ath9k and power management

ath9k and power management

[Davide Pesavento]

Hello,

I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
(with -Dnl80211).
Enabling power management when connected (i.e. 'iwconfig wlan0 power
on'), causes the currently established connection to break and the
machine is no longer able to reconnect to the AP. Right after enabling
pm, dmesg says:

[ 2625.513233] wlan0: beacon loss from AP 00:1b:2f:35:14:78 - sending
probe request
[ 2627.513210] wlan0: no probe response from AP 00:1b:2f:35:14:78 -
disassociating
[ 2627.514430] phy0: Removed STA 00:1b:2f:35:14:78
[ 2627.543407] phy0: Destroyed STA 00:1b:2f:35:14:78
[ 2776.444279] ADDRCONF(NETDEV_UP): wlan0: link is not ready
[ 2776.659861] wlan0: deauthenticating by local choice (reason=3)
[ 2776.831004] ADDRCONF(NETDEV_UP): wlan0: link is not ready

When I turn power management off again, wpa_supplicant is able to
reconnect just fine.
Is this a known bug? If ath9k doesn't support power management
properly, wouldn't it be better to reject the userspace request right
away?

Thanks,
Davide

Re: ath9k and power management

[Kalle Valo]

Davide Pesavento <davidepesa@gmail.com> writes:

> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
> (with -Dnl80211).
> Enabling power management when connected (i.e. 'iwconfig wlan0 power
> on'), causes the currently established connection to break and the
> machine is no longer able to reconnect to the AP.

Try with a timeout, for example three seconds:

iwconfig wlan0 power timeout 3

It might be that ath9k doesn't work when timeout is zero (which 'power
on' effectively does). You need a fairly new wireless-tools to set the
timeout, older versions had a bug with that.

-- 
Kalle Valo

Re: ath9k and power management

[Davide Pesavento]

On Fri, Mar 27, 2009 at 15:42, Kalle Valo <kalle.valo@iki.fi> wrote:
> Davide Pesavento <davidepesa@gmail.com> writes:
>
>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
>> (with -Dnl80211).
>> Enabling power management when connected (i.e. 'iwconfig wlan0 power
>> on'), causes the currently established connection to break and the
>> machine is no longer able to reconnect to the AP.
>
> Try with a timeout, for example three seconds:
>
> iwconfig wlan0 power timeout 3
>
> It might be that ath9k doesn't work when timeout is zero (which 'power
> on' effectively does). You need a fairly new wireless-tools to set the
> timeout, older versions had a bug with that.
>

What do you mean by "fairly new"? I have wireless-tools version 29
installed, which is the latest stable. Do I need to use version 30
beta?

Thanks,
Davide

Re: ath9k and power management

[Davide Pesavento]

On Fri, Mar 27, 2009 at 15:54, Davide Pesavento <davidepesa@gmail.com> wrote:
> On Fri, Mar 27, 2009 at 15:42, Kalle Valo <kalle.valo@iki.fi> wrote:
>> Davide Pesavento <davidepesa@gmail.com> writes:
>>
>>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
>>> (with -Dnl80211).
>>> Enabling power management when connected (i.e. 'iwconfig wlan0 power
>>> on'), causes the currently established connection to break and the
>>> machine is no longer able to reconnect to the AP.
>>
>> Try with a timeout, for example three seconds:
>>
>> iwconfig wlan0 power timeout 3
>>
>> It might be that ath9k doesn't work when timeout is zero (which 'power
>> on' effectively does). You need a fairly new wireless-tools to set the
>> timeout, older versions had a bug with that.
>>
>
> What do you mean by "fairly new"? I have wireless-tools version 29
> installed, which is the latest stable. Do I need to use version 30
> beta?
>

Nevermind, I guess I need the beta version... :-)

# iwconfig wlan0 power timeout 3
Error for wireless request "Set Power Management" (8B2C) :
    invalid argument "3".

Regards,
Davide

Re: ath9k and power management

[Kalle Valo]

Davide Pesavento <davidepesa@gmail.com> writes:

>>> It might be that ath9k doesn't work when timeout is zero (which 'power
>>> on' effectively does). You need a fairly new wireless-tools to set the
>>> timeout, older versions had a bug with that.
>>>
>>
>> What do you mean by "fairly new"? I have wireless-tools version 29
>> installed, which is the latest stable. Do I need to use version 30
>> beta?
>>
>
> Nevermind, I guess I need the beta version... :-)
>
> # iwconfig wlan0 power timeout 3
> Error for wireless request "Set Power Management" (8B2C) :
>     invalid argument "3".

Yeah, you need the beta version. I used version 30-pre7 myself.

-- 
Kalle Valo

Re: ath9k and power management

[Davide Pesavento]

On Fri, Mar 27, 2009 at 15:42, Kalle Valo <kalle.valo@iki.fi> wrote:
> Davide Pesavento <davidepesa@gmail.com> writes:
>
>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
>> (with -Dnl80211).
>> Enabling power management when connected (i.e. 'iwconfig wlan0 power
>> on'), causes the currently established connection to break and the
>> machine is no longer able to reconnect to the AP.
>
> Try with a timeout, for example three seconds:
>
> iwconfig wlan0 power timeout 3
>
> It might be that ath9k doesn't work when timeout is zero (which 'power
> on' effectively does). You need a fairly new wireless-tools to set the
> timeout, older versions had a bug with that.
>

Ok, I installed wireless-tools 30-pre8 and tried with a timeout.
The results are more or less the same as before...

< iwconfig wlan0 power timeout 3 >

[ 2315.293198] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2321.296532] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2327.299864] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2333.303208] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2335.303190] wlan0: no probe response from AP 00:1b:2f:58:30:76 -
disassociating
[ 2335.303431] phy0: Removed STA 00:1b:2f:58:30:76
[ 2335.333248] phy0: Destroyed STA 00:1b:2f:58:30:76
[ 2388.465095] ADDRCONF(NETDEV_UP): wlan0: link is not ready
[ 2388.669843] wlan0: deauthenticating by local choice (reason=3)
[ 2388.840995] ADDRCONF(NETDEV_UP): wlan0: link is not ready

< manual restart of wpa_supplicant >

[ 2392.979891] wlan0: dropped data frame to not associated station
00:00:00:00:00:00
[ 2403.113236] wlan0: dropped data frame to not associated station
00:00:00:00:00:00
[ 2403.705107] wlan0: authenticate with AP 00:1b:2f:58:30:76
[ 2403.707389] wlan0: authenticated
[ 2403.707395] wlan0: associate with AP 00:1b:2f:58:30:76
[ 2403.710568] wlan0: RX AssocResp from 00:1b:2f:58:30:76 (capab=0x431
status=0 aid=1)
[ 2403.710574] wlan0: associated
[ 2403.710603] phy0: Allocated STA 00:1b:2f:58:30:76
[ 2403.710622] phy0: Inserted STA 00:1b:2f:58:30:76
[ 2403.710628] wmaster0: WMM queue=2 aci=0 acm=0 aifs=3 cWmin=15
cWmax=1023 txop=0
[ 2403.710643] wmaster0: WMM queue=3 aci=1 acm=0 aifs=7 cWmin=15
cWmax=1023 txop=0
[ 2403.710656] wmaster0: WMM queue=1 aci=2 acm=0 aifs=2 cWmin=7 cWmax=15 txop=94
[ 2403.710668] wmaster0: WMM queue=0 aci=3 acm=0 aifs=2 cWmin=3 cWmax=7 txop=47
[ 2403.710685] wlan0: switched to short barker preamble
(BSSID=00:1b:2f:58:30:76)
[ 2403.710690] wlan0: switched to short slot time (BSSID=00:1b:2f:58:30:76)
[ 2403.711356] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
[ 2408.535506] ath9k: rx failed to go idle in 10 ms RXSM=0xdeadbeef
[ 2414.393163] wlan0: no IPv6 routers present
[ 2421.536523] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2427.539851] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2433.543195] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2439.546535] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2451.413267] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2461.416589] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[ 2467.419835] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[...]

Regards,
Davide

Re: ath9k and power management

[Luis R. Rodriguez]

On Fri, Mar 27, 2009 at 12:08 PM, Davide Pesavento <davidepesa@gmail.com> wrote:
> On Fri, Mar 27, 2009 at 15:42, Kalle Valo <kalle.valo@iki.fi> wrote:
>> Davide Pesavento <davidepesa@gmail.com> writes:
>>
>>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
>>> (with -Dnl80211).
>>> Enabling power management when connected (i.e. 'iwconfig wlan0 power
>>> on'), causes the currently established connection to break and the
>>> machine is no longer able to reconnect to the AP.
>>
>> Try with a timeout, for example three seconds:
>>
>> iwconfig wlan0 power timeout 3
>>
>> It might be that ath9k doesn't work when timeout is zero (which 'power
>> on' effectively does). You need a fairly new wireless-tools to set the
>> timeout, older versions had a bug with that.
>>
>
> Ok, I installed wireless-tools 30-pre8 and tried with a timeout.
> The results are more or less the same as before...
>
> < iwconfig wlan0 power timeout 3 >
>
> [ 2315.293198] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2321.296532] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2327.299864] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2333.303208] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2335.303190] wlan0: no probe response from AP 00:1b:2f:58:30:76 -
> disassociating
> [ 2335.303431] phy0: Removed STA 00:1b:2f:58:30:76
> [ 2335.333248] phy0: Destroyed STA 00:1b:2f:58:30:76
> [ 2388.465095] ADDRCONF(NETDEV_UP): wlan0: link is not ready
> [ 2388.669843] wlan0: deauthenticating by local choice (reason=3)
> [ 2388.840995] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>
> < manual restart of wpa_supplicant >
>
> [ 2392.979891] wlan0: dropped data frame to not associated station
> 00:00:00:00:00:00
> [ 2403.113236] wlan0: dropped data frame to not associated station
> 00:00:00:00:00:00
> [ 2403.705107] wlan0: authenticate with AP 00:1b:2f:58:30:76
> [ 2403.707389] wlan0: authenticated
> [ 2403.707395] wlan0: associate with AP 00:1b:2f:58:30:76
> [ 2403.710568] wlan0: RX AssocResp from 00:1b:2f:58:30:76 (capab=0x431
> status=0 aid=1)
> [ 2403.710574] wlan0: associated
> [ 2403.710603] phy0: Allocated STA 00:1b:2f:58:30:76
> [ 2403.710622] phy0: Inserted STA 00:1b:2f:58:30:76
> [ 2403.710628] wmaster0: WMM queue=2 aci=0 acm=0 aifs=3 cWmin=15
> cWmax=1023 txop=0
> [ 2403.710643] wmaster0: WMM queue=3 aci=1 acm=0 aifs=7 cWmin=15
> cWmax=1023 txop=0
> [ 2403.710656] wmaster0: WMM queue=1 aci=2 acm=0 aifs=2 cWmin=7 cWmax=15 txop=94
> [ 2403.710668] wmaster0: WMM queue=0 aci=3 acm=0 aifs=2 cWmin=3 cWmax=7 txop=47
> [ 2403.710685] wlan0: switched to short barker preamble
> (BSSID=00:1b:2f:58:30:76)
> [ 2403.710690] wlan0: switched to short slot time (BSSID=00:1b:2f:58:30:76)
> [ 2403.711356] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
> [ 2408.535506] ath9k: rx failed to go idle in 10 ms RXSM=0xdeadbeef
> [ 2414.393163] wlan0: no IPv6 routers present
> [ 2421.536523] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2427.539851] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2433.543195] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2439.546535] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2451.413267] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2461.416589] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [ 2467.419835] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
> probe request
> [...]

Thanks for reporting this, it will be looked at internally.

  Luis

Re: ath9k and power management

[Vivek Natarajan]

On Sat, Mar 28, 2009 at 12:38 AM, Davide Pesavento <davidepesa@gmail.com> wrote:

>>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
>>> (with -Dnl80211).
>>> Enabling power management when connected (i.e. 'iwconfig wlan0 power
>>> on'), causes the currently established connection to break and the
>>> machine is no longer able to reconnect to the AP.

Patch has been sent to wireless-testing. Please report if the issue
still exists after
applying the patch.
The commit log of the patch is

ath9k: No need to abort Rx path when autosleep is enabled.
				
For chipsets supporting autosleep feature, there is no need to abort
Rx engine since they are capable of automatically going back to sleep
after receiving a packet.


Thanks a lot for reporting the issue
Vivek.

Re: ath9k and power management

[Davide Pesavento]

On Mon, Mar 30, 2009 at 11:38, Vivek Natarajan <vivek.natraj@gmail.com> wrote:
> On Sat, Mar 28, 2009 at 12:38 AM, Davide Pesavento <davidepesa@gmail.com> wrote:
>
>>>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
>>>> (with -Dnl80211).
>>>> Enabling power management when connected (i.e. 'iwconfig wlan0 power
>>>> on'), causes the currently established connection to break and the
>>>> machine is no longer able to reconnect to the AP.
>
> Patch has been sent to wireless-testing. Please report if the issue
> still exists after
> applying the patch.
> The commit log of the patch is
>
> ath9k: No need to abort Rx path when autosleep is enabled.
>
> For chipsets supporting autosleep feature, there is no need to abort
> Rx engine since they are capable of automatically going back to sleep
> after receiving a packet.
>
>
> Thanks a lot for reporting the issue
> Vivek.
>

I've applied your patch but unfortunately the behaviour is still the
same as before.

Thanks,
Davide

Re: ath9k and power management

[Vivek Natarajan]

On Mon, Mar 30, 2009 at 7:01 PM, Davide Pesavento <davidepesa@gmail.com> wrote:
> On Mon, Mar 30, 2009 at 11:38, Vivek Natarajan <vivek.natraj@gmail.com> wrote:
>>
>> ath9k: No need to abort Rx path when autosleep is enabled.
>>
>> For chipsets supporting autosleep feature, there is no need to abort
>> Rx engine since they are capable of automatically going back to sleep
>> after receiving a packet.
>
> I've applied your patch but unfortunately the behaviour is still the
> same as before.

One more patch sent:
ath9k: Disable autosleep feature for AR9285 based chipsets.

I hope you are using one of the AR9285 based. I have unit tested and it seems to
work consistently now.

Vivek.

Re: ath9k and power management

[Davide Pesavento]

On Tue, Mar 31, 2009 at 08:54, Vivek Natarajan <vivek.natraj@gmail.com> wrote:
> On Mon, Mar 30, 2009 at 7:01 PM, Davide Pesavento <davidepesa@gmail.com> wrote:
>> On Mon, Mar 30, 2009 at 11:38, Vivek Natarajan <vivek.natraj@gmail.com> wrote:
>>>
>>> ath9k: No need to abort Rx path when autosleep is enabled.
>>>
>>> For chipsets supporting autosleep feature, there is no need to abort
>>> Rx engine since they are capable of automatically going back to sleep
>>> after receiving a packet.
>>
>> I've applied your patch but unfortunately the behaviour is still the
>> same as before.
>
> One more patch sent:
> ath9k: Disable autosleep feature for AR9285 based chipsets.
>
> I hope you are using one of the AR9285 based. I have unit tested and it seems to
> work consistently now.
>

No, I'm not. Sorry for not reporting the hardware type earlier. My
card is identified as:
phy0: Atheros AR5418 MAC/BB Rev:2 AR5133 RF Rev:81:
mem=0xffffc20005020000, irq=17

Thanks,
Davide

Re: ath9k and power management

[Davide Pesavento]

Any progress on this?

Thanks,
Davide

Re: ath9k and power management

[Davide Pesavento]

On Fri, Mar 27, 2009 at 16:27, Davide Pesavento <davidepesa@gmail.com> wrote:
> Hello,
>
> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
> (with -Dnl80211).
> Enabling power management when connected (i.e. 'iwconfig wlan0 power
> on'), causes the currently established connection to break and the
> machine is no longer able to reconnect to the AP. Right after enabling
> pm, dmesg says:
>
> [ 2625.513233] wlan0: beacon loss from AP 00:1b:2f:35:14:78 - sending
> probe request
> [ 2627.513210] wlan0: no probe response from AP 00:1b:2f:35:14:78 -
> disassociating
> [ 2627.514430] phy0: Removed STA 00:1b:2f:35:14:78
> [ 2627.543407] phy0: Destroyed STA 00:1b:2f:35:14:78
> [ 2776.444279] ADDRCONF(NETDEV_UP): wlan0: link is not ready
> [ 2776.659861] wlan0: deauthenticating by local choice (reason=3)
> [ 2776.831004] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>
> When I turn power management off again, wpa_supplicant is able to
> reconnect just fine.
> Is this a known bug? If ath9k doesn't support power management
> properly, wouldn't it be better to reject the userspace request right
> away?
>

This is still unfixed on latest git (2.6.30-rc2-wl). The symptoms are
exactly the same as reported above.

Regards,
Davide

Re: ath9k and power management

[Vivek Natarajan]

On Sat, Apr 18, 2009 at 1:27 AM, Davide Pesavento <davidepesa@gmail.com> wrote:
> On Fri, Mar 27, 2009 at 16:27, Davide Pesavento <davidepesa@gmail.com> wrote:
>> Hello,
>>
>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.9
>> (with -Dnl80211).
>> Enabling power management when connected (i.e. 'iwconfig wlan0 power
>> on'), causes the currently established connection to break and the
>> machine is no longer able to reconnect to the AP. Right after enabling
>> pm, dmesg says:
>>
>> [ 2625.513233] wlan0: beacon loss from AP 00:1b:2f:35:14:78 - sending
>> probe request
>> [ 2627.513210] wlan0: no probe response from AP 00:1b:2f:35:14:78 -
>> disassociating
>> [ 2627.514430] phy0: Removed STA 00:1b:2f:35:14:78
>> [ 2627.543407] phy0: Destroyed STA 00:1b:2f:35:14:78
>> [ 2776.444279] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>> [ 2776.659861] wlan0: deauthenticating by local choice (reason=3)
>> [ 2776.831004] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>>
>> When I turn power management off again, wpa_supplicant is able to
>> reconnect just fine.
>> Is this a known bug? If ath9k doesn't support power management
>> properly, wouldn't it be better to reject the userspace request right
>> away?
>>
>
> This is still unfixed on latest git (2.6.30-rc2-wl). The symptoms are
> exactly the same as reported above.

This is the power save mechanism used in ath9k:
An interrupt is triggered just 4ms before the target beacon time
so that the chip can wake up and receive the beacons. In all the
latest chipsets like AR92XX, this is working fine.I wonder why
AR54XX has some problems related to that.
A quick look into the debug logs shows me that this wake up interrupt
(TIM_TIMER) stops triggering after some time which leads to beacon
loss and disconnection.
Although there is an option of autosleep(in latest versions) where hw can
wakeup automatically and receive beacon instead of a sw trigger call,
the hw may sleep for a longer duration(sometimes)  and we may need to
have some timer mechanism to verify if the hw is hung or not.
I did not have much time to look into this somewhat older chipsets and
I will try to debug more if I manage to get some free time.

Thanks
Vivek

Re: ath9k and power management

[Davide Pesavento]

On Sat, Apr 18, 2009 at 04:48, Vivek Natarajan <vivek.natraj@gmail.com>=
 wrote:
> On Sat, Apr 18, 2009 at 1:27 AM, Davide Pesavento <davidepesa@gmail.c=
om> wrote:
>> On Fri, Mar 27, 2009 at 16:27, Davide Pesavento <davidepesa@gmail.co=
m> wrote:
>>> Hello,
>>>
>>> I'm running a 2.6.29-wl kernel, using ath9k and wpa_supplicant 0.6.=
9
>>> (with -Dnl80211).
>>> Enabling power management when connected (i.e. 'iwconfig wlan0 powe=
r
>>> on'), causes the currently established connection to break and the
>>> machine is no longer able to reconnect to the AP. Right after enabl=
ing
>>> pm, dmesg says:
>>>
>>> [ 2625.513233] wlan0: beacon loss from AP 00:1b:2f:35:14:78 - sendi=
ng
>>> probe request
>>> [ 2627.513210] wlan0: no probe response from AP 00:1b:2f:35:14:78 -
>>> disassociating
>>> [ 2627.514430] phy0: Removed STA 00:1b:2f:35:14:78
>>> [ 2627.543407] phy0: Destroyed STA 00:1b:2f:35:14:78
>>> [ 2776.444279] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>>> [ 2776.659861] wlan0: deauthenticating by local choice (reason=3D3)
>>> [ 2776.831004] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>>>
>>> When I turn power management off again, wpa_supplicant is able to
>>> reconnect just fine.
>>> Is this a known bug? If ath9k doesn't support power management
>>> properly, wouldn't it be better to reject the userspace request rig=
ht
>>> away?
>>>
>>
>> This is still unfixed on latest git (2.6.30-rc2-wl). The symptoms ar=
e
>> exactly the same as reported above.
>
> This is the power save mechanism used in ath9k:
> An interrupt is triggered just 4ms before the target beacon time
> so that the chip can wake up and receive the beacons. In all the
> latest chipsets like AR92XX, this is working fine.I wonder why
> AR54XX has some problems related to that.
> A quick look into the debug logs shows me that this wake up interrupt
> (TIM_TIMER) stops triggering after some time which leads to beacon
> loss and disconnection.
> Although there is an option of autosleep(in latest versions) where hw=
 can
> wakeup automatically and receive beacon instead of a sw trigger call,
> the hw may sleep for a longer duration(sometimes) =C2=A0and we may ne=
ed to
> have some timer mechanism to verify if the hw is hung or not.
> I did not have much time to look into this somewhat older chipsets an=
d
> I will try to debug more if I manage to get some free time.
>
> Thanks
> Vivek
>

I did some more testing with ATH_DBG_RESET and ATH_DBG_BEACON enabled.
Here is dmesg output:

< first I try to connect with PS disabled >

[17772.421740] ath9k: AWAKE -> AWAKE
[17772.658457] ath9k: AWAKE -> AWAKE
[17772.659600] ath9k: AWAKE -> AWAKE
[17772.661119] ath9k: ah->misc_mode 0x4
[17772.662783] ath9k: AWAKE -> AWAKE
[17772.663840] ath9k: AWAKE -> AWAKE
[17772.664968] ath9k: AWAKE -> AWAKE
[17772.666484] ath9k: ah->misc_mode 0x4
[17772.667711] wlan0: authenticate with AP 00:1b:2f:58:30:76
[17772.669950] wlan0: authenticated
[17772.669954] wlan0: associate with AP 00:1b:2f:58:30:76
[17772.673211] wlan0: RX AssocResp from 00:1b:2f:58:30:76 (capab=3D0x43=
1
status=3D0 aid=3D2)
[17772.673215] wlan0: associated
[17772.673233] phy1: Allocated STA 00:1b:2f:58:30:76
[17772.673242] phy1: Inserted STA 00:1b:2f:58:30:76
[17772.673246] wmaster0: WMM queue=3D2 aci=3D0 acm=3D0 aifs=3D3 cWmin=3D=
15
cWmax=3D1023 txop=3D0
[17772.673256] wmaster0: WMM queue=3D3 aci=3D1 acm=3D0 aifs=3D7 cWmin=3D=
15
cWmax=3D1023 txop=3D0
[17772.673265] wmaster0: WMM queue=3D1 aci=3D2 acm=3D0 aifs=3D2 cWmin=3D=
7 cWmax=3D15 txop=3D94
[17772.673273] wmaster0: WMM queue=3D0 aci=3D3 acm=3D0 aifs=3D2 cWmin=3D=
3 cWmax=3D7 txop=3D47
[17772.673283] wlan0: switched to short barker preamble
(BSSID=3D00:1b:2f:58:30:76)
[17772.673286] wlan0: switched to short slot time (BSSID=3D00:1b:2f:58:=
30:76)
[17772.673312] ath9k: tsf: 5595 tsftu: 7
[17772.673315] ath9k: bmiss: 10 sleep: 100 cfp-period: 10000 maxdur: 0 =
next: 100
[17772.673333] ath9k: next DTIM 100
[17772.673336] ath9k: next beacon 100
[17772.673338] ath9k: beacon period 100
[17772.673341] ath9k: DTIM period 10000
[17772.673598] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready

< ok, wpa_supplicant is connected >
< now I execute `iwconfig wlan0 power timeout 1` >

[17836.991721] ath9k: AWAKE -> NETWORK SLEEP
[17836.991742] ath9k: NETWORK SLEEP -> AWAKE
[17836.991838] ath9k: AWAKE -> AWAKE
[17837.011506] ath9k: AWAKE -> AWAKE
[17837.113945] ath9k: AWAKE -> AWAKE
[17837.216346] ath9k: AWAKE -> AWAKE
[17837.318753] ath9k: AWAKE -> AWAKE
[17837.410312] ath9k: AWAKE -> NETWORK SLEEP
[17837.410340] ath9k: NETWORK SLEEP -> AWAKE
[17837.410414] ath9k: AWAKE -> NETWORK SLEEP
[17837.474938] ath9k: NETWORK SLEEP -> AWAKE
[17837.475021] ath9k: AWAKE -> NETWORK SLEEP
[17837.475061] ath9k: NETWORK SLEEP -> AWAKE
[17837.475134] ath9k: AWAKE -> NETWORK SLEEP
[17837.571609] ath9k: NETWORK SLEEP -> AWAKE
[17837.571689] ath9k: AWAKE -> NETWORK SLEEP
[17837.671610] ath9k: NETWORK SLEEP -> AWAKE
[17837.671690] ath9k: AWAKE -> NETWORK SLEEP
[17837.671731] ath9k: NETWORK SLEEP -> AWAKE
[17837.671803] ath9k: AWAKE -> NETWORK SLEEP
[17837.774920] ath9k: NETWORK SLEEP -> AWAKE
[17837.774997] ath9k: AWAKE -> NETWORK SLEEP
[17837.871610] ath9k: NETWORK SLEEP -> AWAKE
[17837.871694] ath9k: AWAKE -> NETWORK SLEEP
[17837.971618] ath9k: NETWORK SLEEP -> AWAKE
[17837.971701] ath9k: AWAKE -> NETWORK SLEEP
[17838.071609] ath9k: NETWORK SLEEP -> AWAKE
[17838.071689] ath9k: AWAKE -> NETWORK SLEEP
[17838.171609] ath9k: NETWORK SLEEP -> AWAKE
[17838.171691] ath9k: AWAKE -> NETWORK SLEEP
[17838.271611] ath9k: NETWORK SLEEP -> AWAKE
[17838.271694] ath9k: AWAKE -> NETWORK SLEEP
[17838.271713] ath9k: NETWORK SLEEP -> AWAKE
[17838.271786] ath9k: AWAKE -> NETWORK SLEEP
[17838.271805] ath9k: NETWORK SLEEP -> AWAKE

< this bouncing between power states goes on for a bit more, until... >

[17840.684940] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request
[17840.684984] ath9k: NETWORK SLEEP -> AWAKE
[17841.685023] ath9k: AWAKE -> AWAKE
[17841.685127] ath9k: AWAKE -> NETWORK SLEEP
[17841.687395] ath9k: NETWORK SLEEP -> AWAKE
[17841.722144] ath9k: AWAKE -> AWAKE
[17841.824534] ath9k: AWAKE -> AWAKE
[17841.926945] ath9k: AWAKE -> AWAKE
[17842.018507] ath9k: AWAKE -> NETWORK SLEEP
[17842.029339] ath9k: NETWORK SLEEP -> AWAKE
[17842.029428] ath9k: AWAKE -> AWAKE
[17842.131762] ath9k: AWAKE -> AWAKE
[17842.234162] ath9k: AWAKE -> AWAKE
[17842.336567] ath9k: AWAKE -> AWAKE
[17842.438969] ath9k: AWAKE -> AWAKE
[17842.530529] ath9k: AWAKE -> NETWORK SLEEP
[17842.530556] ath9k: NETWORK SLEEP -> AWAKE
[17842.530629] ath9k: AWAKE -> NETWORK SLEEP
[17842.571607] ath9k: NETWORK SLEEP -> AWAKE
[17842.571690] ath9k: AWAKE -> NETWORK SLEEP
[17842.571730] ath9k: NETWORK SLEEP -> AWAKE
[17842.571802] ath9k: AWAKE -> NETWORK SLEEP
[17842.671606] ath9k: NETWORK SLEEP -> AWAKE
[17842.671686] ath9k: AWAKE -> NETWORK SLEEP
[17842.671712] ath9k: NETWORK SLEEP -> AWAKE
[17842.671784] ath9k: AWAKE -> NETWORK SLEEP
[17842.771617] ath9k: NETWORK SLEEP -> AWAKE
[17842.771699] ath9k: AWAKE -> NETWORK SLEEP
[17842.771724] ath9k: NETWORK SLEEP -> AWAKE
[17842.771797] ath9k: AWAKE -> NETWORK SLEEP
[17842.871608] ath9k: NETWORK SLEEP -> AWAKE
[17842.871690] ath9k: AWAKE -> NETWORK SLEEP

< and then the whole thing starts over again for another couple of
times, with the hardware rapidly alternating its power state, until it
disconnects >

[17852.691611] wlan0: no probe response from AP 00:1b:2f:58:30:76 -
disassociating
[17852.691695] ath9k: AWAKE -> AWAKE
[17852.691782] phy1: Removed STA 00:1b:2f:58:30:76
[17852.731944] phy1: Destroyed STA 00:1b:2f:58:30:76
[17852.821730] ath9k: AWAKE -> AWAKE
[17852.888413] ath9k: AWAKE -> AWAKE
[17852.955081] ath9k: AWAKE -> AWAKE
[17853.021725] ath9k: AWAKE -> AWAKE
[17853.088395] ath9k: AWAKE -> AWAKE
[17853.155083] ath9k: AWAKE -> AWAKE
[17853.221726] ath9k: AWAKE -> AWAKE
[ ...above message repeated many times... ]
[17857.517104] ath9k: AWAKE -> NETWORK SLEEP
[17857.517117] ath9k: NETWORK SLEEP -> AWAKE
[17857.517280] ath9k: AWAKE -> AWAKE
[17857.518407] ath9k: AWAKE -> AWAKE
[17857.519923] ath9k: ah->misc_mode 0x4
[17857.521323] ath9k: AWAKE -> NETWORK SLEEP
[17857.521336] ath9k: NETWORK SLEEP -> AWAKE
[17857.521405] ath9k: AWAKE -> NETWORK SLEEP
[17857.521424] ath9k: NETWORK SLEEP -> AWAKE
[17857.521493] ath9k: AWAKE -> NETWORK SLEEP
[17857.521509] wlan0: authenticate with AP 00:1b:2f:58:30:76
[17857.521527] ath9k: NETWORK SLEEP -> AWAKE
[17857.522564] ath9k: AWAKE -> AWAKE
[17857.523684] ath9k: AWAKE -> AWAKE
[17857.525216] ath9k: ah->misc_mode 0x4
[17857.526329] ath9k: AWAKE -> AWAKE
[17857.527450] ath9k: AWAKE -> AWAKE
[17857.528972] ath9k: ah->misc_mode 0x4
[17857.529901] wlan0: authenticate with AP 00:1b:2f:58:30:76
[17857.532614] wlan0: authenticated
[17857.532618] wlan0: associate with AP 00:1b:2f:58:30:76
[17857.535619] wlan0: deauthenticated (Reason: 6)
[17858.534801] wlan0: dropped data frame to not associated station
00:1b:2f:58:30:76
[17858.534838] ath9k: AWAKE -> NETWORK SLEEP
[17858.534853] wlan0: direct probe to AP 00:1b:2f:58:30:76 try 1
[17858.534879] ath9k: NETWORK SLEEP -> AWAKE
[17858.538389] wlan0 direct probe responded
[17858.538394] wlan0: authenticate with AP 00:1b:2f:58:30:76
[17858.540608] wlan0: authenticated
[17858.540612] wlan0: associate with AP 00:1b:2f:58:30:76
[17858.543844] wlan0: RX ReassocResp from 00:1b:2f:58:30:76
(capab=3D0x431 status=3D0 aid=3D2)
[17858.543851] wlan0: associated
[17858.543885] phy1: Allocated STA 00:1b:2f:58:30:76
[17858.543897] phy1: Inserted STA 00:1b:2f:58:30:76
[17858.543900] wmaster0: WMM queue=3D2 aci=3D0 acm=3D0 aifs=3D3 cWmin=3D=
15
cWmax=3D1023 txop=3D0
[17858.543913] wmaster0: WMM queue=3D3 aci=3D1 acm=3D0 aifs=3D7 cWmin=3D=
15
cWmax=3D1023 txop=3D0
[17858.543922] wmaster0: WMM queue=3D1 aci=3D2 acm=3D0 aifs=3D2 cWmin=3D=
7 cWmax=3D15 txop=3D94
[17858.543931] wmaster0: WMM queue=3D0 aci=3D3 acm=3D0 aifs=3D2 cWmin=3D=
3 cWmax=3D7 txop=3D47
[17858.543944] wlan0: switched to short barker preamble
(BSSID=3D00:1b:2f:58:30:76)
[17858.543946] wlan0: switched to short slot time (BSSID=3D00:1b:2f:58:=
30:76)
[17858.544358] ath9k: tsf: 18236047462 tsftu: 17808642
[17858.544365] ath9k: bmiss: 10 sleep: 100 cfp-period: 10000 maxdur: 0
next: 17810100
[17858.544384] ath9k: next DTIM 17810100
[17858.544387] ath9k: next beacon 17808700
[17858.544390] ath9k: beacon period 100
[17858.544393] ath9k: DTIM period 10000
[17859.551695] ath9k: AWAKE -> AWAKE
[17859.551803] ath9k: AWAKE -> NETWORK SLEEP
[17859.553670] ath9k: NETWORK SLEEP -> AWAKE
[17859.626507] ath9k: AWAKE -> AWAKE
[17862.241612] ath9k: NETWORK SLEEP -> AWAKE
[17862.241693] ath9k: AWAKE -> NETWORK SLEEP
[17862.241719] ath9k: NETWORK SLEEP -> AWAKE
[17862.241792] ath9k: AWAKE -> NETWORK SLEEP
[17862.341627] ath9k: NETWORK SLEEP -> AWAKE
[17862.341710] ath9k: AWAKE -> NETWORK SLEEP
[17862.341732] ath9k: NETWORK SLEEP -> AWAKE
[17862.341805] ath9k: AWAKE -> NETWORK SLEEP
[17862.441623] ath9k: NETWORK SLEEP -> AWAKE
[17862.441703] ath9k: AWAKE -> NETWORK SLEEP
[17862.541625] ath9k: NETWORK SLEEP -> AWAKE
[ ...above messages repeated many times... ]
[17862.541706] ath9k: AWAKE -> NETWORK SLEEP
[17862.541727] ath9k: NETWORK SLEEP -> AWAKE
[17862.541800] ath9k: AWAKE -> NETWORK SLEEP
[17862.541818] ath9k: NETWORK SLEEP -> AWAKE
[17862.541900] ath9k: AWAKE -> NETWORK SLEEP
[17862.554921] wlan0: beacon loss from AP 00:1b:2f:58:30:76 - sending
probe request

< it is able to reconnect just for a moment, but immediately after
that it fails again in the same way >

Hope it might be useful!

Regards,
Davide
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html