SM4 国密算法原理与 C++跨平台实现详解

SM4 国密算法原理与 C++跨平台实现详解 | 极客日志

算法	密钥长度	分组长度	性能	安全性	优缺点	典型应用场景
SM4	128 位	128 位	软件实现中等效率，硬件优化后性能提升；32 轮迭代复杂度较高	抗差分/线性攻击，符合中国国密标准，国际认可（ISO）	优点：合规性强，设计简洁，对称加解密流程；缺点：国际生态支持有限	中国政务/金融系统、物联网设备、合规性要求场景
AES	128/192/256 位	128 位	硬件加速（如 AES-NI 指令集）下极快；软件实现效率高，轮数少（10/12/14 轮）	抗已知攻击，NIST 认证，行业黄金标准，密钥长度灵活	优点：广泛支持、高效灵活；缺点：依赖硬件加速优化	通用加密（TLS/SSL、磁盘加密、通信协议）
DES	56 位（实际 64 位）	64 位	速度慢（密钥短，16 轮迭代），已过时	已被破解（暴力攻击可行），不安全	优点：历史兼容性；缺点：完全淘汰，仅用于遗留系统	老旧系统兼容（不推荐新项目使用）
3DES	168 位（有效 112 位）	64 位	比 DES 慢 3 倍（3 次 DES 运算）	安全性高于 DES 但低于 AES，NIST 已禁用（2023 年后）	优点：过渡期兼容；缺点：效率低，安全性不足	银行旧系统、传统设备
ChaCha20	256 位	流密码	纯软件实现快（无硬件依赖），适合移动端/低功耗设备	抗侧信道攻击，安全性等同于 AES-256，被 TLS 1.3 广泛采用	优点：移动端高效，抗时序攻击；缺点：分组模式需结合 Poly1305 认证	移动应用、TLS 协议、实时通信

// 示例代码：SM4 密钥扩展（伪代码）
void SM4KeySchedule(uint32_t key[4], uint32_t roundKeys[32]) {
    uint32_t rk = key[0] ^ key[1] ^ key[2] ^ key[3];
    for (int i = 0; i < 32; i++) {
        rk = F(rk, CK[i]); // F 为轮函数，CK 为固定常数
        roundKeys[i] = rk;
    }
}

// sm4.h
#ifndef __SM4_H__
#define __SM4_H__
#include <string.h>

#ifndef GET_ULONG_BE
#define GET_ULONG_BE(n,b,i) {\n    (n)=((unsigned long)(b)[(i)]<<24)|((unsigned long)(b)[(i)+1]<<16)|((unsigned long)(b)[(i)+2]<<8)|((unsigned long)(b)[(i)+3]);\n}
#endif

#ifndef PUT_ULONG_BE
#define PUT_ULONG_BE(n,b,i) {\n    (b)[(i)]=(unsigned char)((n)>>24);\n    (b)[(i)+1]=(unsigned char)((n)>>16);\n    (b)[(i)+2]=(unsigned char)((n)>>8);\n    (b)[(i)+3]=(unsigned char)((n));\n}
#endif

#define SHL(x,n) (((x)&0xFFFFFFFF)<< n)
#define ROTL(x,n) (SHL((x),n)|((x)>>(32- n)))
#define SWAP(a,b){ unsigned long t = a; a = b; b = t; t=0; }

static const unsigned char SboxTable[16][16]={{0xd6,0x90,0xe9,0xfe,0xcc,0xe1,0x3d,0xb7,0x16,0xb6,0x14,0xc2,0x28,0xfb,0x2c,0x05},{0x2b,0x67,0x9a,0x76,0x2a,0xbe,0x04,0xc3,0xaa,0x44,0x13,0x26,0x49,0x86,0x06,0x99},{0x9c,0x42,0x50,0xf4,0x91,0xef,0x98,0x7a,0x33,0x54,0x0b,0x43,0xed,0xcf,0xac,0x62},{0xe4,0xb3,0x1c,0xa9,0xc9,0x08,0xe8,0x95,0x80,0xdf,0x94,0xfa,0x75,0x8f,0x3f,0xa6},{0x47,0x07,0xa7,0xfc,0xf3,0x73,0x17,0xba,0x83,0x59,0x3c,0x19,0xe6,0x85,0x4f,0xa8},{0x68,0x6b,0x81,0xb2,0x71,0x64,0xda,0x8b,0xf8,0xeb,0x0f,0x4b,0x70,0x56,0x9d,0x35},{0x1e,0x24,0x0e,0x5e,0x63,0x58,0xd1,0xa2,0x25,0x22,0x7c,0x3b,0x01,0x21,0x78,0x87},{0xd4,0x00,0x46,0x57,0x9f,0xd3,0x27,0x52,0x4c,0x36,0x02,0xe7,0xa0,0xc4,0xc8,0x9e},{0xea,0xbf,0x8a,0xd2,0x40,0xc7,0x38,0xb5,0xa3,0xf7,0xf2,0xce,0xf9,0x61,0x15,0xa1},{0xe0,0xae,0x5d,0xa4,0x9b,0x34,0x1a,0x55,0xad,0x93,0x32,0x30,0xf5,0x8c,0xb1,0xe3},{0x1d,0xf6,0xe2,0x2e,0x82,0x66,0xca,0x60,0xc0,0x29,0x23,0xab,0x0d,0x53,0x4e,0x6f},{0xd5,0xdb,0x37,0x45,0xde,0xfd,0x8e,0x2f,0x03,0xff,0x6a,0x72,0x6d,0x6c,0x5b,0x51},{0x8d,0x1b,0xaf,0x92,0xbb,0xdd,0xbc,0x7f,0x11,0xd9,0x5c,0x41,0x1f,0x10,0x5a,0xd8},{0x0a,0xc1,0x31,0x88,0xa5,0xcd,0x7b,0xbd,0x2d,0x74,0xd0,0x12,0xb8,0xe5,0xb4,0xb0},{0x89,0x69,0x97,0x4a,0x0c,0x96,0x77,0x7e,0x65,0xb9,0xf1,0x09,0xc5,0x6e,0xc6,0x84},{0x18,0xf0,0x7d,0xec,0x3a,0xdc,0x4d,0x20,0x79,0xee,0x5f,0x3e,0xd7,0xcb,0x39,0x48}};

static const unsigned long FK[4]={0xa3b1bac6,0x56aa3350,0x677d9197,0xb27022dc};
static const unsigned long CK[32]={0x00070e15,0x1c232a31,0x383f464d,0x545b6269,0x70777e85,0x8c939aa1,0xa8afb6bd,0xc4cbd2d9,0xe0e7eef5,0xfc030a11,0x181f262d,0x343b4249,0x50575e65,0x6c737a81,0x888f969d,0xa4abb2b9,0xc0c7ced5,0xdce3eaf1,0xf8ff060d,0x141b2229,0x30373e45,0x4c535a61,0x686f767d,0x848b9299,0xa0a7aeb5,0xbcc3cad1,0xd8dfe6ed,0xf4fb0209,0x10171e25,0x2c333a41,0x484f565d,0x646b7279};

class SM4 {
public:
    static void Enc(const unsigned char* src, unsigned char* dst, size_t& len, unsigned char iv[16], const unsigned char key[16]) {
        size_t i, n;
        unsigned long sk[32];
        SetKey(sk, key);
        i = 16 - (len % 16);
        char* pBuf = nullptr;
        if (i > 0) {
            pBuf = new char[len + i];
            memset(pBuf, (int)i, len + i);
            memcpy(pBuf, src, len);
            src = (unsigned char*)pBuf;
        }
        len += i;
        i = len;
        while (i > 0) {
            for (n = 0; n < 16; n++) {
                dst[n] = src[n] ^ iv[n];
            }
            OneRound(sk, dst, dst);
            memcpy(iv, dst, 16);
            src += 16;
            dst += 16;
            i -= 16;
        }
        delete[] pBuf;
    }

    static void Dec(const unsigned char* src, unsigned char* dst, size_t& len, unsigned char iv[16], const unsigned char key[16]) {
        size_t i, n;
        unsigned long sk[32];
        unsigned char tmp[16];
        SetKey(sk, key);
        for (n = 0; n < 16; n++) {
            SWAP(sk[n], sk[31 - n])
        }
        unsigned char* fst = dst;
        i = len;
        while (i > 0) {
            memcpy(tmp, src, 16);
            OneRound(sk, src, dst);
            for (n = 0; n < 16; n++) {
                dst[n] = dst[n] ^ iv[n];
            }
            memcpy(iv, tmp, 16);
            src += 16;
            dst += 16;
            i -= 16;
        }
        i = fst[len - 1];
        if (i >= 1 && i <= 16) {
            memset(fst + len - i, 0, i);
            len -= i;
        }
    }

private:
    static void SetKey(unsigned long SK[32], const unsigned char key[16]) {
        unsigned long MK[4];
        unsigned long k[36];
        GET_ULONG_BE(MK[0], key, 0);
        GET_ULONG_BE(MK[1], key, 4);
        GET_ULONG_BE(MK[2], key, 8);
        GET_ULONG_BE(MK[3], key, 12);
        k[0] = MK[0] ^ FK[0];
        k[1] = MK[1] ^ FK[1];
        k[2] = MK[2] ^ FK[2];
        k[3] = MK[3] ^ FK[3];
        for (unsigned long i = 0; i < 32; i++) {
            k[i + 4] = k[i] ^ CalciRK(k[i + 1] ^ k[i + 2] ^ k[i + 3] ^ CK[i]);
            SK[i] = k[i + 4];
        }
    }

    static void OneRound(unsigned long sk[32], const unsigned char input[16], unsigned char output[16]) {
        unsigned long i = 0;
        unsigned long ulbuf[36];
        memset(ulbuf, 0, sizeof(ulbuf));
        GET_ULONG_BE(ulbuf[0], input, 0);
        GET_ULONG_BE(ulbuf[1], input, 4);
        GET_ULONG_BE(ulbuf[2], input, 8);
        GET_ULONG_BE(ulbuf[3], input, 12);
        while (i < 32) {
            ulbuf[i + 4] = RoundF(ulbuf[i], ulbuf[i + 1], ulbuf[i + 2], ulbuf[i + 3], sk[i]);
            i++;
        }
        PUT_ULONG_BE(ulbuf[35], output, 0);
        PUT_ULONG_BE(ulbuf[34], output, 4);
        PUT_ULONG_BE(ulbuf[33], output, 8);
        PUT_ULONG_BE(ulbuf[32], output, 12);
    }

    static unsigned long CalciRK(unsigned long ka) {
        unsigned long bb = 0;
        unsigned long rk = 0;
        unsigned char a[4];
        unsigned char b[4];
        PUT_ULONG_BE(ka, a, 0);
        b[0] = SBox(a[0]);
        b[1] = SBox(a[1]);
        b[2] = SBox(a[2]);
        b[3] = SBox(a[3]);
        GET_ULONG_BE(bb, b, 0);
        rk = bb ^ (ROTL(bb, 13)) ^ (ROTL(bb, 23));
        return rk;
    }

    /// <summary>
    /// 8 位输入输出的非线性置换，增强混淆性
    /// </summary>
    /// <param name="inch"></param>
    /// <returns></returns>
    static unsigned char SBox(unsigned char inch) {
        unsigned char* pTable = (unsigned char*)SboxTable;
        unsigned char retVal = (unsigned char)(pTable[inch]);
        return retVal;
    }

    /// <summary>
    /// 线性扩散层，通过循环移位和异或操作扩散比特。
    /// </summary>
    /// <param name="ka"></param>
    /// <returns></returns>
    static unsigned long Linear(unsigned long ka) {
        unsigned long bb = 0;
        unsigned long c = 0;
        unsigned char a[4];
        unsigned char b[4];
        PUT_ULONG_BE(ka, a, 0);
        b[0] = SBox(a[0]);
        b[1] = SBox(a[1]);
        b[2] = SBox(a[2]);
        b[3] = SBox(a[3]);
        GET_ULONG_BE(bb, b, 0);
        c = bb ^ (ROTL(bb, 2)) ^ (ROTL(bb, 10)) ^ (ROTL(bb, 18)) ^ (ROTL(bb, 24));
        return c;
    }

    static unsigned long RoundF(unsigned long x0, unsigned long x1, unsigned long x2, unsigned long x3, unsigned long rk) {
        return (x0 ^ Linear(x1 ^ x2 ^ x3 ^ rk));
    }
};
#endif

// 测试代码
void Test_SM4() {
    const char* pMsg = "Hello, World，I love 中国.";
    unsigned char key[17] = {1,2,3,4,5,6,7,8,'A','B','C','D','E','F','G','H',0};
    unsigned char ciphertext[256] = {0};
    unsigned char iv[16] = {0};
    size_t out_len = strlen(pMsg);
    SM4::Enc((unsigned char*)pMsg, ciphertext, out_len, iv, key);
    printf("明文数据:%d,%s\n加密结果:%d,", (int)strlen(pMsg), pMsg, (int)out_len);
    for (size_t i = 0; i < out_len; i++) {
        printf("%02X ", ciphertext[i]);
    }
    printf("\n");
    unsigned char text[256] = {0};
    size_t text_len = out_len;
    memset(iv, 0, sizeof(iv));
    SM4::Dec((unsigned char*)ciphertext, text, text_len, iv, key);
    text[text_len] = '\0';
    printf("解密结果:%d,%s", (int)text_len, text);
}

明文数据:26,Hello, World，I love 中国.
加密结果:32,9F 0C D6 19 50 33 2F DB 4B 29 10 8E 42 AE 03 C8 01 CB 04 7C 8B 80 E3 C6 B0 22 AF 6D A2 2E 52 67
解密结果:26,Hello, World，I love 中国.

SM4 国密算法原理与 C++跨平台实现详解

引言

一、SM4 算法基础

1.1 算法概述

1.2 核心原理

1.3 核心参数

1.4 核心加密算法流程

二、算法原理详解

2.1 加密流程

2.2 轮函数结构

2.3 加密模式与填充

2.4 对称加密算法对比

三、算法实现技术

3.1 硬件实现

3.2 软件实现

四、安全性分析

4.1 算法设计安全性

4.2 抗攻击能力

4.3 标准化与权威认证

4.4 实际应用中的安全增强

4.5、与 AES 的对比

五、SM4 的 CBC 算法实现（C++）

5.1 源码（`Header-Only`）

代码分析说明

5.2 单元测试

5.2 测试结果

更多推荐文章

相关免费在线工具

SM4 国密算法原理与 C++跨平台实现详解

引言

一、SM4 算法基础

1.1 算法概述

1.2 核心原理

1.3 核心参数

1.4 核心加密算法流程

二、算法原理详解

2.1 加密流程

2.2 轮函数结构

2.3 加密模式与填充

2.4 对称加密算法对比

三、算法实现技术

3.1 硬件实现

3.2 软件实现

四、安全性分析

4.1 算法设计安全性

4.2 抗攻击能力

4.3 标准化与权威认证

4.4 实际应用中的安全增强

4.5、与 AES 的对比

五、SM4 的 CBC 算法实现（C++）

5.1 源码（Header-Only）

代码分析说明

5.2 单元测试

5.2 测试结果

微信扫一扫，关注极客日志

更多推荐文章

相关免费在线工具

5.1 源码（`Header-Only`）