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;
Subject: Re: [Bug] carl9170 + wpa_supplicant: P2P mode doesn't work with
 separate interfaces
From:   =?UTF-8?Q?Frank_Sch=c3=a4fer?= <fschaefer.oss@googlemail.com>
To:     chunkeey@googlemail.com, j@w1.fi
Cc:     linux-wireless@vger.kernel.org
References: <3a9d86b6-744f-e670-8792-9167257edef8@googlemail.com>
 <71051a8e-b8fc-3289-5009-b651e6c1a409@googlemail.com>
Message-ID: <11106990-ab12-3a8b-a203-2af8528ce3a1@googlemail.com>
Date:   Wed, 29 Apr 2020 18:22:44 +0200
User-Agent: Mozilla/5.0 (X11; Linux i686; rv:68.0) Gecko/20100101
 Thunderbird/68.7.0
MIME-Version: 1.0
In-Reply-To: <71051a8e-b8fc-3289-5009-b651e6c1a409@googlemail.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: de-DE
Sender: linux-wireless-owner@vger.kernel.org
Precedence: bulk


Am 27.04.20 um 18:53 schrieb Frank Schäfer:
>
> Am 18.04.20 um 13:20 schrieb Frank Schäfer:
>> Hi,
>>
>> I'm currently testing Miraclecast 
>> (https://github.com/albfan/miraclecast), an open source 
>> Wifi-Display/Miracast implementation.
>> In one of my setups, I'm using a carl9170 device as sink, which fails 
>> with the following debugging output:
>>
>> ...
>> TRACE: wpa: raw message: IFNAME=wlp0s2f1u8 <3>CTRL-EVENT-SCAN-STARTED
>> TRACE: wpa: raw message: <3>P2P-PROV-DISC-PBC-REQ 8e:79:67:11:22:33 
>> p2p_dev_addr=8e:79:67:11:22:33 pri_dev_type=10-0050F204-5 name='MY 
>> DEVICE' config_methods=0x188 dev_capab=0x25 group_capab=0x0
>> TRACE: wpa: raw message: <3>P2P-GO-NEG-REQUEST 8e:79:67:11:22:33 
>> dev_passwd_id=4 go_intent=0
>> DEBUG: supplicant: GO Negotiation Request from 8e:79:67:11:22:33 
>> (supplicant_event_p2p_go_neg_request() in 
>> ../../../src/wifi/wifid-supplicant.c:1065)
>> DEBUG: supplicant: connect to 8e:79:67:11:22:33 via pbc/(null) 
>> (supplicant_peer_connect() in ../../../src/wifi/wifid-supplicant.c:739)
>> TRACE: wpa: raw message: <3>P2P-FIND-STOPPED
>> DEBUG: supplicant: p2p-scanning stopped on wlp0s2f1u8 
>> (supplicant_event_p2p_find_stopped() in 
>> ../../../src/wifi/wifid-supplicant.c:904)
>> TRACE: wpa: raw message: OK
>>
>> TRACE: wpa: raw message: <3>P2P-GO-NEG-SUCCESS role=GO freq=2462 
>> ht40=0 peer_dev=8e:79:67:11:22:33 peer_iface=8e:79:67:11:22:33 
>> wps_method=PBC
>> DEBUG: supplicant: set STA-MAC for 8e:79:67:11:22:33 from <none> to 
>> 8e:79:67:11:22:33 (via GO-NEG-SUCCESS) 
>> (supplicant_event_p2p_go_neg_success() in 
>> ../../../src/wifi/wifid-supplicant.c:1194)
>> TRACE: wpa: raw message: IFNAME=p2p-0 <3>Failed to start AP 
>> functionality
>> DEBUG: supplicant: unhandled wpas-event: IFNAME=p2p-0 <3>Failed to 
>> start AP functionality (supplicant_event() in 
>> ../../../src/wifi/wifid-supplicant.c:1498)
>> TRACE: wpa: raw message: <3>P2P-GROUP-FORMATION-FAILURE
>> DEBUG: supplicant: peer MY DEVICE connection failed 
>> (supplicant_event_p2p_group_formation_failure() in 
>> ../../../src/wifi/wifid-supplicant.c:1318)
>> TRACE: wpa: raw message: <3>P2P-GROUP-REMOVED p2p-0 GO 
>> reason=FORMATION_FAILED
>> DEBUG: supplicant: stray P2P-GROUP-REMOVED: <3>P2P-GROUP-REMOVED 
>> p2p-0 GO reason=FORMATION_FAILED 
>> (supplicant_event_p2p_group_removed() in 
>> ../../../src/wifi/wifid-supplicant.c:1286)
>> DEBUG: supplicant: sent P2P_FIND to wpas on wlp0s2f1u8 
>> (supplicant_p2p_start_scan() in 
>> ../../../src/wifi/wifid-supplicant.c:2036)
>> TRACE: wpa: raw message: OK
>>
>> DEBUG: supplicant: p2p-scanning now active on wlp0s2f1u8 
>> (supplicant_p2p_find_fn() in ../../../src/wifi/wifid-supplicant.c:1917)
>> TRACE: wpa: raw message: IFNAME=wlp0s2f1u8 <3>CTRL-EVENT-SCAN-STARTED
>> ...
>>
>>
>> Searching the web for similar problems leads to the following thread:
>> https://marc.info/?l=hostap&m=135336063022534
>>
>> As suggested there, I have patched the Miraclecast sources to use the 
>> wpa_supplicant configuration parameter p2p_no_group_iface=1.
>> So far, this works fine.
>>
>>
>> Regards,
>> Frank Schäfer
>>
>
> I finally managed to create extended logs and to compare the working 
> ("GOOD") and non-working ("BAD") case.
> See attachment.
> If anything is missing, please let me know.
>
> Hth,
> Frank
>

Ok, this is what I can see happening in carl9170 with regards to the 
interfaces:

1) 1 interface is added+initialized: NL80211_IFTYPE_STATION, p2p=false 
(when miracle-wifid is started)

Non-working case (p2p_no_group_iface=0=default):
2a) 1 interface is added: NL80211_IFTYPE_STATION, p2p=true (apparently 
the p2p group interface)
2b) 1 interface is removed: NL80211_IFTYPE_STATION, p2p=true (apparently 
because setting of the interface flags failed with -EBUSY)
3) 1 interface is added+initialized: NL80211_IFTYPE_AP, p2p=true
    => fails with -EBUSY, because main_vif->p2p is false (main.c, line 656)

Working case (p2p_no_group_iface=1):
2) 1 interface is removed: NL80211_IFTYPE_STATION, p2p=false
3) 1 interface is added+initialized: NL80211_IFTYPE_AP, p2p=true

So if p2p_no_group_iface=1 is used, the initial STA non-p2p interface is 
removed before a new AP+p2p interface is added.
Maybe the same needs to be done in case of p2p_no_group_iface=0 ?

Regards,
Frank