Skip to content

Script generated finite field arithmetic for elliptic curve cryptography

License

Notifications You must be signed in to change notification settings

mcarrickscott/modarith

Repository files navigation

Modarith - python scripts for generation of code for finite field arithmetic

Script generated finite field arithmetic for "fast enough" elliptic curve cryptography

This repository contains Python 3 scripts for the automatic generation of efficent code for multi-precision modular arithmetic, for 16, 32 and 64-bit architectures, in C and in Rust.

The code uses a multi-limb unsaturated-radix representation for big numbers.

For both languages there are two scripts, one specialised for pseudo-Mersenne moduli of the form $2^m-c$ (pseudo.py and pseudo_rust.py), and one that generates efficient and tailored code for other prime shapes, in particular generalised Mersennes like those required for the standardised NIST256, NIST384, and X448 curves, and Montgomery friendly moduli, as sometimes recommended for use with isogenies (monty.py and monty_rust.py).

The code generated includes functions for modular addition, subtraction, multiplication, inversion, quadratic residuosity and square roots. In other words all of the requirements to implement field arithmetic in the context of elliptic curve cryptography.

The code produced by these scripts was used to create the TLSECC library of C and Rust functions which implement all of the elliptic curve cryptography required by the TLS1.3 protocol. See https://github.com/mcarrickscott/TLSECC

Dependencies

As dependencies it is required that an appropriate C compiler (gcc/clang/icx) and/or the Rust compiler (rustc) is included in the path. For the C code also ensure that the clang-format and cppcheck utilities are installed.

Also in the path must be the very useful utility addchain

https://github.com/mmcloughlin/addchain

For accurate timings across a range of architectures for the C code, install Dan Bernstein's libcpucycles utility from https://cpucycles.cr.yp.to/ . It may be necessary to run ldconfig.

Quick start

For a quick start copy the files from here into a working directory, and try

python3 pseudo.py 64 2**255-19
./time

Then 64-bit code for the suggested modulus is generated and tested. An executable that times important functions will be created if the platform allows it. The standalone C timing code is output to time.c, and code for production use is output to field.c

For Rust

python3 pseudo_rust.py 64 2**255-19
./time

In this case the output is directed to files time.rs and field.rs

For more details read the comments in the provided scripts.

As a Proof of Concept, elliptic curve code for RFC7748 is provided in the files rfc7748.c and rfc7748.rs

Read comments in these files for simple build instructions.

(RFC7748 describes an implementation of Diffie-Hellman key exchange on the Montgomery elliptic curves X25519 and X448. You can create your own Montgomery curve using the sagemath script provided in the file bowe.sage)

Modular arithmetic API

Assume the modulus is $p$. The provided functions are

nres() -- Convert a big number to internal format
redc() -- Convert back from internal format, result $\lt p$
modfsb() -- Perform final subtraction to reduce from $\lt 2p$ to $\lt p$
modadd() -- Modular addition, result $\lt 2p$
modsub() -- Modular subtraction, result $\lt 2p$
modneg() -- Modular negation, result $\lt 2p$
modmul() -- Modular multiplication, result $\lt 2p$
modsqr() -- Modular squaring, result $\lt 2p$
modmli() -- Modular multiplication by a small integer, result $\lt 2p$
modcpy() -- Copy a big number
modpro() -- Calculate progenitor, for subsequent use for modular inverses and square roots
modinv() -- Modular inversion
modsqrt() -- Modular square root
modis1() -- Test for equal to unity
modis0() -- Test for equal to zero
modone() -- Set equal to unity
modzer() -- Set equal to zero
modint() -- Convert an integer to internal format
modqr() -- Test for quadratic residue
modcmv() -- Conditional constant time move
modcsw() -- Conditional constant time swap
modshl() -- shift left by bits
modshr() -- shift right by bits
mod2r() -- set to 2^r
modexp() -- export from internal format to byte array
modimp() -- import to internal format from byte array
modsign() -- Extract sign (parity bit)
modcmp() -- Test for equality \

Using the scripts

The scripts can be used out-of-the-box using simple command-line arguments, as in the examples above.

The scripts can also be tailored in various ways at the top of the script, and in the ``user editable area''.

There are default settings for the choice of compiler, choice of using a clock cycle counter, encouragement for inlining certain functions, code formatting, the use of Karatsuba for (maybe) faster multiplication, and, for the C code, an option to ``decorate'' function names to avoid name clashes.

New named moduli can also be provided in the user editable area, and some settings (like radix choice) adapted individually.

Function name decoration may be required to avoid name clashes in C. If using C++ namespaces can be used to avoid this necessity. It is not an issue for Rust.

Constant time

All generated functions are written with the expectation that they will execute in constant time. But high level code is nevertheless at the mercy of both the compiler and the architecture.

It is strongly recommended that the generated assembly language be closely studied to ensure that there are no compiler introduced timing leaks. In particular code generated for the functions modcmv and modcsw should be checked, bearing in mind that they may be inlined by the compiler. If necessary compiler-specific measures should be taken to prevent inlining, and/or place these functions into a separately compiled module.

New - automatic generation of Edwards or Weierstrass curve API code in C

  1. Copy all code from this directory to a working directory.
  2. Run the python script curve.py, selecting curve name and wordlength (32 or 64 bit). The curve will be in either Edwards or Weierstrass form. Edit this script to add your own curve
  3. The script generates 4 files field.c group.c point.h and curve.c
  4. The script automatically injects this code into edwards.c or weierstrass.c and curve.h
  5. Compile and link testcurve.c with edwards.c or weierstrass.c
  6. Run testcurve to test the arithmetic and perform some timings.

The API interface is as indicated in curve.h. The API is completely implemented in edwards.c or weierstrass.c

Make sure to copy fresh copies of edwards.c weierstrass.c and curve.h from source after each test

Support for standard SHA2 and SHA3 hashing algorithms is provided in hash.c and hash.h

Quickstart 1:-

python curve.py 64 ED25519
gcc -O2 testcurve.c edwards.c -lcpucycles -o testcurve
./testcurve

Note that this intermediate API only provides the elliptic curve functionality. A higher level algorithm API (like that provided for Ed448 signature) would use this API while itself providing additional algorithm specific random number and hashing functionality. It may also use the monty.py script to generate code to perform arithmetic modulo the prime group order, if so required by the algorithm.

Quickstart 2:-

python curve.py 64 ED448

Drop group.c into Ed448.c (EdDSA using ED448) where indicated

gcc -O2 Ed448.c edwards.c hash.c -o Ed448
./Ed448

Quickstart 3:-

python curve.py 64 NIST256

Drop group.c into EC256.c (ECDSA using P-256) where indicated

gcc -O2 EC256.c weierstrass.c hash.c -o EC256
./EC256

Rust version

Copy all code from this directory to a working directory, and move to that directory. Create a standard rust project, named "ecc".

cargo new ecc

Replace the default Cargo.toml file with the one provided.

Copy testcurve.rs and hash.rs into the rust project src subdirectory. Delete the default main.rs

In the working directory

python curve_rust.py 64 ED25519

Copy edwards.rs into the rust project src subdirectory.

cd ecc
cargo run --release --bin testcurve

Make sure to copy fresh copies of edwards.rs and weierstrass.rs from source to the working directory after each test.

python curve_rust.py 64 NIST256

Drop group.rs into EC256.rs (ECDSA using P-256) where indicated. Copy EC256.rs and weierstrass.rs to the rust src subdirectory

cd ecc
cargo run --release --bin EC256

Make sure to copy fresh copies of edwards.rs and weierstrass.rs from source to the working directory after each test.

python curve_rust.py 64 ED448

Drop group.rs into Ed448.rs (EdDSA using ED448) where indicated. Copy Ed448.rs and edwards.rs to the rust src subdirectory

cd ecc
cargo run --release --bin Ed448

A lazy reduction optimization

For the Ed448.c example, modular additions and subtractions can be speeded up by first editing monty.py amd monty_rust.py and setting generic=False.

This activates a curve-specific optimization described in https://eprint.iacr.org/2017/437

Create your own ECC scheme

Here we describe the steps involved to create your own ECC based signature scheme in C. The process for Rust is very similar.

The file signature.c contains templates for the implementation of an API for a digital signature scheme. For working examples see Ed448.c and EC256.c

Start with a clean download of all of the files in this directory to a working directory.

Next ensure that the elliptic curve modulus is named (upper case) and supported in either pseudo.py or monty.py. Ensure the elliptic curve group is named (lower case) in monty.py.

Make sure the various elliptic curve constants are given in curve.py under the upper case curve name. Then for a 64-bit build

python curve.py 64 <Upper Case Name>

This will automatically modify or create the files curve.h, group.c and either edwards.c or weierstrass.c.

Now edit the signature.c file. First drop in the code from group.c where indicated. Then access the document that outlines the details of the signature scheme, and complete the API implementation. For working examples refer to Ed448.c and EC256.c

Provide a main program in signature.c which implements some test vectors and compile as

gcc -O2 signature.c <weierstrass.c or edwards.c> hash.c -o signature

Testing

Any implementation should be tested using test vectors. A good source are the test vectors provided by the Wycheproof project - see https://github.com/C2SP/wycheproof

A short python script parse.py is provided which converts the JSON formatted test vectors for ECDSA and EdDSA to a form more easily digestible by a C or Rust program.

About

Script generated finite field arithmetic for elliptic curve cryptography

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published