encoding/asn1: don't parse invalid UTF-8.

Invalid UTF-8 triggers an error when marshaling but, previously, not
when unmarshaling. This means that ASN.1 structures were not
round-tripping.

This change makes invalid UTF-8 in a string marked as UTF-8 to be an
error when Unmarshaling.

Fixes #11126.

Change-Id: Ic37be84d21dc5c03983525e244d955a8b1e1ff14
Reviewed-on: https://go-review.googlesource.com/11056
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>

src/encoding/asn1/asn1.go

@@ -26,6 +26,7 @@ import (
 	"reflect"
 	"strconv"
 	"time"
+	"unicode/utf8"
 )
 
 // A StructuralError suggests that the ASN.1 data is valid, but the Go type
@@ -389,6 +390,9 @@ func parseT61String(bytes []byte) (ret string, err error) {
 // parseUTF8String parses a ASN.1 UTF8String (raw UTF-8) from the given byte
 // array and returns it.
 func parseUTF8String(bytes []byte) (ret string, err error) {
+	if !utf8.Valid(bytes) {
+		return "", errors.New("asn1: invalid UTF-8 string")
+	}
 	return string(bytes), nil
 }
 

src/encoding/asn1/asn1_test.go

@@ -9,6 +9,7 @@ import (
 	"fmt"
 	"math/big"
 	"reflect"
+	"strings"
 	"testing"
 	"time"
 )
@@ -922,3 +923,20 @@ func TestTruncatedExplicitTag(t *testing.T) {
 		t.Error("Unmarshal returned without error")
 	}
 }
+
+type invalidUTF8Test struct {
+	Str string `asn1:"utf8"`
+}
+
+func TestUnmarshalInvalidUTF8(t *testing.T) {
+	data := []byte("0\x05\f\x03a\xc9c")
+	var result invalidUTF8Test
+	_, err := Unmarshal(data, &result)
+
+	const expectedSubstring = "UTF"
+	if err == nil {
+		t.Fatal("Successfully unmarshaled invalid UTF-8 data")
+	} else if !strings.Contains(err.Error(), expectedSubstring) {
+		t.Fatalf("Expected error to mention %q but error was %q", expectedSubstring, err.Error())
+	}
+}