perf: Add AES key caching (#189)

* test: Add AES password benchmark

* perf: Add AES key caching

* style: Lint fixes

go.mod

@@ -7,6 +7,7 @@ require (
 	github.com/bodgit/plumbing v1.3.0
 	github.com/bodgit/windows v1.0.1
 	github.com/hashicorp/go-multierror v1.1.1
+	github.com/hashicorp/golang-lru/v2 v2.0.7
 	github.com/klauspost/compress v1.17.7
 	github.com/pierrec/lz4/v4 v4.1.21
 	github.com/stretchr/testify v1.9.0

go.sum

@@ -66,6 +66,8 @@ github.com/hashicorp/go-multierror v1.1.1 h1:H5DkEtf6CXdFp0N0Em5UCwQpXMWke8IA0+l
 github.com/hashicorp/go-multierror v1.1.1/go.mod h1:iw975J/qwKPdAO1clOe2L8331t/9/fmwbPZ6JB6eMoM=
 github.com/hashicorp/golang-lru v0.5.0/go.mod h1:/m3WP610KZHVQ1SGc6re/UDhFvYD7pJ4Ao+sR/qLZy8=
 github.com/hashicorp/golang-lru v0.5.1/go.mod h1:/m3WP610KZHVQ1SGc6re/UDhFvYD7pJ4Ao+sR/qLZy8=
+github.com/hashicorp/golang-lru/v2 v2.0.7 h1:a+bsQ5rvGLjzHuww6tVxozPZFVghXaHOwFs4luLUK2k=
+github.com/hashicorp/golang-lru/v2 v2.0.7/go.mod h1:QeFd9opnmA6QUJc5vARoKUSoFhyfM2/ZepoAG6RGpeM=
 github.com/ianlancetaylor/demangle v0.0.0-20181102032728-5e5cf60278f6/go.mod h1:aSSvb/t6k1mPoxDqO4vJh6VOCGPwU4O0C2/Eqndh1Sc=
 github.com/jstemmer/go-junit-report v0.0.0-20190106144839-af01ea7f8024/go.mod h1:6v2b51hI/fHJwM22ozAgKL4VKDeJcHhJFhtBdhmNjmU=
 github.com/jstemmer/go-junit-report v0.9.1/go.mod h1:Brl9GWCQeLvo8nXZwPNNblvFj/XSXhF0NWZEnDohbsk=

internal/aes7z/key.go

@@ -4,12 +4,47 @@ import (
 	"bytes"
 	"crypto/sha256"
 	"encoding/binary"
+	"encoding/hex"
 
+	lru "github.com/hashicorp/golang-lru/v2"
+	"go4.org/syncutil"
 	"golang.org/x/text/encoding/unicode"
 	"golang.org/x/text/transform"
 )
 
-func calculateKey(password string, cycles int, salt []byte) []byte {
+type cacheKey struct {
+	password string
+	cycles   int
+	salt     string // []byte isn't comparable
+}
+
+const cacheSize = 10
+
+//nolint:gochecknoglobals
+var (
+	once  syncutil.Once
+	cache *lru.Cache[cacheKey, []byte]
+)
+
+func calculateKey(password string, cycles int, salt []byte) ([]byte, error) {
+	if err := once.Do(func() (err error) {
+		cache, err = lru.New[cacheKey, []byte](cacheSize)
+
+		return
+	}); err != nil {
+		return nil, err
+	}
+
+	ck := cacheKey{
+		password: password,
+		cycles:   cycles,
+		salt:     hex.EncodeToString(salt),
+	}
+
+	if key, ok := cache.Get(ck); ok {
+		return key, nil
+	}
+
 	b := bytes.NewBuffer(salt)
 
 	// Convert password to UTF-16LE
@@ -30,5 +65,7 @@ func calculateKey(password string, cycles int, salt []byte) []byte {
 		copy(key, h.Sum(nil))
 	}
 
-	return key
+	_ = cache.Add(ck, key)
+
+	return key, nil
 }

internal/aes7z/reader.go

@@ -29,7 +29,12 @@ func (rc *readCloser) Close() error {
 }
 
 func (rc *readCloser) Password(p string) error {
-	block, err := aes.NewCipher(calculateKey(p, rc.cycles, rc.salt))
+	key, err := calculateKey(p, rc.cycles, rc.salt)
+	if err != nil {
+		return err
+	}
+
+	block, err := aes.NewCipher(key)
 	if err != nil {
 		return err
 	}

internal/zstd/reader.go

@@ -52,6 +52,7 @@ func NewReader(_ []byte, _ uint64, readers []io.ReadCloser) (io.ReadCloser, erro
 		if r, err = zstd.NewReader(readers[0]); err != nil {
 			return nil, err
 		}
+
 		runtime.SetFinalizer(r, (*zstd.Decoder).Close)
 	}
 

reader.go

@@ -544,7 +544,7 @@ func toValidName(name string) string {
 	return p
 }
 
-//nolint:cyclop
+//nolint:cyclop,funlen
 func (z *Reader) initFileList() {
 	z.fileListOnce.Do(func() {
 		files := make(map[string]int)
@@ -583,12 +583,14 @@ func (z *Reader) initFileList() {
 				isDir: isDir,
 			}
 			z.fileList = append(z.fileList, entry)
+
 			if isDir {
 				knownDirs[name] = idx
 			} else {
 				files[name] = idx
 			}
 		}
+
 		for dir := range dirs {
 			if _, ok := knownDirs[dir]; !ok {
 				if idx, ok := files[dir]; ok {

reader_test.go

@@ -188,6 +188,7 @@ func TestOpenReader(t *testing.T) {
 
 		t.Run(table.name, func(t *testing.T) {
 			t.Parallel()
+
 			r, err := sevenzip.OpenReader(filepath.Join("testdata", table.file))
 			if err != nil {
 				assert.ErrorIs(t, err, table.err)
@@ -243,6 +244,7 @@ func TestOpenReaderWithPassword(t *testing.T) {
 
 		t.Run(table.name, func(t *testing.T) {
 			t.Parallel()
+
 			r, err := sevenzip.OpenReaderWithPassword(filepath.Join("testdata", table.file), table.password)
 			if err != nil {
 				t.Fatal(err)
@@ -362,13 +364,13 @@ func benchmarkArchiveNaiveParallel(b *testing.B, file string, workers int) {
 	}
 }
 
-func benchmarkArchive(b *testing.B, file string, optimised bool) {
+func benchmarkArchive(b *testing.B, file, password string, optimised bool) {
 	b.Helper()
 
 	h := crc32.NewIEEE()
 
 	for n := 0; n < b.N; n++ {
-		r, err := sevenzip.OpenReader(filepath.Join("testdata", file))
+		r, err := sevenzip.OpenReaderWithPassword(filepath.Join("testdata", file), password)
 		if err != nil {
 			b.Fatal(err)
 		}
@@ -382,56 +384,60 @@ func benchmarkArchive(b *testing.B, file string, optimised bool) {
 	}
 }
 
+func BenchmarkAES7z(b *testing.B) {
+	benchmarkArchive(b, "aes7z.7z", "password", true)
+}
+
 func BenchmarkBzip2(b *testing.B) {
-	benchmarkArchive(b, "bzip2.7z", true)
+	benchmarkArchive(b, "bzip2.7z", "", true)
 }
 
 func BenchmarkCopy(b *testing.B) {
-	benchmarkArchive(b, "copy.7z", true)
+	benchmarkArchive(b, "copy.7z", "", true)
 }
 
 func BenchmarkDeflate(b *testing.B) {
-	benchmarkArchive(b, "deflate.7z", true)
+	benchmarkArchive(b, "deflate.7z", "", true)
 }
 
 func BenchmarkDelta(b *testing.B) {
-	benchmarkArchive(b, "delta.7z", true)
+	benchmarkArchive(b, "delta.7z", "", true)
 }
 
 func BenchmarkLZMA(b *testing.B) {
-	benchmarkArchive(b, "lzma.7z", true)
+	benchmarkArchive(b, "lzma.7z", "", true)
 }
 
 func BenchmarkLZMA2(b *testing.B) {
-	benchmarkArchive(b, "lzma2.7z", true)
+	benchmarkArchive(b, "lzma2.7z", "", true)
 }
 
 func BenchmarkBCJ2(b *testing.B) {
-	benchmarkArchive(b, "bcj2.7z", true)
+	benchmarkArchive(b, "bcj2.7z", "", true)
 }
 
 func BenchmarkComplex(b *testing.B) {
-	benchmarkArchive(b, "lzma1900.7z", true)
+	benchmarkArchive(b, "lzma1900.7z", "", true)
 }
 
 func BenchmarkLZ4(b *testing.B) {
-	benchmarkArchive(b, "lz4.7z", true)
+	benchmarkArchive(b, "lz4.7z", "", true)
 }
 
 func BenchmarkBrotli(b *testing.B) {
-	benchmarkArchive(b, "brotli.7z", true)
+	benchmarkArchive(b, "brotli.7z", "", true)
 }
 
 func BenchmarkZstandard(b *testing.B) {
-	benchmarkArchive(b, "zstd.7z", true)
+	benchmarkArchive(b, "zstd.7z", "", true)
 }
 
 func BenchmarkNaiveReader(b *testing.B) {
-	benchmarkArchive(b, "lzma1900.7z", false)
+	benchmarkArchive(b, "lzma1900.7z", "", false)
 }
 
 func BenchmarkOptimisedReader(b *testing.B) {
-	benchmarkArchive(b, "lzma1900.7z", true)
+	benchmarkArchive(b, "lzma1900.7z", "", true)
 }
 
 func BenchmarkNaiveParallelReader(b *testing.B) {
@@ -447,17 +453,17 @@ func BenchmarkParallelReader(b *testing.B) {
 }
 
 func BenchmarkBCJ(b *testing.B) {
-	benchmarkArchive(b, "bcj.7z", true)
+	benchmarkArchive(b, "bcj.7z", "", true)
 }
 
 func BenchmarkPPC(b *testing.B) {
-	benchmarkArchive(b, "ppc.7z", true)
+	benchmarkArchive(b, "ppc.7z", "", true)
 }
 
 func BenchmarkARM(b *testing.B) {
-	benchmarkArchive(b, "arm.7z", true)
+	benchmarkArchive(b, "arm.7z", "", true)
 }
 
 func BenchmarkSPARC(b *testing.B) {
-	benchmarkArchive(b, "sparc.7z", true)
+	benchmarkArchive(b, "sparc.7z", "", true)
 }