Initial push for Sparkle

2 files changed, 737 insertions, 0 deletions

diff --git a/crypto/src/crypto/digests/SparkleDigest.cs b/crypto/src/crypto/digests/SparkleDigest.cs
new file mode 100644
index 000000000..33c6079c8
--- /dev/null
+++ b/crypto/src/crypto/digests/SparkleDigest.cs
@@ -0,0 +1,254 @@
+using System;
+using System.IO;
+using Org.BouncyCastle.Crypto.Utilities;
+using Org.BouncyCastle.Utilities;
+
+/**
+ * Sparkle v1.2, based on the current round 3 submission, https://sparkle-lwc.github.io/
+ * Reference C implementation: https://github.com/cryptolu/sparkle
+ * Specification: https://csrc.nist.gov/CSRC/media/Projects/lightweight-cryptography/documents/finalist-round/updated-spec-doc/sparkle-spec-final.pdf
+ */
+
+namespace Org.BouncyCastle.Crypto.Digests
+{
+    public class SparkleDigest : IDigest
+    {
+        public enum SparkleParameters
+        {
+            ESCH256,
+            ESCH384
+        }
+
+        private readonly uint[] state;
+        private MemoryStream message = new MemoryStream();
+        private readonly int DIGEST_BYTES;
+        private readonly int SPARKLE_STEPS_SLIM;
+        private readonly int SPARKLE_STEPS_BIG;
+        private readonly int STATE_BRANS;
+        private readonly int STATE_WORDS;
+        private readonly int RATE_WORDS;
+        private readonly int RATE_BYTES;
+
+        public SparkleDigest(SparkleParameters sparkleParameters)
+        {
+            int ESCH_DIGEST_LEN;
+            int SPARKLE_STATE;
+            int SPARKLE_RATE = 128;
+            switch (sparkleParameters)
+            {
+                case SparkleParameters.ESCH256:
+                    ESCH_DIGEST_LEN = 256;
+                    SPARKLE_STATE = 384;
+                    SPARKLE_STEPS_SLIM = 7;
+                    SPARKLE_STEPS_BIG = 11;
+                    break;
+                case SparkleParameters.ESCH384:
+                    ESCH_DIGEST_LEN = 384;
+                    SPARKLE_STATE = 512;
+                    SPARKLE_STEPS_SLIM = 8;
+                    SPARKLE_STEPS_BIG = 12;
+                    break;
+                default:
+                    throw new ArgumentException("Invalid definition of SCHWAEMM instance");
+            }
+            STATE_BRANS = SPARKLE_STATE >> 6;
+            STATE_WORDS = SPARKLE_STATE >> 5;
+            RATE_WORDS = SPARKLE_RATE >> 5;
+            RATE_BYTES = SPARKLE_RATE >> 3;
+            DIGEST_BYTES = ESCH_DIGEST_LEN >> 3;
+            state = new uint[STATE_WORDS];
+        }
+
+        private uint ROT(uint x, int n)
+        {
+            return (((x) >> n) | ((x) << (32 - n)));
+        }
+
+        private uint ELL(uint x)
+        {
+            return ROT(((x) ^ ((x) << 16)), 16);
+        }
+
+        private static readonly uint[] RCON = {0xB7E15162, 0xBF715880, 0x38B4DA56, 0x324E7738, 0xBB1185EB, 0x4F7C7B57,
+        0xCFBFA1C8, 0xC2B3293D};
+
+        void sparkle_opt(uint[] state, int brans, int steps)
+        {
+            uint i, j, rc, tmpx, tmpy, x0, y0;
+            for (i = 0; i < steps; i++)
+            {
+                // Add round ant
+                state[1] ^= RCON[i & 7];
+                state[3] ^= i;
+                // ARXBOX layer
+                for (j = 0; j < 2 * brans; j += 2)
+                {
+                    rc = RCON[j >> 1];
+                    state[j] += ROT(state[j + 1], 31);
+                    state[j + 1] ^= ROT(state[j], 24);
+                    state[j] ^= rc;
+                    state[j] += ROT(state[j + 1], 17);
+                    state[j + 1] ^= ROT(state[j], 17);
+                    state[j] ^= rc;
+                    state[j] += state[j + 1];
+                    state[j + 1] ^= ROT(state[j], 31);
+                    state[j] ^= rc;
+                    state[j] += ROT(state[j + 1], 24);
+                    state[j + 1] ^= ROT(state[j], 16);
+                    state[j] ^= rc;
+                }
+                // Linear layer
+                tmpx = x0 = state[0];
+                tmpy = y0 = state[1];
+                for (j = 2; j < brans; j += 2)
+                {
+                    tmpx ^= state[j];
+                    tmpy ^= state[j + 1];
+                }
+                tmpx = ELL(tmpx);
+                tmpy = ELL(tmpy);
+                for (j = 2; j < brans; j += 2)
+                {
+                    state[j - 2] = state[j + brans] ^ state[j] ^ tmpy;
+                    state[j + brans] = state[j];
+                    state[j - 1] = state[j + brans + 1] ^ state[j + 1] ^ tmpx;
+                    state[j + brans + 1] = state[j + 1];
+                }
+                state[brans - 2] = state[brans] ^ x0 ^ tmpy;
+                state[brans] = x0;
+                state[brans - 1] = state[brans + 1] ^ y0 ^ tmpx;
+                state[brans + 1] = y0;
+            }
+        }
+
+        public int GetDigestSize()
+        {
+            return DIGEST_BYTES;
+        }
+
+
+        public void BlockUpdate(byte[] input, int inOff, int len)
+        {
+            message.Write(input, inOff, len);
+        }
+
+
+        public int DoFinal(byte[] output, int outOff)
+        {
+            byte[] input = message.GetBuffer();
+            int inlen = (int)message.Length, i, inOff = 0;
+            uint tmpx, tmpy;
+            // Main Hashing Loop
+            uint[] in32 = Pack.LE_To_UInt32(input, 0, inlen >> 2);
+            while (inlen > RATE_BYTES)
+            {
+                // addition of a buffer block to the state
+                tmpx = 0;
+                tmpy = 0;
+                for (i = 0; i < RATE_WORDS; i += 2)
+                {
+                    tmpx ^= in32[i + (inOff >> 2)];
+                    tmpy ^= in32[i + 1 + (inOff >> 2)];
+                }
+                tmpx = ELL(tmpx);
+                tmpy = ELL(tmpy);
+                for (i = 0; i < RATE_WORDS; i += 2)
+                {
+                    state[i] ^= (in32[i + (inOff >> 2)] ^ tmpy);
+                    state[i + 1] ^= (in32[i + 1 + (inOff >> 2)] ^ tmpx);
+                }
+                for (i = RATE_WORDS; i < (STATE_WORDS / 2); i += 2)
+                {
+                    state[i] ^= tmpy;
+                    state[i + 1] ^= tmpx;
+                }
+                // execute SPARKLE with slim number of steps
+                sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_SLIM);
+                inlen -= RATE_BYTES;
+                inOff += RATE_BYTES;
+            }
+            // Hashing of Last Block
+            // addition of constant M1 or M2 to the state
+            state[STATE_BRANS - 1] ^= ((inlen < RATE_BYTES) ? (1u << 24) : (1u << 25));
+            // addition of last msg block (incl. padding)
+            uint[] buffer = new uint[RATE_WORDS];
+            for (i = 0; i < inlen; ++i)
+            {
+                buffer[i >> 2] |= (input[inOff++] & 0xffu) << ((i & 3) << 3);
+            }
+            if (inlen < RATE_BYTES)
+            {  // padding
+                buffer[i >> 2] |= 0x80u << ((i & 3) << 3);
+            }
+            tmpx = 0;
+            tmpy = 0;
+            for (i = 0; i < RATE_WORDS; i += 2)
+            {
+                tmpx ^= buffer[i];
+                tmpy ^= buffer[i + 1];
+            }
+            tmpx = ELL(tmpx);
+            tmpy = ELL(tmpy);
+            for (i = 0; i < RATE_WORDS; i += 2)
+            {
+                state[i] ^= (buffer[i] ^ tmpy);
+                state[i + 1] ^= (buffer[i + 1] ^ tmpx);
+            }
+            for (i = RATE_WORDS; i < (STATE_WORDS / 2); i += 2)
+            {
+                state[i] ^= tmpy;
+                state[i + 1] ^= tmpx;
+            }
+            // execute SPARKLE with big number of steps
+            sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_BIG);
+            Pack.UInt32_To_LE(state, 0, RATE_WORDS, output, outOff);
+            int outlen = RATE_BYTES;
+            outOff += RATE_BYTES;
+            while (outlen < DIGEST_BYTES)
+            {
+                sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_SLIM);
+                Pack.UInt32_To_LE(state, 0, RATE_WORDS, output, outOff);
+                outlen += RATE_BYTES;
+                outOff += RATE_BYTES;
+            }
+            return DIGEST_BYTES;
+        }
+
+        public String AlgorithmName => "Sparkle Hash";
+
+
+        public void Update(byte input)
+        {
+            message.Write(new byte[] { input }, 0, 1);
+        }
+
+        public void Reset()
+        {
+            message.SetLength(0);
+            Arrays.Fill(state, (byte)0);
+        }
+
+        public int GetByteLength()
+        {
+            return STATE_WORDS;
+        }
+
+#if NETCOREAPP2_1_OR_GREATER || NETSTANDARD2_1_OR_GREATER
+        public void BlockUpdate(ReadOnlySpan<byte> input)
+        {
+            message.Write(input);
+        }
+
+        public int DoFinal(Span<byte> output)
+        {
+            byte[] rv=new byte[DIGEST_BYTES];
+            DoFinal(rv, 0);
+            Reset();
+            rv.AsSpan(0, rv.Length).CopyTo(output);
+            return DIGEST_BYTES;
+        }
+#endif
+
+    }
+}
+
diff --git a/crypto/src/crypto/engines/SparkleEngine.cs b/crypto/src/crypto/engines/SparkleEngine.cs
new file mode 100644
index 000000000..5322ee464
--- /dev/null
+++ b/crypto/src/crypto/engines/SparkleEngine.cs
@@ -0,0 +1,483 @@
+using System;
+using System.IO;
+using Org.BouncyCastle.Crypto.Modes;
+using Org.BouncyCastle.Crypto.Parameters;
+using Org.BouncyCastle.Crypto.Utilities;
+using Org.BouncyCastle.Utilities;
+
+/**
+ * Sparkle v1.2, based on the current round 3 submission, https://sparkle-lwc.github.io/
+ * Reference C implementation: https://github.com/cryptolu/sparkle
+ * Specification: https://csrc.nist.gov/CSRC/media/Projects/lightweight-cryptography/documents/finalist-round/updated-spec-doc/sparkle-spec-final.pdf
+ */
+
+namespace Org.BouncyCastle.Crypto.Engines
+{
+    public class SparkleEngine : IAeadCipher
+    {
+        public enum SparkleParameters
+        {
+            SCHWAEMM128_128,
+            SCHWAEMM256_128,
+            SCHWAEMM192_192,
+            SCHWAEMM256_256
+        }
+        private readonly uint[] state;
+        private readonly uint[] k;
+        private readonly uint[] npub;
+        private readonly byte[] tag;
+        private bool encrypted;
+        private bool aadFinished;
+        private readonly MemoryStream aadData = new MemoryStream();
+        private readonly int SCHWAEMM_KEY_LEN;
+        private readonly int SCHWAEMM_NONCE_LEN;
+        private readonly int SPARKLE_STEPS_SLIM;
+        private readonly int SPARKLE_STEPS_BIG;
+        private readonly int KEY_WORDS;
+        private readonly int TAG_WORDS;
+        private readonly int TAG_BYTES;
+        private readonly int STATE_BRANS;
+        private readonly int STATE_WORDS;
+        private readonly int RATE_WORDS;
+        private readonly int RATE_BYTES;
+        private readonly int CAP_WORDS;
+        private readonly uint _A0;
+        private readonly uint _A1;
+        private readonly uint _M2;
+        private readonly uint _M3;
+
+        public SparkleEngine(SparkleParameters sparkleParameters)
+        {
+            int SPARKLE_STATE;
+            int SCHWAEMM_TAG_LEN;
+            int SPARKLE_CAPACITY;
+            switch (sparkleParameters)
+            {
+                case SparkleParameters.SCHWAEMM128_128:
+                    SCHWAEMM_KEY_LEN = 128;
+                    SCHWAEMM_NONCE_LEN = 128;
+                    SCHWAEMM_TAG_LEN = 128;
+                    SPARKLE_STATE = 256;
+                    SPARKLE_CAPACITY = 128;
+                    SPARKLE_STEPS_SLIM = 7;
+                    SPARKLE_STEPS_BIG = 10;
+                    break;
+                case SparkleParameters.SCHWAEMM256_128:
+                    SCHWAEMM_KEY_LEN = 128;
+                    SCHWAEMM_NONCE_LEN = 256;
+                    SCHWAEMM_TAG_LEN = 128;
+                    SPARKLE_STATE = 384;
+                    SPARKLE_CAPACITY = 128;
+                    SPARKLE_STEPS_SLIM = 7;
+                    SPARKLE_STEPS_BIG = 11;
+                    break;
+                case SparkleParameters.SCHWAEMM192_192:
+                    SCHWAEMM_KEY_LEN = 192;
+                    SCHWAEMM_NONCE_LEN = 192;
+                    SCHWAEMM_TAG_LEN = 192;
+                    SPARKLE_STATE = 384;
+                    SPARKLE_CAPACITY = 192;
+                    SPARKLE_STEPS_SLIM = 7;
+                    SPARKLE_STEPS_BIG = 11;
+                    break;
+                case SparkleParameters.SCHWAEMM256_256:
+                    SCHWAEMM_KEY_LEN = 256;
+                    SCHWAEMM_NONCE_LEN = 256;
+                    SCHWAEMM_TAG_LEN = 256;
+                    SPARKLE_STATE = 512;
+                    SPARKLE_CAPACITY = 256;
+                    SPARKLE_STEPS_SLIM = 8;
+                    SPARKLE_STEPS_BIG = 12;
+                    break;
+                default:
+                    throw new ArgumentException("Invalid definition of SCHWAEMM instance");
+            }
+            KEY_WORDS = SCHWAEMM_KEY_LEN >> 5;
+            TAG_WORDS = SCHWAEMM_TAG_LEN >> 5;
+            TAG_BYTES = SCHWAEMM_TAG_LEN >> 3;
+            STATE_BRANS = SPARKLE_STATE >> 6;
+            STATE_WORDS = SPARKLE_STATE >> 5;
+            RATE_WORDS = SCHWAEMM_NONCE_LEN >> 5;
+            RATE_BYTES = SCHWAEMM_NONCE_LEN >> 3;
+            int CAP_BRANS = SPARKLE_CAPACITY >> 6;
+            CAP_WORDS = SPARKLE_CAPACITY >> 5;
+            _A0 = ((((1u << CAP_BRANS))) << 24);
+            _A1 = (((1u ^ (1u << CAP_BRANS))) << 24);
+            _M2 = (((2u ^ (1u << CAP_BRANS))) << 24);
+            _M3 = (((3u ^ (1u << CAP_BRANS))) << 24);
+            state = new uint[STATE_WORDS];
+            tag = new byte[TAG_BYTES];
+            k = new uint[KEY_WORDS];
+            npub = new uint[RATE_WORDS];
+        }
+
+        private uint ROT(uint x, int n)
+        {
+            return (((x) >> n) | ((x) << (32 - n)));
+        }
+
+        private uint ELL(uint x)
+        {
+            return ROT(((x) ^ ((x) << 16)), 16);
+        }
+
+        private static readonly uint[] RCON = {0xB7E15162, 0xBF715880, 0x38B4DA56, 0x324E7738, 0xBB1185EB, 0x4F7C7B57,
+        0xCFBFA1C8, 0xC2B3293D};
+
+        void sparkle_opt(uint[] state, int brans, int steps)
+        {
+            uint i, j, rc, tmpx, tmpy, x0, y0;
+            for (i = 0; i < steps; i++)
+            {
+                // Add round ant
+                state[1] ^= RCON[i & 7];
+                state[3] ^= i;
+                // ARXBOX layer
+                for (j = 0; j < 2 * brans; j += 2)
+                {
+                    rc = RCON[j >> 1];
+                    state[j] += ROT(state[j + 1], 31);
+                    state[j + 1] ^= ROT(state[j], 24);
+                    state[j] ^= rc;
+                    state[j] += ROT(state[j + 1], 17);
+                    state[j + 1] ^= ROT(state[j], 17);
+                    state[j] ^= rc;
+                    state[j] += state[j + 1];
+                    state[j + 1] ^= ROT(state[j], 31);
+                    state[j] ^= rc;
+                    state[j] += ROT(state[j + 1], 24);
+                    state[j + 1] ^= ROT(state[j], 16);
+                    state[j] ^= rc;
+                }
+                // Linear layer
+                tmpx = x0 = state[0];
+                tmpy = y0 = state[1];
+                for (j = 2; j < brans; j += 2)
+                {
+                    tmpx ^= state[j];
+                    tmpy ^= state[j + 1];
+                }
+                tmpx = ELL(tmpx);
+                tmpy = ELL(tmpy);
+                for (j = 2; j < brans; j += 2)
+                {
+                    state[j - 2] = state[j + brans] ^ state[j] ^ tmpy;
+                    state[j + brans] = state[j];
+                    state[j - 1] = state[j + brans + 1] ^ state[j + 1] ^ tmpx;
+                    state[j + brans + 1] = state[j + 1];
+                }
+                state[brans - 2] = state[brans] ^ x0 ^ tmpy;
+                state[brans] = x0;
+                state[brans - 1] = state[brans + 1] ^ y0 ^ tmpx;
+                state[brans + 1] = y0;
+            }
+        }
+
+        private int CAP_INDEX(int i)
+        {
+            if (RATE_WORDS > CAP_WORDS)
+            {
+                return i & (CAP_WORDS - 1);
+            }
+            return i;
+        }
+
+        // The ProcessAssocData function absorbs the associated data, which becomes
+        // only authenticated but not encrypted, into the state (in blocks of size
+        // RATE_BYTES). Note that this function MUST NOT be called when the length of
+        // the associated data is 0.
+        void ProcessAssocData(uint[] state, byte[] input, int inlen)
+        {
+            // Main Authentication Loop
+            int inOff = 0, i, j;
+            uint tmp;
+            uint[] in32 = Pack.LE_To_UInt32(input, inOff, input.Length >> 2);
+            while (inlen > RATE_BYTES)
+            {
+                // combined Rho and rate-whitening operation
+                // Rho and rate-whitening for the authentication of associated data. The third
+                // parameter indicates whether the uint8_t-pointer 'in' is properly aligned to
+                // permit casting to a int-pointer. If this is the case then array 'in' is
+                // processed directly, otherwise it is first copied to an aligned buffer.
+                for (i = 0, j = RATE_WORDS / 2; i < RATE_WORDS / 2; i++, j++)
+                {
+                    tmp = state[i];
+                    state[i] = state[j] ^ in32[i + (inOff >> 2)] ^ state[RATE_WORDS + i];
+                    state[j] ^= tmp ^ in32[j + (inOff >> 2)] ^ state[RATE_WORDS + CAP_INDEX(j)];
+                }
+                // execute SPARKLE with slim number of steps
+                sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_SLIM);
+                inlen -= RATE_BYTES;
+                inOff += RATE_BYTES;
+            }
+            // Authentication of Last Block
+            // addition of ant A0 or A1 to the state
+            state[STATE_WORDS - 1] ^= ((inlen < RATE_BYTES) ? _A0 : _A1);
+            // combined Rho and rate-whitening (incl. padding)
+            // Rho and rate-whitening for the authentication of the last associated-data
+            // block. Since this last block may require padding, it is always copied to a buffer.
+            uint[] buffer = new uint[RATE_WORDS];
+            for (i = 0; i < inlen; ++i)
+            {
+                buffer[i >> 2] |= (uint)input[inOff++] << ((i & 3) << 3);
+            }
+            if (inlen < RATE_BYTES)
+            {  // padding
+                buffer[i >> 2] |= 0x80u << ((i & 3) << 3);
+            }
+            for (i = 0, j = RATE_WORDS / 2; i < RATE_WORDS / 2; i++, j++)
+            {
+                tmp = state[i];
+                state[i] = state[j] ^ buffer[i] ^ state[RATE_WORDS + i];
+                state[j] ^= tmp ^ buffer[j] ^ state[RATE_WORDS + CAP_INDEX(j)];
+            }
+            // execute SPARKLE with big number of steps
+            sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_BIG);
+        }
+
+        // The ProcessPlainText function encrypts the plaintext (input blocks of size
+        // RATE_BYTES) and generates the respective ciphertext. The uint8_t-array 'input'
+        // contains the plaintext and the ciphertext is written to uint8_t-array 'output'
+        // ('input' and 'output' can be the same array, i.e. they can have the same start
+        // address). Note that this function MUST NOT be called when the length of the
+        // plaintext is 0.
+        void ProcessPlainText(uint[] state, byte[] output, byte[] input, int inOff, int inlen)
+        {
+            // Main Encryption Loop
+            int outOff = 0, i, j;
+            uint tmp1, tmp2;
+            uint[] in32 = Pack.LE_To_UInt32(input, inOff, input.Length >> 2);
+            uint[] out32 = new uint[output.Length >> 2];
+            while (inlen > RATE_BYTES)
+            {
+                // combined Rho and rate-whitening operation
+                // Rho and rate-whitening for the encryption of plaintext. The third parameter
+                // indicates whether the uint8_t-pointers 'input' and 'output' are properly aligned
+                // to permit casting to int-pointers. If this is the case then array 'input'
+                // and 'output' are processed directly, otherwise 'input' is copied to an aligned buffer.
+                for (i = 0, j = RATE_WORDS / 2; i < RATE_WORDS / 2; i++, j++)
+                {
+                    tmp1 = state[i];
+                    tmp2 = state[j];
+                    state[i] = state[j] ^ in32[i + (inOff >> 2)] ^ state[RATE_WORDS + i];
+                    state[j] ^= tmp1 ^ in32[j + (inOff >> 2)] ^ state[RATE_WORDS + CAP_INDEX(j)];
+                    out32[i] = in32[i] ^ tmp1;
+                    out32[j] = in32[j] ^ tmp2;
+                }
+                Pack.UInt32_To_LE(out32, 0, RATE_WORDS, output, outOff);
+                // execute SPARKLE with slim number of steps
+                sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_SLIM);
+                inlen -= RATE_BYTES;
+                outOff += RATE_BYTES;
+                inOff += RATE_BYTES;
+            }
+            // Encryption of Last Block
+            // addition of ant M2 or M3 to the state
+            state[STATE_WORDS - 1] ^= ((inlen < RATE_BYTES) ? _M2 : _M3);
+            // combined Rho and rate-whitening (incl. padding)
+            // Rho and rate-whitening for the encryption of the last plaintext block. Since
+            // this last block may require padding, it is always copied to a buffer.
+            uint[] buffer = new uint[RATE_WORDS];
+            for (i = 0; i < inlen; ++i)
+            {
+                buffer[i >> 2] |= (input[inOff++] & 0xffu) << ((i & 3) << 3);
+            }
+            if (inlen < RATE_BYTES)
+            {  // padding
+                buffer[i >> 2] |= 0x80u << ((i & 3) << 3);
+            }
+            for (i = 0, j = RATE_WORDS / 2; i < RATE_WORDS / 2; i++, j++)
+            {
+                tmp1 = state[i];
+                tmp2 = state[j];
+                state[i] = state[j] ^ buffer[i] ^ state[RATE_WORDS + i];
+                state[j] ^= tmp1 ^ buffer[j] ^ state[RATE_WORDS + CAP_INDEX(j)];
+                buffer[i] ^= tmp1;
+                buffer[j] ^= tmp2;
+            }
+            for (i = 0; i < inlen; ++i)
+            {
+                output[outOff++] = (byte)(buffer[i >> 2] >> ((i & 3) << 3));
+            }
+            // execute SPARKLE with big number of steps
+            sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_BIG);
+        }
+
+        public void Init(bool forEncryption, ICipherParameters param)
+        {
+            /**
+             * Sparkle encryption and decryption is completely symmetrical, so the
+             * 'forEncryption' is irrelevant.
+             */
+            if (!(param is ParametersWithIV))
+            {
+                throw new ArgumentException(
+                    "Sparkle init parameters must include an IV");
+            }
+
+            ParametersWithIV ivParams = (ParametersWithIV)param;
+            byte[] iv = ivParams.GetIV();
+
+            if (iv == null || iv.Length != SCHWAEMM_NONCE_LEN >> 3)
+            {
+                throw new ArgumentException(
+                    "Sparkle requires exactly 16 bytes of IV");
+            }
+            Pack.LE_To_UInt32(iv, 0, npub, 0, RATE_WORDS);
+
+            if (!(ivParams.Parameters is KeyParameter))
+            {
+                throw new ArgumentException(
+                    "Sparkle init parameters must include a key");
+            }
+
+            KeyParameter key = (KeyParameter)ivParams.Parameters;
+            byte[] key8 = key.GetKey();
+            if (key8.Length != SCHWAEMM_KEY_LEN >> 3)
+            {
+                throw new ArgumentException("Sparkle key must be 128 bits long");
+            }
+            Pack.LE_To_UInt32(key8, 0, k, 0, KEY_WORDS);
+
+            reset(false);
+        }
+
+        public void ProcessAadByte(byte input)
+        {
+            aadData.Write(new byte[] { input }, 0, 1);
+        }
+
+
+        public void ProcessAadBytes(byte[] input, int inOff, int len)
+        {
+            aadData.Write(input, inOff, len);
+        }
+
+
+        public int ProcessByte(byte input, byte[] output, int outOff)
+        {
+            return ProcessBytes(new byte[] { input }, 0, 1, output, outOff);
+        }
+
+
+        public int ProcessBytes(byte[] input, int inOff, int len, byte[] output, int outOff)
+        {
+            if (encrypted)
+            {
+                throw new ArgumentException("Sparkle has processed encryption/decryption");
+            }
+            byte[] ad = aadData.GetBuffer();
+            int adsize = (int)aadData.Length;
+            if (adsize != 0)
+            {
+                ProcessAssocData(state, ad, adsize);
+            }
+            if (len != 0)
+            {
+                ProcessPlainText(state, output, input, inOff, len);
+            }
+            return len;
+        }
+
+
+        public int DoFinal(byte[] output, int outOff)
+        {
+            GetMac();
+            Array.Copy(tag, 0, output, outOff, TAG_BYTES);
+            reset(false);
+            return TAG_BYTES;
+        }
+
+        public byte[] GetMac()
+        {
+            if (!aadFinished)
+            {
+                // the key to the capacity part of the state.
+                uint[] buffer = new uint[TAG_WORDS];
+                // to prevent (potentially) unaligned memory accesses
+                Array.Copy(k, 0, buffer, 0, KEY_WORDS);
+                // add key to the capacity-part of the state
+                for (int i = 0; i < KEY_WORDS; i++)
+                {
+                    state[RATE_WORDS + i] ^= buffer[i];
+                }
+                aadFinished = true;
+            }
+            encrypted = true;
+            Pack.UInt32_To_LE(state, RATE_WORDS, TAG_WORDS, tag, 0);
+            return tag;
+        }
+
+
+        public int GetUpdateOutputSize(int len)
+        {
+            return len;
+        }
+
+
+        public int GetOutputSize(int len)
+        {
+            return len + TAG_BYTES;
+        }
+
+
+        public void Reset()
+        {
+            reset(true);
+        }
+
+        private void reset(bool clearMac)
+        {
+            if (clearMac)
+            {
+                Arrays.Fill(tag, (byte)0);
+            }
+            // The Initialize function loads nonce and key into the state and executes the
+            // SPARKLE permutation with the big number of steps.
+            // load nonce into the rate-part of the state
+            Array.Copy(npub, 0, state, 0, RATE_WORDS);
+            // load key into the capacity-part of the sate
+            Array.Copy(k, 0, state, RATE_WORDS, KEY_WORDS);
+            // execute SPARKLE with big number of steps
+            sparkle_opt(state, STATE_BRANS, SPARKLE_STEPS_BIG);
+            aadData.SetLength(0);
+            encrypted = false;
+            aadFinished = false;
+        }
+        public string AlgorithmName => "Sparkle AEAD";
+
+
+
+#if NETCOREAPP2_1_OR_GREATER || NETSTANDARD2_1_OR_GREATER
+        public void ProcessAadBytes(ReadOnlySpan<byte> input)
+        {
+            aadData.Write(input);
+        }
+
+        public int ProcessByte(byte input, Span<byte> output)
+        {
+            byte[] rv = new byte[1];
+            ProcessBytes(new byte[] { input }, 0, 1, rv, 0);
+            rv.AsSpan(0, 1).CopyTo(output);
+            return 1;
+        }
+
+        public int ProcessBytes(ReadOnlySpan<byte> input, Span<byte> output)
+        {
+            byte[] rv = new byte[input.Length];
+            ProcessBytes(input.ToArray(), 0, rv.Length, rv, 0);
+            rv.AsSpan(0, rv.Length).CopyTo(output);
+            return rv.Length;
+        }
+
+        public int DoFinal(Span<byte> output)
+        {
+            byte[] tag = GetMac();
+            tag.AsSpan(0, tag.Length).CopyTo(output);
+            reset(false);
+            return tag.Length;
+        }
+#endif
+    }
+}
+