dwarf/line: fix some bugs with the state machine

Adds a test that compares the output of our state machine with the
output of the debug_line reader in the standard library and checks that
they produce the same output for the debug_line section of grafana as
compiled on macOS (which is the most interesting case since it uses cgo
and therefore goes through dsymutil).

A few bugs were uncovered and fixed:

1. is_stmt was reset improperly after a DW_LNS_end_sequence instruction
2. basic_block, prologue_end and epilogue_begin were not reset after a
   DW_LNS_copy instruction
3. some opcodes were not decoded properly if the debug_line section
   declares fewer standard opcodes than we know about.

Fixes #1282

pkg/dwarf/line/state_machine.go

@@ -5,6 +5,7 @@ import (
 	"encoding/binary"
 	"errors"
 	"fmt"
+	"io"
 
 	"github.com/derekparker/delve/pkg/dwarf/util"
 )
@@ -34,8 +35,6 @@ type StateMachine struct {
 	// the compilation unit)
 	valid bool
 
-	lastOpcodeKind opcodeKind
-
 	started bool
 
 	buf     *bytes.Buffer // remaining instructions
@@ -272,7 +271,7 @@ func (lineInfo *DebugLineInfo) LineToPC(filename string, lineno int) uint64 {
 
 	for {
 		if err := sm.next(); err != nil {
-			if lineInfo.Logf != nil {
+			if lineInfo.Logf != nil && err != io.EOF {
 				lineInfo.Logf("LineToPC error: %v", err)
 			}
 			break
@@ -313,38 +312,37 @@ func (sm *StateMachine) next() error {
 	sm.started = true
 	if sm.valid {
 		sm.lastAddress, sm.lastFile, sm.lastLine = sm.address, sm.file, sm.line
+
+		// valid is set by either a special opcode or a DW_LNS_copy, in both cases
+		// we need to reset basic_block, prologue_end and epilogue_begin
+		sm.basicBlock = false
+		sm.prologueEnd = false
+		sm.epilogueBegin = false
 	}
 	if sm.endSeq {
 		sm.endSeq = false
 		sm.file = sm.dbl.FileNames[0].Path
 		sm.line = 1
 		sm.column = 0
-		sm.isStmt = false
+		sm.isStmt = sm.dbl.Prologue.InitialIsStmt == uint8(1)
 		sm.basicBlock = false
 	}
-	if sm.lastOpcodeKind == specialOpcode {
-		sm.basicBlock = false
-		sm.prologueEnd = false
-		sm.epilogueBegin = false
-	}
 	b, err := sm.buf.ReadByte()
 	if err != nil {
 		return err
 	}
-	if int(b) < len(sm.opcodes) {
+	if b < sm.dbl.Prologue.OpcodeBase {
-		if b == 0 {
+		if int(b) < len(sm.opcodes) {
-			sm.lastOpcodeKind = extendedOpcode
+			sm.valid = false
+			sm.opcodes[b](sm, sm.buf)
 		} else {
-			sm.lastOpcodeKind = standardOpcode
+			// unimplemented standard opcode, read the number of arguments specified
-		}
+			// in the prologue and do nothing with them
-		sm.valid = false
+			opnum := sm.dbl.Prologue.StdOpLengths[b-1]
-		sm.opcodes[b](sm, sm.buf)
+			for i := 0; i < int(opnum); i++ {
-	} else if b < sm.dbl.Prologue.OpcodeBase {
+				util.DecodeSLEB128(sm.buf)
-		// unimplemented standard opcode, read the number of arguments specified
+			}
-		// in the prologue and do nothing with them
+			fmt.Printf("unknown opcode\n")
-		opnum := sm.dbl.Prologue.StdOpLengths[b-1]
-		for i := 0; i < int(opnum); i++ {
-			util.DecodeSLEB128(sm.buf)
 		}
 	} else {
 		execSpecialOpcode(sm, b)
@@ -361,8 +359,6 @@ func execSpecialOpcode(sm *StateMachine, instr byte) {
 	sm.lastDelta = int(sm.dbl.Prologue.LineBase + int8(decoded%sm.dbl.Prologue.LineRange))
 	sm.line += sm.lastDelta
 	sm.address += uint64(decoded/sm.dbl.Prologue.LineRange) * uint64(sm.dbl.Prologue.MinInstrLength)
-	sm.basicBlock = false
-	sm.lastOpcodeKind = specialOpcode
 	sm.valid = true
 }
 
@@ -375,7 +371,6 @@ func execExtendedOpcode(sm *StateMachine, buf *bytes.Buffer) {
 }
 
 func copyfn(sm *StateMachine, buf *bytes.Buffer) {
-	sm.basicBlock = false
 	sm.valid = true
 }
 

pkg/dwarf/line/state_machine_test.go

@@ -0,0 +1,128 @@
+package line
+
+import (
+	"bytes"
+	"compress/gzip"
+	"debug/dwarf"
+	"debug/macho"
+	"fmt"
+	"io"
+	"io/ioutil"
+	"os"
+	"runtime"
+	"testing"
+)
+
+func slurpGzip(path string) ([]byte, error) {
+	fh, err := os.Open(path)
+	if err != nil {
+		return nil, err
+	}
+	defer fh.Close()
+	gzin, err := gzip.NewReader(fh)
+	if err != nil {
+		return nil, err
+	}
+	defer gzin.Close()
+	return ioutil.ReadAll(gzin)
+}
+
+const (
+	newCompileUnit = "NEW COMPILE UNIT"
+	debugLineEnd   = "END"
+)
+
+func TestGrafana(t *testing.T) {
+	// Compares a full execution of our state machine on the debug_line section
+	// of grafana to the output generated using debug/dwarf.LineReader on the
+	// same section.
+
+	if runtime.GOOS == "windows" {
+		t.Skip("filepath.Join ruins this test on windows")
+	}
+
+	debugBytes, err := slurpGzip("_testdata/debug.grafana.debug.gz")
+	if err != nil {
+		t.Fatal(err)
+	}
+	exe, err := macho.NewFile(bytes.NewReader(debugBytes))
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	sec := exe.Section("__debug_line")
+	debugLineBytes, err := sec.Data()
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	data, err := exe.DWARF()
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	debugLineBuffer := bytes.NewBuffer(debugLineBytes)
+	rdr := data.Reader()
+	for {
+		e, err := rdr.Next()
+		if err != nil {
+			t.Fatal(err)
+		}
+		if e == nil {
+			break
+		}
+		rdr.SkipChildren()
+		if e.Tag != dwarf.TagCompileUnit {
+			continue
+		}
+		cuname, _ := e.Val(dwarf.AttrName).(string)
+
+		lineInfo := Parse(e.Val(dwarf.AttrCompDir).(string), debugLineBuffer, t.Logf)
+		sm := newStateMachine(lineInfo, lineInfo.Instructions)
+
+		lnrdr, err := data.LineReader(e)
+		if err != nil {
+			t.Fatal(err)
+		}
+
+		checkCompileUnit(t, cuname, lnrdr, sm)
+	}
+}
+
+func checkCompileUnit(t *testing.T, cuname string, lnrdr *dwarf.LineReader, sm *StateMachine) {
+	var lne dwarf.LineEntry
+	for {
+		if err := sm.next(); err != nil {
+			if err != io.EOF {
+				t.Fatalf("state machine next error: %v", err)
+			}
+			break
+		}
+		if !sm.valid {
+			continue
+		}
+
+		err := lnrdr.Next(&lne)
+		if err == io.EOF {
+			t.Fatalf("line reader ended before our state machine for compile unit %s", cuname)
+		}
+		if err != nil {
+			t.Fatal(err)
+		}
+
+		tgt := fmt.Sprintf("%#x %s:%d isstmt:%v prologue_end:%v epilogue_begin:%v", lne.Address, lne.File.Name, lne.Line, lne.IsStmt, lne.PrologueEnd, lne.EpilogueBegin)
+
+		out := fmt.Sprintf("%#x %s:%d isstmt:%v prologue_end:%v epilogue_begin:%v", sm.address, sm.file, sm.line, sm.isStmt, sm.prologueEnd, sm.epilogueBegin)
+		if out != tgt {
+			t.Errorf("mismatch:\n")
+			t.Errorf("got:\t%s\n", out)
+			t.Errorf("expected:\t%s\n", tgt)
+			t.Fatal("previous error")
+		}
+	}
+
+	err := lnrdr.Next(&lne)
+	if err != io.EOF {
+		t.Fatalf("state machine ended before the line reader for compile unit %s", cuname)
+	}
+}