[chore]: Bump github.com/minio/minio-go/v7 from 7.0.48 to 7.0.49 (#1567)

Bumps [github.com/minio/minio-go/v7](https://github.com/minio/minio-go) from 7.0.48 to 7.0.49.
- [Release notes](https://github.com/minio/minio-go/releases)
- [Commits](https://github.com/minio/minio-go/compare/v7.0.48...v7.0.49)

---
updated-dependencies:
- dependency-name: github.com/minio/minio-go/v7
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

go.mod

@@ -35,7 +35,7 @@ require (
 	github.com/jackc/pgx/v4 v4.17.2
 	github.com/microcosm-cc/bluemonday v1.0.22
 	github.com/miekg/dns v1.1.50
-	github.com/minio/minio-go/v7 v7.0.48
+	github.com/minio/minio-go/v7 v7.0.49
 	github.com/mitchellh/mapstructure v1.5.0
 	github.com/oklog/ulid v1.3.1
 	github.com/spf13/cobra v1.6.1
@@ -89,7 +89,7 @@ require (
 	github.com/dsoprea/go-photoshop-info-format v0.0.0-20200610045659-121dd752914d // indirect
 	github.com/dsoprea/go-png-image-structure/v2 v2.0.0-20210512210324-29b889a6093d // indirect
 	github.com/dsoprea/go-utility/v2 v2.0.0-20200717064901-2fccff4aa15e // indirect
-	github.com/dustin/go-humanize v1.0.0 // indirect
+	github.com/dustin/go-humanize v1.0.1 // indirect
 	github.com/fsnotify/fsnotify v1.6.0 // indirect
 	github.com/gin-contrib/sse v0.1.0 // indirect
 	github.com/go-errors/errors v1.4.1 // indirect
@@ -118,8 +118,8 @@ require (
 	github.com/jinzhu/inflection v1.0.0 // indirect
 	github.com/json-iterator/go v1.1.12 // indirect
 	github.com/kballard/go-shellquote v0.0.0-20180428030007-95032a82bc51 // indirect
-	github.com/klauspost/compress v1.15.9 // indirect
+	github.com/klauspost/compress v1.15.15 // indirect
-	github.com/klauspost/cpuid/v2 v2.1.1 // indirect
+	github.com/klauspost/cpuid/v2 v2.2.3 // indirect
 	github.com/leodido/go-urn v1.2.1 // indirect
 	github.com/magiconair/properties v1.8.7 // indirect
 	github.com/mattn/go-isatty v0.0.17 // indirect

go.sum

@@ -161,8 +161,9 @@ github.com/dsoprea/go-png-image-structure/v2 v2.0.0-20210512210324-29b889a6093d/
 github.com/dsoprea/go-utility v0.0.0-20200711062821-fab8125e9bdf/go.mod h1:95+K3z2L0mqsVYd6yveIv1lmtT3tcQQ3dVakPySffW8=
 github.com/dsoprea/go-utility/v2 v2.0.0-20200717064901-2fccff4aa15e h1:IxIbA7VbCNrwumIYjDoMOdf4KOSkMC6NJE4s8oRbE7E=
 github.com/dsoprea/go-utility/v2 v2.0.0-20200717064901-2fccff4aa15e/go.mod h1:uAzdkPTub5Y9yQwXe8W4m2XuP0tK4a9Q/dantD0+uaU=
-github.com/dustin/go-humanize v1.0.0 h1:VSnTsYCnlFHaM2/igO1h6X3HA71jcobQuxemgkq4zYo=
 github.com/dustin/go-humanize v1.0.0/go.mod h1:HtrtbFcZ19U5GC7JDqmcUSB87Iq5E25KnS6fMYU6eOk=
+github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
+github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
 github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
 github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
 github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
@@ -398,13 +399,13 @@ github.com/kisielk/errcheck v1.5.0/go.mod h1:pFxgyoBC7bSaBwPgfKdkLd5X25qrDl4LWUI
 github.com/kisielk/gotool v1.0.0/go.mod h1:XhKaO+MFFWcvkIS/tQcRk01m1F5IRFswLeQ+oQHNcck=
 github.com/klauspost/compress v1.10.4/go.mod h1:aoV0uJVorq1K+umq18yTdKaF57EivdYsUV+/s2qKfXs=
 github.com/klauspost/compress v1.10.10/go.mod h1:aoV0uJVorq1K+umq18yTdKaF57EivdYsUV+/s2qKfXs=
-github.com/klauspost/compress v1.15.9 h1:wKRjX6JRtDdrE9qwa4b/Cip7ACOshUI4smpCQanqjSY=
+github.com/klauspost/compress v1.15.15 h1:EF27CXIuDsYJ6mmvtBRlEuB2UVOqHG1tAXgZ7yIO+lw=
-github.com/klauspost/compress v1.15.9/go.mod h1:PhcZ0MbTNciWF3rruxRgKxI5NkcHHrHUDtV4Yw2GlzU=
+github.com/klauspost/compress v1.15.15/go.mod h1:ZcK2JAFqKOpnBlxcLsJzYfrS9X1akm9fHZNnD9+Vo/4=
 github.com/klauspost/cpuid/v2 v2.0.1/go.mod h1:FInQzS24/EEf25PyTYn52gqo7WaD8xa0213Md/qVLRg=
 github.com/klauspost/cpuid/v2 v2.0.4/go.mod h1:FInQzS24/EEf25PyTYn52gqo7WaD8xa0213Md/qVLRg=
 github.com/klauspost/cpuid/v2 v2.0.9/go.mod h1:FInQzS24/EEf25PyTYn52gqo7WaD8xa0213Md/qVLRg=
-github.com/klauspost/cpuid/v2 v2.1.1 h1:t0wUqjowdm8ezddV5k0tLWVklVuvLJpoHeb4WBdydm0=
+github.com/klauspost/cpuid/v2 v2.2.3 h1:sxCkb+qR91z4vsqw4vGGZlDgPz3G7gjaLyK3V8y70BU=
-github.com/klauspost/cpuid/v2 v2.1.1/go.mod h1:RVVoqg1df56z8g3pUjL/3lE5UfnlrJX8tyFgg4nqhuY=
+github.com/klauspost/cpuid/v2 v2.2.3/go.mod h1:RVVoqg1df56z8g3pUjL/3lE5UfnlrJX8tyFgg4nqhuY=
 github.com/konsorten/go-windows-terminal-sequences v1.0.1/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
 github.com/konsorten/go-windows-terminal-sequences v1.0.2/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
 github.com/kr/fs v0.1.0/go.mod h1:FFnZGqtBN9Gxj7eW1uZ42v5BccTP0vu6NEaFoC2HwRg=
@@ -443,8 +444,8 @@ github.com/miekg/dns v1.1.50 h1:DQUfb9uc6smULcREF09Uc+/Gd46YWqJd5DbpPE9xkcA=
 github.com/miekg/dns v1.1.50/go.mod h1:e3IlAVfNqAllflbibAZEWOXOQ+Ynzk/dDozDxY7XnME=
 github.com/minio/md5-simd v1.1.2 h1:Gdi1DZK69+ZVMoNHRXJyNcxrMA4dSxoYHZSQbirFg34=
 github.com/minio/md5-simd v1.1.2/go.mod h1:MzdKDxYpY2BT9XQFocsiZf/NKVtR7nkE4RoEpN+20RM=
-github.com/minio/minio-go/v7 v7.0.48 h1:VQtYB/2xHW2SlxqhjRlDpvSiSOfGlyFlXZF1EHARPHM=
+github.com/minio/minio-go/v7 v7.0.49 h1:dE5DfOtnXMXCjr/HWI6zN9vCrY6Sv666qhhiwUMvGV4=
-github.com/minio/minio-go/v7 v7.0.48/go.mod h1:nCrRzjoSUQh8hgKKtu3Y708OLvRLtuASMg2/nvmbarw=
+github.com/minio/minio-go/v7 v7.0.49/go.mod h1:UI34MvQEiob3Cf/gGExGMmzugkM/tNgbFypNDy5LMVc=
 github.com/minio/sha256-simd v1.0.0 h1:v1ta+49hkWZyvaKwrQB8elexRqm6Y0aMLjCNsrYxo6g=
 github.com/minio/sha256-simd v1.0.0/go.mod h1:OuYzVNI5vcoYIAmbIvHPl3N3jUzVedXbKy5RFepssQM=
 github.com/mitchellh/hashstructure/v2 v2.0.2 h1:vGKWl0YJqUNxE8d+h8f6NJLcCJrgbhC4NcD46KavDd4=

vendor/github.com/dustin/go-humanize/.travis.yml

@@ -1,12 +1,12 @@
 sudo: false
 language: go
+go_import_path: github.com/dustin/go-humanize
 go:
-  - 1.3.x
+  - 1.13.x
-  - 1.5.x
+  - 1.14.x
-  - 1.6.x
+  - 1.15.x
-  - 1.7.x
+  - 1.16.x
-  - 1.8.x
+  - stable
-  - 1.9.x
   - master
 matrix:
   allow_failures:
@@ -15,7 +15,7 @@ matrix:
 install:
   - # Do nothing. This is needed to prevent default install action "go get -t -v ./..." from happening here (we want it to happen inside script step).
 script:
-  - go get -t -v ./...
   - diff -u <(echo -n) <(gofmt -d -s .)
-  - go tool vet .
+  - go vet .
+  - go install -v -race ./...
   - go test -v -race ./...

vendor/github.com/dustin/go-humanize/README.markdown

@@ -5,7 +5,7 @@ Just a few functions for helping humanize times and sizes.
 `go get` it as `github.com/dustin/go-humanize`, import it as
 `"github.com/dustin/go-humanize"`, use it as `humanize`.
 
-See [godoc](https://godoc.org/github.com/dustin/go-humanize) for
+See [godoc](https://pkg.go.dev/github.com/dustin/go-humanize) for
 complete documentation.
 
 ## Sizes

vendor/github.com/dustin/go-humanize/bigbytes.go

@@ -28,6 +28,10 @@
 	BigZiByte = (&big.Int{}).Mul(BigEiByte, bigIECExp)
 	// BigYiByte is 1,024 z bytes in bit.Ints
 	BigYiByte = (&big.Int{}).Mul(BigZiByte, bigIECExp)
+	// BigRiByte is 1,024 y bytes in bit.Ints
+	BigRiByte = (&big.Int{}).Mul(BigYiByte, bigIECExp)
+	// BigQiByte is 1,024 r bytes in bit.Ints
+	BigQiByte = (&big.Int{}).Mul(BigRiByte, bigIECExp)
 )
 
 var (
@@ -51,6 +55,10 @@
 	BigZByte = (&big.Int{}).Mul(BigEByte, bigSIExp)
 	// BigYByte is 1,000 SI z bytes in big.Ints
 	BigYByte = (&big.Int{}).Mul(BigZByte, bigSIExp)
+	// BigRByte is 1,000 SI y bytes in big.Ints
+	BigRByte = (&big.Int{}).Mul(BigYByte, bigSIExp)
+	// BigQByte is 1,000 SI r bytes in big.Ints
+	BigQByte = (&big.Int{}).Mul(BigRByte, bigSIExp)
 )
 
 var bigBytesSizeTable = map[string]*big.Int{
@@ -71,6 +79,10 @@
 	"zb":  BigZByte,
 	"yib": BigYiByte,
 	"yb":  BigYByte,
+	"rib": BigRiByte,
+	"rb":  BigRByte,
+	"qib": BigQiByte,
+	"qb":  BigQByte,
 	// Without suffix
 	"":   BigByte,
 	"ki": BigKiByte,
@@ -89,6 +101,10 @@
 	"zi": BigZiByte,
 	"y":  BigYByte,
 	"yi": BigYiByte,
+	"r":  BigRByte,
+	"ri": BigRiByte,
+	"q":  BigQByte,
+	"qi": BigQiByte,
 }
 
 var ten = big.NewInt(10)
@@ -115,7 +131,7 @@ func humanateBigBytes(s, base *big.Int, sizes []string) string {
 //
 // BigBytes(82854982) -> 83 MB
 func BigBytes(s *big.Int) string {
-	sizes := []string{"B", "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB"}
+	sizes := []string{"B", "kB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB", "RB", "QB"}
 	return humanateBigBytes(s, bigSIExp, sizes)
 }
 
@@ -125,7 +141,7 @@ func BigBytes(s *big.Int) string {
 //
 // BigIBytes(82854982) -> 79 MiB
 func BigIBytes(s *big.Int) string {
-	sizes := []string{"B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB"}
+	sizes := []string{"B", "KiB", "MiB", "GiB", "TiB", "PiB", "EiB", "ZiB", "YiB", "RiB", "QiB"}
 	return humanateBigBytes(s, bigIECExp, sizes)
 }
 

vendor/github.com/dustin/go-humanize/commaf.go

@@ -1,3 +1,4 @@
+//go:build go1.6
 // +build go1.6
 
 package humanize

vendor/github.com/dustin/go-humanize/ftoa.go

@@ -6,6 +6,9 @@
 )
 
 func stripTrailingZeros(s string) string {
+	if !strings.ContainsRune(s, '.') {
+		return s
+	}
 	offset := len(s) - 1
 	for offset > 0 {
 		if s[offset] == '.' {

vendor/github.com/dustin/go-humanize/number.go

@@ -73,7 +73,7 @@ func FormatFloat(format string, n float64) string {
 	if n > math.MaxFloat64 {
 		return "Infinity"
 	}
-	if n < -math.MaxFloat64 {
+	if n < (0.0 - math.MaxFloat64) {
 		return "-Infinity"
 	}
 

vendor/github.com/dustin/go-humanize/si.go

@@ -8,6 +8,8 @@
 )
 
 var siPrefixTable = map[float64]string{
+	-30: "q", // quecto
+	-27: "r", // ronto
 	-24: "y", // yocto
 	-21: "z", // zepto
 	-18: "a", // atto
@@ -25,6 +27,8 @@
 	18:  "E", // exa
 	21:  "Z", // zetta
 	24:  "Y", // yotta
+	27:  "R", // ronna
+	30:  "Q", // quetta
 }
 
 var revSIPrefixTable = revfmap(siPrefixTable)

vendor/github.com/klauspost/compress/flate/deflate.go

@@ -131,7 +131,8 @@ func (d *compressor) fillDeflate(b []byte) int {
 	s := d.state
 	if s.index >= 2*windowSize-(minMatchLength+maxMatchLength) {
 		// shift the window by windowSize
-		copy(d.window[:], d.window[windowSize:2*windowSize])
+		//copy(d.window[:], d.window[windowSize:2*windowSize])
+		*(*[windowSize]byte)(d.window) = *(*[windowSize]byte)(d.window[windowSize:])
 		s.index -= windowSize
 		d.windowEnd -= windowSize
 		if d.blockStart >= windowSize {
@@ -293,7 +294,6 @@ func (d *compressor) findMatch(pos int, prevHead int, lookahead int) (length, of
 	}
 	offset = 0
 
-	cGain := 0
 	if d.chain < 100 {
 		for i := prevHead; tries > 0; tries-- {
 			if wEnd == win[i+length] {
@@ -321,10 +321,14 @@ func (d *compressor) findMatch(pos int, prevHead int, lookahead int) (length, of
 		return
 	}
 
+	// Minimum gain to accept a match.
+	cGain := 4
+
 	// Some like it higher (CSV), some like it lower (JSON)
-	const baseCost = 6
+	const baseCost = 3
 	// Base is 4 bytes at with an additional cost.
 	// Matches must be better than this.
+
 	for i := prevHead; tries > 0; tries-- {
 		if wEnd == win[i+length] {
 			n := matchLen(win[i:i+minMatchLook], wPos)
@@ -332,7 +336,7 @@ func (d *compressor) findMatch(pos int, prevHead int, lookahead int) (length, of
 				// Calculate gain. Estimate
 				newGain := d.h.bitLengthRaw(wPos[:n]) - int(offsetExtraBits[offsetCode(uint32(pos-i))]) - baseCost - int(lengthExtraBits[lengthCodes[(n-3)&255]])
 
-				//fmt.Println(n, "gain:", newGain, "prev:", cGain, "raw:", d.h.bitLengthRaw(wPos[:n]))
+				//fmt.Println("gain:", newGain, "prev:", cGain, "raw:", d.h.bitLengthRaw(wPos[:n]), "this-len:", n, "prev-len:", length)
 				if newGain > cGain {
 					length = n
 					offset = pos - i
@@ -373,6 +377,12 @@ func hash4(b []byte) uint32 {
 	return hash4u(binary.LittleEndian.Uint32(b), hashBits)
 }
 
+// hash4 returns the hash of u to fit in a hash table with h bits.
+// Preferably h should be a constant and should always be <32.
+func hash4u(u uint32, h uint8) uint32 {
+	return (u * prime4bytes) >> (32 - h)
+}
+
 // bulkHash4 will compute hashes using the same
 // algorithm as hash4
 func bulkHash4(b []byte, dst []uint32) {
@@ -483,27 +493,103 @@ func (d *compressor) deflateLazy() {
 		}
 
 		if prevLength >= minMatchLength && s.length <= prevLength {
-			// Check for better match at end...
+			// No better match, but check for better match at end...
 			//
-			// checkOff must be >=2 since we otherwise risk checking s.index
+			// Skip forward a number of bytes.
-			// Offset of 2 seems to yield best results.
+			// Offset of 2 seems to yield best results. 3 is sometimes better.
 			const checkOff = 2
+
+			// Check all, except full length
+			if prevLength < maxMatchLength-checkOff {
 				prevIndex := s.index - 1
-			if prevIndex+prevLength+checkOff < s.maxInsertIndex {
+				if prevIndex+prevLength < s.maxInsertIndex {
 					end := lookahead
-				if lookahead > maxMatchLength {
+					if lookahead > maxMatchLength+checkOff {
-					end = maxMatchLength
+						end = maxMatchLength + checkOff
 					}
 					end += prevIndex
-				idx := prevIndex + prevLength - (4 - checkOff)
+
-				h := hash4(d.window[idx:])
+					// Hash at match end.
-				ch2 := int(s.hashHead[h]) - s.hashOffset - prevLength + (4 - checkOff)
+					h := hash4(d.window[prevIndex+prevLength:])
-				if ch2 > minIndex {
+					ch2 := int(s.hashHead[h]) - s.hashOffset - prevLength
-					length := matchLen(d.window[prevIndex:end], d.window[ch2:])
+					if prevIndex-ch2 != prevOffset && ch2 > minIndex+checkOff {
+						length := matchLen(d.window[prevIndex+checkOff:end], d.window[ch2+checkOff:])
 						// It seems like a pure length metric is best.
 						if length > prevLength {
 							prevLength = length
 							prevOffset = prevIndex - ch2
+
+							// Extend back...
+							for i := checkOff - 1; i >= 0; i-- {
+								if prevLength >= maxMatchLength || d.window[prevIndex+i] != d.window[ch2+i] {
+									// Emit tokens we "owe"
+									for j := 0; j <= i; j++ {
+										d.tokens.AddLiteral(d.window[prevIndex+j])
+										if d.tokens.n == maxFlateBlockTokens {
+											// The block includes the current character
+											if d.err = d.writeBlock(&d.tokens, s.index, false); d.err != nil {
+												return
+											}
+											d.tokens.Reset()
+										}
+										s.index++
+										if s.index < s.maxInsertIndex {
+											h := hash4(d.window[s.index:])
+											ch := s.hashHead[h]
+											s.chainHead = int(ch)
+											s.hashPrev[s.index&windowMask] = ch
+											s.hashHead[h] = uint32(s.index + s.hashOffset)
+										}
+									}
+									break
+								} else {
+									prevLength++
+								}
+							}
+						} else if false {
+							// Check one further ahead.
+							// Only rarely better, disabled for now.
+							prevIndex++
+							h := hash4(d.window[prevIndex+prevLength:])
+							ch2 := int(s.hashHead[h]) - s.hashOffset - prevLength
+							if prevIndex-ch2 != prevOffset && ch2 > minIndex+checkOff {
+								length := matchLen(d.window[prevIndex+checkOff:end], d.window[ch2+checkOff:])
+								// It seems like a pure length metric is best.
+								if length > prevLength+checkOff {
+									prevLength = length
+									prevOffset = prevIndex - ch2
+									prevIndex--
+
+									// Extend back...
+									for i := checkOff; i >= 0; i-- {
+										if prevLength >= maxMatchLength || d.window[prevIndex+i] != d.window[ch2+i-1] {
+											// Emit tokens we "owe"
+											for j := 0; j <= i; j++ {
+												d.tokens.AddLiteral(d.window[prevIndex+j])
+												if d.tokens.n == maxFlateBlockTokens {
+													// The block includes the current character
+													if d.err = d.writeBlock(&d.tokens, s.index, false); d.err != nil {
+														return
+													}
+													d.tokens.Reset()
+												}
+												s.index++
+												if s.index < s.maxInsertIndex {
+													h := hash4(d.window[s.index:])
+													ch := s.hashHead[h]
+													s.chainHead = int(ch)
+													s.hashPrev[s.index&windowMask] = ch
+													s.hashHead[h] = uint32(s.index + s.hashOffset)
+												}
+											}
+											break
+										} else {
+											prevLength++
+										}
+									}
+								}
+							}
+						}
 					}
 				}
 			}

vendor/github.com/klauspost/compress/flate/dict_decoder.go

@@ -7,13 +7,13 @@
 // dictDecoder implements the LZ77 sliding dictionary as used in decompression.
 // LZ77 decompresses data through sequences of two forms of commands:
 //
-//	* Literal insertions: Runs of one or more symbols are inserted into the data
+//   - Literal insertions: Runs of one or more symbols are inserted into the data
 //     stream as is. This is accomplished through the writeByte method for a
 //     single symbol, or combinations of writeSlice/writeMark for multiple symbols.
 //     Any valid stream must start with a literal insertion if no preset dictionary
 //     is used.
 //
-//	* Backward copies: Runs of one or more symbols are copied from previously
+//   - Backward copies: Runs of one or more symbols are copied from previously
 //     emitted data. Backward copies come as the tuple (dist, length) where dist
 //     determines how far back in the stream to copy from and length determines how
 //     many bytes to copy. Note that it is valid for the length to be greater than

vendor/github.com/klauspost/compress/flate/fast_encoder.go

@@ -58,17 +58,6 @@ func newFastEnc(level int) fastEnc {
 	prime8bytes = 0xcf1bbcdcb7a56463
 )
 
-func load32(b []byte, i int) uint32 {
-	// Help the compiler eliminate bounds checks on the read so it can be done in a single read.
-	b = b[i:]
-	b = b[:4]
-	return uint32(b[0]) | uint32(b[1])<<8 | uint32(b[2])<<16 | uint32(b[3])<<24
-}
-
-func load64(b []byte, i int) uint64 {
-	return binary.LittleEndian.Uint64(b[i:])
-}
-
 func load3232(b []byte, i int32) uint32 {
 	return binary.LittleEndian.Uint32(b[i:])
 }
@@ -77,10 +66,6 @@ func load6432(b []byte, i int32) uint64 {
 	return binary.LittleEndian.Uint64(b[i:])
 }
 
-func hash(u uint32) uint32 {
-	return (u * 0x1e35a7bd) >> tableShift
-}
-
 type tableEntry struct {
 	offset int32
 }
@@ -104,7 +89,8 @@ func (e *fastGen) addBlock(src []byte) int32 {
 			}
 			// Move down
 			offset := int32(len(e.hist)) - maxMatchOffset
-			copy(e.hist[0:maxMatchOffset], e.hist[offset:])
+			// copy(e.hist[0:maxMatchOffset], e.hist[offset:])
+			*(*[maxMatchOffset]byte)(e.hist) = *(*[maxMatchOffset]byte)(e.hist[offset:])
 			e.cur += offset
 			e.hist = e.hist[:maxMatchOffset]
 		}
@@ -114,39 +100,36 @@ func (e *fastGen) addBlock(src []byte) int32 {
 	return s
 }
 
-// hash4 returns the hash of u to fit in a hash table with h bits.
-// Preferably h should be a constant and should always be <32.
-func hash4u(u uint32, h uint8) uint32 {
-	return (u * prime4bytes) >> (32 - h)
-}
-
 type tableEntryPrev struct {
 	Cur  tableEntry
 	Prev tableEntry
 }
 
-// hash4x64 returns the hash of the lowest 4 bytes of u to fit in a hash table with h bits.
-// Preferably h should be a constant and should always be <32.
-func hash4x64(u uint64, h uint8) uint32 {
-	return (uint32(u) * prime4bytes) >> ((32 - h) & reg8SizeMask32)
-}
-
 // hash7 returns the hash of the lowest 7 bytes of u to fit in a hash table with h bits.
 // Preferably h should be a constant and should always be <64.
 func hash7(u uint64, h uint8) uint32 {
 	return uint32(((u << (64 - 56)) * prime7bytes) >> ((64 - h) & reg8SizeMask64))
 }
 
-// hash8 returns the hash of u to fit in a hash table with h bits.
+// hashLen returns a hash of the lowest mls bytes of with length output bits.
-// Preferably h should be a constant and should always be <64.
+// mls must be >=3 and <=8. Any other value will return hash for 4 bytes.
-func hash8(u uint64, h uint8) uint32 {
+// length should always be < 32.
-	return uint32((u * prime8bytes) >> ((64 - h) & reg8SizeMask64))
+// Preferably length and mls should be a constant for inlining.
+func hashLen(u uint64, length, mls uint8) uint32 {
+	switch mls {
+	case 3:
+		return (uint32(u<<8) * prime3bytes) >> (32 - length)
+	case 5:
+		return uint32(((u << (64 - 40)) * prime5bytes) >> (64 - length))
+	case 6:
+		return uint32(((u << (64 - 48)) * prime6bytes) >> (64 - length))
+	case 7:
+		return uint32(((u << (64 - 56)) * prime7bytes) >> (64 - length))
+	case 8:
+		return uint32((u * prime8bytes) >> (64 - length))
+	default:
+		return (uint32(u) * prime4bytes) >> (32 - length)
 	}
-
-// hash6 returns the hash of the lowest 6 bytes of u to fit in a hash table with h bits.
-// Preferably h should be a constant and should always be <64.
-func hash6(u uint64, h uint8) uint32 {
-	return uint32(((u << (64 - 48)) * prime6bytes) >> ((64 - h) & reg8SizeMask64))
 }
 
 // matchlen will return the match length between offsets and t in src.

vendor/github.com/klauspost/compress/flate/huffman_bit_writer.go

@@ -790,9 +790,11 @@ func (w *huffmanBitWriter) fillTokens() {
 // and offsetEncoding.
 // The number of literal and offset tokens is returned.
 func (w *huffmanBitWriter) indexTokens(t *tokens, filled bool) (numLiterals, numOffsets int) {
-	copy(w.literalFreq[:], t.litHist[:])
+	//copy(w.literalFreq[:], t.litHist[:])
-	copy(w.literalFreq[256:], t.extraHist[:])
+	*(*[256]uint16)(w.literalFreq[:]) = t.litHist
-	copy(w.offsetFreq[:], t.offHist[:offsetCodeCount])
+	//copy(w.literalFreq[256:], t.extraHist[:])
+	*(*[32]uint16)(w.literalFreq[256:]) = t.extraHist
+	w.offsetFreq = t.offHist
 
 	if t.n == 0 {
 		return

vendor/github.com/klauspost/compress/flate/huffman_code.go

@@ -168,12 +168,17 @@ func (h *huffmanEncoder) canReuseBits(freq []uint16) int {
 // The cases of 0, 1, and 2 literals are handled by special case code.
 //
 // list  An array of the literals with non-zero frequencies
+//
 //	and their associated frequencies. The array is in order of increasing
 //	frequency, and has as its last element a special element with frequency
 //	MaxInt32
+//
 // maxBits     The maximum number of bits that should be used to encode any literal.
+//
 //	Must be less than 16.
+//
 // return      An integer array in which array[i] indicates the number of literals
+//
 //	that should be encoded in i bits.
 func (h *huffmanEncoder) bitCounts(list []literalNode, maxBits int32) []int32 {
 	if maxBits >= maxBitsLimit {

vendor/github.com/klauspost/compress/flate/level1.go

@@ -19,6 +19,7 @@ func (e *fastEncL1) Encode(dst *tokens, src []byte) {
 	const (
 		inputMargin            = 12 - 1
 		minNonLiteralBlockSize = 1 + 1 + inputMargin
+		hashBytes              = 5
 	)
 	if debugDeflate && e.cur < 0 {
 		panic(fmt.Sprint("e.cur < 0: ", e.cur))
@@ -68,7 +69,7 @@ func (e *fastEncL1) Encode(dst *tokens, src []byte) {
 	sLimit := int32(len(src) - inputMargin)
 
 	// nextEmit is where in src the next emitLiteral should start from.
-	cv := load3232(src, s)
+	cv := load6432(src, s)
 
 	for {
 		const skipLog = 5
@@ -77,7 +78,7 @@ func (e *fastEncL1) Encode(dst *tokens, src []byte) {
 		nextS := s
 		var candidate tableEntry
 		for {
-			nextHash := hash(cv)
+			nextHash := hashLen(cv, tableBits, hashBytes)
 			candidate = e.table[nextHash]
 			nextS = s + doEvery + (s-nextEmit)>>skipLog
 			if nextS > sLimit {
@@ -86,16 +87,16 @@ func (e *fastEncL1) Encode(dst *tokens, src []byte) {
 
 			now := load6432(src, nextS)
 			e.table[nextHash] = tableEntry{offset: s + e.cur}
-			nextHash = hash(uint32(now))
+			nextHash = hashLen(now, tableBits, hashBytes)
 
 			offset := s - (candidate.offset - e.cur)
-			if offset < maxMatchOffset && cv == load3232(src, candidate.offset-e.cur) {
+			if offset < maxMatchOffset && uint32(cv) == load3232(src, candidate.offset-e.cur) {
 				e.table[nextHash] = tableEntry{offset: nextS + e.cur}
 				break
 			}
 
 			// Do one right away...
-			cv = uint32(now)
+			cv = now
 			s = nextS
 			nextS++
 			candidate = e.table[nextHash]
@@ -103,11 +104,11 @@ func (e *fastEncL1) Encode(dst *tokens, src []byte) {
 			e.table[nextHash] = tableEntry{offset: s + e.cur}
 
 			offset = s - (candidate.offset - e.cur)
-			if offset < maxMatchOffset && cv == load3232(src, candidate.offset-e.cur) {
+			if offset < maxMatchOffset && uint32(cv) == load3232(src, candidate.offset-e.cur) {
 				e.table[nextHash] = tableEntry{offset: nextS + e.cur}
 				break
 			}
-			cv = uint32(now)
+			cv = now
 			s = nextS
 		}
 
@@ -198,9 +199,9 @@ func (e *fastEncL1) Encode(dst *tokens, src []byte) {
 			}
 			if s >= sLimit {
 				// Index first pair after match end.
-				if int(s+l+4) < len(src) {
+				if int(s+l+8) < len(src) {
-					cv := load3232(src, s)
+					cv := load6432(src, s)
-					e.table[hash(cv)] = tableEntry{offset: s + e.cur}
+					e.table[hashLen(cv, tableBits, hashBytes)] = tableEntry{offset: s + e.cur}
 				}
 				goto emitRemainder
 			}
@@ -213,16 +214,16 @@ func (e *fastEncL1) Encode(dst *tokens, src []byte) {
 			// three load32 calls.
 			x := load6432(src, s-2)
 			o := e.cur + s - 2
-			prevHash := hash(uint32(x))
+			prevHash := hashLen(x, tableBits, hashBytes)
 			e.table[prevHash] = tableEntry{offset: o}
 			x >>= 16
-			currHash := hash(uint32(x))
+			currHash := hashLen(x, tableBits, hashBytes)
 			candidate = e.table[currHash]
 			e.table[currHash] = tableEntry{offset: o + 2}
 
 			offset := s - (candidate.offset - e.cur)
 			if offset > maxMatchOffset || uint32(x) != load3232(src, candidate.offset-e.cur) {
-				cv = uint32(x >> 8)
+				cv = x >> 8
 				s++
 				break
 			}

vendor/github.com/klauspost/compress/flate/level2.go

@@ -16,6 +16,7 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 	const (
 		inputMargin            = 12 - 1
 		minNonLiteralBlockSize = 1 + 1 + inputMargin
+		hashBytes              = 5
 	)
 
 	if debugDeflate && e.cur < 0 {
@@ -66,7 +67,7 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 	sLimit := int32(len(src) - inputMargin)
 
 	// nextEmit is where in src the next emitLiteral should start from.
-	cv := load3232(src, s)
+	cv := load6432(src, s)
 	for {
 		// When should we start skipping if we haven't found matches in a long while.
 		const skipLog = 5
@@ -75,7 +76,7 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 		nextS := s
 		var candidate tableEntry
 		for {
-			nextHash := hash4u(cv, bTableBits)
+			nextHash := hashLen(cv, bTableBits, hashBytes)
 			s = nextS
 			nextS = s + doEvery + (s-nextEmit)>>skipLog
 			if nextS > sLimit {
@@ -84,16 +85,16 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 			candidate = e.table[nextHash]
 			now := load6432(src, nextS)
 			e.table[nextHash] = tableEntry{offset: s + e.cur}
-			nextHash = hash4u(uint32(now), bTableBits)
+			nextHash = hashLen(now, bTableBits, hashBytes)
 
 			offset := s - (candidate.offset - e.cur)
-			if offset < maxMatchOffset && cv == load3232(src, candidate.offset-e.cur) {
+			if offset < maxMatchOffset && uint32(cv) == load3232(src, candidate.offset-e.cur) {
 				e.table[nextHash] = tableEntry{offset: nextS + e.cur}
 				break
 			}
 
 			// Do one right away...
-			cv = uint32(now)
+			cv = now
 			s = nextS
 			nextS++
 			candidate = e.table[nextHash]
@@ -101,10 +102,10 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 			e.table[nextHash] = tableEntry{offset: s + e.cur}
 
 			offset = s - (candidate.offset - e.cur)
-			if offset < maxMatchOffset && cv == load3232(src, candidate.offset-e.cur) {
+			if offset < maxMatchOffset && uint32(cv) == load3232(src, candidate.offset-e.cur) {
 				break
 			}
-			cv = uint32(now)
+			cv = now
 		}
 
 		// A 4-byte match has been found. We'll later see if more than 4 bytes
@@ -154,9 +155,9 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 
 			if s >= sLimit {
 				// Index first pair after match end.
-				if int(s+l+4) < len(src) {
+				if int(s+l+8) < len(src) {
-					cv := load3232(src, s)
+					cv := load6432(src, s)
-					e.table[hash4u(cv, bTableBits)] = tableEntry{offset: s + e.cur}
+					e.table[hashLen(cv, bTableBits, hashBytes)] = tableEntry{offset: s + e.cur}
 				}
 				goto emitRemainder
 			}
@@ -164,15 +165,15 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 			// Store every second hash in-between, but offset by 1.
 			for i := s - l + 2; i < s-5; i += 7 {
 				x := load6432(src, i)
-				nextHash := hash4u(uint32(x), bTableBits)
+				nextHash := hashLen(x, bTableBits, hashBytes)
 				e.table[nextHash] = tableEntry{offset: e.cur + i}
 				// Skip one
 				x >>= 16
-				nextHash = hash4u(uint32(x), bTableBits)
+				nextHash = hashLen(x, bTableBits, hashBytes)
 				e.table[nextHash] = tableEntry{offset: e.cur + i + 2}
 				// Skip one
 				x >>= 16
-				nextHash = hash4u(uint32(x), bTableBits)
+				nextHash = hashLen(x, bTableBits, hashBytes)
 				e.table[nextHash] = tableEntry{offset: e.cur + i + 4}
 			}
 
@@ -184,17 +185,17 @@ func (e *fastEncL2) Encode(dst *tokens, src []byte) {
 			// three load32 calls.
 			x := load6432(src, s-2)
 			o := e.cur + s - 2
-			prevHash := hash4u(uint32(x), bTableBits)
+			prevHash := hashLen(x, bTableBits, hashBytes)
-			prevHash2 := hash4u(uint32(x>>8), bTableBits)
+			prevHash2 := hashLen(x>>8, bTableBits, hashBytes)
 			e.table[prevHash] = tableEntry{offset: o}
 			e.table[prevHash2] = tableEntry{offset: o + 1}
-			currHash := hash4u(uint32(x>>16), bTableBits)
+			currHash := hashLen(x>>16, bTableBits, hashBytes)
 			candidate = e.table[currHash]
 			e.table[currHash] = tableEntry{offset: o + 2}
 
 			offset := s - (candidate.offset - e.cur)
 			if offset > maxMatchOffset || uint32(x>>16) != load3232(src, candidate.offset-e.cur) {
-				cv = uint32(x >> 24)
+				cv = x >> 24
 				s++
 				break
 			}

vendor/github.com/klauspost/compress/flate/level3.go

@@ -11,10 +11,11 @@ type fastEncL3 struct {
 // Encode uses a similar algorithm to level 2, will check up to two candidates.
 func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 	const (
-		inputMargin            = 8 - 1
+		inputMargin            = 12 - 1
 		minNonLiteralBlockSize = 1 + 1 + inputMargin
 		tableBits              = 16
 		tableSize              = 1 << tableBits
+		hashBytes              = 5
 	)
 
 	if debugDeflate && e.cur < 0 {
@@ -69,20 +70,20 @@ func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 	sLimit := int32(len(src) - inputMargin)
 
 	// nextEmit is where in src the next emitLiteral should start from.
-	cv := load3232(src, s)
+	cv := load6432(src, s)
 	for {
-		const skipLog = 6
+		const skipLog = 7
 		nextS := s
 		var candidate tableEntry
 		for {
-			nextHash := hash4u(cv, tableBits)
+			nextHash := hashLen(cv, tableBits, hashBytes)
 			s = nextS
 			nextS = s + 1 + (s-nextEmit)>>skipLog
 			if nextS > sLimit {
 				goto emitRemainder
 			}
 			candidates := e.table[nextHash]
-			now := load3232(src, nextS)
+			now := load6432(src, nextS)
 
 			// Safe offset distance until s + 4...
 			minOffset := e.cur + s - (maxMatchOffset - 4)
@@ -96,8 +97,8 @@ func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 				continue
 			}
 
-			if cv == load3232(src, candidate.offset-e.cur) {
+			if uint32(cv) == load3232(src, candidate.offset-e.cur) {
-				if candidates.Prev.offset < minOffset || cv != load3232(src, candidates.Prev.offset-e.cur) {
+				if candidates.Prev.offset < minOffset || uint32(cv) != load3232(src, candidates.Prev.offset-e.cur) {
 					break
 				}
 				// Both match and are valid, pick longest.
@@ -112,7 +113,7 @@ func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 				// We only check if value mismatches.
 				// Offset will always be invalid in other cases.
 				candidate = candidates.Prev
-				if candidate.offset > minOffset && cv == load3232(src, candidate.offset-e.cur) {
+				if candidate.offset > minOffset && uint32(cv) == load3232(src, candidate.offset-e.cur) {
 					break
 				}
 			}
@@ -164,9 +165,9 @@ func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 			if s >= sLimit {
 				t += l
 				// Index first pair after match end.
-				if int(t+4) < len(src) && t > 0 {
+				if int(t+8) < len(src) && t > 0 {
-					cv := load3232(src, t)
+					cv = load6432(src, t)
-					nextHash := hash4u(cv, tableBits)
+					nextHash := hashLen(cv, tableBits, hashBytes)
 					e.table[nextHash] = tableEntryPrev{
 						Prev: e.table[nextHash].Cur,
 						Cur:  tableEntry{offset: e.cur + t},
@@ -176,8 +177,8 @@ func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 			}
 
 			// Store every 5th hash in-between.
-			for i := s - l + 2; i < s-5; i += 5 {
+			for i := s - l + 2; i < s-5; i += 6 {
-				nextHash := hash4u(load3232(src, i), tableBits)
+				nextHash := hashLen(load6432(src, i), tableBits, hashBytes)
 				e.table[nextHash] = tableEntryPrev{
 					Prev: e.table[nextHash].Cur,
 					Cur:  tableEntry{offset: e.cur + i}}
@@ -185,23 +186,23 @@ func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 			// We could immediately start working at s now, but to improve
 			// compression we first update the hash table at s-2 to s.
 			x := load6432(src, s-2)
-			prevHash := hash4u(uint32(x), tableBits)
+			prevHash := hashLen(x, tableBits, hashBytes)
 
 			e.table[prevHash] = tableEntryPrev{
 				Prev: e.table[prevHash].Cur,
 				Cur:  tableEntry{offset: e.cur + s - 2},
 			}
 			x >>= 8
-			prevHash = hash4u(uint32(x), tableBits)
+			prevHash = hashLen(x, tableBits, hashBytes)
 
 			e.table[prevHash] = tableEntryPrev{
 				Prev: e.table[prevHash].Cur,
 				Cur:  tableEntry{offset: e.cur + s - 1},
 			}
 			x >>= 8
-			currHash := hash4u(uint32(x), tableBits)
+			currHash := hashLen(x, tableBits, hashBytes)
 			candidates := e.table[currHash]
-			cv = uint32(x)
+			cv = x
 			e.table[currHash] = tableEntryPrev{
 				Prev: candidates.Cur,
 				Cur:  tableEntry{offset: s + e.cur},
@@ -212,17 +213,17 @@ func (e *fastEncL3) Encode(dst *tokens, src []byte) {
 			minOffset := e.cur + s - (maxMatchOffset - 4)
 
 			if candidate.offset > minOffset {
-				if cv == load3232(src, candidate.offset-e.cur) {
+				if uint32(cv) == load3232(src, candidate.offset-e.cur) {
 					// Found a match...
 					continue
 				}
 				candidate = candidates.Prev
-				if candidate.offset > minOffset && cv == load3232(src, candidate.offset-e.cur) {
+				if candidate.offset > minOffset && uint32(cv) == load3232(src, candidate.offset-e.cur) {
 					// Match at prev...
 					continue
 				}
 			}
-			cv = uint32(x >> 8)
+			cv = x >> 8
 			s++
 			break
 		}

vendor/github.com/klauspost/compress/flate/level4.go

@@ -12,6 +12,7 @@ func (e *fastEncL4) Encode(dst *tokens, src []byte) {
 	const (
 		inputMargin            = 12 - 1
 		minNonLiteralBlockSize = 1 + 1 + inputMargin
+		hashShortBytes         = 4
 	)
 	if debugDeflate && e.cur < 0 {
 		panic(fmt.Sprint("e.cur < 0: ", e.cur))
@@ -80,7 +81,7 @@ func (e *fastEncL4) Encode(dst *tokens, src []byte) {
 		nextS := s
 		var t int32
 		for {
-			nextHashS := hash4x64(cv, tableBits)
+			nextHashS := hashLen(cv, tableBits, hashShortBytes)
 			nextHashL := hash7(cv, tableBits)
 
 			s = nextS
@@ -168,7 +169,7 @@ func (e *fastEncL4) Encode(dst *tokens, src []byte) {
 			// Index first pair after match end.
 			if int(s+8) < len(src) {
 				cv := load6432(src, s)
-				e.table[hash4x64(cv, tableBits)] = tableEntry{offset: s + e.cur}
+				e.table[hashLen(cv, tableBits, hashShortBytes)] = tableEntry{offset: s + e.cur}
 				e.bTable[hash7(cv, tableBits)] = tableEntry{offset: s + e.cur}
 			}
 			goto emitRemainder
@@ -183,7 +184,7 @@ func (e *fastEncL4) Encode(dst *tokens, src []byte) {
 				t2 := tableEntry{offset: t.offset + 1}
 				e.bTable[hash7(cv, tableBits)] = t
 				e.bTable[hash7(cv>>8, tableBits)] = t2
-				e.table[hash4u(uint32(cv>>8), tableBits)] = t2
+				e.table[hashLen(cv>>8, tableBits, hashShortBytes)] = t2
 
 				i += 3
 				for ; i < s-1; i += 3 {
@@ -192,7 +193,7 @@ func (e *fastEncL4) Encode(dst *tokens, src []byte) {
 					t2 := tableEntry{offset: t.offset + 1}
 					e.bTable[hash7(cv, tableBits)] = t
 					e.bTable[hash7(cv>>8, tableBits)] = t2
-					e.table[hash4u(uint32(cv>>8), tableBits)] = t2
+					e.table[hashLen(cv>>8, tableBits, hashShortBytes)] = t2
 				}
 			}
 		}
@@ -201,7 +202,7 @@ func (e *fastEncL4) Encode(dst *tokens, src []byte) {
 		// compression we first update the hash table at s-1 and at s.
 		x := load6432(src, s-1)
 		o := e.cur + s - 1
-		prevHashS := hash4x64(x, tableBits)
+		prevHashS := hashLen(x, tableBits, hashShortBytes)
 		prevHashL := hash7(x, tableBits)
 		e.table[prevHashS] = tableEntry{offset: o}
 		e.bTable[prevHashL] = tableEntry{offset: o}

vendor/github.com/klauspost/compress/flate/level5.go

@@ -12,6 +12,7 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 	const (
 		inputMargin            = 12 - 1
 		minNonLiteralBlockSize = 1 + 1 + inputMargin
+		hashShortBytes         = 4
 	)
 	if debugDeflate && e.cur < 0 {
 		panic(fmt.Sprint("e.cur < 0: ", e.cur))
@@ -88,7 +89,7 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 		var l int32
 		var t int32
 		for {
-			nextHashS := hash4x64(cv, tableBits)
+			nextHashS := hashLen(cv, tableBits, hashShortBytes)
 			nextHashL := hash7(cv, tableBits)
 
 			s = nextS
@@ -105,7 +106,7 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 			eLong := &e.bTable[nextHashL]
 			eLong.Cur, eLong.Prev = entry, eLong.Cur
 
-			nextHashS = hash4x64(next, tableBits)
+			nextHashS = hashLen(next, tableBits, hashShortBytes)
 			nextHashL = hash7(next, tableBits)
 
 			t = lCandidate.Cur.offset - e.cur
@@ -191,14 +192,21 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 
 		// Try to locate a better match by checking the end of best match...
 		if sAt := s + l; l < 30 && sAt < sLimit {
+			// Allow some bytes at the beginning to mismatch.
+			// Sweet spot is 2/3 bytes depending on input.
+			// 3 is only a little better when it is but sometimes a lot worse.
+			// The skipped bytes are tested in Extend backwards,
+			// and still picked up as part of the match if they do.
+			const skipBeginning = 2
 			eLong := e.bTable[hash7(load6432(src, sAt), tableBits)].Cur.offset
-			// Test current
+			t2 := eLong - e.cur - l + skipBeginning
-			t2 := eLong - e.cur - l
+			s2 := s + skipBeginning
-			off := s - t2
+			off := s2 - t2
 			if t2 >= 0 && off < maxMatchOffset && off > 0 {
-				if l2 := e.matchlenLong(s, t2, src); l2 > l {
+				if l2 := e.matchlenLong(s2, t2, src); l2 > l {
 					t = t2
 					l = l2
+					s = s2
 				}
 			}
 		}
@@ -250,7 +258,7 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 			if i < s-1 {
 				cv := load6432(src, i)
 				t := tableEntry{offset: i + e.cur}
-				e.table[hash4x64(cv, tableBits)] = t
+				e.table[hashLen(cv, tableBits, hashShortBytes)] = t
 				eLong := &e.bTable[hash7(cv, tableBits)]
 				eLong.Cur, eLong.Prev = t, eLong.Cur
 
@@ -263,7 +271,7 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 				// We only have enough bits for a short entry at i+2
 				cv >>= 8
 				t = tableEntry{offset: t.offset + 1}
-				e.table[hash4x64(cv, tableBits)] = t
+				e.table[hashLen(cv, tableBits, hashShortBytes)] = t
 
 				// Skip one - otherwise we risk hitting 's'
 				i += 4
@@ -273,7 +281,7 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 					t2 := tableEntry{offset: t.offset + 1}
 					eLong := &e.bTable[hash7(cv, tableBits)]
 					eLong.Cur, eLong.Prev = t, eLong.Cur
-					e.table[hash4u(uint32(cv>>8), tableBits)] = t2
+					e.table[hashLen(cv>>8, tableBits, hashShortBytes)] = t2
 				}
 			}
 		}
@@ -282,7 +290,7 @@ func (e *fastEncL5) Encode(dst *tokens, src []byte) {
 		// compression we first update the hash table at s-1 and at s.
 		x := load6432(src, s-1)
 		o := e.cur + s - 1
-		prevHashS := hash4x64(x, tableBits)
+		prevHashS := hashLen(x, tableBits, hashShortBytes)
 		prevHashL := hash7(x, tableBits)
 		e.table[prevHashS] = tableEntry{offset: o}
 		eLong := &e.bTable[prevHashL]

vendor/github.com/klauspost/compress/flate/level6.go

@@ -12,6 +12,7 @@ func (e *fastEncL6) Encode(dst *tokens, src []byte) {
 	const (
 		inputMargin            = 12 - 1
 		minNonLiteralBlockSize = 1 + 1 + inputMargin
+		hashShortBytes         = 4
 	)
 	if debugDeflate && e.cur < 0 {
 		panic(fmt.Sprint("e.cur < 0: ", e.cur))
@@ -90,7 +91,7 @@ func (e *fastEncL6) Encode(dst *tokens, src []byte) {
 		var l int32
 		var t int32
 		for {
-			nextHashS := hash4x64(cv, tableBits)
+			nextHashS := hashLen(cv, tableBits, hashShortBytes)
 			nextHashL := hash7(cv, tableBits)
 			s = nextS
 			nextS = s + doEvery + (s-nextEmit)>>skipLog
@@ -107,7 +108,7 @@ func (e *fastEncL6) Encode(dst *tokens, src []byte) {
 			eLong.Cur, eLong.Prev = entry, eLong.Cur
 
 			// Calculate hashes of 'next'
-			nextHashS = hash4x64(next, tableBits)
+			nextHashS = hashLen(next, tableBits, hashShortBytes)
 			nextHashL = hash7(next, tableBits)
 
 			t = lCandidate.Cur.offset - e.cur
@@ -213,24 +214,33 @@ func (e *fastEncL6) Encode(dst *tokens, src []byte) {
 
 		// Try to locate a better match by checking the end-of-match...
 		if sAt := s + l; sAt < sLimit {
+			// Allow some bytes at the beginning to mismatch.
+			// Sweet spot is 2/3 bytes depending on input.
+			// 3 is only a little better when it is but sometimes a lot worse.
+			// The skipped bytes are tested in Extend backwards,
+			// and still picked up as part of the match if they do.
+			const skipBeginning = 2
 			eLong := &e.bTable[hash7(load6432(src, sAt), tableBits)]
 			// Test current
-			t2 := eLong.Cur.offset - e.cur - l
+			t2 := eLong.Cur.offset - e.cur - l + skipBeginning
-			off := s - t2
+			s2 := s + skipBeginning
+			off := s2 - t2
 			if off < maxMatchOffset {
 				if off > 0 && t2 >= 0 {
-					if l2 := e.matchlenLong(s, t2, src); l2 > l {
+					if l2 := e.matchlenLong(s2, t2, src); l2 > l {
 						t = t2
 						l = l2
+						s = s2
 					}
 				}
 				// Test next:
-				t2 = eLong.Prev.offset - e.cur - l
+				t2 = eLong.Prev.offset - e.cur - l + skipBeginning
-				off := s - t2
+				off := s2 - t2
 				if off > 0 && off < maxMatchOffset && t2 >= 0 {
-					if l2 := e.matchlenLong(s, t2, src); l2 > l {
+					if l2 := e.matchlenLong(s2, t2, src); l2 > l {
 						t = t2
 						l = l2
+						s = s2
 					}
 				}
 			}
@@ -277,7 +287,7 @@ func (e *fastEncL6) Encode(dst *tokens, src []byte) {
 			// Index after match end.
 			for i := nextS + 1; i < int32(len(src))-8; i += 2 {
 				cv := load6432(src, i)
-				e.table[hash4x64(cv, tableBits)] = tableEntry{offset: i + e.cur}
+				e.table[hashLen(cv, tableBits, hashShortBytes)] = tableEntry{offset: i + e.cur}
 				eLong := &e.bTable[hash7(cv, tableBits)]
 				eLong.Cur, eLong.Prev = tableEntry{offset: i + e.cur}, eLong.Cur
 			}
@@ -292,7 +302,7 @@ func (e *fastEncL6) Encode(dst *tokens, src []byte) {
 				t2 := tableEntry{offset: t.offset + 1}
 				eLong := &e.bTable[hash7(cv, tableBits)]
 				eLong2 := &e.bTable[hash7(cv>>8, tableBits)]
-				e.table[hash4x64(cv, tableBits)] = t
+				e.table[hashLen(cv, tableBits, hashShortBytes)] = t
 				eLong.Cur, eLong.Prev = t, eLong.Cur
 				eLong2.Cur, eLong2.Prev = t2, eLong2.Cur
 			}

vendor/github.com/klauspost/compress/flate/stateless.go

@@ -86,11 +86,19 @@ func StatelessDeflate(out io.Writer, in []byte, eof bool, dict []byte) error {
 		dict = dict[len(dict)-maxStatelessDict:]
 	}
 
+	// For subsequent loops, keep shallow dict reference to avoid alloc+copy.
+	var inDict []byte
+
 	for len(in) > 0 {
 		todo := in
-		if len(todo) > maxStatelessBlock-len(dict) {
+		if len(inDict) > 0 {
+			if len(todo) > maxStatelessBlock-maxStatelessDict {
+				todo = todo[:maxStatelessBlock-maxStatelessDict]
+			}
+		} else if len(todo) > maxStatelessBlock-len(dict) {
 			todo = todo[:maxStatelessBlock-len(dict)]
 		}
+		inOrg := in
 		in = in[len(todo):]
 		uncompressed := todo
 		if len(dict) > 0 {
@@ -102,7 +110,11 @@ func StatelessDeflate(out io.Writer, in []byte, eof bool, dict []byte) error {
 			todo = combined
 		}
 		// Compress
+		if len(inDict) == 0 {
 			statelessEnc(&dst, todo, int16(len(dict)))
+		} else {
+			statelessEnc(&dst, inDict[:maxStatelessDict+len(todo)], maxStatelessDict)
+		}
 		isEof := eof && len(in) == 0
 
 		if dst.n == 0 {
@@ -119,7 +131,8 @@ func StatelessDeflate(out io.Writer, in []byte, eof bool, dict []byte) error {
 		}
 		if len(in) > 0 {
 			// Retain a dict if we have more
-			dict = todo[len(todo)-maxStatelessDict:]
+			inDict = inOrg[len(uncompressed)-maxStatelessDict:]
+			dict = nil
 			dst.Reset()
 		}
 		if bw.err != nil {

vendor/github.com/klauspost/compress/s2/README.md

@@ -325,35 +325,35 @@ The content compressed in this mode is fully compatible with the standard decode
 
 Snappy vs S2 **compression** speed on 16 core (32 thread) computer, using all threads and a single thread (1 CPU):
 
-| File                                                                                                | S2 speed | S2 Throughput | S2 % smaller | S2 "better" | "better" throughput | "better" % smaller |
+| File                                                                                                    | S2 Speed | S2 Throughput | S2 % smaller | S2 "better" | "better" throughput | "better" % smaller |
-|-----------------------------------------------------------------------------------------------------|----------|---------------|--------------|-------------|---------------------|--------------------|
+|---------------------------------------------------------------------------------------------------------|----------|---------------|--------------|-------------|---------------------|--------------------|
-| [rawstudio-mint14.tar](https://files.klauspost.com/compress/rawstudio-mint14.7z)                    | 12.70x   | 10556 MB/s    | 7.35%        | 4.15x       | 3455 MB/s           | 12.79%             |
+| [rawstudio-mint14.tar](https://files.klauspost.com/compress/rawstudio-mint14.7z)                        | 16.33x   | 10556 MB/s    | 8.0%         | 6.04x       | 5252 MB/s           | 14.7%              |
-| (1 CPU)                                                                                             | 1.14x    | 948 MB/s      | -            | 0.42x       | 349 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.08x    | 940 MB/s      | -            | 0.46x       | 400 MB/s            | -                  |
-| [github-june-2days-2019.json](https://files.klauspost.com/compress/github-june-2days-2019.json.zst) | 17.13x   | 14484 MB/s    | 31.60%       | 10.09x      | 8533 MB/s           | 37.71%             |
+| [github-june-2days-2019.json](https://files.klauspost.com/compress/github-june-2days-2019.json.zst)     | 16.51x   | 15224 MB/s    | 31.70%       | 9.47x       | 8734 MB/s           | 37.71%             |
-| (1 CPU)                                                                                             | 1.33x    | 1127 MB/s     | -            | 0.70x       | 589 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.26x    | 1157 MB/s     | -            | 0.60x       | 556 MB/s            | -                  |
-| [github-ranks-backup.bin](https://files.klauspost.com/compress/github-ranks-backup.bin.zst)         | 15.14x   | 12000 MB/s    | -5.79%       | 6.59x       | 5223 MB/s           | 5.80%              |
+| [github-ranks-backup.bin](https://files.klauspost.com/compress/github-ranks-backup.bin.zst)             | 15.14x   | 12598 MB/s    | -5.76%       | 6.23x       | 5675 MB/s           | 3.62%              |
-| (1 CPU)                                                                                             | 1.11x    | 877 MB/s      | -            | 0.47x       | 370 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.02x    | 932 MB/s      | -            | 0.47x       | 432 MB/s            | -                  |
-| [consensus.db.10gb](https://files.klauspost.com/compress/consensus.db.10gb.zst)                     | 14.62x   | 12116 MB/s    | 15.90%       | 5.35x       | 4430 MB/s           | 16.08%             |
+| [consensus.db.10gb](https://files.klauspost.com/compress/consensus.db.10gb.zst)                         | 11.21x   | 12116 MB/s    | 15.95%       | 3.24x       | 3500 MB/s           | 18.00%             |
-| (1 CPU)                                                                                             | 1.38x    | 1146 MB/s     | -            | 0.38x       | 312 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.05x    | 1135 MB/s     | -            | 0.27x       | 292 MB/s            | -                  |
-| [adresser.json](https://files.klauspost.com/compress/adresser.json.zst)                             | 8.83x    | 17579 MB/s    | 43.86%       | 6.54x       | 13011 MB/s          | 47.23%             |
+| [apache.log](https://files.klauspost.com/compress/apache.log.zst)                                       | 8.55x    | 16673 MB/s    | 20.54%       | 5.85x       | 11420 MB/s          | 24.97%             |
-| (1 CPU)                                                                                             | 1.14x    | 2259 MB/s     | -            | 0.74x       | 1475 MB/s           | -                  |
+| (1 CPU)                                                                                                 | 1.91x    | 1771 MB/s     | -            | 0.53x       | 1041 MB/s           | -                  |
-| [gob-stream](https://files.klauspost.com/compress/gob-stream.7z)                                    | 16.72x   | 14019 MB/s    | 24.02%       | 10.11x      | 8477 MB/s           | 30.48%             |
+| [gob-stream](https://files.klauspost.com/compress/gob-stream.7z)                                        | 15.76x   | 14357 MB/s    | 24.01%       | 8.67x       | 7891 MB/s           | 33.68%             |
-| (1 CPU)                                                                                             | 1.24x    | 1043 MB/s     | -            | 0.70x       | 586 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.17x    | 1064 MB/s     | -            | 0.65x       | 595 MB/s            | -                  |
-| [10gb.tar](http://mattmahoney.net/dc/10gb.html)                                                     | 13.33x   | 9254 MB/s     | 1.84%        | 6.75x       | 4686 MB/s           | 6.72%              |
+| [10gb.tar](http://mattmahoney.net/dc/10gb.html)                                                         | 13.33x   | 9835 MB/s     | 2.34%        | 6.85x       | 4863 MB/s           | 9.96%              |
-| (1 CPU)                                                                                             | 0.97x    | 672 MB/s      | -            | 0.53x       | 366 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 0.97x    | 689 MB/s      | -            | 0.55x       | 387 MB/s            | -                  |
-| sharnd.out.2gb                                                                                      | 2.11x    | 12639 MB/s    | 0.01%        | 1.98x       | 11833 MB/s          | 0.01%              |
+| sharnd.out.2gb                                                                                          | 9.11x    | 13213 MB/s    | 0.01%        | 1.49x       | 9184 MB/s           | 0.01%              |
-| (1 CPU)                                                                                             | 0.93x    | 5594 MB/s     | -            | 1.34x       | 8030 MB/s           | -                  |
+| (1 CPU)                                                                                                 | 0.88x    | 5418 MB/s     | -            | 0.77x       | 5417 MB/s           | -                  |
-| [enwik9](http://mattmahoney.net/dc/textdata.html)                                                   | 19.34x   | 8220 MB/s     | 3.98%        | 7.87x       | 3345 MB/s           | 15.82%             |
+| [sofia-air-quality-dataset csv](https://files.klauspost.com/compress/sofia-air-quality-dataset.tar.zst) | 22.00x   | 11477 MB/s    | 18.73%       | 11.15x      | 5817 MB/s           | 27.88%             |
-| (1 CPU)                                                                                             | 1.06x    | 452 MB/s      | -            | 0.50x       | 213 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.23x    | 642 MB/s      | -            | 0.71x       | 642 MB/s            | -                  |
-| [silesia.tar](http://sun.aei.polsl.pl/~sdeor/corpus/silesia.zip)                                    | 10.48x   | 6124 MB/s     | 5.67%        | 3.76x       | 2197 MB/s           | 12.60%             |
+| [silesia.tar](http://sun.aei.polsl.pl/~sdeor/corpus/silesia.zip)                                        | 11.23x   | 6520 MB/s     | 5.9%         | 5.35x       | 3109 MB/s           | 15.88%             |
-| (1 CPU)                                                                                             | 0.97x    | 568 MB/s      | -            | 0.46x       | 271 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.05x    | 607 MB/s      | -            | 0.52x       | 304 MB/s            | -                  |
-| [enwik10](https://encode.su/threads/3315-enwik10-benchmark-results)                                 | 21.07x   | 9020 MB/s     | 6.36%        | 6.91x       | 2959 MB/s           | 16.95%             |
+| [enwik9](https://files.klauspost.com/compress/enwik9.zst)                                               | 19.28x   | 8440 MB/s     | 4.04%        | 9.31x       | 4076 MB/s           | 18.04%             |
-| (1 CPU)                                                                                             | 1.07x    | 460 MB/s      | -            | 0.51x       | 220 MB/s            | -                  |
+| (1 CPU)                                                                                                 | 1.12x    | 488 MB/s      | -            | 0.57x       | 250 MB/s            | -                  |
 
 ### Legend
 
-* `S2 speed`: Speed of S2 compared to Snappy, using 16 cores and 1 core.
+* `S2 Speed`: Speed of S2 compared to Snappy, using 16 cores and 1 core.
-* `S2 throughput`: Throughput of S2 in MB/s. 
+* `S2 Throughput`: Throughput of S2 in MB/s. 
 * `S2 % smaller`: How many percent of the Snappy output size is S2 better.
 * `S2 "better"`: Speed when enabling "better" compression mode in S2 compared to Snappy. 
 * `"better" throughput`: Speed when enabling "better" compression mode in S2 compared to Snappy. 
@@ -361,7 +361,7 @@ Snappy vs S2 **compression** speed on 16 core (32 thread) computer, using all th
 
 There is a good speedup across the board when using a single thread and a significant speedup when using multiple threads.
 
-Machine generated data gets by far the biggest compression boost, with size being being reduced by up to 45% of Snappy size.
+Machine generated data gets by far the biggest compression boost, with size being reduced by up to 35% of Snappy size.
 
 The "better" compression mode sees a good improvement in all cases, but usually at a performance cost.
 
@@ -404,7 +404,7 @@ The "better" compression mode will actively look for shorter matches, which is w
 Without assembly decompression is also very fast; single goroutine decompression speed. No assembly:
 
 | File                           | S2 Throughput | S2 throughput |
-|--------------------------------|--------------|---------------|
+|--------------------------------|---------------|---------------|
 | consensus.db.10gb.s2           | 1.84x         | 2289.8 MB/s   |
 | 10gb.tar.s2                    | 1.30x         | 867.07 MB/s   |
 | rawstudio-mint14.tar.s2        | 1.66x         | 1329.65 MB/s  |
@@ -451,13 +451,13 @@ For this we use [`webdevdata.org-2015-01-07-subset`](https://files.klauspost.com
 53927 files, total input size: 4,014,735,833 bytes. Single goroutine used.
 
 | *                 | Input      | Output     | Reduction  | MB/s       |
-|-------------------|------------|------------|-----------|--------|
+|-------------------|------------|------------|------------|------------|
-| S2                | 4014735833 | 1059723369 | 73.60%    | **934.34** |
+| S2                | 4014735833 | 1059723369 | 73.60%     | **936.73** |
-| S2 Better         | 4014735833 | 969670507  | 75.85%    | 532.70 |
+| S2 Better         | 4014735833 | 961580539  | 76.05%     | 451.10     |
-| S2 Best           | 4014735833 | 906625668  | **77.85%** | 46.84 |
+| S2 Best           | 4014735833 | 899182886  | **77.60%** | 46.84      |
-| Snappy            | 4014735833 | 1128706759 | 71.89%    | 762.59 |
+| Snappy            | 4014735833 | 1128706759 | 71.89%     | 790.15     |
-| S2, Snappy Output | 4014735833 | 1093821420 | 72.75%    | 908.60 |
+| S2, Snappy Output | 4014735833 | 1093823291 | 72.75%     | 936.60     |
-| LZ4               | 4014735833 | 1079259294 | 73.12%    | 526.94 |
+| LZ4               | 4014735833 | 1063768713 | 73.50%     | 452.02     |
 
 S2 delivers both the best single threaded throughput with regular mode and the best compression rate with "best".
 "Better" mode provides the same compression speed as LZ4 with better compression ratio. 
@@ -489,43 +489,24 @@ AMD64 assembly is use for both S2 and Snappy.
 
 | Absolute Perf         | Snappy size | S2 Size | Snappy Speed | S2 Speed    | Snappy dec  | S2 dec      |
 |-----------------------|-------------|---------|--------------|-------------|-------------|-------------|
-| html                  | 22843       | 21111   | 16246 MB/s   | 17438 MB/s  | 40972 MB/s  | 49263 MB/s  |
+| html                  | 22843       | 20868   | 16246 MB/s   | 18617 MB/s  | 40972 MB/s  | 49263 MB/s  |
-| urls.10K              | 335492      | 287326  | 7943 MB/s    | 9693 MB/s   | 22523 MB/s  | 26484 MB/s  |
+| urls.10K              | 335492      | 286541  | 7943 MB/s    | 10201 MB/s  | 22523 MB/s  | 26484 MB/s  |
-| fireworks.jpeg        | 123034      | 123100  | 349544 MB/s  | 273889 MB/s | 718321 MB/s | 827552 MB/s |
+| fireworks.jpeg        | 123034      | 123100  | 349544 MB/s  | 303228 MB/s | 718321 MB/s | 827552 MB/s |
-| fireworks.jpeg (200B) | 146         | 155     | 8869 MB/s    | 17773 MB/s  | 33691 MB/s  | 52421 MB/s  |
+| fireworks.jpeg (200B) | 146         | 155     | 8869 MB/s    | 20180 MB/s  | 33691 MB/s  | 52421 MB/s  |
-| paper-100k.pdf        | 85304       | 84459   | 167546 MB/s  | 101263 MB/s | 326905 MB/s | 291944 MB/s |
+| paper-100k.pdf        | 85304       | 84202   | 167546 MB/s  | 112988 MB/s | 326905 MB/s | 291944 MB/s |
-| html_x_4              | 92234       | 21113   | 15194 MB/s   | 50670 MB/s  | 30843 MB/s  | 32217 MB/s  |
+| html_x_4              | 92234       | 20870   | 15194 MB/s   | 54457 MB/s  | 30843 MB/s  | 32217 MB/s  |
-| alice29.txt           | 88034       | 85975   | 5936 MB/s    | 6139 MB/s   | 12882 MB/s  | 20044 MB/s  |
+| alice29.txt           | 88034       | 85934   | 5936 MB/s    | 6540 MB/s   | 12882 MB/s  | 20044 MB/s  |
-| asyoulik.txt          | 77503       | 79650   | 5517 MB/s    | 6366 MB/s   | 12735 MB/s  | 22806 MB/s  |
+| asyoulik.txt          | 77503       | 79575   | 5517 MB/s    | 6657 MB/s   | 12735 MB/s  | 22806 MB/s  |
-| lcet10.txt            | 234661      | 220670  | 6235 MB/s    | 6067 MB/s   | 14519 MB/s  | 18697 MB/s  |
+| lcet10.txt            | 234661      | 220383  | 6235 MB/s    | 6303 MB/s   | 14519 MB/s  | 18697 MB/s  |
-| plrabn12.txt          | 319267      | 317985  | 5159 MB/s    | 5726 MB/s   | 11923 MB/s  | 19901 MB/s  |
+| plrabn12.txt          | 319267      | 318196  | 5159 MB/s    | 6074 MB/s   | 11923 MB/s  | 19901 MB/s  |
-| geo.protodata         | 23335       | 18690   | 21220 MB/s   | 26529 MB/s  | 56271 MB/s  | 62540 MB/s  |
+| geo.protodata         | 23335       | 18606   | 21220 MB/s   | 25432 MB/s  | 56271 MB/s  | 62540 MB/s  |
-| kppkn.gtb             | 69526       | 65312   | 9732 MB/s    | 8559 MB/s   | 18491 MB/s  | 18969 MB/s  |
+| kppkn.gtb             | 69526       | 65019   | 9732 MB/s    | 8905 MB/s   | 18491 MB/s  | 18969 MB/s  |
-| alice29.txt (128B)    | 80          | 82      | 6691 MB/s    | 15489 MB/s  | 31883 MB/s  | 38874 MB/s  |
+| alice29.txt (128B)    | 80          | 82      | 6691 MB/s    | 17179 MB/s  | 31883 MB/s  | 38874 MB/s  |
-| alice29.txt (1000B)   | 774         | 774     | 12204 MB/s   | 13000 MB/s  | 48056 MB/s  | 52341 MB/s  |
+| alice29.txt (1000B)   | 774         | 774     | 12204 MB/s   | 13273 MB/s  | 48056 MB/s  | 52341 MB/s  |
-| alice29.txt (10000B)  | 6648        | 6933    | 10044 MB/s   | 12806 MB/s  | 32378 MB/s  | 46322 MB/s  |
+| alice29.txt (10000B)  | 6648        | 6933    | 10044 MB/s   | 12824 MB/s  | 32378 MB/s  | 46322 MB/s  |
-| alice29.txt (20000B)  | 12686       | 13574   | 7733 MB/s    | 11210 MB/s  | 30566 MB/s  | 58969 MB/s  |
+| alice29.txt (20000B)  | 12686       | 13516   | 7733 MB/s    | 12160 MB/s  | 30566 MB/s  | 58969 MB/s  |
 
 
-| Relative Perf         | Snappy size | S2 size improved | S2 Speed | S2 Dec Speed |
-|-----------------------|-------------|------------------|----------|--------------|
-| html                  | 22.31%      | 7.58%            | 1.07x    | 1.20x        |
-| urls.10K              | 47.78%      | 14.36%           | 1.22x    | 1.18x        |
-| fireworks.jpeg        | 99.95%      | -0.05%           | 0.78x    | 1.15x        |
-| fireworks.jpeg (200B) | 73.00%      | -6.16%           | 2.00x    | 1.56x        |
-| paper-100k.pdf        | 83.30%      | 0.99%            | 0.60x    | 0.89x        |
-| html_x_4              | 22.52%      | 77.11%           | 3.33x    | 1.04x        |
-| alice29.txt           | 57.88%      | 2.34%            | 1.03x    | 1.56x        |
-| asyoulik.txt          | 61.91%      | -2.77%           | 1.15x    | 1.79x        |
-| lcet10.txt            | 54.99%      | 5.96%            | 0.97x    | 1.29x        |
-| plrabn12.txt          | 66.26%      | 0.40%            | 1.11x    | 1.67x        |
-| geo.protodata         | 19.68%      | 19.91%           | 1.25x    | 1.11x        |
-| kppkn.gtb             | 37.72%      | 6.06%            | 0.88x    | 1.03x        |
-| alice29.txt (128B)    | 62.50%      | -2.50%           | 2.31x    | 1.22x        |
-| alice29.txt (1000B)   | 77.40%      | 0.00%            | 1.07x    | 1.09x        |
-| alice29.txt (10000B)  | 66.48%      | -4.29%           | 1.27x    | 1.43x        |
-| alice29.txt (20000B)  | 63.43%      | -7.00%           | 1.45x    | 1.93x        |
-
 Speed is generally at or above Snappy. Small blocks gets a significant speedup, although at the expense of size. 
 
 Decompression speed is better than Snappy, except in one case. 
@@ -543,43 +524,24 @@ So individual benchmarks should only be seen as a guideline and the overall pict
 
 | Absolute Perf         | Snappy size | Better Size | Snappy Speed | Better Speed | Snappy dec  | Better dec  |
 |-----------------------|-------------|-------------|--------------|--------------|-------------|-------------|
-| html                  | 22843       | 19833       | 16246 MB/s   | 7731 MB/s    | 40972 MB/s  | 40292 MB/s  |
+| html                  | 22843       | 18972       | 16246 MB/s   | 8621 MB/s    | 40972 MB/s  | 40292 MB/s  |
-| urls.10K              | 335492      | 253529      | 7943 MB/s    | 3980 MB/s    | 22523 MB/s  | 20981 MB/s  |
+| urls.10K              | 335492      | 248079      | 7943 MB/s    | 5104 MB/s    | 22523 MB/s  | 20981 MB/s  |
-| fireworks.jpeg        | 123034      | 123100      | 349544 MB/s  | 9760 MB/s    | 718321 MB/s | 823698 MB/s |
+| fireworks.jpeg        | 123034      | 123100      | 349544 MB/s  | 84429 MB/s   | 718321 MB/s | 823698 MB/s |
-| fireworks.jpeg (200B) | 146         | 142         | 8869 MB/s    | 594 MB/s     | 33691 MB/s  | 30101 MB/s  |
+| fireworks.jpeg (200B) | 146         | 149         | 8869 MB/s    | 7125 MB/s    | 33691 MB/s  | 30101 MB/s  |
-| paper-100k.pdf        | 85304       | 82915       | 167546 MB/s  | 7470 MB/s    | 326905 MB/s | 198869 MB/s |
+| paper-100k.pdf        | 85304       | 82887       | 167546 MB/s  | 11087 MB/s   | 326905 MB/s | 198869 MB/s |
-| html_x_4              | 92234       | 19841       | 15194 MB/s   | 23403 MB/s   | 30843 MB/s  | 30937 MB/s  |
+| html_x_4              | 92234       | 18982       | 15194 MB/s   | 29316 MB/s   | 30843 MB/s  | 30937 MB/s  |
-| alice29.txt           | 88034       | 73218       | 5936 MB/s    | 2945 MB/s    | 12882 MB/s  | 16611 MB/s  |
+| alice29.txt           | 88034       | 71611       | 5936 MB/s    | 3709 MB/s    | 12882 MB/s  | 16611 MB/s  |
-| asyoulik.txt          | 77503       | 66844       | 5517 MB/s    | 2739 MB/s    | 12735 MB/s  | 14975 MB/s  |
+| asyoulik.txt          | 77503       | 65941       | 5517 MB/s    | 3380 MB/s    | 12735 MB/s  | 14975 MB/s  |
-| lcet10.txt            | 234661      | 190589      | 6235 MB/s    | 3099 MB/s    | 14519 MB/s  | 16634 MB/s  |
+| lcet10.txt            | 234661      | 184939      | 6235 MB/s    | 3537 MB/s    | 14519 MB/s  | 16634 MB/s  |
-| plrabn12.txt          | 319267      | 270828      | 5159 MB/s    | 2600 MB/s    | 11923 MB/s  | 13382 MB/s  |
+| plrabn12.txt          | 319267      | 264990      | 5159 MB/s    | 2960 MB/s    | 11923 MB/s  | 13382 MB/s  |
-| geo.protodata         | 23335       | 18278       | 21220 MB/s   | 11208 MB/s   | 56271 MB/s  | 57961 MB/s  |
+| geo.protodata         | 23335       | 17689       | 21220 MB/s   | 10859 MB/s   | 56271 MB/s  | 57961 MB/s  |
-| kppkn.gtb             | 69526       | 61851       | 9732 MB/s    | 4556 MB/s    | 18491 MB/s  | 16524 MB/s  |
+| kppkn.gtb             | 69526       | 55398       | 9732 MB/s    | 5206 MB/s    | 18491 MB/s  | 16524 MB/s  |
-| alice29.txt (128B)    | 80          | 81          | 6691 MB/s    | 529 MB/s     | 31883 MB/s  | 34225 MB/s  |
+| alice29.txt (128B)    | 80          | 78          | 6691 MB/s    | 7422 MB/s    | 31883 MB/s  | 34225 MB/s  |
-| alice29.txt (1000B)   | 774         | 748         | 12204 MB/s   | 1943 MB/s    | 48056 MB/s  | 42068 MB/s  |
+| alice29.txt (1000B)   | 774         | 746         | 12204 MB/s   | 5734 MB/s    | 48056 MB/s  | 42068 MB/s  |
-| alice29.txt (10000B)  | 6648        | 6234        | 10044 MB/s   | 2949 MB/s    | 32378 MB/s  | 28813 MB/s  |
+| alice29.txt (10000B)  | 6648        | 6218        | 10044 MB/s   | 6055 MB/s    | 32378 MB/s  | 28813 MB/s  |
-| alice29.txt (20000B)  | 12686       | 11584       | 7733 MB/s    | 2822 MB/s    | 30566 MB/s  | 27315 MB/s  |
+| alice29.txt (20000B)  | 12686       | 11492       | 7733 MB/s    | 3143 MB/s    | 30566 MB/s  | 27315 MB/s  |
 
 
-| Relative Perf         | Snappy size | Better size | Better Speed | Better dec |
-|-----------------------|-------------|-------------|--------------|------------|
-| html                  | 22.31%      | 13.18%      | 0.48x        | 0.98x      |
-| urls.10K              | 47.78%      | 24.43%      | 0.50x        | 0.93x      |
-| fireworks.jpeg        | 99.95%      | -0.05%      | 0.03x        | 1.15x      |
-| fireworks.jpeg (200B) | 73.00%      | 2.74%       | 0.07x        | 0.89x      |
-| paper-100k.pdf        | 83.30%      | 2.80%       | 0.07x        | 0.61x      |
-| html_x_4              | 22.52%      | 78.49%      | 0.04x        | 1.00x      |
-| alice29.txt           | 57.88%      | 16.83%      | 1.54x        | 1.29x      |
-| asyoulik.txt          | 61.91%      | 13.75%      | 0.50x        | 1.18x      |
-| lcet10.txt            | 54.99%      | 18.78%      | 0.50x        | 1.15x      |
-| plrabn12.txt          | 66.26%      | 15.17%      | 0.50x        | 1.12x      |
-| geo.protodata         | 19.68%      | 21.67%      | 0.50x        | 1.03x      |
-| kppkn.gtb             | 37.72%      | 11.04%      | 0.53x        | 0.89x      |
-| alice29.txt (128B)    | 62.50%      | -1.25%      | 0.47x        | 1.07x      |
-| alice29.txt (1000B)   | 77.40%      | 3.36%       | 0.08x        | 0.88x      |
-| alice29.txt (10000B)  | 66.48%      | 6.23%       | 0.16x        | 0.89x      |
-| alice29.txt (20000B)  | 63.43%      | 8.69%       | 0.29x        | 0.89x      |
-
 Except for the mostly incompressible JPEG image compression is better and usually in the 
 double digits in terms of percentage reduction over Snappy.
 
@@ -605,29 +567,29 @@ Some examples compared on 16 core CPU, amd64 assembly used:
 
 ```
 * enwik10
-Default... 10000000000 -> 4761467548 [47.61%]; 1.098s, 8685.6MB/s
+Default... 10000000000 -> 4759950115 [47.60%]; 1.03s, 9263.0MB/s
-Better...  10000000000 -> 4219438251 [42.19%]; 1.925s, 4954.2MB/s
+Better...  10000000000 -> 4084706676 [40.85%]; 2.16s, 4415.4MB/s
-Best...    10000000000 -> 3627364337 [36.27%]; 43.051s, 221.5MB/s
+Best...    10000000000 -> 3615520079 [36.16%]; 42.259s, 225.7MB/s
 
 * github-june-2days-2019.json
-Default... 6273951764 -> 1043196283 [16.63%]; 431ms, 13882.3MB/s
+Default... 6273951764 -> 1041700255 [16.60%]; 431ms, 13882.3MB/s
-Better...  6273951764 -> 949146808 [15.13%]; 547ms, 10938.4MB/s
+Better...  6273951764 -> 945841238 [15.08%]; 547ms, 10938.4MB/s
-Best...    6273951764 -> 832855506 [13.27%]; 9.455s, 632.8MB/s
+Best...    6273951764 -> 826392576 [13.17%]; 9.455s, 632.8MB/s
 
 * nyc-taxi-data-10M.csv
-Default... 3325605752 -> 1095998837 [32.96%]; 324ms, 9788.7MB/s
+Default... 3325605752 -> 1093516949 [32.88%]; 324ms, 9788.7MB/s
-Better...  3325605752 -> 954776589 [28.71%]; 491ms, 6459.4MB/s
+Better...  3325605752 -> 885394158 [26.62%]; 491ms, 6459.4MB/s
-Best...    3325605752 -> 779098746 [23.43%]; 8.29s, 382.6MB/s
+Best...    3325605752 -> 773681257 [23.26%]; 8.29s, 412.0MB/s
 
 * 10gb.tar
-Default... 10065157632 -> 5916578242 [58.78%]; 1.028s, 9337.4MB/s
+Default... 10065157632 -> 5915541066 [58.77%]; 1.028s, 9337.4MB/s
-Better...  10065157632 -> 5649207485 [56.13%]; 1.597s, 6010.6MB/s
+Better...  10065157632 -> 5453844650 [54.19%]; 1.597s, 4862.7MB/s
-Best...    10065157632 -> 5208719802 [51.75%]; 32.78s, 292.8MB/
+Best...    10065157632 -> 5192495021 [51.59%]; 32.78s, 308.2MB/
 
 * consensus.db.10gb
-Default... 10737418240 -> 4562648848 [42.49%]; 882ms, 11610.0MB/s
+Default... 10737418240 -> 4549762344 [42.37%]; 882ms, 12118.4MB/s
-Better...  10737418240 -> 4542428129 [42.30%]; 1.533s, 6679.7MB/s
+Better...  10737418240 -> 4438535064 [41.34%]; 1.533s, 3500.9MB/s
-Best...    10737418240 -> 4244773384 [39.53%]; 42.96s, 238.4MB/s
+Best...    10737418240 -> 4210602774 [39.21%]; 42.96s, 254.4MB/s
 ```
 
 Decompression speed should be around the same as using the 'better' compression mode. 
@@ -649,7 +611,7 @@ Snappy compatible blocks can be generated with the S2 encoder.
 Compression and speed is typically a bit better `MaxEncodedLen` is also smaller for smaller memory usage. Replace 
 
 | Snappy                    | S2 replacement        |
-|----------------------------|-------------------------|
+|---------------------------|-----------------------|
 | snappy.Encode(...)        | s2.EncodeSnappy(...)  |
 | snappy.MaxEncodedLen(...) | s2.MaxEncodedLen(...) |
 
@@ -661,7 +623,7 @@ Comparison of [`webdevdata.org-2015-01-07-subset`](https://files.klauspost.com/c
 53927 files, total input size: 4,014,735,833 bytes. amd64, single goroutine used:
 
 | Encoder               | Size       | MB/s       | Reduction  |
-|-----------------------|------------|------------|------------
+|-----------------------|------------|------------|------------|
 | snappy.Encode         | 1128706759 | 725.59     | 71.89%     |
 | s2.EncodeSnappy       | 1093823291 | **899.16** | 72.75%     |
 | s2.EncodeSnappyBetter | 1001158548 | 578.49     | 75.06%     |
@@ -835,6 +797,13 @@ This is done using the regular "Skip" function:
 
 This will ensure that we are at exactly the offset we want, and reading from `dec` will start at the requested offset.
 
+# Compact storage
+
+For compact storage [RemoveIndexHeaders](https://pkg.go.dev/github.com/klauspost/compress/s2#RemoveIndexHeaders) can be used to remove any redundant info from 
+a serialized index. If you remove the header it must be restored before [Loading](https://pkg.go.dev/github.com/klauspost/compress/s2#Index.Load).
+
+This is expected to save 20 bytes. These can be restored using [RestoreIndexHeaders](https://pkg.go.dev/github.com/klauspost/compress/s2#RestoreIndexHeaders). This removes a layer of security, but is the most compact representation. Returns nil if headers contains errors.
+
 ## Index Format:
 
 Each block is structured as a snappy skippable block, with the chunk ID 0x99.
@@ -845,7 +814,7 @@ Numbers are stored as fixed size little endian values or [zigzag encoded](https:
 with un-encoded value length of 64 bits, unless other limits are specified. 
 
 | Content                              | Format                                                                                                                        |
-|---------------------------------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------------|
+|--------------------------------------|-------------------------------------------------------------------------------------------------------------------------------|
 | ID, `[1]byte`                        | Always 0x99.                                                                                                                  |
 | Data Length, `[3]byte`               | 3 byte little-endian length of the chunk in bytes, following this.                                                            |
 | Header `[6]byte`                     | Header, must be `[115, 50, 105, 100, 120, 0]` or in text: "s2idx\x00".                                                        |
@@ -929,6 +898,7 @@ To decode from any given uncompressed offset `(wantOffset)`:
 
 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`.
@@ -951,10 +921,11 @@ The length is specified by reading the 3-bit length specified in the tag and dec
 | 7      | 65540 + read 3 bytes |
 
 This allows any repeat offset + length to be represented by 2 to 5 bytes.
+It also allows to emit matches longer than 64 bytes with one copy + one repeat instead of several 64 byte copies.
 
 Lengths are stored as little endian values.
 
-The first copy of a block cannot be a repeat offset and the offset is not carried across blocks in streams.
+The first copy of a block cannot be a repeat offset and the offset is reset on every block in streams.
 
 Default streaming block size is 1MB.
 

vendor/github.com/klauspost/compress/s2/decode.go

@@ -952,8 +952,12 @@ func (r *Reader) ReadSeeker(random bool, index []byte) (*ReadSeeker, error) {
 // Seek allows seeking in compressed data.
 func (r *ReadSeeker) Seek(offset int64, whence int) (int64, error) {
 	if r.err != nil {
+		if !errors.Is(r.err, io.EOF) {
 			return 0, r.err
 		}
+		// Reset on EOF
+		r.err = nil
+	}
 	if offset == 0 && whence == io.SeekCurrent {
 		return r.blockStart + int64(r.i), nil
 	}

vendor/github.com/klauspost/compress/s2/decode_other.go

@@ -28,6 +28,9 @@ func s2Decode(dst, src []byte) int {
 
 	// As long as we can read at least 5 bytes...
 	for s < len(src)-5 {
+		// Removing bounds checks is SLOWER, when if doing
+		// in := src[s:s+5]
+		// Checked on Go 1.18
 		switch src[s] & 0x03 {
 		case tagLiteral:
 			x := uint32(src[s] >> 2)
@@ -38,14 +41,19 @@ func s2Decode(dst, src []byte) int {
 				s += 2
 				x = uint32(src[s-1])
 			case x == 61:
+				in := src[s : s+3]
+				x = uint32(in[1]) | uint32(in[2])<<8
 				s += 3
-				x = uint32(src[s-2]) | uint32(src[s-1])<<8
 			case x == 62:
+				in := src[s : s+4]
+				// Load as 32 bit and shift down.
+				x = uint32(in[0]) | uint32(in[1])<<8 | uint32(in[2])<<16 | uint32(in[3])<<24
+				x >>= 8
 				s += 4
-				x = uint32(src[s-3]) | uint32(src[s-2])<<8 | uint32(src[s-1])<<16
 			case x == 63:
+				in := src[s : s+5]
+				x = uint32(in[1]) | uint32(in[2])<<8 | uint32(in[3])<<16 | uint32(in[4])<<24
 				s += 5
-				x = uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24
 			}
 			length = int(x) + 1
 			if length > len(dst)-d || length > len(src)-s || (strconv.IntSize == 32 && length <= 0) {
@@ -62,8 +70,8 @@ func s2Decode(dst, src []byte) int {
 
 		case tagCopy1:
 			s += 2
-			length = int(src[s-2]) >> 2 & 0x7
 			toffset := int(uint32(src[s-2])&0xe0<<3 | uint32(src[s-1]))
+			length = int(src[s-2]) >> 2 & 0x7
 			if toffset == 0 {
 				if debug {
 					fmt.Print("(repeat) ")
@@ -71,14 +79,16 @@ func s2Decode(dst, src []byte) int {
 				// keep last offset
 				switch length {
 				case 5:
+					length = int(src[s]) + 4
 					s += 1
-					length = int(uint32(src[s-1])) + 4
 				case 6:
+					in := src[s : s+2]
+					length = int(uint32(in[0])|(uint32(in[1])<<8)) + (1 << 8)
 					s += 2
-					length = int(uint32(src[s-2])|(uint32(src[s-1])<<8)) + (1 << 8)
 				case 7:
+					in := src[s : s+3]
+					length = int((uint32(in[2])<<16)|(uint32(in[1])<<8)|uint32(in[0])) + (1 << 16)
 					s += 3
-					length = int(uint32(src[s-3])|(uint32(src[s-2])<<8)|(uint32(src[s-1])<<16)) + (1 << 16)
 				default: // 0-> 4
 				}
 			} else {
@@ -86,14 +96,16 @@ func s2Decode(dst, src []byte) int {
 			}
 			length += 4
 		case tagCopy2:
+			in := src[s : s+3]
+			offset = int(uint32(in[1]) | uint32(in[2])<<8)
+			length = 1 + int(in[0])>>2
 			s += 3
-			length = 1 + int(src[s-3])>>2
-			offset = int(uint32(src[s-2]) | uint32(src[s-1])<<8)
 
 		case tagCopy4:
+			in := src[s : s+5]
+			offset = int(uint32(in[1]) | uint32(in[2])<<8 | uint32(in[3])<<16 | uint32(in[4])<<24)
+			length = 1 + int(in[0])>>2
 			s += 5
-			length = 1 + int(src[s-5])>>2
-			offset = int(uint32(src[s-4]) | uint32(src[s-3])<<8 | uint32(src[s-2])<<16 | uint32(src[s-1])<<24)
 		}
 
 		if offset <= 0 || d < offset || length > len(dst)-d {

vendor/github.com/klauspost/compress/s2/encode_all.go

@@ -58,6 +58,7 @@ func encodeGo(dst, src []byte) []byte {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockGo(dst, src []byte) (d int) {

vendor/github.com/klauspost/compress/s2/encode_amd64.go

@@ -8,6 +8,7 @@
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlock(dst, src []byte) (d int) {
@@ -43,6 +44,7 @@ func encodeBlock(dst, src []byte) (d int) {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockBetter(dst, src []byte) (d int) {
@@ -78,6 +80,7 @@ func encodeBlockBetter(dst, src []byte) (d int) {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockSnappy(dst, src []byte) (d int) {
@@ -112,6 +115,7 @@ func encodeBlockSnappy(dst, src []byte) (d int) {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockBetterSnappy(dst, src []byte) (d int) {

vendor/github.com/klauspost/compress/s2/encode_best.go

@@ -15,6 +15,7 @@
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockBest(dst, src []byte) (d int) {
@@ -176,14 +177,21 @@ type match struct {
 						best = bestOf(best, matchAt(getPrev(nextLong), s, uint32(cv), false))
 					}
 					// Search for a match at best match end, see if that is better.
-					if sAt := best.s + best.length; sAt < sLimit {
+					// Allow some bytes at the beginning to mismatch.
-						sBack := best.s
+					// Sweet spot is around 1-2 bytes, but depends on input.
-						backL := best.length
+					// The skipped bytes are tested in Extend backwards,
+					// and still picked up as part of the match if they do.
+					const skipBeginning = 2
+					const skipEnd = 1
+					if sAt := best.s + best.length - skipEnd; sAt < sLimit {
+
+						sBack := best.s + skipBeginning - skipEnd
+						backL := best.length - skipBeginning
 						// Load initial values
 						cv = load64(src, sBack)
-						// Search for mismatch
+
+						// Grab candidates...
 						next := lTable[hash8(load64(src, sAt), lTableBits)]
-						//next := sTable[hash4(load64(src, sAt), sTableBits)]
 
 						if checkAt := getCur(next) - backL; checkAt > 0 {
 							best = bestOf(best, matchAt(checkAt, sBack, uint32(cv), false))
@@ -191,6 +199,16 @@ type match struct {
 						if checkAt := getPrev(next) - backL; checkAt > 0 {
 							best = bestOf(best, matchAt(checkAt, sBack, uint32(cv), false))
 						}
+						// Disabled: Extremely small gain
+						if false {
+							next = sTable[hash4(load64(src, sAt), sTableBits)]
+							if checkAt := getCur(next) - backL; checkAt > 0 {
+								best = bestOf(best, matchAt(checkAt, sBack, uint32(cv), false))
+							}
+							if checkAt := getPrev(next) - backL; checkAt > 0 {
+								best = bestOf(best, matchAt(checkAt, sBack, uint32(cv), false))
+							}
+						}
 					}
 				}
 			}
@@ -288,6 +306,7 @@ type match struct {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockBestSnappy(dst, src []byte) (d int) {
@@ -546,6 +565,7 @@ type match struct {
 // emitCopySize returns the size to encode the offset+length
 //
 // It assumes that:
+//
 //	1 <= offset && offset <= math.MaxUint32
 //	4 <= length && length <= 1 << 24
 func emitCopySize(offset, length int) int {
@@ -584,6 +604,7 @@ func emitCopySize(offset, length int) int {
 // 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 {

vendor/github.com/klauspost/compress/s2/encode_better.go

@@ -42,6 +42,7 @@ func hash8(u uint64, h uint8) uint32 {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockBetterGo(dst, src []byte) (d int) {
@@ -56,7 +57,7 @@ func encodeBlockBetterGo(dst, src []byte) (d int) {
 	// Initialize the hash tables.
 	const (
 		// Long hash matches.
-		lTableBits    = 16
+		lTableBits    = 17
 		maxLTableSize = 1 << lTableBits
 
 		// Short hash matches.
@@ -97,9 +98,26 @@ func encodeBlockBetterGo(dst, src []byte) (d int) {
 			lTable[hashL] = uint32(s)
 			sTable[hashS] = uint32(s)
 
+			valLong := load64(src, candidateL)
+			valShort := load64(src, candidateS)
+
+			// If long matches at least 8 bytes, use that.
+			if cv == valLong {
+				break
+			}
+			if cv == valShort {
+				candidateL = candidateS
+				break
+			}
+
 			// Check repeat at offset checkRep.
 			const checkRep = 1
-			if false && uint32(cv>>(checkRep*8)) == load32(src, s-repeat+checkRep) {
+			// Minimum length of a repeat. Tested with various values.
+			// While 4-5 offers improvements in some, 6 reduces
+			// regressions significantly.
+			const wantRepeatBytes = 6
+			const repeatMask = ((1 << (wantRepeatBytes * 8)) - 1) << (8 * checkRep)
+			if false && repeat > 0 && cv&repeatMask == load64(src, s-repeat)&repeatMask {
 				base := s + checkRep
 				// Extend back
 				for i := base - repeat; base > nextEmit && i > 0 && src[i-1] == src[base-1]; {
@@ -109,8 +127,8 @@ func encodeBlockBetterGo(dst, src []byte) (d int) {
 				d += emitLiteral(dst[d:], src[nextEmit:base])
 
 				// Extend forward
-				candidate := s - repeat + 4 + checkRep
+				candidate := s - repeat + wantRepeatBytes + checkRep
-				s += 4 + checkRep
+				s += wantRepeatBytes + checkRep
 				for s < len(src) {
 					if len(src)-s < 8 {
 						if src[s] == src[candidate] {
@@ -127,28 +145,40 @@ func encodeBlockBetterGo(dst, src []byte) (d int) {
 					s += 8
 					candidate += 8
 				}
-				if nextEmit > 0 {
 				// same as `add := emitCopy(dst[d:], repeat, s-base)` but skips storing offset.
 				d += emitRepeat(dst[d:], repeat, s-base)
-				} else {
-					// First match, cannot be repeat.
-					d += emitCopy(dst[d:], repeat, s-base)
-				}
 				nextEmit = s
 				if s >= sLimit {
 					goto emitRemainder
 				}
+				// Index in-between
+				index0 := base + 1
+				index1 := s - 2
+
+				cv = load64(src, s)
+				for index0 < index1 {
+					cv0 := load64(src, index0)
+					cv1 := load64(src, index1)
+					lTable[hash7(cv0, lTableBits)] = uint32(index0)
+					sTable[hash4(cv0>>8, sTableBits)] = uint32(index0 + 1)
+
+					lTable[hash7(cv1, lTableBits)] = uint32(index1)
+					sTable[hash4(cv1>>8, sTableBits)] = uint32(index1 + 1)
+					index0 += 2
+					index1 -= 2
+				}
 
 				cv = load64(src, s)
 				continue
 			}
 
-			if uint32(cv) == load32(src, candidateL) {
+			// Long likely matches 7, so take that.
+			if uint32(cv) == uint32(valLong) {
 				break
 			}
 
 			// Check our short candidate
-			if uint32(cv) == load32(src, candidateS) {
+			if uint32(cv) == uint32(valShort) {
 				// Try a long candidate at s+1
 				hashL = hash7(cv>>8, lTableBits)
 				candidateL = int(lTable[hashL])
@@ -227,21 +257,29 @@ func encodeBlockBetterGo(dst, src []byte) (d int) {
 			// Do we have space for more, if not bail.
 			return 0
 		}
-		// Index match start+1 (long) and start+2 (short)
+
+		// Index short & long
 		index0 := base + 1
-		// Index match end-2 (long) and end-1 (short)
 		index1 := s - 2
 
 		cv0 := load64(src, index0)
 		cv1 := load64(src, index1)
-		cv = load64(src, s)
 		lTable[hash7(cv0, lTableBits)] = uint32(index0)
-		lTable[hash7(cv0>>8, lTableBits)] = uint32(index0 + 1)
-		lTable[hash7(cv1, lTableBits)] = uint32(index1)
-		lTable[hash7(cv1>>8, lTableBits)] = uint32(index1 + 1)
 		sTable[hash4(cv0>>8, sTableBits)] = uint32(index0 + 1)
-		sTable[hash4(cv0>>16, sTableBits)] = uint32(index0 + 2)
+
+		lTable[hash7(cv1, lTableBits)] = uint32(index1)
 		sTable[hash4(cv1>>8, sTableBits)] = uint32(index1 + 1)
+		index0 += 1
+		index1 -= 1
+		cv = load64(src, s)
+
+		// index every second long in between.
+		for index0 < index1 {
+			lTable[hash7(load64(src, index0), lTableBits)] = uint32(index0)
+			lTable[hash7(load64(src, index1), lTableBits)] = uint32(index1)
+			index0 += 2
+			index1 -= 2
+		}
 	}
 
 emitRemainder:
@@ -260,6 +298,7 @@ func encodeBlockBetterGo(dst, src []byte) (d int) {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src)) &&
 //	minNonLiteralBlockSize <= len(src) && len(src) <= maxBlockSize
 func encodeBlockBetterSnappyGo(dst, src []byte) (d int) {
@@ -402,21 +441,29 @@ func encodeBlockBetterSnappyGo(dst, src []byte) (d int) {
 			// Do we have space for more, if not bail.
 			return 0
 		}
-		// Index match start+1 (long) and start+2 (short)
+
+		// Index short & long
 		index0 := base + 1
-		// Index match end-2 (long) and end-1 (short)
 		index1 := s - 2
 
 		cv0 := load64(src, index0)
 		cv1 := load64(src, index1)
-		cv = load64(src, s)
 		lTable[hash7(cv0, lTableBits)] = uint32(index0)
-		lTable[hash7(cv0>>8, lTableBits)] = uint32(index0 + 1)
-		lTable[hash7(cv1, lTableBits)] = uint32(index1)
-		lTable[hash7(cv1>>8, lTableBits)] = uint32(index1 + 1)
 		sTable[hash4(cv0>>8, sTableBits)] = uint32(index0 + 1)
-		sTable[hash4(cv0>>16, sTableBits)] = uint32(index0 + 2)
+
+		lTable[hash7(cv1, lTableBits)] = uint32(index1)
 		sTable[hash4(cv1>>8, sTableBits)] = uint32(index1 + 1)
+		index0 += 1
+		index1 -= 1
+		cv = load64(src, s)
+
+		// index every second long in between.
+		for index0 < index1 {
+			lTable[hash7(load64(src, index0), lTableBits)] = uint32(index0)
+			lTable[hash7(load64(src, index1), lTableBits)] = uint32(index1)
+			index0 += 2
+			index1 -= 2
+		}
 	}
 
 emitRemainder:

vendor/github.com/klauspost/compress/s2/encode_go.go

@@ -12,6 +12,7 @@
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src))
 func encodeBlock(dst, src []byte) (d int) {
 	if len(src) < minNonLiteralBlockSize {
@@ -25,6 +26,7 @@ func encodeBlock(dst, src []byte) (d int) {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src))
 func encodeBlockBetter(dst, src []byte) (d int) {
 	return encodeBlockBetterGo(dst, src)
@@ -35,6 +37,7 @@ func encodeBlockBetter(dst, src []byte) (d int) {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src))
 func encodeBlockBetterSnappy(dst, src []byte) (d int) {
 	return encodeBlockBetterSnappyGo(dst, src)
@@ -45,6 +48,7 @@ func encodeBlockBetterSnappy(dst, src []byte) (d int) {
 // been written.
 //
 // It also assumes that:
+//
 //	len(dst) >= MaxEncodedLen(len(src))
 func encodeBlockSnappy(dst, src []byte) (d int) {
 	if len(src) < minNonLiteralBlockSize {
@@ -56,6 +60,7 @@ func encodeBlockSnappy(dst, src []byte) (d int) {
 // emitLiteral writes a literal chunk and returns the number of bytes written.
 //
 // It assumes that:
+//
 //	dst is long enough to hold the encoded bytes
 //	0 <= len(lit) && len(lit) <= math.MaxUint32
 func emitLiteral(dst, lit []byte) int {
@@ -146,6 +151,7 @@ func emitRepeat(dst []byte, offset, length int) int {
 // emitCopy writes a copy chunk and returns the number of bytes written.
 //
 // It assumes that:
+//
 //	dst is long enough to hold the encoded bytes
 //	1 <= offset && offset <= math.MaxUint32
 //	4 <= length && length <= 1 << 24
@@ -214,6 +220,7 @@ func emitCopy(dst []byte, offset, length int) int {
 // emitCopyNoRepeat writes a copy chunk and returns the number of bytes written.
 //
 // It assumes that:
+//
 //	dst is long enough to hold the encoded bytes
 //	1 <= offset && offset <= math.MaxUint32
 //	4 <= length && length <= 1 << 24
@@ -273,8 +280,8 @@ func emitCopyNoRepeat(dst []byte, offset, length int) int {
 // matchLen returns how many bytes match in a and b
 //
 // It assumes that:
-//   len(a) <= len(b)
 //
+//	len(a) <= len(b)
 func matchLen(a []byte, b []byte) int {
 	b = b[:len(a)]
 	var checked int

vendor/github.com/klauspost/compress/s2/encodeblock_amd64.go

@@ -1,7 +1,6 @@
 // Code generated by command: go run gen.go -out ../encodeblock_amd64.s -stubs ../encodeblock_amd64.go -pkg=s2. DO NOT EDIT.
 
 //go:build !appengine && !noasm && gc && !noasm
-// +build !appengine,!noasm,gc,!noasm
 
 package s2
 
@@ -150,6 +149,7 @@ func encodeSnappyBetterBlockAsm8B(dst []byte, src []byte) int
 // emitLiteral writes a literal chunk and returns the number of bytes written.
 //
 // It assumes that:
+//
 //	dst is long enough to hold the encoded bytes with margin of 0 bytes
 //	0 <= len(lit) && len(lit) <= math.MaxUint32
 //
@@ -165,6 +165,7 @@ func emitRepeat(dst []byte, offset int, length int) int
 // emitCopy writes a copy chunk and returns the number of bytes written.
 //
 // It assumes that:
+//
 //	dst is long enough to hold the encoded bytes
 //	1 <= offset && offset <= math.MaxUint32
 //	4 <= length && length <= 1 << 24
@@ -175,6 +176,7 @@ func emitCopy(dst []byte, offset int, length int) int
 // emitCopyNoRepeat writes a copy chunk and returns the number of bytes written.
 //
 // It assumes that:
+//
 //	dst is long enough to hold the encoded bytes
 //	1 <= offset && offset <= math.MaxUint32
 //	4 <= length && length <= 1 << 24
@@ -185,6 +187,7 @@ func emitCopyNoRepeat(dst []byte, offset int, length int) int
 // matchLen returns how many bytes match in a and b
 //
 // It assumes that:
+//
 //	len(a) <= len(b)
 //
 //go:noescape

vendor/github.com/klauspost/cpuid/v2/README.md

@@ -17,9 +17,16 @@ Package home: https://github.com/klauspost/cpuid
 ## installing
 
 `go get -u github.com/klauspost/cpuid/v2` using modules.
-
 Drop `v2` for others.
 
+### Homebrew
+
+For macOS/Linux users, you can install via [brew](https://brew.sh/)
+
+```sh
+$ brew install cpuid
+```
+
 ## example
 
 ```Go
@@ -77,10 +84,14 @@ We have Streaming SIMD 2 Extensions
 The `cpuid.CPU` provides access to CPU features. Use `cpuid.CPU.Supports()` to check for CPU features.
 A faster `cpuid.CPU.Has()` is provided which will usually be inlined by the gc compiler.  
 
+To test a larger number of features, they can be combined using `f := CombineFeatures(CMOV, CMPXCHG8, X87, FXSR, MMX, SYSCALL, SSE, SSE2)`, etc.
+This can be using with `cpuid.CPU.HasAll(f)` to quickly test if all features are supported.
+
 Note that for some cpu/os combinations some features will not be detected.
 `amd64` has rather good support and should work reliably on all platforms.
 
-Note that hypervisors may not pass through all CPU features.
+Note that hypervisors may not pass through all CPU features through to the guest OS,
+so even if your host supports a feature it may not be visible on guests.
 
 ## arm64 feature detection
 
@@ -253,6 +264,218 @@ Exit Code 0
 Exit Code 1
 ```
 
+
+## Available flags
+
+### x86 & amd64 
+
+| Feature Flag       | Description                                                                                                                                                                        |
+|--------------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
+| ADX                | Intel ADX (Multi-Precision Add-Carry Instruction Extensions)                                                                                                                       |
+| AESNI              | Advanced Encryption Standard New Instructions                                                                                                                                      |
+| AMD3DNOW           | AMD 3DNOW                                                                                                                                                                          |
+| AMD3DNOWEXT        | AMD 3DNowExt                                                                                                                                                                       |
+| AMXBF16            | Tile computational operations on BFLOAT16 numbers                                                                                                                                  |
+| AMXINT8            | Tile computational operations on 8-bit integers                                                                                                                                    |
+| AMXFP16            | Tile computational operations on FP16 numbers                                                                                                                                      |
+| AMXTILE            | Tile architecture                                                                                                                                                                  |
+| AVX                | AVX functions                                                                                                                                                                      |
+| AVX2               | AVX2 functions                                                                                                                                                                     |
+| AVX512BF16         | AVX-512 BFLOAT16 Instructions                                                                                                                                                      |
+| AVX512BITALG       | AVX-512 Bit Algorithms                                                                                                                                                             |
+| AVX512BW           | AVX-512 Byte and Word Instructions                                                                                                                                                 |
+| AVX512CD           | AVX-512 Conflict Detection Instructions                                                                                                                                            |
+| AVX512DQ           | AVX-512 Doubleword and Quadword Instructions                                                                                                                                       |
+| AVX512ER           | AVX-512 Exponential and Reciprocal Instructions                                                                                                                                    |
+| AVX512F            | AVX-512 Foundation                                                                                                                                                                 |
+| AVX512FP16         | AVX-512 FP16 Instructions                                                                                                                                                          |
+| AVX512IFMA         | AVX-512 Integer Fused Multiply-Add Instructions                                                                                                                                    |
+| AVX512PF           | AVX-512 Prefetch Instructions                                                                                                                                                      |
+| AVX512VBMI         | AVX-512 Vector Bit Manipulation Instructions                                                                                                                                       |
+| AVX512VBMI2        | AVX-512 Vector Bit Manipulation Instructions, Version 2                                                                                                                            |
+| AVX512VL           | AVX-512 Vector Length Extensions                                                                                                                                                   |
+| AVX512VNNI         | AVX-512 Vector Neural Network Instructions                                                                                                                                         |
+| AVX512VP2INTERSECT | AVX-512 Intersect for D/Q                                                                                                                                                          |
+| AVX512VPOPCNTDQ    | AVX-512 Vector Population Count Doubleword and Quadword                                                                                                                            |
+| AVXIFMA            | AVX-IFMA instructions                                                                                                                                                              |
+| AVXNECONVERT       | AVX-NE-CONVERT instructions                                                                                                                                                        |
+| AVXSLOW            | Indicates the CPU performs 2 128 bit operations instead of one                                                                                                                     |
+| AVXVNNI            | AVX (VEX encoded) VNNI neural network instructions                                                                                                                                 |
+| AVXVNNIINT8        | AVX-VNNI-INT8 instructions                                                                                                                                                         |
+| BMI1               | Bit Manipulation Instruction Set 1                                                                                                                                                 |
+| BMI2               | Bit Manipulation Instruction Set 2                                                                                                                                                 |
+| CETIBT             | Intel CET Indirect Branch Tracking                                                                                                                                                 |
+| CETSS              | Intel CET Shadow Stack                                                                                                                                                             |
+| CLDEMOTE           | Cache Line Demote                                                                                                                                                                  |
+| CLMUL              | Carry-less Multiplication                                                                                                                                                          |
+| CLZERO             | CLZERO instruction supported                                                                                                                                                       |
+| CMOV               | i686 CMOV                                                                                                                                                                          |
+| CMPCCXADD          | CMPCCXADD instructions                                                                                                                                                             |
+| CMPSB_SCADBS_SHORT | Fast short CMPSB and SCASB                                                                                                                                                         |
+| CMPXCHG8           | CMPXCHG8 instruction                                                                                                                                                               |
+| CPBOOST            | Core Performance Boost                                                                                                                                                             |
+| CPPC               | AMD: Collaborative Processor Performance Control                                                                                                                                   |
+| CX16               | CMPXCHG16B Instruction                                                                                                                                                             |
+| EFER_LMSLE_UNS     | AMD: =Core::X86::Msr::EFER[LMSLE] is not supported, and MBZ                                                                                                                        |
+| ENQCMD             | Enqueue Command                                                                                                                                                                    |
+| ERMS               | Enhanced REP MOVSB/STOSB                                                                                                                                                           |
+| F16C               | Half-precision floating-point conversion                                                                                                                                           |
+| FLUSH_L1D          | Flush L1D cache                                                                                                                                                                    |
+| FMA3               | Intel FMA 3. Does not imply AVX.                                                                                                                                                   |
+| FMA4               | Bulldozer FMA4 functions                                                                                                                                                           |
+| FP128              | AMD: When set, the internal FP/SIMD execution datapath is 128-bits wide                                                                                                            |
+| FP256              | AMD: When set, the internal FP/SIMD execution datapath is 256-bits wide                                                                                                            |
+| FSRM               | Fast Short Rep Mov                                                                                                                                                                 |
+| FXSR               | FXSAVE, FXRESTOR instructions, CR4 bit 9                                                                                                                                           |
+| FXSROPT            | FXSAVE/FXRSTOR optimizations                                                                                                                                                       |
+| GFNI               | Galois Field New Instructions. May require other features (AVX, AVX512VL,AVX512F) based on usage.                                                                                  |
+| HLE                | Hardware Lock Elision                                                                                                                                                              |
+| HRESET             | If set CPU supports history reset and the IA32_HRESET_ENABLE MSR                                                                                                                   |
+| HTT                | Hyperthreading (enabled)                                                                                                                                                           |
+| HWA                | Hardware assert supported. Indicates support for MSRC001_10                                                                                                                        |
+| HYBRID_CPU         | This part has CPUs of more than one type.                                                                                                                                          |
+| HYPERVISOR         | This bit has been reserved by Intel & AMD for use by hypervisors                                                                                                                   |
+| IA32_ARCH_CAP      | IA32_ARCH_CAPABILITIES MSR (Intel)                                                                                                                                                 |
+| IA32_CORE_CAP      | IA32_CORE_CAPABILITIES MSR                                                                                                                                                         |
+| IBPB               | Indirect Branch Restricted Speculation (IBRS) and Indirect Branch Predictor Barrier (IBPB)                                                                                         |
+| IBRS               | AMD: Indirect Branch Restricted Speculation                                                                                                                                        |
+| IBRS_PREFERRED     | AMD: IBRS is preferred over software solution                                                                                                                                      |
+| IBRS_PROVIDES_SMP  | AMD: IBRS provides Same Mode Protection                                                                                                                                            |
+| IBS                | Instruction Based Sampling (AMD)                                                                                                                                                   |
+| IBSBRNTRGT         | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBSFETCHSAM        | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBSFFV             | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBSOPCNT           | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBSOPCNTEXT        | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBSOPSAM           | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBSRDWROPCNT       | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBSRIPINVALIDCHK   | Instruction Based Sampling Feature (AMD)                                                                                                                                           |
+| IBS_FETCH_CTLX     | AMD: IBS fetch control extended MSR supported                                                                                                                                      |
+| IBS_OPDATA4        | AMD: IBS op data 4 MSR supported                                                                                                                                                   |
+| IBS_OPFUSE         | AMD: Indicates support for IbsOpFuse                                                                                                                                               |
+| IBS_PREVENTHOST    | Disallowing IBS use by the host supported                                                                                                                                          |
+| IBS_ZEN4           | Fetch and Op IBS support IBS extensions added with Zen4                                                                                                                            |
+| INT_WBINVD         | WBINVD/WBNOINVD are interruptible.                                                                                                                                                 |
+| INVLPGB            | NVLPGB and TLBSYNC instruction supported                                                                                                                                           |
+| LAHF               | LAHF/SAHF in long mode                                                                                                                                                             |
+| LAM                | If set, CPU supports Linear Address Masking                                                                                                                                        |
+| LBRVIRT            | LBR virtualization                                                                                                                                                                 |
+| LZCNT              | LZCNT instruction                                                                                                                                                                  |
+| MCAOVERFLOW        | MCA overflow recovery support.                                                                                                                                                     |
+| MCDT_NO            | Processor do not exhibit MXCSR Configuration Dependent Timing behavior and do not need to mitigate it.                                                                             |
+| MCOMMIT            | MCOMMIT instruction supported                                                                                                                                                      |
+| MD_CLEAR           | VERW clears CPU buffers                                                                                                                                                            |
+| MMX                | standard MMX                                                                                                                                                                       |
+| MMXEXT             | SSE integer functions or AMD MMX ext                                                                                                                                               |
+| MOVBE              | MOVBE instruction (big-endian)                                                                                                                                                     |
+| MOVDIR64B          | Move 64 Bytes as Direct Store                                                                                                                                                      |
+| MOVDIRI            | Move Doubleword as Direct Store                                                                                                                                                    |
+| MOVSB_ZL           | Fast Zero-Length MOVSB                                                                                                                                                             |
+| MPX                | Intel MPX (Memory Protection Extensions)                                                                                                                                           |
+| MOVU               | MOVU SSE instructions are more efficient and should be preferred to SSE	MOVL/MOVH. MOVUPS is more efficient than MOVLPS/MOVHPS. MOVUPD is more efficient than MOVLPD/MOVHPD        |
+| MSRIRC             | Instruction Retired Counter MSR available                                                                                                                                          |
+| MSR_PAGEFLUSH      | Page Flush MSR available                                                                                                                                                           |
+| NRIPS              | Indicates support for NRIP save on VMEXIT                                                                                                                                          |
+| NX                 | NX (No-Execute) bit                                                                                                                                                                |
+| OSXSAVE            | XSAVE enabled by OS                                                                                                                                                                |
+| PCONFIG            | PCONFIG for Intel Multi-Key Total Memory Encryption                                                                                                                                |
+| POPCNT             | POPCNT instruction                                                                                                                                                                 |
+| PPIN               | AMD: Protected Processor Inventory Number support. Indicates that Protected Processor Inventory Number (PPIN) capability can be enabled                                            |
+| PREFETCHI          | PREFETCHIT0/1 instructions                                            |
+| PSFD               | AMD: Predictive Store Forward Disable                                                                                                                                              |
+| RDPRU              | RDPRU instruction supported                                                                                                                                                        |
+| RDRAND             | RDRAND instruction is available                                                                                                                                                    |
+| RDSEED             | RDSEED instruction is available                                                                                                                                                    |
+| RDTSCP             | RDTSCP Instruction                                                                                                                                                                 |
+| RTM                | Restricted Transactional Memory                                                                                                                                                    |
+| RTM_ALWAYS_ABORT   | Indicates that the loaded microcode is forcing RTM abort.                                                                                                                          |
+| SERIALIZE          | Serialize Instruction Execution                                                                                                                                                    |
+| SEV                | AMD Secure Encrypted Virtualization supported                                                                                                                                      |
+| SEV_64BIT          | AMD SEV guest execution only allowed from a 64-bit host                                                                                                                            |
+| SEV_ALTERNATIVE    | AMD SEV Alternate Injection supported                                                                                                                                              |
+| SEV_DEBUGSWAP      | Full debug state swap supported for SEV-ES guests                                                                                                                                  |
+| SEV_ES             | AMD SEV Encrypted State supported                                                                                                                                                  |
+| SEV_RESTRICTED     | AMD SEV Restricted Injection supported                                                                                                                                             |
+| SEV_SNP            | AMD SEV Secure Nested Paging supported                                                                                                                                             |
+| SGX                | Software Guard Extensions                                                                                                                                                          |
+| SGXLC              | Software Guard Extensions Launch Control                                                                                                                                           |
+| SHA                | Intel SHA Extensions                                                                                                                                                               |
+| SME                | AMD Secure Memory Encryption supported                                                                                                                                             |
+| SME_COHERENT       | AMD Hardware cache coherency across encryption domains enforced                                                                                                                    |
+| SPEC_CTRL_SSBD     | Speculative Store Bypass Disable                                                                                                                                                   |
+| SRBDS_CTRL         | SRBDS mitigation MSR available                                                                                                                                                     |
+| SSE                | SSE functions                                                                                                                                                                      |
+| SSE2               | P4 SSE functions                                                                                                                                                                   |
+| SSE3               | Prescott SSE3 functions                                                                                                                                                            |
+| SSE4               | Penryn SSE4.1 functions                                                                                                                                                            |
+| SSE42              | Nehalem SSE4.2 functions                                                                                                                                                           |
+| SSE4A              | AMD Barcelona microarchitecture SSE4a instructions                                                                                                                                 |
+| SSSE3              | Conroe SSSE3 functions                                                                                                                                                             |
+| STIBP              | Single Thread Indirect Branch Predictors                                                                                                                                           |
+| STIBP_ALWAYSON     | AMD: Single Thread Indirect Branch Prediction Mode has Enhanced Performance and may be left Always On                                                                              |
+| STOSB_SHORT        | Fast short STOSB                                                                                                                                                                   |
+| SUCCOR             | Software uncorrectable error containment and recovery capability.                                                                                                                  |
+| SVM                | AMD Secure Virtual Machine                                                                                                                                                         |
+| SVMDA              | Indicates support for the SVM decode assists.                                                                                                                                      |
+| SVMFBASID          | SVM, Indicates that TLB flush events, including CR3 writes and CR4.PGE toggles, flush only the current ASID's TLB entries. Also indicates support for the extended VMCBTLB_Control |
+| SVML               | AMD SVM lock. Indicates support for SVM-Lock.                                                                                                                                      |
+| SVMNP              | AMD SVM nested paging                                                                                                                                                              |
+| SVMPF              | SVM pause intercept filter. Indicates support for the pause intercept filter                                                                                                       |
+| SVMPFT             | SVM PAUSE filter threshold. Indicates support for the PAUSE filter cycle count threshold                                                                                           |
+| SYSCALL            | System-Call Extension (SCE): SYSCALL and SYSRET instructions.                                                                                                                      |
+| SYSEE              | SYSENTER and SYSEXIT instructions                                                                                                                                                  |
+| TBM                | AMD Trailing Bit Manipulation                                                                                                                                                      |
+| TLB_FLUSH_NESTED   | AMD: Flushing includes all the nested translations for guest translations                                                                                                          |
+| TME                | Intel Total Memory Encryption. The following MSRs are supported: IA32_TME_CAPABILITY, IA32_TME_ACTIVATE, IA32_TME_EXCLUDE_MASK, and IA32_TME_EXCLUDE_BASE.                         |
+| TOPEXT             | TopologyExtensions: topology extensions support. Indicates support for CPUID Fn8000_001D_EAX_x[N:0]-CPUID Fn8000_001E_EDX.                                                         |
+| TSCRATEMSR         | MSR based TSC rate control. Indicates support for MSR TSC ratio MSRC000_0104                                                                                                       |
+| TSXLDTRK           | Intel TSX Suspend Load Address Tracking                                                                                                                                            |
+| VAES               | Vector AES. AVX(512) versions requires additional checks.                                                                                                                          |
+| VMCBCLEAN          | VMCB clean bits. Indicates support for VMCB clean bits.                                                                                                                            |
+| VMPL               | AMD VM Permission Levels supported                                                                                                                                                 |
+| VMSA_REGPROT       | AMD VMSA Register Protection supported                                                                                                                                             |
+| VMX                | Virtual Machine Extensions                                                                                                                                                         |
+| VPCLMULQDQ         | Carry-Less Multiplication Quadword. Requires AVX for 3 register versions.                                                                                                          |
+| VTE                | AMD Virtual Transparent Encryption supported                                                                                                                                       |
+| WAITPKG            | TPAUSE, UMONITOR, UMWAIT                                                                                                                                                           |
+| WBNOINVD           | Write Back and Do Not Invalidate Cache                                                                                                                                             |
+| X87                | FPU                                                                                                                                                                                |
+| XGETBV1            | Supports XGETBV with ECX = 1                                                                                                                                                       |
+| XOP                | Bulldozer XOP functions                                                                                                                                                            |
+| XSAVE              | XSAVE, XRESTOR, XSETBV, XGETBV                                                                                                                                                     |
+| XSAVEC             | Supports XSAVEC and the compacted form of XRSTOR.                                                                                                                                  |
+| XSAVEOPT           | XSAVEOPT available                                                                                                                                                                 |
+| XSAVES             | Supports XSAVES/XRSTORS and IA32_XSS                                                                                                                                               |
+
+# ARM features:
+
+| Feature Flag | Description                                                      |
+|--------------|------------------------------------------------------------------|
+| AESARM       | AES instructions                                                 |
+| ARMCPUID     | Some CPU ID registers readable at user-level                     |
+| ASIMD        | Advanced SIMD                                                    |
+| ASIMDDP      | SIMD Dot Product                                                 |
+| ASIMDHP      | Advanced SIMD half-precision floating point                      |
+| ASIMDRDM     | Rounding Double Multiply Accumulate/Subtract (SQRDMLAH/SQRDMLSH) |
+| ATOMICS      | Large System Extensions (LSE)                                    |
+| CRC32        | CRC32/CRC32C instructions                                        |
+| DCPOP        | Data cache clean to Point of Persistence (DC CVAP)               |
+| EVTSTRM      | Generic timer                                                    |
+| FCMA         | Floatin point complex number addition and multiplication         |
+| FP           | Single-precision and double-precision floating point             |
+| FPHP         | Half-precision floating point                                    |
+| GPA          | Generic Pointer Authentication                                   |
+| JSCVT        | Javascript-style double->int convert (FJCVTZS)                   |
+| LRCPC        | Weaker release consistency (LDAPR, etc)                          |
+| PMULL        | Polynomial Multiply instructions (PMULL/PMULL2)                  |
+| SHA1         | SHA-1 instructions (SHA1C, etc)                                  |
+| SHA2         | SHA-2 instructions (SHA256H, etc)                                |
+| SHA3         | SHA-3 instructions (EOR3, RAXI, XAR, BCAX)                       |
+| SHA512       | SHA512 instructions                                              |
+| SM3          | SM3 instructions                                                 |
+| SM4          | SM4 instructions                                                 |
+| SVE          | Scalable Vector Extension                                        |
+
 # license
 
 This code is published under an MIT license. See LICENSE file for more information.

vendor/github.com/klauspost/cpuid/v2/cpuid.go

@@ -73,6 +73,7 @@
 	AMD3DNOW                            // AMD 3DNOW
 	AMD3DNOWEXT                         // AMD 3DNowExt
 	AMXBF16                             // Tile computational operations on BFLOAT16 numbers
+	AMXFP16                             // Tile computational operations on FP16 numbers
 	AMXINT8                             // Tile computational operations on 8-bit integers
 	AMXTILE                             // Tile architecture
 	AVX                                 // AVX functions
@@ -93,8 +94,11 @@
 	AVX512VNNI                          // AVX-512 Vector Neural Network Instructions
 	AVX512VP2INTERSECT                  // AVX-512 Intersect for D/Q
 	AVX512VPOPCNTDQ                     // AVX-512 Vector Population Count Doubleword and Quadword
+	AVXIFMA                             // AVX-IFMA instructions
+	AVXNECONVERT                        // AVX-NE-CONVERT instructions
 	AVXSLOW                             // Indicates the CPU performs 2 128 bit operations instead of one
 	AVXVNNI                             // AVX (VEX encoded) VNNI neural network instructions
+	AVXVNNIINT8                         // AVX-VNNI-INT8 instructions
 	BMI1                                // Bit Manipulation Instruction Set 1
 	BMI2                                // Bit Manipulation Instruction Set 2
 	CETIBT                              // Intel CET Indirect Branch Tracking
@@ -103,15 +107,22 @@
 	CLMUL                               // Carry-less Multiplication
 	CLZERO                              // CLZERO instruction supported
 	CMOV                                // i686 CMOV
+	CMPCCXADD                           // CMPCCXADD instructions
 	CMPSB_SCADBS_SHORT                  // Fast short CMPSB and SCASB
 	CMPXCHG8                            // CMPXCHG8 instruction
 	CPBOOST                             // Core Performance Boost
+	CPPC                                // AMD: Collaborative Processor Performance Control
 	CX16                                // CMPXCHG16B Instruction
+	EFER_LMSLE_UNS                      // AMD: =Core::X86::Msr::EFER[LMSLE] is not supported, and MBZ
 	ENQCMD                              // Enqueue Command
 	ERMS                                // Enhanced REP MOVSB/STOSB
 	F16C                                // Half-precision floating-point conversion
+	FLUSH_L1D                           // Flush L1D cache
 	FMA3                                // Intel FMA 3. Does not imply AVX.
 	FMA4                                // Bulldozer FMA4 functions
+	FP128                               // AMD: When set, the internal FP/SIMD execution datapath is no more than 128-bits wide
+	FP256                               // AMD: When set, the internal FP/SIMD execution datapath is no more than 256-bits wide
+	FSRM                                // Fast Short Rep Mov
 	FXSR                                // FXSAVE, FXRESTOR instructions, CR4 bit 9
 	FXSROPT                             // FXSAVE/FXRSTOR optimizations
 	GFNI                                // Galois Field New Instructions. May require other features (AVX, AVX512VL,AVX512F) based on usage.
@@ -119,8 +130,14 @@
 	HRESET                              // If set CPU supports history reset and the IA32_HRESET_ENABLE MSR
 	HTT                                 // Hyperthreading (enabled)
 	HWA                                 // Hardware assert supported. Indicates support for MSRC001_10
+	HYBRID_CPU                          // This part has CPUs of more than one type.
 	HYPERVISOR                          // This bit has been reserved by Intel & AMD for use by hypervisors
+	IA32_ARCH_CAP                       // IA32_ARCH_CAPABILITIES MSR (Intel)
+	IA32_CORE_CAP                       // IA32_CORE_CAPABILITIES MSR
 	IBPB                                // Indirect Branch Restricted Speculation (IBRS) and Indirect Branch Predictor Barrier (IBPB)
+	IBRS                                // AMD: Indirect Branch Restricted Speculation
+	IBRS_PREFERRED                      // AMD: IBRS is preferred over software solution
+	IBRS_PROVIDES_SMP                   // AMD: IBRS provides Same Mode Protection
 	IBS                                 // Instruction Based Sampling (AMD)
 	IBSBRNTRGT                          // Instruction Based Sampling Feature (AMD)
 	IBSFETCHSAM                         // Instruction Based Sampling Feature (AMD)
@@ -130,7 +147,11 @@
 	IBSOPSAM                            // Instruction Based Sampling Feature (AMD)
 	IBSRDWROPCNT                        // Instruction Based Sampling Feature (AMD)
 	IBSRIPINVALIDCHK                    // Instruction Based Sampling Feature (AMD)
+	IBS_FETCH_CTLX                      // AMD: IBS fetch control extended MSR supported
+	IBS_OPDATA4                         // AMD: IBS op data 4 MSR supported
+	IBS_OPFUSE                          // AMD: Indicates support for IbsOpFuse
 	IBS_PREVENTHOST                     // Disallowing IBS use by the host supported
+	IBS_ZEN4                            // AMD: Fetch and Op IBS support IBS extensions added with Zen4
 	INT_WBINVD                          // WBINVD/WBNOINVD are interruptible.
 	INVLPGB                             // NVLPGB and TLBSYNC instruction supported
 	LAHF                                // LAHF/SAHF in long mode
@@ -138,13 +159,16 @@
 	LBRVIRT                             // LBR virtualization
 	LZCNT                               // LZCNT instruction
 	MCAOVERFLOW                         // MCA overflow recovery support.
+	MCDT_NO                             // Processor do not exhibit MXCSR Configuration Dependent Timing behavior and do not need to mitigate it.
 	MCOMMIT                             // MCOMMIT instruction supported
+	MD_CLEAR                            // VERW clears CPU buffers
 	MMX                                 // standard MMX
 	MMXEXT                              // SSE integer functions or AMD MMX ext
 	MOVBE                               // MOVBE instruction (big-endian)
 	MOVDIR64B                           // Move 64 Bytes as Direct Store
 	MOVDIRI                             // Move Doubleword as Direct Store
 	MOVSB_ZL                            // Fast Zero-Length MOVSB
+	MOVU                                // AMD: MOVU SSE instructions are more efficient and should be preferred to SSE	MOVL/MOVH. MOVUPS is more efficient than MOVLPS/MOVHPS. MOVUPD is more efficient than MOVLPD/MOVHPD
 	MPX                                 // Intel MPX (Memory Protection Extensions)
 	MSRIRC                              // Instruction Retired Counter MSR available
 	MSR_PAGEFLUSH                       // Page Flush MSR available
@@ -153,6 +177,9 @@
 	OSXSAVE                             // XSAVE enabled by OS
 	PCONFIG                             // PCONFIG for Intel Multi-Key Total Memory Encryption
 	POPCNT                              // POPCNT instruction
+	PPIN                                // AMD: Protected Processor Inventory Number support. Indicates that Protected Processor Inventory Number (PPIN) capability can be enabled
+	PREFETCHI                           // PREFETCHIT0/1 instructions
+	PSFD                                // AMD: Predictive Store Forward Disable
 	RDPRU                               // RDPRU instruction supported
 	RDRAND                              // RDRAND instruction is available
 	RDSEED                              // RDSEED instruction is available
@@ -172,6 +199,8 @@
 	SHA                                 // Intel SHA Extensions
 	SME                                 // AMD Secure Memory Encryption supported
 	SME_COHERENT                        // AMD Hardware cache coherency across encryption domains enforced
+	SPEC_CTRL_SSBD                      // Speculative Store Bypass Disable
+	SRBDS_CTRL                          // SRBDS mitigation MSR available
 	SSE                                 // SSE functions
 	SSE2                                // P4 SSE functions
 	SSE3                                // Prescott SSE3 functions
@@ -180,6 +209,7 @@
 	SSE4A                               // AMD Barcelona microarchitecture SSE4a instructions
 	SSSE3                               // Conroe SSSE3 functions
 	STIBP                               // Single Thread Indirect Branch Predictors
+	STIBP_ALWAYSON                      // AMD: Single Thread Indirect Branch Prediction Mode has Enhanced Performance and may be left Always On
 	STOSB_SHORT                         // Fast short STOSB
 	SUCCOR                              // Software uncorrectable error containment and recovery capability.
 	SVM                                 // AMD Secure Virtual Machine
@@ -192,8 +222,9 @@
 	SYSCALL                             // System-Call Extension (SCE): SYSCALL and SYSRET instructions.
 	SYSEE                               // SYSENTER and SYSEXIT instructions
 	TBM                                 // AMD Trailing Bit Manipulation
-	TOPEXT                              // TopologyExtensions: topology extensions support. Indicates support for CPUID Fn8000_001D_EAX_x[N:0]-CPUID Fn8000_001E_EDX.
+	TLB_FLUSH_NESTED                    // AMD: Flushing includes all the nested translations for guest translations
 	TME                                 // Intel Total Memory Encryption. The following MSRs are supported: IA32_TME_CAPABILITY, IA32_TME_ACTIVATE, IA32_TME_EXCLUDE_MASK, and IA32_TME_EXCLUDE_BASE.
+	TOPEXT                              // TopologyExtensions: topology extensions support. Indicates support for CPUID Fn8000_001D_EAX_x[N:0]-CPUID Fn8000_001E_EDX.
 	TSCRATEMSR                          // MSR based TSC rate control. Indicates support for MSR TSC ratio MSRC000_0104
 	TSXLDTRK                            // Intel TSX Suspend Load Address Tracking
 	VAES                                // Vector AES. AVX(512) versions requires additional checks.
@@ -358,7 +389,7 @@ func (c CPUInfo) Supports(ids ...FeatureID) bool {
 
 // Has allows for checking a single feature.
 // Should be inlined by the compiler.
-func (c CPUInfo) Has(id FeatureID) bool {
+func (c *CPUInfo) Has(id FeatureID) bool {
 	return c.featureSet.inSet(id)
 }
 
@@ -372,26 +403,47 @@ func (c CPUInfo) AnyOf(ids ...FeatureID) bool {
 	return false
 }
 
+// Features contains several features combined for a fast check using
+// CpuInfo.HasAll
+type Features *flagSet
+
+// CombineFeatures allows to combine several features for a close to constant time lookup.
+func CombineFeatures(ids ...FeatureID) Features {
+	var v flagSet
+	for _, id := range ids {
+		v.set(id)
+	}
+	return &v
+}
+
+func (c *CPUInfo) HasAll(f Features) bool {
+	return c.featureSet.hasSetP(f)
+}
+
 // https://en.wikipedia.org/wiki/X86-64#Microarchitecture_levels
-var level1Features = flagSetWith(CMOV, CMPXCHG8, X87, FXSR, MMX, SYSCALL, SSE, SSE2)
+var oneOfLevel = CombineFeatures(SYSEE, SYSCALL)
-var level2Features = flagSetWith(CMOV, CMPXCHG8, X87, FXSR, MMX, SYSCALL, SSE, SSE2, CX16, LAHF, POPCNT, SSE3, SSE4, SSE42, SSSE3)
+var level1Features = CombineFeatures(CMOV, CMPXCHG8, X87, FXSR, MMX, SSE, SSE2)
-var level3Features = flagSetWith(CMOV, CMPXCHG8, X87, FXSR, MMX, SYSCALL, SSE, SSE2, CX16, LAHF, POPCNT, SSE3, SSE4, SSE42, SSSE3, AVX, AVX2, BMI1, BMI2, F16C, FMA3, LZCNT, MOVBE, OSXSAVE)
+var level2Features = CombineFeatures(CMOV, CMPXCHG8, X87, FXSR, MMX, SSE, SSE2, CX16, LAHF, POPCNT, SSE3, SSE4, SSE42, SSSE3)
-var level4Features = flagSetWith(CMOV, CMPXCHG8, X87, FXSR, MMX, SYSCALL, SSE, SSE2, CX16, LAHF, POPCNT, SSE3, SSE4, SSE42, SSSE3, AVX, AVX2, BMI1, BMI2, F16C, FMA3, LZCNT, MOVBE, OSXSAVE, AVX512F, AVX512BW, AVX512CD, AVX512DQ, AVX512VL)
+var level3Features = CombineFeatures(CMOV, CMPXCHG8, X87, FXSR, MMX, SSE, SSE2, CX16, LAHF, POPCNT, SSE3, SSE4, SSE42, SSSE3, AVX, AVX2, BMI1, BMI2, F16C, FMA3, LZCNT, MOVBE, OSXSAVE)
+var level4Features = CombineFeatures(CMOV, CMPXCHG8, X87, FXSR, MMX, SSE, SSE2, CX16, LAHF, POPCNT, SSE3, SSE4, SSE42, SSSE3, AVX, AVX2, BMI1, BMI2, F16C, FMA3, LZCNT, MOVBE, OSXSAVE, AVX512F, AVX512BW, AVX512CD, AVX512DQ, AVX512VL)
 
 // X64Level returns the microarchitecture level detected on the CPU.
 // If features are lacking or non x64 mode, 0 is returned.
 // See https://en.wikipedia.org/wiki/X86-64#Microarchitecture_levels
 func (c CPUInfo) X64Level() int {
-	if c.featureSet.hasSet(level4Features) {
+	if !c.featureSet.hasOneOf(oneOfLevel) {
+		return 0
+	}
+	if c.featureSet.hasSetP(level4Features) {
 		return 4
 	}
-	if c.featureSet.hasSet(level3Features) {
+	if c.featureSet.hasSetP(level3Features) {
 		return 3
 	}
-	if c.featureSet.hasSet(level2Features) {
+	if c.featureSet.hasSetP(level2Features) {
 		return 2
 	}
-	if c.featureSet.hasSet(level1Features) {
+	if c.featureSet.hasSetP(level1Features) {
 		return 1
 	}
 	return 0
@@ -555,7 +607,7 @@ func (c CPUInfo) VM() bool {
 // flagSet contains detected cpu features and characteristics in an array of flags
 type flagSet [(lastID + flagMask) / flagBits]flags
 
-func (s flagSet) inSet(feat FeatureID) bool {
+func (s *flagSet) inSet(feat FeatureID) bool {
 	return s[feat>>flagBitsLog2]&(1<<(feat&flagMask)) != 0
 }
 
@@ -585,7 +637,7 @@ func (s *flagSet) or(other flagSet) {
 }
 
 // hasSet returns whether all features are present.
-func (s flagSet) hasSet(other flagSet) bool {
+func (s *flagSet) hasSet(other flagSet) bool {
 	for i, v := range other[:] {
 		if s[i]&v != v {
 			return false
@@ -594,8 +646,28 @@ func (s flagSet) hasSet(other flagSet) bool {
 	return true
 }
 
+// hasSet returns whether all features are present.
+func (s *flagSet) hasSetP(other *flagSet) bool {
+	for i, v := range other[:] {
+		if s[i]&v != v {
+			return false
+		}
+	}
+	return true
+}
+
+// hasOneOf returns whether one or more features are present.
+func (s *flagSet) hasOneOf(other *flagSet) bool {
+	for i, v := range other[:] {
+		if s[i]&v != 0 {
+			return true
+		}
+	}
+	return false
+}
+
 // nEnabled will return the number of enabled flags.
-func (s flagSet) nEnabled() (n int) {
+func (s *flagSet) nEnabled() (n int) {
 	for _, v := range s[:] {
 		n += bits.OnesCount64(uint64(v))
 	}
@@ -1093,21 +1165,36 @@ func support() flagSet {
 		fs.setIf(ecx&(1<<30) != 0, SGXLC)
 
 		// CPUID.(EAX=7, ECX=0).EDX
+		fs.setIf(edx&(1<<4) != 0, FSRM)
+		fs.setIf(edx&(1<<9) != 0, SRBDS_CTRL)
+		fs.setIf(edx&(1<<10) != 0, MD_CLEAR)
 		fs.setIf(edx&(1<<11) != 0, RTM_ALWAYS_ABORT)
 		fs.setIf(edx&(1<<14) != 0, SERIALIZE)
+		fs.setIf(edx&(1<<15) != 0, HYBRID_CPU)
 		fs.setIf(edx&(1<<16) != 0, TSXLDTRK)
 		fs.setIf(edx&(1<<18) != 0, PCONFIG)
 		fs.setIf(edx&(1<<20) != 0, CETIBT)
 		fs.setIf(edx&(1<<26) != 0, IBPB)
 		fs.setIf(edx&(1<<27) != 0, STIBP)
+		fs.setIf(edx&(1<<28) != 0, FLUSH_L1D)
+		fs.setIf(edx&(1<<29) != 0, IA32_ARCH_CAP)
+		fs.setIf(edx&(1<<30) != 0, IA32_CORE_CAP)
+		fs.setIf(edx&(1<<31) != 0, SPEC_CTRL_SSBD)
 
-		// CPUID.(EAX=7, ECX=1)
+		// CPUID.(EAX=7, ECX=1).EDX
+		fs.setIf(edx&(1<<4) != 0, AVXVNNIINT8)
+		fs.setIf(edx&(1<<5) != 0, AVXNECONVERT)
+		fs.setIf(edx&(1<<14) != 0, PREFETCHI)
+
+		// CPUID.(EAX=7, ECX=1).EAX
 		eax1, _, _, _ := cpuidex(7, 1)
 		fs.setIf(fs.inSet(AVX) && eax1&(1<<4) != 0, AVXVNNI)
+		fs.setIf(eax1&(1<<7) != 0, CMPCCXADD)
 		fs.setIf(eax1&(1<<10) != 0, MOVSB_ZL)
 		fs.setIf(eax1&(1<<11) != 0, STOSB_SHORT)
 		fs.setIf(eax1&(1<<12) != 0, CMPSB_SCADBS_SHORT)
 		fs.setIf(eax1&(1<<22) != 0, HRESET)
+		fs.setIf(eax1&(1<<23) != 0, AVXIFMA)
 		fs.setIf(eax1&(1<<26) != 0, LAM)
 
 		// Only detect AVX-512 features if XGETBV is supported
@@ -1145,9 +1232,15 @@ func support() flagSet {
 				fs.setIf(edx&(1<<25) != 0, AMXINT8)
 				// eax1 = CPUID.(EAX=7, ECX=1).EAX
 				fs.setIf(eax1&(1<<5) != 0, AVX512BF16)
+				fs.setIf(eax1&(1<<21) != 0, AMXFP16)
 			}
 		}
+
+		// CPUID.(EAX=7, ECX=2)
+		_, _, _, edx = cpuidex(7, 2)
+		fs.setIf(edx&(1<<5) != 0, MCDT_NO)
 	}
+
 	// Processor Extended State Enumeration Sub-leaf (EAX = 0DH, ECX = 1)
 	// EAX
 	// Bit 00: XSAVEOPT is available.
@@ -1212,9 +1305,21 @@ func support() flagSet {
 
 	if maxExtendedFunction() >= 0x80000008 {
 		_, b, _, _ := cpuid(0x80000008)
+		fs.setIf(b&(1<<28) != 0, PSFD)
+		fs.setIf(b&(1<<27) != 0, CPPC)
+		fs.setIf(b&(1<<24) != 0, SPEC_CTRL_SSBD)
+		fs.setIf(b&(1<<23) != 0, PPIN)
+		fs.setIf(b&(1<<21) != 0, TLB_FLUSH_NESTED)
+		fs.setIf(b&(1<<20) != 0, EFER_LMSLE_UNS)
+		fs.setIf(b&(1<<19) != 0, IBRS_PROVIDES_SMP)
+		fs.setIf(b&(1<<18) != 0, IBRS_PREFERRED)
+		fs.setIf(b&(1<<17) != 0, STIBP_ALWAYSON)
+		fs.setIf(b&(1<<15) != 0, STIBP)
+		fs.setIf(b&(1<<14) != 0, IBRS)
+		fs.setIf((b&(1<<13)) != 0, INT_WBINVD)
+		fs.setIf(b&(1<<12) != 0, IBPB)
 		fs.setIf((b&(1<<9)) != 0, WBNOINVD)
 		fs.setIf((b&(1<<8)) != 0, MCOMMIT)
-		fs.setIf((b&(1<<13)) != 0, INT_WBINVD)
 		fs.setIf((b&(1<<4)) != 0, RDPRU)
 		fs.setIf((b&(1<<3)) != 0, INVLPGB)
 		fs.setIf((b&(1<<1)) != 0, MSRIRC)
@@ -1235,6 +1340,13 @@ func support() flagSet {
 		fs.setIf((edx>>12)&1 == 1, SVMPFT)
 	}
 
+	if maxExtendedFunction() >= 0x8000001a {
+		eax, _, _, _ := cpuid(0x8000001a)
+		fs.setIf((eax>>0)&1 == 1, FP128)
+		fs.setIf((eax>>1)&1 == 1, MOVU)
+		fs.setIf((eax>>2)&1 == 1, FP256)
+	}
+
 	if maxExtendedFunction() >= 0x8000001b && fs.inSet(IBS) {
 		eax, _, _, _ := cpuid(0x8000001b)
 		fs.setIf((eax>>0)&1 == 1, IBSFFV)
@@ -1245,6 +1357,10 @@ func support() flagSet {
 		fs.setIf((eax>>5)&1 == 1, IBSBRNTRGT)
 		fs.setIf((eax>>6)&1 == 1, IBSOPCNTEXT)
 		fs.setIf((eax>>7)&1 == 1, IBSRIPINVALIDCHK)
+		fs.setIf((eax>>8)&1 == 1, IBS_OPFUSE)
+		fs.setIf((eax>>9)&1 == 1, IBS_FETCH_CTLX)
+		fs.setIf((eax>>10)&1 == 1, IBS_OPDATA4) // Doc says "Fixed,0. IBS op data 4 MSR supported", but assuming they mean 1.
+		fs.setIf((eax>>11)&1 == 1, IBS_ZEN4)
 	}
 
 	if maxExtendedFunction() >= 0x8000001f && vend == AMD {

vendor/github.com/klauspost/cpuid/v2/featureid_string.go

@@ -13,176 +13,207 @@ func _() {
 	_ = x[AMD3DNOW-3]
 	_ = x[AMD3DNOWEXT-4]
 	_ = x[AMXBF16-5]
-	_ = x[AMXINT8-6]
+	_ = x[AMXFP16-6]
-	_ = x[AMXTILE-7]
+	_ = x[AMXINT8-7]
-	_ = x[AVX-8]
+	_ = x[AMXTILE-8]
-	_ = x[AVX2-9]
+	_ = x[AVX-9]
-	_ = x[AVX512BF16-10]
+	_ = x[AVX2-10]
-	_ = x[AVX512BITALG-11]
+	_ = x[AVX512BF16-11]
-	_ = x[AVX512BW-12]
+	_ = x[AVX512BITALG-12]
-	_ = x[AVX512CD-13]
+	_ = x[AVX512BW-13]
-	_ = x[AVX512DQ-14]
+	_ = x[AVX512CD-14]
-	_ = x[AVX512ER-15]
+	_ = x[AVX512DQ-15]
-	_ = x[AVX512F-16]
+	_ = x[AVX512ER-16]
-	_ = x[AVX512FP16-17]
+	_ = x[AVX512F-17]
-	_ = x[AVX512IFMA-18]
+	_ = x[AVX512FP16-18]
-	_ = x[AVX512PF-19]
+	_ = x[AVX512IFMA-19]
-	_ = x[AVX512VBMI-20]
+	_ = x[AVX512PF-20]
-	_ = x[AVX512VBMI2-21]
+	_ = x[AVX512VBMI-21]
-	_ = x[AVX512VL-22]
+	_ = x[AVX512VBMI2-22]
-	_ = x[AVX512VNNI-23]
+	_ = x[AVX512VL-23]
-	_ = x[AVX512VP2INTERSECT-24]
+	_ = x[AVX512VNNI-24]
-	_ = x[AVX512VPOPCNTDQ-25]
+	_ = x[AVX512VP2INTERSECT-25]
-	_ = x[AVXSLOW-26]
+	_ = x[AVX512VPOPCNTDQ-26]
-	_ = x[AVXVNNI-27]
+	_ = x[AVXIFMA-27]
-	_ = x[BMI1-28]
+	_ = x[AVXNECONVERT-28]
-	_ = x[BMI2-29]
+	_ = x[AVXSLOW-29]
-	_ = x[CETIBT-30]
+	_ = x[AVXVNNI-30]
-	_ = x[CETSS-31]
+	_ = x[AVXVNNIINT8-31]
-	_ = x[CLDEMOTE-32]
+	_ = x[BMI1-32]
-	_ = x[CLMUL-33]
+	_ = x[BMI2-33]
-	_ = x[CLZERO-34]
+	_ = x[CETIBT-34]
-	_ = x[CMOV-35]
+	_ = x[CETSS-35]
-	_ = x[CMPSB_SCADBS_SHORT-36]
+	_ = x[CLDEMOTE-36]
-	_ = x[CMPXCHG8-37]
+	_ = x[CLMUL-37]
-	_ = x[CPBOOST-38]
+	_ = x[CLZERO-38]
-	_ = x[CX16-39]
+	_ = x[CMOV-39]
-	_ = x[ENQCMD-40]
+	_ = x[CMPCCXADD-40]
-	_ = x[ERMS-41]
+	_ = x[CMPSB_SCADBS_SHORT-41]
-	_ = x[F16C-42]
+	_ = x[CMPXCHG8-42]
-	_ = x[FMA3-43]
+	_ = x[CPBOOST-43]
-	_ = x[FMA4-44]
+	_ = x[CPPC-44]
-	_ = x[FXSR-45]
+	_ = x[CX16-45]
-	_ = x[FXSROPT-46]
+	_ = x[EFER_LMSLE_UNS-46]
-	_ = x[GFNI-47]
+	_ = x[ENQCMD-47]
-	_ = x[HLE-48]
+	_ = x[ERMS-48]
-	_ = x[HRESET-49]
+	_ = x[F16C-49]
-	_ = x[HTT-50]
+	_ = x[FLUSH_L1D-50]
-	_ = x[HWA-51]
+	_ = x[FMA3-51]
-	_ = x[HYPERVISOR-52]
+	_ = x[FMA4-52]
-	_ = x[IBPB-53]
+	_ = x[FP128-53]
-	_ = x[IBS-54]
+	_ = x[FP256-54]
-	_ = x[IBSBRNTRGT-55]
+	_ = x[FSRM-55]
-	_ = x[IBSFETCHSAM-56]
+	_ = x[FXSR-56]
-	_ = x[IBSFFV-57]
+	_ = x[FXSROPT-57]
-	_ = x[IBSOPCNT-58]
+	_ = x[GFNI-58]
-	_ = x[IBSOPCNTEXT-59]
+	_ = x[HLE-59]
-	_ = x[IBSOPSAM-60]
+	_ = x[HRESET-60]
-	_ = x[IBSRDWROPCNT-61]
+	_ = x[HTT-61]
-	_ = x[IBSRIPINVALIDCHK-62]
+	_ = x[HWA-62]
-	_ = x[IBS_PREVENTHOST-63]
+	_ = x[HYBRID_CPU-63]
-	_ = x[INT_WBINVD-64]
+	_ = x[HYPERVISOR-64]
-	_ = x[INVLPGB-65]
+	_ = x[IA32_ARCH_CAP-65]
-	_ = x[LAHF-66]
+	_ = x[IA32_CORE_CAP-66]
-	_ = x[LAM-67]
+	_ = x[IBPB-67]
-	_ = x[LBRVIRT-68]
+	_ = x[IBRS-68]
-	_ = x[LZCNT-69]
+	_ = x[IBRS_PREFERRED-69]
-	_ = x[MCAOVERFLOW-70]
+	_ = x[IBRS_PROVIDES_SMP-70]
-	_ = x[MCOMMIT-71]
+	_ = x[IBS-71]
-	_ = x[MMX-72]
+	_ = x[IBSBRNTRGT-72]
-	_ = x[MMXEXT-73]
+	_ = x[IBSFETCHSAM-73]
-	_ = x[MOVBE-74]
+	_ = x[IBSFFV-74]
-	_ = x[MOVDIR64B-75]
+	_ = x[IBSOPCNT-75]
-	_ = x[MOVDIRI-76]
+	_ = x[IBSOPCNTEXT-76]
-	_ = x[MOVSB_ZL-77]
+	_ = x[IBSOPSAM-77]
-	_ = x[MPX-78]
+	_ = x[IBSRDWROPCNT-78]
-	_ = x[MSRIRC-79]
+	_ = x[IBSRIPINVALIDCHK-79]
-	_ = x[MSR_PAGEFLUSH-80]
+	_ = x[IBS_FETCH_CTLX-80]
-	_ = x[NRIPS-81]
+	_ = x[IBS_OPDATA4-81]
-	_ = x[NX-82]
+	_ = x[IBS_OPFUSE-82]
-	_ = x[OSXSAVE-83]
+	_ = x[IBS_PREVENTHOST-83]
-	_ = x[PCONFIG-84]
+	_ = x[IBS_ZEN4-84]
-	_ = x[POPCNT-85]
+	_ = x[INT_WBINVD-85]
-	_ = x[RDPRU-86]
+	_ = x[INVLPGB-86]
-	_ = x[RDRAND-87]
+	_ = x[LAHF-87]
-	_ = x[RDSEED-88]
+	_ = x[LAM-88]
-	_ = x[RDTSCP-89]
+	_ = x[LBRVIRT-89]
-	_ = x[RTM-90]
+	_ = x[LZCNT-90]
-	_ = x[RTM_ALWAYS_ABORT-91]
+	_ = x[MCAOVERFLOW-91]
-	_ = x[SERIALIZE-92]
+	_ = x[MCDT_NO-92]
-	_ = x[SEV-93]
+	_ = x[MCOMMIT-93]
-	_ = x[SEV_64BIT-94]
+	_ = x[MD_CLEAR-94]
-	_ = x[SEV_ALTERNATIVE-95]
+	_ = x[MMX-95]
-	_ = x[SEV_DEBUGSWAP-96]
+	_ = x[MMXEXT-96]
-	_ = x[SEV_ES-97]
+	_ = x[MOVBE-97]
-	_ = x[SEV_RESTRICTED-98]
+	_ = x[MOVDIR64B-98]
-	_ = x[SEV_SNP-99]
+	_ = x[MOVDIRI-99]
-	_ = x[SGX-100]
+	_ = x[MOVSB_ZL-100]
-	_ = x[SGXLC-101]
+	_ = x[MOVU-101]
-	_ = x[SHA-102]
+	_ = x[MPX-102]
-	_ = x[SME-103]
+	_ = x[MSRIRC-103]
-	_ = x[SME_COHERENT-104]
+	_ = x[MSR_PAGEFLUSH-104]
-	_ = x[SSE-105]
+	_ = x[NRIPS-105]
-	_ = x[SSE2-106]
+	_ = x[NX-106]
-	_ = x[SSE3-107]
+	_ = x[OSXSAVE-107]
-	_ = x[SSE4-108]
+	_ = x[PCONFIG-108]
-	_ = x[SSE42-109]
+	_ = x[POPCNT-109]
-	_ = x[SSE4A-110]
+	_ = x[PPIN-110]
-	_ = x[SSSE3-111]
+	_ = x[PREFETCHI-111]
-	_ = x[STIBP-112]
+	_ = x[PSFD-112]
-	_ = x[STOSB_SHORT-113]
+	_ = x[RDPRU-113]
-	_ = x[SUCCOR-114]
+	_ = x[RDRAND-114]
-	_ = x[SVM-115]
+	_ = x[RDSEED-115]
-	_ = x[SVMDA-116]
+	_ = x[RDTSCP-116]
-	_ = x[SVMFBASID-117]
+	_ = x[RTM-117]
-	_ = x[SVML-118]
+	_ = x[RTM_ALWAYS_ABORT-118]
-	_ = x[SVMNP-119]
+	_ = x[SERIALIZE-119]
-	_ = x[SVMPF-120]
+	_ = x[SEV-120]
-	_ = x[SVMPFT-121]
+	_ = x[SEV_64BIT-121]
-	_ = x[SYSCALL-122]
+	_ = x[SEV_ALTERNATIVE-122]
-	_ = x[SYSEE-123]
+	_ = x[SEV_DEBUGSWAP-123]
-	_ = x[TBM-124]
+	_ = x[SEV_ES-124]
-	_ = x[TOPEXT-125]
+	_ = x[SEV_RESTRICTED-125]
-	_ = x[TME-126]
+	_ = x[SEV_SNP-126]
-	_ = x[TSCRATEMSR-127]
+	_ = x[SGX-127]
-	_ = x[TSXLDTRK-128]
+	_ = x[SGXLC-128]
-	_ = x[VAES-129]
+	_ = x[SHA-129]
-	_ = x[VMCBCLEAN-130]
+	_ = x[SME-130]
-	_ = x[VMPL-131]
+	_ = x[SME_COHERENT-131]
-	_ = x[VMSA_REGPROT-132]
+	_ = x[SPEC_CTRL_SSBD-132]
-	_ = x[VMX-133]
+	_ = x[SRBDS_CTRL-133]
-	_ = x[VPCLMULQDQ-134]
+	_ = x[SSE-134]
-	_ = x[VTE-135]
+	_ = x[SSE2-135]
-	_ = x[WAITPKG-136]
+	_ = x[SSE3-136]
-	_ = x[WBNOINVD-137]
+	_ = x[SSE4-137]
-	_ = x[X87-138]
+	_ = x[SSE42-138]
-	_ = x[XGETBV1-139]
+	_ = x[SSE4A-139]
-	_ = x[XOP-140]
+	_ = x[SSSE3-140]
-	_ = x[XSAVE-141]
+	_ = x[STIBP-141]
-	_ = x[XSAVEC-142]
+	_ = x[STIBP_ALWAYSON-142]
-	_ = x[XSAVEOPT-143]
+	_ = x[STOSB_SHORT-143]
-	_ = x[XSAVES-144]
+	_ = x[SUCCOR-144]
-	_ = x[AESARM-145]
+	_ = x[SVM-145]
-	_ = x[ARMCPUID-146]
+	_ = x[SVMDA-146]
-	_ = x[ASIMD-147]
+	_ = x[SVMFBASID-147]
-	_ = x[ASIMDDP-148]
+	_ = x[SVML-148]
-	_ = x[ASIMDHP-149]
+	_ = x[SVMNP-149]
-	_ = x[ASIMDRDM-150]
+	_ = x[SVMPF-150]
-	_ = x[ATOMICS-151]
+	_ = x[SVMPFT-151]
-	_ = x[CRC32-152]
+	_ = x[SYSCALL-152]
-	_ = x[DCPOP-153]
+	_ = x[SYSEE-153]
-	_ = x[EVTSTRM-154]
+	_ = x[TBM-154]
-	_ = x[FCMA-155]
+	_ = x[TLB_FLUSH_NESTED-155]
-	_ = x[FP-156]
+	_ = x[TME-156]
-	_ = x[FPHP-157]
+	_ = x[TOPEXT-157]
-	_ = x[GPA-158]
+	_ = x[TSCRATEMSR-158]
-	_ = x[JSCVT-159]
+	_ = x[TSXLDTRK-159]
-	_ = x[LRCPC-160]
+	_ = x[VAES-160]
-	_ = x[PMULL-161]
+	_ = x[VMCBCLEAN-161]
-	_ = x[SHA1-162]
+	_ = x[VMPL-162]
-	_ = x[SHA2-163]
+	_ = x[VMSA_REGPROT-163]
-	_ = x[SHA3-164]
+	_ = x[VMX-164]
-	_ = x[SHA512-165]
+	_ = x[VPCLMULQDQ-165]
-	_ = x[SM3-166]
+	_ = x[VTE-166]
-	_ = x[SM4-167]
+	_ = x[WAITPKG-167]
-	_ = x[SVE-168]
+	_ = x[WBNOINVD-168]
-	_ = x[lastID-169]
+	_ = x[X87-169]
+	_ = x[XGETBV1-170]
+	_ = x[XOP-171]
+	_ = x[XSAVE-172]
+	_ = x[XSAVEC-173]
+	_ = x[XSAVEOPT-174]
+	_ = x[XSAVES-175]
+	_ = x[AESARM-176]
+	_ = x[ARMCPUID-177]
+	_ = x[ASIMD-178]
+	_ = x[ASIMDDP-179]
+	_ = x[ASIMDHP-180]
+	_ = x[ASIMDRDM-181]
+	_ = x[ATOMICS-182]
+	_ = x[CRC32-183]
+	_ = x[DCPOP-184]
+	_ = x[EVTSTRM-185]
+	_ = x[FCMA-186]
+	_ = x[FP-187]
+	_ = x[FPHP-188]
+	_ = x[GPA-189]
+	_ = x[JSCVT-190]
+	_ = x[LRCPC-191]
+	_ = x[PMULL-192]
+	_ = x[SHA1-193]
+	_ = x[SHA2-194]
+	_ = x[SHA3-195]
+	_ = x[SHA512-196]
+	_ = x[SM3-197]
+	_ = x[SM4-198]
+	_ = x[SVE-199]
+	_ = x[lastID-200]
 	_ = x[firstID-0]
 }
 
-const _FeatureID_name = "firstIDADXAESNIAMD3DNOWAMD3DNOWEXTAMXBF16AMXINT8AMXTILEAVXAVX2AVX512BF16AVX512BITALGAVX512BWAVX512CDAVX512DQAVX512ERAVX512FAVX512FP16AVX512IFMAAVX512PFAVX512VBMIAVX512VBMI2AVX512VLAVX512VNNIAVX512VP2INTERSECTAVX512VPOPCNTDQAVXSLOWAVXVNNIBMI1BMI2CETIBTCETSSCLDEMOTECLMULCLZEROCMOVCMPSB_SCADBS_SHORTCMPXCHG8CPBOOSTCX16ENQCMDERMSF16CFMA3FMA4FXSRFXSROPTGFNIHLEHRESETHTTHWAHYPERVISORIBPBIBSIBSBRNTRGTIBSFETCHSAMIBSFFVIBSOPCNTIBSOPCNTEXTIBSOPSAMIBSRDWROPCNTIBSRIPINVALIDCHKIBS_PREVENTHOSTINT_WBINVDINVLPGBLAHFLAMLBRVIRTLZCNTMCAOVERFLOWMCOMMITMMXMMXEXTMOVBEMOVDIR64BMOVDIRIMOVSB_ZLMPXMSRIRCMSR_PAGEFLUSHNRIPSNXOSXSAVEPCONFIGPOPCNTRDPRURDRANDRDSEEDRDTSCPRTMRTM_ALWAYS_ABORTSERIALIZESEVSEV_64BITSEV_ALTERNATIVESEV_DEBUGSWAPSEV_ESSEV_RESTRICTEDSEV_SNPSGXSGXLCSHASMESME_COHERENTSSESSE2SSE3SSE4SSE42SSE4ASSSE3STIBPSTOSB_SHORTSUCCORSVMSVMDASVMFBASIDSVMLSVMNPSVMPFSVMPFTSYSCALLSYSEETBMTOPEXTTMETSCRATEMSRTSXLDTRKVAESVMCBCLEANVMPLVMSA_REGPROTVMXVPCLMULQDQVTEWAITPKGWBNOINVDX87XGETBV1XOPXSAVEXSAVECXSAVEOPTXSAVESAESARMARMCPUIDASIMDASIMDDPASIMDHPASIMDRDMATOMICSCRC32DCPOPEVTSTRMFCMAFPFPHPGPAJSCVTLRCPCPMULLSHA1SHA2SHA3SHA512SM3SM4SVElastID"
+const _FeatureID_name = "firstIDADXAESNIAMD3DNOWAMD3DNOWEXTAMXBF16AMXFP16AMXINT8AMXTILEAVXAVX2AVX512BF16AVX512BITALGAVX512BWAVX512CDAVX512DQAVX512ERAVX512FAVX512FP16AVX512IFMAAVX512PFAVX512VBMIAVX512VBMI2AVX512VLAVX512VNNIAVX512VP2INTERSECTAVX512VPOPCNTDQAVXIFMAAVXNECONVERTAVXSLOWAVXVNNIAVXVNNIINT8BMI1BMI2CETIBTCETSSCLDEMOTECLMULCLZEROCMOVCMPCCXADDCMPSB_SCADBS_SHORTCMPXCHG8CPBOOSTCPPCCX16EFER_LMSLE_UNSENQCMDERMSF16CFLUSH_L1DFMA3FMA4FP128FP256FSRMFXSRFXSROPTGFNIHLEHRESETHTTHWAHYBRID_CPUHYPERVISORIA32_ARCH_CAPIA32_CORE_CAPIBPBIBRSIBRS_PREFERREDIBRS_PROVIDES_SMPIBSIBSBRNTRGTIBSFETCHSAMIBSFFVIBSOPCNTIBSOPCNTEXTIBSOPSAMIBSRDWROPCNTIBSRIPINVALIDCHKIBS_FETCH_CTLXIBS_OPDATA4IBS_OPFUSEIBS_PREVENTHOSTIBS_ZEN4INT_WBINVDINVLPGBLAHFLAMLBRVIRTLZCNTMCAOVERFLOWMCDT_NOMCOMMITMD_CLEARMMXMMXEXTMOVBEMOVDIR64BMOVDIRIMOVSB_ZLMOVUMPXMSRIRCMSR_PAGEFLUSHNRIPSNXOSXSAVEPCONFIGPOPCNTPPINPREFETCHIPSFDRDPRURDRANDRDSEEDRDTSCPRTMRTM_ALWAYS_ABORTSERIALIZESEVSEV_64BITSEV_ALTERNATIVESEV_DEBUGSWAPSEV_ESSEV_RESTRICTEDSEV_SNPSGXSGXLCSHASMESME_COHERENTSPEC_CTRL_SSBDSRBDS_CTRLSSESSE2SSE3SSE4SSE42SSE4ASSSE3STIBPSTIBP_ALWAYSONSTOSB_SHORTSUCCORSVMSVMDASVMFBASIDSVMLSVMNPSVMPFSVMPFTSYSCALLSYSEETBMTLB_FLUSH_NESTEDTMETOPEXTTSCRATEMSRTSXLDTRKVAESVMCBCLEANVMPLVMSA_REGPROTVMXVPCLMULQDQVTEWAITPKGWBNOINVDX87XGETBV1XOPXSAVEXSAVECXSAVEOPTXSAVESAESARMARMCPUIDASIMDASIMDDPASIMDHPASIMDRDMATOMICSCRC32DCPOPEVTSTRMFCMAFPFPHPGPAJSCVTLRCPCPMULLSHA1SHA2SHA3SHA512SM3SM4SVElastID"
 
-var _FeatureID_index = [...]uint16{0, 7, 10, 15, 23, 34, 41, 48, 55, 58, 62, 72, 84, 92, 100, 108, 116, 123, 133, 143, 151, 161, 172, 180, 190, 208, 223, 230, 237, 241, 245, 251, 256, 264, 269, 275, 279, 297, 305, 312, 316, 322, 326, 330, 334, 338, 342, 349, 353, 356, 362, 365, 368, 378, 382, 385, 395, 406, 412, 420, 431, 439, 451, 467, 482, 492, 499, 503, 506, 513, 518, 529, 536, 539, 545, 550, 559, 566, 574, 577, 583, 596, 601, 603, 610, 617, 623, 628, 634, 640, 646, 649, 665, 674, 677, 686, 701, 714, 720, 734, 741, 744, 749, 752, 755, 767, 770, 774, 778, 782, 787, 792, 797, 802, 813, 819, 822, 827, 836, 840, 845, 850, 856, 863, 868, 871, 877, 880, 890, 898, 902, 911, 915, 927, 930, 940, 943, 950, 958, 961, 968, 971, 976, 982, 990, 996, 1002, 1010, 1015, 1022, 1029, 1037, 1044, 1049, 1054, 1061, 1065, 1067, 1071, 1074, 1079, 1084, 1089, 1093, 1097, 1101, 1107, 1110, 1113, 1116, 1122}
+var _FeatureID_index = [...]uint16{0, 7, 10, 15, 23, 34, 41, 48, 55, 62, 65, 69, 79, 91, 99, 107, 115, 123, 130, 140, 150, 158, 168, 179, 187, 197, 215, 230, 237, 249, 256, 263, 274, 278, 282, 288, 293, 301, 306, 312, 316, 325, 343, 351, 358, 362, 366, 380, 386, 390, 394, 403, 407, 411, 416, 421, 425, 429, 436, 440, 443, 449, 452, 455, 465, 475, 488, 501, 505, 509, 523, 540, 543, 553, 564, 570, 578, 589, 597, 609, 625, 639, 650, 660, 675, 683, 693, 700, 704, 707, 714, 719, 730, 737, 744, 752, 755, 761, 766, 775, 782, 790, 794, 797, 803, 816, 821, 823, 830, 837, 843, 847, 856, 860, 865, 871, 877, 883, 886, 902, 911, 914, 923, 938, 951, 957, 971, 978, 981, 986, 989, 992, 1004, 1018, 1028, 1031, 1035, 1039, 1043, 1048, 1053, 1058, 1063, 1077, 1088, 1094, 1097, 1102, 1111, 1115, 1120, 1125, 1131, 1138, 1143, 1146, 1162, 1165, 1171, 1181, 1189, 1193, 1202, 1206, 1218, 1221, 1231, 1234, 1241, 1249, 1252, 1259, 1262, 1267, 1273, 1281, 1287, 1293, 1301, 1306, 1313, 1320, 1328, 1335, 1340, 1345, 1352, 1356, 1358, 1362, 1365, 1370, 1375, 1380, 1384, 1388, 1392, 1398, 1401, 1404, 1407, 1413}
 
 func (i FeatureID) String() string {
 	if i < 0 || i >= FeatureID(len(_FeatureID_index)-1) {

vendor/github.com/klauspost/cpuid/v2/os_darwin_arm64.go

@@ -83,7 +83,7 @@ func tryToFillCPUInfoFomSysctl(c *CPUInfo) {
 	c.Model = sysctlGetInt(0, "machdep.cpu.model")
 	c.CacheLine = sysctlGetInt64(0, "hw.cachelinesize")
 	c.Cache.L1I = sysctlGetInt64(-1, "hw.l1icachesize")
-	c.Cache.L1D = sysctlGetInt64(-1, "hw.l1icachesize")
+	c.Cache.L1D = sysctlGetInt64(-1, "hw.l1dcachesize")
 	c.Cache.L2 = sysctlGetInt64(-1, "hw.l2cachesize")
 	c.Cache.L3 = sysctlGetInt64(-1, "hw.l3cachesize")
 

vendor/github.com/minio/minio-go/v7/api-put-object.go

@@ -93,6 +93,28 @@ type PutObjectOptions struct {
 	// This can be used for faster uploads on non-seekable or slow-to-seek input.
 	ConcurrentStreamParts bool
 	Internal              AdvancedPutOptions
+
+	customHeaders http.Header
+}
+
+// SetMatchETag if etag matches while PUT MinIO returns an error
+// this is a MinIO specific extension to support optimistic locking
+// semantics.
+func (opts *PutObjectOptions) SetMatchETag(etag string) {
+	if opts.customHeaders == nil {
+		opts.customHeaders = http.Header{}
+	}
+	opts.customHeaders.Set("If-Match", "\""+etag+"\"")
+}
+
+// SetMatchETagExcept if etag does not match while PUT MinIO returns an
+// error this is a MinIO specific extension to support optimistic locking
+// semantics.
+func (opts *PutObjectOptions) SetMatchETagExcept(etag string) {
+	if opts.customHeaders == nil {
+		opts.customHeaders = http.Header{}
+	}
+	opts.customHeaders.Set("If-None-Match", "\""+etag+"\"")
 }
 
 // getNumThreads - gets the number of threads to be used in the multipart
@@ -187,6 +209,12 @@ func (opts PutObjectOptions) Header() (header http.Header) {
 			header.Set("x-amz-meta-"+k, v)
 		}
 	}
+
+	// set any other additional custom headers.
+	for k, v := range opts.customHeaders {
+		header[k] = v
+	}
+
 	return
 }
 

vendor/github.com/minio/minio-go/v7/api.go

@@ -118,7 +118,7 @@ type Options struct {
 // Global constants.
 const (
 	libraryName    = "minio-go"
-	libraryVersion = "v7.0.48"
+	libraryVersion = "v7.0.49"
 )
 
 // User Agent should always following the below style.

vendor/modules.txt

@@ -163,8 +163,8 @@ github.com/dsoprea/go-png-image-structure/v2
 ## explicit; go 1.12
 github.com/dsoprea/go-utility/v2/filesystem
 github.com/dsoprea/go-utility/v2/image
-# github.com/dustin/go-humanize v1.0.0
+# github.com/dustin/go-humanize v1.0.1
-## explicit
+## explicit; go 1.16
 github.com/dustin/go-humanize
 # github.com/fsnotify/fsnotify v1.6.0
 ## explicit; go 1.16
@@ -324,14 +324,14 @@ github.com/json-iterator/go
 # github.com/kballard/go-shellquote v0.0.0-20180428030007-95032a82bc51
 ## explicit
 github.com/kballard/go-shellquote
-# github.com/klauspost/compress v1.15.9
+# github.com/klauspost/compress v1.15.15
-## explicit; go 1.16
+## explicit; go 1.17
 github.com/klauspost/compress/flate
 github.com/klauspost/compress/gzip
 github.com/klauspost/compress/s2
 github.com/klauspost/compress/snappy
 github.com/klauspost/compress/zlib
-# github.com/klauspost/cpuid/v2 v2.1.1
+# github.com/klauspost/cpuid/v2 v2.2.3
 ## explicit; go 1.15
 github.com/klauspost/cpuid/v2
 # github.com/leodido/go-urn v1.2.1
@@ -353,7 +353,7 @@ github.com/miekg/dns
 # github.com/minio/md5-simd v1.1.2
 ## explicit; go 1.14
 github.com/minio/md5-simd
-# github.com/minio/minio-go/v7 v7.0.48
+# github.com/minio/minio-go/v7 v7.0.49
 ## explicit; go 1.17
 github.com/minio/minio-go/v7
 github.com/minio/minio-go/v7/pkg/credentials