Thanks. That's what I figured. In that case, does it make sense to strip out the incorrect MCS information from the radiotap headers of first n-1 subframes, and only let the final subframe carry the correct MCS infomration? It's less confusing IMO.
Regards, Jinghao On Fri, Oct 30, 2015 at 10:17 AM, Adrian Chadd <adr...@freebsd.org> wrote: > Hi, > > IIRC only the data in the last subframe descriptor is "valid". It's > valid for the whole A-MPDU. > > > > -a > > > On 29 October 2015 at 10:46, Jinghao Shi <jingh...@buffalo.edu> wrote: > > Hi, > > > > Recently I used an OpenWRT router (TP-LINK WDR3500, AR9340, ATH9K) to > sniff > > packets, and I found that the MCS information in the radiotap header is > > possible corrupted. > > > > In theory, all A-MPDU subframes should be sent in the same bit rate > > (MCS/BW/GI), since they have the same PPDU header. However, in the packet > > trace, I found the MCS in each A-MPDU subframe's radiotap header is > > different: they are in the same MCS index, but have different BW and GI > > flag. I was able to confirm that only the last subframe has the correct > MCS > > information by looking at the logs at the transmitter side. > > > > Here is a trace snippet to illustrate the problem: > > > > https://www.cloudshark.org/captures/2dad885fee3e (the pcap file is also > > attached). > > > > I was able to trace down to the function that set the MCS information in > > ath_rx_status structure. It is in > > ar9003_mac.c:ath9k_hw_process_rxdesc_edma() (or > > mac.c:ath9k_hw_rxprocdesc()?) It seems the corrupted MCS information is > > passed from the firmware. > > > > So my questions are: > > > > Is this a bug? > > If so, is this caused by the firmware? > > > > > > Thanks, > > Jinghao > > > > _______________________________________________ > > ath9k-devel mailing list > > ath9k-devel@lists.ath9k.org > > https://lists.ath9k.org/mailman/listinfo/ath9k-devel > > >
_______________________________________________ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
