mesh point and ap mode on same interface

Arend van Spriel arend.vanspriel at
Wed Nov 15 02:36:49 PST 2017

On 11/15/2017 11:27 AM, Kalle Valo wrote:
> Arend van Spriel <arend.vanspriel at> writes:
>> On 11/15/2017 11:01 AM, Kalle Valo wrote:
>>> (adding linux-wireless for a cfg80211 question)
>>> Peter Oh <peter.oh at> writes:
>>>> On 11/08/2017 04:49 AM, Harms, Hannes wrote:
>>>>> I found a solution for the problem:
>>>>> When joining the mesh with beacon interval  set  to 100
>>>>> both  mesh and ap interface are coming up.
>>>>>> iw mesh0 mesh join my_mesh beacon-interval 100
>>>>> The default beacon interval is 1000 for mesh interfaces.
>>>>> It seems that QCA9880 is not working
>>>>> with different beacon intervals for ap + mesh.
>>>> Good to hear that working. But mesh beacon interval 1000 which is
>>>> default value should work. Any mesh beacon intervals of multiple of AP
>>>> beacon intervals are supposed to work.
>>>> (In other words, mesh beacon intervals that are not aligned by
>>>> multiple of AP beacon intervals won't work). These are valid for both
>>>> of QCA9880 and QCA4019 as far as I know.
>>> It would be nice if ath10k could print a warning if the beacon intervals
>>> are not compatible. Or is there a way to enforce the beacon interval
>>> with help of cfg80211 or mac80211?
>> There is beacon interval checking in utility function
>> cfg80211_check_combinations() ->
>> cfg80211_iter_combinations() -> cfg80211_calculate_bi_data() for that.
> Nice, that looks useful. Thanks for the tip.

Initally was added by:

commit 0c317a02ca982ca093e71bf07cb562265ba40032
Author: Purushottam Kushwaha <pkushwah at>
Date:   Wed Oct 12 18:26:51 2016 +0530

     cfg80211: support virtual interfaces with different beacon intervals

And last touched by Johannes:

commit 4c8dea638c16141adb046fd2e0cab51dfe43650c
Author: Johannes Berg <johannes.berg at>
Date:   Fri Oct 21 14:25:13 2016 +0200

     cfg80211: validate beacon int as part of iface combinations

The last commit message has some info on how to use in case of AP/mesh.

>> However, it seems brcmfmac is the only user of that function according
>> to lxr (or elixir).
> git-grep says the same.

Yeah. git-grep has not sunk in my system yet ;-)


More information about the ath10k mailing list