mirror of https://github.com/zcash/orchard.git
Apply suggestions from code review
Co-authored-by: str4d <jack@electriccoin.co> Co-authored-by: Daira Hopwood <daira@jacaranda.org>
This commit is contained in:
parent
c074990bb9
commit
137066e056
|
@ -1,5 +1,7 @@
|
|||
# Complete addition
|
||||
|
||||
To implement complete addition inside the circuit, we need to check the following cases:
|
||||
|
||||
$\begin{array}{rcll}
|
||||
\mathcal{O} &+& \mathcal{O} &= \mathcal{O} ✓\\
|
||||
\mathcal{O} &+& (x_q, y_q) &= (x_q, y_q) ✓\\
|
||||
|
@ -9,7 +11,9 @@ $\begin{array}{rcll}
|
|||
(x_p, y_p) &+& (x_q, y_q) &= (x_p, y_p) \;⸭\; (x_q, y_q), \text{if } x_p \neq x_q ✓
|
||||
\end{array}$
|
||||
|
||||
Suppose that we represent $\mathcal{O}$ as $(0, 0)$. ($0$ is not an $x$-coordinate of a valid point because we would need $y^2 = x^3 + 5$, and $5$ is not square in $\mathbb{F}_q$.)
|
||||
We represent $\mathcal{O}$ as $(0, 0)$.
|
||||
|
||||
> $0$ is not an $x$-coordinate of a valid point because we would need $y^2 = x^3 + 5$, and $5$ is not square in $\mathbb{F}_q$.
|
||||
|
||||
$$
|
||||
\begin{aligned}
|
||||
|
@ -26,27 +30,32 @@ For the doubling case, $\lambda$ has to instead be computed as $\frac{3x^2}{2y}$
|
|||
Witness $\lambda, \alpha, \beta, \gamma, \delta, A, B, C, D$.
|
||||
|
||||
$
|
||||
\begin{array}{rcl|l}
|
||||
&&& Meaning \\\hline
|
||||
\begin{array}{rcl|rcl}
|
||||
\text{Constraint} &&& \text{Meaning} \\\hline
|
||||
A \cdot (1-A) &=& 0 & A \in \mathbb{B} \\
|
||||
B \cdot (1-B) &=& 0 & B \in \mathbb{B} \\
|
||||
C \cdot (1-C) &=& 0 & C \in \mathbb{B} \\
|
||||
D \cdot (1-D) &=& 0 & D \in \mathbb{B} \\
|
||||
(x_q - x_p) \cdot \alpha &=& 1-A & x_q = x_p \implies A \\
|
||||
x_p \cdot \beta &=& 1-B & x_p = 0 \implies B \\
|
||||
B \cdot x_p &=& 0 & B \implies x_p = 0 \\
|
||||
x_q \cdot \gamma &=& 1-C & x_q = 0 \implies C \\
|
||||
C \cdot x_q &=& 0 & C \implies x_q = 0 \\
|
||||
(y_q + y_p) \cdot \delta &=& 1-D & y_q = -y_p \implies D \\
|
||||
(x_q - x_p) \cdot ((x_q - x_p) \cdot \lambda - (y_q - y_p)) &=& 0 & x_q \neq x_p \implies \lambda = \frac{y_q - y_p}{x_q - x_p} \\
|
||||
A \cdot \left(2y_p \cdot \lambda - 3{x_p}^2\right) &=& 0 & A \wedge y_p \neq 0 \implies \lambda = \frac{3{x_p}^2}{2y_p} \\
|
||||
(1-B) \cdot (1-C) \cdot (\lambda^2 - x_p - x_q - x_r) + B \cdot (x_r - x_q) &=& 0 & (¬B \wedge ¬C \implies x_r = \lambda^2 - x_p - x_q) \wedge (B \implies x_r = x_q) \\
|
||||
(1-B) \cdot (1-C) \cdot (\lambda \cdot (x_p - x_r) - y_p - y_r) + B \cdot (y_r - y_q) &=& 0 & (¬B \wedge ¬C \implies y_r = \lambda \cdot (x_p - x_r) - y_p) \wedge (B \implies y_r = y_q) \\
|
||||
C \cdot (x_r - x_p) &=& 0 & C \implies x_r = x_p \\
|
||||
C \cdot (y_r - y_p) &=& 0 & C \implies y_r = y_p \\
|
||||
D \cdot x_r &=& 0 & D \implies x_r = 0 \\
|
||||
D \cdot y_r &=& 0 & D \implies y_r = 0 \\
|
||||
(x_q - x_p) \cdot \alpha &=& 1-A & x_q = x_p &\implies& A \\
|
||||
x_p \cdot \beta &=& 1-B & x_p = 0 &\implies& B \\
|
||||
B \cdot x_p &=& 0 & B &\implies& x_p = 0 \\
|
||||
x_q \cdot \gamma &=& 1-C & x_q = 0 &\implies& C \\
|
||||
C \cdot x_q &=& 0 & C &\implies& x_q = 0 \\
|
||||
(y_q + y_p) \cdot \delta &=& 1-D & y_q = -y_p &\implies& D \\
|
||||
(x_q - x_p) \cdot ((x_q - x_p) \cdot \lambda - (y_q - y_p)) &=& 0 & x_q \neq x_p &\implies& \lambda = \frac{y_q - y_p}{x_q - x_p} \\
|
||||
A \cdot \left(2y_p \cdot \lambda - 3{x_p}^2\right) &=& 0 & A \wedge y_p \neq 0 &\implies& \lambda = \frac{3{x_p}^2}{2y_p} \\
|
||||
\\
|
||||
(1-B) \cdot (1-C) \cdot (\lambda^2 - x_p - x_q - x_r) && & (¬B \wedge ¬C &\implies& x_r = \lambda^2 - x_p - x_q) \\
|
||||
+ B \cdot (x_r - x_q) &=& 0 & \wedge (B &\implies& x_r = x_q) \\
|
||||
\\
|
||||
(1-B) \cdot (1-C) \cdot (\lambda \cdot (x_p - x_r) - y_p - y_r) && & (¬B \wedge ¬C &\implies& y_r = \lambda \cdot (x_p - x_r) - y_p) \\
|
||||
+ B \cdot (y_r - y_q) &=& 0 & \wedge (B &\implies& y_r = y_q) \\
|
||||
\\
|
||||
C \cdot (x_r - x_p) &=& 0 & C &\implies& x_r = x_p \\
|
||||
C \cdot (y_r - y_p) &=& 0 & C &\implies& y_r = y_p \\
|
||||
D \cdot x_r &=& 0 & D &\implies& x_r = 0 \\
|
||||
D \cdot y_r &=& 0 & D &\implies& y_r = 0 \\
|
||||
\end{array}
|
||||
$
|
||||
|
||||
Max degree: 4
|
||||
Max degree: 4
|
||||
|
|
|
@ -12,18 +12,22 @@ In most cases, we multiply the fixed bases by $255-$bit scalars from $\mathbb{F}
|
|||
$$\alpha = k_0 + k_1 \cdot (2^3)^1 + \cdots + k_{84} \cdot (2^3)^{84}, k_i \in [0..2^3).$$
|
||||
|
||||
## Load fixed base
|
||||
Then, we precompute multiples of the fixed base $B$ for each window. This takes the form of a window table: $M[0..84][0..7]$ such that:
|
||||
Then, we precompute multiples of the fixed base $B$ for each window. This takes the form of a window table: $M[0..85)[0..8)$ such that:
|
||||
|
||||
- for the first 84 rows $M[0..83][0..7]$: $M[w][k] = [(k+1) \cdot (2^3)^w]B$
|
||||
- in the last row $M[84][0..7]$: $M[w][k] = [k \cdot (2^3)^w - \sum\limits_{j=0}^{83} (2^3)^j]B$
|
||||
- for the first 84 rows $M[0..83][0..7]$: $$M[w][k] = [(k+1) \cdot (2^3)^w]B$$
|
||||
- in the last row $M[84][0..7]$: $$M[w][k] = [k \cdot (2^3)^w - \sum\limits_{j=0}^{83} (2^3)^j]B$$
|
||||
|
||||
The additional $(k + 1)$ term lets us avoid adding the point at infinity in the case $k = 0$. We offset these accumulated terms by subtracting them in the final window $- \sum\limits_{j=0}^{83} (2^3)^j$.
|
||||
The additional $(k + 1)$ term lets us avoid adding the point at infinity in the case $k = 0$. We offset these accumulated terms by subtracting them in the final window, i.e. we subtract $\sum\limits_{j=0}^{83} (2^3)^j$.
|
||||
|
||||
For each window of fixed-base multiples $M[w] = (M[w][0], \cdots, M[w][7]), w \in [0..84]$:
|
||||
- define a Lagrange interpolation polynomial $\mathcal{L}_x(k)$ that maps $k \in [0..7]$ to the $x$-coordinate of the multiple $M[w][k]$, i.e.
|
||||
- $\mathcal{L}_x(k) = ([(k + 1) \cdot 8^w] B)_x$ for $w \in [0..83]$;
|
||||
- $\mathcal{L}_x(k) = ([k \cdot (8)^w - \sum\limits_{j=0}^{83} (8)^j] B)_x$ for $w = 84$; and
|
||||
- find a value $z_w$ such that $z_w + (M[w][k])_y$ is a square $u^2$ in the field, but the wrong-sign $y$-coordinate $z_w - (M[w][k])_y$ does not produce a square.
|
||||
- Define a Lagrange interpolation polynomial $\mathcal{L}_x(k)$ that maps $k \in [0..7]$ to the $x$-coordinate of the multiple $M[w][k]$, i.e.
|
||||
$$
|
||||
\mathcal{L}_x(k) = \begin{cases}
|
||||
([(k + 1) \cdot 8^w] B)_x &\text{for } w \in [0..83]; \\
|
||||
([k \cdot (8)^w - \sum\limits_{j=0}^{83} (8)^j] B)_x &\text{for } w = 84; \text{ and}
|
||||
\end{cases}
|
||||
$$
|
||||
- Find a value $z_w$ such that $z_w + (M[w][k])_y$ is a square $u^2$ in the field, but the wrong-sign $y$-coordinate $z_w - (M[w][k])_y$ does not produce a square.
|
||||
|
||||
Repeating this for all $85$ windows, we end up with:
|
||||
- an $85 \times 8$ table $\mathcal{L}_x$ storing $8$ coefficients interpolating the $x-$coordinate for each window. Each $x$-coordinate interpolation polynomial will be of the form
|
||||
|
@ -36,7 +40,7 @@ We load these precomputed values into fixed columns whenever we do fixed-base sc
|
|||
## Fixed-base scalar multiplication
|
||||
Given a decomposed scalar $\alpha$ and a fixed base $B$, we compute $[\alpha]B$ as such:
|
||||
|
||||
1. For each $k_w, w \in [0..84], k_w \in [0..7]$ in the scalar decomposition,witness the $x$- and $y$-coordinates $(x_w,y_w) = M[w][k_w].$
|
||||
1. For each $k_w, w \in [0..84], k_w \in [0..7]$ in the scalar decomposition, witness the $x$- and $y$-coordinates $(x_w,y_w) = M[w][k_w].$
|
||||
2. Check that $(x_w, y_w)$ is on the curve: $y_w^2 = x_w^3 + b$.
|
||||
3. Witness $u_w$ such that $y_w + z_w = u_w^2$.
|
||||
4. Use [incomplete addition](./incomplete-add.md) to sum the $M[w][k_w]$'s, resulting in $[\alpha]B$.
|
||||
|
|
|
@ -61,7 +61,8 @@ So to entirely avoid exceptional cases, we would need $2^{n+1} + 2^n - 1 < (q-1)
|
|||
|
||||
The first $i$ for which the algorithm using **only** incomplete addition fails is going to be $252$, since $2^{252+1} + 2^{252} - 1 > (q - 1)/2$. We need $n = 254$ to make the wraparound technique above work.
|
||||
|
||||
> sage: q = 0x40000000000000000000000000000000224698fc0994a8dd8c46eb2100000001
|
||||
```python
|
||||
sage: q = 0x40000000000000000000000000000000224698fc0994a8dd8c46eb2100000001
|
||||
sage: 2^253 + 2^252 - 1 < (q-1)//2
|
||||
False
|
||||
sage: 2^252 + 2^251 - 1 < (q-1)//2
|
||||
|
@ -93,10 +94,15 @@ We witness $x_{A,i}, x_{P,i}, x_{A, i+1},$ and $\lambda_{1, i}, \lambda_{2, i},
|
|||
|
||||
1. $
|
||||
\lambda_{2,i}^2 - (x_{A,i+1} + (\lambda_{1,i}^2 - x_{A,i} - x_{P,i}) + x_{A,i}) = 0,
|
||||
$ and
|
||||
$$
|
||||
|
||||
2. $
|
||||
2 \cdot \lambda_{2,i} \cdot (x_{A,i} - x_{A,i+1}) - \big(
|
||||
(\lambda_{1,i} + \lambda_{2,i}) \cdot (x_{A,i} - (\lambda_{1,i}^2 - x_{A,i} - x_{P,i})) +
|
||||
(\lambda_{1,i+1} + \lambda_{2,i+1}) \cdot (x_{A,i+1} - (\lambda_{1,i+1}^2 - x_{A,i+1} - x_{P,i+1}))\big) = 0.
|
||||
$
|
||||
$$
|
||||
\begin{aligned}
|
||||
2 \cdot &\lambda_{2,i} \cdot (x_{A,i} - x_{A,i+1}) - \big( \\
|
||||
&\begin{aligned}
|
||||
(\lambda_{1,i} + \lambda_{2,i}) &\cdot (x_{A,i} - (\lambda_{1,i}^2 - x_{A,i} - x_{P,i})) + \\
|
||||
(\lambda_{1,i+1} + \lambda_{2,i+1}) &\cdot (x_{A,i+1} - (\lambda_{1,i+1}^2 - x_{A,i+1} - x_{P,i+1})) \\
|
||||
\end{aligned} \\
|
||||
\big) &= 0.
|
||||
\end{aligned}
|
||||
$$
|
||||
|
|
Loading…
Reference in New Issue