[chore] update dependencies, bump to Go 1.19.1 (#826)

* update dependencies, bump Go version to 1.19

* bump test image Go version

* update golangci-lint

* update gotosocial-drone-build

* sign

* linting, go fmt

* update swagger docs

* update swagger docs

* whitespace

* update contributing.md

* fuckin whoopsie doopsie

* linterino, linteroni

* fix followrequest test not starting processor

* fix other api/client tests not starting processor

* fix remaining tests where processor not started

* bump go-runners version

* don't check last-webfingered-at, processor may have updated this

* update swagger command

* update bun to latest version

* fix embed to work the same as before with new bun

Signed-off-by: kim <grufwub@gmail.com>
Co-authored-by: tsmethurst <tobi.smethurst@protonmail.com>

8 files changed, 1946 insertions, 452 deletions

diff --git a/vendor/github.com/klauspost/compress/s2/README.md b/vendor/github.com/klauspost/compress/s2/README.md
index 119793456..73c0c462d 100644
--- a/vendor/github.com/klauspost/compress/s2/README.md
+++ b/vendor/github.com/klauspost/compress/s2/README.md
@@ -19,6 +19,7 @@ This is important, so you don't have to worry about spending CPU cycles on alrea
 * Adjustable compression (3 levels) 
 * Concurrent stream compression
 * Faster decompression, even for Snappy compatible content
+* Concurrent Snappy/S2 stream decompression
 * Ability to quickly skip forward in compressed stream
 * Random seeking with indexes
 * Compatible with reading Snappy compressed content
@@ -415,6 +416,25 @@ Without assembly decompression is also very fast; single goroutine decompression
 
 Even though S2 typically compresses better than Snappy, decompression speed is always better. 
 
+### Concurrent Stream Decompression
+
+For full stream decompression S2 offers a [DecodeConcurrent](https://pkg.go.dev/github.com/klauspost/compress/s2#Reader.DecodeConcurrent) 
+that will decode a full stream using multiple goroutines.
+
+Example scaling, AMD Ryzen 3950X, 16 cores, decompression using `s2d -bench=3 <input>`, best of 3: 
+
+| Input                                     | `-cpu=1`   | `-cpu=2`   | `-cpu=4`   | `-cpu=8`   | `-cpu=16`   |
+|-------------------------------------------|------------|------------|------------|------------|-------------|
+| enwik10.snappy                            | 1098.6MB/s | 1819.8MB/s | 3625.6MB/s | 6910.6MB/s | 10818.2MB/s |
+| enwik10.s2                                | 1303.5MB/s | 2606.1MB/s | 4847.9MB/s | 8878.4MB/s | 9592.1MB/s  |
+| sofia-air-quality-dataset.tar.snappy      | 1302.0MB/s | 2165.0MB/s | 4244.5MB/s | 8241.0MB/s | 12920.5MB/s |
+| sofia-air-quality-dataset.tar.s2          | 1399.2MB/s | 2463.2MB/s | 5196.5MB/s | 9639.8MB/s | 11439.5MB/s |
+| sofia-air-quality-dataset.tar.s2 (no asm) | 837.5MB/s  | 1652.6MB/s | 3183.6MB/s | 5945.0MB/s | 9620.7MB/s  |
+
+Scaling can be expected to be pretty linear until memory bandwidth is saturated. 
+
+For now the DecodeConcurrent can only be used for full streams without seeking or combining with regular reads.
+
 ## Block compression
 
 
@@ -873,7 +893,7 @@ for each entry {
     }
     
     // Uncompressed uses previous offset and adds EstBlockSize
-    entry[entryNum].UncompressedOffset = entry[entryNum-1].UncompressedOffset + EstBlockSize
+    entry[entryNum].UncompressedOffset = entry[entryNum-1].UncompressedOffset + EstBlockSize + uOff
 }
 
 
@@ -901,6 +921,14 @@ for each entry {
 }
 ```
 
+To decode from any given uncompressed offset `(wantOffset)`:
+
+* Iterate entries until `entry[n].UncompressedOffset > wantOffset`.
+* Start decoding from `entry[n-1].CompressedOffset`.
+* Discard `entry[n-1].UncompressedOffset - wantOffset` bytes from the decoded stream.
+
+See [using indexes](https://github.com/klauspost/compress/tree/master/s2#using-indexes) for functions that perform the operations with a simpler interface.
+
 # Format Extensions
 
 * Frame [Stream identifier](https://github.com/google/snappy/blob/master/framing_format.txt#L68) changed from `sNaPpY` to `S2sTwO`.
diff --git a/vendor/github.com/klauspost/compress/s2/decode.go b/vendor/github.com/klauspost/compress/s2/decode.go
index 9e7fce885..27c0f3c2c 100644
--- a/vendor/github.com/klauspost/compress/s2/decode.go
+++ b/vendor/github.com/klauspost/compress/s2/decode.go
@@ -11,6 +11,9 @@ import (
 	"fmt"
 	"io"
 	"io/ioutil"
+	"math"
+	"runtime"
+	"sync"
 )
 
 var (
@@ -169,6 +172,14 @@ func ReaderSkippableCB(id uint8, fn func(r io.Reader) error) ReaderOption {
 	}
 }
 
+// ReaderIgnoreCRC will make the reader skip CRC calculation and checks.
+func ReaderIgnoreCRC() ReaderOption {
+	return func(r *Reader) error {
+		r.ignoreCRC = true
+		return nil
+	}
+}
+
 // Reader is an io.Reader that can read Snappy-compressed bytes.
 type Reader struct {
 	r           io.Reader
@@ -191,18 +202,19 @@ type Reader struct {
 	paramsOK       bool
 	snappyFrame    bool
 	ignoreStreamID bool
+	ignoreCRC      bool
 }
 
 // ensureBufferSize will ensure that the buffer can take at least n bytes.
 // If false is returned the buffer exceeds maximum allowed size.
 func (r *Reader) ensureBufferSize(n int) bool {
-	if len(r.buf) >= n {
-		return true
-	}
 	if n > r.maxBufSize {
 		r.err = ErrCorrupt
 		return false
 	}
+	if cap(r.buf) >= n {
+		return true
+	}
 	// Realloc buffer.
 	r.buf = make([]byte, n)
 	return true
@@ -220,6 +232,7 @@ func (r *Reader) Reset(reader io.Reader) {
 	r.err = nil
 	r.i = 0
 	r.j = 0
+	r.blockStart = 0
 	r.readHeader = r.ignoreStreamID
 }
 
@@ -344,7 +357,7 @@ func (r *Reader) Read(p []byte) (int, error) {
 				r.err = err
 				return 0, r.err
 			}
-			if crc(r.decoded[:n]) != checksum {
+			if !r.ignoreCRC && crc(r.decoded[:n]) != checksum {
 				r.err = ErrCRC
 				return 0, r.err
 			}
@@ -385,7 +398,7 @@ func (r *Reader) Read(p []byte) (int, error) {
 			if !r.readFull(r.decoded[:n], false) {
 				return 0, r.err
 			}
-			if crc(r.decoded[:n]) != checksum {
+			if !r.ignoreCRC && crc(r.decoded[:n]) != checksum {
 				r.err = ErrCRC
 				return 0, r.err
 			}
@@ -435,6 +448,259 @@ func (r *Reader) Read(p []byte) (int, error) {
 	}
 }
 
+// DecodeConcurrent will decode the full stream to w.
+// This function should not be combined with reading, seeking or other operations.
+// Up to 'concurrent' goroutines will be used.
+// If <= 0, runtime.NumCPU will be used.
+// On success the number of bytes decompressed nil and is returned.
+// This is mainly intended for bigger streams.
+func (r *Reader) DecodeConcurrent(w io.Writer, concurrent int) (written int64, err error) {
+	if r.i > 0 || r.j > 0 || r.blockStart > 0 {
+		return 0, errors.New("DecodeConcurrent called after ")
+	}
+	if concurrent <= 0 {
+		concurrent = runtime.NumCPU()
+	}
+
+	// Write to output
+	var errMu sync.Mutex
+	var aErr error
+	setErr := func(e error) (ok bool) {
+		errMu.Lock()
+		defer errMu.Unlock()
+		if e == nil {
+			return aErr == nil
+		}
+		if aErr == nil {
+			aErr = e
+		}
+		return false
+	}
+	hasErr := func() (ok bool) {
+		errMu.Lock()
+		v := aErr != nil
+		errMu.Unlock()
+		return v
+	}
+
+	var aWritten int64
+	toRead := make(chan []byte, concurrent)
+	writtenBlocks := make(chan []byte, concurrent)
+	queue := make(chan chan []byte, concurrent)
+	reUse := make(chan chan []byte, concurrent)
+	for i := 0; i < concurrent; i++ {
+		toRead <- make([]byte, 0, r.maxBufSize)
+		writtenBlocks <- make([]byte, 0, r.maxBufSize)
+		reUse <- make(chan []byte, 1)
+	}
+	// Writer
+	var wg sync.WaitGroup
+	wg.Add(1)
+	go func() {
+		defer wg.Done()
+		for toWrite := range queue {
+			entry := <-toWrite
+			reUse <- toWrite
+			if hasErr() {
+				writtenBlocks <- entry
+				continue
+			}
+			n, err := w.Write(entry)
+			want := len(entry)
+			writtenBlocks <- entry
+			if err != nil {
+				setErr(err)
+				continue
+			}
+			if n != want {
+				setErr(io.ErrShortWrite)
+				continue
+			}
+			aWritten += int64(n)
+		}
+	}()
+
+	// Reader
+	defer func() {
+		close(queue)
+		if r.err != nil {
+			err = r.err
+			setErr(r.err)
+		}
+		wg.Wait()
+		if err == nil {
+			err = aErr
+		}
+		written = aWritten
+	}()
+
+	for !hasErr() {
+		if !r.readFull(r.buf[:4], true) {
+			if r.err == io.EOF {
+				r.err = nil
+			}
+			return 0, r.err
+		}
+		chunkType := r.buf[0]
+		if !r.readHeader {
+			if chunkType != chunkTypeStreamIdentifier {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			r.readHeader = true
+		}
+		chunkLen := int(r.buf[1]) | int(r.buf[2])<<8 | int(r.buf[3])<<16
+
+		// The chunk types are specified at
+		// https://github.com/google/snappy/blob/master/framing_format.txt
+		switch chunkType {
+		case chunkTypeCompressedData:
+			r.blockStart += int64(r.j)
+			// Section 4.2. Compressed data (chunk type 0x00).
+			if chunkLen < checksumSize {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			if chunkLen > r.maxBufSize {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			orgBuf := <-toRead
+			buf := orgBuf[:chunkLen]
+
+			if !r.readFull(buf, false) {
+				return 0, r.err
+			}
+
+			checksum := uint32(buf[0]) | uint32(buf[1])<<8 | uint32(buf[2])<<16 | uint32(buf[3])<<24
+			buf = buf[checksumSize:]
+
+			n, err := DecodedLen(buf)
+			if err != nil {
+				r.err = err
+				return 0, r.err
+			}
+			if r.snappyFrame && n > maxSnappyBlockSize {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+
+			if n > r.maxBlock {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			wg.Add(1)
+
+			decoded := <-writtenBlocks
+			entry := <-reUse
+			queue <- entry
+			go func() {
+				defer wg.Done()
+				decoded = decoded[:n]
+				_, err := Decode(decoded, buf)
+				toRead <- orgBuf
+				if err != nil {
+					writtenBlocks <- decoded
+					setErr(err)
+					return
+				}
+				if !r.ignoreCRC && crc(decoded) != checksum {
+					writtenBlocks <- decoded
+					setErr(ErrCRC)
+					return
+				}
+				entry <- decoded
+			}()
+			continue
+
+		case chunkTypeUncompressedData:
+
+			// Section 4.3. Uncompressed data (chunk type 0x01).
+			if chunkLen < checksumSize {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			if chunkLen > r.maxBufSize {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			// Grab write buffer
+			orgBuf := <-writtenBlocks
+			buf := orgBuf[:checksumSize]
+			if !r.readFull(buf, false) {
+				return 0, r.err
+			}
+			checksum := uint32(buf[0]) | uint32(buf[1])<<8 | uint32(buf[2])<<16 | uint32(buf[3])<<24
+			// Read content.
+			n := chunkLen - checksumSize
+
+			if r.snappyFrame && n > maxSnappyBlockSize {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			if n > r.maxBlock {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			// Read uncompressed
+			buf = orgBuf[:n]
+			if !r.readFull(buf, false) {
+				return 0, r.err
+			}
+
+			if !r.ignoreCRC && crc(buf) != checksum {
+				r.err = ErrCRC
+				return 0, r.err
+			}
+			entry := <-reUse
+			queue <- entry
+			entry <- buf
+			continue
+
+		case chunkTypeStreamIdentifier:
+			// Section 4.1. Stream identifier (chunk type 0xff).
+			if chunkLen != len(magicBody) {
+				r.err = ErrCorrupt
+				return 0, r.err
+			}
+			if !r.readFull(r.buf[:len(magicBody)], false) {
+				return 0, r.err
+			}
+			if string(r.buf[:len(magicBody)]) != magicBody {
+				if string(r.buf[:len(magicBody)]) != magicBodySnappy {
+					r.err = ErrCorrupt
+					return 0, r.err
+				} else {
+					r.snappyFrame = true
+				}
+			} else {
+				r.snappyFrame = false
+			}
+			continue
+		}
+
+		if chunkType <= 0x7f {
+			// Section 4.5. Reserved unskippable chunks (chunk types 0x02-0x7f).
+			// fmt.Printf("ERR chunktype: 0x%x\n", chunkType)
+			r.err = ErrUnsupported
+			return 0, r.err
+		}
+		// Section 4.4 Padding (chunk type 0xfe).
+		// Section 4.6. Reserved skippable chunks (chunk types 0x80-0xfd).
+		if chunkLen > maxChunkSize {
+			// fmt.Printf("ERR chunkLen: 0x%x\n", chunkLen)
+			r.err = ErrUnsupported
+			return 0, r.err
+		}
+
+		// fmt.Printf("skippable: ID: 0x%x, len: 0x%x\n", chunkType, chunkLen)
+		if !r.skippable(r.buf, chunkLen, false, chunkType) {
+			return 0, r.err
+		}
+	}
+	return 0, r.err
+}
+
 // Skip will skip n bytes forward in the decompressed output.
 // For larger skips this consumes less CPU and is faster than reading output and discarding it.
 // CRC is not checked on skipped blocks.
@@ -454,7 +720,11 @@ func (r *Reader) Skip(n int64) error {
 			// decoded[i:j] contains decoded bytes that have not yet been passed on.
 			left := int64(r.j - r.i)
 			if left >= n {
-				r.i += int(n)
+				tmp := int64(r.i) + n
+				if tmp > math.MaxInt32 {
+					return errors.New("s2: internal overflow in skip")
+				}
+				r.i = int(tmp)
 				return nil
 			}
 			n -= int64(r.j - r.i)
@@ -526,6 +796,7 @@ func (r *Reader) Skip(n int64) error {
 			} else {
 				// Skip block completely
 				n -= int64(dLen)
+				r.blockStart += int64(dLen)
 				dLen = 0
 			}
 			r.i, r.j = 0, dLen
@@ -656,6 +927,15 @@ func (r *Reader) ReadSeeker(random bool, index []byte) (*ReadSeeker, error) {
 	err = r.index.LoadStream(rs)
 	if err != nil {
 		if err == ErrUnsupported {
+			// If we don't require random seeking, reset input and return.
+			if !random {
+				_, err = rs.Seek(pos, io.SeekStart)
+				if err != nil {
+					return nil, ErrCantSeek{Reason: "resetting stream returned: " + err.Error()}
+				}
+				r.index = nil
+				return &ReadSeeker{Reader: r}, nil
+			}
 			return nil, ErrCantSeek{Reason: "input stream does not contain an index"}
 		}
 		return nil, ErrCantSeek{Reason: "reading index returned: " + err.Error()}
@@ -699,8 +979,16 @@ func (r *ReadSeeker) Seek(offset int64, whence int) (int64, error) {
 	case io.SeekCurrent:
 		offset += r.blockStart + int64(r.i)
 	case io.SeekEnd:
-		offset = -offset
+		if offset > 0 {
+			return 0, errors.New("seek after end of file")
+		}
+		offset = r.index.TotalUncompressed + offset
+	}
+
+	if offset < 0 {
+		return 0, errors.New("seek before start of file")
 	}
+
 	c, u, err := r.index.Find(offset)
 	if err != nil {
 		return r.blockStart + int64(r.i), err
@@ -712,10 +1000,6 @@ func (r *ReadSeeker) Seek(offset int64, whence int) (int64, error) {
 		return 0, err
 	}
 
-	if offset < 0 {
-		offset = r.index.TotalUncompressed + offset
-	}
-
 	r.i = r.j // Remove rest of current block.
 	if u < offset {
 		// Forward inside block
diff --git a/vendor/github.com/klauspost/compress/s2/encode.go b/vendor/github.com/klauspost/compress/s2/encode.go
index 59f992ca6..1aefabf31 100644
--- a/vendor/github.com/klauspost/compress/s2/encode.go
+++ b/vendor/github.com/klauspost/compress/s2/encode.go
@@ -1119,12 +1119,6 @@ func (w *Writer) closeIndex(idx bool) ([]byte, error) {
 			if w.appendIndex {
 				w.written += int64(len(index))
 			}
-			if true {
-				_, err := w.index.Load(index)
-				if err != nil {
-					panic(err)
-				}
-			}
 		}
 
 		if w.pad > 1 {
diff --git a/vendor/github.com/klauspost/compress/s2/encode_best.go b/vendor/github.com/klauspost/compress/s2/encode_best.go
index 448034776..4bc80bc6a 100644
--- a/vendor/github.com/klauspost/compress/s2/encode_best.go
+++ b/vendor/github.com/klauspost/compress/s2/encode_best.go
@@ -370,7 +370,7 @@ func encodeBlockBestSnappy(dst, src []byte) (d int) {
 				}
 				offset := m.s - m.offset
 
-				return score - emitCopySize(offset, m.length)
+				return score - emitCopyNoRepeatSize(offset, m.length)
 			}
 
 			matchAt := func(offset, s int, first uint32) match {
@@ -567,6 +567,10 @@ func emitCopySize(offset, length int) int {
 
 	// Offset no more than 2 bytes.
 	if length > 64 {
+		if offset < 2048 {
+			// Emit 8 bytes, then rest as repeats...
+			return 2 + emitRepeatSize(offset, length-8)
+		}
 		// Emit remaining as repeats, at least 4 bytes remain.
 		return 3 + emitRepeatSize(offset, length-60)
 	}
@@ -577,6 +581,28 @@ func emitCopySize(offset, length int) int {
 	return 2
 }
 
+// emitCopyNoRepeatSize returns the size to encode the offset+length
+//
+// It assumes that:
+//	1 <= offset && offset <= math.MaxUint32
+//	4 <= length && length <= 1 << 24
+func emitCopyNoRepeatSize(offset, length int) int {
+	if offset >= 65536 {
+		return 5 + 5*(length/64)
+	}
+
+	// Offset no more than 2 bytes.
+	if length > 64 {
+		// Emit remaining as repeats, at least 4 bytes remain.
+		return 3 + 3*(length/60)
+	}
+	if length >= 12 || offset >= 2048 {
+		return 3
+	}
+	// Emit the remaining copy, encoded as 2 bytes.
+	return 2
+}
+
 // emitRepeatSize returns the number of bytes required to encode a repeat.
 // Length must be at least 4 and < 1<<24
 func emitRepeatSize(offset, length int) int {
diff --git a/vendor/github.com/klauspost/compress/s2/encode_go.go b/vendor/github.com/klauspost/compress/s2/encode_go.go
index 43d43534e..94784b82a 100644
--- a/vendor/github.com/klauspost/compress/s2/encode_go.go
+++ b/vendor/github.com/klauspost/compress/s2/encode_go.go
@@ -180,14 +180,23 @@ func emitCopy(dst []byte, offset, length int) int {
 
 	// Offset no more than 2 bytes.
 	if length > 64 {
-		// Emit a length 60 copy, encoded as 3 bytes.
-		// Emit remaining as repeat value (minimum 4 bytes).
-		dst[2] = uint8(offset >> 8)
-		dst[1] = uint8(offset)
-		dst[0] = 59<<2 | tagCopy2
-		length -= 60
+		off := 3
+		if offset < 2048 {
+			// emit 8 bytes as tagCopy1, rest as repeats.
+			dst[1] = uint8(offset)
+			dst[0] = uint8(offset>>8)<<5 | uint8(8-4)<<2 | tagCopy1
+			length -= 8
+			off = 2
+		} else {
+			// Emit a length 60 copy, encoded as 3 bytes.
+			// Emit remaining as repeat value (minimum 4 bytes).
+			dst[2] = uint8(offset >> 8)
+			dst[1] = uint8(offset)
+			dst[0] = 59<<2 | tagCopy2
+			length -= 60
+		}
 		// Emit remaining as repeats, at least 4 bytes remain.
-		return 3 + emitRepeat(dst[3:], offset, length)
+		return off + emitRepeat(dst[off:], offset, length)
 	}
 	if length >= 12 || offset >= 2048 {
 		// Emit the remaining copy, encoded as 3 bytes.
diff --git a/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.go b/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.go
index d9312e5b9..88f27c099 100644
--- a/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.go
+++ b/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.go
@@ -5,6 +5,8 @@
 
 package s2
 
+func _dummy_()
+
 // encodeBlockAsm encodes a non-empty src to a guaranteed-large-enough dst.
 // Maximum input 4294967295 bytes.
 // It assumes that the varint-encoded length of the decompressed bytes has already been written.
diff --git a/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.s b/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.s
index 729dbf536..36915d949 100644
--- a/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.s
+++ b/vendor/github.com/klauspost/compress/s2/encodeblock_amd64.s
@@ -5,6 +5,15 @@
 
 #include "textflag.h"
 
+// func _dummy_()
+TEXT ·_dummy_(SB), $0
+#ifdef GOAMD64_v4
+#ifndef GOAMD64_v3
+#define GOAMD64_v3
+#endif
+#endif
+	RET
+
 // func encodeBlockAsm(dst []byte, src []byte) int
 // Requires: BMI, SSE2
 TEXT ·encodeBlockAsm(SB), $65560-56
@@ -253,17 +262,6 @@ matchlen_loopback_repeat_extend_encodeBlockAsm:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -476,6 +474,90 @@ four_bytes_remain_repeat_as_copy_encodeBlockAsm:
 two_byte_offset_repeat_as_copy_encodeBlockAsm:
 	CMPL SI, $0x40
 	JLE  two_byte_offset_short_repeat_as_copy_encodeBlockAsm
+	CMPL DI, $0x00000800
+	JAE  long_offset_short_repeat_as_copy_encodeBlockAsm
+	MOVL $0x00000001, R8
+	LEAL 16(R8), R8
+	MOVB DI, 1(AX)
+	MOVL DI, R9
+	SHRL $0x08, R9
+	SHLL $0x05, R9
+	ORL  R9, R8
+	MOVB R8, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, SI
+
+	// emitRepeat
+	LEAL -4(SI), SI
+	JMP  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+
+emit_repeat_again_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b:
+	MOVL SI, R8
+	LEAL -4(SI), SI
+	CMPL R8, $0x08
+	JLE  repeat_two_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+	CMPL R8, $0x0c
+	JGE  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+	CMPL DI, $0x00000800
+	JLT  repeat_two_offset_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+
+cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b:
+	CMPL SI, $0x00000104
+	JLT  repeat_three_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+	CMPL SI, $0x00010100
+	JLT  repeat_four_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+	CMPL SI, $0x0100ffff
+	JLT  repeat_five_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+	LEAL -16842747(SI), SI
+	MOVW $0x001d, (AX)
+	MOVW $0xfffb, 2(AX)
+	MOVB $0xff, 4(AX)
+	ADDQ $0x05, AX
+	JMP  emit_repeat_again_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b
+
+repeat_five_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b:
+	LEAL -65536(SI), SI
+	MOVL SI, DI
+	MOVW $0x001d, (AX)
+	MOVW SI, 2(AX)
+	SARL $0x10, DI
+	MOVB DI, 4(AX)
+	ADDQ $0x05, AX
+	JMP  repeat_end_emit_encodeBlockAsm
+
+repeat_four_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b:
+	LEAL -256(SI), SI
+	MOVW $0x0019, (AX)
+	MOVW SI, 2(AX)
+	ADDQ $0x04, AX
+	JMP  repeat_end_emit_encodeBlockAsm
+
+repeat_three_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b:
+	LEAL -4(SI), SI
+	MOVW $0x0015, (AX)
+	MOVB SI, 2(AX)
+	ADDQ $0x03, AX
+	JMP  repeat_end_emit_encodeBlockAsm
+
+repeat_two_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b:
+	SHLL $0x02, SI
+	ORL  $0x01, SI
+	MOVW SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm
+
+repeat_two_offset_repeat_as_copy_encodeBlockAsm_emit_copy_short_2b:
+	XORQ R8, R8
+	LEAL 1(R8)(SI*4), SI
+	MOVB DI, 1(AX)
+	SARL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm
+
+long_offset_short_repeat_as_copy_encodeBlockAsm:
 	MOVB $0xee, (AX)
 	MOVW DI, 1(AX)
 	LEAL -60(SI), SI
@@ -791,17 +873,6 @@ matchlen_loopback_match_nolit_encodeBlockAsm:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -944,6 +1015,90 @@ four_bytes_remain_match_nolit_encodeBlockAsm:
 two_byte_offset_match_nolit_encodeBlockAsm:
 	CMPL R10, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBlockAsm
+	CMPL SI, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBlockAsm
+	MOVL $0x00000001, DI
+	LEAL 16(DI), DI
+	MOVB SI, 1(AX)
+	MOVL SI, R8
+	SHRL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, DI
+	MOVB DI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R10
+
+	// emitRepeat
+	LEAL -4(R10), R10
+	JMP  cant_repeat_two_offset_match_nolit_encodeBlockAsm_emit_copy_short_2b
+
+emit_repeat_again_match_nolit_encodeBlockAsm_emit_copy_short_2b:
+	MOVL R10, DI
+	LEAL -4(R10), R10
+	CMPL DI, $0x08
+	JLE  repeat_two_match_nolit_encodeBlockAsm_emit_copy_short_2b
+	CMPL DI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBlockAsm_emit_copy_short_2b
+	CMPL SI, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBlockAsm_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBlockAsm_emit_copy_short_2b:
+	CMPL R10, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBlockAsm_emit_copy_short_2b
+	CMPL R10, $0x00010100
+	JLT  repeat_four_match_nolit_encodeBlockAsm_emit_copy_short_2b
+	CMPL R10, $0x0100ffff
+	JLT  repeat_five_match_nolit_encodeBlockAsm_emit_copy_short_2b
+	LEAL -16842747(R10), R10
+	MOVW $0x001d, (AX)
+	MOVW $0xfffb, 2(AX)
+	MOVB $0xff, 4(AX)
+	ADDQ $0x05, AX
+	JMP  emit_repeat_again_match_nolit_encodeBlockAsm_emit_copy_short_2b
+
+repeat_five_match_nolit_encodeBlockAsm_emit_copy_short_2b:
+	LEAL -65536(R10), R10
+	MOVL R10, SI
+	MOVW $0x001d, (AX)
+	MOVW R10, 2(AX)
+	SARL $0x10, SI
+	MOVB SI, 4(AX)
+	ADDQ $0x05, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm
+
+repeat_four_match_nolit_encodeBlockAsm_emit_copy_short_2b:
+	LEAL -256(R10), R10
+	MOVW $0x0019, (AX)
+	MOVW R10, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm
+
+repeat_three_match_nolit_encodeBlockAsm_emit_copy_short_2b:
+	LEAL -4(R10), R10
+	MOVW $0x0015, (AX)
+	MOVB R10, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm
+
+repeat_two_match_nolit_encodeBlockAsm_emit_copy_short_2b:
+	SHLL $0x02, R10
+	ORL  $0x01, R10
+	MOVW R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm
+
+repeat_two_offset_match_nolit_encodeBlockAsm_emit_copy_short_2b:
+	XORQ DI, DI
+	LEAL 1(DI)(R10*4), R10
+	MOVB SI, 1(AX)
+	SARL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, R10
+	MOVB R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm
+
+long_offset_short_match_nolit_encodeBlockAsm:
 	MOVB $0xee, (AX)
 	MOVW SI, 1(AX)
 	LEAL -60(R10), R10
@@ -1134,17 +1289,36 @@ memmove_emit_remainder_encodeBlockAsm:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm
 
 emit_lit_memmove_emit_remainder_encodeBlockAsm_memmove_move_8through16:
@@ -1466,17 +1640,6 @@ matchlen_loopback_repeat_extend_encodeBlockAsm4MB:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -1667,6 +1830,77 @@ four_bytes_remain_repeat_as_copy_encodeBlockAsm4MB:
 two_byte_offset_repeat_as_copy_encodeBlockAsm4MB:
 	CMPL SI, $0x40
 	JLE  two_byte_offset_short_repeat_as_copy_encodeBlockAsm4MB
+	CMPL DI, $0x00000800
+	JAE  long_offset_short_repeat_as_copy_encodeBlockAsm4MB
+	MOVL $0x00000001, R8
+	LEAL 16(R8), R8
+	MOVB DI, 1(AX)
+	SHRL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, R8
+	MOVB R8, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, SI
+
+	// emitRepeat
+	LEAL -4(SI), SI
+	JMP  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b
+	MOVL SI, R8
+	LEAL -4(SI), SI
+	CMPL R8, $0x08
+	JLE  repeat_two_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b
+	CMPL R8, $0x0c
+	JGE  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b
+	CMPL DI, $0x00000800
+	JLT  repeat_two_offset_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b
+
+cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b:
+	CMPL SI, $0x00000104
+	JLT  repeat_three_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b
+	CMPL SI, $0x00010100
+	JLT  repeat_four_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b
+	LEAL -65536(SI), SI
+	MOVL SI, DI
+	MOVW $0x001d, (AX)
+	MOVW SI, 2(AX)
+	SARL $0x10, DI
+	MOVB DI, 4(AX)
+	ADDQ $0x05, AX
+	JMP  repeat_end_emit_encodeBlockAsm4MB
+
+repeat_four_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b:
+	LEAL -256(SI), SI
+	MOVW $0x0019, (AX)
+	MOVW SI, 2(AX)
+	ADDQ $0x04, AX
+	JMP  repeat_end_emit_encodeBlockAsm4MB
+
+repeat_three_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b:
+	LEAL -4(SI), SI
+	MOVW $0x0015, (AX)
+	MOVB SI, 2(AX)
+	ADDQ $0x03, AX
+	JMP  repeat_end_emit_encodeBlockAsm4MB
+
+repeat_two_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b:
+	SHLL $0x02, SI
+	ORL  $0x01, SI
+	MOVW SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm4MB
+
+repeat_two_offset_repeat_as_copy_encodeBlockAsm4MB_emit_copy_short_2b:
+	XORQ R8, R8
+	LEAL 1(R8)(SI*4), SI
+	MOVB DI, 1(AX)
+	SARL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm4MB
+
+long_offset_short_repeat_as_copy_encodeBlockAsm4MB:
 	MOVB $0xee, (AX)
 	MOVW DI, 1(AX)
 	LEAL -60(SI), SI
@@ -1963,17 +2197,6 @@ matchlen_loopback_match_nolit_encodeBlockAsm4MB:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -2105,6 +2328,77 @@ four_bytes_remain_match_nolit_encodeBlockAsm4MB:
 two_byte_offset_match_nolit_encodeBlockAsm4MB:
 	CMPL R10, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBlockAsm4MB
+	CMPL SI, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBlockAsm4MB
+	MOVL $0x00000001, DI
+	LEAL 16(DI), DI
+	MOVB SI, 1(AX)
+	SHRL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, DI
+	MOVB DI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R10
+
+	// emitRepeat
+	LEAL -4(R10), R10
+	JMP  cant_repeat_two_offset_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b
+	MOVL R10, DI
+	LEAL -4(R10), R10
+	CMPL DI, $0x08
+	JLE  repeat_two_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b
+	CMPL DI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b
+	CMPL SI, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b:
+	CMPL R10, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b
+	CMPL R10, $0x00010100
+	JLT  repeat_four_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b
+	LEAL -65536(R10), R10
+	MOVL R10, SI
+	MOVW $0x001d, (AX)
+	MOVW R10, 2(AX)
+	SARL $0x10, SI
+	MOVB SI, 4(AX)
+	ADDQ $0x05, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm4MB
+
+repeat_four_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b:
+	LEAL -256(R10), R10
+	MOVW $0x0019, (AX)
+	MOVW R10, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm4MB
+
+repeat_three_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b:
+	LEAL -4(R10), R10
+	MOVW $0x0015, (AX)
+	MOVB R10, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm4MB
+
+repeat_two_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b:
+	SHLL $0x02, R10
+	ORL  $0x01, R10
+	MOVW R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm4MB
+
+repeat_two_offset_match_nolit_encodeBlockAsm4MB_emit_copy_short_2b:
+	XORQ DI, DI
+	LEAL 1(DI)(R10*4), R10
+	MOVB SI, 1(AX)
+	SARL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, R10
+	MOVB R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm4MB
+
+long_offset_short_match_nolit_encodeBlockAsm4MB:
 	MOVB $0xee, (AX)
 	MOVW SI, 1(AX)
 	LEAL -60(R10), R10
@@ -2276,17 +2570,36 @@ memmove_emit_remainder_encodeBlockAsm4MB:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm4MB
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm4MB
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm4MB
 
 emit_lit_memmove_emit_remainder_encodeBlockAsm4MB_memmove_move_8through16:
@@ -2597,17 +2910,6 @@ matchlen_loopback_repeat_extend_encodeBlockAsm12B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -2706,6 +3008,65 @@ repeat_as_copy_encodeBlockAsm12B:
 two_byte_offset_repeat_as_copy_encodeBlockAsm12B:
 	CMPL SI, $0x40
 	JLE  two_byte_offset_short_repeat_as_copy_encodeBlockAsm12B
+	CMPL DI, $0x00000800
+	JAE  long_offset_short_repeat_as_copy_encodeBlockAsm12B
+	MOVL $0x00000001, R8
+	LEAL 16(R8), R8
+	MOVB DI, 1(AX)
+	SHRL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, R8
+	MOVB R8, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, SI
+
+	// emitRepeat
+	LEAL -4(SI), SI
+	JMP  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b
+	MOVL SI, R8
+	LEAL -4(SI), SI
+	CMPL R8, $0x08
+	JLE  repeat_two_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b
+	CMPL R8, $0x0c
+	JGE  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b
+	CMPL DI, $0x00000800
+	JLT  repeat_two_offset_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b
+
+cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b:
+	CMPL SI, $0x00000104
+	JLT  repeat_three_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b
+	LEAL -256(SI), SI
+	MOVW $0x0019, (AX)
+	MOVW SI, 2(AX)
+	ADDQ $0x04, AX
+	JMP  repeat_end_emit_encodeBlockAsm12B
+
+repeat_three_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b:
+	LEAL -4(SI), SI
+	MOVW $0x0015, (AX)
+	MOVB SI, 2(AX)
+	ADDQ $0x03, AX
+	JMP  repeat_end_emit_encodeBlockAsm12B
+
+repeat_two_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b:
+	SHLL $0x02, SI
+	ORL  $0x01, SI
+	MOVW SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm12B
+
+repeat_two_offset_repeat_as_copy_encodeBlockAsm12B_emit_copy_short_2b:
+	XORQ R8, R8
+	LEAL 1(R8)(SI*4), SI
+	MOVB DI, 1(AX)
+	SARL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm12B
+
+long_offset_short_repeat_as_copy_encodeBlockAsm12B:
 	MOVB $0xee, (AX)
 	MOVW DI, 1(AX)
 	LEAL -60(SI), SI
@@ -2979,17 +3340,6 @@ matchlen_loopback_match_nolit_encodeBlockAsm12B:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -3041,6 +3391,65 @@ match_nolit_end_encodeBlockAsm12B:
 two_byte_offset_match_nolit_encodeBlockAsm12B:
 	CMPL R10, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBlockAsm12B
+	CMPL SI, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBlockAsm12B
+	MOVL $0x00000001, DI
+	LEAL 16(DI), DI
+	MOVB SI, 1(AX)
+	SHRL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, DI
+	MOVB DI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R10
+
+	// emitRepeat
+	LEAL -4(R10), R10
+	JMP  cant_repeat_two_offset_match_nolit_encodeBlockAsm12B_emit_copy_short_2b
+	MOVL R10, DI
+	LEAL -4(R10), R10
+	CMPL DI, $0x08
+	JLE  repeat_two_match_nolit_encodeBlockAsm12B_emit_copy_short_2b
+	CMPL DI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBlockAsm12B_emit_copy_short_2b
+	CMPL SI, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBlockAsm12B_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBlockAsm12B_emit_copy_short_2b:
+	CMPL R10, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBlockAsm12B_emit_copy_short_2b
+	LEAL -256(R10), R10
+	MOVW $0x0019, (AX)
+	MOVW R10, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm12B
+
+repeat_three_match_nolit_encodeBlockAsm12B_emit_copy_short_2b:
+	LEAL -4(R10), R10
+	MOVW $0x0015, (AX)
+	MOVB R10, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm12B
+
+repeat_two_match_nolit_encodeBlockAsm12B_emit_copy_short_2b:
+	SHLL $0x02, R10
+	ORL  $0x01, R10
+	MOVW R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm12B
+
+repeat_two_offset_match_nolit_encodeBlockAsm12B_emit_copy_short_2b:
+	XORQ DI, DI
+	LEAL 1(DI)(R10*4), R10
+	MOVB SI, 1(AX)
+	SARL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, R10
+	MOVB R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm12B
+
+long_offset_short_match_nolit_encodeBlockAsm12B:
 	MOVB $0xee, (AX)
 	MOVW SI, 1(AX)
 	LEAL -60(R10), R10
@@ -3189,17 +3598,36 @@ memmove_emit_remainder_encodeBlockAsm12B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm12B
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm12B
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm12B
 
 emit_lit_memmove_emit_remainder_encodeBlockAsm12B_memmove_move_8through16:
@@ -3510,17 +3938,6 @@ matchlen_loopback_repeat_extend_encodeBlockAsm10B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -3619,6 +4036,65 @@ repeat_as_copy_encodeBlockAsm10B:
 two_byte_offset_repeat_as_copy_encodeBlockAsm10B:
 	CMPL SI, $0x40
 	JLE  two_byte_offset_short_repeat_as_copy_encodeBlockAsm10B
+	CMPL DI, $0x00000800
+	JAE  long_offset_short_repeat_as_copy_encodeBlockAsm10B
+	MOVL $0x00000001, R8
+	LEAL 16(R8), R8
+	MOVB DI, 1(AX)
+	SHRL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, R8
+	MOVB R8, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, SI
+
+	// emitRepeat
+	LEAL -4(SI), SI
+	JMP  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b
+	MOVL SI, R8
+	LEAL -4(SI), SI
+	CMPL R8, $0x08
+	JLE  repeat_two_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b
+	CMPL R8, $0x0c
+	JGE  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b
+	CMPL DI, $0x00000800
+	JLT  repeat_two_offset_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b
+
+cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b:
+	CMPL SI, $0x00000104
+	JLT  repeat_three_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b
+	LEAL -256(SI), SI
+	MOVW $0x0019, (AX)
+	MOVW SI, 2(AX)
+	ADDQ $0x04, AX
+	JMP  repeat_end_emit_encodeBlockAsm10B
+
+repeat_three_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b:
+	LEAL -4(SI), SI
+	MOVW $0x0015, (AX)
+	MOVB SI, 2(AX)
+	ADDQ $0x03, AX
+	JMP  repeat_end_emit_encodeBlockAsm10B
+
+repeat_two_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b:
+	SHLL $0x02, SI
+	ORL  $0x01, SI
+	MOVW SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm10B
+
+repeat_two_offset_repeat_as_copy_encodeBlockAsm10B_emit_copy_short_2b:
+	XORQ R8, R8
+	LEAL 1(R8)(SI*4), SI
+	MOVB DI, 1(AX)
+	SARL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm10B
+
+long_offset_short_repeat_as_copy_encodeBlockAsm10B:
 	MOVB $0xee, (AX)
 	MOVW DI, 1(AX)
 	LEAL -60(SI), SI
@@ -3892,17 +4368,6 @@ matchlen_loopback_match_nolit_encodeBlockAsm10B:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -3954,6 +4419,65 @@ match_nolit_end_encodeBlockAsm10B:
 two_byte_offset_match_nolit_encodeBlockAsm10B:
 	CMPL R10, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBlockAsm10B
+	CMPL SI, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBlockAsm10B
+	MOVL $0x00000001, DI
+	LEAL 16(DI), DI
+	MOVB SI, 1(AX)
+	SHRL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, DI
+	MOVB DI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R10
+
+	// emitRepeat
+	LEAL -4(R10), R10
+	JMP  cant_repeat_two_offset_match_nolit_encodeBlockAsm10B_emit_copy_short_2b
+	MOVL R10, DI
+	LEAL -4(R10), R10
+	CMPL DI, $0x08
+	JLE  repeat_two_match_nolit_encodeBlockAsm10B_emit_copy_short_2b
+	CMPL DI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBlockAsm10B_emit_copy_short_2b
+	CMPL SI, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBlockAsm10B_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBlockAsm10B_emit_copy_short_2b:
+	CMPL R10, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBlockAsm10B_emit_copy_short_2b
+	LEAL -256(R10), R10
+	MOVW $0x0019, (AX)
+	MOVW R10, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm10B
+
+repeat_three_match_nolit_encodeBlockAsm10B_emit_copy_short_2b:
+	LEAL -4(R10), R10
+	MOVW $0x0015, (AX)
+	MOVB R10, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm10B
+
+repeat_two_match_nolit_encodeBlockAsm10B_emit_copy_short_2b:
+	SHLL $0x02, R10
+	ORL  $0x01, R10
+	MOVW R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm10B
+
+repeat_two_offset_match_nolit_encodeBlockAsm10B_emit_copy_short_2b:
+	XORQ DI, DI
+	LEAL 1(DI)(R10*4), R10
+	MOVB SI, 1(AX)
+	SARL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, R10
+	MOVB R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm10B
+
+long_offset_short_match_nolit_encodeBlockAsm10B:
 	MOVB $0xee, (AX)
 	MOVW SI, 1(AX)
 	LEAL -60(R10), R10
@@ -4102,17 +4626,36 @@ memmove_emit_remainder_encodeBlockAsm10B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm10B
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm10B
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm10B
 
 emit_lit_memmove_emit_remainder_encodeBlockAsm10B_memmove_move_8through16:
@@ -4423,17 +4966,6 @@ matchlen_loopback_repeat_extend_encodeBlockAsm8B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -4528,6 +5060,61 @@ repeat_as_copy_encodeBlockAsm8B:
 two_byte_offset_repeat_as_copy_encodeBlockAsm8B:
 	CMPL SI, $0x40
 	JLE  two_byte_offset_short_repeat_as_copy_encodeBlockAsm8B
+	CMPL DI, $0x00000800
+	JAE  long_offset_short_repeat_as_copy_encodeBlockAsm8B
+	MOVL $0x00000001, R8
+	LEAL 16(R8), R8
+	MOVB DI, 1(AX)
+	SHRL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, R8
+	MOVB R8, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, SI
+
+	// emitRepeat
+	LEAL -4(SI), SI
+	JMP  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm8B_emit_copy_short_2b
+	MOVL SI, DI
+	LEAL -4(SI), SI
+	CMPL DI, $0x08
+	JLE  repeat_two_repeat_as_copy_encodeBlockAsm8B_emit_copy_short_2b
+	CMPL DI, $0x0c
+	JGE  cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm8B_emit_copy_short_2b
+
+cant_repeat_two_offset_repeat_as_copy_encodeBlockAsm8B_emit_copy_short_2b:
+	CMPL SI, $0x00000104
+	JLT  repeat_three_repeat_as_copy_encodeBlockAsm8B_emit_copy_short_2b
+	LEAL -256(SI), SI
+	MOVW $0x0019, (AX)
+	MOVW SI, 2(AX)
+	ADDQ $0x04, AX
+	JMP  repeat_end_emit_encodeBlockAsm8B
+
+repeat_three_repeat_as_copy_encodeBlockAsm8B_emit_copy_short_2b:
+	LEAL -4(SI), SI
+	MOVW $0x0015, (AX)
+	MOVB SI, 2(AX)
+	ADDQ $0x03, AX
+	JMP  repeat_end_emit_encodeBlockAsm8B
+
+repeat_two_repeat_as_copy_encodeBlockAsm8B_emit_copy_short_2b:
+	SHLL $0x02, SI
+	ORL  $0x01, SI
+	MOVW SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm8B
+	XORQ R8, R8
+	LEAL 1(R8)(SI*4), SI
+	MOVB DI, 1(AX)
+	SARL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	JMP  repeat_end_emit_encodeBlockAsm8B
+
+long_offset_short_repeat_as_copy_encodeBlockAsm8B:
 	MOVB $0xee, (AX)
 	MOVW DI, 1(AX)
 	LEAL -60(SI), SI
@@ -4795,17 +5382,6 @@ matchlen_loopback_match_nolit_encodeBlockAsm8B:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -4857,6 +5433,61 @@ match_nolit_end_encodeBlockAsm8B:
 two_byte_offset_match_nolit_encodeBlockAsm8B:
 	CMPL R10, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBlockAsm8B
+	CMPL SI, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBlockAsm8B
+	MOVL $0x00000001, DI
+	LEAL 16(DI), DI
+	MOVB SI, 1(AX)
+	SHRL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, DI
+	MOVB DI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R10
+
+	// emitRepeat
+	LEAL -4(R10), R10
+	JMP  cant_repeat_two_offset_match_nolit_encodeBlockAsm8B_emit_copy_short_2b
+	MOVL R10, SI
+	LEAL -4(R10), R10
+	CMPL SI, $0x08
+	JLE  repeat_two_match_nolit_encodeBlockAsm8B_emit_copy_short_2b
+	CMPL SI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBlockAsm8B_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBlockAsm8B_emit_copy_short_2b:
+	CMPL R10, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBlockAsm8B_emit_copy_short_2b
+	LEAL -256(R10), R10
+	MOVW $0x0019, (AX)
+	MOVW R10, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm8B
+
+repeat_three_match_nolit_encodeBlockAsm8B_emit_copy_short_2b:
+	LEAL -4(R10), R10
+	MOVW $0x0015, (AX)
+	MOVB R10, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm8B
+
+repeat_two_match_nolit_encodeBlockAsm8B_emit_copy_short_2b:
+	SHLL $0x02, R10
+	ORL  $0x01, R10
+	MOVW R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm8B
+	XORQ DI, DI
+	LEAL 1(DI)(R10*4), R10
+	MOVB SI, 1(AX)
+	SARL $0x08, SI
+	SHLL $0x05, SI
+	ORL  SI, R10
+	MOVB R10, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBlockAsm8B
+
+long_offset_short_match_nolit_encodeBlockAsm8B:
 	MOVB $0xee, (AX)
 	MOVW SI, 1(AX)
 	LEAL -60(R10), R10
@@ -4999,17 +5630,36 @@ memmove_emit_remainder_encodeBlockAsm8B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm8B
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm8B
+
+emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeBlockAsm8B
 
 emit_lit_memmove_emit_remainder_encodeBlockAsm8B_memmove_move_8through16:
@@ -5225,17 +5875,6 @@ matchlen_loopback_match_nolit_encodeBetterBlockAsm:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -5541,6 +6180,90 @@ four_bytes_remain_match_nolit_encodeBetterBlockAsm:
 two_byte_offset_match_nolit_encodeBetterBlockAsm:
 	CMPL R12, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBetterBlockAsm
+	CMPL R8, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBetterBlockAsm
+	MOVL $0x00000001, SI
+	LEAL 16(SI), SI
+	MOVB R8, 1(AX)
+	MOVL R8, R9
+	SHRL $0x08, R9
+	SHLL $0x05, R9
+	ORL  R9, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R12
+
+	// emitRepeat
+	LEAL -4(R12), R12
+	JMP  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+
+emit_repeat_again_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b:
+	MOVL R12, SI
+	LEAL -4(R12), R12
+	CMPL SI, $0x08
+	JLE  repeat_two_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+	CMPL SI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+	CMPL R8, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b:
+	CMPL R12, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+	CMPL R12, $0x00010100
+	JLT  repeat_four_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+	CMPL R12, $0x0100ffff
+	JLT  repeat_five_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+	LEAL -16842747(R12), R12
+	MOVW $0x001d, (AX)
+	MOVW $0xfffb, 2(AX)
+	MOVB $0xff, 4(AX)
+	ADDQ $0x05, AX
+	JMP  emit_repeat_again_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b
+
+repeat_five_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b:
+	LEAL -65536(R12), R12
+	MOVL R12, R8
+	MOVW $0x001d, (AX)
+	MOVW R12, 2(AX)
+	SARL $0x10, R8
+	MOVB R8, 4(AX)
+	ADDQ $0x05, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm
+
+repeat_four_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b:
+	LEAL -256(R12), R12
+	MOVW $0x0019, (AX)
+	MOVW R12, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm
+
+repeat_three_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b:
+	LEAL -4(R12), R12
+	MOVW $0x0015, (AX)
+	MOVB R12, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm
+
+repeat_two_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b:
+	SHLL $0x02, R12
+	ORL  $0x01, R12
+	MOVW R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm
+
+repeat_two_offset_match_nolit_encodeBetterBlockAsm_emit_copy_short_2b:
+	XORQ SI, SI
+	LEAL 1(SI)(R12*4), R12
+	MOVB R8, 1(AX)
+	SARL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, R12
+	MOVB R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm
+
+long_offset_short_match_nolit_encodeBetterBlockAsm:
 	MOVB $0xee, (AX)
 	MOVW R8, 1(AX)
 	LEAL -60(R12), R12
@@ -5979,8 +6702,9 @@ memmove_emit_remainder_encodeBetterBlockAsm:
 	MOVL SI, BX
 
 	// genMemMoveShort
-	CMPQ BX, $0x04
-	JLE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_4
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_3
 	CMPQ BX, $0x08
 	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_4through7
 	CMPQ BX, $0x10
@@ -5989,9 +6713,18 @@ memmove_emit_remainder_encodeBetterBlockAsm:
 	JBE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_4:
-	MOVL (CX), SI
-	MOVL SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm
+
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
 	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm
 
 emit_lit_memmove_emit_remainder_encodeBetterBlockAsm_memmove_move_4through7:
@@ -6214,17 +6947,6 @@ matchlen_loopback_match_nolit_encodeBetterBlockAsm4MB:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -6511,6 +7233,77 @@ four_bytes_remain_match_nolit_encodeBetterBlockAsm4MB:
 two_byte_offset_match_nolit_encodeBetterBlockAsm4MB:
 	CMPL R12, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBetterBlockAsm4MB
+	CMPL R8, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBetterBlockAsm4MB
+	MOVL $0x00000001, SI
+	LEAL 16(SI), SI
+	MOVB R8, 1(AX)
+	SHRL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R12
+
+	// emitRepeat
+	LEAL -4(R12), R12
+	JMP  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b
+	MOVL R12, SI
+	LEAL -4(R12), R12
+	CMPL SI, $0x08
+	JLE  repeat_two_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b
+	CMPL SI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b
+	CMPL R8, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b:
+	CMPL R12, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b
+	CMPL R12, $0x00010100
+	JLT  repeat_four_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b
+	LEAL -65536(R12), R12
+	MOVL R12, R8
+	MOVW $0x001d, (AX)
+	MOVW R12, 2(AX)
+	SARL $0x10, R8
+	MOVB R8, 4(AX)
+	ADDQ $0x05, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm4MB
+
+repeat_four_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b:
+	LEAL -256(R12), R12
+	MOVW $0x0019, (AX)
+	MOVW R12, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm4MB
+
+repeat_three_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b:
+	LEAL -4(R12), R12
+	MOVW $0x0015, (AX)
+	MOVB R12, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm4MB
+
+repeat_two_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b:
+	SHLL $0x02, R12
+	ORL  $0x01, R12
+	MOVW R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm4MB
+
+repeat_two_offset_match_nolit_encodeBetterBlockAsm4MB_emit_copy_short_2b:
+	XORQ SI, SI
+	LEAL 1(SI)(R12*4), R12
+	MOVB R8, 1(AX)
+	SARL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, R12
+	MOVB R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm4MB
+
+long_offset_short_match_nolit_encodeBetterBlockAsm4MB:
 	MOVB $0xee, (AX)
 	MOVW R8, 1(AX)
 	LEAL -60(R12), R12
@@ -6911,8 +7704,9 @@ memmove_emit_remainder_encodeBetterBlockAsm4MB:
 	MOVL SI, BX
 
 	// genMemMoveShort
-	CMPQ BX, $0x04
-	JLE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_4
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_3
 	CMPQ BX, $0x08
 	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_4through7
 	CMPQ BX, $0x10
@@ -6921,9 +7715,18 @@ memmove_emit_remainder_encodeBetterBlockAsm4MB:
 	JBE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_4:
-	MOVL (CX), SI
-	MOVL SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm4MB
+
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
 	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm4MB
 
 emit_lit_memmove_emit_remainder_encodeBetterBlockAsm4MB_memmove_move_4through7:
@@ -7138,17 +7941,6 @@ matchlen_loopback_match_nolit_encodeBetterBlockAsm12B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -7335,6 +8127,65 @@ emit_literal_done_match_emit_encodeBetterBlockAsm12B:
 two_byte_offset_match_nolit_encodeBetterBlockAsm12B:
 	CMPL R12, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBetterBlockAsm12B
+	CMPL R8, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBetterBlockAsm12B
+	MOVL $0x00000001, SI
+	LEAL 16(SI), SI
+	MOVB R8, 1(AX)
+	SHRL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R12
+
+	// emitRepeat
+	LEAL -4(R12), R12
+	JMP  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b
+	MOVL R12, SI
+	LEAL -4(R12), R12
+	CMPL SI, $0x08
+	JLE  repeat_two_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b
+	CMPL SI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b
+	CMPL R8, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b:
+	CMPL R12, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b
+	LEAL -256(R12), R12
+	MOVW $0x0019, (AX)
+	MOVW R12, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm12B
+
+repeat_three_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b:
+	LEAL -4(R12), R12
+	MOVW $0x0015, (AX)
+	MOVB R12, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm12B
+
+repeat_two_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b:
+	SHLL $0x02, R12
+	ORL  $0x01, R12
+	MOVW R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm12B
+
+repeat_two_offset_match_nolit_encodeBetterBlockAsm12B_emit_copy_short_2b:
+	XORQ SI, SI
+	LEAL 1(SI)(R12*4), R12
+	MOVB R8, 1(AX)
+	SARL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, R12
+	MOVB R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm12B
+
+long_offset_short_match_nolit_encodeBetterBlockAsm12B:
 	MOVB $0xee, (AX)
 	MOVW R8, 1(AX)
 	LEAL -60(R12), R12
@@ -7689,8 +8540,9 @@ memmove_emit_remainder_encodeBetterBlockAsm12B:
 	MOVL SI, BX
 
 	// genMemMoveShort
-	CMPQ BX, $0x04
-	JLE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_4
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_3
 	CMPQ BX, $0x08
 	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_4through7
 	CMPQ BX, $0x10
@@ -7699,9 +8551,18 @@ memmove_emit_remainder_encodeBetterBlockAsm12B:
 	JBE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_4:
-	MOVL (CX), SI
-	MOVL SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm12B
+
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
 	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm12B
 
 emit_lit_memmove_emit_remainder_encodeBetterBlockAsm12B_memmove_move_4through7:
@@ -7916,17 +8777,6 @@ matchlen_loopback_match_nolit_encodeBetterBlockAsm10B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -8113,6 +8963,65 @@ emit_literal_done_match_emit_encodeBetterBlockAsm10B:
 two_byte_offset_match_nolit_encodeBetterBlockAsm10B:
 	CMPL R12, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBetterBlockAsm10B
+	CMPL R8, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBetterBlockAsm10B
+	MOVL $0x00000001, SI
+	LEAL 16(SI), SI
+	MOVB R8, 1(AX)
+	SHRL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R12
+
+	// emitRepeat
+	LEAL -4(R12), R12
+	JMP  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b
+	MOVL R12, SI
+	LEAL -4(R12), R12
+	CMPL SI, $0x08
+	JLE  repeat_two_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b
+	CMPL SI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b
+	CMPL R8, $0x00000800
+	JLT  repeat_two_offset_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b:
+	CMPL R12, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b
+	LEAL -256(R12), R12
+	MOVW $0x0019, (AX)
+	MOVW R12, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm10B
+
+repeat_three_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b:
+	LEAL -4(R12), R12
+	MOVW $0x0015, (AX)
+	MOVB R12, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm10B
+
+repeat_two_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b:
+	SHLL $0x02, R12
+	ORL  $0x01, R12
+	MOVW R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm10B
+
+repeat_two_offset_match_nolit_encodeBetterBlockAsm10B_emit_copy_short_2b:
+	XORQ SI, SI
+	LEAL 1(SI)(R12*4), R12
+	MOVB R8, 1(AX)
+	SARL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, R12
+	MOVB R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm10B
+
+long_offset_short_match_nolit_encodeBetterBlockAsm10B:
 	MOVB $0xee, (AX)
 	MOVW R8, 1(AX)
 	LEAL -60(R12), R12
@@ -8467,8 +9376,9 @@ memmove_emit_remainder_encodeBetterBlockAsm10B:
 	MOVL SI, BX
 
 	// genMemMoveShort
-	CMPQ BX, $0x04
-	JLE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_4
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_3
 	CMPQ BX, $0x08
 	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_4through7
 	CMPQ BX, $0x10
@@ -8477,9 +9387,18 @@ memmove_emit_remainder_encodeBetterBlockAsm10B:
 	JBE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_4:
-	MOVL (CX), SI
-	MOVL SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm10B
+
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
 	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm10B
 
 emit_lit_memmove_emit_remainder_encodeBetterBlockAsm10B_memmove_move_4through7:
@@ -8694,17 +9613,6 @@ matchlen_loopback_match_nolit_encodeBetterBlockAsm8B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -8891,6 +9799,61 @@ emit_literal_done_match_emit_encodeBetterBlockAsm8B:
 two_byte_offset_match_nolit_encodeBetterBlockAsm8B:
 	CMPL R12, $0x40
 	JLE  two_byte_offset_short_match_nolit_encodeBetterBlockAsm8B
+	CMPL R8, $0x00000800
+	JAE  long_offset_short_match_nolit_encodeBetterBlockAsm8B
+	MOVL $0x00000001, SI
+	LEAL 16(SI), SI
+	MOVB R8, 1(AX)
+	SHRL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, AX
+	SUBL $0x08, R12
+
+	// emitRepeat
+	LEAL -4(R12), R12
+	JMP  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm8B_emit_copy_short_2b
+	MOVL R12, SI
+	LEAL -4(R12), R12
+	CMPL SI, $0x08
+	JLE  repeat_two_match_nolit_encodeBetterBlockAsm8B_emit_copy_short_2b
+	CMPL SI, $0x0c
+	JGE  cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm8B_emit_copy_short_2b
+
+cant_repeat_two_offset_match_nolit_encodeBetterBlockAsm8B_emit_copy_short_2b:
+	CMPL R12, $0x00000104
+	JLT  repeat_three_match_nolit_encodeBetterBlockAsm8B_emit_copy_short_2b
+	LEAL -256(R12), R12
+	MOVW $0x0019, (AX)
+	MOVW R12, 2(AX)
+	ADDQ $0x04, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm8B
+
+repeat_three_match_nolit_encodeBetterBlockAsm8B_emit_copy_short_2b:
+	LEAL -4(R12), R12
+	MOVW $0x0015, (AX)
+	MOVB R12, 2(AX)
+	ADDQ $0x03, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm8B
+
+repeat_two_match_nolit_encodeBetterBlockAsm8B_emit_copy_short_2b:
+	SHLL $0x02, R12
+	ORL  $0x01, R12
+	MOVW R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm8B
+	XORQ SI, SI
+	LEAL 1(SI)(R12*4), R12
+	MOVB R8, 1(AX)
+	SARL $0x08, R8
+	SHLL $0x05, R8
+	ORL  R8, R12
+	MOVB R12, (AX)
+	ADDQ $0x02, AX
+	JMP  match_nolit_emitcopy_end_encodeBetterBlockAsm8B
+
+long_offset_short_match_nolit_encodeBetterBlockAsm8B:
 	MOVB $0xee, (AX)
 	MOVW R8, 1(AX)
 	LEAL -60(R12), R12
@@ -9235,8 +10198,9 @@ memmove_emit_remainder_encodeBetterBlockAsm8B:
 	MOVL SI, BX
 
 	// genMemMoveShort
-	CMPQ BX, $0x04
-	JLE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_4
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_3
 	CMPQ BX, $0x08
 	JB   emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_4through7
 	CMPQ BX, $0x10
@@ -9245,9 +10209,18 @@ memmove_emit_remainder_encodeBetterBlockAsm8B:
 	JBE  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_4:
-	MOVL (CX), SI
-	MOVL SI, (AX)
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm8B
+
+emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
 	JMP  memmove_end_copy_emit_remainder_encodeBetterBlockAsm8B
 
 emit_lit_memmove_emit_remainder_encodeBetterBlockAsm8B_memmove_move_4through7:
@@ -9584,17 +10557,6 @@ matchlen_loopback_repeat_extend_encodeSnappyBlockAsm:
 #ifdef GOAMD64_v3
 	TZCNTQ R10, R10
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R10, R10
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R10, R10
 
@@ -9918,17 +10880,6 @@ matchlen_loopback_match_nolit_encodeSnappyBlockAsm:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -10127,17 +11078,36 @@ memmove_emit_remainder_encodeSnappyBlockAsm:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm
 
 emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm_memmove_move_8through16:
@@ -10448,17 +11418,6 @@ matchlen_loopback_repeat_extend_encodeSnappyBlockAsm64K:
 #ifdef GOAMD64_v3
 	TZCNTQ R10, R10
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R10, R10
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R10, R10
 
@@ -10739,17 +11698,6 @@ matchlen_loopback_match_nolit_encodeSnappyBlockAsm64K:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -10905,17 +11853,36 @@ memmove_emit_remainder_encodeSnappyBlockAsm64K:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm64K
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm64K
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm64K
 
 emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm64K_memmove_move_8through16:
@@ -11226,17 +12193,6 @@ matchlen_loopback_repeat_extend_encodeSnappyBlockAsm12B:
 #ifdef GOAMD64_v3
 	TZCNTQ R10, R10
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R10, R10
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R10, R10
 
@@ -11517,17 +12473,6 @@ matchlen_loopback_match_nolit_encodeSnappyBlockAsm12B:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -11683,17 +12628,36 @@ memmove_emit_remainder_encodeSnappyBlockAsm12B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm12B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm12B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm12B
 
 emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm12B_memmove_move_8through16:
@@ -12004,17 +12968,6 @@ matchlen_loopback_repeat_extend_encodeSnappyBlockAsm10B:
 #ifdef GOAMD64_v3
 	TZCNTQ R10, R10
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R10, R10
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R10, R10
 
@@ -12295,17 +13248,6 @@ matchlen_loopback_match_nolit_encodeSnappyBlockAsm10B:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -12461,17 +13403,36 @@ memmove_emit_remainder_encodeSnappyBlockAsm10B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm10B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm10B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm10B
 
 emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm10B_memmove_move_8through16:
@@ -12782,17 +13743,6 @@ matchlen_loopback_repeat_extend_encodeSnappyBlockAsm8B:
 #ifdef GOAMD64_v3
 	TZCNTQ R10, R10
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R10, R10
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R10, R10
 
@@ -13071,17 +14021,6 @@ matchlen_loopback_match_nolit_encodeSnappyBlockAsm8B:
 #ifdef GOAMD64_v3
 	TZCNTQ R9, R9
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R9, R9
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R9, R9
 
@@ -13235,17 +14174,36 @@ memmove_emit_remainder_encodeSnappyBlockAsm8B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm8B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm8B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBlockAsm8B
 
 emit_lit_memmove_emit_remainder_encodeSnappyBlockAsm8B_memmove_move_8through16:
@@ -13461,17 +14419,6 @@ matchlen_loopback_match_nolit_encodeSnappyBetterBlockAsm:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -13850,17 +14797,36 @@ memmove_emit_remainder_encodeSnappyBetterBlockAsm:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm
 
 emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm_memmove_move_8through16:
@@ -14068,17 +15034,6 @@ matchlen_loopback_match_nolit_encodeSnappyBetterBlockAsm64K:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -14386,17 +15341,36 @@ memmove_emit_remainder_encodeSnappyBetterBlockAsm64K:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm64K
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm64K
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm64K
 
 emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm64K_memmove_move_8through16:
@@ -14604,17 +15578,6 @@ matchlen_loopback_match_nolit_encodeSnappyBetterBlockAsm12B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -14922,17 +15885,36 @@ memmove_emit_remainder_encodeSnappyBetterBlockAsm12B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm12B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm12B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm12B
 
 emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm12B_memmove_move_8through16:
@@ -15140,17 +16122,6 @@ matchlen_loopback_match_nolit_encodeSnappyBetterBlockAsm10B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -15458,17 +16429,36 @@ memmove_emit_remainder_encodeSnappyBetterBlockAsm10B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm10B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm10B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm10B
 
 emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm10B_memmove_move_8through16:
@@ -15676,17 +16666,6 @@ matchlen_loopback_match_nolit_encodeSnappyBetterBlockAsm8B:
 #ifdef GOAMD64_v3
 	TZCNTQ R11, R11
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ R11, R11
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ R11, R11
 
@@ -15992,17 +16971,36 @@ memmove_emit_remainder_encodeSnappyBetterBlockAsm8B:
 	MOVL SI, BX
 
 	// genMemMoveShort
+	CMPQ BX, $0x03
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_1or2
+	JE   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_3
 	CMPQ BX, $0x08
-	JLE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_8
+	JB   emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_4through7
 	CMPQ BX, $0x10
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_8through16
 	CMPQ BX, $0x20
 	JBE  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_17through32
 	JMP  emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_33through64
 
-emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_8:
-	MOVQ (CX), SI
-	MOVQ SI, (AX)
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_1or2:
+	MOVB (CX), SI
+	MOVB -1(CX)(BX*1), CL
+	MOVB SI, (AX)
+	MOVB CL, -1(AX)(BX*1)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm8B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_3:
+	MOVW (CX), SI
+	MOVB 2(CX), CL
+	MOVW SI, (AX)
+	MOVB CL, 2(AX)
+	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm8B
+
+emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_4through7:
+	MOVL (CX), SI
+	MOVL -4(CX)(BX*1), CX
+	MOVL SI, (AX)
+	MOVL CX, -4(AX)(BX*1)
 	JMP  memmove_end_copy_emit_remainder_encodeSnappyBetterBlockAsm8B
 
 emit_lit_memmove_emit_remainder_encodeSnappyBetterBlockAsm8B_memmove_move_8through16:
@@ -16443,6 +17441,97 @@ four_bytes_remain_standalone:
 two_byte_offset_standalone:
 	CMPL DX, $0x40
 	JLE  two_byte_offset_short_standalone
+	CMPL CX, $0x00000800
+	JAE  long_offset_short_standalone
+	MOVL $0x00000001, SI
+	LEAL 16(SI), SI
+	MOVB CL, 1(AX)
+	MOVL CX, DI
+	SHRL $0x08, DI
+	SHLL $0x05, DI
+	ORL  DI, SI
+	MOVB SI, (AX)
+	ADDQ $0x02, BX
+	ADDQ $0x02, AX
+	SUBL $0x08, DX
+
+	// emitRepeat
+	LEAL -4(DX), DX
+	JMP  cant_repeat_two_offset_standalone_emit_copy_short_2b
+
+emit_repeat_again_standalone_emit_copy_short_2b:
+	MOVL DX, SI
+	LEAL -4(DX), DX
+	CMPL SI, $0x08
+	JLE  repeat_two_standalone_emit_copy_short_2b
+	CMPL SI, $0x0c
+	JGE  cant_repeat_two_offset_standalone_emit_copy_short_2b
+	CMPL CX, $0x00000800
+	JLT  repeat_two_offset_standalone_emit_copy_short_2b
+
+cant_repeat_two_offset_standalone_emit_copy_short_2b:
+	CMPL DX, $0x00000104
+	JLT  repeat_three_standalone_emit_copy_short_2b
+	CMPL DX, $0x00010100
+	JLT  repeat_four_standalone_emit_copy_short_2b
+	CMPL DX, $0x0100ffff
+	JLT  repeat_five_standalone_emit_copy_short_2b
+	LEAL -16842747(DX), DX
+	MOVW $0x001d, (AX)
+	MOVW $0xfffb, 2(AX)
+	MOVB $0xff, 4(AX)
+	ADDQ $0x05, AX
+	ADDQ $0x05, BX
+	JMP  emit_repeat_again_standalone_emit_copy_short_2b
+
+repeat_five_standalone_emit_copy_short_2b:
+	LEAL -65536(DX), DX
+	MOVL DX, CX
+	MOVW $0x001d, (AX)
+	MOVW DX, 2(AX)
+	SARL $0x10, CX
+	MOVB CL, 4(AX)
+	ADDQ $0x05, BX
+	ADDQ $0x05, AX
+	JMP  gen_emit_copy_end
+
+repeat_four_standalone_emit_copy_short_2b:
+	LEAL -256(DX), DX
+	MOVW $0x0019, (AX)
+	MOVW DX, 2(AX)
+	ADDQ $0x04, BX
+	ADDQ $0x04, AX
+	JMP  gen_emit_copy_end
+
+repeat_three_standalone_emit_copy_short_2b:
+	LEAL -4(DX), DX
+	MOVW $0x0015, (AX)
+	MOVB DL, 2(AX)
+	ADDQ $0x03, BX
+	ADDQ $0x03, AX
+	JMP  gen_emit_copy_end
+
+repeat_two_standalone_emit_copy_short_2b:
+	SHLL $0x02, DX
+	ORL  $0x01, DX
+	MOVW DX, (AX)
+	ADDQ $0x02, BX
+	ADDQ $0x02, AX
+	JMP  gen_emit_copy_end
+
+repeat_two_offset_standalone_emit_copy_short_2b:
+	XORQ SI, SI
+	LEAL 1(SI)(DX*4), DX
+	MOVB CL, 1(AX)
+	SARL $0x08, CX
+	SHLL $0x05, CX
+	ORL  CX, DX
+	MOVB DL, (AX)
+	ADDQ $0x02, BX
+	ADDQ $0x02, AX
+	JMP  gen_emit_copy_end
+
+long_offset_short_standalone:
 	MOVB $0xee, (AX)
 	MOVW CX, 1(AX)
 	LEAL -60(DX), DX
@@ -16644,17 +17733,6 @@ matchlen_loopback_standalone:
 #ifdef GOAMD64_v3
 	TZCNTQ BX, BX
 
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef GOAMD64_v4
-	TZCNTQ BX, BX
-
-#define TZCNTQ_EMITTED 1
-#endif
-
-#ifdef TZCNTQ_EMITTED
-#undef TZCNTQ_EMITTED
 #else
 	BSFQ BX, BX
 
diff --git a/vendor/github.com/klauspost/compress/s2/index.go b/vendor/github.com/klauspost/compress/s2/index.go
index fd857682e..dd9ecfe71 100644
--- a/vendor/github.com/klauspost/compress/s2/index.go
+++ b/vendor/github.com/klauspost/compress/s2/index.go
@@ -10,6 +10,7 @@ import (
 	"encoding/json"
 	"fmt"
 	"io"
+	"sort"
 )
 
 const (
@@ -100,6 +101,15 @@ func (i *Index) Find(offset int64) (compressedOff, uncompressedOff int64, err er
 	if offset > i.TotalUncompressed {
 		return 0, 0, io.ErrUnexpectedEOF
 	}
+	if len(i.info) > 200 {
+		n := sort.Search(len(i.info), func(n int) bool {
+			return i.info[n].uncompressedOffset > offset
+		})
+		if n == 0 {
+			n = 1
+		}
+		return i.info[n-1].compressedOffset, i.info[n-1].uncompressedOffset, nil
+	}
 	for _, info := range i.info {
 		if info.uncompressedOffset > offset {
 			break
@@ -523,3 +533,66 @@ func (i *Index) JSON() []byte {
 	b, _ := json.MarshalIndent(x, "", "  ")
 	return b
 }
+
+// RemoveIndexHeaders will trim all headers and trailers from a given index.
+// This is expected to save 20 bytes.
+// These can be restored using RestoreIndexHeaders.
+// This removes a layer of security, but is the most compact representation.
+// Returns nil if headers contains errors.
+// The returned slice references the provided slice.
+func RemoveIndexHeaders(b []byte) []byte {
+	const save = 4 + len(S2IndexHeader) + len(S2IndexTrailer) + 4
+	if len(b) <= save {
+		return nil
+	}
+	if b[0] != ChunkTypeIndex {
+		return nil
+	}
+	chunkLen := int(b[1]) | int(b[2])<<8 | int(b[3])<<16
+	b = b[4:]
+
+	// Validate we have enough...
+	if len(b) < chunkLen {
+		return nil
+	}
+	b = b[:chunkLen]
+
+	if !bytes.Equal(b[:len(S2IndexHeader)], []byte(S2IndexHeader)) {
+		return nil
+	}
+	b = b[len(S2IndexHeader):]
+	if !bytes.HasSuffix(b, []byte(S2IndexTrailer)) {
+		return nil
+	}
+	b = bytes.TrimSuffix(b, []byte(S2IndexTrailer))
+
+	if len(b) < 4 {
+		return nil
+	}
+	return b[:len(b)-4]
+}
+
+// RestoreIndexHeaders will index restore headers removed by RemoveIndexHeaders.
+// No error checking is performed on the input.
+// If a 0 length slice is sent, it is returned without modification.
+func RestoreIndexHeaders(in []byte) []byte {
+	if len(in) == 0 {
+		return in
+	}
+	b := make([]byte, 0, 4+len(S2IndexHeader)+len(in)+len(S2IndexTrailer)+4)
+	b = append(b, ChunkTypeIndex, 0, 0, 0)
+	b = append(b, []byte(S2IndexHeader)...)
+	b = append(b, in...)
+
+	var tmp [4]byte
+	binary.LittleEndian.PutUint32(tmp[:], uint32(len(b)+4+len(S2IndexTrailer)))
+	b = append(b, tmp[:4]...)
+	// Trailer
+	b = append(b, []byte(S2IndexTrailer)...)
+
+	chunkLen := len(b) - skippableFrameHeader
+	b[1] = uint8(chunkLen >> 0)
+	b[2] = uint8(chunkLen >> 8)
+	b[3] = uint8(chunkLen >> 16)
+	return b
+}