Any reason to allow MGT TID to be paused by power-save?

Any reason to allow MGT TID to be paused by power-save?

[Ben Greear]

I'm trying to track down a tricky problem with a Mac book and power save
with my ath10k firmware.

At least part of the issue is that when the Mac sleeps, it goes into power-save
state as far as AP is concerned, and then it simply turns itself off.  When it wakes
up a short time later, it then tries to scan and re-associate.  But, if hostapd
hasn't timed out the station (due to longer idle timer), then STA still is treated
as being in power-save in the firmware.  It appears that this causes mgt-frames (probe responses)
not be sent back out of the firmware, so Mac gives up fairly quickly.

I fixed at least some of this by forcing a power-save wake on
receipt of probe requests and auth requests, but some issues remain
for one reason or another.

One thing that came to mind:  Is it *ever* a good idea to pause a mgt-tid due to
power-save?  A second non-pausable TID also exists in the firmware...maybe all
mgt frames should use that?  But if so, then why have a mgt tid at all?

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com

Any reason to allow MGT TID to be paused by power-save?

[Ben Greear]

I'm trying to track down a tricky problem with a Mac book and power save
with my ath10k firmware.

At least part of the issue is that when the Mac sleeps, it goes into power-save
state as far as AP is concerned, and then it simply turns itself off.  When it wakes
up a short time later, it then tries to scan and re-associate.  But, if hostapd
hasn't timed out the station (due to longer idle timer), then STA still is treated
as being in power-save in the firmware.  It appears that this causes mgt-frames (probe responses)
not be sent back out of the firmware, so Mac gives up fairly quickly.

I fixed at least some of this by forcing a power-save wake on
receipt of probe requests and auth requests, but some issues remain
for one reason or another.

One thing that came to mind:  Is it *ever* a good idea to pause a mgt-tid due to
power-save?  A second non-pausable TID also exists in the firmware...maybe all
mgt frames should use that?  But if so, then why have a mgt tid at all?

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: Any reason to allow MGT TID to be paused by power-save?

[Michal Kazior]

On 10 June 2016 at 01:45, Ben Greear <greearb@candelatech.com> wrote:
> I'm trying to track down a tricky problem with a Mac book and power save
> with my ath10k firmware.
>
> At least part of the issue is that when the Mac sleeps, it goes into
> power-save
> state as far as AP is concerned, and then it simply turns itself off.  When
> it wakes
> up a short time later, it then tries to scan and re-associate.  But, if
> hostapd
> hasn't timed out the station (due to longer idle timer), then STA still is
> treated
> as being in power-save in the firmware.  It appears that this causes
> mgt-frames (probe responses)
> not be sent back out of the firmware, so Mac gives up fairly quickly.
>
> I fixed at least some of this by forcing a power-save wake on
> receipt of probe requests and auth requests, but some issues remain
> for one reason or another.
>
> One thing that came to mind:  Is it *ever* a good idea to pause a mgt-tid
> due to
> power-save?

The spec defines some mgmt frames as bufferable (IEEE 802.11-2012
8.2.4.1.7). See ieee80211_is_bufferable_mmpdu() in
include/linux/ieee80211.h.

Action, deauth and disassoc are bufferable. It makes sense anyway.

E.g. On DFS channel you may opt for CSA action frame (instead of CSA
IE in beacons) in which case you do want to buffer frames to clients
that are asleep to give them a chance to notice it, wake up and get
it.


Michał

Re: Any reason to allow MGT TID to be paused by power-save?

[Michal Kazior]

On 10 June 2016 at 01:45, Ben Greear <greearb@candelatech.com> wrote:
> I'm trying to track down a tricky problem with a Mac book and power save
> with my ath10k firmware.
>
> At least part of the issue is that when the Mac sleeps, it goes into
> power-save
> state as far as AP is concerned, and then it simply turns itself off.  When
> it wakes
> up a short time later, it then tries to scan and re-associate.  But, if
> hostapd
> hasn't timed out the station (due to longer idle timer), then STA still is
> treated
> as being in power-save in the firmware.  It appears that this causes
> mgt-frames (probe responses)
> not be sent back out of the firmware, so Mac gives up fairly quickly.
>
> I fixed at least some of this by forcing a power-save wake on
> receipt of probe requests and auth requests, but some issues remain
> for one reason or another.
>
> One thing that came to mind:  Is it *ever* a good idea to pause a mgt-tid
> due to
> power-save?

The spec defines some mgmt frames as bufferable (IEEE 802.11-2012
8.2.4.1.7). See ieee80211_is_bufferable_mmpdu() in
include/linux/ieee80211.h.

Action, deauth and disassoc are bufferable. It makes sense anyway.

E.g. On DFS channel you may opt for CSA action frame (instead of CSA
IE in beacons) in which case you do want to buffer frames to clients
that are asleep to give them a chance to notice it, wake up and get
it.


Michał

_______________________________________________
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k