From 7f145a7e1057f2f2a099827ab9d7e66a31cd1868 Mon Sep 17 00:00:00 2001 From: "dependabot[bot]" <49699333+dependabot[bot]@users.noreply.github.com> Date: Mon, 6 Sep 2021 06:23:51 -0400 Subject: [PATCH] Bump github.com/containerd/stargz-snapshotter/estargz (#431) Bumps [github.com/containerd/stargz-snapshotter/estargz](https://github.com/containerd/stargz-snapshotter) from 0.7.0 to 0.8.0. - [Release notes](https://github.com/containerd/stargz-snapshotter/releases) - [Commits](https://github.com/containerd/stargz-snapshotter/compare/v0.7.0...v0.8.0) --- updated-dependencies: - dependency-name: github.com/containerd/stargz-snapshotter/estargz dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> --- go.mod | 4 +- go.sum | 9 +- .../stargz-snapshotter/estargz/build.go | 61 +- .../stargz-snapshotter/estargz/estargz.go | 436 ++-- .../stargz-snapshotter/estargz/go.mod | 2 +- .../stargz-snapshotter/estargz/go.sum | 6 +- .../stargz-snapshotter/estargz/gzip.go | 216 ++ .../stargz-snapshotter/estargz/testutil.go | 1933 +++++++++++++++++ .../stargz-snapshotter/estargz/types.go | 54 +- vendor/github.com/golang/snappy/.gitignore | 16 - vendor/github.com/golang/snappy/AUTHORS | 18 - vendor/github.com/golang/snappy/CONTRIBUTORS | 41 - vendor/github.com/golang/snappy/README | 107 - .../github.com/golang/snappy/decode_amd64.s | 490 ----- .../github.com/golang/snappy/decode_arm64.s | 494 ----- vendor/github.com/golang/snappy/decode_asm.go | 15 - .../github.com/golang/snappy/encode_amd64.s | 730 ------- .../github.com/golang/snappy/encode_arm64.s | 722 ------ vendor/github.com/golang/snappy/encode_asm.go | 30 - vendor/github.com/golang/snappy/go.mod | 1 - .../klauspost/compress/.gitattributes | 2 + .../github.com/klauspost/compress/.gitignore | 25 + .../klauspost/compress/.goreleaser.yml | 137 ++ vendor/github.com/klauspost/compress/LICENSE | 264 +++ .../github.com/klauspost/compress/README.md | 430 ++++ .../klauspost/compress/compressible.go | 85 + vendor/github.com/klauspost/compress/gen.sh | 4 + vendor/github.com/klauspost/compress/go.mod | 3 + vendor/github.com/klauspost/compress/go.sum | 0 .../klauspost/compress/huff0/compress.go | 64 + .../klauspost/compress/huff0/huff0.go | 62 + .../compress/internal/snapref}/LICENSE | 0 .../compress/internal/snapref}/decode.go | 2 +- .../internal/snapref}/decode_other.go | 4 +- .../compress/internal/snapref}/encode.go | 2 +- .../internal/snapref}/encode_other.go | 4 +- .../compress/internal/snapref}/snappy.go | 4 +- vendor/github.com/klauspost/compress/s2sx.mod | 4 + vendor/github.com/klauspost/compress/s2sx.sum | 0 .../klauspost/compress/zstd/README.md | 32 +- .../klauspost/compress/zstd/decoder.go | 5 +- .../klauspost/compress/zstd/enc_base.go | 4 +- .../klauspost/compress/zstd/enc_best.go | 161 +- .../klauspost/compress/zstd/enc_better.go | 64 +- .../klauspost/compress/zstd/enc_dfast.go | 61 +- .../klauspost/compress/zstd/enc_fast.go | 37 +- .../klauspost/compress/zstd/encoder.go | 31 +- .../compress/zstd/encoder_options.go | 2 +- .../klauspost/compress/zstd/hash.go | 60 +- .../zstd/internal/xxhash/xxhash_amd64.go | 5 +- .../zstd/internal/xxhash/xxhash_other.go | 1 + .../klauspost/compress/zstd/snappy.go | 2 +- .../github.com/klauspost/compress/zstd/zip.go | 9 +- vendor/modules.txt | 10 +- 54 files changed, 3852 insertions(+), 3113 deletions(-) create mode 100644 vendor/github.com/containerd/stargz-snapshotter/estargz/gzip.go create mode 100644 vendor/github.com/containerd/stargz-snapshotter/estargz/testutil.go delete mode 100644 vendor/github.com/golang/snappy/.gitignore delete mode 100644 vendor/github.com/golang/snappy/AUTHORS delete mode 100644 vendor/github.com/golang/snappy/CONTRIBUTORS delete mode 100644 vendor/github.com/golang/snappy/README delete mode 100644 vendor/github.com/golang/snappy/decode_amd64.s delete mode 100644 vendor/github.com/golang/snappy/decode_arm64.s delete mode 100644 vendor/github.com/golang/snappy/decode_asm.go delete mode 100644 vendor/github.com/golang/snappy/encode_amd64.s delete mode 100644 vendor/github.com/golang/snappy/encode_arm64.s delete mode 100644 vendor/github.com/golang/snappy/encode_asm.go delete mode 100644 vendor/github.com/golang/snappy/go.mod create mode 100644 vendor/github.com/klauspost/compress/.gitattributes create mode 100644 vendor/github.com/klauspost/compress/.gitignore create mode 100644 vendor/github.com/klauspost/compress/.goreleaser.yml create mode 100644 vendor/github.com/klauspost/compress/README.md create mode 100644 vendor/github.com/klauspost/compress/compressible.go create mode 100644 vendor/github.com/klauspost/compress/gen.sh create mode 100644 vendor/github.com/klauspost/compress/go.mod create mode 100644 vendor/github.com/klauspost/compress/go.sum rename vendor/github.com/{golang/snappy => klauspost/compress/internal/snapref}/LICENSE (100%) rename vendor/github.com/{golang/snappy => klauspost/compress/internal/snapref}/decode.go (99%) rename vendor/github.com/{golang/snappy => klauspost/compress/internal/snapref}/decode_other.go (98%) rename vendor/github.com/{golang/snappy => klauspost/compress/internal/snapref}/encode.go (99%) rename vendor/github.com/{golang/snappy => klauspost/compress/internal/snapref}/encode_other.go (99%) rename vendor/github.com/{golang/snappy => klauspost/compress/internal/snapref}/snappy.go (96%) create mode 100644 vendor/github.com/klauspost/compress/s2sx.mod create mode 100644 vendor/github.com/klauspost/compress/s2sx.sum diff --git a/go.mod b/go.mod index 2fd86d1f3a..2738cb0d71 100644 --- a/go.mod +++ b/go.mod @@ -3,17 +3,15 @@ module github.com/google/ko go 1.15 require ( - github.com/containerd/stargz-snapshotter/estargz v0.7.0 + github.com/containerd/stargz-snapshotter/estargz v0.8.0 github.com/cpuguy83/go-md2man/v2 v2.0.1 // indirect github.com/docker/docker v20.10.8+incompatible github.com/dprotaso/go-yit v0.0.0-20191028211022-135eb7262960 github.com/evanphx/json-patch/v5 v5.5.0 // indirect github.com/fsnotify/fsnotify v1.5.1 github.com/go-training/helloworld v0.0.0-20200225145412-ba5f4379d78b - github.com/golang/snappy v0.0.4 // indirect github.com/google/go-cmp v0.5.6 github.com/google/go-containerregistry v0.6.0 - github.com/klauspost/compress v1.13.1 // indirect github.com/mattmoor/dep-notify v0.0.0-20190205035814-a45dec370a17 github.com/mattn/go-isatty v0.0.13 // indirect github.com/opencontainers/image-spec v1.0.2-0.20210730191737-8e42a01fb1b7 diff --git a/go.sum b/go.sum index 2878008392..0e854fdf55 100644 --- a/go.sum +++ b/go.sum @@ -185,8 +185,9 @@ github.com/containerd/imgcrypt v1.1.1/go.mod h1:xpLnwiQmEUJPvQoAapeb2SNCxz7Xr6PJ github.com/containerd/nri v0.0.0-20201007170849-eb1350a75164/go.mod h1:+2wGSDGFYfE5+So4M5syatU0N0f0LbWpuqyMi4/BE8c= github.com/containerd/nri v0.0.0-20210316161719-dbaa18c31c14/go.mod h1:lmxnXF6oMkbqs39FiCt1s0R2HSMhcLel9vNL3m4AaeY= github.com/containerd/nri v0.1.0/go.mod h1:lmxnXF6oMkbqs39FiCt1s0R2HSMhcLel9vNL3m4AaeY= -github.com/containerd/stargz-snapshotter/estargz v0.7.0 h1:1d/rydzTywc76lnjJb6qbPCiTiCwts49AzKps/Ecblw= github.com/containerd/stargz-snapshotter/estargz v0.7.0/go.mod h1:83VWDqHnurTKliEB0YvWMiCfLDwv4Cjj1X9Vk98GJZw= +github.com/containerd/stargz-snapshotter/estargz v0.8.0 h1:oA1wx8kTFfImfsT5bScbrZd8gK+WtQnn15q82Djvm0Y= +github.com/containerd/stargz-snapshotter/estargz v0.8.0/go.mod h1:mwIwuwb+D8FX2t45Trwi0hmWmZm5VW7zPP/rekwhWQU= github.com/containerd/ttrpc v0.0.0-20190828154514-0e0f228740de/go.mod h1:PvCDdDGpgqzQIzDW1TphrGLssLDZp2GuS+X5DkEJB8o= github.com/containerd/ttrpc v0.0.0-20190828172938-92c8520ef9f8/go.mod h1:PvCDdDGpgqzQIzDW1TphrGLssLDZp2GuS+X5DkEJB8o= github.com/containerd/ttrpc v0.0.0-20191028202541-4f1b8fe65a5c/go.mod h1:LPm1u0xBw8r8NOKoOdNMeVHSawSsltak+Ihv+etqsE8= @@ -366,8 +367,6 @@ github.com/golang/protobuf v1.5.1/go.mod h1:DopwsBzvsk0Fs44TXzsVbJyPhcCPeIwnvohx github.com/golang/protobuf v1.5.2 h1:ROPKBNFfQgOUMifHyP+KYbvpjbdoFNs+aK7DXlji0Tw= github.com/golang/protobuf v1.5.2/go.mod h1:XVQd3VNwM+JqD3oG2Ue2ip4fOMUkwXdXDdiuN0vRsmY= github.com/golang/snappy v0.0.3/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEWrmP2Q= -github.com/golang/snappy v0.0.4 h1:yAGX7huGHXlcLOEtBnF4w7FQwA26wojNCwOYAEhLjQM= -github.com/golang/snappy v0.0.4/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEWrmP2Q= github.com/google/btree v0.0.0-20180813153112-4030bb1f1f0c/go.mod h1:lNA+9X1NB3Zf8V7Ke586lFgjr2dZNuvo3lPJSGZ5JPQ= github.com/google/btree v1.0.0/go.mod h1:lNA+9X1NB3Zf8V7Ke586lFgjr2dZNuvo3lPJSGZ5JPQ= github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M= @@ -483,8 +482,8 @@ github.com/klauspost/compress v1.11.3/go.mod h1:aoV0uJVorq1K+umq18yTdKaF57EivdYs github.com/klauspost/compress v1.11.13/go.mod h1:aoV0uJVorq1K+umq18yTdKaF57EivdYsUV+/s2qKfXs= github.com/klauspost/compress v1.12.3/go.mod h1:8dP1Hq4DHOhN9w426knH3Rhby4rFm6D8eO+e+Dq5Gzg= github.com/klauspost/compress v1.13.0/go.mod h1:8dP1Hq4DHOhN9w426knH3Rhby4rFm6D8eO+e+Dq5Gzg= -github.com/klauspost/compress v1.13.1 h1:wXr2uRxZTJXHLly6qhJabee5JqIhTRoLBhDOA74hDEQ= -github.com/klauspost/compress v1.13.1/go.mod h1:8dP1Hq4DHOhN9w426knH3Rhby4rFm6D8eO+e+Dq5Gzg= +github.com/klauspost/compress v1.13.5 h1:9O69jUPDcsT9fEm74W92rZL9FQY7rCdaXVneq+yyzl4= +github.com/klauspost/compress v1.13.5/go.mod h1:/3/Vjq9QcHkK5uEr5lBEmyoZ1iFhe47etQ6QUkpK6sk= github.com/konsorten/go-windows-terminal-sequences v1.0.1/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ= github.com/konsorten/go-windows-terminal-sequences v1.0.2/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ= github.com/konsorten/go-windows-terminal-sequences v1.0.3/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ= diff --git a/vendor/github.com/containerd/stargz-snapshotter/estargz/build.go b/vendor/github.com/containerd/stargz-snapshotter/estargz/build.go index fc42b6c348..708b266899 100644 --- a/vendor/github.com/containerd/stargz-snapshotter/estargz/build.go +++ b/vendor/github.com/containerd/stargz-snapshotter/estargz/build.go @@ -26,7 +26,6 @@ import ( "archive/tar" "bytes" "compress/gzip" - "encoding/json" "fmt" "io" "io/ioutil" @@ -48,6 +47,7 @@ type options struct { compressionLevel int prioritizedFiles []string missedPrioritizedFiles *[]string + compression Compression } type Option func(o *options) error @@ -95,6 +95,15 @@ func WithAllowPrioritizeNotFound(missedFiles *[]string) Option { } } +// WithCompression specifies compression algorithm to be used. +// Default is gzip. +func WithCompression(compression Compression) Option { + return func(o *options) error { + o.compression = compression + return nil + } +} + // Blob is an eStargz blob. type Blob struct { io.ReadCloser @@ -126,6 +135,9 @@ func Build(tarBlob *io.SectionReader, opt ...Option) (_ *Blob, rErr error) { return nil, err } } + if opts.compression == nil { + opts.compression = newGzipCompressionWithLevel(opts.compressionLevel) + } layerFiles := newTempFiles() defer func() { if rErr != nil { @@ -155,7 +167,7 @@ func Build(tarBlob *io.SectionReader, opt ...Option) (_ *Blob, rErr error) { if err != nil { return err } - sw := NewWriterLevel(esgzFile, opts.compressionLevel) + sw := NewWriterWithCompressor(esgzFile, opts.compression) sw.ChunkSize = opts.chunkSize if err := sw.AppendTar(readerFromEntries(parts...)); err != nil { return err @@ -187,11 +199,12 @@ func Build(tarBlob *io.SectionReader, opt ...Option) (_ *Blob, rErr error) { diffID := digest.Canonical.Digester() pr, pw := io.Pipe() go func() { - r, err := gzip.NewReader(io.TeeReader(io.MultiReader(append(rs, tocAndFooter)...), pw)) + r, err := opts.compression.Reader(io.TeeReader(io.MultiReader(append(rs, tocAndFooter)...), pw)) if err != nil { pw.CloseWithError(err) return } + defer r.Close() if _, err := io.Copy(diffID.Hash(), r); err != nil { pw.CloseWithError(err) return @@ -213,7 +226,7 @@ func Build(tarBlob *io.SectionReader, opt ...Option) (_ *Blob, rErr error) { // Writers doesn't write TOC and footer to the underlying writers so they can be // combined into a single eStargz and tocAndFooter returned by this function can // be appended at the tail of that combined blob. -func closeWithCombine(compressionLevel int, ws ...*Writer) (tocAndFooter io.Reader, tocDgst digest.Digest, err error) { +func closeWithCombine(compressionLevel int, ws ...*Writer) (tocAndFooterR io.Reader, tocDgst digest.Digest, err error) { if len(ws) == 0 { return nil, "", fmt.Errorf("at least one writer must be passed") } @@ -230,7 +243,7 @@ func closeWithCombine(compressionLevel int, ws ...*Writer) (tocAndFooter io.Read } } var ( - mtoc = new(jtoc) + mtoc = new(JTOC) currentOffset int64 ) mtoc.Version = ws[0].toc.Version @@ -248,40 +261,16 @@ func closeWithCombine(compressionLevel int, ws ...*Writer) (tocAndFooter io.Read currentOffset += w.cw.n } - tocJSON, err := json.MarshalIndent(mtoc, "", "\t") + return tocAndFooter(ws[0].compressor, mtoc, currentOffset) +} + +func tocAndFooter(compressor Compressor, toc *JTOC, offset int64) (io.Reader, digest.Digest, error) { + buf := new(bytes.Buffer) + tocDigest, err := compressor.WriteTOCAndFooter(buf, offset, toc, nil) if err != nil { return nil, "", err } - pr, pw := io.Pipe() - go func() { - zw, _ := gzip.NewWriterLevel(pw, compressionLevel) - tw := tar.NewWriter(zw) - if err := tw.WriteHeader(&tar.Header{ - Typeflag: tar.TypeReg, - Name: TOCTarName, - Size: int64(len(tocJSON)), - }); err != nil { - pw.CloseWithError(err) - return - } - if _, err := tw.Write(tocJSON); err != nil { - pw.CloseWithError(err) - return - } - if err := tw.Close(); err != nil { - pw.CloseWithError(err) - return - } - if err := zw.Close(); err != nil { - pw.CloseWithError(err) - return - } - pw.Close() - }() - return io.MultiReader( - pr, - bytes.NewReader(footerBytes(currentOffset)), - ), digest.FromBytes(tocJSON), nil + return buf, tocDigest, nil } // divideEntries divides passed entries to the parts at least the number specified by the diff --git a/vendor/github.com/containerd/stargz-snapshotter/estargz/estargz.go b/vendor/github.com/containerd/stargz-snapshotter/estargz/estargz.go index c45a7aacaf..e997d9cceb 100644 --- a/vendor/github.com/containerd/stargz-snapshotter/estargz/estargz.go +++ b/vendor/github.com/containerd/stargz-snapshotter/estargz/estargz.go @@ -28,8 +28,6 @@ import ( "bytes" "compress/gzip" "crypto/sha256" - "encoding/binary" - "encoding/json" "fmt" "hash" "io" @@ -37,7 +35,6 @@ import ( "os" "path" "sort" - "strconv" "strings" "sync" "time" @@ -50,7 +47,7 @@ import ( // A Reader permits random access reads from a stargz file. type Reader struct { sr *io.SectionReader - toc *jtoc + toc *JTOC tocDigest digest.Digest // m stores all non-chunk entries, keyed by name. @@ -60,39 +57,116 @@ type Reader struct { // are split up. For a file with a single chunk, it's only // stored in m. chunks map[string][]*TOCEntry + + decompressor Decompressor +} + +type openOpts struct { + tocOffset int64 + decompressors []Decompressor + telemetry *Telemetry +} + +// OpenOption is an option used during opening the layer +type OpenOption func(o *openOpts) error + +// WithTOCOffset option specifies the offset of TOC +func WithTOCOffset(tocOffset int64) OpenOption { + return func(o *openOpts) error { + o.tocOffset = tocOffset + return nil + } +} + +// WithDecompressors option specifies decompressors to use. +// Default is gzip-based decompressor. +func WithDecompressors(decompressors ...Decompressor) OpenOption { + return func(o *openOpts) error { + o.decompressors = decompressors + return nil + } +} + +// WithTelemetry option specifies the telemetry hooks +func WithTelemetry(telemetry *Telemetry) OpenOption { + return func(o *openOpts) error { + o.telemetry = telemetry + return nil + } +} + +// A func which takes start time and records the diff +type MeasureLatencyHook func(time.Time) + +// A struct which defines telemetry hooks. By implementing these hooks you should be able to record +// the latency metrics of the respective steps of estargz open operation. To be used with estargz.OpenWithTelemetry(...) +type Telemetry struct { + GetFooterLatency MeasureLatencyHook // measure time to get stargz footer (in milliseconds) + GetTocLatency MeasureLatencyHook // measure time to GET TOC JSON (in milliseconds) + DeserializeTocLatency MeasureLatencyHook // measure time to deserialize TOC JSON (in milliseconds) } // Open opens a stargz file for reading. +// The behaviour is configurable using options. // // Note that each entry name is normalized as the path that is relative to root. -func Open(sr *io.SectionReader) (*Reader, error) { - tocOff, footerSize, err := OpenFooter(sr) - if err != nil { - return nil, errors.Wrapf(err, "error parsing footer") +func Open(sr *io.SectionReader, opt ...OpenOption) (*Reader, error) { + var opts openOpts + for _, o := range opt { + if err := o(&opts); err != nil { + return nil, err + } } - tocTargz := make([]byte, sr.Size()-tocOff-footerSize) - if _, err := sr.ReadAt(tocTargz, tocOff); err != nil { - return nil, fmt.Errorf("error reading %d byte TOC targz: %v", len(tocTargz), err) + + gzipCompressors := []Decompressor{new(GzipDecompressor), new(legacyGzipDecompressor)} + decompressors := append(gzipCompressors, opts.decompressors...) + + // Determine the size to fetch. Try to fetch as many bytes as possible. + fetchSize := maxFooterSize(sr.Size(), decompressors...) + if maybeTocOffset := opts.tocOffset; maybeTocOffset > fetchSize { + if maybeTocOffset > sr.Size() { + return nil, fmt.Errorf("blob size %d is smaller than the toc offset", sr.Size()) + } + fetchSize = sr.Size() - maybeTocOffset } - zr, err := gzip.NewReader(bytes.NewReader(tocTargz)) - if err != nil { - return nil, fmt.Errorf("malformed TOC gzip header: %v", err) + + start := time.Now() // before getting layer footer + footer := make([]byte, fetchSize) + if _, err := sr.ReadAt(footer, sr.Size()-fetchSize); err != nil { + return nil, fmt.Errorf("error reading footer: %v", err) } - zr.Multistream(false) - tr := tar.NewReader(zr) - h, err := tr.Next() - if err != nil { - return nil, fmt.Errorf("failed to find tar header in TOC gzip stream: %v", err) + if opts.telemetry != nil && opts.telemetry.GetFooterLatency != nil { + opts.telemetry.GetFooterLatency(start) } - if h.Name != TOCTarName { - return nil, fmt.Errorf("TOC tar entry had name %q; expected %q", h.Name, TOCTarName) + + var allErr []error + var found bool + var r *Reader + for _, d := range decompressors { + fSize := d.FooterSize() + fOffset := positive(int64(len(footer)) - fSize) + maybeTocBytes := footer[:fOffset] + tocOffset, tocSize, err := d.ParseFooter(footer[fOffset:]) + if err != nil { + allErr = append(allErr, err) + continue + } + if tocSize <= 0 { + tocSize = sr.Size() - tocOffset - fSize + } + if tocSize < int64(len(maybeTocBytes)) { + maybeTocBytes = maybeTocBytes[:tocSize] + } + r, err = parseTOC(d, sr, tocOffset, tocSize, maybeTocBytes, opts) + if err == nil { + found = true + break + } + allErr = append(allErr, err) } - dgstr := digest.Canonical.Digester() - toc := new(jtoc) - if err := json.NewDecoder(io.TeeReader(tr, dgstr.Hash())).Decode(&toc); err != nil { - return nil, fmt.Errorf("error decoding TOC JSON: %v", err) + if !found { + return nil, errorutil.Aggregate(allErr) } - r := &Reader{sr: sr, toc: toc, tocDigest: dgstr.Digest()} if err := r.initFields(); err != nil { return nil, fmt.Errorf("failed to initialize fields of entries: %v", err) } @@ -100,17 +174,26 @@ func Open(sr *io.SectionReader) (*Reader, error) { } // OpenFooter extracts and parses footer from the given blob. +// only supports gzip-based eStargz. func OpenFooter(sr *io.SectionReader) (tocOffset int64, footerSize int64, rErr error) { if sr.Size() < FooterSize && sr.Size() < legacyFooterSize { return 0, 0, fmt.Errorf("blob size %d is smaller than the footer size", sr.Size()) } - // TODO: read a bigger chunk (1MB?) at once here to hopefully - // get the TOC + footer in one go. var footer [FooterSize]byte if _, err := sr.ReadAt(footer[:], sr.Size()-FooterSize); err != nil { return 0, 0, fmt.Errorf("error reading footer: %v", err) } - return parseFooter(footer[:]) + var allErr []error + for _, d := range []Decompressor{new(GzipDecompressor), new(legacyGzipDecompressor)} { + fSize := d.FooterSize() + fOffset := positive(int64(len(footer)) - fSize) + tocOffset, _, err := d.ParseFooter(footer[fOffset:]) + if err == nil { + return tocOffset, fSize, err + } + allErr = append(allErr, err) + } + return 0, 0, errorutil.Aggregate(allErr) } // initFields populates the Reader from r.toc after decoding it from @@ -252,33 +335,68 @@ func (r *Reader) VerifyTOC(tocDigest digest.Digest) (TOCEntryVerifier, error) { if r.tocDigest != tocDigest { return nil, fmt.Errorf("invalid TOC JSON %q; want %q", r.tocDigest, tocDigest) } - digestMap := make(map[int64]digest.Digest) // map from chunk offset to the digest + + chunkDigestMap := make(map[int64]digest.Digest) // map from chunk offset to the chunk digest + regDigestMap := make(map[int64]digest.Digest) // map from chunk offset to the reg file digest + var chunkDigestMapIncomplete bool + var regDigestMapIncomplete bool + var containsChunk bool for _, e := range r.toc.Entries { - if e.Type == "reg" || e.Type == "chunk" { - if e.Type == "reg" && e.Size == 0 { - continue // ignores empty file - } + if e.Type != "reg" && e.Type != "chunk" { + continue + } - // offset must be unique in stargz blob - if _, ok := digestMap[e.Offset]; ok { - return nil, fmt.Errorf("offset %d found twice", e.Offset) + // offset must be unique in stargz blob + _, dOK := chunkDigestMap[e.Offset] + _, rOK := regDigestMap[e.Offset] + if dOK || rOK { + return nil, fmt.Errorf("offset %d found twice", e.Offset) + } + + if e.Type == "reg" { + if e.Size == 0 { + continue // ignores empty file } - // all chunk entries must contain digest - if e.ChunkDigest == "" { - return nil, fmt.Errorf("ChunkDigest of %q(off=%d) not found in TOC JSON", - e.Name, e.Offset) + // record the digest of regular file payload + if e.Digest != "" { + d, err := digest.Parse(e.Digest) + if err != nil { + return nil, errors.Wrapf(err, + "failed to parse regular file digest %q", e.Digest) + } + regDigestMap[e.Offset] = d + } else { + regDigestMapIncomplete = true } + } else { + containsChunk = true // this layer contains "chunk" entries. + } + // "reg" also can contain ChunkDigest (e.g. when "reg" is the first entry of + // chunked file) + if e.ChunkDigest != "" { d, err := digest.Parse(e.ChunkDigest) if err != nil { - return nil, errors.Wrapf(err, "failed to parse digest %q", e.ChunkDigest) + return nil, errors.Wrapf(err, + "failed to parse chunk digest %q", e.ChunkDigest) } - digestMap[e.Offset] = d + chunkDigestMap[e.Offset] = d + } else { + chunkDigestMapIncomplete = true } } - return &verifier{digestMap: digestMap}, nil + if chunkDigestMapIncomplete { + // Though some chunk digests are not found, if this layer doesn't contain + // "chunk"s and all digest of "reg" files are recorded, we can use them instead. + if !containsChunk && !regDigestMapIncomplete { + return &verifier{digestMap: regDigestMap}, nil + } + return nil, fmt.Errorf("some ChunkDigest not found in TOC JSON") + } + + return &verifier{digestMap: chunkDigestMap}, nil } // verifier is an implementation of TOCEntryVerifier which holds verifiers keyed by @@ -413,17 +531,17 @@ func (fr *fileReader) ReadAt(p []byte, off int64) (n int, err error) { off -= ent.ChunkOffset finalEnt := fr.ents[len(fr.ents)-1] - gzOff := ent.Offset - // gzBytesRemain is the number of compressed gzip bytes in this - // file remaining, over 1+ gzip chunks. - gzBytesRemain := finalEnt.NextOffset() - gzOff + compressedOff := ent.Offset + // compressedBytesRemain is the number of compressed bytes in this + // file remaining, over 1+ chunks. + compressedBytesRemain := finalEnt.NextOffset() - compressedOff - sr := io.NewSectionReader(fr.r.sr, gzOff, gzBytesRemain) + sr := io.NewSectionReader(fr.r.sr, compressedOff, compressedBytesRemain) - const maxGZread = 2 << 20 - var bufSize = maxGZread - if gzBytesRemain < maxGZread { - bufSize = int(gzBytesRemain) + const maxRead = 2 << 20 + var bufSize = maxRead + if compressedBytesRemain < maxRead { + bufSize = int(compressedBytesRemain) } br := bufio.NewReaderSize(sr, bufSize) @@ -431,14 +549,15 @@ func (fr *fileReader) ReadAt(p []byte, off int64) (n int, err error) { return 0, fmt.Errorf("fileReader.ReadAt.peek: %v", err) } - gz, err := gzip.NewReader(br) + dr, err := fr.r.decompressor.Reader(br) if err != nil { - return 0, fmt.Errorf("fileReader.ReadAt.gzipNewReader: %v", err) + return 0, fmt.Errorf("fileReader.ReadAt.decompressor.Reader: %v", err) } - if n, err := io.CopyN(ioutil.Discard, gz, off); n != off || err != nil { + defer dr.Close() + if n, err := io.CopyN(ioutil.Discard, dr, off); n != off || err != nil { return 0, fmt.Errorf("discard of %d bytes = %v, %v", off, n, err) } - return io.ReadFull(gz, p) + return io.ReadFull(dr, p) } // A Writer writes stargz files. @@ -447,14 +566,14 @@ func (fr *fileReader) ReadAt(p []byte, off int64) (n int, err error) { type Writer struct { bw *bufio.Writer cw *countWriter - toc *jtoc + toc *JTOC diffHash hash.Hash // SHA-256 of uncompressed tar - closed bool - gz *gzip.Writer - lastUsername map[int]string - lastGroupname map[int]string - compressionLevel int + closed bool + gz io.WriteCloser + lastUsername map[int]string + lastGroupname map[int]string + compressor Compressor // ChunkSize optionally controls the maximum number of bytes // of data of a regular file that can be written in one gzip @@ -463,16 +582,16 @@ type Writer struct { ChunkSize int } -// currentGzipWriter writes to the current w.gz field, which can +// currentCompressionWriter writes to the current w.gz field, which can // change throughout writing a tar entry. // // Additionally, it updates w's SHA-256 of the uncompressed bytes // of the tar file. -type currentGzipWriter struct{ w *Writer } +type currentCompressionWriter struct{ w *Writer } -func (cgw currentGzipWriter) Write(p []byte) (int, error) { - cgw.w.diffHash.Write(p) - return cgw.w.gz.Write(p) +func (ccw currentCompressionWriter) Write(p []byte) (int, error) { + ccw.w.diffHash.Write(p) + return ccw.w.gz.Write(p) } func (w *Writer) chunkSize() int { @@ -482,26 +601,34 @@ func (w *Writer) chunkSize() int { return w.ChunkSize } -// NewWriter returns a new stargz writer writing to w. +// NewWriter returns a new stargz writer (gzip-based) writing to w. // // The writer must be closed to write its trailing table of contents. func NewWriter(w io.Writer) *Writer { return NewWriterLevel(w, gzip.BestCompression) } -// NewWriterLevel returns a new stargz writer writing to w. +// NewWriterLevel returns a new stargz writer (gzip-based) writing to w. // The compression level is configurable. // // The writer must be closed to write its trailing table of contents. func NewWriterLevel(w io.Writer, compressionLevel int) *Writer { + return NewWriterWithCompressor(w, NewGzipCompressorWithLevel(compressionLevel)) +} + +// NewWriterLevel returns a new stargz writer writing to w. +// The compression method is configurable. +// +// The writer must be closed to write its trailing table of contents. +func NewWriterWithCompressor(w io.Writer, c Compressor) *Writer { bw := bufio.NewWriter(w) cw := &countWriter{w: bw} return &Writer{ - bw: bw, - cw: cw, - toc: &jtoc{Version: 1}, - diffHash: sha256.New(), - compressionLevel: compressionLevel, + bw: bw, + cw: cw, + toc: &JTOC{Version: 1}, + diffHash: sha256.New(), + compressor: c, } } @@ -517,42 +644,16 @@ func (w *Writer) Close() (digest.Digest, error) { return "", err } - // Write the TOC index. - tocOff := w.cw.n - w.gz, _ = gzip.NewWriterLevel(w.cw, w.compressionLevel) - tw := tar.NewWriter(currentGzipWriter{w}) - tocJSON, err := json.MarshalIndent(w.toc, "", "\t") + // Write the TOC index and footer. + tocDigest, err := w.compressor.WriteTOCAndFooter(w.cw, w.cw.n, w.toc, w.diffHash) if err != nil { return "", err } - if err := tw.WriteHeader(&tar.Header{ - Typeflag: tar.TypeReg, - Name: TOCTarName, - Size: int64(len(tocJSON)), - }); err != nil { - return "", err - } - if _, err := tw.Write(tocJSON); err != nil { - return "", err - } - - if err := tw.Close(); err != nil { - return "", err - } - if err := w.closeGz(); err != nil { - return "", err - } - - // And a little footer with pointer to the TOC gzip stream. - if _, err := w.bw.Write(footerBytes(tocOff)); err != nil { - return "", err - } - if err := w.bw.Flush(); err != nil { return "", err } - return digest.FromBytes(tocJSON), nil + return tocDigest, nil } func (w *Writer) closeGz() error { @@ -584,10 +685,11 @@ func (w *Writer) nameIfChanged(mp *map[int]string, id int, name string) string { return name } -func (w *Writer) condOpenGz() { +func (w *Writer) condOpenGz() (err error) { if w.gz == nil { - w.gz, _ = gzip.NewWriterLevel(w.cw, w.compressionLevel) + w.gz, err = w.compressor.Writer(w.cw) } + return } // AppendTar reads the tar or tar.gz file from r and appends @@ -639,8 +741,10 @@ func (w *Writer) AppendTar(r io.Reader) error { ModTime3339: formatModtime(h.ModTime), Xattrs: xattrs, } - w.condOpenGz() - tw := tar.NewWriter(currentGzipWriter{w}) + if err := w.condOpenGz(); err != nil { + return err + } + tw := tar.NewWriter(currentCompressionWriter{w}) if err := tw.WriteHeader(h); err != nil { return err } @@ -699,7 +803,9 @@ func (w *Writer) AppendTar(r io.Reader) error { ent.ChunkOffset = written chunkDigest := digest.Canonical.Digester() - w.condOpenGz() + if err := w.condOpenGz(); err != nil { + return err + } teeChunk := io.TeeReader(tee, chunkDigest.Hash()) if _, err := io.CopyN(tw, teeChunk, chunkSize); err != nil { @@ -732,83 +838,54 @@ func (w *Writer) DiffID() string { return fmt.Sprintf("sha256:%x", w.diffHash.Sum(nil)) } -// footerBytes returns the 51 bytes footer. -func footerBytes(tocOff int64) []byte { - buf := bytes.NewBuffer(make([]byte, 0, FooterSize)) - gz, _ := gzip.NewWriterLevel(buf, gzip.NoCompression) // MUST be NoCompression to keep 51 bytes - - // Extra header indicating the offset of TOCJSON - // https://tools.ietf.org/html/rfc1952#section-2.3.1.1 - header := make([]byte, 4) - header[0], header[1] = 'S', 'G' - subfield := fmt.Sprintf("%016xSTARGZ", tocOff) - binary.LittleEndian.PutUint16(header[2:4], uint16(len(subfield))) // little-endian per RFC1952 - gz.Header.Extra = append(header, []byte(subfield)...) - gz.Close() - if buf.Len() != FooterSize { - panic(fmt.Sprintf("footer buffer = %d, not %d", buf.Len(), FooterSize)) +func maxFooterSize(blobSize int64, decompressors ...Decompressor) (res int64) { + for _, d := range decompressors { + if s := d.FooterSize(); res < s && s <= blobSize { + res = s + } } - return buf.Bytes() + return } -func parseFooter(p []byte) (tocOffset int64, footerSize int64, rErr error) { - var allErr []error - - tocOffset, err := parseEStargzFooter(p) - if err == nil { - return tocOffset, FooterSize, nil - } - allErr = append(allErr, err) - - pad := len(p) - legacyFooterSize - if pad < 0 { - pad = 0 - } - tocOffset, err = parseLegacyFooter(p[pad:]) - if err == nil { - return tocOffset, legacyFooterSize, nil +func parseTOC(d Decompressor, sr *io.SectionReader, tocOff, tocSize int64, tocBytes []byte, opts openOpts) (*Reader, error) { + if len(tocBytes) > 0 { + start := time.Now() + toc, tocDgst, err := d.ParseTOC(bytes.NewReader(tocBytes)) + if err == nil { + if opts.telemetry != nil && opts.telemetry.DeserializeTocLatency != nil { + opts.telemetry.DeserializeTocLatency(start) + } + return &Reader{ + sr: sr, + toc: toc, + tocDigest: tocDgst, + decompressor: d, + }, nil + } } - return 0, 0, errorutil.Aggregate(append(allErr, err)) -} -func parseEStargzFooter(p []byte) (tocOffset int64, err error) { - if len(p) != FooterSize { - return 0, fmt.Errorf("invalid length %d cannot be parsed", len(p)) - } - zr, err := gzip.NewReader(bytes.NewReader(p)) - if err != nil { - return 0, err - } - extra := zr.Header.Extra - si1, si2, subfieldlen, subfield := extra[0], extra[1], extra[2:4], extra[4:] - if si1 != 'S' || si2 != 'G' { - return 0, fmt.Errorf("invalid subfield IDs: %q, %q; want E, S", si1, si2) - } - if slen := binary.LittleEndian.Uint16(subfieldlen); slen != uint16(16+len("STARGZ")) { - return 0, fmt.Errorf("invalid length of subfield %d; want %d", slen, 16+len("STARGZ")) - } - if string(subfield[16:]) != "STARGZ" { - return 0, fmt.Errorf("STARGZ magic string must be included in the footer subfield") + start := time.Now() + tocBytes = make([]byte, tocSize) + if _, err := sr.ReadAt(tocBytes, tocOff); err != nil { + return nil, fmt.Errorf("error reading %d byte TOC targz: %v", len(tocBytes), err) } - return strconv.ParseInt(string(subfield[:16]), 16, 64) -} - -func parseLegacyFooter(p []byte) (tocOffset int64, err error) { - if len(p) != legacyFooterSize { - return 0, fmt.Errorf("legacy: invalid length %d cannot be parsed", len(p)) + if opts.telemetry != nil && opts.telemetry.GetTocLatency != nil { + opts.telemetry.GetTocLatency(start) } - zr, err := gzip.NewReader(bytes.NewReader(p)) + start = time.Now() + toc, tocDgst, err := d.ParseTOC(bytes.NewReader(tocBytes)) if err != nil { - return 0, errors.Wrapf(err, "legacy: failed to get footer gzip reader") - } - extra := zr.Header.Extra - if len(extra) != 16+len("STARGZ") { - return 0, fmt.Errorf("legacy: invalid stargz's extra field size") + return nil, err } - if string(extra[16:]) != "STARGZ" { - return 0, fmt.Errorf("legacy: magic string STARGZ not found") + if opts.telemetry != nil && opts.telemetry.DeserializeTocLatency != nil { + opts.telemetry.DeserializeTocLatency(start) } - return strconv.ParseInt(string(extra[:16]), 16, 64) + return &Reader{ + sr: sr, + toc: toc, + tocDigest: tocDgst, + decompressor: d, + }, nil } func formatModtime(t time.Time) string { @@ -847,3 +924,10 @@ func isGzip(br *bufio.Reader) bool { peek, _ := br.Peek(3) return len(peek) >= 3 && peek[0] == gzipID1 && peek[1] == gzipID2 && peek[2] == gzipDeflate } + +func positive(n int64) int64 { + if n < 0 { + return 0 + } + return n +} diff --git a/vendor/github.com/containerd/stargz-snapshotter/estargz/go.mod b/vendor/github.com/containerd/stargz-snapshotter/estargz/go.mod index bf76d34030..ee6b2e17f4 100644 --- a/vendor/github.com/containerd/stargz-snapshotter/estargz/go.mod +++ b/vendor/github.com/containerd/stargz-snapshotter/estargz/go.mod @@ -3,7 +3,7 @@ module github.com/containerd/stargz-snapshotter/estargz go 1.16 require ( - github.com/klauspost/compress v1.12.3 + github.com/klauspost/compress v1.13.5 github.com/opencontainers/go-digest v1.0.0 github.com/pkg/errors v0.9.1 golang.org/x/sync v0.0.0-20201207232520-09787c993a3a diff --git a/vendor/github.com/containerd/stargz-snapshotter/estargz/go.sum b/vendor/github.com/containerd/stargz-snapshotter/estargz/go.sum index 793b74fda1..66cd2d69c3 100644 --- a/vendor/github.com/containerd/stargz-snapshotter/estargz/go.sum +++ b/vendor/github.com/containerd/stargz-snapshotter/estargz/go.sum @@ -1,7 +1,5 @@ -github.com/golang/snappy v0.0.3 h1:fHPg5GQYlCeLIPB9BZqMVR5nR9A+IM5zcgeTdjMYmLA= -github.com/golang/snappy v0.0.3/go.mod h1:/XxbfmMg8lxefKM7IXC3fBNl/7bRcc72aCRzEWrmP2Q= -github.com/klauspost/compress v1.12.3 h1:G5AfA94pHPysR56qqrkO2pxEexdDzrpFJ6yt/VqWxVU= -github.com/klauspost/compress v1.12.3/go.mod h1:8dP1Hq4DHOhN9w426knH3Rhby4rFm6D8eO+e+Dq5Gzg= +github.com/klauspost/compress v1.13.5 h1:9O69jUPDcsT9fEm74W92rZL9FQY7rCdaXVneq+yyzl4= +github.com/klauspost/compress v1.13.5/go.mod h1:/3/Vjq9QcHkK5uEr5lBEmyoZ1iFhe47etQ6QUkpK6sk= github.com/opencontainers/go-digest v1.0.0 h1:apOUWs51W5PlhuyGyz9FCeeBIOUDA/6nW8Oi/yOhh5U= github.com/opencontainers/go-digest v1.0.0/go.mod h1:0JzlMkj0TRzQZfJkVvzbP0HBR3IKzErnv2BNG4W4MAM= github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4= diff --git a/vendor/github.com/containerd/stargz-snapshotter/estargz/gzip.go b/vendor/github.com/containerd/stargz-snapshotter/estargz/gzip.go new file mode 100644 index 0000000000..efc435e099 --- /dev/null +++ b/vendor/github.com/containerd/stargz-snapshotter/estargz/gzip.go @@ -0,0 +1,216 @@ +/* + Copyright The containerd Authors. + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +*/ + +/* + Copyright 2019 The Go Authors. All rights reserved. + Use of this source code is governed by a BSD-style + license that can be found in the LICENSE file. +*/ + +package estargz + +import ( + "archive/tar" + "bytes" + "compress/gzip" + "encoding/binary" + "encoding/json" + "fmt" + "hash" + "io" + "strconv" + + digest "github.com/opencontainers/go-digest" + "github.com/pkg/errors" +) + +type gzipCompression struct { + *gzipCompressor + *GzipDecompressor +} + +func newGzipCompressionWithLevel(level int) Compression { + return &gzipCompression{ + &gzipCompressor{level}, + &GzipDecompressor{}, + } +} + +func NewGzipCompressorWithLevel(level int) Compressor { + return &gzipCompressor{level} +} + +type gzipCompressor struct { + compressionLevel int +} + +func (gc *gzipCompressor) Writer(w io.Writer) (io.WriteCloser, error) { + return gzip.NewWriterLevel(w, gc.compressionLevel) +} + +func (gc *gzipCompressor) WriteTOCAndFooter(w io.Writer, off int64, toc *JTOC, diffHash hash.Hash) (digest.Digest, error) { + tocJSON, err := json.MarshalIndent(toc, "", "\t") + if err != nil { + return "", err + } + gz, _ := gzip.NewWriterLevel(w, gc.compressionLevel) + gw := io.Writer(gz) + if diffHash != nil { + gw = io.MultiWriter(gz, diffHash) + } + tw := tar.NewWriter(gw) + if err := tw.WriteHeader(&tar.Header{ + Typeflag: tar.TypeReg, + Name: TOCTarName, + Size: int64(len(tocJSON)), + }); err != nil { + return "", err + } + if _, err := tw.Write(tocJSON); err != nil { + return "", err + } + + if err := tw.Close(); err != nil { + return "", err + } + if err := gz.Close(); err != nil { + return "", err + } + if _, err := w.Write(gzipFooterBytes(off)); err != nil { + return "", err + } + return digest.FromBytes(tocJSON), nil +} + +// gzipFooterBytes returns the 51 bytes footer. +func gzipFooterBytes(tocOff int64) []byte { + buf := bytes.NewBuffer(make([]byte, 0, FooterSize)) + gz, _ := gzip.NewWriterLevel(buf, gzip.NoCompression) // MUST be NoCompression to keep 51 bytes + + // Extra header indicating the offset of TOCJSON + // https://tools.ietf.org/html/rfc1952#section-2.3.1.1 + header := make([]byte, 4) + header[0], header[1] = 'S', 'G' + subfield := fmt.Sprintf("%016xSTARGZ", tocOff) + binary.LittleEndian.PutUint16(header[2:4], uint16(len(subfield))) // little-endian per RFC1952 + gz.Header.Extra = append(header, []byte(subfield)...) + gz.Close() + if buf.Len() != FooterSize { + panic(fmt.Sprintf("footer buffer = %d, not %d", buf.Len(), FooterSize)) + } + return buf.Bytes() +} + +type GzipDecompressor struct{} + +func (gz *GzipDecompressor) Reader(r io.Reader) (io.ReadCloser, error) { + return gzip.NewReader(r) +} + +func (gz *GzipDecompressor) ParseTOC(r io.Reader) (toc *JTOC, tocDgst digest.Digest, err error) { + return parseTOCEStargz(r) +} + +func (gz *GzipDecompressor) ParseFooter(p []byte) (tocOffset, tocSize int64, err error) { + if len(p) != FooterSize { + return 0, 0, fmt.Errorf("invalid length %d cannot be parsed", len(p)) + } + zr, err := gzip.NewReader(bytes.NewReader(p)) + if err != nil { + return 0, 0, err + } + defer zr.Close() + extra := zr.Header.Extra + si1, si2, subfieldlen, subfield := extra[0], extra[1], extra[2:4], extra[4:] + if si1 != 'S' || si2 != 'G' { + return 0, 0, fmt.Errorf("invalid subfield IDs: %q, %q; want E, S", si1, si2) + } + if slen := binary.LittleEndian.Uint16(subfieldlen); slen != uint16(16+len("STARGZ")) { + return 0, 0, fmt.Errorf("invalid length of subfield %d; want %d", slen, 16+len("STARGZ")) + } + if string(subfield[16:]) != "STARGZ" { + return 0, 0, fmt.Errorf("STARGZ magic string must be included in the footer subfield") + } + tocOffset, err = strconv.ParseInt(string(subfield[:16]), 16, 64) + if err != nil { + return 0, 0, errors.Wrapf(err, "legacy: failed to parse toc offset") + } + return tocOffset, 0, nil +} + +func (gz *GzipDecompressor) FooterSize() int64 { + return FooterSize +} + +type legacyGzipDecompressor struct{} + +func (gz *legacyGzipDecompressor) Reader(r io.Reader) (io.ReadCloser, error) { + return gzip.NewReader(r) +} + +func (gz *legacyGzipDecompressor) ParseTOC(r io.Reader) (toc *JTOC, tocDgst digest.Digest, err error) { + return parseTOCEStargz(r) +} + +func (gz *legacyGzipDecompressor) ParseFooter(p []byte) (tocOffset, tocSize int64, err error) { + if len(p) != legacyFooterSize { + return 0, 0, fmt.Errorf("legacy: invalid length %d cannot be parsed", len(p)) + } + zr, err := gzip.NewReader(bytes.NewReader(p)) + if err != nil { + return 0, 0, errors.Wrapf(err, "legacy: failed to get footer gzip reader") + } + defer zr.Close() + extra := zr.Header.Extra + if len(extra) != 16+len("STARGZ") { + return 0, 0, fmt.Errorf("legacy: invalid stargz's extra field size") + } + if string(extra[16:]) != "STARGZ" { + return 0, 0, fmt.Errorf("legacy: magic string STARGZ not found") + } + tocOffset, err = strconv.ParseInt(string(extra[:16]), 16, 64) + if err != nil { + return 0, 0, errors.Wrapf(err, "legacy: failed to parse toc offset") + } + return tocOffset, 0, nil +} + +func (gz *legacyGzipDecompressor) FooterSize() int64 { + return legacyFooterSize +} + +func parseTOCEStargz(r io.Reader) (toc *JTOC, tocDgst digest.Digest, err error) { + zr, err := gzip.NewReader(r) + if err != nil { + return nil, "", fmt.Errorf("malformed TOC gzip header: %v", err) + } + defer zr.Close() + zr.Multistream(false) + tr := tar.NewReader(zr) + h, err := tr.Next() + if err != nil { + return nil, "", fmt.Errorf("failed to find tar header in TOC gzip stream: %v", err) + } + if h.Name != TOCTarName { + return nil, "", fmt.Errorf("TOC tar entry had name %q; expected %q", h.Name, TOCTarName) + } + dgstr := digest.Canonical.Digester() + toc = new(JTOC) + if err := json.NewDecoder(io.TeeReader(tr, dgstr.Hash())).Decode(&toc); err != nil { + return nil, "", fmt.Errorf("error decoding TOC JSON: %v", err) + } + return toc, dgstr.Digest(), nil +} diff --git a/vendor/github.com/containerd/stargz-snapshotter/estargz/testutil.go b/vendor/github.com/containerd/stargz-snapshotter/estargz/testutil.go new file mode 100644 index 0000000000..020729b7ed --- /dev/null +++ b/vendor/github.com/containerd/stargz-snapshotter/estargz/testutil.go @@ -0,0 +1,1933 @@ +/* + Copyright The containerd Authors. + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +*/ + +/* + Copyright 2019 The Go Authors. All rights reserved. + Use of this source code is governed by a BSD-style + license that can be found in the LICENSE file. +*/ + +package estargz + +import ( + "archive/tar" + "bytes" + "compress/gzip" + "crypto/sha256" + "encoding/json" + "fmt" + "io" + "io/ioutil" + "os" + "reflect" + "sort" + "strings" + "testing" + "time" + + "github.com/containerd/stargz-snapshotter/estargz/errorutil" + "github.com/klauspost/compress/zstd" + digest "github.com/opencontainers/go-digest" + "github.com/pkg/errors" +) + +// TestingController is Compression with some helper methods necessary for testing. +type TestingController interface { + Compression + CountStreams(*testing.T, []byte) int + DiffIDOf(*testing.T, []byte) string + String() string +} + +// CompressionTestSuite tests this pkg with controllers can build valid eStargz blobs and parse them. +func CompressionTestSuite(t *testing.T, controllers ...TestingController) { + t.Run("testBuild", func(t *testing.T) { t.Parallel(); testBuild(t, controllers...) }) + t.Run("testDigestAndVerify", func(t *testing.T) { t.Parallel(); testDigestAndVerify(t, controllers...) }) + t.Run("testWriteAndOpen", func(t *testing.T) { t.Parallel(); testWriteAndOpen(t, controllers...) }) +} + +const ( + uncompressedType int = iota + gzipType + zstdType +) + +var srcCompressions = []int{ + uncompressedType, + gzipType, + zstdType, +} + +var allowedPrefix = [4]string{"", "./", "/", "../"} + +// testBuild tests the resulting stargz blob built by this pkg has the same +// contents as the normal stargz blob. +func testBuild(t *testing.T, controllers ...TestingController) { + tests := []struct { + name string + chunkSize int + in []tarEntry + }{ + { + name: "regfiles and directories", + chunkSize: 4, + in: tarOf( + file("foo", "test1"), + dir("foo2/"), + file("foo2/bar", "test2", xAttr(map[string]string{"test": "sample"})), + ), + }, + { + name: "empty files", + chunkSize: 4, + in: tarOf( + file("foo", "tttttt"), + file("foo_empty", ""), + file("foo2", "tttttt"), + file("foo_empty2", ""), + file("foo3", "tttttt"), + file("foo_empty3", ""), + file("foo4", "tttttt"), + file("foo_empty4", ""), + file("foo5", "tttttt"), + file("foo_empty5", ""), + file("foo6", "tttttt"), + ), + }, + { + name: "various files", + chunkSize: 4, + in: tarOf( + file("baz.txt", "bazbazbazbazbazbazbaz"), + file("foo.txt", "a"), + symlink("barlink", "test/bar.txt"), + dir("test/"), + dir("dev/"), + blockdev("dev/testblock", 3, 4), + fifo("dev/testfifo"), + chardev("dev/testchar1", 5, 6), + file("test/bar.txt", "testbartestbar", xAttr(map[string]string{"test2": "sample2"})), + dir("test2/"), + link("test2/bazlink", "baz.txt"), + chardev("dev/testchar2", 1, 2), + ), + }, + { + name: "no contents", + chunkSize: 4, + in: tarOf( + file("baz.txt", ""), + symlink("barlink", "test/bar.txt"), + dir("test/"), + dir("dev/"), + blockdev("dev/testblock", 3, 4), + fifo("dev/testfifo"), + chardev("dev/testchar1", 5, 6), + file("test/bar.txt", "", xAttr(map[string]string{"test2": "sample2"})), + dir("test2/"), + link("test2/bazlink", "baz.txt"), + chardev("dev/testchar2", 1, 2), + ), + }, + } + for _, tt := range tests { + for _, srcCompression := range srcCompressions { + srcCompression := srcCompression + for _, cl := range controllers { + cl := cl + for _, prefix := range allowedPrefix { + prefix := prefix + t.Run(tt.name+"-"+fmt.Sprintf("compression=%v-prefix=%q-src=%d", cl, prefix, srcCompression), func(t *testing.T) { + tarBlob := buildTarStatic(t, tt.in, prefix) + // Test divideEntries() + entries, err := sortEntries(tarBlob, nil, nil) // identical order + if err != nil { + t.Fatalf("faield to parse tar: %v", err) + } + var merged []*entry + for _, part := range divideEntries(entries, 4) { + merged = append(merged, part...) + } + if !reflect.DeepEqual(entries, merged) { + for _, e := range entries { + t.Logf("Original: %v", e.header) + } + for _, e := range merged { + t.Logf("Merged: %v", e.header) + } + t.Errorf("divided entries couldn't be merged") + return + } + + // Prepare sample data + wantBuf := new(bytes.Buffer) + sw := NewWriterWithCompressor(wantBuf, cl) + sw.ChunkSize = tt.chunkSize + if err := sw.AppendTar(tarBlob); err != nil { + t.Fatalf("faield to append tar to want stargz: %v", err) + } + if _, err := sw.Close(); err != nil { + t.Fatalf("faield to prepare want stargz: %v", err) + } + wantData := wantBuf.Bytes() + want, err := Open(io.NewSectionReader( + bytes.NewReader(wantData), 0, int64(len(wantData))), + WithDecompressors(cl), + ) + if err != nil { + t.Fatalf("failed to parse the want stargz: %v", err) + } + + // Prepare testing data + rc, err := Build(compressBlob(t, tarBlob, srcCompression), + WithChunkSize(tt.chunkSize), WithCompression(cl)) + if err != nil { + t.Fatalf("faield to build stargz: %v", err) + } + defer rc.Close() + gotBuf := new(bytes.Buffer) + if _, err := io.Copy(gotBuf, rc); err != nil { + t.Fatalf("failed to copy built stargz blob: %v", err) + } + gotData := gotBuf.Bytes() + got, err := Open(io.NewSectionReader( + bytes.NewReader(gotBuf.Bytes()), 0, int64(len(gotData))), + WithDecompressors(cl), + ) + if err != nil { + t.Fatalf("failed to parse the got stargz: %v", err) + } + + // Check DiffID is properly calculated + rc.Close() + diffID := rc.DiffID() + wantDiffID := cl.DiffIDOf(t, gotData) + if diffID.String() != wantDiffID { + t.Errorf("DiffID = %q; want %q", diffID, wantDiffID) + } + + // Compare as stargz + if !isSameVersion(t, cl, wantData, gotData) { + t.Errorf("built stargz hasn't same json") + return + } + if !isSameEntries(t, want, got) { + t.Errorf("built stargz isn't same as the original") + return + } + + // Compare as tar.gz + if !isSameTarGz(t, cl, wantData, gotData) { + t.Errorf("built stargz isn't same tar.gz") + return + } + }) + } + } + } + } +} + +func isSameTarGz(t *testing.T, controller TestingController, a, b []byte) bool { + aGz, err := controller.Reader(bytes.NewReader(a)) + if err != nil { + t.Fatalf("failed to read A") + } + defer aGz.Close() + bGz, err := controller.Reader(bytes.NewReader(b)) + if err != nil { + t.Fatalf("failed to read B") + } + defer bGz.Close() + + // Same as tar's Next() method but ignores landmarks and TOCJSON file + next := func(r *tar.Reader) (h *tar.Header, err error) { + for { + if h, err = r.Next(); err != nil { + return + } + if h.Name != PrefetchLandmark && + h.Name != NoPrefetchLandmark && + h.Name != TOCTarName { + return + } + } + } + + aTar := tar.NewReader(aGz) + bTar := tar.NewReader(bGz) + for { + // Fetch and parse next header. + aH, aErr := next(aTar) + bH, bErr := next(bTar) + if aErr != nil || bErr != nil { + if aErr == io.EOF && bErr == io.EOF { + break + } + t.Fatalf("Failed to parse tar file: A: %v, B: %v", aErr, bErr) + } + if !reflect.DeepEqual(aH, bH) { + t.Logf("different header (A = %v; B = %v)", aH, bH) + return false + + } + aFile, err := ioutil.ReadAll(aTar) + if err != nil { + t.Fatal("failed to read tar payload of A") + } + bFile, err := ioutil.ReadAll(bTar) + if err != nil { + t.Fatal("failed to read tar payload of B") + } + if !bytes.Equal(aFile, bFile) { + t.Logf("different tar payload (A = %q; B = %q)", string(a), string(b)) + return false + } + } + + return true +} + +func isSameVersion(t *testing.T, controller TestingController, a, b []byte) bool { + aJTOC, _, err := parseStargz(io.NewSectionReader(bytes.NewReader(a), 0, int64(len(a))), controller) + if err != nil { + t.Fatalf("failed to parse A: %v", err) + } + bJTOC, _, err := parseStargz(io.NewSectionReader(bytes.NewReader(b), 0, int64(len(b))), controller) + if err != nil { + t.Fatalf("failed to parse B: %v", err) + } + t.Logf("A: TOCJSON: %v", dumpTOCJSON(t, aJTOC)) + t.Logf("B: TOCJSON: %v", dumpTOCJSON(t, bJTOC)) + return aJTOC.Version == bJTOC.Version +} + +func isSameEntries(t *testing.T, a, b *Reader) bool { + aroot, ok := a.Lookup("") + if !ok { + t.Fatalf("failed to get root of A") + } + broot, ok := b.Lookup("") + if !ok { + t.Fatalf("failed to get root of B") + } + aEntry := stargzEntry{aroot, a} + bEntry := stargzEntry{broot, b} + return contains(t, aEntry, bEntry) && contains(t, bEntry, aEntry) +} + +func compressBlob(t *testing.T, src *io.SectionReader, srcCompression int) *io.SectionReader { + buf := new(bytes.Buffer) + var w io.WriteCloser + var err error + if srcCompression == gzipType { + w = gzip.NewWriter(buf) + } else if srcCompression == zstdType { + w, err = zstd.NewWriter(buf) + if err != nil { + t.Fatalf("failed to init zstd writer: %v", err) + } + } else { + return src + } + src.Seek(0, io.SeekStart) + if _, err := io.Copy(w, src); err != nil { + t.Fatalf("failed to compress source") + } + if err := w.Close(); err != nil { + t.Fatalf("failed to finalize compress source") + } + data := buf.Bytes() + return io.NewSectionReader(bytes.NewReader(data), 0, int64(len(data))) + +} + +type stargzEntry struct { + e *TOCEntry + r *Reader +} + +// contains checks if all child entries in "b" are also contained in "a". +// This function also checks if the files/chunks contain the same contents among "a" and "b". +func contains(t *testing.T, a, b stargzEntry) bool { + ae, ar := a.e, a.r + be, br := b.e, b.r + t.Logf("Comparing: %q vs %q", ae.Name, be.Name) + if !equalEntry(ae, be) { + t.Logf("%q != %q: entry: a: %v, b: %v", ae.Name, be.Name, ae, be) + return false + } + if ae.Type == "dir" { + t.Logf("Directory: %q vs %q: %v vs %v", ae.Name, be.Name, + allChildrenName(ae), allChildrenName(be)) + iscontain := true + ae.ForeachChild(func(aBaseName string, aChild *TOCEntry) bool { + // Walk through all files on this stargz file. + + if aChild.Name == PrefetchLandmark || + aChild.Name == NoPrefetchLandmark { + return true // Ignore landmarks + } + + // Ignore a TOCEntry of "./" (formated as "" by stargz lib) on root directory + // because this points to the root directory itself. + if aChild.Name == "" && ae.Name == "" { + return true + } + + bChild, ok := be.LookupChild(aBaseName) + if !ok { + t.Logf("%q (base: %q): not found in b: %v", + ae.Name, aBaseName, allChildrenName(be)) + iscontain = false + return false + } + + childcontain := contains(t, stargzEntry{aChild, a.r}, stargzEntry{bChild, b.r}) + if !childcontain { + t.Logf("%q != %q: non-equal dir", ae.Name, be.Name) + iscontain = false + return false + } + return true + }) + return iscontain + } else if ae.Type == "reg" { + af, err := ar.OpenFile(ae.Name) + if err != nil { + t.Fatalf("failed to open file %q on A: %v", ae.Name, err) + } + bf, err := br.OpenFile(be.Name) + if err != nil { + t.Fatalf("failed to open file %q on B: %v", be.Name, err) + } + + var nr int64 + for nr < ae.Size { + abytes, anext, aok := readOffset(t, af, nr, a) + bbytes, bnext, bok := readOffset(t, bf, nr, b) + if !aok && !bok { + break + } else if !(aok && bok) || anext != bnext { + t.Logf("%q != %q (offset=%d): chunk existence a=%v vs b=%v, anext=%v vs bnext=%v", + ae.Name, be.Name, nr, aok, bok, anext, bnext) + return false + } + nr = anext + if !bytes.Equal(abytes, bbytes) { + t.Logf("%q != %q: different contents %v vs %v", + ae.Name, be.Name, string(abytes), string(bbytes)) + return false + } + } + return true + } + + return true +} + +func allChildrenName(e *TOCEntry) (children []string) { + e.ForeachChild(func(baseName string, _ *TOCEntry) bool { + children = append(children, baseName) + return true + }) + return +} + +func equalEntry(a, b *TOCEntry) bool { + // Here, we selectively compare fileds that we are interested in. + return a.Name == b.Name && + a.Type == b.Type && + a.Size == b.Size && + a.ModTime3339 == b.ModTime3339 && + a.Stat().ModTime().Equal(b.Stat().ModTime()) && // modTime time.Time + a.LinkName == b.LinkName && + a.Mode == b.Mode && + a.UID == b.UID && + a.GID == b.GID && + a.Uname == b.Uname && + a.Gname == b.Gname && + (a.Offset > 0) == (b.Offset > 0) && + (a.NextOffset() > 0) == (b.NextOffset() > 0) && + a.DevMajor == b.DevMajor && + a.DevMinor == b.DevMinor && + a.NumLink == b.NumLink && + reflect.DeepEqual(a.Xattrs, b.Xattrs) && + // chunk-related infomations aren't compared in this function. + // ChunkOffset int64 `json:"chunkOffset,omitempty"` + // ChunkSize int64 `json:"chunkSize,omitempty"` + // children map[string]*TOCEntry + a.Digest == b.Digest +} + +func readOffset(t *testing.T, r *io.SectionReader, offset int64, e stargzEntry) ([]byte, int64, bool) { + ce, ok := e.r.ChunkEntryForOffset(e.e.Name, offset) + if !ok { + return nil, 0, false + } + data := make([]byte, ce.ChunkSize) + t.Logf("Offset: %v, NextOffset: %v", ce.Offset, ce.NextOffset()) + n, err := r.ReadAt(data, ce.ChunkOffset) + if err != nil { + t.Fatalf("failed to read file payload of %q (offset:%d,size:%d): %v", + e.e.Name, ce.ChunkOffset, ce.ChunkSize, err) + } + if int64(n) != ce.ChunkSize { + t.Fatalf("unexpected copied data size %d; want %d", + n, ce.ChunkSize) + } + return data[:n], offset + ce.ChunkSize, true +} + +func dumpTOCJSON(t *testing.T, tocJSON *JTOC) string { + jtocData, err := json.Marshal(*tocJSON) + if err != nil { + t.Fatalf("failed to marshal TOC JSON: %v", err) + } + buf := new(bytes.Buffer) + if _, err := io.Copy(buf, bytes.NewReader(jtocData)); err != nil { + t.Fatalf("failed to read toc json blob: %v", err) + } + return buf.String() +} + +const chunkSize = 3 + +// type check func(t *testing.T, sgzData []byte, tocDigest digest.Digest, dgstMap map[string]digest.Digest, compressionLevel int) +type check func(t *testing.T, sgzData []byte, tocDigest digest.Digest, dgstMap map[string]digest.Digest, controller TestingController) + +// testDigestAndVerify runs specified checks against sample stargz blobs. +func testDigestAndVerify(t *testing.T, controllers ...TestingController) { + tests := []struct { + name string + tarInit func(t *testing.T, dgstMap map[string]digest.Digest) (blob []tarEntry) + checks []check + }{ + { + name: "no-regfile", + tarInit: func(t *testing.T, dgstMap map[string]digest.Digest) (blob []tarEntry) { + return tarOf( + dir("test/"), + ) + }, + checks: []check{ + checkStargzTOC, + checkVerifyTOC, + checkVerifyInvalidStargzFail(buildTarStatic(t, tarOf( + dir("test2/"), // modified + ), allowedPrefix[0])), + }, + }, + { + name: "small-files", + tarInit: func(t *testing.T, dgstMap map[string]digest.Digest) (blob []tarEntry) { + return tarOf( + regDigest(t, "baz.txt", "", dgstMap), + regDigest(t, "foo.txt", "a", dgstMap), + dir("test/"), + regDigest(t, "test/bar.txt", "bbb", dgstMap), + ) + }, + checks: []check{ + checkStargzTOC, + checkVerifyTOC, + checkVerifyInvalidStargzFail(buildTarStatic(t, tarOf( + file("baz.txt", ""), + file("foo.txt", "M"), // modified + dir("test/"), + file("test/bar.txt", "bbb"), + ), allowedPrefix[0])), + // checkVerifyInvalidTOCEntryFail("foo.txt"), // TODO + checkVerifyBrokenContentFail("foo.txt"), + }, + }, + { + name: "big-files", + tarInit: func(t *testing.T, dgstMap map[string]digest.Digest) (blob []tarEntry) { + return tarOf( + regDigest(t, "baz.txt", "bazbazbazbazbazbazbaz", dgstMap), + regDigest(t, "foo.txt", "a", dgstMap), + dir("test/"), + regDigest(t, "test/bar.txt", "testbartestbar", dgstMap), + ) + }, + checks: []check{ + checkStargzTOC, + checkVerifyTOC, + checkVerifyInvalidStargzFail(buildTarStatic(t, tarOf( + file("baz.txt", "bazbazbazMMMbazbazbaz"), // modified + file("foo.txt", "a"), + dir("test/"), + file("test/bar.txt", "testbartestbar"), + ), allowedPrefix[0])), + checkVerifyInvalidTOCEntryFail("test/bar.txt"), + checkVerifyBrokenContentFail("test/bar.txt"), + }, + }, + { + name: "with-non-regfiles", + tarInit: func(t *testing.T, dgstMap map[string]digest.Digest) (blob []tarEntry) { + return tarOf( + regDigest(t, "baz.txt", "bazbazbazbazbazbazbaz", dgstMap), + regDigest(t, "foo.txt", "a", dgstMap), + symlink("barlink", "test/bar.txt"), + dir("test/"), + regDigest(t, "test/bar.txt", "testbartestbar", dgstMap), + dir("test2/"), + link("test2/bazlink", "baz.txt"), + ) + }, + checks: []check{ + checkStargzTOC, + checkVerifyTOC, + checkVerifyInvalidStargzFail(buildTarStatic(t, tarOf( + file("baz.txt", "bazbazbazbazbazbazbaz"), + file("foo.txt", "a"), + symlink("barlink", "test/bar.txt"), + dir("test/"), + file("test/bar.txt", "testbartestbar"), + dir("test2/"), + link("test2/bazlink", "foo.txt"), // modified + ), allowedPrefix[0])), + checkVerifyInvalidTOCEntryFail("test/bar.txt"), + checkVerifyBrokenContentFail("test/bar.txt"), + }, + }, + } + + for _, tt := range tests { + for _, srcCompression := range srcCompressions { + srcCompression := srcCompression + for _, cl := range controllers { + cl := cl + for _, prefix := range allowedPrefix { + prefix := prefix + t.Run(tt.name+"-"+fmt.Sprintf("compression=%v-prefix=%q", cl, prefix), func(t *testing.T) { + // Get original tar file and chunk digests + dgstMap := make(map[string]digest.Digest) + tarBlob := buildTarStatic(t, tt.tarInit(t, dgstMap), prefix) + + rc, err := Build(compressBlob(t, tarBlob, srcCompression), + WithChunkSize(chunkSize), WithCompression(cl)) + if err != nil { + t.Fatalf("failed to convert stargz: %v", err) + } + tocDigest := rc.TOCDigest() + defer rc.Close() + buf := new(bytes.Buffer) + if _, err := io.Copy(buf, rc); err != nil { + t.Fatalf("failed to copy built stargz blob: %v", err) + } + newStargz := buf.Bytes() + // NoPrefetchLandmark is added during `Bulid`, which is expected behaviour. + dgstMap[chunkID(NoPrefetchLandmark, 0, int64(len([]byte{landmarkContents})))] = digest.FromBytes([]byte{landmarkContents}) + + for _, check := range tt.checks { + check(t, newStargz, tocDigest, dgstMap, cl) + } + }) + } + } + } + } +} + +// checkStargzTOC checks the TOC JSON of the passed stargz has the expected +// digest and contains valid chunks. It walks all entries in the stargz and +// checks all chunk digests stored to the TOC JSON match the actual contents. +func checkStargzTOC(t *testing.T, sgzData []byte, tocDigest digest.Digest, dgstMap map[string]digest.Digest, controller TestingController) { + sgz, err := Open( + io.NewSectionReader(bytes.NewReader(sgzData), 0, int64(len(sgzData))), + WithDecompressors(controller), + ) + if err != nil { + t.Errorf("failed to parse converted stargz: %v", err) + return + } + digestMapTOC, err := listDigests(io.NewSectionReader( + bytes.NewReader(sgzData), 0, int64(len(sgzData))), + controller, + ) + if err != nil { + t.Fatalf("failed to list digest: %v", err) + } + found := make(map[string]bool) + for id := range dgstMap { + found[id] = false + } + zr, err := controller.Reader(bytes.NewReader(sgzData)) + if err != nil { + t.Fatalf("failed to decompress converted stargz: %v", err) + } + defer zr.Close() + tr := tar.NewReader(zr) + for { + h, err := tr.Next() + if err != nil { + if err != io.EOF { + t.Errorf("failed to read tar entry: %v", err) + return + } + break + } + if h.Name == TOCTarName { + // Check the digest of TOC JSON based on the actual contents + // It's sure that TOC JSON exists in this archive because + // Open succeeded. + dgstr := digest.Canonical.Digester() + if _, err := io.Copy(dgstr.Hash(), tr); err != nil { + t.Fatalf("failed to calculate digest of TOC JSON: %v", + err) + } + if dgstr.Digest() != tocDigest { + t.Errorf("invalid TOC JSON %q; want %q", tocDigest, dgstr.Digest()) + } + continue + } + if _, ok := sgz.Lookup(h.Name); !ok { + t.Errorf("lost stargz entry %q in the converted TOC", h.Name) + return + } + var n int64 + for n < h.Size { + ce, ok := sgz.ChunkEntryForOffset(h.Name, n) + if !ok { + t.Errorf("lost chunk %q(offset=%d) in the converted TOC", + h.Name, n) + return + } + + // Get the original digest to make sure the file contents are kept unchanged + // from the original tar, during the whole conversion steps. + id := chunkID(h.Name, n, ce.ChunkSize) + want, ok := dgstMap[id] + if !ok { + t.Errorf("Unexpected chunk %q(offset=%d,size=%d): %v", + h.Name, n, ce.ChunkSize, dgstMap) + return + } + found[id] = true + + // Check the file contents + dgstr := digest.Canonical.Digester() + if _, err := io.CopyN(dgstr.Hash(), tr, ce.ChunkSize); err != nil { + t.Fatalf("failed to calculate digest of %q (offset=%d,size=%d)", + h.Name, n, ce.ChunkSize) + } + if want != dgstr.Digest() { + t.Errorf("Invalid contents in converted stargz %q: %q; want %q", + h.Name, dgstr.Digest(), want) + return + } + + // Check the digest stored in TOC JSON + dgstTOC, ok := digestMapTOC[ce.Offset] + if !ok { + t.Errorf("digest of %q(offset=%d,size=%d,chunkOffset=%d) isn't registered", + h.Name, ce.Offset, ce.ChunkSize, ce.ChunkOffset) + } + if want != dgstTOC { + t.Errorf("Invalid digest in TOCEntry %q: %q; want %q", + h.Name, dgstTOC, want) + return + } + + n += ce.ChunkSize + } + } + + for id, ok := range found { + if !ok { + t.Errorf("required chunk %q not found in the converted stargz: %v", id, found) + } + } +} + +// checkVerifyTOC checks the verification works for the TOC JSON of the passed +// stargz. It walks all entries in the stargz and checks the verifications for +// all chunks work. +func checkVerifyTOC(t *testing.T, sgzData []byte, tocDigest digest.Digest, dgstMap map[string]digest.Digest, controller TestingController) { + sgz, err := Open( + io.NewSectionReader(bytes.NewReader(sgzData), 0, int64(len(sgzData))), + WithDecompressors(controller), + ) + if err != nil { + t.Errorf("failed to parse converted stargz: %v", err) + return + } + ev, err := sgz.VerifyTOC(tocDigest) + if err != nil { + t.Errorf("failed to verify stargz: %v", err) + return + } + + found := make(map[string]bool) + for id := range dgstMap { + found[id] = false + } + zr, err := controller.Reader(bytes.NewReader(sgzData)) + if err != nil { + t.Fatalf("failed to decompress converted stargz: %v", err) + } + defer zr.Close() + tr := tar.NewReader(zr) + for { + h, err := tr.Next() + if err != nil { + if err != io.EOF { + t.Errorf("failed to read tar entry: %v", err) + return + } + break + } + if h.Name == TOCTarName { + continue + } + if _, ok := sgz.Lookup(h.Name); !ok { + t.Errorf("lost stargz entry %q in the converted TOC", h.Name) + return + } + var n int64 + for n < h.Size { + ce, ok := sgz.ChunkEntryForOffset(h.Name, n) + if !ok { + t.Errorf("lost chunk %q(offset=%d) in the converted TOC", + h.Name, n) + return + } + + v, err := ev.Verifier(ce) + if err != nil { + t.Errorf("failed to get verifier for %q(offset=%d)", h.Name, n) + } + + found[chunkID(h.Name, n, ce.ChunkSize)] = true + + // Check the file contents + if _, err := io.CopyN(v, tr, ce.ChunkSize); err != nil { + t.Fatalf("failed to get chunk of %q (offset=%d,size=%d)", + h.Name, n, ce.ChunkSize) + } + if !v.Verified() { + t.Errorf("Invalid contents in converted stargz %q (should be succeeded)", + h.Name) + return + } + n += ce.ChunkSize + } + } + + for id, ok := range found { + if !ok { + t.Errorf("required chunk %q not found in the converted stargz: %v", id, found) + } + } +} + +// checkVerifyInvalidTOCEntryFail checks if misconfigured TOC JSON can be +// detected during the verification and the verification returns an error. +func checkVerifyInvalidTOCEntryFail(filename string) check { + return func(t *testing.T, sgzData []byte, tocDigest digest.Digest, dgstMap map[string]digest.Digest, controller TestingController) { + funcs := map[string]rewriteFunc{ + "lost digest in a entry": func(t *testing.T, toc *JTOC, sgz *io.SectionReader) { + var found bool + for _, e := range toc.Entries { + if cleanEntryName(e.Name) == filename { + if e.Type != "reg" && e.Type != "chunk" { + t.Fatalf("entry %q to break must be regfile or chunk", filename) + } + if e.ChunkDigest == "" { + t.Fatalf("entry %q is already invalid", filename) + } + e.ChunkDigest = "" + found = true + } + } + if !found { + t.Fatalf("rewrite target not found") + } + }, + "duplicated entry offset": func(t *testing.T, toc *JTOC, sgz *io.SectionReader) { + var ( + sampleEntry *TOCEntry + targetEntry *TOCEntry + ) + for _, e := range toc.Entries { + if e.Type == "reg" || e.Type == "chunk" { + if cleanEntryName(e.Name) == filename { + targetEntry = e + } else { + sampleEntry = e + } + } + } + if sampleEntry == nil { + t.Fatalf("TOC must contain at least one regfile or chunk entry other than the rewrite target") + } + if targetEntry == nil { + t.Fatalf("rewrite target not found") + } + targetEntry.Offset = sampleEntry.Offset + }, + } + + for name, rFunc := range funcs { + t.Run(name, func(t *testing.T) { + newSgz, newTocDigest := rewriteTOCJSON(t, io.NewSectionReader(bytes.NewReader(sgzData), 0, int64(len(sgzData))), rFunc, controller) + buf := new(bytes.Buffer) + if _, err := io.Copy(buf, newSgz); err != nil { + t.Fatalf("failed to get converted stargz") + } + isgz := buf.Bytes() + + sgz, err := Open( + io.NewSectionReader(bytes.NewReader(isgz), 0, int64(len(isgz))), + WithDecompressors(controller), + ) + if err != nil { + t.Fatalf("failed to parse converted stargz: %v", err) + return + } + _, err = sgz.VerifyTOC(newTocDigest) + if err == nil { + t.Errorf("must fail for invalid TOC") + return + } + }) + } + } +} + +// checkVerifyInvalidStargzFail checks if the verification detects that the +// given stargz file doesn't match to the expected digest and returns error. +func checkVerifyInvalidStargzFail(invalid *io.SectionReader) check { + return func(t *testing.T, sgzData []byte, tocDigest digest.Digest, dgstMap map[string]digest.Digest, controller TestingController) { + rc, err := Build(invalid, WithChunkSize(chunkSize), WithCompression(controller)) + if err != nil { + t.Fatalf("failed to convert stargz: %v", err) + } + defer rc.Close() + buf := new(bytes.Buffer) + if _, err := io.Copy(buf, rc); err != nil { + t.Fatalf("failed to copy built stargz blob: %v", err) + } + mStargz := buf.Bytes() + + sgz, err := Open( + io.NewSectionReader(bytes.NewReader(mStargz), 0, int64(len(mStargz))), + WithDecompressors(controller), + ) + if err != nil { + t.Fatalf("failed to parse converted stargz: %v", err) + return + } + _, err = sgz.VerifyTOC(tocDigest) + if err == nil { + t.Errorf("must fail for invalid TOC") + return + } + } +} + +// checkVerifyBrokenContentFail checks if the verifier detects broken contents +// that doesn't match to the expected digest and returns error. +func checkVerifyBrokenContentFail(filename string) check { + return func(t *testing.T, sgzData []byte, tocDigest digest.Digest, dgstMap map[string]digest.Digest, controller TestingController) { + // Parse stargz file + sgz, err := Open( + io.NewSectionReader(bytes.NewReader(sgzData), 0, int64(len(sgzData))), + WithDecompressors(controller), + ) + if err != nil { + t.Fatalf("failed to parse converted stargz: %v", err) + return + } + ev, err := sgz.VerifyTOC(tocDigest) + if err != nil { + t.Fatalf("failed to verify stargz: %v", err) + return + } + + // Open the target file + sr, err := sgz.OpenFile(filename) + if err != nil { + t.Fatalf("failed to open file %q", filename) + } + ce, ok := sgz.ChunkEntryForOffset(filename, 0) + if !ok { + t.Fatalf("lost chunk %q(offset=%d) in the converted TOC", filename, 0) + return + } + if ce.ChunkSize == 0 { + t.Fatalf("file mustn't be empty") + return + } + data := make([]byte, ce.ChunkSize) + if _, err := sr.ReadAt(data, ce.ChunkOffset); err != nil { + t.Errorf("failed to get data of a chunk of %q(offset=%q)", + filename, ce.ChunkOffset) + } + + // Check the broken chunk (must fail) + v, err := ev.Verifier(ce) + if err != nil { + t.Fatalf("failed to get verifier for %q", filename) + } + broken := append([]byte{^data[0]}, data[1:]...) + if _, err := io.CopyN(v, bytes.NewReader(broken), ce.ChunkSize); err != nil { + t.Fatalf("failed to get chunk of %q (offset=%d,size=%d)", + filename, ce.ChunkOffset, ce.ChunkSize) + } + if v.Verified() { + t.Errorf("verification must fail for broken file chunk %q(org:%q,broken:%q)", + filename, data, broken) + } + } +} + +func chunkID(name string, offset, size int64) string { + return fmt.Sprintf("%s-%d-%d", cleanEntryName(name), offset, size) +} + +type rewriteFunc func(t *testing.T, toc *JTOC, sgz *io.SectionReader) + +func rewriteTOCJSON(t *testing.T, sgz *io.SectionReader, rewrite rewriteFunc, controller TestingController) (newSgz io.Reader, tocDigest digest.Digest) { + decodedJTOC, jtocOffset, err := parseStargz(sgz, controller) + if err != nil { + t.Fatalf("failed to extract TOC JSON: %v", err) + } + + rewrite(t, decodedJTOC, sgz) + + tocFooter, tocDigest, err := tocAndFooter(controller, decodedJTOC, jtocOffset) + if err != nil { + t.Fatalf("failed to create toc and footer: %v", err) + } + + // Reconstruct stargz file with the modified TOC JSON + if _, err := sgz.Seek(0, io.SeekStart); err != nil { + t.Fatalf("failed to reset the seek position of stargz: %v", err) + } + return io.MultiReader( + io.LimitReader(sgz, jtocOffset), // Original stargz (before TOC JSON) + tocFooter, // Rewritten TOC and footer + ), tocDigest +} + +func listDigests(sgz *io.SectionReader, controller TestingController) (map[int64]digest.Digest, error) { + decodedJTOC, _, err := parseStargz(sgz, controller) + if err != nil { + return nil, err + } + digestMap := make(map[int64]digest.Digest) + for _, e := range decodedJTOC.Entries { + if e.Type == "reg" || e.Type == "chunk" { + if e.Type == "reg" && e.Size == 0 { + continue // ignores empty file + } + if e.ChunkDigest == "" { + return nil, fmt.Errorf("ChunkDigest of %q(off=%d) not found in TOC JSON", + e.Name, e.Offset) + } + d, err := digest.Parse(e.ChunkDigest) + if err != nil { + return nil, err + } + digestMap[e.Offset] = d + } + } + return digestMap, nil +} + +func parseStargz(sgz *io.SectionReader, controller TestingController) (decodedJTOC *JTOC, jtocOffset int64, err error) { + fSize := controller.FooterSize() + footer := make([]byte, fSize) + if _, err := sgz.ReadAt(footer, sgz.Size()-fSize); err != nil { + return nil, 0, errors.Wrap(err, "error reading footer") + } + tocOffset, _, err := controller.ParseFooter(footer[positive(int64(len(footer))-fSize):]) + if err != nil { + return nil, 0, errors.Wrapf(err, "failed to parse footer") + } + + // Decode the TOC JSON + tocReader := io.NewSectionReader(sgz, tocOffset, sgz.Size()-tocOffset-fSize) + decodedJTOC, _, err = controller.ParseTOC(tocReader) + if err != nil { + return nil, 0, errors.Wrap(err, "failed to parse TOC") + } + return decodedJTOC, tocOffset, nil +} + +func testWriteAndOpen(t *testing.T, controllers ...TestingController) { + const content = "Some contents" + invalidUtf8 := "\xff\xfe\xfd" + + xAttrFile := xAttr{"foo": "bar", "invalid-utf8": invalidUtf8} + sampleOwner := owner{uid: 50, gid: 100} + + tests := []struct { + name string + chunkSize int + in []tarEntry + want []stargzCheck + wantNumGz int // expected number of streams + }{ + { + name: "empty", + in: tarOf(), + wantNumGz: 2, // TOC + footer + want: checks( + numTOCEntries(0), + ), + }, + { + name: "1dir_1empty_file", + in: tarOf( + dir("foo/"), + file("foo/bar.txt", ""), + ), + wantNumGz: 3, // dir, TOC, footer + want: checks( + numTOCEntries(2), + hasDir("foo/"), + hasFileLen("foo/bar.txt", 0), + entryHasChildren("foo", "bar.txt"), + hasFileDigest("foo/bar.txt", digestFor("")), + ), + }, + { + name: "1dir_1file", + in: tarOf( + dir("foo/"), + file("foo/bar.txt", content, xAttrFile), + ), + wantNumGz: 4, // var dir, foo.txt alone, TOC, footer + want: checks( + numTOCEntries(2), + hasDir("foo/"), + hasFileLen("foo/bar.txt", len(content)), + hasFileDigest("foo/bar.txt", digestFor(content)), + hasFileContentsRange("foo/bar.txt", 0, content), + hasFileContentsRange("foo/bar.txt", 1, content[1:]), + entryHasChildren("", "foo"), + entryHasChildren("foo", "bar.txt"), + hasFileXattrs("foo/bar.txt", "foo", "bar"), + hasFileXattrs("foo/bar.txt", "invalid-utf8", invalidUtf8), + ), + }, + { + name: "2meta_2file", + in: tarOf( + dir("bar/", sampleOwner), + dir("foo/", sampleOwner), + file("foo/bar.txt", content, sampleOwner), + ), + wantNumGz: 4, // both dirs, foo.txt alone, TOC, footer + want: checks( + numTOCEntries(3), + hasDir("bar/"), + hasDir("foo/"), + hasFileLen("foo/bar.txt", len(content)), + entryHasChildren("", "bar", "foo"), + entryHasChildren("foo", "bar.txt"), + hasChunkEntries("foo/bar.txt", 1), + hasEntryOwner("bar/", sampleOwner), + hasEntryOwner("foo/", sampleOwner), + hasEntryOwner("foo/bar.txt", sampleOwner), + ), + }, + { + name: "3dir", + in: tarOf( + dir("bar/"), + dir("foo/"), + dir("foo/bar/"), + ), + wantNumGz: 3, // 3 dirs, TOC, footer + want: checks( + hasDirLinkCount("bar/", 2), + hasDirLinkCount("foo/", 3), + hasDirLinkCount("foo/bar/", 2), + ), + }, + { + name: "symlink", + in: tarOf( + dir("foo/"), + symlink("foo/bar", "../../x"), + ), + wantNumGz: 3, // metas + TOC + footer + want: checks( + numTOCEntries(2), + hasSymlink("foo/bar", "../../x"), + entryHasChildren("", "foo"), + entryHasChildren("foo", "bar"), + ), + }, + { + name: "chunked_file", + chunkSize: 4, + in: tarOf( + dir("foo/"), + file("foo/big.txt", "This "+"is s"+"uch "+"a bi"+"g fi"+"le"), + ), + wantNumGz: 9, + want: checks( + numTOCEntries(7), // 1 for foo dir, 6 for the foo/big.txt file + hasDir("foo/"), + hasFileLen("foo/big.txt", len("This is such a big file")), + hasFileDigest("foo/big.txt", digestFor("This is such a big file")), + hasFileContentsRange("foo/big.txt", 0, "This is such a big file"), + hasFileContentsRange("foo/big.txt", 1, "his is such a big file"), + hasFileContentsRange("foo/big.txt", 2, "is is such a big file"), + hasFileContentsRange("foo/big.txt", 3, "s is such a big file"), + hasFileContentsRange("foo/big.txt", 4, " is such a big file"), + hasFileContentsRange("foo/big.txt", 5, "is such a big file"), + hasFileContentsRange("foo/big.txt", 6, "s such a big file"), + hasFileContentsRange("foo/big.txt", 7, " such a big file"), + hasFileContentsRange("foo/big.txt", 8, "such a big file"), + hasFileContentsRange("foo/big.txt", 9, "uch a big file"), + hasFileContentsRange("foo/big.txt", 10, "ch a big file"), + hasFileContentsRange("foo/big.txt", 11, "h a big file"), + hasFileContentsRange("foo/big.txt", 12, " a big file"), + hasFileContentsRange("foo/big.txt", len("This is such a big file")-1, ""), + hasChunkEntries("foo/big.txt", 6), + ), + }, + { + name: "recursive", + in: tarOf( + dir("/", sampleOwner), + dir("bar/", sampleOwner), + dir("foo/", sampleOwner), + file("foo/bar.txt", content, sampleOwner), + ), + wantNumGz: 4, // dirs, bar.txt alone, TOC, footer + want: checks( + maxDepth(2), // 0: root directory, 1: "foo/", 2: "bar.txt" + ), + }, + { + name: "block_char_fifo", + in: tarOf( + tarEntryFunc(func(w *tar.Writer, prefix string) error { + return w.WriteHeader(&tar.Header{ + Name: prefix + "b", + Typeflag: tar.TypeBlock, + Devmajor: 123, + Devminor: 456, + }) + }), + tarEntryFunc(func(w *tar.Writer, prefix string) error { + return w.WriteHeader(&tar.Header{ + Name: prefix + "c", + Typeflag: tar.TypeChar, + Devmajor: 111, + Devminor: 222, + }) + }), + tarEntryFunc(func(w *tar.Writer, prefix string) error { + return w.WriteHeader(&tar.Header{ + Name: prefix + "f", + Typeflag: tar.TypeFifo, + }) + }), + ), + wantNumGz: 3, + want: checks( + lookupMatch("b", &TOCEntry{Name: "b", Type: "block", DevMajor: 123, DevMinor: 456, NumLink: 1}), + lookupMatch("c", &TOCEntry{Name: "c", Type: "char", DevMajor: 111, DevMinor: 222, NumLink: 1}), + lookupMatch("f", &TOCEntry{Name: "f", Type: "fifo", NumLink: 1}), + ), + }, + { + name: "modes", + in: tarOf( + dir("foo1/", 0755|os.ModeDir|os.ModeSetgid), + file("foo1/bar1", content, 0700|os.ModeSetuid), + file("foo1/bar2", content, 0755|os.ModeSetgid), + dir("foo2/", 0755|os.ModeDir|os.ModeSticky), + file("foo2/bar3", content, 0755|os.ModeSticky), + dir("foo3/", 0755|os.ModeDir), + file("foo3/bar4", content, os.FileMode(0700)), + file("foo3/bar5", content, os.FileMode(0755)), + ), + wantNumGz: 8, // dir, bar1 alone, bar2 alone + dir, bar3 alone + dir, bar4 alone, bar5 alone, TOC, footer + want: checks( + hasMode("foo1/", 0755|os.ModeDir|os.ModeSetgid), + hasMode("foo1/bar1", 0700|os.ModeSetuid), + hasMode("foo1/bar2", 0755|os.ModeSetgid), + hasMode("foo2/", 0755|os.ModeDir|os.ModeSticky), + hasMode("foo2/bar3", 0755|os.ModeSticky), + hasMode("foo3/", 0755|os.ModeDir), + hasMode("foo3/bar4", os.FileMode(0700)), + hasMode("foo3/bar5", os.FileMode(0755)), + ), + }, + } + + for _, tt := range tests { + for _, cl := range controllers { + cl := cl + for _, prefix := range allowedPrefix { + prefix := prefix + t.Run(tt.name+"-"+fmt.Sprintf("compression=%v-prefix=%q", cl, prefix), func(t *testing.T) { + tr, cancel := buildTar(t, tt.in, prefix) + defer cancel() + var stargzBuf bytes.Buffer + w := NewWriterWithCompressor(&stargzBuf, cl) + w.ChunkSize = tt.chunkSize + if err := w.AppendTar(tr); err != nil { + t.Fatalf("Append: %v", err) + } + if _, err := w.Close(); err != nil { + t.Fatalf("Writer.Close: %v", err) + } + b := stargzBuf.Bytes() + + diffID := w.DiffID() + wantDiffID := cl.DiffIDOf(t, b) + if diffID != wantDiffID { + t.Errorf("DiffID = %q; want %q", diffID, wantDiffID) + } + + got := cl.CountStreams(t, b) + if got != tt.wantNumGz { + t.Errorf("number of streams = %d; want %d", got, tt.wantNumGz) + } + + telemetry, checkCalled := newCalledTelemetry() + r, err := Open( + io.NewSectionReader(bytes.NewReader(b), 0, int64(len(b))), + WithDecompressors(cl), + WithTelemetry(telemetry), + ) + if err != nil { + t.Fatalf("stargz.Open: %v", err) + } + if err := checkCalled(); err != nil { + t.Errorf("telemetry failure: %v", err) + } + for _, want := range tt.want { + want.check(t, r) + } + }) + } + } + } +} + +func newCalledTelemetry() (telemetry *Telemetry, check func() error) { + var getFooterLatencyCalled bool + var getTocLatencyCalled bool + var deserializeTocLatencyCalled bool + return &Telemetry{ + func(time.Time) { getFooterLatencyCalled = true }, + func(time.Time) { getTocLatencyCalled = true }, + func(time.Time) { deserializeTocLatencyCalled = true }, + }, func() error { + var allErr []error + if !getFooterLatencyCalled { + allErr = append(allErr, fmt.Errorf("metrics GetFooterLatency isn't called")) + } + if !getTocLatencyCalled { + allErr = append(allErr, fmt.Errorf("metrics GetTocLatency isn't called")) + } + if !deserializeTocLatencyCalled { + allErr = append(allErr, fmt.Errorf("metrics DeserializeTocLatency isn't called")) + } + return errorutil.Aggregate(allErr) + } +} + +func digestFor(content string) string { + sum := sha256.Sum256([]byte(content)) + return fmt.Sprintf("sha256:%x", sum) +} + +type numTOCEntries int + +func (n numTOCEntries) check(t *testing.T, r *Reader) { + if r.toc == nil { + t.Fatal("nil TOC") + } + if got, want := len(r.toc.Entries), int(n); got != want { + t.Errorf("got %d TOC entries; want %d", got, want) + } + t.Logf("got TOC entries:") + for i, ent := range r.toc.Entries { + entj, _ := json.Marshal(ent) + t.Logf(" [%d]: %s\n", i, entj) + } + if t.Failed() { + t.FailNow() + } +} + +func checks(s ...stargzCheck) []stargzCheck { return s } + +type stargzCheck interface { + check(t *testing.T, r *Reader) +} + +type stargzCheckFn func(*testing.T, *Reader) + +func (f stargzCheckFn) check(t *testing.T, r *Reader) { f(t, r) } + +func maxDepth(max int) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + e, ok := r.Lookup("") + if !ok { + t.Fatal("root directory not found") + } + d, err := getMaxDepth(t, e, 0, 10*max) + if err != nil { + t.Errorf("failed to get max depth (wanted %d): %v", max, err) + return + } + if d != max { + t.Errorf("invalid depth %d; want %d", d, max) + return + } + }) +} + +func getMaxDepth(t *testing.T, e *TOCEntry, current, limit int) (max int, rErr error) { + if current > limit { + return -1, fmt.Errorf("walkMaxDepth: exceeds limit: current:%d > limit:%d", + current, limit) + } + max = current + e.ForeachChild(func(baseName string, ent *TOCEntry) bool { + t.Logf("%q(basename:%q) is child of %q\n", ent.Name, baseName, e.Name) + d, err := getMaxDepth(t, ent, current+1, limit) + if err != nil { + rErr = err + return false + } + if d > max { + max = d + } + return true + }) + return +} + +func hasFileLen(file string, wantLen int) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + for _, ent := range r.toc.Entries { + if ent.Name == file { + if ent.Type != "reg" { + t.Errorf("file type of %q is %q; want \"reg\"", file, ent.Type) + } else if ent.Size != int64(wantLen) { + t.Errorf("file size of %q = %d; want %d", file, ent.Size, wantLen) + } + return + } + } + t.Errorf("file %q not found", file) + }) +} + +func hasFileXattrs(file, name, value string) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + for _, ent := range r.toc.Entries { + if ent.Name == file { + if ent.Type != "reg" { + t.Errorf("file type of %q is %q; want \"reg\"", file, ent.Type) + } + if ent.Xattrs == nil { + t.Errorf("file %q has no xattrs", file) + return + } + valueFound, found := ent.Xattrs[name] + if !found { + t.Errorf("file %q has no xattr %q", file, name) + return + } + if string(valueFound) != value { + t.Errorf("file %q has xattr %q with value %q instead of %q", file, name, valueFound, value) + } + + return + } + } + t.Errorf("file %q not found", file) + }) +} + +func hasFileDigest(file string, digest string) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + ent, ok := r.Lookup(file) + if !ok { + t.Fatalf("didn't find TOCEntry for file %q", file) + } + if ent.Digest != digest { + t.Fatalf("Digest(%q) = %q, want %q", file, ent.Digest, digest) + } + }) +} + +func hasFileContentsRange(file string, offset int, want string) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + f, err := r.OpenFile(file) + if err != nil { + t.Fatal(err) + } + got := make([]byte, len(want)) + n, err := f.ReadAt(got, int64(offset)) + if err != nil { + t.Fatalf("ReadAt(len %d, offset %d) = %v, %v", len(got), offset, n, err) + } + if string(got) != want { + t.Fatalf("ReadAt(len %d, offset %d) = %q, want %q", len(got), offset, got, want) + } + }) +} + +func hasChunkEntries(file string, wantChunks int) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + ent, ok := r.Lookup(file) + if !ok { + t.Fatalf("no file for %q", file) + } + if ent.Type != "reg" { + t.Fatalf("file %q has unexpected type %q; want reg", file, ent.Type) + } + chunks := r.getChunks(ent) + if len(chunks) != wantChunks { + t.Errorf("len(r.getChunks(%q)) = %d; want %d", file, len(chunks), wantChunks) + return + } + f := chunks[0] + + var gotChunks []*TOCEntry + var last *TOCEntry + for off := int64(0); off < f.Size; off++ { + e, ok := r.ChunkEntryForOffset(file, off) + if !ok { + t.Errorf("no ChunkEntryForOffset at %d", off) + return + } + if last != e { + gotChunks = append(gotChunks, e) + last = e + } + } + if !reflect.DeepEqual(chunks, gotChunks) { + t.Errorf("gotChunks=%d, want=%d; contents mismatch", len(gotChunks), wantChunks) + } + + // And verify the NextOffset + for i := 0; i < len(gotChunks)-1; i++ { + ci := gotChunks[i] + cnext := gotChunks[i+1] + if ci.NextOffset() != cnext.Offset { + t.Errorf("chunk %d NextOffset %d != next chunk's Offset of %d", i, ci.NextOffset(), cnext.Offset) + } + } + }) +} + +func entryHasChildren(dir string, want ...string) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + want := append([]string(nil), want...) + var got []string + ent, ok := r.Lookup(dir) + if !ok { + t.Fatalf("didn't find TOCEntry for dir node %q", dir) + } + for baseName := range ent.children { + got = append(got, baseName) + } + sort.Strings(got) + sort.Strings(want) + if !reflect.DeepEqual(got, want) { + t.Errorf("children of %q = %q; want %q", dir, got, want) + } + }) +} + +func hasDir(file string) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + for _, ent := range r.toc.Entries { + if ent.Name == cleanEntryName(file) { + if ent.Type != "dir" { + t.Errorf("file type of %q is %q; want \"dir\"", file, ent.Type) + } + return + } + } + t.Errorf("directory %q not found", file) + }) +} + +func hasDirLinkCount(file string, count int) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + for _, ent := range r.toc.Entries { + if ent.Name == cleanEntryName(file) { + if ent.Type != "dir" { + t.Errorf("file type of %q is %q; want \"dir\"", file, ent.Type) + return + } + if ent.NumLink != count { + t.Errorf("link count of %q = %d; want %d", file, ent.NumLink, count) + } + return + } + } + t.Errorf("directory %q not found", file) + }) +} + +func hasMode(file string, mode os.FileMode) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + for _, ent := range r.toc.Entries { + if ent.Name == cleanEntryName(file) { + if ent.Stat().Mode() != mode { + t.Errorf("invalid mode: got %v; want %v", ent.Stat().Mode(), mode) + return + } + return + } + } + t.Errorf("file %q not found", file) + }) +} + +func hasSymlink(file, target string) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + for _, ent := range r.toc.Entries { + if ent.Name == file { + if ent.Type != "symlink" { + t.Errorf("file type of %q is %q; want \"symlink\"", file, ent.Type) + } else if ent.LinkName != target { + t.Errorf("link target of symlink %q is %q; want %q", file, ent.LinkName, target) + } + return + } + } + t.Errorf("symlink %q not found", file) + }) +} + +func lookupMatch(name string, want *TOCEntry) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + e, ok := r.Lookup(name) + if !ok { + t.Fatalf("failed to Lookup entry %q", name) + } + if !reflect.DeepEqual(e, want) { + t.Errorf("entry %q mismatch.\n got: %+v\nwant: %+v\n", name, e, want) + } + + }) +} + +func hasEntryOwner(entry string, owner owner) stargzCheck { + return stargzCheckFn(func(t *testing.T, r *Reader) { + ent, ok := r.Lookup(strings.TrimSuffix(entry, "/")) + if !ok { + t.Errorf("entry %q not found", entry) + return + } + if ent.UID != owner.uid || ent.GID != owner.gid { + t.Errorf("entry %q has invalid owner (uid:%d, gid:%d) instead of (uid:%d, gid:%d)", entry, ent.UID, ent.GID, owner.uid, owner.gid) + return + } + }) +} + +func tarOf(s ...tarEntry) []tarEntry { return s } + +type tarEntry interface { + appendTar(tw *tar.Writer, prefix string) error +} + +type tarEntryFunc func(*tar.Writer, string) error + +func (f tarEntryFunc) appendTar(tw *tar.Writer, prefix string) error { return f(tw, prefix) } + +func buildTar(t *testing.T, ents []tarEntry, prefix string) (r io.Reader, cancel func()) { + pr, pw := io.Pipe() + go func() { + tw := tar.NewWriter(pw) + for _, ent := range ents { + if err := ent.appendTar(tw, prefix); err != nil { + t.Errorf("building input tar: %v", err) + pw.Close() + return + } + } + if err := tw.Close(); err != nil { + t.Errorf("closing write of input tar: %v", err) + } + pw.Close() + }() + return pr, func() { go pr.Close(); go pw.Close() } +} + +func buildTarStatic(t *testing.T, ents []tarEntry, prefix string) *io.SectionReader { + buf := new(bytes.Buffer) + tw := tar.NewWriter(buf) + for _, ent := range ents { + if err := ent.appendTar(tw, prefix); err != nil { + t.Fatalf("building input tar: %v", err) + } + } + if err := tw.Close(); err != nil { + t.Errorf("closing write of input tar: %v", err) + } + data := buf.Bytes() + return io.NewSectionReader(bytes.NewReader(data), 0, int64(len(data))) +} + +func dir(name string, opts ...interface{}) tarEntry { + return tarEntryFunc(func(tw *tar.Writer, prefix string) error { + var o owner + mode := os.FileMode(0755) + for _, opt := range opts { + switch v := opt.(type) { + case owner: + o = v + case os.FileMode: + mode = v + default: + return errors.New("unsupported opt") + } + } + if !strings.HasSuffix(name, "/") { + panic(fmt.Sprintf("missing trailing slash in dir %q ", name)) + } + tm, err := fileModeToTarMode(mode) + if err != nil { + return err + } + return tw.WriteHeader(&tar.Header{ + Typeflag: tar.TypeDir, + Name: prefix + name, + Mode: tm, + Uid: o.uid, + Gid: o.gid, + }) + }) +} + +// xAttr are extended attributes to set on test files created with the file func. +type xAttr map[string]string + +// owner is owner ot set on test files and directories with the file and dir functions. +type owner struct { + uid int + gid int +} + +func file(name, contents string, opts ...interface{}) tarEntry { + return tarEntryFunc(func(tw *tar.Writer, prefix string) error { + var xattrs xAttr + var o owner + mode := os.FileMode(0644) + for _, opt := range opts { + switch v := opt.(type) { + case xAttr: + xattrs = v + case owner: + o = v + case os.FileMode: + mode = v + default: + return errors.New("unsupported opt") + } + } + if strings.HasSuffix(name, "/") { + return fmt.Errorf("bogus trailing slash in file %q", name) + } + tm, err := fileModeToTarMode(mode) + if err != nil { + return err + } + if err := tw.WriteHeader(&tar.Header{ + Typeflag: tar.TypeReg, + Name: prefix + name, + Mode: tm, + Xattrs: xattrs, + Size: int64(len(contents)), + Uid: o.uid, + Gid: o.gid, + }); err != nil { + return err + } + _, err = io.WriteString(tw, contents) + return err + }) +} + +func symlink(name, target string) tarEntry { + return tarEntryFunc(func(tw *tar.Writer, prefix string) error { + return tw.WriteHeader(&tar.Header{ + Typeflag: tar.TypeSymlink, + Name: prefix + name, + Linkname: target, + Mode: 0644, + }) + }) +} + +func link(name string, linkname string) tarEntry { + now := time.Now() + return tarEntryFunc(func(w *tar.Writer, prefix string) error { + return w.WriteHeader(&tar.Header{ + Typeflag: tar.TypeLink, + Name: prefix + name, + Linkname: linkname, + ModTime: now, + AccessTime: now, + ChangeTime: now, + }) + }) +} + +func chardev(name string, major, minor int64) tarEntry { + now := time.Now() + return tarEntryFunc(func(w *tar.Writer, prefix string) error { + return w.WriteHeader(&tar.Header{ + Typeflag: tar.TypeChar, + Name: prefix + name, + Devmajor: major, + Devminor: minor, + ModTime: now, + AccessTime: now, + ChangeTime: now, + }) + }) +} + +func blockdev(name string, major, minor int64) tarEntry { + now := time.Now() + return tarEntryFunc(func(w *tar.Writer, prefix string) error { + return w.WriteHeader(&tar.Header{ + Typeflag: tar.TypeBlock, + Name: prefix + name, + Devmajor: major, + Devminor: minor, + ModTime: now, + AccessTime: now, + ChangeTime: now, + }) + }) +} +func fifo(name string) tarEntry { + now := time.Now() + return tarEntryFunc(func(w *tar.Writer, prefix string) error { + return w.WriteHeader(&tar.Header{ + Typeflag: tar.TypeFifo, + Name: prefix + name, + ModTime: now, + AccessTime: now, + ChangeTime: now, + }) + }) +} + +func prefetchLandmark() tarEntry { + return tarEntryFunc(func(w *tar.Writer, prefix string) error { + if err := w.WriteHeader(&tar.Header{ + Name: PrefetchLandmark, + Typeflag: tar.TypeReg, + Size: int64(len([]byte{landmarkContents})), + }); err != nil { + return err + } + contents := []byte{landmarkContents} + if _, err := io.CopyN(w, bytes.NewReader(contents), int64(len(contents))); err != nil { + return err + } + return nil + }) +} + +func noPrefetchLandmark() tarEntry { + return tarEntryFunc(func(w *tar.Writer, prefix string) error { + if err := w.WriteHeader(&tar.Header{ + Name: NoPrefetchLandmark, + Typeflag: tar.TypeReg, + Size: int64(len([]byte{landmarkContents})), + }); err != nil { + return err + } + contents := []byte{landmarkContents} + if _, err := io.CopyN(w, bytes.NewReader(contents), int64(len(contents))); err != nil { + return err + } + return nil + }) +} + +func regDigest(t *testing.T, name string, contentStr string, digestMap map[string]digest.Digest) tarEntry { + if digestMap == nil { + t.Fatalf("digest map mustn't be nil") + } + content := []byte(contentStr) + + var n int64 + for n < int64(len(content)) { + size := int64(chunkSize) + remain := int64(len(content)) - n + if remain < size { + size = remain + } + dgstr := digest.Canonical.Digester() + if _, err := io.CopyN(dgstr.Hash(), bytes.NewReader(content[n:n+size]), size); err != nil { + t.Fatalf("failed to calculate digest of %q (name=%q,offset=%d,size=%d)", + string(content[n:n+size]), name, n, size) + } + digestMap[chunkID(name, n, size)] = dgstr.Digest() + n += size + } + + return tarEntryFunc(func(w *tar.Writer, prefix string) error { + if err := w.WriteHeader(&tar.Header{ + Typeflag: tar.TypeReg, + Name: prefix + name, + Size: int64(len(content)), + }); err != nil { + return err + } + if _, err := io.CopyN(w, bytes.NewReader(content), int64(len(content))); err != nil { + return err + } + return nil + }) +} + +func fileModeToTarMode(mode os.FileMode) (int64, error) { + h, err := tar.FileInfoHeader(fileInfoOnlyMode(mode), "") + if err != nil { + return 0, err + } + return h.Mode, nil +} + +// fileInfoOnlyMode is os.FileMode that populates only file mode. +type fileInfoOnlyMode os.FileMode + +func (f fileInfoOnlyMode) Name() string { return "" } +func (f fileInfoOnlyMode) Size() int64 { return 0 } +func (f fileInfoOnlyMode) Mode() os.FileMode { return os.FileMode(f) } +func (f fileInfoOnlyMode) ModTime() time.Time { return time.Now() } +func (f fileInfoOnlyMode) IsDir() bool { return os.FileMode(f).IsDir() } +func (f fileInfoOnlyMode) Sys() interface{} { return nil } diff --git a/vendor/github.com/containerd/stargz-snapshotter/estargz/types.go b/vendor/github.com/containerd/stargz-snapshotter/estargz/types.go index 680347a79c..1b1075a646 100644 --- a/vendor/github.com/containerd/stargz-snapshotter/estargz/types.go +++ b/vendor/github.com/containerd/stargz-snapshotter/estargz/types.go @@ -24,6 +24,8 @@ package estargz import ( "archive/tar" + "hash" + "io" "os" "path" "time" @@ -90,8 +92,8 @@ const ( landmarkContents = 0xf ) -// jtoc is the JSON-serialized table of contents index of the files in the stargz file. -type jtoc struct { +// JTOC is the JSON-serialized table of contents index of the files in the stargz file. +type JTOC struct { Version int `json:"version"` Entries []*TOCEntry `json:"entries"` } @@ -262,3 +264,51 @@ type TOCEntryVerifier interface { // contents of the specified TOCEntry. Verifier(ce *TOCEntry) (digest.Verifier, error) } + +// Compression provides the compression helper to be used creating and parsing eStargz. +// This package provides gzip-based Compression by default, but any compression +// algorithm (e.g. zstd) can be used as long as it implements Compression. +type Compression interface { + Compressor + Decompressor +} + +// Compressor represents the helper mothods to be used for creating eStargz. +type Compressor interface { + // Writer returns WriteCloser to be used for writing a chunk to eStargz. + // Everytime a chunk is written, the WriteCloser is closed and Writer is + // called again for writing the next chunk. + Writer(w io.Writer) (io.WriteCloser, error) + + // WriteTOCAndFooter is called to write JTOC to the passed Writer. + // diffHash calculates the DiffID (uncompressed sha256 hash) of the blob + // WriteTOCAndFooter can optionally write anything that affects DiffID calculation + // (e.g. uncompressed TOC JSON). + // + // This function returns tocDgst that represents the digest of TOC that will be used + // to verify this blob when it's parsed. + WriteTOCAndFooter(w io.Writer, off int64, toc *JTOC, diffHash hash.Hash) (tocDgst digest.Digest, err error) +} + +// Deompressor represents the helper mothods to be used for parsing eStargz. +type Decompressor interface { + // Reader returns ReadCloser to be used for decompressing file payload. + Reader(r io.Reader) (io.ReadCloser, error) + + // FooterSize returns the size of the footer of this blob. + FooterSize() int64 + + // ParseFooter parses the footer and returns the offset and (compressed) size of TOC. + // + // Here, tocSize is optional. If tocSize <= 0, it's by default the size of the range + // from tocOffset until the beginning of the footer (blob size - tocOff - FooterSize). + ParseFooter(p []byte) (tocOffset, tocSize int64, err error) + + // ParseTOC parses TOC from the passed reader. The reader provides the partial contents + // of the underlying blob that has the range specified by ParseFooter method. + // + // This function returns tocDgst that represents the digest of TOC that will be used + // to verify this blob. This must match to the value returned from + // Compressor.WriteTOCAndFooter that is used when creating this blob. + ParseTOC(r io.Reader) (toc *JTOC, tocDgst digest.Digest, err error) +} diff --git a/vendor/github.com/golang/snappy/.gitignore b/vendor/github.com/golang/snappy/.gitignore deleted file mode 100644 index 042091d9b3..0000000000 --- a/vendor/github.com/golang/snappy/.gitignore +++ /dev/null @@ -1,16 +0,0 @@ -cmd/snappytool/snappytool -testdata/bench - -# These explicitly listed benchmark data files are for an obsolete version of -# snappy_test.go. -testdata/alice29.txt -testdata/asyoulik.txt -testdata/fireworks.jpeg -testdata/geo.protodata -testdata/html -testdata/html_x_4 -testdata/kppkn.gtb -testdata/lcet10.txt -testdata/paper-100k.pdf -testdata/plrabn12.txt -testdata/urls.10K diff --git a/vendor/github.com/golang/snappy/AUTHORS b/vendor/github.com/golang/snappy/AUTHORS deleted file mode 100644 index 52ccb5a934..0000000000 --- a/vendor/github.com/golang/snappy/AUTHORS +++ /dev/null @@ -1,18 +0,0 @@ -# This is the official list of Snappy-Go authors for copyright purposes. -# This file is distinct from the CONTRIBUTORS files. -# See the latter for an explanation. - -# Names should be added to this file as -# Name or Organization -# The email address is not required for organizations. - -# Please keep the list sorted. - -Amazon.com, Inc -Damian Gryski -Eric Buth -Google Inc. -Jan Mercl <0xjnml@gmail.com> -Klaus Post -Rodolfo Carvalho -Sebastien Binet diff --git a/vendor/github.com/golang/snappy/CONTRIBUTORS b/vendor/github.com/golang/snappy/CONTRIBUTORS deleted file mode 100644 index ea6524ddd0..0000000000 --- a/vendor/github.com/golang/snappy/CONTRIBUTORS +++ /dev/null @@ -1,41 +0,0 @@ -# This is the official list of people who can contribute -# (and typically have contributed) code to the Snappy-Go repository. -# The AUTHORS file lists the copyright holders; this file -# lists people. For example, Google employees are listed here -# but not in AUTHORS, because Google holds the copyright. -# -# The submission process automatically checks to make sure -# that people submitting code are listed in this file (by email address). -# -# Names should be added to this file only after verifying that -# the individual or the individual's organization has agreed to -# the appropriate Contributor License Agreement, found here: -# -# http://code.google.com/legal/individual-cla-v1.0.html -# http://code.google.com/legal/corporate-cla-v1.0.html -# -# The agreement for individuals can be filled out on the web. -# -# When adding J Random Contributor's name to this file, -# either J's name or J's organization's name should be -# added to the AUTHORS file, depending on whether the -# individual or corporate CLA was used. - -# Names should be added to this file like so: -# Name - -# Please keep the list sorted. - -Alex Legg -Damian Gryski -Eric Buth -Jan Mercl <0xjnml@gmail.com> -Jonathan Swinney -Kai Backman -Klaus Post -Marc-Antoine Ruel -Nigel Tao -Rob Pike -Rodolfo Carvalho -Russ Cox -Sebastien Binet diff --git a/vendor/github.com/golang/snappy/README b/vendor/github.com/golang/snappy/README deleted file mode 100644 index cea12879a0..0000000000 --- a/vendor/github.com/golang/snappy/README +++ /dev/null @@ -1,107 +0,0 @@ -The Snappy compression format in the Go programming language. - -To download and install from source: -$ go get github.com/golang/snappy - -Unless otherwise noted, the Snappy-Go source files are distributed -under the BSD-style license found in the LICENSE file. - - - -Benchmarks. - -The golang/snappy benchmarks include compressing (Z) and decompressing (U) ten -or so files, the same set used by the C++ Snappy code (github.com/google/snappy -and note the "google", not "golang"). On an "Intel(R) Core(TM) i7-3770 CPU @ -3.40GHz", Go's GOARCH=amd64 numbers as of 2016-05-29: - -"go test -test.bench=." - -_UFlat0-8 2.19GB/s ± 0% html -_UFlat1-8 1.41GB/s ± 0% urls -_UFlat2-8 23.5GB/s ± 2% jpg -_UFlat3-8 1.91GB/s ± 0% jpg_200 -_UFlat4-8 14.0GB/s ± 1% pdf -_UFlat5-8 1.97GB/s ± 0% html4 -_UFlat6-8 814MB/s ± 0% txt1 -_UFlat7-8 785MB/s ± 0% txt2 -_UFlat8-8 857MB/s ± 0% txt3 -_UFlat9-8 719MB/s ± 1% txt4 -_UFlat10-8 2.84GB/s ± 0% pb -_UFlat11-8 1.05GB/s ± 0% gaviota - -_ZFlat0-8 1.04GB/s ± 0% html -_ZFlat1-8 534MB/s ± 0% urls -_ZFlat2-8 15.7GB/s ± 1% jpg -_ZFlat3-8 740MB/s ± 3% jpg_200 -_ZFlat4-8 9.20GB/s ± 1% pdf -_ZFlat5-8 991MB/s ± 0% html4 -_ZFlat6-8 379MB/s ± 0% txt1 -_ZFlat7-8 352MB/s ± 0% txt2 -_ZFlat8-8 396MB/s ± 1% txt3 -_ZFlat9-8 327MB/s ± 1% txt4 -_ZFlat10-8 1.33GB/s ± 1% pb -_ZFlat11-8 605MB/s ± 1% gaviota - - - -"go test -test.bench=. -tags=noasm" - -_UFlat0-8 621MB/s ± 2% html -_UFlat1-8 494MB/s ± 1% urls -_UFlat2-8 23.2GB/s ± 1% jpg -_UFlat3-8 1.12GB/s ± 1% jpg_200 -_UFlat4-8 4.35GB/s ± 1% pdf -_UFlat5-8 609MB/s ± 0% html4 -_UFlat6-8 296MB/s ± 0% txt1 -_UFlat7-8 288MB/s ± 0% txt2 -_UFlat8-8 309MB/s ± 1% txt3 -_UFlat9-8 280MB/s ± 1% txt4 -_UFlat10-8 753MB/s ± 0% pb -_UFlat11-8 400MB/s ± 0% gaviota - -_ZFlat0-8 409MB/s ± 1% html -_ZFlat1-8 250MB/s ± 1% urls -_ZFlat2-8 12.3GB/s ± 1% jpg -_ZFlat3-8 132MB/s ± 0% jpg_200 -_ZFlat4-8 2.92GB/s ± 0% pdf -_ZFlat5-8 405MB/s ± 1% html4 -_ZFlat6-8 179MB/s ± 1% txt1 -_ZFlat7-8 170MB/s ± 1% txt2 -_ZFlat8-8 189MB/s ± 1% txt3 -_ZFlat9-8 164MB/s ± 1% txt4 -_ZFlat10-8 479MB/s ± 1% pb -_ZFlat11-8 270MB/s ± 1% gaviota - - - -For comparison (Go's encoded output is byte-for-byte identical to C++'s), here -are the numbers from C++ Snappy's - -make CXXFLAGS="-O2 -DNDEBUG -g" clean snappy_unittest.log && cat snappy_unittest.log - -BM_UFlat/0 2.4GB/s html -BM_UFlat/1 1.4GB/s urls -BM_UFlat/2 21.8GB/s jpg -BM_UFlat/3 1.5GB/s jpg_200 -BM_UFlat/4 13.3GB/s pdf -BM_UFlat/5 2.1GB/s html4 -BM_UFlat/6 1.0GB/s txt1 -BM_UFlat/7 959.4MB/s txt2 -BM_UFlat/8 1.0GB/s txt3 -BM_UFlat/9 864.5MB/s txt4 -BM_UFlat/10 2.9GB/s pb -BM_UFlat/11 1.2GB/s gaviota - -BM_ZFlat/0 944.3MB/s html (22.31 %) -BM_ZFlat/1 501.6MB/s urls (47.78 %) -BM_ZFlat/2 14.3GB/s jpg (99.95 %) -BM_ZFlat/3 538.3MB/s jpg_200 (73.00 %) -BM_ZFlat/4 8.3GB/s pdf (83.30 %) -BM_ZFlat/5 903.5MB/s html4 (22.52 %) -BM_ZFlat/6 336.0MB/s txt1 (57.88 %) -BM_ZFlat/7 312.3MB/s txt2 (61.91 %) -BM_ZFlat/8 353.1MB/s txt3 (54.99 %) -BM_ZFlat/9 289.9MB/s txt4 (66.26 %) -BM_ZFlat/10 1.2GB/s pb (19.68 %) -BM_ZFlat/11 527.4MB/s gaviota (37.72 %) diff --git a/vendor/github.com/golang/snappy/decode_amd64.s b/vendor/github.com/golang/snappy/decode_amd64.s deleted file mode 100644 index e6179f65e3..0000000000 --- a/vendor/github.com/golang/snappy/decode_amd64.s +++ /dev/null @@ -1,490 +0,0 @@ -// Copyright 2016 The Go Authors. All rights reserved. -// Use of this source code is governed by a BSD-style -// license that can be found in the LICENSE file. - -// +build !appengine -// +build gc -// +build !noasm - -#include "textflag.h" - -// The asm code generally follows the pure Go code in decode_other.go, except -// where marked with a "!!!". - -// func decode(dst, src []byte) int -// -// All local variables fit into registers. The non-zero stack size is only to -// spill registers and push args when issuing a CALL. The register allocation: -// - AX scratch -// - BX scratch -// - CX length or x -// - DX offset -// - SI &src[s] -// - DI &dst[d] -// + R8 dst_base -// + R9 dst_len -// + R10 dst_base + dst_len -// + R11 src_base -// + R12 src_len -// + R13 src_base + src_len -// - R14 used by doCopy -// - R15 used by doCopy -// -// The registers R8-R13 (marked with a "+") are set at the start of the -// function, and after a CALL returns, and are not otherwise modified. -// -// The d variable is implicitly DI - R8, and len(dst)-d is R10 - DI. -// The s variable is implicitly SI - R11, and len(src)-s is R13 - SI. -TEXT ·decode(SB), NOSPLIT, $48-56 - // Initialize SI, DI and R8-R13. - MOVQ dst_base+0(FP), R8 - MOVQ dst_len+8(FP), R9 - MOVQ R8, DI - MOVQ R8, R10 - ADDQ R9, R10 - MOVQ src_base+24(FP), R11 - MOVQ src_len+32(FP), R12 - MOVQ R11, SI - MOVQ R11, R13 - ADDQ R12, R13 - -loop: - // for s < len(src) - CMPQ SI, R13 - JEQ end - - // CX = uint32(src[s]) - // - // switch src[s] & 0x03 - MOVBLZX (SI), CX - MOVL CX, BX - ANDL $3, BX - CMPL BX, $1 - JAE tagCopy - - // ---------------------------------------- - // The code below handles literal tags. - - // case tagLiteral: - // x := uint32(src[s] >> 2) - // switch - SHRL $2, CX - CMPL CX, $60 - JAE tagLit60Plus - - // case x < 60: - // s++ - INCQ SI - -doLit: - // This is the end of the inner "switch", when we have a literal tag. - // - // We assume that CX == x and x fits in a uint32, where x is the variable - // used in the pure Go decode_other.go code. - - // length = int(x) + 1 - // - // Unlike the pure Go code, we don't need to check if length <= 0 because - // CX can hold 64 bits, so the increment cannot overflow. - INCQ CX - - // Prepare to check if copying length bytes will run past the end of dst or - // src. - // - // AX = len(dst) - d - // BX = len(src) - s - MOVQ R10, AX - SUBQ DI, AX - MOVQ R13, BX - SUBQ SI, BX - - // !!! Try a faster technique for short (16 or fewer bytes) copies. - // - // if length > 16 || len(dst)-d < 16 || len(src)-s < 16 { - // goto callMemmove // Fall back on calling runtime·memmove. - // } - // - // The C++ snappy code calls this TryFastAppend. It also checks len(src)-s - // against 21 instead of 16, because it cannot assume that all of its input - // is contiguous in memory and so it needs to leave enough source bytes to - // read the next tag without refilling buffers, but Go's Decode assumes - // contiguousness (the src argument is a []byte). - CMPQ CX, $16 - JGT callMemmove - CMPQ AX, $16 - JLT callMemmove - CMPQ BX, $16 - JLT callMemmove - - // !!! Implement the copy from src to dst as a 16-byte load and store. - // (Decode's documentation says that dst and src must not overlap.) - // - // This always copies 16 bytes, instead of only length bytes, but that's - // OK. If the input is a valid Snappy encoding then subsequent iterations - // will fix up the overrun. Otherwise, Decode returns a nil []byte (and a - // non-nil error), so the overrun will be ignored. - // - // Note that on amd64, it is legal and cheap to issue unaligned 8-byte or - // 16-byte loads and stores. This technique probably wouldn't be as - // effective on architectures that are fussier about alignment. - MOVOU 0(SI), X0 - MOVOU X0, 0(DI) - - // d += length - // s += length - ADDQ CX, DI - ADDQ CX, SI - JMP loop - -callMemmove: - // if length > len(dst)-d || length > len(src)-s { etc } - CMPQ CX, AX - JGT errCorrupt - CMPQ CX, BX - JGT errCorrupt - - // copy(dst[d:], src[s:s+length]) - // - // This means calling runtime·memmove(&dst[d], &src[s], length), so we push - // DI, SI and CX as arguments. Coincidentally, we also need to spill those - // three registers to the stack, to save local variables across the CALL. - MOVQ DI, 0(SP) - MOVQ SI, 8(SP) - MOVQ CX, 16(SP) - MOVQ DI, 24(SP) - MOVQ SI, 32(SP) - MOVQ CX, 40(SP) - CALL runtime·memmove(SB) - - // Restore local variables: unspill registers from the stack and - // re-calculate R8-R13. - MOVQ 24(SP), DI - MOVQ 32(SP), SI - MOVQ 40(SP), CX - MOVQ dst_base+0(FP), R8 - MOVQ dst_len+8(FP), R9 - MOVQ R8, R10 - ADDQ R9, R10 - MOVQ src_base+24(FP), R11 - MOVQ src_len+32(FP), R12 - MOVQ R11, R13 - ADDQ R12, R13 - - // d += length - // s += length - ADDQ CX, DI - ADDQ CX, SI - JMP loop - -tagLit60Plus: - // !!! This fragment does the - // - // s += x - 58; if uint(s) > uint(len(src)) { etc } - // - // checks. In the asm version, we code it once instead of once per switch case. - ADDQ CX, SI - SUBQ $58, SI - MOVQ SI, BX - SUBQ R11, BX - CMPQ BX, R12 - JA errCorrupt - - // case x == 60: - CMPL CX, $61 - JEQ tagLit61 - JA tagLit62Plus - - // x = uint32(src[s-1]) - MOVBLZX -1(SI), CX - JMP doLit - -tagLit61: - // case x == 61: - // x = uint32(src[s-2]) | uint32(src[s-1])<<8 - MOVWLZX -2(SI), CX - JMP doLit - -tagLit62Plus: - CMPL CX, $62 - JA tagLit63 - - // case x == 62: - // x = uint32(src[s-3]) | uint32(src[s-2])<<8 | uint32(src[s-1])<<16 - MOVWLZX -3(SI), CX - MOVBLZX -1(SI), BX - SHLL $16, BX - ORL BX, CX - JMP doLit - -tagLit63: - // case x == 63: - // x = uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24 - MOVL -4(SI), CX - JMP doLit - -// The code above handles literal tags. -// ---------------------------------------- -// The code below handles copy tags. - -tagCopy4: - // case tagCopy4: - // s += 5 - ADDQ $5, SI - - // if uint(s) > uint(len(src)) { etc } - MOVQ SI, BX - SUBQ R11, BX - CMPQ BX, R12 - JA errCorrupt - - // length = 1 + int(src[s-5])>>2 - SHRQ $2, CX - INCQ CX - - // offset = int(uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24) - MOVLQZX -4(SI), DX - JMP doCopy - -tagCopy2: - // case tagCopy2: - // s += 3 - ADDQ $3, SI - - // if uint(s) > uint(len(src)) { etc } - MOVQ SI, BX - SUBQ R11, BX - CMPQ BX, R12 - JA errCorrupt - - // length = 1 + int(src[s-3])>>2 - SHRQ $2, CX - INCQ CX - - // offset = int(uint32(src[s-2]) | uint32(src[s-1])<<8) - MOVWQZX -2(SI), DX - JMP doCopy - -tagCopy: - // We have a copy tag. We assume that: - // - BX == src[s] & 0x03 - // - CX == src[s] - CMPQ BX, $2 - JEQ tagCopy2 - JA tagCopy4 - - // case tagCopy1: - // s += 2 - ADDQ $2, SI - - // if uint(s) > uint(len(src)) { etc } - MOVQ SI, BX - SUBQ R11, BX - CMPQ BX, R12 - JA errCorrupt - - // offset = int(uint32(src[s-2])&0xe0<<3 | uint32(src[s-1])) - MOVQ CX, DX - ANDQ $0xe0, DX - SHLQ $3, DX - MOVBQZX -1(SI), BX - ORQ BX, DX - - // length = 4 + int(src[s-2])>>2&0x7 - SHRQ $2, CX - ANDQ $7, CX - ADDQ $4, CX - -doCopy: - // This is the end of the outer "switch", when we have a copy tag. - // - // We assume that: - // - CX == length && CX > 0 - // - DX == offset - - // if offset <= 0 { etc } - CMPQ DX, $0 - JLE errCorrupt - - // if d < offset { etc } - MOVQ DI, BX - SUBQ R8, BX - CMPQ BX, DX - JLT errCorrupt - - // if length > len(dst)-d { etc } - MOVQ R10, BX - SUBQ DI, BX - CMPQ CX, BX - JGT errCorrupt - - // forwardCopy(dst[d:d+length], dst[d-offset:]); d += length - // - // Set: - // - R14 = len(dst)-d - // - R15 = &dst[d-offset] - MOVQ R10, R14 - SUBQ DI, R14 - MOVQ DI, R15 - SUBQ DX, R15 - - // !!! Try a faster technique for short (16 or fewer bytes) forward copies. - // - // First, try using two 8-byte load/stores, similar to the doLit technique - // above. Even if dst[d:d+length] and dst[d-offset:] can overlap, this is - // still OK if offset >= 8. Note that this has to be two 8-byte load/stores - // and not one 16-byte load/store, and the first store has to be before the - // second load, due to the overlap if offset is in the range [8, 16). - // - // if length > 16 || offset < 8 || len(dst)-d < 16 { - // goto slowForwardCopy - // } - // copy 16 bytes - // d += length - CMPQ CX, $16 - JGT slowForwardCopy - CMPQ DX, $8 - JLT slowForwardCopy - CMPQ R14, $16 - JLT slowForwardCopy - MOVQ 0(R15), AX - MOVQ AX, 0(DI) - MOVQ 8(R15), BX - MOVQ BX, 8(DI) - ADDQ CX, DI - JMP loop - -slowForwardCopy: - // !!! If the forward copy is longer than 16 bytes, or if offset < 8, we - // can still try 8-byte load stores, provided we can overrun up to 10 extra - // bytes. As above, the overrun will be fixed up by subsequent iterations - // of the outermost loop. - // - // The C++ snappy code calls this technique IncrementalCopyFastPath. Its - // commentary says: - // - // ---- - // - // The main part of this loop is a simple copy of eight bytes at a time - // until we've copied (at least) the requested amount of bytes. However, - // if d and d-offset are less than eight bytes apart (indicating a - // repeating pattern of length < 8), we first need to expand the pattern in - // order to get the correct results. For instance, if the buffer looks like - // this, with the eight-byte and patterns marked as - // intervals: - // - // abxxxxxxxxxxxx - // [------] d-offset - // [------] d - // - // a single eight-byte copy from to will repeat the pattern - // once, after which we can move two bytes without moving : - // - // ababxxxxxxxxxx - // [------] d-offset - // [------] d - // - // and repeat the exercise until the two no longer overlap. - // - // This allows us to do very well in the special case of one single byte - // repeated many times, without taking a big hit for more general cases. - // - // The worst case of extra writing past the end of the match occurs when - // offset == 1 and length == 1; the last copy will read from byte positions - // [0..7] and write to [4..11], whereas it was only supposed to write to - // position 1. Thus, ten excess bytes. - // - // ---- - // - // That "10 byte overrun" worst case is confirmed by Go's - // TestSlowForwardCopyOverrun, which also tests the fixUpSlowForwardCopy - // and finishSlowForwardCopy algorithm. - // - // if length > len(dst)-d-10 { - // goto verySlowForwardCopy - // } - SUBQ $10, R14 - CMPQ CX, R14 - JGT verySlowForwardCopy - -makeOffsetAtLeast8: - // !!! As above, expand the pattern so that offset >= 8 and we can use - // 8-byte load/stores. - // - // for offset < 8 { - // copy 8 bytes from dst[d-offset:] to dst[d:] - // length -= offset - // d += offset - // offset += offset - // // The two previous lines together means that d-offset, and therefore - // // R15, is unchanged. - // } - CMPQ DX, $8 - JGE fixUpSlowForwardCopy - MOVQ (R15), BX - MOVQ BX, (DI) - SUBQ DX, CX - ADDQ DX, DI - ADDQ DX, DX - JMP makeOffsetAtLeast8 - -fixUpSlowForwardCopy: - // !!! Add length (which might be negative now) to d (implied by DI being - // &dst[d]) so that d ends up at the right place when we jump back to the - // top of the loop. Before we do that, though, we save DI to AX so that, if - // length is positive, copying the remaining length bytes will write to the - // right place. - MOVQ DI, AX - ADDQ CX, DI - -finishSlowForwardCopy: - // !!! Repeat 8-byte load/stores until length <= 0. Ending with a negative - // length means that we overrun, but as above, that will be fixed up by - // subsequent iterations of the outermost loop. - CMPQ CX, $0 - JLE loop - MOVQ (R15), BX - MOVQ BX, (AX) - ADDQ $8, R15 - ADDQ $8, AX - SUBQ $8, CX - JMP finishSlowForwardCopy - -verySlowForwardCopy: - // verySlowForwardCopy is a simple implementation of forward copy. In C - // parlance, this is a do/while loop instead of a while loop, since we know - // that length > 0. In Go syntax: - // - // for { - // dst[d] = dst[d - offset] - // d++ - // length-- - // if length == 0 { - // break - // } - // } - MOVB (R15), BX - MOVB BX, (DI) - INCQ R15 - INCQ DI - DECQ CX - JNZ verySlowForwardCopy - JMP loop - -// The code above handles copy tags. -// ---------------------------------------- - -end: - // This is the end of the "for s < len(src)". - // - // if d != len(dst) { etc } - CMPQ DI, R10 - JNE errCorrupt - - // return 0 - MOVQ $0, ret+48(FP) - RET - -errCorrupt: - // return decodeErrCodeCorrupt - MOVQ $1, ret+48(FP) - RET diff --git a/vendor/github.com/golang/snappy/decode_arm64.s b/vendor/github.com/golang/snappy/decode_arm64.s deleted file mode 100644 index 7a3ead17ea..0000000000 --- a/vendor/github.com/golang/snappy/decode_arm64.s +++ /dev/null @@ -1,494 +0,0 @@ -// Copyright 2020 The Go Authors. All rights reserved. -// Use of this source code is governed by a BSD-style -// license that can be found in the LICENSE file. - -// +build !appengine -// +build gc -// +build !noasm - -#include "textflag.h" - -// The asm code generally follows the pure Go code in decode_other.go, except -// where marked with a "!!!". - -// func decode(dst, src []byte) int -// -// All local variables fit into registers. The non-zero stack size is only to -// spill registers and push args when issuing a CALL. The register allocation: -// - R2 scratch -// - R3 scratch -// - R4 length or x -// - R5 offset -// - R6 &src[s] -// - R7 &dst[d] -// + R8 dst_base -// + R9 dst_len -// + R10 dst_base + dst_len -// + R11 src_base -// + R12 src_len -// + R13 src_base + src_len -// - R14 used by doCopy -// - R15 used by doCopy -// -// The registers R8-R13 (marked with a "+") are set at the start of the -// function, and after a CALL returns, and are not otherwise modified. -// -// The d variable is implicitly R7 - R8, and len(dst)-d is R10 - R7. -// The s variable is implicitly R6 - R11, and len(src)-s is R13 - R6. -TEXT ·decode(SB), NOSPLIT, $56-56 - // Initialize R6, R7 and R8-R13. - MOVD dst_base+0(FP), R8 - MOVD dst_len+8(FP), R9 - MOVD R8, R7 - MOVD R8, R10 - ADD R9, R10, R10 - MOVD src_base+24(FP), R11 - MOVD src_len+32(FP), R12 - MOVD R11, R6 - MOVD R11, R13 - ADD R12, R13, R13 - -loop: - // for s < len(src) - CMP R13, R6 - BEQ end - - // R4 = uint32(src[s]) - // - // switch src[s] & 0x03 - MOVBU (R6), R4 - MOVW R4, R3 - ANDW $3, R3 - MOVW $1, R1 - CMPW R1, R3 - BGE tagCopy - - // ---------------------------------------- - // The code below handles literal tags. - - // case tagLiteral: - // x := uint32(src[s] >> 2) - // switch - MOVW $60, R1 - LSRW $2, R4, R4 - CMPW R4, R1 - BLS tagLit60Plus - - // case x < 60: - // s++ - ADD $1, R6, R6 - -doLit: - // This is the end of the inner "switch", when we have a literal tag. - // - // We assume that R4 == x and x fits in a uint32, where x is the variable - // used in the pure Go decode_other.go code. - - // length = int(x) + 1 - // - // Unlike the pure Go code, we don't need to check if length <= 0 because - // R4 can hold 64 bits, so the increment cannot overflow. - ADD $1, R4, R4 - - // Prepare to check if copying length bytes will run past the end of dst or - // src. - // - // R2 = len(dst) - d - // R3 = len(src) - s - MOVD R10, R2 - SUB R7, R2, R2 - MOVD R13, R3 - SUB R6, R3, R3 - - // !!! Try a faster technique for short (16 or fewer bytes) copies. - // - // if length > 16 || len(dst)-d < 16 || len(src)-s < 16 { - // goto callMemmove // Fall back on calling runtime·memmove. - // } - // - // The C++ snappy code calls this TryFastAppend. It also checks len(src)-s - // against 21 instead of 16, because it cannot assume that all of its input - // is contiguous in memory and so it needs to leave enough source bytes to - // read the next tag without refilling buffers, but Go's Decode assumes - // contiguousness (the src argument is a []byte). - CMP $16, R4 - BGT callMemmove - CMP $16, R2 - BLT callMemmove - CMP $16, R3 - BLT callMemmove - - // !!! Implement the copy from src to dst as a 16-byte load and store. - // (Decode's documentation says that dst and src must not overlap.) - // - // This always copies 16 bytes, instead of only length bytes, but that's - // OK. If the input is a valid Snappy encoding then subsequent iterations - // will fix up the overrun. Otherwise, Decode returns a nil []byte (and a - // non-nil error), so the overrun will be ignored. - // - // Note that on arm64, it is legal and cheap to issue unaligned 8-byte or - // 16-byte loads and stores. This technique probably wouldn't be as - // effective on architectures that are fussier about alignment. - LDP 0(R6), (R14, R15) - STP (R14, R15), 0(R7) - - // d += length - // s += length - ADD R4, R7, R7 - ADD R4, R6, R6 - B loop - -callMemmove: - // if length > len(dst)-d || length > len(src)-s { etc } - CMP R2, R4 - BGT errCorrupt - CMP R3, R4 - BGT errCorrupt - - // copy(dst[d:], src[s:s+length]) - // - // This means calling runtime·memmove(&dst[d], &src[s], length), so we push - // R7, R6 and R4 as arguments. Coincidentally, we also need to spill those - // three registers to the stack, to save local variables across the CALL. - MOVD R7, 8(RSP) - MOVD R6, 16(RSP) - MOVD R4, 24(RSP) - MOVD R7, 32(RSP) - MOVD R6, 40(RSP) - MOVD R4, 48(RSP) - CALL runtime·memmove(SB) - - // Restore local variables: unspill registers from the stack and - // re-calculate R8-R13. - MOVD 32(RSP), R7 - MOVD 40(RSP), R6 - MOVD 48(RSP), R4 - MOVD dst_base+0(FP), R8 - MOVD dst_len+8(FP), R9 - MOVD R8, R10 - ADD R9, R10, R10 - MOVD src_base+24(FP), R11 - MOVD src_len+32(FP), R12 - MOVD R11, R13 - ADD R12, R13, R13 - - // d += length - // s += length - ADD R4, R7, R7 - ADD R4, R6, R6 - B loop - -tagLit60Plus: - // !!! This fragment does the - // - // s += x - 58; if uint(s) > uint(len(src)) { etc } - // - // checks. In the asm version, we code it once instead of once per switch case. - ADD R4, R6, R6 - SUB $58, R6, R6 - MOVD R6, R3 - SUB R11, R3, R3 - CMP R12, R3 - BGT errCorrupt - - // case x == 60: - MOVW $61, R1 - CMPW R1, R4 - BEQ tagLit61 - BGT tagLit62Plus - - // x = uint32(src[s-1]) - MOVBU -1(R6), R4 - B doLit - -tagLit61: - // case x == 61: - // x = uint32(src[s-2]) | uint32(src[s-1])<<8 - MOVHU -2(R6), R4 - B doLit - -tagLit62Plus: - CMPW $62, R4 - BHI tagLit63 - - // case x == 62: - // x = uint32(src[s-3]) | uint32(src[s-2])<<8 | uint32(src[s-1])<<16 - MOVHU -3(R6), R4 - MOVBU -1(R6), R3 - ORR R3<<16, R4 - B doLit - -tagLit63: - // case x == 63: - // x = uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24 - MOVWU -4(R6), R4 - B doLit - - // The code above handles literal tags. - // ---------------------------------------- - // The code below handles copy tags. - -tagCopy4: - // case tagCopy4: - // s += 5 - ADD $5, R6, R6 - - // if uint(s) > uint(len(src)) { etc } - MOVD R6, R3 - SUB R11, R3, R3 - CMP R12, R3 - BGT errCorrupt - - // length = 1 + int(src[s-5])>>2 - MOVD $1, R1 - ADD R4>>2, R1, R4 - - // offset = int(uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24) - MOVWU -4(R6), R5 - B doCopy - -tagCopy2: - // case tagCopy2: - // s += 3 - ADD $3, R6, R6 - - // if uint(s) > uint(len(src)) { etc } - MOVD R6, R3 - SUB R11, R3, R3 - CMP R12, R3 - BGT errCorrupt - - // length = 1 + int(src[s-3])>>2 - MOVD $1, R1 - ADD R4>>2, R1, R4 - - // offset = int(uint32(src[s-2]) | uint32(src[s-1])<<8) - MOVHU -2(R6), R5 - B doCopy - -tagCopy: - // We have a copy tag. We assume that: - // - R3 == src[s] & 0x03 - // - R4 == src[s] - CMP $2, R3 - BEQ tagCopy2 - BGT tagCopy4 - - // case tagCopy1: - // s += 2 - ADD $2, R6, R6 - - // if uint(s) > uint(len(src)) { etc } - MOVD R6, R3 - SUB R11, R3, R3 - CMP R12, R3 - BGT errCorrupt - - // offset = int(uint32(src[s-2])&0xe0<<3 | uint32(src[s-1])) - MOVD R4, R5 - AND $0xe0, R5 - MOVBU -1(R6), R3 - ORR R5<<3, R3, R5 - - // length = 4 + int(src[s-2])>>2&0x7 - MOVD $7, R1 - AND R4>>2, R1, R4 - ADD $4, R4, R4 - -doCopy: - // This is the end of the outer "switch", when we have a copy tag. - // - // We assume that: - // - R4 == length && R4 > 0 - // - R5 == offset - - // if offset <= 0 { etc } - MOVD $0, R1 - CMP R1, R5 - BLE errCorrupt - - // if d < offset { etc } - MOVD R7, R3 - SUB R8, R3, R3 - CMP R5, R3 - BLT errCorrupt - - // if length > len(dst)-d { etc } - MOVD R10, R3 - SUB R7, R3, R3 - CMP R3, R4 - BGT errCorrupt - - // forwardCopy(dst[d:d+length], dst[d-offset:]); d += length - // - // Set: - // - R14 = len(dst)-d - // - R15 = &dst[d-offset] - MOVD R10, R14 - SUB R7, R14, R14 - MOVD R7, R15 - SUB R5, R15, R15 - - // !!! Try a faster technique for short (16 or fewer bytes) forward copies. - // - // First, try using two 8-byte load/stores, similar to the doLit technique - // above. Even if dst[d:d+length] and dst[d-offset:] can overlap, this is - // still OK if offset >= 8. Note that this has to be two 8-byte load/stores - // and not one 16-byte load/store, and the first store has to be before the - // second load, due to the overlap if offset is in the range [8, 16). - // - // if length > 16 || offset < 8 || len(dst)-d < 16 { - // goto slowForwardCopy - // } - // copy 16 bytes - // d += length - CMP $16, R4 - BGT slowForwardCopy - CMP $8, R5 - BLT slowForwardCopy - CMP $16, R14 - BLT slowForwardCopy - MOVD 0(R15), R2 - MOVD R2, 0(R7) - MOVD 8(R15), R3 - MOVD R3, 8(R7) - ADD R4, R7, R7 - B loop - -slowForwardCopy: - // !!! If the forward copy is longer than 16 bytes, or if offset < 8, we - // can still try 8-byte load stores, provided we can overrun up to 10 extra - // bytes. As above, the overrun will be fixed up by subsequent iterations - // of the outermost loop. - // - // The C++ snappy code calls this technique IncrementalCopyFastPath. Its - // commentary says: - // - // ---- - // - // The main part of this loop is a simple copy of eight bytes at a time - // until we've copied (at least) the requested amount of bytes. However, - // if d and d-offset are less than eight bytes apart (indicating a - // repeating pattern of length < 8), we first need to expand the pattern in - // order to get the correct results. For instance, if the buffer looks like - // this, with the eight-byte and patterns marked as - // intervals: - // - // abxxxxxxxxxxxx - // [------] d-offset - // [------] d - // - // a single eight-byte copy from to will repeat the pattern - // once, after which we can move two bytes without moving : - // - // ababxxxxxxxxxx - // [------] d-offset - // [------] d - // - // and repeat the exercise until the two no longer overlap. - // - // This allows us to do very well in the special case of one single byte - // repeated many times, without taking a big hit for more general cases. - // - // The worst case of extra writing past the end of the match occurs when - // offset == 1 and length == 1; the last copy will read from byte positions - // [0..7] and write to [4..11], whereas it was only supposed to write to - // position 1. Thus, ten excess bytes. - // - // ---- - // - // That "10 byte overrun" worst case is confirmed by Go's - // TestSlowForwardCopyOverrun, which also tests the fixUpSlowForwardCopy - // and finishSlowForwardCopy algorithm. - // - // if length > len(dst)-d-10 { - // goto verySlowForwardCopy - // } - SUB $10, R14, R14 - CMP R14, R4 - BGT verySlowForwardCopy - -makeOffsetAtLeast8: - // !!! As above, expand the pattern so that offset >= 8 and we can use - // 8-byte load/stores. - // - // for offset < 8 { - // copy 8 bytes from dst[d-offset:] to dst[d:] - // length -= offset - // d += offset - // offset += offset - // // The two previous lines together means that d-offset, and therefore - // // R15, is unchanged. - // } - CMP $8, R5 - BGE fixUpSlowForwardCopy - MOVD (R15), R3 - MOVD R3, (R7) - SUB R5, R4, R4 - ADD R5, R7, R7 - ADD R5, R5, R5 - B makeOffsetAtLeast8 - -fixUpSlowForwardCopy: - // !!! Add length (which might be negative now) to d (implied by R7 being - // &dst[d]) so that d ends up at the right place when we jump back to the - // top of the loop. Before we do that, though, we save R7 to R2 so that, if - // length is positive, copying the remaining length bytes will write to the - // right place. - MOVD R7, R2 - ADD R4, R7, R7 - -finishSlowForwardCopy: - // !!! Repeat 8-byte load/stores until length <= 0. Ending with a negative - // length means that we overrun, but as above, that will be fixed up by - // subsequent iterations of the outermost loop. - MOVD $0, R1 - CMP R1, R4 - BLE loop - MOVD (R15), R3 - MOVD R3, (R2) - ADD $8, R15, R15 - ADD $8, R2, R2 - SUB $8, R4, R4 - B finishSlowForwardCopy - -verySlowForwardCopy: - // verySlowForwardCopy is a simple implementation of forward copy. In C - // parlance, this is a do/while loop instead of a while loop, since we know - // that length > 0. In Go syntax: - // - // for { - // dst[d] = dst[d - offset] - // d++ - // length-- - // if length == 0 { - // break - // } - // } - MOVB (R15), R3 - MOVB R3, (R7) - ADD $1, R15, R15 - ADD $1, R7, R7 - SUB $1, R4, R4 - CBNZ R4, verySlowForwardCopy - B loop - - // The code above handles copy tags. - // ---------------------------------------- - -end: - // This is the end of the "for s < len(src)". - // - // if d != len(dst) { etc } - CMP R10, R7 - BNE errCorrupt - - // return 0 - MOVD $0, ret+48(FP) - RET - -errCorrupt: - // return decodeErrCodeCorrupt - MOVD $1, R2 - MOVD R2, ret+48(FP) - RET diff --git a/vendor/github.com/golang/snappy/decode_asm.go b/vendor/github.com/golang/snappy/decode_asm.go deleted file mode 100644 index 7082b34919..0000000000 --- a/vendor/github.com/golang/snappy/decode_asm.go +++ /dev/null @@ -1,15 +0,0 @@ -// Copyright 2016 The Snappy-Go Authors. All rights reserved. -// Use of this source code is governed by a BSD-style -// license that can be found in the LICENSE file. - -// +build !appengine -// +build gc -// +build !noasm -// +build amd64 arm64 - -package snappy - -// decode has the same semantics as in decode_other.go. -// -//go:noescape -func decode(dst, src []byte) int diff --git a/vendor/github.com/golang/snappy/encode_amd64.s b/vendor/github.com/golang/snappy/encode_amd64.s deleted file mode 100644 index adfd979fe2..0000000000 --- a/vendor/github.com/golang/snappy/encode_amd64.s +++ /dev/null @@ -1,730 +0,0 @@ -// Copyright 2016 The Go Authors. All rights reserved. -// Use of this source code is governed by a BSD-style -// license that can be found in the LICENSE file. - -// +build !appengine -// +build gc -// +build !noasm - -#include "textflag.h" - -// The XXX lines assemble on Go 1.4, 1.5 and 1.7, but not 1.6, due to a -// Go toolchain regression. See https://github.com/golang/go/issues/15426 and -// https://github.com/golang/snappy/issues/29 -// -// As a workaround, the package was built with a known good assembler, and -// those instructions were disassembled by "objdump -d" to yield the -// 4e 0f b7 7c 5c 78 movzwq 0x78(%rsp,%r11,2),%r15 -// style comments, in AT&T asm syntax. Note that rsp here is a physical -// register, not Go/asm's SP pseudo-register (see https://golang.org/doc/asm). -// The instructions were then encoded as "BYTE $0x.." sequences, which assemble -// fine on Go 1.6. - -// The asm code generally follows the pure Go code in encode_other.go, except -// where marked with a "!!!". - -// ---------------------------------------------------------------------------- - -// func emitLiteral(dst, lit []byte) int -// -// All local variables fit into registers. The register allocation: -// - AX len(lit) -// - BX n -// - DX return value -// - DI &dst[i] -// - R10 &lit[0] -// -// The 24 bytes of stack space is to call runtime·memmove. -// -// The unusual register allocation of local variables, such as R10 for the -// source pointer, matches the allocation used at the call site in encodeBlock, -// which makes it easier to manually inline this function. -TEXT ·emitLiteral(SB), NOSPLIT, $24-56 - MOVQ dst_base+0(FP), DI - MOVQ lit_base+24(FP), R10 - MOVQ lit_len+32(FP), AX - MOVQ AX, DX - MOVL AX, BX - SUBL $1, BX - - CMPL BX, $60 - JLT oneByte - CMPL BX, $256 - JLT twoBytes - -threeBytes: - MOVB $0xf4, 0(DI) - MOVW BX, 1(DI) - ADDQ $3, DI - ADDQ $3, DX - JMP memmove - -twoBytes: - MOVB $0xf0, 0(DI) - MOVB BX, 1(DI) - ADDQ $2, DI - ADDQ $2, DX - JMP memmove - -oneByte: - SHLB $2, BX - MOVB BX, 0(DI) - ADDQ $1, DI - ADDQ $1, DX - -memmove: - MOVQ DX, ret+48(FP) - - // copy(dst[i:], lit) - // - // This means calling runtime·memmove(&dst[i], &lit[0], len(lit)), so we push - // DI, R10 and AX as arguments. - MOVQ DI, 0(SP) - MOVQ R10, 8(SP) - MOVQ AX, 16(SP) - CALL runtime·memmove(SB) - RET - -// ---------------------------------------------------------------------------- - -// func emitCopy(dst []byte, offset, length int) int -// -// All local variables fit into registers. The register allocation: -// - AX length -// - SI &dst[0] -// - DI &dst[i] -// - R11 offset -// -// The unusual register allocation of local variables, such as R11 for the -// offset, matches the allocation used at the call site in encodeBlock, which -// makes it easier to manually inline this function. -TEXT ·emitCopy(SB), NOSPLIT, $0-48 - MOVQ dst_base+0(FP), DI - MOVQ DI, SI - MOVQ offset+24(FP), R11 - MOVQ length+32(FP), AX - -loop0: - // for length >= 68 { etc } - CMPL AX, $68 - JLT step1 - - // Emit a length 64 copy, encoded as 3 bytes. - MOVB $0xfe, 0(DI) - MOVW R11, 1(DI) - ADDQ $3, DI - SUBL $64, AX - JMP loop0 - -step1: - // if length > 64 { etc } - CMPL AX, $64 - JLE step2 - - // Emit a length 60 copy, encoded as 3 bytes. - MOVB $0xee, 0(DI) - MOVW R11, 1(DI) - ADDQ $3, DI - SUBL $60, AX - -step2: - // if length >= 12 || offset >= 2048 { goto step3 } - CMPL AX, $12 - JGE step3 - CMPL R11, $2048 - JGE step3 - - // Emit the remaining copy, encoded as 2 bytes. - MOVB R11, 1(DI) - SHRL $8, R11 - SHLB $5, R11 - SUBB $4, AX - SHLB $2, AX - ORB AX, R11 - ORB $1, R11 - MOVB R11, 0(DI) - ADDQ $2, DI - - // Return the number of bytes written. - SUBQ SI, DI - MOVQ DI, ret+40(FP) - RET - -step3: - // Emit the remaining copy, encoded as 3 bytes. - SUBL $1, AX - SHLB $2, AX - ORB $2, AX - MOVB AX, 0(DI) - MOVW R11, 1(DI) - ADDQ $3, DI - - // Return the number of bytes written. - SUBQ SI, DI - MOVQ DI, ret+40(FP) - RET - -// ---------------------------------------------------------------------------- - -// func extendMatch(src []byte, i, j int) int -// -// All local variables fit into registers. The register allocation: -// - DX &src[0] -// - SI &src[j] -// - R13 &src[len(src) - 8] -// - R14 &src[len(src)] -// - R15 &src[i] -// -// The unusual register allocation of local variables, such as R15 for a source -// pointer, matches the allocation used at the call site in encodeBlock, which -// makes it easier to manually inline this function. -TEXT ·extendMatch(SB), NOSPLIT, $0-48 - MOVQ src_base+0(FP), DX - MOVQ src_len+8(FP), R14 - MOVQ i+24(FP), R15 - MOVQ j+32(FP), SI - ADDQ DX, R14 - ADDQ DX, R15 - ADDQ DX, SI - MOVQ R14, R13 - SUBQ $8, R13 - -cmp8: - // As long as we are 8 or more bytes before the end of src, we can load and - // compare 8 bytes at a time. If those 8 bytes are equal, repeat. - CMPQ SI, R13 - JA cmp1 - MOVQ (R15), AX - MOVQ (SI), BX - CMPQ AX, BX - JNE bsf - ADDQ $8, R15 - ADDQ $8, SI - JMP cmp8 - -bsf: - // If those 8 bytes were not equal, XOR the two 8 byte values, and return - // the index of the first byte that differs. The BSF instruction finds the - // least significant 1 bit, the amd64 architecture is little-endian, and - // the shift by 3 converts a bit index to a byte index. - XORQ AX, BX - BSFQ BX, BX - SHRQ $3, BX - ADDQ BX, SI - - // Convert from &src[ret] to ret. - SUBQ DX, SI - MOVQ SI, ret+40(FP) - RET - -cmp1: - // In src's tail, compare 1 byte at a time. - CMPQ SI, R14 - JAE extendMatchEnd - MOVB (R15), AX - MOVB (SI), BX - CMPB AX, BX - JNE extendMatchEnd - ADDQ $1, R15 - ADDQ $1, SI - JMP cmp1 - -extendMatchEnd: - // Convert from &src[ret] to ret. - SUBQ DX, SI - MOVQ SI, ret+40(FP) - RET - -// ---------------------------------------------------------------------------- - -// func encodeBlock(dst, src []byte) (d int) -// -// All local variables fit into registers, other than "var table". The register -// allocation: -// - AX . . -// - BX . . -// - CX 56 shift (note that amd64 shifts by non-immediates must use CX). -// - DX 64 &src[0], tableSize -// - SI 72 &src[s] -// - DI 80 &dst[d] -// - R9 88 sLimit -// - R10 . &src[nextEmit] -// - R11 96 prevHash, currHash, nextHash, offset -// - R12 104 &src[base], skip -// - R13 . &src[nextS], &src[len(src) - 8] -// - R14 . len(src), bytesBetweenHashLookups, &src[len(src)], x -// - R15 112 candidate -// -// The second column (56, 64, etc) is the stack offset to spill the registers -// when calling other functions. We could pack this slightly tighter, but it's -// simpler to have a dedicated spill map independent of the function called. -// -// "var table [maxTableSize]uint16" takes up 32768 bytes of stack space. An -// extra 56 bytes, to call other functions, and an extra 64 bytes, to spill -// local variables (registers) during calls gives 32768 + 56 + 64 = 32888. -TEXT ·encodeBlock(SB), 0, $32888-56 - MOVQ dst_base+0(FP), DI - MOVQ src_base+24(FP), SI - MOVQ src_len+32(FP), R14 - - // shift, tableSize := uint32(32-8), 1<<8 - MOVQ $24, CX - MOVQ $256, DX - -calcShift: - // for ; tableSize < maxTableSize && tableSize < len(src); tableSize *= 2 { - // shift-- - // } - CMPQ DX, $16384 - JGE varTable - CMPQ DX, R14 - JGE varTable - SUBQ $1, CX - SHLQ $1, DX - JMP calcShift - -varTable: - // var table [maxTableSize]uint16 - // - // In the asm code, unlike the Go code, we can zero-initialize only the - // first tableSize elements. Each uint16 element is 2 bytes and each MOVOU - // writes 16 bytes, so we can do only tableSize/8 writes instead of the - // 2048 writes that would zero-initialize all of table's 32768 bytes. - SHRQ $3, DX - LEAQ table-32768(SP), BX - PXOR X0, X0 - -memclr: - MOVOU X0, 0(BX) - ADDQ $16, BX - SUBQ $1, DX - JNZ memclr - - // !!! DX = &src[0] - MOVQ SI, DX - - // sLimit := len(src) - inputMargin - MOVQ R14, R9 - SUBQ $15, R9 - - // !!! Pre-emptively spill CX, DX and R9 to the stack. Their values don't - // change for the rest of the function. - MOVQ CX, 56(SP) - MOVQ DX, 64(SP) - MOVQ R9, 88(SP) - - // nextEmit := 0 - MOVQ DX, R10 - - // s := 1 - ADDQ $1, SI - - // nextHash := hash(load32(src, s), shift) - MOVL 0(SI), R11 - IMULL $0x1e35a7bd, R11 - SHRL CX, R11 - -outer: - // for { etc } - - // skip := 32 - MOVQ $32, R12 - - // nextS := s - MOVQ SI, R13 - - // candidate := 0 - MOVQ $0, R15 - -inner0: - // for { etc } - - // s := nextS - MOVQ R13, SI - - // bytesBetweenHashLookups := skip >> 5 - MOVQ R12, R14 - SHRQ $5, R14 - - // nextS = s + bytesBetweenHashLookups - ADDQ R14, R13 - - // skip += bytesBetweenHashLookups - ADDQ R14, R12 - - // if nextS > sLimit { goto emitRemainder } - MOVQ R13, AX - SUBQ DX, AX - CMPQ AX, R9 - JA emitRemainder - - // candidate = int(table[nextHash]) - // XXX: MOVWQZX table-32768(SP)(R11*2), R15 - // XXX: 4e 0f b7 7c 5c 78 movzwq 0x78(%rsp,%r11,2),%r15 - BYTE $0x4e - BYTE $0x0f - BYTE $0xb7 - BYTE $0x7c - BYTE $0x5c - BYTE $0x78 - - // table[nextHash] = uint16(s) - MOVQ SI, AX - SUBQ DX, AX - - // XXX: MOVW AX, table-32768(SP)(R11*2) - // XXX: 66 42 89 44 5c 78 mov %ax,0x78(%rsp,%r11,2) - BYTE $0x66 - BYTE $0x42 - BYTE $0x89 - BYTE $0x44 - BYTE $0x5c - BYTE $0x78 - - // nextHash = hash(load32(src, nextS), shift) - MOVL 0(R13), R11 - IMULL $0x1e35a7bd, R11 - SHRL CX, R11 - - // if load32(src, s) != load32(src, candidate) { continue } break - MOVL 0(SI), AX - MOVL (DX)(R15*1), BX - CMPL AX, BX - JNE inner0 - -fourByteMatch: - // As per the encode_other.go code: - // - // A 4-byte match has been found. We'll later see etc. - - // !!! Jump to a fast path for short (<= 16 byte) literals. See the comment - // on inputMargin in encode.go. - MOVQ SI, AX - SUBQ R10, AX - CMPQ AX, $16 - JLE emitLiteralFastPath - - // ---------------------------------------- - // Begin inline of the emitLiteral call. - // - // d += emitLiteral(dst[d:], src[nextEmit:s]) - - MOVL AX, BX - SUBL $1, BX - - CMPL BX, $60 - JLT inlineEmitLiteralOneByte - CMPL BX, $256 - JLT inlineEmitLiteralTwoBytes - -inlineEmitLiteralThreeBytes: - MOVB $0xf4, 0(DI) - MOVW BX, 1(DI) - ADDQ $3, DI - JMP inlineEmitLiteralMemmove - -inlineEmitLiteralTwoBytes: - MOVB $0xf0, 0(DI) - MOVB BX, 1(DI) - ADDQ $2, DI - JMP inlineEmitLiteralMemmove - -inlineEmitLiteralOneByte: - SHLB $2, BX - MOVB BX, 0(DI) - ADDQ $1, DI - -inlineEmitLiteralMemmove: - // Spill local variables (registers) onto the stack; call; unspill. - // - // copy(dst[i:], lit) - // - // This means calling runtime·memmove(&dst[i], &lit[0], len(lit)), so we push - // DI, R10 and AX as arguments. - MOVQ DI, 0(SP) - MOVQ R10, 8(SP) - MOVQ AX, 16(SP) - ADDQ AX, DI // Finish the "d +=" part of "d += emitLiteral(etc)". - MOVQ SI, 72(SP) - MOVQ DI, 80(SP) - MOVQ R15, 112(SP) - CALL runtime·memmove(SB) - MOVQ 56(SP), CX - MOVQ 64(SP), DX - MOVQ 72(SP), SI - MOVQ 80(SP), DI - MOVQ 88(SP), R9 - MOVQ 112(SP), R15 - JMP inner1 - -inlineEmitLiteralEnd: - // End inline of the emitLiteral call. - // ---------------------------------------- - -emitLiteralFastPath: - // !!! Emit the 1-byte encoding "uint8(len(lit)-1)<<2". - MOVB AX, BX - SUBB $1, BX - SHLB $2, BX - MOVB BX, (DI) - ADDQ $1, DI - - // !!! Implement the copy from lit to dst as a 16-byte load and store. - // (Encode's documentation says that dst and src must not overlap.) - // - // This always copies 16 bytes, instead of only len(lit) bytes, but that's - // OK. Subsequent iterations will fix up the overrun. - // - // Note that on amd64, it is legal and cheap to issue unaligned 8-byte or - // 16-byte loads and stores. This technique probably wouldn't be as - // effective on architectures that are fussier about alignment. - MOVOU 0(R10), X0 - MOVOU X0, 0(DI) - ADDQ AX, DI - -inner1: - // for { etc } - - // base := s - MOVQ SI, R12 - - // !!! offset := base - candidate - MOVQ R12, R11 - SUBQ R15, R11 - SUBQ DX, R11 - - // ---------------------------------------- - // Begin inline of the extendMatch call. - // - // s = extendMatch(src, candidate+4, s+4) - - // !!! R14 = &src[len(src)] - MOVQ src_len+32(FP), R14 - ADDQ DX, R14 - - // !!! R13 = &src[len(src) - 8] - MOVQ R14, R13 - SUBQ $8, R13 - - // !!! R15 = &src[candidate + 4] - ADDQ $4, R15 - ADDQ DX, R15 - - // !!! s += 4 - ADDQ $4, SI - -inlineExtendMatchCmp8: - // As long as we are 8 or more bytes before the end of src, we can load and - // compare 8 bytes at a time. If those 8 bytes are equal, repeat. - CMPQ SI, R13 - JA inlineExtendMatchCmp1 - MOVQ (R15), AX - MOVQ (SI), BX - CMPQ AX, BX - JNE inlineExtendMatchBSF - ADDQ $8, R15 - ADDQ $8, SI - JMP inlineExtendMatchCmp8 - -inlineExtendMatchBSF: - // If those 8 bytes were not equal, XOR the two 8 byte values, and return - // the index of the first byte that differs. The BSF instruction finds the - // least significant 1 bit, the amd64 architecture is little-endian, and - // the shift by 3 converts a bit index to a byte index. - XORQ AX, BX - BSFQ BX, BX - SHRQ $3, BX - ADDQ BX, SI - JMP inlineExtendMatchEnd - -inlineExtendMatchCmp1: - // In src's tail, compare 1 byte at a time. - CMPQ SI, R14 - JAE inlineExtendMatchEnd - MOVB (R15), AX - MOVB (SI), BX - CMPB AX, BX - JNE inlineExtendMatchEnd - ADDQ $1, R15 - ADDQ $1, SI - JMP inlineExtendMatchCmp1 - -inlineExtendMatchEnd: - // End inline of the extendMatch call. - // ---------------------------------------- - - // ---------------------------------------- - // Begin inline of the emitCopy call. - // - // d += emitCopy(dst[d:], base-candidate, s-base) - - // !!! length := s - base - MOVQ SI, AX - SUBQ R12, AX - -inlineEmitCopyLoop0: - // for length >= 68 { etc } - CMPL AX, $68 - JLT inlineEmitCopyStep1 - - // Emit a length 64 copy, encoded as 3 bytes. - MOVB $0xfe, 0(DI) - MOVW R11, 1(DI) - ADDQ $3, DI - SUBL $64, AX - JMP inlineEmitCopyLoop0 - -inlineEmitCopyStep1: - // if length > 64 { etc } - CMPL AX, $64 - JLE inlineEmitCopyStep2 - - // Emit a length 60 copy, encoded as 3 bytes. - MOVB $0xee, 0(DI) - MOVW R11, 1(DI) - ADDQ $3, DI - SUBL $60, AX - -inlineEmitCopyStep2: - // if length >= 12 || offset >= 2048 { goto inlineEmitCopyStep3 } - CMPL AX, $12 - JGE inlineEmitCopyStep3 - CMPL R11, $2048 - JGE inlineEmitCopyStep3 - - // Emit the remaining copy, encoded as 2 bytes. - MOVB R11, 1(DI) - SHRL $8, R11 - SHLB $5, R11 - SUBB $4, AX - SHLB $2, AX - ORB AX, R11 - ORB $1, R11 - MOVB R11, 0(DI) - ADDQ $2, DI - JMP inlineEmitCopyEnd - -inlineEmitCopyStep3: - // Emit the remaining copy, encoded as 3 bytes. - SUBL $1, AX - SHLB $2, AX - ORB $2, AX - MOVB AX, 0(DI) - MOVW R11, 1(DI) - ADDQ $3, DI - -inlineEmitCopyEnd: - // End inline of the emitCopy call. - // ---------------------------------------- - - // nextEmit = s - MOVQ SI, R10 - - // if s >= sLimit { goto emitRemainder } - MOVQ SI, AX - SUBQ DX, AX - CMPQ AX, R9 - JAE emitRemainder - - // As per the encode_other.go code: - // - // We could immediately etc. - - // x := load64(src, s-1) - MOVQ -1(SI), R14 - - // prevHash := hash(uint32(x>>0), shift) - MOVL R14, R11 - IMULL $0x1e35a7bd, R11 - SHRL CX, R11 - - // table[prevHash] = uint16(s-1) - MOVQ SI, AX - SUBQ DX, AX - SUBQ $1, AX - - // XXX: MOVW AX, table-32768(SP)(R11*2) - // XXX: 66 42 89 44 5c 78 mov %ax,0x78(%rsp,%r11,2) - BYTE $0x66 - BYTE $0x42 - BYTE $0x89 - BYTE $0x44 - BYTE $0x5c - BYTE $0x78 - - // currHash := hash(uint32(x>>8), shift) - SHRQ $8, R14 - MOVL R14, R11 - IMULL $0x1e35a7bd, R11 - SHRL CX, R11 - - // candidate = int(table[currHash]) - // XXX: MOVWQZX table-32768(SP)(R11*2), R15 - // XXX: 4e 0f b7 7c 5c 78 movzwq 0x78(%rsp,%r11,2),%r15 - BYTE $0x4e - BYTE $0x0f - BYTE $0xb7 - BYTE $0x7c - BYTE $0x5c - BYTE $0x78 - - // table[currHash] = uint16(s) - ADDQ $1, AX - - // XXX: MOVW AX, table-32768(SP)(R11*2) - // XXX: 66 42 89 44 5c 78 mov %ax,0x78(%rsp,%r11,2) - BYTE $0x66 - BYTE $0x42 - BYTE $0x89 - BYTE $0x44 - BYTE $0x5c - BYTE $0x78 - - // if uint32(x>>8) == load32(src, candidate) { continue } - MOVL (DX)(R15*1), BX - CMPL R14, BX - JEQ inner1 - - // nextHash = hash(uint32(x>>16), shift) - SHRQ $8, R14 - MOVL R14, R11 - IMULL $0x1e35a7bd, R11 - SHRL CX, R11 - - // s++ - ADDQ $1, SI - - // break out of the inner1 for loop, i.e. continue the outer loop. - JMP outer - -emitRemainder: - // if nextEmit < len(src) { etc } - MOVQ src_len+32(FP), AX - ADDQ DX, AX - CMPQ R10, AX - JEQ encodeBlockEnd - - // d += emitLiteral(dst[d:], src[nextEmit:]) - // - // Push args. - MOVQ DI, 0(SP) - MOVQ $0, 8(SP) // Unnecessary, as the callee ignores it, but conservative. - MOVQ $0, 16(SP) // Unnecessary, as the callee ignores it, but conservative. - MOVQ R10, 24(SP) - SUBQ R10, AX - MOVQ AX, 32(SP) - MOVQ AX, 40(SP) // Unnecessary, as the callee ignores it, but conservative. - - // Spill local variables (registers) onto the stack; call; unspill. - MOVQ DI, 80(SP) - CALL ·emitLiteral(SB) - MOVQ 80(SP), DI - - // Finish the "d +=" part of "d += emitLiteral(etc)". - ADDQ 48(SP), DI - -encodeBlockEnd: - MOVQ dst_base+0(FP), AX - SUBQ AX, DI - MOVQ DI, d+48(FP) - RET diff --git a/vendor/github.com/golang/snappy/encode_arm64.s b/vendor/github.com/golang/snappy/encode_arm64.s deleted file mode 100644 index f8d54adfc5..0000000000 --- a/vendor/github.com/golang/snappy/encode_arm64.s +++ /dev/null @@ -1,722 +0,0 @@ -// Copyright 2020 The Go Authors. All rights reserved. -// Use of this source code is governed by a BSD-style -// license that can be found in the LICENSE file. - -// +build !appengine -// +build gc -// +build !noasm - -#include "textflag.h" - -// The asm code generally follows the pure Go code in encode_other.go, except -// where marked with a "!!!". - -// ---------------------------------------------------------------------------- - -// func emitLiteral(dst, lit []byte) int -// -// All local variables fit into registers. The register allocation: -// - R3 len(lit) -// - R4 n -// - R6 return value -// - R8 &dst[i] -// - R10 &lit[0] -// -// The 32 bytes of stack space is to call runtime·memmove. -// -// The unusual register allocation of local variables, such as R10 for the -// source pointer, matches the allocation used at the call site in encodeBlock, -// which makes it easier to manually inline this function. -TEXT ·emitLiteral(SB), NOSPLIT, $32-56 - MOVD dst_base+0(FP), R8 - MOVD lit_base+24(FP), R10 - MOVD lit_len+32(FP), R3 - MOVD R3, R6 - MOVW R3, R4 - SUBW $1, R4, R4 - - CMPW $60, R4 - BLT oneByte - CMPW $256, R4 - BLT twoBytes - -threeBytes: - MOVD $0xf4, R2 - MOVB R2, 0(R8) - MOVW R4, 1(R8) - ADD $3, R8, R8 - ADD $3, R6, R6 - B memmove - -twoBytes: - MOVD $0xf0, R2 - MOVB R2, 0(R8) - MOVB R4, 1(R8) - ADD $2, R8, R8 - ADD $2, R6, R6 - B memmove - -oneByte: - LSLW $2, R4, R4 - MOVB R4, 0(R8) - ADD $1, R8, R8 - ADD $1, R6, R6 - -memmove: - MOVD R6, ret+48(FP) - - // copy(dst[i:], lit) - // - // This means calling runtime·memmove(&dst[i], &lit[0], len(lit)), so we push - // R8, R10 and R3 as arguments. - MOVD R8, 8(RSP) - MOVD R10, 16(RSP) - MOVD R3, 24(RSP) - CALL runtime·memmove(SB) - RET - -// ---------------------------------------------------------------------------- - -// func emitCopy(dst []byte, offset, length int) int -// -// All local variables fit into registers. The register allocation: -// - R3 length -// - R7 &dst[0] -// - R8 &dst[i] -// - R11 offset -// -// The unusual register allocation of local variables, such as R11 for the -// offset, matches the allocation used at the call site in encodeBlock, which -// makes it easier to manually inline this function. -TEXT ·emitCopy(SB), NOSPLIT, $0-48 - MOVD dst_base+0(FP), R8 - MOVD R8, R7 - MOVD offset+24(FP), R11 - MOVD length+32(FP), R3 - -loop0: - // for length >= 68 { etc } - CMPW $68, R3 - BLT step1 - - // Emit a length 64 copy, encoded as 3 bytes. - MOVD $0xfe, R2 - MOVB R2, 0(R8) - MOVW R11, 1(R8) - ADD $3, R8, R8 - SUB $64, R3, R3 - B loop0 - -step1: - // if length > 64 { etc } - CMP $64, R3 - BLE step2 - - // Emit a length 60 copy, encoded as 3 bytes. - MOVD $0xee, R2 - MOVB R2, 0(R8) - MOVW R11, 1(R8) - ADD $3, R8, R8 - SUB $60, R3, R3 - -step2: - // if length >= 12 || offset >= 2048 { goto step3 } - CMP $12, R3 - BGE step3 - CMPW $2048, R11 - BGE step3 - - // Emit the remaining copy, encoded as 2 bytes. - MOVB R11, 1(R8) - LSRW $3, R11, R11 - AND $0xe0, R11, R11 - SUB $4, R3, R3 - LSLW $2, R3 - AND $0xff, R3, R3 - ORRW R3, R11, R11 - ORRW $1, R11, R11 - MOVB R11, 0(R8) - ADD $2, R8, R8 - - // Return the number of bytes written. - SUB R7, R8, R8 - MOVD R8, ret+40(FP) - RET - -step3: - // Emit the remaining copy, encoded as 3 bytes. - SUB $1, R3, R3 - AND $0xff, R3, R3 - LSLW $2, R3, R3 - ORRW $2, R3, R3 - MOVB R3, 0(R8) - MOVW R11, 1(R8) - ADD $3, R8, R8 - - // Return the number of bytes written. - SUB R7, R8, R8 - MOVD R8, ret+40(FP) - RET - -// ---------------------------------------------------------------------------- - -// func extendMatch(src []byte, i, j int) int -// -// All local variables fit into registers. The register allocation: -// - R6 &src[0] -// - R7 &src[j] -// - R13 &src[len(src) - 8] -// - R14 &src[len(src)] -// - R15 &src[i] -// -// The unusual register allocation of local variables, such as R15 for a source -// pointer, matches the allocation used at the call site in encodeBlock, which -// makes it easier to manually inline this function. -TEXT ·extendMatch(SB), NOSPLIT, $0-48 - MOVD src_base+0(FP), R6 - MOVD src_len+8(FP), R14 - MOVD i+24(FP), R15 - MOVD j+32(FP), R7 - ADD R6, R14, R14 - ADD R6, R15, R15 - ADD R6, R7, R7 - MOVD R14, R13 - SUB $8, R13, R13 - -cmp8: - // As long as we are 8 or more bytes before the end of src, we can load and - // compare 8 bytes at a time. If those 8 bytes are equal, repeat. - CMP R13, R7 - BHI cmp1 - MOVD (R15), R3 - MOVD (R7), R4 - CMP R4, R3 - BNE bsf - ADD $8, R15, R15 - ADD $8, R7, R7 - B cmp8 - -bsf: - // If those 8 bytes were not equal, XOR the two 8 byte values, and return - // the index of the first byte that differs. - // RBIT reverses the bit order, then CLZ counts the leading zeros, the - // combination of which finds the least significant bit which is set. - // The arm64 architecture is little-endian, and the shift by 3 converts - // a bit index to a byte index. - EOR R3, R4, R4 - RBIT R4, R4 - CLZ R4, R4 - ADD R4>>3, R7, R7 - - // Convert from &src[ret] to ret. - SUB R6, R7, R7 - MOVD R7, ret+40(FP) - RET - -cmp1: - // In src's tail, compare 1 byte at a time. - CMP R7, R14 - BLS extendMatchEnd - MOVB (R15), R3 - MOVB (R7), R4 - CMP R4, R3 - BNE extendMatchEnd - ADD $1, R15, R15 - ADD $1, R7, R7 - B cmp1 - -extendMatchEnd: - // Convert from &src[ret] to ret. - SUB R6, R7, R7 - MOVD R7, ret+40(FP) - RET - -// ---------------------------------------------------------------------------- - -// func encodeBlock(dst, src []byte) (d int) -// -// All local variables fit into registers, other than "var table". The register -// allocation: -// - R3 . . -// - R4 . . -// - R5 64 shift -// - R6 72 &src[0], tableSize -// - R7 80 &src[s] -// - R8 88 &dst[d] -// - R9 96 sLimit -// - R10 . &src[nextEmit] -// - R11 104 prevHash, currHash, nextHash, offset -// - R12 112 &src[base], skip -// - R13 . &src[nextS], &src[len(src) - 8] -// - R14 . len(src), bytesBetweenHashLookups, &src[len(src)], x -// - R15 120 candidate -// - R16 . hash constant, 0x1e35a7bd -// - R17 . &table -// - . 128 table -// -// The second column (64, 72, etc) is the stack offset to spill the registers -// when calling other functions. We could pack this slightly tighter, but it's -// simpler to have a dedicated spill map independent of the function called. -// -// "var table [maxTableSize]uint16" takes up 32768 bytes of stack space. An -// extra 64 bytes, to call other functions, and an extra 64 bytes, to spill -// local variables (registers) during calls gives 32768 + 64 + 64 = 32896. -TEXT ·encodeBlock(SB), 0, $32896-56 - MOVD dst_base+0(FP), R8 - MOVD src_base+24(FP), R7 - MOVD src_len+32(FP), R14 - - // shift, tableSize := uint32(32-8), 1<<8 - MOVD $24, R5 - MOVD $256, R6 - MOVW $0xa7bd, R16 - MOVKW $(0x1e35<<16), R16 - -calcShift: - // for ; tableSize < maxTableSize && tableSize < len(src); tableSize *= 2 { - // shift-- - // } - MOVD $16384, R2 - CMP R2, R6 - BGE varTable - CMP R14, R6 - BGE varTable - SUB $1, R5, R5 - LSL $1, R6, R6 - B calcShift - -varTable: - // var table [maxTableSize]uint16 - // - // In the asm code, unlike the Go code, we can zero-initialize only the - // first tableSize elements. Each uint16 element is 2 bytes and each - // iterations writes 64 bytes, so we can do only tableSize/32 writes - // instead of the 2048 writes that would zero-initialize all of table's - // 32768 bytes. This clear could overrun the first tableSize elements, but - // it won't overrun the allocated stack size. - ADD $128, RSP, R17 - MOVD R17, R4 - - // !!! R6 = &src[tableSize] - ADD R6<<1, R17, R6 - -memclr: - STP.P (ZR, ZR), 64(R4) - STP (ZR, ZR), -48(R4) - STP (ZR, ZR), -32(R4) - STP (ZR, ZR), -16(R4) - CMP R4, R6 - BHI memclr - - // !!! R6 = &src[0] - MOVD R7, R6 - - // sLimit := len(src) - inputMargin - MOVD R14, R9 - SUB $15, R9, R9 - - // !!! Pre-emptively spill R5, R6 and R9 to the stack. Their values don't - // change for the rest of the function. - MOVD R5, 64(RSP) - MOVD R6, 72(RSP) - MOVD R9, 96(RSP) - - // nextEmit := 0 - MOVD R6, R10 - - // s := 1 - ADD $1, R7, R7 - - // nextHash := hash(load32(src, s), shift) - MOVW 0(R7), R11 - MULW R16, R11, R11 - LSRW R5, R11, R11 - -outer: - // for { etc } - - // skip := 32 - MOVD $32, R12 - - // nextS := s - MOVD R7, R13 - - // candidate := 0 - MOVD $0, R15 - -inner0: - // for { etc } - - // s := nextS - MOVD R13, R7 - - // bytesBetweenHashLookups := skip >> 5 - MOVD R12, R14 - LSR $5, R14, R14 - - // nextS = s + bytesBetweenHashLookups - ADD R14, R13, R13 - - // skip += bytesBetweenHashLookups - ADD R14, R12, R12 - - // if nextS > sLimit { goto emitRemainder } - MOVD R13, R3 - SUB R6, R3, R3 - CMP R9, R3 - BHI emitRemainder - - // candidate = int(table[nextHash]) - MOVHU 0(R17)(R11<<1), R15 - - // table[nextHash] = uint16(s) - MOVD R7, R3 - SUB R6, R3, R3 - - MOVH R3, 0(R17)(R11<<1) - - // nextHash = hash(load32(src, nextS), shift) - MOVW 0(R13), R11 - MULW R16, R11 - LSRW R5, R11, R11 - - // if load32(src, s) != load32(src, candidate) { continue } break - MOVW 0(R7), R3 - MOVW (R6)(R15), R4 - CMPW R4, R3 - BNE inner0 - -fourByteMatch: - // As per the encode_other.go code: - // - // A 4-byte match has been found. We'll later see etc. - - // !!! Jump to a fast path for short (<= 16 byte) literals. See the comment - // on inputMargin in encode.go. - MOVD R7, R3 - SUB R10, R3, R3 - CMP $16, R3 - BLE emitLiteralFastPath - - // ---------------------------------------- - // Begin inline of the emitLiteral call. - // - // d += emitLiteral(dst[d:], src[nextEmit:s]) - - MOVW R3, R4 - SUBW $1, R4, R4 - - MOVW $60, R2 - CMPW R2, R4 - BLT inlineEmitLiteralOneByte - MOVW $256, R2 - CMPW R2, R4 - BLT inlineEmitLiteralTwoBytes - -inlineEmitLiteralThreeBytes: - MOVD $0xf4, R1 - MOVB R1, 0(R8) - MOVW R4, 1(R8) - ADD $3, R8, R8 - B inlineEmitLiteralMemmove - -inlineEmitLiteralTwoBytes: - MOVD $0xf0, R1 - MOVB R1, 0(R8) - MOVB R4, 1(R8) - ADD $2, R8, R8 - B inlineEmitLiteralMemmove - -inlineEmitLiteralOneByte: - LSLW $2, R4, R4 - MOVB R4, 0(R8) - ADD $1, R8, R8 - -inlineEmitLiteralMemmove: - // Spill local variables (registers) onto the stack; call; unspill. - // - // copy(dst[i:], lit) - // - // This means calling runtime·memmove(&dst[i], &lit[0], len(lit)), so we push - // R8, R10 and R3 as arguments. - MOVD R8, 8(RSP) - MOVD R10, 16(RSP) - MOVD R3, 24(RSP) - - // Finish the "d +=" part of "d += emitLiteral(etc)". - ADD R3, R8, R8 - MOVD R7, 80(RSP) - MOVD R8, 88(RSP) - MOVD R15, 120(RSP) - CALL runtime·memmove(SB) - MOVD 64(RSP), R5 - MOVD 72(RSP), R6 - MOVD 80(RSP), R7 - MOVD 88(RSP), R8 - MOVD 96(RSP), R9 - MOVD 120(RSP), R15 - ADD $128, RSP, R17 - MOVW $0xa7bd, R16 - MOVKW $(0x1e35<<16), R16 - B inner1 - -inlineEmitLiteralEnd: - // End inline of the emitLiteral call. - // ---------------------------------------- - -emitLiteralFastPath: - // !!! Emit the 1-byte encoding "uint8(len(lit)-1)<<2". - MOVB R3, R4 - SUBW $1, R4, R4 - AND $0xff, R4, R4 - LSLW $2, R4, R4 - MOVB R4, (R8) - ADD $1, R8, R8 - - // !!! Implement the copy from lit to dst as a 16-byte load and store. - // (Encode's documentation says that dst and src must not overlap.) - // - // This always copies 16 bytes, instead of only len(lit) bytes, but that's - // OK. Subsequent iterations will fix up the overrun. - // - // Note that on arm64, it is legal and cheap to issue unaligned 8-byte or - // 16-byte loads and stores. This technique probably wouldn't be as - // effective on architectures that are fussier about alignment. - LDP 0(R10), (R0, R1) - STP (R0, R1), 0(R8) - ADD R3, R8, R8 - -inner1: - // for { etc } - - // base := s - MOVD R7, R12 - - // !!! offset := base - candidate - MOVD R12, R11 - SUB R15, R11, R11 - SUB R6, R11, R11 - - // ---------------------------------------- - // Begin inline of the extendMatch call. - // - // s = extendMatch(src, candidate+4, s+4) - - // !!! R14 = &src[len(src)] - MOVD src_len+32(FP), R14 - ADD R6, R14, R14 - - // !!! R13 = &src[len(src) - 8] - MOVD R14, R13 - SUB $8, R13, R13 - - // !!! R15 = &src[candidate + 4] - ADD $4, R15, R15 - ADD R6, R15, R15 - - // !!! s += 4 - ADD $4, R7, R7 - -inlineExtendMatchCmp8: - // As long as we are 8 or more bytes before the end of src, we can load and - // compare 8 bytes at a time. If those 8 bytes are equal, repeat. - CMP R13, R7 - BHI inlineExtendMatchCmp1 - MOVD (R15), R3 - MOVD (R7), R4 - CMP R4, R3 - BNE inlineExtendMatchBSF - ADD $8, R15, R15 - ADD $8, R7, R7 - B inlineExtendMatchCmp8 - -inlineExtendMatchBSF: - // If those 8 bytes were not equal, XOR the two 8 byte values, and return - // the index of the first byte that differs. - // RBIT reverses the bit order, then CLZ counts the leading zeros, the - // combination of which finds the least significant bit which is set. - // The arm64 architecture is little-endian, and the shift by 3 converts - // a bit index to a byte index. - EOR R3, R4, R4 - RBIT R4, R4 - CLZ R4, R4 - ADD R4>>3, R7, R7 - B inlineExtendMatchEnd - -inlineExtendMatchCmp1: - // In src's tail, compare 1 byte at a time. - CMP R7, R14 - BLS inlineExtendMatchEnd - MOVB (R15), R3 - MOVB (R7), R4 - CMP R4, R3 - BNE inlineExtendMatchEnd - ADD $1, R15, R15 - ADD $1, R7, R7 - B inlineExtendMatchCmp1 - -inlineExtendMatchEnd: - // End inline of the extendMatch call. - // ---------------------------------------- - - // ---------------------------------------- - // Begin inline of the emitCopy call. - // - // d += emitCopy(dst[d:], base-candidate, s-base) - - // !!! length := s - base - MOVD R7, R3 - SUB R12, R3, R3 - -inlineEmitCopyLoop0: - // for length >= 68 { etc } - MOVW $68, R2 - CMPW R2, R3 - BLT inlineEmitCopyStep1 - - // Emit a length 64 copy, encoded as 3 bytes. - MOVD $0xfe, R1 - MOVB R1, 0(R8) - MOVW R11, 1(R8) - ADD $3, R8, R8 - SUBW $64, R3, R3 - B inlineEmitCopyLoop0 - -inlineEmitCopyStep1: - // if length > 64 { etc } - MOVW $64, R2 - CMPW R2, R3 - BLE inlineEmitCopyStep2 - - // Emit a length 60 copy, encoded as 3 bytes. - MOVD $0xee, R1 - MOVB R1, 0(R8) - MOVW R11, 1(R8) - ADD $3, R8, R8 - SUBW $60, R3, R3 - -inlineEmitCopyStep2: - // if length >= 12 || offset >= 2048 { goto inlineEmitCopyStep3 } - MOVW $12, R2 - CMPW R2, R3 - BGE inlineEmitCopyStep3 - MOVW $2048, R2 - CMPW R2, R11 - BGE inlineEmitCopyStep3 - - // Emit the remaining copy, encoded as 2 bytes. - MOVB R11, 1(R8) - LSRW $8, R11, R11 - LSLW $5, R11, R11 - SUBW $4, R3, R3 - AND $0xff, R3, R3 - LSLW $2, R3, R3 - ORRW R3, R11, R11 - ORRW $1, R11, R11 - MOVB R11, 0(R8) - ADD $2, R8, R8 - B inlineEmitCopyEnd - -inlineEmitCopyStep3: - // Emit the remaining copy, encoded as 3 bytes. - SUBW $1, R3, R3 - LSLW $2, R3, R3 - ORRW $2, R3, R3 - MOVB R3, 0(R8) - MOVW R11, 1(R8) - ADD $3, R8, R8 - -inlineEmitCopyEnd: - // End inline of the emitCopy call. - // ---------------------------------------- - - // nextEmit = s - MOVD R7, R10 - - // if s >= sLimit { goto emitRemainder } - MOVD R7, R3 - SUB R6, R3, R3 - CMP R3, R9 - BLS emitRemainder - - // As per the encode_other.go code: - // - // We could immediately etc. - - // x := load64(src, s-1) - MOVD -1(R7), R14 - - // prevHash := hash(uint32(x>>0), shift) - MOVW R14, R11 - MULW R16, R11, R11 - LSRW R5, R11, R11 - - // table[prevHash] = uint16(s-1) - MOVD R7, R3 - SUB R6, R3, R3 - SUB $1, R3, R3 - - MOVHU R3, 0(R17)(R11<<1) - - // currHash := hash(uint32(x>>8), shift) - LSR $8, R14, R14 - MOVW R14, R11 - MULW R16, R11, R11 - LSRW R5, R11, R11 - - // candidate = int(table[currHash]) - MOVHU 0(R17)(R11<<1), R15 - - // table[currHash] = uint16(s) - ADD $1, R3, R3 - MOVHU R3, 0(R17)(R11<<1) - - // if uint32(x>>8) == load32(src, candidate) { continue } - MOVW (R6)(R15), R4 - CMPW R4, R14 - BEQ inner1 - - // nextHash = hash(uint32(x>>16), shift) - LSR $8, R14, R14 - MOVW R14, R11 - MULW R16, R11, R11 - LSRW R5, R11, R11 - - // s++ - ADD $1, R7, R7 - - // break out of the inner1 for loop, i.e. continue the outer loop. - B outer - -emitRemainder: - // if nextEmit < len(src) { etc } - MOVD src_len+32(FP), R3 - ADD R6, R3, R3 - CMP R3, R10 - BEQ encodeBlockEnd - - // d += emitLiteral(dst[d:], src[nextEmit:]) - // - // Push args. - MOVD R8, 8(RSP) - MOVD $0, 16(RSP) // Unnecessary, as the callee ignores it, but conservative. - MOVD $0, 24(RSP) // Unnecessary, as the callee ignores it, but conservative. - MOVD R10, 32(RSP) - SUB R10, R3, R3 - MOVD R3, 40(RSP) - MOVD R3, 48(RSP) // Unnecessary, as the callee ignores it, but conservative. - - // Spill local variables (registers) onto the stack; call; unspill. - MOVD R8, 88(RSP) - CALL ·emitLiteral(SB) - MOVD 88(RSP), R8 - - // Finish the "d +=" part of "d += emitLiteral(etc)". - MOVD 56(RSP), R1 - ADD R1, R8, R8 - -encodeBlockEnd: - MOVD dst_base+0(FP), R3 - SUB R3, R8, R8 - MOVD R8, d+48(FP) - RET diff --git a/vendor/github.com/golang/snappy/encode_asm.go b/vendor/github.com/golang/snappy/encode_asm.go deleted file mode 100644 index 107c1e7141..0000000000 --- a/vendor/github.com/golang/snappy/encode_asm.go +++ /dev/null @@ -1,30 +0,0 @@ -// Copyright 2016 The Snappy-Go Authors. All rights reserved. -// Use of this source code is governed by a BSD-style -// license that can be found in the LICENSE file. - -// +build !appengine -// +build gc -// +build !noasm -// +build amd64 arm64 - -package snappy - -// emitLiteral has the same semantics as in encode_other.go. -// -//go:noescape -func emitLiteral(dst, lit []byte) int - -// emitCopy has the same semantics as in encode_other.go. -// -//go:noescape -func emitCopy(dst []byte, offset, length int) int - -// extendMatch has the same semantics as in encode_other.go. -// -//go:noescape -func extendMatch(src []byte, i, j int) int - -// encodeBlock has the same semantics as in encode_other.go. -// -//go:noescape -func encodeBlock(dst, src []byte) (d int) diff --git a/vendor/github.com/golang/snappy/go.mod b/vendor/github.com/golang/snappy/go.mod deleted file mode 100644 index f6406bb2c7..0000000000 --- a/vendor/github.com/golang/snappy/go.mod +++ /dev/null @@ -1 +0,0 @@ -module github.com/golang/snappy diff --git a/vendor/github.com/klauspost/compress/.gitattributes b/vendor/github.com/klauspost/compress/.gitattributes new file mode 100644 index 0000000000..402433593c --- /dev/null +++ b/vendor/github.com/klauspost/compress/.gitattributes @@ -0,0 +1,2 @@ +* -text +*.bin -text -diff diff --git a/vendor/github.com/klauspost/compress/.gitignore b/vendor/github.com/klauspost/compress/.gitignore new file mode 100644 index 0000000000..b35f8449bf --- /dev/null +++ b/vendor/github.com/klauspost/compress/.gitignore @@ -0,0 +1,25 @@ +# Compiled Object files, Static and Dynamic libs (Shared Objects) +*.o +*.a +*.so + +# Folders +_obj +_test + +# Architecture specific extensions/prefixes +*.[568vq] +[568vq].out + +*.cgo1.go +*.cgo2.c +_cgo_defun.c +_cgo_gotypes.go +_cgo_export.* + +_testmain.go + +*.exe +*.test +*.prof +/s2/cmd/_s2sx/sfx-exe diff --git a/vendor/github.com/klauspost/compress/.goreleaser.yml b/vendor/github.com/klauspost/compress/.goreleaser.yml new file mode 100644 index 0000000000..c9014ce1da --- /dev/null +++ b/vendor/github.com/klauspost/compress/.goreleaser.yml @@ -0,0 +1,137 @@ +# This is an example goreleaser.yaml file with some sane defaults. +# Make sure to check the documentation at http://goreleaser.com +before: + hooks: + - ./gen.sh + +builds: + - + id: "s2c" + binary: s2c + main: ./s2/cmd/s2c/main.go + flags: + - -trimpath + env: + - CGO_ENABLED=0 + goos: + - aix + - linux + - freebsd + - netbsd + - windows + - darwin + goarch: + - 386 + - amd64 + - arm + - arm64 + - ppc64 + - ppc64le + - mips64 + - mips64le + goarm: + - 7 + - + id: "s2d" + binary: s2d + main: ./s2/cmd/s2d/main.go + flags: + - -trimpath + env: + - CGO_ENABLED=0 + goos: + - aix + - linux + - freebsd + - netbsd + - windows + - darwin + goarch: + - 386 + - amd64 + - arm + - arm64 + - ppc64 + - ppc64le + - mips64 + - mips64le + goarm: + - 7 + - + id: "s2sx" + binary: s2sx + main: ./s2/cmd/_s2sx/main.go + flags: + - -modfile=s2sx.mod + - -trimpath + env: + - CGO_ENABLED=0 + goos: + - aix + - linux + - freebsd + - netbsd + - windows + - darwin + goarch: + - 386 + - amd64 + - arm + - arm64 + - ppc64 + - ppc64le + - mips64 + - mips64le + goarm: + - 7 + +archives: + - + id: s2-binaries + name_template: "s2-{{ .Os }}_{{ .Arch }}_{{ .Version }}" + replacements: + aix: AIX + darwin: OSX + linux: Linux + windows: Windows + 386: i386 + amd64: x86_64 + freebsd: FreeBSD + netbsd: NetBSD + format_overrides: + - goos: windows + format: zip + files: + - unpack/* + - s2/LICENSE + - s2/README.md +checksum: + name_template: 'checksums.txt' +snapshot: + name_template: "{{ .Tag }}-next" +changelog: + sort: asc + filters: + exclude: + - '^doc:' + - '^docs:' + - '^test:' + - '^tests:' + - '^Update\sREADME.md' + +nfpms: + - + file_name_template: "s2_package_{{ .Version }}_{{ .Os }}_{{ .Arch }}" + vendor: Klaus Post + homepage: https://github.com/klauspost/compress + maintainer: Klaus Post + description: S2 Compression Tool + license: BSD 3-Clause + formats: + - deb + - rpm + replacements: + darwin: Darwin + linux: Linux + freebsd: FreeBSD + amd64: x86_64 diff --git a/vendor/github.com/klauspost/compress/LICENSE b/vendor/github.com/klauspost/compress/LICENSE index 1eb75ef68e..6cd1e96276 100644 --- a/vendor/github.com/klauspost/compress/LICENSE +++ b/vendor/github.com/klauspost/compress/LICENSE @@ -26,3 +26,267 @@ DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +------------------ + +Files: gzhttp/* + + Apache License + Version 2.0, January 2004 + http://www.apache.org/licenses/ + + TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION + + 1. Definitions. + + "License" shall mean the terms and conditions for use, reproduction, + and distribution as defined by Sections 1 through 9 of this document. + + "Licensor" shall mean the copyright owner or entity authorized by + the copyright owner that is granting the License. + + "Legal Entity" shall mean the union of the acting entity and all + other entities that control, are controlled by, or are under common + control with that entity. For the purposes of this definition, + "control" means (i) the power, direct or indirect, to cause the + direction or management of such entity, whether by contract or + otherwise, or (ii) ownership of fifty percent (50%) or more of the + outstanding shares, or (iii) beneficial ownership of such entity. + + "You" (or "Your") shall mean an individual or Legal Entity + exercising permissions granted by this License. + + "Source" form shall mean the preferred form for making modifications, + including but not limited to software source code, documentation + source, and configuration files. + + "Object" form shall mean any form resulting from mechanical + transformation or translation of a Source form, including but + not limited to compiled object code, generated documentation, + and conversions to other media types. + + "Work" shall mean the work of authorship, whether in Source or + Object form, made available under the License, as indicated by a + copyright notice that is included in or attached to the work + (an example is provided in the Appendix below). + + "Derivative Works" shall mean any work, whether in Source or Object + form, that is based on (or derived from) the Work and for which the + editorial revisions, annotations, elaborations, or other modifications + represent, as a whole, an original work of authorship. For the purposes + of this License, Derivative Works shall not include works that remain + separable from, or merely link (or bind by name) to the interfaces of, + the Work and Derivative Works thereof. + + "Contribution" shall mean any work of authorship, including + the original version of the Work and any modifications or additions + to that Work or Derivative Works thereof, that is intentionally + submitted to Licensor for inclusion in the Work by the copyright owner + or by an individual or Legal Entity authorized to submit on behalf of + the copyright owner. For the purposes of this definition, "submitted" + means any form of electronic, verbal, or written communication sent + to the Licensor or its representatives, including but not limited to + communication on electronic mailing lists, source code control systems, + and issue tracking systems that are managed by, or on behalf of, the + Licensor for the purpose of discussing and improving the Work, but + excluding communication that is conspicuously marked or otherwise + designated in writing by the copyright owner as "Not a Contribution." + + "Contributor" shall mean Licensor and any individual or Legal Entity + on behalf of whom a Contribution has been received by Licensor and + subsequently incorporated within the Work. + + 2. Grant of Copyright License. Subject to the terms and conditions of + this License, each Contributor hereby grants to You a perpetual, + worldwide, non-exclusive, no-charge, royalty-free, irrevocable + copyright license to reproduce, prepare Derivative Works of, + publicly display, publicly perform, sublicense, and distribute the + Work and such Derivative Works in Source or Object form. + + 3. Grant of Patent License. Subject to the terms and conditions of + this License, each Contributor hereby grants to You a perpetual, + worldwide, non-exclusive, no-charge, royalty-free, irrevocable + (except as stated in this section) patent license to make, have made, + use, offer to sell, sell, import, and otherwise transfer the Work, + where such license applies only to those patent claims licensable + by such Contributor that are necessarily infringed by their + Contribution(s) alone or by combination of their Contribution(s) + with the Work to which such Contribution(s) was submitted. If You + institute patent litigation against any entity (including a + cross-claim or counterclaim in a lawsuit) alleging that the Work + or a Contribution incorporated within the Work constitutes direct + or contributory patent infringement, then any patent licenses + granted to You under this License for that Work shall terminate + as of the date such litigation is filed. + + 4. Redistribution. You may reproduce and distribute copies of the + Work or Derivative Works thereof in any medium, with or without + modifications, and in Source or Object form, provided that You + meet the following conditions: + + (a) You must give any other recipients of the Work or + Derivative Works a copy of this License; and + + (b) You must cause any modified files to carry prominent notices + stating that You changed the files; and + + (c) You must retain, in the Source form of any Derivative Works + that You distribute, all copyright, patent, trademark, and + attribution notices from the Source form of the Work, + excluding those notices that do not pertain to any part of + the Derivative Works; and + + (d) If the Work includes a "NOTICE" text file as part of its + distribution, then any Derivative Works that You distribute must + include a readable copy of the attribution notices contained + within such NOTICE file, excluding those notices that do not + pertain to any part of the Derivative Works, in at least one + of the following places: within a NOTICE text file distributed + as part of the Derivative Works; within the Source form or + documentation, if provided along with the Derivative Works; or, + within a display generated by the Derivative Works, if and + wherever such third-party notices normally appear. The contents + of the NOTICE file are for informational purposes only and + do not modify the License. You may add Your own attribution + notices within Derivative Works that You distribute, alongside + or as an addendum to the NOTICE text from the Work, provided + that such additional attribution notices cannot be construed + as modifying the License. + + You may add Your own copyright statement to Your modifications and + may provide additional or different license terms and conditions + for use, reproduction, or distribution of Your modifications, or + for any such Derivative Works as a whole, provided Your use, + reproduction, and distribution of the Work otherwise complies with + the conditions stated in this License. + + 5. Submission of Contributions. Unless You explicitly state otherwise, + any Contribution intentionally submitted for inclusion in the Work + by You to the Licensor shall be under the terms and conditions of + this License, without any additional terms or conditions. + Notwithstanding the above, nothing herein shall supersede or modify + the terms of any separate license agreement you may have executed + with Licensor regarding such Contributions. + + 6. Trademarks. This License does not grant permission to use the trade + names, trademarks, service marks, or product names of the Licensor, + except as required for reasonable and customary use in describing the + origin of the Work and reproducing the content of the NOTICE file. + + 7. Disclaimer of Warranty. Unless required by applicable law or + agreed to in writing, Licensor provides the Work (and each + Contributor provides its Contributions) on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or + implied, including, without limitation, any warranties or conditions + of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A + PARTICULAR PURPOSE. You are solely responsible for determining the + appropriateness of using or redistributing the Work and assume any + risks associated with Your exercise of permissions under this License. + + 8. Limitation of Liability. In no event and under no legal theory, + whether in tort (including negligence), contract, or otherwise, + unless required by applicable law (such as deliberate and grossly + negligent acts) or agreed to in writing, shall any Contributor be + liable to You for damages, including any direct, indirect, special, + incidental, or consequential damages of any character arising as a + result of this License or out of the use or inability to use the + Work (including but not limited to damages for loss of goodwill, + work stoppage, computer failure or malfunction, or any and all + other commercial damages or losses), even if such Contributor + has been advised of the possibility of such damages. + + 9. Accepting Warranty or Additional Liability. While redistributing + the Work or Derivative Works thereof, You may choose to offer, + and charge a fee for, acceptance of support, warranty, indemnity, + or other liability obligations and/or rights consistent with this + License. However, in accepting such obligations, You may act only + on Your own behalf and on Your sole responsibility, not on behalf + of any other Contributor, and only if You agree to indemnify, + defend, and hold each Contributor harmless for any liability + incurred by, or claims asserted against, such Contributor by reason + of your accepting any such warranty or additional liability. + + END OF TERMS AND CONDITIONS + + APPENDIX: How to apply the Apache License to your work. + + To apply the Apache License to your work, attach the following + boilerplate notice, with the fields enclosed by brackets "[]" + replaced with your own identifying information. (Don't include + the brackets!) The text should be enclosed in the appropriate + comment syntax for the file format. We also recommend that a + file or class name and description of purpose be included on the + same "printed page" as the copyright notice for easier + identification within third-party archives. + + Copyright 2016-2017 The New York Times Company + + Licensed under the Apache License, Version 2.0 (the "License"); + you may not use this file except in compliance with the License. + You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. + +------------------ + +Files: s2/cmd/internal/readahead/* + +The MIT License (MIT) + +Copyright (c) 2015 Klaus Post + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to deal +in the Software without restriction, including without limitation the rights +to use, copy, modify, merge, publish, distribute, sublicense, and/or sell +copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in all +copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, +OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE +SOFTWARE. + +--------------------- +Files: snappy/* +Files: internal/snapref/* + +Copyright (c) 2011 The Snappy-Go Authors. All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are +met: + + * Redistributions of source code must retain the above copyright +notice, this list of conditions and the following disclaimer. + * Redistributions in binary form must reproduce the above +copyright notice, this list of conditions and the following disclaimer +in the documentation and/or other materials provided with the +distribution. + * Neither the name of Google Inc. nor the names of its +contributors may be used to endorse or promote products derived from +this software without specific prior written permission. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS +"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT +LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR +A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT +OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, +SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT +LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, +DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY +THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT +(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. diff --git a/vendor/github.com/klauspost/compress/README.md b/vendor/github.com/klauspost/compress/README.md new file mode 100644 index 0000000000..235dc7cc68 --- /dev/null +++ b/vendor/github.com/klauspost/compress/README.md @@ -0,0 +1,430 @@ +# compress + +This package provides various compression algorithms. + +* [zstandard](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression and decompression in pure Go. +* [S2](https://github.com/klauspost/compress/tree/master/s2#s2-compression) is a high performance replacement for Snappy. +* Optimized [deflate](https://godoc.org/github.com/klauspost/compress/flate) packages which can be used as a dropin replacement for [gzip](https://godoc.org/github.com/klauspost/compress/gzip), [zip](https://godoc.org/github.com/klauspost/compress/zip) and [zlib](https://godoc.org/github.com/klauspost/compress/zlib). +* [snappy](https://github.com/klauspost/compress/tree/master/snappy) is a drop-in replacement for `github.com/golang/snappy` offering better compression and concurrent streams. +* [huff0](https://github.com/klauspost/compress/tree/master/huff0) and [FSE](https://github.com/klauspost/compress/tree/master/fse) implementations for raw entropy encoding. +* [gzhttp](https://github.com/klauspost/compress/tree/master/gzhttp) Provides client and server wrappers for handling gzipped requests efficiently. +* [pgzip](https://github.com/klauspost/pgzip) is a separate package that provides a very fast parallel gzip implementation. +* [fuzz package](https://github.com/klauspost/compress-fuzz) for fuzz testing all compressors/decompressors here. + +[![Go Reference](https://pkg.go.dev/badge/klauspost/compress.svg)](https://pkg.go.dev/github.com/klauspost/compress?tab=subdirectories) +[![Go](https://github.com/klauspost/compress/actions/workflows/go.yml/badge.svg)](https://github.com/klauspost/compress/actions/workflows/go.yml) +[![Sourcegraph Badge](https://sourcegraph.com/github.com/klauspost/compress/-/badge.svg)](https://sourcegraph.com/github.com/klauspost/compress?badge) + +# changelog + +* Aug 12, 2021 (v1.13.4) + * Add [snappy replacement package](https://github.com/klauspost/compress/tree/master/snappy). + * zstd: Fix incorrect encoding in "best" mode [#415](https://github.com/klauspost/compress/pull/415) + +* Aug 3, 2021 (v1.13.3) + + * zstd: Improve Best compression [#404](https://github.com/klauspost/compress/pull/404) + * zstd: Fix WriteTo error forwarding [#411](https://github.com/klauspost/compress/pull/411) + * gzhttp: Return http.HandlerFunc instead of http.Handler. Unlikely breaking change. [#406](https://github.com/klauspost/compress/pull/406) + * s2sx: Fix max size error [#399](https://github.com/klauspost/compress/pull/399) + * zstd: Add optional stream content size on reset [#401](https://github.com/klauspost/compress/pull/401) + * zstd: use SpeedBestCompression for level >= 10 [#410](https://github.com/klauspost/compress/pull/410) + +* Jun 14, 2021 (v1.13.1) + + * s2: Add full Snappy output support [#396](https://github.com/klauspost/compress/pull/396) + * zstd: Add configurable [Decoder window](https://pkg.go.dev/github.com/klauspost/compress/zstd#WithDecoderMaxWindow) size [#394](https://github.com/klauspost/compress/pull/394) + * gzhttp: Add header to skip compression [#389](https://github.com/klauspost/compress/pull/389) + * s2: Improve speed with bigger output margin [#395](https://github.com/klauspost/compress/pull/395) + +* Jun 3, 2021 (v1.13.0) + * Added [gzhttp](https://github.com/klauspost/compress/tree/master/gzhttp#gzip-handler) which allows wrapping HTTP servers and clients with GZIP compressors. + * zstd: Detect short invalid signatures [#382](https://github.com/klauspost/compress/pull/382) + * zstd: Spawn decoder goroutine only if needed. [#380](https://github.com/klauspost/compress/pull/380) + +* May 25, 2021 (v1.12.3) + * deflate: Better/faster Huffman encoding [#374](https://github.com/klauspost/compress/pull/374) + * deflate: Allocate less for history. [#375](https://github.com/klauspost/compress/pull/375) + * zstd: Forward read errors [#373](https://github.com/klauspost/compress/pull/373) + +* Apr 27, 2021 (v1.12.2) + * zstd: Improve better/best compression [#360](https://github.com/klauspost/compress/pull/360) [#364](https://github.com/klauspost/compress/pull/364) [#365](https://github.com/klauspost/compress/pull/365) + * zstd: Add helpers to compress/decompress zstd inside zip files [#363](https://github.com/klauspost/compress/pull/363) + * deflate: Improve level 5+6 compression [#367](https://github.com/klauspost/compress/pull/367) + * s2: Improve better/best compression [#358](https://github.com/klauspost/compress/pull/358) [#359](https://github.com/klauspost/compress/pull/358) + * s2: Load after checking src limit on amd64. [#362](https://github.com/klauspost/compress/pull/362) + * s2sx: Limit max executable size [#368](https://github.com/klauspost/compress/pull/368) + +* Apr 14, 2021 (v1.12.1) + * snappy package removed. Upstream added as dependency. + * s2: Better compression in "best" mode [#353](https://github.com/klauspost/compress/pull/353) + * s2sx: Add stdin input and detect pre-compressed from signature [#352](https://github.com/klauspost/compress/pull/352) + * s2c/s2d: Add http as possible input [#348](https://github.com/klauspost/compress/pull/348) + * s2c/s2d/s2sx: Always truncate when writing files [#352](https://github.com/klauspost/compress/pull/352) + * zstd: Reduce memory usage further when using [WithLowerEncoderMem](https://pkg.go.dev/github.com/klauspost/compress/zstd#WithLowerEncoderMem) [#346](https://github.com/klauspost/compress/pull/346) + * s2: Fix potential problem with amd64 assembly and profilers [#349](https://github.com/klauspost/compress/pull/349) + +* Mar 26, 2021 (v1.11.13) + * zstd: Big speedup on small dictionary encodes [#344](https://github.com/klauspost/compress/pull/344) [#345](https://github.com/klauspost/compress/pull/345) + * zstd: Add [WithLowerEncoderMem](https://pkg.go.dev/github.com/klauspost/compress/zstd#WithLowerEncoderMem) encoder option [#336](https://github.com/klauspost/compress/pull/336) + * deflate: Improve entropy compression [#338](https://github.com/klauspost/compress/pull/338) + * s2: Clean up and minor performance improvement in best [#341](https://github.com/klauspost/compress/pull/341) + +* Mar 5, 2021 (v1.11.12) + * s2: Add `s2sx` binary that creates [self extracting archives](https://github.com/klauspost/compress/tree/master/s2#s2sx-self-extracting-archives). + * s2: Speed up decompression on non-assembly platforms [#328](https://github.com/klauspost/compress/pull/328) + +* Mar 1, 2021 (v1.11.9) + * s2: Add ARM64 decompression assembly. Around 2x output speed. [#324](https://github.com/klauspost/compress/pull/324) + * s2: Improve "better" speed and efficiency. [#325](https://github.com/klauspost/compress/pull/325) + * s2: Fix binaries. + +* Feb 25, 2021 (v1.11.8) + * s2: Fixed occational out-of-bounds write on amd64. Upgrade recommended. + * s2: Add AMD64 assembly for better mode. 25-50% faster. [#315](https://github.com/klauspost/compress/pull/315) + * s2: Less upfront decoder allocation. [#322](https://github.com/klauspost/compress/pull/322) + * zstd: Faster "compression" of incompressible data. [#314](https://github.com/klauspost/compress/pull/314) + * zip: Fix zip64 headers. [#313](https://github.com/klauspost/compress/pull/313) + +* Jan 14, 2021 (v1.11.7) + * Use Bytes() interface to get bytes across packages. [#309](https://github.com/klauspost/compress/pull/309) + * s2: Add 'best' compression option. [#310](https://github.com/klauspost/compress/pull/310) + * s2: Add ReaderMaxBlockSize, changes `s2.NewReader` signature to include varargs. [#311](https://github.com/klauspost/compress/pull/311) + * s2: Fix crash on small better buffers. [#308](https://github.com/klauspost/compress/pull/308) + * s2: Clean up decoder. [#312](https://github.com/klauspost/compress/pull/312) + +* Jan 7, 2021 (v1.11.6) + * zstd: Make decoder allocations smaller [#306](https://github.com/klauspost/compress/pull/306) + * zstd: Free Decoder resources when Reset is called with a nil io.Reader [#305](https://github.com/klauspost/compress/pull/305) + +* Dec 20, 2020 (v1.11.4) + * zstd: Add Best compression mode [#304](https://github.com/klauspost/compress/pull/304) + * Add header decoder [#299](https://github.com/klauspost/compress/pull/299) + * s2: Add uncompressed stream option [#297](https://github.com/klauspost/compress/pull/297) + * Simplify/speed up small blocks with known max size. [#300](https://github.com/klauspost/compress/pull/300) + * zstd: Always reset literal dict encoder [#303](https://github.com/klauspost/compress/pull/303) + +* Nov 15, 2020 (v1.11.3) + * inflate: 10-15% faster decompression [#293](https://github.com/klauspost/compress/pull/293) + * zstd: Tweak DecodeAll default allocation [#295](https://github.com/klauspost/compress/pull/295) + +* Oct 11, 2020 (v1.11.2) + * s2: Fix out of bounds read in "better" block compression [#291](https://github.com/klauspost/compress/pull/291) + +* Oct 1, 2020 (v1.11.1) + * zstd: Set allLitEntropy true in default configuration [#286](https://github.com/klauspost/compress/pull/286) + +* Sept 8, 2020 (v1.11.0) + * zstd: Add experimental compression [dictionaries](https://github.com/klauspost/compress/tree/master/zstd#dictionaries) [#281](https://github.com/klauspost/compress/pull/281) + * zstd: Fix mixed Write and ReadFrom calls [#282](https://github.com/klauspost/compress/pull/282) + * inflate/gz: Limit variable shifts, ~5% faster decompression [#274](https://github.com/klauspost/compress/pull/274) + +
+ See changes prior to v1.11.0 + +* July 8, 2020 (v1.10.11) + * zstd: Fix extra block when compressing with ReadFrom. [#278](https://github.com/klauspost/compress/pull/278) + * huff0: Also populate compression table when reading decoding table. [#275](https://github.com/klauspost/compress/pull/275) + +* June 23, 2020 (v1.10.10) + * zstd: Skip entropy compression in fastest mode when no matches. [#270](https://github.com/klauspost/compress/pull/270) + +* June 16, 2020 (v1.10.9): + * zstd: API change for specifying dictionaries. See [#268](https://github.com/klauspost/compress/pull/268) + * zip: update CreateHeaderRaw to handle zip64 fields. [#266](https://github.com/klauspost/compress/pull/266) + * Fuzzit tests removed. The service has been purchased and is no longer available. + +* June 5, 2020 (v1.10.8): + * 1.15x faster zstd block decompression. [#265](https://github.com/klauspost/compress/pull/265) + +* June 1, 2020 (v1.10.7): + * Added zstd decompression [dictionary support](https://github.com/klauspost/compress/tree/master/zstd#dictionaries) + * Increase zstd decompression speed up to 1.19x. [#259](https://github.com/klauspost/compress/pull/259) + * Remove internal reset call in zstd compression and reduce allocations. [#263](https://github.com/klauspost/compress/pull/263) + +* May 21, 2020: (v1.10.6) + * zstd: Reduce allocations while decoding. [#258](https://github.com/klauspost/compress/pull/258), [#252](https://github.com/klauspost/compress/pull/252) + * zstd: Stricter decompression checks. + +* April 12, 2020: (v1.10.5) + * s2-commands: Flush output when receiving SIGINT. [#239](https://github.com/klauspost/compress/pull/239) + +* Apr 8, 2020: (v1.10.4) + * zstd: Minor/special case optimizations. [#251](https://github.com/klauspost/compress/pull/251), [#250](https://github.com/klauspost/compress/pull/250), [#249](https://github.com/klauspost/compress/pull/249), [#247](https://github.com/klauspost/compress/pull/247) +* Mar 11, 2020: (v1.10.3) + * s2: Use S2 encoder in pure Go mode for Snappy output as well. [#245](https://github.com/klauspost/compress/pull/245) + * s2: Fix pure Go block encoder. [#244](https://github.com/klauspost/compress/pull/244) + * zstd: Added "better compression" mode. [#240](https://github.com/klauspost/compress/pull/240) + * zstd: Improve speed of fastest compression mode by 5-10% [#241](https://github.com/klauspost/compress/pull/241) + * zstd: Skip creating encoders when not needed. [#238](https://github.com/klauspost/compress/pull/238) + +* Feb 27, 2020: (v1.10.2) + * Close to 50% speedup in inflate (gzip/zip decompression). [#236](https://github.com/klauspost/compress/pull/236) [#234](https://github.com/klauspost/compress/pull/234) [#232](https://github.com/klauspost/compress/pull/232) + * Reduce deflate level 1-6 memory usage up to 59%. [#227](https://github.com/klauspost/compress/pull/227) + +* Feb 18, 2020: (v1.10.1) + * Fix zstd crash when resetting multiple times without sending data. [#226](https://github.com/klauspost/compress/pull/226) + * deflate: Fix dictionary use on level 1-6. [#224](https://github.com/klauspost/compress/pull/224) + * Remove deflate writer reference when closing. [#224](https://github.com/klauspost/compress/pull/224) + +* Feb 4, 2020: (v1.10.0) + * Add optional dictionary to [stateless deflate](https://pkg.go.dev/github.com/klauspost/compress/flate?tab=doc#StatelessDeflate). Breaking change, send `nil` for previous behaviour. [#216](https://github.com/klauspost/compress/pull/216) + * Fix buffer overflow on repeated small block deflate. [#218](https://github.com/klauspost/compress/pull/218) + * Allow copying content from an existing ZIP file without decompressing+compressing. [#214](https://github.com/klauspost/compress/pull/214) + * Added [S2](https://github.com/klauspost/compress/tree/master/s2#s2-compression) AMD64 assembler and various optimizations. Stream speed >10GB/s. [#186](https://github.com/klauspost/compress/pull/186) + +
+ +
+ See changes prior to v1.10.0 + +* Jan 20,2020 (v1.9.8) Optimize gzip/deflate with better size estimates and faster table generation. [#207](https://github.com/klauspost/compress/pull/207) by [luyu6056](https://github.com/luyu6056), [#206](https://github.com/klauspost/compress/pull/206). +* Jan 11, 2020: S2 Encode/Decode will use provided buffer if capacity is big enough. [#204](https://github.com/klauspost/compress/pull/204) +* Jan 5, 2020: (v1.9.7) Fix another zstd regression in v1.9.5 - v1.9.6 removed. +* Jan 4, 2020: (v1.9.6) Regression in v1.9.5 fixed causing corrupt zstd encodes in rare cases. +* Jan 4, 2020: Faster IO in [s2c + s2d commandline tools](https://github.com/klauspost/compress/tree/master/s2#commandline-tools) compression/decompression. [#192](https://github.com/klauspost/compress/pull/192) +* Dec 29, 2019: Removed v1.9.5 since fuzz tests showed a compatibility problem with the reference zstandard decoder. +* Dec 29, 2019: (v1.9.5) zstd: 10-20% faster block compression. [#199](https://github.com/klauspost/compress/pull/199) +* Dec 29, 2019: [zip](https://godoc.org/github.com/klauspost/compress/zip) package updated with latest Go features +* Dec 29, 2019: zstd: Single segment flag condintions tweaked. [#197](https://github.com/klauspost/compress/pull/197) +* Dec 18, 2019: s2: Faster compression when ReadFrom is used. [#198](https://github.com/klauspost/compress/pull/198) +* Dec 10, 2019: s2: Fix repeat length output when just above at 16MB limit. +* Dec 10, 2019: zstd: Add function to get decoder as io.ReadCloser. [#191](https://github.com/klauspost/compress/pull/191) +* Dec 3, 2019: (v1.9.4) S2: limit max repeat length. [#188](https://github.com/klauspost/compress/pull/188) +* Dec 3, 2019: Add [WithNoEntropyCompression](https://godoc.org/github.com/klauspost/compress/zstd#WithNoEntropyCompression) to zstd [#187](https://github.com/klauspost/compress/pull/187) +* Dec 3, 2019: Reduce memory use for tests. Check for leaked goroutines. +* Nov 28, 2019 (v1.9.3) Less allocations in stateless deflate. +* Nov 28, 2019: 5-20% Faster huff0 decode. Impacts zstd as well. [#184](https://github.com/klauspost/compress/pull/184) +* Nov 12, 2019 (v1.9.2) Added [Stateless Compression](#stateless-compression) for gzip/deflate. +* Nov 12, 2019: Fixed zstd decompression of large single blocks. [#180](https://github.com/klauspost/compress/pull/180) +* Nov 11, 2019: Set default [s2c](https://github.com/klauspost/compress/tree/master/s2#commandline-tools) block size to 4MB. +* Nov 11, 2019: Reduce inflate memory use by 1KB. +* Nov 10, 2019: Less allocations in deflate bit writer. +* Nov 10, 2019: Fix inconsistent error returned by zstd decoder. +* Oct 28, 2019 (v1.9.1) ztsd: Fix crash when compressing blocks. [#174](https://github.com/klauspost/compress/pull/174) +* Oct 24, 2019 (v1.9.0) zstd: Fix rare data corruption [#173](https://github.com/klauspost/compress/pull/173) +* Oct 24, 2019 zstd: Fix huff0 out of buffer write [#171](https://github.com/klauspost/compress/pull/171) and always return errors [#172](https://github.com/klauspost/compress/pull/172) +* Oct 10, 2019: Big deflate rewrite, 30-40% faster with better compression [#105](https://github.com/klauspost/compress/pull/105) + +
+ +
+ See changes prior to v1.9.0 + +* Oct 10, 2019: (v1.8.6) zstd: Allow partial reads to get flushed data. [#169](https://github.com/klauspost/compress/pull/169) +* Oct 3, 2019: Fix inconsistent results on broken zstd streams. +* Sep 25, 2019: Added `-rm` (remove source files) and `-q` (no output except errors) to `s2c` and `s2d` [commands](https://github.com/klauspost/compress/tree/master/s2#commandline-tools) +* Sep 16, 2019: (v1.8.4) Add `s2c` and `s2d` [commandline tools](https://github.com/klauspost/compress/tree/master/s2#commandline-tools). +* Sep 10, 2019: (v1.8.3) Fix s2 decoder [Skip](https://godoc.org/github.com/klauspost/compress/s2#Reader.Skip). +* Sep 7, 2019: zstd: Added [WithWindowSize](https://godoc.org/github.com/klauspost/compress/zstd#WithWindowSize), contributed by [ianwilkes](https://github.com/ianwilkes). +* Sep 5, 2019: (v1.8.2) Add [WithZeroFrames](https://godoc.org/github.com/klauspost/compress/zstd#WithZeroFrames) which adds full zero payload block encoding option. +* Sep 5, 2019: Lazy initialization of zstandard predefined en/decoder tables. +* Aug 26, 2019: (v1.8.1) S2: 1-2% compression increase in "better" compression mode. +* Aug 26, 2019: zstd: Check maximum size of Huffman 1X compressed literals while decoding. +* Aug 24, 2019: (v1.8.0) Added [S2 compression](https://github.com/klauspost/compress/tree/master/s2#s2-compression), a high performance replacement for Snappy. +* Aug 21, 2019: (v1.7.6) Fixed minor issues found by fuzzer. One could lead to zstd not decompressing. +* Aug 18, 2019: Add [fuzzit](https://fuzzit.dev/) continuous fuzzing. +* Aug 14, 2019: zstd: Skip incompressible data 2x faster. [#147](https://github.com/klauspost/compress/pull/147) +* Aug 4, 2019 (v1.7.5): Better literal compression. [#146](https://github.com/klauspost/compress/pull/146) +* Aug 4, 2019: Faster zstd compression. [#143](https://github.com/klauspost/compress/pull/143) [#144](https://github.com/klauspost/compress/pull/144) +* Aug 4, 2019: Faster zstd decompression. [#145](https://github.com/klauspost/compress/pull/145) [#143](https://github.com/klauspost/compress/pull/143) [#142](https://github.com/klauspost/compress/pull/142) +* July 15, 2019 (v1.7.4): Fix double EOF block in rare cases on zstd encoder. +* July 15, 2019 (v1.7.3): Minor speedup/compression increase in default zstd encoder. +* July 14, 2019: zstd decoder: Fix decompression error on multiple uses with mixed content. +* July 7, 2019 (v1.7.2): Snappy update, zstd decoder potential race fix. +* June 17, 2019: zstd decompression bugfix. +* June 17, 2019: fix 32 bit builds. +* June 17, 2019: Easier use in modules (less dependencies). +* June 9, 2019: New stronger "default" [zstd](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression mode. Matches zstd default compression ratio. +* June 5, 2019: 20-40% throughput in [zstandard](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression and better compression. +* June 5, 2019: deflate/gzip compression: Reduce memory usage of lower compression levels. +* June 2, 2019: Added [zstandard](https://github.com/klauspost/compress/tree/master/zstd#zstd) compression! +* May 25, 2019: deflate/gzip: 10% faster bit writer, mostly visible in lower levels. +* Apr 22, 2019: [zstd](https://github.com/klauspost/compress/tree/master/zstd#zstd) decompression added. +* Aug 1, 2018: Added [huff0 README](https://github.com/klauspost/compress/tree/master/huff0#huff0-entropy-compression). +* Jul 8, 2018: Added [Performance Update 2018](#performance-update-2018) below. +* Jun 23, 2018: Merged [Go 1.11 inflate optimizations](https://go-review.googlesource.com/c/go/+/102235). Go 1.9 is now required. Backwards compatible version tagged with [v1.3.0](https://github.com/klauspost/compress/releases/tag/v1.3.0). +* Apr 2, 2018: Added [huff0](https://godoc.org/github.com/klauspost/compress/huff0) en/decoder. Experimental for now, API may change. +* Mar 4, 2018: Added [FSE Entropy](https://godoc.org/github.com/klauspost/compress/fse) en/decoder. Experimental for now, API may change. +* Nov 3, 2017: Add compression [Estimate](https://godoc.org/github.com/klauspost/compress#Estimate) function. +* May 28, 2017: Reduce allocations when resetting decoder. +* Apr 02, 2017: Change back to official crc32, since changes were merged in Go 1.7. +* Jan 14, 2017: Reduce stack pressure due to array copies. See [Issue #18625](https://github.com/golang/go/issues/18625). +* Oct 25, 2016: Level 2-4 have been rewritten and now offers significantly better performance than before. +* Oct 20, 2016: Port zlib changes from Go 1.7 to fix zlib writer issue. Please update. +* Oct 16, 2016: Go 1.7 changes merged. Apples to apples this package is a few percent faster, but has a significantly better balance between speed and compression per level. +* Mar 24, 2016: Always attempt Huffman encoding on level 4-7. This improves base 64 encoded data compression. +* Mar 24, 2016: Small speedup for level 1-3. +* Feb 19, 2016: Faster bit writer, level -2 is 15% faster, level 1 is 4% faster. +* Feb 19, 2016: Handle small payloads faster in level 1-3. +* Feb 19, 2016: Added faster level 2 + 3 compression modes. +* Feb 19, 2016: [Rebalanced compression levels](https://blog.klauspost.com/rebalancing-deflate-compression-levels/), so there is a more even progresssion in terms of compression. New default level is 5. +* Feb 14, 2016: Snappy: Merge upstream changes. +* Feb 14, 2016: Snappy: Fix aggressive skipping. +* Feb 14, 2016: Snappy: Update benchmark. +* Feb 13, 2016: Deflate: Fixed assembler problem that could lead to sub-optimal compression. +* Feb 12, 2016: Snappy: Added AMD64 SSE 4.2 optimizations to matching, which makes easy to compress material run faster. Typical speedup is around 25%. +* Feb 9, 2016: Added Snappy package fork. This version is 5-7% faster, much more on hard to compress content. +* Jan 30, 2016: Optimize level 1 to 3 by not considering static dictionary or storing uncompressed. ~4-5% speedup. +* Jan 16, 2016: Optimization on deflate level 1,2,3 compression. +* Jan 8 2016: Merge [CL 18317](https://go-review.googlesource.com/#/c/18317): fix reading, writing of zip64 archives. +* Dec 8 2015: Make level 1 and -2 deterministic even if write size differs. +* Dec 8 2015: Split encoding functions, so hashing and matching can potentially be inlined. 1-3% faster on AMD64. 5% faster on other platforms. +* Dec 8 2015: Fixed rare [one byte out-of bounds read](https://github.com/klauspost/compress/issues/20). Please update! +* Nov 23 2015: Optimization on token writer. ~2-4% faster. Contributed by [@dsnet](https://github.com/dsnet). +* Nov 20 2015: Small optimization to bit writer on 64 bit systems. +* Nov 17 2015: Fixed out-of-bound errors if the underlying Writer returned an error. See [#15](https://github.com/klauspost/compress/issues/15). +* Nov 12 2015: Added [io.WriterTo](https://golang.org/pkg/io/#WriterTo) support to gzip/inflate. +* Nov 11 2015: Merged [CL 16669](https://go-review.googlesource.com/#/c/16669/4): archive/zip: enable overriding (de)compressors per file +* Oct 15 2015: Added skipping on uncompressible data. Random data speed up >5x. + +
+ +# deflate usage + +* [High Throughput Benchmark](http://blog.klauspost.com/go-gzipdeflate-benchmarks/). +* [Small Payload/Webserver Benchmarks](http://blog.klauspost.com/gzip-performance-for-go-webservers/). +* [Linear Time Compression](http://blog.klauspost.com/constant-time-gzipzip-compression/). +* [Re-balancing Deflate Compression Levels](https://blog.klauspost.com/rebalancing-deflate-compression-levels/) + +The packages are drop-in replacements for standard libraries. Simply replace the import path to use them: + +| old import | new import | Documentation +|--------------------|-----------------------------------------|--------------------| +| `compress/gzip` | `github.com/klauspost/compress/gzip` | [gzip](https://pkg.go.dev/github.com/klauspost/compress/gzip?tab=doc) +| `compress/zlib` | `github.com/klauspost/compress/zlib` | [zlib](https://pkg.go.dev/github.com/klauspost/compress/zlib?tab=doc) +| `archive/zip` | `github.com/klauspost/compress/zip` | [zip](https://pkg.go.dev/github.com/klauspost/compress/zip?tab=doc) +| `compress/flate` | `github.com/klauspost/compress/flate` | [flate](https://pkg.go.dev/github.com/klauspost/compress/flate?tab=doc) + +* Optimized [deflate](https://godoc.org/github.com/klauspost/compress/flate) packages which can be used as a dropin replacement for [gzip](https://godoc.org/github.com/klauspost/compress/gzip), [zip](https://godoc.org/github.com/klauspost/compress/zip) and [zlib](https://godoc.org/github.com/klauspost/compress/zlib). + +You may also be interested in [pgzip](https://github.com/klauspost/pgzip), which is a drop in replacement for gzip, which support multithreaded compression on big files and the optimized [crc32](https://github.com/klauspost/crc32) package used by these packages. + +The packages contains the same as the standard library, so you can use the godoc for that: [gzip](http://golang.org/pkg/compress/gzip/), [zip](http://golang.org/pkg/archive/zip/), [zlib](http://golang.org/pkg/compress/zlib/), [flate](http://golang.org/pkg/compress/flate/). + +Currently there is only minor speedup on decompression (mostly CRC32 calculation). + +Memory usage is typically 1MB for a Writer. stdlib is in the same range. +If you expect to have a lot of concurrently allocated Writers consider using +the stateless compress described below. + +# Stateless compression + +This package offers stateless compression as a special option for gzip/deflate. +It will do compression but without maintaining any state between Write calls. + +This means there will be no memory kept between Write calls, but compression and speed will be suboptimal. + +This is only relevant in cases where you expect to run many thousands of compressors concurrently, +but with very little activity. This is *not* intended for regular web servers serving individual requests. + +Because of this, the size of actual Write calls will affect output size. + +In gzip, specify level `-3` / `gzip.StatelessCompression` to enable. + +For direct deflate use, NewStatelessWriter and StatelessDeflate are available. See [documentation](https://godoc.org/github.com/klauspost/compress/flate#NewStatelessWriter) + +A `bufio.Writer` can of course be used to control write sizes. For example, to use a 4KB buffer: + +``` + // replace 'ioutil.Discard' with your output. + gzw, err := gzip.NewWriterLevel(ioutil.Discard, gzip.StatelessCompression) + if err != nil { + return err + } + defer gzw.Close() + + w := bufio.NewWriterSize(gzw, 4096) + defer w.Flush() + + // Write to 'w' +``` + +This will only use up to 4KB in memory when the writer is idle. + +Compression is almost always worse than the fastest compression level +and each write will allocate (a little) memory. + +# Performance Update 2018 + +It has been a while since we have been looking at the speed of this package compared to the standard library, so I thought I would re-do my tests and give some overall recommendations based on the current state. All benchmarks have been performed with Go 1.10 on my Desktop Intel(R) Core(TM) i7-2600 CPU @3.40GHz. Since I last ran the tests, I have gotten more RAM, which means tests with big files are no longer limited by my SSD. + +The raw results are in my [updated spreadsheet](https://docs.google.com/spreadsheets/d/1nuNE2nPfuINCZJRMt6wFWhKpToF95I47XjSsc-1rbPQ/edit?usp=sharing). Due to cgo changes and upstream updates i could not get the cgo version of gzip to compile. Instead I included the [zstd](https://github.com/datadog/zstd) cgo implementation. If I get cgo gzip to work again, I might replace the results in the sheet. + +The columns to take note of are: *MB/s* - the throughput. *Reduction* - the data size reduction in percent of the original. *Rel Speed* relative speed compared to the standard library at the same level. *Smaller* - how many percent smaller is the compressed output compared to stdlib. Negative means the output was bigger. *Loss* means the loss (or gain) in compression as a percentage difference of the input. + +The `gzstd` (standard library gzip) and `gzkp` (this package gzip) only uses one CPU core. [`pgzip`](https://github.com/klauspost/pgzip), [`bgzf`](https://github.com/biogo/hts/tree/master/bgzf) uses all 4 cores. [`zstd`](https://github.com/DataDog/zstd) uses one core, and is a beast (but not Go, yet). + + +## Overall differences. + +There appears to be a roughly 5-10% speed advantage over the standard library when comparing at similar compression levels. + +The biggest difference you will see is the result of [re-balancing](https://blog.klauspost.com/rebalancing-deflate-compression-levels/) the compression levels. I wanted by library to give a smoother transition between the compression levels than the standard library. + +This package attempts to provide a more smooth transition, where "1" is taking a lot of shortcuts, "5" is the reasonable trade-off and "9" is the "give me the best compression", and the values in between gives something reasonable in between. The standard library has big differences in levels 1-4, but levels 5-9 having no significant gains - often spending a lot more time than can be justified by the achieved compression. + +There are links to all the test data in the [spreadsheet](https://docs.google.com/spreadsheets/d/1nuNE2nPfuINCZJRMt6wFWhKpToF95I47XjSsc-1rbPQ/edit?usp=sharing) in the top left field on each tab. + +## Web Content + +This test set aims to emulate typical use in a web server. The test-set is 4GB data in 53k files, and is a mixture of (mostly) HTML, JS, CSS. + +Since level 1 and 9 are close to being the same code, they are quite close. But looking at the levels in-between the differences are quite big. + +Looking at level 6, this package is 88% faster, but will output about 6% more data. For a web server, this means you can serve 88% more data, but have to pay for 6% more bandwidth. You can draw your own conclusions on what would be the most expensive for your case. + +## Object files + +This test is for typical data files stored on a server. In this case it is a collection of Go precompiled objects. They are very compressible. + +The picture is similar to the web content, but with small differences since this is very compressible. Levels 2-3 offer good speed, but is sacrificing quite a bit of compression. + +The standard library seems suboptimal on level 3 and 4 - offering both worse compression and speed than level 6 & 7 of this package respectively. + +## Highly Compressible File + +This is a JSON file with very high redundancy. The reduction starts at 95% on level 1, so in real life terms we are dealing with something like a highly redundant stream of data, etc. + +It is definitely visible that we are dealing with specialized content here, so the results are very scattered. This package does not do very well at levels 1-4, but picks up significantly at level 5 and levels 7 and 8 offering great speed for the achieved compression. + +So if you know you content is extremely compressible you might want to go slightly higher than the defaults. The standard library has a huge gap between levels 3 and 4 in terms of speed (2.75x slowdown), so it offers little "middle ground". + +## Medium-High Compressible + +This is a pretty common test corpus: [enwik9](http://mattmahoney.net/dc/textdata.html). It contains the first 10^9 bytes of the English Wikipedia dump on Mar. 3, 2006. This is a very good test of typical text based compression and more data heavy streams. + +We see a similar picture here as in "Web Content". On equal levels some compression is sacrificed for more speed. Level 5 seems to be the best trade-off between speed and size, beating stdlib level 3 in both. + +## Medium Compressible + +I will combine two test sets, one [10GB file set](http://mattmahoney.net/dc/10gb.html) and a VM disk image (~8GB). Both contain different data types and represent a typical backup scenario. + +The most notable thing is how quickly the standard library drops to very low compression speeds around level 5-6 without any big gains in compression. Since this type of data is fairly common, this does not seem like good behavior. + + +## Un-compressible Content + +This is mainly a test of how good the algorithms are at detecting un-compressible input. The standard library only offers this feature with very conservative settings at level 1. Obviously there is no reason for the algorithms to try to compress input that cannot be compressed. The only downside is that it might skip some compressible data on false detections. + + +## Huffman only compression + +This compression library adds a special compression level, named `HuffmanOnly`, which allows near linear time compression. This is done by completely disabling matching of previous data, and only reduce the number of bits to represent each character. + +This means that often used characters, like 'e' and ' ' (space) in text use the fewest bits to represent, and rare characters like '¤' takes more bits to represent. For more information see [wikipedia](https://en.wikipedia.org/wiki/Huffman_coding) or this nice [video](https://youtu.be/ZdooBTdW5bM). + +Since this type of compression has much less variance, the compression speed is mostly unaffected by the input data, and is usually more than *180MB/s* for a single core. + +The downside is that the compression ratio is usually considerably worse than even the fastest conventional compression. The compression ratio can never be better than 8:1 (12.5%). + +The linear time compression can be used as a "better than nothing" mode, where you cannot risk the encoder to slow down on some content. For comparison, the size of the "Twain" text is *233460 bytes* (+29% vs. level 1) and encode speed is 144MB/s (4.5x level 1). So in this case you trade a 30% size increase for a 4 times speedup. + +For more information see my blog post on [Fast Linear Time Compression](http://blog.klauspost.com/constant-time-gzipzip-compression/). + +This is implemented on Go 1.7 as "Huffman Only" mode, though not exposed for gzip. + + +# license + +This code is licensed under the same conditions as the original Go code. See LICENSE file. diff --git a/vendor/github.com/klauspost/compress/compressible.go b/vendor/github.com/klauspost/compress/compressible.go new file mode 100644 index 0000000000..ea5a692d51 --- /dev/null +++ b/vendor/github.com/klauspost/compress/compressible.go @@ -0,0 +1,85 @@ +package compress + +import "math" + +// Estimate returns a normalized compressibility estimate of block b. +// Values close to zero are likely uncompressible. +// Values above 0.1 are likely to be compressible. +// Values above 0.5 are very compressible. +// Very small lengths will return 0. +func Estimate(b []byte) float64 { + if len(b) < 16 { + return 0 + } + + // Correctly predicted order 1 + hits := 0 + lastMatch := false + var o1 [256]byte + var hist [256]int + c1 := byte(0) + for _, c := range b { + if c == o1[c1] { + // We only count a hit if there was two correct predictions in a row. + if lastMatch { + hits++ + } + lastMatch = true + } else { + lastMatch = false + } + o1[c1] = c + c1 = c + hist[c]++ + } + + // Use x^0.6 to give better spread + prediction := math.Pow(float64(hits)/float64(len(b)), 0.6) + + // Calculate histogram distribution + variance := float64(0) + avg := float64(len(b)) / 256 + + for _, v := range hist { + Δ := float64(v) - avg + variance += Δ * Δ + } + + stddev := math.Sqrt(float64(variance)) / float64(len(b)) + exp := math.Sqrt(1 / float64(len(b))) + + // Subtract expected stddev + stddev -= exp + if stddev < 0 { + stddev = 0 + } + stddev *= 1 + exp + + // Use x^0.4 to give better spread + entropy := math.Pow(stddev, 0.4) + + // 50/50 weight between prediction and histogram distribution + return math.Pow((prediction+entropy)/2, 0.9) +} + +// ShannonEntropyBits returns the number of bits minimum required to represent +// an entropy encoding of the input bytes. +// https://en.wiktionary.org/wiki/Shannon_entropy +func ShannonEntropyBits(b []byte) int { + if len(b) == 0 { + return 0 + } + var hist [256]int + for _, c := range b { + hist[c]++ + } + shannon := float64(0) + invTotal := 1.0 / float64(len(b)) + for _, v := range hist[:] { + if v > 0 { + n := float64(v) + shannon += math.Ceil(-math.Log2(n*invTotal) * n) + } + } + return int(math.Ceil(shannon)) +} diff --git a/vendor/github.com/klauspost/compress/gen.sh b/vendor/github.com/klauspost/compress/gen.sh new file mode 100644 index 0000000000..aff942205f --- /dev/null +++ b/vendor/github.com/klauspost/compress/gen.sh @@ -0,0 +1,4 @@ +#!/bin/sh + +cd s2/cmd/_s2sx/ || exit 1 +go generate . diff --git a/vendor/github.com/klauspost/compress/go.mod b/vendor/github.com/klauspost/compress/go.mod new file mode 100644 index 0000000000..5aa64a436a --- /dev/null +++ b/vendor/github.com/klauspost/compress/go.mod @@ -0,0 +1,3 @@ +module github.com/klauspost/compress + +go 1.15 diff --git a/vendor/github.com/klauspost/compress/go.sum b/vendor/github.com/klauspost/compress/go.sum new file mode 100644 index 0000000000..e69de29bb2 diff --git a/vendor/github.com/klauspost/compress/huff0/compress.go b/vendor/github.com/klauspost/compress/huff0/compress.go index 0823c928ce..8323dc0538 100644 --- a/vendor/github.com/klauspost/compress/huff0/compress.go +++ b/vendor/github.com/klauspost/compress/huff0/compress.go @@ -161,6 +161,70 @@ func compress(in []byte, s *Scratch, compressor func(src []byte) ([]byte, error) return s.Out, false, nil } +// EstimateSizes will estimate the data sizes +func EstimateSizes(in []byte, s *Scratch) (tableSz, dataSz, reuseSz int, err error) { + s, err = s.prepare(in) + if err != nil { + return 0, 0, 0, err + } + + // Create histogram, if none was provided. + tableSz, dataSz, reuseSz = -1, -1, -1 + maxCount := s.maxCount + var canReuse = false + if maxCount == 0 { + maxCount, canReuse = s.countSimple(in) + } else { + canReuse = s.canUseTable(s.prevTable) + } + + // We want the output size to be less than this: + wantSize := len(in) + if s.WantLogLess > 0 { + wantSize -= wantSize >> s.WantLogLess + } + + // Reset for next run. + s.clearCount = true + s.maxCount = 0 + if maxCount >= len(in) { + if maxCount > len(in) { + return 0, 0, 0, fmt.Errorf("maxCount (%d) > length (%d)", maxCount, len(in)) + } + if len(in) == 1 { + return 0, 0, 0, ErrIncompressible + } + // One symbol, use RLE + return 0, 0, 0, ErrUseRLE + } + if maxCount == 1 || maxCount < (len(in)>>7) { + // Each symbol present maximum once or too well distributed. + return 0, 0, 0, ErrIncompressible + } + + // Calculate new table. + err = s.buildCTable() + if err != nil { + return 0, 0, 0, err + } + + if false && !s.canUseTable(s.cTable) { + panic("invalid table generated") + } + + tableSz, err = s.cTable.estTableSize(s) + if err != nil { + return 0, 0, 0, err + } + if canReuse { + reuseSz = s.prevTable.estimateSize(s.count[:s.symbolLen]) + } + dataSz = s.cTable.estimateSize(s.count[:s.symbolLen]) + + // Restore + return tableSz, dataSz, reuseSz, nil +} + func (s *Scratch) compress1X(src []byte) ([]byte, error) { return s.compress1xDo(s.Out, src) } diff --git a/vendor/github.com/klauspost/compress/huff0/huff0.go b/vendor/github.com/klauspost/compress/huff0/huff0.go index 7ec2022b65..3ee00ecb47 100644 --- a/vendor/github.com/klauspost/compress/huff0/huff0.go +++ b/vendor/github.com/klauspost/compress/huff0/huff0.go @@ -245,6 +245,68 @@ func (c cTable) write(s *Scratch) error { return nil } +func (c cTable) estTableSize(s *Scratch) (sz int, err error) { + var ( + // precomputed conversion table + bitsToWeight [tableLogMax + 1]byte + huffLog = s.actualTableLog + // last weight is not saved. + maxSymbolValue = uint8(s.symbolLen - 1) + huffWeight = s.huffWeight[:256] + ) + const ( + maxFSETableLog = 6 + ) + // convert to weight + bitsToWeight[0] = 0 + for n := uint8(1); n < huffLog+1; n++ { + bitsToWeight[n] = huffLog + 1 - n + } + + // Acquire histogram for FSE. + hist := s.fse.Histogram() + hist = hist[:256] + for i := range hist[:16] { + hist[i] = 0 + } + for n := uint8(0); n < maxSymbolValue; n++ { + v := bitsToWeight[c[n].nBits] & 15 + huffWeight[n] = v + hist[v]++ + } + + // FSE compress if feasible. + if maxSymbolValue >= 2 { + huffMaxCnt := uint32(0) + huffMax := uint8(0) + for i, v := range hist[:16] { + if v == 0 { + continue + } + huffMax = byte(i) + if v > huffMaxCnt { + huffMaxCnt = v + } + } + s.fse.HistogramFinished(huffMax, int(huffMaxCnt)) + s.fse.TableLog = maxFSETableLog + b, err := fse.Compress(huffWeight[:maxSymbolValue], s.fse) + if err == nil && len(b) < int(s.symbolLen>>1) { + sz += 1 + len(b) + return sz, nil + } + // Unable to compress (RLE/uncompressible) + } + // write raw values as 4-bits (max : 15) + if maxSymbolValue > (256 - 128) { + // should not happen : likely means source cannot be compressed + return 0, ErrIncompressible + } + // special case, pack weights 4 bits/weight. + sz += 1 + int(maxSymbolValue/2) + return sz, nil +} + // estimateSize returns the estimated size in bytes of the input represented in the // histogram supplied. func (c cTable) estimateSize(hist []uint32) int { diff --git a/vendor/github.com/golang/snappy/LICENSE b/vendor/github.com/klauspost/compress/internal/snapref/LICENSE similarity index 100% rename from vendor/github.com/golang/snappy/LICENSE rename to vendor/github.com/klauspost/compress/internal/snapref/LICENSE diff --git a/vendor/github.com/golang/snappy/decode.go b/vendor/github.com/klauspost/compress/internal/snapref/decode.go similarity index 99% rename from vendor/github.com/golang/snappy/decode.go rename to vendor/github.com/klauspost/compress/internal/snapref/decode.go index 23c6e26c6b..40796a49d6 100644 --- a/vendor/github.com/golang/snappy/decode.go +++ b/vendor/github.com/klauspost/compress/internal/snapref/decode.go @@ -2,7 +2,7 @@ // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. -package snappy +package snapref import ( "encoding/binary" diff --git a/vendor/github.com/golang/snappy/decode_other.go b/vendor/github.com/klauspost/compress/internal/snapref/decode_other.go similarity index 98% rename from vendor/github.com/golang/snappy/decode_other.go rename to vendor/github.com/klauspost/compress/internal/snapref/decode_other.go index 2f672be557..77395a6b8b 100644 --- a/vendor/github.com/golang/snappy/decode_other.go +++ b/vendor/github.com/klauspost/compress/internal/snapref/decode_other.go @@ -2,9 +2,7 @@ // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. -// +build !amd64,!arm64 appengine !gc noasm - -package snappy +package snapref // decode writes the decoding of src to dst. It assumes that the varint-encoded // length of the decompressed bytes has already been read, and that len(dst) diff --git a/vendor/github.com/golang/snappy/encode.go b/vendor/github.com/klauspost/compress/internal/snapref/encode.go similarity index 99% rename from vendor/github.com/golang/snappy/encode.go rename to vendor/github.com/klauspost/compress/internal/snapref/encode.go index 7f23657076..13c6040a5d 100644 --- a/vendor/github.com/golang/snappy/encode.go +++ b/vendor/github.com/klauspost/compress/internal/snapref/encode.go @@ -2,7 +2,7 @@ // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. -package snappy +package snapref import ( "encoding/binary" diff --git a/vendor/github.com/golang/snappy/encode_other.go b/vendor/github.com/klauspost/compress/internal/snapref/encode_other.go similarity index 99% rename from vendor/github.com/golang/snappy/encode_other.go rename to vendor/github.com/klauspost/compress/internal/snapref/encode_other.go index 296d7f0beb..511bba65db 100644 --- a/vendor/github.com/golang/snappy/encode_other.go +++ b/vendor/github.com/klauspost/compress/internal/snapref/encode_other.go @@ -2,9 +2,7 @@ // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. -// +build !amd64,!arm64 appengine !gc noasm - -package snappy +package snapref func load32(b []byte, i int) uint32 { b = b[i : i+4 : len(b)] // Help the compiler eliminate bounds checks on the next line. diff --git a/vendor/github.com/golang/snappy/snappy.go b/vendor/github.com/klauspost/compress/internal/snapref/snappy.go similarity index 96% rename from vendor/github.com/golang/snappy/snappy.go rename to vendor/github.com/klauspost/compress/internal/snapref/snappy.go index ece692ea46..34d01f4aa6 100644 --- a/vendor/github.com/golang/snappy/snappy.go +++ b/vendor/github.com/klauspost/compress/internal/snapref/snappy.go @@ -2,7 +2,7 @@ // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. -// Package snappy implements the Snappy compression format. It aims for very +// Package snapref implements the Snappy compression format. It aims for very // high speeds and reasonable compression. // // There are actually two Snappy formats: block and stream. They are related, @@ -17,7 +17,7 @@ // // The canonical, C++ implementation is at https://github.com/google/snappy and // it only implements the block format. -package snappy // import "github.com/golang/snappy" +package snapref import ( "hash/crc32" diff --git a/vendor/github.com/klauspost/compress/s2sx.mod b/vendor/github.com/klauspost/compress/s2sx.mod new file mode 100644 index 0000000000..2263853fca --- /dev/null +++ b/vendor/github.com/klauspost/compress/s2sx.mod @@ -0,0 +1,4 @@ +module github.com/klauspost/compress + +go 1.16 + diff --git a/vendor/github.com/klauspost/compress/s2sx.sum b/vendor/github.com/klauspost/compress/s2sx.sum new file mode 100644 index 0000000000..e69de29bb2 diff --git a/vendor/github.com/klauspost/compress/zstd/README.md b/vendor/github.com/klauspost/compress/zstd/README.md index 787813fa9e..c8f0f16fc1 100644 --- a/vendor/github.com/klauspost/compress/zstd/README.md +++ b/vendor/github.com/klauspost/compress/zstd/README.md @@ -152,7 +152,7 @@ file out level insize outsize millis mb/s silesia.tar zskp 1 211947520 73101992 643 313.87 silesia.tar zskp 2 211947520 67504318 969 208.38 silesia.tar zskp 3 211947520 64595893 2007 100.68 -silesia.tar zskp 4 211947520 60995370 7691 26.28 +silesia.tar zskp 4 211947520 60995370 8825 22.90 cgo zstd: silesia.tar zstd 1 211947520 73605392 543 371.56 @@ -162,7 +162,7 @@ silesia.tar zstd 9 211947520 60212393 5063 39.92 gzip, stdlib/this package: silesia.tar gzstd 1 211947520 80007735 1654 122.21 -silesia.tar gzkp 1 211947520 80369488 1168 173.06 +silesia.tar gzkp 1 211947520 80136201 1152 175.45 GOB stream of binary data. Highly compressible. https://files.klauspost.com/compress/gob-stream.7z @@ -171,13 +171,15 @@ file out level insize outsize millis mb/s gob-stream zskp 1 1911399616 235022249 3088 590.30 gob-stream zskp 2 1911399616 205669791 3786 481.34 gob-stream zskp 3 1911399616 175034659 9636 189.17 -gob-stream zskp 4 1911399616 167273881 29337 62.13 +gob-stream zskp 4 1911399616 165609838 50369 36.19 + gob-stream zstd 1 1911399616 249810424 2637 691.26 gob-stream zstd 3 1911399616 208192146 3490 522.31 gob-stream zstd 6 1911399616 193632038 6687 272.56 gob-stream zstd 9 1911399616 177620386 16175 112.70 + gob-stream gzstd 1 1911399616 357382641 10251 177.82 -gob-stream gzkp 1 1911399616 362156523 5695 320.08 +gob-stream gzkp 1 1911399616 359753026 5438 335.20 The test data for the Large Text Compression Benchmark is the first 10^9 bytes of the English Wikipedia dump on Mar. 3, 2006. @@ -187,11 +189,13 @@ file out level insize outsize millis mb/s enwik9 zskp 1 1000000000 343848582 3609 264.18 enwik9 zskp 2 1000000000 317276632 5746 165.97 enwik9 zskp 3 1000000000 292243069 12162 78.41 -enwik9 zskp 4 1000000000 275241169 36430 26.18 +enwik9 zskp 4 1000000000 262183768 82837 11.51 + enwik9 zstd 1 1000000000 358072021 3110 306.65 enwik9 zstd 3 1000000000 313734672 4784 199.35 enwik9 zstd 6 1000000000 295138875 10290 92.68 enwik9 zstd 9 1000000000 278348700 28549 33.40 + enwik9 gzstd 1 1000000000 382578136 9604 99.30 enwik9 gzkp 1 1000000000 383825945 6544 145.73 @@ -202,13 +206,15 @@ file out level insize outsize millis mb/s github-june-2days-2019.json zskp 1 6273951764 699045015 10620 563.40 github-june-2days-2019.json zskp 2 6273951764 617881763 11687 511.96 github-june-2days-2019.json zskp 3 6273951764 524340691 34043 175.75 -github-june-2days-2019.json zskp 4 6273951764 503314661 93811 63.78 +github-june-2days-2019.json zskp 4 6273951764 470320075 170190 35.16 + github-june-2days-2019.json zstd 1 6273951764 766284037 8450 708.00 github-june-2days-2019.json zstd 3 6273951764 661889476 10927 547.57 github-june-2days-2019.json zstd 6 6273951764 642756859 22996 260.18 github-june-2days-2019.json zstd 9 6273951764 601974523 52413 114.16 + github-june-2days-2019.json gzstd 1 6273951764 1164400847 29948 199.79 -github-june-2days-2019.json gzkp 1 6273951764 1128755542 19236 311.03 +github-june-2days-2019.json gzkp 1 6273951764 1125417694 21788 274.61 VM Image, Linux mint with a few installed applications: https://files.klauspost.com/compress/rawstudio-mint14.7z @@ -217,13 +223,15 @@ file out level insize outsize millis mb/s rawstudio-mint14.tar zskp 1 8558382592 3667489370 20210 403.84 rawstudio-mint14.tar zskp 2 8558382592 3364592300 31873 256.07 rawstudio-mint14.tar zskp 3 8558382592 3158085214 77675 105.08 -rawstudio-mint14.tar zskp 4 8558382592 3020370044 404956 20.16 +rawstudio-mint14.tar zskp 4 8558382592 2965110639 857750 9.52 + rawstudio-mint14.tar zstd 1 8558382592 3609250104 17136 476.27 rawstudio-mint14.tar zstd 3 8558382592 3341679997 29262 278.92 rawstudio-mint14.tar zstd 6 8558382592 3235846406 77904 104.77 rawstudio-mint14.tar zstd 9 8558382592 3160778861 140946 57.91 + rawstudio-mint14.tar gzstd 1 8558382592 3926257486 57722 141.40 -rawstudio-mint14.tar gzkp 1 8558382592 3970463184 41749 195.49 +rawstudio-mint14.tar gzkp 1 8558382592 3962605659 45113 180.92 CSV data: https://files.klauspost.com/compress/nyc-taxi-data-10M.csv.zst @@ -232,13 +240,15 @@ file out level insize outsize millis mb/s nyc-taxi-data-10M.csv zskp 1 3325605752 641339945 8925 355.35 nyc-taxi-data-10M.csv zskp 2 3325605752 591748091 11268 281.44 nyc-taxi-data-10M.csv zskp 3 3325605752 530289687 25239 125.66 -nyc-taxi-data-10M.csv zskp 4 3325605752 490907191 65939 48.10 +nyc-taxi-data-10M.csv zskp 4 3325605752 476268884 135958 23.33 + nyc-taxi-data-10M.csv zstd 1 3325605752 687399637 8233 385.18 nyc-taxi-data-10M.csv zstd 3 3325605752 598514411 10065 315.07 nyc-taxi-data-10M.csv zstd 6 3325605752 570522953 20038 158.27 nyc-taxi-data-10M.csv zstd 9 3325605752 517554797 64565 49.12 + nyc-taxi-data-10M.csv gzstd 1 3325605752 928656485 23876 132.83 -nyc-taxi-data-10M.csv gzkp 1 3325605752 924718719 16388 193.53 +nyc-taxi-data-10M.csv gzkp 1 3325605752 922257165 16780 189.00 ``` ## Decompressor diff --git a/vendor/github.com/klauspost/compress/zstd/decoder.go b/vendor/github.com/klauspost/compress/zstd/decoder.go index 4d984c3b26..f430f58b57 100644 --- a/vendor/github.com/klauspost/compress/zstd/decoder.go +++ b/vendor/github.com/klauspost/compress/zstd/decoder.go @@ -260,9 +260,10 @@ func (d *Decoder) WriteTo(w io.Writer) (int64, error) { if len(d.current.b) > 0 { n2, err2 := w.Write(d.current.b) n += int64(n2) - if err2 != nil && d.current.err == nil { + if err2 != nil && (d.current.err == nil || d.current.err == io.EOF) { d.current.err = err2 - break + } else if n2 != len(d.current.b) { + d.current.err = io.ErrShortWrite } } if d.current.err != nil { diff --git a/vendor/github.com/klauspost/compress/zstd/enc_base.go b/vendor/github.com/klauspost/compress/zstd/enc_base.go index 60f2986486..295cd602a4 100644 --- a/vendor/github.com/klauspost/compress/zstd/enc_base.go +++ b/vendor/github.com/klauspost/compress/zstd/enc_base.go @@ -38,8 +38,8 @@ func (e *fastBase) AppendCRC(dst []byte) []byte { // WindowSize returns the window size of the encoder, // or a window size small enough to contain the input size, if > 0. -func (e *fastBase) WindowSize(size int) int32 { - if size > 0 && size < int(e.maxMatchOff) { +func (e *fastBase) WindowSize(size int64) int32 { + if size > 0 && size < int64(e.maxMatchOff) { b := int32(1) << uint(bits.Len(uint(size))) // Keep minimum window. if b < 1024 { diff --git a/vendor/github.com/klauspost/compress/zstd/enc_best.go b/vendor/github.com/klauspost/compress/zstd/enc_best.go index b7d4b90047..96028ecd83 100644 --- a/vendor/github.com/klauspost/compress/zstd/enc_best.go +++ b/vendor/github.com/klauspost/compress/zstd/enc_best.go @@ -5,22 +5,61 @@ package zstd import ( + "bytes" "fmt" - "math/bits" + + "github.com/klauspost/compress" ) const ( - bestLongTableBits = 20 // Bits used in the long match table + bestLongTableBits = 22 // Bits used in the long match table bestLongTableSize = 1 << bestLongTableBits // Size of the table + bestLongLen = 8 // Bytes used for table hash // Note: Increasing the short table bits or making the hash shorter // can actually lead to compression degradation since it will 'steal' more from the // long match table and match offsets are quite big. // This greatly depends on the type of input. - bestShortTableBits = 16 // Bits used in the short match table + bestShortTableBits = 18 // Bits used in the short match table bestShortTableSize = 1 << bestShortTableBits // Size of the table + bestShortLen = 4 // Bytes used for table hash + ) +type match struct { + offset int32 + s int32 + length int32 + rep int32 + est int32 +} + +const highScore = 25000 + +// estBits will estimate output bits from predefined tables. +func (m *match) estBits(bitsPerByte int32) { + mlc := mlCode(uint32(m.length - zstdMinMatch)) + var ofc uint8 + if m.rep < 0 { + ofc = ofCode(uint32(m.s-m.offset) + 3) + } else { + ofc = ofCode(uint32(m.rep)) + } + // Cost, excluding + ofTT, mlTT := fsePredefEnc[tableOffsets].ct.symbolTT[ofc], fsePredefEnc[tableMatchLengths].ct.symbolTT[mlc] + + // Add cost of match encoding... + m.est = int32(ofTT.outBits + mlTT.outBits) + m.est += int32(ofTT.deltaNbBits>>16 + mlTT.deltaNbBits>>16) + // Subtract savings compared to literal encoding... + m.est -= (m.length * bitsPerByte) >> 10 + if m.est > 0 { + // Unlikely gain.. + m.length = 0 + m.est = highScore + } +} + // bestFastEncoder uses 2 tables, one for short matches (5 bytes) and one for long matches. // The long match table contains the previous entry with the same hash, // effectively making it a "chain" of length 2. @@ -109,6 +148,14 @@ func (e *bestFastEncoder) Encode(blk *blockEnc, src []byte) { return } + // Use this to estimate literal cost. + // Scaled by 10 bits. + bitsPerByte := int32((compress.ShannonEntropyBits(src) * 1024) / len(src)) + // Huffman can never go < 1 bit/byte + if bitsPerByte < 1024 { + bitsPerByte = 1024 + } + // Override src src = e.hist sLimit := int32(len(src)) - inputMargin @@ -145,51 +192,49 @@ encodeLoop: panic("offset0 was 0") } - type match struct { - offset int32 - s int32 - length int32 - rep int32 - } - matchAt := func(offset int32, s int32, first uint32, rep int32) match { - if s-offset >= e.maxMatchOff || load3232(src, offset) != first { - return match{offset: offset, s: s} - } - return match{offset: offset, s: s, length: 4 + e.matchlen(s+4, offset+4, src), rep: rep} - } - bestOf := func(a, b match) match { - aScore := b.s - a.s + a.length - bScore := a.s - b.s + b.length - if a.rep < 0 { - aScore = aScore - int32(bits.Len32(uint32(a.offset)))/8 - } - if b.rep < 0 { - bScore = bScore - int32(bits.Len32(uint32(b.offset)))/8 - } - if aScore >= bScore { + if a.est+(a.s-b.s)*bitsPerByte>>10 < b.est+(b.s-a.s)*bitsPerByte>>10 { return a } return b } const goodEnough = 100 - nextHashL := hash8(cv, bestLongTableBits) - nextHashS := hash4x64(cv, bestShortTableBits) + nextHashL := hashLen(cv, bestLongTableBits, bestLongLen) + nextHashS := hashLen(cv, bestShortTableBits, bestShortLen) candidateL := e.longTable[nextHashL] candidateS := e.table[nextHashS] + matchAt := func(offset int32, s int32, first uint32, rep int32) match { + if s-offset >= e.maxMatchOff || load3232(src, offset) != first { + return match{s: s, est: highScore} + } + if debugAsserts { + if !bytes.Equal(src[s:s+4], src[offset:offset+4]) { + panic(fmt.Sprintf("first match mismatch: %v != %v, first: %08x", src[s:s+4], src[offset:offset+4], first)) + } + } + m := match{offset: offset, s: s, length: 4 + e.matchlen(s+4, offset+4, src), rep: rep} + m.estBits(bitsPerByte) + return m + } + best := bestOf(matchAt(candidateL.offset-e.cur, s, uint32(cv), -1), matchAt(candidateL.prev-e.cur, s, uint32(cv), -1)) best = bestOf(best, matchAt(candidateS.offset-e.cur, s, uint32(cv), -1)) best = bestOf(best, matchAt(candidateS.prev-e.cur, s, uint32(cv), -1)) + if canRepeat && best.length < goodEnough { - best = bestOf(best, matchAt(s-offset1+1, s+1, uint32(cv>>8), 1)) - best = bestOf(best, matchAt(s-offset2+1, s+1, uint32(cv>>8), 2)) - best = bestOf(best, matchAt(s-offset3+1, s+1, uint32(cv>>8), 3)) + cv32 := uint32(cv >> 8) + spp := s + 1 + best = bestOf(best, matchAt(spp-offset1, spp, cv32, 1)) + best = bestOf(best, matchAt(spp-offset2, spp, cv32, 2)) + best = bestOf(best, matchAt(spp-offset3, spp, cv32, 3)) if best.length > 0 { - best = bestOf(best, matchAt(s-offset1+3, s+3, uint32(cv>>24), 1)) - best = bestOf(best, matchAt(s-offset2+3, s+3, uint32(cv>>24), 2)) - best = bestOf(best, matchAt(s-offset3+3, s+3, uint32(cv>>24), 3)) + cv32 = uint32(cv >> 24) + spp += 2 + best = bestOf(best, matchAt(spp-offset1, spp, cv32, 1)) + best = bestOf(best, matchAt(spp-offset2, spp, cv32, 2)) + best = bestOf(best, matchAt(spp-offset3, spp, cv32, 3)) } } // Load next and check... @@ -209,22 +254,28 @@ encodeLoop: } s++ - candidateS = e.table[hash4x64(cv>>8, bestShortTableBits)] + candidateS = e.table[hashLen(cv>>8, bestShortTableBits, bestShortLen)] cv = load6432(src, s) cv2 := load6432(src, s+1) - candidateL = e.longTable[hash8(cv, bestLongTableBits)] - candidateL2 := e.longTable[hash8(cv2, bestLongTableBits)] + candidateL = e.longTable[hashLen(cv, bestLongTableBits, bestLongLen)] + candidateL2 := e.longTable[hashLen(cv2, bestLongTableBits, bestLongLen)] + // Short at s+1 best = bestOf(best, matchAt(candidateS.offset-e.cur, s, uint32(cv), -1)) + // Long at s+1, s+2 best = bestOf(best, matchAt(candidateL.offset-e.cur, s, uint32(cv), -1)) best = bestOf(best, matchAt(candidateL.prev-e.cur, s, uint32(cv), -1)) best = bestOf(best, matchAt(candidateL2.offset-e.cur, s+1, uint32(cv2), -1)) best = bestOf(best, matchAt(candidateL2.prev-e.cur, s+1, uint32(cv2), -1)) - + if false { + // Short at s+3. + // Too often worse... + best = bestOf(best, matchAt(e.table[hashLen(cv2>>8, bestShortTableBits, bestShortLen)].offset-e.cur, s+2, uint32(cv2>>8), -1)) + } // See if we can find a better match by checking where the current best ends. // Use that offset to see if we can find a better full match. if sAt := best.s + best.length; sAt < sLimit { - nextHashL := hash8(load6432(src, sAt), bestLongTableBits) + nextHashL := hashLen(load6432(src, sAt), bestLongTableBits, bestLongLen) candidateEnd := e.longTable[nextHashL] if pos := candidateEnd.offset - e.cur - best.length; pos >= 0 { bestEnd := bestOf(best, matchAt(pos, best.s, load3232(src, best.s), -1)) @@ -236,6 +287,12 @@ encodeLoop: } } + if debugAsserts { + if !bytes.Equal(src[best.s:best.s+best.length], src[best.offset:best.offset+best.length]) { + panic(fmt.Sprintf("match mismatch: %v != %v", src[best.s:best.s+best.length], src[best.offset:best.offset+best.length])) + } + } + // We have a match, we can store the forward value if best.rep > 0 { s = best.s @@ -284,8 +341,8 @@ encodeLoop: off := index0 + e.cur for index0 < s-1 { cv0 := load6432(src, index0) - h0 := hash8(cv0, bestLongTableBits) - h1 := hash4x64(cv0, bestShortTableBits) + h0 := hashLen(cv0, bestLongTableBits, bestLongLen) + h1 := hashLen(cv0, bestShortTableBits, bestShortLen) e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} e.table[h1] = prevEntry{offset: off, prev: e.table[h1].offset} off++ @@ -311,7 +368,7 @@ encodeLoop: panic(fmt.Sprintf("s (%d) <= t (%d)", s, t)) } - if debugAsserts && canRepeat && int(offset1) > len(src) { + if debugAsserts && int(offset1) > len(src) { panic("invalid offset") } @@ -352,8 +409,8 @@ encodeLoop: // every entry for index0 < s-1 { cv0 := load6432(src, index0) - h0 := hash8(cv0, bestLongTableBits) - h1 := hash4x64(cv0, bestShortTableBits) + h0 := hashLen(cv0, bestLongTableBits, bestLongLen) + h1 := hashLen(cv0, bestShortTableBits, bestShortLen) off := index0 + e.cur e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} e.table[h1] = prevEntry{offset: off, prev: e.table[h1].offset} @@ -374,8 +431,8 @@ encodeLoop: } // Store this, since we have it. - nextHashS := hash4x64(cv, bestShortTableBits) - nextHashL := hash8(cv, bestLongTableBits) + nextHashS := hashLen(cv, bestShortTableBits, bestShortLen) + nextHashL := hashLen(cv, bestLongTableBits, bestLongLen) // We have at least 4 byte match. // No need to check backwards. We come straight from a match @@ -425,7 +482,7 @@ func (e *bestFastEncoder) EncodeNoHist(blk *blockEnc, src []byte) { e.Encode(blk, src) } -// ResetDict will reset and set a dictionary if not nil +// Reset will reset and set a dictionary if not nil func (e *bestFastEncoder) Reset(d *dict, singleBlock bool) { e.resetBase(d, singleBlock) if d == nil { @@ -441,10 +498,10 @@ func (e *bestFastEncoder) Reset(d *dict, singleBlock bool) { const hashLog = bestShortTableBits cv := load6432(d.content, i-e.maxMatchOff) - nextHash := hash4x64(cv, hashLog) // 0 -> 4 - nextHash1 := hash4x64(cv>>8, hashLog) // 1 -> 5 - nextHash2 := hash4x64(cv>>16, hashLog) // 2 -> 6 - nextHash3 := hash4x64(cv>>24, hashLog) // 3 -> 7 + nextHash := hashLen(cv, hashLog, bestShortLen) // 0 -> 4 + nextHash1 := hashLen(cv>>8, hashLog, bestShortLen) // 1 -> 5 + nextHash2 := hashLen(cv>>16, hashLog, bestShortLen) // 2 -> 6 + nextHash3 := hashLen(cv>>24, hashLog, bestShortLen) // 3 -> 7 e.dictTable[nextHash] = prevEntry{ prev: e.dictTable[nextHash].offset, offset: i, @@ -472,7 +529,7 @@ func (e *bestFastEncoder) Reset(d *dict, singleBlock bool) { } if len(d.content) >= 8 { cv := load6432(d.content, 0) - h := hash8(cv, bestLongTableBits) + h := hashLen(cv, bestLongTableBits, bestLongLen) e.dictLongTable[h] = prevEntry{ offset: e.maxMatchOff, prev: e.dictLongTable[h].offset, @@ -482,7 +539,7 @@ func (e *bestFastEncoder) Reset(d *dict, singleBlock bool) { off := 8 // First to read for i := e.maxMatchOff + 1; i < end; i++ { cv = cv>>8 | (uint64(d.content[off]) << 56) - h := hash8(cv, bestLongTableBits) + h := hashLen(cv, bestLongTableBits, bestLongLen) e.dictLongTable[h] = prevEntry{ offset: i, prev: e.dictLongTable[h].offset, diff --git a/vendor/github.com/klauspost/compress/zstd/enc_better.go b/vendor/github.com/klauspost/compress/zstd/enc_better.go index eab7b5083e..602c05ee0c 100644 --- a/vendor/github.com/klauspost/compress/zstd/enc_better.go +++ b/vendor/github.com/klauspost/compress/zstd/enc_better.go @@ -9,6 +9,7 @@ import "fmt" const ( betterLongTableBits = 19 // Bits used in the long match table betterLongTableSize = 1 << betterLongTableBits // Size of the table + betterLongLen = 8 // Bytes used for table hash // Note: Increasing the short table bits or making the hash shorter // can actually lead to compression degradation since it will 'steal' more from the @@ -16,6 +17,7 @@ const ( // This greatly depends on the type of input. betterShortTableBits = 13 // Bits used in the short match table betterShortTableSize = 1 << betterShortTableBits // Size of the table + betterShortLen = 5 // Bytes used for table hash betterLongTableShardCnt = 1 << (betterLongTableBits - dictShardBits) // Number of shards in the table betterLongTableShardSize = betterLongTableSize / betterLongTableShardCnt // Size of an individual shard @@ -154,8 +156,8 @@ encodeLoop: panic("offset0 was 0") } - nextHashS := hash5(cv, betterShortTableBits) - nextHashL := hash8(cv, betterLongTableBits) + nextHashS := hashLen(cv, betterShortTableBits, betterShortLen) + nextHashL := hashLen(cv, betterLongTableBits, betterLongLen) candidateL := e.longTable[nextHashL] candidateS := e.table[nextHashS] @@ -214,10 +216,10 @@ encodeLoop: for index0 < s-1 { cv0 := load6432(src, index0) cv1 := cv0 >> 8 - h0 := hash8(cv0, betterLongTableBits) + h0 := hashLen(cv0, betterLongTableBits, betterLongLen) off := index0 + e.cur e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} - e.table[hash5(cv1, betterShortTableBits)] = tableEntry{offset: off + 1, val: uint32(cv1)} + e.table[hashLen(cv1, betterShortTableBits, betterShortLen)] = tableEntry{offset: off + 1, val: uint32(cv1)} index0 += 2 } cv = load6432(src, s) @@ -275,10 +277,10 @@ encodeLoop: for index0 < s-1 { cv0 := load6432(src, index0) cv1 := cv0 >> 8 - h0 := hash8(cv0, betterLongTableBits) + h0 := hashLen(cv0, betterLongTableBits, betterLongLen) off := index0 + e.cur e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} - e.table[hash5(cv1, betterShortTableBits)] = tableEntry{offset: off + 1, val: uint32(cv1)} + e.table[hashLen(cv1, betterShortTableBits, betterShortLen)] = tableEntry{offset: off + 1, val: uint32(cv1)} index0 += 2 } cv = load6432(src, s) @@ -353,7 +355,7 @@ encodeLoop: // See if we can find a long match at s+1 const checkAt = 1 cv := load6432(src, s+checkAt) - nextHashL = hash8(cv, betterLongTableBits) + nextHashL = hashLen(cv, betterLongTableBits, betterLongLen) candidateL = e.longTable[nextHashL] coffsetL = candidateL.offset - e.cur @@ -413,8 +415,8 @@ encodeLoop: } // Try to find a better match by searching for a long match at the end of the current best match - if true && s+matched < sLimit { - nextHashL := hash8(load6432(src, s+matched), betterLongTableBits) + if s+matched < sLimit { + nextHashL := hashLen(load6432(src, s+matched), betterLongTableBits, betterLongLen) cv := load3232(src, s) candidateL := e.longTable[nextHashL] coffsetL := candidateL.offset - e.cur - matched @@ -495,10 +497,10 @@ encodeLoop: for index0 < s-1 { cv0 := load6432(src, index0) cv1 := cv0 >> 8 - h0 := hash8(cv0, betterLongTableBits) + h0 := hashLen(cv0, betterLongTableBits, betterLongLen) off := index0 + e.cur e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} - e.table[hash5(cv1, betterShortTableBits)] = tableEntry{offset: off + 1, val: uint32(cv1)} + e.table[hashLen(cv1, betterShortTableBits, betterShortLen)] = tableEntry{offset: off + 1, val: uint32(cv1)} index0 += 2 } @@ -516,8 +518,8 @@ encodeLoop: } // Store this, since we have it. - nextHashS := hash5(cv, betterShortTableBits) - nextHashL := hash8(cv, betterLongTableBits) + nextHashS := hashLen(cv, betterShortTableBits, betterShortLen) + nextHashL := hashLen(cv, betterLongTableBits, betterLongLen) // We have at least 4 byte match. // No need to check backwards. We come straight from a match @@ -672,8 +674,8 @@ encodeLoop: panic("offset0 was 0") } - nextHashS := hash5(cv, betterShortTableBits) - nextHashL := hash8(cv, betterLongTableBits) + nextHashS := hashLen(cv, betterShortTableBits, betterShortLen) + nextHashL := hashLen(cv, betterLongTableBits, betterLongLen) candidateL := e.longTable[nextHashL] candidateS := e.table[nextHashS] @@ -734,11 +736,11 @@ encodeLoop: for index0 < s-1 { cv0 := load6432(src, index0) cv1 := cv0 >> 8 - h0 := hash8(cv0, betterLongTableBits) + h0 := hashLen(cv0, betterLongTableBits, betterLongLen) off := index0 + e.cur e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} e.markLongShardDirty(h0) - h1 := hash5(cv1, betterShortTableBits) + h1 := hashLen(cv1, betterShortTableBits, betterShortLen) e.table[h1] = tableEntry{offset: off + 1, val: uint32(cv1)} e.markShortShardDirty(h1) index0 += 2 @@ -798,11 +800,11 @@ encodeLoop: for index0 < s-1 { cv0 := load6432(src, index0) cv1 := cv0 >> 8 - h0 := hash8(cv0, betterLongTableBits) + h0 := hashLen(cv0, betterLongTableBits, betterLongLen) off := index0 + e.cur e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} e.markLongShardDirty(h0) - h1 := hash5(cv1, betterShortTableBits) + h1 := hashLen(cv1, betterShortTableBits, betterShortLen) e.table[h1] = tableEntry{offset: off + 1, val: uint32(cv1)} e.markShortShardDirty(h1) index0 += 2 @@ -879,7 +881,7 @@ encodeLoop: // See if we can find a long match at s+1 const checkAt = 1 cv := load6432(src, s+checkAt) - nextHashL = hash8(cv, betterLongTableBits) + nextHashL = hashLen(cv, betterLongTableBits, betterLongLen) candidateL = e.longTable[nextHashL] coffsetL = candidateL.offset - e.cur @@ -940,7 +942,7 @@ encodeLoop: } // Try to find a better match by searching for a long match at the end of the current best match if s+matched < sLimit { - nextHashL := hash8(load6432(src, s+matched), betterLongTableBits) + nextHashL := hashLen(load6432(src, s+matched), betterLongTableBits, betterLongLen) cv := load3232(src, s) candidateL := e.longTable[nextHashL] coffsetL := candidateL.offset - e.cur - matched @@ -1021,11 +1023,11 @@ encodeLoop: for index0 < s-1 { cv0 := load6432(src, index0) cv1 := cv0 >> 8 - h0 := hash8(cv0, betterLongTableBits) + h0 := hashLen(cv0, betterLongTableBits, betterLongLen) off := index0 + e.cur e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset} e.markLongShardDirty(h0) - h1 := hash5(cv1, betterShortTableBits) + h1 := hashLen(cv1, betterShortTableBits, betterShortLen) e.table[h1] = tableEntry{offset: off + 1, val: uint32(cv1)} e.markShortShardDirty(h1) index0 += 2 @@ -1045,8 +1047,8 @@ encodeLoop: } // Store this, since we have it. - nextHashS := hash5(cv, betterShortTableBits) - nextHashL := hash8(cv, betterLongTableBits) + nextHashS := hashLen(cv, betterShortTableBits, betterShortLen) + nextHashL := hashLen(cv, betterLongTableBits, betterLongLen) // We have at least 4 byte match. // No need to check backwards. We come straight from a match @@ -1113,10 +1115,10 @@ func (e *betterFastEncoderDict) Reset(d *dict, singleBlock bool) { const hashLog = betterShortTableBits cv := load6432(d.content, i-e.maxMatchOff) - nextHash := hash5(cv, hashLog) // 0 -> 4 - nextHash1 := hash5(cv>>8, hashLog) // 1 -> 5 - nextHash2 := hash5(cv>>16, hashLog) // 2 -> 6 - nextHash3 := hash5(cv>>24, hashLog) // 3 -> 7 + nextHash := hashLen(cv, hashLog, betterShortLen) // 0 -> 4 + nextHash1 := hashLen(cv>>8, hashLog, betterShortLen) // 1 -> 5 + nextHash2 := hashLen(cv>>16, hashLog, betterShortLen) // 2 -> 6 + nextHash3 := hashLen(cv>>24, hashLog, betterShortLen) // 3 -> 7 e.dictTable[nextHash] = tableEntry{ val: uint32(cv), offset: i, @@ -1145,7 +1147,7 @@ func (e *betterFastEncoderDict) Reset(d *dict, singleBlock bool) { } if len(d.content) >= 8 { cv := load6432(d.content, 0) - h := hash8(cv, betterLongTableBits) + h := hashLen(cv, betterLongTableBits, betterLongLen) e.dictLongTable[h] = prevEntry{ offset: e.maxMatchOff, prev: e.dictLongTable[h].offset, @@ -1155,7 +1157,7 @@ func (e *betterFastEncoderDict) Reset(d *dict, singleBlock bool) { off := 8 // First to read for i := e.maxMatchOff + 1; i < end; i++ { cv = cv>>8 | (uint64(d.content[off]) << 56) - h := hash8(cv, betterLongTableBits) + h := hashLen(cv, betterLongTableBits, betterLongLen) e.dictLongTable[h] = prevEntry{ offset: i, prev: e.dictLongTable[h].offset, diff --git a/vendor/github.com/klauspost/compress/zstd/enc_dfast.go b/vendor/github.com/klauspost/compress/zstd/enc_dfast.go index 96b21b90e8..d6b3104240 100644 --- a/vendor/github.com/klauspost/compress/zstd/enc_dfast.go +++ b/vendor/github.com/klauspost/compress/zstd/enc_dfast.go @@ -10,6 +10,7 @@ const ( dFastLongTableBits = 17 // Bits used in the long match table dFastLongTableSize = 1 << dFastLongTableBits // Size of the table dFastLongTableMask = dFastLongTableSize - 1 // Mask for table indices. Redundant, but can eliminate bounds checks. + dFastLongLen = 8 // Bytes used for table hash dLongTableShardCnt = 1 << (dFastLongTableBits - dictShardBits) // Number of shards in the table dLongTableShardSize = dFastLongTableSize / tableShardCnt // Size of an individual shard @@ -17,6 +18,8 @@ const ( dFastShortTableBits = tableBits // Bits used in the short match table dFastShortTableSize = 1 << dFastShortTableBits // Size of the table dFastShortTableMask = dFastShortTableSize - 1 // Mask for table indices. Redundant, but can eliminate bounds checks. + dFastShortLen = 5 // Bytes used for table hash + ) type doubleFastEncoder struct { @@ -124,8 +127,8 @@ encodeLoop: panic("offset0 was 0") } - nextHashS := hash5(cv, dFastShortTableBits) - nextHashL := hash8(cv, dFastLongTableBits) + nextHashS := hashLen(cv, dFastShortTableBits, dFastShortLen) + nextHashL := hashLen(cv, dFastLongTableBits, dFastLongLen) candidateL := e.longTable[nextHashL] candidateS := e.table[nextHashS] @@ -208,7 +211,7 @@ encodeLoop: // See if we can find a long match at s+1 const checkAt = 1 cv := load6432(src, s+checkAt) - nextHashL = hash8(cv, dFastLongTableBits) + nextHashL = hashLen(cv, dFastLongTableBits, dFastLongLen) candidateL = e.longTable[nextHashL] coffsetL = s - (candidateL.offset - e.cur) + checkAt @@ -304,16 +307,16 @@ encodeLoop: cv1 := load6432(src, index1) te0 := tableEntry{offset: index0 + e.cur, val: uint32(cv0)} te1 := tableEntry{offset: index1 + e.cur, val: uint32(cv1)} - e.longTable[hash8(cv0, dFastLongTableBits)] = te0 - e.longTable[hash8(cv1, dFastLongTableBits)] = te1 + e.longTable[hashLen(cv0, dFastLongTableBits, dFastLongLen)] = te0 + e.longTable[hashLen(cv1, dFastLongTableBits, dFastLongLen)] = te1 cv0 >>= 8 cv1 >>= 8 te0.offset++ te1.offset++ te0.val = uint32(cv0) te1.val = uint32(cv1) - e.table[hash5(cv0, dFastShortTableBits)] = te0 - e.table[hash5(cv1, dFastShortTableBits)] = te1 + e.table[hashLen(cv0, dFastShortTableBits, dFastShortLen)] = te0 + e.table[hashLen(cv1, dFastShortTableBits, dFastShortLen)] = te1 cv = load6432(src, s) @@ -330,8 +333,8 @@ encodeLoop: } // Store this, since we have it. - nextHashS := hash5(cv, dFastShortTableBits) - nextHashL := hash8(cv, dFastLongTableBits) + nextHashS := hashLen(cv, dFastShortTableBits, dFastShortLen) + nextHashL := hashLen(cv, dFastLongTableBits, dFastLongLen) // We have at least 4 byte match. // No need to check backwards. We come straight from a match @@ -436,8 +439,8 @@ encodeLoop: var t int32 for { - nextHashS := hash5(cv, dFastShortTableBits) - nextHashL := hash8(cv, dFastLongTableBits) + nextHashS := hashLen(cv, dFastShortTableBits, dFastShortLen) + nextHashL := hashLen(cv, dFastLongTableBits, dFastLongLen) candidateL := e.longTable[nextHashL] candidateS := e.table[nextHashS] @@ -521,7 +524,7 @@ encodeLoop: // See if we can find a long match at s+1 const checkAt = 1 cv := load6432(src, s+checkAt) - nextHashL = hash8(cv, dFastLongTableBits) + nextHashL = hashLen(cv, dFastLongTableBits, dFastLongLen) candidateL = e.longTable[nextHashL] coffsetL = s - (candidateL.offset - e.cur) + checkAt @@ -614,16 +617,16 @@ encodeLoop: cv1 := load6432(src, index1) te0 := tableEntry{offset: index0 + e.cur, val: uint32(cv0)} te1 := tableEntry{offset: index1 + e.cur, val: uint32(cv1)} - e.longTable[hash8(cv0, dFastLongTableBits)] = te0 - e.longTable[hash8(cv1, dFastLongTableBits)] = te1 + e.longTable[hashLen(cv0, dFastLongTableBits, dFastLongLen)] = te0 + e.longTable[hashLen(cv1, dFastLongTableBits, dFastLongLen)] = te1 cv0 >>= 8 cv1 >>= 8 te0.offset++ te1.offset++ te0.val = uint32(cv0) te1.val = uint32(cv1) - e.table[hash5(cv0, dFastShortTableBits)] = te0 - e.table[hash5(cv1, dFastShortTableBits)] = te1 + e.table[hashLen(cv0, dFastShortTableBits, dFastShortLen)] = te0 + e.table[hashLen(cv1, dFastShortTableBits, dFastShortLen)] = te1 cv = load6432(src, s) @@ -640,8 +643,8 @@ encodeLoop: } // Store this, since we have it. - nextHashS := hash5(cv1>>8, dFastShortTableBits) - nextHashL := hash8(cv, dFastLongTableBits) + nextHashS := hashLen(cv1>>8, dFastShortTableBits, dFastShortLen) + nextHashL := hashLen(cv, dFastLongTableBits, dFastLongLen) // We have at least 4 byte match. // No need to check backwards. We come straight from a match @@ -782,8 +785,8 @@ encodeLoop: panic("offset0 was 0") } - nextHashS := hash5(cv, dFastShortTableBits) - nextHashL := hash8(cv, dFastLongTableBits) + nextHashS := hashLen(cv, dFastShortTableBits, dFastShortLen) + nextHashL := hashLen(cv, dFastLongTableBits, dFastLongLen) candidateL := e.longTable[nextHashL] candidateS := e.table[nextHashS] @@ -868,7 +871,7 @@ encodeLoop: // See if we can find a long match at s+1 const checkAt = 1 cv := load6432(src, s+checkAt) - nextHashL = hash8(cv, dFastLongTableBits) + nextHashL = hashLen(cv, dFastLongTableBits, dFastLongLen) candidateL = e.longTable[nextHashL] coffsetL = s - (candidateL.offset - e.cur) + checkAt @@ -965,8 +968,8 @@ encodeLoop: cv1 := load6432(src, index1) te0 := tableEntry{offset: index0 + e.cur, val: uint32(cv0)} te1 := tableEntry{offset: index1 + e.cur, val: uint32(cv1)} - longHash1 := hash8(cv0, dFastLongTableBits) - longHash2 := hash8(cv0, dFastLongTableBits) + longHash1 := hashLen(cv0, dFastLongTableBits, dFastLongLen) + longHash2 := hashLen(cv0, dFastLongTableBits, dFastLongLen) e.longTable[longHash1] = te0 e.longTable[longHash2] = te1 e.markLongShardDirty(longHash1) @@ -977,8 +980,8 @@ encodeLoop: te1.offset++ te0.val = uint32(cv0) te1.val = uint32(cv1) - hashVal1 := hash5(cv0, dFastShortTableBits) - hashVal2 := hash5(cv1, dFastShortTableBits) + hashVal1 := hashLen(cv0, dFastShortTableBits, dFastShortLen) + hashVal2 := hashLen(cv1, dFastShortTableBits, dFastShortLen) e.table[hashVal1] = te0 e.markShardDirty(hashVal1) e.table[hashVal2] = te1 @@ -999,8 +1002,8 @@ encodeLoop: } // Store this, since we have it. - nextHashS := hash5(cv, dFastShortTableBits) - nextHashL := hash8(cv, dFastLongTableBits) + nextHashS := hashLen(cv, dFastShortTableBits, dFastShortLen) + nextHashL := hashLen(cv, dFastLongTableBits, dFastLongLen) // We have at least 4 byte match. // No need to check backwards. We come straight from a match @@ -1071,14 +1074,14 @@ func (e *doubleFastEncoderDict) Reset(d *dict, singleBlock bool) { } if len(d.content) >= 8 { cv := load6432(d.content, 0) - e.dictLongTable[hash8(cv, dFastLongTableBits)] = tableEntry{ + e.dictLongTable[hashLen(cv, dFastLongTableBits, dFastLongLen)] = tableEntry{ val: uint32(cv), offset: e.maxMatchOff, } end := int32(len(d.content)) - 8 + e.maxMatchOff for i := e.maxMatchOff + 1; i < end; i++ { cv = cv>>8 | (uint64(d.content[i-e.maxMatchOff+7]) << 56) - e.dictLongTable[hash8(cv, dFastLongTableBits)] = tableEntry{ + e.dictLongTable[hashLen(cv, dFastLongTableBits, dFastLongLen)] = tableEntry{ val: uint32(cv), offset: i, } diff --git a/vendor/github.com/klauspost/compress/zstd/enc_fast.go b/vendor/github.com/klauspost/compress/zstd/enc_fast.go index 2246d286dc..f2502629bc 100644 --- a/vendor/github.com/klauspost/compress/zstd/enc_fast.go +++ b/vendor/github.com/klauspost/compress/zstd/enc_fast.go @@ -11,12 +11,13 @@ import ( ) const ( - tableBits = 15 // Bits used in the table - tableSize = 1 << tableBits // Size of the table - tableShardCnt = 1 << (tableBits - dictShardBits) // Number of shards in the table - tableShardSize = tableSize / tableShardCnt // Size of an individual shard - tableMask = tableSize - 1 // Mask for table indices. Redundant, but can eliminate bounds checks. - maxMatchLength = 131074 + tableBits = 15 // Bits used in the table + tableSize = 1 << tableBits // Size of the table + tableShardCnt = 1 << (tableBits - dictShardBits) // Number of shards in the table + tableShardSize = tableSize / tableShardCnt // Size of an individual shard + tableFastHashLen = 6 + tableMask = tableSize - 1 // Mask for table indices. Redundant, but can eliminate bounds checks. + maxMatchLength = 131074 ) type tableEntry struct { @@ -122,8 +123,8 @@ encodeLoop: panic("offset0 was 0") } - nextHash := hash6(cv, hashLog) - nextHash2 := hash6(cv>>8, hashLog) + nextHash := hashLen(cv, hashLog, tableFastHashLen) + nextHash2 := hashLen(cv>>8, hashLog, tableFastHashLen) candidate := e.table[nextHash] candidate2 := e.table[nextHash2] repIndex := s - offset1 + 2 @@ -301,7 +302,7 @@ encodeLoop: } // Store this, since we have it. - nextHash := hash6(cv, hashLog) + nextHash := hashLen(cv, hashLog, tableFastHashLen) e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)} seq.matchLen = uint32(l) - zstdMinMatch seq.litLen = 0 @@ -405,8 +406,8 @@ encodeLoop: // By not using them for the first 3 matches for { - nextHash := hash6(cv, hashLog) - nextHash2 := hash6(cv>>8, hashLog) + nextHash := hashLen(cv, hashLog, tableFastHashLen) + nextHash2 := hashLen(cv>>8, hashLog, tableFastHashLen) candidate := e.table[nextHash] candidate2 := e.table[nextHash2] repIndex := s - offset1 + 2 @@ -589,7 +590,7 @@ encodeLoop: } // Store this, since we have it. - nextHash := hash6(cv, hashLog) + nextHash := hashLen(cv, hashLog, tableFastHashLen) e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)} seq.matchLen = uint32(l) - zstdMinMatch seq.litLen = 0 @@ -715,8 +716,8 @@ encodeLoop: panic("offset0 was 0") } - nextHash := hash6(cv, hashLog) - nextHash2 := hash6(cv>>8, hashLog) + nextHash := hashLen(cv, hashLog, tableFastHashLen) + nextHash2 := hashLen(cv>>8, hashLog, tableFastHashLen) candidate := e.table[nextHash] candidate2 := e.table[nextHash2] repIndex := s - offset1 + 2 @@ -896,7 +897,7 @@ encodeLoop: } // Store this, since we have it. - nextHash := hash6(cv, hashLog) + nextHash := hashLen(cv, hashLog, tableFastHashLen) e.table[nextHash] = tableEntry{offset: s + e.cur, val: uint32(cv)} e.markShardDirty(nextHash) seq.matchLen = uint32(l) - zstdMinMatch @@ -957,9 +958,9 @@ func (e *fastEncoderDict) Reset(d *dict, singleBlock bool) { const hashLog = tableBits cv := load6432(d.content, i-e.maxMatchOff) - nextHash := hash6(cv, hashLog) // 0 -> 5 - nextHash1 := hash6(cv>>8, hashLog) // 1 -> 6 - nextHash2 := hash6(cv>>16, hashLog) // 2 -> 7 + nextHash := hashLen(cv, hashLog, tableFastHashLen) // 0 -> 5 + nextHash1 := hashLen(cv>>8, hashLog, tableFastHashLen) // 1 -> 6 + nextHash2 := hashLen(cv>>16, hashLog, tableFastHashLen) // 2 -> 7 e.dictTable[nextHash] = tableEntry{ val: uint32(cv), offset: i, diff --git a/vendor/github.com/klauspost/compress/zstd/encoder.go b/vendor/github.com/klauspost/compress/zstd/encoder.go index ea85548fc9..e6e315969b 100644 --- a/vendor/github.com/klauspost/compress/zstd/encoder.go +++ b/vendor/github.com/klauspost/compress/zstd/encoder.go @@ -33,7 +33,7 @@ type encoder interface { Block() *blockEnc CRC() *xxhash.Digest AppendCRC([]byte) []byte - WindowSize(size int) int32 + WindowSize(size int64) int32 UseBlock(*blockEnc) Reset(d *dict, singleBlock bool) } @@ -48,6 +48,8 @@ type encoderState struct { err error writeErr error nWritten int64 + nInput int64 + frameContentSize int64 headerWritten bool eofWritten bool fullFrameWritten bool @@ -120,7 +122,21 @@ func (e *Encoder) Reset(w io.Writer) { s.w = w s.err = nil s.nWritten = 0 + s.nInput = 0 s.writeErr = nil + s.frameContentSize = 0 +} + +// ResetContentSize will reset and set a content size for the next stream. +// If the bytes written does not match the size given an error will be returned +// when calling Close(). +// This is removed when Reset is called. +// Sizes <= 0 results in no content size set. +func (e *Encoder) ResetContentSize(w io.Writer, size int64) { + e.Reset(w) + if size >= 0 { + e.state.frameContentSize = size + } } // Write data to the encoder. @@ -190,6 +206,7 @@ func (e *Encoder) nextBlock(final bool) error { return s.err } s.nWritten += int64(n2) + s.nInput += int64(len(s.filling)) s.current = s.current[:0] s.filling = s.filling[:0] s.headerWritten = true @@ -200,8 +217,8 @@ func (e *Encoder) nextBlock(final bool) error { var tmp [maxHeaderSize]byte fh := frameHeader{ - ContentSize: 0, - WindowSize: uint32(s.encoder.WindowSize(0)), + ContentSize: uint64(s.frameContentSize), + WindowSize: uint32(s.encoder.WindowSize(s.frameContentSize)), SingleSegment: false, Checksum: e.o.crc, DictID: e.o.dict.ID(), @@ -243,6 +260,7 @@ func (e *Encoder) nextBlock(final bool) error { // Move blocks forward. s.filling, s.current, s.previous = s.previous[:0], s.filling, s.current + s.nInput += int64(len(s.current)) s.wg.Add(1) go func(src []byte) { if debugEncoder { @@ -394,6 +412,11 @@ func (e *Encoder) Close() error { if err != nil { return err } + if s.frameContentSize > 0 { + if s.nInput != s.frameContentSize { + return fmt.Errorf("frame content size %d given, but %d bytes was written", s.frameContentSize, s.nInput) + } + } if e.state.fullFrameWritten { return s.err } @@ -470,7 +493,7 @@ func (e *Encoder) EncodeAll(src, dst []byte) []byte { } fh := frameHeader{ ContentSize: uint64(len(src)), - WindowSize: uint32(enc.WindowSize(len(src))), + WindowSize: uint32(enc.WindowSize(int64(len(src)))), SingleSegment: single, Checksum: e.o.crc, DictID: e.o.dict.ID(), diff --git a/vendor/github.com/klauspost/compress/zstd/encoder_options.go b/vendor/github.com/klauspost/compress/zstd/encoder_options.go index 16d4ab63c1..7d29e1d689 100644 --- a/vendor/github.com/klauspost/compress/zstd/encoder_options.go +++ b/vendor/github.com/klauspost/compress/zstd/encoder_options.go @@ -189,7 +189,7 @@ func EncoderLevelFromZstd(level int) EncoderLevel { case level >= 6 && level < 10: return SpeedBetterCompression case level >= 10: - return SpeedBetterCompression + return SpeedBestCompression } return SpeedDefault } diff --git a/vendor/github.com/klauspost/compress/zstd/hash.go b/vendor/github.com/klauspost/compress/zstd/hash.go index 4a752067fc..cf33f29a1b 100644 --- a/vendor/github.com/klauspost/compress/zstd/hash.go +++ b/vendor/github.com/klauspost/compress/zstd/hash.go @@ -13,24 +13,24 @@ const ( prime8bytes = 0xcf1bbcdcb7a56463 ) -// hashLen returns a hash of the lowest l bytes of u for a size size of h bytes. -// l must be >=4 and <=8. Any other value will return hash for 4 bytes. -// h should always be <32. -// Preferably h and l should be a constant. -// FIXME: This does NOT get resolved, if 'mls' is constant, -// so this cannot be used. -func hashLen(u uint64, hashLog, mls uint8) uint32 { +// hashLen returns a hash of the lowest mls bytes of with length output bits. +// mls must be >=3 and <=8. Any other value will return hash for 4 bytes. +// length should always be < 32. +// Preferably length and mls should be a constant for inlining. +func hashLen(u uint64, length, mls uint8) uint32 { switch mls { + case 3: + return (uint32(u<<8) * prime3bytes) >> (32 - length) case 5: - return hash5(u, hashLog) + return uint32(((u << (64 - 40)) * prime5bytes) >> (64 - length)) case 6: - return hash6(u, hashLog) + return uint32(((u << (64 - 48)) * prime6bytes) >> (64 - length)) case 7: - return hash7(u, hashLog) + return uint32(((u << (64 - 56)) * prime7bytes) >> (64 - length)) case 8: - return hash8(u, hashLog) + return uint32((u * prime8bytes) >> (64 - length)) default: - return hash4x64(u, hashLog) + return (uint32(u) * prime4bytes) >> (32 - length) } } @@ -39,39 +39,3 @@ func hashLen(u uint64, hashLog, mls uint8) uint32 { func hash3(u uint32, h uint8) uint32 { return ((u << (32 - 24)) * prime3bytes) >> ((32 - h) & 31) } - -// hash4 returns the hash of u to fit in a hash table with h bits. -// Preferably h should be a constant and should always be <32. -func hash4(u uint32, h uint8) uint32 { - return (u * prime4bytes) >> ((32 - h) & 31) -} - -// hash4x64 returns the hash of the lowest 4 bytes of u to fit in a hash table with h bits. -// Preferably h should be a constant and should always be <32. -func hash4x64(u uint64, h uint8) uint32 { - return (uint32(u) * prime4bytes) >> ((32 - h) & 31) -} - -// hash5 returns the hash of the lowest 5 bytes of u to fit in a hash table with h bits. -// Preferably h should be a constant and should always be <64. -func hash5(u uint64, h uint8) uint32 { - return uint32(((u << (64 - 40)) * prime5bytes) >> ((64 - h) & 63)) -} - -// hash6 returns the hash of the lowest 6 bytes of u to fit in a hash table with h bits. -// Preferably h should be a constant and should always be <64. -func hash6(u uint64, h uint8) uint32 { - return uint32(((u << (64 - 48)) * prime6bytes) >> ((64 - h) & 63)) -} - -// hash7 returns the hash of the lowest 7 bytes of u to fit in a hash table with h bits. -// Preferably h should be a constant and should always be <64. -func hash7(u uint64, h uint8) uint32 { - return uint32(((u << (64 - 56)) * prime7bytes) >> ((64 - h) & 63)) -} - -// hash8 returns the hash of u to fit in a hash table with h bits. -// Preferably h should be a constant and should always be <64. -func hash8(u uint64, h uint8) uint32 { - return uint32((u * prime8bytes) >> ((64 - h) & 63)) -} diff --git a/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_amd64.go b/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_amd64.go index 35318d7c46..3ddbd5c0b0 100644 --- a/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_amd64.go +++ b/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_amd64.go @@ -1,6 +1,5 @@ -// +build !appengine -// +build gc -// +build !purego +//go:build !appengine && gc && !purego +// +build !appengine,gc,!purego package xxhash diff --git a/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_other.go b/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_other.go index 4a5a821603..1f52f296e7 100644 --- a/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_other.go +++ b/vendor/github.com/klauspost/compress/zstd/internal/xxhash/xxhash_other.go @@ -1,3 +1,4 @@ +//go:build !amd64 || appengine || !gc || purego // +build !amd64 appengine !gc purego package xxhash diff --git a/vendor/github.com/klauspost/compress/zstd/snappy.go b/vendor/github.com/klauspost/compress/zstd/snappy.go index 0372b1714a..9e1baad73b 100644 --- a/vendor/github.com/klauspost/compress/zstd/snappy.go +++ b/vendor/github.com/klauspost/compress/zstd/snappy.go @@ -10,8 +10,8 @@ import ( "hash/crc32" "io" - "github.com/golang/snappy" "github.com/klauspost/compress/huff0" + snappy "github.com/klauspost/compress/internal/snapref" ) const ( diff --git a/vendor/github.com/klauspost/compress/zstd/zip.go b/vendor/github.com/klauspost/compress/zstd/zip.go index 9325b928ae..967f29b312 100644 --- a/vendor/github.com/klauspost/compress/zstd/zip.go +++ b/vendor/github.com/klauspost/compress/zstd/zip.go @@ -64,8 +64,9 @@ func (r *pooledZipReader) Close() error { } type pooledZipWriter struct { - mu sync.Mutex // guards Close and Read - enc *Encoder + mu sync.Mutex // guards Close and Read + enc *Encoder + pool *sync.Pool } func (w *pooledZipWriter) Write(p []byte) (n int, err error) { @@ -83,7 +84,7 @@ func (w *pooledZipWriter) Close() error { var err error if w.enc != nil { err = w.enc.Close() - zipReaderPool.Put(w.enc) + w.pool.Put(w.enc) w.enc = nil } return err @@ -104,7 +105,7 @@ func ZipCompressor(opts ...EOption) func(w io.Writer) (io.WriteCloser, error) { return nil, err } } - return &pooledZipWriter{enc: enc}, nil + return &pooledZipWriter{enc: enc, pool: &pool}, nil } } diff --git a/vendor/modules.txt b/vendor/modules.txt index d6179d89cb..6c4cdb29c5 100644 --- a/vendor/modules.txt +++ b/vendor/modules.txt @@ -9,7 +9,7 @@ github.com/alessio/shellescape github.com/containerd/containerd/errdefs github.com/containerd/containerd/log github.com/containerd/containerd/platforms -# github.com/containerd/stargz-snapshotter/estargz v0.7.0 +# github.com/containerd/stargz-snapshotter/estargz v0.8.0 ## explicit github.com/containerd/stargz-snapshotter/estargz github.com/containerd/stargz-snapshotter/estargz/errorutil @@ -78,9 +78,6 @@ github.com/golang/protobuf/ptypes github.com/golang/protobuf/ptypes/any github.com/golang/protobuf/ptypes/duration github.com/golang/protobuf/ptypes/timestamp -# github.com/golang/snappy v0.0.4 -## explicit -github.com/golang/snappy # github.com/google/go-cmp v0.5.6 ## explicit github.com/google/go-cmp/cmp @@ -136,10 +133,11 @@ github.com/hashicorp/hcl/json/scanner github.com/hashicorp/hcl/json/token # github.com/inconshreveable/mousetrap v1.0.0 github.com/inconshreveable/mousetrap -# github.com/klauspost/compress v1.13.1 -## explicit +# github.com/klauspost/compress v1.13.5 +github.com/klauspost/compress github.com/klauspost/compress/fse github.com/klauspost/compress/huff0 +github.com/klauspost/compress/internal/snapref github.com/klauspost/compress/zstd github.com/klauspost/compress/zstd/internal/xxhash # github.com/magiconair/properties v1.8.5