From 6016eed2894713a69e145136ef27cdb13bf3fef2 Mon Sep 17 00:00:00 2001
From: "@RubenKelevra" <cyrond@gmail.com>
Date: Wed, 7 Jul 2021 03:56:44 +0200
Subject: [PATCH] buzhash: reduce target size and cutoff size

Fixes https://github.com/ipfs/go-ipfs/issues/7966
---
 buzhash.go      | 28 +++++++++++++++++++++-------
 buzhash_test.go |  8 +++++---
 parse.go        |  5 ++++-
 3 files changed, 30 insertions(+), 11 deletions(-)

diff --git a/buzhash.go b/buzhash.go
index 83ab019..f8f121a 100644
--- a/buzhash.go
+++ b/buzhash.go
@@ -8,9 +8,11 @@ import (
 )
 
 const (
-	buzMin  = 128 << 10
-	buzMax  = 512 << 10
-	buzMask = 1<<17 - 1
+	buzMinDefault = 16 << 10
+	buzMaxDefault = 64 << 10
+	buzMinLegacy  = 128 << 10
+	buzMaxLegacy  = 512 << 10
+	buzMask       = 1<<17 - 1
 )
 
 type Buzhash struct {
@@ -18,13 +20,20 @@ type Buzhash struct {
 	buf []byte
 	n   int
 
-	err error
+	err       error
+	buzLegacy bool
 }
 
-func NewBuzhash(r io.Reader) *Buzhash {
+func NewBuzhash(r io.Reader, buzLegacy bool) *Buzhash {
+	buzMax := buzMaxDefault
+	if buzLegacy {
+		buzMax = buzMaxLegacy
+	}
+
 	return &Buzhash{
-		r:   r,
-		buf: pool.Get(buzMax),
+		r:         r,
+		buf:       pool.Get(buzMax),
+		buzLegacy: buzLegacy,
 	}
 }
 
@@ -37,6 +46,11 @@ func (b *Buzhash) NextBytes() ([]byte, error) {
 		return nil, b.err
 	}
 
+	buzMin := buzMinDefault
+	if b.buzLegacy {
+		buzMin = buzMinLegacy
+	}
+
 	n, err := io.ReadFull(b.r, b.buf[b.n:])
 	if err != nil {
 		if err == io.ErrUnexpectedEOF || err == io.EOF {
diff --git a/buzhash_test.go b/buzhash_test.go
index 05ad7c3..abb78c4 100644
--- a/buzhash_test.go
+++ b/buzhash_test.go
@@ -17,10 +17,12 @@ func testBuzhashChunking(t *testing.T, buf []byte) (chunkCount int) {
 		t.Fatal(err)
 	}
 
-	r := NewBuzhash(bytes.NewReader(buf))
+	r := NewBuzhash(bytes.NewReader(buf), false)
 
 	var chunks [][]byte
 
+	buzMin := buzMinDefault
+
 	for {
 		chunk, err := r.NextBytes()
 		if err != nil {
@@ -62,14 +64,14 @@ func TestBuzhashChunking(t *testing.T) {
 
 func TestBuzhashChunkReuse(t *testing.T) {
 	newBuzhash := func(r io.Reader) Splitter {
-		return NewBuzhash(r)
+		return NewBuzhash(r, false)
 	}
 	testReuse(t, newBuzhash)
 }
 
 func BenchmarkBuzhash2(b *testing.B) {
 	benchmarkChunker(b, func(r io.Reader) Splitter {
-		return NewBuzhash(r)
+		return NewBuzhash(r, false)
 	})
 }
 
diff --git a/parse.go b/parse.go
index dee8304..32a9e81 100644
--- a/parse.go
+++ b/parse.go
@@ -48,7 +48,10 @@ func FromString(r io.Reader, chunker string) (Splitter, error) {
 		return parseRabinString(r, chunker)
 
 	case chunker == "buzhash":
-		return NewBuzhash(r), nil
+		return NewBuzhash(r, false), nil
+
+	case chunker == "buzhash-legacy":
+		return NewBuzhash(r, true), nil
 
 	default:
 		return nil, fmt.Errorf("unrecognized chunker option: %s", chunker)