neuray API Programmer's Manual

ilibrary_authentication.h File Reference

Description

API component for library authentication functionality.

Code Example

ilibrary_authentication.h

‎/***************************************************************************************************
 * Copyright 2024 NVIDIA Corporation. All rights reserved.
 **************************************************************************************************/

#ifndef MI_NEURAYLIB_ILIBRARY_AUTHENTICATION_H
#define MI_NEURAYLIB_ILIBRARY_AUTHENTICATION_H

#include <mi/base/handle.h>
#include <mi/base/interface_declare.h>
#include <mi/neuraylib/ineuray.h>

#ifdef MI_PLATFORM_WINDOWS
#include <mi/base/miwindows.h>
#else
#include <sys/time.h>
#include <cstdlib>
#include <ctime>
#endif

#include <cstring>

namespace mi {

class IString;

namespace neuraylib {








class ILibrary_authenticator : public
    mi::base::Interface_declare<0x5a7d010a,0x2a65,0x43da,0x92,0xf2,0xcd,0xd9,0xc8,0x4b,0x10,0xd2>
{
public:
    inline static Sint32 
               authenticate(
        const INeuray* library,
        const char* vendor_key, Size vendor_key_length,
        const char* secret_key, Size secret_key_length,
        Sint32 count=1);

    //  Returns a challenge from the library.
    //
    //  Asks the library for a challenge. The authentication process combines the challenge with the
    //  secret key and the vendor key to generate the correct response.
    //
    //  \note This method is internal and should not be called directly. Use the convenience method
    //        #authenticate() instead.
    //
    //  \param buffer          The library will store the challenge here.
    //  \param buffer_length   The actual size of the buffer. If the buffer is not big enough
    //                         generating the challenge generation will fail. Currently, the buffer
    //                         should be able to hold at least 32 bytes.
    //  \return                The necessary size of the needed buffer. The application needs to
    //                         check this value to determine whether its supplied buffer was big
    //                         enough, otherwise the challenge is not valid.
    virtual Size get_challenge( char* buffer, Size buffer_length) = 0;

    //  Submits the calculated response to a challenge.
    //
    //  In addition, the application has to provide a vendor key and a random salt. This is used to
    //  make attacks more difficult.
    //
    //  \note This method is internal and should not be called directly. Use the convenience method
    //        #authenticate() instead.
    //
    //  \param response          The calculated response.
    //  \param response_length   The size of the response.
    //  \param vendor_key        The vendor key assigned to the application writer.
    //  \param vendor_key_length The size of the vendor key.
    //  \param salt              A random salt. Should be a random array of bytes.
    //  \param salt_length       The size of the salt.
    //  \param count             The number of licenses to retrieve.
    virtual void submit_challenge_response(
        const char* response, Size response_length,
        const char* vendor_key, Size vendor_key_length,
        const char* salt, Size salt_length,
        Sint32 count) = 0;

    virtual bool is_trial_license() const = 0;

    virtual Uint64 
               get_trial_seconds_left() const = 0;

    virtual const IString* get_host_id() const = 0;

    virtual const IString* get_last_error_message() const = 0;
};
 // end group mi_neuray_configuration

namespace detail {

// Generates a nonce.
//
// \param[out] buffer       The buffer for the nonce. Needs to be able to hold 32 bytes.
static void generate_nonce( char* buffer);

// Generates the response for a challenge, salt, and secret key.
//
// \param salt              The salt (32 bytes).
// \param challenge         The challenge (32 bytes).
// \param secret_key        The secret key used to calculate the response.
// \param secret_key_length The size of the secret key.
// \param[out] response     The buffer for the response. Needs to be able to hold 32 bytes.
static void calculate_response(
    const char* salt, const char* challenge,
    const char* secret_key, Size secret_key_length, char* response);

// Computes a SHA256 hash value.
//
// \param input           The input for which to compute the SHA256 hash value.
// \param input_length    The size of the input.
// \param[out] buffer     The buffer for the SHA256 hash value. Needs to be able to hold 32 bytes.
static void sha256( const char* input, unsigned int input_length, char* buffer);

} // namespace detail

inline Sint32 
               ILibrary_authenticator::authenticate(
    const INeuray* library,
    const char* vendor_key, Size vendor_key_length,
    const char* secret_key, Size secret_key_length,
    Sint32 count)
{
    if( !library)
        return -1;

    base::Handle< ILibrary_authenticator> authenticator(
        library->get_api_component<ILibrary_authenticator>());
    if( !authenticator.is_valid_interface())
        return -2;

    char challenge[32];
    memset( &challenge[0], 0, 32);
    if( authenticator->get_challenge( challenge, sizeof( challenge)) > sizeof( challenge))
        return -3;

    char salt[32];
    detail::generate_nonce( salt);

    char response[32];
    detail::calculate_response( salt, challenge, secret_key, secret_key_length, response);

    authenticator->submit_challenge_response(
        response, sizeof( response), vendor_key, vendor_key_length, salt, sizeof( salt), count);
    return 0;
}

namespace detail {

void calculate_response(
    const char* salt, const char* challenge,
    const char* secret_key, Size secret_key_length, char* response)
{
    if( secret_key_length > 1024u * 10u)
        return;
    // salt = 32 bytes, challenge = 32 bytes
    Size total_len = 32u + 32u + secret_key_length;
    char* buffer = new char[total_len];
    memcpy( buffer, salt, 32);
    memcpy( buffer + 32, challenge, 32);
    memcpy( buffer + 64, secret_key, secret_key_length);
    sha256( buffer, static_cast<Uint32>( total_len), response);
    delete[] buffer;
}

void generate_nonce( char* buffer)
{
    if( buffer == 0)
        return;
    Uint64 number = 0;
#ifdef MI_PLATFORM_WINDOWS
    srand( GetTickCount());
    LARGE_INTEGER tmp;
    QueryPerformanceCounter( &tmp);
    number += tmp.QuadPart + GetCurrentProcessId() + GetTickCount();
#else
    // Note: icc 13.1 report warning here as an implicit conversion,
    // but this is an explicit conversion.
    srand( static_cast<unsigned>( time(  0)));
    struct timeval tv;
    gettimeofday( &tv, 0);
    number += static_cast<Uint64>( tv.tv_sec + tv.tv_usec);
#endif
    char buf[sizeof( Uint32) + 3*sizeof( Uint64)] = {0};
    int r = rand();
    memcpy( buf, &r, sizeof( Uint32));
    memcpy( buf + sizeof( Uint32), &number, sizeof( Uint64));
    memcpy( buf + sizeof( Uint32) + sizeof( Uint64), &number, sizeof( Uint64));
    number += static_cast<Uint64>( rand());
    memcpy( buf + sizeof( Uint32) + 2*sizeof( Uint64), &number, sizeof( Uint64));
    sha256( buf, static_cast<Uint32>( sizeof( buf)), buffer);
}

// Table of round constants.
// First 32 bits of the fractional parts of the cube roots of the first 64 primes 2..311
static const Uint32 sha256_constants[] = {
    0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5, 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5,
    0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3, 0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174,
    0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc, 0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da,
    0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7, 0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967,
    0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13, 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85,
    0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3, 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070,
    0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5, 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3,
    0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208, 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
};

// Reverses the byte order of the bytes of \p x
static Uint32 flip32( Uint32 x)
{
    return ((((x) & 0xff000000) >> 24) | (((x) & 0x00ff0000) >>  8) |
            (((x) & 0x0000ff00) <<  8) | (((x) & 0x000000ff) << 24));
}

// Rotates the bits of \p x to the right by \p y bits
static Uint32 rightrotate( Uint32 x, Uint32 y)
{
    return (( x >> y) | (x << (32-y)));
}

void sha256( const char* input, unsigned int input_length, char* buffer)
{
    if( (input_length <= 0) || (input == 0) || (buffer == 0))
       return;

    // First 32 bits of the fractional parts of the square roots of the first 8 primes 2..19
    Uint32 state[] = {0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a,
                      0x510e527f, 0x9b05688c, 0x1f83d9ab, 0x5be0cd19};

    // k is the number of '0' bits >= 0 such that the resulting message length is 448 (mod 512)
    unsigned int r = (input_length * 8 + 1) % 512;
    unsigned int k = r > 448 ? 960 - r : 448 - r;

    unsigned int pos = 0;
    for( unsigned int chunk = 0; k != 0; ++chunk) {

        Uint32 W[64] = {0};
        Uint8* ptr = reinterpret_cast<Uint8*>( W);
        unsigned int to_copy = input_length - pos;
        to_copy = to_copy > 64 ? 64 : to_copy;
        if( to_copy > 0) {
            memcpy( ptr, input + pos, to_copy);
            pos += to_copy;
        }

        // If we are at the end of input message
        if( pos == input_length) {
            // If we still have not padded and have space to add a 1, add it
            if( pos/64 == chunk)
                ptr[pos%64] |= static_cast<Uint8>( 0x80);
            // If we can pad and still have space to add the length, add it
            if( (pos*8 + 1 + k) - (chunk*512) <= 448) {
                Uint64 value = input_length * 8;
                ptr = reinterpret_cast<Uint8*>(&W[14]);
                // Note: icc 13.1 report warning for the following
                // code as an implicit conversion, but this is an
                // explicit conversion.
                ptr[0] = static_cast<Uint8>((value >> 56) & 0xff);
                ptr[1] = static_cast<Uint8>((value >> 48) & 0xff);
                ptr[2] = static_cast<Uint8>((value >> 40) & 0xff);
                ptr[3] = static_cast<Uint8>((value >> 32) & 0xff);
                ptr[4] = static_cast<Uint8>((value >> 24) & 0xff);
                ptr[5] = static_cast<Uint8>((value >> 16) & 0xff);
                ptr[6] = static_cast<Uint8>((value >>  8) & 0xff);
                ptr[7] = static_cast<Uint8>( value        & 0xff);
                k = 0;
            }
        }

        // Flip to big endian
        for( int i = 0; i < 16; ++i)
            W[i] = flip32( W[i]);

        // Extend the sixteen 32-bit words into 64 32-bit words
        for( Uint32 i = 16; i < 64; ++i) {
            Uint32 s0 = rightrotate( W[i-15], 7) ^ rightrotate( W[i-15], 18) ^ (W[i-15] >> 3);
            Uint32 s1 = rightrotate( W[i-2], 17) ^ rightrotate( W[i- 2], 19) ^ (W[i-2] >> 10);
            W[i] = W[i-16] + s0 + W[i-7] + s1;
        }

        // Initialize hash value for this chunk
        Uint32 a = state[0];
        Uint32 b = state[1];
        Uint32 c = state[2];
        Uint32 d = state[3];
        Uint32 e = state[4];
        Uint32 f = state[5];
        Uint32 g = state[6];
        Uint32 h = state[7];

        for( Uint32 j = 0; j < 64; ++j) {
            Uint32 s0  = rightrotate( a, 2) ^ rightrotate( a, 13) ^ rightrotate( a, 22);
            Uint32 maj = (a & b) ^ (a & c) ^ (b & c);
            Uint32 t2  = s0 + maj;
            Uint32 s1  = rightrotate( e, 6) ^ rightrotate( e, 11) ^ rightrotate( e, 25);
            Uint32 ch  = (e & f) ^ ((~e) & g);
            Uint32 t1  = h + s1 + ch + sha256_constants[j] + W[j];

            h = g; g = f; f = e; e = d + t1;
            d = c; c = b; b = a; a = t1 + t2;
        }

        // Add this chunk's hash value to result so far
        state[0] += a; state[1] += b; state[2] += c; state[3] += d;
        state[4] += e; state[5] += f; state[6] += g; state[7] += h;
    }

    // Flip to little endian
    for( int i = 0; i < 8; ++i)
        state[i] = flip32( state[i]);

    memcpy( buffer, reinterpret_cast<char*>( state), 32);
}

} // namespace detail

} // namespace neuraylib

} // namespace mi

#endif // MI_NEURAYLIB_ILIBRARY_AUTHENTICATION_H

Namespaces

namespace 
Common namespace for APIs of NVIDIA Advanced Rendering Center GmbH. More...
namespace 
Namespace for the neuray API. More...
namespace mi::neuraylib::detail

Classes

class 
This interface is used for authenticating the application against the library. More...