cmd/internal/obj: minor refactor of wasmimport code

This CL does some minor refactoring of the code handling
wasmimport.
- Put the WasmImport aux reading and writing code together for
  symmetry.
- Define WasmFuncType, embedded in WasmImport. WasmFuncType could
  also be used (later) for wasmexport.
- Move code generation code to a separate function. The containing
  function is already pretty large.
- Simplify linker code a little bit. The loader convention is to
  return the 0 Sym for nonexistent symbol, instead of a separate
  boolean.

No change in generated code. Passes toolstash -cmp
(GOARCH=wasm GOOS=wasip1 go build -toolexec "toolstash -cmp" -a std cmd).

Change-Id: Idc2514f84a08621333841ae4034b81130e0ce411
Reviewed-on: https://go-review.googlesource.com/c/go/+/603135
Reviewed-by: Than McIntosh <thanm@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>

src/cmd/internal/obj/link.go

@@ -32,6 +32,7 @@ package obj
 
 import (
 	"bufio"
+	"bytes"
 	"cmd/internal/dwarf"
 	"cmd/internal/goobj"
 	"cmd/internal/objabi"
@@ -497,7 +498,7 @@ type FuncInfo struct {
 	JumpTables         []JumpTable
 
 	FuncInfoSym *LSym
-	WasmImportSym *LSym
+
 	WasmImport *WasmImport
 
 	sehUnwindInfoSym *LSym
@@ -609,45 +610,118 @@ type WasmImport struct {
 	// Name holds the WASM imported function name specified by the
 	// //go:wasmimport directive.
 	Name string
+
+	WasmFuncType // type of the imported function
+
+	// aux symbol to pass metadata to the linker, serialization of
+	// the fields above.
+	AuxSym *LSym
+}
+
+func (wi *WasmImport) CreateAuxSym() {
+	var b bytes.Buffer
+	wi.Write(&b)
+	p := b.Bytes()
+	wi.AuxSym = &LSym{
+		Type: objabi.SDATA, // doesn't really matter
+		P:    append([]byte(nil), p...),
+		Size: int64(len(p)),
+	}
+}
+
+func (wi *WasmImport) Write(w *bytes.Buffer) {
+	var b [8]byte
+	writeUint32 := func(x uint32) {
+		binary.LittleEndian.PutUint32(b[:], x)
+		w.Write(b[:4])
+	}
+	writeString := func(s string) {
+		writeUint32(uint32(len(s)))
+		w.WriteString(s)
+	}
+	writeString(wi.Module)
+	writeString(wi.Name)
+	wi.WasmFuncType.Write(w)
+}
+
+func (wi *WasmImport) Read(b []byte) {
+	readUint32 := func() uint32 {
+		x := binary.LittleEndian.Uint32(b)
+		b = b[4:]
+		return x
+	}
+	readString := func() string {
+		n := readUint32()
+		s := string(b[:n])
+		b = b[n:]
+		return s
+	}
+	wi.Module = readString()
+	wi.Name = readString()
+	wi.WasmFuncType.Read(b)
+}
+
+// WasmFuncType represents a WebAssembly (WASM) function type with
+// parameters and results translated into WASM types based on the Go function
+// declaration.
+type WasmFuncType struct {
 	// Params holds the imported function parameter fields.
 	Params []WasmField
 	// Results holds the imported function result fields.
 	Results []WasmField
 }
 
-func (wi *WasmImport) CreateSym(ctxt *Link) *LSym {
+func (ft *WasmFuncType) Write(w *bytes.Buffer) {
-	var sym LSym
-
 	var b [8]byte
 	writeByte := func(x byte) {
-		sym.WriteBytes(ctxt, sym.Size, []byte{x})
+		w.WriteByte(x)
 	}
 	writeUint32 := func(x uint32) {
 		binary.LittleEndian.PutUint32(b[:], x)
-		sym.WriteBytes(ctxt, sym.Size, b[:4])
+		w.Write(b[:4])
 	}
 	writeInt64 := func(x int64) {
 		binary.LittleEndian.PutUint64(b[:], uint64(x))
-		sym.WriteBytes(ctxt, sym.Size, b[:])
+		w.Write(b[:])
 	}
-	writeString := func(s string) {
+	writeUint32(uint32(len(ft.Params)))
-		writeUint32(uint32(len(s)))
+	for _, f := range ft.Params {
-		sym.WriteString(ctxt, sym.Size, len(s), s)
-	}
-	writeString(wi.Module)
-	writeString(wi.Name)
-	writeUint32(uint32(len(wi.Params)))
-	for _, f := range wi.Params {
 		writeByte(byte(f.Type))
 		writeInt64(f.Offset)
 	}
-	writeUint32(uint32(len(wi.Results)))
+	writeUint32(uint32(len(ft.Results)))
-	for _, f := range wi.Results {
+	for _, f := range ft.Results {
 		writeByte(byte(f.Type))
 		writeInt64(f.Offset)
 	}
+}
 
-	return &sym
+func (ft *WasmFuncType) Read(b []byte) {
+	readByte := func() byte {
+		x := b[0]
+		b = b[1:]
+		return x
+	}
+	readUint32 := func() uint32 {
+		x := binary.LittleEndian.Uint32(b)
+		b = b[4:]
+		return x
+	}
+	readInt64 := func() int64 {
+		x := binary.LittleEndian.Uint64(b)
+		b = b[8:]
+		return int64(x)
+	}
+	ft.Params = make([]WasmField, readUint32())
+	for i := range ft.Params {
+		ft.Params[i].Type = WasmFieldType(readByte())
+		ft.Params[i].Offset = int64(readInt64())
+	}
+	ft.Results = make([]WasmField, readUint32())
+	for i := range ft.Results {
+		ft.Results[i].Type = WasmFieldType(readByte())
+		ft.Results[i].Offset = int64(readInt64())
+	}
 }
 
 type WasmField struct {

src/cmd/internal/obj/objfile.go

@@ -621,11 +621,11 @@ func (w *writer) Aux(s *LSym) {
 		for _, pcSym := range fn.Pcln.Pcdata {
 			w.aux1(goobj.AuxPcdata, pcSym)
 		}
-		if fn.WasmImportSym != nil {
+		if fn.WasmImport != nil {
-			if fn.WasmImportSym.Size == 0 {
+			if fn.WasmImport.AuxSym.Size == 0 {
 				panic("wasmimport aux sym must have non-zero size")
 			}
-			w.aux1(goobj.AuxWasmImport, fn.WasmImportSym)
+			w.aux1(goobj.AuxWasmImport, fn.WasmImport.AuxSym)
 		}
 	} else if v := s.VarInfo(); v != nil {
 		if v.dwarfInfoSym != nil && v.dwarfInfoSym.Size != 0 {
@@ -732,7 +732,7 @@ func nAuxSym(s *LSym) int {
 		}
 		n += len(fn.Pcln.Pcdata)
 		if fn.WasmImport != nil {
-			if fn.WasmImportSym == nil || fn.WasmImportSym.Size == 0 {
+			if fn.WasmImport.AuxSym == nil || fn.WasmImport.AuxSym.Size == 0 {
 				panic("wasmimport aux sym must exist and have non-zero size")
 			}
 			n++
@@ -797,7 +797,10 @@ func genFuncInfoSyms(ctxt *Link) {
 		fn.FuncInfoSym = isym
 		b.Reset()
 
-		auxsyms := []*LSym{fn.dwarfRangesSym, fn.dwarfLocSym, fn.dwarfDebugLinesSym, fn.dwarfInfoSym, fn.WasmImportSym}
+		auxsyms := []*LSym{fn.dwarfRangesSym, fn.dwarfLocSym, fn.dwarfDebugLinesSym, fn.dwarfInfoSym}
+		if wi := fn.WasmImport; wi != nil {
+			auxsyms = append(auxsyms, wi.AuxSym)
+		}
 		for _, s := range auxsyms {
 			if s == nil || s.Size == 0 {
 				continue

src/cmd/internal/obj/sym.go

@@ -458,7 +458,10 @@ func (ctxt *Link) traverseFuncAux(flag traverseFlag, fsym *LSym, fn func(parent
 		}
 	}
 
-	auxsyms := []*LSym{fninfo.dwarfRangesSym, fninfo.dwarfLocSym, fninfo.dwarfDebugLinesSym, fninfo.dwarfInfoSym, fninfo.WasmImportSym, fninfo.sehUnwindInfoSym}
+	auxsyms := []*LSym{fninfo.dwarfRangesSym, fninfo.dwarfLocSym, fninfo.dwarfDebugLinesSym, fninfo.dwarfInfoSym, fninfo.sehUnwindInfoSym}
+	if wi := fninfo.WasmImport; wi != nil {
+		auxsyms = append(auxsyms, wi.AuxSym)
+	}
 	for _, s := range auxsyms {
 		if s == nil || s.Size == 0 {
 			continue

src/cmd/internal/obj/wasm/wasmobj.go

@@ -188,111 +188,8 @@ func preprocess(ctxt *obj.Link, s *obj.LSym, newprog obj.ProgAlloc) {
 	// If the function exits just to call out to a wasmimport, then
 	// generate the code to translate from our internal Go-stack
 	// based call convention to the native webassembly call convention.
-	if wi := s.Func().WasmImport; wi != nil {
+	if s.Func().WasmImport != nil {
-		s.Func().WasmImportSym = wi.CreateSym(ctxt)
+		genWasmImportWrapper(s, appendp)
-		p := s.Func().Text
-		if p.Link != nil {
-			panic("wrapper functions for WASM imports should not have a body")
-		}
-		to := obj.Addr{
-			Type: obj.TYPE_MEM,
-			Name: obj.NAME_EXTERN,
-			Sym:  s,
-		}
-
-		// If the module that the import is for is our magic "gojs" module, then this
-		// indicates that the called function understands the Go stack-based call convention
-		// so we just pass the stack pointer to it, knowing it will read the params directly
-		// off the stack and push the results into memory based on the stack pointer.
-		if wi.Module == GojsModule {
-			// The called function has a signature of 'func(sp int)'. It has access to the memory
-			// value somewhere to be able to address the memory based on the "sp" value.
-
-			p = appendp(p, AGet, regAddr(REG_SP))
-			p = appendp(p, ACall, to)
-
-			p.Mark = WasmImport
-		} else {
-			if len(wi.Results) > 1 {
-				// TODO(evanphx) implement support for the multi-value proposal:
-				// https://github.com/WebAssembly/multi-value/blob/master/proposals/multi-value/Overview.md
-				panic("invalid results type") // impossible until multi-value proposal has landed
-			}
-			if len(wi.Results) == 1 {
-				// If we have a result (rather than returning nothing at all), then
-				// we'll write the result to the Go stack relative to the current stack pointer.
-				// We cache the current stack pointer value on the wasm stack here and then use
-				// it after the Call instruction to store the result.
-				p = appendp(p, AGet, regAddr(REG_SP))
-			}
-			for _, f := range wi.Params {
-				// Each load instructions will consume the value of sp on the stack, so
-				// we need to read sp for each param. WASM appears to not have a stack dup instruction
-				// (a strange omission for a stack-based VM), if it did, we'd be using the dup here.
-				p = appendp(p, AGet, regAddr(REG_SP))
-
-				// Offset is the location of the param on the Go stack (ie relative to sp).
-				// Because of our call convention, the parameters are located an additional 8 bytes
-				// from sp because we store the return address as an int64 at the bottom of the stack.
-				// Ie the stack looks like [return_addr, param3, param2, param1, etc]
-
-				// Ergo, we add 8 to the true byte offset of the param to skip the return address.
-				loadOffset := f.Offset + 8
-
-				// We're reading the value from the Go stack onto the WASM stack and leaving it there
-				// for CALL to pick them up.
-				switch f.Type {
-				case obj.WasmI32:
-					p = appendp(p, AI32Load, constAddr(loadOffset))
-				case obj.WasmI64:
-					p = appendp(p, AI64Load, constAddr(loadOffset))
-				case obj.WasmF32:
-					p = appendp(p, AF32Load, constAddr(loadOffset))
-				case obj.WasmF64:
-					p = appendp(p, AF64Load, constAddr(loadOffset))
-				case obj.WasmPtr:
-					p = appendp(p, AI64Load, constAddr(loadOffset))
-					p = appendp(p, AI32WrapI64)
-				default:
-					panic("bad param type")
-				}
-			}
-
-			// The call instruction is marked as being for a wasm import so that a later phase
-			// will generate relocation information that allows us to patch this with then
-			// offset of the imported function in the wasm imports.
-			p = appendp(p, ACall, to)
-			p.Mark = WasmImport
-
-			if len(wi.Results) == 1 {
-				f := wi.Results[0]
-
-				// Much like with the params, we need to adjust the offset we store the result value
-				// to by 8 bytes to account for the return address on the Go stack.
-				storeOffset := f.Offset + 8
-
-				// This code is paired the code above that reads the stack pointer onto the wasm
-				// stack. We've done this so we have a consistent view of the sp value as it might
-				// be manipulated by the call and we want to ignore that manipulation here.
-				switch f.Type {
-				case obj.WasmI32:
-					p = appendp(p, AI32Store, constAddr(storeOffset))
-				case obj.WasmI64:
-					p = appendp(p, AI64Store, constAddr(storeOffset))
-				case obj.WasmF32:
-					p = appendp(p, AF32Store, constAddr(storeOffset))
-				case obj.WasmF64:
-					p = appendp(p, AF64Store, constAddr(storeOffset))
-				case obj.WasmPtr:
-					p = appendp(p, AI64ExtendI32U)
-					p = appendp(p, AI64Store, constAddr(storeOffset))
-				default:
-					panic("bad result type")
-				}
-			}
-		}
-
-		p = appendp(p, obj.ARET)
 
 		// It should be 0 already, but we'll set it to 0 anyway just to be sure
 		// that the code below which adds frame expansion code to the function body
@@ -894,6 +791,115 @@ func preprocess(ctxt *obj.Link, s *obj.LSym, newprog obj.ProgAlloc) {
 	}
 }
 
+// Generate function body for wasmimport wrapper function.
+func genWasmImportWrapper(s *obj.LSym, appendp func(p *obj.Prog, as obj.As, args ...obj.Addr) *obj.Prog) {
+	wi := s.Func().WasmImport
+	wi.CreateAuxSym()
+	p := s.Func().Text
+	if p.Link != nil {
+		panic("wrapper functions for WASM imports should not have a body")
+	}
+	to := obj.Addr{
+		Type: obj.TYPE_MEM,
+		Name: obj.NAME_EXTERN,
+		Sym:  s,
+	}
+
+	// If the module that the import is for is our magic "gojs" module, then this
+	// indicates that the called function understands the Go stack-based call convention
+	// so we just pass the stack pointer to it, knowing it will read the params directly
+	// off the stack and push the results into memory based on the stack pointer.
+	if wi.Module == GojsModule {
+		// The called function has a signature of 'func(sp int)'. It has access to the memory
+		// value somewhere to be able to address the memory based on the "sp" value.
+
+		p = appendp(p, AGet, regAddr(REG_SP))
+		p = appendp(p, ACall, to)
+
+		p.Mark = WasmImport
+	} else {
+		if len(wi.Results) > 1 {
+			// TODO(evanphx) implement support for the multi-value proposal:
+			// https://github.com/WebAssembly/multi-value/blob/master/proposals/multi-value/Overview.md
+			panic("invalid results type") // impossible until multi-value proposal has landed
+		}
+		if len(wi.Results) == 1 {
+			// If we have a result (rather than returning nothing at all), then
+			// we'll write the result to the Go stack relative to the current stack pointer.
+			// We cache the current stack pointer value on the wasm stack here and then use
+			// it after the Call instruction to store the result.
+			p = appendp(p, AGet, regAddr(REG_SP))
+		}
+		for _, f := range wi.Params {
+			// Each load instructions will consume the value of sp on the stack, so
+			// we need to read sp for each param. WASM appears to not have a stack dup instruction
+			// (a strange omission for a stack-based VM), if it did, we'd be using the dup here.
+			p = appendp(p, AGet, regAddr(REG_SP))
+
+			// Offset is the location of the param on the Go stack (ie relative to sp).
+			// Because of our call convention, the parameters are located an additional 8 bytes
+			// from sp because we store the return address as an int64 at the bottom of the stack.
+			// Ie the stack looks like [return_addr, param3, param2, param1, etc]
+
+			// Ergo, we add 8 to the true byte offset of the param to skip the return address.
+			loadOffset := f.Offset + 8
+
+			// We're reading the value from the Go stack onto the WASM stack and leaving it there
+			// for CALL to pick them up.
+			switch f.Type {
+			case obj.WasmI32:
+				p = appendp(p, AI32Load, constAddr(loadOffset))
+			case obj.WasmI64:
+				p = appendp(p, AI64Load, constAddr(loadOffset))
+			case obj.WasmF32:
+				p = appendp(p, AF32Load, constAddr(loadOffset))
+			case obj.WasmF64:
+				p = appendp(p, AF64Load, constAddr(loadOffset))
+			case obj.WasmPtr:
+				p = appendp(p, AI64Load, constAddr(loadOffset))
+				p = appendp(p, AI32WrapI64)
+			default:
+				panic("bad param type")
+			}
+		}
+
+		// The call instruction is marked as being for a wasm import so that a later phase
+		// will generate relocation information that allows us to patch this with then
+		// offset of the imported function in the wasm imports.
+		p = appendp(p, ACall, to)
+		p.Mark = WasmImport
+
+		if len(wi.Results) == 1 {
+			f := wi.Results[0]
+
+			// Much like with the params, we need to adjust the offset we store the result value
+			// to by 8 bytes to account for the return address on the Go stack.
+			storeOffset := f.Offset + 8
+
+			// This code is paired the code above that reads the stack pointer onto the wasm
+			// stack. We've done this so we have a consistent view of the sp value as it might
+			// be manipulated by the call and we want to ignore that manipulation here.
+			switch f.Type {
+			case obj.WasmI32:
+				p = appendp(p, AI32Store, constAddr(storeOffset))
+			case obj.WasmI64:
+				p = appendp(p, AI64Store, constAddr(storeOffset))
+			case obj.WasmF32:
+				p = appendp(p, AF32Store, constAddr(storeOffset))
+			case obj.WasmF64:
+				p = appendp(p, AF64Store, constAddr(storeOffset))
+			case obj.WasmPtr:
+				p = appendp(p, AI64ExtendI32U)
+				p = appendp(p, AI64Store, constAddr(storeOffset))
+			default:
+				panic("bad result type")
+			}
+		}
+	}
+
+	p = appendp(p, obj.ARET)
+}
+
 func constAddr(value int64) obj.Addr {
 	return obj.Addr{Type: obj.TYPE_CONST, Offset: value}
 }

src/cmd/link/internal/loader/loader.go

@@ -1620,21 +1620,11 @@ func (l *Loader) Aux(i Sym, j int) Aux {
 // contains the information necessary for the linker to add a WebAssembly
 // import statement.
 // (https://webassembly.github.io/spec/core/syntax/modules.html#imports)
-func (l *Loader) WasmImportSym(fnSymIdx Sym) (Sym, bool) {
+func (l *Loader) WasmImportSym(fnSymIdx Sym) Sym {
 	if l.SymType(fnSymIdx) != sym.STEXT {
 		log.Fatalf("error: non-function sym %d/%s t=%s passed to WasmImportSym", fnSymIdx, l.SymName(fnSymIdx), l.SymType(fnSymIdx).String())
 	}
-	r, li := l.toLocal(fnSymIdx)
+	return l.aux1(fnSymIdx, goobj.AuxWasmImport)
-	auxs := r.Auxs(li)
-	for i := range auxs {
-		a := &auxs[i]
-		switch a.Type() {
-		case goobj.AuxWasmImport:
-			return l.resolve(r, a.Sym()), true
-		}
-	}
-
-	return 0, false
 }
 
 // SEHUnwindSym returns the auxiliary SEH unwind symbol associated with

src/cmd/link/internal/wasm/asm.go

@@ -12,7 +12,6 @@ import (
 	"cmd/link/internal/ld"
 	"cmd/link/internal/loader"
 	"cmd/link/internal/sym"
-	"encoding/binary"
 	"fmt"
 	"internal/abi"
 	"internal/buildcfg"
@@ -61,55 +60,8 @@ type wasmFuncType struct {
 }
 
 func readWasmImport(ldr *loader.Loader, s loader.Sym) obj.WasmImport {
-	reportError := func(err error) { panic(fmt.Sprintf("failed to read WASM import in sym %v: %v", s, err)) }
-
-	data := ldr.Data(s)
-
-	readUint32 := func() (v uint32) {
-		v = binary.LittleEndian.Uint32(data)
-		data = data[4:]
-		return
-	}
-
-	readUint64 := func() (v uint64) {
-		v = binary.LittleEndian.Uint64(data)
-		data = data[8:]
-		return
-	}
-
-	readByte := func() byte {
-		if len(data) == 0 {
-			reportError(io.EOF)
-		}
-
-		b := data[0]
-		data = data[1:]
-		return b
-	}
-
-	readString := func() string {
-		n := readUint32()
-
-		s := string(data[:n])
-
-		data = data[n:]
-
-		return s
-	}
-
 	var wi obj.WasmImport
-	wi.Module = readString()
+	wi.Read(ldr.Data(s))
-	wi.Name = readString()
-	wi.Params = make([]obj.WasmField, readUint32())
-	for i := range wi.Params {
-		wi.Params[i].Type = obj.WasmFieldType(readByte())
-		wi.Params[i].Offset = int64(readUint64())
-	}
-	wi.Results = make([]obj.WasmField, readUint32())
-	for i := range wi.Results {
-		wi.Results[i].Type = obj.WasmFieldType(readByte())
-		wi.Results[i].Offset = int64(readUint64())
-	}
 	return wi
 }
 
@@ -207,8 +159,8 @@ func asmb2(ctxt *ld.Link, ldr *loader.Loader) {
 		for ri := 0; ri < relocs.Count(); ri++ {
 			r := relocs.At(ri)
 			if r.Type() == objabi.R_WASMIMPORT {
-				if lsym, ok := ldr.WasmImportSym(fn); ok {
+				if wsym := ldr.WasmImportSym(fn); wsym != 0 {
-					wi := readWasmImport(ldr, lsym)
+					wi := readWasmImport(ldr, wsym)
 					hostImportMap[fn] = int64(len(hostImports))
 					hostImports = append(hostImports, &wasmFunc{
 						Module: wi.Module,