[debug-info] Add support for llvm.dbg.addr in DIBuilder.

[gottesmm]

Summary

I based this off of the API already create for llvm.dbg.value since both
intrinsics have the same arguments at the API level.

I added some tests exercising the API a little as well as an additional small
test that shows how one can use llvm.dbg.addr to limit the PC range where an
address value is available in the debugger. This is done by calling
llvm.dbg.value with undef and the same metadata info as one used to create the
llvm.dbg.addr.

rdar://83957028

[gottesmm]

Fixed a few typos/clang-formatted.

[gottesmm]

The test failures are in compiler-rt and are not related to this commit.

[aprantl]

This is looking good, thanks!

llvm/lib/IR/DIBuilder.cpp

1062

Is this similar enough to insertDbgValueIntrinsic and insertDbgDeclareIntrinsic that we should factor out a insertDbgIntrinsic(IntrinsicFn *, ...) helper?

[gottesmm]

Made changes as per review requests.

[aprantl]

Thanks, looks good!

llvm/include/llvm/IR/DIBuilder.h

 //===- DIBuilder.h - Debug Information Builder ------------------*- C++ -*-===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 //
namespace llvm {
   class DIBuilder {
     Module &M;
     LLVMContext &VMContext;
 
     DICompileUnit *CUNode;   ///< The one compile unit created by this DIBuiler.
     Function *DeclareFn;     ///< llvm.dbg.declare
     Function *ValueFn;       ///< llvm.dbg.value
     Function *LabelFn;       ///< llvm.dbg.label
+    Function *AddrFn;        ///< llvm.dbg.addr
 
     SmallVector<Metadata *, 4> AllEnumTypes;
     /// Track the RetainTypes, since they can be updated later on.
     SmallVector<TrackingMDNodeRef, 4> AllRetainTypes;
     SmallVector<Metadata *, 4> AllSubprograms;
     SmallVector<Metadata *, 4> AllGVs;
     SmallVector<TrackingMDNodeRef, 4> AllImportedModules;
     /// Map Macro parent (which can be DIMacroFile or nullptr) to a list of
Instruction *insertLabel(DILabel *LabelInfo, const DILocation *DL,
                              BasicBlock *InsertBB, Instruction *InsertBefore);
 
     /// Internal helper for insertDbgValueIntrinsic.
     Instruction *
     insertDbgValueIntrinsic(llvm::Value *Val, DILocalVariable *VarInfo,
                             DIExpression *Expr, const DILocation *DL,
                             BasicBlock *InsertBB, Instruction *InsertBefore);
 
+    /// Internal helper for insertDbgAddrIntrinsic.
+    Instruction *
+    insertDbgAddrIntrinsic(llvm::Value *Val, DILocalVariable *VarInfo,
+                           DIExpression *Expr, const DILocation *DL,
+                           BasicBlock *InsertBB, Instruction *InsertBefore);
+
   public:
     /// Construct a builder for a module.
     ///
     /// If \c AllowUnresolved, collect unresolved nodes attached to the module
     /// in order to resolve cycles during \a finalize().
     ///
     /// If \p CU is given a value other than nullptr, then set \p CUNode to CU.
     explicit DIBuilder(Module &M, bool AllowUnresolved = true,
public:
     /// \param DL           Debug info location.
     /// \param InsertBefore Location for the new intrinsic.
     Instruction *insertDbgValueIntrinsic(llvm::Value *Val,
                                          DILocalVariable *VarInfo,
                                          DIExpression *Expr,
                                          const DILocation *DL,
                                          Instruction *InsertBefore);
 
+    /// Insert a new llvm.dbg.addr intrinsic call.
+    /// \param Addr          llvm::Value of the address
+    /// \param VarInfo      Variable's debug info descriptor.
+    /// \param Expr         A complex location expression.
+    /// \param DL           Debug info location.
+    /// \param InsertAtEnd Location for the new intrinsic.
+    Instruction *insertDbgAddrIntrinsic(llvm::Value *Addr,
+                                        DILocalVariable *VarInfo,
+                                        DIExpression *Expr,
+                                        const DILocation *DL,
+                                        BasicBlock *InsertAtEnd);
+
+    /// Insert a new llvm.dbg.addr intrinsic call.
+    /// \param Addr         llvm::Value of the address.
+    /// \param VarInfo      Variable's debug info descriptor.
+    /// \param Expr         A complex location expression.
+    /// \param DL           Debug info location.
+    /// \param InsertBefore Location for the new intrinsic.
+    Instruction *insertDbgAddrIntrinsic(llvm::Value *Addr,
+                                        DILocalVariable *VarInfo,
+                                        DIExpression *Expr,
+                                        const DILocation *DL,
+                                        Instruction *InsertBefore);
+
     /// Replace the vtable holder in the given type.
     ///
     /// If this creates a self reference, it may orphan some unresolved cycles
     /// in the operands of \c T, so \a DIBuilder needs to track that.
     void replaceVTableHolder(DICompositeType *&T,
                              DIType *VTableHolder);
 
     /// Replace arrays on a composite type.

llvm/lib/IR/DIBuilder.cpp

 //===--- DIBuilder.cpp - Debug Information Builder ------------------------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 //

 
 static cl::opt<bool>
     UseDbgAddr("use-dbg-addr",
                llvm::cl::desc("Use llvm.dbg.addr for all local variables"),
                cl::init(false), cl::Hidden);
 
 DIBuilder::DIBuilder(Module &m, bool AllowUnresolvedNodes, DICompileUnit *CU)
     : M(m), VMContext(M.getContext()), CUNode(CU), DeclareFn(nullptr),
-      ValueFn(nullptr), LabelFn(nullptr),
+      ValueFn(nullptr), LabelFn(nullptr), AddrFn(nullptr),
       AllowUnresolvedNodes(AllowUnresolvedNodes) {
   if (CUNode) {
     if (const auto &ETs = CUNode->getEnumTypes())
       AllEnumTypes.assign(ETs.begin(), ETs.end());
     if (const auto &RTs = CUNode->getRetainedTypes())
       AllRetainTypes.assign(RTs.begin(), RTs.end());
     if (const auto &GVs = CUNode->getGlobalVariables())
       AllGVs.assign(GVs.begin(), GVs.end());

 Instruction *DIBuilder::insertDbgValueIntrinsic(Value *V,
                                                 DILocalVariable *VarInfo,
                                                 DIExpression *Expr,
                                                 const DILocation *DL,
                                                 BasicBlock *InsertAtEnd) {
   return insertDbgValueIntrinsic(V, VarInfo, Expr, DL, InsertAtEnd, nullptr);
 }
 
+Instruction *DIBuilder::insertDbgAddrIntrinsic(Value *V,
+                                               DILocalVariable *VarInfo,
+                                               DIExpression *Expr,
+                                               const DILocation *DL,
+                                               Instruction *InsertBefore) {
+  return insertDbgAddrIntrinsic(
+      V, VarInfo, Expr, DL, InsertBefore ? InsertBefore->getParent() : nullptr,
+      InsertBefore);
+}
+
+Instruction *DIBuilder::insertDbgAddrIntrinsic(Value *V,
+                                               DILocalVariable *VarInfo,
+                                               DIExpression *Expr,
+                                               const DILocation *DL,
+                                               BasicBlock *InsertAtEnd) {
+  return insertDbgAddrIntrinsic(V, VarInfo, Expr, DL, InsertAtEnd, nullptr);
+}
+
 /// Initialize IRBuilder for inserting dbg.declare and dbg.value intrinsics.
 /// This abstracts over the various ways to specify an insert position.
 static void initIRBuilder(IRBuilder<> &Builder, const DILocation *DL,
                           BasicBlock *InsertBB, Instruction *InsertBefore) {
   if (InsertBefore)
     Builder.SetInsertPoint(InsertBefore);
   else if (InsertBB)
     Builder.SetInsertPoint(InsertBB);
Value *Args[] = {getDbgIntrinsicValueImpl(VMContext, V),
                    MetadataAsValue::get(VMContext, VarInfo),
                    MetadataAsValue::get(VMContext, Expr)};
 
   IRBuilder<> B(DL->getContext());
   initIRBuilder(B, DL, InsertBB, InsertBefore);
   return B.CreateCall(ValueFn, Args);
 }
 
+Instruction *DIBuilder::insertDbgAddrIntrinsic(
+    Value *V, DILocalVariable *VarInfo, DIExpression *Expr,
+    const DILocation *DL, BasicBlock *InsertBB, Instruction *InsertBefore) {
+  assert(V && "no value passed to dbg.addr");
+  assert(VarInfo && "empty or invalid DILocalVariable* passed to dbg.value");
+  assert(DL && "Expected debug loc");
+  assert(DL->getScope()->getSubprogram() ==
+             VarInfo->getScope()->getSubprogram() &&
+         "Expected matching subprograms");
+  if (!AddrFn)
+    AddrFn = Intrinsic::getDeclaration(&M, Intrinsic::dbg_addr);
+
+  trackIfUnresolved(VarInfo);
+  trackIfUnresolved(Expr);
+  Value *Args[] = {getDbgIntrinsicValueImpl(VMContext, V),
+                   MetadataAsValue::get(VMContext, VarInfo),
+                   MetadataAsValue::get(VMContext, Expr)};
+
+  IRBuilder<> B(DL->getContext());
+  initIRBuilder(B, DL, InsertBB, InsertBefore);
+  return B.CreateCall(AddrFn, Args);
+}
+
 Instruction *DIBuilder::insertLabel(DILabel *LabelInfo, const DILocation *DL,
                                     BasicBlock *InsertBB,
                                     Instruction *InsertBefore) {
   assert(LabelInfo && "empty or invalid DILabel* passed to dbg.label");
   assert(DL && "Expected debug loc");
   assert(DL->getScope()->getSubprogram() ==
              LabelInfo->getScope()->getSubprogram() &&
          "Expected matching subprograms");

llvm/test/DebugInfo/X86/dbg-addr.ll

 ;; Run twice -- once with DBG_VALUEs, once with instruction referencing.
 ; RUN: llc %s -o %t.s -experimental-debug-variable-locations=false
 ; RUN: llvm-mc -triple x86_64--linux %t.s -filetype=obj -o %t.o
-; RUN: FileCheck < %t.s %s
+; RUN: FileCheck -input-file=%t.s %s
 ; RUN: llvm-dwarfdump %t.o | FileCheck %s --check-prefix=DWARF
 ; RUN: llc %s -o %t.s -experimental-debug-variable-locations=true
 ; RUN: llvm-mc -triple x86_64--linux %t.s -filetype=obj -o %t.o
-; RUN: FileCheck < %t.s %s
+; RUN: FileCheck -input-file=%t.s %s
 ; RUN: llvm-dwarfdump %t.o | FileCheck %s --check-prefix=DWARF
 
 
 ; Unlike dbg.declare, dbg.addr should be lowered to DBG_VALUE instructions. It
 ; is control-dependent.
 
 ; CHECK-LABEL: use_dbg_addr:
 ; CHECK: #DEBUG_VALUE: use_dbg_addr:o <- [$rsp+0]
 ; CHECK-NOT: #DEBUG_VALUE:
 
 ; DWARF: DW_TAG_variable
 ; DWARF-NEXT:              DW_AT_location (DW_OP_fbreg +0)
 ; DWARF-NEXT:              DW_AT_name ("o")
 
+; Make sure that in the second case, we properly get a validity range in the
+; dwarf for the value. This ensures that we can use this technique to invalidate
+; variables.
+
+; CHECK-LABEL: test_dbg_addr_and_dbg_val_undef
+; CHECK: #DEBUG_VALUE: test_dbg_addr_and_dbg_val_undef:second_o <- [$rsp+0]
+; CHECK: #DEBUG_VALUE: test_dbg_addr_and_dbg_val_undef:second_o <- undef
+; CHECK-NOT: #DEBUG_VALUE:
+
+; DWARF: DW_TAG_variable
+; DWARF-NEXT:              DW_AT_location (0x{{[0-9a-z][0-9a-z]*}}:
+; DWARF-NEXT:                 [0x{{[0-9a-z][0-9a-z]*}}, 0x{{[0-9a-z][0-9a-z]*}}): DW_OP_breg7 RSP+0)
+; DWARF-NEXT:              DW_AT_name ("second_o")
 
 ; ModuleID = 't.c'
 source_filename = "t.c"
 target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
 target triple = "x86_64--linux"
 
 %struct.Foo = type { i32 }
 
 ; Function Attrs: noinline nounwind uwtable
 define void @use_dbg_addr() #0 !dbg !7 {
 entry:
   %o = alloca %struct.Foo, align 4
   call void @llvm.dbg.addr(metadata %struct.Foo* %o, metadata !10, metadata !15), !dbg !16
   call void @escape_foo(%struct.Foo* %o), !dbg !17
   ret void, !dbg !18
 }
 
+define void @test_dbg_addr_and_dbg_val_undef() #0 !dbg !117 {
+entry:
+  %o = alloca %struct.Foo, align 4
+  call void @llvm.dbg.addr(metadata %struct.Foo* %o, metadata !1110, metadata !1115), !dbg !1116
+  call void @escape_foo(%struct.Foo* %o), !dbg !1117
+  call void @llvm.dbg.value(metadata %struct.Foo* undef, metadata !1110, metadata !1115), !dbg !1116
+  ret void, !dbg !1118
+}
+
 ; Function Attrs: nounwind readnone speculatable
 declare void @llvm.dbg.addr(metadata, metadata, metadata) #1
+declare void @llvm.dbg.value(metadata, metadata, metadata) #1
 
 declare void @escape_foo(%struct.Foo*)
 
 attributes #0 = { noinline nounwind uwtable }
 attributes #1 = { nounwind readnone speculatable }
 
 !llvm.dbg.cu = !{!0}
 !llvm.module.flags = !{!3, !4, !5}

 !11 = distinct !DICompositeType(tag: DW_TAG_structure_type, name: "Foo", file: !1, line: 1, size: 32, elements: !12)
 !12 = !{!13}
 !13 = !DIDerivedType(tag: DW_TAG_member, name: "x", scope: !11, file: !1, line: 1, baseType: !14, size: 32)
 !14 = !DIBasicType(name: "int", size: 32, encoding: DW_ATE_signed)
 !15 = !DIExpression()
 !16 = !DILocation(line: 4, column: 14, scope: !7)
 !17 = !DILocation(line: 5, column: 3, scope: !7)
 !18 = !DILocation(line: 6, column: 1, scope: !7)
+
+!117 = distinct !DISubprogram(name: "test_dbg_addr_and_dbg_val_undef", scope: !1, file: !1, line: 3, type: !118, isLocal: false, isDefinition: true, scopeLine: 3, flags: DIFlagPrototyped, isOptimized: false, unit: !0, retainedNodes: !2)
+!118 = !DISubroutineType(types: !119)
+!119 = !{null}
+!1110 = !DILocalVariable(name: "second_o", scope: !117, file: !1, line: 4, type: !1111)
+!1111 = distinct !DICompositeType(tag: DW_TAG_structure_type, name: "Foo", file: !1, line: 1, size: 32, elements: !1112)
+!1112 = !{!1113}
+!1113 = !DIDerivedType(tag: DW_TAG_member, name: "x", scope: !1111, file: !1, line: 1, baseType: !1114, size: 32)
+!1114 = !DIBasicType(name: "int", size: 32, encoding: DW_ATE_signed)
+!1115 = !DIExpression()
+!1116 = !DILocation(line: 4, column: 14, scope: !117)
+!1117 = !DILocation(line: 5, column: 3, scope: !117)
+!1118 = !DILocation(line: 6, column: 1, scope: !117)

llvm/unittests/IR/DebugInfoTest.cpp

 //===- llvm/unittest/IR/DebugInfo.cpp - DebugInfo tests -------------------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 
 #include "llvm/IR/DebugInfo.h"
 #include "llvm/ADT/APSInt.h"
 #include "llvm/AsmParser/Parser.h"
 #include "llvm/IR/DIBuilder.h"
 #include "llvm/IR/DebugInfoMetadata.h"
+#include "llvm/IR/IRBuilder.h"
 #include "llvm/IR/IntrinsicInst.h"
 #include "llvm/IR/LLVMContext.h"
+#include "llvm/IR/Metadata.h"
 #include "llvm/IR/Module.h"
 #include "llvm/IR/Verifier.h"
 #include "llvm/Support/SourceMgr.h"
 #include "llvm/Transforms/Utils/Local.h"
 #include "gtest/gtest.h"
 
 using namespace llvm;
 
TEST(DIBuilder, DIEnumerator) {
 
   auto *E1 = DIEnumerator::getIfExists(Ctx, I1, I1.isSigned(), "name");
   EXPECT_TRUE(E1);
 
   auto *E2 = DIEnumerator::getIfExists(Ctx, I2, I1.isSigned(), "name");
   EXPECT_FALSE(E2);
 }
 
+TEST(DIBuilder, createDbgAddr) {
+  LLVMContext C;
+  std::unique_ptr<Module> M = parseIR(C, R"(
+    define void @f() !dbg !6 {
+      %a = alloca i16, align 8
+      ;; It is important that we put the debug marker on the return.
+      ;; We take advantage of that to conjure up a debug loc without
+      ;; having to synthesize one programatically.
+      ret void, !dbg !11
+    }
+    declare void @llvm.dbg.value(metadata, metadata, metadata) #0
+    attributes #0 = { nounwind readnone speculatable willreturn }
+
+    !llvm.dbg.cu = !{!0}
+    !llvm.module.flags = !{!5}
+
+    !0 = distinct !DICompileUnit(language: DW_LANG_C, file: !1, producer: "debugify", isOptimized: true, runtimeVersion: 0, emissionKind: FullDebug, enums: !2)
+    !1 = !DIFile(filename: "t.ll", directory: "/")
+    !2 = !{}
+    !5 = !{i32 2, !"Debug Info Version", i32 3}
+    !6 = distinct !DISubprogram(name: "foo", linkageName: "foo", scope: null, file: !1, line: 1, type: !7, scopeLine: 1, spFlags: DISPFlagDefinition | DISPFlagOptimized, unit: !0, retainedNodes: !8)
+    !7 = !DISubroutineType(types: !2)
+    !8 = !{!9}
+    !9 = !DILocalVariable(name: "1", scope: !6, file: !1, line: 1, type: !10)
+    !10 = !DIBasicType(name: "ty16", size: 16, encoding: DW_ATE_unsigned)
+    !11 = !DILocation(line: 1, column: 1, scope: !6)
+)");
+  auto *F = M->getFunction("f");
+  auto *EntryBlock = &F->getEntryBlock(); 
+
+  auto *CU = cast<DICompileUnit>(M->getNamedMetadata("llvm.dbg.cu")->getOperand(0));
+  auto *Alloca = &*EntryBlock->begin();
+  auto *Ret = EntryBlock->getTerminator();
+
+  auto *SP = cast<DISubprogram>(F->getMetadata(LLVMContext::MD_dbg));
+  auto *File = SP->getFile();
+  std::string Name = "myName";
+  const auto *Loc = Ret->getDebugLoc().get();
+
+  IRBuilder<> Builder(EntryBlock);
+  DIBuilder DIB(*M, true, CU);
+  DIType *DT = DIB.createBasicType("ty16", 16, dwarf::DW_ATE_unsigned);
+
+  DILocalVariable *LocalVar =
+    DIB.createAutoVariable(SP, Name, File, 5/*line*/, DT,
+                             /*AlwaysPreserve=*/true);
+
+  auto *Inst = DIB.insertDbgAddrIntrinsic(Alloca, LocalVar,
+                                          DIB.createExpression(), Loc, Ret);
+
+  DIB.finalize();
+
+  EXPECT_EQ(Inst->getDebugLoc().get(), Loc);
+
+  auto *MD0 = cast<MetadataAsValue>(Inst->getOperand(0))->getMetadata();
+  auto *MD0Local = cast<LocalAsMetadata>(MD0);
+  EXPECT_EQ(MD0Local->getValue(), Alloca);
+  auto *MD1 = cast<MetadataAsValue>(Inst->getOperand(1))->getMetadata();
+  EXPECT_EQ(MD1->getMetadataID(),
+            Metadata::MetadataKind::DILocalVariableKind);
+  auto *MD2 = cast<MetadataAsValue>(Inst->getOperand(2))->getMetadata();
+  auto *MDExp = cast<DIExpression>(MD2);
+  EXPECT_EQ(MDExp->getNumElements(), 0u);
+}
+
 } // end namespace