c - Valgrind 说 "definitely leak"但它是吗？

Question

我正在尝试为井字游戏编写蒙特卡洛树搜索，但在运行时确实存在内存问题(例如，我的计算机内存不足)。

所以我决定用 valgrind 调查情况。

下面是 valgrind 说“绝对泄漏”的代码块之一。

void player_init (Player **p, player_t symbol)
{
  *p = (Player *) malloc (sizeof (Player));
  (**p).symbol = symbol;
  (**p).score = 0;
}

void player_init_all (Player ***p)
{
  *p = (Player **) malloc (sizeof (Player *) * 2);
  for (int i = 0; i < 2; i++)
    {
      (*p)[i] = (Player *) malloc (sizeof (Player));
    }
  player_init (&(*p)[0], PLAYER1);
  player_init (&(*p)[1], PLAYER2);
}

void player_destroy (Player *p)
{
  free (p);
}

其中 Player 和 player_t

typedef char player_t;
typedef struct player Player;
struct player {
    player_t symbol;
    unsigned score;
};

他们就是这样使用的；

int main (int argc, char** argv)
{
    Player **players;
    player_init_all (&players);

    // OTHER FANCY CODE HERE

    for (int i = 0; i < 2; i++)
      player_destroy (players[i]);

    free (players);
    free (board);
    return 0;
}

我是否以错误的方式传递指针？

Valgrind 转储；

==21657== 16 bytes in 1 blocks are definitely lost in loss record 1 of 15
==21657==    at 0x4C2DB8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==21657==    by 0x40147D: player_init_all (main.c:348)
==21657==    by 0x401698: main (main.c:426)
==21657== 
==21657== 16 bytes in 2 blocks are definitely lost in loss record 2 of 15
==21657==    at 0x4C2DB8F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==21657==    by 0x4014AF: player_init_all (main.c:351)
==21657==    by 0x401698: main (main.c:426)

第 348 行 是 player_init_all 的开始，第 351 行 是 中 for 循环的开始>player_init_all

Answer 1

这里有一个相当明显的内存泄漏

void player_init_all (Player ***p)
{
  *p = (Player **) malloc (sizeof (Player *) * 2);
  for (int i = 0; i < 2; i++)
    {
      (*p)[i] = (Player *) malloc (sizeof (Player));
    }
  player_init (&(*p)[0], PLAYER1);
  player_init (&(*p)[1], PLAYER2);
}

您在上述循环中为 Player 对象分配内存。但紧接着 player_init 将再次分配它并覆盖 (*p)[0] 和 (*p)[1] 的值，肯定会泄漏您在上述循环中分配的内容。