From mboxrd@z Thu Jan 1 00:00:00 1970 From: Luis R. Rodriguez Date: Mon, 18 May 2009 10:33:31 -0700 Subject: [ath9k-devel] Linksys WPC300N Wireless-N Notebook Adapter In-Reply-To: <98786.73657.qm@web57902.mail.re3.yahoo.com> References: <98786.73657.qm@web57902.mail.re3.yahoo.com> Message-ID: <20090518173331.GI20637@tesla> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ath9k-devel@lists.ath9k.org On Mon, May 18, 2009 at 09:36:43AM -0700, hong zhang wrote: > > Luis, > > 1. Does ath9k Tx/Rx buffer packets for power save purpose? You mean in AP mode? If you are on a kernel >= 2.6.29 yes. That is even if you are using compat-wireless you need at least 2.6.29 still. > 2. Do TX power and signal level display go same path to mac80211 > through rate control interface? I don't understand your question. > I think iw dev wlan0 station should > display Tx power and signal level as well. 'iw dev wlan0 station dump' does display signal. I get: mcgrof at tesla ~ $ iw dev wlan2 station dump Station 00:0b:85:5b:8d:6c (on wlan2) inactive time: 22628 ms rx bytes: 3042422 rx packets: 13165 tx bytes: 177786 tx packets: 947 signal: -30 dBm tx bitrate: 54.0 MBit/s Keep in mind wireless extensions added a 'quality' term which unfortunately was used differently between drivers. What we add to nl80211 should remain consistant accross drivers and it should be defined properly. If there is some enhancement to this reporting which you can think of please feel free to address this on linux-wireless with your recommendation. The next best thing we have to report signal would be through using the RCPI concept [1], but note that it specifically defines it as "the received RF power in the channel measured over the entire received frame or by other equivalent means which meet the specified accuracy". The last part is a little vague, that of "or by other equivalent means which meet the specified accuracy" and its questionable whether or not all devices support this. Anyway this could just be used by userspace and nl80211 should not have to know about this except maybe whether or not the device claims to support this accuracy terminology, I am also not sure what the status of 802.11k is which the group working on this, so it may be premature to just add it to nl80211/cfg80211. [1] http://www.spinics.net/lists/netdev/msg12028.html Patches are welcomed to add tx power displaying, not sure why we didn't add that. Luis > > > ---henry > > --- On Mon, 5/18/09, Luis R. Rodriguez wrote: > > > From: Luis R. Rodriguez > > Subject: Re: [ath9k-devel] Linksys WPC300N Wireless-N Notebook Adapter > > To: "hong zhang" > > Cc: "Luis R. Rodriguez" , "ath9k-devel at lists.ath9k.org" > > Date: Monday, May 18, 2009, 10:41 AM > > On Sat, May 16, 2009 at 10:51:57PM > > -0700, hong zhang wrote: > > > > > > Luis, > > > > > > Have you experience XJperf (windows graphic version > > iperf) to measure UDP/TCP throughput? I think it does not > > work properly because UDP throughput is always 1Mbites/s. > > > > No, thanfully I don't have a windows box anywhere near my > > home or office, > > that's just a serious security risk. But I see xjperf runs > > in Linux, I'll give > > it a shot some time. > > > > > Also, does "iw dev ath0 station dump" display bit rate > > (I have not tested my WPC300N yet)? > > > > Right now ath9k does not report the proper 11n rate, that > > work needs a > > decent rewrite on rate handling in ath9k. For now please > > use the ath9k > > debugfs file, rcstat. That'll show you counters for every > > MCS rate. > > > > Luis > > > > >