[flang][driver] Add support for -S and implement -c/-emit-obj

This patch adds support for:
  * `-S` in Flang's compiler and frontend drivers,
and implements:
  * `-emit-obj` in Flang's frontend driver and `-c` in Flang's compiler
    driver (this is consistent with Clang).
(these options were already available before, but only as placeholders).
The semantics of these options in Clang and Flang are identical.

The `EmitObjAction` frontend action is renamed as `BackendAction`. This
new name more accurately reflects the fact that this action will
primarily run the code-gen/backend pipeline in LLVM. It also makes more
sense as an action implementing both `-emit-obj` and `-S` (originally,
it was just `-emit-obj`).

`tripleName` from FirContext.cpp is deleted and, when a target triple is
required, `mlir::LLVM::LLVMDialect::getTargetTripleAttrName()` is used
instead. In practice, this means that `fir.triple` is replaced with
`llvm.target_triple`. The former was effectively ignored. The latter is
used when lowering from the LLVM dialect in MLIR to LLVM IR (i.e. it's
embedded in the generated LLVM IR module). The driver can then re-use
it when configuring the backend. With this change, the LLVM IR files
generated by e.g. `tco` will from now on contain the correct target
triple.

The code-gen.f90 test is replaced with code-gen-x86.f90 and
code-gen-aarch64.f90. With 2 seperate files we can verify that
`--target` is correctly taken into account. LIT configuration is updated
to enable e.g.:
```
! REQUIRES: aarch64-registered-target
```

Differential Revision: https://reviews.llvm.org/D120568

Author: banach-space

clang/include/clang/Driver/Options.td

@@ -750,7 +750,7 @@ def Rpass_analysis_EQ : Joined<["-"], "Rpass-analysis=">, Group<R_value_Group>,
            "name matches the given POSIX regular expression">;
 def R_Joined : Joined<["-"], "R">, Group<R_Group>, Flags<[CC1Option, CoreOption]>,
   MetaVarName<"<remark>">, HelpText<"Enable the specified remark">;
-def S : Flag<["-"], "S">, Flags<[NoXarchOption,CC1Option]>, Group<Action_Group>,
+def S : Flag<["-"], "S">, Flags<[NoXarchOption,CC1Option,FlangOption,FC1Option]>, Group<Action_Group>,
   HelpText<"Only run preprocess and compilation steps">;
 def Tbss : JoinedOrSeparate<["-"], "Tbss">, Group<T_Group>,
   MetaVarName<"<addr>">, HelpText<"Set starting address of BSS to <addr>">;

flang/include/flang/Frontend/FrontendActions.h

@@ -192,8 +192,19 @@ class EmitLLVMAction : public CodeGenAction {
   void ExecuteAction() override;
 };
 
-class EmitObjAction : public CodeGenAction {
+class BackendAction : public CodeGenAction {
+public:
+  enum class BackendActionTy {
+    Backend_EmitAssembly, ///< Emit native assembly files
+    Backend_EmitObj ///< Emit native object files
+  };
+
+  BackendAction(BackendActionTy act) : action{act} {};
+
+private:
   void ExecuteAction() override;
+
+  BackendActionTy action;
 };
 
 } // namespace Fortran::frontend

flang/include/flang/Frontend/FrontendOptions.h

@@ -40,6 +40,9 @@ enum ActionKind {
   /// Emit a .o file.
   EmitObj,
 
+  /// Emit a .s file.
+  EmitAssembly,
+
   /// Parse, unparse the parse-tree and output a Fortran source file
   DebugUnparse,
 

flang/lib/Frontend/CMakeLists.txt

@@ -27,6 +27,8 @@ add_flang_library(flangFrontend
   FortranLower
   clangBasic
   clangDriver
+  LLVMAnalysis
+  LLVMTarget
   FIRDialect
   FIRSupport
   FIRBuilder

flang/lib/Frontend/CompilerInvocation.cpp

@@ -153,6 +153,9 @@ static bool ParseFrontendArgs(FrontendOptions &opts, llvm::opt::ArgList &args,
     case clang::driver::options::OPT_emit_obj:
       opts.programAction = EmitObj;
       break;
+    case clang::driver::options::OPT_S:
+      opts.programAction = EmitAssembly;
+      break;
     case clang::driver::options::OPT_fdebug_unparse:
       opts.programAction = DebugUnparse;
       break;

flang/lib/Frontend/FrontendActions.cpp

@@ -31,7 +31,13 @@
 #include "mlir/Pass/PassManager.h"
 #include "mlir/Target/LLVMIR/ModuleTranslation.h"
 #include "llvm/ADT/StringRef.h"
+#include "llvm/Analysis/TargetLibraryInfo.h"
+#include "llvm/Analysis/TargetTransformInfo.h"
+#include "llvm/IR/LegacyPassManager.h"
+#include "llvm/MC/TargetRegistry.h"
+#include "llvm/Passes/PassBuilder.h"
 #include "llvm/Support/ErrorHandling.h"
+#include "llvm/Target/TargetMachine.h"
 #include <clang/Basic/Diagnostic.h>
 #include <memory>
 
@@ -417,7 +423,6 @@ void CodeGenAction::GenerateLLVMIR() {
 
   pm.addPass(std::make_unique<Fortran::lower::VerifierPass>());
   pm.enableVerifier(/*verifyPasses=*/true);
-  mlir::PassPipelineCLParser passPipeline("", "Compiler passes to run");
 
   // Create the pass pipeline
   fir::createMLIRToLLVMPassPipeline(pm);
@@ -490,11 +495,90 @@ void EmitMLIRAction::ExecuteAction() {
   mlirModule->print(*os);
 }
 
-void EmitObjAction::ExecuteAction() {
+void BackendAction::ExecuteAction() {
   CompilerInstance &ci = this->instance();
-  unsigned DiagID = ci.diagnostics().getCustomDiagID(
-      clang::DiagnosticsEngine::Error, "code-generation is not available yet");
-  ci.diagnostics().Report(DiagID);
+  // Generate an LLVM module if it's not already present (it will already be
+  // present if the input file is an LLVM IR/BC file).
+  if (!llvmModule)
+    GenerateLLVMIR();
+
+  // Create `Target`
+  std::string error;
+  const std::string &theTriple = llvmModule->getTargetTriple();
+  const llvm::Target *theTarget =
+      llvm::TargetRegistry::lookupTarget(theTriple, error);
+  // TODO: Make this a diagnostic once `flang-new` can consume LLVM IR files
+  // (in which users could use unsupported triples)
+  assert(theTarget && "Failed to create Target");
+
+  // Create `TargetMachine`
+  std::unique_ptr<llvm::TargetMachine> TM;
+  TM.reset(theTarget->createTargetMachine(theTriple, /*CPU=*/"",
+      /*Features=*/"", llvm::TargetOptions(), llvm::None));
+  assert(TM && "Failed to create TargetMachine");
+  llvmModule->setDataLayout(TM->createDataLayout());
+
+  // If the output stream is a file, generate it and define the corresponding
+  // output stream. If a pre-defined output stream is available, we will use
+  // that instead.
+  //
+  // NOTE: `os` is a smart pointer that will be destroyed at the end of this
+  // method. However, it won't be written to until `CodeGenPasses` is
+  // destroyed. By defining `os` before `CodeGenPasses`, we make sure that the
+  // output stream won't be destroyed before it is written to. This only
+  // applies when an output file is used (i.e. there is no pre-defined output
+  // stream).
+  // TODO: Revisit once the new PM is ready (i.e. when `CodeGenPasses` is
+  // updated to use it).
+  std::unique_ptr<llvm::raw_pwrite_stream> os;
+  if (ci.IsOutputStreamNull()) {
+    // Get the output buffer/file
+    switch (action) {
+    case BackendActionTy::Backend_EmitAssembly:
+      os = ci.CreateDefaultOutputFile(
+          /*Binary=*/false, /*InFile=*/GetCurrentFileOrBufferName(), "s");
+      break;
+    case BackendActionTy::Backend_EmitObj:
+      os = ci.CreateDefaultOutputFile(
+          /*Binary=*/true, /*InFile=*/GetCurrentFileOrBufferName(), "o");
+      break;
+    }
+    if (!os) {
+      unsigned diagID = ci.diagnostics().getCustomDiagID(
+          clang::DiagnosticsEngine::Error, "failed to create the output file");
+      ci.diagnostics().Report(diagID);
+      return;
+    }
+  }
+
+  // Create an LLVM code-gen pass pipeline. Currently only the legacy pass
+  // manager is supported.
+  // TODO: Switch to the new PM once it's available in the backend.
+  llvm::legacy::PassManager CodeGenPasses;
+  CodeGenPasses.add(
+      createTargetTransformInfoWrapperPass(TM->getTargetIRAnalysis()));
+  llvm::Triple triple(theTriple);
+
+  std::unique_ptr<llvm::TargetLibraryInfoImpl> TLII =
+      std::make_unique<llvm::TargetLibraryInfoImpl>(triple);
+  assert(TLII && "Failed to create TargetLibraryInfo");
+  CodeGenPasses.add(new llvm::TargetLibraryInfoWrapperPass(*TLII));
+
+  llvm::CodeGenFileType cgft = (action == BackendActionTy::Backend_EmitAssembly)
+      ? llvm::CodeGenFileType::CGFT_AssemblyFile
+      : llvm::CodeGenFileType::CGFT_ObjectFile;
+  if (TM->addPassesToEmitFile(CodeGenPasses,
+          ci.IsOutputStreamNull() ? *os : ci.GetOutputStream(), nullptr,
+          cgft)) {
+    unsigned diagID =
+        ci.diagnostics().getCustomDiagID(clang::DiagnosticsEngine::Error,
+            "emission of this file type is not supported");
+    ci.diagnostics().Report(diagID);
+    return;
+  }
+
+  // Run the code-gen passes
+  CodeGenPasses.run(*llvmModule);
 }
 
 void InitOnlyAction::ExecuteAction() {

flang/lib/FrontendTool/ExecuteCompilerInvocation.cpp

@@ -38,7 +38,11 @@ static std::unique_ptr<FrontendAction> CreateFrontendBaseAction(
   case EmitLLVM:
     return std::make_unique<EmitLLVMAction>();
   case EmitObj:
-    return std::make_unique<EmitObjAction>();
+    return std::make_unique<BackendAction>(
+        BackendAction::BackendActionTy::Backend_EmitObj);
+  case EmitAssembly:
+    return std::make_unique<BackendAction>(
+        BackendAction::BackendActionTy::Backend_EmitAssembly);
   case DebugUnparse:
     return std::make_unique<DebugUnparseAction>();
   case DebugUnparseNoSema:

flang/lib/Optimizer/Support/FIRContext.cpp

@@ -12,19 +12,20 @@
 
 #include "flang/Optimizer/Support/FIRContext.h"
 #include "flang/Optimizer/Support/KindMapping.h"
+#include "mlir/Dialect/LLVMIR/LLVMDialect.h"
 #include "mlir/IR/BuiltinAttributes.h"
 #include "mlir/IR/BuiltinOps.h"
 #include "llvm/Support/Host.h"
 
-static constexpr const char *tripleName = "fir.triple";
-
 void fir::setTargetTriple(mlir::ModuleOp mod, llvm::StringRef triple) {
   auto target = fir::determineTargetTriple(triple);
-  mod->setAttr(tripleName, mlir::StringAttr::get(mod.getContext(), target));
+  mod->setAttr(mlir::LLVM::LLVMDialect::getTargetTripleAttrName(),
+               mlir::StringAttr::get(mod.getContext(), target));
 }
 
 llvm::Triple fir::getTargetTriple(mlir::ModuleOp mod) {
-  if (auto target = mod->getAttrOfType<mlir::StringAttr>(tripleName))
+  if (auto target = mod->getAttrOfType<mlir::StringAttr>(
+          mlir::LLVM::LLVMDialect::getTargetTripleAttrName()))
     return llvm::Triple(target.getValue());
   return llvm::Triple(llvm::sys::getDefaultTargetTriple());
 }

flang/test/CMakeLists.txt

@@ -46,7 +46,7 @@ set(FLANG_TEST_PARAMS
   flang_site_config=${CMAKE_CURRENT_BINARY_DIR}/lit.site.cfg.py)
 
 set(FLANG_TEST_DEPENDS
-  flang-new llvm-config FileCheck count not module_files fir-opt tco bbc
+  flang-new llvm-config FileCheck count not module_files fir-opt tco bbc llvm-objdump
 )
 
 if (FLANG_INCLUDE_TESTS)

flang/test/Driver/code-gen-aarch64.f90

@@ -0,0 +1,31 @@
+! Test -emit-obj (X86)
+
+! REQUIRES: aarch64-registered-target, x86-registered-target
+
+!-------------
+! RUN COMMANDS
+!-------------
+! RUN: rm -f %t.o
+! RUN: %flang_fc1 -emit-obj -triple aarch64-unknown-linux-gnu %s -o %t.o
+! RUN: llvm-objdump --triple aarch64-unknown-linux-gnu --disassemble-all %t.o | FileCheck %s --check-prefix=CORRECT_TRIPLE
+! RUN: rm -f %t.o
+! RUN: %flang -c --target=aarch64-unknown-linux-gnu %s -o %t.o
+! RUN: llvm-objdump --triple aarch64-unknown-linux-gnu --disassemble-all %t.o | FileCheck %s --check-prefix=CORRECT_TRIPLE
+
+! RUN: %flang -c --target=aarch64-unknown-linux-gnu %s -o %t.o
+! RUN: llvm-objdump --triple x86_64-unknown-linux-gnu --disassemble-all %t.o | FileCheck %s --check-prefix=INCORRECT_TRIPLE
+
+!----------------
+! EXPECTED OUTPUT
+!----------------
+! CORRECT_TRIPLE-LABEL: <_QQmain>:
+! CORRECT_TRIPLE-NEXT:  	ret
+
+! When incorrect triple is used to disassemble, there won't be a ret instruction at all.
+! INCORRECT_TRIPLE-LABEL: <_QQmain>:
+! INCORRECT_TRIPLE-NOT:  	ret
+
+!------
+! INPUT
+!------
+end program

flang/test/Driver/code-gen-x86.f90

@@ -0,0 +1,32 @@
+! Test -emit-obj (X86)
+
+! REQUIRES: x86-registered-target, aarch64-registered-target
+! UNSUPPORTED: darwin, macos
+
+!-------------
+! RUN COMMANDS
+!-------------
+! RUN: rm -f %t.o
+! RUN: %flang_fc1 -triple x86_64-unknown-linux-gnu -emit-obj %s -o %t.o
+! RUN: llvm-objdump --triple x86_64-unknown-linux-gnu --disassemble-all %t.o | FileCheck %s --check-prefix=CORRECT_TRIPLE
+! RUN: rm -f %t.o
+! RUN: %flang --target=x86_64-unknown-linux-gnu -c %s -o %t.o
+! RUN: llvm-objdump --triple x86_64-unknown-linux-gnu --disassemble-all %t.o | FileCheck %s --check-prefix=CORRECT_TRIPLE
+
+! RUN: %flang -c --target=x86_64-unknown-linux-gnu %s -o %t.o
+! RUN: llvm-objdump --triple aarch64-unknown-linux-gnu --disassemble-all %t.o | FileCheck %s --check-prefix=INCORRECT_TRIPLE
+
+!----------------
+! EXPECTED OUTPUT
+!----------------
+! CORRECT_TRIPLE-LABEL: <_QQmain>:
+! CORRECT_TRIPLE-NEXT:  	retq
+
+! When incorrect triple is used to disassemble, there won't be a ret(q) instruction at all.
+! INCORRECT_TRIPLE-LABEL: <_QQmain>:
+! INCORRECT_TRIPLE-NOT:  	ret
+
+!------
+! INPUT
+!------
+end program

flang/test/Driver/code-gen.f90

@@ -55,6 +55,7 @@
 ! CHECK-NEXT: -print-target-triple    Print the normalized target triple
 ! CHECK-NEXT: -P                     Disable linemarker output in -E mode
 ! CHECK-NEXT: -std=<value>           Language standard to compile for
+! CHECK-NEXT: -S                     Only run preprocess and compilation steps
 ! CHECK-NEXT: --target=<value>        Generate code for the given target
 ! CHECK-NEXT: -U <macro>             Undefine macro <macro>
 ! CHECK-NEXT: --version Print version information

flang/test/Driver/driver-help-hidden.f90

@@ -55,6 +55,7 @@
 ! HELP-NEXT: -print-target-triple    Print the normalized target triple
 ! HELP-NEXT: -P                     Disable linemarker output in -E mode
 ! HELP-NEXT: -std=<value>           Language standard to compile for
+! HELP-NEXT: -S                     Only run preprocess and compilation steps
 ! HELP-NEXT: --target=<value>       Generate code for the given target
 ! HELP-NEXT: -U <macro>             Undefine macro <macro>
 ! HELP-NEXT: --version              Print version information
@@ -128,6 +129,7 @@
 ! HELP-FC1-NEXT: -plugin <name>         Use the named plugin action instead of the default action (use "help" to list available options)
 ! HELP-FC1-NEXT: -P                     Disable linemarker output in -E mode
 ! HELP-FC1-NEXT: -std=<value>           Language standard to compile for
+! HELP-FC1-NEXT: -S                     Only run preprocess and compilation steps
 ! HELP-FC1-NEXT: -test-io               Run the InputOuputTest action. Use for development and testing only.
 ! HELP-FC1-NEXT: -triple <value>        Specify target triple (e.g. i686-apple-darwin9)
 ! HELP-FC1-NEXT: -U <macro>             Undefine macro <macro>

flang/test/Driver/driver-help.f90

@@ -0,0 +1,21 @@
+! Test -S (AArch64)
+
+! REQUIRES: aarch64-registered-target
+
+!-------------
+! RUN COMMANDS
+!-------------
+! RUN: %flang_fc1 -S -triple aarch64-unknown-linux-gnu %s -o - | FileCheck %s
+! RUN: %flang -S -target aarch64-unknown-linux-gnu %s -o - | FileCheck %s
+
+!----------------
+! EXPECTED OUTPUT
+!----------------
+! CHECK-LABEL: _QQmain:
+! CHECK-NEXT: .Lfunc_begin0:
+! CHECK: ret
+
+!------
+! INPUT
+!------
+end program

flang/test/Driver/emit-asm-aarch64.f90

@@ -0,0 +1,21 @@
+! Test -S (X86)
+
+! REQUIRES: x86-registered-target
+
+!-------------
+! RUN COMMANDS
+!-------------
+! RUN: %flang_fc1 -S -triple x86_64-unknown-linux-gnu %s -o - | FileCheck %s
+! RUN: %flang -S -target x86_64-unknown-linux-gnu %s -o - | FileCheck %s
+
+!----------------
+! EXPECTED OUTPUT
+!----------------
+! CHECK-LABEL: _QQmain:
+! CHECK-NEXT: .Lfunc_begin0:
+! CHECK: ret
+
+!------
+! INPUT
+!------
+end program

flang/test/Driver/emit-asm-x86.f90

@@ -13,14 +13,15 @@
 ! RUN: %flang -fsyntax-only %s 2>&1 | FileCheck %s --allow-empty
 ! RUN: %flang_fc1 %s 2>&1 | FileCheck %s --allow-empty
 
-! RUN: not %flang  %s 2>&1 | FileCheck %s --check-prefix=NO_FSYNTAX_ONLY
-! RUN: not %flang_fc1 -emit-obj %s 2>&1 | FileCheck %s --check-prefix=NO_FSYNTAX_ONLY
+! RUN: rm -rf %t/non-existent-dir/
+! RUN: not %flang -c %s -o %t/non-existent-dir/syntax-only.o 2>&1 | FileCheck %s --check-prefix=NO_FSYNTAX_ONLY
+! RUN: not %flang_fc1 -emit-obj %s -o %t/non-existent-dir/syntax-only.o 2>&1 | FileCheck %s --check-prefix=NO_FSYNTAX_ONLY
 
 !-----------------
 ! EXPECTED OUTPUT
 !-----------------
 ! CHECK-NOT: error
-! NO_FSYNTAX_ONLY: error: code-generation is not available yet
+! NO_FSYNTAX_ONLY: error: failed to create the output file
 
 !-------
 ! INPUT