[RFC 5/11] Linux FF: Linux effect parameters: time parameters

[Andrew Eikum]

On Sun, Mar 09, 2014 at 11:38:06AM +0100, Elias Vanderstuyft wrote:
> On Wed, Mar 5, 2014 at 4:03 PM, Andrew Eikum <aeikum at codeweavers.com> wrote:
> > On Tue, Mar 04, 2014 at 08:20:09PM +0100, Elias Vanderstuyft wrote:
> >>     So change in dinput/effect_linuxinput.c:488 :
> >>         This->effect.u.periodic.period = tsp->dwPeriod / 1000;
> >>     to :
> >>         if (0 < tsp->dwPeriod && tsp->dwPeriod <= 1000)
> >>             This->effect.u.periodic.period = 1;
> >>         else
> >>             This->effect.u.periodic.period = tsp->dwPeriod / 1000;
> >>
> >
> > Seems good. While we're here, any reason not to use round(3) instead
> > of always rounding down in the >1000 case?
> 
> That's a good question, the reason is that I want to avoid floating
> point arithmetic (round(3)).

Okay. If using floating point is costly, then I agree the tiny (<1ms)
increase in accuracy isn't worth it.

> > Is this how Windows behaves
> > for massive dwPeriod? I would expect E_INVALIDARG, since it's out of
> > the range [0,36000).
> 
> I think you're mixing up dwPeriod with dwPhase. The range of dwPeriod
> is not defined in MSDN.
> 

Yes, you're right. Your updated patch seems fine.

> Maybe we should also change all occurrences of 10000 in a force
> context to DI_FFNOMINALMAX, in the whole FF dinput implementation?
> 

I would support a patch like that, yes.

(I've been busy lately, but I'll get to the rest of your mail sequence
soon!)

Andrew