prog: factor out execution log parsing functionality

It will be needed to reproduction tool.

3 files changed, 158 insertions, 35 deletions

diff --git a/prog/parse.go b/prog/parse.go
new file mode 100644
index 000000000..e4a4e15ad
--- /dev/null
+++ b/prog/parse.go
@@ -0,0 +1,68 @@
+// Copyright 2015 syzkaller project authors. All rights reserved.
+// Use of this source code is governed by Apache 2 LICENSE that can be found in the LICENSE file.
+
+package prog
+
+import (
+	"bytes"
+	"strconv"
+)
+
+// LogEntry describes one program in execution log.
+type LogEntry struct {
+	P     *Prog
+	Proc  int // index of parallel proc
+	Start int // start offset in log
+	End   int // end offset in log
+}
+
+func ParseLog(data []byte) []*LogEntry {
+	var entries []*LogEntry
+	ent := &LogEntry{}
+	var cur []byte
+	for pos := 0; pos < len(data); {
+		nl := bytes.IndexByte(data[pos:], '\n')
+		if nl == -1 {
+			nl = len(data)
+		} else {
+			nl += pos
+		}
+		line := data[pos : nl+1]
+		pos0 := pos
+		pos = nl + 1
+		const delim = "executing program "
+		if delimPos := bytes.Index(line, []byte(delim)); delimPos != -1 {
+			if ent.P != nil && len(ent.P.Calls) != 0 {
+				ent.End = pos0
+				entries = append(entries, ent)
+			}
+			procStart := delimPos + len(delim)
+			procEnd := procStart
+			for procEnd != len(line) && line[procEnd] >= '0' && line[procEnd] <= '9' {
+				procEnd++
+			}
+			proc, _ := strconv.Atoi(string(line[procStart:procEnd]))
+			ent = &LogEntry{
+				Proc:  proc,
+				Start: pos0,
+			}
+			cur = nil
+			continue
+		}
+		if ent == nil {
+			continue
+		}
+		tmp := append(cur, line...)
+		p, err := Deserialize(tmp)
+		if err != nil {
+			continue
+		}
+		cur = tmp
+		ent.P = p
+	}
+	if ent.P != nil && len(ent.P.Calls) != 0 {
+		ent.End = len(data)
+		entries = append(entries, ent)
+	}
+	return entries
+}
diff --git a/prog/parse_test.go b/prog/parse_test.go
new file mode 100644
index 000000000..93aa770cd
--- /dev/null
+++ b/prog/parse_test.go
@@ -0,0 +1,81 @@
+// Copyright 2015 syzkaller project authors. All rights reserved.
+// Use of this source code is governed by Apache 2 LICENSE that can be found in the LICENSE file.
+
+package prog
+
+import (
+	"testing"
+)
+
+func TestParseSingle(t *testing.T) {
+	const execLog = `getpid()
+gettid()	
+`
+	entries := ParseLog([]byte(execLog))
+	if len(entries) != 1 {
+		t.Fatalf("got %v programs, want 1", len(entries))
+	}
+	ent := entries[0]
+	if ent.Start != 0 {
+		t.Fatalf("start offset %v, want 0", ent.Start)
+	}
+	if ent.End != len(execLog) {
+		t.Fatalf("end offset %v, want %v", ent.End, len(execLog))
+	}
+	if ent.Proc != 0 {
+		t.Fatalf("proc %v, want 0", ent.Proc)
+	}
+	want := "getpid-gettid"
+	got := ent.P.String()
+	if got != want {
+		t.Fatalf("bad program: %s, want %s", got, want)
+	}
+}
+
+func TestParseMulti(t *testing.T) {
+	entries := ParseLog([]byte(execLog))
+	if len(entries) != 4 {
+		t.Fatalf("got %v programs, want 4")
+	}
+	off := 0
+	for _, ent := range entries {
+		if off > ent.Start || ent.Start > ent.End || ent.End > len(execLog) {
+			t.Fatalf("bad offsets")
+		}
+	}
+	if entries[0].Proc != 1 || entries[1].Proc != 2 || entries[2].Proc != 33 || entries[3].Proc != 9 {
+		t.Fatalf("bad procs")
+	}
+	if s := entries[0].P.String(); s != "getpid-gettid-munlockall" {
+		t.Fatalf("bad program 0: %s", s)
+	}
+	if s := entries[1].P.String(); s != "getpid-gettid" {
+		t.Fatalf("bad program 1: %s", s)
+	}
+	if s := entries[2].P.String(); s != "gettid-getpid" {
+		t.Fatalf("bad program 2: %s", s)
+	}
+	if s := entries[3].P.String(); s != "munlockall" {
+		t.Fatalf("bad program 3: %s", s)
+	}
+}
+
+const execLog = `
+getpid()
+getpid()
+2015/12/21 12:18:05 executing program 1:
+getpid()
+[ 2351.935478] Modules linked in:
+gettid()
+munlockall()
+2015/12/21 12:18:05 executing program 2:
+[ 2351.935478] Modules linked in:
+getpid()
+gettid()
+2015/12/21 12:18:05 executing program 33:
+gettid()
+getpid()
+[ 2351.935478] Modules linked in:
+2015/12/21 12:18:05 executing program 9:
+munlockall()
+`
diff --git a/tools/syz-execprog/execprog.go b/tools/syz-execprog/execprog.go
index b8dd6ee9a..a637d47ff 100644
--- a/tools/syz-execprog/execprog.go
+++ b/tools/syz-execprog/execprog.go
@@ -6,7 +6,6 @@
 package main
 
 import (
-	"bufio"
 	"bytes"
 	"encoding/binary"
 	"flag"
@@ -41,14 +40,21 @@ var (
 func main() {
 	flag.Parse()
 	if len(flag.Args()) == 0 {
-		fmt.Fprintf(os.Stderr, "usage: execprog [flags] file-with-programs*\n")
+		fmt.Fprintf(os.Stderr, "usage: execprog [flags] file-with-programs+\n")
 		flag.PrintDefaults()
 		os.Exit(1)
 	}
 
 	var progs []*prog.Prog
 	for _, fn := range flag.Args() {
-		progs = append(progs, parseFile(fn)...)
+		data, err := ioutil.ReadFile(fn)
+		if err != nil {
+			log.Fatalf("failed to read log file: %v", err)
+		}
+		entries := prog.ParseLog(data)
+		for _, ent := range entries {
+			progs = append(progs, ent.P)
+		}
 	}
 	log.Printf("parsed %v programs", len(progs))
 	if len(progs) == 0 {
@@ -142,35 +148,3 @@ func main() {
 	}
 	wg.Wait()
 }
-
-func parseFile(fn string) []*prog.Prog {
-	logf, err := os.Open(fn)
-	if err != nil {
-		log.Fatalf("failed to open log file: %v", err)
-	}
-	log.Printf("parsing log %v", fn)
-	s := bufio.NewScanner(logf)
-	var cur []byte
-	var last *prog.Prog
-	var progs []*prog.Prog
-	for s.Scan() {
-		ln := s.Text()
-		tmp := append(cur, ln...)
-		tmp = append(tmp, '\n')
-		p, err := prog.Deserialize(tmp)
-		if err == nil {
-			cur = tmp
-			last = p
-			continue
-		}
-		if last != nil {
-			progs = append(progs, last)
-			last = nil
-			cur = cur[:0]
-		}
-	}
-	if last != nil {
-		progs = append(progs, last)
-	}
-	return progs
-}