Re: [PATCH RFC 1/2] Add polling support to pidfd

From: Oleg Nesterov
Date: Fri Apr 19 2019 - 14:49:24 EST

On 04/18, Jann Horn wrote:
> On Wed, Apr 17, 2019 at 3:09 PM Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
> > On 04/16, Joel Fernandes wrote:
> > > On Tue, Apr 16, 2019 at 02:04:31PM +0200, Oleg Nesterov wrote:
> > > >
> > > > Could you explain when it should return POLLIN? When the whole process exits?
> > >
> > > It returns POLLIN when the task is dead or doesn't exist anymore, or when it
> > > is in a zombie state and there's no other thread in the thread group.
> >
> > IOW, when the whole thread group exits, so it can't be used to monitor sub-threads.
> >
> > just in case... speaking of this patch it doesn't modify proc_tid_base_operations,
> > so you can't poll("/proc/sub-thread-tid") anyway, but iiuc you are going to use
> > the anonymous file returned by CLONE_PIDFD ?
> I don't think procfs works that way. /proc/sub-thread-tid has
> proc_tgid_base_operations despite not being a thread group leader.

Yes, sorry, I meant /proc/pid/task/sub-thread-tid.

But poll("/proc/sub-thread-tid") won't work too, we can't rely on do_notify_parent()
if the task is not a group leader.

> (Yes, that's kinda weird.) AFAICS the WARN_ON_ONCE() in this code can
> be hit trivially, and then the code will misbehave.

Heh, I didn't even notice that WARN_ON_ONCE(task && !thread_group_leader(task)) ;)

> @Joel: I think you'll have to either rewrite this to explicitly bail
> out if you're dealing with a thread group leader, or make the code
> work for threads, too.

The last version of CLONE_PIDFD doesn't allow CLONE_THREAD, so we can forget
about this problem for now.