Outrageously high Quality in "Device List" (Shibby)

Discussion in 'Tomato Firmware' started by Planiwa, Jun 8, 2013.

  1. Planiwa

    Planiwa Network Guru Member

    Much of the time (far more than 10%), the Status > Device List page
    shows, for one or more clients, a Quality (SNR) that is outrageously high (e.g. 100 or more),
    which is clearly wrong.

    In these cases, the RSSI is clearly wrong, often with a large positive value (e.g. 50).

    The error is clearly in the wl driver itself, as can be verified by running

    for INTERFACE in $(nvram get wl_ifnames)
      for CLIENT in $(wl -i $INTERFACE assoclist | sed 's/assoclist//')
        echo $INTERFACE $CLIENT $(wl -i $INTERFACE rssi $CLIENT)
    (This script shows the RSSI for each associated device on each interface/adapter).

    I find it astonishing that no one else appears to have noticed this. I wonder if the newest version still has this interesting feature. (Perhaps people are just happy to see what impossibly good signals they have?! :)

    1.28.0000 MIPSR2-104 K26 USB AIO-64K
    Asus RT-N66U
    wl0: May 29 2012 08:48:50 version
    pharma likes this.
  2. RMerlin

    RMerlin Network Guru Member

    I have seen this problem at random occasions over the months. It is indeed a problem in the wireless driver, restarting the wireless interfaces usually fixes it, otherwise a complete reboot of the router does.

    Unsure what triggers it, I suspect it might be when you send too many requests in a short period of time, leaving the wireless driver "stuck".
  3. Toastman

    Toastman Super Moderator Staff Member Member

    The fault has been around a long time, but there's nothing much we can do about it.
  4. Planiwa

    Planiwa Network Guru Member

    RMerlin, I wonder whether stock FW has this feature? FWIW, I rebooted a router earlier in the day, and the first thing I saw after it came up, were nice multicoloured (2 green bars) clearly bogus Quality displays over > 100 SNR.

    Toastman, sometimes just acknowledging a problem may contribute to its solution. :)

    In any case, once people understand that the wl driver might return a value of 6 or 66 for RSSI instead of -66, they may trust it less to "mitigate" problems. :)
  5. mstombs

    mstombs Network Guru Member

    The newest drivers must be those being used by dd-wrt Linux kernel 3.9 builds? Wonder if they still have the problem?
  6. RMerlin

    RMerlin Network Guru Member

    Not necessarily. You would have to check the actual version string to know for sure, but somehow I doubt that DD-WRT upgraded to the Broadcom SDK 5.110 code.
  7. RMerlin

    RMerlin Network Guru Member

    Stock FW doesn't report the client's RSSI.
  8. Elfew

    Elfew Network Guru Member

    People in Asus know why ;)
  9. mstombs

    mstombs Network Guru Member

    beta k3.9 dd-wrt on n66u reports this

    ____  ___    __        ______ _____        ____  _  _
    | _ \| _ \  \ \      / /  _ \_  _| __  _|___ \| || |
    || | || ||____\ \ /\ / /| |_) || |  \ \ / / __) | || |_
    ||_| ||_||_____\ V  V / |  _ < | |    \ V / / __/|__  _|
    |___/|___/      \_/\_/  |_| \_\|_|    \_/ |_____|  |_|
                          DD-WRT v24-sp2
    BusyBox v1.21.0 (2013-05-27 10:31:33 CEST) built-in shell (ash)
    Enter 'help' for a list of built-in commands.
    root@DD-WRT:~# uname -a
    Linux DD-WRT 3.9.4 #125 Mon May 27 10:19:14 CEST 2013 mips GNU/Linux
    root@DD-WRT:~# wl ver
    6.30 RC163.35
    wl1: May 27 2013 10:18:49 version (r391398)
    Is it OpenWRT driver rather than Broadcom?

    rssi's in wireless survey all look fine -87 to -20
  10. Marcel Tunks

    Marcel Tunks Networkin' Nut Member

    So far the dd-WRT k3.x builds fix a few things but break many, many more. It's definitely still an alpha test. I don't have the skills to determine if it's the wireless drivers, the new kernel, or dd-WRT's stuff causing problems, but either way it currently doesn't seem appropriate for Tomato. Glad they're doing it though, someone needs to be bleeding edge, and some of the features of the new kernel are fun to play with.
  11. RMerlin

    RMerlin Network Guru Member

    6.30 RC163 is Broadcom's driver from their 6.x SDK. This is usually used with 802.11ac devices, 802.11n using the 5.x SDK. BSlayer might have compiled his own driver using the 6.x SDK and made them work with older devices (unless your router is actually a 802.11ac device). I experimented a bit in the past with backporting the 6.xx driver to older devices usually based on the 5.x SDK, and the results were less than stellar. I decided it was a waste of time and moved on.
  12. mstombs

    mstombs Network Guru Member

    Was definitely on an N66u, had a look at latest Asus, Rmerlin, then dd-wrt to end up back at Tomato (without getting out of chair or removing router from wall!).

    Its great some activity in newer kernels in dd-wrt, apparently routing not up to speed yet, but was clearly capable of more than my connection, and yes I'm sure dd-wrt have access to Broadcom sdk's via Buffalo/ commercial deals, and compile time makes it clear the driver built at same time as firmware.

    I can't tell you what driver dd-wrt k3.x on an RT-N16 uses, it just soft-bricks the device, something to do with how they are trying to make the 32kB nvram pretend to be 64kB - a specific upgrade path may get to a working router - until it is reset to defaults!
  13. macgyver

    macgyver Reformed Router Member

    running latest ddwrt 2.6 kernel on E3000 reports this combination which I find to be bad with 5GHz at least on this router
    root@E3000:~# wl ver
    6.30 RC163.1
    wl0: Dec 14 2011 03:57:42 version
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice