Received-SPF: pass (google.com: best guess record for domain of linux-wireless-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Nicolas Cavallari <nicolas.cavallari@green-communications.fr>
To:     =?UTF-8?q?Krzysztof=20Ha=C5=82asa?= <khalasa@piap.pl>,
        Johannes Berg <johannes@sipsolutions.net>
Cc:     linux-wireless@vger.kernel.org, Kalle Valo <kvalo@codeaurora.org>
Subject: [RFC] Allow userspace to reset IBSS stations to fix aggregation issue
Date:   Mon,  9 Dec 2019 11:28:05 +0100
Message-Id: <20191209102808.12014-1-nicolas.cavallari@green-communications.fr>
In-Reply-To: <m34l02mh71.fsf@t19.piap.pl>
References: <m34l02mh71.fsf@t19.piap.pl>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk

I encountered the same issue in an IBSS-RSN network, where quick reboot
of a station would cause issues with aggregation because the kernel is
not aware of the reboot.

I figured out that since wpa_supplicant already detect reboots, the
simplest way to fix it would be for wpa_supplicant to reset the entire
state of the station in the kernel, instead of just resetting keys and
port.

This means extending NL80211_CMD_DEL_STATION to work in IBSS mode too.
My current wpa_supplicant patch tries this new API but fall back to the
old methods if an error is returned.

The changes required for drivers would be as follow:
- mac80211: will just work
- wil6210: does not seem to support IBSS mode
- ath6kl: need a patch
- brcmfmac: need a patch
- mwifiex: need a patch
- qtnfmac: no IBSS support
- staging/wilc1000: no IBSS support
- staging/rtl8723bs: already returns EINVAL in this case.

I made patches for ath6kl, brcmfmac and mwifiex that plainly reject
the request. If there is a better way, i'll glad to hear it.