Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 07 Sep 2014 10:16:06 -0700
From:      Nathan Whitehorn <nwhitehorn@freebsd.org>
To:        Adrian Chadd <adrian@freebsd.org>
Cc:        "freebsd-wireless@freebsd.org" <freebsd-wireless@freebsd.org>
Subject:   Re: Issues with urtwn
Message-ID:  <540C92D6.4030106@freebsd.org>
In-Reply-To: <CAJ-VmokyPcS077wHiP4Mdetms=meqk47v29fKA1edidhorVQpg@mail.gmail.com>
References:  <540C751F.6050202@freebsd.org> <CAJ-VmokyPcS077wHiP4Mdetms=meqk47v29fKA1edidhorVQpg@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help

On 09/07/14 08:28, Adrian Chadd wrote:
> On 7 September 2014 08:09, Nathan Whitehorn <nwhitehorn@freebsd.org> wrote:
>> I've been having some issues with connection stability in urtwn for several
>> months. The usual symptom is that after some period of time the connection
>> will apparently stall. If I'm running ping continuously, for instance, it
>> will at some point stop receiving replies. Then, sometime later, immediately
>> if I use the "reassociate" command in wpa_cli, the connection will fix
>> itself and all the packets I didn't get earlier get delivered at once:
>> hundreds of ping replies, for instance, some with time stamps minutes in the
>> past. No data is actually lost, though.
>>
>> I think the issue is that the driver does not actually support powersave
>> mode (maybe it should?) but reports to the AP that it does:
>>
>>> ifconfig wlan0 list sta (this is on the AP)
>> ADDR               AID CHAN RATE RSSI IDLE  TXSEQ  RXSEQ CAPS FLAG
>> 80:1f:02:cc:47:a9    1   11  11M  8.5    0   5526  55712 EPS AE      RSN
>>
>> I don't know enough about wireless to fix this, but the AP waiting for a
>> powersave poll and never getting one seems consistent with the problem. Is
>> there a simple way just to disable advertising this?
> When it next stalls, check ifconfig wlan0 and ifconfig urtwn0 - see if
> OACTIVE is set.
>
> I know iwn and ath had problems in the past where OACTIVE handling was
> plain broken (wasn't behind locks) and in an SMP, preemptive world
> things got gunked up.
>

OACTIVE is not set when it stalls. It also appears that only the RX path 
is stalled: transmitted packets make it, at least some of the time, to 
their destination when this happens.
-Nathan



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?540C92D6.4030106>