Re: Trouble Associating with ath5k on AR5212

[Will Keaney <keaneyw@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2201 bytes --]

Will Keaney wrote:
> Bob Copeland wrote:
>   
>> On Thu, Apr 02, 2009 at 10:17:31PM -0400, Will Keaney wrote:
>>   
>>     
>>> dmesg shows this:
>>> [38645.639170] ath5k 0000:03:00.0: PCI INT A -> GSI 17 (level, low) ->
>>> IRQ 17
>>> [38645.639199] ath5k 0000:03:00.0: setting latency timer to 64
>>> [38645.639342] ath5k 0000:03:00.0: registered as 'phy5'
>>> [38645.872038] phy5: Selected rate control algorithm 'minstrel'
>>> [38645.882037] Registered led device: ath5k-phy5::rx
>>> [38645.882090] Registered led device: ath5k-phy5::tx
>>> [38645.882097] ath5k phy5: Atheros AR5414 chip found (MAC: 0xa3, PHY: 0x61)
>>> [38646.133614] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>>> [38667.211026] ath5k phy5: noise floor calibration timeout (5785MHz)
>>>
>>> I think the key there is the last line.  There is no further output in
>>> dmesg, regardless of what I do in wpa_supplicant.
>>>     
>>>       
>> Probably a red herring, actually -- but yeah, you should be seeing more
>> output as the association process continues.  That is a pretty frequent
>> and harmless error message, here you are scanning and just changed to
>> a 5.7 ghz channel for a passive scan.
>>
>> Are you getting any results at all from scans?  Is ath5k using 100% of
>> any cpu?
>>   
>>     
> I do get some results on scans, but not as many as with madwifi.  Also,
> it reports dramatically different signal strengths than madwifi (though
> this may be because madwifi sets my maximun Tx power to 17dBm, and ath5k
> sets it to 20).
> I don't see any untoward CPU usage when ath5k is loaded.  wpa_supplicant
> uses about 40% on one core, give or take.
>   
Following up on our conversation in IRC, I found a few other nearby
networks where I could get stronger signal.
ath5k showed much higher signal strengths for all of the networks it
could see than madwifi did.  Although ath5k showed a particular network
with strength between 16-29 dBm, it was unable to connect.
madwifi showed the same network with a strength of 4 dBm, and was able
to connect almost instantly.
As before, I didn't see any additional dmesg output from athk, even with
'dmesg -n7'.

Will



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 261 bytes --]