Re: EHT NSS units question

Ben Greear <[email protected]> Wed, 24 Jul 2024 06:45:47 -0700
Newsgroups org.netbsd.radiotap
Organization Candela Technologies
Message-ID <[email protected]>
On 7/24/24 03:58, Johannes Berg wrote:
> On Thu, 2024-07-18 at 06:58 -0700, Ben Greear wrote:
>> Hello,
>>
>> I would like to know if the units for NSS for EHT user element
>> are zero based (ie, 0x0 means 1NSS) or ones-based (0x1 means 1NSS).
>>
>>   From what I can tell, wireshark is using zero based counting, but
>> mt76 driver is using ones-based counting when filling in the radiotap header.
>>
>> I am not sure which one to fix...
>>
> 
> Looks like we never fully defined this ... Arguably having 4 bits means
> that we intended to have 1-8 (and perhaps 0 mean "unknown"), but having
> another bit for "I knew the NSS" kind of made that unnecessary ...
> 
> But since both wireshark and iwlwifi agree to have 0=1NSS and 1=2NSS I
> guess it's better to clarify that in the spec and fix mt76.

Ok thanks, I have a small patch that fixes mt76, will clean it up and post
to the wifi mailing list.

Thanks,
Ben

-- 
Ben Greear <[email protected]>
Candela Technologies Inc  http://www.candelatech.com