EccPoint_mult about micro-ecc HOT 5 CLOSED

Make sure that you specified all of the curve parameters correctly. Note that the words should be specified in the code in the opposite order since micro-ecc uses a little-endian convention.

from micro-ecc.

yes, I have convert all params to little-endian convertion. for example

p=fffffffe ffffffff ffffffff ffffffff ffffffff 00000000 ffffffff ffffffff
-->#define Curve_P_6 {0xFFFFFFFF, 0xFFFFFFFF, 0x00000000, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFF, 0xFFFFFFFE}

Could you please give me your EMAIL? I send the source code that I modified to you?

from micro-ecc.

[email protected]

from micro-ecc.

There was a bug in your omega_mult() function. Here is the corrected version:

static void omega_mult(uint32_t *p_result, uint32_t *p_right)
{
    /* Multiply by (2^224 + 2^96 - 2^64 + 1). */
    vli_set(p_result, p_right); /* 1 */
    p_result[3 + NUM_ECC_DIGITS] = vli_add(p_result + 3, p_result + 3, p_right); /* 2^96 + 1 */
    p_result[7 + NUM_ECC_DIGITS] = vli_add(p_result + 7, p_result + 7, p_right); /* 2^224 + 2^96 + 1 */
    if(vli_sub(p_result + 2, p_result + 2, p_right))
    { /* Propagate borrow if necessary. */
        uint i;
        for(i = 2 + NUM_ECC_DIGITS; ; ++i)
        {
            --p_result[i];
            if(p_result[i] != 0xffffffff)
            {
                break;
            }
        }
    }
}

I recommend that you try the "Generalized Mersenne Reduction" method in http://www.isys.uni-klu.ac.at/PDF/2001-0126-MT.pdf (algorithms 5 and 6, on page 356). This should result in code similar to the secpr256r1 version (should be significantly faster).

from micro-ecc.

Yes, my signature passed. Thank you very much for your help, I will try that method later.

from micro-ecc.