[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 at 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.