To:     "linux-wireless@vger.kernel.org" <linux-wireless@vger.kernel.org>
From:   Ben Greear <greearb@candelatech.com>
Subject: Block ACK in firmware and software decrypt
Message-ID: <934eb713-1f2e-3d06-e168-dbd7bfe83943@candelatech.com>
Date:   Fri, 8 Mar 2019 04:53:46 -0800
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101
 Thunderbird/45.8.0
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

I have run into a slightly tricky problem related to software decrypt
and block-ack handling in ath10k.

For testing purposes, I need to be able to run ath10k firmware in
software-decrypt mode, but the ath10k firmware also handles block
ack frames.  This becomes a problem when using PMF because then the
firmware cannot decrypt the block-ack setup frames and so blockack
fails.

But, the mac80211 stack can indeed decrypt management frames, so I
was thinking maybe I could have a logic flow that looked like this:

ath10k registers with mac80211 on startup and asks to be notified
   of certain management frames once they are decrypted (block-ack req, response)
ath10k sends encrypted frames up the stack to mac80211
mac80211 decrypts frames
mac80211 remembers ath10k wants to see decrypted mgt frames, and so sends
   decoded block-ack req and response frames back to ath10k driver.
ath10k driver use WMI messaging interface to send contents and meta data
   for the decoded block-ack frames to the firmware.
Firmware processes this as if it received it directly off the wire, letting
   block-ack setup work.

So, this is a special case that would only work in my ath10k driver and
firmware, but does anyone else see another reason to allow something like
this into the upstream kernel?

Thanks,
Ben

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