[dsymutil] Gather local symbol addresses in addition to global in the main executable.

[JDevlieghere]

Summary

Usually local symbols will have their address described in the debug map.
Global symbols have to have their address looked up in the symbol table
of the main executable. By playing with 'ld -r' and export lists, you
can get a symbol described as global by the debug map while actually
being a local symbol as far as the link in concerned. By gathering the
address of local symbols, we fix this issue.

Also, we prefer a global symbol in case of a name collision to preserve
the previous behavior.

Note that using the 'ld -r' tricks, people can actually cause symbol names
collisions that dsymutil has no way to figure out. This fixes the simple
case where there is only one symbol of a given name.

rdar://problem/32826621

[JDevlieghere]

ping

[friss]

LGTM

llvm/test/tools/dsymutil/X86/global_downgraded_to_static.c

+// REQUIRES : system-darwin
+// RUN: dsymutil -oso-prepend-path %p/.. -dump-debug-map %p/../Inputs/global_downgraded_to_static.x86_64 2>&1 | FileCheck %s
+//
+//  To build:
+//    clang -g -c -DFILE1 global_downgraded_to_static.c -o 1.o
+//    clang -g -c -DFILE2 global_downgraded_to_static.c -o 2.o
+//    ld -r -exported_symbol _foo 1.o -o 1.r.o
+//    clang 1.r.o 2.o -o global_downgraded_to_static.x86_64
+
+#if defined(FILE1)
+int global_to_become_static = 42;
+// CHECK: sym: _global_to_become_static,
+// CHECK-SAME: binAddr: 0x0000000100001000
+int foo() {
+  return global_to_become_static;
+}
+#elif defined(FILE2)
+int foo(void);
+int main() {
+  return foo();
+}
+#else
+#error Define FILE1 or FILE2
+#endif

llvm/tools/dsymutil/MachODebugMapParser.cpp

if (!TypeOrErr) {
       // TODO: Actually report errors helpfully.
       consumeError(TypeOrErr.takeError());
       continue;
     }
     SymbolRef::Type Type = *TypeOrErr;
     // Skip undefined and STAB entries.
     if ((Type == SymbolRef::ST_Debug) || (Type == SymbolRef::ST_Unknown))
       continue;
-    // The only symbols of interest are the global variables. These
-    // are the only ones that need to be queried because the address
-    // of common data won't be described in the debug map. All other
-    // addresses should be fetched for the debug map.
+    // In theory, the only symbols of interest are the global variables. These
+    // are the only ones that need to be queried because the address of common
+    // data won't be described in the debug map. All other addresses should be
+    // fetched for the debug map. In reality, by playing with 'ld -r' and
+    // export lists, you can get symbols described as N_GSYM in the debug map,
+    // but associated with a local symbol. Gather all the symbols, but prefer
+    // the global ones.
     uint8_t SymType =
         MainBinary.getSymbolTableEntry(Sym.getRawDataRefImpl()).n_type;
-    if (!(SymType & (MachO::N_EXT | MachO::N_PEXT)))
-      continue;
+    bool Extern = SymType & (MachO::N_EXT | MachO::N_PEXT);
     Expected<section_iterator> SectionOrErr = Sym.getSection();
     if (!SectionOrErr) {
       // TODO: Actually report errors helpfully.
       consumeError(SectionOrErr.takeError());
       continue;
     }
     Section = *SectionOrErr;
     if (Section == MainBinary.section_end() || Section->isText())
       continue;
     uint64_t Addr = Sym.getValue();
     Expected<StringRef> NameOrErr = Sym.getName();
     if (!NameOrErr) {
       // TODO: Actually report errors helpfully.
       consumeError(NameOrErr.takeError());
       continue;
     }
     StringRef Name = *NameOrErr;
     if (Name.size() == 0 || Name[0] == '\0')
       continue;
+    // Override only if the new key is global.
+    if (Extern)
       MainBinarySymbolAddresses[Name] = Addr;
+    else
+      MainBinarySymbolAddresses.try_emplace(Name, Addr);
   }
 }
 
 namespace llvm {
 namespace dsymutil {
 llvm::ErrorOr<std::vector<std::unique_ptr<DebugMap>>>
 parseDebugMap(StringRef InputFile, ArrayRef<std::string> Archs,
               StringRef PrependPath, bool PaperTrailWarnings, bool Verbose,