[lld-macho][re-land] Warn on method name collisions from category definitions

This implements ld64's checks for duplicate method names in categories &
classes.

In addition, this sets us up for implementing Obj-C category merging.
This diff handles the most of the parsing work; what's left is rewriting
those category / class structures.

Numbers for chromium_framework:

             base           diff           difference (95% CI)
  sys_time   2.182 ± 0.027  2.200 ± 0.047  [  -0.2% ..   +1.8%]
  user_time  6.451 ± 0.034  6.479 ± 0.062  [  -0.0% ..   +0.9%]
  wall_time  6.841 ± 0.048  6.885 ± 0.105  [  -0.1% ..   +1.4%]
  samples    33             22

Fixes llvm#54912.

Issues seen with the previous land will be fixed in the next commit.

Reviewed By: #lld-macho, thevinster, oontvoo

Differential Revision: https://reviews.llvm.org/D142916

Author: int3

lld/MachO/Driver.cpp

@@ -1921,6 +1921,8 @@ bool macho::link(ArrayRef<const char *> argsArr, llvm::raw_ostream &stdoutOS,
     if (config->deadStrip)
       markLive();
 
+    objc::checkCategories();
+
     // ICF assumes that all literals have been folded already, so we must run
     // foldIdenticalLiterals before foldIdenticalSections.
     foldIdenticalLiterals();

lld/MachO/InputFiles.cpp

@@ -1254,13 +1254,10 @@ static CIE parseCIE(const InputSection *isec, const EhReader &reader,
     }
   }
   if (personalityAddrOff != 0) {
-    auto personalityRelocIt =
-        llvm::find_if(isec->relocs, [=](const macho::Reloc &r) {
-          return r.offset == personalityAddrOff;
-        });
-    if (personalityRelocIt == isec->relocs.end())
+    const auto *personalityReloc = isec->getRelocAt(personalityAddrOff);
+    if (!personalityReloc)
       reader.failOn(off, "Failed to locate relocation for personality symbol");
-    cie.personalitySymbol = personalityRelocIt->referent.get<macho::Symbol *>();
+    cie.personalitySymbol = personalityReloc->referent.get<macho::Symbol *>();
   }
   return cie;
 }

lld/MachO/InputSection.cpp

@@ -135,6 +135,14 @@ std::string InputSection::getSourceLocation(uint64_t off) const {
   return {};
 }
 
+const Reloc *InputSection::getRelocAt(uint32_t off) const {
+  auto it = llvm::find_if(
+      relocs, [=](const macho::Reloc &r) { return r.offset == off; });
+  if (it == relocs.end())
+    return nullptr;
+  return &*it;
+}
+
 void ConcatInputSection::foldIdentical(ConcatInputSection *copy) {
   align = std::max(align, copy->align);
   copy->live = false;
@@ -259,6 +267,15 @@ const StringPiece &CStringInputSection::getStringPiece(uint64_t off) const {
   return const_cast<CStringInputSection *>(this)->getStringPiece(off);
 }
 
+size_t CStringInputSection::getStringPieceIndex(uint64_t off) const {
+  if (off >= data.size())
+    fatal(toString(this) + ": offset is outside the section");
+
+  auto it =
+      partition_point(pieces, [=](StringPiece p) { return p.inSecOff <= off; });
+  return std::distance(pieces.begin(), it) - 1;
+}
+
 uint64_t CStringInputSection::getOffset(uint64_t off) const {
   const StringPiece &piece = getStringPiece(off);
   uint64_t addend = off - piece.inSecOff;

lld/MachO/InputSection.h

@@ -55,6 +55,8 @@ class InputSection {
   // Return the source line corresponding to an address, or the empty string.
   // Format: Source.cpp:123 (/path/to/Source.cpp:123)
   std::string getSourceLocation(uint64_t off) const;
+  // Return the relocation at \p off, if it exists. This does a linear search.
+  const Reloc *getRelocAt(uint32_t off) const;
   // Whether the data at \p off in this InputSection is live.
   virtual bool isLive(uint64_t off) const = 0;
   virtual void markLive(uint64_t off) = 0;
@@ -218,6 +220,10 @@ class CStringInputSection final : public InputSection {
     return toStringRef(data.slice(begin, end - begin));
   }
 
+  StringRef getStringRefAtOffset(uint64_t off) const {
+    return getStringRef(getStringPieceIndex(off));
+  }
+
   // Returns i'th piece as a CachedHashStringRef. This function is very hot when
   // string merging is enabled, so we want to inline.
   LLVM_ATTRIBUTE_ALWAYS_INLINE
@@ -232,6 +238,9 @@ class CStringInputSection final : public InputSection {
 
   bool deduplicateLiterals = false;
   std::vector<StringPiece> pieces;
+
+private:
+  size_t getStringPieceIndex(uint64_t off) const;
 };
 
 class WordLiteralInputSection final : public InputSection {

lld/MachO/Layout.h

@@ -0,0 +1,74 @@
+//===- Layout.h -----------------------------------------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+
+// Convenience macros for obtaining offsets of members in structs.
+//
+// Usage:
+//
+//   #define FOR_EACH_FOO_FIELD(DO) \
+//     DO(Ptr, bar)                 \
+//     DO(uint32_t, baz)            \
+//   CREATE_LAYOUT_CLASS(Foo, FOR_EACH_FOO_FIELD)
+//   #undef FOR_EACH_FOO_FIELD
+//
+// This will generate
+//
+//   struct FooLayout {
+//     uint32_t barOffset;
+//     uint32_t bazOffset;
+//     uint32_t totalSize;
+//
+//     FooLayout(size_t wordSize) {
+//       if (wordSize == 8)
+//         init<uint64_t>();
+//       else {
+//         assert(wordSize == 4);
+//         init<uint32_t>();
+//       }
+//     }
+//
+//   private:
+//     template <class Ptr> void init() {
+//       FOR_EACH_FIELD(_INIT_OFFSET);
+//       barOffset = offsetof(Layout<Ptr>, bar);
+//       bazOffset = offsetof(Layout<Ptr>, baz);
+//       totalSize = sizeof(Layout<Ptr>);
+//     }
+//     template <class Ptr> struct Layout {
+//       Ptr bar;
+//       uint32_t baz;
+//     };
+//   };
+
+#define _OFFSET_FOR_FIELD(_, name) uint32_t name##Offset;
+#define _INIT_OFFSET(type, name) name##Offset = offsetof(Layout<Ptr>, name);
+#define _LAYOUT_ENTRY(type, name) type name;
+
+#define CREATE_LAYOUT_CLASS(className, FOR_EACH_FIELD)                         \
+  struct className##Layout {                                                   \
+    FOR_EACH_FIELD(_OFFSET_FOR_FIELD)                                          \
+    uint32_t totalSize;                                                        \
+                                                                               \
+    className##Layout(size_t wordSize) {                                       \
+      if (wordSize == 8)                                                       \
+        init<uint64_t>();                                                      \
+      else {                                                                   \
+        assert(wordSize == 4);                                                 \
+        init<uint32_t>();                                                      \
+      }                                                                        \
+    }                                                                          \
+                                                                               \
+  private:                                                                     \
+    template <class Ptr> void init() {                                         \
+      FOR_EACH_FIELD(_INIT_OFFSET);                                            \
+      totalSize = sizeof(Layout<Ptr>);                                         \
+    }                                                                          \
+    template <class Ptr> struct Layout {                                       \
+      FOR_EACH_FIELD(_LAYOUT_ENTRY)                                            \
+    };                                                                         \
+  }