On Fri, 7 Aug 2009, [UTF-8] í??ì? shin hong wrote: > Hi. I am reporting a suspected race but at usb_serial_probe() > in /drivers/usb/serial/usb-serial.c of Linux 2.6.30.4. > Since I do not have much background, please examine the report > and code and let me know your opinions. > > usb_serial_probe() invokes create_serial() while it holds Big Kernel Lock. > However, create_serial() invokes kmalloc() so that BKL might be > released and re-taken due to its releasing-on-block semantics. > For this reason, usb_serial_probe() may have race condition problem. There is no race. The BKL is not used for serializing the call to create_serial(). In fact, the BKL probably isn't needed in usb_serial_probe() at all, and neither is the call to try_module_get(). Alan Stern -- 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
