hash/maphash: add WriteComparable and Comparable

By default, runtime.memhash is used.
When purego is used, reflect is used to generate the same []byte
with the same value, and then hash the []byte.

Fixes golang#54670

Change-Id: Ibd0538a7dfb3d831c5145970cac7c910692bca69

api/next/54670.txt

@@ -0,0 +1,2 @@
+pkg hash/maphash, func Comparable[$0 comparable](Seed, $0) uint64 #54670
+pkg hash/maphash, func WriteComparable[$0 comparable](*Hash, $0) #54670

doc/next/6-stdlib/99-minor/hash/maphash/54670.md

@@ -0,0 +1,2 @@
+New function [Comparable] returns the hash of comparable value v.
+New function [WriteComparable] adds x to the data hashed by [Hash].

src/hash/maphash/maphash.go

@@ -12,6 +12,11 @@
 // (See crypto/sha256 and crypto/sha512 for cryptographic use.)
 package maphash
 
+import (
+	"internal/abi"
+	"internal/byteorder"
+)
+
 // A Seed is a random value that selects the specific hash function
 // computed by a [Hash]. If two Hashes use the same Seeds, they
 // will compute the same hash values for any given input.
@@ -275,3 +280,24 @@ func (h *Hash) Size() int { return 8 }
 
 // BlockSize returns h's block size.
 func (h *Hash) BlockSize() int { return len(h.buf) }
+
+// Comparable returns the hash of comparable value v with the given seed
+// such that Comparable(s, v1) == Comparable(s, v2) if v1 == v2.
+// If v contains a floating-point NaN, then the hash is non-deterministically random.
+func Comparable[T comparable](seed Seed, v T) uint64 {
+	abi.Escape(v)
+	t := abi.TypeOf(v)
+	len := t.Size()
+	if len == 0 {
+		return seed.s
+	}
+	return comparableF(seed.s, v, t)
+}
+
+// WriteComparable adds x to the data hashed by h.
+func WriteComparable[T comparable](h *Hash, x T) {
+	v := Comparable(h.seed, x)
+	var buf [8]byte
+	byteorder.LeAppendUint64(buf[:], v)
+	h.Write(buf[:])
+}

src/hash/maphash/maphash_purego.go

@@ -8,8 +8,10 @@ package maphash
 
 import (
 	"crypto/rand"
+	"internal/abi"
 	"internal/byteorder"
 	"math/bits"
+	"reflect"
 )
 
 func rthash(buf []byte, seed uint64) uint64 {
@@ -92,3 +94,62 @@ func mix(a, b uint64) uint64 {
 	hi, lo := bits.Mul64(a, b)
 	return hi ^ lo
 }
+
+var byteSliceTyp = reflect.TypeFor[[]byte]()
+
+var strSliceTyp = reflect.TypeFor[string]()
+
+func comparableF[T comparable](seed uint64, v T, t *abi.Type) uint64 {
+	vv := reflect.ValueOf(v)
+	typ := vv.Type()
+	if typ == byteSliceTyp {
+		return wyhash(vv.Bytes(), seed, uint64(vv.Len()))
+	}
+	if typ == strSliceTyp {
+		return rthashString(vv.String(), seed)
+	}
+	buf := make([]byte, 0, typ.Size())
+	appendT(buf, vv)
+	return wyhash(buf, seed, uint64(len(buf)))
+}
+
+func appendT(buf []byte, v reflect.Value) []byte {
+	switch v.Kind() {
+	case reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64, reflect.Int:
+		return byteorder.LeAppendUint64(buf, uint64(v.Int()))
+	case reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uint, reflect.Uintptr:
+		return byteorder.LeAppendUint64(buf, v.Uint())
+	case reflect.Array, reflect.Slice:
+		for i := range v.Len() {
+			buf = appendT(buf, v.Index(i))
+		}
+	case reflect.String:
+		return append(buf, v.String()...)
+	case reflect.Struct:
+		for i := range v.NumField() {
+			buf = appendT(buf, v.Field(i))
+		}
+	case reflect.Complex64, reflect.Complex128:
+		c := v.Complex()
+		buf = byteorder.LeAppendUint64(buf, uint64(real(c)))
+		return byteorder.LeAppendUint64(buf, uint64(imag(c)))
+	case reflect.Float32, reflect.Float64:
+		return byteorder.LeAppendUint64(buf, uint64(v.Float()))
+	case reflect.Bool:
+		return byteorder.LeAppendUint16(buf, btoi(v.Bool()))
+	case reflect.UnsafePointer, reflect.Pointer:
+		return byteorder.LeAppendUint64(buf, uint64(v.Pointer()))
+	case reflect.Interface:
+		a := v.InterfaceData()
+		buf = byteorder.LeAppendUint64(buf, uint64(a[0]))
+		return byteorder.LeAppendUint64(buf, uint64(a[1]))
+	}
+	panic("unreachable")
+}
+
+func btoi(b bool) uint16 {
+	if b {
+		return 1
+	}
+	return 0
+}

src/hash/maphash/maphash_runtime.go

@@ -7,6 +7,8 @@
 package maphash
 
 import (
+	"internal/abi"
+	"internal/unsafeheader"
 	"unsafe"
 )
 
@@ -41,3 +43,23 @@ func rthashString(s string, state uint64) uint64 {
 func randUint64() uint64 {
 	return runtime_rand()
 }
+
+func comparableF[T comparable](seed uint64, v T, t *abi.Type) uint64 {
+	k := t.Kind()
+	len := t.Size()
+	ptr := unsafe.Pointer(&v)
+	switch k {
+	case abi.Slice:
+		len = uintptr(((*unsafeheader.Slice)(unsafe.Pointer(&v))).Len) * t.Elem().Size()
+		ptr = ((*unsafeheader.Slice)(unsafe.Pointer(&v))).Data
+	case abi.String:
+		len = uintptr(((*unsafeheader.String)(unsafe.Pointer(&v))).Len)
+		ptr = ((*unsafeheader.String)(unsafe.Pointer(&v))).Data
+	}
+	if unsafe.Sizeof(uintptr(0)) == 8 {
+		return uint64(runtime_memhash(ptr, uintptr(seed), len))
+	}
+	lo := runtime_memhash(ptr, uintptr(seed), len)
+	hi := runtime_memhash(ptr, uintptr(seed>>32), len)
+	return uint64(hi)<<32 | uint64(lo)
+}

src/hash/maphash/maphash_test.go

@@ -210,6 +210,53 @@ func TestSeedFromReset(t *testing.T) {
 	}
 }
 
+func TestCompare(t *testing.T) {
+	var a, b int = 2, 2
+	var pa *int = &a
+	seed := MakeSeed()
+	if Comparable(seed, a) != Comparable(seed, b) {
+		t.Fatal("Comparable(seed, 2) != Comparable(seed, 2)")
+	}
+	old := Comparable(seed, pa)
+	stackGrow(8192)
+	new := Comparable(seed, pa)
+	if old != new {
+		t.Fatal("Comparable(seed, ptr) != Comparable(seed, ptr)")
+	}
+}
+
+//go:noinline
+func stackGrow(dep int) {
+	if dep == 0 {
+		return
+	}
+	var local [1024]byte
+	_ = local
+	stackGrow(dep - 1)
+}
+
+func TestWriteComparable(t *testing.T) {
+	var a, b int = 2, 2
+	var pa *int = &a
+	h1 := Hash{}
+	h2 := Hash{}
+	h1.seed = MakeSeed()
+	h2.seed = h1.seed
+	WriteComparable(&h1, a)
+	WriteComparable(&h2, b)
+	if h1.Sum64() != h1.Sum64() {
+		t.Fatal("WriteComparable(h, 2) != WriteComparable(h, 2)")
+	}
+	WriteComparable(&h1, pa)
+	old := h1.Sum64()
+	stackGrow(8192)
+	WriteComparable(&h2, pa)
+	new := h2.Sum64()
+	if old != new {
+		t.Fatal("WriteComparable(seed, ptr) != WriteComparable(seed, ptr)")
+	}
+}
+
 // Make sure a Hash implements the hash.Hash and hash.Hash64 interfaces.
 var _ hash.Hash = &Hash{}
 var _ hash.Hash64 = &Hash{}