Re: ath11k: IPQ8074: hw2.0: WMI_PEER_STA_KICKOUT after 3900 seconds

[Sven Eckelmann <sven@narfation.org>]

[-- Attachment #1: Type: text/plain, Size: 2032 bytes --]

On Friday, 15 January 2021 15:22:02 CET Sven Eckelmann wrote:
> a rather odd behavior was noticed with ath11k and EAP.
[...]
> Btw. rekeying was disabled (but also setting it to 900s or 3600s doesn't 
> change the behavior at all. So here my hostapd config for completeness:

It seems like it also happens after 3900 seconds with PSK. So following
config will also create this problem:

    driver=nl80211
    logger_syslog=127
    logger_syslog_level=0
    logger_stdout=127
    logger_stdout_level=0
    country_code=US
    ieee80211d=1
    hw_mode=g
    beacon_int=100
    channel=6
    
    ieee80211n=1
    ht_coex=0
    ht_capab=[LDPC][SHORT-GI-20][SHORT-GI-40][TX-STBC][RX-STBC1][MAX-AMSDU-7935][DSSS_CCK-40]
    
    interface=ap2_1
    ctrl_interface=/var/run/hostapd
    ap_isolate=1
    bss_load_update_period=60
    chan_util_avg_period=600
    disassoc_low_ack=1
    preamble=1
    wmm_enabled=1
    ignore_broadcast_ssid=0
    uapsd_advertisement_enabled=1
    utf8_ssid=1
    multi_ap=0
    wpa_passphrase=testtest
    auth_algs=1
    wpa=2
    wpa_pairwise=CCMP
    ssid=psktest
    bridge=br-lan
    wpa_disable_eapol_key_retries=0
    wpa_key_mgmt=WPA-PSK
    okc=0
    disable_pmksa_caching=1
    bssid=f8:d9:b8:2e:37:d0

And this problem is trivial to see on the AP (with a single client) using the 
following patch:

    diff --git a/drivers/net/wireless/ath/ath11k/wmi.c b/drivers/net/wireless/ath/ath11k/wmi.c
    index 9fa6ff9..d1b2c61 100644
    --- a/drivers/net/wireless/ath/ath11k/wmi.c
    +++ b/drivers/net/wireless/ath/ath11k/wmi.c
    @@ -6504,6 +6504,7 @@ static void ath11k_peer_sta_kickout_event(struct ath11k_base *ab, struct sk_buff
     	ath11k_dbg(ab, ATH11K_DBG_WMI, "peer sta kickout event %pM",
     		   arg.mac_addr);
     
    +	WARN_ON(1);
     	ieee80211_report_low_ack(sta, 10);
     
     exit:


Maybe someone knows whether there is already a non upstream (keepalive?) patch 
from QCA in their wlan-open repository (in some of the branches) which I might 
miss.

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 833 bytes --]