[BitcodeReader] Allow reading pointer types from old IR

When opaque pointers are enabled and old IR with typed pointers is read,
the BitcodeReader automatically upgrades all typed pointers to opaque
pointers. This is a lossy conversion, i.e. when a function argument is a
pointer and unused, it’s impossible to reconstruct the original type
behind the pointer.

There are cases where the type information of pointers is needed. One is
reading DXIL, which is bitcode of old LLVM IR and makes a lot of use of
pointers in function signatures.
We’d like to keep using up-to-date llvm to read in and process DXIL, so
in the face of opaque pointers, we need some way to access the type
information of pointers from the read bitcode.

This patch allows extracting type information by supplying functions to
parseBitcodeFile that get called for each function signature or metadata
value. The function can access the type information via the reader’s
type IDs and the getTypeByID and getContainedTypeID functions.
The tests exemplarily shows how type info from pointers can be stored in
metadata for use after the BitcodeReader finished.

Differential Revision: https://reviews.llvm.org/D127728

1 files changed, 23 insertions, 11 deletions

diff --git a/llvm/lib/Bitcode/Reader/MetadataLoader.cpp b/llvm/lib/Bitcode/Reader/MetadataLoader.cpp
index fc452c2..4b5cfed 100644
--- a/llvm/lib/Bitcode/Reader/MetadataLoader.cpp
+++ b/llvm/lib/Bitcode/Reader/MetadataLoader.cpp
@@ -406,7 +406,7 @@ class MetadataLoader::MetadataLoaderImpl {
   BitstreamCursor &Stream;
   LLVMContext &Context;
   Module &TheModule;
-  std::function<Type *(unsigned)> getTypeByID;
+  MetadataLoaderCallbacks Callbacks;
 
   /// Cursor associated with the lazy-loading of Metadata. This is the easy way
   /// to keep around the right "context" (Abbrev list) to be able to jump in
@@ -627,14 +627,15 @@ class MetadataLoader::MetadataLoaderImpl {
     upgradeCUVariables();
   }
 
+  void callMDTypeCallback(Metadata **Val, unsigned TypeID);
+
 public:
   MetadataLoaderImpl(BitstreamCursor &Stream, Module &TheModule,
                      BitcodeReaderValueList &ValueList,
-                     std::function<Type *(unsigned)> getTypeByID,
-                     bool IsImporting)
+                     MetadataLoaderCallbacks Callbacks, bool IsImporting)
       : MetadataList(TheModule.getContext(), Stream.SizeInBytes()),
         ValueList(ValueList), Stream(Stream), Context(TheModule.getContext()),
-        TheModule(TheModule), getTypeByID(std::move(getTypeByID)),
+        TheModule(TheModule), Callbacks(std::move(Callbacks)),
         IsImporting(IsImporting) {}
 
   Error parseMetadata(bool ModuleLevel);
@@ -952,6 +953,14 @@ Expected<bool> MetadataLoader::MetadataLoaderImpl::loadGlobalDeclAttachments() {
   }
 }
 
+void MetadataLoader::MetadataLoaderImpl::callMDTypeCallback(Metadata **Val,
+                                                            unsigned TypeID) {
+  if (Callbacks.MDType) {
+    (*Callbacks.MDType)(Val, TypeID, Callbacks.GetTypeByID,
+                        Callbacks.GetContainedTypeID);
+  }
+}
+
 /// Parse a METADATA_BLOCK. If ModuleLevel is true then we are parsing
 /// module level metadata.
 Error MetadataLoader::MetadataLoaderImpl::parseMetadata(bool ModuleLevel) {
@@ -1221,7 +1230,7 @@ Error MetadataLoader::MetadataLoaderImpl::parseOneMetadata(
     }
 
     unsigned TyID = Record[0];
-    Type *Ty = getTypeByID(TyID);
+    Type *Ty = Callbacks.GetTypeByID(TyID);
     if (Ty->isMetadataTy() || Ty->isVoidTy()) {
       dropRecord();
       break;
@@ -1245,7 +1254,7 @@ Error MetadataLoader::MetadataLoaderImpl::parseOneMetadata(
     SmallVector<Metadata *, 8> Elts;
     for (unsigned i = 0; i != Size; i += 2) {
       unsigned TyID = Record[i];
-      Type *Ty = getTypeByID(TyID);
+      Type *Ty = Callbacks.GetTypeByID(TyID);
       if (!Ty)
         return error("Invalid record");
       if (Ty->isMetadataTy())
@@ -1255,9 +1264,10 @@ Error MetadataLoader::MetadataLoaderImpl::parseOneMetadata(
                                             /*ConstExprInsertBB*/ nullptr);
         if (!V)
           return error("Invalid value reference from old metadata");
-        auto *MD = ValueAsMetadata::get(V);
+        Metadata *MD = ValueAsMetadata::get(V);
         assert(isa<ConstantAsMetadata>(MD) &&
                "Expected non-function-local metadata");
+        callMDTypeCallback(&MD, TyID);
         Elts.push_back(MD);
       } else
         Elts.push_back(nullptr);
@@ -1271,7 +1281,7 @@ Error MetadataLoader::MetadataLoaderImpl::parseOneMetadata(
       return error("Invalid record");
 
     unsigned TyID = Record[0];
-    Type *Ty = getTypeByID(TyID);
+    Type *Ty = Callbacks.GetTypeByID(TyID);
     if (Ty->isMetadataTy() || Ty->isVoidTy())
       return error("Invalid record");
 
@@ -1280,7 +1290,9 @@ Error MetadataLoader::MetadataLoaderImpl::parseOneMetadata(
     if (!V)
       return error("Invalid value reference from metadata");
 
-    MetadataList.assignValue(ValueAsMetadata::get(V), NextMetadataNo);
+    Metadata *MD = ValueAsMetadata::get(V);
+    callMDTypeCallback(&MD, TyID);
+    MetadataList.assignValue(MD, NextMetadataNo);
     NextMetadataNo++;
     break;
   }
@@ -2359,9 +2371,9 @@ MetadataLoader::~MetadataLoader() = default;
 MetadataLoader::MetadataLoader(BitstreamCursor &Stream, Module &TheModule,
                                BitcodeReaderValueList &ValueList,
                                bool IsImporting,
-                               std::function<Type *(unsigned)> getTypeByID)
+                               MetadataLoaderCallbacks Callbacks)
     : Pimpl(std::make_unique<MetadataLoaderImpl>(
-          Stream, TheModule, ValueList, std::move(getTypeByID), IsImporting)) {}
+          Stream, TheModule, ValueList, std::move(Callbacks), IsImporting)) {}
 
 Error MetadataLoader::parseMetadata(bool ModuleLevel) {
   return Pimpl->parseMetadata(ModuleLevel);