On 10/11/2018 1:05 PM, Johannes Berg wrote:
>
>> Thanks for the explanation. The send to same socket does sound more
>> efficient.
>> (In our internal implementation with vendor commands we were forced
>> to send the results as broadcast...)
>
> I suppose we can fix that, in the sense that we can add API to allow
> vendor commands to know the socket to send back to etc.
>
I think that would be useful in general though as far as I know we don't have
any pending feature that requires it right now.
[...]
>>
>> As I remember the driver/FW can easily find the connected channel for
>> connected
>> station. For unconnected station we should probably force this parameter.
>
> Should be able to, yes, but I suppose even userspace can. It just seems
> like extra complexity to impose on the driver, since cfg80211 doesn't
> necessarily have all the right information.
>
Agree with this. I guess we can keep the channel parameter mandatory.
> Actually, hmm, that would imply "use maximum bandwidth" or something?
> And then what if that bandwidth isn't possible with FTM for some reason?
> It's a bit tricky then.
>
DMG always uses the entire channel (2160MHz). We may need to use partial channel
in EDMG but not sure yet. We have an upcoming patch for EDMG support so we can
deal with this later.
>>
>> Good point. I see we currently use 20_NOHT for DMG, guess we can continue
>> using it.
>
> Well, I think it'd make more sense to just enforce the DMG bandwidth
> everywhere, but I won't force the issue over this.
>
Ok. I think using 20_NOHT for now is more consistent since other APIs specify it
for DMG channels.
Thanks,
Lior
