Message-ID: <a5ac66c2bee2f5de7cda68ad13d7008a05975c0b.camel@sipsolutions.net>
Subject: Re: [RFC v4] mac80211: fix rx blockack session race condition
From:   Johannes Berg <johannes@sipsolutions.net>
To:     Jean-Pierre TOSONI <jp.tosoni@acksys.fr>,
        "'linux-wireless@vger.kernel.org'" <linux-wireless@vger.kernel.org>
Date:   Fri, 26 Nov 2021 13:04:47 +0100
In-Reply-To: <DB9PR01MB73541FED9E91AC3005D27DAEE4869@DB9PR01MB7354.eurprd01.prod.exchangelabs.com>
References: <DB9PR01MB73541FED9E91AC3005D27DAEE4869@DB9PR01MB7354.eurprd01.prod.exchangelabs.com>
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.42.1 (3.42.1-1.fc35) 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Thu, 2021-10-28 at 10:25 +0000, Jean-Pierre TOSONI wrote:
 
> +		spin_lock_bh(&rx->sta->ampdu_mlme.rx_offl_lock);
>  		if (!test_bit(tid, rx->sta->ampdu_mlme.agg_session_valid) &&
> -		    !test_and_set_bit(tid, rx->sta->ampdu_mlme.unexpected_agg))
> +		    /* back_req is allowed if the fw just received addba */
> 

If we're going to add an unconditional lock here, I see little reason to
have all the test_bit()/test_and_set_bit() etc.

I really don't think it's _right_ to add an unconditional lock here
though, if it can at all be avoided.

johannes