Merge bitcoin#709: Remove secret-dependant non-constant time operatio…

…n in ecmult_const.

d567b77 Clarify comments about use of rzr on ge functions and abs function. (Gregory Maxwell)
2241ae6 Remove secret-dependant non-constant time operation in ecmult_const. (Gregory Maxwell)

Pull request description:

  ECMULT_CONST_TABLE_GET_GE was branching on its secret input.

  Also makes secp256k1_gej_double_var implemented as a wrapper
   on secp256k1_gej_double_nonzero instead of the other way
   around.  This wasn't a constant time bug but it was fragile
   and could easily become one in the future if the double_var
   algorithm is changed.

ACKs for top commit:
  real-or-random:
    ACK d567b77 I read the diff carefully and tested the code with ECDH enabled and various settings, also on valgrind
  sipa:
    ACK d567b77

Tree-SHA512: f00a921dcc6cc024cfb3ac1a34c1be619b96f1f17ec0ee0f3ff4ea02035ee288e55469491ed3183e2c4e5560cc068c10aafb657dff95a610706e5b9a8cd13966

src/ecmult_const_impl.h

@@ -15,8 +15,10 @@
 /* This is like `ECMULT_TABLE_GET_GE` but is constant time */
 #define ECMULT_CONST_TABLE_GET_GE(r,pre,n,w) do { \
     int m; \
-    int abs_n = (n) * (((n) > 0) * 2 - 1); \
-    int idx_n = abs_n / 2; \
+    /* Extract the sign-bit for a constant time absolute-value. */ \
+    int mask = (n) >> (sizeof(n) * CHAR_BIT - 1); \
+    int abs_n = ((n) + mask) ^ mask; \
+    int idx_n = abs_n >> 1; \
     secp256k1_fe neg_y; \
     VERIFY_CHECK(((n) & 1) == 1); \
     VERIFY_CHECK((n) >= -((1 << ((w)-1)) - 1)); \
@@ -172,6 +174,7 @@ static void secp256k1_ecmult_const(secp256k1_gej *r, const secp256k1_ge *a, cons
         for (i = 0; i < ECMULT_TABLE_SIZE(WINDOW_A); i++) {
             secp256k1_ge_mul_lambda(&pre_a_lam[i], &pre_a[i]);
         }
+
     }
 #endif
 
@@ -195,7 +198,7 @@ static void secp256k1_ecmult_const(secp256k1_gej *r, const secp256k1_ge *a, cons
         int n;
         int j;
         for (j = 0; j < WINDOW_A - 1; ++j) {
-            secp256k1_gej_double_nonzero(r, r, NULL);
+            secp256k1_gej_double_nonzero(r, r);
         }
 
         n = wnaf_1[i];

src/group.h

@@ -95,22 +95,21 @@ static int secp256k1_gej_is_infinity(const secp256k1_gej *a);
 /** Check whether a group element's y coordinate is a quadratic residue. */
 static int secp256k1_gej_has_quad_y_var(const secp256k1_gej *a);
 
-/** Set r equal to the double of a. If rzr is not-NULL, r->z = a->z * *rzr (where infinity means an implicit z = 0).
- * a may not be zero. Constant time. */
-static void secp256k1_gej_double_nonzero(secp256k1_gej *r, const secp256k1_gej *a, secp256k1_fe *rzr);
+/** Set r equal to the double of a, a cannot be infinity. Constant time. */
+static void secp256k1_gej_double_nonzero(secp256k1_gej *r, const secp256k1_gej *a);
 
-/** Set r equal to the double of a. If rzr is not-NULL, r->z = a->z * *rzr (where infinity means an implicit z = 0). */
+/** Set r equal to the double of a. If rzr is not-NULL this sets *rzr such that r->z == a->z * *rzr (where infinity means an implicit z = 0). */
 static void secp256k1_gej_double_var(secp256k1_gej *r, const secp256k1_gej *a, secp256k1_fe *rzr);
 
-/** Set r equal to the sum of a and b. If rzr is non-NULL, r->z = a->z * *rzr (a cannot be infinity in that case). */
+/** Set r equal to the sum of a and b. If rzr is non-NULL this sets *rzr such that r->z == a->z * *rzr (a cannot be infinity in that case). */
 static void secp256k1_gej_add_var(secp256k1_gej *r, const secp256k1_gej *a, const secp256k1_gej *b, secp256k1_fe *rzr);
 
 /** Set r equal to the sum of a and b (with b given in affine coordinates, and not infinity). */
 static void secp256k1_gej_add_ge(secp256k1_gej *r, const secp256k1_gej *a, const secp256k1_ge *b);
 
 /** Set r equal to the sum of a and b (with b given in affine coordinates). This is more efficient
     than secp256k1_gej_add_var. It is identical to secp256k1_gej_add_ge but without constant-time
-    guarantee, and b is allowed to be infinity. If rzr is non-NULL, r->z = a->z * *rzr (a cannot be infinity in that case). */
+    guarantee, and b is allowed to be infinity. If rzr is non-NULL this sets *rzr such that r->z == a->z * *rzr (a cannot be infinity in that case). */
 static void secp256k1_gej_add_ge_var(secp256k1_gej *r, const secp256k1_gej *a, const secp256k1_ge *b, secp256k1_fe *rzr);
 
 /** Set r equal to the sum of a and b (with the inverse of b's Z coordinate passed as bzinv). */

src/group_impl.h

@@ -303,7 +303,7 @@ static int secp256k1_ge_is_valid_var(const secp256k1_ge *a) {
     return secp256k1_fe_equal_var(&y2, &x3);
 }
 
-static void secp256k1_gej_double_var(secp256k1_gej *r, const secp256k1_gej *a, secp256k1_fe *rzr) {
+static SECP256K1_INLINE void secp256k1_gej_double_nonzero(secp256k1_gej *r, const secp256k1_gej *a) {
     /* Operations: 3 mul, 4 sqr, 0 normalize, 12 mul_int/add/negate.
      *
      * Note that there is an implementation described at
@@ -312,29 +312,9 @@ static void secp256k1_gej_double_var(secp256k1_gej *r, const secp256k1_gej *a, s
      * mainly because it requires more normalizations.
      */
     secp256k1_fe t1,t2,t3,t4;
-    /** For secp256k1, 2Q is infinity if and only if Q is infinity. This is because if 2Q = infinity,
-     *  Q must equal -Q, or that Q.y == -(Q.y), or Q.y is 0. For a point on y^2 = x^3 + 7 to have
-     *  y=0, x^3 must be -7 mod p. However, -7 has no cube root mod p.
-     *
-     *  Having said this, if this function receives a point on a sextic twist, e.g. by
-     *  a fault attack, it is possible for y to be 0. This happens for y^2 = x^3 + 6,
-     *  since -6 does have a cube root mod p. For this point, this function will not set
-     *  the infinity flag even though the point doubles to infinity, and the result
-     *  point will be gibberish (z = 0 but infinity = 0).
-     */
-    r->infinity = a->infinity;
-    if (r->infinity) {
-        if (rzr != NULL) {
-            secp256k1_fe_set_int(rzr, 1);
-        }
-        return;
-    }
 
-    if (rzr != NULL) {
-        *rzr = a->y;
-        secp256k1_fe_normalize_weak(rzr);
-        secp256k1_fe_mul_int(rzr, 2);
-    }
+    VERIFY_CHECK(!secp256k1_gej_is_infinity(a));
+    r->infinity = 0;
 
     secp256k1_fe_mul(&r->z, &a->z, &a->y);
     secp256k1_fe_mul_int(&r->z, 2);       /* Z' = 2*Y*Z (2) */
@@ -358,9 +338,32 @@ static void secp256k1_gej_double_var(secp256k1_gej *r, const secp256k1_gej *a, s
     secp256k1_fe_add(&r->y, &t2);         /* Y' = 36*X^3*Y^2 - 27*X^6 - 8*Y^4 (4) */
 }
 
-static SECP256K1_INLINE void secp256k1_gej_double_nonzero(secp256k1_gej *r, const secp256k1_gej *a, secp256k1_fe *rzr) {
-    VERIFY_CHECK(!secp256k1_gej_is_infinity(a));
-    secp256k1_gej_double_var(r, a, rzr);
+static void secp256k1_gej_double_var(secp256k1_gej *r, const secp256k1_gej *a, secp256k1_fe *rzr) {
+    /** For secp256k1, 2Q is infinity if and only if Q is infinity. This is because if 2Q = infinity,
+     *  Q must equal -Q, or that Q.y == -(Q.y), or Q.y is 0. For a point on y^2 = x^3 + 7 to have
+     *  y=0, x^3 must be -7 mod p. However, -7 has no cube root mod p.
+     *
+     *  Having said this, if this function receives a point on a sextic twist, e.g. by
+     *  a fault attack, it is possible for y to be 0. This happens for y^2 = x^3 + 6,
+     *  since -6 does have a cube root mod p. For this point, this function will not set
+     *  the infinity flag even though the point doubles to infinity, and the result
+     *  point will be gibberish (z = 0 but infinity = 0).
+     */
+    if (a->infinity) {
+        r->infinity = 1;
+        if (rzr != NULL) {
+            secp256k1_fe_set_int(rzr, 1);
+        }
+        return;
+    }
+
+    if (rzr != NULL) {
+        *rzr = a->y;
+        secp256k1_fe_normalize_weak(rzr);
+        secp256k1_fe_mul_int(rzr, 2);
+    }
+
+    secp256k1_gej_double_nonzero(r, a);
 }
 
 static void secp256k1_gej_add_var(secp256k1_gej *r, const secp256k1_gej *a, const secp256k1_gej *b, secp256k1_fe *rzr) {

src/tests_exhaustive.c

@@ -142,7 +142,7 @@ void test_exhaustive_addition(const secp256k1_ge *group, const secp256k1_gej *gr
     for (i = 0; i < order; i++) {
         secp256k1_gej tmp;
         if (i > 0) {
-            secp256k1_gej_double_nonzero(&tmp, &groupj[i], NULL);
+            secp256k1_gej_double_nonzero(&tmp, &groupj[i]);
             ge_equals_gej(&group[(2 * i) % order], &tmp);
         }
         secp256k1_gej_double_var(&tmp, &groupj[i], NULL);

src/util.h

@@ -14,6 +14,7 @@
 #include <stdlib.h>
 #include <stdint.h>
 #include <stdio.h>
+#include <limits.h>
 
 typedef struct {
     void (*fn)(const char *text, void* data);