netdev - Re: [PATCH] wcn36xx: handle connection loss indication

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <168f96a5-58ec-fea9-c0d3-61f925bd1129@squareup.com>
Date:   Wed, 1 Sep 2021 11:02:32 -0700
From:   Benjamin Li <benl@...areup.com>
To:     Bryan O'Donoghue <bryan.odonoghue@...aro.org>,
        Loic Poulain <loic.poulain@...aro.org>
Cc:     Kalle Valo <kvalo@...eaurora.org>, stable@...r.kernel.org,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>, wcn36xx@...ts.infradead.org,
        linux-wireless@...r.kernel.org,
        Network Development <netdev@...r.kernel.org>,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] wcn36xx: handle connection loss indication

Thanks for the investigation!

As discussed offline, I will send v2 with Fixes: removed, and Bryan will test and submit a separate patch to add the additional feat_caps for the power_save off case.

Depending on DB410c testing, these feat_caps may need to be gated for WCN3680 only, in which case Loic's patch to re-enable CONNECTION_MONITOR (but gated for WCN3660/3620) would still be needed.

Ben

On 9/1/21 4:56 AM, Bryan O'Donoghue wrote:
> On 01/09/2021 07:40, Loic Poulain wrote:
>> iw wlan0 set power_save off
> 
> I do this on wcn3680b and get no loss of signal
> 
> If I do this though
> 
> diff --git a/drivers/net/wireless/ath/wcn36xx/smd.c b/drivers/net/wireless/ath/wcn36xx/smd.c
> index 03966072f34c..ba613fbb728d 100644
> --- a/drivers/net/wireless/ath/wcn36xx/smd.c
> +++ b/drivers/net/wireless/ath/wcn36xx/smd.c
> @@ -2345,6 +2345,8 @@ int wcn36xx_smd_feature_caps_exchange(struct wcn36xx *wcn)
>                 set_feat_caps(msg_body.feat_caps, DOT11AC);
>                 set_feat_caps(msg_body.feat_caps, ANTENNA_DIVERSITY_SELECTION);
>         }
> +       set_feat_caps(msg_body.feat_caps, IBSS_HEARTBEAT_OFFLOAD);
> +       set_feat_caps(msg_body.feat_caps, WLANACTIVE_OFFLOAD);
> 
>         PREPARE_HAL_BUF(wcn->hal_buf, msg_body);
> 
> @@ -2589,7 +2591,7 @@ static int wcn36xx_smd_missed_beacon_ind(struct wcn36xx *wcn,
>         struct wcn36xx_hal_missed_beacon_ind_msg *rsp = buf;
>         struct ieee80211_vif *vif = NULL;
>         struct wcn36xx_vif *tmp;
> -
> +wcn36xx_info("%s/%d\n", __func__, __LINE__);
>         /* Old FW does not have bss index */
>         if (wcn36xx_is_fw_version(wcn, 1, 2, 2, 24)) {
>                 list_for_each_entry(tmp, &wcn->vif_list, list) {
> @@ -2608,7 +2610,7 @@ static int wcn36xx_smd_missed_beacon_ind(struct wcn36xx *wcn,
> 
>         list_for_each_entry(tmp, &wcn->vif_list, list) {
>                 if (tmp->bss_index == rsp->bss_index) {
> -                       wcn36xx_dbg(WCN36XX_DBG_HAL, "beacon missed bss_index %d\n",
> +                       wcn36xx_info("beacon missed bss_index %d\n",
>                                     rsp->bss_index);
>                         vif = wcn36xx_priv_to_vif(tmp);
>                         ieee80211_connection_loss(vif);
> 
> 
> bingo
> 
> root@...aro-developer:~# iw wlan0 set power_save off
> 
> 
> # pulls plug on AP
> 
> root@...aro-developer:~# [   83.290987] wcn36xx: wcn36xx_smd_missed_beacon_ind/2594
> [   83.291070] wcn36xx: beacon missed bss_index 0
> [   83.295403] wlan0: Connection to AP e2:63:da:9c:a4:bd lost
> 
> I'm not sure if both flags are required but, this is the behavior we want
>