There is a race otherwise where we try to complete a pending IRP but
because the async is writing the report from another thread we didn't
find it and instead ignored it.
Instead we need to atomically check if there was a pending IRP, and if
the queue is empty, or queue the wait.
Later, when a report is going to be marked as pending, and if there's
someone waiting for it already, we instead complete it immediately.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
They are checked against the expected packet anyway, and they otherwise
are enforced to match their specific individual report length, which we
don't have any info about here.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
It now consistently fails on w1064_tsign, and it's called concurrently
with the handler invocation.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This is dependent on timing, and currently fails occasionally both on Windows
and Wine.
Signed-off-by: Zebediah Figura <zfigura@codeweavers.com>
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Adapting it for the IAsyncOperation_boolean interface and use it to
implement IForceFeedbackMotor_Try(Enable|Disable|Reset)Async.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And create it for RawGameController and RacingWheel motors.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
And complete it on the next IOCTL_WINETEST_HID_WAIT_EXPECT ioctl.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Checking that only a few axes get really mapped to the slider axes, and
that we shouldn't use HID_USAGE_SIMULATION_RUDDER if there's already an
Rz mapped axis.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Win7 has spurious device arrival / removal messages for the first polled
HID device created, probably when installing the driver the first time,
and it causes a timeout if we open it before it completed its sequence.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Using a generic EventHandler<IInspectable *> implementation.
This adds several todo_wine because we currently do not implement custom
game controller factories.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
It is important to add RawGameController instances to the vector in the
initialization callback, because some games check the vector, as well
as the Gamepad class vector in the OnGameControllerAdded callback.
This also removes the OnGameControllerAdded failure case, to avoid
leaking controllers that were already added to the controller vector,
and instead rely on the OnGameControllerRemoved call to do the cleanup.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
So that we can access all the controller devices conveniently through
DInput without having to reimplement the whole HID client, or having
to enumerate the devices every time a new device needs to be open,
and being able to workaround the limitations of the device the XInput
controllers are usually exposing.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
This shows that the runtime uses COM aggregation to provide the
IGameController and IGameControllerBatteryInfo implementation, while
expecting the custom controller to implement several other interfaces
depending on the type of provider used as a backend.
Signed-off-by: Rémi Bernon <rbernon@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>