On 08/02, Oleg Nesterov wrote: > > So Adrian, sorry for confusion, I think your patch is fine.
Yes... but do we really need the new CLONE_SET_TID ? set_tid == 0 has no effect, can't we simply check kargs->set_tid != 0 before ns_capable() ? and to remind, I still think alloc_pid() should use idr_is_empty(), but I won't insist. Oleg.
