ide_delay_50ms question

ide_delay_50ms question

[Jani Monoses]

Hi
could anybody tell me why is the mdelay(50) solution better than the
schedule_timeout(HZ/20) in ide_delay_50ms.I suppose because it is bootime
and there is nothing to schedule.
But then the condition should also extend to CONFIG_BLK_DEV_IDECS_MODULE
as well, because when I insert an IDE compactflash card things stop for a
second or so nad I use a modular driver.
And I don't know about removable harddrives but isn't the schedule_timeout
solution better for them as well?


void ide_delay_50ms (void)
{
#ifndef CONFIG_BLK_DEV_IDECS
        mdelay(50);
#else
        __set_current_state(TASK_UNINTERRUPTIBLE);
        schedule_timeout(HZ/20);
#endif /* CONFIG_BLK_DEV_IDECS */
}

Re: ide_delay_50ms question

[David Woodhouse]

jani@astechnix.ro said:
> But then the condition should also extend to CONFIG_BLK_DEV_IDECS_MODUL
> E as well, because when I insert an IDE compactflash card things stop
> for a second or so nad I use a modular driver.

Urgh. Please don't make in-kernel code depend on CONFIG_*_MODULE any more 
than it already does. It's disgusting. 

Why can't we just schedule unconditionally?

--
dwmw2

Re: ide_delay_50ms question

[Alan Cox]

> as well, because when I insert an IDE compactflash card things stop for a
> second or so nad I use a modular driver.
> And I don't know about removable harddrives but isn't the schedule_timeout
> solution better for them as well?

Some of the current drivers call it with interrupts disabled. This is one of
the things that will have to wait until 2.5 is released to really address

Re: ide_delay_50ms question

[Andrzej Krzysztofowicz]

"Alan Cox wrote:"
> > as well, because when I insert an IDE compactflash card things stop for a
> > second or so nad I use a modular driver.
> > And I don't know about removable harddrives but isn't the schedule_timeout
> > solution better for them as well?
> 
> Some of the current drivers call it with interrupts disabled. This is one of
> the things that will have to wait until 2.5 is released to really address

AFAIR it (interrupts disabled) is necessary for old pre-IDE 16-bit
controllers.

Andrzej
-- 
=======================================================================
  Andrzej M. Krzysztofowicz               ankry@mif.pg.gda.pl
  phone (48)(58) 347 14 61
Faculty of Applied Phys. & Math.,   Technical University of Gdansk