Godeps <- golang.org/x/crypto/nacl etc

Godeps/Godeps.json

@@ -75,6 +75,30 @@
 			"Comment": "v2.3-33-g105a1be",
 			"Rev": "105a1beefe3379227a0d61733a29fd2c34f7080c"
 		},
+		{
+			"ImportPath": "golang.org/x/crypto/curve25519",
+			"Rev": "02a186af8b62cb007f392270669b91be5527d39c"
+		},
+		{
+			"ImportPath": "golang.org/x/crypto/nacl/box",
+			"Rev": "02a186af8b62cb007f392270669b91be5527d39c"
+		},
+		{
+			"ImportPath": "golang.org/x/crypto/nacl/secretbox",
+			"Rev": "02a186af8b62cb007f392270669b91be5527d39c"
+		},
+		{
+			"ImportPath": "golang.org/x/crypto/poly1305",
+			"Rev": "02a186af8b62cb007f392270669b91be5527d39c"
+		},
+		{
+			"ImportPath": "golang.org/x/crypto/ripemd160",
+			"Rev": "02a186af8b62cb007f392270669b91be5527d39c"
+		},
+		{
+			"ImportPath": "golang.org/x/crypto/salsa20/salsa",
+			"Rev": "02a186af8b62cb007f392270669b91be5527d39c"
+		},
 		{
 			"ImportPath": "gopkg.in/fatih/set.v0",
 			"Comment": "v0.1.0-3-g27c4092",

Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/session_util.go

@@ -14,10 +14,11 @@ import (
 	"github.com/tendermint/tendermint/Godeps/_workspace/src/github.com/syndtr/goleveldb/leveldb/storage"
 )
 
+// Logging.
+
 type dropper struct {
 	s    *session
-	file storage. // Logging.
-	File
+	file storage.File
 }
 
 func (d dropper) Drop(err error) {

Godeps/_workspace/src/golang.org/x/crypto/curve25519/const_amd64.s

@@ -0,0 +1,20 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// +build amd64,!gccgo,!appengine
+
+DATA ·REDMASK51(SB)/8, $0x0007FFFFFFFFFFFF
+GLOBL ·REDMASK51(SB), 8, $8
+
+DATA ·_121666_213(SB)/8, $996687872
+GLOBL ·_121666_213(SB), 8, $8
+
+DATA ·_2P0(SB)/8, $0xFFFFFFFFFFFDA
+GLOBL ·_2P0(SB), 8, $8
+
+DATA ·_2P1234(SB)/8, $0xFFFFFFFFFFFFE
+GLOBL ·_2P1234(SB), 8, $8

Godeps/_workspace/src/golang.org/x/crypto/curve25519/cswap_amd64.s

@@ -0,0 +1,88 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// +build amd64,!gccgo,!appengine
+
+// func cswap(inout *[5]uint64, v uint64)
+TEXT ·cswap(SB),7,$0
+	MOVQ inout+0(FP),DI
+	MOVQ v+8(FP),SI
+
+	CMPQ SI,$1
+	MOVQ 0(DI),SI
+	MOVQ 80(DI),DX
+	MOVQ 8(DI),CX
+	MOVQ 88(DI),R8
+	MOVQ SI,R9
+	CMOVQEQ DX,SI
+	CMOVQEQ R9,DX
+	MOVQ CX,R9
+	CMOVQEQ R8,CX
+	CMOVQEQ R9,R8
+	MOVQ SI,0(DI)
+	MOVQ DX,80(DI)
+	MOVQ CX,8(DI)
+	MOVQ R8,88(DI)
+	MOVQ 16(DI),SI
+	MOVQ 96(DI),DX
+	MOVQ 24(DI),CX
+	MOVQ 104(DI),R8
+	MOVQ SI,R9
+	CMOVQEQ DX,SI
+	CMOVQEQ R9,DX
+	MOVQ CX,R9
+	CMOVQEQ R8,CX
+	CMOVQEQ R9,R8
+	MOVQ SI,16(DI)
+	MOVQ DX,96(DI)
+	MOVQ CX,24(DI)
+	MOVQ R8,104(DI)
+	MOVQ 32(DI),SI
+	MOVQ 112(DI),DX
+	MOVQ 40(DI),CX
+	MOVQ 120(DI),R8
+	MOVQ SI,R9
+	CMOVQEQ DX,SI
+	CMOVQEQ R9,DX
+	MOVQ CX,R9
+	CMOVQEQ R8,CX
+	CMOVQEQ R9,R8
+	MOVQ SI,32(DI)
+	MOVQ DX,112(DI)
+	MOVQ CX,40(DI)
+	MOVQ R8,120(DI)
+	MOVQ 48(DI),SI
+	MOVQ 128(DI),DX
+	MOVQ 56(DI),CX
+	MOVQ 136(DI),R8
+	MOVQ SI,R9
+	CMOVQEQ DX,SI
+	CMOVQEQ R9,DX
+	MOVQ CX,R9
+	CMOVQEQ R8,CX
+	CMOVQEQ R9,R8
+	MOVQ SI,48(DI)
+	MOVQ DX,128(DI)
+	MOVQ CX,56(DI)
+	MOVQ R8,136(DI)
+	MOVQ 64(DI),SI
+	MOVQ 144(DI),DX
+	MOVQ 72(DI),CX
+	MOVQ 152(DI),R8
+	MOVQ SI,R9
+	CMOVQEQ DX,SI
+	CMOVQEQ R9,DX
+	MOVQ CX,R9
+	CMOVQEQ R8,CX
+	CMOVQEQ R9,R8
+	MOVQ SI,64(DI)
+	MOVQ DX,144(DI)
+	MOVQ CX,72(DI)
+	MOVQ R8,152(DI)
+	MOVQ DI,AX
+	MOVQ SI,DX
+	RET

Godeps/_workspace/src/golang.org/x/crypto/curve25519/curve25519.go

@@ -0,0 +1,841 @@
+// Copyright 2013 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.
+
+// We have a implementation in amd64 assembly so this code is only run on
+// non-amd64 platforms. The amd64 assembly does not support gccgo.
+// +build !amd64 gccgo appengine
+
+package curve25519
+
+// This code is a port of the public domain, "ref10" implementation of
+// curve25519 from SUPERCOP 20130419 by D. J. Bernstein.
+
+// fieldElement represents an element of the field GF(2^255 - 19). An element
+// t, entries t[0]...t[9], represents the integer t[0]+2^26 t[1]+2^51 t[2]+2^77
+// t[3]+2^102 t[4]+...+2^230 t[9]. Bounds on each t[i] vary depending on
+// context.
+type fieldElement [10]int32
+
+func feZero(fe *fieldElement) {
+	for i := range fe {
+		fe[i] = 0
+	}
+}
+
+func feOne(fe *fieldElement) {
+	feZero(fe)
+	fe[0] = 1
+}
+
+func feAdd(dst, a, b *fieldElement) {
+	for i := range dst {
+		dst[i] = a[i] + b[i]
+	}
+}
+
+func feSub(dst, a, b *fieldElement) {
+	for i := range dst {
+		dst[i] = a[i] - b[i]
+	}
+}
+
+func feCopy(dst, src *fieldElement) {
+	for i := range dst {
+		dst[i] = src[i]
+	}
+}
+
+// feCSwap replaces (f,g) with (g,f) if b == 1; replaces (f,g) with (f,g) if b == 0.
+//
+// Preconditions: b in {0,1}.
+func feCSwap(f, g *fieldElement, b int32) {
+	var x fieldElement
+	b = -b
+	for i := range x {
+		x[i] = b & (f[i] ^ g[i])
+	}
+
+	for i := range f {
+		f[i] ^= x[i]
+	}
+	for i := range g {
+		g[i] ^= x[i]
+	}
+}
+
+// load3 reads a 24-bit, little-endian value from in.
+func load3(in []byte) int64 {
+	var r int64
+	r = int64(in[0])
+	r |= int64(in[1]) << 8
+	r |= int64(in[2]) << 16
+	return r
+}
+
+// load4 reads a 32-bit, little-endian value from in.
+func load4(in []byte) int64 {
+	var r int64
+	r = int64(in[0])
+	r |= int64(in[1]) << 8
+	r |= int64(in[2]) << 16
+	r |= int64(in[3]) << 24
+	return r
+}
+
+func feFromBytes(dst *fieldElement, src *[32]byte) {
+	h0 := load4(src[:])
+	h1 := load3(src[4:]) << 6
+	h2 := load3(src[7:]) << 5
+	h3 := load3(src[10:]) << 3
+	h4 := load3(src[13:]) << 2
+	h5 := load4(src[16:])
+	h6 := load3(src[20:]) << 7
+	h7 := load3(src[23:]) << 5
+	h8 := load3(src[26:]) << 4
+	h9 := load3(src[29:]) << 2
+
+	var carry [10]int64
+	carry[9] = (h9 + 1<<24) >> 25
+	h0 += carry[9] * 19
+	h9 -= carry[9] << 25
+	carry[1] = (h1 + 1<<24) >> 25
+	h2 += carry[1]
+	h1 -= carry[1] << 25
+	carry[3] = (h3 + 1<<24) >> 25
+	h4 += carry[3]
+	h3 -= carry[3] << 25
+	carry[5] = (h5 + 1<<24) >> 25
+	h6 += carry[5]
+	h5 -= carry[5] << 25
+	carry[7] = (h7 + 1<<24) >> 25
+	h8 += carry[7]
+	h7 -= carry[7] << 25
+
+	carry[0] = (h0 + 1<<25) >> 26
+	h1 += carry[0]
+	h0 -= carry[0] << 26
+	carry[2] = (h2 + 1<<25) >> 26
+	h3 += carry[2]
+	h2 -= carry[2] << 26
+	carry[4] = (h4 + 1<<25) >> 26
+	h5 += carry[4]
+	h4 -= carry[4] << 26
+	carry[6] = (h6 + 1<<25) >> 26
+	h7 += carry[6]
+	h6 -= carry[6] << 26
+	carry[8] = (h8 + 1<<25) >> 26
+	h9 += carry[8]
+	h8 -= carry[8] << 26
+
+	dst[0] = int32(h0)
+	dst[1] = int32(h1)
+	dst[2] = int32(h2)
+	dst[3] = int32(h3)
+	dst[4] = int32(h4)
+	dst[5] = int32(h5)
+	dst[6] = int32(h6)
+	dst[7] = int32(h7)
+	dst[8] = int32(h8)
+	dst[9] = int32(h9)
+}
+
+// feToBytes marshals h to s.
+// Preconditions:
+//   |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
+//
+// Write p=2^255-19; q=floor(h/p).
+// Basic claim: q = floor(2^(-255)(h + 19 2^(-25)h9 + 2^(-1))).
+//
+// Proof:
+//   Have |h|<=p so |q|<=1 so |19^2 2^(-255) q|<1/4.
+//   Also have |h-2^230 h9|<2^230 so |19 2^(-255)(h-2^230 h9)|<1/4.
+//
+//   Write y=2^(-1)-19^2 2^(-255)q-19 2^(-255)(h-2^230 h9).
+//   Then 0<y<1.
+//
+//   Write r=h-pq.
+//   Have 0<=r<=p-1=2^255-20.
+//   Thus 0<=r+19(2^-255)r<r+19(2^-255)2^255<=2^255-1.
+//
+//   Write x=r+19(2^-255)r+y.
+//   Then 0<x<2^255 so floor(2^(-255)x) = 0 so floor(q+2^(-255)x) = q.
+//
+//   Have q+2^(-255)x = 2^(-255)(h + 19 2^(-25) h9 + 2^(-1))
+//   so floor(2^(-255)(h + 19 2^(-25) h9 + 2^(-1))) = q.
+func feToBytes(s *[32]byte, h *fieldElement) {
+	var carry [10]int32
+
+	q := (19*h[9] + (1 << 24)) >> 25
+	q = (h[0] + q) >> 26
+	q = (h[1] + q) >> 25
+	q = (h[2] + q) >> 26
+	q = (h[3] + q) >> 25
+	q = (h[4] + q) >> 26
+	q = (h[5] + q) >> 25
+	q = (h[6] + q) >> 26
+	q = (h[7] + q) >> 25
+	q = (h[8] + q) >> 26
+	q = (h[9] + q) >> 25
+
+	// Goal: Output h-(2^255-19)q, which is between 0 and 2^255-20.
+	h[0] += 19 * q
+	// Goal: Output h-2^255 q, which is between 0 and 2^255-20.
+
+	carry[0] = h[0] >> 26
+	h[1] += carry[0]
+	h[0] -= carry[0] << 26
+	carry[1] = h[1] >> 25
+	h[2] += carry[1]
+	h[1] -= carry[1] << 25
+	carry[2] = h[2] >> 26
+	h[3] += carry[2]
+	h[2] -= carry[2] << 26
+	carry[3] = h[3] >> 25
+	h[4] += carry[3]
+	h[3] -= carry[3] << 25
+	carry[4] = h[4] >> 26
+	h[5] += carry[4]
+	h[4] -= carry[4] << 26
+	carry[5] = h[5] >> 25
+	h[6] += carry[5]
+	h[5] -= carry[5] << 25
+	carry[6] = h[6] >> 26
+	h[7] += carry[6]
+	h[6] -= carry[6] << 26
+	carry[7] = h[7] >> 25
+	h[8] += carry[7]
+	h[7] -= carry[7] << 25
+	carry[8] = h[8] >> 26
+	h[9] += carry[8]
+	h[8] -= carry[8] << 26
+	carry[9] = h[9] >> 25
+	h[9] -= carry[9] << 25
+	// h10 = carry9
+
+	// Goal: Output h[0]+...+2^255 h10-2^255 q, which is between 0 and 2^255-20.
+	// Have h[0]+...+2^230 h[9] between 0 and 2^255-1;
+	// evidently 2^255 h10-2^255 q = 0.
+	// Goal: Output h[0]+...+2^230 h[9].
+
+	s[0] = byte(h[0] >> 0)
+	s[1] = byte(h[0] >> 8)
+	s[2] = byte(h[0] >> 16)
+	s[3] = byte((h[0] >> 24) | (h[1] << 2))
+	s[4] = byte(h[1] >> 6)
+	s[5] = byte(h[1] >> 14)
+	s[6] = byte((h[1] >> 22) | (h[2] << 3))
+	s[7] = byte(h[2] >> 5)
+	s[8] = byte(h[2] >> 13)
+	s[9] = byte((h[2] >> 21) | (h[3] << 5))
+	s[10] = byte(h[3] >> 3)
+	s[11] = byte(h[3] >> 11)
+	s[12] = byte((h[3] >> 19) | (h[4] << 6))
+	s[13] = byte(h[4] >> 2)
+	s[14] = byte(h[4] >> 10)
+	s[15] = byte(h[4] >> 18)
+	s[16] = byte(h[5] >> 0)
+	s[17] = byte(h[5] >> 8)
+	s[18] = byte(h[5] >> 16)
+	s[19] = byte((h[5] >> 24) | (h[6] << 1))
+	s[20] = byte(h[6] >> 7)
+	s[21] = byte(h[6] >> 15)
+	s[22] = byte((h[6] >> 23) | (h[7] << 3))
+	s[23] = byte(h[7] >> 5)
+	s[24] = byte(h[7] >> 13)
+	s[25] = byte((h[7] >> 21) | (h[8] << 4))
+	s[26] = byte(h[8] >> 4)
+	s[27] = byte(h[8] >> 12)
+	s[28] = byte((h[8] >> 20) | (h[9] << 6))
+	s[29] = byte(h[9] >> 2)
+	s[30] = byte(h[9] >> 10)
+	s[31] = byte(h[9] >> 18)
+}
+
+// feMul calculates h = f * g
+// Can overlap h with f or g.
+//
+// Preconditions:
+//    |f| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+//    |g| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+//
+// Postconditions:
+//    |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
+//
+// Notes on implementation strategy:
+//
+// Using schoolbook multiplication.
+// Karatsuba would save a little in some cost models.
+//
+// Most multiplications by 2 and 19 are 32-bit precomputations;
+// cheaper than 64-bit postcomputations.
+//
+// There is one remaining multiplication by 19 in the carry chain;
+// one *19 precomputation can be merged into this,
+// but the resulting data flow is considerably less clean.
+//
+// There are 12 carries below.
+// 10 of them are 2-way parallelizable and vectorizable.
+// Can get away with 11 carries, but then data flow is much deeper.
+//
+// With tighter constraints on inputs can squeeze carries into int32.
+func feMul(h, f, g *fieldElement) {
+	f0 := f[0]
+	f1 := f[1]
+	f2 := f[2]
+	f3 := f[3]
+	f4 := f[4]
+	f5 := f[5]
+	f6 := f[6]
+	f7 := f[7]
+	f8 := f[8]
+	f9 := f[9]
+	g0 := g[0]
+	g1 := g[1]
+	g2 := g[2]
+	g3 := g[3]
+	g4 := g[4]
+	g5 := g[5]
+	g6 := g[6]
+	g7 := g[7]
+	g8 := g[8]
+	g9 := g[9]
+	g1_19 := 19 * g1 // 1.4*2^29
+	g2_19 := 19 * g2 // 1.4*2^30; still ok
+	g3_19 := 19 * g3
+	g4_19 := 19 * g4
+	g5_19 := 19 * g5
+	g6_19 := 19 * g6
+	g7_19 := 19 * g7
+	g8_19 := 19 * g8
+	g9_19 := 19 * g9
+	f1_2 := 2 * f1
+	f3_2 := 2 * f3
+	f5_2 := 2 * f5
+	f7_2 := 2 * f7
+	f9_2 := 2 * f9
+	f0g0 := int64(f0) * int64(g0)
+	f0g1 := int64(f0) * int64(g1)
+	f0g2 := int64(f0) * int64(g2)
+	f0g3 := int64(f0) * int64(g3)
+	f0g4 := int64(f0) * int64(g4)
+	f0g5 := int64(f0) * int64(g5)
+	f0g6 := int64(f0) * int64(g6)
+	f0g7 := int64(f0) * int64(g7)
+	f0g8 := int64(f0) * int64(g8)
+	f0g9 := int64(f0) * int64(g9)
+	f1g0 := int64(f1) * int64(g0)
+	f1g1_2 := int64(f1_2) * int64(g1)
+	f1g2 := int64(f1) * int64(g2)
+	f1g3_2 := int64(f1_2) * int64(g3)
+	f1g4 := int64(f1) * int64(g4)
+	f1g5_2 := int64(f1_2) * int64(g5)
+	f1g6 := int64(f1) * int64(g6)
+	f1g7_2 := int64(f1_2) * int64(g7)
+	f1g8 := int64(f1) * int64(g8)
+	f1g9_38 := int64(f1_2) * int64(g9_19)
+	f2g0 := int64(f2) * int64(g0)
+	f2g1 := int64(f2) * int64(g1)
+	f2g2 := int64(f2) * int64(g2)
+	f2g3 := int64(f2) * int64(g3)
+	f2g4 := int64(f2) * int64(g4)
+	f2g5 := int64(f2) * int64(g5)
+	f2g6 := int64(f2) * int64(g6)
+	f2g7 := int64(f2) * int64(g7)
+	f2g8_19 := int64(f2) * int64(g8_19)
+	f2g9_19 := int64(f2) * int64(g9_19)
+	f3g0 := int64(f3) * int64(g0)
+	f3g1_2 := int64(f3_2) * int64(g1)
+	f3g2 := int64(f3) * int64(g2)
+	f3g3_2 := int64(f3_2) * int64(g3)
+	f3g4 := int64(f3) * int64(g4)
+	f3g5_2 := int64(f3_2) * int64(g5)
+	f3g6 := int64(f3) * int64(g6)
+	f3g7_38 := int64(f3_2) * int64(g7_19)
+	f3g8_19 := int64(f3) * int64(g8_19)
+	f3g9_38 := int64(f3_2) * int64(g9_19)
+	f4g0 := int64(f4) * int64(g0)
+	f4g1 := int64(f4) * int64(g1)
+	f4g2 := int64(f4) * int64(g2)
+	f4g3 := int64(f4) * int64(g3)
+	f4g4 := int64(f4) * int64(g4)
+	f4g5 := int64(f4) * int64(g5)
+	f4g6_19 := int64(f4) * int64(g6_19)
+	f4g7_19 := int64(f4) * int64(g7_19)
+	f4g8_19 := int64(f4) * int64(g8_19)
+	f4g9_19 := int64(f4) * int64(g9_19)
+	f5g0 := int64(f5) * int64(g0)
+	f5g1_2 := int64(f5_2) * int64(g1)
+	f5g2 := int64(f5) * int64(g2)
+	f5g3_2 := int64(f5_2) * int64(g3)
+	f5g4 := int64(f5) * int64(g4)
+	f5g5_38 := int64(f5_2) * int64(g5_19)
+	f5g6_19 := int64(f5) * int64(g6_19)
+	f5g7_38 := int64(f5_2) * int64(g7_19)
+	f5g8_19 := int64(f5) * int64(g8_19)
+	f5g9_38 := int64(f5_2) * int64(g9_19)
+	f6g0 := int64(f6) * int64(g0)
+	f6g1 := int64(f6) * int64(g1)
+	f6g2 := int64(f6) * int64(g2)
+	f6g3 := int64(f6) * int64(g3)
+	f6g4_19 := int64(f6) * int64(g4_19)
+	f6g5_19 := int64(f6) * int64(g5_19)
+	f6g6_19 := int64(f6) * int64(g6_19)
+	f6g7_19 := int64(f6) * int64(g7_19)
+	f6g8_19 := int64(f6) * int64(g8_19)
+	f6g9_19 := int64(f6) * int64(g9_19)
+	f7g0 := int64(f7) * int64(g0)
+	f7g1_2 := int64(f7_2) * int64(g1)
+	f7g2 := int64(f7) * int64(g2)
+	f7g3_38 := int64(f7_2) * int64(g3_19)
+	f7g4_19 := int64(f7) * int64(g4_19)
+	f7g5_38 := int64(f7_2) * int64(g5_19)
+	f7g6_19 := int64(f7) * int64(g6_19)
+	f7g7_38 := int64(f7_2) * int64(g7_19)
+	f7g8_19 := int64(f7) * int64(g8_19)
+	f7g9_38 := int64(f7_2) * int64(g9_19)
+	f8g0 := int64(f8) * int64(g0)
+	f8g1 := int64(f8) * int64(g1)
+	f8g2_19 := int64(f8) * int64(g2_19)
+	f8g3_19 := int64(f8) * int64(g3_19)
+	f8g4_19 := int64(f8) * int64(g4_19)
+	f8g5_19 := int64(f8) * int64(g5_19)
+	f8g6_19 := int64(f8) * int64(g6_19)
+	f8g7_19 := int64(f8) * int64(g7_19)
+	f8g8_19 := int64(f8) * int64(g8_19)
+	f8g9_19 := int64(f8) * int64(g9_19)
+	f9g0 := int64(f9) * int64(g0)
+	f9g1_38 := int64(f9_2) * int64(g1_19)
+	f9g2_19 := int64(f9) * int64(g2_19)
+	f9g3_38 := int64(f9_2) * int64(g3_19)
+	f9g4_19 := int64(f9) * int64(g4_19)
+	f9g5_38 := int64(f9_2) * int64(g5_19)
+	f9g6_19 := int64(f9) * int64(g6_19)
+	f9g7_38 := int64(f9_2) * int64(g7_19)
+	f9g8_19 := int64(f9) * int64(g8_19)
+	f9g9_38 := int64(f9_2) * int64(g9_19)
+	h0 := f0g0 + f1g9_38 + f2g8_19 + f3g7_38 + f4g6_19 + f5g5_38 + f6g4_19 + f7g3_38 + f8g2_19 + f9g1_38
+	h1 := f0g1 + f1g0 + f2g9_19 + f3g8_19 + f4g7_19 + f5g6_19 + f6g5_19 + f7g4_19 + f8g3_19 + f9g2_19
+	h2 := f0g2 + f1g1_2 + f2g0 + f3g9_38 + f4g8_19 + f5g7_38 + f6g6_19 + f7g5_38 + f8g4_19 + f9g3_38
+	h3 := f0g3 + f1g2 + f2g1 + f3g0 + f4g9_19 + f5g8_19 + f6g7_19 + f7g6_19 + f8g5_19 + f9g4_19
+	h4 := f0g4 + f1g3_2 + f2g2 + f3g1_2 + f4g0 + f5g9_38 + f6g8_19 + f7g7_38 + f8g6_19 + f9g5_38
+	h5 := f0g5 + f1g4 + f2g3 + f3g2 + f4g1 + f5g0 + f6g9_19 + f7g8_19 + f8g7_19 + f9g6_19
+	h6 := f0g6 + f1g5_2 + f2g4 + f3g3_2 + f4g2 + f5g1_2 + f6g0 + f7g9_38 + f8g8_19 + f9g7_38
+	h7 := f0g7 + f1g6 + f2g5 + f3g4 + f4g3 + f5g2 + f6g1 + f7g0 + f8g9_19 + f9g8_19
+	h8 := f0g8 + f1g7_2 + f2g6 + f3g5_2 + f4g4 + f5g3_2 + f6g2 + f7g1_2 + f8g0 + f9g9_38
+	h9 := f0g9 + f1g8 + f2g7 + f3g6 + f4g5 + f5g4 + f6g3 + f7g2 + f8g1 + f9g0
+	var carry [10]int64
+
+	// |h0| <= (1.1*1.1*2^52*(1+19+19+19+19)+1.1*1.1*2^50*(38+38+38+38+38))
+	//   i.e. |h0| <= 1.2*2^59; narrower ranges for h2, h4, h6, h8
+	// |h1| <= (1.1*1.1*2^51*(1+1+19+19+19+19+19+19+19+19))
+	//   i.e. |h1| <= 1.5*2^58; narrower ranges for h3, h5, h7, h9
+
+	carry[0] = (h0 + (1 << 25)) >> 26
+	h1 += carry[0]
+	h0 -= carry[0] << 26
+	carry[4] = (h4 + (1 << 25)) >> 26
+	h5 += carry[4]
+	h4 -= carry[4] << 26
+	// |h0| <= 2^25
+	// |h4| <= 2^25
+	// |h1| <= 1.51*2^58
+	// |h5| <= 1.51*2^58
+
+	carry[1] = (h1 + (1 << 24)) >> 25
+	h2 += carry[1]
+	h1 -= carry[1] << 25
+	carry[5] = (h5 + (1 << 24)) >> 25
+	h6 += carry[5]
+	h5 -= carry[5] << 25
+	// |h1| <= 2^24; from now on fits into int32
+	// |h5| <= 2^24; from now on fits into int32
+	// |h2| <= 1.21*2^59
+	// |h6| <= 1.21*2^59
+
+	carry[2] = (h2 + (1 << 25)) >> 26
+	h3 += carry[2]
+	h2 -= carry[2] << 26
+	carry[6] = (h6 + (1 << 25)) >> 26
+	h7 += carry[6]
+	h6 -= carry[6] << 26
+	// |h2| <= 2^25; from now on fits into int32 unchanged
+	// |h6| <= 2^25; from now on fits into int32 unchanged
+	// |h3| <= 1.51*2^58
+	// |h7| <= 1.51*2^58
+
+	carry[3] = (h3 + (1 << 24)) >> 25
+	h4 += carry[3]
+	h3 -= carry[3] << 25
+	carry[7] = (h7 + (1 << 24)) >> 25
+	h8 += carry[7]
+	h7 -= carry[7] << 25
+	// |h3| <= 2^24; from now on fits into int32 unchanged
+	// |h7| <= 2^24; from now on fits into int32 unchanged
+	// |h4| <= 1.52*2^33
+	// |h8| <= 1.52*2^33
+
+	carry[4] = (h4 + (1 << 25)) >> 26
+	h5 += carry[4]
+	h4 -= carry[4] << 26
+	carry[8] = (h8 + (1 << 25)) >> 26
+	h9 += carry[8]
+	h8 -= carry[8] << 26
+	// |h4| <= 2^25; from now on fits into int32 unchanged
+	// |h8| <= 2^25; from now on fits into int32 unchanged
+	// |h5| <= 1.01*2^24
+	// |h9| <= 1.51*2^58
+
+	carry[9] = (h9 + (1 << 24)) >> 25
+	h0 += carry[9] * 19
+	h9 -= carry[9] << 25
+	// |h9| <= 2^24; from now on fits into int32 unchanged
+	// |h0| <= 1.8*2^37
+
+	carry[0] = (h0 + (1 << 25)) >> 26
+	h1 += carry[0]
+	h0 -= carry[0] << 26
+	// |h0| <= 2^25; from now on fits into int32 unchanged
+	// |h1| <= 1.01*2^24
+
+	h[0] = int32(h0)
+	h[1] = int32(h1)
+	h[2] = int32(h2)
+	h[3] = int32(h3)
+	h[4] = int32(h4)
+	h[5] = int32(h5)
+	h[6] = int32(h6)
+	h[7] = int32(h7)
+	h[8] = int32(h8)
+	h[9] = int32(h9)
+}
+
+// feSquare calculates h = f*f. Can overlap h with f.
+//
+// Preconditions:
+//    |f| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+//
+// Postconditions:
+//    |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
+func feSquare(h, f *fieldElement) {
+	f0 := f[0]
+	f1 := f[1]
+	f2 := f[2]
+	f3 := f[3]
+	f4 := f[4]
+	f5 := f[5]
+	f6 := f[6]
+	f7 := f[7]
+	f8 := f[8]
+	f9 := f[9]
+	f0_2 := 2 * f0
+	f1_2 := 2 * f1
+	f2_2 := 2 * f2
+	f3_2 := 2 * f3
+	f4_2 := 2 * f4
+	f5_2 := 2 * f5
+	f6_2 := 2 * f6
+	f7_2 := 2 * f7
+	f5_38 := 38 * f5 // 1.31*2^30
+	f6_19 := 19 * f6 // 1.31*2^30
+	f7_38 := 38 * f7 // 1.31*2^30
+	f8_19 := 19 * f8 // 1.31*2^30
+	f9_38 := 38 * f9 // 1.31*2^30
+	f0f0 := int64(f0) * int64(f0)
+	f0f1_2 := int64(f0_2) * int64(f1)
+	f0f2_2 := int64(f0_2) * int64(f2)
+	f0f3_2 := int64(f0_2) * int64(f3)
+	f0f4_2 := int64(f0_2) * int64(f4)
+	f0f5_2 := int64(f0_2) * int64(f5)
+	f0f6_2 := int64(f0_2) * int64(f6)
+	f0f7_2 := int64(f0_2) * int64(f7)
+	f0f8_2 := int64(f0_2) * int64(f8)
+	f0f9_2 := int64(f0_2) * int64(f9)
+	f1f1_2 := int64(f1_2) * int64(f1)
+	f1f2_2 := int64(f1_2) * int64(f2)
+	f1f3_4 := int64(f1_2) * int64(f3_2)
+	f1f4_2 := int64(f1_2) * int64(f4)
+	f1f5_4 := int64(f1_2) * int64(f5_2)
+	f1f6_2 := int64(f1_2) * int64(f6)
+	f1f7_4 := int64(f1_2) * int64(f7_2)
+	f1f8_2 := int64(f1_2) * int64(f8)
+	f1f9_76 := int64(f1_2) * int64(f9_38)
+	f2f2 := int64(f2) * int64(f2)
+	f2f3_2 := int64(f2_2) * int64(f3)
+	f2f4_2 := int64(f2_2) * int64(f4)
+	f2f5_2 := int64(f2_2) * int64(f5)
+	f2f6_2 := int64(f2_2) * int64(f6)
+	f2f7_2 := int64(f2_2) * int64(f7)
+	f2f8_38 := int64(f2_2) * int64(f8_19)
+	f2f9_38 := int64(f2) * int64(f9_38)
+	f3f3_2 := int64(f3_2) * int64(f3)
+	f3f4_2 := int64(f3_2) * int64(f4)
+	f3f5_4 := int64(f3_2) * int64(f5_2)
+	f3f6_2 := int64(f3_2) * int64(f6)
+	f3f7_76 := int64(f3_2) * int64(f7_38)
+	f3f8_38 := int64(f3_2) * int64(f8_19)
+	f3f9_76 := int64(f3_2) * int64(f9_38)
+	f4f4 := int64(f4) * int64(f4)
+	f4f5_2 := int64(f4_2) * int64(f5)
+	f4f6_38 := int64(f4_2) * int64(f6_19)
+	f4f7_38 := int64(f4) * int64(f7_38)
+	f4f8_38 := int64(f4_2) * int64(f8_19)
+	f4f9_38 := int64(f4) * int64(f9_38)
+	f5f5_38 := int64(f5) * int64(f5_38)
+	f5f6_38 := int64(f5_2) * int64(f6_19)
+	f5f7_76 := int64(f5_2) * int64(f7_38)
+	f5f8_38 := int64(f5_2) * int64(f8_19)
+	f5f9_76 := int64(f5_2) * int64(f9_38)
+	f6f6_19 := int64(f6) * int64(f6_19)
+	f6f7_38 := int64(f6) * int64(f7_38)
+	f6f8_38 := int64(f6_2) * int64(f8_19)
+	f6f9_38 := int64(f6) * int64(f9_38)
+	f7f7_38 := int64(f7) * int64(f7_38)
+	f7f8_38 := int64(f7_2) * int64(f8_19)
+	f7f9_76 := int64(f7_2) * int64(f9_38)
+	f8f8_19 := int64(f8) * int64(f8_19)
+	f8f9_38 := int64(f8) * int64(f9_38)
+	f9f9_38 := int64(f9) * int64(f9_38)
+	h0 := f0f0 + f1f9_76 + f2f8_38 + f3f7_76 + f4f6_38 + f5f5_38
+	h1 := f0f1_2 + f2f9_38 + f3f8_38 + f4f7_38 + f5f6_38
+	h2 := f0f2_2 + f1f1_2 + f3f9_76 + f4f8_38 + f5f7_76 + f6f6_19
+	h3 := f0f3_2 + f1f2_2 + f4f9_38 + f5f8_38 + f6f7_38
+	h4 := f0f4_2 + f1f3_4 + f2f2 + f5f9_76 + f6f8_38 + f7f7_38
+	h5 := f0f5_2 + f1f4_2 + f2f3_2 + f6f9_38 + f7f8_38
+	h6 := f0f6_2 + f1f5_4 + f2f4_2 + f3f3_2 + f7f9_76 + f8f8_19
+	h7 := f0f7_2 + f1f6_2 + f2f5_2 + f3f4_2 + f8f9_38
+	h8 := f0f8_2 + f1f7_4 + f2f6_2 + f3f5_4 + f4f4 + f9f9_38
+	h9 := f0f9_2 + f1f8_2 + f2f7_2 + f3f6_2 + f4f5_2
+	var carry [10]int64
+
+	carry[0] = (h0 + (1 << 25)) >> 26
+	h1 += carry[0]
+	h0 -= carry[0] << 26
+	carry[4] = (h4 + (1 << 25)) >> 26
+	h5 += carry[4]
+	h4 -= carry[4] << 26
+
+	carry[1] = (h1 + (1 << 24)) >> 25
+	h2 += carry[1]
+	h1 -= carry[1] << 25
+	carry[5] = (h5 + (1 << 24)) >> 25
+	h6 += carry[5]
+	h5 -= carry[5] << 25
+
+	carry[2] = (h2 + (1 << 25)) >> 26
+	h3 += carry[2]
+	h2 -= carry[2] << 26
+	carry[6] = (h6 + (1 << 25)) >> 26
+	h7 += carry[6]
+	h6 -= carry[6] << 26
+
+	carry[3] = (h3 + (1 << 24)) >> 25
+	h4 += carry[3]
+	h3 -= carry[3] << 25
+	carry[7] = (h7 + (1 << 24)) >> 25
+	h8 += carry[7]
+	h7 -= carry[7] << 25
+
+	carry[4] = (h4 + (1 << 25)) >> 26
+	h5 += carry[4]
+	h4 -= carry[4] << 26
+	carry[8] = (h8 + (1 << 25)) >> 26
+	h9 += carry[8]
+	h8 -= carry[8] << 26
+
+	carry[9] = (h9 + (1 << 24)) >> 25
+	h0 += carry[9] * 19
+	h9 -= carry[9] << 25
+
+	carry[0] = (h0 + (1 << 25)) >> 26
+	h1 += carry[0]
+	h0 -= carry[0] << 26
+
+	h[0] = int32(h0)
+	h[1] = int32(h1)
+	h[2] = int32(h2)
+	h[3] = int32(h3)
+	h[4] = int32(h4)
+	h[5] = int32(h5)
+	h[6] = int32(h6)
+	h[7] = int32(h7)
+	h[8] = int32(h8)
+	h[9] = int32(h9)
+}
+
+// feMul121666 calculates h = f * 121666. Can overlap h with f.
+//
+// Preconditions:
+//    |f| bounded by 1.1*2^26,1.1*2^25,1.1*2^26,1.1*2^25,etc.
+//
+// Postconditions:
+//    |h| bounded by 1.1*2^25,1.1*2^24,1.1*2^25,1.1*2^24,etc.
+func feMul121666(h, f *fieldElement) {
+	h0 := int64(f[0]) * 121666
+	h1 := int64(f[1]) * 121666
+	h2 := int64(f[2]) * 121666
+	h3 := int64(f[3]) * 121666
+	h4 := int64(f[4]) * 121666
+	h5 := int64(f[5]) * 121666
+	h6 := int64(f[6]) * 121666
+	h7 := int64(f[7]) * 121666
+	h8 := int64(f[8]) * 121666
+	h9 := int64(f[9]) * 121666
+	var carry [10]int64
+
+	carry[9] = (h9 + (1 << 24)) >> 25
+	h0 += carry[9] * 19
+	h9 -= carry[9] << 25
+	carry[1] = (h1 + (1 << 24)) >> 25
+	h2 += carry[1]
+	h1 -= carry[1] << 25
+	carry[3] = (h3 + (1 << 24)) >> 25
+	h4 += carry[3]
+	h3 -= carry[3] << 25
+	carry[5] = (h5 + (1 << 24)) >> 25
+	h6 += carry[5]
+	h5 -= carry[5] << 25
+	carry[7] = (h7 + (1 << 24)) >> 25
+	h8 += carry[7]
+	h7 -= carry[7] << 25
+
+	carry[0] = (h0 + (1 << 25)) >> 26
+	h1 += carry[0]
+	h0 -= carry[0] << 26
+	carry[2] = (h2 + (1 << 25)) >> 26
+	h3 += carry[2]
+	h2 -= carry[2] << 26
+	carry[4] = (h4 + (1 << 25)) >> 26
+	h5 += carry[4]
+	h4 -= carry[4] << 26
+	carry[6] = (h6 + (1 << 25)) >> 26
+	h7 += carry[6]
+	h6 -= carry[6] << 26
+	carry[8] = (h8 + (1 << 25)) >> 26
+	h9 += carry[8]
+	h8 -= carry[8] << 26
+
+	h[0] = int32(h0)
+	h[1] = int32(h1)
+	h[2] = int32(h2)
+	h[3] = int32(h3)
+	h[4] = int32(h4)
+	h[5] = int32(h5)
+	h[6] = int32(h6)
+	h[7] = int32(h7)
+	h[8] = int32(h8)
+	h[9] = int32(h9)
+}
+
+// feInvert sets out = z^-1.
+func feInvert(out, z *fieldElement) {
+	var t0, t1, t2, t3 fieldElement
+	var i int
+
+	feSquare(&t0, z)
+	for i = 1; i < 1; i++ {
+		feSquare(&t0, &t0)
+	}
+	feSquare(&t1, &t0)
+	for i = 1; i < 2; i++ {
+		feSquare(&t1, &t1)
+	}
+	feMul(&t1, z, &t1)
+	feMul(&t0, &t0, &t1)
+	feSquare(&t2, &t0)
+	for i = 1; i < 1; i++ {
+		feSquare(&t2, &t2)
+	}
+	feMul(&t1, &t1, &t2)
+	feSquare(&t2, &t1)
+	for i = 1; i < 5; i++ {
+		feSquare(&t2, &t2)
+	}
+	feMul(&t1, &t2, &t1)
+	feSquare(&t2, &t1)
+	for i = 1; i < 10; i++ {
+		feSquare(&t2, &t2)
+	}
+	feMul(&t2, &t2, &t1)
+	feSquare(&t3, &t2)
+	for i = 1; i < 20; i++ {
+		feSquare(&t3, &t3)
+	}
+	feMul(&t2, &t3, &t2)
+	feSquare(&t2, &t2)
+	for i = 1; i < 10; i++ {
+		feSquare(&t2, &t2)
+	}
+	feMul(&t1, &t2, &t1)
+	feSquare(&t2, &t1)
+	for i = 1; i < 50; i++ {
+		feSquare(&t2, &t2)
+	}
+	feMul(&t2, &t2, &t1)
+	feSquare(&t3, &t2)
+	for i = 1; i < 100; i++ {
+		feSquare(&t3, &t3)
+	}
+	feMul(&t2, &t3, &t2)
+	feSquare(&t2, &t2)
+	for i = 1; i < 50; i++ {
+		feSquare(&t2, &t2)
+	}
+	feMul(&t1, &t2, &t1)
+	feSquare(&t1, &t1)
+	for i = 1; i < 5; i++ {
+		feSquare(&t1, &t1)
+	}
+	feMul(out, &t1, &t0)
+}
+
+func scalarMult(out, in, base *[32]byte) {
+	var e [32]byte
+
+	copy(e[:], in[:])
+	e[0] &= 248
+	e[31] &= 127
+	e[31] |= 64
+
+	var x1, x2, z2, x3, z3, tmp0, tmp1 fieldElement
+	feFromBytes(&x1, base)
+	feOne(&x2)
+	feCopy(&x3, &x1)
+	feOne(&z3)
+
+	swap := int32(0)
+	for pos := 254; pos >= 0; pos-- {
+		b := e[pos/8] >> uint(pos&7)
+		b &= 1
+		swap ^= int32(b)
+		feCSwap(&x2, &x3, swap)
+		feCSwap(&z2, &z3, swap)
+		swap = int32(b)
+
+		feSub(&tmp0, &x3, &z3)
+		feSub(&tmp1, &x2, &z2)
+		feAdd(&x2, &x2, &z2)
+		feAdd(&z2, &x3, &z3)
+		feMul(&z3, &tmp0, &x2)
+		feMul(&z2, &z2, &tmp1)
+		feSquare(&tmp0, &tmp1)
+		feSquare(&tmp1, &x2)
+		feAdd(&x3, &z3, &z2)
+		feSub(&z2, &z3, &z2)
+		feMul(&x2, &tmp1, &tmp0)
+		feSub(&tmp1, &tmp1, &tmp0)
+		feSquare(&z2, &z2)
+		feMul121666(&z3, &tmp1)
+		feSquare(&x3, &x3)
+		feAdd(&tmp0, &tmp0, &z3)
+		feMul(&z3, &x1, &z2)
+		feMul(&z2, &tmp1, &tmp0)
+	}
+
+	feCSwap(&x2, &x3, swap)
+	feCSwap(&z2, &z3, swap)
+
+	feInvert(&z2, &z2)
+	feMul(&x2, &x2, &z2)
+	feToBytes(out, &x2)
+}

Godeps/_workspace/src/golang.org/x/crypto/curve25519/curve25519_test.go

@@ -0,0 +1,29 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package curve25519
+
+import (
+	"fmt"
+	"testing"
+)
+
+const expectedHex = "89161fde887b2b53de549af483940106ecc114d6982daa98256de23bdf77661a"
+
+func TestBaseScalarMult(t *testing.T) {
+	var a, b [32]byte
+	in := &a
+	out := &b
+	a[0] = 1
+
+	for i := 0; i < 200; i++ {
+		ScalarBaseMult(out, in)
+		in, out = out, in
+	}
+
+	result := fmt.Sprintf("%x", in[:])
+	if result != expectedHex {
+		t.Errorf("incorrect result: got %s, want %s", result, expectedHex)
+	}
+}

Godeps/_workspace/src/golang.org/x/crypto/curve25519/doc.go

@@ -0,0 +1,23 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package curve25519 provides an implementation of scalar multiplication on
+// the elliptic curve known as curve25519. See http://cr.yp.to/ecdh.html
+package curve25519
+
+// basePoint is the x coordinate of the generator of the curve.
+var basePoint = [32]byte{9, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0}
+
+// ScalarMult sets dst to the product in*base where dst and base are the x
+// coordinates of group points and all values are in little-endian form.
+func ScalarMult(dst, in, base *[32]byte) {
+	scalarMult(dst, in, base)
+}
+
+// ScalarBaseMult sets dst to the product in*base where dst and base are the x
+// coordinates of group points, base is the standard generator and all values
+// are in little-endian form.
+func ScalarBaseMult(dst, in *[32]byte) {
+	ScalarMult(dst, in, &basePoint)
+}

Godeps/_workspace/src/golang.org/x/crypto/curve25519/freeze_amd64.s

@@ -0,0 +1,94 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// +build amd64,!gccgo,!appengine
+
+// func freeze(inout *[5]uint64)
+TEXT ·freeze(SB),7,$96-8
+	MOVQ inout+0(FP), DI
+
+	MOVQ SP,R11
+	MOVQ $31,CX
+	NOTQ CX
+	ANDQ CX,SP
+	ADDQ $32,SP
+
+	MOVQ R11,0(SP)
+	MOVQ R12,8(SP)
+	MOVQ R13,16(SP)
+	MOVQ R14,24(SP)
+	MOVQ R15,32(SP)
+	MOVQ BX,40(SP)
+	MOVQ BP,48(SP)
+	MOVQ 0(DI),SI
+	MOVQ 8(DI),DX
+	MOVQ 16(DI),CX
+	MOVQ 24(DI),R8
+	MOVQ 32(DI),R9
+	MOVQ ·REDMASK51(SB),AX
+	MOVQ AX,R10
+	SUBQ $18,R10
+	MOVQ $3,R11
+REDUCELOOP:
+	MOVQ SI,R12
+	SHRQ $51,R12
+	ANDQ AX,SI
+	ADDQ R12,DX
+	MOVQ DX,R12
+	SHRQ $51,R12
+	ANDQ AX,DX
+	ADDQ R12,CX
+	MOVQ CX,R12
+	SHRQ $51,R12
+	ANDQ AX,CX
+	ADDQ R12,R8
+	MOVQ R8,R12
+	SHRQ $51,R12
+	ANDQ AX,R8
+	ADDQ R12,R9
+	MOVQ R9,R12
+	SHRQ $51,R12
+	ANDQ AX,R9
+	IMUL3Q $19,R12,R12
+	ADDQ R12,SI
+	SUBQ $1,R11
+	JA REDUCELOOP
+	MOVQ $1,R12
+	CMPQ R10,SI
+	CMOVQLT R11,R12
+	CMPQ AX,DX
+	CMOVQNE R11,R12
+	CMPQ AX,CX
+	CMOVQNE R11,R12
+	CMPQ AX,R8
+	CMOVQNE R11,R12
+	CMPQ AX,R9
+	CMOVQNE R11,R12
+	NEGQ R12
+	ANDQ R12,AX
+	ANDQ R12,R10
+	SUBQ R10,SI
+	SUBQ AX,DX
+	SUBQ AX,CX
+	SUBQ AX,R8
+	SUBQ AX,R9
+	MOVQ SI,0(DI)
+	MOVQ DX,8(DI)
+	MOVQ CX,16(DI)
+	MOVQ R8,24(DI)
+	MOVQ R9,32(DI)
+	MOVQ 0(SP),R11
+	MOVQ 8(SP),R12
+	MOVQ 16(SP),R13
+	MOVQ 24(SP),R14
+	MOVQ 32(SP),R15
+	MOVQ 40(SP),BX
+	MOVQ 48(SP),BP
+	MOVQ R11,SP
+	MOVQ DI,AX
+	MOVQ SI,DX
+	RET

Godeps/_workspace/src/golang.org/x/crypto/curve25519/mont25519_amd64.go

@@ -0,0 +1,240 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build amd64,!gccgo,!appengine
+
+package curve25519
+
+// These functions are implemented in the .s files. The names of the functions
+// in the rest of the file are also taken from the SUPERCOP sources to help
+// people following along.
+
+//go:noescape
+
+func cswap(inout *[5]uint64, v uint64)
+
+//go:noescape
+
+func ladderstep(inout *[5][5]uint64)
+
+//go:noescape
+
+func freeze(inout *[5]uint64)
+
+//go:noescape
+
+func mul(dest, a, b *[5]uint64)
+
+//go:noescape
+
+func square(out, in *[5]uint64)
+
+// mladder uses a Montgomery ladder to calculate (xr/zr) *= s.
+func mladder(xr, zr *[5]uint64, s *[32]byte) {
+	var work [5][5]uint64
+
+	work[0] = *xr
+	setint(&work[1], 1)
+	setint(&work[2], 0)
+	work[3] = *xr
+	setint(&work[4], 1)
+
+	j := uint(6)
+	var prevbit byte
+
+	for i := 31; i >= 0; i-- {
+		for j < 8 {
+			bit := ((*s)[i] >> j) & 1
+			swap := bit ^ prevbit
+			prevbit = bit
+			cswap(&work[1], uint64(swap))
+			ladderstep(&work)
+			j--
+		}
+		j = 7
+	}
+
+	*xr = work[1]
+	*zr = work[2]
+}
+
+func scalarMult(out, in, base *[32]byte) {
+	var e [32]byte
+	copy(e[:], (*in)[:])
+	e[0] &= 248
+	e[31] &= 127
+	e[31] |= 64
+
+	var t, z [5]uint64
+	unpack(&t, base)
+	mladder(&t, &z, &e)
+	invert(&z, &z)
+	mul(&t, &t, &z)
+	pack(out, &t)
+}
+
+func setint(r *[5]uint64, v uint64) {
+	r[0] = v
+	r[1] = 0
+	r[2] = 0
+	r[3] = 0
+	r[4] = 0
+}
+
+// unpack sets r = x where r consists of 5, 51-bit limbs in little-endian
+// order.
+func unpack(r *[5]uint64, x *[32]byte) {
+	r[0] = uint64(x[0]) |
+		uint64(x[1])<<8 |
+		uint64(x[2])<<16 |
+		uint64(x[3])<<24 |
+		uint64(x[4])<<32 |
+		uint64(x[5])<<40 |
+		uint64(x[6]&7)<<48
+
+	r[1] = uint64(x[6])>>3 |
+		uint64(x[7])<<5 |
+		uint64(x[8])<<13 |
+		uint64(x[9])<<21 |
+		uint64(x[10])<<29 |
+		uint64(x[11])<<37 |
+		uint64(x[12]&63)<<45
+
+	r[2] = uint64(x[12])>>6 |
+		uint64(x[13])<<2 |
+		uint64(x[14])<<10 |
+		uint64(x[15])<<18 |
+		uint64(x[16])<<26 |
+		uint64(x[17])<<34 |
+		uint64(x[18])<<42 |
+		uint64(x[19]&1)<<50
+
+	r[3] = uint64(x[19])>>1 |
+		uint64(x[20])<<7 |
+		uint64(x[21])<<15 |
+		uint64(x[22])<<23 |
+		uint64(x[23])<<31 |
+		uint64(x[24])<<39 |
+		uint64(x[25]&15)<<47
+
+	r[4] = uint64(x[25])>>4 |
+		uint64(x[26])<<4 |
+		uint64(x[27])<<12 |
+		uint64(x[28])<<20 |
+		uint64(x[29])<<28 |
+		uint64(x[30])<<36 |
+		uint64(x[31]&127)<<44
+}
+
+// pack sets out = x where out is the usual, little-endian form of the 5,
+// 51-bit limbs in x.
+func pack(out *[32]byte, x *[5]uint64) {
+	t := *x
+	freeze(&t)
+
+	out[0] = byte(t[0])
+	out[1] = byte(t[0] >> 8)
+	out[2] = byte(t[0] >> 16)
+	out[3] = byte(t[0] >> 24)
+	out[4] = byte(t[0] >> 32)
+	out[5] = byte(t[0] >> 40)
+	out[6] = byte(t[0] >> 48)
+
+	out[6] ^= byte(t[1]<<3) & 0xf8
+	out[7] = byte(t[1] >> 5)
+	out[8] = byte(t[1] >> 13)
+	out[9] = byte(t[1] >> 21)
+	out[10] = byte(t[1] >> 29)
+	out[11] = byte(t[1] >> 37)
+	out[12] = byte(t[1] >> 45)
+
+	out[12] ^= byte(t[2]<<6) & 0xc0
+	out[13] = byte(t[2] >> 2)
+	out[14] = byte(t[2] >> 10)
+	out[15] = byte(t[2] >> 18)
+	out[16] = byte(t[2] >> 26)
+	out[17] = byte(t[2] >> 34)
+	out[18] = byte(t[2] >> 42)
+	out[19] = byte(t[2] >> 50)
+
+	out[19] ^= byte(t[3]<<1) & 0xfe
+	out[20] = byte(t[3] >> 7)
+	out[21] = byte(t[3] >> 15)
+	out[22] = byte(t[3] >> 23)
+	out[23] = byte(t[3] >> 31)
+	out[24] = byte(t[3] >> 39)
+	out[25] = byte(t[3] >> 47)
+
+	out[25] ^= byte(t[4]<<4) & 0xf0
+	out[26] = byte(t[4] >> 4)
+	out[27] = byte(t[4] >> 12)
+	out[28] = byte(t[4] >> 20)
+	out[29] = byte(t[4] >> 28)
+	out[30] = byte(t[4] >> 36)
+	out[31] = byte(t[4] >> 44)
+}
+
+// invert calculates r = x^-1 mod p using Fermat's little theorem.
+func invert(r *[5]uint64, x *[5]uint64) {
+	var z2, z9, z11, z2_5_0, z2_10_0, z2_20_0, z2_50_0, z2_100_0, t [5]uint64
+
+	square(&z2, x)        /* 2 */
+	square(&t, &z2)       /* 4 */
+	square(&t, &t)        /* 8 */
+	mul(&z9, &t, x)       /* 9 */
+	mul(&z11, &z9, &z2)   /* 11 */
+	square(&t, &z11)      /* 22 */
+	mul(&z2_5_0, &t, &z9) /* 2^5 - 2^0 = 31 */
+
+	square(&t, &z2_5_0)      /* 2^6 - 2^1 */
+	for i := 1; i < 5; i++ { /* 2^20 - 2^10 */
+		square(&t, &t)
+	}
+	mul(&z2_10_0, &t, &z2_5_0) /* 2^10 - 2^0 */
+
+	square(&t, &z2_10_0)      /* 2^11 - 2^1 */
+	for i := 1; i < 10; i++ { /* 2^20 - 2^10 */
+		square(&t, &t)
+	}
+	mul(&z2_20_0, &t, &z2_10_0) /* 2^20 - 2^0 */
+
+	square(&t, &z2_20_0)      /* 2^21 - 2^1 */
+	for i := 1; i < 20; i++ { /* 2^40 - 2^20 */
+		square(&t, &t)
+	}
+	mul(&t, &t, &z2_20_0) /* 2^40 - 2^0 */
+
+	square(&t, &t)            /* 2^41 - 2^1 */
+	for i := 1; i < 10; i++ { /* 2^50 - 2^10 */
+		square(&t, &t)
+	}
+	mul(&z2_50_0, &t, &z2_10_0) /* 2^50 - 2^0 */
+
+	square(&t, &z2_50_0)      /* 2^51 - 2^1 */
+	for i := 1; i < 50; i++ { /* 2^100 - 2^50 */
+		square(&t, &t)
+	}
+	mul(&z2_100_0, &t, &z2_50_0) /* 2^100 - 2^0 */
+
+	square(&t, &z2_100_0)      /* 2^101 - 2^1 */
+	for i := 1; i < 100; i++ { /* 2^200 - 2^100 */
+		square(&t, &t)
+	}
+	mul(&t, &t, &z2_100_0) /* 2^200 - 2^0 */
+
+	square(&t, &t)            /* 2^201 - 2^1 */
+	for i := 1; i < 50; i++ { /* 2^250 - 2^50 */
+		square(&t, &t)
+	}
+	mul(&t, &t, &z2_50_0) /* 2^250 - 2^0 */
+
+	square(&t, &t) /* 2^251 - 2^1 */
+	square(&t, &t) /* 2^252 - 2^2 */
+	square(&t, &t) /* 2^253 - 2^3 */
+
+	square(&t, &t) /* 2^254 - 2^4 */
+
+	square(&t, &t)   /* 2^255 - 2^5 */
+	mul(r, &t, &z11) /* 2^255 - 21 */
+}

Godeps/_workspace/src/golang.org/x/crypto/curve25519/mul_amd64.s

@@ -0,0 +1,191 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// +build amd64,!gccgo,!appengine
+
+// func mul(dest, a, b *[5]uint64)
+TEXT ·mul(SB),0,$128-24
+	MOVQ dest+0(FP), DI
+	MOVQ a+8(FP), SI
+	MOVQ b+16(FP), DX
+
+	MOVQ SP,R11
+	MOVQ $31,CX
+	NOTQ CX
+	ANDQ CX,SP
+	ADDQ $32,SP
+
+	MOVQ R11,0(SP)
+	MOVQ R12,8(SP)
+	MOVQ R13,16(SP)
+	MOVQ R14,24(SP)
+	MOVQ R15,32(SP)
+	MOVQ BX,40(SP)
+	MOVQ BP,48(SP)
+	MOVQ DI,56(SP)
+	MOVQ DX,CX
+	MOVQ 24(SI),DX
+	IMUL3Q $19,DX,AX
+	MOVQ AX,64(SP)
+	MULQ 16(CX)
+	MOVQ AX,R8
+	MOVQ DX,R9
+	MOVQ 32(SI),DX
+	IMUL3Q $19,DX,AX
+	MOVQ AX,72(SP)
+	MULQ 8(CX)
+	ADDQ AX,R8
+	ADCQ DX,R9
+	MOVQ 0(SI),AX
+	MULQ 0(CX)
+	ADDQ AX,R8
+	ADCQ DX,R9
+	MOVQ 0(SI),AX
+	MULQ 8(CX)
+	MOVQ AX,R10
+	MOVQ DX,R11
+	MOVQ 0(SI),AX
+	MULQ 16(CX)
+	MOVQ AX,R12
+	MOVQ DX,R13
+	MOVQ 0(SI),AX
+	MULQ 24(CX)
+	MOVQ AX,R14
+	MOVQ DX,R15
+	MOVQ 0(SI),AX
+	MULQ 32(CX)
+	MOVQ AX,BX
+	MOVQ DX,BP
+	MOVQ 8(SI),AX
+	MULQ 0(CX)
+	ADDQ AX,R10
+	ADCQ DX,R11
+	MOVQ 8(SI),AX
+	MULQ 8(CX)
+	ADDQ AX,R12
+	ADCQ DX,R13
+	MOVQ 8(SI),AX
+	MULQ 16(CX)
+	ADDQ AX,R14
+	ADCQ DX,R15
+	MOVQ 8(SI),AX
+	MULQ 24(CX)
+	ADDQ AX,BX
+	ADCQ DX,BP
+	MOVQ 8(SI),DX
+	IMUL3Q $19,DX,AX
+	MULQ 32(CX)
+	ADDQ AX,R8
+	ADCQ DX,R9
+	MOVQ 16(SI),AX
+	MULQ 0(CX)
+	ADDQ AX,R12
+	ADCQ DX,R13
+	MOVQ 16(SI),AX
+	MULQ 8(CX)
+	ADDQ AX,R14
+	ADCQ DX,R15
+	MOVQ 16(SI),AX
+	MULQ 16(CX)
+	ADDQ AX,BX
+	ADCQ DX,BP
+	MOVQ 16(SI),DX
+	IMUL3Q $19,DX,AX
+	MULQ 24(CX)
+	ADDQ AX,R8
+	ADCQ DX,R9
+	MOVQ 16(SI),DX
+	IMUL3Q $19,DX,AX
+	MULQ 32(CX)
+	ADDQ AX,R10
+	ADCQ DX,R11
+	MOVQ 24(SI),AX
+	MULQ 0(CX)
+	ADDQ AX,R14
+	ADCQ DX,R15
+	MOVQ 24(SI),AX
+	MULQ 8(CX)
+	ADDQ AX,BX
+	ADCQ DX,BP
+	MOVQ 64(SP),AX
+	MULQ 24(CX)
+	ADDQ AX,R10
+	ADCQ DX,R11
+	MOVQ 64(SP),AX
+	MULQ 32(CX)
+	ADDQ AX,R12
+	ADCQ DX,R13
+	MOVQ 32(SI),AX
+	MULQ 0(CX)
+	ADDQ AX,BX
+	ADCQ DX,BP
+	MOVQ 72(SP),AX
+	MULQ 16(CX)
+	ADDQ AX,R10
+	ADCQ DX,R11
+	MOVQ 72(SP),AX
+	MULQ 24(CX)
+	ADDQ AX,R12
+	ADCQ DX,R13
+	MOVQ 72(SP),AX
+	MULQ 32(CX)
+	ADDQ AX,R14
+	ADCQ DX,R15
+	MOVQ ·REDMASK51(SB),SI
+	SHLQ $13,R9:R8
+	ANDQ SI,R8
+	SHLQ $13,R11:R10
+	ANDQ SI,R10
+	ADDQ R9,R10
+	SHLQ $13,R13:R12
+	ANDQ SI,R12
+	ADDQ R11,R12
+	SHLQ $13,R15:R14
+	ANDQ SI,R14
+	ADDQ R13,R14
+	SHLQ $13,BP:BX
+	ANDQ SI,BX
+	ADDQ R15,BX
+	IMUL3Q $19,BP,DX
+	ADDQ DX,R8
+	MOVQ R8,DX
+	SHRQ $51,DX
+	ADDQ R10,DX
+	MOVQ DX,CX
+	SHRQ $51,DX
+	ANDQ SI,R8
+	ADDQ R12,DX
+	MOVQ DX,R9
+	SHRQ $51,DX
+	ANDQ SI,CX
+	ADDQ R14,DX
+	MOVQ DX,AX
+	SHRQ $51,DX
+	ANDQ SI,R9
+	ADDQ BX,DX
+	MOVQ DX,R10
+	SHRQ $51,DX
+	ANDQ SI,AX
+	IMUL3Q $19,DX,DX
+	ADDQ DX,R8
+	ANDQ SI,R10
+	MOVQ R8,0(DI)
+	MOVQ CX,8(DI)
+	MOVQ R9,16(DI)
+	MOVQ AX,24(DI)
+	MOVQ R10,32(DI)
+	MOVQ 0(SP),R11
+	MOVQ 8(SP),R12
+	MOVQ 16(SP),R13
+	MOVQ 24(SP),R14
+	MOVQ 32(SP),R15
+	MOVQ 40(SP),BX
+	MOVQ 48(SP),BP
+	MOVQ R11,SP
+	MOVQ DI,AX
+	MOVQ SI,DX
+	RET

Godeps/_workspace/src/golang.org/x/crypto/curve25519/square_amd64.s

@@ -0,0 +1,153 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// +build amd64,!gccgo,!appengine
+
+// func square(out, in *[5]uint64)
+TEXT ·square(SB),7,$96-16
+	MOVQ out+0(FP), DI
+	MOVQ in+8(FP), SI
+
+	MOVQ SP,R11
+	MOVQ $31,CX
+	NOTQ CX
+	ANDQ CX,SP
+	ADDQ $32, SP
+
+	MOVQ R11,0(SP)
+	MOVQ R12,8(SP)
+	MOVQ R13,16(SP)
+	MOVQ R14,24(SP)
+	MOVQ R15,32(SP)
+	MOVQ BX,40(SP)
+	MOVQ BP,48(SP)
+	MOVQ 0(SI),AX
+	MULQ 0(SI)
+	MOVQ AX,CX
+	MOVQ DX,R8
+	MOVQ 0(SI),AX
+	SHLQ $1,AX
+	MULQ 8(SI)
+	MOVQ AX,R9
+	MOVQ DX,R10
+	MOVQ 0(SI),AX
+	SHLQ $1,AX
+	MULQ 16(SI)
+	MOVQ AX,R11
+	MOVQ DX,R12
+	MOVQ 0(SI),AX
+	SHLQ $1,AX
+	MULQ 24(SI)
+	MOVQ AX,R13
+	MOVQ DX,R14
+	MOVQ 0(SI),AX
+	SHLQ $1,AX
+	MULQ 32(SI)
+	MOVQ AX,R15
+	MOVQ DX,BX
+	MOVQ 8(SI),AX
+	MULQ 8(SI)
+	ADDQ AX,R11
+	ADCQ DX,R12
+	MOVQ 8(SI),AX
+	SHLQ $1,AX
+	MULQ 16(SI)
+	ADDQ AX,R13
+	ADCQ DX,R14
+	MOVQ 8(SI),AX
+	SHLQ $1,AX
+	MULQ 24(SI)
+	ADDQ AX,R15
+	ADCQ DX,BX
+	MOVQ 8(SI),DX
+	IMUL3Q $38,DX,AX
+	MULQ 32(SI)
+	ADDQ AX,CX
+	ADCQ DX,R8
+	MOVQ 16(SI),AX
+	MULQ 16(SI)
+	ADDQ AX,R15
+	ADCQ DX,BX
+	MOVQ 16(SI),DX
+	IMUL3Q $38,DX,AX
+	MULQ 24(SI)
+	ADDQ AX,CX
+	ADCQ DX,R8
+	MOVQ 16(SI),DX
+	IMUL3Q $38,DX,AX
+	MULQ 32(SI)
+	ADDQ AX,R9
+	ADCQ DX,R10
+	MOVQ 24(SI),DX
+	IMUL3Q $19,DX,AX
+	MULQ 24(SI)
+	ADDQ AX,R9
+	ADCQ DX,R10
+	MOVQ 24(SI),DX
+	IMUL3Q $38,DX,AX
+	MULQ 32(SI)
+	ADDQ AX,R11
+	ADCQ DX,R12
+	MOVQ 32(SI),DX
+	IMUL3Q $19,DX,AX
+	MULQ 32(SI)
+	ADDQ AX,R13
+	ADCQ DX,R14
+	MOVQ ·REDMASK51(SB),SI
+	SHLQ $13,R8:CX
+	ANDQ SI,CX
+	SHLQ $13,R10:R9
+	ANDQ SI,R9
+	ADDQ R8,R9
+	SHLQ $13,R12:R11
+	ANDQ SI,R11
+	ADDQ R10,R11
+	SHLQ $13,R14:R13
+	ANDQ SI,R13
+	ADDQ R12,R13
+	SHLQ $13,BX:R15
+	ANDQ SI,R15
+	ADDQ R14,R15
+	IMUL3Q $19,BX,DX
+	ADDQ DX,CX
+	MOVQ CX,DX
+	SHRQ $51,DX
+	ADDQ R9,DX
+	ANDQ SI,CX
+	MOVQ DX,R8
+	SHRQ $51,DX
+	ADDQ R11,DX
+	ANDQ SI,R8
+	MOVQ DX,R9
+	SHRQ $51,DX
+	ADDQ R13,DX
+	ANDQ SI,R9
+	MOVQ DX,AX
+	SHRQ $51,DX
+	ADDQ R15,DX
+	ANDQ SI,AX
+	MOVQ DX,R10
+	SHRQ $51,DX
+	IMUL3Q $19,DX,DX
+	ADDQ DX,CX
+	ANDQ SI,R10
+	MOVQ CX,0(DI)
+	MOVQ R8,8(DI)
+	MOVQ R9,16(DI)
+	MOVQ AX,24(DI)
+	MOVQ R10,32(DI)
+	MOVQ 0(SP),R11
+	MOVQ 8(SP),R12
+	MOVQ 16(SP),R13
+	MOVQ 24(SP),R14
+	MOVQ 32(SP),R15
+	MOVQ 40(SP),BX
+	MOVQ 48(SP),BP
+	MOVQ R11,SP
+	MOVQ DI,AX
+	MOVQ SI,DX
+	RET

Godeps/_workspace/src/golang.org/x/crypto/nacl/box/box.go

@@ -0,0 +1,85 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+/*
+Package box authenticates and encrypts messages using public-key cryptography.
+
+Box uses Curve25519, XSalsa20 and Poly1305 to encrypt and authenticate
+messages. The length of messages is not hidden.
+
+It is the caller's responsibility to ensure the uniqueness of nonces—for
+example, by using nonce 1 for the first message, nonce 2 for the second
+message, etc. Nonces are long enough that randomly generated nonces have
+negligible risk of collision.
+
+This package is interoperable with NaCl: http://nacl.cr.yp.to/box.html.
+*/
+package box
+
+import (
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/curve25519"
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/nacl/secretbox"
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa"
+	"io"
+)
+
+// Overhead is the number of bytes of overhead when boxing a message.
+const Overhead = secretbox.Overhead
+
+// GenerateKey generates a new public/private key pair suitable for use with
+// Seal and Open.
+func GenerateKey(rand io.Reader) (publicKey, privateKey *[32]byte, err error) {
+	publicKey = new([32]byte)
+	privateKey = new([32]byte)
+	_, err = io.ReadFull(rand, privateKey[:])
+	if err != nil {
+		publicKey = nil
+		privateKey = nil
+		return
+	}
+
+	curve25519.ScalarBaseMult(publicKey, privateKey)
+	return
+}
+
+var zeros [16]byte
+
+// Precompute calculates the shared key between peersPublicKey and privateKey
+// and writes it to sharedKey. The shared key can be used with
+// OpenAfterPrecomputation and SealAfterPrecomputation to speed up processing
+// when using the same pair of keys repeatedly.
+func Precompute(sharedKey, peersPublicKey, privateKey *[32]byte) {
+	curve25519.ScalarMult(sharedKey, privateKey, peersPublicKey)
+	salsa.HSalsa20(sharedKey, &zeros, sharedKey, &salsa.Sigma)
+}
+
+// Seal appends an encrypted and authenticated copy of message to out, which
+// will be Overhead bytes longer than the original and must not overlap. The
+// nonce must be unique for each distinct message for a given pair of keys.
+func Seal(out, message []byte, nonce *[24]byte, peersPublicKey, privateKey *[32]byte) []byte {
+	var sharedKey [32]byte
+	Precompute(&sharedKey, peersPublicKey, privateKey)
+	return secretbox.Seal(out, message, nonce, &sharedKey)
+}
+
+// SealAfterPrecomputation performs the same actions as Seal, but takes a
+// shared key as generated by Precompute.
+func SealAfterPrecomputation(out, message []byte, nonce *[24]byte, sharedKey *[32]byte) []byte {
+	return secretbox.Seal(out, message, nonce, sharedKey)
+}
+
+// Open authenticates and decrypts a box produced by Seal and appends the
+// message to out, which must not overlap box. The output will be Overhead
+// bytes smaller than box.
+func Open(out, box []byte, nonce *[24]byte, peersPublicKey, privateKey *[32]byte) ([]byte, bool) {
+	var sharedKey [32]byte
+	Precompute(&sharedKey, peersPublicKey, privateKey)
+	return secretbox.Open(out, box, nonce, &sharedKey)
+}
+
+// OpenAfterPrecomputation performs the same actions as Open, but takes a
+// shared key as generated by Precompute.
+func OpenAfterPrecomputation(out, box []byte, nonce *[24]byte, sharedKey *[32]byte) ([]byte, bool) {
+	return secretbox.Open(out, box, nonce, sharedKey)
+}

Godeps/_workspace/src/golang.org/x/crypto/nacl/box/box_test.go

@@ -0,0 +1,78 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package box
+
+import (
+	"bytes"
+	"crypto/rand"
+	"encoding/hex"
+	"testing"
+
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/curve25519"
+)
+
+func TestSealOpen(t *testing.T) {
+	publicKey1, privateKey1, _ := GenerateKey(rand.Reader)
+	publicKey2, privateKey2, _ := GenerateKey(rand.Reader)
+
+	if *privateKey1 == *privateKey2 {
+		t.Fatalf("private keys are equal!")
+	}
+	if *publicKey1 == *publicKey2 {
+		t.Fatalf("public keys are equal!")
+	}
+	message := []byte("test message")
+	var nonce [24]byte
+
+	box := Seal(nil, message, &nonce, publicKey1, privateKey2)
+	opened, ok := Open(nil, box, &nonce, publicKey2, privateKey1)
+	if !ok {
+		t.Fatalf("failed to open box")
+	}
+
+	if !bytes.Equal(opened, message) {
+		t.Fatalf("got %x, want %x", opened, message)
+	}
+
+	for i := range box {
+		box[i] ^= 0x40
+		_, ok := Open(nil, box, &nonce, publicKey2, privateKey1)
+		if ok {
+			t.Fatalf("opened box with byte %d corrupted", i)
+		}
+		box[i] ^= 0x40
+	}
+}
+
+func TestBox(t *testing.T) {
+	var privateKey1, privateKey2 [32]byte
+	for i := range privateKey1[:] {
+		privateKey1[i] = 1
+	}
+	for i := range privateKey2[:] {
+		privateKey2[i] = 2
+	}
+
+	var publicKey1 [32]byte
+	curve25519.ScalarBaseMult(&publicKey1, &privateKey1)
+	var message [64]byte
+	for i := range message[:] {
+		message[i] = 3
+	}
+
+	var nonce [24]byte
+	for i := range nonce[:] {
+		nonce[i] = 4
+	}
+
+	box := Seal(nil, message[:], &nonce, &publicKey1, &privateKey2)
+
+	// expected was generated using the C implementation of NaCl.
+	expected, _ := hex.DecodeString("78ea30b19d2341ebbdba54180f821eec265cf86312549bea8a37652a8bb94f07b78a73ed1708085e6ddd0e943bbdeb8755079a37eb31d86163ce241164a47629c0539f330b4914cd135b3855bc2a2dfc")
+
+	if !bytes.Equal(box, expected) {
+		t.Fatalf("box didn't match, got\n%x\n, expected\n%x", box, expected)
+	}
+}

Godeps/_workspace/src/golang.org/x/crypto/nacl/secretbox/secretbox.go

@@ -0,0 +1,149 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+/*
+Package secretbox encrypts and authenticates small messages.
+
+Secretbox uses XSalsa20 and Poly1305 to encrypt and authenticate messages with
+secret-key cryptography. The length of messages is not hidden.
+
+It is the caller's responsibility to ensure the uniqueness of nonces—for
+example, by using nonce 1 for the first message, nonce 2 for the second
+message, etc. Nonces are long enough that randomly generated nonces have
+negligible risk of collision.
+
+This package is interoperable with NaCl: http://nacl.cr.yp.to/secretbox.html.
+*/
+package secretbox
+
+import (
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/poly1305"
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa"
+)
+
+// Overhead is the number of bytes of overhead when boxing a message.
+const Overhead = poly1305.TagSize
+
+// setup produces a sub-key and Salsa20 counter given a nonce and key.
+func setup(subKey *[32]byte, counter *[16]byte, nonce *[24]byte, key *[32]byte) {
+	// We use XSalsa20 for encryption so first we need to generate a
+	// key and nonce with HSalsa20.
+	var hNonce [16]byte
+	copy(hNonce[:], nonce[:])
+	salsa.HSalsa20(subKey, &hNonce, key, &salsa.Sigma)
+
+	// The final 8 bytes of the original nonce form the new nonce.
+	copy(counter[:], nonce[16:])
+}
+
+// sliceForAppend takes a slice and a requested number of bytes. It returns a
+// slice with the contents of the given slice followed by that many bytes and a
+// second slice that aliases into it and contains only the extra bytes. If the
+// original slice has sufficient capacity then no allocation is performed.
+func sliceForAppend(in []byte, n int) (head, tail []byte) {
+	if total := len(in) + n; cap(in) >= total {
+		head = in[:total]
+	} else {
+		head = make([]byte, total)
+		copy(head, in)
+	}
+	tail = head[len(in):]
+	return
+}
+
+// Seal appends an encrypted and authenticated copy of message to out, which
+// must not overlap message. The key and nonce pair must be unique for each
+// distinct message and the output will be Overhead bytes longer than message.
+func Seal(out, message []byte, nonce *[24]byte, key *[32]byte) []byte {
+	var subKey [32]byte
+	var counter [16]byte
+	setup(&subKey, &counter, nonce, key)
+
+	// The Poly1305 key is generated by encrypting 32 bytes of zeros. Since
+	// Salsa20 works with 64-byte blocks, we also generate 32 bytes of
+	// keystream as a side effect.
+	var firstBlock [64]byte
+	salsa.XORKeyStream(firstBlock[:], firstBlock[:], &counter, &subKey)
+
+	var poly1305Key [32]byte
+	copy(poly1305Key[:], firstBlock[:])
+
+	ret, out := sliceForAppend(out, len(message)+poly1305.TagSize)
+
+	// We XOR up to 32 bytes of message with the keystream generated from
+	// the first block.
+	firstMessageBlock := message
+	if len(firstMessageBlock) > 32 {
+		firstMessageBlock = firstMessageBlock[:32]
+	}
+
+	tagOut := out
+	out = out[poly1305.TagSize:]
+	for i, x := range firstMessageBlock {
+		out[i] = firstBlock[32+i] ^ x
+	}
+	message = message[len(firstMessageBlock):]
+	ciphertext := out
+	out = out[len(firstMessageBlock):]
+
+	// Now encrypt the rest.
+	counter[8] = 1
+	salsa.XORKeyStream(out, message, &counter, &subKey)
+
+	var tag [poly1305.TagSize]byte
+	poly1305.Sum(&tag, ciphertext, &poly1305Key)
+	copy(tagOut, tag[:])
+
+	return ret
+}
+
+// Open authenticates and decrypts a box produced by Seal and appends the
+// message to out, which must not overlap box. The output will be Overhead
+// bytes smaller than box.
+func Open(out []byte, box []byte, nonce *[24]byte, key *[32]byte) ([]byte, bool) {
+	if len(box) < Overhead {
+		return nil, false
+	}
+
+	var subKey [32]byte
+	var counter [16]byte
+	setup(&subKey, &counter, nonce, key)
+
+	// The Poly1305 key is generated by encrypting 32 bytes of zeros. Since
+	// Salsa20 works with 64-byte blocks, we also generate 32 bytes of
+	// keystream as a side effect.
+	var firstBlock [64]byte
+	salsa.XORKeyStream(firstBlock[:], firstBlock[:], &counter, &subKey)
+
+	var poly1305Key [32]byte
+	copy(poly1305Key[:], firstBlock[:])
+	var tag [poly1305.TagSize]byte
+	copy(tag[:], box)
+
+	if !poly1305.Verify(&tag, box[poly1305.TagSize:], &poly1305Key) {
+		return nil, false
+	}
+
+	ret, out := sliceForAppend(out, len(box)-Overhead)
+
+	// We XOR up to 32 bytes of box with the keystream generated from
+	// the first block.
+	box = box[Overhead:]
+	firstMessageBlock := box
+	if len(firstMessageBlock) > 32 {
+		firstMessageBlock = firstMessageBlock[:32]
+	}
+	for i, x := range firstMessageBlock {
+		out[i] = firstBlock[32+i] ^ x
+	}
+
+	box = box[len(firstMessageBlock):]
+	out = out[len(firstMessageBlock):]
+
+	// Now decrypt the rest.
+	counter[8] = 1
+	salsa.XORKeyStream(out, box, &counter, &subKey)
+
+	return ret, true
+}

Godeps/_workspace/src/golang.org/x/crypto/nacl/secretbox/secretbox_test.go

@@ -0,0 +1,91 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package secretbox
+
+import (
+	"bytes"
+	"crypto/rand"
+	"encoding/hex"
+	"testing"
+)
+
+func TestSealOpen(t *testing.T) {
+	var key [32]byte
+	var nonce [24]byte
+
+	rand.Reader.Read(key[:])
+	rand.Reader.Read(nonce[:])
+
+	var box, opened []byte
+
+	for msgLen := 0; msgLen < 128; msgLen += 17 {
+		message := make([]byte, msgLen)
+		rand.Reader.Read(message)
+
+		box = Seal(box[:0], message, &nonce, &key)
+		var ok bool
+		opened, ok = Open(opened[:0], box, &nonce, &key)
+		if !ok {
+			t.Errorf("%d: failed to open box", msgLen)
+			continue
+		}
+
+		if !bytes.Equal(opened, message) {
+			t.Errorf("%d: got %x, expected %x", msgLen, opened, message)
+			continue
+		}
+	}
+
+	for i := range box {
+		box[i] ^= 0x20
+		_, ok := Open(opened[:0], box, &nonce, &key)
+		if ok {
+			t.Errorf("box was opened after corrupting byte %d", i)
+		}
+		box[i] ^= 0x20
+	}
+}
+
+func TestSecretBox(t *testing.T) {
+	var key [32]byte
+	var nonce [24]byte
+	var message [64]byte
+
+	for i := range key[:] {
+		key[i] = 1
+	}
+	for i := range nonce[:] {
+		nonce[i] = 2
+	}
+	for i := range message[:] {
+		message[i] = 3
+	}
+
+	box := Seal(nil, message[:], &nonce, &key)
+	// expected was generated using the C implementation of NaCl.
+	expected, _ := hex.DecodeString("8442bc313f4626f1359e3b50122b6ce6fe66ddfe7d39d14e637eb4fd5b45beadab55198df6ab5368439792a23c87db70acb6156dc5ef957ac04f6276cf6093b84be77ff0849cc33e34b7254d5a8f65ad")
+
+	if !bytes.Equal(box, expected) {
+		t.Fatalf("box didn't match, got\n%x\n, expected\n%x", box, expected)
+	}
+}
+
+func TestAppend(t *testing.T) {
+	var key [32]byte
+	var nonce [24]byte
+	var message [8]byte
+
+	out := make([]byte, 4)
+	box := Seal(out, message[:], &nonce, &key)
+	if !bytes.Equal(box[:4], out[:4]) {
+		t.Fatalf("Seal didn't correctly append")
+	}
+
+	out = make([]byte, 4, 100)
+	box = Seal(out, message[:], &nonce, &key)
+	if !bytes.Equal(box[:4], out[:4]) {
+		t.Fatalf("Seal didn't correctly append with sufficient capacity.")
+	}
+}

Godeps/_workspace/src/golang.org/x/crypto/poly1305/const_amd64.s

@@ -0,0 +1,45 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// +build amd64,!gccgo,!appengine
+
+DATA ·SCALE(SB)/8, $0x37F4000000000000
+GLOBL ·SCALE(SB), 8, $8
+DATA ·TWO32(SB)/8, $0x41F0000000000000
+GLOBL ·TWO32(SB), 8, $8
+DATA ·TWO64(SB)/8, $0x43F0000000000000
+GLOBL ·TWO64(SB), 8, $8
+DATA ·TWO96(SB)/8, $0x45F0000000000000
+GLOBL ·TWO96(SB), 8, $8
+DATA ·ALPHA32(SB)/8, $0x45E8000000000000
+GLOBL ·ALPHA32(SB), 8, $8
+DATA ·ALPHA64(SB)/8, $0x47E8000000000000
+GLOBL ·ALPHA64(SB), 8, $8
+DATA ·ALPHA96(SB)/8, $0x49E8000000000000
+GLOBL ·ALPHA96(SB), 8, $8
+DATA ·ALPHA130(SB)/8, $0x4C08000000000000
+GLOBL ·ALPHA130(SB), 8, $8
+DATA ·DOFFSET0(SB)/8, $0x4330000000000000
+GLOBL ·DOFFSET0(SB), 8, $8
+DATA ·DOFFSET1(SB)/8, $0x4530000000000000
+GLOBL ·DOFFSET1(SB), 8, $8
+DATA ·DOFFSET2(SB)/8, $0x4730000000000000
+GLOBL ·DOFFSET2(SB), 8, $8
+DATA ·DOFFSET3(SB)/8, $0x4930000000000000
+GLOBL ·DOFFSET3(SB), 8, $8
+DATA ·DOFFSET3MINUSTWO128(SB)/8, $0x492FFFFE00000000
+GLOBL ·DOFFSET3MINUSTWO128(SB), 8, $8
+DATA ·HOFFSET0(SB)/8, $0x43300001FFFFFFFB
+GLOBL ·HOFFSET0(SB), 8, $8
+DATA ·HOFFSET1(SB)/8, $0x45300001FFFFFFFE
+GLOBL ·HOFFSET1(SB), 8, $8
+DATA ·HOFFSET2(SB)/8, $0x47300001FFFFFFFE
+GLOBL ·HOFFSET2(SB), 8, $8
+DATA ·HOFFSET3(SB)/8, $0x49300003FFFFFFFE
+GLOBL ·HOFFSET3(SB), 8, $8
+DATA ·ROUNDING(SB)/2, $0x137f
+GLOBL ·ROUNDING(SB), 8, $2

Godeps/_workspace/src/golang.org/x/crypto/poly1305/poly1305.go

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

Godeps/_workspace/src/golang.org/x/crypto/poly1305/poly1305_amd64.s

@@ -0,0 +1,497 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// +build amd64,!gccgo,!appengine
+
+// func poly1305(out *[16]byte, m *byte, mlen uint64, key *[32]key)
+TEXT ·poly1305(SB),0,$224-32
+	MOVQ out+0(FP),DI
+	MOVQ m+8(FP),SI
+	MOVQ mlen+16(FP),DX
+	MOVQ key+24(FP),CX
+
+	MOVQ SP,R11
+	MOVQ $31,R9
+	NOTQ R9
+	ANDQ R9,SP
+	ADDQ $32,SP
+
+	MOVQ R11,32(SP)
+	MOVQ R12,40(SP)
+	MOVQ R13,48(SP)
+	MOVQ R14,56(SP)
+	MOVQ R15,64(SP)
+	MOVQ BX,72(SP)
+	MOVQ BP,80(SP)
+	FLDCW ·ROUNDING(SB)
+	MOVL 0(CX),R8
+	MOVL 4(CX),R9
+	MOVL 8(CX),AX
+	MOVL 12(CX),R10
+	MOVQ DI,88(SP)
+	MOVQ CX,96(SP)
+	MOVL $0X43300000,108(SP)
+	MOVL $0X45300000,116(SP)
+	MOVL $0X47300000,124(SP)
+	MOVL $0X49300000,132(SP)
+	ANDL $0X0FFFFFFF,R8
+	ANDL $0X0FFFFFFC,R9
+	ANDL $0X0FFFFFFC,AX
+	ANDL $0X0FFFFFFC,R10
+	MOVL R8,104(SP)
+	MOVL R9,112(SP)
+	MOVL AX,120(SP)
+	MOVL R10,128(SP)
+	FMOVD 104(SP), F0
+	FSUBD ·DOFFSET0(SB), F0
+	FMOVD 112(SP), F0
+	FSUBD ·DOFFSET1(SB), F0
+	FMOVD 120(SP), F0
+	FSUBD ·DOFFSET2(SB), F0
+	FMOVD 128(SP), F0
+	FSUBD ·DOFFSET3(SB), F0
+	FXCHD F0, F3
+	FMOVDP F0, 136(SP)
+	FXCHD F0, F1
+	FMOVD F0, 144(SP)
+	FMULD ·SCALE(SB), F0
+	FMOVDP F0, 152(SP)
+	FMOVD F0, 160(SP)
+	FMULD ·SCALE(SB), F0
+	FMOVDP F0, 168(SP)
+	FMOVD F0, 176(SP)
+	FMULD ·SCALE(SB), F0
+	FMOVDP F0, 184(SP)
+	FLDZ
+	FLDZ
+	FLDZ
+	FLDZ
+	CMPQ DX,$16
+	JB ADDATMOST15BYTES
+	INITIALATLEAST16BYTES:
+	MOVL 12(SI),DI
+	MOVL 8(SI),CX
+	MOVL 4(SI),R8
+	MOVL 0(SI),R9
+	MOVL DI,128(SP)
+	MOVL CX,120(SP)
+	MOVL R8,112(SP)
+	MOVL R9,104(SP)
+	ADDQ $16,SI
+	SUBQ $16,DX
+	FXCHD F0, F3
+	FADDD 128(SP), F0
+	FSUBD ·DOFFSET3MINUSTWO128(SB), F0
+	FXCHD F0, F1
+	FADDD 112(SP), F0
+	FSUBD ·DOFFSET1(SB), F0
+	FXCHD F0, F2
+	FADDD 120(SP), F0
+	FSUBD ·DOFFSET2(SB), F0
+	FXCHD F0, F3
+	FADDD 104(SP), F0
+	FSUBD ·DOFFSET0(SB), F0
+	CMPQ DX,$16
+	JB MULTIPLYADDATMOST15BYTES
+	MULTIPLYADDATLEAST16BYTES:
+	MOVL 12(SI),DI
+	MOVL 8(SI),CX
+	MOVL 4(SI),R8
+	MOVL 0(SI),R9
+	MOVL DI,128(SP)
+	MOVL CX,120(SP)
+	MOVL R8,112(SP)
+	MOVL R9,104(SP)
+	ADDQ $16,SI
+	SUBQ $16,DX
+	FMOVD ·ALPHA130(SB), F0
+	FADDD F2,F0
+	FSUBD ·ALPHA130(SB), F0
+	FSUBD F0,F2
+	FMULD ·SCALE(SB), F0
+	FMOVD ·ALPHA32(SB), F0
+	FADDD F2,F0
+	FSUBD ·ALPHA32(SB), F0
+	FSUBD F0,F2
+	FXCHD F0, F2
+	FADDDP F0,F1
+	FMOVD ·ALPHA64(SB), F0
+	FADDD F4,F0
+	FSUBD ·ALPHA64(SB), F0
+	FSUBD F0,F4
+	FMOVD ·ALPHA96(SB), F0
+	FADDD F6,F0
+	FSUBD ·ALPHA96(SB), F0
+	FSUBD F0,F6
+	FXCHD F0, F6
+	FADDDP F0,F1
+	FXCHD F0, F3
+	FADDDP F0,F5
+	FXCHD F0, F3
+	FADDDP F0,F1
+	FMOVD 176(SP), F0
+	FMULD F3,F0
+	FMOVD 160(SP), F0
+	FMULD F4,F0
+	FMOVD 144(SP), F0
+	FMULD F5,F0
+	FMOVD 136(SP), F0
+	FMULDP F0,F6
+	FMOVD 160(SP), F0
+	FMULD F4,F0
+	FADDDP F0,F3
+	FMOVD 144(SP), F0
+	FMULD F4,F0
+	FADDDP F0,F2
+	FMOVD 136(SP), F0
+	FMULD F4,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULDP F0,F4
+	FXCHD F0, F3
+	FADDDP F0,F5
+	FMOVD 144(SP), F0
+	FMULD F4,F0
+	FADDDP F0,F2
+	FMOVD 136(SP), F0
+	FMULD F4,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULD F4,F0
+	FADDDP F0,F3
+	FMOVD 168(SP), F0
+	FMULDP F0,F4
+	FXCHD F0, F3
+	FADDDP F0,F4
+	FMOVD 136(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F1
+	FXCHD F0, F3
+	FMOVD 184(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F3
+	FXCHD F0, F1
+	FMOVD 168(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F1
+	FMOVD 152(SP), F0
+	FMULDP F0,F5
+	FXCHD F0, F4
+	FADDDP F0,F1
+	CMPQ DX,$16
+	FXCHD F0, F2
+	FMOVD 128(SP), F0
+	FSUBD ·DOFFSET3MINUSTWO128(SB), F0
+	FADDDP F0,F1
+	FXCHD F0, F1
+	FMOVD 120(SP), F0
+	FSUBD ·DOFFSET2(SB), F0
+	FADDDP F0,F1
+	FXCHD F0, F3
+	FMOVD 112(SP), F0
+	FSUBD ·DOFFSET1(SB), F0
+	FADDDP F0,F1
+	FXCHD F0, F2
+	FMOVD 104(SP), F0
+	FSUBD ·DOFFSET0(SB), F0
+	FADDDP F0,F1
+	JAE MULTIPLYADDATLEAST16BYTES
+	MULTIPLYADDATMOST15BYTES:
+	FMOVD ·ALPHA130(SB), F0
+	FADDD F2,F0
+	FSUBD ·ALPHA130(SB), F0
+	FSUBD F0,F2
+	FMULD ·SCALE(SB), F0
+	FMOVD ·ALPHA32(SB), F0
+	FADDD F2,F0
+	FSUBD ·ALPHA32(SB), F0
+	FSUBD F0,F2
+	FMOVD ·ALPHA64(SB), F0
+	FADDD F5,F0
+	FSUBD ·ALPHA64(SB), F0
+	FSUBD F0,F5
+	FMOVD ·ALPHA96(SB), F0
+	FADDD F7,F0
+	FSUBD ·ALPHA96(SB), F0
+	FSUBD F0,F7
+	FXCHD F0, F7
+	FADDDP F0,F1
+	FXCHD F0, F5
+	FADDDP F0,F1
+	FXCHD F0, F3
+	FADDDP F0,F5
+	FADDDP F0,F1
+	FMOVD 176(SP), F0
+	FMULD F1,F0
+	FMOVD 160(SP), F0
+	FMULD F2,F0
+	FMOVD 144(SP), F0
+	FMULD F3,F0
+	FMOVD 136(SP), F0
+	FMULDP F0,F4
+	FMOVD 160(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F3
+	FMOVD 144(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F2
+	FMOVD 136(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULDP F0,F5
+	FXCHD F0, F4
+	FADDDP F0,F3
+	FMOVD 144(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F2
+	FMOVD 136(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F4
+	FMOVD 168(SP), F0
+	FMULDP F0,F5
+	FXCHD F0, F4
+	FADDDP F0,F2
+	FMOVD 136(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F4
+	FMOVD 168(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F3
+	FMOVD 152(SP), F0
+	FMULDP F0,F5
+	FXCHD F0, F4
+	FADDDP F0,F1
+	ADDATMOST15BYTES:
+	CMPQ DX,$0
+	JE NOMOREBYTES
+	MOVL $0,0(SP)
+	MOVL $0, 4 (SP)
+	MOVL $0, 8 (SP)
+	MOVL $0, 12 (SP)
+	LEAQ 0(SP),DI
+	MOVQ DX,CX
+	REP; MOVSB
+	MOVB $1,0(DI)
+	MOVL  12 (SP),DI
+	MOVL  8 (SP),SI
+	MOVL  4 (SP),DX
+	MOVL 0(SP),CX
+	MOVL DI,128(SP)
+	MOVL SI,120(SP)
+	MOVL DX,112(SP)
+	MOVL CX,104(SP)
+	FXCHD F0, F3
+	FADDD 128(SP), F0
+	FSUBD ·DOFFSET3(SB), F0
+	FXCHD F0, F2
+	FADDD 120(SP), F0
+	FSUBD ·DOFFSET2(SB), F0
+	FXCHD F0, F1
+	FADDD 112(SP), F0
+	FSUBD ·DOFFSET1(SB), F0
+	FXCHD F0, F3
+	FADDD 104(SP), F0
+	FSUBD ·DOFFSET0(SB), F0
+	FMOVD ·ALPHA130(SB), F0
+	FADDD F3,F0
+	FSUBD ·ALPHA130(SB), F0
+	FSUBD F0,F3
+	FMULD ·SCALE(SB), F0
+	FMOVD ·ALPHA32(SB), F0
+	FADDD F2,F0
+	FSUBD ·ALPHA32(SB), F0
+	FSUBD F0,F2
+	FMOVD ·ALPHA64(SB), F0
+	FADDD F6,F0
+	FSUBD ·ALPHA64(SB), F0
+	FSUBD F0,F6
+	FMOVD ·ALPHA96(SB), F0
+	FADDD F5,F0
+	FSUBD ·ALPHA96(SB), F0
+	FSUBD F0,F5
+	FXCHD F0, F4
+	FADDDP F0,F3
+	FXCHD F0, F6
+	FADDDP F0,F1
+	FXCHD F0, F3
+	FADDDP F0,F5
+	FXCHD F0, F3
+	FADDDP F0,F1
+	FMOVD 176(SP), F0
+	FMULD F3,F0
+	FMOVD 160(SP), F0
+	FMULD F4,F0
+	FMOVD 144(SP), F0
+	FMULD F5,F0
+	FMOVD 136(SP), F0
+	FMULDP F0,F6
+	FMOVD 160(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F3
+	FMOVD 144(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F2
+	FMOVD 136(SP), F0
+	FMULD F5,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULDP F0,F5
+	FXCHD F0, F4
+	FADDDP F0,F5
+	FMOVD 144(SP), F0
+	FMULD F6,F0
+	FADDDP F0,F2
+	FMOVD 136(SP), F0
+	FMULD F6,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULD F6,F0
+	FADDDP F0,F4
+	FMOVD 168(SP), F0
+	FMULDP F0,F6
+	FXCHD F0, F5
+	FADDDP F0,F4
+	FMOVD 136(SP), F0
+	FMULD F2,F0
+	FADDDP F0,F1
+	FMOVD 184(SP), F0
+	FMULD F2,F0
+	FADDDP F0,F5
+	FMOVD 168(SP), F0
+	FMULD F2,F0
+	FADDDP F0,F3
+	FMOVD 152(SP), F0
+	FMULDP F0,F2
+	FXCHD F0, F1
+	FADDDP F0,F3
+	FXCHD F0, F3
+	FXCHD F0, F2
+	NOMOREBYTES:
+	MOVL $0,R10
+	FMOVD ·ALPHA130(SB), F0
+	FADDD F4,F0
+	FSUBD ·ALPHA130(SB), F0
+	FSUBD F0,F4
+	FMULD ·SCALE(SB), F0
+	FMOVD ·ALPHA32(SB), F0
+	FADDD F2,F0
+	FSUBD ·ALPHA32(SB), F0
+	FSUBD F0,F2
+	FMOVD ·ALPHA64(SB), F0
+	FADDD F4,F0
+	FSUBD ·ALPHA64(SB), F0
+	FSUBD F0,F4
+	FMOVD ·ALPHA96(SB), F0
+	FADDD F6,F0
+	FSUBD ·ALPHA96(SB), F0
+	FXCHD F0, F6
+	FSUBD F6,F0
+	FXCHD F0, F4
+	FADDDP F0,F3
+	FXCHD F0, F4
+	FADDDP F0,F1
+	FXCHD F0, F2
+	FADDDP F0,F3
+	FXCHD F0, F4
+	FADDDP F0,F3
+	FXCHD F0, F3
+	FADDD ·HOFFSET0(SB), F0
+	FXCHD F0, F3
+	FADDD ·HOFFSET1(SB), F0
+	FXCHD F0, F1
+	FADDD ·HOFFSET2(SB), F0
+	FXCHD F0, F2
+	FADDD ·HOFFSET3(SB), F0
+	FXCHD F0, F3
+	FMOVDP F0, 104(SP)
+	FMOVDP F0, 112(SP)
+	FMOVDP F0, 120(SP)
+	FMOVDP F0, 128(SP)
+	MOVL 108(SP),DI
+	ANDL $63,DI
+	MOVL 116(SP),SI
+	ANDL $63,SI
+	MOVL 124(SP),DX
+	ANDL $63,DX
+	MOVL 132(SP),CX
+	ANDL $63,CX
+	MOVL 112(SP),R8
+	ADDL DI,R8
+	MOVQ R8,112(SP)
+	MOVL 120(SP),DI
+	ADCL SI,DI
+	MOVQ DI,120(SP)
+	MOVL 128(SP),DI
+	ADCL DX,DI
+	MOVQ DI,128(SP)
+	MOVL R10,DI
+	ADCL CX,DI
+	MOVQ DI,136(SP)
+	MOVQ $5,DI
+	MOVL 104(SP),SI
+	ADDL SI,DI
+	MOVQ DI,104(SP)
+	MOVL R10,DI
+	MOVQ 112(SP),DX
+	ADCL DX,DI
+	MOVQ DI,112(SP)
+	MOVL R10,DI
+	MOVQ 120(SP),CX
+	ADCL CX,DI
+	MOVQ DI,120(SP)
+	MOVL R10,DI
+	MOVQ 128(SP),R8
+	ADCL R8,DI
+	MOVQ DI,128(SP)
+	MOVQ $0XFFFFFFFC,DI
+	MOVQ 136(SP),R9
+	ADCL R9,DI
+	SARL $16,DI
+	MOVQ DI,R9
+	XORL $0XFFFFFFFF,R9
+	ANDQ DI,SI
+	MOVQ 104(SP),AX
+	ANDQ R9,AX
+	ORQ AX,SI
+	ANDQ DI,DX
+	MOVQ 112(SP),AX
+	ANDQ R9,AX
+	ORQ AX,DX
+	ANDQ DI,CX
+	MOVQ 120(SP),AX
+	ANDQ R9,AX
+	ORQ AX,CX
+	ANDQ DI,R8
+	MOVQ 128(SP),DI
+	ANDQ R9,DI
+	ORQ DI,R8
+	MOVQ 88(SP),DI
+	MOVQ 96(SP),R9
+	ADDL 16(R9),SI
+	ADCL 20(R9),DX
+	ADCL 24(R9),CX
+	ADCL 28(R9),R8
+	MOVL SI,0(DI)
+	MOVL DX,4(DI)
+	MOVL CX,8(DI)
+	MOVL R8,12(DI)
+	MOVQ 32(SP),R11
+	MOVQ 40(SP),R12
+	MOVQ 48(SP),R13
+	MOVQ 56(SP),R14
+	MOVQ 64(SP),R15
+	MOVQ 72(SP),BX
+	MOVQ 80(SP),BP
+	MOVQ R11,SP
+	RET

Godeps/_workspace/src/golang.org/x/crypto/poly1305/poly1305_arm.s

@@ -0,0 +1,331 @@
+// 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.
+
+// This code was translated into a form compatible with 5a from the public
+// domain source by Andrew Moon: github.com/floodyberry/poly1305-opt/blob/master/app/extensions/poly1305.
+
+// +build arm,!gccgo,!appengine
+
+DATA poly1305_init_constants_armv6<>+0x00(SB)/4, $0x3ffffff
+DATA poly1305_init_constants_armv6<>+0x04(SB)/4, $0x3ffff03
+DATA poly1305_init_constants_armv6<>+0x08(SB)/4, $0x3ffc0ff
+DATA poly1305_init_constants_armv6<>+0x0c(SB)/4, $0x3f03fff
+DATA poly1305_init_constants_armv6<>+0x10(SB)/4, $0x00fffff
+GLOBL poly1305_init_constants_armv6<>(SB), 8, $20
+
+// Warning: the linker may use R11 to synthesize certain instructions. Please
+// take care and verify that no synthetic instructions use it.
+
+TEXT poly1305_init_ext_armv6<>(SB),4,$-4
+  MOVM.DB.W [R4-R11], (R13)
+  MOVM.IA.W (R1), [R2-R5]
+  MOVW $poly1305_init_constants_armv6<>(SB), R7
+  MOVW R2, R8
+  MOVW R2>>26, R9
+  MOVW R3>>20, g
+  MOVW R4>>14, R11
+  MOVW R5>>8, R12
+  ORR R3<<6, R9, R9
+  ORR R4<<12, g, g
+  ORR R5<<18, R11, R11
+  MOVM.IA (R7), [R2-R6]
+  AND R8, R2, R2
+  AND R9, R3, R3
+  AND g, R4, R4
+  AND R11, R5, R5
+  AND R12, R6, R6
+  MOVM.IA.W [R2-R6], (R0)
+  EOR R2, R2, R2
+  EOR R3, R3, R3
+  EOR R4, R4, R4
+  EOR R5, R5, R5
+  EOR R6, R6, R6
+  MOVM.IA.W [R2-R6], (R0)
+  MOVM.IA.W (R1), [R2-R5]
+  MOVM.IA [R2-R6], (R0)
+  MOVM.IA.W (R13), [R4-R11]
+  RET
+
+TEXT poly1305_blocks_armv6<>(SB),4,$-4
+  MOVM.DB.W [R4, R5, R6, R7, R8, R9, g, R11, R14], (R13)
+  SUB $128, R13
+  MOVW R0, 36(R13)
+  MOVW R1, 40(R13)
+  MOVW R2, 44(R13)
+  MOVW R1, R14
+  MOVW R2, R12
+  MOVW 56(R0), R8
+  WORD $0xe1180008 // TST R8, R8 not working see issue 5921
+  EOR R6, R6, R6
+  MOVW.EQ $(1<<24), R6
+  MOVW R6, 32(R13)
+  ADD $64, R13, g
+  MOVM.IA (R0), [R0-R9]
+  MOVM.IA [R0-R4], (g)
+  CMP $16, R12
+  BLO poly1305_blocks_armv6_done
+poly1305_blocks_armv6_mainloop:
+  MOVM.IA.W (R14), [R0-R3]
+  MOVW R0>>26, g
+  MOVW R1>>20, R11
+  MOVW R2>>14, R12
+  MOVW R14, 40(R13)
+  MOVW R3>>8, R4
+  ORR R1<<6, g, g
+  ORR R2<<12, R11, R11
+  ORR R3<<18, R12, R12
+  BIC $0xfc000000, R0, R0
+  BIC $0xfc000000, g, g
+  MOVW 32(R13), R3
+  BIC $0xfc000000, R11, R11
+  BIC $0xfc000000, R12, R12
+  ADD R0, R5, R5
+  ADD g, R6, R6
+  ORR R3, R4, R4
+  ADD R11, R7, R7
+  ADD $64, R13, R14
+  ADD R12, R8, R8
+  ADD R4, R9, R9
+  MOVM.IA (R14), [R0-R4]
+  MULLU R4, R5, (R11, g)
+  MULLU R3, R5, (R14, R12)
+  MULALU R3, R6, (R11, g)
+  MULALU R2, R6, (R14, R12)
+  MULALU R2, R7, (R11, g)
+  MULALU R1, R7, (R14, R12)
+  ADD R4<<2, R4, R4
+  ADD R3<<2, R3, R3
+  MULALU R1, R8, (R11, g)
+  MULALU R0, R8, (R14, R12)
+  MULALU R0, R9, (R11, g)
+  MULALU R4, R9, (R14, R12)
+  MOVW g, 24(R13)
+  MOVW R11, 28(R13)
+  MOVW R12, 16(R13)
+  MOVW R14, 20(R13)
+  MULLU R2, R5, (R11, g)
+  MULLU R1, R5, (R14, R12)
+  MULALU R1, R6, (R11, g)
+  MULALU R0, R6, (R14, R12)
+  MULALU R0, R7, (R11, g)
+  MULALU R4, R7, (R14, R12)
+  ADD R2<<2, R2, R2
+  ADD R1<<2, R1, R1
+  MULALU R4, R8, (R11, g)
+  MULALU R3, R8, (R14, R12)
+  MULALU R3, R9, (R11, g)
+  MULALU R2, R9, (R14, R12)
+  MOVW g, 8(R13)
+  MOVW R11, 12(R13)
+  MOVW R12, 0(R13)
+  MOVW R14, w+4(SP)
+  MULLU R0, R5, (R11, g)
+  MULALU R4, R6, (R11, g)
+  MULALU R3, R7, (R11, g)
+  MULALU R2, R8, (R11, g)
+  MULALU R1, R9, (R11, g)
+  MOVM.IA (R13), [R0-R7]
+  MOVW g>>26, R12
+  MOVW R4>>26, R14
+  ORR R11<<6, R12, R12
+  ORR R5<<6, R14, R14
+  BIC $0xfc000000, g, g
+  BIC $0xfc000000, R4, R4
+  ADD.S R12, R0, R0
+  ADC $0, R1, R1
+  ADD.S R14, R6, R6
+  ADC $0, R7, R7
+  MOVW R0>>26, R12
+  MOVW R6>>26, R14
+  ORR R1<<6, R12, R12
+  ORR R7<<6, R14, R14
+  BIC $0xfc000000, R0, R0
+  BIC $0xfc000000, R6, R6
+  ADD R14<<2, R14, R14
+  ADD.S R12, R2, R2
+  ADC $0, R3, R3
+  ADD R14, g, g
+  MOVW R2>>26, R12
+  MOVW g>>26, R14
+  ORR R3<<6, R12, R12
+  BIC $0xfc000000, g, R5
+  BIC $0xfc000000, R2, R7
+  ADD R12, R4, R4
+  ADD R14, R0, R0
+  MOVW R4>>26, R12
+  BIC $0xfc000000, R4, R8
+  ADD R12, R6, R9
+  MOVW w+44(SP), R12
+  MOVW w+40(SP), R14
+  MOVW R0, R6
+  CMP $32, R12
+  SUB $16, R12, R12
+  MOVW R12, 44(R13)
+  BHS poly1305_blocks_armv6_mainloop
+poly1305_blocks_armv6_done:
+  MOVW 36(R13), R12
+  MOVW R5, 20(R12)
+  MOVW R6, 24(R12)
+  MOVW R7, 28(R12)
+  MOVW R8, 32(R12)
+  MOVW R9, 36(R12)
+  ADD $128, R13, R13
+  MOVM.IA.W (R13), [R4, R5, R6, R7, R8, R9, g, R11, R14]
+  RET
+
+TEXT poly1305_finish_ext_armv6<>(SB),4,$-4
+  MOVM.DB.W [R4, R5, R6, R7, R8, R9, g, R11, R14], (R13)
+  SUB $16, R13, R13
+  MOVW R0, R5
+  MOVW R1, R6
+  MOVW R2, R7
+  MOVW R3, R8
+  AND.S R2, R2, R2
+  BEQ poly1305_finish_ext_armv6_noremaining
+  EOR R0, R0
+  MOVW R13, R9
+  MOVW R0, 0(R13)
+  MOVW R0, 4(R13)
+  MOVW R0, 8(R13)
+  MOVW R0, 12(R13)
+  WORD $0xe3120008 // TST R2, #8 not working see issue 5921
+  BEQ poly1305_finish_ext_armv6_skip8
+  MOVM.IA.W (R1), [g-R11]
+  MOVM.IA.W [g-R11], (R9)
+poly1305_finish_ext_armv6_skip8:
+  WORD $0xe3120004 // TST $4, R2 not working see issue 5921
+  BEQ poly1305_finish_ext_armv6_skip4
+  MOVW.P 4(R1), g
+  MOVW.P g, 4(R9)
+poly1305_finish_ext_armv6_skip4:
+  WORD $0xe3120002 // TST $2, R2 not working see issue 5921
+  BEQ poly1305_finish_ext_armv6_skip2
+  MOVHU.P 2(R1), g
+  MOVH.P g, 2(R9)
+poly1305_finish_ext_armv6_skip2:
+  WORD $0xe3120001 // TST $1, R2 not working see issue 5921
+  BEQ poly1305_finish_ext_armv6_skip1
+  MOVBU.P 1(R1), g
+  MOVBU.P g, 1(R9)
+poly1305_finish_ext_armv6_skip1:
+  MOVW $1, R11
+  MOVBU R11, 0(R9)
+  MOVW R11, 56(R5)
+  MOVW R5, R0
+  MOVW R13, R1
+  MOVW $16, R2
+  BL poly1305_blocks_armv6<>(SB)
+poly1305_finish_ext_armv6_noremaining:
+  MOVW 20(R5), R0
+  MOVW 24(R5), R1
+  MOVW 28(R5), R2
+  MOVW 32(R5), R3
+  MOVW 36(R5), R4
+  MOVW R4>>26, R12
+  BIC $0xfc000000, R4, R4
+  ADD R12<<2, R12, R12
+  ADD R12, R0, R0
+  MOVW R0>>26, R12
+  BIC $0xfc000000, R0, R0
+  ADD R12, R1, R1
+  MOVW R1>>26, R12
+  BIC $0xfc000000, R1, R1
+  ADD R12, R2, R2
+  MOVW R2>>26, R12
+  BIC $0xfc000000, R2, R2
+  ADD R12, R3, R3
+  MOVW R3>>26, R12
+  BIC $0xfc000000, R3, R3
+  ADD R12, R4, R4
+  ADD $5, R0, R6
+  MOVW R6>>26, R12
+  BIC $0xfc000000, R6, R6
+  ADD R12, R1, R7
+  MOVW R7>>26, R12
+  BIC $0xfc000000, R7, R7
+  ADD R12, R2, g
+  MOVW g>>26, R12
+  BIC $0xfc000000, g, g
+  ADD R12, R3, R11
+  MOVW $-(1<<26), R12
+  ADD R11>>26, R12, R12
+  BIC $0xfc000000, R11, R11
+  ADD R12, R4, R14
+  MOVW R14>>31, R12
+  SUB $1, R12
+  AND R12, R6, R6
+  AND R12, R7, R7
+  AND R12, g, g
+  AND R12, R11, R11
+  AND R12, R14, R14
+  MVN R12, R12
+  AND R12, R0, R0
+  AND R12, R1, R1
+  AND R12, R2, R2
+  AND R12, R3, R3
+  AND R12, R4, R4
+  ORR R6, R0, R0
+  ORR R7, R1, R1
+  ORR g, R2, R2
+  ORR R11, R3, R3
+  ORR R14, R4, R4
+  ORR R1<<26, R0, R0
+  MOVW R1>>6, R1
+  ORR R2<<20, R1, R1
+  MOVW R2>>12, R2
+  ORR R3<<14, R2, R2
+  MOVW R3>>18, R3
+  ORR R4<<8, R3, R3
+  MOVW 40(R5), R6
+  MOVW 44(R5), R7
+  MOVW 48(R5), g
+  MOVW 52(R5), R11
+  ADD.S R6, R0, R0
+  ADC.S R7, R1, R1
+  ADC.S g, R2, R2
+  ADC.S R11, R3, R3
+  MOVM.IA [R0-R3], (R8)
+  MOVW R5, R12
+  EOR R0, R0, R0
+  EOR R1, R1, R1
+  EOR R2, R2, R2
+  EOR R3, R3, R3
+  EOR R4, R4, R4
+  EOR R5, R5, R5
+  EOR R6, R6, R6
+  EOR R7, R7, R7
+  MOVM.IA.W [R0-R7], (R12)
+  MOVM.IA [R0-R7], (R12)
+  ADD $16, R13, R13
+  MOVM.IA.W (R13), [R4, R5, R6, R7, R8, R9, g, R11, R14]
+  RET
+
+// func poly1305_auth_armv6(out *[16]byte, m *byte, mlen uint32, key *[32]key)
+TEXT ·poly1305_auth_armv6(SB),0,$280-16
+  MOVW  out+0(FP), R4
+  MOVW  m+4(FP), R5
+  MOVW  mlen+8(FP), R6
+  MOVW  key+12(FP), R7
+
+  MOVW R13, R8
+  BIC $63, R13
+  SUB $64, R13, R13
+  MOVW  R13, R0
+  MOVW  R7, R1
+  BL poly1305_init_ext_armv6<>(SB)
+  BIC.S $15, R6, R2
+  BEQ poly1305_auth_armv6_noblocks
+  MOVW R13, R0
+  MOVW R5, R1
+  ADD R2, R5, R5
+  SUB R2, R6, R6
+  BL poly1305_blocks_armv6<>(SB)
+poly1305_auth_armv6_noblocks:
+  MOVW R13, R0
+  MOVW R5, R1
+  MOVW R6, R2
+  MOVW R4, R3
+  BL poly1305_finish_ext_armv6<>(SB)
+  MOVW R8, R13
+  RET

Godeps/_workspace/src/golang.org/x/crypto/poly1305/poly1305_test.go

@@ -0,0 +1,74 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package poly1305
+
+import (
+	"bytes"
+	"testing"
+)
+
+var testData = []struct {
+	in, k, correct []byte
+}{
+	{
+		[]byte("Hello world!"),
+		[]byte("this is 32-byte key for Poly1305"),
+		[]byte{0xa6, 0xf7, 0x45, 0x00, 0x8f, 0x81, 0xc9, 0x16, 0xa2, 0x0d, 0xcc, 0x74, 0xee, 0xf2, 0xb2, 0xf0},
+	},
+	{
+		make([]byte, 32),
+		[]byte("this is 32-byte key for Poly1305"),
+		[]byte{0x49, 0xec, 0x78, 0x09, 0x0e, 0x48, 0x1e, 0xc6, 0xc2, 0x6b, 0x33, 0xb9, 0x1c, 0xcc, 0x03, 0x07},
+	},
+	{
+		make([]byte, 2007),
+		[]byte("this is 32-byte key for Poly1305"),
+		[]byte{0xda, 0x84, 0xbc, 0xab, 0x02, 0x67, 0x6c, 0x38, 0xcd, 0xb0, 0x15, 0x60, 0x42, 0x74, 0xc2, 0xaa},
+	},
+	{
+		make([]byte, 2007),
+		make([]byte, 32),
+		make([]byte, 16),
+	},
+}
+
+func TestSum(t *testing.T) {
+	var out [16]byte
+	var key [32]byte
+
+	for i, v := range testData {
+		copy(key[:], v.k)
+		Sum(&out, v.in, &key)
+		if !bytes.Equal(out[:], v.correct) {
+			t.Errorf("%d: expected %x, got %x", i, v.correct, out[:])
+		}
+	}
+}
+
+func Benchmark1K(b *testing.B) {
+	b.StopTimer()
+	var out [16]byte
+	var key [32]byte
+	in := make([]byte, 1024)
+	b.SetBytes(int64(len(in)))
+	b.StartTimer()
+
+	for i := 0; i < b.N; i++ {
+		Sum(&out, in, &key)
+	}
+}
+
+func Benchmark64(b *testing.B) {
+	b.StopTimer()
+	var out [16]byte
+	var key [32]byte
+	in := make([]byte, 64)
+	b.SetBytes(int64(len(in)))
+	b.StartTimer()
+
+	for i := 0; i < b.N; i++ {
+		Sum(&out, in, &key)
+	}
+}

Godeps/_workspace/src/golang.org/x/crypto/poly1305/sum_amd64.go

@@ -0,0 +1,24 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build amd64,!gccgo,!appengine
+
+package poly1305
+
+// This function is implemented in poly1305_amd64.s
+
+//go:noescape
+
+func poly1305(out *[16]byte, m *byte, mlen uint64, key *[32]byte)
+
+// Sum generates an authenticator for m using a one-time key and puts the
+// 16-byte result into out. Authenticating two different messages with the same
+// key allows an attacker to forge messages at will.
+func Sum(out *[16]byte, m []byte, key *[32]byte) {
+	var mPtr *byte
+	if len(m) > 0 {
+		mPtr = &m[0]
+	}
+	poly1305(out, mPtr, uint64(len(m)), key)
+}

Godeps/_workspace/src/golang.org/x/crypto/poly1305/sum_arm.go

@@ -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.
+
+// +build arm,!gccgo,!appengine
+
+package poly1305
+
+// This function is implemented in poly1305_arm.s
+
+//go:noescape
+
+func poly1305_auth_armv6(out *[16]byte, m *byte, mlen uint32, key *[32]byte)
+
+// Sum generates an authenticator for m using a one-time key and puts the
+// 16-byte result into out. Authenticating two different messages with the same
+// key allows an attacker to forge messages at will.
+func Sum(out *[16]byte, m []byte, key *[32]byte) {
+	var mPtr *byte
+	if len(m) > 0 {
+		mPtr = &m[0]
+	}
+	poly1305_auth_armv6(out, mPtr, uint32(len(m)), key)
+}

Godeps/_workspace/src/golang.org/x/crypto/ripemd160/ripemd160.go

@@ -0,0 +1,120 @@
+// Copyright 2010 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 ripemd160 implements the RIPEMD-160 hash algorithm.
+package ripemd160
+
+// RIPEMD-160 is designed by by Hans Dobbertin, Antoon Bosselaers, and Bart
+// Preneel with specifications available at:
+// http://homes.esat.kuleuven.be/~cosicart/pdf/AB-9601/AB-9601.pdf.
+
+import (
+	"crypto"
+	"hash"
+)
+
+func init() {
+	crypto.RegisterHash(crypto.RIPEMD160, New)
+}
+
+// The size of the checksum in bytes.
+const Size = 20
+
+// The block size of the hash algorithm in bytes.
+const BlockSize = 64
+
+const (
+	_s0 = 0x67452301
+	_s1 = 0xefcdab89
+	_s2 = 0x98badcfe
+	_s3 = 0x10325476
+	_s4 = 0xc3d2e1f0
+)
+
+// digest represents the partial evaluation of a checksum.
+type digest struct {
+	s  [5]uint32       // running context
+	x  [BlockSize]byte // temporary buffer
+	nx int             // index into x
+	tc uint64          // total count of bytes processed
+}
+
+func (d *digest) Reset() {
+	d.s[0], d.s[1], d.s[2], d.s[3], d.s[4] = _s0, _s1, _s2, _s3, _s4
+	d.nx = 0
+	d.tc = 0
+}
+
+// New returns a new hash.Hash computing the checksum.
+func New() hash.Hash {
+	result := new(digest)
+	result.Reset()
+	return result
+}
+
+func (d *digest) Size() int { return Size }
+
+func (d *digest) BlockSize() int { return BlockSize }
+
+func (d *digest) Write(p []byte) (nn int, err error) {
+	nn = len(p)
+	d.tc += uint64(nn)
+	if d.nx > 0 {
+		n := len(p)
+		if n > BlockSize-d.nx {
+			n = BlockSize - d.nx
+		}
+		for i := 0; i < n; i++ {
+			d.x[d.nx+i] = p[i]
+		}
+		d.nx += n
+		if d.nx == BlockSize {
+			_Block(d, d.x[0:])
+			d.nx = 0
+		}
+		p = p[n:]
+	}
+	n := _Block(d, p)
+	p = p[n:]
+	if len(p) > 0 {
+		d.nx = copy(d.x[:], p)
+	}
+	return
+}
+
+func (d0 *digest) Sum(in []byte) []byte {
+	// Make a copy of d0 so that caller can keep writing and summing.
+	d := *d0
+
+	// Padding.  Add a 1 bit and 0 bits until 56 bytes mod 64.
+	tc := d.tc
+	var tmp [64]byte
+	tmp[0] = 0x80
+	if tc%64 < 56 {
+		d.Write(tmp[0 : 56-tc%64])
+	} else {
+		d.Write(tmp[0 : 64+56-tc%64])
+	}
+
+	// Length in bits.
+	tc <<= 3
+	for i := uint(0); i < 8; i++ {
+		tmp[i] = byte(tc >> (8 * i))
+	}
+	d.Write(tmp[0:8])
+
+	if d.nx != 0 {
+		panic("d.nx != 0")
+	}
+
+	var digest [Size]byte
+	for i, s := range d.s {
+		digest[i*4] = byte(s)
+		digest[i*4+1] = byte(s >> 8)
+		digest[i*4+2] = byte(s >> 16)
+		digest[i*4+3] = byte(s >> 24)
+	}
+
+	return append(in, digest[:]...)
+}

Godeps/_workspace/src/golang.org/x/crypto/ripemd160/ripemd160_test.go

@@ -0,0 +1,64 @@
+// Copyright 2010 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 ripemd160
+
+// Test vectors are from:
+// http://homes.esat.kuleuven.be/~bosselae/ripemd160.html
+
+import (
+	"fmt"
+	"io"
+	"testing"
+)
+
+type mdTest struct {
+	out string
+	in  string
+}
+
+var vectors = [...]mdTest{
+	{"9c1185a5c5e9fc54612808977ee8f548b2258d31", ""},
+	{"0bdc9d2d256b3ee9daae347be6f4dc835a467ffe", "a"},
+	{"8eb208f7e05d987a9b044a8e98c6b087f15a0bfc", "abc"},
+	{"5d0689ef49d2fae572b881b123a85ffa21595f36", "message digest"},
+	{"f71c27109c692c1b56bbdceb5b9d2865b3708dbc", "abcdefghijklmnopqrstuvwxyz"},
+	{"12a053384a9c0c88e405a06c27dcf49ada62eb2b", "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq"},
+	{"b0e20b6e3116640286ed3a87a5713079b21f5189", "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789"},
+	{"9b752e45573d4b39f4dbd3323cab82bf63326bfb", "12345678901234567890123456789012345678901234567890123456789012345678901234567890"},
+}
+
+func TestVectors(t *testing.T) {
+	for i := 0; i < len(vectors); i++ {
+		tv := vectors[i]
+		md := New()
+		for j := 0; j < 3; j++ {
+			if j < 2 {
+				io.WriteString(md, tv.in)
+			} else {
+				io.WriteString(md, tv.in[0:len(tv.in)/2])
+				md.Sum(nil)
+				io.WriteString(md, tv.in[len(tv.in)/2:])
+			}
+			s := fmt.Sprintf("%x", md.Sum(nil))
+			if s != tv.out {
+				t.Fatalf("RIPEMD-160[%d](%s) = %s, expected %s", j, tv.in, s, tv.out)
+			}
+			md.Reset()
+		}
+	}
+}
+
+func TestMillionA(t *testing.T) {
+	md := New()
+	for i := 0; i < 100000; i++ {
+		io.WriteString(md, "aaaaaaaaaa")
+	}
+	out := "52783243c1697bdbe16d37f97f68f08325dc1528"
+	s := fmt.Sprintf("%x", md.Sum(nil))
+	if s != out {
+		t.Fatalf("RIPEMD-160 (1 million 'a') = %s, expected %s", s, out)
+	}
+	md.Reset()
+}

Godeps/_workspace/src/golang.org/x/crypto/ripemd160/ripemd160block.go

@@ -0,0 +1,161 @@
+// Copyright 2010 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.
+
+// RIPEMD-160 block step.
+// In its own file so that a faster assembly or C version
+// can be substituted easily.
+
+package ripemd160
+
+// work buffer indices and roll amounts for one line
+var _n = [80]uint{
+	0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
+	7, 4, 13, 1, 10, 6, 15, 3, 12, 0, 9, 5, 2, 14, 11, 8,
+	3, 10, 14, 4, 9, 15, 8, 1, 2, 7, 0, 6, 13, 11, 5, 12,
+	1, 9, 11, 10, 0, 8, 12, 4, 13, 3, 7, 15, 14, 5, 6, 2,
+	4, 0, 5, 9, 7, 12, 2, 10, 14, 1, 3, 8, 11, 6, 15, 13,
+}
+
+var _r = [80]uint{
+	11, 14, 15, 12, 5, 8, 7, 9, 11, 13, 14, 15, 6, 7, 9, 8,
+	7, 6, 8, 13, 11, 9, 7, 15, 7, 12, 15, 9, 11, 7, 13, 12,
+	11, 13, 6, 7, 14, 9, 13, 15, 14, 8, 13, 6, 5, 12, 7, 5,
+	11, 12, 14, 15, 14, 15, 9, 8, 9, 14, 5, 6, 8, 6, 5, 12,
+	9, 15, 5, 11, 6, 8, 13, 12, 5, 12, 13, 14, 11, 8, 5, 6,
+}
+
+// same for the other parallel one
+var n_ = [80]uint{
+	5, 14, 7, 0, 9, 2, 11, 4, 13, 6, 15, 8, 1, 10, 3, 12,
+	6, 11, 3, 7, 0, 13, 5, 10, 14, 15, 8, 12, 4, 9, 1, 2,
+	15, 5, 1, 3, 7, 14, 6, 9, 11, 8, 12, 2, 10, 0, 4, 13,
+	8, 6, 4, 1, 3, 11, 15, 0, 5, 12, 2, 13, 9, 7, 10, 14,
+	12, 15, 10, 4, 1, 5, 8, 7, 6, 2, 13, 14, 0, 3, 9, 11,
+}
+
+var r_ = [80]uint{
+	8, 9, 9, 11, 13, 15, 15, 5, 7, 7, 8, 11, 14, 14, 12, 6,
+	9, 13, 15, 7, 12, 8, 9, 11, 7, 7, 12, 7, 6, 15, 13, 11,
+	9, 7, 15, 11, 8, 6, 6, 14, 12, 13, 5, 14, 13, 13, 7, 5,
+	15, 5, 8, 11, 14, 14, 6, 14, 6, 9, 12, 9, 12, 5, 15, 8,
+	8, 5, 12, 9, 12, 5, 14, 6, 8, 13, 6, 5, 15, 13, 11, 11,
+}
+
+func _Block(md *digest, p []byte) int {
+	n := 0
+	var x [16]uint32
+	var alpha, beta uint32
+	for len(p) >= BlockSize {
+		a, b, c, d, e := md.s[0], md.s[1], md.s[2], md.s[3], md.s[4]
+		aa, bb, cc, dd, ee := a, b, c, d, e
+		j := 0
+		for i := 0; i < 16; i++ {
+			x[i] = uint32(p[j]) | uint32(p[j+1])<<8 | uint32(p[j+2])<<16 | uint32(p[j+3])<<24
+			j += 4
+		}
+
+		// round 1
+		i := 0
+		for i < 16 {
+			alpha = a + (b ^ c ^ d) + x[_n[i]]
+			s := _r[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + e
+			beta = c<<10 | c>>22
+			a, b, c, d, e = e, alpha, b, beta, d
+
+			// parallel line
+			alpha = aa + (bb ^ (cc | ^dd)) + x[n_[i]] + 0x50a28be6
+			s = r_[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + ee
+			beta = cc<<10 | cc>>22
+			aa, bb, cc, dd, ee = ee, alpha, bb, beta, dd
+
+			i++
+		}
+
+		// round 2
+		for i < 32 {
+			alpha = a + (b&c | ^b&d) + x[_n[i]] + 0x5a827999
+			s := _r[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + e
+			beta = c<<10 | c>>22
+			a, b, c, d, e = e, alpha, b, beta, d
+
+			// parallel line
+			alpha = aa + (bb&dd | cc&^dd) + x[n_[i]] + 0x5c4dd124
+			s = r_[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + ee
+			beta = cc<<10 | cc>>22
+			aa, bb, cc, dd, ee = ee, alpha, bb, beta, dd
+
+			i++
+		}
+
+		// round 3
+		for i < 48 {
+			alpha = a + (b | ^c ^ d) + x[_n[i]] + 0x6ed9eba1
+			s := _r[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + e
+			beta = c<<10 | c>>22
+			a, b, c, d, e = e, alpha, b, beta, d
+
+			// parallel line
+			alpha = aa + (bb | ^cc ^ dd) + x[n_[i]] + 0x6d703ef3
+			s = r_[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + ee
+			beta = cc<<10 | cc>>22
+			aa, bb, cc, dd, ee = ee, alpha, bb, beta, dd
+
+			i++
+		}
+
+		// round 4
+		for i < 64 {
+			alpha = a + (b&d | c&^d) + x[_n[i]] + 0x8f1bbcdc
+			s := _r[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + e
+			beta = c<<10 | c>>22
+			a, b, c, d, e = e, alpha, b, beta, d
+
+			// parallel line
+			alpha = aa + (bb&cc | ^bb&dd) + x[n_[i]] + 0x7a6d76e9
+			s = r_[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + ee
+			beta = cc<<10 | cc>>22
+			aa, bb, cc, dd, ee = ee, alpha, bb, beta, dd
+
+			i++
+		}
+
+		// round 5
+		for i < 80 {
+			alpha = a + (b ^ (c | ^d)) + x[_n[i]] + 0xa953fd4e
+			s := _r[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + e
+			beta = c<<10 | c>>22
+			a, b, c, d, e = e, alpha, b, beta, d
+
+			// parallel line
+			alpha = aa + (bb ^ cc ^ dd) + x[n_[i]]
+			s = r_[i]
+			alpha = (alpha<<s | alpha>>(32-s)) + ee
+			beta = cc<<10 | cc>>22
+			aa, bb, cc, dd, ee = ee, alpha, bb, beta, dd
+
+			i++
+		}
+
+		// combine results
+		dd += c + md.s[1]
+		md.s[1] = md.s[2] + d + ee
+		md.s[2] = md.s[3] + e + aa
+		md.s[3] = md.s[4] + a + bb
+		md.s[4] = md.s[0] + b + cc
+		md.s[0] = dd
+
+		p = p[BlockSize:]
+		n += BlockSize
+	}
+	return n
+}

Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa/hsalsa20.go

@@ -0,0 +1,144 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package salsa provides low-level access to functions in the Salsa family.
+package salsa
+
+// Sigma is the Salsa20 constant for 256-bit keys.
+var Sigma = [16]byte{'e', 'x', 'p', 'a', 'n', 'd', ' ', '3', '2', '-', 'b', 'y', 't', 'e', ' ', 'k'}
+
+// HSalsa20 applies the HSalsa20 core function to a 16-byte input in, 32-byte
+// key k, and 16-byte constant c, and puts the result into the 32-byte array
+// out.
+func HSalsa20(out *[32]byte, in *[16]byte, k *[32]byte, c *[16]byte) {
+	x0 := uint32(c[0]) | uint32(c[1])<<8 | uint32(c[2])<<16 | uint32(c[3])<<24
+	x1 := uint32(k[0]) | uint32(k[1])<<8 | uint32(k[2])<<16 | uint32(k[3])<<24
+	x2 := uint32(k[4]) | uint32(k[5])<<8 | uint32(k[6])<<16 | uint32(k[7])<<24
+	x3 := uint32(k[8]) | uint32(k[9])<<8 | uint32(k[10])<<16 | uint32(k[11])<<24
+	x4 := uint32(k[12]) | uint32(k[13])<<8 | uint32(k[14])<<16 | uint32(k[15])<<24
+	x5 := uint32(c[4]) | uint32(c[5])<<8 | uint32(c[6])<<16 | uint32(c[7])<<24
+	x6 := uint32(in[0]) | uint32(in[1])<<8 | uint32(in[2])<<16 | uint32(in[3])<<24
+	x7 := uint32(in[4]) | uint32(in[5])<<8 | uint32(in[6])<<16 | uint32(in[7])<<24
+	x8 := uint32(in[8]) | uint32(in[9])<<8 | uint32(in[10])<<16 | uint32(in[11])<<24
+	x9 := uint32(in[12]) | uint32(in[13])<<8 | uint32(in[14])<<16 | uint32(in[15])<<24
+	x10 := uint32(c[8]) | uint32(c[9])<<8 | uint32(c[10])<<16 | uint32(c[11])<<24
+	x11 := uint32(k[16]) | uint32(k[17])<<8 | uint32(k[18])<<16 | uint32(k[19])<<24
+	x12 := uint32(k[20]) | uint32(k[21])<<8 | uint32(k[22])<<16 | uint32(k[23])<<24
+	x13 := uint32(k[24]) | uint32(k[25])<<8 | uint32(k[26])<<16 | uint32(k[27])<<24
+	x14 := uint32(k[28]) | uint32(k[29])<<8 | uint32(k[30])<<16 | uint32(k[31])<<24
+	x15 := uint32(c[12]) | uint32(c[13])<<8 | uint32(c[14])<<16 | uint32(c[15])<<24
+
+	for i := 0; i < 20; i += 2 {
+		u := x0 + x12
+		x4 ^= u<<7 | u>>(32-7)
+		u = x4 + x0
+		x8 ^= u<<9 | u>>(32-9)
+		u = x8 + x4
+		x12 ^= u<<13 | u>>(32-13)
+		u = x12 + x8
+		x0 ^= u<<18 | u>>(32-18)
+
+		u = x5 + x1
+		x9 ^= u<<7 | u>>(32-7)
+		u = x9 + x5
+		x13 ^= u<<9 | u>>(32-9)
+		u = x13 + x9
+		x1 ^= u<<13 | u>>(32-13)
+		u = x1 + x13
+		x5 ^= u<<18 | u>>(32-18)
+
+		u = x10 + x6
+		x14 ^= u<<7 | u>>(32-7)
+		u = x14 + x10
+		x2 ^= u<<9 | u>>(32-9)
+		u = x2 + x14
+		x6 ^= u<<13 | u>>(32-13)
+		u = x6 + x2
+		x10 ^= u<<18 | u>>(32-18)
+
+		u = x15 + x11
+		x3 ^= u<<7 | u>>(32-7)
+		u = x3 + x15
+		x7 ^= u<<9 | u>>(32-9)
+		u = x7 + x3
+		x11 ^= u<<13 | u>>(32-13)
+		u = x11 + x7
+		x15 ^= u<<18 | u>>(32-18)
+
+		u = x0 + x3
+		x1 ^= u<<7 | u>>(32-7)
+		u = x1 + x0
+		x2 ^= u<<9 | u>>(32-9)
+		u = x2 + x1
+		x3 ^= u<<13 | u>>(32-13)
+		u = x3 + x2
+		x0 ^= u<<18 | u>>(32-18)
+
+		u = x5 + x4
+		x6 ^= u<<7 | u>>(32-7)
+		u = x6 + x5
+		x7 ^= u<<9 | u>>(32-9)
+		u = x7 + x6
+		x4 ^= u<<13 | u>>(32-13)
+		u = x4 + x7
+		x5 ^= u<<18 | u>>(32-18)
+
+		u = x10 + x9
+		x11 ^= u<<7 | u>>(32-7)
+		u = x11 + x10
+		x8 ^= u<<9 | u>>(32-9)
+		u = x8 + x11
+		x9 ^= u<<13 | u>>(32-13)
+		u = x9 + x8
+		x10 ^= u<<18 | u>>(32-18)
+
+		u = x15 + x14
+		x12 ^= u<<7 | u>>(32-7)
+		u = x12 + x15
+		x13 ^= u<<9 | u>>(32-9)
+		u = x13 + x12
+		x14 ^= u<<13 | u>>(32-13)
+		u = x14 + x13
+		x15 ^= u<<18 | u>>(32-18)
+	}
+	out[0] = byte(x0)
+	out[1] = byte(x0 >> 8)
+	out[2] = byte(x0 >> 16)
+	out[3] = byte(x0 >> 24)
+
+	out[4] = byte(x5)
+	out[5] = byte(x5 >> 8)
+	out[6] = byte(x5 >> 16)
+	out[7] = byte(x5 >> 24)
+
+	out[8] = byte(x10)
+	out[9] = byte(x10 >> 8)
+	out[10] = byte(x10 >> 16)
+	out[11] = byte(x10 >> 24)
+
+	out[12] = byte(x15)
+	out[13] = byte(x15 >> 8)
+	out[14] = byte(x15 >> 16)
+	out[15] = byte(x15 >> 24)
+
+	out[16] = byte(x6)
+	out[17] = byte(x6 >> 8)
+	out[18] = byte(x6 >> 16)
+	out[19] = byte(x6 >> 24)
+
+	out[20] = byte(x7)
+	out[21] = byte(x7 >> 8)
+	out[22] = byte(x7 >> 16)
+	out[23] = byte(x7 >> 24)
+
+	out[24] = byte(x8)
+	out[25] = byte(x8 >> 8)
+	out[26] = byte(x8 >> 16)
+	out[27] = byte(x8 >> 24)
+
+	out[28] = byte(x9)
+	out[29] = byte(x9 >> 8)
+	out[30] = byte(x9 >> 16)
+	out[31] = byte(x9 >> 24)
+}

Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa/salsa2020_amd64.s

@@ -0,0 +1,902 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build amd64,!appengine,!gccgo
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources in SUPERCOP: http://bench.cr.yp.to/supercop.html
+
+// func salsa2020XORKeyStream(out, in *byte, n uint64, nonce, key *byte)
+TEXT ·salsa2020XORKeyStream(SB),0,$512-40
+	MOVQ out+0(FP),DI
+	MOVQ in+8(FP),SI
+	MOVQ n+16(FP),DX
+	MOVQ nonce+24(FP),CX
+	MOVQ key+32(FP),R8
+
+	MOVQ SP,R11
+	MOVQ $31,R9
+	NOTQ R9
+	ANDQ R9,SP
+	ADDQ $32,SP
+
+	MOVQ R11,352(SP)
+	MOVQ R12,360(SP)
+	MOVQ R13,368(SP)
+	MOVQ R14,376(SP)
+	MOVQ R15,384(SP)
+	MOVQ BX,392(SP)
+	MOVQ BP,400(SP)
+	MOVQ DX,R9
+	MOVQ CX,DX
+	MOVQ R8,R10
+	CMPQ R9,$0
+	JBE DONE
+	START:
+	MOVL 20(R10),CX
+	MOVL 0(R10),R8
+	MOVL 0(DX),AX
+	MOVL 16(R10),R11
+	MOVL CX,0(SP)
+	MOVL R8, 4 (SP)
+	MOVL AX, 8 (SP)
+	MOVL R11, 12 (SP)
+	MOVL 8(DX),CX
+	MOVL 24(R10),R8
+	MOVL 4(R10),AX
+	MOVL 4(DX),R11
+	MOVL CX,16(SP)
+	MOVL R8, 20 (SP)
+	MOVL AX, 24 (SP)
+	MOVL R11, 28 (SP)
+	MOVL 12(DX),CX
+	MOVL 12(R10),DX
+	MOVL 28(R10),R8
+	MOVL 8(R10),AX
+	MOVL DX,32(SP)
+	MOVL CX, 36 (SP)
+	MOVL R8, 40 (SP)
+	MOVL AX, 44 (SP)
+	MOVQ $1634760805,DX
+	MOVQ $857760878,CX
+	MOVQ $2036477234,R8
+	MOVQ $1797285236,AX
+	MOVL DX,48(SP)
+	MOVL CX, 52 (SP)
+	MOVL R8, 56 (SP)
+	MOVL AX, 60 (SP)
+	CMPQ R9,$256
+	JB BYTESBETWEEN1AND255
+	MOVOA 48(SP),X0
+	PSHUFL $0X55,X0,X1
+	PSHUFL $0XAA,X0,X2
+	PSHUFL $0XFF,X0,X3
+	PSHUFL $0X00,X0,X0
+	MOVOA X1,64(SP)
+	MOVOA X2,80(SP)
+	MOVOA X3,96(SP)
+	MOVOA X0,112(SP)
+	MOVOA 0(SP),X0
+	PSHUFL $0XAA,X0,X1
+	PSHUFL $0XFF,X0,X2
+	PSHUFL $0X00,X0,X3
+	PSHUFL $0X55,X0,X0
+	MOVOA X1,128(SP)
+	MOVOA X2,144(SP)
+	MOVOA X3,160(SP)
+	MOVOA X0,176(SP)
+	MOVOA 16(SP),X0
+	PSHUFL $0XFF,X0,X1
+	PSHUFL $0X55,X0,X2
+	PSHUFL $0XAA,X0,X0
+	MOVOA X1,192(SP)
+	MOVOA X2,208(SP)
+	MOVOA X0,224(SP)
+	MOVOA 32(SP),X0
+	PSHUFL $0X00,X0,X1
+	PSHUFL $0XAA,X0,X2
+	PSHUFL $0XFF,X0,X0
+	MOVOA X1,240(SP)
+	MOVOA X2,256(SP)
+	MOVOA X0,272(SP)
+	BYTESATLEAST256:
+	MOVL 16(SP),DX
+	MOVL  36 (SP),CX
+	MOVL DX,288(SP)
+	MOVL CX,304(SP)
+	ADDQ $1,DX
+	SHLQ $32,CX
+	ADDQ CX,DX
+	MOVQ DX,CX
+	SHRQ $32,CX
+	MOVL DX, 292 (SP)
+	MOVL CX, 308 (SP)
+	ADDQ $1,DX
+	SHLQ $32,CX
+	ADDQ CX,DX
+	MOVQ DX,CX
+	SHRQ $32,CX
+	MOVL DX, 296 (SP)
+	MOVL CX, 312 (SP)
+	ADDQ $1,DX
+	SHLQ $32,CX
+	ADDQ CX,DX
+	MOVQ DX,CX
+	SHRQ $32,CX
+	MOVL DX, 300 (SP)
+	MOVL CX, 316 (SP)
+	ADDQ $1,DX
+	SHLQ $32,CX
+	ADDQ CX,DX
+	MOVQ DX,CX
+	SHRQ $32,CX
+	MOVL DX,16(SP)
+	MOVL CX, 36 (SP)
+	MOVQ R9,408(SP)
+	MOVQ $20,DX
+	MOVOA 64(SP),X0
+	MOVOA 80(SP),X1
+	MOVOA 96(SP),X2
+	MOVOA 256(SP),X3
+	MOVOA 272(SP),X4
+	MOVOA 128(SP),X5
+	MOVOA 144(SP),X6
+	MOVOA 176(SP),X7
+	MOVOA 192(SP),X8
+	MOVOA 208(SP),X9
+	MOVOA 224(SP),X10
+	MOVOA 304(SP),X11
+	MOVOA 112(SP),X12
+	MOVOA 160(SP),X13
+	MOVOA 240(SP),X14
+	MOVOA 288(SP),X15
+	MAINLOOP1:
+	MOVOA X1,320(SP)
+	MOVOA X2,336(SP)
+	MOVOA X13,X1
+	PADDL X12,X1
+	MOVOA X1,X2
+	PSLLL $7,X1
+	PXOR X1,X14
+	PSRLL $25,X2
+	PXOR X2,X14
+	MOVOA X7,X1
+	PADDL X0,X1
+	MOVOA X1,X2
+	PSLLL $7,X1
+	PXOR X1,X11
+	PSRLL $25,X2
+	PXOR X2,X11
+	MOVOA X12,X1
+	PADDL X14,X1
+	MOVOA X1,X2
+	PSLLL $9,X1
+	PXOR X1,X15
+	PSRLL $23,X2
+	PXOR X2,X15
+	MOVOA X0,X1
+	PADDL X11,X1
+	MOVOA X1,X2
+	PSLLL $9,X1
+	PXOR X1,X9
+	PSRLL $23,X2
+	PXOR X2,X9
+	MOVOA X14,X1
+	PADDL X15,X1
+	MOVOA X1,X2
+	PSLLL $13,X1
+	PXOR X1,X13
+	PSRLL $19,X2
+	PXOR X2,X13
+	MOVOA X11,X1
+	PADDL X9,X1
+	MOVOA X1,X2
+	PSLLL $13,X1
+	PXOR X1,X7
+	PSRLL $19,X2
+	PXOR X2,X7
+	MOVOA X15,X1
+	PADDL X13,X1
+	MOVOA X1,X2
+	PSLLL $18,X1
+	PXOR X1,X12
+	PSRLL $14,X2
+	PXOR X2,X12
+	MOVOA 320(SP),X1
+	MOVOA X12,320(SP)
+	MOVOA X9,X2
+	PADDL X7,X2
+	MOVOA X2,X12
+	PSLLL $18,X2
+	PXOR X2,X0
+	PSRLL $14,X12
+	PXOR X12,X0
+	MOVOA X5,X2
+	PADDL X1,X2
+	MOVOA X2,X12
+	PSLLL $7,X2
+	PXOR X2,X3
+	PSRLL $25,X12
+	PXOR X12,X3
+	MOVOA 336(SP),X2
+	MOVOA X0,336(SP)
+	MOVOA X6,X0
+	PADDL X2,X0
+	MOVOA X0,X12
+	PSLLL $7,X0
+	PXOR X0,X4
+	PSRLL $25,X12
+	PXOR X12,X4
+	MOVOA X1,X0
+	PADDL X3,X0
+	MOVOA X0,X12
+	PSLLL $9,X0
+	PXOR X0,X10
+	PSRLL $23,X12
+	PXOR X12,X10
+	MOVOA X2,X0
+	PADDL X4,X0
+	MOVOA X0,X12
+	PSLLL $9,X0
+	PXOR X0,X8
+	PSRLL $23,X12
+	PXOR X12,X8
+	MOVOA X3,X0
+	PADDL X10,X0
+	MOVOA X0,X12
+	PSLLL $13,X0
+	PXOR X0,X5
+	PSRLL $19,X12
+	PXOR X12,X5
+	MOVOA X4,X0
+	PADDL X8,X0
+	MOVOA X0,X12
+	PSLLL $13,X0
+	PXOR X0,X6
+	PSRLL $19,X12
+	PXOR X12,X6
+	MOVOA X10,X0
+	PADDL X5,X0
+	MOVOA X0,X12
+	PSLLL $18,X0
+	PXOR X0,X1
+	PSRLL $14,X12
+	PXOR X12,X1
+	MOVOA 320(SP),X0
+	MOVOA X1,320(SP)
+	MOVOA X4,X1
+	PADDL X0,X1
+	MOVOA X1,X12
+	PSLLL $7,X1
+	PXOR X1,X7
+	PSRLL $25,X12
+	PXOR X12,X7
+	MOVOA X8,X1
+	PADDL X6,X1
+	MOVOA X1,X12
+	PSLLL $18,X1
+	PXOR X1,X2
+	PSRLL $14,X12
+	PXOR X12,X2
+	MOVOA 336(SP),X12
+	MOVOA X2,336(SP)
+	MOVOA X14,X1
+	PADDL X12,X1
+	MOVOA X1,X2
+	PSLLL $7,X1
+	PXOR X1,X5
+	PSRLL $25,X2
+	PXOR X2,X5
+	MOVOA X0,X1
+	PADDL X7,X1
+	MOVOA X1,X2
+	PSLLL $9,X1
+	PXOR X1,X10
+	PSRLL $23,X2
+	PXOR X2,X10
+	MOVOA X12,X1
+	PADDL X5,X1
+	MOVOA X1,X2
+	PSLLL $9,X1
+	PXOR X1,X8
+	PSRLL $23,X2
+	PXOR X2,X8
+	MOVOA X7,X1
+	PADDL X10,X1
+	MOVOA X1,X2
+	PSLLL $13,X1
+	PXOR X1,X4
+	PSRLL $19,X2
+	PXOR X2,X4
+	MOVOA X5,X1
+	PADDL X8,X1
+	MOVOA X1,X2
+	PSLLL $13,X1
+	PXOR X1,X14
+	PSRLL $19,X2
+	PXOR X2,X14
+	MOVOA X10,X1
+	PADDL X4,X1
+	MOVOA X1,X2
+	PSLLL $18,X1
+	PXOR X1,X0
+	PSRLL $14,X2
+	PXOR X2,X0
+	MOVOA 320(SP),X1
+	MOVOA X0,320(SP)
+	MOVOA X8,X0
+	PADDL X14,X0
+	MOVOA X0,X2
+	PSLLL $18,X0
+	PXOR X0,X12
+	PSRLL $14,X2
+	PXOR X2,X12
+	MOVOA X11,X0
+	PADDL X1,X0
+	MOVOA X0,X2
+	PSLLL $7,X0
+	PXOR X0,X6
+	PSRLL $25,X2
+	PXOR X2,X6
+	MOVOA 336(SP),X2
+	MOVOA X12,336(SP)
+	MOVOA X3,X0
+	PADDL X2,X0
+	MOVOA X0,X12
+	PSLLL $7,X0
+	PXOR X0,X13
+	PSRLL $25,X12
+	PXOR X12,X13
+	MOVOA X1,X0
+	PADDL X6,X0
+	MOVOA X0,X12
+	PSLLL $9,X0
+	PXOR X0,X15
+	PSRLL $23,X12
+	PXOR X12,X15
+	MOVOA X2,X0
+	PADDL X13,X0
+	MOVOA X0,X12
+	PSLLL $9,X0
+	PXOR X0,X9
+	PSRLL $23,X12
+	PXOR X12,X9
+	MOVOA X6,X0
+	PADDL X15,X0
+	MOVOA X0,X12
+	PSLLL $13,X0
+	PXOR X0,X11
+	PSRLL $19,X12
+	PXOR X12,X11
+	MOVOA X13,X0
+	PADDL X9,X0
+	MOVOA X0,X12
+	PSLLL $13,X0
+	PXOR X0,X3
+	PSRLL $19,X12
+	PXOR X12,X3
+	MOVOA X15,X0
+	PADDL X11,X0
+	MOVOA X0,X12
+	PSLLL $18,X0
+	PXOR X0,X1
+	PSRLL $14,X12
+	PXOR X12,X1
+	MOVOA X9,X0
+	PADDL X3,X0
+	MOVOA X0,X12
+	PSLLL $18,X0
+	PXOR X0,X2
+	PSRLL $14,X12
+	PXOR X12,X2
+	MOVOA 320(SP),X12
+	MOVOA 336(SP),X0
+	SUBQ $2,DX
+	JA MAINLOOP1
+	PADDL 112(SP),X12
+	PADDL 176(SP),X7
+	PADDL 224(SP),X10
+	PADDL 272(SP),X4
+	MOVD X12,DX
+	MOVD X7,CX
+	MOVD X10,R8
+	MOVD X4,R9
+	PSHUFL $0X39,X12,X12
+	PSHUFL $0X39,X7,X7
+	PSHUFL $0X39,X10,X10
+	PSHUFL $0X39,X4,X4
+	XORL 0(SI),DX
+	XORL 4(SI),CX
+	XORL 8(SI),R8
+	XORL 12(SI),R9
+	MOVL DX,0(DI)
+	MOVL CX,4(DI)
+	MOVL R8,8(DI)
+	MOVL R9,12(DI)
+	MOVD X12,DX
+	MOVD X7,CX
+	MOVD X10,R8
+	MOVD X4,R9
+	PSHUFL $0X39,X12,X12
+	PSHUFL $0X39,X7,X7
+	PSHUFL $0X39,X10,X10
+	PSHUFL $0X39,X4,X4
+	XORL 64(SI),DX
+	XORL 68(SI),CX
+	XORL 72(SI),R8
+	XORL 76(SI),R9
+	MOVL DX,64(DI)
+	MOVL CX,68(DI)
+	MOVL R8,72(DI)
+	MOVL R9,76(DI)
+	MOVD X12,DX
+	MOVD X7,CX
+	MOVD X10,R8
+	MOVD X4,R9
+	PSHUFL $0X39,X12,X12
+	PSHUFL $0X39,X7,X7
+	PSHUFL $0X39,X10,X10
+	PSHUFL $0X39,X4,X4
+	XORL 128(SI),DX
+	XORL 132(SI),CX
+	XORL 136(SI),R8
+	XORL 140(SI),R9
+	MOVL DX,128(DI)
+	MOVL CX,132(DI)
+	MOVL R8,136(DI)
+	MOVL R9,140(DI)
+	MOVD X12,DX
+	MOVD X7,CX
+	MOVD X10,R8
+	MOVD X4,R9
+	XORL 192(SI),DX
+	XORL 196(SI),CX
+	XORL 200(SI),R8
+	XORL 204(SI),R9
+	MOVL DX,192(DI)
+	MOVL CX,196(DI)
+	MOVL R8,200(DI)
+	MOVL R9,204(DI)
+	PADDL 240(SP),X14
+	PADDL 64(SP),X0
+	PADDL 128(SP),X5
+	PADDL 192(SP),X8
+	MOVD X14,DX
+	MOVD X0,CX
+	MOVD X5,R8
+	MOVD X8,R9
+	PSHUFL $0X39,X14,X14
+	PSHUFL $0X39,X0,X0
+	PSHUFL $0X39,X5,X5
+	PSHUFL $0X39,X8,X8
+	XORL 16(SI),DX
+	XORL 20(SI),CX
+	XORL 24(SI),R8
+	XORL 28(SI),R9
+	MOVL DX,16(DI)
+	MOVL CX,20(DI)
+	MOVL R8,24(DI)
+	MOVL R9,28(DI)
+	MOVD X14,DX
+	MOVD X0,CX
+	MOVD X5,R8
+	MOVD X8,R9
+	PSHUFL $0X39,X14,X14
+	PSHUFL $0X39,X0,X0
+	PSHUFL $0X39,X5,X5
+	PSHUFL $0X39,X8,X8
+	XORL 80(SI),DX
+	XORL 84(SI),CX
+	XORL 88(SI),R8
+	XORL 92(SI),R9
+	MOVL DX,80(DI)
+	MOVL CX,84(DI)
+	MOVL R8,88(DI)
+	MOVL R9,92(DI)
+	MOVD X14,DX
+	MOVD X0,CX
+	MOVD X5,R8
+	MOVD X8,R9
+	PSHUFL $0X39,X14,X14
+	PSHUFL $0X39,X0,X0
+	PSHUFL $0X39,X5,X5
+	PSHUFL $0X39,X8,X8
+	XORL 144(SI),DX
+	XORL 148(SI),CX
+	XORL 152(SI),R8
+	XORL 156(SI),R9
+	MOVL DX,144(DI)
+	MOVL CX,148(DI)
+	MOVL R8,152(DI)
+	MOVL R9,156(DI)
+	MOVD X14,DX
+	MOVD X0,CX
+	MOVD X5,R8
+	MOVD X8,R9
+	XORL 208(SI),DX
+	XORL 212(SI),CX
+	XORL 216(SI),R8
+	XORL 220(SI),R9
+	MOVL DX,208(DI)
+	MOVL CX,212(DI)
+	MOVL R8,216(DI)
+	MOVL R9,220(DI)
+	PADDL 288(SP),X15
+	PADDL 304(SP),X11
+	PADDL 80(SP),X1
+	PADDL 144(SP),X6
+	MOVD X15,DX
+	MOVD X11,CX
+	MOVD X1,R8
+	MOVD X6,R9
+	PSHUFL $0X39,X15,X15
+	PSHUFL $0X39,X11,X11
+	PSHUFL $0X39,X1,X1
+	PSHUFL $0X39,X6,X6
+	XORL 32(SI),DX
+	XORL 36(SI),CX
+	XORL 40(SI),R8
+	XORL 44(SI),R9
+	MOVL DX,32(DI)
+	MOVL CX,36(DI)
+	MOVL R8,40(DI)
+	MOVL R9,44(DI)
+	MOVD X15,DX
+	MOVD X11,CX
+	MOVD X1,R8
+	MOVD X6,R9
+	PSHUFL $0X39,X15,X15
+	PSHUFL $0X39,X11,X11
+	PSHUFL $0X39,X1,X1
+	PSHUFL $0X39,X6,X6
+	XORL 96(SI),DX
+	XORL 100(SI),CX
+	XORL 104(SI),R8
+	XORL 108(SI),R9
+	MOVL DX,96(DI)
+	MOVL CX,100(DI)
+	MOVL R8,104(DI)
+	MOVL R9,108(DI)
+	MOVD X15,DX
+	MOVD X11,CX
+	MOVD X1,R8
+	MOVD X6,R9
+	PSHUFL $0X39,X15,X15
+	PSHUFL $0X39,X11,X11
+	PSHUFL $0X39,X1,X1
+	PSHUFL $0X39,X6,X6
+	XORL 160(SI),DX
+	XORL 164(SI),CX
+	XORL 168(SI),R8
+	XORL 172(SI),R9
+	MOVL DX,160(DI)
+	MOVL CX,164(DI)
+	MOVL R8,168(DI)
+	MOVL R9,172(DI)
+	MOVD X15,DX
+	MOVD X11,CX
+	MOVD X1,R8
+	MOVD X6,R9
+	XORL 224(SI),DX
+	XORL 228(SI),CX
+	XORL 232(SI),R8
+	XORL 236(SI),R9
+	MOVL DX,224(DI)
+	MOVL CX,228(DI)
+	MOVL R8,232(DI)
+	MOVL R9,236(DI)
+	PADDL 160(SP),X13
+	PADDL 208(SP),X9
+	PADDL 256(SP),X3
+	PADDL 96(SP),X2
+	MOVD X13,DX
+	MOVD X9,CX
+	MOVD X3,R8
+	MOVD X2,R9
+	PSHUFL $0X39,X13,X13
+	PSHUFL $0X39,X9,X9
+	PSHUFL $0X39,X3,X3
+	PSHUFL $0X39,X2,X2
+	XORL 48(SI),DX
+	XORL 52(SI),CX
+	XORL 56(SI),R8
+	XORL 60(SI),R9
+	MOVL DX,48(DI)
+	MOVL CX,52(DI)
+	MOVL R8,56(DI)
+	MOVL R9,60(DI)
+	MOVD X13,DX
+	MOVD X9,CX
+	MOVD X3,R8
+	MOVD X2,R9
+	PSHUFL $0X39,X13,X13
+	PSHUFL $0X39,X9,X9
+	PSHUFL $0X39,X3,X3
+	PSHUFL $0X39,X2,X2
+	XORL 112(SI),DX
+	XORL 116(SI),CX
+	XORL 120(SI),R8
+	XORL 124(SI),R9
+	MOVL DX,112(DI)
+	MOVL CX,116(DI)
+	MOVL R8,120(DI)
+	MOVL R9,124(DI)
+	MOVD X13,DX
+	MOVD X9,CX
+	MOVD X3,R8
+	MOVD X2,R9
+	PSHUFL $0X39,X13,X13
+	PSHUFL $0X39,X9,X9
+	PSHUFL $0X39,X3,X3
+	PSHUFL $0X39,X2,X2
+	XORL 176(SI),DX
+	XORL 180(SI),CX
+	XORL 184(SI),R8
+	XORL 188(SI),R9
+	MOVL DX,176(DI)
+	MOVL CX,180(DI)
+	MOVL R8,184(DI)
+	MOVL R9,188(DI)
+	MOVD X13,DX
+	MOVD X9,CX
+	MOVD X3,R8
+	MOVD X2,R9
+	XORL 240(SI),DX
+	XORL 244(SI),CX
+	XORL 248(SI),R8
+	XORL 252(SI),R9
+	MOVL DX,240(DI)
+	MOVL CX,244(DI)
+	MOVL R8,248(DI)
+	MOVL R9,252(DI)
+	MOVQ 408(SP),R9
+	SUBQ $256,R9
+	ADDQ $256,SI
+	ADDQ $256,DI
+	CMPQ R9,$256
+	JAE BYTESATLEAST256
+	CMPQ R9,$0
+	JBE DONE
+	BYTESBETWEEN1AND255:
+	CMPQ R9,$64
+	JAE NOCOPY
+	MOVQ DI,DX
+	LEAQ 416(SP),DI
+	MOVQ R9,CX
+	REP; MOVSB
+	LEAQ 416(SP),DI
+	LEAQ 416(SP),SI
+	NOCOPY:
+	MOVQ R9,408(SP)
+	MOVOA 48(SP),X0
+	MOVOA 0(SP),X1
+	MOVOA 16(SP),X2
+	MOVOA 32(SP),X3
+	MOVOA X1,X4
+	MOVQ $20,CX
+	MAINLOOP2:
+	PADDL X0,X4
+	MOVOA X0,X5
+	MOVOA X4,X6
+	PSLLL $7,X4
+	PSRLL $25,X6
+	PXOR X4,X3
+	PXOR X6,X3
+	PADDL X3,X5
+	MOVOA X3,X4
+	MOVOA X5,X6
+	PSLLL $9,X5
+	PSRLL $23,X6
+	PXOR X5,X2
+	PSHUFL $0X93,X3,X3
+	PXOR X6,X2
+	PADDL X2,X4
+	MOVOA X2,X5
+	MOVOA X4,X6
+	PSLLL $13,X4
+	PSRLL $19,X6
+	PXOR X4,X1
+	PSHUFL $0X4E,X2,X2
+	PXOR X6,X1
+	PADDL X1,X5
+	MOVOA X3,X4
+	MOVOA X5,X6
+	PSLLL $18,X5
+	PSRLL $14,X6
+	PXOR X5,X0
+	PSHUFL $0X39,X1,X1
+	PXOR X6,X0
+	PADDL X0,X4
+	MOVOA X0,X5
+	MOVOA X4,X6
+	PSLLL $7,X4
+	PSRLL $25,X6
+	PXOR X4,X1
+	PXOR X6,X1
+	PADDL X1,X5
+	MOVOA X1,X4
+	MOVOA X5,X6
+	PSLLL $9,X5
+	PSRLL $23,X6
+	PXOR X5,X2
+	PSHUFL $0X93,X1,X1
+	PXOR X6,X2
+	PADDL X2,X4
+	MOVOA X2,X5
+	MOVOA X4,X6
+	PSLLL $13,X4
+	PSRLL $19,X6
+	PXOR X4,X3
+	PSHUFL $0X4E,X2,X2
+	PXOR X6,X3
+	PADDL X3,X5
+	MOVOA X1,X4
+	MOVOA X5,X6
+	PSLLL $18,X5
+	PSRLL $14,X6
+	PXOR X5,X0
+	PSHUFL $0X39,X3,X3
+	PXOR X6,X0
+	PADDL X0,X4
+	MOVOA X0,X5
+	MOVOA X4,X6
+	PSLLL $7,X4
+	PSRLL $25,X6
+	PXOR X4,X3
+	PXOR X6,X3
+	PADDL X3,X5
+	MOVOA X3,X4
+	MOVOA X5,X6
+	PSLLL $9,X5
+	PSRLL $23,X6
+	PXOR X5,X2
+	PSHUFL $0X93,X3,X3
+	PXOR X6,X2
+	PADDL X2,X4
+	MOVOA X2,X5
+	MOVOA X4,X6
+	PSLLL $13,X4
+	PSRLL $19,X6
+	PXOR X4,X1
+	PSHUFL $0X4E,X2,X2
+	PXOR X6,X1
+	PADDL X1,X5
+	MOVOA X3,X4
+	MOVOA X5,X6
+	PSLLL $18,X5
+	PSRLL $14,X6
+	PXOR X5,X0
+	PSHUFL $0X39,X1,X1
+	PXOR X6,X0
+	PADDL X0,X4
+	MOVOA X0,X5
+	MOVOA X4,X6
+	PSLLL $7,X4
+	PSRLL $25,X6
+	PXOR X4,X1
+	PXOR X6,X1
+	PADDL X1,X5
+	MOVOA X1,X4
+	MOVOA X5,X6
+	PSLLL $9,X5
+	PSRLL $23,X6
+	PXOR X5,X2
+	PSHUFL $0X93,X1,X1
+	PXOR X6,X2
+	PADDL X2,X4
+	MOVOA X2,X5
+	MOVOA X4,X6
+	PSLLL $13,X4
+	PSRLL $19,X6
+	PXOR X4,X3
+	PSHUFL $0X4E,X2,X2
+	PXOR X6,X3
+	SUBQ $4,CX
+	PADDL X3,X5
+	MOVOA X1,X4
+	MOVOA X5,X6
+	PSLLL $18,X5
+	PXOR X7,X7
+	PSRLL $14,X6
+	PXOR X5,X0
+	PSHUFL $0X39,X3,X3
+	PXOR X6,X0
+	JA MAINLOOP2
+	PADDL 48(SP),X0
+	PADDL 0(SP),X1
+	PADDL 16(SP),X2
+	PADDL 32(SP),X3
+	MOVD X0,CX
+	MOVD X1,R8
+	MOVD X2,R9
+	MOVD X3,AX
+	PSHUFL $0X39,X0,X0
+	PSHUFL $0X39,X1,X1
+	PSHUFL $0X39,X2,X2
+	PSHUFL $0X39,X3,X3
+	XORL 0(SI),CX
+	XORL 48(SI),R8
+	XORL 32(SI),R9
+	XORL 16(SI),AX
+	MOVL CX,0(DI)
+	MOVL R8,48(DI)
+	MOVL R9,32(DI)
+	MOVL AX,16(DI)
+	MOVD X0,CX
+	MOVD X1,R8
+	MOVD X2,R9
+	MOVD X3,AX
+	PSHUFL $0X39,X0,X0
+	PSHUFL $0X39,X1,X1
+	PSHUFL $0X39,X2,X2
+	PSHUFL $0X39,X3,X3
+	XORL 20(SI),CX
+	XORL 4(SI),R8
+	XORL 52(SI),R9
+	XORL 36(SI),AX
+	MOVL CX,20(DI)
+	MOVL R8,4(DI)
+	MOVL R9,52(DI)
+	MOVL AX,36(DI)
+	MOVD X0,CX
+	MOVD X1,R8
+	MOVD X2,R9
+	MOVD X3,AX
+	PSHUFL $0X39,X0,X0
+	PSHUFL $0X39,X1,X1
+	PSHUFL $0X39,X2,X2
+	PSHUFL $0X39,X3,X3
+	XORL 40(SI),CX
+	XORL 24(SI),R8
+	XORL 8(SI),R9
+	XORL 56(SI),AX
+	MOVL CX,40(DI)
+	MOVL R8,24(DI)
+	MOVL R9,8(DI)
+	MOVL AX,56(DI)
+	MOVD X0,CX
+	MOVD X1,R8
+	MOVD X2,R9
+	MOVD X3,AX
+	XORL 60(SI),CX
+	XORL 44(SI),R8
+	XORL 28(SI),R9
+	XORL 12(SI),AX
+	MOVL CX,60(DI)
+	MOVL R8,44(DI)
+	MOVL R9,28(DI)
+	MOVL AX,12(DI)
+	MOVQ 408(SP),R9
+	MOVL 16(SP),CX
+	MOVL  36 (SP),R8
+	ADDQ $1,CX
+	SHLQ $32,R8
+	ADDQ R8,CX
+	MOVQ CX,R8
+	SHRQ $32,R8
+	MOVL CX,16(SP)
+	MOVL R8, 36 (SP)
+	CMPQ R9,$64
+	JA BYTESATLEAST65
+	JAE BYTESATLEAST64
+	MOVQ DI,SI
+	MOVQ DX,DI
+	MOVQ R9,CX
+	REP; MOVSB
+	BYTESATLEAST64:
+	DONE:
+	MOVQ 352(SP),R11
+	MOVQ 360(SP),R12
+	MOVQ 368(SP),R13
+	MOVQ 376(SP),R14
+	MOVQ 384(SP),R15
+	MOVQ 392(SP),BX
+	MOVQ 400(SP),BP
+	MOVQ R11,SP
+	RET
+	BYTESATLEAST65:
+	SUBQ $64,R9
+	ADDQ $64,DI
+	ADDQ $64,SI
+	JMP BYTESBETWEEN1AND255

Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa/salsa208.go

@@ -0,0 +1,199 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package salsa
+
+// Core208 applies the Salsa20/8 core function to the 64-byte array in and puts
+// the result into the 64-byte array out. The input and output may be the same array.
+func Core208(out *[64]byte, in *[64]byte) {
+	j0 := uint32(in[0]) | uint32(in[1])<<8 | uint32(in[2])<<16 | uint32(in[3])<<24
+	j1 := uint32(in[4]) | uint32(in[5])<<8 | uint32(in[6])<<16 | uint32(in[7])<<24
+	j2 := uint32(in[8]) | uint32(in[9])<<8 | uint32(in[10])<<16 | uint32(in[11])<<24
+	j3 := uint32(in[12]) | uint32(in[13])<<8 | uint32(in[14])<<16 | uint32(in[15])<<24
+	j4 := uint32(in[16]) | uint32(in[17])<<8 | uint32(in[18])<<16 | uint32(in[19])<<24
+	j5 := uint32(in[20]) | uint32(in[21])<<8 | uint32(in[22])<<16 | uint32(in[23])<<24
+	j6 := uint32(in[24]) | uint32(in[25])<<8 | uint32(in[26])<<16 | uint32(in[27])<<24
+	j7 := uint32(in[28]) | uint32(in[29])<<8 | uint32(in[30])<<16 | uint32(in[31])<<24
+	j8 := uint32(in[32]) | uint32(in[33])<<8 | uint32(in[34])<<16 | uint32(in[35])<<24
+	j9 := uint32(in[36]) | uint32(in[37])<<8 | uint32(in[38])<<16 | uint32(in[39])<<24
+	j10 := uint32(in[40]) | uint32(in[41])<<8 | uint32(in[42])<<16 | uint32(in[43])<<24
+	j11 := uint32(in[44]) | uint32(in[45])<<8 | uint32(in[46])<<16 | uint32(in[47])<<24
+	j12 := uint32(in[48]) | uint32(in[49])<<8 | uint32(in[50])<<16 | uint32(in[51])<<24
+	j13 := uint32(in[52]) | uint32(in[53])<<8 | uint32(in[54])<<16 | uint32(in[55])<<24
+	j14 := uint32(in[56]) | uint32(in[57])<<8 | uint32(in[58])<<16 | uint32(in[59])<<24
+	j15 := uint32(in[60]) | uint32(in[61])<<8 | uint32(in[62])<<16 | uint32(in[63])<<24
+
+	x0, x1, x2, x3, x4, x5, x6, x7, x8 := j0, j1, j2, j3, j4, j5, j6, j7, j8
+	x9, x10, x11, x12, x13, x14, x15 := j9, j10, j11, j12, j13, j14, j15
+
+	for i := 0; i < 8; i += 2 {
+		u := x0 + x12
+		x4 ^= u<<7 | u>>(32-7)
+		u = x4 + x0
+		x8 ^= u<<9 | u>>(32-9)
+		u = x8 + x4
+		x12 ^= u<<13 | u>>(32-13)
+		u = x12 + x8
+		x0 ^= u<<18 | u>>(32-18)
+
+		u = x5 + x1
+		x9 ^= u<<7 | u>>(32-7)
+		u = x9 + x5
+		x13 ^= u<<9 | u>>(32-9)
+		u = x13 + x9
+		x1 ^= u<<13 | u>>(32-13)
+		u = x1 + x13
+		x5 ^= u<<18 | u>>(32-18)
+
+		u = x10 + x6
+		x14 ^= u<<7 | u>>(32-7)
+		u = x14 + x10
+		x2 ^= u<<9 | u>>(32-9)
+		u = x2 + x14
+		x6 ^= u<<13 | u>>(32-13)
+		u = x6 + x2
+		x10 ^= u<<18 | u>>(32-18)
+
+		u = x15 + x11
+		x3 ^= u<<7 | u>>(32-7)
+		u = x3 + x15
+		x7 ^= u<<9 | u>>(32-9)
+		u = x7 + x3
+		x11 ^= u<<13 | u>>(32-13)
+		u = x11 + x7
+		x15 ^= u<<18 | u>>(32-18)
+
+		u = x0 + x3
+		x1 ^= u<<7 | u>>(32-7)
+		u = x1 + x0
+		x2 ^= u<<9 | u>>(32-9)
+		u = x2 + x1
+		x3 ^= u<<13 | u>>(32-13)
+		u = x3 + x2
+		x0 ^= u<<18 | u>>(32-18)
+
+		u = x5 + x4
+		x6 ^= u<<7 | u>>(32-7)
+		u = x6 + x5
+		x7 ^= u<<9 | u>>(32-9)
+		u = x7 + x6
+		x4 ^= u<<13 | u>>(32-13)
+		u = x4 + x7
+		x5 ^= u<<18 | u>>(32-18)
+
+		u = x10 + x9
+		x11 ^= u<<7 | u>>(32-7)
+		u = x11 + x10
+		x8 ^= u<<9 | u>>(32-9)
+		u = x8 + x11
+		x9 ^= u<<13 | u>>(32-13)
+		u = x9 + x8
+		x10 ^= u<<18 | u>>(32-18)
+
+		u = x15 + x14
+		x12 ^= u<<7 | u>>(32-7)
+		u = x12 + x15
+		x13 ^= u<<9 | u>>(32-9)
+		u = x13 + x12
+		x14 ^= u<<13 | u>>(32-13)
+		u = x14 + x13
+		x15 ^= u<<18 | u>>(32-18)
+	}
+	x0 += j0
+	x1 += j1
+	x2 += j2
+	x3 += j3
+	x4 += j4
+	x5 += j5
+	x6 += j6
+	x7 += j7
+	x8 += j8
+	x9 += j9
+	x10 += j10
+	x11 += j11
+	x12 += j12
+	x13 += j13
+	x14 += j14
+	x15 += j15
+
+	out[0] = byte(x0)
+	out[1] = byte(x0 >> 8)
+	out[2] = byte(x0 >> 16)
+	out[3] = byte(x0 >> 24)
+
+	out[4] = byte(x1)
+	out[5] = byte(x1 >> 8)
+	out[6] = byte(x1 >> 16)
+	out[7] = byte(x1 >> 24)
+
+	out[8] = byte(x2)
+	out[9] = byte(x2 >> 8)
+	out[10] = byte(x2 >> 16)
+	out[11] = byte(x2 >> 24)
+
+	out[12] = byte(x3)
+	out[13] = byte(x3 >> 8)
+	out[14] = byte(x3 >> 16)
+	out[15] = byte(x3 >> 24)
+
+	out[16] = byte(x4)
+	out[17] = byte(x4 >> 8)
+	out[18] = byte(x4 >> 16)
+	out[19] = byte(x4 >> 24)
+
+	out[20] = byte(x5)
+	out[21] = byte(x5 >> 8)
+	out[22] = byte(x5 >> 16)
+	out[23] = byte(x5 >> 24)
+
+	out[24] = byte(x6)
+	out[25] = byte(x6 >> 8)
+	out[26] = byte(x6 >> 16)
+	out[27] = byte(x6 >> 24)
+
+	out[28] = byte(x7)
+	out[29] = byte(x7 >> 8)
+	out[30] = byte(x7 >> 16)
+	out[31] = byte(x7 >> 24)
+
+	out[32] = byte(x8)
+	out[33] = byte(x8 >> 8)
+	out[34] = byte(x8 >> 16)
+	out[35] = byte(x8 >> 24)
+
+	out[36] = byte(x9)
+	out[37] = byte(x9 >> 8)
+	out[38] = byte(x9 >> 16)
+	out[39] = byte(x9 >> 24)
+
+	out[40] = byte(x10)
+	out[41] = byte(x10 >> 8)
+	out[42] = byte(x10 >> 16)
+	out[43] = byte(x10 >> 24)
+
+	out[44] = byte(x11)
+	out[45] = byte(x11 >> 8)
+	out[46] = byte(x11 >> 16)
+	out[47] = byte(x11 >> 24)
+
+	out[48] = byte(x12)
+	out[49] = byte(x12 >> 8)
+	out[50] = byte(x12 >> 16)
+	out[51] = byte(x12 >> 24)
+
+	out[52] = byte(x13)
+	out[53] = byte(x13 >> 8)
+	out[54] = byte(x13 >> 16)
+	out[55] = byte(x13 >> 24)
+
+	out[56] = byte(x14)
+	out[57] = byte(x14 >> 8)
+	out[58] = byte(x14 >> 16)
+	out[59] = byte(x14 >> 24)
+
+	out[60] = byte(x15)
+	out[61] = byte(x15 >> 8)
+	out[62] = byte(x15 >> 16)
+	out[63] = byte(x15 >> 24)
+}

Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa/salsa20_amd64.go

@@ -0,0 +1,23 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build amd64,!appengine,!gccgo
+
+package salsa
+
+// This function is implemented in salsa2020_amd64.s.
+
+//go:noescape
+
+func salsa2020XORKeyStream(out, in *byte, n uint64, nonce, key *byte)
+
+// XORKeyStream crypts bytes from in to out using the given key and counters.
+// In and out may be the same slice but otherwise should not overlap. Counter
+// contains the raw salsa20 counter bytes (both nonce and block counter).
+func XORKeyStream(out, in []byte, counter *[16]byte, key *[32]byte) {
+	if len(in) == 0 {
+		return
+	}
+	salsa2020XORKeyStream(&out[0], &in[0], uint64(len(in)), &counter[0], &key[0])
+}

Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa/salsa20_ref.go

@@ -0,0 +1,234 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// +build !amd64 appengine gccgo
+
+package salsa
+
+const rounds = 20
+
+// core applies the Salsa20 core function to 16-byte input in, 32-byte key k,
+// and 16-byte constant c, and puts the result into 64-byte array out.
+func core(out *[64]byte, in *[16]byte, k *[32]byte, c *[16]byte) {
+	j0 := uint32(c[0]) | uint32(c[1])<<8 | uint32(c[2])<<16 | uint32(c[3])<<24
+	j1 := uint32(k[0]) | uint32(k[1])<<8 | uint32(k[2])<<16 | uint32(k[3])<<24
+	j2 := uint32(k[4]) | uint32(k[5])<<8 | uint32(k[6])<<16 | uint32(k[7])<<24
+	j3 := uint32(k[8]) | uint32(k[9])<<8 | uint32(k[10])<<16 | uint32(k[11])<<24
+	j4 := uint32(k[12]) | uint32(k[13])<<8 | uint32(k[14])<<16 | uint32(k[15])<<24
+	j5 := uint32(c[4]) | uint32(c[5])<<8 | uint32(c[6])<<16 | uint32(c[7])<<24
+	j6 := uint32(in[0]) | uint32(in[1])<<8 | uint32(in[2])<<16 | uint32(in[3])<<24
+	j7 := uint32(in[4]) | uint32(in[5])<<8 | uint32(in[6])<<16 | uint32(in[7])<<24
+	j8 := uint32(in[8]) | uint32(in[9])<<8 | uint32(in[10])<<16 | uint32(in[11])<<24
+	j9 := uint32(in[12]) | uint32(in[13])<<8 | uint32(in[14])<<16 | uint32(in[15])<<24
+	j10 := uint32(c[8]) | uint32(c[9])<<8 | uint32(c[10])<<16 | uint32(c[11])<<24
+	j11 := uint32(k[16]) | uint32(k[17])<<8 | uint32(k[18])<<16 | uint32(k[19])<<24
+	j12 := uint32(k[20]) | uint32(k[21])<<8 | uint32(k[22])<<16 | uint32(k[23])<<24
+	j13 := uint32(k[24]) | uint32(k[25])<<8 | uint32(k[26])<<16 | uint32(k[27])<<24
+	j14 := uint32(k[28]) | uint32(k[29])<<8 | uint32(k[30])<<16 | uint32(k[31])<<24
+	j15 := uint32(c[12]) | uint32(c[13])<<8 | uint32(c[14])<<16 | uint32(c[15])<<24
+
+	x0, x1, x2, x3, x4, x5, x6, x7, x8 := j0, j1, j2, j3, j4, j5, j6, j7, j8
+	x9, x10, x11, x12, x13, x14, x15 := j9, j10, j11, j12, j13, j14, j15
+
+	for i := 0; i < rounds; i += 2 {
+		u := x0 + x12
+		x4 ^= u<<7 | u>>(32-7)
+		u = x4 + x0
+		x8 ^= u<<9 | u>>(32-9)
+		u = x8 + x4
+		x12 ^= u<<13 | u>>(32-13)
+		u = x12 + x8
+		x0 ^= u<<18 | u>>(32-18)
+
+		u = x5 + x1
+		x9 ^= u<<7 | u>>(32-7)
+		u = x9 + x5
+		x13 ^= u<<9 | u>>(32-9)
+		u = x13 + x9
+		x1 ^= u<<13 | u>>(32-13)
+		u = x1 + x13
+		x5 ^= u<<18 | u>>(32-18)
+
+		u = x10 + x6
+		x14 ^= u<<7 | u>>(32-7)
+		u = x14 + x10
+		x2 ^= u<<9 | u>>(32-9)
+		u = x2 + x14
+		x6 ^= u<<13 | u>>(32-13)
+		u = x6 + x2
+		x10 ^= u<<18 | u>>(32-18)
+
+		u = x15 + x11
+		x3 ^= u<<7 | u>>(32-7)
+		u = x3 + x15
+		x7 ^= u<<9 | u>>(32-9)
+		u = x7 + x3
+		x11 ^= u<<13 | u>>(32-13)
+		u = x11 + x7
+		x15 ^= u<<18 | u>>(32-18)
+
+		u = x0 + x3
+		x1 ^= u<<7 | u>>(32-7)
+		u = x1 + x0
+		x2 ^= u<<9 | u>>(32-9)
+		u = x2 + x1
+		x3 ^= u<<13 | u>>(32-13)
+		u = x3 + x2
+		x0 ^= u<<18 | u>>(32-18)
+
+		u = x5 + x4
+		x6 ^= u<<7 | u>>(32-7)
+		u = x6 + x5
+		x7 ^= u<<9 | u>>(32-9)
+		u = x7 + x6
+		x4 ^= u<<13 | u>>(32-13)
+		u = x4 + x7
+		x5 ^= u<<18 | u>>(32-18)
+
+		u = x10 + x9
+		x11 ^= u<<7 | u>>(32-7)
+		u = x11 + x10
+		x8 ^= u<<9 | u>>(32-9)
+		u = x8 + x11
+		x9 ^= u<<13 | u>>(32-13)
+		u = x9 + x8
+		x10 ^= u<<18 | u>>(32-18)
+
+		u = x15 + x14
+		x12 ^= u<<7 | u>>(32-7)
+		u = x12 + x15
+		x13 ^= u<<9 | u>>(32-9)
+		u = x13 + x12
+		x14 ^= u<<13 | u>>(32-13)
+		u = x14 + x13
+		x15 ^= u<<18 | u>>(32-18)
+	}
+	x0 += j0
+	x1 += j1
+	x2 += j2
+	x3 += j3
+	x4 += j4
+	x5 += j5
+	x6 += j6
+	x7 += j7
+	x8 += j8
+	x9 += j9
+	x10 += j10
+	x11 += j11
+	x12 += j12
+	x13 += j13
+	x14 += j14
+	x15 += j15
+
+	out[0] = byte(x0)
+	out[1] = byte(x0 >> 8)
+	out[2] = byte(x0 >> 16)
+	out[3] = byte(x0 >> 24)
+
+	out[4] = byte(x1)
+	out[5] = byte(x1 >> 8)
+	out[6] = byte(x1 >> 16)
+	out[7] = byte(x1 >> 24)
+
+	out[8] = byte(x2)
+	out[9] = byte(x2 >> 8)
+	out[10] = byte(x2 >> 16)
+	out[11] = byte(x2 >> 24)
+
+	out[12] = byte(x3)
+	out[13] = byte(x3 >> 8)
+	out[14] = byte(x3 >> 16)
+	out[15] = byte(x3 >> 24)
+
+	out[16] = byte(x4)
+	out[17] = byte(x4 >> 8)
+	out[18] = byte(x4 >> 16)
+	out[19] = byte(x4 >> 24)
+
+	out[20] = byte(x5)
+	out[21] = byte(x5 >> 8)
+	out[22] = byte(x5 >> 16)
+	out[23] = byte(x5 >> 24)
+
+	out[24] = byte(x6)
+	out[25] = byte(x6 >> 8)
+	out[26] = byte(x6 >> 16)
+	out[27] = byte(x6 >> 24)
+
+	out[28] = byte(x7)
+	out[29] = byte(x7 >> 8)
+	out[30] = byte(x7 >> 16)
+	out[31] = byte(x7 >> 24)
+
+	out[32] = byte(x8)
+	out[33] = byte(x8 >> 8)
+	out[34] = byte(x8 >> 16)
+	out[35] = byte(x8 >> 24)
+
+	out[36] = byte(x9)
+	out[37] = byte(x9 >> 8)
+	out[38] = byte(x9 >> 16)
+	out[39] = byte(x9 >> 24)
+
+	out[40] = byte(x10)
+	out[41] = byte(x10 >> 8)
+	out[42] = byte(x10 >> 16)
+	out[43] = byte(x10 >> 24)
+
+	out[44] = byte(x11)
+	out[45] = byte(x11 >> 8)
+	out[46] = byte(x11 >> 16)
+	out[47] = byte(x11 >> 24)
+
+	out[48] = byte(x12)
+	out[49] = byte(x12 >> 8)
+	out[50] = byte(x12 >> 16)
+	out[51] = byte(x12 >> 24)
+
+	out[52] = byte(x13)
+	out[53] = byte(x13 >> 8)
+	out[54] = byte(x13 >> 16)
+	out[55] = byte(x13 >> 24)
+
+	out[56] = byte(x14)
+	out[57] = byte(x14 >> 8)
+	out[58] = byte(x14 >> 16)
+	out[59] = byte(x14 >> 24)
+
+	out[60] = byte(x15)
+	out[61] = byte(x15 >> 8)
+	out[62] = byte(x15 >> 16)
+	out[63] = byte(x15 >> 24)
+}
+
+// XORKeyStream crypts bytes from in to out using the given key and counters.
+// In and out may be the same slice but otherwise should not overlap. Counter
+// contains the raw salsa20 counter bytes (both nonce and block counter).
+func XORKeyStream(out, in []byte, counter *[16]byte, key *[32]byte) {
+	var block [64]byte
+	var counterCopy [16]byte
+	copy(counterCopy[:], counter[:])
+
+	for len(in) >= 64 {
+		core(&block, &counterCopy, key, &Sigma)
+		for i, x := range block {
+			out[i] = in[i] ^ x
+		}
+		u := uint32(1)
+		for i := 8; i < 16; i++ {
+			u += uint32(counterCopy[i])
+			counterCopy[i] = byte(u)
+			u >>= 8
+		}
+		in = in[64:]
+		out = out[64:]
+	}
+
+	if len(in) > 0 {
+		core(&block, &counterCopy, key, &Sigma)
+		for i, v := range in {
+			out[i] = v ^ block[i]
+		}
+	}
+}

Godeps/_workspace/src/golang.org/x/crypto/salsa20/salsa/salsa_test.go

@@ -0,0 +1,35 @@
+// Copyright 2012 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package salsa
+
+import "testing"
+
+func TestCore208(t *testing.T) {
+	in := [64]byte{
+		0x7e, 0x87, 0x9a, 0x21, 0x4f, 0x3e, 0xc9, 0x86,
+		0x7c, 0xa9, 0x40, 0xe6, 0x41, 0x71, 0x8f, 0x26,
+		0xba, 0xee, 0x55, 0x5b, 0x8c, 0x61, 0xc1, 0xb5,
+		0x0d, 0xf8, 0x46, 0x11, 0x6d, 0xcd, 0x3b, 0x1d,
+		0xee, 0x24, 0xf3, 0x19, 0xdf, 0x9b, 0x3d, 0x85,
+		0x14, 0x12, 0x1e, 0x4b, 0x5a, 0xc5, 0xaa, 0x32,
+		0x76, 0x02, 0x1d, 0x29, 0x09, 0xc7, 0x48, 0x29,
+		0xed, 0xeb, 0xc6, 0x8d, 0xb8, 0xb8, 0xc2, 0x5e}
+
+	out := [64]byte{
+		0xa4, 0x1f, 0x85, 0x9c, 0x66, 0x08, 0xcc, 0x99,
+		0x3b, 0x81, 0xca, 0xcb, 0x02, 0x0c, 0xef, 0x05,
+		0x04, 0x4b, 0x21, 0x81, 0xa2, 0xfd, 0x33, 0x7d,
+		0xfd, 0x7b, 0x1c, 0x63, 0x96, 0x68, 0x2f, 0x29,
+		0xb4, 0x39, 0x31, 0x68, 0xe3, 0xc9, 0xe6, 0xbc,
+		0xfe, 0x6b, 0xc5, 0xb7, 0xa0, 0x6d, 0x96, 0xba,
+		0xe4, 0x24, 0xcc, 0x10, 0x2c, 0x91, 0x74, 0x5c,
+		0x24, 0xad, 0x67, 0x3d, 0xc7, 0x61, 0x8f, 0x81,
+	}
+
+	Core208(&in, &in)
+	if in != out {
+		t.Errorf("expected %x, got %x", out, in)
+	}
+}

p2p/secret_connection.go

@@ -14,9 +14,9 @@ import (
 	"sync"
 	"time"
 
-	"golang.org/x/crypto/nacl/box"
-	"golang.org/x/crypto/nacl/secretbox"
-	"golang.org/x/crypto/ripemd160"
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/nacl/box"
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/nacl/secretbox"
+	"github.com/tendermint/tendermint/Godeps/_workspace/src/golang.org/x/crypto/ripemd160"
 
 	acm "github.com/tendermint/tendermint/account"
 	bm "github.com/tendermint/tendermint/binary"