[AMDGPU] Implement IR expansion for frem instruction (llvm#130988)

This patch implements a correctly rounded expansion of the frem
instruction in LLVM IR. This is useful for target architectures for
which such an expansion is too involved to be implement in ISel
Lowering. The expansion is based on the code from the AMD device libs
and has been tested successfully against the OpenCL conformance tests on
amdgpu. The expansion is implemented in the preexisting "expand-fp"
pass. It replaces the expansion of "frem" in ISel for the amdgpu target;
it is enabled for targets which do not directly support "frem" and for
which no matching "fmod" LibCall is available.

---------

Co-authored-by: Matt Arsenault <[email protected]>

Author: frederik-h

llvm/include/llvm/CodeGen/ExpandFp.h

@@ -10,6 +10,7 @@
 #define LLVM_CODEGEN_EXPANDFP_H
 
 #include "llvm/IR/PassManager.h"
+#include "llvm/Support/CodeGen.h"
 
 namespace llvm {
 
@@ -18,11 +19,15 @@ class TargetMachine;
 class ExpandFpPass : public PassInfoMixin<ExpandFpPass> {
 private:
   const TargetMachine *TM;
+  CodeGenOptLevel OptLevel;
 
 public:
-  explicit ExpandFpPass(const TargetMachine *TM_) : TM(TM_) {}
+  explicit ExpandFpPass(const TargetMachine *TM, CodeGenOptLevel OptLevel);
 
   PreservedAnalyses run(Function &F, FunctionAnalysisManager &AM);
+
+  void printPipeline(raw_ostream &OS,
+                     function_ref<StringRef(StringRef)> MapClassName2PassName);
 };
 
 } // end namespace llvm

llvm/include/llvm/CodeGen/Passes.h

@@ -550,6 +550,9 @@ LLVM_ABI FunctionPass *createCFIFixup();
 /// Creates CFI Instruction Inserter pass. \see CFIInstrInserter.cpp
 LLVM_ABI FunctionPass *createCFIInstrInserter();
 
+// Expands floating point instructions.
+FunctionPass *createExpandFpPass(CodeGenOptLevel);
+
 /// Creates CFGuard longjmp target identification pass.
 /// \see CFGuardLongjmp.cpp
 LLVM_ABI FunctionPass *createCFGuardLongjmpPass();

llvm/include/llvm/Passes/CodeGenPassBuilder.h

@@ -737,7 +737,7 @@ void CodeGenPassBuilder<Derived, TargetMachineT>::addISelPasses(
 
   addPass(PreISelIntrinsicLoweringPass(&TM));
   addPass(ExpandLargeDivRemPass(&TM));
-  addPass(ExpandFpPass(&TM));
+  addPass(ExpandFpPass(&TM, getOptLevel()));
 
   derived().addIRPasses(addPass);
   derived().addCodeGenPrepare(addPass);