[lld][WebAssembly] Process stub libraries in a loop

[sbc100]

Summary

When stub libraries trigger the fetching of new object files we can
potentially introduce new undefined symbols so process the stub in
loop until no new objects are pulled in.

[sbc100]

• add test

[kripken]

lgtm % question

lld/wasm/Driver.cpp

913

I'm not that familiar with the LLVM codebase - is this a common pattern? What does message do?

[sbc100]

lld/wasm/Driver.cpp

913

This is the way we write to stderr for messages to that are designed to end users to see.

In this case we print some extra info about a symbol if the user passes --trace=<symname>

[kripken]

lld/wasm/Driver.cpp

913

Ah, I see, thanks! Makes sense.

lld/test/wasm/Inputs/libstub.so

 #STUB
 # This is a comment
 foo: foodep1,foodep2
 # This symbols as no dependencies
 bar
+baz: bazdep

lld/test/wasm/stub-library-archive.s

+# RUN: split-file %s %t
+# RUN: llvm-mc -filetype=obj -triple=wasm32-unknown-unknown -o %t.o %t/main.s
+# RUN: llvm-mc -filetype=obj -triple=wasm32-unknown-unknown -o %t/foodeps.o %t/foodeps.s
+# RUN: rm -f %t/libfoodeps.a
+# RUN: llvm-ar rcs %t/libfoodeps.a %t/foodeps.o
+# RUN: wasm-ld %t.o %p/Inputs/libstub.so %t/libfoodeps.a -o %t.wasm
+# RUN: obj2yaml %t.wasm | FileCheck %s
+
+#--- main.s
+
+# The function foo is defined in libstub.so but depends on foodep1 and foodep2
+
+# foodep1 and foodep2 a defined libfoodeps.a(foodeps.o) but this function
+# depeds on baz which is also defined in libstub.so.
+
+.functype foo () -> ()
+
+.globl _start
+_start:
+    .functype _start () -> ()
+    call foo
+    end_function
+
+.globl bazdep
+bazdep:
+    .functype bazdep () -> ()
+    end_function
+
+#--- foodeps.s
+
+.functype baz () -> ()
+
+.globl foodep1
+foodep1:
+  .functype foodep1 () -> ()
+  call baz
+  end_function
+
+.globl foodep2
+foodep2:
+  .functype foodep2 () -> ()
+  end_function
+
+# CHECK:       - Type:            EXPORT
+# CHECK-NEXT:    Exports:
+# CHECK-NEXT:      - Name:            memory
+# CHECK-NEXT:        Kind:            MEMORY
+# CHECK-NEXT:        Index:           0
+# CHECK-NEXT:      - Name:            _start
+# CHECK-NEXT:        Kind:            FUNCTION
+# CHECK-NEXT:        Index:           2
+# CHECK-NEXT:      - Name:            bazdep
+# CHECK-NEXT:        Kind:            FUNCTION
+# CHECK-NEXT:        Index:           3
+# CHECK-NEXT:      - Name:            foodep1
+# CHECK-NEXT:        Kind:            FUNCTION
+# CHECK-NEXT:        Index:           4
+# CHECK-NEXT:      - Name:            foodep2
+# CHECK-NEXT:        Kind:            FUNCTION
+# CHECK-NEXT:        Index:           5

lld/wasm/Driver.cpp

for (auto [name, deps]: stub_file->symbolDependencies) {
         }
       }
     }
   }
 }
 
 static void processStubLibraries() {
   log("-- processStubLibraries");
+  bool depsAdded = false;
+  do {
+    depsAdded = false;
     for (auto &stub_file : symtab->stubFiles) {
       LLVM_DEBUG(llvm::dbgs()
                  << "processing stub file: " << stub_file->getName() << "\n");
       for (auto [name, deps]: stub_file->symbolDependencies) {
         auto* sym = symtab->find(name);
         if (!sym || !sym->isUndefined()) {
-        LLVM_DEBUG(llvm::dbgs() << "stub symbol not needed: " << name << "\n");
+          if (sym && sym->traced)
+            message(toString(stub_file) + ": stub symbol not needed: " + name);
+          else
+            LLVM_DEBUG(llvm::dbgs() << "stub symbol not needed: `" << name << "`\n");
           continue;
         }
         // The first stub library to define a given symbol sets this and
         // definitions in later stub libraries are ignored.
         if (sym->forceImport)
           continue;  // Already handled
         sym->forceImport = true;
         if (sym->traced)
           message(toString(stub_file) + ": importing " + name);
         else
           LLVM_DEBUG(llvm::dbgs()
                      << toString(stub_file) << ": importing " << name << "\n");
         for (const auto dep : deps) {
           auto* needed = symtab->find(dep);
           if (!needed) {
             error(toString(stub_file) + ": undefined symbol: " + dep +
                   ". Required by " + toString(*sym));
           } else if (needed->isUndefined()) {
             error(toString(stub_file) +
                   ": undefined symbol: " + toString(*needed) +
                   ". Required by " + toString(*sym));
           } else {
             if (needed->traced)
               message(toString(stub_file) + ": exported " + toString(*needed) +
                       " due to import of " + name);
             else
               LLVM_DEBUG(llvm::dbgs()
                          << "force export: " << toString(*needed) << "\n");
             needed->forceExport = true;
             if (auto *lazy = dyn_cast<LazySymbol>(needed)) {
+              depsAdded = true;
               lazy->fetch();
               if (!config->whyExtract.empty())
                 config->whyExtractRecords.emplace_back(stub_file->getName(),
                                                        sym->getFile(), *sym);
             }
           }
         }
       }
     }
+  } while (depsAdded);
+
   log("-- done processStubLibraries");
 }
 
 // Reconstructs command line arguments so that so that you can re-run
 // the same command with the same inputs. This is for --reproduce.
 static std::string createResponseFile(const opt::InputArgList &args) {
   SmallString<0> data;
   raw_svector_ostream os(data);