MyFavMirrors/minio
1
0
Fork 0
mirror of https://github.com/minio/minio.git synced 2025-11-10 14:09:48 -05:00
Code Issues Packages Projects Releases Wiki Activity

add SSE-C support for HEAD, GET, PUT (#4894)

Browse Source 
This change adds server-side-encryption support for HEAD, GET and PUT
operations. This PR only addresses single-part PUTs and GETs without
HTTP ranges.

Further this change adds the concept of reserved object metadata which is required
to make encrypted objects tamper-proof and provide API compatibility to AWS S3.
This PR adds the following reserved metadata entries:
- X-Minio-Internal-Server-Side-Encryption-Iv          ('guarantees' tamper-proof property)
- X-Minio-Internal-Server-Side-Encryption-Kdf         (makes Key-MAC computation negotiable in future)
- X-Minio-Internal-Server-Side-Encryption-Key-Mac     (provides AWS S3 API compatibility)

The prefix `X-Minio_Internal` specifies an internal metadata entry which must not
send to clients. All client requests containing a metadata key starting with `X-Minio-Internal`
must also rejected. This is implemented by a generic-handler.

This PR implements SSE-C separated from client-side-encryption (CSE). This cannot decrypt
server-side-encrypted objects on the client-side. However, clients can encrypted the same object
with CSE and SSE-C.

This PR does not address:
 - SSE-C Copy and Copy part
 - SSE-C GET with HTTP ranges
 - SSE-C multipart PUT
 - SSE-C Gateway

Each point must be addressed in a separate PR.

Added to vendor dir:
 - x/crypto/chacha20poly1305
 - x/crypto/poly1305
 - github.com/minio/sio
This commit is contained in:
Andreas Auernhammer
2017-11-08 00:18:59 +01:00
committed by Dee Koder
parent 7e7ae29d89
commit ca6b4773ed
35 changed files with 6321 additions and 197 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
vendor/golang.org/x/crypto/LICENSE generated vendored Normal file
View File

@@ -0,0 +1,27 @@
Copyright (c) 2009 The Go Authors. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following disclaimer
in the documentation and/or other materials provided with the
distribution.
* Neither the name of Google Inc. nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

22
vendor/golang.org/x/crypto/PATENTS generated vendored Normal file
View File

@@ -0,0 +1,22 @@
Additional IP Rights Grant (Patents)
"This implementation" means the copyrightable works distributed by
Google as part of the Go project.
Google hereby grants to You a perpetual, worldwide, non-exclusive,
no-charge, royalty-free, irrevocable (except as stated in this section)
patent license to make, have made, use, offer to sell, sell, import,
transfer and otherwise run, modify and propagate the contents of this
implementation of Go, where such license applies only to those patent
claims, both currently owned or controlled by Google and acquired in
the future, licensable by Google that are necessarily infringed by this
implementation of Go. This grant does not include claims that would be
infringed only as a consequence of further modification of this
implementation. If you or your agent or exclusive licensee institute or
order or agree to the institution of patent litigation against any
entity (including a cross-claim or counterclaim in a lawsuit) alleging
that this implementation of Go or any code incorporated within this
implementation of Go constitutes direct or contributory patent
infringement, or inducement of patent infringement, then any patent
rights granted to you under this License for this implementation of Go
shall terminate as of the date such litigation is filed.

83
vendor/golang.org/x/crypto/chacha20poly1305/chacha20poly1305.go generated vendored Normal file
View File

@@ -0,0 +1,83 @@
// 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 chacha20poly1305 implements the ChaCha20-Poly1305 AEAD as specified in RFC 7539.
package chacha20poly1305 // import "golang.org/x/crypto/chacha20poly1305"
import (
"crypto/cipher"
"errors"
)
const (
// KeySize is the size of the key used by this AEAD, in bytes.
KeySize = 32
// NonceSize is the size of the nonce used with this AEAD, in bytes.
NonceSize = 12
)
type chacha20poly1305 struct {
key [32]byte
}
// New returns a ChaCha20-Poly1305 AEAD that uses the given, 256-bit key.
func New(key []byte) (cipher.AEAD, error) {
if len(key) != KeySize {
return nil, errors.New("chacha20poly1305: bad key length")
}
ret := new(chacha20poly1305)
copy(ret.key[:], key)
return ret, nil
}
func (c *chacha20poly1305) NonceSize() int {
return NonceSize
}
func (c *chacha20poly1305) Overhead() int {
return 16
}
func (c *chacha20poly1305) Seal(dst, nonce, plaintext, additionalData []byte) []byte {
if len(nonce) != NonceSize {
panic("chacha20poly1305: bad nonce length passed to Seal")
}
if uint64(len(plaintext)) > (1<<38)-64 {
panic("chacha20poly1305: plaintext too large")
}
return c.seal(dst, nonce, plaintext, additionalData)
}
var errOpen = errors.New("chacha20poly1305: message authentication failed")
func (c *chacha20poly1305) Open(dst, nonce, ciphertext, additionalData []byte) ([]byte, error) {
if len(nonce) != NonceSize {
panic("chacha20poly1305: bad nonce length passed to Open")
}
if len(ciphertext) < 16 {
return nil, errOpen
}
if uint64(len(ciphertext)) > (1<<38)-48 {
panic("chacha20poly1305: ciphertext too large")
}
return c.open(dst, nonce, ciphertext, additionalData)
}
// sliceForAppend takes a slice and a requested number of bytes. It returns a
// slice with the contents of the given slice followed by that many bytes and a
// second slice that aliases into it and contains only the extra bytes. If the
// original slice has sufficient capacity then no allocation is performed.
func sliceForAppend(in []byte, n int) (head, tail []byte) {
if total := len(in) + n; cap(in) >= total {
head = in[:total]
} else {
head = make([]byte, total)
copy(head, in)
}
tail = head[len(in):]
return
}

127
vendor/golang.org/x/crypto/chacha20poly1305/chacha20poly1305_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,127 @@
// 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.
// +build go1.7,amd64,!gccgo,!appengine
package chacha20poly1305
import "encoding/binary"
//go:noescape
func chacha20Poly1305Open(dst []byte, key []uint32, src, ad []byte) bool
//go:noescape
func chacha20Poly1305Seal(dst []byte, key []uint32, src, ad []byte)
// cpuid is implemented in chacha20poly1305_amd64.s.
func cpuid(eaxArg, ecxArg uint32) (eax, ebx, ecx, edx uint32)
// xgetbv with ecx = 0 is implemented in chacha20poly1305_amd64.s.
func xgetbv() (eax, edx uint32)
var (
useASM bool
useAVX2 bool
)
func init() {
detectCPUFeatures()
}
// detectCPUFeatures is used to detect if cpu instructions
// used by the functions implemented in assembler in
// chacha20poly1305_amd64.s are supported.
func detectCPUFeatures() {
maxID, _, _, _ := cpuid(0, 0)
if maxID < 1 {
return
}
_, _, ecx1, _ := cpuid(1, 0)
haveSSSE3 := isSet(9, ecx1)
useASM = haveSSSE3
haveOSXSAVE := isSet(27, ecx1)
osSupportsAVX := false
// For XGETBV, OSXSAVE bit is required and sufficient.
if haveOSXSAVE {
eax, _ := xgetbv()
// Check if XMM and YMM registers have OS support.
osSupportsAVX = isSet(1, eax) && isSet(2, eax)
}
haveAVX := isSet(28, ecx1) && osSupportsAVX
if maxID < 7 {
return
}
_, ebx7, _, _ := cpuid(7, 0)
haveAVX2 := isSet(5, ebx7) && haveAVX
haveBMI2 := isSet(8, ebx7)
useAVX2 = haveAVX2 && haveBMI2
}
// isSet checks if bit at bitpos is set in value.
func isSet(bitpos uint, value uint32) bool {
return value&(1<<bitpos) != 0
}
// setupState writes a ChaCha20 input matrix to state. See
// https://tools.ietf.org/html/rfc7539#section-2.3.
func setupState(state *[16]uint32, key *[32]byte, nonce []byte) {
state[0] = 0x61707865
state[1] = 0x3320646e
state[2] = 0x79622d32
state[3] = 0x6b206574
state[4] = binary.LittleEndian.Uint32(key[:4])
state[5] = binary.LittleEndian.Uint32(key[4:8])
state[6] = binary.LittleEndian.Uint32(key[8:12])
state[7] = binary.LittleEndian.Uint32(key[12:16])
state[8] = binary.LittleEndian.Uint32(key[16:20])
state[9] = binary.LittleEndian.Uint32(key[20:24])
state[10] = binary.LittleEndian.Uint32(key[24:28])
state[11] = binary.LittleEndian.Uint32(key[28:32])
state[12] = 0
state[13] = binary.LittleEndian.Uint32(nonce[:4])
state[14] = binary.LittleEndian.Uint32(nonce[4:8])
state[15] = binary.LittleEndian.Uint32(nonce[8:12])
}
func (c *chacha20poly1305) seal(dst, nonce, plaintext, additionalData []byte) []byte {
if !useASM {
return c.sealGeneric(dst, nonce, plaintext, additionalData)
}
var state [16]uint32
setupState(&state, &c.key, nonce)
ret, out := sliceForAppend(dst, len(plaintext)+16)
chacha20Poly1305Seal(out[:], state[:], plaintext, additionalData)
return ret
}
func (c *chacha20poly1305) open(dst, nonce, ciphertext, additionalData []byte) ([]byte, error) {
if !useASM {
return c.openGeneric(dst, nonce, ciphertext, additionalData)
}
var state [16]uint32
setupState(&state, &c.key, nonce)
ciphertext = ciphertext[:len(ciphertext)-16]
ret, out := sliceForAppend(dst, len(ciphertext))
if !chacha20Poly1305Open(out, state[:], ciphertext, additionalData) {
for i := range out {
out[i] = 0
}
return nil, errOpen
}
return ret, nil
}

2714
vendor/golang.org/x/crypto/chacha20poly1305/chacha20poly1305_amd64.s generated vendored Normal file
View File

File diff suppressed because it is too large Load Diff

70
vendor/golang.org/x/crypto/chacha20poly1305/chacha20poly1305_generic.go generated vendored Normal file
View File

@@ -0,0 +1,70 @@
// 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 chacha20poly1305
import (
"encoding/binary"
"golang.org/x/crypto/chacha20poly1305/internal/chacha20"
"golang.org/x/crypto/poly1305"
)
func roundTo16(n int) int {
return 16 * ((n + 15) / 16)
}
func (c *chacha20poly1305) sealGeneric(dst, nonce, plaintext, additionalData []byte) []byte {
var counter [16]byte
copy(counter[4:], nonce)
var polyKey [32]byte
chacha20.XORKeyStream(polyKey[:], polyKey[:], &counter, &c.key)
ret, out := sliceForAppend(dst, len(plaintext)+poly1305.TagSize)
counter[0] = 1
chacha20.XORKeyStream(out, plaintext, &counter, &c.key)
polyInput := make([]byte, roundTo16(len(additionalData))+roundTo16(len(plaintext))+8+8)
copy(polyInput, additionalData)
copy(polyInput[roundTo16(len(additionalData)):], out[:len(plaintext)])
binary.LittleEndian.PutUint64(polyInput[len(polyInput)-16:], uint64(len(additionalData)))
binary.LittleEndian.PutUint64(polyInput[len(polyInput)-8:], uint64(len(plaintext)))
var tag [poly1305.TagSize]byte
poly1305.Sum(&tag, polyInput, &polyKey)
copy(out[len(plaintext):], tag[:])
return ret
}
func (c *chacha20poly1305) openGeneric(dst, nonce, ciphertext, additionalData []byte) ([]byte, error) {
var tag [poly1305.TagSize]byte
copy(tag[:], ciphertext[len(ciphertext)-16:])
ciphertext = ciphertext[:len(ciphertext)-16]
var counter [16]byte
copy(counter[4:], nonce)
var polyKey [32]byte
chacha20.XORKeyStream(polyKey[:], polyKey[:], &counter, &c.key)
polyInput := make([]byte, roundTo16(len(additionalData))+roundTo16(len(ciphertext))+8+8)
copy(polyInput, additionalData)
copy(polyInput[roundTo16(len(additionalData)):], ciphertext)
binary.LittleEndian.PutUint64(polyInput[len(polyInput)-16:], uint64(len(additionalData)))
binary.LittleEndian.PutUint64(polyInput[len(polyInput)-8:], uint64(len(ciphertext)))
ret, out := sliceForAppend(dst, len(ciphertext))
if !poly1305.Verify(&tag, polyInput, &polyKey) {
for i := range out {
out[i] = 0
}
return nil, errOpen
}
counter[0] = 1
chacha20.XORKeyStream(out, ciphertext, &counter, &c.key)
return ret, nil
}

15
vendor/golang.org/x/crypto/chacha20poly1305/chacha20poly1305_noasm.go generated vendored Normal file
View File

@@ -0,0 +1,15 @@
// 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.
// +build !amd64 !go1.7 gccgo appengine
package chacha20poly1305
func (c *chacha20poly1305) seal(dst, nonce, plaintext, additionalData []byte) []byte {
return c.sealGeneric(dst, nonce, plaintext, additionalData)
}
func (c *chacha20poly1305) open(dst, nonce, ciphertext, additionalData []byte) ([]byte, error) {
return c.openGeneric(dst, nonce, ciphertext, additionalData)
}

199
vendor/golang.org/x/crypto/chacha20poly1305/internal/chacha20/chacha_generic.go generated vendored Normal file
View File

@@ -0,0 +1,199 @@
// 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 core ChaCha20 function as specified in https://tools.ietf.org/html/rfc7539#section-2.3.
package chacha20
import "encoding/binary"
const rounds = 20
// core applies the ChaCha20 core function to 16-byte input in, 32-byte key k,
// and 16-byte constant c, and puts the result into 64-byte array out.
func core(out *[64]byte, in *[16]byte, k *[32]byte) {
j0 := uint32(0x61707865)
j1 := uint32(0x3320646e)
j2 := uint32(0x79622d32)
j3 := uint32(0x6b206574)
j4 := binary.LittleEndian.Uint32(k[0:4])
j5 := binary.LittleEndian.Uint32(k[4:8])
j6 := binary.LittleEndian.Uint32(k[8:12])
j7 := binary.LittleEndian.Uint32(k[12:16])
j8 := binary.LittleEndian.Uint32(k[16:20])
j9 := binary.LittleEndian.Uint32(k[20:24])
j10 := binary.LittleEndian.Uint32(k[24:28])
j11 := binary.LittleEndian.Uint32(k[28:32])
j12 := binary.LittleEndian.Uint32(in[0:4])
j13 := binary.LittleEndian.Uint32(in[4:8])
j14 := binary.LittleEndian.Uint32(in[8:12])
j15 := binary.LittleEndian.Uint32(in[12:16])
x0, x1, x2, x3, x4, x5, x6, x7 := j0, j1, j2, j3, j4, j5, j6, j7
x8, x9, x10, x11, x12, x13, x14, x15 := j8, j9, j10, j11, j12, j13, j14, j15
for i := 0; i < rounds; i += 2 {
x0 += x4
x12 ^= x0
x12 = (x12 << 16) | (x12 >> (16))
x8 += x12
x4 ^= x8
x4 = (x4 << 12) | (x4 >> (20))
x0 += x4
x12 ^= x0
x12 = (x12 << 8) | (x12 >> (24))
x8 += x12
x4 ^= x8
x4 = (x4 << 7) | (x4 >> (25))
x1 += x5
x13 ^= x1
x13 = (x13 << 16) | (x13 >> 16)
x9 += x13
x5 ^= x9
x5 = (x5 << 12) | (x5 >> 20)
x1 += x5
x13 ^= x1
x13 = (x13 << 8) | (x13 >> 24)
x9 += x13
x5 ^= x9
x5 = (x5 << 7) | (x5 >> 25)
x2 += x6
x14 ^= x2
x14 = (x14 << 16) | (x14 >> 16)
x10 += x14
x6 ^= x10
x6 = (x6 << 12) | (x6 >> 20)
x2 += x6
x14 ^= x2
x14 = (x14 << 8) | (x14 >> 24)
x10 += x14
x6 ^= x10
x6 = (x6 << 7) | (x6 >> 25)
x3 += x7
x15 ^= x3
x15 = (x15 << 16) | (x15 >> 16)
x11 += x15
x7 ^= x11
x7 = (x7 << 12) | (x7 >> 20)
x3 += x7
x15 ^= x3
x15 = (x15 << 8) | (x15 >> 24)
x11 += x15
x7 ^= x11
x7 = (x7 << 7) | (x7 >> 25)
x0 += x5
x15 ^= x0
x15 = (x15 << 16) | (x15 >> 16)
x10 += x15
x5 ^= x10
x5 = (x5 << 12) | (x5 >> 20)
x0 += x5
x15 ^= x0
x15 = (x15 << 8) | (x15 >> 24)
x10 += x15
x5 ^= x10
x5 = (x5 << 7) | (x5 >> 25)
x1 += x6
x12 ^= x1
x12 = (x12 << 16) | (x12 >> 16)
x11 += x12
x6 ^= x11
x6 = (x6 << 12) | (x6 >> 20)
x1 += x6
x12 ^= x1
x12 = (x12 << 8) | (x12 >> 24)
x11 += x12
x6 ^= x11
x6 = (x6 << 7) | (x6 >> 25)
x2 += x7
x13 ^= x2
x13 = (x13 << 16) | (x13 >> 16)
x8 += x13
x7 ^= x8
x7 = (x7 << 12) | (x7 >> 20)
x2 += x7
x13 ^= x2
x13 = (x13 << 8) | (x13 >> 24)
x8 += x13
x7 ^= x8
x7 = (x7 << 7) | (x7 >> 25)
x3 += x4
x14 ^= x3
x14 = (x14 << 16) | (x14 >> 16)
x9 += x14
x4 ^= x9
x4 = (x4 << 12) | (x4 >> 20)
x3 += x4
x14 ^= x3
x14 = (x14 << 8) | (x14 >> 24)
x9 += x14
x4 ^= x9
x4 = (x4 << 7) | (x4 >> 25)
}
x0 += j0
x1 += j1
x2 += j2
x3 += j3
x4 += j4
x5 += j5
x6 += j6
x7 += j7
x8 += j8
x9 += j9
x10 += j10
x11 += j11
x12 += j12
x13 += j13
x14 += j14
x15 += j15
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], x4)
binary.LittleEndian.PutUint32(out[20:24], x5)
binary.LittleEndian.PutUint32(out[24:28], x6)
binary.LittleEndian.PutUint32(out[28:32], x7)
binary.LittleEndian.PutUint32(out[32:36], x8)
binary.LittleEndian.PutUint32(out[36:40], x9)
binary.LittleEndian.PutUint32(out[40:44], x10)
binary.LittleEndian.PutUint32(out[44:48], x11)
binary.LittleEndian.PutUint32(out[48:52], x12)
binary.LittleEndian.PutUint32(out[52:56], x13)
binary.LittleEndian.PutUint32(out[56:60], x14)
binary.LittleEndian.PutUint32(out[60:64], x15)
}
// XORKeyStream crypts bytes from in to out using the given key and counters.
// In and out may be the same slice but otherwise should not overlap. Counter
// contains the raw ChaCha20 counter bytes (i.e. block counter followed by
// nonce).
func XORKeyStream(out, in []byte, counter *[16]byte, key *[32]byte) {
var block [64]byte
var counterCopy [16]byte
copy(counterCopy[:], counter[:])
for len(in) >= 64 {
core(&block, &counterCopy, key)
for i, x := range block {
out[i] = in[i] ^ x
}
u := uint32(1)
for i := 0; i < 4; i++ {
u += uint32(counterCopy[i])
counterCopy[i] = byte(u)
u >>= 8
}
in = in[64:]
out = out[64:]
}
if len(in) > 0 {
core(&block, &counterCopy, key)
for i, v := range in {
out[i] = v ^ block[i]
}
}
}

33
vendor/golang.org/x/crypto/poly1305/poly1305.go generated vendored Normal file
View File

@@ -0,0 +1,33 @@
// Copyright 2012 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 poly1305 implements Poly1305 one-time message authentication code as
specified in https://cr.yp.to/mac/poly1305-20050329.pdf.
Poly1305 is a fast, one-time authentication function. It is infeasible for an
attacker to generate an authenticator for a message without the key. However, a
key must only be used for a single message. Authenticating two different
messages with the same key allows an attacker to forge authenticators for other
messages with the same key.
Poly1305 was originally coupled with AES in order to make Poly1305-AES. AES was
used with a fixed key in order to generate one-time keys from an nonce.
However, in this package AES isn't used and the one-time key is specified
directly.
*/
package poly1305 // import "golang.org/x/crypto/poly1305"
import "crypto/subtle"
// TagSize is the size, in bytes, of a poly1305 authenticator.
const TagSize = 16
// Verify returns true if mac is a valid authenticator for m with the given
// key.
func Verify(mac *[16]byte, m []byte, key *[32]byte) bool {
var tmp [16]byte
Sum(&tmp, m, key)
return subtle.ConstantTimeCompare(tmp[:], mac[:]) == 1
}

22
vendor/golang.org/x/crypto/poly1305/sum_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,22 @@
// Copyright 2012 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 amd64,!gccgo,!appengine
package poly1305
// This function is implemented in sum_amd64.s
//go:noescape
func poly1305(out *[16]byte, m *byte, mlen uint64, key *[32]byte)
// Sum generates an authenticator for m using a one-time key and puts the
// 16-byte result into out. Authenticating two different messages with the same
// key allows an attacker to forge messages at will.
func Sum(out *[16]byte, m []byte, key *[32]byte) {
var mPtr *byte
if len(m) > 0 {
mPtr = &m[0]
}
poly1305(out, mPtr, uint64(len(m)), key)
}

125
vendor/golang.org/x/crypto/poly1305/sum_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,125 @@
// Copyright 2012 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 amd64,!gccgo,!appengine
#include "textflag.h"
#define POLY1305_ADD(msg, h0, h1, h2) \
ADDQ 0(msg), h0; \
ADCQ 8(msg), h1; \
ADCQ $1, h2; \
LEAQ 16(msg), msg
#define POLY1305_MUL(h0, h1, h2, r0, r1, t0, t1, t2, t3) \
MOVQ r0, AX; \
MULQ h0; \
MOVQ AX, t0; \
MOVQ DX, t1; \
MOVQ r0, AX; \
MULQ h1; \
ADDQ AX, t1; \
ADCQ $0, DX; \
MOVQ r0, t2; \
IMULQ h2, t2; \
ADDQ DX, t2; \
\
MOVQ r1, AX; \
MULQ h0; \
ADDQ AX, t1; \
ADCQ $0, DX; \
MOVQ DX, h0; \
MOVQ r1, t3; \
IMULQ h2, t3; \
MOVQ r1, AX; \
MULQ h1; \
ADDQ AX, t2; \
ADCQ DX, t3; \
ADDQ h0, t2; \
ADCQ $0, t3; \
\
MOVQ t0, h0; \
MOVQ t1, h1; \
MOVQ t2, h2; \
ANDQ $3, h2; \
MOVQ t2, t0; \
ANDQ $0xFFFFFFFFFFFFFFFC, t0; \
ADDQ t0, h0; \
ADCQ t3, h1; \
ADCQ $0, h2; \
SHRQ $2, t3, t2; \
SHRQ $2, t3; \
ADDQ t2, h0; \
ADCQ t3, h1; \
ADCQ $0, h2
DATA ·poly1305Mask<>+0x00(SB)/8, $0x0FFFFFFC0FFFFFFF
DATA ·poly1305Mask<>+0x08(SB)/8, $0x0FFFFFFC0FFFFFFC
GLOBL ·poly1305Mask<>(SB), RODATA, $16
// func poly1305(out *[16]byte, m *byte, mlen uint64, key *[32]key)
TEXT ·poly1305(SB), $0-32
MOVQ out+0(FP), DI
MOVQ m+8(FP), SI
MOVQ mlen+16(FP), R15
MOVQ key+24(FP), AX
MOVQ 0(AX), R11
MOVQ 8(AX), R12
ANDQ ·poly1305Mask<>(SB), R11 // r0
ANDQ ·poly1305Mask<>+8(SB), R12 // r1
XORQ R8, R8 // h0
XORQ R9, R9 // h1
XORQ R10, R10 // h2
CMPQ R15, $16
JB bytes_between_0_and_15
loop:
POLY1305_ADD(SI, R8, R9, R10)
multiply:
POLY1305_MUL(R8, R9, R10, R11, R12, BX, CX, R13, R14)
SUBQ $16, R15
CMPQ R15, $16
JAE loop
bytes_between_0_and_15:
TESTQ R15, R15
JZ done
MOVQ $1, BX
XORQ CX, CX
XORQ R13, R13
ADDQ R15, SI
flush_buffer:
SHLQ $8, BX, CX
SHLQ $8, BX
MOVB -1(SI), R13
XORQ R13, BX
DECQ SI
DECQ R15
JNZ flush_buffer
ADDQ BX, R8
ADCQ CX, R9
ADCQ $0, R10
MOVQ $16, R15
JMP multiply
done:
MOVQ R8, AX
MOVQ R9, BX
SUBQ $0xFFFFFFFFFFFFFFFB, AX
SBBQ $0xFFFFFFFFFFFFFFFF, BX
SBBQ $3, R10
CMOVQCS R8, AX
CMOVQCS R9, BX
MOVQ key+24(FP), R8
ADDQ 16(R8), AX
ADCQ 24(R8), BX
MOVQ AX, 0(DI)
MOVQ BX, 8(DI)
RET

22
vendor/golang.org/x/crypto/poly1305/sum_arm.go generated vendored Normal file
View File

@@ -0,0 +1,22 @@
// Copyright 2015 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 arm,!gccgo,!appengine,!nacl
package poly1305
// This function is implemented in sum_arm.s
//go:noescape
func poly1305_auth_armv6(out *[16]byte, m *byte, mlen uint32, key *[32]byte)
// Sum generates an authenticator for m using a one-time key and puts the
// 16-byte result into out. Authenticating two different messages with the same
// key allows an attacker to forge messages at will.
func Sum(out *[16]byte, m []byte, key *[32]byte) {
var mPtr *byte
if len(m) > 0 {
mPtr = &m[0]
}
poly1305_auth_armv6(out, mPtr, uint32(len(m)), key)
}

427
vendor/golang.org/x/crypto/poly1305/sum_arm.s generated vendored Normal file
View File

@@ -0,0 +1,427 @@
// Copyright 2015 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 arm,!gccgo,!appengine,!nacl
#include "textflag.h"
// This code was translated into a form compatible with 5a from the public
// domain source by Andrew Moon: github.com/floodyberry/poly1305-opt/blob/master/app/extensions/poly1305.
DATA ·poly1305_init_constants_armv6<>+0x00(SB)/4, $0x3ffffff
DATA ·poly1305_init_constants_armv6<>+0x04(SB)/4, $0x3ffff03
DATA ·poly1305_init_constants_armv6<>+0x08(SB)/4, $0x3ffc0ff
DATA ·poly1305_init_constants_armv6<>+0x0c(SB)/4, $0x3f03fff
DATA ·poly1305_init_constants_armv6<>+0x10(SB)/4, $0x00fffff
GLOBL ·poly1305_init_constants_armv6<>(SB), 8, $20
// Warning: the linker may use R11 to synthesize certain instructions. Please
// take care and verify that no synthetic instructions use it.
TEXT poly1305_init_ext_armv6<>(SB), NOSPLIT, $0
// Needs 16 bytes of stack and 64 bytes of space pointed to by R0. (It
// might look like it's only 60 bytes of space but the final four bytes
// will be written by another function.) We need to skip over four
// bytes of stack because that's saving the value of 'g'.
ADD $4, R13, R8
MOVM.IB [R4-R7], (R8)
MOVM.IA.W (R1), [R2-R5]
MOVW $·poly1305_init_constants_armv6<>(SB), R7
MOVW R2, R8
MOVW R2>>26, R9
MOVW R3>>20, g
MOVW R4>>14, R11
MOVW R5>>8, R12
ORR R3<<6, R9, R9
ORR R4<<12, g, g
ORR R5<<18, R11, R11
MOVM.IA (R7), [R2-R6]
AND R8, R2, R2
AND R9, R3, R3
AND g, R4, R4
AND R11, R5, R5
AND R12, R6, R6
MOVM.IA.W [R2-R6], (R0)
EOR R2, R2, R2
EOR R3, R3, R3
EOR R4, R4, R4
EOR R5, R5, R5
EOR R6, R6, R6
MOVM.IA.W [R2-R6], (R0)
MOVM.IA.W (R1), [R2-R5]
MOVM.IA [R2-R6], (R0)
ADD $20, R13, R0
MOVM.DA (R0), [R4-R7]
RET
#define MOVW_UNALIGNED(Rsrc, Rdst, Rtmp, offset) \
MOVBU (offset+0)(Rsrc), Rtmp; \
MOVBU Rtmp, (offset+0)(Rdst); \
MOVBU (offset+1)(Rsrc), Rtmp; \
MOVBU Rtmp, (offset+1)(Rdst); \
MOVBU (offset+2)(Rsrc), Rtmp; \
MOVBU Rtmp, (offset+2)(Rdst); \
MOVBU (offset+3)(Rsrc), Rtmp; \
MOVBU Rtmp, (offset+3)(Rdst)
TEXT poly1305_blocks_armv6<>(SB), NOSPLIT, $0
// Needs 24 bytes of stack for saved registers and then 88 bytes of
// scratch space after that. We assume that 24 bytes at (R13) have
// already been used: four bytes for the link register saved in the
// prelude of poly1305_auth_armv6, four bytes for saving the value of g
// in that function and 16 bytes of scratch space used around
// poly1305_finish_ext_armv6_skip1.
ADD $24, R13, R12
MOVM.IB [R4-R8, R14], (R12)
MOVW R0, 88(R13)
MOVW R1, 92(R13)
MOVW R2, 96(R13)
MOVW R1, R14
MOVW R2, R12
MOVW 56(R0), R8
WORD $0xe1180008 // TST R8, R8 not working see issue 5921
EOR R6, R6, R6
MOVW.EQ $(1<<24), R6
MOVW R6, 84(R13)
ADD $116, R13, g
MOVM.IA (R0), [R0-R9]
MOVM.IA [R0-R4], (g)
CMP $16, R12
BLO poly1305_blocks_armv6_done
poly1305_blocks_armv6_mainloop:
WORD $0xe31e0003 // TST R14, #3 not working see issue 5921
BEQ poly1305_blocks_armv6_mainloop_aligned
ADD $100, R13, g
MOVW_UNALIGNED(R14, g, R0, 0)
MOVW_UNALIGNED(R14, g, R0, 4)
MOVW_UNALIGNED(R14, g, R0, 8)
MOVW_UNALIGNED(R14, g, R0, 12)
MOVM.IA (g), [R0-R3]
ADD $16, R14
B poly1305_blocks_armv6_mainloop_loaded
poly1305_blocks_armv6_mainloop_aligned:
MOVM.IA.W (R14), [R0-R3]
poly1305_blocks_armv6_mainloop_loaded:
MOVW R0>>26, g
MOVW R1>>20, R11
MOVW R2>>14, R12
MOVW R14, 92(R13)
MOVW R3>>8, R4
ORR R1<<6, g, g
ORR R2<<12, R11, R11
ORR R3<<18, R12, R12
BIC $0xfc000000, R0, R0
BIC $0xfc000000, g, g
MOVW 84(R13), R3
BIC $0xfc000000, R11, R11
BIC $0xfc000000, R12, R12
ADD R0, R5, R5
ADD g, R6, R6
ORR R3, R4, R4
ADD R11, R7, R7
ADD $116, R13, R14
ADD R12, R8, R8
ADD R4, R9, R9
MOVM.IA (R14), [R0-R4]
MULLU R4, R5, (R11, g)
MULLU R3, R5, (R14, R12)
MULALU R3, R6, (R11, g)
MULALU R2, R6, (R14, R12)
MULALU R2, R7, (R11, g)
MULALU R1, R7, (R14, R12)
ADD R4<<2, R4, R4
ADD R3<<2, R3, R3
MULALU R1, R8, (R11, g)
MULALU R0, R8, (R14, R12)
MULALU R0, R9, (R11, g)
MULALU R4, R9, (R14, R12)
MOVW g, 76(R13)
MOVW R11, 80(R13)
MOVW R12, 68(R13)
MOVW R14, 72(R13)
MULLU R2, R5, (R11, g)
MULLU R1, R5, (R14, R12)
MULALU R1, R6, (R11, g)
MULALU R0, R6, (R14, R12)
MULALU R0, R7, (R11, g)
MULALU R4, R7, (R14, R12)
ADD R2<<2, R2, R2
ADD R1<<2, R1, R1
MULALU R4, R8, (R11, g)
MULALU R3, R8, (R14, R12)
MULALU R3, R9, (R11, g)
MULALU R2, R9, (R14, R12)
MOVW g, 60(R13)
MOVW R11, 64(R13)
MOVW R12, 52(R13)
MOVW R14, 56(R13)
MULLU R0, R5, (R11, g)
MULALU R4, R6, (R11, g)
MULALU R3, R7, (R11, g)
MULALU R2, R8, (R11, g)
MULALU R1, R9, (R11, g)
ADD $52, R13, R0
MOVM.IA (R0), [R0-R7]
MOVW g>>26, R12
MOVW R4>>26, R14
ORR R11<<6, R12, R12
ORR R5<<6, R14, R14
BIC $0xfc000000, g, g
BIC $0xfc000000, R4, R4
ADD.S R12, R0, R0
ADC $0, R1, R1
ADD.S R14, R6, R6
ADC $0, R7, R7
MOVW R0>>26, R12
MOVW R6>>26, R14
ORR R1<<6, R12, R12
ORR R7<<6, R14, R14
BIC $0xfc000000, R0, R0
BIC $0xfc000000, R6, R6
ADD R14<<2, R14, R14
ADD.S R12, R2, R2
ADC $0, R3, R3
ADD R14, g, g
MOVW R2>>26, R12
MOVW g>>26, R14
ORR R3<<6, R12, R12
BIC $0xfc000000, g, R5
BIC $0xfc000000, R2, R7
ADD R12, R4, R4
ADD R14, R0, R0
MOVW R4>>26, R12
BIC $0xfc000000, R4, R8
ADD R12, R6, R9
MOVW 96(R13), R12
MOVW 92(R13), R14
MOVW R0, R6
CMP $32, R12
SUB $16, R12, R12
MOVW R12, 96(R13)
BHS poly1305_blocks_armv6_mainloop
poly1305_blocks_armv6_done:
MOVW 88(R13), R12
MOVW R5, 20(R12)
MOVW R6, 24(R12)
MOVW R7, 28(R12)
MOVW R8, 32(R12)
MOVW R9, 36(R12)
ADD $48, R13, R0
MOVM.DA (R0), [R4-R8, R14]
RET
#define MOVHUP_UNALIGNED(Rsrc, Rdst, Rtmp) \
MOVBU.P 1(Rsrc), Rtmp; \
MOVBU.P Rtmp, 1(Rdst); \
MOVBU.P 1(Rsrc), Rtmp; \
MOVBU.P Rtmp, 1(Rdst)
#define MOVWP_UNALIGNED(Rsrc, Rdst, Rtmp) \
MOVHUP_UNALIGNED(Rsrc, Rdst, Rtmp); \
MOVHUP_UNALIGNED(Rsrc, Rdst, Rtmp)
// func poly1305_auth_armv6(out *[16]byte, m *byte, mlen uint32, key *[32]key)
TEXT ·poly1305_auth_armv6(SB), $196-16
// The value 196, just above, is the sum of 64 (the size of the context
// structure) and 132 (the amount of stack needed).
//
// At this point, the stack pointer (R13) has been moved down. It
// points to the saved link register and there's 196 bytes of free
// space above it.
//
// The stack for this function looks like:
//
// +---------------------
// |
// | 64 bytes of context structure
// |
// +---------------------
// |
// | 112 bytes for poly1305_blocks_armv6
// |
// +---------------------
// | 16 bytes of final block, constructed at
// | poly1305_finish_ext_armv6_skip8
// +---------------------
// | four bytes of saved 'g'
// +---------------------
// | lr, saved by prelude <- R13 points here
// +---------------------
MOVW g, 4(R13)
MOVW out+0(FP), R4
MOVW m+4(FP), R5
MOVW mlen+8(FP), R6
MOVW key+12(FP), R7
ADD $136, R13, R0 // 136 = 4 + 4 + 16 + 112
MOVW R7, R1
// poly1305_init_ext_armv6 will write to the stack from R13+4, but
// that's ok because none of the other values have been written yet.
BL poly1305_init_ext_armv6<>(SB)
BIC.S $15, R6, R2
BEQ poly1305_auth_armv6_noblocks
ADD $136, R13, R0
MOVW R5, R1
ADD R2, R5, R5
SUB R2, R6, R6
BL poly1305_blocks_armv6<>(SB)
poly1305_auth_armv6_noblocks:
ADD $136, R13, R0
MOVW R5, R1
MOVW R6, R2
MOVW R4, R3
MOVW R0, R5
MOVW R1, R6
MOVW R2, R7
MOVW R3, R8
AND.S R2, R2, R2
BEQ poly1305_finish_ext_armv6_noremaining
EOR R0, R0
ADD $8, R13, R9 // 8 = offset to 16 byte scratch space
MOVW R0, (R9)
MOVW R0, 4(R9)
MOVW R0, 8(R9)
MOVW R0, 12(R9)
WORD $0xe3110003 // TST R1, #3 not working see issue 5921
BEQ poly1305_finish_ext_armv6_aligned
WORD $0xe3120008 // TST R2, #8 not working see issue 5921
BEQ poly1305_finish_ext_armv6_skip8
MOVWP_UNALIGNED(R1, R9, g)
MOVWP_UNALIGNED(R1, R9, g)
poly1305_finish_ext_armv6_skip8:
WORD $0xe3120004 // TST $4, R2 not working see issue 5921
BEQ poly1305_finish_ext_armv6_skip4
MOVWP_UNALIGNED(R1, R9, g)
poly1305_finish_ext_armv6_skip4:
WORD $0xe3120002 // TST $2, R2 not working see issue 5921
BEQ poly1305_finish_ext_armv6_skip2
MOVHUP_UNALIGNED(R1, R9, g)
B poly1305_finish_ext_armv6_skip2
poly1305_finish_ext_armv6_aligned:
WORD $0xe3120008 // TST R2, #8 not working see issue 5921
BEQ poly1305_finish_ext_armv6_skip8_aligned
MOVM.IA.W (R1), [g-R11]
MOVM.IA.W [g-R11], (R9)
poly1305_finish_ext_armv6_skip8_aligned:
WORD $0xe3120004 // TST $4, R2 not working see issue 5921
BEQ poly1305_finish_ext_armv6_skip4_aligned
MOVW.P 4(R1), g
MOVW.P g, 4(R9)
poly1305_finish_ext_armv6_skip4_aligned:
WORD $0xe3120002 // TST $2, R2 not working see issue 5921
BEQ poly1305_finish_ext_armv6_skip2
MOVHU.P 2(R1), g
MOVH.P g, 2(R9)
poly1305_finish_ext_armv6_skip2:
WORD $0xe3120001 // TST $1, R2 not working see issue 5921
BEQ poly1305_finish_ext_armv6_skip1
MOVBU.P 1(R1), g
MOVBU.P g, 1(R9)
poly1305_finish_ext_armv6_skip1:
MOVW $1, R11
MOVBU R11, 0(R9)
MOVW R11, 56(R5)
MOVW R5, R0
ADD $8, R13, R1
MOVW $16, R2
BL poly1305_blocks_armv6<>(SB)
poly1305_finish_ext_armv6_noremaining:
MOVW 20(R5), R0
MOVW 24(R5), R1
MOVW 28(R5), R2
MOVW 32(R5), R3
MOVW 36(R5), R4
MOVW R4>>26, R12
BIC $0xfc000000, R4, R4
ADD R12<<2, R12, R12
ADD R12, R0, R0
MOVW R0>>26, R12
BIC $0xfc000000, R0, R0
ADD R12, R1, R1
MOVW R1>>26, R12
BIC $0xfc000000, R1, R1
ADD R12, R2, R2
MOVW R2>>26, R12
BIC $0xfc000000, R2, R2
ADD R12, R3, R3
MOVW R3>>26, R12
BIC $0xfc000000, R3, R3
ADD R12, R4, R4
ADD $5, R0, R6
MOVW R6>>26, R12
BIC $0xfc000000, R6, R6
ADD R12, R1, R7
MOVW R7>>26, R12
BIC $0xfc000000, R7, R7
ADD R12, R2, g
MOVW g>>26, R12
BIC $0xfc000000, g, g
ADD R12, R3, R11
MOVW $-(1<<26), R12
ADD R11>>26, R12, R12
BIC $0xfc000000, R11, R11
ADD R12, R4, R9
MOVW R9>>31, R12
SUB $1, R12
AND R12, R6, R6
AND R12, R7, R7
AND R12, g, g
AND R12, R11, R11
AND R12, R9, R9
MVN R12, R12
AND R12, R0, R0
AND R12, R1, R1
AND R12, R2, R2
AND R12, R3, R3
AND R12, R4, R4
ORR R6, R0, R0
ORR R7, R1, R1
ORR g, R2, R2
ORR R11, R3, R3
ORR R9, R4, R4
ORR R1<<26, R0, R0
MOVW R1>>6, R1
ORR R2<<20, R1, R1
MOVW R2>>12, R2
ORR R3<<14, R2, R2
MOVW R3>>18, R3
ORR R4<<8, R3, R3
MOVW 40(R5), R6
MOVW 44(R5), R7
MOVW 48(R5), g
MOVW 52(R5), R11
ADD.S R6, R0, R0
ADC.S R7, R1, R1
ADC.S g, R2, R2
ADC.S R11, R3, R3
MOVM.IA [R0-R3], (R8)
MOVW R5, R12
EOR R0, R0, R0
EOR R1, R1, R1
EOR R2, R2, R2
EOR R3, R3, R3
EOR R4, R4, R4
EOR R5, R5, R5
EOR R6, R6, R6
EOR R7, R7, R7
MOVM.IA.W [R0-R7], (R12)
MOVM.IA [R0-R7], (R12)
MOVW 4(R13), g
RET

141
vendor/golang.org/x/crypto/poly1305/sum_ref.go generated vendored Normal file
View File

@@ -0,0 +1,141 @@
// Copyright 2012 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 !amd64,!arm gccgo appengine nacl
package poly1305
import "encoding/binary"
// Sum generates an authenticator for msg using a one-time key and puts the
// 16-byte result into out. Authenticating two different messages with the same
// key allows an attacker to forge messages at will.
func Sum(out *[TagSize]byte, msg []byte, key *[32]byte) {
var (
h0, h1, h2, h3, h4 uint32 // the hash accumulators
r0, r1, r2, r3, r4 uint64 // the r part of the key
)
r0 = uint64(binary.LittleEndian.Uint32(key[0:]) & 0x3ffffff)
r1 = uint64((binary.LittleEndian.Uint32(key[3:]) >> 2) & 0x3ffff03)
r2 = uint64((binary.LittleEndian.Uint32(key[6:]) >> 4) & 0x3ffc0ff)
r3 = uint64((binary.LittleEndian.Uint32(key[9:]) >> 6) & 0x3f03fff)
r4 = uint64((binary.LittleEndian.Uint32(key[12:]) >> 8) & 0x00fffff)
R1, R2, R3, R4 := r1*5, r2*5, r3*5, r4*5
for len(msg) >= TagSize {
// h += msg
h0 += binary.LittleEndian.Uint32(msg[0:]) & 0x3ffffff
h1 += (binary.LittleEndian.Uint32(msg[3:]) >> 2) & 0x3ffffff
h2 += (binary.LittleEndian.Uint32(msg[6:]) >> 4) & 0x3ffffff
h3 += (binary.LittleEndian.Uint32(msg[9:]) >> 6) & 0x3ffffff
h4 += (binary.LittleEndian.Uint32(msg[12:]) >> 8) | (1 << 24)
// h *= r
d0 := (uint64(h0) * r0) + (uint64(h1) * R4) + (uint64(h2) * R3) + (uint64(h3) * R2) + (uint64(h4) * R1)
d1 := (d0 >> 26) + (uint64(h0) * r1) + (uint64(h1) * r0) + (uint64(h2) * R4) + (uint64(h3) * R3) + (uint64(h4) * R2)
d2 := (d1 >> 26) + (uint64(h0) * r2) + (uint64(h1) * r1) + (uint64(h2) * r0) + (uint64(h3) * R4) + (uint64(h4) * R3)
d3 := (d2 >> 26) + (uint64(h0) * r3) + (uint64(h1) * r2) + (uint64(h2) * r1) + (uint64(h3) * r0) + (uint64(h4) * R4)
d4 := (d3 >> 26) + (uint64(h0) * r4) + (uint64(h1) * r3) + (uint64(h2) * r2) + (uint64(h3) * r1) + (uint64(h4) * r0)
// h %= p
h0 = uint32(d0) & 0x3ffffff
h1 = uint32(d1) & 0x3ffffff
h2 = uint32(d2) & 0x3ffffff
h3 = uint32(d3) & 0x3ffffff
h4 = uint32(d4) & 0x3ffffff
h0 += uint32(d4>>26) * 5
h1 += h0 >> 26
h0 = h0 & 0x3ffffff
msg = msg[TagSize:]
}
if len(msg) > 0 {
var block [TagSize]byte
off := copy(block[:], msg)
block[off] = 0x01
// h += msg
h0 += binary.LittleEndian.Uint32(block[0:]) & 0x3ffffff
h1 += (binary.LittleEndian.Uint32(block[3:]) >> 2) & 0x3ffffff
h2 += (binary.LittleEndian.Uint32(block[6:]) >> 4) & 0x3ffffff
h3 += (binary.LittleEndian.Uint32(block[9:]) >> 6) & 0x3ffffff
h4 += (binary.LittleEndian.Uint32(block[12:]) >> 8)
// h *= r
d0 := (uint64(h0) * r0) + (uint64(h1) * R4) + (uint64(h2) * R3) + (uint64(h3) * R2) + (uint64(h4) * R1)
d1 := (d0 >> 26) + (uint64(h0) * r1) + (uint64(h1) * r0) + (uint64(h2) * R4) + (uint64(h3) * R3) + (uint64(h4) * R2)
d2 := (d1 >> 26) + (uint64(h0) * r2) + (uint64(h1) * r1) + (uint64(h2) * r0) + (uint64(h3) * R4) + (uint64(h4) * R3)
d3 := (d2 >> 26) + (uint64(h0) * r3) + (uint64(h1) * r2) + (uint64(h2) * r1) + (uint64(h3) * r0) + (uint64(h4) * R4)
d4 := (d3 >> 26) + (uint64(h0) * r4) + (uint64(h1) * r3) + (uint64(h2) * r2) + (uint64(h3) * r1) + (uint64(h4) * r0)
// h %= p
h0 = uint32(d0) & 0x3ffffff
h1 = uint32(d1) & 0x3ffffff
h2 = uint32(d2) & 0x3ffffff
h3 = uint32(d3) & 0x3ffffff
h4 = uint32(d4) & 0x3ffffff
h0 += uint32(d4>>26) * 5
h1 += h0 >> 26
h0 = h0 & 0x3ffffff
}
// h %= p reduction
h2 += h1 >> 26
h1 &= 0x3ffffff
h3 += h2 >> 26
h2 &= 0x3ffffff
h4 += h3 >> 26
h3 &= 0x3ffffff
h0 += 5 * (h4 >> 26)
h4 &= 0x3ffffff
h1 += h0 >> 26
h0 &= 0x3ffffff
// h - p
t0 := h0 + 5
t1 := h1 + (t0 >> 26)
t2 := h2 + (t1 >> 26)
t3 := h3 + (t2 >> 26)
t4 := h4 + (t3 >> 26) - (1 << 26)
t0 &= 0x3ffffff
t1 &= 0x3ffffff
t2 &= 0x3ffffff
t3 &= 0x3ffffff
// select h if h < p else h - p
t_mask := (t4 >> 31) - 1
h_mask := ^t_mask
h0 = (h0 & h_mask) | (t0 & t_mask)
h1 = (h1 & h_mask) | (t1 & t_mask)
h2 = (h2 & h_mask) | (t2 & t_mask)
h3 = (h3 & h_mask) | (t3 & t_mask)
h4 = (h4 & h_mask) | (t4 & t_mask)
// h %= 2^128
h0 |= h1 << 26
h1 = ((h1 >> 6) | (h2 << 20))
h2 = ((h2 >> 12) | (h3 << 14))
h3 = ((h3 >> 18) | (h4 << 8))
// s: the s part of the key
// tag = (h + s) % (2^128)
t := uint64(h0) + uint64(binary.LittleEndian.Uint32(key[16:]))
h0 = uint32(t)
t = uint64(h1) + uint64(binary.LittleEndian.Uint32(key[20:])) + (t >> 32)
h1 = uint32(t)
t = uint64(h2) + uint64(binary.LittleEndian.Uint32(key[24:])) + (t >> 32)
h2 = uint32(t)
t = uint64(h3) + uint64(binary.LittleEndian.Uint32(key[28:])) + (t >> 32)
h3 = uint32(t)
binary.LittleEndian.PutUint32(out[0:], h0)
binary.LittleEndian.PutUint32(out[4:], h1)
binary.LittleEndian.PutUint32(out[8:], h2)
binary.LittleEndian.PutUint32(out[12:], h3)
}