cmd/objdump, cmd/pprof: factor disassembly into cmd/internal/objfile

Moving so that new Go 1.4 pprof can use it.

The old 'GNU objdump workalike' mode for 'go tool objdump'
is now gone, as are the tests for that mode. It was used only
by pre-Go 1.4 pprof. You can still specify an address range on
the command line; you just get the same output format as
you do when dumping the entire binary (without an address
limitation).

LGTM=r
R=r
CC=golang-codereviews, iant
https://golang.org/cl/167320043

4 files changed, 69 insertions, 326 deletions

diff --git a/src/cmd/internal/objfile/objfile.go b/src/cmd/internal/objfile/objfile.go
index 3d4a5d27cd..9227ef387f 100644
--- a/src/cmd/internal/objfile/objfile.go
+++ b/src/cmd/internal/objfile/objfile.go
@@ -9,6 +9,7 @@ import (
 	"debug/gosym"
 	"fmt"
 	"os"
+	"sort"
 )
 
 type rawFile interface {
@@ -62,9 +63,20 @@ func (f *File) Close() error {
 }
 
 func (f *File) Symbols() ([]Sym, error) {
-	return f.raw.symbols()
+	syms, err := f.raw.symbols()
+	if err != nil {
+		return nil, err
+	}
+	sort.Sort(byAddr(syms))
+	return syms, nil
 }
 
+type byAddr []Sym
+
+func (x byAddr) Less(i, j int) bool { return x[i].Addr < x[j].Addr }
+func (x byAddr) Len() int           { return len(x) }
+func (x byAddr) Swap(i, j int)      { x[i], x[j] = x[j], x[i] }
+
 func (f *File) PCLineTable() (*gosym.Table, error) {
 	textStart, symtab, pclntab, err := f.raw.pcln()
 	if err != nil {
diff --git a/src/cmd/objdump/main.go b/src/cmd/objdump/main.go
index 0f125c98bf..708a853702 100644
--- a/src/cmd/objdump/main.go
+++ b/src/cmd/objdump/main.go
@@ -32,24 +32,15 @@
 package main
 
 import (
-	"bufio"
-	"debug/gosym"
-	"encoding/binary"
 	"flag"
 	"fmt"
-	"io"
 	"log"
 	"os"
 	"regexp"
-	"sort"
 	"strconv"
 	"strings"
-	"text/tabwriter"
 
 	"cmd/internal/objfile"
-
-	"cmd/internal/rsc.io/arm/armasm"
-	"cmd/internal/rsc.io/x86/x86asm"
 )
 
 var symregexp = flag.String("s", "", "only dump symbols matching this regexp")
@@ -87,227 +78,30 @@ func main() {
 		log.Fatal(err)
 	}
 
-	syms, err := f.Symbols()
+	dis, err := f.Disasm()
 	if err != nil {
-		log.Fatalf("reading %s: %v", flag.Arg(0), err)
+		log.Fatal("disassemble %s: %v", flag.Arg(0), err)
 	}
 
-	tab, err := f.PCLineTable()
-	if err != nil {
-		log.Fatalf("reading %s: %v", flag.Arg(0), err)
-	}
-
-	textStart, textBytes, err := f.Text()
-	if err != nil {
-		log.Fatalf("reading %s: %v", flag.Arg(0), err)
-	}
-
-	goarch := f.GOARCH()
-
-	disasm := disasms[goarch]
-	if disasm == nil {
-		log.Fatalf("reading %s: unknown architecture", flag.Arg(0))
-	}
-
-	// Filter out section symbols, overwriting syms in place.
-	keep := syms[:0]
-	for _, sym := range syms {
-		switch sym.Name {
-		case "runtime.text", "text", "_text", "runtime.etext", "etext", "_etext":
-			// drop
-		default:
-			keep = append(keep, sym)
-		}
-	}
-	syms = keep
-
-	sort.Sort(ByAddr(syms))
-	lookup := func(addr uint64) (string, uint64) {
-		i := sort.Search(len(syms), func(i int) bool { return addr < syms[i].Addr })
-		if i > 0 {
-			s := syms[i-1]
-			if s.Addr != 0 && s.Addr <= addr && addr < s.Addr+uint64(s.Size) {
-				return s.Name, s.Addr
-			}
-		}
-		return "", 0
-	}
-
-	if flag.NArg() == 1 {
-		// disassembly of entire object - our format
-		dump(tab, lookup, disasm, goarch, syms, textBytes, textStart)
+	switch flag.NArg() {
+	default:
+		usage()
+	case 1:
+		// disassembly of entire object
+		dis.Print(os.Stdout, symRE, 0, ^uint64(0))
 		os.Exit(0)
-	}
-
-	// disassembly of specific piece of object - gnu objdump format for pprof
-	gnuDump(tab, lookup, disasm, textBytes, textStart)
-	os.Exit(0)
-}
-
-// base returns the final element in the path.
-// It works on both Windows and Unix paths.
-func base(path string) string {
-	path = path[strings.LastIndex(path, "/")+1:]
-	path = path[strings.LastIndex(path, `\`)+1:]
-	return path
-}
-
-func dump(tab *gosym.Table, lookup lookupFunc, disasm disasmFunc, goarch string, syms []objfile.Sym, textData []byte, textStart uint64) {
-	stdout := bufio.NewWriter(os.Stdout)
-	defer stdout.Flush()
-
-	printed := false
-	for _, sym := range syms {
-		if (sym.Code != 'T' && sym.Code != 't') || sym.Size == 0 || sym.Addr < textStart || symRE != nil && !symRE.MatchString(sym.Name) {
-			continue
-		}
-		if sym.Addr >= textStart+uint64(len(textData)) || sym.Addr+uint64(sym.Size) > textStart+uint64(len(textData)) {
-			break
-		}
-		if printed {
-			fmt.Fprintf(stdout, "\n")
-		} else {
-			printed = true
-		}
-		file, _, _ := tab.PCToLine(sym.Addr)
-		fmt.Fprintf(stdout, "TEXT %s(SB) %s\n", sym.Name, file)
-		tw := tabwriter.NewWriter(stdout, 1, 8, 1, '\t', 0)
-		start := sym.Addr
-		end := sym.Addr + uint64(sym.Size)
-		for pc := start; pc < end; {
-			i := pc - textStart
-			text, size := disasm(textData[i:end-textStart], pc, lookup)
-			file, line, _ := tab.PCToLine(pc)
-
-			// ARM is word-based, so show actual word hex, not byte hex.
-			// Since ARM is little endian, they're different.
-			if goarch == "arm" && size == 4 {
-				fmt.Fprintf(tw, "\t%s:%d\t%#x\t%08x\t%s\n", base(file), line, pc, binary.LittleEndian.Uint32(textData[i:i+uint64(size)]), text)
-			} else {
-				fmt.Fprintf(tw, "\t%s:%d\t%#x\t%x\t%s\n", base(file), line, pc, textData[i:i+uint64(size)], text)
-			}
-			pc += uint64(size)
-		}
-		tw.Flush()
-	}
-}
-
-func disasm_386(code []byte, pc uint64, lookup lookupFunc) (string, int) {
-	return disasm_x86(code, pc, lookup, 32)
-}
-
-func disasm_amd64(code []byte, pc uint64, lookup lookupFunc) (string, int) {
-	return disasm_x86(code, pc, lookup, 64)
-}
-
-func disasm_x86(code []byte, pc uint64, lookup lookupFunc, arch int) (string, int) {
-	inst, err := x86asm.Decode(code, 64)
-	var text string
-	size := inst.Len
-	if err != nil || size == 0 || inst.Op == 0 {
-		size = 1
-		text = "?"
-	} else {
-		text = x86asm.Plan9Syntax(inst, pc, lookup)
-	}
-	return text, size
-}
-
-type textReader struct {
-	code []byte
-	pc   uint64
-}
-
-func (r textReader) ReadAt(data []byte, off int64) (n int, err error) {
-	if off < 0 || uint64(off) < r.pc {
-		return 0, io.EOF
-	}
-	d := uint64(off) - r.pc
-	if d >= uint64(len(r.code)) {
-		return 0, io.EOF
-	}
-	n = copy(data, r.code[d:])
-	if n < len(data) {
-		err = io.ErrUnexpectedEOF
-	}
-	return
-}
-
-func disasm_arm(code []byte, pc uint64, lookup lookupFunc) (string, int) {
-	inst, err := armasm.Decode(code, armasm.ModeARM)
-	var text string
-	size := inst.Len
-	if err != nil || size == 0 || inst.Op == 0 {
-		size = 4
-		text = "?"
-	} else {
-		text = armasm.Plan9Syntax(inst, pc, lookup, textReader{code, pc})
-	}
-	return text, size
-}
-
-var disasms = map[string]disasmFunc{
-	"386":   disasm_386,
-	"amd64": disasm_amd64,
-	"arm":   disasm_arm,
-}
-
-func gnuDump(tab *gosym.Table, lookup lookupFunc, disasm disasmFunc, textData []byte, textStart uint64) {
-	start, err := strconv.ParseUint(strings.TrimPrefix(flag.Arg(1), "0x"), 16, 64)
-	if err != nil {
-		log.Fatalf("invalid start PC: %v", err)
-	}
-	end, err := strconv.ParseUint(strings.TrimPrefix(flag.Arg(2), "0x"), 16, 64)
-	if err != nil {
-		log.Fatalf("invalid end PC: %v", err)
-	}
-	if start < textStart {
-		start = textStart
-	}
-	if end < start {
-		end = start
-	}
-	if end > textStart+uint64(len(textData)) {
-		end = textStart + uint64(len(textData))
-	}
-
-	stdout := bufio.NewWriter(os.Stdout)
-	defer stdout.Flush()
-
-	// For now, find spans of same PC/line/fn and
-	// emit them as having dummy instructions.
-	var (
-		spanPC   uint64
-		spanFile string
-		spanLine int
-		spanFn   *gosym.Func
-	)
 
-	flush := func(endPC uint64) {
-		if spanPC == 0 {
-			return
-		}
-		fmt.Fprintf(stdout, "%s:%d\n", spanFile, spanLine)
-		for pc := spanPC; pc < endPC; {
-			text, size := disasm(textData[pc-textStart:], pc, lookup)
-			fmt.Fprintf(stdout, " %x: %s\n", pc, text)
-			pc += uint64(size)
+	case 3:
+		// disassembly of PC range
+		start, err := strconv.ParseUint(strings.TrimPrefix(flag.Arg(1), "0x"), 16, 64)
+		if err != nil {
+			log.Fatalf("invalid start PC: %v", err)
 		}
-		spanPC = 0
-	}
-
-	for pc := start; pc < end; pc++ {
-		file, line, fn := tab.PCToLine(pc)
-		if file != spanFile || line != spanLine || fn != spanFn {
-			flush(pc)
-			spanPC, spanFile, spanLine, spanFn = pc, file, line, fn
+		end, err := strconv.ParseUint(strings.TrimPrefix(flag.Arg(2), "0x"), 16, 64)
+		if err != nil {
+			log.Fatalf("invalid end PC: %v", err)
 		}
+		dis.Print(os.Stdout, symRE, start, end)
+		os.Exit(0)
 	}
-	flush(end)
 }
-
-type ByAddr []objfile.Sym
-
-func (x ByAddr) Less(i, j int) bool { return x[i].Addr < x[j].Addr }
-func (x ByAddr) Len() int           { return len(x) }
-func (x ByAddr) Swap(i, j int)      { x[i], x[j] = x[j], x[i] }
diff --git a/src/cmd/objdump/objdump_test.go b/src/cmd/objdump/objdump_test.go
index ffaaa5b437..2bb74663c3 100644
--- a/src/cmd/objdump/objdump_test.go
+++ b/src/cmd/objdump/objdump_test.go
@@ -5,113 +5,15 @@
 package main
 
 import (
-	"bufio"
-	"bytes"
-	"fmt"
 	"io/ioutil"
 	"os"
 	"os/exec"
 	"path/filepath"
 	"runtime"
-	"strconv"
 	"strings"
 	"testing"
 )
 
-func loadSyms(t *testing.T) map[string]string {
-	switch runtime.GOOS {
-	case "android", "nacl":
-		t.Skipf("skipping on %s", runtime.GOOS)
-	}
-
-	cmd := exec.Command("go", "tool", "nm", os.Args[0])
-	out, err := cmd.CombinedOutput()
-	if err != nil {
-		t.Fatalf("go tool nm %v: %v\n%s", os.Args[0], err, string(out))
-	}
-	syms := make(map[string]string)
-	scanner := bufio.NewScanner(bytes.NewReader(out))
-	for scanner.Scan() {
-		f := strings.Fields(scanner.Text())
-		if len(f) < 3 {
-			continue
-		}
-		syms[f[2]] = f[0]
-	}
-	if err := scanner.Err(); err != nil {
-		t.Fatalf("error reading symbols: %v", err)
-	}
-	return syms
-}
-
-func runObjDump(t *testing.T, exe, startaddr, endaddr string) (path, lineno string) {
-	switch runtime.GOOS {
-	case "android", "nacl":
-		t.Skipf("skipping on %s", runtime.GOOS)
-	}
-
-	cmd := exec.Command(exe, os.Args[0], startaddr, endaddr)
-	out, err := cmd.CombinedOutput()
-	if err != nil {
-		t.Fatalf("go tool objdump %v: %v\n%s", os.Args[0], err, string(out))
-	}
-	f := strings.Split(string(out), "\n")
-	if len(f) < 1 {
-		t.Fatal("objdump output must have at least one line")
-	}
-	pathAndLineNo := f[0]
-	f = strings.Split(pathAndLineNo, ":")
-	if runtime.GOOS == "windows" {
-		switch len(f) {
-		case 2:
-			return f[0], f[1]
-		case 3:
-			return f[0] + ":" + f[1], f[2]
-		default:
-			t.Fatalf("no line number found in %q", pathAndLineNo)
-		}
-	}
-	if len(f) != 2 {
-		t.Fatalf("no line number found in %q", pathAndLineNo)
-	}
-	return f[0], f[1]
-}
-
-func testObjDump(t *testing.T, exe, startaddr, endaddr string, line int) {
-	srcPath, srcLineNo := runObjDump(t, exe, startaddr, endaddr)
-	fi1, err := os.Stat("objdump_test.go")
-	if err != nil {
-		t.Fatalf("Stat failed: %v", err)
-	}
-	fi2, err := os.Stat(srcPath)
-	if err != nil {
-		t.Fatalf("Stat failed: %v", err)
-	}
-	if !os.SameFile(fi1, fi2) {
-		t.Fatalf("objdump_test.go and %s are not same file", srcPath)
-	}
-	if srcLineNo != fmt.Sprint(line) {
-		t.Fatalf("line number = %v; want %d", srcLineNo, line)
-	}
-}
-
-func TestObjDump(t *testing.T) {
-	_, _, line, _ := runtime.Caller(0)
-	syms := loadSyms(t)
-
-	tmp, exe := buildObjdump(t)
-	defer os.RemoveAll(tmp)
-
-	startaddr := syms["cmd/objdump.TestObjDump"]
-	addr, err := strconv.ParseUint(startaddr, 16, 64)
-	if err != nil {
-		t.Fatalf("invalid start address %v: %v", startaddr, err)
-	}
-	endaddr := fmt.Sprintf("%x", addr+10)
-	testObjDump(t, exe, startaddr, endaddr, line-1)
-	testObjDump(t, exe, "0x"+startaddr, "0x"+endaddr, line-1)
-}
-
 func buildObjdump(t *testing.T) (tmp, exe string) {
 	switch runtime.GOOS {
 	case "android", "nacl":
diff --git a/src/cmd/pprof/pprof.go b/src/cmd/pprof/pprof.go
index 89a5bb7d22..44f4f6cb72 100644
--- a/src/cmd/pprof/pprof.go
+++ b/src/cmd/pprof/pprof.go
@@ -11,6 +11,7 @@ import (
 	"os"
 	"regexp"
 	"strings"
+	"sync"
 
 	"cmd/internal/objfile"
 	"cmd/pprof/internal/commands"
@@ -100,7 +101,10 @@ func (flags) ExtraUsage() string {
 
 // objTool implements plugin.ObjTool using Go libraries
 // (instead of invoking GNU binutils).
-type objTool struct{}
+type objTool struct {
+	mu          sync.Mutex
+	disasmCache map[string]*objfile.Disasm
+}
 
 func (*objTool) Open(name string, start uint64) (plugin.ObjFile, error) {
 	of, err := objfile.Open(name)
@@ -119,8 +123,39 @@ func (*objTool) Demangle(names []string) (map[string]string, error) {
 	return make(map[string]string), nil
 }
 
-func (*objTool) Disasm(file string, start, end uint64) ([]plugin.Inst, error) {
-	return nil, fmt.Errorf("disassembly not supported")
+func (t *objTool) Disasm(file string, start, end uint64) ([]plugin.Inst, error) {
+	d, err := t.cachedDisasm(file)
+	if err != nil {
+		return nil, err
+	}
+	var asm []plugin.Inst
+	d.Decode(start, end, func(pc, size uint64, file string, line int, text string) {
+		asm = append(asm, plugin.Inst{Addr: pc, File: file, Line: line, Text: text})
+	})
+	return asm, nil
+}
+
+func (t *objTool) cachedDisasm(file string) (*objfile.Disasm, error) {
+	t.mu.Lock()
+	defer t.mu.Unlock()
+	if t.disasmCache == nil {
+		t.disasmCache = make(map[string]*objfile.Disasm)
+	}
+	d := t.disasmCache[file]
+	if d != nil {
+		return d, nil
+	}
+	f, err := objfile.Open(file)
+	if err != nil {
+		return nil, err
+	}
+	d, err = f.Disasm()
+	f.Close()
+	if err != nil {
+		return nil, err
+	}
+	t.disasmCache[file] = d
+	return d, nil
 }
 
 func (*objTool) SetConfig(config string) {