strace lockup when tracing exec in go

[Michal Hocko]

On Fri 23-09-16 16:07:24, Oleg Nesterov wrote:
> On 09/23, Michal Hocko wrote:
> >
> > On Fri 23-09-16 15:21:02, Oleg Nesterov wrote:
> > >
> > > This change is simply wrong no matter what.
> >
> > I've just tried to extend the existing
> >
> > 	/*
> > 	 * Tracers may want to know about even ignored signals.
> > 	 */
> > 	return !t->ptrace;
> >
> > but I probably just do not understand what that actually means. I
> > thought that the tracer is _really_ interested in hearing about the
> > signal.
> 
> Yes, the tracer is really interested to know that a signal was sent to
> the _tracee_, not the tracer ;)

OK, now it makes more sense. I was really scratching my head to
understand this part...
 
> > > We could change do_notify_parent()
> > > to call signal_wake_up() if tsk->ptrace, but see above, this won't help.
> >
> > So does this mean WONTFIX? Can we at least document this behavior? It
> > surely is unexpected.
> 
> No, no, no. Of course this must be fixed. The only problem is that I still
> do not know what should we do. I'll try to return to this problem next week.
> I'm afraid we will need to change de_thread() to wait until all other sub-
> threads have passed exit_notify() or even exit_signals(),

Yes making de_thread completely independent on the state of the tracer
would be a huge improvement. While playing with this test case I
triggered some other interesting hangs (e.g. strace hanging in tty while
trying to print something). So just few interruptible waits is not a
full solution.

> but ooh I don't
> like this. Plus in this case we will need to finally define what
> PTRACE_EVENT_EXIT should actually do.

OK, considering this has been broken for quite some time I do not think
we are in hurry. I am slightly worried about how such a solution would
be stable kernel safe but ohh well.
-- 
Michal Hocko
SUSE Labs