[VOL-5486] Fix deprecated versions
Change-Id: I3e03ea246020547ae75fa92ce8cf5cbba7e8f3bb
Signed-off-by: Abhay Kumar <abhay.kumar@radisys.com>
diff --git a/vendor/github.com/klauspost/compress/zstd/enc_best.go b/vendor/github.com/klauspost/compress/zstd/enc_best.go
index 96028ec..4613724 100644
--- a/vendor/github.com/klauspost/compress/zstd/enc_best.go
+++ b/vendor/github.com/klauspost/compress/zstd/enc_best.go
@@ -34,7 +34,7 @@
est int32
}
-const highScore = 25000
+const highScore = maxMatchLen * 8
// estBits will estimate output bits from predefined tables.
func (m *match) estBits(bitsPerByte int32) {
@@ -43,7 +43,7 @@
if m.rep < 0 {
ofc = ofCode(uint32(m.s-m.offset) + 3)
} else {
- ofc = ofCode(uint32(m.rep))
+ ofc = ofCode(uint32(m.rep) & 3)
}
// Cost, excluding
ofTT, mlTT := fsePredefEnc[tableOffsets].ct.symbolTT[ofc], fsePredefEnc[tableMatchLengths].ct.symbolTT[mlc]
@@ -84,14 +84,10 @@
)
// Protect against e.cur wraparound.
- for e.cur >= bufferReset {
+ for e.cur >= e.bufferReset-int32(len(e.hist)) {
if len(e.hist) == 0 {
- for i := range e.table[:] {
- e.table[i] = prevEntry{}
- }
- for i := range e.longTable[:] {
- e.longTable[i] = prevEntry{}
- }
+ e.table = [bestShortTableSize]prevEntry{}
+ e.longTable = [bestLongTableSize]prevEntry{}
e.cur = e.maxMatchOff
break
}
@@ -139,8 +135,20 @@
break
}
+ // Add block to history
s := e.addBlock(src)
blk.size = len(src)
+
+ // Check RLE first
+ if len(src) > zstdMinMatch {
+ ml := matchLen(src[1:], src)
+ if ml == len(src)-1 {
+ blk.literals = append(blk.literals, src[0])
+ blk.sequences = append(blk.sequences, seq{litLen: 1, matchLen: uint32(len(src)-1) - zstdMinMatch, offset: 1 + 3})
+ return
+ }
+ }
+
if len(src) < minNonLiteralBlockSize {
blk.extraLits = len(src)
blk.literals = blk.literals[:len(src)]
@@ -163,7 +171,6 @@
// nextEmit is where in src the next emitLiteral should start from.
nextEmit := s
- cv := load6432(src, s)
// Relative offsets
offset1 := int32(blk.recentOffsets[0])
@@ -177,7 +184,6 @@
blk.literals = append(blk.literals, src[nextEmit:until]...)
s.litLen = uint32(until - nextEmit)
}
- _ = addLiterals
if debugEncoder {
println("recent offsets:", blk.recentOffsets)
@@ -192,54 +198,104 @@
panic("offset0 was 0")
}
- bestOf := func(a, b match) match {
- if a.est+(a.s-b.s)*bitsPerByte>>10 < b.est+(b.s-a.s)*bitsPerByte>>10 {
- return a
- }
- return b
- }
- const goodEnough = 100
+ const goodEnough = 250
+
+ cv := load6432(src, s)
nextHashL := hashLen(cv, bestLongTableBits, bestLongLen)
nextHashS := hashLen(cv, bestShortTableBits, bestShortLen)
candidateL := e.longTable[nextHashL]
candidateS := e.table[nextHashS]
- matchAt := func(offset int32, s int32, first uint32, rep int32) match {
- if s-offset >= e.maxMatchOff || load3232(src, offset) != first {
- return match{s: s, est: highScore}
+ // Set m to a match at offset if it looks like that will improve compression.
+ improve := func(m *match, offset int32, s int32, first uint32, rep int32) {
+ delta := s - offset
+ if delta >= e.maxMatchOff || delta <= 0 || load3232(src, offset) != first {
+ return
}
- if debugAsserts {
- if !bytes.Equal(src[s:s+4], src[offset:offset+4]) {
- panic(fmt.Sprintf("first match mismatch: %v != %v, first: %08x", src[s:s+4], src[offset:offset+4], first))
+ // Try to quick reject if we already have a long match.
+ if m.length > 16 {
+ left := len(src) - int(m.s+m.length)
+ // If we are too close to the end, keep as is.
+ if left <= 0 {
+ return
+ }
+ checkLen := m.length - (s - m.s) - 8
+ if left > 2 && checkLen > 4 {
+ // Check 4 bytes, 4 bytes from the end of the current match.
+ a := load3232(src, offset+checkLen)
+ b := load3232(src, s+checkLen)
+ if a != b {
+ return
+ }
}
}
- m := match{offset: offset, s: s, length: 4 + e.matchlen(s+4, offset+4, src), rep: rep}
- m.estBits(bitsPerByte)
- return m
+ l := 4 + e.matchlen(s+4, offset+4, src)
+ if m.rep <= 0 {
+ // Extend candidate match backwards as far as possible.
+ // Do not extend repeats as we can assume they are optimal
+ // and offsets change if s == nextEmit.
+ tMin := s - e.maxMatchOff
+ if tMin < 0 {
+ tMin = 0
+ }
+ for offset > tMin && s > nextEmit && src[offset-1] == src[s-1] && l < maxMatchLength {
+ s--
+ offset--
+ l++
+ }
+ }
+ if debugAsserts {
+ if offset >= s {
+ panic(fmt.Sprintf("offset: %d - s:%d - rep: %d - cur :%d - max: %d", offset, s, rep, e.cur, e.maxMatchOff))
+ }
+ if !bytes.Equal(src[s:s+l], src[offset:offset+l]) {
+ panic(fmt.Sprintf("second match mismatch: %v != %v, first: %08x", src[s:s+4], src[offset:offset+4], first))
+ }
+ }
+ cand := match{offset: offset, s: s, length: l, rep: rep}
+ cand.estBits(bitsPerByte)
+ if m.est >= highScore || cand.est-m.est+(cand.s-m.s)*bitsPerByte>>10 < 0 {
+ *m = cand
+ }
}
- best := bestOf(matchAt(candidateL.offset-e.cur, s, uint32(cv), -1), matchAt(candidateL.prev-e.cur, s, uint32(cv), -1))
- best = bestOf(best, matchAt(candidateS.offset-e.cur, s, uint32(cv), -1))
- best = bestOf(best, matchAt(candidateS.prev-e.cur, s, uint32(cv), -1))
+ best := match{s: s, est: highScore}
+ improve(&best, candidateL.offset-e.cur, s, uint32(cv), -1)
+ improve(&best, candidateL.prev-e.cur, s, uint32(cv), -1)
+ improve(&best, candidateS.offset-e.cur, s, uint32(cv), -1)
+ improve(&best, candidateS.prev-e.cur, s, uint32(cv), -1)
if canRepeat && best.length < goodEnough {
- cv32 := uint32(cv >> 8)
- spp := s + 1
- best = bestOf(best, matchAt(spp-offset1, spp, cv32, 1))
- best = bestOf(best, matchAt(spp-offset2, spp, cv32, 2))
- best = bestOf(best, matchAt(spp-offset3, spp, cv32, 3))
- if best.length > 0 {
- cv32 = uint32(cv >> 24)
- spp += 2
- best = bestOf(best, matchAt(spp-offset1, spp, cv32, 1))
- best = bestOf(best, matchAt(spp-offset2, spp, cv32, 2))
- best = bestOf(best, matchAt(spp-offset3, spp, cv32, 3))
+ if s == nextEmit {
+ // Check repeats straight after a match.
+ improve(&best, s-offset2, s, uint32(cv), 1|4)
+ improve(&best, s-offset3, s, uint32(cv), 2|4)
+ if offset1 > 1 {
+ improve(&best, s-(offset1-1), s, uint32(cv), 3|4)
+ }
+ }
+
+ // If either no match or a non-repeat match, check at + 1
+ if best.rep <= 0 {
+ cv32 := uint32(cv >> 8)
+ spp := s + 1
+ improve(&best, spp-offset1, spp, cv32, 1)
+ improve(&best, spp-offset2, spp, cv32, 2)
+ improve(&best, spp-offset3, spp, cv32, 3)
+ if best.rep < 0 {
+ cv32 = uint32(cv >> 24)
+ spp += 2
+ improve(&best, spp-offset1, spp, cv32, 1)
+ improve(&best, spp-offset2, spp, cv32, 2)
+ improve(&best, spp-offset3, spp, cv32, 3)
+ }
}
}
// Load next and check...
e.longTable[nextHashL] = prevEntry{offset: s + e.cur, prev: candidateL.offset}
e.table[nextHashS] = prevEntry{offset: s + e.cur, prev: candidateS.offset}
+ index0 := s + 1
// Look far ahead, unless we have a really long match already...
if best.length < goodEnough {
@@ -249,97 +305,89 @@
if s >= sLimit {
break encodeLoop
}
- cv = load6432(src, s)
continue
}
- s++
candidateS = e.table[hashLen(cv>>8, bestShortTableBits, bestShortLen)]
- cv = load6432(src, s)
- cv2 := load6432(src, s+1)
+ cv = load6432(src, s+1)
+ cv2 := load6432(src, s+2)
candidateL = e.longTable[hashLen(cv, bestLongTableBits, bestLongLen)]
candidateL2 := e.longTable[hashLen(cv2, bestLongTableBits, bestLongLen)]
// Short at s+1
- best = bestOf(best, matchAt(candidateS.offset-e.cur, s, uint32(cv), -1))
+ improve(&best, candidateS.offset-e.cur, s+1, uint32(cv), -1)
// Long at s+1, s+2
- best = bestOf(best, matchAt(candidateL.offset-e.cur, s, uint32(cv), -1))
- best = bestOf(best, matchAt(candidateL.prev-e.cur, s, uint32(cv), -1))
- best = bestOf(best, matchAt(candidateL2.offset-e.cur, s+1, uint32(cv2), -1))
- best = bestOf(best, matchAt(candidateL2.prev-e.cur, s+1, uint32(cv2), -1))
+ improve(&best, candidateL.offset-e.cur, s+1, uint32(cv), -1)
+ improve(&best, candidateL.prev-e.cur, s+1, uint32(cv), -1)
+ improve(&best, candidateL2.offset-e.cur, s+2, uint32(cv2), -1)
+ improve(&best, candidateL2.prev-e.cur, s+2, uint32(cv2), -1)
if false {
// Short at s+3.
// Too often worse...
- best = bestOf(best, matchAt(e.table[hashLen(cv2>>8, bestShortTableBits, bestShortLen)].offset-e.cur, s+2, uint32(cv2>>8), -1))
+ improve(&best, e.table[hashLen(cv2>>8, bestShortTableBits, bestShortLen)].offset-e.cur, s+3, uint32(cv2>>8), -1)
}
- // See if we can find a better match by checking where the current best ends.
- // Use that offset to see if we can find a better full match.
- if sAt := best.s + best.length; sAt < sLimit {
- nextHashL := hashLen(load6432(src, sAt), bestLongTableBits, bestLongLen)
- candidateEnd := e.longTable[nextHashL]
- if pos := candidateEnd.offset - e.cur - best.length; pos >= 0 {
- bestEnd := bestOf(best, matchAt(pos, best.s, load3232(src, best.s), -1))
- if pos := candidateEnd.prev - e.cur - best.length; pos >= 0 {
- bestEnd = bestOf(bestEnd, matchAt(pos, best.s, load3232(src, best.s), -1))
+
+ // Start check at a fixed offset to allow for a few mismatches.
+ // For this compression level 2 yields the best results.
+ // We cannot do this if we have already indexed this position.
+ const skipBeginning = 2
+ if best.s > s-skipBeginning {
+ // See if we can find a better match by checking where the current best ends.
+ // Use that offset to see if we can find a better full match.
+ if sAt := best.s + best.length; sAt < sLimit {
+ nextHashL := hashLen(load6432(src, sAt), bestLongTableBits, bestLongLen)
+ candidateEnd := e.longTable[nextHashL]
+
+ if off := candidateEnd.offset - e.cur - best.length + skipBeginning; off >= 0 {
+ improve(&best, off, best.s+skipBeginning, load3232(src, best.s+skipBeginning), -1)
+ if off := candidateEnd.prev - e.cur - best.length + skipBeginning; off >= 0 {
+ improve(&best, off, best.s+skipBeginning, load3232(src, best.s+skipBeginning), -1)
+ }
}
- best = bestEnd
}
}
}
if debugAsserts {
+ if best.offset >= best.s {
+ panic(fmt.Sprintf("best.offset > s: %d >= %d", best.offset, best.s))
+ }
+ if best.s < nextEmit {
+ panic(fmt.Sprintf("s %d < nextEmit %d", best.s, nextEmit))
+ }
+ if best.offset < s-e.maxMatchOff {
+ panic(fmt.Sprintf("best.offset < s-e.maxMatchOff: %d < %d", best.offset, s-e.maxMatchOff))
+ }
if !bytes.Equal(src[best.s:best.s+best.length], src[best.offset:best.offset+best.length]) {
panic(fmt.Sprintf("match mismatch: %v != %v", src[best.s:best.s+best.length], src[best.offset:best.offset+best.length]))
}
}
// We have a match, we can store the forward value
+ s = best.s
if best.rep > 0 {
- s = best.s
var seq seq
seq.matchLen = uint32(best.length - zstdMinMatch)
+ addLiterals(&seq, best.s)
- // We might be able to match backwards.
- // Extend as long as we can.
- start := best.s
- // We end the search early, so we don't risk 0 literals
- // and have to do special offset treatment.
- startLimit := nextEmit + 1
-
- tMin := s - e.maxMatchOff
- if tMin < 0 {
- tMin = 0
- }
- repIndex := best.offset
- for repIndex > tMin && start > startLimit && src[repIndex-1] == src[start-1] && seq.matchLen < maxMatchLength-zstdMinMatch-1 {
- repIndex--
- start--
- seq.matchLen++
- }
- addLiterals(&seq, start)
-
- // rep 0
- seq.offset = uint32(best.rep)
+ // Repeat. If bit 4 is set, this is a non-lit repeat.
+ seq.offset = uint32(best.rep & 3)
if debugSequences {
- println("repeat sequence", seq, "next s:", s)
+ println("repeat sequence", seq, "next s:", best.s, "off:", best.s-best.offset)
}
blk.sequences = append(blk.sequences, seq)
- // Index match start+1 (long) -> s - 1
- index0 := s
+ // Index old s + 1 -> s - 1
s = best.s + best.length
-
nextEmit = s
- if s >= sLimit {
- if debugEncoder {
- println("repeat ended", s, best.length)
- }
- break encodeLoop
- }
// Index skipped...
+ end := s
+ if s > sLimit+4 {
+ end = sLimit + 4
+ }
off := index0 + e.cur
- for index0 < s-1 {
+ for index0 < end {
cv0 := load6432(src, index0)
h0 := hashLen(cv0, bestLongTableBits, bestLongLen)
h1 := hashLen(cv0, bestShortTableBits, bestShortLen)
@@ -348,19 +396,26 @@
off++
index0++
}
+
switch best.rep {
- case 2:
+ case 2, 4 | 1:
offset1, offset2 = offset2, offset1
- case 3:
+ case 3, 4 | 2:
offset1, offset2, offset3 = offset3, offset1, offset2
+ case 4 | 3:
+ offset1, offset2, offset3 = offset1-1, offset1, offset2
}
- cv = load6432(src, s)
+ if s >= sLimit {
+ if debugEncoder {
+ println("repeat ended", s, best.length)
+ }
+ break encodeLoop
+ }
continue
}
// A 4-byte match has been found. Update recent offsets.
// We'll later see if more than 4 bytes.
- s = best.s
t := best.offset
offset1, offset2, offset3 = s-t, offset1, offset2
@@ -372,22 +427,9 @@
panic("invalid offset")
}
- // Extend the n-byte match as long as possible.
- l := best.length
-
- // Extend backwards
- tMin := s - e.maxMatchOff
- if tMin < 0 {
- tMin = 0
- }
- for t > tMin && s > nextEmit && src[t-1] == src[s-1] && l < maxMatchLength {
- s--
- t--
- l++
- }
-
// Write our sequence
var seq seq
+ l := best.length
seq.litLen = uint32(s - nextEmit)
seq.matchLen = uint32(l - zstdMinMatch)
if seq.litLen > 0 {
@@ -400,65 +442,25 @@
}
blk.sequences = append(blk.sequences, seq)
nextEmit = s
- if s >= sLimit {
- break encodeLoop
+
+ // Index old s + 1 -> s - 1 or sLimit
+ end := s
+ if s > sLimit-4 {
+ end = sLimit - 4
}
- // Index match start+1 (long) -> s - 1
- index0 := s - l + 1
- // every entry
- for index0 < s-1 {
+ off := index0 + e.cur
+ for index0 < end {
cv0 := load6432(src, index0)
h0 := hashLen(cv0, bestLongTableBits, bestLongLen)
h1 := hashLen(cv0, bestShortTableBits, bestShortLen)
- off := index0 + e.cur
e.longTable[h0] = prevEntry{offset: off, prev: e.longTable[h0].offset}
e.table[h1] = prevEntry{offset: off, prev: e.table[h1].offset}
index0++
+ off++
}
-
- cv = load6432(src, s)
- if !canRepeat {
- continue
- }
-
- // Check offset 2
- for {
- o2 := s - offset2
- if load3232(src, o2) != uint32(cv) {
- // Do regular search
- break
- }
-
- // Store this, since we have it.
- nextHashS := hashLen(cv, bestShortTableBits, bestShortLen)
- nextHashL := hashLen(cv, bestLongTableBits, bestLongLen)
-
- // We have at least 4 byte match.
- // No need to check backwards. We come straight from a match
- l := 4 + e.matchlen(s+4, o2+4, src)
-
- e.longTable[nextHashL] = prevEntry{offset: s + e.cur, prev: e.longTable[nextHashL].offset}
- e.table[nextHashS] = prevEntry{offset: s + e.cur, prev: e.table[nextHashS].offset}
- seq.matchLen = uint32(l) - zstdMinMatch
- seq.litLen = 0
-
- // Since litlen is always 0, this is offset 1.
- seq.offset = 1
- s += l
- nextEmit = s
- if debugSequences {
- println("sequence", seq, "next s:", s)
- }
- blk.sequences = append(blk.sequences, seq)
-
- // Swap offset 1 and 2.
- offset1, offset2 = offset2, offset1
- if s >= sLimit {
- // Finished
- break encodeLoop
- }
- cv = load6432(src, s)
+ if s >= sLimit {
+ break encodeLoop
}
}