On Tue, Feb 04, 2025 at 11:36:42PM +0000, Roy Luo wrote:
> device_del() can lead to new work being scheduled in gadget->work
> workqueue. This is observed, for example, with the dwc3 driver with the
> following call stack:
> device_del()
> gadget_unbind_driver()
> usb_gadget_disconnect_locked()
> dwc3_gadget_pullup()
> dwc3_gadget_soft_disconnect()
> usb_gadget_set_state()
> schedule_work(&gadget->work)
>
> Move flush_work() after device_del() to ensure the workqueue is cleaned
> up.
>
> Fixes: 5702f75375aa9 ("usb: gadget: udc-core: move sysfs_notify() to a workqueue")
> Cc: <stable@xxxxxxxxxxxxxxx>
> Signed-off-by: Roy Luo <royluo@xxxxxxxxxx>
> ---
Reviewed-by: Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>
> drivers/usb/gadget/udc/core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/usb/gadget/udc/core.c b/drivers/usb/gadget/udc/core.c
> index a6f46364be65..4b3d5075621a 100644
> --- a/drivers/usb/gadget/udc/core.c
> +++ b/drivers/usb/gadget/udc/core.c
> @@ -1543,8 +1543,8 @@ void usb_del_gadget(struct usb_gadget *gadget)
>
> kobject_uevent(&udc->dev.kobj, KOBJ_REMOVE);
> sysfs_remove_link(&udc->dev.kobj, "gadget");
> - flush_work(&gadget->work);
> device_del(&gadget->dev);
> + flush_work(&gadget->work);
> ida_free(&gadget_id_numbers, gadget->id_number);
> cancel_work_sync(&udc->vbus_work);
> device_unregister(&udc->dev);
>
> base-commit: f286757b644c226b6b31779da95a4fa7ab245ef5
> --
> 2.48.1.362.g079036d154-goog
>
