crypto/internal/mlkem768: new package

This was initially developed at github.com/FiloSottile/mlkem768.

+ 5ce9162 - mlkem768,xwing: add SeedSize <Filippo Valsorda>
+ b43add9 - mlkem768,xwing: add NewKeyFromSeed <Filippo Valsorda>
+ e000fa4 - mlkem768: improve RoundTrip benchmark <Filippo Valsorda>
+ 344d5ee - mlkem768: add exhaustive tests for compress and decompress (#4) <David Buchanan>
+ 08fb36c - mlkem768: do not panic <Filippo Valsorda>
+ 9e9fcc2 - mlkem768: add proposed Wycheproof test vectors <Filippo Valsorda>
+ 5e630b8 - mlkem768: add more tests <Filippo Valsorda>
+ e3fb5df - mlkem768: add TestUnluckyVector <Filippo Valsorda>
+ 3f410e9 - mlkem768: add accumulated pq-crystals vectors <Filippo Valsorda>
+ 9897e2f - mlkem768: add other known test vectors <Filippo Valsorda>
+ cffbfb9 - mlkem768: update sampleNTT comment <Filippo Valsorda>
+ df1b265 - mlkem768: use uint16 reads, simpler bit twiddling <Josh Bleecher Snyder>
+ 50a7fad - mlkem768: unroll ntt inner loop <Josh Bleecher Snyder>
+ cd8140e - mlkem768: avoid extra data copies <Josh Bleecher Snyder>
+ 0c68443 - mlkem768: buffer reads from sha3 <Josh Bleecher Snyder>
+ bb784ff - mlkem768: create README.md <Filippo Valsorda>
+ 35e7ada - mlkem768: add package docs and LICENSE <Filippo Valsorda>
+ 2e6a3df - mlkem768: drop performance optimization notes <Filippo Valsorda>
+ d5449de - mlkem768: add benchmarks <Filippo Valsorda>
+ 3294fee - mlkem768: implement ML-KEM <Filippo Valsorda>
+ 4cb306e - mlkem768: reimplement compress and decompress <Filippo Valsorda>
+ 48e4c4c - mlkem768: fix AHat draft spec typo <Filippo Valsorda>
+ c34ddcf - mlkem768: make better use of constants <Filippo Valsorda>
+ 3b485e1 - mlkem768: initial commit, a full K-PKE implementation <Filippo Valsorda>

Submitting changes on behalf of Josh Bleecher Snyder as authorized at
https://go-review.googlesource.com/c/go/+/547357/comment/61f8433f_04dc9c5d/
and of David Buchanan as authorized at
https://github.com/FiloSottile/mlkem768/pull/4#issuecomment-1975330952.

Updates #64537

Change-Id: I50607336282434d64a1255901b0ef40dbfd47e91
Reviewed-on: https://go-review.googlesource.com/c/go/+/550215
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>

16 files changed, 1848 insertions, 0 deletions

diff --git a/src/vendor/golang.org/x/crypto/sha3/doc.go b/src/vendor/golang.org/x/crypto/sha3/doc.go
new file mode 100644
index 0000000000..decd8cf9bf
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/doc.go
@@ -0,0 +1,62 @@
+// Copyright 2014 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Package sha3 implements the SHA-3 fixed-output-length hash functions and
+// the SHAKE variable-output-length hash functions defined by FIPS-202.
+//
+// Both types of hash function use the "sponge" construction and the Keccak
+// permutation. For a detailed specification see http://keccak.noekeon.org/
+//
+// # Guidance
+//
+// If you aren't sure what function you need, use SHAKE256 with at least 64
+// bytes of output. The SHAKE instances are faster than the SHA3 instances;
+// the latter have to allocate memory to conform to the hash.Hash interface.
+//
+// If you need a secret-key MAC (message authentication code), prepend the
+// secret key to the input, hash with SHAKE256 and read at least 32 bytes of
+// output.
+//
+// # Security strengths
+//
+// The SHA3-x (x equals 224, 256, 384, or 512) functions have a security
+// strength against preimage attacks of x bits. Since they only produce "x"
+// bits of output, their collision-resistance is only "x/2" bits.
+//
+// The SHAKE-256 and -128 functions have a generic security strength of 256 and
+// 128 bits against all attacks, provided that at least 2x bits of their output
+// is used.  Requesting more than 64 or 32 bytes of output, respectively, does
+// not increase the collision-resistance of the SHAKE functions.
+//
+// # The sponge construction
+//
+// A sponge builds a pseudo-random function from a public pseudo-random
+// permutation, by applying the permutation to a state of "rate + capacity"
+// bytes, but hiding "capacity" of the bytes.
+//
+// A sponge starts out with a zero state. To hash an input using a sponge, up
+// to "rate" bytes of the input are XORed into the sponge's state. The sponge
+// is then "full" and the permutation is applied to "empty" it. This process is
+// repeated until all the input has been "absorbed". The input is then padded.
+// The digest is "squeezed" from the sponge in the same way, except that output
+// is copied out instead of input being XORed in.
+//
+// A sponge is parameterized by its generic security strength, which is equal
+// to half its capacity; capacity + rate is equal to the permutation's width.
+// Since the KeccakF-1600 permutation is 1600 bits (200 bytes) wide, this means
+// that the security strength of a sponge instance is equal to (1600 - bitrate) / 2.
+//
+// # Recommendations
+//
+// The SHAKE functions are recommended for most new uses. They can produce
+// output of arbitrary length. SHAKE256, with an output length of at least
+// 64 bytes, provides 256-bit security against all attacks.  The Keccak team
+// recommends it for most applications upgrading from SHA2-512. (NIST chose a
+// much stronger, but much slower, sponge instance for SHA3-512.)
+//
+// The SHA-3 functions are "drop-in" replacements for the SHA-2 functions.
+// They produce output of the same length, with the same security strengths
+// against all attacks. This means, in particular, that SHA3-256 only has
+// 128-bit collision resistance, because its output length is 32 bytes.
+package sha3 // import "golang.org/x/crypto/sha3"
diff --git a/src/vendor/golang.org/x/crypto/sha3/hashes.go b/src/vendor/golang.org/x/crypto/sha3/hashes.go
new file mode 100644
index 0000000000..0d8043fd2a
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/hashes.go
@@ -0,0 +1,97 @@
+// Copyright 2014 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package sha3
+
+// This file provides functions for creating instances of the SHA-3
+// and SHAKE hash functions, as well as utility functions for hashing
+// bytes.
+
+import (
+	"hash"
+)
+
+// New224 creates a new SHA3-224 hash.
+// Its generic security strength is 224 bits against preimage attacks,
+// and 112 bits against collision attacks.
+func New224() hash.Hash {
+	if h := new224Asm(); h != nil {
+		return h
+	}
+	return &state{rate: 144, outputLen: 28, dsbyte: 0x06}
+}
+
+// New256 creates a new SHA3-256 hash.
+// Its generic security strength is 256 bits against preimage attacks,
+// and 128 bits against collision attacks.
+func New256() hash.Hash {
+	if h := new256Asm(); h != nil {
+		return h
+	}
+	return &state{rate: 136, outputLen: 32, dsbyte: 0x06}
+}
+
+// New384 creates a new SHA3-384 hash.
+// Its generic security strength is 384 bits against preimage attacks,
+// and 192 bits against collision attacks.
+func New384() hash.Hash {
+	if h := new384Asm(); h != nil {
+		return h
+	}
+	return &state{rate: 104, outputLen: 48, dsbyte: 0x06}
+}
+
+// New512 creates a new SHA3-512 hash.
+// Its generic security strength is 512 bits against preimage attacks,
+// and 256 bits against collision attacks.
+func New512() hash.Hash {
+	if h := new512Asm(); h != nil {
+		return h
+	}
+	return &state{rate: 72, outputLen: 64, dsbyte: 0x06}
+}
+
+// NewLegacyKeccak256 creates a new Keccak-256 hash.
+//
+// Only use this function if you require compatibility with an existing cryptosystem
+// that uses non-standard padding. All other users should use New256 instead.
+func NewLegacyKeccak256() hash.Hash { return &state{rate: 136, outputLen: 32, dsbyte: 0x01} }
+
+// NewLegacyKeccak512 creates a new Keccak-512 hash.
+//
+// Only use this function if you require compatibility with an existing cryptosystem
+// that uses non-standard padding. All other users should use New512 instead.
+func NewLegacyKeccak512() hash.Hash { return &state{rate: 72, outputLen: 64, dsbyte: 0x01} }
+
+// Sum224 returns the SHA3-224 digest of the data.
+func Sum224(data []byte) (digest [28]byte) {
+	h := New224()
+	h.Write(data)
+	h.Sum(digest[:0])
+	return
+}
+
+// Sum256 returns the SHA3-256 digest of the data.
+func Sum256(data []byte) (digest [32]byte) {
+	h := New256()
+	h.Write(data)
+	h.Sum(digest[:0])
+	return
+}
+
+// Sum384 returns the SHA3-384 digest of the data.
+func Sum384(data []byte) (digest [48]byte) {
+	h := New384()
+	h.Write(data)
+	h.Sum(digest[:0])
+	return
+}
+
+// Sum512 returns the SHA3-512 digest of the data.
+func Sum512(data []byte) (digest [64]byte) {
+	h := New512()
+	h.Write(data)
+	h.Sum(digest[:0])
+	return
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/hashes_generic.go b/src/vendor/golang.org/x/crypto/sha3/hashes_generic.go
new file mode 100644
index 0000000000..fe8c84793c
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/hashes_generic.go
@@ -0,0 +1,27 @@
+// Copyright 2017 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build !gc || purego || !s390x
+
+package sha3
+
+import (
+	"hash"
+)
+
+// new224Asm returns an assembly implementation of SHA3-224 if available,
+// otherwise it returns nil.
+func new224Asm() hash.Hash { return nil }
+
+// new256Asm returns an assembly implementation of SHA3-256 if available,
+// otherwise it returns nil.
+func new256Asm() hash.Hash { return nil }
+
+// new384Asm returns an assembly implementation of SHA3-384 if available,
+// otherwise it returns nil.
+func new384Asm() hash.Hash { return nil }
+
+// new512Asm returns an assembly implementation of SHA3-512 if available,
+// otherwise it returns nil.
+func new512Asm() hash.Hash { return nil }
diff --git a/src/vendor/golang.org/x/crypto/sha3/keccakf.go b/src/vendor/golang.org/x/crypto/sha3/keccakf.go
new file mode 100644
index 0000000000..ce48b1dd3e
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/keccakf.go
@@ -0,0 +1,414 @@
+// Copyright 2014 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build !amd64 || purego || !gc
+
+package sha3
+
+import "math/bits"
+
+// rc stores the round constants for use in the ι step.
+var rc = [24]uint64{
+	0x0000000000000001,
+	0x0000000000008082,
+	0x800000000000808A,
+	0x8000000080008000,
+	0x000000000000808B,
+	0x0000000080000001,
+	0x8000000080008081,
+	0x8000000000008009,
+	0x000000000000008A,
+	0x0000000000000088,
+	0x0000000080008009,
+	0x000000008000000A,
+	0x000000008000808B,
+	0x800000000000008B,
+	0x8000000000008089,
+	0x8000000000008003,
+	0x8000000000008002,
+	0x8000000000000080,
+	0x000000000000800A,
+	0x800000008000000A,
+	0x8000000080008081,
+	0x8000000000008080,
+	0x0000000080000001,
+	0x8000000080008008,
+}
+
+// keccakF1600 applies the Keccak permutation to a 1600b-wide
+// state represented as a slice of 25 uint64s.
+func keccakF1600(a *[25]uint64) {
+	// Implementation translated from Keccak-inplace.c
+	// in the keccak reference code.
+	var t, bc0, bc1, bc2, bc3, bc4, d0, d1, d2, d3, d4 uint64
+
+	for i := 0; i < 24; i += 4 {
+		// Combines the 5 steps in each round into 2 steps.
+		// Unrolls 4 rounds per loop and spreads some steps across rounds.
+
+		// Round 1
+		bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
+		bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
+		bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
+		bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
+		bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
+		d0 = bc4 ^ (bc1<<1 | bc1>>63)
+		d1 = bc0 ^ (bc2<<1 | bc2>>63)
+		d2 = bc1 ^ (bc3<<1 | bc3>>63)
+		d3 = bc2 ^ (bc4<<1 | bc4>>63)
+		d4 = bc3 ^ (bc0<<1 | bc0>>63)
+
+		bc0 = a[0] ^ d0
+		t = a[6] ^ d1
+		bc1 = bits.RotateLeft64(t, 44)
+		t = a[12] ^ d2
+		bc2 = bits.RotateLeft64(t, 43)
+		t = a[18] ^ d3
+		bc3 = bits.RotateLeft64(t, 21)
+		t = a[24] ^ d4
+		bc4 = bits.RotateLeft64(t, 14)
+		a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i]
+		a[6] = bc1 ^ (bc3 &^ bc2)
+		a[12] = bc2 ^ (bc4 &^ bc3)
+		a[18] = bc3 ^ (bc0 &^ bc4)
+		a[24] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[10] ^ d0
+		bc2 = bits.RotateLeft64(t, 3)
+		t = a[16] ^ d1
+		bc3 = bits.RotateLeft64(t, 45)
+		t = a[22] ^ d2
+		bc4 = bits.RotateLeft64(t, 61)
+		t = a[3] ^ d3
+		bc0 = bits.RotateLeft64(t, 28)
+		t = a[9] ^ d4
+		bc1 = bits.RotateLeft64(t, 20)
+		a[10] = bc0 ^ (bc2 &^ bc1)
+		a[16] = bc1 ^ (bc3 &^ bc2)
+		a[22] = bc2 ^ (bc4 &^ bc3)
+		a[3] = bc3 ^ (bc0 &^ bc4)
+		a[9] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[20] ^ d0
+		bc4 = bits.RotateLeft64(t, 18)
+		t = a[1] ^ d1
+		bc0 = bits.RotateLeft64(t, 1)
+		t = a[7] ^ d2
+		bc1 = bits.RotateLeft64(t, 6)
+		t = a[13] ^ d3
+		bc2 = bits.RotateLeft64(t, 25)
+		t = a[19] ^ d4
+		bc3 = bits.RotateLeft64(t, 8)
+		a[20] = bc0 ^ (bc2 &^ bc1)
+		a[1] = bc1 ^ (bc3 &^ bc2)
+		a[7] = bc2 ^ (bc4 &^ bc3)
+		a[13] = bc3 ^ (bc0 &^ bc4)
+		a[19] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[5] ^ d0
+		bc1 = bits.RotateLeft64(t, 36)
+		t = a[11] ^ d1
+		bc2 = bits.RotateLeft64(t, 10)
+		t = a[17] ^ d2
+		bc3 = bits.RotateLeft64(t, 15)
+		t = a[23] ^ d3
+		bc4 = bits.RotateLeft64(t, 56)
+		t = a[4] ^ d4
+		bc0 = bits.RotateLeft64(t, 27)
+		a[5] = bc0 ^ (bc2 &^ bc1)
+		a[11] = bc1 ^ (bc3 &^ bc2)
+		a[17] = bc2 ^ (bc4 &^ bc3)
+		a[23] = bc3 ^ (bc0 &^ bc4)
+		a[4] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[15] ^ d0
+		bc3 = bits.RotateLeft64(t, 41)
+		t = a[21] ^ d1
+		bc4 = bits.RotateLeft64(t, 2)
+		t = a[2] ^ d2
+		bc0 = bits.RotateLeft64(t, 62)
+		t = a[8] ^ d3
+		bc1 = bits.RotateLeft64(t, 55)
+		t = a[14] ^ d4
+		bc2 = bits.RotateLeft64(t, 39)
+		a[15] = bc0 ^ (bc2 &^ bc1)
+		a[21] = bc1 ^ (bc3 &^ bc2)
+		a[2] = bc2 ^ (bc4 &^ bc3)
+		a[8] = bc3 ^ (bc0 &^ bc4)
+		a[14] = bc4 ^ (bc1 &^ bc0)
+
+		// Round 2
+		bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
+		bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
+		bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
+		bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
+		bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
+		d0 = bc4 ^ (bc1<<1 | bc1>>63)
+		d1 = bc0 ^ (bc2<<1 | bc2>>63)
+		d2 = bc1 ^ (bc3<<1 | bc3>>63)
+		d3 = bc2 ^ (bc4<<1 | bc4>>63)
+		d4 = bc3 ^ (bc0<<1 | bc0>>63)
+
+		bc0 = a[0] ^ d0
+		t = a[16] ^ d1
+		bc1 = bits.RotateLeft64(t, 44)
+		t = a[7] ^ d2
+		bc2 = bits.RotateLeft64(t, 43)
+		t = a[23] ^ d3
+		bc3 = bits.RotateLeft64(t, 21)
+		t = a[14] ^ d4
+		bc4 = bits.RotateLeft64(t, 14)
+		a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i+1]
+		a[16] = bc1 ^ (bc3 &^ bc2)
+		a[7] = bc2 ^ (bc4 &^ bc3)
+		a[23] = bc3 ^ (bc0 &^ bc4)
+		a[14] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[20] ^ d0
+		bc2 = bits.RotateLeft64(t, 3)
+		t = a[11] ^ d1
+		bc3 = bits.RotateLeft64(t, 45)
+		t = a[2] ^ d2
+		bc4 = bits.RotateLeft64(t, 61)
+		t = a[18] ^ d3
+		bc0 = bits.RotateLeft64(t, 28)
+		t = a[9] ^ d4
+		bc1 = bits.RotateLeft64(t, 20)
+		a[20] = bc0 ^ (bc2 &^ bc1)
+		a[11] = bc1 ^ (bc3 &^ bc2)
+		a[2] = bc2 ^ (bc4 &^ bc3)
+		a[18] = bc3 ^ (bc0 &^ bc4)
+		a[9] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[15] ^ d0
+		bc4 = bits.RotateLeft64(t, 18)
+		t = a[6] ^ d1
+		bc0 = bits.RotateLeft64(t, 1)
+		t = a[22] ^ d2
+		bc1 = bits.RotateLeft64(t, 6)
+		t = a[13] ^ d3
+		bc2 = bits.RotateLeft64(t, 25)
+		t = a[4] ^ d4
+		bc3 = bits.RotateLeft64(t, 8)
+		a[15] = bc0 ^ (bc2 &^ bc1)
+		a[6] = bc1 ^ (bc3 &^ bc2)
+		a[22] = bc2 ^ (bc4 &^ bc3)
+		a[13] = bc3 ^ (bc0 &^ bc4)
+		a[4] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[10] ^ d0
+		bc1 = bits.RotateLeft64(t, 36)
+		t = a[1] ^ d1
+		bc2 = bits.RotateLeft64(t, 10)
+		t = a[17] ^ d2
+		bc3 = bits.RotateLeft64(t, 15)
+		t = a[8] ^ d3
+		bc4 = bits.RotateLeft64(t, 56)
+		t = a[24] ^ d4
+		bc0 = bits.RotateLeft64(t, 27)
+		a[10] = bc0 ^ (bc2 &^ bc1)
+		a[1] = bc1 ^ (bc3 &^ bc2)
+		a[17] = bc2 ^ (bc4 &^ bc3)
+		a[8] = bc3 ^ (bc0 &^ bc4)
+		a[24] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[5] ^ d0
+		bc3 = bits.RotateLeft64(t, 41)
+		t = a[21] ^ d1
+		bc4 = bits.RotateLeft64(t, 2)
+		t = a[12] ^ d2
+		bc0 = bits.RotateLeft64(t, 62)
+		t = a[3] ^ d3
+		bc1 = bits.RotateLeft64(t, 55)
+		t = a[19] ^ d4
+		bc2 = bits.RotateLeft64(t, 39)
+		a[5] = bc0 ^ (bc2 &^ bc1)
+		a[21] = bc1 ^ (bc3 &^ bc2)
+		a[12] = bc2 ^ (bc4 &^ bc3)
+		a[3] = bc3 ^ (bc0 &^ bc4)
+		a[19] = bc4 ^ (bc1 &^ bc0)
+
+		// Round 3
+		bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
+		bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
+		bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
+		bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
+		bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
+		d0 = bc4 ^ (bc1<<1 | bc1>>63)
+		d1 = bc0 ^ (bc2<<1 | bc2>>63)
+		d2 = bc1 ^ (bc3<<1 | bc3>>63)
+		d3 = bc2 ^ (bc4<<1 | bc4>>63)
+		d4 = bc3 ^ (bc0<<1 | bc0>>63)
+
+		bc0 = a[0] ^ d0
+		t = a[11] ^ d1
+		bc1 = bits.RotateLeft64(t, 44)
+		t = a[22] ^ d2
+		bc2 = bits.RotateLeft64(t, 43)
+		t = a[8] ^ d3
+		bc3 = bits.RotateLeft64(t, 21)
+		t = a[19] ^ d4
+		bc4 = bits.RotateLeft64(t, 14)
+		a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i+2]
+		a[11] = bc1 ^ (bc3 &^ bc2)
+		a[22] = bc2 ^ (bc4 &^ bc3)
+		a[8] = bc3 ^ (bc0 &^ bc4)
+		a[19] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[15] ^ d0
+		bc2 = bits.RotateLeft64(t, 3)
+		t = a[1] ^ d1
+		bc3 = bits.RotateLeft64(t, 45)
+		t = a[12] ^ d2
+		bc4 = bits.RotateLeft64(t, 61)
+		t = a[23] ^ d3
+		bc0 = bits.RotateLeft64(t, 28)
+		t = a[9] ^ d4
+		bc1 = bits.RotateLeft64(t, 20)
+		a[15] = bc0 ^ (bc2 &^ bc1)
+		a[1] = bc1 ^ (bc3 &^ bc2)
+		a[12] = bc2 ^ (bc4 &^ bc3)
+		a[23] = bc3 ^ (bc0 &^ bc4)
+		a[9] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[5] ^ d0
+		bc4 = bits.RotateLeft64(t, 18)
+		t = a[16] ^ d1
+		bc0 = bits.RotateLeft64(t, 1)
+		t = a[2] ^ d2
+		bc1 = bits.RotateLeft64(t, 6)
+		t = a[13] ^ d3
+		bc2 = bits.RotateLeft64(t, 25)
+		t = a[24] ^ d4
+		bc3 = bits.RotateLeft64(t, 8)
+		a[5] = bc0 ^ (bc2 &^ bc1)
+		a[16] = bc1 ^ (bc3 &^ bc2)
+		a[2] = bc2 ^ (bc4 &^ bc3)
+		a[13] = bc3 ^ (bc0 &^ bc4)
+		a[24] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[20] ^ d0
+		bc1 = bits.RotateLeft64(t, 36)
+		t = a[6] ^ d1
+		bc2 = bits.RotateLeft64(t, 10)
+		t = a[17] ^ d2
+		bc3 = bits.RotateLeft64(t, 15)
+		t = a[3] ^ d3
+		bc4 = bits.RotateLeft64(t, 56)
+		t = a[14] ^ d4
+		bc0 = bits.RotateLeft64(t, 27)
+		a[20] = bc0 ^ (bc2 &^ bc1)
+		a[6] = bc1 ^ (bc3 &^ bc2)
+		a[17] = bc2 ^ (bc4 &^ bc3)
+		a[3] = bc3 ^ (bc0 &^ bc4)
+		a[14] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[10] ^ d0
+		bc3 = bits.RotateLeft64(t, 41)
+		t = a[21] ^ d1
+		bc4 = bits.RotateLeft64(t, 2)
+		t = a[7] ^ d2
+		bc0 = bits.RotateLeft64(t, 62)
+		t = a[18] ^ d3
+		bc1 = bits.RotateLeft64(t, 55)
+		t = a[4] ^ d4
+		bc2 = bits.RotateLeft64(t, 39)
+		a[10] = bc0 ^ (bc2 &^ bc1)
+		a[21] = bc1 ^ (bc3 &^ bc2)
+		a[7] = bc2 ^ (bc4 &^ bc3)
+		a[18] = bc3 ^ (bc0 &^ bc4)
+		a[4] = bc4 ^ (bc1 &^ bc0)
+
+		// Round 4
+		bc0 = a[0] ^ a[5] ^ a[10] ^ a[15] ^ a[20]
+		bc1 = a[1] ^ a[6] ^ a[11] ^ a[16] ^ a[21]
+		bc2 = a[2] ^ a[7] ^ a[12] ^ a[17] ^ a[22]
+		bc3 = a[3] ^ a[8] ^ a[13] ^ a[18] ^ a[23]
+		bc4 = a[4] ^ a[9] ^ a[14] ^ a[19] ^ a[24]
+		d0 = bc4 ^ (bc1<<1 | bc1>>63)
+		d1 = bc0 ^ (bc2<<1 | bc2>>63)
+		d2 = bc1 ^ (bc3<<1 | bc3>>63)
+		d3 = bc2 ^ (bc4<<1 | bc4>>63)
+		d4 = bc3 ^ (bc0<<1 | bc0>>63)
+
+		bc0 = a[0] ^ d0
+		t = a[1] ^ d1
+		bc1 = bits.RotateLeft64(t, 44)
+		t = a[2] ^ d2
+		bc2 = bits.RotateLeft64(t, 43)
+		t = a[3] ^ d3
+		bc3 = bits.RotateLeft64(t, 21)
+		t = a[4] ^ d4
+		bc4 = bits.RotateLeft64(t, 14)
+		a[0] = bc0 ^ (bc2 &^ bc1) ^ rc[i+3]
+		a[1] = bc1 ^ (bc3 &^ bc2)
+		a[2] = bc2 ^ (bc4 &^ bc3)
+		a[3] = bc3 ^ (bc0 &^ bc4)
+		a[4] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[5] ^ d0
+		bc2 = bits.RotateLeft64(t, 3)
+		t = a[6] ^ d1
+		bc3 = bits.RotateLeft64(t, 45)
+		t = a[7] ^ d2
+		bc4 = bits.RotateLeft64(t, 61)
+		t = a[8] ^ d3
+		bc0 = bits.RotateLeft64(t, 28)
+		t = a[9] ^ d4
+		bc1 = bits.RotateLeft64(t, 20)
+		a[5] = bc0 ^ (bc2 &^ bc1)
+		a[6] = bc1 ^ (bc3 &^ bc2)
+		a[7] = bc2 ^ (bc4 &^ bc3)
+		a[8] = bc3 ^ (bc0 &^ bc4)
+		a[9] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[10] ^ d0
+		bc4 = bits.RotateLeft64(t, 18)
+		t = a[11] ^ d1
+		bc0 = bits.RotateLeft64(t, 1)
+		t = a[12] ^ d2
+		bc1 = bits.RotateLeft64(t, 6)
+		t = a[13] ^ d3
+		bc2 = bits.RotateLeft64(t, 25)
+		t = a[14] ^ d4
+		bc3 = bits.RotateLeft64(t, 8)
+		a[10] = bc0 ^ (bc2 &^ bc1)
+		a[11] = bc1 ^ (bc3 &^ bc2)
+		a[12] = bc2 ^ (bc4 &^ bc3)
+		a[13] = bc3 ^ (bc0 &^ bc4)
+		a[14] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[15] ^ d0
+		bc1 = bits.RotateLeft64(t, 36)
+		t = a[16] ^ d1
+		bc2 = bits.RotateLeft64(t, 10)
+		t = a[17] ^ d2
+		bc3 = bits.RotateLeft64(t, 15)
+		t = a[18] ^ d3
+		bc4 = bits.RotateLeft64(t, 56)
+		t = a[19] ^ d4
+		bc0 = bits.RotateLeft64(t, 27)
+		a[15] = bc0 ^ (bc2 &^ bc1)
+		a[16] = bc1 ^ (bc3 &^ bc2)
+		a[17] = bc2 ^ (bc4 &^ bc3)
+		a[18] = bc3 ^ (bc0 &^ bc4)
+		a[19] = bc4 ^ (bc1 &^ bc0)
+
+		t = a[20] ^ d0
+		bc3 = bits.RotateLeft64(t, 41)
+		t = a[21] ^ d1
+		bc4 = bits.RotateLeft64(t, 2)
+		t = a[22] ^ d2
+		bc0 = bits.RotateLeft64(t, 62)
+		t = a[23] ^ d3
+		bc1 = bits.RotateLeft64(t, 55)
+		t = a[24] ^ d4
+		bc2 = bits.RotateLeft64(t, 39)
+		a[20] = bc0 ^ (bc2 &^ bc1)
+		a[21] = bc1 ^ (bc3 &^ bc2)
+		a[22] = bc2 ^ (bc4 &^ bc3)
+		a[23] = bc3 ^ (bc0 &^ bc4)
+		a[24] = bc4 ^ (bc1 &^ bc0)
+	}
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/keccakf_amd64.go b/src/vendor/golang.org/x/crypto/sha3/keccakf_amd64.go
new file mode 100644
index 0000000000..b908696be5
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/keccakf_amd64.go
@@ -0,0 +1,13 @@
+// Copyright 2015 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build amd64 && !purego && gc
+
+package sha3
+
+// This function is implemented in keccakf_amd64.s.
+
+//go:noescape
+
+func keccakF1600(a *[25]uint64)
diff --git a/src/vendor/golang.org/x/crypto/sha3/keccakf_amd64.s b/src/vendor/golang.org/x/crypto/sha3/keccakf_amd64.s
new file mode 100644
index 0000000000..1f53938861
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/keccakf_amd64.s
@@ -0,0 +1,390 @@
+// Copyright 2015 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build amd64 && !purego && gc
+
+// This code was translated into a form compatible with 6a from the public
+// domain sources at https://github.com/gvanas/KeccakCodePackage
+
+// Offsets in state
+#define _ba  (0*8)
+#define _be  (1*8)
+#define _bi  (2*8)
+#define _bo  (3*8)
+#define _bu  (4*8)
+#define _ga  (5*8)
+#define _ge  (6*8)
+#define _gi  (7*8)
+#define _go  (8*8)
+#define _gu  (9*8)
+#define _ka (10*8)
+#define _ke (11*8)
+#define _ki (12*8)
+#define _ko (13*8)
+#define _ku (14*8)
+#define _ma (15*8)
+#define _me (16*8)
+#define _mi (17*8)
+#define _mo (18*8)
+#define _mu (19*8)
+#define _sa (20*8)
+#define _se (21*8)
+#define _si (22*8)
+#define _so (23*8)
+#define _su (24*8)
+
+// Temporary registers
+#define rT1  AX
+
+// Round vars
+#define rpState DI
+#define rpStack SP
+
+#define rDa BX
+#define rDe CX
+#define rDi DX
+#define rDo R8
+#define rDu R9
+
+#define rBa R10
+#define rBe R11
+#define rBi R12
+#define rBo R13
+#define rBu R14
+
+#define rCa SI
+#define rCe BP
+#define rCi rBi
+#define rCo rBo
+#define rCu R15
+
+#define MOVQ_RBI_RCE MOVQ rBi, rCe
+#define XORQ_RT1_RCA XORQ rT1, rCa
+#define XORQ_RT1_RCE XORQ rT1, rCe
+#define XORQ_RBA_RCU XORQ rBa, rCu
+#define XORQ_RBE_RCU XORQ rBe, rCu
+#define XORQ_RDU_RCU XORQ rDu, rCu
+#define XORQ_RDA_RCA XORQ rDa, rCa
+#define XORQ_RDE_RCE XORQ rDe, rCe
+
+#define mKeccakRound(iState, oState, rc, B_RBI_RCE, G_RT1_RCA, G_RT1_RCE, G_RBA_RCU, K_RT1_RCA, K_RT1_RCE, K_RBA_RCU, M_RT1_RCA, M_RT1_RCE, M_RBE_RCU, S_RDU_RCU, S_RDA_RCA, S_RDE_RCE) \
+	/* Prepare round */    \
+	MOVQ rCe, rDa;         \
+	ROLQ $1, rDa;          \
+	                       \
+	MOVQ _bi(iState), rCi; \
+	XORQ _gi(iState), rDi; \
+	XORQ rCu, rDa;         \
+	XORQ _ki(iState), rCi; \
+	XORQ _mi(iState), rDi; \
+	XORQ rDi, rCi;         \
+	                       \
+	MOVQ rCi, rDe;         \
+	ROLQ $1, rDe;          \
+	                       \
+	MOVQ _bo(iState), rCo; \
+	XORQ _go(iState), rDo; \
+	XORQ rCa, rDe;         \
+	XORQ _ko(iState), rCo; \
+	XORQ _mo(iState), rDo; \
+	XORQ rDo, rCo;         \
+	                       \
+	MOVQ rCo, rDi;         \
+	ROLQ $1, rDi;          \
+	                       \
+	MOVQ rCu, rDo;         \
+	XORQ rCe, rDi;         \
+	ROLQ $1, rDo;          \
+	                       \
+	MOVQ rCa, rDu;         \
+	XORQ rCi, rDo;         \
+	ROLQ $1, rDu;          \
+	                       \
+	/* Result b */         \
+	MOVQ _ba(iState), rBa; \
+	MOVQ _ge(iState), rBe; \
+	XORQ rCo, rDu;         \
+	MOVQ _ki(iState), rBi; \
+	MOVQ _mo(iState), rBo; \
+	MOVQ _su(iState), rBu; \
+	XORQ rDe, rBe;         \
+	ROLQ $44, rBe;         \
+	XORQ rDi, rBi;         \
+	XORQ rDa, rBa;         \
+	ROLQ $43, rBi;         \
+	                       \
+	MOVQ rBe, rCa;         \
+	MOVQ rc, rT1;          \
+	ORQ  rBi, rCa;         \
+	XORQ rBa, rT1;         \
+	XORQ rT1, rCa;         \
+	MOVQ rCa, _ba(oState); \
+	                       \
+	XORQ rDu, rBu;         \
+	ROLQ $14, rBu;         \
+	MOVQ rBa, rCu;         \
+	ANDQ rBe, rCu;         \
+	XORQ rBu, rCu;         \
+	MOVQ rCu, _bu(oState); \
+	                       \
+	XORQ rDo, rBo;         \
+	ROLQ $21, rBo;         \
+	MOVQ rBo, rT1;         \
+	ANDQ rBu, rT1;         \
+	XORQ rBi, rT1;         \
+	MOVQ rT1, _bi(oState); \
+	                       \
+	NOTQ rBi;              \
+	ORQ  rBa, rBu;         \
+	ORQ  rBo, rBi;         \
+	XORQ rBo, rBu;         \
+	XORQ rBe, rBi;         \
+	MOVQ rBu, _bo(oState); \
+	MOVQ rBi, _be(oState); \
+	B_RBI_RCE;             \
+	                       \
+	/* Result g */         \
+	MOVQ _gu(iState), rBe; \
+	XORQ rDu, rBe;         \
+	MOVQ _ka(iState), rBi; \
+	ROLQ $20, rBe;         \
+	XORQ rDa, rBi;         \
+	ROLQ $3, rBi;          \
+	MOVQ _bo(iState), rBa; \
+	MOVQ rBe, rT1;         \
+	ORQ  rBi, rT1;         \
+	XORQ rDo, rBa;         \
+	MOVQ _me(iState), rBo; \
+	MOVQ _si(iState), rBu; \
+	ROLQ $28, rBa;         \
+	XORQ rBa, rT1;         \
+	MOVQ rT1, _ga(oState); \
+	G_RT1_RCA;             \
+	                       \
+	XORQ rDe, rBo;         \
+	ROLQ $45, rBo;         \
+	MOVQ rBi, rT1;         \
+	ANDQ rBo, rT1;         \
+	XORQ rBe, rT1;         \
+	MOVQ rT1, _ge(oState); \
+	G_RT1_RCE;             \
+	                       \
+	XORQ rDi, rBu;         \
+	ROLQ $61, rBu;         \
+	MOVQ rBu, rT1;         \
+	ORQ  rBa, rT1;         \
+	XORQ rBo, rT1;         \
+	MOVQ rT1, _go(oState); \
+	                       \
+	ANDQ rBe, rBa;         \
+	XORQ rBu, rBa;         \
+	MOVQ rBa, _gu(oState); \
+	NOTQ rBu;              \
+	G_RBA_RCU;             \
+	                       \
+	ORQ  rBu, rBo;         \
+	XORQ rBi, rBo;         \
+	MOVQ rBo, _gi(oState); \
+	                       \
+	/* Result k */         \
+	MOVQ _be(iState), rBa; \
+	MOVQ _gi(iState), rBe; \
+	MOVQ _ko(iState), rBi; \
+	MOVQ _mu(iState), rBo; \
+	MOVQ _sa(iState), rBu; \
+	XORQ rDi, rBe;         \
+	ROLQ $6, rBe;          \
+	XORQ rDo, rBi;         \
+	ROLQ $25, rBi;         \
+	MOVQ rBe, rT1;         \
+	ORQ  rBi, rT1;         \
+	XORQ rDe, rBa;         \
+	ROLQ $1, rBa;          \
+	XORQ rBa, rT1;         \
+	MOVQ rT1, _ka(oState); \
+	K_RT1_RCA;             \
+	                       \
+	XORQ rDu, rBo;         \
+	ROLQ $8, rBo;          \
+	MOVQ rBi, rT1;         \
+	ANDQ rBo, rT1;         \
+	XORQ rBe, rT1;         \
+	MOVQ rT1, _ke(oState); \
+	K_RT1_RCE;             \
+	                       \
+	XORQ rDa, rBu;         \
+	ROLQ $18, rBu;         \
+	NOTQ rBo;              \
+	MOVQ rBo, rT1;         \
+	ANDQ rBu, rT1;         \
+	XORQ rBi, rT1;         \
+	MOVQ rT1, _ki(oState); \
+	                       \
+	MOVQ rBu, rT1;         \
+	ORQ  rBa, rT1;         \
+	XORQ rBo, rT1;         \
+	MOVQ rT1, _ko(oState); \
+	                       \
+	ANDQ rBe, rBa;         \
+	XORQ rBu, rBa;         \
+	MOVQ rBa, _ku(oState); \
+	K_RBA_RCU;             \
+	                       \
+	/* Result m */         \
+	MOVQ _ga(iState), rBe; \
+	XORQ rDa, rBe;         \
+	MOVQ _ke(iState), rBi; \
+	ROLQ $36, rBe;         \
+	XORQ rDe, rBi;         \
+	MOVQ _bu(iState), rBa; \
+	ROLQ $10, rBi;         \
+	MOVQ rBe, rT1;         \
+	MOVQ _mi(iState), rBo; \
+	ANDQ rBi, rT1;         \
+	XORQ rDu, rBa;         \
+	MOVQ _so(iState), rBu; \
+	ROLQ $27, rBa;         \
+	XORQ rBa, rT1;         \
+	MOVQ rT1, _ma(oState); \
+	M_RT1_RCA;             \
+	                       \
+	XORQ rDi, rBo;         \
+	ROLQ $15, rBo;         \
+	MOVQ rBi, rT1;         \
+	ORQ  rBo, rT1;         \
+	XORQ rBe, rT1;         \
+	MOVQ rT1, _me(oState); \
+	M_RT1_RCE;             \
+	                       \
+	XORQ rDo, rBu;         \
+	ROLQ $56, rBu;         \
+	NOTQ rBo;              \
+	MOVQ rBo, rT1;         \
+	ORQ  rBu, rT1;         \
+	XORQ rBi, rT1;         \
+	MOVQ rT1, _mi(oState); \
+	                       \
+	ORQ  rBa, rBe;         \
+	XORQ rBu, rBe;         \
+	MOVQ rBe, _mu(oState); \
+	                       \
+	ANDQ rBa, rBu;         \
+	XORQ rBo, rBu;         \
+	MOVQ rBu, _mo(oState); \
+	M_RBE_RCU;             \
+	                       \
+	/* Result s */         \
+	MOVQ _bi(iState), rBa; \
+	MOVQ _go(iState), rBe; \
+	MOVQ _ku(iState), rBi; \
+	XORQ rDi, rBa;         \
+	MOVQ _ma(iState), rBo; \
+	ROLQ $62, rBa;         \
+	XORQ rDo, rBe;         \
+	MOVQ _se(iState), rBu; \
+	ROLQ $55, rBe;         \
+	                       \
+	XORQ rDu, rBi;         \
+	MOVQ rBa, rDu;         \
+	XORQ rDe, rBu;         \
+	ROLQ $2, rBu;          \
+	ANDQ rBe, rDu;         \
+	XORQ rBu, rDu;         \
+	MOVQ rDu, _su(oState); \
+	                       \
+	ROLQ $39, rBi;         \
+	S_RDU_RCU;             \
+	NOTQ rBe;              \
+	XORQ rDa, rBo;         \
+	MOVQ rBe, rDa;         \
+	ANDQ rBi, rDa;         \
+	XORQ rBa, rDa;         \
+	MOVQ rDa, _sa(oState); \
+	S_RDA_RCA;             \
+	                       \
+	ROLQ $41, rBo;         \
+	MOVQ rBi, rDe;         \
+	ORQ  rBo, rDe;         \
+	XORQ rBe, rDe;         \
+	MOVQ rDe, _se(oState); \
+	S_RDE_RCE;             \
+	                       \
+	MOVQ rBo, rDi;         \
+	MOVQ rBu, rDo;         \
+	ANDQ rBu, rDi;         \
+	ORQ  rBa, rDo;         \
+	XORQ rBi, rDi;         \
+	XORQ rBo, rDo;         \
+	MOVQ rDi, _si(oState); \
+	MOVQ rDo, _so(oState)  \
+
+// func keccakF1600(a *[25]uint64)
+TEXT ·keccakF1600(SB), 0, $200-8
+	MOVQ a+0(FP), rpState
+
+	// Convert the user state into an internal state
+	NOTQ _be(rpState)
+	NOTQ _bi(rpState)
+	NOTQ _go(rpState)
+	NOTQ _ki(rpState)
+	NOTQ _mi(rpState)
+	NOTQ _sa(rpState)
+
+	// Execute the KeccakF permutation
+	MOVQ _ba(rpState), rCa
+	MOVQ _be(rpState), rCe
+	MOVQ _bu(rpState), rCu
+
+	XORQ _ga(rpState), rCa
+	XORQ _ge(rpState), rCe
+	XORQ _gu(rpState), rCu
+
+	XORQ _ka(rpState), rCa
+	XORQ _ke(rpState), rCe
+	XORQ _ku(rpState), rCu
+
+	XORQ _ma(rpState), rCa
+	XORQ _me(rpState), rCe
+	XORQ _mu(rpState), rCu
+
+	XORQ _sa(rpState), rCa
+	XORQ _se(rpState), rCe
+	MOVQ _si(rpState), rDi
+	MOVQ _so(rpState), rDo
+	XORQ _su(rpState), rCu
+
+	mKeccakRound(rpState, rpStack, $0x0000000000000001, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x0000000000008082, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x800000000000808a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x8000000080008000, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x000000000000808b, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x0000000080000001, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x8000000080008081, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x8000000000008009, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x000000000000008a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x0000000000000088, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x0000000080008009, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x000000008000000a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x000000008000808b, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x800000000000008b, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x8000000000008089, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x8000000000008003, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x8000000000008002, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x8000000000000080, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x000000000000800a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x800000008000000a, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x8000000080008081, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x8000000000008080, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpState, rpStack, $0x0000000080000001, MOVQ_RBI_RCE, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBA_RCU, XORQ_RT1_RCA, XORQ_RT1_RCE, XORQ_RBE_RCU, XORQ_RDU_RCU, XORQ_RDA_RCA, XORQ_RDE_RCE)
+	mKeccakRound(rpStack, rpState, $0x8000000080008008, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP, NOP)
+
+	// Revert the internal state to the user state
+	NOTQ _be(rpState)
+	NOTQ _bi(rpState)
+	NOTQ _go(rpState)
+	NOTQ _ki(rpState)
+	NOTQ _mi(rpState)
+	NOTQ _sa(rpState)
+
+	RET
diff --git a/src/vendor/golang.org/x/crypto/sha3/register.go b/src/vendor/golang.org/x/crypto/sha3/register.go
new file mode 100644
index 0000000000..addfd5049b
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/register.go
@@ -0,0 +1,18 @@
+// Copyright 2014 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build go1.4
+
+package sha3
+
+import (
+	"crypto"
+)
+
+func init() {
+	crypto.RegisterHash(crypto.SHA3_224, New224)
+	crypto.RegisterHash(crypto.SHA3_256, New256)
+	crypto.RegisterHash(crypto.SHA3_384, New384)
+	crypto.RegisterHash(crypto.SHA3_512, New512)
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/sha3.go b/src/vendor/golang.org/x/crypto/sha3/sha3.go
new file mode 100644
index 0000000000..4884d172a4
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/sha3.go
@@ -0,0 +1,197 @@
+// Copyright 2014 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package sha3
+
+// spongeDirection indicates the direction bytes are flowing through the sponge.
+type spongeDirection int
+
+const (
+	// spongeAbsorbing indicates that the sponge is absorbing input.
+	spongeAbsorbing spongeDirection = iota
+	// spongeSqueezing indicates that the sponge is being squeezed.
+	spongeSqueezing
+)
+
+const (
+	// maxRate is the maximum size of the internal buffer. SHAKE-256
+	// currently needs the largest buffer.
+	maxRate = 168
+)
+
+type state struct {
+	// Generic sponge components.
+	a    [25]uint64 // main state of the hash
+	buf  []byte     // points into storage
+	rate int        // the number of bytes of state to use
+
+	// dsbyte contains the "domain separation" bits and the first bit of
+	// the padding. Sections 6.1 and 6.2 of [1] separate the outputs of the
+	// SHA-3 and SHAKE functions by appending bitstrings to the message.
+	// Using a little-endian bit-ordering convention, these are "01" for SHA-3
+	// and "1111" for SHAKE, or 00000010b and 00001111b, respectively. Then the
+	// padding rule from section 5.1 is applied to pad the message to a multiple
+	// of the rate, which involves adding a "1" bit, zero or more "0" bits, and
+	// a final "1" bit. We merge the first "1" bit from the padding into dsbyte,
+	// giving 00000110b (0x06) and 00011111b (0x1f).
+	// [1] http://csrc.nist.gov/publications/drafts/fips-202/fips_202_draft.pdf
+	//     "Draft FIPS 202: SHA-3 Standard: Permutation-Based Hash and
+	//      Extendable-Output Functions (May 2014)"
+	dsbyte byte
+
+	storage storageBuf
+
+	// Specific to SHA-3 and SHAKE.
+	outputLen int             // the default output size in bytes
+	state     spongeDirection // whether the sponge is absorbing or squeezing
+}
+
+// BlockSize returns the rate of sponge underlying this hash function.
+func (d *state) BlockSize() int { return d.rate }
+
+// Size returns the output size of the hash function in bytes.
+func (d *state) Size() int { return d.outputLen }
+
+// Reset clears the internal state by zeroing the sponge state and
+// the byte buffer, and setting Sponge.state to absorbing.
+func (d *state) Reset() {
+	// Zero the permutation's state.
+	for i := range d.a {
+		d.a[i] = 0
+	}
+	d.state = spongeAbsorbing
+	d.buf = d.storage.asBytes()[:0]
+}
+
+func (d *state) clone() *state {
+	ret := *d
+	if ret.state == spongeAbsorbing {
+		ret.buf = ret.storage.asBytes()[:len(ret.buf)]
+	} else {
+		ret.buf = ret.storage.asBytes()[d.rate-cap(d.buf) : d.rate]
+	}
+
+	return &ret
+}
+
+// permute applies the KeccakF-1600 permutation. It handles
+// any input-output buffering.
+func (d *state) permute() {
+	switch d.state {
+	case spongeAbsorbing:
+		// If we're absorbing, we need to xor the input into the state
+		// before applying the permutation.
+		xorIn(d, d.buf)
+		d.buf = d.storage.asBytes()[:0]
+		keccakF1600(&d.a)
+	case spongeSqueezing:
+		// If we're squeezing, we need to apply the permutation before
+		// copying more output.
+		keccakF1600(&d.a)
+		d.buf = d.storage.asBytes()[:d.rate]
+		copyOut(d, d.buf)
+	}
+}
+
+// pads appends the domain separation bits in dsbyte, applies
+// the multi-bitrate 10..1 padding rule, and permutes the state.
+func (d *state) padAndPermute(dsbyte byte) {
+	if d.buf == nil {
+		d.buf = d.storage.asBytes()[:0]
+	}
+	// Pad with this instance's domain-separator bits. We know that there's
+	// at least one byte of space in d.buf because, if it were full,
+	// permute would have been called to empty it. dsbyte also contains the
+	// first one bit for the padding. See the comment in the state struct.
+	d.buf = append(d.buf, dsbyte)
+	zerosStart := len(d.buf)
+	d.buf = d.storage.asBytes()[:d.rate]
+	for i := zerosStart; i < d.rate; i++ {
+		d.buf[i] = 0
+	}
+	// This adds the final one bit for the padding. Because of the way that
+	// bits are numbered from the LSB upwards, the final bit is the MSB of
+	// the last byte.
+	d.buf[d.rate-1] ^= 0x80
+	// Apply the permutation
+	d.permute()
+	d.state = spongeSqueezing
+	d.buf = d.storage.asBytes()[:d.rate]
+	copyOut(d, d.buf)
+}
+
+// Write absorbs more data into the hash's state. It panics if any
+// output has already been read.
+func (d *state) Write(p []byte) (written int, err error) {
+	if d.state != spongeAbsorbing {
+		panic("sha3: Write after Read")
+	}
+	if d.buf == nil {
+		d.buf = d.storage.asBytes()[:0]
+	}
+	written = len(p)
+
+	for len(p) > 0 {
+		if len(d.buf) == 0 && len(p) >= d.rate {
+			// The fast path; absorb a full "rate" bytes of input and apply the permutation.
+			xorIn(d, p[:d.rate])
+			p = p[d.rate:]
+			keccakF1600(&d.a)
+		} else {
+			// The slow path; buffer the input until we can fill the sponge, and then xor it in.
+			todo := d.rate - len(d.buf)
+			if todo > len(p) {
+				todo = len(p)
+			}
+			d.buf = append(d.buf, p[:todo]...)
+			p = p[todo:]
+
+			// If the sponge is full, apply the permutation.
+			if len(d.buf) == d.rate {
+				d.permute()
+			}
+		}
+	}
+
+	return
+}
+
+// Read squeezes an arbitrary number of bytes from the sponge.
+func (d *state) Read(out []byte) (n int, err error) {
+	// If we're still absorbing, pad and apply the permutation.
+	if d.state == spongeAbsorbing {
+		d.padAndPermute(d.dsbyte)
+	}
+
+	n = len(out)
+
+	// Now, do the squeezing.
+	for len(out) > 0 {
+		n := copy(out, d.buf)
+		d.buf = d.buf[n:]
+		out = out[n:]
+
+		// Apply the permutation if we've squeezed the sponge dry.
+		if len(d.buf) == 0 {
+			d.permute()
+		}
+	}
+
+	return
+}
+
+// Sum applies padding to the hash state and then squeezes out the desired
+// number of output bytes. It panics if any output has already been read.
+func (d *state) Sum(in []byte) []byte {
+	if d.state != spongeAbsorbing {
+		panic("sha3: Sum after Read")
+	}
+
+	// Make a copy of the original hash so that caller can keep writing
+	// and summing.
+	dup := d.clone()
+	hash := make([]byte, dup.outputLen, 64) // explicit cap to allow stack allocation
+	dup.Read(hash)
+	return append(in, hash...)
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/sha3_s390x.go b/src/vendor/golang.org/x/crypto/sha3/sha3_s390x.go
new file mode 100644
index 0000000000..d861bca528
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/sha3_s390x.go
@@ -0,0 +1,288 @@
+// Copyright 2017 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build gc && !purego
+
+package sha3
+
+// This file contains code for using the 'compute intermediate
+// message digest' (KIMD) and 'compute last message digest' (KLMD)
+// instructions to compute SHA-3 and SHAKE hashes on IBM Z.
+
+import (
+	"hash"
+
+	"golang.org/x/sys/cpu"
+)
+
+// codes represent 7-bit KIMD/KLMD function codes as defined in
+// the Principles of Operation.
+type code uint64
+
+const (
+	// function codes for KIMD/KLMD
+	sha3_224  code = 32
+	sha3_256       = 33
+	sha3_384       = 34
+	sha3_512       = 35
+	shake_128      = 36
+	shake_256      = 37
+	nopad          = 0x100
+)
+
+// kimd is a wrapper for the 'compute intermediate message digest' instruction.
+// src must be a multiple of the rate for the given function code.
+//
+//go:noescape
+func kimd(function code, chain *[200]byte, src []byte)
+
+// klmd is a wrapper for the 'compute last message digest' instruction.
+// src padding is handled by the instruction.
+//
+//go:noescape
+func klmd(function code, chain *[200]byte, dst, src []byte)
+
+type asmState struct {
+	a         [200]byte       // 1600 bit state
+	buf       []byte          // care must be taken to ensure cap(buf) is a multiple of rate
+	rate      int             // equivalent to block size
+	storage   [3072]byte      // underlying storage for buf
+	outputLen int             // output length for full security
+	function  code            // KIMD/KLMD function code
+	state     spongeDirection // whether the sponge is absorbing or squeezing
+}
+
+func newAsmState(function code) *asmState {
+	var s asmState
+	s.function = function
+	switch function {
+	case sha3_224:
+		s.rate = 144
+		s.outputLen = 28
+	case sha3_256:
+		s.rate = 136
+		s.outputLen = 32
+	case sha3_384:
+		s.rate = 104
+		s.outputLen = 48
+	case sha3_512:
+		s.rate = 72
+		s.outputLen = 64
+	case shake_128:
+		s.rate = 168
+		s.outputLen = 32
+	case shake_256:
+		s.rate = 136
+		s.outputLen = 64
+	default:
+		panic("sha3: unrecognized function code")
+	}
+
+	// limit s.buf size to a multiple of s.rate
+	s.resetBuf()
+	return &s
+}
+
+func (s *asmState) clone() *asmState {
+	c := *s
+	c.buf = c.storage[:len(s.buf):cap(s.buf)]
+	return &c
+}
+
+// copyIntoBuf copies b into buf. It will panic if there is not enough space to
+// store all of b.
+func (s *asmState) copyIntoBuf(b []byte) {
+	bufLen := len(s.buf)
+	s.buf = s.buf[:len(s.buf)+len(b)]
+	copy(s.buf[bufLen:], b)
+}
+
+// resetBuf points buf at storage, sets the length to 0 and sets cap to be a
+// multiple of the rate.
+func (s *asmState) resetBuf() {
+	max := (cap(s.storage) / s.rate) * s.rate
+	s.buf = s.storage[:0:max]
+}
+
+// Write (via the embedded io.Writer interface) adds more data to the running hash.
+// It never returns an error.
+func (s *asmState) Write(b []byte) (int, error) {
+	if s.state != spongeAbsorbing {
+		panic("sha3: Write after Read")
+	}
+	length := len(b)
+	for len(b) > 0 {
+		if len(s.buf) == 0 && len(b) >= cap(s.buf) {
+			// Hash the data directly and push any remaining bytes
+			// into the buffer.
+			remainder := len(b) % s.rate
+			kimd(s.function, &s.a, b[:len(b)-remainder])
+			if remainder != 0 {
+				s.copyIntoBuf(b[len(b)-remainder:])
+			}
+			return length, nil
+		}
+
+		if len(s.buf) == cap(s.buf) {
+			// flush the buffer
+			kimd(s.function, &s.a, s.buf)
+			s.buf = s.buf[:0]
+		}
+
+		// copy as much as we can into the buffer
+		n := len(b)
+		if len(b) > cap(s.buf)-len(s.buf) {
+			n = cap(s.buf) - len(s.buf)
+		}
+		s.copyIntoBuf(b[:n])
+		b = b[n:]
+	}
+	return length, nil
+}
+
+// Read squeezes an arbitrary number of bytes from the sponge.
+func (s *asmState) Read(out []byte) (n int, err error) {
+	n = len(out)
+
+	// need to pad if we were absorbing
+	if s.state == spongeAbsorbing {
+		s.state = spongeSqueezing
+
+		// write hash directly into out if possible
+		if len(out)%s.rate == 0 {
+			klmd(s.function, &s.a, out, s.buf) // len(out) may be 0
+			s.buf = s.buf[:0]
+			return
+		}
+
+		// write hash into buffer
+		max := cap(s.buf)
+		if max > len(out) {
+			max = (len(out)/s.rate)*s.rate + s.rate
+		}
+		klmd(s.function, &s.a, s.buf[:max], s.buf)
+		s.buf = s.buf[:max]
+	}
+
+	for len(out) > 0 {
+		// flush the buffer
+		if len(s.buf) != 0 {
+			c := copy(out, s.buf)
+			out = out[c:]
+			s.buf = s.buf[c:]
+			continue
+		}
+
+		// write hash directly into out if possible
+		if len(out)%s.rate == 0 {
+			klmd(s.function|nopad, &s.a, out, nil)
+			return
+		}
+
+		// write hash into buffer
+		s.resetBuf()
+		if cap(s.buf) > len(out) {
+			s.buf = s.buf[:(len(out)/s.rate)*s.rate+s.rate]
+		}
+		klmd(s.function|nopad, &s.a, s.buf, nil)
+	}
+	return
+}
+
+// Sum appends the current hash to b and returns the resulting slice.
+// It does not change the underlying hash state.
+func (s *asmState) Sum(b []byte) []byte {
+	if s.state != spongeAbsorbing {
+		panic("sha3: Sum after Read")
+	}
+
+	// Copy the state to preserve the original.
+	a := s.a
+
+	// Hash the buffer. Note that we don't clear it because we
+	// aren't updating the state.
+	klmd(s.function, &a, nil, s.buf)
+	return append(b, a[:s.outputLen]...)
+}
+
+// Reset resets the Hash to its initial state.
+func (s *asmState) Reset() {
+	for i := range s.a {
+		s.a[i] = 0
+	}
+	s.resetBuf()
+	s.state = spongeAbsorbing
+}
+
+// Size returns the number of bytes Sum will return.
+func (s *asmState) Size() int {
+	return s.outputLen
+}
+
+// BlockSize returns the hash's underlying block size.
+// The Write method must be able to accept any amount
+// of data, but it may operate more efficiently if all writes
+// are a multiple of the block size.
+func (s *asmState) BlockSize() int {
+	return s.rate
+}
+
+// Clone returns a copy of the ShakeHash in its current state.
+func (s *asmState) Clone() ShakeHash {
+	return s.clone()
+}
+
+// new224Asm returns an assembly implementation of SHA3-224 if available,
+// otherwise it returns nil.
+func new224Asm() hash.Hash {
+	if cpu.S390X.HasSHA3 {
+		return newAsmState(sha3_224)
+	}
+	return nil
+}
+
+// new256Asm returns an assembly implementation of SHA3-256 if available,
+// otherwise it returns nil.
+func new256Asm() hash.Hash {
+	if cpu.S390X.HasSHA3 {
+		return newAsmState(sha3_256)
+	}
+	return nil
+}
+
+// new384Asm returns an assembly implementation of SHA3-384 if available,
+// otherwise it returns nil.
+func new384Asm() hash.Hash {
+	if cpu.S390X.HasSHA3 {
+		return newAsmState(sha3_384)
+	}
+	return nil
+}
+
+// new512Asm returns an assembly implementation of SHA3-512 if available,
+// otherwise it returns nil.
+func new512Asm() hash.Hash {
+	if cpu.S390X.HasSHA3 {
+		return newAsmState(sha3_512)
+	}
+	return nil
+}
+
+// newShake128Asm returns an assembly implementation of SHAKE-128 if available,
+// otherwise it returns nil.
+func newShake128Asm() ShakeHash {
+	if cpu.S390X.HasSHA3 {
+		return newAsmState(shake_128)
+	}
+	return nil
+}
+
+// newShake256Asm returns an assembly implementation of SHAKE-256 if available,
+// otherwise it returns nil.
+func newShake256Asm() ShakeHash {
+	if cpu.S390X.HasSHA3 {
+		return newAsmState(shake_256)
+	}
+	return nil
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/sha3_s390x.s b/src/vendor/golang.org/x/crypto/sha3/sha3_s390x.s
new file mode 100644
index 0000000000..826b862c77
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/sha3_s390x.s
@@ -0,0 +1,33 @@
+// Copyright 2017 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build gc && !purego
+
+#include "textflag.h"
+
+// func kimd(function code, chain *[200]byte, src []byte)
+TEXT ·kimd(SB), NOFRAME|NOSPLIT, $0-40
+	MOVD function+0(FP), R0
+	MOVD chain+8(FP), R1
+	LMG  src+16(FP), R2, R3 // R2=base, R3=len
+
+continue:
+	WORD $0xB93E0002 // KIMD --, R2
+	BVS  continue    // continue if interrupted
+	MOVD $0, R0      // reset R0 for pre-go1.8 compilers
+	RET
+
+// func klmd(function code, chain *[200]byte, dst, src []byte)
+TEXT ·klmd(SB), NOFRAME|NOSPLIT, $0-64
+	// TODO: SHAKE support
+	MOVD function+0(FP), R0
+	MOVD chain+8(FP), R1
+	LMG  dst+16(FP), R2, R3 // R2=base, R3=len
+	LMG  src+40(FP), R4, R5 // R4=base, R5=len
+
+continue:
+	WORD $0xB93F0024 // KLMD R2, R4
+	BVS  continue    // continue if interrupted
+	MOVD $0, R0      // reset R0 for pre-go1.8 compilers
+	RET
diff --git a/src/vendor/golang.org/x/crypto/sha3/shake.go b/src/vendor/golang.org/x/crypto/sha3/shake.go
new file mode 100644
index 0000000000..bb69984027
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/shake.go
@@ -0,0 +1,172 @@
+// Copyright 2014 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package sha3
+
+// This file defines the ShakeHash interface, and provides
+// functions for creating SHAKE and cSHAKE instances, as well as utility
+// functions for hashing bytes to arbitrary-length output.
+//
+//
+// SHAKE implementation is based on FIPS PUB 202 [1]
+// cSHAKE implementations is based on NIST SP 800-185 [2]
+//
+// [1] https://nvlpubs.nist.gov/nistpubs/FIPS/NIST.FIPS.202.pdf
+// [2] https://doi.org/10.6028/NIST.SP.800-185
+
+import (
+	"encoding/binary"
+	"hash"
+	"io"
+)
+
+// ShakeHash defines the interface to hash functions that support
+// arbitrary-length output. When used as a plain [hash.Hash], it
+// produces minimum-length outputs that provide full-strength generic
+// security.
+type ShakeHash interface {
+	hash.Hash
+
+	// Read reads more output from the hash; reading affects the hash's
+	// state. (ShakeHash.Read is thus very different from Hash.Sum)
+	// It never returns an error, but subsequent calls to Write or Sum
+	// will panic.
+	io.Reader
+
+	// Clone returns a copy of the ShakeHash in its current state.
+	Clone() ShakeHash
+}
+
+// cSHAKE specific context
+type cshakeState struct {
+	*state // SHA-3 state context and Read/Write operations
+
+	// initBlock is the cSHAKE specific initialization set of bytes. It is initialized
+	// by newCShake function and stores concatenation of N followed by S, encoded
+	// by the method specified in 3.3 of [1].
+	// It is stored here in order for Reset() to be able to put context into
+	// initial state.
+	initBlock []byte
+}
+
+// Consts for configuring initial SHA-3 state
+const (
+	dsbyteShake  = 0x1f
+	dsbyteCShake = 0x04
+	rate128      = 168
+	rate256      = 136
+)
+
+func bytepad(input []byte, w int) []byte {
+	// leftEncode always returns max 9 bytes
+	buf := make([]byte, 0, 9+len(input)+w)
+	buf = append(buf, leftEncode(uint64(w))...)
+	buf = append(buf, input...)
+	padlen := w - (len(buf) % w)
+	return append(buf, make([]byte, padlen)...)
+}
+
+func leftEncode(value uint64) []byte {
+	var b [9]byte
+	binary.BigEndian.PutUint64(b[1:], value)
+	// Trim all but last leading zero bytes
+	i := byte(1)
+	for i < 8 && b[i] == 0 {
+		i++
+	}
+	// Prepend number of encoded bytes
+	b[i-1] = 9 - i
+	return b[i-1:]
+}
+
+func newCShake(N, S []byte, rate, outputLen int, dsbyte byte) ShakeHash {
+	c := cshakeState{state: &state{rate: rate, outputLen: outputLen, dsbyte: dsbyte}}
+
+	// leftEncode returns max 9 bytes
+	c.initBlock = make([]byte, 0, 9*2+len(N)+len(S))
+	c.initBlock = append(c.initBlock, leftEncode(uint64(len(N)*8))...)
+	c.initBlock = append(c.initBlock, N...)
+	c.initBlock = append(c.initBlock, leftEncode(uint64(len(S)*8))...)
+	c.initBlock = append(c.initBlock, S...)
+	c.Write(bytepad(c.initBlock, c.rate))
+	return &c
+}
+
+// Reset resets the hash to initial state.
+func (c *cshakeState) Reset() {
+	c.state.Reset()
+	c.Write(bytepad(c.initBlock, c.rate))
+}
+
+// Clone returns copy of a cSHAKE context within its current state.
+func (c *cshakeState) Clone() ShakeHash {
+	b := make([]byte, len(c.initBlock))
+	copy(b, c.initBlock)
+	return &cshakeState{state: c.clone(), initBlock: b}
+}
+
+// Clone returns copy of SHAKE context within its current state.
+func (c *state) Clone() ShakeHash {
+	return c.clone()
+}
+
+// NewShake128 creates a new SHAKE128 variable-output-length ShakeHash.
+// Its generic security strength is 128 bits against all attacks if at
+// least 32 bytes of its output are used.
+func NewShake128() ShakeHash {
+	if h := newShake128Asm(); h != nil {
+		return h
+	}
+	return &state{rate: rate128, outputLen: 32, dsbyte: dsbyteShake}
+}
+
+// NewShake256 creates a new SHAKE256 variable-output-length ShakeHash.
+// Its generic security strength is 256 bits against all attacks if
+// at least 64 bytes of its output are used.
+func NewShake256() ShakeHash {
+	if h := newShake256Asm(); h != nil {
+		return h
+	}
+	return &state{rate: rate256, outputLen: 64, dsbyte: dsbyteShake}
+}
+
+// NewCShake128 creates a new instance of cSHAKE128 variable-output-length ShakeHash,
+// a customizable variant of SHAKE128.
+// N is used to define functions based on cSHAKE, it can be empty when plain cSHAKE is
+// desired. S is a customization byte string used for domain separation - two cSHAKE
+// computations on same input with different S yield unrelated outputs.
+// When N and S are both empty, this is equivalent to NewShake128.
+func NewCShake128(N, S []byte) ShakeHash {
+	if len(N) == 0 && len(S) == 0 {
+		return NewShake128()
+	}
+	return newCShake(N, S, rate128, 32, dsbyteCShake)
+}
+
+// NewCShake256 creates a new instance of cSHAKE256 variable-output-length ShakeHash,
+// a customizable variant of SHAKE256.
+// N is used to define functions based on cSHAKE, it can be empty when plain cSHAKE is
+// desired. S is a customization byte string used for domain separation - two cSHAKE
+// computations on same input with different S yield unrelated outputs.
+// When N and S are both empty, this is equivalent to NewShake256.
+func NewCShake256(N, S []byte) ShakeHash {
+	if len(N) == 0 && len(S) == 0 {
+		return NewShake256()
+	}
+	return newCShake(N, S, rate256, 64, dsbyteCShake)
+}
+
+// ShakeSum128 writes an arbitrary-length digest of data into hash.
+func ShakeSum128(hash, data []byte) {
+	h := NewShake128()
+	h.Write(data)
+	h.Read(hash)
+}
+
+// ShakeSum256 writes an arbitrary-length digest of data into hash.
+func ShakeSum256(hash, data []byte) {
+	h := NewShake256()
+	h.Write(data)
+	h.Read(hash)
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/shake_generic.go b/src/vendor/golang.org/x/crypto/sha3/shake_generic.go
new file mode 100644
index 0000000000..8d31cf5be2
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/shake_generic.go
@@ -0,0 +1,19 @@
+// Copyright 2017 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build !gc || purego || !s390x
+
+package sha3
+
+// newShake128Asm returns an assembly implementation of SHAKE-128 if available,
+// otherwise it returns nil.
+func newShake128Asm() ShakeHash {
+	return nil
+}
+
+// newShake256Asm returns an assembly implementation of SHAKE-256 if available,
+// otherwise it returns nil.
+func newShake256Asm() ShakeHash {
+	return nil
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/xor.go b/src/vendor/golang.org/x/crypto/sha3/xor.go
new file mode 100644
index 0000000000..7337cca88e
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/xor.go
@@ -0,0 +1,23 @@
+// Copyright 2015 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build (!amd64 && !386 && !ppc64le) || purego
+
+package sha3
+
+// A storageBuf is an aligned array of maxRate bytes.
+type storageBuf [maxRate]byte
+
+func (b *storageBuf) asBytes() *[maxRate]byte {
+	return (*[maxRate]byte)(b)
+}
+
+var (
+	xorIn            = xorInGeneric
+	copyOut          = copyOutGeneric
+	xorInUnaligned   = xorInGeneric
+	copyOutUnaligned = copyOutGeneric
+)
+
+const xorImplementationUnaligned = "generic"
diff --git a/src/vendor/golang.org/x/crypto/sha3/xor_generic.go b/src/vendor/golang.org/x/crypto/sha3/xor_generic.go
new file mode 100644
index 0000000000..8d94771127
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/xor_generic.go
@@ -0,0 +1,28 @@
+// Copyright 2015 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package sha3
+
+import "encoding/binary"
+
+// xorInGeneric xors the bytes in buf into the state; it
+// makes no non-portable assumptions about memory layout
+// or alignment.
+func xorInGeneric(d *state, buf []byte) {
+	n := len(buf) / 8
+
+	for i := 0; i < n; i++ {
+		a := binary.LittleEndian.Uint64(buf)
+		d.a[i] ^= a
+		buf = buf[8:]
+	}
+}
+
+// copyOutGeneric copies uint64s to a byte buffer.
+func copyOutGeneric(d *state, b []byte) {
+	for i := 0; len(b) >= 8; i++ {
+		binary.LittleEndian.PutUint64(b, d.a[i])
+		b = b[8:]
+	}
+}
diff --git a/src/vendor/golang.org/x/crypto/sha3/xor_unaligned.go b/src/vendor/golang.org/x/crypto/sha3/xor_unaligned.go
new file mode 100644
index 0000000000..870e2d16e0
--- /dev/null
+++ b/src/vendor/golang.org/x/crypto/sha3/xor_unaligned.go
@@ -0,0 +1,66 @@
+// Copyright 2015 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+//go:build (amd64 || 386 || ppc64le) && !purego
+
+package sha3
+
+import "unsafe"
+
+// A storageBuf is an aligned array of maxRate bytes.
+type storageBuf [maxRate / 8]uint64
+
+func (b *storageBuf) asBytes() *[maxRate]byte {
+	return (*[maxRate]byte)(unsafe.Pointer(b))
+}
+
+// xorInUnaligned uses unaligned reads and writes to update d.a to contain d.a
+// XOR buf.
+func xorInUnaligned(d *state, buf []byte) {
+	n := len(buf)
+	bw := (*[maxRate / 8]uint64)(unsafe.Pointer(&buf[0]))[: n/8 : n/8]
+	if n >= 72 {
+		d.a[0] ^= bw[0]
+		d.a[1] ^= bw[1]
+		d.a[2] ^= bw[2]
+		d.a[3] ^= bw[3]
+		d.a[4] ^= bw[4]
+		d.a[5] ^= bw[5]
+		d.a[6] ^= bw[6]
+		d.a[7] ^= bw[7]
+		d.a[8] ^= bw[8]
+	}
+	if n >= 104 {
+		d.a[9] ^= bw[9]
+		d.a[10] ^= bw[10]
+		d.a[11] ^= bw[11]
+		d.a[12] ^= bw[12]
+	}
+	if n >= 136 {
+		d.a[13] ^= bw[13]
+		d.a[14] ^= bw[14]
+		d.a[15] ^= bw[15]
+		d.a[16] ^= bw[16]
+	}
+	if n >= 144 {
+		d.a[17] ^= bw[17]
+	}
+	if n >= 168 {
+		d.a[18] ^= bw[18]
+		d.a[19] ^= bw[19]
+		d.a[20] ^= bw[20]
+	}
+}
+
+func copyOutUnaligned(d *state, buf []byte) {
+	ab := (*[maxRate]uint8)(unsafe.Pointer(&d.a[0]))
+	copy(buf, ab[:])
+}
+
+var (
+	xorIn   = xorInUnaligned
+	copyOut = copyOutUnaligned
+)
+
+const xorImplementationUnaligned = "unaligned"
diff --git a/src/vendor/modules.txt b/src/vendor/modules.txt
index 19555379f6..a6de6f5650 100644
--- a/src/vendor/modules.txt
+++ b/src/vendor/modules.txt
@@ -7,6 +7,7 @@ golang.org/x/crypto/cryptobyte/asn1
 golang.org/x/crypto/hkdf
 golang.org/x/crypto/internal/alias
 golang.org/x/crypto/internal/poly1305
+golang.org/x/crypto/sha3
 # golang.org/x/net v0.24.1-0.20240405221309-ec05fdcd7114
 ## explicit; go 1.18
 golang.org/x/net/dns/dnsmessage