cmd/asm: allow def/ref of func ABI when compiling runtime

Function symbols defined and referenced by assembly source currently
always default to ABI0; this patch adds preliminary support for
accepting an explicit ABI selector clause for func defs/refs. This
functionality is currently only enabled when compiling runtime-related
packages (runtime, syscall, reflect). Examples:

  TEXT ·DefinedAbi0Symbol<ABI0>(SB),NOSPLIT,$0
        RET

  TEXT ·DefinedAbi1Symbol<ABIInternal>(SB),NOSPLIT,$0
        CALL    ·AbiZerolSym<ABI0>(SB)
	...
        JMP     ·AbiInternalSym<ABIInternal>(SB)
        RET

Also included is a small change to the code in the compiler that reads
the symabis file emitted by the assembler.

New behavior is currently gated under GOEXPERIMENT=regabi.

Updates #27539, #40724.

Change-Id: Ia22221fe26df0fa002191cfb13bdfaaa38d7df38
Reviewed-on: https://go-review.googlesource.com/c/go/+/260477
Run-TryBot: Than McIntosh <thanm@google.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>
Trust: Than McIntosh <thanm@google.com>

src/cmd/asm/internal/asm/endtoend_test.go

@@ -31,7 +31,7 @@ func testEndToEnd(t *testing.T, goarch, file string) {
 	architecture, ctxt := setArch(goarch)
 	architecture.Init(ctxt)
 	lexer := lex.NewLexer(input)
-	parser := NewParser(ctxt, architecture, lexer)
+	parser := NewParser(ctxt, architecture, lexer, false)
 	pList := new(obj.Plist)
 	var ok bool
 	testOut = new(bytes.Buffer) // The assembler writes test output to this buffer.
@@ -273,7 +273,7 @@ func testErrors(t *testing.T, goarch, file string) {
 	input := filepath.Join("testdata", file+".s")
 	architecture, ctxt := setArch(goarch)
 	lexer := lex.NewLexer(input)
-	parser := NewParser(ctxt, architecture, lexer)
+	parser := NewParser(ctxt, architecture, lexer, false)
 	pList := new(obj.Plist)
 	var ok bool
 	testOut = new(bytes.Buffer) // The assembler writes test output to this buffer.

src/cmd/asm/internal/asm/expr_test.go

@@ -57,7 +57,7 @@ var exprTests = []exprTest{
 }
 
 func TestExpr(t *testing.T) {
-	p := NewParser(nil, nil, nil) // Expression evaluation uses none of these fields of the parser.
+	p := NewParser(nil, nil, nil, false) // Expression evaluation uses none of these fields of the parser.
 	for i, test := range exprTests {
 		p.start(lex.Tokenize(test.input))
 		result := int64(p.expr())
@@ -113,7 +113,7 @@ func TestBadExpr(t *testing.T) {
 }
 
 func runBadTest(i int, test badExprTest, t *testing.T) (err error) {
-	p := NewParser(nil, nil, nil) // Expression evaluation uses none of these fields of the parser.
+	p := NewParser(nil, nil, nil, false) // Expression evaluation uses none of these fields of the parser.
 	p.start(lex.Tokenize(test.input))
 	return tryParse(t, func() {
 		p.expr()

src/cmd/asm/internal/asm/line_test.go

@@ -39,7 +39,7 @@ func testBadInstParser(t *testing.T, goarch string, tests []badInstTest) {
 	for i, test := range tests {
 		arch, ctxt := setArch(goarch)
 		tokenizer := lex.NewTokenizer("", strings.NewReader(test.input+"\n"), nil)
-		parser := NewParser(ctxt, arch, tokenizer)
+		parser := NewParser(ctxt, arch, tokenizer, false)
 
 		err := tryParse(t, func() {
 			parser.Parse()

src/cmd/asm/internal/asm/operand_test.go

@@ -28,7 +28,7 @@ func setArch(goarch string) (*arch.Arch, *obj.Link) {
 
 func newParser(goarch string) *Parser {
 	architecture, ctxt := setArch(goarch)
-	return NewParser(ctxt, architecture, nil)
+	return NewParser(ctxt, architecture, nil, false)
 }
 
 // tryParse executes parse func in panicOnError=true context.
@@ -75,7 +75,12 @@ func testOperandParser(t *testing.T, parser *Parser, tests []operandTest) {
 		parser.start(lex.Tokenize(test.input))
 		addr := obj.Addr{}
 		parser.operand(&addr)
-		result := obj.Dconv(&emptyProg, &addr)
+		var result string
+		if parser.compilingRuntime {
+			result = obj.DconvWithABIDetail(&emptyProg, &addr)
+		} else {
+			result = obj.Dconv(&emptyProg, &addr)
+		}
 		if result != test.output {
 			t.Errorf("fail at %s: got %s; expected %s\n", test.input, result, test.output)
 		}
@@ -86,6 +91,9 @@ func TestAMD64OperandParser(t *testing.T) {
 	parser := newParser("amd64")
 	testOperandParser(t, parser, amd64OperandTests)
 	testBadOperandParser(t, parser, amd64BadOperandTests)
+	parser.compilingRuntime = true
+	testOperandParser(t, parser, amd64RuntimeOperandTests)
+	testBadOperandParser(t, parser, amd64BadOperandRuntimeTests)
 }
 
 func Test386OperandParser(t *testing.T) {
@@ -141,7 +149,7 @@ func TestFuncAddress(t *testing.T) {
 			parser := newParser(sub.arch)
 			for _, test := range sub.tests {
 				parser.start(lex.Tokenize(test.input))
-				name, ok := parser.funcAddress()
+				name, _, ok := parser.funcAddress()
 
 				isFuncSym := strings.HasSuffix(test.input, "(SB)") &&
 					// Ignore static symbols.
@@ -298,6 +306,11 @@ var amd64OperandTests = []operandTest{
 	{"[):[o-FP", ""}, // Issue 12469 - asm hung parsing the o-FP range on non ARM platforms.
 }
 
+var amd64RuntimeOperandTests = []operandTest{
+	{"$bar<ABI0>(SB)", "$bar<ABI0>(SB)"},
+	{"$foo<ABIInternal>(SB)", "$foo<ABIInternal>(SB)"},
+}
+
 var amd64BadOperandTests = []badOperandTest{
 	{"[", "register list: expected ']', found EOF"},
 	{"[4", "register list: bad low register in `[4`"},
@@ -311,6 +324,11 @@ var amd64BadOperandTests = []badOperandTest{
 	{"[X0-X1-X2]", "register list: expected ']' after `[X0-X1`, found '-'"},
 	{"[X0,X3]", "register list: expected '-' after `[X0`, found ','"},
 	{"[X0,X1,X2,X3]", "register list: expected '-' after `[X0`, found ','"},
+	{"$foo<ABI0>", "ABI selector only permitted when compiling runtime, reference was to \"foo\""},
+}
+
+var amd64BadOperandRuntimeTests = []badOperandTest{
+	{"$foo<bletch>", "malformed ABI selector \"bletch\" in reference to \"foo\""},
 }
 
 var x86OperandTests = []operandTest{

src/cmd/asm/internal/asm/parse.go

@@ -43,6 +43,7 @@ type Parser struct {
 	lastProg         *obj.Prog
 	dataAddr         map[string]int64 // Most recent address for DATA for this symbol.
 	isJump           bool             // Instruction being assembled is a jump.
+	compilingRuntime bool
 	errorWriter      io.Writer
 }
 
@@ -51,7 +52,7 @@ type Patch struct {
 	label string
 }
 
-func NewParser(ctxt *obj.Link, ar *arch.Arch, lexer lex.TokenReader) *Parser {
+func NewParser(ctxt *obj.Link, ar *arch.Arch, lexer lex.TokenReader, compilingRuntime bool) *Parser {
 	return &Parser{
 		ctxt:             ctxt,
 		arch:             ar,
@@ -59,6 +60,7 @@ func NewParser(ctxt *obj.Link, ar *arch.Arch, lexer lex.TokenReader) *Parser {
 		labels:           make(map[string]*obj.Prog),
 		dataAddr:         make(map[string]int64),
 		errorWriter:      os.Stderr,
+		compilingRuntime: compilingRuntime,
 	}
 }
 
@@ -310,8 +312,8 @@ func (p *Parser) symDefRef(w io.Writer, word string, operands [][]lex.Token) {
 		// Defines text symbol in operands[0].
 		if len(operands) > 0 {
 			p.start(operands[0])
-			if name, ok := p.funcAddress(); ok {
+			if name, abi, ok := p.funcAddress(); ok {
-				fmt.Fprintf(w, "def %s ABI0\n", name)
+				fmt.Fprintf(w, "def %s %s\n", name, abi)
 			}
 		}
 		return
@@ -329,8 +331,8 @@ func (p *Parser) symDefRef(w io.Writer, word string, operands [][]lex.Token) {
 	// Search for symbol references.
 	for _, op := range operands {
 		p.start(op)
-		if name, ok := p.funcAddress(); ok {
+		if name, abi, ok := p.funcAddress(); ok {
-			fmt.Fprintf(w, "ref %s ABI0\n", name)
+			fmt.Fprintf(w, "ref %s %s\n", name, abi)
 		}
 	}
 }
@@ -765,20 +767,19 @@ func (p *Parser) symbolReference(a *obj.Addr, name string, prefix rune) {
 	case '*':
 		a.Type = obj.TYPE_INDIR
 	}
-	// Weirdness with statics: Might now have "<>".
+
-	isStatic := false
+	// Parse optional <> (indicates a static symbol) or
-	if p.peek() == '<' {
+	// <ABIxxx> (selecting text symbol with specific ABI).
-		isStatic = true
+	doIssueError := true
-		p.next()
+	isStatic, abi := p.symRefAttrs(name, doIssueError)
-		p.get('>')
+
-	}
 	if p.peek() == '+' || p.peek() == '-' {
 		a.Offset = int64(p.expr())
 	}
 	if isStatic {
 		a.Sym = p.ctxt.LookupStatic(name)
 	} else {
-		a.Sym = p.ctxt.Lookup(name)
+		a.Sym = p.ctxt.LookupABI(name, abi)
 	}
 	if p.peek() == scanner.EOF {
 		if prefix == 0 && p.isJump {
@@ -823,12 +824,60 @@ func (p *Parser) setPseudoRegister(addr *obj.Addr, reg string, isStatic bool, pr
 	}
 }
 
+// symRefAttrs parses an optional function symbol attribute clause for
+// the function symbol 'name', logging an error for a malformed
+// attribute clause if 'issueError' is true. The return value is a
+// (boolean, ABI) pair indicating that the named symbol is either
+// static or a particular ABI specification.
+//
+// The expected form of the attribute clause is:
+//
+// empty,           yielding (false, obj.ABI0)
+// "<>",            yielding (true,  obj.ABI0)
+// "<ABI0>"         yielding (false, obj.ABI0)
+// "<ABIInternal>"  yielding (false, obj.ABIInternal)
+//
+// Anything else beginning with "<" logs an error if issueError is
+// true, otherwise returns (false, obj.ABI0).
+//
+func (p *Parser) symRefAttrs(name string, issueError bool) (bool, obj.ABI) {
+	abi := obj.ABI0
+	isStatic := false
+	if p.peek() != '<' {
+		return isStatic, abi
+	}
+	p.next()
+	tok := p.peek()
+	if tok == '>' {
+		isStatic = true
+	} else if tok == scanner.Ident {
+		abistr := p.get(scanner.Ident).String()
+		if !p.compilingRuntime {
+			if issueError {
+				p.errorf("ABI selector only permitted when compiling runtime, reference was to %q", name)
+			}
+		} else {
+			theabi, valid := obj.ParseABI(abistr)
+			if !valid {
+				if issueError {
+					p.errorf("malformed ABI selector %q in reference to %q",
+						abistr, name)
+				}
+			} else {
+				abi = theabi
+			}
+		}
+	}
+	p.get('>')
+	return isStatic, abi
+}
+
 // funcAddress parses an external function address. This is a
 // constrained form of the operand syntax that's always SB-based,
 // non-static, and has at most a simple integer offset:
 //
-//    [$|*]sym[+Int](SB)
+//    [$|*]sym[<abi>][+Int](SB)
-func (p *Parser) funcAddress() (string, bool) {
+func (p *Parser) funcAddress() (string, obj.ABI, bool) {
 	switch p.peek() {
 	case '$', '*':
 		// Skip prefix.
@@ -838,25 +887,32 @@ func (p *Parser) funcAddress() (string, bool) {
 	tok := p.next()
 	name := tok.String()
 	if tok.ScanToken != scanner.Ident || p.atStartOfRegister(name) {
-		return "", false
+		return "", obj.ABI0, false
+	}
+	// Parse optional <> (indicates a static symbol) or
+	// <ABIxxx> (selecting text symbol with specific ABI).
+	noErrMsg := false
+	isStatic, abi := p.symRefAttrs(name, noErrMsg)
+	if isStatic {
+		return "", obj.ABI0, false // This function rejects static symbols.
 	}
 	tok = p.next()
 	if tok.ScanToken == '+' {
 		if p.next().ScanToken != scanner.Int {
-			return "", false
+			return "", obj.ABI0, false
 		}
 		tok = p.next()
 	}
 	if tok.ScanToken != '(' {
-		return "", false
+		return "", obj.ABI0, false
 	}
 	if reg := p.next(); reg.ScanToken != scanner.Ident || reg.String() != "SB" {
-		return "", false
+		return "", obj.ABI0, false
 	}
 	if p.next().ScanToken != ')' || p.peek() != scanner.EOF {
-		return "", false
+		return "", obj.ABI0, false
 	}
-	return name, true
+	return name, abi, true
 }
 
 // registerIndirect parses the general form of a register indirection.

src/cmd/asm/internal/asm/pseudo_test.go

@@ -37,6 +37,7 @@ func TestErroneous(t *testing.T) {
 		{"TEXT", "$0É:0, 0, $1", "expected end of operand, found É"}, // Issue #12467.
 		{"TEXT", "$:0:(SB, 0, $1", "expected '(', found 0"},          // Issue 12468.
 		{"TEXT", "@B(SB),0,$0", "expected '(', found B"},             // Issue 23580.
+		{"TEXT", "foo<ABIInternal>(SB),0", "ABI selector only permitted when compiling runtime, reference was to \"foo\""},
 		{"FUNCDATA", "", "expect two operands for FUNCDATA"},
 		{"FUNCDATA", "(SB ", "expect two operands for FUNCDATA"},
 		{"DATA", "", "expect two operands for DATA"},

src/cmd/asm/main.go

@@ -52,6 +52,7 @@ func main() {
 	case "all", "ret":
 		ctxt.Retpoline = true
 	}
+	compilingRuntime := objabi.IsRuntimePackagePath(*flags.Importpath)
 
 	ctxt.Bso = bufio.NewWriter(os.Stdout)
 	defer ctxt.Bso.Flush()
@@ -74,7 +75,7 @@ func main() {
 	var failedFile string
 	for _, f := range flag.Args() {
 		lexer := lex.NewLexer(f)
-		parser := asm.NewParser(ctxt, architecture, lexer)
+		parser := asm.NewParser(ctxt, architecture, lexer, compilingRuntime)
 		ctxt.DiagFunc = func(format string, args ...interface{}) {
 			diag = true
 			log.Printf(format, args...)

src/cmd/compile/internal/gc/main.go

@@ -968,9 +968,10 @@ func readSymABIs(file, myimportpath string) {
 			if len(parts) != 3 {
 				log.Fatalf(`%s:%d: invalid symabi: syntax is "%s sym abi"`, file, lineNum, parts[0])
 			}
-			sym, abi := parts[1], parts[2]
+			sym, abistr := parts[1], parts[2]
-			if abi != "ABI0" { // Only supported external ABI right now
+			abi, valid := obj.ParseABI(abistr)
-				log.Fatalf(`%s:%d: invalid symabi: unknown abi "%s"`, file, lineNum, abi)
+			if !valid {
+				log.Fatalf(`%s:%d: invalid symabi: unknown abi "%s"`, file, lineNum, abistr)
 			}
 
 			// If the symbol is already prefixed with
@@ -983,9 +984,9 @@ func readSymABIs(file, myimportpath string) {
 
 			// Record for later.
 			if parts[0] == "def" {
-				symabiDefs[sym] = obj.ABI0
+				symabiDefs[sym] = abi
 			} else {
-				symabiRefs[sym] = obj.ABI0
+				symabiRefs[sym] = abi
 			}
 		default:
 			log.Fatalf(`%s:%d: invalid symabi type "%s"`, file, lineNum, parts[0])

src/cmd/internal/obj/link.go

@@ -502,6 +502,20 @@ const (
 	ABICount
 )
 
+// ParseABI converts from a string representation in 'abistr' to the
+// corresponding ABI value. Second return value is TRUE if the
+// abi string is recognized, FALSE otherwise.
+func ParseABI(abistr string) (ABI, bool) {
+	switch abistr {
+	default:
+		return ABI0, false
+	case "ABI0":
+		return ABI0, true
+	case "ABIInternal":
+		return ABIInternal, true
+	}
+}
+
 // Attribute is a set of symbol attributes.
 type Attribute uint32
 

src/cmd/internal/obj/util.go

@@ -210,13 +210,30 @@ func (ctxt *Link) CanReuseProgs() bool {
 	return ctxt.Debugasm == 0
 }
 
+// Dconv accepts an argument 'a' within a prog 'p' and returns a string
+// with a formatted version of the argument.
 func Dconv(p *Prog, a *Addr) string {
 	buf := new(bytes.Buffer)
-	WriteDconv(buf, p, a)
+	writeDconv(buf, p, a, false)
 	return buf.String()
 }
 
+// DconvDconvWithABIDetail accepts an argument 'a' within a prog 'p'
+// and returns a string with a formatted version of the argument, in
+// which text symbols are rendered with explicit ABI selectors.
+func DconvWithABIDetail(p *Prog, a *Addr) string {
+	buf := new(bytes.Buffer)
+	writeDconv(buf, p, a, true)
+	return buf.String()
+}
+
+// WriteDconv accepts an argument 'a' within a prog 'p'
+// and writes a formatted version of the arg to the writer.
 func WriteDconv(w io.Writer, p *Prog, a *Addr) {
+	writeDconv(w, p, a, false)
+}
+
+func writeDconv(w io.Writer, p *Prog, a *Addr, abiDetail bool) {
 	switch a.Type {
 	default:
 		fmt.Fprintf(w, "type=%d", a.Type)
@@ -250,7 +267,7 @@ func WriteDconv(w io.Writer, p *Prog, a *Addr) {
 
 	case TYPE_BRANCH:
 		if a.Sym != nil {
-			fmt.Fprintf(w, "%s(SB)", a.Sym.Name)
+			fmt.Fprintf(w, "%s%s(SB)", a.Sym.Name, abiDecorate(a, abiDetail))
 		} else if a.Target() != nil {
 			fmt.Fprint(w, a.Target().Pc)
 		} else {
@@ -259,7 +276,7 @@ func WriteDconv(w io.Writer, p *Prog, a *Addr) {
 
 	case TYPE_INDIR:
 		io.WriteString(w, "*")
-		a.WriteNameTo(w)
+		a.writeNameTo(w, abiDetail)
 
 	case TYPE_MEM:
 		a.WriteNameTo(w)
@@ -299,7 +316,7 @@ func WriteDconv(w io.Writer, p *Prog, a *Addr) {
 
 	case TYPE_ADDR:
 		io.WriteString(w, "$")
-		a.WriteNameTo(w)
+		a.writeNameTo(w, abiDetail)
 
 	case TYPE_SHIFT:
 		v := int(a.Offset)
@@ -335,6 +352,11 @@ func WriteDconv(w io.Writer, p *Prog, a *Addr) {
 }
 
 func (a *Addr) WriteNameTo(w io.Writer) {
+	a.writeNameTo(w, false)
+}
+
+func (a *Addr) writeNameTo(w io.Writer, abiDetail bool) {
+
 	switch a.Name {
 	default:
 		fmt.Fprintf(w, "name=%d", a.Name)
@@ -356,7 +378,7 @@ func (a *Addr) WriteNameTo(w io.Writer) {
 			reg = Rconv(int(a.Reg))
 		}
 		if a.Sym != nil {
-			fmt.Fprintf(w, "%s%s(%s)", a.Sym.Name, offConv(a.Offset), reg)
+			fmt.Fprintf(w, "%s%s%s(%s)", a.Sym.Name, abiDecorate(a, abiDetail), offConv(a.Offset), reg)
 		} else {
 			fmt.Fprintf(w, "%s(%s)", offConv(a.Offset), reg)
 		}
@@ -596,3 +618,10 @@ func Bool2int(b bool) int {
 	}
 	return i
 }
+
+func abiDecorate(a *Addr, abiDetail bool) string {
+	if !abiDetail || a.Sym == nil {
+		return ""
+	}
+	return fmt.Sprintf("<%s>", a.Sym.ABI())
+}

src/cmd/internal/objabi/path.go

@@ -39,3 +39,25 @@ func PathToPrefix(s string) string {
 
 	return string(p)
 }
+
+// IsRuntimePackagePath examines 'pkgpath' and returns TRUE if it
+// belongs to the collection of "runtime-related" packages, including
+// "runtime" itself, "reflect", "syscall", and the
+// "runtime/internal/*" packages. The compiler and/or assembler in
+// some cases need to be aware of when they are building such a
+// package, for example to enable features such as ABI selectors in
+// assembly sources.
+func IsRuntimePackagePath(pkgpath string) bool {
+	rval := false
+	switch pkgpath {
+	case "runtime":
+		rval = true
+	case "reflect":
+		rval = true
+	case "syscall":
+		rval = true
+	default:
+		rval = strings.HasPrefix(pkgpath, "runtime/internal")
+	}
+	return rval
+}