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 @@ -134,6 +142,7 @@ the ModulePass, CallGraphSCCPass, FunctionPass, or LoopPass, or RegionPass, or 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 @@ -147,12 +156,12 @@ more advanced features are discussed.
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 @@ -160,21 +169,21 @@ 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.
-First, you need to create a new directory somewhere in the LLVM source +
First, configure and build LLVM. This needs to be done directly inside the + LLVM source tree rather than in a separate objects directory. + 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. Next, you must set up a build script + lib/Transforms/Hello. Finally, you must set up a build script (Makefile) that will compile the source code for the new pass. To do this, copy the following into Makefile:
-# Makefile for hello pass @@ -189,33 +198,31 @@ LIBRARYNAME = Hello # dlopen/dlsym on the resulting library. LOADABLE_MODULE = 1 -# Tell the build system which LLVM libraries your pass needs. You'll probably -# need at least LLVMSystem.a, LLVMSupport.a, LLVMCore.a but possibly several -# others too. -LLVMLIBS = LLVMCore.a LLVMSupport.a LLVMSystem.a - # Include the makefile implementation stuff include $(LEVEL)/Makefile.common
This makefile specifies that all of the .cpp files in the current -directory are to be compiled and linked together into a -Debug/lib/Hello.so shared object that can be dynamically loaded by +directory are to be compiled and linked together into a shared object +$(LEVEL)/Debug+Asserts/lib/Hello.so that can be dynamically loaded by the opt or bugpoint tools via their -load options. If your operating system uses a suffix other than .so (such as windows or Mac OS/X), the appropriate extension will be used.
+If you are used CMake to build LLVM, see +Developing an LLVM pass with CMake.
+Now that we have the build scripts set up, we just need to write the code for the pass itself.
Now that we have a way to compile our new pass, we just have to write it. Start out with:
@@ -267,7 +274,7 @@ time.
static char ID;
- Hello() : FunctionPass(&ID) {}
+ Hello() : FunctionPass(ID) {}
This declares pass identifier used by LLVM to identify pass. This allows LLVM to @@ -295,7 +302,7 @@ function.
initialization value is not important.
- RegisterPass<Hello> X("hello", "Hello World Pass",
+ static RegisterPass<Hello> X("hello", "Hello World Pass",
false /* Only looks at CFG */,
false /* Analysis Pass */);
} // end of anonymous namespace
@@ -304,7 +311,7 @@ initialization value is not important.
Lastly, we register our class Hello,
giving it a command line
argument "hello", and a name "Hello World Pass".
-Last two RegisterPass arguments are optional. Their default value is false.
+Last two arguments describe its behavior.
If a pass walks CFG without modifying it then third argument is set to true.
If a pass is an analysis pass, for example dominator tree pass, then true
is supplied as fourth argument.
@@ -322,7 +329,7 @@ is supplied as fourth argument.
struct Hello : public FunctionPass {
static char ID;
- Hello() : FunctionPass(&ID) {}
+ Hello() : FunctionPass(ID) {}
virtual bool runOnFunction(Function &F) {
errs() << "Hello: " << F.getName() << "\n";
@@ -331,13 +338,15 @@ is supplied as fourth argument.
};
char Hello::ID = 0;
- RegisterPass<Hello> X("hello", "Hello World Pass");
+ static RegisterPass<Hello> X("hello", "Hello World Pass", false, false);
}
+
Now that it's all together, compile the file with a simple "gmake" -command in the local directory and you should get a new -"Debug/lib/Hello.so file. Note that everything in this file is +command in the local directory and you should get a new file +"Debug+Asserts/lib/Hello.so" under the top level directory of the LLVM +source tree (not in the local directory). 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.
@@ -345,15 +354,15 @@ them) to be useful.Now that you have a brand new shiny shared object file, we can use the opt command to run an LLVM program through your pass. Because you -registered your pass with the RegisterPass template, you will be able to +registered your pass with RegisterPass, you will be able to use the opt tool to access it, once loaded.
To test it, follow the example at the end of the
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
- The most plain and boring type of pass is the "ImmutablePass"
@@ -482,11 +491,11 @@ invalidated, and are never "run". The "ModulePass"
@@ -508,14 +517,12 @@ DominatorTree for function definitions, not declarations. The "CallGraphSCCPass"
@@ -551,11 +560,9 @@ href="#BasicBlockPass">BasicBlockPass, you should derive from
The runOnSCC method performs the interesting work of the pass, and
@@ -617,12 +623,13 @@ otherwise. In contrast to ModulePass subclasses, FunctionPass
@@ -665,15 +674,14 @@ href="#basiccode">Hello World pass for example). FunctionPass's
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. All LoopPass execute on each loop in the function independent of
all of the other loops in the function. LoopPass processes loops in
@@ -745,19 +756,18 @@ 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's may overload three virtual methods to
+straightforward. LoopPass's 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. RegionPass is similar to 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 didn not.
+ 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.
+
+ 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 is an infrequently used method that is
+called when the pass framework has finished calling runOnRegion for every region in the
+program being compiled. BasicBlockPass's are just like FunctionPass's, except that they must limit
@@ -838,15 +927,14 @@ href="#doInitialization_mod">doInitialization(Module &) and doFinalization(Module &) methods that FunctionPass's have, but also have the following virtual methods that may also be implemented: A MachineFunctionPass is a part of the LLVM code generator that
executes on the machine-dependent representation of each LLVM function in the
-program. A MachineFunctionPass is also a FunctionPass, so all
+program. Code generator passes are registered and initialized specially by
+TargetMachine::addPassesToEmitFile and similar routines, so they
+cannot generally be run from the opt or bugpoint
+commands. A MachineFunctionPass is also a FunctionPass, so all
the restrictions that apply to a FunctionPass also apply to it.
MachineFunctionPasses also have additional restrictions. In particular,
MachineFunctionPasses are not allowed to do any of the following: In the Hello World 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, which requires you to pass at least two
-parameters. The first parameter is the name of the pass that is to be used on
+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 opt or bugpoint). The second argument is the
+example, with opt or bugpoint). The first argument is the
name of the pass, which is to be used for the -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: One of the main responsibilities of the PassManager is to make sure
that passes interact with each other correctly. Because PassManager
@@ -1027,14 +1126,12 @@ specifies. If a pass does not implement the getAnalysisUsage method, it defaults to not
having any prerequisite passes, and invalidating all other passes.
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.
@@ -1076,11 +1176,13 @@ pass is.
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
@@ -1111,22 +1213,13 @@ the fact that it hacks on the CFG.
and: The Pass::getAnalysis<> method is automatically inherited by
your class, providing you with access to the passes that you declared that you
@@ -1196,13 +1291,15 @@ if it is active. For example: 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
@@ -1221,14 +1318,12 @@ between these two extremes for other implementations). To cleanly support
situations like this, the LLVM Pass Infrastructure supports the notion of
Analysis Groups. An Analysis Group is a single simple interface that may be implemented by
multiple different passes. Analysis Groups can be given human readable names
@@ -1246,7 +1341,7 @@ between passes still apply. Although Pass Registration is optional for normal
passes, all analysis group implementations must be registered, and must use the
-RegisterAnalysisGroup template to join the
+INITIALIZE_AG_PASS template to join the
implementation pool. Also, a default implementation of the interface
must be registered with RegisterAnalysisGroup. The RegisterAnalysisGroup template is used to register the analysis
-group itself as well as add pass implementations to the analysis group. First,
-an analysis should be registered, with a human readable name provided for it.
+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": This just shows a class FancyAA that is registered normally, then
-uses the RegisterAnalysisGroup template to "join" the AliasAnalysis
+ This just shows a class FancyAA that
+uses the INITIALIZE_AG_PASS macro both to register and
+to "join" the AliasAnalysis
analysis group. Every implementation of an analysis group should join using
-this template. A single pass may join multiple different analysis groups with
-no problem. Here we show how the default implementation is specified (using the extra
-argument to the RegisterAnalysisGroup template). There must be exactly
+ 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
@@ -1333,13 +1431,15 @@ pass is the default implementation for the interface. The Statistic
class is designed to be an easy way to expose various success
@@ -1351,12 +1451,12 @@ line. See the St
-
+
- The PassManager
@@ -1382,7 +1482,7 @@ results as soon as they are no longer needed.
Now when we run our pass, we get this output: Which shows that we don't accidentally invalidate dominator information
anymore, and therefore do not have to compute it twice. Size matters when constructing production quality tools using llvm,
both for the purposes of distribution, and for regulating the resident code size
@@ -1584,14 +1684,12 @@ 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. There are predefined registries to track instruction scheduling
(RegisterScheduler) and register allocation (RegisterRegAlloc)
@@ -1630,7 +1728,7 @@ form; 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
@@ -1684,13 +1782,15 @@ creator. 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
@@ -1702,14 +1802,12 @@ GDB. First thing you do is start gdb on the opt process: Once you have the basics down, there are a couple of problems that GDB has,
some with solutions, some without. 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: Multiple CPU machines are becoming more common and compilation can never be
fast enough: obviously we should allow for a multithreaded compiler. Because of
@@ -1819,6 +1917,8 @@ Despite that, we have kept the LLVM passes SMP ready, and you should too.
-$ opt -load ../../../Debug/lib/Hello.so -hello < hello.bc > /dev/null
+$ opt -load ../../../Debug+Asserts/lib/Hello.so -hello < hello.bc > /dev/null
Hello: __main
Hello: puts
Hello: main
@@ -380,7 +389,7 @@ interesting way, we just throw away the result of opt (sending it to
opt with the -help option:
-$ opt -load ../../../Debug/lib/Hello.so -help
+$ opt -load ../../../Debug+Asserts/lib/Hello.so -help
OVERVIEW: llvm .bc -> .bc modular optimizer
USAGE: opt [options] <input bitcode>
@@ -408,7 +417,7 @@ the execution time of your pass along with the other passes you queue up. For
example:
-$ opt -load ../../../Debug/lib/Hello.so -hello -time-passes < hello.bc > /dev/null
+$ opt -load ../../../Debug+Asserts/lib/Hello.so -hello -time-passes < hello.bc > /dev/null
Hello: __main
Hello: puts
Hello: main
@@ -435,13 +444,15 @@ about some more details of how they work and how to use them.
virtual bool runOnModule(Module &M) = 0;
@@ -527,12 +534,14 @@ false otherwise.
-
+
+ The doInitialization(CallGraph &) method
+
+
-
virtual bool doInitialization(CallGraph &CG);
@@ -600,14 +606,14 @@ fast).
- virtual bool runOnSCC(const std::vector<CallGraphNode *> &SCCM) = 0;
+ virtual bool runOnSCC(CallGraphSCC &SCC) = 0;
+
+ The doFinalization(CallGraph &) method
+
+
-
virtual bool doFinalization(CallGraph &CG);
@@ -635,12 +642,14 @@ program being compiled.
+
+ The doInitialization(Module &) method
+
+
-
virtual bool doInitialization(Module &M);
@@ -697,11 +705,11 @@ free functions that it needs, adding prototypes to the module if necessary.
virtual bool runOnFunction(Function &F) = 0;
@@ -714,12 +722,13 @@ be returned if the function is modified.
+
+ The doFinalization(Module &) method
+
+
-
virtual bool doFinalization(Module &M);
@@ -732,12 +741,14 @@ program being compiled.
+
+ The doInitialization(Loop *,LPPassManager &) method
+
+
-
virtual bool doInitialization(Loop *, LPPassManager &LPM);
@@ -774,11 +784,11 @@ information.
-
+
-
virtual bool runOnLoop(Loop *, LPPassManager &LPM) = 0;
@@ -792,11 +802,11 @@ should be used to update loop nest.
virtual bool doFinalization();
@@ -809,14 +819,93 @@ program being compiled.
+ The RegionPass class
+
+
+
+
+ The doInitialization(Region *, RGPassManager &) method
+
+
+
+
+ virtual bool doInitialization(Region *, RGPassManager &RGM);
+
+ The runOnRegion method
+
+
+
+ virtual bool runOnRegion(Region *, RGPassManager &RGM) = 0;
+
+ The doFinalization() method
+
+
+
+ virtual bool doFinalization();
+
+
+ The doInitialization(Function &) method
+
+
-
virtual bool doInitialization(Function &F);
@@ -863,11 +951,11 @@ fast).
virtual bool runOnBasicBlock(BasicBlock &BB) = 0;
@@ -881,12 +969,13 @@ if the basic block is modified.
+
+ The doFinalization(Function &) method
+
+
-
virtual bool doFinalization(Function &F);
@@ -900,39 +989,46 @@ finalization.
-
-
+
+ The runOnMachineFunction(MachineFunction &MF) method
+
+
-
virtual bool runOnMachineFunction(MachineFunction &MF) = 0;
@@ -953,23 +1049,26 @@ remember, you may not modify the LLVM Function or its contents from a
The print method
-
virtual void print(std::ostream &O, const Module *M) const;
@@ -1004,13 +1101,15 @@ depended on.
virtual void getAnalysisUsage(AnalysisUsage &Info) const;
@@ -1050,11 +1147,14 @@ object:
+
+ The AnalysisUsage::addRequired<>
+ and AnalysisUsage::addRequiredTransitive<> methods
+
+
+
+
+
+ The AnalysisUsage::addPreserved<> method
+
+
-
- // This is an example implementation from an analysis, which does not modify
- // the program at all, yet has a prerequisite.
- void PostDominanceFrontier::getAnalysisUsage(AnalysisUsage &AU) const {
- AU.setPreservesAll();
- AU.addRequired<PostDominatorTree>();
- }
-
+
+ Example implementations of getAnalysisUsage
+
+
-
// This example modifies the program, but does not modify the CFG
@@ -1139,12 +1232,14 @@ the fact that it hacks on the CFG.
+
+ The getAnalysis<> and
+ getAnalysisIfAvailable<> methods
+
+
-
namespace {
- // Analysis Group implementations must be registered normally...
- RegisterPass<FancyAA>
- B("somefancyaa", "A more complex alias analysis implementation");
-
// Declare that we implement the AliasAnalysis interface
- RegisterAnalysisGroup<AliasAnalysis> C(B);
+ 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?
+ );
}
namespace {
- // Analysis Group implementations must be registered normally...
- RegisterPass<BasicAliasAnalysis>
- D("basicaa", "Basic Alias Analysis (default AA impl)");
-
// Declare that we implement the AliasAnalysis interface
- RegisterAnalysisGroup<AliasAnalysis, true> E(D);
+ 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?
+ );
}
Pass Statistics
-
-$ opt -load ../../../Debug/lib/Hello.so -gcse -licm --debug-pass=Structure < hello.bc > /dev/null
+$ opt -load ../../../Debug+Asserts/lib/Hello.so -gcse -licm --debug-pass=Structure < hello.bc > /dev/null
Module Pass Manager
Function Pass Manager
Dominator Set Construction
@@ -1454,7 +1554,7 @@ passes.
World pass in between the two passes:
-$ opt -load ../../../Debug/lib/Hello.so -gcse -hello -licm --debug-pass=Structure < hello.bc > /dev/null
+$ opt -load ../../../Debug+Asserts/lib/Hello.so -gcse -hello -licm --debug-pass=Structure < hello.bc > /dev/null
Module Pass Manager
Function Pass Manager
Dominator Set Construction
@@ -1495,7 +1595,7 @@ href="#getAnalysisUsage">getAnalysisUsage method to our pass:
-$ opt -load ../../../Debug/lib/Hello.so -gcse -hello -licm --debug-pass=Structure < hello.bc > /dev/null
+$ opt -load ../../../Debug+Asserts/lib/Hello.so -gcse -hello -licm --debug-pass=Structure < hello.bc > /dev/null
Pass Arguments: -gcse -hello -licm
Module Pass Manager
Function Pass Manager
@@ -1523,14 +1623,12 @@ Hello: main
virtual void releaseMemory();
@@ -1551,13 +1649,15 @@ class, before the next call of run* in your pass.
$ llc -help
...
- -regalloc - Register allocator to use: (default = linearscan)
+ -regalloc - Register allocator to use (default=linearscan)
=linearscan - linear scan register allocator
=local - local register allocator
=simple - simple register allocator
@@ -1652,11 +1750,11 @@ call line to llvm/Codegen/LinkAllCodegenComponents.h.
-
+
-
(gdb) break llvm::PassManager::run
Breakpoint 1 at 0x2413bc: file Pass.cpp, line 70.
-(gdb) run test.bc -load $(LLVMTOP)/llvm/Debug/lib/[libname].so -[passoption]
-Starting program: opt test.bc -load $(LLVMTOP)/llvm/Debug/lib/[libname].so -[passoption]
+(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)
@@ -1750,11 +1848,11 @@ or do other standard debugging stuff.
@@ -1828,7 +1928,7 @@ Despite that, we have kept the LLVM passes SMP ready, and you should too.
src="http://www.w3.org/Icons/valid-html401-blue" alt="Valid HTML 4.01">
Chris Lattner
- The LLVM Compiler Infrastructure
+ The LLVM Compiler Infrastructure
Last modified: $Date$