Hi, Lukas and Nick > I'm sending my own patch in a hope it will finally make it through. Thanks. I did brief test with the patch and those test results show the throughput issues (receiving throughput of ath5k) was fixed. It's so great. Thanks a lot, Lukas. (but with my cases, the problem was for 11g not for 11a )
While the receiving issue was solved, throughput of xmit from ath5k platform went down to 9Mbps by some reason. (my test problem ?) <sample config> PC(iperf client) --> Ubiquity RS-Pro(ath5k) --> Ubiquity RS-Pro(ath5k) --> PC(iperf server) I think this has nothing to do with the patch ( should not) , and might be caused by my some mistakes. I will check this again tomorrow. anyway, thanks a lot Lukas. takayuki Kaiso https://lists.ath5k.org/pipermail/ath5k-devel/2009-September/002730.html > < test scenario > > - iperf (UDP) between the PC(A) connected to Soekris 4826 with > Madwifi-0.9.4 and > another PC(B) connected to Ubiquity RS-Pro with Ath5k > (compat-wireless-2009-0820) > - IBSS ad-hoc wireless between Soekris 4826 and Ubiquity RS-Pro > > < result > > 1. when iperf packets go from PC(B) to PC(A), iperf UDP receiver > showed up to 16Mbps > and loss is 0.0% --> looks so great . > > 2. change the flow direction and when iperf packets go from PC(A) to > PC(B), > iperf UDP receiver showed following results. (in this case ath5 > is receiver side ) > (iperf -c x.x.x.x -u -b 12M -t 30 ) > > - 90% loss (when sender(madwifi) set as "rate 54M fixed") > - 90% loss (when sender(madwifi) set as "rate 48M fixed") > - 1% ~ 3% loss (when sender(madwifi) set as "rate 36M fixed") > - 0.2% ~ 0.7% loss (when sender(madwifi) set as "rate 24M fixed") > - 5%~20% (when sender(madwifi) set as "rate auto") > > On 16.11.2009 15:06 Luka's( Turek wrote: > >> In my first tests I found a problem which is basically a showstopper for >> us: performance in 802.11a mode is very bad. It's some problem with >> reception at higher rates, at 48Mbit there's 30% to 70% packetloss, at >> 54Mbit even more. >> > > I've traced the bug to I/Q calibration. A bit-mask is applied to a signed > value, so the result is a nonsense, which breaks packet reception. A patch > fixing it was already sent, but probably got lost somewhere: > https://lists.ath5k.org/pipermail/ath5k-devel/2009-September/002730.html > > I'm sending my own patch in a hope it will finally make it through. > > Lukas Turek > > ------------------------------------------------------------------------ > > _______________________________________________ > ath5k-devel mailing list > ath5k-devel@lists.ath5k.org > https://lists.ath5k.org/mailman/listinfo/ath5k-devel > -- ***************************************** 株式会社 シンクチューブ 海藻 敬之 tka...@thinktube.com 〒658-0032 神戸市東灘区向洋町中6−9 KFMビル 4E-10 Phone: 078-857-8390 Fax: 078-857-8389 www.thinktube.com
_______________________________________________ ath5k-devel mailing list ath5k-devel@lists.ath5k.org https://lists.ath5k.org/mailman/listinfo/ath5k-devel