Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[DebugInfo] Add num_extra_inhabitants to debug info #112590

Open
wants to merge 2 commits into
base: main
Choose a base branch
from

Conversation

augusto2112
Copy link
Contributor

@augusto2112 augusto2112 commented Oct 16, 2024

An extra inhabitant is a bit pattern that does not represent a valid
value for instances of a given type. The number of extra inhabitants is
the number of those bit configurations.

This is used by Swift to save space when composing types. For example,
because Bool only needs 2 bit patterns to represent all of its values
(true and false), an Optional only occupies 1 byte in memory by
using a bit configuration that is unused by Bool. Which bit patterns are
unused are part of the ABI of the language.

Since Swift generics are not monomorphized, by using dynamic libraries
you can have generic types whose size, alignment, etc, are known only
at runtime (which is why this feature is needed).

This patch adds num_extra_inhabitants to LLVM-IR debug info and in DWARF
as an Apple extension.

@llvmbot
Copy link
Collaborator

llvmbot commented Oct 16, 2024

@llvm/pr-subscribers-debuginfo

Author: Augusto Noronha (augusto2112)

Changes

An extra inhabitant is a bit pattern that does not represent a valid value for instances of a given type. The number of extra inhabitants is the number of those bit configurations.

This is used by Swift to save space when composing types. For example, because Bool only needs 2 bit patterns to represent all of its values (true and false), an Optional<Bool> only occupies 1 byte in memory by using a bit configuration that is unused by Bool.

This patch adds num_extra_inhabitants to LLVM-IR debug info and in DWARF as an Apple extension.


Patch is 55.36 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/112590.diff

15 Files Affected:

  • (modified) llvm/include/llvm/BinaryFormat/Dwarf.def (+1)
  • (modified) llvm/include/llvm/IR/DIBuilder.h (+10-2)
  • (modified) llvm/include/llvm/IR/DebugInfoMetadata.h (+73-54)
  • (modified) llvm/lib/AsmParser/LLParser.cpp (+11-7)
  • (modified) llvm/lib/Bitcode/Reader/MetadataLoader.cpp (+9-7)
  • (modified) llvm/lib/Bitcode/Writer/BitcodeWriter.cpp (+2)
  • (modified) llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp (+8)
  • (modified) llvm/lib/IR/AsmWriter.cpp (+2)
  • (modified) llvm/lib/IR/DIBuilder.cpp (+9-5)
  • (modified) llvm/lib/IR/DebugInfoMetadata.cpp (+28-25)
  • (modified) llvm/lib/IR/LLVMContextImpl.h (+18-8)
  • (modified) llvm/test/Assembler/debug-info.ll (+8-2)
  • (added) llvm/test/DebugInfo/AArch64/num_extra_inhabitants.ll (+30)
  • (modified) llvm/unittests/IR/DebugTypeODRUniquingTest.cpp (+18-17)
  • (modified) llvm/unittests/IR/MetadataTest.cpp (+19-14)
diff --git a/llvm/include/llvm/BinaryFormat/Dwarf.def b/llvm/include/llvm/BinaryFormat/Dwarf.def
index d55947fc5103ac..66b397b1935943 100644
--- a/llvm/include/llvm/BinaryFormat/Dwarf.def
+++ b/llvm/include/llvm/BinaryFormat/Dwarf.def
@@ -636,6 +636,7 @@ HANDLE_DW_AT(0x3fed, APPLE_property, 0, APPLE)
 HANDLE_DW_AT(0x3fee, APPLE_objc_direct, 0, APPLE)
 HANDLE_DW_AT(0x3fef, APPLE_sdk, 0, APPLE)
 HANDLE_DW_AT(0x3ff0, APPLE_origin, 0, APPLE)
+HANDLE_DW_AT(0x3ff1, APPLE_num_extra_inhabitants, 0, APPLE)
 
 // Attribute form encodings.
 HANDLE_DW_FORM(0x01, addr, 2, DWARF)
diff --git a/llvm/include/llvm/IR/DIBuilder.h b/llvm/include/llvm/IR/DIBuilder.h
index a5655f630fc4f0..6b311b61152b2a 100644
--- a/llvm/include/llvm/IR/DIBuilder.h
+++ b/llvm/include/llvm/IR/DIBuilder.h
@@ -225,9 +225,13 @@ namespace llvm {
     /// \param SizeInBits  Size of the type.
     /// \param Encoding    DWARF encoding code, e.g., dwarf::DW_ATE_float.
     /// \param Flags       Optional DWARF attributes, e.g., DW_AT_endianity.
+    /// \param NumExtraInhabitants The number of extra inhabitants of the type.
+    /// An extra inhabitant is a bit pattern that does not represent a valid
+    /// value for instances of a given type.
     DIBasicType *createBasicType(StringRef Name, uint64_t SizeInBits,
                                  unsigned Encoding,
-                                 DINode::DIFlags Flags = DINode::FlagZero);
+                                 DINode::DIFlags Flags = DINode::FlagZero,
+                                 uint32_t NumExtraInhabitants = 0);
 
     /// Create debugging information entry for a string
     /// type.
@@ -484,11 +488,15 @@ namespace llvm {
     /// \param Elements     Struct elements.
     /// \param RunTimeLang  Optional parameter, Objective-C runtime version.
     /// \param UniqueIdentifier A unique identifier for the struct.
+    /// \param NumExtraInhabitants The number of extra inhabitants of the type.
+    /// An extra inhabitant is a bit pattern that does not represent a valid
+    /// value for instances of a given type.
     DICompositeType *createStructType(
         DIScope *Scope, StringRef Name, DIFile *File, unsigned LineNumber,
         uint64_t SizeInBits, uint32_t AlignInBits, DINode::DIFlags Flags,
         DIType *DerivedFrom, DINodeArray Elements, unsigned RunTimeLang = 0,
-        DIType *VTableHolder = nullptr, StringRef UniqueIdentifier = "");
+        DIType *VTableHolder = nullptr, StringRef UniqueIdentifier = "", 
+        uint32_t NumExtraInhabitants = 0);
 
     /// Create debugging information entry for an union.
     /// \param Scope        Scope in which this union is defined.
diff --git a/llvm/include/llvm/IR/DebugInfoMetadata.h b/llvm/include/llvm/IR/DebugInfoMetadata.h
index d2b4e900438d37..1043a5cf24bd76 100644
--- a/llvm/include/llvm/IR/DebugInfoMetadata.h
+++ b/llvm/include/llvm/IR/DebugInfoMetadata.h
@@ -712,31 +712,38 @@ class DIType : public DIScope {
   DIFlags Flags;
   uint64_t SizeInBits;
   uint64_t OffsetInBits;
+  uint32_t NumExtraInhabitants;
 
 protected:
   DIType(LLVMContext &C, unsigned ID, StorageType Storage, unsigned Tag,
          unsigned Line, uint64_t SizeInBits, uint32_t AlignInBits,
-         uint64_t OffsetInBits, DIFlags Flags, ArrayRef<Metadata *> Ops)
+         uint64_t OffsetInBits, uint32_t NumExtraInhabitants, DIFlags Flags,
+         ArrayRef<Metadata *> Ops)
       : DIScope(C, ID, Storage, Tag, Ops) {
-    init(Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    init(Line, SizeInBits, AlignInBits, OffsetInBits, NumExtraInhabitants,
+         Flags);
   }
   ~DIType() = default;
 
   void init(unsigned Line, uint64_t SizeInBits, uint32_t AlignInBits,
-            uint64_t OffsetInBits, DIFlags Flags) {
+            uint64_t OffsetInBits, uint32_t NumExtraInhabitants,
+            DIFlags Flags) {
     this->Line = Line;
     this->Flags = Flags;
     this->SizeInBits = SizeInBits;
     this->SubclassData32 = AlignInBits;
     this->OffsetInBits = OffsetInBits;
+    this->NumExtraInhabitants = NumExtraInhabitants;
   }
 
   /// Change fields in place.
   void mutate(unsigned Tag, unsigned Line, uint64_t SizeInBits,
-              uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags) {
+              uint32_t AlignInBits, uint64_t OffsetInBits,
+              uint32_t NumExtraInhabitants, DIFlags Flags) {
     assert(isDistinct() && "Only distinct nodes can mutate");
     setTag(Tag);
-    init(Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    init(Line, SizeInBits, AlignInBits, OffsetInBits, NumExtraInhabitants,
+         Flags);
   }
 
 public:
@@ -749,6 +756,7 @@ class DIType : public DIScope {
   uint32_t getAlignInBits() const;
   uint32_t getAlignInBytes() const { return getAlignInBits() / CHAR_BIT; }
   uint64_t getOffsetInBits() const { return OffsetInBits; }
+  uint32_t getNumExtraInhabitants() const { return NumExtraInhabitants; }
   DIFlags getFlags() const { return Flags; }
 
   DIScope *getScope() const { return cast_or_null<DIScope>(getRawScope()); }
@@ -820,49 +828,55 @@ class DIBasicType : public DIType {
 
   DIBasicType(LLVMContext &C, StorageType Storage, unsigned Tag,
               uint64_t SizeInBits, uint32_t AlignInBits, unsigned Encoding,
-              DIFlags Flags, ArrayRef<Metadata *> Ops)
+              uint32_t NumExtraInhabitants, DIFlags Flags,
+              ArrayRef<Metadata *> Ops)
       : DIType(C, DIBasicTypeKind, Storage, Tag, 0, SizeInBits, AlignInBits, 0,
-               Flags, Ops),
+               NumExtraInhabitants, Flags, Ops),
         Encoding(Encoding) {}
   ~DIBasicType() = default;
 
   static DIBasicType *getImpl(LLVMContext &Context, unsigned Tag,
                               StringRef Name, uint64_t SizeInBits,
                               uint32_t AlignInBits, unsigned Encoding,
-                              DIFlags Flags, StorageType Storage,
-                              bool ShouldCreate = true) {
+                              uint32_t NumExtraInhabitants, DIFlags Flags,
+                              StorageType Storage, bool ShouldCreate = true) {
     return getImpl(Context, Tag, getCanonicalMDString(Context, Name),
-                   SizeInBits, AlignInBits, Encoding, Flags, Storage,
-                   ShouldCreate);
+                   SizeInBits, AlignInBits, Encoding, NumExtraInhabitants,
+                   Flags, Storage, ShouldCreate);
   }
   static DIBasicType *getImpl(LLVMContext &Context, unsigned Tag,
                               MDString *Name, uint64_t SizeInBits,
                               uint32_t AlignInBits, unsigned Encoding,
-                              DIFlags Flags, StorageType Storage,
-                              bool ShouldCreate = true);
+                              uint32_t NumExtraInhabitants, DIFlags Flags,
+                              StorageType Storage, bool ShouldCreate = true);
 
   TempDIBasicType cloneImpl() const {
     return getTemporary(getContext(), getTag(), getName(), getSizeInBits(),
-                        getAlignInBits(), getEncoding(), getFlags());
+                        getAlignInBits(), getEncoding(),
+                        getNumExtraInhabitants(), getFlags());
   }
 
 public:
   DEFINE_MDNODE_GET(DIBasicType, (unsigned Tag, StringRef Name),
-                    (Tag, Name, 0, 0, 0, FlagZero))
+                    (Tag, Name, 0, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, StringRef Name, uint64_t SizeInBits),
-                    (Tag, Name, SizeInBits, 0, 0, FlagZero))
+                    (Tag, Name, SizeInBits, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, MDString *Name, uint64_t SizeInBits),
-                    (Tag, Name, SizeInBits, 0, 0, FlagZero))
+                    (Tag, Name, SizeInBits, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, StringRef Name, uint64_t SizeInBits,
-                     uint32_t AlignInBits, unsigned Encoding, DIFlags Flags),
-                    (Tag, Name, SizeInBits, AlignInBits, Encoding, Flags))
+                     uint32_t AlignInBits, unsigned Encoding,
+                     uint32_t NumExtraInhabitants, DIFlags Flags),
+                    (Tag, Name, SizeInBits, AlignInBits, Encoding,
+                     NumExtraInhabitants, Flags))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, MDString *Name, uint64_t SizeInBits,
-                     uint32_t AlignInBits, unsigned Encoding, DIFlags Flags),
-                    (Tag, Name, SizeInBits, AlignInBits, Encoding, Flags))
+                     uint32_t AlignInBits, unsigned Encoding,
+                     uint32_t NumExtraInhabitants, DIFlags Flags),
+                    (Tag, Name, SizeInBits, AlignInBits, Encoding,
+                     NumExtraInhabitants, Flags))
 
   TempDIBasicType clone() const { return cloneImpl(); }
 
@@ -890,7 +904,7 @@ class DIStringType : public DIType {
                uint64_t SizeInBits, uint32_t AlignInBits, unsigned Encoding,
                ArrayRef<Metadata *> Ops)
       : DIType(C, DIStringTypeKind, Storage, Tag, 0, SizeInBits, AlignInBits, 0,
-               FlagZero, Ops),
+               0, FlagZero, Ops),
         Encoding(Encoding) {}
   ~DIStringType() = default;
 
@@ -1016,7 +1030,7 @@ class DIDerivedType : public DIType {
                 std::optional<PtrAuthData> PtrAuthData, DIFlags Flags,
                 ArrayRef<Metadata *> Ops)
       : DIType(C, DIDerivedTypeKind, Storage, Tag, Line, SizeInBits,
-               AlignInBits, OffsetInBits, Flags, Ops),
+               AlignInBits, OffsetInBits, 0, Flags, Ops),
         DWARFAddressSpace(DWARFAddressSpace) {
     if (PtrAuthData)
       SubclassData32 = PtrAuthData->RawData;
@@ -1157,39 +1171,43 @@ class DICompositeType : public DIType {
 
   DICompositeType(LLVMContext &C, StorageType Storage, unsigned Tag,
                   unsigned Line, unsigned RuntimeLang, uint64_t SizeInBits,
-                  uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags,
+                  uint32_t AlignInBits, uint64_t OffsetInBits,
+                  uint32_t NumExtraInhabitants, DIFlags Flags,
                   ArrayRef<Metadata *> Ops)
       : DIType(C, DICompositeTypeKind, Storage, Tag, Line, SizeInBits,
-               AlignInBits, OffsetInBits, Flags, Ops),
+               AlignInBits, OffsetInBits, NumExtraInhabitants, Flags, Ops),
         RuntimeLang(RuntimeLang) {}
   ~DICompositeType() = default;
 
   /// Change fields in place.
   void mutate(unsigned Tag, unsigned Line, unsigned RuntimeLang,
               uint64_t SizeInBits, uint32_t AlignInBits, uint64_t OffsetInBits,
-              DIFlags Flags) {
+              uint32_t NumExtraInhabitants, DIFlags Flags) {
     assert(isDistinct() && "Only distinct nodes can mutate");
     assert(getRawIdentifier() && "Only ODR-uniqued nodes should mutate");
     this->RuntimeLang = RuntimeLang;
-    DIType::mutate(Tag, Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    DIType::mutate(Tag, Line, SizeInBits, AlignInBits, OffsetInBits,
+                   NumExtraInhabitants, Flags);
   }
 
   static DICompositeType *
   getImpl(LLVMContext &Context, unsigned Tag, StringRef Name, Metadata *File,
           unsigned Line, DIScope *Scope, DIType *BaseType, uint64_t SizeInBits,
-          uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags,
-          DINodeArray Elements, unsigned RuntimeLang, DIType *VTableHolder,
+          uint32_t AlignInBits, uint64_t OffsetInBits,
+          uint32_t NumExtraInhabitants, DIFlags Flags, DINodeArray Elements,
+          unsigned RuntimeLang, DIType *VTableHolder,
           DITemplateParameterArray TemplateParams, StringRef Identifier,
           DIDerivedType *Discriminator, Metadata *DataLocation,
           Metadata *Associated, Metadata *Allocated, Metadata *Rank,
           DINodeArray Annotations, StorageType Storage,
           bool ShouldCreate = true) {
-    return getImpl(
-        Context, Tag, getCanonicalMDString(Context, Name), File, Line, Scope,
-        BaseType, SizeInBits, AlignInBits, OffsetInBits, Flags, Elements.get(),
-        RuntimeLang, VTableHolder, TemplateParams.get(),
-        getCanonicalMDString(Context, Identifier), Discriminator, DataLocation,
-        Associated, Allocated, Rank, Annotations.get(), Storage, ShouldCreate);
+    return getImpl(Context, Tag, getCanonicalMDString(Context, Name), File,
+                   Line, Scope, BaseType, SizeInBits, AlignInBits, OffsetInBits,
+                   Flags, Elements.get(), RuntimeLang, VTableHolder,
+                   TemplateParams.get(),
+                   getCanonicalMDString(Context, Identifier), Discriminator,
+                   DataLocation, Associated, Allocated, Rank, Annotations.get(),
+                   NumExtraInhabitants, Storage, ShouldCreate);
   }
   static DICompositeType *
   getImpl(LLVMContext &Context, unsigned Tag, MDString *Name, Metadata *File,
@@ -1199,7 +1217,8 @@ class DICompositeType : public DIType {
           Metadata *VTableHolder, Metadata *TemplateParams,
           MDString *Identifier, Metadata *Discriminator, Metadata *DataLocation,
           Metadata *Associated, Metadata *Allocated, Metadata *Rank,
-          Metadata *Annotations, StorageType Storage, bool ShouldCreate = true);
+          Metadata *Annotations, uint32_t NumExtraInhabitants,
+          StorageType Storage, bool ShouldCreate = true);
 
   TempDICompositeType cloneImpl() const {
     return getTemporary(
@@ -1208,7 +1227,7 @@ class DICompositeType : public DIType {
         getFlags(), getElements(), getRuntimeLang(), getVTableHolder(),
         getTemplateParams(), getIdentifier(), getDiscriminator(),
         getRawDataLocation(), getRawAssociated(), getRawAllocated(),
-        getRawRank(), getAnnotations());
+        getRawRank(), getAnnotations(), getNumExtraInhabitants());
   }
 
 public:
@@ -1222,11 +1241,11 @@ class DICompositeType : public DIType {
        StringRef Identifier = "", DIDerivedType *Discriminator = nullptr,
        Metadata *DataLocation = nullptr, Metadata *Associated = nullptr,
        Metadata *Allocated = nullptr, Metadata *Rank = nullptr,
-       DINodeArray Annotations = nullptr),
+       DINodeArray Annotations = nullptr, uint32_t NumExtraInhabitants = 0),
       (Tag, Name, File, Line, Scope, BaseType, SizeInBits, AlignInBits,
-       OffsetInBits, Flags, Elements, RuntimeLang, VTableHolder, TemplateParams,
-       Identifier, Discriminator, DataLocation, Associated, Allocated, Rank,
-       Annotations))
+       OffsetInBits, NumExtraInhabitants, Flags, Elements, RuntimeLang,
+       VTableHolder, TemplateParams, Identifier, Discriminator, DataLocation,
+       Associated, Allocated, Rank, Annotations))
   DEFINE_MDNODE_GET(
       DICompositeType,
       (unsigned Tag, MDString *Name, Metadata *File, unsigned Line,
@@ -1236,11 +1255,12 @@ class DICompositeType : public DIType {
        Metadata *TemplateParams = nullptr, MDString *Identifier = nullptr,
        Metadata *Discriminator = nullptr, Metadata *DataLocation = nullptr,
        Metadata *Associated = nullptr, Metadata *Allocated = nullptr,
-       Metadata *Rank = nullptr, Metadata *Annotations = nullptr),
+       Metadata *Rank = nullptr, Metadata *Annotations = nullptr,
+       uint32_t NumExtraInhabitants = 0),
       (Tag, Name, File, Line, Scope, BaseType, SizeInBits, AlignInBits,
        OffsetInBits, Flags, Elements, RuntimeLang, VTableHolder, TemplateParams,
        Identifier, Discriminator, DataLocation, Associated, Allocated, Rank,
-       Annotations))
+       Annotations, NumExtraInhabitants))
 
   TempDICompositeType clone() const { return cloneImpl(); }
 
@@ -1255,8 +1275,8 @@ class DICompositeType : public DIType {
   getODRType(LLVMContext &Context, MDString &Identifier, unsigned Tag,
              MDString *Name, Metadata *File, unsigned Line, Metadata *Scope,
              Metadata *BaseType, uint64_t SizeInBits, uint32_t AlignInBits,
-             uint64_t OffsetInBits, DIFlags Flags, Metadata *Elements,
-             unsigned RuntimeLang, Metadata *VTableHolder,
+             uint64_t OffsetInBits, uint32_t NumExtraInhabitants, DIFlags Flags,
+             Metadata *Elements, unsigned RuntimeLang, Metadata *VTableHolder,
              Metadata *TemplateParams, Metadata *Discriminator,
              Metadata *DataLocation, Metadata *Associated, Metadata *Allocated,
              Metadata *Rank, Metadata *Annotations);
@@ -1272,15 +1292,14 @@ class DICompositeType : public DIType {
   ///
   /// If not \a LLVMContext::isODRUniquingDebugTypes(), this function returns
   /// nullptr.
-  static DICompositeType *
-  buildODRType(LLVMContext &Context, MDString &Identifier, unsigned Tag,
-               MDString *Name, Metadata *File, unsigned Line, Metadata *Scope,
-               Metadata *BaseType, uint64_t SizeInBits, uint32_t AlignInBits,
-               uint64_t OffsetInBits, DIFlags Flags, Metadata *Elements,
-               unsigned RuntimeLang, Metadata *VTableHolder,
-               Metadata *TemplateParams, Metadata *Discriminator,
-               Metadata *DataLocation, Metadata *Associated,
-               Metadata *Allocated, Metadata *Rank, Metadata *Annotations);
+  static DICompositeType *buildODRType(
+      LLVMContext &Context, MDString &Identifier, unsigned Tag, MDString *Name,
+      Metadata *File, unsigned Line, Metadata *Scope, Metadata *BaseType,
+      uint64_t SizeInBits, uint32_t AlignInBits, uint64_t OffsetInBits,
+      uint32_t NumExtraInhabitants, DIFlags Flags, Metadata *Elements,
+      unsigned RuntimeLang, Metadata *VTableHolder, Metadata *TemplateParams,
+      Metadata *Discriminator, Metadata *DataLocation, Metadata *Associated,
+      Metadata *Allocated, Metadata *Rank, Metadata *Annotations);
 
   DIType *getBaseType() const { return cast_or_null<DIType>(getRawBaseType()); }
   DINodeArray getElements() const {
diff --git a/llvm/lib/AsmParser/LLParser.cpp b/llvm/lib/AsmParser/LLParser.cpp
index 5b9bddeb7cfe82..af9155636a2239 100644
--- a/llvm/lib/AsmParser/LLParser.cpp
+++ b/llvm/lib/AsmParser/LLParser.cpp
@@ -5329,12 +5329,14 @@ bool LLParser::parseDIBasicType(MDNode *&Result, bool IsDistinct) {
   OPTIONAL(size, MDUnsignedField, (0, UINT64_MAX));                            \
   OPTIONAL(align, MDUnsignedField, (0, UINT32_MAX));                           \
   OPTIONAL(encoding, DwarfAttEncodingField, );                                 \
+  OPTIONAL(num_extra_inhabitants, MDUnsignedField, (0, UINT32_MAX));           \
   OPTIONAL(flags, DIFlagField, );
   PARSE_MD_FIELDS();
 #undef VISIT_MD_FIELDS
 
   Result = GET_OR_DISTINCT(DIBasicType, (Context, tag.Val, name.Val, size.Val,
-                                         align.Val, encoding.Val, flags.Val));
+                                         align.Val, encoding.Val,
+                                         num_extra_inhabitants.Val, flags.Val));
   return false;
 }
 
@@ -5432,7 +5434,8 @@ bool LLParser::parseDICompositeType(MDNode *&Result, bool IsDistinct) {
   OPTIONAL(associated, MDField, );                                             \
   OPTIONAL(allocated, MDField, );                                              \
   OPTIONAL(rank, MDSignedOrMDField, );                                         \
-  OPTIONAL(annotations, MDField, );
+  OPTIONAL(annotations, MDField, );                                            \
+  OPTIONAL(num_extra_inhabitants, MDUnsignedField, (0, UINT32_MAX));           
   PARSE_MD_FIELDS();
 #undef VISIT_MD_FIELDS
 
@@ -5447,10 +5450,11 @@ bool LLParser::parseDICompositeType(MDNode *&Result, bool IsDistinct) {
   if (identifier.Val)
     if (auto *CT = DICompositeType::buildODRType(
             Context, *identifier.Val, tag.Val, name.Val, file.Val, line.Val,
-            scope.Val, baseType.Val, size.Val, align.Val, offset.Val, flags.Val,
-            elements.Val, runtimeLang.Val, vtableHolder.Val, templateParams.Val,
-            discriminator.Val, dataLocation.Val, associated.Val, allocated.Val,
-            Rank, annotations.Val)) {
+            scope.Val, baseType.Val, size.Val, align.Val, offset.Val,
+            num_extra_inhabitants.Val, flags.Val, elements.Val, runtimeLang.Val,
+            vtableHolder.Val, templateParams.Val, discriminator.Val,
+            dataLocation.Val, asso...
[truncated]

@llvmbot
Copy link
Collaborator

llvmbot commented Oct 16, 2024

@llvm/pr-subscribers-llvm-ir

Author: Augusto Noronha (augusto2112)

Changes

An extra inhabitant is a bit pattern that does not represent a valid value for instances of a given type. The number of extra inhabitants is the number of those bit configurations.

This is used by Swift to save space when composing types. For example, because Bool only needs 2 bit patterns to represent all of its values (true and false), an Optional<Bool> only occupies 1 byte in memory by using a bit configuration that is unused by Bool.

This patch adds num_extra_inhabitants to LLVM-IR debug info and in DWARF as an Apple extension.


Patch is 55.36 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/112590.diff

15 Files Affected:

  • (modified) llvm/include/llvm/BinaryFormat/Dwarf.def (+1)
  • (modified) llvm/include/llvm/IR/DIBuilder.h (+10-2)
  • (modified) llvm/include/llvm/IR/DebugInfoMetadata.h (+73-54)
  • (modified) llvm/lib/AsmParser/LLParser.cpp (+11-7)
  • (modified) llvm/lib/Bitcode/Reader/MetadataLoader.cpp (+9-7)
  • (modified) llvm/lib/Bitcode/Writer/BitcodeWriter.cpp (+2)
  • (modified) llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp (+8)
  • (modified) llvm/lib/IR/AsmWriter.cpp (+2)
  • (modified) llvm/lib/IR/DIBuilder.cpp (+9-5)
  • (modified) llvm/lib/IR/DebugInfoMetadata.cpp (+28-25)
  • (modified) llvm/lib/IR/LLVMContextImpl.h (+18-8)
  • (modified) llvm/test/Assembler/debug-info.ll (+8-2)
  • (added) llvm/test/DebugInfo/AArch64/num_extra_inhabitants.ll (+30)
  • (modified) llvm/unittests/IR/DebugTypeODRUniquingTest.cpp (+18-17)
  • (modified) llvm/unittests/IR/MetadataTest.cpp (+19-14)
diff --git a/llvm/include/llvm/BinaryFormat/Dwarf.def b/llvm/include/llvm/BinaryFormat/Dwarf.def
index d55947fc5103ac..66b397b1935943 100644
--- a/llvm/include/llvm/BinaryFormat/Dwarf.def
+++ b/llvm/include/llvm/BinaryFormat/Dwarf.def
@@ -636,6 +636,7 @@ HANDLE_DW_AT(0x3fed, APPLE_property, 0, APPLE)
 HANDLE_DW_AT(0x3fee, APPLE_objc_direct, 0, APPLE)
 HANDLE_DW_AT(0x3fef, APPLE_sdk, 0, APPLE)
 HANDLE_DW_AT(0x3ff0, APPLE_origin, 0, APPLE)
+HANDLE_DW_AT(0x3ff1, APPLE_num_extra_inhabitants, 0, APPLE)
 
 // Attribute form encodings.
 HANDLE_DW_FORM(0x01, addr, 2, DWARF)
diff --git a/llvm/include/llvm/IR/DIBuilder.h b/llvm/include/llvm/IR/DIBuilder.h
index a5655f630fc4f0..6b311b61152b2a 100644
--- a/llvm/include/llvm/IR/DIBuilder.h
+++ b/llvm/include/llvm/IR/DIBuilder.h
@@ -225,9 +225,13 @@ namespace llvm {
     /// \param SizeInBits  Size of the type.
     /// \param Encoding    DWARF encoding code, e.g., dwarf::DW_ATE_float.
     /// \param Flags       Optional DWARF attributes, e.g., DW_AT_endianity.
+    /// \param NumExtraInhabitants The number of extra inhabitants of the type.
+    /// An extra inhabitant is a bit pattern that does not represent a valid
+    /// value for instances of a given type.
     DIBasicType *createBasicType(StringRef Name, uint64_t SizeInBits,
                                  unsigned Encoding,
-                                 DINode::DIFlags Flags = DINode::FlagZero);
+                                 DINode::DIFlags Flags = DINode::FlagZero,
+                                 uint32_t NumExtraInhabitants = 0);
 
     /// Create debugging information entry for a string
     /// type.
@@ -484,11 +488,15 @@ namespace llvm {
     /// \param Elements     Struct elements.
     /// \param RunTimeLang  Optional parameter, Objective-C runtime version.
     /// \param UniqueIdentifier A unique identifier for the struct.
+    /// \param NumExtraInhabitants The number of extra inhabitants of the type.
+    /// An extra inhabitant is a bit pattern that does not represent a valid
+    /// value for instances of a given type.
     DICompositeType *createStructType(
         DIScope *Scope, StringRef Name, DIFile *File, unsigned LineNumber,
         uint64_t SizeInBits, uint32_t AlignInBits, DINode::DIFlags Flags,
         DIType *DerivedFrom, DINodeArray Elements, unsigned RunTimeLang = 0,
-        DIType *VTableHolder = nullptr, StringRef UniqueIdentifier = "");
+        DIType *VTableHolder = nullptr, StringRef UniqueIdentifier = "", 
+        uint32_t NumExtraInhabitants = 0);
 
     /// Create debugging information entry for an union.
     /// \param Scope        Scope in which this union is defined.
diff --git a/llvm/include/llvm/IR/DebugInfoMetadata.h b/llvm/include/llvm/IR/DebugInfoMetadata.h
index d2b4e900438d37..1043a5cf24bd76 100644
--- a/llvm/include/llvm/IR/DebugInfoMetadata.h
+++ b/llvm/include/llvm/IR/DebugInfoMetadata.h
@@ -712,31 +712,38 @@ class DIType : public DIScope {
   DIFlags Flags;
   uint64_t SizeInBits;
   uint64_t OffsetInBits;
+  uint32_t NumExtraInhabitants;
 
 protected:
   DIType(LLVMContext &C, unsigned ID, StorageType Storage, unsigned Tag,
          unsigned Line, uint64_t SizeInBits, uint32_t AlignInBits,
-         uint64_t OffsetInBits, DIFlags Flags, ArrayRef<Metadata *> Ops)
+         uint64_t OffsetInBits, uint32_t NumExtraInhabitants, DIFlags Flags,
+         ArrayRef<Metadata *> Ops)
       : DIScope(C, ID, Storage, Tag, Ops) {
-    init(Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    init(Line, SizeInBits, AlignInBits, OffsetInBits, NumExtraInhabitants,
+         Flags);
   }
   ~DIType() = default;
 
   void init(unsigned Line, uint64_t SizeInBits, uint32_t AlignInBits,
-            uint64_t OffsetInBits, DIFlags Flags) {
+            uint64_t OffsetInBits, uint32_t NumExtraInhabitants,
+            DIFlags Flags) {
     this->Line = Line;
     this->Flags = Flags;
     this->SizeInBits = SizeInBits;
     this->SubclassData32 = AlignInBits;
     this->OffsetInBits = OffsetInBits;
+    this->NumExtraInhabitants = NumExtraInhabitants;
   }
 
   /// Change fields in place.
   void mutate(unsigned Tag, unsigned Line, uint64_t SizeInBits,
-              uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags) {
+              uint32_t AlignInBits, uint64_t OffsetInBits,
+              uint32_t NumExtraInhabitants, DIFlags Flags) {
     assert(isDistinct() && "Only distinct nodes can mutate");
     setTag(Tag);
-    init(Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    init(Line, SizeInBits, AlignInBits, OffsetInBits, NumExtraInhabitants,
+         Flags);
   }
 
 public:
@@ -749,6 +756,7 @@ class DIType : public DIScope {
   uint32_t getAlignInBits() const;
   uint32_t getAlignInBytes() const { return getAlignInBits() / CHAR_BIT; }
   uint64_t getOffsetInBits() const { return OffsetInBits; }
+  uint32_t getNumExtraInhabitants() const { return NumExtraInhabitants; }
   DIFlags getFlags() const { return Flags; }
 
   DIScope *getScope() const { return cast_or_null<DIScope>(getRawScope()); }
@@ -820,49 +828,55 @@ class DIBasicType : public DIType {
 
   DIBasicType(LLVMContext &C, StorageType Storage, unsigned Tag,
               uint64_t SizeInBits, uint32_t AlignInBits, unsigned Encoding,
-              DIFlags Flags, ArrayRef<Metadata *> Ops)
+              uint32_t NumExtraInhabitants, DIFlags Flags,
+              ArrayRef<Metadata *> Ops)
       : DIType(C, DIBasicTypeKind, Storage, Tag, 0, SizeInBits, AlignInBits, 0,
-               Flags, Ops),
+               NumExtraInhabitants, Flags, Ops),
         Encoding(Encoding) {}
   ~DIBasicType() = default;
 
   static DIBasicType *getImpl(LLVMContext &Context, unsigned Tag,
                               StringRef Name, uint64_t SizeInBits,
                               uint32_t AlignInBits, unsigned Encoding,
-                              DIFlags Flags, StorageType Storage,
-                              bool ShouldCreate = true) {
+                              uint32_t NumExtraInhabitants, DIFlags Flags,
+                              StorageType Storage, bool ShouldCreate = true) {
     return getImpl(Context, Tag, getCanonicalMDString(Context, Name),
-                   SizeInBits, AlignInBits, Encoding, Flags, Storage,
-                   ShouldCreate);
+                   SizeInBits, AlignInBits, Encoding, NumExtraInhabitants,
+                   Flags, Storage, ShouldCreate);
   }
   static DIBasicType *getImpl(LLVMContext &Context, unsigned Tag,
                               MDString *Name, uint64_t SizeInBits,
                               uint32_t AlignInBits, unsigned Encoding,
-                              DIFlags Flags, StorageType Storage,
-                              bool ShouldCreate = true);
+                              uint32_t NumExtraInhabitants, DIFlags Flags,
+                              StorageType Storage, bool ShouldCreate = true);
 
   TempDIBasicType cloneImpl() const {
     return getTemporary(getContext(), getTag(), getName(), getSizeInBits(),
-                        getAlignInBits(), getEncoding(), getFlags());
+                        getAlignInBits(), getEncoding(),
+                        getNumExtraInhabitants(), getFlags());
   }
 
 public:
   DEFINE_MDNODE_GET(DIBasicType, (unsigned Tag, StringRef Name),
-                    (Tag, Name, 0, 0, 0, FlagZero))
+                    (Tag, Name, 0, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, StringRef Name, uint64_t SizeInBits),
-                    (Tag, Name, SizeInBits, 0, 0, FlagZero))
+                    (Tag, Name, SizeInBits, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, MDString *Name, uint64_t SizeInBits),
-                    (Tag, Name, SizeInBits, 0, 0, FlagZero))
+                    (Tag, Name, SizeInBits, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, StringRef Name, uint64_t SizeInBits,
-                     uint32_t AlignInBits, unsigned Encoding, DIFlags Flags),
-                    (Tag, Name, SizeInBits, AlignInBits, Encoding, Flags))
+                     uint32_t AlignInBits, unsigned Encoding,
+                     uint32_t NumExtraInhabitants, DIFlags Flags),
+                    (Tag, Name, SizeInBits, AlignInBits, Encoding,
+                     NumExtraInhabitants, Flags))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, MDString *Name, uint64_t SizeInBits,
-                     uint32_t AlignInBits, unsigned Encoding, DIFlags Flags),
-                    (Tag, Name, SizeInBits, AlignInBits, Encoding, Flags))
+                     uint32_t AlignInBits, unsigned Encoding,
+                     uint32_t NumExtraInhabitants, DIFlags Flags),
+                    (Tag, Name, SizeInBits, AlignInBits, Encoding,
+                     NumExtraInhabitants, Flags))
 
   TempDIBasicType clone() const { return cloneImpl(); }
 
@@ -890,7 +904,7 @@ class DIStringType : public DIType {
                uint64_t SizeInBits, uint32_t AlignInBits, unsigned Encoding,
                ArrayRef<Metadata *> Ops)
       : DIType(C, DIStringTypeKind, Storage, Tag, 0, SizeInBits, AlignInBits, 0,
-               FlagZero, Ops),
+               0, FlagZero, Ops),
         Encoding(Encoding) {}
   ~DIStringType() = default;
 
@@ -1016,7 +1030,7 @@ class DIDerivedType : public DIType {
                 std::optional<PtrAuthData> PtrAuthData, DIFlags Flags,
                 ArrayRef<Metadata *> Ops)
       : DIType(C, DIDerivedTypeKind, Storage, Tag, Line, SizeInBits,
-               AlignInBits, OffsetInBits, Flags, Ops),
+               AlignInBits, OffsetInBits, 0, Flags, Ops),
         DWARFAddressSpace(DWARFAddressSpace) {
     if (PtrAuthData)
       SubclassData32 = PtrAuthData->RawData;
@@ -1157,39 +1171,43 @@ class DICompositeType : public DIType {
 
   DICompositeType(LLVMContext &C, StorageType Storage, unsigned Tag,
                   unsigned Line, unsigned RuntimeLang, uint64_t SizeInBits,
-                  uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags,
+                  uint32_t AlignInBits, uint64_t OffsetInBits,
+                  uint32_t NumExtraInhabitants, DIFlags Flags,
                   ArrayRef<Metadata *> Ops)
       : DIType(C, DICompositeTypeKind, Storage, Tag, Line, SizeInBits,
-               AlignInBits, OffsetInBits, Flags, Ops),
+               AlignInBits, OffsetInBits, NumExtraInhabitants, Flags, Ops),
         RuntimeLang(RuntimeLang) {}
   ~DICompositeType() = default;
 
   /// Change fields in place.
   void mutate(unsigned Tag, unsigned Line, unsigned RuntimeLang,
               uint64_t SizeInBits, uint32_t AlignInBits, uint64_t OffsetInBits,
-              DIFlags Flags) {
+              uint32_t NumExtraInhabitants, DIFlags Flags) {
     assert(isDistinct() && "Only distinct nodes can mutate");
     assert(getRawIdentifier() && "Only ODR-uniqued nodes should mutate");
     this->RuntimeLang = RuntimeLang;
-    DIType::mutate(Tag, Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    DIType::mutate(Tag, Line, SizeInBits, AlignInBits, OffsetInBits,
+                   NumExtraInhabitants, Flags);
   }
 
   static DICompositeType *
   getImpl(LLVMContext &Context, unsigned Tag, StringRef Name, Metadata *File,
           unsigned Line, DIScope *Scope, DIType *BaseType, uint64_t SizeInBits,
-          uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags,
-          DINodeArray Elements, unsigned RuntimeLang, DIType *VTableHolder,
+          uint32_t AlignInBits, uint64_t OffsetInBits,
+          uint32_t NumExtraInhabitants, DIFlags Flags, DINodeArray Elements,
+          unsigned RuntimeLang, DIType *VTableHolder,
           DITemplateParameterArray TemplateParams, StringRef Identifier,
           DIDerivedType *Discriminator, Metadata *DataLocation,
           Metadata *Associated, Metadata *Allocated, Metadata *Rank,
           DINodeArray Annotations, StorageType Storage,
           bool ShouldCreate = true) {
-    return getImpl(
-        Context, Tag, getCanonicalMDString(Context, Name), File, Line, Scope,
-        BaseType, SizeInBits, AlignInBits, OffsetInBits, Flags, Elements.get(),
-        RuntimeLang, VTableHolder, TemplateParams.get(),
-        getCanonicalMDString(Context, Identifier), Discriminator, DataLocation,
-        Associated, Allocated, Rank, Annotations.get(), Storage, ShouldCreate);
+    return getImpl(Context, Tag, getCanonicalMDString(Context, Name), File,
+                   Line, Scope, BaseType, SizeInBits, AlignInBits, OffsetInBits,
+                   Flags, Elements.get(), RuntimeLang, VTableHolder,
+                   TemplateParams.get(),
+                   getCanonicalMDString(Context, Identifier), Discriminator,
+                   DataLocation, Associated, Allocated, Rank, Annotations.get(),
+                   NumExtraInhabitants, Storage, ShouldCreate);
   }
   static DICompositeType *
   getImpl(LLVMContext &Context, unsigned Tag, MDString *Name, Metadata *File,
@@ -1199,7 +1217,8 @@ class DICompositeType : public DIType {
           Metadata *VTableHolder, Metadata *TemplateParams,
           MDString *Identifier, Metadata *Discriminator, Metadata *DataLocation,
           Metadata *Associated, Metadata *Allocated, Metadata *Rank,
-          Metadata *Annotations, StorageType Storage, bool ShouldCreate = true);
+          Metadata *Annotations, uint32_t NumExtraInhabitants,
+          StorageType Storage, bool ShouldCreate = true);
 
   TempDICompositeType cloneImpl() const {
     return getTemporary(
@@ -1208,7 +1227,7 @@ class DICompositeType : public DIType {
         getFlags(), getElements(), getRuntimeLang(), getVTableHolder(),
         getTemplateParams(), getIdentifier(), getDiscriminator(),
         getRawDataLocation(), getRawAssociated(), getRawAllocated(),
-        getRawRank(), getAnnotations());
+        getRawRank(), getAnnotations(), getNumExtraInhabitants());
   }
 
 public:
@@ -1222,11 +1241,11 @@ class DICompositeType : public DIType {
        StringRef Identifier = "", DIDerivedType *Discriminator = nullptr,
        Metadata *DataLocation = nullptr, Metadata *Associated = nullptr,
        Metadata *Allocated = nullptr, Metadata *Rank = nullptr,
-       DINodeArray Annotations = nullptr),
+       DINodeArray Annotations = nullptr, uint32_t NumExtraInhabitants = 0),
       (Tag, Name, File, Line, Scope, BaseType, SizeInBits, AlignInBits,
-       OffsetInBits, Flags, Elements, RuntimeLang, VTableHolder, TemplateParams,
-       Identifier, Discriminator, DataLocation, Associated, Allocated, Rank,
-       Annotations))
+       OffsetInBits, NumExtraInhabitants, Flags, Elements, RuntimeLang,
+       VTableHolder, TemplateParams, Identifier, Discriminator, DataLocation,
+       Associated, Allocated, Rank, Annotations))
   DEFINE_MDNODE_GET(
       DICompositeType,
       (unsigned Tag, MDString *Name, Metadata *File, unsigned Line,
@@ -1236,11 +1255,12 @@ class DICompositeType : public DIType {
        Metadata *TemplateParams = nullptr, MDString *Identifier = nullptr,
        Metadata *Discriminator = nullptr, Metadata *DataLocation = nullptr,
        Metadata *Associated = nullptr, Metadata *Allocated = nullptr,
-       Metadata *Rank = nullptr, Metadata *Annotations = nullptr),
+       Metadata *Rank = nullptr, Metadata *Annotations = nullptr,
+       uint32_t NumExtraInhabitants = 0),
       (Tag, Name, File, Line, Scope, BaseType, SizeInBits, AlignInBits,
        OffsetInBits, Flags, Elements, RuntimeLang, VTableHolder, TemplateParams,
        Identifier, Discriminator, DataLocation, Associated, Allocated, Rank,
-       Annotations))
+       Annotations, NumExtraInhabitants))
 
   TempDICompositeType clone() const { return cloneImpl(); }
 
@@ -1255,8 +1275,8 @@ class DICompositeType : public DIType {
   getODRType(LLVMContext &Context, MDString &Identifier, unsigned Tag,
              MDString *Name, Metadata *File, unsigned Line, Metadata *Scope,
              Metadata *BaseType, uint64_t SizeInBits, uint32_t AlignInBits,
-             uint64_t OffsetInBits, DIFlags Flags, Metadata *Elements,
-             unsigned RuntimeLang, Metadata *VTableHolder,
+             uint64_t OffsetInBits, uint32_t NumExtraInhabitants, DIFlags Flags,
+             Metadata *Elements, unsigned RuntimeLang, Metadata *VTableHolder,
              Metadata *TemplateParams, Metadata *Discriminator,
              Metadata *DataLocation, Metadata *Associated, Metadata *Allocated,
              Metadata *Rank, Metadata *Annotations);
@@ -1272,15 +1292,14 @@ class DICompositeType : public DIType {
   ///
   /// If not \a LLVMContext::isODRUniquingDebugTypes(), this function returns
   /// nullptr.
-  static DICompositeType *
-  buildODRType(LLVMContext &Context, MDString &Identifier, unsigned Tag,
-               MDString *Name, Metadata *File, unsigned Line, Metadata *Scope,
-               Metadata *BaseType, uint64_t SizeInBits, uint32_t AlignInBits,
-               uint64_t OffsetInBits, DIFlags Flags, Metadata *Elements,
-               unsigned RuntimeLang, Metadata *VTableHolder,
-               Metadata *TemplateParams, Metadata *Discriminator,
-               Metadata *DataLocation, Metadata *Associated,
-               Metadata *Allocated, Metadata *Rank, Metadata *Annotations);
+  static DICompositeType *buildODRType(
+      LLVMContext &Context, MDString &Identifier, unsigned Tag, MDString *Name,
+      Metadata *File, unsigned Line, Metadata *Scope, Metadata *BaseType,
+      uint64_t SizeInBits, uint32_t AlignInBits, uint64_t OffsetInBits,
+      uint32_t NumExtraInhabitants, DIFlags Flags, Metadata *Elements,
+      unsigned RuntimeLang, Metadata *VTableHolder, Metadata *TemplateParams,
+      Metadata *Discriminator, Metadata *DataLocation, Metadata *Associated,
+      Metadata *Allocated, Metadata *Rank, Metadata *Annotations);
 
   DIType *getBaseType() const { return cast_or_null<DIType>(getRawBaseType()); }
   DINodeArray getElements() const {
diff --git a/llvm/lib/AsmParser/LLParser.cpp b/llvm/lib/AsmParser/LLParser.cpp
index 5b9bddeb7cfe82..af9155636a2239 100644
--- a/llvm/lib/AsmParser/LLParser.cpp
+++ b/llvm/lib/AsmParser/LLParser.cpp
@@ -5329,12 +5329,14 @@ bool LLParser::parseDIBasicType(MDNode *&Result, bool IsDistinct) {
   OPTIONAL(size, MDUnsignedField, (0, UINT64_MAX));                            \
   OPTIONAL(align, MDUnsignedField, (0, UINT32_MAX));                           \
   OPTIONAL(encoding, DwarfAttEncodingField, );                                 \
+  OPTIONAL(num_extra_inhabitants, MDUnsignedField, (0, UINT32_MAX));           \
   OPTIONAL(flags, DIFlagField, );
   PARSE_MD_FIELDS();
 #undef VISIT_MD_FIELDS
 
   Result = GET_OR_DISTINCT(DIBasicType, (Context, tag.Val, name.Val, size.Val,
-                                         align.Val, encoding.Val, flags.Val));
+                                         align.Val, encoding.Val,
+                                         num_extra_inhabitants.Val, flags.Val));
   return false;
 }
 
@@ -5432,7 +5434,8 @@ bool LLParser::parseDICompositeType(MDNode *&Result, bool IsDistinct) {
   OPTIONAL(associated, MDField, );                                             \
   OPTIONAL(allocated, MDField, );                                              \
   OPTIONAL(rank, MDSignedOrMDField, );                                         \
-  OPTIONAL(annotations, MDField, );
+  OPTIONAL(annotations, MDField, );                                            \
+  OPTIONAL(num_extra_inhabitants, MDUnsignedField, (0, UINT32_MAX));           
   PARSE_MD_FIELDS();
 #undef VISIT_MD_FIELDS
 
@@ -5447,10 +5450,11 @@ bool LLParser::parseDICompositeType(MDNode *&Result, bool IsDistinct) {
   if (identifier.Val)
     if (auto *CT = DICompositeType::buildODRType(
             Context, *identifier.Val, tag.Val, name.Val, file.Val, line.Val,
-            scope.Val, baseType.Val, size.Val, align.Val, offset.Val, flags.Val,
-            elements.Val, runtimeLang.Val, vtableHolder.Val, templateParams.Val,
-            discriminator.Val, dataLocation.Val, associated.Val, allocated.Val,
-            Rank, annotations.Val)) {
+            scope.Val, baseType.Val, size.Val, align.Val, offset.Val,
+            num_extra_inhabitants.Val, flags.Val, elements.Val, runtimeLang.Val,
+            vtableHolder.Val, templateParams.Val, discriminator.Val,
+            dataLocation.Val, asso...
[truncated]

@llvmbot
Copy link
Collaborator

llvmbot commented Oct 16, 2024

@llvm/pr-subscribers-llvm-binary-utilities

Author: Augusto Noronha (augusto2112)

Changes

An extra inhabitant is a bit pattern that does not represent a valid value for instances of a given type. The number of extra inhabitants is the number of those bit configurations.

This is used by Swift to save space when composing types. For example, because Bool only needs 2 bit patterns to represent all of its values (true and false), an Optional<Bool> only occupies 1 byte in memory by using a bit configuration that is unused by Bool.

This patch adds num_extra_inhabitants to LLVM-IR debug info and in DWARF as an Apple extension.


Patch is 55.36 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/112590.diff

15 Files Affected:

  • (modified) llvm/include/llvm/BinaryFormat/Dwarf.def (+1)
  • (modified) llvm/include/llvm/IR/DIBuilder.h (+10-2)
  • (modified) llvm/include/llvm/IR/DebugInfoMetadata.h (+73-54)
  • (modified) llvm/lib/AsmParser/LLParser.cpp (+11-7)
  • (modified) llvm/lib/Bitcode/Reader/MetadataLoader.cpp (+9-7)
  • (modified) llvm/lib/Bitcode/Writer/BitcodeWriter.cpp (+2)
  • (modified) llvm/lib/CodeGen/AsmPrinter/DwarfUnit.cpp (+8)
  • (modified) llvm/lib/IR/AsmWriter.cpp (+2)
  • (modified) llvm/lib/IR/DIBuilder.cpp (+9-5)
  • (modified) llvm/lib/IR/DebugInfoMetadata.cpp (+28-25)
  • (modified) llvm/lib/IR/LLVMContextImpl.h (+18-8)
  • (modified) llvm/test/Assembler/debug-info.ll (+8-2)
  • (added) llvm/test/DebugInfo/AArch64/num_extra_inhabitants.ll (+30)
  • (modified) llvm/unittests/IR/DebugTypeODRUniquingTest.cpp (+18-17)
  • (modified) llvm/unittests/IR/MetadataTest.cpp (+19-14)
diff --git a/llvm/include/llvm/BinaryFormat/Dwarf.def b/llvm/include/llvm/BinaryFormat/Dwarf.def
index d55947fc5103ac..66b397b1935943 100644
--- a/llvm/include/llvm/BinaryFormat/Dwarf.def
+++ b/llvm/include/llvm/BinaryFormat/Dwarf.def
@@ -636,6 +636,7 @@ HANDLE_DW_AT(0x3fed, APPLE_property, 0, APPLE)
 HANDLE_DW_AT(0x3fee, APPLE_objc_direct, 0, APPLE)
 HANDLE_DW_AT(0x3fef, APPLE_sdk, 0, APPLE)
 HANDLE_DW_AT(0x3ff0, APPLE_origin, 0, APPLE)
+HANDLE_DW_AT(0x3ff1, APPLE_num_extra_inhabitants, 0, APPLE)
 
 // Attribute form encodings.
 HANDLE_DW_FORM(0x01, addr, 2, DWARF)
diff --git a/llvm/include/llvm/IR/DIBuilder.h b/llvm/include/llvm/IR/DIBuilder.h
index a5655f630fc4f0..6b311b61152b2a 100644
--- a/llvm/include/llvm/IR/DIBuilder.h
+++ b/llvm/include/llvm/IR/DIBuilder.h
@@ -225,9 +225,13 @@ namespace llvm {
     /// \param SizeInBits  Size of the type.
     /// \param Encoding    DWARF encoding code, e.g., dwarf::DW_ATE_float.
     /// \param Flags       Optional DWARF attributes, e.g., DW_AT_endianity.
+    /// \param NumExtraInhabitants The number of extra inhabitants of the type.
+    /// An extra inhabitant is a bit pattern that does not represent a valid
+    /// value for instances of a given type.
     DIBasicType *createBasicType(StringRef Name, uint64_t SizeInBits,
                                  unsigned Encoding,
-                                 DINode::DIFlags Flags = DINode::FlagZero);
+                                 DINode::DIFlags Flags = DINode::FlagZero,
+                                 uint32_t NumExtraInhabitants = 0);
 
     /// Create debugging information entry for a string
     /// type.
@@ -484,11 +488,15 @@ namespace llvm {
     /// \param Elements     Struct elements.
     /// \param RunTimeLang  Optional parameter, Objective-C runtime version.
     /// \param UniqueIdentifier A unique identifier for the struct.
+    /// \param NumExtraInhabitants The number of extra inhabitants of the type.
+    /// An extra inhabitant is a bit pattern that does not represent a valid
+    /// value for instances of a given type.
     DICompositeType *createStructType(
         DIScope *Scope, StringRef Name, DIFile *File, unsigned LineNumber,
         uint64_t SizeInBits, uint32_t AlignInBits, DINode::DIFlags Flags,
         DIType *DerivedFrom, DINodeArray Elements, unsigned RunTimeLang = 0,
-        DIType *VTableHolder = nullptr, StringRef UniqueIdentifier = "");
+        DIType *VTableHolder = nullptr, StringRef UniqueIdentifier = "", 
+        uint32_t NumExtraInhabitants = 0);
 
     /// Create debugging information entry for an union.
     /// \param Scope        Scope in which this union is defined.
diff --git a/llvm/include/llvm/IR/DebugInfoMetadata.h b/llvm/include/llvm/IR/DebugInfoMetadata.h
index d2b4e900438d37..1043a5cf24bd76 100644
--- a/llvm/include/llvm/IR/DebugInfoMetadata.h
+++ b/llvm/include/llvm/IR/DebugInfoMetadata.h
@@ -712,31 +712,38 @@ class DIType : public DIScope {
   DIFlags Flags;
   uint64_t SizeInBits;
   uint64_t OffsetInBits;
+  uint32_t NumExtraInhabitants;
 
 protected:
   DIType(LLVMContext &C, unsigned ID, StorageType Storage, unsigned Tag,
          unsigned Line, uint64_t SizeInBits, uint32_t AlignInBits,
-         uint64_t OffsetInBits, DIFlags Flags, ArrayRef<Metadata *> Ops)
+         uint64_t OffsetInBits, uint32_t NumExtraInhabitants, DIFlags Flags,
+         ArrayRef<Metadata *> Ops)
       : DIScope(C, ID, Storage, Tag, Ops) {
-    init(Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    init(Line, SizeInBits, AlignInBits, OffsetInBits, NumExtraInhabitants,
+         Flags);
   }
   ~DIType() = default;
 
   void init(unsigned Line, uint64_t SizeInBits, uint32_t AlignInBits,
-            uint64_t OffsetInBits, DIFlags Flags) {
+            uint64_t OffsetInBits, uint32_t NumExtraInhabitants,
+            DIFlags Flags) {
     this->Line = Line;
     this->Flags = Flags;
     this->SizeInBits = SizeInBits;
     this->SubclassData32 = AlignInBits;
     this->OffsetInBits = OffsetInBits;
+    this->NumExtraInhabitants = NumExtraInhabitants;
   }
 
   /// Change fields in place.
   void mutate(unsigned Tag, unsigned Line, uint64_t SizeInBits,
-              uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags) {
+              uint32_t AlignInBits, uint64_t OffsetInBits,
+              uint32_t NumExtraInhabitants, DIFlags Flags) {
     assert(isDistinct() && "Only distinct nodes can mutate");
     setTag(Tag);
-    init(Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    init(Line, SizeInBits, AlignInBits, OffsetInBits, NumExtraInhabitants,
+         Flags);
   }
 
 public:
@@ -749,6 +756,7 @@ class DIType : public DIScope {
   uint32_t getAlignInBits() const;
   uint32_t getAlignInBytes() const { return getAlignInBits() / CHAR_BIT; }
   uint64_t getOffsetInBits() const { return OffsetInBits; }
+  uint32_t getNumExtraInhabitants() const { return NumExtraInhabitants; }
   DIFlags getFlags() const { return Flags; }
 
   DIScope *getScope() const { return cast_or_null<DIScope>(getRawScope()); }
@@ -820,49 +828,55 @@ class DIBasicType : public DIType {
 
   DIBasicType(LLVMContext &C, StorageType Storage, unsigned Tag,
               uint64_t SizeInBits, uint32_t AlignInBits, unsigned Encoding,
-              DIFlags Flags, ArrayRef<Metadata *> Ops)
+              uint32_t NumExtraInhabitants, DIFlags Flags,
+              ArrayRef<Metadata *> Ops)
       : DIType(C, DIBasicTypeKind, Storage, Tag, 0, SizeInBits, AlignInBits, 0,
-               Flags, Ops),
+               NumExtraInhabitants, Flags, Ops),
         Encoding(Encoding) {}
   ~DIBasicType() = default;
 
   static DIBasicType *getImpl(LLVMContext &Context, unsigned Tag,
                               StringRef Name, uint64_t SizeInBits,
                               uint32_t AlignInBits, unsigned Encoding,
-                              DIFlags Flags, StorageType Storage,
-                              bool ShouldCreate = true) {
+                              uint32_t NumExtraInhabitants, DIFlags Flags,
+                              StorageType Storage, bool ShouldCreate = true) {
     return getImpl(Context, Tag, getCanonicalMDString(Context, Name),
-                   SizeInBits, AlignInBits, Encoding, Flags, Storage,
-                   ShouldCreate);
+                   SizeInBits, AlignInBits, Encoding, NumExtraInhabitants,
+                   Flags, Storage, ShouldCreate);
   }
   static DIBasicType *getImpl(LLVMContext &Context, unsigned Tag,
                               MDString *Name, uint64_t SizeInBits,
                               uint32_t AlignInBits, unsigned Encoding,
-                              DIFlags Flags, StorageType Storage,
-                              bool ShouldCreate = true);
+                              uint32_t NumExtraInhabitants, DIFlags Flags,
+                              StorageType Storage, bool ShouldCreate = true);
 
   TempDIBasicType cloneImpl() const {
     return getTemporary(getContext(), getTag(), getName(), getSizeInBits(),
-                        getAlignInBits(), getEncoding(), getFlags());
+                        getAlignInBits(), getEncoding(),
+                        getNumExtraInhabitants(), getFlags());
   }
 
 public:
   DEFINE_MDNODE_GET(DIBasicType, (unsigned Tag, StringRef Name),
-                    (Tag, Name, 0, 0, 0, FlagZero))
+                    (Tag, Name, 0, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, StringRef Name, uint64_t SizeInBits),
-                    (Tag, Name, SizeInBits, 0, 0, FlagZero))
+                    (Tag, Name, SizeInBits, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, MDString *Name, uint64_t SizeInBits),
-                    (Tag, Name, SizeInBits, 0, 0, FlagZero))
+                    (Tag, Name, SizeInBits, 0, 0, 0, FlagZero))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, StringRef Name, uint64_t SizeInBits,
-                     uint32_t AlignInBits, unsigned Encoding, DIFlags Flags),
-                    (Tag, Name, SizeInBits, AlignInBits, Encoding, Flags))
+                     uint32_t AlignInBits, unsigned Encoding,
+                     uint32_t NumExtraInhabitants, DIFlags Flags),
+                    (Tag, Name, SizeInBits, AlignInBits, Encoding,
+                     NumExtraInhabitants, Flags))
   DEFINE_MDNODE_GET(DIBasicType,
                     (unsigned Tag, MDString *Name, uint64_t SizeInBits,
-                     uint32_t AlignInBits, unsigned Encoding, DIFlags Flags),
-                    (Tag, Name, SizeInBits, AlignInBits, Encoding, Flags))
+                     uint32_t AlignInBits, unsigned Encoding,
+                     uint32_t NumExtraInhabitants, DIFlags Flags),
+                    (Tag, Name, SizeInBits, AlignInBits, Encoding,
+                     NumExtraInhabitants, Flags))
 
   TempDIBasicType clone() const { return cloneImpl(); }
 
@@ -890,7 +904,7 @@ class DIStringType : public DIType {
                uint64_t SizeInBits, uint32_t AlignInBits, unsigned Encoding,
                ArrayRef<Metadata *> Ops)
       : DIType(C, DIStringTypeKind, Storage, Tag, 0, SizeInBits, AlignInBits, 0,
-               FlagZero, Ops),
+               0, FlagZero, Ops),
         Encoding(Encoding) {}
   ~DIStringType() = default;
 
@@ -1016,7 +1030,7 @@ class DIDerivedType : public DIType {
                 std::optional<PtrAuthData> PtrAuthData, DIFlags Flags,
                 ArrayRef<Metadata *> Ops)
       : DIType(C, DIDerivedTypeKind, Storage, Tag, Line, SizeInBits,
-               AlignInBits, OffsetInBits, Flags, Ops),
+               AlignInBits, OffsetInBits, 0, Flags, Ops),
         DWARFAddressSpace(DWARFAddressSpace) {
     if (PtrAuthData)
       SubclassData32 = PtrAuthData->RawData;
@@ -1157,39 +1171,43 @@ class DICompositeType : public DIType {
 
   DICompositeType(LLVMContext &C, StorageType Storage, unsigned Tag,
                   unsigned Line, unsigned RuntimeLang, uint64_t SizeInBits,
-                  uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags,
+                  uint32_t AlignInBits, uint64_t OffsetInBits,
+                  uint32_t NumExtraInhabitants, DIFlags Flags,
                   ArrayRef<Metadata *> Ops)
       : DIType(C, DICompositeTypeKind, Storage, Tag, Line, SizeInBits,
-               AlignInBits, OffsetInBits, Flags, Ops),
+               AlignInBits, OffsetInBits, NumExtraInhabitants, Flags, Ops),
         RuntimeLang(RuntimeLang) {}
   ~DICompositeType() = default;
 
   /// Change fields in place.
   void mutate(unsigned Tag, unsigned Line, unsigned RuntimeLang,
               uint64_t SizeInBits, uint32_t AlignInBits, uint64_t OffsetInBits,
-              DIFlags Flags) {
+              uint32_t NumExtraInhabitants, DIFlags Flags) {
     assert(isDistinct() && "Only distinct nodes can mutate");
     assert(getRawIdentifier() && "Only ODR-uniqued nodes should mutate");
     this->RuntimeLang = RuntimeLang;
-    DIType::mutate(Tag, Line, SizeInBits, AlignInBits, OffsetInBits, Flags);
+    DIType::mutate(Tag, Line, SizeInBits, AlignInBits, OffsetInBits,
+                   NumExtraInhabitants, Flags);
   }
 
   static DICompositeType *
   getImpl(LLVMContext &Context, unsigned Tag, StringRef Name, Metadata *File,
           unsigned Line, DIScope *Scope, DIType *BaseType, uint64_t SizeInBits,
-          uint32_t AlignInBits, uint64_t OffsetInBits, DIFlags Flags,
-          DINodeArray Elements, unsigned RuntimeLang, DIType *VTableHolder,
+          uint32_t AlignInBits, uint64_t OffsetInBits,
+          uint32_t NumExtraInhabitants, DIFlags Flags, DINodeArray Elements,
+          unsigned RuntimeLang, DIType *VTableHolder,
           DITemplateParameterArray TemplateParams, StringRef Identifier,
           DIDerivedType *Discriminator, Metadata *DataLocation,
           Metadata *Associated, Metadata *Allocated, Metadata *Rank,
           DINodeArray Annotations, StorageType Storage,
           bool ShouldCreate = true) {
-    return getImpl(
-        Context, Tag, getCanonicalMDString(Context, Name), File, Line, Scope,
-        BaseType, SizeInBits, AlignInBits, OffsetInBits, Flags, Elements.get(),
-        RuntimeLang, VTableHolder, TemplateParams.get(),
-        getCanonicalMDString(Context, Identifier), Discriminator, DataLocation,
-        Associated, Allocated, Rank, Annotations.get(), Storage, ShouldCreate);
+    return getImpl(Context, Tag, getCanonicalMDString(Context, Name), File,
+                   Line, Scope, BaseType, SizeInBits, AlignInBits, OffsetInBits,
+                   Flags, Elements.get(), RuntimeLang, VTableHolder,
+                   TemplateParams.get(),
+                   getCanonicalMDString(Context, Identifier), Discriminator,
+                   DataLocation, Associated, Allocated, Rank, Annotations.get(),
+                   NumExtraInhabitants, Storage, ShouldCreate);
   }
   static DICompositeType *
   getImpl(LLVMContext &Context, unsigned Tag, MDString *Name, Metadata *File,
@@ -1199,7 +1217,8 @@ class DICompositeType : public DIType {
           Metadata *VTableHolder, Metadata *TemplateParams,
           MDString *Identifier, Metadata *Discriminator, Metadata *DataLocation,
           Metadata *Associated, Metadata *Allocated, Metadata *Rank,
-          Metadata *Annotations, StorageType Storage, bool ShouldCreate = true);
+          Metadata *Annotations, uint32_t NumExtraInhabitants,
+          StorageType Storage, bool ShouldCreate = true);
 
   TempDICompositeType cloneImpl() const {
     return getTemporary(
@@ -1208,7 +1227,7 @@ class DICompositeType : public DIType {
         getFlags(), getElements(), getRuntimeLang(), getVTableHolder(),
         getTemplateParams(), getIdentifier(), getDiscriminator(),
         getRawDataLocation(), getRawAssociated(), getRawAllocated(),
-        getRawRank(), getAnnotations());
+        getRawRank(), getAnnotations(), getNumExtraInhabitants());
   }
 
 public:
@@ -1222,11 +1241,11 @@ class DICompositeType : public DIType {
        StringRef Identifier = "", DIDerivedType *Discriminator = nullptr,
        Metadata *DataLocation = nullptr, Metadata *Associated = nullptr,
        Metadata *Allocated = nullptr, Metadata *Rank = nullptr,
-       DINodeArray Annotations = nullptr),
+       DINodeArray Annotations = nullptr, uint32_t NumExtraInhabitants = 0),
       (Tag, Name, File, Line, Scope, BaseType, SizeInBits, AlignInBits,
-       OffsetInBits, Flags, Elements, RuntimeLang, VTableHolder, TemplateParams,
-       Identifier, Discriminator, DataLocation, Associated, Allocated, Rank,
-       Annotations))
+       OffsetInBits, NumExtraInhabitants, Flags, Elements, RuntimeLang,
+       VTableHolder, TemplateParams, Identifier, Discriminator, DataLocation,
+       Associated, Allocated, Rank, Annotations))
   DEFINE_MDNODE_GET(
       DICompositeType,
       (unsigned Tag, MDString *Name, Metadata *File, unsigned Line,
@@ -1236,11 +1255,12 @@ class DICompositeType : public DIType {
        Metadata *TemplateParams = nullptr, MDString *Identifier = nullptr,
        Metadata *Discriminator = nullptr, Metadata *DataLocation = nullptr,
        Metadata *Associated = nullptr, Metadata *Allocated = nullptr,
-       Metadata *Rank = nullptr, Metadata *Annotations = nullptr),
+       Metadata *Rank = nullptr, Metadata *Annotations = nullptr,
+       uint32_t NumExtraInhabitants = 0),
       (Tag, Name, File, Line, Scope, BaseType, SizeInBits, AlignInBits,
        OffsetInBits, Flags, Elements, RuntimeLang, VTableHolder, TemplateParams,
        Identifier, Discriminator, DataLocation, Associated, Allocated, Rank,
-       Annotations))
+       Annotations, NumExtraInhabitants))
 
   TempDICompositeType clone() const { return cloneImpl(); }
 
@@ -1255,8 +1275,8 @@ class DICompositeType : public DIType {
   getODRType(LLVMContext &Context, MDString &Identifier, unsigned Tag,
              MDString *Name, Metadata *File, unsigned Line, Metadata *Scope,
              Metadata *BaseType, uint64_t SizeInBits, uint32_t AlignInBits,
-             uint64_t OffsetInBits, DIFlags Flags, Metadata *Elements,
-             unsigned RuntimeLang, Metadata *VTableHolder,
+             uint64_t OffsetInBits, uint32_t NumExtraInhabitants, DIFlags Flags,
+             Metadata *Elements, unsigned RuntimeLang, Metadata *VTableHolder,
              Metadata *TemplateParams, Metadata *Discriminator,
              Metadata *DataLocation, Metadata *Associated, Metadata *Allocated,
              Metadata *Rank, Metadata *Annotations);
@@ -1272,15 +1292,14 @@ class DICompositeType : public DIType {
   ///
   /// If not \a LLVMContext::isODRUniquingDebugTypes(), this function returns
   /// nullptr.
-  static DICompositeType *
-  buildODRType(LLVMContext &Context, MDString &Identifier, unsigned Tag,
-               MDString *Name, Metadata *File, unsigned Line, Metadata *Scope,
-               Metadata *BaseType, uint64_t SizeInBits, uint32_t AlignInBits,
-               uint64_t OffsetInBits, DIFlags Flags, Metadata *Elements,
-               unsigned RuntimeLang, Metadata *VTableHolder,
-               Metadata *TemplateParams, Metadata *Discriminator,
-               Metadata *DataLocation, Metadata *Associated,
-               Metadata *Allocated, Metadata *Rank, Metadata *Annotations);
+  static DICompositeType *buildODRType(
+      LLVMContext &Context, MDString &Identifier, unsigned Tag, MDString *Name,
+      Metadata *File, unsigned Line, Metadata *Scope, Metadata *BaseType,
+      uint64_t SizeInBits, uint32_t AlignInBits, uint64_t OffsetInBits,
+      uint32_t NumExtraInhabitants, DIFlags Flags, Metadata *Elements,
+      unsigned RuntimeLang, Metadata *VTableHolder, Metadata *TemplateParams,
+      Metadata *Discriminator, Metadata *DataLocation, Metadata *Associated,
+      Metadata *Allocated, Metadata *Rank, Metadata *Annotations);
 
   DIType *getBaseType() const { return cast_or_null<DIType>(getRawBaseType()); }
   DINodeArray getElements() const {
diff --git a/llvm/lib/AsmParser/LLParser.cpp b/llvm/lib/AsmParser/LLParser.cpp
index 5b9bddeb7cfe82..af9155636a2239 100644
--- a/llvm/lib/AsmParser/LLParser.cpp
+++ b/llvm/lib/AsmParser/LLParser.cpp
@@ -5329,12 +5329,14 @@ bool LLParser::parseDIBasicType(MDNode *&Result, bool IsDistinct) {
   OPTIONAL(size, MDUnsignedField, (0, UINT64_MAX));                            \
   OPTIONAL(align, MDUnsignedField, (0, UINT32_MAX));                           \
   OPTIONAL(encoding, DwarfAttEncodingField, );                                 \
+  OPTIONAL(num_extra_inhabitants, MDUnsignedField, (0, UINT32_MAX));           \
   OPTIONAL(flags, DIFlagField, );
   PARSE_MD_FIELDS();
 #undef VISIT_MD_FIELDS
 
   Result = GET_OR_DISTINCT(DIBasicType, (Context, tag.Val, name.Val, size.Val,
-                                         align.Val, encoding.Val, flags.Val));
+                                         align.Val, encoding.Val,
+                                         num_extra_inhabitants.Val, flags.Val));
   return false;
 }
 
@@ -5432,7 +5434,8 @@ bool LLParser::parseDICompositeType(MDNode *&Result, bool IsDistinct) {
   OPTIONAL(associated, MDField, );                                             \
   OPTIONAL(allocated, MDField, );                                              \
   OPTIONAL(rank, MDSignedOrMDField, );                                         \
-  OPTIONAL(annotations, MDField, );
+  OPTIONAL(annotations, MDField, );                                            \
+  OPTIONAL(num_extra_inhabitants, MDUnsignedField, (0, UINT32_MAX));           
   PARSE_MD_FIELDS();
 #undef VISIT_MD_FIELDS
 
@@ -5447,10 +5450,11 @@ bool LLParser::parseDICompositeType(MDNode *&Result, bool IsDistinct) {
   if (identifier.Val)
     if (auto *CT = DICompositeType::buildODRType(
             Context, *identifier.Val, tag.Val, name.Val, file.Val, line.Val,
-            scope.Val, baseType.Val, size.Val, align.Val, offset.Val, flags.Val,
-            elements.Val, runtimeLang.Val, vtableHolder.Val, templateParams.Val,
-            discriminator.Val, dataLocation.Val, associated.Val, allocated.Val,
-            Rank, annotations.Val)) {
+            scope.Val, baseType.Val, size.Val, align.Val, offset.Val,
+            num_extra_inhabitants.Val, flags.Val, elements.Val, runtimeLang.Val,
+            vtableHolder.Val, templateParams.Val, discriminator.Val,
+            dataLocation.Val, asso...
[truncated]

Copy link

github-actions bot commented Oct 16, 2024

✅ With the latest revision this PR passed the C/C++ code formatter.

Copy link
Collaborator

@pogo59 pogo59 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A count of invalid bit patterns doesn't seem that helpful to the debugger. What would it do with this information? How would it know what the actual invalid bit patterns are? For example, an enum with two enumerators with values equal to 0 and 5 would have exactly the same number of "extra inhabitants" as bool.

@augusto2112
Copy link
Contributor Author

What would it do with this information?

This information is used to calculate the layout of types. For example, since Bool has extra inhabitants, the debugger needs that information to calculate that Optional will only use 1 byte.

Since Swift generics are not monomorphized, by using dynamic libraries you can have generic types whose size, alignment, etc, are known only at runtime (which is why this feature is needed).

How would it know what the actual invalid bit patterns are? For example, an enum with two enumerators with values equal to 0 and 5 would have exactly the same number of "extra inhabitants" as bool.

This is part of the ABI of the language (which is why I added the attribute as an Apple extension). You are correct that if you have an enum with two enumerators equal to 0 and 5 you can't know which bit patterns are extra inhabitants. But for enums that don't specify what the integer representation is, two enumerators means that all bit patterns outside 0 and 1 are extra inhabitants.

@dwblaikie
Copy link
Collaborator

Can you include some of these details in the description/commit message?
(this bit, or something like it, especially: "Since Swift generics are not monomorphized, by using dynamic libraries you can have generic types whose size, alignment, etc, are known only at runtime (which is why this feature is needed).")

@augusto2112
Copy link
Contributor Author

Can you include some of these details in the description/commit message? (this bit, or something like it, especially: "Since Swift generics are not monomorphized, by using dynamic libraries you can have generic types whose size, alignment, etc, are known only at runtime (which is why this feature is needed).")

Ok, I updated the commit message.

@dwblaikie
Copy link
Collaborator

Oh, do you have a link to the ABI documentation that describes/shows how "num extra inhabitants" would be enough to know which bit patterns are available for extra inhabitants?

An extra inhabitant is a bit pattern that does not represent a valid
value for instances of a given type. The number of extra inhabitants is
the number of those bit configurations.

This is used by Swift to save space when composing types. For example,
because Bool only needs 2 bit patterns to represent all of its values
(true and false), an Optional<Bool> only occupies 1 byte in memory by
using a bit configuration that is unused by Bool. Which bit patterns are
unused are part of the ABI of the language.

Since Swift generics are not monomorphized, by using dynamic libraries
you can have generic types whose size, alignment, etc, are known only
at runtime (which is why this feature is needed).

This patch adds num_extra_inhabitants to LLVM-IR debug info and in DWARF
as an Apple extension.
@augusto2112
Copy link
Contributor Author

Oh, do you have a link to the ABI documentation that describes/shows how "num extra inhabitants" would be enough to know which bit patterns are available for extra inhabitants?

There's some documentation at https://github.com/swiftlang/swift/blob/main/docs/ABIStabilityManifesto.md and https://github.com/swiftlang/swift/blob/main/docs/ABI/TypeLayout.rst, although it doesn't seem to be complete, unfortunately.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants