Re: 6.0 BETA1 if_ath problems

From: James Snow (snow+freebsd-current_at_teardrop.org)
Date: 07/25/05

  • Next message: Julian Elischer: "installing 6.0 over the network"
    Date: Sun, 24 Jul 2005 20:24:33 -0400
    To: Sam Leffler <sam@errno.com>
    
    

    I'm sorry. I was grumpy when I wrote this.

    On Sat, Jul 23, 2005 at 04:19:45PM -0700, Sam Leffler wrote:
    > James Snow wrote:
    > >On Fri, Jul 22, 2005 at 10:09:48PM +0200, Toni Schmidbauer wrote:
    > >
    > >>so it seems things are getting better under 6.0.
    > >
    > >
    > >Unfortunately, I'm having the opposite experience. My X40's ath seems to
    > >get worse with each update. Right now it downs the ath0 interface
    > >at random intervals.
    >
    > I have one of almost every version of atheros part and I do not see the
    > problems you allude to; not that you've provided any useful info here.
    > In general I tend to ignore complaints of the form "it don't work".

    I have no useful information here since, as mentioned above, it's been
    completely random thus far. It may also be something that was fixed in
    my most recent update; time will tell.

    > >Also, when in adhoc mode it will sometimes spew errors, to the point of
    > >slowing the machine down. Repeatedly doing an 'ifconfig down' and
    > >reconfiguring the interface will eventually stabilize it.
    >
    > Again, nothing useful here. I just fixed a couple of adhoc mode
    > problems yesterday that are likely unrelated. I had a network of 4
    > machines setup in adhoc mode; 2 ath freebsd systems, one powerbook with
    > tiger, and one freebsd system with wi. All were in the same room. I saw
    > no problems in several hours of testing. When the 11b station was not
    > in the network I routinely got 28Mb/s for tcp netperf between stations
    > (this is 11g).

    This at least I can reproduce. It's only a problem when I'm first trying
    to bring up the adhoc network. Once it's up, (i.e., another node as
    associated with it) it stays up and is problem free.

    dev.ath.0.debug: 0xc0080 => 0xc00e0<reset,mode,beacon,state,node>

    I configure ath0 as follows:

    ifconfig ath0 172.16.1.1 netmask 255.255.255.252 ssid x40 wepmode off \
    mediaopt adhoc

    And eventually I get stuck in this loop:

    ath0: hardware error; resetting
    ath_draintxq: beacon queue 0x1eb8ac00
    ath_tx_stopdma: tx queue [0] 0, link 0
    ath_tx_stopdma: tx queue [1] 0, link 0
    ath_tx_stopdma: tx queue [2] 0, link 0
    ath_tx_stopdma: tx queue [3] 0, link 0
    ath_tx_stopdma: tx queue [8] 0, link 0
    ath_stoprecv: rx queue 0x1868dc, link 0xc15f28b0
    ath_mode_init: RX filter 0x97, MC filter 00000001:00000040
    ath_beacon_config: nexttbtt 100 intval 100 (100)
    ath_beacon_setup: m 0xc1835200 len 58
    ath0: hardware error; resetting

    athstats:

    66 hardware error interrupts
    56 mib overflow interrupts
    36 tx management frames
    110 tx frames discarded prior to association
    182 long on-chip tx retries
    33 tx frames with no ack marked
    26131 rx failed 'cuz of bad CRC
    2082 rx failed 'cuz of PHY err
        346 OFDM timing
        199 CCK timing
        1537 CCK restart
    68 beacons transmitted
    1196 periodic calibrations
    71582 rate control checks
    52129 switched default/rx antenna
    Antenna profile:
    [1] tx 2838 rx 269272
    [2] tx 16585 rx 280757

    80211stats:

    97454 rx from wrong bssid
    125 rx discard 'cuz dup
    30 rx discard 'cuz mcast echo
    81100 rx discard mgt frames
    349350 rx beacon frames
    24 rx element unknown
    24 rx frame chan mismatch
    5 active scans started

    This is the device:

    ath0: <Atheros 5212> mem 0xd0200000-0xd020ffff irq 11 at device 2.0 on pci2
    ath0: Ethernet address: 00:0e:9b:6e:60:fc
    ath0: mac 5.9 phy 4.3 radio 3.6

    Anything else I can provide, please let me know.

    > I am not saying the code is perfect. I'm aware of a couple of issues
    > that I'm working on but in general I am not seeing significant issues
    > with ath devices. If you or anyone else wants to help fix problems you
    > need to provide useful information such as the mac/phy version for the
    > part and the output from 80211stats and/or athstats. It also helps to
    > enable debugging msgs; usually at the 802.11 layer using 80211debug.

    Again, I apologize for the tone of my last email. I was just being
    generally bitchy and it was certainly not directed at you.

    -Snow

    _______________________________________________
    freebsd-current@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-current
    To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org"


  • Next message: Julian Elischer: "installing 6.0 over the network"

    Relevant Pages

    • Re: 802.11 RSSI value displayed from ifconfig(8)
      ... number of associations denied by wlan acl) information, ... tx frames discarded prior to association ... 1744404 tx stopped 'cuz no xmit buffer ... 4959 rx discard 'cuz acl policy ...
      (freebsd-stable)
    • Re: wpa drops out
      ... Selecting BSS from priority group 0 ... 78 rx discard 'cuz mcast echo ... 5064 rx discard mgt frames ...
      (freebsd-current)
    • Re: Dispatcher
      ... The thread getting the frames puts them in some sort of queue. ... Each recipient lives in its own thread and is basically waiting for a ... > b) a listener attached which is there to stay (needs to receive in order, ...
      (microsoft.public.dotnet.languages.csharp)
    • Re: AVStream: DMA & Process() QUEUEing
      ... How can the QUEUE be empty when I have cloned stream pointers on a Frames? ... > As I said you should do KsPinAttemptProcessing when the DMA completes. ...
      (microsoft.public.development.device.drivers)
    • Re: writing files in the Tick events when each write takes longer than the interval
      ... In my post I should have said: it is much like saving the frames of a ... Is it possible to use a message queue to track the writes? ... Seth Rowe http://sethrowe.blogspot.com/ ... yes you would need to save each bitmap into the queue ...
      (microsoft.public.dotnet.languages.vb)