"len" comes from the USB transfer and it's probably correct. The thing is that we already have similar checks like: if (data[i] >= serial->num_ports) { So adding a sanity test here matches the rest of the code and is a good idea. Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx> --- v2: No change. Rebased. diff --git a/drivers/usb/serial/keyspan.c b/drivers/usb/serial/keyspan.c index 3d303f0..eb30d7b 100644 --- a/drivers/usb/serial/keyspan.c +++ b/drivers/usb/serial/keyspan.c @@ -727,14 +727,15 @@ static void usa49wg_indat_callback(struct urb *urb) if ((data[i] & 0x80) == 0) { /* no error on any byte */ i++; - for (x = 1; x < len ; ++x) + for (x = 1; x < len && i < urb->actual_length; ++x) tty_insert_flip_char(&port->port, data[i++], 0); } else { /* * some bytes had errors, every byte has status */ - for (x = 0; x + 1 < len; x += 2) { + for (x = 0; x + 1 < len && + i + 1 < urb->actual_length; x += 2) { int stat = data[i], flag = 0; if (stat & RXERROR_OVERRUN) -- To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html