I am implementing with shared memory.
In order to access the same shared memory, whenever the shared memory address assigned pid and the current pid are different, the address is reassigned through system V's shmat.
When the child process terminates or the parent process deletes the shared memory, it shmdt the last shmat address and shmctl the shared memory id.

I think this is correct behavior for spawn, but is this also correct behavior for fork?
If I use fork, I expect the same memory access without shmat even though the pid is different.

If it was a process created with fork, when it had the id and address of the shared memory, I thought that the child process should either run shmdt on the existing address or update the pid.

1. Is there a way to check if the id and address of the shared memory that the process has are valid?
2. Does shmdt in a forked child process not affect the parent process? (Does the parent process still access to shared memory well?)
3. Is it necessary to shmdt all shared memory owned by the parent process in the child process created by fork in order for normal shmctl to proceed?