Index: test/wasm/undefined-weak-call.ll =================================================================== --- /dev/null +++ test/wasm/undefined-weak-call.ll @@ -0,0 +1,112 @@ +; RUN: llc -filetype=obj %s -o %t.o +; RUN: wasm-ld --check-signatures --no-entry %t.o -o %t.wasm +; RUN: obj2yaml %t.wasm | FileCheck %s + +; Check that calling an undefined weak function generates an appropriate stub +; that will fail at runtime with "unreachable". + +target triple = "wasm32-unknown-unknown-wasm" + +declare extern_weak void @weakFunc1() +declare extern_weak void @weakFunc2() ; same signature +declare extern_weak void @weakFunc3(i32 %arg) ; different +declare extern_weak void @weakFunc4() ; should be GC'd as not called + +define i32 @callWeakFuncs() { + call void @weakFunc1() + call void @weakFunc2() + call void @weakFunc3(i32 2) + %addr1 = ptrtoint void ()* @weakFunc1 to i32 + %addr4 = ptrtoint void ()* @weakFunc4 to i32 + %sum = add i32 %addr1, %addr4 + ret i32 %sum +} + +; CHECK: --- !WASM +; CHECK-NEXT: FileHeader: +; CHECK-NEXT: Version: 0x00000001 +; CHECK-NEXT: Sections: +; CHECK-NEXT: - Type: TYPE +; CHECK-NEXT: Signatures: +; CHECK-NEXT: - Index: 0 +; CHECK-NEXT: ReturnType: I32 +; CHECK-NEXT: ParamTypes: +; CHECK-NEXT: - Index: 1 +; CHECK-NEXT: ReturnType: NORESULT +; CHECK-NEXT: ParamTypes: +; CHECK-NEXT: - Index: 2 +; CHECK-NEXT: ReturnType: NORESULT +; CHECK-NEXT: ParamTypes: +; CHECK-NEXT: - I32 +; CHECK-NEXT: - Type: FUNCTION +; CHECK-NEXT: FunctionTypes: [ 0, 1, 1, 2 ] +; CHECK-NEXT: - Type: TABLE +; CHECK-NEXT: Tables: +; CHECK-NEXT: - ElemType: ANYFUNC +; CHECK-NEXT: Limits: +; CHECK-NEXT: Flags: [ HAS_MAX ] +; CHECK-NEXT: Initial: 0x00000001 +; CHECK-NEXT: Maximum: 0x00000001 +; CHECK-NEXT: - Type: MEMORY +; CHECK-NEXT: Memories: +; CHECK-NEXT: - Initial: 0x00000002 +; CHECK-NEXT: - Type: GLOBAL +; CHECK-NEXT: Globals: +; CHECK-NEXT: - Index: 0 +; CHECK-NEXT: Type: I32 +; CHECK-NEXT: Mutable: true +; CHECK-NEXT: InitExpr: +; CHECK-NEXT: Opcode: I32_CONST +; CHECK-NEXT: Value: 66560 +; CHECK-NEXT: - Index: 1 +; CHECK-NEXT: Type: I32 +; CHECK-NEXT: Mutable: false +; CHECK-NEXT: InitExpr: +; CHECK-NEXT: Opcode: I32_CONST +; CHECK-NEXT: Value: 66560 +; CHECK-NEXT: - Index: 2 +; CHECK-NEXT: Type: I32 +; CHECK-NEXT: Mutable: false +; CHECK-NEXT: InitExpr: +; CHECK-NEXT: Opcode: I32_CONST +; CHECK-NEXT: Value: 1024 +; CHECK-NEXT: - Type: EXPORT +; CHECK-NEXT: Exports: +; CHECK-NEXT: - Name: memory +; CHECK-NEXT: Kind: MEMORY +; CHECK-NEXT: Index: 0 +; CHECK-NEXT: - Name: __heap_base +; CHECK-NEXT: Kind: GLOBAL +; CHECK-NEXT: Index: 1 +; CHECK-NEXT: - Name: __data_end +; CHECK-NEXT: Kind: GLOBAL +; CHECK-NEXT: Index: 2 +; CHECK-NEXT: - Name: callWeakFuncs +; CHECK-NEXT: Kind: FUNCTION +; CHECK-NEXT: Index: 0 +; CHECK-NEXT: - Type: CODE +; CHECK-NEXT: Functions: +; CHECK-NEXT: - Index: 0 +; CHECK-NEXT: Locals: +; CHECK-NEXT: Body: 10828080800010828080800041021083808080004180808080004180808080006A0B +; CHECK-NEXT: - Index: 1 +; CHECK-NEXT: Locals: +; CHECK-NEXT: Body: 0B +; CHECK-NEXT: - Index: 2 +; CHECK-NEXT: Locals: +; CHECK-NEXT: Body: 000B +; CHECK-NEXT: - Index: 3 +; CHECK-NEXT: Locals: +; CHECK-NEXT: Body: 000B +; CHECK-NEXT: - Type: CUSTOM +; CHECK-NEXT: Name: name +; CHECK-NEXT: FunctionNames: +; CHECK-NEXT: - Index: 0 +; CHECK-NEXT: Name: callWeakFuncs +; CHECK-NEXT: - Index: 1 +; CHECK-NEXT: Name: __wasm_call_ctors +; CHECK-NEXT: - Index: 2 +; CHECK-NEXT: Name: __undefined +; CHECK-NEXT: - Index: 3 +; CHECK-NEXT: Name: __undefined +; CHECK-NEXT: ... Index: wasm/Driver.cpp =================================================================== --- wasm/Driver.cpp +++ wasm/Driver.cpp @@ -215,6 +215,46 @@ return Arg->getValue(); } +static const uint8_t UnreachableFn[] = { + 0x03 /* ULEB length */, 0x00 /* ULEB num locals */, + 0x00 /* opcode unreachable */, 0x0b /* opcode end */ +}; + +// For weak undefined functions, there may be "call" instructions that reference +// the symbol. In this case, we need to synthesise a dummy/stub function that +// will abort at runtime, so that relocations can still provided an operand to +// the call instruction that passes Wasm validation. +static void handleWeakUndefines() { + DenseMap UndefinedFunctions; + for (Symbol *Sym : Symtab->getSymbols()) { + if (!Sym->isUndefined() || !Sym->isWeak()) + continue; + auto *FuncSym = dyn_cast(Sym); + if (!FuncSym) + continue; + + // It is possible for undefined functions not to have a signature (eg. if + // added via "--undefined"), but weak undefined ones do have a signature. + assert(FuncSym->getFunctionType()); + const WasmSignature &Sig = *FuncSym->getFunctionType(); + + // Add a synthetic dummy for weak undefined functions. These dummies will + // be GC'd if not used as the target of any "call" instructions. + InputFunction *&Func = UndefinedFunctions[Sig]; + if (!Func) { + Func = make(Sig, UnreachableFn, "__undefined"); + Symtab->SyntheticFunctions.emplace_back(Func); + // Ensure it compares equal to the null pointer, and so that table relocs + // don't pull in the stub body (only call-operand relocs should do that). + Func->setTableIndex(0); + } + // Hide our dummy to prevent export, and mark as local since the name is + // not unique. + uint32_t Flags = WASM_SYMBOL_VISIBILITY_HIDDEN | WASM_SYMBOL_BINDING_LOCAL; + replaceSymbol(Sym, Sym->getName(), Flags, nullptr, Func); + } +} + void LinkerDriver::link(ArrayRef ArgsArr) { WasmOptTable Parser; opt::InputArgList Args = Parser.parse(ArgsArr.slice(1)); @@ -327,6 +367,10 @@ for (InputFile *F : Files) Symtab->addFile(F); + // Add synthetic dummies for weak undefined functions. + if (!Config->Relocatable) + handleWeakUndefines(); + // Make sure we have resolved all symbols. if (!Config->Relocatable && !Config->AllowUndefined) { Symtab->reportRemainingUndefines(); Index: wasm/InputFiles.cpp =================================================================== --- wasm/InputFiles.cpp +++ wasm/InputFiles.cpp @@ -65,14 +65,8 @@ uint32_t ObjFile::calcNewValue(const WasmRelocation &Reloc) const { switch (Reloc.Type) { case R_WEBASSEMBLY_TABLE_INDEX_I32: - case R_WEBASSEMBLY_TABLE_INDEX_SLEB: { - // The null case is possible, if you take the address of a weak function - // that's simply not supplied. - FunctionSymbol *Sym = getFunctionSymbol(Reloc.Index); - if (Sym->hasTableIndex()) - return Sym->getTableIndex(); - return 0; - } + case R_WEBASSEMBLY_TABLE_INDEX_SLEB: + return getFunctionSymbol(Reloc.Index)->getTableIndex(); case R_WEBASSEMBLY_MEMORY_ADDR_SLEB: case R_WEBASSEMBLY_MEMORY_ADDR_I32: case R_WEBASSEMBLY_MEMORY_ADDR_LEB: Index: wasm/MarkLive.cpp =================================================================== --- wasm/MarkLive.cpp +++ wasm/MarkLive.cpp @@ -73,8 +73,19 @@ InputChunk *C = Q.pop_back_val(); for (const WasmRelocation Reloc : C->getRelocations()) { - if (Reloc.Type != R_WEBASSEMBLY_TYPE_INDEX_LEB) - Enqueue(C->File->getSymbol(Reloc.Index)); + if (Reloc.Type == R_WEBASSEMBLY_TYPE_INDEX_LEB) + continue; + Symbol *Sym = C->File->getSymbol(Reloc.Index); + // Don't mark functions live if we're taking the address of a function + // that won't actually go in the function table (has index zero). This + // is the case for some synthetic functions. + if (Reloc.Type == R_WEBASSEMBLY_TABLE_INDEX_SLEB || + Reloc.Type == R_WEBASSEMBLY_TABLE_INDEX_I32) { + FunctionSymbol *FuncSym = cast(Sym); + if (FuncSym->hasTableIndex() && FuncSym->getTableIndex() == 0) + continue; + } + Enqueue(Sym); } }