4.4-stable review patch. If anyone has any objections, please let me know.
------------------
[ Upstream commit 1ff5fa3c6732f08e01ae12f12286d4728c9e4d86 ]
After initially connecting a wired Xbox 360 controller or sending it a command to change LEDs, a status/response packet is interpreted as controller input. This causes the state of buttons represented in byte 2 of the controller data packet to be incorrect until the next valid input packet. Wireless Xbox 360 controllers are not affected.
Writing a new value to the LED device while holding the Start button and running jstest is sufficient to reproduce this bug. An event will come through with the Start button released.
Xboxdrv also won't attempt to read controller input from a packet where byte 0 is non-zero. It also checks that byte 1 is 0x14, but that value differs between wired and wireless controllers and this code is shared by both. I think just checking byte 0 is enough to eliminate unwanted packets.
The following are some examples of 3-byte status packets I saw: 01 03 02 02 03 00 03 03 03 08 03 00
Signed-off-by: Cameron Gutman aicommander@gmail.com Signed-off-by: Pavel Rojtberg rojtberg@gmail.com Cc: stable@vger.kernel.org Signed-off-by: Dmitry Torokhov dmitry.torokhov@gmail.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/input/joystick/xpad.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/drivers/input/joystick/xpad.c b/drivers/input/joystick/xpad.c index 5f47e062cada..a49be8bc5819 100644 --- a/drivers/input/joystick/xpad.c +++ b/drivers/input/joystick/xpad.c @@ -459,6 +459,10 @@ static void xpad_process_packet(struct usb_xpad *xpad, u16 cmd, unsigned char *d static void xpad360_process_packet(struct usb_xpad *xpad, struct input_dev *dev, u16 cmd, unsigned char *data) { + /* valid pad data */ + if (data[0] != 0x00) + return; + /* digital pad */ if (xpad->mapping & MAP_DPAD_TO_BUTTONS) { /* dpad as buttons (left, right, up, down) */