Extending LLVM: Adding instructions, intrinsics, types, etc.
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.
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 LLVM-dev 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
many LLVM analyses and transformations, so it may be quite a bit of
work.
Adding an intrinsic function 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.
Before you invest a significant amount of effort into a non-trivial
extension, ask on the list 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.
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.
- llvm/docs/LangRef.html:
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.
- llvm/include/llvm/Intrinsics.h:
add an enum in the llvm::Intrinsic namespace
- llvm/lib/VMCore/Verifier.cpp:
Add code to check the invariants of the intrinsic are respected.
- llvm/lib/VMCore/Function.cpp (Function::getIntrinsicID()):
Identify the new intrinsic function, returning the enum for the intrinsic
that you added.
- llvm/lib/Analysis/BasicAliasAnalysis.cpp: If the new intrinsic does
not access memory or does not write to memory, add it to the relevant list
of functions.
- llvm/lib/Transforms/Utils/Local.cpp: If it is possible to constant
fold your intrinsic, add support to it in the canConstantFoldCallTo and
ConstantFoldCall functions.
- Test your intrinsic
- llvm/test/Regression/*: add your test cases to the test suite
Once the intrinsic has been added to the system, you must add code generator
support for it. Generally you must do the following steps:
- Add support to the C backend in lib/Target/CBackend/
- 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 visitCallInst. 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.
- Add a enum value for the SelectionDAG node in
include/llvm/CodeGen/SelectionDAGNodes.h
- Also, add code to lib/CodeGen/SelectionDAG/SelectionDAG.cpp (and
SelectionDAGPrinter.cpp) to print the node.
- Add code to SelectionDAG/SelectionDAGISel.cpp to recognize the
intrinsic.
- Presumably the intrinsic should be recognized and turned into the node you
added above.
- Add code to SelectionDAG/LegalizeDAG.cpp to legalize, promote, and
expand the node as necessary.
- If the intrinsic can be expanded to primitive operations, legalize can break
the node down into other elementary operations that are be supported.
- Add target-specific support to specific code generators.
- Extend the code generators you are interested in to recognize and support
the node, emitting the code you want.
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. llvm.ctpop) to see how they are recognized and turned
into a node by SelectionDAGISel.cpp, legalized by
LegalizeDAG.cpp, then finally emitted by the various code generators.
WARNING: adding instructions changes the bytecode
format, and it will take some effort to maintain compatibility with
the previous version. Only add an instruction if it is absolutely
necessary.
- llvm/include/llvm/Instruction.def:
add a number for your instruction and an enum name
- llvm/include/llvm/Instructions.h:
add a definition for the class that will represent your instruction
- llvm/include/llvm/Support/InstVisitor.h:
add a prototype for a visitor to your new instruction type
- llvm/lib/AsmParser/Lexer.l:
add a new token to parse your instruction from assembly text file
- llvm/lib/AsmParser/llvmAsmParser.y:
add the grammar on how your instruction can be read and what it will
construct as a result
- llvm/lib/Bytecode/Reader/Reader.cpp:
add a case for your instruction and how it will be parsed from bytecode
- llvm/lib/VMCore/Instruction.cpp:
add a case for how your instruction will be printed out to assembly
- llvm/lib/VMCore/Instructions.cpp:
implement the class you defined in
llvm/include/llvm/Instructions.h
- Test your instruction
- llvm/lib/Target/*:
Add support for your instruction to code generators, or add a lowering
pass.
- llvm/test/Regression/*: add your test cases to the test suite.
Also, you need to implement (or modify) any analyses or passes that you want
to understand this new instruction.
WARNING: adding new types changes the bytecode
format, and will break compatibility with currently-existing LLVM
installations. Only add new types if it is absolutely necessary.