lkml.org 
[lkml]   [2011]   [Mar]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 7/8] job control: Notify the real parent of job control events regardless of ptrace
Hey, Oleg.

On Tue, Mar 22, 2011 at 08:44:16PM +0100, Oleg Nesterov wrote:
> On 03/22, Tejun Heo wrote:
> >
> > On Mon, Mar 21, 2011 at 06:43:06PM +0100, Oleg Nesterov wrote:
> > > Hmm... in fact I can't convince myself we really need to look at
> > > child->group_leader, will recheck... Anyway, this is minor too.
> >
> > Care to elaborate?
>
> child->real_parent must be equal to child->group_leader->real_parent.

Okay.

> What we need is ptrace_reparented(). But as I said, it should be
> fixed first (I think) and this needs a separate patch/discussion.
>
> So yes, we need a new helper.
>
> Well, and look at get_signal_to_deliver(),
>
> leader = current->group_leader;
> if (task_ptrace(leader) && !real_parent_is_ptracer(leader))
> do_notify_parent_cldstop(leader, true, why);
>
> Indeed. But in fact task_ptrace(leader) is not needed. If
> real_parent_is_ptracer(tsk) returns false, then tsk must be ptraced.
>
> So we could write
>
> if (!real_parent_is_ptracer(leader))
> do_notify_parent_cldstop(leader, true, why);
>
> but now this looks confusing because of the naming. Compare with
>
> if (ptrace_reparented(leader))
> do_notify_parent_cldstop(leader);

I see but I don't think ptrace_reparented() is a good name for
!same_thread_group(child->parent, child->real_parent). It gets too
confusing with child->parent != child->real_parent test.

Anyways, I'll leave the renaming for later patch and just update the
->group_leader part.

Thanks.

--
tejun


\
 
 \ /
  Last update: 2011-03-23 10:19    [from the cache]
©2003-2011 Jasper Spaans