svn commit: r336203 - in head: contrib/wpa contrib/wpa/hostapd contrib/wpa/hs20/client contrib/wpa/patches contrib/wpa/src/ap contrib/wpa/src/common contrib/wpa/src/crypto contrib/wpa/src/drivers c...

Fri Jul 20 00:25:09 UTC 2018

On Thu, Jul 19, 2018 at 6:21 PM, Kyle Evans <kevans at freebsd.org> wrote:
> On Thu, Jul 19, 2018 at 4:33 PM, Cy Schubert <Cy.Schubert at cschubert.com> wrote:
>> In message <201807192114.w6JLEapA097589 at slippy.cwsent.com>, Cy Schubert
>> writes:
>>> In message <17042686.Mc0X0P6XHu at asus.theweb.org.ua>, "Oleg V. Nauman"
>>> writes:
>>> > On Thursday, July 19, 2018 4:54:42 PM EEST Cy Schubert wrote:
>>> > > In message <CACNAnaEaMtRfHTXyCcGGmVP_37=gjLfDjPD5yd1gZqmpC0Z0Rg at mail.gma
>>> > > il.com>
>>> > >
>>> > > , Kyle Evans writes:
>>> > > > On Thu, Jul 19, 2018 at 7:13 AM, Niclas Zeising
>>> > > >
>>> > > > <zeising+freebsd at daemonic.se> wrote:
>>> > > > > [ sending this again since I missed the list the first time, apologie
>>> s
>>> > > > > if
>>> > > > > anyone receives a duplicate ]
>>> > > > >
>>> > > > > On 07/19/18 13:57, Kyle Evans wrote:
>>> > > > >> On Thu, Jul 19, 2018 at 4:51 AM, Alexey Dokuchaev <danfe at freebsd.org
>>> >
>>> > > > >>
>>> > > > >> wrote:
>>> > > > >>> On Thu, Jul 19, 2018 at 11:48:03AM +0300, Andrey V. Elsukov wrote:
>>> > > > >>>> ...
>>> > > > >>>> Yesterday I updated my notebook (with iwm(4)) and also noticed tha
>>> t
>>> > > > >>>> wi-fi connection periodically breaks. /etc/rc.d/wpa_supplicant
>>> > > > >>>> restart
>>> > > > >>>> wlan0 helps. After your message I reinstalled wpa_supplicant from
>>> ol
>>> > d
>>> > > > >>>> source and now it works stable already about 2 hours.
>>> > > > >>>
>>> > > > >>> So, right now, we have broken wpa_supplicant(8) in -CURRENT? :-/
>>> > > > >>
>>> > > > >> Well, "broken". It's incredibly stable outside of rekeying events, a
>>> nd
>>> > > > >> further testing shows that I don't actually notice these disconnects
>>> > > > >> most of the time because it reassociates fast enough. I noticed it t
>>> he
>>> > > > >> first time because apparently I had both SSIDs from my AP uncommente
>>> d
>>> > > > >> in my wpa_supplicant.conf and it decided at that point to connect to
>>> > > > >> the other one, which took a little longer.
>>> > > > >>
>>> > > > >> Contrary to Andrey's report, though, I don't have to kick
>>> > > > >> wpa_supplicant at all. It will reassociate on its own every single
>>> > > > >> time.
>>> > > > >
>>> > > > > Hi!
>>> > > > > I have the exact same problem as Andrey, with the same driver.  I've
>>> no
>>> > t
>>> > > > > investigated very much, but when using the 2.8 wpa_supplicant the wif
>>> i
>>> > > > > network dies after a little while, and I have to restart it (usually
>>> > > > > with
>>> > > > > /etc/rc.d/netif restart).  Then it works for a little while, before
>>> > > > > going
>>> > > > > down again.  With the old wpa_supplicant I didn't have this problem.
>>> > > > >
>>> > > > > I don't have very much else to add except noting that I'm affected as
>>> > > > > well.
>>> > > > > I haven't had time to debug it properly (which is why I've never
>>> > > > > reported
>>> > > > > it)
>>> > > >
>>> > > > I plan on trying out the latest from upstream beyond the patch Cy sent
>>> > > > along earlier to see if it's perhaps been addressed elsewhere in the
>>> > > > past two years since this release was made.
>>> > >
>>> > > A point of reference. I've had no issues here with any of the networks
>>> > > I use. All the networks I use are either WPA-PSK or open. The last
>>> > > WPA-EAP I used was at former $JOB a few years ago. However, at the Link
>>> > > Lounge just outside where $JOB is at my wifi would disconnect every 30
>>> > > minutes using our old wpa 2.5, requiring a netif restart. 2.6 resolved
>>> > > that issue.
>>> > >
>>> > > Upline git commit 0adc9b28b39d414d5febfff752f6a1576f785c85 also looks
>>> > > interesting.
>>> > >
>>> > > ommit 0adc9b28b39d414d5febfff752f6a1576f785c85
>>> > > Author: Jouni Malinen <j at w1.fi>
>>> > > Date:   Sun Oct 1 12:32:57 2017 +0300
>>> > >
>>> > >     Fix PTK rekeying to generate a new ANonce
>>> > >
>>> > >     The Authenticator state machine path for PTK rekeying ended up
>>> > > bypassing
>>> > >     the AUTHENTICATION2 state where a new ANonce is generated when going
>>> > >     directly to the PTKSTART state since there is no need to try to
>>> > >     determine the PMK again in such a case. This is far from ideal
>>> > > since the
>>> > >     new PTK would depend on a new nonce only from the supplicant.
>>> > >
>>> > >     Fix this by generating a new ANonce when moving to the PTKSTART
>>> > > state
>>> > >     for the purpose of starting new 4-way handshake to rekey PTK.
>>> > >
>>> > >     Signed-off-by: Jouni Malinen <j at w1.fi>
>>> > >
>>> > >
>>> > > I suspect a timeout because reason=1 in Kyle's log.
>>> >
>>> >
>>> >  I have two systems experienced wifi connection issues after recent HEAD
>>> > update.
>>> >  Both of them experiencing frequent up/down wlan0 events on boot so wireles
>>> s
>>> > connection can not negotiate DHCP requests, possibly due to fact that both
>>> > connecting to the same AP.
>>> >  AP capabilities list:
>>> >
>>> > *****   f8:1a:67:56:16:16    1   54M  -74:-96   100 EPS  WPA WME ATH WPS
>>> >
>>> > Interesting enough that switching wpa_supplicant to version 2.6 from ports
>>> > fixes that issue completely.
>>> >
>>> > Hopefully it helps.
>>> >
>>> >  Thank you.
>>>
>>> I've imported all the patches in the port, from our upline into base.
>>> Some were already committed to
>>> --
>>> Cheers,
>>> Cy Schubert <Cy.Schubert at cschubert.com>
>>> FreeBSD UNIX:  <cy at FreeBSD.org>   Web:  http://www.FreeBSD.org
>>>
>>>       The need of the many outweighs the greed of the few.
>>>  base 2.5 others not. This should bring base up to par with the port,
>>> address the remaining security issues, and probably fix this thread too.
>>
>> exmh. I had my cursor in the wrong place when I hit send.
>>
>> I've imported all the patches in the port, from our upline into base.
>> Some were already committed to base 2.5 others not. This should bring
>> base up to par with the port, address the remaining security issues,
>> and probably fix this thread too.
>>
>
> FWIW- with ports 2.6 I've confirmed that instead of the reassociation I get:
>
> Jul 19 18:17:30 shiva wpa_supplicant[34199]: wlan0: WPA: Group
> rekeying completed with ... [GTK=CCMP]
>
> I'll try with base 2.6 now that you've updated with all of these patches.

Alright, base 2.6 is still no good here. I note that there's still
some diff between ports and base [1] (about 252 lines of diff to sort
through, nothing serious... I removed the obviously-for-libressl
diff).

Some of it looks kind of suspicious, but I'd guess the changes in
./src/rsn_supp/wpa.c are mostly what make the difference for me.   How
much of this really needs to stick around, given that ports
wpa_supplicant is actually pretty stable?

[1] https://people.freebsd.org/~kevans/wpa.diff