Re: [PATCH] Input: serio - re-add thaw (and add freeze)

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Tue, Feb 16, 2010 at 04:48:30PM +0000, Alan Jenkins wrote:
> On 2/4/10, Alan Jenkins <alan-jenkins@tuffmail.co.uk> wrote:
> > Dmitry Torokhov wrote:
> >> On Wed, Feb 03, 2010 at 06:59:54PM +0000, Alan Jenkins wrote:
> >>
> >>> 633aae2 "Input: i8042 - switch to using dev_pm_ops"
> >>> removed handling for PMSG_THAW, causing obscure breakage.
> >>>
> >>> It can break if you press keys during hibernation,
> >>> which causes subsequent keypresses to be lost -
> >>> so you can't cancel s2disk by pressing backspace -
> >>> and then just before system poweroff you get
> >>> "psmouse.c: Failed to deactivate mouse".
> >>>
> >>> So let's add the thaw handler back.
> >>>
> >>> Also set the freeze handler.  It looks like PMSG_FREEZE
> >>> didn't do anything in the past, but I think that must
> >>> have been an oversight.
> >>>
> >>
> >> We do not need to do anything special at freeze time that is why there
> >> wasn't freeze handler. And so thaw should be pretty much noop as well.
> 
> Ok, so I did some more mindless fiddling.  It seems that all it needs
> is to call the interrupt handler at thaw time.
> 

Yep, this makes more sense. If you happen to press the key(s) while
interrupts are off the keyboard controller becomes "jammed". Your sign
off please?

> Regards
> Alan
> 
> diff --git a/drivers/input/serio/i8042.c b/drivers/input/serio/i8042.c
> index d84a36e..b54aee7 100644
> --- a/drivers/input/serio/i8042.c
> +++ b/drivers/input/serio/i8042.c
> @@ -1161,9 +1161,17 @@ static int i8042_pm_restore(struct device *dev)
>  	return 0;
>  }
> 
> +static int i8042_pm_thaw(struct device *dev)
> +{
> +	i8042_interrupt(0, NULL);
> +
> +	return 0;
> +}
> +
>  static const struct dev_pm_ops i8042_pm_ops = {
>  	.suspend	= i8042_pm_reset,
>  	.resume		= i8042_pm_restore,
> +	.thaw		= i8042_pm_thaw,
>  	.poweroff	= i8042_pm_reset,
>  	.restore	= i8042_pm_restore,
>  };

-- 
Dmitry