Initial checkin of the "extract" utility

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@2716 91177308-0d34-0410-b5e6-96231b3b80d8

3 files changed, 214 insertions, 0 deletions

diff --git a/tools/extract/Makefile b/tools/extract/Makefile
new file mode 100644
index 0000000..f9209d1
--- /dev/null
+++ b/tools/extract/Makefile
@@ -0,0 +1,6 @@
+LEVEL = ../..
+
+TOOLNAME = extract
+USEDLIBS = bcreader bcwriter transforms ipo analysis transformutils ipa vmcore support
+
+include $(LEVEL)/Makefile.common
diff --git a/tools/extract/extract.cpp b/tools/extract/extract.cpp
new file mode 100644
index 0000000..7a6a1c1
--- /dev/null
+++ b/tools/extract/extract.cpp
@@ -0,0 +1,104 @@
+//===----------------------------------------------------------------------===//
+// LLVM extract Utility
+//
+// This utility changes the input module to only contain a single function,
+// which is primarily used for debugging transformations.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Module.h"
+#include "llvm/PassManager.h"
+#include "llvm/Bytecode/Reader.h"
+#include "llvm/Bytecode/WriteBytecodePass.h"
+#include "llvm/GlobalVariable.h"
+#include "llvm/Function.h"
+#include "llvm/Transforms/IPO/GlobalDCE.h"
+#include "llvm/Transforms/ConstantMerge.h"
+#include "llvm/Transforms/CleanupGCCOutput.h"
+#include "Support/CommandLine.h"
+#include <memory>
+
+static cl::String InputFilename("", "Specify input bytecode file", 0, "-");
+static cl::String ExtractFunc("func", "Specify function to extract", 0, "main");
+
+struct FunctionExtractorPass : public Pass {
+  const char *getPassName() const { return "Function Extractor"; }
+
+  bool run(Module *M) {
+    // Mark all global variables to be internal
+    for (Module::giterator I = M->gbegin(), E = M->gend(); I != E; ++I)
+      (*I)->setInternalLinkage(true);
+
+    Function *Named = 0;
+
+    // Loop over all of the functions in the module, dropping all references in
+    // functions that are not the named function.
+    for (Module::iterator I = M->begin(), E = M->end(); I != E;)
+      // Check to see if this is the named function!
+      if (!Named && (*I)->getName() == ExtractFunc) {
+        // Yes, it is.  Keep track of it...
+        Named = *I;
+
+        // Remove the named function from the module.
+        M->getFunctionList().remove(I);
+        E = M->end();
+      } else {
+        // Nope it's not the named function, delete the body of the function
+        (*I)->dropAllReferences();
+        ++I;
+      }
+
+    // All of the functions that still have uses now must be used by global
+    // variables or the named function.  Loop through them and create a new,
+    // external function for the used ones... making all uses point to the new
+    // functions.
+    std::vector<Function*> NewFunctions;
+    
+    for (Module::iterator I = M->begin(), E = M->end(); I != E; ++I)
+      if (!(*I)->use_empty()) {
+        Function *New = new Function((*I)->getFunctionType(), false,
+                                     (*I)->getName());
+        (*I)->replaceAllUsesWith(New);
+        NewFunctions.push_back(New);
+      }
+    
+    // Now the module only has unused functions with their references dropped.
+    // Delete them all now!
+    M->getFunctionList().delete_all();
+
+    // Re-insert the named function...
+    if (Named)
+      M->getFunctionList().push_back(Named);
+    else
+      std::cerr << "Warning: Function '" << ExtractFunc << "' not found!\n";
+    
+    // Insert all of the function stubs...
+    M->getFunctionList().insert(M->end(), NewFunctions.begin(),
+                                NewFunctions.end());
+    return true;
+  }
+};
+
+
+int main(int argc, char **argv) {
+  cl::ParseCommandLineOptions(argc, argv, " llvm extractor\n");
+
+  std::auto_ptr<Module> M(ParseBytecodeFile(InputFilename));
+  if (M.get() == 0) {
+    std::cerr << "bytecode didn't read correctly.\n";
+    return 1;
+  }
+
+  // In addition to just parsing the input from GCC, we also want to spiff it up
+  // a little bit.  Do this now.
+  //
+  PassManager Passes;
+  Passes.add(new FunctionExtractorPass());
+  Passes.add(createGlobalDCEPass());              // Delete unreachable globals
+  Passes.add(createConstantMergePass());          // Merge dup global constants
+  Passes.add(createCleanupGCCOutputPass());       // Fix gccisms
+  Passes.add(new WriteBytecodePass(&std::cout));  // Write bytecode to file...
+
+  Passes.run(M.get());
+  return 0;
+}
diff --git a/tools/llvm-extract/llvm-extract.cpp b/tools/llvm-extract/llvm-extract.cpp
new file mode 100644
index 0000000..7a6a1c1
--- /dev/null
+++ b/tools/llvm-extract/llvm-extract.cpp
@@ -0,0 +1,104 @@
+//===----------------------------------------------------------------------===//
+// LLVM extract Utility
+//
+// This utility changes the input module to only contain a single function,
+// which is primarily used for debugging transformations.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Module.h"
+#include "llvm/PassManager.h"
+#include "llvm/Bytecode/Reader.h"
+#include "llvm/Bytecode/WriteBytecodePass.h"
+#include "llvm/GlobalVariable.h"
+#include "llvm/Function.h"
+#include "llvm/Transforms/IPO/GlobalDCE.h"
+#include "llvm/Transforms/ConstantMerge.h"
+#include "llvm/Transforms/CleanupGCCOutput.h"
+#include "Support/CommandLine.h"
+#include <memory>
+
+static cl::String InputFilename("", "Specify input bytecode file", 0, "-");
+static cl::String ExtractFunc("func", "Specify function to extract", 0, "main");
+
+struct FunctionExtractorPass : public Pass {
+  const char *getPassName() const { return "Function Extractor"; }
+
+  bool run(Module *M) {
+    // Mark all global variables to be internal
+    for (Module::giterator I = M->gbegin(), E = M->gend(); I != E; ++I)
+      (*I)->setInternalLinkage(true);
+
+    Function *Named = 0;
+
+    // Loop over all of the functions in the module, dropping all references in
+    // functions that are not the named function.
+    for (Module::iterator I = M->begin(), E = M->end(); I != E;)
+      // Check to see if this is the named function!
+      if (!Named && (*I)->getName() == ExtractFunc) {
+        // Yes, it is.  Keep track of it...
+        Named = *I;
+
+        // Remove the named function from the module.
+        M->getFunctionList().remove(I);
+        E = M->end();
+      } else {
+        // Nope it's not the named function, delete the body of the function
+        (*I)->dropAllReferences();
+        ++I;
+      }
+
+    // All of the functions that still have uses now must be used by global
+    // variables or the named function.  Loop through them and create a new,
+    // external function for the used ones... making all uses point to the new
+    // functions.
+    std::vector<Function*> NewFunctions;
+    
+    for (Module::iterator I = M->begin(), E = M->end(); I != E; ++I)
+      if (!(*I)->use_empty()) {
+        Function *New = new Function((*I)->getFunctionType(), false,
+                                     (*I)->getName());
+        (*I)->replaceAllUsesWith(New);
+        NewFunctions.push_back(New);
+      }
+    
+    // Now the module only has unused functions with their references dropped.
+    // Delete them all now!
+    M->getFunctionList().delete_all();
+
+    // Re-insert the named function...
+    if (Named)
+      M->getFunctionList().push_back(Named);
+    else
+      std::cerr << "Warning: Function '" << ExtractFunc << "' not found!\n";
+    
+    // Insert all of the function stubs...
+    M->getFunctionList().insert(M->end(), NewFunctions.begin(),
+                                NewFunctions.end());
+    return true;
+  }
+};
+
+
+int main(int argc, char **argv) {
+  cl::ParseCommandLineOptions(argc, argv, " llvm extractor\n");
+
+  std::auto_ptr<Module> M(ParseBytecodeFile(InputFilename));
+  if (M.get() == 0) {
+    std::cerr << "bytecode didn't read correctly.\n";
+    return 1;
+  }
+
+  // In addition to just parsing the input from GCC, we also want to spiff it up
+  // a little bit.  Do this now.
+  //
+  PassManager Passes;
+  Passes.add(new FunctionExtractorPass());
+  Passes.add(createGlobalDCEPass());              // Delete unreachable globals
+  Passes.add(createConstantMergePass());          // Merge dup global constants
+  Passes.add(createCleanupGCCOutputPass());       // Fix gccisms
+  Passes.add(new WriteBytecodePass(&std::cout));  // Write bytecode to file...
+
+  Passes.run(M.get());
+  return 0;
+}