1 <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
2 "http://www.w3.org/TR/html4/strict.dtd">
5 <title>Extending LLVM: Adding instructions, intrinsics, types, etc.</title>
6 <link rel="stylesheet" href="llvm.css" type="text/css">
11 <div class="doc_title">
12 Extending LLVM: Adding instructions, intrinsics, types, etc.
16 <li><a href="#introduction">Introduction and Warning</a></li>
17 <li><a href="#intrinsic">Adding a new intrinsic function</a></li>
18 <li><a href="#instruction">Adding a new instruction</a></li>
19 <li><a href="#type">Adding a new type</a>
21 <li><a href="#fund_type">Adding a new fundamental type</a></li>
22 <li><a href="#derived_type">Adding a new derived type</a></li>
26 <div class="doc_author">
27 <p>Written by <a href="http://misha.brukman.net">Misha Brukman</a>,
28 Brad Jones, and <a href="http://nondot.org/sabre">Chris Lattner</a></p>
31 <!-- *********************************************************************** -->
32 <div class="doc_section">
33 <a name="introduction">Introduction and Warning</a>
35 <!-- *********************************************************************** -->
37 <div class="doc_text">
39 <p>During the course of using LLVM, you may wish to customize it for your
40 research project or for experimentation. At this point, you may realize that
41 you need to add something to LLVM, whether it be a new fundamental type, a new
42 intrinsic function, or a whole new instruction.</p>
44 <p>When you come to this realization, stop and think. Do you really need to
45 extend LLVM? Is it a new fundamental capability that LLVM does not support at
46 its current incarnation or can it be synthesized from already pre-existing LLVM
47 elements? If you are not sure, ask on the <a
48 href="http://mail.cs.uiuc.edu/mailman/listinfo/llvmdev">LLVM-dev</a> list. The
49 reason is that extending LLVM will get involved as you need to update all the
50 different passes that you intend to use with your extension, and there are
51 <em>many</em> LLVM analyses and transformations, so it may be quite a bit of
54 <p>Adding an <a href="#intrinsic">intrinsic function</a> is easier than adding
55 an instruction, and is transparent to optimization passes which treat it as an
56 unanalyzable function. If your added functionality can be expressed as a
57 function call, an intrinsic function is the method of choice for LLVM
60 <p>Before you invest a significant amount of effort into a non-trivial
61 extension, <span class="doc_warning">ask on the list</span> if what you are
62 looking to do can be done with already-existing infrastructure, or if maybe
63 someone else is already working on it. You will save yourself a lot of time and
64 effort by doing so.</p>
68 <!-- *********************************************************************** -->
69 <div class="doc_section">
70 <a name="intrinsic">Adding a new intrinsic function</a>
72 <!-- *********************************************************************** -->
74 <div class="doc_text">
76 <p>Adding a new intrinsic function to LLVM is much easier than adding a new
77 instruction. Almost all extensions to LLVM should start as an intrinsic
78 function and then be turned into an instruction if warranted.</p>
81 <li><tt>llvm/docs/LangRef.html</tt>:
82 Document the intrinsic. Decide whether it is code generator specific and
83 what the restrictions are. Talk to other people about it so that you are
84 sure it's a good idea.</li>
86 <li><tt>llvm/include/llvm/Intrinsics.h</tt>:
87 add an enum in the <tt>llvm::Intrinsic</tt> namespace</li>
89 <li><tt>llvm/lib/VMCore/Verifier.cpp</tt>:
90 Add code to check the invariants of the intrinsic are respected.</li>
92 <li><tt>llvm/lib/VMCore/Function.cpp (<tt>Function::getIntrinsicID()</tt>)</tt>:
93 Identify the new intrinsic function, returning the enum for the intrinsic
96 <li><tt>llvm/lib/Analysis/BasicAliasAnalysis.cpp</tt>: If the new intrinsic does
97 not access memory or does not write to memory, add it to the relevant list
100 <li><tt>llvm/lib/Analysis/ConstantFolding.cpp</tt>: If it is possible to
101 constant fold your intrinsic, add support to it in the
102 <tt>canConstantFoldCallTo</tt> and <tt>ConstantFoldCall</tt> functions.</li>
104 <li><tt>llvm/lib/Transforms/Utils/Local.cpp</tt>: If your intrinsic has no side-
105 effects, add it to the list of intrinsics in the
106 <tt>isInstructionTriviallyDead</tt> function.</li>
108 <li>Test your intrinsic</li>
110 <li><tt>llvm/test/Regression/*</tt>: add your test cases to the test suite</li>
113 <p>Once the intrinsic has been added to the system, you must add code generator
114 support for it. Generally you must do the following steps:</p>
117 <dt>Add support to the C backend in <tt>lib/Target/CBackend/</tt></dt>
119 <dd>Depending on the intrinsic, there are a few ways to implement this. First,
120 if it makes sense to lower the intrinsic to an expanded sequence of C code in
121 all cases, just emit the expansion in <tt>visitCallInst</tt>. Second, if the
122 intrinsic has some way to express it with GCC (or any other compiler)
123 extensions, it can be conditionally supported based on the compiler compiling
124 the CBE output (see llvm.prefetch for an example). Third, if the intrinsic
125 really has no way to be lowered, just have the code generator emit code that
126 prints an error message and calls abort if executed.
129 <dt>Add a enum value for the SelectionDAG node in
130 <tt>include/llvm/CodeGen/SelectionDAGNodes.h</tt></dt>
132 <dd>Also, add code to <tt>lib/CodeGen/SelectionDAG/SelectionDAG.cpp</tt> (and
133 <tt>SelectionDAGPrinter.cpp</tt>) to print the node.</dd>
135 <dt>Add code to <tt>SelectionDAG/SelectionDAGISel.cpp</tt> to recognize the
138 <dd>Presumably the intrinsic should be recognized and turned into the node you
141 <dt>Add code to <tt>SelectionDAG/LegalizeDAG.cpp</tt> to <a
142 href="CodeGenerator.html#selectiondag_legalize">legalize, promote, and
143 expand</a> the node as necessary.</dt>
145 <dd>If the intrinsic can be expanded to primitive operations, legalize can break
146 the node down into other elementary operations that are be supported.</dd>
148 <dt>Add target-specific support to specific code generators.</dt>
150 <dd>Extend the code generators you are interested in to recognize and support
151 the node, emitting the code you want.</dd>
155 Unfortunately, the process of extending the code generator to support a new node
156 is not extremely well documented. As such, it is often helpful to look at other
157 intrinsics (e.g. <tt>llvm.ctpop</tt>) to see how they are recognized and turned
158 into a node by <tt>SelectionDAGISel.cpp</tt>, legalized by
159 <tt>LegalizeDAG.cpp</tt>, then finally emitted by the various code generators.
164 <!-- *********************************************************************** -->
165 <div class="doc_section">
166 <a name="instruction">Adding a new instruction</a>
168 <!-- *********************************************************************** -->
170 <div class="doc_text">
172 <p><span class="doc_warning">WARNING: adding instructions changes the bytecode
173 format, and it will take some effort to maintain compatibility with
174 the previous version.</span> Only add an instruction if it is absolutely
179 <li><tt>llvm/include/llvm/Instruction.def</tt>:
180 add a number for your instruction and an enum name</li>
182 <li><tt>llvm/include/llvm/Instructions.h</tt>:
183 add a definition for the class that will represent your instruction</li>
185 <li><tt>llvm/include/llvm/Support/InstVisitor.h</tt>:
186 add a prototype for a visitor to your new instruction type</li>
188 <li><tt>llvm/lib/AsmParser/Lexer.l</tt>:
189 add a new token to parse your instruction from assembly text file</li>
191 <li><tt>llvm/lib/AsmParser/llvmAsmParser.y</tt>:
192 add the grammar on how your instruction can be read and what it will
193 construct as a result</li>
195 <li><tt>llvm/lib/Bytecode/Reader/Reader.cpp</tt>:
196 add a case for your instruction and how it will be parsed from bytecode</li>
198 <li><tt>llvm/lib/VMCore/Instruction.cpp</tt>:
199 add a case for how your instruction will be printed out to assembly</li>
201 <li><tt>llvm/lib/VMCore/Instructions.cpp</tt>:
202 implement the class you defined in
203 <tt>llvm/include/llvm/Instructions.h</tt></li>
205 <li>Test your instruction</li>
207 <li><tt>llvm/lib/Target/*</tt>:
208 Add support for your instruction to code generators, or add a lowering
211 <li><tt>llvm/test/Regression/*</tt>: add your test cases to the test suite.</li>
215 <p>Also, you need to implement (or modify) any analyses or passes that you want
216 to understand this new instruction.</p>
221 <!-- *********************************************************************** -->
222 <div class="doc_section">
223 <a name="type">Adding a new type</a>
225 <!-- *********************************************************************** -->
227 <div class="doc_text">
229 <p><span class="doc_warning">WARNING: adding new types changes the bytecode
230 format, and will break compatibility with currently-existing LLVM
231 installations.</span> Only add new types if it is absolutely necessary.</p>
235 <!-- ======================================================================= -->
236 <div class="doc_subsection">
237 <a name="fund_type">Adding a fundamental type</a>
240 <div class="doc_text">
244 <li><tt>llvm/include/llvm/Type.h</tt>:
245 add enum for the new type; add static <tt>Type*</tt> for this type</li>
247 <li><tt>llvm/lib/VMCore/Type.cpp</tt>:
248 add mapping from <tt>TypeID</tt> => <tt>Type*</tt>;
249 initialize the static <tt>Type*</tt></li>
251 <li><tt>llvm/lib/AsmReader/Lexer.l</tt>:
252 add ability to parse in the type from text assembly</li>
254 <li><tt>llvm/lib/AsmReader/llvmAsmParser.y</tt>:
255 add a token for that type</li>
261 <!-- ======================================================================= -->
262 <div class="doc_subsection">
263 <a name="derived_type">Adding a derived type</a>
266 <div class="doc_text">
269 <li><tt>llvm/include/llvm/Type.h</tt>:
270 add enum for the new type; add a forward declaration of the type
273 <li><tt>llvm/include/llvm/DerivedTypes.h</tt>:
274 add new class to represent new class in the hierarchy; add forward
275 declaration to the TypeMap value type</li>
277 <li><tt>llvm/lib/VMCore/Type.cpp</tt>:
278 add support for derived type to:
279 <div class="doc_code">
281 std::string getTypeDescription(const Type &Ty,
282 std::vector<const Type*> &TypeStack)
283 bool TypesEqual(const Type *Ty, const Type *Ty2,
284 std::map<const Type*, const Type*> & EqTypes)
287 add necessary member functions for type, and factory methods</li>
289 <li><tt>llvm/lib/AsmReader/Lexer.l</tt>:
290 add ability to parse in the type from text assembly</li>
292 <li><tt>llvm/lib/ByteCode/Writer/Writer.cpp</tt>:
293 modify <tt>void BytecodeWriter::outputType(const Type *T)</tt> to serialize
296 <li><tt>llvm/lib/ByteCode/Reader/Reader.cpp</tt>:
297 modify <tt>const Type *BytecodeReader::ParseType()</tt> to read your data
300 <li><tt>llvm/lib/VMCore/AsmWriter.cpp</tt>:
302 <div class="doc_code">
304 void calcTypeName(const Type *Ty,
305 std::vector<const Type*> &TypeStack,
306 std::map<const Type*,std::string> &TypeNames,
307 std::string & Result)
310 to output the new derived type
318 <!-- *********************************************************************** -->
322 <a href="http://jigsaw.w3.org/css-validator/check/referer"><img
323 src="http://jigsaw.w3.org/css-validator/images/vcss" alt="Valid CSS!"></a>
324 <a href="http://validator.w3.org/check/referer"><img
325 src="http://www.w3.org/Icons/valid-html401" alt="Valid HTML 4.01!" /></a>
327 <a href="http://llvm.cs.uiuc.edu">The LLVM Compiler Infrastructure</a>
329 Last modified: $Date$