Hi Alan, On 12/06/2016 09:21 PM, Alan Stern wrote: > Krzysztof: > > Your patch doesn't address a bug that has been present in > usb_add_gadget_udc_release() for some time. If the bind callback fails > with any error other than -EPROBE_DEFER, the driver remains unbound to > the gadget but it is not added back on to the > gadget_driver_pending_list. As a result, when the driver is > unregistered, the list_del() call will cause problems. >
You are absolutely right. I have never claimed that I fixed anything more than described in commit msg;) > There's a good reason not to add the driver back on to the list: If > it's not working then we don't want to keep trying to probe it over and > over again. I think the best thing to do would be > list_del_init(&driver->pending), so that at least unregistration won't > cause an oops. > Sounds reasonable. I will submit a patch for this. > There's another thing. If the bind callback fails, why not continue > the loop in check_pending_gadget_drivers()? Maybe another pending > driver will succeed. (Of course, to do this would require using > list_for_each_entry_safe instead of list_for_each_entry.) Yah. True. That's how probing drivers for devices usually works. I'll submit a patch for this. @Felipe Please ignore this patch. I will post it together with fixes suggested by Alan in a series so they can be easily applied without ordering problems. Best regards, -- Krzysztof Opasiak Samsung R&D Institute Poland Samsung Electronics -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
