On 07/28, Roland McGrath wrote:
>
> > Introduce the empty inline tracehook_finish_jctl() helper called by
> > do_signal_stop() after wakeup.
>
> You could possibly roll this into the tracehook_notify_jctl rework too.
> Or send it upstream separately, either way.
I'd prefer to send it separately, just to avoid the "how this hook is
connected to subtle changes in do_signal_stop" questions/discussion.
> > +/**
> > + * tracehook_finish_jctl - report about return from job control stop
> > + *
> > + * This is called by do_signal_stop() after wakeup.
>
> + * Called without locks.
>
> Just for consistency/completeness of tracehook.h kerneldoc.
OK, will re-send.
> > } while (try_to_freeze());
> > +
> > + tracehook_finish_jctl();
> > /*
> > * Now we don't run again until continued.
> > */
>
> Move the hook after this comment, since the comment is a "stopped now"
> comment and the hook is the "continued again" hook.
Hmm... shouldn't we move this comment above schedule() ? And, perhaps,
/* Now we don't run again until continued or killed */
?
Oleg.
