[CodeGen] Support generation of TBAA info in the new format

[kosarev]

Summary

Now that the MDBuilder helpers generating TBAA type and access descriptors in the new format are in place, we can teach clang to use them when requested.

[rjmccall]

Tests?

[kosarev]

There are more tests that we will need to rewrite during migration to the new format than I think we could handle with a single patch. I was going to address them with separate patches. Or, I can rewrite some most basic of them as part of this patch.

[rjmccall]

Rewriting some of the most basic tests would be fine. Please either use new FileCheck lines or clone the existing tests, since we don't really know how long this transition will last.

[hfinkel]

In D41394#959715, @rjmccall wrote:

Rewriting some of the most basic tests would be fine. Please either use new FileCheck lines or clone the existing tests, since we don't really know how long this transition will last.

+1

Otherwise, this looks fine to me.

[kosarev]

Added tests.

[rjmccall]

You can pass multiple -check-prefix arguments to FileCheck and it'll match all of them. You can use that to make your test change simpler: make the existing RUN check for both PATH and OLD-PATH and the new RUN check for both PATH and NEW-PATH, then change all the existing metadata matches to OLD-PATH.

[kosarev]

Updated as suggested.

[rjmccall]

That's great, thanks. LGTM.

lib/CodeGen/CodeGenTBAA.cpp

llvm::MDNode *CodeGenTBAA::getRoot() {
   }
 
   return Root;
 }
 
 llvm::MDNode *CodeGenTBAA::createScalarTypeNode(StringRef Name,
                                                 llvm::MDNode *Parent,
                                                 uint64_t Size) {
-  (void)Size; // TODO: Support generation of size-aware type nodes.
+  if (CodeGenOpts.NewStructPathTBAA) {
+    llvm::Metadata *Id = MDHelper.createString(Name);
+    return MDHelper.createTBAATypeNode(Parent, Size, Id);
+  }
   return MDHelper.createTBAAScalarTypeNode(Name, Parent);
 }
 
 llvm::MDNode *CodeGenTBAA::getChar() {
   // Define the root of the tree for user-accessible memory. C and C++
   // give special powers to char and certain similar types. However,
   // these special powers only cover user-accessible memory, and doesn't
   // include things like vtables.
if (auto *TTy = dyn_cast<RecordType>(Ty)) {
     if (Features.CPlusPlus) {
       // Don't use the mangler for C code.
       llvm::raw_svector_ostream Out(OutName);
       MContext.mangleTypeName(QualType(Ty, 0), Out);
     } else {
       OutName = RD->getName();
     }
 
-    // TODO: Support size-aware type nodes and create one here for the
-    // given aggregate type.
+    if (CodeGenOpts.NewStructPathTBAA) {
+      llvm::MDNode *Parent = getChar();
+      uint64_t Size = Context.getTypeSizeInChars(Ty).getQuantity();
+      llvm::Metadata *Id = MDHelper.createString(OutName);
+      return MDHelper.createTBAATypeNode(Parent, Size, Id, Fields);
+    }
 
     // Create the struct type node with a vector of pairs (offset, type).
     SmallVector<std::pair<llvm::MDNode*, uint64_t>, 4> OffsetsAndTypes;
     for (const auto &Field : Fields)
         OffsetsAndTypes.push_back(std::make_pair(Field.Type, Field.Offset));
     return MDHelper.createTBAAStructTypeNode(OutName, OffsetsAndTypes);
   }
 
llvm::MDNode *CodeGenTBAA::getAccessTagInfo(TBAAAccessInfo Info) {
   llvm::MDNode *&N = AccessTagMetadataCache[Info];
   if (N)
     return N;
 
   if (!Info.BaseType) {
     Info.BaseType = Info.AccessType;
     assert(!Info.Offset && "Nonzero offset for an access with no base type!");
   }
+  if (CodeGenOpts.NewStructPathTBAA) {
+    return N = MDHelper.createTBAAAccessTag(Info.BaseType, Info.AccessType,
+                                            Info.Offset, Info.Size);
+  }
   return N = MDHelper.createTBAAStructTagNode(Info.BaseType, Info.AccessType,
                                               Info.Offset);
 }
 
 TBAAAccessInfo CodeGenTBAA::mergeTBAAInfoForCast(TBAAAccessInfo SourceInfo,
                                                  TBAAAccessInfo TargetInfo) {
   if (SourceInfo.isMayAlias() || TargetInfo.isMayAlias())
     return TBAAAccessInfo::getMayAliasInfo();

test/CodeGen/tbaa.cpp

 // RUN: %clang_cc1 -triple x86_64-apple-darwin -O1 -no-struct-path-tbaa -disable-llvm-passes %s -emit-llvm -o - | FileCheck %s
-// RUN: %clang_cc1 -triple x86_64-apple-darwin -O1 -disable-llvm-passes %s -emit-llvm -o - | FileCheck %s -check-prefix=PATH
+// RUN: %clang_cc1 -triple x86_64-apple-darwin -O1 -disable-llvm-passes %s -emit-llvm -o - | FileCheck %s -check-prefixes=PATH,OLD-PATH
+// RUN: %clang_cc1 -triple x86_64-apple-darwin -O1 -new-struct-path-tbaa -disable-llvm-passes %s -emit-llvm -o - | FileCheck %s -check-prefixes=PATH,NEW-PATH
 // RUN: %clang_cc1 -triple x86_64-apple-darwin -O0 -disable-llvm-passes %s -emit-llvm -o - | FileCheck %s -check-prefix=NO-TBAA
 // RUN: %clang_cc1 -triple x86_64-apple-darwin -O1 -relaxed-aliasing -disable-llvm-passes %s -emit-llvm -o - | FileCheck %s -check-prefix=NO-TBAA
 // Test TBAA metadata generated by front-end.
 //
 // NO-TBAA-NOT: !tbaa
 
 typedef unsigned char uint8_t;
 typedef unsigned short uint16_t;

 // CHECK: [[TYPE_char:!.*]] = !{!"omnipotent char", [[TAG_cxx_tbaa:!.*]],
 // CHECK: [[TAG_cxx_tbaa]] = !{!"Simple C++ TBAA"}
 // CHECK: [[TAG_i32]] = !{[[TYPE_i32:!.*]], [[TYPE_i32]], i64 0}
 // CHECK: [[TYPE_i32]] = !{!"int", [[TYPE_char]],
 // CHECK: [[TAG_i16]] = !{[[TYPE_i16:!.*]], [[TYPE_i16]], i64 0}
 // CHECK: [[TYPE_i16]] = !{!"short", [[TYPE_char]],
 // CHECK: [[TAG_char]] = !{[[TYPE_char]], [[TYPE_char]], i64 0}
 
-// PATH: [[TYPE_CHAR:!.*]] = !{!"omnipotent char", !
-// PATH: [[TAG_i32]] = !{[[TYPE_INT:!.*]], [[TYPE_INT]], i64 0}
-// PATH: [[TYPE_INT]] = !{!"int", [[TYPE_CHAR]]
-// PATH: [[TAG_A_f32]] = !{[[TYPE_A:!.*]], [[TYPE_INT]], i64 4}
-// PATH: [[TYPE_A]] = !{!"_ZTS7StructA", [[TYPE_SHORT:!.*]], i64 0, [[TYPE_INT]], i64 4, [[TYPE_SHORT]], i64 8, [[TYPE_INT]], i64 12}
-// PATH: [[TYPE_SHORT:!.*]] = !{!"short", [[TYPE_CHAR]]
-// PATH: [[TAG_A_f16]] = !{[[TYPE_A]], [[TYPE_SHORT]], i64 0}
-// PATH: [[TAG_B_a_f32]] = !{[[TYPE_B:!.*]], [[TYPE_INT]], i64 8}
-// PATH: [[TYPE_B]] = !{!"_ZTS7StructB", [[TYPE_SHORT]], i64 0, [[TYPE_A]], i64 4, [[TYPE_INT]], i64 20}
-// PATH: [[TAG_B_a_f16]] = !{[[TYPE_B]], [[TYPE_SHORT]], i64 4}
-// PATH: [[TAG_B_f32]] = !{[[TYPE_B]], [[TYPE_INT]], i64 20}
-// PATH: [[TAG_B_a_f32_2]] = !{[[TYPE_B]], [[TYPE_INT]], i64 16}
-// PATH: [[TAG_S_f32]] = !{[[TYPE_S:!.*]], [[TYPE_INT]], i64 4}
-// PATH: [[TYPE_S]] = !{!"_ZTS7StructS", [[TYPE_SHORT]], i64 0, [[TYPE_INT]], i64 4}
-// PATH: [[TAG_S_f16]] = !{[[TYPE_S]], [[TYPE_SHORT]], i64 0}
-// PATH: [[TAG_S2_f32]] = !{[[TYPE_S2:!.*]], [[TYPE_INT]], i64 4}
-// PATH: [[TYPE_S2]] = !{!"_ZTS8StructS2", [[TYPE_SHORT]], i64 0, [[TYPE_INT]], i64 4}
-// PATH: [[TAG_S2_f16]] = !{[[TYPE_S2]], [[TYPE_SHORT]], i64 0}
-// PATH: [[TAG_C_b_a_f32]] = !{[[TYPE_C:!.*]], [[TYPE_INT]], i64 12}
-// PATH: [[TYPE_C]] = !{!"_ZTS7StructC", [[TYPE_SHORT]], i64 0, [[TYPE_B]], i64 4, [[TYPE_INT]], i64 28}
-// PATH: [[TAG_D_b_a_f32]] = !{[[TYPE_D:!.*]], [[TYPE_INT]], i64 12}
-// PATH: [[TYPE_D]] = !{!"_ZTS7StructD", [[TYPE_SHORT]], i64 0, [[TYPE_B]], i64 4, [[TYPE_INT]], i64 28, [[TYPE_CHAR]], i64 32}
-// PATH: [[TAG_five_b]] = !{[[TYPE_five:!.*]], [[TYPE_CHAR]], i64 1}
-// PATH: [[TYPE_five]] = !{!"_ZTS4five", [[TYPE_CHAR]], i64 0, [[TYPE_INT]], i64 1, [[TYPE_CHAR]], i64 1, [[TYPE_CHAR]], i64 2}
-// PATH: [[TAG_six_b]] = !{[[TYPE_six:!.*]], [[TYPE_CHAR]], i64 4}
-// PATH: [[TYPE_six]] = !{!"_ZTS3six", [[TYPE_CHAR]], i64 0, [[TYPE_INT]], i64 4, [[TYPE_CHAR]], i64 4, [[TYPE_CHAR]], i64 5}
+// OLD-PATH: [[TYPE_CHAR:!.*]] = !{!"omnipotent char", !
+// OLD-PATH: [[TAG_i32]] = !{[[TYPE_INT:!.*]], [[TYPE_INT]], i64 0}
+// OLD-PATH: [[TYPE_INT]] = !{!"int", [[TYPE_CHAR]]
+// OLD-PATH: [[TAG_A_f32]] = !{[[TYPE_A:!.*]], [[TYPE_INT]], i64 4}
+// OLD-PATH: [[TYPE_A]] = !{!"_ZTS7StructA", [[TYPE_SHORT:!.*]], i64 0, [[TYPE_INT]], i64 4, [[TYPE_SHORT]], i64 8, [[TYPE_INT]], i64 12}
+// OLD-PATH: [[TYPE_SHORT:!.*]] = !{!"short", [[TYPE_CHAR]]
+// OLD-PATH: [[TAG_A_f16]] = !{[[TYPE_A]], [[TYPE_SHORT]], i64 0}
+// OLD-PATH: [[TAG_B_a_f32]] = !{[[TYPE_B:!.*]], [[TYPE_INT]], i64 8}
+// OLD-PATH: [[TYPE_B]] = !{!"_ZTS7StructB", [[TYPE_SHORT]], i64 0, [[TYPE_A]], i64 4, [[TYPE_INT]], i64 20}
+// OLD-PATH: [[TAG_B_a_f16]] = !{[[TYPE_B]], [[TYPE_SHORT]], i64 4}
+// OLD-PATH: [[TAG_B_f32]] = !{[[TYPE_B]], [[TYPE_INT]], i64 20}
+// OLD-PATH: [[TAG_B_a_f32_2]] = !{[[TYPE_B]], [[TYPE_INT]], i64 16}
+// OLD-PATH: [[TAG_S_f32]] = !{[[TYPE_S:!.*]], [[TYPE_INT]], i64 4}
+// OLD-PATH: [[TYPE_S]] = !{!"_ZTS7StructS", [[TYPE_SHORT]], i64 0, [[TYPE_INT]], i64 4}
+// OLD-PATH: [[TAG_S_f16]] = !{[[TYPE_S]], [[TYPE_SHORT]], i64 0}
+// OLD-PATH: [[TAG_S2_f32]] = !{[[TYPE_S2:!.*]], [[TYPE_INT]], i64 4}
+// OLD-PATH: [[TYPE_S2]] = !{!"_ZTS8StructS2", [[TYPE_SHORT]], i64 0, [[TYPE_INT]], i64 4}
+// OLD-PATH: [[TAG_S2_f16]] = !{[[TYPE_S2]], [[TYPE_SHORT]], i64 0}
+// OLD-PATH: [[TAG_C_b_a_f32]] = !{[[TYPE_C:!.*]], [[TYPE_INT]], i64 12}
+// OLD-PATH: [[TYPE_C]] = !{!"_ZTS7StructC", [[TYPE_SHORT]], i64 0, [[TYPE_B]], i64 4, [[TYPE_INT]], i64 28}
+// OLD-PATH: [[TAG_D_b_a_f32]] = !{[[TYPE_D:!.*]], [[TYPE_INT]], i64 12}
+// OLD-PATH: [[TYPE_D]] = !{!"_ZTS7StructD", [[TYPE_SHORT]], i64 0, [[TYPE_B]], i64 4, [[TYPE_INT]], i64 28, [[TYPE_CHAR]], i64 32}
+// OLD-PATH: [[TAG_five_b]] = !{[[TYPE_five:!.*]], [[TYPE_CHAR]], i64 1}
+// OLD-PATH: [[TYPE_five]] = !{!"_ZTS4five", [[TYPE_CHAR]], i64 0, [[TYPE_INT]], i64 1, [[TYPE_CHAR]], i64 1, [[TYPE_CHAR]], i64 2}
+// OLD-PATH: [[TAG_six_b]] = !{[[TYPE_six:!.*]], [[TYPE_CHAR]], i64 4}
+// OLD-PATH: [[TYPE_six]] = !{!"_ZTS3six", [[TYPE_CHAR]], i64 0, [[TYPE_INT]], i64 4, [[TYPE_CHAR]], i64 4, [[TYPE_CHAR]], i64 5}
+
+// NEW-PATH-DAG: [[ROOT:!.*]] = !{!"Simple C++ TBAA"}
+// NEW-PATH-DAG: [[TYPE_char:!.*]] = !{[[ROOT]], i64 1, !"omnipotent char"}
+// NEW-PATH-DAG: [[TYPE_short:!.*]] = !{[[TYPE_char]], i64 2, !"short"}
+// NEW-PATH-DAG: [[TYPE_int:!.*]] = !{[[TYPE_char]], i64 4, !"int"}
+// NEW-PATH-DAG: [[TAG_i32:!.*]] = !{[[TYPE_int]], [[TYPE_int]], i64 0, i64 4}
+// NEW-PATH-DAG: [[TYPE_A:!.*]] = !{[[TYPE_char]], i64 16, !"_ZTS7StructA", [[TYPE_short]], i64 0, i64 2, [[TYPE_int]], i64 4, i64 4, !12, i64 8, i64 2, [[TYPE_int]], i64 12, i64 4}
+// NEW-PATH-DAG: [[TAG_A_f16]] = !{[[TYPE_A]], [[TYPE_short]], i64 0, i64 2}
+// NEW-PATH-DAG: [[TAG_A_f32]] = !{[[TYPE_A]], [[TYPE_int]], i64 4, i64 4}
+// NEW-PATH-DAG: [[TYPE_B:!.*]] = !{[[TYPE_char]], i64 24, !"_ZTS7StructB", [[TYPE_short]], i64 0, i64 2, [[TYPE_A]], i64 4, i64 16, [[TYPE_int]], i64 20, i64 4}
+// NEW-PATH-DAG: [[TAG_B_a_f16]] = !{[[TYPE_B]], [[TYPE_short]], i64 4, i64 2}
+// NEW-PATH-DAG: [[TAG_B_a_f32]] = !{[[TYPE_B]], [[TYPE_int]], i64 8, i64 4}
+// NEW-PATH-DAG: [[TAG_B_f32]] = !{[[TYPE_B]], [[TYPE_int]], i64 20, i64 4}
+// NEW-PATH-DAG: [[TAG_B_a_f32_2]] = !{[[TYPE_B]], [[TYPE_int]], i64 16, i64 4}
+// NEW-PATH-DAG: [[TYPE_S:!.*]] = !{[[TYPE_char]], i64 8, !"_ZTS7StructS", [[TYPE_short]], i64 0, i64 2, [[TYPE_int]], i64 4, i64 4}
+// NEW-PATH-DAG: [[TAG_S_f16]] = !{[[TYPE_S]], [[TYPE_short]], i64 0, i64 2}
+// NEW-PATH-DAG: [[TAG_S_f32]] = !{[[TYPE_S]], [[TYPE_int]], i64 4, i64 4}
+// NEW-PATH-DAG: [[TYPE_S2:!.*]] = !{[[TYPE_char]], i64 8, !"_ZTS8StructS2", [[TYPE_short]], i64 0, i64 2, [[TYPE_int]], i64 4, i64 4}
+// NEW-PATH-DAG: [[TAG_S2_f16]] = !{[[TYPE_S2]], [[TYPE_short]], i64 0, i64 2}
+// NEW-PATH-DAG: [[TAG_S2_f32]] = !{[[TYPE_S2]], [[TYPE_int]], i64 4, i64 4}
+// NEW-PATH-DAG: [[TYPE_C:!.*]] = !{[[TYPE_char]], i64 32, !"_ZTS7StructC", [[TYPE_short]], i64 0, i64 2, [[TYPE_B]], i64 4, i64 24, [[TYPE_int]], i64 28, i64 4}
+// NEW-PATH-DAG: [[TAG_C_b_a_f32]] = !{[[TYPE_C]], [[TYPE_int]], i64 12, i64 4}
+// NEW-PATH-DAG: [[TYPE_D:!.*]] = !{[[TYPE_char]], i64 36, !"_ZTS7StructD", [[TYPE_short]], i64 0, i64 2, [[TYPE_B]], i64 4, i64 24, [[TYPE_int]], i64 28, i64 4, [[TYPE_char]], i64 32, i64 1}
+// NEW-PATH-DAG: [[TAG_D_b_a_f32]] = !{[[TYPE_D]], [[TYPE_int]], i64 12, i64 4}
+// NEW-PATH-DAG: [[TYPE_five:!.*]] = !{[[TYPE_char]], i64 3, !"_ZTS4five", [[TYPE_char]], i64 0, i64 1, [[TYPE_int]], i64 1, i64 4, [[TYPE_char]], i64 1, i64 1, [[TYPE_char]], i64 2, i64 1}
+// NEW-PATH-DAG: [[TAG_five_b]] = !{[[TYPE_five]], [[TYPE_char]], i64 1, i64 1}
+// NEW-PATH-DAG: [[TYPE_six:!.*]] = !{[[TYPE_char]], i64 6, !"_ZTS3six", [[TYPE_char]], i64 0, i64 1, [[TYPE_int]], i64 4, i64 4, [[TYPE_char]], i64 4, i64 1, [[TYPE_char]], i64 5, i64 1}
+// NEW-PATH-DAG: [[TAG_six_b]] = !{[[TYPE_six]], [[TYPE_char]], i64 4, i64 1}