My crazy idea at the moment is that since things can spend inordinate amounts of time in the txqueue in wireless in particular, that moving (tcp and packet) timestamping to where it enters the queue rather than hits the wireless device would make a difference. Unless that's already how it works. http://www.bufferbloat.net/issues/304#change-1186 The secondary fallout of that is an easier way to expire (probably rescheduled) packets that have spent too much time in queue. Eric had put up a (rfc) tc filter a while back that dumped a timestamp into the private block of the skb... but... I kind of like the idea of reusing the existing hwtstamps concept for the period "queue entrance to queue exit", as a queue is a 'device', IMHO. from linux/skbuff.h * hwtstamps can only be compared against other hwtstamps from * the same device. * * This structure is attached to packets as part of the * &skb_shared_info. Use skb_hwtstamps() to get a pointer. */ struct skb_shared_hwtstamps { ktime_t hwtstamp; ktime_t syststamp; }; Sign me, lost in a myriad of data paths on a sunday afternoon, wearing an asbestos suit -- Dave Täht SKYPE: davetaht http://www.bufferbloat.net -- 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
