Message-ID: <5251bf43-a980-3398-80d6-9689f0efc926@nbd.name>
Date:   Sun, 21 Nov 2021 20:46:20 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:91.0)
 Gecko/20100101 Thunderbird/91.3.2
Subject: Re: Bug/Regression - Ralink RT2800 kernel deference issue since
 kernel 5.14
Content-Language: en-US
To:     Robert W <rwbugreport@lost-in-the-void.net>, xl@wp.pl
Cc:     linux-wireless@vger.kernel.org
References: <c07b4142fb725ed87a2cef530bae9ee7@lost-in-the-void.net>
From:   Felix Fietkau <nbd@nbd.name>
In-Reply-To: <c07b4142fb725ed87a2cef530bae9ee7@lost-in-the-void.net>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2021-11-12 22:21, Robert W wrote:
> Hi,
> 
> I have recently tried to upgrade a system to run the 5.14 kernel and
> noticed that I was getting hangs on the system which seem to relate to
> this wireless dongle which I am running hostapd against to provide a
> small AP. Since that time kernel 5.15 was released and I have tested
> that kernel as well and the problem persists.  The issue sometimes
> occurs within an hour and other times it takes a few hours. I have tried
> to narrow down what is causing the issue.  I thought it might be a
> change elsewhere outside the mac80211/driver area so I tried to run the
> wireless backports 5.15-rc6 on top of the last stable kernel of 5.13 so
> as to have the smallest number of changes on the system. The backports
> crashes with the same error.  I have included below the netconsole
> output and the gdb of the kernel module in question. The capture was
> from when I was using the backports, but the error  is the same.  I hope
> the pastebins are okay to use on the list.  I did try to get some
> assistance on linux-wireless IRC and some one asked for the outputs I
> had for the information so I am reusing them here.
> 
> The following pastebin is the data sent to a netconsole instance.
> 
> https://pastebin.com/UxRrTtUh
> 
> The gdb debug of the kernel module suggests this section of code.
> 
> https://pastebin.com/iMzxWEVU
> 
> I am able to run further tests if required.  I am not a programmer, but
> I can usually apply code patches and build the resulting code. Please
> let me know if require any further information.

Please try this patch:

diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index 278945e3e08a..e19f6e246642 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -1822,15 +1822,15 @@ static int invoke_tx_handlers_late(struct ieee80211_tx_data *tx)
  	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb);
  	ieee80211_tx_result res = TX_CONTINUE;
  
+	if (!ieee80211_hw_check(&tx->local->hw, HAS_RATE_CONTROL))
+		CALL_TXH(ieee80211_tx_h_rate_ctrl);
+
  	if (unlikely(info->flags & IEEE80211_TX_INTFL_RETRANSMISSION)) {
  		__skb_queue_tail(&tx->skbs, tx->skb);
  		tx->skb = NULL;
  		goto txh_done;
  	}
  
-	if (!ieee80211_hw_check(&tx->local->hw, HAS_RATE_CONTROL))
-		CALL_TXH(ieee80211_tx_h_rate_ctrl);
-
  	CALL_TXH(ieee80211_tx_h_michael_mic_add);
  	CALL_TXH(ieee80211_tx_h_sequence);
  	CALL_TXH(ieee80211_tx_h_fragment);