vendor: add uncommitted files from previous change
Nick Craig-Wood
parent
92662baceb
commit
e0d41da3e3
|
|
@ -0,0 +1,17 @@
|
|||
// Copyright 2018 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build go1.11
|
||||
// +build !gccgo,!appengine
|
||||
|
||||
package chacha20
|
||||
|
||||
const bufSize = 256
|
||||
|
||||
//go:noescape
|
||||
func xorKeyStreamVX(dst, src []byte, key *[8]uint32, nonce *[3]uint32, counter *uint32)
|
||||
|
||||
func (c *Cipher) xorKeyStreamBlocks(dst, src []byte) {
|
||||
xorKeyStreamVX(dst, src, &c.key, &c.nonce, &c.counter)
|
||||
}
|
||||
|
|
@ -0,0 +1,308 @@
|
|||
// Copyright 2018 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build go1.11
|
||||
// +build !gccgo,!appengine
|
||||
|
||||
#include "textflag.h"
|
||||
|
||||
#define NUM_ROUNDS 10
|
||||
|
||||
// func xorKeyStreamVX(dst, src []byte, key *[8]uint32, nonce *[3]uint32, counter *uint32)
|
||||
TEXT ·xorKeyStreamVX(SB), NOSPLIT, $0
|
||||
MOVD dst+0(FP), R1
|
||||
MOVD src+24(FP), R2
|
||||
MOVD src_len+32(FP), R3
|
||||
MOVD key+48(FP), R4
|
||||
MOVD nonce+56(FP), R6
|
||||
MOVD counter+64(FP), R7
|
||||
|
||||
MOVD $·constants(SB), R10
|
||||
MOVD $·incRotMatrix(SB), R11
|
||||
|
||||
MOVW (R7), R20
|
||||
|
||||
AND $~255, R3, R13
|
||||
ADD R2, R13, R12 // R12 for block end
|
||||
AND $255, R3, R13
|
||||
loop:
|
||||
MOVD $NUM_ROUNDS, R21
|
||||
VLD1 (R11), [V30.S4, V31.S4]
|
||||
|
||||
// load contants
|
||||
// VLD4R (R10), [V0.S4, V1.S4, V2.S4, V3.S4]
|
||||
WORD $0x4D60E940
|
||||
|
||||
// load keys
|
||||
// VLD4R 16(R4), [V4.S4, V5.S4, V6.S4, V7.S4]
|
||||
WORD $0x4DFFE884
|
||||
// VLD4R 16(R4), [V8.S4, V9.S4, V10.S4, V11.S4]
|
||||
WORD $0x4DFFE888
|
||||
SUB $32, R4
|
||||
|
||||
// load counter + nonce
|
||||
// VLD1R (R7), [V12.S4]
|
||||
WORD $0x4D40C8EC
|
||||
|
||||
// VLD3R (R6), [V13.S4, V14.S4, V15.S4]
|
||||
WORD $0x4D40E8CD
|
||||
|
||||
// update counter
|
||||
VADD V30.S4, V12.S4, V12.S4
|
||||
|
||||
chacha:
|
||||
// V0..V3 += V4..V7
|
||||
// V12..V15 <<<= ((V12..V15 XOR V0..V3), 16)
|
||||
VADD V0.S4, V4.S4, V0.S4
|
||||
VADD V1.S4, V5.S4, V1.S4
|
||||
VADD V2.S4, V6.S4, V2.S4
|
||||
VADD V3.S4, V7.S4, V3.S4
|
||||
VEOR V12.B16, V0.B16, V12.B16
|
||||
VEOR V13.B16, V1.B16, V13.B16
|
||||
VEOR V14.B16, V2.B16, V14.B16
|
||||
VEOR V15.B16, V3.B16, V15.B16
|
||||
VREV32 V12.H8, V12.H8
|
||||
VREV32 V13.H8, V13.H8
|
||||
VREV32 V14.H8, V14.H8
|
||||
VREV32 V15.H8, V15.H8
|
||||
// V8..V11 += V12..V15
|
||||
// V4..V7 <<<= ((V4..V7 XOR V8..V11), 12)
|
||||
VADD V8.S4, V12.S4, V8.S4
|
||||
VADD V9.S4, V13.S4, V9.S4
|
||||
VADD V10.S4, V14.S4, V10.S4
|
||||
VADD V11.S4, V15.S4, V11.S4
|
||||
VEOR V8.B16, V4.B16, V16.B16
|
||||
VEOR V9.B16, V5.B16, V17.B16
|
||||
VEOR V10.B16, V6.B16, V18.B16
|
||||
VEOR V11.B16, V7.B16, V19.B16
|
||||
VSHL $12, V16.S4, V4.S4
|
||||
VSHL $12, V17.S4, V5.S4
|
||||
VSHL $12, V18.S4, V6.S4
|
||||
VSHL $12, V19.S4, V7.S4
|
||||
VSRI $20, V16.S4, V4.S4
|
||||
VSRI $20, V17.S4, V5.S4
|
||||
VSRI $20, V18.S4, V6.S4
|
||||
VSRI $20, V19.S4, V7.S4
|
||||
|
||||
// V0..V3 += V4..V7
|
||||
// V12..V15 <<<= ((V12..V15 XOR V0..V3), 8)
|
||||
VADD V0.S4, V4.S4, V0.S4
|
||||
VADD V1.S4, V5.S4, V1.S4
|
||||
VADD V2.S4, V6.S4, V2.S4
|
||||
VADD V3.S4, V7.S4, V3.S4
|
||||
VEOR V12.B16, V0.B16, V12.B16
|
||||
VEOR V13.B16, V1.B16, V13.B16
|
||||
VEOR V14.B16, V2.B16, V14.B16
|
||||
VEOR V15.B16, V3.B16, V15.B16
|
||||
VTBL V31.B16, [V12.B16], V12.B16
|
||||
VTBL V31.B16, [V13.B16], V13.B16
|
||||
VTBL V31.B16, [V14.B16], V14.B16
|
||||
VTBL V31.B16, [V15.B16], V15.B16
|
||||
|
||||
// V8..V11 += V12..V15
|
||||
// V4..V7 <<<= ((V4..V7 XOR V8..V11), 7)
|
||||
VADD V12.S4, V8.S4, V8.S4
|
||||
VADD V13.S4, V9.S4, V9.S4
|
||||
VADD V14.S4, V10.S4, V10.S4
|
||||
VADD V15.S4, V11.S4, V11.S4
|
||||
VEOR V8.B16, V4.B16, V16.B16
|
||||
VEOR V9.B16, V5.B16, V17.B16
|
||||
VEOR V10.B16, V6.B16, V18.B16
|
||||
VEOR V11.B16, V7.B16, V19.B16
|
||||
VSHL $7, V16.S4, V4.S4
|
||||
VSHL $7, V17.S4, V5.S4
|
||||
VSHL $7, V18.S4, V6.S4
|
||||
VSHL $7, V19.S4, V7.S4
|
||||
VSRI $25, V16.S4, V4.S4
|
||||
VSRI $25, V17.S4, V5.S4
|
||||
VSRI $25, V18.S4, V6.S4
|
||||
VSRI $25, V19.S4, V7.S4
|
||||
|
||||
// V0..V3 += V5..V7, V4
|
||||
// V15,V12-V14 <<<= ((V15,V12-V14 XOR V0..V3), 16)
|
||||
VADD V0.S4, V5.S4, V0.S4
|
||||
VADD V1.S4, V6.S4, V1.S4
|
||||
VADD V2.S4, V7.S4, V2.S4
|
||||
VADD V3.S4, V4.S4, V3.S4
|
||||
VEOR V15.B16, V0.B16, V15.B16
|
||||
VEOR V12.B16, V1.B16, V12.B16
|
||||
VEOR V13.B16, V2.B16, V13.B16
|
||||
VEOR V14.B16, V3.B16, V14.B16
|
||||
VREV32 V12.H8, V12.H8
|
||||
VREV32 V13.H8, V13.H8
|
||||
VREV32 V14.H8, V14.H8
|
||||
VREV32 V15.H8, V15.H8
|
||||
|
||||
// V10 += V15; V5 <<<= ((V10 XOR V5), 12)
|
||||
// ...
|
||||
VADD V15.S4, V10.S4, V10.S4
|
||||
VADD V12.S4, V11.S4, V11.S4
|
||||
VADD V13.S4, V8.S4, V8.S4
|
||||
VADD V14.S4, V9.S4, V9.S4
|
||||
VEOR V10.B16, V5.B16, V16.B16
|
||||
VEOR V11.B16, V6.B16, V17.B16
|
||||
VEOR V8.B16, V7.B16, V18.B16
|
||||
VEOR V9.B16, V4.B16, V19.B16
|
||||
VSHL $12, V16.S4, V5.S4
|
||||
VSHL $12, V17.S4, V6.S4
|
||||
VSHL $12, V18.S4, V7.S4
|
||||
VSHL $12, V19.S4, V4.S4
|
||||
VSRI $20, V16.S4, V5.S4
|
||||
VSRI $20, V17.S4, V6.S4
|
||||
VSRI $20, V18.S4, V7.S4
|
||||
VSRI $20, V19.S4, V4.S4
|
||||
|
||||
// V0 += V5; V15 <<<= ((V0 XOR V15), 8)
|
||||
// ...
|
||||
VADD V5.S4, V0.S4, V0.S4
|
||||
VADD V6.S4, V1.S4, V1.S4
|
||||
VADD V7.S4, V2.S4, V2.S4
|
||||
VADD V4.S4, V3.S4, V3.S4
|
||||
VEOR V0.B16, V15.B16, V15.B16
|
||||
VEOR V1.B16, V12.B16, V12.B16
|
||||
VEOR V2.B16, V13.B16, V13.B16
|
||||
VEOR V3.B16, V14.B16, V14.B16
|
||||
VTBL V31.B16, [V12.B16], V12.B16
|
||||
VTBL V31.B16, [V13.B16], V13.B16
|
||||
VTBL V31.B16, [V14.B16], V14.B16
|
||||
VTBL V31.B16, [V15.B16], V15.B16
|
||||
|
||||
// V10 += V15; V5 <<<= ((V10 XOR V5), 7)
|
||||
// ...
|
||||
VADD V15.S4, V10.S4, V10.S4
|
||||
VADD V12.S4, V11.S4, V11.S4
|
||||
VADD V13.S4, V8.S4, V8.S4
|
||||
VADD V14.S4, V9.S4, V9.S4
|
||||
VEOR V10.B16, V5.B16, V16.B16
|
||||
VEOR V11.B16, V6.B16, V17.B16
|
||||
VEOR V8.B16, V7.B16, V18.B16
|
||||
VEOR V9.B16, V4.B16, V19.B16
|
||||
VSHL $7, V16.S4, V5.S4
|
||||
VSHL $7, V17.S4, V6.S4
|
||||
VSHL $7, V18.S4, V7.S4
|
||||
VSHL $7, V19.S4, V4.S4
|
||||
VSRI $25, V16.S4, V5.S4
|
||||
VSRI $25, V17.S4, V6.S4
|
||||
VSRI $25, V18.S4, V7.S4
|
||||
VSRI $25, V19.S4, V4.S4
|
||||
|
||||
SUB $1, R21
|
||||
CBNZ R21, chacha
|
||||
|
||||
// VLD4R (R10), [V16.S4, V17.S4, V18.S4, V19.S4]
|
||||
WORD $0x4D60E950
|
||||
|
||||
// VLD4R 16(R4), [V20.S4, V21.S4, V22.S4, V23.S4]
|
||||
WORD $0x4DFFE894
|
||||
VADD V30.S4, V12.S4, V12.S4
|
||||
VADD V16.S4, V0.S4, V0.S4
|
||||
VADD V17.S4, V1.S4, V1.S4
|
||||
VADD V18.S4, V2.S4, V2.S4
|
||||
VADD V19.S4, V3.S4, V3.S4
|
||||
// VLD4R 16(R4), [V24.S4, V25.S4, V26.S4, V27.S4]
|
||||
WORD $0x4DFFE898
|
||||
// restore R4
|
||||
SUB $32, R4
|
||||
|
||||
// load counter + nonce
|
||||
// VLD1R (R7), [V28.S4]
|
||||
WORD $0x4D40C8FC
|
||||
// VLD3R (R6), [V29.S4, V30.S4, V31.S4]
|
||||
WORD $0x4D40E8DD
|
||||
|
||||
VADD V20.S4, V4.S4, V4.S4
|
||||
VADD V21.S4, V5.S4, V5.S4
|
||||
VADD V22.S4, V6.S4, V6.S4
|
||||
VADD V23.S4, V7.S4, V7.S4
|
||||
VADD V24.S4, V8.S4, V8.S4
|
||||
VADD V25.S4, V9.S4, V9.S4
|
||||
VADD V26.S4, V10.S4, V10.S4
|
||||
VADD V27.S4, V11.S4, V11.S4
|
||||
VADD V28.S4, V12.S4, V12.S4
|
||||
VADD V29.S4, V13.S4, V13.S4
|
||||
VADD V30.S4, V14.S4, V14.S4
|
||||
VADD V31.S4, V15.S4, V15.S4
|
||||
|
||||
VZIP1 V1.S4, V0.S4, V16.S4
|
||||
VZIP2 V1.S4, V0.S4, V17.S4
|
||||
VZIP1 V3.S4, V2.S4, V18.S4
|
||||
VZIP2 V3.S4, V2.S4, V19.S4
|
||||
VZIP1 V5.S4, V4.S4, V20.S4
|
||||
VZIP2 V5.S4, V4.S4, V21.S4
|
||||
VZIP1 V7.S4, V6.S4, V22.S4
|
||||
VZIP2 V7.S4, V6.S4, V23.S4
|
||||
VZIP1 V9.S4, V8.S4, V24.S4
|
||||
VZIP2 V9.S4, V8.S4, V25.S4
|
||||
VZIP1 V11.S4, V10.S4, V26.S4
|
||||
VZIP2 V11.S4, V10.S4, V27.S4
|
||||
VZIP1 V13.S4, V12.S4, V28.S4
|
||||
VZIP2 V13.S4, V12.S4, V29.S4
|
||||
VZIP1 V15.S4, V14.S4, V30.S4
|
||||
VZIP2 V15.S4, V14.S4, V31.S4
|
||||
VZIP1 V18.D2, V16.D2, V0.D2
|
||||
VZIP2 V18.D2, V16.D2, V4.D2
|
||||
VZIP1 V19.D2, V17.D2, V8.D2
|
||||
VZIP2 V19.D2, V17.D2, V12.D2
|
||||
VLD1.P 64(R2), [V16.B16, V17.B16, V18.B16, V19.B16]
|
||||
|
||||
VZIP1 V22.D2, V20.D2, V1.D2
|
||||
VZIP2 V22.D2, V20.D2, V5.D2
|
||||
VZIP1 V23.D2, V21.D2, V9.D2
|
||||
VZIP2 V23.D2, V21.D2, V13.D2
|
||||
VLD1.P 64(R2), [V20.B16, V21.B16, V22.B16, V23.B16]
|
||||
VZIP1 V26.D2, V24.D2, V2.D2
|
||||
VZIP2 V26.D2, V24.D2, V6.D2
|
||||
VZIP1 V27.D2, V25.D2, V10.D2
|
||||
VZIP2 V27.D2, V25.D2, V14.D2
|
||||
VLD1.P 64(R2), [V24.B16, V25.B16, V26.B16, V27.B16]
|
||||
VZIP1 V30.D2, V28.D2, V3.D2
|
||||
VZIP2 V30.D2, V28.D2, V7.D2
|
||||
VZIP1 V31.D2, V29.D2, V11.D2
|
||||
VZIP2 V31.D2, V29.D2, V15.D2
|
||||
VLD1.P 64(R2), [V28.B16, V29.B16, V30.B16, V31.B16]
|
||||
VEOR V0.B16, V16.B16, V16.B16
|
||||
VEOR V1.B16, V17.B16, V17.B16
|
||||
VEOR V2.B16, V18.B16, V18.B16
|
||||
VEOR V3.B16, V19.B16, V19.B16
|
||||
VST1.P [V16.B16, V17.B16, V18.B16, V19.B16], 64(R1)
|
||||
VEOR V4.B16, V20.B16, V20.B16
|
||||
VEOR V5.B16, V21.B16, V21.B16
|
||||
VEOR V6.B16, V22.B16, V22.B16
|
||||
VEOR V7.B16, V23.B16, V23.B16
|
||||
VST1.P [V20.B16, V21.B16, V22.B16, V23.B16], 64(R1)
|
||||
VEOR V8.B16, V24.B16, V24.B16
|
||||
VEOR V9.B16, V25.B16, V25.B16
|
||||
VEOR V10.B16, V26.B16, V26.B16
|
||||
VEOR V11.B16, V27.B16, V27.B16
|
||||
VST1.P [V24.B16, V25.B16, V26.B16, V27.B16], 64(R1)
|
||||
VEOR V12.B16, V28.B16, V28.B16
|
||||
VEOR V13.B16, V29.B16, V29.B16
|
||||
VEOR V14.B16, V30.B16, V30.B16
|
||||
VEOR V15.B16, V31.B16, V31.B16
|
||||
VST1.P [V28.B16, V29.B16, V30.B16, V31.B16], 64(R1)
|
||||
|
||||
ADD $4, R20
|
||||
MOVW R20, (R7) // update counter
|
||||
|
||||
CMP R2, R12
|
||||
BGT loop
|
||||
|
||||
RET
|
||||
|
||||
|
||||
DATA ·constants+0x00(SB)/4, $0x61707865
|
||||
DATA ·constants+0x04(SB)/4, $0x3320646e
|
||||
DATA ·constants+0x08(SB)/4, $0x79622d32
|
||||
DATA ·constants+0x0c(SB)/4, $0x6b206574
|
||||
GLOBL ·constants(SB), NOPTR|RODATA, $32
|
||||
|
||||
DATA ·incRotMatrix+0x00(SB)/4, $0x00000000
|
||||
DATA ·incRotMatrix+0x04(SB)/4, $0x00000001
|
||||
DATA ·incRotMatrix+0x08(SB)/4, $0x00000002
|
||||
DATA ·incRotMatrix+0x0c(SB)/4, $0x00000003
|
||||
DATA ·incRotMatrix+0x10(SB)/4, $0x02010003
|
||||
DATA ·incRotMatrix+0x14(SB)/4, $0x06050407
|
||||
DATA ·incRotMatrix+0x18(SB)/4, $0x0A09080B
|
||||
DATA ·incRotMatrix+0x1c(SB)/4, $0x0E0D0C0F
|
||||
GLOBL ·incRotMatrix(SB), NOPTR|RODATA, $32
|
||||
|
|
@ -0,0 +1,364 @@
|
|||
// Copyright 2016 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// Package chacha20 implements the ChaCha20 and XChaCha20 encryption algorithms
|
||||
// as specified in RFC 8439 and draft-irtf-cfrg-xchacha-01.
|
||||
package chacha20
|
||||
|
||||
import (
|
||||
"crypto/cipher"
|
||||
"encoding/binary"
|
||||
"errors"
|
||||
"math/bits"
|
||||
|
||||
"golang.org/x/crypto/internal/subtle"
|
||||
)
|
||||
|
||||
const (
|
||||
// KeySize is the size of the key used by this cipher, in bytes.
|
||||
KeySize = 32
|
||||
|
||||
// NonceSize is the size of the nonce used with the standard variant of this
|
||||
// cipher, in bytes.
|
||||
//
|
||||
// Note that this is too short to be safely generated at random if the same
|
||||
// key is reused more than 2³² times.
|
||||
NonceSize = 12
|
||||
|
||||
// NonceSizeX is the size of the nonce used with the XChaCha20 variant of
|
||||
// this cipher, in bytes.
|
||||
NonceSizeX = 24
|
||||
)
|
||||
|
||||
// Cipher is a stateful instance of ChaCha20 or XChaCha20 using a particular key
|
||||
// and nonce. A *Cipher implements the cipher.Stream interface.
|
||||
type Cipher struct {
|
||||
// The ChaCha20 state is 16 words: 4 constant, 8 of key, 1 of counter
|
||||
// (incremented after each block), and 3 of nonce.
|
||||
key [8]uint32
|
||||
counter uint32
|
||||
nonce [3]uint32
|
||||
|
||||
// The last len bytes of buf are leftover key stream bytes from the previous
|
||||
// XORKeyStream invocation. The size of buf depends on how many blocks are
|
||||
// computed at a time.
|
||||
buf [bufSize]byte
|
||||
len int
|
||||
|
||||
// The counter-independent results of the first round are cached after they
|
||||
// are computed the first time.
|
||||
precompDone bool
|
||||
p1, p5, p9, p13 uint32
|
||||
p2, p6, p10, p14 uint32
|
||||
p3, p7, p11, p15 uint32
|
||||
}
|
||||
|
||||
var _ cipher.Stream = (*Cipher)(nil)
|
||||
|
||||
// NewUnauthenticatedCipher creates a new ChaCha20 stream cipher with the given
|
||||
// 32 bytes key and a 12 or 24 bytes nonce. If a nonce of 24 bytes is provided,
|
||||
// the XChaCha20 construction will be used. It returns an error if key or nonce
|
||||
// have any other length.
|
||||
//
|
||||
// Note that ChaCha20, like all stream ciphers, is not authenticated and allows
|
||||
// attackers to silently tamper with the plaintext. For this reason, it is more
|
||||
// appropriate as a building block than as a standalone encryption mechanism.
|
||||
// Instead, consider using package golang.org/x/crypto/chacha20poly1305.
|
||||
func NewUnauthenticatedCipher(key, nonce []byte) (*Cipher, error) {
|
||||
// This function is split into a wrapper so that the Cipher allocation will
|
||||
// be inlined, and depending on how the caller uses the return value, won't
|
||||
// escape to the heap.
|
||||
c := &Cipher{}
|
||||
return newUnauthenticatedCipher(c, key, nonce)
|
||||
}
|
||||
|
||||
func newUnauthenticatedCipher(c *Cipher, key, nonce []byte) (*Cipher, error) {
|
||||
if len(key) != KeySize {
|
||||
return nil, errors.New("chacha20: wrong key size")
|
||||
}
|
||||
if len(nonce) == NonceSizeX {
|
||||
// XChaCha20 uses the ChaCha20 core to mix 16 bytes of the nonce into a
|
||||
// derived key, allowing it to operate on a nonce of 24 bytes. See
|
||||
// draft-irtf-cfrg-xchacha-01, Section 2.3.
|
||||
key, _ = HChaCha20(key, nonce[0:16])
|
||||
cNonce := make([]byte, NonceSize)
|
||||
copy(cNonce[4:12], nonce[16:24])
|
||||
nonce = cNonce
|
||||
} else if len(nonce) != NonceSize {
|
||||
return nil, errors.New("chacha20: wrong nonce size")
|
||||
}
|
||||
|
||||
c.key = [8]uint32{
|
||||
binary.LittleEndian.Uint32(key[0:4]),
|
||||
binary.LittleEndian.Uint32(key[4:8]),
|
||||
binary.LittleEndian.Uint32(key[8:12]),
|
||||
binary.LittleEndian.Uint32(key[12:16]),
|
||||
binary.LittleEndian.Uint32(key[16:20]),
|
||||
binary.LittleEndian.Uint32(key[20:24]),
|
||||
binary.LittleEndian.Uint32(key[24:28]),
|
||||
binary.LittleEndian.Uint32(key[28:32]),
|
||||
}
|
||||
c.nonce = [3]uint32{
|
||||
binary.LittleEndian.Uint32(nonce[0:4]),
|
||||
binary.LittleEndian.Uint32(nonce[4:8]),
|
||||
binary.LittleEndian.Uint32(nonce[8:12]),
|
||||
}
|
||||
return c, nil
|
||||
}
|
||||
|
||||
// The constant first 4 words of the ChaCha20 state.
|
||||
const (
|
||||
j0 uint32 = 0x61707865 // expa
|
||||
j1 uint32 = 0x3320646e // nd 3
|
||||
j2 uint32 = 0x79622d32 // 2-by
|
||||
j3 uint32 = 0x6b206574 // te k
|
||||
)
|
||||
|
||||
const blockSize = 64
|
||||
|
||||
// quarterRound is the core of ChaCha20. It shuffles the bits of 4 state words.
|
||||
// It's executed 4 times for each of the 20 ChaCha20 rounds, operating on all 16
|
||||
// words each round, in columnar or diagonal groups of 4 at a time.
|
||||
func quarterRound(a, b, c, d uint32) (uint32, uint32, uint32, uint32) {
|
||||
a += b
|
||||
d ^= a
|
||||
d = bits.RotateLeft32(d, 16)
|
||||
c += d
|
||||
b ^= c
|
||||
b = bits.RotateLeft32(b, 12)
|
||||
a += b
|
||||
d ^= a
|
||||
d = bits.RotateLeft32(d, 8)
|
||||
c += d
|
||||
b ^= c
|
||||
b = bits.RotateLeft32(b, 7)
|
||||
return a, b, c, d
|
||||
}
|
||||
|
||||
// XORKeyStream XORs each byte in the given slice with a byte from the
|
||||
// cipher's key stream. Dst and src must overlap entirely or not at all.
|
||||
//
|
||||
// If len(dst) < len(src), XORKeyStream will panic. It is acceptable
|
||||
// to pass a dst bigger than src, and in that case, XORKeyStream will
|
||||
// only update dst[:len(src)] and will not touch the rest of dst.
|
||||
//
|
||||
// Multiple calls to XORKeyStream behave as if the concatenation of
|
||||
// the src buffers was passed in a single run. That is, Cipher
|
||||
// maintains state and does not reset at each XORKeyStream call.
|
||||
func (s *Cipher) XORKeyStream(dst, src []byte) {
|
||||
if len(src) == 0 {
|
||||
return
|
||||
}
|
||||
if len(dst) < len(src) {
|
||||
panic("chacha20: output smaller than input")
|
||||
}
|
||||
dst = dst[:len(src)]
|
||||
if subtle.InexactOverlap(dst, src) {
|
||||
panic("chacha20: invalid buffer overlap")
|
||||
}
|
||||
|
||||
// First, drain any remaining key stream from a previous XORKeyStream.
|
||||
if s.len != 0 {
|
||||
keyStream := s.buf[bufSize-s.len:]
|
||||
if len(src) < len(keyStream) {
|
||||
keyStream = keyStream[:len(src)]
|
||||
}
|
||||
_ = src[len(keyStream)-1] // bounds check elimination hint
|
||||
for i, b := range keyStream {
|
||||
dst[i] = src[i] ^ b
|
||||
}
|
||||
s.len -= len(keyStream)
|
||||
src = src[len(keyStream):]
|
||||
dst = dst[len(keyStream):]
|
||||
}
|
||||
|
||||
const blocksPerBuf = bufSize / blockSize
|
||||
numBufs := (uint64(len(src)) + bufSize - 1) / bufSize
|
||||
if uint64(s.counter)+numBufs*blocksPerBuf >= 1<<32 {
|
||||
panic("chacha20: counter overflow")
|
||||
}
|
||||
|
||||
// xorKeyStreamBlocks implementations expect input lengths that are a
|
||||
// multiple of bufSize. Platform-specific ones process multiple blocks at a
|
||||
// time, so have bufSizes that are a multiple of blockSize.
|
||||
|
||||
rem := len(src) % bufSize
|
||||
full := len(src) - rem
|
||||
|
||||
if full > 0 {
|
||||
s.xorKeyStreamBlocks(dst[:full], src[:full])
|
||||
}
|
||||
|
||||
// If we have a partial (multi-)block, pad it for xorKeyStreamBlocks, and
|
||||
// keep the leftover keystream for the next XORKeyStream invocation.
|
||||
if rem > 0 {
|
||||
s.buf = [bufSize]byte{}
|
||||
copy(s.buf[:], src[full:])
|
||||
s.xorKeyStreamBlocks(s.buf[:], s.buf[:])
|
||||
s.len = bufSize - copy(dst[full:], s.buf[:])
|
||||
}
|
||||
}
|
||||
|
||||
func (s *Cipher) xorKeyStreamBlocksGeneric(dst, src []byte) {
|
||||
if len(dst) != len(src) || len(dst)%blockSize != 0 {
|
||||
panic("chacha20: internal error: wrong dst and/or src length")
|
||||
}
|
||||
|
||||
// To generate each block of key stream, the initial cipher state
|
||||
// (represented below) is passed through 20 rounds of shuffling,
|
||||
// alternatively applying quarterRounds by columns (like 1, 5, 9, 13)
|
||||
// or by diagonals (like 1, 6, 11, 12).
|
||||
//
|
||||
// 0:cccccccc 1:cccccccc 2:cccccccc 3:cccccccc
|
||||
// 4:kkkkkkkk 5:kkkkkkkk 6:kkkkkkkk 7:kkkkkkkk
|
||||
// 8:kkkkkkkk 9:kkkkkkkk 10:kkkkkkkk 11:kkkkkkkk
|
||||
// 12:bbbbbbbb 13:nnnnnnnn 14:nnnnnnnn 15:nnnnnnnn
|
||||
//
|
||||
// c=constant k=key b=blockcount n=nonce
|
||||
var (
|
||||
c0, c1, c2, c3 = j0, j1, j2, j3
|
||||
c4, c5, c6, c7 = s.key[0], s.key[1], s.key[2], s.key[3]
|
||||
c8, c9, c10, c11 = s.key[4], s.key[5], s.key[6], s.key[7]
|
||||
_, c13, c14, c15 = s.counter, s.nonce[0], s.nonce[1], s.nonce[2]
|
||||
)
|
||||
|
||||
// Three quarters of the first round don't depend on the counter, so we can
|
||||
// calculate them here, and reuse them for multiple blocks in the loop, and
|
||||
// for future XORKeyStream invocations.
|
||||
if !s.precompDone {
|
||||
s.p1, s.p5, s.p9, s.p13 = quarterRound(c1, c5, c9, c13)
|
||||
s.p2, s.p6, s.p10, s.p14 = quarterRound(c2, c6, c10, c14)
|
||||
s.p3, s.p7, s.p11, s.p15 = quarterRound(c3, c7, c11, c15)
|
||||
s.precompDone = true
|
||||
}
|
||||
|
||||
for i := 0; i < len(src); i += blockSize {
|
||||
// The remainder of the first column round.
|
||||
fcr0, fcr4, fcr8, fcr12 := quarterRound(c0, c4, c8, s.counter)
|
||||
|
||||
// The second diagonal round.
|
||||
x0, x5, x10, x15 := quarterRound(fcr0, s.p5, s.p10, s.p15)
|
||||
x1, x6, x11, x12 := quarterRound(s.p1, s.p6, s.p11, fcr12)
|
||||
x2, x7, x8, x13 := quarterRound(s.p2, s.p7, fcr8, s.p13)
|
||||
x3, x4, x9, x14 := quarterRound(s.p3, fcr4, s.p9, s.p14)
|
||||
|
||||
// The remaining 18 rounds.
|
||||
for i := 0; i < 9; i++ {
|
||||
// Column round.
|
||||
x0, x4, x8, x12 = quarterRound(x0, x4, x8, x12)
|
||||
x1, x5, x9, x13 = quarterRound(x1, x5, x9, x13)
|
||||
x2, x6, x10, x14 = quarterRound(x2, x6, x10, x14)
|
||||
x3, x7, x11, x15 = quarterRound(x3, x7, x11, x15)
|
||||
|
||||
// Diagonal round.
|
||||
x0, x5, x10, x15 = quarterRound(x0, x5, x10, x15)
|
||||
x1, x6, x11, x12 = quarterRound(x1, x6, x11, x12)
|
||||
x2, x7, x8, x13 = quarterRound(x2, x7, x8, x13)
|
||||
x3, x4, x9, x14 = quarterRound(x3, x4, x9, x14)
|
||||
}
|
||||
|
||||
// Finally, add back the initial state to generate the key stream.
|
||||
x0 += c0
|
||||
x1 += c1
|
||||
x2 += c2
|
||||
x3 += c3
|
||||
x4 += c4
|
||||
x5 += c5
|
||||
x6 += c6
|
||||
x7 += c7
|
||||
x8 += c8
|
||||
x9 += c9
|
||||
x10 += c10
|
||||
x11 += c11
|
||||
x12 += s.counter
|
||||
x13 += c13
|
||||
x14 += c14
|
||||
x15 += c15
|
||||
|
||||
s.counter += 1
|
||||
if s.counter == 0 {
|
||||
panic("chacha20: internal error: counter overflow")
|
||||
}
|
||||
|
||||
in, out := src[i:], dst[i:]
|
||||
in, out = in[:blockSize], out[:blockSize] // bounds check elimination hint
|
||||
|
||||
// XOR the key stream with the source and write out the result.
|
||||
xor(out[0:], in[0:], x0)
|
||||
xor(out[4:], in[4:], x1)
|
||||
xor(out[8:], in[8:], x2)
|
||||
xor(out[12:], in[12:], x3)
|
||||
xor(out[16:], in[16:], x4)
|
||||
xor(out[20:], in[20:], x5)
|
||||
xor(out[24:], in[24:], x6)
|
||||
xor(out[28:], in[28:], x7)
|
||||
xor(out[32:], in[32:], x8)
|
||||
xor(out[36:], in[36:], x9)
|
||||
xor(out[40:], in[40:], x10)
|
||||
xor(out[44:], in[44:], x11)
|
||||
xor(out[48:], in[48:], x12)
|
||||
xor(out[52:], in[52:], x13)
|
||||
xor(out[56:], in[56:], x14)
|
||||
xor(out[60:], in[60:], x15)
|
||||
}
|
||||
}
|
||||
|
||||
// HChaCha20 uses the ChaCha20 core to generate a derived key from a 32 bytes
|
||||
// key and a 16 bytes nonce. It returns an error if key or nonce have any other
|
||||
// length. It is used as part of the XChaCha20 construction.
|
||||
func HChaCha20(key, nonce []byte) ([]byte, error) {
|
||||
// This function is split into a wrapper so that the slice allocation will
|
||||
// be inlined, and depending on how the caller uses the return value, won't
|
||||
// escape to the heap.
|
||||
out := make([]byte, 32)
|
||||
return hChaCha20(out, key, nonce)
|
||||
}
|
||||
|
||||
func hChaCha20(out, key, nonce []byte) ([]byte, error) {
|
||||
if len(key) != KeySize {
|
||||
return nil, errors.New("chacha20: wrong HChaCha20 key size")
|
||||
}
|
||||
if len(nonce) != 16 {
|
||||
return nil, errors.New("chacha20: wrong HChaCha20 nonce size")
|
||||
}
|
||||
|
||||
x0, x1, x2, x3 := j0, j1, j2, j3
|
||||
x4 := binary.LittleEndian.Uint32(key[0:4])
|
||||
x5 := binary.LittleEndian.Uint32(key[4:8])
|
||||
x6 := binary.LittleEndian.Uint32(key[8:12])
|
||||
x7 := binary.LittleEndian.Uint32(key[12:16])
|
||||
x8 := binary.LittleEndian.Uint32(key[16:20])
|
||||
x9 := binary.LittleEndian.Uint32(key[20:24])
|
||||
x10 := binary.LittleEndian.Uint32(key[24:28])
|
||||
x11 := binary.LittleEndian.Uint32(key[28:32])
|
||||
x12 := binary.LittleEndian.Uint32(nonce[0:4])
|
||||
x13 := binary.LittleEndian.Uint32(nonce[4:8])
|
||||
x14 := binary.LittleEndian.Uint32(nonce[8:12])
|
||||
x15 := binary.LittleEndian.Uint32(nonce[12:16])
|
||||
|
||||
for i := 0; i < 10; i++ {
|
||||
// Diagonal round.
|
||||
x0, x4, x8, x12 = quarterRound(x0, x4, x8, x12)
|
||||
x1, x5, x9, x13 = quarterRound(x1, x5, x9, x13)
|
||||
x2, x6, x10, x14 = quarterRound(x2, x6, x10, x14)
|
||||
x3, x7, x11, x15 = quarterRound(x3, x7, x11, x15)
|
||||
|
||||
// Column round.
|
||||
x0, x5, x10, x15 = quarterRound(x0, x5, x10, x15)
|
||||
x1, x6, x11, x12 = quarterRound(x1, x6, x11, x12)
|
||||
x2, x7, x8, x13 = quarterRound(x2, x7, x8, x13)
|
||||
x3, x4, x9, x14 = quarterRound(x3, x4, x9, x14)
|
||||
}
|
||||
|
||||
_ = out[31] // bounds check elimination hint
|
||||
binary.LittleEndian.PutUint32(out[0:4], x0)
|
||||
binary.LittleEndian.PutUint32(out[4:8], x1)
|
||||
binary.LittleEndian.PutUint32(out[8:12], x2)
|
||||
binary.LittleEndian.PutUint32(out[12:16], x3)
|
||||
binary.LittleEndian.PutUint32(out[16:20], x12)
|
||||
binary.LittleEndian.PutUint32(out[20:24], x13)
|
||||
binary.LittleEndian.PutUint32(out[24:28], x14)
|
||||
binary.LittleEndian.PutUint32(out[28:32], x15)
|
||||
return out, nil
|
||||
}
|
||||
|
|
@ -0,0 +1,13 @@
|
|||
// Copyright 2018 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build !arm64,!s390x,!ppc64le arm64,!go1.11 gccgo appengine
|
||||
|
||||
package chacha20
|
||||
|
||||
const bufSize = blockSize
|
||||
|
||||
func (s *Cipher) xorKeyStreamBlocks(dst, src []byte) {
|
||||
s.xorKeyStreamBlocksGeneric(dst, src)
|
||||
}
|
||||
|
|
@ -0,0 +1,16 @@
|
|||
// Copyright 2019 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build !gccgo,!appengine
|
||||
|
||||
package chacha20
|
||||
|
||||
const bufSize = 256
|
||||
|
||||
//go:noescape
|
||||
func chaCha20_ctr32_vsx(out, inp *byte, len int, key *[8]uint32, counter *uint32)
|
||||
|
||||
func (c *Cipher) xorKeyStreamBlocks(dst, src []byte) {
|
||||
chaCha20_ctr32_vsx(&dst[0], &src[0], len(src), &c.key, &c.counter)
|
||||
}
|
||||
|
|
@ -0,0 +1,449 @@
|
|||
// Copyright 2019 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// Based on CRYPTOGAMS code with the following comment:
|
||||
// # ====================================================================
|
||||
// # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
|
||||
// # project. The module is, however, dual licensed under OpenSSL and
|
||||
// # CRYPTOGAMS licenses depending on where you obtain it. For further
|
||||
// # details see http://www.openssl.org/~appro/cryptogams/.
|
||||
// # ====================================================================
|
||||
|
||||
// Code for the perl script that generates the ppc64 assembler
|
||||
// can be found in the cryptogams repository at the link below. It is based on
|
||||
// the original from openssl.
|
||||
|
||||
// https://github.com/dot-asm/cryptogams/commit/a60f5b50ed908e91
|
||||
|
||||
// The differences in this and the original implementation are
|
||||
// due to the calling conventions and initialization of constants.
|
||||
|
||||
// +build !gccgo,!appengine
|
||||
|
||||
#include "textflag.h"
|
||||
|
||||
#define OUT R3
|
||||
#define INP R4
|
||||
#define LEN R5
|
||||
#define KEY R6
|
||||
#define CNT R7
|
||||
#define TMP R15
|
||||
|
||||
#define CONSTBASE R16
|
||||
#define BLOCKS R17
|
||||
|
||||
DATA consts<>+0x00(SB)/8, $0x3320646e61707865
|
||||
DATA consts<>+0x08(SB)/8, $0x6b20657479622d32
|
||||
DATA consts<>+0x10(SB)/8, $0x0000000000000001
|
||||
DATA consts<>+0x18(SB)/8, $0x0000000000000000
|
||||
DATA consts<>+0x20(SB)/8, $0x0000000000000004
|
||||
DATA consts<>+0x28(SB)/8, $0x0000000000000000
|
||||
DATA consts<>+0x30(SB)/8, $0x0a0b08090e0f0c0d
|
||||
DATA consts<>+0x38(SB)/8, $0x0203000106070405
|
||||
DATA consts<>+0x40(SB)/8, $0x090a0b080d0e0f0c
|
||||
DATA consts<>+0x48(SB)/8, $0x0102030005060704
|
||||
DATA consts<>+0x50(SB)/8, $0x6170786561707865
|
||||
DATA consts<>+0x58(SB)/8, $0x6170786561707865
|
||||
DATA consts<>+0x60(SB)/8, $0x3320646e3320646e
|
||||
DATA consts<>+0x68(SB)/8, $0x3320646e3320646e
|
||||
DATA consts<>+0x70(SB)/8, $0x79622d3279622d32
|
||||
DATA consts<>+0x78(SB)/8, $0x79622d3279622d32
|
||||
DATA consts<>+0x80(SB)/8, $0x6b2065746b206574
|
||||
DATA consts<>+0x88(SB)/8, $0x6b2065746b206574
|
||||
DATA consts<>+0x90(SB)/8, $0x0000000100000000
|
||||
DATA consts<>+0x98(SB)/8, $0x0000000300000002
|
||||
GLOBL consts<>(SB), RODATA, $0xa0
|
||||
|
||||
//func chaCha20_ctr32_vsx(out, inp *byte, len int, key *[8]uint32, counter *uint32)
|
||||
TEXT ·chaCha20_ctr32_vsx(SB),NOSPLIT,$64-40
|
||||
MOVD out+0(FP), OUT
|
||||
MOVD inp+8(FP), INP
|
||||
MOVD len+16(FP), LEN
|
||||
MOVD key+24(FP), KEY
|
||||
MOVD counter+32(FP), CNT
|
||||
|
||||
// Addressing for constants
|
||||
MOVD $consts<>+0x00(SB), CONSTBASE
|
||||
MOVD $16, R8
|
||||
MOVD $32, R9
|
||||
MOVD $48, R10
|
||||
MOVD $64, R11
|
||||
SRD $6, LEN, BLOCKS
|
||||
// V16
|
||||
LXVW4X (CONSTBASE)(R0), VS48
|
||||
ADD $80,CONSTBASE
|
||||
|
||||
// Load key into V17,V18
|
||||
LXVW4X (KEY)(R0), VS49
|
||||
LXVW4X (KEY)(R8), VS50
|
||||
|
||||
// Load CNT, NONCE into V19
|
||||
LXVW4X (CNT)(R0), VS51
|
||||
|
||||
// Clear V27
|
||||
VXOR V27, V27, V27
|
||||
|
||||
// V28
|
||||
LXVW4X (CONSTBASE)(R11), VS60
|
||||
|
||||
// splat slot from V19 -> V26
|
||||
VSPLTW $0, V19, V26
|
||||
|
||||
VSLDOI $4, V19, V27, V19
|
||||
VSLDOI $12, V27, V19, V19
|
||||
|
||||
VADDUWM V26, V28, V26
|
||||
|
||||
MOVD $10, R14
|
||||
MOVD R14, CTR
|
||||
|
||||
loop_outer_vsx:
|
||||
// V0, V1, V2, V3
|
||||
LXVW4X (R0)(CONSTBASE), VS32
|
||||
LXVW4X (R8)(CONSTBASE), VS33
|
||||
LXVW4X (R9)(CONSTBASE), VS34
|
||||
LXVW4X (R10)(CONSTBASE), VS35
|
||||
|
||||
// splat values from V17, V18 into V4-V11
|
||||
VSPLTW $0, V17, V4
|
||||
VSPLTW $1, V17, V5
|
||||
VSPLTW $2, V17, V6
|
||||
VSPLTW $3, V17, V7
|
||||
VSPLTW $0, V18, V8
|
||||
VSPLTW $1, V18, V9
|
||||
VSPLTW $2, V18, V10
|
||||
VSPLTW $3, V18, V11
|
||||
|
||||
// VOR
|
||||
VOR V26, V26, V12
|
||||
|
||||
// splat values from V19 -> V13, V14, V15
|
||||
VSPLTW $1, V19, V13
|
||||
VSPLTW $2, V19, V14
|
||||
VSPLTW $3, V19, V15
|
||||
|
||||
// splat const values
|
||||
VSPLTISW $-16, V27
|
||||
VSPLTISW $12, V28
|
||||
VSPLTISW $8, V29
|
||||
VSPLTISW $7, V30
|
||||
|
||||
loop_vsx:
|
||||
VADDUWM V0, V4, V0
|
||||
VADDUWM V1, V5, V1
|
||||
VADDUWM V2, V6, V2
|
||||
VADDUWM V3, V7, V3
|
||||
|
||||
VXOR V12, V0, V12
|
||||
VXOR V13, V1, V13
|
||||
VXOR V14, V2, V14
|
||||
VXOR V15, V3, V15
|
||||
|
||||
VRLW V12, V27, V12
|
||||
VRLW V13, V27, V13
|
||||
VRLW V14, V27, V14
|
||||
VRLW V15, V27, V15
|
||||
|
||||
VADDUWM V8, V12, V8
|
||||
VADDUWM V9, V13, V9
|
||||
VADDUWM V10, V14, V10
|
||||
VADDUWM V11, V15, V11
|
||||
|
||||
VXOR V4, V8, V4
|
||||
VXOR V5, V9, V5
|
||||
VXOR V6, V10, V6
|
||||
VXOR V7, V11, V7
|
||||
|
||||
VRLW V4, V28, V4
|
||||
VRLW V5, V28, V5
|
||||
VRLW V6, V28, V6
|
||||
VRLW V7, V28, V7
|
||||
|
||||
VADDUWM V0, V4, V0
|
||||
VADDUWM V1, V5, V1
|
||||
VADDUWM V2, V6, V2
|
||||
VADDUWM V3, V7, V3
|
||||
|
||||
VXOR V12, V0, V12
|
||||
VXOR V13, V1, V13
|
||||
VXOR V14, V2, V14
|
||||
VXOR V15, V3, V15
|
||||
|
||||
VRLW V12, V29, V12
|
||||
VRLW V13, V29, V13
|
||||
VRLW V14, V29, V14
|
||||
VRLW V15, V29, V15
|
||||
|
||||
VADDUWM V8, V12, V8
|
||||
VADDUWM V9, V13, V9
|
||||
VADDUWM V10, V14, V10
|
||||
VADDUWM V11, V15, V11
|
||||
|
||||
VXOR V4, V8, V4
|
||||
VXOR V5, V9, V5
|
||||
VXOR V6, V10, V6
|
||||
VXOR V7, V11, V7
|
||||
|
||||
VRLW V4, V30, V4
|
||||
VRLW V5, V30, V5
|
||||
VRLW V6, V30, V6
|
||||
VRLW V7, V30, V7
|
||||
|
||||
VADDUWM V0, V5, V0
|
||||
VADDUWM V1, V6, V1
|
||||
VADDUWM V2, V7, V2
|
||||
VADDUWM V3, V4, V3
|
||||
|
||||
VXOR V15, V0, V15
|
||||
VXOR V12, V1, V12
|
||||
VXOR V13, V2, V13
|
||||
VXOR V14, V3, V14
|
||||
|
||||
VRLW V15, V27, V15
|
||||
VRLW V12, V27, V12
|
||||
VRLW V13, V27, V13
|
||||
VRLW V14, V27, V14
|
||||
|
||||
VADDUWM V10, V15, V10
|
||||
VADDUWM V11, V12, V11
|
||||
VADDUWM V8, V13, V8
|
||||
VADDUWM V9, V14, V9
|
||||
|
||||
VXOR V5, V10, V5
|
||||
VXOR V6, V11, V6
|
||||
VXOR V7, V8, V7
|
||||
VXOR V4, V9, V4
|
||||
|
||||
VRLW V5, V28, V5
|
||||
VRLW V6, V28, V6
|
||||
VRLW V7, V28, V7
|
||||
VRLW V4, V28, V4
|
||||
|
||||
VADDUWM V0, V5, V0
|
||||
VADDUWM V1, V6, V1
|
||||
VADDUWM V2, V7, V2
|
||||
VADDUWM V3, V4, V3
|
||||
|
||||
VXOR V15, V0, V15
|
||||
VXOR V12, V1, V12
|
||||
VXOR V13, V2, V13
|
||||
VXOR V14, V3, V14
|
||||
|
||||
VRLW V15, V29, V15
|
||||
VRLW V12, V29, V12
|
||||
VRLW V13, V29, V13
|
||||
VRLW V14, V29, V14
|
||||
|
||||
VADDUWM V10, V15, V10
|
||||
VADDUWM V11, V12, V11
|
||||
VADDUWM V8, V13, V8
|
||||
VADDUWM V9, V14, V9
|
||||
|
||||
VXOR V5, V10, V5
|
||||
VXOR V6, V11, V6
|
||||
VXOR V7, V8, V7
|
||||
VXOR V4, V9, V4
|
||||
|
||||
VRLW V5, V30, V5
|
||||
VRLW V6, V30, V6
|
||||
VRLW V7, V30, V7
|
||||
VRLW V4, V30, V4
|
||||
BC 16, LT, loop_vsx
|
||||
|
||||
VADDUWM V12, V26, V12
|
||||
|
||||
WORD $0x13600F8C // VMRGEW V0, V1, V27
|
||||
WORD $0x13821F8C // VMRGEW V2, V3, V28
|
||||
|
||||
WORD $0x10000E8C // VMRGOW V0, V1, V0
|
||||
WORD $0x10421E8C // VMRGOW V2, V3, V2
|
||||
|
||||
WORD $0x13A42F8C // VMRGEW V4, V5, V29
|
||||
WORD $0x13C63F8C // VMRGEW V6, V7, V30
|
||||
|
||||
XXPERMDI VS32, VS34, $0, VS33
|
||||
XXPERMDI VS32, VS34, $3, VS35
|
||||
XXPERMDI VS59, VS60, $0, VS32
|
||||
XXPERMDI VS59, VS60, $3, VS34
|
||||
|
||||
WORD $0x10842E8C // VMRGOW V4, V5, V4
|
||||
WORD $0x10C63E8C // VMRGOW V6, V7, V6
|
||||
|
||||
WORD $0x13684F8C // VMRGEW V8, V9, V27
|
||||
WORD $0x138A5F8C // VMRGEW V10, V11, V28
|
||||
|
||||
XXPERMDI VS36, VS38, $0, VS37
|
||||
XXPERMDI VS36, VS38, $3, VS39
|
||||
XXPERMDI VS61, VS62, $0, VS36
|
||||
XXPERMDI VS61, VS62, $3, VS38
|
||||
|
||||
WORD $0x11084E8C // VMRGOW V8, V9, V8
|
||||
WORD $0x114A5E8C // VMRGOW V10, V11, V10
|
||||
|
||||
WORD $0x13AC6F8C // VMRGEW V12, V13, V29
|
||||
WORD $0x13CE7F8C // VMRGEW V14, V15, V30
|
||||
|
||||
XXPERMDI VS40, VS42, $0, VS41
|
||||
XXPERMDI VS40, VS42, $3, VS43
|
||||
XXPERMDI VS59, VS60, $0, VS40
|
||||
XXPERMDI VS59, VS60, $3, VS42
|
||||
|
||||
WORD $0x118C6E8C // VMRGOW V12, V13, V12
|
||||
WORD $0x11CE7E8C // VMRGOW V14, V15, V14
|
||||
|
||||
VSPLTISW $4, V27
|
||||
VADDUWM V26, V27, V26
|
||||
|
||||
XXPERMDI VS44, VS46, $0, VS45
|
||||
XXPERMDI VS44, VS46, $3, VS47
|
||||
XXPERMDI VS61, VS62, $0, VS44
|
||||
XXPERMDI VS61, VS62, $3, VS46
|
||||
|
||||
VADDUWM V0, V16, V0
|
||||
VADDUWM V4, V17, V4
|
||||
VADDUWM V8, V18, V8
|
||||
VADDUWM V12, V19, V12
|
||||
|
||||
CMPU LEN, $64
|
||||
BLT tail_vsx
|
||||
|
||||
// Bottom of loop
|
||||
LXVW4X (INP)(R0), VS59
|
||||
LXVW4X (INP)(R8), VS60
|
||||
LXVW4X (INP)(R9), VS61
|
||||
LXVW4X (INP)(R10), VS62
|
||||
|
||||
VXOR V27, V0, V27
|
||||
VXOR V28, V4, V28
|
||||
VXOR V29, V8, V29
|
||||
VXOR V30, V12, V30
|
||||
|
||||
STXVW4X VS59, (OUT)(R0)
|
||||
STXVW4X VS60, (OUT)(R8)
|
||||
ADD $64, INP
|
||||
STXVW4X VS61, (OUT)(R9)
|
||||
ADD $-64, LEN
|
||||
STXVW4X VS62, (OUT)(R10)
|
||||
ADD $64, OUT
|
||||
BEQ done_vsx
|
||||
|
||||
VADDUWM V1, V16, V0
|
||||
VADDUWM V5, V17, V4
|
||||
VADDUWM V9, V18, V8
|
||||
VADDUWM V13, V19, V12
|
||||
|
||||
CMPU LEN, $64
|
||||
BLT tail_vsx
|
||||
|
||||
LXVW4X (INP)(R0), VS59
|
||||
LXVW4X (INP)(R8), VS60
|
||||
LXVW4X (INP)(R9), VS61
|
||||
LXVW4X (INP)(R10), VS62
|
||||
VXOR V27, V0, V27
|
||||
|
||||
VXOR V28, V4, V28
|
||||
VXOR V29, V8, V29
|
||||
VXOR V30, V12, V30
|
||||
|
||||
STXVW4X VS59, (OUT)(R0)
|
||||
STXVW4X VS60, (OUT)(R8)
|
||||
ADD $64, INP
|
||||
STXVW4X VS61, (OUT)(R9)
|
||||
ADD $-64, LEN
|
||||
STXVW4X VS62, (OUT)(V10)
|
||||
ADD $64, OUT
|
||||
BEQ done_vsx
|
||||
|
||||
VADDUWM V2, V16, V0
|
||||
VADDUWM V6, V17, V4
|
||||
VADDUWM V10, V18, V8
|
||||
VADDUWM V14, V19, V12
|
||||
|
||||
CMPU LEN, $64
|
||||
BLT tail_vsx
|
||||
|
||||
LXVW4X (INP)(R0), VS59
|
||||
LXVW4X (INP)(R8), VS60
|
||||
LXVW4X (INP)(R9), VS61
|
||||
LXVW4X (INP)(R10), VS62
|
||||
|
||||
VXOR V27, V0, V27
|
||||
VXOR V28, V4, V28
|
||||
VXOR V29, V8, V29
|
||||
VXOR V30, V12, V30
|
||||
|
||||
STXVW4X VS59, (OUT)(R0)
|
||||
STXVW4X VS60, (OUT)(R8)
|
||||
ADD $64, INP
|
||||
STXVW4X VS61, (OUT)(R9)
|
||||
ADD $-64, LEN
|
||||
STXVW4X VS62, (OUT)(R10)
|
||||
ADD $64, OUT
|
||||
BEQ done_vsx
|
||||
|
||||
VADDUWM V3, V16, V0
|
||||
VADDUWM V7, V17, V4
|
||||
VADDUWM V11, V18, V8
|
||||
VADDUWM V15, V19, V12
|
||||
|
||||
CMPU LEN, $64
|
||||
BLT tail_vsx
|
||||
|
||||
LXVW4X (INP)(R0), VS59
|
||||
LXVW4X (INP)(R8), VS60
|
||||
LXVW4X (INP)(R9), VS61
|
||||
LXVW4X (INP)(R10), VS62
|
||||
|
||||
VXOR V27, V0, V27
|
||||
VXOR V28, V4, V28
|
||||
VXOR V29, V8, V29
|
||||
VXOR V30, V12, V30
|
||||
|
||||
STXVW4X VS59, (OUT)(R0)
|
||||
STXVW4X VS60, (OUT)(R8)
|
||||
ADD $64, INP
|
||||
STXVW4X VS61, (OUT)(R9)
|
||||
ADD $-64, LEN
|
||||
STXVW4X VS62, (OUT)(R10)
|
||||
ADD $64, OUT
|
||||
|
||||
MOVD $10, R14
|
||||
MOVD R14, CTR
|
||||
BNE loop_outer_vsx
|
||||
|
||||
done_vsx:
|
||||
// Increment counter by number of 64 byte blocks
|
||||
MOVD (CNT), R14
|
||||
ADD BLOCKS, R14
|
||||
MOVD R14, (CNT)
|
||||
RET
|
||||
|
||||
tail_vsx:
|
||||
ADD $32, R1, R11
|
||||
MOVD LEN, CTR
|
||||
|
||||
// Save values on stack to copy from
|
||||
STXVW4X VS32, (R11)(R0)
|
||||
STXVW4X VS36, (R11)(R8)
|
||||
STXVW4X VS40, (R11)(R9)
|
||||
STXVW4X VS44, (R11)(R10)
|
||||
ADD $-1, R11, R12
|
||||
ADD $-1, INP
|
||||
ADD $-1, OUT
|
||||
|
||||
looptail_vsx:
|
||||
// Copying the result to OUT
|
||||
// in bytes.
|
||||
MOVBZU 1(R12), KEY
|
||||
MOVBZU 1(INP), TMP
|
||||
XOR KEY, TMP, KEY
|
||||
MOVBU KEY, 1(OUT)
|
||||
BC 16, LT, looptail_vsx
|
||||
|
||||
// Clear the stack values
|
||||
STXVW4X VS48, (R11)(R0)
|
||||
STXVW4X VS48, (R11)(R8)
|
||||
STXVW4X VS48, (R11)(R9)
|
||||
STXVW4X VS48, (R11)(R10)
|
||||
BR done_vsx
|
||||
|
|
@ -0,0 +1,26 @@
|
|||
// Copyright 2018 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build !gccgo,!appengine
|
||||
|
||||
package chacha20
|
||||
|
||||
import "golang.org/x/sys/cpu"
|
||||
|
||||
var haveAsm = cpu.S390X.HasVX
|
||||
|
||||
const bufSize = 256
|
||||
|
||||
// xorKeyStreamVX is an assembly implementation of XORKeyStream. It must only
|
||||
// be called when the vector facility is available. Implementation in asm_s390x.s.
|
||||
//go:noescape
|
||||
func xorKeyStreamVX(dst, src []byte, key *[8]uint32, nonce *[3]uint32, counter *uint32)
|
||||
|
||||
func (c *Cipher) xorKeyStreamBlocks(dst, src []byte) {
|
||||
if cpu.S390X.HasVX {
|
||||
xorKeyStreamVX(dst, src, &c.key, &c.nonce, &c.counter)
|
||||
} else {
|
||||
c.xorKeyStreamBlocksGeneric(dst, src)
|
||||
}
|
||||
}
|
||||
|
|
@ -0,0 +1,224 @@
|
|||
// Copyright 2018 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build !gccgo,!appengine
|
||||
|
||||
#include "go_asm.h"
|
||||
#include "textflag.h"
|
||||
|
||||
// This is an implementation of the ChaCha20 encryption algorithm as
|
||||
// specified in RFC 7539. It uses vector instructions to compute
|
||||
// 4 keystream blocks in parallel (256 bytes) which are then XORed
|
||||
// with the bytes in the input slice.
|
||||
|
||||
GLOBL ·constants<>(SB), RODATA|NOPTR, $32
|
||||
// BSWAP: swap bytes in each 4-byte element
|
||||
DATA ·constants<>+0x00(SB)/4, $0x03020100
|
||||
DATA ·constants<>+0x04(SB)/4, $0x07060504
|
||||
DATA ·constants<>+0x08(SB)/4, $0x0b0a0908
|
||||
DATA ·constants<>+0x0c(SB)/4, $0x0f0e0d0c
|
||||
// J0: [j0, j1, j2, j3]
|
||||
DATA ·constants<>+0x10(SB)/4, $0x61707865
|
||||
DATA ·constants<>+0x14(SB)/4, $0x3320646e
|
||||
DATA ·constants<>+0x18(SB)/4, $0x79622d32
|
||||
DATA ·constants<>+0x1c(SB)/4, $0x6b206574
|
||||
|
||||
#define BSWAP V5
|
||||
#define J0 V6
|
||||
#define KEY0 V7
|
||||
#define KEY1 V8
|
||||
#define NONCE V9
|
||||
#define CTR V10
|
||||
#define M0 V11
|
||||
#define M1 V12
|
||||
#define M2 V13
|
||||
#define M3 V14
|
||||
#define INC V15
|
||||
#define X0 V16
|
||||
#define X1 V17
|
||||
#define X2 V18
|
||||
#define X3 V19
|
||||
#define X4 V20
|
||||
#define X5 V21
|
||||
#define X6 V22
|
||||
#define X7 V23
|
||||
#define X8 V24
|
||||
#define X9 V25
|
||||
#define X10 V26
|
||||
#define X11 V27
|
||||
#define X12 V28
|
||||
#define X13 V29
|
||||
#define X14 V30
|
||||
#define X15 V31
|
||||
|
||||
#define NUM_ROUNDS 20
|
||||
|
||||
#define ROUND4(a0, a1, a2, a3, b0, b1, b2, b3, c0, c1, c2, c3, d0, d1, d2, d3) \
|
||||
VAF a1, a0, a0 \
|
||||
VAF b1, b0, b0 \
|
||||
VAF c1, c0, c0 \
|
||||
VAF d1, d0, d0 \
|
||||
VX a0, a2, a2 \
|
||||
VX b0, b2, b2 \
|
||||
VX c0, c2, c2 \
|
||||
VX d0, d2, d2 \
|
||||
VERLLF $16, a2, a2 \
|
||||
VERLLF $16, b2, b2 \
|
||||
VERLLF $16, c2, c2 \
|
||||
VERLLF $16, d2, d2 \
|
||||
VAF a2, a3, a3 \
|
||||
VAF b2, b3, b3 \
|
||||
VAF c2, c3, c3 \
|
||||
VAF d2, d3, d3 \
|
||||
VX a3, a1, a1 \
|
||||
VX b3, b1, b1 \
|
||||
VX c3, c1, c1 \
|
||||
VX d3, d1, d1 \
|
||||
VERLLF $12, a1, a1 \
|
||||
VERLLF $12, b1, b1 \
|
||||
VERLLF $12, c1, c1 \
|
||||
VERLLF $12, d1, d1 \
|
||||
VAF a1, a0, a0 \
|
||||
VAF b1, b0, b0 \
|
||||
VAF c1, c0, c0 \
|
||||
VAF d1, d0, d0 \
|
||||
VX a0, a2, a2 \
|
||||
VX b0, b2, b2 \
|
||||
VX c0, c2, c2 \
|
||||
VX d0, d2, d2 \
|
||||
VERLLF $8, a2, a2 \
|
||||
VERLLF $8, b2, b2 \
|
||||
VERLLF $8, c2, c2 \
|
||||
VERLLF $8, d2, d2 \
|
||||
VAF a2, a3, a3 \
|
||||
VAF b2, b3, b3 \
|
||||
VAF c2, c3, c3 \
|
||||
VAF d2, d3, d3 \
|
||||
VX a3, a1, a1 \
|
||||
VX b3, b1, b1 \
|
||||
VX c3, c1, c1 \
|
||||
VX d3, d1, d1 \
|
||||
VERLLF $7, a1, a1 \
|
||||
VERLLF $7, b1, b1 \
|
||||
VERLLF $7, c1, c1 \
|
||||
VERLLF $7, d1, d1
|
||||
|
||||
#define PERMUTE(mask, v0, v1, v2, v3) \
|
||||
VPERM v0, v0, mask, v0 \
|
||||
VPERM v1, v1, mask, v1 \
|
||||
VPERM v2, v2, mask, v2 \
|
||||
VPERM v3, v3, mask, v3
|
||||
|
||||
#define ADDV(x, v0, v1, v2, v3) \
|
||||
VAF x, v0, v0 \
|
||||
VAF x, v1, v1 \
|
||||
VAF x, v2, v2 \
|
||||
VAF x, v3, v3
|
||||
|
||||
#define XORV(off, dst, src, v0, v1, v2, v3) \
|
||||
VLM off(src), M0, M3 \
|
||||
PERMUTE(BSWAP, v0, v1, v2, v3) \
|
||||
VX v0, M0, M0 \
|
||||
VX v1, M1, M1 \
|
||||
VX v2, M2, M2 \
|
||||
VX v3, M3, M3 \
|
||||
VSTM M0, M3, off(dst)
|
||||
|
||||
#define SHUFFLE(a, b, c, d, t, u, v, w) \
|
||||
VMRHF a, c, t \ // t = {a[0], c[0], a[1], c[1]}
|
||||
VMRHF b, d, u \ // u = {b[0], d[0], b[1], d[1]}
|
||||
VMRLF a, c, v \ // v = {a[2], c[2], a[3], c[3]}
|
||||
VMRLF b, d, w \ // w = {b[2], d[2], b[3], d[3]}
|
||||
VMRHF t, u, a \ // a = {a[0], b[0], c[0], d[0]}
|
||||
VMRLF t, u, b \ // b = {a[1], b[1], c[1], d[1]}
|
||||
VMRHF v, w, c \ // c = {a[2], b[2], c[2], d[2]}
|
||||
VMRLF v, w, d // d = {a[3], b[3], c[3], d[3]}
|
||||
|
||||
// func xorKeyStreamVX(dst, src []byte, key *[8]uint32, nonce *[3]uint32, counter *uint32)
|
||||
TEXT ·xorKeyStreamVX(SB), NOSPLIT, $0
|
||||
MOVD $·constants<>(SB), R1
|
||||
MOVD dst+0(FP), R2 // R2=&dst[0]
|
||||
LMG src+24(FP), R3, R4 // R3=&src[0] R4=len(src)
|
||||
MOVD key+48(FP), R5 // R5=key
|
||||
MOVD nonce+56(FP), R6 // R6=nonce
|
||||
MOVD counter+64(FP), R7 // R7=counter
|
||||
|
||||
// load BSWAP and J0
|
||||
VLM (R1), BSWAP, J0
|
||||
|
||||
// setup
|
||||
MOVD $95, R0
|
||||
VLM (R5), KEY0, KEY1
|
||||
VLL R0, (R6), NONCE
|
||||
VZERO M0
|
||||
VLEIB $7, $32, M0
|
||||
VSRLB M0, NONCE, NONCE
|
||||
|
||||
// initialize counter values
|
||||
VLREPF (R7), CTR
|
||||
VZERO INC
|
||||
VLEIF $1, $1, INC
|
||||
VLEIF $2, $2, INC
|
||||
VLEIF $3, $3, INC
|
||||
VAF INC, CTR, CTR
|
||||
VREPIF $4, INC
|
||||
|
||||
chacha:
|
||||
VREPF $0, J0, X0
|
||||
VREPF $1, J0, X1
|
||||
VREPF $2, J0, X2
|
||||
VREPF $3, J0, X3
|
||||
VREPF $0, KEY0, X4
|
||||
VREPF $1, KEY0, X5
|
||||
VREPF $2, KEY0, X6
|
||||
VREPF $3, KEY0, X7
|
||||
VREPF $0, KEY1, X8
|
||||
VREPF $1, KEY1, X9
|
||||
VREPF $2, KEY1, X10
|
||||
VREPF $3, KEY1, X11
|
||||
VLR CTR, X12
|
||||
VREPF $1, NONCE, X13
|
||||
VREPF $2, NONCE, X14
|
||||
VREPF $3, NONCE, X15
|
||||
|
||||
MOVD $(NUM_ROUNDS/2), R1
|
||||
|
||||
loop:
|
||||
ROUND4(X0, X4, X12, X8, X1, X5, X13, X9, X2, X6, X14, X10, X3, X7, X15, X11)
|
||||
ROUND4(X0, X5, X15, X10, X1, X6, X12, X11, X2, X7, X13, X8, X3, X4, X14, X9)
|
||||
|
||||
ADD $-1, R1
|
||||
BNE loop
|
||||
|
||||
// decrement length
|
||||
ADD $-256, R4
|
||||
|
||||
// rearrange vectors
|
||||
SHUFFLE(X0, X1, X2, X3, M0, M1, M2, M3)
|
||||
ADDV(J0, X0, X1, X2, X3)
|
||||
SHUFFLE(X4, X5, X6, X7, M0, M1, M2, M3)
|
||||
ADDV(KEY0, X4, X5, X6, X7)
|
||||
SHUFFLE(X8, X9, X10, X11, M0, M1, M2, M3)
|
||||
ADDV(KEY1, X8, X9, X10, X11)
|
||||
VAF CTR, X12, X12
|
||||
SHUFFLE(X12, X13, X14, X15, M0, M1, M2, M3)
|
||||
ADDV(NONCE, X12, X13, X14, X15)
|
||||
|
||||
// increment counters
|
||||
VAF INC, CTR, CTR
|
||||
|
||||
// xor keystream with plaintext
|
||||
XORV(0*64, R2, R3, X0, X4, X8, X12)
|
||||
XORV(1*64, R2, R3, X1, X5, X9, X13)
|
||||
XORV(2*64, R2, R3, X2, X6, X10, X14)
|
||||
XORV(3*64, R2, R3, X3, X7, X11, X15)
|
||||
|
||||
// increment pointers
|
||||
MOVD $256(R2), R2
|
||||
MOVD $256(R3), R3
|
||||
|
||||
CMPBNE R4, $0, chacha
|
||||
|
||||
VSTEF $0, CTR, (R7)
|
||||
RET
|
||||
|
|
@ -0,0 +1,41 @@
|
|||
// Copyright 2018 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found src the LICENSE file.
|
||||
|
||||
package chacha20
|
||||
|
||||
import "runtime"
|
||||
|
||||
// Platforms that have fast unaligned 32-bit little endian accesses.
|
||||
const unaligned = runtime.GOARCH == "386" ||
|
||||
runtime.GOARCH == "amd64" ||
|
||||
runtime.GOARCH == "arm64" ||
|
||||
runtime.GOARCH == "ppc64le" ||
|
||||
runtime.GOARCH == "s390x"
|
||||
|
||||
// xor reads a little endian uint32 from src, XORs it with u and
|
||||
// places the result in little endian byte order in dst.
|
||||
func xor(dst, src []byte, u uint32) {
|
||||
_, _ = src[3], dst[3] // eliminate bounds checks
|
||||
if unaligned {
|
||||
// The compiler should optimize this code into
|
||||
// 32-bit unaligned little endian loads and stores.
|
||||
// TODO: delete once the compiler does a reliably
|
||||
// good job with the generic code below.
|
||||
// See issue #25111 for more details.
|
||||
v := uint32(src[0])
|
||||
v |= uint32(src[1]) << 8
|
||||
v |= uint32(src[2]) << 16
|
||||
v |= uint32(src[3]) << 24
|
||||
v ^= u
|
||||
dst[0] = byte(v)
|
||||
dst[1] = byte(v >> 8)
|
||||
dst[2] = byte(v >> 16)
|
||||
dst[3] = byte(v >> 24)
|
||||
} else {
|
||||
dst[0] = src[0] ^ byte(u)
|
||||
dst[1] = src[1] ^ byte(u>>8)
|
||||
dst[2] = src[2] ^ byte(u>>16)
|
||||
dst[3] = src[3] ^ byte(u>>24)
|
||||
}
|
||||
}
|
||||
|
|
@ -0,0 +1,39 @@
|
|||
// Copyright 2019 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build !go1.13
|
||||
|
||||
package poly1305
|
||||
|
||||
// Generic fallbacks for the math/bits intrinsics, copied from
|
||||
// src/math/bits/bits.go. They were added in Go 1.12, but Add64 and Sum64 had
|
||||
// variable time fallbacks until Go 1.13.
|
||||
|
||||
func bitsAdd64(x, y, carry uint64) (sum, carryOut uint64) {
|
||||
sum = x + y + carry
|
||||
carryOut = ((x & y) | ((x | y) &^ sum)) >> 63
|
||||
return
|
||||
}
|
||||
|
||||
func bitsSub64(x, y, borrow uint64) (diff, borrowOut uint64) {
|
||||
diff = x - y - borrow
|
||||
borrowOut = ((^x & y) | (^(x ^ y) & diff)) >> 63
|
||||
return
|
||||
}
|
||||
|
||||
func bitsMul64(x, y uint64) (hi, lo uint64) {
|
||||
const mask32 = 1<<32 - 1
|
||||
x0 := x & mask32
|
||||
x1 := x >> 32
|
||||
y0 := y & mask32
|
||||
y1 := y >> 32
|
||||
w0 := x0 * y0
|
||||
t := x1*y0 + w0>>32
|
||||
w1 := t & mask32
|
||||
w2 := t >> 32
|
||||
w1 += x0 * y1
|
||||
hi = x1*y1 + w2 + w1>>32
|
||||
lo = x * y
|
||||
return
|
||||
}
|
||||
|
|
@ -0,0 +1,21 @@
|
|||
// Copyright 2019 The Go Authors. All rights reserved.
|
||||
// Use of this source code is governed by a BSD-style
|
||||
// license that can be found in the LICENSE file.
|
||||
|
||||
// +build go1.13
|
||||
|
||||
package poly1305
|
||||
|
||||
import "math/bits"
|
||||
|
||||
func bitsAdd64(x, y, carry uint64) (sum, carryOut uint64) {
|
||||
return bits.Add64(x, y, carry)
|
||||
}
|
||||
|
||||
func bitsSub64(x, y, borrow uint64) (diff, borrowOut uint64) {
|
||||
return bits.Sub64(x, y, borrow)
|
||||
}
|
||||
|
||||
func bitsMul64(x, y uint64) (hi, lo uint64) {
|
||||
return bits.Mul64(x, y)
|
||||
}
|
||||
Loading…
Reference in New Issue