docs/ExtendingLLVM.html - external/github.com/llvm-mirror/llvm - Git at Google

 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
                       "http://www.w3.org/TR/html4/strict.dtd">
 <html>
 <head>
   <title>Extending LLVM: Adding instructions, intrinsics, types, etc.</title>
   <link rel="stylesheet" href="llvm.css" type="text/css">
 </head>

 <body>

 <div class="doc_title">
   Extending LLVM: Adding instructions, intrinsics, types, etc.
 </div>

 <ol>
   <li><a href="#introduction">Introduction and Warning</a></li>
   <li><a href="#intrinsic">Adding a new intrinsic function</a></li>
   <li><a href="#instruction">Adding a new instruction</a></li>
   <li><a href="#type">Adding a new type</a>
   <ol>
     <li><a href="#fund_type">Adding a new fundamental type</a></li>
     <li><a href="#derived_type">Adding a new derived type</a></li>
   </ol></li>
 </ol>

 <div class="doc_author">
   <p>Written by <a href="http://misha.brukman.net">Misha Brukman</a>,
   Brad Jones, and <a href="http://nondot.org/sabre">Chris Lattner</a></p>
 </div>

 <!-- *********************************************************************** -->
 <div class="doc_section">
   <a name="introduction">Introduction and Warning</a>
 </div>
 <!-- *********************************************************************** -->

 <div class="doc_text">

 <p>During the course of using LLVM, you may wish to customize it for your
 research project or for experimentation. At this point, you may realize that
 you need to add something to LLVM, whether it be a new fundamental type, a new
 intrinsic function, or a whole new instruction.</p>

 <p>When you come to this realization, stop and think. Do you really need to
 extend LLVM? Is it a new fundamental capability that LLVM does not support at
 its current incarnation or can it be synthesized from already pre-existing LLVM
 elements? If you are not sure, ask on the <a
 href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM-dev</a> list. The
 reason is that extending LLVM will get involved as you need to update all the
 different passes that you intend to use with your extension, and there are
 <em>many</em> LLVM analyses and transformations, so it may be quite a bit of
 work.</p>

 <p>Adding an <a href="#intrinsic">intrinsic function</a> is easier than adding
 an instruction, and is transparent to optimization passes which treat it as an
 unanalyzable function.  If your added functionality can be expressed as a
 function call, an intrinsic function is the method of choice for LLVM
 extension.</p>

 <p>Before you invest a significant amount of effort into a non-trivial
 extension, <span class="doc_warning">ask on the list</span> if what you are
 looking to do can be done with already-existing infrastructure, or if maybe
 someone else is already working on it. You will save yourself a lot of time and
 effort by doing so.</p>

 </div>

 <!-- *********************************************************************** -->
 <div class="doc_section">
   <a name="intrinsic">Adding a new intrinsic function</a>
 </div>
 <!-- *********************************************************************** -->

 <div class="doc_text">

 <p>Adding a new intrinsic function to LLVM is much easier than adding a new
 instruction.  Almost all extensions to LLVM should start as an intrinsic
 function and then be turned into an instruction if warranted.</p>

 <ol>
 <li><tt>llvm/docs/LangRef.html</tt>:
     Document the intrinsic.  Decide whether it is code generator specific and
     what the restrictions are.  Talk to other people about it so that you are
     sure it's a good idea.</li>

 <li><tt>llvm/include/llvm/Intrinsics.h</tt>:
     add an enum in the <tt>llvm::Intrinsic</tt> namespace</li>

 <li><tt>llvm/lib/VMCore/Verifier.cpp</tt>:
     Add code to check the invariants of the intrinsic are respected.</li>

 <li><tt>llvm/lib/VMCore/Function.cpp (<tt>Function::getIntrinsicID()</tt>)</tt>:
     Identify the new intrinsic function, returning the enum for the intrinsic
     that you added.</li>

 <li><tt>llvm/lib/Analysis/BasicAliasAnalysis.cpp</tt>: If the new intrinsic does
     not access memory or does not write to memory, add it to the relevant list
     of functions.</li>

 <li><tt>llvm/lib/Transforms/Utils/Local.cpp</tt>: If it is possible to constant
 fold your intrinsic, add support to it in the <tt>canConstantFoldCallTo</tt> and
 <tt>ConstantFoldCall</tt> functions.</li>

 <li>Test your intrinsic</li>

 <li><tt>llvm/test/Regression/*</tt>: add your test cases to the test suite</li>
 </ol>

 <p>Once the intrinsic has been added to the system, you must add code generator
 support for it.  Generally you must do the following steps:</p>

 <dl>
 <dt>Add support to the C backend in <tt>lib/Target/CBackend/</tt></dt>

 <dd>Depending on the intrinsic, there are a few ways to implement this.  First,
 if it makes sense to lower the intrinsic to an expanded sequence of C code in
 all cases, just emit the expansion in <tt>visitCallInst</tt>.  Second, if the
 intrinsic has some way to express it with GCC (or any other compiler)
 extensions, it can be conditionally supported based on the compiler compiling
 the CBE output (see llvm.prefetch for an example).  Third, if the intrinsic
 really has no way to be lowered, just have the code generator emit code that
 prints an error message and calls abort if executed.
 </dd>

 <dt>Add a enum value for the SelectionDAG node in
 <tt>include/llvm/CodeGen/SelectionDAGNodes.h</tt></dt>

 <dd>Also, add code to <tt>lib/CodeGen/SelectionDAG/SelectionDAG.cpp</tt> (and
 <tt>SelectionDAGPrinter.cpp</tt>) to print the node.</dd>

 <dt>Add code to <tt>SelectionDAG/SelectionDAGISel.cpp</tt> to recognize the
 intrinsic.</dt>

 <dd>Presumably the intrinsic should be recognized and turned into the node you
 added above.</dd>

 <dt>Add code to <tt>SelectionDAG/LegalizeDAG.cpp</tt> to <a
 href="CodeGenerator.html#selectiondag_legalize">legalize, promote, and
 expand</a> the node as necessary.</dt>

 <dd>If the intrinsic can be expanded to primitive operations, legalize can break
 the node down into other elementary operations that are be supported.</dd>

 <dt>Add target-specific support to specific code generators.</dt>

 <dd>Extend the code generators you are interested in to recognize and support
 the node, emitting the code you want.</dd>
 </dl>

 <p>
 Unfortunately, the process of extending the code generator to support a new node
 is not extremely well documented.  As such, it is often helpful to look at other
 intrinsics (e.g. <tt>llvm.ctpop</tt>) to see how they are recognized and turned
 into a node by <tt>SelectionDAGISel.cpp</tt>, legalized by
 <tt>LegalizeDAG.cpp</tt>, then finally emitted by the various code generators.
 </p>

 </div>

 <!-- *********************************************************************** -->
 <div class="doc_section">
   <a name="instruction">Adding a new instruction</a>
 </div>
 <!-- *********************************************************************** -->

 <div class="doc_text">

 <p><span class="doc_warning">WARNING: adding instructions changes the bytecode
 format, and it will take some effort to maintain compatibility with
 the previous version.</span> Only add an instruction if it is absolutely
 necessary.</p>

 <ol>

 <li><tt>llvm/include/llvm/Instruction.def</tt>:
     add a number for your instruction and an enum name</li>

 <li><tt>llvm/include/llvm/Instructions.h</tt>:
     add a definition for the class that will represent your instruction</li>

 <li><tt>llvm/include/llvm/Support/InstVisitor.h</tt>:
     add a prototype for a visitor to your new instruction type</li>

 <li><tt>llvm/lib/AsmParser/Lexer.l</tt>:
     add a new token to parse your instruction from assembly text file</li>

 <li><tt>llvm/lib/AsmParser/llvmAsmParser.y</tt>:
     add the grammar on how your instruction can be read and what it will
     construct as a result</li>

 <li><tt>llvm/lib/Bytecode/Reader/Reader.cpp</tt>:
     add a case for your instruction and how it will be parsed from bytecode</li>

 <li><tt>llvm/lib/VMCore/Instruction.cpp</tt>:
     add a case for how your instruction will be printed out to assembly</li>

 <li><tt>llvm/lib/VMCore/Instructions.cpp</tt>:
     implement the class you defined in
     <tt>llvm/include/llvm/Instructions.h</tt></li>

 <li>Test your instruction</li>

 <li><tt>llvm/lib/Target/*</tt>:
     Add support for your instruction to code generators, or add a lowering
     pass.</li>

 <li><tt>llvm/test/Regression/*</tt>: add your test cases to the test suite.</li>

 </ol>

 <p>Also, you need to implement (or modify) any analyses or passes that you want
 to understand this new instruction.</p>

 </div>


 <!-- *********************************************************************** -->
 <div class="doc_section">
   <a name="type">Adding a new type</a>
 </div>
 <!-- *********************************************************************** -->

 <div class="doc_text">

 <p><span class="doc_warning">WARNING: adding new types changes the bytecode
 format, and will break compatibility with currently-existing LLVM
 installations.</span> Only add new types if it is absolutely necessary.</p>

 </div>

 <!-- ======================================================================= -->
 <div class="doc_subsection">
   <a name="fund_type">Adding a fundamental type</a>
 </div>

 <div class="doc_text">

 <ol>

 <li><tt>llvm/include/llvm/Type.h</tt>:
     add enum for the new type; add static <tt>Type*</tt> for this type</li>

 <li><tt>llvm/lib/VMCore/Type.cpp</tt>:
     add mapping from <tt>TypeID</tt> =&gt; <tt>Type*</tt>;
     initialize the static <tt>Type*</tt></li>

 <li><tt>llvm/lib/AsmReader/Lexer.l</tt>:
     add ability to parse in the type from text assembly</li>

 <li><tt>llvm/lib/AsmReader/llvmAsmParser.y</tt>:
     add a token for that type</li>

 </ol>

 </div>

 <!-- ======================================================================= -->
 <div class="doc_subsection">
   <a name="derived_type">Adding a derived type</a>
 </div>

 <div class="doc_text">

 <ol>
 <li><tt>llvm/include/llvm/Type.h</tt>:
     add enum for the new type; add a forward declaration of the type
     also</li>

 <li><tt>llvm/include/llvm/DerivedType.h</tt>:
     add new class to represent new class in the hierarchy; add forward
     declaration to the TypeMap value type</li>

 <li><tt>llvm/lib/VMCore/Type.cpp</tt>:
     add support for derived type to:
 <div class="doc_code">
 <pre>
 std::string getTypeDescription(const Type &amp;Ty,
   std::vector&lt;const Type*&gt; &amp;TypeStack)
 bool TypesEqual(const Type *Ty, const Type *Ty2,
   std::map&lt;const Type*, const Type*&gt; &amp; EqTypes)
 </pre>
 </div>
     add necessary member functions for type, and factory methods</li>

 <li><tt>llvm/lib/AsmReader/Lexer.l</tt>:
     add ability to parse in the type from text assembly</li>

 <li><tt>llvm/lib/ByteCode/Writer/Writer.cpp</tt>:
     modify <tt>void BytecodeWriter::outputType(const Type *T)</tt> to serialize
     your type</li>

 <li><tt>llvm/lib/ByteCode/Reader/Reader.cpp</tt>:
     modify <tt>const Type *BytecodeReader::ParseType()</tt> to read your data
     type</li>

 <li><tt>llvm/lib/VMCore/AsmWriter.cpp</tt>:
     modify
 <div class="doc_code">
 <pre>
 void calcTypeName(const Type *Ty,
                   std::vector&lt;const Type*&gt; &amp;TypeStack,
                   std::map&lt;const Type*,std::string&gt; &amp;TypeNames,
                   std::string &amp; Result)
 </pre>
 </div>
     to output the new derived type
 </li>


 </ol>

 </div>

 <!-- *********************************************************************** -->

 <hr>
 <address>
   <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
   src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
   <a href="http://validator.w3.org/check/referer"><img
   src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>

   <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a>
   <br>
   Last modified: $Date$
 </address>

 </body>
 </html>
	<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
	"http://www.w3.org/TR/html4/strict.dtd">
	<html>
	<head>
	<title>Extending LLVM: Adding instructions, intrinsics, types, etc.</title>
	<link rel="stylesheet" href="llvm.css" type="text/css">
	</head>

	<body>

	<div class="doc_title">
	Extending LLVM: Adding instructions, intrinsics, types, etc.
	</div>

	<ol>
	<li><a href="#introduction">Introduction and Warning</a></li>
	<li><a href="#intrinsic">Adding a new intrinsic function</a></li>
	<li><a href="#instruction">Adding a new instruction</a></li>
	<li><a href="#type">Adding a new type</a>
	<ol>
	<li><a href="#fund_type">Adding a new fundamental type</a></li>
	<li><a href="#derived_type">Adding a new derived type</a></li>
	</ol></li>
	</ol>

	<div class="doc_author">
	<p>Written by <a href="http://misha.brukman.net">Misha Brukman</a>,
	Brad Jones, and <a href="http://nondot.org/sabre">Chris Lattner</a></p>
	</div>

	<!-- *********************************************************************** -->
	<div class="doc_section">
	<a name="introduction">Introduction and Warning</a>
	</div>
	<!-- *********************************************************************** -->

	<div class="doc_text">

	<p>During the course of using LLVM, you may wish to customize it for your
	research project or for experimentation. At this point, you may realize that
	you need to add something to LLVM, whether it be a new fundamental type, a new
	intrinsic function, or a whole new instruction.</p>

	<p>When you come to this realization, stop and think. Do you really need to
	extend LLVM? Is it a new fundamental capability that LLVM does not support at
	its current incarnation or can it be synthesized from already pre-existing LLVM
	elements? If you are not sure, ask on the <a
	href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM-dev</a> list. The
	reason is that extending LLVM will get involved as you need to update all the
	different passes that you intend to use with your extension, and there are
	<em>many</em> LLVM analyses and transformations, so it may be quite a bit of
	work.</p>

	<p>Adding an <a href="#intrinsic">intrinsic function</a> is easier than adding
	an instruction, and is transparent to optimization passes which treat it as an
	unanalyzable function. If your added functionality can be expressed as a
	function call, an intrinsic function is the method of choice for LLVM
	extension.</p>

	<p>Before you invest a significant amount of effort into a non-trivial
	extension, <span class="doc_warning">ask on the list</span> if what you are
	looking to do can be done with already-existing infrastructure, or if maybe
	someone else is already working on it. You will save yourself a lot of time and
	effort by doing so.</p>

	</div>

	<!-- *********************************************************************** -->
	<div class="doc_section">
	<a name="intrinsic">Adding a new intrinsic function</a>
	</div>
	<!-- *********************************************************************** -->

	<div class="doc_text">

	<p>Adding a new intrinsic function to LLVM is much easier than adding a new
	instruction. Almost all extensions to LLVM should start as an intrinsic
	function and then be turned into an instruction if warranted.</p>

	<ol>
	<li><tt>llvm/docs/LangRef.html</tt>:
	Document the intrinsic. Decide whether it is code generator specific and
	what the restrictions are. Talk to other people about it so that you are
	sure it's a good idea.</li>

	<li><tt>llvm/include/llvm/Intrinsics.h</tt>:
	add an enum in the <tt>llvm::Intrinsic</tt> namespace</li>

	<li><tt>llvm/lib/VMCore/Verifier.cpp</tt>:
	Add code to check the invariants of the intrinsic are respected.</li>

	<li><tt>llvm/lib/VMCore/Function.cpp (<tt>Function::getIntrinsicID()</tt>)</tt>:
	Identify the new intrinsic function, returning the enum for the intrinsic
	that you added.</li>

	<li><tt>llvm/lib/Analysis/BasicAliasAnalysis.cpp</tt>: If the new intrinsic does
	not access memory or does not write to memory, add it to the relevant list
	of functions.</li>

	<li><tt>llvm/lib/Transforms/Utils/Local.cpp</tt>: If it is possible to constant
	fold your intrinsic, add support to it in the <tt>canConstantFoldCallTo</tt> and
	<tt>ConstantFoldCall</tt> functions.</li>

	<li>Test your intrinsic</li>

	<li><tt>llvm/test/Regression/*</tt>: add your test cases to the test suite</li>
	</ol>

	<p>Once the intrinsic has been added to the system, you must add code generator
	support for it. Generally you must do the following steps:</p>

	<dl>
	<dt>Add support to the C backend in <tt>lib/Target/CBackend/</tt></dt>

	<dd>Depending on the intrinsic, there are a few ways to implement this. First,
	if it makes sense to lower the intrinsic to an expanded sequence of C code in
	all cases, just emit the expansion in <tt>visitCallInst</tt>. Second, if the
	intrinsic has some way to express it with GCC (or any other compiler)
	extensions, it can be conditionally supported based on the compiler compiling
	the CBE output (see llvm.prefetch for an example). Third, if the intrinsic
	really has no way to be lowered, just have the code generator emit code that
	prints an error message and calls abort if executed.
	</dd>

	<dt>Add a enum value for the SelectionDAG node in
	<tt>include/llvm/CodeGen/SelectionDAGNodes.h</tt></dt>

	<dd>Also, add code to <tt>lib/CodeGen/SelectionDAG/SelectionDAG.cpp</tt> (and
	<tt>SelectionDAGPrinter.cpp</tt>) to print the node.</dd>

	<dt>Add code to <tt>SelectionDAG/SelectionDAGISel.cpp</tt> to recognize the
	intrinsic.</dt>

	<dd>Presumably the intrinsic should be recognized and turned into the node you
	added above.</dd>

	<dt>Add code to <tt>SelectionDAG/LegalizeDAG.cpp</tt> to <a
	href="CodeGenerator.html#selectiondag_legalize">legalize, promote, and
	expand</a> the node as necessary.</dt>

	<dd>If the intrinsic can be expanded to primitive operations, legalize can break
	the node down into other elementary operations that are be supported.</dd>

	<dt>Add target-specific support to specific code generators.</dt>

	<dd>Extend the code generators you are interested in to recognize and support
	the node, emitting the code you want.</dd>
	</dl>

	<p>
	Unfortunately, the process of extending the code generator to support a new node
	is not extremely well documented. As such, it is often helpful to look at other
	intrinsics (e.g. <tt>llvm.ctpop</tt>) to see how they are recognized and turned
	into a node by <tt>SelectionDAGISel.cpp</tt>, legalized by
	<tt>LegalizeDAG.cpp</tt>, then finally emitted by the various code generators.
	</p>

	</div>

	<!-- *********************************************************************** -->
	<div class="doc_section">
	<a name="instruction">Adding a new instruction</a>
	</div>
	<!-- *********************************************************************** -->

	<div class="doc_text">

	<p><span class="doc_warning">WARNING: adding instructions changes the bytecode
	format, and it will take some effort to maintain compatibility with
	the previous version.</span> Only add an instruction if it is absolutely
	necessary.</p>

	<ol>

	<li><tt>llvm/include/llvm/Instruction.def</tt>:
	add a number for your instruction and an enum name</li>

	<li><tt>llvm/include/llvm/Instructions.h</tt>:
	add a definition for the class that will represent your instruction</li>

	<li><tt>llvm/include/llvm/Support/InstVisitor.h</tt>:
	add a prototype for a visitor to your new instruction type</li>

	<li><tt>llvm/lib/AsmParser/Lexer.l</tt>:
	add a new token to parse your instruction from assembly text file</li>

	<li><tt>llvm/lib/AsmParser/llvmAsmParser.y</tt>:
	add the grammar on how your instruction can be read and what it will
	construct as a result</li>

	<li><tt>llvm/lib/Bytecode/Reader/Reader.cpp</tt>:
	add a case for your instruction and how it will be parsed from bytecode</li>

	<li><tt>llvm/lib/VMCore/Instruction.cpp</tt>:
	add a case for how your instruction will be printed out to assembly</li>

	<li><tt>llvm/lib/VMCore/Instructions.cpp</tt>:
	implement the class you defined in
	<tt>llvm/include/llvm/Instructions.h</tt></li>

	<li>Test your instruction</li>

	<li><tt>llvm/lib/Target/*</tt>:
	Add support for your instruction to code generators, or add a lowering
	pass.</li>

	<li><tt>llvm/test/Regression/*</tt>: add your test cases to the test suite.</li>

	</ol>

	<p>Also, you need to implement (or modify) any analyses or passes that you want
	to understand this new instruction.</p>

	</div>


	<!-- *********************************************************************** -->
	<div class="doc_section">
	<a name="type">Adding a new type</a>
	</div>
	<!-- *********************************************************************** -->

	<div class="doc_text">

	<p><span class="doc_warning">WARNING: adding new types changes the bytecode
	format, and will break compatibility with currently-existing LLVM
	installations.</span> Only add new types if it is absolutely necessary.</p>

	</div>

	<!-- ======================================================================= -->
	<div class="doc_subsection">
	<a name="fund_type">Adding a fundamental type</a>
	</div>

	<div class="doc_text">

	<ol>

	<li><tt>llvm/include/llvm/Type.h</tt>:
	add enum for the new type; add static <tt>Type*</tt> for this type</li>

	<li><tt>llvm/lib/VMCore/Type.cpp</tt>:
	add mapping from <tt>TypeID</tt> => <tt>Type*</tt>;
	initialize the static <tt>Type*</tt></li>

	<li><tt>llvm/lib/AsmReader/Lexer.l</tt>:
	add ability to parse in the type from text assembly</li>

	<li><tt>llvm/lib/AsmReader/llvmAsmParser.y</tt>:
	add a token for that type</li>

	</ol>

	</div>

	<!-- ======================================================================= -->
	<div class="doc_subsection">
	<a name="derived_type">Adding a derived type</a>
	</div>

	<div class="doc_text">

	<ol>
	<li><tt>llvm/include/llvm/Type.h</tt>:
	add enum for the new type; add a forward declaration of the type
	also</li>

	<li><tt>llvm/include/llvm/DerivedType.h</tt>:
	add new class to represent new class in the hierarchy; add forward
	declaration to the TypeMap value type</li>

	<li><tt>llvm/lib/VMCore/Type.cpp</tt>:
	add support for derived type to:
	<div class="doc_code">
	<pre>
	std::string getTypeDescription(const Type &Ty,
	std::vector<const Type*> &TypeStack)
	bool TypesEqual(const Type Ty, const Type Ty2,
	std::map<const Type, const Type> & EqTypes)
	</pre>
	</div>
	add necessary member functions for type, and factory methods</li>

	<li><tt>llvm/lib/AsmReader/Lexer.l</tt>:
	add ability to parse in the type from text assembly</li>

	<li><tt>llvm/lib/ByteCode/Writer/Writer.cpp</tt>:
	modify <tt>void BytecodeWriter::outputType(const Type *T)</tt> to serialize
	your type</li>

	<li><tt>llvm/lib/ByteCode/Reader/Reader.cpp</tt>:
	modify <tt>const Type *BytecodeReader::ParseType()</tt> to read your data
	type</li>

	<li><tt>llvm/lib/VMCore/AsmWriter.cpp</tt>:
	modify
	<div class="doc_code">
	<pre>
	void calcTypeName(const Type *Ty,
	std::vector<const Type*> &TypeStack,
	std::map<const Type*,std::string> &TypeNames,
	std::string & Result)
	</pre>
	</div>
	to output the new derived type
	</li>


	</ol>

	</div>

	<!-- *********************************************************************** -->

	<hr>
	<address>
	<a href="http://jigsaw.w3.org/css-validator/check/referer"><img
	src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
	<a href="http://validator.w3.org/check/referer"><img
	src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>

	<a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a>
	<br>
	Last modified: $Date$
	</address>

	</body>
	</html>