frost/secp256k1-zkp
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add exhaustive tests for group arithmetic, signing, and ecmult on a small group

Browse Source 
If you compile without ./configure --enable-exhaustive-tests=no,
this will create a binary ./exhaustive_tests which will execute
every function possible on a group of small order obtained by
moving to a twist of our curve and locating a generator of small
order.

Currently defaults to order 13, though by changing some #ifdefs
you can get a couple other ones. (Currently 199, which will take
forever to run, and 14, which won't work because it's composite.)

TODO exhaustive tests for the various modules
This commit is contained in:
Andrew Poelstra
2016-07-07 10:11:30 +00:00
parent 20b8877be1
commit 83836a9547
9 changed files with 360 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

157
src/tests_exhaustive.c
View File

@@ -1,5 +1,5 @@
/**********************************************************************
* Copyright (c) 2015 Andrew Poelstra *
* Copyright (c) 2016 Andrew Poelstra *
* Distributed under the MIT software license, see the accompanying *
* file COPYING or http://www.opensource.org/licenses/mit-license.php.*
**********************************************************************/
@@ -13,8 +13,12 @@
#include <time.h>
#undef USE_ECMULT_STATIC_PRECOMPUTATION
#ifndef EXHAUSTIVE_TEST_ORDER
#define EXHAUSTIVE_TEST_ORDER 199
/* see group_impl.h for allowable values */
#define EXHAUSTIVE_TEST_ORDER 13
#define EXHAUSTIVE_TEST_LAMBDA 9 /* cube root of 1 mod 13 */
#endif
#include "include/secp256k1.h"
@@ -60,6 +64,37 @@ void random_fe(secp256k1_fe *x) {
}
/** END stolen from tests.c */
int secp256k1_nonce_function_smallint(unsigned char *nonce32, const unsigned char *msg32,
const unsigned char *key32, const unsigned char *algo16,
void *data, unsigned int attempt) {
secp256k1_scalar s;
int *idata = data;
(void)msg32;
(void)key32;
(void)algo16;
/* Some nonces cannot be used because they'd cause s and/or r to be zero.
* The signing function has retry logic here that just re-calls the nonce
* function with an increased `attempt`. So if attempt > 0 this means we
* need to change the nonce to avoid an infinite loop. */
if (attempt > 0) {
(*idata)++;
}
secp256k1_scalar_set_int(&s, *idata);
secp256k1_scalar_get_b32(nonce32, &s);
return 1;
}
#ifdef USE_ENDOMORPHISM
void test_exhaustive_endomorphism(const secp256k1_ge *group, int order) {
int i;
for (i = 0; i < order; i++) {
secp256k1_ge res;
secp256k1_ge_mul_lambda(&res, &group[i]);
ge_equals_ge(&group[i * EXHAUSTIVE_TEST_LAMBDA % EXHAUSTIVE_TEST_ORDER], &res);
}
}
#endif
void test_exhaustive_addition(const secp256k1_ge *group, const secp256k1_gej *groupj, int order) {
int i, j;
@@ -120,26 +155,90 @@ void test_exhaustive_addition(const secp256k1_ge *group, const secp256k1_gej *gr
}
}
void test_exhaustive_ecmult(secp256k1_context *ctx, secp256k1_ge *group, secp256k1_gej *groupj, int order) {
int i, j;
const int r_log = secp256k1_rand32() % order; /* TODO be less biased */
for (j = 0; j < order; j++) {
for (i = 0; i < order; i++) {
secp256k1_gej tmp;
secp256k1_scalar na, ng;
secp256k1_scalar_set_int(&na, i);
secp256k1_scalar_set_int(&ng, j);
void test_exhaustive_ecmult(const secp256k1_context *ctx, const secp256k1_ge *group, const secp256k1_gej *groupj, int order) {
int i, j, r_log;
for (r_log = 1; r_log < order; r_log++) {
for (j = 0; j < order; j++) {
for (i = 0; i < order; i++) {
secp256k1_gej tmp;
secp256k1_scalar na, ng;
secp256k1_scalar_set_int(&na, i);
secp256k1_scalar_set_int(&ng, j);
secp256k1_ecmult(&ctx->ecmult_ctx, &tmp, &groupj[r_log], &na, &ng);
ge_equals_gej(&group[(i * r_log + j) % order], &tmp);
secp256k1_ecmult(&ctx->ecmult_ctx, &tmp, &groupj[r_log], &na, &ng);
ge_equals_gej(&group[(i * r_log + j) % order], &tmp);
/* TODO we cannot exhaustively test ecmult_const as it does a scalar
* negation for even numbers, and our code is not designed to handle
* such a small scalar modulus. */
if (i > 0) {
secp256k1_ecmult_const(&tmp, &group[i], &ng);
ge_equals_gej(&group[(i * j) % order], &tmp);
}
}
}
}
}
void r_from_k(secp256k1_scalar *r, const secp256k1_ge *group, int k) {
secp256k1_fe x;
unsigned char x_bin[32];
k %= EXHAUSTIVE_TEST_ORDER;
x = group[k].x;
secp256k1_fe_normalize(&x);
secp256k1_fe_get_b32(x_bin, &x);
secp256k1_scalar_set_b32(r, x_bin, NULL);
}
/* hee hee hee */
int solve_discrete_log(const secp256k1_scalar *x_coord, const secp256k1_ge *group, int order) {
int i;
for (i = 0; i < order; i++) {
secp256k1_scalar check_x;
r_from_k(&check_x, group, i);
if (*x_coord == check_x) {
return i;
}
}
return -1;
}
void test_exhaustive_sign(const secp256k1_context *ctx, const secp256k1_ge *group, int order) {
int i, j, k;
/* Loop */
for (i = 1; i < order; i++) { /* message */
for (j = 1; j < order; j++) { /* key */
for (k = 1; k < order; k++) { /* nonce */
secp256k1_ecdsa_signature sig;
secp256k1_scalar sk, msg, r, s, expected_r;
unsigned char sk32[32], msg32[32];
secp256k1_scalar_set_int(&msg, i);
secp256k1_scalar_set_int(&sk, j);
secp256k1_scalar_get_b32(sk32, &sk);
secp256k1_scalar_get_b32(msg32, &msg);
secp256k1_ecdsa_sign(ctx, &sig, msg32, sk32, secp256k1_nonce_function_smallint, &k);
secp256k1_ecdsa_signature_load(ctx, &r, &s, &sig);
/* Note that we compute expected_r *after* signing -- this is important
* because our nonce-computing function function might change k during
* signing. */
r_from_k(&expected_r, group, k);
CHECK(r == expected_r);
CHECK((k * s) % order == (i + r * j) % order ||
(k * (EXHAUSTIVE_TEST_ORDER - s)) % order == (i + r * j) % order);
}
}
}
/* We would like to verify zero-knowledge here by counting how often every
* possible (s, r) tuple appears, but because the group order is larger
* than the field order, when coercing the x-values to scalar values, some
* appear more often than others, so we are actually not zero-knowledge.
* (This effect also appears in the real code, but the difference is on the
* order of 1/2^128th the field order, so the deviation is not useful to a
* computationally bounded attacker.)
*/
}
int main(void) {
int i;
secp256k1_gej groupj[EXHAUSTIVE_TEST_ORDER];
@@ -151,18 +250,42 @@ int main(void) {
/* TODO set z = 1, then do num_tests runs with random z values */
/* Generate the entire group */
secp256k1_ge_set_infinity(&group[0]);
secp256k1_gej_set_infinity(&groupj[0]);
secp256k1_ge_set_gej(&group[0], &groupj[0]);
for (i = 1; i < EXHAUSTIVE_TEST_ORDER; i++) {
/* Set a different random z-value for each Jacobian point */
secp256k1_fe z;
random_fe(&z);
secp256k1_gej_add_ge(&groupj[i], &groupj[i - 1], &secp256k1_ge_const_g);
secp256k1_ge_set_gej(&group[i], &groupj[i]);
secp256k1_gej_rescale(&groupj[i], &z);
/* Verify against ecmult_gen */
{
secp256k1_scalar scalar_i;
secp256k1_gej generatedj;
secp256k1_ge generated;
secp256k1_scalar_set_int(&scalar_i, i);
secp256k1_ecmult_gen(&ctx->ecmult_gen_ctx, &generatedj, &scalar_i);
secp256k1_ge_set_gej(&generated, &generatedj);
CHECK(group[i].infinity == 0);
CHECK(generated.infinity == 0);
CHECK(secp256k1_fe_equal_var(&generated.x, &group[i].x));
CHECK(secp256k1_fe_equal_var(&generated.y, &group[i].y));
}
}
/* Run the tests */
test_exhaustive_sign(ctx, group, EXHAUSTIVE_TEST_ORDER);
/* cannot exhaustively test verify, since our verify code
* depends on the field order being less than twice the
* group order */
#ifdef USE_ENDOMORPHISM
test_exhaustive_endomorphism(group, EXHAUSTIVE_TEST_ORDER);
#endif
test_exhaustive_addition(group, groupj, EXHAUSTIVE_TEST_ORDER);
test_exhaustive_ecmult(ctx, group, groupj, EXHAUSTIVE_TEST_ORDER);