admin/kubesphere
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: kubesphere 4.0 (#6115)

Browse Source 
* feat: kubesphere 4.0

Signed-off-by: ci-bot <ci-bot@kubesphere.io>

* feat: kubesphere 4.0

Signed-off-by: ci-bot <ci-bot@kubesphere.io>

---------

Signed-off-by: ci-bot <ci-bot@kubesphere.io>
Co-authored-by: ks-ci-bot <ks-ci-bot@example.com>
Co-authored-by: joyceliu <joyceliu@yunify.com>
This commit is contained in:
KubeSphere CI Bot
2024-09-06 11:05:52 +08:00
committed by GitHub
parent b5015ec7b9
commit 447a51f08b
8557 changed files with 546695 additions and 1146174 deletions
Download Patch File Download Diff File Expand all files Collapse all files

283
vendor/golang.org/x/crypto/argon2/argon2.go generated vendored Normal file
View File

@@ -0,0 +1,283 @@
// Copyright 2017 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 argon2 implements the key derivation function Argon2.
// Argon2 was selected as the winner of the Password Hashing Competition and can
// be used to derive cryptographic keys from passwords.
//
// For a detailed specification of Argon2 see [1].
//
// If you aren't sure which function you need, use Argon2id (IDKey) and
// the parameter recommendations for your scenario.
//
// # Argon2i
//
// Argon2i (implemented by Key) is the side-channel resistant version of Argon2.
// It uses data-independent memory access, which is preferred for password
// hashing and password-based key derivation. Argon2i requires more passes over
// memory than Argon2id to protect from trade-off attacks. The recommended
// parameters (taken from [2]) for non-interactive operations are time=3 and to
// use the maximum available memory.
//
// # Argon2id
//
// Argon2id (implemented by IDKey) is a hybrid version of Argon2 combining
// Argon2i and Argon2d. It uses data-independent memory access for the first
// half of the first iteration over the memory and data-dependent memory access
// for the rest. Argon2id is side-channel resistant and provides better brute-
// force cost savings due to time-memory tradeoffs than Argon2i. The recommended
// parameters for non-interactive operations (taken from [2]) are time=1 and to
// use the maximum available memory.
//
// [1] https://github.com/P-H-C/phc-winner-argon2/blob/master/argon2-specs.pdf
// [2] https://tools.ietf.org/html/draft-irtf-cfrg-argon2-03#section-9.3
package argon2
import (
"encoding/binary"
"sync"
"golang.org/x/crypto/blake2b"
)
// The Argon2 version implemented by this package.
const Version = 0x13
const (
argon2d = iota
argon2i
argon2id
)
// Key derives a key from the password, salt, and cost parameters using Argon2i
// returning a byte slice of length keyLen that can be used as cryptographic
// key. The CPU cost and parallelism degree must be greater than zero.
//
// For example, you can get a derived key for e.g. AES-256 (which needs a
// 32-byte key) by doing:
//
// key := argon2.Key([]byte("some password"), salt, 3, 32*1024, 4, 32)
//
// The draft RFC recommends[2] time=3, and memory=32*1024 is a sensible number.
// If using that amount of memory (32 MB) is not possible in some contexts then
// the time parameter can be increased to compensate.
//
// The time parameter specifies the number of passes over the memory and the
// memory parameter specifies the size of the memory in KiB. For example
// memory=32*1024 sets the memory cost to ~32 MB. The number of threads can be
// adjusted to the number of available CPUs. The cost parameters should be
// increased as memory latency and CPU parallelism increases. Remember to get a
// good random salt.
func Key(password, salt []byte, time, memory uint32, threads uint8, keyLen uint32) []byte {
return deriveKey(argon2i, password, salt, nil, nil, time, memory, threads, keyLen)
}
// IDKey derives a key from the password, salt, and cost parameters using
// Argon2id returning a byte slice of length keyLen that can be used as
// cryptographic key. The CPU cost and parallelism degree must be greater than
// zero.
//
// For example, you can get a derived key for e.g. AES-256 (which needs a
// 32-byte key) by doing:
//
// key := argon2.IDKey([]byte("some password"), salt, 1, 64*1024, 4, 32)
//
// The draft RFC recommends[2] time=1, and memory=64*1024 is a sensible number.
// If using that amount of memory (64 MB) is not possible in some contexts then
// the time parameter can be increased to compensate.
//
// The time parameter specifies the number of passes over the memory and the
// memory parameter specifies the size of the memory in KiB. For example
// memory=64*1024 sets the memory cost to ~64 MB. The number of threads can be
// adjusted to the numbers of available CPUs. The cost parameters should be
// increased as memory latency and CPU parallelism increases. Remember to get a
// good random salt.
func IDKey(password, salt []byte, time, memory uint32, threads uint8, keyLen uint32) []byte {
return deriveKey(argon2id, password, salt, nil, nil, time, memory, threads, keyLen)
}
func deriveKey(mode int, password, salt, secret, data []byte, time, memory uint32, threads uint8, keyLen uint32) []byte {
if time < 1 {
panic("argon2: number of rounds too small")
}
if threads < 1 {
panic("argon2: parallelism degree too low")
}
h0 := initHash(password, salt, secret, data, time, memory, uint32(threads), keyLen, mode)
memory = memory / (syncPoints * uint32(threads)) * (syncPoints * uint32(threads))
if memory < 2*syncPoints*uint32(threads) {
memory = 2 * syncPoints * uint32(threads)
}
B := initBlocks(&h0, memory, uint32(threads))
processBlocks(B, time, memory, uint32(threads), mode)
return extractKey(B, memory, uint32(threads), keyLen)
}
const (
blockLength = 128
syncPoints = 4
)
type block [blockLength]uint64
func initHash(password, salt, key, data []byte, time, memory, threads, keyLen uint32, mode int) [blake2b.Size + 8]byte {
var (
h0 [blake2b.Size + 8]byte
params [24]byte
tmp [4]byte
)
b2, _ := blake2b.New512(nil)
binary.LittleEndian.PutUint32(params[0:4], threads)
binary.LittleEndian.PutUint32(params[4:8], keyLen)
binary.LittleEndian.PutUint32(params[8:12], memory)
binary.LittleEndian.PutUint32(params[12:16], time)
binary.LittleEndian.PutUint32(params[16:20], uint32(Version))
binary.LittleEndian.PutUint32(params[20:24], uint32(mode))
b2.Write(params[:])
binary.LittleEndian.PutUint32(tmp[:], uint32(len(password)))
b2.Write(tmp[:])
b2.Write(password)
binary.LittleEndian.PutUint32(tmp[:], uint32(len(salt)))
b2.Write(tmp[:])
b2.Write(salt)
binary.LittleEndian.PutUint32(tmp[:], uint32(len(key)))
b2.Write(tmp[:])
b2.Write(key)
binary.LittleEndian.PutUint32(tmp[:], uint32(len(data)))
b2.Write(tmp[:])
b2.Write(data)
b2.Sum(h0[:0])
return h0
}
func initBlocks(h0 *[blake2b.Size + 8]byte, memory, threads uint32) []block {
var block0 [1024]byte
B := make([]block, memory)
for lane := uint32(0); lane < threads; lane++ {
j := lane * (memory / threads)
binary.LittleEndian.PutUint32(h0[blake2b.Size+4:], lane)
binary.LittleEndian.PutUint32(h0[blake2b.Size:], 0)
blake2bHash(block0[:], h0[:])
for i := range B[j+0] {
B[j+0][i] = binary.LittleEndian.Uint64(block0[i*8:])
}
binary.LittleEndian.PutUint32(h0[blake2b.Size:], 1)
blake2bHash(block0[:], h0[:])
for i := range B[j+1] {
B[j+1][i] = binary.LittleEndian.Uint64(block0[i*8:])
}
}
return B
}
func processBlocks(B []block, time, memory, threads uint32, mode int) {
lanes := memory / threads
segments := lanes / syncPoints
processSegment := func(n, slice, lane uint32, wg *sync.WaitGroup) {
var addresses, in, zero block
if mode == argon2i || (mode == argon2id && n == 0 && slice < syncPoints/2) {
in[0] = uint64(n)
in[1] = uint64(lane)
in[2] = uint64(slice)
in[3] = uint64(memory)
in[4] = uint64(time)
in[5] = uint64(mode)
}
index := uint32(0)
if n == 0 && slice == 0 {
index = 2 // we have already generated the first two blocks
if mode == argon2i || mode == argon2id {
in[6]++
processBlock(&addresses, &in, &zero)
processBlock(&addresses, &addresses, &zero)
}
}
offset := lane*lanes + slice*segments + index
var random uint64
for index < segments {
prev := offset - 1
if index == 0 && slice == 0 {
prev += lanes // last block in lane
}
if mode == argon2i || (mode == argon2id && n == 0 && slice < syncPoints/2) {
if index%blockLength == 0 {
in[6]++
processBlock(&addresses, &in, &zero)
processBlock(&addresses, &addresses, &zero)
}
random = addresses[index%blockLength]
} else {
random = B[prev][0]
}
newOffset := indexAlpha(random, lanes, segments, threads, n, slice, lane, index)
processBlockXOR(&B[offset], &B[prev], &B[newOffset])
index, offset = index+1, offset+1
}
wg.Done()
}
for n := uint32(0); n < time; n++ {
for slice := uint32(0); slice < syncPoints; slice++ {
var wg sync.WaitGroup
for lane := uint32(0); lane < threads; lane++ {
wg.Add(1)
go processSegment(n, slice, lane, &wg)
}
wg.Wait()
}
}
}
func extractKey(B []block, memory, threads, keyLen uint32) []byte {
lanes := memory / threads
for lane := uint32(0); lane < threads-1; lane++ {
for i, v := range B[(lane*lanes)+lanes-1] {
B[memory-1][i] ^= v
}
}
var block [1024]byte
for i, v := range B[memory-1] {
binary.LittleEndian.PutUint64(block[i*8:], v)
}
key := make([]byte, keyLen)
blake2bHash(key, block[:])
return key
}
func indexAlpha(rand uint64, lanes, segments, threads, n, slice, lane, index uint32) uint32 {
refLane := uint32(rand>>32) % threads
if n == 0 && slice == 0 {
refLane = lane
}
m, s := 3*segments, ((slice+1)%syncPoints)*segments
if lane == refLane {
m += index
}
if n == 0 {
m, s = slice*segments, 0
if slice == 0 || lane == refLane {
m += index
}
}
if index == 0 || lane == refLane {
m--
}
return phi(rand, uint64(m), uint64(s), refLane, lanes)
}
func phi(rand, m, s uint64, lane, lanes uint32) uint32 {
p := rand & 0xFFFFFFFF
p = (p * p) >> 32
p = (p * m) >> 32
return lane*lanes + uint32((s+m-(p+1))%uint64(lanes))
}

53
vendor/golang.org/x/crypto/argon2/blake2b.go generated vendored Normal file
View File

@@ -0,0 +1,53 @@
// Copyright 2017 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 argon2
import (
"encoding/binary"
"hash"
"golang.org/x/crypto/blake2b"
)
// blake2bHash computes an arbitrary long hash value of in
// and writes the hash to out.
func blake2bHash(out []byte, in []byte) {
var b2 hash.Hash
if n := len(out); n < blake2b.Size {
b2, _ = blake2b.New(n, nil)
} else {
b2, _ = blake2b.New512(nil)
}
var buffer [blake2b.Size]byte
binary.LittleEndian.PutUint32(buffer[:4], uint32(len(out)))
b2.Write(buffer[:4])
b2.Write(in)
if len(out) <= blake2b.Size {
b2.Sum(out[:0])
return
}
outLen := len(out)
b2.Sum(buffer[:0])
b2.Reset()
copy(out, buffer[:32])
out = out[32:]
for len(out) > blake2b.Size {
b2.Write(buffer[:])
b2.Sum(buffer[:0])
copy(out, buffer[:32])
out = out[32:]
b2.Reset()
}
if outLen%blake2b.Size > 0 { // outLen > 64
r := ((outLen + 31) / 32) - 2 // ⌈τ /32⌉-2
b2, _ = blake2b.New(outLen-32*r, nil)
}
b2.Write(buffer[:])
b2.Sum(out[:0])
}

61
vendor/golang.org/x/crypto/argon2/blamka_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,61 @@
// Copyright 2017 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.
//go:build amd64 && gc && !purego
// +build amd64,gc,!purego
package argon2
import "golang.org/x/sys/cpu"
func init() {
useSSE4 = cpu.X86.HasSSE41
}
//go:noescape
func mixBlocksSSE2(out, a, b, c *block)
//go:noescape
func xorBlocksSSE2(out, a, b, c *block)
//go:noescape
func blamkaSSE4(b *block)
func processBlockSSE(out, in1, in2 *block, xor bool) {
var t block
mixBlocksSSE2(&t, in1, in2, &t)
if useSSE4 {
blamkaSSE4(&t)
} else {
for i := 0; i < blockLength; i += 16 {
blamkaGeneric(
&t[i+0], &t[i+1], &t[i+2], &t[i+3],
&t[i+4], &t[i+5], &t[i+6], &t[i+7],
&t[i+8], &t[i+9], &t[i+10], &t[i+11],
&t[i+12], &t[i+13], &t[i+14], &t[i+15],
)
}
for i := 0; i < blockLength/8; i += 2 {
blamkaGeneric(
&t[i], &t[i+1], &t[16+i], &t[16+i+1],
&t[32+i], &t[32+i+1], &t[48+i], &t[48+i+1],
&t[64+i], &t[64+i+1], &t[80+i], &t[80+i+1],
&t[96+i], &t[96+i+1], &t[112+i], &t[112+i+1],
)
}
}
if xor {
xorBlocksSSE2(out, in1, in2, &t)
} else {
mixBlocksSSE2(out, in1, in2, &t)
}
}
func processBlock(out, in1, in2 *block) {
processBlockSSE(out, in1, in2, false)
}
func processBlockXOR(out, in1, in2 *block) {
processBlockSSE(out, in1, in2, true)
}

244
vendor/golang.org/x/crypto/argon2/blamka_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,244 @@
// Copyright 2017 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.
//go:build amd64 && gc && !purego
// +build amd64,gc,!purego
#include "textflag.h"
DATA ·c40<>+0x00(SB)/8, $0x0201000706050403
DATA ·c40<>+0x08(SB)/8, $0x0a09080f0e0d0c0b
GLOBL ·c40<>(SB), (NOPTR+RODATA), $16
DATA ·c48<>+0x00(SB)/8, $0x0100070605040302
DATA ·c48<>+0x08(SB)/8, $0x09080f0e0d0c0b0a
GLOBL ·c48<>(SB), (NOPTR+RODATA), $16
#define SHUFFLE(v2, v3, v4, v5, v6, v7, t1, t2) \
MOVO v4, t1; \
MOVO v5, v4; \
MOVO t1, v5; \
MOVO v6, t1; \
PUNPCKLQDQ v6, t2; \
PUNPCKHQDQ v7, v6; \
PUNPCKHQDQ t2, v6; \
PUNPCKLQDQ v7, t2; \
MOVO t1, v7; \
MOVO v2, t1; \
PUNPCKHQDQ t2, v7; \
PUNPCKLQDQ v3, t2; \
PUNPCKHQDQ t2, v2; \
PUNPCKLQDQ t1, t2; \
PUNPCKHQDQ t2, v3
#define SHUFFLE_INV(v2, v3, v4, v5, v6, v7, t1, t2) \
MOVO v4, t1; \
MOVO v5, v4; \
MOVO t1, v5; \
MOVO v2, t1; \
PUNPCKLQDQ v2, t2; \
PUNPCKHQDQ v3, v2; \
PUNPCKHQDQ t2, v2; \
PUNPCKLQDQ v3, t2; \
MOVO t1, v3; \
MOVO v6, t1; \
PUNPCKHQDQ t2, v3; \
PUNPCKLQDQ v7, t2; \
PUNPCKHQDQ t2, v6; \
PUNPCKLQDQ t1, t2; \
PUNPCKHQDQ t2, v7
#define HALF_ROUND(v0, v1, v2, v3, v4, v5, v6, v7, t0, c40, c48) \
MOVO v0, t0; \
PMULULQ v2, t0; \
PADDQ v2, v0; \
PADDQ t0, v0; \
PADDQ t0, v0; \
PXOR v0, v6; \
PSHUFD $0xB1, v6, v6; \
MOVO v4, t0; \
PMULULQ v6, t0; \
PADDQ v6, v4; \
PADDQ t0, v4; \
PADDQ t0, v4; \
PXOR v4, v2; \
PSHUFB c40, v2; \
MOVO v0, t0; \
PMULULQ v2, t0; \
PADDQ v2, v0; \
PADDQ t0, v0; \
PADDQ t0, v0; \
PXOR v0, v6; \
PSHUFB c48, v6; \
MOVO v4, t0; \
PMULULQ v6, t0; \
PADDQ v6, v4; \
PADDQ t0, v4; \
PADDQ t0, v4; \
PXOR v4, v2; \
MOVO v2, t0; \
PADDQ v2, t0; \
PSRLQ $63, v2; \
PXOR t0, v2; \
MOVO v1, t0; \
PMULULQ v3, t0; \
PADDQ v3, v1; \
PADDQ t0, v1; \
PADDQ t0, v1; \
PXOR v1, v7; \
PSHUFD $0xB1, v7, v7; \
MOVO v5, t0; \
PMULULQ v7, t0; \
PADDQ v7, v5; \
PADDQ t0, v5; \
PADDQ t0, v5; \
PXOR v5, v3; \
PSHUFB c40, v3; \
MOVO v1, t0; \
PMULULQ v3, t0; \
PADDQ v3, v1; \
PADDQ t0, v1; \
PADDQ t0, v1; \
PXOR v1, v7; \
PSHUFB c48, v7; \
MOVO v5, t0; \
PMULULQ v7, t0; \
PADDQ v7, v5; \
PADDQ t0, v5; \
PADDQ t0, v5; \
PXOR v5, v3; \
MOVO v3, t0; \
PADDQ v3, t0; \
PSRLQ $63, v3; \
PXOR t0, v3
#define LOAD_MSG_0(block, off) \
MOVOU 8*(off+0)(block), X0; \
MOVOU 8*(off+2)(block), X1; \
MOVOU 8*(off+4)(block), X2; \
MOVOU 8*(off+6)(block), X3; \
MOVOU 8*(off+8)(block), X4; \
MOVOU 8*(off+10)(block), X5; \
MOVOU 8*(off+12)(block), X6; \
MOVOU 8*(off+14)(block), X7
#define STORE_MSG_0(block, off) \
MOVOU X0, 8*(off+0)(block); \
MOVOU X1, 8*(off+2)(block); \
MOVOU X2, 8*(off+4)(block); \
MOVOU X3, 8*(off+6)(block); \
MOVOU X4, 8*(off+8)(block); \
MOVOU X5, 8*(off+10)(block); \
MOVOU X6, 8*(off+12)(block); \
MOVOU X7, 8*(off+14)(block)
#define LOAD_MSG_1(block, off) \
MOVOU 8*off+0*8(block), X0; \
MOVOU 8*off+16*8(block), X1; \
MOVOU 8*off+32*8(block), X2; \
MOVOU 8*off+48*8(block), X3; \
MOVOU 8*off+64*8(block), X4; \
MOVOU 8*off+80*8(block), X5; \
MOVOU 8*off+96*8(block), X6; \
MOVOU 8*off+112*8(block), X7
#define STORE_MSG_1(block, off) \
MOVOU X0, 8*off+0*8(block); \
MOVOU X1, 8*off+16*8(block); \
MOVOU X2, 8*off+32*8(block); \
MOVOU X3, 8*off+48*8(block); \
MOVOU X4, 8*off+64*8(block); \
MOVOU X5, 8*off+80*8(block); \
MOVOU X6, 8*off+96*8(block); \
MOVOU X7, 8*off+112*8(block)
#define BLAMKA_ROUND_0(block, off, t0, t1, c40, c48) \
LOAD_MSG_0(block, off); \
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, t0, c40, c48); \
SHUFFLE(X2, X3, X4, X5, X6, X7, t0, t1); \
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, t0, c40, c48); \
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, t0, t1); \
STORE_MSG_0(block, off)
#define BLAMKA_ROUND_1(block, off, t0, t1, c40, c48) \
LOAD_MSG_1(block, off); \
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, t0, c40, c48); \
SHUFFLE(X2, X3, X4, X5, X6, X7, t0, t1); \
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, t0, c40, c48); \
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, t0, t1); \
STORE_MSG_1(block, off)
// func blamkaSSE4(b *block)
TEXT ·blamkaSSE4(SB), 4, $0-8
MOVQ b+0(FP), AX
MOVOU ·c40<>(SB), X10
MOVOU ·c48<>(SB), X11
BLAMKA_ROUND_0(AX, 0, X8, X9, X10, X11)
BLAMKA_ROUND_0(AX, 16, X8, X9, X10, X11)
BLAMKA_ROUND_0(AX, 32, X8, X9, X10, X11)
BLAMKA_ROUND_0(AX, 48, X8, X9, X10, X11)
BLAMKA_ROUND_0(AX, 64, X8, X9, X10, X11)
BLAMKA_ROUND_0(AX, 80, X8, X9, X10, X11)
BLAMKA_ROUND_0(AX, 96, X8, X9, X10, X11)
BLAMKA_ROUND_0(AX, 112, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 0, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 2, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 4, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 6, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 8, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 10, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 12, X8, X9, X10, X11)
BLAMKA_ROUND_1(AX, 14, X8, X9, X10, X11)
RET
// func mixBlocksSSE2(out, a, b, c *block)
TEXT ·mixBlocksSSE2(SB), 4, $0-32
MOVQ out+0(FP), DX
MOVQ a+8(FP), AX
MOVQ b+16(FP), BX
MOVQ a+24(FP), CX
MOVQ $128, BP
loop:
MOVOU 0(AX), X0
MOVOU 0(BX), X1
MOVOU 0(CX), X2
PXOR X1, X0
PXOR X2, X0
MOVOU X0, 0(DX)
ADDQ $16, AX
ADDQ $16, BX
ADDQ $16, CX
ADDQ $16, DX
SUBQ $2, BP
JA loop
RET
// func xorBlocksSSE2(out, a, b, c *block)
TEXT ·xorBlocksSSE2(SB), 4, $0-32
MOVQ out+0(FP), DX
MOVQ a+8(FP), AX
MOVQ b+16(FP), BX
MOVQ a+24(FP), CX
MOVQ $128, BP
loop:
MOVOU 0(AX), X0
MOVOU 0(BX), X1
MOVOU 0(CX), X2
MOVOU 0(DX), X3
PXOR X1, X0
PXOR X2, X0
PXOR X3, X0
MOVOU X0, 0(DX)
ADDQ $16, AX
ADDQ $16, BX
ADDQ $16, CX
ADDQ $16, DX
SUBQ $2, BP
JA loop
RET

163
vendor/golang.org/x/crypto/argon2/blamka_generic.go generated vendored Normal file
View File

@@ -0,0 +1,163 @@
// Copyright 2017 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 argon2
var useSSE4 bool
func processBlockGeneric(out, in1, in2 *block, xor bool) {
var t block
for i := range t {
t[i] = in1[i] ^ in2[i]
}
for i := 0; i < blockLength; i += 16 {
blamkaGeneric(
&t[i+0], &t[i+1], &t[i+2], &t[i+3],
&t[i+4], &t[i+5], &t[i+6], &t[i+7],
&t[i+8], &t[i+9], &t[i+10], &t[i+11],
&t[i+12], &t[i+13], &t[i+14], &t[i+15],
)
}
for i := 0; i < blockLength/8; i += 2 {
blamkaGeneric(
&t[i], &t[i+1], &t[16+i], &t[16+i+1],
&t[32+i], &t[32+i+1], &t[48+i], &t[48+i+1],
&t[64+i], &t[64+i+1], &t[80+i], &t[80+i+1],
&t[96+i], &t[96+i+1], &t[112+i], &t[112+i+1],
)
}
if xor {
for i := range t {
out[i] ^= in1[i] ^ in2[i] ^ t[i]
}
} else {
for i := range t {
out[i] = in1[i] ^ in2[i] ^ t[i]
}
}
}
func blamkaGeneric(t00, t01, t02, t03, t04, t05, t06, t07, t08, t09, t10, t11, t12, t13, t14, t15 *uint64) {
v00, v01, v02, v03 := *t00, *t01, *t02, *t03
v04, v05, v06, v07 := *t04, *t05, *t06, *t07
v08, v09, v10, v11 := *t08, *t09, *t10, *t11
v12, v13, v14, v15 := *t12, *t13, *t14, *t15
v00 += v04 + 2*uint64(uint32(v00))*uint64(uint32(v04))
v12 ^= v00
v12 = v12>>32 | v12<<32
v08 += v12 + 2*uint64(uint32(v08))*uint64(uint32(v12))
v04 ^= v08
v04 = v04>>24 | v04<<40
v00 += v04 + 2*uint64(uint32(v00))*uint64(uint32(v04))
v12 ^= v00
v12 = v12>>16 | v12<<48
v08 += v12 + 2*uint64(uint32(v08))*uint64(uint32(v12))
v04 ^= v08
v04 = v04>>63 | v04<<1
v01 += v05 + 2*uint64(uint32(v01))*uint64(uint32(v05))
v13 ^= v01
v13 = v13>>32 | v13<<32
v09 += v13 + 2*uint64(uint32(v09))*uint64(uint32(v13))
v05 ^= v09
v05 = v05>>24 | v05<<40
v01 += v05 + 2*uint64(uint32(v01))*uint64(uint32(v05))
v13 ^= v01
v13 = v13>>16 | v13<<48
v09 += v13 + 2*uint64(uint32(v09))*uint64(uint32(v13))
v05 ^= v09
v05 = v05>>63 | v05<<1
v02 += v06 + 2*uint64(uint32(v02))*uint64(uint32(v06))
v14 ^= v02
v14 = v14>>32 | v14<<32
v10 += v14 + 2*uint64(uint32(v10))*uint64(uint32(v14))
v06 ^= v10
v06 = v06>>24 | v06<<40
v02 += v06 + 2*uint64(uint32(v02))*uint64(uint32(v06))
v14 ^= v02
v14 = v14>>16 | v14<<48
v10 += v14 + 2*uint64(uint32(v10))*uint64(uint32(v14))
v06 ^= v10
v06 = v06>>63 | v06<<1
v03 += v07 + 2*uint64(uint32(v03))*uint64(uint32(v07))
v15 ^= v03
v15 = v15>>32 | v15<<32
v11 += v15 + 2*uint64(uint32(v11))*uint64(uint32(v15))
v07 ^= v11
v07 = v07>>24 | v07<<40
v03 += v07 + 2*uint64(uint32(v03))*uint64(uint32(v07))
v15 ^= v03
v15 = v15>>16 | v15<<48
v11 += v15 + 2*uint64(uint32(v11))*uint64(uint32(v15))
v07 ^= v11
v07 = v07>>63 | v07<<1
v00 += v05 + 2*uint64(uint32(v00))*uint64(uint32(v05))
v15 ^= v00
v15 = v15>>32 | v15<<32
v10 += v15 + 2*uint64(uint32(v10))*uint64(uint32(v15))
v05 ^= v10
v05 = v05>>24 | v05<<40
v00 += v05 + 2*uint64(uint32(v00))*uint64(uint32(v05))
v15 ^= v00
v15 = v15>>16 | v15<<48
v10 += v15 + 2*uint64(uint32(v10))*uint64(uint32(v15))
v05 ^= v10
v05 = v05>>63 | v05<<1
v01 += v06 + 2*uint64(uint32(v01))*uint64(uint32(v06))
v12 ^= v01
v12 = v12>>32 | v12<<32
v11 += v12 + 2*uint64(uint32(v11))*uint64(uint32(v12))
v06 ^= v11
v06 = v06>>24 | v06<<40
v01 += v06 + 2*uint64(uint32(v01))*uint64(uint32(v06))
v12 ^= v01
v12 = v12>>16 | v12<<48
v11 += v12 + 2*uint64(uint32(v11))*uint64(uint32(v12))
v06 ^= v11
v06 = v06>>63 | v06<<1
v02 += v07 + 2*uint64(uint32(v02))*uint64(uint32(v07))
v13 ^= v02
v13 = v13>>32 | v13<<32
v08 += v13 + 2*uint64(uint32(v08))*uint64(uint32(v13))
v07 ^= v08
v07 = v07>>24 | v07<<40
v02 += v07 + 2*uint64(uint32(v02))*uint64(uint32(v07))
v13 ^= v02
v13 = v13>>16 | v13<<48
v08 += v13 + 2*uint64(uint32(v08))*uint64(uint32(v13))
v07 ^= v08
v07 = v07>>63 | v07<<1
v03 += v04 + 2*uint64(uint32(v03))*uint64(uint32(v04))
v14 ^= v03
v14 = v14>>32 | v14<<32
v09 += v14 + 2*uint64(uint32(v09))*uint64(uint32(v14))
v04 ^= v09
v04 = v04>>24 | v04<<40
v03 += v04 + 2*uint64(uint32(v03))*uint64(uint32(v04))
v14 ^= v03
v14 = v14>>16 | v14<<48
v09 += v14 + 2*uint64(uint32(v09))*uint64(uint32(v14))
v04 ^= v09
v04 = v04>>63 | v04<<1
*t00, *t01, *t02, *t03 = v00, v01, v02, v03
*t04, *t05, *t06, *t07 = v04, v05, v06, v07
*t08, *t09, *t10, *t11 = v08, v09, v10, v11
*t12, *t13, *t14, *t15 = v12, v13, v14, v15
}

16
vendor/golang.org/x/crypto/argon2/blamka_ref.go generated vendored Normal file
View File

@@ -0,0 +1,16 @@
// Copyright 2017 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.
//go:build !amd64 || purego || !gc
// +build !amd64 purego !gc
package argon2
func processBlock(out, in1, in2 *block) {
processBlockGeneric(out, in1, in2, false)
}
func processBlockXOR(out, in1, in2 *block) {
processBlockGeneric(out, in1, in2, true)
}

291
vendor/golang.org/x/crypto/blake2b/blake2b.go generated vendored Normal file
View File

@@ -0,0 +1,291 @@
// 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 blake2b implements the BLAKE2b hash algorithm defined by RFC 7693
// and the extendable output function (XOF) BLAKE2Xb.
//
// BLAKE2b is optimized for 64-bit platforms—including NEON-enabled ARMs—and
// produces digests of any size between 1 and 64 bytes.
// For a detailed specification of BLAKE2b see https://blake2.net/blake2.pdf
// and for BLAKE2Xb see https://blake2.net/blake2x.pdf
//
// If you aren't sure which function you need, use BLAKE2b (Sum512 or New512).
// If you need a secret-key MAC (message authentication code), use the New512
// function with a non-nil key.
//
// BLAKE2X is a construction to compute hash values larger than 64 bytes. It
// can produce hash values between 0 and 4 GiB.
package blake2b
import (
"encoding/binary"
"errors"
"hash"
)
const (
// The blocksize of BLAKE2b in bytes.
BlockSize = 128
// The hash size of BLAKE2b-512 in bytes.
Size = 64
// The hash size of BLAKE2b-384 in bytes.
Size384 = 48
// The hash size of BLAKE2b-256 in bytes.
Size256 = 32
)
var (
useAVX2 bool
useAVX bool
useSSE4 bool
)
var (
errKeySize = errors.New("blake2b: invalid key size")
errHashSize = errors.New("blake2b: invalid hash size")
)
var iv = [8]uint64{
0x6a09e667f3bcc908, 0xbb67ae8584caa73b, 0x3c6ef372fe94f82b, 0xa54ff53a5f1d36f1,
0x510e527fade682d1, 0x9b05688c2b3e6c1f, 0x1f83d9abfb41bd6b, 0x5be0cd19137e2179,
}
// Sum512 returns the BLAKE2b-512 checksum of the data.
func Sum512(data []byte) [Size]byte {
var sum [Size]byte
checkSum(&sum, Size, data)
return sum
}
// Sum384 returns the BLAKE2b-384 checksum of the data.
func Sum384(data []byte) [Size384]byte {
var sum [Size]byte
var sum384 [Size384]byte
checkSum(&sum, Size384, data)
copy(sum384[:], sum[:Size384])
return sum384
}
// Sum256 returns the BLAKE2b-256 checksum of the data.
func Sum256(data []byte) [Size256]byte {
var sum [Size]byte
var sum256 [Size256]byte
checkSum(&sum, Size256, data)
copy(sum256[:], sum[:Size256])
return sum256
}
// New512 returns a new hash.Hash computing the BLAKE2b-512 checksum. A non-nil
// key turns the hash into a MAC. The key must be between zero and 64 bytes long.
func New512(key []byte) (hash.Hash, error) { return newDigest(Size, key) }
// New384 returns a new hash.Hash computing the BLAKE2b-384 checksum. A non-nil
// key turns the hash into a MAC. The key must be between zero and 64 bytes long.
func New384(key []byte) (hash.Hash, error) { return newDigest(Size384, key) }
// New256 returns a new hash.Hash computing the BLAKE2b-256 checksum. A non-nil
// key turns the hash into a MAC. The key must be between zero and 64 bytes long.
func New256(key []byte) (hash.Hash, error) { return newDigest(Size256, key) }
// New returns a new hash.Hash computing the BLAKE2b checksum with a custom length.
// A non-nil key turns the hash into a MAC. The key must be between zero and 64 bytes long.
// The hash size can be a value between 1 and 64 but it is highly recommended to use
// values equal or greater than:
// - 32 if BLAKE2b is used as a hash function (The key is zero bytes long).
// - 16 if BLAKE2b is used as a MAC function (The key is at least 16 bytes long).
// When the key is nil, the returned hash.Hash implements BinaryMarshaler
// and BinaryUnmarshaler for state (de)serialization as documented by hash.Hash.
func New(size int, key []byte) (hash.Hash, error) { return newDigest(size, key) }
func newDigest(hashSize int, key []byte) (*digest, error) {
if hashSize < 1 || hashSize > Size {
return nil, errHashSize
}
if len(key) > Size {
return nil, errKeySize
}
d := &digest{
size: hashSize,
keyLen: len(key),
}
copy(d.key[:], key)
d.Reset()
return d, nil
}
func checkSum(sum *[Size]byte, hashSize int, data []byte) {
h := iv
h[0] ^= uint64(hashSize) | (1 << 16) | (1 << 24)
var c [2]uint64
if length := len(data); length > BlockSize {
n := length &^ (BlockSize - 1)
if length == n {
n -= BlockSize
}
hashBlocks(&h, &c, 0, data[:n])
data = data[n:]
}
var block [BlockSize]byte
offset := copy(block[:], data)
remaining := uint64(BlockSize - offset)
if c[0] < remaining {
c[1]--
}
c[0] -= remaining
hashBlocks(&h, &c, 0xFFFFFFFFFFFFFFFF, block[:])
for i, v := range h[:(hashSize+7)/8] {
binary.LittleEndian.PutUint64(sum[8*i:], v)
}
}
type digest struct {
h [8]uint64
c [2]uint64
size int
block [BlockSize]byte
offset int
key [BlockSize]byte
keyLen int
}
const (
magic = "b2b"
marshaledSize = len(magic) + 8*8 + 2*8 + 1 + BlockSize + 1
)
func (d *digest) MarshalBinary() ([]byte, error) {
if d.keyLen != 0 {
return nil, errors.New("crypto/blake2b: cannot marshal MACs")
}
b := make([]byte, 0, marshaledSize)
b = append(b, magic...)
for i := 0; i < 8; i++ {
b = appendUint64(b, d.h[i])
}
b = appendUint64(b, d.c[0])
b = appendUint64(b, d.c[1])
// Maximum value for size is 64
b = append(b, byte(d.size))
b = append(b, d.block[:]...)
b = append(b, byte(d.offset))
return b, nil
}
func (d *digest) UnmarshalBinary(b []byte) error {
if len(b) < len(magic) || string(b[:len(magic)]) != magic {
return errors.New("crypto/blake2b: invalid hash state identifier")
}
if len(b) != marshaledSize {
return errors.New("crypto/blake2b: invalid hash state size")
}
b = b[len(magic):]
for i := 0; i < 8; i++ {
b, d.h[i] = consumeUint64(b)
}
b, d.c[0] = consumeUint64(b)
b, d.c[1] = consumeUint64(b)
d.size = int(b[0])
b = b[1:]
copy(d.block[:], b[:BlockSize])
b = b[BlockSize:]
d.offset = int(b[0])
return nil
}
func (d *digest) BlockSize() int { return BlockSize }
func (d *digest) Size() int { return d.size }
func (d *digest) Reset() {
d.h = iv
d.h[0] ^= uint64(d.size) | (uint64(d.keyLen) << 8) | (1 << 16) | (1 << 24)
d.offset, d.c[0], d.c[1] = 0, 0, 0
if d.keyLen > 0 {
d.block = d.key
d.offset = BlockSize
}
}
func (d *digest) Write(p []byte) (n int, err error) {
n = len(p)
if d.offset > 0 {
remaining := BlockSize - d.offset
if n <= remaining {
d.offset += copy(d.block[d.offset:], p)
return
}
copy(d.block[d.offset:], p[:remaining])
hashBlocks(&d.h, &d.c, 0, d.block[:])
d.offset = 0
p = p[remaining:]
}
if length := len(p); length > BlockSize {
nn := length &^ (BlockSize - 1)
if length == nn {
nn -= BlockSize
}
hashBlocks(&d.h, &d.c, 0, p[:nn])
p = p[nn:]
}
if len(p) > 0 {
d.offset += copy(d.block[:], p)
}
return
}
func (d *digest) Sum(sum []byte) []byte {
var hash [Size]byte
d.finalize(&hash)
return append(sum, hash[:d.size]...)
}
func (d *digest) finalize(hash *[Size]byte) {
var block [BlockSize]byte
copy(block[:], d.block[:d.offset])
remaining := uint64(BlockSize - d.offset)
c := d.c
if c[0] < remaining {
c[1]--
}
c[0] -= remaining
h := d.h
hashBlocks(&h, &c, 0xFFFFFFFFFFFFFFFF, block[:])
for i, v := range h {
binary.LittleEndian.PutUint64(hash[8*i:], v)
}
}
func appendUint64(b []byte, x uint64) []byte {
var a [8]byte
binary.BigEndian.PutUint64(a[:], x)
return append(b, a[:]...)
}
func appendUint32(b []byte, x uint32) []byte {
var a [4]byte
binary.BigEndian.PutUint32(a[:], x)
return append(b, a[:]...)
}
func consumeUint64(b []byte) ([]byte, uint64) {
x := binary.BigEndian.Uint64(b)
return b[8:], x
}
func consumeUint32(b []byte) ([]byte, uint32) {
x := binary.BigEndian.Uint32(b)
return b[4:], x
}

38
vendor/golang.org/x/crypto/blake2b/blake2bAVX2_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,38 @@
// 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.
//go:build go1.7 && amd64 && gc && !purego
// +build go1.7,amd64,gc,!purego
package blake2b
import "golang.org/x/sys/cpu"
func init() {
useAVX2 = cpu.X86.HasAVX2
useAVX = cpu.X86.HasAVX
useSSE4 = cpu.X86.HasSSE41
}
//go:noescape
func hashBlocksAVX2(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte)
//go:noescape
func hashBlocksAVX(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte)
//go:noescape
func hashBlocksSSE4(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte)
func hashBlocks(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte) {
switch {
case useAVX2:
hashBlocksAVX2(h, c, flag, blocks)
case useAVX:
hashBlocksAVX(h, c, flag, blocks)
case useSSE4:
hashBlocksSSE4(h, c, flag, blocks)
default:
hashBlocksGeneric(h, c, flag, blocks)
}
}

745
vendor/golang.org/x/crypto/blake2b/blake2bAVX2_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,745 @@
// 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.
//go:build go1.7 && amd64 && gc && !purego
// +build go1.7,amd64,gc,!purego
#include "textflag.h"
DATA ·AVX2_iv0<>+0x00(SB)/8, $0x6a09e667f3bcc908
DATA ·AVX2_iv0<>+0x08(SB)/8, $0xbb67ae8584caa73b
DATA ·AVX2_iv0<>+0x10(SB)/8, $0x3c6ef372fe94f82b
DATA ·AVX2_iv0<>+0x18(SB)/8, $0xa54ff53a5f1d36f1
GLOBL ·AVX2_iv0<>(SB), (NOPTR+RODATA), $32
DATA ·AVX2_iv1<>+0x00(SB)/8, $0x510e527fade682d1
DATA ·AVX2_iv1<>+0x08(SB)/8, $0x9b05688c2b3e6c1f
DATA ·AVX2_iv1<>+0x10(SB)/8, $0x1f83d9abfb41bd6b
DATA ·AVX2_iv1<>+0x18(SB)/8, $0x5be0cd19137e2179
GLOBL ·AVX2_iv1<>(SB), (NOPTR+RODATA), $32
DATA ·AVX2_c40<>+0x00(SB)/8, $0x0201000706050403
DATA ·AVX2_c40<>+0x08(SB)/8, $0x0a09080f0e0d0c0b
DATA ·AVX2_c40<>+0x10(SB)/8, $0x0201000706050403
DATA ·AVX2_c40<>+0x18(SB)/8, $0x0a09080f0e0d0c0b
GLOBL ·AVX2_c40<>(SB), (NOPTR+RODATA), $32
DATA ·AVX2_c48<>+0x00(SB)/8, $0x0100070605040302
DATA ·AVX2_c48<>+0x08(SB)/8, $0x09080f0e0d0c0b0a
DATA ·AVX2_c48<>+0x10(SB)/8, $0x0100070605040302
DATA ·AVX2_c48<>+0x18(SB)/8, $0x09080f0e0d0c0b0a
GLOBL ·AVX2_c48<>(SB), (NOPTR+RODATA), $32
DATA ·AVX_iv0<>+0x00(SB)/8, $0x6a09e667f3bcc908
DATA ·AVX_iv0<>+0x08(SB)/8, $0xbb67ae8584caa73b
GLOBL ·AVX_iv0<>(SB), (NOPTR+RODATA), $16
DATA ·AVX_iv1<>+0x00(SB)/8, $0x3c6ef372fe94f82b
DATA ·AVX_iv1<>+0x08(SB)/8, $0xa54ff53a5f1d36f1
GLOBL ·AVX_iv1<>(SB), (NOPTR+RODATA), $16
DATA ·AVX_iv2<>+0x00(SB)/8, $0x510e527fade682d1
DATA ·AVX_iv2<>+0x08(SB)/8, $0x9b05688c2b3e6c1f
GLOBL ·AVX_iv2<>(SB), (NOPTR+RODATA), $16
DATA ·AVX_iv3<>+0x00(SB)/8, $0x1f83d9abfb41bd6b
DATA ·AVX_iv3<>+0x08(SB)/8, $0x5be0cd19137e2179
GLOBL ·AVX_iv3<>(SB), (NOPTR+RODATA), $16
DATA ·AVX_c40<>+0x00(SB)/8, $0x0201000706050403
DATA ·AVX_c40<>+0x08(SB)/8, $0x0a09080f0e0d0c0b
GLOBL ·AVX_c40<>(SB), (NOPTR+RODATA), $16
DATA ·AVX_c48<>+0x00(SB)/8, $0x0100070605040302
DATA ·AVX_c48<>+0x08(SB)/8, $0x09080f0e0d0c0b0a
GLOBL ·AVX_c48<>(SB), (NOPTR+RODATA), $16
#define VPERMQ_0x39_Y1_Y1 BYTE $0xc4; BYTE $0xe3; BYTE $0xfd; BYTE $0x00; BYTE $0xc9; BYTE $0x39
#define VPERMQ_0x93_Y1_Y1 BYTE $0xc4; BYTE $0xe3; BYTE $0xfd; BYTE $0x00; BYTE $0xc9; BYTE $0x93
#define VPERMQ_0x4E_Y2_Y2 BYTE $0xc4; BYTE $0xe3; BYTE $0xfd; BYTE $0x00; BYTE $0xd2; BYTE $0x4e
#define VPERMQ_0x93_Y3_Y3 BYTE $0xc4; BYTE $0xe3; BYTE $0xfd; BYTE $0x00; BYTE $0xdb; BYTE $0x93
#define VPERMQ_0x39_Y3_Y3 BYTE $0xc4; BYTE $0xe3; BYTE $0xfd; BYTE $0x00; BYTE $0xdb; BYTE $0x39
#define ROUND_AVX2(m0, m1, m2, m3, t, c40, c48) \
VPADDQ m0, Y0, Y0; \
VPADDQ Y1, Y0, Y0; \
VPXOR Y0, Y3, Y3; \
VPSHUFD $-79, Y3, Y3; \
VPADDQ Y3, Y2, Y2; \
VPXOR Y2, Y1, Y1; \
VPSHUFB c40, Y1, Y1; \
VPADDQ m1, Y0, Y0; \
VPADDQ Y1, Y0, Y0; \
VPXOR Y0, Y3, Y3; \
VPSHUFB c48, Y3, Y3; \
VPADDQ Y3, Y2, Y2; \
VPXOR Y2, Y1, Y1; \
VPADDQ Y1, Y1, t; \
VPSRLQ $63, Y1, Y1; \
VPXOR t, Y1, Y1; \
VPERMQ_0x39_Y1_Y1; \
VPERMQ_0x4E_Y2_Y2; \
VPERMQ_0x93_Y3_Y3; \
VPADDQ m2, Y0, Y0; \
VPADDQ Y1, Y0, Y0; \
VPXOR Y0, Y3, Y3; \
VPSHUFD $-79, Y3, Y3; \
VPADDQ Y3, Y2, Y2; \
VPXOR Y2, Y1, Y1; \
VPSHUFB c40, Y1, Y1; \
VPADDQ m3, Y0, Y0; \
VPADDQ Y1, Y0, Y0; \
VPXOR Y0, Y3, Y3; \
VPSHUFB c48, Y3, Y3; \
VPADDQ Y3, Y2, Y2; \
VPXOR Y2, Y1, Y1; \
VPADDQ Y1, Y1, t; \
VPSRLQ $63, Y1, Y1; \
VPXOR t, Y1, Y1; \
VPERMQ_0x39_Y3_Y3; \
VPERMQ_0x4E_Y2_Y2; \
VPERMQ_0x93_Y1_Y1
#define VMOVQ_SI_X11_0 BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x1E
#define VMOVQ_SI_X12_0 BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x26
#define VMOVQ_SI_X13_0 BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x2E
#define VMOVQ_SI_X14_0 BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x36
#define VMOVQ_SI_X15_0 BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x3E
#define VMOVQ_SI_X11(n) BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x5E; BYTE $n
#define VMOVQ_SI_X12(n) BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x66; BYTE $n
#define VMOVQ_SI_X13(n) BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x6E; BYTE $n
#define VMOVQ_SI_X14(n) BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x76; BYTE $n
#define VMOVQ_SI_X15(n) BYTE $0xC5; BYTE $0x7A; BYTE $0x7E; BYTE $0x7E; BYTE $n
#define VPINSRQ_1_SI_X11_0 BYTE $0xC4; BYTE $0x63; BYTE $0xA1; BYTE $0x22; BYTE $0x1E; BYTE $0x01
#define VPINSRQ_1_SI_X12_0 BYTE $0xC4; BYTE $0x63; BYTE $0x99; BYTE $0x22; BYTE $0x26; BYTE $0x01
#define VPINSRQ_1_SI_X13_0 BYTE $0xC4; BYTE $0x63; BYTE $0x91; BYTE $0x22; BYTE $0x2E; BYTE $0x01
#define VPINSRQ_1_SI_X14_0 BYTE $0xC4; BYTE $0x63; BYTE $0x89; BYTE $0x22; BYTE $0x36; BYTE $0x01
#define VPINSRQ_1_SI_X15_0 BYTE $0xC4; BYTE $0x63; BYTE $0x81; BYTE $0x22; BYTE $0x3E; BYTE $0x01
#define VPINSRQ_1_SI_X11(n) BYTE $0xC4; BYTE $0x63; BYTE $0xA1; BYTE $0x22; BYTE $0x5E; BYTE $n; BYTE $0x01
#define VPINSRQ_1_SI_X12(n) BYTE $0xC4; BYTE $0x63; BYTE $0x99; BYTE $0x22; BYTE $0x66; BYTE $n; BYTE $0x01
#define VPINSRQ_1_SI_X13(n) BYTE $0xC4; BYTE $0x63; BYTE $0x91; BYTE $0x22; BYTE $0x6E; BYTE $n; BYTE $0x01
#define VPINSRQ_1_SI_X14(n) BYTE $0xC4; BYTE $0x63; BYTE $0x89; BYTE $0x22; BYTE $0x76; BYTE $n; BYTE $0x01
#define VPINSRQ_1_SI_X15(n) BYTE $0xC4; BYTE $0x63; BYTE $0x81; BYTE $0x22; BYTE $0x7E; BYTE $n; BYTE $0x01
#define VMOVQ_R8_X15 BYTE $0xC4; BYTE $0x41; BYTE $0xF9; BYTE $0x6E; BYTE $0xF8
#define VPINSRQ_1_R9_X15 BYTE $0xC4; BYTE $0x43; BYTE $0x81; BYTE $0x22; BYTE $0xF9; BYTE $0x01
// load msg: Y12 = (i0, i1, i2, i3)
// i0, i1, i2, i3 must not be 0
#define LOAD_MSG_AVX2_Y12(i0, i1, i2, i3) \
VMOVQ_SI_X12(i0*8); \
VMOVQ_SI_X11(i2*8); \
VPINSRQ_1_SI_X12(i1*8); \
VPINSRQ_1_SI_X11(i3*8); \
VINSERTI128 $1, X11, Y12, Y12
// load msg: Y13 = (i0, i1, i2, i3)
// i0, i1, i2, i3 must not be 0
#define LOAD_MSG_AVX2_Y13(i0, i1, i2, i3) \
VMOVQ_SI_X13(i0*8); \
VMOVQ_SI_X11(i2*8); \
VPINSRQ_1_SI_X13(i1*8); \
VPINSRQ_1_SI_X11(i3*8); \
VINSERTI128 $1, X11, Y13, Y13
// load msg: Y14 = (i0, i1, i2, i3)
// i0, i1, i2, i3 must not be 0
#define LOAD_MSG_AVX2_Y14(i0, i1, i2, i3) \
VMOVQ_SI_X14(i0*8); \
VMOVQ_SI_X11(i2*8); \
VPINSRQ_1_SI_X14(i1*8); \
VPINSRQ_1_SI_X11(i3*8); \
VINSERTI128 $1, X11, Y14, Y14
// load msg: Y15 = (i0, i1, i2, i3)
// i0, i1, i2, i3 must not be 0
#define LOAD_MSG_AVX2_Y15(i0, i1, i2, i3) \
VMOVQ_SI_X15(i0*8); \
VMOVQ_SI_X11(i2*8); \
VPINSRQ_1_SI_X15(i1*8); \
VPINSRQ_1_SI_X11(i3*8); \
VINSERTI128 $1, X11, Y15, Y15
#define LOAD_MSG_AVX2_0_2_4_6_1_3_5_7_8_10_12_14_9_11_13_15() \
VMOVQ_SI_X12_0; \
VMOVQ_SI_X11(4*8); \
VPINSRQ_1_SI_X12(2*8); \
VPINSRQ_1_SI_X11(6*8); \
VINSERTI128 $1, X11, Y12, Y12; \
LOAD_MSG_AVX2_Y13(1, 3, 5, 7); \
LOAD_MSG_AVX2_Y14(8, 10, 12, 14); \
LOAD_MSG_AVX2_Y15(9, 11, 13, 15)
#define LOAD_MSG_AVX2_14_4_9_13_10_8_15_6_1_0_11_5_12_2_7_3() \
LOAD_MSG_AVX2_Y12(14, 4, 9, 13); \
LOAD_MSG_AVX2_Y13(10, 8, 15, 6); \
VMOVQ_SI_X11(11*8); \
VPSHUFD $0x4E, 0*8(SI), X14; \
VPINSRQ_1_SI_X11(5*8); \
VINSERTI128 $1, X11, Y14, Y14; \
LOAD_MSG_AVX2_Y15(12, 2, 7, 3)
#define LOAD_MSG_AVX2_11_12_5_15_8_0_2_13_10_3_7_9_14_6_1_4() \
VMOVQ_SI_X11(5*8); \
VMOVDQU 11*8(SI), X12; \
VPINSRQ_1_SI_X11(15*8); \
VINSERTI128 $1, X11, Y12, Y12; \
VMOVQ_SI_X13(8*8); \
VMOVQ_SI_X11(2*8); \
VPINSRQ_1_SI_X13_0; \
VPINSRQ_1_SI_X11(13*8); \
VINSERTI128 $1, X11, Y13, Y13; \
LOAD_MSG_AVX2_Y14(10, 3, 7, 9); \
LOAD_MSG_AVX2_Y15(14, 6, 1, 4)
#define LOAD_MSG_AVX2_7_3_13_11_9_1_12_14_2_5_4_15_6_10_0_8() \
LOAD_MSG_AVX2_Y12(7, 3, 13, 11); \
LOAD_MSG_AVX2_Y13(9, 1, 12, 14); \
LOAD_MSG_AVX2_Y14(2, 5, 4, 15); \
VMOVQ_SI_X15(6*8); \
VMOVQ_SI_X11_0; \
VPINSRQ_1_SI_X15(10*8); \
VPINSRQ_1_SI_X11(8*8); \
VINSERTI128 $1, X11, Y15, Y15
#define LOAD_MSG_AVX2_9_5_2_10_0_7_4_15_14_11_6_3_1_12_8_13() \
LOAD_MSG_AVX2_Y12(9, 5, 2, 10); \
VMOVQ_SI_X13_0; \
VMOVQ_SI_X11(4*8); \
VPINSRQ_1_SI_X13(7*8); \
VPINSRQ_1_SI_X11(15*8); \
VINSERTI128 $1, X11, Y13, Y13; \
LOAD_MSG_AVX2_Y14(14, 11, 6, 3); \
LOAD_MSG_AVX2_Y15(1, 12, 8, 13)
#define LOAD_MSG_AVX2_2_6_0_8_12_10_11_3_4_7_15_1_13_5_14_9() \
VMOVQ_SI_X12(2*8); \
VMOVQ_SI_X11_0; \
VPINSRQ_1_SI_X12(6*8); \
VPINSRQ_1_SI_X11(8*8); \
VINSERTI128 $1, X11, Y12, Y12; \
LOAD_MSG_AVX2_Y13(12, 10, 11, 3); \
LOAD_MSG_AVX2_Y14(4, 7, 15, 1); \
LOAD_MSG_AVX2_Y15(13, 5, 14, 9)
#define LOAD_MSG_AVX2_12_1_14_4_5_15_13_10_0_6_9_8_7_3_2_11() \
LOAD_MSG_AVX2_Y12(12, 1, 14, 4); \
LOAD_MSG_AVX2_Y13(5, 15, 13, 10); \
VMOVQ_SI_X14_0; \
VPSHUFD $0x4E, 8*8(SI), X11; \
VPINSRQ_1_SI_X14(6*8); \
VINSERTI128 $1, X11, Y14, Y14; \
LOAD_MSG_AVX2_Y15(7, 3, 2, 11)
#define LOAD_MSG_AVX2_13_7_12_3_11_14_1_9_5_15_8_2_0_4_6_10() \
LOAD_MSG_AVX2_Y12(13, 7, 12, 3); \
LOAD_MSG_AVX2_Y13(11, 14, 1, 9); \
LOAD_MSG_AVX2_Y14(5, 15, 8, 2); \
VMOVQ_SI_X15_0; \
VMOVQ_SI_X11(6*8); \
VPINSRQ_1_SI_X15(4*8); \
VPINSRQ_1_SI_X11(10*8); \
VINSERTI128 $1, X11, Y15, Y15
#define LOAD_MSG_AVX2_6_14_11_0_15_9_3_8_12_13_1_10_2_7_4_5() \
VMOVQ_SI_X12(6*8); \
VMOVQ_SI_X11(11*8); \
VPINSRQ_1_SI_X12(14*8); \
VPINSRQ_1_SI_X11_0; \
VINSERTI128 $1, X11, Y12, Y12; \
LOAD_MSG_AVX2_Y13(15, 9, 3, 8); \
VMOVQ_SI_X11(1*8); \
VMOVDQU 12*8(SI), X14; \
VPINSRQ_1_SI_X11(10*8); \
VINSERTI128 $1, X11, Y14, Y14; \
VMOVQ_SI_X15(2*8); \
VMOVDQU 4*8(SI), X11; \
VPINSRQ_1_SI_X15(7*8); \
VINSERTI128 $1, X11, Y15, Y15
#define LOAD_MSG_AVX2_10_8_7_1_2_4_6_5_15_9_3_13_11_14_12_0() \
LOAD_MSG_AVX2_Y12(10, 8, 7, 1); \
VMOVQ_SI_X13(2*8); \
VPSHUFD $0x4E, 5*8(SI), X11; \
VPINSRQ_1_SI_X13(4*8); \
VINSERTI128 $1, X11, Y13, Y13; \
LOAD_MSG_AVX2_Y14(15, 9, 3, 13); \
VMOVQ_SI_X15(11*8); \
VMOVQ_SI_X11(12*8); \
VPINSRQ_1_SI_X15(14*8); \
VPINSRQ_1_SI_X11_0; \
VINSERTI128 $1, X11, Y15, Y15
// func hashBlocksAVX2(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte)
TEXT ·hashBlocksAVX2(SB), 4, $320-48 // frame size = 288 + 32 byte alignment
MOVQ h+0(FP), AX
MOVQ c+8(FP), BX
MOVQ flag+16(FP), CX
MOVQ blocks_base+24(FP), SI
MOVQ blocks_len+32(FP), DI
MOVQ SP, DX
ADDQ $31, DX
ANDQ $~31, DX
MOVQ CX, 16(DX)
XORQ CX, CX
MOVQ CX, 24(DX)
VMOVDQU ·AVX2_c40<>(SB), Y4
VMOVDQU ·AVX2_c48<>(SB), Y5
VMOVDQU 0(AX), Y8
VMOVDQU 32(AX), Y9
VMOVDQU ·AVX2_iv0<>(SB), Y6
VMOVDQU ·AVX2_iv1<>(SB), Y7
MOVQ 0(BX), R8
MOVQ 8(BX), R9
MOVQ R9, 8(DX)
loop:
ADDQ $128, R8
MOVQ R8, 0(DX)
CMPQ R8, $128
JGE noinc
INCQ R9
MOVQ R9, 8(DX)
noinc:
VMOVDQA Y8, Y0
VMOVDQA Y9, Y1
VMOVDQA Y6, Y2
VPXOR 0(DX), Y7, Y3
LOAD_MSG_AVX2_0_2_4_6_1_3_5_7_8_10_12_14_9_11_13_15()
VMOVDQA Y12, 32(DX)
VMOVDQA Y13, 64(DX)
VMOVDQA Y14, 96(DX)
VMOVDQA Y15, 128(DX)
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_14_4_9_13_10_8_15_6_1_0_11_5_12_2_7_3()
VMOVDQA Y12, 160(DX)
VMOVDQA Y13, 192(DX)
VMOVDQA Y14, 224(DX)
VMOVDQA Y15, 256(DX)
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_11_12_5_15_8_0_2_13_10_3_7_9_14_6_1_4()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_7_3_13_11_9_1_12_14_2_5_4_15_6_10_0_8()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_9_5_2_10_0_7_4_15_14_11_6_3_1_12_8_13()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_2_6_0_8_12_10_11_3_4_7_15_1_13_5_14_9()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_12_1_14_4_5_15_13_10_0_6_9_8_7_3_2_11()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_13_7_12_3_11_14_1_9_5_15_8_2_0_4_6_10()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_6_14_11_0_15_9_3_8_12_13_1_10_2_7_4_5()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
LOAD_MSG_AVX2_10_8_7_1_2_4_6_5_15_9_3_13_11_14_12_0()
ROUND_AVX2(Y12, Y13, Y14, Y15, Y10, Y4, Y5)
ROUND_AVX2(32(DX), 64(DX), 96(DX), 128(DX), Y10, Y4, Y5)
ROUND_AVX2(160(DX), 192(DX), 224(DX), 256(DX), Y10, Y4, Y5)
VPXOR Y0, Y8, Y8
VPXOR Y1, Y9, Y9
VPXOR Y2, Y8, Y8
VPXOR Y3, Y9, Y9
LEAQ 128(SI), SI
SUBQ $128, DI
JNE loop
MOVQ R8, 0(BX)
MOVQ R9, 8(BX)
VMOVDQU Y8, 0(AX)
VMOVDQU Y9, 32(AX)
VZEROUPPER
RET
#define VPUNPCKLQDQ_X2_X2_X15 BYTE $0xC5; BYTE $0x69; BYTE $0x6C; BYTE $0xFA
#define VPUNPCKLQDQ_X3_X3_X15 BYTE $0xC5; BYTE $0x61; BYTE $0x6C; BYTE $0xFB
#define VPUNPCKLQDQ_X7_X7_X15 BYTE $0xC5; BYTE $0x41; BYTE $0x6C; BYTE $0xFF
#define VPUNPCKLQDQ_X13_X13_X15 BYTE $0xC4; BYTE $0x41; BYTE $0x11; BYTE $0x6C; BYTE $0xFD
#define VPUNPCKLQDQ_X14_X14_X15 BYTE $0xC4; BYTE $0x41; BYTE $0x09; BYTE $0x6C; BYTE $0xFE
#define VPUNPCKHQDQ_X15_X2_X2 BYTE $0xC4; BYTE $0xC1; BYTE $0x69; BYTE $0x6D; BYTE $0xD7
#define VPUNPCKHQDQ_X15_X3_X3 BYTE $0xC4; BYTE $0xC1; BYTE $0x61; BYTE $0x6D; BYTE $0xDF
#define VPUNPCKHQDQ_X15_X6_X6 BYTE $0xC4; BYTE $0xC1; BYTE $0x49; BYTE $0x6D; BYTE $0xF7
#define VPUNPCKHQDQ_X15_X7_X7 BYTE $0xC4; BYTE $0xC1; BYTE $0x41; BYTE $0x6D; BYTE $0xFF
#define VPUNPCKHQDQ_X15_X3_X2 BYTE $0xC4; BYTE $0xC1; BYTE $0x61; BYTE $0x6D; BYTE $0xD7
#define VPUNPCKHQDQ_X15_X7_X6 BYTE $0xC4; BYTE $0xC1; BYTE $0x41; BYTE $0x6D; BYTE $0xF7
#define VPUNPCKHQDQ_X15_X13_X3 BYTE $0xC4; BYTE $0xC1; BYTE $0x11; BYTE $0x6D; BYTE $0xDF
#define VPUNPCKHQDQ_X15_X13_X7 BYTE $0xC4; BYTE $0xC1; BYTE $0x11; BYTE $0x6D; BYTE $0xFF
#define SHUFFLE_AVX() \
VMOVDQA X6, X13; \
VMOVDQA X2, X14; \
VMOVDQA X4, X6; \
VPUNPCKLQDQ_X13_X13_X15; \
VMOVDQA X5, X4; \
VMOVDQA X6, X5; \
VPUNPCKHQDQ_X15_X7_X6; \
VPUNPCKLQDQ_X7_X7_X15; \
VPUNPCKHQDQ_X15_X13_X7; \
VPUNPCKLQDQ_X3_X3_X15; \
VPUNPCKHQDQ_X15_X2_X2; \
VPUNPCKLQDQ_X14_X14_X15; \
VPUNPCKHQDQ_X15_X3_X3; \
#define SHUFFLE_AVX_INV() \
VMOVDQA X2, X13; \
VMOVDQA X4, X14; \
VPUNPCKLQDQ_X2_X2_X15; \
VMOVDQA X5, X4; \
VPUNPCKHQDQ_X15_X3_X2; \
VMOVDQA X14, X5; \
VPUNPCKLQDQ_X3_X3_X15; \
VMOVDQA X6, X14; \
VPUNPCKHQDQ_X15_X13_X3; \
VPUNPCKLQDQ_X7_X7_X15; \
VPUNPCKHQDQ_X15_X6_X6; \
VPUNPCKLQDQ_X14_X14_X15; \
VPUNPCKHQDQ_X15_X7_X7; \
#define HALF_ROUND_AVX(v0, v1, v2, v3, v4, v5, v6, v7, m0, m1, m2, m3, t0, c40, c48) \
VPADDQ m0, v0, v0; \
VPADDQ v2, v0, v0; \
VPADDQ m1, v1, v1; \
VPADDQ v3, v1, v1; \
VPXOR v0, v6, v6; \
VPXOR v1, v7, v7; \
VPSHUFD $-79, v6, v6; \
VPSHUFD $-79, v7, v7; \
VPADDQ v6, v4, v4; \
VPADDQ v7, v5, v5; \
VPXOR v4, v2, v2; \
VPXOR v5, v3, v3; \
VPSHUFB c40, v2, v2; \
VPSHUFB c40, v3, v3; \
VPADDQ m2, v0, v0; \
VPADDQ v2, v0, v0; \
VPADDQ m3, v1, v1; \
VPADDQ v3, v1, v1; \
VPXOR v0, v6, v6; \
VPXOR v1, v7, v7; \
VPSHUFB c48, v6, v6; \
VPSHUFB c48, v7, v7; \
VPADDQ v6, v4, v4; \
VPADDQ v7, v5, v5; \
VPXOR v4, v2, v2; \
VPXOR v5, v3, v3; \
VPADDQ v2, v2, t0; \
VPSRLQ $63, v2, v2; \
VPXOR t0, v2, v2; \
VPADDQ v3, v3, t0; \
VPSRLQ $63, v3, v3; \
VPXOR t0, v3, v3
// load msg: X12 = (i0, i1), X13 = (i2, i3), X14 = (i4, i5), X15 = (i6, i7)
// i0, i1, i2, i3, i4, i5, i6, i7 must not be 0
#define LOAD_MSG_AVX(i0, i1, i2, i3, i4, i5, i6, i7) \
VMOVQ_SI_X12(i0*8); \
VMOVQ_SI_X13(i2*8); \
VMOVQ_SI_X14(i4*8); \
VMOVQ_SI_X15(i6*8); \
VPINSRQ_1_SI_X12(i1*8); \
VPINSRQ_1_SI_X13(i3*8); \
VPINSRQ_1_SI_X14(i5*8); \
VPINSRQ_1_SI_X15(i7*8)
// load msg: X12 = (0, 2), X13 = (4, 6), X14 = (1, 3), X15 = (5, 7)
#define LOAD_MSG_AVX_0_2_4_6_1_3_5_7() \
VMOVQ_SI_X12_0; \
VMOVQ_SI_X13(4*8); \
VMOVQ_SI_X14(1*8); \
VMOVQ_SI_X15(5*8); \
VPINSRQ_1_SI_X12(2*8); \
VPINSRQ_1_SI_X13(6*8); \
VPINSRQ_1_SI_X14(3*8); \
VPINSRQ_1_SI_X15(7*8)
// load msg: X12 = (1, 0), X13 = (11, 5), X14 = (12, 2), X15 = (7, 3)
#define LOAD_MSG_AVX_1_0_11_5_12_2_7_3() \
VPSHUFD $0x4E, 0*8(SI), X12; \
VMOVQ_SI_X13(11*8); \
VMOVQ_SI_X14(12*8); \
VMOVQ_SI_X15(7*8); \
VPINSRQ_1_SI_X13(5*8); \
VPINSRQ_1_SI_X14(2*8); \
VPINSRQ_1_SI_X15(3*8)
// load msg: X12 = (11, 12), X13 = (5, 15), X14 = (8, 0), X15 = (2, 13)
#define LOAD_MSG_AVX_11_12_5_15_8_0_2_13() \
VMOVDQU 11*8(SI), X12; \
VMOVQ_SI_X13(5*8); \
VMOVQ_SI_X14(8*8); \
VMOVQ_SI_X15(2*8); \
VPINSRQ_1_SI_X13(15*8); \
VPINSRQ_1_SI_X14_0; \
VPINSRQ_1_SI_X15(13*8)
// load msg: X12 = (2, 5), X13 = (4, 15), X14 = (6, 10), X15 = (0, 8)
#define LOAD_MSG_AVX_2_5_4_15_6_10_0_8() \
VMOVQ_SI_X12(2*8); \
VMOVQ_SI_X13(4*8); \
VMOVQ_SI_X14(6*8); \
VMOVQ_SI_X15_0; \
VPINSRQ_1_SI_X12(5*8); \
VPINSRQ_1_SI_X13(15*8); \
VPINSRQ_1_SI_X14(10*8); \
VPINSRQ_1_SI_X15(8*8)
// load msg: X12 = (9, 5), X13 = (2, 10), X14 = (0, 7), X15 = (4, 15)
#define LOAD_MSG_AVX_9_5_2_10_0_7_4_15() \
VMOVQ_SI_X12(9*8); \
VMOVQ_SI_X13(2*8); \
VMOVQ_SI_X14_0; \
VMOVQ_SI_X15(4*8); \
VPINSRQ_1_SI_X12(5*8); \
VPINSRQ_1_SI_X13(10*8); \
VPINSRQ_1_SI_X14(7*8); \
VPINSRQ_1_SI_X15(15*8)
// load msg: X12 = (2, 6), X13 = (0, 8), X14 = (12, 10), X15 = (11, 3)
#define LOAD_MSG_AVX_2_6_0_8_12_10_11_3() \
VMOVQ_SI_X12(2*8); \
VMOVQ_SI_X13_0; \
VMOVQ_SI_X14(12*8); \
VMOVQ_SI_X15(11*8); \
VPINSRQ_1_SI_X12(6*8); \
VPINSRQ_1_SI_X13(8*8); \
VPINSRQ_1_SI_X14(10*8); \
VPINSRQ_1_SI_X15(3*8)
// load msg: X12 = (0, 6), X13 = (9, 8), X14 = (7, 3), X15 = (2, 11)
#define LOAD_MSG_AVX_0_6_9_8_7_3_2_11() \
MOVQ 0*8(SI), X12; \
VPSHUFD $0x4E, 8*8(SI), X13; \
MOVQ 7*8(SI), X14; \
MOVQ 2*8(SI), X15; \
VPINSRQ_1_SI_X12(6*8); \
VPINSRQ_1_SI_X14(3*8); \
VPINSRQ_1_SI_X15(11*8)
// load msg: X12 = (6, 14), X13 = (11, 0), X14 = (15, 9), X15 = (3, 8)
#define LOAD_MSG_AVX_6_14_11_0_15_9_3_8() \
MOVQ 6*8(SI), X12; \
MOVQ 11*8(SI), X13; \
MOVQ 15*8(SI), X14; \
MOVQ 3*8(SI), X15; \
VPINSRQ_1_SI_X12(14*8); \
VPINSRQ_1_SI_X13_0; \
VPINSRQ_1_SI_X14(9*8); \
VPINSRQ_1_SI_X15(8*8)
// load msg: X12 = (5, 15), X13 = (8, 2), X14 = (0, 4), X15 = (6, 10)
#define LOAD_MSG_AVX_5_15_8_2_0_4_6_10() \
MOVQ 5*8(SI), X12; \
MOVQ 8*8(SI), X13; \
MOVQ 0*8(SI), X14; \
MOVQ 6*8(SI), X15; \
VPINSRQ_1_SI_X12(15*8); \
VPINSRQ_1_SI_X13(2*8); \
VPINSRQ_1_SI_X14(4*8); \
VPINSRQ_1_SI_X15(10*8)
// load msg: X12 = (12, 13), X13 = (1, 10), X14 = (2, 7), X15 = (4, 5)
#define LOAD_MSG_AVX_12_13_1_10_2_7_4_5() \
VMOVDQU 12*8(SI), X12; \
MOVQ 1*8(SI), X13; \
MOVQ 2*8(SI), X14; \
VPINSRQ_1_SI_X13(10*8); \
VPINSRQ_1_SI_X14(7*8); \
VMOVDQU 4*8(SI), X15
// load msg: X12 = (15, 9), X13 = (3, 13), X14 = (11, 14), X15 = (12, 0)
#define LOAD_MSG_AVX_15_9_3_13_11_14_12_0() \
MOVQ 15*8(SI), X12; \
MOVQ 3*8(SI), X13; \
MOVQ 11*8(SI), X14; \
MOVQ 12*8(SI), X15; \
VPINSRQ_1_SI_X12(9*8); \
VPINSRQ_1_SI_X13(13*8); \
VPINSRQ_1_SI_X14(14*8); \
VPINSRQ_1_SI_X15_0
// func hashBlocksAVX(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte)
TEXT ·hashBlocksAVX(SB), 4, $288-48 // frame size = 272 + 16 byte alignment
MOVQ h+0(FP), AX
MOVQ c+8(FP), BX
MOVQ flag+16(FP), CX
MOVQ blocks_base+24(FP), SI
MOVQ blocks_len+32(FP), DI
MOVQ SP, R10
ADDQ $15, R10
ANDQ $~15, R10
VMOVDQU ·AVX_c40<>(SB), X0
VMOVDQU ·AVX_c48<>(SB), X1
VMOVDQA X0, X8
VMOVDQA X1, X9
VMOVDQU ·AVX_iv3<>(SB), X0
VMOVDQA X0, 0(R10)
XORQ CX, 0(R10) // 0(R10) = ·AVX_iv3 ^ (CX || 0)
VMOVDQU 0(AX), X10
VMOVDQU 16(AX), X11
VMOVDQU 32(AX), X2
VMOVDQU 48(AX), X3
MOVQ 0(BX), R8
MOVQ 8(BX), R9
loop:
ADDQ $128, R8
CMPQ R8, $128
JGE noinc
INCQ R9
noinc:
VMOVQ_R8_X15
VPINSRQ_1_R9_X15
VMOVDQA X10, X0
VMOVDQA X11, X1
VMOVDQU ·AVX_iv0<>(SB), X4
VMOVDQU ·AVX_iv1<>(SB), X5
VMOVDQU ·AVX_iv2<>(SB), X6
VPXOR X15, X6, X6
VMOVDQA 0(R10), X7
LOAD_MSG_AVX_0_2_4_6_1_3_5_7()
VMOVDQA X12, 16(R10)
VMOVDQA X13, 32(R10)
VMOVDQA X14, 48(R10)
VMOVDQA X15, 64(R10)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX(8, 10, 12, 14, 9, 11, 13, 15)
VMOVDQA X12, 80(R10)
VMOVDQA X13, 96(R10)
VMOVDQA X14, 112(R10)
VMOVDQA X15, 128(R10)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX(14, 4, 9, 13, 10, 8, 15, 6)
VMOVDQA X12, 144(R10)
VMOVDQA X13, 160(R10)
VMOVDQA X14, 176(R10)
VMOVDQA X15, 192(R10)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX_1_0_11_5_12_2_7_3()
VMOVDQA X12, 208(R10)
VMOVDQA X13, 224(R10)
VMOVDQA X14, 240(R10)
VMOVDQA X15, 256(R10)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX_11_12_5_15_8_0_2_13()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX(10, 3, 7, 9, 14, 6, 1, 4)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX(7, 3, 13, 11, 9, 1, 12, 14)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX_2_5_4_15_6_10_0_8()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX_9_5_2_10_0_7_4_15()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX(14, 11, 6, 3, 1, 12, 8, 13)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX_2_6_0_8_12_10_11_3()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX(4, 7, 15, 1, 13, 5, 14, 9)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX(12, 1, 14, 4, 5, 15, 13, 10)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX_0_6_9_8_7_3_2_11()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX(13, 7, 12, 3, 11, 14, 1, 9)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX_5_15_8_2_0_4_6_10()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX_6_14_11_0_15_9_3_8()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX_12_13_1_10_2_7_4_5()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
LOAD_MSG_AVX(10, 8, 7, 1, 2, 4, 6, 5)
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX()
LOAD_MSG_AVX_15_9_3_13_11_14_12_0()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, X12, X13, X14, X15, X15, X8, X9)
SHUFFLE_AVX_INV()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, 16(R10), 32(R10), 48(R10), 64(R10), X15, X8, X9)
SHUFFLE_AVX()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, 80(R10), 96(R10), 112(R10), 128(R10), X15, X8, X9)
SHUFFLE_AVX_INV()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, 144(R10), 160(R10), 176(R10), 192(R10), X15, X8, X9)
SHUFFLE_AVX()
HALF_ROUND_AVX(X0, X1, X2, X3, X4, X5, X6, X7, 208(R10), 224(R10), 240(R10), 256(R10), X15, X8, X9)
SHUFFLE_AVX_INV()
VMOVDQU 32(AX), X14
VMOVDQU 48(AX), X15
VPXOR X0, X10, X10
VPXOR X1, X11, X11
VPXOR X2, X14, X14
VPXOR X3, X15, X15
VPXOR X4, X10, X10
VPXOR X5, X11, X11
VPXOR X6, X14, X2
VPXOR X7, X15, X3
VMOVDQU X2, 32(AX)
VMOVDQU X3, 48(AX)
LEAQ 128(SI), SI
SUBQ $128, DI
JNE loop
VMOVDQU X10, 0(AX)
VMOVDQU X11, 16(AX)
MOVQ R8, 0(BX)
MOVQ R9, 8(BX)
VZEROUPPER
RET

25
vendor/golang.org/x/crypto/blake2b/blake2b_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,25 @@
// 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.
//go:build !go1.7 && amd64 && gc && !purego
// +build !go1.7,amd64,gc,!purego
package blake2b
import "golang.org/x/sys/cpu"
func init() {
useSSE4 = cpu.X86.HasSSE41
}
//go:noescape
func hashBlocksSSE4(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte)
func hashBlocks(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte) {
if useSSE4 {
hashBlocksSSE4(h, c, flag, blocks)
} else {
hashBlocksGeneric(h, c, flag, blocks)
}
}

279
vendor/golang.org/x/crypto/blake2b/blake2b_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,279 @@
// 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.
//go:build amd64 && gc && !purego
// +build amd64,gc,!purego
#include "textflag.h"
DATA ·iv0<>+0x00(SB)/8, $0x6a09e667f3bcc908
DATA ·iv0<>+0x08(SB)/8, $0xbb67ae8584caa73b
GLOBL ·iv0<>(SB), (NOPTR+RODATA), $16
DATA ·iv1<>+0x00(SB)/8, $0x3c6ef372fe94f82b
DATA ·iv1<>+0x08(SB)/8, $0xa54ff53a5f1d36f1
GLOBL ·iv1<>(SB), (NOPTR+RODATA), $16
DATA ·iv2<>+0x00(SB)/8, $0x510e527fade682d1
DATA ·iv2<>+0x08(SB)/8, $0x9b05688c2b3e6c1f
GLOBL ·iv2<>(SB), (NOPTR+RODATA), $16
DATA ·iv3<>+0x00(SB)/8, $0x1f83d9abfb41bd6b
DATA ·iv3<>+0x08(SB)/8, $0x5be0cd19137e2179
GLOBL ·iv3<>(SB), (NOPTR+RODATA), $16
DATA ·c40<>+0x00(SB)/8, $0x0201000706050403
DATA ·c40<>+0x08(SB)/8, $0x0a09080f0e0d0c0b
GLOBL ·c40<>(SB), (NOPTR+RODATA), $16
DATA ·c48<>+0x00(SB)/8, $0x0100070605040302
DATA ·c48<>+0x08(SB)/8, $0x09080f0e0d0c0b0a
GLOBL ·c48<>(SB), (NOPTR+RODATA), $16
#define SHUFFLE(v2, v3, v4, v5, v6, v7, t1, t2) \
MOVO v4, t1; \
MOVO v5, v4; \
MOVO t1, v5; \
MOVO v6, t1; \
PUNPCKLQDQ v6, t2; \
PUNPCKHQDQ v7, v6; \
PUNPCKHQDQ t2, v6; \
PUNPCKLQDQ v7, t2; \
MOVO t1, v7; \
MOVO v2, t1; \
PUNPCKHQDQ t2, v7; \
PUNPCKLQDQ v3, t2; \
PUNPCKHQDQ t2, v2; \
PUNPCKLQDQ t1, t2; \
PUNPCKHQDQ t2, v3
#define SHUFFLE_INV(v2, v3, v4, v5, v6, v7, t1, t2) \
MOVO v4, t1; \
MOVO v5, v4; \
MOVO t1, v5; \
MOVO v2, t1; \
PUNPCKLQDQ v2, t2; \
PUNPCKHQDQ v3, v2; \
PUNPCKHQDQ t2, v2; \
PUNPCKLQDQ v3, t2; \
MOVO t1, v3; \
MOVO v6, t1; \
PUNPCKHQDQ t2, v3; \
PUNPCKLQDQ v7, t2; \
PUNPCKHQDQ t2, v6; \
PUNPCKLQDQ t1, t2; \
PUNPCKHQDQ t2, v7
#define HALF_ROUND(v0, v1, v2, v3, v4, v5, v6, v7, m0, m1, m2, m3, t0, c40, c48) \
PADDQ m0, v0; \
PADDQ m1, v1; \
PADDQ v2, v0; \
PADDQ v3, v1; \
PXOR v0, v6; \
PXOR v1, v7; \
PSHUFD $0xB1, v6, v6; \
PSHUFD $0xB1, v7, v7; \
PADDQ v6, v4; \
PADDQ v7, v5; \
PXOR v4, v2; \
PXOR v5, v3; \
PSHUFB c40, v2; \
PSHUFB c40, v3; \
PADDQ m2, v0; \
PADDQ m3, v1; \
PADDQ v2, v0; \
PADDQ v3, v1; \
PXOR v0, v6; \
PXOR v1, v7; \
PSHUFB c48, v6; \
PSHUFB c48, v7; \
PADDQ v6, v4; \
PADDQ v7, v5; \
PXOR v4, v2; \
PXOR v5, v3; \
MOVOU v2, t0; \
PADDQ v2, t0; \
PSRLQ $63, v2; \
PXOR t0, v2; \
MOVOU v3, t0; \
PADDQ v3, t0; \
PSRLQ $63, v3; \
PXOR t0, v3
#define LOAD_MSG(m0, m1, m2, m3, src, i0, i1, i2, i3, i4, i5, i6, i7) \
MOVQ i0*8(src), m0; \
PINSRQ $1, i1*8(src), m0; \
MOVQ i2*8(src), m1; \
PINSRQ $1, i3*8(src), m1; \
MOVQ i4*8(src), m2; \
PINSRQ $1, i5*8(src), m2; \
MOVQ i6*8(src), m3; \
PINSRQ $1, i7*8(src), m3
// func hashBlocksSSE4(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte)
TEXT ·hashBlocksSSE4(SB), 4, $288-48 // frame size = 272 + 16 byte alignment
MOVQ h+0(FP), AX
MOVQ c+8(FP), BX
MOVQ flag+16(FP), CX
MOVQ blocks_base+24(FP), SI
MOVQ blocks_len+32(FP), DI
MOVQ SP, R10
ADDQ $15, R10
ANDQ $~15, R10
MOVOU ·iv3<>(SB), X0
MOVO X0, 0(R10)
XORQ CX, 0(R10) // 0(R10) = ·iv3 ^ (CX || 0)
MOVOU ·c40<>(SB), X13
MOVOU ·c48<>(SB), X14
MOVOU 0(AX), X12
MOVOU 16(AX), X15
MOVQ 0(BX), R8
MOVQ 8(BX), R9
loop:
ADDQ $128, R8
CMPQ R8, $128
JGE noinc
INCQ R9
noinc:
MOVQ R8, X8
PINSRQ $1, R9, X8
MOVO X12, X0
MOVO X15, X1
MOVOU 32(AX), X2
MOVOU 48(AX), X3
MOVOU ·iv0<>(SB), X4
MOVOU ·iv1<>(SB), X5
MOVOU ·iv2<>(SB), X6
PXOR X8, X6
MOVO 0(R10), X7
LOAD_MSG(X8, X9, X10, X11, SI, 0, 2, 4, 6, 1, 3, 5, 7)
MOVO X8, 16(R10)
MOVO X9, 32(R10)
MOVO X10, 48(R10)
MOVO X11, 64(R10)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 8, 10, 12, 14, 9, 11, 13, 15)
MOVO X8, 80(R10)
MOVO X9, 96(R10)
MOVO X10, 112(R10)
MOVO X11, 128(R10)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 14, 4, 9, 13, 10, 8, 15, 6)
MOVO X8, 144(R10)
MOVO X9, 160(R10)
MOVO X10, 176(R10)
MOVO X11, 192(R10)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 1, 0, 11, 5, 12, 2, 7, 3)
MOVO X8, 208(R10)
MOVO X9, 224(R10)
MOVO X10, 240(R10)
MOVO X11, 256(R10)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 11, 12, 5, 15, 8, 0, 2, 13)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 10, 3, 7, 9, 14, 6, 1, 4)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 7, 3, 13, 11, 9, 1, 12, 14)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 2, 5, 4, 15, 6, 10, 0, 8)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 9, 5, 2, 10, 0, 7, 4, 15)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 14, 11, 6, 3, 1, 12, 8, 13)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 2, 6, 0, 8, 12, 10, 11, 3)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 4, 7, 15, 1, 13, 5, 14, 9)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 12, 1, 14, 4, 5, 15, 13, 10)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 0, 6, 9, 8, 7, 3, 2, 11)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 13, 7, 12, 3, 11, 14, 1, 9)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 5, 15, 8, 2, 0, 4, 6, 10)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 6, 14, 11, 0, 15, 9, 3, 8)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 12, 13, 1, 10, 2, 7, 4, 5)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 10, 8, 7, 1, 2, 4, 6, 5)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
LOAD_MSG(X8, X9, X10, X11, SI, 15, 9, 3, 13, 11, 14, 12, 0)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, X8, X9, X10, X11, X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, 16(R10), 32(R10), 48(R10), 64(R10), X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, 80(R10), 96(R10), 112(R10), 128(R10), X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, 144(R10), 160(R10), 176(R10), 192(R10), X11, X13, X14)
SHUFFLE(X2, X3, X4, X5, X6, X7, X8, X9)
HALF_ROUND(X0, X1, X2, X3, X4, X5, X6, X7, 208(R10), 224(R10), 240(R10), 256(R10), X11, X13, X14)
SHUFFLE_INV(X2, X3, X4, X5, X6, X7, X8, X9)
MOVOU 32(AX), X10
MOVOU 48(AX), X11
PXOR X0, X12
PXOR X1, X15
PXOR X2, X10
PXOR X3, X11
PXOR X4, X12
PXOR X5, X15
PXOR X6, X10
PXOR X7, X11
MOVOU X10, 32(AX)
MOVOU X11, 48(AX)
LEAQ 128(SI), SI
SUBQ $128, DI
JNE loop
MOVOU X12, 0(AX)
MOVOU X15, 16(AX)
MOVQ R8, 0(BX)
MOVQ R9, 8(BX)
RET

182
vendor/golang.org/x/crypto/blake2b/blake2b_generic.go generated vendored Normal file
View File

@@ -0,0 +1,182 @@
// 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 blake2b
import (
"encoding/binary"
"math/bits"
)
// the precomputed values for BLAKE2b
// there are 12 16-byte arrays - one for each round
// the entries are calculated from the sigma constants.
var precomputed = [12][16]byte{
{0, 2, 4, 6, 1, 3, 5, 7, 8, 10, 12, 14, 9, 11, 13, 15},
{14, 4, 9, 13, 10, 8, 15, 6, 1, 0, 11, 5, 12, 2, 7, 3},
{11, 12, 5, 15, 8, 0, 2, 13, 10, 3, 7, 9, 14, 6, 1, 4},
{7, 3, 13, 11, 9, 1, 12, 14, 2, 5, 4, 15, 6, 10, 0, 8},
{9, 5, 2, 10, 0, 7, 4, 15, 14, 11, 6, 3, 1, 12, 8, 13},
{2, 6, 0, 8, 12, 10, 11, 3, 4, 7, 15, 1, 13, 5, 14, 9},
{12, 1, 14, 4, 5, 15, 13, 10, 0, 6, 9, 8, 7, 3, 2, 11},
{13, 7, 12, 3, 11, 14, 1, 9, 5, 15, 8, 2, 0, 4, 6, 10},
{6, 14, 11, 0, 15, 9, 3, 8, 12, 13, 1, 10, 2, 7, 4, 5},
{10, 8, 7, 1, 2, 4, 6, 5, 15, 9, 3, 13, 11, 14, 12, 0},
{0, 2, 4, 6, 1, 3, 5, 7, 8, 10, 12, 14, 9, 11, 13, 15}, // equal to the first
{14, 4, 9, 13, 10, 8, 15, 6, 1, 0, 11, 5, 12, 2, 7, 3}, // equal to the second
}
func hashBlocksGeneric(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte) {
var m [16]uint64
c0, c1 := c[0], c[1]
for i := 0; i < len(blocks); {
c0 += BlockSize
if c0 < BlockSize {
c1++
}
v0, v1, v2, v3, v4, v5, v6, v7 := h[0], h[1], h[2], h[3], h[4], h[5], h[6], h[7]
v8, v9, v10, v11, v12, v13, v14, v15 := iv[0], iv[1], iv[2], iv[3], iv[4], iv[5], iv[6], iv[7]
v12 ^= c0
v13 ^= c1
v14 ^= flag
for j := range m {
m[j] = binary.LittleEndian.Uint64(blocks[i:])
i += 8
}
for j := range precomputed {
s := &(precomputed[j])
v0 += m[s[0]]
v0 += v4
v12 ^= v0
v12 = bits.RotateLeft64(v12, -32)
v8 += v12
v4 ^= v8
v4 = bits.RotateLeft64(v4, -24)
v1 += m[s[1]]
v1 += v5
v13 ^= v1
v13 = bits.RotateLeft64(v13, -32)
v9 += v13
v5 ^= v9
v5 = bits.RotateLeft64(v5, -24)
v2 += m[s[2]]
v2 += v6
v14 ^= v2
v14 = bits.RotateLeft64(v14, -32)
v10 += v14
v6 ^= v10
v6 = bits.RotateLeft64(v6, -24)
v3 += m[s[3]]
v3 += v7
v15 ^= v3
v15 = bits.RotateLeft64(v15, -32)
v11 += v15
v7 ^= v11
v7 = bits.RotateLeft64(v7, -24)
v0 += m[s[4]]
v0 += v4
v12 ^= v0
v12 = bits.RotateLeft64(v12, -16)
v8 += v12
v4 ^= v8
v4 = bits.RotateLeft64(v4, -63)
v1 += m[s[5]]
v1 += v5
v13 ^= v1
v13 = bits.RotateLeft64(v13, -16)
v9 += v13
v5 ^= v9
v5 = bits.RotateLeft64(v5, -63)
v2 += m[s[6]]
v2 += v6
v14 ^= v2
v14 = bits.RotateLeft64(v14, -16)
v10 += v14
v6 ^= v10
v6 = bits.RotateLeft64(v6, -63)
v3 += m[s[7]]
v3 += v7
v15 ^= v3
v15 = bits.RotateLeft64(v15, -16)
v11 += v15
v7 ^= v11
v7 = bits.RotateLeft64(v7, -63)
v0 += m[s[8]]
v0 += v5
v15 ^= v0
v15 = bits.RotateLeft64(v15, -32)
v10 += v15
v5 ^= v10
v5 = bits.RotateLeft64(v5, -24)
v1 += m[s[9]]
v1 += v6
v12 ^= v1
v12 = bits.RotateLeft64(v12, -32)
v11 += v12
v6 ^= v11
v6 = bits.RotateLeft64(v6, -24)
v2 += m[s[10]]
v2 += v7
v13 ^= v2
v13 = bits.RotateLeft64(v13, -32)
v8 += v13
v7 ^= v8
v7 = bits.RotateLeft64(v7, -24)
v3 += m[s[11]]
v3 += v4
v14 ^= v3
v14 = bits.RotateLeft64(v14, -32)
v9 += v14
v4 ^= v9
v4 = bits.RotateLeft64(v4, -24)
v0 += m[s[12]]
v0 += v5
v15 ^= v0
v15 = bits.RotateLeft64(v15, -16)
v10 += v15
v5 ^= v10
v5 = bits.RotateLeft64(v5, -63)
v1 += m[s[13]]
v1 += v6
v12 ^= v1
v12 = bits.RotateLeft64(v12, -16)
v11 += v12
v6 ^= v11
v6 = bits.RotateLeft64(v6, -63)
v2 += m[s[14]]
v2 += v7
v13 ^= v2
v13 = bits.RotateLeft64(v13, -16)
v8 += v13
v7 ^= v8
v7 = bits.RotateLeft64(v7, -63)
v3 += m[s[15]]
v3 += v4
v14 ^= v3
v14 = bits.RotateLeft64(v14, -16)
v9 += v14
v4 ^= v9
v4 = bits.RotateLeft64(v4, -63)
}
h[0] ^= v0 ^ v8
h[1] ^= v1 ^ v9
h[2] ^= v2 ^ v10
h[3] ^= v3 ^ v11
h[4] ^= v4 ^ v12
h[5] ^= v5 ^ v13
h[6] ^= v6 ^ v14
h[7] ^= v7 ^ v15
}
c[0], c[1] = c0, c1
}

12
vendor/golang.org/x/crypto/blake2b/blake2b_ref.go generated vendored Normal file
View File

@@ -0,0 +1,12 @@
// 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.
//go:build !amd64 || purego || !gc
// +build !amd64 purego !gc
package blake2b
func hashBlocks(h *[8]uint64, c *[2]uint64, flag uint64, blocks []byte) {
hashBlocksGeneric(h, c, flag, blocks)
}

177
vendor/golang.org/x/crypto/blake2b/blake2x.go generated vendored Normal file
View File

@@ -0,0 +1,177 @@
// Copyright 2017 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 blake2b
import (
"encoding/binary"
"errors"
"io"
)
// XOF defines the interface to hash functions that
// support arbitrary-length output.
type XOF interface {
// Write absorbs more data into the hash's state. It panics if called
// after Read.
io.Writer
// Read reads more output from the hash. It returns io.EOF if the limit
// has been reached.
io.Reader
// Clone returns a copy of the XOF in its current state.
Clone() XOF
// Reset resets the XOF to its initial state.
Reset()
}
// OutputLengthUnknown can be used as the size argument to NewXOF to indicate
// the length of the output is not known in advance.
const OutputLengthUnknown = 0
// magicUnknownOutputLength is a magic value for the output size that indicates
// an unknown number of output bytes.
const magicUnknownOutputLength = (1 << 32) - 1
// maxOutputLength is the absolute maximum number of bytes to produce when the
// number of output bytes is unknown.
const maxOutputLength = (1 << 32) * 64
// NewXOF creates a new variable-output-length hash. The hash either produce a
// known number of bytes (1 <= size < 2**32-1), or an unknown number of bytes
// (size == OutputLengthUnknown). In the latter case, an absolute limit of
// 256GiB applies.
//
// A non-nil key turns the hash into a MAC. The key must between
// zero and 32 bytes long.
func NewXOF(size uint32, key []byte) (XOF, error) {
if len(key) > Size {
return nil, errKeySize
}
if size == magicUnknownOutputLength {
// 2^32-1 indicates an unknown number of bytes and thus isn't a
// valid length.
return nil, errors.New("blake2b: XOF length too large")
}
if size == OutputLengthUnknown {
size = magicUnknownOutputLength
}
x := &xof{
d: digest{
size: Size,
keyLen: len(key),
},
length: size,
}
copy(x.d.key[:], key)
x.Reset()
return x, nil
}
type xof struct {
d digest
length uint32
remaining uint64
cfg, root, block [Size]byte
offset int
nodeOffset uint32
readMode bool
}
func (x *xof) Write(p []byte) (n int, err error) {
if x.readMode {
panic("blake2b: write to XOF after read")
}
return x.d.Write(p)
}
func (x *xof) Clone() XOF {
clone := *x
return &clone
}
func (x *xof) Reset() {
x.cfg[0] = byte(Size)
binary.LittleEndian.PutUint32(x.cfg[4:], uint32(Size)) // leaf length
binary.LittleEndian.PutUint32(x.cfg[12:], x.length) // XOF length
x.cfg[17] = byte(Size) // inner hash size
x.d.Reset()
x.d.h[1] ^= uint64(x.length) << 32
x.remaining = uint64(x.length)
if x.remaining == magicUnknownOutputLength {
x.remaining = maxOutputLength
}
x.offset, x.nodeOffset = 0, 0
x.readMode = false
}
func (x *xof) Read(p []byte) (n int, err error) {
if !x.readMode {
x.d.finalize(&x.root)
x.readMode = true
}
if x.remaining == 0 {
return 0, io.EOF
}
n = len(p)
if uint64(n) > x.remaining {
n = int(x.remaining)
p = p[:n]
}
if x.offset > 0 {
blockRemaining := Size - x.offset
if n < blockRemaining {
x.offset += copy(p, x.block[x.offset:])
x.remaining -= uint64(n)
return
}
copy(p, x.block[x.offset:])
p = p[blockRemaining:]
x.offset = 0
x.remaining -= uint64(blockRemaining)
}
for len(p) >= Size {
binary.LittleEndian.PutUint32(x.cfg[8:], x.nodeOffset)
x.nodeOffset++
x.d.initConfig(&x.cfg)
x.d.Write(x.root[:])
x.d.finalize(&x.block)
copy(p, x.block[:])
p = p[Size:]
x.remaining -= uint64(Size)
}
if todo := len(p); todo > 0 {
if x.remaining < uint64(Size) {
x.cfg[0] = byte(x.remaining)
}
binary.LittleEndian.PutUint32(x.cfg[8:], x.nodeOffset)
x.nodeOffset++
x.d.initConfig(&x.cfg)
x.d.Write(x.root[:])
x.d.finalize(&x.block)
x.offset = copy(p, x.block[:todo])
x.remaining -= uint64(todo)
}
return
}
func (d *digest) initConfig(cfg *[Size]byte) {
d.offset, d.c[0], d.c[1] = 0, 0, 0
for i := range d.h {
d.h[i] = iv[i] ^ binary.LittleEndian.Uint64(cfg[i*8:])
}
}

33
vendor/golang.org/x/crypto/blake2b/register.go generated vendored Normal file
View File

@@ -0,0 +1,33 @@
// Copyright 2017 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.
//go:build go1.9
// +build go1.9
package blake2b
import (
"crypto"
"hash"
)
func init() {
newHash256 := func() hash.Hash {
h, _ := New256(nil)
return h
}
newHash384 := func() hash.Hash {
h, _ := New384(nil)
return h
}
newHash512 := func() hash.Hash {
h, _ := New512(nil)
return h
}
crypto.RegisterHash(crypto.BLAKE2b_256, newHash256)
crypto.RegisterHash(crypto.BLAKE2b_384, newHash384)
crypto.RegisterHash(crypto.BLAKE2b_512, newHash512)
}

93
vendor/golang.org/x/crypto/hkdf/hkdf.go generated vendored Normal file
View File

@@ -0,0 +1,93 @@
// Copyright 2014 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 hkdf implements the HMAC-based Extract-and-Expand Key Derivation
// Function (HKDF) as defined in RFC 5869.
//
// HKDF is a cryptographic key derivation function (KDF) with the goal of
// expanding limited input keying material into one or more cryptographically
// strong secret keys.
package hkdf // import "golang.org/x/crypto/hkdf"
import (
"crypto/hmac"
"errors"
"hash"
"io"
)
// Extract generates a pseudorandom key for use with Expand from an input secret
// and an optional independent salt.
//
// Only use this function if you need to reuse the extracted key with multiple
// Expand invocations and different context values. Most common scenarios,
// including the generation of multiple keys, should use New instead.
func Extract(hash func() hash.Hash, secret, salt []byte) []byte {
if salt == nil {
salt = make([]byte, hash().Size())
}
extractor := hmac.New(hash, salt)
extractor.Write(secret)
return extractor.Sum(nil)
}
type hkdf struct {
expander hash.Hash
size int
info []byte
counter byte
prev []byte
buf []byte
}
func (f *hkdf) Read(p []byte) (int, error) {
// Check whether enough data can be generated
need := len(p)
remains := len(f.buf) + int(255-f.counter+1)*f.size
if remains < need {
return 0, errors.New("hkdf: entropy limit reached")
}
// Read any leftover from the buffer
n := copy(p, f.buf)
p = p[n:]
// Fill the rest of the buffer
for len(p) > 0 {
f.expander.Reset()
f.expander.Write(f.prev)
f.expander.Write(f.info)
f.expander.Write([]byte{f.counter})
f.prev = f.expander.Sum(f.prev[:0])
f.counter++
// Copy the new batch into p
f.buf = f.prev
n = copy(p, f.buf)
p = p[n:]
}
// Save leftovers for next run
f.buf = f.buf[n:]
return need, nil
}
// Expand returns a Reader, from which keys can be read, using the given
// pseudorandom key and optional context info, skipping the extraction step.
//
// The pseudorandomKey should have been generated by Extract, or be a uniformly
// random or pseudorandom cryptographically strong key. See RFC 5869, Section
// 3.3. Most common scenarios will want to use New instead.
func Expand(hash func() hash.Hash, pseudorandomKey, info []byte) io.Reader {
expander := hmac.New(hash, pseudorandomKey)
return &hkdf{expander, expander.Size(), info, 1, nil, nil}
}
// New returns a Reader, from which keys can be read, using the given hash,
// secret, salt and context info. Salt and info can be nil.
func New(hash func() hash.Hash, secret, salt, info []byte) io.Reader {
prk := Extract(hash, secret, salt)
return Expand(hash, prk, info)
}

62
vendor/golang.org/x/crypto/sha3/doc.go generated vendored Normal file
View File

@@ -0,0 +1,62 @@
// Copyright 2014 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 sha3 implements the SHA-3 fixed-output-length hash functions and
// the SHAKE variable-output-length hash functions defined by FIPS-202.
//
// Both types of hash function use the "sponge" construction and the Keccak
// permutation. For a detailed specification see http://keccak.noekeon.org/
//
// # Guidance
//
// If you aren't sure what function you need, use SHAKE256 with at least 64
// bytes of output. The SHAKE instances are faster than the SHA3 instances;
// the latter have to allocate memory to conform to the hash.Hash interface.
//
// If you need a secret-key MAC (message authentication code), prepend the
// secret key to the input, hash with SHAKE256 and read at least 32 bytes of
// output.
//
// # Security strengths
//
// The SHA3-x (x equals 224, 256, 384, or 512) functions have a security
// strength against preimage attacks of x bits. Since they only produce "x"
// bits of output, their collision-resistance is only "x/2" bits.
//
// The SHAKE-256 and -128 functions have a generic security strength of 256 and
// 128 bits against all attacks, provided that at least 2x bits of their output
// is used. Requesting more than 64 or 32 bytes of output, respectively, does
// not increase the collision-resistance of the SHAKE functions.
//
// # The sponge construction
//
// A sponge builds a pseudo-random function from a public pseudo-random
// permutation, by applying the permutation to a state of "rate + capacity"
// bytes, but hiding "capacity" of the bytes.
//
// A sponge starts out with a zero state. To hash an input using a sponge, up
// to "rate" bytes of the input are XORed into the sponge's state. The sponge
// is then "full" and the permutation is applied to "empty" it. This process is
// repeated until all the input has been "absorbed". The input is then padded.
// The digest is "squeezed" from the sponge in the same way, except that output
// is copied out instead of input being XORed in.
//
// A sponge is parameterized by its generic security strength, which is equal
// to half its capacity; capacity + rate is equal to the permutation's width.
// Since the KeccakF-1600 permutation is 1600 bits (200 bytes) wide, this means
// that the security strength of a sponge instance is equal to (1600 - bitrate) / 2.
//
// # Recommendations
//
// The SHAKE functions are recommended for most new uses. They can produce
// output of arbitrary length. SHAKE256, with an output length of at least
// 64 bytes, provides 256-bit security against all attacks. The Keccak team
// recommends it for most applications upgrading from SHA2-512. (NIST chose a
// much stronger, but much slower, sponge instance for SHA3-512.)
//
// The SHA-3 functions are "drop-in" replacements for the SHA-2 functions.
// They produce output of the same length, with the same security strengths
// against all attacks. This means, in particular, that SHA3-256 only has
// 128-bit collision resistance, because its output length is 32 bytes.
package sha3 // import "golang.org/x/crypto/sha3"

97
vendor/golang.org/x/crypto/sha3/hashes.go generated vendored Normal file
View File

@@ -0,0 +1,97 @@
// Copyright 2014 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 sha3
// This file provides functions for creating instances of the SHA-3
// and SHAKE hash functions, as well as utility functions for hashing
// bytes.
import (
"hash"
)
// New224 creates a new SHA3-224 hash.
// Its generic security strength is 224 bits against preimage attacks,
// and 112 bits against collision attacks.
func New224() hash.Hash {
if h := new224Asm(); h != nil {
return h
}
return &state{rate: 144, outputLen: 28, dsbyte: 0x06}
}
// New256 creates a new SHA3-256 hash.
// Its generic security strength is 256 bits against preimage attacks,
// and 128 bits against collision attacks.
func New256() hash.Hash {
if h := new256Asm(); h != nil {
return h
}
return &state{rate: 136, outputLen: 32, dsbyte: 0x06}
}
// New384 creates a new SHA3-384 hash.
// Its generic security strength is 384 bits against preimage attacks,
// and 192 bits against collision attacks.
func New384() hash.Hash {
if h := new384Asm(); h != nil {
return h
}
return &state{rate: 104, outputLen: 48, dsbyte: 0x06}
}
// New512 creates a new SHA3-512 hash.
// Its generic security strength is 512 bits against preimage attacks,
// and 256 bits against collision attacks.
func New512() hash.Hash {
if h := new512Asm(); h != nil {
return h
}
return &state{rate: 72, outputLen: 64, dsbyte: 0x06}
}
// NewLegacyKeccak256 creates a new Keccak-256 hash.
//
// Only use this function if you require compatibility with an existing cryptosystem
// that uses non-standard padding. All other users should use New256 instead.
func NewLegacyKeccak256() hash.Hash { return &state{rate: 136, outputLen: 32, dsbyte: 0x01} }
// NewLegacyKeccak512 creates a new Keccak-512 hash.
//
// Only use this function if you require compatibility with an existing cryptosystem
// that uses non-standard padding. All other users should use New512 instead.
func NewLegacyKeccak512() hash.Hash { return &state{rate: 72, outputLen: 64, dsbyte: 0x01} }
// Sum224 returns the SHA3-224 digest of the data.
func Sum224(data []byte) (digest [28]byte) {
h := New224()
h.Write(data)
h.Sum(digest[:0])
return
}
// Sum256 returns the SHA3-256 digest of the data.
func Sum256(data []byte) (digest [32]byte) {
h := New256()
h.Write(data)
h.Sum(digest[:0])
return
}
// Sum384 returns the SHA3-384 digest of the data.
func Sum384(data []byte) (digest [48]byte) {
h := New384()
h.Write(data)
h.Sum(digest[:0])
return
}
// Sum512 returns the SHA3-512 digest of the data.
func Sum512(data []byte) (digest [64]byte) {
h := New512()
h.Write(data)
h.Sum(digest[:0])
return
}

28
vendor/golang.org/x/crypto/sha3/hashes_generic.go generated vendored Normal file
View File

@@ -0,0 +1,28 @@
// Copyright 2017 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.
//go:build !gc || purego || !s390x
// +build !gc purego !s390x
package sha3
import (
"hash"
)
// new224Asm returns an assembly implementation of SHA3-224 if available,
// otherwise it returns nil.
func new224Asm() hash.Hash { return nil }
// new256Asm returns an assembly implementation of SHA3-256 if available,
// otherwise it returns nil.
func new256Asm() hash.Hash { return nil }
// new384Asm returns an assembly implementation of SHA3-384 if available,
// otherwise it returns nil.
func new384Asm() hash.Hash { return nil }
// new512Asm returns an assembly implementation of SHA3-512 if available,
// otherwise it returns nil.
func new512Asm() hash.Hash { return nil }

415
vendor/golang.org/x/crypto/sha3/keccakf.go generated vendored Normal file
View File

@@ -0,0 +1,415 @@
// Copyright 2014 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.
//go:build !amd64 || purego || !gc
// +build !amd64 purego !gc
package sha3
import "math/bits"
// rc stores the round constants for use in the ι step.
var rc = [24]uint64{
0x0000000000000001,
0x0000000000008082,
0x800000000000808A,
0x8000000080008000,
0x000000000000808B,
0x0000000080000001,
0x8000000080008081,
0x8000000000008009,
0x000000000000008A,
0x0000000000000088,
0x0000000080008009,
0x000000008000000A,
0x000000008000808B,
0x800000000000008B,
0x8000000000008089,
0x8000000000008003,
0x8000000000008002,
0x8000000000000080,
0x000000000000800A,
0x800000008000000A,
0x8000000080008081,
0x8000000000008080,
0x0000000080000001,
0x8000000080008008,
}
// keccakF1600 applies the Keccak permutation to a 1600b-wide
// state represented as a slice of 25 uint64s.
func keccakF1600(a *[25]uint64) {
// Implementation translated from Keccak-inplace.c
// in the keccak reference code.
var t, bc0, bc1, bc2, bc3, bc4, d0, d1, d2, d3, d4 uint64
for i := 0; i < 24; i += 4 {
// Combines the 5 steps in each round into 2 steps.
// Unrolls 4 rounds per loop and spreads some steps across rounds.
// Round 1
bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
d0 = bc4 ^ (bc1<<1 | bc1>>63)
d1 = bc0 ^ (bc2<<1 | bc2>>63)
d2 = bc1 ^ (bc3<<1 | bc3>>63)
d3 = bc2 ^ (bc4<<1 | bc4>>63)
d4 = bc3 ^ (bc0<<1 | bc0>>63)
bc0 = a[0] ^ d0
t = a[6] ^ d1
bc1 = bits.RotateLeft64(t, 44)
t = a[12] ^ d2
bc2 = bits.RotateLeft64(t, 43)
t = a[18] ^ d3
bc3 = bits.RotateLeft64(t, 21)
t = a[24] ^ d4
bc4 = bits.RotateLeft64(t, 14)
a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i]
a[6] = bc1 ^ (bc3 &^ bc2)
a[12] = bc2 ^ (bc4 &^ bc3)
a[18] = bc3 ^ (bc0 &^ bc4)
a[24] = bc4 ^ (bc1 &^ bc0)
t = a[10] ^ d0
bc2 = bits.RotateLeft64(t, 3)
t = a[16] ^ d1
bc3 = bits.RotateLeft64(t, 45)
t = a[22] ^ d2
bc4 = bits.RotateLeft64(t, 61)
t = a[3] ^ d3
bc0 = bits.RotateLeft64(t, 28)
t = a[9] ^ d4
bc1 = bits.RotateLeft64(t, 20)
a[10] = bc0 ^ (bc2 &^ bc1)
a[16] = bc1 ^ (bc3 &^ bc2)
a[22] = bc2 ^ (bc4 &^ bc3)
a[3] = bc3 ^ (bc0 &^ bc4)
a[9] = bc4 ^ (bc1 &^ bc0)
t = a[20] ^ d0
bc4 = bits.RotateLeft64(t, 18)
t = a[1] ^ d1
bc0 = bits.RotateLeft64(t, 1)
t = a[7] ^ d2
bc1 = bits.RotateLeft64(t, 6)
t = a[13] ^ d3
bc2 = bits.RotateLeft64(t, 25)
t = a[19] ^ d4
bc3 = bits.RotateLeft64(t, 8)
a[20] = bc0 ^ (bc2 &^ bc1)
a[1] = bc1 ^ (bc3 &^ bc2)
a[7] = bc2 ^ (bc4 &^ bc3)
a[13] = bc3 ^ (bc0 &^ bc4)
a[19] = bc4 ^ (bc1 &^ bc0)
t = a[5] ^ d0
bc1 = bits.RotateLeft64(t, 36)
t = a[11] ^ d1
bc2 = bits.RotateLeft64(t, 10)
t = a[17] ^ d2
bc3 = bits.RotateLeft64(t, 15)
t = a[23] ^ d3
bc4 = bits.RotateLeft64(t, 56)
t = a[4] ^ d4
bc0 = bits.RotateLeft64(t, 27)
a[5] = bc0 ^ (bc2 &^ bc1)
a[11] = bc1 ^ (bc3 &^ bc2)
a[17] = bc2 ^ (bc4 &^ bc3)
a[23] = bc3 ^ (bc0 &^ bc4)
a[4] = bc4 ^ (bc1 &^ bc0)
t = a[15] ^ d0
bc3 = bits.RotateLeft64(t, 41)
t = a[21] ^ d1
bc4 = bits.RotateLeft64(t, 2)
t = a[2] ^ d2
bc0 = bits.RotateLeft64(t, 62)
t = a[8] ^ d3
bc1 = bits.RotateLeft64(t, 55)
t = a[14] ^ d4
bc2 = bits.RotateLeft64(t, 39)
a[15] = bc0 ^ (bc2 &^ bc1)
a[21] = bc1 ^ (bc3 &^ bc2)
a[2] = bc2 ^ (bc4 &^ bc3)
a[8] = bc3 ^ (bc0 &^ bc4)
a[14] = bc4 ^ (bc1 &^ bc0)
// Round 2
bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
d0 = bc4 ^ (bc1<<1 | bc1>>63)
d1 = bc0 ^ (bc2<<1 | bc2>>63)
d2 = bc1 ^ (bc3<<1 | bc3>>63)
d3 = bc2 ^ (bc4<<1 | bc4>>63)
d4 = bc3 ^ (bc0<<1 | bc0>>63)
bc0 = a[0] ^ d0
t = a[16] ^ d1
bc1 = bits.RotateLeft64(t, 44)
t = a[7] ^ d2
bc2 = bits.RotateLeft64(t, 43)
t = a[23] ^ d3
bc3 = bits.RotateLeft64(t, 21)
t = a[14] ^ d4
bc4 = bits.RotateLeft64(t, 14)
a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i+1]
a[16] = bc1 ^ (bc3 &^ bc2)
a[7] = bc2 ^ (bc4 &^ bc3)
a[23] = bc3 ^ (bc0 &^ bc4)
a[14] = bc4 ^ (bc1 &^ bc0)
t = a[20] ^ d0
bc2 = bits.RotateLeft64(t, 3)
t = a[11] ^ d1
bc3 = bits.RotateLeft64(t, 45)
t = a[2] ^ d2
bc4 = bits.RotateLeft64(t, 61)
t = a[18] ^ d3
bc0 = bits.RotateLeft64(t, 28)
t = a[9] ^ d4
bc1 = bits.RotateLeft64(t, 20)
a[20] = bc0 ^ (bc2 &^ bc1)
a[11] = bc1 ^ (bc3 &^ bc2)
a[2] = bc2 ^ (bc4 &^ bc3)
a[18] = bc3 ^ (bc0 &^ bc4)
a[9] = bc4 ^ (bc1 &^ bc0)
t = a[15] ^ d0
bc4 = bits.RotateLeft64(t, 18)
t = a[6] ^ d1
bc0 = bits.RotateLeft64(t, 1)
t = a[22] ^ d2
bc1 = bits.RotateLeft64(t, 6)
t = a[13] ^ d3
bc2 = bits.RotateLeft64(t, 25)
t = a[4] ^ d4
bc3 = bits.RotateLeft64(t, 8)
a[15] = bc0 ^ (bc2 &^ bc1)
a[6] = bc1 ^ (bc3 &^ bc2)
a[22] = bc2 ^ (bc4 &^ bc3)
a[13] = bc3 ^ (bc0 &^ bc4)
a[4] = bc4 ^ (bc1 &^ bc0)
t = a[10] ^ d0
bc1 = bits.RotateLeft64(t, 36)
t = a[1] ^ d1
bc2 = bits.RotateLeft64(t, 10)
t = a[17] ^ d2
bc3 = bits.RotateLeft64(t, 15)
t = a[8] ^ d3
bc4 = bits.RotateLeft64(t, 56)
t = a[24] ^ d4
bc0 = bits.RotateLeft64(t, 27)
a[10] = bc0 ^ (bc2 &^ bc1)
a[1] = bc1 ^ (bc3 &^ bc2)
a[17] = bc2 ^ (bc4 &^ bc3)
a[8] = bc3 ^ (bc0 &^ bc4)
a[24] = bc4 ^ (bc1 &^ bc0)
t = a[5] ^ d0
bc3 = bits.RotateLeft64(t, 41)
t = a[21] ^ d1
bc4 = bits.RotateLeft64(t, 2)
t = a[12] ^ d2
bc0 = bits.RotateLeft64(t, 62)
t = a[3] ^ d3
bc1 = bits.RotateLeft64(t, 55)
t = a[19] ^ d4
bc2 = bits.RotateLeft64(t, 39)
a[5] = bc0 ^ (bc2 &^ bc1)
a[21] = bc1 ^ (bc3 &^ bc2)
a[12] = bc2 ^ (bc4 &^ bc3)
a[3] = bc3 ^ (bc0 &^ bc4)
a[19] = bc4 ^ (bc1 &^ bc0)
// Round 3
bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
d0 = bc4 ^ (bc1<<1 | bc1>>63)
d1 = bc0 ^ (bc2<<1 | bc2>>63)
d2 = bc1 ^ (bc3<<1 | bc3>>63)
d3 = bc2 ^ (bc4<<1 | bc4>>63)
d4 = bc3 ^ (bc0<<1 | bc0>>63)
bc0 = a[0] ^ d0
t = a[11] ^ d1
bc1 = bits.RotateLeft64(t, 44)
t = a[22] ^ d2
bc2 = bits.RotateLeft64(t, 43)
t = a[8] ^ d3
bc3 = bits.RotateLeft64(t, 21)
t = a[19] ^ d4
bc4 = bits.RotateLeft64(t, 14)
a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i+2]
a[11] = bc1 ^ (bc3 &^ bc2)
a[22] = bc2 ^ (bc4 &^ bc3)
a[8] = bc3 ^ (bc0 &^ bc4)
a[19] = bc4 ^ (bc1 &^ bc0)
t = a[15] ^ d0
bc2 = bits.RotateLeft64(t, 3)
t = a[1] ^ d1
bc3 = bits.RotateLeft64(t, 45)
t = a[12] ^ d2
bc4 = bits.RotateLeft64(t, 61)
t = a[23] ^ d3
bc0 = bits.RotateLeft64(t, 28)
t = a[9] ^ d4
bc1 = bits.RotateLeft64(t, 20)
a[15] = bc0 ^ (bc2 &^ bc1)
a[1] = bc1 ^ (bc3 &^ bc2)
a[12] = bc2 ^ (bc4 &^ bc3)
a[23] = bc3 ^ (bc0 &^ bc4)
a[9] = bc4 ^ (bc1 &^ bc0)
t = a[5] ^ d0
bc4 = bits.RotateLeft64(t, 18)
t = a[16] ^ d1
bc0 = bits.RotateLeft64(t, 1)
t = a[2] ^ d2
bc1 = bits.RotateLeft64(t, 6)
t = a[13] ^ d3
bc2 = bits.RotateLeft64(t, 25)
t = a[24] ^ d4
bc3 = bits.RotateLeft64(t, 8)
a[5] = bc0 ^ (bc2 &^ bc1)
a[16] = bc1 ^ (bc3 &^ bc2)
a[2] = bc2 ^ (bc4 &^ bc3)
a[13] = bc3 ^ (bc0 &^ bc4)
a[24] = bc4 ^ (bc1 &^ bc0)
t = a[20] ^ d0
bc1 = bits.RotateLeft64(t, 36)
t = a[6] ^ d1
bc2 = bits.RotateLeft64(t, 10)
t = a[17] ^ d2
bc3 = bits.RotateLeft64(t, 15)
t = a[3] ^ d3
bc4 = bits.RotateLeft64(t, 56)
t = a[14] ^ d4
bc0 = bits.RotateLeft64(t, 27)
a[20] = bc0 ^ (bc2 &^ bc1)
a[6] = bc1 ^ (bc3 &^ bc2)
a[17] = bc2 ^ (bc4 &^ bc3)
a[3] = bc3 ^ (bc0 &^ bc4)
a[14] = bc4 ^ (bc1 &^ bc0)
t = a[10] ^ d0
bc3 = bits.RotateLeft64(t, 41)
t = a[21] ^ d1
bc4 = bits.RotateLeft64(t, 2)
t = a[7] ^ d2
bc0 = bits.RotateLeft64(t, 62)
t = a[18] ^ d3
bc1 = bits.RotateLeft64(t, 55)
t = a[4] ^ d4
bc2 = bits.RotateLeft64(t, 39)
a[10] = bc0 ^ (bc2 &^ bc1)
a[21] = bc1 ^ (bc3 &^ bc2)
a[7] = bc2 ^ (bc4 &^ bc3)
a[18] = bc3 ^ (bc0 &^ bc4)
a[4] = bc4 ^ (bc1 &^ bc0)
// Round 4
bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
d0 = bc4 ^ (bc1<<1 | bc1>>63)
d1 = bc0 ^ (bc2<<1 | bc2>>63)
d2 = bc1 ^ (bc3<<1 | bc3>>63)
d3 = bc2 ^ (bc4<<1 | bc4>>63)
d4 = bc3 ^ (bc0<<1 | bc0>>63)
bc0 = a[0] ^ d0
t = a[1] ^ d1
bc1 = bits.RotateLeft64(t, 44)
t = a[2] ^ d2
bc2 = bits.RotateLeft64(t, 43)
t = a[3] ^ d3
bc3 = bits.RotateLeft64(t, 21)
t = a[4] ^ d4
bc4 = bits.RotateLeft64(t, 14)
a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i+3]
a[1] = bc1 ^ (bc3 &^ bc2)
a[2] = bc2 ^ (bc4 &^ bc3)
a[3] = bc3 ^ (bc0 &^ bc4)
a[4] = bc4 ^ (bc1 &^ bc0)
t = a[5] ^ d0
bc2 = bits.RotateLeft64(t, 3)
t = a[6] ^ d1
bc3 = bits.RotateLeft64(t, 45)
t = a[7] ^ d2
bc4 = bits.RotateLeft64(t, 61)
t = a[8] ^ d3
bc0 = bits.RotateLeft64(t, 28)
t = a[9] ^ d4
bc1 = bits.RotateLeft64(t, 20)
a[5] = bc0 ^ (bc2 &^ bc1)
a[6] = bc1 ^ (bc3 &^ bc2)
a[7] = bc2 ^ (bc4 &^ bc3)
a[8] = bc3 ^ (bc0 &^ bc4)
a[9] = bc4 ^ (bc1 &^ bc0)
t = a[10] ^ d0
bc4 = bits.RotateLeft64(t, 18)
t = a[11] ^ d1
bc0 = bits.RotateLeft64(t, 1)
t = a[12] ^ d2
bc1 = bits.RotateLeft64(t, 6)
t = a[13] ^ d3
bc2 = bits.RotateLeft64(t, 25)
t = a[14] ^ d4
bc3 = bits.RotateLeft64(t, 8)
a[10] = bc0 ^ (bc2 &^ bc1)
a[11] = bc1 ^ (bc3 &^ bc2)
a[12] = bc2 ^ (bc4 &^ bc3)
a[13] = bc3 ^ (bc0 &^ bc4)
a[14] = bc4 ^ (bc1 &^ bc0)
t = a[15] ^ d0
bc1 = bits.RotateLeft64(t, 36)
t = a[16] ^ d1
bc2 = bits.RotateLeft64(t, 10)
t = a[17] ^ d2
bc3 = bits.RotateLeft64(t, 15)
t = a[18] ^ d3
bc4 = bits.RotateLeft64(t, 56)
t = a[19] ^ d4
bc0 = bits.RotateLeft64(t, 27)
a[15] = bc0 ^ (bc2 &^ bc1)
a[16] = bc1 ^ (bc3 &^ bc2)
a[17] = bc2 ^ (bc4 &^ bc3)
a[18] = bc3 ^ (bc0 &^ bc4)
a[19] = bc4 ^ (bc1 &^ bc0)
t = a[20] ^ d0
bc3 = bits.RotateLeft64(t, 41)
t = a[21] ^ d1
bc4 = bits.RotateLeft64(t, 2)
t = a[22] ^ d2
bc0 = bits.RotateLeft64(t, 62)
t = a[23] ^ d3
bc1 = bits.RotateLeft64(t, 55)
t = a[24] ^ d4
bc2 = bits.RotateLeft64(t, 39)
a[20] = bc0 ^ (bc2 &^ bc1)
a[21] = bc1 ^ (bc3 &^ bc2)
a[22] = bc2 ^ (bc4 &^ bc3)
a[23] = bc3 ^ (bc0 &^ bc4)
a[24] = bc4 ^ (bc1 &^ bc0)
}
}

14
vendor/golang.org/x/crypto/sha3/keccakf_amd64.go generated vendored Normal file
View File

@@ -0,0 +1,14 @@
// 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.
//go:build amd64 && !purego && gc
// +build amd64,!purego,gc
package sha3
// This function is implemented in keccakf_amd64.s.
//go:noescape
func keccakF1600(a *[25]uint64)

391
vendor/golang.org/x/crypto/sha3/keccakf_amd64.s generated vendored Normal file
View File

@@ -0,0 +1,391 @@
// 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.
//go:build amd64 && !purego && gc
// +build amd64,!purego,gc
// This code was translated into a form compatible with 6a from the public
// domain sources at https://github.com/gvanas/KeccakCodePackage
// Offsets in state
#define _ba (0*8)
#define _be (1*8)
#define _bi (2*8)
#define _bo (3*8)
#define _bu (4*8)
#define _ga (5*8)
#define _ge (6*8)
#define _gi (7*8)
#define _go (8*8)
#define _gu (9*8)
#define _ka (10*8)
#define _ke (11*8)
#define _ki (12*8)
#define _ko (13*8)
#define _ku (14*8)
#define _ma (15*8)
#define _me (16*8)
#define _mi (17*8)
#define _mo (18*8)
#define _mu (19*8)
#define _sa (20*8)
#define _se (21*8)
#define _si (22*8)
#define _so (23*8)
#define _su (24*8)
// Temporary registers
#define rT1 AX
// Round vars
#define rpState DI
#define rpStack SP
#define rDa BX
#define rDe CX
#define rDi DX
#define rDo R8
#define rDu R9
#define rBa R10
#define rBe R11
#define rBi R12
#define rBo R13
#define rBu R14
#define rCa SI
#define rCe BP
#define rCi rBi
#define rCo rBo
#define rCu R15
#define MOVQ_RBI_RCE MOVQ rBi, rCe
#define XORQ_RT1_RCA XORQ rT1, rCa
#define XORQ_RT1_RCE XORQ rT1, rCe
#define XORQ_RBA_RCU XORQ rBa, rCu
#define XORQ_RBE_RCU XORQ rBe, rCu
#define XORQ_RDU_RCU XORQ rDu, rCu
#define XORQ_RDA_RCA XORQ rDa, rCa
#define XORQ_RDE_RCE XORQ rDe, rCe
#define mKeccakRound(iState, oState, rc, B_RBI_RCE, G_RT1_RCA, G_RT1_RCE, G_RBA_RCU, K_RT1_RCA, K_RT1_RCE, K_RBA_RCU, M_RT1_RCA, M_RT1_RCE, M_RBE_RCU, S_RDU_RCU, S_RDA_RCA, S_RDE_RCE) \
/* Prepare round */ \
MOVQ rCe, rDa; \
ROLQ $1, rDa; \
\
MOVQ _bi(iState), rCi; \
XORQ _gi(iState), rDi; \
XORQ rCu, rDa; \
XORQ _ki(iState), rCi; \
XORQ _mi(iState), rDi; \
XORQ rDi, rCi; \
\
MOVQ rCi, rDe; \
ROLQ $1, rDe; \
\
MOVQ _bo(iState), rCo; \
XORQ _go(iState), rDo; \
XORQ rCa, rDe; \
XORQ _ko(iState), rCo; \
XORQ _mo(iState), rDo; \
XORQ rDo, rCo; \
\
MOVQ rCo, rDi; \
ROLQ $1, rDi; \
\
MOVQ rCu, rDo; \
XORQ rCe, rDi; \
ROLQ $1, rDo; \
\
MOVQ rCa, rDu; \
XORQ rCi, rDo; \
ROLQ $1, rDu; \
\
/* Result b */ \
MOVQ _ba(iState), rBa; \
MOVQ _ge(iState), rBe; \
XORQ rCo, rDu; \
MOVQ _ki(iState), rBi; \
MOVQ _mo(iState), rBo; \
MOVQ _su(iState), rBu; \
XORQ rDe, rBe; \
ROLQ $44, rBe; \
XORQ rDi, rBi; \
XORQ rDa, rBa; \
ROLQ $43, rBi; \
\
MOVQ rBe, rCa; \
MOVQ rc, rT1; \
ORQ rBi, rCa; \
XORQ rBa, rT1; \
XORQ rT1, rCa; \
MOVQ rCa, _ba(oState); \
\
XORQ rDu, rBu; \
ROLQ $14, rBu; \
MOVQ rBa, rCu; \
ANDQ rBe, rCu; \
XORQ rBu, rCu; \
MOVQ rCu, _bu(oState); \
\
XORQ rDo, rBo; \
ROLQ $21, rBo; \
MOVQ rBo, rT1; \
ANDQ rBu, rT1; \
XORQ rBi, rT1; \
MOVQ rT1, _bi(oState); \
\
NOTQ rBi; \
ORQ rBa, rBu; \
ORQ rBo, rBi; \
XORQ rBo, rBu; \
XORQ rBe, rBi; \
MOVQ rBu, _bo(oState); \
MOVQ rBi, _be(oState); \
B_RBI_RCE; \
\
/* Result g */ \
MOVQ _gu(iState), rBe; \
XORQ rDu, rBe; \
MOVQ _ka(iState), rBi; \
ROLQ $20, rBe; \
XORQ rDa, rBi; \
ROLQ $3, rBi; \
MOVQ _bo(iState), rBa; \
MOVQ rBe, rT1; \
ORQ rBi, rT1; \
XORQ rDo, rBa; \
MOVQ _me(iState), rBo; \
MOVQ _si(iState), rBu; \
ROLQ $28, rBa; \
XORQ rBa, rT1; \
MOVQ rT1, _ga(oState); \
G_RT1_RCA; \
\
XORQ rDe, rBo; \
ROLQ $45, rBo; \
MOVQ rBi, rT1; \
ANDQ rBo, rT1; \
XORQ rBe, rT1; \
MOVQ rT1, _ge(oState); \
G_RT1_RCE; \
\
XORQ rDi, rBu; \
ROLQ $61, rBu; \
MOVQ rBu, rT1; \
ORQ rBa, rT1; \
XORQ rBo, rT1; \
MOVQ rT1, _go(oState); \
\
ANDQ rBe, rBa; \
XORQ rBu, rBa; \
MOVQ rBa, _gu(oState); \
NOTQ rBu; \
G_RBA_RCU; \
\
ORQ rBu, rBo; \
XORQ rBi, rBo; \
MOVQ rBo, _gi(oState); \
\
/* Result k */ \
MOVQ _be(iState), rBa; \
MOVQ _gi(iState), rBe; \
MOVQ _ko(iState), rBi; \
MOVQ _mu(iState), rBo; \
MOVQ _sa(iState), rBu; \
XORQ rDi, rBe; \
ROLQ $6, rBe; \
XORQ rDo, rBi; \
ROLQ $25, rBi; \
MOVQ rBe, rT1; \
ORQ rBi, rT1; \
XORQ rDe, rBa; \
ROLQ $1, rBa; \
XORQ rBa, rT1; \
MOVQ rT1, _ka(oState); \
K_RT1_RCA; \
\
XORQ rDu, rBo; \
ROLQ $8, rBo; \
MOVQ rBi, rT1; \
ANDQ rBo, rT1; \
XORQ rBe, rT1; \
MOVQ rT1, _ke(oState); \
K_RT1_RCE; \
\
XORQ rDa, rBu; \
ROLQ $18, rBu; \
NOTQ rBo; \
MOVQ rBo, rT1; \
ANDQ rBu, rT1; \
XORQ rBi, rT1; \
MOVQ rT1, _ki(oState); \
\
MOVQ rBu, rT1; \
ORQ rBa, rT1; \
XORQ rBo, rT1; \
MOVQ rT1, _ko(oState); \
\
ANDQ rBe, rBa; \
XORQ rBu, rBa; \
MOVQ rBa, _ku(oState); \
K_RBA_RCU; \
\
/* Result m */ \
MOVQ _ga(iState), rBe; \
XORQ rDa, rBe; \
MOVQ _ke(iState), rBi; \
ROLQ $36, rBe; \
XORQ rDe, rBi; \
MOVQ _bu(iState), rBa; \
ROLQ $10, rBi; \
MOVQ rBe, rT1; \
MOVQ _mi(iState), rBo; \
ANDQ rBi, rT1; \
XORQ rDu, rBa; \
MOVQ _so(iState), rBu; \
ROLQ $27, rBa; \
XORQ rBa, rT1; \
MOVQ rT1, _ma(oState); \
M_RT1_RCA; \
\
XORQ rDi, rBo; \
ROLQ $15, rBo; \
MOVQ rBi, rT1; \
ORQ rBo, rT1; \
XORQ rBe, rT1; \
MOVQ rT1, _me(oState); \
M_RT1_RCE; \
\
XORQ rDo, rBu; \
ROLQ $56, rBu; \
NOTQ rBo; \
MOVQ rBo, rT1; \
ORQ rBu, rT1; \
XORQ rBi, rT1; \
MOVQ rT1, _mi(oState); \
\
ORQ rBa, rBe; \
XORQ rBu, rBe; \
MOVQ rBe, _mu(oState); \
\
ANDQ rBa, rBu; \
XORQ rBo, rBu; \
MOVQ rBu, _mo(oState); \
M_RBE_RCU; \
\
/* Result s */ \
MOVQ _bi(iState), rBa; \
MOVQ _go(iState), rBe; \
MOVQ _ku(iState), rBi; \
XORQ rDi, rBa; \
MOVQ _ma(iState), rBo; \
ROLQ $62, rBa; \
XORQ rDo, rBe; \
MOVQ _se(iState), rBu; \
ROLQ $55, rBe; \
\
XORQ rDu, rBi; \
MOVQ rBa, rDu; \
XORQ rDe, rBu; \
ROLQ $2, rBu; \
ANDQ rBe, rDu; \
XORQ rBu, rDu; \
MOVQ rDu, _su(oState); \
\
ROLQ $39, rBi; \
S_RDU_RCU; \
NOTQ rBe; \
XORQ rDa, rBo; \
MOVQ rBe, rDa; \
ANDQ rBi, rDa; \
XORQ rBa, rDa; \
MOVQ rDa, _sa(oState); \
S_RDA_RCA; \
\
ROLQ $41, rBo; \
MOVQ rBi, rDe; \
ORQ rBo, rDe; \
XORQ rBe, rDe; \
MOVQ rDe, _se(oState); \
S_RDE_RCE; \
\
MOVQ rBo, rDi; \
MOVQ rBu, rDo; \
ANDQ rBu, rDi; \
ORQ rBa, rDo; \
XORQ rBi, rDi; \
XORQ rBo, rDo; \
MOVQ rDi, _si(oState); \
MOVQ rDo, _so(oState) \
// func keccakF1600(state *[25]uint64)
TEXT ·keccakF1600(SB), 0, $200-8
MOVQ state+0(FP), rpState
// Convert the user state into an internal state
NOTQ _be(rpState)
NOTQ _bi(rpState)
NOTQ _go(rpState)
NOTQ _ki(rpState)
NOTQ _mi(rpState)
NOTQ _sa(rpState)
// Execute the KeccakF permutation
MOVQ _ba(rpState), rCa
MOVQ _be(rpState), rCe
MOVQ _bu(rpState), rCu
XORQ _ga(rpState), rCa
XORQ _ge(rpState), rCe
XORQ _gu(rpState), rCu
XORQ _ka(rpState), rCa
XORQ _ke(rpState), rCe
XORQ _ku(rpState), rCu
XORQ _ma(rpState), rCa
XORQ _me(rpState), rCe
XORQ _mu(rpState), rCu
XORQ _sa(rpState), rCa
XORQ _se(rpState), rCe
MOVQ _si(rpState), rDi
MOVQ _so(rpState), rDo
XORQ _su(rpState), rCu
mKeccakRound(rpState, rpStack, $0x0000000000000001, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x0000000000008082, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x800000000000808a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x8000000080008000, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x000000000000808b, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x0000000080000001, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x8000000080008081, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x8000000000008009, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x000000000000008a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x0000000000000088, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x0000000080008009, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x000000008000000a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x000000008000808b, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x800000000000008b, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x8000000000008089, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x8000000000008003, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x8000000000008002, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x8000000000000080, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x000000000000800a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x800000008000000a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x8000000080008081, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x8000000000008080, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpState, rpStack, $0x0000000080000001, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
mKeccakRound(rpStack, rpState, $0x8000000080008008, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP)
// Revert the internal state to the user state
NOTQ _be(rpState)
NOTQ _bi(rpState)
NOTQ _go(rpState)
NOTQ _ki(rpState)
NOTQ _mi(rpState)
NOTQ _sa(rpState)
RET

19
vendor/golang.org/x/crypto/sha3/register.go generated vendored Normal file
View File

@@ -0,0 +1,19 @@
// Copyright 2014 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.
//go:build go1.4
// +build go1.4
package sha3
import (
"crypto"
)
func init() {
crypto.RegisterHash(crypto.SHA3_224, New224)
crypto.RegisterHash(crypto.SHA3_256, New256)
crypto.RegisterHash(crypto.SHA3_384, New384)
crypto.RegisterHash(crypto.SHA3_512, New512)
}

193
vendor/golang.org/x/crypto/sha3/sha3.go generated vendored Normal file
View File

@@ -0,0 +1,193 @@
// Copyright 2014 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 sha3
// spongeDirection indicates the direction bytes are flowing through the sponge.
type spongeDirection int
const (
// spongeAbsorbing indicates that the sponge is absorbing input.
spongeAbsorbing spongeDirection = iota
// spongeSqueezing indicates that the sponge is being squeezed.
spongeSqueezing
)
const (
// maxRate is the maximum size of the internal buffer. SHAKE-256
// currently needs the largest buffer.
maxRate = 168
)
type state struct {
// Generic sponge components.
a [25]uint64 // main state of the hash
buf []byte // points into storage
rate int // the number of bytes of state to use
// dsbyte contains the "domain separation" bits and the first bit of
// the padding. Sections 6.1 and 6.2 of [1] separate the outputs of the
// SHA-3 and SHAKE functions by appending bitstrings to the message.
// Using a little-endian bit-ordering convention, these are "01" for SHA-3
// and "1111" for SHAKE, or 00000010b and 00001111b, respectively. Then the
// padding rule from section 5.1 is applied to pad the message to a multiple
// of the rate, which involves adding a "1" bit, zero or more "0" bits, and
// a final "1" bit. We merge the first "1" bit from the padding into dsbyte,
// giving 00000110b (0x06) and 00011111b (0x1f).
// [1] http://csrc.nist.gov/publications/drafts/fips-202/fips_202_draft.pdf
// "Draft FIPS 202: SHA-3 Standard: Permutation-Based Hash and
// Extendable-Output Functions (May 2014)"
dsbyte byte
storage storageBuf
// Specific to SHA-3 and SHAKE.
outputLen int // the default output size in bytes
state spongeDirection // whether the sponge is absorbing or squeezing
}
// BlockSize returns the rate of sponge underlying this hash function.
func (d *state) BlockSize() int { return d.rate }
// Size returns the output size of the hash function in bytes.
func (d *state) Size() int { return d.outputLen }
// Reset clears the internal state by zeroing the sponge state and
// the byte buffer, and setting Sponge.state to absorbing.
func (d *state) Reset() {
// Zero the permutation's state.
for i := range d.a {
d.a[i] = 0
}
d.state = spongeAbsorbing
d.buf = d.storage.asBytes()[:0]
}
func (d *state) clone() *state {
ret := *d
if ret.state == spongeAbsorbing {
ret.buf = ret.storage.asBytes()[:len(ret.buf)]
} else {
ret.buf = ret.storage.asBytes()[d.rate-cap(d.buf) : d.rate]
}
return &ret
}
// permute applies the KeccakF-1600 permutation. It handles
// any input-output buffering.
func (d *state) permute() {
switch d.state {
case spongeAbsorbing:
// If we're absorbing, we need to xor the input into the state
// before applying the permutation.
xorIn(d, d.buf)
d.buf = d.storage.asBytes()[:0]
keccakF1600(&d.a)
case spongeSqueezing:
// If we're squeezing, we need to apply the permutation before
// copying more output.
keccakF1600(&d.a)
d.buf = d.storage.asBytes()[:d.rate]
copyOut(d, d.buf)
}
}
// pads appends the domain separation bits in dsbyte, applies
// the multi-bitrate 10..1 padding rule, and permutes the state.
func (d *state) padAndPermute(dsbyte byte) {
if d.buf == nil {
d.buf = d.storage.asBytes()[:0]
}
// Pad with this instance's domain-separator bits. We know that there's
// at least one byte of space in d.buf because, if it were full,
// permute would have been called to empty it. dsbyte also contains the
// first one bit for the padding. See the comment in the state struct.
d.buf = append(d.buf, dsbyte)
zerosStart := len(d.buf)
d.buf = d.storage.asBytes()[:d.rate]
for i := zerosStart; i < d.rate; i++ {
d.buf[i] = 0
}
// This adds the final one bit for the padding. Because of the way that
// bits are numbered from the LSB upwards, the final bit is the MSB of
// the last byte.
d.buf[d.rate-1] ^= 0x80
// Apply the permutation
d.permute()
d.state = spongeSqueezing
d.buf = d.storage.asBytes()[:d.rate]
copyOut(d, d.buf)
}
// Write absorbs more data into the hash's state. It produces an error
// if more data is written to the ShakeHash after writing
func (d *state) Write(p []byte) (written int, err error) {
if d.state != spongeAbsorbing {
panic("sha3: write to sponge after read")
}
if d.buf == nil {
d.buf = d.storage.asBytes()[:0]
}
written = len(p)
for len(p) > 0 {
if len(d.buf) == 0 && len(p) >= d.rate {
// The fast path; absorb a full "rate" bytes of input and apply the permutation.
xorIn(d, p[:d.rate])
p = p[d.rate:]
keccakF1600(&d.a)
} else {
// The slow path; buffer the input until we can fill the sponge, and then xor it in.
todo := d.rate - len(d.buf)
if todo > len(p) {
todo = len(p)
}
d.buf = append(d.buf, p[:todo]...)
p = p[todo:]
// If the sponge is full, apply the permutation.
if len(d.buf) == d.rate {
d.permute()
}
}
}
return
}
// Read squeezes an arbitrary number of bytes from the sponge.
func (d *state) Read(out []byte) (n int, err error) {
// If we're still absorbing, pad and apply the permutation.
if d.state == spongeAbsorbing {
d.padAndPermute(d.dsbyte)
}
n = len(out)
// Now, do the squeezing.
for len(out) > 0 {
n := copy(out, d.buf)
d.buf = d.buf[n:]
out = out[n:]
// Apply the permutation if we've squeezed the sponge dry.
if len(d.buf) == 0 {
d.permute()
}
}
return
}
// Sum applies padding to the hash state and then squeezes out the desired
// number of output bytes.
func (d *state) Sum(in []byte) []byte {
// Make a copy of the original hash so that caller can keep writing
// and summing.
dup := d.clone()
hash := make([]byte, dup.outputLen)
dup.Read(hash)
return append(in, hash...)
}

287
vendor/golang.org/x/crypto/sha3/sha3_s390x.go generated vendored Normal file
View File

@@ -0,0 +1,287 @@
// Copyright 2017 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.
//go:build gc && !purego
// +build gc,!purego
package sha3
// This file contains code for using the 'compute intermediate
// message digest' (KIMD) and 'compute last message digest' (KLMD)
// instructions to compute SHA-3 and SHAKE hashes on IBM Z.
import (
"hash"
"golang.org/x/sys/cpu"
)
// codes represent 7-bit KIMD/KLMD function codes as defined in
// the Principles of Operation.
type code uint64
const (
// function codes for KIMD/KLMD
sha3_224 code = 32
sha3_256 = 33
sha3_384 = 34
sha3_512 = 35
shake_128 = 36
shake_256 = 37
nopad = 0x100
)
// kimd is a wrapper for the 'compute intermediate message digest' instruction.
// src must be a multiple of the rate for the given function code.
//
//go:noescape
func kimd(function code, chain *[200]byte, src []byte)
// klmd is a wrapper for the 'compute last message digest' instruction.
// src padding is handled by the instruction.
//
//go:noescape
func klmd(function code, chain *[200]byte, dst, src []byte)
type asmState struct {
a [200]byte // 1600 bit state
buf []byte // care must be taken to ensure cap(buf) is a multiple of rate
rate int // equivalent to block size
storage [3072]byte // underlying storage for buf
outputLen int // output length if fixed, 0 if not
function code // KIMD/KLMD function code
state spongeDirection // whether the sponge is absorbing or squeezing
}
func newAsmState(function code) *asmState {
var s asmState
s.function = function
switch function {
case sha3_224:
s.rate = 144
s.outputLen = 28
case sha3_256:
s.rate = 136
s.outputLen = 32
case sha3_384:
s.rate = 104
s.outputLen = 48
case sha3_512:
s.rate = 72
s.outputLen = 64
case shake_128:
s.rate = 168
case shake_256:
s.rate = 136
default:
panic("sha3: unrecognized function code")
}
// limit s.buf size to a multiple of s.rate
s.resetBuf()
return &s
}
func (s *asmState) clone() *asmState {
c := *s
c.buf = c.storage[:len(s.buf):cap(s.buf)]
return &c
}
// copyIntoBuf copies b into buf. It will panic if there is not enough space to
// store all of b.
func (s *asmState) copyIntoBuf(b []byte) {
bufLen := len(s.buf)
s.buf = s.buf[:len(s.buf)+len(b)]
copy(s.buf[bufLen:], b)
}
// resetBuf points buf at storage, sets the length to 0 and sets cap to be a
// multiple of the rate.
func (s *asmState) resetBuf() {
max := (cap(s.storage) / s.rate) * s.rate
s.buf = s.storage[:0:max]
}
// Write (via the embedded io.Writer interface) adds more data to the running hash.
// It never returns an error.
func (s *asmState) Write(b []byte) (int, error) {
if s.state != spongeAbsorbing {
panic("sha3: write to sponge after read")
}
length := len(b)
for len(b) > 0 {
if len(s.buf) == 0 && len(b) >= cap(s.buf) {
// Hash the data directly and push any remaining bytes
// into the buffer.
remainder := len(b) % s.rate
kimd(s.function, &s.a, b[:len(b)-remainder])
if remainder != 0 {
s.copyIntoBuf(b[len(b)-remainder:])
}
return length, nil
}
if len(s.buf) == cap(s.buf) {
// flush the buffer
kimd(s.function, &s.a, s.buf)
s.buf = s.buf[:0]
}
// copy as much as we can into the buffer
n := len(b)
if len(b) > cap(s.buf)-len(s.buf) {
n = cap(s.buf) - len(s.buf)
}
s.copyIntoBuf(b[:n])
b = b[n:]
}
return length, nil
}
// Read squeezes an arbitrary number of bytes from the sponge.
func (s *asmState) Read(out []byte) (n int, err error) {
n = len(out)
// need to pad if we were absorbing
if s.state == spongeAbsorbing {
s.state = spongeSqueezing
// write hash directly into out if possible
if len(out)%s.rate == 0 {
klmd(s.function, &s.a, out, s.buf) // len(out) may be 0
s.buf = s.buf[:0]
return
}
// write hash into buffer
max := cap(s.buf)
if max > len(out) {
max = (len(out)/s.rate)*s.rate + s.rate
}
klmd(s.function, &s.a, s.buf[:max], s.buf)
s.buf = s.buf[:max]
}
for len(out) > 0 {
// flush the buffer
if len(s.buf) != 0 {
c := copy(out, s.buf)
out = out[c:]
s.buf = s.buf[c:]
continue
}
// write hash directly into out if possible
if len(out)%s.rate == 0 {
klmd(s.function|nopad, &s.a, out, nil)
return
}
// write hash into buffer
s.resetBuf()
if cap(s.buf) > len(out) {
s.buf = s.buf[:(len(out)/s.rate)*s.rate+s.rate]
}
klmd(s.function|nopad, &s.a, s.buf, nil)
}
return
}
// Sum appends the current hash to b and returns the resulting slice.
// It does not change the underlying hash state.
func (s *asmState) Sum(b []byte) []byte {
if s.outputLen == 0 {
panic("sha3: cannot call Sum on SHAKE functions")
}
// Copy the state to preserve the original.
a := s.a
// Hash the buffer. Note that we don't clear it because we
// aren't updating the state.
klmd(s.function, &a, nil, s.buf)
return append(b, a[:s.outputLen]...)
}
// Reset resets the Hash to its initial state.
func (s *asmState) Reset() {
for i := range s.a {
s.a[i] = 0
}
s.resetBuf()
s.state = spongeAbsorbing
}
// Size returns the number of bytes Sum will return.
func (s *asmState) Size() int {
return s.outputLen
}
// BlockSize returns the hash's underlying block size.
// The Write method must be able to accept any amount
// of data, but it may operate more efficiently if all writes
// are a multiple of the block size.
func (s *asmState) BlockSize() int {
return s.rate
}
// Clone returns a copy of the ShakeHash in its current state.
func (s *asmState) Clone() ShakeHash {
return s.clone()
}
// new224Asm returns an assembly implementation of SHA3-224 if available,
// otherwise it returns nil.
func new224Asm() hash.Hash {
if cpu.S390X.HasSHA3 {
return newAsmState(sha3_224)
}
return nil
}
// new256Asm returns an assembly implementation of SHA3-256 if available,
// otherwise it returns nil.
func new256Asm() hash.Hash {
if cpu.S390X.HasSHA3 {
return newAsmState(sha3_256)
}
return nil
}
// new384Asm returns an assembly implementation of SHA3-384 if available,
// otherwise it returns nil.
func new384Asm() hash.Hash {
if cpu.S390X.HasSHA3 {
return newAsmState(sha3_384)
}
return nil
}
// new512Asm returns an assembly implementation of SHA3-512 if available,
// otherwise it returns nil.
func new512Asm() hash.Hash {
if cpu.S390X.HasSHA3 {
return newAsmState(sha3_512)
}
return nil
}
// newShake128Asm returns an assembly implementation of SHAKE-128 if available,
// otherwise it returns nil.
func newShake128Asm() ShakeHash {
if cpu.S390X.HasSHA3 {
return newAsmState(shake_128)
}
return nil
}
// newShake256Asm returns an assembly implementation of SHAKE-256 if available,
// otherwise it returns nil.
func newShake256Asm() ShakeHash {
if cpu.S390X.HasSHA3 {
return newAsmState(shake_256)
}
return nil
}

34
vendor/golang.org/x/crypto/sha3/sha3_s390x.s generated vendored Normal file
View File

@@ -0,0 +1,34 @@
// Copyright 2017 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.
//go:build gc && !purego
// +build gc,!purego
#include "textflag.h"
// func kimd(function code, chain *[200]byte, src []byte)
TEXT ·kimd(SB), NOFRAME|NOSPLIT, $0-40
MOVD function+0(FP), R0
MOVD chain+8(FP), R1
LMG src+16(FP), R2, R3 // R2=base, R3=len
continue:
WORD $0xB93E0002 // KIMD --, R2
BVS continue // continue if interrupted
MOVD $0, R0 // reset R0 for pre-go1.8 compilers
RET
// func klmd(function code, chain *[200]byte, dst, src []byte)
TEXT ·klmd(SB), NOFRAME|NOSPLIT, $0-64
// TODO: SHAKE support
MOVD function+0(FP), R0
MOVD chain+8(FP), R1
LMG dst+16(FP), R2, R3 // R2=base, R3=len
LMG src+40(FP), R4, R5 // R4=base, R5=len
continue:
WORD $0xB93F0024 // KLMD R2, R4
BVS continue // continue if interrupted
MOVD $0, R0 // reset R0 for pre-go1.8 compilers
RET

173
vendor/golang.org/x/crypto/sha3/shake.go generated vendored Normal file
View File

@@ -0,0 +1,173 @@
// Copyright 2014 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 sha3
// This file defines the ShakeHash interface, and provides
// functions for creating SHAKE and cSHAKE instances, as well as utility
// functions for hashing bytes to arbitrary-length output.
//
//
// SHAKE implementation is based on FIPS PUB 202 [1]
// cSHAKE implementations is based on NIST SP 800-185 [2]
//
// [1] https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.202.pdf
// [2] https://doi.org/10.6028/NIST.SP.800-185
import (
"encoding/binary"
"io"
)
// ShakeHash defines the interface to hash functions that
// support arbitrary-length output.
type ShakeHash interface {
// Write absorbs more data into the hash's state. It panics if input is
// written to it after output has been read from it.
io.Writer
// Read reads more output from the hash; reading affects the hash's
// state. (ShakeHash.Read is thus very different from Hash.Sum)
// It never returns an error.
io.Reader
// Clone returns a copy of the ShakeHash in its current state.
Clone() ShakeHash
// Reset resets the ShakeHash to its initial state.
Reset()
}
// cSHAKE specific context
type cshakeState struct {
*state // SHA-3 state context and Read/Write operations
// initBlock is the cSHAKE specific initialization set of bytes. It is initialized
// by newCShake function and stores concatenation of N followed by S, encoded
// by the method specified in 3.3 of [1].
// It is stored here in order for Reset() to be able to put context into
// initial state.
initBlock []byte
}
// Consts for configuring initial SHA-3 state
const (
dsbyteShake = 0x1f
dsbyteCShake = 0x04
rate128 = 168
rate256 = 136
)
func bytepad(input []byte, w int) []byte {
// leftEncode always returns max 9 bytes
buf := make([]byte, 0, 9+len(input)+w)
buf = append(buf, leftEncode(uint64(w))...)
buf = append(buf, input...)
padlen := w - (len(buf) % w)
return append(buf, make([]byte, padlen)...)
}
func leftEncode(value uint64) []byte {
var b [9]byte
binary.BigEndian.PutUint64(b[1:], value)
// Trim all but last leading zero bytes
i := byte(1)
for i < 8 && b[i] == 0 {
i++
}
// Prepend number of encoded bytes
b[i-1] = 9 - i
return b[i-1:]
}
func newCShake(N, S []byte, rate int, dsbyte byte) ShakeHash {
c := cshakeState{state: &state{rate: rate, dsbyte: dsbyte}}
// leftEncode returns max 9 bytes
c.initBlock = make([]byte, 0, 9*2+len(N)+len(S))
c.initBlock = append(c.initBlock, leftEncode(uint64(len(N)*8))...)
c.initBlock = append(c.initBlock, N...)
c.initBlock = append(c.initBlock, leftEncode(uint64(len(S)*8))...)
c.initBlock = append(c.initBlock, S...)
c.Write(bytepad(c.initBlock, c.rate))
return &c
}
// Reset resets the hash to initial state.
func (c *cshakeState) Reset() {
c.state.Reset()
c.Write(bytepad(c.initBlock, c.rate))
}
// Clone returns copy of a cSHAKE context within its current state.
func (c *cshakeState) Clone() ShakeHash {
b := make([]byte, len(c.initBlock))
copy(b, c.initBlock)
return &cshakeState{state: c.clone(), initBlock: b}
}
// Clone returns copy of SHAKE context within its current state.
func (c *state) Clone() ShakeHash {
return c.clone()
}
// NewShake128 creates a new SHAKE128 variable-output-length ShakeHash.
// Its generic security strength is 128 bits against all attacks if at
// least 32 bytes of its output are used.
func NewShake128() ShakeHash {
if h := newShake128Asm(); h != nil {
return h
}
return &state{rate: rate128, dsbyte: dsbyteShake}
}
// NewShake256 creates a new SHAKE256 variable-output-length ShakeHash.
// Its generic security strength is 256 bits against all attacks if
// at least 64 bytes of its output are used.
func NewShake256() ShakeHash {
if h := newShake256Asm(); h != nil {
return h
}
return &state{rate: rate256, dsbyte: dsbyteShake}
}
// NewCShake128 creates a new instance of cSHAKE128 variable-output-length ShakeHash,
// a customizable variant of SHAKE128.
// N is used to define functions based on cSHAKE, it can be empty when plain cSHAKE is
// desired. S is a customization byte string used for domain separation - two cSHAKE
// computations on same input with different S yield unrelated outputs.
// When N and S are both empty, this is equivalent to NewShake128.
func NewCShake128(N, S []byte) ShakeHash {
if len(N) == 0 && len(S) == 0 {
return NewShake128()
}
return newCShake(N, S, rate128, dsbyteCShake)
}
// NewCShake256 creates a new instance of cSHAKE256 variable-output-length ShakeHash,
// a customizable variant of SHAKE256.
// N is used to define functions based on cSHAKE, it can be empty when plain cSHAKE is
// desired. S is a customization byte string used for domain separation - two cSHAKE
// computations on same input with different S yield unrelated outputs.
// When N and S are both empty, this is equivalent to NewShake256.
func NewCShake256(N, S []byte) ShakeHash {
if len(N) == 0 && len(S) == 0 {
return NewShake256()
}
return newCShake(N, S, rate256, dsbyteCShake)
}
// ShakeSum128 writes an arbitrary-length digest of data into hash.
func ShakeSum128(hash, data []byte) {
h := NewShake128()
h.Write(data)
h.Read(hash)
}
// ShakeSum256 writes an arbitrary-length digest of data into hash.
func ShakeSum256(hash, data []byte) {
h := NewShake256()
h.Write(data)
h.Read(hash)
}

20
vendor/golang.org/x/crypto/sha3/shake_generic.go generated vendored Normal file
View File

@@ -0,0 +1,20 @@
// Copyright 2017 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.
//go:build !gc || purego || !s390x
// +build !gc purego !s390x
package sha3
// newShake128Asm returns an assembly implementation of SHAKE-128 if available,
// otherwise it returns nil.
func newShake128Asm() ShakeHash {
return nil
}
// newShake256Asm returns an assembly implementation of SHAKE-256 if available,
// otherwise it returns nil.
func newShake256Asm() ShakeHash {
return nil
}

24
vendor/golang.org/x/crypto/sha3/xor.go generated vendored Normal file
View File

@@ -0,0 +1,24 @@
// 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.
//go:build (!amd64 && !386 && !ppc64le) || purego
// +build !amd64,!386,!ppc64le purego
package sha3
// A storageBuf is an aligned array of maxRate bytes.
type storageBuf [maxRate]byte
func (b *storageBuf) asBytes() *[maxRate]byte {
return (*[maxRate]byte)(b)
}
var (
xorIn = xorInGeneric
copyOut = copyOutGeneric
xorInUnaligned = xorInGeneric
copyOutUnaligned = copyOutGeneric
)
const xorImplementationUnaligned = "generic"

28
vendor/golang.org/x/crypto/sha3/xor_generic.go generated vendored Normal file
View File

@@ -0,0 +1,28 @@
// 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.
package sha3
import "encoding/binary"
// xorInGeneric xors the bytes in buf into the state; it
// makes no non-portable assumptions about memory layout
// or alignment.
func xorInGeneric(d *state, buf []byte) {
n := len(buf) / 8
for i := 0; i < n; i++ {
a := binary.LittleEndian.Uint64(buf)
d.a[i] ^= a
buf = buf[8:]
}
}
// copyOutGeneric copies uint64s to a byte buffer.
func copyOutGeneric(d *state, b []byte) {
for i := 0; len(b) >= 8; i++ {
binary.LittleEndian.PutUint64(b, d.a[i])
b = b[8:]
}
}

68
vendor/golang.org/x/crypto/sha3/xor_unaligned.go generated vendored Normal file
View File

@@ -0,0 +1,68 @@
// 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.
//go:build (amd64 || 386 || ppc64le) && !purego
// +build amd64 386 ppc64le
// +build !purego
package sha3
import "unsafe"
// A storageBuf is an aligned array of maxRate bytes.
type storageBuf [maxRate / 8]uint64
func (b *storageBuf) asBytes() *[maxRate]byte {
return (*[maxRate]byte)(unsafe.Pointer(b))
}
// xorInUnaligned uses unaligned reads and writes to update d.a to contain d.a
// XOR buf.
func xorInUnaligned(d *state, buf []byte) {
n := len(buf)
bw := (*[maxRate / 8]uint64)(unsafe.Pointer(&buf[0]))[: n/8 : n/8]
if n >= 72 {
d.a[0] ^= bw[0]
d.a[1] ^= bw[1]
d.a[2] ^= bw[2]
d.a[3] ^= bw[3]
d.a[4] ^= bw[4]
d.a[5] ^= bw[5]
d.a[6] ^= bw[6]
d.a[7] ^= bw[7]
d.a[8] ^= bw[8]
}
if n >= 104 {
d.a[9] ^= bw[9]
d.a[10] ^= bw[10]
d.a[11] ^= bw[11]
d.a[12] ^= bw[12]
}
if n >= 136 {
d.a[13] ^= bw[13]
d.a[14] ^= bw[14]
d.a[15] ^= bw[15]
d.a[16] ^= bw[16]
}
if n >= 144 {
d.a[17] ^= bw[17]
}
if n >= 168 {
d.a[18] ^= bw[18]
d.a[19] ^= bw[19]
d.a[20] ^= bw[20]
}
}
func copyOutUnaligned(d *state, buf []byte) {
ab := (*[maxRate]uint8)(unsafe.Pointer(&d.a[0]))
copy(buf, ab[:])
}
var (
xorIn = xorInUnaligned
copyOut = copyOutUnaligned
)
const xorImplementationUnaligned = "unaligned"