It seems that only choose_devnum() was not ready to process more hub events at the same time. All should be fine if we take bus->usb_address0_mutex there. It will make sure that more devnums will not be chosen for the given bus and the related devices at the same time. Signed-off-by: Petr Mladek <pmladek@xxxxxxx> --- drivers/usb/core/hub.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c index 6c1a79a50576..69f9f1c1d9c1 100644 --- a/drivers/usb/core/hub.c +++ b/drivers/usb/core/hub.c @@ -2042,7 +2042,8 @@ static void choose_devnum(struct usb_device *udev) int devnum; struct usb_bus *bus = udev->bus; - /* If hub_wq ever becomes multithreaded, this will need a lock */ + /* be safe when more hub events are proceed in parallel */ + mutex_lock(&bus->usb_address0_mutex); if (udev->wusb) { devnum = udev->portnum + 1; BUG_ON(test_bit(devnum, bus->devmap.devicemap)); @@ -2060,6 +2061,7 @@ static void choose_devnum(struct usb_device *udev) set_bit(devnum, bus->devmap.devicemap); udev->devnum = devnum; } + mutex_unlock(&bus->usb_address0_mutex); } static void release_devnum(struct usb_device *udev) @@ -5162,11 +5164,8 @@ int usb_hub_init(void) * USB-PERSIST port handover. Otherwise it might see that a full-speed * device was gone before the EHCI controller had handed its port * over to the companion full-speed controller. - * - * Also we use ordered workqueue because the code is not ready - * for parallel execution of hub events, see choose_devnum(). */ - hub_wq = alloc_ordered_workqueue("usb_hub_wq", WQ_FREEZABLE); + hub_wq = alloc_workqueue("usb_hub_wq", WQ_FREEZABLE, 0); if (hub_wq) return 0; -- 1.8.4 -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html