1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Wireless Measurements from WL

Discussion in 'Tomato Firmware' started by Planiwa, Mar 12, 2012.

  1. Planiwa

    Planiwa LI Guru Member

    The Tomato GUI pages "Device List" and "Wireless Survey", as well as "Overview" show various statistics about the Wireless network and environment.

    Here we will explore how to obtain that information from the wl command, so that we can use it for logging. diagnosing, and other processing, such as determining the effects of changes in the network.

    I will share some of my own findings, and I hope that others will contribute as well.

    Here are some esential wl commands :
    Code:
    wl noise
    wl rssi $MAC
    wl sta_info $MAC
     
    wl scan
        sleep 5
     
    wl channels
    wl channel
     
    wl scanresults
    
    The wl results can be reformatted for readability:
    Code:
    Noise: -93
     
    Wifi clients detail report:
    RSSI: 14 -35 85 -45 18 -49 5B -51 A7 -61
                                                                  TXP TXF RXUP RXMBP TXR RXR
    5B idle 0 seconds in network 2696 seconds AUTH ASSO AUTHO N_CAP 49289 0 37722 574 5 78
    85 idle 4 seconds in network 3862 seconds AUTH ASSO AUTHO PS N_CAP 250208 47 145217 1586 58 1
    18 idle 47 seconds in network 5474 seconds AUTH ASSO AUTHO _____ 194957 266 111404 1057 1 1
    A7 idle 0 seconds in network 8841 seconds AUTH ASSO AUTHO BRCM N_CAP 291603 1 378336 271 5 6
    14 idle 0 seconds in network 32471 seconds AUTH ASSO AUTHO _____ 409112 104 241959 744 36 24 
    Here is a Wireless Survey:
    Code:
        Channels: 1 2 3 4 5 6 7 8 9 10 11
    Interference: 0 0 0 0 0 0 0 0 0 0 0  
    .....                                
    Current mac channel 1 target channel 1                                             
                                          
    Wireless network scan results:        
                                          
     -53 dBm        noise: -67 dBm  Channel: 10             N: SGI20 SGI40  "The Trial"
     -60 dBm        noise: -91 dBm  Channel: 11             N: SGI40        "rotorouter"
     -61 dBm        noise: -70 dBm  Channel: 6              N: SGI20 SGI40  "linksys1090"
     -61 dBm        noise: -91 dBm  Channel: 11             N:      "BELL427"            
     -62 dBm        noise: -85 dBm  Channel: 1              N: SGI20        "Brasil"     
     -64 dBm        noise: -94 dBm  Channel: 1              N: SGI20        "Brasil"     
     -70 dBm        noise: -70 dBm  Channel: 4              N: 40Mhz SGI20 SGI40    "HELEN-PC_Network"
     -72 dBm        noise: -82 dBm  Channel: 6      "qef.com"                                         
     -77 dBm        noise: -85 dBm  Channel: 1      "Brunswick0114"                                   
     -83 dBm        noise: -85 dBm  Channel: 1      "micasa"                                          
    
     
  2. mvsgeek

    mvsgeek Addicted to LI Member

    I'd like to know how the "Interference Level" on the Overview page is derived. Is it calculated from other values like noise and rssi? I've scoured various wl syntax pages without success. Also, your Wireless Survey above shows interference per channel, but I don't see that line when I run the "wl channels" command :

    /tmp/home/root# wl channels
    1 2 3 4 5 6 7 8 9 10 11
    What am I missing?
     
  3. Planiwa

    Planiwa LI Guru Member

    The Overview pages probably uses the current channel:

    wl chanim_state $(set -- $(wl channel);echo $8)


    For per-channel interference:

    for i in $(wl channels);do echo -n "$(wl chanim_state $i) ";done;echo


    wl channels # only shows the available channels. It serves as a header in the display.

    BTW, one could calculate quality in the above scan report like this:

    Code:
    :; :; grep noise /tmp/wr.html |awk '{print "Quality: ", $1-$4, $0}'|sort -nr -k2
    Quality:  32  -60 dBmnoise: -92 dBmChannel: 4N: 40Mhz SGI20 SGI40"HELEN-PC_Network"
    Quality:  30  -62 dBmnoise: -92 dBmChannel: 6N: SGI20 SGI40"linksys1090"
    Quality:  25  -67 dBmnoise: -92 dBmChannel: 6"qef.com"
    Quality:  24  -49 dBmnoise: -73 dBmChannel: 10N: SGI20 SGI40"The Trial"
    Quality:  23  -59 dBmnoise: -82 dBmChannel: 1N: SGI20"Brasil"
    Quality:  20  -56 dBmnoise: -76 dBmChannel: 1N: SGI20"Brasil"
    Quality:  10  -78 dBmnoise: -88 dBmChannel: 6"linksys252"
    Quality:  5  -77 dBmnoise: -82 dBmChannel: 1"goya"
    Quality:  5  -77 dBmnoise: -82 dBmChannel: 1"Brunswick0114"
    Quality:  1  -81 dBmnoise: -82 dBmChannel: 1"Apple Network ab369f"
    Quality:  -1  -83 dBmnoise: -82 dBmChannel: 1"wooligan"
    :;
    
    Note that "The Trial" has the strongest RSSI, but also much noise, so that it ranks 4th in quality.
     
  4. mvsgeek

    mvsgeek Addicted to LI Member

    Thank you! None of the wl reference pages I looked at had the "chanim_state" option, and it's exactly what I was looking for. It works for Toastman build 7483, but not for 7633 or Vanilla Tomato 1.28. No big deal, it's interference at the main router (7483) that interests me.
     
  5. Planiwa

    Planiwa LI Guru Member

    Let's look at a scan result:
    Code:
    Wireless network scan results:
    Qual: 34  -49 dBmnoise: -83 dBmChannel: 10N: SGI20 SGI40"The Trial"
    Qual: 29  -63 dBmnoise: -92 dBmChannel: 4N: 40Mhz SGI20 SGI40"HELEN-PC_Network"
    Qual: 28  -54 dBmnoise: -82 dBmChannel: 10N: SGI20 SGI40""
    Qual: 28  -52 dBmnoise: -80 dBmChannel: 11N: SGI40"rotorouter"
    Qual: 27  -65 dBmnoise: -92 dBmChannel: 6N: SGI20 SGI40"linksys1090"
    Qual: 20  -56 dBmnoise: -76 dBmChannel: 1N: SGI20"Brasil"
    Qual: 18  -58 dBmnoise: -76 dBmChannel: 1N: SGI20"Brasil"
    Qual: 17  -63 dBmnoise: -80 dBmChannel: 11N:"BELL427"
    
    If these were the RSSI and Noise values for Associated Stations, the Quality ranking would make sense. We want low noise, high strength.
    But it seems to me that when it comes to Interference, we want low noise and low strength.
    So, instead of "quality", we should calculate an "interference" value that reflects that.
    Perhaps, just add Noise and RSSI?:
    Interference = 100 + Noise + RSSI, so that we would get:
    Code:
    Interference: 68 Qual: 34  -49 dBm noise: -83 dBm Channel: 10  N: SGI20 SGI40  "The Trial" 
    Interference: 68 Qual: 28  -52 dBm noise: -80 dBm Channel: 11  N: SGI40  "rotorouter" 
    Interference: 68 Qual: 20  -56 dBm noise: -76 dBm  Channel: 1  N: SGI20  "Brasil" 
    Interference: 66 Qual: 18  -58 dBm noise: -76 dBm  Channel: 1  N: SGI20  "Brasil" 
    Interference: 64 Qual: 28  -54 dBm noise: -82 dBm  Channel: 10  N: SGI20 SGI40  "" 
    Interference: 57 Qual: 17  -63 dBm noise: -80 dBm Channel: 11  N:  "BELL427" 
    Interference: 45 Qual: 29  -63 dBm noise: -92 dBm Channel: 4  N: 40Mhz SGI20 SGI40  "HELEN-PC_Network" 
    Interference: 43 Qual: 27  -65 dBm noise: -92 dBm  Channel: 6  N: SGI20 SGI40  "linksys1090" 
    
    Here is a current scan:
    Code:
    Interference: 92 Qual: 34  -37 dBmnoise: -71 dBmChannel: 1N: SGI20"Brasil"
    Interference: 72 Qual: 14  -57 dBmnoise: -71 dBmChannel: 1N: SGI20"Brasil"
    Interference: 70 Qual: 36  -47 dBmnoise: -83 dBmChannel: 10N: SGI20 SGI40"The Trial"
    Interference: 68 Qual: 10  -61 dBmnoise: -71 dBmChannel: 4N: 40Mhz SGI20 SGI40"HELEN-PC_Network"
    Interference: 65 Qual: 29  -53 dBmnoise: -82 dBmChannel: 11N: SGI40"rotorouter"
    Interference: 56 Qual: 20  -62 dBmnoise: -82 dBmChannel: 11N:"BELL427"
    Interference: 47 Qual: 31  -61 dBmnoise: -92 dBmChannel: 6N: SGI20 SGI40"linksys1090"
    Interference: 31 Qual: 15  -77 dBmnoise: -92 dBmChannel: 10"unclefester"
    
    (The nasty posting software removes tabs.)

    Are there any views on this?
    How long should one wait for a scan?
    Does a scan disrupt the wireless network?
    Why is no one concerned that Tomato's "hardware default" power is 1500mW?
    Or that wl's "set tx power to default" does nothing?

    What exactly does "wl phy_rxiqest" mean?
     
  6. Planiwa

    Planiwa LI Guru Member

    So, monitoring the wireless network remotely is difficult.
    If there is a responsive wireless host, we can get data. But if the wireless hosts are unable to connect, there is no data. And absence of data might mean that everyone is outside enjoying the warm weather.

    (Worst case is when someone calls to say they can get a wireless connection, but can't get to the Internet, and it turns out that their PC has connected to a strange access point that happens to be someone's print server. ;-))

    So, how can I get a host that I can manage on the wireless network? Here's an idea:

    Set up an old WRT in WIFI client mode, and also connect it via ether to the router, so that I can still manage it, even when the wifi fails. Hm. Now to work out the datails. :)
     
  7. Planiwa

    Planiwa LI Guru Member

    Here is an interesting data point:
    Code:
        Channels: 1 2 3 4 5 6 7 8 9 10 11
    Interference: 0 0 0 0 0 0 0 1 1 0 1
    .....
    Current mac channel 1 target channel 1
     
    Wireless network scan results:
    Interference: 63 Qual:  9  -64 dBmnoise: -73 dBmChannel: 4 N: 40Mhz SGI20 SGI40"HELEN-PC_Network"
    Interference: 63 Qual: 21  -58 dBmnoise: -79 dBmChannel: 1 N: SGI20"Brasil"
    Interference: 62 Qual: 16  -61 dBmnoise: -77 dBmChannel: 1 N: SGI20"Brasil"
    Interference: 54 Qual: 38  -54 dBmnoise: -92 dBmChannel: 10 N: SGI20 SGI40"The Trial"
    Interference: 50 Qual: 34  -58 dBmnoise: -92 dBmChannel: 6 N: SGI20 SGI40"linksys1090"
    Interference: 28 Qual: 12  -80 dBmnoise: -92 dBmChannel: 1 "Brunswick0114"
    
    wl says "severe interference" on channels 8, 9, and 11.
    wl also says that it sees no radios on those channels.

    Hm.
     
  8. Planiwa

    Planiwa LI Guru Member

    About wl and nvram -- it appears that in NVRAM,

    wl contains the default settings
    wl0 contains the active settings

    Note that what what wl calls interference, nvram calls mitigation. Thus:
    Code:
      wl mitigation
    wl: Unsupported
      wl interference
    Auto Wireless LAN Interference mitigation is enabled and not active, and noise reduction is enabled. (mode 4)
      nvram find interference
      nvram find mitigation
    wl0_mitigation=3
    wl_mitigation=0
      nvram set wl0_mitigation=4
      nvram find mitigation
    wl0_mitigation=4
    wl_mitigation=0
    
     
  9. Planiwa

    Planiwa LI Guru Member

    Is it possible that no one is interested in monitoring the levels of signal strength, noise, and interference over time? I've started recording the following data so that information can be derived from it later:

    TIME NOISE STATION-RSSI ... INTERFERED-CHANNEL ...

    Code:
    03201200 -76 A2-51
    03201201 -80 A2-51 4 5 6 11
    03201202 -80 A2-50 4 5 6 11
    03201203 -80 A2-51 4 5 6 11
    03201204 -80 A2-51 4 5 6 11
    03201205 -79 A2-52 6
    03201206 -83 A2-52 5
    03201207 -80 A2-50 5
    03201208 -81 A2-48 5
    03201209 -81 A2-49 5
    03201210 -81 A2-56
    03201211 -79 A2-55
    03201212 -81 A2-51
    03201213 -73 A2-49
    03201214 -74 A2-53
    03201215 -74 A2-52
    03201216 -76 A2-49 6
    03201217 -73 A2-57 6
    03201218 -73 A2-55 6
    03201219 -74 A2-55 6
    03201220 -74 A2-57 6
    03201221 -74 A2-56
    03201222 -75 14-47 A2-55
    03201223 -73 14-46 A2-55
    03201224 -74 5B-53 14-44 A2-55
    03201225 -77 5B-54 14-44 A2-55 5
    03201226 -76 5B-54 14-42 A2-56 5
    03201227 -80 5B-59 14-47 A2-56 5
    03201228 -74 5B-54 14-51 A2-56 5
    03201229 -75 5B-56 14-54 A2-56 5
    03201230 -79 5B-55 14-45 A2-58
    
    It seems that the noise level can get pretty high (up to -69dBm).

    The extreme values over the past 12 hours have been:

    Code:
      sort -k2 wlmon |head
    03200634 -69
    03200635 -72
    03200704 -72
    03200712 -72
    03200829 -72
    03200840 -72
    03200808 -72  5 6
    03200809 -72  5 6
    03200955 -72 A2-47 6
    03200945 -72 A2-48 6
      sort -k2 wlmon |tail
    03200332 -94 18-48 71-56 6
    03200522 -94 18-49 4
    03200341 -94 18-49 71-61
    03200329 -94 18-49 71-63 5
    03200334 -94 18-51 71-62 6
    03200423 -94 18-56 71-54
    03192318 -94 85-54 18-50 A7-61 DE-57 14-42 6
    03200336 -95 18-48 71-61
    03200348 -95 18-49 71-56
    03200344 -95 18-49 71-63
    
    Similarly, STA 14, a subnotebook, moves between the best and the worst locations -- next door to the router and the basement. STA DE is the furthest from the router, 2 floors up, and closest to the routers across the street.

    Code:
      egrep '14-30|14-65' wlmon
    03192349 -92 A2-58 71-53 85-58 18-57 A7-68 14-65 3
    03200135 -89 18-49 85-58 5B-58 A2-44 71-52 14-30
      egrep 'DE-52|DE-66' wlmon
    03192210 -91 5B-54 DE-66 14-57 6
    03192211 -91 5B-54 DE-66 14-58
    03192212 -92 5B-56 DE-66 14-55
    03192213 -90 5B-55 DE-66 14-60
    03192214 -92 5B-56 DE-66 14-58
    03192215 -91 A7-53 5B-54 DE-66 14-56 6
    03192217 -91 A7-57 5B-55 DE-66 14-60
    03192301 -90 A7-50 DE-66 14-43 3 4 5
    03192302 -92 A7-49 DE-66 14-44 3 4 5
    03192303 -90 A7-51 DE-66 14-43 3 4 5
    03192304 -92 A7-51 DE-66 14-44 3 4 5
    03192305 -91 A7-53 DE-66 14-43
    03192312 -90 A7-56 DE-66 14-45
    03192314 -90 A7-57 DE-66 14-43
    03192339 -90 A2-50 71-53 85-50 18-54 A7-57 DE-66 14-41 11
    03201035 -74 A2-51 DE-52
    03201037 -78 A2-51 DE-52
    03201040 -77 A2-51 DE-52
    
    It seems that higher signal strength correlates with higher noise. In the case of DE, which is stationary, the difference between Noise and RSSI (=Quality) is pretty stable at 25.
     
  10. Planiwa

    Planiwa LI Guru Member

    Two more observations. I had set "wl interference 4". But between 0610 and 0615 something had changed that to 3. No one was on the network or on the system. Nothing ran except wlmon. The only peculiar thing is that just then, the Noise increased by 8dB and kept rising, from a steady level of -91 at 0610 up to -69 by 0634:

    Code:
    03200607 -91               
    03200608 -91               
    03200609 -91               
    03200610 -91                 
    03200611 -83               
    03200612 -83               
    03200613 -83               
    03200614 -84               
    03200615 -82                       
    03200616 -82                     
    03200617 -81                     
    03200618 -83                     
    03200619 -84                     
    03200620 -84                           
    03200621 -84                         
    03200622 -84                         
    03200623 -83                         
    03200624 -81                         
    03200625 -81                     
    03200626 -74                       
    03200627 -73                       
    03200628 -75                       
    03200629 -75                       
    03200630 -74                     
    03200631 -77                     
    03200632 -75                     
    03200633 -74                     
    03200634 -69                     
    03200635 -72                     
    03200636 -74                     
    03200637 -73                   
    
    Perhaps the change in setting affects what wl reports?
    Or the change in noise caused something to change the setting?
    Or something cause the noise and changed the setting?
    Or the absence of interference changes the interference-mitigation setting?

    Perhaps wl should listen to Bessie Smith and Fats Waller -- paraphrasing Clarence Williams:
    "When you set me, daddy, I stay set." :)

    Also, there is an interference storm raging now, with interference on up to 5 channels:

    Code:
    03201300 -83 14-45 A2-48 1 5 9
    03201301 -82 14-43 A2-48 5 6
    03201302 -84 14-36 A2-48 5 6
    03201303 -82 14-49 A2-48 5 6
    03201304 -82 14-53 A2-49 5 6
    03201305 -83 14-53 A2-48 5 7 8
    03201306 -84 14-47 A2-48 2 8
    03201307 -82 14-49 A2-48 2 8
    03201308 -79 5B-49 14-47 A2-48 2 8
    03201309 -81 5B-47 14-52 A2-48 2 8
    03201310 -79 5B-48 14-50 A2-48 3 4 6 8
    03201311 -79 5B-45 14-54 A2-48 1 3 6 7 8
    03201312 -80 5B-43 14-49 A2-48 1 3 6 7 8
    03201313 -80 5B-49 14-49 A2-48 1 3 6 7 8
    03201314 -81 5B-49 14-48 A2-50 1 3 6 7 8
    03201315 -79 5B-52 14-50 A2-49 1 4 6 7
    03201316 -83 5B-50 14-42 A2-48 3 4
    03201317 -84 5B-49 14-47 A2-50 3 4
    03201318 -79 5B-45 14-50 A2-49 3 4
    03201319 -80 5B-47 14-49 A2-49 3 4
    03201320 -77 5B-47 14-49 A2-49 3 4
    03201321 -83 5B-51 14-50 A2-48 2 4 5 6
    03201322 -81 5B-47 14-57 A2-52 2 4 5 6
    03201323 -78 5B-46 14-52 A2-51 2 4 5 6
    03201324 -78 5B-48 14-48 A2-50 2 4 5 6
    03201325 -81 5B-53 14-53 A2-50 2 4 5 6 7
    03201326 -82 5B-53 14-51 A2-50 4 5 7
    03201327 -83 5B-54 14-50 A2-50 4 5 7
    03201328 -83 5B-55 14-50 A2-49 4 5 7
    03201329 -78 5B-49 14-49 A2-48 4 5 7
    03201330 -78 5B-49 14-50 A2-49 1 2 4 5 7
    03201331 -82 5B-47 14-51 A2-49 6 7
    03201332 -78 5B-49 14-50 A2-49 6 7
    03201333 -82 5B-47 14-50 A2-49 6 7
    03201334 -79 5B-48 14-57 A2-49 6 7
    03201335 -79 5B-47 14-51 A2-49 1 3 4 6 7
    03201336 -80 5B-48 14-48 A2-49 3 4 5
    03201337 -81 5B-47 14-54 A2-49 3 4 5
    03201338 -81 5B-47 14-49 A2-51 3 4 5
    03201339 -80 5B-48 14-50 A2-51 3 4 5
    03201340 -78 5B-50 14-42 A2-50 8 11
    03201341 -80 5B-50 14-47 A2-49 1 4 6 7
    03201342 -82 5B-49 14-54 A2-49 1 4 6 7
    03201343 -76 5B-50 14-49 A2-53 1 4 6 7
    03201344 -76 5B-47 14-46 A2-48 1 4 6 7 11
    03201345 -81 5B-49 14-51 A2-47 1 4 6
    03201346 -79 5B-48 14-46 A2-53 4 6
    03201347 -75 5B-47 14-47 A2-55 4 6
    03201348 -80 5B-50 14-49 A2-51 4 6
    03201349 -81 5B-51 14-54 A2-46 4 6
    03201350 -83 5B-50 14-46 A2-46 8
    03201351 -83 5B-50 14-53 A2-46 2 4 9
    03201352 -80 5B-53 14-51 A2-47 2 4 9
    03201353 -78 5B-49 14-54 A2-45 2 4 9
    03201354 -80 5B-48 14-44 A2-47 2 4 9
    03201355 -81 14-47 A2-53 6
    03201356 -79 14-49 A2-47 1 4
    03201357 -80 14-50 A2-53 1 4
    03201358 -82 14-48 A2-47 1 4
    03201359 -82 14-49 A2-47 1 4
    03201400 -83 14-50 A2-53 1 2 6 7 8
    
    Here is the bigest one of the day -- Interference on 8 channels:

    Code:
    03201620 -78 5B-29 18-46 71-61 14-57 A2-52
    03201621 -80 5B-46 18-43 71-67 14-53 A2-52
    03201622 -80 5B-47 18-43 71-64 14-50 A2-52
    03201623 -83 5B-48 18-47 71-62 14-52 A2-53
    03201624 -80 5B-47 18-44 71-59 14-53 A2-53
    03201625 -79 5B-48 18-48 71-60 14-54 A2-53 6
    03201626 -80 5B-47 18-44 71-62 14-55 A2-52 6
    03201627 -84 5B-47 18-46 71-69 14-53 A2-52 6
    03201628 -79 5B-47 18-44 71-58 14-45 A2-52 6
    03201629 -81 5B-52 18-45 71-60 14-53 A2-53 6
    03201630 -78 5B-50 18-44 71-67 14-52 A2-53 2
    03201631 -81 5B-50 18-48 71-65 14-52 A2-53 2
    03201632 -79 5B-51 18-47 71-63 14-52 A2-52 2
    03201633 -82 5B-52 18-49 71-63 14-55 A2-52 2
    03201634 -80 5B-49 18-43 71-68 14-57 A2-52 2
    03201635 -80 5B-50 18-44 71-61 14-54 A2-53 3 4 5 7
    03201636 -80 5B-48 18-44 71-67 14-51 A2-53 3 4 5 7
    03201637 -82 5B-47 18-44 71-73 14-49 A2-52 3 4 5 7
    03201638 -80 5B-48 18-43 71-61 14-54 A2-51 3 4 5 7
    03201639 -82 5B-50 18-48 71-67 14-51 A2-51 3 4 5 7
    03201640 -82 5B-48 18-52 71-62 14-56 A2-51 1 2 3 4 5 6 8 9
    03201641 -81 5B-45 18-41 71-57 14-58 A2-52 1 2 3 4 5 6 8 9
    03201642 -82 5B-48 18-43 71-67 14-56 A2-53 1 2 3 4 5 6 8 9
    03201643 -78 5B-48 18-43 71-60 14-52 A2-53 1 2 3 4 5 6 8 9
    03201644 -81 5B-52 18-47 71-55 14-53 A2-53 1 2 3 4 5 6 8 9
    03201645 -83 5B-50 18-47 71-54 14-50 A2-53 1 3 9
    03201646 -79 5B-47 18-43 71-63 14-50 A2-53 1 3 9
    03201647 -82 5B-50 18-46 71-71 14-56 A2-53 1 3 9
    03201648 -81 5B-51 18-47 71-55 14-50 A2-51 1 3 9
    03201649 -82 5B-51 18-49 71-59 14-55 A2-51 1 3 9
    03201650 -81 5B-48 18-43 71-61 14-51 A2-53 3 4 5
    03201651 -83 5B-53 18-49 71-61 14-51 A2-52 3 4 5
    03201652 -83 5B-51 18-50 71-60 14-50 A2-52 3 4 5
    03201653 -81 18-44 5B-52 71-60 14-51 A2-52 3 4 5
    03201654 -83 18-46 5B-50 71-60 14-50 A2-52 3 4 5
    03201655 -81 18-46 5B-52 71-61 14-52 A2-52 3 4
    03201656 -81 18-43 5B-48 71-63 14-50 A2-53 3 4
    03201657 -80 18-43 5B-51 71-53 14-55 A2-52 3 4
    03201658 -80 18-49 5B-55 71-54 14-50 A2-52 3 4
    03201659 -80 18-42 5B-52 71-54 14-53 A2-53 3 4
    03201700 -83 18-49 5B-51 71-53 14-49 A2-52
    
     
  11. Planiwa

    Planiwa LI Guru Member

    I am happy to report that in the Network Status web page that is available to all the LAN users at http:/router/

    The following is now included (based on the wlmon data collected and shown above):
    Code:
    Last   10 minutes. Noise: Mean: -82 dBm, Max -77 dBm, at 1938 ; Interference: 150% , Max 3 at 1944 .
    Last  100 minutes. Noise: Mean: -81 dBm, Max -77 dBm, at 1938 ; Interference: 59% , Max 3 at 1944 .
    Last 1000 minutes. Noise: Mean: -81 dBm, Max -69 dBm, at 0634 ; Interference: 96% , Max 8 at 1644 .
    
     
  12. Planiwa

    Planiwa LI Guru Member

    I now have strong evidence that suggests that wl is broken!

    Once again wl changed its IM setting spontaneously from 4 to 3; this time it did so at 2230:

    Code:
    03212220 -90 5B-54 18-46 71-56 14-52 5 6
    03212221 -94 18-45 A7-60 5B-56 71-56 14-50 5 6
    03212222 -84 18-45 A7-66 5B-56 71-57 14-54 5 6
    03212223 -88 18-46 A7-63 5B-56 71-58 14-42 5 6
    03212224 -85 18-46 A7-62 5B-57 71-57 14-41 5 6
    03212225 -87 18-45 A7-66 5B-57 71-57 14-47
    03212226 -88 18-46 A7-66 5B-57 71-56 14-43
    03212227 -88 18-45 A7-64 5B-57 71-56 14-39
    03212228 -89 A7-53 18-45 5B-57 71-57 14-42
    03212229 -88 A7-53 18-45 5B-57 71-55 14-42 1
    03212230 -80 71-52 5B-54 18-51 A7-53 1 3 5 7
    03212231 -82 71-56 5B-50 18-49 A7-53 1 3 5 7
    03212232 -80 71-52 5B-52 18-49 A7-53 1 3 5 7
    03212233 -80 71-53 5B-50 18-48 A7-53 1 3 5 7
    03212234 -81 71-53 5B-50 18-48 A7-52 1 3 5 7
    03212235 -79 A7-53 71-51 5B-52 18-49 1 2 3 4 7 8 9
    03212236 -80 A7-53 71-51 5B-51 18-48 1 2 3 4 7 8 9
    03212237 -80 A7-53 71-52 5B-52 18-48 1 2 3 4 7 8 9
    03212238 -79 5B-52 A7-53 71-51 18-49 1 2 3 4 7 8 9
    03212239 -80 5B-50 A7-52 71-51 18-49 1 2 3 4 7 8 9
    03212240 -80 5B-55 A7-55 71-52 18-49 2 4 5 6 7
    03212241 -81 5B-50 A7-57 71-55 18-49 2 4 5 6 7
    03212242 -82 A7-58 18-50 5B-49 71-55 2 4 5 6 7
    03212243 -79 A7-57 18-50 5B-49 71-56 2 4 5 6 7
    03212244 -80 A7-55 18-50 5B-51 71-52 2 4 5 6 7
    03212245 -79 5B-52 A7-54 18-50 71-53 11
    03212246 -83 A7-54 14-37 71-56 5B-49 18-46 11
    03212247 -86 A7-54 14-42 71-56 5B-50 18-46 11
    03212248 -74 A7-54 14-39 71-56 5B-50 18-50 11
    03212249 -81 A7-56 14-40 71-56 5B-49 18-49 11
    03212250 -78 A7-52 14-41 71-56 5B-50 18-46
    
    You will note that immediately after the change, Noise went up 8 dBm and Interference went from none to 7 within 5 minutes.

    You may recall that on 0320 the IM was at 3 between 0615 and 2100. Just now it was 3 from 2229 until I reset tit to 4 at 0000. You can clearly see that both Noise and Interference increased during those times by examining these timeline charts:

    Code:
    Noise Timeline (in -dBm) -- Maximum noise measured during hour.
    ddmmHHMM 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
     
    03192205                                                                  82 79
    03200000 84 84 78 84 85 85 69 72 72 72 72 72 73 75 69 75 77 76 78 77 72 83 81 84
    03210000 83 84 83 82 83 83 84 84 83 83 84 84 84 81 84 84 74 83 82 83 84 84 73 72
    03220000 84
     
    Interference Timeline -- MeanMax -- severely interfered channels during hour.
    ddmmHHMM 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23
     
    03192205                                                                  22 33
    03200000 33 23 11 11 22 22 02 12 22 22 22 33 45 55 67 22 48 22 12 23 23 11 11 11
    03210000 11 11 01 11 11 00 00 11 11 01 33 00 11 00 11 14 11 46 55 11 12 00 47 11
    03220000 01
    
    So, Mr. wl has some explaining to do:

    1. Why, when Interference is set to 4, does wl spontaneously change it to 3?

    3. Why does the Noise drop significantly when IM=3?

    4. Why does the Interference increase significantly when IM=3?

    5. What is the meaning of this message:

    Auto Wireless LAN Interference mitigation is enabled and not active, and noise reduction is enabled. (mode 4)

    What is meant by "active"? I have never seen it say "active", even when there was much noise or much interference. So, how is this different from:

    Auto Wireless LAN Interference mitigation is enabled and not active. (mode 3)

    And why can't I get it to be "active", whatever that means?

    I should mention that I always run with NOCOMMIT on, and that ...
    Code:
      NVRAM find mitigat
        Searching in flashed NVRAM:
    wl_mitigation=0
    wl0_mitigation=3
        Searching in working nvram:
    wl0_mitigation=4
    wl_mitigation=0
    
    [Tomato v1.28.7494 MIPSR2-Toastman-VLAN-RT K26 USB VPN-NOCAT]
     
  13. Planiwa

    Planiwa LI Guru Member

    Just in case anyone is reading this thread ... ;-)

    It happened again. Right under my nose. One minute it was "4" and the next it was "3". All by itself.
    I wonder what "wl interference_override" means in this context?

    Another observation: When there is severe interference, it tends to be in "off-side" channels, such as 2, 3, 5.
    And, scans at the time show no other radios on those channels.

    This suggests non-WLAN sources of interference.

    Of the options:

    interference
    Get/Set interference mitigation mode. Choices are:
    0 = none
    1 = non wlan
    2 = wlan manual
    3 = wlan automatic
    4 = wlan automatic with noise reduction

    Might #1 be a better choice? Why or why not?

    The data clearly shows that 4 is better than 3, not only for noise, but also for interference.

    I may have to look at the wl source. Shudder. Brrr.
     
  14. Planiwa

    Planiwa LI Guru Member

    In the last message I had asked if "1" might be better than "4". Today I experienced a "wifi unusable except right next to the router" event:

    (In kitchen, above router):

    Code:
      Size Received/Sent Dups? Loss    Min  Mean  Max  Dev  Trate RSSI Noi  MMDD_HH:MMSS
     
    1420 B    257/1000 P      74.3%  185.6  7397 12207 3362      5  -53 -90  0403_14:2528
    1420 B    153/1000 P      84.7%  2564.5  3434  4172  386      5  -51 -90  0403_14:2541
    1420 B    167/1000 P      83.3%  469.8  4627  6205 1393      5  -51 -91  0403_14:2555
    1420 B    156/1000 P      84.4%  3406.4  4542  5450  607      5  -54 -89  0403_14:2611
    1420 B    276/1000 P      72.4%  258.9  3010  4775 1217      5  -51 -89  0403_14:2624
    1420 B    260/1000 P      74.0%  1020.3  2431  3600  553      5  -53 -87  0403_14:2635
    1420 B    236/1000 P      76.4%    5.6  2475  3825  834      5  -52 -88  0403_14:2647
    1420 B    275/1000 P      72.5%    6.5  2121  3918 1206      5  -52 -88  0403_14:2700
    1420 B    353/1000 P      64.7%    12.9  1436  2754  715      5  -55 -90  0403_14:2711
    1420 B    388/1000 P      61.2%    5.7  5638  9442 3216      5  -56 -88  0403_14:2725
    1420 B    340/1000 P      66.0%    20.0  2024  3753 1034      5  -54 -89  0403_14:2740
    1420 B    471/1000 P      52.9%    7.0  2810  5055 1495      5  -56 -90  0403_14:2752
    1420 B    432/1000 P      56.8%    7.3  1958  3051  953      5  -55 -89  0403_14:2803
    1420 B    468/1000 P      53.2%    6.5  2768  4982 1603      5  -55 -89  0403_14:2817
    1420 B    254/1000 P      74.6%    14.6  2250  4359 1394      5  -54 -90  0403_14:2827
    1420 B    208/1000 P      79.2%  589.3  4321  6868 2321      5  -55 -90  0403_14:2843
    1420 B    235/1000 P      76.5%  429.3  3005  4981 1344      5  -56 -91  0403_14:2855
    
    So, I changed "wl interference 1" (was 4). The result was immediate and dramatic:
    Code:
      Size Received/Sent Dups? Loss    Min  Mean  Max  Dev  Trate RSSI Noi  MMDD_HH:MMSS
    1420 B    533/1000 P      46.7%    1.8  962  2961  761    59  -56 -88  0403_14:2905
    1420 B  1000/1000 P      0.0%    1.0    3    50    4    117  -56 -89  0403_14:2911
    1420 B  1000/1000 P      0.0%    1.0    4    56    6    117  -57 -89  0403_14:2917
    1420 B  1000/1000 P      0.0%    1.0    4    47    5    104  -57 -89  0403_14:2925
    1420 B  1000/1000 P      0.0%    1.0    3    42    4    104  -56 -89  0403_14:2931
    1420 B  1000/1000 P      0.0%    0.9    3    36    4    104  -55 -90  0403_14:2936
    1420 B  1000/1000 P      0.0%    1.0    4    53    6    78  -56 -89  0403_14:2942
    1420 B  1000/1000 P      0.0%    1.0    3    26    2    145  -52 -90  0403_14:2948
    1420 B  1000/1000 P      0.0%    1.0    5    67    7    145  -56 -90  0403_14:2953
    1420 B  1000/1000 P      0.0%    1.0    4    44    4    117  -57 -86  0403_14:2959
    1420 B  1000/1000 P      0.0%    1.0    4    45    5    117  -54 -87  0403_14:3004
    1420 B  1000/1000 P      0.0%    1.0    5    54    6    117  -54 -89  0403_14:3010
    1420 B  1000/1000 P      0.0%    1.0    6  100  10    52  -57 -89  0403_14:3015
    1420 B  1000/1000 P      0.0%    1.0    3    46    4    104  -54 -90  0403_14:3024
    1420 B  1000/1000 P      0.0%    1.0    5    91    9    104  -57 -89  0403_14:3029
    1420 B  1000/1000 P      0.0%    1.0    4    50    5    78  -57 -89  0403_14:3035
    1420 B  1000/1000 P      0.0%    0.9    3    75    6    117  -56 -89  0403_14:3041
    1420 B  1000/1000 P      0.0%    1.0    3    17    2    117  -56 -90  0403_14:3046
    
    Packet loss went from 70% to 0%.
    Latency from over 5000ms to under 5ms.
    Data rate from under 10 to over 100Mb/s.

    [EDIT: fixed typo "wl interference 1"]
     
  15. mvsgeek

    mvsgeek Addicted to LI Member

    *Cough...raises hand* I'm reading it, and trying to digest as much as my aging brain cells will absorb.

    Is "wl interference a" valid? I get an error. Did you mean "0" (none)? Interesting that the dramatic improvement was not accompanied by any significant change in RSSI or Noise.

    I have an ongoing need to identify sources of interference in my rural environment - sometimes it's as if a giant microwave oven has been fired up in the field between me and the main router. Wet trees don't help either, especially in spring and summer.

    Good luck with the wl source - I'd offer to help, but it's probably not written in BAL or COBOL;)
     
  16. bluenote

    bluenote Networkin' Nut Member

    I'd appreciate the cliff's notes for these results as well.
    Can you draw any early conclusions about what you've found?
     
  17. Planiwa

    Planiwa LI Guru Member

    @mvsgeek thanks for the typo alert. Should be "1", of course.

    As for COBOL, I tend to agree with Dijkstra, although nowadays I would apply the sentiment to Linux, and in this context I am leaving out GNU, because GNU's not the problem. :) Enough of that. I did have the privilege of meeting Admiral Grace Hopper -- with her handy light-nanosecond. :)


    As for the wl command -- it is a good example of a horrible program -- poorly designed, undocumented, poorly implemented, apparently understood by no one, prone to intermittent, unpredictable, difficult to detect error states, and, of course, mission-critical, and without an alternative option.

    In this thread I am trying to discover how to use the wl command to measure and control wireless parameters.

    The obvious questions include:

    What useful information (about the wireless network) can we get with the wl command?
    What concepts and assumptions underlie the essential workings of this command?
    What do some of the poorly worded pieces of "documentation" mean?
    Under what circumstances wl settings get unset unexpectedly and apparently "by themselves"?
    How does this command and its universe interact with other configuration-commands, such as nvram?

    (BTW, I have some very nice nvram tools now, to examine and compare the flashed and the cached nvram, and to check if it has changed.)
    . . .

    Clearly this sort of work is not of a casual nature. It is a work in ptogress. Hopefully there can be some collaboration to advance the state of understanding of these matters.

    I'll be happy to clarify anything about my findings so far that may be unclear.
     
  18. bluenote

    bluenote Networkin' Nut Member

    So you found that using the auto setting degraded performance, and the "other than wifi" interference setting increased your performance significantly.
    How about compared to "none" setting?
     
  19. Planiwa

    Planiwa LI Guru Member

    At this point we should read again Toastman's advice -- avoid any of these settings:

    Interference Mitigation looks like a Solution to the problem of Wifi instability. But it can actually aggravate the problem to the point of complete breakdown.

    All settings other than None (0) are hazardous. They may look like they are working for a time. Then they cause extreme slowdowns, disconnects, inability to connect, and disappearance of the access point from view altogether.
     
  20. mvsgeek

    mvsgeek Addicted to LI Member

    This is exactly what I observed after changing the setting from "none" to "non wlan" (1). No disappearances, but several of my WDS slaves failed to re-connect. Also major slowdown on the successfully re-connected slaves. Since setting it back to "none" everybody is happy again.
     
  21. Planiwa

    Planiwa LI Guru Member

    It happened again: A few hours ago, wl interference mitigation spontaneously switched from 0 to 3.

    The NVRAM setting, in both the flashed NVRAM and the cached NVRAM was: wl0_mitigation=4.

    I have now set it to 0 and commited.

    I have lomg suspected that wl messes with the NVRAM.

    Let's see what happens next.
     
  22. Planiwa

    Planiwa LI Guru Member

    What is chanim_mode and how does it relate to chanim_state detection?

    Code:
      wl chanim_mode ?
    chanim_mode
    get/set channel interference measure (chanim) mode
    Usage: wl chanim_mode <value>
    value: 0 - disabled; 1 - detection only; 2 - detection and avoidance
     
      wl chanim_mode 1
     
      for i in $(wl channels);do echo -n "$(wl chanim_state $i) ";done;echo
    0 0 0 0 0 1 0 0 0 0 0 
      
      wl chanim_mode 0
     
      for i in $(wl channels);do echo -n "$(wl chanim_state $i) ";done;echo
    0 0 0 0 0 1 0 0 0 0 0 
     
      wl chanim_mode
    CHANIM mode: disabled.
    
     
  23. Planiwa

    Planiwa LI Guru Member

    It happened again. Between 10:51 and 10:56.
    The wl interference setting spontaneously went from 0 to 3. NVRAM settings are 0:
    Code:
      nvr miti
         Searching in Flashed NVRAM: 
    wl_mitigation=0
    wl0_mitigation=0
         Searching in working nvram: 
    wl_mitigation=0
    wl0_mitigation=0
     
      wl chanim_mode
    CHANIM mode: disabled.
      wl interference
    Auto Wireless LAN Interference mitigation is enabled and not active. (mode 3)
    
    Am I the only one who is experiencing this bizarre behaviour by wl?
    Or am I the only one who is aware of it?
     
  24. Planiwa

    Planiwa LI Guru Member

    So, I now have a very nice script (wrs) that explores wl readings and settings, does a client report and a networks report, and also reports selected settings, both from flash nvram and from working nvram.

    It is meant for sys admins who wish to explore the wl command for insight and understanding of their wireless network, and perhaps collaborate with others.

    (It is not meant for casual users who just want to be told what to do.)

    Depending on how much interest there is, I may post it here or by PM. It is 140 lines, and the output is about 75-175 lines.
     
  25. ryzhov_al

    ryzhov_al Networkin' Nut Member

    Please, share your scirpt with us.
     
  26. Planiwa

    Planiwa LI Guru Member

    So I have a general question and a particular question.

    Is there a forum where people are actually having a conversation about these things? Does the OpenWRT project have such a forum?

    The particular question:

    wl chanim_mode

    What is this? How does it relate to

    wl interference

    and how does that relate to

    wl interference_override

    The wl "help text" (1328 lines for 323 subcommands) confusingly states:

    Code:
    # channel interference measure (chanim) mode
    # 0 - disabled; 1 - detection only; 2 - detection and avoidance
     
    # interference mitigation mode.
    # 0= none 1= non wlan 2= wlan manual 3= wlan automatic 4= wlan automatic with noise reduction
     
    # interference mitigation override.
    # 0 = no interference mitigation 1 = non wlan 2 = wlan manual 3 = wlan automatic
    # 4 = wlan automatic with noise reduction -1 = remove override, override disabled
    
     
  27. azdps

    azdps LI Guru Member

    Very interesting read so far Planiwa. Keep posting more info. Unfortunately I have nothing to add.
     
  28. Planiwa

    Planiwa LI Guru Member

    Here are some wl subcommands that appear to be if interest.
    Some require arguments. Use "wl -h foo" for hints about subcommand foo.
    (Some of these have corresponding nvram setting.)

    Code:
    wl scan
    wl scanresults
    wl chanim_state
     
    wl sta_info
    wl noise
    wl phy_rxiqest
    wl channel
    wl country
    wl chanim_mode
    wl interference
    wl interference_override
    wl gmode
    wl protection_control
    wl gmode_protection
    wl gmode_protection_control
    wl gmode_protection_override
    wl nmode_protection
    wl mimo_txbw
    wl wme
    wl wme_apsd
    wl wme_maxbw_params
    wl wme_tx_params
    wl frameburst
    wl plcphdr
    wl abminrate
    wl txpwr1
    wl -h txpwr1
    wl txpwrlimit
    wl powerindex
    wl ver
    
     

Share This Page