Merge 8.6

203 files changed, 6320 insertions, 6635 deletions

diff --git a/libtommath/README.md b/libtommath/README.md
index 0e57a4d..be5b207 100644
--- a/libtommath/README.md
+++ b/libtommath/README.md
@@ -4,22 +4,41 @@ This is the git repository for [LibTomMath](http://www.libtom.net/LibTomMath/),
 
 ## Build Status
 
+### Travis CI
+
 master: [![Build Status](https://api.travis-ci.org/libtom/libtommath.png?branch=master)](https://travis-ci.org/libtom/libtommath)
 
 develop: [![Build Status](https://api.travis-ci.org/libtom/libtommath.png?branch=develop)](https://travis-ci.org/libtom/libtommath)
 
+### AppVeyor
+
+master: [![Build status](https://ci.appveyor.com/api/projects/status/b80lpolw3i8m6hsh/branch/master?svg=true)](https://ci.appveyor.com/project/libtom/libtommath/branch/master)
+
+develop: [![Build status](https://ci.appveyor.com/api/projects/status/b80lpolw3i8m6hsh/branch/develop?svg=true)](https://ci.appveyor.com/project/libtom/libtommath/branch/develop)
+
+### ABI Laboratory
+
 API/ABI changes: [check here](https://abi-laboratory.pro/tracker/timeline/libtommath/)
 
 ## Summary
 
 The `develop` branch contains the in-development version. Stable releases are tagged.
 
-Documentation is built from the LaTeX file `bn.tex`. There is also limited documentation in `tommath.h`. There is also a document, `tommath.pdf`, which describes the goals of the project and many of the algorithms used.
+Documentation is built from the LaTeX file `bn.tex`. There is also limited documentation in `tommath.h`.
+There is also a document, `tommath.pdf`, which describes the goals of the project and many of the algorithms used.
 
-The project can be build by using `make`. Along with the usual `make`, `make clean` and `make install`, there are several other build targets, see the makefile for details. There are also makefiles for certain specific platforms.
+The project can be build by using `make`. Along with the usual `make`, `make clean` and `make install`,
+there are several other build targets, see the makefile for details.
+There are also makefiles for certain specific platforms.
 
 ## Testing
 
 Tests are located in `demo/` and can be built in two flavors.
-* `make test` creates a test binary that is intended to be run against `mtest`. `mtest` can be built with `make mtest` and test execution is done like `./mtest/mtest | ./test`. `mtest` is creating test vectors using an alternative MPI library and `test` is consuming these vectors to verify correct behavior of ltm
-* `make test_standalone` creates a stand-alone test binary that executes several test routines.
+* `make test` creates a stand-alone test binary that executes several test routines.
+* `make mtest_opponent` creates a test binary that is intended to be run against `mtest`.
+  `mtest` can be built with `make mtest` and test execution is done like `./mtest/mtest | ./mtest_opponent`.
+  `mtest` is creating test vectors using an alternative MPI library and `test` is consuming these vectors to verify correct behavior of ltm
+
+## Building and Installing
+
+Building is straightforward for GNU Linux only, the section "Building LibTomMath" in the documentation in `doc/bn.pdf` has the details.
diff --git a/libtommath/appveyor.yml b/libtommath/appveyor.yml
new file mode 100644
index 0000000..187a09a
--- /dev/null
+++ b/libtommath/appveyor.yml
@@ -0,0 +1,20 @@
+version: 1.2.0-{build}
+branches:
+  only:
+  - master
+  - develop
+  - /^release/
+  - /^travis/
+image:
+- Visual Studio 2019
+- Visual Studio 2017
+- Visual Studio 2015
+build_script:
+- cmd: >-
+    if "Visual Studio 2019"=="%APPVEYOR_BUILD_WORKER_IMAGE%" call "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Auxiliary\Build\vcvars64.bat"
+        if "Visual Studio 2017"=="%APPVEYOR_BUILD_WORKER_IMAGE%" call "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Auxiliary\Build\vcvars64.bat"
+        if "Visual Studio 2015"=="%APPVEYOR_BUILD_WORKER_IMAGE%" call "C:\Program Files\Microsoft SDKs\Windows\v7.1\Bin\SetEnv.cmd" /x64
+        if "Visual Studio 2015"=="%APPVEYOR_BUILD_WORKER_IMAGE%" call "C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\vcvarsall.bat" x86_amd64
+        nmake -f makefile.msvc all
+test_script:
+- cmd: test.exe
diff --git a/libtommath/astylerc b/libtommath/astylerc
index 5d63f7a..c5ff779 100644
--- a/libtommath/astylerc
+++ b/libtommath/astylerc
@@ -4,6 +4,9 @@
 # usage:
 #       astyle --options=astylerc *.[ch]
 
+# Do not create backup, annonying in the times of git
+suffix=none
+
 ## Bracket Style Options
 style=kr
 
diff --git a/libtommath/bn_cutoffs.c b/libtommath/bn_cutoffs.c
new file mode 100644
index 0000000..b02ab71
--- /dev/null
+++ b/libtommath/bn_cutoffs.c
@@ -0,0 +1,14 @@
+#include "tommath_private.h"
+#ifdef BN_CUTOFFS_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+#ifndef MP_FIXED_CUTOFFS
+#include "tommath_cutoffs.h"
+int KARATSUBA_MUL_CUTOFF = MP_DEFAULT_KARATSUBA_MUL_CUTOFF,
+    KARATSUBA_SQR_CUTOFF = MP_DEFAULT_KARATSUBA_SQR_CUTOFF,
+    TOOM_MUL_CUTOFF = MP_DEFAULT_TOOM_MUL_CUTOFF,
+    TOOM_SQR_CUTOFF = MP_DEFAULT_TOOM_SQR_CUTOFF;
+#endif
+
+#endif
diff --git a/libtommath/bn_deprecated.c b/libtommath/bn_deprecated.c
new file mode 100644
index 0000000..456f8a2
--- /dev/null
+++ b/libtommath/bn_deprecated.c
@@ -0,0 +1,321 @@
+#include "tommath_private.h"
+#ifdef BN_DEPRECATED_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+#ifdef BN_MP_GET_BIT_C
+int mp_get_bit(const mp_int *a, int b)
+{
+   if (b < 0) {
+      return MP_VAL;
+   }
+   return (s_mp_get_bit(a, (unsigned int)b) == MP_YES) ? MP_YES : MP_NO;
+}
+#endif
+#ifdef BN_MP_JACOBI_C
+mp_err mp_jacobi(const mp_int *a, const mp_int *n, int *c)
+{
+   if (a->sign == MP_NEG) {
+      return MP_VAL;
+   }
+   if (mp_cmp_d(n, 0uL) != MP_GT) {
+      return MP_VAL;
+   }
+   return mp_kronecker(a, n, c);
+}
+#endif
+#ifdef BN_MP_PRIME_RANDOM_EX_C
+mp_err mp_prime_random_ex(mp_int *a, int t, int size, int flags, private_mp_prime_callback cb, void *dat)
+{
+   return s_mp_prime_random_ex(a, t, size, flags, cb, dat);
+}
+#endif
+#ifdef BN_MP_RAND_DIGIT_C
+mp_err mp_rand_digit(mp_digit *r)
+{
+   mp_err err = s_mp_rand_source(r, sizeof(mp_digit));
+   *r &= MP_MASK;
+   return err;
+}
+#endif
+#ifdef BN_FAST_MP_INVMOD_C
+mp_err fast_mp_invmod(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return s_mp_invmod_fast(a, b, c);
+}
+#endif
+#ifdef BN_FAST_MP_MONTGOMERY_REDUCE_C
+mp_err fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
+{
+   return s_mp_montgomery_reduce_fast(x, n, rho);
+}
+#endif
+#ifdef BN_FAST_S_MP_MUL_DIGS_C
+mp_err fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
+{
+   return s_mp_mul_digs_fast(a, b, c, digs);
+}
+#endif
+#ifdef BN_FAST_S_MP_MUL_HIGH_DIGS_C
+mp_err fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
+{
+   return s_mp_mul_high_digs_fast(a, b, c, digs);
+}
+#endif
+#ifdef BN_FAST_S_MP_SQR_C
+mp_err fast_s_mp_sqr(const mp_int *a, mp_int *b)
+{
+   return s_mp_sqr_fast(a, b);
+}
+#endif
+#ifdef BN_MP_BALANCE_MUL_C
+mp_err mp_balance_mul(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return s_mp_balance_mul(a, b, c);
+}
+#endif
+#ifdef BN_MP_EXPTMOD_FAST_C
+mp_err mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode)
+{
+   return s_mp_exptmod_fast(G, X, P, Y, redmode);
+}
+#endif
+#ifdef BN_MP_INVMOD_SLOW_C
+mp_err mp_invmod_slow(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return s_mp_invmod_slow(a, b, c);
+}
+#endif
+#ifdef BN_MP_KARATSUBA_MUL_C
+mp_err mp_karatsuba_mul(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return s_mp_karatsuba_mul(a, b, c);
+}
+#endif
+#ifdef BN_MP_KARATSUBA_SQR_C
+mp_err mp_karatsuba_sqr(const mp_int *a, mp_int *b)
+{
+   return s_mp_karatsuba_sqr(a, b);
+}
+#endif
+#ifdef BN_MP_TOOM_MUL_C
+mp_err mp_toom_mul(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return s_mp_toom_mul(a, b, c);
+}
+#endif
+#ifdef BN_MP_TOOM_SQR_C
+mp_err mp_toom_sqr(const mp_int *a, mp_int *b)
+{
+   return s_mp_toom_sqr(a, b);
+}
+#endif
+#ifdef S_MP_REVERSE_C
+void bn_reverse(unsigned char *s, int len)
+{
+   if (len > 0) {
+      s_mp_reverse(s, (size_t)len);
+   }
+}
+#endif
+#ifdef BN_MP_TC_AND_C
+mp_err mp_tc_and(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return mp_and(a, b, c);
+}
+#endif
+#ifdef BN_MP_TC_OR_C
+mp_err mp_tc_or(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return mp_or(a, b, c);
+}
+#endif
+#ifdef BN_MP_TC_XOR_C
+mp_err mp_tc_xor(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   return mp_xor(a, b, c);
+}
+#endif
+#ifdef BN_MP_TC_DIV_2D_C
+mp_err mp_tc_div_2d(const mp_int *a, int b, mp_int *c)
+{
+   return mp_signed_rsh(a, b, c);
+}
+#endif
+#ifdef BN_MP_INIT_SET_INT_C
+mp_err mp_init_set_int(mp_int *a, unsigned long b)
+{
+   return mp_init_u32(a, (uint32_t)b);
+}
+#endif
+#ifdef BN_MP_SET_INT_C
+mp_err mp_set_int(mp_int *a, unsigned long b)
+{
+   mp_set_u32(a, (uint32_t)b);
+   return MP_OKAY;
+}
+#endif
+#ifdef BN_MP_SET_LONG_C
+mp_err mp_set_long(mp_int *a, unsigned long b)
+{
+   mp_set_ul(a, b);
+   return MP_OKAY;
+}
+#endif
+#ifdef BN_MP_SET_LONG_LONG_C
+mp_err mp_set_long_long(mp_int *a, unsigned long long b)
+{
+   mp_set_ull(a, b);
+   return MP_OKAY;
+}
+#endif
+#ifdef BN_MP_GET_INT_C
+unsigned long mp_get_int(const mp_int *a)
+{
+   return (unsigned long)mp_get_mag_u32(a);
+}
+#endif
+#ifdef BN_MP_GET_LONG_C
+unsigned long mp_get_long(const mp_int *a)
+{
+   return (unsigned long)mp_get_mag_ul(a);
+}
+#endif
+#ifdef BN_MP_GET_LONG_LONG_C
+unsigned long long mp_get_long_long(const mp_int *a)
+{
+   return mp_get_mag_ull(a);
+}
+#endif
+#ifdef BN_MP_PRIME_IS_DIVISIBLE_C
+mp_err mp_prime_is_divisible(const mp_int *a, mp_bool *result)
+{
+   return s_mp_prime_is_divisible(a, result);
+}
+#endif
+#ifdef BN_MP_EXPT_D_EX_C
+mp_err mp_expt_d_ex(const mp_int *a, mp_digit b, mp_int *c, int fast)
+{
+   (void)fast;
+   if (b > MP_MIN(MP_DIGIT_MAX, UINT32_MAX)) {
+      return MP_VAL;
+   }
+   return mp_expt_u32(a, (uint32_t)b, c);
+}
+#endif
+#ifdef BN_MP_EXPT_D_C
+mp_err mp_expt_d(const mp_int *a, mp_digit b, mp_int *c)
+{
+   if (b > MP_MIN(MP_DIGIT_MAX, UINT32_MAX)) {
+      return MP_VAL;
+   }
+   return mp_expt_u32(a, (uint32_t)b, c);
+}
+#endif
+#ifdef BN_MP_N_ROOT_EX_C
+mp_err mp_n_root_ex(const mp_int *a, mp_digit b, mp_int *c, int fast)
+{
+   (void)fast;
+   if (b > MP_MIN(MP_DIGIT_MAX, UINT32_MAX)) {
+      return MP_VAL;
+   }
+   return mp_root_u32(a, (unsigned int)b, c);
+}
+#endif
+#ifdef BN_MP_N_ROOT_C
+mp_err mp_n_root(const mp_int *a, mp_digit b, mp_int *c)
+{
+   if (b > MP_MIN(MP_DIGIT_MAX, UINT32_MAX)) {
+      return MP_VAL;
+   }
+   return mp_root_u32(a, (unsigned int)b, c);
+}
+#endif
+#ifdef BN_MP_UNSIGNED_BIN_SIZE_C
+int mp_unsigned_bin_size(const mp_int *a)
+{
+   return (int)mp_ubin_size(a);
+}
+#endif
+#ifdef BN_MP_READ_UNSIGNED_BIN_C
+mp_err mp_read_unsigned_bin(mp_int *a, const unsigned char *b, int c)
+{
+   return mp_from_ubin(a, b, (size_t) c);
+}
+#endif
+#ifdef BN_MP_TO_UNSIGNED_BIN_C
+mp_err mp_to_unsigned_bin(const mp_int *a, unsigned char *b)
+{
+   return mp_to_ubin(a, b, SIZE_MAX, NULL);
+}
+#endif
+#ifdef BN_MP_TO_UNSIGNED_BIN_N_C
+mp_err mp_to_unsigned_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen)
+{
+   size_t n = mp_ubin_size(a);
+   if (*outlen < (unsigned long)n) {
+      return MP_VAL;
+   }
+   *outlen = (unsigned long)n;
+   return mp_to_ubin(a, b, n, NULL);
+}
+#endif
+#ifdef BN_MP_SIGNED_BIN_SIZE_C
+int mp_signed_bin_size(const mp_int *a)
+{
+   return (int)mp_sbin_size(a);
+}
+#endif
+#ifdef BN_MP_READ_SIGNED_BIN_C
+mp_err mp_read_signed_bin(mp_int *a, const unsigned char *b, int c)
+{
+   return mp_from_sbin(a, b, (size_t) c);
+}
+#endif
+#ifdef BN_MP_TO_SIGNED_BIN_C
+mp_err mp_to_signed_bin(const mp_int *a, unsigned char *b)
+{
+   return mp_to_sbin(a, b, SIZE_MAX, NULL);
+}
+#endif
+#ifdef BN_MP_TO_SIGNED_BIN_N_C
+mp_err mp_to_signed_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen)
+{
+   size_t n = mp_sbin_size(a);
+   if (*outlen < (unsigned long)n) {
+      return MP_VAL;
+   }
+   *outlen = (unsigned long)n;
+   return mp_to_sbin(a, b, n, NULL);
+}
+#endif
+#ifdef BN_MP_TORADIX_N_C
+mp_err mp_toradix_n(const mp_int *a, char *str, int radix, int maxlen)
+{
+   if (maxlen < 0) {
+      return MP_VAL;
+   }
+   return mp_to_radix(a, str, (size_t)maxlen, NULL, radix);
+}
+#endif
+#ifdef BN_MP_TORADIX_C
+mp_err mp_toradix(const mp_int *a, char *str, int radix)
+{
+   return mp_to_radix(a, str, SIZE_MAX, NULL, radix);
+}
+#endif
+#ifdef BN_MP_IMPORT_C
+mp_err mp_import(mp_int *rop, size_t count, int order, size_t size, int endian, size_t nails,
+                 const void *op)
+{
+   return mp_unpack(rop, count, order, size, endian, nails, op);
+}
+#endif
+#ifdef BN_MP_EXPORT_C
+mp_err mp_export(void *rop, size_t *countp, int order, size_t size,
+                 int endian, size_t nails, const mp_int *op)
+{
+   return mp_pack(rop, SIZE_MAX, countp, order, size, endian, nails, op);
+}
+#endif
+#endif
diff --git a/libtommath/bn_error.c b/libtommath/bn_error.c
deleted file mode 100644
index 697875f..0000000
--- a/libtommath/bn_error.c
+++ /dev/null
@@ -1,44 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_ERROR_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-static const struct {
-   int code;
-   const char *msg;
-} msgs[] = {
-   { MP_OKAY, "Successful" },
-   { MP_MEM,  "Out of heap" },
-   { MP_VAL,  "Value out of range" }
-};
-
-/* return a char * string for a given code */
-const char *mp_error_to_string(int code)
-{
-   size_t x;
-
-   /* scan the lookup table for the given message */
-   for (x = 0; x < (sizeof(msgs) / sizeof(msgs[0])); x++) {
-      if (msgs[x].code == code) {
-         return msgs[x].msg;
-      }
-   }
-
-   /* generic reply for invalid code */
-   return "Invalid error code";
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_fast_mp_invmod.c b/libtommath/bn_fast_mp_invmod.c
deleted file mode 100644
index 3c8088f..0000000
--- a/libtommath/bn_fast_mp_invmod.c
+++ /dev/null
@@ -1,160 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_FAST_MP_INVMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* computes the modular inverse via binary extended euclidean algorithm,
- * that is c = 1/a mod b
- *
- * Based on slow invmod except this is optimized for the case where b is
- * odd as per HAC Note 14.64 on pp. 610
- */
-int fast_mp_invmod(const mp_int *a, const mp_int *b, mp_int *c)
-{
-   mp_int  x, y, u, v, B, D;
-   int     res, neg;
-
-   /* 2. [modified] b must be odd   */
-   if (mp_iseven(b) == MP_YES) {
-      return MP_VAL;
-   }
-
-   /* init all our temps */
-   if ((res = mp_init_multi(&x, &y, &u, &v, &B, &D, NULL)) != MP_OKAY) {
-      return res;
-   }
-
-   /* x == modulus, y == value to invert */
-   if ((res = mp_copy(b, &x)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* we need y = |a| */
-   if ((res = mp_mod(a, b, &y)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* if one of x,y is zero return an error! */
-   if ((mp_iszero(&x) == MP_YES) || (mp_iszero(&y) == MP_YES)) {
-      res = MP_VAL;
-      goto LBL_ERR;
-   }
-
-   /* 3. u=x, v=y, A=1, B=0, C=0,D=1 */
-   if ((res = mp_copy(&x, &u)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_copy(&y, &v)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_set(&D, 1uL);
-
-top:
-   /* 4.  while u is even do */
-   while (mp_iseven(&u) == MP_YES) {
-      /* 4.1 u = u/2 */
-      if ((res = mp_div_2(&u, &u)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      /* 4.2 if B is odd then */
-      if (mp_isodd(&B) == MP_YES) {
-         if ((res = mp_sub(&B, &x, &B)) != MP_OKAY) {
-            goto LBL_ERR;
-         }
-      }
-      /* B = B/2 */
-      if ((res = mp_div_2(&B, &B)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* 5.  while v is even do */
-   while (mp_iseven(&v) == MP_YES) {
-      /* 5.1 v = v/2 */
-      if ((res = mp_div_2(&v, &v)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      /* 5.2 if D is odd then */
-      if (mp_isodd(&D) == MP_YES) {
-         /* D = (D-x)/2 */
-         if ((res = mp_sub(&D, &x, &D)) != MP_OKAY) {
-            goto LBL_ERR;
-         }
-      }
-      /* D = D/2 */
-      if ((res = mp_div_2(&D, &D)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* 6.  if u >= v then */
-   if (mp_cmp(&u, &v) != MP_LT) {
-      /* u = u - v, B = B - D */
-      if ((res = mp_sub(&u, &v, &u)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-
-      if ((res = mp_sub(&B, &D, &B)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   } else {
-      /* v - v - u, D = D - B */
-      if ((res = mp_sub(&v, &u, &v)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-
-      if ((res = mp_sub(&D, &B, &D)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* if not zero goto step 4 */
-   if (mp_iszero(&u) == MP_NO) {
-      goto top;
-   }
-
-   /* now a = C, b = D, gcd == g*v */
-
-   /* if v != 1 then there is no inverse */
-   if (mp_cmp_d(&v, 1uL) != MP_EQ) {
-      res = MP_VAL;
-      goto LBL_ERR;
-   }
-
-   /* b is now the inverse */
-   neg = a->sign;
-   while (D.sign == MP_NEG) {
-      if ((res = mp_add(&D, b, &D)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* too big */
-   while (mp_cmp_mag(&D, b) != MP_LT) {
-      if ((res = mp_sub(&D, b, &D)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   mp_exch(&D, c);
-   c->sign = neg;
-   res = MP_OKAY;
-
-LBL_ERR:
-   mp_clear_multi(&x, &y, &u, &v, &B, &D, NULL);
-   return res;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_2expt.c b/libtommath/bn_mp_2expt.c
index 42f5746..0ae3df1 100644
--- a/libtommath/bn_mp_2expt.c
+++ b/libtommath/bn_mp_2expt.c
@@ -1,44 +1,31 @@
 #include "tommath_private.h"
 #ifdef BN_MP_2EXPT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* computes a = 2**b
  *
  * Simple algorithm which zeroes the int, grows it then just sets one bit
  * as required.
  */
-int mp_2expt(mp_int *a, int b)
+mp_err mp_2expt(mp_int *a, int b)
 {
-   int     res;
+   mp_err    err;
 
    /* zero a as per default */
    mp_zero(a);
 
    /* grow a to accomodate the single bit */
-   if ((res = mp_grow(a, (b / DIGIT_BIT) + 1)) != MP_OKAY) {
-      return res;
+   if ((err = mp_grow(a, (b / MP_DIGIT_BIT) + 1)) != MP_OKAY) {
+      return err;
    }
 
    /* set the used count of where the bit will go */
-   a->used = (b / DIGIT_BIT) + 1;
+   a->used = (b / MP_DIGIT_BIT) + 1;
 
    /* put the single bit in its place */
-   a->dp[b / DIGIT_BIT] = (mp_digit)1 << (mp_digit)(b % DIGIT_BIT);
+   a->dp[b / MP_DIGIT_BIT] = (mp_digit)1 << (mp_digit)(b % MP_DIGIT_BIT);
 
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_abs.c b/libtommath/bn_mp_abs.c
index f12d261..00900bb 100644
--- a/libtommath/bn_mp_abs.c
+++ b/libtommath/bn_mp_abs.c
@@ -1,29 +1,20 @@
 #include "tommath_private.h"
 #ifdef BN_MP_ABS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* b = |a|
  *
  * Simple function copies the input and fixes the sign to positive
  */
-int mp_abs(const mp_int *a, mp_int *b)
+mp_err mp_abs(const mp_int *a, mp_int *b)
 {
-   int     res;
+   mp_err     err;
 
    /* copy a to b */
    if (a != b) {
-      if ((res = mp_copy(a, b)) != MP_OKAY) {
-         return res;
+      if ((err = mp_copy(a, b)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -33,7 +24,3 @@ int mp_abs(const mp_int *a, mp_int *b)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_add.c b/libtommath/bn_mp_add.c
index f04388a..dfa78de 100644
--- a/libtommath/bn_mp_add.c
+++ b/libtommath/bn_mp_add.c
@@ -1,21 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_ADD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* high level addition (handles signs) */
-int mp_add(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_add(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int     sa, sb, res;
+   mp_sign sa, sb;
+   mp_err err;
 
    /* get sign of both inputs */
    sa = a->sign;
@@ -26,7 +18,7 @@ int mp_add(const mp_int *a, const mp_int *b, mp_int *c)
       /* both positive or both negative */
       /* add their magnitudes, copy the sign */
       c->sign = sa;
-      res = s_mp_add(a, b, c);
+      err = s_mp_add(a, b, c);
    } else {
       /* one positive, the other negative */
       /* subtract the one with the greater magnitude from */
@@ -34,17 +26,13 @@ int mp_add(const mp_int *a, const mp_int *b, mp_int *c)
       /* the sign of the one with the greater magnitude. */
       if (mp_cmp_mag(a, b) == MP_LT) {
          c->sign = sb;
-         res = s_mp_sub(b, a, c);
+         err = s_mp_sub(b, a, c);
       } else {
          c->sign = sa;
-         res = s_mp_sub(a, b, c);
+         err = s_mp_sub(a, b, c);
       }
    }
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_add_d.c b/libtommath/bn_mp_add_d.c
index ecdb791..f301575 100644
--- a/libtommath/bn_mp_add_d.c
+++ b/libtommath/bn_mp_add_d.c
@@ -1,27 +1,19 @@
 #include "tommath_private.h"
 #ifdef BN_MP_ADD_D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* single digit addition */
-int mp_add_d(const mp_int *a, mp_digit b, mp_int *c)
+mp_err mp_add_d(const mp_int *a, mp_digit b, mp_int *c)
 {
-   int     res, ix, oldused;
-   mp_digit *tmpa, *tmpc, mu;
+   mp_err     err;
+   int ix, oldused;
+   mp_digit *tmpa, *tmpc;
 
    /* grow c as required */
    if (c->alloc < (a->used + 1)) {
-      if ((res = mp_grow(c, a->used + 1)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(c, a->used + 1)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -32,7 +24,7 @@ int mp_add_d(const mp_int *a, mp_digit b, mp_int *c)
       a_.sign = MP_ZPOS;
 
       /* c = |a| - b */
-      res = mp_sub_d(&a_, b, c);
+      err = mp_sub_d(&a_, b, c);
 
       /* fix sign  */
       c->sign = MP_NEG;
@@ -40,7 +32,7 @@ int mp_add_d(const mp_int *a, mp_digit b, mp_int *c)
       /* clamp */
       mp_clamp(c);
 
-      return res;
+      return err;
    }
 
    /* old number of used digits in c */
@@ -54,17 +46,11 @@ int mp_add_d(const mp_int *a, mp_digit b, mp_int *c)
 
    /* if a is positive */
    if (a->sign == MP_ZPOS) {
-      /* add digit, after this we're propagating
-       * the carry.
-       */
-      *tmpc   = *tmpa++ + b;
-      mu      = *tmpc >> DIGIT_BIT;
-      *tmpc++ &= MP_MASK;
-
-      /* now handle rest of the digits */
-      for (ix = 1; ix < a->used; ix++) {
+      /* add digits, mu is carry */
+      mp_digit mu = b;
+      for (ix = 0; ix < a->used; ix++) {
          *tmpc   = *tmpa++ + mu;
-         mu      = *tmpc >> DIGIT_BIT;
+         mu      = *tmpc >> MP_DIGIT_BIT;
          *tmpc++ &= MP_MASK;
       }
       /* set final carry */
@@ -94,16 +80,10 @@ int mp_add_d(const mp_int *a, mp_digit b, mp_int *c)
    c->sign = MP_ZPOS;
 
    /* now zero to oldused */
-   while (ix++ < oldused) {
-      *tmpc++ = 0;
-   }
+   MP_ZERO_DIGITS(tmpc, oldused - ix);
    mp_clamp(c);
 
    return MP_OKAY;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_addmod.c b/libtommath/bn_mp_addmod.c
index f8e4dda..1dcfb67 100644
--- a/libtommath/bn_mp_addmod.c
+++ b/libtommath/bn_mp_addmod.c
@@ -1,37 +1,25 @@
 #include "tommath_private.h"
 #ifdef BN_MP_ADDMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* d = a + b (mod c) */
-int mp_addmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d)
+mp_err mp_addmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d)
 {
-   int     res;
+   mp_err  err;
    mp_int  t;
 
-   if ((res = mp_init(&t)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init(&t)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_add(a, b, &t)) != MP_OKAY) {
-      mp_clear(&t);
-      return res;
+   if ((err = mp_add(a, b, &t)) != MP_OKAY) {
+      goto LBL_ERR;
    }
-   res = mp_mod(&t, c, d);
+   err = mp_mod(&t, c, d);
+
+LBL_ERR:
    mp_clear(&t);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_and.c b/libtommath/bn_mp_and.c
index 10fe28b..c259f8d 100644
--- a/libtommath/bn_mp_and.c
+++ b/libtommath/bn_mp_and.c
@@ -6,7 +6,7 @@
 /* two complement and */
 mp_err mp_and(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int used = MAX(a->used, b->used) + 1, i;
+   int used = MP_MAX(a->used, b->used) + 1, i;
    mp_err err;
    mp_digit ac = 1, bc = 1, cc = 1;
    mp_sign csign = ((a->sign == MP_NEG) && (b->sign == MP_NEG)) ? MP_NEG : MP_ZPOS;
diff --git a/libtommath/bn_mp_clamp.c b/libtommath/bn_mp_clamp.c
index 0953f4b..ac23bfd 100644
--- a/libtommath/bn_mp_clamp.c
+++ b/libtommath/bn_mp_clamp.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_CLAMP_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* trim unused digits
  *
@@ -34,7 +25,3 @@ void mp_clamp(mp_int *a)
    }
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_clear.c b/libtommath/bn_mp_clear.c
index b8e724c..ff78324 100644
--- a/libtommath/bn_mp_clear.c
+++ b/libtommath/bn_mp_clear.c
@@ -1,31 +1,15 @@
 #include "tommath_private.h"
 #ifdef BN_MP_CLEAR_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* clear one (frees)  */
 void mp_clear(mp_int *a)
 {
-   int i;
-
    /* only do anything if a hasn't been freed previously */
    if (a->dp != NULL) {
-      /* first zero the digits */
-      for (i = 0; i < a->used; i++) {
-         a->dp[i] = 0;
-      }
-
       /* free ram */
-      XFREE(a->dp, sizeof (mp_digit) * (size_t)a->alloc);
+      MP_FREE_DIGITS(a->dp, a->alloc);
 
       /* reset members to make debugging easier */
       a->dp    = NULL;
@@ -34,7 +18,3 @@ void mp_clear(mp_int *a)
    }
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_clear_multi.c b/libtommath/bn_mp_clear_multi.c
index c96b4ac..794e45f 100644
--- a/libtommath/bn_mp_clear_multi.c
+++ b/libtommath/bn_mp_clear_multi.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_CLEAR_MULTI_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #include <stdarg.h>
 
@@ -26,7 +17,3 @@ void mp_clear_multi(mp_int *mp, ...)
    va_end(args);
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_cnt_lsb.c b/libtommath/bn_mp_cnt_lsb.c
index 5d9b327..4b2d206 100644
--- a/libtommath/bn_mp_cnt_lsb.c
+++ b/libtommath/bn_mp_cnt_lsb.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_CNT_LSB_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 static const int lnz[16] = {
    4, 0, 1, 0, 2, 0, 1, 0, 3, 0, 1, 0, 2, 0, 1, 0
@@ -23,14 +14,14 @@ int mp_cnt_lsb(const mp_int *a)
    mp_digit q, qq;
 
    /* easy out */
-   if (mp_iszero(a) == MP_YES) {
+   if (MP_IS_ZERO(a)) {
       return 0;
    }
 
    /* scan lower digits until non-zero */
    for (x = 0; (x < a->used) && (a->dp[x] == 0u); x++) {}
    q = a->dp[x];
-   x *= DIGIT_BIT;
+   x *= MP_DIGIT_BIT;
 
    /* now scan this digit until a 1 is found */
    if ((q & 1u) == 0u) {
@@ -44,7 +35,3 @@ int mp_cnt_lsb(const mp_int *a)
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_complement.c b/libtommath/bn_mp_complement.c
index 5a5a969..fef1423 100644
--- a/libtommath/bn_mp_complement.c
+++ b/libtommath/bn_mp_complement.c
@@ -1,25 +1,12 @@
 #include "tommath_private.h"
 #ifdef BN_MP_COMPLEMENT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* b = ~a */
-int mp_complement(const mp_int *a, mp_int *b)
+mp_err mp_complement(const mp_int *a, mp_int *b)
 {
-   int res = mp_neg(a, b);
-   return (res == MP_OKAY) ? mp_sub_d(b, 1uL, b) : res;
+   mp_err err = mp_neg(a, b);
+   return (err == MP_OKAY) ? mp_sub_d(b, 1uL, b) : err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_copy.c b/libtommath/bn_mp_copy.c
index 51e0239..e72fcf6 100644
--- a/libtommath/bn_mp_copy.c
+++ b/libtommath/bn_mp_copy.c
@@ -1,21 +1,14 @@
 #include "tommath_private.h"
 #ifdef BN_MP_COPY_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* copy, b = a */
-int mp_copy(const mp_int *a, mp_int *b)
+mp_err mp_copy(const mp_int *a, mp_int *b)
 {
-   int     res, n;
+   int n;
+   mp_digit *tmpa, *tmpb;
+   mp_err err;
 
    /* if dst == src do nothing */
    if (a == b) {
@@ -24,41 +17,31 @@ int mp_copy(const mp_int *a, mp_int *b)
 
    /* grow dest */
    if (b->alloc < a->used) {
-      if ((res = mp_grow(b, a->used)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(b, a->used)) != MP_OKAY) {
+         return err;
       }
    }
 
    /* zero b and copy the parameters over */
-   {
-      mp_digit *tmpa, *tmpb;
+   /* pointer aliases */
 
-      /* pointer aliases */
+   /* source */
+   tmpa = a->dp;
 
-      /* source */
-      tmpa = a->dp;
+   /* destination */
+   tmpb = b->dp;
 
-      /* destination */
-      tmpb = b->dp;
-
-      /* copy all the digits */
-      for (n = 0; n < a->used; n++) {
-         *tmpb++ = *tmpa++;
-      }
-
-      /* clear high digits */
-      for (; n < b->used; n++) {
-         *tmpb++ = 0;
-      }
+   /* copy all the digits */
+   for (n = 0; n < a->used; n++) {
+      *tmpb++ = *tmpa++;
    }
 
+   /* clear high digits */
+   MP_ZERO_DIGITS(tmpb, b->used - n);
+
    /* copy used count and sign */
    b->used = a->used;
    b->sign = a->sign;
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_count_bits.c b/libtommath/bn_mp_count_bits.c
index f7a05df..b7c2cad 100644
--- a/libtommath/bn_mp_count_bits.c
+++ b/libtommath/bn_mp_count_bits.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_COUNT_BITS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* returns the number of bits in an int */
 int mp_count_bits(const mp_int *a)
@@ -19,23 +10,19 @@ int mp_count_bits(const mp_int *a)
    mp_digit q;
 
    /* shortcut */
-   if (a->used == 0) {
+   if (MP_IS_ZERO(a)) {
       return 0;
    }
 
    /* get number of digits and add that */
-   r = (a->used - 1) * DIGIT_BIT;
+   r = (a->used - 1) * MP_DIGIT_BIT;
 
    /* take the last digit and count the bits in it */
    q = a->dp[a->used - 1];
-   while (q > (mp_digit)0) {
+   while (q > 0u) {
       ++r;
-      q >>= (mp_digit)1;
+      q >>= 1u;
    }
    return r;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_decr.c b/libtommath/bn_mp_decr.c
new file mode 100644
index 0000000..c6a1572
--- /dev/null
+++ b/libtommath/bn_mp_decr.c
@@ -0,0 +1,34 @@
+#include "tommath_private.h"
+#ifdef BN_MP_DECR_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* Decrement "a" by one like "a--". Changes input! */
+mp_err mp_decr(mp_int *a)
+{
+   if (MP_IS_ZERO(a)) {
+      mp_set(a,1uL);
+      a->sign = MP_NEG;
+      return MP_OKAY;
+   } else if (a->sign == MP_NEG) {
+      mp_err err;
+      a->sign = MP_ZPOS;
+      if ((err = mp_incr(a)) != MP_OKAY) {
+         return err;
+      }
+      /* There is no -0 in LTM */
+      if (!MP_IS_ZERO(a)) {
+         a->sign = MP_NEG;
+      }
+      return MP_OKAY;
+   } else if (a->dp[0] > 1uL) {
+      a->dp[0]--;
+      if (a->dp[0] == 0u) {
+         mp_zero(a);
+      }
+      return MP_OKAY;
+   } else {
+      return mp_sub_d(a, 1uL,a);
+   }
+}
+#endif
diff --git a/libtommath/bn_mp_div.c b/libtommath/bn_mp_div.c
index 44e3cb9..71de55b 100644
--- a/libtommath/bn_mp_div.c
+++ b/libtommath/bn_mp_div.c
@@ -1,69 +1,55 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DIV_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #ifdef BN_MP_DIV_SMALL
 
 /* slower bit-bang division... also smaller */
-int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
+mp_err mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
 {
    mp_int ta, tb, tq, q;
-   int    res, n, n2;
+   int     n, n2;
+   mp_err err;
 
    /* is divisor zero ? */
-   if (mp_iszero(b) == MP_YES) {
+   if (MP_IS_ZERO(b)) {
       return MP_VAL;
    }
 
    /* if a < b then q=0, r = a */
    if (mp_cmp_mag(a, b) == MP_LT) {
       if (d != NULL) {
-         res = mp_copy(a, d);
+         err = mp_copy(a, d);
       } else {
-         res = MP_OKAY;
+         err = MP_OKAY;
       }
       if (c != NULL) {
          mp_zero(c);
       }
-      return res;
+      return err;
    }
 
    /* init our temps */
-   if ((res = mp_init_multi(&ta, &tb, &tq, &q, NULL)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_multi(&ta, &tb, &tq, &q, NULL)) != MP_OKAY) {
+      return err;
    }
 
 
    mp_set(&tq, 1uL);
    n = mp_count_bits(a) - mp_count_bits(b);
-   if (((res = mp_abs(a, &ta)) != MP_OKAY) ||
-       ((res = mp_abs(b, &tb)) != MP_OKAY) ||
-       ((res = mp_mul_2d(&tb, n, &tb)) != MP_OKAY) ||
-       ((res = mp_mul_2d(&tq, n, &tq)) != MP_OKAY)) {
-      goto LBL_ERR;
-   }
+   if ((err = mp_abs(a, &ta)) != MP_OKAY)                         goto LBL_ERR;
+   if ((err = mp_abs(b, &tb)) != MP_OKAY)                         goto LBL_ERR;
+   if ((err = mp_mul_2d(&tb, n, &tb)) != MP_OKAY)                 goto LBL_ERR;
+   if ((err = mp_mul_2d(&tq, n, &tq)) != MP_OKAY)                 goto LBL_ERR;
 
    while (n-- >= 0) {
       if (mp_cmp(&tb, &ta) != MP_GT) {
-         if (((res = mp_sub(&ta, &tb, &ta)) != MP_OKAY) ||
-             ((res = mp_add(&q, &tq, &q)) != MP_OKAY)) {
-            goto LBL_ERR;
-         }
-      }
-      if (((res = mp_div_2d(&tb, 1, &tb, NULL)) != MP_OKAY) ||
-          ((res = mp_div_2d(&tq, 1, &tq, NULL)) != MP_OKAY)) {
-         goto LBL_ERR;
+         if ((err = mp_sub(&ta, &tb, &ta)) != MP_OKAY)            goto LBL_ERR;
+         if ((err = mp_add(&q, &tq, &q)) != MP_OKAY)              goto LBL_ERR;
       }
+      if ((err = mp_div_2d(&tb, 1, &tb, NULL)) != MP_OKAY)        goto LBL_ERR;
+      if ((err = mp_div_2d(&tq, 1, &tq, NULL)) != MP_OKAY)        goto LBL_ERR;
    }
 
    /* now q == quotient and ta == remainder */
@@ -71,15 +57,15 @@ int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
    n2 = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;
    if (c != NULL) {
       mp_exch(c, &q);
-      c->sign  = (mp_iszero(c) == MP_YES) ? MP_ZPOS : n2;
+      c->sign  = MP_IS_ZERO(c) ? MP_ZPOS : n2;
    }
    if (d != NULL) {
       mp_exch(d, &ta);
-      d->sign = (mp_iszero(d) == MP_YES) ? MP_ZPOS : n;
+      d->sign = MP_IS_ZERO(d) ? MP_ZPOS : n;
    }
 LBL_ERR:
    mp_clear_multi(&ta, &tb, &tq, &q, NULL);
-   return res;
+   return err;
 }
 
 #else
@@ -97,64 +83,54 @@ LBL_ERR:
  * The overall algorithm is as described as
  * 14.20 from HAC but fixed to treat these cases.
 */
-int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
+mp_err mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
 {
    mp_int  q, x, y, t1, t2;
-   int     res, n, t, i, norm, neg;
+   int     n, t, i, norm;
+   mp_sign neg;
+   mp_err  err;
 
    /* is divisor zero ? */
-   if (mp_iszero(b) == MP_YES) {
+   if (MP_IS_ZERO(b)) {
       return MP_VAL;
    }
 
    /* if a < b then q=0, r = a */
    if (mp_cmp_mag(a, b) == MP_LT) {
       if (d != NULL) {
-         res = mp_copy(a, d);
+         err = mp_copy(a, d);
       } else {
-         res = MP_OKAY;
+         err = MP_OKAY;
       }
       if (c != NULL) {
          mp_zero(c);
       }
-      return res;
+      return err;
    }
 
-   if ((res = mp_init_size(&q, a->used + 2)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&q, a->used + 2)) != MP_OKAY) {
+      return err;
    }
    q.used = a->used + 2;
 
-   if ((res = mp_init(&t1)) != MP_OKAY) {
-      goto LBL_Q;
-   }
+   if ((err = mp_init(&t1)) != MP_OKAY)                           goto LBL_Q;
 
-   if ((res = mp_init(&t2)) != MP_OKAY) {
-      goto LBL_T1;
-   }
+   if ((err = mp_init(&t2)) != MP_OKAY)                           goto LBL_T1;
 
-   if ((res = mp_init_copy(&x, a)) != MP_OKAY) {
-      goto LBL_T2;
-   }
+   if ((err = mp_init_copy(&x, a)) != MP_OKAY)                    goto LBL_T2;
 
-   if ((res = mp_init_copy(&y, b)) != MP_OKAY) {
-      goto LBL_X;
-   }
+   if ((err = mp_init_copy(&y, b)) != MP_OKAY)                    goto LBL_X;
 
    /* fix the sign */
    neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;
    x.sign = y.sign = MP_ZPOS;
 
-   /* normalize both x and y, ensure that y >= b/2, [b == 2**DIGIT_BIT] */
-   norm = mp_count_bits(&y) % DIGIT_BIT;
-   if (norm < (DIGIT_BIT - 1)) {
-      norm = (DIGIT_BIT - 1) - norm;
-      if ((res = mp_mul_2d(&x, norm, &x)) != MP_OKAY) {
-         goto LBL_Y;
-      }
-      if ((res = mp_mul_2d(&y, norm, &y)) != MP_OKAY) {
-         goto LBL_Y;
-      }
+   /* normalize both x and y, ensure that y >= b/2, [b == 2**MP_DIGIT_BIT] */
+   norm = mp_count_bits(&y) % MP_DIGIT_BIT;
+   if (norm < (MP_DIGIT_BIT - 1)) {
+      norm = (MP_DIGIT_BIT - 1) - norm;
+      if ((err = mp_mul_2d(&x, norm, &x)) != MP_OKAY)             goto LBL_Y;
+      if ((err = mp_mul_2d(&y, norm, &y)) != MP_OKAY)             goto LBL_Y;
    } else {
       norm = 0;
    }
@@ -164,15 +140,12 @@ int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
    t = y.used - 1;
 
    /* while (x >= y*b**n-t) do { q[n-t] += 1; x -= y*b**{n-t} } */
-   if ((res = mp_lshd(&y, n - t)) != MP_OKAY) { /* y = y*b**{n-t} */
-      goto LBL_Y;
-   }
+   /* y = y*b**{n-t} */
+   if ((err = mp_lshd(&y, n - t)) != MP_OKAY)                     goto LBL_Y;
 
    while (mp_cmp(&x, &y) != MP_LT) {
       ++(q.dp[n - t]);
-      if ((res = mp_sub(&x, &y, &x)) != MP_OKAY) {
-         goto LBL_Y;
-      }
+      if ((err = mp_sub(&x, &y, &x)) != MP_OKAY)                  goto LBL_Y;
    }
 
    /* reset y by shifting it back down */
@@ -187,10 +160,10 @@ int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
       /* step 3.1 if xi == yt then set q{i-t-1} to b-1,
        * otherwise set q{i-t-1} to (xi*b + x{i-1})/yt */
       if (x.dp[i] == y.dp[t]) {
-         q.dp[(i - t) - 1] = ((mp_digit)1 << (mp_digit)DIGIT_BIT) - (mp_digit)1;
+         q.dp[(i - t) - 1] = ((mp_digit)1 << (mp_digit)MP_DIGIT_BIT) - (mp_digit)1;
       } else {
          mp_word tmp;
-         tmp = (mp_word)x.dp[i] << (mp_word)DIGIT_BIT;
+         tmp = (mp_word)x.dp[i] << (mp_word)MP_DIGIT_BIT;
          tmp |= (mp_word)x.dp[i - 1];
          tmp /= (mp_word)y.dp[t];
          if (tmp > (mp_word)MP_MASK) {
@@ -213,41 +186,27 @@ int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
          t1.dp[0] = ((t - 1) < 0) ? 0u : y.dp[t - 1];
          t1.dp[1] = y.dp[t];
          t1.used = 2;
-         if ((res = mp_mul_d(&t1, q.dp[(i - t) - 1], &t1)) != MP_OKAY) {
-            goto LBL_Y;
-         }
+         if ((err = mp_mul_d(&t1, q.dp[(i - t) - 1], &t1)) != MP_OKAY) goto LBL_Y;
 
          /* find right hand */
          t2.dp[0] = ((i - 2) < 0) ? 0u : x.dp[i - 2];
-         t2.dp[1] = ((i - 1) < 0) ? 0u : x.dp[i - 1];
+         t2.dp[1] = x.dp[i - 1]; /* i >= 1 always holds */
          t2.dp[2] = x.dp[i];
          t2.used = 3;
       } while (mp_cmp_mag(&t1, &t2) == MP_GT);
 
       /* step 3.3 x = x - q{i-t-1} * y * b**{i-t-1} */
-      if ((res = mp_mul_d(&y, q.dp[(i - t) - 1], &t1)) != MP_OKAY) {
-         goto LBL_Y;
-      }
+      if ((err = mp_mul_d(&y, q.dp[(i - t) - 1], &t1)) != MP_OKAY) goto LBL_Y;
 
-      if ((res = mp_lshd(&t1, (i - t) - 1)) != MP_OKAY) {
-         goto LBL_Y;
-      }
+      if ((err = mp_lshd(&t1, (i - t) - 1)) != MP_OKAY)           goto LBL_Y;
 
-      if ((res = mp_sub(&x, &t1, &x)) != MP_OKAY) {
-         goto LBL_Y;
-      }
+      if ((err = mp_sub(&x, &t1, &x)) != MP_OKAY)                 goto LBL_Y;
 
       /* if x < 0 then { x = x + y*b**{i-t-1}; q{i-t-1} -= 1; } */
       if (x.sign == MP_NEG) {
-         if ((res = mp_copy(&y, &t1)) != MP_OKAY) {
-            goto LBL_Y;
-         }
-         if ((res = mp_lshd(&t1, (i - t) - 1)) != MP_OKAY) {
-            goto LBL_Y;
-         }
-         if ((res = mp_add(&x, &t1, &x)) != MP_OKAY) {
-            goto LBL_Y;
-         }
+         if ((err = mp_copy(&y, &t1)) != MP_OKAY)                 goto LBL_Y;
+         if ((err = mp_lshd(&t1, (i - t) - 1)) != MP_OKAY)        goto LBL_Y;
+         if ((err = mp_add(&x, &t1, &x)) != MP_OKAY)              goto LBL_Y;
 
          q.dp[(i - t) - 1] = (q.dp[(i - t) - 1] - 1uL) & MP_MASK;
       }
@@ -267,13 +226,11 @@ int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d)
    }
 
    if (d != NULL) {
-      if ((res = mp_div_2d(&x, norm, &x, NULL)) != MP_OKAY) {
-         goto LBL_Y;
-      }
+      if ((err = mp_div_2d(&x, norm, &x, NULL)) != MP_OKAY)       goto LBL_Y;
       mp_exch(&x, d);
    }
 
-   res = MP_OKAY;
+   err = MP_OKAY;
 
 LBL_Y:
    mp_clear(&y);
@@ -285,13 +242,9 @@ LBL_T1:
    mp_clear(&t1);
 LBL_Q:
    mp_clear(&q);
-   return res;
+   return err;
 }
 
 #endif
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_div_2.c b/libtommath/bn_mp_div_2.c
index e679d00..f56ea81 100644
--- a/libtommath/bn_mp_div_2.c
+++ b/libtommath/bn_mp_div_2.c
@@ -1,65 +1,49 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DIV_2_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* b = a/2 */
-int mp_div_2(const mp_int *a, mp_int *b)
+mp_err mp_div_2(const mp_int *a, mp_int *b)
 {
-   int     x, res, oldused;
+   int     x, oldused;
+   mp_digit r, rr, *tmpa, *tmpb;
+   mp_err err;
 
    /* copy */
    if (b->alloc < a->used) {
-      if ((res = mp_grow(b, a->used)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(b, a->used)) != MP_OKAY) {
+         return err;
       }
    }
 
    oldused = b->used;
    b->used = a->used;
-   {
-      mp_digit r, rr, *tmpa, *tmpb;
 
-      /* source alias */
-      tmpa = a->dp + b->used - 1;
+   /* source alias */
+   tmpa = a->dp + b->used - 1;
 
-      /* dest alias */
-      tmpb = b->dp + b->used - 1;
+   /* dest alias */
+   tmpb = b->dp + b->used - 1;
 
-      /* carry */
-      r = 0;
-      for (x = b->used - 1; x >= 0; x--) {
-         /* get the carry for the next iteration */
-         rr = *tmpa & 1u;
+   /* carry */
+   r = 0;
+   for (x = b->used - 1; x >= 0; x--) {
+      /* get the carry for the next iteration */
+      rr = *tmpa & 1u;
 
-         /* shift the current digit, add in carry and store */
-         *tmpb-- = (*tmpa-- >> 1) | (r << (DIGIT_BIT - 1));
+      /* shift the current digit, add in carry and store */
+      *tmpb-- = (*tmpa-- >> 1) | (r << (MP_DIGIT_BIT - 1));
 
-         /* forward carry to next iteration */
-         r = rr;
-      }
-
-      /* zero excess digits */
-      tmpb = b->dp + b->used;
-      for (x = b->used; x < oldused; x++) {
-         *tmpb++ = 0;
-      }
+      /* forward carry to next iteration */
+      r = rr;
    }
+
+   /* zero excess digits */
+   MP_ZERO_DIGITS(b->dp + b->used, oldused - b->used);
+
    b->sign = a->sign;
    mp_clamp(b);
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_div_2d.c b/libtommath/bn_mp_div_2d.c
index 912faaf..c47d5ce 100644
--- a/libtommath/bn_mp_div_2d.c
+++ b/libtommath/bn_mp_div_2d.c
@@ -1,52 +1,44 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DIV_2D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* shift right by a certain bit count (store quotient in c, optional remainder in d) */
-int mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d)
+mp_err mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d)
 {
    mp_digit D, r, rr;
-   int     x, res;
+   int     x;
+   mp_err err;
 
    /* if the shift count is <= 0 then we do no work */
    if (b <= 0) {
-      res = mp_copy(a, c);
+      err = mp_copy(a, c);
       if (d != NULL) {
          mp_zero(d);
       }
-      return res;
+      return err;
    }
 
    /* copy */
-   if ((res = mp_copy(a, c)) != MP_OKAY) {
-      return res;
+   if ((err = mp_copy(a, c)) != MP_OKAY) {
+      return err;
    }
    /* 'a' should not be used after here - it might be the same as d */
 
    /* get the remainder */
    if (d != NULL) {
-      if ((res = mp_mod_2d(a, b, d)) != MP_OKAY) {
-         return res;
+      if ((err = mp_mod_2d(a, b, d)) != MP_OKAY) {
+         return err;
       }
    }
 
    /* shift by as many digits in the bit count */
-   if (b >= DIGIT_BIT) {
-      mp_rshd(c, b / DIGIT_BIT);
+   if (b >= MP_DIGIT_BIT) {
+      mp_rshd(c, b / MP_DIGIT_BIT);
    }
 
-   /* shift any bit count < DIGIT_BIT */
-   D = (mp_digit)(b % DIGIT_BIT);
+   /* shift any bit count < MP_DIGIT_BIT */
+   D = (mp_digit)(b % MP_DIGIT_BIT);
    if (D != 0u) {
       mp_digit *tmpc, mask, shift;
 
@@ -54,7 +46,7 @@ int mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d)
       mask = ((mp_digit)1 << D) - 1uL;
 
       /* shift for lsb */
-      shift = (mp_digit)DIGIT_BIT - D;
+      shift = (mp_digit)MP_DIGIT_BIT - D;
 
       /* alias */
       tmpc = c->dp + (c->used - 1);
@@ -77,7 +69,3 @@ int mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_div_3.c b/libtommath/bn_mp_div_3.c
index 33a3432..3a23fdf 100644
--- a/libtommath/bn_mp_div_3.c
+++ b/libtommath/bn_mp_div_3.c
@@ -1,41 +1,33 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DIV_3_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* divide by three (based on routine from MPI and the GMP manual) */
-int mp_div_3(const mp_int *a, mp_int *c, mp_digit *d)
+mp_err mp_div_3(const mp_int *a, mp_int *c, mp_digit *d)
 {
    mp_int   q;
    mp_word  w, t;
    mp_digit b;
-   int      res, ix;
+   mp_err   err;
+   int      ix;
 
-   /* b = 2**DIGIT_BIT / 3 */
-   b = ((mp_word)1 << (mp_word)DIGIT_BIT) / (mp_word)3;
+   /* b = 2**MP_DIGIT_BIT / 3 */
+   b = ((mp_word)1 << (mp_word)MP_DIGIT_BIT) / (mp_word)3;
 
-   if ((res = mp_init_size(&q, a->used)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&q, a->used)) != MP_OKAY) {
+      return err;
    }
 
    q.used = a->used;
    q.sign = a->sign;
    w = 0;
    for (ix = a->used - 1; ix >= 0; ix--) {
-      w = (w << (mp_word)DIGIT_BIT) | (mp_word)a->dp[ix];
+      w = (w << (mp_word)MP_DIGIT_BIT) | (mp_word)a->dp[ix];
 
       if (w >= 3u) {
          /* multiply w by [1/3] */
-         t = (w * (mp_word)b) >> (mp_word)DIGIT_BIT;
+         t = (w * (mp_word)b) >> (mp_word)MP_DIGIT_BIT;
 
          /* now subtract 3 * [w/3] from w, to get the remainder */
          w -= t+t+t;
@@ -65,11 +57,7 @@ int mp_div_3(const mp_int *a, mp_int *c, mp_digit *d)
    }
    mp_clear(&q);
 
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_div_d.c b/libtommath/bn_mp_div_d.c
index 943521d..b9d718b 100644
--- a/libtommath/bn_mp_div_d.c
+++ b/libtommath/bn_mp_div_d.c
@@ -1,24 +1,16 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DIV_D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* single digit division (based on routine from MPI) */
-int mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d)
+mp_err mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d)
 {
    mp_int  q;
    mp_word w;
    mp_digit t;
-   int     res, ix;
+   mp_err err;
+   int ix;
 
    /* cannot divide by zero */
    if (b == 0u) {
@@ -26,7 +18,7 @@ int mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d)
    }
 
    /* quick outs */
-   if ((b == 1u) || (mp_iszero(a) == MP_YES)) {
+   if ((b == 1u) || MP_IS_ZERO(a)) {
       if (d != NULL) {
          *d = 0;
       }
@@ -37,11 +29,10 @@ int mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d)
    }
 
    /* power of two ? */
-   if (((b & (b-1)) == 0)) {
-      for (ix = 1; ix < DIGIT_BIT; ix++) {
-         if (b == (((mp_digit)1)<<ix)) {
-            break;
-         }
+   if ((b & (b - 1u)) == 0u) {
+      ix = 1;
+      while ((ix < MP_DIGIT_BIT) && (b != (((mp_digit)1)<<ix))) {
+         ix++;
       }
       if (d != NULL) {
          *d = a->dp[0] & (((mp_digit)1<<(mp_digit)ix) - 1uL);
@@ -52,23 +43,21 @@ int mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d)
       return MP_OKAY;
    }
 
-#ifdef BN_MP_DIV_3_C
    /* three? */
-   if (b == 3u) {
+   if (MP_HAS(MP_DIV_3) && (b == 3u)) {
       return mp_div_3(a, c, d);
    }
-#endif
 
    /* no easy answer [c'est la vie].  Just division */
-   if ((res = mp_init_size(&q, a->used)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&q, a->used)) != MP_OKAY) {
+      return err;
    }
 
    q.used = a->used;
    q.sign = a->sign;
    w = 0;
    for (ix = a->used - 1; ix >= 0; ix--) {
-      w = (w << (mp_word)DIGIT_BIT) | (mp_word)a->dp[ix];
+      w = (w << (mp_word)MP_DIGIT_BIT) | (mp_word)a->dp[ix];
 
       if (w >= b) {
          t = (mp_digit)(w / b);
@@ -89,11 +78,7 @@ int mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d)
    }
    mp_clear(&q);
 
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_dr_is_modulus.c b/libtommath/bn_mp_dr_is_modulus.c
index 4d0c8ce..83760ea 100644
--- a/libtommath/bn_mp_dr_is_modulus.c
+++ b/libtommath/bn_mp_dr_is_modulus.c
@@ -1,25 +1,16 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DR_IS_MODULUS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* determines if a number is a valid DR modulus */
-int mp_dr_is_modulus(const mp_int *a)
+mp_bool mp_dr_is_modulus(const mp_int *a)
 {
    int ix;
 
    /* must be at least two digits */
    if (a->used < 2) {
-      return 0;
+      return MP_NO;
    }
 
    /* must be of the form b**k - a [a <= b] so all
@@ -27,14 +18,10 @@ int mp_dr_is_modulus(const mp_int *a)
     */
    for (ix = 1; ix < a->used; ix++) {
       if (a->dp[ix] != MP_MASK) {
-         return 0;
+         return MP_NO;
       }
    }
-   return 1;
+   return MP_YES;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_dr_reduce.c b/libtommath/bn_mp_dr_reduce.c
index da24d17..ffc33a6 100644
--- a/libtommath/bn_mp_dr_reduce.c
+++ b/libtommath/bn_mp_dr_reduce.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DR_REDUCE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* reduce "x" in place modulo "n" using the Diminished Radix algorithm.
  *
@@ -26,9 +17,10 @@
  *
  * Input x must be in the range 0 <= x <= (n-1)**2
  */
-int mp_dr_reduce(mp_int *x, const mp_int *n, mp_digit k)
+mp_err mp_dr_reduce(mp_int *x, const mp_int *n, mp_digit k)
 {
-   int      err, i, m;
+   mp_err      err;
+   int i, m;
    mp_word  r;
    mp_digit mu, *tmpx1, *tmpx2;
 
@@ -60,16 +52,14 @@ top:
    for (i = 0; i < m; i++) {
       r         = ((mp_word)*tmpx2++ * (mp_word)k) + *tmpx1 + mu;
       *tmpx1++  = (mp_digit)(r & MP_MASK);
-      mu        = (mp_digit)(r >> ((mp_word)DIGIT_BIT));
+      mu        = (mp_digit)(r >> ((mp_word)MP_DIGIT_BIT));
    }
 
    /* set final carry */
    *tmpx1++ = mu;
 
    /* zero words above m */
-   for (i = m + 1; i < x->used; i++) {
-      *tmpx1++ = 0;
-   }
+   MP_ZERO_DIGITS(tmpx1, (x->used - m) - 1);
 
    /* clamp, sub and return */
    mp_clamp(x);
@@ -86,7 +76,3 @@ top:
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_dr_setup.c b/libtommath/bn_mp_dr_setup.c
index f8c7e7e..32d5f38 100644
--- a/libtommath/bn_mp_dr_setup.c
+++ b/libtommath/bn_mp_dr_setup.c
@@ -1,28 +1,15 @@
 #include "tommath_private.h"
 #ifdef BN_MP_DR_SETUP_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* determines the setup value */
 void mp_dr_setup(const mp_int *a, mp_digit *d)
 {
-   /* the casts are required if DIGIT_BIT is one less than
-    * the number of bits in a mp_digit [e.g. DIGIT_BIT==31]
+   /* the casts are required if MP_DIGIT_BIT is one less than
+    * the number of bits in a mp_digit [e.g. MP_DIGIT_BIT==31]
     */
-   *d = (mp_digit)(((mp_word)1 << (mp_word)DIGIT_BIT) - (mp_word)a->dp[0]);
+   *d = (mp_digit)(((mp_word)1 << (mp_word)MP_DIGIT_BIT) - (mp_word)a->dp[0]);
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_error_to_string.c b/libtommath/bn_mp_error_to_string.c
new file mode 100644
index 0000000..2e2adb0
--- /dev/null
+++ b/libtommath/bn_mp_error_to_string.c
@@ -0,0 +1,27 @@
+#include "tommath_private.h"
+#ifdef BN_MP_ERROR_TO_STRING_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* return a char * string for a given code */
+const char *mp_error_to_string(mp_err code)
+{
+   switch (code) {
+   case MP_OKAY:
+      return "Successful";
+   case MP_ERR:
+      return "Unknown error";
+   case MP_MEM:
+      return "Out of heap";
+   case MP_VAL:
+      return "Value out of range";
+   case MP_ITER:
+      return "Max. iterations reached";
+   case MP_BUF:
+      return "Buffer overflow";
+   default:
+      return "Invalid error code";
+   }
+}
+
+#endif
diff --git a/libtommath/bn_mp_exch.c b/libtommath/bn_mp_exch.c
index 2f33877..552094c 100644
--- a/libtommath/bn_mp_exch.c
+++ b/libtommath/bn_mp_exch.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_EXCH_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* swap the elements of two integers, for cases where you can't simply swap the
  * mp_int pointers around
@@ -24,7 +15,3 @@ void mp_exch(mp_int *a, mp_int *b)
    *b = t;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_export.c b/libtommath/bn_mp_export.c
deleted file mode 100644
index ea48e90..0000000
--- a/libtommath/bn_mp_export.c
+++ /dev/null
@@ -1,84 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_EXPORT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* based on gmp's mpz_export.
- * see http://gmplib.org/manual/Integer-Import-and-Export.html
- */
-int mp_export(void *rop, size_t *countp, int order, size_t size,
-              int endian, size_t nails, const mp_int *op)
-{
-   int result;
-   size_t odd_nails, nail_bytes, i, j, bits, count;
-   unsigned char odd_nail_mask;
-
-   mp_int t;
-
-   if ((result = mp_init_copy(&t, op)) != MP_OKAY) {
-      return result;
-   }
-
-   if (endian == 0) {
-      union {
-         unsigned int i;
-         char c[4];
-      } lint;
-      lint.i = 0x01020304;
-
-      endian = (lint.c[0] == '\x04') ? -1 : 1;
-   }
-
-   odd_nails = (nails % 8u);
-   odd_nail_mask = 0xff;
-   for (i = 0; i < odd_nails; ++i) {
-      odd_nail_mask ^= (unsigned char)(1u << (7u - i));
-   }
-   nail_bytes = nails / 8u;
-
-   bits = (size_t)mp_count_bits(&t);
-   count = (bits / ((size * 8u) - nails)) + (((bits % ((size * 8u) - nails)) != 0u) ? 1u : 0u);
-
-   for (i = 0; i < count; ++i) {
-      for (j = 0; j < size; ++j) {
-         unsigned char *byte = (unsigned char *)rop +
-                               (((order == -1) ? i : ((count - 1u) - i)) * size) +
-                               ((endian == -1) ? j : ((size - 1u) - j));
-
-         if (j >= (size - nail_bytes)) {
-            *byte = 0;
-            continue;
-         }
-
-         *byte = (unsigned char)((j == ((size - nail_bytes) - 1u)) ? (t.dp[0] & odd_nail_mask) : (t.dp[0] & 0xFFuL));
-
-         if ((result = mp_div_2d(&t, (j == ((size - nail_bytes) - 1u)) ? (int)(8u - odd_nails) : 8, &t, NULL)) != MP_OKAY) {
-            mp_clear(&t);
-            return result;
-         }
-      }
-   }
-
-   mp_clear(&t);
-
-   if (countp != NULL) {
-      *countp = count;
-   }
-
-   return MP_OKAY;
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_expt_d.c b/libtommath/bn_mp_expt_d.c
deleted file mode 100644
index e0df09c..0000000
--- a/libtommath/bn_mp_expt_d.c
+++ /dev/null
@@ -1,25 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_EXPT_D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* wrapper function for mp_expt_d_ex() */
-int mp_expt_d(const mp_int *a, mp_digit b, mp_int *c)
-{
-   return mp_expt_d_ex(a, b, c, 0);
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_expt_d_ex.c b/libtommath/bn_mp_expt_d_ex.c
deleted file mode 100644
index 5a6f7b2..0000000
--- a/libtommath/bn_mp_expt_d_ex.c
+++ /dev/null
@@ -1,79 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_EXPT_D_EX_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* calculate c = a**b  using a square-multiply algorithm */
-int mp_expt_d_ex(const mp_int *a, mp_digit b, mp_int *c, int fast)
-{
-   int     res;
-   unsigned int x;
-
-   mp_int  g;
-
-   if ((res = mp_init_copy(&g, a)) != MP_OKAY) {
-      return res;
-   }
-
-   /* set initial result */
-   mp_set(c, 1uL);
-
-   if (fast != 0) {
-      while (b > 0u) {
-         /* if the bit is set multiply */
-         if ((b & 1u) != 0u) {
-            if ((res = mp_mul(c, &g, c)) != MP_OKAY) {
-               mp_clear(&g);
-               return res;
-            }
-         }
-
-         /* square */
-         if (b > 1u) {
-            if ((res = mp_sqr(&g, &g)) != MP_OKAY) {
-               mp_clear(&g);
-               return res;
-            }
-         }
-
-         /* shift to next bit */
-         b >>= 1;
-      }
-   } else {
-      for (x = 0; x < (unsigned)DIGIT_BIT; x++) {
-         /* square */
-         if ((res = mp_sqr(c, c)) != MP_OKAY) {
-            mp_clear(&g);
-            return res;
-         }
-
-         /* if the bit is set multiply */
-         if ((b & ((mp_digit)1 << (DIGIT_BIT - 1))) != 0u) {
-            if ((res = mp_mul(c, &g, c)) != MP_OKAY) {
-               mp_clear(&g);
-               return res;
-            }
-         }
-
-         /* shift to next bit */
-         b <<= 1;
-      }
-   } /* if ... else */
-
-   mp_clear(&g);
-   return MP_OKAY;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_expt_u32.c b/libtommath/bn_mp_expt_u32.c
new file mode 100644
index 0000000..67c8fd2
--- /dev/null
+++ b/libtommath/bn_mp_expt_u32.c
@@ -0,0 +1,46 @@
+#include "tommath_private.h"
+#ifdef BN_MP_EXPT_U32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* calculate c = a**b  using a square-multiply algorithm */
+mp_err mp_expt_u32(const mp_int *a, unsigned int b, mp_int *c)
+{
+   mp_err err;
+
+   mp_int  g;
+
+   if ((err = mp_init_copy(&g, a)) != MP_OKAY) {
+      return err;
+   }
+
+   /* set initial result */
+   mp_set(c, 1uL);
+
+   while (b > 0u) {
+      /* if the bit is set multiply */
+      if ((b & 1u) != 0u) {
+         if ((err = mp_mul(c, &g, c)) != MP_OKAY) {
+            goto LBL_ERR;
+         }
+      }
+
+      /* square */
+      if (b > 1u) {
+         if ((err = mp_sqr(&g, &g)) != MP_OKAY) {
+            goto LBL_ERR;
+         }
+      }
+
+      /* shift to next bit */
+      b >>= 1;
+   }
+
+   err = MP_OKAY;
+
+LBL_ERR:
+   mp_clear(&g);
+   return err;
+}
+
+#endif
diff --git a/libtommath/bn_mp_exptmod.c b/libtommath/bn_mp_exptmod.c
index c400b7e..5f811eb 100644
--- a/libtommath/bn_mp_exptmod.c
+++ b/libtommath/bn_mp_exptmod.c
@@ -1,24 +1,14 @@
 #include "tommath_private.h"
 #ifdef BN_MP_EXPTMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* this is a shell function that calls either the normal or Montgomery
  * exptmod functions.  Originally the call to the montgomery code was
  * embedded in the normal function but that wasted alot of stack space
  * for nothing (since 99% of the time the Montgomery code would be called)
  */
-int mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y)
+mp_err mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y)
 {
    int dr;
 
@@ -29,81 +19,58 @@ int mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y)
 
    /* if exponent X is negative we have to recurse */
    if (X->sign == MP_NEG) {
-#ifdef BN_MP_INVMOD_C
       mp_int tmpG, tmpX;
-      int err;
+      mp_err err;
 
-      /* first compute 1/G mod P */
-      if ((err = mp_init(&tmpG)) != MP_OKAY) {
+      if (!MP_HAS(MP_INVMOD)) {
+         return MP_VAL;
+      }
+
+      if ((err = mp_init_multi(&tmpG, &tmpX, NULL)) != MP_OKAY) {
          return err;
       }
+
+      /* first compute 1/G mod P */
       if ((err = mp_invmod(G, P, &tmpG)) != MP_OKAY) {
-         mp_clear(&tmpG);
-         return err;
+         goto LBL_ERR;
       }
 
       /* now get |X| */
-      if ((err = mp_init(&tmpX)) != MP_OKAY) {
-         mp_clear(&tmpG);
-         return err;
-      }
       if ((err = mp_abs(X, &tmpX)) != MP_OKAY) {
-         mp_clear_multi(&tmpG, &tmpX, NULL);
-         return err;
+         goto LBL_ERR;
       }
 
       /* and now compute (1/G)**|X| instead of G**X [X < 0] */
       err = mp_exptmod(&tmpG, &tmpX, P, Y);
+LBL_ERR:
       mp_clear_multi(&tmpG, &tmpX, NULL);
       return err;
-#else
-      /* no invmod */
-      return MP_VAL;
-#endif
    }
 
    /* modified diminished radix reduction */
-#if defined(BN_MP_REDUCE_IS_2K_L_C) && defined(BN_MP_REDUCE_2K_L_C) && defined(BN_S_MP_EXPTMOD_C)
-   if (mp_reduce_is_2k_l(P) == MP_YES) {
+   if (MP_HAS(MP_REDUCE_IS_2K_L) && MP_HAS(MP_REDUCE_2K_L) && MP_HAS(S_MP_EXPTMOD) &&
+       (mp_reduce_is_2k_l(P) == MP_YES)) {
       return s_mp_exptmod(G, X, P, Y, 1);
    }
-#endif
 
-#ifdef BN_MP_DR_IS_MODULUS_C
-   /* is it a DR modulus? */
-   dr = mp_dr_is_modulus(P);
-#else
-   /* default to no */
-   dr = 0;
-#endif
+   /* is it a DR modulus? default to no */
+   dr = (MP_HAS(MP_DR_IS_MODULUS) && (mp_dr_is_modulus(P) == MP_YES)) ? 1 : 0;
 
-#ifdef BN_MP_REDUCE_IS_2K_C
    /* if not, is it a unrestricted DR modulus? */
-   if (dr == 0) {
-      dr = mp_reduce_is_2k(P) << 1;
+   if (MP_HAS(MP_REDUCE_IS_2K) && (dr == 0)) {
+      dr = (mp_reduce_is_2k(P) == MP_YES) ? 2 : 0;
    }
-#endif
 
    /* if the modulus is odd or dr != 0 use the montgomery method */
-#ifdef BN_MP_EXPTMOD_FAST_C
-   if ((mp_isodd(P) == MP_YES) || (dr !=  0)) {
-      return mp_exptmod_fast(G, X, P, Y, dr);
-   } else {
-#endif
-#ifdef BN_S_MP_EXPTMOD_C
+   if (MP_HAS(S_MP_EXPTMOD_FAST) && (MP_IS_ODD(P) || (dr != 0))) {
+      return s_mp_exptmod_fast(G, X, P, Y, dr);
+   } else if (MP_HAS(S_MP_EXPTMOD)) {
       /* otherwise use the generic Barrett reduction technique */
       return s_mp_exptmod(G, X, P, Y, 0);
-#else
+   } else {
       /* no exptmod for evens */
       return MP_VAL;
-#endif
-#ifdef BN_MP_EXPTMOD_FAST_C
    }
-#endif
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_exteuclid.c b/libtommath/bn_mp_exteuclid.c
index c23a6c1..faf47ba 100644
--- a/libtommath/bn_mp_exteuclid.c
+++ b/libtommath/bn_mp_exteuclid.c
@@ -1,24 +1,15 @@
 #include "tommath_private.h"
 #ifdef BN_MP_EXTEUCLID_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* Extended euclidean algorithm of (a, b) produces
    a*u1 + b*u2 = u3
  */
-int mp_exteuclid(const mp_int *a, const mp_int *b, mp_int *U1, mp_int *U2, mp_int *U3)
+mp_err mp_exteuclid(const mp_int *a, const mp_int *b, mp_int *U1, mp_int *U2, mp_int *U3)
 {
    mp_int u1, u2, u3, v1, v2, v3, t1, t2, t3, q, tmp;
-   int err;
+   mp_err err;
 
    if ((err = mp_init_multi(&u1, &u2, &u3, &v1, &v2, &v3, &t1, &t2, &t3, &q, &tmp, NULL)) != MP_OKAY) {
       return err;
@@ -26,77 +17,41 @@ int mp_exteuclid(const mp_int *a, const mp_int *b, mp_int *U1, mp_int *U2, mp_in
 
    /* initialize, (u1,u2,u3) = (1,0,a) */
    mp_set(&u1, 1uL);
-   if ((err = mp_copy(a, &u3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
+   if ((err = mp_copy(a, &u3)) != MP_OKAY)                        goto LBL_ERR;
 
    /* initialize, (v1,v2,v3) = (0,1,b) */
    mp_set(&v2, 1uL);
-   if ((err = mp_copy(b, &v3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
+   if ((err = mp_copy(b, &v3)) != MP_OKAY)                        goto LBL_ERR;
 
    /* loop while v3 != 0 */
-   while (mp_iszero(&v3) == MP_NO) {
+   while (!MP_IS_ZERO(&v3)) {
       /* q = u3/v3 */
-      if ((err = mp_div(&u3, &v3, &q, NULL)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
+      if ((err = mp_div(&u3, &v3, &q, NULL)) != MP_OKAY)          goto LBL_ERR;
 
       /* (t1,t2,t3) = (u1,u2,u3) - (v1,v2,v3)q */
-      if ((err = mp_mul(&v1, &q, &tmp)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_sub(&u1, &tmp, &t1)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_mul(&v2, &q, &tmp)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_sub(&u2, &tmp, &t2)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_mul(&v3, &q, &tmp)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_sub(&u3, &tmp, &t3)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
+      if ((err = mp_mul(&v1, &q, &tmp)) != MP_OKAY)               goto LBL_ERR;
+      if ((err = mp_sub(&u1, &tmp, &t1)) != MP_OKAY)              goto LBL_ERR;
+      if ((err = mp_mul(&v2, &q, &tmp)) != MP_OKAY)               goto LBL_ERR;
+      if ((err = mp_sub(&u2, &tmp, &t2)) != MP_OKAY)              goto LBL_ERR;
+      if ((err = mp_mul(&v3, &q, &tmp)) != MP_OKAY)               goto LBL_ERR;
+      if ((err = mp_sub(&u3, &tmp, &t3)) != MP_OKAY)              goto LBL_ERR;
 
       /* (u1,u2,u3) = (v1,v2,v3) */
-      if ((err = mp_copy(&v1, &u1)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_copy(&v2, &u2)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_copy(&v3, &u3)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
+      if ((err = mp_copy(&v1, &u1)) != MP_OKAY)                   goto LBL_ERR;
+      if ((err = mp_copy(&v2, &u2)) != MP_OKAY)                   goto LBL_ERR;
+      if ((err = mp_copy(&v3, &u3)) != MP_OKAY)                   goto LBL_ERR;
 
       /* (v1,v2,v3) = (t1,t2,t3) */
-      if ((err = mp_copy(&t1, &v1)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_copy(&t2, &v2)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_copy(&t3, &v3)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
+      if ((err = mp_copy(&t1, &v1)) != MP_OKAY)                   goto LBL_ERR;
+      if ((err = mp_copy(&t2, &v2)) != MP_OKAY)                   goto LBL_ERR;
+      if ((err = mp_copy(&t3, &v3)) != MP_OKAY)                   goto LBL_ERR;
    }
 
    /* make sure U3 >= 0 */
    if (u3.sign == MP_NEG) {
-      if ((err = mp_neg(&u1, &u1)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_neg(&u2, &u2)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((err = mp_neg(&u3, &u3)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
+      if ((err = mp_neg(&u1, &u1)) != MP_OKAY)                    goto LBL_ERR;
+      if ((err = mp_neg(&u2, &u2)) != MP_OKAY)                    goto LBL_ERR;
+      if ((err = mp_neg(&u3, &u3)) != MP_OKAY)                    goto LBL_ERR;
    }
 
    /* copy result out */
@@ -116,7 +71,3 @@ LBL_ERR:
    return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_fread.c b/libtommath/bn_mp_fread.c
index 9c935cb..52ea773 100644
--- a/libtommath/bn_mp_fread.c
+++ b/libtommath/bn_mp_fread.c
@@ -1,29 +1,17 @@
 #include "tommath_private.h"
 #ifdef BN_MP_FREAD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
-#ifndef LTM_NO_FILE
+#ifndef MP_NO_FILE
 /* read a bigint from a file stream in ASCII */
-int mp_fread(mp_int *a, int radix, FILE *stream)
+mp_err mp_fread(mp_int *a, int radix, FILE *stream)
 {
-   int err, ch, neg, y;
-   unsigned pos;
-
-   /* clear a */
-   mp_zero(a);
+   mp_err err;
+   mp_sign neg;
 
    /* if first digit is - then set negative */
-   ch = fgetc(stream);
+   int ch = fgetc(stream);
    if (ch == (int)'-') {
       neg = MP_NEG;
       ch = fgetc(stream);
@@ -31,8 +19,17 @@ int mp_fread(mp_int *a, int radix, FILE *stream)
       neg = MP_ZPOS;
    }
 
-   for (;;) {
-      pos = (unsigned)(ch - (int)'(');
+   /* no digits, return error */
+   if (ch == EOF) {
+      return MP_ERR;
+   }
+
+   /* clear a */
+   mp_zero(a);
+
+   do {
+      int y;
+      unsigned pos = (unsigned)(ch - (int)'(');
       if (mp_s_rmap_reverse_sz < pos) {
          break;
       }
@@ -50,10 +47,9 @@ int mp_fread(mp_int *a, int radix, FILE *stream)
       if ((err = mp_add_d(a, (mp_digit)y, a)) != MP_OKAY) {
          return err;
       }
+   } while ((ch = fgetc(stream)) != EOF);
 
-      ch = fgetc(stream);
-   }
-   if (mp_cmp_d(a, 0uL) != MP_EQ) {
+   if (a->used != 0) {
       a->sign = neg;
    }
 
@@ -62,7 +58,3 @@ int mp_fread(mp_int *a, int radix, FILE *stream)
 #endif
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_from_sbin.c b/libtommath/bn_mp_from_sbin.c
new file mode 100644
index 0000000..20e4597
--- /dev/null
+++ b/libtommath/bn_mp_from_sbin.c
@@ -0,0 +1,25 @@
+#include "tommath_private.h"
+#ifdef BN_MP_FROM_SBIN_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* read signed bin, big endian, first byte is 0==positive or 1==negative */
+mp_err mp_from_sbin(mp_int *a, const unsigned char *buf, size_t size)
+{
+   mp_err err;
+
+   /* read magnitude */
+   if ((err = mp_from_ubin(a, buf + 1, size - 1u)) != MP_OKAY) {
+      return err;
+   }
+
+   /* first byte is 0 for positive, non-zero for negative */
+   if (buf[0] == (unsigned char)0) {
+      a->sign = MP_ZPOS;
+   } else {
+      a->sign = MP_NEG;
+   }
+
+   return MP_OKAY;
+}
+#endif
diff --git a/libtommath/bn_mp_from_ubin.c b/libtommath/bn_mp_from_ubin.c
new file mode 100644
index 0000000..7f73cbc
--- /dev/null
+++ b/libtommath/bn_mp_from_ubin.c
@@ -0,0 +1,39 @@
+#include "tommath_private.h"
+#ifdef BN_MP_FROM_UBIN_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* reads a unsigned char array, assumes the msb is stored first [big endian] */
+mp_err mp_from_ubin(mp_int *a, const unsigned char *buf, size_t size)
+{
+   mp_err err;
+
+   /* make sure there are at least two digits */
+   if (a->alloc < 2) {
+      if ((err = mp_grow(a, 2)) != MP_OKAY) {
+         return err;
+      }
+   }
+
+   /* zero the int */
+   mp_zero(a);
+
+   /* read the bytes in */
+   while (size-- > 0u) {
+      if ((err = mp_mul_2d(a, 8, a)) != MP_OKAY) {
+         return err;
+      }
+
+#ifndef MP_8BIT
+      a->dp[0] |= *buf++;
+      a->used += 1;
+#else
+      a->dp[0] = (*buf & MP_MASK);
+      a->dp[1] |= ((*buf++ >> 7) & 1u);
+      a->used += 2;
+#endif
+   }
+   mp_clamp(a);
+   return MP_OKAY;
+}
+#endif
diff --git a/libtommath/bn_mp_fwrite.c b/libtommath/bn_mp_fwrite.c
index 85a942f..abe2e67 100644
--- a/libtommath/bn_mp_fwrite.c
+++ b/libtommath/bn_mp_fwrite.c
@@ -1,51 +1,45 @@
 #include "tommath_private.h"
 #ifdef BN_MP_FWRITE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-#ifndef LTM_NO_FILE
-int mp_fwrite(const mp_int *a, int radix, FILE *stream)
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+#ifndef MP_NO_FILE
+mp_err mp_fwrite(const mp_int *a, int radix, FILE *stream)
 {
    char *buf;
-   int err, len, x;
-
-   if ((err = mp_radix_size(a, radix, &len)) != MP_OKAY) {
-      return err;
+   mp_err err;
+   int len;
+   size_t written;
+
+   /* TODO: this function is not in this PR */
+   if (MP_HAS(MP_RADIX_SIZE_OVERESTIMATE)) {
+      /* if ((err = mp_radix_size_overestimate(&t, base, &len)) != MP_OKAY)      goto LBL_ERR; */
+   } else {
+      if ((err = mp_radix_size(a, radix, &len)) != MP_OKAY) {
+         return err;
+      }
    }
 
-   buf = (char *) XMALLOC((size_t)len);
+   buf = (char *) MP_MALLOC((size_t)len);
    if (buf == NULL) {
       return MP_MEM;
    }
 
-   if ((err = mp_toradix(a, buf, radix)) != MP_OKAY) {
-      XFREE(buf, len);
-      return err;
+   if ((err = mp_to_radix(a, buf, (size_t)len, &written, radix)) != MP_OKAY) {
+      goto LBL_ERR;
    }
 
-   for (x = 0; x < len; x++) {
-      if (fputc((int)buf[x], stream) == EOF) {
-         XFREE(buf, len);
-         return MP_VAL;
-      }
+   if (fwrite(buf, written, 1uL, stream) != 1uL) {
+      err = MP_ERR;
+      goto LBL_ERR;
    }
+   err = MP_OKAY;
+
 
-   XFREE(buf, len);
-   return MP_OKAY;
+LBL_ERR:
+   MP_FREE_BUFFER(buf, (size_t)len);
+   return err;
 }
 #endif
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_gcd.c b/libtommath/bn_mp_gcd.c
index 05030c2..53029ba 100644
--- a/libtommath/bn_mp_gcd.c
+++ b/libtommath/bn_mp_gcd.c
@@ -1,37 +1,29 @@
 #include "tommath_private.h"
 #ifdef BN_MP_GCD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* Greatest Common Divisor using the binary method */
-int mp_gcd(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_gcd(const mp_int *a, const mp_int *b, mp_int *c)
 {
    mp_int  u, v;
-   int     k, u_lsb, v_lsb, res;
+   int     k, u_lsb, v_lsb;
+   mp_err err;
 
    /* either zero than gcd is the largest */
-   if (mp_iszero(a) == MP_YES) {
+   if (MP_IS_ZERO(a)) {
       return mp_abs(b, c);
    }
-   if (mp_iszero(b) == MP_YES) {
+   if (MP_IS_ZERO(b)) {
       return mp_abs(a, c);
    }
 
    /* get copies of a and b we can modify */
-   if ((res = mp_init_copy(&u, a)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_copy(&u, a)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_init_copy(&v, b)) != MP_OKAY) {
+   if ((err = mp_init_copy(&v, b)) != MP_OKAY) {
       goto LBL_U;
    }
 
@@ -41,33 +33,33 @@ int mp_gcd(const mp_int *a, const mp_int *b, mp_int *c)
    /* B1.  Find the common power of two for u and v */
    u_lsb = mp_cnt_lsb(&u);
    v_lsb = mp_cnt_lsb(&v);
-   k     = MIN(u_lsb, v_lsb);
+   k     = MP_MIN(u_lsb, v_lsb);
 
    if (k > 0) {
       /* divide the power of two out */
-      if ((res = mp_div_2d(&u, k, &u, NULL)) != MP_OKAY) {
+      if ((err = mp_div_2d(&u, k, &u, NULL)) != MP_OKAY) {
          goto LBL_V;
       }
 
-      if ((res = mp_div_2d(&v, k, &v, NULL)) != MP_OKAY) {
+      if ((err = mp_div_2d(&v, k, &v, NULL)) != MP_OKAY) {
          goto LBL_V;
       }
    }
 
    /* divide any remaining factors of two out */
    if (u_lsb != k) {
-      if ((res = mp_div_2d(&u, u_lsb - k, &u, NULL)) != MP_OKAY) {
+      if ((err = mp_div_2d(&u, u_lsb - k, &u, NULL)) != MP_OKAY) {
          goto LBL_V;
       }
    }
 
    if (v_lsb != k) {
-      if ((res = mp_div_2d(&v, v_lsb - k, &v, NULL)) != MP_OKAY) {
+      if ((err = mp_div_2d(&v, v_lsb - k, &v, NULL)) != MP_OKAY) {
          goto LBL_V;
       }
    }
 
-   while (mp_iszero(&v) == MP_NO) {
+   while (!MP_IS_ZERO(&v)) {
       /* make sure v is the largest */
       if (mp_cmp_mag(&u, &v) == MP_GT) {
          /* swap u and v to make sure v is >= u */
@@ -75,30 +67,26 @@ int mp_gcd(const mp_int *a, const mp_int *b, mp_int *c)
       }
 
       /* subtract smallest from largest */
-      if ((res = s_mp_sub(&v, &u, &v)) != MP_OKAY) {
+      if ((err = s_mp_sub(&v, &u, &v)) != MP_OKAY) {
          goto LBL_V;
       }
 
       /* Divide out all factors of two */
-      if ((res = mp_div_2d(&v, mp_cnt_lsb(&v), &v, NULL)) != MP_OKAY) {
+      if ((err = mp_div_2d(&v, mp_cnt_lsb(&v), &v, NULL)) != MP_OKAY) {
          goto LBL_V;
       }
    }
 
    /* multiply by 2**k which we divided out at the beginning */
-   if ((res = mp_mul_2d(&u, k, c)) != MP_OKAY) {
+   if ((err = mp_mul_2d(&u, k, c)) != MP_OKAY) {
       goto LBL_V;
    }
    c->sign = MP_ZPOS;
-   res = MP_OKAY;
+   err = MP_OKAY;
 LBL_V:
    mp_clear(&u);
 LBL_U:
    mp_clear(&v);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_get_double.c b/libtommath/bn_mp_get_double.c
index 629eae3..c9b1b19 100644
--- a/libtommath/bn_mp_get_double.c
+++ b/libtommath/bn_mp_get_double.c
@@ -1,22 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_GET_DOUBLE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 double mp_get_double(const mp_int *a)
 {
    int i;
    double d = 0.0, fac = 1.0;
-   for (i = 0; i < DIGIT_BIT; ++i) {
+   for (i = 0; i < MP_DIGIT_BIT; ++i) {
       fac *= 2.0;
    }
    for (i = a->used; i --> 0;) {
@@ -25,7 +16,3 @@ double mp_get_double(const mp_int *a)
    return (a->sign == MP_NEG) ? -d : d;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_get_i32.c b/libtommath/bn_mp_get_i32.c
new file mode 100644
index 0000000..030b657
--- /dev/null
+++ b/libtommath/bn_mp_get_i32.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_I32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_SIGNED(mp_get_i32, mp_get_mag_u32, int32_t, uint32_t)
+#endif
diff --git a/libtommath/bn_mp_get_i64.c b/libtommath/bn_mp_get_i64.c
new file mode 100644
index 0000000..969c8d2
--- /dev/null
+++ b/libtommath/bn_mp_get_i64.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_I64_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_SIGNED(mp_get_i64, mp_get_mag_u64, int64_t, uint64_t)
+#endif
diff --git a/libtommath/bn_mp_get_int.c b/libtommath/bn_mp_get_int.c
deleted file mode 100644
index d9c7a11..0000000
--- a/libtommath/bn_mp_get_int.c
+++ /dev/null
@@ -1,25 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_GET_INT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* get the lower 32-bits of an mp_int */
-unsigned long mp_get_int(const mp_int *a)
-{
-   /* force result to 32-bits always so it is consistent on non 32-bit platforms */
-   return mp_get_long(a) & 0xFFFFFFFFUL;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_get_l.c b/libtommath/bn_mp_get_l.c
new file mode 100644
index 0000000..55d78ec
--- /dev/null
+++ b/libtommath/bn_mp_get_l.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_L_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_SIGNED(mp_get_l, mp_get_mag_ul, long, unsigned long)
+#endif
diff --git a/libtommath/bn_mp_get_ll.c b/libtommath/bn_mp_get_ll.c
new file mode 100644
index 0000000..2687534
--- /dev/null
+++ b/libtommath/bn_mp_get_ll.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_LL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_SIGNED(mp_get_ll, mp_get_mag_ull, long long, unsigned long long)
+#endif
diff --git a/libtommath/bn_mp_get_long.c b/libtommath/bn_mp_get_long.c
deleted file mode 100644
index b95bb8a..0000000
--- a/libtommath/bn_mp_get_long.c
+++ /dev/null
@@ -1,42 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_GET_LONG_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* get the lower unsigned long of an mp_int, platform dependent */
-unsigned long mp_get_long(const mp_int *a)
-{
-   int i;
-   unsigned long res;
-
-   if (IS_ZERO(a)) {
-      return 0;
-   }
-
-   /* get number of digits of the lsb we have to read */
-   i = MIN(a->used, (((CHAR_BIT * (int)sizeof(unsigned long)) + DIGIT_BIT - 1) / DIGIT_BIT)) - 1;
-
-   /* get most significant digit of result */
-   res = (unsigned long)a->dp[i];
-
-#if (ULONG_MAX != 0xFFFFFFFFUL) || (DIGIT_BIT < 32)
-   while (--i >= 0) {
-      res = (res << DIGIT_BIT) | (unsigned long)a->dp[i];
-   }
-#endif
-   return res;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_get_long_long.c b/libtommath/bn_mp_get_long_long.c
deleted file mode 100644
index 333c6d1..0000000
--- a/libtommath/bn_mp_get_long_long.c
+++ /dev/null
@@ -1,42 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_GET_LONG_LONG_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* get the lower unsigned long long of an mp_int, platform dependent */
-Tcl_WideUInt mp_get_long_long(const mp_int *a)
-{
-   int i;
-   Tcl_WideUInt res;
-
-   if (IS_ZERO(a)) {
-      return 0;
-   }
-
-   /* get number of digits of the lsb we have to read */
-   i = MIN(a->used, (((CHAR_BIT * (int)sizeof(Tcl_WideUInt)) + DIGIT_BIT - 1) / DIGIT_BIT)) - 1;
-
-   /* get most significant digit of result */
-   res = (Tcl_WideUInt)a->dp[i];
-
-#if DIGIT_BIT < 64
-   while (--i >= 0) {
-      res = (res << DIGIT_BIT) | (Tcl_WideUInt)a->dp[i];
-   }
-#endif
-   return res;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_get_mag_u32.c b/libtommath/bn_mp_get_mag_u32.c
new file mode 100644
index 0000000..d77189b
--- /dev/null
+++ b/libtommath/bn_mp_get_mag_u32.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_MAG_U32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_MAG(mp_get_mag_u32, uint32_t)
+#endif
diff --git a/libtommath/bn_mp_get_mag_u64.c b/libtommath/bn_mp_get_mag_u64.c
new file mode 100644
index 0000000..36dd73f
--- /dev/null
+++ b/libtommath/bn_mp_get_mag_u64.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_MAG_U64_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_MAG(mp_get_mag_u64, uint64_t)
+#endif
diff --git a/libtommath/bn_mp_get_mag_ul.c b/libtommath/bn_mp_get_mag_ul.c
new file mode 100644
index 0000000..e8819ae
--- /dev/null
+++ b/libtommath/bn_mp_get_mag_ul.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_MAG_UL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_MAG(mp_get_mag_ul, unsigned long)
+#endif
diff --git a/libtommath/bn_mp_get_mag_ull.c b/libtommath/bn_mp_get_mag_ull.c
new file mode 100644
index 0000000..63a2741
--- /dev/null
+++ b/libtommath/bn_mp_get_mag_ull.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_GET_MAG_ULL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_GET_MAG(mp_get_mag_ull, unsigned long long)
+#endif
diff --git a/libtommath/bn_mp_grow.c b/libtommath/bn_mp_grow.c
index b120194..9e904c5 100644
--- a/libtommath/bn_mp_grow.c
+++ b/libtommath/bn_mp_grow.c
@@ -1,37 +1,25 @@
 #include "tommath_private.h"
 #ifdef BN_MP_GROW_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* grow as required */
-int mp_grow(mp_int *a, int size)
+mp_err mp_grow(mp_int *a, int size)
 {
    int     i;
    mp_digit *tmp;
 
    /* if the alloc size is smaller alloc more ram */
    if (a->alloc < size) {
-      /* ensure there are always at least MP_PREC digits extra on top */
-      size += (MP_PREC * 2) - (size % MP_PREC);
-
       /* reallocate the array a->dp
        *
        * We store the return in a temporary variable
        * in case the operation failed we don't want
        * to overwrite the dp member of a.
        */
-      tmp = (mp_digit *) XREALLOC(a->dp,
-                                  (size_t)a->alloc * sizeof (mp_digit),
-                                  (size_t)size * sizeof(mp_digit));
+      tmp = (mp_digit *) MP_REALLOC(a->dp,
+                                    (size_t)a->alloc * sizeof(mp_digit),
+                                    (size_t)size * sizeof(mp_digit));
       if (tmp == NULL) {
          /* reallocation failed but "a" is still valid [can be freed] */
          return MP_MEM;
@@ -43,14 +31,8 @@ int mp_grow(mp_int *a, int size)
       /* zero excess digits */
       i        = a->alloc;
       a->alloc = size;
-      for (; i < a->alloc; i++) {
-         a->dp[i] = 0;
-      }
+      MP_ZERO_DIGITS(a->dp + i, a->alloc - i);
    }
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_import.c b/libtommath/bn_mp_import.c
deleted file mode 100644
index 066c5b3..0000000
--- a/libtommath/bn_mp_import.c
+++ /dev/null
@@ -1,68 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_IMPORT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* based on gmp's mpz_import.
- * see http://gmplib.org/manual/Integer-Import-and-Export.html
- */
-int mp_import(mp_int *rop, size_t count, int order, size_t size,
-              int endian, size_t nails, const void *op)
-{
-   int result;
-   size_t odd_nails, nail_bytes, i, j;
-   unsigned char odd_nail_mask;
-
-   mp_zero(rop);
-
-   if (endian == 0) {
-      union {
-         unsigned int i;
-         char c[4];
-      } lint;
-      lint.i = 0x01020304;
-
-      endian = (lint.c[0] == '\x04') ? -1 : 1;
-   }
-
-   odd_nails = (nails % 8u);
-   odd_nail_mask = 0xff;
-   for (i = 0; i < odd_nails; ++i) {
-      odd_nail_mask ^= (unsigned char)(1u << (7u - i));
-   }
-   nail_bytes = nails / 8u;
-
-   for (i = 0; i < count; ++i) {
-      for (j = 0; j < (size - nail_bytes); ++j) {
-         unsigned char byte = *((unsigned char *)op +
-                                (((order == 1) ? i : ((count - 1u) - i)) * size) +
-                                ((endian == 1) ? (j + nail_bytes) : (((size - 1u) - j) - nail_bytes)));
-
-         if ((result = mp_mul_2d(rop, (j == 0u) ? (int)(8u - odd_nails) : 8, rop)) != MP_OKAY) {
-            return result;
-         }
-
-         rop->dp[0] |= (j == 0u) ? (mp_digit)(byte & odd_nail_mask) : (mp_digit)byte;
-         rop->used  += 1;
-      }
-   }
-
-   mp_clamp(rop);
-
-   return MP_OKAY;
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_incr.c b/libtommath/bn_mp_incr.c
new file mode 100644
index 0000000..7695ac7
--- /dev/null
+++ b/libtommath/bn_mp_incr.c
@@ -0,0 +1,30 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INCR_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* Increment "a" by one like "a++". Changes input! */
+mp_err mp_incr(mp_int *a)
+{
+   if (MP_IS_ZERO(a)) {
+      mp_set(a,1uL);
+      return MP_OKAY;
+   } else if (a->sign == MP_NEG) {
+      mp_err err;
+      a->sign = MP_ZPOS;
+      if ((err = mp_decr(a)) != MP_OKAY) {
+         return err;
+      }
+      /* There is no -0 in LTM */
+      if (!MP_IS_ZERO(a)) {
+         a->sign = MP_NEG;
+      }
+      return MP_OKAY;
+   } else if (a->dp[0] < MP_DIGIT_MAX) {
+      a->dp[0]++;
+      return MP_OKAY;
+   } else {
+      return mp_add_d(a, 1uL,a);
+   }
+}
+#endif
diff --git a/libtommath/bn_mp_init.c b/libtommath/bn_mp_init.c
index 3c0c489..2eb7924 100644
--- a/libtommath/bn_mp_init.c
+++ b/libtommath/bn_mp_init.c
@@ -1,33 +1,17 @@
 #include "tommath_private.h"
 #ifdef BN_MP_INIT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* init a new mp_int */
-int mp_init(mp_int *a)
+mp_err mp_init(mp_int *a)
 {
-   int i;
-
    /* allocate memory required and clear it */
-   a->dp = (mp_digit *) XMALLOC(MP_PREC * sizeof(mp_digit));
+   a->dp = (mp_digit *) MP_CALLOC((size_t)MP_PREC, sizeof(mp_digit));
    if (a->dp == NULL) {
       return MP_MEM;
    }
 
-   /* set the digits to zero */
-   for (i = 0; i < MP_PREC; i++) {
-      a->dp[i] = 0;
-   }
-
    /* set the used to zero, allocated digits to the default precision
     * and sign to positive */
    a->used  = 0;
@@ -37,7 +21,3 @@ int mp_init(mp_int *a)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_init_copy.c b/libtommath/bn_mp_init_copy.c
index 4739a98..1888203 100644
--- a/libtommath/bn_mp_init_copy.c
+++ b/libtommath/bn_mp_init_copy.c
@@ -1,34 +1,21 @@
 #include "tommath_private.h"
 #ifdef BN_MP_INIT_COPY_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* creates "a" then copies b into it */
-int mp_init_copy(mp_int *a, const mp_int *b)
+mp_err mp_init_copy(mp_int *a, const mp_int *b)
 {
-   int     res;
+   mp_err     err;
 
-   if ((res = mp_init_size(a, b->used)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(a, b->used)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_copy(b, a)) != MP_OKAY) {
+   if ((err = mp_copy(b, a)) != MP_OKAY) {
       mp_clear(a);
    }
 
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_init_i32.c b/libtommath/bn_mp_init_i32.c
new file mode 100644
index 0000000..bc4de8d
--- /dev/null
+++ b/libtommath/bn_mp_init_i32.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_I32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_i32, mp_set_i32, int32_t)
+#endif
diff --git a/libtommath/bn_mp_init_i64.c b/libtommath/bn_mp_init_i64.c
new file mode 100644
index 0000000..2fa1516
--- /dev/null
+++ b/libtommath/bn_mp_init_i64.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_I64_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_i64, mp_set_i64, int64_t)
+#endif
diff --git a/libtommath/bn_mp_init_l.c b/libtommath/bn_mp_init_l.c
new file mode 100644
index 0000000..bc380b5
--- /dev/null
+++ b/libtommath/bn_mp_init_l.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_L_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_l, mp_set_l, long)
+#endif
diff --git a/libtommath/bn_mp_init_ll.c b/libtommath/bn_mp_init_ll.c
new file mode 100644
index 0000000..dc7c4a4
--- /dev/null
+++ b/libtommath/bn_mp_init_ll.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_LL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_ll, mp_set_ll, long long)
+#endif
diff --git a/libtommath/bn_mp_init_multi.c b/libtommath/bn_mp_init_multi.c
index 7f8bd04..d8390b5 100644
--- a/libtommath/bn_mp_init_multi.c
+++ b/libtommath/bn_mp_init_multi.c
@@ -1,22 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_INIT_MULTI_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #include <stdarg.h>
 
-int mp_init_multi(mp_int *mp, ...)
+mp_err mp_init_multi(mp_int *mp, ...)
 {
-   mp_err res = MP_OKAY;      /* Assume ok until proven otherwise */
+   mp_err err = MP_OKAY;      /* Assume ok until proven otherwise */
    int n = 0;                 /* Number of ok inits */
    mp_int *cur_arg = mp;
    va_list args;
@@ -37,18 +28,14 @@ int mp_init_multi(mp_int *mp, ...)
             cur_arg = va_arg(clean_args, mp_int *);
          }
          va_end(clean_args);
-         res = MP_MEM;
+         err = MP_MEM;
          break;
       }
       n++;
       cur_arg = va_arg(args, mp_int *);
    }
    va_end(args);
-   return res;                /* Assumed ok, if error flagged above. */
+   return err;                /* Assumed ok, if error flagged above. */
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_init_set.c b/libtommath/bn_mp_init_set.c
index 36606af..5068f2b 100644
--- a/libtommath/bn_mp_init_set.c
+++ b/libtommath/bn_mp_init_set.c
@@ -1,21 +1,12 @@
 #include "tommath_private.h"
 #ifdef BN_MP_INIT_SET_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* initialize and set a digit */
-int mp_init_set(mp_int *a, mp_digit b)
+mp_err mp_init_set(mp_int *a, mp_digit b)
 {
-   int err;
+   mp_err err;
    if ((err = mp_init(a)) != MP_OKAY) {
       return err;
    }
@@ -23,7 +14,3 @@ int mp_init_set(mp_int *a, mp_digit b)
    return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_init_set_int.c b/libtommath/bn_mp_init_set_int.c
deleted file mode 100644
index 7d81811..0000000
--- a/libtommath/bn_mp_init_set_int.c
+++ /dev/null
@@ -1,28 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_INIT_SET_INT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* initialize and set a digit */
-int mp_init_set_int(mp_int *a, unsigned long b)
-{
-   int err;
-   if ((err = mp_init(a)) != MP_OKAY) {
-      return err;
-   }
-   return mp_set_int(a, b);
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_init_size.c b/libtommath/bn_mp_init_size.c
index 1becb23..d622687 100644
--- a/libtommath/bn_mp_init_size.c
+++ b/libtommath/bn_mp_init_size.c
@@ -1,27 +1,15 @@
 #include "tommath_private.h"
 #ifdef BN_MP_INIT_SIZE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* init an mp_init for a given size */
-int mp_init_size(mp_int *a, int size)
+mp_err mp_init_size(mp_int *a, int size)
 {
-   int x;
-
-   /* pad size so there are always extra digits */
-   size += (MP_PREC * 2) - (size % MP_PREC);
+   size = MP_MAX(MP_MIN_PREC, size);
 
    /* alloc mem */
-   a->dp = (mp_digit *) XMALLOC((size_t)size * sizeof(mp_digit));
+   a->dp = (mp_digit *) MP_CALLOC((size_t)size, sizeof(mp_digit));
    if (a->dp == NULL) {
       return MP_MEM;
    }
@@ -31,15 +19,6 @@ int mp_init_size(mp_int *a, int size)
    a->alloc = size;
    a->sign  = MP_ZPOS;
 
-   /* zero the digits */
-   for (x = 0; x < size; x++) {
-      a->dp[x] = 0;
-   }
-
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_init_u32.c b/libtommath/bn_mp_init_u32.c
new file mode 100644
index 0000000..015d89b
--- /dev/null
+++ b/libtommath/bn_mp_init_u32.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_U32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_u32, mp_set_u32, uint32_t)
+#endif
diff --git a/libtommath/bn_mp_init_u64.c b/libtommath/bn_mp_init_u64.c
new file mode 100644
index 0000000..2b35f7e
--- /dev/null
+++ b/libtommath/bn_mp_init_u64.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_U64_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_u64, mp_set_u64, uint64_t)
+#endif
diff --git a/libtommath/bn_mp_init_ul.c b/libtommath/bn_mp_init_ul.c
new file mode 100644
index 0000000..5164f72
--- /dev/null
+++ b/libtommath/bn_mp_init_ul.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_UL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_ul, mp_set_ul, unsigned long)
+#endif
diff --git a/libtommath/bn_mp_init_ull.c b/libtommath/bn_mp_init_ull.c
new file mode 100644
index 0000000..84110c0
--- /dev/null
+++ b/libtommath/bn_mp_init_ull.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_INIT_ULL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_INIT_INT(mp_init_ull, mp_set_ull, unsigned long long)
+#endif
diff --git a/libtommath/bn_mp_invmod.c b/libtommath/bn_mp_invmod.c
index f1a482d..7b35a24 100644
--- a/libtommath/bn_mp_invmod.c
+++ b/libtommath/bn_mp_invmod.c
@@ -1,40 +1,23 @@
 #include "tommath_private.h"
 #ifdef BN_MP_INVMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* hac 14.61, pp608 */
-int mp_invmod(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_invmod(const mp_int *a, const mp_int *b, mp_int *c)
 {
    /* b cannot be negative and has to be >1 */
    if ((b->sign == MP_NEG) || (mp_cmp_d(b, 1uL) != MP_GT)) {
       return MP_VAL;
    }
 
-#ifdef BN_FAST_MP_INVMOD_C
    /* if the modulus is odd we can use a faster routine instead */
-   if ((mp_isodd(b) == MP_YES)) {
-      return fast_mp_invmod(a, b, c);
+   if (MP_HAS(S_MP_INVMOD_FAST) && MP_IS_ODD(b)) {
+      return s_mp_invmod_fast(a, b, c);
    }
-#endif
 
-#ifdef BN_MP_INVMOD_SLOW_C
-   return mp_invmod_slow(a, b, c);
-#else
-   return MP_VAL;
-#endif
+   return MP_HAS(S_MP_INVMOD_SLOW)
+          ? s_mp_invmod_slow(a, b, c)
+          : MP_VAL;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_invmod_slow.c b/libtommath/bn_mp_invmod_slow.c
deleted file mode 100644
index e60cf04..0000000
--- a/libtommath/bn_mp_invmod_slow.c
+++ /dev/null
@@ -1,173 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_INVMOD_SLOW_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* hac 14.61, pp608 */
-int mp_invmod_slow(const mp_int *a, const mp_int *b, mp_int *c)
-{
-   mp_int  x, y, u, v, A, B, C, D;
-   int     res;
-
-   /* b cannot be negative */
-   if ((b->sign == MP_NEG) || (mp_iszero(b) == MP_YES)) {
-      return MP_VAL;
-   }
-
-   /* init temps */
-   if ((res = mp_init_multi(&x, &y, &u, &v,
-                            &A, &B, &C, &D, NULL)) != MP_OKAY) {
-      return res;
-   }
-
-   /* x = a, y = b */
-   if ((res = mp_mod(a, b, &x)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_copy(b, &y)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* 2. [modified] if x,y are both even then return an error! */
-   if ((mp_iseven(&x) == MP_YES) && (mp_iseven(&y) == MP_YES)) {
-      res = MP_VAL;
-      goto LBL_ERR;
-   }
-
-   /* 3. u=x, v=y, A=1, B=0, C=0,D=1 */
-   if ((res = mp_copy(&x, &u)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_copy(&y, &v)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_set(&A, 1uL);
-   mp_set(&D, 1uL);
-
-top:
-   /* 4.  while u is even do */
-   while (mp_iseven(&u) == MP_YES) {
-      /* 4.1 u = u/2 */
-      if ((res = mp_div_2(&u, &u)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      /* 4.2 if A or B is odd then */
-      if ((mp_isodd(&A) == MP_YES) || (mp_isodd(&B) == MP_YES)) {
-         /* A = (A+y)/2, B = (B-x)/2 */
-         if ((res = mp_add(&A, &y, &A)) != MP_OKAY) {
-            goto LBL_ERR;
-         }
-         if ((res = mp_sub(&B, &x, &B)) != MP_OKAY) {
-            goto LBL_ERR;
-         }
-      }
-      /* A = A/2, B = B/2 */
-      if ((res = mp_div_2(&A, &A)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((res = mp_div_2(&B, &B)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* 5.  while v is even do */
-   while (mp_iseven(&v) == MP_YES) {
-      /* 5.1 v = v/2 */
-      if ((res = mp_div_2(&v, &v)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      /* 5.2 if C or D is odd then */
-      if ((mp_isodd(&C) == MP_YES) || (mp_isodd(&D) == MP_YES)) {
-         /* C = (C+y)/2, D = (D-x)/2 */
-         if ((res = mp_add(&C, &y, &C)) != MP_OKAY) {
-            goto LBL_ERR;
-         }
-         if ((res = mp_sub(&D, &x, &D)) != MP_OKAY) {
-            goto LBL_ERR;
-         }
-      }
-      /* C = C/2, D = D/2 */
-      if ((res = mp_div_2(&C, &C)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-      if ((res = mp_div_2(&D, &D)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* 6.  if u >= v then */
-   if (mp_cmp(&u, &v) != MP_LT) {
-      /* u = u - v, A = A - C, B = B - D */
-      if ((res = mp_sub(&u, &v, &u)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-
-      if ((res = mp_sub(&A, &C, &A)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-
-      if ((res = mp_sub(&B, &D, &B)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   } else {
-      /* v - v - u, C = C - A, D = D - B */
-      if ((res = mp_sub(&v, &u, &v)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-
-      if ((res = mp_sub(&C, &A, &C)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-
-      if ((res = mp_sub(&D, &B, &D)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* if not zero goto step 4 */
-   if (mp_iszero(&u) == MP_NO)
-      goto top;
-
-   /* now a = C, b = D, gcd == g*v */
-
-   /* if v != 1 then there is no inverse */
-   if (mp_cmp_d(&v, 1uL) != MP_EQ) {
-      res = MP_VAL;
-      goto LBL_ERR;
-   }
-
-   /* if its too low */
-   while (mp_cmp_d(&C, 0uL) == MP_LT) {
-      if ((res = mp_add(&C, b, &C)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* too big */
-   while (mp_cmp_mag(&C, b) != MP_LT) {
-      if ((res = mp_sub(&C, b, &C)) != MP_OKAY) {
-         goto LBL_ERR;
-      }
-   }
-
-   /* C is now the inverse */
-   mp_exch(&C, c);
-   res = MP_OKAY;
-LBL_ERR:
-   mp_clear_multi(&x, &y, &u, &v, &A, &B, &C, &D, NULL);
-   return res;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_is_square.c b/libtommath/bn_mp_is_square.c
index 1dd1d6c..69e77a2 100644
--- a/libtommath/bn_mp_is_square.c
+++ b/libtommath/bn_mp_is_square.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_IS_SQUARE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* Check if remainders are possible squares - fast exclude non-squares */
 static const char rem_128[128] = {
@@ -35,9 +26,9 @@ static const char rem_105[105] = {
 };
 
 /* Store non-zero to ret if arg is square, and zero if not */
-int mp_is_square(const mp_int *arg, int *ret)
+mp_err mp_is_square(const mp_int *arg, mp_bool *ret)
 {
-   int           res;
+   mp_err        err;
    mp_digit      c;
    mp_int        t;
    unsigned long r;
@@ -49,33 +40,33 @@ int mp_is_square(const mp_int *arg, int *ret)
       return MP_VAL;
    }
 
-   if (IS_ZERO(arg)) {
+   if (MP_IS_ZERO(arg)) {
       return MP_OKAY;
    }
 
-   /* First check mod 128 (suppose that DIGIT_BIT is at least 7) */
+   /* First check mod 128 (suppose that MP_DIGIT_BIT is at least 7) */
    if (rem_128[127u & arg->dp[0]] == (char)1) {
       return MP_OKAY;
    }
 
    /* Next check mod 105 (3*5*7) */
-   if ((res = mp_mod_d(arg, 105uL, &c)) != MP_OKAY) {
-      return res;
+   if ((err = mp_mod_d(arg, 105uL, &c)) != MP_OKAY) {
+      return err;
    }
    if (rem_105[c] == (char)1) {
       return MP_OKAY;
    }
 
 
-   if ((res = mp_init_set_int(&t, 11L*13L*17L*19L*23L*29L*31L)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_u32(&t, 11u*13u*17u*19u*23u*29u*31u)) != MP_OKAY) {
+      return err;
    }
-   if ((res = mp_mod(arg, &t, &t)) != MP_OKAY) {
+   if ((err = mp_mod(arg, &t, &t)) != MP_OKAY) {
       goto LBL_ERR;
    }
-   r = mp_get_int(&t);
+   r = mp_get_u32(&t);
    /* Check for other prime modules, note it's not an ERROR but we must
-    * free "t" so the easiest way is to goto LBL_ERR.  We know that res
+    * free "t" so the easiest way is to goto LBL_ERR.  We know that err
     * is already equal to MP_OKAY from the mp_mod call
     */
    if (((1uL<<(r%11uL)) & 0x5C4uL) != 0uL)         goto LBL_ERR;
@@ -87,20 +78,16 @@ int mp_is_square(const mp_int *arg, int *ret)
    if (((1uL<<(r%31uL)) & 0x6DE2B848uL) != 0uL)    goto LBL_ERR;
 
    /* Final check - is sqr(sqrt(arg)) == arg ? */
-   if ((res = mp_sqrt(arg, &t)) != MP_OKAY) {
+   if ((err = mp_sqrt(arg, &t)) != MP_OKAY) {
       goto LBL_ERR;
    }
-   if ((res = mp_sqr(&t, &t)) != MP_OKAY) {
+   if ((err = mp_sqr(&t, &t)) != MP_OKAY) {
       goto LBL_ERR;
    }
 
    *ret = (mp_cmp_mag(&t, arg) == MP_EQ) ? MP_YES : MP_NO;
 LBL_ERR:
    mp_clear(&t);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_iseven.c b/libtommath/bn_mp_iseven.c
new file mode 100644
index 0000000..5cb9622
--- /dev/null
+++ b/libtommath/bn_mp_iseven.c
@@ -0,0 +1,10 @@
+#include "tommath_private.h"
+#ifdef BN_MP_ISEVEN_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+mp_bool mp_iseven(const mp_int *a)
+{
+   return MP_IS_EVEN(a) ? MP_YES : MP_NO;
+}
+#endif
diff --git a/libtommath/bn_mp_isodd.c b/libtommath/bn_mp_isodd.c
new file mode 100644
index 0000000..bf17646
--- /dev/null
+++ b/libtommath/bn_mp_isodd.c
@@ -0,0 +1,10 @@
+#include "tommath_private.h"
+#ifdef BN_MP_ISODD_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+mp_bool mp_isodd(const mp_int *a)
+{
+   return MP_IS_ODD(a) ? MP_YES : MP_NO;
+}
+#endif
diff --git a/libtommath/bn_mp_jacobi.c b/libtommath/bn_mp_jacobi.c
deleted file mode 100644
index 1eb3dd4..0000000
--- a/libtommath/bn_mp_jacobi.c
+++ /dev/null
@@ -1,36 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_JACOBI_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* computes the jacobi c = (a | n) (or Legendre if n is prime)
- * Kept for legacy reasons, please use mp_kronecker() instead
- */
-int mp_jacobi(const mp_int *a, const mp_int *n, int *c)
-{
-   /* if a < 0 return MP_VAL */
-   if (mp_isneg(a) == MP_YES) {
-      return MP_VAL;
-   }
-
-   /* if n <= 0 return MP_VAL */
-   if (mp_cmp_d(n, 0uL) != MP_GT) {
-      return MP_VAL;
-   }
-
-   return mp_kronecker(a, n, c);
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_kronecker.c b/libtommath/bn_mp_kronecker.c
index a20fa74..525a820 100644
--- a/libtommath/bn_mp_kronecker.c
+++ b/libtommath/bn_mp_kronecker.c
@@ -1,17 +1,8 @@
 #include "tommath_private.h"
 #ifdef BN_MP_KRONECKER_C
 
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /*
    Kronecker symbol (a|p)
@@ -26,43 +17,41 @@
      publisher={Springer Science \& Business Media}
     }
  */
-int mp_kronecker(const mp_int *a, const mp_int *p, int *c)
+mp_err mp_kronecker(const mp_int *a, const mp_int *p, int *c)
 {
    mp_int a1, p1, r;
-
-   int e = MP_OKAY;
+   mp_err err;
    int v, k;
 
    static const int table[8] = {0, 1, 0, -1, 0, -1, 0, 1};
 
-   if (mp_iszero(p) != MP_NO) {
+   if (MP_IS_ZERO(p)) {
       if ((a->used == 1) && (a->dp[0] == 1u)) {
          *c = 1;
-         return e;
       } else {
          *c = 0;
-         return e;
       }
+      return MP_OKAY;
    }
 
-   if ((mp_iseven(a) != MP_NO) && (mp_iseven(p) != MP_NO)) {
+   if (MP_IS_EVEN(a) && MP_IS_EVEN(p)) {
       *c = 0;
-      return e;
+      return MP_OKAY;
    }
 
-   if ((e = mp_init_copy(&a1, a)) != MP_OKAY) {
-      return e;
+   if ((err = mp_init_copy(&a1, a)) != MP_OKAY) {
+      return err;
    }
-   if ((e = mp_init_copy(&p1, p)) != MP_OKAY) {
+   if ((err = mp_init_copy(&p1, p)) != MP_OKAY) {
       goto LBL_KRON_0;
    }
 
    v = mp_cnt_lsb(&p1);
-   if ((e = mp_div_2d(&p1, v, &p1, NULL)) != MP_OKAY) {
+   if ((err = mp_div_2d(&p1, v, &p1, NULL)) != MP_OKAY) {
       goto LBL_KRON_1;
    }
 
-   if ((v & 0x1) == 0) {
+   if ((v & 1) == 0) {
       k = 1;
    } else {
       k = table[a->dp[0] & 7u];
@@ -75,12 +64,12 @@ int mp_kronecker(const mp_int *a, const mp_int *p, int *c)
       }
    }
 
-   if ((e = mp_init(&r)) != MP_OKAY) {
+   if ((err = mp_init(&r)) != MP_OKAY) {
       goto LBL_KRON_1;
    }
 
    for (;;) {
-      if (mp_iszero(&a1) != MP_NO) {
+      if (MP_IS_ZERO(&a1)) {
          if (mp_cmp_d(&p1, 1uL) == MP_EQ) {
             *c = k;
             goto LBL_KRON;
@@ -91,11 +80,11 @@ int mp_kronecker(const mp_int *a, const mp_int *p, int *c)
       }
 
       v = mp_cnt_lsb(&a1);
-      if ((e = mp_div_2d(&a1, v, &a1, NULL)) != MP_OKAY) {
+      if ((err = mp_div_2d(&a1, v, &a1, NULL)) != MP_OKAY) {
          goto LBL_KRON;
       }
 
-      if ((v & 0x1) == 1) {
+      if ((v & 1) == 1) {
          k = k * table[p1.dp[0] & 7u];
       }
 
@@ -115,14 +104,14 @@ int mp_kronecker(const mp_int *a, const mp_int *p, int *c)
          }
       }
 
-      if ((e = mp_copy(&a1, &r)) != MP_OKAY) {
+      if ((err = mp_copy(&a1, &r)) != MP_OKAY) {
          goto LBL_KRON;
       }
       r.sign = MP_ZPOS;
-      if ((e = mp_mod(&p1, &r, &a1)) != MP_OKAY) {
+      if ((err = mp_mod(&p1, &r, &a1)) != MP_OKAY) {
          goto LBL_KRON;
       }
-      if ((e = mp_copy(&r, &p1)) != MP_OKAY) {
+      if ((err = mp_copy(&r, &p1)) != MP_OKAY) {
          goto LBL_KRON;
       }
    }
@@ -134,11 +123,7 @@ LBL_KRON_1:
 LBL_KRON_0:
    mp_clear(&a1);
 
-   return e;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_lcm.c b/libtommath/bn_mp_lcm.c
index cb9fa3d..c32b269 100644
--- a/libtommath/bn_mp_lcm.c
+++ b/libtommath/bn_mp_lcm.c
@@ -1,46 +1,37 @@
 #include "tommath_private.h"
 #ifdef BN_MP_LCM_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* computes least common multiple as |a*b|/(a, b) */
-int mp_lcm(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_lcm(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int     res;
+   mp_err  err;
    mp_int  t1, t2;
 
 
-   if ((res = mp_init_multi(&t1, &t2, NULL)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_multi(&t1, &t2, NULL)) != MP_OKAY) {
+      return err;
    }
 
    /* t1 = get the GCD of the two inputs */
-   if ((res = mp_gcd(a, b, &t1)) != MP_OKAY) {
+   if ((err = mp_gcd(a, b, &t1)) != MP_OKAY) {
       goto LBL_T;
    }
 
    /* divide the smallest by the GCD */
    if (mp_cmp_mag(a, b) == MP_LT) {
       /* store quotient in t2 such that t2 * b is the LCM */
-      if ((res = mp_div(a, &t1, &t2, NULL)) != MP_OKAY) {
+      if ((err = mp_div(a, &t1, &t2, NULL)) != MP_OKAY) {
          goto LBL_T;
       }
-      res = mp_mul(b, &t2, c);
+      err = mp_mul(b, &t2, c);
    } else {
       /* store quotient in t2 such that t2 * a is the LCM */
-      if ((res = mp_div(b, &t1, &t2, NULL)) != MP_OKAY) {
+      if ((err = mp_div(b, &t1, &t2, NULL)) != MP_OKAY) {
          goto LBL_T;
       }
-      res = mp_mul(a, &t2, c);
+      err = mp_mul(a, &t2, c);
    }
 
    /* fix the sign to positive */
@@ -48,10 +39,6 @@ int mp_lcm(const mp_int *a, const mp_int *b, mp_int *c)
 
 LBL_T:
    mp_clear_multi(&t1, &t2, NULL);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_log_u32.c b/libtommath/bn_mp_log_u32.c
new file mode 100644
index 0000000..f507b1d
--- /dev/null
+++ b/libtommath/bn_mp_log_u32.c
@@ -0,0 +1,180 @@
+#include "tommath_private.h"
+#ifdef BN_MP_LOG_U32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* Compute log_{base}(a) */
+static mp_word s_pow(mp_word base, mp_word exponent)
+{
+   mp_word result = 1;
+   while (exponent != 0u) {
+      if ((exponent & 1u) == 1u) {
+         result *= base;
+      }
+      exponent >>= 1;
+      base *= base;
+   }
+
+   return result;
+}
+
+static mp_digit s_digit_ilogb(mp_digit base, mp_digit n)
+{
+   mp_word bracket_low = 1, bracket_mid, bracket_high, N;
+   mp_digit ret, high = 1uL, low = 0uL, mid;
+
+   if (n < base) {
+      return 0uL;
+   }
+   if (n == base) {
+      return 1uL;
+   }
+
+   bracket_high = (mp_word) base ;
+   N = (mp_word) n;
+
+   while (bracket_high < N) {
+      low = high;
+      bracket_low = bracket_high;
+      high <<= 1;
+      bracket_high *= bracket_high;
+   }
+
+   while (((mp_digit)(high - low)) > 1uL) {
+      mid = (low + high) >> 1;
+      bracket_mid = bracket_low * s_pow(base, (mp_word)(mid - low));
+
+      if (N < bracket_mid) {
+         high = mid ;
+         bracket_high = bracket_mid ;
+      }
+      if (N > bracket_mid) {
+         low = mid ;
+         bracket_low = bracket_mid ;
+      }
+      if (N == bracket_mid) {
+         return (mp_digit) mid;
+      }
+   }
+
+   if (bracket_high == N) {
+      ret = high;
+   } else {
+      ret = low;
+   }
+
+   return ret;
+}
+
+/* TODO: output could be "int" because the output of mp_radix_size is int, too,
+         as is the output of mp_bitcount.
+         With the same problem: max size is INT_MAX * MP_DIGIT not INT_MAX only!
+*/
+mp_err mp_log_u32(const mp_int *a, unsigned int base, unsigned int *c)
+{
+   mp_err err;
+   mp_ord cmp;
+   unsigned int high, low, mid;
+   mp_int bracket_low, bracket_high, bracket_mid, t, bi_base;
+
+   err = MP_OKAY;
+
+   if (a->sign == MP_NEG) {
+      return MP_VAL;
+   }
+
+   if (MP_IS_ZERO(a)) {
+      return MP_VAL;
+   }
+
+   if (base < 2u) {
+      return MP_VAL;
+   }
+
+   /* A small shortcut for bases that are powers of two. */
+   if ((base & (base - 1u)) == 0u) {
+      int y, bit_count;
+      for (y=0; (y < 7) && ((base & 1u) == 0u); y++) {
+         base >>= 1;
+      }
+      bit_count = mp_count_bits(a) - 1;
+      *c = (unsigned int)(bit_count/y);
+      return MP_OKAY;
+   }
+
+   if (a->used == 1) {
+      *c = (unsigned int)s_digit_ilogb(base, a->dp[0]);
+      return err;
+   }
+
+   cmp = mp_cmp_d(a, base);
+   if ((cmp == MP_LT) || (cmp == MP_EQ)) {
+      *c = cmp == MP_EQ;
+      return err;
+   }
+
+   if ((err =
+           mp_init_multi(&bracket_low, &bracket_high,
+                         &bracket_mid, &t, &bi_base, NULL)) != MP_OKAY) {
+      return err;
+   }
+
+   low = 0u;
+   mp_set(&bracket_low, 1uL);
+   high = 1u;
+
+   mp_set(&bracket_high, base);
+
+   /*
+       A kind of Giant-step/baby-step algorithm.
+       Idea shamelessly stolen from https://programmingpraxis.com/2010/05/07/integer-logarithms/2/
+       The effect is asymptotic, hence needs benchmarks to test if the Giant-step should be skipped
+       for small n.
+    */
+   while (mp_cmp(&bracket_high, a) == MP_LT) {
+      low = high;
+      if ((err = mp_copy(&bracket_high, &bracket_low)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      high <<= 1;
+      if ((err = mp_sqr(&bracket_high, &bracket_high)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+   }
+   mp_set(&bi_base, base);
+
+   while ((high - low) > 1u) {
+      mid = (high + low) >> 1;
+
+      if ((err = mp_expt_u32(&bi_base, mid - low, &t)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      if ((err = mp_mul(&bracket_low, &t, &bracket_mid)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      cmp = mp_cmp(a, &bracket_mid);
+      if (cmp == MP_LT) {
+         high = mid;
+         mp_exch(&bracket_mid, &bracket_high);
+      }
+      if (cmp == MP_GT) {
+         low = mid;
+         mp_exch(&bracket_mid, &bracket_low);
+      }
+      if (cmp == MP_EQ) {
+         *c = mid;
+         goto LBL_END;
+      }
+   }
+
+   *c = (mp_cmp(&bracket_high, a) == MP_EQ) ? high : low;
+
+LBL_END:
+LBL_ERR:
+   mp_clear_multi(&bracket_low, &bracket_high, &bracket_mid,
+                  &t, &bi_base, NULL);
+   return err;
+}
+
+
+#endif
diff --git a/libtommath/bn_mp_lshd.c b/libtommath/bn_mp_lshd.c
index 6762a10..8234580 100644
--- a/libtommath/bn_mp_lshd.c
+++ b/libtommath/bn_mp_lshd.c
@@ -1,68 +1,51 @@
 #include "tommath_private.h"
 #ifdef BN_MP_LSHD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* shift left a certain amount of digits */
-int mp_lshd(mp_int *a, int b)
+mp_err mp_lshd(mp_int *a, int b)
 {
-   int     x, res;
+   int x;
+   mp_err err;
+   mp_digit *top, *bottom;
 
    /* if its less than zero return */
    if (b <= 0) {
       return MP_OKAY;
    }
    /* no need to shift 0 around */
-   if (mp_iszero(a) == MP_YES) {
+   if (MP_IS_ZERO(a)) {
       return MP_OKAY;
    }
 
    /* grow to fit the new digits */
    if (a->alloc < (a->used + b)) {
-      if ((res = mp_grow(a, a->used + b)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(a, a->used + b)) != MP_OKAY) {
+         return err;
       }
    }
 
-   {
-      mp_digit *top, *bottom;
+   /* increment the used by the shift amount then copy upwards */
+   a->used += b;
 
-      /* increment the used by the shift amount then copy upwards */
-      a->used += b;
+   /* top */
+   top = a->dp + a->used - 1;
 
-      /* top */
-      top = a->dp + a->used - 1;
+   /* base */
+   bottom = (a->dp + a->used - 1) - b;
 
-      /* base */
-      bottom = (a->dp + a->used - 1) - b;
+   /* much like mp_rshd this is implemented using a sliding window
+    * except the window goes the otherway around.  Copying from
+    * the bottom to the top.  see bn_mp_rshd.c for more info.
+    */
+   for (x = a->used - 1; x >= b; x--) {
+      *top-- = *bottom--;
+   }
 
-      /* much like mp_rshd this is implemented using a sliding window
-       * except the window goes the otherway around.  Copying from
-       * the bottom to the top.  see bn_mp_rshd.c for more info.
-       */
-      for (x = a->used - 1; x >= b; x--) {
-         *top-- = *bottom--;
-      }
+   /* zero the lower digits */
+   MP_ZERO_DIGITS(a->dp, b);
 
-      /* zero the lower digits */
-      top = a->dp;
-      for (x = 0; x < b; x++) {
-         *top++ = 0;
-      }
-   }
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mod.c b/libtommath/bn_mp_mod.c
index fa022a7..8fbfe08 100644
--- a/libtommath/bn_mp_mod.c
+++ b/libtommath/bn_mp_mod.c
@@ -1,44 +1,31 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* c = a mod b, 0 <= c < b if b > 0, b < c <= 0 if b < 0 */
-int mp_mod(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_mod(const mp_int *a, const mp_int *b, mp_int *c)
 {
    mp_int  t;
-   int     res;
+   mp_err  err;
 
-   if ((res = mp_init_size(&t, b->used)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&t, b->used)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_div(a, b, NULL, &t)) != MP_OKAY) {
-      mp_clear(&t);
-      return res;
+   if ((err = mp_div(a, b, NULL, &t)) != MP_OKAY) {
+      goto LBL_ERR;
    }
 
-   if ((mp_iszero(&t) != MP_NO) || (t.sign == b->sign)) {
-      res = MP_OKAY;
+   if (MP_IS_ZERO(&t) || (t.sign == b->sign)) {
+      err = MP_OKAY;
       mp_exch(&t, c);
    } else {
-      res = mp_add(b, &t, c);
+      err = mp_add(b, &t, c);
    }
 
+LBL_ERR:
    mp_clear(&t);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mod_2d.c b/libtommath/bn_mp_mod_2d.c
index 759198b..5bf57a1 100644
--- a/libtommath/bn_mp_mod_2d.c
+++ b/libtommath/bn_mp_mod_2d.c
@@ -1,21 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MOD_2D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* calc a value mod 2**b */
-int mp_mod_2d(const mp_int *a, int b, mp_int *c)
+mp_err mp_mod_2d(const mp_int *a, int b, mp_int *c)
 {
-   int     x, res;
+   int x;
+   mp_err err;
 
    /* if b is <= 0 then zero the int */
    if (b <= 0) {
@@ -24,28 +16,23 @@ int mp_mod_2d(const mp_int *a, int b, mp_int *c)
    }
 
    /* if the modulus is larger than the value than return */
-   if (b >= (a->used * DIGIT_BIT)) {
-      res = mp_copy(a, c);
-      return res;
+   if (b >= (a->used * MP_DIGIT_BIT)) {
+      return mp_copy(a, c);
    }
 
    /* copy */
-   if ((res = mp_copy(a, c)) != MP_OKAY) {
-      return res;
+   if ((err = mp_copy(a, c)) != MP_OKAY) {
+      return err;
    }
 
    /* zero digits above the last digit of the modulus */
-   for (x = (b / DIGIT_BIT) + (((b % DIGIT_BIT) == 0) ? 0 : 1); x < c->used; x++) {
-      c->dp[x] = 0;
-   }
+   x = (b / MP_DIGIT_BIT) + (((b % MP_DIGIT_BIT) == 0) ? 0 : 1);
+   MP_ZERO_DIGITS(c->dp + x, c->used - x);
+
    /* clear the digit that is not completely outside/inside the modulus */
-   c->dp[b / DIGIT_BIT] &=
-      ((mp_digit)1 << (mp_digit)(b % DIGIT_BIT)) - (mp_digit)1;
+   c->dp[b / MP_DIGIT_BIT] &=
+      ((mp_digit)1 << (mp_digit)(b % MP_DIGIT_BIT)) - (mp_digit)1;
    mp_clamp(c);
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mod_d.c b/libtommath/bn_mp_mod_d.c
index f58b6b5..0b6c12a 100644
--- a/libtommath/bn_mp_mod_d.c
+++ b/libtommath/bn_mp_mod_d.c
@@ -1,23 +1,10 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MOD_D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
-int mp_mod_d(const mp_int *a, mp_digit b, mp_digit *c)
+mp_err mp_mod_d(const mp_int *a, mp_digit b, mp_digit *c)
 {
    return mp_div_d(a, b, NULL, c);
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_montgomery_calc_normalization.c b/libtommath/bn_mp_montgomery_calc_normalization.c
index 848378c..8379789 100644
--- a/libtommath/bn_mp_montgomery_calc_normalization.c
+++ b/libtommath/bn_mp_montgomery_calc_normalization.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MONTGOMERY_CALC_NORMALIZATION_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /*
  * shifts with subtractions when the result is greater than b.
@@ -18,16 +9,17 @@
  * The method is slightly modified to shift B unconditionally upto just under
  * the leading bit of b.  This saves alot of multiple precision shifting.
  */
-int mp_montgomery_calc_normalization(mp_int *a, const mp_int *b)
+mp_err mp_montgomery_calc_normalization(mp_int *a, const mp_int *b)
 {
-   int     x, bits, res;
+   int    x, bits;
+   mp_err err;
 
    /* how many bits of last digit does b use */
-   bits = mp_count_bits(b) % DIGIT_BIT;
+   bits = mp_count_bits(b) % MP_DIGIT_BIT;
 
    if (b->used > 1) {
-      if ((res = mp_2expt(a, ((b->used - 1) * DIGIT_BIT) + bits - 1)) != MP_OKAY) {
-         return res;
+      if ((err = mp_2expt(a, ((b->used - 1) * MP_DIGIT_BIT) + bits - 1)) != MP_OKAY) {
+         return err;
       }
    } else {
       mp_set(a, 1uL);
@@ -36,13 +28,13 @@ int mp_montgomery_calc_normalization(mp_int *a, const mp_int *b)
 
 
    /* now compute C = A * B mod b */
-   for (x = bits - 1; x < (int)DIGIT_BIT; x++) {
-      if ((res = mp_mul_2(a, a)) != MP_OKAY) {
-         return res;
+   for (x = bits - 1; x < (int)MP_DIGIT_BIT; x++) {
+      if ((err = mp_mul_2(a, a)) != MP_OKAY) {
+         return err;
       }
       if (mp_cmp_mag(a, b) != MP_LT) {
-         if ((res = s_mp_sub(a, b, a)) != MP_OKAY) {
-            return res;
+         if ((err = s_mp_sub(a, b, a)) != MP_OKAY) {
+            return err;
          }
       }
    }
@@ -50,7 +42,3 @@ int mp_montgomery_calc_normalization(mp_int *a, const mp_int *b)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_montgomery_reduce.c b/libtommath/bn_mp_montgomery_reduce.c
index 382c7cc..3ce1f15 100644
--- a/libtommath/bn_mp_montgomery_reduce.c
+++ b/libtommath/bn_mp_montgomery_reduce.c
@@ -1,21 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MONTGOMERY_REDUCE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* computes xR**-1 == x (mod N) via Montgomery Reduction */
-int mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
+mp_err mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
 {
-   int     ix, res, digs;
+   int      ix, digs;
+   mp_err   err;
    mp_digit mu;
 
    /* can the fast reduction [comba] method be used?
@@ -25,17 +17,16 @@ int mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
     * are fixed up in the inner loop.
     */
    digs = (n->used * 2) + 1;
-   if ((digs < (int)MP_WARRAY) &&
-       (x->used <= (int)MP_WARRAY) &&
-       (n->used <
-        (int)(1u << (((size_t)CHAR_BIT * sizeof(mp_word)) - (2u * (size_t)DIGIT_BIT))))) {
-      return fast_mp_montgomery_reduce(x, n, rho);
+   if ((digs < PRIVATE_MP_WARRAY) &&
+       (x->used <= PRIVATE_MP_WARRAY) &&
+       (n->used < MP_MAXFAST)) {
+      return s_mp_montgomery_reduce_fast(x, n, rho);
    }
 
    /* grow the input as required */
    if (x->alloc < digs) {
-      if ((res = mp_grow(x, digs)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(x, digs)) != MP_OKAY) {
+         return err;
       }
    }
    x->used = digs;
@@ -73,7 +64,7 @@ int mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
                       (mp_word)u + (mp_word)*tmpx;
 
             /* get carry */
-            u       = (mp_digit)(r >> (mp_word)DIGIT_BIT);
+            u       = (mp_digit)(r >> (mp_word)MP_DIGIT_BIT);
 
             /* fix digit */
             *tmpx++ = (mp_digit)(r & (mp_word)MP_MASK);
@@ -84,7 +75,7 @@ int mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
          /* propagate carries upwards as required*/
          while (u != 0u) {
             *tmpx   += u;
-            u        = *tmpx >> DIGIT_BIT;
+            u        = *tmpx >> MP_DIGIT_BIT;
             *tmpx++ &= MP_MASK;
          }
       }
@@ -109,7 +100,3 @@ int mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_montgomery_setup.c b/libtommath/bn_mp_montgomery_setup.c
index 26c632a..39f6e9d 100644
--- a/libtommath/bn_mp_montgomery_setup.c
+++ b/libtommath/bn_mp_montgomery_setup.c
@@ -1,19 +1,10 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MONTGOMERY_SETUP_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* setups the montgomery reduction stuff */
-int mp_montgomery_setup(const mp_int *n, mp_digit *rho)
+mp_err mp_montgomery_setup(const mp_int *n, mp_digit *rho)
 {
    mp_digit x, b;
 
@@ -44,12 +35,8 @@ int mp_montgomery_setup(const mp_int *n, mp_digit *rho)
 #endif
 
    /* rho = -1/m mod b */
-   *rho = (mp_digit)(((mp_word)1 << (mp_word)DIGIT_BIT) - x) & MP_MASK;
+   *rho = (mp_digit)(((mp_word)1 << (mp_word)MP_DIGIT_BIT) - x) & MP_MASK;
 
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mul.c b/libtommath/bn_mp_mul.c
index f83b1b7..c5e0438 100644
--- a/libtommath/bn_mp_mul.c
+++ b/libtommath/bn_mp_mul.c
@@ -1,64 +1,52 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MUL_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* high level multiplication (handles sign) */
-int mp_mul(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_mul(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int     res, neg;
-   neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;
+   mp_err err;
+   int min_len = MP_MIN(a->used, b->used),
+       max_len = MP_MAX(a->used, b->used),
+       digs = a->used + b->used + 1;
+   mp_sign neg = (a->sign == b->sign) ? MP_ZPOS : MP_NEG;
 
-   /* use Toom-Cook? */
-#ifdef BN_MP_TOOM_MUL_C
-   if (MIN(a->used, b->used) >= TOOM_MUL_CUTOFF) {
-      res = mp_toom_mul(a, b, c);
-   } else
-#endif
-#ifdef BN_MP_KARATSUBA_MUL_C
-      /* use Karatsuba? */
-      if (MIN(a->used, b->used) >= KARATSUBA_MUL_CUTOFF) {
-         res = mp_karatsuba_mul(a, b, c);
-      } else
-#endif
-      {
-         /* can we use the fast multiplier?
-          *
-          * The fast multiplier can be used if the output will
-          * have less than MP_WARRAY digits and the number of
-          * digits won't affect carry propagation
-          */
-         int     digs = a->used + b->used + 1;
-
-#ifdef BN_FAST_S_MP_MUL_DIGS_C
-         if ((digs < (int)MP_WARRAY) &&
-             (MIN(a->used, b->used) <=
-              (int)(1u << (((size_t)CHAR_BIT * sizeof(mp_word)) - (2u * (size_t)DIGIT_BIT))))) {
-            res = fast_s_mp_mul_digs(a, b, c, digs);
-         } else
-#endif
-         {
-#ifdef BN_S_MP_MUL_DIGS_C
-            res = s_mp_mul(a, b, c); /* uses s_mp_mul_digs */
-#else
-            res = MP_VAL;
-#endif
-         }
-      }
+   if (MP_HAS(S_MP_BALANCE_MUL) &&
+       /* Check sizes. The smaller one needs to be larger than the Karatsuba cut-off.
+        * The bigger one needs to be at least about one MP_KARATSUBA_MUL_CUTOFF bigger
+        * to make some sense, but it depends on architecture, OS, position of the
+        * stars... so YMMV.
+        * Using it to cut the input into slices small enough for fast_s_mp_mul_digs
+        * was actually slower on the author's machine, but YMMV.
+        */
+       (min_len >= MP_KARATSUBA_MUL_CUTOFF) &&
+       ((max_len / 2) >= MP_KARATSUBA_MUL_CUTOFF) &&
+       /* Not much effect was observed below a ratio of 1:2, but again: YMMV. */
+       (max_len >= (2 * min_len))) {
+      err = s_mp_balance_mul(a,b,c);
+   } else if (MP_HAS(S_MP_TOOM_MUL) &&
+              (min_len >= MP_TOOM_MUL_CUTOFF)) {
+      err = s_mp_toom_mul(a, b, c);
+   } else if (MP_HAS(S_MP_KARATSUBA_MUL) &&
+              (min_len >= MP_KARATSUBA_MUL_CUTOFF)) {
+      err = s_mp_karatsuba_mul(a, b, c);
+   } else if (MP_HAS(S_MP_MUL_DIGS_FAST) &&
+              /* can we use the fast multiplier?
+               *
+               * The fast multiplier can be used if the output will
+               * have less than MP_WARRAY digits and the number of
+               * digits won't affect carry propagation
+               */
+              (digs < PRIVATE_MP_WARRAY) &&
+              (min_len <= MP_MAXFAST)) {
+      err = s_mp_mul_digs_fast(a, b, c, digs);
+   } else if (MP_HAS(S_MP_MUL_DIGS)) {
+      err = s_mp_mul_digs(a, b, c, digs);
+   } else {
+      err = MP_VAL;
+   }
    c->sign = (c->used > 0) ? neg : MP_ZPOS;
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mul_2.c b/libtommath/bn_mp_mul_2.c
index 2ed5516..bc0691a 100644
--- a/libtommath/bn_mp_mul_2.c
+++ b/libtommath/bn_mp_mul_2.c
@@ -1,26 +1,18 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MUL_2_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* b = a*2 */
-int mp_mul_2(const mp_int *a, mp_int *b)
+mp_err mp_mul_2(const mp_int *a, mp_int *b)
 {
-   int     x, res, oldused;
+   int     x, oldused;
+   mp_err err;
 
    /* grow to accomodate result */
    if (b->alloc < (a->used + 1)) {
-      if ((res = mp_grow(b, a->used + 1)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(b, a->used + 1)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -43,7 +35,7 @@ int mp_mul_2(const mp_int *a, mp_int *b)
          /* get what will be the *next* carry bit from the
           * MSB of the current digit
           */
-         rr = *tmpa >> (mp_digit)(DIGIT_BIT - 1);
+         rr = *tmpa >> (mp_digit)(MP_DIGIT_BIT - 1);
 
          /* now shift up this digit, add in the carry [from the previous] */
          *tmpb++ = ((*tmpa++ << 1uL) | r) & MP_MASK;
@@ -64,16 +56,9 @@ int mp_mul_2(const mp_int *a, mp_int *b)
       /* now zero any excess digits on the destination
        * that we didn't write to
        */
-      tmpb = b->dp + b->used;
-      for (x = b->used; x < oldused; x++) {
-         *tmpb++ = 0;
-      }
+      MP_ZERO_DIGITS(b->dp + b->used, oldused - b->used);
    }
    b->sign = a->sign;
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mul_2d.c b/libtommath/bn_mp_mul_2d.c
index 9ea548d..87354de 100644
--- a/libtommath/bn_mp_mul_2d.c
+++ b/libtommath/bn_mp_mul_2d.c
@@ -1,45 +1,36 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MUL_2D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* shift left by a certain bit count */
-int mp_mul_2d(const mp_int *a, int b, mp_int *c)
+mp_err mp_mul_2d(const mp_int *a, int b, mp_int *c)
 {
    mp_digit d;
-   int      res;
+   mp_err   err;
 
    /* copy */
    if (a != c) {
-      if ((res = mp_copy(a, c)) != MP_OKAY) {
-         return res;
+      if ((err = mp_copy(a, c)) != MP_OKAY) {
+         return err;
       }
    }
 
-   if (c->alloc < (c->used + (b / DIGIT_BIT) + 1)) {
-      if ((res = mp_grow(c, c->used + (b / DIGIT_BIT) + 1)) != MP_OKAY) {
-         return res;
+   if (c->alloc < (c->used + (b / MP_DIGIT_BIT) + 1)) {
+      if ((err = mp_grow(c, c->used + (b / MP_DIGIT_BIT) + 1)) != MP_OKAY) {
+         return err;
       }
    }
 
    /* shift by as many digits in the bit count */
-   if (b >= DIGIT_BIT) {
-      if ((res = mp_lshd(c, b / DIGIT_BIT)) != MP_OKAY) {
-         return res;
+   if (b >= MP_DIGIT_BIT) {
+      if ((err = mp_lshd(c, b / MP_DIGIT_BIT)) != MP_OKAY) {
+         return err;
       }
    }
 
-   /* shift any bit count < DIGIT_BIT */
-   d = (mp_digit)(b % DIGIT_BIT);
+   /* shift any bit count < MP_DIGIT_BIT */
+   d = (mp_digit)(b % MP_DIGIT_BIT);
    if (d != 0u) {
       mp_digit *tmpc, shift, mask, r, rr;
       int x;
@@ -48,7 +39,7 @@ int mp_mul_2d(const mp_int *a, int b, mp_int *c)
       mask = ((mp_digit)1 << d) - (mp_digit)1;
 
       /* shift for msbs */
-      shift = (mp_digit)DIGIT_BIT - d;
+      shift = (mp_digit)MP_DIGIT_BIT - d;
 
       /* alias */
       tmpc = c->dp;
@@ -76,7 +67,3 @@ int mp_mul_2d(const mp_int *a, int b, mp_int *c)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mul_d.c b/libtommath/bn_mp_mul_d.c
index 936e133..b56dfa3 100644
--- a/libtommath/bn_mp_mul_d.c
+++ b/libtommath/bn_mp_mul_d.c
@@ -1,28 +1,20 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MUL_D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* multiply by a digit */
-int mp_mul_d(const mp_int *a, mp_digit b, mp_int *c)
+mp_err mp_mul_d(const mp_int *a, mp_digit b, mp_int *c)
 {
    mp_digit u, *tmpa, *tmpc;
    mp_word  r;
-   int      ix, res, olduse;
+   mp_err   err;
+   int      ix, olduse;
 
    /* make sure c is big enough to hold a*b */
    if (c->alloc < (a->used + 1)) {
-      if ((res = mp_grow(c, a->used + 1)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(c, a->used + 1)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -50,7 +42,7 @@ int mp_mul_d(const mp_int *a, mp_digit b, mp_int *c)
       *tmpc++ = (mp_digit)(r & (mp_word)MP_MASK);
 
       /* send carry into next iteration */
-      u       = (mp_digit)(r >> (mp_word)DIGIT_BIT);
+      u       = (mp_digit)(r >> (mp_word)MP_DIGIT_BIT);
    }
 
    /* store final carry [if any] and increment ix offset  */
@@ -58,9 +50,7 @@ int mp_mul_d(const mp_int *a, mp_digit b, mp_int *c)
    ++ix;
 
    /* now zero digits above the top */
-   while (ix++ < olduse) {
-      *tmpc++ = 0;
-   }
+   MP_ZERO_DIGITS(tmpc, olduse - ix);
 
    /* set used count */
    c->used = a->used + 1;
@@ -69,7 +59,3 @@ int mp_mul_d(const mp_int *a, mp_digit b, mp_int *c)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_mulmod.c b/libtommath/bn_mp_mulmod.c
index 4192452..160d162 100644
--- a/libtommath/bn_mp_mulmod.c
+++ b/libtommath/bn_mp_mulmod.c
@@ -1,37 +1,25 @@
 #include "tommath_private.h"
 #ifdef BN_MP_MULMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* d = a * b (mod c) */
-int mp_mulmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d)
+mp_err mp_mulmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d)
 {
-   int     res;
-   mp_int  t;
+   mp_err err;
+   mp_int t;
 
-   if ((res = mp_init_size(&t, c->used)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&t, c->used)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_mul(a, b, &t)) != MP_OKAY) {
-      mp_clear(&t);
-      return res;
+   if ((err = mp_mul(a, b, &t)) != MP_OKAY) {
+      goto LBL_ERR;
    }
-   res = mp_mod(&t, c, d);
+   err = mp_mod(&t, c, d);
+
+LBL_ERR:
    mp_clear(&t);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_n_root.c b/libtommath/bn_mp_n_root.c
deleted file mode 100644
index c14771f..0000000
--- a/libtommath/bn_mp_n_root.c
+++ /dev/null
@@ -1,27 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_N_ROOT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* wrapper function for mp_n_root_ex()
- * computes c = (a)**(1/b) such that (c)**b <= a and (c+1)**b > a
- */
-int mp_n_root(const mp_int *a, mp_digit b, mp_int *c)
-{
-   return mp_n_root_ex(a, b, c, 0);
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_n_root_ex.c b/libtommath/bn_mp_n_root_ex.c
deleted file mode 100644
index ebc08ba..0000000
--- a/libtommath/bn_mp_n_root_ex.c
+++ /dev/null
@@ -1,129 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_N_ROOT_EX_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* find the n'th root of an integer
- *
- * Result found such that (c)**b <= a and (c+1)**b > a
- *
- * This algorithm uses Newton's approximation
- * x[i+1] = x[i] - f(x[i])/f'(x[i])
- * which will find the root in log(N) time where
- * each step involves a fair bit.  This is not meant to
- * find huge roots [square and cube, etc].
- */
-int mp_n_root_ex(const mp_int *a, mp_digit b, mp_int *c, int fast)
-{
-   mp_int  t1, t2, t3, a_;
-   int     res;
-
-   /* input must be positive if b is even */
-   if (((b & 1u) == 0u) && (a->sign == MP_NEG)) {
-      return MP_VAL;
-   }
-
-   if ((res = mp_init(&t1)) != MP_OKAY) {
-      return res;
-   }
-
-   if ((res = mp_init(&t2)) != MP_OKAY) {
-      goto LBL_T1;
-   }
-
-   if ((res = mp_init(&t3)) != MP_OKAY) {
-      goto LBL_T2;
-   }
-
-   /* if a is negative fudge the sign but keep track */
-   a_ = *a;
-   a_.sign = MP_ZPOS;
-
-   /* t2 = 2 */
-   mp_set(&t2, 2uL);
-
-   do {
-      /* t1 = t2 */
-      if ((res = mp_copy(&t2, &t1)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-
-      /* t2 = t1 - ((t1**b - a) / (b * t1**(b-1))) */
-
-      /* t3 = t1**(b-1) */
-      if ((res = mp_expt_d_ex(&t1, b - 1u, &t3, fast)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-
-      /* numerator */
-      /* t2 = t1**b */
-      if ((res = mp_mul(&t3, &t1, &t2)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-
-      /* t2 = t1**b - a */
-      if ((res = mp_sub(&t2, &a_, &t2)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-
-      /* denominator */
-      /* t3 = t1**(b-1) * b  */
-      if ((res = mp_mul_d(&t3, b, &t3)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-
-      /* t3 = (t1**b - a)/(b * t1**(b-1)) */
-      if ((res = mp_div(&t2, &t3, &t3, NULL)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-
-      if ((res = mp_sub(&t1, &t3, &t2)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-   }  while (mp_cmp(&t1, &t2) != MP_EQ);
-
-   /* result can be off by a few so check */
-   for (;;) {
-      if ((res = mp_expt_d_ex(&t1, b, &t2, fast)) != MP_OKAY) {
-         goto LBL_T3;
-      }
-
-      if (mp_cmp(&t2, &a_) == MP_GT) {
-         if ((res = mp_sub_d(&t1, 1uL, &t1)) != MP_OKAY) {
-            goto LBL_T3;
-         }
-      } else {
-         break;
-      }
-   }
-
-   /* set the result */
-   mp_exch(&t1, c);
-
-   /* set the sign of the result */
-   c->sign = a->sign;
-
-   res = MP_OKAY;
-
-LBL_T3:
-   mp_clear(&t3);
-LBL_T2:
-   mp_clear(&t2);
-LBL_T1:
-   mp_clear(&t1);
-   return res;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_neg.c b/libtommath/bn_mp_neg.c
index 9020525..264d900 100644
--- a/libtommath/bn_mp_neg.c
+++ b/libtommath/bn_mp_neg.c
@@ -1,28 +1,19 @@
 #include "tommath_private.h"
 #ifdef BN_MP_NEG_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* b = -a */
-int mp_neg(const mp_int *a, mp_int *b)
+mp_err mp_neg(const mp_int *a, mp_int *b)
 {
-   int     res;
+   mp_err err;
    if (a != b) {
-      if ((res = mp_copy(a, b)) != MP_OKAY) {
-         return res;
+      if ((err = mp_copy(a, b)) != MP_OKAY) {
+         return err;
       }
    }
 
-   if (mp_iszero(b) != MP_YES) {
+   if (!MP_IS_ZERO(b)) {
       b->sign = (a->sign == MP_ZPOS) ? MP_NEG : MP_ZPOS;
    } else {
       b->sign = MP_ZPOS;
@@ -31,7 +22,3 @@ int mp_neg(const mp_int *a, mp_int *b)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_or.c b/libtommath/bn_mp_or.c
index 78a6b08..cdacbfb 100644
--- a/libtommath/bn_mp_or.c
+++ b/libtommath/bn_mp_or.c
@@ -6,7 +6,7 @@
 /* two complement or */
 mp_err mp_or(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int used = MAX(a->used, b->used) + 1, i;
+   int used = MP_MAX(a->used, b->used) + 1, i;
    mp_err err;
    mp_digit ac = 1, bc = 1, cc = 1;
    mp_sign csign = ((a->sign == MP_NEG) || (b->sign == MP_NEG)) ? MP_NEG : MP_ZPOS;
diff --git a/libtommath/bn_mp_pack.c b/libtommath/bn_mp_pack.c
new file mode 100644
index 0000000..6e00b6f
--- /dev/null
+++ b/libtommath/bn_mp_pack.c
@@ -0,0 +1,69 @@
+#include "tommath_private.h"
+#ifdef BN_MP_PACK_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* based on gmp's mpz_export.
+ * see http://gmplib.org/manual/Integer-Import-and-Export.html
+ */
+mp_err mp_pack(void *rop, size_t maxcount, size_t *written, mp_order order, size_t size,
+               mp_endian endian, size_t nails, const mp_int *op)
+{
+   mp_err err;
+   size_t odd_nails, nail_bytes, i, j, count;
+   unsigned char odd_nail_mask;
+
+   mp_int t;
+
+   count = mp_pack_count(op, nails, size);
+
+   if (count > maxcount) {
+      return MP_BUF;
+   }
+
+   if ((err = mp_init_copy(&t, op)) != MP_OKAY) {
+      return err;
+   }
+
+   if (endian == MP_NATIVE_ENDIAN) {
+      MP_GET_ENDIANNESS(endian);
+   }
+
+   odd_nails = (nails % 8u);
+   odd_nail_mask = 0xff;
+   for (i = 0u; i < odd_nails; ++i) {
+      odd_nail_mask ^= (unsigned char)(1u << (7u - i));
+   }
+   nail_bytes = nails / 8u;
+
+   for (i = 0u; i < count; ++i) {
+      for (j = 0u; j < size; ++j) {
+         unsigned char *byte = (unsigned char *)rop +
+                               (((order == MP_LSB_FIRST) ? i : ((count - 1u) - i)) * size) +
+                               ((endian == MP_LITTLE_ENDIAN) ? j : ((size - 1u) - j));
+
+         if (j >= (size - nail_bytes)) {
+            *byte = 0;
+            continue;
+         }
+
+         *byte = (unsigned char)((j == ((size - nail_bytes) - 1u)) ? (t.dp[0] & odd_nail_mask) : (t.dp[0] & 0xFFuL));
+
+         if ((err = mp_div_2d(&t, (j == ((size - nail_bytes) - 1u)) ? (int)(8u - odd_nails) : 8, &t, NULL)) != MP_OKAY) {
+            goto LBL_ERR;
+         }
+
+      }
+   }
+
+   if (written != NULL) {
+      *written = count;
+   }
+   err = MP_OKAY;
+
+LBL_ERR:
+   mp_clear(&t);
+   return err;
+}
+
+#endif
diff --git a/libtommath/bn_mp_pack_count.c b/libtommath/bn_mp_pack_count.c
new file mode 100644
index 0000000..dfecdf9
--- /dev/null
+++ b/libtommath/bn_mp_pack_count.c
@@ -0,0 +1,12 @@
+#include "tommath_private.h"
+#ifdef BN_MP_PACK_COUNT_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+size_t mp_pack_count(const mp_int *a, size_t nails, size_t size)
+{
+   size_t bits = (size_t)mp_count_bits(a);
+   return ((bits / ((size * 8u) - nails)) + (((bits % ((size * 8u) - nails)) != 0u) ? 1u : 0u));
+}
+
+#endif
diff --git a/libtommath/bn_mp_prime_fermat.c b/libtommath/bn_mp_prime_fermat.c
index 63ced96..af3e884 100644
--- a/libtommath/bn_mp_prime_fermat.c
+++ b/libtommath/bn_mp_prime_fermat.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_PRIME_FERMAT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* performs one Fermat test.
  *
@@ -20,10 +11,10 @@
  *
  * Sets result to 1 if the congruence holds, or zero otherwise.
  */
-int mp_prime_fermat(const mp_int *a, const mp_int *b, int *result)
+mp_err mp_prime_fermat(const mp_int *a, const mp_int *b, mp_bool *result)
 {
    mp_int  t;
-   int     err;
+   mp_err  err;
 
    /* default to composite  */
    *result = MP_NO;
@@ -54,7 +45,3 @@ LBL_T:
    return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_frobenius_underwood.c b/libtommath/bn_mp_prime_frobenius_underwood.c
index 4ceb51e..253e8d5 100644
--- a/libtommath/bn_mp_prime_frobenius_underwood.c
+++ b/libtommath/bn_mp_prime_frobenius_underwood.c
@@ -1,22 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_PRIME_FROBENIUS_UNDERWOOD_C
 
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /*
  *  See file bn_mp_prime_is_prime.c or the documentation in doc/bn.tex for the details
  */
-#ifndef LTM_USE_FIPS_ONLY
+#ifndef LTM_USE_ONLY_MR
 
 #ifdef MP_8BIT
 /*
@@ -32,17 +23,17 @@
 #else
 #define LTM_FROBENIUS_UNDERWOOD_A 32764
 #endif
-int mp_prime_frobenius_underwood(const mp_int *N, int *result)
+mp_err mp_prime_frobenius_underwood(const mp_int *N, mp_bool *result)
 {
    mp_int T1z, T2z, Np1z, sz, tz;
 
-   int a, ap2, length, i, j, isset;
-   int e;
+   int a, ap2, length, i, j;
+   mp_err err;
 
    *result = MP_NO;
 
-   if ((e = mp_init_multi(&T1z, &T2z, &Np1z, &sz, &tz, NULL)) != MP_OKAY) {
-      return e;
+   if ((err = mp_init_multi(&T1z, &T2z, &Np1z, &sz, &tz, NULL)) != MP_OKAY) {
+      return err;
    }
 
    for (a = 0; a < LTM_FROBENIUS_UNDERWOOD_A; a++) {
@@ -52,21 +43,13 @@ int mp_prime_frobenius_underwood(const mp_int *N, int *result)
          continue;
       }
       /* (32764^2 - 4) < 2^31, no bigint for >MP_8BIT needed) */
-      if ((e = mp_set_long(&T1z, (unsigned long)a)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
+      mp_set_u32(&T1z, (uint32_t)a);
 
-      if ((e = mp_sqr(&T1z, &T1z)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
+      if ((err = mp_sqr(&T1z, &T1z)) != MP_OKAY)                  goto LBL_FU_ERR;
 
-      if ((e = mp_sub_d(&T1z, 4uL, &T1z)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
+      if ((err = mp_sub_d(&T1z, 4uL, &T1z)) != MP_OKAY)           goto LBL_FU_ERR;
 
-      if ((e = mp_kronecker(&T1z, N, &j)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
+      if ((err = mp_kronecker(&T1z, N, &j)) != MP_OKAY)           goto LBL_FU_ERR;
 
       if (j == -1) {
          break;
@@ -79,26 +62,18 @@ int mp_prime_frobenius_underwood(const mp_int *N, int *result)
    }
    /* Tell it a composite and set return value accordingly */
    if (a >= LTM_FROBENIUS_UNDERWOOD_A) {
-      e = MP_ITER;
+      err = MP_ITER;
       goto LBL_FU_ERR;
    }
    /* Composite if N and (a+4)*(2*a+5) are not coprime */
-   if ((e = mp_set_long(&T1z, (unsigned long)((a+4)*((2*a)+5)))) != MP_OKAY) {
-      goto LBL_FU_ERR;
-   }
+   mp_set_u32(&T1z, (uint32_t)((a+4)*((2*a)+5)));
 
-   if ((e = mp_gcd(N, &T1z, &T1z)) != MP_OKAY) {
-      goto LBL_FU_ERR;
-   }
+   if ((err = mp_gcd(N, &T1z, &T1z)) != MP_OKAY)                  goto LBL_FU_ERR;
 
-   if (!((T1z.used == 1) && (T1z.dp[0] == 1u))) {
-      goto LBL_FU_ERR;
-   }
+   if (!((T1z.used == 1) && (T1z.dp[0] == 1u)))                   goto LBL_FU_ERR;
 
    ap2 = a + 2;
-   if ((e = mp_add_d(N, 1uL, &Np1z)) != MP_OKAY) {
-      goto LBL_FU_ERR;
-   }
+   if ((err = mp_add_d(N, 1uL, &Np1z)) != MP_OKAY)                goto LBL_FU_ERR;
 
    mp_set(&sz, 1uL);
    mp_set(&tz, 2uL);
@@ -110,89 +85,48 @@ int mp_prime_frobenius_underwood(const mp_int *N, int *result)
        * tz   = ((tz-sz)*(tz+sz))%N;
        * sz   = temp;
        */
-      if ((e = mp_mul_2(&tz, &T2z)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
+      if ((err = mp_mul_2(&tz, &T2z)) != MP_OKAY)                 goto LBL_FU_ERR;
 
       /* a = 0 at about 50% of the cases (non-square and odd input) */
       if (a != 0) {
-         if ((e = mp_mul_d(&sz, (mp_digit)a, &T1z)) != MP_OKAY) {
-            goto LBL_FU_ERR;
-         }
-         if ((e = mp_add(&T1z, &T2z, &T2z)) != MP_OKAY) {
-            goto LBL_FU_ERR;
-         }
+         if ((err = mp_mul_d(&sz, (mp_digit)a, &T1z)) != MP_OKAY) goto LBL_FU_ERR;
+         if ((err = mp_add(&T1z, &T2z, &T2z)) != MP_OKAY)         goto LBL_FU_ERR;
       }
 
-      if ((e = mp_mul(&T2z, &sz, &T1z)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
-      if ((e = mp_sub(&tz, &sz, &T2z)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
-      if ((e = mp_add(&sz, &tz, &sz)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
-      if ((e = mp_mul(&sz, &T2z, &tz)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
-      if ((e = mp_mod(&tz, N, &tz)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
-      if ((e = mp_mod(&T1z, N, &sz)) != MP_OKAY) {
-         goto LBL_FU_ERR;
-      }
-      if ((isset = mp_get_bit(&Np1z, i)) == MP_VAL) {
-         e = isset;
-         goto LBL_FU_ERR;
-      }
-      if (isset == MP_YES) {
+      if ((err = mp_mul(&T2z, &sz, &T1z)) != MP_OKAY)             goto LBL_FU_ERR;
+      if ((err = mp_sub(&tz, &sz, &T2z)) != MP_OKAY)              goto LBL_FU_ERR;
+      if ((err = mp_add(&sz, &tz, &sz)) != MP_OKAY)               goto LBL_FU_ERR;
+      if ((err = mp_mul(&sz, &T2z, &tz)) != MP_OKAY)              goto LBL_FU_ERR;
+      if ((err = mp_mod(&tz, N, &tz)) != MP_OKAY)                 goto LBL_FU_ERR;
+      if ((err = mp_mod(&T1z, N, &sz)) != MP_OKAY)                goto LBL_FU_ERR;
+      if (s_mp_get_bit(&Np1z, (unsigned int)i) == MP_YES) {
          /*
           *  temp = (a+2) * sz + tz
           *  tz   = 2 * tz - sz
           *  sz   = temp
           */
          if (a == 0) {
-            if ((e = mp_mul_2(&sz, &T1z)) != MP_OKAY) {
-               goto LBL_FU_ERR;
-            }
+            if ((err = mp_mul_2(&sz, &T1z)) != MP_OKAY)           goto LBL_FU_ERR;
          } else {
-            if ((e = mp_mul_d(&sz, (mp_digit)ap2, &T1z)) != MP_OKAY) {
-               goto LBL_FU_ERR;
-            }
-         }
-         if ((e = mp_add(&T1z, &tz, &T1z)) != MP_OKAY) {
-            goto LBL_FU_ERR;
-         }
-         if ((e = mp_mul_2(&tz, &T2z)) != MP_OKAY) {
-            goto LBL_FU_ERR;
-         }
-         if ((e = mp_sub(&T2z, &sz, &tz)) != MP_OKAY) {
-            goto LBL_FU_ERR;
+            if ((err = mp_mul_d(&sz, (mp_digit)ap2, &T1z)) != MP_OKAY) goto LBL_FU_ERR;
          }
+         if ((err = mp_add(&T1z, &tz, &T1z)) != MP_OKAY)          goto LBL_FU_ERR;
+         if ((err = mp_mul_2(&tz, &T2z)) != MP_OKAY)              goto LBL_FU_ERR;
+         if ((err = mp_sub(&T2z, &sz, &tz)) != MP_OKAY)           goto LBL_FU_ERR;
          mp_exch(&sz, &T1z);
       }
    }
 
-   if ((e = mp_set_long(&T1z, (unsigned long)((2 * a) + 5))) != MP_OKAY) {
-      goto LBL_FU_ERR;
-   }
-   if ((e = mp_mod(&T1z, N, &T1z)) != MP_OKAY) {
-      goto LBL_FU_ERR;
-   }
-   if ((mp_iszero(&sz) != MP_NO) && (mp_cmp(&tz, &T1z) == MP_EQ)) {
+   mp_set_u32(&T1z, (uint32_t)((2 * a) + 5));
+   if ((err = mp_mod(&T1z, N, &T1z)) != MP_OKAY)                  goto LBL_FU_ERR;
+   if (MP_IS_ZERO(&sz) && (mp_cmp(&tz, &T1z) == MP_EQ)) {
       *result = MP_YES;
-      goto LBL_FU_ERR;
    }
 
 LBL_FU_ERR:
    mp_clear_multi(&tz, &sz, &Np1z, &T2z, &T1z, NULL);
-   return e;
+   return err;
 }
 
 #endif
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_is_divisible.c b/libtommath/bn_mp_prime_is_divisible.c
deleted file mode 100644
index 0e6e2f3..0000000
--- a/libtommath/bn_mp_prime_is_divisible.c
+++ /dev/null
@@ -1,47 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_PRIME_IS_DIVISIBLE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* determines if an integers is divisible by one
- * of the first PRIME_SIZE primes or not
- *
- * sets result to 0 if not, 1 if yes
- */
-int mp_prime_is_divisible(const mp_int *a, int *result)
-{
-   int     err, ix;
-   mp_digit res;
-
-   /* default to not */
-   *result = MP_NO;
-
-   for (ix = 0; ix < PRIME_SIZE; ix++) {
-      /* what is a mod LBL_prime_tab[ix] */
-      if ((err = mp_mod_d(a, ltm_prime_tab[ix], &res)) != MP_OKAY) {
-         return err;
-      }
-
-      /* is the residue zero? */
-      if (res == 0u) {
-         *result = MP_YES;
-         return MP_OKAY;
-      }
-   }
-
-   return MP_OKAY;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_is_prime.c b/libtommath/bn_mp_prime_is_prime.c
index 15637c8..7f9fc0b 100644
--- a/libtommath/bn_mp_prime_is_prime.c
+++ b/libtommath/bn_mp_prime_is_prime.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_PRIME_IS_PRIME_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* portable integer log of two with small footprint */
 static unsigned int s_floor_ilog2(int value)
@@ -23,61 +14,58 @@ static unsigned int s_floor_ilog2(int value)
 }
 
 
-int mp_prime_is_prime(const mp_int *a, int t, int *result)
+mp_err mp_prime_is_prime(const mp_int *a, int t, mp_bool *result)
 {
    mp_int  b;
-   int     ix, err, res, p_max = 0, size_a, len;
+   int     ix, p_max = 0, size_a, len;
+   mp_bool res;
+   mp_err  err;
    unsigned int fips_rand, mask;
 
    /* default to no */
    *result = MP_NO;
 
-   /* valid value of t? */
-   if (t > PRIME_SIZE) {
-      return MP_VAL;
-   }
-
    /* Some shortcuts */
    /* N > 3 */
    if (a->used == 1) {
       if ((a->dp[0] == 0u) || (a->dp[0] == 1u)) {
-         *result = 0;
+         *result = MP_NO;
          return MP_OKAY;
       }
       if (a->dp[0] == 2u) {
-         *result = 1;
+         *result = MP_YES;
          return MP_OKAY;
       }
    }
 
    /* N must be odd */
-   if (mp_iseven(a) == MP_YES) {
+   if (MP_IS_EVEN(a)) {
       return MP_OKAY;
    }
    /* N is not a perfect square: floor(sqrt(N))^2 != N */
    if ((err = mp_is_square(a, &res)) != MP_OKAY) {
       return err;
    }
-   if (res != 0) {
+   if (res != MP_NO) {
       return MP_OKAY;
    }
 
    /* is the input equal to one of the primes in the table? */
-   for (ix = 0; ix < PRIME_SIZE; ix++) {
-      if (mp_cmp_d(a, ltm_prime_tab[ix]) == MP_EQ) {
+   for (ix = 0; ix < PRIVATE_MP_PRIME_TAB_SIZE; ix++) {
+      if (mp_cmp_d(a, s_mp_prime_tab[ix]) == MP_EQ) {
          *result = MP_YES;
          return MP_OKAY;
       }
    }
 #ifdef MP_8BIT
    /* The search in the loop above was exhaustive in this case */
-   if ((a->used == 1) && (PRIME_SIZE >= 31)) {
+   if ((a->used == 1) && (PRIVATE_MP_PRIME_TAB_SIZE >= 31)) {
       return MP_OKAY;
    }
 #endif
 
    /* first perform trial division */
-   if ((err = mp_prime_is_divisible(a, &res)) != MP_OKAY) {
+   if ((err = s_mp_prime_is_divisible(a, &res)) != MP_OKAY) {
       return err;
    }
 
@@ -114,10 +102,10 @@ int mp_prime_is_prime(const mp_int *a, int t, int *result)
 
    /*
     * Both, the Frobenius-Underwood test and the the Lucas-Selfridge test are quite
-    * slow so if speed is an issue, define LTM_USE_FIPS_ONLY to use M-R tests with
+    * slow so if speed is an issue, define LTM_USE_ONLY_MR to use M-R tests with
     * bases 2, 3 and t random bases.
     */
-#ifndef LTM_USE_FIPS_ONLY
+#ifndef LTM_USE_ONLY_MR
    if (t >= 0) {
       /*
        * Use a Frobenius-Underwood test instead of the Lucas-Selfridge test for
@@ -149,44 +137,14 @@ int mp_prime_is_prime(const mp_int *a, int t, int *result)
    }
 
    /*
-      abs(t) extra rounds of M-R to extend the range of primes it can find if t < 0.
       Only recommended if the input range is known to be < 3317044064679887385961981
 
-      It uses the bases for a deterministic M-R test if input < 3317044064679887385961981
+      It uses the bases necessary for a deterministic M-R test if the input is
+      smaller than  3317044064679887385961981
       The caller has to check the size.
-
-      Not for cryptographic use because with known bases strong M-R pseudoprimes can
-      be constructed. Use at least one M-R test with a random base (t >= 1).
-
-      The 1119 bit large number
-
-      80383745745363949125707961434194210813883768828755814583748891752229742737653\
-      33652186502336163960045457915042023603208766569966760987284043965408232928738\
-      79185086916685732826776177102938969773947016708230428687109997439976544144845\
-      34115587245063340927902227529622941498423068816854043264575340183297861112989\
-      60644845216191652872597534901
-
-      has been constructed by F. Arnault (F. Arnault, "Rabin-Miller primality test:
-      composite numbers which pass it.",  Mathematics of Computation, 1995, 64. Jg.,
-      Nr. 209, S. 355-361), is a semiprime with the two factors
-
-      40095821663949960541830645208454685300518816604113250877450620473800321707011\
-      96242716223191597219733582163165085358166969145233813917169287527980445796800\
-      452592031836601
-
-      20047910831974980270915322604227342650259408302056625438725310236900160853505\
-      98121358111595798609866791081582542679083484572616906958584643763990222898400\
-      226296015918301
-
-      and it is a strong pseudoprime to all forty-six prime M-R bases up to 200
-
-      It does not fail the strong Bailley-PSP test as implemented here, it is just
-      given as an example, if not the reason to use the BPSW-test instead of M-R-tests
-      with a sequence of primes 2...n.
-
+      TODO: can be made a bit finer grained but comparing is not free.
    */
    if (t < 0) {
-      t = -t;
       /*
           Sorenson, Jonathan; Webster, Jonathan (2015).
            "Strong Pseudoprimes to Twelve Prime Bases".
@@ -212,18 +170,9 @@ int mp_prime_is_prime(const mp_int *a, int t, int *result)
          }
       }
 
-      /* for compatibility with the current API (well, compatible within a sign's width) */
-      if (p_max < t) {
-         p_max = t;
-      }
-
-      if (p_max > PRIME_SIZE) {
-         err = MP_VAL;
-         goto LBL_B;
-      }
       /* we did bases 2 and 3  already, skip them */
       for (ix = 2; ix < p_max; ix++) {
-         mp_set(&b, ltm_prime_tab[ix]);
+         mp_set(&b, s_mp_prime_tab[ix]);
          if ((err = mp_prime_miller_rabin(a, &b, &res)) != MP_OKAY) {
             goto LBL_B;
          }
@@ -296,19 +245,19 @@ int mp_prime_is_prime(const mp_int *a, int t, int *result)
           * One 8-bit digit is too small, so concatenate two if the size of
           * unsigned int allows for it.
           */
-         if (((sizeof(unsigned int) * CHAR_BIT)/2) >= (sizeof(mp_digit) * CHAR_BIT)) {
+         if ((MP_SIZEOF_BITS(unsigned int)/2) >= MP_SIZEOF_BITS(mp_digit)) {
             if ((err = mp_rand(&b, 1)) != MP_OKAY) {
                goto LBL_B;
             }
-            fips_rand <<= sizeof(mp_digit) * CHAR_BIT;
+            fips_rand <<= MP_SIZEOF_BITS(mp_digit);
             fips_rand |= (unsigned int) b.dp[0];
             fips_rand &= mask;
          }
 #endif
-         if (fips_rand > (unsigned int)(INT_MAX - DIGIT_BIT)) {
-            len = INT_MAX / DIGIT_BIT;
+         if (fips_rand > (unsigned int)(INT_MAX - MP_DIGIT_BIT)) {
+            len = INT_MAX / MP_DIGIT_BIT;
          } else {
-            len = (((int)fips_rand + DIGIT_BIT) / DIGIT_BIT);
+            len = (((int)fips_rand + MP_DIGIT_BIT) / MP_DIGIT_BIT);
          }
          /*  Unlikely. */
          if (len < 0) {
@@ -332,16 +281,15 @@ int mp_prime_is_prime(const mp_int *a, int t, int *result)
          }
          /*
           * That number might got too big and the witness has to be
-          * smaller than or equal to "a"
+          * smaller than "a"
           */
          len = mp_count_bits(&b);
-         if (len > size_a) {
-            len = len - size_a;
+         if (len >= size_a) {
+            len = (len - size_a) + 1;
             if ((err = mp_div_2d(&b, len, &b, NULL)) != MP_OKAY) {
                goto LBL_B;
             }
          }
-
          /* Although the chance for b <= 3 is miniscule, try again. */
          if (mp_cmp_d(&b, 3uL) != MP_GT) {
             ix--;
@@ -364,7 +312,3 @@ LBL_B:
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_miller_rabin.c b/libtommath/bn_mp_prime_miller_rabin.c
index a12e533..96470db 100644
--- a/libtommath/bn_mp_prime_miller_rabin.c
+++ b/libtommath/bn_mp_prime_miller_rabin.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_PRIME_MILLER_RABIN_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* Miller-Rabin test of "a" to the base of "b" as described in
  * HAC pp. 139 Algorithm 4.24
@@ -19,10 +10,11 @@
  * Randomly the chance of error is no more than 1/4 and often
  * very much lower.
  */
-int mp_prime_miller_rabin(const mp_int *a, const mp_int *b, int *result)
+mp_err mp_prime_miller_rabin(const mp_int *a, const mp_int *b, mp_bool *result)
 {
    mp_int  n1, y, r;
-   int     s, j, err;
+   mp_err  err;
+   int     s, j;
 
    /* default */
    *result = MP_NO;
@@ -97,7 +89,3 @@ LBL_N1:
    return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_next_prime.c b/libtommath/bn_mp_prime_next_prime.c
index 28256ca..d656565 100644
--- a/libtommath/bn_mp_prime_next_prime.c
+++ b/libtommath/bn_mp_prime_next_prime.c
@@ -1,63 +1,43 @@
 #include "tommath_private.h"
 #ifdef BN_MP_PRIME_NEXT_PRIME_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* finds the next prime after the number "a" using "t" trials
  * of Miller-Rabin.
  *
  * bbs_style = 1 means the prime must be congruent to 3 mod 4
  */
-int mp_prime_next_prime(mp_int *a, int t, int bbs_style)
+mp_err mp_prime_next_prime(mp_int *a, int t, int bbs_style)
 {
-   int      err, res = MP_NO, x, y;
-   mp_digit res_tab[PRIME_SIZE], step, kstep;
+   int      x, y;
+   mp_ord   cmp;
+   mp_err   err;
+   mp_bool  res = MP_NO;
+   mp_digit res_tab[PRIVATE_MP_PRIME_TAB_SIZE], step, kstep;
    mp_int   b;
 
    /* force positive */
    a->sign = MP_ZPOS;
 
    /* simple algo if a is less than the largest prime in the table */
-   if (mp_cmp_d(a, ltm_prime_tab[PRIME_SIZE-1]) == MP_LT) {
-      /* find which prime it is bigger than */
-      for (x = PRIME_SIZE - 2; x >= 0; x--) {
-         if (mp_cmp_d(a, ltm_prime_tab[x]) != MP_LT) {
-            if (bbs_style == 1) {
-               /* ok we found a prime smaller or
-                * equal [so the next is larger]
-                *
-                * however, the prime must be
-                * congruent to 3 mod 4
-                */
-               if ((ltm_prime_tab[x + 1] & 3u) != 3u) {
-                  /* scan upwards for a prime congruent to 3 mod 4 */
-                  for (y = x + 1; y < PRIME_SIZE; y++) {
-                     if ((ltm_prime_tab[y] & 3u) == 3u) {
-                        mp_set(a, ltm_prime_tab[y]);
-                        return MP_OKAY;
-                     }
-                  }
-               }
+   if (mp_cmp_d(a, s_mp_prime_tab[PRIVATE_MP_PRIME_TAB_SIZE-1]) == MP_LT) {
+      /* find which prime it is bigger than "a" */
+      for (x = 0; x < PRIVATE_MP_PRIME_TAB_SIZE; x++) {
+         cmp = mp_cmp_d(a, s_mp_prime_tab[x]);
+         if (cmp == MP_EQ) {
+            continue;
+         }
+         if (cmp != MP_GT) {
+            if ((bbs_style == 1) && ((s_mp_prime_tab[x] & 3u) != 3u)) {
+               /* try again until we get a prime congruent to 3 mod 4 */
+               continue;
             } else {
-               mp_set(a, ltm_prime_tab[x + 1]);
+               mp_set(a, s_mp_prime_tab[x]);
                return MP_OKAY;
             }
          }
       }
-      /* at this point a maybe 1 */
-      if (mp_cmp_d(a, 1uL) == MP_EQ) {
-         mp_set(a, 2uL);
-         return MP_OKAY;
-      }
       /* fall through to the sieve */
    }
 
@@ -75,10 +55,10 @@ int mp_prime_next_prime(mp_int *a, int t, int bbs_style)
       if ((a->dp[0] & 3u) != 3u) {
          if ((err = mp_sub_d(a, (a->dp[0] & 3u) + 1u, a)) != MP_OKAY) {
             return err;
-         };
+         }
       }
    } else {
-      if (mp_iseven(a) == MP_YES) {
+      if (MP_IS_EVEN(a)) {
          /* force odd */
          if ((err = mp_sub_d(a, 1uL, a)) != MP_OKAY) {
             return err;
@@ -87,8 +67,8 @@ int mp_prime_next_prime(mp_int *a, int t, int bbs_style)
    }
 
    /* generate the restable */
-   for (x = 1; x < PRIME_SIZE; x++) {
-      if ((err = mp_mod_d(a, ltm_prime_tab[x], res_tab + x)) != MP_OKAY) {
+   for (x = 1; x < PRIVATE_MP_PRIME_TAB_SIZE; x++) {
+      if ((err = mp_mod_d(a, s_mp_prime_tab[x], res_tab + x)) != MP_OKAY) {
          return err;
       }
    }
@@ -109,13 +89,13 @@ int mp_prime_next_prime(mp_int *a, int t, int bbs_style)
          step += kstep;
 
          /* compute the new residue without using division */
-         for (x = 1; x < PRIME_SIZE; x++) {
+         for (x = 1; x < PRIVATE_MP_PRIME_TAB_SIZE; x++) {
             /* add the step to each residue */
             res_tab[x] += kstep;
 
             /* subtract the modulus [instead of using division] */
-            if (res_tab[x] >= ltm_prime_tab[x]) {
-               res_tab[x]  -= ltm_prime_tab[x];
+            if (res_tab[x] >= s_mp_prime_tab[x]) {
+               res_tab[x]  -= s_mp_prime_tab[x];
             }
 
             /* set flag if zero */
@@ -123,15 +103,15 @@ int mp_prime_next_prime(mp_int *a, int t, int bbs_style)
                y = 1;
             }
          }
-      } while ((y == 1) && (step < (((mp_digit)1 << DIGIT_BIT) - kstep)));
+      } while ((y == 1) && (step < (((mp_digit)1 << MP_DIGIT_BIT) - kstep)));
 
       /* add the step */
       if ((err = mp_add_d(a, step, a)) != MP_OKAY) {
          goto LBL_ERR;
       }
 
-      /* if didn't pass sieve and step == MAX then skip test */
-      if ((y == 1) && (step >= (((mp_digit)1 << DIGIT_BIT) - kstep))) {
+      /* if didn't pass sieve and step == MP_MAX then skip test */
+      if ((y == 1) && (step >= (((mp_digit)1 << MP_DIGIT_BIT) - kstep))) {
          continue;
       }
 
@@ -150,7 +130,3 @@ LBL_ERR:
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_rabin_miller_trials.c b/libtommath/bn_mp_prime_rabin_miller_trials.c
index 1c0a748..8bbaf6c 100644
--- a/libtommath/bn_mp_prime_rabin_miller_trials.c
+++ b/libtommath/bn_mp_prime_rabin_miller_trials.c
@@ -1,39 +1,34 @@
 #include "tommath_private.h"
 #ifdef BN_MP_PRIME_RABIN_MILLER_TRIALS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 static const struct {
    int k, t;
 } sizes[] = {
-   {    80,    -1 }, /* Use deterministic algorithm for size <= 80 bits */
-   {    81,    39 },
-   {    96,    37 },
-   {   128,    32 },
-   {   160,    27 },
-   {   192,    21 },
-   {   256,    16 },
-   {   384,    10 },
-   {   512,     7 },
-   {   640,     6 },
-   {   768,     5 },
-   {   896,     4 },
-   {  1024,     4 },
-   {  2048,     2 },
-   {  4096,     1 },
+   {    80, -1 }, /* Use deterministic algorithm for size <= 80 bits */
+   {    81, 37 }, /* max. error = 2^(-96)*/
+   {    96, 32 }, /* max. error = 2^(-96)*/
+   {   128, 40 }, /* max. error = 2^(-112)*/
+   {   160, 35 }, /* max. error = 2^(-112)*/
+   {   256, 27 }, /* max. error = 2^(-128)*/
+   {   384, 16 }, /* max. error = 2^(-128)*/
+   {   512, 18 }, /* max. error = 2^(-160)*/
+   {   768, 11 }, /* max. error = 2^(-160)*/
+   {   896, 10 }, /* max. error = 2^(-160)*/
+   {  1024, 12 }, /* max. error = 2^(-192)*/
+   {  1536, 8  }, /* max. error = 2^(-192)*/
+   {  2048, 6  }, /* max. error = 2^(-192)*/
+   {  3072, 4  }, /* max. error = 2^(-192)*/
+   {  4096, 5  }, /* max. error = 2^(-256)*/
+   {  5120, 4  }, /* max. error = 2^(-256)*/
+   {  6144, 4  }, /* max. error = 2^(-256)*/
+   {  8192, 3  }, /* max. error = 2^(-256)*/
+   {  9216, 3  }, /* max. error = 2^(-256)*/
+   { 10240, 2  }  /* For bigger keysizes use always at least 2 Rounds */
 };
 
-/* returns # of RM trials required for a given bit size and max. error of 2^(-96)*/
+/* returns # of RM trials required for a given bit size */
 int mp_prime_rabin_miller_trials(int size)
 {
    int x;
@@ -45,12 +40,8 @@ int mp_prime_rabin_miller_trials(int size)
          return (x == 0) ? sizes[0].t : sizes[x - 1].t;
       }
    }
-   return sizes[x-1].t + 1;
+   return sizes[x-1].t;
 }
 
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_random_ex.c b/libtommath/bn_mp_prime_rand.c
index 0ca29ec..4530e9a 100644
--- a/libtommath/bn_mp_prime_random_ex.c
+++ b/libtommath/bn_mp_prime_rand.c
@@ -1,24 +1,15 @@
 #include "tommath_private.h"
-#ifdef BN_MP_PRIME_RANDOM_EX_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_MP_PRIME_RAND_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* makes a truly random prime of a given size (bits),
  *
  * Flags are as follows:
  *
- *   LTM_PRIME_BBS      - make prime congruent to 3 mod 4
- *   LTM_PRIME_SAFE     - make sure (p-1)/2 is prime as well (implies LTM_PRIME_BBS)
- *   LTM_PRIME_2MSB_ON  - make the 2nd highest bit one
+ *   MP_PRIME_BBS      - make prime congruent to 3 mod 4
+ *   MP_PRIME_SAFE     - make sure (p-1)/2 is prime as well (implies MP_PRIME_BBS)
+ *   MP_PRIME_2MSB_ON  - make the 2nd highest bit one
  *
  * You have to supply a callback which fills in a buffer with random bytes.  "dat" is a parameter you can
  * have passed to the callback (e.g. a state or something).  This function doesn't use "dat" itself
@@ -27,44 +18,46 @@
  */
 
 /* This is possibly the mother of all prime generation functions, muahahahahaha! */
-int mp_prime_random_ex(mp_int *a, int t, int size, int flags, ltm_prime_callback cb, void *dat)
+mp_err s_mp_prime_random_ex(mp_int *a, int t, int size, int flags, private_mp_prime_callback cb, void *dat)
 {
    unsigned char *tmp, maskAND, maskOR_msb, maskOR_lsb;
-   int res, err, bsize, maskOR_msb_offset;
+   int bsize, maskOR_msb_offset;
+   mp_bool res;
+   mp_err err;
 
    /* sanity check the input */
    if ((size <= 1) || (t <= 0)) {
       return MP_VAL;
    }
 
-   /* LTM_PRIME_SAFE implies LTM_PRIME_BBS */
-   if ((flags & LTM_PRIME_SAFE) != 0) {
-      flags |= LTM_PRIME_BBS;
+   /* MP_PRIME_SAFE implies MP_PRIME_BBS */
+   if ((flags & MP_PRIME_SAFE) != 0) {
+      flags |= MP_PRIME_BBS;
    }
 
    /* calc the byte size */
    bsize = (size>>3) + ((size&7)?1:0);
 
    /* we need a buffer of bsize bytes */
-   tmp = (unsigned char *) XMALLOC((size_t)bsize);
+   tmp = (unsigned char *) MP_MALLOC((size_t)bsize);
    if (tmp == NULL) {
       return MP_MEM;
    }
 
    /* calc the maskAND value for the MSbyte*/
-   maskAND = ((size&7) == 0) ? 0xFF : (unsigned char)(0xFF >> (8 - (size & 7)));
+   maskAND = ((size&7) == 0) ? 0xFFu : (unsigned char)(0xFFu >> (8 - (size & 7)));
 
    /* calc the maskOR_msb */
    maskOR_msb        = 0;
    maskOR_msb_offset = ((size & 7) == 1) ? 1 : 0;
-   if ((flags & LTM_PRIME_2MSB_ON) != 0) {
+   if ((flags & MP_PRIME_2MSB_ON) != 0) {
       maskOR_msb       |= (unsigned char)(0x80 >> ((9 - size) & 7));
    }
 
    /* get the maskOR_lsb */
-   maskOR_lsb         = 1;
-   if ((flags & LTM_PRIME_BBS) != 0) {
-      maskOR_lsb     |= 3;
+   maskOR_lsb         = 1u;
+   if ((flags & MP_PRIME_BBS) != 0) {
+      maskOR_lsb     |= 3u;
    }
 
    do {
@@ -83,7 +76,8 @@ int mp_prime_random_ex(mp_int *a, int t, int size, int flags, ltm_prime_callback
       tmp[bsize-1]             |= maskOR_lsb;
 
       /* read it in */
-      if ((err = mp_read_unsigned_bin(a, tmp, bsize)) != MP_OKAY) {
+      /* TODO: casting only for now until all lengths have been changed to the type "size_t"*/
+      if ((err = mp_from_ubin(a, tmp, (size_t)bsize)) != MP_OKAY) {
          goto error;
       }
 
@@ -95,7 +89,7 @@ int mp_prime_random_ex(mp_int *a, int t, int size, int flags, ltm_prime_callback
          continue;
       }
 
-      if ((flags & LTM_PRIME_SAFE) != 0) {
+      if ((flags & MP_PRIME_SAFE) != 0) {
          /* see if (a-1)/2 is prime */
          if ((err = mp_sub_d(a, 1uL, a)) != MP_OKAY) {
             goto error;
@@ -111,7 +105,7 @@ int mp_prime_random_ex(mp_int *a, int t, int size, int flags, ltm_prime_callback
       }
    } while (res == MP_NO);
 
-   if ((flags & LTM_PRIME_SAFE) != 0) {
+   if ((flags & MP_PRIME_SAFE) != 0) {
       /* restore a to the original value */
       if ((err = mp_mul_2(a, a)) != MP_OKAY) {
          goto error;
@@ -123,13 +117,25 @@ int mp_prime_random_ex(mp_int *a, int t, int size, int flags, ltm_prime_callback
 
    err = MP_OKAY;
 error:
-   XFREE(tmp, bsize);
+   MP_FREE_BUFFER(tmp, (size_t)bsize);
    return err;
 }
 
+static int s_mp_rand_cb(unsigned char *dst, int len, void *dat)
+{
+   (void)dat;
+   if (len <= 0) {
+      return len;
+   }
+   if (s_mp_rand_source(dst, (size_t)len) != MP_OKAY) {
+      return 0;
+   }
+   return len;
+}
+
+mp_err mp_prime_rand(mp_int *a, int t, int size, int flags)
+{
+   return s_mp_prime_random_ex(a, t, size, flags, s_mp_rand_cb, NULL);
+}
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_prime_strong_lucas_selfridge.c b/libtommath/bn_mp_prime_strong_lucas_selfridge.c
index 5a94f8e..b50bbcd 100644
--- a/libtommath/bn_mp_prime_strong_lucas_selfridge.c
+++ b/libtommath/bn_mp_prime_strong_lucas_selfridge.c
@@ -1,22 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_PRIME_STRONG_LUCAS_SELFRIDGE_C
 
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /*
  *  See file bn_mp_prime_is_prime.c or the documentation in doc/bn.tex for the details
  */
-#ifndef LTM_USE_FIPS_ONLY
+#ifndef LTM_USE_ONLY_MR
 
 /*
  *  8-bit is just too small. You can try the Frobenius test
@@ -28,33 +19,21 @@
  * multiply bigint a with int d and put the result in c
  * Like mp_mul_d() but with a signed long as the small input
  */
-static int s_mp_mul_si(const mp_int *a, long d, mp_int *c)
+static mp_err s_mp_mul_si(const mp_int *a, int32_t d, mp_int *c)
 {
    mp_int t;
-   int err, neg = 0;
+   mp_err err;
 
    if ((err = mp_init(&t)) != MP_OKAY) {
       return err;
    }
-   if (d < 0) {
-      neg = 1;
-      d = -d;
-   }
 
    /*
     * mp_digit might be smaller than a long, which excludes
     * the use of mp_mul_d() here.
     */
-   if ((err = mp_set_long(&t, (unsigned long) d)) != MP_OKAY) {
-      goto LBL_MPMULSI_ERR;
-   }
-   if ((err = mp_mul(a, &t, c)) != MP_OKAY) {
-      goto LBL_MPMULSI_ERR;
-   }
-   if (neg ==  1) {
-      c->sign = (a->sign == MP_NEG) ? MP_ZPOS: MP_NEG;
-   }
-LBL_MPMULSI_ERR:
+   mp_set_i32(&t, d);
+   err = mp_mul(a, &t, c);
    mp_clear(&t);
    return err;
 }
@@ -75,14 +54,14 @@ LBL_MPMULSI_ERR:
     (If that name sounds familiar, he is the guy who found the fdiv bug in the
      Pentium (P5x, I think) Intel processor)
 */
-int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
+mp_err mp_prime_strong_lucas_selfridge(const mp_int *a, mp_bool *result)
 {
    /* CZ TODO: choose better variable names! */
    mp_int Dz, gcd, Np1, Uz, Vz, U2mz, V2mz, Qmz, Q2mz, Qkdz, T1z, T2z, T3z, T4z, Q2kdz;
    /* CZ TODO: Some of them need the full 32 bit, hence the (temporary) exclusion of MP_8BIT */
    int32_t D, Ds, J, sign, P, Q, r, s, u, Nbits;
-   int e;
-   int isset, oddness;
+   mp_err err;
+   mp_bool oddness;
 
    *result = MP_NO;
    /*
@@ -93,9 +72,9 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
    included.
    */
 
-   if ((e = mp_init_multi(&Dz, &gcd, &Np1, &Uz, &Vz, &U2mz, &V2mz, &Qmz, &Q2mz, &Qkdz, &T1z, &T2z, &T3z, &T4z, &Q2kdz,
-                          NULL)) != MP_OKAY) {
-      return e;
+   if ((err = mp_init_multi(&Dz, &gcd, &Np1, &Uz, &Vz, &U2mz, &V2mz, &Qmz, &Q2mz, &Qkdz, &T1z, &T2z, &T3z, &T4z, &Q2kdz,
+                            NULL)) != MP_OKAY) {
+      return err;
    }
 
    D = 5;
@@ -104,12 +83,9 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
    for (;;) {
       Ds   = sign * D;
       sign = -sign;
-      if ((e = mp_set_long(&Dz, (unsigned long)D)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_gcd(a, &Dz, &gcd)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
+      mp_set_u32(&Dz, (uint32_t)D);
+      if ((err = mp_gcd(a, &Dz, &gcd)) != MP_OKAY)                goto LBL_LS_ERR;
+
       /* if 1 < GCD < N then N is composite with factor "D", and
          Jacobi(D,N) is technically undefined (but often returned
          as zero). */
@@ -119,9 +95,7 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
       if (Ds < 0) {
          Dz.sign = MP_NEG;
       }
-      if ((e = mp_kronecker(&Dz, a, &J)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
+      if ((err = mp_kronecker(&Dz, a, &J)) != MP_OKAY)            goto LBL_LS_ERR;
 
       if (J == -1) {
          break;
@@ -129,7 +103,7 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
       D += 2;
 
       if (D > (INT_MAX - 2)) {
-         e = MP_VAL;
+         err = MP_VAL;
          goto LBL_LS_ERR;
       }
    }
@@ -169,9 +143,7 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
       Baillie-PSW test based on the strong Lucas-Selfridge test
       should be more reliable. */
 
-   if ((e = mp_add_d(a, 1uL, &Np1)) != MP_OKAY) {
-      goto LBL_LS_ERR;
-   }
+   if ((err = mp_add_d(a, 1uL, &Np1)) != MP_OKAY)                 goto LBL_LS_ERR;
    s = mp_cnt_lsb(&Np1);
 
    /* CZ
@@ -181,9 +153,7 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
     * dividing an even number by two does not produce
     * any leftovers.
     */
-   if ((e = mp_div_2d(&Np1, s, &Dz, NULL)) != MP_OKAY) {
-      goto LBL_LS_ERR;
-   }
+   if ((err = mp_div_2d(&Np1, s, &Dz, NULL)) != MP_OKAY)          goto LBL_LS_ERR;
    /* We must now compute U_d and V_d. Since d is odd, the accumulated
       values U and V are initialized to U_1 and V_1 (if the target
       index were even, U and V would be initialized instead to U_0=0
@@ -200,34 +170,10 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
    mp_set(&U2mz, 1uL);  /* U_1 */
    mp_set(&V2mz, (mp_digit)P);  /* V_1 */
 
-   if (Q < 0) {
-      Q = -Q;
-      if ((e = mp_set_long(&Qmz, (unsigned long)Q)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_mul_2(&Qmz, &Q2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      /* Initializes calculation of Q^d */
-      if ((e = mp_set_long(&Qkdz, (unsigned long)Q)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      Qmz.sign = MP_NEG;
-      Q2mz.sign = MP_NEG;
-      Qkdz.sign = MP_NEG;
-      Q = -Q;
-   } else {
-      if ((e = mp_set_long(&Qmz, (unsigned long)Q)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_mul_2(&Qmz, &Q2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      /* Initializes calculation of Q^d */
-      if ((e = mp_set_long(&Qkdz, (unsigned long)Q)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-   }
+   mp_set_i32(&Qmz, Q);
+   if ((err = mp_mul_2(&Qmz, &Q2mz)) != MP_OKAY)                  goto LBL_LS_ERR;
+   /* Initializes calculation of Q^d */
+   mp_set_i32(&Qkdz, Q);
 
    Nbits = mp_count_bits(&Dz);
 
@@ -240,37 +186,20 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
        * V_2m = V_m*V_m - 2*Q^m
        */
 
-      if ((e = mp_mul(&U2mz, &V2mz, &U2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_mod(&U2mz, a, &U2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_sqr(&V2mz, &V2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_sub(&V2mz, &Q2mz, &V2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_mod(&V2mz, a, &V2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
+      if ((err = mp_mul(&U2mz, &V2mz, &U2mz)) != MP_OKAY)         goto LBL_LS_ERR;
+      if ((err = mp_mod(&U2mz, a, &U2mz)) != MP_OKAY)             goto LBL_LS_ERR;
+      if ((err = mp_sqr(&V2mz, &V2mz)) != MP_OKAY)                goto LBL_LS_ERR;
+      if ((err = mp_sub(&V2mz, &Q2mz, &V2mz)) != MP_OKAY)         goto LBL_LS_ERR;
+      if ((err = mp_mod(&V2mz, a, &V2mz)) != MP_OKAY)             goto LBL_LS_ERR;
+
       /* Must calculate powers of Q for use in V_2m, also for Q^d later */
-      if ((e = mp_sqr(&Qmz, &Qmz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
+      if ((err = mp_sqr(&Qmz, &Qmz)) != MP_OKAY)                  goto LBL_LS_ERR;
+
       /* prevents overflow */ /* CZ  still necessary without a fixed prealloc'd mem.? */
-      if ((e = mp_mod(&Qmz, a, &Qmz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_mul_2(&Qmz, &Q2mz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((isset = mp_get_bit(&Dz, u)) == MP_VAL) {
-         e = isset;
-         goto LBL_LS_ERR;
-      }
-      if (isset == MP_YES) {
+      if ((err = mp_mod(&Qmz, a, &Qmz)) != MP_OKAY)               goto LBL_LS_ERR;
+      if ((err = mp_mul_2(&Qmz, &Q2mz)) != MP_OKAY)               goto LBL_LS_ERR;
+
+      if (s_mp_get_bit(&Dz, (unsigned int)u) == MP_YES) {
          /* Formulas for addition of indices (carried out mod N);
           *
           * U_(m+n) = (U_m*V_n + U_n*V_m)/2
@@ -278,79 +207,46 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
           *
           * Be careful with division by 2 (mod N)!
           */
-         if ((e = mp_mul(&U2mz, &Vz, &T1z)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_mul(&Uz, &V2mz, &T2z)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_mul(&V2mz, &Vz, &T3z)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_mul(&U2mz, &Uz, &T4z)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = s_mp_mul_si(&T4z, (long)Ds, &T4z)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_add(&T1z, &T2z, &Uz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if (mp_isodd(&Uz) != MP_NO) {
-            if ((e = mp_add(&Uz, a, &Uz)) != MP_OKAY) {
-               goto LBL_LS_ERR;
-            }
+         if ((err = mp_mul(&U2mz, &Vz, &T1z)) != MP_OKAY)         goto LBL_LS_ERR;
+         if ((err = mp_mul(&Uz, &V2mz, &T2z)) != MP_OKAY)         goto LBL_LS_ERR;
+         if ((err = mp_mul(&V2mz, &Vz, &T3z)) != MP_OKAY)         goto LBL_LS_ERR;
+         if ((err = mp_mul(&U2mz, &Uz, &T4z)) != MP_OKAY)         goto LBL_LS_ERR;
+         if ((err = s_mp_mul_si(&T4z, Ds, &T4z)) != MP_OKAY)      goto LBL_LS_ERR;
+         if ((err = mp_add(&T1z, &T2z, &Uz)) != MP_OKAY)          goto LBL_LS_ERR;
+         if (MP_IS_ODD(&Uz)) {
+            if ((err = mp_add(&Uz, a, &Uz)) != MP_OKAY)           goto LBL_LS_ERR;
          }
          /* CZ
           * This should round towards negative infinity because
           * Thomas R. Nicely used GMP's mpz_fdiv_q_2exp().
           * But mp_div_2() does not do so, it is truncating instead.
           */
-         oddness = mp_isodd(&Uz);
-         if ((e = mp_div_2(&Uz, &Uz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
+         oddness = MP_IS_ODD(&Uz) ? MP_YES : MP_NO;
+         if ((err = mp_div_2(&Uz, &Uz)) != MP_OKAY)               goto LBL_LS_ERR;
          if ((Uz.sign == MP_NEG) && (oddness != MP_NO)) {
-            if ((e = mp_sub_d(&Uz, 1uL, &Uz)) != MP_OKAY) {
-               goto LBL_LS_ERR;
-            }
-         }
-         if ((e = mp_add(&T3z, &T4z, &Vz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
+            if ((err = mp_sub_d(&Uz, 1uL, &Uz)) != MP_OKAY)       goto LBL_LS_ERR;
          }
-         if (mp_isodd(&Vz) != MP_NO) {
-            if ((e = mp_add(&Vz, a, &Vz)) != MP_OKAY) {
-               goto LBL_LS_ERR;
-            }
-         }
-         oddness = mp_isodd(&Vz);
-         if ((e = mp_div_2(&Vz, &Vz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
+         if ((err = mp_add(&T3z, &T4z, &Vz)) != MP_OKAY)          goto LBL_LS_ERR;
+         if (MP_IS_ODD(&Vz)) {
+            if ((err = mp_add(&Vz, a, &Vz)) != MP_OKAY)           goto LBL_LS_ERR;
          }
+         oddness = MP_IS_ODD(&Vz) ? MP_YES : MP_NO;
+         if ((err = mp_div_2(&Vz, &Vz)) != MP_OKAY)               goto LBL_LS_ERR;
          if ((Vz.sign == MP_NEG) && (oddness != MP_NO)) {
-            if ((e = mp_sub_d(&Vz, 1uL, &Vz)) != MP_OKAY) {
-               goto LBL_LS_ERR;
-            }
-         }
-         if ((e = mp_mod(&Uz, a, &Uz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_mod(&Vz, a, &Vz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
+            if ((err = mp_sub_d(&Vz, 1uL, &Vz)) != MP_OKAY)       goto LBL_LS_ERR;
          }
+         if ((err = mp_mod(&Uz, a, &Uz)) != MP_OKAY)              goto LBL_LS_ERR;
+         if ((err = mp_mod(&Vz, a, &Vz)) != MP_OKAY)              goto LBL_LS_ERR;
+
          /* Calculating Q^d for later use */
-         if ((e = mp_mul(&Qkdz, &Qmz, &Qkdz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_mod(&Qkdz, a, &Qkdz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
+         if ((err = mp_mul(&Qkdz, &Qmz, &Qkdz)) != MP_OKAY)       goto LBL_LS_ERR;
+         if ((err = mp_mod(&Qkdz, a, &Qkdz)) != MP_OKAY)          goto LBL_LS_ERR;
       }
    }
 
    /* If U_d or V_d is congruent to 0 mod N, then N is a prime or a
       strong Lucas pseudoprime. */
-   if ((mp_iszero(&Uz) != MP_NO) || (mp_iszero(&Vz) != MP_NO)) {
+   if (MP_IS_ZERO(&Uz) || MP_IS_ZERO(&Vz)) {
       *result = MP_YES;
       goto LBL_LS_ERR;
    }
@@ -367,45 +263,27 @@ int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result)
       Lucas pseudoprime. */
 
    /* Initialize 2*Q^(d*2^r) for V_2m */
-   if ((e = mp_mul_2(&Qkdz, &Q2kdz)) != MP_OKAY) {
-      goto LBL_LS_ERR;
-   }
+   if ((err = mp_mul_2(&Qkdz, &Q2kdz)) != MP_OKAY)                goto LBL_LS_ERR;
 
    for (r = 1; r < s; r++) {
-      if ((e = mp_sqr(&Vz, &Vz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_sub(&Vz, &Q2kdz, &Vz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if ((e = mp_mod(&Vz, a, &Vz)) != MP_OKAY) {
-         goto LBL_LS_ERR;
-      }
-      if (mp_iszero(&Vz) != MP_NO) {
+      if ((err = mp_sqr(&Vz, &Vz)) != MP_OKAY)                    goto LBL_LS_ERR;
+      if ((err = mp_sub(&Vz, &Q2kdz, &Vz)) != MP_OKAY)            goto LBL_LS_ERR;
+      if ((err = mp_mod(&Vz, a, &Vz)) != MP_OKAY)                 goto LBL_LS_ERR;
+      if (MP_IS_ZERO(&Vz)) {
          *result = MP_YES;
          goto LBL_LS_ERR;
       }
       /* Calculate Q^{d*2^r} for next r (final iteration irrelevant). */
       if (r < (s - 1)) {
-         if ((e = mp_sqr(&Qkdz, &Qkdz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_mod(&Qkdz, a, &Qkdz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
-         if ((e = mp_mul_2(&Qkdz, &Q2kdz)) != MP_OKAY) {
-            goto LBL_LS_ERR;
-         }
+         if ((err = mp_sqr(&Qkdz, &Qkdz)) != MP_OKAY)             goto LBL_LS_ERR;
+         if ((err = mp_mod(&Qkdz, a, &Qkdz)) != MP_OKAY)          goto LBL_LS_ERR;
+         if ((err = mp_mul_2(&Qkdz, &Q2kdz)) != MP_OKAY)          goto LBL_LS_ERR;
       }
    }
 LBL_LS_ERR:
    mp_clear_multi(&Q2kdz, &T4z, &T3z, &T2z, &T1z, &Qkdz, &Q2mz, &Qmz, &V2mz, &U2mz, &Vz, &Uz, &Np1, &gcd, &Dz, NULL);
-   return e;
+   return err;
 }
 #endif
 #endif
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_radix_size.c b/libtommath/bn_mp_radix_size.c
index 8583faa..b96f487 100644
--- a/libtommath/bn_mp_radix_size.c
+++ b/libtommath/bn_mp_radix_size.c
@@ -1,22 +1,14 @@
 #include "tommath_private.h"
 #ifdef BN_MP_RADIX_SIZE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
-/* returns size of ASCII reprensentation */
-int mp_radix_size(const mp_int *a, int radix, int *size)
+/* returns size of ASCII representation */
+mp_err mp_radix_size(const mp_int *a, int radix, int *size)
 {
-   int     res, digs;
-   mp_int  t;
+   mp_err  err;
+   int digs;
+   mp_int   t;
    mp_digit d;
 
    *size = 0;
@@ -26,14 +18,14 @@ int mp_radix_size(const mp_int *a, int radix, int *size)
       return MP_VAL;
    }
 
-   if (mp_iszero(a) == MP_YES) {
+   if (MP_IS_ZERO(a)) {
       *size = 2;
       return MP_OKAY;
    }
 
    /* special case for binary */
    if (radix == 2) {
-      *size = mp_count_bits(a) + ((a->sign == MP_NEG) ? 1 : 0) + 1;
+      *size = (mp_count_bits(a) + ((a->sign == MP_NEG) ? 1 : 0) + 1);
       return MP_OKAY;
    }
 
@@ -46,30 +38,28 @@ int mp_radix_size(const mp_int *a, int radix, int *size)
    }
 
    /* init a copy of the input */
-   if ((res = mp_init_copy(&t, a)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_copy(&t, a)) != MP_OKAY) {
+      return err;
    }
 
    /* force temp to positive */
    t.sign = MP_ZPOS;
 
    /* fetch out all of the digits */
-   while (mp_iszero(&t) == MP_NO) {
-      if ((res = mp_div_d(&t, (mp_digit)radix, &t, &d)) != MP_OKAY) {
-         mp_clear(&t);
-         return res;
+   while (!MP_IS_ZERO(&t)) {
+      if ((err = mp_div_d(&t, (mp_digit)radix, &t, &d)) != MP_OKAY) {
+         goto LBL_ERR;
       }
       ++digs;
    }
-   mp_clear(&t);
 
    /* return digs + 1, the 1 is for the NULL byte that would be required. */
    *size = digs + 1;
-   return MP_OKAY;
+   err = MP_OKAY;
+
+LBL_ERR:
+   mp_clear(&t);
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_radix_smap.c b/libtommath/bn_mp_radix_smap.c
index f0b743e..eb4765a 100644
--- a/libtommath/bn_mp_radix_smap.c
+++ b/libtommath/bn_mp_radix_smap.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_RADIX_SMAP_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* chars used in radix conversions */
 const char *const mp_s_rmap = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz+/";
@@ -29,7 +20,3 @@ const unsigned char mp_s_rmap_reverse[] = {
 };
 const size_t mp_s_rmap_reverse_sz = sizeof(mp_s_rmap_reverse);
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_rand.c b/libtommath/bn_mp_rand.c
index 17aa5a2..7e9052c 100644
--- a/libtommath/bn_mp_rand.c
+++ b/libtommath/bn_mp_rand.c
@@ -1,222 +1,46 @@
 #include "tommath_private.h"
 #ifdef BN_MP_RAND_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
-/* First the OS-specific special cases
- * - *BSD
- * - Windows
- */
-#if defined(__FreeBSD__) || defined(__OpenBSD__) || defined(__NetBSD__) || defined(__DragonFly__)
-#define MP_ARC4RANDOM
-#define MP_GEN_RANDOM_MAX     0xffffffffu
-#define MP_GEN_RANDOM_SHIFT   32
+mp_err(*s_mp_rand_source)(void *out, size_t size) = s_mp_rand_platform;
 
-static int s_read_arc4random(mp_digit *p)
+void mp_rand_source(mp_err(*source)(void *out, size_t size))
 {
-   mp_digit d = 0, msk = 0;
-   do {
-      d <<= MP_GEN_RANDOM_SHIFT;
-      d |= ((mp_digit) arc4random());
-      msk <<= MP_GEN_RANDOM_SHIFT;
-      msk |= (MP_MASK & MP_GEN_RANDOM_MAX);
-   } while ((MP_MASK & msk) != MP_MASK);
-   *p = d;
-   return MP_OKAY;
-}
-#endif
-
-#if defined(_WIN32) || defined(_WIN32_WCE)
-#define MP_WIN_CSP
-
-#ifndef _WIN32_WINNT
-#define _WIN32_WINNT 0x0400
-#endif
-#ifdef _WIN32_WCE
-#define UNDER_CE
-#define ARM
-#endif
-
-#define WIN32_LEAN_AND_MEAN
-#include <windows.h>
-#include <wincrypt.h>
-
-static HCRYPTPROV hProv = 0;
-
-static void s_cleanup_win_csp(void)
-{
-   CryptReleaseContext(hProv, 0);
-   hProv = 0;
-}
-
-static int s_read_win_csp(mp_digit *p)
-{
-   int ret = -1;
-   if (hProv == 0) {
-      if (!CryptAcquireContext(&hProv, NULL, MS_DEF_PROV, PROV_RSA_FULL,
-                               (CRYPT_VERIFYCONTEXT | CRYPT_MACHINE_KEYSET)) &&
-          !CryptAcquireContext(&hProv, NULL, MS_DEF_PROV, PROV_RSA_FULL,
-                               CRYPT_VERIFYCONTEXT | CRYPT_MACHINE_KEYSET | CRYPT_NEWKEYSET)) {
-         hProv = 0;
-         return ret;
-      }
-      atexit(s_cleanup_win_csp);
-   }
-   if (CryptGenRandom(hProv, sizeof(*p), (void *)p) == TRUE) {
-      ret = MP_OKAY;
-   }
-   return ret;
+   s_mp_rand_source = (source == NULL) ? s_mp_rand_platform : source;
 }
-#endif /* WIN32 */
 
-#if !defined(MP_WIN_CSP) && defined(__linux__) && defined(__GLIBC_PREREQ)
-#if __GLIBC_PREREQ(2, 25)
-#define MP_GETRANDOM
-#include <sys/random.h>
-#include <errno.h>
-
-static int s_read_getrandom(mp_digit *p)
+mp_err mp_rand(mp_int *a, int digits)
 {
-   int ret;
-   do {
-      ret = getrandom(p, sizeof(*p), 0);
-   } while ((ret == -1) && (errno == EINTR));
-   if (ret == sizeof(*p)) return MP_OKAY;
-   return -1;
-}
-#endif
-#endif
-
-/* We assume all platforms besides windows provide "/dev/urandom".
- * In case yours doesn't, define MP_NO_DEV_URANDOM at compile-time.
- */
-#if !defined(MP_WIN_CSP) && !defined(MP_NO_DEV_URANDOM)
-#ifndef MP_DEV_URANDOM
-#define MP_DEV_URANDOM "/dev/urandom"
-#endif
-#include <fcntl.h>
-#include <errno.h>
-#include <unistd.h>
-
-static int s_read_dev_urandom(mp_digit *p)
-{
-   ssize_t r;
-   int fd;
-   do {
-      fd = open(MP_DEV_URANDOM, O_RDONLY);
-   } while ((fd == -1) && (errno == EINTR));
-   if (fd == -1) return -1;
-   do {
-      r = read(fd, p, sizeof(*p));
-   } while ((r == -1) && (errno == EINTR));
-   close(fd);
-   if (r != sizeof(*p)) return -1;
-   return MP_OKAY;
-}
-#endif
-
-#if defined(MP_PRNG_ENABLE_LTM_RNG)
-unsigned long (*ltm_rng)(unsigned char *out, unsigned long outlen, void (*callback)(void));
-void (*ltm_rng_callback)(void);
-
-static int s_read_ltm_rng(mp_digit *p)
-{
-   unsigned long ret;
-   if (ltm_rng == NULL) return -1;
-   ret = ltm_rng((void *)p, sizeof(*p), ltm_rng_callback);
-   if (ret != sizeof(*p)) return -1;
-   return MP_OKAY;
-}
-#endif
-
-static int s_rand_digit(mp_digit *p)
-{
-   int ret = -1;
-
-#if defined(MP_ARC4RANDOM)
-   ret = s_read_arc4random(p);
-   if (ret == MP_OKAY) return ret;
-#endif
-
-#if defined(MP_WIN_CSP)
-   ret = s_read_win_csp(p);
-   if (ret == MP_OKAY) return ret;
-#else
-
-#if defined(MP_GETRANDOM)
-   ret = s_read_getrandom(p);
-   if (ret == MP_OKAY) return ret;
-#endif
-#if defined(MP_DEV_URANDOM)
-   ret = s_read_dev_urandom(p);
-   if (ret == MP_OKAY) return ret;
-#endif
-
-#endif /* MP_WIN_CSP */
-
-#if defined(MP_PRNG_ENABLE_LTM_RNG)
-   ret = s_read_ltm_rng(p);
-   if (ret == MP_OKAY) return ret;
-#endif
-
-   return ret;
-}
-
-/* makes a pseudo-random int of a given size */
-int mp_rand_digit(mp_digit *r)
-{
-   int ret = s_rand_digit(r);
-   *r &= MP_MASK;
-   return ret;
-}
-
-int mp_rand(mp_int *a, int digits)
-{
-   int     res;
-   mp_digit d;
+   int i;
+   mp_err err;
 
    mp_zero(a);
+
    if (digits <= 0) {
       return MP_OKAY;
    }
 
-   /* first place a random non-zero digit */
-   do {
-      if (mp_rand_digit(&d) != MP_OKAY) {
-         return MP_VAL;
-      }
-   } while (d == 0u);
+   if ((err = mp_grow(a, digits)) != MP_OKAY) {
+      return err;
+   }
 
-   if ((res = mp_add_d(a, d, a)) != MP_OKAY) {
-      return res;
+   if ((err = s_mp_rand_source(a->dp, (size_t)digits * sizeof(mp_digit))) != MP_OKAY) {
+      return err;
    }
 
-   while (--digits > 0) {
-      if ((res = mp_lshd(a, 1)) != MP_OKAY) {
-         return res;
+   /* TODO: We ensure that the highest digit is nonzero. Should this be removed? */
+   while ((a->dp[digits - 1] & MP_MASK) == 0u) {
+      if ((err = s_mp_rand_source(a->dp + digits - 1, sizeof(mp_digit))) != MP_OKAY) {
+         return err;
       }
+   }
 
-      if (mp_rand_digit(&d) != MP_OKAY) {
-         return MP_VAL;
-      }
-      if ((res = mp_add_d(a, d, a)) != MP_OKAY) {
-         return res;
-      }
+   a->used = digits;
+   for (i = 0; i < digits; ++i) {
+      a->dp[i] &= MP_MASK;
    }
 
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_read_radix.c b/libtommath/bn_mp_read_radix.c
index a8723b7..de18e06 100644
--- a/libtommath/bn_mp_read_radix.c
+++ b/libtommath/bn_mp_read_radix.c
@@ -1,25 +1,18 @@
 #include "tommath_private.h"
 #ifdef BN_MP_READ_RADIX_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #define MP_TOUPPER(c) ((((c) >= 'a') && ((c) <= 'z')) ? (((c) + 'A') - 'a') : (c))
 
 /* read a string [ASCII] in a given radix */
-int mp_read_radix(mp_int *a, const char *str, int radix)
+mp_err mp_read_radix(mp_int *a, const char *str, int radix)
 {
-   int     y, res, neg;
+   mp_err   err;
+   int      y;
+   mp_sign  neg;
    unsigned pos;
-   char    ch;
+   char     ch;
 
    /* zero the digit bignum */
    mp_zero(a);
@@ -62,11 +55,11 @@ int mp_read_radix(mp_int *a, const char *str, int radix)
       if ((y == 0xff) || (y >= radix)) {
          break;
       }
-      if ((res = mp_mul_d(a, (mp_digit)radix, a)) != MP_OKAY) {
-         return res;
+      if ((err = mp_mul_d(a, (mp_digit)radix, a)) != MP_OKAY) {
+         return err;
       }
-      if ((res = mp_add_d(a, (mp_digit)y, a)) != MP_OKAY) {
-         return res;
+      if ((err = mp_add_d(a, (mp_digit)y, a)) != MP_OKAY) {
+         return err;
       }
       ++str;
    }
@@ -78,13 +71,9 @@ int mp_read_radix(mp_int *a, const char *str, int radix)
    }
 
    /* set the sign only if a != 0 */
-   if (mp_iszero(a) != MP_YES) {
+   if (!MP_IS_ZERO(a)) {
       a->sign = neg;
    }
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_read_signed_bin.c b/libtommath/bn_mp_read_signed_bin.c
deleted file mode 100644
index e97a1d0..0000000
--- a/libtommath/bn_mp_read_signed_bin.c
+++ /dev/null
@@ -1,38 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_READ_SIGNED_BIN_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* read signed bin, big endian, first byte is 0==positive or 1==negative */
-int mp_read_signed_bin(mp_int *a, const unsigned char *b, int c)
-{
-   int     res;
-
-   /* read magnitude */
-   if ((res = mp_read_unsigned_bin(a, b + 1, c - 1)) != MP_OKAY) {
-      return res;
-   }
-
-   /* first byte is 0 for positive, non-zero for negative */
-   if (b[0] == (unsigned char)0) {
-      a->sign = MP_ZPOS;
-   } else {
-      a->sign = MP_NEG;
-   }
-
-   return MP_OKAY;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_read_unsigned_bin.c b/libtommath/bn_mp_read_unsigned_bin.c
deleted file mode 100644
index 648762a..0000000
--- a/libtommath/bn_mp_read_unsigned_bin.c
+++ /dev/null
@@ -1,52 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_READ_UNSIGNED_BIN_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* reads a unsigned char array, assumes the msb is stored first [big endian] */
-int mp_read_unsigned_bin(mp_int *a, const unsigned char *b, int c)
-{
-   int     res;
-
-   /* make sure there are at least two digits */
-   if (a->alloc < 2) {
-      if ((res = mp_grow(a, 2)) != MP_OKAY) {
-         return res;
-      }
-   }
-
-   /* zero the int */
-   mp_zero(a);
-
-   /* read the bytes in */
-   while (c-- > 0) {
-      if ((res = mp_mul_2d(a, 8, a)) != MP_OKAY) {
-         return res;
-      }
-
-#ifndef MP_8BIT
-      a->dp[0] |= *b++;
-      a->used += 1;
-#else
-      a->dp[0] = (*b & MP_MASK);
-      a->dp[1] |= ((*b++ >> 7) & 1u);
-      a->used += 2;
-#endif
-   }
-   mp_clamp(a);
-   return MP_OKAY;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce.c b/libtommath/bn_mp_reduce.c
index cbf8641..3c669d4 100644
--- a/libtommath/bn_mp_reduce.c
+++ b/libtommath/bn_mp_reduce.c
@@ -1,86 +1,76 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* reduces x mod m, assumes 0 < x < m**2, mu is
  * precomputed via mp_reduce_setup.
  * From HAC pp.604 Algorithm 14.42
  */
-int mp_reduce(mp_int *x, const mp_int *m, const mp_int *mu)
+mp_err mp_reduce(mp_int *x, const mp_int *m, const mp_int *mu)
 {
    mp_int  q;
-   int     res, um = m->used;
+   mp_err  err;
+   int     um = m->used;
 
    /* q = x */
-   if ((res = mp_init_copy(&q, x)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_copy(&q, x)) != MP_OKAY) {
+      return err;
    }
 
    /* q1 = x / b**(k-1)  */
    mp_rshd(&q, um - 1);
 
    /* according to HAC this optimization is ok */
-   if ((mp_digit)um > ((mp_digit)1 << (DIGIT_BIT - 1))) {
-      if ((res = mp_mul(&q, mu, &q)) != MP_OKAY) {
+   if ((mp_digit)um > ((mp_digit)1 << (MP_DIGIT_BIT - 1))) {
+      if ((err = mp_mul(&q, mu, &q)) != MP_OKAY) {
          goto CLEANUP;
       }
-   } else {
-#ifdef BN_S_MP_MUL_HIGH_DIGS_C
-      if ((res = s_mp_mul_high_digs(&q, mu, &q, um)) != MP_OKAY) {
+   } else if (MP_HAS(S_MP_MUL_HIGH_DIGS)) {
+      if ((err = s_mp_mul_high_digs(&q, mu, &q, um)) != MP_OKAY) {
          goto CLEANUP;
       }
-#elif defined(BN_FAST_S_MP_MUL_HIGH_DIGS_C)
-      if ((res = fast_s_mp_mul_high_digs(&q, mu, &q, um)) != MP_OKAY) {
+   } else if (MP_HAS(S_MP_MUL_HIGH_DIGS_FAST)) {
+      if ((err = s_mp_mul_high_digs_fast(&q, mu, &q, um)) != MP_OKAY) {
          goto CLEANUP;
       }
-#else
-      {
-         res = MP_VAL;
-         goto CLEANUP;
-      }
-#endif
+   } else {
+      err = MP_VAL;
+      goto CLEANUP;
    }
 
    /* q3 = q2 / b**(k+1) */
    mp_rshd(&q, um + 1);
 
    /* x = x mod b**(k+1), quick (no division) */
-   if ((res = mp_mod_2d(x, DIGIT_BIT * (um + 1), x)) != MP_OKAY) {
+   if ((err = mp_mod_2d(x, MP_DIGIT_BIT * (um + 1), x)) != MP_OKAY) {
       goto CLEANUP;
    }
 
    /* q = q * m mod b**(k+1), quick (no division) */
-   if ((res = s_mp_mul_digs(&q, m, &q, um + 1)) != MP_OKAY) {
+   if ((err = s_mp_mul_digs(&q, m, &q, um + 1)) != MP_OKAY) {
       goto CLEANUP;
    }
 
    /* x = x - q */
-   if ((res = mp_sub(x, &q, x)) != MP_OKAY) {
+   if ((err = mp_sub(x, &q, x)) != MP_OKAY) {
       goto CLEANUP;
    }
 
    /* If x < 0, add b**(k+1) to it */
    if (mp_cmp_d(x, 0uL) == MP_LT) {
       mp_set(&q, 1uL);
-      if ((res = mp_lshd(&q, um + 1)) != MP_OKAY)
+      if ((err = mp_lshd(&q, um + 1)) != MP_OKAY) {
          goto CLEANUP;
-      if ((res = mp_add(x, &q, x)) != MP_OKAY)
+      }
+      if ((err = mp_add(x, &q, x)) != MP_OKAY) {
          goto CLEANUP;
+      }
    }
 
    /* Back off if it's too big */
    while (mp_cmp(x, m) != MP_LT) {
-      if ((res = s_mp_sub(x, m, x)) != MP_OKAY) {
+      if ((err = s_mp_sub(x, m, x)) != MP_OKAY) {
          goto CLEANUP;
       }
    }
@@ -88,10 +78,6 @@ int mp_reduce(mp_int *x, const mp_int *m, const mp_int *mu)
 CLEANUP:
    mp_clear(&q);
 
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce_2k.c b/libtommath/bn_mp_reduce_2k.c
index af673e6..1cea6cb 100644
--- a/libtommath/bn_mp_reduce_2k.c
+++ b/libtommath/bn_mp_reduce_2k.c
@@ -1,48 +1,40 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_2K_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* reduces a modulo n where n is of the form 2**p - d */
-int mp_reduce_2k(mp_int *a, const mp_int *n, mp_digit d)
+mp_err mp_reduce_2k(mp_int *a, const mp_int *n, mp_digit d)
 {
    mp_int q;
-   int    p, res;
+   mp_err err;
+   int    p;
 
-   if ((res = mp_init(&q)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init(&q)) != MP_OKAY) {
+      return err;
    }
 
    p = mp_count_bits(n);
 top:
    /* q = a/2**p, a = a mod 2**p */
-   if ((res = mp_div_2d(a, p, &q, a)) != MP_OKAY) {
+   if ((err = mp_div_2d(a, p, &q, a)) != MP_OKAY) {
       goto LBL_ERR;
    }
 
    if (d != 1u) {
       /* q = q * d */
-      if ((res = mp_mul_d(&q, d, &q)) != MP_OKAY) {
+      if ((err = mp_mul_d(&q, d, &q)) != MP_OKAY) {
          goto LBL_ERR;
       }
    }
 
    /* a = a + q */
-   if ((res = s_mp_add(a, &q, a)) != MP_OKAY) {
+   if ((err = s_mp_add(a, &q, a)) != MP_OKAY) {
       goto LBL_ERR;
    }
 
    if (mp_cmp_mag(a, n) != MP_LT) {
-      if ((res = s_mp_sub(a, n, a)) != MP_OKAY) {
+      if ((err = s_mp_sub(a, n, a)) != MP_OKAY) {
          goto LBL_ERR;
       }
       goto top;
@@ -50,11 +42,7 @@ top:
 
 LBL_ERR:
    mp_clear(&q);
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce_2k_l.c b/libtommath/bn_mp_reduce_2k_l.c
index afdc321..6a9f3d3 100644
--- a/libtommath/bn_mp_reduce_2k_l.c
+++ b/libtommath/bn_mp_reduce_2k_l.c
@@ -1,49 +1,41 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_2K_L_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* reduces a modulo n where n is of the form 2**p - d
    This differs from reduce_2k since "d" can be larger
    than a single digit.
 */
-int mp_reduce_2k_l(mp_int *a, const mp_int *n, const mp_int *d)
+mp_err mp_reduce_2k_l(mp_int *a, const mp_int *n, const mp_int *d)
 {
    mp_int q;
-   int    p, res;
+   mp_err err;
+   int    p;
 
-   if ((res = mp_init(&q)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init(&q)) != MP_OKAY) {
+      return err;
    }
 
    p = mp_count_bits(n);
 top:
    /* q = a/2**p, a = a mod 2**p */
-   if ((res = mp_div_2d(a, p, &q, a)) != MP_OKAY) {
+   if ((err = mp_div_2d(a, p, &q, a)) != MP_OKAY) {
       goto LBL_ERR;
    }
 
    /* q = q * d */
-   if ((res = mp_mul(&q, d, &q)) != MP_OKAY) {
+   if ((err = mp_mul(&q, d, &q)) != MP_OKAY) {
       goto LBL_ERR;
    }
 
    /* a = a + q */
-   if ((res = s_mp_add(a, &q, a)) != MP_OKAY) {
+   if ((err = s_mp_add(a, &q, a)) != MP_OKAY) {
       goto LBL_ERR;
    }
 
    if (mp_cmp_mag(a, n) != MP_LT) {
-      if ((res = s_mp_sub(a, n, a)) != MP_OKAY) {
+      if ((err = s_mp_sub(a, n, a)) != MP_OKAY) {
          goto LBL_ERR;
       }
       goto top;
@@ -51,11 +43,7 @@ top:
 
 LBL_ERR:
    mp_clear(&q);
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce_2k_setup.c b/libtommath/bn_mp_reduce_2k_setup.c
index 166a965..2eaf7ad 100644
--- a/libtommath/bn_mp_reduce_2k_setup.c
+++ b/libtommath/bn_mp_reduce_2k_setup.c
@@ -1,36 +1,28 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_2K_SETUP_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* determines the setup value */
-int mp_reduce_2k_setup(const mp_int *a, mp_digit *d)
+mp_err mp_reduce_2k_setup(const mp_int *a, mp_digit *d)
 {
-   int res, p;
+   mp_err err;
    mp_int tmp;
+   int    p;
 
-   if ((res = mp_init(&tmp)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init(&tmp)) != MP_OKAY) {
+      return err;
    }
 
    p = mp_count_bits(a);
-   if ((res = mp_2expt(&tmp, p)) != MP_OKAY) {
+   if ((err = mp_2expt(&tmp, p)) != MP_OKAY) {
       mp_clear(&tmp);
-      return res;
+      return err;
    }
 
-   if ((res = s_mp_sub(&tmp, a, &tmp)) != MP_OKAY) {
+   if ((err = s_mp_sub(&tmp, a, &tmp)) != MP_OKAY) {
       mp_clear(&tmp);
-      return res;
+      return err;
    }
 
    *d = tmp.dp[0];
@@ -38,7 +30,3 @@ int mp_reduce_2k_setup(const mp_int *a, mp_digit *d)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce_2k_setup_l.c b/libtommath/bn_mp_reduce_2k_setup_l.c
index 5584b48..4f9aa14 100644
--- a/libtommath/bn_mp_reduce_2k_setup_l.c
+++ b/libtommath/bn_mp_reduce_2k_setup_l.c
@@ -1,41 +1,28 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_2K_SETUP_L_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* determines the setup value */
-int mp_reduce_2k_setup_l(const mp_int *a, mp_int *d)
+mp_err mp_reduce_2k_setup_l(const mp_int *a, mp_int *d)
 {
-   int    res;
+   mp_err err;
    mp_int tmp;
 
-   if ((res = mp_init(&tmp)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init(&tmp)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_2expt(&tmp, mp_count_bits(a))) != MP_OKAY) {
+   if ((err = mp_2expt(&tmp, mp_count_bits(a))) != MP_OKAY) {
       goto LBL_ERR;
    }
 
-   if ((res = s_mp_sub(&tmp, a, d)) != MP_OKAY) {
+   if ((err = s_mp_sub(&tmp, a, d)) != MP_OKAY) {
       goto LBL_ERR;
    }
 
 LBL_ERR:
    mp_clear(&tmp);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce_is_2k.c b/libtommath/bn_mp_reduce_is_2k.c
index 8be985e..a9f4f9f 100644
--- a/libtommath/bn_mp_reduce_is_2k.c
+++ b/libtommath/bn_mp_reduce_is_2k.c
@@ -1,19 +1,10 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_IS_2K_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* determines if mp_reduce_2k can be used */
-int mp_reduce_is_2k(const mp_int *a)
+mp_bool mp_reduce_is_2k(const mp_int *a)
 {
    int ix, iy, iw;
    mp_digit iz;
@@ -28,22 +19,20 @@ int mp_reduce_is_2k(const mp_int *a)
       iw = 1;
 
       /* Test every bit from the second digit up, must be 1 */
-      for (ix = DIGIT_BIT; ix < iy; ix++) {
+      for (ix = MP_DIGIT_BIT; ix < iy; ix++) {
          if ((a->dp[iw] & iz) == 0u) {
             return MP_NO;
          }
          iz <<= 1;
-         if (iz > (mp_digit)MP_MASK) {
+         if (iz > MP_DIGIT_MAX) {
             ++iw;
             iz = 1;
          }
       }
+      return MP_YES;
+   } else {
+      return MP_YES;
    }
-   return MP_YES;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce_is_2k_l.c b/libtommath/bn_mp_reduce_is_2k_l.c
index da4aeda..4bc69be 100644
--- a/libtommath/bn_mp_reduce_is_2k_l.c
+++ b/libtommath/bn_mp_reduce_is_2k_l.c
@@ -1,19 +1,10 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_IS_2K_L_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* determines if reduce_2k_l can be used */
-int mp_reduce_is_2k_l(const mp_int *a)
+mp_bool mp_reduce_is_2k_l(const mp_int *a)
 {
    int ix, iy;
 
@@ -24,18 +15,14 @@ int mp_reduce_is_2k_l(const mp_int *a)
    } else if (a->used > 1) {
       /* if more than half of the digits are -1 we're sold */
       for (iy = ix = 0; ix < a->used; ix++) {
-         if (a->dp[ix] == MP_MASK) {
+         if (a->dp[ix] == MP_DIGIT_MAX) {
             ++iy;
          }
       }
       return (iy >= (a->used/2)) ? MP_YES : MP_NO;
-
+   } else {
+      return MP_NO;
    }
-   return MP_NO;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_reduce_setup.c b/libtommath/bn_mp_reduce_setup.c
index 134d8a3..f02160f 100644
--- a/libtommath/bn_mp_reduce_setup.c
+++ b/libtommath/bn_mp_reduce_setup.c
@@ -1,31 +1,17 @@
 #include "tommath_private.h"
 #ifdef BN_MP_REDUCE_SETUP_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* pre-calculate the value required for Barrett reduction
  * For a given modulus "b" it calulates the value required in "a"
  */
-int mp_reduce_setup(mp_int *a, const mp_int *b)
+mp_err mp_reduce_setup(mp_int *a, const mp_int *b)
 {
-   int     res;
-
-   if ((res = mp_2expt(a, b->used * 2 * DIGIT_BIT)) != MP_OKAY) {
-      return res;
+   mp_err err;
+   if ((err = mp_2expt(a, b->used * 2 * MP_DIGIT_BIT)) != MP_OKAY) {
+      return err;
    }
    return mp_div(a, b, a, NULL);
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_root_u32.c b/libtommath/bn_mp_root_u32.c
new file mode 100644
index 0000000..b60cf26
--- /dev/null
+++ b/libtommath/bn_mp_root_u32.c
@@ -0,0 +1,139 @@
+#include "tommath_private.h"
+#ifdef BN_MP_ROOT_U32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* find the n'th root of an integer
+ *
+ * Result found such that (c)**b <= a and (c+1)**b > a
+ *
+ * This algorithm uses Newton's approximation
+ * x[i+1] = x[i] - f(x[i])/f'(x[i])
+ * which will find the root in log(N) time where
+ * each step involves a fair bit.
+ */
+mp_err mp_root_u32(const mp_int *a, unsigned int b, mp_int *c)
+{
+   mp_int t1, t2, t3, a_;
+   mp_ord cmp;
+   int    ilog2;
+   mp_err err;
+
+   /* input must be positive if b is even */
+   if (((b & 1u) == 0u) && (a->sign == MP_NEG)) {
+      return MP_VAL;
+   }
+
+   if ((err = mp_init_multi(&t1, &t2, &t3, NULL)) != MP_OKAY) {
+      return err;
+   }
+
+   /* if a is negative fudge the sign but keep track */
+   a_ = *a;
+   a_.sign = MP_ZPOS;
+
+   /* Compute seed: 2^(log_2(n)/b + 2)*/
+   ilog2 = mp_count_bits(a);
+
+   /*
+     If "b" is larger than INT_MAX it is also larger than
+     log_2(n) because the bit-length of the "n" is measured
+     with an int and hence the root is always < 2 (two).
+   */
+   if (b > (unsigned int)(INT_MAX/2)) {
+      mp_set(c, 1uL);
+      c->sign = a->sign;
+      err = MP_OKAY;
+      goto LBL_ERR;
+   }
+
+   /* "b" is smaller than INT_MAX, we can cast safely */
+   if (ilog2 < (int)b) {
+      mp_set(c, 1uL);
+      c->sign = a->sign;
+      err = MP_OKAY;
+      goto LBL_ERR;
+   }
+   ilog2 =  ilog2 / ((int)b);
+   if (ilog2 == 0) {
+      mp_set(c, 1uL);
+      c->sign = a->sign;
+      err = MP_OKAY;
+      goto LBL_ERR;
+   }
+   /* Start value must be larger than root */
+   ilog2 += 2;
+   if ((err = mp_2expt(&t2,ilog2)) != MP_OKAY)                    goto LBL_ERR;
+   do {
+      /* t1 = t2 */
+      if ((err = mp_copy(&t2, &t1)) != MP_OKAY)                   goto LBL_ERR;
+
+      /* t2 = t1 - ((t1**b - a) / (b * t1**(b-1))) */
+
+      /* t3 = t1**(b-1) */
+      if ((err = mp_expt_u32(&t1, b - 1u, &t3)) != MP_OKAY)       goto LBL_ERR;
+
+      /* numerator */
+      /* t2 = t1**b */
+      if ((err = mp_mul(&t3, &t1, &t2)) != MP_OKAY)               goto LBL_ERR;
+
+      /* t2 = t1**b - a */
+      if ((err = mp_sub(&t2, &a_, &t2)) != MP_OKAY)               goto LBL_ERR;
+
+      /* denominator */
+      /* t3 = t1**(b-1) * b  */
+      if ((err = mp_mul_d(&t3, b, &t3)) != MP_OKAY)               goto LBL_ERR;
+
+      /* t3 = (t1**b - a)/(b * t1**(b-1)) */
+      if ((err = mp_div(&t2, &t3, &t3, NULL)) != MP_OKAY)         goto LBL_ERR;
+
+      if ((err = mp_sub(&t1, &t3, &t2)) != MP_OKAY)               goto LBL_ERR;
+
+      /*
+          Number of rounds is at most log_2(root). If it is more it
+          got stuck, so break out of the loop and do the rest manually.
+       */
+      if (ilog2-- == 0) {
+         break;
+      }
+   }  while (mp_cmp(&t1, &t2) != MP_EQ);
+
+   /* result can be off by a few so check */
+   /* Loop beneath can overshoot by one if found root is smaller than actual root */
+   for (;;) {
+      if ((err = mp_expt_u32(&t1, b, &t2)) != MP_OKAY)            goto LBL_ERR;
+      cmp = mp_cmp(&t2, &a_);
+      if (cmp == MP_EQ) {
+         err = MP_OKAY;
+         goto LBL_ERR;
+      }
+      if (cmp == MP_LT) {
+         if ((err = mp_add_d(&t1, 1uL, &t1)) != MP_OKAY)          goto LBL_ERR;
+      } else {
+         break;
+      }
+   }
+   /* correct overshoot from above or from recurrence */
+   for (;;) {
+      if ((err = mp_expt_u32(&t1, b, &t2)) != MP_OKAY)            goto LBL_ERR;
+      if (mp_cmp(&t2, &a_) == MP_GT) {
+         if ((err = mp_sub_d(&t1, 1uL, &t1)) != MP_OKAY)          goto LBL_ERR;
+      } else {
+         break;
+      }
+   }
+
+   /* set the result */
+   mp_exch(&t1, c);
+
+   /* set the sign of the result */
+   c->sign = a->sign;
+
+   err = MP_OKAY;
+
+LBL_ERR:
+   mp_clear_multi(&t1, &t2, &t3, NULL);
+   return err;
+}
+
+#endif
diff --git a/libtommath/bn_mp_rshd.c b/libtommath/bn_mp_rshd.c
index 61ab8c0..bb8743e 100644
--- a/libtommath/bn_mp_rshd.c
+++ b/libtommath/bn_mp_rshd.c
@@ -1,21 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_RSHD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* shift right a certain amount of digits */
 void mp_rshd(mp_int *a, int b)
 {
    int     x;
+   mp_digit *bottom, *top;
 
    /* if b <= 0 then ignore it */
    if (b <= 0) {
@@ -28,42 +20,32 @@ void mp_rshd(mp_int *a, int b)
       return;
    }
 
-   {
-      mp_digit *bottom, *top;
+   /* shift the digits down */
 
-      /* shift the digits down */
+   /* bottom */
+   bottom = a->dp;
 
-      /* bottom */
-      bottom = a->dp;
+   /* top [offset into digits] */
+   top = a->dp + b;
 
-      /* top [offset into digits] */
-      top = a->dp + b;
+   /* this is implemented as a sliding window where
+    * the window is b-digits long and digits from
+    * the top of the window are copied to the bottom
+    *
+    * e.g.
 
-      /* this is implemented as a sliding window where
-       * the window is b-digits long and digits from
-       * the top of the window are copied to the bottom
-       *
-       * e.g.
-
-       b-2 | b-1 | b0 | b1 | b2 | ... | bb |   ---->
-                   /\                   |      ---->
-                    \-------------------/      ---->
-       */
-      for (x = 0; x < (a->used - b); x++) {
-         *bottom++ = *top++;
-      }
-
-      /* zero the top digits */
-      for (; x < a->used; x++) {
-         *bottom++ = 0;
-      }
+    b-2 | b-1 | b0 | b1 | b2 | ... | bb |   ---->
+                /\                   |      ---->
+                 \-------------------/      ---->
+    */
+   for (x = 0; x < (a->used - b); x++) {
+      *bottom++ = *top++;
    }
 
+   /* zero the top digits */
+   MP_ZERO_DIGITS(bottom, a->used - x);
+
    /* remove excess digits */
    a->used -= b;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_sbin_size.c b/libtommath/bn_mp_sbin_size.c
new file mode 100644
index 0000000..e0993d6
--- /dev/null
+++ b/libtommath/bn_mp_sbin_size.c
@@ -0,0 +1,11 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SBIN_SIZE_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* get the size for an signed equivalent */
+size_t mp_sbin_size(const mp_int *a)
+{
+   return 1u + mp_ubin_size(a);
+}
+#endif
diff --git a/libtommath/bn_mp_set.c b/libtommath/bn_mp_set.c
index 590a100..44ac6df 100644
--- a/libtommath/bn_mp_set.c
+++ b/libtommath/bn_mp_set.c
@@ -1,26 +1,14 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SET_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* set to a digit */
 void mp_set(mp_int *a, mp_digit b)
 {
-   mp_zero(a);
    a->dp[0] = b & MP_MASK;
+   a->sign  = MP_ZPOS;
    a->used  = (a->dp[0] != 0u) ? 1 : 0;
+   MP_ZERO_DIGITS(a->dp + a->used, a->alloc - a->used);
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_set_double.c b/libtommath/bn_mp_set_double.c
index 12f8dad..a42fc70 100644
--- a/libtommath/bn_mp_set_double.c
+++ b/libtommath/bn_mp_set_double.c
@@ -1,47 +1,36 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SET_DOUBLE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #if defined(__STDC_IEC_559__) || defined(__GCC_IEC_559)
-int mp_set_double(mp_int *a, double b)
+mp_err mp_set_double(mp_int *a, double b)
 {
-   unsigned long long frac;
-   int exp, res;
+   uint64_t frac;
+   int exp;
+   mp_err err;
    union {
       double   dbl;
-      unsigned long long bits;
+      uint64_t bits;
    } cast;
    cast.dbl = b;
 
-   exp = (int)((unsigned)(cast.bits >> 52) & 0x7FFU);
-   frac = (cast.bits & ((1ULL << 52) - 1ULL)) | (1ULL << 52);
+   exp = (int)((unsigned)(cast.bits >> 52) & 0x7FFu);
+   frac = (cast.bits & ((1uLL << 52) - 1uLL)) | (1uLL << 52);
 
    if (exp == 0x7FF) { /* +-inf, NaN */
       return MP_VAL;
    }
    exp -= 1023 + 52;
 
-   res = mp_set_long_long(a, frac);
-   if (res != MP_OKAY) {
-      return res;
-   }
+   mp_set_u64(a, frac);
 
-   res = (exp < 0) ? mp_div_2d(a, -exp, a, NULL) : mp_mul_2d(a, exp, a);
-   if (res != MP_OKAY) {
-      return res;
+   err = (exp < 0) ? mp_div_2d(a, -exp, a, NULL) : mp_mul_2d(a, exp, a);
+   if (err != MP_OKAY) {
+      return err;
    }
 
-   if (((cast.bits >> 63) != 0ULL) && !IS_ZERO(a)) {
+   if (((cast.bits >> 63) != 0uLL) && !MP_IS_ZERO(a)) {
       a->sign = MP_NEG;
    }
 
@@ -56,7 +45,3 @@ int mp_set_double(mp_int *a, double b)
 #  endif
 #endif
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_set_i32.c b/libtommath/bn_mp_set_i32.c
new file mode 100644
index 0000000..df4513d
--- /dev/null
+++ b/libtommath/bn_mp_set_i32.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_I32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_SIGNED(mp_set_i32, mp_set_u32, int32_t, uint32_t)
+#endif
diff --git a/libtommath/bn_mp_set_i64.c b/libtommath/bn_mp_set_i64.c
new file mode 100644
index 0000000..395103b
--- /dev/null
+++ b/libtommath/bn_mp_set_i64.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_I64_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_SIGNED(mp_set_i64, mp_set_u64, int64_t, uint64_t)
+#endif
diff --git a/libtommath/bn_mp_set_int.c b/libtommath/bn_mp_set_int.c
deleted file mode 100644
index 4f01e25..0000000
--- a/libtommath/bn_mp_set_int.c
+++ /dev/null
@@ -1,45 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_SET_INT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* set a 32-bit const */
-int mp_set_int(mp_int *a, unsigned long b)
-{
-   int     x, res;
-
-   mp_zero(a);
-
-   /* set four bits at a time */
-   for (x = 0; x < 8; x++) {
-      /* shift the number up four bits */
-      if ((res = mp_mul_2d(a, 4, a)) != MP_OKAY) {
-         return res;
-      }
-
-      /* OR in the top four bits of the source */
-      a->dp[0] |= (mp_digit)(b >> 28) & 15uL;
-
-      /* shift the source up to the next four bits */
-      b <<= 4;
-
-      /* ensure that digits are not clamped off */
-      a->used += 1;
-   }
-   mp_clamp(a);
-   return MP_OKAY;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_set_l.c b/libtommath/bn_mp_set_l.c
new file mode 100644
index 0000000..1e445fb
--- /dev/null
+++ b/libtommath/bn_mp_set_l.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_L_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_SIGNED(mp_set_l, mp_set_ul, long, unsigned long)
+#endif
diff --git a/libtommath/bn_mp_set_ll.c b/libtommath/bn_mp_set_ll.c
new file mode 100644
index 0000000..3e2324f
--- /dev/null
+++ b/libtommath/bn_mp_set_ll.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_LL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_SIGNED(mp_set_ll, mp_set_ull, long long, unsigned long long)
+#endif
diff --git a/libtommath/bn_mp_set_long.c b/libtommath/bn_mp_set_long.c
deleted file mode 100644
index 35be8e7..0000000
--- a/libtommath/bn_mp_set_long.c
+++ /dev/null
@@ -1,21 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_SET_LONG_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* set a platform dependent unsigned long int */
-MP_SET_XLONG(mp_set_long, unsigned long)
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_set_long_long.c b/libtommath/bn_mp_set_long_long.c
deleted file mode 100644
index 633c4df..0000000
--- a/libtommath/bn_mp_set_long_long.c
+++ /dev/null
@@ -1,21 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_SET_LONG_LONG_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* set a platform dependent unsigned long long int */
-MP_SET_XLONG(mp_set_long_long, Tcl_WideUInt)
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_set_u32.c b/libtommath/bn_mp_set_u32.c
new file mode 100644
index 0000000..18ba5e1
--- /dev/null
+++ b/libtommath/bn_mp_set_u32.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_U32_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_UNSIGNED(mp_set_u32, uint32_t)
+#endif
diff --git a/libtommath/bn_mp_set_u64.c b/libtommath/bn_mp_set_u64.c
new file mode 100644
index 0000000..88fab6c
--- /dev/null
+++ b/libtommath/bn_mp_set_u64.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_U64_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_UNSIGNED(mp_set_u64, uint64_t)
+#endif
diff --git a/libtommath/bn_mp_set_ul.c b/libtommath/bn_mp_set_ul.c
new file mode 100644
index 0000000..adfd85c
--- /dev/null
+++ b/libtommath/bn_mp_set_ul.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_UL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_UNSIGNED(mp_set_ul, unsigned long)
+#endif
diff --git a/libtommath/bn_mp_set_ull.c b/libtommath/bn_mp_set_ull.c
new file mode 100644
index 0000000..8fbc1bd
--- /dev/null
+++ b/libtommath/bn_mp_set_ull.c
@@ -0,0 +1,7 @@
+#include "tommath_private.h"
+#ifdef BN_MP_SET_ULL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+MP_SET_UNSIGNED(mp_set_ull, unsigned long long)
+#endif
diff --git a/libtommath/bn_mp_shrink.c b/libtommath/bn_mp_shrink.c
index fa30184..cf27ed9 100644
--- a/libtommath/bn_mp_shrink.c
+++ b/libtommath/bn_mp_shrink.c
@@ -1,40 +1,22 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SHRINK_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* shrink a bignum */
-int mp_shrink(mp_int *a)
+mp_err mp_shrink(mp_int *a)
 {
    mp_digit *tmp;
-   int used = 1;
-
-   if (a->used > 0) {
-      used = a->used;
-   }
-
-   if (a->alloc != used) {
-      if ((tmp = (mp_digit *) XREALLOC(a->dp,
-                                       (size_t)a->alloc * sizeof (mp_digit),
-                                       (size_t)used * sizeof(mp_digit))) == NULL) {
+   int alloc = MP_MAX(MP_MIN_PREC, a->used);
+   if (a->alloc != alloc) {
+      if ((tmp = (mp_digit *) MP_REALLOC(a->dp,
+                                         (size_t)a->alloc * sizeof(mp_digit),
+                                         (size_t)alloc * sizeof(mp_digit))) == NULL) {
          return MP_MEM;
       }
       a->dp    = tmp;
-      a->alloc = used;
+      a->alloc = alloc;
    }
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_signed_bin_size.c b/libtommath/bn_mp_signed_bin_size.c
deleted file mode 100644
index 89cd43e..0000000
--- a/libtommath/bn_mp_signed_bin_size.c
+++ /dev/null
@@ -1,24 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_SIGNED_BIN_SIZE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* get the size for an signed equivalent */
-int mp_signed_bin_size(const mp_int *a)
-{
-   return 1 + mp_unsigned_bin_size(a);
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_sqr.c b/libtommath/bn_mp_sqr.c
index 63bb2e2..571b25d 100644
--- a/libtommath/bn_mp_sqr.c
+++ b/libtommath/bn_mp_sqr.c
@@ -1,56 +1,28 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SQR_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* computes b = a*a */
-int mp_sqr(const mp_int *a, mp_int *b)
+mp_err mp_sqr(const mp_int *a, mp_int *b)
 {
-   int     res;
-
-#ifdef BN_MP_TOOM_SQR_C
-   /* use Toom-Cook? */
-   if (a->used >= TOOM_SQR_CUTOFF) {
-      res = mp_toom_sqr(a, b);
-      /* Karatsuba? */
-   } else
-#endif
-#ifdef BN_MP_KARATSUBA_SQR_C
-      if (a->used >= KARATSUBA_SQR_CUTOFF) {
-         res = mp_karatsuba_sqr(a, b);
-      } else
-#endif
-      {
-#ifdef BN_FAST_S_MP_SQR_C
-         /* can we use the fast comba multiplier? */
-         if ((((a->used * 2) + 1) < (int)MP_WARRAY) &&
-             (a->used <
-              (int)(1u << (((sizeof(mp_word) * (size_t)CHAR_BIT) - (2u * (size_t)DIGIT_BIT)) - 1u)))) {
-            res = fast_s_mp_sqr(a, b);
-         } else
-#endif
-         {
-#ifdef BN_S_MP_SQR_C
-            res = s_mp_sqr(a, b);
-#else
-            res = MP_VAL;
-#endif
-         }
-      }
+   mp_err err;
+   if (MP_HAS(S_MP_TOOM_SQR) && /* use Toom-Cook? */
+       (a->used >= MP_TOOM_SQR_CUTOFF)) {
+      err = s_mp_toom_sqr(a, b);
+   } else if (MP_HAS(S_MP_KARATSUBA_SQR) &&  /* Karatsuba? */
+              (a->used >= MP_KARATSUBA_SQR_CUTOFF)) {
+      err = s_mp_karatsuba_sqr(a, b);
+   } else if (MP_HAS(S_MP_SQR_FAST) && /* can we use the fast comba multiplier? */
+              (((a->used * 2) + 1) < PRIVATE_MP_WARRAY) &&
+              (a->used < (MP_MAXFAST / 2))) {
+      err = s_mp_sqr_fast(a, b);
+   } else if (MP_HAS(S_MP_SQR)) {
+      err = s_mp_sqr(a, b);
+   } else {
+      err = MP_VAL;
+   }
    b->sign = MP_ZPOS;
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_sqrmod.c b/libtommath/bn_mp_sqrmod.c
index 953829e..626ea2c 100644
--- a/libtommath/bn_mp_sqrmod.c
+++ b/libtommath/bn_mp_sqrmod.c
@@ -1,37 +1,25 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SQRMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* c = a * a (mod b) */
-int mp_sqrmod(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_sqrmod(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int     res;
+   mp_err  err;
    mp_int  t;
 
-   if ((res = mp_init(&t)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init(&t)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_sqr(a, &t)) != MP_OKAY) {
-      mp_clear(&t);
-      return res;
+   if ((err = mp_sqr(a, &t)) != MP_OKAY) {
+      goto LBL_ERR;
    }
-   res = mp_mod(&t, b, c);
+   err = mp_mod(&t, b, c);
+
+LBL_ERR:
    mp_clear(&t);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_sqrt.c b/libtommath/bn_mp_sqrt.c
index 116fb14..aaeac1d 100644
--- a/libtommath/bn_mp_sqrt.c
+++ b/libtommath/bn_mp_sqrt.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SQRT_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #ifndef NO_FLOATING_POINT
 #include <math.h>
@@ -20,9 +11,9 @@
 #endif
 
 /* this function is less generic than mp_n_root, simpler and faster */
-int mp_sqrt(const mp_int *arg, mp_int *ret)
+mp_err mp_sqrt(const mp_int *arg, mp_int *ret)
 {
-   int res;
+   mp_err err;
    mp_int t1, t2;
 #ifndef NO_FLOATING_POINT
    int i, j, k;
@@ -36,7 +27,7 @@ int mp_sqrt(const mp_int *arg, mp_int *ret)
    }
 
    /* easy out */
-   if (mp_iszero(arg) == MP_YES) {
+   if (MP_IS_ZERO(arg)) {
       mp_zero(ret);
       return MP_OKAY;
    }
@@ -45,11 +36,11 @@ int mp_sqrt(const mp_int *arg, mp_int *ret)
 
    i = (arg->used / 2) - 1;
    j = 2 * i;
-   if ((res = mp_init_size(&t1, i+2)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&t1, i+2)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_init(&t2)) != MP_OKAY) {
+   if ((err = mp_init(&t2)) != MP_OKAY) {
       goto E2;
    }
 
@@ -99,11 +90,11 @@ int mp_sqrt(const mp_int *arg, mp_int *ret)
 
 #else
 
-   if ((res = mp_init_copy(&t1, arg)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_copy(&t1, arg)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_init(&t2)) != MP_OKAY) {
+   if ((err = mp_init(&t2)) != MP_OKAY) {
       goto E2;
    }
 
@@ -113,24 +104,24 @@ int mp_sqrt(const mp_int *arg, mp_int *ret)
 #endif
 
    /* t1 > 0  */
-   if ((res = mp_div(arg, &t1, &t2, NULL)) != MP_OKAY) {
+   if ((err = mp_div(arg, &t1, &t2, NULL)) != MP_OKAY) {
       goto E1;
    }
-   if ((res = mp_add(&t1, &t2, &t1)) != MP_OKAY) {
+   if ((err = mp_add(&t1, &t2, &t1)) != MP_OKAY) {
       goto E1;
    }
-   if ((res = mp_div_2(&t1, &t1)) != MP_OKAY) {
+   if ((err = mp_div_2(&t1, &t1)) != MP_OKAY) {
       goto E1;
    }
    /* And now t1 > sqrt(arg) */
    do {
-      if ((res = mp_div(arg, &t1, &t2, NULL)) != MP_OKAY) {
+      if ((err = mp_div(arg, &t1, &t2, NULL)) != MP_OKAY) {
          goto E1;
       }
-      if ((res = mp_add(&t1, &t2, &t1)) != MP_OKAY) {
+      if ((err = mp_add(&t1, &t2, &t1)) != MP_OKAY) {
          goto E1;
       }
-      if ((res = mp_div_2(&t1, &t1)) != MP_OKAY) {
+      if ((err = mp_div_2(&t1, &t1)) != MP_OKAY) {
          goto E1;
       }
       /* t1 >= sqrt(arg) >= t2 at this point */
@@ -142,11 +133,7 @@ E1:
    mp_clear(&t2);
 E2:
    mp_clear(&t1);
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_sqrtmod_prime.c b/libtommath/bn_mp_sqrtmod_prime.c
index cc4da3b..a833ed7 100644
--- a/libtommath/bn_mp_sqrtmod_prime.c
+++ b/libtommath/bn_mp_sqrtmod_prime.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SQRTMOD_PRIME_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* Tonelli-Shanks algorithm
  * https://en.wikipedia.org/wiki/Tonelli%E2%80%93Shanks_algorithm
@@ -18,9 +9,10 @@
  *
  */
 
-int mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret)
+mp_err mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret)
 {
-   int res, legendre;
+   mp_err err;
+   int legendre;
    mp_int t1, C, Q, S, Z, M, T, R, two;
    mp_digit i;
 
@@ -30,90 +22,89 @@ int mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret)
       return MP_OKAY;
    }
    if (mp_cmp_d(prime, 2uL) == MP_EQ)                            return MP_VAL; /* prime must be odd */
-   if ((res = mp_jacobi(n, prime, &legendre)) != MP_OKAY)        return res;
+   if ((err = mp_kronecker(n, prime, &legendre)) != MP_OKAY)        return err;
    if (legendre == -1)                                           return MP_VAL; /* quadratic non-residue mod prime */
 
-   if ((res = mp_init_multi(&t1, &C, &Q, &S, &Z, &M, &T, &R, &two, NULL)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_multi(&t1, &C, &Q, &S, &Z, &M, &T, &R, &two, NULL)) != MP_OKAY) {
+      return err;
    }
 
    /* SPECIAL CASE: if prime mod 4 == 3
-    * compute directly: res = n^(prime+1)/4 mod prime
+    * compute directly: err = n^(prime+1)/4 mod prime
     * Handbook of Applied Cryptography algorithm 3.36
     */
-   if ((res = mp_mod_d(prime, 4uL, &i)) != MP_OKAY)               goto cleanup;
+   if ((err = mp_mod_d(prime, 4uL, &i)) != MP_OKAY)               goto cleanup;
    if (i == 3u) {
-      if ((res = mp_add_d(prime, 1uL, &t1)) != MP_OKAY)           goto cleanup;
-      if ((res = mp_div_2(&t1, &t1)) != MP_OKAY)                  goto cleanup;
-      if ((res = mp_div_2(&t1, &t1)) != MP_OKAY)                  goto cleanup;
-      if ((res = mp_exptmod(n, &t1, prime, ret)) != MP_OKAY)      goto cleanup;
-      res = MP_OKAY;
+      if ((err = mp_add_d(prime, 1uL, &t1)) != MP_OKAY)           goto cleanup;
+      if ((err = mp_div_2(&t1, &t1)) != MP_OKAY)                  goto cleanup;
+      if ((err = mp_div_2(&t1, &t1)) != MP_OKAY)                  goto cleanup;
+      if ((err = mp_exptmod(n, &t1, prime, ret)) != MP_OKAY)      goto cleanup;
+      err = MP_OKAY;
       goto cleanup;
    }
 
    /* NOW: Tonelli-Shanks algorithm */
 
    /* factor out powers of 2 from prime-1, defining Q and S as: prime-1 = Q*2^S */
-   if ((res = mp_copy(prime, &Q)) != MP_OKAY)                    goto cleanup;
-   if ((res = mp_sub_d(&Q, 1uL, &Q)) != MP_OKAY)                 goto cleanup;
+   if ((err = mp_copy(prime, &Q)) != MP_OKAY)                    goto cleanup;
+   if ((err = mp_sub_d(&Q, 1uL, &Q)) != MP_OKAY)                 goto cleanup;
    /* Q = prime - 1 */
    mp_zero(&S);
    /* S = 0 */
-   while (mp_iseven(&Q) != MP_NO) {
-      if ((res = mp_div_2(&Q, &Q)) != MP_OKAY)                    goto cleanup;
+   while (MP_IS_EVEN(&Q)) {
+      if ((err = mp_div_2(&Q, &Q)) != MP_OKAY)                    goto cleanup;
       /* Q = Q / 2 */
-      if ((res = mp_add_d(&S, 1uL, &S)) != MP_OKAY)               goto cleanup;
+      if ((err = mp_add_d(&S, 1uL, &S)) != MP_OKAY)               goto cleanup;
       /* S = S + 1 */
    }
 
    /* find a Z such that the Legendre symbol (Z|prime) == -1 */
-   if ((res = mp_set_int(&Z, 2uL)) != MP_OKAY)                    goto cleanup;
+   mp_set_u32(&Z, 2u);
    /* Z = 2 */
-   while (1) {
-      if ((res = mp_jacobi(&Z, prime, &legendre)) != MP_OKAY)     goto cleanup;
+   for (;;) {
+      if ((err = mp_kronecker(&Z, prime, &legendre)) != MP_OKAY)     goto cleanup;
       if (legendre == -1) break;
-      if ((res = mp_add_d(&Z, 1uL, &Z)) != MP_OKAY)               goto cleanup;
+      if ((err = mp_add_d(&Z, 1uL, &Z)) != MP_OKAY)               goto cleanup;
       /* Z = Z + 1 */
    }
 
-   if ((res = mp_exptmod(&Z, &Q, prime, &C)) != MP_OKAY)         goto cleanup;
+   if ((err = mp_exptmod(&Z, &Q, prime, &C)) != MP_OKAY)         goto cleanup;
    /* C = Z ^ Q mod prime */
-   if ((res = mp_add_d(&Q, 1uL, &t1)) != MP_OKAY)                goto cleanup;
-   if ((res = mp_div_2(&t1, &t1)) != MP_OKAY)                    goto cleanup;
+   if ((err = mp_add_d(&Q, 1uL, &t1)) != MP_OKAY)                goto cleanup;
+   if ((err = mp_div_2(&t1, &t1)) != MP_OKAY)                    goto cleanup;
    /* t1 = (Q + 1) / 2 */
-   if ((res = mp_exptmod(n, &t1, prime, &R)) != MP_OKAY)         goto cleanup;
+   if ((err = mp_exptmod(n, &t1, prime, &R)) != MP_OKAY)         goto cleanup;
    /* R = n ^ ((Q + 1) / 2) mod prime */
-   if ((res = mp_exptmod(n, &Q, prime, &T)) != MP_OKAY)          goto cleanup;
+   if ((err = mp_exptmod(n, &Q, prime, &T)) != MP_OKAY)          goto cleanup;
    /* T = n ^ Q mod prime */
-   if ((res = mp_copy(&S, &M)) != MP_OKAY)                       goto cleanup;
+   if ((err = mp_copy(&S, &M)) != MP_OKAY)                       goto cleanup;
    /* M = S */
-   if ((res = mp_set_int(&two, 2uL)) != MP_OKAY)                 goto cleanup;
+   mp_set_u32(&two, 2u);
 
-   res = MP_VAL;
-   while (1) {
-      if ((res = mp_copy(&T, &t1)) != MP_OKAY)                    goto cleanup;
+   for (;;) {
+      if ((err = mp_copy(&T, &t1)) != MP_OKAY)                    goto cleanup;
       i = 0;
-      while (1) {
+      for (;;) {
          if (mp_cmp_d(&t1, 1uL) == MP_EQ) break;
-         if ((res = mp_exptmod(&t1, &two, prime, &t1)) != MP_OKAY) goto cleanup;
+         if ((err = mp_exptmod(&t1, &two, prime, &t1)) != MP_OKAY) goto cleanup;
          i++;
       }
       if (i == 0u) {
-         if ((res = mp_copy(&R, ret)) != MP_OKAY)                  goto cleanup;
-         res = MP_OKAY;
+         if ((err = mp_copy(&R, ret)) != MP_OKAY)                  goto cleanup;
+         err = MP_OKAY;
          goto cleanup;
       }
-      if ((res = mp_sub_d(&M, i, &t1)) != MP_OKAY)                goto cleanup;
-      if ((res = mp_sub_d(&t1, 1uL, &t1)) != MP_OKAY)             goto cleanup;
-      if ((res = mp_exptmod(&two, &t1, prime, &t1)) != MP_OKAY)   goto cleanup;
+      if ((err = mp_sub_d(&M, i, &t1)) != MP_OKAY)                goto cleanup;
+      if ((err = mp_sub_d(&t1, 1uL, &t1)) != MP_OKAY)             goto cleanup;
+      if ((err = mp_exptmod(&two, &t1, prime, &t1)) != MP_OKAY)   goto cleanup;
       /* t1 = 2 ^ (M - i - 1) */
-      if ((res = mp_exptmod(&C, &t1, prime, &t1)) != MP_OKAY)     goto cleanup;
+      if ((err = mp_exptmod(&C, &t1, prime, &t1)) != MP_OKAY)     goto cleanup;
       /* t1 = C ^ (2 ^ (M - i - 1)) mod prime */
-      if ((res = mp_sqrmod(&t1, prime, &C)) != MP_OKAY)           goto cleanup;
+      if ((err = mp_sqrmod(&t1, prime, &C)) != MP_OKAY)           goto cleanup;
       /* C = (t1 * t1) mod prime */
-      if ((res = mp_mulmod(&R, &t1, prime, &R)) != MP_OKAY)       goto cleanup;
+      if ((err = mp_mulmod(&R, &t1, prime, &R)) != MP_OKAY)       goto cleanup;
       /* R = (R * t1) mod prime */
-      if ((res = mp_mulmod(&T, &C, prime, &T)) != MP_OKAY)        goto cleanup;
+      if ((err = mp_mulmod(&T, &C, prime, &T)) != MP_OKAY)        goto cleanup;
       /* T = (T * C) mod prime */
       mp_set(&M, i);
       /* M = i */
@@ -121,11 +112,7 @@ int mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret)
 
 cleanup:
    mp_clear_multi(&t1, &C, &Q, &S, &Z, &M, &T, &R, &two, NULL);
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_sub.c b/libtommath/bn_mp_sub.c
index df31951..c1ea39e 100644
--- a/libtommath/bn_mp_sub.c
+++ b/libtommath/bn_mp_sub.c
@@ -1,24 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SUB_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* high level subtraction (handles signs) */
-int mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int     sa, sb, res;
-
-   sa = a->sign;
-   sb = b->sign;
+   mp_sign sa = a->sign, sb = b->sign;
+   mp_err err;
 
    if (sa != sb) {
       /* subtract a negative from a positive, OR */
@@ -26,7 +15,7 @@ int mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
       /* In either case, ADD their magnitudes, */
       /* and use the sign of the first number. */
       c->sign = sa;
-      res = s_mp_add(a, b, c);
+      err = s_mp_add(a, b, c);
    } else {
       /* subtract a positive from a positive, OR */
       /* subtract a negative from a negative. */
@@ -36,20 +25,16 @@ int mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
          /* Copy the sign from the first */
          c->sign = sa;
          /* The first has a larger or equal magnitude */
-         res = s_mp_sub(a, b, c);
+         err = s_mp_sub(a, b, c);
       } else {
          /* The result has the *opposite* sign from */
          /* the first number. */
          c->sign = (sa == MP_ZPOS) ? MP_NEG : MP_ZPOS;
          /* The second has a larger magnitude */
-         res = s_mp_sub(b, a, c);
+         err = s_mp_sub(b, a, c);
       }
    }
-   return res;
+   return err;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_sub_d.c b/libtommath/bn_mp_sub_d.c
index d8ac250..3ebf9b4 100644
--- a/libtommath/bn_mp_sub_d.c
+++ b/libtommath/bn_mp_sub_d.c
@@ -1,27 +1,19 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SUB_D_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* single digit subtraction */
-int mp_sub_d(const mp_int *a, mp_digit b, mp_int *c)
+mp_err mp_sub_d(const mp_int *a, mp_digit b, mp_int *c)
 {
-   mp_digit *tmpa, *tmpc, mu;
-   int       res, ix, oldused;
+   mp_digit *tmpa, *tmpc;
+   mp_err    err;
+   int       ix, oldused;
 
    /* grow c as required */
    if (c->alloc < (a->used + 1)) {
-      if ((res = mp_grow(c, a->used + 1)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(c, a->used + 1)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -31,13 +23,13 @@ int mp_sub_d(const mp_int *a, mp_digit b, mp_int *c)
    if (a->sign == MP_NEG) {
       mp_int a_ = *a;
       a_.sign = MP_ZPOS;
-      res     = mp_add_d(&a_, b, c);
+      err     = mp_add_d(&a_, b, c);
       c->sign = MP_NEG;
 
       /* clamp */
       mp_clamp(c);
 
-      return res;
+      return err;
    }
 
    /* setup regs */
@@ -58,33 +50,25 @@ int mp_sub_d(const mp_int *a, mp_digit b, mp_int *c)
       c->sign = MP_NEG;
       c->used = 1;
    } else {
+      mp_digit mu = b;
+
       /* positive/size */
       c->sign = MP_ZPOS;
       c->used = a->used;
 
-      /* subtract first digit */
-      *tmpc    = *tmpa++ - b;
-      mu       = *tmpc >> ((sizeof(mp_digit) * (size_t)CHAR_BIT) - 1u);
-      *tmpc++ &= MP_MASK;
-
-      /* handle rest of the digits */
-      for (ix = 1; ix < a->used; ix++) {
+      /* subtract digits, mu is carry */
+      for (ix = 0; ix < a->used; ix++) {
          *tmpc    = *tmpa++ - mu;
-         mu       = *tmpc >> ((sizeof(mp_digit) * (size_t)CHAR_BIT) - 1u);
+         mu       = *tmpc >> (MP_SIZEOF_BITS(mp_digit) - 1u);
          *tmpc++ &= MP_MASK;
       }
    }
 
    /* zero excess digits */
-   while (ix++ < oldused) {
-      *tmpc++ = 0;
-   }
+   MP_ZERO_DIGITS(tmpc, oldused - ix);
+
    mp_clamp(c);
    return MP_OKAY;
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_submod.c b/libtommath/bn_mp_submod.c
index ba9ee6f..5ebd374 100644
--- a/libtommath/bn_mp_submod.c
+++ b/libtommath/bn_mp_submod.c
@@ -1,38 +1,25 @@
 #include "tommath_private.h"
 #ifdef BN_MP_SUBMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* d = a - b (mod c) */
-int mp_submod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d)
+mp_err mp_submod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d)
 {
-   int     res;
-   mp_int  t;
+   mp_err err;
+   mp_int t;
 
-
-   if ((res = mp_init(&t)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init(&t)) != MP_OKAY) {
+      return err;
    }
 
-   if ((res = mp_sub(a, b, &t)) != MP_OKAY) {
-      mp_clear(&t);
-      return res;
+   if ((err = mp_sub(a, b, &t)) != MP_OKAY) {
+      goto LBL_ERR;
    }
-   res = mp_mod(&t, c, d);
+   err = mp_mod(&t, c, d);
+
+LBL_ERR:
    mp_clear(&t);
-   return res;
+   return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_to_radix.c b/libtommath/bn_mp_to_radix.c
new file mode 100644
index 0000000..7fa86ca
--- /dev/null
+++ b/libtommath/bn_mp_to_radix.c
@@ -0,0 +1,84 @@
+#include "tommath_private.h"
+#ifdef BN_MP_TO_RADIX_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* stores a bignum as a ASCII string in a given radix (2..64)
+ *
+ * Stores upto "size - 1" chars and always a NULL byte, puts the number of characters
+ * written, including the '\0', in "written".
+ */
+mp_err mp_to_radix(const mp_int *a, char *str, size_t maxlen, size_t *written, int radix)
+{
+   size_t  digs;
+   mp_err  err;
+   mp_int  t;
+   mp_digit d;
+   char   *_s = str;
+
+   /* check range of radix and size*/
+   if (maxlen < 2u) {
+      return MP_BUF;
+   }
+   if ((radix < 2) || (radix > 64)) {
+      return MP_VAL;
+   }
+
+   /* quick out if its zero */
+   if (MP_IS_ZERO(a)) {
+      *str++ = '0';
+      *str = '\0';
+      if (written != NULL) {
+         *written = 2u;
+      }
+      return MP_OKAY;
+   }
+
+   if ((err = mp_init_copy(&t, a)) != MP_OKAY) {
+      return err;
+   }
+
+   /* if it is negative output a - */
+   if (t.sign == MP_NEG) {
+      /* we have to reverse our digits later... but not the - sign!! */
+      ++_s;
+
+      /* store the flag and mark the number as positive */
+      *str++ = '-';
+      t.sign = MP_ZPOS;
+
+      /* subtract a char */
+      --maxlen;
+   }
+   digs = 0u;
+   while (!MP_IS_ZERO(&t)) {
+      if (--maxlen < 1u) {
+         /* no more room */
+         err = MP_BUF;
+         goto LBL_ERR;
+      }
+      if ((err = mp_div_d(&t, (mp_digit)radix, &t, &d)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      *str++ = mp_s_rmap[d];
+      ++digs;
+   }
+   /* reverse the digits of the string.  In this case _s points
+    * to the first digit [exluding the sign] of the number
+    */
+   s_mp_reverse((unsigned char *)_s, digs);
+
+   /* append a NULL so the string is properly terminated */
+   *str = '\0';
+   digs++;
+
+   if (written != NULL) {
+      *written = (a->sign == MP_NEG) ? (digs + 1u): digs;
+   }
+
+LBL_ERR:
+   mp_clear(&t);
+   return err;
+}
+
+#endif
diff --git a/libtommath/bn_mp_to_sbin.c b/libtommath/bn_mp_to_sbin.c
new file mode 100644
index 0000000..dbaf53e
--- /dev/null
+++ b/libtommath/bn_mp_to_sbin.c
@@ -0,0 +1,22 @@
+#include "tommath_private.h"
+#ifdef BN_MP_TO_SBIN_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* store in signed [big endian] format */
+mp_err mp_to_sbin(const mp_int *a, unsigned char *buf, size_t maxlen, size_t *written)
+{
+   mp_err err;
+   if (maxlen == 0u) {
+      return MP_BUF;
+   }
+   if ((err = mp_to_ubin(a, buf + 1, maxlen - 1u, written)) != MP_OKAY) {
+      return err;
+   }
+   if (written != NULL) {
+      (*written)++;
+   }
+   buf[0] = (a->sign == MP_ZPOS) ? (unsigned char)0 : (unsigned char)1;
+   return MP_OKAY;
+}
+#endif
diff --git a/libtommath/bn_mp_to_signed_bin.c b/libtommath/bn_mp_to_signed_bin.c
deleted file mode 100644
index 04e3b84..0000000
--- a/libtommath/bn_mp_to_signed_bin.c
+++ /dev/null
@@ -1,30 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TO_SIGNED_BIN_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* store in signed [big endian] format */
-int mp_to_signed_bin(const mp_int *a, unsigned char *b)
-{
-   int     res;
-
-   if ((res = mp_to_unsigned_bin(a, b + 1)) != MP_OKAY) {
-      return res;
-   }
-   b[0] = (a->sign == MP_ZPOS) ? (unsigned char)0 : (unsigned char)1;
-   return MP_OKAY;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_to_signed_bin_n.c b/libtommath/bn_mp_to_signed_bin_n.c
deleted file mode 100644
index d13fede..0000000
--- a/libtommath/bn_mp_to_signed_bin_n.c
+++ /dev/null
@@ -1,28 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TO_SIGNED_BIN_N_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* store in signed [big endian] format */
-int mp_to_signed_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen)
-{
-   if (*outlen < (unsigned long)mp_signed_bin_size(a)) {
-      return MP_VAL;
-   }
-   *outlen = (unsigned long)mp_signed_bin_size(a);
-   return mp_to_signed_bin(a, b);
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_to_ubin.c b/libtommath/bn_mp_to_ubin.c
new file mode 100644
index 0000000..4913c3a
--- /dev/null
+++ b/libtommath/bn_mp_to_ubin.c
@@ -0,0 +1,42 @@
+#include "tommath_private.h"
+#ifdef BN_MP_TO_UBIN_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* store in unsigned [big endian] format */
+mp_err mp_to_ubin(const mp_int *a, unsigned char *buf, size_t maxlen, size_t *written)
+{
+   size_t  x, count;
+   mp_err  err;
+   mp_int  t;
+
+   size_t size = (size_t)mp_count_bits(a);
+   count = (size / 8u) + (((size & 7u) != 0u) ? 1u : 0u);
+   if (count > maxlen) {
+      return MP_BUF;
+   }
+
+   if ((err = mp_init_copy(&t, a)) != MP_OKAY) {
+      return err;
+   }
+
+   for (x = count; x --> 0u;) {
+#ifndef MP_8BIT
+      buf[x] = (unsigned char)(t.dp[0] & 255u);
+#else
+      buf[x] = (unsigned char)(t.dp[0] | ((t.dp[1] & 1u) << 7));
+#endif
+      if ((err = mp_div_2d(&t, 8, &t, NULL)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+   }
+
+   if (written != NULL) {
+      *written = count;
+   }
+
+LBL_ERR:
+   mp_clear(&t);
+   return err;
+}
+#endif
diff --git a/libtommath/bn_mp_to_unsigned_bin.c b/libtommath/bn_mp_to_unsigned_bin.c
deleted file mode 100644
index ab57514..0000000
--- a/libtommath/bn_mp_to_unsigned_bin.c
+++ /dev/null
@@ -1,45 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TO_UNSIGNED_BIN_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* store in unsigned [big endian] format */
-int mp_to_unsigned_bin(const mp_int *a, unsigned char *b)
-{
-   int     x, res;
-   mp_int  t;
-
-   if ((res = mp_init_copy(&t, a)) != MP_OKAY) {
-      return res;
-   }
-
-   x = 0;
-   while (mp_iszero(&t) == MP_NO) {
-#ifndef MP_8BIT
-      b[x++] = (unsigned char)(t.dp[0] & 255u);
-#else
-      b[x++] = (unsigned char)(t.dp[0] | ((t.dp[1] & 1u) << 7));
-#endif
-      if ((res = mp_div_2d(&t, 8, &t, NULL)) != MP_OKAY) {
-         mp_clear(&t);
-         return res;
-      }
-   }
-   bn_reverse(b, x);
-   mp_clear(&t);
-   return MP_OKAY;
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_to_unsigned_bin_n.c b/libtommath/bn_mp_to_unsigned_bin_n.c
deleted file mode 100644
index c53e7fb..0000000
--- a/libtommath/bn_mp_to_unsigned_bin_n.c
+++ /dev/null
@@ -1,28 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TO_UNSIGNED_BIN_N_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* store in unsigned [big endian] format */
-int mp_to_unsigned_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen)
-{
-   if (*outlen < (unsigned long)mp_unsigned_bin_size(a)) {
-      return MP_VAL;
-   }
-   *outlen = (unsigned long)mp_unsigned_bin_size(a);
-   return mp_to_unsigned_bin(a, b);
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_toom_mul.c b/libtommath/bn_mp_toom_mul.c
deleted file mode 100644
index 32b5e43..0000000
--- a/libtommath/bn_mp_toom_mul.c
+++ /dev/null
@@ -1,283 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TOOM_MUL_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* multiplication using the Toom-Cook 3-way algorithm
- *
- * Much more complicated than Karatsuba but has a lower
- * asymptotic running time of O(N**1.464).  This algorithm is
- * only particularly useful on VERY large inputs
- * (we're talking 1000s of digits here...).
-*/
-int mp_toom_mul(const mp_int *a, const mp_int *b, mp_int *c)
-{
-   mp_int w0, w1, w2, w3, w4, tmp1, tmp2, a0, a1, a2, b0, b1, b2;
-   int res, B;
-
-   /* init temps */
-   if ((res = mp_init_multi(&w0, &w1, &w2, &w3, &w4,
-                            &a0, &a1, &a2, &b0, &b1,
-                            &b2, &tmp1, &tmp2, NULL)) != MP_OKAY) {
-      return res;
-   }
-
-   /* B */
-   B = MIN(a->used, b->used) / 3;
-
-   /* a = a2 * B**2 + a1 * B + a0 */
-   if ((res = mp_mod_2d(a, DIGIT_BIT * B, &a0)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_copy(a, &a1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_rshd(&a1, B);
-   if ((res = mp_mod_2d(&a1, DIGIT_BIT * B, &a1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_copy(a, &a2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_rshd(&a2, B*2);
-
-   /* b = b2 * B**2 + b1 * B + b0 */
-   if ((res = mp_mod_2d(b, DIGIT_BIT * B, &b0)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_copy(b, &b1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_rshd(&b1, B);
-   (void)mp_mod_2d(&b1, DIGIT_BIT * B, &b1);
-
-   if ((res = mp_copy(b, &b2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_rshd(&b2, B*2);
-
-   /* w0 = a0*b0 */
-   if ((res = mp_mul(&a0, &b0, &w0)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* w4 = a2 * b2 */
-   if ((res = mp_mul(&a2, &b2, &w4)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* w1 = (a2 + 2(a1 + 2a0))(b2 + 2(b1 + 2b0)) */
-   if ((res = mp_mul_2(&a0, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a2, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_mul_2(&b0, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp2, &b1, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_mul_2(&tmp2, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp2, &b2, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_mul(&tmp1, &tmp2, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* w3 = (a0 + 2(a1 + 2a2))(b0 + 2(b1 + 2b2)) */
-   if ((res = mp_mul_2(&a2, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_mul_2(&b2, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp2, &b1, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_mul_2(&tmp2, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp2, &b0, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_mul(&tmp1, &tmp2, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-
-   /* w2 = (a2 + a1 + a0)(b2 + b1 + b0) */
-   if ((res = mp_add(&a2, &a1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&b2, &b1, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp2, &b0, &tmp2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_mul(&tmp1, &tmp2, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* now solve the matrix
-
-      0  0  0  0  1
-      1  2  4  8  16
-      1  1  1  1  1
-      16 8  4  2  1
-      1  0  0  0  0
-
-      using 12 subtractions, 4 shifts,
-             2 small divisions and 1 small multiplication
-    */
-
-   /* r1 - r4 */
-   if ((res = mp_sub(&w1, &w4, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - r0 */
-   if ((res = mp_sub(&w3, &w0, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1/2 */
-   if ((res = mp_div_2(&w1, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3/2 */
-   if ((res = mp_div_2(&w3, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r2 - r0 - r4 */
-   if ((res = mp_sub(&w2, &w0, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w2, &w4, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1 - r2 */
-   if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - r2 */
-   if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1 - 8r0 */
-   if ((res = mp_mul_2d(&w0, 3, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w1, &tmp1, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - 8r4 */
-   if ((res = mp_mul_2d(&w4, 3, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w3, &tmp1, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* 3r2 - r1 - r3 */
-   if ((res = mp_mul_d(&w2, 3uL, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w2, &w1, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w2, &w3, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1 - r2 */
-   if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - r2 */
-   if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1/3 */
-   if ((res = mp_div_3(&w1, &w1, NULL)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3/3 */
-   if ((res = mp_div_3(&w3, &w3, NULL)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* at this point shift W[n] by B*n */
-   if ((res = mp_lshd(&w1, 1*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_lshd(&w2, 2*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_lshd(&w3, 3*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_lshd(&w4, 4*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_add(&w0, &w1, c)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&w2, &w3, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&w4, &tmp1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, c, c)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-LBL_ERR:
-   mp_clear_multi(&w0, &w1, &w2, &w3, &w4,
-                  &a0, &a1, &a2, &b0, &b1,
-                  &b2, &tmp1, &tmp2, NULL);
-   return res;
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_toom_sqr.c b/libtommath/bn_mp_toom_sqr.c
deleted file mode 100644
index 8595db5..0000000
--- a/libtommath/bn_mp_toom_sqr.c
+++ /dev/null
@@ -1,224 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TOOM_SQR_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* squaring using Toom-Cook 3-way algorithm */
-int mp_toom_sqr(const mp_int *a, mp_int *b)
-{
-   mp_int w0, w1, w2, w3, w4, tmp1, a0, a1, a2;
-   int res, B;
-
-   /* init temps */
-   if ((res = mp_init_multi(&w0, &w1, &w2, &w3, &w4, &a0, &a1, &a2, &tmp1, NULL)) != MP_OKAY) {
-      return res;
-   }
-
-   /* B */
-   B = a->used / 3;
-
-   /* a = a2 * B**2 + a1 * B + a0 */
-   if ((res = mp_mod_2d(a, DIGIT_BIT * B, &a0)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_copy(a, &a1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_rshd(&a1, B);
-   if ((res = mp_mod_2d(&a1, DIGIT_BIT * B, &a1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_copy(a, &a2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   mp_rshd(&a2, B*2);
-
-   /* w0 = a0*a0 */
-   if ((res = mp_sqr(&a0, &w0)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* w4 = a2 * a2 */
-   if ((res = mp_sqr(&a2, &w4)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* w1 = (a2 + 2(a1 + 2a0))**2 */
-   if ((res = mp_mul_2(&a0, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a2, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_sqr(&tmp1, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* w3 = (a0 + 2(a1 + 2a2))**2 */
-   if ((res = mp_mul_2(&a2, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_mul_2(&tmp1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_sqr(&tmp1, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-
-   /* w2 = (a2 + a1 + a0)**2 */
-   if ((res = mp_add(&a2, &a1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, &a0, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sqr(&tmp1, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* now solve the matrix
-
-      0  0  0  0  1
-      1  2  4  8  16
-      1  1  1  1  1
-      16 8  4  2  1
-      1  0  0  0  0
-
-      using 12 subtractions, 4 shifts, 2 small divisions and 1 small multiplication.
-    */
-
-   /* r1 - r4 */
-   if ((res = mp_sub(&w1, &w4, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - r0 */
-   if ((res = mp_sub(&w3, &w0, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1/2 */
-   if ((res = mp_div_2(&w1, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3/2 */
-   if ((res = mp_div_2(&w3, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r2 - r0 - r4 */
-   if ((res = mp_sub(&w2, &w0, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w2, &w4, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1 - r2 */
-   if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - r2 */
-   if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1 - 8r0 */
-   if ((res = mp_mul_2d(&w0, 3, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w1, &tmp1, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - 8r4 */
-   if ((res = mp_mul_2d(&w4, 3, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w3, &tmp1, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* 3r2 - r1 - r3 */
-   if ((res = mp_mul_d(&w2, 3uL, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w2, &w1, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_sub(&w2, &w3, &w2)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1 - r2 */
-   if ((res = mp_sub(&w1, &w2, &w1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3 - r2 */
-   if ((res = mp_sub(&w3, &w2, &w3)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r1/3 */
-   if ((res = mp_div_3(&w1, &w1, NULL)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   /* r3/3 */
-   if ((res = mp_div_3(&w3, &w3, NULL)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   /* at this point shift W[n] by B*n */
-   if ((res = mp_lshd(&w1, 1*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_lshd(&w2, 2*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_lshd(&w3, 3*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_lshd(&w4, 4*B)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-   if ((res = mp_add(&w0, &w1, b)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&w2, &w3, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&w4, &tmp1, &tmp1)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-   if ((res = mp_add(&tmp1, b, b)) != MP_OKAY) {
-      goto LBL_ERR;
-   }
-
-LBL_ERR:
-   mp_clear_multi(&w0, &w1, &w2, &w3, &w4, &a0, &a1, &a2, &tmp1, NULL);
-   return res;
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_toradix.c b/libtommath/bn_mp_toradix.c
deleted file mode 100644
index c6e1c65..0000000
--- a/libtommath/bn_mp_toradix.c
+++ /dev/null
@@ -1,72 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TORADIX_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* stores a bignum as a ASCII string in a given radix (2..64) */
-int mp_toradix(const mp_int *a, char *str, int radix)
-{
-   int     res, digs;
-   mp_int  t;
-   mp_digit d;
-   char   *_s = str;
-
-   /* check range of the radix */
-   if ((radix < 2) || (radix > 64)) {
-      return MP_VAL;
-   }
-
-   /* quick out if its zero */
-   if (mp_iszero(a) == MP_YES) {
-      *str++ = '0';
-      *str = '\0';
-      return MP_OKAY;
-   }
-
-   if ((res = mp_init_copy(&t, a)) != MP_OKAY) {
-      return res;
-   }
-
-   /* if it is negative output a - */
-   if (t.sign == MP_NEG) {
-      ++_s;
-      *str++ = '-';
-      t.sign = MP_ZPOS;
-   }
-
-   digs = 0;
-   while (mp_iszero(&t) == MP_NO) {
-      if ((res = mp_div_d(&t, (mp_digit)radix, &t, &d)) != MP_OKAY) {
-         mp_clear(&t);
-         return res;
-      }
-      *str++ = mp_s_rmap[d];
-      ++digs;
-   }
-
-   /* reverse the digits of the string.  In this case _s points
-    * to the first digit [exluding the sign] of the number]
-    */
-   bn_reverse((unsigned char *)_s, digs);
-
-   /* append a NULL so the string is properly terminated */
-   *str = '\0';
-
-   mp_clear(&t);
-   return MP_OKAY;
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_toradix_n.c b/libtommath/bn_mp_toradix_n.c
deleted file mode 100644
index 84431f2..0000000
--- a/libtommath/bn_mp_toradix_n.c
+++ /dev/null
@@ -1,85 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_TORADIX_N_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* stores a bignum as a ASCII string in a given radix (2..64)
- *
- * Stores upto maxlen-1 chars and always a NULL byte
- */
-int mp_toradix_n(const mp_int *a, char *str, int radix, int maxlen)
-{
-   int     res, digs;
-   mp_int  t;
-   mp_digit d;
-   char   *_s = str;
-
-   /* check range of the maxlen, radix */
-   if ((maxlen < 2) || (radix < 2) || (radix > 64)) {
-      return MP_VAL;
-   }
-
-   /* quick out if its zero */
-   if (mp_iszero(a) == MP_YES) {
-      *str++ = '0';
-      *str = '\0';
-      return MP_OKAY;
-   }
-
-   if ((res = mp_init_copy(&t, a)) != MP_OKAY) {
-      return res;
-   }
-
-   /* if it is negative output a - */
-   if (t.sign == MP_NEG) {
-      /* we have to reverse our digits later... but not the - sign!! */
-      ++_s;
-
-      /* store the flag and mark the number as positive */
-      *str++ = '-';
-      t.sign = MP_ZPOS;
-
-      /* subtract a char */
-      --maxlen;
-   }
-
-   digs = 0;
-   while (mp_iszero(&t) == MP_NO) {
-      if (--maxlen < 1) {
-         /* no more room */
-         break;
-      }
-      if ((res = mp_div_d(&t, (mp_digit)radix, &t, &d)) != MP_OKAY) {
-         mp_clear(&t);
-         return res;
-      }
-      *str++ = mp_s_rmap[d];
-      ++digs;
-   }
-
-   /* reverse the digits of the string.  In this case _s points
-    * to the first digit [exluding the sign] of the number
-    */
-   bn_reverse((unsigned char *)_s, digs);
-
-   /* append a NULL so the string is properly terminated */
-   *str = '\0';
-
-   mp_clear(&t);
-   return MP_OKAY;
-}
-
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_ubin_size.c b/libtommath/bn_mp_ubin_size.c
new file mode 100644
index 0000000..21230b4
--- /dev/null
+++ b/libtommath/bn_mp_ubin_size.c
@@ -0,0 +1,12 @@
+#include "tommath_private.h"
+#ifdef BN_MP_UBIN_SIZE_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* get the size for an unsigned equivalent */
+size_t mp_ubin_size(const mp_int *a)
+{
+   size_t size = (size_t)mp_count_bits(a);
+   return (size / 8u) + (((size & 7u) != 0u) ? 1u : 0u);
+}
+#endif
diff --git a/libtommath/bn_mp_unpack.c b/libtommath/bn_mp_unpack.c
new file mode 100644
index 0000000..d4eb90e
--- /dev/null
+++ b/libtommath/bn_mp_unpack.c
@@ -0,0 +1,49 @@
+#include "tommath_private.h"
+#ifdef BN_MP_UNPACK_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* based on gmp's mpz_import.
+ * see http://gmplib.org/manual/Integer-Import-and-Export.html
+ */
+mp_err mp_unpack(mp_int *rop, size_t count, mp_order order, size_t size,
+                 mp_endian endian, size_t nails, const void *op)
+{
+   mp_err err;
+   size_t odd_nails, nail_bytes, i, j;
+   unsigned char odd_nail_mask;
+
+   mp_zero(rop);
+
+   if (endian == MP_NATIVE_ENDIAN) {
+      MP_GET_ENDIANNESS(endian);
+   }
+
+   odd_nails = (nails % 8u);
+   odd_nail_mask = 0xff;
+   for (i = 0; i < odd_nails; ++i) {
+      odd_nail_mask ^= (unsigned char)(1u << (7u - i));
+   }
+   nail_bytes = nails / 8u;
+
+   for (i = 0; i < count; ++i) {
+      for (j = 0; j < (size - nail_bytes); ++j) {
+         unsigned char byte = *((const unsigned char *)op +
+                                (((order == MP_MSB_FIRST) ? i : ((count - 1u) - i)) * size) +
+                                ((endian == MP_BIG_ENDIAN) ? (j + nail_bytes) : (((size - 1u) - j) - nail_bytes)));
+
+         if ((err = mp_mul_2d(rop, (j == 0u) ? (int)(8u - odd_nails) : 8, rop)) != MP_OKAY) {
+            return err;
+         }
+
+         rop->dp[0] |= (j == 0u) ? (mp_digit)(byte & odd_nail_mask) : (mp_digit)byte;
+         rop->used  += 1;
+      }
+   }
+
+   mp_clamp(rop);
+
+   return MP_OKAY;
+}
+
+#endif
diff --git a/libtommath/bn_mp_unsigned_bin_size.c b/libtommath/bn_mp_unsigned_bin_size.c
deleted file mode 100644
index d716c8f..0000000
--- a/libtommath/bn_mp_unsigned_bin_size.c
+++ /dev/null
@@ -1,25 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_MP_UNSIGNED_BIN_SIZE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* get the size for an unsigned equivalent */
-int mp_unsigned_bin_size(const mp_int *a)
-{
-   int     size = mp_count_bits(a);
-   return (size / 8) + ((((unsigned)size & 7u) != 0u) ? 1 : 0);
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_xor.c b/libtommath/bn_mp_xor.c
index 073985e..71e7ca1 100644
--- a/libtommath/bn_mp_xor.c
+++ b/libtommath/bn_mp_xor.c
@@ -6,7 +6,7 @@
 /* two complement xor */
 mp_err mp_xor(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int used = MAX(a->used, b->used) + 1, i;
+   int used = MP_MAX(a->used, b->used) + 1, i;
    mp_err err;
    mp_digit ac = 1, bc = 1, cc = 1;
    mp_sign csign = (a->sign != b->sign) ? MP_NEG : MP_ZPOS;
diff --git a/libtommath/bn_mp_zero.c b/libtommath/bn_mp_zero.c
index 89f7c29..72a255e 100644
--- a/libtommath/bn_mp_zero.c
+++ b/libtommath/bn_mp_zero.c
@@ -1,33 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_MP_ZERO_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* set to zero */
 void mp_zero(mp_int *a)
 {
-   int       n;
-   mp_digit *tmp;
-
    a->sign = MP_ZPOS;
    a->used = 0;
-
-   tmp = a->dp;
-   for (n = 0; n < a->alloc; n++) {
-      *tmp++ = 0;
-   }
+   MP_ZERO_DIGITS(a->dp, a->alloc);
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_prime_tab.c b/libtommath/bn_prime_tab.c
index 5c0e192..a6c07f8 100644
--- a/libtommath/bn_prime_tab.c
+++ b/libtommath/bn_prime_tab.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
 #ifdef BN_PRIME_TAB_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 const mp_digit ltm_prime_tab[] = {
    0x0002, 0x0003, 0x0005, 0x0007, 0x000B, 0x000D, 0x0011, 0x0013,
@@ -52,8 +43,19 @@ const mp_digit ltm_prime_tab[] = {
    0x062B, 0x062F, 0x063D, 0x0641, 0x0647, 0x0649, 0x064D, 0x0653
 #endif
 };
+
+#if defined(__GNUC__) && __GNUC__ >= 4
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wdeprecated-declarations"
+const mp_digit *s_mp_prime_tab = ltm_prime_tab;
+#pragma GCC diagnostic pop
+#elif defined(_MSC_VER) && _MSC_VER >= 1500
+#pragma warning(push)
+#pragma warning(disable: 4996)
+const mp_digit *s_mp_prime_tab = ltm_prime_tab;
+#pragma warning(pop)
+#else
+const mp_digit *s_mp_prime_tab = ltm_prime_tab;
 #endif
 
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
+#endif
diff --git a/libtommath/bn_reverse.c b/libtommath/bn_reverse.c
deleted file mode 100644
index 2990528..0000000
--- a/libtommath/bn_reverse.c
+++ /dev/null
@@ -1,35 +0,0 @@
-#include "tommath_private.h"
-#ifdef BN_REVERSE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* reverse an array, used for radix code */
-void bn_reverse(unsigned char *s, int len)
-{
-   int     ix, iy;
-   unsigned char t;
-
-   ix = 0;
-   iy = len - 1;
-   while (ix < iy) {
-      t     = s[ix];
-      s[ix] = s[iy];
-      s[iy] = t;
-      ++ix;
-      --iy;
-   }
-}
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_add.c b/libtommath/bn_s_mp_add.c
index 979e470..c946aa8 100644
--- a/libtommath/bn_s_mp_add.c
+++ b/libtommath/bn_s_mp_add.c
@@ -1,22 +1,14 @@
 #include "tommath_private.h"
 #ifdef BN_S_MP_ADD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* low level addition, based on HAC pp.594, Algorithm 14.7 */
-int s_mp_add(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err s_mp_add(const mp_int *a, const mp_int *b, mp_int *c)
 {
    const mp_int *x;
-   int     olduse, res, min, max;
+   mp_err err;
+   int     olduse, min, max;
 
    /* find sizes, we let |a| <= |b| which means we have to sort
     * them.  "x" will point to the input with the most digits
@@ -33,8 +25,8 @@ int s_mp_add(const mp_int *a, const mp_int *b, mp_int *c)
 
    /* init result */
    if (c->alloc < (max + 1)) {
-      if ((res = mp_grow(c, max + 1)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(c, max + 1)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -64,7 +56,7 @@ int s_mp_add(const mp_int *a, const mp_int *b, mp_int *c)
          *tmpc = *tmpa++ + *tmpb++ + u;
 
          /* U = carry bit of T[i] */
-         u = *tmpc >> (mp_digit)DIGIT_BIT;
+         u = *tmpc >> (mp_digit)MP_DIGIT_BIT;
 
          /* take away carry bit from T[i] */
          *tmpc++ &= MP_MASK;
@@ -79,7 +71,7 @@ int s_mp_add(const mp_int *a, const mp_int *b, mp_int *c)
             *tmpc = x->dp[i] + u;
 
             /* U = carry bit of T[i] */
-            u = *tmpc >> (mp_digit)DIGIT_BIT;
+            u = *tmpc >> (mp_digit)MP_DIGIT_BIT;
 
             /* take away carry bit from T[i] */
             *tmpc++ &= MP_MASK;
@@ -90,16 +82,10 @@ int s_mp_add(const mp_int *a, const mp_int *b, mp_int *c)
       *tmpc++ = u;
 
       /* clear digits above oldused */
-      for (i = c->used; i < olduse; i++) {
-         *tmpc++ = 0;
-      }
+      MP_ZERO_DIGITS(tmpc, olduse - c->used);
    }
 
    mp_clamp(c);
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_balance_mul.c b/libtommath/bn_s_mp_balance_mul.c
new file mode 100644
index 0000000..7ece5d7
--- /dev/null
+++ b/libtommath/bn_s_mp_balance_mul.c
@@ -0,0 +1,81 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_BALANCE_MUL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* single-digit multiplication with the smaller number as the single-digit */
+mp_err s_mp_balance_mul(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   int count, len_a, len_b, nblocks, i, j, bsize;
+   mp_int a0, tmp, A, B, r;
+   mp_err err;
+
+   len_a = a->used;
+   len_b = b->used;
+
+   nblocks = MP_MAX(a->used, b->used) / MP_MIN(a->used, b->used);
+   bsize = MP_MIN(a->used, b->used) ;
+
+   if ((err = mp_init_size(&a0, bsize + 2)) != MP_OKAY) {
+      return err;
+   }
+   if ((err = mp_init_multi(&tmp, &r, NULL)) != MP_OKAY) {
+      mp_clear(&a0);
+      return err;
+   }
+
+   /* Make sure that A is the larger one*/
+   if (len_a < len_b) {
+      B = *a;
+      A = *b;
+   } else {
+      A = *a;
+      B = *b;
+   }
+
+   for (i = 0, j=0; i < nblocks; i++) {
+      /* Cut a slice off of a */
+      a0.used = 0;
+      for (count = 0; count < bsize; count++) {
+         a0.dp[count] = A.dp[ j++ ];
+         a0.used++;
+      }
+      mp_clamp(&a0);
+      /* Multiply with b */
+      if ((err = mp_mul(&a0, &B, &tmp)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      /* Shift tmp to the correct position */
+      if ((err = mp_lshd(&tmp, bsize * i)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      /* Add to output. No carry needed */
+      if ((err = mp_add(&r, &tmp, &r)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+   }
+   /* The left-overs; there are always left-overs */
+   if (j < A.used) {
+      a0.used = 0;
+      for (count = 0; j < A.used; count++) {
+         a0.dp[count] = A.dp[ j++ ];
+         a0.used++;
+      }
+      mp_clamp(&a0);
+      if ((err = mp_mul(&a0, &B, &tmp)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      if ((err = mp_lshd(&tmp, bsize * i)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+      if ((err = mp_add(&r, &tmp, &r)) != MP_OKAY) {
+         goto LBL_ERR;
+      }
+   }
+
+   mp_exch(&r,c);
+LBL_ERR:
+   mp_clear_multi(&a0, &tmp, &r,NULL);
+   return err;
+}
+#endif
diff --git a/libtommath/bn_s_mp_exptmod.c b/libtommath/bn_s_mp_exptmod.c
index b22cde8..c3bfa95 100644
--- a/libtommath/bn_s_mp_exptmod.c
+++ b/libtommath/bn_s_mp_exptmod.c
@@ -1,29 +1,23 @@
 #include "tommath_private.h"
 #ifdef BN_S_MP_EXPTMOD_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #ifdef MP_LOW_MEM
 #   define TAB_SIZE 32
+#   define MAX_WINSIZE 5
 #else
 #   define TAB_SIZE 256
+#   define MAX_WINSIZE 0
 #endif
 
-int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode)
+mp_err s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode)
 {
    mp_int  M[TAB_SIZE], res, mu;
    mp_digit buf;
-   int     err, bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
-   int (*redux)(mp_int *x, const mp_int *m, const mp_int *mu);
+   mp_err   err;
+   int      bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
+   mp_err(*redux)(mp_int *x, const mp_int *m, const mp_int *mu);
 
    /* find window size */
    x = mp_count_bits(X);
@@ -43,11 +37,7 @@ int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, i
       winsize = 8;
    }
 
-#ifdef MP_LOW_MEM
-   if (winsize > 5) {
-      winsize = 5;
-   }
-#endif
+   winsize = MAX_WINSIZE ? MP_MIN(MAX_WINSIZE, winsize) : winsize;
 
    /* init M array */
    /* init first cell */
@@ -67,19 +57,13 @@ int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, i
    }
 
    /* create mu, used for Barrett reduction */
-   if ((err = mp_init(&mu)) != MP_OKAY) {
-      goto LBL_M;
-   }
+   if ((err = mp_init(&mu)) != MP_OKAY)                           goto LBL_M;
 
    if (redmode == 0) {
-      if ((err = mp_reduce_setup(&mu, P)) != MP_OKAY) {
-         goto LBL_MU;
-      }
+      if ((err = mp_reduce_setup(&mu, P)) != MP_OKAY)             goto LBL_MU;
       redux = mp_reduce;
    } else {
-      if ((err = mp_reduce_2k_setup_l(P, &mu)) != MP_OKAY) {
-         goto LBL_MU;
-      }
+      if ((err = mp_reduce_2k_setup_l(P, &mu)) != MP_OKAY)        goto LBL_MU;
       redux = mp_reduce_2k_l;
    }
 
@@ -91,46 +75,32 @@ int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, i
     * The first half of the table is not
     * computed though accept for M[0] and M[1]
     */
-   if ((err = mp_mod(G, P, &M[1])) != MP_OKAY) {
-      goto LBL_MU;
-   }
+   if ((err = mp_mod(G, P, &M[1])) != MP_OKAY)                    goto LBL_MU;
 
    /* compute the value at M[1<<(winsize-1)] by squaring
     * M[1] (winsize-1) times
     */
-   if ((err = mp_copy(&M[1], &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) {
-      goto LBL_MU;
-   }
+   if ((err = mp_copy(&M[1], &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) goto LBL_MU;
 
    for (x = 0; x < (winsize - 1); x++) {
       /* square it */
       if ((err = mp_sqr(&M[(size_t)1 << (winsize - 1)],
-                        &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) {
-         goto LBL_MU;
-      }
+                        &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) goto LBL_MU;
 
       /* reduce modulo P */
-      if ((err = redux(&M[(size_t)1 << (winsize - 1)], P, &mu)) != MP_OKAY) {
-         goto LBL_MU;
-      }
+      if ((err = redux(&M[(size_t)1 << (winsize - 1)], P, &mu)) != MP_OKAY) goto LBL_MU;
    }
 
    /* create upper table, that is M[x] = M[x-1] * M[1] (mod P)
     * for x = (2**(winsize - 1) + 1) to (2**winsize - 1)
     */
    for (x = (1 << (winsize - 1)) + 1; x < (1 << winsize); x++) {
-      if ((err = mp_mul(&M[x - 1], &M[1], &M[x])) != MP_OKAY) {
-         goto LBL_MU;
-      }
-      if ((err = redux(&M[x], P, &mu)) != MP_OKAY) {
-         goto LBL_MU;
-      }
+      if ((err = mp_mul(&M[x - 1], &M[1], &M[x])) != MP_OKAY)     goto LBL_MU;
+      if ((err = redux(&M[x], P, &mu)) != MP_OKAY)                goto LBL_MU;
    }
 
    /* setup result */
-   if ((err = mp_init(&res)) != MP_OKAY) {
-      goto LBL_MU;
-   }
+   if ((err = mp_init(&res)) != MP_OKAY)                          goto LBL_MU;
    mp_set(&res, 1uL);
 
    /* set initial mode and bit cnt */
@@ -150,11 +120,11 @@ int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, i
          }
          /* read next digit and reset the bitcnt */
          buf    = X->dp[digidx--];
-         bitcnt = (int)DIGIT_BIT;
+         bitcnt = (int)MP_DIGIT_BIT;
       }
 
       /* grab the next msb from the exponent */
-      y     = (buf >> (mp_digit)(DIGIT_BIT - 1)) & 1;
+      y     = (buf >> (mp_digit)(MP_DIGIT_BIT - 1)) & 1uL;
       buf <<= (mp_digit)1;
 
       /* if the bit is zero and mode == 0 then we ignore it
@@ -168,12 +138,8 @@ int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, i
 
       /* if the bit is zero and mode == 1 then we square */
       if ((mode == 1) && (y == 0)) {
-         if ((err = mp_sqr(&res, &res)) != MP_OKAY) {
-            goto LBL_RES;
-         }
-         if ((err = redux(&res, P, &mu)) != MP_OKAY) {
-            goto LBL_RES;
-         }
+         if ((err = mp_sqr(&res, &res)) != MP_OKAY)               goto LBL_RES;
+         if ((err = redux(&res, P, &mu)) != MP_OKAY)              goto LBL_RES;
          continue;
       }
 
@@ -185,21 +151,13 @@ int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, i
          /* ok window is filled so square as required and multiply  */
          /* square first */
          for (x = 0; x < winsize; x++) {
-            if ((err = mp_sqr(&res, &res)) != MP_OKAY) {
-               goto LBL_RES;
-            }
-            if ((err = redux(&res, P, &mu)) != MP_OKAY) {
-               goto LBL_RES;
-            }
+            if ((err = mp_sqr(&res, &res)) != MP_OKAY)            goto LBL_RES;
+            if ((err = redux(&res, P, &mu)) != MP_OKAY)           goto LBL_RES;
          }
 
          /* then multiply */
-         if ((err = mp_mul(&res, &M[bitbuf], &res)) != MP_OKAY) {
-            goto LBL_RES;
-         }
-         if ((err = redux(&res, P, &mu)) != MP_OKAY) {
-            goto LBL_RES;
-         }
+         if ((err = mp_mul(&res, &M[bitbuf], &res)) != MP_OKAY)  goto LBL_RES;
+         if ((err = redux(&res, P, &mu)) != MP_OKAY)             goto LBL_RES;
 
          /* empty window and reset */
          bitcpy = 0;
@@ -212,22 +170,14 @@ int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, i
    if ((mode == 2) && (bitcpy > 0)) {
       /* square then multiply if the bit is set */
       for (x = 0; x < bitcpy; x++) {
-         if ((err = mp_sqr(&res, &res)) != MP_OKAY) {
-            goto LBL_RES;
-         }
-         if ((err = redux(&res, P, &mu)) != MP_OKAY) {
-            goto LBL_RES;
-         }
+         if ((err = mp_sqr(&res, &res)) != MP_OKAY)               goto LBL_RES;
+         if ((err = redux(&res, P, &mu)) != MP_OKAY)              goto LBL_RES;
 
          bitbuf <<= 1;
          if ((bitbuf & (1 << winsize)) != 0) {
             /* then multiply */
-            if ((err = mp_mul(&res, &M[1], &res)) != MP_OKAY) {
-               goto LBL_RES;
-            }
-            if ((err = redux(&res, P, &mu)) != MP_OKAY) {
-               goto LBL_RES;
-            }
+            if ((err = mp_mul(&res, &M[1], &res)) != MP_OKAY)     goto LBL_RES;
+            if ((err = redux(&res, P, &mu)) != MP_OKAY)           goto LBL_RES;
          }
       }
    }
@@ -246,7 +196,3 @@ LBL_M:
    return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_exptmod_fast.c b/libtommath/bn_s_mp_exptmod_fast.c
index 4de9c5f..3a0dc5b 100644
--- a/libtommath/bn_mp_exptmod_fast.c
+++ b/libtommath/bn_s_mp_exptmod_fast.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
-#ifdef BN_MP_EXPTMOD_FAST_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_S_MP_EXPTMOD_FAST_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* computes Y == G**X mod P, HAC pp.616, Algorithm 14.85
  *
@@ -22,21 +13,24 @@
 
 #ifdef MP_LOW_MEM
 #   define TAB_SIZE 32
+#   define MAX_WINSIZE 5
 #else
 #   define TAB_SIZE 256
+#   define MAX_WINSIZE 0
 #endif
 
-int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode)
+mp_err s_mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode)
 {
    mp_int  M[TAB_SIZE], res;
    mp_digit buf, mp;
-   int     err, bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
+   int     bitbuf, bitcpy, bitcnt, mode, digidx, x, y, winsize;
+   mp_err   err;
 
    /* use a pointer to the reduction algorithm.  This allows us to use
     * one of many reduction algorithms without modding the guts of
     * the code with if statements everywhere.
     */
-   int (*redux)(mp_int *x, const mp_int *n, mp_digit rho);
+   mp_err(*redux)(mp_int *x, const mp_int *n, mp_digit rho);
 
    /* find window size */
    x = mp_count_bits(X);
@@ -56,11 +50,7 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
       winsize = 8;
    }
 
-#ifdef MP_LOW_MEM
-   if (winsize > 5) {
-      winsize = 5;
-   }
-#endif
+   winsize = MAX_WINSIZE ? MP_MIN(MAX_WINSIZE, winsize) : winsize;
 
    /* init M array */
    /* init first cell */
@@ -81,58 +71,46 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
 
    /* determine and setup reduction code */
    if (redmode == 0) {
-#ifdef BN_MP_MONTGOMERY_SETUP_C
-      /* now setup montgomery  */
-      if ((err = mp_montgomery_setup(P, &mp)) != MP_OKAY) {
+      if (MP_HAS(MP_MONTGOMERY_SETUP)) {
+         /* now setup montgomery  */
+         if ((err = mp_montgomery_setup(P, &mp)) != MP_OKAY)      goto LBL_M;
+      } else {
+         err = MP_VAL;
          goto LBL_M;
       }
-#else
-      err = MP_VAL;
-      goto LBL_M;
-#endif
 
       /* automatically pick the comba one if available (saves quite a few calls/ifs) */
-#ifdef BN_FAST_MP_MONTGOMERY_REDUCE_C
-      if ((((P->used * 2) + 1) < (int)MP_WARRAY) &&
-          (P->used < (1 << ((CHAR_BIT * sizeof(mp_word)) - (2 * DIGIT_BIT))))) {
-         redux = fast_mp_montgomery_reduce;
-      } else
-#endif
-      {
-#ifdef BN_MP_MONTGOMERY_REDUCE_C
+      if (MP_HAS(S_MP_MONTGOMERY_REDUCE_FAST) &&
+          (((P->used * 2) + 1) < PRIVATE_MP_WARRAY) &&
+          (P->used < MP_MAXFAST)) {
+         redux = s_mp_montgomery_reduce_fast;
+      } else if (MP_HAS(MP_MONTGOMERY_REDUCE)) {
          /* use slower baseline Montgomery method */
          redux = mp_montgomery_reduce;
-#else
+      } else {
          err = MP_VAL;
          goto LBL_M;
-#endif
       }
    } else if (redmode == 1) {
-#if defined(BN_MP_DR_SETUP_C) && defined(BN_MP_DR_REDUCE_C)
-      /* setup DR reduction for moduli of the form B**k - b */
-      mp_dr_setup(P, &mp);
-      redux = mp_dr_reduce;
-#else
-      err = MP_VAL;
-      goto LBL_M;
-#endif
-   } else {
-#if defined(BN_MP_REDUCE_2K_SETUP_C) && defined(BN_MP_REDUCE_2K_C)
-      /* setup DR reduction for moduli of the form 2**k - b */
-      if ((err = mp_reduce_2k_setup(P, &mp)) != MP_OKAY) {
+      if (MP_HAS(MP_DR_SETUP) && MP_HAS(MP_DR_REDUCE)) {
+         /* setup DR reduction for moduli of the form B**k - b */
+         mp_dr_setup(P, &mp);
+         redux = mp_dr_reduce;
+      } else {
+         err = MP_VAL;
          goto LBL_M;
       }
+   } else if (MP_HAS(MP_REDUCE_2K_SETUP) && MP_HAS(MP_REDUCE_2K)) {
+      /* setup DR reduction for moduli of the form 2**k - b */
+      if ((err = mp_reduce_2k_setup(P, &mp)) != MP_OKAY)          goto LBL_M;
       redux = mp_reduce_2k;
-#else
+   } else {
       err = MP_VAL;
       goto LBL_M;
-#endif
    }
 
    /* setup result */
-   if ((err = mp_init_size(&res, P->alloc)) != MP_OKAY) {
-      goto LBL_M;
-   }
+   if ((err = mp_init_size(&res, P->alloc)) != MP_OKAY)           goto LBL_M;
 
    /* create M table
     *
@@ -142,49 +120,33 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
     */
 
    if (redmode == 0) {
-#ifdef BN_MP_MONTGOMERY_CALC_NORMALIZATION_C
-      /* now we need R mod m */
-      if ((err = mp_montgomery_calc_normalization(&res, P)) != MP_OKAY) {
-         goto LBL_RES;
-      }
+      if (MP_HAS(MP_MONTGOMERY_CALC_NORMALIZATION)) {
+         /* now we need R mod m */
+         if ((err = mp_montgomery_calc_normalization(&res, P)) != MP_OKAY) goto LBL_RES;
 
-      /* now set M[1] to G * R mod m */
-      if ((err = mp_mulmod(G, &res, P, &M[1])) != MP_OKAY) {
+         /* now set M[1] to G * R mod m */
+         if ((err = mp_mulmod(G, &res, P, &M[1])) != MP_OKAY)     goto LBL_RES;
+      } else {
+         err = MP_VAL;
          goto LBL_RES;
       }
-#else
-      err = MP_VAL;
-      goto LBL_RES;
-#endif
    } else {
       mp_set(&res, 1uL);
-      if ((err = mp_mod(G, P, &M[1])) != MP_OKAY) {
-         goto LBL_RES;
-      }
+      if ((err = mp_mod(G, P, &M[1])) != MP_OKAY)                 goto LBL_RES;
    }
 
    /* compute the value at M[1<<(winsize-1)] by squaring M[1] (winsize-1) times */
-   if ((err = mp_copy(&M[1], &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) {
-      goto LBL_RES;
-   }
+   if ((err = mp_copy(&M[1], &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) goto LBL_RES;
 
    for (x = 0; x < (winsize - 1); x++) {
-      if ((err = mp_sqr(&M[(size_t)1 << (winsize - 1)], &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) {
-         goto LBL_RES;
-      }
-      if ((err = redux(&M[(size_t)1 << (winsize - 1)], P, mp)) != MP_OKAY) {
-         goto LBL_RES;
-      }
+      if ((err = mp_sqr(&M[(size_t)1 << (winsize - 1)], &M[(size_t)1 << (winsize - 1)])) != MP_OKAY) goto LBL_RES;
+      if ((err = redux(&M[(size_t)1 << (winsize - 1)], P, mp)) != MP_OKAY) goto LBL_RES;
    }
 
    /* create upper table */
    for (x = (1 << (winsize - 1)) + 1; x < (1 << winsize); x++) {
-      if ((err = mp_mul(&M[x - 1], &M[1], &M[x])) != MP_OKAY) {
-         goto LBL_RES;
-      }
-      if ((err = redux(&M[x], P, mp)) != MP_OKAY) {
-         goto LBL_RES;
-      }
+      if ((err = mp_mul(&M[x - 1], &M[1], &M[x])) != MP_OKAY)     goto LBL_RES;
+      if ((err = redux(&M[x], P, mp)) != MP_OKAY)                 goto LBL_RES;
    }
 
    /* set initial mode and bit cnt */
@@ -204,11 +166,11 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
          }
          /* read next digit and reset bitcnt */
          buf    = X->dp[digidx--];
-         bitcnt = (int)DIGIT_BIT;
+         bitcnt = (int)MP_DIGIT_BIT;
       }
 
       /* grab the next msb from the exponent */
-      y     = (mp_digit)(buf >> (DIGIT_BIT - 1)) & 1;
+      y     = (mp_digit)(buf >> (MP_DIGIT_BIT - 1)) & 1uL;
       buf <<= (mp_digit)1;
 
       /* if the bit is zero and mode == 0 then we ignore it
@@ -222,12 +184,8 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
 
       /* if the bit is zero and mode == 1 then we square */
       if ((mode == 1) && (y == 0)) {
-         if ((err = mp_sqr(&res, &res)) != MP_OKAY) {
-            goto LBL_RES;
-         }
-         if ((err = redux(&res, P, mp)) != MP_OKAY) {
-            goto LBL_RES;
-         }
+         if ((err = mp_sqr(&res, &res)) != MP_OKAY)               goto LBL_RES;
+         if ((err = redux(&res, P, mp)) != MP_OKAY)               goto LBL_RES;
          continue;
       }
 
@@ -239,21 +197,13 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
          /* ok window is filled so square as required and multiply  */
          /* square first */
          for (x = 0; x < winsize; x++) {
-            if ((err = mp_sqr(&res, &res)) != MP_OKAY) {
-               goto LBL_RES;
-            }
-            if ((err = redux(&res, P, mp)) != MP_OKAY) {
-               goto LBL_RES;
-            }
+            if ((err = mp_sqr(&res, &res)) != MP_OKAY)            goto LBL_RES;
+            if ((err = redux(&res, P, mp)) != MP_OKAY)            goto LBL_RES;
          }
 
          /* then multiply */
-         if ((err = mp_mul(&res, &M[bitbuf], &res)) != MP_OKAY) {
-            goto LBL_RES;
-         }
-         if ((err = redux(&res, P, mp)) != MP_OKAY) {
-            goto LBL_RES;
-         }
+         if ((err = mp_mul(&res, &M[bitbuf], &res)) != MP_OKAY)   goto LBL_RES;
+         if ((err = redux(&res, P, mp)) != MP_OKAY)               goto LBL_RES;
 
          /* empty window and reset */
          bitcpy = 0;
@@ -266,23 +216,15 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
    if ((mode == 2) && (bitcpy > 0)) {
       /* square then multiply if the bit is set */
       for (x = 0; x < bitcpy; x++) {
-         if ((err = mp_sqr(&res, &res)) != MP_OKAY) {
-            goto LBL_RES;
-         }
-         if ((err = redux(&res, P, mp)) != MP_OKAY) {
-            goto LBL_RES;
-         }
+         if ((err = mp_sqr(&res, &res)) != MP_OKAY)               goto LBL_RES;
+         if ((err = redux(&res, P, mp)) != MP_OKAY)               goto LBL_RES;
 
          /* get next bit of the window */
          bitbuf <<= 1;
          if ((bitbuf & (1 << winsize)) != 0) {
             /* then multiply */
-            if ((err = mp_mul(&res, &M[1], &res)) != MP_OKAY) {
-               goto LBL_RES;
-            }
-            if ((err = redux(&res, P, mp)) != MP_OKAY) {
-               goto LBL_RES;
-            }
+            if ((err = mp_mul(&res, &M[1], &res)) != MP_OKAY)     goto LBL_RES;
+            if ((err = redux(&res, P, mp)) != MP_OKAY)            goto LBL_RES;
          }
       }
    }
@@ -294,9 +236,7 @@ int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y
        * to reduce one more time to cancel out the factor
        * of R.
        */
-      if ((err = redux(&res, P, mp)) != MP_OKAY) {
-         goto LBL_RES;
-      }
+      if ((err = redux(&res, P, mp)) != MP_OKAY)                  goto LBL_RES;
    }
 
    /* swap res with Y */
@@ -312,8 +252,3 @@ LBL_M:
    return err;
 }
 #endif
-
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_get_bit.c b/libtommath/bn_s_mp_get_bit.c
index da9ccbb..28598df 100644
--- a/libtommath/bn_s_mp_get_bit.c
+++ b/libtommath/bn_s_mp_get_bit.c
@@ -5,16 +5,16 @@
 /* SPDX-License-Identifier: Unlicense */
 
 /* Get bit at position b and return MP_YES if the bit is 1, MP_NO if it is 0 */
-mp_bool s_mp_get_bit(const mp_int *a, int b)
+mp_bool s_mp_get_bit(const mp_int *a, unsigned int b)
 {
    mp_digit bit;
-   int limb = (int)((unsigned)b / MP_DIGIT_BIT);
+   int limb = (int)(b / MP_DIGIT_BIT);
 
    if (limb >= a->used) {
       return MP_NO;
    }
 
-   bit = (mp_digit)1 << ((unsigned)b % MP_DIGIT_BIT);
+   bit = (mp_digit)1 << (b % MP_DIGIT_BIT);
    return ((a->dp[limb] & bit) != 0u) ? MP_YES : MP_NO;
 }
 
diff --git a/libtommath/bn_s_mp_invmod_fast.c b/libtommath/bn_s_mp_invmod_fast.c
new file mode 100644
index 0000000..677d7ab
--- /dev/null
+++ b/libtommath/bn_s_mp_invmod_fast.c
@@ -0,0 +1,118 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_INVMOD_FAST_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* computes the modular inverse via binary extended euclidean algorithm,
+ * that is c = 1/a mod b
+ *
+ * Based on slow invmod except this is optimized for the case where b is
+ * odd as per HAC Note 14.64 on pp. 610
+ */
+mp_err s_mp_invmod_fast(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   mp_int  x, y, u, v, B, D;
+   mp_sign neg;
+   mp_err  err;
+
+   /* 2. [modified] b must be odd   */
+   if (MP_IS_EVEN(b)) {
+      return MP_VAL;
+   }
+
+   /* init all our temps */
+   if ((err = mp_init_multi(&x, &y, &u, &v, &B, &D, NULL)) != MP_OKAY) {
+      return err;
+   }
+
+   /* x == modulus, y == value to invert */
+   if ((err = mp_copy(b, &x)) != MP_OKAY)                         goto LBL_ERR;
+
+   /* we need y = |a| */
+   if ((err = mp_mod(a, b, &y)) != MP_OKAY)                       goto LBL_ERR;
+
+   /* if one of x,y is zero return an error! */
+   if (MP_IS_ZERO(&x) || MP_IS_ZERO(&y)) {
+      err = MP_VAL;
+      goto LBL_ERR;
+   }
+
+   /* 3. u=x, v=y, A=1, B=0, C=0,D=1 */
+   if ((err = mp_copy(&x, &u)) != MP_OKAY)                        goto LBL_ERR;
+   if ((err = mp_copy(&y, &v)) != MP_OKAY)                        goto LBL_ERR;
+   mp_set(&D, 1uL);
+
+top:
+   /* 4.  while u is even do */
+   while (MP_IS_EVEN(&u)) {
+      /* 4.1 u = u/2 */
+      if ((err = mp_div_2(&u, &u)) != MP_OKAY)                    goto LBL_ERR;
+
+      /* 4.2 if B is odd then */
+      if (MP_IS_ODD(&B)) {
+         if ((err = mp_sub(&B, &x, &B)) != MP_OKAY)               goto LBL_ERR;
+      }
+      /* B = B/2 */
+      if ((err = mp_div_2(&B, &B)) != MP_OKAY)                    goto LBL_ERR;
+   }
+
+   /* 5.  while v is even do */
+   while (MP_IS_EVEN(&v)) {
+      /* 5.1 v = v/2 */
+      if ((err = mp_div_2(&v, &v)) != MP_OKAY)                    goto LBL_ERR;
+
+      /* 5.2 if D is odd then */
+      if (MP_IS_ODD(&D)) {
+         /* D = (D-x)/2 */
+         if ((err = mp_sub(&D, &x, &D)) != MP_OKAY)               goto LBL_ERR;
+      }
+      /* D = D/2 */
+      if ((err = mp_div_2(&D, &D)) != MP_OKAY)                    goto LBL_ERR;
+   }
+
+   /* 6.  if u >= v then */
+   if (mp_cmp(&u, &v) != MP_LT) {
+      /* u = u - v, B = B - D */
+      if ((err = mp_sub(&u, &v, &u)) != MP_OKAY)                  goto LBL_ERR;
+
+      if ((err = mp_sub(&B, &D, &B)) != MP_OKAY)                  goto LBL_ERR;
+   } else {
+      /* v - v - u, D = D - B */
+      if ((err = mp_sub(&v, &u, &v)) != MP_OKAY)                  goto LBL_ERR;
+
+      if ((err = mp_sub(&D, &B, &D)) != MP_OKAY)                  goto LBL_ERR;
+   }
+
+   /* if not zero goto step 4 */
+   if (!MP_IS_ZERO(&u)) {
+      goto top;
+   }
+
+   /* now a = C, b = D, gcd == g*v */
+
+   /* if v != 1 then there is no inverse */
+   if (mp_cmp_d(&v, 1uL) != MP_EQ) {
+      err = MP_VAL;
+      goto LBL_ERR;
+   }
+
+   /* b is now the inverse */
+   neg = a->sign;
+   while (D.sign == MP_NEG) {
+      if ((err = mp_add(&D, b, &D)) != MP_OKAY)                   goto LBL_ERR;
+   }
+
+   /* too big */
+   while (mp_cmp_mag(&D, b) != MP_LT) {
+      if ((err = mp_sub(&D, b, &D)) != MP_OKAY)                   goto LBL_ERR;
+   }
+
+   mp_exch(&D, c);
+   c->sign = neg;
+   err = MP_OKAY;
+
+LBL_ERR:
+   mp_clear_multi(&x, &y, &u, &v, &B, &D, NULL);
+   return err;
+}
+#endif
diff --git a/libtommath/bn_s_mp_invmod_slow.c b/libtommath/bn_s_mp_invmod_slow.c
new file mode 100644
index 0000000..4c5db33
--- /dev/null
+++ b/libtommath/bn_s_mp_invmod_slow.c
@@ -0,0 +1,119 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_INVMOD_SLOW_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* hac 14.61, pp608 */
+mp_err s_mp_invmod_slow(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   mp_int  x, y, u, v, A, B, C, D;
+   mp_err  err;
+
+   /* b cannot be negative */
+   if ((b->sign == MP_NEG) || MP_IS_ZERO(b)) {
+      return MP_VAL;
+   }
+
+   /* init temps */
+   if ((err = mp_init_multi(&x, &y, &u, &v,
+                            &A, &B, &C, &D, NULL)) != MP_OKAY) {
+      return err;
+   }
+
+   /* x = a, y = b */
+   if ((err = mp_mod(a, b, &x)) != MP_OKAY)                       goto LBL_ERR;
+   if ((err = mp_copy(b, &y)) != MP_OKAY)                         goto LBL_ERR;
+
+   /* 2. [modified] if x,y are both even then return an error! */
+   if (MP_IS_EVEN(&x) && MP_IS_EVEN(&y)) {
+      err = MP_VAL;
+      goto LBL_ERR;
+   }
+
+   /* 3. u=x, v=y, A=1, B=0, C=0,D=1 */
+   if ((err = mp_copy(&x, &u)) != MP_OKAY)                        goto LBL_ERR;
+   if ((err = mp_copy(&y, &v)) != MP_OKAY)                        goto LBL_ERR;
+   mp_set(&A, 1uL);
+   mp_set(&D, 1uL);
+
+top:
+   /* 4.  while u is even do */
+   while (MP_IS_EVEN(&u)) {
+      /* 4.1 u = u/2 */
+      if ((err = mp_div_2(&u, &u)) != MP_OKAY)                    goto LBL_ERR;
+
+      /* 4.2 if A or B is odd then */
+      if (MP_IS_ODD(&A) || MP_IS_ODD(&B)) {
+         /* A = (A+y)/2, B = (B-x)/2 */
+         if ((err = mp_add(&A, &y, &A)) != MP_OKAY)               goto LBL_ERR;
+         if ((err = mp_sub(&B, &x, &B)) != MP_OKAY)               goto LBL_ERR;
+      }
+      /* A = A/2, B = B/2 */
+      if ((err = mp_div_2(&A, &A)) != MP_OKAY)                    goto LBL_ERR;
+      if ((err = mp_div_2(&B, &B)) != MP_OKAY)                    goto LBL_ERR;
+   }
+
+   /* 5.  while v is even do */
+   while (MP_IS_EVEN(&v)) {
+      /* 5.1 v = v/2 */
+      if ((err = mp_div_2(&v, &v)) != MP_OKAY)                    goto LBL_ERR;
+
+      /* 5.2 if C or D is odd then */
+      if (MP_IS_ODD(&C) || MP_IS_ODD(&D)) {
+         /* C = (C+y)/2, D = (D-x)/2 */
+         if ((err = mp_add(&C, &y, &C)) != MP_OKAY)               goto LBL_ERR;
+         if ((err = mp_sub(&D, &x, &D)) != MP_OKAY)               goto LBL_ERR;
+      }
+      /* C = C/2, D = D/2 */
+      if ((err = mp_div_2(&C, &C)) != MP_OKAY)                    goto LBL_ERR;
+      if ((err = mp_div_2(&D, &D)) != MP_OKAY)                    goto LBL_ERR;
+   }
+
+   /* 6.  if u >= v then */
+   if (mp_cmp(&u, &v) != MP_LT) {
+      /* u = u - v, A = A - C, B = B - D */
+      if ((err = mp_sub(&u, &v, &u)) != MP_OKAY)                  goto LBL_ERR;
+
+      if ((err = mp_sub(&A, &C, &A)) != MP_OKAY)                  goto LBL_ERR;
+
+      if ((err = mp_sub(&B, &D, &B)) != MP_OKAY)                  goto LBL_ERR;
+   } else {
+      /* v - v - u, C = C - A, D = D - B */
+      if ((err = mp_sub(&v, &u, &v)) != MP_OKAY)                  goto LBL_ERR;
+
+      if ((err = mp_sub(&C, &A, &C)) != MP_OKAY)                  goto LBL_ERR;
+
+      if ((err = mp_sub(&D, &B, &D)) != MP_OKAY)                  goto LBL_ERR;
+   }
+
+   /* if not zero goto step 4 */
+   if (!MP_IS_ZERO(&u)) {
+      goto top;
+   }
+
+   /* now a = C, b = D, gcd == g*v */
+
+   /* if v != 1 then there is no inverse */
+   if (mp_cmp_d(&v, 1uL) != MP_EQ) {
+      err = MP_VAL;
+      goto LBL_ERR;
+   }
+
+   /* if its too low */
+   while (mp_cmp_d(&C, 0uL) == MP_LT) {
+      if ((err = mp_add(&C, b, &C)) != MP_OKAY)                   goto LBL_ERR;
+   }
+
+   /* too big */
+   while (mp_cmp_mag(&C, b) != MP_LT) {
+      if ((err = mp_sub(&C, b, &C)) != MP_OKAY)                   goto LBL_ERR;
+   }
+
+   /* C is now the inverse */
+   mp_exch(&C, c);
+   err = MP_OKAY;
+LBL_ERR:
+   mp_clear_multi(&x, &y, &u, &v, &A, &B, &C, &D, NULL);
+   return err;
+}
+#endif
diff --git a/libtommath/bn_mp_karatsuba_mul.c b/libtommath/bn_s_mp_karatsuba_mul.c
index cb75bca..85899fb 100644
--- a/libtommath/bn_mp_karatsuba_mul.c
+++ b/libtommath/bn_s_mp_karatsuba_mul.c
@@ -1,21 +1,12 @@
 #include "tommath_private.h"
-#ifdef BN_MP_KARATSUBA_MUL_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_S_MP_KARATSUBA_MUL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* c = |a| * |b| using Karatsuba Multiplication using
  * three half size multiplications
  *
- * Let B represent the radix [e.g. 2**DIGIT_BIT] and
+ * Let B represent the radix [e.g. 2**MP_DIGIT_BIT] and
  * let n represent half of the number of digits in
  * the min(a,b)
  *
@@ -41,37 +32,42 @@
  * Generally though the overhead of this method doesn't pay off
  * until a certain size (N ~ 80) is reached.
  */
-int mp_karatsuba_mul(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err s_mp_karatsuba_mul(const mp_int *a, const mp_int *b, mp_int *c)
 {
    mp_int  x0, x1, y0, y1, t1, x0y0, x1y1;
-   int     B, err;
-
-   /* default the return code to an error */
-   err = MP_MEM;
+   int     B;
+   mp_err  err = MP_MEM; /* default the return code to an error */
 
    /* min # of digits */
-   B = MIN(a->used, b->used);
+   B = MP_MIN(a->used, b->used);
 
    /* now divide in two */
    B = B >> 1;
 
    /* init copy all the temps */
-   if (mp_init_size(&x0, B) != MP_OKAY)
+   if (mp_init_size(&x0, B) != MP_OKAY) {
       goto LBL_ERR;
-   if (mp_init_size(&x1, a->used - B) != MP_OKAY)
+   }
+   if (mp_init_size(&x1, a->used - B) != MP_OKAY) {
       goto X0;
-   if (mp_init_size(&y0, B) != MP_OKAY)
+   }
+   if (mp_init_size(&y0, B) != MP_OKAY) {
       goto X1;
-   if (mp_init_size(&y1, b->used - B) != MP_OKAY)
+   }
+   if (mp_init_size(&y1, b->used - B) != MP_OKAY) {
       goto Y0;
+   }
 
    /* init temps */
-   if (mp_init_size(&t1, B * 2) != MP_OKAY)
+   if (mp_init_size(&t1, B * 2) != MP_OKAY) {
       goto Y1;
-   if (mp_init_size(&x0y0, B * 2) != MP_OKAY)
+   }
+   if (mp_init_size(&x0y0, B * 2) != MP_OKAY) {
       goto T1;
-   if (mp_init_size(&x1y1, B * 2) != MP_OKAY)
+   }
+   if (mp_init_size(&x1y1, B * 2) != MP_OKAY) {
       goto X0Y0;
+   }
 
    /* now shift the digits */
    x0.used = y0.used = B;
@@ -114,35 +110,46 @@ int mp_karatsuba_mul(const mp_int *a, const mp_int *b, mp_int *c)
 
    /* now calc the products x0y0 and x1y1 */
    /* after this x0 is no longer required, free temp [x0==t2]! */
-   if (mp_mul(&x0, &y0, &x0y0) != MP_OKAY)
+   if (mp_mul(&x0, &y0, &x0y0) != MP_OKAY) {
       goto X1Y1;          /* x0y0 = x0*y0 */
-   if (mp_mul(&x1, &y1, &x1y1) != MP_OKAY)
+   }
+   if (mp_mul(&x1, &y1, &x1y1) != MP_OKAY) {
       goto X1Y1;          /* x1y1 = x1*y1 */
+   }
 
    /* now calc x1+x0 and y1+y0 */
-   if (s_mp_add(&x1, &x0, &t1) != MP_OKAY)
+   if (s_mp_add(&x1, &x0, &t1) != MP_OKAY) {
       goto X1Y1;          /* t1 = x1 - x0 */
-   if (s_mp_add(&y1, &y0, &x0) != MP_OKAY)
+   }
+   if (s_mp_add(&y1, &y0, &x0) != MP_OKAY) {
       goto X1Y1;          /* t2 = y1 - y0 */
-   if (mp_mul(&t1, &x0, &t1) != MP_OKAY)
+   }
+   if (mp_mul(&t1, &x0, &t1) != MP_OKAY) {
       goto X1Y1;          /* t1 = (x1 + x0) * (y1 + y0) */
+   }
 
    /* add x0y0 */
-   if (mp_add(&x0y0, &x1y1, &x0) != MP_OKAY)
+   if (mp_add(&x0y0, &x1y1, &x0) != MP_OKAY) {
       goto X1Y1;          /* t2 = x0y0 + x1y1 */
-   if (s_mp_sub(&t1, &x0, &t1) != MP_OKAY)
+   }
+   if (s_mp_sub(&t1, &x0, &t1) != MP_OKAY) {
       goto X1Y1;          /* t1 = (x1+x0)*(y1+y0) - (x1y1 + x0y0) */
+   }
 
    /* shift by B */
-   if (mp_lshd(&t1, B) != MP_OKAY)
+   if (mp_lshd(&t1, B) != MP_OKAY) {
       goto X1Y1;          /* t1 = (x0y0 + x1y1 - (x1-x0)*(y1-y0))<<B */
-   if (mp_lshd(&x1y1, B * 2) != MP_OKAY)
+   }
+   if (mp_lshd(&x1y1, B * 2) != MP_OKAY) {
       goto X1Y1;          /* x1y1 = x1y1 << 2*B */
+   }
 
-   if (mp_add(&x0y0, &t1, &t1) != MP_OKAY)
+   if (mp_add(&x0y0, &t1, &t1) != MP_OKAY) {
       goto X1Y1;          /* t1 = x0y0 + t1 */
-   if (mp_add(&t1, &x1y1, c) != MP_OKAY)
+   }
+   if (mp_add(&t1, &x1y1, c) != MP_OKAY) {
       goto X1Y1;          /* t1 = x0y0 + t1 + x1y1 */
+   }
 
    /* Algorithm succeeded set the return code to MP_OKAY */
    err = MP_OKAY;
@@ -165,7 +172,3 @@ LBL_ERR:
    return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_mp_karatsuba_sqr.c b/libtommath/bn_s_mp_karatsuba_sqr.c
index c219a37..f132d07 100644
--- a/libtommath/bn_mp_karatsuba_sqr.c
+++ b/libtommath/bn_s_mp_karatsuba_sqr.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
-#ifdef BN_MP_KARATSUBA_SQR_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_S_MP_KARATSUBA_SQR_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* Karatsuba squaring, computes b = a*a using three
  * half size squarings
@@ -19,12 +10,11 @@
  * is essentially the same algorithm but merely
  * tuned to perform recursive squarings.
  */
-int mp_karatsuba_sqr(const mp_int *a, mp_int *b)
+mp_err s_mp_karatsuba_sqr(const mp_int *a, mp_int *b)
 {
    mp_int  x0, x1, t1, t2, x0x0, x1x1;
-   int     B, err;
-
-   err = MP_MEM;
+   int     B;
+   mp_err  err = MP_MEM;
 
    /* min # of digits */
    B = a->used;
@@ -118,7 +108,3 @@ LBL_ERR:
    return err;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_fast_mp_montgomery_reduce.c b/libtommath/bn_s_mp_montgomery_reduce_fast.c
index eb5d90b..8c3fe5e 100644
--- a/libtommath/bn_fast_mp_montgomery_reduce.c
+++ b/libtommath/bn_s_mp_montgomery_reduce_fast.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
-#ifdef BN_FAST_MP_MONTGOMERY_REDUCE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_S_MP_MONTGOMERY_REDUCE_FAST_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* computes xR**-1 == x (mod N) via Montgomery Reduction
  *
@@ -20,12 +11,13 @@
  *
  * Based on Algorithm 14.32 on pp.601 of HAC.
 */
-int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
+mp_err s_mp_montgomery_reduce_fast(mp_int *x, const mp_int *n, mp_digit rho)
 {
-   int     ix, res, olduse;
-   mp_word W[MP_WARRAY];
+   int     ix, olduse;
+   mp_err  err;
+   mp_word W[PRIVATE_MP_WARRAY];
 
-   if (x->used > (int)MP_WARRAY) {
+   if (x->used > PRIVATE_MP_WARRAY) {
       return MP_VAL;
    }
 
@@ -34,8 +26,8 @@ int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
 
    /* grow a as required */
    if (x->alloc < (n->used + 1)) {
-      if ((res = mp_grow(x, n->used + 1)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(x, n->used + 1)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -58,8 +50,8 @@ int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
       }
 
       /* zero the high words of W[a->used..m->used*2] */
-      for (; ix < ((n->used * 2) + 1); ix++) {
-         *_W++ = 0;
+      if (ix < ((n->used * 2) + 1)) {
+         MP_ZERO_BUFFER(_W, sizeof(mp_word) * (size_t)(((n->used * 2) + 1) - ix));
       }
    }
 
@@ -108,7 +100,7 @@ int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
       }
 
       /* now fix carry for next digit, W[ix+1] */
-      W[ix + 1] += W[ix] >> (mp_word)DIGIT_BIT;
+      W[ix + 1] += W[ix] >> (mp_word)MP_DIGIT_BIT;
    }
 
    /* now we have to propagate the carries and
@@ -127,8 +119,8 @@ int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
       /* alias for next word, where the carry goes */
       _W = W + ++ix;
 
-      for (; ix <= ((n->used * 2) + 1); ix++) {
-         *_W++ += *_W1++ >> (mp_word)DIGIT_BIT;
+      for (; ix < ((n->used * 2) + 1); ix++) {
+         *_W++ += *_W1++ >> (mp_word)MP_DIGIT_BIT;
       }
 
       /* copy out, A = A/b**n
@@ -151,9 +143,7 @@ int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
       /* zero oldused digits, if the input a was larger than
        * m->used+1 we'll have to clear the digits
        */
-      for (; ix < olduse; ix++) {
-         *tmpx++ = 0;
-      }
+      MP_ZERO_DIGITS(tmpx, olduse - ix);
    }
 
    /* set the max used and clamp */
@@ -167,7 +157,3 @@ int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_mul_digs.c b/libtommath/bn_s_mp_mul_digs.c
index 332e974..6c1114e 100644
--- a/libtommath/bn_s_mp_mul_digs.c
+++ b/libtommath/bn_s_mp_mul_digs.c
@@ -1,38 +1,29 @@
 #include "tommath_private.h"
 #ifdef BN_S_MP_MUL_DIGS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* multiplies |a| * |b| and only computes upto digs digits of result
  * HAC pp. 595, Algorithm 14.12  Modified so you can control how
  * many digits of output are created.
  */
-int s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
+mp_err s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
 {
    mp_int  t;
-   int     res, pa, pb, ix, iy;
+   mp_err  err;
+   int     pa, pb, ix, iy;
    mp_digit u;
    mp_word r;
    mp_digit tmpx, *tmpt, *tmpy;
 
    /* can we use the fast multiplier? */
-   if ((digs < (int)MP_WARRAY) &&
-       (MIN(a->used, b->used) <
-        (int)(1u << (((size_t)CHAR_BIT * sizeof(mp_word)) - (2u * (size_t)DIGIT_BIT))))) {
-      return fast_s_mp_mul_digs(a, b, c, digs);
+   if ((digs < PRIVATE_MP_WARRAY) &&
+       (MP_MIN(a->used, b->used) < MP_MAXFAST)) {
+      return s_mp_mul_digs_fast(a, b, c, digs);
    }
 
-   if ((res = mp_init_size(&t, digs)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&t, digs)) != MP_OKAY) {
+      return err;
    }
    t.used = digs;
 
@@ -43,7 +34,7 @@ int s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
       u = 0;
 
       /* limit ourselves to making digs digits of output */
-      pb = MIN(b->used, digs - ix);
+      pb = MP_MIN(b->used, digs - ix);
 
       /* setup some aliases */
       /* copy of the digit from a used within the nested loop */
@@ -66,7 +57,7 @@ int s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
          *tmpt++ = (mp_digit)(r & (mp_word)MP_MASK);
 
          /* get the carry word from the result */
-         u       = (mp_digit)(r >> (mp_word)DIGIT_BIT);
+         u       = (mp_digit)(r >> (mp_word)MP_DIGIT_BIT);
       }
       /* set carry if it is placed below digs */
       if ((ix + iy) < digs) {
@@ -81,7 +72,3 @@ int s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_fast_s_mp_mul_digs.c b/libtommath/bn_s_mp_mul_digs_fast.c
index 4736799..090a099 100644
--- a/libtommath/bn_fast_s_mp_mul_digs.c
+++ b/libtommath/bn_s_mp_mul_digs_fast.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
-#ifdef BN_FAST_S_MP_MUL_DIGS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_S_MP_MUL_DIGS_FAST_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* Fast (comba) multiplier
  *
@@ -28,21 +19,22 @@
  * Based on Algorithm 14.12 on pp.595 of HAC.
  *
  */
-int fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
+mp_err s_mp_mul_digs_fast(const mp_int *a, const mp_int *b, mp_int *c, int digs)
 {
-   int     olduse, res, pa, ix, iz;
-   mp_digit W[MP_WARRAY];
+   int      olduse, pa, ix, iz;
+   mp_err   err;
+   mp_digit W[PRIVATE_MP_WARRAY];
    mp_word  _W;
 
    /* grow the destination as required */
    if (c->alloc < digs) {
-      if ((res = mp_grow(c, digs)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(c, digs)) != MP_OKAY) {
+         return err;
       }
    }
 
    /* number of output digits to produce */
-   pa = MIN(digs, a->used + b->used);
+   pa = MP_MIN(digs, a->used + b->used);
 
    /* clear the carry */
    _W = 0;
@@ -52,7 +44,7 @@ int fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
       mp_digit *tmpx, *tmpy;
 
       /* get offsets into the two bignums */
-      ty = MIN(b->used-1, ix);
+      ty = MP_MIN(b->used-1, ix);
       tx = ix - ty;
 
       /* setup temp aliases */
@@ -62,7 +54,7 @@ int fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
       /* this is the number of times the loop will iterrate, essentially
          while (tx++ < a->used && ty-- >= 0) { ... }
        */
-      iy = MIN(a->used-tx, ty+1);
+      iy = MP_MIN(a->used-tx, ty+1);
 
       /* execute loop */
       for (iz = 0; iz < iy; ++iz) {
@@ -74,7 +66,7 @@ int fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
       W[ix] = (mp_digit)_W & MP_MASK;
 
       /* make next carry */
-      _W = _W >> (mp_word)DIGIT_BIT;
+      _W = _W >> (mp_word)MP_DIGIT_BIT;
    }
 
    /* setup dest */
@@ -90,15 +82,9 @@ int fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
       }
 
       /* clear unused digits [that existed in the old copy of c] */
-      for (; ix < olduse; ix++) {
-         *tmpc++ = 0;
-      }
+      MP_ZERO_DIGITS(tmpc, olduse - ix);
    }
    mp_clamp(c);
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_mul_high_digs.c b/libtommath/bn_s_mp_mul_high_digs.c
index 509682b..5c5d721 100644
--- a/libtommath/bn_s_mp_mul_high_digs.c
+++ b/libtommath/bn_s_mp_mul_high_digs.c
@@ -1,38 +1,29 @@
 #include "tommath_private.h"
 #ifdef BN_S_MP_MUL_HIGH_DIGS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* multiplies |a| * |b| and does not compute the lower digs digits
  * [meant to get the higher part of the product]
  */
-int s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
+mp_err s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
 {
-   mp_int  t;
-   int     res, pa, pb, ix, iy;
+   mp_int   t;
+   int      pa, pb, ix, iy;
+   mp_err   err;
    mp_digit u;
-   mp_word r;
+   mp_word  r;
    mp_digit tmpx, *tmpt, *tmpy;
 
    /* can we use the fast multiplier? */
-#ifdef BN_FAST_S_MP_MUL_HIGH_DIGS_C
-   if (((a->used + b->used + 1) < (int)MP_WARRAY)
-       && (MIN(a->used, b->used) < (int)(1u << (((size_t)CHAR_BIT * sizeof(mp_word)) - (2u * (size_t)DIGIT_BIT))))) {
-      return fast_s_mp_mul_high_digs(a, b, c, digs);
+   if (MP_HAS(S_MP_MUL_HIGH_DIGS_FAST)
+       && ((a->used + b->used + 1) < PRIVATE_MP_WARRAY)
+       && (MP_MIN(a->used, b->used) < MP_MAXFAST)) {
+      return s_mp_mul_high_digs_fast(a, b, c, digs);
    }
-#endif
 
-   if ((res = mp_init_size(&t, a->used + b->used + 1)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&t, a->used + b->used + 1)) != MP_OKAY) {
+      return err;
    }
    t.used = a->used + b->used + 1;
 
@@ -61,7 +52,7 @@ int s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
          *tmpt++ = (mp_digit)(r & (mp_word)MP_MASK);
 
          /* carry the carry */
-         u       = (mp_digit)(r >> (mp_word)DIGIT_BIT);
+         u       = (mp_digit)(r >> (mp_word)MP_DIGIT_BIT);
       }
       *tmpt = u;
    }
@@ -71,7 +62,3 @@ int s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_fast_s_mp_mul_high_digs.c b/libtommath/bn_s_mp_mul_high_digs_fast.c
index 06c076c..a74adc6 100644
--- a/libtommath/bn_fast_s_mp_mul_high_digs.c
+++ b/libtommath/bn_s_mp_mul_high_digs_fast.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
-#ifdef BN_FAST_S_MP_MUL_HIGH_DIGS_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_S_MP_MUL_HIGH_DIGS_FAST_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* this is a modified version of fast_s_mul_digs that only produces
  * output digits *above* digs.  See the comments for fast_s_mul_digs
@@ -21,17 +12,18 @@
  *
  * Based on Algorithm 14.12 on pp.595 of HAC.
  */
-int fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs)
+mp_err s_mp_mul_high_digs_fast(const mp_int *a, const mp_int *b, mp_int *c, int digs)
 {
-   int     olduse, res, pa, ix, iz;
-   mp_digit W[MP_WARRAY];
+   int     olduse, pa, ix, iz;
+   mp_err   err;
+   mp_digit W[PRIVATE_MP_WARRAY];
    mp_word  _W;
 
    /* grow the destination as required */
    pa = a->used + b->used;
    if (c->alloc < pa) {
-      if ((res = mp_grow(c, pa)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(c, pa)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -43,7 +35,7 @@ int fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int dig
       mp_digit *tmpx, *tmpy;
 
       /* get offsets into the two bignums */
-      ty = MIN(b->used-1, ix);
+      ty = MP_MIN(b->used-1, ix);
       tx = ix - ty;
 
       /* setup temp aliases */
@@ -53,7 +45,7 @@ int fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int dig
       /* this is the number of times the loop will iterrate, essentially its
          while (tx++ < a->used && ty-- >= 0) { ... }
        */
-      iy = MIN(a->used-tx, ty+1);
+      iy = MP_MIN(a->used-tx, ty+1);
 
       /* execute loop */
       for (iz = 0; iz < iy; iz++) {
@@ -64,7 +56,7 @@ int fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int dig
       W[ix] = (mp_digit)_W & MP_MASK;
 
       /* make next carry */
-      _W = _W >> (mp_word)DIGIT_BIT;
+      _W = _W >> (mp_word)MP_DIGIT_BIT;
    }
 
    /* setup dest */
@@ -81,15 +73,9 @@ int fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int dig
       }
 
       /* clear unused digits [that existed in the old copy of c] */
-      for (; ix < olduse; ix++) {
-         *tmpc++ = 0;
-      }
+      MP_ZERO_DIGITS(tmpc, olduse - ix);
    }
    mp_clamp(c);
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_prime_is_divisible.c b/libtommath/bn_s_mp_prime_is_divisible.c
new file mode 100644
index 0000000..ffd5093
--- /dev/null
+++ b/libtommath/bn_s_mp_prime_is_divisible.c
@@ -0,0 +1,35 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_PRIME_IS_DIVISIBLE_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* determines if an integers is divisible by one
+ * of the first PRIME_SIZE primes or not
+ *
+ * sets result to 0 if not, 1 if yes
+ */
+mp_err s_mp_prime_is_divisible(const mp_int *a, mp_bool *result)
+{
+   int      ix;
+   mp_err   err;
+   mp_digit res;
+
+   /* default to not */
+   *result = MP_NO;
+
+   for (ix = 0; ix < PRIVATE_MP_PRIME_TAB_SIZE; ix++) {
+      /* what is a mod LBL_prime_tab[ix] */
+      if ((err = mp_mod_d(a, s_mp_prime_tab[ix], &res)) != MP_OKAY) {
+         return err;
+      }
+
+      /* is the residue zero? */
+      if (res == 0u) {
+         *result = MP_YES;
+         return MP_OKAY;
+      }
+   }
+
+   return MP_OKAY;
+}
+#endif
diff --git a/libtommath/bn_s_mp_rand_jenkins.c b/libtommath/bn_s_mp_rand_jenkins.c
new file mode 100644
index 0000000..da0771c
--- /dev/null
+++ b/libtommath/bn_s_mp_rand_jenkins.c
@@ -0,0 +1,52 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_RAND_JENKINS_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* Bob Jenkins' http://burtleburtle.net/bob/rand/smallprng.html */
+/* Chosen for speed and a good "mix" */
+typedef struct {
+   uint64_t a;
+   uint64_t b;
+   uint64_t c;
+   uint64_t d;
+} ranctx;
+
+static ranctx jenkins_x;
+
+#define rot(x,k) (((x)<<(k))|((x)>>(64-(k))))
+static uint64_t s_rand_jenkins_val(void)
+{
+   uint64_t e = jenkins_x.a - rot(jenkins_x.b, 7);
+   jenkins_x.a = jenkins_x.b ^ rot(jenkins_x.c, 13);
+   jenkins_x.b = jenkins_x.c + rot(jenkins_x.d, 37);
+   jenkins_x.c = jenkins_x.d + e;
+   jenkins_x.d = e + jenkins_x.a;
+   return jenkins_x.d;
+}
+
+void s_mp_rand_jenkins_init(uint64_t seed)
+{
+   uint64_t i;
+   jenkins_x.a = 0xf1ea5eedULL;
+   jenkins_x.b = jenkins_x.c = jenkins_x.d = seed;
+   for (i = 0uLL; i < 20uLL; ++i) {
+      (void)s_rand_jenkins_val();
+   }
+}
+
+mp_err s_mp_rand_jenkins(void *p, size_t n)
+{
+   char *q = (char *)p;
+   while (n > 0u) {
+      int i;
+      uint64_t x = s_rand_jenkins_val();
+      for (i = 0; (i < 8) && (n > 0u); ++i, --n) {
+         *q++ = (char)(x & 0xFFuLL);
+         x >>= 8;
+      }
+   }
+   return MP_OKAY;
+}
+
+#endif
diff --git a/libtommath/bn_s_mp_rand_platform.c b/libtommath/bn_s_mp_rand_platform.c
new file mode 100644
index 0000000..27339bf
--- /dev/null
+++ b/libtommath/bn_s_mp_rand_platform.c
@@ -0,0 +1,148 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_RAND_PLATFORM_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* First the OS-specific special cases
+ * - *BSD
+ * - Windows
+ */
+#if defined(__FreeBSD__) || defined(__OpenBSD__) || defined(__NetBSD__) || defined(__DragonFly__)
+#define BN_S_READ_ARC4RANDOM_C
+static mp_err s_read_arc4random(void *p, size_t n)
+{
+   arc4random_buf(p, n);
+   return MP_OKAY;
+}
+#endif
+
+#if defined(_WIN32) || defined(_WIN32_WCE)
+#define BN_S_READ_WINCSP_C
+
+#ifndef _WIN32_WINNT
+#define _WIN32_WINNT 0x0400
+#endif
+#ifdef _WIN32_WCE
+#define UNDER_CE
+#define ARM
+#endif
+
+#define WIN32_LEAN_AND_MEAN
+#include <windows.h>
+#include <wincrypt.h>
+
+static mp_err s_read_wincsp(void *p, size_t n)
+{
+   static HCRYPTPROV hProv = 0;
+   if (hProv == 0) {
+      HCRYPTPROV h = 0;
+      if (!CryptAcquireContext(&h, NULL, MS_DEF_PROV, PROV_RSA_FULL,
+                               (CRYPT_VERIFYCONTEXT | CRYPT_MACHINE_KEYSET)) &&
+          !CryptAcquireContext(&h, NULL, MS_DEF_PROV, PROV_RSA_FULL,
+                               CRYPT_VERIFYCONTEXT | CRYPT_MACHINE_KEYSET | CRYPT_NEWKEYSET)) {
+         return MP_ERR;
+      }
+      hProv = h;
+   }
+   return CryptGenRandom(hProv, (DWORD)n, (BYTE *)p) == TRUE ? MP_OKAY : MP_ERR;
+}
+#endif /* WIN32 */
+
+#if !defined(BN_S_READ_WINCSP_C) && defined(__linux__) && defined(__GLIBC_PREREQ)
+#if __GLIBC_PREREQ(2, 25)
+#define BN_S_READ_GETRANDOM_C
+#include <sys/random.h>
+#include <errno.h>
+
+static mp_err s_read_getrandom(void *p, size_t n)
+{
+   char *q = (char *)p;
+   while (n > 0u) {
+      ssize_t ret = getrandom(q, n, 0);
+      if (ret < 0) {
+         if (errno == EINTR) {
+            continue;
+         }
+         return MP_ERR;
+      }
+      q += ret;
+      n -= (size_t)ret;
+   }
+   return MP_OKAY;
+}
+#endif
+#endif
+
+/* We assume all platforms besides windows provide "/dev/urandom".
+ * In case yours doesn't, define MP_NO_DEV_URANDOM at compile-time.
+ */
+#if !defined(BN_S_READ_WINCSP_C) && !defined(MP_NO_DEV_URANDOM)
+#define BN_S_READ_URANDOM_C
+#ifndef MP_DEV_URANDOM
+#define MP_DEV_URANDOM "/dev/urandom"
+#endif
+#include <fcntl.h>
+#include <errno.h>
+#include <unistd.h>
+
+static mp_err s_read_urandom(void *p, size_t n)
+{
+   int fd;
+   char *q = (char *)p;
+
+   do {
+      fd = open(MP_DEV_URANDOM, O_RDONLY);
+   } while ((fd == -1) && (errno == EINTR));
+   if (fd == -1) return MP_ERR;
+
+   while (n > 0u) {
+      ssize_t ret = read(fd, p, n);
+      if (ret < 0) {
+         if (errno == EINTR) {
+            continue;
+         }
+         close(fd);
+         return MP_ERR;
+      }
+      q += ret;
+      n -= (size_t)ret;
+   }
+
+   close(fd);
+   return MP_OKAY;
+}
+#endif
+
+#if defined(MP_PRNG_ENABLE_LTM_RNG)
+#define BN_S_READ_LTM_RNG
+unsigned long (*ltm_rng)(unsigned char *out, unsigned long outlen, void (*callback)(void));
+void (*ltm_rng_callback)(void);
+
+static mp_err s_read_ltm_rng(void *p, size_t n)
+{
+   unsigned long res;
+   if (ltm_rng == NULL) return MP_ERR;
+   res = ltm_rng(p, n, ltm_rng_callback);
+   if (res != n) return MP_ERR;
+   return MP_OKAY;
+}
+#endif
+
+mp_err s_read_arc4random(void *p, size_t n);
+mp_err s_read_wincsp(void *p, size_t n);
+mp_err s_read_getrandom(void *p, size_t n);
+mp_err s_read_urandom(void *p, size_t n);
+mp_err s_read_ltm_rng(void *p, size_t n);
+
+mp_err s_mp_rand_platform(void *p, size_t n)
+{
+   mp_err err = MP_ERR;
+   if ((err != MP_OKAY) && MP_HAS(S_READ_ARC4RANDOM)) err = s_read_arc4random(p, n);
+   if ((err != MP_OKAY) && MP_HAS(S_READ_WINCSP))     err = s_read_wincsp(p, n);
+   if ((err != MP_OKAY) && MP_HAS(S_READ_GETRANDOM))  err = s_read_getrandom(p, n);
+   if ((err != MP_OKAY) && MP_HAS(S_READ_URANDOM))    err = s_read_urandom(p, n);
+   if ((err != MP_OKAY) && MP_HAS(S_READ_LTM_RNG))    err = s_read_ltm_rng(p, n);
+   return err;
+}
+
+#endif
diff --git a/libtommath/bn_s_mp_reverse.c b/libtommath/bn_s_mp_reverse.c
new file mode 100644
index 0000000..c549e60
--- /dev/null
+++ b/libtommath/bn_s_mp_reverse.c
@@ -0,0 +1,22 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_REVERSE_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* reverse an array, used for radix code */
+void s_mp_reverse(unsigned char *s, size_t len)
+{
+   size_t   ix, iy;
+   unsigned char t;
+
+   ix = 0u;
+   iy = len - 1u;
+   while (ix < iy) {
+      t     = s[ix];
+      s[ix] = s[iy];
+      s[iy] = t;
+      ++ix;
+      --iy;
+   }
+}
+#endif
diff --git a/libtommath/bn_s_mp_sqr.c b/libtommath/bn_s_mp_sqr.c
index b3d0fd0..505c9f0 100644
--- a/libtommath/bn_s_mp_sqr.c
+++ b/libtommath/bn_s_mp_sqr.c
@@ -1,28 +1,20 @@
 #include "tommath_private.h"
 #ifdef BN_S_MP_SQR_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* low level squaring, b = a*a, HAC pp.596-597, Algorithm 14.16 */
-int s_mp_sqr(const mp_int *a, mp_int *b)
+mp_err s_mp_sqr(const mp_int *a, mp_int *b)
 {
-   mp_int  t;
-   int     res, ix, iy, pa;
-   mp_word r;
+   mp_int   t;
+   int      ix, iy, pa;
+   mp_err   err;
+   mp_word  r;
    mp_digit u, tmpx, *tmpt;
 
    pa = a->used;
-   if ((res = mp_init_size(&t, (2 * pa) + 1)) != MP_OKAY) {
-      return res;
+   if ((err = mp_init_size(&t, (2 * pa) + 1)) != MP_OKAY) {
+      return err;
    }
 
    /* default used is maximum possible size */
@@ -38,7 +30,7 @@ int s_mp_sqr(const mp_int *a, mp_int *b)
       t.dp[ix+ix] = (mp_digit)(r & (mp_word)MP_MASK);
 
       /* get the carry */
-      u           = (mp_digit)(r >> (mp_word)DIGIT_BIT);
+      u           = (mp_digit)(r >> (mp_word)MP_DIGIT_BIT);
 
       /* left hand side of A[ix] * A[iy] */
       tmpx        = a->dp[ix];
@@ -59,13 +51,13 @@ int s_mp_sqr(const mp_int *a, mp_int *b)
          *tmpt++ = (mp_digit)(r & (mp_word)MP_MASK);
 
          /* get carry */
-         u       = (mp_digit)(r >> (mp_word)DIGIT_BIT);
+         u       = (mp_digit)(r >> (mp_word)MP_DIGIT_BIT);
       }
       /* propagate upwards */
       while (u != 0uL) {
          r       = (mp_word)*tmpt + (mp_word)u;
          *tmpt++ = (mp_digit)(r & (mp_word)MP_MASK);
-         u       = (mp_digit)(r >> (mp_word)DIGIT_BIT);
+         u       = (mp_digit)(r >> (mp_word)MP_DIGIT_BIT);
       }
    }
 
@@ -75,7 +67,3 @@ int s_mp_sqr(const mp_int *a, mp_int *b)
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_fast_s_mp_sqr.c b/libtommath/bn_s_mp_sqr_fast.c
index 5be8e9d..d5090fe 100644
--- a/libtommath/bn_fast_s_mp_sqr.c
+++ b/libtommath/bn_s_mp_sqr_fast.c
@@ -1,16 +1,7 @@
 #include "tommath_private.h"
-#ifdef BN_FAST_S_MP_SQR_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+#ifdef BN_S_MP_SQR_FAST_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* the jist of squaring...
  * you do like mult except the offset of the tmpx [one that
@@ -22,17 +13,18 @@
 After that loop you do the squares and add them in.
 */
 
-int fast_s_mp_sqr(const mp_int *a, mp_int *b)
+mp_err s_mp_sqr_fast(const mp_int *a, mp_int *b)
 {
-   int       olduse, res, pa, ix, iz;
-   mp_digit   W[MP_WARRAY], *tmpx;
+   int       olduse, pa, ix, iz;
+   mp_digit  W[PRIVATE_MP_WARRAY], *tmpx;
    mp_word   W1;
+   mp_err    err;
 
    /* grow the destination as required */
    pa = a->used + a->used;
    if (b->alloc < pa) {
-      if ((res = mp_grow(b, pa)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(b, pa)) != MP_OKAY) {
+         return err;
       }
    }
 
@@ -47,7 +39,7 @@ int fast_s_mp_sqr(const mp_int *a, mp_int *b)
       _W = 0;
 
       /* get offsets into the two bignums */
-      ty = MIN(a->used-1, ix);
+      ty = MP_MIN(a->used-1, ix);
       tx = ix - ty;
 
       /* setup temp aliases */
@@ -57,13 +49,13 @@ int fast_s_mp_sqr(const mp_int *a, mp_int *b)
       /* this is the number of times the loop will iterrate, essentially
          while (tx++ < a->used && ty-- >= 0) { ... }
        */
-      iy = MIN(a->used-tx, ty+1);
+      iy = MP_MIN(a->used-tx, ty+1);
 
       /* now for squaring tx can never equal ty
        * we halve the distance since they approach at a rate of 2x
        * and we have to round because odd cases need to be executed
        */
-      iy = MIN(iy, ((ty-tx)+1)>>1);
+      iy = MP_MIN(iy, ((ty-tx)+1)>>1);
 
       /* execute loop */
       for (iz = 0; iz < iy; iz++) {
@@ -79,10 +71,10 @@ int fast_s_mp_sqr(const mp_int *a, mp_int *b)
       }
 
       /* store it */
-      W[ix] = _W & MP_MASK;
+      W[ix] = (mp_digit)_W & MP_MASK;
 
       /* make next carry */
-      W1 = _W >> (mp_word)DIGIT_BIT;
+      W1 = _W >> (mp_word)MP_DIGIT_BIT;
    }
 
    /* setup dest */
@@ -97,15 +89,9 @@ int fast_s_mp_sqr(const mp_int *a, mp_int *b)
       }
 
       /* clear unused digits [that existed in the old copy of c] */
-      for (; ix < olduse; ix++) {
-         *tmpb++ = 0;
-      }
+      MP_ZERO_DIGITS(tmpb, olduse - ix);
    }
    mp_clamp(b);
    return MP_OKAY;
 }
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_sub.c b/libtommath/bn_s_mp_sub.c
index 88e44dc..5672dab 100644
--- a/libtommath/bn_s_mp_sub.c
+++ b/libtommath/bn_s_mp_sub.c
@@ -1,21 +1,13 @@
 #include "tommath_private.h"
 #ifdef BN_S_MP_SUB_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* low level subtraction (assumes |a| > |b|), HAC pp.595 Algorithm 14.9 */
-int s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
+mp_err s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
 {
-   int     olduse, res, min, max;
+   int    olduse, min, max;
+   mp_err err;
 
    /* find sizes */
    min = b->used;
@@ -23,8 +15,8 @@ int s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
 
    /* init result */
    if (c->alloc < max) {
-      if ((res = mp_grow(c, max)) != MP_OKAY) {
-         return res;
+      if ((err = mp_grow(c, max)) != MP_OKAY) {
+         return err;
       }
    }
    olduse = c->used;
@@ -50,7 +42,7 @@ int s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
           * if a carry does occur it will propagate all the way to the
           * MSB.  As a result a single shift is enough to get the carry
           */
-         u = *tmpc >> (((size_t)CHAR_BIT * sizeof(mp_digit)) - 1u);
+         u = *tmpc >> (MP_SIZEOF_BITS(mp_digit) - 1u);
 
          /* Clear carry from T[i] */
          *tmpc++ &= MP_MASK;
@@ -62,16 +54,14 @@ int s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
          *tmpc = *tmpa++ - u;
 
          /* U = carry bit of T[i] */
-         u = *tmpc >> (((size_t)CHAR_BIT * sizeof(mp_digit)) - 1u);
+         u = *tmpc >> (MP_SIZEOF_BITS(mp_digit) - 1u);
 
          /* Clear carry from T[i] */
          *tmpc++ &= MP_MASK;
       }
 
       /* clear digits above used (since we may not have grown result above) */
-      for (i = c->used; i < olduse; i++) {
-         *tmpc++ = 0;
-      }
+      MP_ZERO_DIGITS(tmpc, olduse - c->used);
    }
 
    mp_clamp(c);
@@ -79,7 +69,3 @@ int s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c)
 }
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/bn_s_mp_toom_mul.c b/libtommath/bn_s_mp_toom_mul.c
new file mode 100644
index 0000000..86901b0
--- /dev/null
+++ b/libtommath/bn_s_mp_toom_mul.c
@@ -0,0 +1,215 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_TOOM_MUL_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* multiplication using the Toom-Cook 3-way algorithm
+ *
+ * Much more complicated than Karatsuba but has a lower
+ * asymptotic running time of O(N**1.464).  This algorithm is
+ * only particularly useful on VERY large inputs
+ * (we're talking 1000s of digits here...).
+*/
+
+/*
+   This file contains code from J. Arndt's book  "Matters Computational"
+   and the accompanying FXT-library with permission of the author.
+*/
+
+/*
+   Setup from
+
+     Chung, Jaewook, and M. Anwar Hasan. "Asymmetric squaring formulae."
+     18th IEEE Symposium on Computer Arithmetic (ARITH'07). IEEE, 2007.
+
+   The interpolation from above needed one temporary variable more
+   than the interpolation here:
+
+     Bodrato, Marco, and Alberto Zanoni. "What about Toom-Cook matrices optimality."
+     Centro Vito Volterra Universita di Roma Tor Vergata (2006)
+*/
+
+mp_err s_mp_toom_mul(const mp_int *a, const mp_int *b, mp_int *c)
+{
+   mp_int S1, S2, T1, a0, a1, a2, b0, b1, b2;
+   int B, count;
+   mp_err err;
+
+   /* init temps */
+   if ((err = mp_init_multi(&S1, &S2, &T1, NULL)) != MP_OKAY) {
+      return err;
+   }
+
+   /* B */
+   B = MP_MIN(a->used, b->used) / 3;
+
+   /** a = a2 * x^2 + a1 * x + a0; */
+   if ((err = mp_init_size(&a0, B)) != MP_OKAY)                   goto LBL_ERRa0;
+
+   for (count = 0; count < B; count++) {
+      a0.dp[count] = a->dp[count];
+      a0.used++;
+   }
+   mp_clamp(&a0);
+   if ((err = mp_init_size(&a1, B)) != MP_OKAY)                   goto LBL_ERRa1;
+   for (; count < (2 * B); count++) {
+      a1.dp[count - B] = a->dp[count];
+      a1.used++;
+   }
+   mp_clamp(&a1);
+   if ((err = mp_init_size(&a2, B + (a->used - (3 * B)))) != MP_OKAY) goto LBL_ERRa2;
+   for (; count < a->used; count++) {
+      a2.dp[count - (2 * B)] = a->dp[count];
+      a2.used++;
+   }
+   mp_clamp(&a2);
+
+   /** b = b2 * x^2 + b1 * x + b0; */
+   if ((err = mp_init_size(&b0, B)) != MP_OKAY)                   goto LBL_ERRb0;
+   for (count = 0; count < B; count++) {
+      b0.dp[count] = b->dp[count];
+      b0.used++;
+   }
+   mp_clamp(&b0);
+   if ((err = mp_init_size(&b1, B)) != MP_OKAY)                   goto LBL_ERRb1;
+   for (; count < (2 * B); count++) {
+      b1.dp[count - B] = b->dp[count];
+      b1.used++;
+   }
+   mp_clamp(&b1);
+   if ((err = mp_init_size(&b2, B + (b->used - (3 * B)))) != MP_OKAY) goto LBL_ERRb2;
+   for (; count < b->used; count++) {
+      b2.dp[count - (2 * B)] = b->dp[count];
+      b2.used++;
+   }
+   mp_clamp(&b2);
+
+   /** \\ S1 = (a2+a1+a0) * (b2+b1+b0); */
+   /** T1 = a2 + a1; */
+   if ((err = mp_add(&a2, &a1, &T1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** S2 = T1 + a0; */
+   if ((err = mp_add(&T1, &a0, &S2)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** c = b2 + b1; */
+   if ((err = mp_add(&b2, &b1, c)) != MP_OKAY)                    goto LBL_ERR;
+
+   /** S1 = c + b0; */
+   if ((err = mp_add(c, &b0, &S1)) != MP_OKAY)                    goto LBL_ERR;
+
+   /** S1 = S1 * S2; */
+   if ((err = mp_mul(&S1, &S2, &S1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** \\S2 = (4*a2+2*a1+a0) * (4*b2+2*b1+b0); */
+   /** T1 = T1 + a2; */
+   if ((err = mp_add(&T1, &a2, &T1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** T1 = T1 << 1; */
+   if ((err = mp_mul_2(&T1, &T1)) != MP_OKAY)                     goto LBL_ERR;
+
+   /** T1 = T1 + a0; */
+   if ((err = mp_add(&T1, &a0, &T1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** c = c + b2; */
+   if ((err = mp_add(c, &b2, c)) != MP_OKAY)                      goto LBL_ERR;
+
+   /** c = c << 1; */
+   if ((err = mp_mul_2(c, c)) != MP_OKAY)                         goto LBL_ERR;
+
+   /** c = c + b0; */
+   if ((err = mp_add(c, &b0, c)) != MP_OKAY)                      goto LBL_ERR;
+
+   /** S2 = T1 * c; */
+   if ((err = mp_mul(&T1, c, &S2)) != MP_OKAY)                    goto LBL_ERR;
+
+   /** \\S3 = (a2-a1+a0) * (b2-b1+b0); */
+   /** a1 = a2 - a1; */
+   if ((err = mp_sub(&a2, &a1, &a1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** a1 = a1 + a0; */
+   if ((err = mp_add(&a1, &a0, &a1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** b1 = b2 - b1; */
+   if ((err = mp_sub(&b2, &b1, &b1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** b1 = b1 + b0; */
+   if ((err = mp_add(&b1, &b0, &b1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** a1 = a1 * b1; */
+   if ((err = mp_mul(&a1, &b1, &a1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** b1 = a2 * b2; */
+   if ((err = mp_mul(&a2, &b2, &b1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** \\S2 = (S2 - S3)/3; */
+   /** S2 = S2 - a1; */
+   if ((err = mp_sub(&S2, &a1, &S2)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** S2 = S2 / 3; \\ this is an exact division  */
+   if ((err = mp_div_3(&S2, &S2, NULL)) != MP_OKAY)               goto LBL_ERR;
+
+   /** a1 = S1 - a1; */
+   if ((err = mp_sub(&S1, &a1, &a1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** a1 = a1 >> 1; */
+   if ((err = mp_div_2(&a1, &a1)) != MP_OKAY)                     goto LBL_ERR;
+
+   /** a0 = a0 * b0; */
+   if ((err = mp_mul(&a0, &b0, &a0)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** S1 = S1 - a0; */
+   if ((err = mp_sub(&S1, &a0, &S1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** S2 = S2 - S1; */
+   if ((err = mp_sub(&S2, &S1, &S2)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** S2 = S2 >> 1; */
+   if ((err = mp_div_2(&S2, &S2)) != MP_OKAY)                     goto LBL_ERR;
+
+   /** S1 = S1 - a1; */
+   if ((err = mp_sub(&S1, &a1, &S1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** S1 = S1 - b1; */
+   if ((err = mp_sub(&S1, &b1, &S1)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** T1 = b1 << 1; */
+   if ((err = mp_mul_2(&b1, &T1)) != MP_OKAY)                     goto LBL_ERR;
+
+   /** S2 = S2 - T1; */
+   if ((err = mp_sub(&S2, &T1, &S2)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** a1 = a1 - S2; */
+   if ((err = mp_sub(&a1, &S2, &a1)) != MP_OKAY)                  goto LBL_ERR;
+
+
+   /** P = b1*x^4+ S2*x^3+ S1*x^2+ a1*x + a0; */
+   if ((err = mp_lshd(&b1, 4 * B)) != MP_OKAY)                    goto LBL_ERR;
+   if ((err = mp_lshd(&S2, 3 * B)) != MP_OKAY)                    goto LBL_ERR;
+   if ((err = mp_add(&b1, &S2, &b1)) != MP_OKAY)                  goto LBL_ERR;
+   if ((err = mp_lshd(&S1, 2 * B)) != MP_OKAY)                    goto LBL_ERR;
+   if ((err = mp_add(&b1, &S1, &b1)) != MP_OKAY)                  goto LBL_ERR;
+   if ((err = mp_lshd(&a1, 1 * B)) != MP_OKAY)                    goto LBL_ERR;
+   if ((err = mp_add(&b1, &a1, &b1)) != MP_OKAY)                  goto LBL_ERR;
+   if ((err = mp_add(&b1, &a0, c)) != MP_OKAY)                    goto LBL_ERR;
+
+   /** a * b - P */
+
+
+LBL_ERR:
+   mp_clear(&b2);
+LBL_ERRb2:
+   mp_clear(&b1);
+LBL_ERRb1:
+   mp_clear(&b0);
+LBL_ERRb0:
+   mp_clear(&a2);
+LBL_ERRa2:
+   mp_clear(&a1);
+LBL_ERRa1:
+   mp_clear(&a0);
+LBL_ERRa0:
+   mp_clear_multi(&S1, &S2, &T1, NULL);
+   return err;
+}
+
+#endif
diff --git a/libtommath/bn_s_mp_toom_sqr.c b/libtommath/bn_s_mp_toom_sqr.c
new file mode 100644
index 0000000..f2ffb30
--- /dev/null
+++ b/libtommath/bn_s_mp_toom_sqr.c
@@ -0,0 +1,147 @@
+#include "tommath_private.h"
+#ifdef BN_S_MP_TOOM_SQR_C
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+/* squaring using Toom-Cook 3-way algorithm */
+
+/*
+   This file contains code from J. Arndt's book  "Matters Computational"
+   and the accompanying FXT-library with permission of the author.
+*/
+
+/* squaring using Toom-Cook 3-way algorithm */
+/*
+   Setup and interpolation from algorithm SQR_3 in
+
+     Chung, Jaewook, and M. Anwar Hasan. "Asymmetric squaring formulae."
+     18th IEEE Symposium on Computer Arithmetic (ARITH'07). IEEE, 2007.
+
+*/
+mp_err s_mp_toom_sqr(const mp_int *a, mp_int *b)
+{
+   mp_int S0, a0, a1, a2;
+   mp_digit *tmpa, *tmpc;
+   int B, count;
+   mp_err err;
+
+
+   /* init temps */
+   if ((err = mp_init(&S0)) != MP_OKAY) {
+      return err;
+   }
+
+   /* B */
+   B = a->used / 3;
+
+   /** a = a2 * x^2 + a1 * x + a0; */
+   if ((err = mp_init_size(&a0, B)) != MP_OKAY)                   goto LBL_ERRa0;
+
+   a0.used = B;
+   if ((err = mp_init_size(&a1, B)) != MP_OKAY)                   goto LBL_ERRa1;
+   a1.used = B;
+   if ((err = mp_init_size(&a2, B + (a->used - (3 * B)))) != MP_OKAY) goto LBL_ERRa2;
+
+   tmpa = a->dp;
+   tmpc = a0.dp;
+   for (count = 0; count < B; count++) {
+      *tmpc++ = *tmpa++;
+   }
+   tmpc = a1.dp;
+   for (; count < (2 * B); count++) {
+      *tmpc++ = *tmpa++;
+   }
+   tmpc = a2.dp;
+   for (; count < a->used; count++) {
+      *tmpc++ = *tmpa++;
+      a2.used++;
+   }
+   mp_clamp(&a0);
+   mp_clamp(&a1);
+   mp_clamp(&a2);
+
+   /** S0 = a0^2;  */
+   if ((err = mp_sqr(&a0, &S0)) != MP_OKAY)                       goto LBL_ERR;
+
+   /** \\S1 = (a2 + a1 + a0)^2 */
+   /** \\S2 = (a2 - a1 + a0)^2  */
+   /** \\S1 = a0 + a2; */
+   /** a0 = a0 + a2; */
+   if ((err = mp_add(&a0, &a2, &a0)) != MP_OKAY)                  goto LBL_ERR;
+   /** \\S2 = S1 - a1; */
+   /** b = a0 - a1; */
+   if ((err = mp_sub(&a0, &a1, b)) != MP_OKAY)                    goto LBL_ERR;
+   /** \\S1 = S1 + a1; */
+   /** a0 = a0 + a1; */
+   if ((err = mp_add(&a0, &a1, &a0)) != MP_OKAY)                  goto LBL_ERR;
+   /** \\S1 = S1^2;  */
+   /** a0 = a0^2; */
+   if ((err = mp_sqr(&a0, &a0)) != MP_OKAY)                       goto LBL_ERR;
+   /** \\S2 = S2^2;  */
+   /** b = b^2; */
+   if ((err = mp_sqr(b, b)) != MP_OKAY)                           goto LBL_ERR;
+
+   /** \\ S3 = 2 * a1 * a2  */
+   /** \\S3 = a1 * a2;  */
+   /** a1 = a1 * a2; */
+   if ((err = mp_mul(&a1, &a2, &a1)) != MP_OKAY)                  goto LBL_ERR;
+   /** \\S3 = S3 << 1;  */
+   /** a1 = a1 << 1; */
+   if ((err = mp_mul_2(&a1, &a1)) != MP_OKAY)                     goto LBL_ERR;
+
+   /** \\S4 = a2^2;  */
+   /** a2 = a2^2; */
+   if ((err = mp_sqr(&a2, &a2)) != MP_OKAY)                       goto LBL_ERR;
+
+   /** \\ tmp = (S1 + S2)/2  */
+   /** \\tmp = S1 + S2; */
+   /** b = a0 + b; */
+   if ((err = mp_add(&a0, b, b)) != MP_OKAY)                      goto LBL_ERR;
+   /** \\tmp = tmp >> 1; */
+   /** b = b >> 1; */
+   if ((err = mp_div_2(b, b)) != MP_OKAY)                         goto LBL_ERR;
+
+   /** \\ S1 = S1 - tmp - S3  */
+   /** \\S1 = S1 - tmp; */
+   /** a0 = a0 - b; */
+   if ((err = mp_sub(&a0, b, &a0)) != MP_OKAY)                    goto LBL_ERR;
+   /** \\S1 = S1 - S3;  */
+   /** a0 = a0 - a1; */
+   if ((err = mp_sub(&a0, &a1, &a0)) != MP_OKAY)                  goto LBL_ERR;
+
+   /** \\S2 = tmp - S4 -S0  */
+   /** \\S2 = tmp - S4;  */
+   /** b = b - a2; */
+   if ((err = mp_sub(b, &a2, b)) != MP_OKAY)                      goto LBL_ERR;
+   /** \\S2 = S2 - S0;  */
+   /** b = b - S0; */
+   if ((err = mp_sub(b, &S0, b)) != MP_OKAY)                      goto LBL_ERR;
+
+
+   /** \\P = S4*x^4 + S3*x^3 + S2*x^2 + S1*x + S0; */
+   /** P = a2*x^4 + a1*x^3 + b*x^2 + a0*x + S0; */
+
+   if ((err = mp_lshd(&a2, 4 * B)) != MP_OKAY)                    goto LBL_ERR;
+   if ((err = mp_lshd(&a1, 3 * B)) != MP_OKAY)                    goto LBL_ERR;
+   if ((err = mp_lshd(b, 2 * B)) != MP_OKAY)                      goto LBL_ERR;
+   if ((err = mp_lshd(&a0, 1 * B)) != MP_OKAY)                    goto LBL_ERR;
+   if ((err = mp_add(&a2, &a1, &a2)) != MP_OKAY)                  goto LBL_ERR;
+   if ((err = mp_add(&a2, b, b)) != MP_OKAY)                      goto LBL_ERR;
+   if ((err = mp_add(b, &a0, b)) != MP_OKAY)                      goto LBL_ERR;
+   if ((err = mp_add(b, &S0, b)) != MP_OKAY)                      goto LBL_ERR;
+   /** a^2 - P  */
+
+
+LBL_ERR:
+   mp_clear(&a2);
+LBL_ERRa2:
+   mp_clear(&a1);
+LBL_ERRa1:
+   mp_clear(&a0);
+LBL_ERRa0:
+   mp_clear(&S0);
+
+   return err;
+}
+
+#endif
diff --git a/libtommath/bncore.c b/libtommath/bncore.c
deleted file mode 100644
index c97b8e1..0000000
--- a/libtommath/bncore.c
+++ /dev/null
@@ -1,33 +0,0 @@
-#include "tommath_private.h"
-#ifdef BNCORE_C
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
-
-/* Known optimal configurations
-
- CPU                    /Compiler     /MUL CUTOFF/SQR CUTOFF
--------------------------------------------------------------
- Intel P4 Northwood     /GCC v3.4.1   /        88/       128/LTM 0.32 ;-)
- AMD Athlon64           /GCC v3.4.4   /        80/       120/LTM 0.35
-
-*/
-
-int     KARATSUBA_MUL_CUTOFF = 80,      /* Min. number of digits before Karatsuba multiplication is used. */
-        KARATSUBA_SQR_CUTOFF = 120,     /* Min. number of digits before Karatsuba squaring is used. */
-
-        TOOM_MUL_CUTOFF      = 350,      /* no optimal values of these are known yet so set em high */
-        TOOM_SQR_CUTOFF      = 400;
-#endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/changes.txt b/libtommath/changes.txt
index aa0c64c..ebf7382 100644
--- a/libtommath/changes.txt
+++ b/libtommath/changes.txt
@@ -1,3 +1,40 @@
+XXX XXth, 2019
+v1.2.0
+       -- A huge refactoring of the library happened - renaming,
+          deprecating and replacing existing functions by improved API's.
+
+          All deprecated functions, macros and symbols are only marked as such
+          so this version is still API and ABI compatible to v1.x.
+
+       -- Daniel Mendler was pushing for those changes and contributing a load of patches,
+          refactorings, code reviews and whatnotelse.
+       -- Christoph Zurnieden re-worked internals of the library, improved the performance,
+          did code reviews and wrote documentation.
+       -- Francois Perrad did some refactoring and took again care of linting the sources and
+          provided all fixes.
+       -- Jan Nijtmans, Karel Miko and Joachim Breitner contributed various patches.
+
+       -- Private symbols can now be hidden for the shared library builds, disabled by default.
+       -- All API's follow a single code style, are prefixed the same etc.
+       -- Unified, safer and improved API's
+       -- Less magic numbers - return values (where appropriate) and most flags are now enums,
+          this was implemented in a backwards compatible way where return values were int.
+       -- API's with return values are now by default marked as "warn on unsused result", this
+          can be disabled if required (which will most likely hide bugs), c.f. MP_WUR in tommath.h
+       -- Provide a whole set of setters&getters for different primitive types (long, uint32_t, etc.)
+       -- All those primitive setters are now optimized.
+       -- It's possible to automatically tune the cutoff values for Karatsuba&Toom-Cook
+       -- The custom allocators which were formerly known as XMALLOC(), XFREE() etc. are now available
+          as MP_MALLOC(), MP_REALLOC(), MP_CALLOC() and MP_FREE(). MP_REALLOC() and MP_FREE() now also
+          provide the allocated size to ease the usage of simple allocators without tracking.
+       -- Building is now also possible with MSVC 2015, 2017 and 2019 (use makefile.msvc)
+       -- Added mp_decr() and mp_incr()
+       -- Added mp_log_u32()
+       -- Improved prime-checking
+       -- Improved Toom-Cook multiplication
+       -- Removed the LTM book (`make docs` now builds the user manual)
+
+
 Jan 28th, 2019
 v1.1.0
        -- Christoph Zurnieden contributed FIPS 186.4 compliant
diff --git a/libtommath/helper.pl b/libtommath/helper.pl
new file mode 100755
index 0000000..e60c1a7
--- /dev/null
+++ b/libtommath/helper.pl
@@ -0,0 +1,482 @@
+#!/usr/bin/env perl
+
+use strict;
+use warnings;
+
+use Getopt::Long;
+use File::Find 'find';
+use File::Basename 'basename';
+use File::Glob 'bsd_glob';
+
+sub read_file {
+  my $f = shift;
+  open my $fh, "<", $f or die "FATAL: read_rawfile() cannot open file '$f': $!";
+  binmode $fh;
+  return do { local $/; <$fh> };
+}
+
+sub write_file {
+  my ($f, $data) = @_;
+  die "FATAL: write_file() no data" unless defined $data;
+  open my $fh, ">", $f or die "FATAL: write_file() cannot open file '$f': $!";
+  binmode $fh;
+  print $fh $data or die "FATAL: write_file() cannot write to '$f': $!";
+  close $fh or die "FATAL: write_file() cannot close '$f': $!";
+  return;
+}
+
+sub sanitize_comments {
+  my($content) = @_;
+  $content =~ s{/\*(.*?)\*/}{my $x=$1; $x =~ s/\w/x/g; "/*$x*/";}egs;
+  return $content;
+}
+
+sub check_source {
+  my @all_files = (
+        bsd_glob("makefile*"),
+        bsd_glob("*.{h,c,sh,pl}"),
+        bsd_glob("*/*.{h,c,sh,pl}"),
+  );
+
+  my $fails = 0;
+  for my $file (sort @all_files) {
+    my $troubles = {};
+    my $lineno = 1;
+    my $content = read_file($file);
+    $content = sanitize_comments $content;
+    push @{$troubles->{crlf_line_end}}, '?' if $content =~ /\r/;
+    for my $l (split /\n/, $content) {
+      push @{$troubles->{merge_conflict}},     $lineno if $l =~ /^(<<<<<<<|=======|>>>>>>>)([^<=>]|$)/;
+      push @{$troubles->{trailing_space}},     $lineno if $l =~ / $/;
+      push @{$troubles->{tab}},                $lineno if $l =~ /\t/ && basename($file) !~ /^makefile/i;
+      push @{$troubles->{non_ascii_char}},     $lineno if $l =~ /[^[:ascii:]]/;
+      push @{$troubles->{cpp_comment}},        $lineno if $file =~ /\.(c|h)$/ && ($l =~ /\s\/\// || $l =~ /\/\/\s/);
+      # we prefer using XMALLOC, XFREE, XREALLOC, XCALLOC ...
+      push @{$troubles->{unwanted_malloc}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bmalloc\s*\(/;
+      push @{$troubles->{unwanted_realloc}},   $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\brealloc\s*\(/;
+      push @{$troubles->{unwanted_calloc}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bcalloc\s*\(/;
+      push @{$troubles->{unwanted_free}},      $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bfree\s*\(/;
+      # and we probably want to also avoid the following
+      push @{$troubles->{unwanted_memcpy}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bmemcpy\s*\(/;
+      push @{$troubles->{unwanted_memset}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bmemset\s*\(/;
+      push @{$troubles->{unwanted_memcpy}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bmemcpy\s*\(/;
+      push @{$troubles->{unwanted_memmove}},   $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bmemmove\s*\(/;
+      push @{$troubles->{unwanted_memcmp}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bmemcmp\s*\(/;
+      push @{$troubles->{unwanted_strcmp}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bstrcmp\s*\(/;
+      push @{$troubles->{unwanted_strcpy}},    $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bstrcpy\s*\(/;
+      push @{$troubles->{unwanted_strncpy}},   $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bstrncpy\s*\(/;
+      push @{$troubles->{unwanted_clock}},     $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bclock\s*\(/;
+      push @{$troubles->{unwanted_qsort}},     $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bqsort\s*\(/;
+      push @{$troubles->{sizeof_no_brackets}}, $lineno if $file =~ /^[^\/]+\.c$/ && $l =~ /\bsizeof\s*[^\(]/;
+      if ($file =~ m|^[^\/]+\.c$| && $l =~ /^static(\s+[a-zA-Z0-9_]+)+\s+([a-zA-Z0-9_]+)\s*\(/) {
+        my $funcname = $2;
+        # static functions should start with s_
+        push @{$troubles->{staticfunc_name}}, "$lineno($funcname)" if $funcname !~ /^s_/;
+      }
+      $lineno++;
+    }
+    for my $k (sort keys %$troubles) {
+      warn "[$k] $file line:" . join(",", @{$troubles->{$k}}) . "\n";
+      $fails++;
+    }
+  }
+
+  warn( $fails > 0 ? "check-source:    FAIL $fails\n" : "check-source:    PASS\n" );
+  return $fails;
+}
+
+sub check_comments {
+  my $fails = 0;
+  my $first_comment = <<'MARKER';
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+MARKER
+  #my @all_files = (bsd_glob("*.{h,c}"), bsd_glob("*/*.{h,c}"));
+  my @all_files = (bsd_glob("*.{h,c}"));
+  for my $f (@all_files) {
+    my $txt = read_file($f);
+    if ($txt !~ /\Q$first_comment\E/s) {
+      warn "[first_comment] $f\n";
+      $fails++;
+    }
+  }
+  warn( $fails > 0 ? "check-comments:  FAIL $fails\n" : "check-comments:  PASS\n" );
+  return $fails;
+}
+
+sub check_doc {
+  my $fails = 0;
+  my $tex = read_file('doc/bn.tex');
+  my $tmh = read_file('tommath.h');
+  my @functions = $tmh =~ /\n\s*[a-zA-Z0-9_* ]+?(mp_[a-z0-9_]+)\s*\([^\)]+\)\s*;/sg;
+  my @macros    = $tmh =~ /\n\s*#define\s+([a-z0-9_]+)\s*\([^\)]+\)/sg;
+  for my $n (sort @functions) {
+    (my $nn = $n) =~ s/_/\\_/g; # mp_sub_d >> mp\_sub\_d
+    if ($tex !~ /index\Q{$nn}\E/) {
+      warn "[missing_doc_for_function] $n\n";
+      $fails++
+    }
+  }
+  for my $n (sort @macros) {
+    (my $nn = $n) =~ s/_/\\_/g; # mp_iszero >> mp\_iszero
+    if ($tex !~ /index\Q{$nn}\E/) {
+      warn "[missing_doc_for_macro] $n\n";
+      $fails++
+    }
+  }
+  warn( $fails > 0 ? "check_doc:       FAIL $fails\n" : "check-doc:       PASS\n" );
+  return $fails;
+}
+
+sub prepare_variable {
+  my ($varname, @list) = @_;
+  my $output = "$varname=";
+  my $len = length($output);
+  foreach my $obj (sort @list) {
+    $len = $len + length $obj;
+    $obj =~ s/\*/\$/;
+    if ($len > 100) {
+      $output .= "\\\n";
+      $len = length $obj;
+    }
+    $output .= $obj . ' ';
+  }
+  $output =~ s/ $//;
+  return $output;
+}
+
+sub prepare_msvc_files_xml {
+  my ($all, $exclude_re, $targets) = @_;
+  my $last = [];
+  my $depth = 2;
+
+  # sort files in the same order as visual studio (ugly, I know)
+  my @parts = ();
+  for my $orig (@$all) {
+    my $p = $orig;
+    $p =~ s|/|/~|g;
+    $p =~ s|/~([^/]+)$|/$1|g;
+    my @l = map { sprintf "% -99s", $_ } split /\//, $p;
+    push @parts, [ $orig, join(':', @l) ];
+  }
+  my @sorted = map { $_->[0] } sort { $a->[1] cmp $b->[1] } @parts;
+
+  my $files = "<Files>\r\n";
+  for my $full (@sorted) {
+    my @items = split /\//, $full; # split by '/'
+    $full =~ s|/|\\|g;             # replace '/' bt '\'
+    shift @items; # drop first one (src)
+    pop @items;   # drop last one (filename.ext)
+    my $current = \@items;
+    if (join(':', @$current) ne join(':', @$last)) {
+      my $common = 0;
+      $common++ while ($last->[$common] && $current->[$common] && $last->[$common] eq $current->[$common]);
+      my $back = @$last - $common;
+      if ($back > 0) {
+        $files .= ("\t" x --$depth) . "</Filter>\r\n" for (1..$back);
+      }
+      my $fwd = [ @$current ]; splice(@$fwd, 0, $common);
+      for my $i (0..scalar(@$fwd) - 1) {
+        $files .= ("\t" x $depth) . "<Filter\r\n";
+        $files .= ("\t" x $depth) . "\tName=\"$fwd->[$i]\"\r\n";
+        $files .= ("\t" x $depth) . "\t>\r\n";
+        $depth++;
+      }
+      $last = $current;
+    }
+    $files .= ("\t" x $depth) . "<File\r\n";
+    $files .= ("\t" x $depth) . "\tRelativePath=\"$full\"\r\n";
+    $files .= ("\t" x $depth) . "\t>\r\n";
+    if ($full =~ $exclude_re) {
+      for (@$targets) {
+        $files .= ("\t" x $depth) . "\t<FileConfiguration\r\n";
+        $files .= ("\t" x $depth) . "\t\tName=\"$_\"\r\n";
+        $files .= ("\t" x $depth) . "\t\tExcludedFromBuild=\"true\"\r\n";
+        $files .= ("\t" x $depth) . "\t\t>\r\n";
+        $files .= ("\t" x $depth) . "\t\t<Tool\r\n";
+        $files .= ("\t" x $depth) . "\t\t\tName=\"VCCLCompilerTool\"\r\n";
+        $files .= ("\t" x $depth) . "\t\t\tAdditionalIncludeDirectories=\"\"\r\n";
+        $files .= ("\t" x $depth) . "\t\t\tPreprocessorDefinitions=\"\"\r\n";
+        $files .= ("\t" x $depth) . "\t\t/>\r\n";
+        $files .= ("\t" x $depth) . "\t</FileConfiguration>\r\n";
+      }
+    }
+    $files .= ("\t" x $depth) . "</File>\r\n";
+  }
+  $files .= ("\t" x --$depth) . "</Filter>\r\n" for (@$last);
+  $files .= "\t</Files>";
+  return $files;
+}
+
+sub patch_file {
+  my ($content, @variables) = @_;
+  for my $v (@variables) {
+    if ($v =~ /^([A-Z0-9_]+)\s*=.*$/si) {
+      my $name = $1;
+      $content =~ s/\n\Q$name\E\b.*?[^\\]\n/\n$v\n/s;
+    }
+    else {
+      die "patch_file failed: " . substr($v, 0, 30) . "..";
+    }
+  }
+  return $content;
+}
+
+sub process_makefiles {
+  my $write = shift;
+  my $changed_count = 0;
+  my @o = map { my $x = $_; $x =~ s/\.c$/.o/; $x } bsd_glob("*.c");
+  my @all = bsd_glob("*.{c,h}");
+
+  my $var_o = prepare_variable("OBJECTS", @o);
+  (my $var_obj = $var_o) =~ s/\.o\b/.obj/sg;
+
+  # update MSVC project files
+  my $msvc_files = prepare_msvc_files_xml(\@all, qr/NOT_USED_HERE/, ['Debug|Win32', 'Release|Win32', 'Debug|x64', 'Release|x64']);
+  for my $m (qw/libtommath_VS2008.vcproj/) {
+    my $old = read_file($m);
+    my $new = $old;
+    $new =~ s|<Files>.*</Files>|$msvc_files|s;
+    if ($old ne $new) {
+      write_file($m, $new) if $write;
+      warn "changed: $m\n";
+      $changed_count++;
+    }
+  }
+
+  # update OBJECTS + HEADERS in makefile*
+  for my $m (qw/ makefile makefile.shared makefile_include.mk makefile.msvc makefile.unix makefile.mingw /) {
+    my $old = read_file($m);
+    my $new = $m eq 'makefile.msvc' ? patch_file($old, $var_obj)
+                                    : patch_file($old, $var_o);
+    if ($old ne $new) {
+      write_file($m, $new) if $write;
+      warn "changed: $m\n";
+      $changed_count++;
+    }
+  }
+
+  if ($write) {
+    return 0; # no failures
+  }
+  else {
+    warn( $changed_count > 0 ? "check-makefiles: FAIL $changed_count\n" : "check-makefiles: PASS\n" );
+    return $changed_count;
+  }
+}
+
+sub draw_func
+{
+   my ($deplist, $depmap, $out, $indent, $funcslist) = @_;
+   my @funcs = split ',', $funcslist;
+   # try this if you want to have a look at a minimized version of the callgraph without all the trivial functions
+   #if ($deplist =~ /$funcs[0]/ || $funcs[0] =~ /BN_MP_(ADD|SUB|CLEAR|CLEAR_\S+|DIV|MUL|COPY|ZERO|GROW|CLAMP|INIT|INIT_\S+|SET|ABS|CMP|CMP_D|EXCH)_C/) {
+   if ($deplist =~ /$funcs[0]/) {
+      return $deplist;
+   } else {
+      $deplist = $deplist . $funcs[0];
+   }
+   if ($indent == 0) {
+   } elsif ($indent >= 1) {
+      print {$out} '|   ' x ($indent - 1) . '+--->';
+   }
+   print {$out} $funcs[0] . "\n";
+   shift @funcs;
+   my $olddeplist = $deplist;
+   foreach my $i (@funcs) {
+      $deplist = draw_func($deplist, $depmap, $out, $indent + 1, ${$depmap}{$i}) if exists ${$depmap}{$i};
+   }
+   return $olddeplist;
+}
+
+sub update_dep
+{
+    #open class file and write preamble
+    open(my $class, '>', 'tommath_class.h') or die "Couldn't open tommath_class.h for writing\n";
+    print {$class} << 'EOS';
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
+#if !(defined(LTM1) && defined(LTM2) && defined(LTM3))
+#define LTM_INSIDE
+#if defined(LTM2)
+#   define LTM3
+#endif
+#if defined(LTM1)
+#   define LTM2
+#endif
+#define LTM1
+#if defined(LTM_ALL)
+EOS
+
+    foreach my $filename (glob 'bn*.c') {
+        my $define = $filename;
+
+        print "Processing $filename\n";
+
+        # convert filename to upper case so we can use it as a define
+        $define =~ tr/[a-z]/[A-Z]/;
+        $define =~ tr/\./_/;
+        print {$class} "#   define $define\n";
+
+        # now copy text and apply #ifdef as required
+        my $apply = 0;
+        open(my $src, '<', $filename);
+        open(my $out, '>', 'tmp');
+
+        # first line will be the #ifdef
+        my $line = <$src>;
+        if ($line =~ /include/) {
+            print {$out} $line;
+        } else {
+            print {$out} << "EOS";
+#include "tommath_private.h"
+#ifdef $define
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+$line
+EOS
+            $apply = 1;
+        }
+        while (<$src>) {
+            if ($_ !~ /tommath\.h/) {
+                print {$out} $_;
+            }
+        }
+        if ($apply == 1) {
+            print {$out} "#endif\n";
+        }
+        close $src;
+        close $out;
+
+        unlink $filename;
+        rename 'tmp', $filename;
+    }
+    print {$class} "#endif\n#endif\n";
+
+    # now do classes
+    my %depmap;
+    foreach my $filename (glob 'bn*.c') {
+        my $content;
+        if ($filename =~ "bn_deprecated.c") {
+            open(my $src, '<', $filename) or die "Can't open source file!\n";
+            read $src, $content, -s $src;
+            close $src;
+        } else {
+            my $cc = $ENV{'CC'} || 'gcc';
+            $content = `$cc -E -x c -DLTM_ALL $filename`;
+            $content =~ s/^# 1 "$filename".*?^# 2 "$filename"//ms;
+        }
+
+        # convert filename to upper case so we can use it as a define
+        $filename =~ tr/[a-z]/[A-Z]/;
+        $filename =~ tr/\./_/;
+
+        print {$class} "#if defined($filename)\n";
+        my $list = $filename;
+
+        # strip comments
+        $content =~ s{/\*.*?\*/}{}gs;
+
+        # scan for mp_* and make classes
+        my @deps = ();
+        foreach my $line (split /\n/, $content) {
+            while ($line =~ /(fast_)?(s_)?mp\_[a-z_0-9]*((?=\;)|(?=\())|(?<=\()mp\_[a-z_0-9]*(?=\()/g) {
+                my $a = $&;
+                next if $a eq "mp_err";
+                $a =~ tr/[a-z]/[A-Z]/;
+                $a = 'BN_' . $a . '_C';
+                push @deps, $a;
+            }
+        }
+        @deps = sort(@deps);
+        foreach my $a (@deps) {
+            if ($list !~ /$a/) {
+                print {$class} "#   define $a\n";
+            }
+            $list = $list . ',' . $a;
+        }
+        $depmap{$filename} = $list;
+
+        print {$class} "#endif\n\n";
+    }
+
+    print {$class} << 'EOS';
+#ifdef LTM_INSIDE
+#undef LTM_INSIDE
+#ifdef LTM3
+#   define LTM_LAST
+#endif
+
+#include "tommath_superclass.h"
+#include "tommath_class.h"
+#else
+#   define LTM_LAST
+#endif
+EOS
+    close $class;
+
+    #now let's make a cool call graph...
+
+    open(my $out, '>', 'callgraph.txt');
+    foreach (sort keys %depmap) {
+        draw_func("", \%depmap, $out, 0, $depmap{$_});
+        print {$out} "\n\n";
+    }
+    close $out;
+
+    return 0;
+}
+
+sub generate_def {
+    my @files = split /\n/, `git ls-files`;
+    @files = grep(/\.c/, @files);
+    @files = map { my $x = $_; $x =~ s/^bn_|\.c$//g; $x; } @files;
+    @files = grep(!/mp_radix_smap/, @files);
+
+    push(@files, qw(mp_set_int mp_set_long mp_set_long_long mp_get_int mp_get_long mp_get_long_long mp_init_set_int));
+
+    my $files = join("\n    ", sort(grep(/^mp_/, @files)));
+    write_file "tommath.def", "; libtommath
+;
+; Use this command to produce a 32-bit .lib file, for use in any MSVC version
+;   lib -machine:X86 -name:libtommath.dll -def:tommath.def -out:tommath.lib
+; Use this command to produce a 64-bit .lib file, for use in any MSVC version
+;   lib -machine:X64 -name:libtommath.dll -def:tommath.def -out:tommath.lib
+;
+EXPORTS
+    $files
+";
+    return 0;
+}
+
+sub die_usage {
+  die <<"MARKER";
+usage: $0 -s   OR   $0 --check-source
+       $0 -o   OR   $0 --check-comments
+       $0 -m   OR   $0 --check-makefiles
+       $0 -a   OR   $0 --check-all
+       $0 -u   OR   $0 --update-files
+MARKER
+}
+
+GetOptions( "s|check-source"        => \my $check_source,
+            "o|check-comments"      => \my $check_comments,
+            "m|check-makefiles"     => \my $check_makefiles,
+            "d|check-doc"           => \my $check_doc,
+            "a|check-all"           => \my $check_all,
+            "u|update-files"        => \my $update_files,
+            "h|help"                => \my $help
+          ) or die_usage;
+
+my $failure;
+$failure ||= check_source()       if $check_all || $check_source;
+$failure ||= check_comments()     if $check_all || $check_comments;
+$failure ||= check_doc()          if $check_doc; # temporarily excluded from --check-all
+$failure ||= process_makefiles(0) if $check_all || $check_makefiles;
+$failure ||= process_makefiles(1) if $update_files;
+$failure ||= update_dep()         if $update_files;
+$failure ||= generate_def()       if $update_files;
+
+die_usage unless defined $failure;
+exit $failure ? 1 : 0;
diff --git a/libtommath/libtommath_VS2008.vcproj b/libtommath/libtommath_VS2008.vcproj
index 5b2637b..67cc89b 100644
--- a/libtommath/libtommath_VS2008.vcproj
+++ b/libtommath/libtommath_VS2008.vcproj
@@ -313,27 +313,11 @@
 	</References>

 	<Files>

 		<File

-			RelativePath="bn_error.c"

+			RelativePath="bn_cutoffs.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_fast_mp_invmod.c"

-			>

-		</File>

-		<File

-			RelativePath="bn_fast_mp_montgomery_reduce.c"

-			>

-		</File>

-		<File

-			RelativePath="bn_fast_s_mp_mul_digs.c"

-			>

-		</File>

-		<File

-			RelativePath="bn_fast_s_mp_mul_high_digs.c"

-			>

-		</File>

-		<File

-			RelativePath="bn_fast_s_mp_sqr.c"

+			RelativePath="bn_deprecated.c"

 			>

 		</File>

 		<File

@@ -401,6 +385,10 @@
 			>

 		</File>

 		<File

+			RelativePath="bn_mp_decr.c"

+			>

+		</File>

+		<File

 			RelativePath="bn_mp_div.c"

 			>

 		</File>

@@ -433,35 +421,35 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_exch.c"

+			RelativePath="bn_mp_error_to_string.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_export.c"

+			RelativePath="bn_mp_exch.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_expt_d.c"

+			RelativePath="bn_mp_expt_u32.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_expt_d_ex.c"

+			RelativePath="bn_mp_exptmod.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_exptmod.c"

+			RelativePath="bn_mp_exteuclid.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_exptmod_fast.c"

+			RelativePath="bn_mp_fread.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_exteuclid.c"

+			RelativePath="bn_mp_from_sbin.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_fread.c"

+			RelativePath="bn_mp_from_ubin.c"

 			>

 		</File>

 		<File

@@ -473,23 +461,39 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_get_bit.c"

+			RelativePath="bn_mp_get_double.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_get_double.c"

+			RelativePath="bn_mp_get_i32.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_get_i64.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_get_l.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_get_ll.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_get_int.c"

+			RelativePath="bn_mp_get_mag_u32.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_get_long.c"

+			RelativePath="bn_mp_get_mag_u64.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_get_long_long.c"

+			RelativePath="bn_mp_get_mag_ul.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_get_mag_ull.c"

 			>

 		</File>

 		<File

@@ -497,7 +501,7 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_import.c"

+			RelativePath="bn_mp_incr.c"

 			>

 		</File>

 		<File

@@ -509,6 +513,22 @@
 			>

 		</File>

 		<File

+			RelativePath="bn_mp_init_i32.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_init_i64.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_init_l.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_init_ll.c"

+			>

+		</File>

+		<File

 			RelativePath="bn_mp_init_multi.c"

 			>

 		</File>

@@ -517,35 +537,39 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_init_set_int.c"

+			RelativePath="bn_mp_init_size.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_init_size.c"

+			RelativePath="bn_mp_init_u32.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_invmod.c"

+			RelativePath="bn_mp_init_u64.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_invmod_slow.c"

+			RelativePath="bn_mp_init_ul.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_is_square.c"

+			RelativePath="bn_mp_init_ull.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_invmod.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_jacobi.c"

+			RelativePath="bn_mp_is_square.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_karatsuba_mul.c"

+			RelativePath="bn_mp_iseven.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_karatsuba_sqr.c"

+			RelativePath="bn_mp_isodd.c"

 			>

 		</File>

 		<File

@@ -557,6 +581,10 @@
 			>

 		</File>

 		<File

+			RelativePath="bn_mp_log_u32.c"

+			>

+		</File>

+		<File

 			RelativePath="bn_mp_lshd.c"

 			>

 		</File>

@@ -605,19 +633,19 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_n_root.c"

+			RelativePath="bn_mp_neg.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_n_root_ex.c"

+			RelativePath="bn_mp_or.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_neg.c"

+			RelativePath="bn_mp_pack.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_or.c"

+			RelativePath="bn_mp_pack_count.c"

 			>

 		</File>

 		<File

@@ -629,10 +657,6 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_prime_is_divisible.c"

-			>

-		</File>

-		<File

 			RelativePath="bn_mp_prime_is_prime.c"

 			>

 		</File>

@@ -649,7 +673,7 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_prime_random_ex.c"

+			RelativePath="bn_mp_prime_rand.c"

 			>

 		</File>

 		<File

@@ -673,14 +697,6 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_read_signed_bin.c"

-			>

-		</File>

-		<File

-			RelativePath="bn_mp_read_unsigned_bin.c"

-			>

-		</File>

-		<File

 			RelativePath="bn_mp_reduce.c"

 			>

 		</File>

@@ -713,10 +729,18 @@
 			>

 		</File>

 		<File

+			RelativePath="bn_mp_root_u32.c"

+			>

+		</File>

+		<File

 			RelativePath="bn_mp_rshd.c"

 			>

 		</File>

 		<File

+			RelativePath="bn_mp_sbin_size.c"

+			>

+		</File>

+		<File

 			RelativePath="bn_mp_set.c"

 			>

 		</File>

@@ -725,15 +749,35 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_set_int.c"

+			RelativePath="bn_mp_set_i32.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_set_i64.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_set_long.c"

+			RelativePath="bn_mp_set_l.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_set_long_long.c"

+			RelativePath="bn_mp_set_ll.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_set_u32.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_set_u64.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_set_ul.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_mp_set_ull.c"

 			>

 		</File>

 		<File

@@ -741,7 +785,7 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_signed_bin_size.c"

+			RelativePath="bn_mp_signed_rsh.c"

 			>

 		</File>

 		<File

@@ -773,83 +817,83 @@
 			>

 		</File>

 		<File

-			RelativePath="bn_mp_tc_and.c"

+			RelativePath="bn_mp_to_radix.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_tc_div_2d.c"

+			RelativePath="bn_mp_to_sbin.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_tc_or.c"

+			RelativePath="bn_mp_to_ubin.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_tc_xor.c"

+			RelativePath="bn_mp_ubin_size.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_to_signed_bin.c"

+			RelativePath="bn_mp_unpack.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_to_signed_bin_n.c"

+			RelativePath="bn_mp_xor.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_to_unsigned_bin.c"

+			RelativePath="bn_mp_zero.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_to_unsigned_bin_n.c"

+			RelativePath="bn_prime_tab.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_toom_mul.c"

+			RelativePath="bn_s_mp_add.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_toom_sqr.c"

+			RelativePath="bn_s_mp_balance_mul.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_toradix.c"

+			RelativePath="bn_s_mp_exptmod.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_toradix_n.c"

+			RelativePath="bn_s_mp_exptmod_fast.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_unsigned_bin_size.c"

+			RelativePath="bn_s_mp_get_bit.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_xor.c"

+			RelativePath="bn_s_mp_invmod_fast.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_mp_zero.c"

+			RelativePath="bn_s_mp_invmod_slow.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_prime_tab.c"

+			RelativePath="bn_s_mp_karatsuba_mul.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_reverse.c"

+			RelativePath="bn_s_mp_karatsuba_sqr.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_s_mp_add.c"

+			RelativePath="bn_s_mp_montgomery_reduce_fast.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_s_mp_exptmod.c"

+			RelativePath="bn_s_mp_mul_digs.c"

 			>

 		</File>

 		<File

-			RelativePath="bn_s_mp_mul_digs.c"

+			RelativePath="bn_s_mp_mul_digs_fast.c"

 			>

 		</File>

 		<File

@@ -857,15 +901,43 @@
 			>

 		</File>

 		<File

+			RelativePath="bn_s_mp_mul_high_digs_fast.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_s_mp_prime_is_divisible.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_s_mp_rand_jenkins.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_s_mp_rand_platform.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_s_mp_reverse.c"

+			>

+		</File>

+		<File

 			RelativePath="bn_s_mp_sqr.c"

 			>

 		</File>

 		<File

+			RelativePath="bn_s_mp_sqr_fast.c"

+			>

+		</File>

+		<File

 			RelativePath="bn_s_mp_sub.c"

 			>

 		</File>

 		<File

-			RelativePath="bncore.c"

+			RelativePath="bn_s_mp_toom_mul.c"

+			>

+		</File>

+		<File

+			RelativePath="bn_s_mp_toom_sqr.c"

 			>

 		</File>

 		<File

@@ -877,6 +949,10 @@
 			>

 		</File>

 		<File

+			RelativePath="tommath_cutoffs.h"

+			>

+		</File>

+		<File

 			RelativePath="tommath_private.h"

 			>

 		</File>

diff --git a/libtommath/makefile b/libtommath/makefile
index a07c274..be9fac6 100644
--- a/libtommath/makefile
+++ b/libtommath/makefile
@@ -17,46 +17,47 @@ coverage: LIBNAME:=-Wl,--whole-archive $(LIBNAME)  -Wl,--no-whole-archive
 
 include makefile_include.mk
 
-%.o: %.c
+%.o: %.c $(HEADERS)
 ifneq ($V,1)
 	@echo "   * ${CC} $@"
 endif
-	${silent} ${CC} -c ${CFLAGS} $< -o $@
+	${silent} ${CC} -c ${LTM_CFLAGS} $< -o $@
 
 LCOV_ARGS=--directory .
 
 #START_INS
-OBJECTS=bn_error.o bn_fast_mp_invmod.o bn_fast_mp_montgomery_reduce.o bn_fast_s_mp_mul_digs.o \
-bn_fast_s_mp_mul_high_digs.o bn_fast_s_mp_sqr.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o \
-bn_mp_addmod.o bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o \
-bn_mp_cmp_mag.o bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_div.o \
-bn_mp_div_2.o bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o \
-bn_mp_dr_setup.o bn_mp_exch.o bn_mp_export.o bn_mp_expt_d.o bn_mp_expt_d_ex.o bn_mp_exptmod.o \
-bn_mp_exptmod_fast.o bn_mp_exteuclid.o bn_mp_fread.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_bit.o \
-bn_mp_get_double.o bn_mp_get_int.o bn_mp_get_long.o bn_mp_get_long_long.o bn_mp_grow.o bn_mp_import.o \
-bn_mp_init.o bn_mp_init_copy.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_set_int.o bn_mp_init_size.o \
-bn_mp_invmod.o bn_mp_invmod_slow.o bn_mp_is_square.o bn_mp_jacobi.o bn_mp_karatsuba_mul.o \
-bn_mp_karatsuba_sqr.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_lshd.o bn_mp_mod.o bn_mp_mod_2d.o bn_mp_mod_d.o \
-bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o bn_mp_montgomery_setup.o bn_mp_mul.o \
-bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_n_root.o bn_mp_n_root_ex.o bn_mp_neg.o \
-bn_mp_or.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o bn_mp_prime_is_divisible.o \
+OBJECTS=bn_cutoffs.o bn_deprecated.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o bn_mp_addmod.o \
+bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o bn_mp_cmp_mag.o \
+bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_decr.o bn_mp_div.o bn_mp_div_2.o \
+bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o bn_mp_dr_setup.o \
+bn_mp_error_to_string.o bn_mp_exch.o bn_mp_expt_u32.o bn_mp_exptmod.o bn_mp_exteuclid.o bn_mp_fread.o \
+bn_mp_from_sbin.o bn_mp_from_ubin.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_double.o bn_mp_get_i32.o \
+bn_mp_get_i64.o bn_mp_get_l.o bn_mp_get_ll.o bn_mp_get_mag_u32.o bn_mp_get_mag_u64.o bn_mp_get_mag_ul.o \
+bn_mp_get_mag_ull.o bn_mp_grow.o bn_mp_incr.o bn_mp_init.o bn_mp_init_copy.o bn_mp_init_i32.o \
+bn_mp_init_i64.o bn_mp_init_l.o bn_mp_init_ll.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_size.o \
+bn_mp_init_u32.o bn_mp_init_u64.o bn_mp_init_ul.o bn_mp_init_ull.o bn_mp_invmod.o bn_mp_is_square.o \
+bn_mp_iseven.o bn_mp_isodd.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_log_u32.o bn_mp_lshd.o bn_mp_mod.o \
+bn_mp_mod_2d.o bn_mp_mod_d.o bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o \
+bn_mp_montgomery_setup.o bn_mp_mul.o bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_neg.o \
+bn_mp_or.o bn_mp_pack.o bn_mp_pack_count.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o \
 bn_mp_prime_is_prime.o bn_mp_prime_miller_rabin.o bn_mp_prime_next_prime.o \
-bn_mp_prime_rabin_miller_trials.o bn_mp_prime_random_ex.o bn_mp_prime_strong_lucas_selfridge.o \
-bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_read_signed_bin.o \
-bn_mp_read_unsigned_bin.o bn_mp_reduce.o bn_mp_reduce_2k.o bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o \
-bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_rshd.o \
-bn_mp_set.o bn_mp_set_double.o bn_mp_set_int.o bn_mp_set_long.o bn_mp_set_long_long.o bn_mp_shrink.o \
-bn_mp_signed_bin_size.o bn_mp_sqr.o bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o \
-bn_mp_sub_d.o bn_mp_submod.o bn_mp_tc_and.o bn_mp_tc_div_2d.o bn_mp_tc_or.o bn_mp_tc_xor.o \
-bn_mp_to_signed_bin.o bn_mp_to_signed_bin_n.o bn_mp_to_unsigned_bin.o bn_mp_to_unsigned_bin_n.o \
-bn_mp_toom_mul.o bn_mp_toom_sqr.o bn_mp_toradix.o bn_mp_toradix_n.o bn_mp_unsigned_bin_size.o bn_mp_xor.o \
-bn_mp_zero.o bn_prime_tab.o bn_reverse.o bn_s_mp_add.o bn_s_mp_exptmod.o bn_s_mp_mul_digs.o \
-bn_s_mp_mul_high_digs.o bn_s_mp_sqr.o bn_s_mp_sub.o bncore.o
+bn_mp_prime_rabin_miller_trials.o bn_mp_prime_rand.o bn_mp_prime_strong_lucas_selfridge.o \
+bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_reduce.o bn_mp_reduce_2k.o \
+bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o \
+bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_root_u32.o bn_mp_rshd.o bn_mp_sbin_size.o bn_mp_set.o \
+bn_mp_set_double.o bn_mp_set_i32.o bn_mp_set_i64.o bn_mp_set_l.o bn_mp_set_ll.o bn_mp_set_u32.o \
+bn_mp_set_u64.o bn_mp_set_ul.o bn_mp_set_ull.o bn_mp_shrink.o bn_mp_signed_rsh.o bn_mp_sqr.o \
+bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o bn_mp_sub_d.o bn_mp_submod.o \
+bn_mp_to_radix.o bn_mp_to_sbin.o bn_mp_to_ubin.o bn_mp_ubin_size.o bn_mp_unpack.o bn_mp_xor.o bn_mp_zero.o \
+bn_prime_tab.o bn_s_mp_add.o bn_s_mp_balance_mul.o bn_s_mp_exptmod.o bn_s_mp_exptmod_fast.o \
+bn_s_mp_get_bit.o bn_s_mp_invmod_fast.o bn_s_mp_invmod_slow.o bn_s_mp_karatsuba_mul.o \
+bn_s_mp_karatsuba_sqr.o bn_s_mp_montgomery_reduce_fast.o bn_s_mp_mul_digs.o bn_s_mp_mul_digs_fast.o \
+bn_s_mp_mul_high_digs.o bn_s_mp_mul_high_digs_fast.o bn_s_mp_prime_is_divisible.o \
+bn_s_mp_rand_jenkins.o bn_s_mp_rand_platform.o bn_s_mp_reverse.o bn_s_mp_sqr.o bn_s_mp_sqr_fast.o \
+bn_s_mp_sub.o bn_s_mp_toom_mul.o bn_s_mp_toom_sqr.o
 
 #END_INS
 
-$(OBJECTS): $(HEADERS)
-
 $(LIBNAME):  $(OBJECTS)
 	$(AR) $(ARFLAGS) $@ $(OBJECTS)
 	$(RANLIB) $@
@@ -76,11 +77,11 @@ profiled:
 #make a single object profiled library
 profiled_single:
 	perl gen.pl
-	$(CC) $(CFLAGS) -fprofile-arcs -DTESTING -c mpi.c -o mpi.o
-	$(CC) $(CFLAGS) -DTESTING -DTIMER demo/timing.c mpi.o -lgcov -o timing
+	$(CC) $(LTM_CFLAGS) -fprofile-arcs -DTESTING -c mpi.c -o mpi.o
+	$(CC) $(LTM_CFLAGS) -DTESTING -DTIMER demo/timing.c mpi.o -lgcov -o timing
 	./timing
 	rm -f *.o timing
-	$(CC) $(CFLAGS) -fbranch-probabilities -DTESTING -c mpi.c -o mpi.o
+	$(CC) $(LTM_CFLAGS) -fbranch-probabilities -DTESTING -c mpi.c -o mpi.o
 	$(AR) $(ARFLAGS) $(LIBNAME) mpi.o
 	ranlib $(LIBNAME)
 
@@ -94,30 +95,37 @@ uninstall:
 	rm $(DESTDIR)$(LIBPATH)/$(LIBNAME)
 	rm $(HEADERS_PUB:%=$(DESTDIR)$(INCPATH)/%)
 
-test: $(LIBNAME) demo/demo.o
-	$(CC) $(CFLAGS) demo/demo.o $(LIBNAME) $(LFLAGS) -o test
+test_standalone: test
+	@echo "test_standalone is deprecated, please use make-target 'test'"
+
+DEMOS=test mtest_opponent
+
+define DEMO_template
+$(1): demo/$(1).o demo/shared.o $$(LIBNAME)
+	$$(CC) $$(LTM_CFLAGS) $$(LTM_LFLAGS) $$^ -o $$@
+endef
 
-test_standalone: $(LIBNAME) demo/demo.o
-	$(CC) $(CFLAGS) demo/demo.o $(LIBNAME) $(LFLAGS) -o test
+$(foreach demo, $(strip $(DEMOS)), $(eval $(call DEMO_template,$(demo))))
 
 .PHONY: mtest
 mtest:
-	cd mtest ; $(CC) $(CFLAGS) -O0 mtest.c $(LFLAGS) -o mtest
+	cd mtest ; $(CC) $(LTM_CFLAGS) -O0 mtest.c $(LTM_LFLAGS) -o mtest
 
 timing: $(LIBNAME) demo/timing.c
-	$(CC) $(CFLAGS) -DTIMER demo/timing.c $(LIBNAME) $(LFLAGS) -o timing
+	$(CC) $(LTM_CFLAGS) -DTIMER demo/timing.c $(LIBNAME) $(LTM_LFLAGS) -o timing
+
+tune: $(LIBNAME)
+	$(MAKE) -C etc tune CFLAGS="$(LTM_CFLAGS)"
+	$(MAKE)
 
 # You have to create a file .coveralls.yml with the content "repo_token: <the token>"
 # in the base folder to be able to submit to coveralls
 coveralls: lcov
 	coveralls-lcov
 
-docdvi poster docs mandvi manual:
+docs manual:
 	$(MAKE) -C doc/ $@ V=$(V)
 
-pretty:
-	perl pretty.build
-
 .PHONY: pre_gen
 pre_gen:
 	mkdir -p pre_gen
@@ -125,7 +133,7 @@ pre_gen:
 	sed -e 's/[[:blank:]]*$$//' mpi.c > pre_gen/mpi.c
 	rm mpi.c
 
-zipup: clean astyle new_file manual poster docs
+zipup: clean astyle new_file docs
 	@# Update the index, so diff-index won't fail in case the pdf has been created.
 	@#   As the pdf creation modifies the tex files, git sometimes detects the
 	@#   modified files, but misses that it's put back to its original version.
@@ -137,22 +145,21 @@ zipup: clean astyle new_file manual poster docs
 	@echo 'fixme check'
 	-@(find libtommath-$(VERSION)/ -type f | xargs grep 'FIXM[E]') && echo '############## BEWARE: the "fixme" marker was found !!! ##############' || true
 	mkdir -p libtommath-$(VERSION)/doc
-	cp doc/bn.pdf doc/tommath.pdf doc/poster.pdf libtommath-$(VERSION)/doc/
+	cp doc/bn.pdf libtommath-$(VERSION)/doc/
 	$(MAKE) -C libtommath-$(VERSION)/ pre_gen
 	tar -c libtommath-$(VERSION)/ | xz -6e -c - > ltm-$(VERSION).tar.xz
 	zip -9rq ltm-$(VERSION).zip libtommath-$(VERSION)
 	cp doc/bn.pdf bn-$(VERSION).pdf
-	cp doc/tommath.pdf tommath-$(VERSION).pdf
 	rm -rf libtommath-$(VERSION)
 	gpg -b -a ltm-$(VERSION).tar.xz
 	gpg -b -a ltm-$(VERSION).zip
 
 new_file:
-	bash updatemakes.sh
-	perl dep.pl
+	perl helper.pl --update-files
 
 perlcritic:
 	perlcritic *.pl doc/*.pl
 
 astyle:
-	astyle --options=astylerc $(OBJECTS:.o=.c) tommath*.h demo/*.c etc/*.c mtest/mtest.c
+	@echo "   * run astyle on all sources"
+	@astyle --options=astylerc --formatted $(OBJECTS:.o=.c) tommath*.h demo/*.c etc/*.c mtest/mtest.c
diff --git a/libtommath/makefile.mingw b/libtommath/makefile.mingw
index ec0de2b..7eee57d 100644
--- a/libtommath/makefile.mingw
+++ b/libtommath/makefile.mingw
@@ -1,6 +1,6 @@
 # MAKEFILE for MS Windows (mingw + gcc + gmake)
 #
-# BEWARE: variable OBJECTS is updated via ./updatemakes.sh
+# BEWARE: variable OBJECTS is updated via helper.pl
 
 ### USAGE:
 # Open a command prompt with gcc + gmake in PATH and start:
@@ -21,7 +21,7 @@ LDFLAGS   =
 
 #Compilation flags
 LTM_CFLAGS  = -I. $(CFLAGS)
-LTM_LDFLAGS = $(LDFLAGS)
+LTM_LDFLAGS = $(LDFLAGS) -static-libgcc
 
 #Libraries to be created
 LIBMAIN_S =libtommath.a
@@ -29,36 +29,38 @@ LIBMAIN_I =libtommath.dll.a
 LIBMAIN_D =libtommath.dll
 
 #List of objects to compile (all goes to libtommath.a)
-OBJECTS=bn_error.o bn_fast_mp_invmod.o bn_fast_mp_montgomery_reduce.o bn_fast_s_mp_mul_digs.o \
-bn_fast_s_mp_mul_high_digs.o bn_fast_s_mp_sqr.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o \
-bn_mp_addmod.o bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o \
-bn_mp_cmp_mag.o bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_div.o \
-bn_mp_div_2.o bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o \
-bn_mp_dr_setup.o bn_mp_exch.o bn_mp_export.o bn_mp_expt_d.o bn_mp_expt_d_ex.o bn_mp_exptmod.o \
-bn_mp_exptmod_fast.o bn_mp_exteuclid.o bn_mp_fread.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_bit.o \
-bn_mp_get_double.o bn_mp_get_int.o bn_mp_get_long.o bn_mp_get_long_long.o bn_mp_grow.o bn_mp_import.o \
-bn_mp_init.o bn_mp_init_copy.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_set_int.o bn_mp_init_size.o \
-bn_mp_invmod.o bn_mp_invmod_slow.o bn_mp_is_square.o bn_mp_jacobi.o bn_mp_karatsuba_mul.o \
-bn_mp_karatsuba_sqr.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_lshd.o bn_mp_mod.o bn_mp_mod_2d.o bn_mp_mod_d.o \
-bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o bn_mp_montgomery_setup.o bn_mp_mul.o \
-bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_n_root.o bn_mp_n_root_ex.o bn_mp_neg.o \
-bn_mp_or.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o bn_mp_prime_is_divisible.o \
+OBJECTS=bn_cutoffs.o bn_deprecated.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o bn_mp_addmod.o \
+bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o bn_mp_cmp_mag.o \
+bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_decr.o bn_mp_div.o bn_mp_div_2.o \
+bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o bn_mp_dr_setup.o \
+bn_mp_error_to_string.o bn_mp_exch.o bn_mp_expt_u32.o bn_mp_exptmod.o bn_mp_exteuclid.o bn_mp_fread.o \
+bn_mp_from_sbin.o bn_mp_from_ubin.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_double.o bn_mp_get_i32.o \
+bn_mp_get_i64.o bn_mp_get_l.o bn_mp_get_ll.o bn_mp_get_mag_u32.o bn_mp_get_mag_u64.o bn_mp_get_mag_ul.o \
+bn_mp_get_mag_ull.o bn_mp_grow.o bn_mp_incr.o bn_mp_init.o bn_mp_init_copy.o bn_mp_init_i32.o \
+bn_mp_init_i64.o bn_mp_init_l.o bn_mp_init_ll.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_size.o \
+bn_mp_init_u32.o bn_mp_init_u64.o bn_mp_init_ul.o bn_mp_init_ull.o bn_mp_invmod.o bn_mp_is_square.o \
+bn_mp_iseven.o bn_mp_isodd.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_log_u32.o bn_mp_lshd.o bn_mp_mod.o \
+bn_mp_mod_2d.o bn_mp_mod_d.o bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o \
+bn_mp_montgomery_setup.o bn_mp_mul.o bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_neg.o \
+bn_mp_or.o bn_mp_pack.o bn_mp_pack_count.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o \
 bn_mp_prime_is_prime.o bn_mp_prime_miller_rabin.o bn_mp_prime_next_prime.o \
-bn_mp_prime_rabin_miller_trials.o bn_mp_prime_random_ex.o bn_mp_prime_strong_lucas_selfridge.o \
-bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_read_signed_bin.o \
-bn_mp_read_unsigned_bin.o bn_mp_reduce.o bn_mp_reduce_2k.o bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o \
-bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_rshd.o \
-bn_mp_set.o bn_mp_set_double.o bn_mp_set_int.o bn_mp_set_long.o bn_mp_set_long_long.o bn_mp_shrink.o \
-bn_mp_signed_bin_size.o bn_mp_sqr.o bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o \
-bn_mp_sub_d.o bn_mp_submod.o bn_mp_tc_and.o bn_mp_tc_div_2d.o bn_mp_tc_or.o bn_mp_tc_xor.o \
-bn_mp_to_signed_bin.o bn_mp_to_signed_bin_n.o bn_mp_to_unsigned_bin.o bn_mp_to_unsigned_bin_n.o \
-bn_mp_toom_mul.o bn_mp_toom_sqr.o bn_mp_toradix.o bn_mp_toradix_n.o bn_mp_unsigned_bin_size.o bn_mp_xor.o \
-bn_mp_zero.o bn_prime_tab.o bn_reverse.o bn_s_mp_add.o bn_s_mp_exptmod.o bn_s_mp_mul_digs.o \
-bn_s_mp_mul_high_digs.o bn_s_mp_sqr.o bn_s_mp_sub.o bncore.o
-
-HEADERS_PUB=tommath.h tommath_class.h tommath_superclass.h
-
-HEADERS=tommath_private.h $(HEADERS_PUB)
+bn_mp_prime_rabin_miller_trials.o bn_mp_prime_rand.o bn_mp_prime_strong_lucas_selfridge.o \
+bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_reduce.o bn_mp_reduce_2k.o \
+bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o \
+bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_root_u32.o bn_mp_rshd.o bn_mp_sbin_size.o bn_mp_set.o \
+bn_mp_set_double.o bn_mp_set_i32.o bn_mp_set_i64.o bn_mp_set_l.o bn_mp_set_ll.o bn_mp_set_u32.o \
+bn_mp_set_u64.o bn_mp_set_ul.o bn_mp_set_ull.o bn_mp_shrink.o bn_mp_signed_rsh.o bn_mp_sqr.o \
+bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o bn_mp_sub_d.o bn_mp_submod.o \
+bn_mp_to_radix.o bn_mp_to_sbin.o bn_mp_to_ubin.o bn_mp_ubin_size.o bn_mp_unpack.o bn_mp_xor.o bn_mp_zero.o \
+bn_prime_tab.o bn_s_mp_add.o bn_s_mp_balance_mul.o bn_s_mp_exptmod.o bn_s_mp_exptmod_fast.o \
+bn_s_mp_get_bit.o bn_s_mp_invmod_fast.o bn_s_mp_invmod_slow.o bn_s_mp_karatsuba_mul.o \
+bn_s_mp_karatsuba_sqr.o bn_s_mp_montgomery_reduce_fast.o bn_s_mp_mul_digs.o bn_s_mp_mul_digs_fast.o \
+bn_s_mp_mul_high_digs.o bn_s_mp_mul_high_digs_fast.o bn_s_mp_prime_is_divisible.o \
+bn_s_mp_rand_jenkins.o bn_s_mp_rand_platform.o bn_s_mp_reverse.o bn_s_mp_sqr.o bn_s_mp_sqr_fast.o \
+bn_s_mp_sub.o bn_s_mp_toom_mul.o bn_s_mp_toom_sqr.o
+
+HEADERS_PUB=tommath.h
+HEADERS=tommath_private.h tommath_class.h tommath_superclass.h tommath_cutoffs.h $(HEADERS_PUB)
 
 #The default rule for make builds the libtommath.a library (static)
 default: $(LIBMAIN_S)
@@ -79,14 +81,19 @@ $(LIBMAIN_D) $(LIBMAIN_I): $(OBJECTS)
 	$(CC) -s -shared -o $(LIBMAIN_D) $^ -Wl,--enable-auto-import,--export-all -Wl,--out-implib=$(LIBMAIN_I) $(LTM_LDFLAGS)
 	$(STRIP) -S $(LIBMAIN_D)
 
-#Build test_standalone suite
-test.exe: $(LIBMAIN_S) demo/demo.c
-	$(CC) $(LTM_CFLAGS) $(LTM_LDFLAGS) demo/demo.c $(LIBMAIN_S) -DLTM_DEMO_TEST_VS_MTEST=0 -o $@
+#Build test suite
+test.exe: demo/shared.o demo/test.o $(LIBMAIN_S)
+	$(CC) $(LTM_CFLAGS) $(LTM_LDFLAGS) $^ -o $@
 	@echo NOTICE: start the tests by launching test.exe
 
 test_standalone: test.exe
+	@echo test_standalone is deprecated, please use make-target 'test.exe'
 
-all: $(LIBMAIN_S) test_standalone
+all: $(LIBMAIN_S) test.exe
+
+tune: $(LIBNAME_S)
+	$(MAKE) -C etc tune
+	$(MAKE)
 
 clean:
 	@-cmd /c del /Q /S *.o *.a *.exe *.dll 2>nul
@@ -100,7 +107,3 @@ install: $(LIBMAIN_S) $(LIBMAIN_I) $(LIBMAIN_D)
 	copy /Y $(LIBMAIN_I) "$(PREFIX)\lib"
 	copy /Y $(LIBMAIN_D) "$(PREFIX)\bin"
 	copy /Y tommath*.h "$(PREFIX)\include"
-
-# ref:         $Format:%D$
-# git commit:  $Format:%H$
-# commit time: $Format:%ai$
diff --git a/libtommath/makefile.msvc b/libtommath/makefile.msvc
index 50db449..aa8d8be 100644
--- a/libtommath/makefile.msvc
+++ b/libtommath/makefile.msvc
@@ -1,6 +1,6 @@
 # MAKEFILE for MS Windows (nmake + Windows SDK)
 #
-# BEWARE: variable OBJECTS is updated via ./updatemakes.sh
+# BEWARE: variable OBJECTS is updated via helper.pl
 
 ### USAGE:
 # Open a command prompt with WinSDK variables set and start:
@@ -14,43 +14,45 @@ PREFIX    = c:\devel
 CFLAGS    = /Ox
 
 #Compilation flags
-LTM_CFLAGS  = /nologo /I./ /D_CRT_SECURE_NO_WARNINGS /D_CRT_NONSTDC_NO_DEPRECATE /W3 $(CFLAGS)
+LTM_CFLAGS  = /nologo /I./ /D_CRT_SECURE_NO_WARNINGS /D_CRT_NONSTDC_NO_DEPRECATE /D__STDC_WANT_SECURE_LIB__=1 /D_CRT_HAS_CXX17=0 /Wall /wd4146 /wd4127 /wd4668 /wd4710 /wd4711 /wd4820 /wd5045 /WX $(CFLAGS)
 LTM_LDFLAGS = advapi32.lib
 
 #Libraries to be created (this makefile builds only static libraries)
 LIBMAIN_S =tommath.lib
 
 #List of objects to compile (all goes to tommath.lib)
-OBJECTS=bn_error.obj bn_fast_mp_invmod.obj bn_fast_mp_montgomery_reduce.obj bn_fast_s_mp_mul_digs.obj \
-bn_fast_s_mp_mul_high_digs.obj bn_fast_s_mp_sqr.obj bn_mp_2expt.obj bn_mp_abs.obj bn_mp_add.obj bn_mp_add_d.obj \
-bn_mp_addmod.obj bn_mp_and.obj bn_mp_clamp.obj bn_mp_clear.obj bn_mp_clear_multi.obj bn_mp_cmp.obj bn_mp_cmp_d.obj \
-bn_mp_cmp_mag.obj bn_mp_cnt_lsb.obj bn_mp_complement.obj bn_mp_copy.obj bn_mp_count_bits.obj bn_mp_div.obj \
-bn_mp_div_2.obj bn_mp_div_2d.obj bn_mp_div_3.obj bn_mp_div_d.obj bn_mp_dr_is_modulus.obj bn_mp_dr_reduce.obj \
-bn_mp_dr_setup.obj bn_mp_exch.obj bn_mp_export.obj bn_mp_expt_d.obj bn_mp_expt_d_ex.obj bn_mp_exptmod.obj \
-bn_mp_exptmod_fast.obj bn_mp_exteuclid.obj bn_mp_fread.obj bn_mp_fwrite.obj bn_mp_gcd.obj bn_mp_get_bit.obj \
-bn_mp_get_double.obj bn_mp_get_int.obj bn_mp_get_long.obj bn_mp_get_long_long.obj bn_mp_grow.obj bn_mp_import.obj \
-bn_mp_init.obj bn_mp_init_copy.obj bn_mp_init_multi.obj bn_mp_init_set.obj bn_mp_init_set_int.obj bn_mp_init_size.obj \
-bn_mp_invmod.obj bn_mp_invmod_slow.obj bn_mp_is_square.obj bn_mp_jacobi.obj bn_mp_karatsuba_mul.obj \
-bn_mp_karatsuba_sqr.obj bn_mp_kronecker.obj bn_mp_lcm.obj bn_mp_lshd.obj bn_mp_mod.obj bn_mp_mod_2d.obj bn_mp_mod_d.obj \
-bn_mp_montgomery_calc_normalization.obj bn_mp_montgomery_reduce.obj bn_mp_montgomery_setup.obj bn_mp_mul.obj \
-bn_mp_mul_2.obj bn_mp_mul_2d.obj bn_mp_mul_d.obj bn_mp_mulmod.obj bn_mp_n_root.obj bn_mp_n_root_ex.obj bn_mp_neg.obj \
-bn_mp_or.obj bn_mp_prime_fermat.obj bn_mp_prime_frobenius_underwood.obj bn_mp_prime_is_divisible.obj \
+OBJECTS=bn_cutoffs.obj bn_deprecated.obj bn_mp_2expt.obj bn_mp_abs.obj bn_mp_add.obj bn_mp_add_d.obj bn_mp_addmod.obj \
+bn_mp_and.obj bn_mp_clamp.obj bn_mp_clear.obj bn_mp_clear_multi.obj bn_mp_cmp.obj bn_mp_cmp_d.obj bn_mp_cmp_mag.obj \
+bn_mp_cnt_lsb.obj bn_mp_complement.obj bn_mp_copy.obj bn_mp_count_bits.obj bn_mp_decr.obj bn_mp_div.obj bn_mp_div_2.obj \
+bn_mp_div_2d.obj bn_mp_div_3.obj bn_mp_div_d.obj bn_mp_dr_is_modulus.obj bn_mp_dr_reduce.obj bn_mp_dr_setup.obj \
+bn_mp_error_to_string.obj bn_mp_exch.obj bn_mp_expt_u32.obj bn_mp_exptmod.obj bn_mp_exteuclid.obj bn_mp_fread.obj \
+bn_mp_from_sbin.obj bn_mp_from_ubin.obj bn_mp_fwrite.obj bn_mp_gcd.obj bn_mp_get_double.obj bn_mp_get_i32.obj \
+bn_mp_get_i64.obj bn_mp_get_l.obj bn_mp_get_ll.obj bn_mp_get_mag_u32.obj bn_mp_get_mag_u64.obj bn_mp_get_mag_ul.obj \
+bn_mp_get_mag_ull.obj bn_mp_grow.obj bn_mp_incr.obj bn_mp_init.obj bn_mp_init_copy.obj bn_mp_init_i32.obj \
+bn_mp_init_i64.obj bn_mp_init_l.obj bn_mp_init_ll.obj bn_mp_init_multi.obj bn_mp_init_set.obj bn_mp_init_size.obj \
+bn_mp_init_u32.obj bn_mp_init_u64.obj bn_mp_init_ul.obj bn_mp_init_ull.obj bn_mp_invmod.obj bn_mp_is_square.obj \
+bn_mp_iseven.obj bn_mp_isodd.obj bn_mp_kronecker.obj bn_mp_lcm.obj bn_mp_log_u32.obj bn_mp_lshd.obj bn_mp_mod.obj \
+bn_mp_mod_2d.obj bn_mp_mod_d.obj bn_mp_montgomery_calc_normalization.obj bn_mp_montgomery_reduce.obj \
+bn_mp_montgomery_setup.obj bn_mp_mul.obj bn_mp_mul_2.obj bn_mp_mul_2d.obj bn_mp_mul_d.obj bn_mp_mulmod.obj bn_mp_neg.obj \
+bn_mp_or.obj bn_mp_pack.obj bn_mp_pack_count.obj bn_mp_prime_fermat.obj bn_mp_prime_frobenius_underwood.obj \
 bn_mp_prime_is_prime.obj bn_mp_prime_miller_rabin.obj bn_mp_prime_next_prime.obj \
-bn_mp_prime_rabin_miller_trials.obj bn_mp_prime_random_ex.obj bn_mp_prime_strong_lucas_selfridge.obj \
-bn_mp_radix_size.obj bn_mp_radix_smap.obj bn_mp_rand.obj bn_mp_read_radix.obj bn_mp_read_signed_bin.obj \
-bn_mp_read_unsigned_bin.obj bn_mp_reduce.obj bn_mp_reduce_2k.obj bn_mp_reduce_2k_l.obj bn_mp_reduce_2k_setup.obj \
-bn_mp_reduce_2k_setup_l.obj bn_mp_reduce_is_2k.obj bn_mp_reduce_is_2k_l.obj bn_mp_reduce_setup.obj bn_mp_rshd.obj \
-bn_mp_set.obj bn_mp_set_double.obj bn_mp_set_int.obj bn_mp_set_long.obj bn_mp_set_long_long.obj bn_mp_shrink.obj \
-bn_mp_signed_bin_size.obj bn_mp_sqr.obj bn_mp_sqrmod.obj bn_mp_sqrt.obj bn_mp_sqrtmod_prime.obj bn_mp_sub.obj \
-bn_mp_sub_d.obj bn_mp_submod.obj bn_mp_tc_and.obj bn_mp_tc_div_2d.obj bn_mp_tc_or.obj bn_mp_tc_xor.obj \
-bn_mp_to_signed_bin.obj bn_mp_to_signed_bin_n.obj bn_mp_to_unsigned_bin.obj bn_mp_to_unsigned_bin_n.obj \
-bn_mp_toom_mul.obj bn_mp_toom_sqr.obj bn_mp_toradix.obj bn_mp_toradix_n.obj bn_mp_unsigned_bin_size.obj bn_mp_xor.obj \
-bn_mp_zero.obj bn_prime_tab.obj bn_reverse.obj bn_s_mp_add.obj bn_s_mp_exptmod.obj bn_s_mp_mul_digs.obj \
-bn_s_mp_mul_high_digs.obj bn_s_mp_sqr.obj bn_s_mp_sub.obj bncore.obj
-
-HEADERS_PUB=tommath.h tommath_class.h tommath_superclass.h
-
-HEADERS=tommath_private.h $(HEADERS_PUB)
+bn_mp_prime_rabin_miller_trials.obj bn_mp_prime_rand.obj bn_mp_prime_strong_lucas_selfridge.obj \
+bn_mp_radix_size.obj bn_mp_radix_smap.obj bn_mp_rand.obj bn_mp_read_radix.obj bn_mp_reduce.obj bn_mp_reduce_2k.obj \
+bn_mp_reduce_2k_l.obj bn_mp_reduce_2k_setup.obj bn_mp_reduce_2k_setup_l.obj bn_mp_reduce_is_2k.obj \
+bn_mp_reduce_is_2k_l.obj bn_mp_reduce_setup.obj bn_mp_root_u32.obj bn_mp_rshd.obj bn_mp_sbin_size.obj bn_mp_set.obj \
+bn_mp_set_double.obj bn_mp_set_i32.obj bn_mp_set_i64.obj bn_mp_set_l.obj bn_mp_set_ll.obj bn_mp_set_u32.obj \
+bn_mp_set_u64.obj bn_mp_set_ul.obj bn_mp_set_ull.obj bn_mp_shrink.obj bn_mp_signed_rsh.obj bn_mp_sqr.obj \
+bn_mp_sqrmod.obj bn_mp_sqrt.obj bn_mp_sqrtmod_prime.obj bn_mp_sub.obj bn_mp_sub_d.obj bn_mp_submod.obj \
+bn_mp_to_radix.obj bn_mp_to_sbin.obj bn_mp_to_ubin.obj bn_mp_ubin_size.obj bn_mp_unpack.obj bn_mp_xor.obj bn_mp_zero.obj \
+bn_prime_tab.obj bn_s_mp_add.obj bn_s_mp_balance_mul.obj bn_s_mp_exptmod.obj bn_s_mp_exptmod_fast.obj \
+bn_s_mp_get_bit.obj bn_s_mp_invmod_fast.obj bn_s_mp_invmod_slow.obj bn_s_mp_karatsuba_mul.obj \
+bn_s_mp_karatsuba_sqr.obj bn_s_mp_montgomery_reduce_fast.obj bn_s_mp_mul_digs.obj bn_s_mp_mul_digs_fast.obj \
+bn_s_mp_mul_high_digs.obj bn_s_mp_mul_high_digs_fast.obj bn_s_mp_prime_is_divisible.obj \
+bn_s_mp_rand_jenkins.obj bn_s_mp_rand_platform.obj bn_s_mp_reverse.obj bn_s_mp_sqr.obj bn_s_mp_sqr_fast.obj \
+bn_s_mp_sub.obj bn_s_mp_toom_mul.obj bn_s_mp_toom_sqr.obj
+
+HEADERS_PUB=tommath.h
+HEADERS=tommath_private.h tommath_class.h tommath_superclass.h tommath_cutoffs.h $(HEADERS_PUB)
 
 #The default rule for make builds the tommath.lib library (static)
 default: $(LIBMAIN_S)
@@ -61,18 +63,23 @@ $(OBJECTS): $(HEADERS)
 .c.obj:
 	$(CC) $(LTM_CFLAGS) /c $< /Fo$@
 
-#Create tomcrypt.lib
+#Create tommath.lib
 $(LIBMAIN_S): $(OBJECTS)
 	lib /out:$(LIBMAIN_S) $(OBJECTS)
 
-#Build test_standalone suite
-test.exe: $(LIBMAIN_S) demo/demo.c
-	cl $(LTM_CFLAGS) $(TOBJECTS) $(LIBMAIN_S) $(LTM_LDFLAGS) demo/demo.c /DLTM_DEMO_TEST_VS_MTEST=0 /Fe$@
+#Build test suite
+test.exe: $(LIBMAIN_S) demo/shared.obj demo/test.obj
+	cl $(LTM_CFLAGS) $(TOBJECTS) $(LIBMAIN_S) $(LTM_LDFLAGS) demo/shared.c demo/test.c /Fe$@
 	@echo NOTICE: start the tests by launching test.exe
 
 test_standalone: test.exe
+	@echo test_standalone is deprecated, please use make-target 'test.exe'
 
-all: $(LIBMAIN_S) test_standalone
+all: $(LIBMAIN_S) test.exe
+
+tune: $(LIBMAIN_S)
+	$(MAKE) -C etc tune
+	$(MAKE)
 
 clean:
 	@-cmd /c del /Q /S *.OBJ *.LIB *.EXE *.DLL 2>nul
@@ -84,7 +91,3 @@ install: $(LIBMAIN_S)
 	cmd /c if not exist "$(PREFIX)\include" mkdir "$(PREFIX)\include"
 	copy /Y $(LIBMAIN_S) "$(PREFIX)\lib"
 	copy /Y tommath*.h "$(PREFIX)\include"
-
-# ref:         $Format:%D$
-# git commit:  $Format:%H$
-# commit time: $Format:%ai$
diff --git a/libtommath/makefile.shared b/libtommath/makefile.shared
index 3955f83..6802107 100644
--- a/libtommath/makefile.shared
+++ b/libtommath/makefile.shared
@@ -18,48 +18,52 @@ ifndef LIBTOOL
   endif
 endif
 LTCOMPILE = $(LIBTOOL) --mode=compile --tag=CC $(CC)
+LTLINK = $(LIBTOOL) --mode=link --tag=CC $(CC)
 
 LCOV_ARGS=--directory .libs --directory .
 
 #START_INS
-OBJECTS=bn_error.o bn_fast_mp_invmod.o bn_fast_mp_montgomery_reduce.o bn_fast_s_mp_mul_digs.o \
-bn_fast_s_mp_mul_high_digs.o bn_fast_s_mp_sqr.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o \
-bn_mp_addmod.o bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o \
-bn_mp_cmp_mag.o bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_div.o \
-bn_mp_div_2.o bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o \
-bn_mp_dr_setup.o bn_mp_exch.o bn_mp_export.o bn_mp_expt_d.o bn_mp_expt_d_ex.o bn_mp_exptmod.o \
-bn_mp_exptmod_fast.o bn_mp_exteuclid.o bn_mp_fread.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_bit.o \
-bn_mp_get_double.o bn_mp_get_int.o bn_mp_get_long.o bn_mp_get_long_long.o bn_mp_grow.o bn_mp_import.o \
-bn_mp_init.o bn_mp_init_copy.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_set_int.o bn_mp_init_size.o \
-bn_mp_invmod.o bn_mp_invmod_slow.o bn_mp_is_square.o bn_mp_jacobi.o bn_mp_karatsuba_mul.o \
-bn_mp_karatsuba_sqr.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_lshd.o bn_mp_mod.o bn_mp_mod_2d.o bn_mp_mod_d.o \
-bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o bn_mp_montgomery_setup.o bn_mp_mul.o \
-bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_n_root.o bn_mp_n_root_ex.o bn_mp_neg.o \
-bn_mp_or.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o bn_mp_prime_is_divisible.o \
+OBJECTS=bn_cutoffs.o bn_deprecated.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o bn_mp_addmod.o \
+bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o bn_mp_cmp_mag.o \
+bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_decr.o bn_mp_div.o bn_mp_div_2.o \
+bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o bn_mp_dr_setup.o \
+bn_mp_error_to_string.o bn_mp_exch.o bn_mp_expt_u32.o bn_mp_exptmod.o bn_mp_exteuclid.o bn_mp_fread.o \
+bn_mp_from_sbin.o bn_mp_from_ubin.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_double.o bn_mp_get_i32.o \
+bn_mp_get_i64.o bn_mp_get_l.o bn_mp_get_ll.o bn_mp_get_mag_u32.o bn_mp_get_mag_u64.o bn_mp_get_mag_ul.o \
+bn_mp_get_mag_ull.o bn_mp_grow.o bn_mp_incr.o bn_mp_init.o bn_mp_init_copy.o bn_mp_init_i32.o \
+bn_mp_init_i64.o bn_mp_init_l.o bn_mp_init_ll.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_size.o \
+bn_mp_init_u32.o bn_mp_init_u64.o bn_mp_init_ul.o bn_mp_init_ull.o bn_mp_invmod.o bn_mp_is_square.o \
+bn_mp_iseven.o bn_mp_isodd.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_log_u32.o bn_mp_lshd.o bn_mp_mod.o \
+bn_mp_mod_2d.o bn_mp_mod_d.o bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o \
+bn_mp_montgomery_setup.o bn_mp_mul.o bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_neg.o \
+bn_mp_or.o bn_mp_pack.o bn_mp_pack_count.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o \
 bn_mp_prime_is_prime.o bn_mp_prime_miller_rabin.o bn_mp_prime_next_prime.o \
-bn_mp_prime_rabin_miller_trials.o bn_mp_prime_random_ex.o bn_mp_prime_strong_lucas_selfridge.o \
-bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_read_signed_bin.o \
-bn_mp_read_unsigned_bin.o bn_mp_reduce.o bn_mp_reduce_2k.o bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o \
-bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_rshd.o \
-bn_mp_set.o bn_mp_set_double.o bn_mp_set_int.o bn_mp_set_long.o bn_mp_set_long_long.o bn_mp_shrink.o \
-bn_mp_signed_bin_size.o bn_mp_sqr.o bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o \
-bn_mp_sub_d.o bn_mp_submod.o bn_mp_tc_and.o bn_mp_tc_div_2d.o bn_mp_tc_or.o bn_mp_tc_xor.o \
-bn_mp_to_signed_bin.o bn_mp_to_signed_bin_n.o bn_mp_to_unsigned_bin.o bn_mp_to_unsigned_bin_n.o \
-bn_mp_toom_mul.o bn_mp_toom_sqr.o bn_mp_toradix.o bn_mp_toradix_n.o bn_mp_unsigned_bin_size.o bn_mp_xor.o \
-bn_mp_zero.o bn_prime_tab.o bn_reverse.o bn_s_mp_add.o bn_s_mp_exptmod.o bn_s_mp_mul_digs.o \
-bn_s_mp_mul_high_digs.o bn_s_mp_sqr.o bn_s_mp_sub.o bncore.o
+bn_mp_prime_rabin_miller_trials.o bn_mp_prime_rand.o bn_mp_prime_strong_lucas_selfridge.o \
+bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_reduce.o bn_mp_reduce_2k.o \
+bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o \
+bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_root_u32.o bn_mp_rshd.o bn_mp_sbin_size.o bn_mp_set.o \
+bn_mp_set_double.o bn_mp_set_i32.o bn_mp_set_i64.o bn_mp_set_l.o bn_mp_set_ll.o bn_mp_set_u32.o \
+bn_mp_set_u64.o bn_mp_set_ul.o bn_mp_set_ull.o bn_mp_shrink.o bn_mp_signed_rsh.o bn_mp_sqr.o \
+bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o bn_mp_sub_d.o bn_mp_submod.o \
+bn_mp_to_radix.o bn_mp_to_sbin.o bn_mp_to_ubin.o bn_mp_ubin_size.o bn_mp_unpack.o bn_mp_xor.o bn_mp_zero.o \
+bn_prime_tab.o bn_s_mp_add.o bn_s_mp_balance_mul.o bn_s_mp_exptmod.o bn_s_mp_exptmod_fast.o \
+bn_s_mp_get_bit.o bn_s_mp_invmod_fast.o bn_s_mp_invmod_slow.o bn_s_mp_karatsuba_mul.o \
+bn_s_mp_karatsuba_sqr.o bn_s_mp_montgomery_reduce_fast.o bn_s_mp_mul_digs.o bn_s_mp_mul_digs_fast.o \
+bn_s_mp_mul_high_digs.o bn_s_mp_mul_high_digs_fast.o bn_s_mp_prime_is_divisible.o \
+bn_s_mp_rand_jenkins.o bn_s_mp_rand_platform.o bn_s_mp_reverse.o bn_s_mp_sqr.o bn_s_mp_sqr_fast.o \
+bn_s_mp_sub.o bn_s_mp_toom_mul.o bn_s_mp_toom_sqr.o
 
 #END_INS
 
 objs: $(OBJECTS)
 
-.c.o:
-	$(LTCOMPILE) $(CFLAGS) $(LDFLAGS) -o $@ -c $<
+.c.o: $(HEADERS)
+	$(LTCOMPILE) $(LTM_CFLAGS) $(LTM_LDFLAGS) -o $@ -c $<
 
 LOBJECTS = $(OBJECTS:.o=.lo)
 
 $(LIBNAME):  $(OBJECTS)
-	$(LIBTOOL) --mode=link --tag=CC $(CC) $(LDFLAGS) $(LOBJECTS) -o $(LIBNAME) -rpath $(LIBPATH) -version-info $(VERSION_SO) $(LIBTOOLFLAGS)
+	$(LTLINK) $(LTM_LDFLAGS) $(LOBJECTS) -o $(LIBNAME) -rpath $(LIBPATH) -version-info $(VERSION_SO) $(LTM_LIBTOOLFLAGS)
 
 install: $(LIBNAME)
 	install -d $(DESTDIR)$(LIBPATH)
@@ -75,17 +79,21 @@ uninstall:
 	rm $(HEADERS_PUB:%=$(DESTDIR)$(INCPATH)/%)
 	rm $(DESTDIR)$(LIBPATH)/pkgconfig/libtommath.pc
 
-test: $(LIBNAME) demo/demo.o
-	$(CC) $(CFLAGS) -c demo/demo.c -o demo/demo.o
-	$(LIBTOOL) --mode=link $(CC) $(LDFLAGS) -o test demo/demo.o $(LIBNAME)
+test_standalone: test
+	@echo "test_standalone is deprecated, please use make-target 'test'"
 
-test_standalone: $(LIBNAME) demo/demo.o
-	$(CC) $(CFLAGS) -c demo/demo.c -o demo/demo.o
-	$(LIBTOOL) --mode=link $(CC) $(LDFLAGS) -o test demo/demo.o $(LIBNAME)
+test mtest_opponent: demo/shared.o $(LIBNAME) | demo/test.o demo/mtest_opponent.o
+	$(LTLINK) $(LTM_LDFLAGS) demo/$@.o $^ -o $@
 
 .PHONY: mtest
 mtest:
-	cd mtest ; $(CC) $(CFLAGS) $(LDFLAGS) mtest.c -o mtest
+	cd mtest ; $(CC) $(LTM_CFLAGS) -O0 mtest.c $(LTM_LDFLAGS) -o mtest
 
 timing: $(LIBNAME) demo/timing.c
-	$(LIBTOOL) --mode=link $(CC) $(CFLAGS) $(LDFLAGS) -DTIMER demo/timing.c $(LIBNAME) -o timing
+	$(LTLINK) $(LTM_CFLAGS) $(LTM_LDFLAGS) -DTIMER demo/timing.c $(LIBNAME) -o timing
+
+tune: $(LIBNAME)
+	$(LTCOMPILE) $(LTM_CFLAGS) -c etc/tune.c -o etc/tune.o
+	$(LTLINK) $(LTM_LDFLAGS) -o etc/tune etc/tune.o $(LIBNAME)
+	cd etc/; /bin/sh tune_it.sh; cd ..
+	$(MAKE) -f makefile.shared
diff --git a/libtommath/makefile.unix b/libtommath/makefile.unix
index b89cf47..4cefc7e 100644
--- a/libtommath/makefile.unix
+++ b/libtommath/makefile.unix
@@ -21,7 +21,7 @@ RANLIB    = ranlib
 CFLAGS    = -O2
 LDFLAGS   =
 
-VERSION   = 1.1.0
+VERSION   = 1.2.0
 
 #Compilation flags
 LTM_CFLAGS  = -I. $(CFLAGS)
@@ -30,36 +30,38 @@ LTM_LDFLAGS = $(LDFLAGS)
 #Library to be created (this makefile builds only static library)
 LIBMAIN_S = libtommath.a
 
-OBJECTS=bn_error.o bn_fast_mp_invmod.o bn_fast_mp_montgomery_reduce.o bn_fast_s_mp_mul_digs.o \
-bn_fast_s_mp_mul_high_digs.o bn_fast_s_mp_sqr.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o \
-bn_mp_addmod.o bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o \
-bn_mp_cmp_mag.o bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_div.o \
-bn_mp_div_2.o bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o \
-bn_mp_dr_setup.o bn_mp_exch.o bn_mp_export.o bn_mp_expt_d.o bn_mp_expt_d_ex.o bn_mp_exptmod.o \
-bn_mp_exptmod_fast.o bn_mp_exteuclid.o bn_mp_fread.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_bit.o \
-bn_mp_get_double.o bn_mp_get_int.o bn_mp_get_long.o bn_mp_get_long_long.o bn_mp_grow.o bn_mp_import.o \
-bn_mp_init.o bn_mp_init_copy.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_set_int.o bn_mp_init_size.o \
-bn_mp_invmod.o bn_mp_invmod_slow.o bn_mp_is_square.o bn_mp_jacobi.o bn_mp_karatsuba_mul.o \
-bn_mp_karatsuba_sqr.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_lshd.o bn_mp_mod.o bn_mp_mod_2d.o bn_mp_mod_d.o \
-bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o bn_mp_montgomery_setup.o bn_mp_mul.o \
-bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_n_root.o bn_mp_n_root_ex.o bn_mp_neg.o \
-bn_mp_or.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o bn_mp_prime_is_divisible.o \
+OBJECTS=bn_cutoffs.o bn_deprecated.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o bn_mp_addmod.o \
+bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o bn_mp_cmp_mag.o \
+bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_decr.o bn_mp_div.o bn_mp_div_2.o \
+bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o bn_mp_dr_setup.o \
+bn_mp_error_to_string.o bn_mp_exch.o bn_mp_expt_u32.o bn_mp_exptmod.o bn_mp_exteuclid.o bn_mp_fread.o \
+bn_mp_from_sbin.o bn_mp_from_ubin.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_double.o bn_mp_get_i32.o \
+bn_mp_get_i64.o bn_mp_get_l.o bn_mp_get_ll.o bn_mp_get_mag_u32.o bn_mp_get_mag_u64.o bn_mp_get_mag_ul.o \
+bn_mp_get_mag_ull.o bn_mp_grow.o bn_mp_incr.o bn_mp_init.o bn_mp_init_copy.o bn_mp_init_i32.o \
+bn_mp_init_i64.o bn_mp_init_l.o bn_mp_init_ll.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_size.o \
+bn_mp_init_u32.o bn_mp_init_u64.o bn_mp_init_ul.o bn_mp_init_ull.o bn_mp_invmod.o bn_mp_is_square.o \
+bn_mp_iseven.o bn_mp_isodd.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_log_u32.o bn_mp_lshd.o bn_mp_mod.o \
+bn_mp_mod_2d.o bn_mp_mod_d.o bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o \
+bn_mp_montgomery_setup.o bn_mp_mul.o bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_neg.o \
+bn_mp_or.o bn_mp_pack.o bn_mp_pack_count.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o \
 bn_mp_prime_is_prime.o bn_mp_prime_miller_rabin.o bn_mp_prime_next_prime.o \
-bn_mp_prime_rabin_miller_trials.o bn_mp_prime_random_ex.o bn_mp_prime_strong_lucas_selfridge.o \
-bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_read_signed_bin.o \
-bn_mp_read_unsigned_bin.o bn_mp_reduce.o bn_mp_reduce_2k.o bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o \
-bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_rshd.o \
-bn_mp_set.o bn_mp_set_double.o bn_mp_set_int.o bn_mp_set_long.o bn_mp_set_long_long.o bn_mp_shrink.o \
-bn_mp_signed_bin_size.o bn_mp_sqr.o bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o \
-bn_mp_sub_d.o bn_mp_submod.o bn_mp_tc_and.o bn_mp_tc_div_2d.o bn_mp_tc_or.o bn_mp_tc_xor.o \
-bn_mp_to_signed_bin.o bn_mp_to_signed_bin_n.o bn_mp_to_unsigned_bin.o bn_mp_to_unsigned_bin_n.o \
-bn_mp_toom_mul.o bn_mp_toom_sqr.o bn_mp_toradix.o bn_mp_toradix_n.o bn_mp_unsigned_bin_size.o bn_mp_xor.o \
-bn_mp_zero.o bn_prime_tab.o bn_reverse.o bn_s_mp_add.o bn_s_mp_exptmod.o bn_s_mp_mul_digs.o \
-bn_s_mp_mul_high_digs.o bn_s_mp_sqr.o bn_s_mp_sub.o bncore.o
-
-HEADERS_PUB=tommath.h tommath_class.h tommath_superclass.h
-
-HEADERS=tommath_private.h $(HEADERS_PUB)
+bn_mp_prime_rabin_miller_trials.o bn_mp_prime_rand.o bn_mp_prime_strong_lucas_selfridge.o \
+bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_reduce.o bn_mp_reduce_2k.o \
+bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o \
+bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_root_u32.o bn_mp_rshd.o bn_mp_sbin_size.o bn_mp_set.o \
+bn_mp_set_double.o bn_mp_set_i32.o bn_mp_set_i64.o bn_mp_set_l.o bn_mp_set_ll.o bn_mp_set_u32.o \
+bn_mp_set_u64.o bn_mp_set_ul.o bn_mp_set_ull.o bn_mp_shrink.o bn_mp_signed_rsh.o bn_mp_sqr.o \
+bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o bn_mp_sub_d.o bn_mp_submod.o \
+bn_mp_to_radix.o bn_mp_to_sbin.o bn_mp_to_ubin.o bn_mp_ubin_size.o bn_mp_unpack.o bn_mp_xor.o bn_mp_zero.o \
+bn_prime_tab.o bn_s_mp_add.o bn_s_mp_balance_mul.o bn_s_mp_exptmod.o bn_s_mp_exptmod_fast.o \
+bn_s_mp_get_bit.o bn_s_mp_invmod_fast.o bn_s_mp_invmod_slow.o bn_s_mp_karatsuba_mul.o \
+bn_s_mp_karatsuba_sqr.o bn_s_mp_montgomery_reduce_fast.o bn_s_mp_mul_digs.o bn_s_mp_mul_digs_fast.o \
+bn_s_mp_mul_high_digs.o bn_s_mp_mul_high_digs_fast.o bn_s_mp_prime_is_divisible.o \
+bn_s_mp_rand_jenkins.o bn_s_mp_rand_platform.o bn_s_mp_reverse.o bn_s_mp_sqr.o bn_s_mp_sqr_fast.o \
+bn_s_mp_sub.o bn_s_mp_toom_mul.o bn_s_mp_toom_sqr.o
+
+HEADERS_PUB=tommath.h
+HEADERS=tommath_private.h tommath_class.h tommath_superclass.h tommath_cutoffs.h $(HEADERS_PUB)
 
 #The default rule for make builds the libtommath.a library (static)
 default: $(LIBMAIN_S)
@@ -78,18 +80,23 @@ $(LIBMAIN_S): $(OBJECTS)
 	$(RANLIB) $@
 
 #Build test_standalone suite
-test: $(LIBMAIN_S) demo/demo.c
-	$(CC) $(LTM_CFLAGS) $(LTM_LDFLAGS) demo/demo.c $(LIBMAIN_S) -DLTM_DEMO_TEST_VS_MTEST=0 -o $@
+test: demo/shared.o demo/test.o $(LIBMAIN_S)
+	$(CC) $(LTM_CFLAGS) $(LTM_LDFLAGS) $^ -o $@
 	@echo "NOTICE: start the tests by: ./test"
 
 test_standalone: test
+	@echo "test_standalone is deprecated, please use make-target 'test'"
 
-all: $(LIBMAIN_S) test_standalone
+all: $(LIBMAIN_S) test
+
+tune: $(LIBMAIN_S)
+	$(MAKE) -C etc tune
+	$(MAKE)
 
 #NOTE: this makefile works also on cygwin, thus we need to delete *.exe
 clean:
 	-@rm -f $(OBJECTS) $(LIBMAIN_S)
-	-@rm -f demo/demo.o test test.exe
+	-@rm -f demo/main.o demo/opponent.o demo/test.o test test.exe
 
 #Install the library + headers
 install: $(LIBMAIN_S)
@@ -97,7 +104,3 @@ install: $(LIBMAIN_S)
 	@cp $(LIBMAIN_S) $(DESTDIR)$(LIBPATH)/
 	@cp $(HEADERS_PUB) $(DESTDIR)$(INCPATH)/
 	@sed -e 's,^prefix=.*,prefix=$(PREFIX),' -e 's,^Version:.*,Version: $(VERSION),' libtommath.pc.in > $(DESTDIR)$(LIBPATH)/pkgconfig/libtommath.pc
-
-# ref:         $Format:%D$
-# git commit:  $Format:%H$
-# commit time: $Format:%ai$
diff --git a/libtommath/makefile_include.mk b/libtommath/makefile_include.mk
index ec2205b..7b025e8 100644
--- a/libtommath/makefile_include.mk
+++ b/libtommath/makefile_include.mk
@@ -3,9 +3,9 @@
 #
 
 #version of library
-VERSION=1.1.0
-VERSION_PC=1.1.0
-VERSION_SO=2:0:1
+VERSION=1.2.0
+VERSION_PC=1.2.0
+VERSION_SO=3:0:2
 
 PLATFORM := $(shell uname | sed -e 's/_.*//')
 
@@ -47,52 +47,71 @@ else
 endif
 endif
 
-CFLAGS += -I./ -Wall -Wsign-compare -Wextra -Wshadow
+LTM_CFLAGS += -I./ -Wall -Wsign-compare -Wextra -Wshadow
+
+ifdef SANITIZER
+LTM_CFLAGS += -fsanitize=undefined -fno-sanitize-recover=all -fno-sanitize=float-divide-by-zero
+endif
 
 ifndef NO_ADDTL_WARNINGS
 # additional warnings
-CFLAGS += -Wsystem-headers -Wdeclaration-after-statement -Wbad-function-cast -Wcast-align
-CFLAGS += -Wstrict-prototypes -Wpointer-arith
+LTM_CFLAGS += -Wdeclaration-after-statement -Wbad-function-cast -Wcast-align
+LTM_CFLAGS += -Wstrict-prototypes -Wpointer-arith
+endif
+
+ifdef CONV_WARNINGS
+LTM_CFLAGS += -std=c89 -Wconversion -Wsign-conversion
+ifeq ($(CONV_WARNINGS), strict)
+LTM_CFLAGS += -DMP_USE_ENUMS -Wc++-compat
+endif
+else
+LTM_CFLAGS += -Wsystem-headers
 endif
 
 ifdef COMPILE_DEBUG
 #debug
-CFLAGS += -g3
-else
+LTM_CFLAGS += -g3
+endif
 
 ifdef COMPILE_SIZE
 #for size
-CFLAGS += -Os
+LTM_CFLAGS += -Os
 else
 
 ifndef IGNORE_SPEED
 #for speed
-CFLAGS += -O3 -funroll-loops
+LTM_CFLAGS += -O3 -funroll-loops
 
 #x86 optimizations [should be valid for any GCC install though]
-CFLAGS  += -fomit-frame-pointer
+LTM_CFLAGS  += -fomit-frame-pointer
 endif
 
 endif # COMPILE_SIZE
-endif # COMPILE_DEBUG
 
 ifneq ($(findstring clang,$(CC)),)
-CFLAGS += -Wno-typedef-redefinition -Wno-tautological-compare -Wno-builtin-requires-header
+LTM_CFLAGS += -Wno-typedef-redefinition -Wno-tautological-compare -Wno-builtin-requires-header
 endif
 ifneq ($(findstring mingw,$(CC)),)
-CFLAGS += -Wno-shadow
+LTM_CFLAGS += -Wno-shadow
 endif
 ifeq ($(PLATFORM), Darwin)
-CFLAGS += -Wno-nullability-completeness
+LTM_CFLAGS += -Wno-nullability-completeness
 endif
 ifeq ($(PLATFORM), CYGWIN)
 LIBTOOLFLAGS += -no-undefined
 endif
 
+# add in the standard FLAGS
+LTM_CFLAGS += $(CFLAGS)
+LTM_LFLAGS += $(LFLAGS)
+LTM_LDFLAGS += $(LDFLAGS)
+LTM_LIBTOOLFLAGS += $(LIBTOOLFLAGS)
+
+
 ifeq ($(PLATFORM),FreeBSD)
   _ARCH := $(shell sysctl -b hw.machine_arch)
 else
-  _ARCH := $(shell arch)
+  _ARCH := $(shell uname -m)
 endif
 
 # adjust coverage set
@@ -104,10 +123,8 @@ else
    COVERAGE_APP = ./test
 endif
 
-HEADERS_PUB=tommath.h tommath_class.h tommath_superclass.h
-HEADERS=tommath_private.h $(HEADERS_PUB)
-
-test_standalone: CFLAGS+=-DLTM_DEMO_TEST_VS_MTEST=0
+HEADERS_PUB=tommath.h
+HEADERS=tommath_private.h tommath_class.h tommath_superclass.h tommath_cutoffs.h $(HEADERS_PUB)
 
 #LIBPATH  The directory for libtommath to be installed to.
 #INCPATH  The directory to install the header files for libtommath.
@@ -120,9 +137,9 @@ DATAPATH ?= $(PREFIX)/share/doc/libtommath/pdf
 
 #make the code coverage of the library
 #
-coverage: CFLAGS += -fprofile-arcs -ftest-coverage -DTIMING_NO_LOGS
-coverage: LFLAGS += -lgcov
-coverage: LDFLAGS += -lgcov
+coverage: LTM_CFLAGS += -fprofile-arcs -ftest-coverage -DTIMING_NO_LOGS
+coverage: LTM_LFLAGS += -lgcov
+coverage: LTM_LDFLAGS += -lgcov
 
 coverage: $(COVERAGE)
 	$(COVERAGE_APP)
@@ -141,8 +158,9 @@ cleancov-clean:
 cleancov: cleancov-clean clean
 
 clean:
-	rm -f *.gcda *.gcno *.gcov *.bat *.o *.a *.obj *.lib *.exe *.dll etclib/*.o demo/demo.o test timing mpitest mtest/mtest mtest/mtest.exe \
-        *.idx *.toc *.log *.aux *.dvi *.lof *.ind *.ilg *.ps *.log *.s mpi.c *.da *.dyn *.dpi tommath.tex `find . -type f | grep [~] | xargs` *.lo *.la
-	rm -rf .libs/
+	rm -f *.gcda *.gcno *.gcov *.bat *.o *.a *.obj *.lib *.exe *.dll etclib/*.o \
+				demo/*.o test timing mtest_opponent mtest/mtest mtest/mtest.exe tuning_list \
+				*.s mpi.c *.da *.dyn *.dpi tommath.tex `find . -type f | grep [~] | xargs` *.lo *.la
+	rm -rf .libs/ demo/.libs
 	${MAKE} -C etc/ clean MAKE=${MAKE}
 	${MAKE} -C doc/ clean MAKE=${MAKE}
diff --git a/libtommath/tommath.def b/libtommath/tommath.def
new file mode 100644
index 0000000..229fae4
--- /dev/null
+++ b/libtommath/tommath.def
@@ -0,0 +1,145 @@
+; libtommath
+;
+; Use this command to produce a 32-bit .lib file, for use in any MSVC version
+;   lib -machine:X86 -name:libtommath.dll -def:tommath.def -out:tommath.lib
+; Use this command to produce a 64-bit .lib file, for use in any MSVC version
+;   lib -machine:X64 -name:libtommath.dll -def:tommath.def -out:tommath.lib
+;
+EXPORTS
+    mp_2expt
+    mp_abs
+    mp_add
+    mp_add_d
+    mp_addmod
+    mp_and
+    mp_clamp
+    mp_clear
+    mp_clear_multi
+    mp_cmp
+    mp_cmp_d
+    mp_cmp_mag
+    mp_cnt_lsb
+    mp_complement
+    mp_copy
+    mp_count_bits
+    mp_decr
+    mp_div
+    mp_div_2
+    mp_div_2d
+    mp_div_3
+    mp_div_d
+    mp_dr_is_modulus
+    mp_dr_reduce
+    mp_dr_setup
+    mp_error_to_string
+    mp_exch
+    mp_expt_u32
+    mp_exptmod
+    mp_exteuclid
+    mp_fread
+    mp_from_sbin
+    mp_from_ubin
+    mp_fwrite
+    mp_gcd
+    mp_get_double
+    mp_get_i32
+    mp_get_i64
+    mp_get_int
+    mp_get_l
+    mp_get_ll
+    mp_get_long
+    mp_get_long_long
+    mp_get_mag_u32
+    mp_get_mag_u64
+    mp_get_mag_ul
+    mp_get_mag_ull
+    mp_grow
+    mp_incr
+    mp_init
+    mp_init_copy
+    mp_init_i32
+    mp_init_i64
+    mp_init_l
+    mp_init_ll
+    mp_init_multi
+    mp_init_set
+    mp_init_set_int
+    mp_init_size
+    mp_init_u32
+    mp_init_u64
+    mp_init_ul
+    mp_init_ull
+    mp_invmod
+    mp_is_square
+    mp_iseven
+    mp_isodd
+    mp_kronecker
+    mp_lcm
+    mp_log_u32
+    mp_lshd
+    mp_mod
+    mp_mod_2d
+    mp_mod_d
+    mp_montgomery_calc_normalization
+    mp_montgomery_reduce
+    mp_montgomery_setup
+    mp_mul
+    mp_mul_2
+    mp_mul_2d
+    mp_mul_d
+    mp_mulmod
+    mp_neg
+    mp_or
+    mp_pack
+    mp_pack_count
+    mp_prime_fermat
+    mp_prime_frobenius_underwood
+    mp_prime_is_prime
+    mp_prime_miller_rabin
+    mp_prime_next_prime
+    mp_prime_rabin_miller_trials
+    mp_prime_rand
+    mp_prime_strong_lucas_selfridge
+    mp_radix_size
+    mp_rand
+    mp_read_radix
+    mp_reduce
+    mp_reduce_2k
+    mp_reduce_2k_l
+    mp_reduce_2k_setup
+    mp_reduce_2k_setup_l
+    mp_reduce_is_2k
+    mp_reduce_is_2k_l
+    mp_reduce_setup
+    mp_root_u32
+    mp_rshd
+    mp_sbin_size
+    mp_set
+    mp_set_double
+    mp_set_i32
+    mp_set_i64
+    mp_set_int
+    mp_set_l
+    mp_set_ll
+    mp_set_long
+    mp_set_long_long
+    mp_set_u32
+    mp_set_u64
+    mp_set_ul
+    mp_set_ull
+    mp_shrink
+    mp_signed_rsh
+    mp_sqr
+    mp_sqrmod
+    mp_sqrt
+    mp_sqrtmod_prime
+    mp_sub
+    mp_sub_d
+    mp_submod
+    mp_to_radix
+    mp_to_sbin
+    mp_to_ubin
+    mp_ubin_size
+    mp_unpack
+    mp_xor
+    mp_zero
diff --git a/libtommath/tommath.h b/libtommath/tommath.h
index 5fd149c..63f8190 100644
--- a/libtommath/tommath.h
+++ b/libtommath/tommath.h
@@ -4,18 +4,35 @@
 #ifndef BN_H_
 #define BN_H_
 
-#include <stdio.h>
-#include <stdlib.h>
+#ifndef MP_NO_STDINT
+#  include <stdint.h>
+#endif
+#include <stddef.h>
 #include <limits.h>
 
-#include "tommath_class.h"
+#ifdef LTM_NO_FILE
+#  warning LTM_NO_FILE has been deprecated, use MP_NO_FILE.
+#  define MP_NO_FILE
+#endif
+
+#ifndef MP_NO_FILE
+#  include <stdio.h>
+#endif
+
+#ifdef MP_8BIT
+#  ifdef _MSC_VER
+#    pragma message("8-bit (MP_8BIT) support is deprecated and will be dropped completely in the next version.")
+#  else
+#    warning "8-bit (MP_8BIT) support is deprecated and will be dropped completely in the next version."
+#  endif
+#endif
 
 #ifdef __cplusplus
 extern "C" {
 #endif
 
 /* MS Visual C++ doesn't have a 128bit type for words, so fall back to 32bit MPI's (where words are 64bit) */
-#if defined(_WIN32) || defined(__LLP64__) || defined(__e2k__) || defined(__LCC__)
+#if (defined(_WIN32) || defined(__LLP64__) || defined(__e2k__) || defined(__LCC__)) && !defined(MP_64BIT)
 #   define MP_32BIT
 #endif
 
@@ -26,8 +43,8 @@ extern "C" {
     defined(__sparcv9) || defined(__sparc_v9__) || defined(__sparc64__) || \
     defined(__ia64) || defined(__ia64__) || defined(__itanium__) || defined(_M_IA64) || \
     defined(__LP64__) || defined(_LP64) || defined(__64BIT__)
-#   if !(defined(MP_32BIT) || defined(MP_16BIT) || defined(MP_8BIT))
-#      if defined(__GNUC__)
+#   if !(defined(MP_64BIT) || defined(MP_32BIT) || defined(MP_16BIT) || defined(MP_8BIT))
+#      if defined(__GNUC__) && !defined(__hppa)
 /* we support 128bit integers only via: __attribute__((mode(TI))) */
 #         define MP_64BIT
 #      else
@@ -37,77 +54,127 @@ extern "C" {
 #   endif
 #endif
 
-typedef unsigned long long Tcl_WideUInt;
+#ifdef MP_DIGIT_BIT
+#   error Defining MP_DIGIT_BIT is disallowed, use MP_8/16/31/32/64BIT
+#endif
 
 /* some default configurations.
  *
- * A "mp_digit" must be able to hold DIGIT_BIT + 1 bits
- * A "mp_word" must be able to hold 2*DIGIT_BIT + 1 bits
+ * A "mp_digit" must be able to hold MP_DIGIT_BIT + 1 bits
+ * A "mp_word" must be able to hold 2*MP_DIGIT_BIT + 1 bits
  *
  * At the very least a mp_digit must be able to hold 7 bits
  * [any size beyond that is ok provided it doesn't overflow the data type]
  */
+
 #ifdef MP_8BIT
 typedef unsigned char        mp_digit;
 typedef unsigned short       mp_word;
-#   define MP_SIZEOF_MP_DIGIT 1
-#   ifdef DIGIT_BIT
-#      error You must not define DIGIT_BIT when using MP_8BIT
-#   endif
+#   define MP_DIGIT_BIT 7
 #elif defined(MP_16BIT)
 typedef unsigned short       mp_digit;
 typedef unsigned int         mp_word;
-#   define MP_SIZEOF_MP_DIGIT 2
-#   ifdef DIGIT_BIT
-#      error You must not define DIGIT_BIT when using MP_16BIT
-#   endif
+#   define MP_DIGIT_BIT 15
 #elif defined(MP_64BIT)
 /* for GCC only on supported platforms */
-typedef unsigned long long   mp_digit;
+typedef Tcl_WideUInt   mp_digit;
+#if defined(__GNUC__)
 typedef unsigned long        mp_word __attribute__((mode(TI)));
-#   define DIGIT_BIT 60
+#endif
+#   define MP_DIGIT_BIT 60
 #else
-/* this is the default case, 28-bit digits */
-
-/* this is to make porting into LibTomCrypt easier :-) */
 typedef unsigned int         mp_digit;
-typedef unsigned long long   mp_word;
-
+typedef Tcl_WideUInt   mp_word;
 #   ifdef MP_31BIT
-/* this is an extension that uses 31-bit digits */
-#      define DIGIT_BIT 31
+/*
+ * This is an extension that uses 31-bit digits.
+ * Please be aware that not all functions support this size, especially s_mp_mul_digs_fast
+ * will be reduced to work on small numbers only:
+ * Up to 8 limbs, 248 bits instead of up to 512 limbs, 15872 bits with MP_28BIT.
+ */
+#      define MP_DIGIT_BIT 31
 #   else
 /* default case is 28-bit digits, defines MP_28BIT as a handy macro to test */
-#      define DIGIT_BIT 28
+#      define MP_DIGIT_BIT 28
 #      define MP_28BIT
 #   endif
 #endif
 
-#define MP_MASK          ((((mp_digit)1)<<((mp_digit)DIGIT_BIT))-((mp_digit)1))
+#define MP_MASK          ((((mp_digit)1)<<((mp_digit)MP_DIGIT_BIT))-((mp_digit)1))
+#define MP_DIGIT_MAX     MP_MASK
 
-/* equalities */
+/* Primality generation flags */
+#define MP_PRIME_BBS      0x0001 /* BBS style prime */
+#define MP_PRIME_SAFE     0x0002 /* Safe prime (p-1)/2 == prime */
+#define MP_PRIME_2MSB_ON  0x0008 /* force 2nd MSB to 1 */
+
+#ifdef MP_USE_ENUMS
+typedef enum {
+   MP_ZPOS = 0,   /* positive */
+   MP_NEG = 1     /* negative */
+} mp_sign;
+typedef enum {
+   MP_LT = -1,    /* less than */
+   MP_EQ = 0,     /* equal */
+   MP_GT = 1      /* greater than */
+} mp_ord;
+typedef enum {
+   MP_NO = 0,
+   MP_YES = 1
+} mp_bool;
+typedef enum {
+   MP_OKAY  = 0,   /* no error */
+   MP_ERR   = -1,  /* unknown error */
+   MP_MEM   = -2,  /* out of mem */
+   MP_VAL   = -3,  /* invalid input */
+   MP_ITER  = -4,  /* maximum iterations reached */
+   MP_BUF   = -5   /* buffer overflow, supplied buffer too small */
+} mp_err;
+typedef enum {
+   MP_LSB_FIRST = -1,
+   MP_MSB_FIRST =  1
+} mp_order;
+typedef enum {
+   MP_LITTLE_ENDIAN  = -1,
+   MP_NATIVE_ENDIAN  =  0,
+   MP_BIG_ENDIAN     =  1
+} mp_endian;
+#else
+typedef int mp_sign;
+#define MP_ZPOS       0   /* positive integer */
+#define MP_NEG        1   /* negative */
+typedef int mp_ord;
 #define MP_LT        -1   /* less than */
 #define MP_EQ         0   /* equal to */
 #define MP_GT         1   /* greater than */
-
-#define MP_ZPOS       0   /* positive integer */
-#define MP_NEG        1   /* negative */
-
-#define MP_OKAY       0   /* ok result */
+typedef int mp_bool;
+#define MP_YES        1
+#define MP_NO         0
+typedef int mp_err;
+#define MP_OKAY       0   /* no error */
+#define MP_ERR        -1  /* unknown error */
 #define MP_MEM        -2  /* out of mem */
 #define MP_VAL        -3  /* invalid input */
-#define MP_RANGE      MP_VAL
-#define MP_ITER       -4  /* Max. iterations reached */
-
-#define MP_YES        1   /* yes response */
-#define MP_NO         0   /* no response */
+#define MP_ITER       -4  /* maximum iterations reached */
+#define MP_BUF        -5  /* buffer overflow, supplied buffer too small */
+typedef int mp_order;
+#define MP_LSB_FIRST -1
+#define MP_MSB_FIRST  1
+typedef int mp_endian;
+#define MP_LITTLE_ENDIAN  -1
+#define MP_NATIVE_ENDIAN  0
+#define MP_BIG_ENDIAN     1
+#endif
 
-/* Primality generation flags */
-#define LTM_PRIME_BBS      0x0001 /* BBS style prime */
-#define LTM_PRIME_SAFE     0x0002 /* Safe prime (p-1)/2 == prime */
-#define LTM_PRIME_2MSB_ON  0x0008 /* force 2nd MSB to 1 */
+/* tunable cutoffs */
 
-typedef int           mp_err;
+#ifndef MP_FIXED_CUTOFFS
+extern int
+KARATSUBA_MUL_CUTOFF,
+KARATSUBA_SQR_CUTOFF,
+TOOM_MUL_CUTOFF,
+TOOM_SQR_CUTOFF;
+#endif
 
 /* define this to use lower memory usage routines (exptmods mostly) */
 /* #define MP_LOW_MEM */
@@ -115,106 +182,204 @@ typedef int           mp_err;
 /* default precision */
 #ifndef MP_PREC
 #   ifndef MP_LOW_MEM
-#      define MP_PREC 32        /* default digits of precision */
+#      define PRIVATE_MP_PREC 32        /* default digits of precision */
+#   elif defined(MP_8BIT)
+#      define PRIVATE_MP_PREC 16        /* default digits of precision */
 #   else
-#      define MP_PREC 8         /* default digits of precision */
+#      define PRIVATE_MP_PREC 8         /* default digits of precision */
 #   endif
+#   define MP_PREC (MP_DEPRECATED_PRAGMA("MP_PREC is an internal macro") PRIVATE_MP_PREC)
 #endif
 
+/* size of comba arrays, should be at least 2 * 2**(BITS_PER_WORD - BITS_PER_DIGIT*2) */
+#define PRIVATE_MP_WARRAY (int)(1 << (((CHAR_BIT * (int)sizeof(mp_word)) - (2 * MP_DIGIT_BIT)) + 1))
+
+#if defined(__GNUC__) && __GNUC__ >= 4
+#   define MP_NULL_TERMINATED __attribute__((sentinel))
+#else
+#   define MP_NULL_TERMINATED
+#endif
+
+/*
+ * MP_WUR - warn unused result
+ * ---------------------------
+ *
+ * The result of functions annotated with MP_WUR must be
+ * checked and cannot be ignored.
+ *
+ * Most functions in libtommath return an error code.
+ * This error code must be checked in order to prevent crashes or invalid
+ * results.
+ *
+ * If you still want to avoid the error checks for quick and dirty programs
+ * without robustness guarantees, you can `#define MP_WUR` before including
+ * tommath.h, disabling the warnings.
+ */
+#ifndef MP_WUR
+#  if defined(__GNUC__) && __GNUC__ >= 4
+#     define MP_WUR __attribute__((warn_unused_result))
+#  else
+#     define MP_WUR
+#  endif
+#endif
+
+#if defined(__GNUC__) && (__GNUC__ * 100 + __GNUC_MINOR__ >= 405)
+#  define MP_DEPRECATED(x) __attribute__((deprecated("replaced by " #x)))
+#  define PRIVATE_MP_DEPRECATED_PRAGMA(s) _Pragma(#s)
+#  define MP_DEPRECATED_PRAGMA(s) PRIVATE_MP_DEPRECATED_PRAGMA(GCC warning s)
+#elif defined(_MSC_VER) && _MSC_VER >= 1500
+#  define MP_DEPRECATED(x) __declspec(deprecated("replaced by " #x))
+#  define MP_DEPRECATED_PRAGMA(s) __pragma(message(s))
+#else
+#  define MP_DEPRECATED(s)
+#  define MP_DEPRECATED_PRAGMA(s)
+#endif
+
+#define DIGIT_BIT   (MP_DEPRECATED_PRAGMA("DIGIT_BIT macro is deprecated, MP_DIGIT_BIT instead") MP_DIGIT_BIT)
+#define USED(m)     (MP_DEPRECATED_PRAGMA("USED macro is deprecated, use z->used instead") (m)->used)
+#define DIGIT(m, k) (MP_DEPRECATED_PRAGMA("DIGIT macro is deprecated, use z->dp instead") (m)->dp[(k)])
+#define SIGN(m)     (MP_DEPRECATED_PRAGMA("SIGN macro is deprecated, use z->sign instead") (m)->sign)
+
 /* the infamous mp_int structure */
 typedef struct  {
-   int used, alloc, sign;
+   int used, alloc;
+   mp_sign sign;
    mp_digit *dp;
 } mp_int;
 
 /* callback for mp_prime_random, should fill dst with random bytes and return how many read [upto len] */
-typedef int ltm_prime_callback(unsigned char *dst, int len, void *dat);
-
+typedef int private_mp_prime_callback(unsigned char *dst, int len, void *dat);
+typedef private_mp_prime_callback MP_DEPRECATED(mp_rand_source) ltm_prime_callback;
 
 /* error code to char* string */
-const char *mp_error_to_string(int code);
+const char *mp_error_to_string(mp_err code) MP_WUR;
 
 /* ---> init and deinit bignum functions <--- */
 /* init a bignum */
-int mp_init(mp_int *a);
+mp_err mp_init(mp_int *a) MP_WUR;
 
 /* free a bignum */
 void mp_clear(mp_int *a);
 
 /* init a null terminated series of arguments */
-int mp_init_multi(mp_int *mp, ...);
+mp_err mp_init_multi(mp_int *mp, ...) MP_NULL_TERMINATED MP_WUR;
 
 /* clear a null terminated series of arguments */
-void mp_clear_multi(mp_int *mp, ...);
+void mp_clear_multi(mp_int *mp, ...) MP_NULL_TERMINATED;
 
 /* exchange two ints */
 void mp_exch(mp_int *a, mp_int *b);
 
 /* shrink ram required for a bignum */
-mp_err mp_shrink(mp_int *a);
+mp_err mp_shrink(mp_int *a) MP_WUR;
 
 /* grow an int to a given size */
-mp_err mp_grow(mp_int *a, int size);
+mp_err mp_grow(mp_int *a, int size) MP_WUR;
 
 /* init to a given number of digits */
-mp_err mp_init_size(mp_int *a, int size);
+mp_err mp_init_size(mp_int *a, int size) MP_WUR;
 
 /* ---> Basic Manipulations <--- */
 #define mp_iszero(a) (((a)->used == 0) ? MP_YES : MP_NO)
-#define mp_iseven(a) (((a)->used == 0 || (((a)->dp[0] & 1) == 0)) ? MP_YES : MP_NO)
-#define mp_isodd(a)  (((a)->used > 0 && (((a)->dp[0] & 1) == 1)) ? MP_YES : MP_NO)
+mp_bool mp_iseven(const mp_int *a) MP_WUR;
+mp_bool mp_isodd(const mp_int *a) MP_WUR;
 #define mp_isneg(a)  (((a)->sign != MP_ZPOS) ? MP_YES : MP_NO)
 
 /* set to zero */
 void mp_zero(mp_int *a);
 
-/* set to a digit */
+/* get and set doubles */
+double mp_get_double(const mp_int *a) MP_WUR;
+mp_err mp_set_double(mp_int *a, double b) MP_WUR;
+
+/* get integer, set integer and init with integer (int32_t) */
+#ifndef MP_NO_STDINT
+int32_t mp_get_i32(const mp_int *a) MP_WUR;
+void mp_set_i32(mp_int *a, int32_t b);
+mp_err mp_init_i32(mp_int *a, int32_t b) MP_WUR;
+
+/* get integer, set integer and init with integer, behaves like two complement for negative numbers (uint32_t) */
+#define mp_get_u32(a) ((uint32_t)mp_get_i32(a))
+void mp_set_u32(mp_int *a, uint32_t b);
+mp_err mp_init_u32(mp_int *a, uint32_t b) MP_WUR;
+
+/* get integer, set integer and init with integer (int64_t) */
+int64_t mp_get_i64(const mp_int *a) MP_WUR;
+void mp_set_i64(mp_int *a, int64_t b);
+mp_err mp_init_i64(mp_int *a, int64_t b) MP_WUR;
+
+/* get integer, set integer and init with integer, behaves like two complement for negative numbers (uint64_t) */
+#define mp_get_u64(a) ((uint64_t)mp_get_i64(a))
+void mp_set_u64(mp_int *a, uint64_t b);
+mp_err mp_init_u64(mp_int *a, uint64_t b) MP_WUR;
+
+/* get magnitude */
+uint32_t mp_get_mag_u32(const mp_int *a) MP_WUR;
+uint64_t mp_get_mag_u64(const mp_int *a) MP_WUR;
+#endif
+unsigned long mp_get_mag_ul(const mp_int *a) MP_WUR;
+Tcl_WideUInt mp_get_mag_ull(const mp_int *a) MP_WUR;
+
+/* get integer, set integer (long) */
+long mp_get_l(const mp_int *a) MP_WUR;
+void mp_set_l(mp_int *a, long b);
+mp_err mp_init_l(mp_int *a, long b) MP_WUR;
+
+/* get integer, set integer (unsigned long) */
+#define mp_get_ul(a) ((unsigned long)mp_get_l(a))
+void mp_set_ul(mp_int *a, unsigned long b);
+mp_err mp_init_ul(mp_int *a, unsigned long b) MP_WUR;
+
+/* get integer, set integer (Tcl_WideInt) */
+Tcl_WideInt mp_get_ll(const mp_int *a) MP_WUR;
+void mp_set_ll(mp_int *a, Tcl_WideInt b);
+mp_err mp_init_ll(mp_int *a, Tcl_WideInt b) MP_WUR;
+
+/* get integer, set integer (Tcl_WideUInt) */
+#define mp_get_ull(a) ((Tcl_WideUInt)mp_get_ll(a))
+void mp_set_ull(mp_int *a, Tcl_WideUInt b);
+mp_err mp_init_ull(mp_int *a, Tcl_WideUInt b) MP_WUR;
+
+/* set to single unsigned digit, up to MP_DIGIT_MAX */
 void mp_set(mp_int *a, mp_digit b);
+mp_err mp_init_set(mp_int *a, mp_digit b) MP_WUR;
 
-/* set a double */
-int mp_set_double(mp_int *a, double b);
-
-/* set a 32-bit const */
-int mp_set_int(mp_int *a, unsigned long b);
-
-/* set a platform dependent unsigned long value */
-int mp_set_long(mp_int *a, unsigned long b);
-
-/* set a platform dependent unsigned long long value */
-int mp_set_long_long(mp_int *a, unsigned long long b);
-
-/* get a double */
-double mp_get_double(const mp_int *a);
-
-/* get a 32-bit value */
-unsigned long mp_get_int(const mp_int *a);
-
-/* get a platform dependent unsigned long value */
-unsigned long mp_get_long(const mp_int *a);
-
-/* get a platform dependent unsigned long long value */
-unsigned long long mp_get_long_long(const mp_int *a);
-
-/* initialize and set a digit */
-int mp_init_set(mp_int *a, mp_digit b);
-
-/* initialize and set 32-bit value */
-int mp_init_set_int(mp_int *a, unsigned long b);
+/* get integer, set integer and init with integer (deprecated) */
+MP_DEPRECATED(mp_get_mag_u32/mp_get_u32) unsigned long mp_get_int(const mp_int *a) MP_WUR;
+MP_DEPRECATED(mp_get_mag_ul/mp_get_ul) unsigned long mp_get_long(const mp_int *a) MP_WUR;
+MP_DEPRECATED(mp_get_mag_ull/mp_get_ull) Tcl_WideUInt mp_get_long_long(const mp_int *a) MP_WUR;
+MP_DEPRECATED(mp_set_ul) mp_err mp_set_int(mp_int *a, unsigned long b);
+MP_DEPRECATED(mp_set_ul) mp_err mp_set_long(mp_int *a, unsigned long b);
+MP_DEPRECATED(mp_set_ull) mp_err mp_set_long_long(mp_int *a, Tcl_WideUInt b);
+MP_DEPRECATED(mp_init_ul) mp_err mp_init_set_int(mp_int *a, unsigned long b) MP_WUR;
 
 /* copy, b = a */
-int mp_copy(const mp_int *a, mp_int *b);
+mp_err mp_copy(const mp_int *a, mp_int *b) MP_WUR;
 
 /* inits and copies, a = b */
-int mp_init_copy(mp_int *a, const mp_int *b);
+mp_err mp_init_copy(mp_int *a, const mp_int *b) MP_WUR;
 
 /* trim unused digits */
 void mp_clamp(mp_int *a);
 
-/* import binary data */
-int mp_import(mp_int *rop, size_t count, int order, size_t size, int endian, size_t nails, const void *op);
 
 /* export binary data */
-int mp_export(void *rop, size_t *countp, int order, size_t size, int endian, size_t nails, const mp_int *op);
+MP_DEPRECATED(mp_pack) mp_err mp_export(void *rop, size_t *countp, int order, size_t size,
+                                        int endian, size_t nails, const mp_int *op) MP_WUR;
+
+/* import binary data */
+MP_DEPRECATED(mp_unpack) mp_err mp_import(mp_int *rop, size_t count, int order,
+      size_t size, int endian, size_t nails,
+      const void *op) MP_WUR;
+
+/* unpack binary data */
+mp_err mp_unpack(mp_int *rop, size_t count, mp_order order, size_t size, mp_endian endian,
+                 size_t nails, const void *op) MP_WUR;
+
+/* pack binary data */
+size_t mp_pack_count(const mp_int *a, size_t nails, size_t size) MP_WUR;
+mp_err mp_pack(void *rop, size_t maxcount, size_t *written, mp_order order, size_t size,
+               mp_endian endian, size_t nails, const mp_int *op) MP_WUR;
 
 /* ---> digit manipulation <--- */
 
@@ -222,37 +387,43 @@ int mp_export(void *rop, size_t *countp, int order, size_t size, int endian, siz
 void mp_rshd(mp_int *a, int b);
 
 /* left shift by "b" digits */
-int mp_lshd(mp_int *a, int b);
+mp_err mp_lshd(mp_int *a, int b) MP_WUR;
 
 /* c = a / 2**b, implemented as c = a >> b */
-int mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d);
+mp_err mp_div_2d(const mp_int *a, int b, mp_int *c, mp_int *d) MP_WUR;
 
 /* b = a/2 */
-int mp_div_2(const mp_int *a, mp_int *b);
+mp_err mp_div_2(const mp_int *a, mp_int *b) MP_WUR;
+
+/* a/3 => 3c + d == a */
+mp_err mp_div_3(const mp_int *a, mp_int *c, mp_digit *d) MP_WUR;
 
 /* c = a * 2**b, implemented as c = a << b */
-int mp_mul_2d(const mp_int *a, int b, mp_int *c);
+mp_err mp_mul_2d(const mp_int *a, int b, mp_int *c) MP_WUR;
 
 /* b = a*2 */
-int mp_mul_2(const mp_int *a, mp_int *b);
+mp_err mp_mul_2(const mp_int *a, mp_int *b) MP_WUR;
 
 /* c = a mod 2**b */
-int mp_mod_2d(const mp_int *a, int b, mp_int *c);
+mp_err mp_mod_2d(const mp_int *a, int b, mp_int *c) MP_WUR;
 
 /* computes a = 2**b */
-int mp_2expt(mp_int *a, int b);
+mp_err mp_2expt(mp_int *a, int b) MP_WUR;
 
 /* Counts the number of lsbs which are zero before the first zero bit */
-int mp_cnt_lsb(const mp_int *a);
+int mp_cnt_lsb(const mp_int *a) MP_WUR;
 
 /* I Love Earth! */
 
 /* makes a pseudo-random mp_int of a given size */
-int mp_rand(mp_int *a, int digits);
+mp_err mp_rand(mp_int *a, int digits) MP_WUR;
 /* makes a pseudo-random small int of a given size */
-int mp_rand_digit(mp_digit *r);
+MP_DEPRECATED(mp_rand) mp_err mp_rand_digit(mp_digit *r) MP_WUR;
+/* use custom random data source instead of source provided the platform */
+void mp_rand_source(mp_err(*source)(void *out, size_t size));
 
 #ifdef MP_PRNG_ENABLE_LTM_RNG
+#  warning MP_PRNG_ENABLE_LTM_RNG has been deprecated, use mp_rand_source instead.
 /* A last resort to provide random data on systems without any of the other
  * implemented ways to gather entropy.
  * It is compatible with `rng_get_bytes()` from libtomcrypt so you could
@@ -262,232 +433,230 @@ extern void (*ltm_rng_callback)(void);
 #endif
 
 /* ---> binary operations <--- */
-/* c = a XOR b  */
-int mp_xor(const mp_int *a, const mp_int *b, mp_int *c);
-
-/* c = a OR b */
-int mp_or(const mp_int *a, const mp_int *b, mp_int *c);
-
-/* c = a AND b */
-int mp_and(const mp_int *a, const mp_int *b, mp_int *c);
 
 /* Checks the bit at position b and returns MP_YES
-   if the bit is 1, MP_NO if it is 0 and MP_VAL
-   in case of error */
-int mp_get_bit(const mp_int *a, int b);
+ * if the bit is 1, MP_NO if it is 0 and MP_VAL
+ * in case of error
+ */
+MP_DEPRECATED(s_mp_get_bit) int mp_get_bit(const mp_int *a, int b) MP_WUR;
 
 /* c = a XOR b (two complement) */
-int mp_tc_xor(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(mp_xor) mp_err mp_tc_xor(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+mp_err mp_xor(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* c = a OR b (two complement) */
-int mp_tc_or(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(mp_or) mp_err mp_tc_or(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+mp_err mp_or(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* c = a AND b (two complement) */
-int mp_tc_and(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(mp_and) mp_err mp_tc_and(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+mp_err mp_and(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
-/* right shift (two complement) */
-int mp_signed_rsh(const mp_int *a, int b, mp_int *c);
+/* b = ~a (bitwise not, two complement) */
+mp_err mp_complement(const mp_int *a, mp_int *b) MP_WUR;
 
-/* ---> Basic arithmetic <--- */
+/* right shift with sign extension */
+MP_DEPRECATED(mp_signed_rsh) mp_err mp_tc_div_2d(const mp_int *a, int b, mp_int *c) MP_WUR;
+mp_err mp_signed_rsh(const mp_int *a, int b, mp_int *c) MP_WUR;
 
-/* b = ~a */
-int mp_complement(const mp_int *a, mp_int *b);
+/* ---> Basic arithmetic <--- */
 
 /* b = -a */
-int mp_neg(const mp_int *a, mp_int *b);
+mp_err mp_neg(const mp_int *a, mp_int *b) MP_WUR;
 
 /* b = |a| */
-int mp_abs(const mp_int *a, mp_int *b);
+mp_err mp_abs(const mp_int *a, mp_int *b) MP_WUR;
 
 /* compare a to b */
-int mp_cmp(const mp_int *a, const mp_int *b);
+mp_ord mp_cmp(const mp_int *a, const mp_int *b) MP_WUR;
 
 /* compare |a| to |b| */
-int mp_cmp_mag(const mp_int *a, const mp_int *b);
+mp_ord mp_cmp_mag(const mp_int *a, const mp_int *b) MP_WUR;
 
 /* c = a + b */
-int mp_add(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_add(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* c = a - b */
-int mp_sub(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_sub(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* c = a * b */
-int mp_mul(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_mul(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* b = a*a  */
-int mp_sqr(const mp_int *a, mp_int *b);
+mp_err mp_sqr(const mp_int *a, mp_int *b) MP_WUR;
 
 /* a/b => cb + d == a */
-int mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d);
+mp_err mp_div(const mp_int *a, const mp_int *b, mp_int *c, mp_int *d) MP_WUR;
 
 /* c = a mod b, 0 <= c < b  */
-int mp_mod(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_mod(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+
+/* Increment "a" by one like "a++". Changes input! */
+mp_err mp_incr(mp_int *a) MP_WUR;
+
+/* Decrement "a" by one like "a--". Changes input! */
+mp_err mp_decr(mp_int *a) MP_WUR;
 
 /* ---> single digit functions <--- */
 
 /* compare against a single digit */
-int mp_cmp_d(const mp_int *a, mp_digit b);
+mp_ord mp_cmp_d(const mp_int *a, mp_digit b) MP_WUR;
 
 /* c = a + b */
-int mp_add_d(const mp_int *a, mp_digit b, mp_int *c);
+mp_err mp_add_d(const mp_int *a, mp_digit b, mp_int *c) MP_WUR;
 
 /* c = a - b */
-int mp_sub_d(const mp_int *a, mp_digit b, mp_int *c);
+mp_err mp_sub_d(const mp_int *a, mp_digit b, mp_int *c) MP_WUR;
 
 /* c = a * b */
-int mp_mul_d(const mp_int *a, mp_digit b, mp_int *c);
+mp_err mp_mul_d(const mp_int *a, mp_digit b, mp_int *c) MP_WUR;
 
 /* a/b => cb + d == a */
-int mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d);
-
-/* a/3 => 3c + d == a */
-int mp_div_3(const mp_int *a, mp_int *c, mp_digit *d);
-
-/* c = a**b */
-int mp_expt_d(const mp_int *a, mp_digit b, mp_int *c);
-int mp_expt_d_ex(const mp_int *a, mp_digit b, mp_int *c, int fast);
+mp_err mp_div_d(const mp_int *a, mp_digit b, mp_int *c, mp_digit *d) MP_WUR;
 
 /* c = a mod b, 0 <= c < b  */
-int mp_mod_d(const mp_int *a, mp_digit b, mp_digit *c);
+mp_err mp_mod_d(const mp_int *a, mp_digit b, mp_digit *c) MP_WUR;
 
 /* ---> number theory <--- */
 
 /* d = a + b (mod c) */
-int mp_addmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d);
+mp_err mp_addmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d) MP_WUR;
 
 /* d = a - b (mod c) */
-int mp_submod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d);
+mp_err mp_submod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d) MP_WUR;
 
 /* d = a * b (mod c) */
-int mp_mulmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d);
+mp_err mp_mulmod(const mp_int *a, const mp_int *b, const mp_int *c, mp_int *d) MP_WUR;
 
 /* c = a * a (mod b) */
-int mp_sqrmod(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_sqrmod(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* c = 1/a (mod b) */
-int mp_invmod(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_invmod(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* c = (a, b) */
-int mp_gcd(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_gcd(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* produces value such that U1*a + U2*b = U3 */
-int mp_exteuclid(const mp_int *a, const mp_int *b, mp_int *U1, mp_int *U2, mp_int *U3);
+mp_err mp_exteuclid(const mp_int *a, const mp_int *b, mp_int *U1, mp_int *U2, mp_int *U3) MP_WUR;
 
 /* c = [a, b] or (a*b)/(a, b) */
-int mp_lcm(const mp_int *a, const mp_int *b, mp_int *c);
+mp_err mp_lcm(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
 
 /* finds one of the b'th root of a, such that |c|**b <= |a|
  *
  * returns error if a < 0 and b is even
  */
-int mp_n_root(const mp_int *a, mp_digit b, mp_int *c);
-int mp_n_root_ex(const mp_int *a, mp_digit b, mp_int *c, int fast);
+mp_err mp_root_u32(const mp_int *a, unsigned int b, mp_int *c) MP_WUR;
+MP_DEPRECATED(mp_root_u32) mp_err mp_n_root(const mp_int *a, mp_digit b, mp_int *c) MP_WUR;
+MP_DEPRECATED(mp_root_u32) mp_err mp_n_root_ex(const mp_int *a, mp_digit b, mp_int *c, int fast) MP_WUR;
 
 /* special sqrt algo */
-int mp_sqrt(const mp_int *arg, mp_int *ret);
+mp_err mp_sqrt(const mp_int *arg, mp_int *ret) MP_WUR;
 
 /* special sqrt (mod prime) */
-int mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret);
+mp_err mp_sqrtmod_prime(const mp_int *n, const mp_int *prime, mp_int *ret) MP_WUR;
 
 /* is number a square? */
-int mp_is_square(const mp_int *arg, int *ret);
+mp_err mp_is_square(const mp_int *arg, mp_bool *ret) MP_WUR;
 
 /* computes the jacobi c = (a | n) (or Legendre if b is prime)  */
-int mp_jacobi(const mp_int *a, const mp_int *n, int *c);
+MP_DEPRECATED(mp_kronecker) mp_err mp_jacobi(const mp_int *a, const mp_int *n, int *c) MP_WUR;
 
 /* computes the Kronecker symbol c = (a | p) (like jacobi() but with {a,p} in Z */
-int mp_kronecker(const mp_int *a, const mp_int *p, int *c);
+mp_err mp_kronecker(const mp_int *a, const mp_int *p, int *c) MP_WUR;
 
 /* used to setup the Barrett reduction for a given modulus b */
-int mp_reduce_setup(mp_int *a, const mp_int *b);
+mp_err mp_reduce_setup(mp_int *a, const mp_int *b) MP_WUR;
 
 /* Barrett Reduction, computes a (mod b) with a precomputed value c
  *
  * Assumes that 0 < x <= m*m, note if 0 > x > -(m*m) then you can merely
  * compute the reduction as -1 * mp_reduce(mp_abs(x)) [pseudo code].
  */
-int mp_reduce(mp_int *x, const mp_int *m, const mp_int *mu);
+mp_err mp_reduce(mp_int *x, const mp_int *m, const mp_int *mu) MP_WUR;
 
 /* setups the montgomery reduction */
-int mp_montgomery_setup(const mp_int *n, mp_digit *rho);
+mp_err mp_montgomery_setup(const mp_int *n, mp_digit *rho) MP_WUR;
 
 /* computes a = B**n mod b without division or multiplication useful for
  * normalizing numbers in a Montgomery system.
  */
-int mp_montgomery_calc_normalization(mp_int *a, const mp_int *b);
+mp_err mp_montgomery_calc_normalization(mp_int *a, const mp_int *b) MP_WUR;
 
 /* computes x/R == x (mod N) via Montgomery Reduction */
-int mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho);
+mp_err mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho) MP_WUR;
 
 /* returns 1 if a is a valid DR modulus */
-int mp_dr_is_modulus(const mp_int *a);
+mp_bool mp_dr_is_modulus(const mp_int *a) MP_WUR;
 
 /* sets the value of "d" required for mp_dr_reduce */
 void mp_dr_setup(const mp_int *a, mp_digit *d);
 
 /* reduces a modulo n using the Diminished Radix method */
-int mp_dr_reduce(mp_int *x, const mp_int *n, mp_digit k);
+mp_err mp_dr_reduce(mp_int *x, const mp_int *n, mp_digit k) MP_WUR;
 
 /* returns true if a can be reduced with mp_reduce_2k */
-int mp_reduce_is_2k(const mp_int *a);
+mp_bool mp_reduce_is_2k(const mp_int *a) MP_WUR;
 
 /* determines k value for 2k reduction */
-int mp_reduce_2k_setup(const mp_int *a, mp_digit *d);
+mp_err mp_reduce_2k_setup(const mp_int *a, mp_digit *d) MP_WUR;
 
 /* reduces a modulo b where b is of the form 2**p - k [0 <= a] */
-int mp_reduce_2k(mp_int *a, const mp_int *n, mp_digit d);
+mp_err mp_reduce_2k(mp_int *a, const mp_int *n, mp_digit d) MP_WUR;
 
 /* returns true if a can be reduced with mp_reduce_2k_l */
-int mp_reduce_is_2k_l(const mp_int *a);
+mp_bool mp_reduce_is_2k_l(const mp_int *a) MP_WUR;
 
 /* determines k value for 2k reduction */
-int mp_reduce_2k_setup_l(const mp_int *a, mp_int *d);
+mp_err mp_reduce_2k_setup_l(const mp_int *a, mp_int *d) MP_WUR;
 
 /* reduces a modulo b where b is of the form 2**p - k [0 <= a] */
-int mp_reduce_2k_l(mp_int *a, const mp_int *n, const mp_int *d);
+mp_err mp_reduce_2k_l(mp_int *a, const mp_int *n, const mp_int *d) MP_WUR;
 
 /* Y = G**X (mod P) */
-int mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y);
+mp_err mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y) MP_WUR;
 
 /* ---> Primes <--- */
 
 /* number of primes */
 #ifdef MP_8BIT
-#  define PRIME_SIZE 31
+#  define PRIVATE_MP_PRIME_TAB_SIZE 31
 #else
-#  define PRIME_SIZE 256
+#  define PRIVATE_MP_PRIME_TAB_SIZE 256
 #endif
+#define PRIME_SIZE (MP_DEPRECATED_PRAGMA("PRIME_SIZE has been made internal") PRIVATE_MP_PRIME_TAB_SIZE)
 
 /* table of first PRIME_SIZE primes */
-extern const mp_digit ltm_prime_tab[PRIME_SIZE];
+MP_DEPRECATED(internal) extern const mp_digit ltm_prime_tab[PRIVATE_MP_PRIME_TAB_SIZE];
 
 /* result=1 if a is divisible by one of the first PRIME_SIZE primes */
-int mp_prime_is_divisible(const mp_int *a, int *result);
+MP_DEPRECATED(mp_prime_is_prime) mp_err mp_prime_is_divisible(const mp_int *a, mp_bool *result) MP_WUR;
 
 /* performs one Fermat test of "a" using base "b".
  * Sets result to 0 if composite or 1 if probable prime
  */
-int mp_prime_fermat(const mp_int *a, const mp_int *b, int *result);
+mp_err mp_prime_fermat(const mp_int *a, const mp_int *b, mp_bool *result) MP_WUR;
 
 /* performs one Miller-Rabin test of "a" using base "b".
  * Sets result to 0 if composite or 1 if probable prime
  */
-int mp_prime_miller_rabin(const mp_int *a, const mp_int *b, int *result);
+mp_err mp_prime_miller_rabin(const mp_int *a, const mp_int *b, mp_bool *result) MP_WUR;
 
 /* This gives [for a given bit size] the number of trials required
  * such that Miller-Rabin gives a prob of failure lower than 2^-96
  */
-int mp_prime_rabin_miller_trials(int size);
+int mp_prime_rabin_miller_trials(int size) MP_WUR;
 
 /* performs one strong Lucas-Selfridge test of "a".
  * Sets result to 0 if composite or 1 if probable prime
  */
-int mp_prime_strong_lucas_selfridge(const mp_int *a, int *result);
+mp_err mp_prime_strong_lucas_selfridge(const mp_int *a, mp_bool *result) MP_WUR;
 
 /* performs one Frobenius test of "a" as described by Paul Underwood.
  * Sets result to 0 if composite or 1 if probable prime
  */
-int mp_prime_frobenius_underwood(const mp_int *N, int *result);
+mp_err mp_prime_frobenius_underwood(const mp_int *N, mp_bool *result) MP_WUR;
 
 /* performs t random rounds of Miller-Rabin on "a" additional to
  * bases 2 and 3.  Also performs an initial sieve of trial
@@ -503,14 +672,14 @@ int mp_prime_frobenius_underwood(const mp_int *N, int *result);
  *
  * Sets result to 1 if probably prime, 0 otherwise
  */
-int mp_prime_is_prime(const mp_int *a, int t, int *result);
+mp_err mp_prime_is_prime(const mp_int *a, int t, mp_bool *result) MP_WUR;
 
 /* finds the next prime after the number "a" using "t" trials
  * of Miller-Rabin.
  *
  * bbs_style = 1 means the prime must be congruent to 3 mod 4
  */
-int mp_prime_next_prime(mp_int *a, int t, int bbs_style);
+mp_err mp_prime_next_prime(mp_int *a, int t, int bbs_style) MP_WUR;
 
 /* makes a truly random prime of a given size (bytes),
  * call with bbs = 1 if you want it to be congruent to 3 mod 4
@@ -521,65 +690,85 @@ int mp_prime_next_prime(mp_int *a, int t, int bbs_style);
  *
  * The prime generated will be larger than 2^(8*size).
  */
-#define mp_prime_random(a, t, size, bbs, cb, dat) mp_prime_random_ex(a, t, ((size) * 8) + 1, (bbs==1)?LTM_PRIME_BBS:0, cb, dat)
+#define mp_prime_random(a, t, size, bbs, cb, dat) (MP_DEPRECATED_PRAGMA("mp_prime_random has been deprecated, use mp_prime_rand instead") mp_prime_random_ex(a, t, ((size) * 8) + 1, (bbs==1)?MP_PRIME_BBS:0, cb, dat))
 
 /* makes a truly random prime of a given size (bits),
  *
  * Flags are as follows:
  *
- *   LTM_PRIME_BBS      - make prime congruent to 3 mod 4
- *   LTM_PRIME_SAFE     - make sure (p-1)/2 is prime as well (implies LTM_PRIME_BBS)
- *   LTM_PRIME_2MSB_ON  - make the 2nd highest bit one
+ *   MP_PRIME_BBS      - make prime congruent to 3 mod 4
+ *   MP_PRIME_SAFE     - make sure (p-1)/2 is prime as well (implies MP_PRIME_BBS)
+ *   MP_PRIME_2MSB_ON  - make the 2nd highest bit one
  *
  * You have to supply a callback which fills in a buffer with random bytes.  "dat" is a parameter you can
  * have passed to the callback (e.g. a state or something).  This function doesn't use "dat" itself
  * so it can be NULL
  *
  */
-int mp_prime_random_ex(mp_int *a, int t, int size, int flags, ltm_prime_callback cb, void *dat);
+MP_DEPRECATED(mp_prime_rand) mp_err mp_prime_random_ex(mp_int *a, int t, int size, int flags,
+      private_mp_prime_callback cb, void *dat) MP_WUR;
+mp_err mp_prime_rand(mp_int *a, int t, int size, int flags) MP_WUR;
+
+/* Integer logarithm to integer base */
+mp_err mp_log_u32(const mp_int *a, unsigned int base, unsigned int *c) MP_WUR;
+
+/* c = a**b */
+mp_err mp_expt_u32(const mp_int *a, unsigned int b, mp_int *c) MP_WUR;
+MP_DEPRECATED(mp_expt_u32) mp_err mp_expt_d(const mp_int *a, mp_digit b, mp_int *c) MP_WUR;
+MP_DEPRECATED(mp_expt_u32) mp_err mp_expt_d_ex(const mp_int *a, mp_digit b, mp_int *c, int fast) MP_WUR;
 
 /* ---> radix conversion <--- */
-int mp_count_bits(const mp_int *a);
-
-int mp_unsigned_bin_size(const mp_int *a);
-int mp_read_unsigned_bin(mp_int *a, const unsigned char *b, int c);
-int mp_to_unsigned_bin(const mp_int *a, unsigned char *b);
-int mp_to_unsigned_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen);
-
-int mp_signed_bin_size(const mp_int *a);
-int mp_read_signed_bin(mp_int *a, const unsigned char *b, int c);
-int mp_to_signed_bin(const mp_int *a,  unsigned char *b);
-int mp_to_signed_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen);
-
-int mp_read_radix(mp_int *a, const char *str, int radix);
-int mp_toradix(const mp_int *a, char *str, int radix);
-int mp_toradix_n(const mp_int *a, char *str, int radix, int maxlen);
-int mp_radix_size(const mp_int *a, int radix, int *size);
-
-#ifndef LTM_NO_FILE
-int mp_fread(mp_int *a, int radix, FILE *stream);
-int mp_fwrite(const mp_int *a, int radix, FILE *stream);
+int mp_count_bits(const mp_int *a) MP_WUR;
+
+
+MP_DEPRECATED(mp_ubin_size) int mp_unsigned_bin_size(const mp_int *a) MP_WUR;
+MP_DEPRECATED(mp_from_ubin) mp_err mp_read_unsigned_bin(mp_int *a, const unsigned char *b, int c) MP_WUR;
+MP_DEPRECATED(mp_to_ubin) mp_err mp_to_unsigned_bin(const mp_int *a, unsigned char *b) MP_WUR;
+MP_DEPRECATED(mp_to_ubin) mp_err mp_to_unsigned_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen) MP_WUR;
+
+MP_DEPRECATED(mp_sbin_size) int mp_signed_bin_size(const mp_int *a) MP_WUR;
+MP_DEPRECATED(mp_from_sbin) mp_err mp_read_signed_bin(mp_int *a, const unsigned char *b, int c) MP_WUR;
+MP_DEPRECATED(mp_to_sbin) mp_err mp_to_signed_bin(const mp_int *a,  unsigned char *b) MP_WUR;
+MP_DEPRECATED(mp_to_sbin) mp_err mp_to_signed_bin_n(const mp_int *a, unsigned char *b, unsigned long *outlen) MP_WUR;
+
+size_t mp_ubin_size(const mp_int *a) MP_WUR;
+mp_err mp_from_ubin(mp_int *a, const unsigned char *buf, size_t size) MP_WUR;
+mp_err mp_to_ubin(const mp_int *a, unsigned char *buf, size_t maxlen, size_t *written) MP_WUR;
+
+size_t mp_sbin_size(const mp_int *a) MP_WUR;
+mp_err mp_from_sbin(mp_int *a, const unsigned char *buf, size_t size) MP_WUR;
+mp_err mp_to_sbin(const mp_int *a, unsigned char *buf, size_t maxlen, size_t *written) MP_WUR;
+
+mp_err mp_read_radix(mp_int *a, const char *str, int radix) MP_WUR;
+MP_DEPRECATED(mp_to_radix) mp_err mp_toradix(const mp_int *a, char *str, int radix) MP_WUR;
+MP_DEPRECATED(mp_to_radix) mp_err mp_toradix_n(const mp_int *a, char *str, int radix, int maxlen) MP_WUR;
+mp_err mp_to_radix(const mp_int *a, char *str, size_t maxlen, size_t *written, int radix) MP_WUR;
+mp_err mp_radix_size(const mp_int *a, int radix, int *size) MP_WUR;
+
+#ifndef MP_NO_FILE
+mp_err mp_fread(mp_int *a, int radix, FILE *stream) MP_WUR;
+mp_err mp_fwrite(const mp_int *a, int radix, FILE *stream) MP_WUR;
 #endif
 
-#define mp_read_raw(mp, str, len) mp_read_signed_bin((mp), (str), (len))
-#define mp_raw_size(mp)           mp_signed_bin_size(mp)
-#define mp_toraw(mp, str)         mp_to_signed_bin((mp), (str))
-#define mp_read_mag(mp, str, len) mp_read_unsigned_bin((mp), (str), (len))
-#define mp_mag_size(mp)           mp_unsigned_bin_size(mp)
-#define mp_tomag(mp, str)         mp_to_unsigned_bin((mp), (str))
+#define mp_read_raw(mp, str, len) (MP_DEPRECATED_PRAGMA("replaced by mp_read_signed_bin") mp_read_signed_bin((mp), (str), (len)))
+#define mp_raw_size(mp)           (MP_DEPRECATED_PRAGMA("replaced by mp_signed_bin_size") mp_signed_bin_size(mp))
+#define mp_toraw(mp, str)         (MP_DEPRECATED_PRAGMA("replaced by mp_to_signed_bin") mp_to_signed_bin((mp), (str)))
+#define mp_read_mag(mp, str, len) (MP_DEPRECATED_PRAGMA("replaced by mp_read_unsigned_bin") mp_read_unsigned_bin((mp), (str), (len))
+#define mp_mag_size(mp)           (MP_DEPRECATED_PRAGMA("replaced by mp_unsigned_bin_size") mp_unsigned_bin_size(mp))
+#define mp_tomag(mp, str)         (MP_DEPRECATED_PRAGMA("replaced by mp_to_unsigned_bin") mp_to_unsigned_bin((mp), (str)))
+
+#define mp_tobinary(M, S)  (MP_DEPRECATED_PRAGMA("replaced by mp_to_binary")  mp_toradix((M), (S), 2))
+#define mp_tooctal(M, S)   (MP_DEPRECATED_PRAGMA("replaced by mp_to_octal")   mp_toradix((M), (S), 8))
+#define mp_todecimal(M, S) (MP_DEPRECATED_PRAGMA("replaced by mp_to_decimal") mp_toradix((M), (S), 10))
+#define mp_tohex(M, S)     (MP_DEPRECATED_PRAGMA("replaced by mp_to_hex")     mp_toradix((M), (S), 16))
 
-#define mp_tobinary(M, S)  mp_toradix((M), (S), 2)
-#define mp_tooctal(M, S)   mp_toradix((M), (S), 8)
-#define mp_todecimal(M, S) mp_toradix((M), (S), 10)
-#define mp_tohex(M, S)     mp_toradix((M), (S), 16)
+#define mp_to_binary(M, S, N)  mp_to_radix((M), (S), (N), NULL, 2)
+#define mp_to_octal(M, S, N)   mp_to_radix((M), (S), (N), NULL, 8)
+#define mp_to_decimal(M, S, N) mp_to_radix((M), (S), (N), NULL, 10)
+#define mp_to_hex(M, S, N)     mp_to_radix((M), (S), (N), NULL, 16)
 
 #ifdef __cplusplus
 }
 #endif
 
 #endif
-
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/tommath_class.h b/libtommath/tommath_class.h
index b3c09fe..52ba585 100644
--- a/libtommath/tommath_class.h
+++ b/libtommath/tommath_class.h
@@ -1,16 +1,8 @@
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 #if !(defined(LTM1) && defined(LTM2) && defined(LTM3))
+#define LTM_INSIDE
 #if defined(LTM2)
 #   define LTM3
 #endif
@@ -19,12 +11,8 @@
 #endif
 #define LTM1
 #if defined(LTM_ALL)
-#   define BN_ERROR_C
-#   define BN_FAST_MP_INVMOD_C
-#   define BN_FAST_MP_MONTGOMERY_REDUCE_C
-#   define BN_FAST_S_MP_MUL_DIGS_C
-#   define BN_FAST_S_MP_MUL_HIGH_DIGS_C
-#   define BN_FAST_S_MP_SQR_C
+#   define BN_CUTOFFS_C
+#   define BN_DEPRECATED_C
 #   define BN_MP_2EXPT_C
 #   define BN_MP_ABS_C
 #   define BN_MP_ADD_C
@@ -41,6 +29,7 @@
 #   define BN_MP_COMPLEMENT_C
 #   define BN_MP_COPY_C
 #   define BN_MP_COUNT_BITS_C
+#   define BN_MP_DECR_C
 #   define BN_MP_DIV_C
 #   define BN_MP_DIV_2_C
 #   define BN_MP_DIV_2D_C
@@ -49,37 +38,47 @@
 #   define BN_MP_DR_IS_MODULUS_C
 #   define BN_MP_DR_REDUCE_C
 #   define BN_MP_DR_SETUP_C
+#   define BN_MP_ERROR_TO_STRING_C
 #   define BN_MP_EXCH_C
-#   define BN_MP_EXPORT_C
-#   define BN_MP_EXPT_D_C
-#   define BN_MP_EXPT_D_EX_C
+#   define BN_MP_EXPT_U32_C
 #   define BN_MP_EXPTMOD_C
-#   define BN_MP_EXPTMOD_FAST_C
 #   define BN_MP_EXTEUCLID_C
 #   define BN_MP_FREAD_C
+#   define BN_MP_FROM_SBIN_C
+#   define BN_MP_FROM_UBIN_C
 #   define BN_MP_FWRITE_C
 #   define BN_MP_GCD_C
-#   define BN_S_MP_GET_BIT_C
 #   define BN_MP_GET_DOUBLE_C
-#   define BN_MP_GET_INT_C
-#   define BN_MP_GET_LONG_C
-#   define BN_MP_GET_LONG_LONG_C
+#   define BN_MP_GET_I32_C
+#   define BN_MP_GET_I64_C
+#   define BN_MP_GET_L_C
+#   define BN_MP_GET_LL_C
+#   define BN_MP_GET_MAG_U32_C
+#   define BN_MP_GET_MAG_U64_C
+#   define BN_MP_GET_MAG_UL_C
+#   define BN_MP_GET_MAG_ULL_C
 #   define BN_MP_GROW_C
-#   define BN_MP_IMPORT_C
+#   define BN_MP_INCR_C
 #   define BN_MP_INIT_C
 #   define BN_MP_INIT_COPY_C
+#   define BN_MP_INIT_I32_C
+#   define BN_MP_INIT_I64_C
+#   define BN_MP_INIT_L_C
+#   define BN_MP_INIT_LL_C
 #   define BN_MP_INIT_MULTI_C
 #   define BN_MP_INIT_SET_C
-#   define BN_MP_INIT_SET_INT_C
 #   define BN_MP_INIT_SIZE_C
+#   define BN_MP_INIT_U32_C
+#   define BN_MP_INIT_U64_C
+#   define BN_MP_INIT_UL_C
+#   define BN_MP_INIT_ULL_C
 #   define BN_MP_INVMOD_C
-#   define BN_MP_INVMOD_SLOW_C
 #   define BN_MP_IS_SQUARE_C
-#   define BN_MP_JACOBI_C
-#   define BN_MP_KARATSUBA_MUL_C
-#   define BN_MP_KARATSUBA_SQR_C
+#   define BN_MP_ISEVEN_C
+#   define BN_MP_ISODD_C
 #   define BN_MP_KRONECKER_C
 #   define BN_MP_LCM_C
+#   define BN_MP_LOG_U32_C
 #   define BN_MP_LSHD_C
 #   define BN_MP_MOD_C
 #   define BN_MP_MOD_2D_C
@@ -92,25 +91,22 @@
 #   define BN_MP_MUL_2D_C
 #   define BN_MP_MUL_D_C
 #   define BN_MP_MULMOD_C
-#   define BN_MP_N_ROOT_C
-#   define BN_MP_N_ROOT_EX_C
 #   define BN_MP_NEG_C
 #   define BN_MP_OR_C
+#   define BN_MP_PACK_C
+#   define BN_MP_PACK_COUNT_C
 #   define BN_MP_PRIME_FERMAT_C
 #   define BN_MP_PRIME_FROBENIUS_UNDERWOOD_C
-#   define BN_MP_PRIME_IS_DIVISIBLE_C
 #   define BN_MP_PRIME_IS_PRIME_C
 #   define BN_MP_PRIME_MILLER_RABIN_C
 #   define BN_MP_PRIME_NEXT_PRIME_C
 #   define BN_MP_PRIME_RABIN_MILLER_TRIALS_C
-#   define BN_MP_PRIME_RANDOM_EX_C
+#   define BN_MP_PRIME_RAND_C
 #   define BN_MP_PRIME_STRONG_LUCAS_SELFRIDGE_C
 #   define BN_MP_RADIX_SIZE_C
 #   define BN_MP_RADIX_SMAP_C
 #   define BN_MP_RAND_C
 #   define BN_MP_READ_RADIX_C
-#   define BN_MP_READ_SIGNED_BIN_C
-#   define BN_MP_READ_UNSIGNED_BIN_C
 #   define BN_MP_REDUCE_C
 #   define BN_MP_REDUCE_2K_C
 #   define BN_MP_REDUCE_2K_L_C
@@ -119,14 +115,21 @@
 #   define BN_MP_REDUCE_IS_2K_C
 #   define BN_MP_REDUCE_IS_2K_L_C
 #   define BN_MP_REDUCE_SETUP_C
+#   define BN_MP_ROOT_U32_C
 #   define BN_MP_RSHD_C
+#   define BN_MP_SBIN_SIZE_C
 #   define BN_MP_SET_C
 #   define BN_MP_SET_DOUBLE_C
-#   define BN_MP_SET_INT_C
-#   define BN_MP_SET_LONG_C
-#   define BN_MP_SET_LONG_LONG_C
+#   define BN_MP_SET_I32_C
+#   define BN_MP_SET_I64_C
+#   define BN_MP_SET_L_C
+#   define BN_MP_SET_LL_C
+#   define BN_MP_SET_U32_C
+#   define BN_MP_SET_U64_C
+#   define BN_MP_SET_UL_C
+#   define BN_MP_SET_ULL_C
 #   define BN_MP_SHRINK_C
-#   define BN_MP_SIGNED_BIN_SIZE_C
+#   define BN_MP_SIGNED_RSH_C
 #   define BN_MP_SQR_C
 #   define BN_MP_SQRMOD_C
 #   define BN_MP_SQRT_C
@@ -134,79 +137,132 @@
 #   define BN_MP_SUB_C
 #   define BN_MP_SUB_D_C
 #   define BN_MP_SUBMOD_C
-#   define BN_MP_TC_AND_C
-#   define BN_MP_SIGNED_RSH_C
-#   define BN_MP_TC_OR_C
-#   define BN_MP_TC_XOR_C
-#   define BN_MP_TO_SIGNED_BIN_C
-#   define BN_MP_TO_SIGNED_BIN_N_C
-#   define BN_MP_TO_UNSIGNED_BIN_C
-#   define BN_MP_TO_UNSIGNED_BIN_N_C
-#   define BN_MP_TOOM_MUL_C
-#   define BN_MP_TOOM_SQR_C
-#   define BN_MP_TORADIX_C
-#   define BN_MP_TORADIX_N_C
-#   define BN_MP_UNSIGNED_BIN_SIZE_C
+#   define BN_MP_TO_RADIX_C
+#   define BN_MP_TO_SBIN_C
+#   define BN_MP_TO_UBIN_C
+#   define BN_MP_UBIN_SIZE_C
+#   define BN_MP_UNPACK_C
 #   define BN_MP_XOR_C
 #   define BN_MP_ZERO_C
 #   define BN_PRIME_TAB_C
-#   define BN_REVERSE_C
 #   define BN_S_MP_ADD_C
+#   define BN_S_MP_BALANCE_MUL_C
 #   define BN_S_MP_EXPTMOD_C
+#   define BN_S_MP_EXPTMOD_FAST_C
+#   define BN_S_MP_GET_BIT_C
+#   define BN_S_MP_INVMOD_FAST_C
+#   define BN_S_MP_INVMOD_SLOW_C
+#   define BN_S_MP_KARATSUBA_MUL_C
+#   define BN_S_MP_KARATSUBA_SQR_C
+#   define BN_S_MP_MONTGOMERY_REDUCE_FAST_C
 #   define BN_S_MP_MUL_DIGS_C
+#   define BN_S_MP_MUL_DIGS_FAST_C
 #   define BN_S_MP_MUL_HIGH_DIGS_C
+#   define BN_S_MP_MUL_HIGH_DIGS_FAST_C
+#   define BN_S_MP_PRIME_IS_DIVISIBLE_C
+#   define BN_S_MP_RAND_JENKINS_C
+#   define BN_S_MP_RAND_PLATFORM_C
+#   define BN_S_MP_REVERSE_C
 #   define BN_S_MP_SQR_C
+#   define BN_S_MP_SQR_FAST_C
 #   define BN_S_MP_SUB_C
-#   define BNCORE_C
+#   define BN_S_MP_TOOM_MUL_C
+#   define BN_S_MP_TOOM_SQR_C
 #endif
-#if defined(BN_ERROR_C)
-#   define BN_MP_ERROR_TO_STRING_C
 #endif
-
-#if defined(BN_FAST_MP_INVMOD_C)
-#   define BN_MP_ISEVEN_C
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_COPY_C
-#   define BN_MP_MOD_C
-#   define BN_MP_ISZERO_C
-#   define BN_MP_SET_C
-#   define BN_MP_DIV_2_C
-#   define BN_MP_ISODD_C
-#   define BN_MP_SUB_C
-#   define BN_MP_CMP_C
-#   define BN_MP_CMP_D_C
-#   define BN_MP_ADD_C
-#   define BN_MP_CMP_MAG_C
-#   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_MULTI_C
-#endif
-
-#if defined(BN_FAST_MP_MONTGOMERY_REDUCE_C)
-#   define BN_MP_GROW_C
-#   define BN_MP_RSHD_C
-#   define BN_MP_CLAMP_C
-#   define BN_MP_CMP_MAG_C
-#   define BN_S_MP_SUB_C
-#endif
-
-#if defined(BN_FAST_S_MP_MUL_DIGS_C)
-#   define BN_MP_GROW_C
-#   define BN_MP_CLAMP_C
+#if defined(BN_CUTOFFS_C)
 #endif
 
-#if defined(BN_FAST_S_MP_MUL_HIGH_DIGS_C)
-#   define BN_MP_GROW_C
-#   define BN_MP_CLAMP_C
-#endif
-
-#if defined(BN_FAST_S_MP_SQR_C)
-#   define BN_MP_GROW_C
-#   define BN_MP_CLAMP_C
+#if defined(BN_DEPRECATED_C)
+#   define BN_FAST_MP_INVMOD_C
+#   define BN_FAST_MP_MONTGOMERY_REDUCE_C
+#   define BN_FAST_S_MP_MUL_DIGS_C
+#   define BN_FAST_S_MP_MUL_HIGH_DIGS_C
+#   define BN_FAST_S_MP_SQR_C
+#   define BN_MP_AND_C
+#   define BN_MP_BALANCE_MUL_C
+#   define BN_MP_CMP_D_C
+#   define BN_MP_EXPORT_C
+#   define BN_MP_EXPTMOD_FAST_C
+#   define BN_MP_EXPT_D_C
+#   define BN_MP_EXPT_D_EX_C
+#   define BN_MP_EXPT_U32_C
+#   define BN_MP_FROM_SBIN_C
+#   define BN_MP_FROM_UBIN_C
+#   define BN_MP_GET_BIT_C
+#   define BN_MP_GET_INT_C
+#   define BN_MP_GET_LONG_C
+#   define BN_MP_GET_LONG_LONG_C
+#   define BN_MP_GET_MAG_U32_C
+#   define BN_MP_GET_MAG_ULL_C
+#   define BN_MP_GET_MAG_UL_C
+#   define BN_MP_IMPORT_C
+#   define BN_MP_INIT_SET_INT_C
+#   define BN_MP_INIT_U32_C
+#   define BN_MP_INVMOD_SLOW_C
+#   define BN_MP_JACOBI_C
+#   define BN_MP_KARATSUBA_MUL_C
+#   define BN_MP_KARATSUBA_SQR_C
+#   define BN_MP_KRONECKER_C
+#   define BN_MP_N_ROOT_C
+#   define BN_MP_N_ROOT_EX_C
+#   define BN_MP_OR_C
+#   define BN_MP_PACK_C
+#   define BN_MP_PRIME_IS_DIVISIBLE_C
+#   define BN_MP_PRIME_RANDOM_EX_C
+#   define BN_MP_RAND_DIGIT_C
+#   define BN_MP_READ_SIGNED_BIN_C
+#   define BN_MP_READ_UNSIGNED_BIN_C
+#   define BN_MP_ROOT_U32_C
+#   define BN_MP_SBIN_SIZE_C
+#   define BN_MP_SET_INT_C
+#   define BN_MP_SET_LONG_C
+#   define BN_MP_SET_LONG_LONG_C
+#   define BN_MP_SET_U32_C
+#   define BN_MP_SET_U64_C
+#   define BN_MP_SIGNED_BIN_SIZE_C
+#   define BN_MP_SIGNED_RSH_C
+#   define BN_MP_TC_AND_C
+#   define BN_MP_TC_DIV_2D_C
+#   define BN_MP_TC_OR_C
+#   define BN_MP_TC_XOR_C
+#   define BN_MP_TOOM_MUL_C
+#   define BN_MP_TOOM_SQR_C
+#   define BN_MP_TORADIX_C
+#   define BN_MP_TORADIX_N_C
+#   define BN_MP_TO_RADIX_C
+#   define BN_MP_TO_SBIN_C
+#   define BN_MP_TO_SIGNED_BIN_C
+#   define BN_MP_TO_SIGNED_BIN_N_C
+#   define BN_MP_TO_UBIN_C
+#   define BN_MP_TO_UNSIGNED_BIN_C
+#   define BN_MP_TO_UNSIGNED_BIN_N_C
+#   define BN_MP_UBIN_SIZE_C
+#   define BN_MP_UNPACK_C
+#   define BN_MP_UNSIGNED_BIN_SIZE_C
+#   define BN_MP_XOR_C
+#   define BN_S_MP_BALANCE_MUL_C
+#   define BN_S_MP_EXPTMOD_FAST_C
+#   define BN_S_MP_GET_BIT_C
+#   define BN_S_MP_INVMOD_FAST_C
+#   define BN_S_MP_INVMOD_SLOW_C
+#   define BN_S_MP_KARATSUBA_MUL_C
+#   define BN_S_MP_KARATSUBA_SQR_C
+#   define BN_S_MP_MONTGOMERY_REDUCE_FAST_C
+#   define BN_S_MP_MUL_DIGS_FAST_C
+#   define BN_S_MP_MUL_HIGH_DIGS_FAST_C
+#   define BN_S_MP_PRIME_IS_DIVISIBLE_C
+#   define BN_S_MP_PRIME_RANDOM_EX_C
+#   define BN_S_MP_RAND_SOURCE_C
+#   define BN_S_MP_REVERSE_C
+#   define BN_S_MP_SQR_FAST_C
+#   define BN_S_MP_TOOM_MUL_C
+#   define BN_S_MP_TOOM_SQR_C
 #endif
 
 #if defined(BN_MP_2EXPT_C)
-#   define BN_MP_ZERO_C
 #   define BN_MP_GROW_C
+#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_ABS_C)
@@ -214,29 +270,27 @@
 #endif
 
 #if defined(BN_MP_ADD_C)
-#   define BN_S_MP_ADD_C
 #   define BN_MP_CMP_MAG_C
+#   define BN_S_MP_ADD_C
 #   define BN_S_MP_SUB_C
 #endif
 
 #if defined(BN_MP_ADD_D_C)
+#   define BN_MP_CLAMP_C
 #   define BN_MP_GROW_C
 #   define BN_MP_SUB_D_C
-#   define BN_MP_CLAMP_C
 #endif
 
 #if defined(BN_MP_ADDMOD_C)
-#   define BN_MP_INIT_C
 #   define BN_MP_ADD_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_C
 #   define BN_MP_MOD_C
 #endif
 
 #if defined(BN_MP_AND_C)
-#   define BN_MP_INIT_COPY_C
 #   define BN_MP_CLAMP_C
-#   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_GROW_C
 #endif
 
 #if defined(BN_MP_CLAMP_C)
@@ -260,7 +314,6 @@
 #endif
 
 #if defined(BN_MP_CNT_LSB_C)
-#   define BN_MP_ISZERO_C
 #endif
 
 #if defined(BN_MP_COMPLEMENT_C)
@@ -275,212 +328,221 @@
 #if defined(BN_MP_COUNT_BITS_C)
 #endif
 
+#if defined(BN_MP_DECR_C)
+#   define BN_MP_INCR_C
+#   define BN_MP_SET_C
+#   define BN_MP_SUB_D_C
+#   define BN_MP_ZERO_C
+#endif
+
 #if defined(BN_MP_DIV_C)
-#   define BN_MP_ISZERO_C
+#   define BN_MP_ADD_C
+#   define BN_MP_CLAMP_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_C
 #   define BN_MP_CMP_MAG_C
 #   define BN_MP_COPY_C
-#   define BN_MP_ZERO_C
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_SET_C
 #   define BN_MP_COUNT_BITS_C
-#   define BN_MP_ABS_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_CMP_C
-#   define BN_MP_SUB_C
-#   define BN_MP_ADD_C
 #   define BN_MP_DIV_2D_C
 #   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_MULTI_C
-#   define BN_MP_INIT_SIZE_C
 #   define BN_MP_INIT_C
 #   define BN_MP_INIT_COPY_C
+#   define BN_MP_INIT_SIZE_C
 #   define BN_MP_LSHD_C
-#   define BN_MP_RSHD_C
+#   define BN_MP_MUL_2D_C
 #   define BN_MP_MUL_D_C
-#   define BN_MP_CLAMP_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_RSHD_C
+#   define BN_MP_SUB_C
+#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_DIV_2_C)
-#   define BN_MP_GROW_C
 #   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
 #endif
 
 #if defined(BN_MP_DIV_2D_C)
+#   define BN_MP_CLAMP_C
 #   define BN_MP_COPY_C
-#   define BN_MP_ZERO_C
 #   define BN_MP_MOD_2D_C
 #   define BN_MP_RSHD_C
-#   define BN_MP_CLAMP_C
+#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_DIV_3_C)
-#   define BN_MP_INIT_SIZE_C
 #   define BN_MP_CLAMP_C
-#   define BN_MP_EXCH_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_INIT_SIZE_C
 #endif
 
 #if defined(BN_MP_DIV_D_C)
-#   define BN_MP_ISZERO_C
+#   define BN_MP_CLAMP_C
+#   define BN_MP_CLEAR_C
 #   define BN_MP_COPY_C
 #   define BN_MP_DIV_2D_C
 #   define BN_MP_DIV_3_C
-#   define BN_MP_INIT_SIZE_C
-#   define BN_MP_CLAMP_C
 #   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_SIZE_C
 #endif
 
 #if defined(BN_MP_DR_IS_MODULUS_C)
 #endif
 
 #if defined(BN_MP_DR_REDUCE_C)
-#   define BN_MP_GROW_C
 #   define BN_MP_CLAMP_C
 #   define BN_MP_CMP_MAG_C
+#   define BN_MP_GROW_C
 #   define BN_S_MP_SUB_C
 #endif
 
 #if defined(BN_MP_DR_SETUP_C)
 #endif
 
-#if defined(BN_MP_EXCH_C)
+#if defined(BN_MP_ERROR_TO_STRING_C)
 #endif
 
-#if defined(BN_MP_EXPORT_C)
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_COUNT_BITS_C
-#   define BN_MP_DIV_2D_C
-#   define BN_MP_CLEAR_C
-#endif
-
-#if defined(BN_MP_EXPT_D_C)
-#   define BN_MP_EXPT_D_EX_C
+#if defined(BN_MP_EXCH_C)
 #endif
 
-#if defined(BN_MP_EXPT_D_EX_C)
+#if defined(BN_MP_EXPT_U32_C)
+#   define BN_MP_CLEAR_C
 #   define BN_MP_INIT_COPY_C
-#   define BN_MP_SET_C
 #   define BN_MP_MUL_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_SET_C
 #   define BN_MP_SQR_C
 #endif
 
 #if defined(BN_MP_EXPTMOD_C)
-#   define BN_MP_INIT_C
-#   define BN_MP_INVMOD_C
-#   define BN_MP_CLEAR_C
 #   define BN_MP_ABS_C
 #   define BN_MP_CLEAR_MULTI_C
-#   define BN_MP_REDUCE_IS_2K_L_C
-#   define BN_S_MP_EXPTMOD_C
 #   define BN_MP_DR_IS_MODULUS_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_INVMOD_C
 #   define BN_MP_REDUCE_IS_2K_C
-#   define BN_MP_ISODD_C
-#   define BN_MP_EXPTMOD_FAST_C
-#endif
-
-#if defined(BN_MP_EXPTMOD_FAST_C)
-#   define BN_MP_COUNT_BITS_C
-#   define BN_MP_INIT_SIZE_C
-#   define BN_MP_CLEAR_C
-#   define BN_MP_MONTGOMERY_SETUP_C
-#   define BN_FAST_MP_MONTGOMERY_REDUCE_C
-#   define BN_MP_MONTGOMERY_REDUCE_C
-#   define BN_MP_DR_SETUP_C
-#   define BN_MP_DR_REDUCE_C
-#   define BN_MP_REDUCE_2K_SETUP_C
-#   define BN_MP_REDUCE_2K_C
-#   define BN_MP_MONTGOMERY_CALC_NORMALIZATION_C
-#   define BN_MP_MULMOD_C
-#   define BN_MP_SET_C
-#   define BN_MP_MOD_C
-#   define BN_MP_COPY_C
-#   define BN_MP_SQR_C
-#   define BN_MP_MUL_C
-#   define BN_MP_EXCH_C
+#   define BN_MP_REDUCE_IS_2K_L_C
+#   define BN_S_MP_EXPTMOD_C
+#   define BN_S_MP_EXPTMOD_FAST_C
 #endif
 
 #if defined(BN_MP_EXTEUCLID_C)
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_SET_C
+#   define BN_MP_CLEAR_MULTI_C
 #   define BN_MP_COPY_C
-#   define BN_MP_ISZERO_C
 #   define BN_MP_DIV_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_INIT_MULTI_C
 #   define BN_MP_MUL_C
-#   define BN_MP_SUB_C
 #   define BN_MP_NEG_C
-#   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_SET_C
+#   define BN_MP_SUB_C
 #endif
 
 #if defined(BN_MP_FREAD_C)
-#   define BN_MP_ZERO_C
-#   define BN_MP_S_RMAP_REVERSE_SZ_C
-#   define BN_MP_S_RMAP_REVERSE_C
-#   define BN_MP_MUL_D_C
 #   define BN_MP_ADD_D_C
-#   define BN_MP_CMP_D_C
+#   define BN_MP_MUL_D_C
+#   define BN_MP_ZERO_C
+#endif
+
+#if defined(BN_MP_FROM_SBIN_C)
+#   define BN_MP_FROM_UBIN_C
+#endif
+
+#if defined(BN_MP_FROM_UBIN_C)
+#   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
+#   define BN_MP_MUL_2D_C
+#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_FWRITE_C)
 #   define BN_MP_RADIX_SIZE_C
-#   define BN_MP_TORADIX_C
+#   define BN_MP_TO_RADIX_C
 #endif
 
 #if defined(BN_MP_GCD_C)
-#   define BN_MP_ISZERO_C
 #   define BN_MP_ABS_C
-#   define BN_MP_INIT_COPY_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_MAG_C
 #   define BN_MP_CNT_LSB_C
 #   define BN_MP_DIV_2D_C
-#   define BN_MP_CMP_MAG_C
 #   define BN_MP_EXCH_C
-#   define BN_S_MP_SUB_C
+#   define BN_MP_INIT_COPY_C
 #   define BN_MP_MUL_2D_C
-#   define BN_MP_CLEAR_C
+#   define BN_S_MP_SUB_C
 #endif
 
-#if defined(BN_S_MP_GET_BIT_C)
-#   define BN_MP_ISZERO_C
+#if defined(BN_MP_GET_DOUBLE_C)
 #endif
 
-#if defined(BN_MP_GET_DOUBLE_C)
-#   define BN_MP_ISNEG_C
+#if defined(BN_MP_GET_I32_C)
+#   define BN_MP_GET_MAG_U32_C
+#endif
+
+#if defined(BN_MP_GET_I64_C)
+#   define BN_MP_GET_MAG_U64_C
+#endif
+
+#if defined(BN_MP_GET_L_C)
+#   define BN_MP_GET_MAG_UL_C
+#endif
+
+#if defined(BN_MP_GET_LL_C)
+#   define BN_MP_GET_MAG_ULL_C
 #endif
 
-#if defined(BN_MP_GET_INT_C)
+#if defined(BN_MP_GET_MAG_U32_C)
 #endif
 
-#if defined(BN_MP_GET_LONG_C)
+#if defined(BN_MP_GET_MAG_U64_C)
 #endif
 
-#if defined(BN_MP_GET_LONG_LONG_C)
+#if defined(BN_MP_GET_MAG_UL_C)
+#endif
+
+#if defined(BN_MP_GET_MAG_ULL_C)
 #endif
 
 #if defined(BN_MP_GROW_C)
 #endif
 
-#if defined(BN_MP_IMPORT_C)
-#   define BN_MP_ZERO_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_CLAMP_C
+#if defined(BN_MP_INCR_C)
+#   define BN_MP_ADD_D_C
+#   define BN_MP_DECR_C
+#   define BN_MP_SET_C
 #endif
 
 #if defined(BN_MP_INIT_C)
 #endif
 
 #if defined(BN_MP_INIT_COPY_C)
-#   define BN_MP_INIT_SIZE_C
-#   define BN_MP_COPY_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_COPY_C
+#   define BN_MP_INIT_SIZE_C
 #endif
 
-#if defined(BN_MP_INIT_MULTI_C)
-#   define BN_MP_ERR_C
+#if defined(BN_MP_INIT_I32_C)
+#   define BN_MP_INIT_C
+#   define BN_MP_SET_I32_C
+#endif
+
+#if defined(BN_MP_INIT_I64_C)
 #   define BN_MP_INIT_C
+#   define BN_MP_SET_I64_C
+#endif
+
+#if defined(BN_MP_INIT_L_C)
+#   define BN_MP_INIT_C
+#   define BN_MP_SET_L_C
+#endif
+
+#if defined(BN_MP_INIT_LL_C)
+#   define BN_MP_INIT_C
+#   define BN_MP_SET_LL_C
+#endif
+
+#if defined(BN_MP_INIT_MULTI_C)
 #   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_C
 #endif
 
 #if defined(BN_MP_INIT_SET_C)
@@ -488,119 +550,102 @@
 #   define BN_MP_SET_C
 #endif
 
-#if defined(BN_MP_INIT_SET_INT_C)
+#if defined(BN_MP_INIT_SIZE_C)
+#endif
+
+#if defined(BN_MP_INIT_U32_C)
 #   define BN_MP_INIT_C
-#   define BN_MP_SET_INT_C
+#   define BN_MP_SET_U32_C
 #endif
 
-#if defined(BN_MP_INIT_SIZE_C)
+#if defined(BN_MP_INIT_U64_C)
 #   define BN_MP_INIT_C
+#   define BN_MP_SET_U64_C
 #endif
 
-#if defined(BN_MP_INVMOD_C)
-#   define BN_MP_CMP_D_C
-#   define BN_MP_ISODD_C
-#   define BN_FAST_MP_INVMOD_C
-#   define BN_MP_INVMOD_SLOW_C
+#if defined(BN_MP_INIT_UL_C)
+#   define BN_MP_INIT_C
+#   define BN_MP_SET_UL_C
 #endif
 
-#if defined(BN_MP_INVMOD_SLOW_C)
-#   define BN_MP_ISZERO_C
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_MOD_C
-#   define BN_MP_COPY_C
-#   define BN_MP_ISEVEN_C
-#   define BN_MP_SET_C
-#   define BN_MP_DIV_2_C
-#   define BN_MP_ISODD_C
-#   define BN_MP_ADD_C
-#   define BN_MP_SUB_C
-#   define BN_MP_CMP_C
+#if defined(BN_MP_INIT_ULL_C)
+#   define BN_MP_INIT_C
+#   define BN_MP_SET_ULL_C
+#endif
+
+#if defined(BN_MP_INVMOD_C)
 #   define BN_MP_CMP_D_C
-#   define BN_MP_CMP_MAG_C
-#   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_MULTI_C
+#   define BN_S_MP_INVMOD_FAST_C
+#   define BN_S_MP_INVMOD_SLOW_C
 #endif
 
 #if defined(BN_MP_IS_SQUARE_C)
-#   define BN_MP_MOD_D_C
-#   define BN_MP_INIT_SET_INT_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_MAG_C
+#   define BN_MP_GET_I32_C
+#   define BN_MP_INIT_U32_C
 #   define BN_MP_MOD_C
-#   define BN_MP_GET_INT_C
+#   define BN_MP_MOD_D_C
 #   define BN_MP_SQRT_C
 #   define BN_MP_SQR_C
-#   define BN_MP_CMP_MAG_C
-#   define BN_MP_CLEAR_C
 #endif
 
-#if defined(BN_MP_JACOBI_C)
-#   define BN_MP_KRONECKER_C
-#   define BN_MP_ISNEG_C
-#   define BN_MP_CMP_D_C
+#if defined(BN_MP_ISEVEN_C)
 #endif
 
-#if defined(BN_MP_KARATSUBA_MUL_C)
-#   define BN_MP_MUL_C
-#   define BN_MP_INIT_SIZE_C
-#   define BN_MP_CLAMP_C
-#   define BN_S_MP_ADD_C
-#   define BN_MP_ADD_C
-#   define BN_S_MP_SUB_C
-#   define BN_MP_LSHD_C
-#   define BN_MP_CLEAR_C
-#endif
-
-#if defined(BN_MP_KARATSUBA_SQR_C)
-#   define BN_MP_INIT_SIZE_C
-#   define BN_MP_CLAMP_C
-#   define BN_MP_SQR_C
-#   define BN_S_MP_ADD_C
-#   define BN_S_MP_SUB_C
-#   define BN_MP_LSHD_C
-#   define BN_MP_ADD_C
-#   define BN_MP_CLEAR_C
+#if defined(BN_MP_ISODD_C)
 #endif
 
 #if defined(BN_MP_KRONECKER_C)
-#   define BN_MP_ISZERO_C
-#   define BN_MP_ISEVEN_C
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_CNT_LSB_C
-#   define BN_MP_DIV_2D_C
+#   define BN_MP_CLEAR_C
 #   define BN_MP_CMP_D_C
+#   define BN_MP_CNT_LSB_C
 #   define BN_MP_COPY_C
+#   define BN_MP_DIV_2D_C
+#   define BN_MP_INIT_C
+#   define BN_MP_INIT_COPY_C
 #   define BN_MP_MOD_C
-#   define BN_MP_CLEAR_C
 #endif
 
 #if defined(BN_MP_LCM_C)
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_GCD_C
+#   define BN_MP_CLEAR_MULTI_C
 #   define BN_MP_CMP_MAG_C
 #   define BN_MP_DIV_C
+#   define BN_MP_GCD_C
+#   define BN_MP_INIT_MULTI_C
 #   define BN_MP_MUL_C
+#endif
+
+#if defined(BN_MP_LOG_U32_C)
 #   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_CMP_C
+#   define BN_MP_CMP_D_C
+#   define BN_MP_COPY_C
+#   define BN_MP_COUNT_BITS_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_EXPT_U32_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_MUL_C
+#   define BN_MP_SET_C
+#   define BN_MP_SQR_C
 #endif
 
 #if defined(BN_MP_LSHD_C)
-#   define BN_MP_ISZERO_C
 #   define BN_MP_GROW_C
-#   define BN_MP_RSHD_C
 #endif
 
 #if defined(BN_MP_MOD_C)
-#   define BN_MP_INIT_SIZE_C
-#   define BN_MP_DIV_C
+#   define BN_MP_ADD_C
 #   define BN_MP_CLEAR_C
-#   define BN_MP_ISZERO_C
+#   define BN_MP_DIV_C
 #   define BN_MP_EXCH_C
-#   define BN_MP_ADD_C
+#   define BN_MP_INIT_SIZE_C
 #endif
 
 #if defined(BN_MP_MOD_2D_C)
-#   define BN_MP_ZERO_C
-#   define BN_MP_COPY_C
 #   define BN_MP_CLAMP_C
+#   define BN_MP_COPY_C
+#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_MOD_D_C)
@@ -608,20 +653,20 @@
 #endif
 
 #if defined(BN_MP_MONTGOMERY_CALC_NORMALIZATION_C)
-#   define BN_MP_COUNT_BITS_C
 #   define BN_MP_2EXPT_C
-#   define BN_MP_SET_C
-#   define BN_MP_MUL_2_C
 #   define BN_MP_CMP_MAG_C
+#   define BN_MP_COUNT_BITS_C
+#   define BN_MP_MUL_2_C
+#   define BN_MP_SET_C
 #   define BN_S_MP_SUB_C
 #endif
 
 #if defined(BN_MP_MONTGOMERY_REDUCE_C)
-#   define BN_FAST_MP_MONTGOMERY_REDUCE_C
-#   define BN_MP_GROW_C
 #   define BN_MP_CLAMP_C
-#   define BN_MP_RSHD_C
 #   define BN_MP_CMP_MAG_C
+#   define BN_MP_GROW_C
+#   define BN_MP_RSHD_C
+#   define BN_S_MP_MONTGOMERY_REDUCE_FAST_C
 #   define BN_S_MP_SUB_C
 #endif
 
@@ -629,11 +674,11 @@
 #endif
 
 #if defined(BN_MP_MUL_C)
-#   define BN_MP_TOOM_MUL_C
-#   define BN_MP_KARATSUBA_MUL_C
-#   define BN_FAST_S_MP_MUL_DIGS_C
-#   define BN_S_MP_MUL_C
+#   define BN_S_MP_BALANCE_MUL_C
+#   define BN_S_MP_KARATSUBA_MUL_C
 #   define BN_S_MP_MUL_DIGS_C
+#   define BN_S_MP_MUL_DIGS_FAST_C
+#   define BN_S_MP_TOOM_MUL_C
 #endif
 
 #if defined(BN_MP_MUL_2_C)
@@ -641,276 +686,238 @@
 #endif
 
 #if defined(BN_MP_MUL_2D_C)
+#   define BN_MP_CLAMP_C
 #   define BN_MP_COPY_C
 #   define BN_MP_GROW_C
 #   define BN_MP_LSHD_C
-#   define BN_MP_CLAMP_C
 #endif
 
 #if defined(BN_MP_MUL_D_C)
-#   define BN_MP_GROW_C
 #   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
 #endif
 
 #if defined(BN_MP_MULMOD_C)
-#   define BN_MP_INIT_SIZE_C
-#   define BN_MP_MUL_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_SIZE_C
 #   define BN_MP_MOD_C
-#endif
-
-#if defined(BN_MP_N_ROOT_C)
-#   define BN_MP_N_ROOT_EX_C
-#endif
-
-#if defined(BN_MP_N_ROOT_EX_C)
-#   define BN_MP_INIT_C
-#   define BN_MP_SET_C
-#   define BN_MP_COPY_C
-#   define BN_MP_EXPT_D_EX_C
 #   define BN_MP_MUL_C
-#   define BN_MP_SUB_C
-#   define BN_MP_MUL_D_C
-#   define BN_MP_DIV_C
-#   define BN_MP_CMP_C
-#   define BN_MP_SUB_D_C
-#   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_C
 #endif
 
 #if defined(BN_MP_NEG_C)
 #   define BN_MP_COPY_C
-#   define BN_MP_ISZERO_C
 #endif
 
 #if defined(BN_MP_OR_C)
-#   define BN_MP_INIT_COPY_C
 #   define BN_MP_CLAMP_C
-#   define BN_MP_EXCH_C
+#   define BN_MP_GROW_C
+#endif
+
+#if defined(BN_MP_PACK_C)
 #   define BN_MP_CLEAR_C
+#   define BN_MP_DIV_2D_C
+#   define BN_MP_INIT_COPY_C
+#   define BN_MP_PACK_COUNT_C
+#endif
+
+#if defined(BN_MP_PACK_COUNT_C)
+#   define BN_MP_COUNT_BITS_C
 #endif
 
 #if defined(BN_MP_PRIME_FERMAT_C)
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_C
 #   define BN_MP_CMP_D_C
-#   define BN_MP_INIT_C
 #   define BN_MP_EXPTMOD_C
-#   define BN_MP_CMP_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_C
 #endif
 
 #if defined(BN_MP_PRIME_FROBENIUS_UNDERWOOD_C)
-#   define BN_MP_PRIME_IS_PRIME_C
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_SET_LONG_C
-#   define BN_MP_SQR_C
-#   define BN_MP_SUB_D_C
-#   define BN_MP_KRONECKER_C
-#   define BN_MP_GCD_C
+#   define BN_MP_ADD_C
 #   define BN_MP_ADD_D_C
-#   define BN_MP_SET_C
+#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_CMP_C
 #   define BN_MP_COUNT_BITS_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_GCD_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_KRONECKER_C
+#   define BN_MP_MOD_C
 #   define BN_MP_MUL_2_C
-#   define BN_MP_MUL_D_C
-#   define BN_MP_ADD_C
 #   define BN_MP_MUL_C
+#   define BN_MP_MUL_D_C
+#   define BN_MP_SET_C
+#   define BN_MP_SET_U32_C
+#   define BN_MP_SQR_C
 #   define BN_MP_SUB_C
-#   define BN_MP_MOD_C
+#   define BN_MP_SUB_D_C
 #   define BN_S_MP_GET_BIT_C
-#   define BN_MP_EXCH_C
-#   define BN_MP_ISZERO_C
-#   define BN_MP_CMP_C
-#   define BN_MP_CLEAR_MULTI_C
-#endif
-
-#if defined(BN_MP_PRIME_IS_DIVISIBLE_C)
-#   define BN_MP_MOD_D_C
 #endif
 
 #if defined(BN_MP_PRIME_IS_PRIME_C)
-#   define BN_MP_ISEVEN_C
-#   define BN_MP_IS_SQUARE_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_C
 #   define BN_MP_CMP_D_C
-#   define BN_MP_PRIME_IS_DIVISIBLE_C
+#   define BN_MP_COUNT_BITS_C
+#   define BN_MP_DIV_2D_C
 #   define BN_MP_INIT_SET_C
+#   define BN_MP_IS_SQUARE_C
 #   define BN_MP_PRIME_MILLER_RABIN_C
-#   define BN_MP_PRIME_FROBENIUS_UNDERWOOD_C
 #   define BN_MP_PRIME_STRONG_LUCAS_SELFRIDGE_C
+#   define BN_MP_RAND_C
 #   define BN_MP_READ_RADIX_C
-#   define BN_MP_CMP_C
 #   define BN_MP_SET_C
-#   define BN_MP_COUNT_BITS_C
-#   define BN_MP_RAND_C
-#   define BN_MP_DIV_2D_C
-#   define BN_MP_CLEAR_C
+#   define BN_S_MP_PRIME_IS_DIVISIBLE_C
 #endif
 
 #if defined(BN_MP_PRIME_MILLER_RABIN_C)
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_C
 #   define BN_MP_CMP_D_C
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_SUB_D_C
 #   define BN_MP_CNT_LSB_C
 #   define BN_MP_DIV_2D_C
 #   define BN_MP_EXPTMOD_C
-#   define BN_MP_CMP_C
+#   define BN_MP_INIT_C
+#   define BN_MP_INIT_COPY_C
 #   define BN_MP_SQRMOD_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_SUB_D_C
 #endif
 
 #if defined(BN_MP_PRIME_NEXT_PRIME_C)
+#   define BN_MP_ADD_D_C
+#   define BN_MP_CLEAR_C
 #   define BN_MP_CMP_D_C
-#   define BN_MP_SET_C
-#   define BN_MP_SUB_D_C
-#   define BN_MP_ISEVEN_C
-#   define BN_MP_MOD_D_C
 #   define BN_MP_INIT_C
-#   define BN_MP_ADD_D_C
+#   define BN_MP_MOD_D_C
 #   define BN_MP_PRIME_IS_PRIME_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_SET_C
+#   define BN_MP_SUB_D_C
 #endif
 
 #if defined(BN_MP_PRIME_RABIN_MILLER_TRIALS_C)
 #endif
 
-#if defined(BN_MP_PRIME_RANDOM_EX_C)
-#   define BN_MP_READ_UNSIGNED_BIN_C
-#   define BN_MP_PRIME_IS_PRIME_C
-#   define BN_MP_SUB_D_C
+#if defined(BN_MP_PRIME_RAND_C)
+#   define BN_MP_ADD_D_C
 #   define BN_MP_DIV_2_C
+#   define BN_MP_FROM_UBIN_C
 #   define BN_MP_MUL_2_C
-#   define BN_MP_ADD_D_C
+#   define BN_MP_PRIME_IS_PRIME_C
+#   define BN_MP_SUB_D_C
+#   define BN_S_MP_PRIME_RANDOM_EX_C
+#   define BN_S_MP_RAND_CB_C
+#   define BN_S_MP_RAND_SOURCE_C
 #endif
 
 #if defined(BN_MP_PRIME_STRONG_LUCAS_SELFRIDGE_C)
-#   define BN_MP_PRIME_IS_PRIME_C
-#   define BN_MP_MUL_D_C
-#   define BN_S_MP_MUL_SI_C
-#   define BN_MP_INIT_C
-#   define BN_MP_SET_LONG_C
-#   define BN_MP_MUL_C
+#   define BN_MP_ADD_C
+#   define BN_MP_ADD_D_C
 #   define BN_MP_CLEAR_C
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_GCD_C
-#   define BN_MP_CMP_D_C
+#   define BN_MP_CLEAR_MULTI_C
 #   define BN_MP_CMP_C
-#   define BN_MP_KRONECKER_C
-#   define BN_MP_ADD_D_C
+#   define BN_MP_CMP_D_C
 #   define BN_MP_CNT_LSB_C
-#   define BN_MP_DIV_2D_C
-#   define BN_MP_SET_C
-#   define BN_MP_MUL_2_C
 #   define BN_MP_COUNT_BITS_C
+#   define BN_MP_DIV_2D_C
+#   define BN_MP_DIV_2_C
+#   define BN_MP_GCD_C
+#   define BN_MP_INIT_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_KRONECKER_C
 #   define BN_MP_MOD_C
+#   define BN_MP_MUL_2_C
+#   define BN_MP_MUL_C
+#   define BN_MP_SET_C
+#   define BN_MP_SET_I32_C
+#   define BN_MP_SET_U32_C
 #   define BN_MP_SQR_C
 #   define BN_MP_SUB_C
-#   define BN_S_MP_GET_BIT_C
-#   define BN_MP_ADD_C
-#   define BN_MP_ISODD_C
-#   define BN_MP_DIV_2_C
 #   define BN_MP_SUB_D_C
-#   define BN_MP_ISZERO_C
-#   define BN_MP_CLEAR_MULTI_C
+#   define BN_S_MP_GET_BIT_C
+#   define BN_S_MP_MUL_SI_C
 #endif
 
 #if defined(BN_MP_RADIX_SIZE_C)
-#   define BN_MP_ISZERO_C
+#   define BN_MP_CLEAR_C
 #   define BN_MP_COUNT_BITS_C
-#   define BN_MP_INIT_COPY_C
 #   define BN_MP_DIV_D_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_COPY_C
 #endif
 
 #if defined(BN_MP_RADIX_SMAP_C)
-#   define BN_MP_S_RMAP_C
-#   define BN_MP_S_RMAP_REVERSE_C
-#   define BN_MP_S_RMAP_REVERSE_SZ_C
 #endif
 
 #if defined(BN_MP_RAND_C)
-#   define BN_MP_RAND_DIGIT_C
+#   define BN_MP_GROW_C
+#   define BN_MP_RAND_SOURCE_C
 #   define BN_MP_ZERO_C
-#   define BN_MP_ADD_D_C
-#   define BN_MP_LSHD_C
+#   define BN_S_MP_RAND_PLATFORM_C
+#   define BN_S_MP_RAND_SOURCE_C
 #endif
 
 #if defined(BN_MP_READ_RADIX_C)
-#   define BN_MP_ZERO_C
-#   define BN_MP_S_RMAP_REVERSE_SZ_C
-#   define BN_MP_S_RMAP_REVERSE_C
-#   define BN_MP_MUL_D_C
 #   define BN_MP_ADD_D_C
-#   define BN_MP_ISZERO_C
-#endif
-
-#if defined(BN_MP_READ_SIGNED_BIN_C)
-#   define BN_MP_READ_UNSIGNED_BIN_C
-#endif
-
-#if defined(BN_MP_READ_UNSIGNED_BIN_C)
-#   define BN_MP_GROW_C
+#   define BN_MP_MUL_D_C
 #   define BN_MP_ZERO_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_CLAMP_C
 #endif
 
 #if defined(BN_MP_REDUCE_C)
-#   define BN_MP_REDUCE_SETUP_C
+#   define BN_MP_ADD_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_C
+#   define BN_MP_CMP_D_C
 #   define BN_MP_INIT_COPY_C
-#   define BN_MP_RSHD_C
-#   define BN_MP_MUL_C
-#   define BN_S_MP_MUL_HIGH_DIGS_C
-#   define BN_FAST_S_MP_MUL_HIGH_DIGS_C
+#   define BN_MP_LSHD_C
 #   define BN_MP_MOD_2D_C
-#   define BN_S_MP_MUL_DIGS_C
-#   define BN_MP_SUB_C
-#   define BN_MP_CMP_D_C
+#   define BN_MP_MUL_C
+#   define BN_MP_RSHD_C
 #   define BN_MP_SET_C
-#   define BN_MP_LSHD_C
-#   define BN_MP_ADD_C
-#   define BN_MP_CMP_C
+#   define BN_MP_SUB_C
+#   define BN_S_MP_MUL_DIGS_C
+#   define BN_S_MP_MUL_HIGH_DIGS_C
+#   define BN_S_MP_MUL_HIGH_DIGS_FAST_C
 #   define BN_S_MP_SUB_C
-#   define BN_MP_CLEAR_C
 #endif
 
 #if defined(BN_MP_REDUCE_2K_C)
-#   define BN_MP_INIT_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_MAG_C
 #   define BN_MP_COUNT_BITS_C
 #   define BN_MP_DIV_2D_C
+#   define BN_MP_INIT_C
 #   define BN_MP_MUL_D_C
 #   define BN_S_MP_ADD_C
-#   define BN_MP_CMP_MAG_C
 #   define BN_S_MP_SUB_C
-#   define BN_MP_CLEAR_C
 #endif
 
 #if defined(BN_MP_REDUCE_2K_L_C)
-#   define BN_MP_INIT_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CMP_MAG_C
 #   define BN_MP_COUNT_BITS_C
 #   define BN_MP_DIV_2D_C
+#   define BN_MP_INIT_C
 #   define BN_MP_MUL_C
 #   define BN_S_MP_ADD_C
-#   define BN_MP_CMP_MAG_C
 #   define BN_S_MP_SUB_C
-#   define BN_MP_CLEAR_C
 #endif
 
 #if defined(BN_MP_REDUCE_2K_SETUP_C)
-#   define BN_MP_INIT_C
-#   define BN_MP_COUNT_BITS_C
 #   define BN_MP_2EXPT_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_COUNT_BITS_C
+#   define BN_MP_INIT_C
 #   define BN_S_MP_SUB_C
 #endif
 
 #if defined(BN_MP_REDUCE_2K_SETUP_L_C)
-#   define BN_MP_INIT_C
 #   define BN_MP_2EXPT_C
+#   define BN_MP_CLEAR_C
 #   define BN_MP_COUNT_BITS_C
+#   define BN_MP_INIT_C
 #   define BN_S_MP_SUB_C
-#   define BN_MP_CLEAR_C
 #endif
 
 #if defined(BN_MP_REDUCE_IS_2K_C)
-#   define BN_MP_REDUCE_2K_C
 #   define BN_MP_COUNT_BITS_C
 #endif
 
@@ -922,302 +929,391 @@
 #   define BN_MP_DIV_C
 #endif
 
+#if defined(BN_MP_ROOT_U32_C)
+#   define BN_MP_2EXPT_C
+#   define BN_MP_ADD_D_C
+#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_CMP_C
+#   define BN_MP_COPY_C
+#   define BN_MP_COUNT_BITS_C
+#   define BN_MP_DIV_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_EXPT_U32_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_MUL_C
+#   define BN_MP_MUL_D_C
+#   define BN_MP_SET_C
+#   define BN_MP_SUB_C
+#   define BN_MP_SUB_D_C
+#endif
+
 #if defined(BN_MP_RSHD_C)
 #   define BN_MP_ZERO_C
 #endif
 
+#if defined(BN_MP_SBIN_SIZE_C)
+#   define BN_MP_UBIN_SIZE_C
+#endif
+
 #if defined(BN_MP_SET_C)
-#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_SET_DOUBLE_C)
-#   define BN_MP_SET_LONG_LONG_C
 #   define BN_MP_DIV_2D_C
 #   define BN_MP_MUL_2D_C
-#   define BN_MP_ISZERO_C
+#   define BN_MP_SET_U64_C
 #endif
 
-#if defined(BN_MP_SET_INT_C)
-#   define BN_MP_ZERO_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_CLAMP_C
+#if defined(BN_MP_SET_I32_C)
+#   define BN_MP_SET_U32_C
+#endif
+
+#if defined(BN_MP_SET_I64_C)
+#   define BN_MP_SET_U64_C
+#endif
+
+#if defined(BN_MP_SET_L_C)
+#   define BN_MP_SET_UL_C
+#endif
+
+#if defined(BN_MP_SET_LL_C)
+#   define BN_MP_SET_ULL_C
+#endif
+
+#if defined(BN_MP_SET_U32_C)
+#endif
+
+#if defined(BN_MP_SET_U64_C)
 #endif
 
-#if defined(BN_MP_SET_LONG_C)
+#if defined(BN_MP_SET_UL_C)
 #endif
 
-#if defined(BN_MP_SET_LONG_LONG_C)
+#if defined(BN_MP_SET_ULL_C)
 #endif
 
 #if defined(BN_MP_SHRINK_C)
 #endif
 
-#if defined(BN_MP_SIGNED_BIN_SIZE_C)
-#   define BN_MP_UNSIGNED_BIN_SIZE_C
+#if defined(BN_MP_SIGNED_RSH_C)
+#   define BN_MP_ADD_D_C
+#   define BN_MP_DIV_2D_C
+#   define BN_MP_SUB_D_C
 #endif
 
 #if defined(BN_MP_SQR_C)
-#   define BN_MP_TOOM_SQR_C
-#   define BN_MP_KARATSUBA_SQR_C
-#   define BN_FAST_S_MP_SQR_C
+#   define BN_S_MP_KARATSUBA_SQR_C
 #   define BN_S_MP_SQR_C
+#   define BN_S_MP_SQR_FAST_C
+#   define BN_S_MP_TOOM_SQR_C
 #endif
 
 #if defined(BN_MP_SQRMOD_C)
-#   define BN_MP_INIT_C
-#   define BN_MP_SQR_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_C
 #   define BN_MP_MOD_C
+#   define BN_MP_SQR_C
 #endif
 
 #if defined(BN_MP_SQRT_C)
-#   define BN_MP_N_ROOT_C
-#   define BN_MP_ISZERO_C
-#   define BN_MP_ZERO_C
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_RSHD_C
-#   define BN_MP_DIV_C
 #   define BN_MP_ADD_C
-#   define BN_MP_DIV_2_C
+#   define BN_MP_CLEAR_C
 #   define BN_MP_CMP_MAG_C
+#   define BN_MP_DIV_2_C
+#   define BN_MP_DIV_C
 #   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_C
+#   define BN_MP_INIT_COPY_C
+#   define BN_MP_RSHD_C
+#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_SQRTMOD_PRIME_C)
-#   define BN_MP_CMP_D_C
-#   define BN_MP_ZERO_C
-#   define BN_MP_JACOBI_C
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_MOD_D_C
 #   define BN_MP_ADD_D_C
+#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_CMP_D_C
+#   define BN_MP_COPY_C
 #   define BN_MP_DIV_2_C
 #   define BN_MP_EXPTMOD_C
-#   define BN_MP_COPY_C
-#   define BN_MP_SUB_D_C
-#   define BN_MP_ISEVEN_C
-#   define BN_MP_SET_INT_C
-#   define BN_MP_SQRMOD_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_KRONECKER_C
+#   define BN_MP_MOD_D_C
 #   define BN_MP_MULMOD_C
 #   define BN_MP_SET_C
-#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_SET_U32_C
+#   define BN_MP_SQRMOD_C
+#   define BN_MP_SUB_D_C
+#   define BN_MP_ZERO_C
 #endif
 
 #if defined(BN_MP_SUB_C)
-#   define BN_S_MP_ADD_C
 #   define BN_MP_CMP_MAG_C
+#   define BN_S_MP_ADD_C
 #   define BN_S_MP_SUB_C
 #endif
 
 #if defined(BN_MP_SUB_D_C)
-#   define BN_MP_GROW_C
 #   define BN_MP_ADD_D_C
 #   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
 #endif
 
 #if defined(BN_MP_SUBMOD_C)
-#   define BN_MP_INIT_C
-#   define BN_MP_SUB_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_C
 #   define BN_MP_MOD_C
+#   define BN_MP_SUB_C
 #endif
 
-#if defined(BN_MP_TC_AND_C)
-#   define BN_MP_ISNEG_C
-#   define BN_MP_COUNT_BITS_C
-#   define BN_MP_INIT_SET_INT_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_INIT_C
-#   define BN_MP_ADD_C
+#if defined(BN_MP_TO_RADIX_C)
 #   define BN_MP_CLEAR_C
-#   define BN_MP_AND_C
-#   define BN_MP_SUB_C
+#   define BN_MP_DIV_D_C
+#   define BN_MP_INIT_COPY_C
+#   define BN_S_MP_REVERSE_C
 #endif
 
-#if defined(BN_MP_SIGNED_RSH_C)
-#   define BN_MP_ISNEG_C
-#   define BN_MP_DIV_2D_C
-#   define BN_MP_ADD_D_C
-#   define BN_MP_SUB_D_C
+#if defined(BN_MP_TO_SBIN_C)
+#   define BN_MP_TO_UBIN_C
 #endif
 
-#if defined(BN_MP_TC_OR_C)
-#   define BN_MP_ISNEG_C
-#   define BN_MP_COUNT_BITS_C
-#   define BN_MP_INIT_SET_INT_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_INIT_C
-#   define BN_MP_ADD_C
+#if defined(BN_MP_TO_UBIN_C)
 #   define BN_MP_CLEAR_C
-#   define BN_MP_OR_C
-#   define BN_MP_SUB_C
+#   define BN_MP_DIV_2D_C
+#   define BN_MP_INIT_COPY_C
+#   define BN_MP_UBIN_SIZE_C
 #endif
 
-#if defined(BN_MP_TC_XOR_C)
-#   define BN_MP_ISNEG_C
+#if defined(BN_MP_UBIN_SIZE_C)
 #   define BN_MP_COUNT_BITS_C
-#   define BN_MP_INIT_SET_INT_C
+#endif
+
+#if defined(BN_MP_UNPACK_C)
+#   define BN_MP_CLAMP_C
 #   define BN_MP_MUL_2D_C
-#   define BN_MP_INIT_C
-#   define BN_MP_ADD_C
-#   define BN_MP_CLEAR_C
-#   define BN_MP_XOR_C
-#   define BN_MP_SUB_C
+#   define BN_MP_ZERO_C
 #endif
 
-#if defined(BN_MP_TO_SIGNED_BIN_C)
-#   define BN_MP_TO_UNSIGNED_BIN_C
+#if defined(BN_MP_XOR_C)
+#   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
 #endif
 
-#if defined(BN_MP_TO_SIGNED_BIN_N_C)
-#   define BN_MP_SIGNED_BIN_SIZE_C
-#   define BN_MP_TO_SIGNED_BIN_C
+#if defined(BN_MP_ZERO_C)
 #endif
 
-#if defined(BN_MP_TO_UNSIGNED_BIN_C)
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_ISZERO_C
-#   define BN_MP_DIV_2D_C
-#   define BN_MP_CLEAR_C
+#if defined(BN_PRIME_TAB_C)
 #endif
 
-#if defined(BN_MP_TO_UNSIGNED_BIN_N_C)
-#   define BN_MP_UNSIGNED_BIN_SIZE_C
-#   define BN_MP_TO_UNSIGNED_BIN_C
+#if defined(BN_S_MP_ADD_C)
+#   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
 #endif
 
-#if defined(BN_MP_TOOM_MUL_C)
-#   define BN_MP_INIT_MULTI_C
-#   define BN_MP_MOD_2D_C
-#   define BN_MP_COPY_C
-#   define BN_MP_RSHD_C
-#   define BN_MP_MUL_C
-#   define BN_MP_MUL_2_C
+#if defined(BN_S_MP_BALANCE_MUL_C)
 #   define BN_MP_ADD_C
-#   define BN_MP_SUB_C
-#   define BN_MP_DIV_2_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_MUL_D_C
-#   define BN_MP_DIV_3_C
-#   define BN_MP_LSHD_C
+#   define BN_MP_CLAMP_C
+#   define BN_MP_CLEAR_C
 #   define BN_MP_CLEAR_MULTI_C
-#endif
-
-#if defined(BN_MP_TOOM_SQR_C)
+#   define BN_MP_EXCH_C
 #   define BN_MP_INIT_MULTI_C
-#   define BN_MP_MOD_2D_C
-#   define BN_MP_COPY_C
-#   define BN_MP_RSHD_C
-#   define BN_MP_SQR_C
-#   define BN_MP_MUL_2_C
-#   define BN_MP_ADD_C
-#   define BN_MP_SUB_C
-#   define BN_MP_DIV_2_C
-#   define BN_MP_MUL_2D_C
-#   define BN_MP_MUL_D_C
-#   define BN_MP_DIV_3_C
+#   define BN_MP_INIT_SIZE_C
 #   define BN_MP_LSHD_C
-#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_MUL_C
 #endif
 
-#if defined(BN_MP_TORADIX_C)
-#   define BN_MP_ISZERO_C
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_DIV_D_C
+#if defined(BN_S_MP_EXPTMOD_C)
 #   define BN_MP_CLEAR_C
-#   define BN_MP_S_RMAP_C
+#   define BN_MP_COPY_C
+#   define BN_MP_COUNT_BITS_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_INIT_C
+#   define BN_MP_MOD_C
+#   define BN_MP_MUL_C
+#   define BN_MP_REDUCE_2K_L_C
+#   define BN_MP_REDUCE_2K_SETUP_L_C
+#   define BN_MP_REDUCE_C
+#   define BN_MP_REDUCE_SETUP_C
+#   define BN_MP_SET_C
+#   define BN_MP_SQR_C
 #endif
 
-#if defined(BN_MP_TORADIX_N_C)
-#   define BN_MP_ISZERO_C
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_DIV_D_C
+#if defined(BN_S_MP_EXPTMOD_FAST_C)
 #   define BN_MP_CLEAR_C
-#   define BN_MP_S_RMAP_C
+#   define BN_MP_COPY_C
+#   define BN_MP_COUNT_BITS_C
+#   define BN_MP_DR_REDUCE_C
+#   define BN_MP_DR_SETUP_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_INIT_SIZE_C
+#   define BN_MP_MOD_C
+#   define BN_MP_MONTGOMERY_CALC_NORMALIZATION_C
+#   define BN_MP_MONTGOMERY_REDUCE_C
+#   define BN_MP_MONTGOMERY_SETUP_C
+#   define BN_MP_MULMOD_C
+#   define BN_MP_MUL_C
+#   define BN_MP_REDUCE_2K_C
+#   define BN_MP_REDUCE_2K_SETUP_C
+#   define BN_MP_SET_C
+#   define BN_MP_SQR_C
+#   define BN_S_MP_MONTGOMERY_REDUCE_FAST_C
 #endif
 
-#if defined(BN_MP_UNSIGNED_BIN_SIZE_C)
-#   define BN_MP_COUNT_BITS_C
+#if defined(BN_S_MP_GET_BIT_C)
 #endif
 
-#if defined(BN_MP_XOR_C)
-#   define BN_MP_INIT_COPY_C
-#   define BN_MP_CLAMP_C
+#if defined(BN_S_MP_INVMOD_FAST_C)
+#   define BN_MP_ADD_C
+#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_CMP_C
+#   define BN_MP_CMP_D_C
+#   define BN_MP_CMP_MAG_C
+#   define BN_MP_COPY_C
+#   define BN_MP_DIV_2_C
 #   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_MOD_C
+#   define BN_MP_SET_C
+#   define BN_MP_SUB_C
 #endif
 
-#if defined(BN_MP_ZERO_C)
+#if defined(BN_S_MP_INVMOD_SLOW_C)
+#   define BN_MP_ADD_C
+#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_CMP_C
+#   define BN_MP_CMP_D_C
+#   define BN_MP_CMP_MAG_C
+#   define BN_MP_COPY_C
+#   define BN_MP_DIV_2_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_MOD_C
+#   define BN_MP_SET_C
+#   define BN_MP_SUB_C
 #endif
 
-#if defined(BN_PRIME_TAB_C)
+#if defined(BN_S_MP_KARATSUBA_MUL_C)
+#   define BN_MP_ADD_C
+#   define BN_MP_CLAMP_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_SIZE_C
+#   define BN_MP_LSHD_C
+#   define BN_MP_MUL_C
+#   define BN_S_MP_ADD_C
+#   define BN_S_MP_SUB_C
 #endif
 
-#if defined(BN_REVERSE_C)
+#if defined(BN_S_MP_KARATSUBA_SQR_C)
+#   define BN_MP_ADD_C
+#   define BN_MP_CLAMP_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_INIT_SIZE_C
+#   define BN_MP_LSHD_C
+#   define BN_MP_SQR_C
+#   define BN_S_MP_ADD_C
+#   define BN_S_MP_SUB_C
 #endif
 
-#if defined(BN_S_MP_ADD_C)
-#   define BN_MP_GROW_C
+#if defined(BN_S_MP_MONTGOMERY_REDUCE_FAST_C)
 #   define BN_MP_CLAMP_C
+#   define BN_MP_CMP_MAG_C
+#   define BN_MP_GROW_C
+#   define BN_S_MP_SUB_C
 #endif
 
-#if defined(BN_S_MP_EXPTMOD_C)
-#   define BN_MP_COUNT_BITS_C
-#   define BN_MP_INIT_C
+#if defined(BN_S_MP_MUL_DIGS_C)
+#   define BN_MP_CLAMP_C
 #   define BN_MP_CLEAR_C
-#   define BN_MP_REDUCE_SETUP_C
-#   define BN_MP_REDUCE_C
-#   define BN_MP_REDUCE_2K_SETUP_L_C
-#   define BN_MP_REDUCE_2K_L_C
-#   define BN_MP_MOD_C
-#   define BN_MP_COPY_C
-#   define BN_MP_SQR_C
-#   define BN_MP_MUL_C
-#   define BN_MP_SET_C
 #   define BN_MP_EXCH_C
+#   define BN_MP_INIT_SIZE_C
+#   define BN_S_MP_MUL_DIGS_FAST_C
 #endif
 
-#if defined(BN_S_MP_MUL_DIGS_C)
-#   define BN_FAST_S_MP_MUL_DIGS_C
-#   define BN_MP_INIT_SIZE_C
+#if defined(BN_S_MP_MUL_DIGS_FAST_C)
 #   define BN_MP_CLAMP_C
-#   define BN_MP_EXCH_C
-#   define BN_MP_CLEAR_C
+#   define BN_MP_GROW_C
 #endif
 
 #if defined(BN_S_MP_MUL_HIGH_DIGS_C)
-#   define BN_FAST_S_MP_MUL_HIGH_DIGS_C
-#   define BN_MP_INIT_SIZE_C
 #   define BN_MP_CLAMP_C
-#   define BN_MP_EXCH_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_INIT_SIZE_C
+#   define BN_S_MP_MUL_HIGH_DIGS_FAST_C
+#endif
+
+#if defined(BN_S_MP_MUL_HIGH_DIGS_FAST_C)
+#   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
+#endif
+
+#if defined(BN_S_MP_PRIME_IS_DIVISIBLE_C)
+#   define BN_MP_MOD_D_C
+#endif
+
+#if defined(BN_S_MP_RAND_JENKINS_C)
+#   define BN_S_MP_RAND_JENKINS_INIT_C
+#endif
+
+#if defined(BN_S_MP_RAND_PLATFORM_C)
+#endif
+
+#if defined(BN_S_MP_REVERSE_C)
 #endif
 
 #if defined(BN_S_MP_SQR_C)
-#   define BN_MP_INIT_SIZE_C
 #   define BN_MP_CLAMP_C
-#   define BN_MP_EXCH_C
 #   define BN_MP_CLEAR_C
+#   define BN_MP_EXCH_C
+#   define BN_MP_INIT_SIZE_C
+#endif
+
+#if defined(BN_S_MP_SQR_FAST_C)
+#   define BN_MP_CLAMP_C
+#   define BN_MP_GROW_C
 #endif
 
 #if defined(BN_S_MP_SUB_C)
+#   define BN_MP_CLAMP_C
 #   define BN_MP_GROW_C
+#endif
+
+#if defined(BN_S_MP_TOOM_MUL_C)
+#   define BN_MP_ADD_C
 #   define BN_MP_CLAMP_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_CLEAR_MULTI_C
+#   define BN_MP_DIV_2_C
+#   define BN_MP_DIV_3_C
+#   define BN_MP_INIT_MULTI_C
+#   define BN_MP_INIT_SIZE_C
+#   define BN_MP_LSHD_C
+#   define BN_MP_MUL_2_C
+#   define BN_MP_MUL_C
+#   define BN_MP_SUB_C
 #endif
 
-#if defined(BNCORE_C)
+#if defined(BN_S_MP_TOOM_SQR_C)
+#   define BN_MP_ADD_C
+#   define BN_MP_CLAMP_C
+#   define BN_MP_CLEAR_C
+#   define BN_MP_DIV_2_C
+#   define BN_MP_INIT_C
+#   define BN_MP_INIT_SIZE_C
+#   define BN_MP_LSHD_C
+#   define BN_MP_MUL_2_C
+#   define BN_MP_MUL_C
+#   define BN_MP_SQR_C
+#   define BN_MP_SUB_C
 #endif
 
+#ifdef LTM_INSIDE
+#undef LTM_INSIDE
 #ifdef LTM3
 #   define LTM_LAST
 #endif
 
-#include <tommath_superclass.h>
-#include <tommath_class.h>
+#include "tommath_superclass.h"
+#include "tommath_class.h"
 #else
 #   define LTM_LAST
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
diff --git a/libtommath/tommath_cutoffs.h b/libtommath/tommath_cutoffs.h
new file mode 100644
index 0000000..a65a9b3
--- /dev/null
+++ b/libtommath/tommath_cutoffs.h
@@ -0,0 +1,13 @@
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+/*
+   Current values evaluated on an AMD A8-6600K (64-bit).
+   Type "make tune" to optimize them for your machine but
+   be aware that it may take a long time. It took 2:30 minutes
+   on the aforementioned machine for example.
+ */
+
+#define MP_DEFAULT_KARATSUBA_MUL_CUTOFF 80
+#define MP_DEFAULT_KARATSUBA_SQR_CUTOFF 120
+#define MP_DEFAULT_TOOM_MUL_CUTOFF      350
+#define MP_DEFAULT_TOOM_SQR_CUTOFF      400
diff --git a/libtommath/tommath_private.h b/libtommath/tommath_private.h
index b3600a0..3867ac6 100644
--- a/libtommath/tommath_private.h
+++ b/libtommath/tommath_private.h
@@ -1,114 +1,299 @@
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
+
 #ifndef TOMMATH_PRIV_H_
 #define TOMMATH_PRIV_H_
 
 #include <tommath.h>
+#include "tommath_class.h"
 
-#ifndef MIN
-#define MIN(x, y) (((x) < (y)) ? (x) : (y))
+/*
+ * Private symbols
+ * ---------------
+ *
+ * On Unix symbols can be marked as hidden if libtommath is compiled
+ * as a shared object. By default, symbols are visible.
+ * As of now, this feature is opt-in via the MP_PRIVATE_SYMBOLS define.
+ *
+ * On Win32 a .def file must be used to specify the exported symbols.
+ */
+#if defined (MP_PRIVATE_SYMBOLS) && defined(__GNUC__) && __GNUC__ >= 4
+#   define MP_PRIVATE __attribute__ ((visibility ("hidden")))
+#else
+#   define MP_PRIVATE
 #endif
 
-#ifndef MAX
-#define MAX(x, y) (((x) > (y)) ? (x) : (y))
+/* Hardening libtommath
+ * --------------------
+ *
+ * By default memory is zeroed before calling
+ * MP_FREE to avoid leaking data. This is good
+ * practice in cryptographical applications.
+ *
+ * Note however that memory allocators used
+ * in cryptographical applications can often
+ * be configured by itself to clear memory,
+ * rendering the clearing in tommath unnecessary.
+ * See for example https://github.com/GrapheneOS/hardened_malloc
+ * and the option CONFIG_ZERO_ON_FREE.
+ *
+ * Furthermore there are applications which
+ * value performance more and want this
+ * feature to be disabled. For such applications
+ * define MP_NO_ZERO_ON_FREE during compilation.
+ */
+#ifdef MP_NO_ZERO_ON_FREE
+#  define MP_FREE_BUFFER(mem, size)   MP_FREE((mem), (size))
+#  define MP_FREE_DIGITS(mem, digits) MP_FREE((mem), sizeof (mp_digit) * (size_t)(digits))
+#else
+#  define MP_FREE_BUFFER(mem, size)                     \
+do {                                                    \
+   size_t fs_ = (size);                                 \
+   void* fm_ = (mem);                                   \
+   if (fm_ != NULL) {                                   \
+      MP_ZERO_BUFFER(fm_, fs_);                         \
+      MP_FREE(fm_, fs_);                                \
+   }                                                    \
+} while (0)
+#  define MP_FREE_DIGITS(mem, digits)                   \
+do {                                                    \
+   int fd_ = (digits);                                  \
+   void* fm_ = (mem);                                   \
+   if (fm_ != NULL) {                                   \
+      size_t fs_ = sizeof (mp_digit) * (size_t)fd_;     \
+      MP_ZERO_BUFFER(fm_, fs_);                         \
+      MP_FREE(fm_, fs_);                                \
+   }                                                    \
+} while (0)
 #endif
 
-#ifdef __cplusplus
-extern "C" {
+#ifdef MP_USE_MEMSET
+#  include <string.h>
+#  define MP_ZERO_BUFFER(mem, size)   memset((mem), 0, (size))
+#  define MP_ZERO_DIGITS(mem, digits)                   \
+do {                                                    \
+   int zd_ = (digits);                                  \
+   if (zd_ > 0) {                                       \
+      memset((mem), 0, sizeof(mp_digit) * (size_t)zd_); \
+   }                                                    \
+} while (0)
+#else
+#  define MP_ZERO_BUFFER(mem, size)                     \
+do {                                                    \
+   size_t zs_ = (size);                                 \
+   char* zm_ = (char*)(mem);                            \
+   while (zs_-- > 0u) {                                 \
+      *zm_++ = '\0';                                    \
+   }                                                    \
+} while (0)
+#  define MP_ZERO_DIGITS(mem, digits)                   \
+do {                                                    \
+   int zd_ = (digits);                                  \
+   mp_digit* zm_ = (mem);                               \
+   while (zd_-- > 0) {                                  \
+      *zm_++ = 0;                                       \
+   }                                                    \
+} while (0)
+#endif
+
+/* Tunable cutoffs
+ * ---------------
+ *
+ *  - In the default settings, a cutoff X can be modified at runtime
+ *    by adjusting the corresponding X_CUTOFF variable.
+ *
+ *  - Tunability of the library can be disabled at compile time
+ *    by defining the MP_FIXED_CUTOFFS macro.
+ *
+ *  - There is an additional file tommath_cutoffs.h, which defines
+ *    the default cutoffs. These can be adjusted manually or by the
+ *    autotuner.
+ *
+ */
+
+#ifdef MP_FIXED_CUTOFFS
+#  include "tommath_cutoffs.h"
+#  define MP_KARATSUBA_MUL_CUTOFF MP_DEFAULT_KARATSUBA_MUL_CUTOFF
+#  define MP_KARATSUBA_SQR_CUTOFF MP_DEFAULT_KARATSUBA_SQR_CUTOFF
+#  define MP_TOOM_MUL_CUTOFF      MP_DEFAULT_TOOM_MUL_CUTOFF
+#  define MP_TOOM_SQR_CUTOFF      MP_DEFAULT_TOOM_SQR_CUTOFF
+#else
+#  define MP_KARATSUBA_MUL_CUTOFF KARATSUBA_MUL_CUTOFF
+#  define MP_KARATSUBA_SQR_CUTOFF KARATSUBA_SQR_CUTOFF
+#  define MP_TOOM_MUL_CUTOFF      TOOM_MUL_CUTOFF
+#  define MP_TOOM_SQR_CUTOFF      TOOM_SQR_CUTOFF
 #endif
 
 /* define heap macros */
-#ifndef XMALLOC
+#ifndef MP_MALLOC
 /* default to libc stuff */
-#   define XMALLOC(size)                   malloc(size)
-#   define XFREE(mem, size)                free(mem)
-#   define XREALLOC(mem, oldsize, newsize) realloc(mem, newsize)
+#   include <stdlib.h>
+#   define MP_MALLOC(size)                   malloc(size)
+#   define MP_REALLOC(mem, oldsize, newsize) realloc((mem), (newsize))
+#   define MP_CALLOC(nmemb, size)            calloc((nmemb), (size))
+#   define MP_FREE(mem, size)                free(mem)
 #elif 0
 /* prototypes for our heap functions */
-extern void *XMALLOC(size_t size);
-extern void *XREALLOC(void *mem, size_t oldsize, size_t newsize);
-extern void XFREE(void *mem, size_t size);
+extern void *MP_MALLOC(size_t size);
+extern void *MP_REALLOC(void *mem, size_t oldsize, size_t newsize);
+extern void *MP_CALLOC(size_t nmemb, size_t size);
+extern void MP_FREE(void *mem, size_t size);
+#endif
+
+/* feature detection macro */
+#ifdef _MSC_VER
+/* Prevent false positive: not enough arguments for function-like macro invocation */
+#pragma warning(disable: 4003)
 #endif
+#define MP_STRINGIZE(x)  MP__STRINGIZE(x)
+#define MP__STRINGIZE(x) ""#x""
+#define MP_HAS(x)        (sizeof(MP_STRINGIZE(BN_##x##_C)) == 1u)
 
-/* you'll have to tune these... */
-#define KARATSUBA_MUL_CUTOFF 80      /* Min. number of digits before Karatsuba multiplication is used. */
-#define KARATSUBA_SQR_CUTOFF 120     /* Min. number of digits before Karatsuba squaring is used. */
-#define TOOM_MUL_CUTOFF      350     /* no optimal values of these are known yet so set em high */
-#define TOOM_SQR_CUTOFF      400
+#define MP_MIN(x, y) (((x) < (y)) ? (x) : (y))
+#define MP_MAX(x, y) (((x) > (y)) ? (x) : (y))
 
-/* size of comba arrays, should be at least 2 * 2**(BITS_PER_WORD - BITS_PER_DIGIT*2) */
-#define MP_WARRAY               (1u << (((sizeof(mp_word) * CHAR_BIT) - (2 * DIGIT_BIT)) + 1))
+/* Static assertion */
+#define MP_STATIC_ASSERT(msg, cond) typedef char mp_static_assert_##msg[(cond) ? 1 : -1];
 
 /* ---> Basic Manipulations <--- */
-#define IS_ZERO(a) ((a)->used == 0)
-#define IS_EVEN(a) (((a)->used == 0) || (((a)->dp[0] & 1u) == 0u))
-#define IS_ODD(a)  (((a)->used > 0) && (((a)->dp[0] & 1u) == 1u))
+#define MP_IS_ZERO(a) ((a)->used == 0)
+#define MP_IS_EVEN(a) (((a)->used == 0) || (((a)->dp[0] & 1u) == 0u))
+#define MP_IS_ODD(a)  (((a)->used > 0) && (((a)->dp[0] & 1u) == 1u))
+
+#define MP_SIZEOF_BITS(type)    ((size_t)CHAR_BIT * sizeof(type))
+#define MP_MAXFAST              (int)(1uL << (MP_SIZEOF_BITS(mp_word) - (2u * (size_t)MP_DIGIT_BIT)))
+
+/* TODO: Remove PRIVATE_MP_PREC as soon as deprecated MP_PREC is removed from tommath.h */
+#ifdef PRIVATE_MP_PREC
+#   undef MP_PREC
+#   define MP_PREC PRIVATE_MP_PREC
+#endif
+
+/* Minimum number of available digits in mp_int, MP_PREC >= MP_MIN_PREC */
+#define MP_MIN_PREC ((((int)MP_SIZEOF_BITS(Tcl_WideInt) + MP_DIGIT_BIT) - 1) / MP_DIGIT_BIT)
+
+MP_STATIC_ASSERT(prec_geq_min_prec, MP_PREC >= MP_MIN_PREC)
+
+/* random number source */
+extern MP_PRIVATE mp_err(*s_mp_rand_source)(void *out, size_t size);
 
 /* lowlevel functions, do not call! */
-int s_mp_add(const mp_int *a, const mp_int *b, mp_int *c);
-int s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c);
-#define s_mp_mul(a, b, c) s_mp_mul_digs(a, b, c, (a)->used + (b)->used + 1)
-int fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs);
-int s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs);
-int fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs);
-int s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs);
-int fast_s_mp_sqr(const mp_int *a, mp_int *b);
-int s_mp_sqr(const mp_int *a, mp_int *b);
-int mp_karatsuba_mul(const mp_int *a, const mp_int *b, mp_int *c);
-int mp_toom_mul(const mp_int *a, const mp_int *b, mp_int *c);
-int mp_karatsuba_sqr(const mp_int *a, mp_int *b);
-int mp_toom_sqr(const mp_int *a, mp_int *b);
-int fast_mp_invmod(const mp_int *a, const mp_int *b, mp_int *c);
-int mp_invmod_slow(const mp_int *a, const mp_int *b, mp_int *c);
-int fast_mp_montgomery_reduce(mp_int *x, const mp_int *n, mp_digit rho);
-int mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode);
-int s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode);
-void bn_reverse(unsigned char *s, int len);
-
-extern const char *const mp_s_rmap;
-extern const unsigned char mp_s_rmap_reverse[];
-extern const size_t mp_s_rmap_reverse_sz;
-
-/* Fancy macro to set an MPI from another type.
- * There are several things assumed:
- *  x is the counter
- *  a is the pointer to the MPI
- *  b is the original value that should be set in the MPI.
- */
-#define MP_SET_XLONG(func_name, type)                    \
-int func_name (mp_int * a, type b)                       \
-{                                                        \
-   int x = 0;                                            \
-   int new_size = (((CHAR_BIT * sizeof(type)) + DIGIT_BIT) - 1) / DIGIT_BIT; \
-   int res = mp_grow(a, new_size);                       \
-   if (res == MP_OKAY) {                                 \
-     mp_zero(a);                                         \
-     while (b != 0u) {                                   \
-        a->dp[x++] = ((mp_digit)b & MP_MASK);            \
-        if ((CHAR_BIT * sizeof (b)) <= DIGIT_BIT) { break; } \
-        b >>= ((CHAR_BIT * sizeof (b)) <= DIGIT_BIT ? 0 : DIGIT_BIT); \
-     }                                                   \
-     a->used = x;                                        \
-   }                                                     \
-   return res;                                           \
-}
-
-#ifdef __cplusplus
-}
+MP_PRIVATE mp_bool s_mp_get_bit(const mp_int *a, unsigned int b);
+MP_PRIVATE mp_err s_mp_add(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+MP_PRIVATE mp_err s_mp_sub(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+MP_PRIVATE mp_err s_mp_mul_digs_fast(const mp_int *a, const mp_int *b, mp_int *c, int digs) MP_WUR;
+MP_PRIVATE mp_err s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs) MP_WUR;
+MP_PRIVATE mp_err s_mp_mul_high_digs_fast(const mp_int *a, const mp_int *b, mp_int *c, int digs) MP_WUR;
+MP_PRIVATE mp_err s_mp_mul_high_digs(const mp_int *a, const mp_int *b, mp_int *c, int digs) MP_WUR;
+MP_PRIVATE mp_err s_mp_sqr_fast(const mp_int *a, mp_int *b) MP_WUR;
+MP_PRIVATE mp_err s_mp_sqr(const mp_int *a, mp_int *b) MP_WUR;
+MP_PRIVATE mp_err s_mp_balance_mul(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+MP_PRIVATE mp_err s_mp_karatsuba_mul(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+MP_PRIVATE mp_err s_mp_toom_mul(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+MP_PRIVATE mp_err s_mp_karatsuba_sqr(const mp_int *a, mp_int *b) MP_WUR;
+MP_PRIVATE mp_err s_mp_toom_sqr(const mp_int *a, mp_int *b) MP_WUR;
+MP_PRIVATE mp_err s_mp_invmod_fast(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+MP_PRIVATE mp_err s_mp_invmod_slow(const mp_int *a, const mp_int *b, mp_int *c) MP_WUR;
+MP_PRIVATE mp_err s_mp_montgomery_reduce_fast(mp_int *x, const mp_int *n, mp_digit rho) MP_WUR;
+MP_PRIVATE mp_err s_mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode) MP_WUR;
+MP_PRIVATE mp_err s_mp_exptmod(const mp_int *G, const mp_int *X, const mp_int *P, mp_int *Y, int redmode) MP_WUR;
+MP_PRIVATE mp_err s_mp_rand_platform(void *p, size_t n) MP_WUR;
+MP_PRIVATE mp_err s_mp_prime_random_ex(mp_int *a, int t, int size, int flags, private_mp_prime_callback cb, void *dat);
+MP_PRIVATE void s_mp_reverse(unsigned char *s, size_t len);
+MP_PRIVATE mp_err s_mp_prime_is_divisible(const mp_int *a, mp_bool *result);
+
+/* TODO: jenkins prng is not thread safe as of now */
+MP_PRIVATE mp_err s_mp_rand_jenkins(void *p, size_t n) MP_WUR;
+#ifndef MP_NO_STDINT
+MP_PRIVATE void s_mp_rand_jenkins_init(uint64_t seed);
 #endif
 
+extern MP_PRIVATE const char *const mp_s_rmap;
+extern MP_PRIVATE const unsigned char mp_s_rmap_reverse[];
+extern MP_PRIVATE const size_t mp_s_rmap_reverse_sz;
+extern MP_PRIVATE const mp_digit *s_mp_prime_tab;
+
+/* deprecated functions */
+#if 0
+MP_DEPRECATED(s_mp_invmod_fast) mp_err fast_mp_invmod(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(s_mp_montgomery_reduce_fast) mp_err fast_mp_montgomery_reduce(mp_int *x, const mp_int *n,
+      mp_digit rho);
+MP_DEPRECATED(s_mp_mul_digs_fast) mp_err fast_s_mp_mul_digs(const mp_int *a, const mp_int *b, mp_int *c,
+      int digs);
+MP_DEPRECATED(s_mp_mul_high_digs_fast) mp_err fast_s_mp_mul_high_digs(const mp_int *a, const mp_int *b,
+      mp_int *c,
+      int digs);
+MP_DEPRECATED(s_mp_sqr_fast) mp_err fast_s_mp_sqr(const mp_int *a, mp_int *b);
+MP_DEPRECATED(s_mp_balance_mul) mp_err mp_balance_mul(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(s_mp_exptmod_fast) mp_err mp_exptmod_fast(const mp_int *G, const mp_int *X, const mp_int *P,
+      mp_int *Y,
+      int redmode);
+MP_DEPRECATED(s_mp_invmod_slow) mp_err mp_invmod_slow(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(s_mp_karatsuba_mul) mp_err mp_karatsuba_mul(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(s_mp_karatsuba_sqr) mp_err mp_karatsuba_sqr(const mp_int *a, mp_int *b);
+MP_DEPRECATED(s_mp_toom_mul) mp_err mp_toom_mul(const mp_int *a, const mp_int *b, mp_int *c);
+MP_DEPRECATED(s_mp_toom_sqr) mp_err mp_toom_sqr(const mp_int *a, mp_int *b);
+MP_DEPRECATED(s_mp_reverse) void bn_reverse(unsigned char *s, int len);
 #endif
 
+#define MP_GET_ENDIANNESS(x) \
+   do{\
+      int16_t n = 0x1;                                          \
+      char *p = (char *)&n;                                     \
+      x = (p[0] == '\x01') ? MP_LITTLE_ENDIAN : MP_BIG_ENDIAN;  \
+   } while (0)
+
+/* code-generating macros */
+#define MP_SET_UNSIGNED(name, type)                                                    \
+    void name(mp_int * a, type b)                                                      \
+    {                                                                                  \
+        int i = 0;                                                                     \
+        while (b != 0u) {                                                              \
+            a->dp[i++] = ((mp_digit)b & MP_MASK);                                      \
+            if (MP_SIZEOF_BITS(type) <= MP_DIGIT_BIT) { break; }                       \
+            b >>= ((MP_SIZEOF_BITS(type) <= MP_DIGIT_BIT) ? 0 : MP_DIGIT_BIT);         \
+        }                                                                              \
+        a->used = i;                                                                   \
+        a->sign = MP_ZPOS;                                                             \
+        MP_ZERO_DIGITS(a->dp + a->used, a->alloc - a->used);                           \
+    }
+
+#define MP_SET_SIGNED(name, uname, type, utype)          \
+    void name(mp_int * a, type b)                        \
+    {                                                    \
+        uname(a, (b < 0) ? -(utype)b : (utype)b);        \
+        if (b < 0) { a->sign = MP_NEG; }                 \
+    }
 
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */
+#define MP_INIT_INT(name , set, type)                    \
+    mp_err name(mp_int * a, type b)                      \
+    {                                                    \
+        mp_err err;                                      \
+        if ((err = mp_init(a)) != MP_OKAY) {             \
+            return err;                                  \
+        }                                                \
+        set(a, b);                                       \
+        return MP_OKAY;                                  \
+    }
+
+#define MP_GET_MAG(name, type)                                                         \
+    type name(const mp_int* a)                                                         \
+    {                                                                                  \
+        unsigned i = MP_MIN((unsigned)a->used, (unsigned)((MP_SIZEOF_BITS(type) + MP_DIGIT_BIT - 1) / MP_DIGIT_BIT)); \
+        type res = 0u;                                                                 \
+        while (i --> 0u) {                                                             \
+            res <<= ((MP_SIZEOF_BITS(type) <= MP_DIGIT_BIT) ? 0 : MP_DIGIT_BIT);       \
+            res |= (type)a->dp[i];                                                     \
+            if (MP_SIZEOF_BITS(type) <= MP_DIGIT_BIT) { break; }                       \
+        }                                                                              \
+        return res;                                                                    \
+    }
+
+#define MP_GET_SIGNED(name, mag, type, utype)                 \
+    type name(const mp_int* a)                                \
+    {                                                         \
+        utype res = mag(a);                                   \
+        return (a->sign == MP_NEG) ? (type)-res : (type)res;  \
+    }
+
+#endif
diff --git a/libtommath/tommath_superclass.h b/libtommath/tommath_superclass.h
index 7b98ed6..d88bce9 100644
--- a/libtommath/tommath_superclass.h
+++ b/libtommath/tommath_superclass.h
@@ -1,22 +1,16 @@
-/* LibTomMath, multiple-precision integer library -- Tom St Denis
- *
- * LibTomMath is a library that provides multiple-precision
- * integer arithmetic as well as number theoretic functionality.
- *
- * The library was designed directly after the MPI library by
- * Michael Fromberger but has been written from scratch with
- * additional optimizations in place.
- *
- * SPDX-License-Identifier: Unlicense
- */
+/* LibTomMath, multiple-precision integer library -- Tom St Denis */
+/* SPDX-License-Identifier: Unlicense */
 
 /* super class file for PK algos */
 
 /* default ... include all MPI */
+#ifndef LTM_NOTHING
 #define LTM_ALL
+#endif
 
 /* RSA only (does not support DH/DSA/ECC) */
 /* #define SC_RSA_1 */
+/* #define SC_RSA_1_WITH_TESTS */
 
 /* For reference.... On an Athlon64 optimizing for speed...
 
@@ -24,65 +18,93 @@
 
 */
 
+#ifdef SC_RSA_1_WITH_TESTS
+#   define BN_MP_ERROR_TO_STRING_C
+#   define BN_MP_FREAD_C
+#   define BN_MP_FWRITE_C
+#   define BN_MP_INCR_C
+#   define BN_MP_ISEVEN_C
+#   define BN_MP_ISODD_C
+#   define BN_MP_NEG_C
+#   define BN_MP_PRIME_FROBENIUS_UNDERWOOD_C
+#   define BN_MP_RADIX_SIZE_C
+#   define BN_MP_RAND_C
+#   define BN_MP_REDUCE_C
+#   define BN_MP_REDUCE_2K_L_C
+#   define BN_MP_FROM_SBIN_C
+#   define BN_MP_ROOT_U32_C
+#   define BN_MP_SET_L_C
+#   define BN_MP_SET_UL_C
+#   define BN_MP_SBIN_SIZE_C
+#   define BN_MP_TO_RADIX_C
+#   define BN_MP_TO_SBIN_C
+#   define BN_S_MP_RAND_JENKINS_C
+#   define BN_S_MP_RAND_PLATFORM_C
+#endif
+
 /* Works for RSA only, mpi.o is 68KiB */
-#ifdef SC_RSA_1
-#   define BN_MP_SHRINK_C
-#   define BN_MP_LCM_C
-#   define BN_MP_PRIME_RANDOM_EX_C
-#   define BN_MP_INVMOD_C
-#   define BN_MP_GCD_C
-#   define BN_MP_MOD_C
-#   define BN_MP_MULMOD_C
+#if defined(SC_RSA_1) || defined (SC_RSA_1_WITH_TESTS)
+#   define BN_CUTOFFS_C
 #   define BN_MP_ADDMOD_C
+#   define BN_MP_CLEAR_MULTI_C
 #   define BN_MP_EXPTMOD_C
-#   define BN_MP_SET_INT_C
+#   define BN_MP_GCD_C
 #   define BN_MP_INIT_MULTI_C
-#   define BN_MP_CLEAR_MULTI_C
-#   define BN_MP_UNSIGNED_BIN_SIZE_C
-#   define BN_MP_TO_UNSIGNED_BIN_C
+#   define BN_MP_INVMOD_C
+#   define BN_MP_LCM_C
+#   define BN_MP_MOD_C
 #   define BN_MP_MOD_D_C
+#   define BN_MP_MULMOD_C
+#   define BN_MP_PRIME_IS_PRIME_C
 #   define BN_MP_PRIME_RABIN_MILLER_TRIALS_C
-#   define BN_REVERSE_C
+#   define BN_MP_PRIME_RAND_C
+#   define BN_MP_RADIX_SMAP_C
+#   define BN_MP_SET_INT_C
+#   define BN_MP_SHRINK_C
+#   define BN_MP_TO_UNSIGNED_BIN_C
+#   define BN_MP_UNSIGNED_BIN_SIZE_C
 #   define BN_PRIME_TAB_C
+#   define BN_S_MP_REVERSE_C
 
 /* other modifiers */
 #   define BN_MP_DIV_SMALL                    /* Slower division, not critical */
 
+
 /* here we are on the last pass so we turn things off.  The functions classes are still there
  * but we remove them specifically from the build.  This also invokes tweaks in functions
  * like removing support for even moduli, etc...
  */
 #   ifdef LTM_LAST
-#      undef BN_MP_TOOM_MUL_C
-#      undef BN_MP_TOOM_SQR_C
-#      undef BN_MP_KARATSUBA_MUL_C
-#      undef BN_MP_KARATSUBA_SQR_C
-#      undef BN_MP_REDUCE_C
-#      undef BN_MP_REDUCE_SETUP_C
 #      undef BN_MP_DR_IS_MODULUS_C
 #      undef BN_MP_DR_SETUP_C
 #      undef BN_MP_DR_REDUCE_C
-#      undef BN_MP_REDUCE_IS_2K_C
+#      undef BN_MP_DIV_3_C
 #      undef BN_MP_REDUCE_2K_SETUP_C
 #      undef BN_MP_REDUCE_2K_C
+#      undef BN_MP_REDUCE_IS_2K_C
+#      undef BN_MP_REDUCE_SETUP_C
+#      undef BN_S_MP_BALANCE_MUL_C
 #      undef BN_S_MP_EXPTMOD_C
-#      undef BN_MP_DIV_3_C
+#      undef BN_S_MP_INVMOD_FAST_C
+#      undef BN_S_MP_KARATSUBA_MUL_C
+#      undef BN_S_MP_KARATSUBA_SQR_C
 #      undef BN_S_MP_MUL_HIGH_DIGS_C
-#      undef BN_FAST_S_MP_MUL_HIGH_DIGS_C
-#      undef BN_FAST_MP_INVMOD_C
+#      undef BN_S_MP_MUL_HIGH_DIGS_FAST_C
+#      undef BN_S_MP_TOOM_MUL_C
+#      undef BN_S_MP_TOOM_SQR_C
+
+#      ifndef SC_RSA_1_WITH_TESTS
+#         undef BN_MP_REDUCE_C
+#      endif
 
 /* To safely undefine these you have to make sure your RSA key won't exceed the Comba threshold
  * which is roughly 255 digits [7140 bits for 32-bit machines, 15300 bits for 64-bit machines]
  * which means roughly speaking you can handle upto 2536-bit RSA keys with these defined without
  * trouble.
  */
+#      undef BN_MP_MONTGOMERY_REDUCE_C
 #      undef BN_S_MP_MUL_DIGS_C
 #      undef BN_S_MP_SQR_C
-#      undef BN_MP_MONTGOMERY_REDUCE_C
 #   endif
 
 #endif
-
-/* ref:         $Format:%D$ */
-/* git commit:  $Format:%H$ */
-/* commit time: $Format:%ai$ */