Re: [PATCH] recalc_sigpending_tsk fixes

2007-05-16 Thread Roland McGrath
> We already discussed this, this is not so important, but how about > > void recalc_sigpending_and_wake(struct task_struct *t) > { > int was_pending = signal_pending(t); > > if (recalc_sigpending_tsk(t) && !was_pending) > signal_wake_

Re: [PATCH] recalc_sigpending_tsk fixes

2007-05-16 Thread Oleg Nesterov
On 05/16, Roland McGrath wrote: > > + * After recalculating TIF_SIGPENDING, we need to make sure the task wakes > up. > + * This is superfluous when called on current, the wakeup is a harmless > no-op. > + */ > +void recalc_sigpending_and_wake(struct task_struct *t) > +{ > + if (recalc_sigpe

[PATCH] recalc_sigpending_tsk fixes

2007-05-16 Thread Roland McGrath
Steve Hawkes discovered a problem where recalc_sigpending_tsk was called in do_sigaction but no signal_wake_up call was made, preventing later signals from waking up blocked threads with TIF_SIGPENDING already set. In fact, the few other calls to recalc_sigpending_tsk outside the signals code are