Race could happen with QOS : let's consider queue 0 low prio and queue 1 high prio - queue 1 send packet lot's of packet (iv32 = 0, iv32=1, iv32=2) - queue 0 send a packet with iv32=1 and iv16=0xffff (or near wrap) If queue 0 packet with iv32=1 is received after queue 1 with iv32=2, we will - send tkip key for iv32=2 to hardware - send tkip key for iv32=1 to hardware and then we should wait wrap to iv32=3 to have a chance to send correct key to hardware. With this patch, we only send tkip key to hardware on the first wrap of a queue. With the previous example, we will only send tkip key for iv32=2 to hardware. Gregor -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
