WebDawg | 22 May 17:38 2015

Re: Wireless AC



So looking into some of these devices, is there some type of HCL?  I see a nice organized list, but very little about hostapd compatibility.
HostAP mailing list
HostAP <at> lists.shmoo.com
volker.obhof.w | 22 May 11:47 2015

Modify capability information field in assocication request

I just search everywhere in wpa_supplicant source code to modify the capability information field in association request but I didn’t find anything.


In file driver_nl80211.c there is a function called wpa_driver_nl80211_associate() where a association request is set. Then in function nl80211_connect_common() a new association request frame is built.


But nowhere is set the data for the capability information field.


Can anybody help me to find out which struct and function I have to use to modify capability information field in wpa_supplicant?

Kommanditgesellschaft, Sitz: Bruchsal, RG Mannheim HRA 230970
Komplementärin: SEW-EURODRIVE Verwaltungs-GmbH, Sitz: Bruchsal, RG Mannheim HRB 230207

Gesellschafter: Jürgen Blickle, Rainer Blickle
Geschäftsführer: Jürgen Blickle (Vorsitzender), Johann Soder, Dr. Jürgen Zanghellini
HostAP mailing list
HostAP <at> lists.shmoo.com
WebDawg | 21 May 19:45 2015

Wireless AC

I wanted to setup a Wireless AC access point and was shopping for the best chipset/hardware.  I am going to be building an access point and I was thinking pcie would be the best interface for the card.

Can someone tell me the best card to get?  I was looking for something that I should expect to be supported in the future by hostapd and its associated driver.

What is the most popular fastest AC card?
HostAP mailing list
HostAP <at> lists.shmoo.com
Brad Smith | 21 May 19:32 2015

Troubleshooting tips?

Hello all,

I'm in the US, working remotely with a tech school in Kenya to provision a micro PC as a hostapd-powered wireless access point that provides net access when available (bandwidth can get pretty flaky out there), and locally hosts offline versions of resources like wikipedia, MOOCs, etc regardless. 

I tested everything at home and had five or six devices of various types all streaming video from it with no problems, but now they're testing it in the field and reporting that they start having connectivity problems after one or two devices.

I'm a Linux/dev guy so I can handle technical docs and whatnot if people have something to link me to, but I have almost no experience with this level of wireless networking, so I'm hoping someone here would be willing to help me out to further a good cause by offering some tips on how to troubleshoot this sort of problem. 

Here's what we know:
  • Problems don't appear to be associated with any particular client device or class of devices (laptops, phones, etc).
  • Problems are intermittent, so a machine will fail to connect to the WAP for a while, but then 30 minutes later will work fine... until it doesn't any more.
  • There are other WAPs in the area, but they're far enough away that ours is by far the strongest signal.
  • Hardware-wise, we're using a cubietruck. It's actually a bit unclear exactly what wifi hardware is in there, but a user wiki says: 

    There is an AMPAK AP6210 Wifi+BT chip on board with PCB antenna, but it is not clear what exactly is inside:

    • Broadcom BCM43362 + BCM20710
    • Broadcom BCM4330 / BCM40183 Based Chipset
    • Cubietech says it is BCM4329/BCM40181
  • We're using the bcmdhd driver with op_mode=2 on Ubuntu Trusty (ARM)
  • Our hostapd.conf is pretty basic:
  • I think they may have added lines to enable WAP as well. I'll ask for a copy of what exactly they're using.
Does anyone have suggestions for where to start in troubleshooting this? 

HostAP mailing list
HostAP <at> lists.shmoo.com
Yury Shvedov | 21 May 08:24 2015

Multiple wlans' bssids


according to this many other sources and my experience of hostapd's code modifying the mac address of base virtual interface must be multiple to 0x8 at the last octet. After my simple modifications of hostapd's code this this requirement gone, and hostapd uses original address of wireless interface for the first bss (hostapd_iface::bss[0]). But when I'm trying to add another bss, driver fails to create it with the error like "Could not set interface wlan1-2 hwaddr: Invalid argument".

Can someone please explain me what the reason of such behaviour? Could I

HostAP mailing list
HostAP <at> lists.shmoo.com
Harsha vardhan | 20 May 06:40 2015

PEAP-TLS Configuration file

Hi Team,

Can anyone give me support on how to test PEAP-TLS method with wpa_supplicant.

HostAP mailing list
HostAP <at> lists.shmoo.com
Harsha vardhan | 19 May 12:36 2015

Request for PEAP-TLS configuration file

Dear Team,

Can you please provide wpa_supplicant config file PEAP-TLS method.

Thanks in advance.

HostAP mailing list
HostAP <at> lists.shmoo.com
John Ernberg | 19 May 08:09 2015

Using wpa_supplicant.conf with removable WiFi modules?


We're trying to use wpa_supplicant 2.4 with our removable WiFi module 
for P2P connections. It requires some device specific settings that we 
specified in the .conf file. The problem is that due to how 
configurations are assigned to interfaces, we have to restart 
wpa_supplicant every time we re-insert the module. This is not 
preferable to us.

Is there a way to assign these settings to wpa_supplicant so they become 
global for the entire execution period? If not, is there any reason as 
to why it was designed like this?

The settings are these:


Thank you in advance.

Best regards // John Ernberg
Min Zhang | 15 May 00:02 2015

[PATCH] WPS: Exclude missing UUID-E from PBC overlap detection

Missing UUID-E shouldn't result in PBC overlap. An AP selected as
PBC registrar can Beacon w/o UUID-E attr as not required, but only
include unique UUID-E in its Probe Resp. The Andriod PBC scan seemly
just uses Beacon, so the NULL uuid check results in a false overlap.

Also commit 44cd430f877a558744def067ab38c4f9e4b283fc "Fixed PBC
overlap detection to handle case of missing UUID-E" tried but still
ended up PBC overlap on missing UUID-E.

The patch effectively exclude Beacon Frame from PBC overlap
detection. It is inline with WPS PBC overlap spec that only mentioned
Probe not Beacon. Another way is let PBC scan uses Probe Resp so it
can extract the actual UUID-E to compare.

Note the WPS spec also only suggested Selected Registrar as PBC
overlap detection. Andriod added UUID-E matching on top which I think
it is better since it allows unrelated PBC overlap.

Signed-off-by: Min Zhang <mzhang <at> roku.com>
  wpa_supplicant/wps_supplicant.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/wpa_supplicant/wps_supplicant.c b/wpa_supplicant/wps_supplicant.c
index 40a5c69..a2162a4 100644
--- a/wpa_supplicant/wps_supplicant.c
+++ b/wpa_supplicant/wps_supplicant.c
 <at>  <at>  -1697,7 +1697,7  <at>  <at>  int wpas_wps_scan_pbc_overlap(struct wpa_supplicant *wpa_s,
  		uuid = wps_get_uuid_e(ie);
  		wpa_hexdump(MSG_DEBUG, "WPS: UUID of the other BSS",
  			    uuid, UUID_LEN);
-		if (sel_uuid == NULL || uuid == NULL ||
+		if (sel_uuid != NULL && uuid != NULL &&
  		    os_memcmp(sel_uuid, uuid, UUID_LEN) != 0) {
  			ret = 1; /* PBC overlap */
  			wpa_msg(wpa_s, MSG_INFO, "WPS: PBC overlap detected: "

Peer, Ilan | 14 May 12:40 2015

RE: p2p bug/feature when using intel 7260 and 5GHz


This is opposite to the pervious case. In this case the ath device is initiating a p2p_connection forcing
the use of frequency 5180 but with go_intent=1, while the Intel device is responding with go_intent=15.
Thus, in this case the Intel device is going to be the P2P GO, but according to the regulatory channel
information on the Intel device only 1..11 are allowed for use for beaconing, so this scenario will fail.

Did you try the opposite?



> -----Original Message-----
> From: Janusz Dziedzic [mailto:janusz.dziedzic <at> tieto.com]
> Sent: Thursday, May 14, 2015 13:22
> To: Peer, Ilan
> Cc: hostap <at> lists.shmoo.com
> Subject: Re: p2p bug/feature when using intel 7260 and 5GHz
> On 14 May 2015 at 08:01, Peer, Ilan <ilan.peer <at> intel.com> wrote:
> > Hi Janusz,
> >
> > Did you set p2p_add_cli_chan in the P2P Device configuration file? Without
> setting it, passive-scan channels will not be added in the P2P channels.
> >
> > It this still fails, can you please send me wpa_supplicant log file and also the
> output of iw list?
> >
> attached logs
Janusz Dziedzic | 14 May 07:17 2015

p2p bug/feature when using intel 7260 and 5GHz


A) Connection to autonomous GO works fine:
t1-ath9k <at> wlan0 send: P2P_GROUP_ADD freq=5180
t1-ath9k <at> wlan0 recv: OK

t1-ath9k <at> wlan0 wait_msg: P2P-GROUP-STARTED (30)
t1-ath9k <at> wlan0 get: <3>P2P-GROUP-STARTED p2p-wlan0-0 GO
ssid="DIRECT-SG" freq=5180 passphrase="x7M4ScIq"

t2-intel-7260 <at> p2p-dev-wlan3 send: P2P_CONNECT 2c:d0:5a:d3:f1:e9 12345670 join
2-intel-7260 <at> p2p-dev-wlan3 get: <3>P2P-GROUP-STARTED p2p-wlan3-0
client ssid="DIRECT-SG" freq=5180

B) But when GO negotiation and intel with go_intent=1
t1-ath9k <at> wlan0 send: P2P_CONNECT 0c:8b:fd:1e:58:9e pbc freq=5180  go_intent=15
t2-intel-7260 <at> p2p-dev-wlan3 wait_msg: P2P-GO-NEG-REQUEST 2c:d0:5a:d3:f1:e9 (60)
t2-intel-7260 <at> p2p-dev-wlan3 send: P2P_CONNECT 2c:d0:5a:d3:f1:e9 pbc go_intent=1
t2-intel-7260 <at> p2p-dev-wlan3 get: <3>P2P-GO-NEG-FAILURE status=7

GO (ath9k) reject GO-NEG and in logs I see:

P2P: Own reg_classes 1 peer reg_classes 1 intersection reg_classes 0
P2P: No common channels found

While intel will be p2p client (no beaconnig required), why we reject
this? Are we to strict in this case?