Date: Tue, 30 Jul 2013 10:19:26 -0700 From: Adrian Chadd <adrian@freebsd.org> To: Cedric GROSS <cg@cgross.info> Cc: freebsd-wireless@freebsd.org Subject: Re: Wifi rates Message-ID: <CAJ-Vmon0fZc6UieB7P2D5OOMDkZHP2PYJvFc2D8Sk7K-gmxWiQ@mail.gmail.com> In-Reply-To: <000c01ce8d07$d10a6550$731f2ff0$@info> References: <00b101ce8c8b$3db5f200$b921d600$@info> <CAJ-Vmo=y609_w2dMyLkgOpd24poRJKqn=LPW%2BQGi7z346xTZrQ@mail.gmail.com> <000c01ce8d07$d10a6550$731f2ff0$@info>
next in thread | previous in thread | raw e-mail | index | archive | help
On 30 July 2013 02:33, Cedric GROSS <cg@cgross.info> wrote: > [root@Test]/root#ifconfig -v wlan0 list sta > ADDR AID CHAN RATE RSSI IDLE TXSEQ RXSEQ CAPS FLAG > 00:24:d4:97:80:20 2 5 6M 18.5 0 1162 58480 EP AQEHTRS > SSID<WLenine> RATES<B2,B4,B11, > B22,12,18,24,36> DSPARMS<5> ERP<0x4> XRATES<48,72,96,108> RSN<v1 mc:AES-CCMP > uc:AES-CCMP km:8021X-PS > K> HTCAP<cap 0x6c param 0x3 mcsset[0-23,32] extcap 0x0 txbf 0x0 antenna 0x0> > HTINFO<ctl 5, 0,7,0,0 b > asicmcs[]> WME<qosinfo 0x0 BE[aifsn 3 cwmin 4 cwmax 10 txop 0] BK[aifsn 7 > cwmin 4 cwmax 10 txop 0] V > O[aifsn 2 cwmin 3 cwmax 4 txop 94] VI[aifsn 2 cwmin 2 cwmax 3 txop 47]> Right, there's HT elements in there, so it looks likely you properly negotiated HT. > And the amrr output : > Test kernel: wlan0: [00:24:d4:97:80:20] amrr_node_init: non-11n node > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: nrates=0, initial rate 0 That's the initial "I don't know what you are yet" selection. > Test kernel: wlan0: [00:24:d4:97:80:20] amrr_node_init: 11n node > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: nrates=16, initial rate 128 Here it negotiated 11n. The initial rate is 0x80, so MCS 0. > Test kernel: wlan0: [00:24:d4:97:80:20] recv auth frame with algorithm 0 seq > 2 > Test kernel: wlan0: [00:24:d4:97:80:20] amrr_node_init: 11n node > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: nrates=16, initial rate 128 > Test kernel: wlan0: link state changed to UP > Test dhclient: New IP Address (wlan0): 192.168.1.20 > Test dhclient: New Subnet Mask (wlan0): 255.255.255.0 > Test dhclient: New Broadcast Address (wlan0): 192.168.1.255 > Test dhclient: New Routers (wlan0): 192.168.1.254 > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: current rate 15, txcnt=12, > retrycnt=144 The rate index starts at MCS0. > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR decreasing rate 14 (txcnt=12 > retrycnt=144) > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: current rate 14, txcnt=12, > retrycnt=192 > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR decreasing rate 13 (txcnt=12 > retrycnt=192) > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: current rate 13, txcnt=11, > retrycnt=176 > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR decreasing rate 12 (txcnt=11 > retrycnt=176) > ... > And so on until > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: current rate 0, txcnt=11, > retrycnt=176 > Test last message repeated 2 times > Test kernel: wlan0: [00:24:d4:97:80:20] AMRR: current rate 0, txcnt=12, > retrycnt=192 And since there's errors at all transmission rates (look at the retrycnt), it eventually drops you down to 1mb. > What is 'nrates' and 'rate' in start of output ? Is nrates an index in rate > table ? Yes, look in ieee80211_phy.c . > After the dhclient, I started a ftp file transfer. As you can see rate > decrease to 0 very quickly. > Is the 'current rate' same as initial nrates value ? Nope. The debugging isn't very clear and I should fix it. But it's starting in the 11n rate table at MCS0 and it's going down very quickly into the non-11n rates. > I've found this for MCS : > http://www.digitalairwireless.com/wireless-blog/recent/demystifying-modulati > on-and-coding-scheme-index-values.html MCS is the 11n encoding scheme. So you're seeing significant transmit retries. That's worth investigating. (We still have to merge more code into -HEAD though.) -adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-Vmon0fZc6UieB7P2D5OOMDkZHP2PYJvFc2D8Sk7K-gmxWiQ>