Stop pretending we support multiple LLVM versions

Author: nicolasvasilache

tc/core/polyhedral/llvm_jit.cc

@@ -36,34 +36,6 @@ using namespace llvm;
 
 namespace tc {
 
-#if LLVM_VERSION_MAJOR <= 6
-
-Jit::Jit()
-    : TM_(EngineBuilder().selectTarget()),
-      DL_(TM_->createDataLayout()),
-      objectLayer_([]() { return std::make_shared<SectionMemoryManager>(); }),
-      compileLayer_(objectLayer_, orc::SimpleCompiler(*TM_)) {}
-
-void Jit::addModule(std::shared_ptr<Module> M) {
-  M->setTargetTriple(TM_->getTargetTriple().str());
-  auto Resolver = orc::createLambdaResolver(
-      [&](const std::string& Name) {
-        if (auto Sym = compileLayer_.findSymbol(Name, false))
-          return Sym;
-        return JITSymbol(nullptr);
-      },
-      [](const std::string& Name) {
-        if (auto SymAddr = RTDyldMemoryManager::getSymbolAddressInProcess(Name))
-          return JITSymbol(SymAddr, JITSymbolFlags::Exported);
-        return JITSymbol(nullptr);
-      });
-
-  auto res = compileLayer_.addModule(M, std::move(Resolver));
-  TC_CHECK(res) << "Failed to jit compile.";
-}
-
-#else
-
 Jit::Jit()
     : ES(),
       Resolver(llvm::orc::createLegacyLookupResolver(
@@ -87,22 +59,14 @@ Jit::Jit()
             return llvm::orc::RTDyldObjectLinkingLayer::Resources{
                 std::make_shared<SectionMemoryManager>(), Resolver};
           }),
-      compileLayer_(objectLayer_, orc::SimpleCompiler(*TM_)) {
-  std::string err;
-
-  sys::DynamicLibrary::LoadLibraryPermanently(nullptr, &err);
-  if (err != "") {
-    throw std::runtime_error("Failed to find cilkrts: " + err);
-  }
-}
+      compileLayer_(objectLayer_, orc::SimpleCompiler(*TM_)) {}
 
 void Jit::addModule(std::shared_ptr<Module> M) {
   M->setTargetTriple(TM_->getTargetTriple().str());
   auto K = ES.allocateVModule();
   llvm::Error res = compileLayer_.addModule(K, CloneModule(*M));
   TC_CHECK(!res) << "Failed to jit compile.";
 }
-#endif
 
 std::shared_ptr<Module> Jit::codegenScop(
     const std::string& specializedName,