Re: [PATCH 00/10] mac80211 patches from our internal tree 2017-08-05

[Kalle Valo <kvalo@codeaurora.org>]

Luca Coelho <luca@coelho.fi> writes:

> On Mon, 2017-08-07 at 12:06 +0300, Kalle Valo wrote:
>> Luca Coelho <luca@coelho.fi> writes:
>> 
>> > Here are some pending mac80211 patches from our internal tree.
>> > 
>> > The "mac80211: add api to start ba session timer expired flow" patch
>> > is needed by an iwlwifi patch that I want to send for -fixes, so it
>> > would have to be applied to -fixes as well.
>> 
>> We are getting to the later stages of the release cycle and I'm raising
>> the bar for wireless-drivers even higher. How serious iwlwifi bug is
>> that fixing?
>
> The problem is a bad degradation in throughput with our new 9000 family
> of devices because when aggregation times out we stop the aggregation
> internally in the firmware but don't send a delba.  Then on the AP side,
>  also with our driver, we don't handle the timeout as we should, so
> aggregations the devices get out of sync and BA sessions are not
> possible anymore, limiting our throughput to ~30Mbps, in some specific
> internal tests.

Ok.

>> > Should I send a separate patchset with these two so they can be both
>> > applied at the same time (either in the mac80211 or in
>> > wireless-drivers tree)?
>> 
>> Now that Johannes is away and I'm taking any urgent mac80211 patches, I
>> think the best approach is that you include mac80211 patch in the same
>> patchset as the iwlwifi patches destined for wireless-drivers.
>
> Okay, if you think the (one-liner) iwlwifi driver fix is -rc'able, I'll
> resend this in a patchset including both changes.

If the iwlwifi patch is a oneliner it doesn't sound too bad, but I
reserve the right to change my mind :)

But try to submit the pull request in the next few days so that I can
submit the patches forward by end of this week. And I think it's easiest
that you apply the mac80211 patch directly to your tree and I just pull
from you.

-- 
Kalle Valo