Re: [RFC 00/11] multi-channel support

Re: [RFC 00/11] multi-channel support

[Johannes Berg]

Hi,

Thanks for the code! I've started to work on this as well, and will take
a look at your code soon. I'm currently a bit sick though so it might be
a couple of days.

> Work still needs to be done:
>  * powersave per-vif
>  * queue locking per-vif
>  * offchannel rework (hw_config, work_work)
>  * and a bit more

Have you my thoughts at
http://thread.gmane.org/gmane.linux.kernel.wireless.general/86070
about the queue issue?

> Questions:
> 
>  * monitor interfaces:
>    Currently ieee80211_set_channel gets netdev==NULL when iface is
>    a monitor. Is there a particular reason behind it?

This is my biggest question too. I need to sit down and think through
the different monitor mode use cases to be able to answer it.

>  * ieee80211_hw_config:
>    Should we extend it to take ieee80211_sub_if_data or should we
>    use ieee80211_bss_info_change_notify? If so, is ieee80211_hw_config
>    eventually to be removed?

That's the other question I've asked myself too :-)
I think extending it doesn't make a lot of sense. Most of the parameters
can be moved or removed, if they aren't already dead (like
IEEE80211_CONF_CHANGE_LISTEN_INTERVAL.)

POWER might be an interesting question -- not sure if all devices are
going to support per-vif TX power and if we should require that. Ditto
for SMPS. IDLE is already per interface in addition, and goes together
with MONITOR and how we handle monitor mode in the future (see above.)
That leaves just RETRY_LIMITS which, to be honest, I have no idea about
right now.

johannes

Re: [RFC 00/11] multi-channel support

[Michał Kazior]

Johannes Berg wrote:
> Hi,
>
> Thanks for the code! I've started to work on this as well, and will take
> a look at your code soon. I'm currently a bit sick though so it might be
> a couple of days.
>
>> Work still needs to be done:
>>   * powersave per-vif
>>   * queue locking per-vif
>>   * offchannel rework (hw_config, work_work)
>>   * and a bit more
>
> Have you my thoughts at
> http://thread.gmane.org/gmane.linux.kernel.wireless.general/86070
> about the queue issue?

Yes, I've read it. I did think about different approaches to handling 
queues but I decided to do the 'least changes' one for now. I have some 
patches for a simple splitting of tx pending stuff to per-vif and queue 
locking function vif variants. I can send them tomorrow once I clean 
them up.

It seems I failed to use msmtp correctly and my emails ended up only at 
your mailbox. I'll resend them to the mailing list alone. I'm sorry 
about that.


-- Pozdrawiam / Best Regards, Michał Kazior.