Well timed removal of USB cable leaves g_serial unusable. * It is possible that the g_serial driver can recieve a request to handle SET_LINE_CODING in acm_setup() when the end-point is invalid. Further more, acm_setup() will setup a completition handler for the request which results in the request becoming stuck. Not certain what is racing where, but replacing the completion request with a memcpy() operation appears to avoid the problem completely.
Signed-off-by: Mark Ferrell <[email protected]> --- drivers/usb/gadget/f_acm.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/usb/gadget/f_acm.c b/drivers/usb/gadget/f_acm.c index 4b7e33e..6cd2b69 100644 --- a/drivers/usb/gadget/f_acm.c +++ b/drivers/usb/gadget/f_acm.c @@ -363,9 +363,11 @@ static int acm_setup(struct usb_function *f, const struct usb_ctrlrequest *ctrl) || w_index != acm->ctrl_id) goto invalid; - value = w_length; - cdev->gadget->ep0->driver_data = acm; - req->complete = acm_complete_set_line_coding; + if (req->actual == sizeof(acm->port_line_coding)) { + value = min_t(unsigned, w_length, + sizeof(struct usb_cdc_line_coding)); + memcpy(&acm->port_line_coding, req->buf, value); + } break; /* GET_LINE_CODING ... return what host sent, or initial value */ -- tg: (37af71c..) upstream/usb/gadget/g_serial/bad_request (depends on: master) -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
