Re: [PATCH v3] do_wait: make PIDTYPE_PID case O(1) instead of O(n)

[Eric W. Biederman]

Oleg Nesterov <oleg@xxxxxxxxxx> writes:

> On 03/10, Eric W. Biederman wrote:
>>
>> Jim Newsome <jnewsome@xxxxxxxxxxxxxx> writes:
>>
>> > +static int do_wait_pid(struct wait_opts *wo)
>> > +{
>> > +	struct task_struct *target = pid_task(wo->wo_pid, PIDTYPE_PID);
>>                                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>> This is subtle change in behavior.
>>
>> Today on the task->children list we only place thread group leaders.
>
> Aaah, yes, thanks Eric!
>
>> So the code either needs a thread_group_leader filter on target before
>> the ptrace=0 case or we need to use "pid_task(wo->wo_pid, PIDTYPE_TGID)"
>> and "pid_task(wo->wo_pid, PIDTYPE_PID)" for the "ptrace=1" case.
>
> Agreed,
>
>> I would like to make thread_group_leaders go away
>
> Hmm, why?

Mostly because we have class of very nasty bugs to fix because code
thinks one thread is special.

There has been and I think still is code that mishandles zombie thread
group leaders.

Particularly nasty are zombie thread group leaders after userspace has
called setresuid in a way that changes signal permissions.

Eric