Re: [PATCH v8 2/3] pid: Introduce pidfd_getfd syscall

[Matthew Wilcox <willy@infradead.org>]

On Fri, Jan 03, 2020 at 08:29:27AM -0800, Sargun Dhillon wrote:
> +++ b/kernel/pid.c
> @@ -578,3 +578,93 @@ void __init pid_idr_init(void)
>  	init_pid_ns.pid_cachep = KMEM_CACHE(pid,
>  			SLAB_HWCACHE_ALIGN | SLAB_PANIC | SLAB_ACCOUNT);
>  }
> +
> +static struct file *__pidfd_fget(struct task_struct *task, int fd)
> +{
> +	struct file *file;
> +	int ret;
> +
> +	ret = mutex_lock_killable(&task->signal->cred_guard_mutex);
> +	if (ret)
> +		return ERR_PTR(ret);
> +
> +	if (ptrace_may_access(task, PTRACE_MODE_ATTACH_REALCREDS))
> +		file = fget_task(task, fd);
> +	else
> +		file = ERR_PTR(-EPERM);
> +
> +	mutex_unlock(&task->signal->cred_guard_mutex);
> +
> +	return file ?: ERR_PTR(-EBADF);
> +}
> +
> +static int pidfd_getfd(struct pid *pid, int fd)
> +{
> +	struct task_struct *task;
> +	struct file *file;
> +	int ret;
> +
> +	task = get_pid_task(pid, PIDTYPE_PID);
> +	if (!task)
> +		return -ESRCH;
> +
> +	file = __pidfd_fget(task, fd);
> +	put_task_struct(task);
> +	if (IS_ERR(file))
> +		return PTR_ERR(file);
> +
> +	ret = security_file_receive(file);
> +	if (ret) {
> +		fput(file);
> +		return ret;
> +	}
> +
> +	ret = get_unused_fd_flags(O_CLOEXEC);
> +	if (ret < 0)
> +		fput(file);
> +	else
> +		fd_install(ret, file);
> +
> +	return ret;
> +}
> +
> +/**
> + * sys_pidfd_getfd() - Get a file descriptor from another process
> + *
> + * @pidfd:	the pidfd file descriptor of the process
> + * @fd:		the file descriptor number to get
> + * @flags:	flags on how to get the fd (reserved)
> + *
> + * This syscall gets a copy of a file descriptor from another process
> + * based on the pidfd, and file descriptor number. It requires that
> + * the calling process has the ability to ptrace the process represented
> + * by the pidfd. The process which is having its file descriptor copied
> + * is otherwise unaffected.
> + *
> + * Return: On success, a cloexec file descriptor is returned.
> + *         On error, a negative errno number will be returned.
> + */

We don't usually kernel-doc syscalls.  They should have manpages instead.

> +SYSCALL_DEFINE3(pidfd_getfd, int, pidfd, int, fd,
> +		unsigned int, flags)
> +{
> +	struct pid *pid;
> +	struct fd f;
> +	int ret;
> +
> +	/* flags is currently unused - make sure it's unset */
> +	if (flags)
> +		return -EINVAL;

Is EINVAL the right errno here?  Often we use ENOSYS for bad flags to
syscalls.

> +	f = fdget(pidfd);
> +	if (!f.file)
> +		return -EBADF;
> +
> +	pid = pidfd_pid(f.file);
> +	if (IS_ERR(pid))
> +		ret = PTR_ERR(pid);
> +	else
> +		ret = pidfd_getfd(pid, fd);

You can simplify this by having pidfd_pid() return ERR_PTR(-EBADF) if
!f.file, and having pidfd_getfd() return PTR_ERR() if IS_ERR(pid).  Then
this function looks like:

	if (flags)
		return -EINVAL;

	f = fdget(pidfd);
	pid = pidfd_pid(f.file);
	ret = pidfd_getfd(pid, fd);
	fdput(f);
	return ret;

You could even eliminate the 'pid' variable and just do:

	ret = pidfd_getfd(pidfd_pid(f.file), fd);

but that's a step too far for me.

It's unfortunate that -EBADF might mean that either the first or second
argument is a bad fd number.  I'm not sure I have a good alternative though.