ecmult: Use size_t for array indices

[real-or-random]

I don't think the current code is incorrect, but using size_t improves readability because the type makes it clear that we're dealing with array indices.

Also, making the result of the ECMULT_TABLE_SIZE macro (hopefully) a size_t fixes a compiler warning on MSVC, see #1791.

[hebasto]

Concept ACK.

I came to the same idea while working on #1791, but left it for later :)

[hebasto]

ACK 9e7aede, I have reviewed the code and it looks OK.

However, there are still more cases to address:

• size_t i in:

  secp256k1/src/ecmult_impl.h

  Lines 339 to 348 in 2d9137c

  	for (np = 0; np < no; ++np) {
  	if (i < state->ps[np].bits_na_1 && (n = state->ps[np].wnaf_na_1[i])) {
  	secp256k1_ecmult_table_get_ge(&tmpa, state->pre_a + np * ECMULT_TABLE_SIZE(WINDOW_A), n, WINDOW_A);
  	secp256k1_gej_add_ge_var(r, r, &tmpa, NULL);
  	}
  	if (i < state->ps[np].bits_na_lam && (n = state->ps[np].wnaf_na_lam[i])) {
  	secp256k1_ecmult_table_get_ge_lambda(&tmpa, state->pre_a + np * ECMULT_TABLE_SIZE(WINDOW_A), state->aux + np * ECMULT_TABLE_SIZE(WINDOW_A), n, WINDOW_A);
  	secp256k1_gej_add_ge_var(r, r, &tmpa, NULL);
  	}
  	}

• size_t m in:

  secp256k1/src/ecmult_const_impl.h

  Lines 93 to 98 in 2d9137c

  	for (m = 1; m < ECMULT_CONST_TABLE_SIZE; m++) { \
  	/* This loop is used to avoid secret data in array indices. See
  	* the comment in ecmult_gen_impl.h for rationale. */ \
  	secp256k1_fe_cmov(&(r)->x, &(pre)[m].x, m == index); \
  	secp256k1_fe_cmov(&(r)->y, &(pre)[m].y, m == index); \
  	} \

• size_t i in:

  secp256k1/src/ecmult_const_impl.h

  Lines 228 to 230 in 2d9137c

  	for (i = 0; i < ECMULT_CONST_TABLE_SIZE; i++) {
  	secp256k1_ge_mul_lambda(&pre_a_lam[i], &pre_a[i]);
  	}

Also, making the result of the ECMULT_TABLE_SIZE macro (hopefully) a size_t fixes a compiler warning on MSVC, see #1791.

It does.

[real-or-random]

• size_t i in:

  secp256k1/src/ecmult_impl.h

  Lines 339 to 348 in 2d9137c

  	for (np = 0; np < no; ++np) {
  	if (i < state->ps[np].bits_na_1 && (n = state->ps[np].wnaf_na_1[i])) {
  	secp256k1_ecmult_table_get_ge(&tmpa, state->pre_a + np * ECMULT_TABLE_SIZE(WINDOW_A), n, WINDOW_A);
  	secp256k1_gej_add_ge_var(r, r, &tmpa, NULL);
  	}
  	if (i < state->ps[np].bits_na_lam && (n = state->ps[np].wnaf_na_lam[i])) {
  	secp256k1_ecmult_table_get_ge_lambda(&tmpa, state->pre_a + np * ECMULT_TABLE_SIZE(WINDOW_A), state->aux + np * ECMULT_TABLE_SIZE(WINDOW_A), n, WINDOW_A);
  	secp256k1_gej_add_ge_var(r, r, &tmpa, NULL);
  	}
  	}

I think there's no perfect solution here, but fixing this one will open up the entire business of cleaning up types in secp256k1_ecmult_wnaf (see #1795), and I think this should be a separate PR.

The same is true for the stuff in ecmult_gen; I think this could be a separate PR.

[hebasto]

re-ACK 47eb709.

[jonasnick]

ACK 47eb709

[real-or-random]

@theStack Want to review this?

[theStack]

ACK 47eb709