[lld-macho] Make it possible to re-export .tbd files

[int3]

Summary

Two things needed fixing for that to work:

1. getName() no longer returns null for DylibFiles constructed from TAPIs
2. markSubLibrary() now accepts .tbd as a possible extension

[smeenai]

LGTM

[int3]

fix ASAN

[thakis]

One of the N patches that landed 9:45 pm eastern yesterday seems to break check-lld on mac: http://45.33.8.238/mac/19510/step_10.txt

Please take a look and revert for now if it takes a while to fix.

[int3]

Hmm, the buildbot seems to be okay with it though: http://lab.llvm.org:8011/builders/lld-x86_64-darwin

[int3]

I'm not able to reproduce that failure locally either :/

lld/MachO/Driver.cpp

if (!symbol.empty()) {
         entry.anyObjectFile = std::max(entry.anyObjectFile, priority);
     }
 
     --priority;
   }
 }
 
 // We expect sub-library names of the form "libfoo", which will match a dylib
-// with a path of .*/libfoo.dylib.
+// with a path of .*/libfoo.{dylib, tbd}.
+// XXX ld64 seems to ignore the extension entirely when matching sub-libraries;
+// I'm not sure what the use case for that is.
 static bool markSubLibrary(StringRef searchName) {
   for (InputFile *file : inputFiles) {
     if (auto *dylibFile = dyn_cast<DylibFile>(file)) {
       StringRef filename = path::filename(dylibFile->getName());
-      if (filename.consume_front(searchName) && filename == ".dylib") {
+      if (filename.consume_front(searchName) &&
+          (filename == ".dylib" || filename == ".tbd")) {
         dylibFile->reexport = true;
         return true;
       }
     }
   }
   return false;
 }
 

lld/MachO/InputFiles.h

enum Kind {
     ObjKind,
     DylibKind,
     ArchiveKind,
     OpaqueKind,
   };
 
   virtual ~InputFile() = default;
   Kind kind() const { return fileKind; }
-  StringRef getName() const { return mb.getBufferIdentifier(); }
+  StringRef getName() const { return name; }
 
   MemoryBufferRef mb;
   std::vector<Symbol *> symbols;
   ArrayRef<llvm::MachO::section_64> sectionHeaders;
   std::vector<SubsectionMap> subsections;
 
 protected:
-  InputFile(Kind kind, MemoryBufferRef mb) : mb(mb), fileKind(kind) {}
+  InputFile(Kind kind, MemoryBufferRef mb)
+      : mb(mb), fileKind(kind), name(mb.getBufferIdentifier()) {}
+
+  InputFile(Kind kind, const llvm::MachO::InterfaceFile &interface)
+      : fileKind(kind), name(interface.getPath()) {}
 
   void parseSections(ArrayRef<llvm::MachO::section_64>);
 
   void parseSymbols(ArrayRef<lld::structs::nlist_64> nList, const char *strtab,
                     bool subsectionsViaSymbols);
 
   void parseRelocations(const llvm::MachO::section_64 &, SubsectionMap &);
 
 private:
   const Kind fileKind;
+  const StringRef name;
 };
 
 // .o file
 class ObjFile : public InputFile {
 public:
   explicit ObjFile(MemoryBufferRef mb);
   static bool classof(const InputFile *f) { return f->kind() == ObjKind; }
 };
 
 // command-line -sectcreate file
 class OpaqueFile : public InputFile {
 public:
   explicit OpaqueFile(MemoryBufferRef mb, StringRef segName,
                       StringRef sectName);
   static bool classof(const InputFile *f) { return f->kind() == OpaqueKind; }
 };
 
 // .dylib file
 class DylibFile : public InputFile {
 public:
-  explicit DylibFile(const llvm::MachO::InterfaceFile &interface,
-                     DylibFile *umbrella = nullptr);
-
   // Mach-O dylibs can re-export other dylibs as sub-libraries, meaning that the
   // symbols in those sub-libraries will be available under the umbrella
   // library's namespace. Those sub-libraries can also have their own
   // re-exports. When loading a re-exported dylib, `umbrella` should be set to
   // the root dylib to ensure symbols in the child library are correctly bound
   // to the root. On the other hand, if a dylib is being directly loaded
   // (through an -lfoo flag), then `umbrella` should be a nullptr.
   explicit DylibFile(MemoryBufferRef mb, DylibFile *umbrella = nullptr);
 
+  explicit DylibFile(const llvm::MachO::InterfaceFile &interface,
+                     DylibFile *umbrella = nullptr);
+
   static bool classof(const InputFile *f) { return f->kind() == DylibKind; }
 
   StringRef dylibName;
   uint64_t ordinal = 0; // Ordinal numbering starts from 1, so 0 is a sentinel
   bool reexport = false;
   std::vector<DylibFile *> reexported;
 };
 

lld/MachO/InputFiles.cpp

if (!buffer) {
       error("unable to read re-exported dylib at " + reexportPath);
       return;
     }
     reexported.push_back(make<DylibFile>(*buffer, umbrella));
   }
 }
 
 DylibFile::DylibFile(const InterfaceFile &interface, DylibFile *umbrella)
-    : InputFile(DylibKind, MemoryBufferRef()) {
+    : InputFile(DylibKind, interface) {
   if (umbrella == nullptr)
     umbrella = this;
 
   dylibName = saver.save(interface.getInstallName());
   auto addSymbol = [&](const Twine &name) -> void {
     symbols.push_back(symtab->addDylib(saver.save(name), umbrella,
                                        /*isWeakDef=*/false,
                                        /*isTlv=*/false));

lld/test/MachO/Inputs/MacOSX.sdk/usr/lib/libc++.tbd

+--- !tapi-tbd-v3
+archs:            [ i386, x86_64 ]
+uuids:            [ 'i386: 00000000-0000-0000-0000-000000000000', 'x86_64: 00000000-0000-0000-0000-0
+0000000001' ]
+platform:         macosx
+install-name:     '/usr/lib/libc++.dylib'
+current-version:  1281
+exports:
+  - archs:        [ i386, x86_64 ]
+...

lld/test/MachO/reexport-stub.s

+# REQUIRES: x86
+# RUN: mkdir -p %t
+
+## This test verifies that a non-TBD dylib can re-export a TBD library.
+
+# RUN: echo "" | llvm-mc -filetype=obj -triple=x86_64-apple-darwin -o %t/reexporter.o
+# RUN: lld -flavor darwinnew -dylib -syslibroot %S/Inputs/MacOSX.sdk -lc++ -sub_library libc++ \
+# RUN:   %t/reexporter.o -o %t/libreexporter.dylib
+# RUN: llvm-objdump --macho --all-headers %t/libreexporter.dylib | FileCheck %s --check-prefix=DYLIB-HEADERS
+# DYLIB-HEADERS:     cmd     LC_REEXPORT_DYLIB
+# DYLIB-HEADERS-NOT: Load command
+# DYLIB-HEADERS:     name    /usr/lib/libc++.dylib