Configuring WiFi6 in hostapd

Tue Sep 8 18:53:19 EDT 2020

On Wed, Sep 9, 2020 at 2:18 AM Sugandh Huthanahally Mohan
<sugandhgowda at gmail.com> wrote:
>
> On Wed, Sep 9, 2020 at 12:27 AM Sugandh Huthanahally Mohan
> <sugandhgowda at gmail.com> wrote:
> > >
> > > I have captured connection between a WiFi6 AP and a WiFi6 STA in wireshark.
> > >I have attached a wireshark pcap file.
> > > I could see HE capabilities exchange in Association request and response.
> > Yes, this looks fine to me except for VHT IEs in the probe and
> > association response in 2.4GHz.
> > Maybe this is what is confusing, can you please disable ieee80211ac=0
> > in the hostapd.conf and try?
> I disabled iee80211ac and I am not seeing VHT capabilities in Association
> frames but the data rate remains the same.I am attaching wireshark capture.
okay, thanks.
> > The iw phy output is below. I couldn't see HE connection in this output.
> >
> > $iw dev wlp3s0 link
> > Connected to 8c:c6:81:69:86:65 (on wlp3s0)
> >     SSID: WiFi_AX200
> >     freq: 2437
> >     RX: 27838 bytes (199 packets)
> >     TX: 13319 bytes (97 packets)
> >     signal: -34 dBm
> >     rx bitrate: 26.0 MBit/s MCS 3
> >     tx bitrate: 2.0 MBit/s
> >
> >     bss flags:    short-slot-time
> >     dtim period:    2
> >     beacon int:    100
> >
> > > The link doesn't show any specific details related to HE other than
> > > the MCS which can vary, try running
> > > an iperf to see HE MCSes in a clean channel.
>
> Could you please explain what this means.
> I want to know what clean channel means and where actually I can see HE MCSes.

Sure. The link command doesn't printout any in capability information explicitly
 except for the MCS details. And figuring out the mode of connection (11n/11ax)
from the MCS is tricky as the MCS varies across depending on the rate
control algorithm
(for iwlwifi: hardware). So, I was suggesting to use iperf (the more
data you pass, more
chance of the rate control algo to go to higher MCSes) and also by
"clean channels" I was
referring to quieter channels where there are no other BSSes or no
other traffic.

But, before we go into debugging data path, I still don't understand
why wpa_supplicant
isn't saying it is a HE association?

Ah, After staring at the code and using blame, a crucial fix missed
out hostap-2.9 windows,
the fix was to properly validate the HE capabilities, I was looking at
the latest cod and
everything looked fine, hence took time.

hostap_2_9-> 7th August
0497e4148197fb8fdf92b5c901ce06df1bfa548c ->13th August

Please use the latest HEAD of hostap.git, you need to compile it
manually and try the
test and let me know how it goes. (make sure all proper configs are enabled)

> > On Wed, Sep 2, 2020 at 3:39 PM Krishna Chaitanya
> > <chaitanya.mgit at gmail.com> wrote:
> > >
> > > On Wed, Sep 2, 2020 at 6:38 PM Sugandh Huthanahally Mohan
> > > <sugandhgowda at gmail.com> wrote:
> > > >
> > > > On Wed, Sep 2, 2020 at 3:16 PM Sugandh Huthanahally Mohan
> > > > <sugandhgowda at gmail.com> wrote:
> > > > >
> > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > >
> > > > > >
> > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > different draft implementation. At the very least we should start
> > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > the IEs.
> > > > >
> > > > > I am using wpa_supplicant in STA to connect  to AP.
> > > > > I am attaching hostapd logs and wpa_supplicant logs.
> > > > > I am observing that the STA is connecting and disconnecting when
> > > > > I try to connect STA to AP.
> > > > I guess this is a new STA with wpa_supplicant, right? Is it also AX200 based
> > > > chipset? In the STA mode, the kernel adds the HE IEs (either mac80211
> > > > or the driver),
> > > > so, but interestingly in your previous dmesg, I don't see any mac80211 logs.
> > > >
> > > > So, need dmesg logs that capture the connection (look for mac80211 logs).
> > > > >
> > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > >
> > > > > > On Wed, Sep 2, 2020 at 1:23 AM Sugandh Huthanahally Mohan
> > > > > > <sugandhgowda at gmail.com> wrote:
> > > > > > >
> > > > > > > Hi Krishna,
> > > > > > >
> > > > > > > Currently I am not running wpa_supplicant.
> > > > > > > Is it necessary for hostapd?
> > > > > >
> > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > different draft implementation. At the very least we should start
> > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > the IEs.
> > > >
> > > > Yes this is a new STA connecting AP with wpa_supplicant.
> > > > It is also an AX200.
> > > > I don’t see mac80211 logs on the STA dmesg.
> > > > I am attaching dmesg of the station.
> > >
> > > These are from mac80211:
> > > [  152.255545] wlp3s0: authenticate with 8c:c6:81:69:86:65
> > > [  152.258347] wlp3s0: send auth to 8c:c6:81:69:86:65 (try 1/3)
> > > [  152.283971] wlp3s0: authenticated
> > > [  152.286624] wlp3s0: associate with 8c:c6:81:69:86:65 (try 1/3)
> > > [  152.293456] wlp3s0: RX AssocResp from 8c:c6:81:69:86:65
> > > (capab=0x411 status=0 aid=1)
> > > [  152.299672] wlp3s0: associated
> > > [  152.364843] IPv6: ADDRCONF(NETDEV_CHANGE): wlp3s0: link becomes ready
> > >
> > > I was hoping to see some prints disabling HE, but looks fine. So,
> > > still, nothing from the logs
> > > to say why HE is disabled in the STA (why it isn't sending HE IEs in
> > > assoc resp).
> > >
> > > Only Sniffer should tell us without resorting to Whitebox debug.
> > > >
> > > > On Wed, Sep 2, 2020 at 1:26 PM Krishna Chaitanya
> > > > <chaitanya.mgit at gmail.com> wrote:
> > > > >
> > > > > On Wed, Sep 2, 2020 at 3:16 PM Sugandh Huthanahally Mohan
> > > > > <sugandhgowda at gmail.com> wrote:
> > > > > >
> > > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > > >
> > > > > > >
> > > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > > different draft implementation. At the very least we should start
> > > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > > the IEs.
> > > > > >
> > > > > > I am using wpa_supplicant in STA to connect  to AP.
> > > > > > I am attaching hostapd logs and wpa_supplicant logs.
> > > > > > I am observing that the STA is connecting and disconnecting when
> > > > > > I try to connect STA to AP.
> > > > > I guess this is a new STA with wpa_supplicant, right? Is it also AX200 based
> > > > > chipset? In the STA mode, the kernel adds the HE IEs (either mac80211
> > > > > or the driver),
> > > > > so, but interestingly in your previous dmesg, I don't see any mac80211 logs.
> > > > >
> > > > > So, need dmesg logs that capture the connection (look for mac80211 logs).
> > > > > >
> > > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > > >
> > > > > > > On Wed, Sep 2, 2020 at 1:23 AM Sugandh Huthanahally Mohan
> > > > > > > <sugandhgowda at gmail.com> wrote:
> > > > > > > >
> > > > > > > > Hi Krishna,
> > > > > > > >
> > > > > > > > Currently I am not running wpa_supplicant.
> > > > > > > > Is it necessary for hostapd?
> > > > > > >
> > > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > > different draft implementation. At the very least we should start
> > > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > > the IEs.
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Thanks,
> > > > > Regards,
> > > > > Chaitanya T K.
> > >
> > >
> > >
> > > --
> > > Thanks,
> > > Regards,
> > > Chaitanya T K.
>
>
> On Tue, Sep 8, 2020 at 9:55 PM Krishna Chaitanya
> <chaitanya.mgit at gmail.com> wrote:
> >
> > On Wed, Sep 9, 2020 at 12:27 AM Sugandh Huthanahally Mohan
> > <sugandhgowda at gmail.com> wrote:
> > >
> > > I have captured connection between a WiFi6 AP and a WiFi6 STA in wireshark.
> > > I have attached a wireshark pcap file.
> > > I could see HE capabilities exchange in Association request and response.
> > Yes, this looks fine to me except for VHT IEs in the probe and
> > association response in 2.4GHz.
> > Maybe this is what is confusing, can you please disable ieee80211ac=0
> > in the hostapd.conf and try?
> >
> > > The iw phy output is below. I couldn't see HE connection in this output.
> > >
> > > $iw dev wlp3s0 link
> > > Connected to 8c:c6:81:69:86:65 (on wlp3s0)
> > >     SSID: WiFi_AX200
> > >     freq: 2437
> > >     RX: 27838 bytes (199 packets)
> > >     TX: 13319 bytes (97 packets)
> > >     signal: -34 dBm
> > >     rx bitrate: 26.0 MBit/s MCS 3
> > >     tx bitrate: 2.0 MBit/s
> > >
> > >     bss flags:    short-slot-time
> > >     dtim period:    2
> > >     beacon int:    100
> > >
> > The link doesn't show any specific details related to HE other than
> > the MCS which can vary, try running
> > an iperf to see HE MCSes in a clean channel.
> > >
> > > On Wed, Sep 2, 2020 at 3:39 PM Krishna Chaitanya
> > > <chaitanya.mgit at gmail.com> wrote:
> > > >
> > > > On Wed, Sep 2, 2020 at 6:38 PM Sugandh Huthanahally Mohan
> > > > <sugandhgowda at gmail.com> wrote:
> > > > >
> > > > > On Wed, Sep 2, 2020 at 3:16 PM Sugandh Huthanahally Mohan
> > > > > <sugandhgowda at gmail.com> wrote:
> > > > > >
> > > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > > >
> > > > > > >
> > > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > > different draft implementation. At the very least we should start
> > > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > > the IEs.
> > > > > >
> > > > > > I am using wpa_supplicant in STA to connect  to AP.
> > > > > > I am attaching hostapd logs and wpa_supplicant logs.
> > > > > > I am observing that the STA is connecting and disconnecting when
> > > > > > I try to connect STA to AP.
> > > > > I guess this is a new STA with wpa_supplicant, right? Is it also AX200 based
> > > > > chipset? In the STA mode, the kernel adds the HE IEs (either mac80211
> > > > > or the driver),
> > > > > so, but interestingly in your previous dmesg, I don't see any mac80211 logs.
> > > > >
> > > > > So, need dmesg logs that capture the connection (look for mac80211 logs).
> > > > > >
> > > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > > >
> > > > > > > On Wed, Sep 2, 2020 at 1:23 AM Sugandh Huthanahally Mohan
> > > > > > > <sugandhgowda at gmail.com> wrote:
> > > > > > > >
> > > > > > > > Hi Krishna,
> > > > > > > >
> > > > > > > > Currently I am not running wpa_supplicant.
> > > > > > > > Is it necessary for hostapd?
> > > > > > >
> > > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > > different draft implementation. At the very least we should start
> > > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > > the IEs.
> > > > >
> > > > > Yes this is a new STA connecting AP with wpa_supplicant.
> > > > > It is also an AX200.
> > > > > I don’t see mac80211 logs on the STA dmesg.
> > > > > I am attaching dmesg of the station.
> > > >
> > > > These are from mac80211:
> > > > [  152.255545] wlp3s0: authenticate with 8c:c6:81:69:86:65
> > > > [  152.258347] wlp3s0: send auth to 8c:c6:81:69:86:65 (try 1/3)
> > > > [  152.283971] wlp3s0: authenticated
> > > > [  152.286624] wlp3s0: associate with 8c:c6:81:69:86:65 (try 1/3)
> > > > [  152.293456] wlp3s0: RX AssocResp from 8c:c6:81:69:86:65
> > > > (capab=0x411 status=0 aid=1)
> > > > [  152.299672] wlp3s0: associated
> > > > [  152.364843] IPv6: ADDRCONF(NETDEV_CHANGE): wlp3s0: link becomes ready
> > > >
> > > > I was hoping to see some prints disabling HE, but looks fine. So,
> > > > still, nothing from the logs
> > > > to say why HE is disabled in the STA (why it isn't sending HE IEs in
> > > > assoc resp).
> > > >
> > > > Only Sniffer should tell us without resorting to Whitebox debug.
> > > > >
> > > > > On Wed, Sep 2, 2020 at 1:26 PM Krishna Chaitanya
> > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > >
> > > > > > On Wed, Sep 2, 2020 at 3:16 PM Sugandh Huthanahally Mohan
> > > > > > <sugandhgowda at gmail.com> wrote:
> > > > > > >
> > > > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > > > >
> > > > > > > >
> > > > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > > > different draft implementation. At the very least we should start
> > > > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > > > the IEs.
> > > > > > >
> > > > > > > I am using wpa_supplicant in STA to connect  to AP.
> > > > > > > I am attaching hostapd logs and wpa_supplicant logs.
> > > > > > > I am observing that the STA is connecting and disconnecting when
> > > > > > > I try to connect STA to AP.
> > > > > > I guess this is a new STA with wpa_supplicant, right? Is it also AX200 based
> > > > > > chipset? In the STA mode, the kernel adds the HE IEs (either mac80211
> > > > > > or the driver),
> > > > > > so, but interestingly in your previous dmesg, I don't see any mac80211 logs.
> > > > > >
> > > > > > So, need dmesg logs that capture the connection (look for mac80211 logs).
> > > > > > >
> > > > > > > On Tue, Sep 1, 2020 at 10:09 PM Krishna Chaitanya
> > > > > > > <chaitanya.mgit at gmail.com> wrote:
> > > > > > > >
> > > > > > > > On Wed, Sep 2, 2020 at 1:23 AM Sugandh Huthanahally Mohan
> > > > > > > > <sugandhgowda at gmail.com> wrote:
> > > > > > > > >
> > > > > > > > > Hi Krishna,
> > > > > > > > >
> > > > > > > > > Currently I am not running wpa_supplicant.
> > > > > > > > > Is it necessary for hostapd?
> > > > > > > >
> > > > > > > > Well, not necessarily, but hostap side it looks fine, so,
> > > > > > > > we need to see why the STA isn't responding well, probably a
> > > > > > > > different draft implementation. At the very least we should start
> > > > > > > > with a sniffer capture of the association b/w the stations to check
> > > > > > > > the IEs.
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Thanks,
> > > > > > Regards,
> > > > > > Chaitanya T K.
> > > >
> > > >
> > > >
> > > > --
> > > > Thanks,
> > > > Regards,
> > > > Chaitanya T K.