Hi,
Felipe Balbi wrote:
Hi,
Thinh Nguyen Thinh.Nguyen@synopsys.com writes:
If the gadget had already started, don't try to start again. Otherwise, we may request the same threaded irq with the same dev_id, it will mess up the interrupt freeing logic. This can happen if a user tries to trigger a soft-connect from soft_connect sysfs multiple times. Check to make sure that the gadget had started before proceeding to request threaded irq. Fix this by checking if there's bounded gadget driver.
Looks like this should be fixed at the framework level, otherwise we will have to patch every single UDC. After that is done, we can remove the dwc->gadget_driver check from here.
Sure. We can do that.
Thanks, Thinh