cgo: only include the symbols that are necessary (recursively)

Only try to convert the C symbols to their Go equivalents that are
actually referenced by the Go code with C.<somesymbol>. This avoids
having to support all possible C types, which is difficult because of
oddities like `typedef void` or `__builtin_va_list`. Especially
__builtin_va_list, which varies between targets.

loader/cgo.go

@@ -23,6 +23,7 @@ type fileInfo struct {
 	typedefs        map[string]*typedefInfo
 	elaboratedTypes map[string]ast.Expr
 	importCPos      token.Pos
+	missingSymbols  map[string]struct{}
 }
 
 // functionInfo stores some information about a Cgo function found by libclang
@@ -60,7 +61,22 @@ var cgoAliases = map[string]string{
 	"C.uint32_t":  "uint32",
 	"C.uint64_t":  "uint64",
 	"C.uintptr_t": "uintptr",
-	"C.__builtin_va_list": "uintptr", // dummy value until fully implemented
+}
+
+// cgoBuiltinAliases are handled specially because they only exist on the Go
+// side of CGo, not on the CGo (they're prefixed with "_Cgo_" there).
+var cgoBuiltinAliases = map[string]struct{}{
+	"char":      struct{}{},
+	"schar":     struct{}{},
+	"uchar":     struct{}{},
+	"short":     struct{}{},
+	"ushort":    struct{}{},
+	"int":       struct{}{},
+	"uint":      struct{}{},
+	"long":      struct{}{},
+	"ulong":     struct{}{},
+	"longlong":  struct{}{},
+	"ulonglong": struct{}{},
 }
 
 // cgoTypes lists some C types with ambiguous sizes that must be retrieved
@@ -91,6 +107,13 @@ func (p *Package) processCgo(filename string, f *ast.File, cflags []string) []er
 		globals:         map[string]*globalInfo{},
 		typedefs:        map[string]*typedefInfo{},
 		elaboratedTypes: map[string]ast.Expr{},
+		missingSymbols:  map[string]struct{}{},
+	}
+
+	// Find all C.* symbols.
+	f = astutil.Apply(f, info.findMissingCGoNames, nil).(*ast.File)
+	for name := range cgoBuiltinAliases {
+		info.missingSymbols["_Cgo_"+name] = struct{}{}
 	}
 
 	// Find `import "C"` statements in the file.
@@ -222,6 +245,9 @@ func (info *fileInfo) addFuncDecls() {
 //         // ...
 //     )
 func (info *fileInfo) addFuncPtrDecls() {
+	if len(info.functions) == 0 {
+		return
+	}
 	gen := &ast.GenDecl{
 		TokPos: info.importCPos,
 		Tok:    token.VAR,
@@ -270,6 +296,9 @@ func (info *fileInfo) addFuncPtrDecls() {
 //         // ...
 //     )
 func (info *fileInfo) addVarDecls() {
+	if len(info.globals) == 0 {
+		return
+	}
 	gen := &ast.GenDecl{
 		TokPos: info.importCPos,
 		Tok:    token.VAR,
@@ -346,6 +375,9 @@ func (info *fileInfo) addTypeAliases() {
 }
 
 func (info *fileInfo) addTypedefs() {
+	if len(info.typedefs) == 0 {
+		return
+	}
 	gen := &ast.GenDecl{
 		TokPos: info.importCPos,
 		Tok:    token.TYPE,
@@ -394,6 +426,9 @@ func (info *fileInfo) addTypedefs() {
 // See also:
 // https://en.cppreference.com/w/cpp/language/elaborated_type_specifier
 func (info *fileInfo) addElaboratedTypes() {
+	if len(info.elaboratedTypes) == 0 {
+		return
+	}
 	gen := &ast.GenDecl{
 		TokPos: info.importCPos,
 		Tok:    token.TYPE,
@@ -424,6 +459,27 @@ func (info *fileInfo) addElaboratedTypes() {
 	info.Decls = append(info.Decls, gen)
 }
 
+// findMissingCGoNames traverses the AST and finds all C.something names. Only
+// these symbols are extracted from the parsed C AST and converted to the Go
+// equivalent.
+func (info *fileInfo) findMissingCGoNames(cursor *astutil.Cursor) bool {
+	switch node := cursor.Node().(type) {
+	case *ast.SelectorExpr:
+		x, ok := node.X.(*ast.Ident)
+		if !ok {
+			return true
+		}
+		if x.Name == "C" {
+			name := node.Sel.Name
+			if _, ok := cgoBuiltinAliases[name]; ok {
+				name = "_Cgo_" + name
+			}
+			info.missingSymbols[name] = struct{}{}
+		}
+	}
+	return true
+}
+
 // walker replaces all "C".<something> expressions to literal "C.<something>"
 // expressions. Such expressions are impossible to write in Go (a dot cannot be
 // used in the middle of a name) so in practice all C identifiers live in a

loader/libclang.go

@@ -168,6 +168,9 @@ func tinygo_clang_globals_visitor(c, parent C.GoCXCursor, client_data C.CXClient
 	switch kind {
 	case C.CXCursor_FunctionDecl:
 		name := getString(C.tinygo_clang_getCursorSpelling(c))
+		if _, required := info.missingSymbols[name]; !required {
+			return C.CXChildVisit_Continue
+		}
 		cursorType := C.tinygo_clang_getCursorType(c)
 		if C.clang_isFunctionTypeVariadic(cursorType) != 0 {
 			return C.CXChildVisit_Continue // not supported
@@ -199,58 +202,23 @@ func tinygo_clang_globals_visitor(c, parent C.GoCXCursor, client_data C.CXClient
 		}
 	case C.CXCursor_StructDecl:
 		typ := C.tinygo_clang_getCursorType(c)
+		name := getString(C.tinygo_clang_getCursorSpelling(c))
+		if _, required := info.missingSymbols["struct_"+name]; !required {
+			return C.CXChildVisit_Continue
+		}
 		info.makeASTType(typ, pos)
 	case C.CXCursor_TypedefDecl:
 		typedefType := C.tinygo_clang_getCursorType(c)
 		name := getString(C.clang_getTypedefName(typedefType))
-		underlyingType := C.tinygo_clang_getTypedefDeclUnderlyingType(c)
-		expr := info.makeASTType(underlyingType, pos)
-		if strings.HasPrefix(name, "_Cgo_") {
-			expr := expr.(*ast.Ident)
-			typeSize := C.clang_Type_getSizeOf(underlyingType)
-			switch expr.Name {
-			case "C.char":
-				if typeSize != 1 {
-					// This happens for some very special purpose architectures
-					// (DSPs etc.) that are not currently targeted.
-					// https://www.embecosm.com/2017/04/18/non-8-bit-char-support-in-clang-and-llvm/
-					panic("unknown char width")
-				}
-				switch underlyingType.kind {
-				case C.CXType_Char_S:
-					expr.Name = "int8"
-				case C.CXType_Char_U:
-					expr.Name = "uint8"
-				}
-			case "C.schar", "C.short", "C.int", "C.long", "C.longlong":
-				switch typeSize {
-				case 1:
-					expr.Name = "int8"
-				case 2:
-					expr.Name = "int16"
-				case 4:
-					expr.Name = "int32"
-				case 8:
-					expr.Name = "int64"
-				}
-			case "C.uchar", "C.ushort", "C.uint", "C.ulong", "C.ulonglong":
-				switch typeSize {
-				case 1:
-					expr.Name = "uint8"
-				case 2:
-					expr.Name = "uint16"
-				case 4:
-					expr.Name = "uint32"
-				case 8:
-					expr.Name = "uint64"
-				}
-			}
-		}
-		info.typedefs[name] = &typedefInfo{
-			typeExpr: expr,
+		if _, required := info.missingSymbols[name]; !required {
+			return C.CXChildVisit_Continue
 		}
+		info.makeASTType(typedefType, pos)
 	case C.CXCursor_VarDecl:
 		name := getString(C.tinygo_clang_getCursorSpelling(c))
+		if _, required := info.missingSymbols[name]; !required {
+			return C.CXChildVisit_Continue
+		}
 		cursorType := C.tinygo_clang_getCursorType(c)
 		info.globals[name] = &globalInfo{
 			typeExpr: info.makeASTType(cursorType, pos),
@@ -394,10 +362,60 @@ func (info *fileInfo) makeASTType(typ C.CXType, pos token.Pos) ast.Expr {
 			},
 		}
 	case C.CXType_Typedef:
-		typedefName := getString(C.clang_getTypedefName(typ))
+		name := getString(C.clang_getTypedefName(typ))
+		if _, ok := info.typedefs[name]; !ok {
+			info.typedefs[name] = nil // don't recurse
+			c := C.tinygo_clang_getTypeDeclaration(typ)
+			underlyingType := C.tinygo_clang_getTypedefDeclUnderlyingType(c)
+			expr := info.makeASTType(underlyingType, pos)
+			if strings.HasPrefix(name, "_Cgo_") {
+				expr := expr.(*ast.Ident)
+				typeSize := C.clang_Type_getSizeOf(underlyingType)
+				switch expr.Name {
+				case "C.char":
+					if typeSize != 1 {
+						// This happens for some very special purpose architectures
+						// (DSPs etc.) that are not currently targeted.
+						// https://www.embecosm.com/2017/04/18/non-8-bit-char-support-in-clang-and-llvm/
+						panic("unknown char width")
+					}
+					switch underlyingType.kind {
+					case C.CXType_Char_S:
+						expr.Name = "int8"
+					case C.CXType_Char_U:
+						expr.Name = "uint8"
+					}
+				case "C.schar", "C.short", "C.int", "C.long", "C.longlong":
+					switch typeSize {
+					case 1:
+						expr.Name = "int8"
+					case 2:
+						expr.Name = "int16"
+					case 4:
+						expr.Name = "int32"
+					case 8:
+						expr.Name = "int64"
+					}
+				case "C.uchar", "C.ushort", "C.uint", "C.ulong", "C.ulonglong":
+					switch typeSize {
+					case 1:
+						expr.Name = "uint8"
+					case 2:
+						expr.Name = "uint16"
+					case 4:
+						expr.Name = "uint32"
+					case 8:
+						expr.Name = "uint64"
+					}
+				}
+			}
+			info.typedefs[name] = &typedefInfo{
+				typeExpr: expr,
+			}
+		}
 		return &ast.Ident{
 			NamePos: pos,
-			Name:    "C." + typedefName,
+			Name:    "C." + name,
 		}
 	case C.CXType_Elaborated:
 		underlying := C.clang_Type_getNamedType(typ)

testdata/cgo/main.h

@@ -2,12 +2,17 @@
 #include <stdint.h>
 
 typedef short myint;
+typedef short unusedTypedef;
 int add(int a, int b);
+int unusedFunction(void);
 typedef int (*binop_t) (int, int);
 int doCallback(int a, int b, binop_t cb);
 typedef int * intPointer;
 void store(int value, int *ptr);
 
+// this signature should not be included by CGo
+void unusedFunction2(int x, __builtin_va_list args);
+
 typedef struct collection {
 	short         s;
 	long          l;
@@ -37,6 +42,7 @@ void unionSetData(short f0, short f1, short f2);
 
 // test globals and datatypes
 extern int global;
+extern int unusedGlobal;
 extern bool globalBool;
 extern bool globalBool2;
 extern float globalFloat;