Re: [PATCH 6/6] freezer: kill unused set_freezable_with_signal()

[Oleg Nesterov <oleg@redhat.com>]

On 09/05, Tejun Heo wrote:
>
> Hello, Oleg.
>
> On Sun, Sep 04, 2011 at 08:46:26PM +0200, Oleg Nesterov wrote:
> > > @@ -72,10 +72,6 @@ bool __refrigerator(bool check_kthr_stop)
> > >  		schedule();
> > >  	}
> > >
> > > -	spin_lock_irq(&current->sighand->siglock);
> > > -	recalc_sigpending(); /* We sent fake signal, clean it up */
> > > -	spin_unlock_irq(&current->sighand->siglock);
> > > -
> >
> > Why? This recalc_sigpending() makes sense imho. Otherwise the user-space
> > tasks (almost) always return with TIF_SIGPENDING. May be we can do this
> > under "if (PF_KTRHREAD)".
>
> PF_KTHREAD no longer gets TIF_SIGPENDING set, so...

Yes,

> > For example. Suppose the user-space task does wait_event_freezable()...
> >
> > Hmm. OTOH, wait_event_freezable() looks wrong anyway... So probably
> > this doesn't matter. ptrace_stop/get_signal_to_deliver doesn't need
> > this, probably we do not care about the other callers.
>
> Can you elaborate on it being wrong?  Do you mean the possibility of
> leaking spurious TIF_SIGPENDING?

Perhaps it is correct... Just I do not understand what it should do.
I thought it is "wait_for_event && do_not_block_freezer". And at first
glance the code looks as if it tries to do this. Say, in the "likely"
case we restart wait_event_interruptible() after refrigerator().

But this looks racy. Suppose that freezing() is already false when
try_to_freeze() or __refrigerator() is called. Say, cgroup_freezer does
freeze_task() + __thaw_task(). Why it returns -ERESTARTSYS in this case?

And if it can be used by the userspace thread, then we should probably
do recalc_sigpending() somewhere, otherwise wait_event_freezable() will
always return -ERESTARTSYS after __refrigerator().

Oleg.