XInitThreads引发内存泄漏，如何主动释放其分配的内存？

我在用minilibx（X11的封装库），为了让库支持线程，我在初始化函数里加了XInitThreads()。编译出的程序叫fdf，用Valgrind（命令：valgrind --log-file=v.log --leak-check=full ./fdf test2.fdf）检测内存泄漏时，发现XInitThreads分配的内存没被释放。有没有主动释放这块内存的方法？程序结束时需要调用什么函数来释放XInitThreads分配的内存？

系统信息（Ubuntu） #

Linux 6.11.0-1-t2-jammy #1 SMP PREEMPT_DYNAMIC Tue Sep 17 06:47:39 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux

Valgrind日志（v.log） #

==1546656== Memcheck, a memory error detector
==1546656== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==1546656== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info
==1546656== Command: ./fdf test2.fdf
==1546656== Parent PID: 1464122
==1546656== 
==1546656== 
==1546656== HEAP SUMMARY:
==1546656==     in use at exit: 120 bytes in 3 blocks
==1546656==   total heap usage: 53,837 allocs, 53,834 frees, 1,199,360 bytes allocated
==1546656== 
==1546656== 40 bytes in 1 blocks are still reachable in loss record 1 of 3
==1546656==    at 0x4848899: malloc (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==1546656==    by 0x48B09E9: XInitThreads (in /usr/lib/x86_64-linux-gnu/libX11.so.6.4.0)
==1546656==    by 0x11422A: mlx_init (in /home/baschnit/repos/42fdf_new/fdf)
==1546656==    by 0x10AAB1: main (main.c:59)
==1546656== 
==1546656== 40 bytes in 1 blocks are still reachable in loss record 2 of 3
==1546656==    at 0x4848899: malloc (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==1546656==    by 0x48B0A06: XInitThreads (in /usr/lib/x86_64-linux-gnu/libX11.so.6.4.0)
==1546656==    by 0x11422A: mlx_init (in /home/baschnit/repos/42fdf_new/fdf)
==1546656==    by 0x10AAB1: main (main.c:59)
==1546656== 
==1546656== 40 bytes in 1 blocks are still reachable in loss record 3 of 3
==1546656==    at 0x4848899: malloc (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==1546656==    by 0x48B0A23: XInitThreads (in /usr/lib/x86_64-linux-gnu/libX11.so.6.4.0)
==1546656==    by 0x11422A: mlx_init (in /home/baschnit/repos/42fdf_new/fdf)
==1546656==    by 0x10AAB1: main (main.c:59)
==1546656== 
==1546656== LEAK SUMMARY:
==1546656==    definitely lost: 0 bytes in 0 blocks
==1546656==    indirectly lost: 0 bytes in 0 blocks
==1546656==      possibly lost: 0 bytes in 0 blocks
==1546656==    still reachable: 120 bytes in 3 blocks
==1546656==         suppressed: 0 bytes in 0 blocks
==1546656== 
==1546656== For lists of detected and suppressed errors, rerun with: -s
==1546656== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 1 from 1)

最小复现步骤 #

• 从Git下载minilibx到libx文件夹
• 在父文件夹创建main.c文件，用命令编译：cc main.c libx/libmlx.a -Ilibx -lX11 -lXext
• 执行检测：valgrind --leak-check=full --show-leak-kinds=all ./a.out

main.c代码 #

#include <stdlib.h>
#include <X11/Xlib.h>
#include "mlx.h"

int main(void)
{
    void *mlx;

    XInitThreads();
    mlx = mlx_init();
    mlx_destroy_display(mlx);
    free(mlx);
}

首先明确：X11没有提供官方函数来释放XInitThreads()分配的内存。

从Valgrind日志能看到，这些内存标记为still reachable，意味着程序退出时内存是可访问的，但没有被主动释放。这类内存通常会由操作系统在进程终止后自动回收，不属于严重的内存泄漏问题。

为什么没有释放函数？ #

XInitThreads()初始化的是Xlib的全局线程安全结构，这些结构设计为在整个进程生命周期中存在，X11库本身不会提供销毁接口——因为一旦启用线程支持，进程中所有Xlib操作都依赖这些结构，直到进程结束。

怎么处理Valgrind的报告？ #

如果不想看到这些报告，有两个方案：

1. 忽略still reachable类型的泄漏：这类泄漏不会导致内存增长，进程结束后系统会回收，对程序运行没有影响。
2. 添加Valgrind抑制规则：创建一个抑制文件，让Valgrind忽略XInitThreads相关的内存分配。示例抑制规则如下：

   {
      <XInitThreads_leak>
      Memcheck:Leak
      match-leak-kinds: reachable
      fun:malloc
      fun:XInitThreads
   }
   

   保存为xinitthr.supp，然后运行Valgrind时加上--suppressions=xinitthr.supp参数即可。

注意事项 #

不要尝试手动释放这些内存——你没有办法安全获取这些内存的指针，强行释放会导致程序崩溃或未定义行为。

内容的提问来源于stack exchange，提问作者Bastian