Implement High-Quality Random Number Generation Using AES-CTR Mode with OpenSSL and AES-NI Support

[Knogle]

Implement Cryptographically Secure PRNG Using AES-256-CTR with OpenSSL and AES-NI Support

This commit introduces a new cryptographically secure pseudorandom number generator (CSPRNG) for Nwipe, utilizing the AES-256 algorithm in Counter mode (CTR). By integrating with OpenSSL and leveraging AES-NI hardware acceleration when available, this implementation provides high-quality random numbers essential for cryptographic operations and secure data wiping.

Algorithms have been added in order to ensure high PRNG quality in case of some issue with the OpenSSL library.
4K of random data will be generated during init with the actual seed, and checked for randomness and entropy.

Furthermore i suggest closing #604 in favor of this one, as there is no benefit if we implement AES.

int nwipe_aes_ctr_prng_init(NWIPE_PRNG_INIT_SIGNATURE)
{
    /* rest of the code here */

    // Validate the PRNG output to ensure it is functioning correctly.
    if (aes_ctr_prng_validate((aes_ctr_state_t*)*state) != 0)
    {
        nwipe_log(NWIPE_LOG_FATAL, "AES CTR PRNG validation failed.");
        aes_ctr_prng_general_cleanup((aes_ctr_state_t*)*state);
        free(*state);
        *state = NULL;
        return -1;
    }

    /* rest of the code here */
}

Actual validation.

/* Validates the pseudorandom data generation by generating test data and performing statistical tests.
   This function generates test data and performs:
   - Bit Frequency Test
   - Byte Frequency Test
   - Entropy Calculation
   - Checks for any obvious patterns
   - Thresholds are set to detect non-random behavior.
   - state: Pointer to the initialized AES CTR PRNG state.
   Returns 0 on success, -1 on failure. */
int aes_ctr_prng_validate(aes_ctr_state_t* state)
{
    assert(state != NULL);

    const size_t test_data_size = 4096;  // 4KB of data
    unsigned char* test_buffer = malloc(test_data_size);
    if (!test_buffer)
    {
        nwipe_log(NWIPE_LOG_ERROR, "Validation failed: Unable to allocate memory for test buffer.");
        return -1;
    }
    memset(test_buffer, 0, test_data_size);  // Zero out the buffer
    int outlen = 0;  // Length of data produced by encryption
    size_t total_generated = 0;

    // Generate 4KB of pseudorandom data
    while (total_generated < test_data_size)
    {
        size_t chunk_size = (test_data_size - total_generated);
        if (chunk_size > 512)
            chunk_size = 512;  // Generate data in chunks to avoid large allocations

        if (EVP_EncryptUpdate(state->ctx, test_buffer + total_generated, &outlen,
                              test_buffer + total_generated, chunk_size) != 1)
        {
            nwipe_log(NWIPE_LOG_ERROR,
                      "Failed to generate pseudorandom numbers during validation, error code: %s.",
                      ERR_error_string(ERR_get_error(), NULL));  // Log generation failure
            free(test_buffer);
            return -1;  // Handle error
        }
        total_generated += outlen;
    }

    // Proceed with statistical tests (unchanged from previous implementation)

    // ... [Rest of the validation code] ...

    free(test_buffer);
    return 0;  // Validation successful
}

Key Features:

• AES-256-CTR Mode for strong security and unpredictability.
• Integration with OpenSSL for robust cryptographic operations.
• Automatic detection and utilization of AES-NI instructions for performance enhancement.
• Dynamic linking of OpenSSL to benefit from security updates without recompilation.
• Comprehensive validation step generating 4 KB of data and performing statistical tests:
  • Bit frequency test (±2% deviation allowed).
  • Shannon entropy calculation (minimum 7.5 bits per byte).
  • Pattern detection to identify anomalies.

Implementation Details:

• Uses SHA-256 to derive a 256-bit key from the seed.
• Initializes AES-256-CTR context via OpenSSL's EVP interfaces.
• Generates pseudorandom data by encrypting sequential counter values.
• Includes thorough error handling and resource management.

Statistical Test Results:

• All tests in the NIST Statistical Test Suite passed.
• TestU01 SmallCrush battery tests passed.
• • BigCrush battery tests passed.

This enhancement improves Nwipe's ability to meet stringent security requirements and provides a robust foundation for secure data wiping operations.
For further reference check #559