Message-ID: <fc263550-13a6-4cfa-c0b6-c8b2f7bccf96@nbd.name>
Date:   Tue, 14 Mar 2023 22:55:07 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0)
 Gecko/20100101 Thunderbird/102.8.0
Subject: Re: [PATCH] wifi: mac80211: Serialize
 ieee80211_handle_wake_tx_queue()
Content-Language: en-US
To:     Alexander Wetzel <alexander@wetzel-home.de>,
        johannes@sipsolutions.net
Cc:     linux-wireless@vger.kernel.org, Thomas Mann <rauchwolke@gmx.net>,
        stable@vger.kernel.org
References: <20230314211122.111688-1-alexander@wetzel-home.de>
From:   Felix Fietkau <nbd@nbd.name>
In-Reply-To: <20230314211122.111688-1-alexander@wetzel-home.de>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 14.03.23 22:11, Alexander Wetzel wrote:
> ieee80211_handle_wake_tx_queue must not run concurrent multiple times.
> It calls ieee80211_txq_schedule_start() and the drivers migrated to iTXQ
> do not expect overlapping drv_tx() calls.
> 
> This fixes 'c850e31f79f0 ("wifi: mac80211: add internal handler for
> wake_tx_queue")', which introduced ieee80211_handle_wake_tx_queue.
> Drivers started to use it with 'a790cc3a4fad ("wifi: mac80211: add
> wake_tx_queue callback to drivers")'.
> But only after fixing an independent bug with
> '4444bc2116ae ("wifi: mac80211: Proper mark iTXQs for resumption")'
> problematic concurrent calls really happened and exposed the initial
> issue.
> 
> Fixes: c850e31f79f0 ("wifi: mac80211: add internal handler for wake_tx_queue")
> Reported-by: Thomas Mann <rauchwolke@gmx.net>
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=217119
> Link: https://lore.kernel.org/r/b8efebc6-4399-d0b8-b2a0-66843314616b@leemhuis.info/
> Link: https://lore.kernel.org/r/b7445607128a6b9ed7c17fcdcf3679bfaf4aaea.camel@sipsolutions.net>
> CC: <stable@vger.kernel.org>
> Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
> ---
> 
> @Thomas
> Would be good when you can test that patch again.
> But it would be really strange if it's not working, too...
> 
> @Johannes
> Based on your last mail you prefer to hard serialize it and not use a
> spin lock per AC. So I kept that part from the first patch.

This is missing the spin_lock_init() call.

- Felix