diff --git a/CMakeLists.txt b/CMakeLists.txt
index 99bf31f1f..d4a236f3f 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -179,6 +179,7 @@ endif(ENABLE_ZLIB_SUPPORT)
 
 add_subdirectory(library)
 add_subdirectory(include)
+add_subdirectory(tinycrypt)
 
 if(ENABLE_PROGRAMS)
     add_subdirectory(programs)
diff --git a/doxygen/mbedtls.doxyfile b/doxygen/mbedtls.doxyfile
index 574db8d46..e30b18431 100644
--- a/doxygen/mbedtls.doxyfile
+++ b/doxygen/mbedtls.doxyfile
@@ -696,7 +696,9 @@ RECURSIVE              = YES
 # Note that relative paths are relative to the directory from which doxygen is
 # run.
 
-EXCLUDE                =
+EXCLUDE                = ./../include/tinycrypt/ecc.h \
+                         ./../include/tinycrypt/ecc_dh.h \
+                         ./../include/tinycrypt/ecc_dsa.h
 
 # The EXCLUDE_SYMLINKS tag can be used to select whether or not files or
 # directories that are symbolic links (a Unix file system feature) are excluded
diff --git a/include/CMakeLists.txt b/include/CMakeLists.txt
index 1b581a54d..862fa9298 100644
--- a/include/CMakeLists.txt
+++ b/include/CMakeLists.txt
@@ -3,6 +3,7 @@ option(INSTALL_MBEDTLS_HEADERS "Install mbed TLS headers." ON)
 if(INSTALL_MBEDTLS_HEADERS)
 
     file(GLOB headers "mbedtls/*.h")
+    file(GLOB headers "tinycrypt/*.h")
 
     install(FILES ${headers}
         DESTINATION include/mbedtls
diff --git a/include/mbedtls/check_config.h b/include/mbedtls/check_config.h
index b86e5807e..e73beac8d 100644
--- a/include/mbedtls/check_config.h
+++ b/include/mbedtls/check_config.h
@@ -87,6 +87,10 @@
 #error "MBEDTLS_CMAC_C defined, but not all prerequisites"
 #endif
 
+#if defined(MBEDTLS_USE_UECC) && defined(MBEDTLS_NO_64BIT_MULTIPLICATION)
+#error "MBEDTLS_USE_UECC defined, but it cannot be defined with MBEDTLS_NO_64BIT_MULTIPLICATION"
+#endif
+
 #if defined(MBEDTLS_NIST_KW_C) && \
     ( !defined(MBEDTLS_AES_C) || !defined(MBEDTLS_CIPHER_C) )
 #error "MBEDTLS_NIST_KW_C defined, but not all prerequisites"
diff --git a/include/mbedtls/config.h b/include/mbedtls/config.h
index 654f9725e..4d82f629d 100644
--- a/include/mbedtls/config.h
+++ b/include/mbedtls/config.h
@@ -2254,6 +2254,20 @@
  */
 #define MBEDTLS_ECP_C
 
+/**
+ * \def MBEDTLS_USE_UECC
+ *
+ * Enable the tinycrypt ECC library.
+ *
+ * Module:  tinycrypt/ecc.c
+ *          tinycrypt/ecc_dh.c
+ *          tinycrypt/ecc_dsa.c
+ *
+ * This module provides alternative ECC handling functions replacing
+ * native MBEDTLS ECP module.
+ */
+//#define MBEDTLS_USE_UECC
+
 /**
  * \def MBEDTLS_ENTROPY_C
  *
diff --git a/include/tinycrypt/ecc.h b/include/tinycrypt/ecc.h
new file mode 100644
index 000000000..ede754448
--- /dev/null
+++ b/include/tinycrypt/ecc.h
@@ -0,0 +1,547 @@
+/* ecc.h - TinyCrypt interface to common ECC functions */
+
+/* Copyright (c) 2014, Kenneth MacKay
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * * Redistributions of source code must retain the above copyright notice, this
+ *   list of conditions and the following disclaimer.
+ *
+ * * Redistributions in binary form must reproduce the above copyright notice,
+ *   this list of conditions and the following disclaimer in the documentation
+ *   and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ *  Copyright (C) 2017 by Intel Corporation, All Rights Reserved.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions are met:
+ *
+ *    - Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *
+ *    - Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ *    - Neither the name of Intel Corporation nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ *  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ *  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ *  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ *  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ *  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ *  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ *  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ *  POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/**
+ * @file
+ * @brief -- Interface to common ECC functions.
+ *
+ *  Overview: This software is an implementation of common functions
+ *            necessary to elliptic curve cryptography. This implementation uses
+ *            curve NIST p-256.
+ *
+ *  Security: The curve NIST p-256 provides approximately 128 bits of security.
+ *
+ */
+
+#if defined(MBEDTLS_USE_UECC)
+#ifndef __TC_UECC_H__
+#define __TC_UECC_H__
+
+#include <stdint.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* Word size (4 bytes considering 32-bits architectures) */
+#define uECC_WORD_SIZE 4
+
+/* setting max number of calls to prng: */
+#ifndef uECC_RNG_MAX_TRIES
+#define uECC_RNG_MAX_TRIES 64
+#endif
+
+/* defining data types to store word and bit counts: */
+typedef int8_t wordcount_t;
+typedef int16_t bitcount_t;
+/* defining data type for comparison result: */
+typedef int8_t cmpresult_t;
+/* defining data type to store ECC coordinate/point in 32bits words: */
+typedef unsigned int uECC_word_t;
+/* defining data type to store an ECC coordinate/point in 64bits words: */
+typedef uint64_t uECC_dword_t;
+
+/* defining masks useful for ecc computations: */
+#define HIGH_BIT_SET 0x80000000
+#define uECC_WORD_BITS 32
+#define uECC_WORD_BITS_SHIFT 5
+#define uECC_WORD_BITS_MASK 0x01F
+
+/* Number of words of 32 bits to represent an element of the the curve p-256: */
+#define NUM_ECC_WORDS 8
+/* Number of bytes to represent an element of the the curve p-256: */
+#define NUM_ECC_BYTES (uECC_WORD_SIZE*NUM_ECC_WORDS)
+
+/* structure that represents an elliptic curve (e.g. p256):*/
+struct uECC_Curve_t;
+typedef const struct uECC_Curve_t * uECC_Curve;
+struct uECC_Curve_t {
+  wordcount_t num_words;
+  wordcount_t num_bytes;
+  bitcount_t num_n_bits;
+  uECC_word_t p[NUM_ECC_WORDS];
+  uECC_word_t n[NUM_ECC_WORDS];
+  uECC_word_t G[NUM_ECC_WORDS * 2];
+  uECC_word_t b[NUM_ECC_WORDS];
+  void (*double_jacobian)(uECC_word_t * X1, uECC_word_t * Y1, uECC_word_t * Z1,
+	uECC_Curve curve);
+  void (*x_side)(uECC_word_t *result, const uECC_word_t *x, uECC_Curve curve);
+  void (*mmod_fast)(uECC_word_t *result, uECC_word_t *product);
+};
+
+/*
+ * @brief computes doubling of point ion jacobian coordinates, in place.
+ * @param X1 IN/OUT -- x coordinate
+ * @param Y1 IN/OUT -- y coordinate
+ * @param Z1 IN/OUT -- z coordinate
+ * @param curve IN -- elliptic curve
+ */
+void double_jacobian_default(uECC_word_t * X1, uECC_word_t * Y1,
+			     uECC_word_t * Z1, uECC_Curve curve);
+
+/*
+ * @brief Computes x^3 + ax + b. result must not overlap x.
+ * @param result OUT -- x^3 + ax + b
+ * @param x IN -- value of x
+ * @param curve IN -- elliptic curve
+ */
+void x_side_default(uECC_word_t *result, const uECC_word_t *x,
+		    uECC_Curve curve);
+
+/*
+ * @brief Computes result = product % curve_p
+ * from http://www.nsa.gov/ia/_files/nist-routines.pdf
+ * @param result OUT -- product % curve_p
+ * @param product IN -- value to be reduced mod curve_p
+ */
+void vli_mmod_fast_secp256r1(unsigned int *result, unsigned int *product);
+
+/* Bytes to words ordering: */
+#define BYTES_TO_WORDS_8(a, b, c, d, e, f, g, h) 0x##d##c##b##a, 0x##h##g##f##e
+#define BYTES_TO_WORDS_4(a, b, c, d) 0x##d##c##b##a
+#define BITS_TO_WORDS(num_bits) \
+	((num_bits + ((uECC_WORD_SIZE * 8) - 1)) / (uECC_WORD_SIZE * 8))
+#define BITS_TO_BYTES(num_bits) ((num_bits + 7) / 8)
+
+/* definition of curve NIST p-256: */
+static const struct uECC_Curve_t curve_secp256r1 = {
+	NUM_ECC_WORDS,
+	NUM_ECC_BYTES,
+	256, /* num_n_bits */ {
+		BYTES_TO_WORDS_8(FF, FF, FF, FF, FF, FF, FF, FF),
+		BYTES_TO_WORDS_8(FF, FF, FF, FF, 00, 00, 00, 00),
+        	BYTES_TO_WORDS_8(00, 00, 00, 00, 00, 00, 00, 00),
+        	BYTES_TO_WORDS_8(01, 00, 00, 00, FF, FF, FF, FF)
+	}, {
+		BYTES_TO_WORDS_8(51, 25, 63, FC, C2, CA, B9, F3),
+            	BYTES_TO_WORDS_8(84, 9E, 17, A7, AD, FA, E6, BC),
+            	BYTES_TO_WORDS_8(FF, FF, FF, FF, FF, FF, FF, FF),
+            	BYTES_TO_WORDS_8(00, 00, 00, 00, FF, FF, FF, FF)
+	}, {
+		BYTES_TO_WORDS_8(96, C2, 98, D8, 45, 39, A1, F4),
+                BYTES_TO_WORDS_8(A0, 33, EB, 2D, 81, 7D, 03, 77),
+                BYTES_TO_WORDS_8(F2, 40, A4, 63, E5, E6, BC, F8),
+                BYTES_TO_WORDS_8(47, 42, 2C, E1, F2, D1, 17, 6B),
+
+                BYTES_TO_WORDS_8(F5, 51, BF, 37, 68, 40, B6, CB),
+                BYTES_TO_WORDS_8(CE, 5E, 31, 6B, 57, 33, CE, 2B),
+                BYTES_TO_WORDS_8(16, 9E, 0F, 7C, 4A, EB, E7, 8E),
+                BYTES_TO_WORDS_8(9B, 7F, 1A, FE, E2, 42, E3, 4F)
+	}, {
+		BYTES_TO_WORDS_8(4B, 60, D2, 27, 3E, 3C, CE, 3B),
+                BYTES_TO_WORDS_8(F6, B0, 53, CC, B0, 06, 1D, 65),
+                BYTES_TO_WORDS_8(BC, 86, 98, 76, 55, BD, EB, B3),
+                BYTES_TO_WORDS_8(E7, 93, 3A, AA, D8, 35, C6, 5A)
+	},
+        &double_jacobian_default,
+        &x_side_default,
+        &vli_mmod_fast_secp256r1
+};
+
+uECC_Curve uECC_secp256r1(void);
+
+/*
+ * @brief Generates a random integer in the range 0 < random < top.
+ * Both random and top have num_words words.
+ * @param random OUT -- random integer in the range 0 < random < top
+ * @param top IN -- upper limit
+ * @param num_words IN -- number of words
+ * @return a random integer in the range 0 < random < top
+ */
+int uECC_generate_random_int(uECC_word_t *random, const uECC_word_t *top,
+			     wordcount_t num_words);
+
+
+/* uECC_RNG_Function type
+ * The RNG function should fill 'size' random bytes into 'dest'. It should
+ * return 1 if 'dest' was filled with random data, or 0 if the random data could
+ * not be generated. The filled-in values should be either truly random, or from
+ * a cryptographically-secure PRNG.
+ *
+ * A correctly functioning RNG function must be set (using uECC_set_rng())
+ * before calling uECC_make_key() or uECC_sign().
+ *
+ * Setting a correctly functioning RNG function improves the resistance to
+ * side-channel attacks for uECC_shared_secret().
+ *
+ * A correct RNG function is set by default. If you are building on another
+ * POSIX-compliant system that supports /dev/random or /dev/urandom, you can
+ * define uECC_POSIX to use the predefined RNG.
+ */
+typedef int(*uECC_RNG_Function)(uint8_t *dest, unsigned int size);
+
+/*
+ * @brief Set the function that will be used to generate random bytes. The RNG
+ * function should return 1 if the random data was generated, or 0 if the random
+ * data could not be generated.
+ *
+ * @note On platforms where there is no predefined RNG function, this must be
+ * called before uECC_make_key() or uECC_sign() are used.
+ *
+ * @param rng_function IN -- function that will be used to generate random bytes
+ */
+void uECC_set_rng(uECC_RNG_Function rng_function);
+
+/*
+ * @brief provides current uECC_RNG_Function.
+ * @return Returns the function that will be used to generate random bytes.
+ */
+uECC_RNG_Function uECC_get_rng(void);
+
+/*
+ * @brief computes the size of a private key for the curve in bytes.
+ * @param curve IN -- elliptic curve
+ * @return size of a private key for the curve in bytes.
+ */
+int uECC_curve_private_key_size(uECC_Curve curve);
+
+/*
+ * @brief computes the size of a public key for the curve in bytes.
+ * @param curve IN -- elliptic curve
+ * @return the size of a public key for the curve in bytes.
+ */
+int uECC_curve_public_key_size(uECC_Curve curve);
+
+/*
+ * @brief Compute the corresponding public key for a private key.
+ * @param private_key IN -- The private key to compute the public key for
+ * @param public_key OUT -- Will be filled in with the corresponding public key
+ * @param curve
+ * @return Returns 1 if key was computed successfully, 0 if an error occurred.
+ */
+int uECC_compute_public_key(const uint8_t *private_key,
+			    uint8_t *public_key, uECC_Curve curve);
+
+/*
+ * @brief Compute public-key.
+ * @return corresponding public-key.
+ * @param result OUT -- public-key
+ * @param private_key IN -- private-key
+ * @param curve IN -- elliptic curve
+ */
+uECC_word_t EccPoint_compute_public_key(uECC_word_t *result,
+					uECC_word_t *private_key, uECC_Curve curve);
+
+/*
+ * @brief Regularize the bitcount for the private key so that attackers cannot
+ * use a side channel attack to learn the number of leading zeros.
+ * @return Regularized k
+ * @param k IN -- private-key
+ * @param k0 IN/OUT -- regularized k
+ * @param k1 IN/OUT -- regularized k
+ * @param curve IN -- elliptic curve
+ */
+uECC_word_t regularize_k(const uECC_word_t * const k, uECC_word_t *k0,
+			 uECC_word_t *k1, uECC_Curve curve);
+
+/*
+ * @brief Point multiplication algorithm using Montgomery's ladder with co-Z
+ * coordinates. See http://eprint.iacr.org/2011/338.pdf.
+ * @note Result may overlap point.
+ * @param result OUT -- returns scalar*point
+ * @param point IN -- elliptic curve point
+ * @param scalar IN -- scalar
+ * @param initial_Z IN -- initial value for z
+ * @param num_bits IN -- number of bits in scalar
+ * @param curve IN -- elliptic curve
+ */
+void EccPoint_mult(uECC_word_t * result, const uECC_word_t * point,
+		   const uECC_word_t * scalar, const uECC_word_t * initial_Z,
+		   bitcount_t num_bits, uECC_Curve curve);
+
+/*
+ * @brief Constant-time comparison to zero - secure way to compare long integers
+ * @param vli IN -- very long integer
+ * @param num_words IN -- number of words in the vli
+ * @return 1 if vli == 0, 0 otherwise.
+ */
+uECC_word_t uECC_vli_isZero(const uECC_word_t *vli, wordcount_t num_words);
+
+/*
+ * @brief Check if 'point' is the point at infinity
+ * @param point IN -- elliptic curve point
+ * @param curve IN -- elliptic curve
+ * @return if 'point' is the point at infinity, 0 otherwise.
+ */
+uECC_word_t EccPoint_isZero(const uECC_word_t *point, uECC_Curve curve);
+
+/*
+ * @brief computes the sign of left - right, in constant time.
+ * @param left IN -- left term to be compared
+ * @param right IN -- right term to be compared
+ * @param num_words IN -- number of words
+ * @return the sign of left - right
+ */
+cmpresult_t uECC_vli_cmp(const uECC_word_t *left, const uECC_word_t *right,
+			 wordcount_t num_words);
+
+/*
+ * @brief computes sign of left - right, not in constant time.
+ * @note should not be used if inputs are part of a secret
+ * @param left IN -- left term to be compared
+ * @param right IN -- right term to be compared
+ * @param num_words IN -- number of words
+ * @return the sign of left - right
+ */
+cmpresult_t uECC_vli_cmp_unsafe(const uECC_word_t *left, const uECC_word_t *right,
+				wordcount_t num_words);
+
+/*
+ * @brief Computes result = (left - right) % mod.
+ * @note Assumes that (left < mod) and (right < mod), and that result does not
+ * overlap mod.
+ * @param result OUT -- (left - right) % mod
+ * @param left IN -- leftright term in modular subtraction
+ * @param right IN -- right term in modular subtraction
+ * @param mod IN -- mod
+ * @param num_words IN -- number of words
+ */
+void uECC_vli_modSub(uECC_word_t *result, const uECC_word_t *left,
+		     const uECC_word_t *right, const uECC_word_t *mod,
+		     wordcount_t num_words);
+
+/*
+ * @brief Computes P' = (x1', y1', Z3), P + Q = (x3, y3, Z3) or
+ * P => P', Q => P + Q
+ * @note assumes Input P = (x1, y1, Z), Q = (x2, y2, Z)
+ * @param X1 IN -- x coordinate of P
+ * @param Y1 IN -- y coordinate of P
+ * @param X2 IN -- x coordinate of Q
+ * @param Y2 IN -- y coordinate of Q
+ * @param curve IN -- elliptic curve
+ */
+void XYcZ_add(uECC_word_t * X1, uECC_word_t * Y1, uECC_word_t * X2,
+	      uECC_word_t * Y2, uECC_Curve curve);
+
+/*
+ * @brief Computes (x1 * z^2, y1 * z^3)
+ * @param X1 IN -- previous x1 coordinate
+ * @param Y1 IN -- previous y1 coordinate
+ * @param Z IN -- z value
+ * @param curve IN -- elliptic curve
+ */
+void apply_z(uECC_word_t * X1, uECC_word_t * Y1, const uECC_word_t * const Z,
+	     uECC_Curve curve);
+
+/*
+ * @brief Check if bit is set.
+ * @return Returns nonzero if bit 'bit' of vli is set.
+ * @warning It is assumed that the value provided in 'bit' is within the
+ * boundaries of the word-array 'vli'.
+ * @note The bit ordering layout assumed for vli is: {31, 30, ..., 0},
+ * {63, 62, ..., 32}, {95, 94, ..., 64}, {127, 126,..., 96} for a vli consisting
+ * of 4 uECC_word_t elements.
+ */
+uECC_word_t uECC_vli_testBit(const uECC_word_t *vli, bitcount_t bit);
+
+/*
+ * @brief Computes result = product % mod, where product is 2N words long.
+ * @param result OUT -- product % mod
+ * @param mod IN -- module
+ * @param num_words IN -- number of words
+ * @warning Currently only designed to work for curve_p or curve_n.
+ */
+void uECC_vli_mmod(uECC_word_t *result, uECC_word_t *product,
+		   const uECC_word_t *mod, wordcount_t num_words);
+
+/*
+ * @brief Computes modular product (using curve->mmod_fast)
+ * @param result OUT -- (left * right) mod % curve_p
+ * @param left IN -- left term in product
+ * @param right IN -- right term in product
+ * @param curve IN -- elliptic curve
+ */
+void uECC_vli_modMult_fast(uECC_word_t *result, const uECC_word_t *left,
+			   const uECC_word_t *right, uECC_Curve curve);
+
+/*
+ * @brief Computes result = left - right.
+ * @note Can modify in place.
+ * @param result OUT -- left - right
+ * @param left IN -- left term in subtraction
+ * @param right IN -- right term in subtraction
+ * @param num_words IN -- number of words
+ * @return borrow
+ */
+uECC_word_t uECC_vli_sub(uECC_word_t *result, const uECC_word_t *left,
+			 const uECC_word_t *right, wordcount_t num_words);
+
+/*
+ * @brief Constant-time comparison function(secure way to compare long ints)
+ * @param left IN -- left term in comparison
+ * @param right IN -- right term in comparison
+ * @param num_words IN -- number of words
+ * @return Returns 0 if left == right, 1 otherwise.
+ */
+uECC_word_t uECC_vli_equal(const uECC_word_t *left, const uECC_word_t *right,
+			   wordcount_t num_words);
+
+/*
+ * @brief Computes (left * right) % mod
+ * @param result OUT -- (left * right) % mod
+ * @param left IN -- left term in product
+ * @param right IN -- right term in product
+ * @param mod IN -- mod
+ * @param num_words IN -- number of words
+ */
+void uECC_vli_modMult(uECC_word_t *result, const uECC_word_t *left,
+		      const uECC_word_t *right, const uECC_word_t *mod,
+	              wordcount_t num_words);
+
+/*
+ * @brief Computes (1 / input) % mod
+ * @note All VLIs are the same size.
+ * @note See "Euclid's GCD to Montgomery Multiplication to the Great Divide"
+ * @param result OUT -- (1 / input) % mod
+ * @param input IN -- value to be modular inverted
+ * @param mod IN -- mod
+ * @param num_words -- number of words
+ */
+void uECC_vli_modInv(uECC_word_t *result, const uECC_word_t *input,
+		     const uECC_word_t *mod, wordcount_t num_words);
+
+/*
+ * @brief Sets dest = src.
+ * @param dest OUT -- destination buffer
+ * @param src IN --  origin buffer
+ * @param num_words IN -- number of words
+ */
+void uECC_vli_set(uECC_word_t *dest, const uECC_word_t *src,
+		  wordcount_t num_words);
+
+/*
+ * @brief Computes (left + right) % mod.
+ * @note Assumes that (left < mod) and right < mod), and that result does not
+ * overlap mod.
+ * @param result OUT -- (left + right) % mod.
+ * @param left IN -- left term in addition
+ * @param right IN -- right term in addition
+ * @param mod IN -- mod
+ * @param num_words IN -- number of words
+ */
+void uECC_vli_modAdd(uECC_word_t *result,  const uECC_word_t *left,
+    		     const uECC_word_t *right, const uECC_word_t *mod,
+   		     wordcount_t num_words);
+
+/*
+ * @brief Counts the number of bits required to represent vli.
+ * @param vli IN -- very long integer
+ * @param max_words IN -- number of words
+ * @return number of bits in given vli
+ */
+bitcount_t uECC_vli_numBits(const uECC_word_t *vli, 
+			    const wordcount_t max_words);
+
+/*
+ * @brief Erases (set to 0) vli
+ * @param vli IN -- very long integer
+ * @param num_words IN -- number of words
+ */
+void uECC_vli_clear(uECC_word_t *vli, wordcount_t num_words);
+
+/*
+ * @brief check if it is a valid point in the curve
+ * @param point IN -- point to be checked
+ * @param curve IN -- elliptic curve
+ * @return 0 if point is valid
+ * @exception returns -1 if it is a point at infinity
+ * @exception returns -2 if x or y is smaller than p,
+ * @exception returns -3 if y^2 != x^3 + ax + b.
+ */
+int uECC_valid_point(const uECC_word_t *point, uECC_Curve curve);
+
+/*
+ * @brief Check if a public key is valid.
+ * @param public_key IN -- The public key to be checked.
+ * @return returns 0 if the public key is valid
+ * @exception returns -1 if it is a point at infinity
+ * @exception returns -2 if x or y is smaller than p,
+ * @exception returns -3 if y^2 != x^3 + ax + b.
+ * @exception returns -4 if public key is the group generator.
+ *
+ * @note Note that you are not required to check for a valid public key before
+ * using any other uECC functions. However, you may wish to avoid spending CPU
+ * time computing a shared secret or verifying a signature using an invalid
+ * public key.
+ */
+int uECC_valid_public_key(const uint8_t *public_key, uECC_Curve curve);
+
+ /*
+  * @brief Converts an integer in uECC native format to big-endian bytes.
+  * @param bytes OUT -- bytes representation
+  * @param num_bytes IN -- number of bytes
+  * @param native IN -- uECC native representation
+  */
+void uECC_vli_nativeToBytes(uint8_t *bytes, int num_bytes,
+    			    const unsigned int *native);
+
+/*
+ * @brief Converts big-endian bytes to an integer in uECC native format.
+ * @param native OUT -- uECC native representation
+ * @param bytes IN -- bytes representation
+ * @param num_bytes IN -- number of bytes
+ */
+void uECC_vli_bytesToNative(unsigned int *native, const uint8_t *bytes,
+			    int num_bytes);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* __TC_UECC_H__ */
+#endif /* MBEDTLS_USE_UECC */
diff --git a/include/tinycrypt/ecc_dh.h b/include/tinycrypt/ecc_dh.h
new file mode 100644
index 000000000..18a4fd269
--- /dev/null
+++ b/include/tinycrypt/ecc_dh.h
@@ -0,0 +1,133 @@
+/* ecc_dh.h - TinyCrypt interface to EC-DH implementation */
+
+/*
+ * Copyright (c) 2014, Kenneth MacKay
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * * Redistributions of source code must retain the above copyright notice, this
+ *   list of conditions and the following disclaimer.
+ *
+ * * Redistributions in binary form must reproduce the above copyright notice,
+ *   this list of conditions and the following disclaimer in the documentation
+ *   and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/* Copyright (C) 2017 by Intel Corporation, All Rights Reserved.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions are met:
+ *
+ *    - Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *
+ *    - Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ *    - Neither the name of Intel Corporation nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ *  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ *  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ *  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ *  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ *  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ *  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ *  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ *  POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/**
+ * @file
+ * @brief -- Interface to EC-DH implementation.
+ *
+ *  Overview: This software is an implementation of EC-DH. This implementation
+ *            uses curve NIST p-256.
+ *
+ *  Security: The curve NIST p-256 provides approximately 128 bits of security.
+ */
+
+#if defined(MBEDTLS_USE_UECC)
+#ifndef __TC_ECC_DH_H__
+#define __TC_ECC_DH_H__
+
+#include <tinycrypt/ecc.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * @brief Create a public/private key pair.
+ * @return returns TC_CRYPTO_SUCCESS (1) if the key pair was generated successfully
+ *         returns TC_CRYPTO_FAIL (0) if error while generating key pair
+ *
+ * @param p_public_key OUT -- Will be filled in with the public key. Must be at
+ * least 2 * the curve size (in bytes) long. For curve secp256r1, p_public_key
+ * must be 64 bytes long.
+ * @param p_private_key OUT -- Will be filled in with the private key. Must be as
+ * long as the curve order (for secp256r1, p_private_key must be 32 bytes long).
+ *
+ * @note side-channel countermeasure: algorithm strengthened against timing
+ * attack.
+ * @warning A cryptographically-secure PRNG function must be set (using
+ * uECC_set_rng()) before calling uECC_make_key().
+ */
+int uECC_make_key(uint8_t *p_public_key, uint8_t *p_private_key, uECC_Curve curve);
+
+#ifdef ENABLE_TESTS
+
+/**
+ * @brief Create a public/private key pair given a specific d.
+ *
+ * @note THIS FUNCTION SHOULD BE CALLED ONLY FOR TEST PURPOSES. Refer to
+ * uECC_make_key() function for real applications.
+ */
+int uECC_make_key_with_d(uint8_t *p_public_key, uint8_t *p_private_key,
+    			 unsigned int *d, uECC_Curve curve);
+#endif
+
+/**
+ * @brief Compute a shared secret given your secret key and someone else's
+ * public key.
+ * @return returns TC_CRYPTO_SUCCESS (1) if the shared secret was computed successfully
+ *         returns TC_CRYPTO_FAIL (0) otherwise
+ *
+ * @param p_secret OUT -- Will be filled in with the shared secret value. Must be
+ * the same size as the curve size (for curve secp256r1, secret must be 32 bytes
+ * long.
+ * @param p_public_key IN -- The public key of the remote party.
+ * @param p_private_key IN -- Your private key.
+ *
+ * @warning It is recommended to use the output of uECC_shared_secret() as the
+ * input of a recommended Key Derivation Function (see NIST SP 800-108) in
+ * order to produce a cryptographically secure symmetric key.
+ */
+int uECC_shared_secret(const uint8_t *p_public_key, const uint8_t *p_private_key,
+		       uint8_t *p_secret, uECC_Curve curve);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* __TC_ECC_DH_H__ */
+#endif /* MBEDTLS_USE_UECC */
diff --git a/include/tinycrypt/ecc_dsa.h b/include/tinycrypt/ecc_dsa.h
new file mode 100644
index 000000000..806534001
--- /dev/null
+++ b/include/tinycrypt/ecc_dsa.h
@@ -0,0 +1,141 @@
+/* ecc_dh.h - TinyCrypt interface to EC-DSA implementation */
+
+/*
+ * Copyright (c) 2014, Kenneth MacKay
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * * Redistributions of source code must retain the above copyright notice, this
+ *   list of conditions and the following disclaimer.
+ *
+ * * Redistributions in binary form must reproduce the above copyright notice,
+ *   this list of conditions and the following disclaimer in the documentation
+ *   and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ * Copyright (C) 2017 by Intel Corporation, All Rights Reserved.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions are met:
+ *
+ *    - Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *
+ *    - Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ *    - Neither the name of Intel Corporation nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ *  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ *  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ *  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ *  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ *  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ *  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ *  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ *  POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/**
+ * @file
+ * @brief -- Interface to EC-DSA implementation.
+ *
+ *  Overview: This software is an implementation of EC-DSA. This implementation
+ *            uses curve NIST p-256.
+ *
+ *  Security: The curve NIST p-256 provides approximately 128 bits of security.
+ *
+ *  Usage:  - To sign: Compute a hash of the data you wish to sign (SHA-2 is
+ *          recommended) and pass it in to ecdsa_sign function along with your
+ *          private key and a random number. You must use a new non-predictable
+ *          random number to generate each new signature.
+ *          - To verify a signature: Compute the hash of the signed data using
+ *          the same hash as the signer and pass it to this function along with
+ *          the signer's public key and the signature values (r and s).
+ */
+
+#if defined(MBEDTLS_USE_UECC)
+#ifndef __TC_ECC_DSA_H__
+#define __TC_ECC_DSA_H__
+
+#include <tinycrypt/ecc.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * @brief Generate an ECDSA signature for a given hash value.
+ * @return returns TC_CRYPTO_SUCCESS (1) if the signature generated successfully
+ *         returns TC_CRYPTO_FAIL (0) if an error occurred.
+ *
+ * @param p_private_key IN -- Your private key.
+ * @param p_message_hash IN -- The hash of the message to sign.
+ * @param p_hash_size IN -- The size of p_message_hash in bytes.
+ * @param p_signature OUT -- Will be filled in with the signature value. Must be
+ * at least 2 * curve size long (for secp256r1, signature must be 64 bytes long).
+ *
+ * @warning A cryptographically-secure PRNG function must be set (using
+ * uECC_set_rng()) before calling uECC_sign().
+ * @note Usage: Compute a hash of the data you wish to sign (SHA-2 is
+ * recommended) and pass it in to this function along with your private key.
+ * @note side-channel countermeasure: algorithm strengthened against timing
+ * attack.
+ */
+int uECC_sign(const uint8_t *p_private_key, const uint8_t *p_message_hash,
+	      unsigned p_hash_size, uint8_t *p_signature, uECC_Curve curve);
+
+#ifdef ENABLE_TESTS
+/*
+ * THIS FUNCTION SHOULD BE CALLED FOR TEST PURPOSES ONLY.
+ * Refer to uECC_sign() function for real applications.
+ */
+int uECC_sign_with_k(const uint8_t *private_key, const uint8_t *message_hash,
+		     unsigned int hash_size, uECC_word_t *k, uint8_t *signature,
+		     uECC_Curve curve);
+#endif
+
+/**
+ * @brief Verify an ECDSA signature.
+ * @return returns TC_SUCCESS (1) if the signature is valid
+ * 	   returns TC_FAIL (0) if the signature is invalid.
+ *
+ * @param p_public_key IN -- The signer's public key.
+ * @param p_message_hash IN -- The hash of the signed data.
+ * @param p_hash_size IN -- The size of p_message_hash in bytes.
+ * @param p_signature IN -- The signature values.
+ *
+ * @note Usage: Compute the hash of the signed data using the same hash as the
+ * signer and pass it to this function along with the signer's public key and
+ * the signature values (hash_size and signature).
+ */
+int uECC_verify(const uint8_t *p_public_key, const uint8_t *p_message_hash,
+		unsigned int p_hash_size, const uint8_t *p_signature, uECC_Curve curve);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* __TC_ECC_DSA_H__ */
+#endif /* MBEDTLS_USE_UECC */
diff --git a/library/CMakeLists.txt b/library/CMakeLists.txt
index f3c804481..182e3a66b 100644
--- a/library/CMakeLists.txt
+++ b/library/CMakeLists.txt
@@ -131,6 +131,8 @@ if(LINK_WITH_PTHREAD)
     set(libs ${libs} pthread)
 endif()
 
+set(libs ${libs} tinycrypt)
+
 if (NOT USE_STATIC_MBEDTLS_LIBRARY AND NOT USE_SHARED_MBEDTLS_LIBRARY)
     message(FATAL_ERROR "Need to choose static or shared mbedtls build!")
 endif(NOT USE_STATIC_MBEDTLS_LIBRARY AND NOT USE_SHARED_MBEDTLS_LIBRARY)
diff --git a/library/Makefile b/library/Makefile
index 430c59881..d653aa0ef 100644
--- a/library/Makefile
+++ b/library/Makefile
@@ -63,6 +63,8 @@ DLEXT = dylib
 endif
 endif
 
+VPATH = ../tinycrypt
+
 OBJS_CRYPTO=	aes.o		aesni.o		arc4.o		\
 		aria.o		asn1parse.o	asn1write.o	\
 		base64.o	bignum.o	blowfish.o	\
@@ -84,7 +86,8 @@ OBJS_CRYPTO=	aes.o		aesni.o		arc4.o		\
 		ripemd160.o	rsa_internal.o	rsa.o  		\
 		sha1.o		sha256.o	sha512.o	\
 		threading.o	timing.o	version.o	\
-		version_features.o		xtea.o
+		version_features.o		xtea.o		\
+		ecc.o		ecc_dh.o	ecc_dsa.o
 
 OBJS_X509=	certs.o		pkcs11.o	x509.o		\
 		x509_create.o	x509_crl.o	x509_crt.o	\
diff --git a/library/version_features.c b/library/version_features.c
index 24143d052..7dbe10796 100644
--- a/library/version_features.c
+++ b/library/version_features.c
@@ -612,6 +612,9 @@ static const char *features[] = {
 #if defined(MBEDTLS_ECP_C)
     "MBEDTLS_ECP_C",
 #endif /* MBEDTLS_ECP_C */
+#if defined(MBEDTLS_USE_UECC)
+    "MBEDTLS_USE_UECC",
+#endif /* MBEDTLS_USE_UECC */
 #if defined(MBEDTLS_ENTROPY_C)
     "MBEDTLS_ENTROPY_C",
 #endif /* MBEDTLS_ENTROPY_C */
diff --git a/programs/ssl/query_config.c b/programs/ssl/query_config.c
index 6e281977e..266216f0f 100644
--- a/programs/ssl/query_config.c
+++ b/programs/ssl/query_config.c
@@ -1674,6 +1674,14 @@ int query_config( const char *config )
     }
 #endif /* MBEDTLS_ECP_C */
 
+#if defined(MBEDTLS_USE_UECC)
+    if( strcmp( "MBEDTLS_USE_UECC", config ) == 0 )
+    {
+        MACRO_EXPANSION_TO_STR( MBEDTLS_USE_UECC );
+        return( 0 );
+    }
+#endif /* MBEDTLS_USE_UECC */
+
 #if defined(MBEDTLS_ENTROPY_C)
     if( strcmp( "MBEDTLS_ENTROPY_C", config ) == 0 )
     {
diff --git a/scripts/config.pl b/scripts/config.pl
index 42ec6f81b..d5c2ed88c 100755
--- a/scripts/config.pl
+++ b/scripts/config.pl
@@ -98,6 +98,7 @@ MBEDTLS_ZLIB_SUPPORT
 MBEDTLS_PKCS11_C
 MBEDTLS_NO_UDBL_DIVISION
 MBEDTLS_NO_64BIT_MULTIPLICATION
+MBEDTLS_USE_UECC
 _ALT\s*$
 );
 
diff --git a/tests/scripts/all.sh b/tests/scripts/all.sh
index 53fbc7e25..d43d4784c 100755
--- a/tests/scripts/all.sh
+++ b/tests/scripts/all.sh
@@ -1030,6 +1030,19 @@ component_test_no_64bit_multiplication () {
     make test
 }
 
+component_build_uecc_cmake () {
+    msg "build: uecc native, cmake"
+    scripts/config.pl set MBEDTLS_USE_UECC
+    CC=gcc cmake .
+    make
+}
+
+component_build_uecc_make () {
+    msg "build: uecc native, make"
+    scripts/config.pl set MBEDTLS_USE_UECC
+    make CC=gcc CFLAGS='-Werror -O1'
+}
+
 component_build_arm_none_eabi_gcc () {
     msg "build: arm-none-eabi-gcc, make" # ~ 10s
     scripts/config.pl full
@@ -1125,6 +1138,15 @@ component_build_armcc () {
     armc6_build_test "--target=aarch64-arm-none-eabi -march=armv8.2-a"
 }
 
+component_build_armcc_uecc_baremetal () {
+    msg "build: ARM Compiler 5, make with uecc and baremetal"
+    scripts/config.pl baremetal
+    scripts/config.pl set MBEDTLS_USE_UECC
+
+    make CC="$ARMC5_CC" AR="$ARMC5_AR" WARNING_CFLAGS='--strict --c99' lib
+    make clean
+}
+
 component_test_allow_sha1 () {
     msg "build: allow SHA1 in certificates by default"
     scripts/config.pl set MBEDTLS_TLS_DEFAULT_ALLOW_SHA1_IN_CERTIFICATES
diff --git a/tests/scripts/check-files.py b/tests/scripts/check-files.py
index 005a077c7..2228f3573 100755
--- a/tests/scripts/check-files.py
+++ b/tests/scripts/check-files.py
@@ -181,7 +181,7 @@ class IntegrityChecker(object):
             ".c", ".h", ".sh", ".pl", ".py", ".md", ".function", ".data",
             "Makefile", "CMakeLists.txt", "ChangeLog"
         )
-        self.excluded_directories = ['.git', 'mbed-os']
+        self.excluded_directories = ['.git', 'mbed-os', 'tinycrypt']
         self.excluded_paths = list(map(os.path.normpath, [
             'cov-int',
             'examples',
diff --git a/tests/scripts/check-names.sh b/tests/scripts/check-names.sh
index f18a162cc..6d9ab0ea3 100755
--- a/tests/scripts/check-names.sh
+++ b/tests/scripts/check-names.sh
@@ -26,14 +26,21 @@ tests/scripts/list-symbols.sh
 FAIL=0
 
 printf "\nExported symbols declared in header: "
-UNDECLARED=$( diff exported-symbols identifiers | sed -n -e 's/^< //p' )
-if [ "x$UNDECLARED" = "x" ]; then
+UNDECLARED=$(diff exported-symbols identifiers | sed -n -e 's/^< //p') > undeclared
+
+FILTERED=$( diff tests/scripts/whitelist undeclared | sed -n -e 's/^< //p')
+
+if [ "x$UNDECLARED" != "x" ]; then
+if [ "x$FILTERED" = "x" ]; then
     echo "PASS"
 else
     echo "FAIL"
-    echo "$UNDECLARED"
+    echo "$FILTERED"
     FAIL=1
 fi
+else
+    echo "PASS"
+fi
 
 diff macros identifiers | sed -n -e 's/< //p' > actual-macros
 
@@ -84,7 +91,7 @@ fi
 
 printf "\nOverall: "
 if [ "$FAIL" -eq 0 ]; then
-    rm macros actual-macros enum-consts identifiers exported-symbols
+    rm macros actual-macros enum-consts identifiers exported-symbols undeclared
     echo "PASSED"
     exit 0
 else
diff --git a/tests/scripts/whitelist b/tests/scripts/whitelist
new file mode 100644
index 000000000..99c406096
--- /dev/null
+++ b/tests/scripts/whitelist
@@ -0,0 +1,38 @@
+EccPoint_compute_public_key
+EccPoint_isZero
+EccPoint_mult
+regularize_k
+uECC_compute_public_key
+uECC_curve_private_key_size
+uECC_curve_public_key_size
+uECC_generate_random_int
+uECC_get_rng
+uECC_make_key
+uECC_make_key_with_d
+uECC_secp256r1
+uECC_set_rng
+uECC_shared_secret
+uECC_sign
+uECC_sign_with_k
+uECC_valid_point
+uECC_valid_public_key
+uECC_verify
+uECC_vli_bytesToNative
+uECC_vli_clear
+uECC_vli_cmp
+uECC_vli_cmp_unsafe
+uECC_vli_equal
+uECC_vli_isZero
+uECC_vli_mmod
+uECC_vli_modAdd
+uECC_vli_modInv
+uECC_vli_modMult
+uECC_vli_modMult_fast
+uECC_vli_modSub
+uECC_vli_nativeToBytes
+uECC_vli_numBits
+uECC_vli_set
+uECC_vli_sub
+uECC_vli_testBit
+vli_mmod_fast_secp256r1
+x_side_default
diff --git a/tinycrypt/CMakeLists.txt b/tinycrypt/CMakeLists.txt
new file mode 100644
index 000000000..7674d83be
--- /dev/null
+++ b/tinycrypt/CMakeLists.txt
@@ -0,0 +1,7 @@
+set(src_tinycrypt
+    ecc_dh.c
+    ecc_dsa.c
+    ecc.c
+)
+
+add_library(tinycrypt STATIC ${src_tinycrypt})
diff --git a/tinycrypt/ecc.c b/tinycrypt/ecc.c
new file mode 100644
index 000000000..090f691e9
--- /dev/null
+++ b/tinycrypt/ecc.c
@@ -0,0 +1,943 @@
+/* ecc.c - TinyCrypt implementation of common ECC functions */
+
+/*
+ * Copyright (c) 2014, Kenneth MacKay
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright notice,
+ * this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright notice,
+ * this list of conditions and the following disclaimer in the documentation
+ * and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ * ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ *  Copyright (C) 2017 by Intel Corporation, All Rights Reserved.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions are met:
+ *
+ *    - Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *
+ *    - Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ *    - Neither the name of Intel Corporation nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ *  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ *  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ *  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ *  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ *  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ *  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ *  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ *  POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#if defined(MBEDTLS_USE_UECC)
+#include <tinycrypt/ecc.h>
+#include <string.h>
+
+/* IMPORTANT: Make sure a cryptographically-secure PRNG is set and the platform
+ * has access to enough entropy in order to feed the PRNG regularly. */
+#if default_RNG_defined
+static uECC_RNG_Function g_rng_function = &default_CSPRNG;
+#else
+static uECC_RNG_Function g_rng_function = 0;
+#endif
+
+void uECC_set_rng(uECC_RNG_Function rng_function)
+{
+	g_rng_function = rng_function;
+}
+
+uECC_RNG_Function uECC_get_rng(void)
+{
+	return g_rng_function;
+}
+
+int uECC_curve_private_key_size(uECC_Curve curve)
+{
+	return BITS_TO_BYTES(curve->num_n_bits);
+}
+
+int uECC_curve_public_key_size(uECC_Curve curve)
+{
+	return 2 * curve->num_bytes;
+}
+
+void uECC_vli_clear(uECC_word_t *vli, wordcount_t num_words)
+{
+	wordcount_t i;
+	for (i = 0; i < num_words; ++i) {
+		 vli[i] = 0;
+	}
+}
+
+uECC_word_t uECC_vli_isZero(const uECC_word_t *vli, wordcount_t num_words)
+{
+	uECC_word_t bits = 0;
+	wordcount_t i;
+	for (i = 0; i < num_words; ++i) {
+		bits |= vli[i];
+	}
+	return (bits == 0);
+}
+
+uECC_word_t uECC_vli_testBit(const uECC_word_t *vli, bitcount_t bit)
+{
+	return (vli[bit >> uECC_WORD_BITS_SHIFT] &
+		((uECC_word_t)1 << (bit & uECC_WORD_BITS_MASK)));
+}
+
+/* Counts the number of words in vli. */
+static wordcount_t vli_numDigits(const uECC_word_t *vli,
+				 const wordcount_t max_words)
+{
+
+	wordcount_t i;
+	/* Search from the end until we find a non-zero digit. We do it in reverse
+	 * because we expect that most digits will be nonzero. */
+	for (i = max_words - 1; i >= 0 && vli[i] == 0; --i) {
+	}
+
+	return (i + 1);
+}
+
+bitcount_t uECC_vli_numBits(const uECC_word_t *vli,
+			    const wordcount_t max_words)
+{
+
+	uECC_word_t i;
+	uECC_word_t digit;
+
+	wordcount_t num_digits = vli_numDigits(vli, max_words);
+	if (num_digits == 0) {
+		return 0;
+	}
+
+	digit = vli[num_digits - 1];
+	for (i = 0; digit; ++i) {
+		digit >>= 1;
+	}
+
+	return (((bitcount_t)(num_digits - 1) << uECC_WORD_BITS_SHIFT) + i);
+}
+
+void uECC_vli_set(uECC_word_t *dest, const uECC_word_t *src,
+		  wordcount_t num_words)
+{
+	wordcount_t i;
+
+	for (i = 0; i < num_words; ++i) {
+		dest[i] = src[i];
+  	}
+}
+
+cmpresult_t uECC_vli_cmp_unsafe(const uECC_word_t *left,
+				const uECC_word_t *right,
+				wordcount_t num_words)
+{
+	wordcount_t i;
+
+	for (i = num_words - 1; i >= 0; --i) {
+		if (left[i] > right[i]) {
+			return 1;
+		} else if (left[i] < right[i]) {
+			return -1;
+		}
+	}
+	return 0;
+}
+
+uECC_word_t uECC_vli_equal(const uECC_word_t *left, const uECC_word_t *right,
+			   wordcount_t num_words)
+{
+
+	uECC_word_t diff = 0;
+	wordcount_t i;
+
+	for (i = num_words - 1; i >= 0; --i) {
+		diff |= (left[i] ^ right[i]);
+	}
+	return !(diff == 0);
+}
+
+uECC_word_t cond_set(uECC_word_t p_true, uECC_word_t p_false, unsigned int cond)
+{
+	return (p_true*(cond)) | (p_false*(!cond));
+}
+
+/* Computes result = left - right, returning borrow, in constant time.
+ * Can modify in place. */
+uECC_word_t uECC_vli_sub(uECC_word_t *result, const uECC_word_t *left,
+			 const uECC_word_t *right, wordcount_t num_words)
+{
+	uECC_word_t borrow = 0;
+	wordcount_t i;
+	for (i = 0; i < num_words; ++i) {
+		uECC_word_t diff = left[i] - right[i] - borrow;
+		uECC_word_t val = (diff > left[i]);
+		borrow = cond_set(val, borrow, (diff != left[i]));
+
+		result[i] = diff;
+	}
+	return borrow;
+}
+
+/* Computes result = left + right, returning carry, in constant time.
+ * Can modify in place. */
+static uECC_word_t uECC_vli_add(uECC_word_t *result, const uECC_word_t *left,
+				const uECC_word_t *right, wordcount_t num_words)
+{
+	uECC_word_t carry = 0;
+	wordcount_t i;
+	for (i = 0; i < num_words; ++i) {
+		uECC_word_t sum = left[i] + right[i] + carry;
+		uECC_word_t val = (sum < left[i]);
+		carry = cond_set(val, carry, (sum != left[i]));
+		result[i] = sum;
+	}
+	return carry;
+}
+
+cmpresult_t uECC_vli_cmp(const uECC_word_t *left, const uECC_word_t *right,
+			 wordcount_t num_words)
+{
+	uECC_word_t tmp[NUM_ECC_WORDS];
+	uECC_word_t neg = !!uECC_vli_sub(tmp, left, right, num_words);
+	uECC_word_t equal = uECC_vli_isZero(tmp, num_words);
+	return (!equal - 2 * neg);
+}
+
+/* Computes vli = vli >> 1. */
+static void uECC_vli_rshift1(uECC_word_t *vli, wordcount_t num_words)
+{
+	uECC_word_t *end = vli;
+	uECC_word_t carry = 0;
+
+	vli += num_words;
+	while (vli-- > end) {
+		uECC_word_t temp = *vli;
+		*vli = (temp >> 1) | carry;
+		carry = temp << (uECC_WORD_BITS - 1);
+	}
+}
+
+static void muladd(uECC_word_t a, uECC_word_t b, uECC_word_t *r0,
+		   uECC_word_t *r1, uECC_word_t *r2)
+{
+
+	uECC_dword_t p = (uECC_dword_t)a * b;
+	uECC_dword_t r01 = ((uECC_dword_t)(*r1) << uECC_WORD_BITS) | *r0;
+	r01 += p;
+	*r2 += (r01 < p);
+	*r1 = r01 >> uECC_WORD_BITS;
+	*r0 = (uECC_word_t)r01;
+
+}
+
+/* Computes result = left * right. Result must be 2 * num_words long. */
+static void uECC_vli_mult(uECC_word_t *result, const uECC_word_t *left,
+			  const uECC_word_t *right, wordcount_t num_words)
+{
+
+	uECC_word_t r0 = 0;
+	uECC_word_t r1 = 0;
+	uECC_word_t r2 = 0;
+	wordcount_t i, k;
+
+	/* Compute each digit of result in sequence, maintaining the carries. */
+	for (k = 0; k < num_words; ++k) {
+
+		for (i = 0; i <= k; ++i) {
+			muladd(left[i], right[k - i], &r0, &r1, &r2);
+		}
+
+		result[k] = r0;
+		r0 = r1;
+		r1 = r2;
+		r2 = 0;
+	}
+
+	for (k = num_words; k < num_words * 2 - 1; ++k) {
+
+		for (i = (k + 1) - num_words; i < num_words; ++i) {
+			muladd(left[i], right[k - i], &r0, &r1, &r2);
+		}
+		result[k] = r0;
+		r0 = r1;
+		r1 = r2;
+		r2 = 0;
+	}
+	result[num_words * 2 - 1] = r0;
+}
+
+void uECC_vli_modAdd(uECC_word_t *result, const uECC_word_t *left,
+		     const uECC_word_t *right, const uECC_word_t *mod,
+		     wordcount_t num_words)
+{
+	uECC_word_t carry = uECC_vli_add(result, left, right, num_words);
+	if (carry || uECC_vli_cmp_unsafe(mod, result, num_words) != 1) {
+	/* result > mod (result = mod + remainder), so subtract mod to get
+	 * remainder. */
+		uECC_vli_sub(result, result, mod, num_words);
+	}
+}
+
+void uECC_vli_modSub(uECC_word_t *result, const uECC_word_t *left,
+		     const uECC_word_t *right, const uECC_word_t *mod,
+		     wordcount_t num_words)
+{
+	uECC_word_t l_borrow = uECC_vli_sub(result, left, right, num_words);
+	if (l_borrow) {
+		/* In this case, result == -diff == (max int) - diff. Since -x % d == d - x,
+		 * we can get the correct result from result + mod (with overflow). */
+		uECC_vli_add(result, result, mod, num_words);
+	}
+}
+
+/* Computes result = product % mod, where product is 2N words long. */
+/* Currently only designed to work for curve_p or curve_n. */
+void uECC_vli_mmod(uECC_word_t *result, uECC_word_t *product,
+    		   const uECC_word_t *mod, wordcount_t num_words)
+{
+	uECC_word_t mod_multiple[2 * NUM_ECC_WORDS];
+	uECC_word_t tmp[2 * NUM_ECC_WORDS];
+	uECC_word_t *v[2] = {tmp, product};
+	uECC_word_t index;
+
+	/* Shift mod so its highest set bit is at the maximum position. */
+	bitcount_t shift = (num_words * 2 * uECC_WORD_BITS) -
+			   uECC_vli_numBits(mod, num_words);
+	wordcount_t word_shift = shift / uECC_WORD_BITS;
+	wordcount_t bit_shift = shift % uECC_WORD_BITS;
+	uECC_word_t carry = 0;
+	uECC_vli_clear(mod_multiple, word_shift);
+	if (bit_shift > 0) {
+		for(index = 0; index < (uECC_word_t)num_words; ++index) {
+			mod_multiple[word_shift + index] = (mod[index] << bit_shift) | carry;
+			carry = mod[index] >> (uECC_WORD_BITS - bit_shift);
+		}
+	} else {
+		uECC_vli_set(mod_multiple + word_shift, mod, num_words);
+	}
+
+	for (index = 1; shift >= 0; --shift) {
+		uECC_word_t borrow = 0;
+		wordcount_t i;
+		for (i = 0; i < num_words * 2; ++i) {
+			uECC_word_t diff = v[index][i] - mod_multiple[i] - borrow;
+			if (diff != v[index][i]) {
+				borrow = (diff > v[index][i]);
+			}
+			v[1 - index][i] = diff;
+		}
+		/* Swap the index if there was no borrow */
+		index = !(index ^ borrow);
+		uECC_vli_rshift1(mod_multiple, num_words);
+		mod_multiple[num_words - 1] |= mod_multiple[num_words] <<
+					       (uECC_WORD_BITS - 1);
+		uECC_vli_rshift1(mod_multiple + num_words, num_words);
+	}
+	uECC_vli_set(result, v[index], num_words);
+}
+
+void uECC_vli_modMult(uECC_word_t *result, const uECC_word_t *left,
+		      const uECC_word_t *right, const uECC_word_t *mod,
+		      wordcount_t num_words)
+{
+	uECC_word_t product[2 * NUM_ECC_WORDS];
+	uECC_vli_mult(product, left, right, num_words);
+	uECC_vli_mmod(result, product, mod, num_words);
+}
+
+void uECC_vli_modMult_fast(uECC_word_t *result, const uECC_word_t *left,
+			   const uECC_word_t *right, uECC_Curve curve)
+{
+	uECC_word_t product[2 * NUM_ECC_WORDS];
+	uECC_vli_mult(product, left, right, curve->num_words);
+
+	curve->mmod_fast(result, product);
+}
+
+static void uECC_vli_modSquare_fast(uECC_word_t *result,
+				    const uECC_word_t *left,
+				    uECC_Curve curve)
+{
+	uECC_vli_modMult_fast(result, left, left, curve);
+}
+
+
+#define EVEN(vli) (!(vli[0] & 1))
+
+static void vli_modInv_update(uECC_word_t *uv,
+			      const uECC_word_t *mod,
+			      wordcount_t num_words)
+{
+
+	uECC_word_t carry = 0;
+
+	if (!EVEN(uv)) {
+		carry = uECC_vli_add(uv, uv, mod, num_words);
+	}
+	uECC_vli_rshift1(uv, num_words);
+	if (carry) {
+		uv[num_words - 1] |= HIGH_BIT_SET;
+	}
+}
+
+void uECC_vli_modInv(uECC_word_t *result, const uECC_word_t *input,
+		     const uECC_word_t *mod, wordcount_t num_words)
+{
+	uECC_word_t a[NUM_ECC_WORDS], b[NUM_ECC_WORDS];
+	uECC_word_t u[NUM_ECC_WORDS], v[NUM_ECC_WORDS];
+	cmpresult_t cmpResult;
+
+	if (uECC_vli_isZero(input, num_words)) {
+		uECC_vli_clear(result, num_words);
+		return;
+	}
+
+	uECC_vli_set(a, input, num_words);
+	uECC_vli_set(b, mod, num_words);
+	uECC_vli_clear(u, num_words);
+	u[0] = 1;
+	uECC_vli_clear(v, num_words);
+	while ((cmpResult = uECC_vli_cmp_unsafe(a, b, num_words)) != 0) {
+		if (EVEN(a)) {
+			uECC_vli_rshift1(a, num_words);
+      			vli_modInv_update(u, mod, num_words);
+    		} else if (EVEN(b)) {
+			uECC_vli_rshift1(b, num_words);
+			vli_modInv_update(v, mod, num_words);
+		} else if (cmpResult > 0) {
+			uECC_vli_sub(a, a, b, num_words);
+			uECC_vli_rshift1(a, num_words);
+			if (uECC_vli_cmp_unsafe(u, v, num_words) < 0) {
+        			uECC_vli_add(u, u, mod, num_words);
+      			}
+      			uECC_vli_sub(u, u, v, num_words);
+      			vli_modInv_update(u, mod, num_words);
+    		} else {
+      			uECC_vli_sub(b, b, a, num_words);
+      			uECC_vli_rshift1(b, num_words);
+      			if (uECC_vli_cmp_unsafe(v, u, num_words) < 0) {
+        			uECC_vli_add(v, v, mod, num_words);
+      			}
+      			uECC_vli_sub(v, v, u, num_words);
+      			vli_modInv_update(v, mod, num_words);
+    		}
+  	}
+  	uECC_vli_set(result, u, num_words);
+}
+
+/* ------ Point operations ------ */
+
+void double_jacobian_default(uECC_word_t * X1, uECC_word_t * Y1,
+			     uECC_word_t * Z1, uECC_Curve curve)
+{
+	/* t1 = X, t2 = Y, t3 = Z */
+	uECC_word_t t4[NUM_ECC_WORDS];
+	uECC_word_t t5[NUM_ECC_WORDS];
+	wordcount_t num_words = curve->num_words;
+
+	if (uECC_vli_isZero(Z1, num_words)) {
+		return;
+	}
+
+	uECC_vli_modSquare_fast(t4, Y1, curve);   /* t4 = y1^2 */
+	uECC_vli_modMult_fast(t5, X1, t4, curve); /* t5 = x1*y1^2 = A */
+	uECC_vli_modSquare_fast(t4, t4, curve);   /* t4 = y1^4 */
+	uECC_vli_modMult_fast(Y1, Y1, Z1, curve); /* t2 = y1*z1 = z3 */
+	uECC_vli_modSquare_fast(Z1, Z1, curve);   /* t3 = z1^2 */
+
+	uECC_vli_modAdd(X1, X1, Z1, curve->p, num_words); /* t1 = x1 + z1^2 */
+	uECC_vli_modAdd(Z1, Z1, Z1, curve->p, num_words); /* t3 = 2*z1^2 */
+	uECC_vli_modSub(Z1, X1, Z1, curve->p, num_words); /* t3 = x1 - z1^2 */
+	uECC_vli_modMult_fast(X1, X1, Z1, curve); /* t1 = x1^2 - z1^4 */
+
+	uECC_vli_modAdd(Z1, X1, X1, curve->p, num_words); /* t3 = 2*(x1^2 - z1^4) */
+	uECC_vli_modAdd(X1, X1, Z1, curve->p, num_words); /* t1 = 3*(x1^2 - z1^4) */
+	if (uECC_vli_testBit(X1, 0)) {
+		uECC_word_t l_carry = uECC_vli_add(X1, X1, curve->p, num_words);
+		uECC_vli_rshift1(X1, num_words);
+		X1[num_words - 1] |= l_carry << (uECC_WORD_BITS - 1);
+	} else {
+		uECC_vli_rshift1(X1, num_words);
+	}
+
+	/* t1 = 3/2*(x1^2 - z1^4) = B */
+	uECC_vli_modSquare_fast(Z1, X1, curve); /* t3 = B^2 */
+	uECC_vli_modSub(Z1, Z1, t5, curve->p, num_words); /* t3 = B^2 - A */
+	uECC_vli_modSub(Z1, Z1, t5, curve->p, num_words); /* t3 = B^2 - 2A = x3 */
+	uECC_vli_modSub(t5, t5, Z1, curve->p, num_words); /* t5 = A - x3 */
+	uECC_vli_modMult_fast(X1, X1, t5, curve); /* t1 = B * (A - x3) */
+	/* t4 = B * (A - x3) - y1^4 = y3: */
+	uECC_vli_modSub(t4, X1, t4, curve->p, num_words);
+
+	uECC_vli_set(X1, Z1, num_words);
+	uECC_vli_set(Z1, Y1, num_words);
+	uECC_vli_set(Y1, t4, num_words);
+}
+
+void x_side_default(uECC_word_t *result,
+		    const uECC_word_t *x,
+		    uECC_Curve curve)
+{
+	uECC_word_t _3[NUM_ECC_WORDS] = {3}; /* -a = 3 */
+	wordcount_t num_words = curve->num_words;
+
+	uECC_vli_modSquare_fast(result, x, curve); /* r = x^2 */
+	uECC_vli_modSub(result, result, _3, curve->p, num_words); /* r = x^2 - 3 */
+	uECC_vli_modMult_fast(result, result, x, curve); /* r = x^3 - 3x */
+	/* r = x^3 - 3x + b: */
+	uECC_vli_modAdd(result, result, curve->b, curve->p, num_words);
+}
+
+uECC_Curve uECC_secp256r1(void)
+{
+	return &curve_secp256r1;
+}
+
+void vli_mmod_fast_secp256r1(unsigned int *result, unsigned int*product)
+{
+	unsigned int tmp[NUM_ECC_WORDS];
+	int carry;
+
+	/* t */
+	uECC_vli_set(result, product, NUM_ECC_WORDS);
+
+	/* s1 */
+	tmp[0] = tmp[1] = tmp[2] = 0;
+	tmp[3] = product[11];
+	tmp[4] = product[12];
+	tmp[5] = product[13];
+	tmp[6] = product[14];
+	tmp[7] = product[15];
+	carry = uECC_vli_add(tmp, tmp, tmp, NUM_ECC_WORDS);
+	carry += uECC_vli_add(result, result, tmp, NUM_ECC_WORDS);
+
+	/* s2 */
+	tmp[3] = product[12];
+	tmp[4] = product[13];
+	tmp[5] = product[14];
+	tmp[6] = product[15];
+	tmp[7] = 0;
+	carry += uECC_vli_add(tmp, tmp, tmp, NUM_ECC_WORDS);
+	carry += uECC_vli_add(result, result, tmp, NUM_ECC_WORDS);
+
+	/* s3 */
+	tmp[0] = product[8];
+	tmp[1] = product[9];
+	tmp[2] = product[10];
+	tmp[3] = tmp[4] = tmp[5] = 0;
+	tmp[6] = product[14];
+	tmp[7] = product[15];
+  	carry += uECC_vli_add(result, result, tmp, NUM_ECC_WORDS);
+
+	/* s4 */
+	tmp[0] = product[9];
+	tmp[1] = product[10];
+	tmp[2] = product[11];
+	tmp[3] = product[13];
+	tmp[4] = product[14];
+	tmp[5] = product[15];
+	tmp[6] = product[13];
+	tmp[7] = product[8];
+	carry += uECC_vli_add(result, result, tmp, NUM_ECC_WORDS);
+
+	/* d1 */
+	tmp[0] = product[11];
+	tmp[1] = product[12];
+	tmp[2] = product[13];
+	tmp[3] = tmp[4] = tmp[5] = 0;
+	tmp[6] = product[8];
+	tmp[7] = product[10];
+	carry -= uECC_vli_sub(result, result, tmp, NUM_ECC_WORDS);
+
+	/* d2 */
+	tmp[0] = product[12];
+	tmp[1] = product[13];
+	tmp[2] = product[14];
+	tmp[3] = product[15];
+	tmp[4] = tmp[5] = 0;
+	tmp[6] = product[9];
+	tmp[7] = product[11];
+	carry -= uECC_vli_sub(result, result, tmp, NUM_ECC_WORDS);
+
+	/* d3 */
+	tmp[0] = product[13];
+	tmp[1] = product[14];
+	tmp[2] = product[15];
+	tmp[3] = product[8];
+	tmp[4] = product[9];
+	tmp[5] = product[10];
+	tmp[6] = 0;
+	tmp[7] = product[12];
+	carry -= uECC_vli_sub(result, result, tmp, NUM_ECC_WORDS);
+
+	/* d4 */
+	tmp[0] = product[14];
+	tmp[1] = product[15];
+	tmp[2] = 0;
+	tmp[3] = product[9];
+	tmp[4] = product[10];
+	tmp[5] = product[11];
+	tmp[6] = 0;
+	tmp[7] = product[13];
+	carry -= uECC_vli_sub(result, result, tmp, NUM_ECC_WORDS);
+
+	if (carry < 0) {
+		do {
+			carry += uECC_vli_add(result, result, curve_secp256r1.p, NUM_ECC_WORDS);
+		}
+		while (carry < 0);
+	} else  {
+		while (carry || 
+		       uECC_vli_cmp_unsafe(curve_secp256r1.p, result, NUM_ECC_WORDS) != 1) {
+			carry -= uECC_vli_sub(result, result, curve_secp256r1.p, NUM_ECC_WORDS);
+		}
+	}
+}
+
+uECC_word_t EccPoint_isZero(const uECC_word_t *point, uECC_Curve curve)
+{
+	return uECC_vli_isZero(point, curve->num_words * 2);
+}
+
+void apply_z(uECC_word_t * X1, uECC_word_t * Y1, const uECC_word_t * const Z,
+	     uECC_Curve curve)
+{
+	uECC_word_t t1[NUM_ECC_WORDS];
+
+	uECC_vli_modSquare_fast(t1, Z, curve);    /* z^2 */
+	uECC_vli_modMult_fast(X1, X1, t1, curve); /* x1 * z^2 */
+	uECC_vli_modMult_fast(t1, t1, Z, curve);  /* z^3 */
+	uECC_vli_modMult_fast(Y1, Y1, t1, curve); /* y1 * z^3 */
+}
+
+/* P = (x1, y1) => 2P, (x2, y2) => P' */
+static void XYcZ_initial_double(uECC_word_t * X1, uECC_word_t * Y1,
+				uECC_word_t * X2, uECC_word_t * Y2,
+				const uECC_word_t * const initial_Z,
+				uECC_Curve curve)
+{
+	uECC_word_t z[NUM_ECC_WORDS];
+	wordcount_t num_words = curve->num_words;
+	if (initial_Z) {
+		uECC_vli_set(z, initial_Z, num_words);
+	} else {
+		uECC_vli_clear(z, num_words);
+		z[0] = 1;
+	}
+
+	uECC_vli_set(X2, X1, num_words);
+	uECC_vli_set(Y2, Y1, num_words);
+
+	apply_z(X1, Y1, z, curve);
+	curve->double_jacobian(X1, Y1, z, curve);
+	apply_z(X2, Y2, z, curve);
+}
+
+void XYcZ_add(uECC_word_t * X1, uECC_word_t * Y1,
+	      uECC_word_t * X2, uECC_word_t * Y2,
+	      uECC_Curve curve)
+{
+	/* t1 = X1, t2 = Y1, t3 = X2, t4 = Y2 */
+	uECC_word_t t5[NUM_ECC_WORDS];
+	wordcount_t num_words = curve->num_words;
+
+	uECC_vli_modSub(t5, X2, X1, curve->p, num_words); /* t5 = x2 - x1 */
+	uECC_vli_modSquare_fast(t5, t5, curve); /* t5 = (x2 - x1)^2 = A */
+	uECC_vli_modMult_fast(X1, X1, t5, curve); /* t1 = x1*A = B */
+	uECC_vli_modMult_fast(X2, X2, t5, curve); /* t3 = x2*A = C */
+	uECC_vli_modSub(Y2, Y2, Y1, curve->p, num_words); /* t4 = y2 - y1 */
+	uECC_vli_modSquare_fast(t5, Y2, curve); /* t5 = (y2 - y1)^2 = D */
+
+	uECC_vli_modSub(t5, t5, X1, curve->p, num_words); /* t5 = D - B */
+	uECC_vli_modSub(t5, t5, X2, curve->p, num_words); /* t5 = D - B - C = x3 */
+	uECC_vli_modSub(X2, X2, X1, curve->p, num_words); /* t3 = C - B */
+	uECC_vli_modMult_fast(Y1, Y1, X2, curve); /* t2 = y1*(C - B) */
+	uECC_vli_modSub(X2, X1, t5, curve->p, num_words); /* t3 = B - x3 */
+	uECC_vli_modMult_fast(Y2, Y2, X2, curve); /* t4 = (y2 - y1)*(B - x3) */
+	uECC_vli_modSub(Y2, Y2, Y1, curve->p, num_words); /* t4 = y3 */
+
+	uECC_vli_set(X2, t5, num_words);
+}
+
+/* Input P = (x1, y1, Z), Q = (x2, y2, Z)
+   Output P + Q = (x3, y3, Z3), P - Q = (x3', y3', Z3)
+   or P => P - Q, Q => P + Q
+ */
+static void XYcZ_addC(uECC_word_t * X1, uECC_word_t * Y1,
+		      uECC_word_t * X2, uECC_word_t * Y2,
+		      uECC_Curve curve)
+{
+	/* t1 = X1, t2 = Y1, t3 = X2, t4 = Y2 */
+	uECC_word_t t5[NUM_ECC_WORDS];
+	uECC_word_t t6[NUM_ECC_WORDS];
+	uECC_word_t t7[NUM_ECC_WORDS];
+	wordcount_t num_words = curve->num_words;
+
+	uECC_vli_modSub(t5, X2, X1, curve->p, num_words); /* t5 = x2 - x1 */
+	uECC_vli_modSquare_fast(t5, t5, curve); /* t5 = (x2 - x1)^2 = A */
+	uECC_vli_modMult_fast(X1, X1, t5, curve); /* t1 = x1*A = B */
+	uECC_vli_modMult_fast(X2, X2, t5, curve); /* t3 = x2*A = C */
+	uECC_vli_modAdd(t5, Y2, Y1, curve->p, num_words); /* t5 = y2 + y1 */
+	uECC_vli_modSub(Y2, Y2, Y1, curve->p, num_words); /* t4 = y2 - y1 */
+
+	uECC_vli_modSub(t6, X2, X1, curve->p, num_words); /* t6 = C - B */
+	uECC_vli_modMult_fast(Y1, Y1, t6, curve); /* t2 = y1 * (C - B) = E */
+	uECC_vli_modAdd(t6, X1, X2, curve->p, num_words); /* t6 = B + C */
+	uECC_vli_modSquare_fast(X2, Y2, curve); /* t3 = (y2 - y1)^2 = D */
+	uECC_vli_modSub(X2, X2, t6, curve->p, num_words); /* t3 = D - (B + C) = x3 */
+
+	uECC_vli_modSub(t7, X1, X2, curve->p, num_words); /* t7 = B - x3 */
+	uECC_vli_modMult_fast(Y2, Y2, t7, curve); /* t4 = (y2 - y1)*(B - x3) */
+	/* t4 = (y2 - y1)*(B - x3) - E = y3: */
+	uECC_vli_modSub(Y2, Y2, Y1, curve->p, num_words);
+
+	uECC_vli_modSquare_fast(t7, t5, curve); /* t7 = (y2 + y1)^2 = F */
+	uECC_vli_modSub(t7, t7, t6, curve->p, num_words); /* t7 = F - (B + C) = x3' */
+	uECC_vli_modSub(t6, t7, X1, curve->p, num_words); /* t6 = x3' - B */
+	uECC_vli_modMult_fast(t6, t6, t5, curve); /* t6 = (y2+y1)*(x3' - B) */
+	/* t2 = (y2+y1)*(x3' - B) - E = y3': */
+	uECC_vli_modSub(Y1, t6, Y1, curve->p, num_words);
+
+	uECC_vli_set(X1, t7, num_words);
+}
+
+void EccPoint_mult(uECC_word_t * result, const uECC_word_t * point,
+		   const uECC_word_t * scalar,
+		   const uECC_word_t * initial_Z,
+		   bitcount_t num_bits, uECC_Curve curve) 
+{
+	/* R0 and R1 */
+	uECC_word_t Rx[2][NUM_ECC_WORDS];
+	uECC_word_t Ry[2][NUM_ECC_WORDS];
+	uECC_word_t z[NUM_ECC_WORDS];
+	bitcount_t i;
+	uECC_word_t nb;
+	wordcount_t num_words = curve->num_words;
+
+	uECC_vli_set(Rx[1], point, num_words);
+  	uECC_vli_set(Ry[1], point + num_words, num_words);
+
+	XYcZ_initial_double(Rx[1], Ry[1], Rx[0], Ry[0], initial_Z, curve);
+
+	for (i = num_bits - 2; i > 0; --i) {
+		nb = !uECC_vli_testBit(scalar, i);
+		XYcZ_addC(Rx[1 - nb], Ry[1 - nb], Rx[nb], Ry[nb], curve);
+		XYcZ_add(Rx[nb], Ry[nb], Rx[1 - nb], Ry[1 - nb], curve);
+	}
+
+	nb = !uECC_vli_testBit(scalar, 0);
+	XYcZ_addC(Rx[1 - nb], Ry[1 - nb], Rx[nb], Ry[nb], curve);
+
+	/* Find final 1/Z value. */
+	uECC_vli_modSub(z, Rx[1], Rx[0], curve->p, num_words); /* X1 - X0 */
+	uECC_vli_modMult_fast(z, z, Ry[1 - nb], curve); /* Yb * (X1 - X0) */
+	uECC_vli_modMult_fast(z, z, point, curve); /* xP * Yb * (X1 - X0) */
+	uECC_vli_modInv(z, z, curve->p, num_words); /* 1 / (xP * Yb * (X1 - X0))*/
+	/* yP / (xP * Yb * (X1 - X0)) */
+	uECC_vli_modMult_fast(z, z, point + num_words, curve);
+	/* Xb * yP / (xP * Yb * (X1 - X0)) */
+	uECC_vli_modMult_fast(z, z, Rx[1 - nb], curve);
+	/* End 1/Z calculation */
+
+	XYcZ_add(Rx[nb], Ry[nb], Rx[1 - nb], Ry[1 - nb], curve);
+	apply_z(Rx[0], Ry[0], z, curve);
+
+	uECC_vli_set(result, Rx[0], num_words);
+	uECC_vli_set(result + num_words, Ry[0], num_words);
+}
+
+uECC_word_t regularize_k(const uECC_word_t * const k, uECC_word_t *k0,
+			 uECC_word_t *k1, uECC_Curve curve)
+{
+
+	wordcount_t num_n_words = BITS_TO_WORDS(curve->num_n_bits);
+
+	bitcount_t num_n_bits = curve->num_n_bits;
+
+	uECC_word_t carry = uECC_vli_add(k0, k, curve->n, num_n_words) ||
+			     (num_n_bits < ((bitcount_t)num_n_words * uECC_WORD_SIZE * 8) &&
+			     uECC_vli_testBit(k0, num_n_bits));
+
+	uECC_vli_add(k1, k0, curve->n, num_n_words);
+
+	return carry;
+}
+
+uECC_word_t EccPoint_compute_public_key(uECC_word_t *result,
+					uECC_word_t *private_key,
+					uECC_Curve curve)
+{
+
+	uECC_word_t tmp1[NUM_ECC_WORDS];
+ 	uECC_word_t tmp2[NUM_ECC_WORDS];
+	uECC_word_t *p2[2] = {tmp1, tmp2};
+	uECC_word_t carry;
+
+	/* Regularize the bitcount for the private key so that attackers cannot
+	 * use a side channel attack to learn the number of leading zeros. */
+	carry = regularize_k(private_key, tmp1, tmp2, curve);
+
+	EccPoint_mult(result, curve->G, p2[!carry], 0, curve->num_n_bits + 1, curve);
+
+	if (EccPoint_isZero(result, curve)) {
+		return 0;
+	}
+	return 1;
+}
+
+/* Converts an integer in uECC native format to big-endian bytes. */
+void uECC_vli_nativeToBytes(uint8_t *bytes, int num_bytes,
+			    const unsigned int *native)
+{
+	wordcount_t i;
+	for (i = 0; i < num_bytes; ++i) {
+		unsigned b = num_bytes - 1 - i;
+		bytes[i] = native[b / uECC_WORD_SIZE] >> (8 * (b % uECC_WORD_SIZE));
+	}
+}
+
+/* Converts big-endian bytes to an integer in uECC native format. */
+void uECC_vli_bytesToNative(unsigned int *native, const uint8_t *bytes,
+			    int num_bytes)
+{
+	wordcount_t i;
+	uECC_vli_clear(native, (num_bytes + (uECC_WORD_SIZE - 1)) / uECC_WORD_SIZE);
+	for (i = 0; i < num_bytes; ++i) {
+		unsigned b = num_bytes - 1 - i;
+		native[b / uECC_WORD_SIZE] |=
+			(uECC_word_t)bytes[i] << (8 * (b % uECC_WORD_SIZE));
+  	}
+}
+
+int uECC_generate_random_int(uECC_word_t *random, const uECC_word_t *top,
+			     wordcount_t num_words)
+{
+	uECC_word_t mask = (uECC_word_t)-1;
+	uECC_word_t tries;
+	bitcount_t num_bits = uECC_vli_numBits(top, num_words);
+
+	if (!g_rng_function) {
+		return 0;
+	}
+
+	for (tries = 0; tries < uECC_RNG_MAX_TRIES; ++tries) {
+		if (!g_rng_function((uint8_t *)random, num_words * uECC_WORD_SIZE)) {
+      			return 0;
+    		}
+		random[num_words - 1] &=
+        		mask >> ((bitcount_t)(num_words * uECC_WORD_SIZE * 8 - num_bits));
+		if (!uECC_vli_isZero(random, num_words) &&
+			uECC_vli_cmp(top, random, num_words) == 1) {
+			return 1;
+		}
+	}
+	return 0;
+}
+
+
+int uECC_valid_point(const uECC_word_t *point, uECC_Curve curve)
+{
+	uECC_word_t tmp1[NUM_ECC_WORDS];
+	uECC_word_t tmp2[NUM_ECC_WORDS];
+	wordcount_t num_words = curve->num_words;
+
+	/* The point at infinity is invalid. */
+	if (EccPoint_isZero(point, curve)) {
+		return -1;
+	}
+
+	/* x and y must be smaller than p. */
+	if (uECC_vli_cmp_unsafe(curve->p, point, num_words) != 1 ||
+		uECC_vli_cmp_unsafe(curve->p, point + num_words, num_words) != 1) {
+		return -2;
+	}
+
+	uECC_vli_modSquare_fast(tmp1, point + num_words, curve);
+	curve->x_side(tmp2, point, curve); /* tmp2 = x^3 + ax + b */
+
+	/* Make sure that y^2 == x^3 + ax + b */
+	if (uECC_vli_equal(tmp1, tmp2, num_words) != 0)
+		return -3;
+
+	return 0;
+}
+
+int uECC_valid_public_key(const uint8_t *public_key, uECC_Curve curve)
+{
+
+	uECC_word_t _public[NUM_ECC_WORDS * 2];
+
+	uECC_vli_bytesToNative(_public, public_key, curve->num_bytes);
+	uECC_vli_bytesToNative(
+	_public + curve->num_words,
+	public_key + curve->num_bytes,
+	curve->num_bytes);
+
+	if (uECC_vli_cmp_unsafe(_public, curve->G, NUM_ECC_WORDS * 2) == 0) {
+		return -4;
+	}
+
+	return uECC_valid_point(_public, curve);
+}
+
+int uECC_compute_public_key(const uint8_t *private_key, uint8_t *public_key,
+			    uECC_Curve curve)
+{
+
+	uECC_word_t _private[NUM_ECC_WORDS];
+	uECC_word_t _public[NUM_ECC_WORDS * 2];
+
+	uECC_vli_bytesToNative(
+	_private,
+	private_key,
+	BITS_TO_BYTES(curve->num_n_bits));
+
+	/* Make sure the private key is in the range [1, n-1]. */
+	if (uECC_vli_isZero(_private, BITS_TO_WORDS(curve->num_n_bits))) {
+		return 0;
+	}
+
+	if (uECC_vli_cmp(curve->n, _private, BITS_TO_WORDS(curve->num_n_bits)) != 1) {
+		return 0;
+	}
+
+	/* Compute public key. */
+	if (!EccPoint_compute_public_key(_public, _private, curve)) {
+		return 0;
+	}
+
+	uECC_vli_nativeToBytes(public_key, curve->num_bytes, _public);
+	uECC_vli_nativeToBytes(
+	public_key +
+	curve->num_bytes, curve->num_bytes, _public + curve->num_words);
+	return 1;
+}
+#else
+typedef int mbedtls_dummy_uecc_def;
+#endif /* MBEDTLS_USE_UECC */
+
diff --git a/tinycrypt/ecc_dh.c b/tinycrypt/ecc_dh.c
new file mode 100644
index 000000000..5d0a52f77
--- /dev/null
+++ b/tinycrypt/ecc_dh.c
@@ -0,0 +1,201 @@
+/* ec_dh.c - TinyCrypt implementation of EC-DH */
+
+/*
+ * Copyright (c) 2014, Kenneth MacKay
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *  * Redistributions of source code must retain the above copyright notice,
+ *    this list of conditions and the following disclaimer.
+ *  * Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ *  Copyright (C) 2017 by Intel Corporation, All Rights Reserved.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions are met:
+ *
+ *    - Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *
+ *    - Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ *    - Neither the name of Intel Corporation nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ *  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ *  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ *  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ *  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ *  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ *  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ *  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ *  POSSIBILITY OF SUCH DAMAGE.
+ */
+#if defined(MBEDTLS_USE_UECC)
+#include <tinycrypt/ecc.h>
+#include <tinycrypt/ecc_dh.h>
+#include <string.h>
+#include "mbedtls/platform_util.h"
+
+#if default_RNG_defined
+static uECC_RNG_Function g_rng_function = &default_CSPRNG;
+#else
+static uECC_RNG_Function g_rng_function = 0;
+#endif
+
+int uECC_make_key_with_d(uint8_t *public_key, uint8_t *private_key,
+			 unsigned int *d, uECC_Curve curve)
+{
+
+	uECC_word_t _private[NUM_ECC_WORDS];
+	uECC_word_t _public[NUM_ECC_WORDS * 2];
+
+	/* This function is designed for test purposes-only (such as validating NIST
+	 * test vectors) as it uses a provided value for d instead of generating
+	 * it uniformly at random. */
+	memcpy (_private, d, NUM_ECC_BYTES);
+
+	/* Computing public-key from private: */
+	if (EccPoint_compute_public_key(_public, _private, curve)) {
+
+		/* Converting buffers to correct bit order: */
+		uECC_vli_nativeToBytes(private_key,
+				       BITS_TO_BYTES(curve->num_n_bits),
+				       _private);
+		uECC_vli_nativeToBytes(public_key,
+				       curve->num_bytes,
+				       _public);
+		uECC_vli_nativeToBytes(public_key + curve->num_bytes,
+				       curve->num_bytes,
+				       _public + curve->num_words);
+
+		/* erasing temporary buffer used to store secret: */
+		memset(_private, 0, NUM_ECC_BYTES);
+
+		return 1;
+	}
+	return 0;
+}
+
+int uECC_make_key(uint8_t *public_key, uint8_t *private_key, uECC_Curve curve)
+{
+
+	uECC_word_t _random[NUM_ECC_WORDS * 2];
+	uECC_word_t _private[NUM_ECC_WORDS];
+	uECC_word_t _public[NUM_ECC_WORDS * 2];
+	uECC_word_t tries;
+
+	for (tries = 0; tries < uECC_RNG_MAX_TRIES; ++tries) {
+		/* Generating _private uniformly at random: */
+		uECC_RNG_Function rng_function = uECC_get_rng();
+		if (!rng_function ||
+			!rng_function((uint8_t *)_random, 2 * NUM_ECC_WORDS*uECC_WORD_SIZE)) {
+        		return 0;
+		}
+
+		/* computing modular reduction of _random (see FIPS 186.4 B.4.1): */
+		uECC_vli_mmod(_private, _random, curve->n, BITS_TO_WORDS(curve->num_n_bits));
+
+		/* Computing public-key from private: */
+		if (EccPoint_compute_public_key(_public, _private, curve)) {
+
+			/* Converting buffers to correct bit order: */
+			uECC_vli_nativeToBytes(private_key,
+					       BITS_TO_BYTES(curve->num_n_bits),
+					       _private);
+			uECC_vli_nativeToBytes(public_key,
+					       curve->num_bytes,
+					       _public);
+			uECC_vli_nativeToBytes(public_key + curve->num_bytes,
+ 					       curve->num_bytes,
+					       _public + curve->num_words);
+
+			/* erasing temporary buffer that stored secret: */
+			memset(_private, 0, NUM_ECC_BYTES);
+
+      			return 1;
+    		}
+  	}
+	return 0;
+}
+
+int uECC_shared_secret(const uint8_t *public_key, const uint8_t *private_key,
+		       uint8_t *secret, uECC_Curve curve)
+{
+
+	uECC_word_t _public[NUM_ECC_WORDS * 2];
+	uECC_word_t _private[NUM_ECC_WORDS];
+
+	uECC_word_t tmp[NUM_ECC_WORDS];
+	uECC_word_t *p2[2] = {_private, tmp};
+	uECC_word_t *initial_Z = 0;
+	uECC_word_t carry;
+	wordcount_t num_words = curve->num_words;
+	wordcount_t num_bytes = curve->num_bytes;
+	int r;
+
+	/* Converting buffers to correct bit order: */
+	uECC_vli_bytesToNative(_private,
+      			       private_key,
+			       BITS_TO_BYTES(curve->num_n_bits));
+	uECC_vli_bytesToNative(_public,
+      			       public_key,
+			       num_bytes);
+	uECC_vli_bytesToNative(_public + num_words,
+			       public_key + num_bytes,
+			       num_bytes);
+
+	/* Regularize the bitcount for the private key so that attackers cannot use a
+	 * side channel attack to learn the number of leading zeros. */
+	carry = regularize_k(_private, _private, tmp, curve);
+
+	/* If an RNG function was specified, try to get a random initial Z value to
+	 * improve protection against side-channel attacks. */
+	if (g_rng_function) {
+		if (!uECC_generate_random_int(p2[carry], curve->p, num_words)) {
+			r = 0;
+			goto clear_and_out;
+    		}
+    		initial_Z = p2[carry];
+  	}
+
+	EccPoint_mult(_public, _public, p2[!carry], initial_Z, curve->num_n_bits + 1,
+		      curve);
+
+	uECC_vli_nativeToBytes(secret, num_bytes, _public);
+	r = !EccPoint_isZero(_public, curve);
+
+clear_and_out:
+	/* erasing temporary buffer used to store secret: */
+	mbedtls_platform_zeroize(p2, sizeof(p2));
+	mbedtls_platform_zeroize(tmp, sizeof(tmp));
+	mbedtls_platform_zeroize(_private, sizeof(_private));
+
+	return r;
+}
+#else
+typedef int mbedtls_dummy_uecc_def;
+#endif /* MBEDTLS_USE_UECC */
diff --git a/tinycrypt/ecc_dsa.c b/tinycrypt/ecc_dsa.c
new file mode 100644
index 000000000..ff8a78a3a
--- /dev/null
+++ b/tinycrypt/ecc_dsa.c
@@ -0,0 +1,297 @@
+/* ec_dsa.c - TinyCrypt implementation of EC-DSA */
+
+/* Copyright (c) 2014, Kenneth MacKay
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *  * Redistributions of source code must retain the above copyright notice,
+ *    this list of conditions and the following disclaimer.
+ *  * Redistributions in binary form must reproduce the above copyright notice,
+ *    this list of conditions and the following disclaimer in the documentation
+ *    and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.*/
+
+/*
+ *  Copyright (C) 2017 by Intel Corporation, All Rights Reserved.
+ *
+ *  Redistribution and use in source and binary forms, with or without
+ *  modification, are permitted provided that the following conditions are met:
+ *
+ *    - Redistributions of source code must retain the above copyright notice,
+ *     this list of conditions and the following disclaimer.
+ *
+ *    - Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ *    - Neither the name of Intel Corporation nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ *  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ *  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ *  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ *  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ *  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ *  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ *  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ *  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ *  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ *  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ *  POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#if defined(MBEDTLS_USE_UECC)
+#include <tinycrypt/ecc.h>
+#include <tinycrypt/ecc_dsa.h>
+
+#if default_RNG_defined
+static uECC_RNG_Function g_rng_function = &default_CSPRNG;
+#else
+static uECC_RNG_Function g_rng_function = 0;
+#endif
+
+static void bits2int(uECC_word_t *native, const uint8_t *bits,
+		     unsigned bits_size, uECC_Curve curve)
+{
+	unsigned num_n_bytes = BITS_TO_BYTES(curve->num_n_bits);
+	unsigned num_n_words = BITS_TO_WORDS(curve->num_n_bits);
+	int shift;
+	uECC_word_t carry;
+	uECC_word_t *ptr;
+
+	if (bits_size > num_n_bytes) {
+		bits_size = num_n_bytes;
+	}
+
+	uECC_vli_clear(native, num_n_words);
+	uECC_vli_bytesToNative(native, bits, bits_size);
+	if (bits_size * 8 <= (unsigned)curve->num_n_bits) {
+		return;
+	}
+	shift = bits_size * 8 - curve->num_n_bits;
+	carry = 0;
+	ptr = native + num_n_words;
+	while (ptr-- > native) {
+		uECC_word_t temp = *ptr;
+		*ptr = (temp >> shift) | carry;
+		carry = temp << (uECC_WORD_BITS - shift);
+	}
+
+	/* Reduce mod curve_n */
+	if (uECC_vli_cmp_unsafe(curve->n, native, num_n_words) != 1) {
+		uECC_vli_sub(native, native, curve->n, num_n_words);
+	}
+}
+
+int uECC_sign_with_k(const uint8_t *private_key, const uint8_t *message_hash,
+		     unsigned hash_size, uECC_word_t *k, uint8_t *signature,
+		     uECC_Curve curve)
+{
+
+	uECC_word_t tmp[NUM_ECC_WORDS];
+	uECC_word_t s[NUM_ECC_WORDS];
+	uECC_word_t *k2[2] = {tmp, s};
+	uECC_word_t p[NUM_ECC_WORDS * 2];
+	uECC_word_t carry;
+	wordcount_t num_words = curve->num_words;
+	wordcount_t num_n_words = BITS_TO_WORDS(curve->num_n_bits);
+	bitcount_t num_n_bits = curve->num_n_bits;
+
+	/* Make sure 0 < k < curve_n */
+  	if (uECC_vli_isZero(k, num_words) ||
+	    uECC_vli_cmp(curve->n, k, num_n_words) != 1) {
+		return 0;
+	}
+
+	carry = regularize_k(k, tmp, s, curve);
+	EccPoint_mult(p, curve->G, k2[!carry], 0, num_n_bits + 1, curve);
+	if (uECC_vli_isZero(p, num_words)) {
+		return 0;
+	}
+
+	/* If an RNG function was specified, get a random number
+	to prevent side channel analysis of k. */
+	if (!g_rng_function) {
+		uECC_vli_clear(tmp, num_n_words);
+		tmp[0] = 1;
+	}
+	else if (!uECC_generate_random_int(tmp, curve->n, num_n_words)) {
+		return 0;
+	}
+
+	/* Prevent side channel analysis of uECC_vli_modInv() to determine
+	bits of k / the private key by premultiplying by a random number */
+	uECC_vli_modMult(k, k, tmp, curve->n, num_n_words); /* k' = rand * k */
+	uECC_vli_modInv(k, k, curve->n, num_n_words);       /* k = 1 / k' */
+	uECC_vli_modMult(k, k, tmp, curve->n, num_n_words); /* k = 1 / k */
+
+	uECC_vli_nativeToBytes(signature, curve->num_bytes, p); /* store r */
+
+	/* tmp = d: */
+	uECC_vli_bytesToNative(tmp, private_key, BITS_TO_BYTES(curve->num_n_bits));
+
+	s[num_n_words - 1] = 0;
+	uECC_vli_set(s, p, num_words);
+	uECC_vli_modMult(s, tmp, s, curve->n, num_n_words); /* s = r*d */
+
+	bits2int(tmp, message_hash, hash_size, curve);
+	uECC_vli_modAdd(s, tmp, s, curve->n, num_n_words); /* s = e + r*d */
+	uECC_vli_modMult(s, s, k, curve->n, num_n_words);  /* s = (e + r*d) / k */
+	if (uECC_vli_numBits(s, num_n_words) > (bitcount_t)curve->num_bytes * 8) {
+		return 0;
+	}
+
+	uECC_vli_nativeToBytes(signature + curve->num_bytes, curve->num_bytes, s);
+	return 1;
+}
+
+int uECC_sign(const uint8_t *private_key, const uint8_t *message_hash,
+	      unsigned hash_size, uint8_t *signature, uECC_Curve curve)
+{
+	      uECC_word_t _random[2*NUM_ECC_WORDS];
+	      uECC_word_t k[NUM_ECC_WORDS];
+	      uECC_word_t tries;
+
+	for (tries = 0; tries < uECC_RNG_MAX_TRIES; ++tries) {
+		/* Generating _random uniformly at random: */
+		uECC_RNG_Function rng_function = uECC_get_rng();
+		if (!rng_function ||
+		    !rng_function((uint8_t *)_random, 2*NUM_ECC_WORDS*uECC_WORD_SIZE)) {
+			return 0;
+		}
+
+		// computing k as modular reduction of _random (see FIPS 186.4 B.5.1):
+		uECC_vli_mmod(k, _random, curve->n, BITS_TO_WORDS(curve->num_n_bits));
+
+		if (uECC_sign_with_k(private_key, message_hash, hash_size, k, signature, 
+		    curve)) {
+			return 1;
+		}
+	}
+	return 0;
+}
+
+static bitcount_t smax(bitcount_t a, bitcount_t b)
+{
+	return (a > b ? a : b);
+}
+
+int uECC_verify(const uint8_t *public_key, const uint8_t *message_hash,
+		unsigned hash_size, const uint8_t *signature,
+	        uECC_Curve curve)
+{
+
+	uECC_word_t u1[NUM_ECC_WORDS], u2[NUM_ECC_WORDS];
+	uECC_word_t z[NUM_ECC_WORDS];
+	uECC_word_t sum[NUM_ECC_WORDS * 2];
+	uECC_word_t rx[NUM_ECC_WORDS];
+	uECC_word_t ry[NUM_ECC_WORDS];
+	uECC_word_t tx[NUM_ECC_WORDS];
+	uECC_word_t ty[NUM_ECC_WORDS];
+	uECC_word_t tz[NUM_ECC_WORDS];
+	const uECC_word_t *points[4];
+	const uECC_word_t *point;
+	bitcount_t num_bits;
+	bitcount_t i;
+
+	uECC_word_t _public[NUM_ECC_WORDS * 2];
+	uECC_word_t r[NUM_ECC_WORDS], s[NUM_ECC_WORDS];
+	wordcount_t num_words = curve->num_words;
+	wordcount_t num_n_words = BITS_TO_WORDS(curve->num_n_bits);
+
+	rx[num_n_words - 1] = 0;
+	r[num_n_words - 1] = 0;
+	s[num_n_words - 1] = 0;
+
+	uECC_vli_bytesToNative(_public, public_key, curve->num_bytes);
+	uECC_vli_bytesToNative(_public + num_words, public_key + curve->num_bytes,
+			       curve->num_bytes);
+	uECC_vli_bytesToNative(r, signature, curve->num_bytes);
+	uECC_vli_bytesToNative(s, signature + curve->num_bytes, curve->num_bytes);
+
+	/* r, s must not be 0. */
+	if (uECC_vli_isZero(r, num_words) || uECC_vli_isZero(s, num_words)) {
+		return 0;
+	}
+
+	/* r, s must be < n. */
+	if (uECC_vli_cmp_unsafe(curve->n, r, num_n_words) != 1 ||
+	    uECC_vli_cmp_unsafe(curve->n, s, num_n_words) != 1) {
+		return 0;
+	}
+
+	/* Calculate u1 and u2. */
+	uECC_vli_modInv(z, s, curve->n, num_n_words); /* z = 1/s */
+	u1[num_n_words - 1] = 0;
+	bits2int(u1, message_hash, hash_size, curve);
+	uECC_vli_modMult(u1, u1, z, curve->n, num_n_words); /* u1 = e/s */
+	uECC_vli_modMult(u2, r, z, curve->n, num_n_words); /* u2 = r/s */
+
+	/* Calculate sum = G + Q. */
+	uECC_vli_set(sum, _public, num_words);
+	uECC_vli_set(sum + num_words, _public + num_words, num_words);
+	uECC_vli_set(tx, curve->G, num_words);
+	uECC_vli_set(ty, curve->G + num_words, num_words);
+	uECC_vli_modSub(z, sum, tx, curve->p, num_words); /* z = x2 - x1 */
+	XYcZ_add(tx, ty, sum, sum + num_words, curve);
+	uECC_vli_modInv(z, z, curve->p, num_words); /* z = 1/z */
+	apply_z(sum, sum + num_words, z, curve);
+
+	/* Use Shamir's trick to calculate u1*G + u2*Q */
+	points[0] = 0;
+	points[1] = curve->G;
+	points[2] = _public;
+	points[3] = sum;
+	num_bits = smax(uECC_vli_numBits(u1, num_n_words),
+	uECC_vli_numBits(u2, num_n_words));
+
+	point = points[(!!uECC_vli_testBit(u1, num_bits - 1)) |
+                       ((!!uECC_vli_testBit(u2, num_bits - 1)) << 1)];
+	uECC_vli_set(rx, point, num_words);
+	uECC_vli_set(ry, point + num_words, num_words);
+	uECC_vli_clear(z, num_words);
+	z[0] = 1;
+
+	for (i = num_bits - 2; i >= 0; --i) {
+		uECC_word_t index;
+		curve->double_jacobian(rx, ry, z, curve);
+
+		index = (!!uECC_vli_testBit(u1, i)) | ((!!uECC_vli_testBit(u2, i)) << 1);
+		point = points[index];
+		if (point) {
+			uECC_vli_set(tx, point, num_words);
+			uECC_vli_set(ty, point + num_words, num_words);
+			apply_z(tx, ty, z, curve);
+			uECC_vli_modSub(tz, rx, tx, curve->p, num_words); /* Z = x2 - x1 */
+			XYcZ_add(tx, ty, rx, ry, curve);
+			uECC_vli_modMult_fast(z, z, tz, curve);
+		}
+  	}
+
+	uECC_vli_modInv(z, z, curve->p, num_words); /* Z = 1/Z */
+	apply_z(rx, ry, z, curve);
+
+	/* v = x1 (mod n) */
+	if (uECC_vli_cmp_unsafe(curve->n, rx, num_n_words) != 1) {
+		uECC_vli_sub(rx, rx, curve->n, num_n_words);
+	}
+
+	/* Accept only if v == r. */
+	return (int)(uECC_vli_equal(rx, r, num_words) == 0);
+}
+#else
+typedef int mbedtls_dummy_uecc_def;
+#endif /* MBEDTLS_USE_UECC */