Re: [PATCH v3 1/1] pidfd: implement PIDFD_THREAD flag for pidfd_open()

[Tycho Andersen <tycho@tycho.pizza>]

Hi Oleg,

On Wed, Jan 31, 2024 at 02:26:02PM +0100, Oleg Nesterov wrote:
> With this flag:
> 
> 	- pidfd_open() doesn't require that the target task must be
> 	  a thread-group leader
> 
> 	- pidfd_poll() succeeds when the task exits and becomes a
> 	  zombie (iow, passes exit_notify()), even if it is a leader
> 	  and thread-group is not empty.
> 
> 	  This means that the behaviour of pidfd_poll(PIDFD_THREAD,
> 	  pid-of-group-leader) is not well defined if it races with
> 	  exec() from its sub-thread; pidfd_poll() can succeed or not
> 	  depending on whether pidfd_task_exited() is called before
> 	  or after exchange_tids().
> 
> 	  Perhaps we can improve this behaviour later, pidfd_poll()
> 	  can probably take sig->group_exec_task into account. But
> 	  this doesn't really differ from the case when the leader
> 	  exits before other threads (so pidfd_poll() succeeds) and
> 	  then another thread execs and pidfd_poll() will block again.

I don't have a strong opinion here; leaving it "undefined" for now is
fine with me.

> @@ -2173,7 +2195,9 @@ static int __pidfd_prepare(struct pid *pid, unsigned int flags, struct file **re
>   */
>  int pidfd_prepare(struct pid *pid, unsigned int flags, struct file **ret)
>  {
> -	if (!pid || !pid_has_task(pid, PIDTYPE_TGID))
> +	bool thread = flags & PIDFD_THREAD;
> +
> +	if (!pid || !pid_has_task(pid, thread ? PIDTYPE_PID : PIDTYPE_TGID));

Small typo here, trailing ;. When I fix that, tests pass for me.

Assuming that's fixed up:

Reviewed-by: Tycho Andersen <tandersen@netflix.com>
Tested-by: Tycho Andersen <tandersen@netflix.com>

Thanks for your help!

Tycho