Messages in this thread | | | Date | Tue, 24 Mar 2015 20:02:29 +0100 | From | Oleg Nesterov <> | Subject | Re: [PATCH] mm: fix lockdep build in rcu-protected get_mm_exe_file() |
| |
On 03/24, Konstantin Khlebnikov wrote: > > On Tue, Mar 24, 2015 at 9:10 PM, Oleg Nesterov <oleg@redhat.com> wrote: > >>> > >>> "atomic_read(&mm->mm_users) <= 1" looks a bit more "safe". But again, > >>> I won't insist. > >> > >> Not so safe: this will race with get_task_mm(). > > > > How? > > I mean rcu/lockdep debug migh race with get_task_mm() and generate > false-positive warning about non-protected rcu_dereference.
Still can't understand, I think it can't... and if it could, then this warning would not be false positive.
Anut this doesn't matter because we seem to agree this check should go away.
> > Yeees, probably rcu_dereference_raw() would be even better. set_mm_exe_file() > > must be called only if nobody but us can access this mm. > > Yep.
Great. Davidlohr will you agree?
Oleg.
| |