[PATCH] hostapd: take configured antenna set into consideration

Ben Greear greearb at candelatech.com
Fri Jan 22 08:06:06 PST 2016

On 01/22/2016 07:25 AM, Felix Fietkau wrote:
> On 2016-01-22 15:57, Ben Greear wrote:
>> On 01/22/2016 03:13 AM, Felix Fietkau wrote:
>>> On 2015-12-09 21:20, greearb at candelatech.com wrote:
>>>> From: Ben Greear <greearb at candelatech.com>
>>>> If user has configured TX antennas to be less than what
>>>> hardware advertises, the MCS reported by hardware will
>>>> be too large.  So, remove MCS sets accordingly.
>>>> Signed-off-by: Ben Greear <greearb at candelatech.com>
>>> Why is it necessary to do this in user space? Shouldn't the driver
>>> recalculate its MCS capabilities and let hostapd use the updated
>>> information.
>>> IIRC that's what ath9k is doing at the moment.
>> I guess that could work too, but long term, I think hostapd should have
>> a lot more control over this anyway:  Maybe one vAP would like to be 1x1
>> and another 3x3 on the same radio (different vdevs) for instance.  I'd
>> use this feature for testing, but perhaps others could take advantage
>> of it in other ways.
> I don't think that makes sense, and it doesn't really match up with
> hardware capabilities either. You can't just enable/disable antennas per
> VAP. And if you just want to advertise different MCS capabilities for
> testing, then antenna configuration is a pretty bad and hackish way to
> achieve that goal.

I meant more that I wanted to over-ride the MCS that an AP advertised, so
one vAP might advertise MCS of 1x1 and another might be 3x3.  Underlying
radio would be set to 3x3.

In the end, it would emulate vAP with different NSS, just implemented

Assuming rate-ctrl was properly tied into hostapd and/or configured
elsewhere, I think it could work fine.  Similar to how the MCS overrides
and such works for wpa_supplicant.

>> I haven't tested ath9k in this scenario...have you been able to confirm
>> that the beacons show proper MCS values with stock hostapd when you lower
>> the NSS?
> Yes, I just tested it with the current code and the MCS information is
> correct. Running iw phy0 info will also show reduced HT capabilities.

I guess I still like to make hostapd act as I patched it, but sounds like
someone should fix ath10k as well.  Maybe I'll find some motivation to do
that soon.


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

More information about the Hostap mailing list