c - 我将PID放入数组中，并且valgrind告诉我超出范围或分配的内存太少

Question

我坚持了valgrind用我的代码报告的这个问题。我正在向全局数组中添加pids（这样做是因为我想在一个信号处理程序中打印pids）。
valgrind似乎对将pid放入数组的代码行有问题：

PIDs[i] = cPID;

这是我的代码：

pid_t *PIDs;

void funct(int i) {
     //doing unrelated stuff

     rc = fork();
     if(rc==-1) {
          //error
     }
     else if(rc==0) {
          //child (this runs first b/c I used sleep on parent)
          rc=getpid();
          //execvp call
     }
     else {
          PIDs[i]=rc;
     }
}

int main() {

     //calculate no. of PIDs - I checked this and it reports the correct number
     int i = 8; 

     //allocate memory for PIDs array
     PIDs = malloc(sizeof(pid_t)*number_of_PIDS); //number_of_PIDs report 8
     funct(i);

     free(PIDs);
}

Valgrind错误如下：

==6310== LEAK SUMMARY:
==6310==    definitely lost: 0 bytes in 0 blocks
==6310==    indirectly lost: 0 bytes in 0 blocks
==6310==      possibly lost: 0 bytes in 0 blocks
==6310==    still reachable: 24 bytes in 1 blocks
==6310==         suppressed: 0 bytes in 0 blocks
==6310== Reachable blocks (those to which a pointer was found) are not shown.
==6310== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==6310==
==6310== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
==6310==
==6310== 1 errors in context 1 of 1:
==6310== Invalid write of size 4
==6310==    at 0x4010F3: funct
==6310==    by 0x401496: main
==6310==  Address 0x51db388 is 0 bytes after a block of size 24 alloc'd
==6310==    at 0x4C2A0B0: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==6310==    by 0x40134F: main
==6310==
==6310== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)

我做错什么了吗？或者有没有其他的建议去哪里寻找valgrind错误？

Answer 1

地址0x51db388是大小为24的块后的0字节
不可能是8码。24除以8等于3，pid_t几乎肯定是4。
您可能为6个pid_t分配了空间，但最终却使用了9个（包括0到8个）。