usb: renesas: silence uninitialized variable report in usbhsg_recip_r…

…un_handle()

In drivers/usb/renesas_usbhs/mod_gadget.c::usbhsg_recip_run_handle()
the Coverity Prevent checker currently flags a warning about possibly
uninitialized use of 'ret' i usbhsg_recip_run_handle(). It does this
since it assumes we take one of the non-default branches in the switch
and then subsequently take the false branch in the 'if (func)' case
below. This exact scenario will never happen, but Coverity can't see
that for some reason. This patch initializes 'ret' to '0' when it is
declared which should shut up this report and won't really hurt - so
why not? At least then it's clear that 'ret' is always initialized..

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Felipe Balbi <balbi@ti.com>

drivers/usb/renesas_usbhs/mod_gadget.c

@@ -425,7 +425,7 @@ static int usbhsg_recip_run_handle(struct usbhs_priv *priv,
 	struct usbhs_pipe *pipe;
 	int recip = ctrl->bRequestType & USB_RECIP_MASK;
 	int nth = le16_to_cpu(ctrl->wIndex) & USB_ENDPOINT_NUMBER_MASK;
-	int ret;
+	int ret = 0;
 	int (*func)(struct usbhs_priv *priv, struct usbhsg_uep *uep,
 		    struct usb_ctrlrequest *ctrl);
 	char *msg;