[𝘀𝗽𝗿] changes introduced through rebase

Created using spr 1.3.6-beta.1

[skip ci]

Author: necipfazil

llvm/include/llvm/CodeGen/AsmPrinter.h

@@ -356,6 +356,13 @@ class AsmPrinter : public MachineFunctionPass {
     DwarfUsesRelocationsAcrossSections = Enable;
   }
 
+  /// Generate and emit labels for callees of the indirect callsites which will
+  /// be used to populate the .callgraph section.
+  void emitIndirectCalleeLabels(
+      FunctionInfo &FuncInfo,
+      const MachineFunction::CallSiteInfoMap &CallSitesInfoMap,
+      MachineInstr &MI);
+
   //===------------------------------------------------------------------===//
   // XRay instrumentation implementation.
   //===------------------------------------------------------------------===//

llvm/include/llvm/CodeGen/MachineFunction.h

@@ -534,11 +534,12 @@ class LLVM_ABI MachineFunction {
     unsigned TargetFlags;
   };
 
+  using CallSiteInfoMap = DenseMap<const MachineInstr *, CallSiteInfo>;
+
 private:
   Delegate *TheDelegate = nullptr;
   GISelChangeObserver *Observer = nullptr;
 
-  using CallSiteInfoMap = DenseMap<const MachineInstr *, CallSiteInfo>;
   /// Map a call instruction to call site arguments forwarding info.
   CallSiteInfoMap CallSitesInfo;
 

llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp

@@ -1666,7 +1666,7 @@ static ConstantInt *extractNumericCGTypeId(const Function &F) {
   return ConstantInt::get(Int64Ty, TypeIdVal);
 }
 
-/// Emits call graph section.
+/// Emits .callgraph section.
 void AsmPrinter::emitCallGraphSection(const MachineFunction &MF,
                                       FunctionInfo &FuncInfo) {
   if (!MF.getTarget().Options.EmitCallGraphSection)
@@ -1675,7 +1675,7 @@ void AsmPrinter::emitCallGraphSection(const MachineFunction &MF,
   // Switch to the call graph section for the function
   MCSection *FuncCGSection =
       getObjFileLowering().getCallGraphSection(*getCurrentSection());
-  assert(FuncCGSection && "null call graph section");
+  assert(FuncCGSection && "null callgraph section");
   OutStreamer->pushSection();
   OutStreamer->switchSection(FuncCGSection);
 
@@ -1728,7 +1728,7 @@ void AsmPrinter::emitCallGraphSection(const MachineFunction &MF,
   OutStreamer->emitInt64(CallSiteLabels.size());
 
   // Emit the type id and call site label pairs.
-  for (const auto& [TypeId, Label] : CallSiteLabels) {
+  for (const auto &[TypeId, Label] : CallSiteLabels) {
     OutStreamer->emitInt64(TypeId);
     OutStreamer->emitSymbolValue(Label, TM.getProgramPointerSize());
   }
@@ -1867,6 +1867,30 @@ static StringRef getMIMnemonic(const MachineInstr &MI, MCStreamer &Streamer) {
   return Name;
 }
 
+void AsmPrinter::emitIndirectCalleeLabels(
+    FunctionInfo &FuncInfo,
+    const MachineFunction::CallSiteInfoMap &CallSitesInfoMap,
+    MachineInstr &MI) {
+  // Only indirect calls have type identifiers set.
+  const auto &CallSiteInfo = CallSitesInfoMap.find(&MI);
+  if (CallSiteInfo == CallSitesInfoMap.end())
+    return;
+  if (CallSiteInfo->second.CalleeTypeIds.empty())
+    return;
+
+  for (auto *CalleeTypeId : CallSiteInfo->second.CalleeTypeIds) {
+    // Emit label.
+    MCSymbol *S = MF->getContext().createTempSymbol();
+    OutStreamer->emitLabel(S);
+
+    // Get numeric callee_type id value.
+    uint64_t CalleeTypeIdVal = CalleeTypeId->getZExtValue();
+
+    // Add to function's callsite labels.
+    FuncInfo.CallSiteLabels.emplace_back(CalleeTypeIdVal, S);
+  }
+}
+
 /// EmitFunctionBody - This method emits the body and trailer for a
 /// function.
 void AsmPrinter::emitFunctionBody() {
@@ -2045,28 +2069,9 @@ void AsmPrinter::emitFunctionBody() {
         break;
       }
 
-      // FIXME: Some indirect calls can get lowered to jump instructions,
-      // resulting in emitting labels for them. The extra information can
-      // be neglected while disassembling but still takes space in the binary.
-      if (TM.Options.EmitCallGraphSection && MI.isCall()) {
-        // Only indirect calls have type identifiers set.
-        const auto &CallSiteInfo = CallSitesInfoMap.find(&MI);
-        if (CallSiteInfo != CallSitesInfoMap.end()) {
-          if (!CallSiteInfo->second.CalleeTypeIds.empty()) {
-            for (auto *CalleeTypeId : CallSiteInfo->second.CalleeTypeIds) {
-              // Emit label.
-              MCSymbol *S = MF->getContext().createTempSymbol();
-              OutStreamer->emitLabel(S);
-
-              // Get numeric callee_type id value.
-              uint64_t CalleeTypeIdVal = CalleeTypeId->getZExtValue();
-
-              // Add to function's callsite labels.
-              FuncInfo.CallSiteLabels.emplace_back(CalleeTypeIdVal, S);
-            }
-          }
-        }
-      }
+      if (TM.Options.EmitCallGraphSection && MI.isCall())
+        emitIndirectCalleeLabels(FuncInfo, CallSitesInfoMap, MI);
+
       // If there is a post-instruction symbol, emit a label for it here.
       if (MCSymbol *S = MI.getPostInstrSymbol())
         OutStreamer->emitLabel(S);

llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp

@@ -4111,6 +4111,11 @@ Instruction *InstCombinerImpl::visitCallBase(CallBase &Call) {
             Call, Builder.CreateBitOrPointerCast(ReturnedArg, CallTy));
     }
 
+  // Drop unnecessary callee_type metadata from calls that were converted
+  // into direct calls.
+  if (Call.getMetadata(LLVMContext::MD_callee_type) && !Call.isIndirectCall())
+    Call.setMetadata(LLVMContext::MD_callee_type, nullptr);
+
   // Drop unnecessary kcfi operand bundles from calls that were converted
   // into direct calls.
   auto Bundle = Call.getOperandBundle(LLVMContext::OB_kcfi);

llvm/test/CodeGen/AArch64/callsite-emit-calleetypeid-tailcall.ll

@@ -0,0 +1,19 @@
+;; Tests that call site callee type ids can be extracted and set from
+;; callee_type metadata for indirect tail calls.
+
+;; Verify the exact calleeTypeId value to ensure it is not garbage but the value
+;; computed as the type id from the callee_type metadata.
+; RUN: llc --call-graph-section -mtriple aarch64-linux-gnu < %s -stop-after=finalize-isel -o - | FileCheck %s
+
+define dso_local noundef i32 @_Z13call_indirectPFicEc(ptr noundef readonly captures(none) %func, i8 noundef signext %x) local_unnamed_addr !type !0 {
+entry:
+  ; CHECK: callSites:
+  ; CHECK-NEXT: - { bb: {{.*}}, offset: {{.*}}, fwdArgRegs: [], calleeTypeIds:
+  ; CHECK-NEXT: [ 3498816979441845844 ] }
+  %call = tail call noundef i32 %func(i8 noundef signext %x), !callee_type !1
+  ret i32 %call
+}
+
+!0 = !{i64 0, !"_ZTSFiPvcE.generalized"}
+!1 = !{!2}
+!2 = !{i64 0, !"_ZTSFicE.generalized"}

llvm/test/CodeGen/AArch64/call-site-info-typeid.ll renamed to llvm/test/CodeGen/AArch64/callsite-emit-calleetypeid.ll

@@ -0,0 +1,19 @@
+;; Tests that call site callee type ids can be extracted and set from
+;; callee_type metadata for indirect tail calls.
+
+;; Verify the exact calleeTypeId value to ensure it is not garbage but the value
+;; computed as the type id from the callee_type metadata.
+; RUN: llc --call-graph-section -mtriple arm-linux-gnu < %s -stop-after=finalize-isel -o - | FileCheck %s
+
+define dso_local noundef i32 @_Z13call_indirectPFicEc(ptr noundef readonly captures(none) %func, i8 noundef signext %x) local_unnamed_addr !type !0 {
+entry:
+  ; CHECK: callSites:
+  ; CHECK-NEXT: - { bb: {{.*}}, offset: {{.*}}, fwdArgRegs: [], calleeTypeIds:
+  ; CHECK-NEXT: [ 3498816979441845844 ] }
+  %call = tail call noundef i32 %func(i8 noundef signext %x), !callee_type !1
+  ret i32 %call
+}
+
+!0 = !{i64 0, !"_ZTSFiPvcE.generalized"}
+!1 = !{!2}
+!2 = !{i64 0, !"_ZTSFicE.generalized"}

llvm/test/CodeGen/ARM/callsite-emit-calleetypeid-tailcall.ll

@@ -13,11 +13,11 @@
 # CHECK-NEXT: - { bb: {{.*}}, offset: {{.*}}, fwdArgRegs: []
 
 --- |  
-  declare noundef i32 @_Z4fizzii(i32 noundef %x, i32 noundef %y)
-  
-  declare noundef i32 @_Z4buzzii(i32 noundef %x, i32 noundef %y)
-  
-  define dso_local noundef i32 @_Z3barii(i32 noundef %x, i32 noundef %y) {
+  declare !type !0 noundef i32 @_Z4fizzii(i32 noundef %x, i32 noundef %y)
+
+  declare !type !0 noundef i32 @_Z4buzzii(i32 noundef %x, i32 noundef %y)
+
+  define dso_local noundef i32 @_Z3barii(i32 noundef %x, i32 noundef %y) !type !0 {
   entry:
     %x.addr = alloca i32, align 4
     %y.addr = alloca i32, align 4
@@ -33,7 +33,7 @@
     ret i32 %sub
   }
 
-  define dso_local noundef i32 @_Z3fooii(i32 noundef %x, i32 noundef %y) {
+  define dso_local noundef i32 @_Z3fooii(i32 noundef %x, i32 noundef %y) !type !0 {
   entry:
     %x.addr = alloca i32, align 4
     %y.addr = alloca i32, align 4