tests/fuzzers: update fuzzers to be based on go-native fuzzing (#28352)

This change modifies the fuzzers to use the native golang fuzzing framework instead of go-fuzz

crypto/blake2b/blake2b_f_fuzz.go

@@ -1,58 +0,0 @@
-//go:build gofuzz
-// +build gofuzz
-
-package blake2b
-
-import (
-	"encoding/binary"
-)
-
-func Fuzz(data []byte) int {
-	// Make sure the data confirms to the input model
-	if len(data) != 211 {
-		return 0
-	}
-	// Parse everything and call all the implementations
-	var (
-		rounds = binary.BigEndian.Uint16(data[0:2])
-
-		h [8]uint64
-		m [16]uint64
-		t [2]uint64
-		f uint64
-	)
-	for i := 0; i < 8; i++ {
-		offset := 2 + i*8
-		h[i] = binary.LittleEndian.Uint64(data[offset : offset+8])
-	}
-	for i := 0; i < 16; i++ {
-		offset := 66 + i*8
-		m[i] = binary.LittleEndian.Uint64(data[offset : offset+8])
-	}
-	t[0] = binary.LittleEndian.Uint64(data[194:202])
-	t[1] = binary.LittleEndian.Uint64(data[202:210])
-
-	if data[210]%2 == 1 { // Avoid spinning the fuzzer to hit 0/1
-		f = 0xFFFFFFFFFFFFFFFF
-	}
-	// Run the blake2b compression on all instruction sets and cross reference
-	want := h
-	fGeneric(&want, &m, t[0], t[1], f, uint64(rounds))
-
-	have := h
-	fSSE4(&have, &m, t[0], t[1], f, uint64(rounds))
-	if have != want {
-		panic("SSE4 mismatches generic algo")
-	}
-	have = h
-	fAVX(&have, &m, t[0], t[1], f, uint64(rounds))
-	if have != want {
-		panic("AVX mismatches generic algo")
-	}
-	have = h
-	fAVX2(&have, &m, t[0], t[1], f, uint64(rounds))
-	if have != want {
-		panic("AVX2 mismatches generic algo")
-	}
-	return 1
-}

crypto/blake2b/blake2b_f_test.go

@@ -1,6 +1,7 @@
 package blake2b
 
 import (
+	"encoding/binary"
 	"fmt"
 	"reflect"
 	"testing"
@@ -57,3 +58,60 @@ var testVectorsF = []testVector{
 		},
 	},
 }
+
+func Fuzz(f *testing.F) {
+	f.Fuzz(func(t *testing.T, data []byte) {
+		fuzz(data)
+	})
+}
+
+func fuzz(data []byte) {
+	// Make sure the data confirms to the input model
+	if len(data) != 211 {
+		return
+	}
+	// Parse everything and call all the implementations
+	var (
+		rounds = binary.BigEndian.Uint16(data[0:2])
+
+		h [8]uint64
+		m [16]uint64
+		t [2]uint64
+		f uint64
+	)
+
+	for i := 0; i < 8; i++ {
+		offset := 2 + i*8
+		h[i] = binary.LittleEndian.Uint64(data[offset : offset+8])
+	}
+	for i := 0; i < 16; i++ {
+		offset := 66 + i*8
+		m[i] = binary.LittleEndian.Uint64(data[offset : offset+8])
+	}
+	t[0] = binary.LittleEndian.Uint64(data[194:202])
+	t[1] = binary.LittleEndian.Uint64(data[202:210])
+
+	if data[210]%2 == 1 { // Avoid spinning the fuzzer to hit 0/1
+		f = 0xFFFFFFFFFFFFFFFF
+	}
+
+	// Run the blake2b compression on all instruction sets and cross reference
+	want := h
+	fGeneric(&want, &m, t[0], t[1], f, uint64(rounds))
+
+	have := h
+	fSSE4(&have, &m, t[0], t[1], f, uint64(rounds))
+	if have != want {
+		panic("SSE4 mismatches generic algo")
+	}
+	have = h
+	fAVX(&have, &m, t[0], t[1], f, uint64(rounds))
+	if have != want {
+		panic("AVX mismatches generic algo")
+	}
+	have = h
+	fAVX2(&have, &m, t[0], t[1], f, uint64(rounds))
+	if have != want {
+		panic("AVX2 mismatches generic algo")
+	}
+}