diff --git a/library/bignum.c b/library/bignum.c
index d3a1b00d52..2421c1a3ec 100644
--- a/library/bignum.c
+++ b/library/bignum.c
@@ -1136,7 +1136,8 @@ int mbedtls_mpi_mul_mpi(mbedtls_mpi *X, const mbedtls_mpi *A, const mbedtls_mpi
     MPI_VALIDATE_RET(A != NULL);
     MPI_VALIDATE_RET(B != NULL);
 
-    mbedtls_mpi_init(&TA); mbedtls_mpi_init(&TB);
+    mbedtls_mpi_init(&TA);
+    mbedtls_mpi_init(&TB);
 
     if (X == A) {
         MBEDTLS_MPI_CHK(mbedtls_mpi_copy(&TA, A)); A = &TA;
@@ -1166,13 +1167,7 @@ int mbedtls_mpi_mul_mpi(mbedtls_mpi *X, const mbedtls_mpi *A, const mbedtls_mpi
     MBEDTLS_MPI_CHK(mbedtls_mpi_grow(X, i + j));
     MBEDTLS_MPI_CHK(mbedtls_mpi_lset(X, 0));
 
-    for (size_t k = 0; k < j; k++) {
-        /* We know that there cannot be any carry-out since we're
-         * iterating from bottom to top. */
-        (void) mbedtls_mpi_core_mla(X->p + k, i + 1,
-                                    A->p, i,
-                                    B->p[k]);
-    }
+    mbedtls_mpi_core_mul(X->p, A->p, i, B->p, j);
 
     /* If the result is 0, we don't shortcut the operation, which reduces
      * but does not eliminate side channels leaking the zero-ness. We do
diff --git a/library/bignum_core.c b/library/bignum_core.c
index 1ec5340b7d..1ba4142c7e 100644
--- a/library/bignum_core.c
+++ b/library/bignum_core.c
@@ -448,13 +448,15 @@ mbedtls_mpi_uint mbedtls_mpi_core_mla(mbedtls_mpi_uint *d, size_t d_len,
     return c;
 }
 
-void MPI_CORE(mul)( mbedtls_mpi_uint *X,
-                    const mbedtls_mpi_uint *A, size_t a,
-                    const mbedtls_mpi_uint *B, size_t b )
+void mbedtls_mpi_core_mul(mbedtls_mpi_uint *X,
+                          const mbedtls_mpi_uint *A, size_t A_limbs,
+                          const mbedtls_mpi_uint *B, size_t B_limbs)
 {
-    memset( X, 0, ( a + b ) * ciL );
-    for( size_t i=0; i < b; i++ )
-        (void) mbedtls_mpi_core_mla( X + i, a + 1, A, a, B[i] );
+    memset(X, 0, (A_limbs + B_limbs) * ciL);
+
+    for (size_t i = 0; i < B_limbs; i++) {
+        (void) mbedtls_mpi_core_mla(X + i, A_limbs + 1, A, A_limbs, B[i]);
+    }
 }
 
 /*
diff --git a/library/bignum_core.h b/library/bignum_core.h
index f66db8fd43..3a111600b8 100644
--- a/library/bignum_core.h
+++ b/library/bignum_core.h
@@ -398,24 +398,22 @@ mbedtls_mpi_uint mbedtls_mpi_core_mla(mbedtls_mpi_uint *X, size_t X_limbs,
                                       const mbedtls_mpi_uint *A, size_t A_limbs,
                                       mbedtls_mpi_uint b);
 
-#define MPI_CORE(func) mbedtls_mpi_core_ ## func ## _minimal
-
 /**
  * \brief Perform a known-size multiplication
  *
- * \param[out] X        The pointer to the (little-endian) array
- *                      representing the product of \p a and \p b.
- *                      This must be of length \p a + \p b.
- * \param[in] A         The pointer to the (little-endian) array
- *                      representing the first factor.
- * \param a             The number of limbs in \p A.
- * \param[in] B         The pointer to the (little-endian) array
- *                      representing the second factor.
- * \param b             The number of limbs in \p B.
+ * \param[out] X     The pointer to the (little-endian) array to receive
+ *                   the product of \p A_limbs and \p B_limbs.
+ *                   This must be of length \p A_limbs + \p B_limbs.
+ * \param[in] A      The pointer to the (little-endian) array
+ *                   representing the first factor.
+ * \param A_limbs    The number of limbs in \p A.
+ * \param[in] B      The pointer to the (little-endian) array
+ *                   representing the second factor.
+ * \param B_limbs    The number of limbs in \p B.
  */
-void MPI_CORE(mul)( mbedtls_mpi_uint *X,
-                    const mbedtls_mpi_uint *A, size_t a,
-                    const mbedtls_mpi_uint *B, size_t b );
+void mbedtls_mpi_core_mul(mbedtls_mpi_uint *X,
+                          const mbedtls_mpi_uint *A, size_t A_limbs,
+                          const mbedtls_mpi_uint *B, size_t B_limbs);
 
 /**
  * \brief Calculate initialisation value for fast Montgomery modular