6.6-stable review patch. If anyone has any objections, please let me know. ------------------ From: John Keeping <jkeeping@xxxxxxxxxxxxxxxxx> commit df9158826b00e53f42c67d62c887a84490d80a0a upstream. Registering a gadget driver is expected to complete synchronously and immediately after calling driver_register() this function checks that the driver has bound so as to return an error. Set PROBE_FORCE_SYNCHRONOUS to ensure this is the case even when asynchronous probing is set as the default. Fixes: fc274c1e99731 ("USB: gadget: Add a new bus for gadgets") Cc: stable@xxxxxxxxxxxxxxx Signed-off-by: John Keeping <jkeeping@xxxxxxxxxxxxxxxxx> Link: https://lore.kernel.org/r/20240913102325.2826261-1-jkeeping@xxxxxxxxxxxxxxxxx Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/usb/gadget/udc/core.c | 1 + 1 file changed, 1 insertion(+) --- a/drivers/usb/gadget/udc/core.c +++ b/drivers/usb/gadget/udc/core.c @@ -1671,6 +1671,7 @@ int usb_gadget_register_driver_owner(str driver->driver.bus = &gadget_bus_type; driver->driver.owner = owner; driver->driver.mod_name = mod_name; + driver->driver.probe_type = PROBE_FORCE_SYNCHRONOUS; ret = driver_register(&driver->driver); if (ret) { pr_warn("%s: driver registration failed: %d\n",