Is this normal: ath9k: Two wiphys trying to scan at the same time

Is this normal: ath9k: Two wiphys trying to scan at the same time

[ASIC Felix]

Hi,

is this normal:

  ath9k: Two wiphys trying to scan at the same time

afaik, I did not enable virtual phy stuff. Connection appears to be
working okay. I've been seeing this for quite some time.

Setup: Laptop, Atheros MB92, kernel wireless-testing master-2009-09-23


wlan0: deauthenticated from 00:0b:85:6f:20:8c (Reason: 1)
ath9k: Two wiphys trying to scan at the same time
wlan0: direct probe to AP 00:0b:85:6f:20:8c (try 1)
wlan0 direct probe responded
wlan0: authenticate with AP 00:0b:85:6f:20:8c (try 1)
wlan0: authenticated
wlan0: associate with AP 00:0b:85:6f:20:8c (try 1)
wlan0: RX ReassocResp from 00:0b:85:6f:20:8c (capab=0x431 status=0
aid=5)
wlan0: associated

Best regards,
Felix

Re: Is this normal: ath9k: Two wiphys trying to scan at the same time

[Jouni Malinen]

On Thu, Sep 24, 2009 at 01:51:02PM -0700, ASIC Felix wrote:

> is this normal:
> 
>   ath9k: Two wiphys trying to scan at the same time
> 
> afaik, I did not enable virtual phy stuff. Connection appears to be
> working okay. I've been seeing this for quite some time.

That should only be showing up if virtual wiphys are used. In this case,
it should not really be showing up and is likely indicating a bug
somewhere (ath9k or mac80211). I don't know how to reproduce this
easily, so I have not looked into more details yet. It would be
interesting to add printk calls into ath9k_sw_scan_start() and
ath9k_sw_scan_complete() just after the mutex_lock() call to figure out
whether mac80211 ends up calling sw_scan_start() driver handler twice in
a row which would be one possible way to get this triggered. The other
way to hit this would be if ath9k were to somehow not update aphy->state
correctly when the scan has been completed.

-- 
Jouni Malinen                                            PGP id EFC895FA