On Sun, 8 Mar 2009, Roel Kluin wrote: > I think this is what you wanted? It compiles with defconfig, checkpatch doesn't > complain, nor sparse anymore, but please review anyways. One thing could be changed; otherwise it looks okay. > urb->dev->maxchild is an int, shuld it be converted to unsigned as well? One thing at a time! :-) You can submit another patch to change maxchild, if you want. Since its value is never bigger than 1000 and usually is considerably smaller than 32, the signedness doesn't matter. > transfer_buffer_length and actual_length have become unsigned, therefore some > additional conversion of local variables, function arguments and print > specifications is desired. A test for a negative urb->transfer_buffer_length has > become obsolete as well. > usb_internal_control_msg() returns both error values and a length. To ensure > that it does not return an unsigned length that is converted into a negative, > the returned length is limited to INT_MAX. Instead of doing it this way, it would be simpler and more robust to change the test for a negative urb->transfer_buffer_length in submit_urb() into a test for urb->transfer_buffer_length > INT_MAX. (Strictly speaking, this shouldn't be INT_MAX -- it should be the largest possible 32-bit signed integer. But until 64-bit integers become the default we will be safe...) Then we would know that the length returned by usb_internal_control_msg() could never be too large. 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
