Re: [RFC][PATCH v4 1/2] printk: Make printk() completely async

[Sergey Senozhatsky]

On (03/16/16 11:10), Byungchul Park wrote:
> > > What about this?
> > > 
> > > int need_flush_console;
> > > 
> > > 	while(1) {
> > > 		set_current_state(TASK_INTERRUPTIBLE);
> > > 		if (!need_flush_console)
> > > 			schedule();
> 
>                 else // This shoule be here, IMHO. Hm?

do __set_current_state(TASK_RUNNING) only if need_flush_console == true?
why? I see no problems with Petr's code.

> > > 		__set_current_state(TASK_RUNNING);
> > > 
> > > 		need_flush_console =  false;
> > > 
> > > > +		console_lock();
> > > > +		console_unlock();
> > > > +	}

	-ss