3、关键点来了：L 是线性的
L 定义 ： L(x) = x ^ (x<<<2) ^ (x<<<10) ^ (x<<<18) ^ (x<<<24)

这是一个线性变换, 满足：
L(A ^ B) = L(A) ^ L(B)

// ====== T-table ======
static uint32_t T0[256],T1[256],T2[256],T3[256];

// 预读 S盒, 对每个字节先进行 S+L 计算
void init_ttable(){
    for(int i=0;i<256;i++){
        uint32_t b=Sbox[i];
        T0[i]=L(b<<24);
        T1[i]=L(b<<16);
        T2[i]=L(b<<8);
        T3[i]=L(b);
    }
}

// ====== T-table 加密 ======
void sm4_encrypt_t(uint32_t in[4], uint32_t rk[32], uint32_t out[4]) {
    uint32_t x0 = in[0];
    uint32_t x1 = in[1];
    uint32_t x2 = in[2];
    uint32_t x3 = in[3];

    // 32轮计算 X4 = X0 ^ T(X1 ^ X2 ^ X3 ^ rk)
    for (int i = 0; i < 32; i++) {
        // 先整体 XOR
        uint32_t tmp = x1 ^ x2 ^ x3 ^ rk[i];

        // 再拆字节（这是 T-table 的入口）
        uint32_t a = (tmp >> 24) & 0xff;
        uint32_t b = (tmp >> 16) & 0xff;
        uint32_t c = (tmp >> 8)  & 0xff;
        uint32_t d = tmp & 0xff;

        // T-table 查表
        uint32_t t = T0[a] ^ T1[b] ^ T2[c] ^ T3[d];

        // Feistel 更新
        uint32_t newX = x0 ^ t;

        // 滚动
        x0 = x1;
        x1 = x2;
        x2 = x3;
        x3 = newX;
    }

    // 反序输出
    out[0] = x3;
    out[1] = x2;
    out[2] = x1;
    out[3] = x0;
}

// ====== T-table 解密 ======
void sm4_decrypt_t(uint32_t in[4], uint32_t rk[32], uint32_t out[4]) {
    uint32_t x0 = in[0];
    uint32_t x1 = in[1];
    uint32_t x2 = in[2];
    uint32_t x3 = in[3];

    for (int i = 0; i < 32; i++) {
        uint32_t tmp = x1 ^ x2 ^ x3 ^ rk[31 - i];

        uint32_t a = (tmp >> 24) & 0xff;
        uint32_t b = (tmp >> 16) & 0xff;
        uint32_t c = (tmp >> 8)  & 0xff;
        uint32_t d = tmp & 0xff;

        uint32_t t = T0[a] ^ T1[b] ^ T2[c] ^ T3[d];

        uint32_t newX = x0 ^ t;

        x0 = x1;
        x1 = x2;
        x2 = x3;
        x3 = newX;
    }

    out[0] = x3;
    out[1] = x2;
    out[2] = x1;
    out[3] = x0;
}

// ====== 测试 ======
int main(){
    uint32_t MK[4]={0x01234567,0x89abcdef,0xfedcba98,0x76543210};
    uint32_t P[4]={0x01234567,0x89abcdef,0xfedcba98,0x76543210};

    uint32_t rk[32],C[4],D[4];

    key_schedule(MK,rk);
    init_ttable();

    sm4_encrypt_t(P,rk,C);
    sm4_decrypt_t(C,rk,D);

    printf("Plain : %08x %08x %08x %08x\n",P[0],P[1],P[2],P[3]);
    printf("Cipher: %08x %08x %08x %08x\n",C[0],C[1],C[2],C[3]);
    printf("Dec   : %08x %08x %08x %08x\n",D[0],D[1],D[2],D[3]);

    return 0;
}