Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Sender: kvalo=codeaurora.org@mg.codeaurora.org
DMARC-Filter: OpenDMARC Filter v1.3.2 smtp.codeaurora.org E307AC433CA
From:   Kalle Valo <kvalo@codeaurora.org>
To:     Peter Oh <peter.oh@eero.com>
Cc:     Dan Carpenter <dan.carpenter@oracle.com>,
        kernel-janitors@vger.kernel.org,
        Carl Huang <cjhuang@codeaurora.org>,
        linux-wireless@vger.kernel.org, ath11k@lists.infradead.org
Subject: Re: [PATCH] ath11k: fix a locking bug in ath11k_mac_op_start()
References: <YBfy+zc3XkiyIe6t@mwanda>
        <896a68dd-02ec-4fe3-3fbf-ec1bd2decefe@eero.com>
Date:   Tue, 02 Feb 2021 06:32:44 +0200
In-Reply-To: <896a68dd-02ec-4fe3-3fbf-ec1bd2decefe@eero.com> (Peter Oh's
        message of "Mon, 1 Feb 2021 13:47:27 -0800")
Message-ID: <875z3b84yb.fsf@codeaurora.org>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.5 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Peter Oh <peter.oh@eero.com> writes:

> On 2/1/21 4:24 AM, Dan Carpenter wrote:
>> This error path leads to a Smatch warning:
>>
>>      drivers/net/wireless/ath/ath11k/mac.c:4269 ath11k_mac_op_start()
>>      error: double unlocked '&ar->conf_mutex' (orig line 4251)
>>
>> We're not holding the lock when we do the "goto err;" so it leads to a
>> double unlock.  We should hold the lock because the error path sets
>> "ar->state" so the right fix is to take the lock before doing the goto.
>>
>> Fixes: c83c500b55b6 ("ath11k: enable idle power save mode")
>> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
>> ---
>>   drivers/net/wireless/ath/ath11k/mac.c | 1 +
>>   1 file changed, 1 insertion(+)
>>
>> diff --git a/drivers/net/wireless/ath/ath11k/mac.c b/drivers/net/wireles=
s/ath/ath11k/mac.c
>> index c1608f64ea95..12e981e9e3d7 100644
>> --- a/drivers/net/wireless/ath/ath11k/mac.c
>> +++ b/drivers/net/wireless/ath/ath11k/mac.c
>> @@ -4259,6 +4259,7 @@ static int ath11k_mac_op_start(struct ieee80211_hw=
 *hw)
>>   						1, pdev->pdev_id);
>>   		if (ret) {
>>   			ath11k_err(ab, "failed to enable idle ps: %d\n", ret);
>> +			mutex_lock(&ar->conf_mutex);
>>   			goto err;
>>   		}
>>   	}
>
> It seems moving idle_ps condition in between ath11k_wmi_pdev_lro_cfg()
> and mutex_unlock()
> =C2=A0is a better way in this case.

I agree. In Dan's approach I'm worried about the race when we unlock and
lock again, the state can change in that case.

--=20
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatc=
hes