original except that the instruction has no parent (i.e. it's not
+  embedded into a BasicBlock_), and it has no name.
+
+.. _Constant:
+
+The ``Constant`` class and subclasses
+-------------------------------------
+
+Constant represents a base class for different types of constants.  It is
+subclassed by ConstantInt, ConstantArray, etc. for representing the various
+types of Constants.  GlobalValue_ is also a subclass, which represents the
+address of a global variable or function.
+
+.. _s_Constant:
+
+Important Subclasses of Constant
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* ConstantInt : This subclass of Constant represents an integer constant of
+  any width.
+
+  * ``const APInt& getValue() const``: Returns the underlying
+    value of this constant, an APInt value.
+
+  * ``int64_t getSExtValue() const``: Converts the underlying APInt value to an
+    int64_t via sign extension.  If the value (not the bit width) of the APInt
+    is too large to fit in an int64_t, an assertion will result.  For this
+    reason, use of this method is discouraged.
+
+  * ``uint64_t getZExtValue() const``: Converts the underlying APInt value
+    to a uint64_t via zero extension.  IF the value (not the bit width) of the
+    APInt is too large to fit in a uint64_t, an assertion will result.  For this
+    reason, use of this method is discouraged.
+
+  * ``static ConstantInt* get(const APInt& Val)``: Returns the ConstantInt
+    object that represents the value provided by ``Val``.  The type is implied
+    as the IntegerType that corresponds to the bit width of ``Val``.
+
+  * ``static ConstantInt* get(const Type *Ty, uint64_t Val)``: Returns the
+    ConstantInt object that represents the value provided by ``Val`` for integer
+    type ``Ty``.
+
+* ConstantFP : This class represents a floating point constant.
+
+  * ``double getValue() const``: Returns the underlying value of this constant.
+
+* ConstantArray : This represents a constant array.
+
+  * ``const std::vector<Use> &getValues() const``: Returns a vector of
+    component constants that makeup this array.
+
+* ConstantStruct : This represents a constant struct.
+
+  * ``const std::vector<Use> &getValues() const``: Returns a vector of
+    component constants that makeup this array.
+
+* GlobalValue : This represents either a global variable or a function.  In
+  either case, the value is a constant fixed address (after linking).
+
+.. _GlobalValue:
+
+The ``GlobalValue`` class
+-------------------------
+
+``#include "llvm/IR/GlobalValue.h"``
+
+header source: `GlobalValue.h
+<http://llvm.org/doxygen/GlobalValue_8h-source.html>`_
+
+doxygen info: `GlobalValue Class
+<http://llvm.org/doxygen/classllvm_1_1GlobalValue.html>`_
+
+Superclasses: Constant_, User_, Value_
+
+Global values ( GlobalVariable_\ s or :ref:`Function <c_Function>`\ s) are the
+only LLVM values that are visible in the bodies of all :ref:`Function
+<c_Function>`\ s.  Because they are visible at global scope, they are also
+subject to linking with other globals defined in different translation units.
+To control the linking process, ``GlobalValue``\ s know their linkage rules.
+Specifically, ``GlobalValue``\ s know whether they have internal or external
+linkage, as defined by the ``LinkageTypes`` enumeration.
+
+If a ``GlobalValue`` has internal linkage (equivalent to being ``static`` in C),
+it is not visible to code outside the current translation unit, and does not
+participate in linking.  If it has external linkage, it is visible to external
+code, and does participate in linking.  In addition to linkage information,
+``GlobalValue``\ s keep track of which Module_ they are currently part of.
+
+Because ``GlobalValue``\ s are memory objects, they are always referred to by
+their **address**.  As such, the Type_ of a global is always a pointer to its
+contents.  It is important to remember this when using the ``GetElementPtrInst``
+instruction because this pointer must be dereferenced first.  For example, if
+you have a ``GlobalVariable`` (a subclass of ``GlobalValue)`` that is an array
+of 24 ints, type ``[24 x i32]``, then the ``GlobalVariable`` is a pointer to
+that array.  Although the address of the first element of this array and the
+value of the ``GlobalVariable`` are the same, they have different types.  The
+``GlobalVariable``'s type is ``[24 x i32]``.  The first element's type is
+``i32.`` Because of this, accessing a global value requires you to dereference
+the pointer with ``GetElementPtrInst`` first, then its elements can be accessed.
+This is explained in the `LLVM Language Reference Manual
+<LangRef.html#globalvars>`_.
+
+.. _m_GlobalValue:
+
+Important Public Members of the ``GlobalValue`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* | ``bool hasInternalLinkage() const``
+  | ``bool hasExternalLinkage() const``
+  | ``void setInternalLinkage(bool HasInternalLinkage)``
+
+  These methods manipulate the linkage characteristics of the ``GlobalValue``.
+
+* ``Module *getParent()``
+
+  This returns the Module_ that the
+  GlobalValue is currently embedded into.
+
+.. _c_Function:
+
+The ``Function`` class
+----------------------
+
+``#include "llvm/IR/Function.h"``
+
+header source: `Function.h <http://llvm.org/doxygen/Function_8h-source.html>`_
+
+doxygen info: `Function Class
+<http://llvm.org/doxygen/classllvm_1_1Function.html>`_
+
+Superclasses: GlobalValue_, Constant_, User_, Value_
+
+The ``Function`` class represents a single procedure in LLVM.  It is actually
+one of the more complex classes in the LLVM hierarchy because it must keep track
+of a large amount of data.  The ``Function`` class keeps track of a list of
+BasicBlock_\ s, a list of formal Argument_\ s, and a SymbolTable_.
+
+The list of BasicBlock_\ s is the most commonly used part of ``Function``
+objects.  The list imposes an implicit ordering of the blocks in the function,
+which indicate how the code will be laid out by the backend.  Additionally, the
+first BasicBlock_ is the implicit entry node for the ``Function``.  It is not
+legal in LLVM to explicitly branch to this initial block.  There are no implicit
+exit nodes, and in fact there may be multiple exit nodes from a single
+``Function``.  If the BasicBlock_ list is empty, this indicates that the
+``Function`` is actually a function declaration: the actual body of the function
+hasn't been linked in yet.
+
+In addition to a list of BasicBlock_\ s, the ``Function`` class also keeps track
+of the list of formal Argument_\ s that the function receives.  This container
+manages the lifetime of the Argument_ nodes, just like the BasicBlock_ list does
+for the BasicBlock_\ s.
+
+The SymbolTable_ is a very rarely used LLVM feature that is only used when you
+have to look up a value by name.  Aside from that, the SymbolTable_ is used
+internally to make sure that there are not conflicts between the names of
+Instruction_\ s, BasicBlock_\ s, or Argument_\ s in the function body.
+
+Note that ``Function`` is a GlobalValue_ and therefore also a Constant_.  The
+value of the function is its address (after linking) which is guaranteed to be
+constant.
+
+.. _m_Function:
+
+Important Public Members of the ``Function``
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* ``Function(const FunctionType *Ty, LinkageTypes Linkage,
+  const std::string &N = "", Module* Parent = 0)``
+
+  Constructor used when you need to create new ``Function``\ s to add the
+  program.  The constructor must specify the type of the function to create and
+  what type of linkage the function should have.  The FunctionType_ argument
+  specifies the formal arguments and return value for the function.  The same
+  FunctionType_ value can be used to create multiple functions.  The ``Parent``
+  argument specifies the Module in which the function is defined.  If this
+  argument is provided, the function will automatically be inserted into that
+  module's list of functions.
+
+* ``bool isDeclaration()``
+
+  Return whether or not the ``Function`` has a body defined.  If the function is
+  "external", it does not have a body, and thus must be resolved by linking with
+  a function defined in a different translation unit.
+
+* | ``Function::iterator`` - Typedef for basic block list iterator
+  | ``Function::const_iterator`` - Typedef for const_iterator.
+  | ``begin()``, ``end()``, ``size()``, ``empty()``
+
+  These are forwarding methods that make it easy to access the contents of a
+  ``Function`` object's BasicBlock_ list.
+
+* ``Function::BasicBlockListType &getBasicBlockList()``
+
+  Returns the list of BasicBlock_\ s.  This is necessary to use when you need to
+  update the list or perform a complex action that doesn't have a forwarding
+  method.
+
+* | ``Function::arg_iterator`` - Typedef for the argument list iterator
+  | ``Function::const_arg_iterator`` - Typedef for const_iterator.
+  | ``arg_begin()``, ``arg_end()``, ``arg_size()``, ``arg_empty()``
+
+  These are forwarding methods that make it easy to access the contents of a
+  ``Function`` object's Argument_ list.
+
+* ``Function::ArgumentListType &getArgumentList()``
+
+  Returns the list of Argument_.  This is necessary to use when you need to
+  update the list or perform a complex action that doesn't have a forwarding
+  method.
+
+* ``BasicBlock &getEntryBlock()``
+
+  Returns the entry ``BasicBlock`` for the function.  Because the entry block
+  for the function is always the first block, this returns the first block of
+  the ``Function``.
+
+* | ``Type *getReturnType()``
+  | ``FunctionType *getFunctionType()``
+
+  This traverses the Type_ of the ``Function`` and returns the return type of
+  the function, or the FunctionType_ of the actual function.
+
+* ``SymbolTable *getSymbolTable()``
+
+  Return a pointer to the SymbolTable_ for this ``Function``.
+
+.. _GlobalVariable:
+
+The ``GlobalVariable`` class
+----------------------------
+
+``#include "llvm/IR/GlobalVariable.h"``
+
+header source: `GlobalVariable.h
+<http://llvm.org/doxygen/GlobalVariable_8h-source.html>`_
+
+doxygen info: `GlobalVariable Class
+<http://llvm.org/doxygen/classllvm_1_1GlobalVariable.html>`_
+
+Superclasses: GlobalValue_, Constant_, User_, Value_
+
+Global variables are represented with the (surprise surprise) ``GlobalVariable``
+class.  Like functions, ``GlobalVariable``\ s are also subclasses of
+GlobalValue_, and as such are always referenced by their address (global values
+must live in memory, so their "name" refers to their constant address).  See
+GlobalValue_ for more on this.  Global variables may have an initial value
+(which must be a Constant_), and if they have an initializer, they may be marked
+as "constant" themselves (indicating that their contents never change at
+runtime).
+
+.. _m_GlobalVariable:
+
+Important Public Members of the ``GlobalVariable`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* ``GlobalVariable(const Type *Ty, bool isConstant, LinkageTypes &Linkage,
+  Constant *Initializer = 0, const std::string &Name = "", Module* Parent = 0)``
+
+  Create a new global variable of the specified type.  If ``isConstant`` is true
+  then the global variable will be marked as unchanging for the program.  The
+  Linkage parameter specifies the type of linkage (internal, external, weak,
+  linkonce, appending) for the variable.  If the linkage is InternalLinkage,
+  WeakAnyLinkage, WeakODRLinkage, LinkOnceAnyLinkage or LinkOnceODRLinkage, then
+  the resultant global variable will have internal linkage.  AppendingLinkage
+  concatenates together all instances (in different translation units) of the
+  variable into a single variable but is only applicable to arrays.  See the
+  `LLVM Language Reference <LangRef.html#modulestructure>`_ for further details
+  on linkage types.  Optionally an initializer, a name, and the module to put
+  the variable into may be specified for the global variable as well.
+
+* ``bool isConstant() const``
+
+  Returns true if this is a global variable that is known not to be modified at
+  runtime.
+
+* ``bool hasInitializer()``
+
+  Returns true if this ``GlobalVariable`` has an intializer.
+
+* ``Constant *getInitializer()``
+
+  Returns the initial value for a ``GlobalVariable``.  It is not legal to call
+  this method if there is no initializer.
+
+.. _BasicBlock:
+
+The ``BasicBlock`` class
+------------------------
+
+``#include "llvm/IR/BasicBlock.h"``
+
+header source: `BasicBlock.h
+<http://llvm.org/doxygen/BasicBlock_8h-source.html>`_
+
+doxygen info: `BasicBlock Class
+<http://llvm.org/doxygen/classllvm_1_1BasicBlock.html>`_
+
+Superclass: Value_
+
+This class represents a single entry single exit section of the code, commonly
+known as a basic block by the compiler community.  The ``BasicBlock`` class
+maintains a list of Instruction_\ s, which form the body of the block.  Matching
+the language definition, the last element of this list of instructions is always
+a terminator instruction (a subclass of the TerminatorInst_ class).
+
+In addition to tracking the list of instructions that make up the block, the
+``BasicBlock`` class also keeps track of the :ref:`Function <c_Function>` that
+it is embedded into.
+
+Note that ``BasicBlock``\ s themselves are Value_\ s, because they are
+referenced by instructions like branches and can go in the switch tables.
+``BasicBlock``\ s have type ``label``.
+
+.. _m_BasicBlock:
+
+Important Public Members of the ``BasicBlock`` class
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+* ``BasicBlock(const std::string &Name = "", Function *Parent = 0)``
+
+  The ``BasicBlock`` constructor is used to create new basic blocks for
+  insertion into a function.  The constructor optionally takes a name for the
+  new block, and a :ref:`Function <c_Function>` to insert it into.  If the
+  ``Parent`` parameter is specified, the new ``BasicBlock`` is automatically
+  inserted at the end of the specified :ref:`Function <c_Function>`, if not
+  specified, the BasicBlock must be manually inserted into the :ref:`Function
+  <c_Function>`.
+
+* | ``BasicBlock::iterator`` - Typedef for instruction list iterator
+  | ``BasicBlock::const_iterator`` - Typedef for const_iterator.
+  | ``begin()``, ``end()``, ``front()``, ``back()``,
+    ``size()``, ``empty()``
+    STL-style functions for accessing the instruction list.
+
+  These methods and typedefs are forwarding functions that have the same
+  semantics as the standard library methods of the same names.  These methods
+  expose the underlying instruction list of a basic block in a way that is easy
+  to manipulate.  To get the full complement of container operations (including
+  operations to update the list), you must use the ``getInstList()`` method.
+
+* ``BasicBlock::InstListType &getInstList()``
+
+  This method is used to get access to the underlying container that actually
+  holds the Instructions.  This method must be used when there isn't a
+  forwarding function in the ``BasicBlock`` class for the operation that you
+  would like to perform.  Because there are no forwarding functions for
+  "updating" operations, you need to use this if you want to update the contents
+  of a ``BasicBlock``.
+
+* ``Function *getParent()``
+
+  Returns a pointer to :ref:`Function <c_Function>` the block is embedded into,
+  or a null pointer if it is homeless.
+
+* ``TerminatorInst *getTerminator()``
+
+  Returns a pointer to the terminator instruction that appears at the end of the
+  ``BasicBlock``.  If there is no terminator instruction, or if the last
+  instruction in the block is not a terminator, then a null pointer is returned.
+
+.. _Argument:
+
+The ``Argument`` class
+----------------------
+
+This subclass of Value defines the interface for incoming formal arguments to a
+function.  A Function maintains a list of its formal arguments.  An argument has
+a pointer to the parent Function.
+
+

Added: www-releases/trunk/4.0.0/docs/_sources/Projects.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/Projects.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/Projects.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/Projects.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,257 @@
+========================
+Creating an LLVM Project
+========================
+
+.. contents::
+   :local:
+
+Overview
+========
+
+The LLVM build system is designed to facilitate the building of third party
+projects that use LLVM header files, libraries, and tools.  In order to use
+these facilities, a ``Makefile`` from a project must do the following things:
+
+* Set ``make`` variables. There are several variables that a ``Makefile`` needs
+  to set to use the LLVM build system:
+
+  * ``PROJECT_NAME`` - The name by which your project is known.
+  * ``LLVM_SRC_ROOT`` - The root of the LLVM source tree.
+  * ``LLVM_OBJ_ROOT`` - The root of the LLVM object tree.
+  * ``PROJ_SRC_ROOT`` - The root of the project's source tree.
+  * ``PROJ_OBJ_ROOT`` - The root of the project's object tree.
+  * ``PROJ_INSTALL_ROOT`` - The root installation directory.
+  * ``LEVEL`` - The relative path from the current directory to the
+    project's root ``($PROJ_OBJ_ROOT)``.
+
+* Include ``Makefile.config`` from ``$(LLVM_OBJ_ROOT)``.
+
+* Include ``Makefile.rules`` from ``$(LLVM_SRC_ROOT)``.
+
+There are two ways that you can set all of these variables:
+
+* You can write your own ``Makefiles`` which hard-code these values.
+
+* You can use the pre-made LLVM sample project. This sample project includes
+  ``Makefiles``, a configure script that can be used to configure the location
+  of LLVM, and the ability to support multiple object directories from a single
+  source directory.
+
+If you want to devise your own build system, studying other projects and LLVM
+``Makefiles`` will probably provide enough information on how to write your own
+``Makefiles``.
+
+Source Tree Layout
+==================
+
+In order to use the LLVM build system, you will want to organize your source
+code so that it can benefit from the build system's features.  Mainly, you want
+your source tree layout to look similar to the LLVM source tree layout.
+
+Underneath your top level directory, you should have the following directories:
+
+**lib**
+
+    This subdirectory should contain all of your library source code.  For each
+    library that you build, you will have one directory in **lib** that will
+    contain that library's source code.
+
+    Libraries can be object files, archives, or dynamic libraries.  The **lib**
+    directory is just a convenient place for libraries as it places them all in
+    a directory from which they can be linked later.
+
+**include**
+
+    This subdirectory should contain any header files that are global to your
+    project. By global, we mean that they are used by more than one library or
+    executable of your project.
+
+    By placing your header files in **include**, they will be found
+    automatically by the LLVM build system.  For example, if you have a file
+    **include/jazz/note.h**, then your source files can include it simply with
+    **#include "jazz/note.h"**.
+
+**tools**
+
+    This subdirectory should contain all of your source code for executables.
+    For each program that you build, you will have one directory in **tools**
+    that will contain that program's source code.
+
+**test**
+
+    This subdirectory should contain tests that verify that your code works
+    correctly.  Automated tests are especially useful.
+
+    Currently, the LLVM build system provides basic support for tests. The LLVM
+    system provides the following:
+
+* LLVM contains regression tests in ``llvm/test``.  These tests are run by the
+  :doc:`Lit <CommandGuide/lit>` testing tool.  This test procedure uses ``RUN``
+  lines in the actual test case to determine how to run the test.  See the
+  :doc:`TestingGuide` for more details.
+
+* LLVM contains an optional package called ``llvm-test``, which provides
+  benchmarks and programs that are known to compile with the Clang front
+  end. You can use these programs to test your code, gather statistical
+  information, and compare it to the current LLVM performance statistics.
+  
+  Currently, there is no way to hook your tests directly into the ``llvm/test``
+  testing harness. You will simply need to find a way to use the source
+  provided within that directory on your own.
+
+Typically, you will want to build your **lib** directory first followed by your
+**tools** directory.
+
+Writing LLVM Style Makefiles
+============================
+
+The LLVM build system provides a convenient way to build libraries and
+executables.  Most of your project Makefiles will only need to define a few
+variables.  Below is a list of the variables one can set and what they can
+do:
+
+Required Variables
+------------------
+
+``LEVEL``
+
+    This variable is the relative path from this ``Makefile`` to the top
+    directory of your project's source code.  For example, if your source code
+    is in ``/tmp/src``, then the ``Makefile`` in ``/tmp/src/jump/high``
+    would set ``LEVEL`` to ``"../.."``.
+
+Variables for Building Subdirectories
+-------------------------------------
+
+``DIRS``
+
+    This is a space separated list of subdirectories that should be built.  They
+    will be built, one at a time, in the order specified.
+
+``PARALLEL_DIRS``
+
+    This is a list of directories that can be built in parallel. These will be
+    built after the directories in DIRS have been built.
+
+``OPTIONAL_DIRS``
+
+    This is a list of directories that can be built if they exist, but will not
+    cause an error if they do not exist.  They are built serially in the order
+    in which they are listed.
+
+Variables for Building Libraries
+--------------------------------
+
+``LIBRARYNAME``
+
+    This variable contains the base name of the library that will be built.  For
+    example, to build a library named ``libsample.a``, ``LIBRARYNAME`` should
+    be set to ``sample``.
+
+``BUILD_ARCHIVE``
+
+    By default, a library is a ``.o`` file that is linked directly into a
+    program.  To build an archive (also known as a static library), set the
+    ``BUILD_ARCHIVE`` variable.
+
+``SHARED_LIBRARY``
+
+    If ``SHARED_LIBRARY`` is defined in your Makefile, a shared (or dynamic)
+    library will be built.
+
+Variables for Building Programs
+-------------------------------
+
+``TOOLNAME``
+
+    This variable contains the name of the program that will be built.  For
+    example, to build an executable named ``sample``, ``TOOLNAME`` should be set
+    to ``sample``.
+
+``USEDLIBS``
+
+    This variable holds a space separated list of libraries that should be
+    linked into the program.  These libraries must be libraries that come from
+    your **lib** directory.  The libraries must be specified without their
+    ``lib`` prefix.  For example, to link ``libsample.a``, you would set
+    ``USEDLIBS`` to ``sample.a``.
+
+    Note that this works only for statically linked libraries.
+
+``LLVMLIBS``
+
+    This variable holds a space separated list of libraries that should be
+    linked into the program.  These libraries must be LLVM libraries.  The
+    libraries must be specified without their ``lib`` prefix.  For example, to
+    link with a driver that performs an IR transformation you might set
+    ``LLVMLIBS`` to this minimal set of libraries ``LLVMSupport.a LLVMCore.a
+    LLVMBitReader.a LLVMAsmParser.a LLVMAnalysis.a LLVMTransformUtils.a
+    LLVMScalarOpts.a LLVMTarget.a``.
+
+    Note that this works only for statically linked libraries. LLVM is split
+    into a large number of static libraries, and the list of libraries you
+    require may be much longer than the list above. To see a full list of
+    libraries use: ``llvm-config --libs all``.  Using ``LINK_COMPONENTS`` as
+    described below, obviates the need to set ``LLVMLIBS``.
+
+``LINK_COMPONENTS``
+
+    This variable holds a space separated list of components that the LLVM
+    ``Makefiles`` pass to the ``llvm-config`` tool to generate a link line for
+    the program. For example, to link with all LLVM libraries use
+    ``LINK_COMPONENTS = all``.
+
+``LIBS``
+
+    To link dynamic libraries, add ``-l<library base name>`` to the ``LIBS``
+    variable.  The LLVM build system will look in the same places for dynamic
+    libraries as it does for static libraries.
+
+    For example, to link ``libsample.so``, you would have the following line in
+    your ``Makefile``:
+
+        .. code-block:: makefile
+
+          LIBS += -lsample
+
+Note that ``LIBS`` must occur in the Makefile after the inclusion of
+``Makefile.common``.
+
+Miscellaneous Variables
+-----------------------
+
+``CFLAGS`` & ``CPPFLAGS``
+
+    This variable can be used to add options to the C and C++ compiler,
+    respectively.  It is typically used to add options that tell the compiler
+    the location of additional directories to search for header files.
+
+    It is highly suggested that you append to ``CFLAGS`` and ``CPPFLAGS`` as
+    opposed to overwriting them.  The master ``Makefiles`` may already have
+    useful options in them that you may not want to overwrite.
+
+Placement of Object Code
+========================
+
+The final location of built libraries and executables will depend upon whether
+you do a ``Debug``, ``Release``, or ``Profile`` build.
+
+Libraries
+
+    All libraries (static and dynamic) will be stored in
+    ``PROJ_OBJ_ROOT/<type>/lib``, where *type* is ``Debug``, ``Release``, or
+    ``Profile`` for a debug, optimized, or profiled build, respectively.
+
+Executables
+
+    All executables will be stored in ``PROJ_OBJ_ROOT/<type>/bin``, where *type*
+    is ``Debug``, ``Release``, or ``Profile`` for a debug, optimized, or
+    profiled build, respectively.
+
+Further Help
+============
+
+If you have any questions or need any help creating an LLVM project, the LLVM
+team would be more than happy to help.  You can always post your questions to
+the `LLVM Developers Mailing List
+<http://lists.llvm.org/pipermail/llvm-dev/>`_.

Added: www-releases/trunk/4.0.0/docs/_sources/Proposals/GitHubMove.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/Proposals/GitHubMove.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/Proposals/GitHubMove.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/Proposals/GitHubMove.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,868 @@
+==============================
+Moving LLVM Projects to GitHub
+==============================
+
+.. contents:: Table of Contents
+  :depth: 4
+  :local:
+
+Introduction
+============
+
+This is a proposal to move our current revision control system from our own
+hosted Subversion to GitHub. Below are the financial and technical arguments as
+to why we are proposing such a move and how people (and validation
+infrastructure) will continue to work with a Git-based LLVM.
+
+There will be a survey pointing at this document which we'll use to gauge the
+community's reaction and, if we collectively decide to move, the time-frame. Be
+sure to make your view count.
+
+Additionally, we will discuss this during a BoF at the next US LLVM Developer
+meeting (http://llvm.org/devmtg/2016-11/).
+
+What This Proposal is *Not* About
+=================================
+
+Changing the development policy.
+
+This proposal relates only to moving the hosting of our source-code repository
+from SVN hosted on our own servers to Git hosted on GitHub. We are not proposing
+using GitHub's issue tracker, pull-requests, or code-review.
+
+Contributers will continue to earn commit access on demand under the Developer
+Policy, except that that a GitHub account will be required instead of SVN
+username/password-hash.
+
+Why Git, and Why GitHub?
+========================
+
+Why Move At All?
+----------------
+
+This discussion began because we currently host our own Subversion server
+and Git mirror on a voluntary basis. The LLVM Foundation sponsors the server and
+provides limited support, but there is only so much it can do.
+
+Volunteers are not sysadmins themselves, but compiler engineers that happen
+to know a thing or two about hosting servers. We also don't have 24/7 support,
+and we sometimes wake up to see that continuous integration is broken because
+the SVN server is either down or unresponsive.
+
+We should take advantage of one of the services out there (GitHub, GitLab,
+and BitBucket, among others) that offer better service (24/7 stability, disk
+space, Git server, code browsing, forking facilities, etc) for free.
+
+Why Git?
+--------
+
+Many new coders nowadays start with Git, and a lot of people have never used
+SVN, CVS, or anything else. Websites like GitHub have changed the landscape
+of open source contributions, reducing the cost of first contribution and
+fostering collaboration.
+
+Git is also the version control many LLVM developers use. Despite the
+sources being stored in a SVN server, these developers are already using Git
+through the Git-SVN integration.
+
+Git allows you to:
+
+* Commit, squash, merge, and fork locally without touching the remote server.
+* Maintain local branches, enabling multiple threads of development.
+* Collaborate on these branches (e.g. through your own fork of llvm on GitHub).
+* Inspect the repository history (blame, log, bisect) without Internet access.
+* Maintain remote forks and branches on Git hosting services and
+  integrate back to the main repository.
+
+In addition, because Git seems to be replacing many OSS projects' version
+control systems, there are many tools that are built over Git.
+Future tooling may support Git first (if not only).
+
+Why GitHub?
+-----------
+
+GitHub, like GitLab and BitBucket, provides free code hosting for open source
+projects. Any of these could replace the code-hosting infrastructure that we
+have today.
+
+These services also have a dedicated team to monitor, migrate, improve and
+distribute the contents of the repositories depending on region and load.
+
+GitHub has one important advantage over GitLab and
+BitBucket: it offers read-write **SVN** access to the repository
+(https://github.com/blog/626-announcing-svn-support).
+This would enable people to continue working post-migration as though our code
+were still canonically in an SVN repository.
+
+In addition, there are already multiple LLVM mirrors on GitHub, indicating that
+part of our community has already settled there.
+
+On Managing Revision Numbers with Git
+-------------------------------------
+
+The current SVN repository hosts all the LLVM sub-projects alongside each other.
+A single revision number (e.g. r123456) thus identifies a consistent version of
+all LLVM sub-projects.
+
+Git does not use sequential integer revision number but instead uses a hash to
+identify each commit. (Linus mentioned that the lack of such revision number
+is "the only real design mistake" in Git [TorvaldRevNum]_.)
+
+The loss of a sequential integer revision number has been a sticking point in
+past discussions about Git:
+
+- "The 'branch' I most care about is mainline, and losing the ability to say
+  'fixed in r1234' (with some sort of monotonically increasing number) would
+  be a tragic loss." [LattnerRevNum]_
+- "I like those results sorted by time and the chronology should be obvious, but
+  timestamps are incredibly cumbersome and make it difficult to verify that a
+  given checkout matches a given set of results." [TrickRevNum]_
+- "There is still the major regression with unreadable version numbers.
+  Given the amount of Bugzilla traffic with 'Fixed in...', that's a
+  non-trivial issue." [JSonnRevNum]_
+- "Sequential IDs are important for LNT and llvmlab bisection tool." [MatthewsRevNum]_.
+
+However, Git can emulate this increasing revision number:
+``git rev-list --count <commit-hash>``. This identifier is unique only
+within a single branch, but this means the tuple `(num, branch-name)` uniquely
+identifies a commit.
+
+We can thus use this revision number to ensure that e.g. `clang -v` reports a
+user-friendly revision number (e.g. `master-12345` or `4.0-5321`), addressing
+the objections raised above with respect to this aspect of Git.
+
+What About Branches and Merges?
+-------------------------------
+
+In contrast to SVN, Git makes branching easy. Git's commit history is
+represented as a DAG, a departure from SVN's linear history. However, we propose
+to mandate making merge commits illegal in our canonical Git repository.
+
+Unfortunately, GitHub does not support server side hooks to enforce such a
+policy.  We must rely on the community to avoid pushing merge commits.
+
+GitHub offers a feature called `Status Checks`: a branch protected by
+`status checks` requires commits to be whitelisted before the push can happen.
+We could supply a pre-push hook on the client side that would run and check the
+history, before whitelisting the commit being pushed [statuschecks]_.
+However this solution would be somewhat fragile (how do you update a script
+installed on every developer machine?) and prevents SVN access to the
+repository.
+
+What About Commit Emails?
+-------------------------
+
+We will need a new bot to send emails for each commit. This proposal leaves the
+email format unchanged besides the commit URL.
+
+Straw Man Migration Plan
+========================
+
+Step #1 : Before The Move
+-------------------------
+
+1. Update docs to mention the move, so people are aware of what is going on.
+2. Set up a read-only version of the GitHub project, mirroring our current SVN
+   repository.
+3. Add the required bots to implement the commit emails, as well as the
+   umbrella repository update (if the multirepo is selected) or the read-only
+   Git views for the sub-projects (if the monorepo is selected).
+
+Step #2 : Git Move
+------------------
+
+4. Update the buildbots to pick up updates and commits from the GitHub
+   repository. Not all bots have to migrate at this point, but it'll help
+   provide infrastructure testing.
+5. Update Phabricator to pick up commits from the GitHub repository.
+6. LNT and llvmlab have to be updated: they rely on unique monotonically
+   increasing integer across branch [MatthewsRevNum]_.
+7. Instruct downstream integrators to pick up commits from the GitHub
+   repository.
+8. Review and prepare an update for the LLVM documentation.
+
+Until this point nothing has changed for developers, it will just
+boil down to a lot of work for buildbot and other infrastructure
+owners.
+
+The migration will pause here until all dependencies have cleared, and all
+problems have been solved.
+
+Step #3: Write Access Move
+--------------------------
+
+9. Collect developers' GitHub account information, and add them to the project.
+10. Switch the SVN repository to read-only and allow pushes to the GitHub repository.
+11. Update the documentation.
+12. Mirror Git to SVN.
+
+Step #4 : Post Move
+-------------------
+
+13. Archive the SVN repository.
+14. Update links on the LLVM website pointing to viewvc/klaus/phab etc. to
+    point to GitHub instead.
+
+One or Multiple Repositories?
+=============================
+
+There are two major variants for how to structure our Git repository: The
+"multirepo" and the "monorepo".
+
+Multirepo Variant
+-----------------
+
+This variant recommends moving each LLVM sub-project to a separate Git
+repository. This mimics the existing official read-only Git repositories
+(e.g., http://llvm.org/git/compiler-rt.git), and creates new canonical
+repositories for each sub-project.
+
+This will allow the individual sub-projects to remain distinct: a
+developer interested only in compiler-rt can checkout only this repository,
+build it, and work in isolation of the other sub-projects.
+
+A key need is to be able to check out multiple projects (i.e. lldb+clang+llvm or
+clang+llvm+libcxx for example) at a specific revision.
+
+A tuple of revisions (one entry per repository) accurately describes the state
+across the sub-projects.
+For example, a given version of clang would be
+*<LLVM-12345, clang-5432, libcxx-123, etc.>*.
+
+Umbrella Repository
+^^^^^^^^^^^^^^^^^^^
+
+To make this more convenient, a separate *umbrella* repository will be
+provided. This repository will be used for the sole purpose of understanding
+the sequence in which commits were pushed to the different repositories and to
+provide a single revision number.
+
+This umbrella repository will be read-only and continuously updated
+to record the above tuple. The proposed form to record this is to use Git
+[submodules]_, possibly along with a set of scripts to help check out a
+specific revision of the LLVM distribution.
+
+A regular LLVM developer does not need to interact with the umbrella repository
+-- the individual repositories can be checked out independently -- but you would
+need to use the umbrella repository to bisect multiple sub-projects at the same
+time, or to check-out old revisions of LLVM with another sub-project at a
+consistent state.
+
+This umbrella repository will be updated automatically by a bot (running on
+notice from a webhook on every push, and periodically) on a per commit basis: a
+single commit in the umbrella repository would match a single commit in a
+sub-project.
+
+Living Downstream
+^^^^^^^^^^^^^^^^^
+
+Downstream SVN users can use the read/write SVN bridges with the following
+caveats:
+
+ * Be prepared for a one-time change to the upstream revision numbers.
+ * The upstream sub-project revision numbers will no longer be in sync.
+
+Downstream Git users can continue without any major changes, with the minor
+change of upstreaming using `git push` instead of `git svn dcommit`.
+
+Git users also have the option of adopting an umbrella repository downstream.
+The tooling for the upstream umbrella can easily be reused for downstream needs,
+incorporating extra sub-projects and branching in parallel with sub-project
+branches.
+
+Multirepo Preview
+^^^^^^^^^^^^^^^^^
+
+As a preview (disclaimer: this rough prototype, not polished and not
+representative of the final solution), you can look at the following:
+
+  * Repository: https://github.com/llvm-beanz/llvm-submodules
+  * Update bot: http://beanz-bot.com:8180/jenkins/job/submodule-update/
+
+Concerns
+^^^^^^^^
+
+ * Because GitHub does not allow server-side hooks, and because there is no
+   "push timestamp" in Git, the umbrella repository sequence isn't totally
+   exact: commits from different repositories pushed around the same time can
+   appear in different orders. However, we don't expect it to be the common case
+   or to cause serious issues in practice.
+ * You can't have a single cross-projects commit that would update both LLVM and
+   other sub-projects (something that can be achieved now). It would be possible
+   to establish a protocol whereby users add a special token to their commit
+   messages that causes the umbrella repo's updater bot to group all of them
+   into a single revision.
+ * Another option is to group commits that were pushed closely enough together
+   in the umbrella repository. This has the advantage of allowing cross-project
+   commits, and is less sensitive to mis-ordering commits. However, this has the
+   potential to group unrelated commits together, especially if the bot goes
+   down and needs to catch up.
+ * This variant relies on heavier tooling. But the current prototype shows that
+   it is not out-of-reach.
+ * Submodules don't have a good reputation / are complicating the command line.
+   However, in the proposed setup, a regular developer will seldom interact with
+   submodules directly, and certainly never update them.
+ * Refactoring across projects is not friendly: taking some functions from clang
+   to make it part of a utility in libSupport wouldn't carry the history of the
+   code in the llvm repo, preventing recursively applying `git blame` for
+   instance. However, this is not very different than how most people are
+   Interacting with the repository today, by splitting such change in multiple
+   commits.
+
+Workflows
+^^^^^^^^^
+
+ * :ref:`Checkout/Clone a Single Project, without Commit Access <workflow-checkout-commit>`.
+ * :ref:`Checkout/Clone a Single Project, with Commit Access <workflow-multicheckout-nocommit>`.
+ * :ref:`Checkout/Clone Multiple Projects, with Commit Access <workflow-multicheckout-multicommit>`.
+ * :ref:`Commit an API Change in LLVM and Update the Sub-projects <workflow-cross-repo-commit>`.
+ * :ref:`Branching/Stashing/Updating for Local Development or Experiments <workflow-multi-branching>`.
+ * :ref:`Bisecting <workflow-multi-bisecting>`.
+
+Monorepo Variant
+----------------
+
+This variant recommends moving all LLVM sub-projects to a single Git repository,
+similar to https://github.com/llvm-project/llvm-project.
+This would mimic an export of the current SVN repository, with each sub-project
+having its own top-level directory.
+Not all sub-projects are used for building toolchains. In practice, www/
+and test-suite/ will probably stay out of the monorepo.
+
+Putting all sub-projects in a single checkout makes cross-project refactoring
+naturally simple:
+
+ * New sub-projects can be trivially split out for better reuse and/or layering
+   (e.g., to allow libSupport and/or LIT to be used by runtimes without adding a
+   dependency on LLVM).
+ * Changing an API in LLVM and upgrading the sub-projects will always be done in
+   a single commit, designing away a common source of temporary build breakage.
+ * Moving code across sub-project (during refactoring for instance) in a single
+   commit enables accurate `git blame` when tracking code change history.
+ * Tooling based on `git grep` works natively across sub-projects, allowing to
+   easier find refactoring opportunities across projects (for example reusing a
+   datastructure initially in LLDB by moving it into libSupport).
+ * Having all the sources present encourages maintaining the other sub-projects
+   when changing API.
+
+Finally, the monorepo maintains the property of the existing SVN repository that
+the sub-projects move synchronously, and a single revision number (or commit
+hash) identifies the state of the development across all projects.
+
+.. _build_single_project:
+
+Building a single sub-project
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Nobody will be forced to build unnecessary projects.  The exact structure
+is TBD, but making it trivial to configure builds for a single sub-project
+(or a subset of sub-projects) is a hard requirement.
+
+As an example, it could look like the following::
+
+  mkdir build && cd build
+  # Configure only LLVM (default)
+  cmake path/to/monorepo
+  # Configure LLVM and lld
+  cmake path/to/monorepo -DLLVM_ENABLE_PROJECTS=lld
+  # Configure LLVM and clang
+  cmake path/to/monorepo -DLLVM_ENABLE_PROJECTS=clang
+
+.. _git-svn-mirror:
+
+Read/write sub-project mirrors
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+With the Monorepo, the existing single-subproject mirrors (e.g.
+http://llvm.org/git/compiler-rt.git) with git-svn read-write access would
+continue to be maintained: developers would continue to be able to use the
+existing single-subproject git repositories as they do today, with *no changes
+to workflow*. Everything (git fetch, git svn dcommit, etc.) could continue to
+work identically to how it works today. The monorepo can be set-up such that the
+SVN revision number matches the SVN revision in the GitHub SVN-bridge.
+
+Living Downstream
+^^^^^^^^^^^^^^^^^
+
+Downstream SVN users can use the read/write SVN bridge. The SVN revision
+number can be preserved in the monorepo, minimizing the impact.
+
+Downstream Git users can continue without any major changes, by using the
+git-svn mirrors on top of the SVN bridge.
+
+Git users can also work upstream with monorepo even if their downstream
+fork has split repositories.  They can apply patches in the appropriate
+subdirectories of the monorepo using, e.g., `git am --directory=...`, or
+plain `diff` and `patch`.
+
+Alternatively, Git users can migrate their own fork to the monorepo.  As a
+demonstration, we've migrated the "CHERI" fork to the monorepo in two ways:
+
+ * Using a script that rewrites history (including merges) so that it looks
+   like the fork always lived in the monorepo [LebarCHERI]_.  The upside of
+   this is when you check out an old revision, you get a copy of all llvm
+   sub-projects at a consistent revision.  (For instance, if it's a clang
+   fork, when you check out an old revision you'll get a consistent version
+   of llvm proper.)  The downside is that this changes the fork's commit
+   hashes.
+
+ * Merging the fork into the monorepo [AminiCHERI]_.  This preserves the
+   fork's commit hashes, but when you check out an old commit you only get
+   the one sub-project.
+
+Monorepo Preview
+^^^^^^^^^^^^^^^^^
+
+As a preview (disclaimer: this rough prototype, not polished and not
+representative of the final solution), you can look at the following:
+
+  * Full Repository: https://github.com/joker-eph/llvm-project
+  * Single sub-project view with *SVN write access* to the full repo:
+    https://github.com/joker-eph/compiler-rt
+
+Concerns
+^^^^^^^^
+
+ * Using the monolithic repository may add overhead for those contributing to a
+   standalone sub-project, particularly on runtimes like libcxx and compiler-rt
+   that don't rely on LLVM; currently, a fresh clone of libcxx is only 15MB (vs.
+   1GB for the monorepo), and the commit rate of LLVM may cause more frequent
+   `git push` collisions when upstreaming. Affected contributors can continue to
+   use the SVN bridge or the single-subproject Git mirrors with git-svn for
+   read-write.
+ * Using the monolithic repository may add overhead for those *integrating* a
+   standalone sub-project, even if they aren't contributing to it, due to the
+   same disk space concern as the point above. The availability of the
+   sub-project Git mirror addesses this, even without SVN access.
+ * Preservation of the existing read/write SVN-based workflows relies on the
+   GitHub SVN bridge, which is an extra dependency.  Maintaining this locks us
+   into GitHub and could restrict future workflow changes.
+
+Workflows
+^^^^^^^^^
+
+ * :ref:`Checkout/Clone a Single Project, without Commit Access <workflow-checkout-commit>`.
+ * :ref:`Checkout/Clone a Single Project, with Commit Access <workflow-monocheckout-nocommit>`.
+ * :ref:`Checkout/Clone Multiple Projects, with Commit Access <workflow-monocheckout-multicommit>`.
+ * :ref:`Commit an API Change in LLVM and Update the Sub-projects <workflow-cross-repo-commit>`.
+ * :ref:`Branching/Stashing/Updating for Local Development or Experiments <workflow-mono-branching>`.
+ * :ref:`Bisecting <workflow-mono-bisecting>`.
+
+Multi/Mono Hybrid Variant
+-------------------------
+
+This variant recommends moving only the LLVM sub-projects that are *rev-locked*
+to LLVM into a monorepo (clang, lld, lldb, ...), following the multirepo
+proposal for the rest.  While neither variant recommends combining sub-projects
+like www/ and test-suite/ (which are completely standalone), this goes further
+and keeps sub-projects like libcxx and compiler-rt in their own distinct
+repositories.
+
+Concerns
+^^^^^^^^
+
+ * This has most disadvantages of multirepo and monorepo, without bringing many
+   of the advantages.
+ * Downstream have to upgrade to the monorepo structure, but only partially. So
+   they will keep the infrastructure to integrate the other separate
+   sub-projects.
+ * All projects that use LIT for testing are effectively rev-locked to LLVM.
+   Furthermore, some runtimes (like compiler-rt) are rev-locked with Clang.
+   It's not clear where to draw the lines.
+
+
+Workflow Before/After
+=====================
+
+This section goes through a few examples of workflows, intended to illustrate
+how end-users or developers would interact with the repository for
+various use-cases.
+
+.. _workflow-checkout-commit:
+
+Checkout/Clone a Single Project, without Commit Access
+------------------------------------------------------
+
+Except the URL, nothing changes. The possibilities today are::
+
+  svn co http://llvm.org/svn/llvm-project/llvm/trunk llvm
+  # or with Git
+  git clone http://llvm.org/git/llvm.git
+
+After the move to GitHub, you would do either::
+
+  git clone https://github.com/llvm-project/llvm.git
+  # or using the GitHub svn native bridge
+  svn co https://github.com/llvm-project/llvm/trunk
+
+The above works for both the monorepo and the multirepo, as we'll maintain the
+existing read-only views of the individual sub-projects.
+
+Checkout/Clone a Single Project, with Commit Access
+---------------------------------------------------
+
+Currently
+^^^^^^^^^
+
+::
+
+  # direct SVN checkout
+  svn co https://user@llvm.org/svn/llvm-project/llvm/trunk llvm
+  # or using the read-only Git view, with git-svn
+  git clone http://llvm.org/git/llvm.git
+  cd llvm
+  git svn init https://llvm.org/svn/llvm-project/llvm/trunk --username=<username>
+  git config svn-remote.svn.fetch :refs/remotes/origin/master
+  git svn rebase -l  # -l avoids fetching ahead of the git mirror.
+
+Commits are performed using `svn commit` or with the sequence `git commit` and
+`git svn dcommit`.
+
+.. _workflow-multicheckout-nocommit:
+
+Multirepo Variant
+^^^^^^^^^^^^^^^^^
+
+With the multirepo variant, nothing changes but the URL, and commits can be
+performed using `svn commit` or `git commit` and `git push`::
+
+  git clone https://github.com/llvm/llvm.git llvm
+  # or using the GitHub svn native bridge
+  svn co https://github.com/llvm/llvm/trunk/ llvm
+
+.. _workflow-monocheckout-nocommit:
+
+Monorepo Variant
+^^^^^^^^^^^^^^^^
+
+With the monorepo variant, there are a few options, depending on your
+constraints. First, you could just clone the full repository::
+
+  git clone https://github.com/llvm/llvm-projects.git llvm
+  # or using the GitHub svn native bridge
+  svn co https://github.com/llvm/llvm-projects/trunk/ llvm
+
+At this point you have every sub-project (llvm, clang, lld, lldb, ...), which
+:ref:`doesn't imply you have to build all of them <build_single_project>`. You
+can still build only compiler-rt for instance. In this way it's not different
+from someone who would check out all the projects with SVN today.
+
+You can commit as normal using `git commit` and `git push` or `svn commit`, and
+read the history for a single project (`git log libcxx` for example).
+
+Secondly, there are a few options to avoid checking out all the sources.
+
+**Using the GitHub SVN bridge**
+
+The GitHub SVN native bridge allows to checkout a subdirectory directly:
+
+  svn co https://github.com/llvm/llvm-projects/trunk/compiler-rt compiler-rt  —username=...
+
+This checks out only compiler-rt and provides commit access using "svn commit",
+in the same way as it would do today.
+
+**Using a Subproject Git Nirror**
+
+You can use *git-svn* and one of the sub-project mirrors::
+
+  # Clone from the single read-only Git repo
+  git clone http://llvm.org/git/llvm.git
+  cd llvm
+  # Configure the SVN remote and initialize the svn metadata
+  $ git svn init https://github.com/joker-eph/llvm-project/trunk/llvm —username=...
+  git config svn-remote.svn.fetch :refs/remotes/origin/master
+  git svn rebase -l
+
+In this case the repository contains only a single sub-project, and commits can
+be made using `git svn dcommit`, again exactly as we do today.
+
+**Using a Sparse Checkouts**
+
+You can hide the other directories using a Git sparse checkout::
+
+  git config core.sparseCheckout true
+  echo /compiler-rt > .git/info/sparse-checkout
+  git read-tree -mu HEAD
+
+The data for all sub-projects is still in your `.git` directory, but in your
+checkout, you only see `compiler-rt`.
+Before you push, you'll need to fetch and rebase (`git pull --rebase`) as
+usual.
+
+Note that when you fetch you'll likely pull in changes to sub-projects you don't
+care about. If you are using spasre checkout, the files from other projects
+won't appear on your disk. The only effect is that your commit hash changes.
+
+You can check whether the changes in the last fetch are relevant to your commit
+by running::
+
+  git log origin/master@{1}..origin/master -- libcxx
+
+This command can be hidden in a script so that `git llvmpush` would perform all
+these steps, fail only if such a dependent change exists, and show immediately
+the change that prevented the push. An immediate repeat of the command would
+(almost) certainly result in a successful push.
+Note that today with SVN or git-svn, this step is not possible since the
+"rebase" implicitly happens while committing (unless a conflict occurs).
+
+Checkout/Clone Multiple Projects, with Commit Access
+----------------------------------------------------
+
+Let's look how to assemble llvm+clang+libcxx at a given revision.
+
+Currently
+^^^^^^^^^
+
+::
+
+  svn co http://llvm.org/svn/llvm-project/llvm/trunk llvm -r $REVISION
+  cd llvm/tools
+  svn co http://llvm.org/svn/llvm-project/clang/trunk clang -r $REVISION
+  cd ../projects
+  svn co http://llvm.org/svn/llvm-project/libcxx/trunk libcxx -r $REVISION
+
+Or using git-svn::
+
+  git clone http://llvm.org/git/llvm.git
+  cd llvm/
+  git svn init https://llvm.org/svn/llvm-project/llvm/trunk --username=<username>
+  git config svn-remote.svn.fetch :refs/remotes/origin/master
+  git svn rebase -l
+  git checkout `git svn find-rev -B r258109`
+  cd tools
+  git clone http://llvm.org/git/clang.git
+  cd clang/
+  git svn init https://llvm.org/svn/llvm-project/clang/trunk --username=<username>
+  git config svn-remote.svn.fetch :refs/remotes/origin/master
+  git svn rebase -l
+  git checkout `git svn find-rev -B r258109`
+  cd ../../projects/
+  git clone http://llvm.org/git/libcxx.git
+  cd libcxx
+  git svn init https://llvm.org/svn/llvm-project/libcxx/trunk --username=<username>
+  git config svn-remote.svn.fetch :refs/remotes/origin/master
+  git svn rebase -l
+  git checkout `git svn find-rev -B r258109`
+
+Note that the list would be longer with more sub-projects.
+
+.. _workflow-multicheckout-multicommit:
+
+Multirepo Variant
+^^^^^^^^^^^^^^^^^
+
+With the multirepo variant, the umbrella repository will be used. This is
+where the mapping from a single revision number to the individual repositories
+revisions is stored.::
+
+  git clone https://github.com/llvm-beanz/llvm-submodules
+  cd llvm-submodules
+  git checkout $REVISION
+  git submodule init
+  git submodule update clang llvm libcxx
+  # the list of sub-project is optional, `git submodule update` would get them all.
+
+At this point the clang, llvm, and libcxx individual repositories are cloned
+and stored alongside each other. There are CMake flags to describe the directory
+structure; alternatively, you can just symlink `clang` to `llvm/tools/clang`,
+etc.
+
+Another option is to checkout repositories based on the commit timestamp::
+
+  git checkout `git rev-list -n 1 --before="2009-07-27 13:37" master`
+
+.. _workflow-monocheckout-multicommit:
+
+Monorepo Variant
+^^^^^^^^^^^^^^^^
+
+The repository contains natively the source for every sub-projects at the right
+revision, which makes this straightforward::
+
+  git clone https://github.com/llvm/llvm-projects.git llvm-projects
+  cd llvm-projects
+  git checkout $REVISION
+
+As before, at this point clang, llvm, and libcxx are stored in directories
+alongside each other.
+
+.. _workflow-cross-repo-commit:
+
+Commit an API Change in LLVM and Update the Sub-projects
+--------------------------------------------------------
+
+Today this is possible, even though not common (at least not documented) for
+subversion users and for git-svn users. For example, few Git users try to update
+LLD or Clang in the same commit as they change an LLVM API.
+
+The multirepo variant does not address this: one would have to commit and push
+separately in every individual repository. It would be possible to establish a
+protocol whereby users add a special token to their commit messages that causes
+the umbrella repo's updater bot to group all of them into a single revision.
+
+The monorepo variant handles this natively.
+
+Branching/Stashing/Updating for Local Development or Experiments
+----------------------------------------------------------------
+
+Currently
+^^^^^^^^^
+
+SVN does not allow this use case, but developers that are currently using
+git-svn can do it. Let's look in practice what it means when dealing with
+multiple sub-projects.
+
+To update the repository to tip of trunk::
+
+  git pull
+  cd tools/clang
+  git pull
+  cd ../../projects/libcxx
+  git pull
+
+To create a new branch::
+
+  git checkout -b MyBranch
+  cd tools/clang
+  git checkout -b MyBranch
+  cd ../../projects/libcxx
+  git checkout -b MyBranch
+
+To switch branches::
+
+  git checkout AnotherBranch
+  cd tools/clang
+  git checkout AnotherBranch
+  cd ../../projects/libcxx
+  git checkout AnotherBranch
+
+.. _workflow-multi-branching:
+
+Multirepo Variant
+^^^^^^^^^^^^^^^^^
+
+The multirepo works the same as the current Git workflow: every command needs
+to be applied to each of the individual repositories.
+However, the umbrella repository makes this easy using `git submodule foreach`
+to replicate a command on all the individual repositories (or submodules
+in this case):
+
+To create a new branch::
+
+  git submodule foreach git checkout -b MyBranch
+
+To switch branches::
+
+  git submodule foreach git checkout AnotherBranch
+
+.. _workflow-mono-branching:
+
+Monorepo Variant
+^^^^^^^^^^^^^^^^
+
+Regular Git commands are sufficient, because everything is in a single
+repository:
+
+To update the repository to tip of trunk::
+
+  git pull
+
+To create a new branch::
+
+  git checkout -b MyBranch
+
+To switch branches::
+
+  git checkout AnotherBranch
+
+Bisecting
+---------
+
+Assuming a developer is looking for a bug in clang (or lld, or lldb, ...).
+
+Currently
+^^^^^^^^^
+
+SVN does not have builtin bisection support, but the single revision across
+sub-projects makes it possible to script around.
+
+Using the existing Git read-only view of the repositories, it is possible to use
+the native Git bisection script over the llvm repository, and use some scripting
+to synchronize the clang repository to match the llvm revision.
+
+.. _workflow-multi-bisecting:
+
+Multirepo Variant
+^^^^^^^^^^^^^^^^^
+
+With the multi-repositories variant, the cross-repository synchronization is
+achieved using the umbrella repository. This repository contains only
+submodules for the other sub-projects. The native Git bisection can be used on
+the umbrella repository directly. A subtlety is that the bisect script itself
+needs to make sure the submodules are updated accordingly.
+
+For example, to find which commit introduces a regression where clang-3.9
+crashes but not clang-3.8 passes, one should be able to simply do::
+
+  git bisect start release_39 release_38
+  git bisect run ./bisect_script.sh
+
+With the `bisect_script.sh` script being::
+
+  #!/bin/sh
+  cd $UMBRELLA_DIRECTORY
+  git submodule update llvm clang libcxx #....
+  cd $BUILD_DIR
+
+  ninja clang || exit 125   # an exit code of 125 asks "git bisect"
+                            # to "skip" the current commit
+
+  ./bin/clang some_crash_test.cpp
+
+When the `git bisect run` command returns, the umbrella repository is set to
+the state where the regression is introduced. The commit diff in the umbrella
+indicate which submodule was updated, and the last commit in this sub-projects
+is the one that the bisect found.
+
+.. _workflow-mono-bisecting:
+
+Monorepo Variant
+^^^^^^^^^^^^^^^^
+
+Bisecting on the monorepo is straightforward, and very similar to the above,
+except that the bisection script does not need to include the
+`git submodule update` step.
+
+The same example, finding which commit introduces a regression where clang-3.9
+crashes but not clang-3.8 passes, will look like::
+
+  git bisect start release_39 release_38
+  git bisect run ./bisect_script.sh
+
+With the `bisect_script.sh` script being::
+
+  #!/bin/sh
+  cd $BUILD_DIR
+
+  ninja clang || exit 125   # an exit code of 125 asks "git bisect"
+                            # to "skip" the current commit
+
+  ./bin/clang some_crash_test.cpp
+
+Also, since the monorepo handles commits update across multiple projects, you're
+less like to encounter a build failure where a commit change an API in LLVM and
+another later one "fixes" the build in clang.
+
+
+References
+==========
+
+.. [LattnerRevNum] Chris Lattner, http://lists.llvm.org/pipermail/llvm-dev/2011-July/041739.html
+.. [TrickRevNum] Andrew Trick, http://lists.llvm.org/pipermail/llvm-dev/2011-July/041721.html
+.. [JSonnRevNum] Joerg Sonnenberg, http://lists.llvm.org/pipermail/llvm-dev/2011-July/041688.html
+.. [TorvaldRevNum] Linus Torvald, http://git.661346.n2.nabble.com/Git-commit-generation-numbers-td6584414.html
+.. [MatthewsRevNum] Chris Matthews, http://lists.llvm.org/pipermail/cfe-dev/2016-July/049886.html
+.. [submodules] Git submodules, https://git-scm.com/book/en/v2/Git-Tools-Submodules)
+.. [statuschecks] GitHub status-checks, https://help.github.com/articles/about-required-status-checks/
+.. [LebarCHERI] Port *CHERI* to a single repository rewriting history, http://lists.llvm.org/pipermail/llvm-dev/2016-July/102787.html
+.. [AminiCHERI] Port *CHERI* to a single repository preserving history, http://lists.llvm.org/pipermail/llvm-dev/2016-July/102804.html

Added: www-releases/trunk/4.0.0/docs/_sources/ReleaseNotes.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/ReleaseNotes.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/ReleaseNotes.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/ReleaseNotes.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,349 @@
+========================
+LLVM 4.0.0 Release Notes
+========================
+
+.. contents::
+    :local:
+
+Introduction
+============
+
+This document contains the release notes for the LLVM Compiler Infrastructure,
+release 4.0.0.  Here we describe the status of LLVM, including major improvements
+from the previous release, improvements in various subprojects of LLVM, and
+some of the current users of the code.  All LLVM releases may be downloaded
+from the `LLVM releases web site <http://llvm.org/releases/>`_.
+
+For more information about LLVM, including information about the latest
+release, please check out the `main LLVM web site <http://llvm.org/>`_.  If you
+have questions or comments, the `LLVM Developer's Mailing List
+<http://lists.llvm.org/mailman/listinfo/llvm-dev>`_ is a good place to send
+them.
+
+New Versioning Scheme
+=====================
+Starting with this release, LLVM is using a
+`new versioning scheme <http://blog.llvm.org/2016/12/llvms-new-versioning-scheme.html>`_,
+increasing the major version number with each major release. Stable updates to
+this release will be versioned 4.0.x, and the next major release, six months
+from now, will be version 5.0.0.
+
+Non-comprehensive list of changes in this release
+=================================================
+* The minimum compiler version required for building LLVM has been raised to
+  4.8 for GCC and 2015 for Visual Studio.
+
+* The C API functions ``LLVMAddFunctionAttr``, ``LLVMGetFunctionAttr``,
+  ``LLVMRemoveFunctionAttr``, ``LLVMAddAttribute``, ``LLVMRemoveAttribute``,
+  ``LLVMGetAttribute``, ``LLVMAddInstrAttribute`` and
+  ``LLVMRemoveInstrAttribute`` have been removed.
+
+* The C API enum ``LLVMAttribute`` has been deleted.
+
+* The definition and uses of ``LLVM_ATRIBUTE_UNUSED_RESULT`` in the LLVM source
+  were replaced with ``LLVM_NODISCARD``, which matches the C++17 ``[[nodiscard]]``
+  semantics rather than gcc's ``__attribute__((warn_unused_result))``.
+
+* The Timer related APIs now expect a Name and Description. When upgrading code
+  the previously used names should become descriptions and a short name in the
+  style of a programming language identifier should be added.
+
+* LLVM now handles ``invariant.group`` across different basic blocks, which makes
+  it possible to devirtualize virtual calls inside loops.
+
+* The aggressive dead code elimination phase ("adce") now removes
+  branches which do not effect program behavior. Loops are retained by
+  default since they may be infinite but these can also be removed
+  with LLVM option ``-adce-remove-loops`` when the loop body otherwise has
+  no live operations.
+
+* The llvm-cov tool can now export coverage data as json. Its html output mode
+  has also improved.
+
+Improvements to ThinLTO (-flto=thin)
+------------------------------------
+Integration with profile data (PGO). When available, profile data
+enables more accurate function importing decisions, as well as
+cross-module indirect call promotion.
+
+Significant build-time and binary-size improvements when compiling with
+debug info (-g).
+
+LLVM Coroutines
+---------------
+
+Experimental support for :doc:`Coroutines` was added, which can be enabled
+with ``-enable-coroutines`` in ``opt`` the command tool or using the
+``addCoroutinePassesToExtensionPoints`` API when building the optimization
+pipeline.
+
+For more information on LLVM Coroutines and the LLVM implementation, see
+`2016 LLVM Developers’ Meeting talk on LLVM Coroutines
+<http://llvm.org/devmtg/2016-11/#talk4>`_.
+
+Regcall and Vectorcall Calling Conventions
+--------------------------------------------------
+
+Support was added for ``_regcall`` calling convention.
+Existing ``__vectorcall`` calling convention support was extended to include
+correct handling of HVAs.
+
+The ``__vectorcall`` calling convention was introduced by Microsoft to
+enhance register usage when passing parameters.
+For more information please read `__vectorcall documentation
+<https://msdn.microsoft.com/en-us/library/dn375768.aspx>`_.
+
+The ``__regcall`` calling convention was introduced by Intel to
+optimize parameter transfer on function call.
+This calling convention ensures that as many values as possible are
+passed or returned in registers.
+For more information please read `__regcall documentation
+<https://software.intel.com/en-us/node/693069>`_.
+
+Code Generation Testing
+-----------------------
+
+Passes that work on the machine instruction representation can be tested with
+the .mir serialization format. ``llc`` supports the ``-run-pass``,
+``-stop-after``, ``-stop-before``, ``-start-after``, ``-start-before`` to
+run a single pass of the code generation pipeline, or to stop or start the code
+generation pipeline at a given point.
+
+Additional information can be found in the :doc:`MIRLangRef`. The format is
+used by the tests ending in ``.mir`` in the ``test/CodeGen`` directory.
+
+This feature is available since 2015. It is used more often lately and was not
+mentioned in the release notes yet.
+
+Intrusive list API overhaul
+---------------------------
+
+The intrusive list infrastructure was substantially rewritten over the last
+couple of releases, primarily to excise undefined behaviour.  The biggest
+changes landed in this release.
+
+* ``simple_ilist<T>`` is a lower-level intrusive list that never takes
+  ownership of its nodes.  New intrusive-list clients should consider using it
+  instead of ``ilist<T>``.
+
+  * ``ilist_tag<class>`` allows a single data type to be inserted into two
+    parallel intrusive lists.  A type can inherit twice from ``ilist_node``,
+    first using ``ilist_node<T,ilist_tag<A>>`` (enabling insertion into
+    ``simple_ilist<T,ilist_tag<A>>``) and second using
+    ``ilist_node<T,ilist_tag<B>>`` (enabling insertion into
+    ``simple_ilist<T,ilist_tag<B>>``), where ``A`` and ``B`` are arbitrary
+    types.
+
+  * ``ilist_sentinel_tracking<bool>`` controls whether an iterator knows
+    whether it's pointing at the sentinel (``end()``).  By default, sentinel
+    tracking is on when ABI-breaking checks are enabled, and off otherwise;
+    this is used for an assertion when dereferencing ``end()`` (this assertion
+    triggered often in practice, and many backend bugs were fixed).  Explicitly
+    turning on sentinel tracking also enables ``iterator::isEnd()``.  This is
+    used by ``MachineInstrBundleIterator`` to iterate over bundles.
+
+* ``ilist<T>`` is built on top of ``simple_ilist<T>``, and supports the same
+  configuration options.  As before (and unlike ``simple_ilist<T>``),
+  ``ilist<T>`` takes ownership of its nodes.  However, it no longer supports
+  *allocating* nodes, and is now equivalent to ``iplist<T>``.  ``iplist<T>``
+  will likely be removed in the future.
+
+  * ``ilist<T>`` now always uses ``ilist_traits<T>``.  Instead of passing a
+    custom traits class in via a template parameter, clients that want to
+    customize the traits should specialize ``ilist_traits<T>``.  Clients that
+    want to avoid ownership can specialize ``ilist_alloc_traits<T>`` to inherit
+    from ``ilist_noalloc_traits<T>`` (or to do something funky); clients that
+    need callbacks can specialize ``ilist_callback_traits<T>`` directly.
+
+* The underlying data structure is now a simple recursive linked list.  The
+  sentinel node contains only a "next" (``begin()``) and "prev" (``rbegin()``)
+  pointer and is stored in the same allocation as ``simple_ilist<T>``.
+  Previously, it was malloc-allocated on-demand by default, although the
+  now-defunct ``ilist_sentinel_traits<T>`` was sometimes specialized to avoid
+  this.
+
+* The ``reverse_iterator`` class no longer uses ``std::reverse_iterator``.
+  Instead, it now has a handle to the same node that it dereferences to.
+  Reverse iterators now have the same iterator invalidation semantics as
+  forward iterators.
+
+  * ``iterator`` and ``reverse_iterator`` have explicit conversion constructors
+    that match ``std::reverse_iterator``'s off-by-one semantics, so that
+    reversing the end points of an iterator range results in the same range
+    (albeit in reverse).  I.e., ``reverse_iterator(begin())`` equals
+    ``rend()``.
+
+  * ``iterator::getReverse()`` and ``reverse_iterator::getReverse()`` return an
+    iterator that dereferences to the *same* node.  I.e.,
+    ``begin().getReverse()`` equals ``--rend()``.
+
+  * ``ilist_node<T>::getIterator()`` and
+    ``ilist_node<T>::getReverseIterator()`` return the forward and reverse
+    iterators that dereference to the current node.  I.e.,
+    ``begin()->getIterator()`` equals ``begin()`` and
+    ``rbegin()->getReverseIterator()`` equals ``rbegin()``.
+
+* ``iterator`` now stores an ``ilist_node_base*`` instead of a ``T*``.  The
+  implicit conversions between ``ilist<T>::iterator`` and ``T*`` have been
+  removed.  Clients may use ``N->getIterator()`` (if not ``nullptr``) or
+  ``&*I`` (if not ``end()``); alternatively, clients may refactor to use
+  references for known-good nodes.
+
+Changes to the ARM Targets
+--------------------------
+
+**During this release the AArch64 target has:**
+
+* Gained support for ILP32 relocations.
+* Gained support for XRay.
+* Made even more progress on GlobalISel. There is still some work left before
+  it is production-ready though.
+* Refined the support for Qualcomm's Falkor and Samsung's Exynos CPUs.
+* Learned a few new tricks for lowering multiplications by constants, folding
+  spilled/refilled copies etc.
+
+**During this release the ARM target has:**
+
+* Gained support for ROPI (read-only position independence) and RWPI
+  (read-write position independence), which can be used to remove the need for
+  a dynamic linker.
+* Gained support for execute-only code, which is placed in pages without read
+  permissions.
+* Gained a machine scheduler for Cortex-R52.
+* Gained support for XRay.
+* Gained Thumb1 implementations for several compiler-rt builtins. It also
+  has some support for building the builtins for HF targets.
+* Started using the generic bitreverse intrinsic instead of rbit.
+* Gained very basic support for GlobalISel.
+
+A lot of work has also been done in LLD for ARM, which now supports more
+relocations and TLS.
+
+Note: From the next release (5.0), the "vulcan" target will be renamed to
+"thunderx2t99", including command line options, assembly directives, etc. This
+release (4.0) will be the last one to accept "vulcan" as its name.
+
+Changes to the AVR Target
+-----------------------------
+
+This marks the first release where the AVR backend has been completely merged
+from a fork into LLVM trunk. The backend is still marked experimental, but
+is generally quite usable. All downstream development has halted on
+`GitHub <https://github.com/avr-llvm/llvm>`_, and changes now go directly into
+LLVM trunk.
+
+* Instruction selector and pseudo instruction expansion pass landed
+* `read_register` and `write_register` intrinsics are now supported
+* Support stack stores greater than 63-bytes from the bottom of the stack
+* A number of assertion errors have been fixed
+* Support stores to `undef` locations
+* Very basic support for the target has been added to clang
+* Small optimizations to some 16-bit boolean expressions
+
+Most of the work behind the scenes has been on correctness of generated
+assembly, and also fixing some assertions we would hit on some well-formed
+inputs.
+
+Changes to the MIPS Target
+-----------------------------
+
+**During this release the MIPS target has:**
+
+* IAS is now enabled by default for Debian mips64el.
+* Added support for the two operand form for many instructions.
+* Added the following macros: unaligned load/store, seq, double word load/store for O32.
+* Improved the parsing of complex memory offset expressions.
+* Enabled the integrated assembler by default for Debian mips64el.
+* Added a generic scheduler based on the interAptiv CPU.
+* Added support for thread local relocations.
+* Added recip, rsqrt, evp, dvp, synci instructions in IAS.
+* Optimized the generation of constants from some cases.
+
+**The following issues have been fixed:**
+
+* Thread local debug information is correctly recorded.
+* MSA intrinsics are now range checked.
+* Fixed an issue with MSA and the no-odd-spreg abi.
+* Fixed some corner cases in handling forbidden slots for MIPSR6.
+* Fixed an issue with jumps not being converted to relative branches for assembly.
+* Fixed the handling of local symbols and jal instruction.
+* N32/N64 no longer have their relocation tables sorted as per their ABIs.
+* Fixed a crash when half-precision floating point conversion MSA intrinsics are used.
+* Fixed several crashes involving FastISel.
+* Corrected the corrected definitions for aui/daui/dahi/dati for MIPSR6.
+
+Changes to the X86 Target
+-------------------------
+
+**During this release the X86 target has:**
+
+* Added support AMD Ryzen (znver1) CPUs.
+* Gained support for using VEX encoding on AVX-512 CPUs to reduce code size when possible.
+* Improved AVX-512 codegen.
+
+Changes to the OCaml bindings
+-----------------------------
+
+* The attribute API was completely overhauled, following the changes
+  to the C API.
+
+
+External Open Source Projects Using LLVM 4.0.0
+==============================================
+
+LDC - the LLVM-based D compiler
+-------------------------------
+
+`D <http://dlang.org>`_ is a language with C-like syntax and static typing. It
+pragmatically combines efficiency, control, and modeling power, with safety and
+programmer productivity. D supports powerful concepts like Compile-Time Function
+Execution (CTFE) and Template Meta-Programming, provides an innovative approach
+to concurrency and offers many classical paradigms.
+
+`LDC <http://wiki.dlang.org/LDC>`_ uses the frontend from the reference compiler
+combined with LLVM as backend to produce efficient native code. LDC targets
+x86/x86_64 systems like Linux, OS X, FreeBSD and Windows and also Linux on ARM
+and PowerPC (32/64 bit). Ports to other architectures like AArch64 and MIPS64
+are underway.
+
+Portable Computing Language (pocl)
+----------------------------------
+
+In addition to producing an easily portable open source OpenCL
+implementation, another major goal of `pocl <http://pocl.sourceforge.net/>`_
+is improving performance portability of OpenCL programs with
+compiler optimizations, reducing the need for target-dependent manual
+optimizations. An important part of pocl is a set of LLVM passes used to
+statically parallelize multiple work-items with the kernel compiler, even in
+the presence of work-group barriers. This enables static parallelization of
+the fine-grained static concurrency in the work groups in multiple ways.
+
+TTA-based Co-design Environment (TCE)
+-------------------------------------
+
+`TCE <http://tce.cs.tut.fi/>`_ is a toolset for designing customized
+processors based on the Transport Triggered Architecture (TTA).
+The toolset provides a complete co-design flow from C/C++
+programs down to synthesizable VHDL/Verilog and parallel program binaries.
+Processor customization points include register files, function units,
+supported operations, and the interconnection network.
+
+TCE uses Clang and LLVM for C/C++/OpenCL C language support, target independent
+optimizations and also for parts of code generation. It generates new
+LLVM-based code generators "on the fly" for the designed TTA processors and
+loads them in to the compiler backend as runtime libraries to avoid
+per-target recompilation of larger parts of the compiler chain.
+
+
+Additional Information
+======================
+
+A wide variety of additional information is available on the `LLVM web page
+<http://llvm.org/>`_, in particular in the `documentation
+<http://llvm.org/docs/>`_ section.  The web page also contains versions of the
+API documentation which is up-to-date with the Subversion version of the source
+code.  You can access versions of these documents specific to this release by
+going into the ``llvm/docs/`` directory in the LLVM tree.
+
+If you have any questions or comments about LLVM, please feel free to contact
+us via the `mailing lists <http://llvm.org/docs/#maillist>`_.

Added: www-releases/trunk/4.0.0/docs/_sources/ReleaseProcess.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/ReleaseProcess.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/ReleaseProcess.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/ReleaseProcess.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,210 @@
+=============================
+How To Validate a New Release
+=============================
+
+.. contents::
+   :local:
+   :depth: 1
+
+Introduction
+============
+
+This document contains information about testing the release candidates that will
+ultimately be the next LLVM release. For more information on how to manage the
+actual release, please refer to :doc:`HowToReleaseLLVM`.
+
+Overview of the Release Process
+-------------------------------
+
+Once the release process starts, the Release Manager will ask for volunteers,
+and it'll be the role of each volunteer to:
+
+* Test and benchmark the previous release
+
+* Test and benchmark each release candidate, comparing to the previous release and candidates
+
+* Identify, reduce and report every regression found during tests and benchmarks
+
+* Make sure the critical bugs get fixed and merged to the next release candidate
+
+Not all bugs or regressions are show-stoppers and it's a bit of a grey area what
+should be fixed before the next candidate and what can wait until the next release.
+
+It'll depend on:
+
+* The severity of the bug, how many people it affects and if it's a regression or a
+  known bug. Known bugs are "unsupported features" and some bugs can be disabled if
+  they have been implemented recently.
+
+* The stage in the release. Less critical bugs should be considered to be fixed between
+  RC1 and RC2, but not so much at the end of it.
+
+* If it's a correctness or a performance regression. Performance regression tends to be
+  taken more lightly than correctness.
+
+.. _scripts:
+
+Scripts
+=======
+
+The scripts are in the ``utils/release`` directory.
+
+test-release.sh
+---------------
+
+This script will check-out, configure and compile LLVM+Clang (+ most add-ons, like ``compiler-rt``,
+``libcxx``, ``libomp`` and ``clang-extra-tools``) in three stages, and will test the final stage.
+It'll have installed the final binaries on the Phase3/Releasei(+Asserts) directory, and
+that's the one you should use for the test-suite and other external tests.
+
+To run the script on a specific release candidate run::
+
+   ./test-release.sh \
+        -release 3.3 \
+        -rc 1 \
+        -no-64bit \
+        -test-asserts \
+        -no-compare-files
+
+Each system will require different options. For instance, x86_64 will obviously not need
+``-no-64bit`` while 32-bit systems will, or the script will fail.
+
+The important flags to get right are:
+
+* On the pre-release, you should change ``-rc 1`` to ``-final``. On RC2, change it to ``-rc 2`` and so on.
+
+* On non-release testing, you can use ``-final`` in conjunction with ``-no-checkout``, but you'll have to
+  create the ``final`` directory by hand and link the correct source dir to ``final/llvm.src``.
+
+* For release candidates, you need ``-test-asserts``, or it won't create a "Release+Asserts" directory,
+  which is needed for release testing and benchmarking. This will take twice as long.
+
+* On the final candidate you just need Release builds, and that's the binary directory you'll have to pack.
+
+This script builds three phases of Clang+LLVM twice each (Release and Release+Asserts), so use
+screen or nohup to avoid headaches, since it'll take a long time.
+
+Use the ``--help`` option to see all the options and chose it according to your needs.
+
+
+findRegressions-nightly.py
+--------------------------
+
+TODO
+
+.. _test-suite:
+
+Test Suite
+==========
+
+.. contents::
+   :local:
+
+Follow the `LNT Quick Start Guide <http://llvm.org/docs/lnt/quickstart.html>`__ link on how to set-up the test-suite
+
+The binary location you'll have to use for testing is inside the ``rcN/Phase3/Release+Asserts/llvmCore-REL-RC.install``.
+Link that directory to an easier location and run the test-suite.
+
+An example on the run command line, assuming you created a link from the correct
+install directory to ``~/devel/llvm/install``::
+
+   ./sandbox/bin/python sandbox/bin/lnt runtest \
+       nt \
+       -j4 \
+       --sandbox sandbox \
+       --test-suite ~/devel/llvm/test/test-suite \
+       --cc ~/devel/llvm/install/bin/clang \
+       --cxx ~/devel/llvm/install/bin/clang++
+
+It should have no new regressions, compared to the previous release or release candidate. You don't need to fix
+all the bugs in the test-suite, since they're not necessarily meant to pass on all architectures all the time. This is
+due to the nature of the result checking, which relies on direct comparison, and most of the time, the failures are
+related to bad output checking, rather than bad code generation.
+
+If the errors are in LLVM itself, please report every single regression found as blocker, and all the other bugs
+as important, but not necessarily blocking the release to proceed. They can be set as "known failures" and to be
+fix on a future date.
+
+.. _pre-release-process:
+
+Pre-Release Process
+===================
+
+.. contents::
+   :local:
+
+When the release process is announced on the mailing list, you should prepare
+for the testing, by applying the same testing you'll do on the release candidates,
+on the previous release.
+
+You should:
+
+* Download the previous release sources from http://llvm.org/releases/download.html.
+
+* Run the test-release.sh script on ``final`` mode (change ``-rc 1`` to ``-final``).
+
+* Once all three stages are done, it'll test the final stage.
+
+* Using the ``Phase3/Release+Asserts/llvmCore-MAJ.MIN-final.install`` base, run the test-suite.
+
+If the final phase's ``make check-all`` failed, it's a good idea to also test the
+intermediate stages by going on the obj directory and running ``make check-all`` to find
+if there's at least one stage that passes (helps when reducing the error for bug report
+purposes).
+
+.. _release-process:
+
+Release Process
+===============
+
+.. contents::
+   :local:
+
+When the Release Manager sends you the release candidate, download all sources,
+unzip on the same directory (there will be sym-links from the appropriate places
+to them), and run the release test as above.
+
+You should:
+
+* Download the current candidate sources from where the release manager points you
+  (ex. http://llvm.org/pre-releases/3.3/rc1/).
+
+* Repeat the steps above with ``-rc 1``, ``-rc 2`` etc modes and run the test-suite
+  the same way.
+
+* Compare the results, report all errors on Bugzilla and publish the binary blob
+  where the release manager can grab it.
+
+Once the release manages announces that the latest candidate is the good one, you
+have to pack the ``Release`` (no Asserts) install directory on ``Phase3`` and that
+will be the official binary.
+
+* Rename (or link) ``clang+llvm-REL-ARCH-ENV`` to the .install directory
+
+* Tar that into the same name with ``.tar.gz`` extensioan from outside the directory
+
+* Make it available for the release manager to download
+
+.. _bug-reporting:
+
+Bug Reporting Process
+=====================
+
+.. contents::
+   :local:
+
+If you found regressions or failures when comparing a release candidate with the
+previous release, follow the rules below:
+
+* Critical bugs on compilation should be fixed as soon as possible, possibly before
+  releasing the binary blobs.
+
+* Check-all tests should be fixed before the next release candidate, but can wait
+  until the test-suite run is finished.
+
+* Bugs in the test suite or unimportant check-all tests can be fixed in between
+  release candidates.
+
+* New features or recent big changes, when close to the release, should have done
+  in a way that it's easy to disable. If they misbehave, prefer disabling them than
+  releasing an unstable (but untested) binary package.

Added: www-releases/trunk/4.0.0/docs/_sources/ReportingGuide.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/ReportingGuide.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/ReportingGuide.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/ReportingGuide.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,143 @@
+===============
+Reporting Guide
+===============
+
+.. note::
+
+   This document is currently a **DRAFT** document while it is being discussed
+   by the community.
+
+If you believe someone is violating the :doc:`code of conduct <CodeOfConduct>`
+you can always report it to the LLVM Foundation Code of Conduct Advisory
+Committee by emailing conduct at llvm.org. **All reports will be kept
+confidential.** This isn't a public list and only `members`_ of the advisory
+committee will receive the report.
+
+If you believe anyone is in **physical danger**, please notify appropriate law
+enforcement first. If you are unsure what law enforcement agency is
+appropriate, please include this in your report and we will attempt to notify
+them.
+
+If the violation occurs at an event such as a Developer Meeting and requires
+immediate attention, you can also reach out to any of the event organizers or
+staff. Event organizers and staff will be prepared to handle the incident and
+able to help. If you cannot find one of the organizers, the venue staff can
+locate one for you. We will also post detailed contact information for specific
+events as part of each events' information. In person reports will still be
+kept confidential exactly as above, but also feel free to (anonymously if
+needed) email conduct at llvm.org.
+
+.. note::
+   The LLVM community has long handled inappropriate behavior on its own, using
+   both private communication and public responses. Nothing in this document is
+   intended to discourage this self enforcement of community norms. Instead,
+   the mechanisms described here are intended to supplement any self
+   enforcement within the community. They provide avenues for handling severe
+   cases or cases where the reporting party does not wish to respond directly
+   for any reason.
+
+Filing a report
+===============
+
+Reports can be as formal or informal as needed for the situation at hand. If
+possible, please include as much information as you can. If you feel
+comfortable, please consider including:
+
+* Your contact info (so we can get in touch with you if we need to follow up).
+* Names (real, nicknames, or pseudonyms) of any individuals involved. If there
+  were other witnesses besides you, please try to include them as well.
+* When and where the incident occurred. Please be as specific as possible.
+* Your account of what occurred. If there is a publicly available record (e.g.
+  a mailing list archive or a public IRC logger) please include a link.
+* Any extra context you believe existed for the incident.
+* If you believe this incident is ongoing.
+* Any other information you believe we should have.
+
+What happens after you file a report?
+=====================================
+
+You will receive an email from the advisory committee acknowledging receipt
+within 24 hours (and we will aim to respond much quicker than that).
+
+The advisory committee will immediately meet to review the incident and try to
+determine:
+
+* What happened and who was involved.
+* Whether this event constitutes a code of conduct violation.
+* Whether this is an ongoing situation, or if there is a threat to anyone's
+  physical safety.
+
+If this is determined to be an ongoing incident or a threat to physical safety,
+the working groups' immediate priority will be to protect everyone involved.
+This means we may delay an "official" response until we believe that the
+situation has ended and that everyone is physically safe.
+
+The working group will try to contact other parties involved or witnessing the
+event to gain clarity on what happened and understand any different
+perspectives.
+
+Once the advisory committee has a complete account of the events they will make
+a decision as to how to respond. Responses may include:
+
+* Nothing, if we determine no violation occurred or it has already been
+  appropriately resolved.
+* Providing either moderation or mediation to ongoing interactions (where
+  appropriate, safe, and desired by both parties).
+* A private reprimand from the working group to the individuals involved.
+* An imposed vacation (i.e. asking someone to "take a week off" from a mailing
+  list or IRC).
+* A public reprimand.
+* A permanent or temporary ban from some or all LLVM spaces (mailing lists,
+  IRC, etc.)
+* Involvement of relevant law enforcement if appropriate.
+
+If the situation is not resolved within one week, we'll respond within one week
+to the original reporter with an update and explanation.
+
+Once we've determined our response, we will separately contact the original
+reporter and other individuals to let them know what actions (if any) we'll be
+taking. We will take into account feedback from the individuals involved on the
+appropriateness of our response, but we don't guarantee we'll act on it.
+
+After any incident, the advisory committee will make a report on the situation
+to the LLVM Foundation board. The board may choose to make a public statement
+about the incident. If that's the case, the identities of anyone involved will
+remain confidential unless instructed by those inviduals otherwise.
+
+Appealing
+=========
+
+Only permanent resolutions (such as bans) or requests for public actions may be
+appealed. To appeal a decision of the working group, contact the LLVM
+Foundation board at board at llvm.org with your appeal and the board will review
+the case.
+
+In general, it is **not** appropriate to appeal a particular decision on
+a public mailing list. Doing so would involve disclosure of information which
+whould be confidential. Disclosing this kind of information publicly may be
+considered a separate and (potentially) more serious violation of the Code of
+Conduct. This is not meant to limit discussion of the Code of Conduct, the
+advisory board itself, or the appropriateness of responses in general, but
+**please** refrain from mentioning specific facts about cases without the
+explicit permission of all parties involved.
+
+.. _members:
+
+Members of the Code of Conduct Advisory Committee
+=================================================
+
+The members serving on the advisory committee are listed here with contact
+information in case you are more comfortable talking directly to a specific
+member of the committee.
+
+.. note::
+
+   FIXME: When we form the initial advisory committee, the members names and private contact info need to be added here.
+
+
+
+(This text is based on the `Django Project`_ Code of Conduct, which is in turn
+based on wording from the `Speak Up! project`_.)
+
+.. _Django Project: https://www.djangoproject.com/conduct/
+.. _Speak Up! project: http://speakup.io/coc.html

Added: www-releases/trunk/4.0.0/docs/_sources/ScudoHardenedAllocator.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/ScudoHardenedAllocator.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/ScudoHardenedAllocator.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/ScudoHardenedAllocator.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,167 @@
+========================
+Scudo Hardened Allocator
+========================
+
+.. contents::
+   :local:
+   :depth: 1
+
+Introduction
+============
+
+The Scudo Hardened Allocator is a user-mode allocator based on LLVM Sanitizer's
+CombinedAllocator, which aims at providing additional mitigations against heap
+based vulnerabilities, while maintaining good performance.
+
+The name "Scudo" has been retained from the initial implementation (Escudo
+meaning Shield in Spanish and Portuguese).
+
+Design
+======
+
+Chunk Header
+------------
+Every chunk of heap memory will be preceded by a chunk header. This has two
+purposes, the first one being to store various information about the chunk,
+the second one being to detect potential heap overflows. In order to achieve
+this, the header will be checksumed, involving the pointer to the chunk itself
+and a global secret. Any corruption of the header will be detected when said
+header is accessed, and the process terminated.
+
+The following information is stored in the header:
+
+- the 16-bit checksum;
+- the user requested size for that chunk, which is necessary for reallocation
+  purposes;
+- the state of the chunk (available, allocated or quarantined);
+- the allocation type (malloc, new, new[] or memalign), to detect potential
+  mismatches in the allocation APIs used;
+- whether or not the chunk is offseted (ie: if the chunk beginning is different
+  than the backend allocation beginning, which is most often the case with some
+  aligned allocations);
+- the associated offset;
+- a 16-bit salt.
+
+On x64, which is currently the only architecture supported, the header fits
+within 16-bytes, which works nicely with the minimum alignment requirements.
+
+The checksum is computed as a CRC32 (requiring the SSE 4.2 instruction set)
+of the global secret, the chunk pointer itself, and the 16 bytes of header with
+the checksum field zeroed out.
+
+The header is atomically loaded and stored to prevent races (this requires
+platform support such as the cmpxchg16b instruction). This is important as two
+consecutive chunks could belong to different threads. We also want to avoid
+any type of double fetches of information located in the header, and use local
+copies of the header for this purpose.
+
+Delayed Freelist
+-----------------
+A delayed freelist allows us to not return a chunk directly to the backend, but
+to keep it aside for a while. Once a criterion is met, the delayed freelist is
+emptied, and the quarantined chunks are returned to the backend. This helps
+mitigate use-after-free vulnerabilities by reducing the determinism of the
+allocation and deallocation patterns.
+
+This feature is using the Sanitizer's Quarantine as its base, and the amount of
+memory that it can hold is configurable by the user (see the Options section
+below).
+
+Randomness
+----------
+It is important for the allocator to not make use of fixed addresses. We use
+the dynamic base option for the SizeClassAllocator, allowing us to benefit
+from the randomness of mmap.
+
+Usage
+=====
+
+Library
+-------
+The allocator static library can be built from the LLVM build tree thanks to
+the ``scudo`` CMake rule. The associated tests can be exercised thanks to the
+``check-scudo`` CMake rule.
+
+Linking the static library to your project can require the use of the
+``whole-archive`` linker flag (or equivalent), depending on your linker.
+Additional flags might also be necessary.
+
+Your linked binary should now make use of the Scudo allocation and deallocation
+functions.
+
+You may also build Scudo like this: 
+
+.. code::
+
+  cd $LLVM/projects/compiler-rt/lib
+  clang++ -fPIC -std=c++11 -msse4.2 -mcx16 -O2 -I. scudo/*.cpp \
+    $(\ls sanitizer_common/*.{cc,S} | grep -v "sanitizer_termination\|sanitizer_common_nolibc") \
+    -shared -o scudo-allocator.so -lpthread
+
+and then use it with existing binaries as follows:
+
+.. code::
+
+  LD_PRELOAD=`pwd`/scudo-allocator.so ./a.out
+
+Options
+-------
+Several aspects of the allocator can be configured through the following ways:
+
+- by defining a ``__scudo_default_options`` function in one's program that
+  returns the options string to be parsed. Said function must have the following
+  prototype: ``extern "C" const char* __scudo_default_options()``.
+
+- through the environment variable SCUDO_OPTIONS, containing the options string
+  to be parsed. Options defined this way will override any definition made
+  through ``__scudo_default_options``;
+
+The options string follows a syntax similar to ASan, where distinct options
+can be assigned in the same string, separated by colons.
+
+For example, using the environment variable:
+
+.. code::
+
+  SCUDO_OPTIONS="DeleteSizeMismatch=1:QuarantineSizeMb=16" ./a.out
+
+Or using the function:
+
+.. code::
+
+  extern "C" const char *__scudo_default_options() {
+    return "DeleteSizeMismatch=1:QuarantineSizeMb=16";
+  }
+
+
+The following options are available:
+
++-----------------------------+---------+------------------------------------------------+
+| Option                      | Default | Description                                    |
++-----------------------------+---------+------------------------------------------------+
+| QuarantineSizeMb            | 64      | The size (in Mb) of quarantine used to delay   |
+|                             |         | the actual deallocation of chunks. Lower value |
+|                             |         | may reduce memory usage but decrease the       |
+|                             |         | effectiveness of the mitigation; a negative    |
+|                             |         | value will fallback to a default of 64Mb.      |
++-----------------------------+---------+------------------------------------------------+
+| ThreadLocalQuarantineSizeKb | 1024    | The size (in Kb) of per-thread cache use to    |
+|                             |         | offload the global quarantine. Lower value may |
+|                             |         | reduce memory usage but might increase         |
+|                             |         | contention on the global quarantine.           |
++-----------------------------+---------+------------------------------------------------+
+| DeallocationTypeMismatch    | true    | Whether or not we report errors on             |
+|                             |         | malloc/delete, new/free, new/delete[], etc.    |
++-----------------------------+---------+------------------------------------------------+
+| DeleteSizeMismatch          | true    | Whether or not we report errors on mismatch    |
+|                             |         | between sizes of new and delete.               |
++-----------------------------+---------+------------------------------------------------+
+| ZeroContents                | false   | Whether or not we zero chunk contents on       |
+|                             |         | allocation and deallocation.                   |
++-----------------------------+---------+------------------------------------------------+
+
+Allocator related common Sanitizer options can also be passed through Scudo
+options, such as ``allocator_may_return_null``. A detailed list including those
+can be found here:
+https://github.com/google/sanitizers/wiki/SanitizerCommonFlags.
+

Added: www-releases/trunk/4.0.0/docs/_sources/SegmentedStacks.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/SegmentedStacks.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/SegmentedStacks.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/SegmentedStacks.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,77 @@
+========================
+Segmented Stacks in LLVM
+========================
+
+.. contents::
+   :local:
+
+Introduction
+============
+
+Segmented stack allows stack space to be allocated incrementally than as a
+monolithic chunk (of some worst case size) at thread initialization. This is
+done by allocating stack blocks (henceforth called *stacklets*) and linking them
+into a doubly linked list. The function prologue is responsible for checking if
+the current stacklet has enough space for the function to execute; and if not,
+call into the libgcc runtime to allocate more stack space. Segmented stacks are
+enabled with the ``"split-stack"`` attribute on LLVM functions.
+
+The runtime functionality is `already there in libgcc
+<http://gcc.gnu.org/wiki/SplitStacks>`_.
+
+Implementation Details
+======================
+
+.. _allocating stacklets:
+
+Allocating Stacklets
+--------------------
+
+As mentioned above, the function prologue checks if the current stacklet has
+enough space. The current approach is to use a slot in the TCB to store the
+current stack limit (minus the amount of space needed to allocate a new block) -
+this slot's offset is again dictated by ``libgcc``. The generated
+assembly looks like this on x86-64:
+
+.. code-block:: text
+
+    leaq     -8(%rsp), %r10
+    cmpq     %fs:112,  %r10
+    jg       .LBB0_2
+
+    # More stack space needs to be allocated
+    movabsq  $8, %r10   # The amount of space needed
+    movabsq  $0, %r11   # The total size of arguments passed on stack
+    callq    __morestack
+    ret                 # The reason for this extra return is explained below
+  .LBB0_2:
+    # Usual prologue continues here
+
+The size of function arguments on the stack needs to be passed to
+``__morestack`` (this function is implemented in ``libgcc``) since that number
+of bytes has to be copied from the previous stacklet to the current one. This is
+so that SP (and FP) relative addressing of function arguments work as expected.
+
+The unusual ``ret`` is needed to have the function which made a call to
+``__morestack`` return correctly. ``__morestack``, instead of returning, calls
+into ``.LBB0_2``. This is possible since both, the size of the ``ret``
+instruction and the PC of call to ``__morestack`` are known. When the function
+body returns, control is transferred back to ``__morestack``. ``__morestack``
+then de-allocates the new stacklet, restores the correct SP value, and does a
+second return, which returns control to the correct caller.
+
+Variable Sized Allocas
+----------------------
+
+The section on `allocating stacklets`_ automatically assumes that every stack
+frame will be of fixed size. However, LLVM allows the use of the ``llvm.alloca``
+intrinsic to allocate dynamically sized blocks of memory on the stack. When
+faced with such a variable-sized alloca, code is generated to:
+
+* Check if the current stacklet has enough space. If yes, just bump the SP, like
+  in the normal case.
+* If not, generate a call to ``libgcc``, which allocates the memory from the
+  heap.
+
+The memory allocated from the heap is linked into a list in the current
+stacklet, and freed along with the same. This prevents a memory leak.

Added: www-releases/trunk/4.0.0/docs/_sources/SourceLevelDebugging.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/SourceLevelDebugging.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/SourceLevelDebugging.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/SourceLevelDebugging.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,1419 @@
+================================
+Source Level Debugging with LLVM
+================================
+
+.. contents::
+   :local:
+
+Introduction
+============
+
+This document is the central repository for all information pertaining to debug
+information in LLVM.  It describes the :ref:`actual format that the LLVM debug
+information takes <format>`, which is useful for those interested in creating
+front-ends or dealing directly with the information.  Further, this document
+provides specific examples of what debug information for C/C++ looks like.
+
+Philosophy behind LLVM debugging information
+--------------------------------------------
+
+The idea of the LLVM debugging information is to capture how the important
+pieces of the source-language's Abstract Syntax Tree map onto LLVM code.
+Several design aspects have shaped the solution that appears here.  The
+important ones are:
+
+* Debugging information should have very little impact on the rest of the
+  compiler.  No transformations, analyses, or code generators should need to
+  be modified because of debugging information.
+
+* LLVM optimizations should interact in :ref:`well-defined and easily described
+  ways <intro_debugopt>` with the debugging information.
+
+* Because LLVM is designed to support arbitrary programming languages,
+  LLVM-to-LLVM tools should not need to know anything about the semantics of
+  the source-level-language.
+
+* Source-level languages are often **widely** different from one another.
+  LLVM should not put any restrictions of the flavor of the source-language,
+  and the debugging information should work with any language.
+
+* With code generator support, it should be possible to use an LLVM compiler
+  to compile a program to native machine code and standard debugging
+  formats.  This allows compatibility with traditional machine-code level
+  debuggers, like GDB or DBX.
+
+The approach used by the LLVM implementation is to use a small set of
+:ref:`intrinsic functions <format_common_intrinsics>` to define a mapping
+between LLVM program objects and the source-level objects.  The description of
+the source-level program is maintained in LLVM metadata in an
+:ref:`implementation-defined format <ccxx_frontend>` (the C/C++ front-end
+currently uses working draft 7 of the `DWARF 3 standard
+<http://www.eagercon.com/dwarf/dwarf3std.htm>`_).
+
+When a program is being debugged, a debugger interacts with the user and turns
+the stored debug information into source-language specific information.  As
+such, a debugger must be aware of the source-language, and is thus tied to a
+specific language or family of languages.
+
+Debug information consumers
+---------------------------
+
+The role of debug information is to provide meta information normally stripped
+away during the compilation process.  This meta information provides an LLVM
+user a relationship between generated code and the original program source
+code.
+
+Currently, there are two backend consumers of debug info: DwarfDebug and
+CodeViewDebug. DwarfDebug produces DWARF suitable for use with GDB, LLDB, and
+other DWARF-based debuggers. :ref:`CodeViewDebug <codeview>` produces CodeView,
+the Microsoft debug info format, which is usable with Microsoft debuggers such
+as Visual Studio and WinDBG. LLVM's debug information format is mostly derived
+from and inspired by DWARF, but it is feasible to translate into other target
+debug info formats such as STABS.
+
+It would also be reasonable to use debug information to feed profiling tools
+for analysis of generated code, or, tools for reconstructing the original
+source from generated code.
+
+.. _intro_debugopt:
+
+Debugging optimized code
+------------------------
+
+An extremely high priority of LLVM debugging information is to make it interact
+well with optimizations and analysis.  In particular, the LLVM debug
+information provides the following guarantees:
+
+* LLVM debug information **always provides information to accurately read
+  the source-level state of the program**, regardless of which LLVM
+  optimizations have been run, and without any modification to the
+  optimizations themselves.  However, some optimizations may impact the
+  ability to modify the current state of the program with a debugger, such
+  as setting program variables, or calling functions that have been
+  deleted.
+
+* As desired, LLVM optimizations can be upgraded to be aware of debugging
+  information, allowing them to update the debugging information as they
+  perform aggressive optimizations.  This means that, with effort, the LLVM
+  optimizers could optimize debug code just as well as non-debug code.
+
+* LLVM debug information does not prevent optimizations from
+  happening (for example inlining, basic block reordering/merging/cleanup,
+  tail duplication, etc).
+
+* LLVM debug information is automatically optimized along with the rest of
+  the program, using existing facilities.  For example, duplicate
+  information is automatically merged by the linker, and unused information
+  is automatically removed.
+
+Basically, the debug information allows you to compile a program with
+"``-O0 -g``" and get full debug information, allowing you to arbitrarily modify
+the program as it executes from a debugger.  Compiling a program with
+"``-O3 -g``" gives you full debug information that is always available and
+accurate for reading (e.g., you get accurate stack traces despite tail call
+elimination and inlining), but you might lose the ability to modify the program
+and call functions which were optimized out of the program, or inlined away
+completely.
+
+The :ref:`LLVM test suite <test-suite-quickstart>` provides a framework to test
+optimizer's handling of debugging information.  It can be run like this:
+
+.. code-block:: bash
+
+  % cd llvm/projects/test-suite/MultiSource/Benchmarks  # or some other level
+  % make TEST=dbgopt
+
+This will test impact of debugging information on optimization passes.  If
+debugging information influences optimization passes then it will be reported
+as a failure.  See :doc:`TestingGuide` for more information on LLVM test
+infrastructure and how to run various tests.
+
+.. _format:
+
+Debugging information format
+============================
+
+LLVM debugging information has been carefully designed to make it possible for
+the optimizer to optimize the program and debugging information without
+necessarily having to know anything about debugging information.  In
+particular, the use of metadata avoids duplicated debugging information from
+the beginning, and the global dead code elimination pass automatically deletes
+debugging information for a function if it decides to delete the function.
+
+To do this, most of the debugging information (descriptors for types,
+variables, functions, source files, etc) is inserted by the language front-end
+in the form of LLVM metadata.
+
+Debug information is designed to be agnostic about the target debugger and
+debugging information representation (e.g. DWARF/Stabs/etc).  It uses a generic
+pass to decode the information that represents variables, types, functions,
+namespaces, etc: this allows for arbitrary source-language semantics and
+type-systems to be used, as long as there is a module written for the target
+debugger to interpret the information.
+
+To provide basic functionality, the LLVM debugger does have to make some
+assumptions about the source-level language being debugged, though it keeps
+these to a minimum.  The only common features that the LLVM debugger assumes
+exist are `source files <LangRef.html#difile>`_, and `program objects
+<LangRef.html#diglobalvariable>`_.  These abstract objects are used by a
+debugger to form stack traces, show information about local variables, etc.
+
+This section of the documentation first describes the representation aspects
+common to any source-language.  :ref:`ccxx_frontend` describes the data layout
+conventions used by the C and C++ front-ends.
+
+Debug information descriptors are `specialized metadata nodes
+<LangRef.html#specialized-metadata>`_, first-class subclasses of ``Metadata``.
+
+.. _format_common_intrinsics:
+
+Debugger intrinsic functions
+----------------------------
+
+LLVM uses several intrinsic functions (name prefixed with "``llvm.dbg``") to
+provide debug information at various points in generated code.
+
+``llvm.dbg.declare``
+^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: llvm
+
+  void @llvm.dbg.declare(metadata, metadata, metadata)
+
+This intrinsic provides information about a local element (e.g., variable).
+The first argument is metadata holding the alloca for the variable.  The second
+argument is a `local variable <LangRef.html#dilocalvariable>`_ containing a
+description of the variable.  The third argument is a `complex expression
+<LangRef.html#diexpression>`_.
+
+``llvm.dbg.value``
+^^^^^^^^^^^^^^^^^^
+
+.. code-block:: llvm
+
+  void @llvm.dbg.value(metadata, i64, metadata, metadata)
+
+This intrinsic provides information when a user source variable is set to a new
+value.  The first argument is the new value (wrapped as metadata).  The second
+argument is the offset in the user source variable where the new value is
+written.  The third argument is a `local variable
+<LangRef.html#dilocalvariable>`_ containing a description of the variable.  The
+fourth argument is a `complex expression <LangRef.html#diexpression>`_.
+
+Object lifetimes and scoping
+============================
+
+In many languages, the local variables in functions can have their lifetimes or
+scopes limited to a subset of a function.  In the C family of languages, for
+example, variables are only live (readable and writable) within the source
+block that they are defined in.  In functional languages, values are only
+readable after they have been defined.  Though this is a very obvious concept,
+it is non-trivial to model in LLVM, because it has no notion of scoping in this
+sense, and does not want to be tied to a language's scoping rules.
+
+In order to handle this, the LLVM debug format uses the metadata attached to
+llvm instructions to encode line number and scoping information.  Consider the
+following C fragment, for example:
+
+.. code-block:: c
+
+  1.  void foo() {
+  2.    int X = 21;
+  3.    int Y = 22;
+  4.    {
+  5.      int Z = 23;
+  6.      Z = X;
+  7.    }
+  8.    X = Y;
+  9.  }
+
+Compiled to LLVM, this function would be represented like this:
+
+.. code-block:: text
+
+  ; Function Attrs: nounwind ssp uwtable
+  define void @foo() #0 !dbg !4 {
+  entry:
+    %X = alloca i32, align 4
+    %Y = alloca i32, align 4
+    %Z = alloca i32, align 4
+    call void @llvm.dbg.declare(metadata i32* %X, metadata !11, metadata !13), !dbg !14
+    store i32 21, i32* %X, align 4, !dbg !14
+    call void @llvm.dbg.declare(metadata i32* %Y, metadata !15, metadata !13), !dbg !16
+    store i32 22, i32* %Y, align 4, !dbg !16
+    call void @llvm.dbg.declare(metadata i32* %Z, metadata !17, metadata !13), !dbg !19
+    store i32 23, i32* %Z, align 4, !dbg !19
+    %0 = load i32, i32* %X, align 4, !dbg !20
+    store i32 %0, i32* %Z, align 4, !dbg !21
+    %1 = load i32, i32* %Y, align 4, !dbg !22
+    store i32 %1, i32* %X, align 4, !dbg !23
+    ret void, !dbg !24
+  }
+
+  ; Function Attrs: nounwind readnone
+  declare void @llvm.dbg.declare(metadata, metadata, metadata) #1
+
+  attributes #0 = { nounwind ssp uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="true" "no-frame-pointer-elim-non-leaf" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
+  attributes #1 = { nounwind readnone }
+
+  !llvm.dbg.cu = !{!0}
+  !llvm.module.flags = !{!7, !8, !9}
+  !llvm.ident = !{!10}
+
+  !0 = !DICompileUnit(language: DW_LANG_C99, file: !1, producer: "clang version 3.7.0 (trunk 231150) (llvm/trunk 231154)", isOptimized: false, runtimeVersion: 0, emissionKind: FullDebug, enums: !2, retainedTypes: !2, subprograms: !3, globals: !2, imports: !2)
+  !1 = !DIFile(filename: "/dev/stdin", directory: "/Users/dexonsmith/data/llvm/debug-info")
+  !2 = !{}
+  !3 = !{!4}
+  !4 = distinct !DISubprogram(name: "foo", scope: !1, file: !1, line: 1, type: !5, isLocal: false, isDefinition: true, scopeLine: 1, isOptimized: false, variables: !2)
+  !5 = !DISubroutineType(types: !6)
+  !6 = !{null}
+  !7 = !{i32 2, !"Dwarf Version", i32 2}
+  !8 = !{i32 2, !"Debug Info Version", i32 3}
+  !9 = !{i32 1, !"PIC Level", i32 2}
+  !10 = !{!"clang version 3.7.0 (trunk 231150) (llvm/trunk 231154)"}
+  !11 = !DILocalVariable(name: "X", scope: !4, file: !1, line: 2, type: !12)
+  !12 = !DIBasicType(name: "int", size: 32, align: 32, encoding: DW_ATE_signed)
+  !13 = !DIExpression()
+  !14 = !DILocation(line: 2, column: 9, scope: !4)
+  !15 = !DILocalVariable(name: "Y", scope: !4, file: !1, line: 3, type: !12)
+  !16 = !DILocation(line: 3, column: 9, scope: !4)
+  !17 = !DILocalVariable(name: "Z", scope: !18, file: !1, line: 5, type: !12)
+  !18 = distinct !DILexicalBlock(scope: !4, file: !1, line: 4, column: 5)
+  !19 = !DILocation(line: 5, column: 11, scope: !18)
+  !20 = !DILocation(line: 6, column: 11, scope: !18)
+  !21 = !DILocation(line: 6, column: 9, scope: !18)
+  !22 = !DILocation(line: 8, column: 9, scope: !4)
+  !23 = !DILocation(line: 8, column: 7, scope: !4)
+  !24 = !DILocation(line: 9, column: 3, scope: !4)
+
+
+This example illustrates a few important details about LLVM debugging
+information.  In particular, it shows how the ``llvm.dbg.declare`` intrinsic and
+location information, which are attached to an instruction, are applied
+together to allow a debugger to analyze the relationship between statements,
+variable definitions, and the code used to implement the function.
+
+.. code-block:: llvm
+
+  call void @llvm.dbg.declare(metadata i32* %X, metadata !11, metadata !13), !dbg !14
+    ; [debug line = 2:7] [debug variable = X]
+
+The first intrinsic ``%llvm.dbg.declare`` encodes debugging information for the
+variable ``X``.  The metadata ``!dbg !14`` attached to the intrinsic provides
+scope information for the variable ``X``.
+
+.. code-block:: text
+
+  !14 = !DILocation(line: 2, column: 9, scope: !4)
+  !4 = distinct !DISubprogram(name: "foo", scope: !1, file: !1, line: 1, type: !5,
+                              isLocal: false, isDefinition: true, scopeLine: 1,
+                              isOptimized: false, variables: !2)
+
+Here ``!14`` is metadata providing `location information
+<LangRef.html#dilocation>`_.  In this example, scope is encoded by ``!4``, a
+`subprogram descriptor <LangRef.html#disubprogram>`_.  This way the location
+information attached to the intrinsics indicates that the variable ``X`` is
+declared at line number 2 at a function level scope in function ``foo``.
+
+Now lets take another example.
+
+.. code-block:: llvm
+
+  call void @llvm.dbg.declare(metadata i32* %Z, metadata !17, metadata !13), !dbg !19
+    ; [debug line = 5:9] [debug variable = Z]
+
+The third intrinsic ``%llvm.dbg.declare`` encodes debugging information for
+variable ``Z``.  The metadata ``!dbg !19`` attached to the intrinsic provides
+scope information for the variable ``Z``.
+
+.. code-block:: text
+
+  !18 = distinct !DILexicalBlock(scope: !4, file: !1, line: 4, column: 5)
+  !19 = !DILocation(line: 5, column: 11, scope: !18)
+
+Here ``!19`` indicates that ``Z`` is declared at line number 5 and column
+number 0 inside of lexical scope ``!18``.  The lexical scope itself resides
+inside of subprogram ``!4`` described above.
+
+The scope information attached with each instruction provides a straightforward
+way to find instructions covered by a scope.
+
+.. _ccxx_frontend:
+
+C/C++ front-end specific debug information
+==========================================
+
+The C and C++ front-ends represent information about the program in a format
+that is effectively identical to `DWARF 3.0
+<http://www.eagercon.com/dwarf/dwarf3std.htm>`_ in terms of information
+content.  This allows code generators to trivially support native debuggers by
+generating standard dwarf information, and contains enough information for
+non-dwarf targets to translate it as needed.
+
+This section describes the forms used to represent C and C++ programs.  Other
+languages could pattern themselves after this (which itself is tuned to
+representing programs in the same way that DWARF 3 does), or they could choose
+to provide completely different forms if they don't fit into the DWARF model.
+As support for debugging information gets added to the various LLVM
+source-language front-ends, the information used should be documented here.
+
+The following sections provide examples of a few C/C++ constructs and the debug
+information that would best describe those constructs.  The canonical
+references are the ``DIDescriptor`` classes defined in
+``include/llvm/IR/DebugInfo.h`` and the implementations of the helper functions
+in ``lib/IR/DIBuilder.cpp``.
+
+C/C++ source file information
+-----------------------------
+
+``llvm::Instruction`` provides easy access to metadata attached with an
+instruction.  One can extract line number information encoded in LLVM IR using
+``Instruction::getDebugLoc()`` and ``DILocation::getLine()``.
+
+.. code-block:: c++
+
+  if (DILocation *Loc = I->getDebugLoc()) { // Here I is an LLVM instruction
+    unsigned Line = Loc->getLine();
+    StringRef File = Loc->getFilename();
+    StringRef Dir = Loc->getDirectory();
+  }
+
+C/C++ global variable information
+---------------------------------
+
+Given an integer global variable declared as follows:
+
+.. code-block:: c
+
+  _Alignas(8) int MyGlobal = 100;
+
+a C/C++ front-end would generate the following descriptors:
+
+.. code-block:: text
+
+  ;;
+  ;; Define the global itself.
+  ;;
+  @MyGlobal = global i32 100, align 8, !dbg !0
+
+  ;;
+  ;; List of debug info of globals
+  ;;
+  !llvm.dbg.cu = !{!1}
+
+  ;; Some unrelated metadata.
+  !llvm.module.flags = !{!6, !7}
+  !llvm.ident = !{!8}
+
+  ;; Define the global variable itself
+  !0 = distinct !DIGlobalVariable(name: "MyGlobal", scope: !1, file: !2, line: 1, type: !5, isLocal: false, isDefinition: true, align: 64)
+
+  ;; Define the compile unit.
+  !1 = distinct !DICompileUnit(language: DW_LANG_C99, file: !2,
+                               producer: "clang version 4.0.0 (http://llvm.org/git/clang.git ae4deadbea242e8ea517eef662c30443f75bd086) (http://llvm.org/git/llvm.git 818b4c1539df3e51dc7e62c89ead4abfd348827d)",
+                               isOptimized: false, runtimeVersion: 0, emissionKind: FullDebug,
+                               enums: !3, globals: !4)
+
+  ;;
+  ;; Define the file
+  ;;
+  !2 = !DIFile(filename: "/dev/stdin",
+               directory: "/Users/dexonsmith/data/llvm/debug-info")
+
+  ;; An empty array.
+  !3 = !{}
+
+  ;; The Array of Global Variables
+  !4 = !{!0}
+
+  ;;
+  ;; Define the type
+  ;;
+  !5 = !DIBasicType(name: "int", size: 32, encoding: DW_ATE_signed)
+
+  ;; Dwarf version to output.
+  !6 = !{i32 2, !"Dwarf Version", i32 4}
+
+  ;; Debug info schema version.
+  !7 = !{i32 2, !"Debug Info Version", i32 3}
+
+  ;; Compiler identification
+  !8 = !{!"clang version 4.0.0 (http://llvm.org/git/clang.git ae4deadbea242e8ea517eef662c30443f75bd086) (http://llvm.org/git/llvm.git 818b4c1539df3e51dc7e62c89ead4abfd348827d)"}
+
+
+The align value in DIGlobalVariable description specifies variable alignment in
+case it was forced by C11 _Alignas(), C++11 alignas() keywords or compiler
+attribute __attribute__((aligned ())). In other case (when this field is missing)
+alignment is considered default. This is used when producing DWARF output
+for DW_AT_alignment value.
+
+C/C++ function information
+--------------------------
+
+Given a function declared as follows:
+
+.. code-block:: c
+
+  int main(int argc, char *argv[]) {
+    return 0;
+  }
+
+a C/C++ front-end would generate the following descriptors:
+
+.. code-block:: text
+
+  ;;
+  ;; Define the anchor for subprograms.
+  ;;
+  !4 = !DISubprogram(name: "main", scope: !1, file: !1, line: 1, type: !5,
+                     isLocal: false, isDefinition: true, scopeLine: 1,
+                     flags: DIFlagPrototyped, isOptimized: false,
+                     variables: !2)
+
+  ;;
+  ;; Define the subprogram itself.
+  ;;
+  define i32 @main(i32 %argc, i8** %argv) !dbg !4 {
+  ...
+  }
+
+Debugging information format
+============================
+
+Debugging Information Extension for Objective C Properties
+----------------------------------------------------------
+
+Introduction
+^^^^^^^^^^^^
+
+Objective C provides a simpler way to declare and define accessor methods using
+declared properties.  The language provides features to declare a property and
+to let compiler synthesize accessor methods.
+
+The debugger lets developer inspect Objective C interfaces and their instance
+variables and class variables.  However, the debugger does not know anything
+about the properties defined in Objective C interfaces.  The debugger consumes
+information generated by compiler in DWARF format.  The format does not support
+encoding of Objective C properties.  This proposal describes DWARF extensions to
+encode Objective C properties, which the debugger can use to let developers
+inspect Objective C properties.
+
+Proposal
+^^^^^^^^
+
+Objective C properties exist separately from class members.  A property can be
+defined only by "setter" and "getter" selectors, and be calculated anew on each
+access.  Or a property can just be a direct access to some declared ivar.
+Finally it can have an ivar "automatically synthesized" for it by the compiler,
+in which case the property can be referred to in user code directly using the
+standard C dereference syntax as well as through the property "dot" syntax, but
+there is no entry in the ``@interface`` declaration corresponding to this ivar.
+
+To facilitate debugging, these properties we will add a new DWARF TAG into the
+``DW_TAG_structure_type`` definition for the class to hold the description of a
+given property, and a set of DWARF attributes that provide said description.
+The property tag will also contain the name and declared type of the property.
+
+If there is a related ivar, there will also be a DWARF property attribute placed
+in the ``DW_TAG_member`` DIE for that ivar referring back to the property TAG
+for that property.  And in the case where the compiler synthesizes the ivar
+directly, the compiler is expected to generate a ``DW_TAG_member`` for that
+ivar (with the ``DW_AT_artificial`` set to 1), whose name will be the name used
+to access this ivar directly in code, and with the property attribute pointing
+back to the property it is backing.
+
+The following examples will serve as illustration for our discussion:
+
+.. code-block:: objc
+
+  @interface I1 {
+    int n2;
+  }
+
+  @property int p1;
+  @property int p2;
+  @end
+
+  @implementation I1
+  @synthesize p1;
+  @synthesize p2 = n2;
+  @end
+
+This produces the following DWARF (this is a "pseudo dwarfdump" output):
+
+.. code-block:: none
+
+  0x00000100:  TAG_structure_type [7] *
+                 AT_APPLE_runtime_class( 0x10 )
+                 AT_name( "I1" )
+                 AT_decl_file( "Objc_Property.m" )
+                 AT_decl_line( 3 )
+
+  0x00000110    TAG_APPLE_property
+                  AT_name ( "p1" )
+                  AT_type ( {0x00000150} ( int ) )
+
+  0x00000120:   TAG_APPLE_property
+                  AT_name ( "p2" )
+                  AT_type ( {0x00000150} ( int ) )
+
+  0x00000130:   TAG_member [8]
+                  AT_name( "_p1" )
+                  AT_APPLE_property ( {0x00000110} "p1" )
+                  AT_type( {0x00000150} ( int ) )
+                  AT_artificial ( 0x1 )
+
+  0x00000140:    TAG_member [8]
+                   AT_name( "n2" )
+                   AT_APPLE_property ( {0x00000120} "p2" )
+                   AT_type( {0x00000150} ( int ) )
+
+  0x00000150:  AT_type( ( int ) )
+
+Note, the current convention is that the name of the ivar for an
+auto-synthesized property is the name of the property from which it derives
+with an underscore prepended, as is shown in the example.  But we actually
+don't need to know this convention, since we are given the name of the ivar
+directly.
+
+Also, it is common practice in ObjC to have different property declarations in
+the @interface and @implementation - e.g. to provide a read-only property in
+the interface,and a read-write interface in the implementation.  In that case,
+the compiler should emit whichever property declaration will be in force in the
+current translation unit.
+
+Developers can decorate a property with attributes which are encoded using
+``DW_AT_APPLE_property_attribute``.
+
+.. code-block:: objc
+
+  @property (readonly, nonatomic) int pr;
+
+.. code-block:: none
+
+  TAG_APPLE_property [8]
+    AT_name( "pr" )
+    AT_type ( {0x00000147} (int) )
+    AT_APPLE_property_attribute (DW_APPLE_PROPERTY_readonly, DW_APPLE_PROPERTY_nonatomic)
+
+The setter and getter method names are attached to the property using
+``DW_AT_APPLE_property_setter`` and ``DW_AT_APPLE_property_getter`` attributes.
+
+.. code-block:: objc
+
+  @interface I1
+  @property (setter=myOwnP3Setter:) int p3;
+  -(void)myOwnP3Setter:(int)a;
+  @end
+
+  @implementation I1
+  @synthesize p3;
+  -(void)myOwnP3Setter:(int)a{ }
+  @end
+
+The DWARF for this would be:
+
+.. code-block:: none
+
+  0x000003bd: TAG_structure_type [7] *
+                AT_APPLE_runtime_class( 0x10 )
+                AT_name( "I1" )
+                AT_decl_file( "Objc_Property.m" )
+                AT_decl_line( 3 )
+
+  0x000003cd      TAG_APPLE_property
+                    AT_name ( "p3" )
+                    AT_APPLE_property_setter ( "myOwnP3Setter:" )
+                    AT_type( {0x00000147} ( int ) )
+
+  0x000003f3:     TAG_member [8]
+                    AT_name( "_p3" )
+                    AT_type ( {0x00000147} ( int ) )
+                    AT_APPLE_property ( {0x000003cd} )
+                    AT_artificial ( 0x1 )
+
+New DWARF Tags
+^^^^^^^^^^^^^^
+
++-----------------------+--------+
+| TAG                   | Value  |
++=======================+========+
+| DW_TAG_APPLE_property | 0x4200 |
++-----------------------+--------+
+
+New DWARF Attributes
+^^^^^^^^^^^^^^^^^^^^
+
++--------------------------------+--------+-----------+
+| Attribute                      | Value  | Classes   |
++================================+========+===========+
+| DW_AT_APPLE_property           | 0x3fed | Reference |
++--------------------------------+--------+-----------+
+| DW_AT_APPLE_property_getter    | 0x3fe9 | String    |
++--------------------------------+--------+-----------+
+| DW_AT_APPLE_property_setter    | 0x3fea | String    |
++--------------------------------+--------+-----------+
+| DW_AT_APPLE_property_attribute | 0x3feb | Constant  |
++--------------------------------+--------+-----------+
+
+New DWARF Constants
+^^^^^^^^^^^^^^^^^^^
+
++--------------------------------------+-------+
+| Name                                 | Value |
++======================================+=======+
+| DW_APPLE_PROPERTY_readonly           | 0x01  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_getter             | 0x02  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_assign             | 0x04  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_readwrite          | 0x08  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_retain             | 0x10  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_copy               | 0x20  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_nonatomic          | 0x40  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_setter             | 0x80  |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_atomic             | 0x100 |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_weak               | 0x200 |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_strong             | 0x400 |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_unsafe_unretained  | 0x800 |
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_nullability        | 0x1000|
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_null_resettable    | 0x2000|
++--------------------------------------+-------+
+| DW_APPLE_PROPERTY_class              | 0x4000|
++--------------------------------------+-------+
+
+Name Accelerator Tables
+-----------------------
+
+Introduction
+^^^^^^^^^^^^
+
+The "``.debug_pubnames``" and "``.debug_pubtypes``" formats are not what a
+debugger needs.  The "``pub``" in the section name indicates that the entries
+in the table are publicly visible names only.  This means no static or hidden
+functions show up in the "``.debug_pubnames``".  No static variables or private
+class variables are in the "``.debug_pubtypes``".  Many compilers add different
+things to these tables, so we can't rely upon the contents between gcc, icc, or
+clang.
+
+The typical query given by users tends not to match up with the contents of
+these tables.  For example, the DWARF spec states that "In the case of the name
+of a function member or static data member of a C++ structure, class or union,
+the name presented in the "``.debug_pubnames``" section is not the simple name
+given by the ``DW_AT_name attribute`` of the referenced debugging information
+entry, but rather the fully qualified name of the data or function member."
+So the only names in these tables for complex C++ entries is a fully
+qualified name.  Debugger users tend not to enter their search strings as
+"``a::b::c(int,const Foo&) const``", but rather as "``c``", "``b::c``" , or
+"``a::b::c``".  So the name entered in the name table must be demangled in
+order to chop it up appropriately and additional names must be manually entered
+into the table to make it effective as a name lookup table for debuggers to
+use.
+
+All debuggers currently ignore the "``.debug_pubnames``" table as a result of
+its inconsistent and useless public-only name content making it a waste of
+space in the object file.  These tables, when they are written to disk, are not
+sorted in any way, leaving every debugger to do its own parsing and sorting.
+These tables also include an inlined copy of the string values in the table
+itself making the tables much larger than they need to be on disk, especially
+for large C++ programs.
+
+Can't we just fix the sections by adding all of the names we need to this
+table? No, because that is not what the tables are defined to contain and we
+won't know the difference between the old bad tables and the new good tables.
+At best we could make our own renamed sections that contain all of the data we
+need.
+
+These tables are also insufficient for what a debugger like LLDB needs.  LLDB
+uses clang for its expression parsing where LLDB acts as a PCH.  LLDB is then
+often asked to look for type "``foo``" or namespace "``bar``", or list items in
+namespace "``baz``".  Namespaces are not included in the pubnames or pubtypes
+tables.  Since clang asks a lot of questions when it is parsing an expression,
+we need to be very fast when looking up names, as it happens a lot.  Having new
+accelerator tables that are optimized for very quick lookups will benefit this
+type of debugging experience greatly.
+
+We would like to generate name lookup tables that can be mapped into memory
+from disk, and used as is, with little or no up-front parsing.  We would also
+be able to control the exact content of these different tables so they contain
+exactly what we need.  The Name Accelerator Tables were designed to fix these
+issues.  In order to solve these issues we need to:
+
+* Have a format that can be mapped into memory from disk and used as is
+* Lookups should be very fast
+* Extensible table format so these tables can be made by many producers
+* Contain all of the names needed for typical lookups out of the box
+* Strict rules for the contents of tables
+
+Table size is important and the accelerator table format should allow the reuse
+of strings from common string tables so the strings for the names are not
+duplicated.  We also want to make sure the table is ready to be used as-is by
+simply mapping the table into memory with minimal header parsing.
+
+The name lookups need to be fast and optimized for the kinds of lookups that
+debuggers tend to do.  Optimally we would like to touch as few parts of the
+mapped table as possible when doing a name lookup and be able to quickly find
+the name entry we are looking for, or discover there are no matches.  In the
+case of debuggers we optimized for lookups that fail most of the time.
+
+Each table that is defined should have strict rules on exactly what is in the
+accelerator tables and documented so clients can rely on the content.
+
+Hash Tables
+^^^^^^^^^^^
+
+Standard Hash Tables
+""""""""""""""""""""
+
+Typical hash tables have a header, buckets, and each bucket points to the
+bucket contents:
+
+.. code-block:: none
+
+  .------------.
+  |  HEADER    |
+  |------------|
+  |  BUCKETS   |
+  |------------|
+  |  DATA      |
+  `------------'
+
+The BUCKETS are an array of offsets to DATA for each hash:
+
+.. code-block:: none
+
+  .------------.
+  | 0x00001000 | BUCKETS[0]
+  | 0x00002000 | BUCKETS[1]
+  | 0x00002200 | BUCKETS[2]
+  | 0x000034f0 | BUCKETS[3]
+  |            | ...
+  | 0xXXXXXXXX | BUCKETS[n_buckets]
+  '------------'
+
+So for ``bucket[3]`` in the example above, we have an offset into the table
+0x000034f0 which points to a chain of entries for the bucket.  Each bucket must
+contain a next pointer, full 32 bit hash value, the string itself, and the data
+for the current string value.
+
+.. code-block:: none
+
+              .------------.
+  0x000034f0: | 0x00003500 | next pointer
+              | 0x12345678 | 32 bit hash
+              | "erase"    | string value
+              | data[n]    | HashData for this bucket
+              |------------|
+  0x00003500: | 0x00003550 | next pointer
+              | 0x29273623 | 32 bit hash
+              | "dump"     | string value
+              | data[n]    | HashData for this bucket
+              |------------|
+  0x00003550: | 0x00000000 | next pointer
+              | 0x82638293 | 32 bit hash
+              | "main"     | string value
+              | data[n]    | HashData for this bucket
+              `------------'
+
+The problem with this layout for debuggers is that we need to optimize for the
+negative lookup case where the symbol we're searching for is not present.  So
+if we were to lookup "``printf``" in the table above, we would make a 32-bit
+hash for "``printf``", it might match ``bucket[3]``.  We would need to go to
+the offset 0x000034f0 and start looking to see if our 32 bit hash matches.  To
+do so, we need to read the next pointer, then read the hash, compare it, and
+skip to the next bucket.  Each time we are skipping many bytes in memory and
+touching new pages just to do the compare on the full 32 bit hash.  All of
+these accesses then tell us that we didn't have a match.
+
+Name Hash Tables
+""""""""""""""""
+
+To solve the issues mentioned above we have structured the hash tables a bit
+differently: a header, buckets, an array of all unique 32 bit hash values,
+followed by an array of hash value data offsets, one for each hash value, then
+the data for all hash values:
+
+.. code-block:: none
+
+  .-------------.
+  |  HEADER     |
+  |-------------|
+  |  BUCKETS    |
+  |-------------|
+  |  HASHES     |
+  |-------------|
+  |  OFFSETS    |
+  |-------------|
+  |  DATA       |
+  `-------------'
+
+The ``BUCKETS`` in the name tables are an index into the ``HASHES`` array.  By
+making all of the full 32 bit hash values contiguous in memory, we allow
+ourselves to efficiently check for a match while touching as little memory as
+possible.  Most often checking the 32 bit hash values is as far as the lookup
+goes.  If it does match, it usually is a match with no collisions.  So for a
+table with "``n_buckets``" buckets, and "``n_hashes``" unique 32 bit hash
+values, we can clarify the contents of the ``BUCKETS``, ``HASHES`` and
+``OFFSETS`` as:
+
+.. code-block:: none
+
+  .-------------------------.
+  |  HEADER.magic           | uint32_t
+  |  HEADER.version         | uint16_t
+  |  HEADER.hash_function   | uint16_t
+  |  HEADER.bucket_count    | uint32_t
+  |  HEADER.hashes_count    | uint32_t
+  |  HEADER.header_data_len | uint32_t
+  |  HEADER_DATA            | HeaderData
+  |-------------------------|
+  |  BUCKETS                | uint32_t[n_buckets] // 32 bit hash indexes
+  |-------------------------|
+  |  HASHES                 | uint32_t[n_hashes] // 32 bit hash values
+  |-------------------------|
+  |  OFFSETS                | uint32_t[n_hashes] // 32 bit offsets to hash value data
+  |-------------------------|
+  |  ALL HASH DATA          |
+  `-------------------------'
+
+So taking the exact same data from the standard hash example above we end up
+with:
+
+.. code-block:: none
+
+              .------------.
+              | HEADER     |
+              |------------|
+              |          0 | BUCKETS[0]
+              |          2 | BUCKETS[1]
+              |          5 | BUCKETS[2]
+              |          6 | BUCKETS[3]
+              |            | ...
+              |        ... | BUCKETS[n_buckets]
+              |------------|
+              | 0x........ | HASHES[0]
+              | 0x........ | HASHES[1]
+              | 0x........ | HASHES[2]
+              | 0x........ | HASHES[3]
+              | 0x........ | HASHES[4]
+              | 0x........ | HASHES[5]
+              | 0x12345678 | HASHES[6]    hash for BUCKETS[3]
+              | 0x29273623 | HASHES[7]    hash for BUCKETS[3]
+              | 0x82638293 | HASHES[8]    hash for BUCKETS[3]
+              | 0x........ | HASHES[9]
+              | 0x........ | HASHES[10]
+              | 0x........ | HASHES[11]
+              | 0x........ | HASHES[12]
+              | 0x........ | HASHES[13]
+              | 0x........ | HASHES[n_hashes]
+              |------------|
+              | 0x........ | OFFSETS[0]
+              | 0x........ | OFFSETS[1]
+              | 0x........ | OFFSETS[2]
+              | 0x........ | OFFSETS[3]
+              | 0x........ | OFFSETS[4]
+              | 0x........ | OFFSETS[5]
+              | 0x000034f0 | OFFSETS[6]   offset for BUCKETS[3]
+              | 0x00003500 | OFFSETS[7]   offset for BUCKETS[3]
+              | 0x00003550 | OFFSETS[8]   offset for BUCKETS[3]
+              | 0x........ | OFFSETS[9]
+              | 0x........ | OFFSETS[10]
+              | 0x........ | OFFSETS[11]
+              | 0x........ | OFFSETS[12]
+              | 0x........ | OFFSETS[13]
+              | 0x........ | OFFSETS[n_hashes]
+              |------------|
+              |            |
+              |            |
+              |            |
+              |            |
+              |            |
+              |------------|
+  0x000034f0: | 0x00001203 | .debug_str ("erase")
+              | 0x00000004 | A 32 bit array count - number of HashData with name "erase"
+              | 0x........ | HashData[0]
+              | 0x........ | HashData[1]
+              | 0x........ | HashData[2]
+              | 0x........ | HashData[3]
+              | 0x00000000 | String offset into .debug_str (terminate data for hash)
+              |------------|
+  0x00003500: | 0x00001203 | String offset into .debug_str ("collision")
+              | 0x00000002 | A 32 bit array count - number of HashData with name "collision"
+              | 0x........ | HashData[0]
+              | 0x........ | HashData[1]
+              | 0x00001203 | String offset into .debug_str ("dump")
+              | 0x00000003 | A 32 bit array count - number of HashData with name "dump"
+              | 0x........ | HashData[0]
+              | 0x........ | HashData[1]
+              | 0x........ | HashData[2]
+              | 0x00000000 | String offset into .debug_str (terminate data for hash)
+              |------------|
+  0x00003550: | 0x00001203 | String offset into .debug_str ("main")
+              | 0x00000009 | A 32 bit array count - number of HashData with name "main"
+              | 0x........ | HashData[0]
+              | 0x........ | HashData[1]
+              | 0x........ | HashData[2]
+              | 0x........ | HashData[3]
+              | 0x........ | HashData[4]
+              | 0x........ | HashData[5]
+              | 0x........ | HashData[6]
+              | 0x........ | HashData[7]
+              | 0x........ | HashData[8]
+              | 0x00000000 | String offset into .debug_str (terminate data for hash)
+              `------------'
+
+So we still have all of the same data, we just organize it more efficiently for
+debugger lookup.  If we repeat the same "``printf``" lookup from above, we
+would hash "``printf``" and find it matches ``BUCKETS[3]`` by taking the 32 bit
+hash value and modulo it by ``n_buckets``.  ``BUCKETS[3]`` contains "6" which
+is the index into the ``HASHES`` table.  We would then compare any consecutive
+32 bit hashes values in the ``HASHES`` array as long as the hashes would be in
+``BUCKETS[3]``.  We do this by verifying that each subsequent hash value modulo
+``n_buckets`` is still 3.  In the case of a failed lookup we would access the
+memory for ``BUCKETS[3]``, and then compare a few consecutive 32 bit hashes
+before we know that we have no match.  We don't end up marching through
+multiple words of memory and we really keep the number of processor data cache
+lines being accessed as small as possible.
+
+The string hash that is used for these lookup tables is the Daniel J.
+Bernstein hash which is also used in the ELF ``GNU_HASH`` sections.  It is a
+very good hash for all kinds of names in programs with very few hash
+collisions.
+
+Empty buckets are designated by using an invalid hash index of ``UINT32_MAX``.
+
+Details
+^^^^^^^
+
+These name hash tables are designed to be generic where specializations of the
+table get to define additional data that goes into the header ("``HeaderData``"),
+how the string value is stored ("``KeyType``") and the content of the data for each
+hash value.
+
+Header Layout
+"""""""""""""
+
+The header has a fixed part, and the specialized part.  The exact format of the
+header is:
+
+.. code-block:: c
+
+  struct Header
+  {
+    uint32_t   magic;           // 'HASH' magic value to allow endian detection
+    uint16_t   version;         // Version number
+    uint16_t   hash_function;   // The hash function enumeration that was used
+    uint32_t   bucket_count;    // The number of buckets in this hash table
+    uint32_t   hashes_count;    // The total number of unique hash values and hash data offsets in this table
+    uint32_t   header_data_len; // The bytes to skip to get to the hash indexes (buckets) for correct alignment
+                                // Specifically the length of the following HeaderData field - this does not
+                                // include the size of the preceding fields
+    HeaderData header_data;     // Implementation specific header data
+  };
+
+The header starts with a 32 bit "``magic``" value which must be ``'HASH'``
+encoded as an ASCII integer.  This allows the detection of the start of the
+hash table and also allows the table's byte order to be determined so the table
+can be correctly extracted.  The "``magic``" value is followed by a 16 bit
+``version`` number which allows the table to be revised and modified in the
+future.  The current version number is 1. ``hash_function`` is a ``uint16_t``
+enumeration that specifies which hash function was used to produce this table.
+The current values for the hash function enumerations include:
+
+.. code-block:: c
+
+  enum HashFunctionType
+  {
+    eHashFunctionDJB = 0u, // Daniel J Bernstein hash function
+  };
+
+``bucket_count`` is a 32 bit unsigned integer that represents how many buckets
+are in the ``BUCKETS`` array.  ``hashes_count`` is the number of unique 32 bit
+hash values that are in the ``HASHES`` array, and is the same number of offsets
+are contained in the ``OFFSETS`` array.  ``header_data_len`` specifies the size
+in bytes of the ``HeaderData`` that is filled in by specialized versions of
+this table.
+
+Fixed Lookup
+""""""""""""
+
+The header is followed by the buckets, hashes, offsets, and hash value data.
+
+.. code-block:: c
+
+  struct FixedTable
+  {
+    uint32_t buckets[Header.bucket_count];  // An array of hash indexes into the "hashes[]" array below
+    uint32_t hashes [Header.hashes_count];  // Every unique 32 bit hash for the entire table is in this table
+    uint32_t offsets[Header.hashes_count];  // An offset that corresponds to each item in the "hashes[]" array above
+  };
+
+``buckets`` is an array of 32 bit indexes into the ``hashes`` array.  The
+``hashes`` array contains all of the 32 bit hash values for all names in the
+hash table.  Each hash in the ``hashes`` table has an offset in the ``offsets``
+array that points to the data for the hash value.
+
+This table setup makes it very easy to repurpose these tables to contain
+different data, while keeping the lookup mechanism the same for all tables.
+This layout also makes it possible to save the table to disk and map it in
+later and do very efficient name lookups with little or no parsing.
+
+DWARF lookup tables can be implemented in a variety of ways and can store a lot
+of information for each name.  We want to make the DWARF tables extensible and
+able to store the data efficiently so we have used some of the DWARF features
+that enable efficient data storage to define exactly what kind of data we store
+for each name.
+
+The ``HeaderData`` contains a definition of the contents of each HashData chunk.
+We might want to store an offset to all of the debug information entries (DIEs)
+for each name.  To keep things extensible, we create a list of items, or
+Atoms, that are contained in the data for each name.  First comes the type of
+the data in each atom:
+
+.. code-block:: c
+
+  enum AtomType
+  {
+    eAtomTypeNULL       = 0u,
+    eAtomTypeDIEOffset  = 1u,   // DIE offset, check form for encoding
+    eAtomTypeCUOffset   = 2u,   // DIE offset of the compiler unit header that contains the item in question
+    eAtomTypeTag        = 3u,   // DW_TAG_xxx value, should be encoded as DW_FORM_data1 (if no tags exceed 255) or DW_FORM_data2
+    eAtomTypeNameFlags  = 4u,   // Flags from enum NameFlags
+    eAtomTypeTypeFlags  = 5u,   // Flags from enum TypeFlags
+  };
+
+The enumeration values and their meanings are:
+
+.. code-block:: none
+
+  eAtomTypeNULL       - a termination atom that specifies the end of the atom list
+  eAtomTypeDIEOffset  - an offset into the .debug_info section for the DWARF DIE for this name
+  eAtomTypeCUOffset   - an offset into the .debug_info section for the CU that contains the DIE
+  eAtomTypeDIETag     - The DW_TAG_XXX enumeration value so you don't have to parse the DWARF to see what it is
+  eAtomTypeNameFlags  - Flags for functions and global variables (isFunction, isInlined, isExternal...)
+  eAtomTypeTypeFlags  - Flags for types (isCXXClass, isObjCClass, ...)
+
+Then we allow each atom type to define the atom type and how the data for each
+atom type data is encoded:
+
+.. code-block:: c
+
+  struct Atom
+  {
+    uint16_t type;  // AtomType enum value
+    uint16_t form;  // DWARF DW_FORM_XXX defines
+  };
+
+The ``form`` type above is from the DWARF specification and defines the exact
+encoding of the data for the Atom type.  See the DWARF specification for the
+``DW_FORM_`` definitions.
+
+.. code-block:: c
+
+  struct HeaderData
+  {
+    uint32_t die_offset_base;
+    uint32_t atom_count;
+    Atoms    atoms[atom_count0];
+  };
+
+``HeaderData`` defines the base DIE offset that should be added to any atoms
+that are encoded using the ``DW_FORM_ref1``, ``DW_FORM_ref2``,
+``DW_FORM_ref4``, ``DW_FORM_ref8`` or ``DW_FORM_ref_udata``.  It also defines
+what is contained in each ``HashData`` object -- ``Atom.form`` tells us how large
+each field will be in the ``HashData`` and the ``Atom.type`` tells us how this data
+should be interpreted.
+
+For the current implementations of the "``.apple_names``" (all functions +
+globals), the "``.apple_types``" (names of all types that are defined), and
+the "``.apple_namespaces``" (all namespaces), we currently set the ``Atom``
+array to be:
+
+.. code-block:: c
+
+  HeaderData.atom_count = 1;
+  HeaderData.atoms[0].type = eAtomTypeDIEOffset;
+  HeaderData.atoms[0].form = DW_FORM_data4;
+
+This defines the contents to be the DIE offset (eAtomTypeDIEOffset) that is
+encoded as a 32 bit value (DW_FORM_data4).  This allows a single name to have
+multiple matching DIEs in a single file, which could come up with an inlined
+function for instance.  Future tables could include more information about the
+DIE such as flags indicating if the DIE is a function, method, block,
+or inlined.
+
+The KeyType for the DWARF table is a 32 bit string table offset into the
+".debug_str" table.  The ".debug_str" is the string table for the DWARF which
+may already contain copies of all of the strings.  This helps make sure, with
+help from the compiler, that we reuse the strings between all of the DWARF
+sections and keeps the hash table size down.  Another benefit to having the
+compiler generate all strings as DW_FORM_strp in the debug info, is that
+DWARF parsing can be made much faster.
+
+After a lookup is made, we get an offset into the hash data.  The hash data
+needs to be able to deal with 32 bit hash collisions, so the chunk of data
+at the offset in the hash data consists of a triple:
+
+.. code-block:: c
+
+  uint32_t str_offset
+  uint32_t hash_data_count
+  HashData[hash_data_count]
+
+If "str_offset" is zero, then the bucket contents are done. 99.9% of the
+hash data chunks contain a single item (no 32 bit hash collision):
+
+.. code-block:: none
+
+  .------------.
+  | 0x00001023 | uint32_t KeyType (.debug_str[0x0001023] => "main")
+  | 0x00000004 | uint32_t HashData count
+  | 0x........ | uint32_t HashData[0] DIE offset
+  | 0x........ | uint32_t HashData[1] DIE offset
+  | 0x........ | uint32_t HashData[2] DIE offset
+  | 0x........ | uint32_t HashData[3] DIE offset
+  | 0x00000000 | uint32_t KeyType (end of hash chain)
+  `------------'
+
+If there are collisions, you will have multiple valid string offsets:
+
+.. code-block:: none
+
+  .------------.
+  | 0x00001023 | uint32_t KeyType (.debug_str[0x0001023] => "main")
+  | 0x00000004 | uint32_t HashData count
+  | 0x........ | uint32_t HashData[0] DIE offset
+  | 0x........ | uint32_t HashData[1] DIE offset
+  | 0x........ | uint32_t HashData[2] DIE offset
+  | 0x........ | uint32_t HashData[3] DIE offset
+  | 0x00002023 | uint32_t KeyType (.debug_str[0x0002023] => "print")
+  | 0x00000002 | uint32_t HashData count
+  | 0x........ | uint32_t HashData[0] DIE offset
+  | 0x........ | uint32_t HashData[1] DIE offset
+  | 0x00000000 | uint32_t KeyType (end of hash chain)
+  `------------'
+
+Current testing with real world C++ binaries has shown that there is around 1
+32 bit hash collision per 100,000 name entries.
+
+Contents
+^^^^^^^^
+
+As we said, we want to strictly define exactly what is included in the
+different tables.  For DWARF, we have 3 tables: "``.apple_names``",
+"``.apple_types``", and "``.apple_namespaces``".
+
+"``.apple_names``" sections should contain an entry for each DWARF DIE whose
+``DW_TAG`` is a ``DW_TAG_label``, ``DW_TAG_inlined_subroutine``, or
+``DW_TAG_subprogram`` that has address attributes: ``DW_AT_low_pc``,
+``DW_AT_high_pc``, ``DW_AT_ranges`` or ``DW_AT_entry_pc``.  It also contains
+``DW_TAG_variable`` DIEs that have a ``DW_OP_addr`` in the location (global and
+static variables).  All global and static variables should be included,
+including those scoped within functions and classes.  For example using the
+following code:
+
+.. code-block:: c
+
+  static int var = 0;
+
+  void f ()
+  {
+    static int var = 0;
+  }
+
+Both of the static ``var`` variables would be included in the table.  All
+functions should emit both their full names and their basenames.  For C or C++,
+the full name is the mangled name (if available) which is usually in the
+``DW_AT_MIPS_linkage_name`` attribute, and the ``DW_AT_name`` contains the
+function basename.  If global or static variables have a mangled name in a
+``DW_AT_MIPS_linkage_name`` attribute, this should be emitted along with the
+simple name found in the ``DW_AT_name`` attribute.
+
+"``.apple_types``" sections should contain an entry for each DWARF DIE whose
+tag is one of:
+
+* DW_TAG_array_type
+* DW_TAG_class_type
+* DW_TAG_enumeration_type
+* DW_TAG_pointer_type
+* DW_TAG_reference_type
+* DW_TAG_string_type
+* DW_TAG_structure_type
+* DW_TAG_subroutine_type
+* DW_TAG_typedef
+* DW_TAG_union_type
+* DW_TAG_ptr_to_member_type
+* DW_TAG_set_type
+* DW_TAG_subrange_type
+* DW_TAG_base_type
+* DW_TAG_const_type
+* DW_TAG_file_type
+* DW_TAG_namelist
+* DW_TAG_packed_type
+* DW_TAG_volatile_type
+* DW_TAG_restrict_type
+* DW_TAG_atomic_type
+* DW_TAG_interface_type
+* DW_TAG_unspecified_type
+* DW_TAG_shared_type
+
+Only entries with a ``DW_AT_name`` attribute are included, and the entry must
+not be a forward declaration (``DW_AT_declaration`` attribute with a non-zero
+value).  For example, using the following code:
+
+.. code-block:: c
+
+  int main ()
+  {
+    int *b = 0;
+    return *b;
+  }
+
+We get a few type DIEs:
+
+.. code-block:: none
+
+  0x00000067:     TAG_base_type [5]
+                  AT_encoding( DW_ATE_signed )
+                  AT_name( "int" )
+                  AT_byte_size( 0x04 )
+
+  0x0000006e:     TAG_pointer_type [6]
+                  AT_type( {0x00000067} ( int ) )
+                  AT_byte_size( 0x08 )
+
+The DW_TAG_pointer_type is not included because it does not have a ``DW_AT_name``.
+
+"``.apple_namespaces``" section should contain all ``DW_TAG_namespace`` DIEs.
+If we run into a namespace that has no name this is an anonymous namespace, and
+the name should be output as "``(anonymous namespace)``" (without the quotes).
+Why?  This matches the output of the ``abi::cxa_demangle()`` that is in the
+standard C++ library that demangles mangled names.
+
+
+Language Extensions and File Format Changes
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Objective-C Extensions
+""""""""""""""""""""""
+
+"``.apple_objc``" section should contain all ``DW_TAG_subprogram`` DIEs for an
+Objective-C class.  The name used in the hash table is the name of the
+Objective-C class itself.  If the Objective-C class has a category, then an
+entry is made for both the class name without the category, and for the class
+name with the category.  So if we have a DIE at offset 0x1234 with a name of
+method "``-[NSString(my_additions) stringWithSpecialString:]``", we would add
+an entry for "``NSString``" that points to DIE 0x1234, and an entry for
+"``NSString(my_additions)``" that points to 0x1234.  This allows us to quickly
+track down all Objective-C methods for an Objective-C class when doing
+expressions.  It is needed because of the dynamic nature of Objective-C where
+anyone can add methods to a class.  The DWARF for Objective-C methods is also
+emitted differently from C++ classes where the methods are not usually
+contained in the class definition, they are scattered about across one or more
+compile units.  Categories can also be defined in different shared libraries.
+So we need to be able to quickly find all of the methods and class functions
+given the Objective-C class name, or quickly find all methods and class
+functions for a class + category name.  This table does not contain any
+selector names, it just maps Objective-C class names (or class names +
+category) to all of the methods and class functions.  The selectors are added
+as function basenames in the "``.debug_names``" section.
+
+In the "``.apple_names``" section for Objective-C functions, the full name is
+the entire function name with the brackets ("``-[NSString
+stringWithCString:]``") and the basename is the selector only
+("``stringWithCString:``").
+
+Mach-O Changes
+""""""""""""""
+
+The sections names for the apple hash tables are for non-mach-o files.  For
+mach-o files, the sections should be contained in the ``__DWARF`` segment with
+names as follows:
+
+* "``.apple_names``" -> "``__apple_names``"
+* "``.apple_types``" -> "``__apple_types``"
+* "``.apple_namespaces``" -> "``__apple_namespac``" (16 character limit)
+* "``.apple_objc``" -> "``__apple_objc``"
+
+.. _codeview:
+
+CodeView Debug Info Format
+==========================
+
+LLVM supports emitting CodeView, the Microsoft debug info format, and this
+section describes the design and implementation of that support.
+
+Format Background
+-----------------
+
+CodeView as a format is clearly oriented around C++ debugging, and in C++, the
+majority of debug information tends to be type information. Therefore, the
+overriding design constraint of CodeView is the separation of type information
+from other "symbol" information so that type information can be efficiently
+merged across translation units. Both type information and symbol information is
+generally stored as a sequence of records, where each record begins with a
+16-bit record size and a 16-bit record kind.
+
+Type information is usually stored in the ``.debug$T`` section of the object
+file.  All other debug info, such as line info, string table, symbol info, and
+inlinee info, is stored in one or more ``.debug$S`` sections. There may only be
+one ``.debug$T`` section per object file, since all other debug info refers to
+it. If a PDB (enabled by the ``/Zi`` MSVC option) was used during compilation,
+the ``.debug$T`` section will contain only an ``LF_TYPESERVER2`` record pointing
+to the PDB. When using PDBs, symbol information appears to remain in the object
+file ``.debug$S`` sections.
+
+Type records are referred to by their index, which is the number of records in
+the stream before a given record plus ``0x1000``. Many common basic types, such
+as the basic integral types and unqualified pointers to them, are represented
+using type indices less than ``0x1000``. Such basic types are built in to
+CodeView consumers and do not require type records.
+
+Each type record may only contain type indices that are less than its own type
+index. This ensures that the graph of type stream references is acyclic. While
+the source-level type graph may contain cycles through pointer types (consider a
+linked list struct), these cycles are removed from the type stream by always
+referring to the forward declaration record of user-defined record types. Only
+"symbol" records in the ``.debug$S`` streams may refer to complete,
+non-forward-declaration type records.
+
+Working with CodeView
+---------------------
+
+These are instructions for some common tasks for developers working to improve
+LLVM's CodeView support. Most of them revolve around using the CodeView dumper
+embedded in ``llvm-readobj``.
+
+* Testing MSVC's output::
+
+    $ cl -c -Z7 foo.cpp # Use /Z7 to keep types in the object file
+    $ llvm-readobj -codeview foo.obj
+
+* Getting LLVM IR debug info out of Clang::
+
+    $ clang -g -gcodeview --target=x86_64-windows-msvc foo.cpp -S -emit-llvm
+
+  Use this to generate LLVM IR for LLVM test cases.
+
+* Generate and dump CodeView from LLVM IR metadata::
+
+    $ llc foo.ll -filetype=obj -o foo.obj
+    $ llvm-readobj -codeview foo.obj > foo.txt
+
+  Use this pattern in lit test cases and FileCheck the output of llvm-readobj
+
+Improving LLVM's CodeView support is a process of finding interesting type
+records, constructing a C++ test case that makes MSVC emit those records,
+dumping the records, understanding them, and then generating equivalent records
+in LLVM's backend.

Added: www-releases/trunk/4.0.0/docs/_sources/SphinxQuickstartTemplate.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/SphinxQuickstartTemplate.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/SphinxQuickstartTemplate.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/SphinxQuickstartTemplate.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,160 @@
+==========================
+Sphinx Quickstart Template
+==========================
+
+Introduction and Quickstart
+===========================
+
+This document is meant to get you writing documentation as fast as possible
+even if you have no previous experience with Sphinx. The goal is to take
+someone in the state of "I want to write documentation and get it added to
+LLVM's docs" and turn that into useful documentation mailed to llvm-commits
+with as little nonsense as possible.
+
+You can find this document in ``docs/SphinxQuickstartTemplate.rst``. You
+should copy it, open the new file in your text editor, write your docs, and
+then send the new document to llvm-commits for review.
+
+Focus on *content*. It is easy to fix the Sphinx (reStructuredText) syntax
+later if necessary, although reStructuredText tries to imitate common
+plain-text conventions so it should be quite natural. A basic knowledge of
+reStructuredText syntax is useful when writing the document, so the last
+~half of this document (starting with `Example Section`_) gives examples
+which should cover 99% of use cases.
+
+Let me say that again: focus on *content*. But if you really need to verify
+Sphinx's output, see ``docs/README.txt`` for information.
+
+Once you have finished with the content, please send the ``.rst`` file to
+llvm-commits for review.
+
+Guidelines
+==========
+
+Try to answer the following questions in your first section:
+
+#. Why would I want to read this document?
+
+#. What should I know to be able to follow along with this document?
+
+#. What will I have learned by the end of this document?
+
+Common names for the first section are ``Introduction``, ``Overview``, or
+``Background``.
+
+If possible, make your document a "how to". Give it a name ``HowTo*.rst``
+like the other "how to" documents. This format is usually the easiest
+for another person to understand and also the most useful.
+
+You generally should not be writing documentation other than a "how to"
+unless there is already a "how to" about your topic. The reason for this
+is that without a "how to" document to read first, it is difficult for a
+person to understand a more advanced document.
+
+Focus on content (yes, I had to say it again).
+
+The rest of this document shows example reStructuredText markup constructs
+that are meant to be read by you in your text editor after you have copied
+this file into a new file for the documentation you are about to write.
+
+Example Section
+===============
+
+Your text can be *emphasized*, **bold**, or ``monospace``.
+
+Use blank lines to separate paragraphs.
+
+Headings (like ``Example Section`` just above) give your document its
+structure. Use the same kind of adornments (e.g. ``======`` vs. ``------``)
+as are used in this document. The adornment must be the same length as the
+text above it. For Vim users, variations of ``yypVr=`` might be handy.
+
+Example Subsection
+------------------
+
+Make a link `like this <http://llvm.org/>`_. There is also a more
+sophisticated syntax which `can be more readable`_ for longer links since
+it disrupts the flow less. You can put the ``.. _`link text`: <URL>`` block
+pretty much anywhere later in the document.
+
+.. _`can be more readable`: http://en.wikipedia.org/wiki/LLVM
+
+Lists can be made like this:
+
+#. A list starting with ``#.`` will be automatically numbered.
+
+#. This is a second list element.
+
+   #. Use indentation to create nested lists.
+
+You can also use unordered lists.
+
+* Stuff.
+
+  + Deeper stuff.
+
+* More stuff.
+
+Example Subsubsection
+^^^^^^^^^^^^^^^^^^^^^
+
+You can make blocks of code like this:
+
+.. code-block:: c++
+
+   int main() {
+     return 0;
+   }
+
+For a shell session, use a ``console`` code block (some existing docs use
+``bash``):
+
+.. code-block:: console
+
+   $ echo "Goodbye cruel world!"
+   $ rm -rf /
+
+If you need to show LLVM IR use the ``llvm`` code block.
+
+.. code-block:: llvm
+
+   define i32 @test1() {
+   entry:
+     ret i32 0
+   }
+
+Some other common code blocks you might need are ``c``, ``objc``, ``make``,
+and ``cmake``. If you need something beyond that, you can look at the `full
+list`_ of supported code blocks.
+
+.. _`full list`: http://pygments.org/docs/lexers/
+
+However, don't waste time fiddling with syntax highlighting when you could
+be adding meaningful content. When in doubt, show preformatted text
+without any syntax highlighting like this:
+
+::
+
+                          .
+                           +:.
+                       ..:: ::
+                    .++:+:: ::+:.:.
+                   .:+           :
+            ::.::..::            .+.
+          ..:+    ::              :
+    ......+:.                    ..
+          :++.    ..              :
+            .+:::+::              :
+            ..   . .+            ::
+                     +.:      .::+.
+                      ...+. .: .
+                         .++:..
+                          ...
+
+Hopefully you won't need to be this deep
+""""""""""""""""""""""""""""""""""""""""
+
+If you need to do fancier things than what has been shown in this document,
+you can mail the list or check Sphinx's `reStructuredText Primer`_.
+
+.. _`reStructuredText Primer`: http://sphinx.pocoo.org/rest.html

Added: www-releases/trunk/4.0.0/docs/_sources/StackMaps.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/StackMaps.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/StackMaps.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/StackMaps.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,511 @@
+===================================
+Stack maps and patch points in LLVM
+===================================
+
+.. contents::
+   :local:
+   :depth: 2
+
+Definitions
+===========
+
+In this document we refer to the "runtime" collectively as all
+components that serve as the LLVM client, including the LLVM IR
+generator, object code consumer, and code patcher.
+
+A stack map records the location of ``live values`` at a particular
+instruction address. These ``live values`` do not refer to all the
+LLVM values live across the stack map. Instead, they are only the
+values that the runtime requires to be live at this point. For
+example, they may be the values the runtime will need to resume
+program execution at that point independent of the compiled function
+containing the stack map.
+
+LLVM emits stack map data into the object code within a designated
+:ref:`stackmap-section`. This stack map data contains a record for
+each stack map. The record stores the stack map's instruction address
+and contains a entry for each mapped value. Each entry encodes a
+value's location as a register, stack offset, or constant.
+
+A patch point is an instruction address at which space is reserved for
+patching a new instruction sequence at run time. Patch points look
+much like calls to LLVM. They take arguments that follow a calling
+convention and may return a value. They also imply stack map
+generation, which allows the runtime to locate the patchpoint and
+find the location of ``live values`` at that point.
+
+Motivation
+==========
+
+This functionality is currently experimental but is potentially useful
+in a variety of settings, the most obvious being a runtime (JIT)
+compiler. Example applications of the patchpoint intrinsics are
+implementing an inline call cache for polymorphic method dispatch or
+optimizing the retrieval of properties in dynamically typed languages
+such as JavaScript.
+
+The intrinsics documented here are currently used by the JavaScript
+compiler within the open source WebKit project, see the `FTL JIT
+<https://trac.webkit.org/wiki/FTLJIT>`_, but they are designed to be
+used whenever stack maps or code patching are needed. Because the
+intrinsics have experimental status, compatibility across LLVM
+releases is not guaranteed.
+
+The stack map functionality described in this document is separate
+from the functionality described in
+:ref:`stack-map`. `GCFunctionMetadata` provides the location of
+pointers into a collected heap captured by the `GCRoot` intrinsic,
+which can also be considered a "stack map". Unlike the stack maps
+defined above, the `GCFunctionMetadata` stack map interface does not
+provide a way to associate live register values of arbitrary type with
+an instruction address, nor does it specify a format for the resulting
+stack map. The stack maps described here could potentially provide
+richer information to a garbage collecting runtime, but that usage
+will not be discussed in this document.
+
+Intrinsics
+==========
+
+The following two kinds of intrinsics can be used to implement stack
+maps and patch points: ``llvm.experimental.stackmap`` and
+``llvm.experimental.patchpoint``. Both kinds of intrinsics generate a
+stack map record, and they both allow some form of code patching. They
+can be used independently (i.e. ``llvm.experimental.patchpoint``
+implicitly generates a stack map without the need for an additional
+call to ``llvm.experimental.stackmap``). The choice of which to use
+depends on whether it is necessary to reserve space for code patching
+and whether any of the intrinsic arguments should be lowered according
+to calling conventions. ``llvm.experimental.stackmap`` does not
+reserve any space, nor does it expect any call arguments. If the
+runtime patches code at the stack map's address, it will destructively
+overwrite the program text. This is unlike
+``llvm.experimental.patchpoint``, which reserves space for in-place
+patching without overwriting surrounding code. The
+``llvm.experimental.patchpoint`` intrinsic also lowers a specified
+number of arguments according to its calling convention. This allows
+patched code to make in-place function calls without marshaling.
+
+Each instance of one of these intrinsics generates a stack map record
+in the :ref:`stackmap-section`. The record includes an ID, allowing
+the runtime to uniquely identify the stack map, and the offset within
+the code from the beginning of the enclosing function.
+
+'``llvm.experimental.stackmap``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare void
+        @llvm.experimental.stackmap(i64 <id>, i32 <numShadowBytes>, ...)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.stackmap``' intrinsic records the location of
+specified values in the stack map without generating any code.
+
+Operands:
+"""""""""
+
+The first operand is an ID to be encoded within the stack map. The
+second operand is the number of shadow bytes following the
+intrinsic. The variable number of operands that follow are the ``live
+values`` for which locations will be recorded in the stack map.
+
+To use this intrinsic as a bare-bones stack map, with no code patching
+support, the number of shadow bytes can be set to zero.
+
+Semantics:
+""""""""""
+
+The stack map intrinsic generates no code in place, unless nops are
+needed to cover its shadow (see below). However, its offset from
+function entry is stored in the stack map. This is the relative
+instruction address immediately following the instructions that
+precede the stack map.
+
+The stack map ID allows a runtime to locate the desired stack map
+record. LLVM passes this ID through directly to the stack map
+record without checking uniqueness.
+
+LLVM guarantees a shadow of instructions following the stack map's
+instruction offset during which neither the end of the basic block nor
+another call to ``llvm.experimental.stackmap`` or
+``llvm.experimental.patchpoint`` may occur. This allows the runtime to
+patch the code at this point in response to an event triggered from
+outside the code. The code for instructions following the stack map
+may be emitted in the stack map's shadow, and these instructions may
+be overwritten by destructive patching. Without shadow bytes, this
+destructive patching could overwrite program text or data outside the
+current function. We disallow overlapping stack map shadows so that
+the runtime does not need to consider this corner case.
+
+For example, a stack map with 8 byte shadow:
+
+.. code-block:: llvm
+
+  call void @runtime()
+  call void (i64, i32, ...)* @llvm.experimental.stackmap(i64 77, i32 8,
+                                                         i64* %ptr)
+  %val = load i64* %ptr
+  %add = add i64 %val, 3
+  ret i64 %add
+
+May require one byte of nop-padding:
+
+.. code-block:: none
+
+  0x00 callq _runtime
+  0x05 nop                <--- stack map address
+  0x06 movq (%rdi), %rax
+  0x07 addq $3, %rax
+  0x0a popq %rdx
+  0x0b ret                <---- end of 8-byte shadow
+
+Now, if the runtime needs to invalidate the compiled code, it may
+patch 8 bytes of code at the stack map's address at follows:
+
+.. code-block:: none
+
+  0x00 callq _runtime
+  0x05 movl  $0xffff, %rax <--- patched code at stack map address
+  0x0a callq *%rax         <---- end of 8-byte shadow
+
+This way, after the normal call to the runtime returns, the code will
+execute a patched call to a special entry point that can rebuild a
+stack frame from the values located by the stack map.
+
+'``llvm.experimental.patchpoint.*``' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare void
+        @llvm.experimental.patchpoint.void(i64 <id>, i32 <numBytes>,
+                                           i8* <target>, i32 <numArgs>, ...)
+      declare i64
+        @llvm.experimental.patchpoint.i64(i64 <id>, i32 <numBytes>,
+                                          i8* <target>, i32 <numArgs>, ...)
+
+Overview:
+"""""""""
+
+The '``llvm.experimental.patchpoint.*``' intrinsics creates a function
+call to the specified ``<target>`` and records the location of specified
+values in the stack map.
+
+Operands:
+"""""""""
+
+The first operand is an ID, the second operand is the number of bytes
+reserved for the patchable region, the third operand is the target
+address of a function (optionally null), and the fourth operand
+specifies how many of the following variable operands are considered
+function call arguments. The remaining variable number of operands are
+the ``live values`` for which locations will be recorded in the stack
+map.
+
+Semantics:
+""""""""""
+
+The patch point intrinsic generates a stack map. It also emits a
+function call to the address specified by ``<target>`` if the address
+is not a constant null. The function call and its arguments are
+lowered according to the calling convention specified at the
+intrinsic's callsite. Variants of the intrinsic with non-void return
+type also return a value according to calling convention.
+
+On PowerPC, note that ``<target>`` must be the ABI function pointer for the
+intended target of the indirect call. Specifically, when compiling for the
+ELF V1 ABI, ``<target>`` is the function-descriptor address normally used as
+the C/C++ function-pointer representation.
+
+Requesting zero patch point arguments is valid. In this case, all
+variable operands are handled just like
+``llvm.experimental.stackmap.*``. The difference is that space will
+still be reserved for patching, a call will be emitted, and a return
+value is allowed.
+
+The location of the arguments are not normally recorded in the stack
+map because they are already fixed by the calling convention. The
+remaining ``live values`` will have their location recorded, which
+could be a register, stack location, or constant. A special calling
+convention has been introduced for use with stack maps, anyregcc,
+which forces the arguments to be loaded into registers but allows
+those register to be dynamically allocated. These argument registers
+will have their register locations recorded in the stack map in
+addition to the remaining ``live values``.
+
+The patch point also emits nops to cover at least ``<numBytes>`` of
+instruction encoding space. Hence, the client must ensure that
+``<numBytes>`` is enough to encode a call to the target address on the
+supported targets. If the call target is constant null, then there is
+no minimum requirement. A zero-byte null target patchpoint is
+valid.
+
+The runtime may patch the code emitted for the patch point, including
+the call sequence and nops. However, the runtime may not assume
+anything about the code LLVM emits within the reserved space. Partial
+patching is not allowed. The runtime must patch all reserved bytes,
+padding with nops if necessary.
+
+This example shows a patch point reserving 15 bytes, with one argument
+in $rdi, and a return value in $rax per native calling convention:
+
+.. code-block:: llvm
+
+  %target = inttoptr i64 -281474976710654 to i8*
+  %val = call i64 (i64, i32, ...)*
+           @llvm.experimental.patchpoint.i64(i64 78, i32 15,
+                                             i8* %target, i32 1, i64* %ptr)
+  %add = add i64 %val, 3
+  ret i64 %add
+
+May generate:
+
+.. code-block:: none
+
+  0x00 movabsq $0xffff000000000002, %r11 <--- patch point address
+  0x0a callq   *%r11
+  0x0d nop
+  0x0e nop                               <--- end of reserved 15-bytes
+  0x0f addq    $0x3, %rax
+  0x10 movl    %rax, 8(%rsp)
+
+Note that no stack map locations will be recorded. If the patched code
+sequence does not need arguments fixed to specific calling convention
+registers, then the ``anyregcc`` convention may be used:
+
+.. code-block:: none
+
+  %val = call anyregcc @llvm.experimental.patchpoint(i64 78, i32 15,
+                                                     i8* %target, i32 1,
+                                                     i64* %ptr)
+
+The stack map now indicates the location of the %ptr argument and
+return value:
+
+.. code-block:: none
+
+  Stack Map: ID=78, Loc0=%r9 Loc1=%r8
+
+The patch code sequence may now use the argument that happened to be
+allocated in %r8 and return a value allocated in %r9:
+
+.. code-block:: none
+
+  0x00 movslq 4(%r8) %r9              <--- patched code at patch point address
+  0x03 nop
+  ...
+  0x0e nop                            <--- end of reserved 15-bytes
+  0x0f addq    $0x3, %r9
+  0x10 movl    %r9, 8(%rsp)
+
+.. _stackmap-format:
+
+Stack Map Format
+================
+
+The existence of a stack map or patch point intrinsic within an LLVM
+Module forces code emission to create a :ref:`stackmap-section`. The
+format of this section follows:
+
+.. code-block:: none
+
+  Header {
+    uint8  : Stack Map Version (current version is 2)
+    uint8  : Reserved (expected to be 0)
+    uint16 : Reserved (expected to be 0)
+  }
+  uint32 : NumFunctions
+  uint32 : NumConstants
+  uint32 : NumRecords
+  StkSizeRecord[NumFunctions] {
+    uint64 : Function Address
+    uint64 : Stack Size
+    uint64 : Record Count
+  }
+  Constants[NumConstants] {
+    uint64 : LargeConstant
+  }
+  StkMapRecord[NumRecords] {
+    uint64 : PatchPoint ID
+    uint32 : Instruction Offset
+    uint16 : Reserved (record flags)
+    uint16 : NumLocations
+    Location[NumLocations] {
+      uint8  : Register | Direct | Indirect | Constant | ConstantIndex
+      uint8  : Reserved (location flags)
+      uint16 : Dwarf RegNum
+      int32  : Offset or SmallConstant
+    }
+    uint16 : Padding
+    uint16 : NumLiveOuts
+    LiveOuts[NumLiveOuts]
+      uint16 : Dwarf RegNum
+      uint8  : Reserved
+      uint8  : Size in Bytes
+    }
+    uint32 : Padding (only if required to align to 8 byte)
+  }
+
+The first byte of each location encodes a type that indicates how to
+interpret the ``RegNum`` and ``Offset`` fields as follows:
+
+======== ========== =================== ===========================
+Encoding Type       Value               Description
+-------- ---------- ------------------- ---------------------------
+0x1      Register   Reg                 Value in a register
+0x2      Direct     Reg + Offset        Frame index value
+0x3      Indirect   [Reg + Offset]      Spilled value
+0x4      Constant   Offset              Small constant
+0x5      ConstIndex Constants[Offset]   Large constant
+======== ========== =================== ===========================
+
+In the common case, a value is available in a register, and the
+``Offset`` field will be zero. Values spilled to the stack are encoded
+as ``Indirect`` locations. The runtime must load those values from a
+stack address, typically in the form ``[BP + Offset]``. If an
+``alloca`` value is passed directly to a stack map intrinsic, then
+LLVM may fold the frame index into the stack map as an optimization to
+avoid allocating a register or stack slot. These frame indices will be
+encoded as ``Direct`` locations in the form ``BP + Offset``. LLVM may
+also optimize constants by emitting them directly in the stack map,
+either in the ``Offset`` of a ``Constant`` location or in the constant
+pool, referred to by ``ConstantIndex`` locations.
+
+At each callsite, a "liveout" register list is also recorded. These
+are the registers that are live across the stackmap and therefore must
+be saved by the runtime. This is an important optimization when the
+patchpoint intrinsic is used with a calling convention that by default
+preserves most registers as callee-save.
+
+Each entry in the liveout register list contains a DWARF register
+number and size in bytes. The stackmap format deliberately omits
+specific subregister information. Instead the runtime must interpret
+this information conservatively. For example, if the stackmap reports
+one byte at ``%rax``, then the value may be in either ``%al`` or
+``%ah``. It doesn't matter in practice, because the runtime will
+simply save ``%rax``. However, if the stackmap reports 16 bytes at
+``%ymm0``, then the runtime can safely optimize by saving only
+``%xmm0``.
+
+The stack map format is a contract between an LLVM SVN revision and
+the runtime. It is currently experimental and may change in the short
+term, but minimizing the need to update the runtime is
+important. Consequently, the stack map design is motivated by
+simplicity and extensibility. Compactness of the representation is
+secondary because the runtime is expected to parse the data
+immediately after compiling a module and encode the information in its
+own format. Since the runtime controls the allocation of sections, it
+can reuse the same stack map space for multiple modules.
+
+Stackmap support is currently only implemented for 64-bit
+platforms. However, a 32-bit implementation should be able to use the
+same format with an insignificant amount of wasted space.
+
+.. _stackmap-section:
+
+Stack Map Section
+^^^^^^^^^^^^^^^^^
+
+A JIT compiler can easily access this section by providing its own
+memory manager via the LLVM C API
+``LLVMCreateSimpleMCJITMemoryManager()``. When creating the memory
+manager, the JIT provides a callback:
+``LLVMMemoryManagerAllocateDataSectionCallback()``. When LLVM creates
+this section, it invokes the callback and passes the section name. The
+JIT can record the in-memory address of the section at this time and
+later parse it to recover the stack map data.
+
+On Darwin, the stack map section name is "__llvm_stackmaps". The
+segment name is "__LLVM_STACKMAPS".
+
+Stack Map Usage
+===============
+
+The stack map support described in this document can be used to
+precisely determine the location of values at a specific position in
+the code. LLVM does not maintain any mapping between those values and
+any higher-level entity. The runtime must be able to interpret the
+stack map record given only the ID, offset, and the order of the
+locations, records, and functions, which LLVM preserves.
+
+Note that this is quite different from the goal of debug information,
+which is a best-effort attempt to track the location of named
+variables at every instruction.
+
+An important motivation for this design is to allow a runtime to
+commandeer a stack frame when execution reaches an instruction address
+associated with a stack map. The runtime must be able to rebuild a
+stack frame and resume program execution using the information
+provided by the stack map. For example, execution may resume in an
+interpreter or a recompiled version of the same function.
+
+This usage restricts LLVM optimization. Clearly, LLVM must not move
+stores across a stack map. However, loads must also be handled
+conservatively. If the load may trigger an exception, hoisting it
+above a stack map could be invalid. For example, the runtime may
+determine that a load is safe to execute without a type check given
+the current state of the type system. If the type system changes while
+some activation of the load's function exists on the stack, the load
+becomes unsafe. The runtime can prevent subsequent execution of that
+load by immediately patching any stack map location that lies between
+the current call site and the load (typically, the runtime would
+simply patch all stack map locations to invalidate the function). If
+the compiler had hoisted the load above the stack map, then the
+program could crash before the runtime could take back control.
+
+To enforce these semantics, stackmap and patchpoint intrinsics are
+considered to potentially read and write all memory. This may limit
+optimization more than some clients desire. This limitation may be
+avoided by marking the call site as "readonly". In the future we may
+also allow meta-data to be added to the intrinsic call to express
+aliasing, thereby allowing optimizations to hoist certain loads above
+stack maps.
+
+Direct Stack Map Entries
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+As shown in :ref:`stackmap-section`, a Direct stack map location
+records the address of frame index. This address is itself the value
+that the runtime requested. This differs from Indirect locations,
+which refer to a stack locations from which the requested values must
+be loaded. Direct locations can communicate the address if an alloca,
+while Indirect locations handle register spills.
+
+For example:
+
+.. code-block:: none
+
+  entry:
+    %a = alloca i64...
+    llvm.experimental.stackmap(i64 <ID>, i32 <shadowBytes>, i64* %a)
+
+The runtime can determine this alloca's relative location on the
+stack immediately after compilation, or at any time thereafter. This
+differs from Register and Indirect locations, because the runtime can
+only read the values in those locations when execution reaches the
+instruction address of the stack map.
+
+This functionality requires LLVM to treat entry-block allocas
+specially when they are directly consumed by an intrinsics. (This is
+the same requirement imposed by the llvm.gcroot intrinsic.) LLVM
+transformations must not substitute the alloca with any intervening
+value. This can be verified by the runtime simply by checking that the
+stack map's location is a Direct location type.
+
+
+Supported Architectures
+=======================
+
+Support for StackMap generation and the related intrinsics requires 
+some code for each backend.  Today, only a subset of LLVM's backends 
+are supported.  The currently supported architectures are X86_64, 
+PowerPC, and Aarch64.

Added: www-releases/trunk/4.0.0/docs/_sources/Statepoints.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/Statepoints.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/Statepoints.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/Statepoints.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,841 @@
+=====================================
+Garbage Collection Safepoints in LLVM
+=====================================
+
+.. contents::
+   :local:
+   :depth: 2
+
+Status
+=======
+
+This document describes a set of experimental extensions to LLVM. Use
+with caution.  Because the intrinsics have experimental status,
+compatibility across LLVM releases is not guaranteed.
+
+LLVM currently supports an alternate mechanism for conservative
+garbage collection support using the ``gcroot`` intrinsic.  The mechanism
+described here shares little in common with the alternate ``gcroot``
+implementation and it is hoped that this mechanism will eventually
+replace the gc_root mechanism.
+
+Overview
+========
+
+To collect dead objects, garbage collectors must be able to identify
+any references to objects contained within executing code, and,
+depending on the collector, potentially update them.  The collector
+does not need this information at all points in code - that would make
+the problem much harder - but only at well-defined points in the
+execution known as 'safepoints' For most collectors, it is sufficient
+to track at least one copy of each unique pointer value.  However, for
+a collector which wishes to relocate objects directly reachable from
+running code, a higher standard is required.
+
+One additional challenge is that the compiler may compute intermediate
+results ("derived pointers") which point outside of the allocation or
+even into the middle of another allocation.  The eventual use of this
+intermediate value must yield an address within the bounds of the
+allocation, but such "exterior derived pointers" may be visible to the
+collector.  Given this, a garbage collector can not safely rely on the
+runtime value of an address to indicate the object it is associated
+with.  If the garbage collector wishes to move any object, the
+compiler must provide a mapping, for each pointer, to an indication of
+its allocation.
+
+To simplify the interaction between a collector and the compiled code,
+most garbage collectors are organized in terms of three abstractions:
+load barriers, store barriers, and safepoints.
+
+#. A load barrier is a bit of code executed immediately after the
+   machine load instruction, but before any use of the value loaded.
+   Depending on the collector, such a barrier may be needed for all
+   loads, merely loads of a particular type (in the original source
+   language), or none at all.
+
+#. Analogously, a store barrier is a code fragment that runs
+   immediately before the machine store instruction, but after the
+   computation of the value stored.  The most common use of a store
+   barrier is to update a 'card table' in a generational garbage
+   collector.
+
+#. A safepoint is a location at which pointers visible to the compiled
+   code (i.e. currently in registers or on the stack) are allowed to
+   change.  After the safepoint completes, the actual pointer value
+   may differ, but the 'object' (as seen by the source language)
+   pointed to will not.
+
+  Note that the term 'safepoint' is somewhat overloaded.  It refers to
+  both the location at which the machine state is parsable and the
+  coordination protocol involved in bring application threads to a
+  point at which the collector can safely use that information.  The
+  term "statepoint" as used in this document refers exclusively to the
+  former.
+
+This document focuses on the last item - compiler support for
+safepoints in generated code.  We will assume that an outside
+mechanism has decided where to place safepoints.  From our
+perspective, all safepoints will be function calls.  To support
+relocation of objects directly reachable from values in compiled code,
+the collector must be able to:
+
+#. identify every copy of a pointer (including copies introduced by
+   the compiler itself) at the safepoint,
+#. identify which object each pointer relates to, and
+#. potentially update each of those copies.
+
+This document describes the mechanism by which an LLVM based compiler
+can provide this information to a language runtime/collector, and
+ensure that all pointers can be read and updated if desired.  The
+heart of the approach is to construct (or rewrite) the IR in a manner
+where the possible updates performed by the garbage collector are
+explicitly visible in the IR.  Doing so requires that we:
+
+#. create a new SSA value for each potentially relocated pointer, and
+   ensure that no uses of the original (non relocated) value is
+   reachable after the safepoint,
+#. specify the relocation in a way which is opaque to the compiler to
+   ensure that the optimizer can not introduce new uses of an
+   unrelocated value after a statepoint. This prevents the optimizer
+   from performing unsound optimizations.
+#. recording a mapping of live pointers (and the allocation they're
+   associated with) for each statepoint.
+
+At the most abstract level, inserting a safepoint can be thought of as
+replacing a call instruction with a call to a multiple return value
+function which both calls the original target of the call, returns
+it's result, and returns updated values for any live pointers to
+garbage collected objects.
+
+  Note that the task of identifying all live pointers to garbage
+  collected values, transforming the IR to expose a pointer giving the
+  base object for every such live pointer, and inserting all the
+  intrinsics correctly is explicitly out of scope for this document.
+  The recommended approach is to use the :ref:`utility passes 
+  <statepoint-utilities>` described below. 
+
+This abstract function call is concretely represented by a sequence of
+intrinsic calls known collectively as a "statepoint relocation sequence".
+
+Let's consider a simple call in LLVM IR:
+
+.. code-block:: llvm
+
+  define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj) 
+         gc "statepoint-example" {
+    call void ()* @foo()
+    ret i8 addrspace(1)* %obj
+  }
+
+Depending on our language we may need to allow a safepoint during the execution 
+of ``foo``. If so, we need to let the collector update local values in the 
+current frame.  If we don't, we'll be accessing a potential invalid reference 
+once we eventually return from the call.
+
+In this example, we need to relocate the SSA value ``%obj``.  Since we can't 
+actually change the value in the SSA value ``%obj``, we need to introduce a new 
+SSA value ``%obj.relocated`` which represents the potentially changed value of
+``%obj`` after the safepoint and update any following uses appropriately.  The 
+resulting relocation sequence is:
+
+.. code-block:: text
+
+  define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj) 
+         gc "statepoint-example" {
+    %0 = call token (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0, i8 addrspace(1)* %obj)
+    %obj.relocated = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token %0, i32 7, i32 7)
+    ret i8 addrspace(1)* %obj.relocated
+  }
+
+Ideally, this sequence would have been represented as a M argument, N
+return value function (where M is the number of values being
+relocated + the original call arguments and N is the original return
+value + each relocated value), but LLVM does not easily support such a
+representation.
+
+Instead, the statepoint intrinsic marks the actual site of the
+safepoint or statepoint.  The statepoint returns a token value (which
+exists only at compile time).  To get back the original return value
+of the call, we use the ``gc.result`` intrinsic.  To get the relocation
+of each pointer in turn, we use the ``gc.relocate`` intrinsic with the
+appropriate index.  Note that both the ``gc.relocate`` and ``gc.result`` are
+tied to the statepoint.  The combination forms a "statepoint relocation 
+sequence" and represents the entirety of a parseable call or 'statepoint'.
+
+When lowered, this example would generate the following x86 assembly:
+
+.. code-block:: gas
+  
+	  .globl	test1
+	  .align	16, 0x90
+	  pushq	%rax
+	  callq	foo
+  .Ltmp1:
+	  movq	(%rsp), %rax  # This load is redundant (oops!)
+	  popq	%rdx
+	  retq
+
+Each of the potentially relocated values has been spilled to the
+stack, and a record of that location has been recorded to the
+:ref:`Stack Map section <stackmap-section>`.  If the garbage collector
+needs to update any of these pointers during the call, it knows
+exactly what to change.
+
+The relevant parts of the StackMap section for our example are:
+
+.. code-block:: gas
+  
+  # This describes the call site
+  # Stack Maps: callsite 2882400000
+	  .quad	2882400000
+	  .long	.Ltmp1-test1
+	  .short	0
+  # .. 8 entries skipped ..
+  # This entry describes the spill slot which is directly addressable
+  # off RSP with offset 0.  Given the value was spilled with a pushq, 
+  # that makes sense.
+  # Stack Maps:   Loc 8: Direct RSP     [encoding: .byte 2, .byte 8, .short 7, .int 0]
+	  .byte	2
+	  .byte	8
+	  .short	7
+	  .long	0
+
+This example was taken from the tests for the :ref:`RewriteStatepointsForGC` utility pass.  As such, it's full StackMap can be easily examined with the following command.
+
+.. code-block:: bash
+
+  opt -rewrite-statepoints-for-gc test/Transforms/RewriteStatepointsForGC/basics.ll -S | llc -debug-only=stackmaps
+
+Base & Derived Pointers
+^^^^^^^^^^^^^^^^^^^^^^^
+
+A "base pointer" is one which points to the starting address of an allocation
+(object).  A "derived pointer" is one which is offset from a base pointer by
+some amount.  When relocating objects, a garbage collector needs to be able 
+to relocate each derived pointer associated with an allocation to the same 
+offset from the new address.
+
+"Interior derived pointers" remain within the bounds of the allocation 
+they're associated with.  As a result, the base object can be found at 
+runtime provided the bounds of allocations are known to the runtime system.
+
+"Exterior derived pointers" are outside the bounds of the associated object;
+they may even fall within *another* allocations address range.  As a result,
+there is no way for a garbage collector to determine which allocation they 
+are associated with at runtime and compiler support is needed.
+
+The ``gc.relocate`` intrinsic supports an explicit operand for describing the
+allocation associated with a derived pointer.  This operand is frequently 
+referred to as the base operand, but does not strictly speaking have to be
+a base pointer, but it does need to lie within the bounds of the associated
+allocation.  Some collectors may require that the operand be an actual base
+pointer rather than merely an internal derived pointer. Note that during 
+lowering both the base and derived pointer operands are required to be live 
+over the associated call safepoint even if the base is otherwise unused 
+afterwards.
+
+If we extend our previous example to include a pointless derived pointer, 
+we get:
+
+.. code-block:: text
+
+  define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj) 
+         gc "statepoint-example" {
+    %gep = getelementptr i8, i8 addrspace(1)* %obj, i64 20000
+    %token = call token (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0, i8 addrspace(1)* %obj, i8 addrspace(1)* %gep)
+    %obj.relocated = call i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token %token, i32 7, i32 7)
+    %gep.relocated = call i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token %token, i32 7, i32 8)
+    %p = getelementptr i8, i8 addrspace(1)* %gep, i64 -20000
+    ret i8 addrspace(1)* %p
+  }
+
+Note that in this example %p and %obj.relocate are the same address and we
+could replace one with the other, potentially removing the derived pointer
+from the live set at the safepoint entirely.
+
+.. _gc_transition_args:
+
+GC Transitions
+^^^^^^^^^^^^^^^^^^
+
+As a practical consideration, many garbage-collected systems allow code that is
+collector-aware ("managed code") to call code that is not collector-aware
+("unmanaged code"). It is common that such calls must also be safepoints, since
+it is desirable to allow the collector to run during the execution of
+unmanaged code. Furthermore, it is common that coordinating the transition from
+managed to unmanaged code requires extra code generation at the call site to
+inform the collector of the transition. In order to support these needs, a
+statepoint may be marked as a GC transition, and data that is necessary to
+perform the transition (if any) may be provided as additional arguments to the
+statepoint.
+
+  Note that although in many cases statepoints may be inferred to be GC
+  transitions based on the function symbols involved (e.g. a call from a
+  function with GC strategy "foo" to a function with GC strategy "bar"),
+  indirect calls that are also GC transitions must also be supported. This
+  requirement is the driving force behind the decision to require that GC
+  transitions are explicitly marked.
+
+Let's revisit the sample given above, this time treating the call to ``@foo``
+as a GC transition. Depending on our target, the transition code may need to
+access some extra state in order to inform the collector of the transition.
+Let's assume a hypothetical GC--somewhat unimaginatively named "hypothetical-gc"
+--that requires that a TLS variable must be written to before and after a call
+to unmanaged code. The resulting relocation sequence is:
+
+.. code-block:: text
+
+  @flag = thread_local global i32 0, align 4
+
+  define i8 addrspace(1)* @test1(i8 addrspace(1) *%obj)
+         gc "hypothetical-gc" {
+
+    %0 = call token (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 0, i32 0, void ()* @foo, i32 0, i32 1, i32* @Flag, i32 0, i8 addrspace(1)* %obj)
+    %obj.relocated = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token %0, i32 7, i32 7)
+    ret i8 addrspace(1)* %obj.relocated
+  }
+
+During lowering, this will result in a instruction selection DAG that looks
+something like:
+
+::
+
+  CALLSEQ_START
+  ...
+  GC_TRANSITION_START (lowered i32 *@Flag), SRCVALUE i32* Flag
+  STATEPOINT
+  GC_TRANSITION_END (lowered i32 *@Flag), SRCVALUE i32 *Flag
+  ...
+  CALLSEQ_END
+
+In order to generate the necessary transition code, the backend for each target
+supported by "hypothetical-gc" must be modified to lower ``GC_TRANSITION_START``
+and ``GC_TRANSITION_END`` nodes appropriately when the "hypothetical-gc"
+strategy is in use for a particular function. Assuming that such lowering has
+been added for X86, the generated assembly would be:
+
+.. code-block:: gas
+
+	  .globl	test1
+	  .align	16, 0x90
+	  pushq	%rax
+	  movl $1, %fs:Flag at TPOFF
+	  callq	foo
+	  movl $0, %fs:Flag at TPOFF
+  .Ltmp1:
+	  movq	(%rsp), %rax  # This load is redundant (oops!)
+	  popq	%rdx
+	  retq
+
+Note that the design as presented above is not fully implemented: in particular,
+strategy-specific lowering is not present, and all GC transitions are emitted as
+as single no-op before and after the call instruction. These no-ops are often
+removed by the backend during dead machine instruction elimination.
+
+
+Intrinsics
+===========
+
+'llvm.experimental.gc.statepoint' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare token
+        @llvm.experimental.gc.statepoint(i64 <id>, i32 <num patch bytes>,
+                       func_type <target>, 
+                       i64 <#call args>, i64 <flags>,
+                       ... (call parameters),
+                       i64 <# transition args>, ... (transition parameters),
+                       i64 <# deopt args>, ... (deopt parameters),
+                       ... (gc parameters))
+
+Overview:
+"""""""""
+
+The statepoint intrinsic represents a call which is parse-able by the
+runtime.
+
+Operands:
+"""""""""
+
+The 'id' operand is a constant integer that is reported as the ID
+field in the generated stackmap.  LLVM does not interpret this
+parameter in any way and its meaning is up to the statepoint user to
+decide.  Note that LLVM is free to duplicate code containing
+statepoint calls, and this may transform IR that had a unique 'id' per
+lexical call to statepoint to IR that does not.
+
+If 'num patch bytes' is non-zero then the call instruction
+corresponding to the statepoint is not emitted and LLVM emits 'num
+patch bytes' bytes of nops in its place.  LLVM will emit code to
+prepare the function arguments and retrieve the function return value
+in accordance to the calling convention; the former before the nop
+sequence and the latter after the nop sequence.  It is expected that
+the user will patch over the 'num patch bytes' bytes of nops with a
+calling sequence specific to their runtime before executing the
+generated machine code.  There are no guarantees with respect to the
+alignment of the nop sequence.  Unlike :doc:`StackMaps` statepoints do
+not have a concept of shadow bytes.  Note that semantically the
+statepoint still represents a call or invoke to 'target', and the nop
+sequence after patching is expected to represent an operation
+equivalent to a call or invoke to 'target'.
+
+The 'target' operand is the function actually being called.  The
+target can be specified as either a symbolic LLVM function, or as an
+arbitrary Value of appropriate function type.  Note that the function
+type must match the signature of the callee and the types of the 'call
+parameters' arguments.
+
+The '#call args' operand is the number of arguments to the actual
+call.  It must exactly match the number of arguments passed in the
+'call parameters' variable length section.
+
+The 'flags' operand is used to specify extra information about the
+statepoint. This is currently only used to mark certain statepoints
+as GC transitions. This operand is a 64-bit integer with the following
+layout, where bit 0 is the least significant bit:
+
+  +-------+---------------------------------------------------+
+  | Bit # | Usage                                             |
+  +=======+===================================================+
+  |     0 | Set if the statepoint is a GC transition, cleared |
+  |       | otherwise.                                        |
+  +-------+---------------------------------------------------+
+  |  1-63 | Reserved for future use; must be cleared.         |
+  +-------+---------------------------------------------------+
+
+The 'call parameters' arguments are simply the arguments which need to
+be passed to the call target.  They will be lowered according to the
+specified calling convention and otherwise handled like a normal call
+instruction.  The number of arguments must exactly match what is
+specified in '# call args'.  The types must match the signature of
+'target'.
+
+The 'transition parameters' arguments contain an arbitrary list of
+Values which need to be passed to GC transition code. They will be
+lowered and passed as operands to the appropriate GC_TRANSITION nodes
+in the selection DAG. It is assumed that these arguments must be
+available before and after (but not necessarily during) the execution
+of the callee. The '# transition args' field indicates how many operands
+are to be interpreted as 'transition parameters'.
+
+The 'deopt parameters' arguments contain an arbitrary list of Values
+which is meaningful to the runtime.  The runtime may read any of these
+values, but is assumed not to modify them.  If the garbage collector
+might need to modify one of these values, it must also be listed in
+the 'gc pointer' argument list.  The '# deopt args' field indicates
+how many operands are to be interpreted as 'deopt parameters'.
+
+The 'gc parameters' arguments contain every pointer to a garbage
+collector object which potentially needs to be updated by the garbage
+collector.  Note that the argument list must explicitly contain a base
+pointer for every derived pointer listed.  The order of arguments is
+unimportant.  Unlike the other variable length parameter sets, this
+list is not length prefixed.
+
+Semantics:
+""""""""""
+
+A statepoint is assumed to read and write all memory.  As a result,
+memory operations can not be reordered past a statepoint.  It is
+illegal to mark a statepoint as being either 'readonly' or 'readnone'.
+
+Note that legal IR can not perform any memory operation on a 'gc
+pointer' argument of the statepoint in a location statically reachable
+from the statepoint.  Instead, the explicitly relocated value (from a
+``gc.relocate``) must be used.
+
+'llvm.experimental.gc.result' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare type*
+        @llvm.experimental.gc.result(token %statepoint_token)
+
+Overview:
+"""""""""
+
+``gc.result`` extracts the result of the original call instruction
+which was replaced by the ``gc.statepoint``.  The ``gc.result``
+intrinsic is actually a family of three intrinsics due to an
+implementation limitation.  Other than the type of the return value,
+the semantics are the same.
+
+Operands:
+"""""""""
+
+The first and only argument is the ``gc.statepoint`` which starts
+the safepoint sequence of which this ``gc.result`` is a part.
+Despite the typing of this as a generic token, *only* the value defined 
+by a ``gc.statepoint`` is legal here.
+
+Semantics:
+""""""""""
+
+The ``gc.result`` represents the return value of the call target of
+the ``statepoint``.  The type of the ``gc.result`` must exactly match
+the type of the target.  If the call target returns void, there will
+be no ``gc.result``.
+
+A ``gc.result`` is modeled as a 'readnone' pure function.  It has no
+side effects since it is just a projection of the return value of the
+previous call represented by the ``gc.statepoint``.
+
+'llvm.experimental.gc.relocate' Intrinsic
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Syntax:
+"""""""
+
+::
+
+      declare <pointer type>
+        @llvm.experimental.gc.relocate(token %statepoint_token, 
+                                       i32 %base_offset, 
+                                       i32 %pointer_offset)
+
+Overview:
+"""""""""
+
+A ``gc.relocate`` returns the potentially relocated value of a pointer
+at the safepoint.
+
+Operands:
+"""""""""
+
+The first argument is the ``gc.statepoint`` which starts the
+safepoint sequence of which this ``gc.relocation`` is a part.
+Despite the typing of this as a generic token, *only* the value defined 
+by a ``gc.statepoint`` is legal here.
+
+The second argument is an index into the statepoints list of arguments
+which specifies the allocation for the pointer being relocated.
+This index must land within the 'gc parameter' section of the
+statepoint's argument list.  The associated value must be within the
+object with which the pointer being relocated is associated. The optimizer
+is free to change *which* interior derived pointer is reported, provided that
+it does not replace an actual base pointer with another interior derived 
+pointer.  Collectors are allowed to rely on the base pointer operand 
+remaining an actual base pointer if so constructed.
+
+The third argument is an index into the statepoint's list of arguments
+which specify the (potentially) derived pointer being relocated.  It
+is legal for this index to be the same as the second argument
+if-and-only-if a base pointer is being relocated. This index must land
+within the 'gc parameter' section of the statepoint's argument list.
+
+Semantics:
+""""""""""
+
+The return value of ``gc.relocate`` is the potentially relocated value
+of the pointer specified by it's arguments.  It is unspecified how the
+value of the returned pointer relates to the argument to the
+``gc.statepoint`` other than that a) it points to the same source
+language object with the same offset, and b) the 'based-on'
+relationship of the newly relocated pointers is a projection of the
+unrelocated pointers.  In particular, the integer value of the pointer
+returned is unspecified.
+
+A ``gc.relocate`` is modeled as a ``readnone`` pure function.  It has no
+side effects since it is just a way to extract information about work
+done during the actual call modeled by the ``gc.statepoint``.
+
+.. _statepoint-stackmap-format:
+
+Stack Map Format
+================
+
+Locations for each pointer value which may need read and/or updated by
+the runtime or collector are provided via the :ref:`Stack Map format
+<stackmap-format>` specified in the PatchPoint documentation.
+
+Each statepoint generates the following Locations:
+
+* Constant which describes the calling convention of the call target. This
+  constant is a valid :ref:`calling convention identifier <callingconv>` for
+  the version of LLVM used to generate the stackmap. No additional compatibility
+  guarantees are made for this constant over what LLVM provides elsewhere w.r.t.
+  these identifiers.
+* Constant which describes the flags passed to the statepoint intrinsic
+* Constant which describes number of following deopt *Locations* (not
+  operands)
+* Variable number of Locations, one for each deopt parameter listed in
+  the IR statepoint (same number as described by previous Constant).  At 
+  the moment, only deopt parameters with a bitwidth of 64 bits or less 
+  are supported.  Values of a type larger than 64 bits can be specified 
+  and reported only if a) the value is constant at the call site, and b) 
+  the constant can be represented with less than 64 bits (assuming zero 
+  extension to the original bitwidth).
+* Variable number of relocation records, each of which consists of 
+  exactly two Locations.  Relocation records are described in detail
+  below.
+
+Each relocation record provides sufficient information for a collector to 
+relocate one or more derived pointers.  Each record consists of a pair of 
+Locations.  The second element in the record represents the pointer (or 
+pointers) which need updated.  The first element in the record provides a 
+pointer to the base of the object with which the pointer(s) being relocated is
+associated.  This information is required for handling generalized derived 
+pointers since a pointer may be outside the bounds of the original allocation,
+but still needs to be relocated with the allocation.  Additionally:
+
+* It is guaranteed that the base pointer must also appear explicitly as a 
+  relocation pair if used after the statepoint. 
+* There may be fewer relocation records then gc parameters in the IR
+  statepoint. Each *unique* pair will occur at least once; duplicates
+  are possible.  
+* The Locations within each record may either be of pointer size or a 
+  multiple of pointer size.  In the later case, the record must be 
+  interpreted as describing a sequence of pointers and their corresponding 
+  base pointers. If the Location is of size N x sizeof(pointer), then
+  there will be N records of one pointer each contained within the Location.
+  Both Locations in a pair can be assumed to be of the same size.
+
+Note that the Locations used in each section may describe the same
+physical location.  e.g. A stack slot may appear as a deopt location,
+a gc base pointer, and a gc derived pointer.
+
+The LiveOut section of the StkMapRecord will be empty for a statepoint
+record.
+
+Safepoint Semantics & Verification
+==================================
+
+The fundamental correctness property for the compiled code's
+correctness w.r.t. the garbage collector is a dynamic one.  It must be
+the case that there is no dynamic trace such that a operation
+involving a potentially relocated pointer is observably-after a
+safepoint which could relocate it.  'observably-after' is this usage
+means that an outside observer could observe this sequence of events
+in a way which precludes the operation being performed before the
+safepoint.
+
+To understand why this 'observable-after' property is required,
+consider a null comparison performed on the original copy of a
+relocated pointer.  Assuming that control flow follows the safepoint,
+there is no way to observe externally whether the null comparison is
+performed before or after the safepoint.  (Remember, the original
+Value is unmodified by the safepoint.)  The compiler is free to make
+either scheduling choice.
+
+The actual correctness property implemented is slightly stronger than
+this.  We require that there be no *static path* on which a
+potentially relocated pointer is 'observably-after' it may have been
+relocated.  This is slightly stronger than is strictly necessary (and
+thus may disallow some otherwise valid programs), but greatly
+simplifies reasoning about correctness of the compiled code.
+
+By construction, this property will be upheld by the optimizer if
+correctly established in the source IR.  This is a key invariant of
+the design.
+
+The existing IR Verifier pass has been extended to check most of the
+local restrictions on the intrinsics mentioned in their respective
+documentation.  The current implementation in LLVM does not check the
+key relocation invariant, but this is ongoing work on developing such
+a verifier.  Please ask on llvm-dev if you're interested in
+experimenting with the current version.
+
+.. _statepoint-utilities:
+
+Utility Passes for Safepoint Insertion
+======================================
+
+.. _RewriteStatepointsForGC:
+
+RewriteStatepointsForGC
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+The pass RewriteStatepointsForGC transforms a functions IR by replacing a 
+``gc.statepoint`` (with an optional ``gc.result``) with a full relocation 
+sequence, including all required ``gc.relocates``.  To function, the pass 
+requires that the GC strategy specified for the function be able to reliably 
+distinguish between GC references and non-GC references in IR it is given.
+
+As an example, given this code:
+
+.. code-block:: text
+
+  define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj) 
+         gc "statepoint-example" {
+    call token (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 5, i32 0, i32 -1, i32 0, i32 0, i32 0)
+    ret i8 addrspace(1)* %obj
+  }
+
+The pass would produce this IR:
+
+.. code-block:: text
+
+  define i8 addrspace(1)* @test1(i8 addrspace(1)* %obj) 
+         gc "statepoint-example" {
+    %0 = call token (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 5, i32 0, i32 -1, i32 0, i32 0, i32 0, i8 addrspace(1)* %obj)
+    %obj.relocated = call coldcc i8 addrspace(1)* @llvm.experimental.gc.relocate.p1i8(token %0, i32 12, i32 12)
+    ret i8 addrspace(1)* %obj.relocated
+  }
+
+In the above examples, the addrspace(1) marker on the pointers is the mechanism
+that the ``statepoint-example`` GC strategy uses to distinguish references from
+non references.  Address space 1 is not globally reserved for this purpose.
+
+This pass can be used an utility function by a language frontend that doesn't 
+want to manually reason about liveness, base pointers, or relocation when 
+constructing IR.  As currently implemented, RewriteStatepointsForGC must be 
+run after SSA construction (i.e. mem2ref).
+
+RewriteStatepointsForGC will ensure that appropriate base pointers are listed
+for every relocation created.  It will do so by duplicating code as needed to
+propagate the base pointer associated with each pointer being relocated to
+the appropriate safepoints.  The implementation assumes that the following 
+IR constructs produce base pointers: loads from the heap, addresses of global 
+variables, function arguments, function return values. Constant pointers (such
+as null) are also assumed to be base pointers.  In practice, this constraint
+can be relaxed to producing interior derived pointers provided the target 
+collector can find the associated allocation from an arbitrary interior 
+derived pointer.
+
+In practice, RewriteStatepointsForGC can be run much later in the pass 
+pipeline, after most optimization is already done.  This helps to improve 
+the quality of the generated code when compiled with garbage collection support.
+In the long run, this is the intended usage model.  At this time, a few details
+have yet to be worked out about the semantic model required to guarantee this 
+is always correct.  As such, please use with caution and report bugs.
+
+.. _PlaceSafepoints:
+
+PlaceSafepoints
+^^^^^^^^^^^^^^^^
+
+The pass PlaceSafepoints transforms a function's IR by replacing any call or 
+invoke instructions with appropriate ``gc.statepoint`` and ``gc.result`` pairs,
+and inserting safepoint polls sufficient to ensure running code checks for a 
+safepoint request on a timely manner.  This pass is expected to be run before 
+RewriteStatepointsForGC and thus does not produce full relocation sequences.  
+
+As an example, given input IR of the following:
+
+.. code-block:: llvm
+
+  define void @test() gc "statepoint-example" {
+    call void @foo()
+    ret void
+  }
+
+  declare void @do_safepoint()
+  define void @gc.safepoint_poll() {
+    call void @do_safepoint()
+    ret void
+  }
+
+
+This pass would produce the following IR:
+
+.. code-block:: text
+
+  define void @test() gc "statepoint-example" {
+    %safepoint_token = call token (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @do_safepoint, i32 0, i32 0, i32 0, i32 0)
+    %safepoint_token1 = call token (i64, i32, void ()*, i32, i32, ...)* @llvm.experimental.gc.statepoint.p0f_isVoidf(i64 2882400000, i32 0, void ()* @foo, i32 0, i32 0, i32 0, i32 0)
+    ret void
+  }
+
+In this case, we've added an (unconditional) entry safepoint poll and converted the call into a ``gc.statepoint``.  Note that despite appearances, the entry poll is not necessarily redundant.  We'd have to know that ``foo`` and ``test`` were not mutually recursive for the poll to be redundant.  In practice, you'd probably want to your poll definition to contain a conditional branch of some form.
+
+
+At the moment, PlaceSafepoints can insert safepoint polls at method entry and 
+loop backedges locations.  Extending this to work with return polls would be 
+straight forward if desired.
+
+PlaceSafepoints includes a number of optimizations to avoid placing safepoint 
+polls at particular sites unless needed to ensure timely execution of a poll 
+under normal conditions.  PlaceSafepoints does not attempt to ensure timely 
+execution of a poll under worst case conditions such as heavy system paging.
+
+The implementation of a safepoint poll action is specified by looking up a 
+function of the name ``gc.safepoint_poll`` in the containing Module.  The body
+of this function is inserted at each poll site desired.  While calls or invokes
+inside this method are transformed to a ``gc.statepoints``, recursive poll 
+insertion is not performed.
+
+By default PlaceSafepoints passes in ``0xABCDEF00`` as the statepoint
+ID and ``0`` as the number of patchable bytes to the newly constructed
+``gc.statepoint``.  These values can be configured on a per-callsite
+basis using the attributes ``"statepoint-id"`` and
+``"statepoint-num-patch-bytes"``.  If a call site is marked with a
+``"statepoint-id"`` function attribute and its value is a positive
+integer (represented as a string), then that value is used as the ID
+of the newly constructed ``gc.statepoint``.  If a call site is marked
+with a ``"statepoint-num-patch-bytes"`` function attribute and its
+value is a positive integer, then that value is used as the 'num patch
+bytes' parameter of the newly constructed ``gc.statepoint``.  The
+``"statepoint-id"`` and ``"statepoint-num-patch-bytes"`` attributes
+are not propagated to the ``gc.statepoint`` call or invoke if they
+could be successfully parsed.
+
+If you are scheduling the RewriteStatepointsForGC pass late in the pass order,
+you should probably schedule this pass immediately before it.  The exception 
+would be if you need to preserve abstract frame information (e.g. for
+deoptimization or introspection) at safepoints.  In that case, ask on the 
+llvm-dev mailing list for suggestions.
+
+
+Supported Architectures
+=======================
+
+Support for statepoint generation requires some code for each backend.
+Today, only X86_64 is supported.  
+
+Problem Areas and Active Work
+=============================
+
+#. As the existing users of the late rewriting model have matured, we've found
+   cases where the optimizer breaks the assumption that an SSA value of
+   gc-pointer type actually contains a gc-pointer and vice-versa.  We need to
+   clarify our expectations and propose at least one small IR change.  (Today,
+   the gc-pointer distinction is managed via address spaces.  This turns out
+   not to be quite strong enough.)
+
+#. Support for languages which allow unmanaged pointers to garbage collected
+   objects (i.e. pass a pointer to an object to a C routine) via pinning.
+
+#. Support for garbage collected objects allocated on the stack.  Specifically,
+   allocas are always assumed to be in address space 0 and we need a
+   cast/promotion operator to let rewriting identify them.
+
+#. The current statepoint lowering is known to be somewhat poor.  In the very
+   long term, we'd like to integrate statepoints with the register allocator;
+   in the near term this is unlikely to happen.  We've found the quality of
+   lowering to be relatively unimportant as hot-statepoints are almost always
+   inliner bugs.
+
+#. Concerns have been raised that the statepoint representation results in a
+   large amount of IR being produced for some examples and that this
+   contributes to higher than expected memory usage and compile times.  There's
+   no immediate plans to make changes due to this, but alternate models may be
+   explored in the future.
+
+#. Relocations along exceptional paths are currently broken in ToT.  In
+   particular, there is current no way to represent a rethrow on a path which
+   also has relocations.  See `this llvm-dev discussion
+   <https://groups.google.com/forum/#!topic/llvm-dev/AE417XjgxvI>`_ for more
+   detail.
+
+Bugs and Enhancements
+=====================
+
+Currently known bugs and enhancements under consideration can be
+tracked by performing a `bugzilla search
+<http://llvm.org/bugs/buglist.cgi?cmdtype=runnamed&namedcmd=Statepoint%20Bugs&list_id=64342>`_
+for [Statepoint] in the summary field. When filing new bugs, please
+use this tag so that interested parties see the newly filed bug.  As
+with most LLVM features, design discussions take place on `llvm-dev
+<http://lists.llvm.org/mailman/listinfo/llvm-dev>`_, and patches
+should be sent to `llvm-commits
+<http://lists.llvm.org/mailman/listinfo/llvm-commits>`_ for review.
+

Added: www-releases/trunk/4.0.0/docs/_sources/SystemLibrary.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/SystemLibrary.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/SystemLibrary.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/SystemLibrary.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,247 @@
+==============
+System Library
+==============
+
+Abstract
+========
+
+This document provides some details on LLVM's System Library, located in the
+source at ``lib/System`` and ``include/llvm/System``. The library's purpose is
+to shield LLVM from the differences between operating systems for the few
+services LLVM needs from the operating system. Much of LLVM is written using
+portability features of standard C++. However, in a few areas, system dependent
+facilities are needed and the System Library is the wrapper around those system
+calls.
+
+By centralizing LLVM's use of operating system interfaces, we make it possible
+for the LLVM tool chain and runtime libraries to be more easily ported to new
+platforms since (theoretically) only ``lib/System`` needs to be ported.  This
+library also unclutters the rest of LLVM from #ifdef use and special cases for
+specific operating systems. Such uses are replaced with simple calls to the
+interfaces provided in ``include/llvm/System``.
+
+Note that the System Library is not intended to be a complete operating system
+wrapper (such as the Adaptive Communications Environment (ACE) or Apache
+Portable Runtime (APR)), but only provides the functionality necessary to
+support LLVM.
+
+The System Library was written by Reid Spencer who formulated the design based
+on similar work originating from the eXtensible Programming System (XPS).
+Several people helped with the effort; especially, Jeff Cohen and Henrik Bach
+on the Win32 port.
+
+Keeping LLVM Portable
+=====================
+
+In order to keep LLVM portable, LLVM developers should adhere to a set of
+portability rules associated with the System Library. Adherence to these rules
+should help the System Library achieve its goal of shielding LLVM from the
+variations in operating system interfaces and doing so efficiently.  The
+following sections define the rules needed to fulfill this objective.
+
+Don't Include System Headers
+----------------------------
+
+Except in ``lib/System``, no LLVM source code should directly ``#include`` a
+system header. Care has been taken to remove all such ``#includes`` from LLVM
+while ``lib/System`` was being developed.  Specifically this means that header
+files like "``unistd.h``", "``windows.h``", "``stdio.h``", and "``string.h``"
+are forbidden to be included by LLVM source code outside the implementation of
+``lib/System``.
+
+To obtain system-dependent functionality, existing interfaces to the system
+found in ``include/llvm/System`` should be used. If an appropriate interface is
+not available, it should be added to ``include/llvm/System`` and implemented in
+``lib/System`` for all supported platforms.
+
+Don't Expose System Headers
+---------------------------
+
+The System Library must shield LLVM from **all** system headers. To obtain
+system level functionality, LLVM source must ``#include "llvm/System/Thing.h"``
+and nothing else. This means that ``Thing.h`` cannot expose any system header
+files. This protects LLVM from accidentally using system specific functionality
+and only allows it via the ``lib/System`` interface.
+
+Use Standard C Headers
+----------------------
+
+The **standard** C headers (the ones beginning with "c") are allowed to be
+exposed through the ``lib/System`` interface. These headers and the things they
+declare are considered to be platform agnostic. LLVM source files may include
+them directly or obtain their inclusion through ``lib/System`` interfaces.
+
+Use Standard C++ Headers
+------------------------
+
+The **standard** C++ headers from the standard C++ library and standard
+template library may be exposed through the ``lib/System`` interface. These
+headers and the things they declare are considered to be platform agnostic.
+LLVM source files may include them or obtain their inclusion through
+``lib/System`` interfaces.
+
+High Level Interface
+--------------------
+
+The entry points specified in the interface of ``lib/System`` must be aimed at
+completing some reasonably high level task needed by LLVM. We do not want to
+simply wrap each operating system call. It would be preferable to wrap several
+operating system calls that are always used in conjunction with one another by
+LLVM.
+
+For example, consider what is needed to execute a program, wait for it to
+complete, and return its result code. On Unix, this involves the following
+operating system calls: ``getenv``, ``fork``, ``execve``, and ``wait``. The
+correct thing for ``lib/System`` to provide is a function, say
+``ExecuteProgramAndWait``, that implements the functionality completely.  what
+we don't want is wrappers for the operating system calls involved.
+
+There must **not** be a one-to-one relationship between operating system
+calls and the System library's interface. Any such interface function will be
+suspicious.
+
+No Unused Functionality
+-----------------------
+
+There must be no functionality specified in the interface of ``lib/System``
+that isn't actually used by LLVM. We're not writing a general purpose operating
+system wrapper here, just enough to satisfy LLVM's needs. And, LLVM doesn't
+need much. This design goal aims to keep the ``lib/System`` interface small and
+understandable which should foster its actual use and adoption.
+
+No Duplicate Implementations
+----------------------------
+
+The implementation of a function for a given platform must be written exactly
+once. This implies that it must be possible to apply a function's
+implementation to multiple operating systems if those operating systems can
+share the same implementation. This rule applies to the set of operating
+systems supported for a given class of operating system (e.g. Unix, Win32).
+
+No Virtual Methods
+------------------
+
+The System Library interfaces can be called quite frequently by LLVM. In order
+to make those calls as efficient as possible, we discourage the use of virtual
+methods. There is no need to use inheritance for implementation differences, it
+just adds complexity. The ``#include`` mechanism works just fine.
+
+No Exposed Functions
+--------------------
+
+Any functions defined by system libraries (i.e. not defined by ``lib/System``)
+must not be exposed through the ``lib/System`` interface, even if the header
+file for that function is not exposed. This prevents inadvertent use of system
+specific functionality.
+
+For example, the ``stat`` system call is notorious for having variations in the
+data it provides. ``lib/System`` must not declare ``stat`` nor allow it to be
+declared. Instead it should provide its own interface to discovering
+information about files and directories. Those interfaces may be implemented in
+terms of ``stat`` but that is strictly an implementation detail. The interface
+provided by the System Library must be implemented on all platforms (even those
+without ``stat``).
+
+No Exposed Data
+---------------
+
+Any data defined by system libraries (i.e. not defined by ``lib/System``) must
+not be exposed through the ``lib/System`` interface, even if the header file
+for that function is not exposed. As with functions, this prevents inadvertent
+use of data that might not exist on all platforms.
+
+Minimize Soft Errors
+--------------------
+
+Operating system interfaces will generally provide error results for every
+little thing that could go wrong. In almost all cases, you can divide these
+error results into two groups: normal/good/soft and abnormal/bad/hard. That is,
+some of the errors are simply information like "file not found", "insufficient
+privileges", etc. while other errors are much harder like "out of space", "bad
+disk sector", or "system call interrupted". We'll call the first group "*soft*"
+errors and the second group "*hard*" errors.
+
+``lib/System`` must always attempt to minimize soft errors.  This is a design
+requirement because the minimization of soft errors can affect the granularity
+and the nature of the interface. In general, if you find that you're wanting to
+throw soft errors, you must review the granularity of the interface because it
+is likely you're trying to implement something that is too low level. The rule
+of thumb is to provide interface functions that **can't** fail, except when
+faced with hard errors.
+
+For a trivial example, suppose we wanted to add an "``OpenFileForWriting``"
+function. For many operating systems, if the file doesn't exist, attempting to
+open the file will produce an error.  However, ``lib/System`` should not simply
+throw that error if it occurs because its a soft error. The problem is that the
+interface function, ``OpenFileForWriting`` is too low level. It should be
+``OpenOrCreateFileForWriting``. In the case of the soft "doesn't exist" error,
+this function would just create it and then open it for writing.
+
+This design principle needs to be maintained in ``lib/System`` because it
+avoids the propagation of soft error handling throughout the rest of LLVM.
+Hard errors will generally just cause a termination for an LLVM tool so don't
+be bashful about throwing them.
+
+Rules of thumb:
+
+#. Don't throw soft errors, only hard errors.
+
+#. If you're tempted to throw a soft error, re-think the interface.
+
+#. Handle internally the most common normal/good/soft error conditions
+   so the rest of LLVM doesn't have to.
+
+No throw Specifications
+-----------------------
+
+None of the ``lib/System`` interface functions may be declared with C++
+``throw()`` specifications on them. This requirement makes sure that the
+compiler does not insert additional exception handling code into the interface
+functions. This is a performance consideration: ``lib/System`` functions are at
+the bottom of many call chains and as such can be frequently called. We need
+them to be as efficient as possible.  However, no routines in the system
+library should actually throw exceptions.
+
+Code Organization
+-----------------
+
+Implementations of the System Library interface are separated by their general
+class of operating system. Currently only Unix and Win32 classes are defined
+but more could be added for other operating system classifications.  To
+distinguish which implementation to compile, the code in ``lib/System`` uses
+the ``LLVM_ON_UNIX`` and ``LLVM_ON_WIN32`` ``#defines`` provided via configure
+through the ``llvm/Config/config.h`` file. Each source file in ``lib/System``,
+after implementing the generic (operating system independent) functionality
+needs to include the correct implementation using a set of
+``#if defined(LLVM_ON_XYZ)`` directives. For example, if we had
+``lib/System/File.cpp``, we'd expect to see in that file:
+
+.. code-block:: c++
+
+  #if defined(LLVM_ON_UNIX)
+  #include "Unix/File.cpp"
+  #endif
+  #if defined(LLVM_ON_WIN32)
+  #include "Win32/File.cpp"
+  #endif
+
+The implementation in ``lib/System/Unix/File.cpp`` should handle all Unix
+variants. The implementation in ``lib/System/Win32/File.cpp`` should handle all
+Win32 variants.  What this does is quickly differentiate the basic class of
+operating system that will provide the implementation. The specific details for
+a given platform must still be determined through the use of ``#ifdef``.
+
+Consistent Semantics
+--------------------
+
+The implementation of a ``lib/System`` interface can vary drastically between
+platforms. That's okay as long as the end result of the interface function is
+the same. For example, a function to create a directory is pretty straight
+forward on all operating system. System V IPC on the other hand isn't even
+supported on all platforms. Instead of "supporting" System V IPC,
+``lib/System`` should provide an interface to the basic concept of
+inter-process communications. The implementations might use System V IPC if
+that was available or named pipes, or whatever gets the job done effectively
+for a given operating system.  In all cases, the interface and the
+implementation must be semantically consistent.
+

Added: www-releases/trunk/4.0.0/docs/_sources/TableGen/BackEnds.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TableGen/BackEnds.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TableGen/BackEnds.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TableGen/BackEnds.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,428 @@
+=================
+TableGen BackEnds
+=================
+
+.. contents::
+   :local:
+
+Introduction
+============
+
+TableGen backends are at the core of TableGen's functionality. The source files
+provide the semantics to a generated (in memory) structure, but it's up to the
+backend to print this out in a way that is meaningful to the user (normally a
+C program including a file or a textual list of warnings, options and error
+messages).
+
+TableGen is used by both LLVM and Clang with very different goals. LLVM uses it
+as a way to automate the generation of massive amounts of information regarding
+instructions, schedules, cores and architecture features. Some backends generate
+output that is consumed by more than one source file, so they need to be created
+in a way that is easy to use pre-processor tricks. Some backends can also print
+C code structures, so that they can be directly included as-is.
+
+Clang, on the other hand, uses it mainly for diagnostic messages (errors,
+warnings, tips) and attributes, so more on the textual end of the scale.
+
+LLVM BackEnds
+=============
+
+.. warning::
+   This document is raw. Each section below needs three sub-sections: description
+   of its purpose with a list of users, output generated from generic input, and
+   finally why it needed a new backend (in case there's something similar).
+
+Overall, each backend will take the same TableGen file type and transform into
+similar output for different targets/uses. There is an implicit contract between
+the TableGen files, the back-ends and their users.
+
+For instance, a global contract is that each back-end produces macro-guarded
+sections. Based on whether the file is included by a header or a source file,
+or even in which context of each file the include is being used, you have
+todefine a macro just before including it, to get the right output:
+
+.. code-block:: c++
+
+  #define GET_REGINFO_TARGET_DESC
+  #include "ARMGenRegisterInfo.inc"
+
+And just part of the generated file would be included. This is useful if
+you need the same information in multiple formats (instantiation, initialization,
+getter/setter functions, etc) from the same source TableGen file without having
+to re-compile the TableGen file multiple times.
+
+Sometimes, multiple macros might be defined before the same include file to
+output multiple blocks:
+
+.. code-block:: c++
+
+  #define GET_REGISTER_MATCHER
+  #define GET_SUBTARGET_FEATURE_NAME
+  #define GET_MATCHER_IMPLEMENTATION
+  #include "ARMGenAsmMatcher.inc"
+
+The macros will be undef'd automatically as they're used, in the include file.
+
+On all LLVM back-ends, the ``llvm-tblgen`` binary will be executed on the root
+TableGen file ``<Target>.td``, which should include all others. This guarantees
+that all information needed is accessible, and that no duplication is needed
+in the TableGen files.
+
+CodeEmitter
+-----------
+
+**Purpose**: CodeEmitterGen uses the descriptions of instructions and their fields to
+construct an automated code emitter: a function that, given a MachineInstr,
+returns the (currently, 32-bit unsigned) value of the instruction.
+
+**Output**: C++ code, implementing the target's CodeEmitter
+class by overriding the virtual functions as ``<Target>CodeEmitter::function()``.
+
+**Usage**: Used to include directly at the end of ``<Target>MCCodeEmitter.cpp``.
+
+RegisterInfo
+------------
+
+**Purpose**: This tablegen backend is responsible for emitting a description of a target
+register file for a code generator.  It uses instances of the Register,
+RegisterAliases, and RegisterClass classes to gather this information.
+
+**Output**: C++ code with enums and structures representing the register mappings,
+properties, masks, etc.
+
+**Usage**: Both on ``<Target>BaseRegisterInfo`` and ``<Target>MCTargetDesc`` (headers
+and source files) with macros defining in which they are for declaration vs.
+initialization issues.
+
+InstrInfo
+---------
+
+**Purpose**: This tablegen backend is responsible for emitting a description of the target
+instruction set for the code generator. (what are the differences from CodeEmitter?)
+
+**Output**: C++ code with enums and structures representing the instruction mappings,
+properties, masks, etc.
+
+**Usage**: Both on ``<Target>BaseInstrInfo`` and ``<Target>MCTargetDesc`` (headers
+and source files) with macros defining in which they are for declaration vs.
+initialization issues.
+
+AsmWriter
+---------
+
+**Purpose**: Emits an assembly printer for the current target.
+
+**Output**: Implementation of ``<Target>InstPrinter::printInstruction()``, among
+other things.
+
+**Usage**: Included directly into ``InstPrinter/<Target>InstPrinter.cpp``.
+
+AsmMatcher
+----------
+
+**Purpose**: Emits a target specifier matcher for
+converting parsed assembly operands in the MCInst structures. It also
+emits a matcher for custom operand parsing. Extensive documentation is
+written on the ``AsmMatcherEmitter.cpp`` file.
+
+**Output**: Assembler parsers' matcher functions, declarations, etc.
+
+**Usage**: Used in back-ends' ``AsmParser/<Target>AsmParser.cpp`` for
+building the AsmParser class.
+
+Disassembler
+------------
+
+**Purpose**: Contains disassembler table emitters for various
+architectures. Extensive documentation is written on the
+``DisassemblerEmitter.cpp`` file.
+
+**Output**: Decoding tables, static decoding functions, etc.
+
+**Usage**: Directly included in ``Disassembler/<Target>Disassembler.cpp``
+to cater for all default decodings, after all hand-made ones.
+
+PseudoLowering
+--------------
+
+**Purpose**: Generate pseudo instruction lowering.
+
+**Output**: Implements ``<Target>AsmPrinter::emitPseudoExpansionLowering()``.
+
+**Usage**: Included directly into ``<Target>AsmPrinter.cpp``.
+
+CallingConv
+-----------
+
+**Purpose**: Responsible for emitting descriptions of the calling
+conventions supported by this target.
+
+**Output**: Implement static functions to deal with calling conventions
+chained by matching styles, returning false on no match.
+
+**Usage**: Used in ISelLowering and FastIsel as function pointers to
+implementation returned by a CC selection function.
+
+DAGISel
+-------
+
+**Purpose**: Generate a DAG instruction selector.
+
+**Output**: Creates huge functions for automating DAG selection.
+
+**Usage**: Included in ``<Target>ISelDAGToDAG.cpp`` inside the target's
+implementation of ``SelectionDAGISel``.
+
+DFAPacketizer
+-------------
+
+**Purpose**: This class parses the Schedule.td file and produces an API that
+can be used to reason about whether an instruction can be added to a packet
+on a VLIW architecture. The class internally generates a deterministic finite
+automaton (DFA) that models all possible mappings of machine instructions
+to functional units as instructions are added to a packet.
+
+**Output**: Scheduling tables for GPU back-ends (Hexagon, AMD).
+
+**Usage**: Included directly on ``<Target>InstrInfo.cpp``.
+
+FastISel
+--------
+
+**Purpose**: This tablegen backend emits code for use by the "fast"
+instruction selection algorithm. See the comments at the top of
+lib/CodeGen/SelectionDAG/FastISel.cpp for background. This file
+scans through the target's tablegen instruction-info files
+and extracts instructions with obvious-looking patterns, and it emits
+code to look up these instructions by type and operator.
+
+**Output**: Generates ``Predicate`` and ``FastEmit`` methods.
+
+**Usage**: Implements private methods of the targets' implementation
+of ``FastISel`` class.
+
+Subtarget
+---------
+
+**Purpose**: Generate subtarget enumerations.
+
+**Output**: Enums, globals, local tables for sub-target information.
+
+**Usage**: Populates ``<Target>Subtarget`` and
+``MCTargetDesc/<Target>MCTargetDesc`` files (both headers and source).
+
+Intrinsic
+---------
+
+**Purpose**: Generate (target) intrinsic information.
+
+OptParserDefs
+-------------
+
+**Purpose**: Print enum values for a class.
+
+CTags
+-----
+
+**Purpose**: This tablegen backend emits an index of definitions in ctags(1)
+format. A helper script, utils/TableGen/tdtags, provides an easier-to-use
+interface; run 'tdtags -H' for documentation.
+
+Clang BackEnds
+==============
+
+ClangAttrClasses
+----------------
+
+**Purpose**: Creates Attrs.inc, which contains semantic attribute class
+declarations for any attribute in ``Attr.td`` that has not set ``ASTNode = 0``.
+This file is included as part of ``Attr.h``.
+
+ClangAttrParserStringSwitches
+-----------------------------
+
+**Purpose**: Creates AttrParserStringSwitches.inc, which contains
+StringSwitch::Case statements for parser-related string switches. Each switch
+is given its own macro (such as ``CLANG_ATTR_ARG_CONTEXT_LIST``, or
+``CLANG_ATTR_IDENTIFIER_ARG_LIST``), which is expected to be defined before
+including AttrParserStringSwitches.inc, and undefined after.
+
+ClangAttrImpl
+-------------
+
+**Purpose**: Creates AttrImpl.inc, which contains semantic attribute class
+definitions for any attribute in ``Attr.td`` that has not set ``ASTNode = 0``.
+This file is included as part of ``AttrImpl.cpp``.
+
+ClangAttrList
+-------------
+
+**Purpose**: Creates AttrList.inc, which is used when a list of semantic
+attribute identifiers is required. For instance, ``AttrKinds.h`` includes this
+file to generate the list of ``attr::Kind`` enumeration values. This list is
+separated out into multiple categories: attributes, inheritable attributes, and
+inheritable parameter attributes. This categorization happens automatically
+based on information in ``Attr.td`` and is used to implement the ``classof``
+functionality required for ``dyn_cast`` and similar APIs.
+
+ClangAttrPCHRead
+----------------
+
+**Purpose**: Creates AttrPCHRead.inc, which is used to deserialize attributes
+in the ``ASTReader::ReadAttributes`` function.
+
+ClangAttrPCHWrite
+-----------------
+
+**Purpose**: Creates AttrPCHWrite.inc, which is used to serialize attributes in
+the ``ASTWriter::WriteAttributes`` function.
+
+ClangAttrSpellings
+---------------------
+
+**Purpose**: Creates AttrSpellings.inc, which is used to implement the
+``__has_attribute`` feature test macro.
+
+ClangAttrSpellingListIndex
+--------------------------
+
+**Purpose**: Creates AttrSpellingListIndex.inc, which is used to map parsed
+attribute spellings (including which syntax or scope was used) to an attribute
+spelling list index. These spelling list index values are internal
+implementation details exposed via
+``AttributeList::getAttributeSpellingListIndex``.
+
+ClangAttrVisitor
+-------------------
+
+**Purpose**: Creates AttrVisitor.inc, which is used when implementing 
+recursive AST visitors.
+
+ClangAttrTemplateInstantiate
+----------------------------
+
+**Purpose**: Creates AttrTemplateInstantiate.inc, which implements the
+``instantiateTemplateAttribute`` function, used when instantiating a template
+that requires an attribute to be cloned.
+
+ClangAttrParsedAttrList
+-----------------------
+
+**Purpose**: Creates AttrParsedAttrList.inc, which is used to generate the
+``AttributeList::Kind`` parsed attribute enumeration.
+
+ClangAttrParsedAttrImpl
+-----------------------
+
+**Purpose**: Creates AttrParsedAttrImpl.inc, which is used by
+``AttributeList.cpp`` to implement several functions on the ``AttributeList``
+class. This functionality is implemented via the ``AttrInfoMap ParsedAttrInfo``
+array, which contains one element per parsed attribute object.
+
+ClangAttrParsedAttrKinds
+------------------------
+
+**Purpose**: Creates AttrParsedAttrKinds.inc, which is used to implement the
+``AttributeList::getKind`` function, mapping a string (and syntax) to a parsed
+attribute ``AttributeList::Kind`` enumeration.
+
+ClangAttrDump
+-------------
+
+**Purpose**: Creates AttrDump.inc, which dumps information about an attribute.
+It is used to implement ``ASTDumper::dumpAttr``.
+
+ClangDiagsDefs
+--------------
+
+Generate Clang diagnostics definitions.
+
+ClangDiagGroups
+---------------
+
+Generate Clang diagnostic groups.
+
+ClangDiagsIndexName
+-------------------
+
+Generate Clang diagnostic name index.
+
+ClangCommentNodes
+-----------------
+
+Generate Clang AST comment nodes.
+
+ClangDeclNodes
+--------------
+
+Generate Clang AST declaration nodes.
+
+ClangStmtNodes
+--------------
+
+Generate Clang AST statement nodes.
+
+ClangSACheckers
+---------------
+
+Generate Clang Static Analyzer checkers.
+
+ClangCommentHTMLTags
+--------------------
+
+Generate efficient matchers for HTML tag names that are used in documentation comments.
+
+ClangCommentHTMLTagsProperties
+------------------------------
+
+Generate efficient matchers for HTML tag properties.
+
+ClangCommentHTMLNamedCharacterReferences
+----------------------------------------
+
+Generate function to translate named character references to UTF-8 sequences.
+
+ClangCommentCommandInfo
+-----------------------
+
+Generate command properties for commands that are used in documentation comments.
+
+ClangCommentCommandList
+-----------------------
+
+Generate list of commands that are used in documentation comments.
+
+ArmNeon
+-------
+
+Generate arm_neon.h for clang.
+
+ArmNeonSema
+-----------
+
+Generate ARM NEON sema support for clang.
+
+ArmNeonTest
+-----------
+
+Generate ARM NEON tests for clang.
+
+AttrDocs
+--------
+
+**Purpose**: Creates ``AttributeReference.rst`` from ``AttrDocs.td``, and is
+used for documenting user-facing attributes.
+
+How to write a back-end
+=======================
+
+TODO.
+
+Until we get a step-by-step HowTo for writing TableGen backends, you can at
+least grab the boilerplate (build system, new files, etc.) from Clang's
+r173931.
+
+TODO: How they work, how to write one.  This section should not contain details
+about any particular backend, except maybe ``-print-enums`` as an example.  This
+should highlight the APIs in ``TableGen/Record.h``.
+

Added: www-releases/trunk/4.0.0/docs/_sources/TableGen/Deficiencies.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TableGen/Deficiencies.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TableGen/Deficiencies.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TableGen/Deficiencies.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,31 @@
+=====================
+TableGen Deficiencies
+=====================
+
+.. contents::
+   :local:
+
+Introduction
+============
+
+Despite being very generic, TableGen has some deficiencies that have been
+pointed out numerous times. The common theme is that, while TableGen allows
+you to build Domain-Specific-Languages, the final languages that you create
+lack the power of other DSLs, which in turn increase considerably the size
+and complexity of TableGen files.
+
+At the same time, TableGen allows you to create virtually any meaning of
+the basic concepts via custom-made back-ends, which can pervert the original
+design and make it very hard for newcomers to understand it.
+
+There are some in favour of extending the semantics even more, but making sure
+back-ends adhere to strict rules. Others suggesting we should move to more
+powerful DSLs designed with specific purposes, or even re-using existing
+DSLs.
+
+Known Problems
+==============
+
+TODO: Add here frequently asked questions about why TableGen doesn't do
+what you want, how it might, and how we could extend/restrict it to
+be more use friendly.

Added: www-releases/trunk/4.0.0/docs/_sources/TableGen/LangIntro.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TableGen/LangIntro.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TableGen/LangIntro.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TableGen/LangIntro.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,612 @@
+==============================
+TableGen Language Introduction
+==============================
+
+.. contents::
+   :local:
+
+.. warning::
+   This document is extremely rough. If you find something lacking, please
+   fix it, file a documentation bug, or ask about it on llvm-dev.
+
+Introduction
+============
+
+This document is not meant to be a normative spec about the TableGen language
+in and of itself (i.e. how to understand a given construct in terms of how
+it affects the final set of records represented by the TableGen file). For
+the formal language specification, see :doc:`LangRef`.
+
+TableGen syntax
+===============
+
+TableGen doesn't care about the meaning of data (that is up to the backend to
+define), but it does care about syntax, and it enforces a simple type system.
+This section describes the syntax and the constructs allowed in a TableGen file.
+
+TableGen primitives
+-------------------
+
+TableGen comments
+^^^^^^^^^^^^^^^^^
+
+TableGen supports C++ style "``//``" comments, which run to the end of the
+line, and it also supports **nestable** "``/* */``" comments.
+
+.. _TableGen type:
+
+The TableGen type system
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+TableGen files are strongly typed, in a simple (but complete) type-system.
+These types are used to perform automatic conversions, check for errors, and to
+help interface designers constrain the input that they allow.  Every `value
+definition`_ is required to have an associated type.
+
+TableGen supports a mixture of very low-level types (such as ``bit``) and very
+high-level types (such as ``dag``).  This flexibility is what allows it to
+describe a wide range of information conveniently and compactly.  The TableGen
+types are:
+
+``bit``
+    A 'bit' is a boolean value that can hold either 0 or 1.
+
+``int``
+    The 'int' type represents a simple 32-bit integer value, such as 5.
+
+``string``
+    The 'string' type represents an ordered sequence of characters of arbitrary
+    length.
+
+``bits<n>``
+    A 'bits' type is an arbitrary, but fixed, size integer that is broken up
+    into individual bits.  This type is useful because it can handle some bits
+    being defined while others are undefined.
+
+``list<ty>``
+    This type represents a list whose elements are some other type.  The
+    contained type is arbitrary: it can even be another list type.
+
+Class type
+    Specifying a class name in a type context means that the defined value must
+    be a subclass of the specified class.  This is useful in conjunction with
+    the ``list`` type, for example, to constrain the elements of the list to a
+    common base class (e.g., a ``list<Register>`` can only contain definitions
+    derived from the "``Register``" class).
+
+``dag``
+    This type represents a nestable directed graph of elements.
+
+To date, these types have been sufficient for describing things that TableGen
+has been used for, but it is straight-forward to extend this list if needed.
+
+.. _TableGen expressions:
+
+TableGen values and expressions
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+TableGen allows for a pretty reasonable number of different expression forms
+when building up values.  These forms allow the TableGen file to be written in a
+natural syntax and flavor for the application.  The current expression forms
+supported include:
+
+``?``
+    uninitialized field
+
+``0b1001011``
+    binary integer value.
+    Note that this is sized by the number of bits given and will not be
+    silently extended/truncated.
+
+``7``
+    decimal integer value
+
+``0x7F``
+    hexadecimal integer value
+
+``"foo"``
+    string value
+
+``[{ ... }]``
+    usually called a "code fragment", but is just a multiline string literal
+
+``[ X, Y, Z ]<type>``
+    list value.  <type> is the type of the list element and is usually optional.
+    In rare cases, TableGen is unable to deduce the element type in which case
+    the user must specify it explicitly.
+
+``{ a, b, 0b10 }``
+    initializer for a "bits<4>" value.
+    1-bit from "a", 1-bit from "b", 2-bits from 0b10.
+
+``value``
+    value reference
+
+``value{17}``
+    access to one bit of a value
+
+``value{15-17}``
+    access to multiple bits of a value
+
+``DEF``
+    reference to a record definition
+
+``CLASS<val list>``
+    reference to a new anonymous definition of CLASS with the specified template
+    arguments.
+
+``X.Y``
+    reference to the subfield of a value
+
+``list[4-7,17,2-3]``
+    A slice of the 'list' list, including elements 4,5,6,7,17,2, and 3 from it.
+    Elements may be included multiple times.
+
+``foreach <var> = [ <list> ] in { <body> }``
+
+``foreach <var> = [ <list> ] in <def>``
+    Replicate <body> or <def>, replacing instances of <var> with each value
+    in <list>.  <var> is scoped at the level of the ``foreach`` loop and must
+    not conflict with any other object introduced in <body> or <def>.  Currently
+    only ``def``\s are expanded within <body>.
+
+``foreach <var> = 0-15 in ...``
+
+``foreach <var> = {0-15,32-47} in ...``
+    Loop over ranges of integers. The braces are required for multiple ranges.
+
+``(DEF a, b)``
+    a dag value.  The first element is required to be a record definition, the
+    remaining elements in the list may be arbitrary other values, including
+    nested ```dag``' values.
+
+``!listconcat(a, b, ...)``
+    A list value that is the result of concatenating the 'a' and 'b' lists.
+    The lists must have the same element type.
+    More than two arguments are accepted with the result being the concatenation
+    of all the lists given.
+
+``!strconcat(a, b, ...)``
+    A string value that is the result of concatenating the 'a' and 'b' strings.
+    More than two arguments are accepted with the result being the concatenation
+    of all the strings given.
+
+``str1#str2``
+    "#" (paste) is a shorthand for !strconcat.  It may concatenate things that
+    are not quoted strings, in which case an implicit !cast<string> is done on
+    the operand of the paste.
+
+``!cast<type>(a)``
+    A symbol of type *type* obtained by looking up the string 'a' in the symbol
+    table.  If the type of 'a' does not match *type*, TableGen aborts with an
+    error. !cast<string> is a special case in that the argument must be an
+    object defined by a 'def' construct.
+
+``!subst(a, b, c)``
+    If 'a' and 'b' are of string type or are symbol references, substitute 'b'
+    for 'a' in 'c.'  This operation is analogous to $(subst) in GNU make.
+
+``!foreach(a, b, c)``
+    For each member of dag or list 'b' apply operator 'c.'  'a' is a dummy
+    variable that should be declared as a member variable of an instantiated
+    class.  This operation is analogous to $(foreach) in GNU make.
+
+``!head(a)``
+    The first element of list 'a.'
+
+``!tail(a)``
+    The 2nd-N elements of list 'a.'
+
+``!empty(a)``
+    An integer {0,1} indicating whether list 'a' is empty.
+
+``!if(a,b,c)``
+  'b' if the result of 'int' or 'bit' operator 'a' is nonzero, 'c' otherwise.
+
+``!eq(a,b)``
+    'bit 1' if string a is equal to string b, 0 otherwise.  This only operates
+    on string, int and bit objects.  Use !cast<string> to compare other types of
+    objects.
+
+``!shl(a,b)`` ``!srl(a,b)`` ``!sra(a,b)`` ``!add(a,b)`` ``!and(a,b)``
+    The usual binary and arithmetic operators.
+
+Note that all of the values have rules specifying how they convert to values
+for different types.  These rules allow you to assign a value like "``7``"
+to a "``bits<4>``" value, for example.
+
+Classes and definitions
+-----------------------
+
+As mentioned in the :doc:`introduction <index>`, classes and definitions (collectively known as
+'records') in TableGen are the main high-level unit of information that TableGen
+collects.  Records are defined with a ``def`` or ``class`` keyword, the record
+name, and an optional list of "`template arguments`_".  If the record has
+superclasses, they are specified as a comma separated list that starts with a
+colon character ("``:``").  If `value definitions`_ or `let expressions`_ are
+needed for the class, they are enclosed in curly braces ("``{}``"); otherwise,
+the record ends with a semicolon.
+
+Here is a simple TableGen file:
+
+.. code-block:: text
+
+  class C { bit V = 1; }
+  def X : C;
+  def Y : C {
+    string Greeting = "hello";
+  }
+
+This example defines two definitions, ``X`` and ``Y``, both of which derive from
+the ``C`` class.  Because of this, they both get the ``V`` bit value.  The ``Y``
+definition also gets the Greeting member as well.
+
+In general, classes are useful for collecting together the commonality between a
+group of records and isolating it in a single place.  Also, classes permit the
+specification of default values for their subclasses, allowing the subclasses to
+override them as they wish.
+
+.. _value definition:
+.. _value definitions:
+
+Value definitions
+^^^^^^^^^^^^^^^^^
+
+Value definitions define named entries in records.  A value must be defined
+before it can be referred to as the operand for another value definition or
+before the value is reset with a `let expression`_.  A value is defined by
+specifying a `TableGen type`_ and a name.  If an initial value is available, it
+may be specified after the type with an equal sign.  Value definitions require
+terminating semicolons.
+
+.. _let expression:
+.. _let expressions:
+.. _"let" expressions within a record:
+
+'let' expressions
+^^^^^^^^^^^^^^^^^
+
+A record-level let expression is used to change the value of a value definition
+in a record.  This is primarily useful when a superclass defines a value that a
+derived class or definition wants to override.  Let expressions consist of the
+'``let``' keyword followed by a value name, an equal sign ("``=``"), and a new
+value.  For example, a new class could be added to the example above, redefining
+the ``V`` field for all of its subclasses:
+
+.. code-block:: text
+
+  class D : C { let V = 0; }
+  def Z : D;
+
+In this case, the ``Z`` definition will have a zero value for its ``V`` value,
+despite the fact that it derives (indirectly) from the ``C`` class, because the
+``D`` class overrode its value.
+
+.. _template arguments:
+
+Class template arguments
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+TableGen permits the definition of parameterized classes as well as normal
+concrete classes.  Parameterized TableGen classes specify a list of variable
+bindings (which may optionally have defaults) that are bound when used.  Here is
+a simple example:
+
+.. code-block:: text
+
+  class FPFormat<bits<3> val> {
+    bits<3> Value = val;
+  }
+  def NotFP      : FPFormat<0>;
+  def ZeroArgFP  : FPFormat<1>;
+  def OneArgFP   : FPFormat<2>;
+  def OneArgFPRW : FPFormat<3>;
+  def TwoArgFP   : FPFormat<4>;
+  def CompareFP  : FPFormat<5>;
+  def CondMovFP  : FPFormat<6>;
+  def SpecialFP  : FPFormat<7>;
+
+In this case, template arguments are used as a space efficient way to specify a
+list of "enumeration values", each with a "``Value``" field set to the specified
+integer.
+
+The more esoteric forms of `TableGen expressions`_ are useful in conjunction
+with template arguments.  As an example:
+
+.. code-block:: text
+
+  class ModRefVal<bits<2> val> {
+    bits<2> Value = val;
+  }
+
+  def None   : ModRefVal<0>;
+  def Mod    : ModRefVal<1>;
+  def Ref    : ModRefVal<2>;
+  def ModRef : ModRefVal<3>;
+
+  class Value<ModRefVal MR> {
+    // Decode some information into a more convenient format, while providing
+    // a nice interface to the user of the "Value" class.
+    bit isMod = MR.Value{0};
+    bit isRef = MR.Value{1};
+
+    // other stuff...
+  }
+
+  // Example uses
+  def bork : Value<Mod>;
+  def zork : Value<Ref>;
+  def hork : Value<ModRef>;
+
+This is obviously a contrived example, but it shows how template arguments can
+be used to decouple the interface provided to the user of the class from the
+actual internal data representation expected by the class.  In this case,
+running ``llvm-tblgen`` on the example prints the following definitions:
+
+.. code-block:: text
+
+  def bork {      // Value
+    bit isMod = 1;
+    bit isRef = 0;
+  }
+  def hork {      // Value
+    bit isMod = 1;
+    bit isRef = 1;
+  }
+  def zork {      // Value
+    bit isMod = 0;
+    bit isRef = 1;
+  }
+
+This shows that TableGen was able to dig into the argument and extract a piece
+of information that was requested by the designer of the "Value" class.  For
+more realistic examples, please see existing users of TableGen, such as the X86
+backend.
+
+Multiclass definitions and instances
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+While classes with template arguments are a good way to factor commonality
+between two instances of a definition, multiclasses allow a convenient notation
+for defining multiple definitions at once (instances of implicitly constructed
+classes).  For example, consider an 3-address instruction set whose instructions
+come in two forms: "``reg = reg op reg``" and "``reg = reg op imm``"
+(e.g. SPARC). In this case, you'd like to specify in one place that this
+commonality exists, then in a separate place indicate what all the ops are.
+
+Here is an example TableGen fragment that shows this idea:
+
+.. code-block:: text
+
+  def ops;
+  def GPR;
+  def Imm;
+  class inst<int opc, string asmstr, dag operandlist>;
+
+  multiclass ri_inst<int opc, string asmstr> {
+    def _rr : inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
+                   (ops GPR:$dst, GPR:$src1, GPR:$src2)>;
+    def _ri : inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
+                   (ops GPR:$dst, GPR:$src1, Imm:$src2)>;
+  }
+
+  // Instantiations of the ri_inst multiclass.
+  defm ADD : ri_inst<0b111, "add">;
+  defm SUB : ri_inst<0b101, "sub">;
+  defm MUL : ri_inst<0b100, "mul">;
+  ...
+
+The name of the resultant definitions has the multidef fragment names appended
+to them, so this defines ``ADD_rr``, ``ADD_ri``, ``SUB_rr``, etc.  A defm may
+inherit from multiple multiclasses, instantiating definitions from each
+multiclass.  Using a multiclass this way is exactly equivalent to instantiating
+the classes multiple times yourself, e.g. by writing:
+
+.. code-block:: text
+
+  def ops;
+  def GPR;
+  def Imm;
+  class inst<int opc, string asmstr, dag operandlist>;
+
+  class rrinst<int opc, string asmstr>
+    : inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
+           (ops GPR:$dst, GPR:$src1, GPR:$src2)>;
+
+  class riinst<int opc, string asmstr>
+    : inst<opc, !strconcat(asmstr, " $dst, $src1, $src2"),
+           (ops GPR:$dst, GPR:$src1, Imm:$src2)>;
+
+  // Instantiations of the ri_inst multiclass.
+  def ADD_rr : rrinst<0b111, "add">;
+  def ADD_ri : riinst<0b111, "add">;
+  def SUB_rr : rrinst<0b101, "sub">;
+  def SUB_ri : riinst<0b101, "sub">;
+  def MUL_rr : rrinst<0b100, "mul">;
+  def MUL_ri : riinst<0b100, "mul">;
+  ...
+
+A ``defm`` can also be used inside a multiclass providing several levels of
+multiclass instantiations.
+
+.. code-block:: text
+
+  class Instruction<bits<4> opc, string Name> {
+    bits<4> opcode = opc;
+    string name = Name;
+  }
+
+  multiclass basic_r<bits<4> opc> {
+    def rr : Instruction<opc, "rr">;
+    def rm : Instruction<opc, "rm">;
+  }
+
+  multiclass basic_s<bits<4> opc> {
+    defm SS : basic_r<opc>;
+    defm SD : basic_r<opc>;
+    def X : Instruction<opc, "x">;
+  }
+
+  multiclass basic_p<bits<4> opc> {
+    defm PS : basic_r<opc>;
+    defm PD : basic_r<opc>;
+    def Y : Instruction<opc, "y">;
+  }
+
+  defm ADD : basic_s<0xf>, basic_p<0xf>;
+  ...
+
+  // Results
+  def ADDPDrm { ...
+  def ADDPDrr { ...
+  def ADDPSrm { ...
+  def ADDPSrr { ...
+  def ADDSDrm { ...
+  def ADDSDrr { ...
+  def ADDY { ...
+  def ADDX { ...
+
+``defm`` declarations can inherit from classes too, the rule to follow is that
+the class list must start after the last multiclass, and there must be at least
+one multiclass before them.
+
+.. code-block:: text
+
+  class XD { bits<4> Prefix = 11; }
+  class XS { bits<4> Prefix = 12; }
+
+  class I<bits<4> op> {
+    bits<4> opcode = op;
+  }
+
+  multiclass R {
+    def rr : I<4>;
+    def rm : I<2>;
+  }
+
+  multiclass Y {
+    defm SS : R, XD;
+    defm SD : R, XS;
+  }
+
+  defm Instr : Y;
+
+  // Results
+  def InstrSDrm {
+    bits<4> opcode = { 0, 0, 1, 0 };
+    bits<4> Prefix = { 1, 1, 0, 0 };
+  }
+  ...
+  def InstrSSrr {
+    bits<4> opcode = { 0, 1, 0, 0 };
+    bits<4> Prefix = { 1, 0, 1, 1 };
+  }
+
+File scope entities
+-------------------
+
+File inclusion
+^^^^^^^^^^^^^^
+
+TableGen supports the '``include``' token, which textually substitutes the
+specified file in place of the include directive.  The filename should be
+specified as a double quoted string immediately after the '``include``' keyword.
+Example:
+
+.. code-block:: text
+
+  include "foo.td"
+
+'let' expressions
+^^^^^^^^^^^^^^^^^
+
+"Let" expressions at file scope are similar to `"let" expressions within a
+record`_, except they can specify a value binding for multiple records at a
+time, and may be useful in certain other cases.  File-scope let expressions are
+really just another way that TableGen allows the end-user to factor out
+commonality from the records.
+
+File-scope "let" expressions take a comma-separated list of bindings to apply,
+and one or more records to bind the values in.  Here are some examples:
+
+.. code-block:: text
+
+  let isTerminator = 1, isReturn = 1, isBarrier = 1, hasCtrlDep = 1 in
+    def RET : I<0xC3, RawFrm, (outs), (ins), "ret", [(X86retflag 0)]>;
+
+  let isCall = 1 in
+    // All calls clobber the non-callee saved registers...
+    let Defs = [EAX, ECX, EDX, FP0, FP1, FP2, FP3, FP4, FP5, FP6, ST0,
+                MM0, MM1, MM2, MM3, MM4, MM5, MM6, MM7,
+                XMM0, XMM1, XMM2, XMM3, XMM4, XMM5, XMM6, XMM7, EFLAGS] in {
+      def CALLpcrel32 : Ii32<0xE8, RawFrm, (outs), (ins i32imm:$dst,variable_ops),
+                             "call\t${dst:call}", []>;
+      def CALL32r     : I<0xFF, MRM2r, (outs), (ins GR32:$dst, variable_ops),
+                          "call\t{*}$dst", [(X86call GR32:$dst)]>;
+      def CALL32m     : I<0xFF, MRM2m, (outs), (ins i32mem:$dst, variable_ops),
+                          "call\t{*}$dst", []>;
+    }
+
+File-scope "let" expressions are often useful when a couple of definitions need
+to be added to several records, and the records do not otherwise need to be
+opened, as in the case with the ``CALL*`` instructions above.
+
+It's also possible to use "let" expressions inside multiclasses, providing more
+ways to factor out commonality from the records, specially if using several
+levels of multiclass instantiations. This also avoids the need of using "let"
+expressions within subsequent records inside a multiclass.
+
+.. code-block:: text
+
+  multiclass basic_r<bits<4> opc> {
+    let Predicates = [HasSSE2] in {
+      def rr : Instruction<opc, "rr">;
+      def rm : Instruction<opc, "rm">;
+    }
+    let Predicates = [HasSSE3] in
+      def rx : Instruction<opc, "rx">;
+  }
+
+  multiclass basic_ss<bits<4> opc> {
+    let IsDouble = 0 in
+      defm SS : basic_r<opc>;
+
+    let IsDouble = 1 in
+      defm SD : basic_r<opc>;
+  }
+
+  defm ADD : basic_ss<0xf>;
+
+Looping
+^^^^^^^
+
+TableGen supports the '``foreach``' block, which textually replicates the loop
+body, substituting iterator values for iterator references in the body.
+Example:
+
+.. code-block:: text
+
+  foreach i = [0, 1, 2, 3] in {
+    def R#i : Register<...>;
+    def F#i : Register<...>;
+  }
+
+This will create objects ``R0``, ``R1``, ``R2`` and ``R3``.  ``foreach`` blocks
+may be nested. If there is only one item in the body the braces may be
+elided:
+
+.. code-block:: text
+
+  foreach i = [0, 1, 2, 3] in
+    def R#i : Register<...>;
+
+Code Generator backend info
+===========================
+
+Expressions used by code generator to describe instructions and isel patterns:
+
+``(implicit a)``
+    an implicitly defined physical register.  This tells the dag instruction
+    selection emitter the input pattern's extra definitions matches implicit
+    physical register definitions.
+

Added: www-releases/trunk/4.0.0/docs/_sources/TableGen/LangRef.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TableGen/LangRef.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TableGen/LangRef.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TableGen/LangRef.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,390 @@
+===========================
+TableGen Language Reference
+===========================
+
+.. contents::
+   :local:
+
+.. warning::
+   This document is extremely rough. If you find something lacking, please
+   fix it, file a documentation bug, or ask about it on llvm-dev.
+
+Introduction
+============
+
+This document is meant to be a normative spec about the TableGen language
+in and of itself (i.e. how to understand a given construct in terms of how
+it affects the final set of records represented by the TableGen file). If
+you are unsure if this document is really what you are looking for, please
+read the :doc:`introduction to TableGen <index>` first.
+
+Notation
+========
+
+The lexical and syntax notation used here is intended to imitate
+`Python's`_. In particular, for lexical definitions, the productions
+operate at the character level and there is no implied whitespace between
+elements. The syntax definitions operate at the token level, so there is
+implied whitespace between tokens.
+
+.. _`Python's`: http://docs.python.org/py3k/reference/introduction.html#notation
+
+Lexical Analysis
+================
+
+TableGen supports BCPL (``// ...``) and nestable C-style (``/* ... */``)
+comments.
+
+The following is a listing of the basic punctuation tokens::
+
+   - + [ ] { } ( ) < > : ; .  = ? #
+
+Numeric literals take one of the following forms:
+
+.. TableGen actually will lex some pretty strange sequences an interpret
+   them as numbers. What is shown here is an attempt to approximate what it
+   "should" accept.
+
+.. productionlist::
+   TokInteger: `DecimalInteger` | `HexInteger` | `BinInteger`
+   DecimalInteger: ["+" | "-"] ("0"..."9")+
+   HexInteger: "0x" ("0"..."9" | "a"..."f" | "A"..."F")+
+   BinInteger: "0b" ("0" | "1")+
+
+One aspect to note is that the :token:`DecimalInteger` token *includes* the
+``+`` or ``-``, as opposed to having ``+`` and ``-`` be unary operators as
+most languages do.
+
+Also note that :token:`BinInteger` creates a value of type ``bits<n>``
+(where ``n`` is the number of bits).  This will implicitly convert to
+integers when needed.
+
+TableGen has identifier-like tokens:
+
+.. productionlist::
+   ualpha: "a"..."z" | "A"..."Z" | "_"
+   TokIdentifier: ("0"..."9")* `ualpha` (`ualpha` | "0"..."9")*
+   TokVarName: "$" `ualpha` (`ualpha` |  "0"..."9")*
+
+Note that unlike most languages, TableGen allows :token:`TokIdentifier` to
+begin with a number. In case of ambiguity, a token will be interpreted as a
+numeric literal rather than an identifier.
+
+TableGen also has two string-like literals:
+
+.. productionlist::
+   TokString: '"' <non-'"' characters and C-like escapes> '"'
+   TokCodeFragment: "[{" <shortest text not containing "}]"> "}]"
+
+:token:`TokCodeFragment` is essentially a multiline string literal
+delimited by ``[{`` and ``}]``.
+
+.. note::
+   The current implementation accepts the following C-like escapes::
+
+      \\ \' \" \t \n
+
+TableGen also has the following keywords::
+
+   bit   bits      class   code         dag
+   def   foreach   defm    field        in
+   int   let       list    multiclass   string
+
+TableGen also has "bang operators" which have a
+wide variety of meanings:
+
+.. productionlist::
+   BangOperator: one of
+               :!eq     !if      !head    !tail      !con
+               :!add    !shl     !sra     !srl       !and
+               :!or     !empty   !subst   !foreach   !strconcat
+               :!cast   !listconcat
+
+
+Syntax
+======
+
+TableGen has an ``include`` mechanism. It does not play a role in the
+syntax per se, since it is lexically replaced with the contents of the
+included file.
+
+.. productionlist::
+   IncludeDirective: "include" `TokString`
+
+TableGen's top-level production consists of "objects".
+
+.. productionlist::
+   TableGenFile: `Object`*
+   Object: `Class` | `Def` | `Defm` | `Let` | `MultiClass` | `Foreach`
+
+``class``\es
+------------
+
+.. productionlist::
+   Class: "class" `TokIdentifier` [`TemplateArgList`] `ObjectBody`
+
+A ``class`` declaration creates a record which other records can inherit
+from. A class can be parametrized by a list of "template arguments", whose
+values can be used in the class body.
+
+A given class can only be defined once. A ``class`` declaration is
+considered to define the class if any of the following is true:
+
+.. break ObjectBody into its consituents so that they are present here?
+
+#. The :token:`TemplateArgList` is present.
+#. The :token:`Body` in the :token:`ObjectBody` is present and is not empty.
+#. The :token:`BaseClassList` in the :token:`ObjectBody` is present.
+
+You can declare an empty class by giving and empty :token:`TemplateArgList`
+and an empty :token:`ObjectBody`. This can serve as a restricted form of
+forward declaration: note that records deriving from the forward-declared
+class will inherit no fields from it since the record expansion is done
+when the record is parsed.
+
+.. productionlist::
+   TemplateArgList: "<" `Declaration` ("," `Declaration`)* ">"
+
+Declarations
+------------
+
+.. Omitting mention of arcane "field" prefix to discourage its use.
+
+The declaration syntax is pretty much what you would expect as a C++
+programmer.
+
+.. productionlist::
+   Declaration: `Type` `TokIdentifier` ["=" `Value`]
+
+It assigns the value to the identifier.
+
+Types
+-----
+
+.. productionlist::
+   Type: "string" | "code" | "bit" | "int" | "dag"
+       :| "bits" "<" `TokInteger` ">"
+       :| "list" "<" `Type` ">"
+       :| `ClassID`
+   ClassID: `TokIdentifier`
+
+Both ``string`` and ``code`` correspond to the string type; the difference
+is purely to indicate programmer intention.
+
+The :token:`ClassID` must identify a class that has been previously
+declared or defined.
+
+Values
+------
+
+.. productionlist::
+   Value: `SimpleValue` `ValueSuffix`*
+   ValueSuffix: "{" `RangeList` "}"
+              :| "[" `RangeList` "]"
+              :| "." `TokIdentifier`
+   RangeList: `RangePiece` ("," `RangePiece`)*
+   RangePiece: `TokInteger`
+             :| `TokInteger` "-" `TokInteger`
+             :| `TokInteger` `TokInteger`
+
+The peculiar last form of :token:`RangePiece` is due to the fact that the
+"``-``" is included in the :token:`TokInteger`, hence ``1-5`` gets lexed as
+two consecutive :token:`TokInteger`'s, with values ``1`` and ``-5``,
+instead of "1", "-", and "5".
+The :token:`RangeList` can be thought of as specifying "list slice" in some
+contexts.
+
+
+:token:`SimpleValue` has a number of forms:
+
+
+.. productionlist::
+   SimpleValue: `TokIdentifier`
+
+The value will be the variable referenced by the identifier. It can be one
+of:
+
+.. The code for this is exceptionally abstruse. These examples are a
+   best-effort attempt.
+
+* name of a ``def``, such as the use of ``Bar`` in::
+
+     def Bar : SomeClass {
+       int X = 5;
+     }
+
+     def Foo {
+       SomeClass Baz = Bar;
+     }
+
+* value local to a ``def``, such as the use of ``Bar`` in::
+
+     def Foo {
+       int Bar = 5;
+       int Baz = Bar;
+     }
+
+* a template arg of a ``class``, such as the use of ``Bar`` in::
+
+     class Foo<int Bar> {
+       int Baz = Bar;
+     }
+
+* value local to a ``multiclass``, such as the use of ``Bar`` in::
+
+     multiclass Foo {
+       int Bar = 5;
+       int Baz = Bar;
+     }
+
+* a template arg to a ``multiclass``, such as the use of ``Bar`` in::
+
+     multiclass Foo<int Bar> {
+       int Baz = Bar;
+     }
+
+.. productionlist::
+   SimpleValue: `TokInteger`
+
+This represents the numeric value of the integer.
+
+.. productionlist::
+   SimpleValue: `TokString`+
+
+Multiple adjacent string literals are concatenated like in C/C++. The value
+is the concatenation of the strings.
+
+.. productionlist::
+   SimpleValue: `TokCodeFragment`
+
+The value is the string value of the code fragment.
+
+.. productionlist::
+   SimpleValue: "?"
+
+``?`` represents an "unset" initializer.
+
+.. productionlist::
+   SimpleValue: "{" `ValueList` "}"
+   ValueList: [`ValueListNE`]
+   ValueListNE: `Value` ("," `Value`)*
+
+This represents a sequence of bits, as would be used to initialize a
+``bits<n>`` field (where ``n`` is the number of bits).
+
+.. productionlist::
+   SimpleValue: `ClassID` "<" `ValueListNE` ">"
+
+This generates a new anonymous record definition (as would be created by an
+unnamed ``def`` inheriting from the given class with the given template
+arguments) and the value is the value of that record definition.
+
+.. productionlist::
+   SimpleValue: "[" `ValueList` "]" ["<" `Type` ">"]
+
+A list initializer. The optional :token:`Type` can be used to indicate a
+specific element type, otherwise the element type will be deduced from the
+given values.
+
+.. The initial `DagArg` of the dag must start with an identifier or
+   !cast, but this is more of an implementation detail and so for now just
+   leave it out.
+
+.. productionlist::
+   SimpleValue: "(" `DagArg` `DagArgList` ")"
+   DagArgList: `DagArg` ("," `DagArg`)*
+   DagArg: `Value` [":" `TokVarName`] | `TokVarName`
+
+The initial :token:`DagArg` is called the "operator" of the dag.
+
+.. productionlist::
+   SimpleValue: `BangOperator` ["<" `Type` ">"] "(" `ValueListNE` ")"
+
+Bodies
+------
+
+.. productionlist::
+   ObjectBody: `BaseClassList` `Body`
+   BaseClassList: [":" `BaseClassListNE`]
+   BaseClassListNE: `SubClassRef` ("," `SubClassRef`)*
+   SubClassRef: (`ClassID` | `MultiClassID`) ["<" `ValueList` ">"]
+   DefmID: `TokIdentifier`
+
+The version with the :token:`MultiClassID` is only valid in the
+:token:`BaseClassList` of a ``defm``.
+The :token:`MultiClassID` should be the name of a ``multiclass``.
+
+.. put this somewhere else
+
+It is after parsing the base class list that the "let stack" is applied.
+
+.. productionlist::
+   Body: ";" | "{" BodyList "}"
+   BodyList: BodyItem*
+   BodyItem: `Declaration` ";"
+           :| "let" `TokIdentifier` [`RangeList`] "=" `Value` ";"
+
+The ``let`` form allows overriding the value of an inherited field.
+
+``def``
+-------
+
+.. TODO::
+   There can be pastes in the names here, like ``#NAME#``. Look into that
+   and document it (it boils down to ParseIDValue with IDParseMode ==
+   ParseNameMode). ParseObjectName calls into the general ParseValue, with
+   the only different from "arbitrary expression parsing" being IDParseMode
+   == Mode.
+
+.. productionlist::
+   Def: "def" `TokIdentifier` `ObjectBody`
+
+Defines a record whose name is given by the :token:`TokIdentifier`. The
+fields of the record are inherited from the base classes and defined in the
+body.
+
+Special handling occurs if this ``def`` appears inside a ``multiclass`` or
+a ``foreach``.
+
+``defm``
+--------
+
+.. productionlist::
+   Defm: "defm" `TokIdentifier` ":" `BaseClassListNE` ";"
+
+Note that in the :token:`BaseClassList`, all of the ``multiclass``'s must
+precede any ``class``'s that appear.
+
+``foreach``
+-----------
+
+.. productionlist::
+   Foreach: "foreach" `Declaration` "in" "{" `Object`* "}"
+          :| "foreach" `Declaration` "in" `Object`
+
+The value assigned to the variable in the declaration is iterated over and
+the object or object list is reevaluated with the variable set at each
+iterated value.
+
+Top-Level ``let``
+-----------------
+
+.. productionlist::
+   Let:  "let" `LetList` "in" "{" `Object`* "}"
+      :| "let" `LetList` "in" `Object`
+   LetList: `LetItem` ("," `LetItem`)*
+   LetItem: `TokIdentifier` [`RangeList`] "=" `Value`
+
+This is effectively equivalent to ``let`` inside the body of a record
+except that it applies to multiple records at a time. The bindings are
+applied at the end of parsing the base classes of a record.
+
+``multiclass``
+--------------
+
+.. productionlist::
+   MultiClass: "multiclass" `TokIdentifier` [`TemplateArgList`]
+             : [":" `BaseMultiClassList`] "{" `MultiClassObject`+ "}"
+   BaseMultiClassList: `MultiClassID` ("," `MultiClassID`)*
+   MultiClassID: `TokIdentifier`
+   MultiClassObject: `Def` | `Defm` | `Let` | `Foreach`

Added: www-releases/trunk/4.0.0/docs/_sources/TableGen/index.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TableGen/index.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TableGen/index.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TableGen/index.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,308 @@
+========
+TableGen
+========
+
+.. contents::
+   :local:
+
+.. toctree::
+   :hidden:
+
+   BackEnds
+   LangRef
+   LangIntro
+   Deficiencies
+
+Introduction
+============
+
+TableGen's purpose is to help a human develop and maintain records of
+domain-specific information.  Because there may be a large number of these
+records, it is specifically designed to allow writing flexible descriptions and
+for common features of these records to be factored out.  This reduces the
+amount of duplication in the description, reduces the chance of error, and makes
+it easier to structure domain specific information.
+
+The core part of TableGen parses a file, instantiates the declarations, and
+hands the result off to a domain-specific `backend`_ for processing.
+
+The current major users of TableGen are :doc:`../CodeGenerator`
+and the
+`Clang diagnostics and attributes <http://clang.llvm.org/docs/UsersManual.html#controlling-errors-and-warnings>`_.
+
+Note that if you work on TableGen much, and use emacs or vim, that you can find
+an emacs "TableGen mode" and a vim language file in the ``llvm/utils/emacs`` and
+``llvm/utils/vim`` directories of your LLVM distribution, respectively.
+
+.. _intro:
+
+
+The TableGen program
+====================
+
+TableGen files are interpreted by the TableGen program: `llvm-tblgen` available
+on your build directory under `bin`. It is not installed in the system (or where
+your sysroot is set to), since it has no use beyond LLVM's build process.
+
+Running TableGen
+----------------
+
+TableGen runs just like any other LLVM tool.  The first (optional) argument
+specifies the file to read.  If a filename is not specified, ``llvm-tblgen``
+reads from standard input.
+
+To be useful, one of the `backends`_ must be used.  These backends are
+selectable on the command line (type '``llvm-tblgen -help``' for a list).  For
+example, to get a list of all of the definitions that subclass a particular type
+(which can be useful for building up an enum list of these records), use the
+``-print-enums`` option:
+
+.. code-block:: bash
+
+  $ llvm-tblgen X86.td -print-enums -class=Register
+  AH, AL, AX, BH, BL, BP, BPL, BX, CH, CL, CX, DH, DI, DIL, DL, DX, EAX, EBP, EBX,
+  ECX, EDI, EDX, EFLAGS, EIP, ESI, ESP, FP0, FP1, FP2, FP3, FP4, FP5, FP6, IP,
+  MM0, MM1, MM2, MM3, MM4, MM5, MM6, MM7, R10, R10B, R10D, R10W, R11, R11B, R11D,
+  R11W, R12, R12B, R12D, R12W, R13, R13B, R13D, R13W, R14, R14B, R14D, R14W, R15,
+  R15B, R15D, R15W, R8, R8B, R8D, R8W, R9, R9B, R9D, R9W, RAX, RBP, RBX, RCX, RDI,
+  RDX, RIP, RSI, RSP, SI, SIL, SP, SPL, ST0, ST1, ST2, ST3, ST4, ST5, ST6, ST7,
+  XMM0, XMM1, XMM10, XMM11, XMM12, XMM13, XMM14, XMM15, XMM2, XMM3, XMM4, XMM5,
+  XMM6, XMM7, XMM8, XMM9,
+
+  $ llvm-tblgen X86.td -print-enums -class=Instruction 
+  ABS_F, ABS_Fp32, ABS_Fp64, ABS_Fp80, ADC32mi, ADC32mi8, ADC32mr, ADC32ri,
+  ADC32ri8, ADC32rm, ADC32rr, ADC64mi32, ADC64mi8, ADC64mr, ADC64ri32, ADC64ri8,
+  ADC64rm, ADC64rr, ADD16mi, ADD16mi8, ADD16mr, ADD16ri, ADD16ri8, ADD16rm,
+  ADD16rr, ADD32mi, ADD32mi8, ADD32mr, ADD32ri, ADD32ri8, ADD32rm, ADD32rr,
+  ADD64mi32, ADD64mi8, ADD64mr, ADD64ri32, ...
+
+The default backend prints out all of the records.
+
+If you plan to use TableGen, you will most likely have to write a `backend`_
+that extracts the information specific to what you need and formats it in the
+appropriate way.
+
+Example
+-------
+
+With no other arguments, `llvm-tblgen` parses the specified file and prints out all
+of the classes, then all of the definitions.  This is a good way to see what the
+various definitions expand to fully.  Running this on the ``X86.td`` file prints
+this (at the time of this writing):
+
+.. code-block:: text
+
+  ...
+  def ADD32rr {   // Instruction X86Inst I
+    string Namespace = "X86";
+    dag OutOperandList = (outs GR32:$dst);
+    dag InOperandList = (ins GR32:$src1, GR32:$src2);
+    string AsmString = "add{l}\t{$src2, $dst|$dst, $src2}";
+    list<dag> Pattern = [(set GR32:$dst, (add GR32:$src1, GR32:$src2))];
+    list<Register> Uses = [];
+    list<Register> Defs = [EFLAGS];
+    list<Predicate> Predicates = [];
+    int CodeSize = 3;
+    int AddedComplexity = 0;
+    bit isReturn = 0;
+    bit isBranch = 0;
+    bit isIndirectBranch = 0;
+    bit isBarrier = 0;
+    bit isCall = 0;
+    bit canFoldAsLoad = 0;
+    bit mayLoad = 0;
+    bit mayStore = 0;
+    bit isImplicitDef = 0;
+    bit isConvertibleToThreeAddress = 1;
+    bit isCommutable = 1;
+    bit isTerminator = 0;
+    bit isReMaterializable = 0;
+    bit isPredicable = 0;
+    bit hasDelaySlot = 0;
+    bit usesCustomInserter = 0;
+    bit hasCtrlDep = 0;
+    bit isNotDuplicable = 0;
+    bit hasSideEffects = 0;
+    InstrItinClass Itinerary = NoItinerary;
+    string Constraints = "";
+    string DisableEncoding = "";
+    bits<8> Opcode = { 0, 0, 0, 0, 0, 0, 0, 1 };
+    Format Form = MRMDestReg;
+    bits<6> FormBits = { 0, 0, 0, 0, 1, 1 };
+    ImmType ImmT = NoImm;
+    bits<3> ImmTypeBits = { 0, 0, 0 };
+    bit hasOpSizePrefix = 0;
+    bit hasAdSizePrefix = 0;
+    bits<4> Prefix = { 0, 0, 0, 0 };
+    bit hasREX_WPrefix = 0;
+    FPFormat FPForm = ?;
+    bits<3> FPFormBits = { 0, 0, 0 };
+  }
+  ...
+
+This definition corresponds to the 32-bit register-register ``add`` instruction
+of the x86 architecture.  ``def ADD32rr`` defines a record named
+``ADD32rr``, and the comment at the end of the line indicates the superclasses
+of the definition.  The body of the record contains all of the data that
+TableGen assembled for the record, indicating that the instruction is part of
+the "X86" namespace, the pattern indicating how the instruction is selected by
+the code generator, that it is a two-address instruction, has a particular
+encoding, etc.  The contents and semantics of the information in the record are
+specific to the needs of the X86 backend, and are only shown as an example.
+
+As you can see, a lot of information is needed for every instruction supported
+by the code generator, and specifying it all manually would be unmaintainable,
+prone to bugs, and tiring to do in the first place.  Because we are using
+TableGen, all of the information was derived from the following definition:
+
+.. code-block:: text
+
+  let Defs = [EFLAGS],
+      isCommutable = 1,                  // X = ADD Y,Z --> X = ADD Z,Y
+      isConvertibleToThreeAddress = 1 in // Can transform into LEA.
+  def ADD32rr  : I<0x01, MRMDestReg, (outs GR32:$dst),
+                                     (ins GR32:$src1, GR32:$src2),
+                   "add{l}\t{$src2, $dst|$dst, $src2}",
+                   [(set GR32:$dst, (add GR32:$src1, GR32:$src2))]>;
+
+This definition makes use of the custom class ``I`` (extended from the custom
+class ``X86Inst``), which is defined in the X86-specific TableGen file, to
+factor out the common features that instructions of its class share.  A key
+feature of TableGen is that it allows the end-user to define the abstractions
+they prefer to use when describing their information.
+
+Each ``def`` record has a special entry called "NAME".  This is the name of the
+record ("``ADD32rr``" above).  In the general case ``def`` names can be formed
+from various kinds of string processing expressions and ``NAME`` resolves to the
+final value obtained after resolving all of those expressions.  The user may
+refer to ``NAME`` anywhere she desires to use the ultimate name of the ``def``.
+``NAME`` should not be defined anywhere else in user code to avoid conflicts.
+
+Syntax
+======
+
+TableGen has a syntax that is loosely based on C++ templates, with built-in
+types and specification. In addition, TableGen's syntax introduces some
+automation concepts like multiclass, foreach, let, etc.
+
+Basic concepts
+--------------
+
+TableGen files consist of two key parts: 'classes' and 'definitions', both of
+which are considered 'records'.
+
+**TableGen records** have a unique name, a list of values, and a list of
+superclasses.  The list of values is the main data that TableGen builds for each
+record; it is this that holds the domain specific information for the
+application.  The interpretation of this data is left to a specific `backend`_,
+but the structure and format rules are taken care of and are fixed by
+TableGen.
+
+**TableGen definitions** are the concrete form of 'records'.  These generally do
+not have any undefined values, and are marked with the '``def``' keyword.
+
+.. code-block:: text
+
+  def FeatureFPARMv8 : SubtargetFeature<"fp-armv8", "HasFPARMv8", "true",
+                                        "Enable ARMv8 FP">;
+
+In this example, FeatureFPARMv8 is ``SubtargetFeature`` record initialised
+with some values. The names of the classes are defined via the
+keyword `class` either on the same file or some other included. Most target
+TableGen files include the generic ones in ``include/llvm/Target``.
+
+**TableGen classes** are abstract records that are used to build and describe
+other records.  These classes allow the end-user to build abstractions for
+either the domain they are targeting (such as "Register", "RegisterClass", and
+"Instruction" in the LLVM code generator) or for the implementor to help factor
+out common properties of records (such as "FPInst", which is used to represent
+floating point instructions in the X86 backend).  TableGen keeps track of all of
+the classes that are used to build up a definition, so the backend can find all
+definitions of a particular class, such as "Instruction".
+
+.. code-block:: text
+
+ class ProcNoItin<string Name, list<SubtargetFeature> Features>
+       : Processor<Name, NoItineraries, Features>;
+  
+Here, the class ProcNoItin, receiving parameters `Name` of type `string` and
+a list of target features is specializing the class Processor by passing the
+arguments down as well as hard-coding NoItineraries.
+
+**TableGen multiclasses** are groups of abstract records that are instantiated
+all at once.  Each instantiation can result in multiple TableGen definitions.
+If a multiclass inherits from another multiclass, the definitions in the
+sub-multiclass become part of the current multiclass, as if they were declared
+in the current multiclass.
+
+.. code-block:: text
+
+  multiclass ro_signed_pats<string T, string Rm, dag Base, dag Offset, dag Extend,
+                          dag address, ValueType sty> {
+  def : Pat<(i32 (!cast<SDNode>("sextload" # sty) address)),
+            (!cast<Instruction>("LDRS" # T # "w_" # Rm # "_RegOffset")
+              Base, Offset, Extend)>;
+
+  def : Pat<(i64 (!cast<SDNode>("sextload" # sty) address)),
+            (!cast<Instruction>("LDRS" # T # "x_" # Rm # "_RegOffset")
+              Base, Offset, Extend)>;
+  }
+
+  defm : ro_signed_pats<"B", Rm, Base, Offset, Extend,
+                        !foreach(decls.pattern, address,
+                                 !subst(SHIFT, imm_eq0, decls.pattern)),
+                        i8>;
+
+
+
+See the :doc:`TableGen Language Introduction <LangIntro>` for more generic
+information on the usage of the language, and the
+:doc:`TableGen Language Reference <LangRef>` for more in-depth description
+of the formal language specification.
+
+.. _backend:
+.. _backends:
+
+TableGen backends
+=================
+
+TableGen files have no real meaning without a back-end. The default operation
+of running ``llvm-tblgen`` is to print the information in a textual format, but
+that's only useful for debugging of the TableGen files themselves. The power
+in TableGen is, however, to interpret the source files into an internal 
+representation that can be generated into anything you want.
+
+Current usage of TableGen is to create huge include files with tables that you
+can either include directly (if the output is in the language you're coding),
+or be used in pre-processing via macros surrounding the include of the file.
+
+Direct output can be used if the back-end already prints a table in C format
+or if the output is just a list of strings (for error and warning messages).
+Pre-processed output should be used if the same information needs to be used
+in different contexts (like Instruction names), so your back-end should print
+a meta-information list that can be shaped into different compile-time formats.
+
+See the `TableGen BackEnds <BackEnds.html>`_ for more information.
+
+TableGen Deficiencies
+=====================
+
+Despite being very generic, TableGen has some deficiencies that have been
+pointed out numerous times. The common theme is that, while TableGen allows
+you to build Domain-Specific-Languages, the final languages that you create
+lack the power of other DSLs, which in turn increase considerably the size
+and complexity of TableGen files.
+
+At the same time, TableGen allows you to create virtually any meaning of
+the basic concepts via custom-made back-ends, which can pervert the original
+design and make it very hard for newcomers to understand the evil TableGen
+file.
+
+There are some in favour of extending the semantics even more, but making sure
+back-ends adhere to strict rules. Others are suggesting we should move to less,
+more powerful DSLs designed with specific purposes, or even re-using existing
+DSLs.
+
+Either way, this is a discussion that will likely span across several years,
+if not decades. You can read more in the `TableGen Deficiencies <Deficiencies.html>`_
+document.

Added: www-releases/trunk/4.0.0/docs/_sources/TableGenFundamentals.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TableGenFundamentals.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TableGenFundamentals.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TableGenFundamentals.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,10 @@
+=====================
+TableGen Fundamentals
+=====================
+
+Moved
+=====
+
+The TableGen fundamentals documentation has moved to a directory on its own
+and is now available at :doc:`TableGen/index`. Please, change your links to
+that page.

Added: www-releases/trunk/4.0.0/docs/_sources/TestSuiteMakefileGuide.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TestSuiteMakefileGuide.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TestSuiteMakefileGuide.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TestSuiteMakefileGuide.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,346 @@
+=====================
+LLVM test-suite Guide
+=====================
+
+.. contents::
+   :local:
+
+Overview
+========
+
+This document describes the features of the Makefile-based LLVM
+test-suite as well as the cmake based replacement. This way of interacting
+with the test-suite is deprecated in favor of running the test-suite using LNT,
+but may continue to prove useful for some users. See the Testing
+Guide's :ref:`test-suite Quickstart <test-suite-quickstart>` section for more
+information.
+
+Test suite Structure
+====================
+
+The ``test-suite`` module contains a number of programs that can be
+compiled with LLVM and executed. These programs are compiled using the
+native compiler and various LLVM backends. The output from the program
+compiled with the native compiler is assumed correct; the results from
+the other programs are compared to the native program output and pass if
+they match.
+
+When executing tests, it is usually a good idea to start out with a
+subset of the available tests or programs. This makes test run times
+smaller at first and later on this is useful to investigate individual
+test failures. To run some test only on a subset of programs, simply
+change directory to the programs you want tested and run ``gmake``
+there. Alternatively, you can run a different test using the ``TEST``
+variable to change what tests or run on the selected programs (see below
+for more info).
+
+In addition for testing correctness, the ``test-suite`` directory also
+performs timing tests of various LLVM optimizations. It also records
+compilation times for the compilers and the JIT. This information can be
+used to compare the effectiveness of LLVM's optimizations and code
+generation.
+
+``test-suite`` tests are divided into three types of tests: MultiSource,
+SingleSource, and External.
+
+-  ``test-suite/SingleSource``
+
+   The SingleSource directory contains test programs that are only a
+   single source file in size. These are usually small benchmark
+   programs or small programs that calculate a particular value. Several
+   such programs are grouped together in each directory.
+
+-  ``test-suite/MultiSource``
+
+   The MultiSource directory contains subdirectories which contain
+   entire programs with multiple source files. Large benchmarks and
+   whole applications go here.
+
+-  ``test-suite/External``
+
+   The External directory contains Makefiles for building code that is
+   external to (i.e., not distributed with) LLVM. The most prominent
+   members of this directory are the SPEC 95 and SPEC 2000 benchmark
+   suites. The ``External`` directory does not contain these actual
+   tests, but only the Makefiles that know how to properly compile these
+   programs from somewhere else. The presence and location of these
+   external programs is configured by the test-suite ``configure``
+   script.
+
+Each tree is then subdivided into several categories, including
+applications, benchmarks, regression tests, code that is strange
+grammatically, etc. These organizations should be relatively self
+explanatory.
+
+Some tests are known to fail. Some are bugs that we have not fixed yet;
+others are features that we haven't added yet (or may never add). In the
+regression tests, the result for such tests will be XFAIL (eXpected
+FAILure). In this way, you can tell the difference between an expected
+and unexpected failure.
+
+The tests in the test suite have no such feature at this time. If the
+test passes, only warnings and other miscellaneous output will be
+generated. If a test fails, a large <program> FAILED message will be
+displayed. This will help you separate benign warnings from actual test
+failures.
+
+Running the test suite via CMake
+================================
+
+To run the test suite, you need to use the following steps:
+
+#. The test suite uses the lit test runner to run the test-suite,
+   you need to have lit installed first.  Check out LLVM and install lit:
+   
+   .. code-block:: bash
+
+       % svn co http://llvm.org/svn/llvm-project/llvm/trunk llvm
+       % cd llvm/utils/lit
+       % sudo python setup.py install # Or without sudo, install in virtual-env.
+       running install
+       running bdist_egg
+       running egg_info
+       writing lit.egg-info/PKG-INFO
+       ...
+       % lit --version
+       lit 0.5.0dev
+
+#. Check out the ``test-suite`` module with:
+
+   .. code-block:: bash
+
+       % svn co http://llvm.org/svn/llvm-project/test-suite/trunk test-suite
+
+#. Use CMake to configure the test suite in a new directory. You cannot build
+   the test suite in the source tree.
+
+   .. code-block:: bash
+   
+       % mkdir test-suite-build
+       % cd test-suite-build
+       % cmake ../test-suite
+
+#. Build the benchmarks, using the makefiles CMake generated.
+
+.. code-block:: bash
+
+    % make
+    Scanning dependencies of target timeit-target
+    [  0%] Building C object tools/CMakeFiles/timeit-target.dir/timeit.c.o
+    [  0%] Linking C executable timeit-target
+    [  0%] Built target timeit-target
+    Scanning dependencies of target fpcmp-host
+    [  0%] [TEST_SUITE_HOST_CC] Building host executable fpcmp
+    [  0%] Built target fpcmp-host
+    Scanning dependencies of target timeit-host
+    [  0%] [TEST_SUITE_HOST_CC] Building host executable timeit
+    [  0%] Built target timeit-host
+
+    
+#. Run the tests with lit:
+
+.. code-block:: bash
+
+    % lit -v -j 1 . -o results.json
+    -- Testing: 474 tests, 1 threads --
+    PASS: test-suite :: MultiSource/Applications/ALAC/decode/alacconvert-decode.test (1 of 474)
+    ********** TEST 'test-suite :: MultiSource/Applications/ALAC/decode/alacconvert-decode.test' RESULTS **********
+    compile_time: 0.2192 
+    exec_time: 0.0462 
+    hash: "59620e187c6ac38b36382685ccd2b63b" 
+    size: 83348 
+    **********
+    PASS: test-suite :: MultiSource/Applications/ALAC/encode/alacconvert-encode.test (2 of 474)
+
+
+Running the test suite via Makefiles (deprecated)
+=================================================
+
+First, all tests are executed within the LLVM object directory tree.
+They *are not* executed inside of the LLVM source tree. This is because
+the test suite creates temporary files during execution.
+
+To run the test suite, you need to use the following steps:
+
+#. ``cd`` into the ``llvm/projects`` directory in your source tree.
+#. Check out the ``test-suite`` module with:
+
+   .. code-block:: bash
+
+       % svn co http://llvm.org/svn/llvm-project/test-suite/trunk test-suite
+
+   This will get the test suite into ``llvm/projects/test-suite``.
+
+#. Configure and build ``llvm``.
+
+#. Configure and build ``llvm-gcc``.
+
+#. Install ``llvm-gcc`` somewhere.
+
+#. *Re-configure* ``llvm`` from the top level of each build tree (LLVM
+   object directory tree) in which you want to run the test suite, just
+   as you do before building LLVM.
+
+   During the *re-configuration*, you must either: (1) have ``llvm-gcc``
+   you just built in your path, or (2) specify the directory where your
+   just-built ``llvm-gcc`` is installed using
+   ``--with-llvmgccdir=$LLVM_GCC_DIR``.
+
+   You must also tell the configure machinery that the test suite is
+   available so it can be configured for your build tree:
+
+   .. code-block:: bash
+
+       % cd $LLVM_OBJ_ROOT ; $LLVM_SRC_ROOT/configure [--with-llvmgccdir=$LLVM_GCC_DIR]
+
+   [Remember that ``$LLVM_GCC_DIR`` is the directory where you
+   *installed* llvm-gcc, not its src or obj directory.]
+
+#. You can now run the test suite from your build tree as follows:
+
+   .. code-block:: bash
+
+       % cd $LLVM_OBJ_ROOT/projects/test-suite
+       % make
+
+Note that the second and third steps only need to be done once. After
+you have the suite checked out and configured, you don't need to do it
+again (unless the test code or configure script changes).
+
+Configuring External Tests
+--------------------------
+
+In order to run the External tests in the ``test-suite`` module, you
+must specify *--with-externals*. This must be done during the
+*re-configuration* step (see above), and the ``llvm`` re-configuration
+must recognize the previously-built ``llvm-gcc``. If any of these is
+missing or neglected, the External tests won't work.
+
+* *--with-externals*
+
+* *--with-externals=<directory>*
+
+This tells LLVM where to find any external tests. They are expected to
+be in specifically named subdirectories of <``directory``>. If
+``directory`` is left unspecified, ``configure`` uses the default value
+``/home/vadve/shared/benchmarks/speccpu2000/benchspec``. Subdirectory
+names known to LLVM include:
+
+* spec95
+
+* speccpu2000
+
+* speccpu2006
+
+* povray31
+
+Others are added from time to time, and can be determined from
+``configure``.
+
+Running different tests
+-----------------------
+
+In addition to the regular "whole program" tests, the ``test-suite``
+module also provides a mechanism for compiling the programs in different
+ways. If the variable TEST is defined on the ``gmake`` command line, the
+test system will include a Makefile named
+``TEST.<value of TEST variable>.Makefile``. This Makefile can modify
+build rules to yield different results.
+
+For example, the LLVM nightly tester uses ``TEST.nightly.Makefile`` to
+create the nightly test reports. To run the nightly tests, run
+``gmake TEST=nightly``.
+
+There are several TEST Makefiles available in the tree. Some of them are
+designed for internal LLVM research and will not work outside of the
+LLVM research group. They may still be valuable, however, as a guide to
+writing your own TEST Makefile for any optimization or analysis passes
+that you develop with LLVM.
+
+Generating test output
+----------------------
+
+There are a number of ways to run the tests and generate output. The
+most simple one is simply running ``gmake`` with no arguments. This will
+compile and run all programs in the tree using a number of different
+methods and compare results. Any failures are reported in the output,
+but are likely drowned in the other output. Passes are not reported
+explicitly.
+
+Somewhat better is running ``gmake TEST=sometest test``, which runs the
+specified test and usually adds per-program summaries to the output
+(depending on which sometest you use). For example, the ``nightly`` test
+explicitly outputs TEST-PASS or TEST-FAIL for every test after each
+program. Though these lines are still drowned in the output, it's easy
+to grep the output logs in the Output directories.
+
+Even better are the ``report`` and ``report.format`` targets (where
+``format`` is one of ``html``, ``csv``, ``text`` or ``graphs``). The
+exact contents of the report are dependent on which ``TEST`` you are
+running, but the text results are always shown at the end of the run and
+the results are always stored in the ``report.<type>.format`` file (when
+running with ``TEST=<type>``). The ``report`` also generate a file
+called ``report.<type>.raw.out`` containing the output of the entire
+test run.
+
+Writing custom tests for the test suite
+---------------------------------------
+
+Assuming you can run the test suite, (e.g.
+"``gmake TEST=nightly report``" should work), it is really easy to run
+optimizations or code generator components against every program in the
+tree, collecting statistics or running custom checks for correctness. At
+base, this is how the nightly tester works, it's just one example of a
+general framework.
+
+Lets say that you have an LLVM optimization pass, and you want to see
+how many times it triggers. First thing you should do is add an LLVM
+`statistic <ProgrammersManual.html#Statistic>`_ to your pass, which will
+tally counts of things you care about.
+
+Following this, you can set up a test and a report that collects these
+and formats them for easy viewing. This consists of two files, a
+"``test-suite/TEST.XXX.Makefile``" fragment (where XXX is the name of
+your test) and a "``test-suite/TEST.XXX.report``" file that indicates
+how to format the output into a table. There are many example reports of
+various levels of sophistication included with the test suite, and the
+framework is very general.
+
+If you are interested in testing an optimization pass, check out the
+"libcalls" test as an example. It can be run like this:
+
+.. code-block:: bash
+
+    % cd llvm/projects/test-suite/MultiSource/Benchmarks  # or some other level
+    % make TEST=libcalls report
+
+This will do a bunch of stuff, then eventually print a table like this:
+
+::
+
+    Name                                  | total | #exit |
+    ...
+    FreeBench/analyzer/analyzer           | 51    | 6     |
+    FreeBench/fourinarow/fourinarow       | 1     | 1     |
+    FreeBench/neural/neural               | 19    | 9     |
+    FreeBench/pifft/pifft                 | 5     | 3     |
+    MallocBench/cfrac/cfrac               | 1     | *     |
+    MallocBench/espresso/espresso         | 52    | 12    |
+    MallocBench/gs/gs                     | 4     | *     |
+    Prolangs-C/TimberWolfMC/timberwolfmc  | 302   | *     |
+    Prolangs-C/agrep/agrep                | 33    | 12    |
+    Prolangs-C/allroots/allroots          | *     | *     |
+    Prolangs-C/assembler/assembler        | 47    | *     |
+    Prolangs-C/bison/mybison              | 74    | *     |
+    ...
+
+This basically is grepping the -stats output and displaying it in a
+table. You can also use the "TEST=libcalls report.html" target to get
+the table in HTML form, similarly for report.csv and report.tex.
+
+The source for this is in ``test-suite/TEST.libcalls.*``. The format is
+pretty simple: the Makefile indicates how to run the test (in this case,
+"``opt -simplify-libcalls -stats``"), and the report contains one line
+for each column of the output. The first value is the header for the
+column and the second is the regex to grep the output of the command
+for. There are lots of example reports that can do fancy stuff.

Added: www-releases/trunk/4.0.0/docs/_sources/TestingGuide.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TestingGuide.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TestingGuide.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TestingGuide.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,619 @@
+=================================
+LLVM Testing Infrastructure Guide
+=================================
+
+.. contents::
+   :local:
+
+.. toctree::
+   :hidden:
+
+   TestSuiteMakefileGuide
+
+Overview
+========
+
+This document is the reference manual for the LLVM testing
+infrastructure. It documents the structure of the LLVM testing
+infrastructure, the tools needed to use it, and how to add and run
+tests.
+
+Requirements
+============
+
+In order to use the LLVM testing infrastructure, you will need all of the
+software required to build LLVM, as well as `Python <http://python.org>`_ 2.7 or
+later.
+
+If you intend to run the :ref:`test-suite <test-suite-overview>`, you will also
+need a development version of zlib (zlib1g-dev is known to work on several Linux
+distributions).
+
+LLVM testing infrastructure organization
+========================================
+
+The LLVM testing infrastructure contains two major categories of tests:
+regression tests and whole programs. The regression tests are contained
+inside the LLVM repository itself under ``llvm/test`` and are expected
+to always pass -- they should be run before every commit.
+
+The whole programs tests are referred to as the "LLVM test suite" (or
+"test-suite") and are in the ``test-suite`` module in subversion. For
+historical reasons, these tests are also referred to as the "nightly
+tests" in places, which is less ambiguous than "test-suite" and remains
+in use although we run them much more often than nightly.
+
+Regression tests
+----------------
+
+The regression tests are small pieces of code that test a specific
+feature of LLVM or trigger a specific bug in LLVM. The language they are
+written in depends on the part of LLVM being tested. These tests are driven by
+the :doc:`Lit <CommandGuide/lit>` testing tool (which is part of LLVM), and
+are located in the ``llvm/test`` directory.
+
+Typically when a bug is found in LLVM, a regression test containing just
+enough code to reproduce the problem should be written and placed
+somewhere underneath this directory. For example, it can be a small
+piece of LLVM IR distilled from an actual application or benchmark.
+
+``test-suite``
+--------------
+
+The test suite contains whole programs, which are pieces of code which
+can be compiled and linked into a stand-alone program that can be
+executed. These programs are generally written in high level languages
+such as C or C++.
+
+These programs are compiled using a user specified compiler and set of
+flags, and then executed to capture the program output and timing
+information. The output of these programs is compared to a reference
+output to ensure that the program is being compiled correctly.
+
+In addition to compiling and executing programs, whole program tests
+serve as a way of benchmarking LLVM performance, both in terms of the
+efficiency of the programs generated as well as the speed with which
+LLVM compiles, optimizes, and generates code.
+
+The test-suite is located in the ``test-suite`` Subversion module.
+
+Debugging Information tests
+---------------------------
+
+The test suite contains tests to check quality of debugging information.
+The test are written in C based languages or in LLVM assembly language.
+
+These tests are compiled and run under a debugger. The debugger output
+is checked to validate of debugging information. See README.txt in the
+test suite for more information . This test suite is located in the
+``debuginfo-tests`` Subversion module.
+
+Quick start
+===========
+
+The tests are located in two separate Subversion modules. The
+regressions tests are in the main "llvm" module under the directory
+``llvm/test`` (so you get these tests for free with the main LLVM tree).
+Use ``make check-all`` to run the regression tests after building LLVM.
+
+The more comprehensive test suite that includes whole programs in C and C++
+is in the ``test-suite`` module. See :ref:`test-suite Quickstart
+<test-suite-quickstart>` for more information on running these tests.
+
+Regression tests
+----------------
+
+To run all of the LLVM regression tests use the check-llvm target:
+
+.. code-block:: bash
+
+    % make check-llvm
+
+If you have `Clang <http://clang.llvm.org/>`_ checked out and built, you
+can run the LLVM and Clang tests simultaneously using:
+
+.. code-block:: bash
+
+    % make check-all
+
+To run the tests with Valgrind (Memcheck by default), use the ``LIT_ARGS`` make
+variable to pass the required options to lit. For example, you can use:
+
+.. code-block:: bash
+
+    % make check LIT_ARGS="-v --vg --vg-leak"
+
+to enable testing with valgrind and with leak checking enabled.
+
+To run individual tests or subsets of tests, you can use the ``llvm-lit``
+script which is built as part of LLVM. For example, to run the
+``Integer/BitPacked.ll`` test by itself you can run:
+
+.. code-block:: bash
+
+    % llvm-lit ~/llvm/test/Integer/BitPacked.ll 
+
+or to run all of the ARM CodeGen tests:
+
+.. code-block:: bash
+
+    % llvm-lit ~/llvm/test/CodeGen/ARM
+
+For more information on using the :program:`lit` tool, see ``llvm-lit --help``
+or the :doc:`lit man page <CommandGuide/lit>`.
+
+Debugging Information tests
+---------------------------
+
+To run debugging information tests simply checkout the tests inside
+clang/test directory.
+
+.. code-block:: bash
+
+    % cd clang/test
+    % svn co http://llvm.org/svn/llvm-project/debuginfo-tests/trunk debuginfo-tests
+
+These tests are already set up to run as part of clang regression tests.
+
+Regression test structure
+=========================
+
+The LLVM regression tests are driven by :program:`lit` and are located in the
+``llvm/test`` directory.
+
+This directory contains a large array of small tests that exercise
+various features of LLVM and to ensure that regressions do not occur.
+The directory is broken into several sub-directories, each focused on a
+particular area of LLVM.
+
+Writing new regression tests
+----------------------------
+
+The regression test structure is very simple, but does require some
+information to be set. This information is gathered via ``configure``
+and is written to a file, ``test/lit.site.cfg`` in the build directory.
+The ``llvm/test`` Makefile does this work for you.
+
+In order for the regression tests to work, each directory of tests must
+have a ``lit.local.cfg`` file. :program:`lit` looks for this file to determine
+how to run the tests. This file is just Python code and thus is very
+flexible, but we've standardized it for the LLVM regression tests. If
+you're adding a directory of tests, just copy ``lit.local.cfg`` from
+another directory to get running. The standard ``lit.local.cfg`` simply
+specifies which files to look in for tests. Any directory that contains
+only directories does not need the ``lit.local.cfg`` file. Read the :doc:`Lit
+documentation <CommandGuide/lit>` for more information.
+
+Each test file must contain lines starting with "RUN:" that tell :program:`lit`
+how to run it. If there are no RUN lines, :program:`lit` will issue an error
+while running a test.
+
+RUN lines are specified in the comments of the test program using the
+keyword ``RUN`` followed by a colon, and lastly the command (pipeline)
+to execute. Together, these lines form the "script" that :program:`lit`
+executes to run the test case. The syntax of the RUN lines is similar to a
+shell's syntax for pipelines including I/O redirection and variable
+substitution. However, even though these lines may *look* like a shell
+script, they are not. RUN lines are interpreted by :program:`lit`.
+Consequently, the syntax differs from shell in a few ways. You can specify
+as many RUN lines as needed.
+
+:program:`lit` performs substitution on each RUN line to replace LLVM tool names
+with the full paths to the executable built for each tool (in
+``$(LLVM_OBJ_ROOT)/$(BuildMode)/bin)``. This ensures that :program:`lit` does
+not invoke any stray LLVM tools in the user's path during testing.
+
+Each RUN line is executed on its own, distinct from other lines unless
+its last character is ``\``. This continuation character causes the RUN
+line to be concatenated with the next one. In this way you can build up
+long pipelines of commands without making huge line lengths. The lines
+ending in ``\`` are concatenated until a RUN line that doesn't end in
+``\`` is found. This concatenated set of RUN lines then constitutes one
+execution. :program:`lit` will substitute variables and arrange for the pipeline
+to be executed. If any process in the pipeline fails, the entire line (and
+test case) fails too.
+
+Below is an example of legal RUN lines in a ``.ll`` file:
+
+.. code-block:: llvm
+
+    ; RUN: llvm-as < %s | llvm-dis > %t1
+    ; RUN: llvm-dis < %s.bc-13 > %t2
+    ; RUN: diff %t1 %t2
+
+As with a Unix shell, the RUN lines permit pipelines and I/O
+redirection to be used.
+
+There are some quoting rules that you must pay attention to when writing
+your RUN lines. In general nothing needs to be quoted. :program:`lit` won't
+strip off any quote characters so they will get passed to the invoked program.
+To avoid this use curly braces to tell :program:`lit` that it should treat
+everything enclosed as one value.
+
+In general, you should strive to keep your RUN lines as simple as possible,
+using them only to run tools that generate textual output you can then examine.
+The recommended way to examine output to figure out if the test passes is using
+the :doc:`FileCheck tool <CommandGuide/FileCheck>`. *[The usage of grep in RUN
+lines is deprecated - please do not send or commit patches that use it.]*
+
+Put related tests into a single file rather than having a separate file per
+test. Check if there are files already covering your feature and consider
+adding your code there instead of creating a new file.
+
+Extra files
+-----------
+
+If your test requires extra files besides the file containing the ``RUN:``
+lines, the idiomatic place to put them is in a subdirectory ``Inputs``.
+You can then refer to the extra files as ``%S/Inputs/foo.bar``.
+
+For example, consider ``test/Linker/ident.ll``. The directory structure is
+as follows::
+
+  test/
+    Linker/
+      ident.ll
+      Inputs/
+        ident.a.ll
+        ident.b.ll
+
+For convenience, these are the contents:
+
+.. code-block:: llvm
+
+  ;;;;; ident.ll:
+
+  ; RUN: llvm-link %S/Inputs/ident.a.ll %S/Inputs/ident.b.ll -S | FileCheck %s
+
+  ; Verify that multiple input llvm.ident metadata are linked together.
+
+  ; CHECK-DAG: !llvm.ident = !{!0, !1, !2}
+  ; CHECK-DAG: "Compiler V1"
+  ; CHECK-DAG: "Compiler V2"
+  ; CHECK-DAG: "Compiler V3"
+
+  ;;;;; Inputs/ident.a.ll:
+
+  !llvm.ident = !{!0, !1}
+  !0 = metadata !{metadata !"Compiler V1"}
+  !1 = metadata !{metadata !"Compiler V2"}
+
+  ;;;;; Inputs/ident.b.ll:
+
+  !llvm.ident = !{!0}
+  !0 = metadata !{metadata !"Compiler V3"}
+
+For symmetry reasons, ``ident.ll`` is just a dummy file that doesn't
+actually participate in the test besides holding the ``RUN:`` lines.
+
+.. note::
+
+  Some existing tests use ``RUN: true`` in extra files instead of just
+  putting the extra files in an ``Inputs/`` directory. This pattern is
+  deprecated.
+
+Fragile tests
+-------------
+
+It is easy to write a fragile test that would fail spuriously if the tool being
+tested outputs a full path to the input file.  For example, :program:`opt` by
+default outputs a ``ModuleID``:
+
+.. code-block:: console
+
+  $ cat example.ll
+  define i32 @main() nounwind {
+      ret i32 0
+  }
+
+  $ opt -S /path/to/example.ll
+  ; ModuleID = '/path/to/example.ll'
+
+  define i32 @main() nounwind {
+      ret i32 0
+  }
+
+``ModuleID`` can unexpetedly match against ``CHECK`` lines.  For example:
+
+.. code-block:: llvm
+
+  ; RUN: opt -S %s | FileCheck
+
+  define i32 @main() nounwind {
+      ; CHECK-NOT: load
+      ret i32 0
+  }
+
+This test will fail if placed into a ``download`` directory.
+
+To make your tests robust, always use ``opt ... < %s`` in the RUN line.
+:program:`opt` does not output a ``ModuleID`` when input comes from stdin.
+
+Platform-Specific Tests
+-----------------------
+
+Whenever adding tests that require the knowledge of a specific platform,
+either related to code generated, specific output or back-end features,
+you must make sure to isolate the features, so that buildbots that
+run on different architectures (and don't even compile all back-ends),
+don't fail.
+
+The first problem is to check for target-specific output, for example sizes
+of structures, paths and architecture names, for example:
+
+* Tests containing Windows paths will fail on Linux and vice-versa.
+* Tests that check for ``x86_64`` somewhere in the text will fail anywhere else.
+* Tests where the debug information calculates the size of types and structures.
+
+Also, if the test rely on any behaviour that is coded in any back-end, it must
+go in its own directory. So, for instance, code generator tests for ARM go
+into ``test/CodeGen/ARM`` and so on. Those directories contain a special
+``lit`` configuration file that ensure all tests in that directory will
+only run if a specific back-end is compiled and available.
+
+For instance, on ``test/CodeGen/ARM``, the ``lit.local.cfg`` is:
+
+.. code-block:: python
+
+  config.suffixes = ['.ll', '.c', '.cpp', '.test']
+  if not 'ARM' in config.root.targets:
+    config.unsupported = True
+
+Other platform-specific tests are those that depend on a specific feature
+of a specific sub-architecture, for example only to Intel chips that support ``AVX2``.
+
+For instance, ``test/CodeGen/X86/psubus.ll`` tests three sub-architecture
+variants:
+
+.. code-block:: llvm
+
+  ; RUN: llc -mcpu=core2 < %s | FileCheck %s -check-prefix=SSE2
+  ; RUN: llc -mcpu=corei7-avx < %s | FileCheck %s -check-prefix=AVX1
+  ; RUN: llc -mcpu=core-avx2 < %s | FileCheck %s -check-prefix=AVX2
+
+And the checks are different:
+
+.. code-block:: llvm
+
+  ; SSE2: @test1
+  ; SSE2: psubusw LCPI0_0(%rip), %xmm0
+  ; AVX1: @test1
+  ; AVX1: vpsubusw LCPI0_0(%rip), %xmm0, %xmm0
+  ; AVX2: @test1
+  ; AVX2: vpsubusw LCPI0_0(%rip), %xmm0, %xmm0
+
+So, if you're testing for a behaviour that you know is platform-specific or
+depends on special features of sub-architectures, you must add the specific
+triple, test with the specific FileCheck and put it into the specific
+directory that will filter out all other architectures.
+
+REQUIRES and REQUIRES-ANY directive
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Some tests can be enabled only in specific situation - like having
+debug build. Use ``REQUIRES`` directive to specify those requirements.
+
+.. code-block:: llvm
+
+    ; This test will be only enabled in the build with asserts
+    ; REQUIRES: asserts
+
+You can separate requirements by a comma.
+``REQUIRES`` means all listed requirements must be satisfied.
+``REQUIRES-ANY`` means at least one must be satisfied.
+
+List of features that can be used in ``REQUIRES`` and ``REQUIRES-ANY`` can be
+found in lit.cfg files.
+
+Substitutions
+-------------
+
+Besides replacing LLVM tool names the following substitutions are performed in
+RUN lines:
+
+``%%``
+   Replaced by a single ``%``. This allows escaping other substitutions.
+
+``%s``
+   File path to the test case's source. This is suitable for passing on the
+   command line as the input to an LLVM tool.
+
+   Example: ``/home/user/llvm/test/MC/ELF/foo_test.s``
+
+``%S``
+   Directory path to the test case's source.
+
+   Example: ``/home/user/llvm/test/MC/ELF``
+
+``%t``
+   File path to a temporary file name that could be used for this test case.
+   The file name won't conflict with other test cases. You can append to it
+   if you need multiple temporaries. This is useful as the destination of
+   some redirected output.
+
+   Example: ``/home/user/llvm.build/test/MC/ELF/Output/foo_test.s.tmp``
+
+``%T``
+   Directory of ``%t``.
+
+   Example: ``/home/user/llvm.build/test/MC/ELF/Output``
+
+``%{pathsep}``
+
+   Expands to the path separator, i.e. ``:`` (or ``;`` on Windows).
+
+
+**LLVM-specific substitutions:**
+
+``%shlibext``
+   The suffix for the host platforms shared library files. This includes the
+   period as the first character.
+
+   Example: ``.so`` (Linux), ``.dylib`` (OS X), ``.dll`` (Windows)
+
+``%exeext``
+   The suffix for the host platforms executable files. This includes the
+   period as the first character.
+
+   Example: ``.exe`` (Windows), empty on Linux.
+
+``%(line)``, ``%(line+<number>)``, ``%(line-<number>)``
+   The number of the line where this substitution is used, with an optional
+   integer offset. This can be used in tests with multiple RUN lines, which
+   reference test file's line numbers.
+
+
+**Clang-specific substitutions:**
+
+``%clang``
+   Invokes the Clang driver.
+
+``%clang_cpp``
+   Invokes the Clang driver for C++.
+
+``%clang_cl``
+   Invokes the CL-compatible Clang driver.
+
+``%clangxx``
+   Invokes the G++-compatible Clang driver.
+
+``%clang_cc1``
+   Invokes the Clang frontend.
+
+``%itanium_abi_triple``, ``%ms_abi_triple``
+   These substitutions can be used to get the current target triple adjusted to
+   the desired ABI. For example, if the test suite is running with the
+   ``i686-pc-win32`` target, ``%itanium_abi_triple`` will expand to
+   ``i686-pc-mingw32``. This allows a test to run with a specific ABI without
+   constraining it to a specific triple.
+
+To add more substituations, look at ``test/lit.cfg`` or ``lit.local.cfg``.
+
+
+Options
+-------
+
+The llvm lit configuration allows to customize some things with user options:
+
+``llc``, ``opt``, ...
+    Substitute the respective llvm tool name with a custom command line. This
+    allows to specify custom paths and default arguments for these tools.
+    Example:
+
+    % llvm-lit "-Dllc=llc -verify-machineinstrs"
+
+``run_long_tests``
+    Enable the execution of long running tests.
+
+``llvm_site_config``
+    Load the specified lit configuration instead of the default one.
+
+
+Other Features
+--------------
+
+To make RUN line writing easier, there are several helper programs. These
+helpers are in the PATH when running tests, so you can just call them using
+their name. For example:
+
+``not``
+   This program runs its arguments and then inverts the result code from it.
+   Zero result codes become 1. Non-zero result codes become 0.
+
+Sometimes it is necessary to mark a test case as "expected fail" or
+XFAIL. You can easily mark a test as XFAIL just by including ``XFAIL:``
+on a line near the top of the file. This signals that the test case
+should succeed if the test fails. Such test cases are counted separately
+by the testing tool. To specify an expected fail, use the XFAIL keyword
+in the comments of the test program followed by a colon and one or more
+failure patterns. Each failure pattern can be either ``*`` (to specify
+fail everywhere), or a part of a target triple (indicating the test
+should fail on that platform), or the name of a configurable feature
+(for example, ``loadable_module``). If there is a match, the test is
+expected to fail. If not, the test is expected to succeed. To XFAIL
+everywhere just specify ``XFAIL: *``. Here is an example of an ``XFAIL``
+line:
+
+.. code-block:: llvm
+
+    ; XFAIL: darwin,sun
+
+To make the output more useful, :program:`lit` will scan
+the lines of the test case for ones that contain a pattern that matches
+``PR[0-9]+``. This is the syntax for specifying a PR (Problem Report) number
+that is related to the test case. The number after "PR" specifies the
+LLVM bugzilla number. When a PR number is specified, it will be used in
+the pass/fail reporting. This is useful to quickly get some context when
+a test fails.
+
+Finally, any line that contains "END." will cause the special
+interpretation of lines to terminate. This is generally done right after
+the last RUN: line. This has two side effects:
+
+(a) it prevents special interpretation of lines that are part of the test
+    program, not the instructions to the test case, and
+
+(b) it speeds things up for really big test cases by avoiding
+    interpretation of the remainder of the file.
+
+.. _test-suite-overview:
+
+``test-suite`` Overview
+=======================
+
+The ``test-suite`` module contains a number of programs that can be
+compiled and executed. The ``test-suite`` includes reference outputs for
+all of the programs, so that the output of the executed program can be
+checked for correctness.
+
+``test-suite`` tests are divided into three types of tests: MultiSource,
+SingleSource, and External.
+
+-  ``test-suite/SingleSource``
+
+   The SingleSource directory contains test programs that are only a
+   single source file in size. These are usually small benchmark
+   programs or small programs that calculate a particular value. Several
+   such programs are grouped together in each directory.
+
+-  ``test-suite/MultiSource``
+
+   The MultiSource directory contains subdirectories which contain
+   entire programs with multiple source files. Large benchmarks and
+   whole applications go here.
+
+-  ``test-suite/External``
+
+   The External directory contains Makefiles for building code that is
+   external to (i.e., not distributed with) LLVM. The most prominent
+   members of this directory are the SPEC 95 and SPEC 2000 benchmark
+   suites. The ``External`` directory does not contain these actual
+   tests, but only the Makefiles that know how to properly compile these
+   programs from somewhere else. When using ``LNT``, use the
+   ``--test-externals`` option to include these tests in the results.
+
+.. _test-suite-quickstart:
+
+``test-suite`` Quickstart
+-------------------------
+
+The modern way of running the ``test-suite`` is focused on testing and
+benchmarking complete compilers using the
+`LNT <http://llvm.org/docs/lnt>`_ testing infrastructure.
+
+For more information on using LNT to execute the ``test-suite``, please
+see the `LNT Quickstart <http://llvm.org/docs/lnt/quickstart.html>`_
+documentation.
+
+``test-suite`` Makefiles
+------------------------
+
+Historically, the ``test-suite`` was executed using a complicated setup
+of Makefiles. The LNT based approach above is recommended for most
+users, but there are some testing scenarios which are not supported by
+the LNT approach. In addition, LNT currently uses the Makefile setup
+under the covers and so developers who are interested in how LNT works
+under the hood may want to understand the Makefile based setup.
+
+For more information on the ``test-suite`` Makefile setup, please see
+the :doc:`Test Suite Makefile Guide <TestSuiteMakefileGuide>`.

Added: www-releases/trunk/4.0.0/docs/_sources/TypeMetadata.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/TypeMetadata.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/TypeMetadata.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/TypeMetadata.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,226 @@
+=============
+Type Metadata
+=============
+
+Type metadata is a mechanism that allows IR modules to co-operatively build
+pointer sets corresponding to addresses within a given set of globals. LLVM's
+`control flow integrity`_ implementation uses this metadata to efficiently
+check (at each call site) that a given address corresponds to either a
+valid vtable or function pointer for a given class or function type, and its
+whole-program devirtualization pass uses the metadata to identify potential
+callees for a given virtual call.
+
+To use the mechanism, a client creates metadata nodes with two elements:
+
+1. a byte offset into the global (generally zero for functions)
+2. a metadata object representing an identifier for the type
+
+These metadata nodes are associated with globals by using global object
+metadata attachments with the ``!type`` metadata kind.
+
+Each type identifier must exclusively identify either global variables
+or functions.
+
+.. admonition:: Limitation
+
+  The current implementation only supports attaching metadata to functions on
+  the x86-32 and x86-64 architectures.
+
+An intrinsic, :ref:`llvm.type.test <type.test>`, is used to test whether a
+given pointer is associated with a type identifier.
+
+.. _control flow integrity: http://clang.llvm.org/docs/ControlFlowIntegrity.html
+
+Representing Type Information using Type Metadata
+=================================================
+
+This section describes how Clang represents C++ type information associated with
+virtual tables using type metadata.
+
+Consider the following inheritance hierarchy:
+
+.. code-block:: c++
+
+  struct A {
+    virtual void f();
+  };
+
+  struct B : A {
+    virtual void f();
+    virtual void g();
+  };
+
+  struct C {
+    virtual void h();
+  };
+
+  struct D : A, C {
+    virtual void f();
+    virtual void h();
+  };
+
+The virtual table objects for A, B, C and D look like this (under the Itanium ABI):
+
+.. csv-table:: Virtual Table Layout for A, B, C, D
+  :header: Class, 0, 1, 2, 3, 4, 5, 6
+
+  A, A::offset-to-top, &A::rtti, &A::f
+  B, B::offset-to-top, &B::rtti, &B::f, &B::g
+  C, C::offset-to-top, &C::rtti, &C::h
+  D, D::offset-to-top, &D::rtti, &D::f, &D::h, D::offset-to-top, &D::rtti, thunk for &D::h
+
+When an object of type A is constructed, the address of ``&A::f`` in A's
+virtual table object is stored in the object's vtable pointer.  In ABI parlance
+this address is known as an `address point`_. Similarly, when an object of type
+B is constructed, the address of ``&B::f`` is stored in the vtable pointer. In
+this way, the vtable in B's virtual table object is compatible with A's vtable.
+
+D is a little more complicated, due to the use of multiple inheritance. Its
+virtual table object contains two vtables, one compatible with A's vtable and
+the other compatible with C's vtable. Objects of type D contain two virtual
+pointers, one belonging to the A subobject and containing the address of
+the vtable compatible with A's vtable, and the other belonging to the C
+subobject and containing the address of the vtable compatible with C's vtable.
+
+The full set of compatibility information for the above class hierarchy is
+shown below. The following table shows the name of a class, the offset of an
+address point within that class's vtable and the name of one of the classes
+with which that address point is compatible.
+
+.. csv-table:: Type Offsets for A, B, C, D
+  :header: VTable for, Offset, Compatible Class
+
+  A, 16, A
+  B, 16, A
+   ,   , B
+  C, 16, C
+  D, 16, A
+   ,   , D
+   , 48, C
+
+The next step is to encode this compatibility information into the IR. The way
+this is done is to create type metadata named after each of the compatible
+classes, with which we associate each of the compatible address points in
+each vtable. For example, these type metadata entries encode the compatibility
+information for the above hierarchy:
+
+::
+
+  @_ZTV1A = constant [...], !type !0
+  @_ZTV1B = constant [...], !type !0, !type !1
+  @_ZTV1C = constant [...], !type !2
+  @_ZTV1D = constant [...], !type !0, !type !3, !type !4
+
+  !0 = !{i64 16, !"_ZTS1A"}
+  !1 = !{i64 16, !"_ZTS1B"}
+  !2 = !{i64 16, !"_ZTS1C"}
+  !3 = !{i64 16, !"_ZTS1D"}
+  !4 = !{i64 48, !"_ZTS1C"}
+
+With this type metadata, we can now use the ``llvm.type.test`` intrinsic to
+test whether a given pointer is compatible with a type identifier. Working
+backwards, if ``llvm.type.test`` returns true for a particular pointer,
+we can also statically determine the identities of the virtual functions
+that a particular virtual call may call. For example, if a program assumes
+a pointer to be a member of ``!"_ZST1A"``, we know that the address can
+be only be one of ``_ZTV1A+16``, ``_ZTV1B+16`` or ``_ZTV1D+16`` (i.e. the
+address points of the vtables of A, B and D respectively). If we then load
+an address from that pointer, we know that the address can only be one of
+``&A::f``, ``&B::f`` or ``&D::f``.
+
+.. _address point: https://mentorembedded.github.io/cxx-abi/abi.html#vtable-general
+
+Testing Addresses For Type Membership
+=====================================
+
+If a program tests an address using ``llvm.type.test``, this will cause
+a link-time optimization pass, ``LowerTypeTests``, to replace calls to this
+intrinsic with efficient code to perform type member tests. At a high level,
+the pass will lay out referenced globals in a consecutive memory region in
+the object file, construct bit vectors that map onto that memory region,
+and generate code at each of the ``llvm.type.test`` call sites to test
+pointers against those bit vectors. Because of the layout manipulation, the
+globals' definitions must be available at LTO time. For more information,
+see the `control flow integrity design document`_.
+
+A type identifier that identifies functions is transformed into a jump table,
+which is a block of code consisting of one branch instruction for each
+of the functions associated with the type identifier that branches to the
+target function. The pass will redirect any taken function addresses to the
+corresponding jump table entry. In the object file's symbol table, the jump
+table entries take the identities of the original functions, so that addresses
+taken outside the module will pass any verification done inside the module.
+
+Jump tables may call external functions, so their definitions need not
+be available at LTO time. Note that if an externally defined function is
+associated with a type identifier, there is no guarantee that its identity
+within the module will be the same as its identity outside of the module,
+as the former will be the jump table entry if a jump table is necessary.
+
+The `GlobalLayoutBuilder`_ class is responsible for laying out the globals
+efficiently to minimize the sizes of the underlying bitsets.
+
+.. _control flow integrity design document: http://clang.llvm.org/docs/ControlFlowIntegrityDesign.html
+
+:Example:
+
+::
+
+    target datalayout = "e-p:32:32"
+
+    @a = internal global i32 0, !type !0
+    @b = internal global i32 0, !type !0, !type !1
+    @c = internal global i32 0, !type !1
+    @d = internal global [2 x i32] [i32 0, i32 0], !type !2
+
+    define void @e() !type !3 {
+      ret void
+    }
+
+    define void @f() {
+      ret void
+    }
+
+    declare void @g() !type !3
+
+    !0 = !{i32 0, !"typeid1"}
+    !1 = !{i32 0, !"typeid2"}
+    !2 = !{i32 4, !"typeid2"}
+    !3 = !{i32 0, !"typeid3"}
+
+    declare i1 @llvm.type.test(i8* %ptr, metadata %typeid) nounwind readnone
+
+    define i1 @foo(i32* %p) {
+      %pi8 = bitcast i32* %p to i8*
+      %x = call i1 @llvm.type.test(i8* %pi8, metadata !"typeid1")
+      ret i1 %x
+    }
+
+    define i1 @bar(i32* %p) {
+      %pi8 = bitcast i32* %p to i8*
+      %x = call i1 @llvm.type.test(i8* %pi8, metadata !"typeid2")
+      ret i1 %x
+    }
+
+    define i1 @baz(void ()* %p) {
+      %pi8 = bitcast void ()* %p to i8*
+      %x = call i1 @llvm.type.test(i8* %pi8, metadata !"typeid3")
+      ret i1 %x
+    }
+
+    define void @main() {
+      %a1 = call i1 @foo(i32* @a) ; returns 1
+      %b1 = call i1 @foo(i32* @b) ; returns 1
+      %c1 = call i1 @foo(i32* @c) ; returns 0
+      %a2 = call i1 @bar(i32* @a) ; returns 0
+      %b2 = call i1 @bar(i32* @b) ; returns 1
+      %c2 = call i1 @bar(i32* @c) ; returns 1
+      %d02 = call i1 @bar(i32* getelementptr ([2 x i32]* @d, i32 0, i32 0)) ; returns 0
+      %d12 = call i1 @bar(i32* getelementptr ([2 x i32]* @d, i32 0, i32 1)) ; returns 1
+      %e = call i1 @baz(void ()* @e) ; returns 1
+      %f = call i1 @baz(void ()* @f) ; returns 0
+      %g = call i1 @baz(void ()* @g) ; returns 1
+      ret void
+    }
+
+.. _GlobalLayoutBuilder: http://llvm.org/klaus/llvm/blob/master/include/llvm/Transforms/IPO/LowerTypeTests.h

Added: www-releases/trunk/4.0.0/docs/_sources/Vectorizers.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/Vectorizers.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/Vectorizers.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/Vectorizers.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,426 @@
+==========================
+Auto-Vectorization in LLVM
+==========================
+
+.. contents::
+   :local:
+
+LLVM has two vectorizers: The :ref:`Loop Vectorizer <loop-vectorizer>`,
+which operates on Loops, and the :ref:`SLP Vectorizer
+<slp-vectorizer>`. These vectorizers
+focus on different optimization opportunities and use different techniques.
+The SLP vectorizer merges multiple scalars that are found in the code into
+vectors while the Loop Vectorizer widens instructions in loops
+to operate on multiple consecutive iterations.
+
+Both the Loop Vectorizer and the SLP Vectorizer are enabled by default.
+
+.. _loop-vectorizer:
+
+The Loop Vectorizer
+===================
+
+Usage
+-----
+
+The Loop Vectorizer is enabled by default, but it can be disabled
+through clang using the command line flag:
+
+.. code-block:: console
+
+   $ clang ... -fno-vectorize  file.c
+
+Command line flags
+^^^^^^^^^^^^^^^^^^
+
+The loop vectorizer uses a cost model to decide on the optimal vectorization factor
+and unroll factor. However, users of the vectorizer can force the vectorizer to use
+specific values. Both 'clang' and 'opt' support the flags below.
+
+Users can control the vectorization SIMD width using the command line flag "-force-vector-width".
+
+.. code-block:: console
+
+  $ clang  -mllvm -force-vector-width=8 ...
+  $ opt -loop-vectorize -force-vector-width=8 ...
+
+Users can control the unroll factor using the command line flag "-force-vector-unroll"
+
+.. code-block:: console
+
+  $ clang  -mllvm -force-vector-unroll=2 ...
+  $ opt -loop-vectorize -force-vector-unroll=2 ...
+
+Pragma loop hint directives
+^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``#pragma clang loop`` directive allows loop vectorization hints to be
+specified for the subsequent for, while, do-while, or c++11 range-based for
+loop. The directive allows vectorization and interleaving to be enabled or
+disabled. Vector width as well as interleave count can also be manually
+specified. The following example explicitly enables vectorization and
+interleaving:
+
+.. code-block:: c++
+
+  #pragma clang loop vectorize(enable) interleave(enable)
+  while(...) {
+    ...
+  }
+
+The following example implicitly enables vectorization and interleaving by
+specifying a vector width and interleaving count:
+
+.. code-block:: c++
+
+  #pragma clang loop vectorize_width(2) interleave_count(2)
+  for(...) {
+    ...
+  }
+
+See the Clang
+`language extensions
+<http://clang.llvm.org/docs/LanguageExtensions.html#extensions-for-loop-hint-optimizations>`_
+for details.
+
+Diagnostics
+-----------
+
+Many loops cannot be vectorized including loops with complicated control flow,
+unvectorizable types, and unvectorizable calls. The loop vectorizer generates
+optimization remarks which can be queried using command line options to identify
+and diagnose loops that are skipped by the loop-vectorizer.
+
+Optimization remarks are enabled using:
+
+``-Rpass=loop-vectorize`` identifies loops that were successfully vectorized.
+
+``-Rpass-missed=loop-vectorize`` identifies loops that failed vectorization and
+indicates if vectorization was specified.
+
+``-Rpass-analysis=loop-vectorize`` identifies the statements that caused
+vectorization to fail.
+
+Consider the following loop:
+
+.. code-block:: c++
+
+  #pragma clang loop vectorize(enable)
+  for (int i = 0; i < Length; i++) {
+    switch(A[i]) {
+    case 0: A[i] = i*2; break;
+    case 1: A[i] = i;   break;
+    default: A[i] = 0;
+    }
+  }
+
+The command line ``-Rpass-missed=loop-vectorized`` prints the remark:
+
+.. code-block:: console
+
+  no_switch.cpp:4:5: remark: loop not vectorized: vectorization is explicitly enabled [-Rpass-missed=loop-vectorize]
+
+And the command line ``-Rpass-analysis=loop-vectorize`` indicates that the
+switch statement cannot be vectorized.
+
+.. code-block:: console
+
+  no_switch.cpp:4:5: remark: loop not vectorized: loop contains a switch statement [-Rpass-analysis=loop-vectorize]
+    switch(A[i]) {
+    ^
+
+To ensure line and column numbers are produced include the command line options
+``-gline-tables-only`` and ``-gcolumn-info``. See the Clang `user manual
+<http://clang.llvm.org/docs/UsersManual.html#options-to-emit-optimization-reports>`_
+for details
+
+Features
+--------
+
+The LLVM Loop Vectorizer has a number of features that allow it to vectorize
+complex loops.
+
+Loops with unknown trip count
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The Loop Vectorizer supports loops with an unknown trip count.
+In the loop below, the iteration ``start`` and ``finish`` points are unknown,
+and the Loop Vectorizer has a mechanism to vectorize loops that do not start
+at zero. In this example, 'n' may not be a multiple of the vector width, and
+the vectorizer has to execute the last few iterations as scalar code. Keeping
+a scalar copy of the loop increases the code size.
+
+.. code-block:: c++
+
+  void bar(float *A, float* B, float K, int start, int end) {
+    for (int i = start; i < end; ++i)
+      A[i] *= B[i] + K;
+  }
+
+Runtime Checks of Pointers
+^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+In the example below, if the pointers A and B point to consecutive addresses,
+then it is illegal to vectorize the code because some elements of A will be
+written before they are read from array B.
+
+Some programmers use the 'restrict' keyword to notify the compiler that the
+pointers are disjointed, but in our example, the Loop Vectorizer has no way of
+knowing that the pointers A and B are unique. The Loop Vectorizer handles this
+loop by placing code that checks, at runtime, if the arrays A and B point to
+disjointed memory locations. If arrays A and B overlap, then the scalar version
+of the loop is executed.
+
+.. code-block:: c++
+
+  void bar(float *A, float* B, float K, int n) {
+    for (int i = 0; i < n; ++i)
+      A[i] *= B[i] + K;
+  }
+
+
+Reductions
+^^^^^^^^^^
+
+In this example the ``sum`` variable is used by consecutive iterations of
+the loop. Normally, this would prevent vectorization, but the vectorizer can
+detect that 'sum' is a reduction variable. The variable 'sum' becomes a vector
+of integers, and at the end of the loop the elements of the array are added
+together to create the correct result. We support a number of different
+reduction operations, such as addition, multiplication, XOR, AND and OR.
+
+.. code-block:: c++
+
+  int foo(int *A, int *B, int n) {
+    unsigned sum = 0;
+    for (int i = 0; i < n; ++i)
+      sum += A[i] + 5;
+    return sum;
+  }
+
+We support floating point reduction operations when `-ffast-math` is used.
+
+Inductions
+^^^^^^^^^^
+
+In this example the value of the induction variable ``i`` is saved into an
+array. The Loop Vectorizer knows to vectorize induction variables.
+
+.. code-block:: c++
+
+  void bar(float *A, float* B, float K, int n) {
+    for (int i = 0; i < n; ++i)
+      A[i] = i;
+  }
+
+If Conversion
+^^^^^^^^^^^^^
+
+The Loop Vectorizer is able to "flatten" the IF statement in the code and
+generate a single stream of instructions. The Loop Vectorizer supports any
+control flow in the innermost loop. The innermost loop may contain complex
+nesting of IFs, ELSEs and even GOTOs.
+
+.. code-block:: c++
+
+  int foo(int *A, int *B, int n) {
+    unsigned sum = 0;
+    for (int i = 0; i < n; ++i)
+      if (A[i] > B[i])
+        sum += A[i] + 5;
+    return sum;
+  }
+
+Pointer Induction Variables
+^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+This example uses the "accumulate" function of the standard c++ library. This
+loop uses C++ iterators, which are pointers, and not integer indices.
+The Loop Vectorizer detects pointer induction variables and can vectorize
+this loop. This feature is important because many C++ programs use iterators.
+
+.. code-block:: c++
+
+  int baz(int *A, int n) {
+    return std::accumulate(A, A + n, 0);
+  }
+
+Reverse Iterators
+^^^^^^^^^^^^^^^^^
+
+The Loop Vectorizer can vectorize loops that count backwards.
+
+.. code-block:: c++
+
+  int foo(int *A, int *B, int n) {
+    for (int i = n; i > 0; --i)
+      A[i] +=1;
+  }
+
+Scatter / Gather
+^^^^^^^^^^^^^^^^
+
+The Loop Vectorizer can vectorize code that becomes a sequence of scalar instructions 
+that scatter/gathers memory.
+
+.. code-block:: c++
+
+  int foo(int * A, int * B, int n) {
+    for (intptr_t i = 0; i < n; ++i)
+        A[i] += B[i * 4];
+  }
+
+In many situations the cost model will inform LLVM that this is not beneficial
+and LLVM will only vectorize such code if forced with "-mllvm -force-vector-width=#".
+
+Vectorization of Mixed Types
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The Loop Vectorizer can vectorize programs with mixed types. The Vectorizer
+cost model can estimate the cost of the type conversion and decide if
+vectorization is profitable.
+
+.. code-block:: c++
+
+  int foo(int *A, char *B, int n, int k) {
+    for (int i = 0; i < n; ++i)
+      A[i] += 4 * B[i];
+  }
+
+Global Structures Alias Analysis
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Access to global structures can also be vectorized, with alias analysis being
+used to make sure accesses don't alias. Run-time checks can also be added on
+pointer access to structure members.
+
+Many variations are supported, but some that rely on undefined behaviour being
+ignored (as other compilers do) are still being left un-vectorized.
+
+.. code-block:: c++
+
+  struct { int A[100], K, B[100]; } Foo;
+
+  int foo() {
+    for (int i = 0; i < 100; ++i)
+      Foo.A[i] = Foo.B[i] + 100;
+  }
+
+Vectorization of function calls
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The Loop Vectorize can vectorize intrinsic math functions.
+See the table below for a list of these functions.
+
++-----+-----+---------+
+| pow | exp |  exp2   |
++-----+-----+---------+
+| sin | cos |  sqrt   |
++-----+-----+---------+
+| log |log2 |  log10  |
++-----+-----+---------+
+|fabs |floor|  ceil   |
++-----+-----+---------+
+|fma  |trunc|nearbyint|
++-----+-----+---------+
+|     |     | fmuladd |
++-----+-----+---------+
+
+The loop vectorizer knows about special instructions on the target and will
+vectorize a loop containing a function call that maps to the instructions. For
+example, the loop below will be vectorized on Intel x86 if the SSE4.1 roundps
+instruction is available.
+
+.. code-block:: c++
+
+  void foo(float *f) {
+    for (int i = 0; i != 1024; ++i)
+      f[i] = floorf(f[i]);
+  }
+
+Partial unrolling during vectorization
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Modern processors feature multiple execution units, and only programs that contain a
+high degree of parallelism can fully utilize the entire width of the machine. 
+The Loop Vectorizer increases the instruction level parallelism (ILP) by 
+performing partial-unrolling of loops.
+
+In the example below the entire array is accumulated into the variable 'sum'.
+This is inefficient because only a single execution port can be used by the processor.
+By unrolling the code the Loop Vectorizer allows two or more execution ports
+to be used simultaneously.
+
+.. code-block:: c++
+
+  int foo(int *A, int *B, int n) {
+    unsigned sum = 0;
+    for (int i = 0; i < n; ++i)
+        sum += A[i];
+    return sum;
+  }
+
+The Loop Vectorizer uses a cost model to decide when it is profitable to unroll loops.
+The decision to unroll the loop depends on the register pressure and the generated code size. 
+
+Performance
+-----------
+
+This section shows the execution time of Clang on a simple benchmark:
+`gcc-loops <http://llvm.org/viewvc/llvm-project/test-suite/trunk/SingleSource/UnitTests/Vectorizer/>`_.
+This benchmarks is a collection of loops from the GCC autovectorization
+`page <http://gcc.gnu.org/projects/tree-ssa/vectorization.html>`_ by Dorit Nuzman.
+
+The chart below compares GCC-4.7, ICC-13, and Clang-SVN with and without loop vectorization at -O3, tuned for "corei7-avx", running on a Sandybridge iMac.
+The Y-axis shows the time in msec. Lower is better. The last column shows the geomean of all the kernels.
+
+.. image:: gcc-loops.png
+
+And Linpack-pc with the same configuration. Result is Mflops, higher is better.
+
+.. image:: linpack-pc.png
+
+.. _slp-vectorizer:
+
+The SLP Vectorizer
+==================
+
+Details
+-------
+
+The goal of SLP vectorization (a.k.a. superword-level parallelism) is
+to combine similar independent instructions
+into vector instructions. Memory accesses, arithmetic operations, comparison
+operations, PHI-nodes, can all be vectorized using this technique.
+
+For example, the following function performs very similar operations on its
+inputs (a1, b1) and (a2, b2). The basic-block vectorizer may combine these
+into vector operations.
+
+.. code-block:: c++
+
+  void foo(int a1, int a2, int b1, int b2, int *A) {
+    A[0] = a1*(a1 + b1)/b1 + 50*b1/a1;
+    A[1] = a2*(a2 + b2)/b2 + 50*b2/a2;
+  }
+
+The SLP-vectorizer processes the code bottom-up, across basic blocks, in search of scalars to combine.
+
+Usage
+------
+
+The SLP Vectorizer is enabled by default, but it can be disabled
+through clang using the command line flag:
+
+.. code-block:: console
+
+   $ clang -fno-slp-vectorize file.c
+
+LLVM has a second basic block vectorization phase
+which is more compile-time intensive (The BB vectorizer). This optimization
+can be enabled through clang using the command line flag:
+
+.. code-block:: console
+
+   $ clang -fslp-vectorize-aggressive file.c
+

Added: www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMBackend.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMBackend.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMBackend.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMBackend.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,1939 @@
+=======================
+Writing an LLVM Backend
+=======================
+
+.. toctree::
+   :hidden:
+
+   HowToUseInstrMappings
+
+.. contents::
+   :local:
+
+Introduction
+============
+
+This document describes techniques for writing compiler backends that convert
+the LLVM Intermediate Representation (IR) to code for a specified machine or
+other languages.  Code intended for a specific machine can take the form of
+either assembly code or binary code (usable for a JIT compiler).
+
+The backend of LLVM features a target-independent code generator that may
+create output for several types of target CPUs --- including X86, PowerPC,
+ARM, and SPARC.  The backend may also be used to generate code targeted at SPUs
+of the Cell processor or GPUs to support the execution of compute kernels.
+
+The document focuses on existing examples found in subdirectories of
+``llvm/lib/Target`` in a downloaded LLVM release.  In particular, this document
+focuses on the example of creating a static compiler (one that emits text
+assembly) for a SPARC target, because SPARC has fairly standard
+characteristics, such as a RISC instruction set and straightforward calling
+conventions.
+
+Audience
+--------
+
+The audience for this document is anyone who needs to write an LLVM backend to
+generate code for a specific hardware or software target.
+
+Prerequisite Reading
+--------------------
+
+These essential documents must be read before reading this document:
+
+* `LLVM Language Reference Manual <LangRef.html>`_ --- a reference manual for
+  the LLVM assembly language.
+
+* :doc:`CodeGenerator` --- a guide to the components (classes and code
+  generation algorithms) for translating the LLVM internal representation into
+  machine code for a specified target.  Pay particular attention to the
+  descriptions of code generation stages: Instruction Selection, Scheduling and
+  Formation, SSA-based Optimization, Register Allocation, Prolog/Epilog Code
+  Insertion, Late Machine Code Optimizations, and Code Emission.
+
+* :doc:`TableGen/index` --- a document that describes the TableGen
+  (``tblgen``) application that manages domain-specific information to support
+  LLVM code generation.  TableGen processes input from a target description
+  file (``.td`` suffix) and generates C++ code that can be used for code
+  generation.
+
+* :doc:`WritingAnLLVMPass` --- The assembly printer is a ``FunctionPass``, as
+  are several ``SelectionDAG`` processing steps.
+
+To follow the SPARC examples in this document, have a copy of `The SPARC
+Architecture Manual, Version 8 <http://www.sparc.org/standards/V8.pdf>`_ for
+reference.  For details about the ARM instruction set, refer to the `ARM
+Architecture Reference Manual <http://infocenter.arm.com/>`_.  For more about
+the GNU Assembler format (``GAS``), see `Using As
+<http://sourceware.org/binutils/docs/as/index.html>`_, especially for the
+assembly printer.  "Using As" contains a list of target machine dependent
+features.
+
+Basic Steps
+-----------
+
+To write a compiler backend for LLVM that converts the LLVM IR to code for a
+specified target (machine or other language), follow these steps:
+
+* Create a subclass of the ``TargetMachine`` class that describes
+  characteristics of your target machine.  Copy existing examples of specific
+  ``TargetMachine`` class and header files; for example, start with
+  ``SparcTargetMachine.cpp`` and ``SparcTargetMachine.h``, but change the file
+  names for your target.  Similarly, change code that references "``Sparc``" to
+  reference your target.
+
+* Describe the register set of the target.  Use TableGen to generate code for
+  register definition, register aliases, and register classes from a
+  target-specific ``RegisterInfo.td`` input file.  You should also write
+  additional code for a subclass of the ``TargetRegisterInfo`` class that
+  represents the class register file data used for register allocation and also
+  describes the interactions between registers.
+
+* Describe the instruction set of the target.  Use TableGen to generate code
+  for target-specific instructions from target-specific versions of
+  ``TargetInstrFormats.td`` and ``TargetInstrInfo.td``.  You should write
+  additional code for a subclass of the ``TargetInstrInfo`` class to represent
+  machine instructions supported by the target machine.
+
+* Describe the selection and conversion of the LLVM IR from a Directed Acyclic
+  Graph (DAG) representation of instructions to native target-specific
+  instructions.  Use TableGen to generate code that matches patterns and
+  selects instructions based on additional information in a target-specific
+  version of ``TargetInstrInfo.td``.  Write code for ``XXXISelDAGToDAG.cpp``,
+  where ``XXX`` identifies the specific target, to perform pattern matching and
+  DAG-to-DAG instruction selection.  Also write code in ``XXXISelLowering.cpp``
+  to replace or remove operations and data types that are not supported
+  natively in a SelectionDAG.
+
+* Write code for an assembly printer that converts LLVM IR to a GAS format for
+  your target machine.  You should add assembly strings to the instructions
+  defined in your target-specific version of ``TargetInstrInfo.td``.  You
+  should also write code for a subclass of ``AsmPrinter`` that performs the
+  LLVM-to-assembly conversion and a trivial subclass of ``TargetAsmInfo``.
+
+* Optionally, add support for subtargets (i.e., variants with different
+  capabilities).  You should also write code for a subclass of the
+  ``TargetSubtarget`` class, which allows you to use the ``-mcpu=`` and
+  ``-mattr=`` command-line options.
+
+* Optionally, add JIT support and create a machine code emitter (subclass of
+  ``TargetJITInfo``) that is used to emit binary code directly into memory.
+
+In the ``.cpp`` and ``.h``. files, initially stub up these methods and then
+implement them later.  Initially, you may not know which private members that
+the class will need and which components will need to be subclassed.
+
+Preliminaries
+-------------
+
+To actually create your compiler backend, you need to create and modify a few
+files.  The absolute minimum is discussed here.  But to actually use the LLVM
+target-independent code generator, you must perform the steps described in the
+:doc:`LLVM Target-Independent Code Generator <CodeGenerator>` document.
+
+First, you should create a subdirectory under ``lib/Target`` to hold all the
+files related to your target.  If your target is called "Dummy", create the
+directory ``lib/Target/Dummy``.
+
+In this new directory, create a ``CMakeLists.txt``.  It is easiest to copy a
+``CMakeLists.txt`` of another target and modify it.  It should at least contain
+the ``LLVM_TARGET_DEFINITIONS`` variable. The library can be named ``LLVMDummy``
+(for example, see the MIPS target).  Alternatively, you can split the library
+into ``LLVMDummyCodeGen`` and ``LLVMDummyAsmPrinter``, the latter of which
+should be implemented in a subdirectory below ``lib/Target/Dummy`` (for example,
+see the PowerPC target).
+
+Note that these two naming schemes are hardcoded into ``llvm-config``.  Using
+any other naming scheme will confuse ``llvm-config`` and produce a lot of
+(seemingly unrelated) linker errors when linking ``llc``.
+
+To make your target actually do something, you need to implement a subclass of
+``TargetMachine``.  This implementation should typically be in the file
+``lib/Target/DummyTargetMachine.cpp``, but any file in the ``lib/Target``
+directory will be built and should work.  To use LLVM's target independent code
+generator, you should do what all current machine backends do: create a
+subclass of ``LLVMTargetMachine``.  (To create a target from scratch, create a
+subclass of ``TargetMachine``.)
+
+To get LLVM to actually build and link your target, you need to run ``cmake``
+with ``-DLLVM_EXPERIMENTAL_TARGETS_TO_BUILD=Dummy``. This will build your
+target without needing to add it to the list of all the targets.
+
+Once your target is stable, you can add it to the ``LLVM_ALL_TARGETS`` variable
+located in the main ``CMakeLists.txt``.
+
+Target Machine
+==============
+
+``LLVMTargetMachine`` is designed as a base class for targets implemented with
+the LLVM target-independent code generator.  The ``LLVMTargetMachine`` class
+should be specialized by a concrete target class that implements the various
+virtual methods.  ``LLVMTargetMachine`` is defined as a subclass of
+``TargetMachine`` in ``include/llvm/Target/TargetMachine.h``.  The
+``TargetMachine`` class implementation (``TargetMachine.cpp``) also processes
+numerous command-line options.
+
+To create a concrete target-specific subclass of ``LLVMTargetMachine``, start
+by copying an existing ``TargetMachine`` class and header.  You should name the
+files that you create to reflect your specific target.  For instance, for the
+SPARC target, name the files ``SparcTargetMachine.h`` and
+``SparcTargetMachine.cpp``.
+
+For a target machine ``XXX``, the implementation of ``XXXTargetMachine`` must
+have access methods to obtain objects that represent target components.  These
+methods are named ``get*Info``, and are intended to obtain the instruction set
+(``getInstrInfo``), register set (``getRegisterInfo``), stack frame layout
+(``getFrameInfo``), and similar information.  ``XXXTargetMachine`` must also
+implement the ``getDataLayout`` method to access an object with target-specific
+data characteristics, such as data type size and alignment requirements.
+
+For instance, for the SPARC target, the header file ``SparcTargetMachine.h``
+declares prototypes for several ``get*Info`` and ``getDataLayout`` methods that
+simply return a class member.
+
+.. code-block:: c++
+
+  namespace llvm {
+
+  class Module;
+
+  class SparcTargetMachine : public LLVMTargetMachine {
+    const DataLayout DataLayout;       // Calculates type size & alignment
+    SparcSubtarget Subtarget;
+    SparcInstrInfo InstrInfo;
+    TargetFrameInfo FrameInfo;
+
+  protected:
+    virtual const TargetAsmInfo *createTargetAsmInfo() const;
+
+  public:
+    SparcTargetMachine(const Module &M, const std::string &FS);
+
+    virtual const SparcInstrInfo *getInstrInfo() const {return &InstrInfo; }
+    virtual const TargetFrameInfo *getFrameInfo() const {return &FrameInfo; }
+    virtual const TargetSubtarget *getSubtargetImpl() const{return &Subtarget; }
+    virtual const TargetRegisterInfo *getRegisterInfo() const {
+      return &InstrInfo.getRegisterInfo();
+    }
+    virtual const DataLayout *getDataLayout() const { return &DataLayout; }
+    static unsigned getModuleMatchQuality(const Module &M);
+
+    // Pass Pipeline Configuration
+    virtual bool addInstSelector(PassManagerBase &PM, bool Fast);
+    virtual bool addPreEmitPass(PassManagerBase &PM, bool Fast);
+  };
+
+  } // end namespace llvm
+
+* ``getInstrInfo()``
+* ``getRegisterInfo()``
+* ``getFrameInfo()``
+* ``getDataLayout()``
+* ``getSubtargetImpl()``
+
+For some targets, you also need to support the following methods:
+
+* ``getTargetLowering()``
+* ``getJITInfo()``
+
+Some architectures, such as GPUs, do not support jumping to an arbitrary
+program location and implement branching using masked execution and loop using
+special instructions around the loop body. In order to avoid CFG modifications
+that introduce irreducible control flow not handled by such hardware, a target
+must call `setRequiresStructuredCFG(true)` when being initialized.
+
+In addition, the ``XXXTargetMachine`` constructor should specify a
+``TargetDescription`` string that determines the data layout for the target
+machine, including characteristics such as pointer size, alignment, and
+endianness.  For example, the constructor for ``SparcTargetMachine`` contains
+the following:
+
+.. code-block:: c++
+
+  SparcTargetMachine::SparcTargetMachine(const Module &M, const std::string &FS)
+    : DataLayout("E-p:32:32-f128:128:128"),
+      Subtarget(M, FS), InstrInfo(Subtarget),
+      FrameInfo(TargetFrameInfo::StackGrowsDown, 8, 0) {
+  }
+
+Hyphens separate portions of the ``TargetDescription`` string.
+
+* An upper-case "``E``" in the string indicates a big-endian target data model.
+  A lower-case "``e``" indicates little-endian.
+
+* "``p:``" is followed by pointer information: size, ABI alignment, and
+  preferred alignment.  If only two figures follow "``p:``", then the first
+  value is pointer size, and the second value is both ABI and preferred
+  alignment.
+
+* Then a letter for numeric type alignment: "``i``", "``f``", "``v``", or
+  "``a``" (corresponding to integer, floating point, vector, or aggregate).
+  "``i``", "``v``", or "``a``" are followed by ABI alignment and preferred
+  alignment. "``f``" is followed by three values: the first indicates the size
+  of a long double, then ABI alignment, and then ABI preferred alignment.
+
+Target Registration
+===================
+
+You must also register your target with the ``TargetRegistry``, which is what
+other LLVM tools use to be able to lookup and use your target at runtime.  The
+``TargetRegistry`` can be used directly, but for most targets there are helper
+templates which should take care of the work for you.
+
+All targets should declare a global ``Target`` object which is used to
+represent the target during registration.  Then, in the target's ``TargetInfo``
+library, the target should define that object and use the ``RegisterTarget``
+template to register the target.  For example, the Sparc registration code
+looks like this:
+
+.. code-block:: c++
+
+  Target llvm::getTheSparcTarget();
+
+  extern "C" void LLVMInitializeSparcTargetInfo() {
+    RegisterTarget<Triple::sparc, /*HasJIT=*/false>
+      X(getTheSparcTarget(), "sparc", "Sparc");
+  }
+
+This allows the ``TargetRegistry`` to look up the target by name or by target
+triple.  In addition, most targets will also register additional features which
+are available in separate libraries.  These registration steps are separate,
+because some clients may wish to only link in some parts of the target --- the
+JIT code generator does not require the use of the assembler printer, for
+example.  Here is an example of registering the Sparc assembly printer:
+
+.. code-block:: c++
+
+  extern "C" void LLVMInitializeSparcAsmPrinter() {
+    RegisterAsmPrinter<SparcAsmPrinter> X(getTheSparcTarget());
+  }
+
+For more information, see "`llvm/Target/TargetRegistry.h
+</doxygen/TargetRegistry_8h-source.html>`_".
+
+Register Set and Register Classes
+=================================
+
+You should describe a concrete target-specific class that represents the
+register file of a target machine.  This class is called ``XXXRegisterInfo``
+(where ``XXX`` identifies the target) and represents the class register file
+data that is used for register allocation.  It also describes the interactions
+between registers.
+
+You also need to define register classes to categorize related registers.  A
+register class should be added for groups of registers that are all treated the
+same way for some instruction.  Typical examples are register classes for
+integer, floating-point, or vector registers.  A register allocator allows an
+instruction to use any register in a specified register class to perform the
+instruction in a similar manner.  Register classes allocate virtual registers
+to instructions from these sets, and register classes let the
+target-independent register allocator automatically choose the actual
+registers.
+
+Much of the code for registers, including register definition, register
+aliases, and register classes, is generated by TableGen from
+``XXXRegisterInfo.td`` input files and placed in ``XXXGenRegisterInfo.h.inc``
+and ``XXXGenRegisterInfo.inc`` output files.  Some of the code in the
+implementation of ``XXXRegisterInfo`` requires hand-coding.
+
+Defining a Register
+-------------------
+
+The ``XXXRegisterInfo.td`` file typically starts with register definitions for
+a target machine.  The ``Register`` class (specified in ``Target.td``) is used
+to define an object for each register.  The specified string ``n`` becomes the
+``Name`` of the register.  The basic ``Register`` object does not have any
+subregisters and does not specify any aliases.
+
+.. code-block:: text
+
+  class Register<string n> {
+    string Namespace = "";
+    string AsmName = n;
+    string Name = n;
+    int SpillSize = 0;
+    int SpillAlignment = 0;
+    list<Register> Aliases = [];
+    list<Register> SubRegs = [];
+    list<int> DwarfNumbers = [];
+  }
+
+For example, in the ``X86RegisterInfo.td`` file, there are register definitions
+that utilize the ``Register`` class, such as:
+
+.. code-block:: text
+
+  def AL : Register<"AL">, DwarfRegNum<[0, 0, 0]>;
+
+This defines the register ``AL`` and assigns it values (with ``DwarfRegNum``)
+that are used by ``gcc``, ``gdb``, or a debug information writer to identify a
+register.  For register ``AL``, ``DwarfRegNum`` takes an array of 3 values
+representing 3 different modes: the first element is for X86-64, the second for
+exception handling (EH) on X86-32, and the third is generic. -1 is a special
+Dwarf number that indicates the gcc number is undefined, and -2 indicates the
+register number is invalid for this mode.
+
+From the previously described line in the ``X86RegisterInfo.td`` file, TableGen
+generates this code in the ``X86GenRegisterInfo.inc`` file:
+
+.. code-block:: c++
+
+  static const unsigned GR8[] = { X86::AL, ... };
+
+  const unsigned AL_AliasSet[] = { X86::AX, X86::EAX, X86::RAX, 0 };
+
+  const TargetRegisterDesc RegisterDescriptors[] = {
+    ...
+  { "AL", "AL", AL_AliasSet, Empty_SubRegsSet, Empty_SubRegsSet, AL_SuperRegsSet }, ...
+
+From the register info file, TableGen generates a ``TargetRegisterDesc`` object
+for each register.  ``TargetRegisterDesc`` is defined in
+``include/llvm/Target/TargetRegisterInfo.h`` with the following fields:
+
+.. code-block:: c++
+
+  struct TargetRegisterDesc {
+    const char     *AsmName;      // Assembly language name for the register
+    const char     *Name;         // Printable name for the reg (for debugging)
+    const unsigned *AliasSet;     // Register Alias Set
+    const unsigned *SubRegs;      // Sub-register set
+    const unsigned *ImmSubRegs;   // Immediate sub-register set
+    const unsigned *SuperRegs;    // Super-register set
+  };
+
+TableGen uses the entire target description file (``.td``) to determine text
+names for the register (in the ``AsmName`` and ``Name`` fields of
+``TargetRegisterDesc``) and the relationships of other registers to the defined
+register (in the other ``TargetRegisterDesc`` fields).  In this example, other
+definitions establish the registers "``AX``", "``EAX``", and "``RAX``" as
+aliases for one another, so TableGen generates a null-terminated array
+(``AL_AliasSet``) for this register alias set.
+
+The ``Register`` class is commonly used as a base class for more complex
+classes.  In ``Target.td``, the ``Register`` class is the base for the
+``RegisterWithSubRegs`` class that is used to define registers that need to
+specify subregisters in the ``SubRegs`` list, as shown here:
+
+.. code-block:: text
+
+  class RegisterWithSubRegs<string n, list<Register> subregs> : Register<n> {
+    let SubRegs = subregs;
+  }
+
+In ``SparcRegisterInfo.td``, additional register classes are defined for SPARC:
+a ``Register`` subclass, ``SparcReg``, and further subclasses: ``Ri``, ``Rf``,
+and ``Rd``.  SPARC registers are identified by 5-bit ID numbers, which is a
+feature common to these subclasses.  Note the use of "``let``" expressions to
+override values that are initially defined in a superclass (such as ``SubRegs``
+field in the ``Rd`` class).
+
+.. code-block:: text
+
+  class SparcReg<string n> : Register<n> {
+    field bits<5> Num;
+    let Namespace = "SP";
+  }
+  // Ri - 32-bit integer registers
+  class Ri<bits<5> num, string n> :
+  SparcReg<n> {
+    let Num = num;
+  }
+  // Rf - 32-bit floating-point registers
+  class Rf<bits<5> num, string n> :
+  SparcReg<n> {
+    let Num = num;
+  }
+  // Rd - Slots in the FP register file for 64-bit floating-point values.
+  class Rd<bits<5> num, string n, list<Register> subregs> : SparcReg<n> {
+    let Num = num;
+    let SubRegs = subregs;
+  }
+
+In the ``SparcRegisterInfo.td`` file, there are register definitions that
+utilize these subclasses of ``Register``, such as:
+
+.. code-block:: text
+
+  def G0 : Ri< 0, "G0">, DwarfRegNum<[0]>;
+  def G1 : Ri< 1, "G1">, DwarfRegNum<[1]>;
+  ...
+  def F0 : Rf< 0, "F0">, DwarfRegNum<[32]>;
+  def F1 : Rf< 1, "F1">, DwarfRegNum<[33]>;
+  ...
+  def D0 : Rd< 0, "F0", [F0, F1]>, DwarfRegNum<[32]>;
+  def D1 : Rd< 2, "F2", [F2, F3]>, DwarfRegNum<[34]>;
+
+The last two registers shown above (``D0`` and ``D1``) are double-precision
+floating-point registers that are aliases for pairs of single-precision
+floating-point sub-registers.  In addition to aliases, the sub-register and
+super-register relationships of the defined register are in fields of a
+register's ``TargetRegisterDesc``.
+
+Defining a Register Class
+-------------------------
+
+The ``RegisterClass`` class (specified in ``Target.td``) is used to define an
+object that represents a group of related registers and also defines the
+default allocation order of the registers.  A target description file
+``XXXRegisterInfo.td`` that uses ``Target.td`` can construct register classes
+using the following class:
+
+.. code-block:: text
+
+  class RegisterClass<string namespace,
+  list<ValueType> regTypes, int alignment, dag regList> {
+    string Namespace = namespace;
+    list<ValueType> RegTypes = regTypes;
+    int Size = 0;  // spill size, in bits; zero lets tblgen pick the size
+    int Alignment = alignment;
+
+    // CopyCost is the cost of copying a value between two registers
+    // default value 1 means a single instruction
+    // A negative value means copying is extremely expensive or impossible
+    int CopyCost = 1;
+    dag MemberList = regList;
+
+    // for register classes that are subregisters of this class
+    list<RegisterClass> SubRegClassList = [];
+
+    code MethodProtos = [{}];  // to insert arbitrary code
+    code MethodBodies = [{}];
+  }
+
+To define a ``RegisterClass``, use the following 4 arguments:
+
+* The first argument of the definition is the name of the namespace.
+
+* The second argument is a list of ``ValueType`` register type values that are
+  defined in ``include/llvm/CodeGen/ValueTypes.td``.  Defined values include
+  integer types (such as ``i16``, ``i32``, and ``i1`` for Boolean),
+  floating-point types (``f32``, ``f64``), and vector types (for example,
+  ``v8i16`` for an ``8 x i16`` vector).  All registers in a ``RegisterClass``
+  must have the same ``ValueType``, but some registers may store vector data in
+  different configurations.  For example a register that can process a 128-bit
+  vector may be able to handle 16 8-bit integer elements, 8 16-bit integers, 4
+  32-bit integers, and so on.
+
+* The third argument of the ``RegisterClass`` definition specifies the
+  alignment required of the registers when they are stored or loaded to
+  memory.
+
+* The final argument, ``regList``, specifies which registers are in this class.
+  If an alternative allocation order method is not specified, then ``regList``
+  also defines the order of allocation used by the register allocator.  Besides
+  simply listing registers with ``(add R0, R1, ...)``, more advanced set
+  operators are available.  See ``include/llvm/Target/Target.td`` for more
+  information.
+
+In ``SparcRegisterInfo.td``, three ``RegisterClass`` objects are defined:
+``FPRegs``, ``DFPRegs``, and ``IntRegs``.  For all three register classes, the
+first argument defines the namespace with the string "``SP``".  ``FPRegs``
+defines a group of 32 single-precision floating-point registers (``F0`` to
+``F31``); ``DFPRegs`` defines a group of 16 double-precision registers
+(``D0-D15``).
+
+.. code-block:: text
+
+  // F0, F1, F2, ..., F31
+  def FPRegs : RegisterClass<"SP", [f32], 32, (sequence "F%u", 0, 31)>;
+
+  def DFPRegs : RegisterClass<"SP", [f64], 64,
+                              (add D0, D1, D2, D3, D4, D5, D6, D7, D8,
+                                   D9, D10, D11, D12, D13, D14, D15)>;
+
+  def IntRegs : RegisterClass<"SP", [i32], 32,
+      (add L0, L1, L2, L3, L4, L5, L6, L7,
+           I0, I1, I2, I3, I4, I5,
+           O0, O1, O2, O3, O4, O5, O7,
+           G1,
+           // Non-allocatable regs:
+           G2, G3, G4,
+           O6,        // stack ptr
+           I6,        // frame ptr
+           I7,        // return address
+           G0,        // constant zero
+           G5, G6, G7 // reserved for kernel
+      )>;
+
+Using ``SparcRegisterInfo.td`` with TableGen generates several output files
+that are intended for inclusion in other source code that you write.
+``SparcRegisterInfo.td`` generates ``SparcGenRegisterInfo.h.inc``, which should
+be included in the header file for the implementation of the SPARC register
+implementation that you write (``SparcRegisterInfo.h``).  In
+``SparcGenRegisterInfo.h.inc`` a new structure is defined called
+``SparcGenRegisterInfo`` that uses ``TargetRegisterInfo`` as its base.  It also
+specifies types, based upon the defined register classes: ``DFPRegsClass``,
+``FPRegsClass``, and ``IntRegsClass``.
+
+``SparcRegisterInfo.td`` also generates ``SparcGenRegisterInfo.inc``, which is
+included at the bottom of ``SparcRegisterInfo.cpp``, the SPARC register
+implementation.  The code below shows only the generated integer registers and
+associated register classes.  The order of registers in ``IntRegs`` reflects
+the order in the definition of ``IntRegs`` in the target description file.
+
+.. code-block:: c++
+
+  // IntRegs Register Class...
+  static const unsigned IntRegs[] = {
+    SP::L0, SP::L1, SP::L2, SP::L3, SP::L4, SP::L5,
+    SP::L6, SP::L7, SP::I0, SP::I1, SP::I2, SP::I3,
+    SP::I4, SP::I5, SP::O0, SP::O1, SP::O2, SP::O3,
+    SP::O4, SP::O5, SP::O7, SP::G1, SP::G2, SP::G3,
+    SP::G4, SP::O6, SP::I6, SP::I7, SP::G0, SP::G5,
+    SP::G6, SP::G7,
+  };
+
+  // IntRegsVTs Register Class Value Types...
+  static const MVT::ValueType IntRegsVTs[] = {
+    MVT::i32, MVT::Other
+  };
+
+  namespace SP {   // Register class instances
+    DFPRegsClass    DFPRegsRegClass;
+    FPRegsClass     FPRegsRegClass;
+    IntRegsClass    IntRegsRegClass;
+  ...
+    // IntRegs Sub-register Classess...
+    static const TargetRegisterClass* const IntRegsSubRegClasses [] = {
+      NULL
+    };
+  ...
+    // IntRegs Super-register Classess...
+    static const TargetRegisterClass* const IntRegsSuperRegClasses [] = {
+      NULL
+    };
+  ...
+    // IntRegs Register Class sub-classes...
+    static const TargetRegisterClass* const IntRegsSubclasses [] = {
+      NULL
+    };
+  ...
+    // IntRegs Register Class super-classes...
+    static const TargetRegisterClass* const IntRegsSuperclasses [] = {
+      NULL
+    };
+
+    IntRegsClass::IntRegsClass() : TargetRegisterClass(IntRegsRegClassID,
+      IntRegsVTs, IntRegsSubclasses, IntRegsSuperclasses, IntRegsSubRegClasses,
+      IntRegsSuperRegClasses, 4, 4, 1, IntRegs, IntRegs + 32) {}
+  }
+
+The register allocators will avoid using reserved registers, and callee saved
+registers are not used until all the volatile registers have been used.  That
+is usually good enough, but in some cases it may be necessary to provide custom
+allocation orders.
+
+Implement a subclass of ``TargetRegisterInfo``
+----------------------------------------------
+
+The final step is to hand code portions of ``XXXRegisterInfo``, which
+implements the interface described in ``TargetRegisterInfo.h`` (see
+:ref:`TargetRegisterInfo`).  These functions return ``0``, ``NULL``, or
+``false``, unless overridden.  Here is a list of functions that are overridden
+for the SPARC implementation in ``SparcRegisterInfo.cpp``:
+
+* ``getCalleeSavedRegs`` --- Returns a list of callee-saved registers in the
+  order of the desired callee-save stack frame offset.
+
+* ``getReservedRegs`` --- Returns a bitset indexed by physical register
+  numbers, indicating if a particular register is unavailable.
+
+* ``hasFP`` --- Return a Boolean indicating if a function should have a
+  dedicated frame pointer register.
+
+* ``eliminateCallFramePseudoInstr`` --- If call frame setup or destroy pseudo
+  instructions are used, this can be called to eliminate them.
+
+* ``eliminateFrameIndex`` --- Eliminate abstract frame indices from
+  instructions that may use them.
+
+* ``emitPrologue`` --- Insert prologue code into the function.
+
+* ``emitEpilogue`` --- Insert epilogue code into the function.
+
+.. _instruction-set:
+
+Instruction Set
+===============
+
+During the early stages of code generation, the LLVM IR code is converted to a
+``SelectionDAG`` with nodes that are instances of the ``SDNode`` class
+containing target instructions.  An ``SDNode`` has an opcode, operands, type
+requirements, and operation properties.  For example, is an operation
+commutative, does an operation load from memory.  The various operation node
+types are described in the ``include/llvm/CodeGen/SelectionDAGNodes.h`` file
+(values of the ``NodeType`` enum in the ``ISD`` namespace).
+
+TableGen uses the following target description (``.td``) input files to
+generate much of the code for instruction definition:
+
+* ``Target.td`` --- Where the ``Instruction``, ``Operand``, ``InstrInfo``, and
+  other fundamental classes are defined.
+
+* ``TargetSelectionDAG.td`` --- Used by ``SelectionDAG`` instruction selection
+  generators, contains ``SDTC*`` classes (selection DAG type constraint),
+  definitions of ``SelectionDAG`` nodes (such as ``imm``, ``cond``, ``bb``,
+  ``add``, ``fadd``, ``sub``), and pattern support (``Pattern``, ``Pat``,
+  ``PatFrag``, ``PatLeaf``, ``ComplexPattern``.
+
+* ``XXXInstrFormats.td`` --- Patterns for definitions of target-specific
+  instructions.
+
+* ``XXXInstrInfo.td`` --- Target-specific definitions of instruction templates,
+  condition codes, and instructions of an instruction set.  For architecture
+  modifications, a different file name may be used.  For example, for Pentium
+  with SSE instruction, this file is ``X86InstrSSE.td``, and for Pentium with
+  MMX, this file is ``X86InstrMMX.td``.
+
+There is also a target-specific ``XXX.td`` file, where ``XXX`` is the name of
+the target.  The ``XXX.td`` file includes the other ``.td`` input files, but
+its contents are only directly important for subtargets.
+
+You should describe a concrete target-specific class ``XXXInstrInfo`` that
+represents machine instructions supported by a target machine.
+``XXXInstrInfo`` contains an array of ``XXXInstrDescriptor`` objects, each of
+which describes one instruction.  An instruction descriptor defines:
+
+* Opcode mnemonic
+* Number of operands
+* List of implicit register definitions and uses
+* Target-independent properties (such as memory access, is commutable)
+* Target-specific flags
+
+The Instruction class (defined in ``Target.td``) is mostly used as a base for
+more complex instruction classes.
+
+.. code-block:: text
+
+  class Instruction {
+    string Namespace = "";
+    dag OutOperandList;    // A dag containing the MI def operand list.
+    dag InOperandList;     // A dag containing the MI use operand list.
+    string AsmString = ""; // The .s format to print the instruction with.
+    list<dag> Pattern;     // Set to the DAG pattern for this instruction.
+    list<Register> Uses = [];
+    list<Register> Defs = [];
+    list<Predicate> Predicates = [];  // predicates turned into isel match code
+    ... remainder not shown for space ...
+  }
+
+A ``SelectionDAG`` node (``SDNode``) should contain an object representing a
+target-specific instruction that is defined in ``XXXInstrInfo.td``.  The
+instruction objects should represent instructions from the architecture manual
+of the target machine (such as the SPARC Architecture Manual for the SPARC
+target).
+
+A single instruction from the architecture manual is often modeled as multiple
+target instructions, depending upon its operands.  For example, a manual might
+describe an add instruction that takes a register or an immediate operand.  An
+LLVM target could model this with two instructions named ``ADDri`` and
+``ADDrr``.
+
+You should define a class for each instruction category and define each opcode
+as a subclass of the category with appropriate parameters such as the fixed
+binary encoding of opcodes and extended opcodes.  You should map the register
+bits to the bits of the instruction in which they are encoded (for the JIT).
+Also you should specify how the instruction should be printed when the
+automatic assembly printer is used.
+
+As is described in the SPARC Architecture Manual, Version 8, there are three
+major 32-bit formats for instructions.  Format 1 is only for the ``CALL``
+instruction.  Format 2 is for branch on condition codes and ``SETHI`` (set high
+bits of a register) instructions.  Format 3 is for other instructions.
+
+Each of these formats has corresponding classes in ``SparcInstrFormat.td``.
+``InstSP`` is a base class for other instruction classes.  Additional base
+classes are specified for more precise formats: for example in
+``SparcInstrFormat.td``, ``F2_1`` is for ``SETHI``, and ``F2_2`` is for
+branches.  There are three other base classes: ``F3_1`` for register/register
+operations, ``F3_2`` for register/immediate operations, and ``F3_3`` for
+floating-point operations.  ``SparcInstrInfo.td`` also adds the base class
+``Pseudo`` for synthetic SPARC instructions.
+
+``SparcInstrInfo.td`` largely consists of operand and instruction definitions
+for the SPARC target.  In ``SparcInstrInfo.td``, the following target
+description file entry, ``LDrr``, defines the Load Integer instruction for a
+Word (the ``LD`` SPARC opcode) from a memory address to a register.  The first
+parameter, the value 3 (``11``\ :sub:`2`), is the operation value for this
+category of operation.  The second parameter (``000000``\ :sub:`2`) is the
+specific operation value for ``LD``/Load Word.  The third parameter is the
+output destination, which is a register operand and defined in the ``Register``
+target description file (``IntRegs``).
+
+.. code-block:: text
+
+  def LDrr : F3_1 <3, 0b000000, (outs IntRegs:$dst), (ins MEMrr:$addr),
+                   "ld [$addr], $dst",
+                   [(set i32:$dst, (load ADDRrr:$addr))]>;
+
+The fourth parameter is the input source, which uses the address operand
+``MEMrr`` that is defined earlier in ``SparcInstrInfo.td``:
+
+.. code-block:: text
+
+  def MEMrr : Operand<i32> {
+    let PrintMethod = "printMemOperand";
+    let MIOperandInfo = (ops IntRegs, IntRegs);
+  }
+
+The fifth parameter is a string that is used by the assembly printer and can be
+left as an empty string until the assembly printer interface is implemented.
+The sixth and final parameter is the pattern used to match the instruction
+during the SelectionDAG Select Phase described in :doc:`CodeGenerator`.
+This parameter is detailed in the next section, :ref:`instruction-selector`.
+
+Instruction class definitions are not overloaded for different operand types,
+so separate versions of instructions are needed for register, memory, or
+immediate value operands.  For example, to perform a Load Integer instruction
+for a Word from an immediate operand to a register, the following instruction
+class is defined:
+
+.. code-block:: text
+
+  def LDri : F3_2 <3, 0b000000, (outs IntRegs:$dst), (ins MEMri:$addr),
+                   "ld [$addr], $dst",
+                   [(set i32:$dst, (load ADDRri:$addr))]>;
+
+Writing these definitions for so many similar instructions can involve a lot of
+cut and paste.  In ``.td`` files, the ``multiclass`` directive enables the
+creation of templates to define several instruction classes at once (using the
+``defm`` directive).  For example in ``SparcInstrInfo.td``, the ``multiclass``
+pattern ``F3_12`` is defined to create 2 instruction classes each time
+``F3_12`` is invoked:
+
+.. code-block:: text
+
+  multiclass F3_12 <string OpcStr, bits<6> Op3Val, SDNode OpNode> {
+    def rr  : F3_1 <2, Op3Val,
+                   (outs IntRegs:$dst), (ins IntRegs:$b, IntRegs:$c),
+                   !strconcat(OpcStr, " $b, $c, $dst"),
+                   [(set i32:$dst, (OpNode i32:$b, i32:$c))]>;
+    def ri  : F3_2 <2, Op3Val,
+                   (outs IntRegs:$dst), (ins IntRegs:$b, i32imm:$c),
+                   !strconcat(OpcStr, " $b, $c, $dst"),
+                   [(set i32:$dst, (OpNode i32:$b, simm13:$c))]>;
+  }
+
+So when the ``defm`` directive is used for the ``XOR`` and ``ADD``
+instructions, as seen below, it creates four instruction objects: ``XORrr``,
+``XORri``, ``ADDrr``, and ``ADDri``.
+
+.. code-block:: text
+
+  defm XOR   : F3_12<"xor", 0b000011, xor>;
+  defm ADD   : F3_12<"add", 0b000000, add>;
+
+``SparcInstrInfo.td`` also includes definitions for condition codes that are
+referenced by branch instructions.  The following definitions in
+``SparcInstrInfo.td`` indicate the bit location of the SPARC condition code.
+For example, the 10\ :sup:`th` bit represents the "greater than" condition for
+integers, and the 22\ :sup:`nd` bit represents the "greater than" condition for
+floats.
+
+.. code-block:: text
+
+  def ICC_NE  : ICC_VAL< 9>;  // Not Equal
+  def ICC_E   : ICC_VAL< 1>;  // Equal
+  def ICC_G   : ICC_VAL<10>;  // Greater
+  ...
+  def FCC_U   : FCC_VAL<23>;  // Unordered
+  def FCC_G   : FCC_VAL<22>;  // Greater
+  def FCC_UG  : FCC_VAL<21>;  // Unordered or Greater
+  ...
+
+(Note that ``Sparc.h`` also defines enums that correspond to the same SPARC
+condition codes.  Care must be taken to ensure the values in ``Sparc.h``
+correspond to the values in ``SparcInstrInfo.td``.  I.e., ``SPCC::ICC_NE = 9``,
+``SPCC::FCC_U = 23`` and so on.)
+
+Instruction Operand Mapping
+---------------------------
+
+The code generator backend maps instruction operands to fields in the
+instruction.  Operands are assigned to unbound fields in the instruction in the
+order they are defined.  Fields are bound when they are assigned a value.  For
+example, the Sparc target defines the ``XNORrr`` instruction as a ``F3_1``
+format instruction having three operands.
+
+.. code-block:: text
+
+  def XNORrr  : F3_1<2, 0b000111,
+                     (outs IntRegs:$dst), (ins IntRegs:$b, IntRegs:$c),
+                     "xnor $b, $c, $dst",
+                     [(set i32:$dst, (not (xor i32:$b, i32:$c)))]>;
+
+The instruction templates in ``SparcInstrFormats.td`` show the base class for
+``F3_1`` is ``InstSP``.
+
+.. code-block:: text
+
+  class InstSP<dag outs, dag ins, string asmstr, list<dag> pattern> : Instruction {
+    field bits<32> Inst;
+    let Namespace = "SP";
+    bits<2> op;
+    let Inst{31-30} = op;
+    dag OutOperandList = outs;
+    dag InOperandList = ins;
+    let AsmString   = asmstr;
+    let Pattern = pattern;
+  }
+
+``InstSP`` leaves the ``op`` field unbound.
+
+.. code-block:: text
+
+  class F3<dag outs, dag ins, string asmstr, list<dag> pattern>
+      : InstSP<outs, ins, asmstr, pattern> {
+    bits<5> rd;
+    bits<6> op3;
+    bits<5> rs1;
+    let op{1} = 1;   // Op = 2 or 3
+    let Inst{29-25} = rd;
+    let Inst{24-19} = op3;
+    let Inst{18-14} = rs1;
+  }
+
+``F3`` binds the ``op`` field and defines the ``rd``, ``op3``, and ``rs1``
+fields.  ``F3`` format instructions will bind the operands ``rd``, ``op3``, and
+``rs1`` fields.
+
+.. code-block:: text
+
+  class F3_1<bits<2> opVal, bits<6> op3val, dag outs, dag ins,
+             string asmstr, list<dag> pattern> : F3<outs, ins, asmstr, pattern> {
+    bits<8> asi = 0; // asi not currently used
+    bits<5> rs2;
+    let op         = opVal;
+    let op3        = op3val;
+    let Inst{13}   = 0;     // i field = 0
+    let Inst{12-5} = asi;   // address space identifier
+    let Inst{4-0}  = rs2;
+  }
+
+``F3_1`` binds the ``op3`` field and defines the ``rs2`` fields.  ``F3_1``
+format instructions will bind the operands to the ``rd``, ``rs1``, and ``rs2``
+fields.  This results in the ``XNORrr`` instruction binding ``$dst``, ``$b``,
+and ``$c`` operands to the ``rd``, ``rs1``, and ``rs2`` fields respectively.
+
+Instruction Operand Name Mapping
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+TableGen will also generate a function called getNamedOperandIdx() which
+can be used to look up an operand's index in a MachineInstr based on its
+TableGen name.  Setting the UseNamedOperandTable bit in an instruction's
+TableGen definition will add all of its operands to an enumeration in the
+llvm::XXX:OpName namespace and also add an entry for it into the OperandMap
+table, which can be queried using getNamedOperandIdx()
+
+.. code-block:: text
+
+  int DstIndex = SP::getNamedOperandIdx(SP::XNORrr, SP::OpName::dst); // => 0
+  int BIndex = SP::getNamedOperandIdx(SP::XNORrr, SP::OpName::b);     // => 1
+  int CIndex = SP::getNamedOperandIdx(SP::XNORrr, SP::OpName::c);     // => 2
+  int DIndex = SP::getNamedOperandIdx(SP::XNORrr, SP::OpName::d);     // => -1
+
+  ...
+
+The entries in the OpName enum are taken verbatim from the TableGen definitions,
+so operands with lowercase names will have lower case entries in the enum.
+
+To include the getNamedOperandIdx() function in your backend, you will need
+to define a few preprocessor macros in XXXInstrInfo.cpp and XXXInstrInfo.h.
+For example:
+
+XXXInstrInfo.cpp:
+
+.. code-block:: c++ 
+
+  #define GET_INSTRINFO_NAMED_OPS // For getNamedOperandIdx() function
+  #include "XXXGenInstrInfo.inc"
+
+XXXInstrInfo.h:
+
+.. code-block:: c++
+
+  #define GET_INSTRINFO_OPERAND_ENUM // For OpName enum
+  #include "XXXGenInstrInfo.inc"
+
+  namespace XXX {
+    int16_t getNamedOperandIdx(uint16_t Opcode, uint16_t NamedIndex);
+  } // End namespace XXX
+
+Instruction Operand Types
+^^^^^^^^^^^^^^^^^^^^^^^^^
+
+TableGen will also generate an enumeration consisting of all named Operand
+types defined in the backend, in the llvm::XXX::OpTypes namespace.
+Some common immediate Operand types (for instance i8, i32, i64, f32, f64)
+are defined for all targets in ``include/llvm/Target/Target.td``, and are
+available in each Target's OpTypes enum.  Also, only named Operand types appear
+in the enumeration: anonymous types are ignored.
+For example, the X86 backend defines ``brtarget`` and ``brtarget8``, both
+instances of the TableGen ``Operand`` class, which represent branch target
+operands:
+
+.. code-block:: text
+
+  def brtarget : Operand<OtherVT>;
+  def brtarget8 : Operand<OtherVT>;
+
+This results in:
+
+.. code-block:: c++
+
+  namespace X86 {
+  namespace OpTypes {
+  enum OperandType {
+    ...
+    brtarget,
+    brtarget8,
+    ...
+    i32imm,
+    i64imm,
+    ...
+    OPERAND_TYPE_LIST_END
+  } // End namespace OpTypes
+  } // End namespace X86
+
+In typical TableGen fashion, to use the enum, you will need to define a
+preprocessor macro:
+
+.. code-block:: c++
+
+  #define GET_INSTRINFO_OPERAND_TYPES_ENUM // For OpTypes enum
+  #include "XXXGenInstrInfo.inc"
+
+
+Instruction Scheduling
+----------------------
+
+Instruction itineraries can be queried using MCDesc::getSchedClass(). The
+value can be named by an enumemation in llvm::XXX::Sched namespace generated
+by TableGen in XXXGenInstrInfo.inc. The name of the schedule classes are
+the same as provided in XXXSchedule.td plus a default NoItinerary class.
+
+Instruction Relation Mapping
+----------------------------
+
+This TableGen feature is used to relate instructions with each other.  It is
+particularly useful when you have multiple instruction formats and need to
+switch between them after instruction selection.  This entire feature is driven
+by relation models which can be defined in ``XXXInstrInfo.td`` files
+according to the target-specific instruction set.  Relation models are defined
+using ``InstrMapping`` class as a base.  TableGen parses all the models
+and generates instruction relation maps using the specified information.
+Relation maps are emitted as tables in the ``XXXGenInstrInfo.inc`` file
+along with the functions to query them.  For the detailed information on how to
+use this feature, please refer to :doc:`HowToUseInstrMappings`.
+
+Implement a subclass of ``TargetInstrInfo``
+-------------------------------------------
+
+The final step is to hand code portions of ``XXXInstrInfo``, which implements
+the interface described in ``TargetInstrInfo.h`` (see :ref:`TargetInstrInfo`).
+These functions return ``0`` or a Boolean or they assert, unless overridden.
+Here's a list of functions that are overridden for the SPARC implementation in
+``SparcInstrInfo.cpp``:
+
+* ``isLoadFromStackSlot`` --- If the specified machine instruction is a direct
+  load from a stack slot, return the register number of the destination and the
+  ``FrameIndex`` of the stack slot.
+
+* ``isStoreToStackSlot`` --- If the specified machine instruction is a direct
+  store to a stack slot, return the register number of the destination and the
+  ``FrameIndex`` of the stack slot.
+
+* ``copyPhysReg`` --- Copy values between a pair of physical registers.
+
+* ``storeRegToStackSlot`` --- Store a register value to a stack slot.
+
+* ``loadRegFromStackSlot`` --- Load a register value from a stack slot.
+
+* ``storeRegToAddr`` --- Store a register value to memory.
+
+* ``loadRegFromAddr`` --- Load a register value from memory.
+
+* ``foldMemoryOperand`` --- Attempt to combine instructions of any load or
+  store instruction for the specified operand(s).
+
+Branch Folding and If Conversion
+--------------------------------
+
+Performance can be improved by combining instructions or by eliminating
+instructions that are never reached.  The ``AnalyzeBranch`` method in
+``XXXInstrInfo`` may be implemented to examine conditional instructions and
+remove unnecessary instructions.  ``AnalyzeBranch`` looks at the end of a
+machine basic block (MBB) for opportunities for improvement, such as branch
+folding and if conversion.  The ``BranchFolder`` and ``IfConverter`` machine
+function passes (see the source files ``BranchFolding.cpp`` and
+``IfConversion.cpp`` in the ``lib/CodeGen`` directory) call ``AnalyzeBranch``
+to improve the control flow graph that represents the instructions.
+
+Several implementations of ``AnalyzeBranch`` (for ARM, Alpha, and X86) can be
+examined as models for your own ``AnalyzeBranch`` implementation.  Since SPARC
+does not implement a useful ``AnalyzeBranch``, the ARM target implementation is
+shown below.
+
+``AnalyzeBranch`` returns a Boolean value and takes four parameters:
+
+* ``MachineBasicBlock &MBB`` --- The incoming block to be examined.
+
+* ``MachineBasicBlock *&TBB`` --- A destination block that is returned.  For a
+  conditional branch that evaluates to true, ``TBB`` is the destination.
+
+* ``MachineBasicBlock *&FBB`` --- For a conditional branch that evaluates to
+  false, ``FBB`` is returned as the destination.
+
+* ``std::vector<MachineOperand> &Cond`` --- List of operands to evaluate a
+  condition for a conditional branch.
+
+In the simplest case, if a block ends without a branch, then it falls through
+to the successor block.  No destination blocks are specified for either ``TBB``
+or ``FBB``, so both parameters return ``NULL``.  The start of the
+``AnalyzeBranch`` (see code below for the ARM target) shows the function
+parameters and the code for the simplest case.
+
+.. code-block:: c++
+
+  bool ARMInstrInfo::AnalyzeBranch(MachineBasicBlock &MBB,
+                                   MachineBasicBlock *&TBB,
+                                   MachineBasicBlock *&FBB,
+                                   std::vector<MachineOperand> &Cond) const
+  {
+    MachineBasicBlock::iterator I = MBB.end();
+    if (I == MBB.begin() || !isUnpredicatedTerminator(--I))
+      return false;
+
+If a block ends with a single unconditional branch instruction, then
+``AnalyzeBranch`` (shown below) should return the destination of that branch in
+the ``TBB`` parameter.
+
+.. code-block:: c++
+
+    if (LastOpc == ARM::B || LastOpc == ARM::tB) {
+      TBB = LastInst->getOperand(0).getMBB();
+      return false;
+    }
+
+If a block ends with two unconditional branches, then the second branch is
+never reached.  In that situation, as shown below, remove the last branch
+instruction and return the penultimate branch in the ``TBB`` parameter.
+
+.. code-block:: c++
+
+    if ((SecondLastOpc == ARM::B || SecondLastOpc == ARM::tB) &&
+        (LastOpc == ARM::B || LastOpc == ARM::tB)) {
+      TBB = SecondLastInst->getOperand(0).getMBB();
+      I = LastInst;
+      I->eraseFromParent();
+      return false;
+    }
+
+A block may end with a single conditional branch instruction that falls through
+to successor block if the condition evaluates to false.  In that case,
+``AnalyzeBranch`` (shown below) should return the destination of that
+conditional branch in the ``TBB`` parameter and a list of operands in the
+``Cond`` parameter to evaluate the condition.
+
+.. code-block:: c++
+
+    if (LastOpc == ARM::Bcc || LastOpc == ARM::tBcc) {
+      // Block ends with fall-through condbranch.
+      TBB = LastInst->getOperand(0).getMBB();
+      Cond.push_back(LastInst->getOperand(1));
+      Cond.push_back(LastInst->getOperand(2));
+      return false;
+    }
+
+If a block ends with both a conditional branch and an ensuing unconditional
+branch, then ``AnalyzeBranch`` (shown below) should return the conditional
+branch destination (assuming it corresponds to a conditional evaluation of
+"``true``") in the ``TBB`` parameter and the unconditional branch destination
+in the ``FBB`` (corresponding to a conditional evaluation of "``false``").  A
+list of operands to evaluate the condition should be returned in the ``Cond``
+parameter.
+
+.. code-block:: c++
+
+    unsigned SecondLastOpc = SecondLastInst->getOpcode();
+
+    if ((SecondLastOpc == ARM::Bcc && LastOpc == ARM::B) ||
+        (SecondLastOpc == ARM::tBcc && LastOpc == ARM::tB)) {
+      TBB =  SecondLastInst->getOperand(0).getMBB();
+      Cond.push_back(SecondLastInst->getOperand(1));
+      Cond.push_back(SecondLastInst->getOperand(2));
+      FBB = LastInst->getOperand(0).getMBB();
+      return false;
+    }
+
+For the last two cases (ending with a single conditional branch or ending with
+one conditional and one unconditional branch), the operands returned in the
+``Cond`` parameter can be passed to methods of other instructions to create new
+branches or perform other operations.  An implementation of ``AnalyzeBranch``
+requires the helper methods ``RemoveBranch`` and ``InsertBranch`` to manage
+subsequent operations.
+
+``AnalyzeBranch`` should return false indicating success in most circumstances.
+``AnalyzeBranch`` should only return true when the method is stumped about what
+to do, for example, if a block has three terminating branches.
+``AnalyzeBranch`` may return true if it encounters a terminator it cannot
+handle, such as an indirect branch.
+
+.. _instruction-selector:
+
+Instruction Selector
+====================
+
+LLVM uses a ``SelectionDAG`` to represent LLVM IR instructions, and nodes of
+the ``SelectionDAG`` ideally represent native target instructions.  During code
+generation, instruction selection passes are performed to convert non-native
+DAG instructions into native target-specific instructions.  The pass described
+in ``XXXISelDAGToDAG.cpp`` is used to match patterns and perform DAG-to-DAG
+instruction selection.  Optionally, a pass may be defined (in
+``XXXBranchSelector.cpp``) to perform similar DAG-to-DAG operations for branch
+instructions.  Later, the code in ``XXXISelLowering.cpp`` replaces or removes
+operations and data types not supported natively (legalizes) in a
+``SelectionDAG``.
+
+TableGen generates code for instruction selection using the following target
+description input files:
+
+* ``XXXInstrInfo.td`` --- Contains definitions of instructions in a
+  target-specific instruction set, generates ``XXXGenDAGISel.inc``, which is
+  included in ``XXXISelDAGToDAG.cpp``.
+
+* ``XXXCallingConv.td`` --- Contains the calling and return value conventions
+  for the target architecture, and it generates ``XXXGenCallingConv.inc``,
+  which is included in ``XXXISelLowering.cpp``.
+
+The implementation of an instruction selection pass must include a header that
+declares the ``FunctionPass`` class or a subclass of ``FunctionPass``.  In
+``XXXTargetMachine.cpp``, a Pass Manager (PM) should add each instruction
+selection pass into the queue of passes to run.
+
+The LLVM static compiler (``llc``) is an excellent tool for visualizing the
+contents of DAGs.  To display the ``SelectionDAG`` before or after specific
+processing phases, use the command line options for ``llc``, described at
+:ref:`SelectionDAG-Process`.
+
+To describe instruction selector behavior, you should add patterns for lowering
+LLVM code into a ``SelectionDAG`` as the last parameter of the instruction
+definitions in ``XXXInstrInfo.td``.  For example, in ``SparcInstrInfo.td``,
+this entry defines a register store operation, and the last parameter describes
+a pattern with the store DAG operator.
+
+.. code-block:: text
+
+  def STrr  : F3_1< 3, 0b000100, (outs), (ins MEMrr:$addr, IntRegs:$src),
+                   "st $src, [$addr]", [(store i32:$src, ADDRrr:$addr)]>;
+
+``ADDRrr`` is a memory mode that is also defined in ``SparcInstrInfo.td``:
+
+.. code-block:: text
+
+  def ADDRrr : ComplexPattern<i32, 2, "SelectADDRrr", [], []>;
+
+The definition of ``ADDRrr`` refers to ``SelectADDRrr``, which is a function
+defined in an implementation of the Instructor Selector (such as
+``SparcISelDAGToDAG.cpp``).
+
+In ``lib/Target/TargetSelectionDAG.td``, the DAG operator for store is defined
+below:
+
+.. code-block:: text
+
+  def store : PatFrag<(ops node:$val, node:$ptr),
+                      (st node:$val, node:$ptr), [{
+    if (StoreSDNode *ST = dyn_cast<StoreSDNode>(N))
+      return !ST->isTruncatingStore() &&
+             ST->getAddressingMode() == ISD::UNINDEXED;
+    return false;
+  }]>;
+
+``XXXInstrInfo.td`` also generates (in ``XXXGenDAGISel.inc``) the
+``SelectCode`` method that is used to call the appropriate processing method
+for an instruction.  In this example, ``SelectCode`` calls ``Select_ISD_STORE``
+for the ``ISD::STORE`` opcode.
+
+.. code-block:: c++
+
+  SDNode *SelectCode(SDValue N) {
+    ...
+    MVT::ValueType NVT = N.getNode()->getValueType(0);
+    switch (N.getOpcode()) {
+    case ISD::STORE: {
+      switch (NVT) {
+      default:
+        return Select_ISD_STORE(N);
+        break;
+      }
+      break;
+    }
+    ...
+
+The pattern for ``STrr`` is matched, so elsewhere in ``XXXGenDAGISel.inc``,
+code for ``STrr`` is created for ``Select_ISD_STORE``.  The ``Emit_22`` method
+is also generated in ``XXXGenDAGISel.inc`` to complete the processing of this
+instruction.
+
+.. code-block:: c++
+
+  SDNode *Select_ISD_STORE(const SDValue &N) {
+    SDValue Chain = N.getOperand(0);
+    if (Predicate_store(N.getNode())) {
+      SDValue N1 = N.getOperand(1);
+      SDValue N2 = N.getOperand(2);
+      SDValue CPTmp0;
+      SDValue CPTmp1;
+
+      // Pattern: (st:void i32:i32:$src,
+      //           ADDRrr:i32:$addr)<<P:Predicate_store>>
+      // Emits: (STrr:void ADDRrr:i32:$addr, IntRegs:i32:$src)
+      // Pattern complexity = 13  cost = 1  size = 0
+      if (SelectADDRrr(N, N2, CPTmp0, CPTmp1) &&
+          N1.getNode()->getValueType(0) == MVT::i32 &&
+          N2.getNode()->getValueType(0) == MVT::i32) {
+        return Emit_22(N, SP::STrr, CPTmp0, CPTmp1);
+      }
+  ...
+
+The SelectionDAG Legalize Phase
+-------------------------------
+
+The Legalize phase converts a DAG to use types and operations that are natively
+supported by the target.  For natively unsupported types and operations, you
+need to add code to the target-specific ``XXXTargetLowering`` implementation to
+convert unsupported types and operations to supported ones.
+
+In the constructor for the ``XXXTargetLowering`` class, first use the
+``addRegisterClass`` method to specify which types are supported and which
+register classes are associated with them.  The code for the register classes
+are generated by TableGen from ``XXXRegisterInfo.td`` and placed in
+``XXXGenRegisterInfo.h.inc``.  For example, the implementation of the
+constructor for the SparcTargetLowering class (in ``SparcISelLowering.cpp``)
+starts with the following code:
+
+.. code-block:: c++
+
+  addRegisterClass(MVT::i32, SP::IntRegsRegisterClass);
+  addRegisterClass(MVT::f32, SP::FPRegsRegisterClass);
+  addRegisterClass(MVT::f64, SP::DFPRegsRegisterClass);
+
+You should examine the node types in the ``ISD`` namespace
+(``include/llvm/CodeGen/SelectionDAGNodes.h``) and determine which operations
+the target natively supports.  For operations that do **not** have native
+support, add a callback to the constructor for the ``XXXTargetLowering`` class,
+so the instruction selection process knows what to do.  The ``TargetLowering``
+class callback methods (declared in ``llvm/Target/TargetLowering.h``) are:
+
+* ``setOperationAction`` --- General operation.
+* ``setLoadExtAction`` --- Load with extension.
+* ``setTruncStoreAction`` --- Truncating store.
+* ``setIndexedLoadAction`` --- Indexed load.
+* ``setIndexedStoreAction`` --- Indexed store.
+* ``setConvertAction`` --- Type conversion.
+* ``setCondCodeAction`` --- Support for a given condition code.
+
+Note: on older releases, ``setLoadXAction`` is used instead of
+``setLoadExtAction``.  Also, on older releases, ``setCondCodeAction`` may not
+be supported.  Examine your release to see what methods are specifically
+supported.
+
+These callbacks are used to determine that an operation does or does not work
+with a specified type (or types).  And in all cases, the third parameter is a
+``LegalAction`` type enum value: ``Promote``, ``Expand``, ``Custom``, or
+``Legal``.  ``SparcISelLowering.cpp`` contains examples of all four
+``LegalAction`` values.
+
+Promote
+^^^^^^^
+
+For an operation without native support for a given type, the specified type
+may be promoted to a larger type that is supported.  For example, SPARC does
+not support a sign-extending load for Boolean values (``i1`` type), so in
+``SparcISelLowering.cpp`` the third parameter below, ``Promote``, changes
+``i1`` type values to a large type before loading.
+
+.. code-block:: c++
+
+  setLoadExtAction(ISD::SEXTLOAD, MVT::i1, Promote);
+
+Expand
+^^^^^^
+
+For a type without native support, a value may need to be broken down further,
+rather than promoted.  For an operation without native support, a combination
+of other operations may be used to similar effect.  In SPARC, the
+floating-point sine and cosine trig operations are supported by expansion to
+other operations, as indicated by the third parameter, ``Expand``, to
+``setOperationAction``:
+
+.. code-block:: c++
+
+  setOperationAction(ISD::FSIN, MVT::f32, Expand);
+  setOperationAction(ISD::FCOS, MVT::f32, Expand);
+
+Custom
+^^^^^^
+
+For some operations, simple type promotion or operation expansion may be
+insufficient.  In some cases, a special intrinsic function must be implemented.
+
+For example, a constant value may require special treatment, or an operation
+may require spilling and restoring registers in the stack and working with
+register allocators.
+
+As seen in ``SparcISelLowering.cpp`` code below, to perform a type conversion
+from a floating point value to a signed integer, first the
+``setOperationAction`` should be called with ``Custom`` as the third parameter:
+
+.. code-block:: c++
+
+  setOperationAction(ISD::FP_TO_SINT, MVT::i32, Custom);
+
+In the ``LowerOperation`` method, for each ``Custom`` operation, a case
+statement should be added to indicate what function to call.  In the following
+code, an ``FP_TO_SINT`` opcode will call the ``LowerFP_TO_SINT`` method:
+
+.. code-block:: c++
+
+  SDValue SparcTargetLowering::LowerOperation(SDValue Op, SelectionDAG &DAG) {
+    switch (Op.getOpcode()) {
+    case ISD::FP_TO_SINT: return LowerFP_TO_SINT(Op, DAG);
+    ...
+    }
+  }
+
+Finally, the ``LowerFP_TO_SINT`` method is implemented, using an FP register to
+convert the floating-point value to an integer.
+
+.. code-block:: c++
+
+  static SDValue LowerFP_TO_SINT(SDValue Op, SelectionDAG &DAG) {
+    assert(Op.getValueType() == MVT::i32);
+    Op = DAG.getNode(SPISD::FTOI, MVT::f32, Op.getOperand(0));
+    return DAG.getNode(ISD::BITCAST, MVT::i32, Op);
+  }
+
+Legal
+^^^^^
+
+The ``Legal`` ``LegalizeAction`` enum value simply indicates that an operation
+**is** natively supported.  ``Legal`` represents the default condition, so it
+is rarely used.  In ``SparcISelLowering.cpp``, the action for ``CTPOP`` (an
+operation to count the bits set in an integer) is natively supported only for
+SPARC v9.  The following code enables the ``Expand`` conversion technique for
+non-v9 SPARC implementations.
+
+.. code-block:: c++
+
+  setOperationAction(ISD::CTPOP, MVT::i32, Expand);
+  ...
+  if (TM.getSubtarget<SparcSubtarget>().isV9())
+    setOperationAction(ISD::CTPOP, MVT::i32, Legal);
+
+Calling Conventions
+-------------------
+
+To support target-specific calling conventions, ``XXXGenCallingConv.td`` uses
+interfaces (such as ``CCIfType`` and ``CCAssignToReg``) that are defined in
+``lib/Target/TargetCallingConv.td``.  TableGen can take the target descriptor
+file ``XXXGenCallingConv.td`` and generate the header file
+``XXXGenCallingConv.inc``, which is typically included in
+``XXXISelLowering.cpp``.  You can use the interfaces in
+``TargetCallingConv.td`` to specify:
+
+* The order of parameter allocation.
+
+* Where parameters and return values are placed (that is, on the stack or in
+  registers).
+
+* Which registers may be used.
+
+* Whether the caller or callee unwinds the stack.
+
+The following example demonstrates the use of the ``CCIfType`` and
+``CCAssignToReg`` interfaces.  If the ``CCIfType`` predicate is true (that is,
+if the current argument is of type ``f32`` or ``f64``), then the action is
+performed.  In this case, the ``CCAssignToReg`` action assigns the argument
+value to the first available register: either ``R0`` or ``R1``.
+
+.. code-block:: text
+
+  CCIfType<[f32,f64], CCAssignToReg<[R0, R1]>>
+
+``SparcCallingConv.td`` contains definitions for a target-specific return-value
+calling convention (``RetCC_Sparc32``) and a basic 32-bit C calling convention
+(``CC_Sparc32``).  The definition of ``RetCC_Sparc32`` (shown below) indicates
+which registers are used for specified scalar return types.  A single-precision
+float is returned to register ``F0``, and a double-precision float goes to
+register ``D0``.  A 32-bit integer is returned in register ``I0`` or ``I1``.
+
+.. code-block:: text
+
+  def RetCC_Sparc32 : CallingConv<[
+    CCIfType<[i32], CCAssignToReg<[I0, I1]>>,
+    CCIfType<[f32], CCAssignToReg<[F0]>>,
+    CCIfType<[f64], CCAssignToReg<[D0]>>
+  ]>;
+
+The definition of ``CC_Sparc32`` in ``SparcCallingConv.td`` introduces
+``CCAssignToStack``, which assigns the value to a stack slot with the specified
+size and alignment.  In the example below, the first parameter, 4, indicates
+the size of the slot, and the second parameter, also 4, indicates the stack
+alignment along 4-byte units.  (Special cases: if size is zero, then the ABI
+size is used; if alignment is zero, then the ABI alignment is used.)
+
+.. code-block:: text
+
+  def CC_Sparc32 : CallingConv<[
+    // All arguments get passed in integer registers if there is space.
+    CCIfType<[i32, f32, f64], CCAssignToReg<[I0, I1, I2, I3, I4, I5]>>,
+    CCAssignToStack<4, 4>
+  ]>;
+
+``CCDelegateTo`` is another commonly used interface, which tries to find a
+specified sub-calling convention, and, if a match is found, it is invoked.  In
+the following example (in ``X86CallingConv.td``), the definition of
+``RetCC_X86_32_C`` ends with ``CCDelegateTo``.  After the current value is
+assigned to the register ``ST0`` or ``ST1``, the ``RetCC_X86Common`` is
+invoked.
+
+.. code-block:: text
+
+  def RetCC_X86_32_C : CallingConv<[
+    CCIfType<[f32], CCAssignToReg<[ST0, ST1]>>,
+    CCIfType<[f64], CCAssignToReg<[ST0, ST1]>>,
+    CCDelegateTo<RetCC_X86Common>
+  ]>;
+
+``CCIfCC`` is an interface that attempts to match the given name to the current
+calling convention.  If the name identifies the current calling convention,
+then a specified action is invoked.  In the following example (in
+``X86CallingConv.td``), if the ``Fast`` calling convention is in use, then
+``RetCC_X86_32_Fast`` is invoked.  If the ``SSECall`` calling convention is in
+use, then ``RetCC_X86_32_SSE`` is invoked.
+
+.. code-block:: text
+
+  def RetCC_X86_32 : CallingConv<[
+    CCIfCC<"CallingConv::Fast", CCDelegateTo<RetCC_X86_32_Fast>>,
+    CCIfCC<"CallingConv::X86_SSECall", CCDelegateTo<RetCC_X86_32_SSE>>,
+    CCDelegateTo<RetCC_X86_32_C>
+  ]>;
+
+Other calling convention interfaces include:
+
+* ``CCIf <predicate, action>`` --- If the predicate matches, apply the action.
+
+* ``CCIfInReg <action>`` --- If the argument is marked with the "``inreg``"
+  attribute, then apply the action.
+
+* ``CCIfNest <action>`` --- If the argument is marked with the "``nest``"
+  attribute, then apply the action.
+
+* ``CCIfNotVarArg <action>`` --- If the current function does not take a
+  variable number of arguments, apply the action.
+
+* ``CCAssignToRegWithShadow <registerList, shadowList>`` --- similar to
+  ``CCAssignToReg``, but with a shadow list of registers.
+
+* ``CCPassByVal <size, align>`` --- Assign value to a stack slot with the
+  minimum specified size and alignment.
+
+* ``CCPromoteToType <type>`` --- Promote the current value to the specified
+  type.
+
+* ``CallingConv <[actions]>`` --- Define each calling convention that is
+  supported.
+
+Assembly Printer
+================
+
+During the code emission stage, the code generator may utilize an LLVM pass to
+produce assembly output.  To do this, you want to implement the code for a
+printer that converts LLVM IR to a GAS-format assembly language for your target
+machine, using the following steps:
+
+* Define all the assembly strings for your target, adding them to the
+  instructions defined in the ``XXXInstrInfo.td`` file.  (See
+  :ref:`instruction-set`.)  TableGen will produce an output file
+  (``XXXGenAsmWriter.inc``) with an implementation of the ``printInstruction``
+  method for the ``XXXAsmPrinter`` class.
+
+* Write ``XXXTargetAsmInfo.h``, which contains the bare-bones declaration of
+  the ``XXXTargetAsmInfo`` class (a subclass of ``TargetAsmInfo``).
+
+* Write ``XXXTargetAsmInfo.cpp``, which contains target-specific values for
+  ``TargetAsmInfo`` properties and sometimes new implementations for methods.
+
+* Write ``XXXAsmPrinter.cpp``, which implements the ``AsmPrinter`` class that
+  performs the LLVM-to-assembly conversion.
+
+The code in ``XXXTargetAsmInfo.h`` is usually a trivial declaration of the
+``XXXTargetAsmInfo`` class for use in ``XXXTargetAsmInfo.cpp``.  Similarly,
+``XXXTargetAsmInfo.cpp`` usually has a few declarations of ``XXXTargetAsmInfo``
+replacement values that override the default values in ``TargetAsmInfo.cpp``.
+For example in ``SparcTargetAsmInfo.cpp``:
+
+.. code-block:: c++
+
+  SparcTargetAsmInfo::SparcTargetAsmInfo(const SparcTargetMachine &TM) {
+    Data16bitsDirective = "\t.half\t";
+    Data32bitsDirective = "\t.word\t";
+    Data64bitsDirective = 0;  // .xword is only supported by V9.
+    ZeroDirective = "\t.skip\t";
+    CommentString = "!";
+    ConstantPoolSection = "\t.section \".rodata\",#alloc\n";
+  }
+
+The X86 assembly printer implementation (``X86TargetAsmInfo``) is an example
+where the target specific ``TargetAsmInfo`` class uses an overridden methods:
+``ExpandInlineAsm``.
+
+A target-specific implementation of ``AsmPrinter`` is written in
+``XXXAsmPrinter.cpp``, which implements the ``AsmPrinter`` class that converts
+the LLVM to printable assembly.  The implementation must include the following
+headers that have declarations for the ``AsmPrinter`` and
+``MachineFunctionPass`` classes.  The ``MachineFunctionPass`` is a subclass of
+``FunctionPass``.
+
+.. code-block:: c++
+
+  #include "llvm/CodeGen/AsmPrinter.h"
+  #include "llvm/CodeGen/MachineFunctionPass.h"
+
+As a ``FunctionPass``, ``AsmPrinter`` first calls ``doInitialization`` to set
+up the ``AsmPrinter``.  In ``SparcAsmPrinter``, a ``Mangler`` object is
+instantiated to process variable names.
+
+In ``XXXAsmPrinter.cpp``, the ``runOnMachineFunction`` method (declared in
+``MachineFunctionPass``) must be implemented for ``XXXAsmPrinter``.  In
+``MachineFunctionPass``, the ``runOnFunction`` method invokes
+``runOnMachineFunction``.  Target-specific implementations of
+``runOnMachineFunction`` differ, but generally do the following to process each
+machine function:
+
+* Call ``SetupMachineFunction`` to perform initialization.
+
+* Call ``EmitConstantPool`` to print out (to the output stream) constants which
+  have been spilled to memory.
+
+* Call ``EmitJumpTableInfo`` to print out jump tables used by the current
+  function.
+
+* Print out the label for the current function.
+
+* Print out the code for the function, including basic block labels and the
+  assembly for the instruction (using ``printInstruction``)
+
+The ``XXXAsmPrinter`` implementation must also include the code generated by
+TableGen that is output in the ``XXXGenAsmWriter.inc`` file.  The code in
+``XXXGenAsmWriter.inc`` contains an implementation of the ``printInstruction``
+method that may call these methods:
+
+* ``printOperand``
+* ``printMemOperand``
+* ``printCCOperand`` (for conditional statements)
+* ``printDataDirective``
+* ``printDeclare``
+* ``printImplicitDef``
+* ``printInlineAsm``
+
+The implementations of ``printDeclare``, ``printImplicitDef``,
+``printInlineAsm``, and ``printLabel`` in ``AsmPrinter.cpp`` are generally
+adequate for printing assembly and do not need to be overridden.
+
+The ``printOperand`` method is implemented with a long ``switch``/``case``
+statement for the type of operand: register, immediate, basic block, external
+symbol, global address, constant pool index, or jump table index.  For an
+instruction with a memory address operand, the ``printMemOperand`` method
+should be implemented to generate the proper output.  Similarly,
+``printCCOperand`` should be used to print a conditional operand.
+
+``doFinalization`` should be overridden in ``XXXAsmPrinter``, and it should be
+called to shut down the assembly printer.  During ``doFinalization``, global
+variables and constants are printed to output.
+
+Subtarget Support
+=================
+
+Subtarget support is used to inform the code generation process of instruction
+set variations for a given chip set.  For example, the LLVM SPARC
+implementation provided covers three major versions of the SPARC microprocessor
+architecture: Version 8 (V8, which is a 32-bit architecture), Version 9 (V9, a
+64-bit architecture), and the UltraSPARC architecture.  V8 has 16
+double-precision floating-point registers that are also usable as either 32
+single-precision or 8 quad-precision registers.  V8 is also purely big-endian.
+V9 has 32 double-precision floating-point registers that are also usable as 16
+quad-precision registers, but cannot be used as single-precision registers.
+The UltraSPARC architecture combines V9 with UltraSPARC Visual Instruction Set
+extensions.
+
+If subtarget support is needed, you should implement a target-specific
+``XXXSubtarget`` class for your architecture.  This class should process the
+command-line options ``-mcpu=`` and ``-mattr=``.
+
+TableGen uses definitions in the ``Target.td`` and ``Sparc.td`` files to
+generate code in ``SparcGenSubtarget.inc``.  In ``Target.td``, shown below, the
+``SubtargetFeature`` interface is defined.  The first 4 string parameters of
+the ``SubtargetFeature`` interface are a feature name, an attribute set by the
+feature, the value of the attribute, and a description of the feature.  (The
+fifth parameter is a list of features whose presence is implied, and its
+default value is an empty array.)
+
+.. code-block:: text
+
+  class SubtargetFeature<string n, string a, string v, string d,
+                         list<SubtargetFeature> i = []> {
+    string Name = n;
+    string Attribute = a;
+    string Value = v;
+    string Desc = d;
+    list<SubtargetFeature> Implies = i;
+  }
+
+In the ``Sparc.td`` file, the ``SubtargetFeature`` is used to define the
+following features.
+
+.. code-block:: text
+
+  def FeatureV9 : SubtargetFeature<"v9", "IsV9", "true",
+                       "Enable SPARC-V9 instructions">;
+  def FeatureV8Deprecated : SubtargetFeature<"deprecated-v8",
+                       "V8DeprecatedInsts", "true",
+                       "Enable deprecated V8 instructions in V9 mode">;
+  def FeatureVIS : SubtargetFeature<"vis", "IsVIS", "true",
+                       "Enable UltraSPARC Visual Instruction Set extensions">;
+
+Elsewhere in ``Sparc.td``, the ``Proc`` class is defined and then is used to
+define particular SPARC processor subtypes that may have the previously
+described features.
+
+.. code-block:: text
+
+  class Proc<string Name, list<SubtargetFeature> Features>
+    : Processor<Name, NoItineraries, Features>;
+
+  def : Proc<"generic",         []>;
+  def : Proc<"v8",              []>;
+  def : Proc<"supersparc",      []>;
+  def : Proc<"sparclite",       []>;
+  def : Proc<"f934",            []>;
+  def : Proc<"hypersparc",      []>;
+  def : Proc<"sparclite86x",    []>;
+  def : Proc<"sparclet",        []>;
+  def : Proc<"tsc701",          []>;
+  def : Proc<"v9",              [FeatureV9]>;
+  def : Proc<"ultrasparc",      [FeatureV9, FeatureV8Deprecated]>;
+  def : Proc<"ultrasparc3",     [FeatureV9, FeatureV8Deprecated]>;
+  def : Proc<"ultrasparc3-vis", [FeatureV9, FeatureV8Deprecated, FeatureVIS]>;
+
+From ``Target.td`` and ``Sparc.td`` files, the resulting
+``SparcGenSubtarget.inc`` specifies enum values to identify the features,
+arrays of constants to represent the CPU features and CPU subtypes, and the
+``ParseSubtargetFeatures`` method that parses the features string that sets
+specified subtarget options.  The generated ``SparcGenSubtarget.inc`` file
+should be included in the ``SparcSubtarget.cpp``.  The target-specific
+implementation of the ``XXXSubtarget`` method should follow this pseudocode:
+
+.. code-block:: c++
+
+  XXXSubtarget::XXXSubtarget(const Module &M, const std::string &FS) {
+    // Set the default features
+    // Determine default and user specified characteristics of the CPU
+    // Call ParseSubtargetFeatures(FS, CPU) to parse the features string
+    // Perform any additional operations
+  }
+
+JIT Support
+===========
+
+The implementation of a target machine optionally includes a Just-In-Time (JIT)
+code generator that emits machine code and auxiliary structures as binary
+output that can be written directly to memory.  To do this, implement JIT code
+generation by performing the following steps:
+
+* Write an ``XXXCodeEmitter.cpp`` file that contains a machine function pass
+  that transforms target-machine instructions into relocatable machine
+  code.
+
+* Write an ``XXXJITInfo.cpp`` file that implements the JIT interfaces for
+  target-specific code-generation activities, such as emitting machine code and
+  stubs.
+
+* Modify ``XXXTargetMachine`` so that it provides a ``TargetJITInfo`` object
+  through its ``getJITInfo`` method.
+
+There are several different approaches to writing the JIT support code.  For
+instance, TableGen and target descriptor files may be used for creating a JIT
+code generator, but are not mandatory.  For the Alpha and PowerPC target
+machines, TableGen is used to generate ``XXXGenCodeEmitter.inc``, which
+contains the binary coding of machine instructions and the
+``getBinaryCodeForInstr`` method to access those codes.  Other JIT
+implementations do not.
+
+Both ``XXXJITInfo.cpp`` and ``XXXCodeEmitter.cpp`` must include the
+``llvm/CodeGen/MachineCodeEmitter.h`` header file that defines the
+``MachineCodeEmitter`` class containing code for several callback functions
+that write data (in bytes, words, strings, etc.) to the output stream.
+
+Machine Code Emitter
+--------------------
+
+In ``XXXCodeEmitter.cpp``, a target-specific of the ``Emitter`` class is
+implemented as a function pass (subclass of ``MachineFunctionPass``).  The
+target-specific implementation of ``runOnMachineFunction`` (invoked by
+``runOnFunction`` in ``MachineFunctionPass``) iterates through the
+``MachineBasicBlock`` calls ``emitInstruction`` to process each instruction and
+emit binary code.  ``emitInstruction`` is largely implemented with case
+statements on the instruction types defined in ``XXXInstrInfo.h``.  For
+example, in ``X86CodeEmitter.cpp``, the ``emitInstruction`` method is built
+around the following ``switch``/``case`` statements:
+
+.. code-block:: c++
+
+  switch (Desc->TSFlags & X86::FormMask) {
+  case X86II::Pseudo:  // for not yet implemented instructions
+     ...               // or pseudo-instructions
+     break;
+  case X86II::RawFrm:  // for instructions with a fixed opcode value
+     ...
+     break;
+  case X86II::AddRegFrm: // for instructions that have one register operand
+     ...                 // added to their opcode
+     break;
+  case X86II::MRMDestReg:// for instructions that use the Mod/RM byte
+     ...                 // to specify a destination (register)
+     break;
+  case X86II::MRMDestMem:// for instructions that use the Mod/RM byte
+     ...                 // to specify a destination (memory)
+     break;
+  case X86II::MRMSrcReg: // for instructions that use the Mod/RM byte
+     ...                 // to specify a source (register)
+     break;
+  case X86II::MRMSrcMem: // for instructions that use the Mod/RM byte
+     ...                 // to specify a source (memory)
+     break;
+  case X86II::MRM0r: case X86II::MRM1r:  // for instructions that operate on
+  case X86II::MRM2r: case X86II::MRM3r:  // a REGISTER r/m operand and
+  case X86II::MRM4r: case X86II::MRM5r:  // use the Mod/RM byte and a field
+  case X86II::MRM6r: case X86II::MRM7r:  // to hold extended opcode data
+     ...
+     break;
+  case X86II::MRM0m: case X86II::MRM1m:  // for instructions that operate on
+  case X86II::MRM2m: case X86II::MRM3m:  // a MEMORY r/m operand and
+  case X86II::MRM4m: case X86II::MRM5m:  // use the Mod/RM byte and a field
+  case X86II::MRM6m: case X86II::MRM7m:  // to hold extended opcode data
+     ...
+     break;
+  case X86II::MRMInitReg: // for instructions whose source and
+     ...                  // destination are the same register
+     break;
+  }
+
+The implementations of these case statements often first emit the opcode and
+then get the operand(s).  Then depending upon the operand, helper methods may
+be called to process the operand(s).  For example, in ``X86CodeEmitter.cpp``,
+for the ``X86II::AddRegFrm`` case, the first data emitted (by ``emitByte``) is
+the opcode added to the register operand.  Then an object representing the
+machine operand, ``MO1``, is extracted.  The helper methods such as
+``isImmediate``, ``isGlobalAddress``, ``isExternalSymbol``,
+``isConstantPoolIndex``, and ``isJumpTableIndex`` determine the operand type.
+(``X86CodeEmitter.cpp`` also has private methods such as ``emitConstant``,
+``emitGlobalAddress``, ``emitExternalSymbolAddress``, ``emitConstPoolAddress``,
+and ``emitJumpTableAddress`` that emit the data into the output stream.)
+
+.. code-block:: c++
+
+  case X86II::AddRegFrm:
+    MCE.emitByte(BaseOpcode + getX86RegNum(MI.getOperand(CurOp++).getReg()));
+
+    if (CurOp != NumOps) {
+      const MachineOperand &MO1 = MI.getOperand(CurOp++);
+      unsigned Size = X86InstrInfo::sizeOfImm(Desc);
+      if (MO1.isImmediate())
+        emitConstant(MO1.getImm(), Size);
+      else {
+        unsigned rt = Is64BitMode ? X86::reloc_pcrel_word
+          : (IsPIC ? X86::reloc_picrel_word : X86::reloc_absolute_word);
+        if (Opcode == X86::MOV64ri)
+          rt = X86::reloc_absolute_dword;  // FIXME: add X86II flag?
+        if (MO1.isGlobalAddress()) {
+          bool NeedStub = isa<Function>(MO1.getGlobal());
+          bool isLazy = gvNeedsLazyPtr(MO1.getGlobal());
+          emitGlobalAddress(MO1.getGlobal(), rt, MO1.getOffset(), 0,
+                            NeedStub, isLazy);
+        } else if (MO1.isExternalSymbol())
+          emitExternalSymbolAddress(MO1.getSymbolName(), rt);
+        else if (MO1.isConstantPoolIndex())
+          emitConstPoolAddress(MO1.getIndex(), rt);
+        else if (MO1.isJumpTableIndex())
+          emitJumpTableAddress(MO1.getIndex(), rt);
+      }
+    }
+    break;
+
+In the previous example, ``XXXCodeEmitter.cpp`` uses the variable ``rt``, which
+is a ``RelocationType`` enum that may be used to relocate addresses (for
+example, a global address with a PIC base offset).  The ``RelocationType`` enum
+for that target is defined in the short target-specific ``XXXRelocations.h``
+file.  The ``RelocationType`` is used by the ``relocate`` method defined in
+``XXXJITInfo.cpp`` to rewrite addresses for referenced global symbols.
+
+For example, ``X86Relocations.h`` specifies the following relocation types for
+the X86 addresses.  In all four cases, the relocated value is added to the
+value already in memory.  For ``reloc_pcrel_word`` and ``reloc_picrel_word``,
+there is an additional initial adjustment.
+
+.. code-block:: c++
+
+  enum RelocationType {
+    reloc_pcrel_word = 0,    // add reloc value after adjusting for the PC loc
+    reloc_picrel_word = 1,   // add reloc value after adjusting for the PIC base
+    reloc_absolute_word = 2, // absolute relocation; no additional adjustment
+    reloc_absolute_dword = 3 // absolute relocation; no additional adjustment
+  };
+
+Target JIT Info
+---------------
+
+``XXXJITInfo.cpp`` implements the JIT interfaces for target-specific
+code-generation activities, such as emitting machine code and stubs.  At
+minimum, a target-specific version of ``XXXJITInfo`` implements the following:
+
+* ``getLazyResolverFunction`` --- Initializes the JIT, gives the target a
+  function that is used for compilation.
+
+* ``emitFunctionStub`` --- Returns a native function with a specified address
+  for a callback function.
+
+* ``relocate`` --- Changes the addresses of referenced globals, based on
+  relocation types.
+
+* Callback function that are wrappers to a function stub that is used when the
+  real target is not initially known.
+
+``getLazyResolverFunction`` is generally trivial to implement.  It makes the
+incoming parameter as the global ``JITCompilerFunction`` and returns the
+callback function that will be used a function wrapper.  For the Alpha target
+(in ``AlphaJITInfo.cpp``), the ``getLazyResolverFunction`` implementation is
+simply:
+
+.. code-block:: c++
+
+  TargetJITInfo::LazyResolverFn AlphaJITInfo::getLazyResolverFunction(
+                                              JITCompilerFn F) {
+    JITCompilerFunction = F;
+    return AlphaCompilationCallback;
+  }
+
+For the X86 target, the ``getLazyResolverFunction`` implementation is a little
+more complicated, because it returns a different callback function for
+processors with SSE instructions and XMM registers.
+
+The callback function initially saves and later restores the callee register
+values, incoming arguments, and frame and return address.  The callback
+function needs low-level access to the registers or stack, so it is typically
+implemented with assembler.
+

Added: www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMPass.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMPass.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMPass.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/WritingAnLLVMPass.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,1434 @@
+====================
+Writing an LLVM Pass
+====================
+
+.. contents::
+    :local:
+
+Introduction --- What is a pass?
+================================
+
+The LLVM Pass Framework is an important part of the LLVM system, because LLVM
+passes are where most of the interesting parts of the compiler exist.  Passes
+perform the transformations and optimizations that make up the compiler, they
+build the analysis results that are used by these transformations, and they
+are, above all, a structuring technique for compiler code.
+
+All LLVM passes are subclasses of the `Pass
+<http://llvm.org/doxygen/classllvm_1_1Pass.html>`_ class, which implement
+functionality by overriding virtual methods inherited from ``Pass``.  Depending
+on how your pass works, you should inherit from the :ref:`ModulePass
+<writing-an-llvm-pass-ModulePass>` , :ref:`CallGraphSCCPass
+<writing-an-llvm-pass-CallGraphSCCPass>`, :ref:`FunctionPass
+<writing-an-llvm-pass-FunctionPass>` , or :ref:`LoopPass
+<writing-an-llvm-pass-LoopPass>`, or :ref:`RegionPass
+<writing-an-llvm-pass-RegionPass>`, or :ref:`BasicBlockPass
+<writing-an-llvm-pass-BasicBlockPass>` classes, which gives the system more
+information about what your pass does, and how it can be combined with other
+passes.  One of the main features of the LLVM Pass Framework is that it
+schedules passes to run in an efficient way based on the constraints that your
+pass meets (which are indicated by which class they derive from).
+
+We start by showing you how to construct a pass, everything from setting up the
+code, to compiling, loading, and executing it.  After the basics are down, more
+advanced features are discussed.
+
+Quick Start --- Writing hello world
+===================================
+
+Here we describe how to write the "hello world" of passes.  The "Hello" pass is
+designed to simply print out the name of non-external functions that exist in
+the program being compiled.  It does not modify the program at all, it just
+inspects it.  The source code and files for this pass are available in the LLVM
+source tree in the ``lib/Transforms/Hello`` directory.
+
+.. _writing-an-llvm-pass-makefile:
+
+Setting up the build environment
+--------------------------------
+
+First, configure and build LLVM.  Next, you need to create a new directory
+somewhere in the LLVM source base.  For this example, we'll assume that you
+made ``lib/Transforms/Hello``.  Finally, you must set up a build script
+that will compile the source code for the new pass.  To do this,
+copy the following into ``CMakeLists.txt``:
+
+.. code-block:: cmake
+
+  add_llvm_loadable_module( LLVMHello
+    Hello.cpp
+  
+    PLUGIN_TOOL
+    opt
+    )
+
+and the following line into ``lib/Transforms/CMakeLists.txt``:
+
+.. code-block:: cmake
+
+  add_subdirectory(Hello)
+
+(Note that there is already a directory named ``Hello`` with a sample "Hello"
+pass; you may play with it -- in which case you don't need to modify any
+``CMakeLists.txt`` files -- or, if you want to create everything from scratch,
+use another name.)
+
+This build script specifies that ``Hello.cpp`` file in the current directory
+is to be compiled and linked into a shared object ``$(LEVEL)/lib/LLVMHello.so`` that
+can be dynamically loaded by the :program:`opt` tool via its :option:`-load`
+option. If your operating system uses a suffix other than ``.so`` (such as
+Windows or Mac OS X), the appropriate extension will be used.
+
+Now that we have the build scripts set up, we just need to write the code for
+the pass itself.
+
+.. _writing-an-llvm-pass-basiccode:
+
+Basic code required
+-------------------
+
+Now that we have a way to compile our new pass, we just have to write it.
+Start out with:
+
+.. code-block:: c++
+
+  #include "llvm/Pass.h"
+  #include "llvm/IR/Function.h"
+  #include "llvm/Support/raw_ostream.h"
+
+Which are needed because we are writing a `Pass
+<http://llvm.org/doxygen/classllvm_1_1Pass.html>`_, we are operating on
+`Function <http://llvm.org/doxygen/classllvm_1_1Function.html>`_\ s, and we will
+be doing some printing.
+
+Next we have:
+
+.. code-block:: c++
+
+  using namespace llvm;
+
+... which is required because the functions from the include files live in the
+llvm namespace.
+
+Next we have:
+
+.. code-block:: c++
+
+  namespace {
+
+... which starts out an anonymous namespace.  Anonymous namespaces are to C++
+what the "``static``" keyword is to C (at global scope).  It makes the things
+declared inside of the anonymous namespace visible only to the current file.
+If you're not familiar with them, consult a decent C++ book for more
+information.
+
+Next, we declare our pass itself:
+
+.. code-block:: c++
+
+  struct Hello : public FunctionPass {
+
+This declares a "``Hello``" class that is a subclass of :ref:`FunctionPass
+<writing-an-llvm-pass-FunctionPass>`.  The different builtin pass subclasses
+are described in detail :ref:`later <writing-an-llvm-pass-pass-classes>`, but
+for now, know that ``FunctionPass`` operates on a function at a time.
+
+.. code-block:: c++
+
+    static char ID;
+    Hello() : FunctionPass(ID) {}
+
+This declares pass identifier used by LLVM to identify pass.  This allows LLVM
+to avoid using expensive C++ runtime information.
+
+.. code-block:: c++
+
+    bool runOnFunction(Function &F) override {
+      errs() << "Hello: ";
+      errs().write_escaped(F.getName()) << '\n';
+      return false;
+    }
+  }; // end of struct Hello
+  }  // end of anonymous namespace
+
+We declare a :ref:`runOnFunction <writing-an-llvm-pass-runOnFunction>` method,
+which overrides an abstract virtual method inherited from :ref:`FunctionPass
+<writing-an-llvm-pass-FunctionPass>`.  This is where we are supposed to do our
+thing, so we just print out our message with the name of each function.
+
+.. code-block:: c++
+
+  char Hello::ID = 0;
+
+We initialize pass ID here.  LLVM uses ID's address to identify a pass, so
+initialization value is not important.
+
+.. code-block:: c++
+
+  static RegisterPass<Hello> X("hello", "Hello World Pass",
+                               false /* Only looks at CFG */,
+                               false /* Analysis Pass */);
+
+Lastly, we :ref:`register our class <writing-an-llvm-pass-registration>`
+``Hello``, giving it a command line argument "``hello``", and a name "Hello
+World Pass".  The last two arguments describe its behavior: if a pass walks CFG
+without modifying it then the third argument is set to ``true``; if a pass is
+an analysis pass, for example dominator tree pass, then ``true`` is supplied as
+the fourth argument.
+
+As a whole, the ``.cpp`` file looks like:
+
+.. code-block:: c++
+
+  #include "llvm/Pass.h"
+  #include "llvm/IR/Function.h"
+  #include "llvm/Support/raw_ostream.h"
+  
+  using namespace llvm;
+  
+  namespace {
+  struct Hello : public FunctionPass {
+    static char ID;
+    Hello() : FunctionPass(ID) {}
+  
+    bool runOnFunction(Function &F) override {
+      errs() << "Hello: ";
+      errs().write_escaped(F.getName()) << '\n';
+      return false;
+    }
+  }; // end of struct Hello
+  }  // end of anonymous namespace
+  
+  char Hello::ID = 0;
+  static RegisterPass<Hello> X("hello", "Hello World Pass",
+                               false /* Only looks at CFG */,
+                               false /* Analysis Pass */);
+
+Now that it's all together, compile the file with a simple "``gmake``" command
+from the top level of your build directory and you should get a new file
+"``lib/LLVMHello.so``".  Note that everything in this file is
+contained in an anonymous namespace --- this reflects the fact that passes
+are self contained units that do not need external interfaces (although they
+can have them) to be useful.
+
+Running a pass with ``opt``
+---------------------------
+
+Now that you have a brand new shiny shared object file, we can use the
+:program:`opt` command to run an LLVM program through your pass.  Because you
+registered your pass with ``RegisterPass``, you will be able to use the
+:program:`opt` tool to access it, once loaded.
+
+To test it, follow the example at the end of the :doc:`GettingStarted` to
+compile "Hello World" to LLVM.  We can now run the bitcode file (hello.bc) for
+the program through our transformation like this (or course, any bitcode file
+will work):
+
+.. code-block:: console
+
+  $ opt -load lib/LLVMHello.so -hello < hello.bc > /dev/null
+  Hello: __main
+  Hello: puts
+  Hello: main
+
+The :option:`-load` option specifies that :program:`opt` should load your pass
+as a shared object, which makes "``-hello``" a valid command line argument
+(which is one reason you need to :ref:`register your pass
+<writing-an-llvm-pass-registration>`).  Because the Hello pass does not modify
+the program in any interesting way, we just throw away the result of
+:program:`opt` (sending it to ``/dev/null``).
+
+To see what happened to the other string you registered, try running
+:program:`opt` with the :option:`-help` option:
+
+.. code-block:: console
+
+  $ opt -load lib/LLVMHello.so -help
+  OVERVIEW: llvm .bc -> .bc modular optimizer and analysis printer
+
+  USAGE: opt [subcommand] [options] <input bitcode file>
+
+  OPTIONS:
+    Optimizations available:
+  ...
+      -guard-widening           - Widen guards
+      -gvn                      - Global Value Numbering
+      -gvn-hoist                - Early GVN Hoisting of Expressions
+      -hello                    - Hello World Pass
+      -indvars                  - Induction Variable Simplification
+      -inferattrs               - Infer set function attributes
+  ...
+
+The pass name gets added as the information string for your pass, giving some
+documentation to users of :program:`opt`.  Now that you have a working pass,
+you would go ahead and make it do the cool transformations you want.  Once you
+get it all working and tested, it may become useful to find out how fast your
+pass is.  The :ref:`PassManager <writing-an-llvm-pass-passmanager>` provides a
+nice command line option (:option:`--time-passes`) that allows you to get
+information about the execution time of your pass along with the other passes
+you queue up.  For example:
+
+.. code-block:: console
+
+  $ opt -load lib/LLVMHello.so -hello -time-passes < hello.bc > /dev/null
+  Hello: __main
+  Hello: puts
+  Hello: main
+  ===-------------------------------------------------------------------------===
+                        ... Pass execution timing report ...
+  ===-------------------------------------------------------------------------===
+    Total Execution Time: 0.0007 seconds (0.0005 wall clock)
+  
+     ---User Time---   --User+System--   ---Wall Time---  --- Name ---
+     0.0004 ( 55.3%)   0.0004 ( 55.3%)   0.0004 ( 75.7%)  Bitcode Writer
+     0.0003 ( 44.7%)   0.0003 ( 44.7%)   0.0001 ( 13.6%)  Hello World Pass
+     0.0000 (  0.0%)   0.0000 (  0.0%)   0.0001 ( 10.7%)  Module Verifier
+     0.0007 (100.0%)   0.0007 (100.0%)   0.0005 (100.0%)  Total
+
+As you can see, our implementation above is pretty fast.  The additional
+passes listed are automatically inserted by the :program:`opt` tool to verify
+that the LLVM emitted by your pass is still valid and well formed LLVM, which
+hasn't been broken somehow.
+
+Now that you have seen the basics of the mechanics behind passes, we can talk
+about some more details of how they work and how to use them.
+
+.. _writing-an-llvm-pass-pass-classes:
+
+Pass classes and requirements
+=============================
+
+One of the first things that you should do when designing a new pass is to
+decide what class you should subclass for your pass.  The :ref:`Hello World
+<writing-an-llvm-pass-basiccode>` example uses the :ref:`FunctionPass
+<writing-an-llvm-pass-FunctionPass>` class for its implementation, but we did
+not discuss why or when this should occur.  Here we talk about the classes
+available, from the most general to the most specific.
+
+When choosing a superclass for your ``Pass``, you should choose the **most
+specific** class possible, while still being able to meet the requirements
+listed.  This gives the LLVM Pass Infrastructure information necessary to
+optimize how passes are run, so that the resultant compiler isn't unnecessarily
+slow.
+
+The ``ImmutablePass`` class
+---------------------------
+
+The most plain and boring type of pass is the "`ImmutablePass
+<http://llvm.org/doxygen/classllvm_1_1ImmutablePass.html>`_" class.  This pass
+type is used for passes that do not have to be run, do not change state, and
+never need to be updated.  This is not a normal type of transformation or
+analysis, but can provide information about the current compiler configuration.
+
+Although this pass class is very infrequently used, it is important for
+providing information about the current target machine being compiled for, and
+other static information that can affect the various transformations.
+
+``ImmutablePass``\ es never invalidate other transformations, are never
+invalidated, and are never "run".
+
+.. _writing-an-llvm-pass-ModulePass:
+
+The ``ModulePass`` class
+------------------------
+
+The `ModulePass <http://llvm.org/doxygen/classllvm_1_1ModulePass.html>`_ class
+is the most general of all superclasses that you can use.  Deriving from
+``ModulePass`` indicates that your pass uses the entire program as a unit,
+referring to function bodies in no predictable order, or adding and removing
+functions.  Because nothing is known about the behavior of ``ModulePass``
+subclasses, no optimization can be done for their execution.
+
+A module pass can use function level passes (e.g. dominators) using the
+``getAnalysis`` interface ``getAnalysis<DominatorTree>(llvm::Function *)`` to
+provide the function to retrieve analysis result for, if the function pass does
+not require any module or immutable passes.  Note that this can only be done
+for functions for which the analysis ran, e.g. in the case of dominators you
+should only ask for the ``DominatorTree`` for function definitions, not
+declarations.
+
+To write a correct ``ModulePass`` subclass, derive from ``ModulePass`` and
+overload the ``runOnModule`` method with the following signature:
+
+The ``runOnModule`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool runOnModule(Module &M) = 0;
+
+The ``runOnModule`` method performs the interesting work of the pass.  It
+should return ``true`` if the module was modified by the transformation and
+``false`` otherwise.
+
+.. _writing-an-llvm-pass-CallGraphSCCPass:
+
+The ``CallGraphSCCPass`` class
+------------------------------
+
+The `CallGraphSCCPass
+<http://llvm.org/doxygen/classllvm_1_1CallGraphSCCPass.html>`_ is used by
+passes that need to traverse the program bottom-up on the call graph (callees
+before callers).  Deriving from ``CallGraphSCCPass`` provides some mechanics
+for building and traversing the ``CallGraph``, but also allows the system to
+optimize execution of ``CallGraphSCCPass``\ es.  If your pass meets the
+requirements outlined below, and doesn't meet the requirements of a
+:ref:`FunctionPass <writing-an-llvm-pass-FunctionPass>` or :ref:`BasicBlockPass
+<writing-an-llvm-pass-BasicBlockPass>`, you should derive from
+``CallGraphSCCPass``.
+
+``TODO``: explain briefly what SCC, Tarjan's algo, and B-U mean.
+
+To be explicit, CallGraphSCCPass subclasses are:
+
+#. ... *not allowed* to inspect or modify any ``Function``\ s other than those
+   in the current SCC and the direct callers and direct callees of the SCC.
+#. ... *required* to preserve the current ``CallGraph`` object, updating it to
+   reflect any changes made to the program.
+#. ... *not allowed* to add or remove SCC's from the current Module, though
+   they may change the contents of an SCC.
+#. ... *allowed* to add or remove global variables from the current Module.
+#. ... *allowed* to maintain state across invocations of :ref:`runOnSCC
+   <writing-an-llvm-pass-runOnSCC>` (including global data).
+
+Implementing a ``CallGraphSCCPass`` is slightly tricky in some cases because it
+has to handle SCCs with more than one node in it.  All of the virtual methods
+described below should return ``true`` if they modified the program, or
+``false`` if they didn't.
+
+The ``doInitialization(CallGraph &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doInitialization(CallGraph &CG);
+
+The ``doInitialization`` method is allowed to do most of the things that
+``CallGraphSCCPass``\ es are not allowed to do.  They can add and remove
+functions, get pointers to functions, etc.  The ``doInitialization`` method is
+designed to do simple initialization type of stuff that does not depend on the
+SCCs being processed.  The ``doInitialization`` method call is not scheduled to
+overlap with any other pass executions (thus it should be very fast).
+
+.. _writing-an-llvm-pass-runOnSCC:
+
+The ``runOnSCC`` method
+^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool runOnSCC(CallGraphSCC &SCC) = 0;
+
+The ``runOnSCC`` method performs the interesting work of the pass, and should
+return ``true`` if the module was modified by the transformation, ``false``
+otherwise.
+
+The ``doFinalization(CallGraph &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doFinalization(CallGraph &CG);
+
+The ``doFinalization`` method is an infrequently used method that is called
+when the pass framework has finished calling :ref:`runOnSCC
+<writing-an-llvm-pass-runOnSCC>` for every SCC in the program being compiled.
+
+.. _writing-an-llvm-pass-FunctionPass:
+
+The ``FunctionPass`` class
+--------------------------
+
+In contrast to ``ModulePass`` subclasses, `FunctionPass
+<http://llvm.org/doxygen/classllvm_1_1Pass.html>`_ subclasses do have a
+predictable, local behavior that can be expected by the system.  All
+``FunctionPass`` execute on each function in the program independent of all of
+the other functions in the program.  ``FunctionPass``\ es do not require that
+they are executed in a particular order, and ``FunctionPass``\ es do not modify
+external functions.
+
+To be explicit, ``FunctionPass`` subclasses are not allowed to:
+
+#. Inspect or modify a ``Function`` other than the one currently being processed.
+#. Add or remove ``Function``\ s from the current ``Module``.
+#. Add or remove global variables from the current ``Module``.
+#. Maintain state across invocations of :ref:`runOnFunction
+   <writing-an-llvm-pass-runOnFunction>` (including global data).
+
+Implementing a ``FunctionPass`` is usually straightforward (See the :ref:`Hello
+World <writing-an-llvm-pass-basiccode>` pass for example).
+``FunctionPass``\ es may overload three virtual methods to do their work.  All
+of these methods should return ``true`` if they modified the program, or
+``false`` if they didn't.
+
+.. _writing-an-llvm-pass-doInitialization-mod:
+
+The ``doInitialization(Module &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doInitialization(Module &M);
+
+The ``doInitialization`` method is allowed to do most of the things that
+``FunctionPass``\ es are not allowed to do.  They can add and remove functions,
+get pointers to functions, etc.  The ``doInitialization`` method is designed to
+do simple initialization type of stuff that does not depend on the functions
+being processed.  The ``doInitialization`` method call is not scheduled to
+overlap with any other pass executions (thus it should be very fast).
+
+A good example of how this method should be used is the `LowerAllocations
+<http://llvm.org/doxygen/LowerAllocations_8cpp-source.html>`_ pass.  This pass
+converts ``malloc`` and ``free`` instructions into platform dependent
+``malloc()`` and ``free()`` function calls.  It uses the ``doInitialization``
+method to get a reference to the ``malloc`` and ``free`` functions that it
+needs, adding prototypes to the module if necessary.
+
+.. _writing-an-llvm-pass-runOnFunction:
+
+The ``runOnFunction`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool runOnFunction(Function &F) = 0;
+
+The ``runOnFunction`` method must be implemented by your subclass to do the
+transformation or analysis work of your pass.  As usual, a ``true`` value
+should be returned if the function is modified.
+
+.. _writing-an-llvm-pass-doFinalization-mod:
+
+The ``doFinalization(Module &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doFinalization(Module &M);
+
+The ``doFinalization`` method is an infrequently used method that is called
+when the pass framework has finished calling :ref:`runOnFunction
+<writing-an-llvm-pass-runOnFunction>` for every function in the program being
+compiled.
+
+.. _writing-an-llvm-pass-LoopPass:
+
+The ``LoopPass`` class
+----------------------
+
+All ``LoopPass`` execute on each loop in the function independent of all of the
+other loops in the function.  ``LoopPass`` processes loops in loop nest order
+such that outer most loop is processed last.
+
+``LoopPass`` subclasses are allowed to update loop nest using ``LPPassManager``
+interface.  Implementing a loop pass is usually straightforward.
+``LoopPass``\ es may overload three virtual methods to do their work.  All
+these methods should return ``true`` if they modified the program, or ``false``
+if they didn't.
+
+A ``LoopPass`` subclass which is intended to run as part of the main loop pass
+pipeline needs to preserve all of the same *function* analyses that the other
+loop passes in its pipeline require. To make that easier,
+a ``getLoopAnalysisUsage`` function is provided by ``LoopUtils.h``. It can be
+called within the subclass's ``getAnalysisUsage`` override to get consistent
+and correct behavior. Analogously, ``INITIALIZE_PASS_DEPENDENCY(LoopPass)``
+will initialize this set of function analyses.
+
+The ``doInitialization(Loop *, LPPassManager &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doInitialization(Loop *, LPPassManager &LPM);
+
+The ``doInitialization`` method is designed to do simple initialization type of
+stuff that does not depend on the functions being processed.  The
+``doInitialization`` method call is not scheduled to overlap with any other
+pass executions (thus it should be very fast).  ``LPPassManager`` interface
+should be used to access ``Function`` or ``Module`` level analysis information.
+
+.. _writing-an-llvm-pass-runOnLoop:
+
+The ``runOnLoop`` method
+^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool runOnLoop(Loop *, LPPassManager &LPM) = 0;
+
+The ``runOnLoop`` method must be implemented by your subclass to do the
+transformation or analysis work of your pass.  As usual, a ``true`` value
+should be returned if the function is modified.  ``LPPassManager`` interface
+should be used to update loop nest.
+
+The ``doFinalization()`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doFinalization();
+
+The ``doFinalization`` method is an infrequently used method that is called
+when the pass framework has finished calling :ref:`runOnLoop
+<writing-an-llvm-pass-runOnLoop>` for every loop in the program being compiled.
+
+.. _writing-an-llvm-pass-RegionPass:
+
+The ``RegionPass`` class
+------------------------
+
+``RegionPass`` is similar to :ref:`LoopPass <writing-an-llvm-pass-LoopPass>`,
+but executes on each single entry single exit region in the function.
+``RegionPass`` processes regions in nested order such that the outer most
+region is processed last.
+
+``RegionPass`` subclasses are allowed to update the region tree by using the
+``RGPassManager`` interface.  You may overload three virtual methods of
+``RegionPass`` to implement your own region pass.  All these methods should
+return ``true`` if they modified the program, or ``false`` if they did not.
+
+The ``doInitialization(Region *, RGPassManager &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doInitialization(Region *, RGPassManager &RGM);
+
+The ``doInitialization`` method is designed to do simple initialization type of
+stuff that does not depend on the functions being processed.  The
+``doInitialization`` method call is not scheduled to overlap with any other
+pass executions (thus it should be very fast).  ``RPPassManager`` interface
+should be used to access ``Function`` or ``Module`` level analysis information.
+
+.. _writing-an-llvm-pass-runOnRegion:
+
+The ``runOnRegion`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool runOnRegion(Region *, RGPassManager &RGM) = 0;
+
+The ``runOnRegion`` method must be implemented by your subclass to do the
+transformation or analysis work of your pass.  As usual, a true value should be
+returned if the region is modified.  ``RGPassManager`` interface should be used to
+update region tree.
+
+The ``doFinalization()`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doFinalization();
+
+The ``doFinalization`` method is an infrequently used method that is called
+when the pass framework has finished calling :ref:`runOnRegion
+<writing-an-llvm-pass-runOnRegion>` for every region in the program being
+compiled.
+
+.. _writing-an-llvm-pass-BasicBlockPass:
+
+The ``BasicBlockPass`` class
+----------------------------
+
+``BasicBlockPass``\ es are just like :ref:`FunctionPass's
+<writing-an-llvm-pass-FunctionPass>` , except that they must limit their scope
+of inspection and modification to a single basic block at a time.  As such,
+they are **not** allowed to do any of the following:
+
+#. Modify or inspect any basic blocks outside of the current one.
+#. Maintain state across invocations of :ref:`runOnBasicBlock
+   <writing-an-llvm-pass-runOnBasicBlock>`.
+#. Modify the control flow graph (by altering terminator instructions)
+#. Any of the things forbidden for :ref:`FunctionPasses
+   <writing-an-llvm-pass-FunctionPass>`.
+
+``BasicBlockPass``\ es are useful for traditional local and "peephole"
+optimizations.  They may override the same :ref:`doInitialization(Module &)
+<writing-an-llvm-pass-doInitialization-mod>` and :ref:`doFinalization(Module &)
+<writing-an-llvm-pass-doFinalization-mod>` methods that :ref:`FunctionPass's
+<writing-an-llvm-pass-FunctionPass>` have, but also have the following virtual
+methods that may also be implemented:
+
+The ``doInitialization(Function &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool doInitialization(Function &F);
+
+The ``doInitialization`` method is allowed to do most of the things that
+``BasicBlockPass``\ es are not allowed to do, but that ``FunctionPass``\ es
+can.  The ``doInitialization`` method is designed to do simple initialization
+that does not depend on the ``BasicBlock``\ s being processed.  The
+``doInitialization`` method call is not scheduled to overlap with any other
+pass executions (thus it should be very fast).
+
+.. _writing-an-llvm-pass-runOnBasicBlock:
+
+The ``runOnBasicBlock`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool runOnBasicBlock(BasicBlock &BB) = 0;
+
+Override this function to do the work of the ``BasicBlockPass``.  This function
+is not allowed to inspect or modify basic blocks other than the parameter, and
+are not allowed to modify the CFG.  A ``true`` value must be returned if the
+basic block is modified.
+
+The ``doFinalization(Function &)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+    virtual bool doFinalization(Function &F);
+
+The ``doFinalization`` method is an infrequently used method that is called
+when the pass framework has finished calling :ref:`runOnBasicBlock
+<writing-an-llvm-pass-runOnBasicBlock>` for every ``BasicBlock`` in the program
+being compiled.  This can be used to perform per-function finalization.
+
+The ``MachineFunctionPass`` class
+---------------------------------
+
+A ``MachineFunctionPass`` is a part of the LLVM code generator that executes on
+the machine-dependent representation of each LLVM function in the program.
+
+Code generator passes are registered and initialized specially by
+``TargetMachine::addPassesToEmitFile`` and similar routines, so they cannot
+generally be run from the :program:`opt` or :program:`bugpoint` commands.
+
+A ``MachineFunctionPass`` is also a ``FunctionPass``, so all the restrictions
+that apply to a ``FunctionPass`` also apply to it.  ``MachineFunctionPass``\ es
+also have additional restrictions.  In particular, ``MachineFunctionPass``\ es
+are not allowed to do any of the following:
+
+#. Modify or create any LLVM IR ``Instruction``\ s, ``BasicBlock``\ s,
+   ``Argument``\ s, ``Function``\ s, ``GlobalVariable``\ s,
+   ``GlobalAlias``\ es, or ``Module``\ s.
+#. Modify a ``MachineFunction`` other than the one currently being processed.
+#. Maintain state across invocations of :ref:`runOnMachineFunction
+   <writing-an-llvm-pass-runOnMachineFunction>` (including global data).
+
+.. _writing-an-llvm-pass-runOnMachineFunction:
+
+The ``runOnMachineFunction(MachineFunction &MF)`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual bool runOnMachineFunction(MachineFunction &MF) = 0;
+
+``runOnMachineFunction`` can be considered the main entry point of a
+``MachineFunctionPass``; that is, you should override this method to do the
+work of your ``MachineFunctionPass``.
+
+The ``runOnMachineFunction`` method is called on every ``MachineFunction`` in a
+``Module``, so that the ``MachineFunctionPass`` may perform optimizations on
+the machine-dependent representation of the function.  If you want to get at
+the LLVM ``Function`` for the ``MachineFunction`` you're working on, use
+``MachineFunction``'s ``getFunction()`` accessor method --- but remember, you
+may not modify the LLVM ``Function`` or its contents from a
+``MachineFunctionPass``.
+
+.. _writing-an-llvm-pass-registration:
+
+Pass registration
+-----------------
+
+In the :ref:`Hello World <writing-an-llvm-pass-basiccode>` example pass we
+illustrated how pass registration works, and discussed some of the reasons that
+it is used and what it does.  Here we discuss how and why passes are
+registered.
+
+As we saw above, passes are registered with the ``RegisterPass`` template.  The
+template parameter is the name of the pass that is to be used on the command
+line to specify that the pass should be added to a program (for example, with
+:program:`opt` or :program:`bugpoint`).  The first argument is the name of the
+pass, which is to be used for the :option:`-help` output of programs, as well
+as for debug output generated by the `--debug-pass` option.
+
+If you want your pass to be easily dumpable, you should implement the virtual
+print method:
+
+The ``print`` method
+^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual void print(llvm::raw_ostream &O, const Module *M) const;
+
+The ``print`` method must be implemented by "analyses" in order to print a
+human readable version of the analysis results.  This is useful for debugging
+an analysis itself, as well as for other people to figure out how an analysis
+works.  Use the opt ``-analyze`` argument to invoke this method.
+
+The ``llvm::raw_ostream`` parameter specifies the stream to write the results
+on, and the ``Module`` parameter gives a pointer to the top level module of the
+program that has been analyzed.  Note however that this pointer may be ``NULL``
+in certain circumstances (such as calling the ``Pass::dump()`` from a
+debugger), so it should only be used to enhance debug output, it should not be
+depended on.
+
+.. _writing-an-llvm-pass-interaction:
+
+Specifying interactions between passes
+--------------------------------------
+
+One of the main responsibilities of the ``PassManager`` is to make sure that
+passes interact with each other correctly.  Because ``PassManager`` tries to
+:ref:`optimize the execution of passes <writing-an-llvm-pass-passmanager>` it
+must know how the passes interact with each other and what dependencies exist
+between the various passes.  To track this, each pass can declare the set of
+passes that are required to be executed before the current pass, and the passes
+which are invalidated by the current pass.
+
+Typically this functionality is used to require that analysis results are
+computed before your pass is run.  Running arbitrary transformation passes can
+invalidate the computed analysis results, which is what the invalidation set
+specifies.  If a pass does not implement the :ref:`getAnalysisUsage
+<writing-an-llvm-pass-getAnalysisUsage>` method, it defaults to not having any
+prerequisite passes, and invalidating **all** other passes.
+
+.. _writing-an-llvm-pass-getAnalysisUsage:
+
+The ``getAnalysisUsage`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual void getAnalysisUsage(AnalysisUsage &Info) const;
+
+By implementing the ``getAnalysisUsage`` method, the required and invalidated
+sets may be specified for your transformation.  The implementation should fill
+in the `AnalysisUsage
+<http://llvm.org/doxygen/classllvm_1_1AnalysisUsage.html>`_ object with
+information about which passes are required and not invalidated.  To do this, a
+pass may call any of the following methods on the ``AnalysisUsage`` object:
+
+The ``AnalysisUsage::addRequired<>`` and ``AnalysisUsage::addRequiredTransitive<>`` methods
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+If your pass requires a previous pass to be executed (an analysis for example),
+it can use one of these methods to arrange for it to be run before your pass.
+LLVM has many different types of analyses and passes that can be required,
+spanning the range from ``DominatorSet`` to ``BreakCriticalEdges``.  Requiring
+``BreakCriticalEdges``, for example, guarantees that there will be no critical
+edges in the CFG when your pass has been run.
+
+Some analyses chain to other analyses to do their job.  For example, an
+`AliasAnalysis <AliasAnalysis>` implementation is required to :ref:`chain
+<aliasanalysis-chaining>` to other alias analysis passes.  In cases where
+analyses chain, the ``addRequiredTransitive`` method should be used instead of
+the ``addRequired`` method.  This informs the ``PassManager`` that the
+transitively required pass should be alive as long as the requiring pass is.
+
+The ``AnalysisUsage::addPreserved<>`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+One of the jobs of the ``PassManager`` is to optimize how and when analyses are
+run.  In particular, it attempts to avoid recomputing data unless it needs to.
+For this reason, passes are allowed to declare that they preserve (i.e., they
+don't invalidate) an existing analysis if it's available.  For example, a
+simple constant folding pass would not modify the CFG, so it can't possibly
+affect the results of dominator analysis.  By default, all passes are assumed
+to invalidate all others.
+
+The ``AnalysisUsage`` class provides several methods which are useful in
+certain circumstances that are related to ``addPreserved``.  In particular, the
+``setPreservesAll`` method can be called to indicate that the pass does not
+modify the LLVM program at all (which is true for analyses), and the
+``setPreservesCFG`` method can be used by transformations that change
+instructions in the program but do not modify the CFG or terminator
+instructions (note that this property is implicitly set for
+:ref:`BasicBlockPass <writing-an-llvm-pass-BasicBlockPass>`\ es).
+
+``addPreserved`` is particularly useful for transformations like
+``BreakCriticalEdges``.  This pass knows how to update a small set of loop and
+dominator related analyses if they exist, so it can preserve them, despite the
+fact that it hacks on the CFG.
+
+Example implementations of ``getAnalysisUsage``
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  // This example modifies the program, but does not modify the CFG
+  void LICM::getAnalysisUsage(AnalysisUsage &AU) const {
+    AU.setPreservesCFG();
+    AU.addRequired<LoopInfoWrapperPass>();
+  }
+
+.. _writing-an-llvm-pass-getAnalysis:
+
+The ``getAnalysis<>`` and ``getAnalysisIfAvailable<>`` methods
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``Pass::getAnalysis<>`` method is automatically inherited by your class,
+providing you with access to the passes that you declared that you required
+with the :ref:`getAnalysisUsage <writing-an-llvm-pass-getAnalysisUsage>`
+method.  It takes a single template argument that specifies which pass class
+you want, and returns a reference to that pass.  For example:
+
+.. code-block:: c++
+
+  bool LICM::runOnFunction(Function &F) {
+    LoopInfo &LI = getAnalysis<LoopInfoWrapperPass>().getLoopInfo();
+    //...
+  }
+
+This method call returns a reference to the pass desired.  You may get a
+runtime assertion failure if you attempt to get an analysis that you did not
+declare as required in your :ref:`getAnalysisUsage
+<writing-an-llvm-pass-getAnalysisUsage>` implementation.  This method can be
+called by your ``run*`` method implementation, or by any other local method
+invoked by your ``run*`` method.
+
+A module level pass can use function level analysis info using this interface.
+For example:
+
+.. code-block:: c++
+
+  bool ModuleLevelPass::runOnModule(Module &M) {
+    //...
+    DominatorTree &DT = getAnalysis<DominatorTree>(Func);
+    //...
+  }
+
+In above example, ``runOnFunction`` for ``DominatorTree`` is called by pass
+manager before returning a reference to the desired pass.
+
+If your pass is capable of updating analyses if they exist (e.g.,
+``BreakCriticalEdges``, as described above), you can use the
+``getAnalysisIfAvailable`` method, which returns a pointer to the analysis if
+it is active.  For example:
+
+.. code-block:: c++
+
+  if (DominatorSet *DS = getAnalysisIfAvailable<DominatorSet>()) {
+    // A DominatorSet is active.  This code will update it.
+  }
+
+Implementing Analysis Groups
+----------------------------
+
+Now that we understand the basics of how passes are defined, how they are used,
+and how they are required from other passes, it's time to get a little bit
+fancier.  All of the pass relationships that we have seen so far are very
+simple: one pass depends on one other specific pass to be run before it can
+run.  For many applications, this is great, for others, more flexibility is
+required.
+
+In particular, some analyses are defined such that there is a single simple
+interface to the analysis results, but multiple ways of calculating them.
+Consider alias analysis for example.  The most trivial alias analysis returns
+"may alias" for any alias query.  The most sophisticated analysis a
+flow-sensitive, context-sensitive interprocedural analysis that can take a
+significant amount of time to execute (and obviously, there is a lot of room
+between these two extremes for other implementations).  To cleanly support
+situations like this, the LLVM Pass Infrastructure supports the notion of
+Analysis Groups.
+
+Analysis Group Concepts
+^^^^^^^^^^^^^^^^^^^^^^^
+
+An Analysis Group is a single simple interface that may be implemented by
+multiple different passes.  Analysis Groups can be given human readable names
+just like passes, but unlike passes, they need not derive from the ``Pass``
+class.  An analysis group may have one or more implementations, one of which is
+the "default" implementation.
+
+Analysis groups are used by client passes just like other passes are: the
+``AnalysisUsage::addRequired()`` and ``Pass::getAnalysis()`` methods.  In order
+to resolve this requirement, the :ref:`PassManager
+<writing-an-llvm-pass-passmanager>` scans the available passes to see if any
+implementations of the analysis group are available.  If none is available, the
+default implementation is created for the pass to use.  All standard rules for
+:ref:`interaction between passes <writing-an-llvm-pass-interaction>` still
+apply.
+
+Although :ref:`Pass Registration <writing-an-llvm-pass-registration>` is
+optional for normal passes, all analysis group implementations must be
+registered, and must use the :ref:`INITIALIZE_AG_PASS
+<writing-an-llvm-pass-RegisterAnalysisGroup>` template to join the
+implementation pool.  Also, a default implementation of the interface **must**
+be registered with :ref:`RegisterAnalysisGroup
+<writing-an-llvm-pass-RegisterAnalysisGroup>`.
+
+As a concrete example of an Analysis Group in action, consider the
+`AliasAnalysis <http://llvm.org/doxygen/classllvm_1_1AliasAnalysis.html>`_
+analysis group.  The default implementation of the alias analysis interface
+(the `basicaa <http://llvm.org/doxygen/structBasicAliasAnalysis.html>`_ pass)
+just does a few simple checks that don't require significant analysis to
+compute (such as: two different globals can never alias each other, etc).
+Passes that use the `AliasAnalysis
+<http://llvm.org/doxygen/classllvm_1_1AliasAnalysis.html>`_ interface (for
+example the `gvn <http://llvm.org/doxygen/classllvm_1_1GVN.html>`_ pass), do not
+care which implementation of alias analysis is actually provided, they just use
+the designated interface.
+
+From the user's perspective, commands work just like normal.  Issuing the
+command ``opt -gvn ...`` will cause the ``basicaa`` class to be instantiated
+and added to the pass sequence.  Issuing the command ``opt -somefancyaa -gvn
+...`` will cause the ``gvn`` pass to use the ``somefancyaa`` alias analysis
+(which doesn't actually exist, it's just a hypothetical example) instead.
+
+.. _writing-an-llvm-pass-RegisterAnalysisGroup:
+
+Using ``RegisterAnalysisGroup``
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The ``RegisterAnalysisGroup`` template is used to register the analysis group
+itself, while the ``INITIALIZE_AG_PASS`` is used to add pass implementations to
+the analysis group.  First, an analysis group should be registered, with a
+human readable name provided for it.  Unlike registration of passes, there is
+no command line argument to be specified for the Analysis Group Interface
+itself, because it is "abstract":
+
+.. code-block:: c++
+
+  static RegisterAnalysisGroup<AliasAnalysis> A("Alias Analysis");
+
+Once the analysis is registered, passes can declare that they are valid
+implementations of the interface by using the following code:
+
+.. code-block:: c++
+
+  namespace {
+    // Declare that we implement the AliasAnalysis interface
+    INITIALIZE_AG_PASS(FancyAA, AliasAnalysis , "somefancyaa",
+        "A more complex alias analysis implementation",
+        false,  // Is CFG Only?
+        true,   // Is Analysis?
+        false); // Is default Analysis Group implementation?
+  }
+
+This just shows a class ``FancyAA`` that uses the ``INITIALIZE_AG_PASS`` macro
+both to register and to "join" the `AliasAnalysis
+<http://llvm.org/doxygen/classllvm_1_1AliasAnalysis.html>`_ analysis group.
+Every implementation of an analysis group should join using this macro.
+
+.. code-block:: c++
+
+  namespace {
+    // Declare that we implement the AliasAnalysis interface
+    INITIALIZE_AG_PASS(BasicAA, AliasAnalysis, "basicaa",
+        "Basic Alias Analysis (default AA impl)",
+        false, // Is CFG Only?
+        true,  // Is Analysis?
+        true); // Is default Analysis Group implementation?
+  }
+
+Here we show how the default implementation is specified (using the final
+argument to the ``INITIALIZE_AG_PASS`` template).  There must be exactly one
+default implementation available at all times for an Analysis Group to be used.
+Only default implementation can derive from ``ImmutablePass``.  Here we declare
+that the `BasicAliasAnalysis
+<http://llvm.org/doxygen/structBasicAliasAnalysis.html>`_ pass is the default
+implementation for the interface.
+
+Pass Statistics
+===============
+
+The `Statistic <http://llvm.org/doxygen/Statistic_8h-source.html>`_ class is
+designed to be an easy way to expose various success metrics from passes.
+These statistics are printed at the end of a run, when the :option:`-stats`
+command line option is enabled on the command line.  See the :ref:`Statistics
+section <Statistic>` in the Programmer's Manual for details.
+
+.. _writing-an-llvm-pass-passmanager:
+
+What PassManager does
+---------------------
+
+The `PassManager <http://llvm.org/doxygen/PassManager_8h-source.html>`_ `class
+<http://llvm.org/doxygen/classllvm_1_1PassManager.html>`_ takes a list of
+passes, ensures their :ref:`prerequisites <writing-an-llvm-pass-interaction>`
+are set up correctly, and then schedules passes to run efficiently.  All of the
+LLVM tools that run passes use the PassManager for execution of these passes.
+
+The PassManager does two main things to try to reduce the execution time of a
+series of passes:
+
+#. **Share analysis results.**  The ``PassManager`` attempts to avoid
+   recomputing analysis results as much as possible.  This means keeping track
+   of which analyses are available already, which analyses get invalidated, and
+   which analyses are needed to be run for a pass.  An important part of work
+   is that the ``PassManager`` tracks the exact lifetime of all analysis
+   results, allowing it to :ref:`free memory
+   <writing-an-llvm-pass-releaseMemory>` allocated to holding analysis results
+   as soon as they are no longer needed.
+
+#. **Pipeline the execution of passes on the program.**  The ``PassManager``
+   attempts to get better cache and memory usage behavior out of a series of
+   passes by pipelining the passes together.  This means that, given a series
+   of consecutive :ref:`FunctionPass <writing-an-llvm-pass-FunctionPass>`, it
+   will execute all of the :ref:`FunctionPass
+   <writing-an-llvm-pass-FunctionPass>` on the first function, then all of the
+   :ref:`FunctionPasses <writing-an-llvm-pass-FunctionPass>` on the second
+   function, etc... until the entire program has been run through the passes.
+
+   This improves the cache behavior of the compiler, because it is only
+   touching the LLVM program representation for a single function at a time,
+   instead of traversing the entire program.  It reduces the memory consumption
+   of compiler, because, for example, only one `DominatorSet
+   <http://llvm.org/doxygen/classllvm_1_1DominatorSet.html>`_ needs to be
+   calculated at a time.  This also makes it possible to implement some
+   :ref:`interesting enhancements <writing-an-llvm-pass-SMP>` in the future.
+
+The effectiveness of the ``PassManager`` is influenced directly by how much
+information it has about the behaviors of the passes it is scheduling.  For
+example, the "preserved" set is intentionally conservative in the face of an
+unimplemented :ref:`getAnalysisUsage <writing-an-llvm-pass-getAnalysisUsage>`
+method.  Not implementing when it should be implemented will have the effect of
+not allowing any analysis results to live across the execution of your pass.
+
+The ``PassManager`` class exposes a ``--debug-pass`` command line options that
+is useful for debugging pass execution, seeing how things work, and diagnosing
+when you should be preserving more analyses than you currently are.  (To get
+information about all of the variants of the ``--debug-pass`` option, just type
+"``opt -help-hidden``").
+
+By using the --debug-pass=Structure option, for example, we can see how our
+:ref:`Hello World <writing-an-llvm-pass-basiccode>` pass interacts with other
+passes.  Lets try it out with the gvn and licm passes:
+
+.. code-block:: console
+
+  $ opt -load lib/LLVMHello.so -gvn -licm --debug-pass=Structure < hello.bc > /dev/null
+  ModulePass Manager
+    FunctionPass Manager
+      Dominator Tree Construction
+      Basic Alias Analysis (stateless AA impl)
+      Function Alias Analysis Results
+      Memory Dependence Analysis
+      Global Value Numbering
+      Natural Loop Information
+      Canonicalize natural loops
+      Loop-Closed SSA Form Pass
+      Basic Alias Analysis (stateless AA impl)
+      Function Alias Analysis Results
+      Scalar Evolution Analysis
+      Loop Pass Manager
+        Loop Invariant Code Motion
+      Module Verifier
+    Bitcode Writer
+
+This output shows us when passes are constructed.
+Here we see that GVN uses dominator tree information to do its job.  The LICM pass
+uses natural loop information, which uses dominator tree as well.
+
+After the LICM pass, the module verifier runs (which is automatically added by
+the :program:`opt` tool), which uses the dominator tree to check that the
+resultant LLVM code is well formed. Note that the dominator tree is computed
+once, and shared by three passes.
+
+Lets see how this changes when we run the :ref:`Hello World
+<writing-an-llvm-pass-basiccode>` pass in between the two passes:
+
+.. code-block:: console
+
+  $ opt -load lib/LLVMHello.so -gvn -hello -licm --debug-pass=Structure < hello.bc > /dev/null
+  ModulePass Manager
+    FunctionPass Manager
+      Dominator Tree Construction
+      Basic Alias Analysis (stateless AA impl)
+      Function Alias Analysis Results
+      Memory Dependence Analysis
+      Global Value Numbering
+      Hello World Pass
+      Dominator Tree Construction
+      Natural Loop Information
+      Canonicalize natural loops
+      Loop-Closed SSA Form Pass
+      Basic Alias Analysis (stateless AA impl)
+      Function Alias Analysis Results
+      Scalar Evolution Analysis
+      Loop Pass Manager
+        Loop Invariant Code Motion
+      Module Verifier
+    Bitcode Writer
+  Hello: __main
+  Hello: puts
+  Hello: main
+
+Here we see that the :ref:`Hello World <writing-an-llvm-pass-basiccode>` pass
+has killed the Dominator Tree pass, even though it doesn't modify the code at
+all!  To fix this, we need to add the following :ref:`getAnalysisUsage
+<writing-an-llvm-pass-getAnalysisUsage>` method to our pass:
+
+.. code-block:: c++
+
+  // We don't modify the program, so we preserve all analyses
+  void getAnalysisUsage(AnalysisUsage &AU) const override {
+    AU.setPreservesAll();
+  }
+
+Now when we run our pass, we get this output:
+
+.. code-block:: console
+
+  $ opt -load lib/LLVMHello.so -gvn -hello -licm --debug-pass=Structure < hello.bc > /dev/null
+  Pass Arguments:  -gvn -hello -licm
+  ModulePass Manager
+    FunctionPass Manager
+      Dominator Tree Construction
+      Basic Alias Analysis (stateless AA impl)
+      Function Alias Analysis Results
+      Memory Dependence Analysis
+      Global Value Numbering
+      Hello World Pass
+      Natural Loop Information
+      Canonicalize natural loops
+      Loop-Closed SSA Form Pass
+      Basic Alias Analysis (stateless AA impl)
+      Function Alias Analysis Results
+      Scalar Evolution Analysis
+      Loop Pass Manager
+        Loop Invariant Code Motion
+      Module Verifier
+    Bitcode Writer
+  Hello: __main
+  Hello: puts
+  Hello: main
+
+Which shows that we don't accidentally invalidate dominator information
+anymore, and therefore do not have to compute it twice.
+
+.. _writing-an-llvm-pass-releaseMemory:
+
+The ``releaseMemory`` method
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: c++
+
+  virtual void releaseMemory();
+
+The ``PassManager`` automatically determines when to compute analysis results,
+and how long to keep them around for.  Because the lifetime of the pass object
+itself is effectively the entire duration of the compilation process, we need
+some way to free analysis results when they are no longer useful.  The
+``releaseMemory`` virtual method is the way to do this.
+
+If you are writing an analysis or any other pass that retains a significant
+amount of state (for use by another pass which "requires" your pass and uses
+the :ref:`getAnalysis <writing-an-llvm-pass-getAnalysis>` method) you should
+implement ``releaseMemory`` to, well, release the memory allocated to maintain
+this internal state.  This method is called after the ``run*`` method for the
+class, before the next call of ``run*`` in your pass.
+
+Registering dynamically loaded passes
+=====================================
+
+*Size matters* when constructing production quality tools using LLVM, both for
+the purposes of distribution, and for regulating the resident code size when
+running on the target system.  Therefore, it becomes desirable to selectively
+use some passes, while omitting others and maintain the flexibility to change
+configurations later on.  You want to be able to do all this, and, provide
+feedback to the user.  This is where pass registration comes into play.
+
+The fundamental mechanisms for pass registration are the
+``MachinePassRegistry`` class and subclasses of ``MachinePassRegistryNode``.
+
+An instance of ``MachinePassRegistry`` is used to maintain a list of
+``MachinePassRegistryNode`` objects.  This instance maintains the list and
+communicates additions and deletions to the command line interface.
+
+An instance of ``MachinePassRegistryNode`` subclass is used to maintain
+information provided about a particular pass.  This information includes the
+command line name, the command help string and the address of the function used
+to create an instance of the pass.  A global static constructor of one of these
+instances *registers* with a corresponding ``MachinePassRegistry``, the static
+destructor *unregisters*.  Thus a pass that is statically linked in the tool
+will be registered at start up.  A dynamically loaded pass will register on
+load and unregister at unload.
+
+Using existing registries
+-------------------------
+
+There are predefined registries to track instruction scheduling
+(``RegisterScheduler``) and register allocation (``RegisterRegAlloc``) machine
+passes.  Here we will describe how to *register* a register allocator machine
+pass.
+
+Implement your register allocator machine pass.  In your register allocator
+``.cpp`` file add the following include:
+
+.. code-block:: c++
+
+  #include "llvm/CodeGen/RegAllocRegistry.h"
+
+Also in your register allocator ``.cpp`` file, define a creator function in the
+form:
+
+.. code-block:: c++
+
+  FunctionPass *createMyRegisterAllocator() {
+    return new MyRegisterAllocator();
+  }
+
+Note that the signature of this function should match the type of
+``RegisterRegAlloc::FunctionPassCtor``.  In the same file add the "installing"
+declaration, in the form:
+
+.. code-block:: c++
+
+  static RegisterRegAlloc myRegAlloc("myregalloc",
+                                     "my register allocator help string",
+                                     createMyRegisterAllocator);
+
+Note the two spaces prior to the help string produces a tidy result on the
+:option:`-help` query.
+
+.. code-block:: console
+
+  $ llc -help
+    ...
+    -regalloc                    - Register allocator to use (default=linearscan)
+      =linearscan                -   linear scan register allocator
+      =local                     -   local register allocator
+      =simple                    -   simple register allocator
+      =myregalloc                -   my register allocator help string
+    ...
+
+And that's it.  The user is now free to use ``-regalloc=myregalloc`` as an
+option.  Registering instruction schedulers is similar except use the
+``RegisterScheduler`` class.  Note that the
+``RegisterScheduler::FunctionPassCtor`` is significantly different from
+``RegisterRegAlloc::FunctionPassCtor``.
+
+To force the load/linking of your register allocator into the
+:program:`llc`/:program:`lli` tools, add your creator function's global
+declaration to ``Passes.h`` and add a "pseudo" call line to
+``llvm/Codegen/LinkAllCodegenComponents.h``.
+
+Creating new registries
+-----------------------
+
+The easiest way to get started is to clone one of the existing registries; we
+recommend ``llvm/CodeGen/RegAllocRegistry.h``.  The key things to modify are
+the class name and the ``FunctionPassCtor`` type.
+
+Then you need to declare the registry.  Example: if your pass registry is
+``RegisterMyPasses`` then define:
+
+.. code-block:: c++
+
+  MachinePassRegistry RegisterMyPasses::Registry;
+
+And finally, declare the command line option for your passes.  Example:
+
+.. code-block:: c++
+
+  cl::opt<RegisterMyPasses::FunctionPassCtor, false,
+          RegisterPassParser<RegisterMyPasses> >
+  MyPassOpt("mypass",
+            cl::init(&createDefaultMyPass),
+            cl::desc("my pass option help"));
+
+Here the command option is "``mypass``", with ``createDefaultMyPass`` as the
+default creator.
+
+Using GDB with dynamically loaded passes
+----------------------------------------
+
+Unfortunately, using GDB with dynamically loaded passes is not as easy as it
+should be.  First of all, you can't set a breakpoint in a shared object that
+has not been loaded yet, and second of all there are problems with inlined
+functions in shared objects.  Here are some suggestions to debugging your pass
+with GDB.
+
+For sake of discussion, I'm going to assume that you are debugging a
+transformation invoked by :program:`opt`, although nothing described here
+depends on that.
+
+Setting a breakpoint in your pass
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+First thing you do is start gdb on the opt process:
+
+.. code-block:: console
+
+  $ gdb opt
+  GNU gdb 5.0
+  Copyright 2000 Free Software Foundation, Inc.
+  GDB is free software, covered by the GNU General Public License, and you are
+  welcome to change it and/or distribute copies of it under certain conditions.
+  Type "show copying" to see the conditions.
+  There is absolutely no warranty for GDB.  Type "show warranty" for details.
+  This GDB was configured as "sparc-sun-solaris2.6"...
+  (gdb)
+
+Note that :program:`opt` has a lot of debugging information in it, so it takes
+time to load.  Be patient.  Since we cannot set a breakpoint in our pass yet
+(the shared object isn't loaded until runtime), we must execute the process,
+and have it stop before it invokes our pass, but after it has loaded the shared
+object.  The most foolproof way of doing this is to set a breakpoint in
+``PassManager::run`` and then run the process with the arguments you want:
+
+.. code-block:: console
+
+  $ (gdb) break llvm::PassManager::run
+  Breakpoint 1 at 0x2413bc: file Pass.cpp, line 70.
+  (gdb) run test.bc -load $(LLVMTOP)/llvm/Debug+Asserts/lib/[libname].so -[passoption]
+  Starting program: opt test.bc -load $(LLVMTOP)/llvm/Debug+Asserts/lib/[libname].so -[passoption]
+  Breakpoint 1, PassManager::run (this=0xffbef174, M=@0x70b298) at Pass.cpp:70
+  70      bool PassManager::run(Module &M) { return PM->run(M); }
+  (gdb)
+
+Once the :program:`opt` stops in the ``PassManager::run`` method you are now
+free to set breakpoints in your pass so that you can trace through execution or
+do other standard debugging stuff.
+
+Miscellaneous Problems
+^^^^^^^^^^^^^^^^^^^^^^
+
+Once you have the basics down, there are a couple of problems that GDB has,
+some with solutions, some without.
+
+* Inline functions have bogus stack information.  In general, GDB does a pretty
+  good job getting stack traces and stepping through inline functions.  When a
+  pass is dynamically loaded however, it somehow completely loses this
+  capability.  The only solution I know of is to de-inline a function (move it
+  from the body of a class to a ``.cpp`` file).
+
+* Restarting the program breaks breakpoints.  After following the information
+  above, you have succeeded in getting some breakpoints planted in your pass.
+  Next thing you know, you restart the program (i.e., you type "``run``" again),
+  and you start getting errors about breakpoints being unsettable.  The only
+  way I have found to "fix" this problem is to delete the breakpoints that are
+  already set in your pass, run the program, and re-set the breakpoints once
+  execution stops in ``PassManager::run``.
+
+Hopefully these tips will help with common case debugging situations.  If you'd
+like to contribute some tips of your own, just contact `Chris
+<mailto:sabre at nondot.org>`_.
+
+Future extensions planned
+-------------------------
+
+Although the LLVM Pass Infrastructure is very capable as it stands, and does
+some nifty stuff, there are things we'd like to add in the future.  Here is
+where we are going:
+
+.. _writing-an-llvm-pass-SMP:
+
+Multithreaded LLVM
+^^^^^^^^^^^^^^^^^^
+
+Multiple CPU machines are becoming more common and compilation can never be
+fast enough: obviously we should allow for a multithreaded compiler.  Because
+of the semantics defined for passes above (specifically they cannot maintain
+state across invocations of their ``run*`` methods), a nice clean way to
+implement a multithreaded compiler would be for the ``PassManager`` class to
+create multiple instances of each pass object, and allow the separate instances
+to be hacking on different parts of the program at the same time.
+
+This implementation would prevent each of the passes from having to implement
+multithreaded constructs, requiring only the LLVM core to have locking in a few
+places (for global resources).  Although this is a simple extension, we simply
+haven't had time (or multiprocessor machines, thus a reason) to implement this.
+Despite that, we have kept the LLVM passes SMP ready, and you should too.
+

Added: www-releases/trunk/4.0.0/docs/_sources/XRay.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/XRay.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/XRay.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/XRay.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,232 @@
+====================
+XRay Instrumentation
+====================
+
+:Version: 1 as of 2016-11-08
+
+.. contents::
+   :local:
+
+
+Introduction
+============
+
+XRay is a function call tracing system which combines compiler-inserted
+instrumentation points and a runtime library that can dynamically enable and
+disable the instrumentation.
+
+More high level information about XRay can be found in the `XRay whitepaper`_.
+
+This document describes how to use XRay as implemented in LLVM.
+
+XRay in LLVM
+============
+
+XRay consists of three main parts:
+
+- Compiler-inserted instrumentation points.
+- A runtime library for enabling/disabling tracing at runtime.
+- A suite of tools for analysing the traces.
+
+  **NOTE:** As of the time of this writing, XRay is only available for x86_64
+  and arm7 32-bit (no-thumb) Linux.
+
+The compiler-inserted instrumentation points come in the form of nop-sleds in
+the final generated binary, and an ELF section named ``xray_instr_map`` which
+contains entries pointing to these instrumentation points. The runtime library
+relies on being able to access the entries of the ``xray_instr_map``, and
+overwrite the instrumentation points at runtime.
+
+Using XRay
+==========
+
+You can use XRay in a couple of ways:
+
+- Instrumenting your C/C++/Objective-C/Objective-C++ application.
+- Generating LLVM IR with the correct function attributes.
+
+The rest of this section covers these main ways and later on how to customise
+what XRay does in an XRay-instrumented binary.
+
+Instrumenting your C/C++/Objective-C Application
+------------------------------------------------
+
+The easiest way of getting XRay instrumentation for your application is by
+enabling the ``-fxray-instrument`` flag in your clang invocation.
+
+For example:
+
+::
+
+  clang -fxray-instrument ..
+
+By default, functions that have at least 200 instructions will get XRay
+instrumentation points. You can tweak that number through the
+``-fxray-instruction-threshold=`` flag:
+
+::
+
+  clang -fxray-instrument -fxray-instruction-threshold=1 ..
+
+You can also specifically instrument functions in your binary to either always
+or never be instrumented using source-level attributes. You can do it using the
+GCC-style attributes or C++11-style attributes.
+
+.. code-block:: c++
+
+    [[clang::xray_always_intrument]] void always_instrumented();
+
+    [[clang::xray_never_instrument]] void never_instrumented();
+
+    void alt_always_instrumented() __attribute__((xray_always_intrument));
+
+    void alt_never_instrumented() __attribute__((xray_never_instrument));
+
+When linking a binary, you can either manually link in the `XRay Runtime
+Library`_ or use ``clang`` to link it in automatically with the
+``-fxray-instrument`` flag.
+
+LLVM Function Attribute
+-----------------------
+
+If you're using LLVM IR directly, you can add the ``function-instrument``
+string attribute to your functions, to get the similar effect that the
+C/C++/Objective-C source-level attributes would get:
+
+.. code-block:: llvm
+
+    define i32 @always_instrument() uwtable "function-instrument"="xray-always" {
+      ; ...
+    }
+
+    define i32 @never_instrument() uwtable "function-instrument"="xray-never" {
+      ; ...
+    }
+
+You can also set the ``xray-instruction-threshold`` attribute and provide a
+numeric string value for how many instructions should be in the function before
+it gets instrumented.
+
+.. code-block:: llvm
+
+    define i32 @maybe_instrument() uwtable "xray-instruction-threshold"="2" {
+      ; ...
+    }
+
+XRay Runtime Library
+--------------------
+
+The XRay Runtime Library is part of the compiler-rt project, which implements
+the runtime components that perform the patching and unpatching of inserted
+instrumentation points. When you use ``clang`` to link your binaries and the
+``-fxray-instrument`` flag, it will automatically link in the XRay runtime.
+
+The default implementation of the XRay runtime will enable XRay instrumentation
+before ``main`` starts, which works for applications that have a short
+lifetime. This implementation also records all function entry and exit events
+which may result in a lot of records in the resulting trace.
+
+Also by default the filename of the XRay trace is ``xray-log.XXXXXX`` where the
+``XXXXXX`` part is randomly generated.
+
+These options can be controlled through the ``XRAY_OPTIONS`` environment
+variable, where we list down the options and their defaults below.
+
++-------------------+-----------------+---------------+------------------------+
+| Option            | Type            | Default       | Description            |
++===================+=================+===============+========================+
+| patch_premain     | ``bool``        | ``true``      | Whether to patch       |
+|                   |                 |               | instrumentation points |
+|                   |                 |               | before main.           |
++-------------------+-----------------+---------------+------------------------+
+| xray_naive_log    | ``bool``        | ``true``      | Whether to install     |
+|                   |                 |               | the naive log          |
+|                   |                 |               | implementation.        |
++-------------------+-----------------+---------------+------------------------+
+| xray_logfile_base | ``const char*`` | ``xray-log.`` | Filename base for the  |
+|                   |                 |               | XRay logfile.          |
++-------------------+-----------------+---------------+------------------------+
+
+If you choose to not use the default logging implementation that comes with the
+XRay runtime and/or control when/how the XRay instrumentation runs, you may use
+the XRay APIs directly for doing so. To do this, you'll need to include the
+``xray_interface.h`` from the compiler-rt ``xray`` directory. The important API
+functions we list below:
+
+- ``__xray_set_handler(void (*entry)(int32_t, XRayEntryType))``: Install your
+  own logging handler for when an event is encountered. See
+  ``xray/xray_interface.h`` for more details.
+- ``__xray_remove_handler()``: Removes whatever the installed handler is.
+- ``__xray_patch()``: Patch all the instrumentation points defined in the
+  binary.
+- ``__xray_unpatch()``: Unpatch the instrumentation points defined in the
+  binary.
+
+There are some requirements on the logging handler to be installed for the
+thread-safety of operations to be performed by the XRay runtime library:
+
+- The function should be thread-safe, as multiple threads may be invoking the
+  function at the same time. If the logging function needs to do
+  synchronisation, it must do so internally as XRay does not provide any
+  synchronisation guarantees outside from the atomicity of updates to the
+  pointer.
+- The pointer provided to ``__xray_set_handler(...)`` must be live even after
+  calls to ``__xray_remove_handler()`` and ``__xray_unpatch()`` have succeeded.
+  XRay cannot guarantee that all threads that have ever gotten a copy of the
+  pointer will not invoke the function.
+
+
+Trace Analysis Tools
+--------------------
+
+We currently have the beginnings of a trace analysis tool in LLVM, which can be
+found in the ``tools/llvm-xray`` directory. The ``llvm-xray`` tool currently
+supports the following subcommands:
+
+- ``extract``: Extract the instrumentation map from a binary, and return it as
+  YAML.
+
+
+Future Work
+===========
+
+There are a number of ongoing efforts for expanding the toolset building around
+the XRay instrumentation system.
+
+Flight Data Recorder Mode
+-------------------------
+
+The `XRay whitepaper`_ mentions a mode for when events are kept in memory, and
+have the traces be dumped on demand through a triggering API. This work is
+currently ongoing.
+
+Trace Analysis
+--------------
+
+There are a few more subcommands making its way to the ``llvm-xray`` tool, that
+are currently under review:
+
+- ``convert``: Turns an XRay trace from one format to another. Currently
+  supporting conversion from the binary XRay log to YAML.
+- ``account``: Do function call accounting based on data in the XRay log.
+
+We have more subcommands and modes that we're thinking of developing, in the
+following forms:
+
+- ``stack``: Reconstruct the function call stacks in a timeline.
+- ``convert``: Converting from one version of the XRay log to another (higher)
+  version, and converting to other trace formats (i.e. Chrome Trace Viewer,
+  pprof, etc.).
+- ``graph``: Generate a function call graph with relative timings and distributions.
+
+More Platforms
+--------------
+
+Since XRay is only currently available in x86_64 and arm7 32-bit (no-thumb)
+running Linux, we're looking to supporting more platforms (architectures and
+operating systems).
+
+.. References...
+
+.. _`XRay whitepaper`: http://research.google.com/pubs/pub45287.html
+

Added: www-releases/trunk/4.0.0/docs/_sources/YamlIO.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/YamlIO.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/YamlIO.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/YamlIO.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,1035 @@
+=====================
+YAML I/O
+=====================
+
+.. contents::
+   :local:
+
+Introduction to YAML
+====================
+
+YAML is a human readable data serialization language.  The full YAML language 
+spec can be read at `yaml.org 
+<http://www.yaml.org/spec/1.2/spec.html#Introduction>`_.  The simplest form of
+yaml is just "scalars", "mappings", and "sequences".  A scalar is any number
+or string.  The pound/hash symbol (#) begins a comment line.   A mapping is 
+a set of key-value pairs where the key ends with a colon.  For example:
+
+.. code-block:: yaml
+
+     # a mapping
+     name:      Tom
+     hat-size:  7
+     
+A sequence is a list of items where each item starts with a leading dash ('-'). 
+For example:
+
+.. code-block:: yaml
+
+     # a sequence
+     - x86
+     - x86_64
+     - PowerPC
+
+You can combine mappings and sequences by indenting.  For example a sequence
+of mappings in which one of the mapping values is itself a sequence:
+
+.. code-block:: yaml
+
+     # a sequence of mappings with one key's value being a sequence
+     - name:      Tom
+       cpus:
+        - x86
+        - x86_64
+     - name:      Bob
+       cpus:
+        - x86
+     - name:      Dan
+       cpus:
+        - PowerPC
+        - x86
+
+Sometime sequences are known to be short and the one entry per line is too
+verbose, so YAML offers an alternate syntax for sequences called a "Flow
+Sequence" in which you put comma separated sequence elements into square 
+brackets.  The above example could then be simplified to :
+
+
+.. code-block:: yaml
+
+     # a sequence of mappings with one key's value being a flow sequence
+     - name:      Tom
+       cpus:      [ x86, x86_64 ]
+     - name:      Bob
+       cpus:      [ x86 ]
+     - name:      Dan
+       cpus:      [ PowerPC, x86 ]
+
+
+Introduction to YAML I/O
+========================
+
+The use of indenting makes the YAML easy for a human to read and understand,
+but having a program read and write YAML involves a lot of tedious details.
+The YAML I/O library structures and simplifies reading and writing YAML 
+documents.
+
+YAML I/O assumes you have some "native" data structures which you want to be
+able to dump as YAML and recreate from YAML.  The first step is to try 
+writing example YAML for your data structures. You may find after looking at 
+possible YAML representations that a direct mapping of your data structures
+to YAML is not very readable.  Often the fields are not in the order that
+a human would find readable.  Or the same information is replicated in multiple
+locations, making it hard for a human to write such YAML correctly.  
+
+In relational database theory there is a design step called normalization in 
+which you reorganize fields and tables.  The same considerations need to 
+go into the design of your YAML encoding.  But, you may not want to change
+your existing native data structures.  Therefore, when writing out YAML
+there may be a normalization step, and when reading YAML there would be a
+corresponding denormalization step.  
+
+YAML I/O uses a non-invasive, traits based design.  YAML I/O defines some 
+abstract base templates.  You specialize those templates on your data types.
+For instance, if you have an enumerated type FooBar you could specialize 
+ScalarEnumerationTraits on that type and define the enumeration() method:
+
+.. code-block:: c++
+
+    using llvm::yaml::ScalarEnumerationTraits;
+    using llvm::yaml::IO;
+
+    template <>
+    struct ScalarEnumerationTraits<FooBar> {
+      static void enumeration(IO &io, FooBar &value) {
+      ...
+      }
+    };
+
+
+As with all YAML I/O template specializations, the ScalarEnumerationTraits is used for 
+both reading and writing YAML. That is, the mapping between in-memory enum
+values and the YAML string representation is only in one place.
+This assures that the code for writing and parsing of YAML stays in sync.
+
+To specify a YAML mappings, you define a specialization on 
+llvm::yaml::MappingTraits.
+If your native data structure happens to be a struct that is already normalized,
+then the specialization is simple.  For example:
+
+.. code-block:: c++
+   
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+    
+    template <>
+    struct MappingTraits<Person> {
+      static void mapping(IO &io, Person &info) {
+        io.mapRequired("name",         info.name);
+        io.mapOptional("hat-size",     info.hatSize);
+      }
+    };
+
+
+A YAML sequence is automatically inferred if you data type has begin()/end()
+iterators and a push_back() method.  Therefore any of the STL containers
+(such as std::vector<>) will automatically translate to YAML sequences.
+
+Once you have defined specializations for your data types, you can 
+programmatically use YAML I/O to write a YAML document:
+
+.. code-block:: c++
+   
+    using llvm::yaml::Output;
+
+    Person tom;
+    tom.name = "Tom";
+    tom.hatSize = 8;
+    Person dan;
+    dan.name = "Dan";
+    dan.hatSize = 7;
+    std::vector<Person> persons;
+    persons.push_back(tom);
+    persons.push_back(dan);
+    
+    Output yout(llvm::outs());
+    yout << persons;
+   
+This would write the following:
+
+.. code-block:: yaml
+
+     - name:      Tom
+       hat-size:  8
+     - name:      Dan
+       hat-size:  7
+
+And you can also read such YAML documents with the following code:
+
+.. code-block:: c++
+
+    using llvm::yaml::Input;
+
+    typedef std::vector<Person> PersonList;
+    std::vector<PersonList> docs;
+    
+    Input yin(document.getBuffer());
+    yin >> docs;
+    
+    if ( yin.error() )
+      return;
+    
+    // Process read document
+    for ( PersonList &pl : docs ) {
+      for ( Person &person : pl ) {
+        cout << "name=" << person.name;
+      }
+    }
+  
+One other feature of YAML is the ability to define multiple documents in a 
+single file.  That is why reading YAML produces a vector of your document type.
+
+
+
+Error Handling
+==============
+
+When parsing a YAML document, if the input does not match your schema (as 
+expressed in your XxxTraits<> specializations).  YAML I/O 
+will print out an error message and your Input object's error() method will 
+return true. For instance the following document:
+
+.. code-block:: yaml
+
+     - name:      Tom
+       shoe-size: 12
+     - name:      Dan
+       hat-size:  7
+
+Has a key (shoe-size) that is not defined in the schema.  YAML I/O will 
+automatically generate this error:
+
+.. code-block:: yaml
+
+    YAML:2:2: error: unknown key 'shoe-size'
+      shoe-size:       12
+      ^~~~~~~~~
+
+Similar errors are produced for other input not conforming to the schema.
+
+
+Scalars
+=======
+
+YAML scalars are just strings (i.e. not a sequence or mapping).  The YAML I/O
+library provides support for translating between YAML scalars and specific
+C++ types.
+
+
+Built-in types
+--------------
+The following types have built-in support in YAML I/O:
+
+* bool
+* float
+* double
+* StringRef
+* std::string
+* int64_t
+* int32_t
+* int16_t
+* int8_t
+* uint64_t
+* uint32_t
+* uint16_t
+* uint8_t
+
+That is, you can use those types in fields of MappingTraits or as element type
+in sequence.  When reading, YAML I/O will validate that the string found
+is convertible to that type and error out if not.
+
+
+Unique types
+------------
+Given that YAML I/O is trait based, the selection of how to convert your data
+to YAML is based on the type of your data.  But in C++ type matching, typedefs
+do not generate unique type names.  That means if you have two typedefs of
+unsigned int, to YAML I/O both types look exactly like unsigned int.  To
+facilitate make unique type names, YAML I/O provides a macro which is used
+like a typedef on built-in types, but expands to create a class with conversion
+operators to and from the base type.  For example:
+
+.. code-block:: c++
+
+    LLVM_YAML_STRONG_TYPEDEF(uint32_t, MyFooFlags)
+    LLVM_YAML_STRONG_TYPEDEF(uint32_t, MyBarFlags)
+
+This generates two classes MyFooFlags and MyBarFlags which you can use in your
+native data structures instead of uint32_t. They are implicitly 
+converted to and from uint32_t.  The point of creating these unique types
+is that you can now specify traits on them to get different YAML conversions.
+
+Hex types
+---------
+An example use of a unique type is that YAML I/O provides fixed sized unsigned
+integers that are written with YAML I/O as hexadecimal instead of the decimal
+format used by the built-in integer types:
+
+* Hex64
+* Hex32
+* Hex16
+* Hex8
+
+You can use llvm::yaml::Hex32 instead of uint32_t and the only different will
+be that when YAML I/O writes out that type it will be formatted in hexadecimal.
+
+
+ScalarEnumerationTraits
+-----------------------
+YAML I/O supports translating between in-memory enumerations and a set of string
+values in YAML documents. This is done by specializing ScalarEnumerationTraits<>
+on your enumeration type and define a enumeration() method. 
+For instance, suppose you had an enumeration of CPUs and a struct with it as 
+a field:
+
+.. code-block:: c++
+
+    enum CPUs {
+      cpu_x86_64  = 5,
+      cpu_x86     = 7,
+      cpu_PowerPC = 8
+    };
+    
+    struct Info {
+      CPUs      cpu;
+      uint32_t  flags;
+    };
+    
+To support reading and writing of this enumeration, you can define a 
+ScalarEnumerationTraits specialization on CPUs, which can then be used 
+as a field type: 
+
+.. code-block:: c++
+
+    using llvm::yaml::ScalarEnumerationTraits;
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+
+    template <>
+    struct ScalarEnumerationTraits<CPUs> {
+      static void enumeration(IO &io, CPUs &value) {
+        io.enumCase(value, "x86_64",  cpu_x86_64);
+        io.enumCase(value, "x86",     cpu_x86);
+        io.enumCase(value, "PowerPC", cpu_PowerPC);
+      }
+    };
+ 
+    template <>
+    struct MappingTraits<Info> {
+      static void mapping(IO &io, Info &info) {
+        io.mapRequired("cpu",       info.cpu);
+        io.mapOptional("flags",     info.flags, 0);
+      }
+    };
+
+When reading YAML, if the string found does not match any of the strings
+specified by enumCase() methods, an error is automatically generated.
+When writing YAML, if the value being written does not match any of the values
+specified by the enumCase() methods, a runtime assertion is triggered.
+  
+
+BitValue
+--------
+Another common data structure in C++ is a field where each bit has a unique
+meaning.  This is often used in a "flags" field.  YAML I/O has support for
+converting such fields to a flow sequence.   For instance suppose you 
+had the following bit flags defined:
+
+.. code-block:: c++
+
+    enum {
+      flagsPointy = 1
+      flagsHollow = 2
+      flagsFlat   = 4
+      flagsRound  = 8
+    };
+
+    LLVM_YAML_STRONG_TYPEDEF(uint32_t, MyFlags)
+    
+To support reading and writing of MyFlags, you specialize ScalarBitSetTraits<>
+on MyFlags and provide the bit values and their names.   
+
+.. code-block:: c++
+
+    using llvm::yaml::ScalarBitSetTraits;
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+
+    template <>
+    struct ScalarBitSetTraits<MyFlags> {
+      static void bitset(IO &io, MyFlags &value) {
+        io.bitSetCase(value, "hollow",  flagHollow);
+        io.bitSetCase(value, "flat",    flagFlat);
+        io.bitSetCase(value, "round",   flagRound);
+        io.bitSetCase(value, "pointy",  flagPointy);
+      }
+    };
+    
+    struct Info {
+      StringRef   name;
+      MyFlags     flags;
+    };
+    
+    template <>
+    struct MappingTraits<Info> {
+      static void mapping(IO &io, Info& info) {
+        io.mapRequired("name",  info.name);
+        io.mapRequired("flags", info.flags);
+       }
+    };
+
+With the above, YAML I/O (when writing) will test mask each value in the 
+bitset trait against the flags field, and each that matches will
+cause the corresponding string to be added to the flow sequence.  The opposite
+is done when reading and any unknown string values will result in a error. With 
+the above schema, a same valid YAML document is:
+
+.. code-block:: yaml
+
+    name:    Tom
+    flags:   [ pointy, flat ]
+
+Sometimes a "flags" field might contains an enumeration part
+defined by a bit-mask.
+
+.. code-block:: c++
+
+    enum {
+      flagsFeatureA = 1,
+      flagsFeatureB = 2,
+      flagsFeatureC = 4,
+
+      flagsCPUMask = 24,
+
+      flagsCPU1 = 8,
+      flagsCPU2 = 16
+    };
+
+To support reading and writing such fields, you need to use the maskedBitSet()
+method and provide the bit values, their names and the enumeration mask.
+
+.. code-block:: c++
+
+    template <>
+    struct ScalarBitSetTraits<MyFlags> {
+      static void bitset(IO &io, MyFlags &value) {
+        io.bitSetCase(value, "featureA",  flagsFeatureA);
+        io.bitSetCase(value, "featureB",  flagsFeatureB);
+        io.bitSetCase(value, "featureC",  flagsFeatureC);
+        io.maskedBitSetCase(value, "CPU1",  flagsCPU1, flagsCPUMask);
+        io.maskedBitSetCase(value, "CPU2",  flagsCPU2, flagsCPUMask);
+      }
+    };
+
+YAML I/O (when writing) will apply the enumeration mask to the flags field,
+and compare the result and values from the bitset. As in case of a regular
+bitset, each that matches will cause the corresponding string to be added
+to the flow sequence.
+
+Custom Scalar
+-------------
+Sometimes for readability a scalar needs to be formatted in a custom way. For
+instance your internal data structure may use a integer for time (seconds since
+some epoch), but in YAML it would be much nicer to express that integer in 
+some time format (e.g. 4-May-2012 10:30pm).  YAML I/O has a way to support  
+custom formatting and parsing of scalar types by specializing ScalarTraits<> on
+your data type.  When writing, YAML I/O will provide the native type and
+your specialization must create a temporary llvm::StringRef.  When reading,
+YAML I/O will provide an llvm::StringRef of scalar and your specialization
+must convert that to your native data type.  An outline of a custom scalar type
+looks like:
+
+.. code-block:: c++
+
+    using llvm::yaml::ScalarTraits;
+    using llvm::yaml::IO;
+
+    template <>
+    struct ScalarTraits<MyCustomType> {
+      static void output(const MyCustomType &value, void*,
+                         llvm::raw_ostream &out) {
+        out << value;  // do custom formatting here
+      }
+      static StringRef input(StringRef scalar, void*, MyCustomType &value) {
+        // do custom parsing here.  Return the empty string on success,
+        // or an error message on failure.
+        return StringRef();
+      }
+      // Determine if this scalar needs quotes.
+      static bool mustQuote(StringRef) { return true; }
+    };
+
+Block Scalars
+-------------
+
+YAML block scalars are string literals that are represented in YAML using the
+literal block notation, just like the example shown below:
+
+.. code-block:: yaml
+
+    text: |
+      First line
+      Second line
+
+The YAML I/O library provides support for translating between YAML block scalars
+and specific C++ types by allowing you to specialize BlockScalarTraits<> on
+your data type. The library doesn't provide any built-in support for block
+scalar I/O for types like std::string and llvm::StringRef as they are already
+supported by YAML I/O and use the ordinary scalar notation by default.
+
+BlockScalarTraits specializations are very similar to the
+ScalarTraits specialization - YAML I/O will provide the native type and your
+specialization must create a temporary llvm::StringRef when writing, and
+it will also provide an llvm::StringRef that has the value of that block scalar
+and your specialization must convert that to your native data type when reading.
+An example of a custom type with an appropriate specialization of
+BlockScalarTraits is shown below:
+
+.. code-block:: c++
+
+    using llvm::yaml::BlockScalarTraits;
+    using llvm::yaml::IO;
+
+    struct MyStringType {
+      std::string Str;
+    };
+
+    template <>
+    struct BlockScalarTraits<MyStringType> {
+      static void output(const MyStringType &Value, void *Ctxt,
+                         llvm::raw_ostream &OS) {
+        OS << Value.Str;
+      }
+
+      static StringRef input(StringRef Scalar, void *Ctxt,
+                             MyStringType &Value) {
+        Value.Str = Scalar.str();
+        return StringRef();
+      }
+    };
+
+    
+
+Mappings
+========
+
+To be translated to or from a YAML mapping for your type T you must specialize  
+llvm::yaml::MappingTraits on T and implement the "void mapping(IO &io, T&)" 
+method. If your native data structures use pointers to a class everywhere,
+you can specialize on the class pointer.  Examples:
+
+.. code-block:: c++
+   
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+    
+    // Example of struct Foo which is used by value
+    template <>
+    struct MappingTraits<Foo> {
+      static void mapping(IO &io, Foo &foo) {
+        io.mapOptional("size",      foo.size);
+      ...
+      }
+    };
+
+    // Example of struct Bar which is natively always a pointer
+    template <>
+    struct MappingTraits<Bar*> {
+      static void mapping(IO &io, Bar *&bar) {
+        io.mapOptional("size",    bar->size);
+      ...
+      }
+    };
+
+
+No Normalization
+----------------
+
+The mapping() method is responsible, if needed, for normalizing and 
+denormalizing. In a simple case where the native data structure requires no 
+normalization, the mapping method just uses mapOptional() or mapRequired() to 
+bind the struct's fields to YAML key names.  For example:
+
+.. code-block:: c++
+   
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+    
+    template <>
+    struct MappingTraits<Person> {
+      static void mapping(IO &io, Person &info) {
+        io.mapRequired("name",         info.name);
+        io.mapOptional("hat-size",     info.hatSize);
+      }
+    };
+
+
+Normalization
+----------------
+
+When [de]normalization is required, the mapping() method needs a way to access
+normalized values as fields. To help with this, there is
+a template MappingNormalization<> which you can then use to automatically
+do the normalization and denormalization.  The template is used to create
+a local variable in your mapping() method which contains the normalized keys.
+
+Suppose you have native data type 
+Polar which specifies a position in polar coordinates (distance, angle):
+
+.. code-block:: c++
+   
+    struct Polar {
+      float distance;
+      float angle;
+    };
+
+but you've decided the normalized YAML for should be in x,y coordinates. That 
+is, you want the yaml to look like:
+
+.. code-block:: yaml
+
+    x:   10.3
+    y:   -4.7
+
+You can support this by defining a MappingTraits that normalizes the polar
+coordinates to x,y coordinates when writing YAML and denormalizes x,y 
+coordinates into polar when reading YAML.  
+
+.. code-block:: c++
+   
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+        
+    template <>
+    struct MappingTraits<Polar> {
+      
+      class NormalizedPolar {
+      public:
+        NormalizedPolar(IO &io)
+          : x(0.0), y(0.0) {
+        }
+        NormalizedPolar(IO &, Polar &polar)
+          : x(polar.distance * cos(polar.angle)), 
+            y(polar.distance * sin(polar.angle)) {
+        }
+        Polar denormalize(IO &) {
+          return Polar(sqrt(x*x+y*y), arctan(x,y));
+        }
+         
+        float        x;
+        float        y;
+      };
+
+      static void mapping(IO &io, Polar &polar) {
+        MappingNormalization<NormalizedPolar, Polar> keys(io, polar);
+        
+        io.mapRequired("x",    keys->x);
+        io.mapRequired("y",    keys->y);
+      }
+    };
+
+When writing YAML, the local variable "keys" will be a stack allocated 
+instance of NormalizedPolar, constructed from the supplied polar object which
+initializes it x and y fields.  The mapRequired() methods then write out the x
+and y values as key/value pairs.  
+
+When reading YAML, the local variable "keys" will be a stack allocated instance
+of NormalizedPolar, constructed by the empty constructor.  The mapRequired 
+methods will find the matching key in the YAML document and fill in the x and y 
+fields of the NormalizedPolar object keys. At the end of the mapping() method
+when the local keys variable goes out of scope, the denormalize() method will
+automatically be called to convert the read values back to polar coordinates,
+and then assigned back to the second parameter to mapping().
+
+In some cases, the normalized class may be a subclass of the native type and
+could be returned by the denormalize() method, except that the temporary
+normalized instance is stack allocated.  In these cases, the utility template
+MappingNormalizationHeap<> can be used instead.  It just like 
+MappingNormalization<> except that it heap allocates the normalized object
+when reading YAML.  It never destroys the normalized object.  The denormalize()
+method can this return "this".
+
+
+Default values
+--------------
+Within a mapping() method, calls to io.mapRequired() mean that that key is 
+required to exist when parsing YAML documents, otherwise YAML I/O will issue an 
+error.
+
+On the other hand, keys registered with io.mapOptional() are allowed to not 
+exist in the YAML document being read.  So what value is put in the field 
+for those optional keys? 
+There are two steps to how those optional fields are filled in. First, the  
+second parameter to the mapping() method is a reference to a native class.  That
+native class must have a default constructor.  Whatever value the default
+constructor initially sets for an optional field will be that field's value.
+Second, the mapOptional() method has an optional third parameter.  If provided
+it is the value that mapOptional() should set that field to if the YAML document  
+does not have that key.  
+
+There is one important difference between those two ways (default constructor
+and third parameter to mapOptional). When YAML I/O generates a YAML document, 
+if the mapOptional() third parameter is used, if the actual value being written
+is the same as (using ==) the default value, then that key/value is not written.
+
+
+Order of Keys
+--------------
+
+When writing out a YAML document, the keys are written in the order that the
+calls to mapRequired()/mapOptional() are made in the mapping() method. This
+gives you a chance to write the fields in an order that a human reader of
+the YAML document would find natural.  This may be different that the order
+of the fields in the native class.
+
+When reading in a YAML document, the keys in the document can be in any order, 
+but they are processed in the order that the calls to mapRequired()/mapOptional() 
+are made in the mapping() method.  That enables some interesting 
+functionality.  For instance, if the first field bound is the cpu and the second
+field bound is flags, and the flags are cpu specific, you can programmatically
+switch how the flags are converted to and from YAML based on the cpu.  
+This works for both reading and writing. For example:
+
+.. code-block:: c++
+
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+    
+    struct Info {
+      CPUs        cpu;
+      uint32_t    flags;
+    };
+
+    template <>
+    struct MappingTraits<Info> {
+      static void mapping(IO &io, Info &info) {
+        io.mapRequired("cpu",       info.cpu);
+        // flags must come after cpu for this to work when reading yaml
+        if ( info.cpu == cpu_x86_64 )
+          io.mapRequired("flags",  *(My86_64Flags*)info.flags);
+        else
+          io.mapRequired("flags",  *(My86Flags*)info.flags);
+     }
+    };
+
+
+Tags
+----
+
+The YAML syntax supports tags as a way to specify the type of a node before
+it is parsed. This allows dynamic types of nodes.  But the YAML I/O model uses
+static typing, so there are limits to how you can use tags with the YAML I/O
+model. Recently, we added support to YAML I/O for checking/setting the optional 
+tag on a map. Using this functionality it is even possbile to support different 
+mappings, as long as they are convertable.  
+
+To check a tag, inside your mapping() method you can use io.mapTag() to specify
+what the tag should be.  This will also add that tag when writing yaml.
+
+Validation
+----------
+
+Sometimes in a yaml map, each key/value pair is valid, but the combination is
+not.  This is similar to something having no syntax errors, but still having
+semantic errors.  To support semantic level checking, YAML I/O allows
+an optional ``validate()`` method in a MappingTraits template specialization.  
+
+When parsing yaml, the ``validate()`` method is call *after* all key/values in 
+the map have been processed. Any error message returned by the ``validate()`` 
+method during input will be printed just a like a syntax error would be printed.
+When writing yaml, the ``validate()`` method is called *before* the yaml 
+key/values  are written.  Any error during output will trigger an ``assert()`` 
+because it is a programming error to have invalid struct values.
+
+
+.. code-block:: c++
+
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+    
+    struct Stuff {
+      ...
+    };
+
+    template <>
+    struct MappingTraits<Stuff> {
+      static void mapping(IO &io, Stuff &stuff) {
+      ...
+      }
+      static StringRef validate(IO &io, Stuff &stuff) {
+        // Look at all fields in 'stuff' and if there
+        // are any bad values return a string describing
+        // the error.  Otherwise return an empty string.
+        return StringRef();
+      }
+    };
+
+Flow Mapping
+------------
+A YAML "flow mapping" is a mapping that uses the inline notation
+(e.g { x: 1, y: 0 } ) when written to YAML. To specify that a type should be
+written in YAML using flow mapping, your MappingTraits specialization should
+add "static const bool flow = true;". For instance:
+
+.. code-block:: c++
+
+    using llvm::yaml::MappingTraits;
+    using llvm::yaml::IO;
+
+    struct Stuff {
+      ...
+    };
+
+    template <>
+    struct MappingTraits<Stuff> {
+      static void mapping(IO &io, Stuff &stuff) {
+        ...
+      }
+
+      static const bool flow = true;
+    }
+
+Flow mappings are subject to line wrapping according to the Output object
+configuration.
+
+Sequence
+========
+
+To be translated to or from a YAML sequence for your type T you must specialize
+llvm::yaml::SequenceTraits on T and implement two methods:
+``size_t size(IO &io, T&)`` and
+``T::value_type& element(IO &io, T&, size_t indx)``.  For example:
+
+.. code-block:: c++
+
+  template <>
+  struct SequenceTraits<MySeq> {
+    static size_t size(IO &io, MySeq &list) { ... }
+    static MySeqEl &element(IO &io, MySeq &list, size_t index) { ... }
+  };
+
+The size() method returns how many elements are currently in your sequence.
+The element() method returns a reference to the i'th element in the sequence. 
+When parsing YAML, the element() method may be called with an index one bigger
+than the current size.  Your element() method should allocate space for one
+more element (using default constructor if element is a C++ object) and returns
+a reference to that new allocated space.  
+
+
+Flow Sequence
+-------------
+A YAML "flow sequence" is a sequence that when written to YAML it uses the 
+inline notation (e.g [ foo, bar ] ).  To specify that a sequence type should
+be written in YAML as a flow sequence, your SequenceTraits specialization should
+add "static const bool flow = true;".  For instance:
+
+.. code-block:: c++
+
+  template <>
+  struct SequenceTraits<MyList> {
+    static size_t size(IO &io, MyList &list) { ... }
+    static MyListEl &element(IO &io, MyList &list, size_t index) { ... }
+    
+    // The existence of this member causes YAML I/O to use a flow sequence
+    static const bool flow = true;
+  };
+
+With the above, if you used MyList as the data type in your native data 
+structures, then when converted to YAML, a flow sequence of integers 
+will be used (e.g. [ 10, -3, 4 ]).
+
+Flow sequences are subject to line wrapping according to the Output object
+configuration.
+
+Utility Macros
+--------------
+Since a common source of sequences is std::vector<>, YAML I/O provides macros:
+LLVM_YAML_IS_SEQUENCE_VECTOR() and LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR() which
+can be used to easily specify SequenceTraits<> on a std::vector type.  YAML 
+I/O does not partial specialize SequenceTraits on std::vector<> because that
+would force all vectors to be sequences.  An example use of the macros:
+
+.. code-block:: c++
+
+  std::vector<MyType1>;
+  std::vector<MyType2>;
+  LLVM_YAML_IS_SEQUENCE_VECTOR(MyType1)
+  LLVM_YAML_IS_FLOW_SEQUENCE_VECTOR(MyType2)
+
+
+
+Document List
+=============
+
+YAML allows you to define multiple "documents" in a single YAML file.  Each 
+new document starts with a left aligned "---" token.  The end of all documents
+is denoted with a left aligned "..." token.  Many users of YAML will never
+have need for multiple documents.  The top level node in their YAML schema
+will be a mapping or sequence. For those cases, the following is not needed.
+But for cases where you do want multiple documents, you can specify a
+trait for you document list type.  The trait has the same methods as 
+SequenceTraits but is named DocumentListTraits.  For example:
+
+.. code-block:: c++
+
+  template <>
+  struct DocumentListTraits<MyDocList> {
+    static size_t size(IO &io, MyDocList &list) { ... }
+    static MyDocType element(IO &io, MyDocList &list, size_t index) { ... }
+  };
+
+
+User Context Data
+=================
+When an llvm::yaml::Input or llvm::yaml::Output object is created their 
+constructors take an optional "context" parameter.  This is a pointer to 
+whatever state information you might need.  
+
+For instance, in a previous example we showed how the conversion type for a 
+flags field could be determined at runtime based on the value of another field 
+in the mapping. But what if an inner mapping needs to know some field value
+of an outer mapping?  That is where the "context" parameter comes in. You
+can set values in the context in the outer map's mapping() method and
+retrieve those values in the inner map's mapping() method.
+
+The context value is just a void*.  All your traits which use the context 
+and operate on your native data types, need to agree what the context value
+actually is.  It could be a pointer to an object or struct which your various
+traits use to shared context sensitive information.
+
+
+Output
+======
+
+The llvm::yaml::Output class is used to generate a YAML document from your 
+in-memory data structures, using traits defined on your data types.  
+To instantiate an Output object you need an llvm::raw_ostream, an optional 
+context pointer and an optional wrapping column:
+
+.. code-block:: c++
+
+      class Output : public IO {
+      public:
+        Output(llvm::raw_ostream &, void *context = NULL, int WrapColumn = 70);
+    
+Once you have an Output object, you can use the C++ stream operator on it
+to write your native data as YAML. One thing to recall is that a YAML file
+can contain multiple "documents".  If the top level data structure you are
+streaming as YAML is a mapping, scalar, or sequence, then Output assumes you
+are generating one document and wraps the mapping output 
+with  "``---``" and trailing "``...``".  
+
+The WrapColumn parameter will cause the flow mappings and sequences to
+line-wrap when they go over the supplied column. Pass 0 to completely
+suppress the wrapping.
+
+.. code-block:: c++
+   
+    using llvm::yaml::Output;
+
+    void dumpMyMapDoc(const MyMapType &info) {
+      Output yout(llvm::outs());
+      yout << info;
+    }
+
+The above could produce output like:
+
+.. code-block:: yaml
+
+     ---
+     name:      Tom
+     hat-size:  7
+     ...
+
+On the other hand, if the top level data structure you are streaming as YAML
+has a DocumentListTraits specialization, then Output walks through each element
+of your DocumentList and generates a "---" before the start of each element
+and ends with a "...".
+
+.. code-block:: c++
+   
+    using llvm::yaml::Output;
+
+    void dumpMyMapDoc(const MyDocListType &docList) {
+      Output yout(llvm::outs());
+      yout << docList;
+    }
+
+The above could produce output like:
+
+.. code-block:: yaml
+
+     ---
+     name:      Tom
+     hat-size:  7
+     ---
+     name:      Tom
+     shoe-size:  11
+     ...
+
+Input
+=====
+
+The llvm::yaml::Input class is used to parse YAML document(s) into your native
+data structures. To instantiate an Input
+object you need a StringRef to the entire YAML file, and optionally a context 
+pointer:
+
+.. code-block:: c++
+
+      class Input : public IO {
+      public:
+        Input(StringRef inputContent, void *context=NULL);
+    
+Once you have an Input object, you can use the C++ stream operator to read
+the document(s).  If you expect there might be multiple YAML documents in
+one file, you'll need to specialize DocumentListTraits on a list of your
+document type and stream in that document list type.  Otherwise you can
+just stream in the document type.  Also, you can check if there was 
+any syntax errors in the YAML be calling the error() method on the Input
+object.  For example:
+
+.. code-block:: c++
+   
+     // Reading a single document
+     using llvm::yaml::Input;
+
+     Input yin(mb.getBuffer());
+     
+     // Parse the YAML file
+     MyDocType theDoc;
+     yin >> theDoc;
+
+     // Check for error
+     if ( yin.error() )
+       return;
+  
+      
+.. code-block:: c++
+   
+     // Reading multiple documents in one file
+     using llvm::yaml::Input;
+
+     LLVM_YAML_IS_DOCUMENT_LIST_VECTOR(std::vector<MyDocType>)
+     
+     Input yin(mb.getBuffer());
+     
+     // Parse the YAML file
+     std::vector<MyDocType> theDocList;
+     yin >> theDocList;
+
+     // Check for error
+     if ( yin.error() )
+       return;
+
+

Added: www-releases/trunk/4.0.0/docs/_sources/index.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/index.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/index.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/index.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,534 @@
+Overview
+========
+
+The LLVM compiler infrastructure supports a wide range of projects, from
+industrial strength compilers to specialized JIT applications to small
+research projects.
+
+Similarly, documentation is broken down into several high-level groupings
+targeted at different audiences:
+
+LLVM Design & Overview
+======================
+
+Several introductory papers and presentations.
+
+.. toctree::
+   :hidden:
+
+   LangRef
+
+:doc:`LangRef`
+  Defines the LLVM intermediate representation.
+
+`Introduction to the LLVM Compiler`__
+  Presentation providing a users introduction to LLVM.
+
+  .. __: http://llvm.org/pubs/2008-10-04-ACAT-LLVM-Intro.html
+
+`Intro to LLVM`__
+  Book chapter providing a compiler hacker's introduction to LLVM.
+
+  .. __: http://www.aosabook.org/en/llvm.html
+
+
+`LLVM: A Compilation Framework for Lifelong Program Analysis & Transformation`__
+  Design overview.
+
+  .. __: http://llvm.org/pubs/2004-01-30-CGO-LLVM.html
+
+`LLVM: An Infrastructure for Multi-Stage Optimization`__
+  More details (quite old now).
+
+  .. __: http://llvm.org/pubs/2002-12-LattnerMSThesis.html
+
+`Publications mentioning LLVM <http://llvm.org/pubs>`_
+   ..
+
+User Guides
+===========
+
+For those new to the LLVM system.
+
+NOTE: If you are a user who is only interested in using LLVM-based
+compilers, you should look into `Clang <http://clang.llvm.org>`_ or
+`DragonEgg <http://dragonegg.llvm.org>`_ instead. The documentation here is
+intended for users who have a need to work with the intermediate LLVM
+representation.
+
+.. toctree::
+   :hidden:
+
+   CMake
+   CMakePrimer
+   AdvancedBuilds
+   HowToBuildOnARM
+   HowToCrossCompileLLVM
+   CommandGuide/index
+   GettingStarted
+   GettingStartedVS
+   FAQ
+   Lexicon
+   HowToAddABuilder
+   yaml2obj
+   HowToSubmitABug
+   SphinxQuickstartTemplate
+   Phabricator
+   TestingGuide
+   tutorial/index
+   ReleaseNotes
+   Passes
+   YamlIO
+   GetElementPtr
+   Frontend/PerformanceTips
+   MCJITDesignAndImplementation
+   CodeOfConduct
+   CompileCudaWithLLVM
+   ReportingGuide
+
+:doc:`GettingStarted`
+   Discusses how to get up and running quickly with the LLVM infrastructure.
+   Everything from unpacking and compilation of the distribution to execution
+   of some tools.
+
+:doc:`CMake`
+   An addendum to the main Getting Started guide for those using the `CMake
+   build system <http://www.cmake.org>`_.
+
+:doc:`HowToBuildOnARM`
+   Notes on building and testing LLVM/Clang on ARM.
+
+:doc:`HowToCrossCompileLLVM`
+   Notes on cross-building and testing LLVM/Clang.
+
+:doc:`GettingStartedVS`
+   An addendum to the main Getting Started guide for those using Visual Studio
+   on Windows.
+
+:doc:`tutorial/index`
+   Tutorials about using LLVM. Includes a tutorial about making a custom
+   language with LLVM.
+
+:doc:`LLVM Command Guide <CommandGuide/index>`
+   A reference manual for the LLVM command line utilities ("man" pages for LLVM
+   tools).
+
+:doc:`Passes`
+   A list of optimizations and analyses implemented in LLVM.
+
+:doc:`FAQ`
+   A list of common questions and problems and their solutions.
+
+:doc:`Release notes for the current release <ReleaseNotes>`
+   This describes new features, known bugs, and other limitations.
+
+:doc:`HowToSubmitABug`
+   Instructions for properly submitting information about any bugs you run into
+   in the LLVM system.
+
+:doc:`SphinxQuickstartTemplate`
+  A template + tutorial for writing new Sphinx documentation. It is meant
+  to be read in source form.
+
+:doc:`LLVM Testing Infrastructure Guide <TestingGuide>`
+   A reference manual for using the LLVM testing infrastructure.
+
+`How to build the C, C++, ObjC, and ObjC++ front end`__
+   Instructions for building the clang front-end from source.
+
+   .. __: http://clang.llvm.org/get_started.html
+
+:doc:`Lexicon`
+   Definition of acronyms, terms and concepts used in LLVM.
+
+:doc:`HowToAddABuilder`
+   Instructions for adding new builder to LLVM buildbot master.
+
+:doc:`YamlIO`
+   A reference guide for using LLVM's YAML I/O library.
+
+:doc:`GetElementPtr`
+  Answers to some very frequent questions about LLVM's most frequently
+  misunderstood instruction.
+
+:doc:`Frontend/PerformanceTips`
+   A collection of tips for frontend authors on how to generate IR 
+   which LLVM is able to effectively optimize.
+
+
+Programming Documentation
+=========================
+
+For developers of applications which use LLVM as a library.
+
+.. toctree::
+   :hidden:
+
+   Atomics
+   CodingStandards
+   CommandLine
+   CompilerWriterInfo
+   ExtendingLLVM
+   HowToSetUpLLVMStyleRTTI
+   ProgrammersManual
+   Extensions
+   LibFuzzer
+   ScudoHardenedAllocator
+   OptBisect
+
+:doc:`LLVM Language Reference Manual <LangRef>`
+  Defines the LLVM intermediate representation and the assembly form of the
+  different nodes.
+
+:doc:`Atomics`
+  Information about LLVM's concurrency model.
+
+:doc:`ProgrammersManual`
+  Introduction to the general layout of the LLVM sourcebase, important classes
+  and APIs, and some tips & tricks.
+
+:doc:`Extensions`
+  LLVM-specific extensions to tools and formats LLVM seeks compatibility with.
+
+:doc:`CommandLine`
+  Provides information on using the command line parsing library.
+
+:doc:`CodingStandards`
+  Details the LLVM coding standards and provides useful information on writing
+  efficient C++ code.
+
+:doc:`HowToSetUpLLVMStyleRTTI`
+  How to make ``isa<>``, ``dyn_cast<>``, etc. available for clients of your
+  class hierarchy.
+
+:doc:`ExtendingLLVM`
+  Look here to see how to add instructions and intrinsics to LLVM.
+
+`Doxygen generated documentation <http://llvm.org/doxygen/>`_
+  (`classes <http://llvm.org/doxygen/inherits.html>`_)
+  (`tarball <http://llvm.org/doxygen/doxygen.tar.gz>`_)
+
+`Documentation for Go bindings <http://godoc.org/llvm.org/llvm/bindings/go/llvm>`_
+
+`ViewVC Repository Browser <http://llvm.org/viewvc/>`_
+   ..
+
+:doc:`CompilerWriterInfo`
+  A list of helpful links for compiler writers.
+
+:doc:`LibFuzzer`
+  A library for writing in-process guided fuzzers.
+
+:doc:`ScudoHardenedAllocator`
+  A library that implements a security-hardened `malloc()`.
+
+:doc:`OptBisect`
+  A command line option for debugging optimization-induced failures.
+
+Subsystem Documentation
+=======================
+
+For API clients and LLVM developers.
+
+.. toctree::
+   :hidden:
+
+   AliasAnalysis
+   MemorySSA
+   BitCodeFormat
+   BlockFrequencyTerminology
+   BranchWeightMetadata
+   Bugpoint
+   CodeGenerator
+   ExceptionHandling
+   LinkTimeOptimization
+   SegmentedStacks
+   TableGenFundamentals
+   TableGen/index
+   DebuggingJITedCode
+   GoldPlugin
+   MarkedUpDisassembly
+   SystemLibrary
+   SourceLevelDebugging
+   Vectorizers
+   WritingAnLLVMBackend
+   GarbageCollection
+   WritingAnLLVMPass
+   HowToUseAttributes
+   NVPTXUsage
+   AMDGPUUsage
+   StackMaps
+   InAlloca
+   BigEndianNEON
+   CoverageMappingFormat
+   Statepoints
+   MergeFunctions
+   TypeMetadata
+   FaultMaps
+   MIRLangRef
+   Coroutines
+   GlobalISel
+   XRay
+   PDB/index
+
+:doc:`WritingAnLLVMPass`
+   Information on how to write LLVM transformations and analyses.
+
+:doc:`WritingAnLLVMBackend`
+   Information on how to write LLVM backends for machine targets.
+
+:doc:`CodeGenerator`
+   The design and implementation of the LLVM code generator.  Useful if you are
+   working on retargetting LLVM to a new architecture, designing a new codegen
+   pass, or enhancing existing components.
+
+:doc:`Machine IR (MIR) Format Reference Manual <MIRLangRef>`
+   A reference manual for the MIR serialization format, which is used to test
+   LLVM's code generation passes.
+
+:doc:`TableGen <TableGen/index>`
+   Describes the TableGen tool, which is used heavily by the LLVM code
+   generator.
+
+:doc:`AliasAnalysis`
+   Information on how to write a new alias analysis implementation or how to
+   use existing analyses.
+
+:doc:`MemorySSA`
+   Information about the MemorySSA utility in LLVM, as well as how to use it.
+
+:doc:`GarbageCollection`
+   The interfaces source-language compilers should use for compiling GC'd
+   programs.
+
+:doc:`Source Level Debugging with LLVM <SourceLevelDebugging>`
+   This document describes the design and philosophy behind the LLVM
+   source-level debugger.
+
+:doc:`Vectorizers`
+   This document describes the current status of vectorization in LLVM.
+
+:doc:`ExceptionHandling`
+   This document describes the design and implementation of exception handling
+   in LLVM.
+
+:doc:`Bugpoint`
+   Automatic bug finder and test-case reducer description and usage
+   information.
+
+:doc:`BitCodeFormat`
+   This describes the file format and encoding used for LLVM "bc" files.
+
+:doc:`System Library <SystemLibrary>`
+   This document describes the LLVM System Library (``lib/System``) and
+   how to keep LLVM source code portable
+
+:doc:`LinkTimeOptimization`
+   This document describes the interface between LLVM intermodular optimizer
+   and the linker and its design
+
+:doc:`GoldPlugin`
+   How to build your programs with link-time optimization on Linux.
+
+:doc:`DebuggingJITedCode`
+   How to debug JITed code with GDB.
+
+:doc:`MCJITDesignAndImplementation`
+   Describes the inner workings of MCJIT execution engine.
+
+:doc:`BranchWeightMetadata`
+   Provides information about Branch Prediction Information.
+
+:doc:`BlockFrequencyTerminology`
+   Provides information about terminology used in the ``BlockFrequencyInfo``
+   analysis pass.
+
+:doc:`SegmentedStacks`
+   This document describes segmented stacks and how they are used in LLVM.
+
+:doc:`MarkedUpDisassembly`
+   This document describes the optional rich disassembly output syntax.
+
+:doc:`HowToUseAttributes`
+  Answers some questions about the new Attributes infrastructure.
+
+:doc:`NVPTXUsage`
+   This document describes using the NVPTX back-end to compile GPU kernels.
+
+:doc:`AMDGPUUsage`
+   This document describes how to use the AMDGPU back-end.
+
+:doc:`StackMaps`
+  LLVM support for mapping instruction addresses to the location of
+  values and allowing code to be patched.
+
+:doc:`BigEndianNEON`
+  LLVM's support for generating NEON instructions on big endian ARM targets is
+  somewhat nonintuitive. This document explains the implementation and rationale.
+
+:doc:`CoverageMappingFormat`
+  This describes the format and encoding used for LLVM’s code coverage mapping.
+
+:doc:`Statepoints`
+  This describes a set of experimental extensions for garbage
+  collection support.
+
+:doc:`MergeFunctions`
+  Describes functions merging optimization.
+
+:doc:`InAlloca`
+  Description of the ``inalloca`` argument attribute.
+
+:doc:`FaultMaps`
+  LLVM support for folding control flow into faulting machine instructions.
+
+:doc:`CompileCudaWithLLVM`
+  LLVM support for CUDA.
+
+:doc:`Coroutines`
+  LLVM support for coroutines.
+
+:doc:`GlobalISel`
+  This describes the prototype instruction selection replacement, GlobalISel.
+
+:doc:`XRay`
+  High-level documentation of how to use XRay in LLVM.
+
+:doc:`The Microsoft PDB File Format <PDB/index>`
+  A detailed description of the Microsoft PDB (Program Database) file format.
+
+Development Process Documentation
+=================================
+
+Information about LLVM's development process.
+
+.. toctree::
+   :hidden:
+
+   DeveloperPolicy
+   Projects
+   LLVMBuild
+   HowToReleaseLLVM
+   Packaging
+   ReleaseProcess
+   Phabricator
+
+:doc:`DeveloperPolicy`
+   The LLVM project's policy towards developers and their contributions.
+
+:doc:`Projects`
+  How-to guide and templates for new projects that *use* the LLVM
+  infrastructure.  The templates (directory organization, Makefiles, and test
+  tree) allow the project code to be located outside (or inside) the ``llvm/``
+  tree, while using LLVM header files and libraries.
+
+:doc:`LLVMBuild`
+  Describes the LLVMBuild organization and files used by LLVM to specify
+  component descriptions.
+
+:doc:`HowToReleaseLLVM`
+  This is a guide to preparing LLVM releases. Most developers can ignore it.
+
+:doc:`ReleaseProcess`
+  This is a guide to validate a new release, during the release process. Most developers can ignore it.
+
+:doc:`Packaging`
+   Advice on packaging LLVM into a distribution.
+
+:doc:`Phabricator`
+   Describes how to use the Phabricator code review tool hosted on
+   http://reviews.llvm.org/ and its command line interface, Arcanist.
+
+Community
+=========
+
+LLVM has a thriving community of friendly and helpful developers.
+The two primary communication mechanisms in the LLVM community are mailing
+lists and IRC.
+
+Mailing Lists
+-------------
+
+If you can't find what you need in these docs, try consulting the mailing
+lists.
+
+`Developer's List (llvm-dev)`__
+  This list is for people who want to be included in technical discussions of
+  LLVM. People post to this list when they have questions about writing code
+  for or using the LLVM tools. It is relatively low volume.
+
+  .. __: http://lists.llvm.org/mailman/listinfo/llvm-dev
+
+`Commits Archive (llvm-commits)`__
+  This list contains all commit messages that are made when LLVM developers
+  commit code changes to the repository. It also serves as a forum for
+  patch review (i.e. send patches here). It is useful for those who want to
+  stay on the bleeding edge of LLVM development. This list is very high
+  volume.
+
+  .. __: http://lists.llvm.org/pipermail/llvm-commits/
+
+`Bugs & Patches Archive (llvm-bugs)`__
+  This list gets emailed every time a bug is opened and closed. It is
+  higher volume than the LLVM-dev list.
+
+  .. __: http://lists.llvm.org/pipermail/llvm-bugs/
+
+`Test Results Archive (llvm-testresults)`__
+  A message is automatically sent to this list by every active nightly tester
+  when it completes.  As such, this list gets email several times each day,
+  making it a high volume list.
+
+  .. __: http://lists.llvm.org/pipermail/llvm-testresults/
+
+`LLVM Announcements List (llvm-announce)`__
+  This is a low volume list that provides important announcements regarding
+  LLVM.  It gets email about once a month.
+
+  .. __: http://lists.llvm.org/mailman/listinfo/llvm-announce
+
+IRC
+---
+
+Users and developers of the LLVM project (including subprojects such as Clang)
+can be found in #llvm on `irc.oftc.net <irc://irc.oftc.net/llvm>`_.
+
+This channel has several bots.
+
+* Buildbot reporters
+
+  * llvmbb - Bot for the main LLVM buildbot master.
+    http://lab.llvm.org:8011/console
+  * bb-chapuni - An individually run buildbot master. http://bb.pgr.jp/console
+  * smooshlab - Apple's internal buildbot master.
+
+* robot - Bugzilla linker. %bug <number>
+
+* clang-bot - A `geordi <http://www.eelis.net/geordi/>`_ instance running
+  near-trunk clang instead of gcc.
+
+Community wide proposals
+------------------------
+
+Proposals for massive changes in how the community behaves and how the work flow
+can be better.
+
+.. toctree::
+   :hidden:
+
+   CodeOfConduct
+   Proposals/GitHubMove
+
+:doc:`CodeOfConduct`
+   Proposal to adopt a code of conduct on the LLVM social spaces (lists, events,
+   IRC, etc).
+
+:doc:`Proposals/GitHubMove`
+   Proposal to move from SVN/Git to GitHub.
+
+
+Indices and tables
+==================
+
+* :ref:`genindex`
+* :ref:`search`

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT1.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT1.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT1.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT1.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,374 @@
+=======================================================
+Building a JIT: Starting out with KaleidoscopeJIT
+=======================================================
+
+.. contents::
+   :local:
+
+Chapter 1 Introduction
+======================
+
+Welcome to Chapter 1 of the "Building an ORC-based JIT in LLVM" tutorial. This
+tutorial runs through the implementation of a JIT compiler using LLVM's
+On-Request-Compilation (ORC) APIs. It begins with a simplified version of the
+KaleidoscopeJIT class used in the
+`Implementing a language with LLVM <LangImpl1.html>`_ tutorials and then
+introduces new features like optimization, lazy compilation and remote
+execution.
+
+The goal of this tutorial is to introduce you to LLVM's ORC JIT APIs, show how
+these APIs interact with other parts of LLVM, and to teach you how to recombine
+them to build a custom JIT that is suited to your use-case.
+
+The structure of the tutorial is:
+
+- Chapter #1: Investigate the simple KaleidoscopeJIT class. This will
+  introduce some of the basic concepts of the ORC JIT APIs, including the
+  idea of an ORC *Layer*.
+
+- `Chapter #2 <BuildingAJIT2.html>`_: Extend the basic KaleidoscopeJIT by adding
+  a new layer that will optimize IR and generated code.
+
+- `Chapter #3 <BuildingAJIT3.html>`_: Further extend the JIT by adding a
+  Compile-On-Demand layer to lazily compile IR.
+
+- `Chapter #4 <BuildingAJIT4.html>`_: Improve the laziness of our JIT by
+  replacing the Compile-On-Demand layer with a custom layer that uses the ORC
+  Compile Callbacks API directly to defer IR-generation until functions are
+  called.
+
+- `Chapter #5 <BuildingAJIT5.html>`_: Add process isolation by JITing code into
+  a remote process with reduced privileges using the JIT Remote APIs.
+
+To provide input for our JIT we will use the Kaleidoscope REPL from
+`Chapter 7 <LangImpl7.html>`_ of the "Implementing a language in LLVM tutorial",
+with one minor modification: We will remove the FunctionPassManager from the
+code for that chapter and replace it with optimization support in our JIT class
+in Chapter #2.
+
+Finally, a word on API generations: ORC is the 3rd generation of LLVM JIT API.
+It was preceded by MCJIT, and before that by the (now deleted) legacy JIT.
+These tutorials don't assume any experience with these earlier APIs, but
+readers acquainted with them will see many familiar elements. Where appropriate
+we will make this connection with the earlier APIs explicit to help people who
+are transitioning from them to ORC.
+
+JIT API Basics
+==============
+
+The purpose of a JIT compiler is to compile code "on-the-fly" as it is needed,
+rather than compiling whole programs to disk ahead of time as a traditional
+compiler does. To support that aim our initial, bare-bones JIT API will be:
+
+1. Handle addModule(Module &M) -- Make the given IR module available for
+   execution.
+2. JITSymbol findSymbol(const std::string &Name) -- Search for pointers to
+   symbols (functions or variables) that have been added to the JIT.
+3. void removeModule(Handle H) -- Remove a module from the JIT, releasing any
+   memory that had been used for the compiled code.
+
+A basic use-case for this API, executing the 'main' function from a module,
+will look like:
+
+.. code-block:: c++
+
+  std::unique_ptr<Module> M = buildModule();
+  JIT J;
+  Handle H = J.addModule(*M);
+  int (*Main)(int, char*[]) =
+    (int(*)(int, char*[])J.findSymbol("main").getAddress();
+  int Result = Main();
+  J.removeModule(H);
+
+The APIs that we build in these tutorials will all be variations on this simple
+theme. Behind the API we will refine the implementation of the JIT to add
+support for optimization and lazy compilation. Eventually we will extend the
+API itself to allow higher-level program representations (e.g. ASTs) to be
+added to the JIT.
+
+KaleidoscopeJIT
+===============
+
+In the previous section we described our API, now we examine a simple
+implementation of it: The KaleidoscopeJIT class [1]_ that was used in the
+`Implementing a language with LLVM <LangImpl1.html>`_ tutorials. We will use
+the REPL code from `Chapter 7 <LangImpl7.html>`_ of that tutorial to supply the
+input for our JIT: Each time the user enters an expression the REPL will add a
+new IR module containing the code for that expression to the JIT. If the
+expression is a top-level expression like '1+1' or 'sin(x)', the REPL will also
+use the findSymbol method of our JIT class find and execute the code for the
+expression, and then use the removeModule method to remove the code again
+(since there's no way to re-invoke an anonymous expression). In later chapters
+of this tutorial we'll modify the REPL to enable new interactions with our JIT
+class, but for now we will take this setup for granted and focus our attention on
+the implementation of our JIT itself.
+
+Our KaleidoscopeJIT class is defined in the KaleidoscopeJIT.h header. After the
+usual include guards and #includes [2]_, we get to the definition of our class:
+
+.. code-block:: c++
+
+  #ifndef LLVM_EXECUTIONENGINE_ORC_KALEIDOSCOPEJIT_H
+  #define LLVM_EXECUTIONENGINE_ORC_KALEIDOSCOPEJIT_H
+
+  #include "llvm/ExecutionEngine/ExecutionEngine.h"
+  #include "llvm/ExecutionEngine/RTDyldMemoryManager.h"
+  #include "llvm/ExecutionEngine/Orc/CompileUtils.h"
+  #include "llvm/ExecutionEngine/Orc/IRCompileLayer.h"
+  #include "llvm/ExecutionEngine/Orc/LambdaResolver.h"
+  #include "llvm/ExecutionEngine/Orc/ObjectLinkingLayer.h"
+  #include "llvm/IR/Mangler.h"
+  #include "llvm/Support/DynamicLibrary.h"
+
+  namespace llvm {
+  namespace orc {
+
+  class KaleidoscopeJIT {
+  private:
+
+    std::unique_ptr<TargetMachine> TM;
+    const DataLayout DL;
+    ObjectLinkingLayer<> ObjectLayer;
+    IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
+
+  public:
+
+    typedef decltype(CompileLayer)::ModuleSetHandleT ModuleHandleT;
+
+Our class begins with four members: A TargetMachine, TM, which will be used
+to build our LLVM compiler instance; A DataLayout, DL, which will be used for
+symbol mangling (more on that later), and two ORC *layers*: an
+ObjectLinkingLayer and a IRCompileLayer. We'll be talking more about layers in
+the next chapter, but for now you can think of them as analogous to LLVM
+Passes: they wrap up useful JIT utilities behind an easy to compose interface.
+The first layer, ObjectLinkingLayer, is the foundation of our JIT: it takes
+in-memory object files produced by a compiler and links them on the fly to make
+them executable. This JIT-on-top-of-a-linker design was introduced in MCJIT,
+however the linker was hidden inside the MCJIT class. In ORC we expose the
+linker so that clients can access and configure it directly if they need to. In
+this tutorial our ObjectLinkingLayer will just be used to support the next layer
+in our stack: the IRCompileLayer, which will be responsible for taking LLVM IR,
+compiling it, and passing the resulting in-memory object files down to the
+object linking layer below.
+
+That's it for member variables, after that we have a single typedef:
+ModuleHandle. This is the handle type that will be returned from our JIT's
+addModule method, and can be passed to the removeModule method to remove a
+module. The IRCompileLayer class already provides a convenient handle type
+(IRCompileLayer::ModuleSetHandleT), so we just alias our ModuleHandle to this.
+
+.. code-block:: c++
+
+  KaleidoscopeJIT()
+      : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),
+    CompileLayer(ObjectLayer, SimpleCompiler(*TM)) {
+    llvm::sys::DynamicLibrary::LoadLibraryPermanently(nullptr);
+  }
+
+  TargetMachine &getTargetMachine() { return *TM; }
+
+Next up we have our class constructor. We begin by initializing TM using the
+EngineBuilder::selectTarget helper method, which constructs a TargetMachine for
+the current process. Next we use our newly created TargetMachine to initialize
+DL, our DataLayout. Then we initialize our IRCompileLayer. Our IRCompile layer
+needs two things: (1) A reference to our object linking layer, and (2) a
+compiler instance to use to perform the actual compilation from IR to object
+files. We use the off-the-shelf SimpleCompiler instance for now. Finally, in
+the body of the constructor, we call the DynamicLibrary::LoadLibraryPermanently
+method with a nullptr argument. Normally the LoadLibraryPermanently method is
+called with the path of a dynamic library to load, but when passed a null
+pointer it will 'load' the host process itself, making its exported symbols
+available for execution.
+
+.. code-block:: c++
+
+  ModuleHandle addModule(std::unique_ptr<Module> M) {
+    // Build our symbol resolver:
+    // Lambda 1: Look back into the JIT itself to find symbols that are part of
+    //           the same "logical dylib".
+    // Lambda 2: Search for external symbols in the host process.
+    auto Resolver = createLambdaResolver(
+        [&](const std::string &Name) {
+          if (auto Sym = CompileLayer.findSymbol(Name, false))
+            return Sym;
+          return JITSymbol(nullptr);
+        },
+        [](const std::string &S) {
+          if (auto SymAddr =
+                RTDyldMemoryManager::getSymbolAddressInProcess(Name))
+            return JITSymbol(SymAddr, JITSymbolFlags::Exported);
+          return JITSymbol(nullptr);
+        });
+
+    // Build a singlton module set to hold our module.
+    std::vector<std::unique_ptr<Module>> Ms;
+    Ms.push_back(std::move(M));
+
+    // Add the set to the JIT with the resolver we created above and a newly
+    // created SectionMemoryManager.
+    return CompileLayer.addModuleSet(std::move(Ms),
+                                     make_unique<SectionMemoryManager>(),
+                                     std::move(Resolver));
+  }
+
+Now we come to the first of our JIT API methods: addModule. This method is
+responsible for adding IR to the JIT and making it available for execution. In
+this initial implementation of our JIT we will make our modules "available for
+execution" by adding them straight to the IRCompileLayer, which will
+immediately compile them. In later chapters we will teach our JIT to be lazier
+and instead add the Modules to a "pending" list to be compiled if and when they
+are first executed.
+
+To add our module to the IRCompileLayer we need to supply two auxiliary objects
+(as well as the module itself): a memory manager and a symbol resolver.  The
+memory manager will be responsible for managing the memory allocated to JIT'd
+machine code, setting memory permissions, and registering exception handling
+tables (if the JIT'd code uses exceptions). For our memory manager we will use
+the SectionMemoryManager class: another off-the-shelf utility that provides all
+the basic functionality we need. The second auxiliary class, the symbol
+resolver, is more interesting for us. It exists to tell the JIT where to look
+when it encounters an *external symbol* in the module we are adding.  External
+symbols are any symbol not defined within the module itself, including calls to
+functions outside the JIT and calls to functions defined in other modules that
+have already been added to the JIT. It may seem as though modules added to the
+JIT should "know about one another" by default, but since we would still have to
+supply a symbol resolver for references to code outside the JIT it turns out to
+be easier to just re-use this one mechanism for all symbol resolution. This has
+the added benefit that the user has full control over the symbol resolution
+process. Should we search for definitions within the JIT first, then fall back
+on external definitions? Or should we prefer external definitions where
+available and only JIT code if we don't already have an available
+implementation? By using a single symbol resolution scheme we are free to choose
+whatever makes the most sense for any given use case.
+
+Building a symbol resolver is made especially easy by the *createLambdaResolver*
+function. This function takes two lambdas [3]_ and returns a JITSymbolResolver
+instance. The first lambda is used as the implementation of the resolver's
+findSymbolInLogicalDylib method, which searches for symbol definitions that
+should be thought of as being part of the same "logical" dynamic library as this
+Module. If you are familiar with static linking: this means that
+findSymbolInLogicalDylib should expose symbols with common linkage and hidden
+visibility. If all this sounds foreign you can ignore the details and just
+remember that this is the first method that the linker will use to try to find a
+symbol definition. If the findSymbolInLogicalDylib method returns a null result
+then the linker will call the second symbol resolver method, called findSymbol,
+which searches for symbols that should be thought of as external to (but
+visibile from) the module and its logical dylib. In this tutorial we will adopt
+the following simple scheme: All modules added to the JIT will behave as if they
+were linked into a single, ever-growing logical dylib. To implement this our
+first lambda (the one defining findSymbolInLogicalDylib) will just search for
+JIT'd code by calling the CompileLayer's findSymbol method. If we don't find a
+symbol in the JIT itself we'll fall back to our second lambda, which implements
+findSymbol. This will use the RTDyldMemoyrManager::getSymbolAddressInProcess
+method to search for the symbol within the program itself. If we can't find a
+symbol definition via either of these paths the JIT will refuse to accept our
+module, returning a "symbol not found" error.
+
+Now that we've built our symbol resolver we're ready to add our module to the
+JIT. We do this by calling the CompileLayer's addModuleSet method [4]_. Since
+we only have a single Module and addModuleSet expects a collection, we will
+create a vector of modules and add our module as the only member. Since we
+have already typedef'd our ModuleHandle type to be the same as the
+CompileLayer's handle type, we can return the handle from addModuleSet
+directly from our addModule method.
+
+.. code-block:: c++
+
+  JITSymbol findSymbol(const std::string Name) {
+    std::string MangledName;
+    raw_string_ostream MangledNameStream(MangledName);
+    Mangler::getNameWithPrefix(MangledNameStream, Name, DL);
+    return CompileLayer.findSymbol(MangledNameStream.str(), true);
+  }
+
+  void removeModule(ModuleHandle H) {
+    CompileLayer.removeModuleSet(H);
+  }
+
+Now that we can add code to our JIT, we need a way to find the symbols we've
+added to it. To do that we call the findSymbol method on our IRCompileLayer,
+but with a twist: We have to *mangle* the name of the symbol we're searching
+for first. The reason for this is that the ORC JIT components use mangled
+symbols internally the same way a static compiler and linker would, rather
+than using plain IR symbol names. The kind of mangling will depend on the
+DataLayout, which in turn depends on the target platform. To allow us to
+remain portable and search based on the un-mangled name, we just re-produce
+this mangling ourselves.
+
+We now come to the last method in our JIT API: removeModule. This method is
+responsible for destructing the MemoryManager and SymbolResolver that were
+added with a given module, freeing any resources they were using in the
+process. In our Kaleidoscope demo we rely on this method to remove the module
+representing the most recent top-level expression, preventing it from being
+treated as a duplicate definition when the next top-level expression is
+entered. It is generally good to free any module that you know you won't need
+to call further, just to free up the resources dedicated to it. However, you
+don't strictly need to do this: All resources will be cleaned up when your
+JIT class is destructed, if the haven't been freed before then.
+
+This brings us to the end of Chapter 1 of Building a JIT. You now have a basic
+but fully functioning JIT stack that you can use to take LLVM IR and make it
+executable within the context of your JIT process. In the next chapter we'll
+look at how to extend this JIT to produce better quality code, and in the
+process take a deeper look at the ORC layer concept.
+
+`Next: Extending the KaleidoscopeJIT <BuildingAJIT2.html>`_
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example. To build this
+example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core orc native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/BuildingAJIT/Chapter1/KaleidoscopeJIT.h
+   :language: c++
+
+.. [1] Actually we use a cut-down version of KaleidoscopeJIT that makes a
+       simplifying assumption: symbols cannot be re-defined. This will make it
+       impossible to re-define symbols in the REPL, but will make our symbol
+       lookup logic simpler. Re-introducing support for symbol redefinition is
+       left as an exercise for the reader. (The KaleidoscopeJIT.h used in the
+       original tutorials will be a helpful reference).
+
+.. [2] +-----------------------+-----------------------------------------------+
+       |         File          |               Reason for inclusion            |
+       +=======================+===============================================+
+       |   ExecutionEngine.h   | Access to the EngineBuilder::selectTarget     |
+       |                       | method.                                       |
+       +-----------------------+-----------------------------------------------+
+       |                       | Access to the                                 |
+       | RTDyldMemoryManager.h | RTDyldMemoryManager::getSymbolAddressInProcess|
+       |                       | method.                                       |
+       +-----------------------+-----------------------------------------------+
+       |    CompileUtils.h     | Provides the SimpleCompiler class.            |
+       +-----------------------+-----------------------------------------------+
+       |   IRCompileLayer.h    | Provides the IRCompileLayer class.            |
+       +-----------------------+-----------------------------------------------+
+       |                       | Access the createLambdaResolver function,     |
+       |   LambdaResolver.h    | which provides easy construction of symbol    |
+       |                       | resolvers.                                    |
+       +-----------------------+-----------------------------------------------+
+       |  ObjectLinkingLayer.h | Provides the ObjectLinkingLayer class.        |
+       +-----------------------+-----------------------------------------------+
+       |       Mangler.h       | Provides the Mangler class for platform       |
+       |                       | specific name-mangling.                       |
+       +-----------------------+-----------------------------------------------+
+       |   DynamicLibrary.h    | Provides the DynamicLibrary class, which      |
+       |                       | makes symbols in the host process searchable. |
+       +-----------------------+-----------------------------------------------+
+
+.. [3] Actually they don't have to be lambdas, any object with a call operator
+       will do, including plain old functions or std::functions.
+
+.. [4] ORC layers accept sets of Modules, rather than individual ones, so that
+       all Modules in the set could be co-located by the memory manager, though
+       this feature is not yet implemented.

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT2.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT2.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT2.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT2.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,336 @@
+=====================================================================
+Building a JIT: Adding Optimizations -- An introduction to ORC Layers
+=====================================================================
+
+.. contents::
+   :local:
+
+**This tutorial is under active development. It is incomplete and details may
+change frequently.** Nonetheless we invite you to try it out as it stands, and
+we welcome any feedback.
+
+Chapter 2 Introduction
+======================
+
+Welcome to Chapter 2 of the "Building an ORC-based JIT in LLVM" tutorial. In
+`Chapter 1 <BuildingAJIT1.html>`_ of this series we examined a basic JIT
+class, KaleidoscopeJIT, that could take LLVM IR modules as input and produce
+executable code in memory. KaleidoscopeJIT was able to do this with relatively
+little code by composing two off-the-shelf *ORC layers*: IRCompileLayer and
+ObjectLinkingLayer, to do much of the heavy lifting.
+
+In this layer we'll learn more about the ORC layer concept by using a new layer,
+IRTransformLayer, to add IR optimization support to KaleidoscopeJIT.
+
+Optimizing Modules using the IRTransformLayer
+=============================================
+
+In `Chapter 4 <LangImpl4.html>`_ of the "Implementing a language with LLVM"
+tutorial series the llvm *FunctionPassManager* is introduced as a means for
+optimizing LLVM IR. Interested readers may read that chapter for details, but
+in short: to optimize a Module we create an llvm::FunctionPassManager
+instance, configure it with a set of optimizations, then run the PassManager on
+a Module to mutate it into a (hopefully) more optimized but semantically
+equivalent form. In the original tutorial series the FunctionPassManager was
+created outside the KaleidoscopeJIT and modules were optimized before being
+added to it. In this Chapter we will make optimization a phase of our JIT
+instead. For now this will provide us a motivation to learn more about ORC
+layers, but in the long term making optimization part of our JIT will yield an
+important benefit: When we begin lazily compiling code (i.e. deferring
+compilation of each function until the first time it's run), having
+optimization managed by our JIT will allow us to optimize lazily too, rather
+than having to do all our optimization up-front.
+
+To add optimization support to our JIT we will take the KaleidoscopeJIT from
+Chapter 1 and compose an ORC *IRTransformLayer* on top. We will look at how the
+IRTransformLayer works in more detail below, but the interface is simple: the
+constructor for this layer takes a reference to the layer below (as all layers
+do) plus an *IR optimization function* that it will apply to each Module that
+is added via addModuleSet:
+
+.. code-block:: c++
+
+  class KaleidoscopeJIT {
+  private:
+    std::unique_ptr<TargetMachine> TM;
+    const DataLayout DL;
+    ObjectLinkingLayer<> ObjectLayer;
+    IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
+
+    typedef std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>
+      OptimizeFunction;
+
+    IRTransformLayer<decltype(CompileLayer), OptimizeFunction> OptimizeLayer;
+
+  public:
+    typedef decltype(OptimizeLayer)::ModuleSetHandleT ModuleHandle;
+
+    KaleidoscopeJIT()
+        : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),
+          CompileLayer(ObjectLayer, SimpleCompiler(*TM)),
+          OptimizeLayer(CompileLayer,
+                        [this](std::unique_ptr<Module> M) {
+                          return optimizeModule(std::move(M));
+                        }) {
+      llvm::sys::DynamicLibrary::LoadLibraryPermanently(nullptr);
+    }
+
+Our extended KaleidoscopeJIT class starts out the same as it did in Chapter 1,
+but after the CompileLayer we introduce a typedef for our optimization function.
+In this case we use a std::function (a handy wrapper for "function-like" things)
+from a single unique_ptr<Module> input to a std::unique_ptr<Module> output. With
+our optimization function typedef in place we can declare our OptimizeLayer,
+which sits on top of our CompileLayer.
+
+To initialize our OptimizeLayer we pass it a reference to the CompileLayer
+below (standard practice for layers), and we initialize the OptimizeFunction
+using a lambda that calls out to an "optimizeModule" function that we will
+define below.
+
+.. code-block:: c++
+
+  // ...
+  auto Resolver = createLambdaResolver(
+      [&](const std::string &Name) {
+        if (auto Sym = OptimizeLayer.findSymbol(Name, false))
+          return Sym;
+        return JITSymbol(nullptr);
+      },
+  // ...
+
+.. code-block:: c++
+
+  // ...
+  return OptimizeLayer.addModuleSet(std::move(Ms),
+                                    make_unique<SectionMemoryManager>(),
+                                    std::move(Resolver));
+  // ...
+
+.. code-block:: c++
+
+  // ...
+  return OptimizeLayer.findSymbol(MangledNameStream.str(), true);
+  // ...
+
+.. code-block:: c++
+
+  // ...
+  OptimizeLayer.removeModuleSet(H);
+  // ...
+
+Next we need to replace references to 'CompileLayer' with references to
+OptimizeLayer in our key methods: addModule, findSymbol, and removeModule. In
+addModule we need to be careful to replace both references: the findSymbol call
+inside our resolver, and the call through to addModuleSet.
+
+.. code-block:: c++
+
+  std::unique_ptr<Module> optimizeModule(std::unique_ptr<Module> M) {
+    // Create a function pass manager.
+    auto FPM = llvm::make_unique<legacy::FunctionPassManager>(M.get());
+
+    // Add some optimizations.
+    FPM->add(createInstructionCombiningPass());
+    FPM->add(createReassociatePass());
+    FPM->add(createGVNPass());
+    FPM->add(createCFGSimplificationPass());
+    FPM->doInitialization();
+
+    // Run the optimizations over all functions in the module being added to
+    // the JIT.
+    for (auto &F : *M)
+      FPM->run(F);
+
+    return M;
+  }
+
+At the bottom of our JIT we add a private method to do the actual optimization:
+*optimizeModule*. This function sets up a FunctionPassManager, adds some passes
+to it, runs it over every function in the module, and then returns the mutated
+module. The specific optimizations are the same ones used in
+`Chapter 4 <LangImpl4.html>`_ of the "Implementing a language with LLVM"
+tutorial series. Readers may visit that chapter for a more in-depth
+discussion of these, and of IR optimization in general.
+
+And that's it in terms of changes to KaleidoscopeJIT: When a module is added via
+addModule the OptimizeLayer will call our optimizeModule function before passing
+the transformed module on to the CompileLayer below. Of course, we could have
+called optimizeModule directly in our addModule function and not gone to the
+bother of using the IRTransformLayer, but doing so gives us another opportunity
+to see how layers compose. It also provides a neat entry point to the *layer*
+concept itself, because IRTransformLayer turns out to be one of the simplest
+implementations of the layer concept that can be devised:
+
+.. code-block:: c++
+
+  template <typename BaseLayerT, typename TransformFtor>
+  class IRTransformLayer {
+  public:
+    typedef typename BaseLayerT::ModuleSetHandleT ModuleSetHandleT;
+
+    IRTransformLayer(BaseLayerT &BaseLayer,
+                     TransformFtor Transform = TransformFtor())
+      : BaseLayer(BaseLayer), Transform(std::move(Transform)) {}
+
+    template <typename ModuleSetT, typename MemoryManagerPtrT,
+              typename SymbolResolverPtrT>
+    ModuleSetHandleT addModuleSet(ModuleSetT Ms,
+                                  MemoryManagerPtrT MemMgr,
+                                  SymbolResolverPtrT Resolver) {
+
+      for (auto I = Ms.begin(), E = Ms.end(); I != E; ++I)
+        *I = Transform(std::move(*I));
+
+      return BaseLayer.addModuleSet(std::move(Ms), std::move(MemMgr),
+                                  std::move(Resolver));
+    }
+
+    void removeModuleSet(ModuleSetHandleT H) { BaseLayer.removeModuleSet(H); }
+
+    JITSymbol findSymbol(const std::string &Name, bool ExportedSymbolsOnly) {
+      return BaseLayer.findSymbol(Name, ExportedSymbolsOnly);
+    }
+
+    JITSymbol findSymbolIn(ModuleSetHandleT H, const std::string &Name,
+                           bool ExportedSymbolsOnly) {
+      return BaseLayer.findSymbolIn(H, Name, ExportedSymbolsOnly);
+    }
+
+    void emitAndFinalize(ModuleSetHandleT H) {
+      BaseLayer.emitAndFinalize(H);
+    }
+
+    TransformFtor& getTransform() { return Transform; }
+
+    const TransformFtor& getTransform() const { return Transform; }
+
+  private:
+    BaseLayerT &BaseLayer;
+    TransformFtor Transform;
+  };
+
+This is the whole definition of IRTransformLayer, from
+``llvm/include/llvm/ExecutionEngine/Orc/IRTransformLayer.h``, stripped of its
+comments. It is a template class with two template arguments: ``BaesLayerT`` and
+``TransformFtor`` that provide the type of the base layer and the type of the
+"transform functor" (in our case a std::function) respectively. This class is
+concerned with two very simple jobs: (1) Running every IR Module that is added
+with addModuleSet through the transform functor, and (2) conforming to the ORC
+layer interface. The interface consists of one typedef and five methods:
+
++------------------+-----------------------------------------------------------+
+|     Interface    |                         Description                       |
++==================+===========================================================+
+|                  | Provides a handle that can be used to identify a module   |
+| ModuleSetHandleT | set when calling findSymbolIn, removeModuleSet, or        |
+|                  | emitAndFinalize.                                          |
++------------------+-----------------------------------------------------------+
+|                  | Takes a given set of Modules and makes them "available    |
+|                  | for execution. This means that symbols in those modules   |
+|                  | should be searchable via findSymbol and findSymbolIn, and |
+|                  | the address of the symbols should be read/writable (for   |
+|                  | data symbols), or executable (for function symbols) after |
+|                  | JITSymbol::getAddress() is called. Note: This means that  |
+|   addModuleSet   | addModuleSet doesn't have to compile (or do any other     |
+|                  | work) up-front. It *can*, like IRCompileLayer, act        |
+|                  | eagerly, but it can also simply record the module and     |
+|                  | take no further action until somebody calls               |
+|                  | JITSymbol::getAddress(). In IRTransformLayer's case       |
+|                  | addModuleSet eagerly applies the transform functor to     |
+|                  | each module in the set, then passes the resulting set     |
+|                  | of mutated modules down to the layer below.               |
++------------------+-----------------------------------------------------------+
+|                  | Removes a set of modules from the JIT. Code or data       |
+|  removeModuleSet | defined in these modules will no longer be available, and |
+|                  | the memory holding the JIT'd definitions will be freed.   |
++------------------+-----------------------------------------------------------+
+|                  | Searches for the named symbol in all modules that have    |
+|                  | previously been added via addModuleSet (and not yet       |
+|    findSymbol    | removed by a call to removeModuleSet). In                 |
+|                  | IRTransformLayer we just pass the query on to the layer   |
+|                  | below. In our REPL this is our default way to search for  |
+|                  | function definitions.                                     |
++------------------+-----------------------------------------------------------+
+|                  | Searches for the named symbol in the module set indicated |
+|                  | by the given ModuleSetHandleT. This is just an optimized  |
+|                  | search, better for lookup-speed when you know exactly     |
+|                  | a symbol definition should be found. In IRTransformLayer  |
+|   findSymbolIn   | we just pass this query on to the layer below. In our     |
+|                  | REPL we use this method to search for functions           |
+|                  | representing top-level expressions, since we know exactly |
+|                  | where we'll find them: in the top-level expression module |
+|                  | we just added.                                            |
++------------------+-----------------------------------------------------------+
+|                  | Forces all of the actions required to make the code and   |
+|                  | data in a module set (represented by a ModuleSetHandleT)  |
+|                  | accessible. Behaves as if some symbol in the set had been |
+|                  | searched for and JITSymbol::getSymbolAddress called. This |
+| emitAndFinalize  | is rarely needed, but can be useful when dealing with     |
+|                  | layers that usually behave lazily if the user wants to    |
+|                  | trigger early compilation (for example, to use idle CPU   |
+|                  | time to eagerly compile code in the background).          |
++------------------+-----------------------------------------------------------+
+
+This interface attempts to capture the natural operations of a JIT (with some
+wrinkles like emitAndFinalize for performance), similar to the basic JIT API
+operations we identified in Chapter 1. Conforming to the layer concept allows
+classes to compose neatly by implementing their behaviors in terms of the these
+same operations, carried out on the layer below. For example, an eager layer
+(like IRTransformLayer) can implement addModuleSet by running each module in the
+set through its transform up-front and immediately passing the result to the
+layer below. A lazy layer, by contrast, could implement addModuleSet by
+squirreling away the modules doing no other up-front work, but applying the
+transform (and calling addModuleSet on the layer below) when the client calls
+findSymbol instead. The JIT'd program behavior will be the same either way, but
+these choices will have different performance characteristics: Doing work
+eagerly means the JIT takes longer up-front, but proceeds smoothly once this is
+done. Deferring work allows the JIT to get up-and-running quickly, but will
+force the JIT to pause and wait whenever some code or data is needed that hasn't
+already been processed.
+
+Our current REPL is eager: Each function definition is optimized and compiled as
+soon as it's typed in. If we were to make the transform layer lazy (but not
+change things otherwise) we could defer optimization until the first time we
+reference a function in a top-level expression (see if you can figure out why,
+then check out the answer below [1]_). In the next chapter, however we'll
+introduce fully lazy compilation, in which function's aren't compiled until
+they're first called at run-time. At this point the trade-offs get much more
+interesting: the lazier we are, the quicker we can start executing the first
+function, but the more often we'll have to pause to compile newly encountered
+functions. If we only code-gen lazily, but optimize eagerly, we'll have a slow
+startup (which everything is optimized) but relatively short pauses as each
+function just passes through code-gen. If we both optimize and code-gen lazily
+we can start executing the first function more quickly, but we'll have longer
+pauses as each function has to be both optimized and code-gen'd when it's first
+executed. Things become even more interesting if we consider interproceedural
+optimizations like inlining, which must be performed eagerly. These are
+complex trade-offs, and there is no one-size-fits all solution to them, but by
+providing composable layers we leave the decisions to the person implementing
+the JIT, and make it easy for them to experiment with different configurations.
+
+`Next: Adding Per-function Lazy Compilation <BuildingAJIT3.html>`_
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example with an
+IRTransformLayer added to enable optimization. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core orc native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/BuildingAJIT/Chapter2/KaleidoscopeJIT.h
+   :language: c++
+
+.. [1] When we add our top-level expression to the JIT, any calls to functions
+       that we defined earlier will appear to the ObjectLinkingLayer as
+       external symbols. The ObjectLinkingLayer will call the SymbolResolver
+       that we defined in addModuleSet, which in turn calls findSymbol on the
+       OptimizeLayer, at which point even a lazy transform layer will have to
+       do its work.

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT3.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT3.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT3.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT3.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,185 @@
+=============================================
+Building a JIT: Per-function Lazy Compilation
+=============================================
+
+.. contents::
+   :local:
+
+**This tutorial is under active development. It is incomplete and details may
+change frequently.** Nonetheless we invite you to try it out as it stands, and
+we welcome any feedback.
+
+Chapter 3 Introduction
+======================
+
+Welcome to Chapter 3 of the "Building an ORC-based JIT in LLVM" tutorial. This
+chapter discusses lazy JITing and shows you how to enable it by adding an ORC
+CompileOnDemand layer the JIT from `Chapter 2 <BuildingAJIT2.html>`_.
+
+Lazy Compilation
+================
+
+When we add a module to the KaleidoscopeJIT class from Chapter 2 it is
+immediately optimized, compiled and linked for us by the IRTransformLayer,
+IRCompileLayer and ObjectLinkingLayer respectively. This scheme, where all the
+work to make a Module executable is done up front, is simple to understand and
+its performance characteristics are easy to reason about. However, it will lead
+to very high startup times if the amount of code to be compiled is large, and
+may also do a lot of unnecessary compilation if only a few compiled functions
+are ever called at runtime. A truly "just-in-time" compiler should allow us to
+defer the compilation of any given function until the moment that function is
+first called, improving launch times and eliminating redundant work. In fact,
+the ORC APIs provide us with a layer to lazily compile LLVM IR:
+*CompileOnDemandLayer*.
+
+The CompileOnDemandLayer class conforms to the layer interface described in
+Chapter 2, but its addModuleSet method behaves quite differently from the layers
+we have seen so far: rather than doing any work up front, it just scans the
+Modules being added and arranges for each function in them to be compiled the
+first time it is called. To do this, the CompileOnDemandLayer creates two small
+utilities for each function that it scans: a *stub* and a *compile
+callback*. The stub is a pair of a function pointer (which will be pointed at
+the function's implementation once the function has been compiled) and an
+indirect jump through the pointer. By fixing the address of the indirect jump
+for the lifetime of the program we can give the function a permanent "effective
+address", one that can be safely used for indirection and function pointer
+comparison even if the function's implementation is never compiled, or if it is
+compiled more than once (due to, for example, recompiling the function at a
+higher optimization level) and changes address. The second utility, the compile
+callback, represents a re-entry point from the program into the compiler that
+will trigger compilation and then execution of a function. By initializing the
+function's stub to point at the function's compile callback, we enable lazy
+compilation: The first attempted call to the function will follow the function
+pointer and trigger the compile callback instead. The compile callback will
+compile the function, update the function pointer for the stub, then execute
+the function. On all subsequent calls to the function, the function pointer
+will point at the already-compiled function, so there is no further overhead
+from the compiler. We will look at this process in more detail in the next
+chapter of this tutorial, but for now we'll trust the CompileOnDemandLayer to
+set all the stubs and callbacks up for us. All we need to do is to add the
+CompileOnDemandLayer to the top of our stack and we'll get the benefits of
+lazy compilation. We just need a few changes to the source:
+
+.. code-block:: c++
+
+  ...
+  #include "llvm/ExecutionEngine/SectionMemoryManager.h"
+  #include "llvm/ExecutionEngine/Orc/CompileOnDemandLayer.h"
+  #include "llvm/ExecutionEngine/Orc/CompileUtils.h"
+  ...
+
+  ...
+  class KaleidoscopeJIT {
+  private:
+    std::unique_ptr<TargetMachine> TM;
+    const DataLayout DL;
+    std::unique_ptr<JITCompileCallbackManager> CompileCallbackManager;
+    ObjectLinkingLayer<> ObjectLayer;
+    IRCompileLayer<decltype(ObjectLayer)> CompileLayer;
+
+    typedef std::function<std::unique_ptr<Module>(std::unique_ptr<Module>)>
+      OptimizeFunction;
+
+    IRTransformLayer<decltype(CompileLayer), OptimizeFunction> OptimizeLayer;
+    CompileOnDemandLayer<decltype(OptimizeLayer)> CODLayer;
+
+  public:
+    typedef decltype(CODLayer)::ModuleSetHandleT ModuleHandle;
+
+First we need to include the CompileOnDemandLayer.h header, then add two new
+members: a std::unique_ptr<CompileCallbackManager> and a CompileOnDemandLayer,
+to our class. The CompileCallbackManager member is used by the CompileOnDemandLayer
+to create the compile callback needed for each function.
+
+.. code-block:: c++
+
+  KaleidoscopeJIT()
+      : TM(EngineBuilder().selectTarget()), DL(TM->createDataLayout()),
+        CompileLayer(ObjectLayer, SimpleCompiler(*TM)),
+        OptimizeLayer(CompileLayer,
+                      [this](std::unique_ptr<Module> M) {
+                        return optimizeModule(std::move(M));
+                      }),
+        CompileCallbackManager(
+            orc::createLocalCompileCallbackManager(TM->getTargetTriple(), 0)),
+        CODLayer(OptimizeLayer,
+                 [this](Function &F) { return std::set<Function*>({&F}); },
+                 *CompileCallbackManager,
+                 orc::createLocalIndirectStubsManagerBuilder(
+                   TM->getTargetTriple())) {
+    llvm::sys::DynamicLibrary::LoadLibraryPermanently(nullptr);
+  }
+
+Next we have to update our constructor to initialize the new members. To create
+an appropriate compile callback manager we use the
+createLocalCompileCallbackManager function, which takes a TargetMachine and a
+JITTargetAddress to call if it receives a request to compile an unknown
+function.  In our simple JIT this situation is unlikely to come up, so we'll
+cheat and just pass '0' here. In a production quality JIT you could give the
+address of a function that throws an exception in order to unwind the JIT'd
+code's stack.
+
+Now we can construct our CompileOnDemandLayer. Following the pattern from
+previous layers we start by passing a reference to the next layer down in our
+stack -- the OptimizeLayer. Next we need to supply a 'partitioning function':
+when a not-yet-compiled function is called, the CompileOnDemandLayer will call
+this function to ask us what we would like to compile. At a minimum we need to
+compile the function being called (given by the argument to the partitioning
+function), but we could also request that the CompileOnDemandLayer compile other
+functions that are unconditionally called (or highly likely to be called) from
+the function being called. For KaleidoscopeJIT we'll keep it simple and just
+request compilation of the function that was called. Next we pass a reference to
+our CompileCallbackManager. Finally, we need to supply an "indirect stubs
+manager builder": a utility function that constructs IndirectStubManagers, which
+are in turn used to build the stubs for the functions in each module. The
+CompileOnDemandLayer will call the indirect stub manager builder once for each
+call to addModuleSet, and use the resulting indirect stubs manager to create
+stubs for all functions in all modules in the set. If/when the module set is
+removed from the JIT the indirect stubs manager will be deleted, freeing any
+memory allocated to the stubs. We supply this function by using the
+createLocalIndirectStubsManagerBuilder utility.
+
+.. code-block:: c++
+
+  // ...
+          if (auto Sym = CODLayer.findSymbol(Name, false))
+  // ...
+  return CODLayer.addModuleSet(std::move(Ms),
+                               make_unique<SectionMemoryManager>(),
+                               std::move(Resolver));
+  // ...
+
+  // ...
+  return CODLayer.findSymbol(MangledNameStream.str(), true);
+  // ...
+
+  // ...
+  CODLayer.removeModuleSet(H);
+  // ...
+
+Finally, we need to replace the references to OptimizeLayer in our addModule,
+findSymbol, and removeModule methods. With that, we're up and running.
+
+**To be done:**
+
+** Chapter conclusion.**
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example with a CompileOnDemand
+layer added to enable lazy function-at-a-time compilation. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core orc native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/BuildingAJIT/Chapter3/KaleidoscopeJIT.h
+   :language: c++
+
+`Next: Extreme Laziness -- Using Compile Callbacks to JIT directly from ASTs <BuildingAJIT4.html>`_

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT4.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT4.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT4.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT4.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,48 @@
+===========================================================================
+Building a JIT: Extreme Laziness - Using Compile Callbacks to JIT from ASTs
+===========================================================================
+
+.. contents::
+   :local:
+
+**This tutorial is under active development. It is incomplete and details may
+change frequently.** Nonetheless we invite you to try it out as it stands, and
+we welcome any feedback.
+
+Chapter 4 Introduction
+======================
+
+Welcome to Chapter 4 of the "Building an ORC-based JIT in LLVM" tutorial. This
+chapter introduces the Compile Callbacks and Indirect Stubs APIs and shows how
+they can be used to replace the CompileOnDemand layer from
+`Chapter 3 <BuildingAJIT3.html>`_ with a custom lazy-JITing scheme that JITs
+directly from Kaleidoscope ASTs.
+
+**To be done:**
+
+**(1) Describe the drawbacks of JITing from IR (have to compile to IR first,
+which reduces the benefits of laziness).**
+
+**(2) Describe CompileCallbackManagers and IndirectStubManagers in detail.**
+
+**(3) Run through the implementation of addFunctionAST.**
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example that JITs lazily from
+Kaleidoscope ASTS. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core orc native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/BuildingAJIT/Chapter4/KaleidoscopeJIT.h
+   :language: c++
+
+`Next: Remote-JITing -- Process-isolation and laziness-at-a-distance <BuildingAJIT5.html>`_

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT5.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT5.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT5.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/BuildingAJIT5.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,55 @@
+=============================================================================
+Building a JIT: Remote-JITing -- Process Isolation and Laziness at a Distance
+=============================================================================
+
+.. contents::
+   :local:
+
+**This tutorial is under active development. It is incomplete and details may
+change frequently.** Nonetheless we invite you to try it out as it stands, and
+we welcome any feedback.
+
+Chapter 5 Introduction
+======================
+
+Welcome to Chapter 5 of the "Building an ORC-based JIT in LLVM" tutorial. This
+chapter introduces the ORC RemoteJIT Client/Server APIs and shows how to use
+them to build a JIT stack that will execute its code via a communications
+channel with a different process. This can be a separate process on the same
+machine, a process on a different machine, or even a process on a different
+platform/architecture. The code builds on top of the lazy-AST-compiling JIT
+stack from `Chapter 4 <BuildingAJIT3.html>`_.
+
+**To be done -- this is going to be a long one:**
+
+**(1) Introduce channels, RPC, RemoteJIT Client and Server APIs**
+
+**(2) Describe the client code in greater detail. Discuss modifications of the
+KaleidoscopeJIT class, and the REPL itself.**
+
+**(3) Describe the server code.**
+
+**(4) Describe how to run the demo.**
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example that JITs lazily from
+Kaleidoscope ASTS. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core orc native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code for the modified KaleidoscopeJIT:
+
+.. literalinclude:: ../../examples/Kaleidoscope/BuildingAJIT/Chapter5/KaleidoscopeJIT.h
+   :language: c++
+
+And the code for the JIT server:
+
+.. literalinclude:: ../../examples/Kaleidoscope/BuildingAJIT/Chapter5/Server/server.cpp
+   :language: c++

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl01.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl01.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl01.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl01.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,293 @@
+=================================================
+Kaleidoscope: Tutorial Introduction and the Lexer
+=================================================
+
+.. contents::
+   :local:
+
+Tutorial Introduction
+=====================
+
+Welcome to the "Implementing a language with LLVM" tutorial. This
+tutorial runs through the implementation of a simple language, showing
+how fun and easy it can be. This tutorial will get you up and started as
+well as help to build a framework you can extend to other languages. The
+code in this tutorial can also be used as a playground to hack on other
+LLVM specific things.
+
+The goal of this tutorial is to progressively unveil our language,
+describing how it is built up over time. This will let us cover a fairly
+broad range of language design and LLVM-specific usage issues, showing
+and explaining the code for it all along the way, without overwhelming
+you with tons of details up front.
+
+It is useful to point out ahead of time that this tutorial is really
+about teaching compiler techniques and LLVM specifically, *not* about
+teaching modern and sane software engineering principles. In practice,
+this means that we'll take a number of shortcuts to simplify the
+exposition. For example, the code uses global variables
+all over the place, doesn't use nice design patterns like
+`visitors <http://en.wikipedia.org/wiki/Visitor_pattern>`_, etc... but
+it is very simple. If you dig in and use the code as a basis for future
+projects, fixing these deficiencies shouldn't be hard.
+
+I've tried to put this tutorial together in a way that makes chapters
+easy to skip over if you are already familiar with or are uninterested
+in the various pieces. The structure of the tutorial is:
+
+-  `Chapter #1 <#language>`_: Introduction to the Kaleidoscope
+   language, and the definition of its Lexer - This shows where we are
+   going and the basic functionality that we want it to do. In order to
+   make this tutorial maximally understandable and hackable, we choose
+   to implement everything in C++ instead of using lexer and parser
+   generators. LLVM obviously works just fine with such tools, feel free
+   to use one if you prefer.
+-  `Chapter #2 <LangImpl02.html>`_: Implementing a Parser and AST -
+   With the lexer in place, we can talk about parsing techniques and
+   basic AST construction. This tutorial describes recursive descent
+   parsing and operator precedence parsing. Nothing in Chapters 1 or 2
+   is LLVM-specific, the code doesn't even link in LLVM at this point.
+   :)
+-  `Chapter #3 <LangImpl03.html>`_: Code generation to LLVM IR - With
+   the AST ready, we can show off how easy generation of LLVM IR really
+   is.
+-  `Chapter #4 <LangImpl04.html>`_: Adding JIT and Optimizer Support
+   - Because a lot of people are interested in using LLVM as a JIT,
+   we'll dive right into it and show you the 3 lines it takes to add JIT
+   support. LLVM is also useful in many other ways, but this is one
+   simple and "sexy" way to show off its power. :)
+-  `Chapter #5 <LangImpl05.html>`_: Extending the Language: Control
+   Flow - With the language up and running, we show how to extend it
+   with control flow operations (if/then/else and a 'for' loop). This
+   gives us a chance to talk about simple SSA construction and control
+   flow.
+-  `Chapter #6 <LangImpl06.html>`_: Extending the Language:
+   User-defined Operators - This is a silly but fun chapter that talks
+   about extending the language to let the user program define their own
+   arbitrary unary and binary operators (with assignable precedence!).
+   This lets us build a significant piece of the "language" as library
+   routines.
+-  `Chapter #7 <LangImpl07.html>`_: Extending the Language: Mutable
+   Variables - This chapter talks about adding user-defined local
+   variables along with an assignment operator. The interesting part
+   about this is how easy and trivial it is to construct SSA form in
+   LLVM: no, LLVM does *not* require your front-end to construct SSA
+   form!
+-  `Chapter #8 <LangImpl08.html>`_: Compiling to Object Files - This
+   chapter explains how to take LLVM IR and compile it down to object
+   files.
+-  `Chapter #9 <LangImpl09.html>`_: Extending the Language: Debug
+   Information - Having built a decent little programming language with
+   control flow, functions and mutable variables, we consider what it
+   takes to add debug information to standalone executables. This debug
+   information will allow you to set breakpoints in Kaleidoscope
+   functions, print out argument variables, and call functions - all
+   from within the debugger!
+-  `Chapter #10 <LangImpl10.html>`_: Conclusion and other useful LLVM
+   tidbits - This chapter wraps up the series by talking about
+   potential ways to extend the language, but also includes a bunch of
+   pointers to info about "special topics" like adding garbage
+   collection support, exceptions, debugging, support for "spaghetti
+   stacks", and a bunch of other tips and tricks.
+
+By the end of the tutorial, we'll have written a bit less than 1000 lines
+of non-comment, non-blank, lines of code. With this small amount of
+code, we'll have built up a very reasonable compiler for a non-trivial
+language including a hand-written lexer, parser, AST, as well as code
+generation support with a JIT compiler. While other systems may have
+interesting "hello world" tutorials, I think the breadth of this
+tutorial is a great testament to the strengths of LLVM and why you
+should consider it if you're interested in language or compiler design.
+
+A note about this tutorial: we expect you to extend the language and
+play with it on your own. Take the code and go crazy hacking away at it,
+compilers don't need to be scary creatures - it can be a lot of fun to
+play with languages!
+
+The Basic Language
+==================
+
+This tutorial will be illustrated with a toy language that we'll call
+"`Kaleidoscope <http://en.wikipedia.org/wiki/Kaleidoscope>`_" (derived
+from "meaning beautiful, form, and view"). Kaleidoscope is a procedural
+language that allows you to define functions, use conditionals, math,
+etc. Over the course of the tutorial, we'll extend Kaleidoscope to
+support the if/then/else construct, a for loop, user defined operators,
+JIT compilation with a simple command line interface, etc.
+
+Because we want to keep things simple, the only datatype in Kaleidoscope
+is a 64-bit floating point type (aka 'double' in C parlance). As such,
+all values are implicitly double precision and the language doesn't
+require type declarations. This gives the language a very nice and
+simple syntax. For example, the following simple example computes
+`Fibonacci numbers: <http://en.wikipedia.org/wiki/Fibonacci_number>`_
+
+::
+
+    # Compute the x'th fibonacci number.
+    def fib(x)
+      if x < 3 then
+        1
+      else
+        fib(x-1)+fib(x-2)
+
+    # This expression will compute the 40th number.
+    fib(40)
+
+We also allow Kaleidoscope to call into standard library functions (the
+LLVM JIT makes this completely trivial). This means that you can use the
+'extern' keyword to define a function before you use it (this is also
+useful for mutually recursive functions). For example:
+
+::
+
+    extern sin(arg);
+    extern cos(arg);
+    extern atan2(arg1 arg2);
+
+    atan2(sin(.4), cos(42))
+
+A more interesting example is included in Chapter 6 where we write a
+little Kaleidoscope application that `displays a Mandelbrot
+Set <LangImpl06.html#kicking-the-tires>`_ at various levels of magnification.
+
+Lets dive into the implementation of this language!
+
+The Lexer
+=========
+
+When it comes to implementing a language, the first thing needed is the
+ability to process a text file and recognize what it says. The
+traditional way to do this is to use a
+"`lexer <http://en.wikipedia.org/wiki/Lexical_analysis>`_" (aka
+'scanner') to break the input up into "tokens". Each token returned by
+the lexer includes a token code and potentially some metadata (e.g. the
+numeric value of a number). First, we define the possibilities:
+
+.. code-block:: c++
+
+    // The lexer returns tokens [0-255] if it is an unknown character, otherwise one
+    // of these for known things.
+    enum Token {
+      tok_eof = -1,
+
+      // commands
+      tok_def = -2,
+      tok_extern = -3,
+
+      // primary
+      tok_identifier = -4,
+      tok_number = -5,
+    };
+
+    static std::string IdentifierStr; // Filled in if tok_identifier
+    static double NumVal;             // Filled in if tok_number
+
+Each token returned by our lexer will either be one of the Token enum
+values or it will be an 'unknown' character like '+', which is returned
+as its ASCII value. If the current token is an identifier, the
+``IdentifierStr`` global variable holds the name of the identifier. If
+the current token is a numeric literal (like 1.0), ``NumVal`` holds its
+value. Note that we use global variables for simplicity, this is not the
+best choice for a real language implementation :).
+
+The actual implementation of the lexer is a single function named
+``gettok``. The ``gettok`` function is called to return the next token
+from standard input. Its definition starts as:
+
+.. code-block:: c++
+
+    /// gettok - Return the next token from standard input.
+    static int gettok() {
+      static int LastChar = ' ';
+
+      // Skip any whitespace.
+      while (isspace(LastChar))
+        LastChar = getchar();
+
+``gettok`` works by calling the C ``getchar()`` function to read
+characters one at a time from standard input. It eats them as it
+recognizes them and stores the last character read, but not processed,
+in LastChar. The first thing that it has to do is ignore whitespace
+between tokens. This is accomplished with the loop above.
+
+The next thing ``gettok`` needs to do is recognize identifiers and
+specific keywords like "def". Kaleidoscope does this with this simple
+loop:
+
+.. code-block:: c++
+
+      if (isalpha(LastChar)) { // identifier: [a-zA-Z][a-zA-Z0-9]*
+        IdentifierStr = LastChar;
+        while (isalnum((LastChar = getchar())))
+          IdentifierStr += LastChar;
+
+        if (IdentifierStr == "def")
+          return tok_def;
+        if (IdentifierStr == "extern")
+          return tok_extern;
+        return tok_identifier;
+      }
+
+Note that this code sets the '``IdentifierStr``' global whenever it
+lexes an identifier. Also, since language keywords are matched by the
+same loop, we handle them here inline. Numeric values are similar:
+
+.. code-block:: c++
+
+      if (isdigit(LastChar) || LastChar == '.') {   // Number: [0-9.]+
+        std::string NumStr;
+        do {
+          NumStr += LastChar;
+          LastChar = getchar();
+        } while (isdigit(LastChar) || LastChar == '.');
+
+        NumVal = strtod(NumStr.c_str(), 0);
+        return tok_number;
+      }
+
+This is all pretty straight-forward code for processing input. When
+reading a numeric value from input, we use the C ``strtod`` function to
+convert it to a numeric value that we store in ``NumVal``. Note that
+this isn't doing sufficient error checking: it will incorrectly read
+"1.23.45.67" and handle it as if you typed in "1.23". Feel free to
+extend it :). Next we handle comments:
+
+.. code-block:: c++
+
+      if (LastChar == '#') {
+        // Comment until end of line.
+        do
+          LastChar = getchar();
+        while (LastChar != EOF && LastChar != '\n' && LastChar != '\r');
+
+        if (LastChar != EOF)
+          return gettok();
+      }
+
+We handle comments by skipping to the end of the line and then return
+the next token. Finally, if the input doesn't match one of the above
+cases, it is either an operator character like '+' or the end of the
+file. These are handled with this code:
+
+.. code-block:: c++
+
+      // Check for end of file.  Don't eat the EOF.
+      if (LastChar == EOF)
+        return tok_eof;
+
+      // Otherwise, just return the character as its ascii value.
+      int ThisChar = LastChar;
+      LastChar = getchar();
+      return ThisChar;
+    }
+
+With this, we have the complete lexer for the basic Kaleidoscope
+language (the `full code listing <LangImpl02.html#full-code-listing>`_ for the Lexer
+is available in the `next chapter <LangImpl02.html>`_ of the tutorial).
+Next we'll `build a simple parser that uses this to build an Abstract
+Syntax Tree <LangImpl02.html>`_. When we have that, we'll include a
+driver so that you can use the lexer and parser together.
+
+`Next: Implementing a Parser and AST <LangImpl02.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl02.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl02.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl02.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl02.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,735 @@
+===========================================
+Kaleidoscope: Implementing a Parser and AST
+===========================================
+
+.. contents::
+   :local:
+
+Chapter 2 Introduction
+======================
+
+Welcome to Chapter 2 of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. This chapter shows you how to use the
+lexer, built in `Chapter 1 <LangImpl1.html>`_, to build a full
+`parser <http://en.wikipedia.org/wiki/Parsing>`_ for our Kaleidoscope
+language. Once we have a parser, we'll define and build an `Abstract
+Syntax Tree <http://en.wikipedia.org/wiki/Abstract_syntax_tree>`_ (AST).
+
+The parser we will build uses a combination of `Recursive Descent
+Parsing <http://en.wikipedia.org/wiki/Recursive_descent_parser>`_ and
+`Operator-Precedence
+Parsing <http://en.wikipedia.org/wiki/Operator-precedence_parser>`_ to
+parse the Kaleidoscope language (the latter for binary expressions and
+the former for everything else). Before we get to parsing though, lets
+talk about the output of the parser: the Abstract Syntax Tree.
+
+The Abstract Syntax Tree (AST)
+==============================
+
+The AST for a program captures its behavior in such a way that it is
+easy for later stages of the compiler (e.g. code generation) to
+interpret. We basically want one object for each construct in the
+language, and the AST should closely model the language. In
+Kaleidoscope, we have expressions, a prototype, and a function object.
+We'll start with expressions first:
+
+.. code-block:: c++
+
+    /// ExprAST - Base class for all expression nodes.
+    class ExprAST {
+    public:
+      virtual ~ExprAST() {}
+    };
+
+    /// NumberExprAST - Expression class for numeric literals like "1.0".
+    class NumberExprAST : public ExprAST {
+      double Val;
+
+    public:
+      NumberExprAST(double Val) : Val(Val) {}
+    };
+
+The code above shows the definition of the base ExprAST class and one
+subclass which we use for numeric literals. The important thing to note
+about this code is that the NumberExprAST class captures the numeric
+value of the literal as an instance variable. This allows later phases
+of the compiler to know what the stored numeric value is.
+
+Right now we only create the AST, so there are no useful accessor
+methods on them. It would be very easy to add a virtual method to pretty
+print the code, for example. Here are the other expression AST node
+definitions that we'll use in the basic form of the Kaleidoscope
+language:
+
+.. code-block:: c++
+
+    /// VariableExprAST - Expression class for referencing a variable, like "a".
+    class VariableExprAST : public ExprAST {
+      std::string Name;
+
+    public:
+      VariableExprAST(const std::string &Name) : Name(Name) {}
+    };
+
+    /// BinaryExprAST - Expression class for a binary operator.
+    class BinaryExprAST : public ExprAST {
+      char Op;
+      std::unique_ptr<ExprAST> LHS, RHS;
+
+    public:
+      BinaryExprAST(char op, std::unique_ptr<ExprAST> LHS,
+                    std::unique_ptr<ExprAST> RHS)
+        : Op(op), LHS(std::move(LHS)), RHS(std::move(RHS)) {}
+    };
+
+    /// CallExprAST - Expression class for function calls.
+    class CallExprAST : public ExprAST {
+      std::string Callee;
+      std::vector<std::unique_ptr<ExprAST>> Args;
+
+    public:
+      CallExprAST(const std::string &Callee,
+                  std::vector<std::unique_ptr<ExprAST>> Args)
+        : Callee(Callee), Args(std::move(Args)) {}
+    };
+
+This is all (intentionally) rather straight-forward: variables capture
+the variable name, binary operators capture their opcode (e.g. '+'), and
+calls capture a function name as well as a list of any argument
+expressions. One thing that is nice about our AST is that it captures
+the language features without talking about the syntax of the language.
+Note that there is no discussion about precedence of binary operators,
+lexical structure, etc.
+
+For our basic language, these are all of the expression nodes we'll
+define. Because it doesn't have conditional control flow, it isn't
+Turing-complete; we'll fix that in a later installment. The two things
+we need next are a way to talk about the interface to a function, and a
+way to talk about functions themselves:
+
+.. code-block:: c++
+
+    /// PrototypeAST - This class represents the "prototype" for a function,
+    /// which captures its name, and its argument names (thus implicitly the number
+    /// of arguments the function takes).
+    class PrototypeAST {
+      std::string Name;
+      std::vector<std::string> Args;
+
+    public:
+      PrototypeAST(const std::string &name, std::vector<std::string> Args)
+        : Name(name), Args(std::move(Args)) {}
+    };
+
+    /// FunctionAST - This class represents a function definition itself.
+    class FunctionAST {
+      std::unique_ptr<PrototypeAST> Proto;
+      std::unique_ptr<ExprAST> Body;
+
+    public:
+      FunctionAST(std::unique_ptr<PrototypeAST> Proto,
+                  std::unique_ptr<ExprAST> Body)
+        : Proto(std::move(Proto)), Body(std::move(Body)) {}
+    };
+
+In Kaleidoscope, functions are typed with just a count of their
+arguments. Since all values are double precision floating point, the
+type of each argument doesn't need to be stored anywhere. In a more
+aggressive and realistic language, the "ExprAST" class would probably
+have a type field.
+
+With this scaffolding, we can now talk about parsing expressions and
+function bodies in Kaleidoscope.
+
+Parser Basics
+=============
+
+Now that we have an AST to build, we need to define the parser code to
+build it. The idea here is that we want to parse something like "x+y"
+(which is returned as three tokens by the lexer) into an AST that could
+be generated with calls like this:
+
+.. code-block:: c++
+
+      auto LHS = llvm::make_unique<VariableExprAST>("x");
+      auto RHS = llvm::make_unique<VariableExprAST>("y");
+      auto Result = std::make_unique<BinaryExprAST>('+', std::move(LHS),
+                                                    std::move(RHS));
+
+In order to do this, we'll start by defining some basic helper routines:
+
+.. code-block:: c++
+
+    /// CurTok/getNextToken - Provide a simple token buffer.  CurTok is the current
+    /// token the parser is looking at.  getNextToken reads another token from the
+    /// lexer and updates CurTok with its results.
+    static int CurTok;
+    static int getNextToken() {
+      return CurTok = gettok();
+    }
+
+This implements a simple token buffer around the lexer. This allows us
+to look one token ahead at what the lexer is returning. Every function
+in our parser will assume that CurTok is the current token that needs to
+be parsed.
+
+.. code-block:: c++
+
+
+    /// LogError* - These are little helper functions for error handling.
+    std::unique_ptr<ExprAST> LogError(const char *Str) {
+      fprintf(stderr, "LogError: %s\n", Str);
+      return nullptr;
+    }
+    std::unique_ptr<PrototypeAST> LogErrorP(const char *Str) {
+      LogError(Str);
+      return nullptr;
+    }
+
+The ``LogError`` routines are simple helper routines that our parser will
+use to handle errors. The error recovery in our parser will not be the
+best and is not particular user-friendly, but it will be enough for our
+tutorial. These routines make it easier to handle errors in routines
+that have various return types: they always return null.
+
+With these basic helper functions, we can implement the first piece of
+our grammar: numeric literals.
+
+Basic Expression Parsing
+========================
+
+We start with numeric literals, because they are the simplest to
+process. For each production in our grammar, we'll define a function
+which parses that production. For numeric literals, we have:
+
+.. code-block:: c++
+
+    /// numberexpr ::= number
+    static std::unique_ptr<ExprAST> ParseNumberExpr() {
+      auto Result = llvm::make_unique<NumberExprAST>(NumVal);
+      getNextToken(); // consume the number
+      return std::move(Result);
+    }
+
+This routine is very simple: it expects to be called when the current
+token is a ``tok_number`` token. It takes the current number value,
+creates a ``NumberExprAST`` node, advances the lexer to the next token,
+and finally returns.
+
+There are some interesting aspects to this. The most important one is
+that this routine eats all of the tokens that correspond to the
+production and returns the lexer buffer with the next token (which is
+not part of the grammar production) ready to go. This is a fairly
+standard way to go for recursive descent parsers. For a better example,
+the parenthesis operator is defined like this:
+
+.. code-block:: c++
+
+    /// parenexpr ::= '(' expression ')'
+    static std::unique_ptr<ExprAST> ParseParenExpr() {
+      getNextToken(); // eat (.
+      auto V = ParseExpression();
+      if (!V)
+        return nullptr;
+
+      if (CurTok != ')')
+        return LogError("expected ')'");
+      getNextToken(); // eat ).
+      return V;
+    }
+
+This function illustrates a number of interesting things about the
+parser:
+
+1) It shows how we use the LogError routines. When called, this function
+expects that the current token is a '(' token, but after parsing the
+subexpression, it is possible that there is no ')' waiting. For example,
+if the user types in "(4 x" instead of "(4)", the parser should emit an
+error. Because errors can occur, the parser needs a way to indicate that
+they happened: in our parser, we return null on an error.
+
+2) Another interesting aspect of this function is that it uses recursion
+by calling ``ParseExpression`` (we will soon see that
+``ParseExpression`` can call ``ParseParenExpr``). This is powerful
+because it allows us to handle recursive grammars, and keeps each
+production very simple. Note that parentheses do not cause construction
+of AST nodes themselves. While we could do it this way, the most
+important role of parentheses are to guide the parser and provide
+grouping. Once the parser constructs the AST, parentheses are not
+needed.
+
+The next simple production is for handling variable references and
+function calls:
+
+.. code-block:: c++
+
+    /// identifierexpr
+    ///   ::= identifier
+    ///   ::= identifier '(' expression* ')'
+    static std::unique_ptr<ExprAST> ParseIdentifierExpr() {
+      std::string IdName = IdentifierStr;
+
+      getNextToken();  // eat identifier.
+
+      if (CurTok != '(') // Simple variable ref.
+        return llvm::make_unique<VariableExprAST>(IdName);
+
+      // Call.
+      getNextToken();  // eat (
+      std::vector<std::unique_ptr<ExprAST>> Args;
+      if (CurTok != ')') {
+        while (1) {
+          if (auto Arg = ParseExpression())
+            Args.push_back(std::move(Arg));
+          else
+            return nullptr;
+
+          if (CurTok == ')')
+            break;
+
+          if (CurTok != ',')
+            return LogError("Expected ')' or ',' in argument list");
+          getNextToken();
+        }
+      }
+
+      // Eat the ')'.
+      getNextToken();
+
+      return llvm::make_unique<CallExprAST>(IdName, std::move(Args));
+    }
+
+This routine follows the same style as the other routines. (It expects
+to be called if the current token is a ``tok_identifier`` token). It
+also has recursion and error handling. One interesting aspect of this is
+that it uses *look-ahead* to determine if the current identifier is a
+stand alone variable reference or if it is a function call expression.
+It handles this by checking to see if the token after the identifier is
+a '(' token, constructing either a ``VariableExprAST`` or
+``CallExprAST`` node as appropriate.
+
+Now that we have all of our simple expression-parsing logic in place, we
+can define a helper function to wrap it together into one entry point.
+We call this class of expressions "primary" expressions, for reasons
+that will become more clear `later in the
+tutorial <LangImpl6.html#user-defined-unary-operators>`_. In order to parse an arbitrary
+primary expression, we need to determine what sort of expression it is:
+
+.. code-block:: c++
+
+    /// primary
+    ///   ::= identifierexpr
+    ///   ::= numberexpr
+    ///   ::= parenexpr
+    static std::unique_ptr<ExprAST> ParsePrimary() {
+      switch (CurTok) {
+      default:
+        return LogError("unknown token when expecting an expression");
+      case tok_identifier:
+        return ParseIdentifierExpr();
+      case tok_number:
+        return ParseNumberExpr();
+      case '(':
+        return ParseParenExpr();
+      }
+    }
+
+Now that you see the definition of this function, it is more obvious why
+we can assume the state of CurTok in the various functions. This uses
+look-ahead to determine which sort of expression is being inspected, and
+then parses it with a function call.
+
+Now that basic expressions are handled, we need to handle binary
+expressions. They are a bit more complex.
+
+Binary Expression Parsing
+=========================
+
+Binary expressions are significantly harder to parse because they are
+often ambiguous. For example, when given the string "x+y\*z", the parser
+can choose to parse it as either "(x+y)\*z" or "x+(y\*z)". With common
+definitions from mathematics, we expect the later parse, because "\*"
+(multiplication) has higher *precedence* than "+" (addition).
+
+There are many ways to handle this, but an elegant and efficient way is
+to use `Operator-Precedence
+Parsing <http://en.wikipedia.org/wiki/Operator-precedence_parser>`_.
+This parsing technique uses the precedence of binary operators to guide
+recursion. To start with, we need a table of precedences:
+
+.. code-block:: c++
+
+    /// BinopPrecedence - This holds the precedence for each binary operator that is
+    /// defined.
+    static std::map<char, int> BinopPrecedence;
+
+    /// GetTokPrecedence - Get the precedence of the pending binary operator token.
+    static int GetTokPrecedence() {
+      if (!isascii(CurTok))
+        return -1;
+
+      // Make sure it's a declared binop.
+      int TokPrec = BinopPrecedence[CurTok];
+      if (TokPrec <= 0) return -1;
+      return TokPrec;
+    }
+
+    int main() {
+      // Install standard binary operators.
+      // 1 is lowest precedence.
+      BinopPrecedence['<'] = 10;
+      BinopPrecedence['+'] = 20;
+      BinopPrecedence['-'] = 20;
+      BinopPrecedence['*'] = 40;  // highest.
+      ...
+    }
+
+For the basic form of Kaleidoscope, we will only support 4 binary
+operators (this can obviously be extended by you, our brave and intrepid
+reader). The ``GetTokPrecedence`` function returns the precedence for
+the current token, or -1 if the token is not a binary operator. Having a
+map makes it easy to add new operators and makes it clear that the
+algorithm doesn't depend on the specific operators involved, but it
+would be easy enough to eliminate the map and do the comparisons in the
+``GetTokPrecedence`` function. (Or just use a fixed-size array).
+
+With the helper above defined, we can now start parsing binary
+expressions. The basic idea of operator precedence parsing is to break
+down an expression with potentially ambiguous binary operators into
+pieces. Consider, for example, the expression "a+b+(c+d)\*e\*f+g".
+Operator precedence parsing considers this as a stream of primary
+expressions separated by binary operators. As such, it will first parse
+the leading primary expression "a", then it will see the pairs [+, b]
+[+, (c+d)] [\*, e] [\*, f] and [+, g]. Note that because parentheses are
+primary expressions, the binary expression parser doesn't need to worry
+about nested subexpressions like (c+d) at all.
+
+To start, an expression is a primary expression potentially followed by
+a sequence of [binop,primaryexpr] pairs:
+
+.. code-block:: c++
+
+    /// expression
+    ///   ::= primary binoprhs
+    ///
+    static std::unique_ptr<ExprAST> ParseExpression() {
+      auto LHS = ParsePrimary();
+      if (!LHS)
+        return nullptr;
+
+      return ParseBinOpRHS(0, std::move(LHS));
+    }
+
+``ParseBinOpRHS`` is the function that parses the sequence of pairs for
+us. It takes a precedence and a pointer to an expression for the part
+that has been parsed so far. Note that "x" is a perfectly valid
+expression: As such, "binoprhs" is allowed to be empty, in which case it
+returns the expression that is passed into it. In our example above, the
+code passes the expression for "a" into ``ParseBinOpRHS`` and the
+current token is "+".
+
+The precedence value passed into ``ParseBinOpRHS`` indicates the
+*minimal operator precedence* that the function is allowed to eat. For
+example, if the current pair stream is [+, x] and ``ParseBinOpRHS`` is
+passed in a precedence of 40, it will not consume any tokens (because
+the precedence of '+' is only 20). With this in mind, ``ParseBinOpRHS``
+starts with:
+
+.. code-block:: c++
+
+    /// binoprhs
+    ///   ::= ('+' primary)*
+    static std::unique_ptr<ExprAST> ParseBinOpRHS(int ExprPrec,
+                                                  std::unique_ptr<ExprAST> LHS) {
+      // If this is a binop, find its precedence.
+      while (1) {
+        int TokPrec = GetTokPrecedence();
+
+        // If this is a binop that binds at least as tightly as the current binop,
+        // consume it, otherwise we are done.
+        if (TokPrec < ExprPrec)
+          return LHS;
+
+This code gets the precedence of the current token and checks to see if
+if is too low. Because we defined invalid tokens to have a precedence of
+-1, this check implicitly knows that the pair-stream ends when the token
+stream runs out of binary operators. If this check succeeds, we know
+that the token is a binary operator and that it will be included in this
+expression:
+
+.. code-block:: c++
+
+        // Okay, we know this is a binop.
+        int BinOp = CurTok;
+        getNextToken();  // eat binop
+
+        // Parse the primary expression after the binary operator.
+        auto RHS = ParsePrimary();
+        if (!RHS)
+          return nullptr;
+
+As such, this code eats (and remembers) the binary operator and then
+parses the primary expression that follows. This builds up the whole
+pair, the first of which is [+, b] for the running example.
+
+Now that we parsed the left-hand side of an expression and one pair of
+the RHS sequence, we have to decide which way the expression associates.
+In particular, we could have "(a+b) binop unparsed" or "a + (b binop
+unparsed)". To determine this, we look ahead at "binop" to determine its
+precedence and compare it to BinOp's precedence (which is '+' in this
+case):
+
+.. code-block:: c++
+
+        // If BinOp binds less tightly with RHS than the operator after RHS, let
+        // the pending operator take RHS as its LHS.
+        int NextPrec = GetTokPrecedence();
+        if (TokPrec < NextPrec) {
+
+If the precedence of the binop to the right of "RHS" is lower or equal
+to the precedence of our current operator, then we know that the
+parentheses associate as "(a+b) binop ...". In our example, the current
+operator is "+" and the next operator is "+", we know that they have the
+same precedence. In this case we'll create the AST node for "a+b", and
+then continue parsing:
+
+.. code-block:: c++
+
+          ... if body omitted ...
+        }
+
+        // Merge LHS/RHS.
+        LHS = llvm::make_unique<BinaryExprAST>(BinOp, std::move(LHS),
+                                               std::move(RHS));
+      }  // loop around to the top of the while loop.
+    }
+
+In our example above, this will turn "a+b+" into "(a+b)" and execute the
+next iteration of the loop, with "+" as the current token. The code
+above will eat, remember, and parse "(c+d)" as the primary expression,
+which makes the current pair equal to [+, (c+d)]. It will then evaluate
+the 'if' conditional above with "\*" as the binop to the right of the
+primary. In this case, the precedence of "\*" is higher than the
+precedence of "+" so the if condition will be entered.
+
+The critical question left here is "how can the if condition parse the
+right hand side in full"? In particular, to build the AST correctly for
+our example, it needs to get all of "(c+d)\*e\*f" as the RHS expression
+variable. The code to do this is surprisingly simple (code from the
+above two blocks duplicated for context):
+
+.. code-block:: c++
+
+        // If BinOp binds less tightly with RHS than the operator after RHS, let
+        // the pending operator take RHS as its LHS.
+        int NextPrec = GetTokPrecedence();
+        if (TokPrec < NextPrec) {
+          RHS = ParseBinOpRHS(TokPrec+1, std::move(RHS));
+          if (!RHS)
+            return nullptr;
+        }
+        // Merge LHS/RHS.
+        LHS = llvm::make_unique<BinaryExprAST>(BinOp, std::move(LHS),
+                                               std::move(RHS));
+      }  // loop around to the top of the while loop.
+    }
+
+At this point, we know that the binary operator to the RHS of our
+primary has higher precedence than the binop we are currently parsing.
+As such, we know that any sequence of pairs whose operators are all
+higher precedence than "+" should be parsed together and returned as
+"RHS". To do this, we recursively invoke the ``ParseBinOpRHS`` function
+specifying "TokPrec+1" as the minimum precedence required for it to
+continue. In our example above, this will cause it to return the AST
+node for "(c+d)\*e\*f" as RHS, which is then set as the RHS of the '+'
+expression.
+
+Finally, on the next iteration of the while loop, the "+g" piece is
+parsed and added to the AST. With this little bit of code (14
+non-trivial lines), we correctly handle fully general binary expression
+parsing in a very elegant way. This was a whirlwind tour of this code,
+and it is somewhat subtle. I recommend running through it with a few
+tough examples to see how it works.
+
+This wraps up handling of expressions. At this point, we can point the
+parser at an arbitrary token stream and build an expression from it,
+stopping at the first token that is not part of the expression. Next up
+we need to handle function definitions, etc.
+
+Parsing the Rest
+================
+
+The next thing missing is handling of function prototypes. In
+Kaleidoscope, these are used both for 'extern' function declarations as
+well as function body definitions. The code to do this is
+straight-forward and not very interesting (once you've survived
+expressions):
+
+.. code-block:: c++
+
+    /// prototype
+    ///   ::= id '(' id* ')'
+    static std::unique_ptr<PrototypeAST> ParsePrototype() {
+      if (CurTok != tok_identifier)
+        return LogErrorP("Expected function name in prototype");
+
+      std::string FnName = IdentifierStr;
+      getNextToken();
+
+      if (CurTok != '(')
+        return LogErrorP("Expected '(' in prototype");
+
+      // Read the list of argument names.
+      std::vector<std::string> ArgNames;
+      while (getNextToken() == tok_identifier)
+        ArgNames.push_back(IdentifierStr);
+      if (CurTok != ')')
+        return LogErrorP("Expected ')' in prototype");
+
+      // success.
+      getNextToken();  // eat ')'.
+
+      return llvm::make_unique<PrototypeAST>(FnName, std::move(ArgNames));
+    }
+
+Given this, a function definition is very simple, just a prototype plus
+an expression to implement the body:
+
+.. code-block:: c++
+
+    /// definition ::= 'def' prototype expression
+    static std::unique_ptr<FunctionAST> ParseDefinition() {
+      getNextToken();  // eat def.
+      auto Proto = ParsePrototype();
+      if (!Proto) return nullptr;
+
+      if (auto E = ParseExpression())
+        return llvm::make_unique<FunctionAST>(std::move(Proto), std::move(E));
+      return nullptr;
+    }
+
+In addition, we support 'extern' to declare functions like 'sin' and
+'cos' as well as to support forward declaration of user functions. These
+'extern's are just prototypes with no body:
+
+.. code-block:: c++
+
+    /// external ::= 'extern' prototype
+    static std::unique_ptr<PrototypeAST> ParseExtern() {
+      getNextToken();  // eat extern.
+      return ParsePrototype();
+    }
+
+Finally, we'll also let the user type in arbitrary top-level expressions
+and evaluate them on the fly. We will handle this by defining anonymous
+nullary (zero argument) functions for them:
+
+.. code-block:: c++
+
+    /// toplevelexpr ::= expression
+    static std::unique_ptr<FunctionAST> ParseTopLevelExpr() {
+      if (auto E = ParseExpression()) {
+        // Make an anonymous proto.
+        auto Proto = llvm::make_unique<PrototypeAST>("", std::vector<std::string>());
+        return llvm::make_unique<FunctionAST>(std::move(Proto), std::move(E));
+      }
+      return nullptr;
+    }
+
+Now that we have all the pieces, let's build a little driver that will
+let us actually *execute* this code we've built!
+
+The Driver
+==========
+
+The driver for this simply invokes all of the parsing pieces with a
+top-level dispatch loop. There isn't much interesting here, so I'll just
+include the top-level loop. See `below <#full-code-listing>`_ for full code in the
+"Top-Level Parsing" section.
+
+.. code-block:: c++
+
+    /// top ::= definition | external | expression | ';'
+    static void MainLoop() {
+      while (1) {
+        fprintf(stderr, "ready> ");
+        switch (CurTok) {
+        case tok_eof:
+          return;
+        case ';': // ignore top-level semicolons.
+          getNextToken();
+          break;
+        case tok_def:
+          HandleDefinition();
+          break;
+        case tok_extern:
+          HandleExtern();
+          break;
+        default:
+          HandleTopLevelExpression();
+          break;
+        }
+      }
+    }
+
+The most interesting part of this is that we ignore top-level
+semicolons. Why is this, you ask? The basic reason is that if you type
+"4 + 5" at the command line, the parser doesn't know whether that is the
+end of what you will type or not. For example, on the next line you
+could type "def foo..." in which case 4+5 is the end of a top-level
+expression. Alternatively you could type "\* 6", which would continue
+the expression. Having top-level semicolons allows you to type "4+5;",
+and the parser will know you are done.
+
+Conclusions
+===========
+
+With just under 400 lines of commented code (240 lines of non-comment,
+non-blank code), we fully defined our minimal language, including a
+lexer, parser, and AST builder. With this done, the executable will
+validate Kaleidoscope code and tell us if it is grammatically invalid.
+For example, here is a sample interaction:
+
+.. code-block:: bash
+
+    $ ./a.out
+    ready> def foo(x y) x+foo(y, 4.0);
+    Parsed a function definition.
+    ready> def foo(x y) x+y y;
+    Parsed a function definition.
+    Parsed a top-level expr
+    ready> def foo(x y) x+y );
+    Parsed a function definition.
+    Error: unknown token when expecting an expression
+    ready> extern sin(a);
+    ready> Parsed an extern
+    ready> ^D
+    $
+
+There is a lot of room for extension here. You can define new AST nodes,
+extend the language in many ways, etc. In the `next
+installment <LangImpl03.html>`_, we will describe how to generate LLVM
+Intermediate Representation (IR) from the AST.
+
+Full Code Listing
+=================
+
+Here is the complete code listing for this and the previous chapter.
+Note that it is fully self-contained: you don't need LLVM or any
+external libraries at all for this. (Besides the C and C++ standard
+libraries, of course.) To build this, just compile with:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g -O3 toy.cpp
+    # Run
+    ./a.out
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter2/toy.cpp
+   :language: c++
+
+`Next: Implementing Code Generation to LLVM IR <LangImpl03.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl03.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl03.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl03.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl03.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,567 @@
+========================================
+Kaleidoscope: Code generation to LLVM IR
+========================================
+
+.. contents::
+   :local:
+
+Chapter 3 Introduction
+======================
+
+Welcome to Chapter 3 of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. This chapter shows you how to transform
+the `Abstract Syntax Tree <LangImpl2.html>`_, built in Chapter 2, into
+LLVM IR. This will teach you a little bit about how LLVM does things, as
+well as demonstrate how easy it is to use. It's much more work to build
+a lexer and parser than it is to generate LLVM IR code. :)
+
+**Please note**: the code in this chapter and later require LLVM 3.7 or
+later. LLVM 3.6 and before will not work with it. Also note that you
+need to use a version of this tutorial that matches your LLVM release:
+If you are using an official LLVM release, use the version of the
+documentation included with your release or on the `llvm.org releases
+page <http://llvm.org/releases/>`_.
+
+Code Generation Setup
+=====================
+
+In order to generate LLVM IR, we want some simple setup to get started.
+First we define virtual code generation (codegen) methods in each AST
+class:
+
+.. code-block:: c++
+
+    /// ExprAST - Base class for all expression nodes.
+    class ExprAST {
+    public:
+      virtual ~ExprAST() {}
+      virtual Value *codegen() = 0;
+    };
+
+    /// NumberExprAST - Expression class for numeric literals like "1.0".
+    class NumberExprAST : public ExprAST {
+      double Val;
+
+    public:
+      NumberExprAST(double Val) : Val(Val) {}
+      virtual Value *codegen();
+    };
+    ...
+
+The codegen() method says to emit IR for that AST node along with all
+the things it depends on, and they all return an LLVM Value object.
+"Value" is the class used to represent a "`Static Single Assignment
+(SSA) <http://en.wikipedia.org/wiki/Static_single_assignment_form>`_
+register" or "SSA value" in LLVM. The most distinct aspect of SSA values
+is that their value is computed as the related instruction executes, and
+it does not get a new value until (and if) the instruction re-executes.
+In other words, there is no way to "change" an SSA value. For more
+information, please read up on `Static Single
+Assignment <http://en.wikipedia.org/wiki/Static_single_assignment_form>`_
+- the concepts are really quite natural once you grok them.
+
+Note that instead of adding virtual methods to the ExprAST class
+hierarchy, it could also make sense to use a `visitor
+pattern <http://en.wikipedia.org/wiki/Visitor_pattern>`_ or some other
+way to model this. Again, this tutorial won't dwell on good software
+engineering practices: for our purposes, adding a virtual method is
+simplest.
+
+The second thing we want is an "LogError" method like we used for the
+parser, which will be used to report errors found during code generation
+(for example, use of an undeclared parameter):
+
+.. code-block:: c++
+
+    static LLVMContext TheContext;
+    static IRBuilder<> Builder(TheContext);
+    static std::unique_ptr<Module> TheModule;
+    static std::map<std::string, Value *> NamedValues;
+
+    Value *LogErrorV(const char *Str) {
+      LogError(Str);
+      return nullptr;
+    }
+
+The static variables will be used during code generation. ``TheContext``
+is an opaque object that owns a lot of core LLVM data structures, such as
+the type and constant value tables. We don't need to understand it in
+detail, we just need a single instance to pass into APIs that require it.
+
+The ``Builder`` object is a helper object that makes it easy to generate
+LLVM instructions. Instances of the
+`IRBuilder <http://llvm.org/doxygen/IRBuilder_8h-source.html>`_
+class template keep track of the current place to insert instructions
+and has methods to create new instructions.
+
+``TheModule`` is an LLVM construct that contains functions and global
+variables. In many ways, it is the top-level structure that the LLVM IR
+uses to contain code. It will own the memory for all of the IR that we
+generate, which is why the codegen() method returns a raw Value\*,
+rather than a unique_ptr<Value>.
+
+The ``NamedValues`` map keeps track of which values are defined in the
+current scope and what their LLVM representation is. (In other words, it
+is a symbol table for the code). In this form of Kaleidoscope, the only
+things that can be referenced are function parameters. As such, function
+parameters will be in this map when generating code for their function
+body.
+
+With these basics in place, we can start talking about how to generate
+code for each expression. Note that this assumes that the ``Builder``
+has been set up to generate code *into* something. For now, we'll assume
+that this has already been done, and we'll just use it to emit code.
+
+Expression Code Generation
+==========================
+
+Generating LLVM code for expression nodes is very straightforward: less
+than 45 lines of commented code for all four of our expression nodes.
+First we'll do numeric literals:
+
+.. code-block:: c++
+
+    Value *NumberExprAST::codegen() {
+      return ConstantFP::get(LLVMContext, APFloat(Val));
+    }
+
+In the LLVM IR, numeric constants are represented with the
+``ConstantFP`` class, which holds the numeric value in an ``APFloat``
+internally (``APFloat`` has the capability of holding floating point
+constants of Arbitrary Precision). This code basically just creates
+and returns a ``ConstantFP``. Note that in the LLVM IR that constants
+are all uniqued together and shared. For this reason, the API uses the
+"foo::get(...)" idiom instead of "new foo(..)" or "foo::Create(..)".
+
+.. code-block:: c++
+
+    Value *VariableExprAST::codegen() {
+      // Look this variable up in the function.
+      Value *V = NamedValues[Name];
+      if (!V)
+        LogErrorV("Unknown variable name");
+      return V;
+    }
+
+References to variables are also quite simple using LLVM. In the simple
+version of Kaleidoscope, we assume that the variable has already been
+emitted somewhere and its value is available. In practice, the only
+values that can be in the ``NamedValues`` map are function arguments.
+This code simply checks to see that the specified name is in the map (if
+not, an unknown variable is being referenced) and returns the value for
+it. In future chapters, we'll add support for `loop induction
+variables <LangImpl5.html#for-loop-expression>`_ in the symbol table, and for `local
+variables <LangImpl7.html#user-defined-local-variables>`_.
+
+.. code-block:: c++
+
+    Value *BinaryExprAST::codegen() {
+      Value *L = LHS->codegen();
+      Value *R = RHS->codegen();
+      if (!L || !R)
+        return nullptr;
+
+      switch (Op) {
+      case '+':
+        return Builder.CreateFAdd(L, R, "addtmp");
+      case '-':
+        return Builder.CreateFSub(L, R, "subtmp");
+      case '*':
+        return Builder.CreateFMul(L, R, "multmp");
+      case '<':
+        L = Builder.CreateFCmpULT(L, R, "cmptmp");
+        // Convert bool 0/1 to double 0.0 or 1.0
+        return Builder.CreateUIToFP(L, Type::getDoubleTy(LLVMContext),
+                                    "booltmp");
+      default:
+        return LogErrorV("invalid binary operator");
+      }
+    }
+
+Binary operators start to get more interesting. The basic idea here is
+that we recursively emit code for the left-hand side of the expression,
+then the right-hand side, then we compute the result of the binary
+expression. In this code, we do a simple switch on the opcode to create
+the right LLVM instruction.
+
+In the example above, the LLVM builder class is starting to show its
+value. IRBuilder knows where to insert the newly created instruction,
+all you have to do is specify what instruction to create (e.g. with
+``CreateFAdd``), which operands to use (``L`` and ``R`` here) and
+optionally provide a name for the generated instruction.
+
+One nice thing about LLVM is that the name is just a hint. For instance,
+if the code above emits multiple "addtmp" variables, LLVM will
+automatically provide each one with an increasing, unique numeric
+suffix. Local value names for instructions are purely optional, but it
+makes it much easier to read the IR dumps.
+
+`LLVM instructions <../LangRef.html#instruction-reference>`_ are constrained by strict
+rules: for example, the Left and Right operators of an `add
+instruction <../LangRef.html#add-instruction>`_ must have the same type, and the
+result type of the add must match the operand types. Because all values
+in Kaleidoscope are doubles, this makes for very simple code for add,
+sub and mul.
+
+On the other hand, LLVM specifies that the `fcmp
+instruction <../LangRef.html#fcmp-instruction>`_ always returns an 'i1' value (a
+one bit integer). The problem with this is that Kaleidoscope wants the
+value to be a 0.0 or 1.0 value. In order to get these semantics, we
+combine the fcmp instruction with a `uitofp
+instruction <../LangRef.html#uitofp-to-instruction>`_. This instruction converts its
+input integer into a floating point value by treating the input as an
+unsigned value. In contrast, if we used the `sitofp
+instruction <../LangRef.html#sitofp-to-instruction>`_, the Kaleidoscope '<' operator
+would return 0.0 and -1.0, depending on the input value.
+
+.. code-block:: c++
+
+    Value *CallExprAST::codegen() {
+      // Look up the name in the global module table.
+      Function *CalleeF = TheModule->getFunction(Callee);
+      if (!CalleeF)
+        return LogErrorV("Unknown function referenced");
+
+      // If argument mismatch error.
+      if (CalleeF->arg_size() != Args.size())
+        return LogErrorV("Incorrect # arguments passed");
+
+      std::vector<Value *> ArgsV;
+      for (unsigned i = 0, e = Args.size(); i != e; ++i) {
+        ArgsV.push_back(Args[i]->codegen());
+        if (!ArgsV.back())
+          return nullptr;
+      }
+
+      return Builder.CreateCall(CalleeF, ArgsV, "calltmp");
+    }
+
+Code generation for function calls is quite straightforward with LLVM. The code
+above initially does a function name lookup in the LLVM Module's symbol table.
+Recall that the LLVM Module is the container that holds the functions we are
+JIT'ing. By giving each function the same name as what the user specifies, we
+can use the LLVM symbol table to resolve function names for us.
+
+Once we have the function to call, we recursively codegen each argument
+that is to be passed in, and create an LLVM `call
+instruction <../LangRef.html#call-instruction>`_. Note that LLVM uses the native C
+calling conventions by default, allowing these calls to also call into
+standard library functions like "sin" and "cos", with no additional
+effort.
+
+This wraps up our handling of the four basic expressions that we have so
+far in Kaleidoscope. Feel free to go in and add some more. For example,
+by browsing the `LLVM language reference <../LangRef.html>`_ you'll find
+several other interesting instructions that are really easy to plug into
+our basic framework.
+
+Function Code Generation
+========================
+
+Code generation for prototypes and functions must handle a number of
+details, which make their code less beautiful than expression code
+generation, but allows us to illustrate some important points. First,
+lets talk about code generation for prototypes: they are used both for
+function bodies and external function declarations. The code starts
+with:
+
+.. code-block:: c++
+
+    Function *PrototypeAST::codegen() {
+      // Make the function type:  double(double,double) etc.
+      std::vector<Type*> Doubles(Args.size(),
+                                 Type::getDoubleTy(LLVMContext));
+      FunctionType *FT =
+        FunctionType::get(Type::getDoubleTy(LLVMContext), Doubles, false);
+
+      Function *F =
+        Function::Create(FT, Function::ExternalLinkage, Name, TheModule);
+
+This code packs a lot of power into a few lines. Note first that this
+function returns a "Function\*" instead of a "Value\*". Because a
+"prototype" really talks about the external interface for a function
+(not the value computed by an expression), it makes sense for it to
+return the LLVM Function it corresponds to when codegen'd.
+
+The call to ``FunctionType::get`` creates the ``FunctionType`` that
+should be used for a given Prototype. Since all function arguments in
+Kaleidoscope are of type double, the first line creates a vector of "N"
+LLVM double types. It then uses the ``Functiontype::get`` method to
+create a function type that takes "N" doubles as arguments, returns one
+double as a result, and that is not vararg (the false parameter
+indicates this). Note that Types in LLVM are uniqued just like Constants
+are, so you don't "new" a type, you "get" it.
+
+The final line above actually creates the IR Function corresponding to
+the Prototype. This indicates the type, linkage and name to use, as
+well as which module to insert into. "`external
+linkage <../LangRef.html#linkage>`_" means that the function may be
+defined outside the current module and/or that it is callable by
+functions outside the module. The Name passed in is the name the user
+specified: since "``TheModule``" is specified, this name is registered
+in "``TheModule``"s symbol table.
+
+.. code-block:: c++
+
+  // Set names for all arguments.
+  unsigned Idx = 0;
+  for (auto &Arg : F->args())
+    Arg.setName(Args[Idx++]);
+
+  return F;
+
+Finally, we set the name of each of the function's arguments according to the
+names given in the Prototype. This step isn't strictly necessary, but keeping
+the names consistent makes the IR more readable, and allows subsequent code to
+refer directly to the arguments for their names, rather than having to look up
+them up in the Prototype AST.
+
+At this point we have a function prototype with no body. This is how LLVM IR
+represents function declarations. For extern statements in Kaleidoscope, this
+is as far as we need to go. For function definitions however, we need to
+codegen and attach a function body.
+
+.. code-block:: c++
+
+  Function *FunctionAST::codegen() {
+      // First, check for an existing function from a previous 'extern' declaration.
+    Function *TheFunction = TheModule->getFunction(Proto->getName());
+
+    if (!TheFunction)
+      TheFunction = Proto->codegen();
+
+    if (!TheFunction)
+      return nullptr;
+
+    if (!TheFunction->empty())
+      return (Function*)LogErrorV("Function cannot be redefined.");
+
+
+For function definitions, we start by searching TheModule's symbol table for an
+existing version of this function, in case one has already been created using an
+'extern' statement. If Module::getFunction returns null then no previous version
+exists, so we'll codegen one from the Prototype. In either case, we want to
+assert that the function is empty (i.e. has no body yet) before we start.
+
+.. code-block:: c++
+
+  // Create a new basic block to start insertion into.
+  BasicBlock *BB = BasicBlock::Create(LLVMContext, "entry", TheFunction);
+  Builder.SetInsertPoint(BB);
+
+  // Record the function arguments in the NamedValues map.
+  NamedValues.clear();
+  for (auto &Arg : TheFunction->args())
+    NamedValues[Arg.getName()] = &Arg;
+
+Now we get to the point where the ``Builder`` is set up. The first line
+creates a new `basic block <http://en.wikipedia.org/wiki/Basic_block>`_
+(named "entry"), which is inserted into ``TheFunction``. The second line
+then tells the builder that new instructions should be inserted into the
+end of the new basic block. Basic blocks in LLVM are an important part
+of functions that define the `Control Flow
+Graph <http://en.wikipedia.org/wiki/Control_flow_graph>`_. Since we
+don't have any control flow, our functions will only contain one block
+at this point. We'll fix this in `Chapter 5 <LangImpl5.html>`_ :).
+
+Next we add the function arguments to the NamedValues map (after first clearing
+it out) so that they're accessible to ``VariableExprAST`` nodes.
+
+.. code-block:: c++
+
+      if (Value *RetVal = Body->codegen()) {
+        // Finish off the function.
+        Builder.CreateRet(RetVal);
+
+        // Validate the generated code, checking for consistency.
+        verifyFunction(*TheFunction);
+
+        return TheFunction;
+      }
+
+Once the insertion point has been set up and the NamedValues map populated,
+we call the ``codegen()`` method for the root expression of the function. If no
+error happens, this emits code to compute the expression into the entry block
+and returns the value that was computed. Assuming no error, we then create an
+LLVM `ret instruction <../LangRef.html#ret-instruction>`_, which completes the function.
+Once the function is built, we call ``verifyFunction``, which is
+provided by LLVM. This function does a variety of consistency checks on
+the generated code, to determine if our compiler is doing everything
+right. Using this is important: it can catch a lot of bugs. Once the
+function is finished and validated, we return it.
+
+.. code-block:: c++
+
+      // Error reading body, remove function.
+      TheFunction->eraseFromParent();
+      return nullptr;
+    }
+
+The only piece left here is handling of the error case. For simplicity,
+we handle this by merely deleting the function we produced with the
+``eraseFromParent`` method. This allows the user to redefine a function
+that they incorrectly typed in before: if we didn't delete it, it would
+live in the symbol table, with a body, preventing future redefinition.
+
+This code does have a bug, though: If the ``FunctionAST::codegen()`` method
+finds an existing IR Function, it does not validate its signature against the
+definition's own prototype. This means that an earlier 'extern' declaration will
+take precedence over the function definition's signature, which can cause
+codegen to fail, for instance if the function arguments are named differently.
+There are a number of ways to fix this bug, see what you can come up with! Here
+is a testcase:
+
+::
+
+    extern foo(a);     # ok, defines foo.
+    def foo(b) b;      # Error: Unknown variable name. (decl using 'a' takes precedence).
+
+Driver Changes and Closing Thoughts
+===================================
+
+For now, code generation to LLVM doesn't really get us much, except that
+we can look at the pretty IR calls. The sample code inserts calls to
+codegen into the "``HandleDefinition``", "``HandleExtern``" etc
+functions, and then dumps out the LLVM IR. This gives a nice way to look
+at the LLVM IR for simple functions. For example:
+
+::
+
+    ready> 4+5;
+    Read top-level expression:
+    define double @0() {
+    entry:
+      ret double 9.000000e+00
+    }
+
+Note how the parser turns the top-level expression into anonymous
+functions for us. This will be handy when we add `JIT
+support <LangImpl4.html#adding-a-jit-compiler>`_ in the next chapter. Also note that the
+code is very literally transcribed, no optimizations are being performed
+except simple constant folding done by IRBuilder. We will `add
+optimizations <LangImpl4.html#trivial-constant-folding>`_ explicitly in the next
+chapter.
+
+::
+
+    ready> def foo(a b) a*a + 2*a*b + b*b;
+    Read function definition:
+    define double @foo(double %a, double %b) {
+    entry:
+      %multmp = fmul double %a, %a
+      %multmp1 = fmul double 2.000000e+00, %a
+      %multmp2 = fmul double %multmp1, %b
+      %addtmp = fadd double %multmp, %multmp2
+      %multmp3 = fmul double %b, %b
+      %addtmp4 = fadd double %addtmp, %multmp3
+      ret double %addtmp4
+    }
+
+This shows some simple arithmetic. Notice the striking similarity to the
+LLVM builder calls that we use to create the instructions.
+
+::
+
+    ready> def bar(a) foo(a, 4.0) + bar(31337);
+    Read function definition:
+    define double @bar(double %a) {
+    entry:
+      %calltmp = call double @foo(double %a, double 4.000000e+00)
+      %calltmp1 = call double @bar(double 3.133700e+04)
+      %addtmp = fadd double %calltmp, %calltmp1
+      ret double %addtmp
+    }
+
+This shows some function calls. Note that this function will take a long
+time to execute if you call it. In the future we'll add conditional
+control flow to actually make recursion useful :).
+
+::
+
+    ready> extern cos(x);
+    Read extern:
+    declare double @cos(double)
+
+    ready> cos(1.234);
+    Read top-level expression:
+    define double @1() {
+    entry:
+      %calltmp = call double @cos(double 1.234000e+00)
+      ret double %calltmp
+    }
+
+This shows an extern for the libm "cos" function, and a call to it.
+
+.. TODO:: Abandon Pygments' horrible `llvm` lexer. It just totally gives up
+   on highlighting this due to the first line.
+
+::
+
+    ready> ^D
+    ; ModuleID = 'my cool jit'
+
+    define double @0() {
+    entry:
+      %addtmp = fadd double 4.000000e+00, 5.000000e+00
+      ret double %addtmp
+    }
+
+    define double @foo(double %a, double %b) {
+    entry:
+      %multmp = fmul double %a, %a
+      %multmp1 = fmul double 2.000000e+00, %a
+      %multmp2 = fmul double %multmp1, %b
+      %addtmp = fadd double %multmp, %multmp2
+      %multmp3 = fmul double %b, %b
+      %addtmp4 = fadd double %addtmp, %multmp3
+      ret double %addtmp4
+    }
+
+    define double @bar(double %a) {
+    entry:
+      %calltmp = call double @foo(double %a, double 4.000000e+00)
+      %calltmp1 = call double @bar(double 3.133700e+04)
+      %addtmp = fadd double %calltmp, %calltmp1
+      ret double %addtmp
+    }
+
+    declare double @cos(double)
+
+    define double @1() {
+    entry:
+      %calltmp = call double @cos(double 1.234000e+00)
+      ret double %calltmp
+    }
+
+When you quit the current demo, it dumps out the IR for the entire
+module generated. Here you can see the big picture with all the
+functions referencing each other.
+
+This wraps up the third chapter of the Kaleidoscope tutorial. Up next,
+we'll describe how to `add JIT codegen and optimizer
+support <LangImpl4.html>`_ to this so we can actually start running
+code!
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example, enhanced with
+the LLVM code generator. Because this uses the LLVM libraries, we need
+to link them in. To do this, we use the
+`llvm-config <http://llvm.org/cmds/llvm-config.html>`_ tool to inform
+our makefile/command line about which options to use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g -O3 toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core` -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter3/toy.cpp
+   :language: c++
+
+`Next: Adding JIT and Optimizer Support <LangImpl04.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl04.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl04.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl04.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl04.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,610 @@
+==============================================
+Kaleidoscope: Adding JIT and Optimizer Support
+==============================================
+
+.. contents::
+   :local:
+
+Chapter 4 Introduction
+======================
+
+Welcome to Chapter 4 of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. Chapters 1-3 described the implementation
+of a simple language and added support for generating LLVM IR. This
+chapter describes two new techniques: adding optimizer support to your
+language, and adding JIT compiler support. These additions will
+demonstrate how to get nice, efficient code for the Kaleidoscope
+language.
+
+Trivial Constant Folding
+========================
+
+Our demonstration for Chapter 3 is elegant and easy to extend.
+Unfortunately, it does not produce wonderful code. The IRBuilder,
+however, does give us obvious optimizations when compiling simple code:
+
+::
+
+    ready> def test(x) 1+2+x;
+    Read function definition:
+    define double @test(double %x) {
+    entry:
+            %addtmp = fadd double 3.000000e+00, %x
+            ret double %addtmp
+    }
+
+This code is not a literal transcription of the AST built by parsing the
+input. That would be:
+
+::
+
+    ready> def test(x) 1+2+x;
+    Read function definition:
+    define double @test(double %x) {
+    entry:
+            %addtmp = fadd double 2.000000e+00, 1.000000e+00
+            %addtmp1 = fadd double %addtmp, %x
+            ret double %addtmp1
+    }
+
+Constant folding, as seen above, in particular, is a very common and
+very important optimization: so much so that many language implementors
+implement constant folding support in their AST representation.
+
+With LLVM, you don't need this support in the AST. Since all calls to
+build LLVM IR go through the LLVM IR builder, the builder itself checked
+to see if there was a constant folding opportunity when you call it. If
+so, it just does the constant fold and return the constant instead of
+creating an instruction.
+
+Well, that was easy :). In practice, we recommend always using
+``IRBuilder`` when generating code like this. It has no "syntactic
+overhead" for its use (you don't have to uglify your compiler with
+constant checks everywhere) and it can dramatically reduce the amount of
+LLVM IR that is generated in some cases (particular for languages with a
+macro preprocessor or that use a lot of constants).
+
+On the other hand, the ``IRBuilder`` is limited by the fact that it does
+all of its analysis inline with the code as it is built. If you take a
+slightly more complex example:
+
+::
+
+    ready> def test(x) (1+2+x)*(x+(1+2));
+    ready> Read function definition:
+    define double @test(double %x) {
+    entry:
+            %addtmp = fadd double 3.000000e+00, %x
+            %addtmp1 = fadd double %x, 3.000000e+00
+            %multmp = fmul double %addtmp, %addtmp1
+            ret double %multmp
+    }
+
+In this case, the LHS and RHS of the multiplication are the same value.
+We'd really like to see this generate "``tmp = x+3; result = tmp*tmp;``"
+instead of computing "``x+3``" twice.
+
+Unfortunately, no amount of local analysis will be able to detect and
+correct this. This requires two transformations: reassociation of
+expressions (to make the add's lexically identical) and Common
+Subexpression Elimination (CSE) to delete the redundant add instruction.
+Fortunately, LLVM provides a broad range of optimizations that you can
+use, in the form of "passes".
+
+LLVM Optimization Passes
+========================
+
+LLVM provides many optimization passes, which do many different sorts of
+things and have different tradeoffs. Unlike other systems, LLVM doesn't
+hold to the mistaken notion that one set of optimizations is right for
+all languages and for all situations. LLVM allows a compiler implementor
+to make complete decisions about what optimizations to use, in which
+order, and in what situation.
+
+As a concrete example, LLVM supports both "whole module" passes, which
+look across as large of body of code as they can (often a whole file,
+but if run at link time, this can be a substantial portion of the whole
+program). It also supports and includes "per-function" passes which just
+operate on a single function at a time, without looking at other
+functions. For more information on passes and how they are run, see the
+`How to Write a Pass <../WritingAnLLVMPass.html>`_ document and the
+`List of LLVM Passes <../Passes.html>`_.
+
+For Kaleidoscope, we are currently generating functions on the fly, one
+at a time, as the user types them in. We aren't shooting for the
+ultimate optimization experience in this setting, but we also want to
+catch the easy and quick stuff where possible. As such, we will choose
+to run a few per-function optimizations as the user types the function
+in. If we wanted to make a "static Kaleidoscope compiler", we would use
+exactly the code we have now, except that we would defer running the
+optimizer until the entire file has been parsed.
+
+In order to get per-function optimizations going, we need to set up a
+`FunctionPassManager <../WritingAnLLVMPass.html#what-passmanager-doesr>`_ to hold
+and organize the LLVM optimizations that we want to run. Once we have
+that, we can add a set of optimizations to run. We'll need a new
+FunctionPassManager for each module that we want to optimize, so we'll
+write a function to create and initialize both the module and pass manager
+for us:
+
+.. code-block:: c++
+
+    void InitializeModuleAndPassManager(void) {
+      // Open a new module.
+      Context LLVMContext;
+      TheModule = llvm::make_unique<Module>("my cool jit", LLVMContext);
+      TheModule->setDataLayout(TheJIT->getTargetMachine().createDataLayout());
+
+      // Create a new pass manager attached to it.
+      TheFPM = llvm::make_unique<FunctionPassManager>(TheModule.get());
+
+      // Provide basic AliasAnalysis support for GVN.
+      TheFPM.add(createBasicAliasAnalysisPass());
+      // Do simple "peephole" optimizations and bit-twiddling optzns.
+      TheFPM.add(createInstructionCombiningPass());
+      // Reassociate expressions.
+      TheFPM.add(createReassociatePass());
+      // Eliminate Common SubExpressions.
+      TheFPM.add(createGVNPass());
+      // Simplify the control flow graph (deleting unreachable blocks, etc).
+      TheFPM.add(createCFGSimplificationPass());
+
+      TheFPM.doInitialization();
+    }
+
+This code initializes the global module ``TheModule``, and the function pass
+manager ``TheFPM``, which is attached to ``TheModule``. Once the pass manager is
+set up, we use a series of "add" calls to add a bunch of LLVM passes.
+
+In this case, we choose to add five passes: one analysis pass (alias analysis),
+and four optimization passes. The passes we choose here are a pretty standard set
+of "cleanup" optimizations that are useful for a wide variety of code. I won't
+delve into what they do but, believe me, they are a good starting place :).
+
+Once the PassManager is set up, we need to make use of it. We do this by
+running it after our newly created function is constructed (in
+``FunctionAST::codegen()``), but before it is returned to the client:
+
+.. code-block:: c++
+
+      if (Value *RetVal = Body->codegen()) {
+        // Finish off the function.
+        Builder.CreateRet(RetVal);
+
+        // Validate the generated code, checking for consistency.
+        verifyFunction(*TheFunction);
+
+        // Optimize the function.
+        TheFPM->run(*TheFunction);
+
+        return TheFunction;
+      }
+
+As you can see, this is pretty straightforward. The
+``FunctionPassManager`` optimizes and updates the LLVM Function\* in
+place, improving (hopefully) its body. With this in place, we can try
+our test above again:
+
+::
+
+    ready> def test(x) (1+2+x)*(x+(1+2));
+    ready> Read function definition:
+    define double @test(double %x) {
+    entry:
+            %addtmp = fadd double %x, 3.000000e+00
+            %multmp = fmul double %addtmp, %addtmp
+            ret double %multmp
+    }
+
+As expected, we now get our nicely optimized code, saving a floating
+point add instruction from every execution of this function.
+
+LLVM provides a wide variety of optimizations that can be used in
+certain circumstances. Some `documentation about the various
+passes <../Passes.html>`_ is available, but it isn't very complete.
+Another good source of ideas can come from looking at the passes that
+``Clang`` runs to get started. The "``opt``" tool allows you to
+experiment with passes from the command line, so you can see if they do
+anything.
+
+Now that we have reasonable code coming out of our front-end, lets talk
+about executing it!
+
+Adding a JIT Compiler
+=====================
+
+Code that is available in LLVM IR can have a wide variety of tools
+applied to it. For example, you can run optimizations on it (as we did
+above), you can dump it out in textual or binary forms, you can compile
+the code to an assembly file (.s) for some target, or you can JIT
+compile it. The nice thing about the LLVM IR representation is that it
+is the "common currency" between many different parts of the compiler.
+
+In this section, we'll add JIT compiler support to our interpreter. The
+basic idea that we want for Kaleidoscope is to have the user enter
+function bodies as they do now, but immediately evaluate the top-level
+expressions they type in. For example, if they type in "1 + 2;", we
+should evaluate and print out 3. If they define a function, they should
+be able to call it from the command line.
+
+In order to do this, we first declare and initialize the JIT. This is
+done by adding a global variable ``TheJIT``, and initializing it in
+``main``:
+
+.. code-block:: c++
+
+    static std::unique_ptr<KaleidoscopeJIT> TheJIT;
+    ...
+    int main() {
+      ..
+      TheJIT = llvm::make_unique<KaleidoscopeJIT>();
+
+      // Run the main "interpreter loop" now.
+      MainLoop();
+
+      return 0;
+    }
+
+The KaleidoscopeJIT class is a simple JIT built specifically for these
+tutorials. In later chapters we will look at how it works and extend it with
+new features, but for now we will take it as given. Its API is very simple::
+``addModule`` adds an LLVM IR module to the JIT, making its functions
+available for execution; ``removeModule`` removes a module, freeing any
+memory associated with the code in that module; and ``findSymbol`` allows us
+to look up pointers to the compiled code.
+
+We can take this simple API and change our code that parses top-level expressions to
+look like this:
+
+.. code-block:: c++
+
+    static void HandleTopLevelExpression() {
+      // Evaluate a top-level expression into an anonymous function.
+      if (auto FnAST = ParseTopLevelExpr()) {
+        if (FnAST->codegen()) {
+
+          // JIT the module containing the anonymous expression, keeping a handle so
+          // we can free it later.
+          auto H = TheJIT->addModule(std::move(TheModule));
+          InitializeModuleAndPassManager();
+
+          // Search the JIT for the __anon_expr symbol.
+          auto ExprSymbol = TheJIT->findSymbol("__anon_expr");
+          assert(ExprSymbol && "Function not found");
+
+          // Get the symbol's address and cast it to the right type (takes no
+          // arguments, returns a double) so we can call it as a native function.
+          double (*FP)() = (double (*)())(intptr_t)ExprSymbol.getAddress();
+          fprintf(stderr, "Evaluated to %f\n", FP());
+
+          // Delete the anonymous expression module from the JIT.
+          TheJIT->removeModule(H);
+        }
+
+If parsing and codegen succeeed, the next step is to add the module containing
+the top-level expression to the JIT. We do this by calling addModule, which
+triggers code generation for all the functions in the module, and returns a
+handle that can be used to remove the module from the JIT later. Once the module
+has been added to the JIT it can no longer be modified, so we also open a new
+module to hold subsequent code by calling ``InitializeModuleAndPassManager()``.
+
+Once we've added the module to the JIT we need to get a pointer to the final
+generated code. We do this by calling the JIT's findSymbol method, and passing
+the name of the top-level expression function: ``__anon_expr``. Since we just
+added this function, we assert that findSymbol returned a result.
+
+Next, we get the in-memory address of the ``__anon_expr`` function by calling
+``getAddress()`` on the symbol. Recall that we compile top-level expressions
+into a self-contained LLVM function that takes no arguments and returns the
+computed double. Because the LLVM JIT compiler matches the native platform ABI,
+this means that you can just cast the result pointer to a function pointer of
+that type and call it directly. This means, there is no difference between JIT
+compiled code and native machine code that is statically linked into your
+application.
+
+Finally, since we don't support re-evaluation of top-level expressions, we
+remove the module from the JIT when we're done to free the associated memory.
+Recall, however, that the module we created a few lines earlier (via
+``InitializeModuleAndPassManager``) is still open and waiting for new code to be
+added.
+
+With just these two changes, lets see how Kaleidoscope works now!
+
+::
+
+    ready> 4+5;
+    Read top-level expression:
+    define double @0() {
+    entry:
+      ret double 9.000000e+00
+    }
+
+    Evaluated to 9.000000
+
+Well this looks like it is basically working. The dump of the function
+shows the "no argument function that always returns double" that we
+synthesize for each top-level expression that is typed in. This
+demonstrates very basic functionality, but can we do more?
+
+::
+
+    ready> def testfunc(x y) x + y*2;
+    Read function definition:
+    define double @testfunc(double %x, double %y) {
+    entry:
+      %multmp = fmul double %y, 2.000000e+00
+      %addtmp = fadd double %multmp, %x
+      ret double %addtmp
+    }
+
+    ready> testfunc(4, 10);
+    Read top-level expression:
+    define double @1() {
+    entry:
+      %calltmp = call double @testfunc(double 4.000000e+00, double 1.000000e+01)
+      ret double %calltmp
+    }
+
+    Evaluated to 24.000000
+
+    ready> testfunc(5, 10);
+    ready> LLVM ERROR: Program used external function 'testfunc' which could not be resolved!
+
+
+Function definitions and calls also work, but something went very wrong on that
+last line. The call looks valid, so what happened? As you may have guessed from
+the the API a Module is a unit of allocation for the JIT, and testfunc was part
+of the same module that contained anonymous expression. When we removed that
+module from the JIT to free the memory for the anonymous expression, we deleted
+the definition of ``testfunc`` along with it. Then, when we tried to call
+testfunc a second time, the JIT could no longer find it.
+
+The easiest way to fix this is to put the anonymous expression in a separate
+module from the rest of the function definitions. The JIT will happily resolve
+function calls across module boundaries, as long as each of the functions called
+has a prototype, and is added to the JIT before it is called. By putting the
+anonymous expression in a different module we can delete it without affecting
+the rest of the functions.
+
+In fact, we're going to go a step further and put every function in its own
+module. Doing so allows us to exploit a useful property of the KaleidoscopeJIT
+that will make our environment more REPL-like: Functions can be added to the
+JIT more than once (unlike a module where every function must have a unique
+definition). When you look up a symbol in KaleidoscopeJIT it will always return
+the most recent definition:
+
+::
+
+    ready> def foo(x) x + 1;
+    Read function definition:
+    define double @foo(double %x) {
+    entry:
+      %addtmp = fadd double %x, 1.000000e+00
+      ret double %addtmp
+    }
+
+    ready> foo(2);
+    Evaluated to 3.000000
+
+    ready> def foo(x) x + 2;
+    define double @foo(double %x) {
+    entry:
+      %addtmp = fadd double %x, 2.000000e+00
+      ret double %addtmp
+    }
+
+    ready> foo(2);
+    Evaluated to 4.000000
+
+
+To allow each function to live in its own module we'll need a way to
+re-generate previous function declarations into each new module we open:
+
+.. code-block:: c++
+
+    static std::unique_ptr<KaleidoscopeJIT> TheJIT;
+
+    ...
+
+    Function *getFunction(std::string Name) {
+      // First, see if the function has already been added to the current module.
+      if (auto *F = TheModule->getFunction(Name))
+        return F;
+
+      // If not, check whether we can codegen the declaration from some existing
+      // prototype.
+      auto FI = FunctionProtos.find(Name);
+      if (FI != FunctionProtos.end())
+        return FI->second->codegen();
+
+      // If no existing prototype exists, return null.
+      return nullptr;
+    }
+
+    ...
+
+    Value *CallExprAST::codegen() {
+      // Look up the name in the global module table.
+      Function *CalleeF = getFunction(Callee);
+
+    ...
+
+    Function *FunctionAST::codegen() {
+      // Transfer ownership of the prototype to the FunctionProtos map, but keep a
+      // reference to it for use below.
+      auto &P = *Proto;
+      FunctionProtos[Proto->getName()] = std::move(Proto);
+      Function *TheFunction = getFunction(P.getName());
+      if (!TheFunction)
+        return nullptr;
+
+
+To enable this, we'll start by adding a new global, ``FunctionProtos``, that
+holds the most recent prototype for each function. We'll also add a convenience
+method, ``getFunction()``, to replace calls to ``TheModule->getFunction()``.
+Our convenience method searches ``TheModule`` for an existing function
+declaration, falling back to generating a new declaration from FunctionProtos if
+it doesn't find one. In ``CallExprAST::codegen()`` we just need to replace the
+call to ``TheModule->getFunction()``. In ``FunctionAST::codegen()`` we need to
+update the FunctionProtos map first, then call ``getFunction()``. With this
+done, we can always obtain a function declaration in the current module for any
+previously declared function.
+
+We also need to update HandleDefinition and HandleExtern:
+
+.. code-block:: c++
+
+    static void HandleDefinition() {
+      if (auto FnAST = ParseDefinition()) {
+        if (auto *FnIR = FnAST->codegen()) {
+          fprintf(stderr, "Read function definition:");
+          FnIR->dump();
+          TheJIT->addModule(std::move(TheModule));
+          InitializeModuleAndPassManager();
+        }
+      } else {
+        // Skip token for error recovery.
+         getNextToken();
+      }
+    }
+
+    static void HandleExtern() {
+      if (auto ProtoAST = ParseExtern()) {
+        if (auto *FnIR = ProtoAST->codegen()) {
+          fprintf(stderr, "Read extern: ");
+          FnIR->dump();
+          FunctionProtos[ProtoAST->getName()] = std::move(ProtoAST);
+        }
+      } else {
+        // Skip token for error recovery.
+        getNextToken();
+      }
+    }
+
+In HandleDefinition, we add two lines to transfer the newly defined function to
+the JIT and open a new module. In HandleExtern, we just need to add one line to
+add the prototype to FunctionProtos.
+
+With these changes made, lets try our REPL again (I removed the dump of the
+anonymous functions this time, you should get the idea by now :) :
+
+::
+
+    ready> def foo(x) x + 1;
+    ready> foo(2);
+    Evaluated to 3.000000
+
+    ready> def foo(x) x + 2;
+    ready> foo(2);
+    Evaluated to 4.000000
+
+It works!
+
+Even with this simple code, we get some surprisingly powerful capabilities -
+check this out:
+
+::
+
+    ready> extern sin(x);
+    Read extern:
+    declare double @sin(double)
+
+    ready> extern cos(x);
+    Read extern:
+    declare double @cos(double)
+
+    ready> sin(1.0);
+    Read top-level expression:
+    define double @2() {
+    entry:
+      ret double 0x3FEAED548F090CEE
+    }
+
+    Evaluated to 0.841471
+
+    ready> def foo(x) sin(x)*sin(x) + cos(x)*cos(x);
+    Read function definition:
+    define double @foo(double %x) {
+    entry:
+      %calltmp = call double @sin(double %x)
+      %multmp = fmul double %calltmp, %calltmp
+      %calltmp2 = call double @cos(double %x)
+      %multmp4 = fmul double %calltmp2, %calltmp2
+      %addtmp = fadd double %multmp, %multmp4
+      ret double %addtmp
+    }
+
+    ready> foo(4.0);
+    Read top-level expression:
+    define double @3() {
+    entry:
+      %calltmp = call double @foo(double 4.000000e+00)
+      ret double %calltmp
+    }
+
+    Evaluated to 1.000000
+
+Whoa, how does the JIT know about sin and cos? The answer is surprisingly
+simple: The KaleidoscopeJIT has a straightforward symbol resolution rule that
+it uses to find symbols that aren't available in any given module: First
+it searches all the modules that have already been added to the JIT, from the
+most recent to the oldest, to find the newest definition. If no definition is
+found inside the JIT, it falls back to calling "``dlsym("sin")``" on the
+Kaleidoscope process itself. Since "``sin``" is defined within the JIT's
+address space, it simply patches up calls in the module to call the libm
+version of ``sin`` directly.
+
+In the future we'll see how tweaking this symbol resolution rule can be used to
+enable all sorts of useful features, from security (restricting the set of
+symbols available to JIT'd code), to dynamic code generation based on symbol
+names, and even lazy compilation.
+
+One immediate benefit of the symbol resolution rule is that we can now extend
+the language by writing arbitrary C++ code to implement operations. For example,
+if we add:
+
+.. code-block:: c++
+
+    /// putchard - putchar that takes a double and returns 0.
+    extern "C" double putchard(double X) {
+      fputc((char)X, stderr);
+      return 0;
+    }
+
+Now we can produce simple output to the console by using things like:
+"``extern putchard(x); putchard(120);``", which prints a lowercase 'x'
+on the console (120 is the ASCII code for 'x'). Similar code could be
+used to implement file I/O, console input, and many other capabilities
+in Kaleidoscope.
+
+This completes the JIT and optimizer chapter of the Kaleidoscope
+tutorial. At this point, we can compile a non-Turing-complete
+programming language, optimize and JIT compile it in a user-driven way.
+Next up we'll look into `extending the language with control flow
+constructs <LangImpl5.html>`_, tackling some interesting LLVM IR issues
+along the way.
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example, enhanced with
+the LLVM JIT and optimizer. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core mcjit native` -O3 -o toy
+    # Run
+    ./toy
+
+If you are compiling this on Linux, make sure to add the "-rdynamic"
+option as well. This makes sure that the external functions are resolved
+properly at runtime.
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter4/toy.cpp
+   :language: c++
+
+`Next: Extending the language: control flow <LangImpl05.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl05.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl05.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl05.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl05.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,790 @@
+==================================================
+Kaleidoscope: Extending the Language: Control Flow
+==================================================
+
+.. contents::
+   :local:
+
+Chapter 5 Introduction
+======================
+
+Welcome to Chapter 5 of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. Parts 1-4 described the implementation of
+the simple Kaleidoscope language and included support for generating
+LLVM IR, followed by optimizations and a JIT compiler. Unfortunately, as
+presented, Kaleidoscope is mostly useless: it has no control flow other
+than call and return. This means that you can't have conditional
+branches in the code, significantly limiting its power. In this episode
+of "build that compiler", we'll extend Kaleidoscope to have an
+if/then/else expression plus a simple 'for' loop.
+
+If/Then/Else
+============
+
+Extending Kaleidoscope to support if/then/else is quite straightforward.
+It basically requires adding support for this "new" concept to the
+lexer, parser, AST, and LLVM code emitter. This example is nice, because
+it shows how easy it is to "grow" a language over time, incrementally
+extending it as new ideas are discovered.
+
+Before we get going on "how" we add this extension, lets talk about
+"what" we want. The basic idea is that we want to be able to write this
+sort of thing:
+
+::
+
+    def fib(x)
+      if x < 3 then
+        1
+      else
+        fib(x-1)+fib(x-2);
+
+In Kaleidoscope, every construct is an expression: there are no
+statements. As such, the if/then/else expression needs to return a value
+like any other. Since we're using a mostly functional form, we'll have
+it evaluate its conditional, then return the 'then' or 'else' value
+based on how the condition was resolved. This is very similar to the C
+"?:" expression.
+
+The semantics of the if/then/else expression is that it evaluates the
+condition to a boolean equality value: 0.0 is considered to be false and
+everything else is considered to be true. If the condition is true, the
+first subexpression is evaluated and returned, if the condition is
+false, the second subexpression is evaluated and returned. Since
+Kaleidoscope allows side-effects, this behavior is important to nail
+down.
+
+Now that we know what we "want", lets break this down into its
+constituent pieces.
+
+Lexer Extensions for If/Then/Else
+---------------------------------
+
+The lexer extensions are straightforward. First we add new enum values
+for the relevant tokens:
+
+.. code-block:: c++
+
+      // control
+      tok_if = -6,
+      tok_then = -7,
+      tok_else = -8,
+
+Once we have that, we recognize the new keywords in the lexer. This is
+pretty simple stuff:
+
+.. code-block:: c++
+
+        ...
+        if (IdentifierStr == "def")
+          return tok_def;
+        if (IdentifierStr == "extern")
+          return tok_extern;
+        if (IdentifierStr == "if")
+          return tok_if;
+        if (IdentifierStr == "then")
+          return tok_then;
+        if (IdentifierStr == "else")
+          return tok_else;
+        return tok_identifier;
+
+AST Extensions for If/Then/Else
+-------------------------------
+
+To represent the new expression we add a new AST node for it:
+
+.. code-block:: c++
+
+    /// IfExprAST - Expression class for if/then/else.
+    class IfExprAST : public ExprAST {
+      std::unique_ptr<ExprAST> Cond, Then, Else;
+
+    public:
+      IfExprAST(std::unique_ptr<ExprAST> Cond, std::unique_ptr<ExprAST> Then,
+                std::unique_ptr<ExprAST> Else)
+        : Cond(std::move(Cond)), Then(std::move(Then)), Else(std::move(Else)) {}
+      virtual Value *codegen();
+    };
+
+The AST node just has pointers to the various subexpressions.
+
+Parser Extensions for If/Then/Else
+----------------------------------
+
+Now that we have the relevant tokens coming from the lexer and we have
+the AST node to build, our parsing logic is relatively straightforward.
+First we define a new parsing function:
+
+.. code-block:: c++
+
+    /// ifexpr ::= 'if' expression 'then' expression 'else' expression
+    static std::unique_ptr<ExprAST> ParseIfExpr() {
+      getNextToken();  // eat the if.
+
+      // condition.
+      auto Cond = ParseExpression();
+      if (!Cond)
+        return nullptr;
+
+      if (CurTok != tok_then)
+        return LogError("expected then");
+      getNextToken();  // eat the then
+
+      auto Then = ParseExpression();
+      if (!Then)
+        return nullptr;
+
+      if (CurTok != tok_else)
+        return LogError("expected else");
+
+      getNextToken();
+
+      auto Else = ParseExpression();
+      if (!Else)
+        return nullptr;
+
+      return llvm::make_unique<IfExprAST>(std::move(Cond), std::move(Then),
+                                          std::move(Else));
+    }
+
+Next we hook it up as a primary expression:
+
+.. code-block:: c++
+
+    static std::unique_ptr<ExprAST> ParsePrimary() {
+      switch (CurTok) {
+      default:
+        return LogError("unknown token when expecting an expression");
+      case tok_identifier:
+        return ParseIdentifierExpr();
+      case tok_number:
+        return ParseNumberExpr();
+      case '(':
+        return ParseParenExpr();
+      case tok_if:
+        return ParseIfExpr();
+      }
+    }
+
+LLVM IR for If/Then/Else
+------------------------
+
+Now that we have it parsing and building the AST, the final piece is
+adding LLVM code generation support. This is the most interesting part
+of the if/then/else example, because this is where it starts to
+introduce new concepts. All of the code above has been thoroughly
+described in previous chapters.
+
+To motivate the code we want to produce, lets take a look at a simple
+example. Consider:
+
+::
+
+    extern foo();
+    extern bar();
+    def baz(x) if x then foo() else bar();
+
+If you disable optimizations, the code you'll (soon) get from
+Kaleidoscope looks like this:
+
+.. code-block:: llvm
+
+    declare double @foo()
+
+    declare double @bar()
+
+    define double @baz(double %x) {
+    entry:
+      %ifcond = fcmp one double %x, 0.000000e+00
+      br i1 %ifcond, label %then, label %else
+
+    then:       ; preds = %entry
+      %calltmp = call double @foo()
+      br label %ifcont
+
+    else:       ; preds = %entry
+      %calltmp1 = call double @bar()
+      br label %ifcont
+
+    ifcont:     ; preds = %else, %then
+      %iftmp = phi double [ %calltmp, %then ], [ %calltmp1, %else ]
+      ret double %iftmp
+    }
+
+To visualize the control flow graph, you can use a nifty feature of the
+LLVM '`opt <http://llvm.org/cmds/opt.html>`_' tool. If you put this LLVM
+IR into "t.ll" and run "``llvm-as < t.ll | opt -analyze -view-cfg``", `a
+window will pop up <../ProgrammersManual.html#viewing-graphs-while-debugging-code>`_ and you'll
+see this graph:
+
+.. figure:: LangImpl05-cfg.png
+   :align: center
+   :alt: Example CFG
+
+   Example CFG
+
+Another way to get this is to call "``F->viewCFG()``" or
+"``F->viewCFGOnly()``" (where F is a "``Function*``") either by
+inserting actual calls into the code and recompiling or by calling these
+in the debugger. LLVM has many nice features for visualizing various
+graphs.
+
+Getting back to the generated code, it is fairly simple: the entry block
+evaluates the conditional expression ("x" in our case here) and compares
+the result to 0.0 with the "``fcmp one``" instruction ('one' is "Ordered
+and Not Equal"). Based on the result of this expression, the code jumps
+to either the "then" or "else" blocks, which contain the expressions for
+the true/false cases.
+
+Once the then/else blocks are finished executing, they both branch back
+to the 'ifcont' block to execute the code that happens after the
+if/then/else. In this case the only thing left to do is to return to the
+caller of the function. The question then becomes: how does the code
+know which expression to return?
+
+The answer to this question involves an important SSA operation: the
+`Phi
+operation <http://en.wikipedia.org/wiki/Static_single_assignment_form>`_.
+If you're not familiar with SSA, `the wikipedia
+article <http://en.wikipedia.org/wiki/Static_single_assignment_form>`_
+is a good introduction and there are various other introductions to it
+available on your favorite search engine. The short version is that
+"execution" of the Phi operation requires "remembering" which block
+control came from. The Phi operation takes on the value corresponding to
+the input control block. In this case, if control comes in from the
+"then" block, it gets the value of "calltmp". If control comes from the
+"else" block, it gets the value of "calltmp1".
+
+At this point, you are probably starting to think "Oh no! This means my
+simple and elegant front-end will have to start generating SSA form in
+order to use LLVM!". Fortunately, this is not the case, and we strongly
+advise *not* implementing an SSA construction algorithm in your
+front-end unless there is an amazingly good reason to do so. In
+practice, there are two sorts of values that float around in code
+written for your average imperative programming language that might need
+Phi nodes:
+
+#. Code that involves user variables: ``x = 1; x = x + 1;``
+#. Values that are implicit in the structure of your AST, such as the
+   Phi node in this case.
+
+In `Chapter 7 <LangImpl7.html>`_ of this tutorial ("mutable variables"),
+we'll talk about #1 in depth. For now, just believe me that you don't
+need SSA construction to handle this case. For #2, you have the choice
+of using the techniques that we will describe for #1, or you can insert
+Phi nodes directly, if convenient. In this case, it is really
+easy to generate the Phi node, so we choose to do it directly.
+
+Okay, enough of the motivation and overview, lets generate code!
+
+Code Generation for If/Then/Else
+--------------------------------
+
+In order to generate code for this, we implement the ``codegen`` method
+for ``IfExprAST``:
+
+.. code-block:: c++
+
+    Value *IfExprAST::codegen() {
+      Value *CondV = Cond->codegen();
+      if (!CondV)
+        return nullptr;
+
+      // Convert condition to a bool by comparing equal to 0.0.
+      CondV = Builder.CreateFCmpONE(
+          CondV, ConstantFP::get(LLVMContext, APFloat(0.0)), "ifcond");
+
+This code is straightforward and similar to what we saw before. We emit
+the expression for the condition, then compare that value to zero to get
+a truth value as a 1-bit (bool) value.
+
+.. code-block:: c++
+
+      Function *TheFunction = Builder.GetInsertBlock()->getParent();
+
+      // Create blocks for the then and else cases.  Insert the 'then' block at the
+      // end of the function.
+      BasicBlock *ThenBB =
+          BasicBlock::Create(LLVMContext, "then", TheFunction);
+      BasicBlock *ElseBB = BasicBlock::Create(LLVMContext, "else");
+      BasicBlock *MergeBB = BasicBlock::Create(LLVMContext, "ifcont");
+
+      Builder.CreateCondBr(CondV, ThenBB, ElseBB);
+
+This code creates the basic blocks that are related to the if/then/else
+statement, and correspond directly to the blocks in the example above.
+The first line gets the current Function object that is being built. It
+gets this by asking the builder for the current BasicBlock, and asking
+that block for its "parent" (the function it is currently embedded
+into).
+
+Once it has that, it creates three blocks. Note that it passes
+"TheFunction" into the constructor for the "then" block. This causes the
+constructor to automatically insert the new block into the end of the
+specified function. The other two blocks are created, but aren't yet
+inserted into the function.
+
+Once the blocks are created, we can emit the conditional branch that
+chooses between them. Note that creating new blocks does not implicitly
+affect the IRBuilder, so it is still inserting into the block that the
+condition went into. Also note that it is creating a branch to the
+"then" block and the "else" block, even though the "else" block isn't
+inserted into the function yet. This is all ok: it is the standard way
+that LLVM supports forward references.
+
+.. code-block:: c++
+
+      // Emit then value.
+      Builder.SetInsertPoint(ThenBB);
+
+      Value *ThenV = Then->codegen();
+      if (!ThenV)
+        return nullptr;
+
+      Builder.CreateBr(MergeBB);
+      // Codegen of 'Then' can change the current block, update ThenBB for the PHI.
+      ThenBB = Builder.GetInsertBlock();
+
+After the conditional branch is inserted, we move the builder to start
+inserting into the "then" block. Strictly speaking, this call moves the
+insertion point to be at the end of the specified block. However, since
+the "then" block is empty, it also starts out by inserting at the
+beginning of the block. :)
+
+Once the insertion point is set, we recursively codegen the "then"
+expression from the AST. To finish off the "then" block, we create an
+unconditional branch to the merge block. One interesting (and very
+important) aspect of the LLVM IR is that it `requires all basic blocks
+to be "terminated" <../LangRef.html#functionstructure>`_ with a `control
+flow instruction <../LangRef.html#terminators>`_ such as return or
+branch. This means that all control flow, *including fall throughs* must
+be made explicit in the LLVM IR. If you violate this rule, the verifier
+will emit an error.
+
+The final line here is quite subtle, but is very important. The basic
+issue is that when we create the Phi node in the merge block, we need to
+set up the block/value pairs that indicate how the Phi will work.
+Importantly, the Phi node expects to have an entry for each predecessor
+of the block in the CFG. Why then, are we getting the current block when
+we just set it to ThenBB 5 lines above? The problem is that the "Then"
+expression may actually itself change the block that the Builder is
+emitting into if, for example, it contains a nested "if/then/else"
+expression. Because calling ``codegen()`` recursively could arbitrarily change
+the notion of the current block, we are required to get an up-to-date
+value for code that will set up the Phi node.
+
+.. code-block:: c++
+
+      // Emit else block.
+      TheFunction->getBasicBlockList().push_back(ElseBB);
+      Builder.SetInsertPoint(ElseBB);
+
+      Value *ElseV = Else->codegen();
+      if (!ElseV)
+        return nullptr;
+
+      Builder.CreateBr(MergeBB);
+      // codegen of 'Else' can change the current block, update ElseBB for the PHI.
+      ElseBB = Builder.GetInsertBlock();
+
+Code generation for the 'else' block is basically identical to codegen
+for the 'then' block. The only significant difference is the first line,
+which adds the 'else' block to the function. Recall previously that the
+'else' block was created, but not added to the function. Now that the
+'then' and 'else' blocks are emitted, we can finish up with the merge
+code:
+
+.. code-block:: c++
+
+      // Emit merge block.
+      TheFunction->getBasicBlockList().push_back(MergeBB);
+      Builder.SetInsertPoint(MergeBB);
+      PHINode *PN =
+        Builder.CreatePHI(Type::getDoubleTy(LLVMContext), 2, "iftmp");
+
+      PN->addIncoming(ThenV, ThenBB);
+      PN->addIncoming(ElseV, ElseBB);
+      return PN;
+    }
+
+The first two lines here are now familiar: the first adds the "merge"
+block to the Function object (it was previously floating, like the else
+block above). The second changes the insertion point so that newly
+created code will go into the "merge" block. Once that is done, we need
+to create the PHI node and set up the block/value pairs for the PHI.
+
+Finally, the CodeGen function returns the phi node as the value computed
+by the if/then/else expression. In our example above, this returned
+value will feed into the code for the top-level function, which will
+create the return instruction.
+
+Overall, we now have the ability to execute conditional code in
+Kaleidoscope. With this extension, Kaleidoscope is a fairly complete
+language that can calculate a wide variety of numeric functions. Next up
+we'll add another useful expression that is familiar from non-functional
+languages...
+
+'for' Loop Expression
+=====================
+
+Now that we know how to add basic control flow constructs to the
+language, we have the tools to add more powerful things. Lets add
+something more aggressive, a 'for' expression:
+
+::
+
+     extern putchard(char)
+     def printstar(n)
+       for i = 1, i < n, 1.0 in
+         putchard(42);  # ascii 42 = '*'
+
+     # print 100 '*' characters
+     printstar(100);
+
+This expression defines a new variable ("i" in this case) which iterates
+from a starting value, while the condition ("i < n" in this case) is
+true, incrementing by an optional step value ("1.0" in this case). If
+the step value is omitted, it defaults to 1.0. While the loop is true,
+it executes its body expression. Because we don't have anything better
+to return, we'll just define the loop as always returning 0.0. In the
+future when we have mutable variables, it will get more useful.
+
+As before, lets talk about the changes that we need to Kaleidoscope to
+support this.
+
+Lexer Extensions for the 'for' Loop
+-----------------------------------
+
+The lexer extensions are the same sort of thing as for if/then/else:
+
+.. code-block:: c++
+
+      ... in enum Token ...
+      // control
+      tok_if = -6, tok_then = -7, tok_else = -8,
+      tok_for = -9, tok_in = -10
+
+      ... in gettok ...
+      if (IdentifierStr == "def")
+        return tok_def;
+      if (IdentifierStr == "extern")
+        return tok_extern;
+      if (IdentifierStr == "if")
+        return tok_if;
+      if (IdentifierStr == "then")
+        return tok_then;
+      if (IdentifierStr == "else")
+        return tok_else;
+      if (IdentifierStr == "for")
+        return tok_for;
+      if (IdentifierStr == "in")
+        return tok_in;
+      return tok_identifier;
+
+AST Extensions for the 'for' Loop
+---------------------------------
+
+The AST node is just as simple. It basically boils down to capturing the
+variable name and the constituent expressions in the node.
+
+.. code-block:: c++
+
+    /// ForExprAST - Expression class for for/in.
+    class ForExprAST : public ExprAST {
+      std::string VarName;
+      std::unique_ptr<ExprAST> Start, End, Step, Body;
+
+    public:
+      ForExprAST(const std::string &VarName, std::unique_ptr<ExprAST> Start,
+                 std::unique_ptr<ExprAST> End, std::unique_ptr<ExprAST> Step,
+                 std::unique_ptr<ExprAST> Body)
+        : VarName(VarName), Start(std::move(Start)), End(std::move(End)),
+          Step(std::move(Step)), Body(std::move(Body)) {}
+      virtual Value *codegen();
+    };
+
+Parser Extensions for the 'for' Loop
+------------------------------------
+
+The parser code is also fairly standard. The only interesting thing here
+is handling of the optional step value. The parser code handles it by
+checking to see if the second comma is present. If not, it sets the step
+value to null in the AST node:
+
+.. code-block:: c++
+
+    /// forexpr ::= 'for' identifier '=' expr ',' expr (',' expr)? 'in' expression
+    static std::unique_ptr<ExprAST> ParseForExpr() {
+      getNextToken();  // eat the for.
+
+      if (CurTok != tok_identifier)
+        return LogError("expected identifier after for");
+
+      std::string IdName = IdentifierStr;
+      getNextToken();  // eat identifier.
+
+      if (CurTok != '=')
+        return LogError("expected '=' after for");
+      getNextToken();  // eat '='.
+
+
+      auto Start = ParseExpression();
+      if (!Start)
+        return nullptr;
+      if (CurTok != ',')
+        return LogError("expected ',' after for start value");
+      getNextToken();
+
+      auto End = ParseExpression();
+      if (!End)
+        return nullptr;
+
+      // The step value is optional.
+      std::unique_ptr<ExprAST> Step;
+      if (CurTok == ',') {
+        getNextToken();
+        Step = ParseExpression();
+        if (!Step)
+          return nullptr;
+      }
+
+      if (CurTok != tok_in)
+        return LogError("expected 'in' after for");
+      getNextToken();  // eat 'in'.
+
+      auto Body = ParseExpression();
+      if (!Body)
+        return nullptr;
+
+      return llvm::make_unique<ForExprAST>(IdName, std::move(Start),
+                                           std::move(End), std::move(Step),
+                                           std::move(Body));
+    }
+
+LLVM IR for the 'for' Loop
+--------------------------
+
+Now we get to the good part: the LLVM IR we want to generate for this
+thing. With the simple example above, we get this LLVM IR (note that
+this dump is generated with optimizations disabled for clarity):
+
+.. code-block:: llvm
+
+    declare double @putchard(double)
+
+    define double @printstar(double %n) {
+    entry:
+      ; initial value = 1.0 (inlined into phi)
+      br label %loop
+
+    loop:       ; preds = %loop, %entry
+      %i = phi double [ 1.000000e+00, %entry ], [ %nextvar, %loop ]
+      ; body
+      %calltmp = call double @putchard(double 4.200000e+01)
+      ; increment
+      %nextvar = fadd double %i, 1.000000e+00
+
+      ; termination test
+      %cmptmp = fcmp ult double %i, %n
+      %booltmp = uitofp i1 %cmptmp to double
+      %loopcond = fcmp one double %booltmp, 0.000000e+00
+      br i1 %loopcond, label %loop, label %afterloop
+
+    afterloop:      ; preds = %loop
+      ; loop always returns 0.0
+      ret double 0.000000e+00
+    }
+
+This loop contains all the same constructs we saw before: a phi node,
+several expressions, and some basic blocks. Lets see how this fits
+together.
+
+Code Generation for the 'for' Loop
+----------------------------------
+
+The first part of codegen is very simple: we just output the start
+expression for the loop value:
+
+.. code-block:: c++
+
+    Value *ForExprAST::codegen() {
+      // Emit the start code first, without 'variable' in scope.
+      Value *StartVal = Start->codegen();
+      if (StartVal == 0) return 0;
+
+With this out of the way, the next step is to set up the LLVM basic
+block for the start of the loop body. In the case above, the whole loop
+body is one block, but remember that the body code itself could consist
+of multiple blocks (e.g. if it contains an if/then/else or a for/in
+expression).
+
+.. code-block:: c++
+
+      // Make the new basic block for the loop header, inserting after current
+      // block.
+      Function *TheFunction = Builder.GetInsertBlock()->getParent();
+      BasicBlock *PreheaderBB = Builder.GetInsertBlock();
+      BasicBlock *LoopBB =
+          BasicBlock::Create(LLVMContext, "loop", TheFunction);
+
+      // Insert an explicit fall through from the current block to the LoopBB.
+      Builder.CreateBr(LoopBB);
+
+This code is similar to what we saw for if/then/else. Because we will
+need it to create the Phi node, we remember the block that falls through
+into the loop. Once we have that, we create the actual block that starts
+the loop and create an unconditional branch for the fall-through between
+the two blocks.
+
+.. code-block:: c++
+
+      // Start insertion in LoopBB.
+      Builder.SetInsertPoint(LoopBB);
+
+      // Start the PHI node with an entry for Start.
+      PHINode *Variable = Builder.CreatePHI(Type::getDoubleTy(LLVMContext),
+                                            2, VarName.c_str());
+      Variable->addIncoming(StartVal, PreheaderBB);
+
+Now that the "preheader" for the loop is set up, we switch to emitting
+code for the loop body. To begin with, we move the insertion point and
+create the PHI node for the loop induction variable. Since we already
+know the incoming value for the starting value, we add it to the Phi
+node. Note that the Phi will eventually get a second value for the
+backedge, but we can't set it up yet (because it doesn't exist!).
+
+.. code-block:: c++
+
+      // Within the loop, the variable is defined equal to the PHI node.  If it
+      // shadows an existing variable, we have to restore it, so save it now.
+      Value *OldVal = NamedValues[VarName];
+      NamedValues[VarName] = Variable;
+
+      // Emit the body of the loop.  This, like any other expr, can change the
+      // current BB.  Note that we ignore the value computed by the body, but don't
+      // allow an error.
+      if (!Body->codegen())
+        return nullptr;
+
+Now the code starts to get more interesting. Our 'for' loop introduces a
+new variable to the symbol table. This means that our symbol table can
+now contain either function arguments or loop variables. To handle this,
+before we codegen the body of the loop, we add the loop variable as the
+current value for its name. Note that it is possible that there is a
+variable of the same name in the outer scope. It would be easy to make
+this an error (emit an error and return null if there is already an
+entry for VarName) but we choose to allow shadowing of variables. In
+order to handle this correctly, we remember the Value that we are
+potentially shadowing in ``OldVal`` (which will be null if there is no
+shadowed variable).
+
+Once the loop variable is set into the symbol table, the code
+recursively codegen's the body. This allows the body to use the loop
+variable: any references to it will naturally find it in the symbol
+table.
+
+.. code-block:: c++
+
+      // Emit the step value.
+      Value *StepVal = nullptr;
+      if (Step) {
+        StepVal = Step->codegen();
+        if (!StepVal)
+          return nullptr;
+      } else {
+        // If not specified, use 1.0.
+        StepVal = ConstantFP::get(LLVMContext, APFloat(1.0));
+      }
+
+      Value *NextVar = Builder.CreateFAdd(Variable, StepVal, "nextvar");
+
+Now that the body is emitted, we compute the next value of the iteration
+variable by adding the step value, or 1.0 if it isn't present.
+'``NextVar``' will be the value of the loop variable on the next
+iteration of the loop.
+
+.. code-block:: c++
+
+      // Compute the end condition.
+      Value *EndCond = End->codegen();
+      if (!EndCond)
+        return nullptr;
+
+      // Convert condition to a bool by comparing equal to 0.0.
+      EndCond = Builder.CreateFCmpONE(
+          EndCond, ConstantFP::get(LLVMContext, APFloat(0.0)), "loopcond");
+
+Finally, we evaluate the exit value of the loop, to determine whether
+the loop should exit. This mirrors the condition evaluation for the
+if/then/else statement.
+
+.. code-block:: c++
+
+      // Create the "after loop" block and insert it.
+      BasicBlock *LoopEndBB = Builder.GetInsertBlock();
+      BasicBlock *AfterBB =
+          BasicBlock::Create(LLVMContext, "afterloop", TheFunction);
+
+      // Insert the conditional branch into the end of LoopEndBB.
+      Builder.CreateCondBr(EndCond, LoopBB, AfterBB);
+
+      // Any new code will be inserted in AfterBB.
+      Builder.SetInsertPoint(AfterBB);
+
+With the code for the body of the loop complete, we just need to finish
+up the control flow for it. This code remembers the end block (for the
+phi node), then creates the block for the loop exit ("afterloop"). Based
+on the value of the exit condition, it creates a conditional branch that
+chooses between executing the loop again and exiting the loop. Any
+future code is emitted in the "afterloop" block, so it sets the
+insertion position to it.
+
+.. code-block:: c++
+
+      // Add a new entry to the PHI node for the backedge.
+      Variable->addIncoming(NextVar, LoopEndBB);
+
+      // Restore the unshadowed variable.
+      if (OldVal)
+        NamedValues[VarName] = OldVal;
+      else
+        NamedValues.erase(VarName);
+
+      // for expr always returns 0.0.
+      return Constant::getNullValue(Type::getDoubleTy(LLVMContext));
+    }
+
+The final code handles various cleanups: now that we have the "NextVar"
+value, we can add the incoming value to the loop PHI node. After that,
+we remove the loop variable from the symbol table, so that it isn't in
+scope after the for loop. Finally, code generation of the for loop
+always returns 0.0, so that is what we return from
+``ForExprAST::codegen()``.
+
+With this, we conclude the "adding control flow to Kaleidoscope" chapter
+of the tutorial. In this chapter we added two control flow constructs,
+and used them to motivate a couple of aspects of the LLVM IR that are
+important for front-end implementors to know. In the next chapter of our
+saga, we will get a bit crazier and add `user-defined
+operators <LangImpl6.html>`_ to our poor innocent language.
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example, enhanced with
+the if/then/else and for expressions.. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core mcjit native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter5/toy.cpp
+   :language: c++
+
+`Next: Extending the language: user-defined operators <LangImpl06.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl06.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl06.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl06.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl06.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,768 @@
+============================================================
+Kaleidoscope: Extending the Language: User-defined Operators
+============================================================
+
+.. contents::
+   :local:
+
+Chapter 6 Introduction
+======================
+
+Welcome to Chapter 6 of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. At this point in our tutorial, we now
+have a fully functional language that is fairly minimal, but also
+useful. There is still one big problem with it, however. Our language
+doesn't have many useful operators (like division, logical negation, or
+even any comparisons besides less-than).
+
+This chapter of the tutorial takes a wild digression into adding
+user-defined operators to the simple and beautiful Kaleidoscope
+language. This digression now gives us a simple and ugly language in
+some ways, but also a powerful one at the same time. One of the great
+things about creating your own language is that you get to decide what
+is good or bad. In this tutorial we'll assume that it is okay to use
+this as a way to show some interesting parsing techniques.
+
+At the end of this tutorial, we'll run through an example Kaleidoscope
+application that `renders the Mandelbrot set <#kicking-the-tires>`_. This gives an
+example of what you can build with Kaleidoscope and its feature set.
+
+User-defined Operators: the Idea
+================================
+
+The "operator overloading" that we will add to Kaleidoscope is more
+general than languages like C++. In C++, you are only allowed to
+redefine existing operators: you can't programmatically change the
+grammar, introduce new operators, change precedence levels, etc. In this
+chapter, we will add this capability to Kaleidoscope, which will let the
+user round out the set of operators that are supported.
+
+The point of going into user-defined operators in a tutorial like this
+is to show the power and flexibility of using a hand-written parser.
+Thus far, the parser we have been implementing uses recursive descent
+for most parts of the grammar and operator precedence parsing for the
+expressions. See `Chapter 2 <LangImpl2.html>`_ for details. Without
+using operator precedence parsing, it would be very difficult to allow
+the programmer to introduce new operators into the grammar: the grammar
+is dynamically extensible as the JIT runs.
+
+The two specific features we'll add are programmable unary operators
+(right now, Kaleidoscope has no unary operators at all) as well as
+binary operators. An example of this is:
+
+::
+
+    # Logical unary not.
+    def unary!(v)
+      if v then
+        0
+      else
+        1;
+
+    # Define > with the same precedence as <.
+    def binary> 10 (LHS RHS)
+      RHS < LHS;
+
+    # Binary "logical or", (note that it does not "short circuit")
+    def binary| 5 (LHS RHS)
+      if LHS then
+        1
+      else if RHS then
+        1
+      else
+        0;
+
+    # Define = with slightly lower precedence than relationals.
+    def binary= 9 (LHS RHS)
+      !(LHS < RHS | LHS > RHS);
+
+Many languages aspire to being able to implement their standard runtime
+library in the language itself. In Kaleidoscope, we can implement
+significant parts of the language in the library!
+
+We will break down implementation of these features into two parts:
+implementing support for user-defined binary operators and adding unary
+operators.
+
+User-defined Binary Operators
+=============================
+
+Adding support for user-defined binary operators is pretty simple with
+our current framework. We'll first add support for the unary/binary
+keywords:
+
+.. code-block:: c++
+
+    enum Token {
+      ...
+      // operators
+      tok_binary = -11,
+      tok_unary = -12
+    };
+    ...
+    static int gettok() {
+    ...
+        if (IdentifierStr == "for")
+          return tok_for;
+        if (IdentifierStr == "in")
+          return tok_in;
+        if (IdentifierStr == "binary")
+          return tok_binary;
+        if (IdentifierStr == "unary")
+          return tok_unary;
+        return tok_identifier;
+
+This just adds lexer support for the unary and binary keywords, like we
+did in `previous chapters <LangImpl5.html#lexer-extensions-for-if-then-else>`_. One nice thing
+about our current AST, is that we represent binary operators with full
+generalisation by using their ASCII code as the opcode. For our extended
+operators, we'll use this same representation, so we don't need any new
+AST or parser support.
+
+On the other hand, we have to be able to represent the definitions of
+these new operators, in the "def binary\| 5" part of the function
+definition. In our grammar so far, the "name" for the function
+definition is parsed as the "prototype" production and into the
+``PrototypeAST`` AST node. To represent our new user-defined operators
+as prototypes, we have to extend the ``PrototypeAST`` AST node like
+this:
+
+.. code-block:: c++
+
+    /// PrototypeAST - This class represents the "prototype" for a function,
+    /// which captures its argument names as well as if it is an operator.
+    class PrototypeAST {
+      std::string Name;
+      std::vector<std::string> Args;
+      bool IsOperator;
+      unsigned Precedence;  // Precedence if a binary op.
+
+    public:
+      PrototypeAST(const std::string &name, std::vector<std::string> Args,
+                   bool IsOperator = false, unsigned Prec = 0)
+      : Name(name), Args(std::move(Args)), IsOperator(IsOperator),
+        Precedence(Prec) {}
+
+      bool isUnaryOp() const { return IsOperator && Args.size() == 1; }
+      bool isBinaryOp() const { return IsOperator && Args.size() == 2; }
+
+      char getOperatorName() const {
+        assert(isUnaryOp() || isBinaryOp());
+        return Name[Name.size()-1];
+      }
+
+      unsigned getBinaryPrecedence() const { return Precedence; }
+
+      Function *codegen();
+    };
+
+Basically, in addition to knowing a name for the prototype, we now keep
+track of whether it was an operator, and if it was, what precedence
+level the operator is at. The precedence is only used for binary
+operators (as you'll see below, it just doesn't apply for unary
+operators). Now that we have a way to represent the prototype for a
+user-defined operator, we need to parse it:
+
+.. code-block:: c++
+
+    /// prototype
+    ///   ::= id '(' id* ')'
+    ///   ::= binary LETTER number? (id, id)
+    static std::unique_ptr<PrototypeAST> ParsePrototype() {
+      std::string FnName;
+
+      unsigned Kind = 0;  // 0 = identifier, 1 = unary, 2 = binary.
+      unsigned BinaryPrecedence = 30;
+
+      switch (CurTok) {
+      default:
+        return LogErrorP("Expected function name in prototype");
+      case tok_identifier:
+        FnName = IdentifierStr;
+        Kind = 0;
+        getNextToken();
+        break;
+      case tok_binary:
+        getNextToken();
+        if (!isascii(CurTok))
+          return LogErrorP("Expected binary operator");
+        FnName = "binary";
+        FnName += (char)CurTok;
+        Kind = 2;
+        getNextToken();
+
+        // Read the precedence if present.
+        if (CurTok == tok_number) {
+          if (NumVal < 1 || NumVal > 100)
+            return LogErrorP("Invalid precedecnce: must be 1..100");
+          BinaryPrecedence = (unsigned)NumVal;
+          getNextToken();
+        }
+        break;
+      }
+
+      if (CurTok != '(')
+        return LogErrorP("Expected '(' in prototype");
+
+      std::vector<std::string> ArgNames;
+      while (getNextToken() == tok_identifier)
+        ArgNames.push_back(IdentifierStr);
+      if (CurTok != ')')
+        return LogErrorP("Expected ')' in prototype");
+
+      // success.
+      getNextToken();  // eat ')'.
+
+      // Verify right number of names for operator.
+      if (Kind && ArgNames.size() != Kind)
+        return LogErrorP("Invalid number of operands for operator");
+
+      return llvm::make_unique<PrototypeAST>(FnName, std::move(ArgNames), Kind != 0,
+                                             BinaryPrecedence);
+    }
+
+This is all fairly straightforward parsing code, and we have already
+seen a lot of similar code in the past. One interesting part about the
+code above is the couple lines that set up ``FnName`` for binary
+operators. This builds names like "binary@" for a newly defined "@"
+operator. This then takes advantage of the fact that symbol names in the
+LLVM symbol table are allowed to have any character in them, including
+embedded nul characters.
+
+The next interesting thing to add, is codegen support for these binary
+operators. Given our current structure, this is a simple addition of a
+default case for our existing binary operator node:
+
+.. code-block:: c++
+
+    Value *BinaryExprAST::codegen() {
+      Value *L = LHS->codegen();
+      Value *R = RHS->codegen();
+      if (!L || !R)
+        return nullptr;
+
+      switch (Op) {
+      case '+':
+        return Builder.CreateFAdd(L, R, "addtmp");
+      case '-':
+        return Builder.CreateFSub(L, R, "subtmp");
+      case '*':
+        return Builder.CreateFMul(L, R, "multmp");
+      case '<':
+        L = Builder.CreateFCmpULT(L, R, "cmptmp");
+        // Convert bool 0/1 to double 0.0 or 1.0
+        return Builder.CreateUIToFP(L, Type::getDoubleTy(LLVMContext),
+                                    "booltmp");
+      default:
+        break;
+      }
+
+      // If it wasn't a builtin binary operator, it must be a user defined one. Emit
+      // a call to it.
+      Function *F = TheModule->getFunction(std::string("binary") + Op);
+      assert(F && "binary operator not found!");
+
+      Value *Ops[2] = { L, R };
+      return Builder.CreateCall(F, Ops, "binop");
+    }
+
+As you can see above, the new code is actually really simple. It just
+does a lookup for the appropriate operator in the symbol table and
+generates a function call to it. Since user-defined operators are just
+built as normal functions (because the "prototype" boils down to a
+function with the right name) everything falls into place.
+
+The final piece of code we are missing, is a bit of top-level magic:
+
+.. code-block:: c++
+
+    Function *FunctionAST::codegen() {
+      NamedValues.clear();
+
+      Function *TheFunction = Proto->codegen();
+      if (!TheFunction)
+        return nullptr;
+
+      // If this is an operator, install it.
+      if (Proto->isBinaryOp())
+        BinopPrecedence[Proto->getOperatorName()] = Proto->getBinaryPrecedence();
+
+      // Create a new basic block to start insertion into.
+      BasicBlock *BB = BasicBlock::Create(LLVMContext, "entry", TheFunction);
+      Builder.SetInsertPoint(BB);
+
+      if (Value *RetVal = Body->codegen()) {
+        ...
+
+Basically, before codegening a function, if it is a user-defined
+operator, we register it in the precedence table. This allows the binary
+operator parsing logic we already have in place to handle it. Since we
+are working on a fully-general operator precedence parser, this is all
+we need to do to "extend the grammar".
+
+Now we have useful user-defined binary operators. This builds a lot on
+the previous framework we built for other operators. Adding unary
+operators is a bit more challenging, because we don't have any framework
+for it yet - lets see what it takes.
+
+User-defined Unary Operators
+============================
+
+Since we don't currently support unary operators in the Kaleidoscope
+language, we'll need to add everything to support them. Above, we added
+simple support for the 'unary' keyword to the lexer. In addition to
+that, we need an AST node:
+
+.. code-block:: c++
+
+    /// UnaryExprAST - Expression class for a unary operator.
+    class UnaryExprAST : public ExprAST {
+      char Opcode;
+      std::unique_ptr<ExprAST> Operand;
+
+    public:
+      UnaryExprAST(char Opcode, std::unique_ptr<ExprAST> Operand)
+        : Opcode(Opcode), Operand(std::move(Operand)) {}
+      virtual Value *codegen();
+    };
+
+This AST node is very simple and obvious by now. It directly mirrors the
+binary operator AST node, except that it only has one child. With this,
+we need to add the parsing logic. Parsing a unary operator is pretty
+simple: we'll add a new function to do it:
+
+.. code-block:: c++
+
+    /// unary
+    ///   ::= primary
+    ///   ::= '!' unary
+    static std::unique_ptr<ExprAST> ParseUnary() {
+      // If the current token is not an operator, it must be a primary expr.
+      if (!isascii(CurTok) || CurTok == '(' || CurTok == ',')
+        return ParsePrimary();
+
+      // If this is a unary operator, read it.
+      int Opc = CurTok;
+      getNextToken();
+      if (auto Operand = ParseUnary())
+        return llvm::unique_ptr<UnaryExprAST>(Opc, std::move(Operand));
+      return nullptr;
+    }
+
+The grammar we add is pretty straightforward here. If we see a unary
+operator when parsing a primary operator, we eat the operator as a
+prefix and parse the remaining piece as another unary operator. This
+allows us to handle multiple unary operators (e.g. "!!x"). Note that
+unary operators can't have ambiguous parses like binary operators can,
+so there is no need for precedence information.
+
+The problem with this function, is that we need to call ParseUnary from
+somewhere. To do this, we change previous callers of ParsePrimary to
+call ParseUnary instead:
+
+.. code-block:: c++
+
+    /// binoprhs
+    ///   ::= ('+' unary)*
+    static std::unique_ptr<ExprAST> ParseBinOpRHS(int ExprPrec,
+                                                  std::unique_ptr<ExprAST> LHS) {
+      ...
+        // Parse the unary expression after the binary operator.
+        auto RHS = ParseUnary();
+        if (!RHS)
+          return nullptr;
+      ...
+    }
+    /// expression
+    ///   ::= unary binoprhs
+    ///
+    static std::unique_ptr<ExprAST> ParseExpression() {
+      auto LHS = ParseUnary();
+      if (!LHS)
+        return nullptr;
+
+      return ParseBinOpRHS(0, std::move(LHS));
+    }
+
+With these two simple changes, we are now able to parse unary operators
+and build the AST for them. Next up, we need to add parser support for
+prototypes, to parse the unary operator prototype. We extend the binary
+operator code above with:
+
+.. code-block:: c++
+
+    /// prototype
+    ///   ::= id '(' id* ')'
+    ///   ::= binary LETTER number? (id, id)
+    ///   ::= unary LETTER (id)
+    static std::unique_ptr<PrototypeAST> ParsePrototype() {
+      std::string FnName;
+
+      unsigned Kind = 0;  // 0 = identifier, 1 = unary, 2 = binary.
+      unsigned BinaryPrecedence = 30;
+
+      switch (CurTok) {
+      default:
+        return LogErrorP("Expected function name in prototype");
+      case tok_identifier:
+        FnName = IdentifierStr;
+        Kind = 0;
+        getNextToken();
+        break;
+      case tok_unary:
+        getNextToken();
+        if (!isascii(CurTok))
+          return LogErrorP("Expected unary operator");
+        FnName = "unary";
+        FnName += (char)CurTok;
+        Kind = 1;
+        getNextToken();
+        break;
+      case tok_binary:
+        ...
+
+As with binary operators, we name unary operators with a name that
+includes the operator character. This assists us at code generation
+time. Speaking of, the final piece we need to add is codegen support for
+unary operators. It looks like this:
+
+.. code-block:: c++
+
+    Value *UnaryExprAST::codegen() {
+      Value *OperandV = Operand->codegen();
+      if (!OperandV)
+        return nullptr;
+
+      Function *F = TheModule->getFunction(std::string("unary")+Opcode);
+      if (!F)
+        return LogErrorV("Unknown unary operator");
+
+      return Builder.CreateCall(F, OperandV, "unop");
+    }
+
+This code is similar to, but simpler than, the code for binary
+operators. It is simpler primarily because it doesn't need to handle any
+predefined operators.
+
+Kicking the Tires
+=================
+
+It is somewhat hard to believe, but with a few simple extensions we've
+covered in the last chapters, we have grown a real-ish language. With
+this, we can do a lot of interesting things, including I/O, math, and a
+bunch of other things. For example, we can now add a nice sequencing
+operator (printd is defined to print out the specified value and a
+newline):
+
+::
+
+    ready> extern printd(x);
+    Read extern:
+    declare double @printd(double)
+
+    ready> def binary : 1 (x y) 0;  # Low-precedence operator that ignores operands.
+    ..
+    ready> printd(123) : printd(456) : printd(789);
+    123.000000
+    456.000000
+    789.000000
+    Evaluated to 0.000000
+
+We can also define a bunch of other "primitive" operations, such as:
+
+::
+
+    # Logical unary not.
+    def unary!(v)
+      if v then
+        0
+      else
+        1;
+
+    # Unary negate.
+    def unary-(v)
+      0-v;
+
+    # Define > with the same precedence as <.
+    def binary> 10 (LHS RHS)
+      RHS < LHS;
+
+    # Binary logical or, which does not short circuit.
+    def binary| 5 (LHS RHS)
+      if LHS then
+        1
+      else if RHS then
+        1
+      else
+        0;
+
+    # Binary logical and, which does not short circuit.
+    def binary& 6 (LHS RHS)
+      if !LHS then
+        0
+      else
+        !!RHS;
+
+    # Define = with slightly lower precedence than relationals.
+    def binary = 9 (LHS RHS)
+      !(LHS < RHS | LHS > RHS);
+
+    # Define ':' for sequencing: as a low-precedence operator that ignores operands
+    # and just returns the RHS.
+    def binary : 1 (x y) y;
+
+Given the previous if/then/else support, we can also define interesting
+functions for I/O. For example, the following prints out a character
+whose "density" reflects the value passed in: the lower the value, the
+denser the character:
+
+::
+
+    ready>
+
+    extern putchard(char)
+    def printdensity(d)
+      if d > 8 then
+        putchard(32)  # ' '
+      else if d > 4 then
+        putchard(46)  # '.'
+      else if d > 2 then
+        putchard(43)  # '+'
+      else
+        putchard(42); # '*'
+    ...
+    ready> printdensity(1): printdensity(2): printdensity(3):
+           printdensity(4): printdensity(5): printdensity(9):
+           putchard(10);
+    **++.
+    Evaluated to 0.000000
+
+Based on these simple primitive operations, we can start to define more
+interesting things. For example, here's a little function that solves
+for the number of iterations it takes a function in the complex plane to
+converge:
+
+::
+
+    # Determine whether the specific location diverges.
+    # Solve for z = z^2 + c in the complex plane.
+    def mandelconverger(real imag iters creal cimag)
+      if iters > 255 | (real*real + imag*imag > 4) then
+        iters
+      else
+        mandelconverger(real*real - imag*imag + creal,
+                        2*real*imag + cimag,
+                        iters+1, creal, cimag);
+
+    # Return the number of iterations required for the iteration to escape
+    def mandelconverge(real imag)
+      mandelconverger(real, imag, 0, real, imag);
+
+This "``z = z2 + c``" function is a beautiful little creature that is
+the basis for computation of the `Mandelbrot
+Set <http://en.wikipedia.org/wiki/Mandelbrot_set>`_. Our
+``mandelconverge`` function returns the number of iterations that it
+takes for a complex orbit to escape, saturating to 255. This is not a
+very useful function by itself, but if you plot its value over a
+two-dimensional plane, you can see the Mandelbrot set. Given that we are
+limited to using putchard here, our amazing graphical output is limited,
+but we can whip together something using the density plotter above:
+
+::
+
+    # Compute and plot the mandelbrot set with the specified 2 dimensional range
+    # info.
+    def mandelhelp(xmin xmax xstep   ymin ymax ystep)
+      for y = ymin, y < ymax, ystep in (
+        (for x = xmin, x < xmax, xstep in
+           printdensity(mandelconverge(x,y)))
+        : putchard(10)
+      )
+
+    # mandel - This is a convenient helper function for plotting the mandelbrot set
+    # from the specified position with the specified Magnification.
+    def mandel(realstart imagstart realmag imagmag)
+      mandelhelp(realstart, realstart+realmag*78, realmag,
+                 imagstart, imagstart+imagmag*40, imagmag);
+
+Given this, we can try plotting out the mandelbrot set! Lets try it out:
+
+::
+
+    ready> mandel(-2.3, -1.3, 0.05, 0.07);
+    *******************************+++++++++++*************************************
+    *************************+++++++++++++++++++++++*******************************
+    **********************+++++++++++++++++++++++++++++****************************
+    *******************+++++++++++++++++++++.. ...++++++++*************************
+    *****************++++++++++++++++++++++.... ...+++++++++***********************
+    ***************+++++++++++++++++++++++.....   ...+++++++++*********************
+    **************+++++++++++++++++++++++....     ....+++++++++********************
+    *************++++++++++++++++++++++......      .....++++++++*******************
+    ************+++++++++++++++++++++.......       .......+++++++******************
+    ***********+++++++++++++++++++....                ... .+++++++*****************
+    **********+++++++++++++++++.......                     .+++++++****************
+    *********++++++++++++++...........                    ...+++++++***************
+    ********++++++++++++............                      ...++++++++**************
+    ********++++++++++... ..........                        .++++++++**************
+    *******+++++++++.....                                   .+++++++++*************
+    *******++++++++......                                  ..+++++++++*************
+    *******++++++.......                                   ..+++++++++*************
+    *******+++++......                                     ..+++++++++*************
+    *******.... ....                                      ...+++++++++*************
+    *******.... .                                         ...+++++++++*************
+    *******+++++......                                    ...+++++++++*************
+    *******++++++.......                                   ..+++++++++*************
+    *******++++++++......                                   .+++++++++*************
+    *******+++++++++.....                                  ..+++++++++*************
+    ********++++++++++... ..........                        .++++++++**************
+    ********++++++++++++............                      ...++++++++**************
+    *********++++++++++++++..........                     ...+++++++***************
+    **********++++++++++++++++........                     .+++++++****************
+    **********++++++++++++++++++++....                ... ..+++++++****************
+    ***********++++++++++++++++++++++.......       .......++++++++*****************
+    ************+++++++++++++++++++++++......      ......++++++++******************
+    **************+++++++++++++++++++++++....      ....++++++++********************
+    ***************+++++++++++++++++++++++.....   ...+++++++++*********************
+    *****************++++++++++++++++++++++....  ...++++++++***********************
+    *******************+++++++++++++++++++++......++++++++*************************
+    *********************++++++++++++++++++++++.++++++++***************************
+    *************************+++++++++++++++++++++++*******************************
+    ******************************+++++++++++++************************************
+    *******************************************************************************
+    *******************************************************************************
+    *******************************************************************************
+    Evaluated to 0.000000
+    ready> mandel(-2, -1, 0.02, 0.04);
+    **************************+++++++++++++++++++++++++++++++++++++++++++++++++++++
+    ***********************++++++++++++++++++++++++++++++++++++++++++++++++++++++++
+    *********************+++++++++++++++++++++++++++++++++++++++++++++++++++++++++.
+    *******************+++++++++++++++++++++++++++++++++++++++++++++++++++++++++...
+    *****************+++++++++++++++++++++++++++++++++++++++++++++++++++++++++.....
+    ***************++++++++++++++++++++++++++++++++++++++++++++++++++++++++........
+    **************++++++++++++++++++++++++++++++++++++++++++++++++++++++...........
+    ************+++++++++++++++++++++++++++++++++++++++++++++++++++++..............
+    ***********++++++++++++++++++++++++++++++++++++++++++++++++++........        .
+    **********++++++++++++++++++++++++++++++++++++++++++++++.............
+    ********+++++++++++++++++++++++++++++++++++++++++++..................
+    *******+++++++++++++++++++++++++++++++++++++++.......................
+    ******+++++++++++++++++++++++++++++++++++...........................
+    *****++++++++++++++++++++++++++++++++............................
+    *****++++++++++++++++++++++++++++...............................
+    ****++++++++++++++++++++++++++......   .........................
+    ***++++++++++++++++++++++++.........     ......    ...........
+    ***++++++++++++++++++++++............
+    **+++++++++++++++++++++..............
+    **+++++++++++++++++++................
+    *++++++++++++++++++.................
+    *++++++++++++++++............ ...
+    *++++++++++++++..............
+    *+++....++++................
+    *..........  ...........
+    *
+    *..........  ...........
+    *+++....++++................
+    *++++++++++++++..............
+    *++++++++++++++++............ ...
+    *++++++++++++++++++.................
+    **+++++++++++++++++++................
+    **+++++++++++++++++++++..............
+    ***++++++++++++++++++++++............
+    ***++++++++++++++++++++++++.........     ......    ...........
+    ****++++++++++++++++++++++++++......   .........................
+    *****++++++++++++++++++++++++++++...............................
+    *****++++++++++++++++++++++++++++++++............................
+    ******+++++++++++++++++++++++++++++++++++...........................
+    *******+++++++++++++++++++++++++++++++++++++++.......................
+    ********+++++++++++++++++++++++++++++++++++++++++++..................
+    Evaluated to 0.000000
+    ready> mandel(-0.9, -1.4, 0.02, 0.03);
+    *******************************************************************************
+    *******************************************************************************
+    *******************************************************************************
+    **********+++++++++++++++++++++************************************************
+    *+++++++++++++++++++++++++++++++++++++++***************************************
+    +++++++++++++++++++++++++++++++++++++++++++++**********************************
+    ++++++++++++++++++++++++++++++++++++++++++++++++++*****************************
+    ++++++++++++++++++++++++++++++++++++++++++++++++++++++*************************
+    +++++++++++++++++++++++++++++++++++++++++++++++++++++++++**********************
+    +++++++++++++++++++++++++++++++++.........++++++++++++++++++*******************
+    +++++++++++++++++++++++++++++++....   ......+++++++++++++++++++****************
+    +++++++++++++++++++++++++++++.......  ........+++++++++++++++++++**************
+    ++++++++++++++++++++++++++++........   ........++++++++++++++++++++************
+    +++++++++++++++++++++++++++.........     ..  ...+++++++++++++++++++++**********
+    ++++++++++++++++++++++++++...........        ....++++++++++++++++++++++********
+    ++++++++++++++++++++++++.............       .......++++++++++++++++++++++******
+    +++++++++++++++++++++++.............        ........+++++++++++++++++++++++****
+    ++++++++++++++++++++++...........           ..........++++++++++++++++++++++***
+    ++++++++++++++++++++...........                .........++++++++++++++++++++++*
+    ++++++++++++++++++............                  ...........++++++++++++++++++++
+    ++++++++++++++++...............                 .............++++++++++++++++++
+    ++++++++++++++.................                 ...............++++++++++++++++
+    ++++++++++++..................                  .................++++++++++++++
+    +++++++++..................                      .................+++++++++++++
+    ++++++........        .                               .........  ..++++++++++++
+    ++............                                         ......    ....++++++++++
+    ..............                                                    ...++++++++++
+    ..............                                                    ....+++++++++
+    ..............                                                    .....++++++++
+    .............                                                    ......++++++++
+    ...........                                                     .......++++++++
+    .........                                                       ........+++++++
+    .........                                                       ........+++++++
+    .........                                                           ....+++++++
+    ........                                                             ...+++++++
+    .......                                                              ...+++++++
+                                                                        ....+++++++
+                                                                       .....+++++++
+                                                                        ....+++++++
+                                                                        ....+++++++
+                                                                        ....+++++++
+    Evaluated to 0.000000
+    ready> ^D
+
+At this point, you may be starting to realize that Kaleidoscope is a
+real and powerful language. It may not be self-similar :), but it can be
+used to plot things that are!
+
+With this, we conclude the "adding user-defined operators" chapter of
+the tutorial. We have successfully augmented our language, adding the
+ability to extend the language in the library, and we have shown how
+this can be used to build a simple but interesting end-user application
+in Kaleidoscope. At this point, Kaleidoscope can build a variety of
+applications that are functional and can call functions with
+side-effects, but it can't actually define and mutate a variable itself.
+
+Strikingly, variable mutation is an important feature of some languages,
+and it is not at all obvious how to `add support for mutable
+variables <LangImpl7.html>`_ without having to add an "SSA construction"
+phase to your front-end. In the next chapter, we will describe how you
+can add variable mutation without building SSA in your front-end.
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example, enhanced with
+the if/then/else and for expressions.. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core mcjit native` -O3 -o toy
+    # Run
+    ./toy
+
+On some platforms, you will need to specify -rdynamic or
+-Wl,--export-dynamic when linking. This ensures that symbols defined in
+the main executable are exported to the dynamic linker and so are
+available for symbol resolution at run time. This is not needed if you
+compile your support code into a shared library, although doing that
+will cause problems on Windows.
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter6/toy.cpp
+   :language: c++
+
+`Next: Extending the language: mutable variables / SSA
+construction <LangImpl07.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl07.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl07.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl07.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl07.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,881 @@
+=======================================================
+Kaleidoscope: Extending the Language: Mutable Variables
+=======================================================
+
+.. contents::
+   :local:
+
+Chapter 7 Introduction
+======================
+
+Welcome to Chapter 7 of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. In chapters 1 through 6, we've built a
+very respectable, albeit simple, `functional programming
+language <http://en.wikipedia.org/wiki/Functional_programming>`_. In our
+journey, we learned some parsing techniques, how to build and represent
+an AST, how to build LLVM IR, and how to optimize the resultant code as
+well as JIT compile it.
+
+While Kaleidoscope is interesting as a functional language, the fact
+that it is functional makes it "too easy" to generate LLVM IR for it. In
+particular, a functional language makes it very easy to build LLVM IR
+directly in `SSA
+form <http://en.wikipedia.org/wiki/Static_single_assignment_form>`_.
+Since LLVM requires that the input code be in SSA form, this is a very
+nice property and it is often unclear to newcomers how to generate code
+for an imperative language with mutable variables.
+
+The short (and happy) summary of this chapter is that there is no need
+for your front-end to build SSA form: LLVM provides highly tuned and
+well tested support for this, though the way it works is a bit
+unexpected for some.
+
+Why is this a hard problem?
+===========================
+
+To understand why mutable variables cause complexities in SSA
+construction, consider this extremely simple C example:
+
+.. code-block:: c
+
+    int G, H;
+    int test(_Bool Condition) {
+      int X;
+      if (Condition)
+        X = G;
+      else
+        X = H;
+      return X;
+    }
+
+In this case, we have the variable "X", whose value depends on the path
+executed in the program. Because there are two different possible values
+for X before the return instruction, a PHI node is inserted to merge the
+two values. The LLVM IR that we want for this example looks like this:
+
+.. code-block:: llvm
+
+    @G = weak global i32 0   ; type of @G is i32*
+    @H = weak global i32 0   ; type of @H is i32*
+
+    define i32 @test(i1 %Condition) {
+    entry:
+      br i1 %Condition, label %cond_true, label %cond_false
+
+    cond_true:
+      %X.0 = load i32* @G
+      br label %cond_next
+
+    cond_false:
+      %X.1 = load i32* @H
+      br label %cond_next
+
+    cond_next:
+      %X.2 = phi i32 [ %X.1, %cond_false ], [ %X.0, %cond_true ]
+      ret i32 %X.2
+    }
+
+In this example, the loads from the G and H global variables are
+explicit in the LLVM IR, and they live in the then/else branches of the
+if statement (cond\_true/cond\_false). In order to merge the incoming
+values, the X.2 phi node in the cond\_next block selects the right value
+to use based on where control flow is coming from: if control flow comes
+from the cond\_false block, X.2 gets the value of X.1. Alternatively, if
+control flow comes from cond\_true, it gets the value of X.0. The intent
+of this chapter is not to explain the details of SSA form. For more
+information, see one of the many `online
+references <http://en.wikipedia.org/wiki/Static_single_assignment_form>`_.
+
+The question for this article is "who places the phi nodes when lowering
+assignments to mutable variables?". The issue here is that LLVM
+*requires* that its IR be in SSA form: there is no "non-ssa" mode for
+it. However, SSA construction requires non-trivial algorithms and data
+structures, so it is inconvenient and wasteful for every front-end to
+have to reproduce this logic.
+
+Memory in LLVM
+==============
+
+The 'trick' here is that while LLVM does require all register values to
+be in SSA form, it does not require (or permit) memory objects to be in
+SSA form. In the example above, note that the loads from G and H are
+direct accesses to G and H: they are not renamed or versioned. This
+differs from some other compiler systems, which do try to version memory
+objects. In LLVM, instead of encoding dataflow analysis of memory into
+the LLVM IR, it is handled with `Analysis
+Passes <../WritingAnLLVMPass.html>`_ which are computed on demand.
+
+With this in mind, the high-level idea is that we want to make a stack
+variable (which lives in memory, because it is on the stack) for each
+mutable object in a function. To take advantage of this trick, we need
+to talk about how LLVM represents stack variables.
+
+In LLVM, all memory accesses are explicit with load/store instructions,
+and it is carefully designed not to have (or need) an "address-of"
+operator. Notice how the type of the @G/@H global variables is actually
+"i32\*" even though the variable is defined as "i32". What this means is
+that @G defines *space* for an i32 in the global data area, but its
+*name* actually refers to the address for that space. Stack variables
+work the same way, except that instead of being declared with global
+variable definitions, they are declared with the `LLVM alloca
+instruction <../LangRef.html#alloca-instruction>`_:
+
+.. code-block:: llvm
+
+    define i32 @example() {
+    entry:
+      %X = alloca i32           ; type of %X is i32*.
+      ...
+      %tmp = load i32* %X       ; load the stack value %X from the stack.
+      %tmp2 = add i32 %tmp, 1   ; increment it
+      store i32 %tmp2, i32* %X  ; store it back
+      ...
+
+This code shows an example of how you can declare and manipulate a stack
+variable in the LLVM IR. Stack memory allocated with the alloca
+instruction is fully general: you can pass the address of the stack slot
+to functions, you can store it in other variables, etc. In our example
+above, we could rewrite the example to use the alloca technique to avoid
+using a PHI node:
+
+.. code-block:: llvm
+
+    @G = weak global i32 0   ; type of @G is i32*
+    @H = weak global i32 0   ; type of @H is i32*
+
+    define i32 @test(i1 %Condition) {
+    entry:
+      %X = alloca i32           ; type of %X is i32*.
+      br i1 %Condition, label %cond_true, label %cond_false
+
+    cond_true:
+      %X.0 = load i32* @G
+      store i32 %X.0, i32* %X   ; Update X
+      br label %cond_next
+
+    cond_false:
+      %X.1 = load i32* @H
+      store i32 %X.1, i32* %X   ; Update X
+      br label %cond_next
+
+    cond_next:
+      %X.2 = load i32* %X       ; Read X
+      ret i32 %X.2
+    }
+
+With this, we have discovered a way to handle arbitrary mutable
+variables without the need to create Phi nodes at all:
+
+#. Each mutable variable becomes a stack allocation.
+#. Each read of the variable becomes a load from the stack.
+#. Each update of the variable becomes a store to the stack.
+#. Taking the address of a variable just uses the stack address
+   directly.
+
+While this solution has solved our immediate problem, it introduced
+another one: we have now apparently introduced a lot of stack traffic
+for very simple and common operations, a major performance problem.
+Fortunately for us, the LLVM optimizer has a highly-tuned optimization
+pass named "mem2reg" that handles this case, promoting allocas like this
+into SSA registers, inserting Phi nodes as appropriate. If you run this
+example through the pass, for example, you'll get:
+
+.. code-block:: bash
+
+    $ llvm-as < example.ll | opt -mem2reg | llvm-dis
+    @G = weak global i32 0
+    @H = weak global i32 0
+
+    define i32 @test(i1 %Condition) {
+    entry:
+      br i1 %Condition, label %cond_true, label %cond_false
+
+    cond_true:
+      %X.0 = load i32* @G
+      br label %cond_next
+
+    cond_false:
+      %X.1 = load i32* @H
+      br label %cond_next
+
+    cond_next:
+      %X.01 = phi i32 [ %X.1, %cond_false ], [ %X.0, %cond_true ]
+      ret i32 %X.01
+    }
+
+The mem2reg pass implements the standard "iterated dominance frontier"
+algorithm for constructing SSA form and has a number of optimizations
+that speed up (very common) degenerate cases. The mem2reg optimization
+pass is the answer to dealing with mutable variables, and we highly
+recommend that you depend on it. Note that mem2reg only works on
+variables in certain circumstances:
+
+#. mem2reg is alloca-driven: it looks for allocas and if it can handle
+   them, it promotes them. It does not apply to global variables or heap
+   allocations.
+#. mem2reg only looks for alloca instructions in the entry block of the
+   function. Being in the entry block guarantees that the alloca is only
+   executed once, which makes analysis simpler.
+#. mem2reg only promotes allocas whose uses are direct loads and stores.
+   If the address of the stack object is passed to a function, or if any
+   funny pointer arithmetic is involved, the alloca will not be
+   promoted.
+#. mem2reg only works on allocas of `first
+   class <../LangRef.html#first-class-types>`_ values (such as pointers,
+   scalars and vectors), and only if the array size of the allocation is
+   1 (or missing in the .ll file). mem2reg is not capable of promoting
+   structs or arrays to registers. Note that the "sroa" pass is
+   more powerful and can promote structs, "unions", and arrays in many
+   cases.
+
+All of these properties are easy to satisfy for most imperative
+languages, and we'll illustrate it below with Kaleidoscope. The final
+question you may be asking is: should I bother with this nonsense for my
+front-end? Wouldn't it be better if I just did SSA construction
+directly, avoiding use of the mem2reg optimization pass? In short, we
+strongly recommend that you use this technique for building SSA form,
+unless there is an extremely good reason not to. Using this technique
+is:
+
+-  Proven and well tested: clang uses this technique
+   for local mutable variables. As such, the most common clients of LLVM
+   are using this to handle a bulk of their variables. You can be sure
+   that bugs are found fast and fixed early.
+-  Extremely Fast: mem2reg has a number of special cases that make it
+   fast in common cases as well as fully general. For example, it has
+   fast-paths for variables that are only used in a single block,
+   variables that only have one assignment point, good heuristics to
+   avoid insertion of unneeded phi nodes, etc.
+-  Needed for debug info generation: `Debug information in
+   LLVM <../SourceLevelDebugging.html>`_ relies on having the address of
+   the variable exposed so that debug info can be attached to it. This
+   technique dovetails very naturally with this style of debug info.
+
+If nothing else, this makes it much easier to get your front-end up and
+running, and is very simple to implement. Let's extend Kaleidoscope with
+mutable variables now!
+
+Mutable Variables in Kaleidoscope
+=================================
+
+Now that we know the sort of problem we want to tackle, let's see what
+this looks like in the context of our little Kaleidoscope language.
+We're going to add two features:
+
+#. The ability to mutate variables with the '=' operator.
+#. The ability to define new variables.
+
+While the first item is really what this is about, we only have
+variables for incoming arguments as well as for induction variables, and
+redefining those only goes so far :). Also, the ability to define new
+variables is a useful thing regardless of whether you will be mutating
+them. Here's a motivating example that shows how we could use these:
+
+::
+
+    # Define ':' for sequencing: as a low-precedence operator that ignores operands
+    # and just returns the RHS.
+    def binary : 1 (x y) y;
+
+    # Recursive fib, we could do this before.
+    def fib(x)
+      if (x < 3) then
+        1
+      else
+        fib(x-1)+fib(x-2);
+
+    # Iterative fib.
+    def fibi(x)
+      var a = 1, b = 1, c in
+      (for i = 3, i < x in
+         c = a + b :
+         a = b :
+         b = c) :
+      b;
+
+    # Call it.
+    fibi(10);
+
+In order to mutate variables, we have to change our existing variables
+to use the "alloca trick". Once we have that, we'll add our new
+operator, then extend Kaleidoscope to support new variable definitions.
+
+Adjusting Existing Variables for Mutation
+=========================================
+
+The symbol table in Kaleidoscope is managed at code generation time by
+the '``NamedValues``' map. This map currently keeps track of the LLVM
+"Value\*" that holds the double value for the named variable. In order
+to support mutation, we need to change this slightly, so that
+``NamedValues`` holds the *memory location* of the variable in question.
+Note that this change is a refactoring: it changes the structure of the
+code, but does not (by itself) change the behavior of the compiler. All
+of these changes are isolated in the Kaleidoscope code generator.
+
+At this point in Kaleidoscope's development, it only supports variables
+for two things: incoming arguments to functions and the induction
+variable of 'for' loops. For consistency, we'll allow mutation of these
+variables in addition to other user-defined variables. This means that
+these will both need memory locations.
+
+To start our transformation of Kaleidoscope, we'll change the
+NamedValues map so that it maps to AllocaInst\* instead of Value\*. Once
+we do this, the C++ compiler will tell us what parts of the code we need
+to update:
+
+.. code-block:: c++
+
+    static std::map<std::string, AllocaInst*> NamedValues;
+
+Also, since we will need to create these alloca's, we'll use a helper
+function that ensures that the allocas are created in the entry block of
+the function:
+
+.. code-block:: c++
+
+    /// CreateEntryBlockAlloca - Create an alloca instruction in the entry block of
+    /// the function.  This is used for mutable variables etc.
+    static AllocaInst *CreateEntryBlockAlloca(Function *TheFunction,
+                                              const std::string &VarName) {
+      IRBuilder<> TmpB(&TheFunction->getEntryBlock(),
+                     TheFunction->getEntryBlock().begin());
+      return TmpB.CreateAlloca(Type::getDoubleTy(LLVMContext), 0,
+                               VarName.c_str());
+    }
+
+This funny looking code creates an IRBuilder object that is pointing at
+the first instruction (.begin()) of the entry block. It then creates an
+alloca with the expected name and returns it. Because all values in
+Kaleidoscope are doubles, there is no need to pass in a type to use.
+
+With this in place, the first functionality change we want to make is to
+variable references. In our new scheme, variables live on the stack, so
+code generating a reference to them actually needs to produce a load
+from the stack slot:
+
+.. code-block:: c++
+
+    Value *VariableExprAST::codegen() {
+      // Look this variable up in the function.
+      Value *V = NamedValues[Name];
+      if (!V)
+        return LogErrorV("Unknown variable name");
+
+      // Load the value.
+      return Builder.CreateLoad(V, Name.c_str());
+    }
+
+As you can see, this is pretty straightforward. Now we need to update
+the things that define the variables to set up the alloca. We'll start
+with ``ForExprAST::codegen()`` (see the `full code listing <#id1>`_ for
+the unabridged code):
+
+.. code-block:: c++
+
+      Function *TheFunction = Builder.GetInsertBlock()->getParent();
+
+      // Create an alloca for the variable in the entry block.
+      AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName);
+
+        // Emit the start code first, without 'variable' in scope.
+      Value *StartVal = Start->codegen();
+      if (!StartVal)
+        return nullptr;
+
+      // Store the value into the alloca.
+      Builder.CreateStore(StartVal, Alloca);
+      ...
+
+      // Compute the end condition.
+      Value *EndCond = End->codegen();
+      if (!EndCond)
+        return nullptr;
+
+      // Reload, increment, and restore the alloca.  This handles the case where
+      // the body of the loop mutates the variable.
+      Value *CurVar = Builder.CreateLoad(Alloca);
+      Value *NextVar = Builder.CreateFAdd(CurVar, StepVal, "nextvar");
+      Builder.CreateStore(NextVar, Alloca);
+      ...
+
+This code is virtually identical to the code `before we allowed mutable
+variables <LangImpl5.html#code-generation-for-the-for-loop>`_. The big difference is that we
+no longer have to construct a PHI node, and we use load/store to access
+the variable as needed.
+
+To support mutable argument variables, we need to also make allocas for
+them. The code for this is also pretty simple:
+
+.. code-block:: c++
+
+    /// CreateArgumentAllocas - Create an alloca for each argument and register the
+    /// argument in the symbol table so that references to it will succeed.
+    void PrototypeAST::CreateArgumentAllocas(Function *F) {
+      Function::arg_iterator AI = F->arg_begin();
+      for (unsigned Idx = 0, e = Args.size(); Idx != e; ++Idx, ++AI) {
+        // Create an alloca for this variable.
+        AllocaInst *Alloca = CreateEntryBlockAlloca(F, Args[Idx]);
+
+        // Store the initial value into the alloca.
+        Builder.CreateStore(AI, Alloca);
+
+        // Add arguments to variable symbol table.
+        NamedValues[Args[Idx]] = Alloca;
+      }
+    }
+
+For each argument, we make an alloca, store the input value to the
+function into the alloca, and register the alloca as the memory location
+for the argument. This method gets invoked by ``FunctionAST::codegen()``
+right after it sets up the entry block for the function.
+
+The final missing piece is adding the mem2reg pass, which allows us to
+get good codegen once again:
+
+.. code-block:: c++
+
+        // Set up the optimizer pipeline.  Start with registering info about how the
+        // target lays out data structures.
+        OurFPM.add(new DataLayout(*TheExecutionEngine->getDataLayout()));
+        // Promote allocas to registers.
+        OurFPM.add(createPromoteMemoryToRegisterPass());
+        // Do simple "peephole" optimizations and bit-twiddling optzns.
+        OurFPM.add(createInstructionCombiningPass());
+        // Reassociate expressions.
+        OurFPM.add(createReassociatePass());
+
+It is interesting to see what the code looks like before and after the
+mem2reg optimization runs. For example, this is the before/after code
+for our recursive fib function. Before the optimization:
+
+.. code-block:: llvm
+
+    define double @fib(double %x) {
+    entry:
+      %x1 = alloca double
+      store double %x, double* %x1
+      %x2 = load double* %x1
+      %cmptmp = fcmp ult double %x2, 3.000000e+00
+      %booltmp = uitofp i1 %cmptmp to double
+      %ifcond = fcmp one double %booltmp, 0.000000e+00
+      br i1 %ifcond, label %then, label %else
+
+    then:       ; preds = %entry
+      br label %ifcont
+
+    else:       ; preds = %entry
+      %x3 = load double* %x1
+      %subtmp = fsub double %x3, 1.000000e+00
+      %calltmp = call double @fib(double %subtmp)
+      %x4 = load double* %x1
+      %subtmp5 = fsub double %x4, 2.000000e+00
+      %calltmp6 = call double @fib(double %subtmp5)
+      %addtmp = fadd double %calltmp, %calltmp6
+      br label %ifcont
+
+    ifcont:     ; preds = %else, %then
+      %iftmp = phi double [ 1.000000e+00, %then ], [ %addtmp, %else ]
+      ret double %iftmp
+    }
+
+Here there is only one variable (x, the input argument) but you can
+still see the extremely simple-minded code generation strategy we are
+using. In the entry block, an alloca is created, and the initial input
+value is stored into it. Each reference to the variable does a reload
+from the stack. Also, note that we didn't modify the if/then/else
+expression, so it still inserts a PHI node. While we could make an
+alloca for it, it is actually easier to create a PHI node for it, so we
+still just make the PHI.
+
+Here is the code after the mem2reg pass runs:
+
+.. code-block:: llvm
+
+    define double @fib(double %x) {
+    entry:
+      %cmptmp = fcmp ult double %x, 3.000000e+00
+      %booltmp = uitofp i1 %cmptmp to double
+      %ifcond = fcmp one double %booltmp, 0.000000e+00
+      br i1 %ifcond, label %then, label %else
+
+    then:
+      br label %ifcont
+
+    else:
+      %subtmp = fsub double %x, 1.000000e+00
+      %calltmp = call double @fib(double %subtmp)
+      %subtmp5 = fsub double %x, 2.000000e+00
+      %calltmp6 = call double @fib(double %subtmp5)
+      %addtmp = fadd double %calltmp, %calltmp6
+      br label %ifcont
+
+    ifcont:     ; preds = %else, %then
+      %iftmp = phi double [ 1.000000e+00, %then ], [ %addtmp, %else ]
+      ret double %iftmp
+    }
+
+This is a trivial case for mem2reg, since there are no redefinitions of
+the variable. The point of showing this is to calm your tension about
+inserting such blatent inefficiencies :).
+
+After the rest of the optimizers run, we get:
+
+.. code-block:: llvm
+
+    define double @fib(double %x) {
+    entry:
+      %cmptmp = fcmp ult double %x, 3.000000e+00
+      %booltmp = uitofp i1 %cmptmp to double
+      %ifcond = fcmp ueq double %booltmp, 0.000000e+00
+      br i1 %ifcond, label %else, label %ifcont
+
+    else:
+      %subtmp = fsub double %x, 1.000000e+00
+      %calltmp = call double @fib(double %subtmp)
+      %subtmp5 = fsub double %x, 2.000000e+00
+      %calltmp6 = call double @fib(double %subtmp5)
+      %addtmp = fadd double %calltmp, %calltmp6
+      ret double %addtmp
+
+    ifcont:
+      ret double 1.000000e+00
+    }
+
+Here we see that the simplifycfg pass decided to clone the return
+instruction into the end of the 'else' block. This allowed it to
+eliminate some branches and the PHI node.
+
+Now that all symbol table references are updated to use stack variables,
+we'll add the assignment operator.
+
+New Assignment Operator
+=======================
+
+With our current framework, adding a new assignment operator is really
+simple. We will parse it just like any other binary operator, but handle
+it internally (instead of allowing the user to define it). The first
+step is to set a precedence:
+
+.. code-block:: c++
+
+     int main() {
+       // Install standard binary operators.
+       // 1 is lowest precedence.
+       BinopPrecedence['='] = 2;
+       BinopPrecedence['<'] = 10;
+       BinopPrecedence['+'] = 20;
+       BinopPrecedence['-'] = 20;
+
+Now that the parser knows the precedence of the binary operator, it
+takes care of all the parsing and AST generation. We just need to
+implement codegen for the assignment operator. This looks like:
+
+.. code-block:: c++
+
+    Value *BinaryExprAST::codegen() {
+      // Special case '=' because we don't want to emit the LHS as an expression.
+      if (Op == '=') {
+        // Assignment requires the LHS to be an identifier.
+        VariableExprAST *LHSE = dynamic_cast<VariableExprAST*>(LHS.get());
+        if (!LHSE)
+          return LogErrorV("destination of '=' must be a variable");
+
+Unlike the rest of the binary operators, our assignment operator doesn't
+follow the "emit LHS, emit RHS, do computation" model. As such, it is
+handled as a special case before the other binary operators are handled.
+The other strange thing is that it requires the LHS to be a variable. It
+is invalid to have "(x+1) = expr" - only things like "x = expr" are
+allowed.
+
+.. code-block:: c++
+
+        // Codegen the RHS.
+        Value *Val = RHS->codegen();
+        if (!Val)
+          return nullptr;
+
+        // Look up the name.
+        Value *Variable = NamedValues[LHSE->getName()];
+        if (!Variable)
+          return LogErrorV("Unknown variable name");
+
+        Builder.CreateStore(Val, Variable);
+        return Val;
+      }
+      ...
+
+Once we have the variable, codegen'ing the assignment is
+straightforward: we emit the RHS of the assignment, create a store, and
+return the computed value. Returning a value allows for chained
+assignments like "X = (Y = Z)".
+
+Now that we have an assignment operator, we can mutate loop variables
+and arguments. For example, we can now run code like this:
+
+::
+
+    # Function to print a double.
+    extern printd(x);
+
+    # Define ':' for sequencing: as a low-precedence operator that ignores operands
+    # and just returns the RHS.
+    def binary : 1 (x y) y;
+
+    def test(x)
+      printd(x) :
+      x = 4 :
+      printd(x);
+
+    test(123);
+
+When run, this example prints "123" and then "4", showing that we did
+actually mutate the value! Okay, we have now officially implemented our
+goal: getting this to work requires SSA construction in the general
+case. However, to be really useful, we want the ability to define our
+own local variables, let's add this next!
+
+User-defined Local Variables
+============================
+
+Adding var/in is just like any other extension we made to
+Kaleidoscope: we extend the lexer, the parser, the AST and the code
+generator. The first step for adding our new 'var/in' construct is to
+extend the lexer. As before, this is pretty trivial, the code looks like
+this:
+
+.. code-block:: c++
+
+    enum Token {
+      ...
+      // var definition
+      tok_var = -13
+    ...
+    }
+    ...
+    static int gettok() {
+    ...
+        if (IdentifierStr == "in")
+          return tok_in;
+        if (IdentifierStr == "binary")
+          return tok_binary;
+        if (IdentifierStr == "unary")
+          return tok_unary;
+        if (IdentifierStr == "var")
+          return tok_var;
+        return tok_identifier;
+    ...
+
+The next step is to define the AST node that we will construct. For
+var/in, it looks like this:
+
+.. code-block:: c++
+
+    /// VarExprAST - Expression class for var/in
+    class VarExprAST : public ExprAST {
+      std::vector<std::pair<std::string, std::unique_ptr<ExprAST>>> VarNames;
+      std::unique_ptr<ExprAST> Body;
+
+    public:
+      VarExprAST(std::vector<std::pair<std::string, std::unique_ptr<ExprAST>>> VarNames,
+                 std::unique_ptr<ExprAST> body)
+      : VarNames(std::move(VarNames)), Body(std::move(Body)) {}
+
+      virtual Value *codegen();
+    };
+
+var/in allows a list of names to be defined all at once, and each name
+can optionally have an initializer value. As such, we capture this
+information in the VarNames vector. Also, var/in has a body, this body
+is allowed to access the variables defined by the var/in.
+
+With this in place, we can define the parser pieces. The first thing we
+do is add it as a primary expression:
+
+.. code-block:: c++
+
+    /// primary
+    ///   ::= identifierexpr
+    ///   ::= numberexpr
+    ///   ::= parenexpr
+    ///   ::= ifexpr
+    ///   ::= forexpr
+    ///   ::= varexpr
+    static std::unique_ptr<ExprAST> ParsePrimary() {
+      switch (CurTok) {
+      default:
+        return LogError("unknown token when expecting an expression");
+      case tok_identifier:
+        return ParseIdentifierExpr();
+      case tok_number:
+        return ParseNumberExpr();
+      case '(':
+        return ParseParenExpr();
+      case tok_if:
+        return ParseIfExpr();
+      case tok_for:
+        return ParseForExpr();
+      case tok_var:
+        return ParseVarExpr();
+      }
+    }
+
+Next we define ParseVarExpr:
+
+.. code-block:: c++
+
+    /// varexpr ::= 'var' identifier ('=' expression)?
+    //                    (',' identifier ('=' expression)?)* 'in' expression
+    static std::unique_ptr<ExprAST> ParseVarExpr() {
+      getNextToken();  // eat the var.
+
+      std::vector<std::pair<std::string, std::unique_ptr<ExprAST>>> VarNames;
+
+      // At least one variable name is required.
+      if (CurTok != tok_identifier)
+        return LogError("expected identifier after var");
+
+The first part of this code parses the list of identifier/expr pairs
+into the local ``VarNames`` vector.
+
+.. code-block:: c++
+
+      while (1) {
+        std::string Name = IdentifierStr;
+        getNextToken();  // eat identifier.
+
+        // Read the optional initializer.
+        std::unique_ptr<ExprAST> Init;
+        if (CurTok == '=') {
+          getNextToken(); // eat the '='.
+
+          Init = ParseExpression();
+          if (!Init) return nullptr;
+        }
+
+        VarNames.push_back(std::make_pair(Name, std::move(Init)));
+
+        // End of var list, exit loop.
+        if (CurTok != ',') break;
+        getNextToken(); // eat the ','.
+
+        if (CurTok != tok_identifier)
+          return LogError("expected identifier list after var");
+      }
+
+Once all the variables are parsed, we then parse the body and create the
+AST node:
+
+.. code-block:: c++
+
+      // At this point, we have to have 'in'.
+      if (CurTok != tok_in)
+        return LogError("expected 'in' keyword after 'var'");
+      getNextToken();  // eat 'in'.
+
+      auto Body = ParseExpression();
+      if (!Body)
+        return nullptr;
+
+      return llvm::make_unique<VarExprAST>(std::move(VarNames),
+                                           std::move(Body));
+    }
+
+Now that we can parse and represent the code, we need to support
+emission of LLVM IR for it. This code starts out with:
+
+.. code-block:: c++
+
+    Value *VarExprAST::codegen() {
+      std::vector<AllocaInst *> OldBindings;
+
+      Function *TheFunction = Builder.GetInsertBlock()->getParent();
+
+      // Register all variables and emit their initializer.
+      for (unsigned i = 0, e = VarNames.size(); i != e; ++i) {
+        const std::string &VarName = VarNames[i].first;
+        ExprAST *Init = VarNames[i].second.get();
+
+Basically it loops over all the variables, installing them one at a
+time. For each variable we put into the symbol table, we remember the
+previous value that we replace in OldBindings.
+
+.. code-block:: c++
+
+        // Emit the initializer before adding the variable to scope, this prevents
+        // the initializer from referencing the variable itself, and permits stuff
+        // like this:
+        //  var a = 1 in
+        //    var a = a in ...   # refers to outer 'a'.
+        Value *InitVal;
+        if (Init) {
+          InitVal = Init->codegen();
+          if (!InitVal)
+            return nullptr;
+        } else { // If not specified, use 0.0.
+          InitVal = ConstantFP::get(LLVMContext, APFloat(0.0));
+        }
+
+        AllocaInst *Alloca = CreateEntryBlockAlloca(TheFunction, VarName);
+        Builder.CreateStore(InitVal, Alloca);
+
+        // Remember the old variable binding so that we can restore the binding when
+        // we unrecurse.
+        OldBindings.push_back(NamedValues[VarName]);
+
+        // Remember this binding.
+        NamedValues[VarName] = Alloca;
+      }
+
+There are more comments here than code. The basic idea is that we emit
+the initializer, create the alloca, then update the symbol table to
+point to it. Once all the variables are installed in the symbol table,
+we evaluate the body of the var/in expression:
+
+.. code-block:: c++
+
+      // Codegen the body, now that all vars are in scope.
+      Value *BodyVal = Body->codegen();
+      if (!BodyVal)
+        return nullptr;
+
+Finally, before returning, we restore the previous variable bindings:
+
+.. code-block:: c++
+
+      // Pop all our variables from scope.
+      for (unsigned i = 0, e = VarNames.size(); i != e; ++i)
+        NamedValues[VarNames[i].first] = OldBindings[i];
+
+      // Return the body computation.
+      return BodyVal;
+    }
+
+The end result of all of this is that we get properly scoped variable
+definitions, and we even (trivially) allow mutation of them :).
+
+With this, we completed what we set out to do. Our nice iterative fib
+example from the intro compiles and runs just fine. The mem2reg pass
+optimizes all of our stack variables into SSA registers, inserting PHI
+nodes where needed, and our front-end remains simple: no "iterated
+dominance frontier" computation anywhere in sight.
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example, enhanced with
+mutable variables and var/in support. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core mcjit native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter7/toy.cpp
+   :language: c++
+
+`Next: Compiling to Object Code <LangImpl08.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl08.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl08.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl08.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl08.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,218 @@
+========================================
+ Kaleidoscope: Compiling to Object Code
+========================================
+
+.. contents::
+   :local:
+
+Chapter 8 Introduction
+======================
+
+Welcome to Chapter 8 of the "`Implementing a language with LLVM
+<index.html>`_" tutorial. This chapter describes how to compile our
+language down to object files.
+
+Choosing a target
+=================
+
+LLVM has native support for cross-compilation. You can compile to the
+architecture of your current machine, or just as easily compile for
+other architectures. In this tutorial, we'll target the current
+machine.
+
+To specify the architecture that you want to target, we use a string
+called a "target triple". This takes the form
+``<arch><sub>-<vendor>-<sys>-<abi>`` (see the `cross compilation docs
+<http://clang.llvm.org/docs/CrossCompilation.html#target-triple>`_).
+
+As an example, we can see what clang thinks is our current target
+triple:
+
+::
+
+    $ clang --version | grep Target
+    Target: x86_64-unknown-linux-gnu
+
+Running this command may show something different on your machine as
+you might be using a different architecture or operating system to me.
+
+Fortunately, we don't need to hard-code a target triple to target the
+current machine. LLVM provides ``sys::getDefaultTargetTriple``, which
+returns the target triple of the current machine.
+
+.. code-block:: c++
+
+    auto TargetTriple = sys::getDefaultTargetTriple();
+
+LLVM doesn't require us to to link in all the target
+functionality. For example, if we're just using the JIT, we don't need
+the assembly printers. Similarly, if we're only targeting certain
+architectures, we can only link in the functionality for those
+architectures.
+
+For this example, we'll initialize all the targets for emitting object
+code.
+
+.. code-block:: c++
+
+    InitializeAllTargetInfos();
+    InitializeAllTargets();
+    InitializeAllTargetMCs();
+    InitializeAllAsmParsers();
+    InitializeAllAsmPrinters();
+
+We can now use our target triple to get a ``Target``:
+
+.. code-block:: c++
+
+  std::string Error;
+  auto Target = TargetRegistry::lookupTarget(TargetTriple, Error);
+
+  // Print an error and exit if we couldn't find the requested target.
+  // This generally occurs if we've forgotten to initialise the
+  // TargetRegistry or we have a bogus target triple.
+  if (!Target) {
+    errs() << Error;
+    return 1;
+  }
+
+Target Machine
+==============
+
+We will also need a ``TargetMachine``. This class provides a complete
+machine description of the machine we're targeting. If we want to
+target a specific feature (such as SSE) or a specific CPU (such as
+Intel's Sandylake), we do so now.
+
+To see which features and CPUs that LLVM knows about, we can use
+``llc``. For example, let's look at x86:
+
+::
+
+    $ llvm-as < /dev/null | llc -march=x86 -mattr=help
+    Available CPUs for this target:
+
+      amdfam10      - Select the amdfam10 processor.
+      athlon        - Select the athlon processor.
+      athlon-4      - Select the athlon-4 processor.
+      ...
+
+    Available features for this target:
+
+      16bit-mode            - 16-bit mode (i8086).
+      32bit-mode            - 32-bit mode (80386).
+      3dnow                 - Enable 3DNow! instructions.
+      3dnowa                - Enable 3DNow! Athlon instructions.
+      ...
+
+For our example, we'll use the generic CPU without any additional
+features, options or relocation model.
+
+.. code-block:: c++
+
+  auto CPU = "generic";
+  auto Features = "";
+
+  TargetOptions opt;
+  auto RM = Optional<Reloc::Model>();
+  auto TargetMachine = Target->createTargetMachine(TargetTriple, CPU, Features, opt, RM);
+
+
+Configuring the Module
+======================
+
+We're now ready to configure our module, to specify the target and
+data layout. This isn't strictly necessary, but the `frontend
+performance guide <../Frontend/PerformanceTips.html>`_ recommends
+this. Optimizations benefit from knowing about the target and data
+layout.
+
+.. code-block:: c++
+
+  TheModule->setDataLayout(TargetMachine->createDataLayout());
+  TheModule->setTargetTriple(TargetTriple);   
+  
+Emit Object Code
+================
+
+We're ready to emit object code! Let's define where we want to write
+our file to:
+
+.. code-block:: c++
+
+  auto Filename = "output.o";
+  std::error_code EC;
+  raw_fd_ostream dest(Filename, EC, sys::fs::F_None);
+
+  if (EC) {
+    errs() << "Could not open file: " << EC.message();
+    return 1;
+  }
+
+Finally, we define a pass that emits object code, then we run that
+pass:
+
+.. code-block:: c++
+
+  legacy::PassManager pass;
+  auto FileType = TargetMachine::CGFT_ObjectFile;
+
+  if (TargetMachine->addPassesToEmitFile(pass, dest, FileType)) {
+    errs() << "TargetMachine can't emit a file of this type";
+    return 1;
+  }
+
+  pass.run(*TheModule);
+  dest.flush();
+
+Putting It All Together
+=======================
+
+Does it work? Let's give it a try. We need to compile our code, but
+note that the arguments to ``llvm-config`` are different to the previous chapters.
+
+::
+
+    $ clang++ -g -O3 toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs all` -o toy
+
+Let's run it, and define a simple ``average`` function. Press Ctrl-D
+when you're done.
+
+::
+   
+    $ ./toy
+    ready> def average(x y) (x + y) * 0.5;
+    ^D
+    Wrote output.o
+
+We have an object file! To test it, let's write a simple program and
+link it with our output. Here's the source code:
+
+.. code-block:: c++
+
+    #include <iostream>
+
+    extern "C" {
+        double average(double, double);
+    }
+
+    int main() {
+        std::cout << "average of 3.0 and 4.0: " << average(3.0, 4.0) << std::endl;
+    }
+
+We link our program to output.o and check the result is what we
+expected:
+
+::
+
+    $ clang++ main.cpp output.o -o main
+    $ ./main
+    average of 3.0 and 4.0: 3.5
+
+Full Code Listing
+=================
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter8/toy.cpp
+   :language: c++
+
+`Next: Adding Debug Information <LangImpl09.html>`_

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl09.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl09.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl09.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl09.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,462 @@
+======================================
+Kaleidoscope: Adding Debug Information
+======================================
+
+.. contents::
+   :local:
+
+Chapter 9 Introduction
+======================
+
+Welcome to Chapter 9 of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. In chapters 1 through 8, we've built a
+decent little programming language with functions and variables.
+What happens if something goes wrong though, how do you debug your
+program?
+
+Source level debugging uses formatted data that helps a debugger
+translate from binary and the state of the machine back to the
+source that the programmer wrote. In LLVM we generally use a format
+called `DWARF <http://dwarfstd.org>`_. DWARF is a compact encoding
+that represents types, source locations, and variable locations. 
+
+The short summary of this chapter is that we'll go through the
+various things you have to add to a programming language to
+support debug info, and how you translate that into DWARF.
+
+Caveat: For now we can't debug via the JIT, so we'll need to compile
+our program down to something small and standalone. As part of this
+we'll make a few modifications to the running of the language and
+how programs are compiled. This means that we'll have a source file
+with a simple program written in Kaleidoscope rather than the
+interactive JIT. It does involve a limitation that we can only
+have one "top level" command at a time to reduce the number of
+changes necessary.
+
+Here's the sample program we'll be compiling:
+
+.. code-block:: python
+
+   def fib(x)
+     if x < 3 then
+       1
+     else
+       fib(x-1)+fib(x-2);
+
+   fib(10)
+
+
+Why is this a hard problem?
+===========================
+
+Debug information is a hard problem for a few different reasons - mostly
+centered around optimized code. First, optimization makes keeping source
+locations more difficult. In LLVM IR we keep the original source location
+for each IR level instruction on the instruction. Optimization passes
+should keep the source locations for newly created instructions, but merged
+instructions only get to keep a single location - this can cause jumping
+around when stepping through optimized programs. Secondly, optimization
+can move variables in ways that are either optimized out, shared in memory
+with other variables, or difficult to track. For the purposes of this
+tutorial we're going to avoid optimization (as you'll see with one of the
+next sets of patches).
+
+Ahead-of-Time Compilation Mode
+==============================
+
+To highlight only the aspects of adding debug information to a source
+language without needing to worry about the complexities of JIT debugging
+we're going to make a few changes to Kaleidoscope to support compiling
+the IR emitted by the front end into a simple standalone program that
+you can execute, debug, and see results.
+
+First we make our anonymous function that contains our top level
+statement be our "main":
+
+.. code-block:: udiff
+
+  -    auto Proto = llvm::make_unique<PrototypeAST>("", std::vector<std::string>());
+  +    auto Proto = llvm::make_unique<PrototypeAST>("main", std::vector<std::string>());
+
+just with the simple change of giving it a name.
+
+Then we're going to remove the command line code wherever it exists:
+
+.. code-block:: udiff
+
+  @@ -1129,7 +1129,6 @@ static void HandleTopLevelExpression() {
+   /// top ::= definition | external | expression | ';'
+   static void MainLoop() {
+     while (1) {
+  -    fprintf(stderr, "ready> ");
+       switch (CurTok) {
+       case tok_eof:
+         return;
+  @@ -1184,7 +1183,6 @@ int main() {
+     BinopPrecedence['*'] = 40; // highest.
+ 
+     // Prime the first token.
+  -  fprintf(stderr, "ready> ");
+     getNextToken();
+ 
+Lastly we're going to disable all of the optimization passes and the JIT so
+that the only thing that happens after we're done parsing and generating
+code is that the llvm IR goes to standard error:
+
+.. code-block:: udiff
+
+  @@ -1108,17 +1108,8 @@ static void HandleExtern() {
+   static void HandleTopLevelExpression() {
+     // Evaluate a top-level expression into an anonymous function.
+     if (auto FnAST = ParseTopLevelExpr()) {
+  -    if (auto *FnIR = FnAST->codegen()) {
+  -      // We're just doing this to make sure it executes.
+  -      TheExecutionEngine->finalizeObject();
+  -      // JIT the function, returning a function pointer.
+  -      void *FPtr = TheExecutionEngine->getPointerToFunction(FnIR);
+  -
+  -      // Cast it to the right type (takes no arguments, returns a double) so we
+  -      // can call it as a native function.
+  -      double (*FP)() = (double (*)())(intptr_t)FPtr;
+  -      // Ignore the return value for this.
+  -      (void)FP;
+  +    if (!F->codegen()) {
+  +      fprintf(stderr, "Error generating code for top level expr");
+       }
+     } else {
+       // Skip token for error recovery.
+  @@ -1439,11 +1459,11 @@ int main() {
+     // target lays out data structures.
+     TheModule->setDataLayout(TheExecutionEngine->getDataLayout());
+     OurFPM.add(new DataLayoutPass());
+  +#if 0
+     OurFPM.add(createBasicAliasAnalysisPass());
+     // Promote allocas to registers.
+     OurFPM.add(createPromoteMemoryToRegisterPass());
+  @@ -1218,7 +1210,7 @@ int main() {
+     OurFPM.add(createGVNPass());
+     // Simplify the control flow graph (deleting unreachable blocks, etc).
+     OurFPM.add(createCFGSimplificationPass());
+  -
+  +  #endif
+     OurFPM.doInitialization();
+ 
+     // Set the global so the code gen can use this.
+
+This relatively small set of changes get us to the point that we can compile
+our piece of Kaleidoscope language down to an executable program via this
+command line:
+
+.. code-block:: bash
+
+  Kaleidoscope-Ch9 < fib.ks | & clang -x ir -
+
+which gives an a.out/a.exe in the current working directory.
+
+Compile Unit
+============
+
+The top level container for a section of code in DWARF is a compile unit.
+This contains the type and function data for an individual translation unit
+(read: one file of source code). So the first thing we need to do is
+construct one for our fib.ks file.
+
+DWARF Emission Setup
+====================
+
+Similar to the ``IRBuilder`` class we have a
+`DIBuilder <http://llvm.org/doxygen/classllvm_1_1DIBuilder.html>`_ class
+that helps in constructing debug metadata for an llvm IR file. It
+corresponds 1:1 similarly to ``IRBuilder`` and llvm IR, but with nicer names.
+Using it does require that you be more familiar with DWARF terminology than
+you needed to be with ``IRBuilder`` and ``Instruction`` names, but if you
+read through the general documentation on the
+`Metadata Format <http://llvm.org/docs/SourceLevelDebugging.html>`_ it
+should be a little more clear. We'll be using this class to construct all
+of our IR level descriptions. Construction for it takes a module so we
+need to construct it shortly after we construct our module. We've left it
+as a global static variable to make it a bit easier to use.
+
+Next we're going to create a small container to cache some of our frequent
+data. The first will be our compile unit, but we'll also write a bit of
+code for our one type since we won't have to worry about multiple typed
+expressions:
+
+.. code-block:: c++
+
+  static DIBuilder *DBuilder;
+
+  struct DebugInfo {
+    DICompileUnit *TheCU;
+    DIType *DblTy;
+
+    DIType *getDoubleTy();
+  } KSDbgInfo;
+
+  DIType *DebugInfo::getDoubleTy() {
+    if (DblTy.isValid())
+      return DblTy;
+
+    DblTy = DBuilder->createBasicType("double", 64, 64, dwarf::DW_ATE_float);
+    return DblTy;
+  }
+
+And then later on in ``main`` when we're constructing our module:
+
+.. code-block:: c++
+
+  DBuilder = new DIBuilder(*TheModule);
+
+  KSDbgInfo.TheCU = DBuilder->createCompileUnit(
+      dwarf::DW_LANG_C, "fib.ks", ".", "Kaleidoscope Compiler", 0, "", 0);
+
+There are a couple of things to note here. First, while we're producing a
+compile unit for a language called Kaleidoscope we used the language
+constant for C. This is because a debugger wouldn't necessarily understand
+the calling conventions or default ABI for a language it doesn't recognize
+and we follow the C ABI in our llvm code generation so it's the closest
+thing to accurate. This ensures we can actually call functions from the
+debugger and have them execute. Secondly, you'll see the "fib.ks" in the
+call to ``createCompileUnit``. This is a default hard coded value since
+we're using shell redirection to put our source into the Kaleidoscope
+compiler. In a usual front end you'd have an input file name and it would
+go there.
+
+One last thing as part of emitting debug information via DIBuilder is that
+we need to "finalize" the debug information. The reasons are part of the
+underlying API for DIBuilder, but make sure you do this near the end of
+main:
+
+.. code-block:: c++
+
+  DBuilder->finalize();
+
+before you dump out the module.
+
+Functions
+=========
+
+Now that we have our ``Compile Unit`` and our source locations, we can add
+function definitions to the debug info. So in ``PrototypeAST::codegen()`` we
+add a few lines of code to describe a context for our subprogram, in this
+case the "File", and the actual definition of the function itself.
+
+So the context:
+
+.. code-block:: c++
+
+  DIFile *Unit = DBuilder->createFile(KSDbgInfo.TheCU.getFilename(),
+                                      KSDbgInfo.TheCU.getDirectory());
+
+giving us an DIFile and asking the ``Compile Unit`` we created above for the
+directory and filename where we are currently. Then, for now, we use some
+source locations of 0 (since our AST doesn't currently have source location
+information) and construct our function definition:
+
+.. code-block:: c++
+
+  DIScope *FContext = Unit;
+  unsigned LineNo = 0;
+  unsigned ScopeLine = 0;
+  DISubprogram *SP = DBuilder->createFunction(
+      FContext, Name, StringRef(), Unit, LineNo,
+      CreateFunctionType(Args.size(), Unit), false /* internal linkage */,
+      true /* definition */, ScopeLine, DINode::FlagPrototyped, false);
+  F->setSubprogram(SP);
+
+and we now have an DISubprogram that contains a reference to all of our
+metadata for the function.
+
+Source Locations
+================
+
+The most important thing for debug information is accurate source location -
+this makes it possible to map your source code back. We have a problem though,
+Kaleidoscope really doesn't have any source location information in the lexer
+or parser so we'll need to add it.
+
+.. code-block:: c++
+
+   struct SourceLocation {
+     int Line;
+     int Col;
+   };
+   static SourceLocation CurLoc;
+   static SourceLocation LexLoc = {1, 0};
+
+   static int advance() {
+     int LastChar = getchar();
+
+     if (LastChar == '\n' || LastChar == '\r') {
+       LexLoc.Line++;
+       LexLoc.Col = 0;
+     } else
+       LexLoc.Col++;
+     return LastChar;
+   }
+
+In this set of code we've added some functionality on how to keep track of the
+line and column of the "source file". As we lex every token we set our current
+current "lexical location" to the assorted line and column for the beginning
+of the token. We do this by overriding all of the previous calls to
+``getchar()`` with our new ``advance()`` that keeps track of the information
+and then we have added to all of our AST classes a source location:
+
+.. code-block:: c++
+
+   class ExprAST {
+     SourceLocation Loc;
+
+     public:
+       ExprAST(SourceLocation Loc = CurLoc) : Loc(Loc) {}
+       virtual ~ExprAST() {}
+       virtual Value* codegen() = 0;
+       int getLine() const { return Loc.Line; }
+       int getCol() const { return Loc.Col; }
+       virtual raw_ostream &dump(raw_ostream &out, int ind) {
+         return out << ':' << getLine() << ':' << getCol() << '\n';
+       }
+
+that we pass down through when we create a new expression:
+
+.. code-block:: c++
+
+   LHS = llvm::make_unique<BinaryExprAST>(BinLoc, BinOp, std::move(LHS),
+                                          std::move(RHS));
+
+giving us locations for each of our expressions and variables.
+
+From this we can make sure to tell ``DIBuilder`` when we're at a new source
+location so it can use that when we generate the rest of our code and make
+sure that each instruction has source location information. We do this
+by constructing another small function:
+
+.. code-block:: c++
+
+  void DebugInfo::emitLocation(ExprAST *AST) {
+    DIScope *Scope;
+    if (LexicalBlocks.empty())
+      Scope = TheCU;
+    else
+      Scope = LexicalBlocks.back();
+    Builder.SetCurrentDebugLocation(
+        DebugLoc::get(AST->getLine(), AST->getCol(), Scope));
+  }
+
+that both tells the main ``IRBuilder`` where we are, but also what scope
+we're in. Since we've just created a function above we can either be in
+the main file scope (like when we created our function), or now we can be
+in the function scope we just created. To represent this we create a stack
+of scopes:
+
+.. code-block:: c++
+
+   std::vector<DIScope *> LexicalBlocks;
+   std::map<const PrototypeAST *, DIScope *> FnScopeMap;
+
+and keep a map of each function to the scope that it represents (an
+DISubprogram is also an DIScope).
+
+Then we make sure to:
+
+.. code-block:: c++
+
+   KSDbgInfo.emitLocation(this);
+
+emit the location every time we start to generate code for a new AST, and
+also:
+
+.. code-block:: c++
+
+  KSDbgInfo.FnScopeMap[this] = SP;
+
+store the scope (function) when we create it and use it:
+
+  KSDbgInfo.LexicalBlocks.push_back(&KSDbgInfo.FnScopeMap[Proto]);
+
+when we start generating the code for each function.
+
+also, don't forget to pop the scope back off of your scope stack at the
+end of the code generation for the function:
+
+.. code-block:: c++
+
+  // Pop off the lexical block for the function since we added it
+  // unconditionally.
+  KSDbgInfo.LexicalBlocks.pop_back();
+
+Variables
+=========
+
+Now that we have functions, we need to be able to print out the variables
+we have in scope. Let's get our function arguments set up so we can get
+decent backtraces and see how our functions are being called. It isn't
+a lot of code, and we generally handle it when we're creating the
+argument allocas in ``PrototypeAST::CreateArgumentAllocas``.
+
+.. code-block:: c++
+
+  DIScope *Scope = KSDbgInfo.LexicalBlocks.back();
+  DIFile *Unit = DBuilder->createFile(KSDbgInfo.TheCU.getFilename(),
+                                      KSDbgInfo.TheCU.getDirectory());
+  DILocalVariable D = DBuilder->createParameterVariable(
+      Scope, Args[Idx], Idx + 1, Unit, Line, KSDbgInfo.getDoubleTy(), true);
+
+  DBuilder->insertDeclare(Alloca, D, DBuilder->createExpression(),
+                          DebugLoc::get(Line, 0, Scope),
+                          Builder.GetInsertBlock());
+
+Here we're doing a few things. First, we're grabbing our current scope
+for the variable so we can say what range of code our variable is valid
+through. Second, we're creating the variable, giving it the scope,
+the name, source location, type, and since it's an argument, the argument
+index. Third, we create an ``lvm.dbg.declare`` call to indicate at the IR
+level that we've got a variable in an alloca (and it gives a starting
+location for the variable), and setting a source location for the
+beginning of the scope on the declare.
+
+One interesting thing to note at this point is that various debuggers have
+assumptions based on how code and debug information was generated for them
+in the past. In this case we need to do a little bit of a hack to avoid
+generating line information for the function prologue so that the debugger
+knows to skip over those instructions when setting a breakpoint. So in
+``FunctionAST::CodeGen`` we add a couple of lines:
+
+.. code-block:: c++
+
+  // Unset the location for the prologue emission (leading instructions with no
+  // location in a function are considered part of the prologue and the debugger
+  // will run past them when breaking on a function)
+  KSDbgInfo.emitLocation(nullptr);
+
+and then emit a new location when we actually start generating code for the
+body of the function:
+
+.. code-block:: c++
+
+  KSDbgInfo.emitLocation(Body);
+
+With this we have enough debug information to set breakpoints in functions,
+print out argument variables, and call functions. Not too bad for just a
+few simple lines of code!
+
+Full Code Listing
+=================
+
+Here is the complete code listing for our running example, enhanced with
+debug information. To build this example, use:
+
+.. code-block:: bash
+
+    # Compile
+    clang++ -g toy.cpp `llvm-config --cxxflags --ldflags --system-libs --libs core mcjit native` -O3 -o toy
+    # Run
+    ./toy
+
+Here is the code:
+
+.. literalinclude:: ../../examples/Kaleidoscope/Chapter9/toy.cpp
+   :language: c++
+
+`Next: Conclusion and other useful LLVM tidbits <LangImpl10.html>`_
+

Added: www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl10.txt
URL: http://llvm.org/viewvc/llvm-project/www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl10.txt?rev=297634&view=auto
==============================================================================
--- www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl10.txt (added)
+++ www-releases/trunk/4.0.0/docs/_sources/tutorial/LangImpl10.txt Mon Mar 13 11:30:12 2017
@@ -0,0 +1,259 @@
+======================================================
+Kaleidoscope: Conclusion and other useful LLVM tidbits
+======================================================
+
+.. contents::
+   :local:
+
+Tutorial Conclusion
+===================
+
+Welcome to the final chapter of the "`Implementing a language with
+LLVM <index.html>`_" tutorial. In the course of this tutorial, we have
+grown our little Kaleidoscope language from being a useless toy, to
+being a semi-interesting (but probably still useless) toy. :)
+
+It is interesting to see how far we've come, and how little code it has
+taken. We built the entire lexer, parser, AST, code generator, an
+interactive run-loop (with a JIT!), and emitted debug information in
+standalone executables - all in under 1000 lines of (non-comment/non-blank)
+code.
+
+Our little language supports a couple of interesting features: it
+supports user defined binary and unary operators, it uses JIT
+compilation for immediate evaluation, and it supports a few control flow
+constructs with SSA construction.
+
+Part of the idea of this tutorial was to show you how easy and fun it
+can be to define, build, and play with languages. Building a compiler
+need not be a scary or mystical process! Now that you've seen some of
+the basics, I strongly encourage you to take the code and hack on it.
+For example, try adding:
+
+-  **global variables** - While global variables have questional value
+   in modern software engineering, they are often useful when putting
+   together quick little hacks like the Kaleidoscope compiler itself.
+   Fortunately, our current setup makes it very easy to add global
+   variables: just have value lookup check to see if an unresolved
+   variable is in the global variable symbol table before rejecting it.
+   To create a new global variable, make an instance of the LLVM
+   ``GlobalVariable`` class.
+-  **typed variables** - Kaleidoscope currently only supports variables
+   of type double. This gives the language a very nice elegance, because
+   only supporting one type means that you never have to specify types.
+   Different languages have different ways of handling this. The easiest
+   way is to require the user to specify types for every variable
+   definition, and record the type of the variable in the symbol table
+   along with its Value\*.
+-  **arrays, structs, vectors, etc** - Once you add types, you can start
+   extending the type system in all sorts of interesting ways. Simple
+   arrays are very easy and are quite useful for many different
+   applications. Adding them is mostly an exercise in learning how the
+   LLVM `getelementptr <../LangRef.html#getelementptr-instruction>`_ instruction
+   works: it is so nifty/unconventional, it `has its own
+   FAQ <../GetElementPtr.html>`_!
+-  **standard runtime** - Our current language allows the user to access
+   arbitrary external functions, and we use it for things like "printd"
+   and "putchard". As you extend the language to add higher-level
+   constructs, often these constructs make the most sense if they are
+   lowered to calls into a language-supplied runtime. For example, if
+   you add hash tables to the language, it would probably make sense to
+   add the routines to a runtime, instead of inlining them all the way.
+-  **memory management** - Currently we can only access the stack in
+   Kaleidoscope. It would also be useful to be able to allocate heap
+   memory, either with calls to the standard libc malloc/free interface
+   or with a garbage collector. If you would like to use garbage
+   collection, note that LLVM fully supports `Accurate Garbage
+   Collection <../GarbageCollection.html>`_ including algorithms that
+   move objects and need to scan/update the stack.
+-  **exception handling support** - LLVM supports generation of `zero
+   cost exceptions <../ExceptionHandling.html>`_ which interoperate with
+   code compiled in other languages. You could also generate code by
+   implicitly making every function return an error value and checking
+   it. You could also make explicit use of setjmp/longjmp. There are
+   many different ways to go here.
+-  **object orientation, generics, database access, complex numbers,
+   geometric programming, ...** - Really, there is no end of crazy
+   features that you can add to the language.
+-  **unusual domains** - We've been talking about applying LLVM to a
+   domain that many people are interested in: building a compiler for a
+   specific language. However, there are many other domains that can use
+   compiler technology that are not typically considered. For example,
+   LLVM has been used to implement OpenGL graphics acceleration,
+   translate C++ code to ActionScript, and many other cute and clever
+   things. Maybe you will be the first to JIT compile a regular
+   expression interpreter into native code with LLVM?
+
+Have fun - try doing something crazy and unusual. Building a language
+like everyone else always has, is much less fun than trying something a
+little crazy or off the wall and seeing how it turns out. If you get
+stuck or want to talk about it, feel free to email the `llvm-dev mailing
+list <http://lists.llvm.org/mailman/listinfo/llvm-dev>`_: it has lots
+of people who are interested in languages and are often willing to help
+out.
+
+Before we end this tutorial, I want to talk about some "tips and tricks"
+for generating LLVM IR. These are some of the more subtle things that
+may not be obvious, but are very useful if you want to take advantage of
+LLVM's capabilities.
+
+Properties of the LLVM IR
+=========================
+
+We have a couple of common questions about code in the LLVM IR form -
+let's just get these out of the way right now, shall we?
+
+Target Independence
+-------------------
+
+Kaleidoscope is an example of a "portable language": any program written
+in Kaleidoscope will work the same way on any target that it runs on.
+Many other languages have this property, e.g. lisp, java, haskell,
+javascript, python, etc (note that while these languages are portable,
+not all their libraries are).
+
+One nice aspect of LLVM is that it is often capable of preserving target
+independence in the IR: you can take the LLVM IR for a
+Kaleidoscope-compiled program and run it on any target that LLVM
+supports, even emitting C code and compiling that on targets that LLVM
+doesn't support natively. You can trivially tell that the Kaleidoscope
+compiler generates target-independent code because it never queries for
+any target-specific information when generating code.
+
+The fact that LLVM provides a compact, target-independent,
+representation for code gets a lot of people excited. Unfortunately,
+these people are usually thinking about C or a language from the C
+family when they are asking questions about language portability. I say
+"unfortunately", because there is really no way to make (fully general)
+C code portable, other than shipping the source code around (and of
+course, C source code is not actually portable in general either - ever
+port a really old application from 32- to 64-bits?).
+
+The problem with C (again, in its full generality) is that it is heavily
+laden with target specific assumptions. As one simple example, the
+preprocessor often destructively removes target-independence from the
+code when it processes the input text:
+
+.. code-block:: c
+
+    #ifdef __i386__
+      int X = 1;
+    #else
+      int X = 42;
+    #endif
+
+While it is possible to engineer more and more complex solutions to
+problems like this, it cannot be solved in full generality in a way that
+is better than shipping the actual source code.
+
+That said, there are interesting subsets of C that can be made portable.
+If you are willing to fix primitive types to a fixed size (say int =
+32-bits, and long = 64-bits), don't care about ABI compatibility with
+existing binaries, and are willing to give up some other minor features,
+you can have portable code. This can make sense for specialized domains
+such as an in-kernel language.
+
+Safety Guarantees
+-----------------
+
+Many of the languages above are also "safe" languages: it is impossible
+for a program written in Java to corrupt its address space and crash the
+process (assuming the JVM has no bugs). Safety is an interesting
+property that requires a combination of language design, runtime
+support, and often operating system support.
+
+It is certainly possible to implement a safe language in LLVM, but LLVM
+IR does not itself guarantee safety. The LLVM IR allows unsafe pointer
+casts, use after free bugs, buffer over-runs, and a variety of other
+problems. Safety needs to be implemented as a layer on top of LLVM and,
+conveniently, several groups have investigated this. Ask on the `llvm-dev
+mailing list <http://lists.llvm.org/mailman/listinfo/llvm-dev>`_ if
+you are interested in more details.
+
+Language-Specific Optimizations
+-------------------------------
+
+One thing about LLVM that turns off many people is that it does not
+solve all the world's problems in one system (sorry 'world hunger',
+someone else will have to solve you some other day). One specific
+complaint is that people perceive LLVM as being incapable of performing
+high-level language-specific optimization: LLVM "loses too much
+information".
+
+Unfortunately, this is really not the place to give you a full and
+unified version of "Chris Lattner's theory of compiler design". Instead,
+I'll make a few observations:
+
+First, you're right that LLVM does lose information. For example, as of
+this writing, there is no way to distinguish in the LLVM IR whether an
+SSA-value came from a C "int" or a C "long" on an ILP32 machine (other
+than debug info). Both get compiled down to an 'i32' value and the
+information about what it came from is lost. The more general issue
+here, is that the LLVM type system uses "structural equivalence" instead
+of "name equivalence". Another place this surprises people is if you
+have two types in a high-level language that have the same structure
+(e.g. two different structs that have a single int field): these types
+will compile down into a single LLVM type and it will be impossible to
+tell what it came from.
+
+Second, while LLVM does lose information, LLVM is not a fixed target: we
+continue to enhance and improve it in many different ways. In addition
+to adding new features (LLVM did not always support exceptions or debug
+info), we also extend the IR to capture important information for
+optimization (e.g. whether an argument is sign or zero extended,
+information about pointers aliasing, etc). Many of the enhancements are
+user-driven: people want LLVM to include some specific feature, so they
+go ahead and extend it.
+
+Third, it is *possible and easy* to add language-specific optimizations,
+and you have a number of choices in how to do it. As one trivial
+example, it is easy to add language-specific optimization passes that
+"know" things about code compiled for a language. In the case of the C
+family, there is an optimization pass that "knows" about the standard C
+library functions. If you call "exit(0)" in main(), it knows that it is
+safe to optimize that into "return 0;" because C specifies what the
+'exit' function does.
+
+In addition to simple library knowledge, it is possible to embed a
+variety of other language-specific information into the LLVM IR. If you
+have a specific need and run into a wall, please bring the topic up on
+the llvm-dev list. At the very worst, you can always treat LLVM as if it
+were a "dumb code generator" and implement the high-level optimizations
+you desire in your front-end, on the language-specific AST.
+
+Tips and Tricks
+===============
+
+There is a variety of useful tips and tricks that you come to know after
+working on/with LLVM that aren't obvious at first glance. Instead of
+letting everyone rediscover them, this section talks about some of these
+issues.
+
+Implementing portable offsetof/sizeof
+-------------------------------------
+
+One interesting thing that comes up, if you are trying to keep the code
+generated by your compiler "target independent", is that you often need
+to know the size of some LLVM type or the offset of some field in an
+llvm structure. For example, you might need to pass the size of a type
+into a function that allocates memory.
+
+Unfortunately, this can vary widely across targets: for example the
+width of a pointer is trivially target-specific. However, there is a
+`clever way to use the getelementptr
+instruction <http://nondot.org/sabre/LLVMNotes/SizeOf-OffsetOf-VariableSizedStructs.txt>`_
+that allows you to compute this in a portable way.
+
+Garbage Collected Stack Frames
+------------------------------
+
+Some languages want to explicitly manage their stack frames, often so
+that they are garbage collected or to allow easy implementation of
+closures. There are often better ways to implement these features than
+explicit stack frames, but `LLVM does support
+them, <http://nondot.org/sabre/LLVMNotes/ExplicitlyManagedStackFrames.txt>`_
+if you want. It requires your front-end to convert the code into
+`Continuation Passing
+Style <http://en.wikipedia.org/wiki/Continuation-passing_style>`_ and
+the use of tail calls (which LLVM also supports).
+




More information about the llvm-commits mailing list