Re: [PATCH 2/5] dinput: Stop and unload effect when freeing it.

[Daniel Remenak]

On Sat, Jul 5, 2008 at 9:24 PM, Vitaliy Margolen
<wine-patches(a)kievinfo.com> wrote:
<snip>
> +        LinuxInputEffectImpl_Stop(iface);
> +        LinuxInputEffectImpl_Unload(iface);
> <snip>

Vitaliy,

Native behavior is for Unload to stop the effect also (see msdn or
test it yourself), so the call to Stop here is unnecessary.  If the
call to Stop is required, it must be done in Unload, not Release, as
it's not unusual to unload an effect without releasing it when a game
has a large number of available effect instances.

It is my understanding that no explicit call to Stop is needed, either
in Release or Unload, as evdev has the same behavior documented (see
section 3.3 of ff.txt in the kernel docs).

The call to Unload is certainly correct, and I don't think there's
anything strictly wrong with calling Stop; just make sure it's not
covering up a bug in Unload.

Regards,
Daniel Remenak

[PATCH 2/5] dinput: Stop and unload effect when freeing it.

[Vitaliy Margolen]

Daniel Remenak wrote:
> On Sat, Jul 5, 2008 at 9:24 PM, Vitaliy Margolen
> <wine-patches(a)kievinfo.com> wrote:
> <snip>
> > +        LinuxInputEffectImpl_Stop(iface);
> > +        LinuxInputEffectImpl_Unload(iface);
> > <snip>
>
> Vitaliy,
>
> Native behavior is for Unload to stop the effect also (see msdn or
> test it yourself), so the call to Stop here is unnecessary.  If the
> call to Stop is required, it must be done in Unload, not Release, as
> it's not unusual to unload an effect without releasing it when a game
> has a large number of available effect instances.
>
> It is my understanding that no explicit call to Stop is needed, either
> in Release or Unload, as evdev has the same behavior documented (see
> section 3.3 of ff.txt in the kernel docs).
>
> The call to Unload is certainly correct, and I don't think there's
> anything strictly wrong with calling Stop; just make sure it's not
> covering up a bug in Unload.

You probably right. I did not experiment a lot with force feedback. Was 
trying to apply the same assumptions that exist in other places in dinput - 
anything that's being freed first have to stop then remove itself from the 
system.

Just trying to prevent any possible bugs that could be caused by not 
stopping/unloading force feedback effects.

Vitaliy.