hi uli! just a few ideas:
using SWBA would be an option, but maybe you could use the same DMA mechanism that is used now for beacons - but send a different frame. that would ensure the packet goes out in time, and the other end should not sync if it's not a beacon frame. you can specify the time it is getting sent out by using ath5k_hw_init_beacon(ah, nexttbtt, intval) in a 64sec window (the timer is 16bit in TU if i remember correctly). i'm not sure if it's possible to disable carrier sense at all, but the esiest way to avoid the problem is testing on an otherwise quiet 5GHz channel. please let me know about your progress - i'm eager to find a way to measure when the rx timestamp is taken exactly. bruno On Tuesday 19 February 2008 19:00:38 Ulrich Meis wrote: > Hi, > > For some experiments I'd like to send out frames at a precise point in > time - no matter what's happening on the channel. For example, in order > to find out how synchronized stations really are beacons cannot be used > because the TSF is constantly adjusted. Being able to send out frames at > a particular point in time would be helpful there. > > I'm thinking I'd have to set cwmin/cwmax to 0 on the queue and for the > most precise timing I should probably use the SWBA interrupt. > > Now the problem is how do I disable carrier sense? I remember a comment > from Luis about the RSSI threshold, but if I tune that, I'll probably > not be able to receive any frames anymore since they'd be treated as > noise, right? > > If you have any ideas, please enlighten me! > > Thanks, > > Uli > _______________________________________________ > ath5k-devel mailing list > [email protected] > https://lists.ath5k.org/mailman/listinfo/ath5k-devel _______________________________________________ ath5k-devel mailing list [email protected] https://lists.ath5k.org/mailman/listinfo/ath5k-devel
