Bug 30857 - wpa_supplicant can´t connect though ciphers shall be supported
Summary: wpa_supplicant can´t connect though ciphers shall be supported
Status: RESOLVED OLD
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 8
Hardware: All Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Mageia Bug Squad
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-09-17 00:18 CEST by Elmar Stellnberger
Modified: 2024-06-22 19:37 CEST (History)
4 users (show)

See Also:
Source RPM:
CVE:
Status comment:


Attachments
iwlist wlp58s0 scanning (26.25 KB, text/plain)
2022-09-17 00:19 CEST, Elmar Stellnberger
Details
lspci: Atheros wireless adapter (2.13 KB, text/plain)
2022-09-17 00:19 CEST, Elmar Stellnberger
Details
wpa_pareus2.conf (243 bytes, text/plain)
2022-09-17 00:20 CEST, Elmar Stellnberger
Details
wpa_supplicant2.msg (39.95 KB, text/plain)
2022-09-17 00:21 CEST, Elmar Stellnberger
Details
wpa_pareus3.conf (139 bytes, text/plain)
2022-09-17 00:21 CEST, Elmar Stellnberger
Details
wpa_supplicant3.msg (31.69 KB, text/plain)
2022-09-17 00:21 CEST, Elmar Stellnberger
Details
lspci -vn (14.31 KB, text/plain)
2022-09-17 23:40 CEST, Elmar Stellnberger
Details
iw dev wlp58s0 scan (new non-deprecated scan tool) (41.16 KB, text/plain)
2022-09-18 22:27 CEST, Elmar Stellnberger
Details
wpa_supplicant messages for empty essid and wpa_pareus2.conf (32.08 KB, text/plain)
2022-09-19 10:26 CEST, Elmar Stellnberger
Details
wpa_pareus1.conf (195 bytes, text/plain)
2022-09-19 10:28 CEST, Elmar Stellnberger
Details
wpa_supplicant messages for empty essid and wpa_pareus1.conf (32.37 KB, text/plain)
2022-09-19 10:29 CEST, Elmar Stellnberger
Details
iwlist wlp2s0 scanning, 2023-05-18, Atombook (25.55 KB, text/plain)
2023-05-18 08:52 CEST, Elmar Stellnberger
Details
output of wpa_supplicant & iwconfig: access point associated, IP given (1.46 KB, text/plain)
2023-05-18 08:54 CEST, Elmar Stellnberger
Details
dmesg of successful connection (1.60 KB, text/plain)
2023-05-18 08:55 CEST, Elmar Stellnberger
Details
wpa_pareus1.conf I used for setting up the successful connection (214 bytes, text/plain)
2023-05-18 08:57 CEST, Elmar Stellnberger
Details
iw dev wlp2s0 scan, 2023-05-18 (43.37 KB, text/plain)
2023-05-18 09:12 CEST, Elmar Stellnberger
Details

Description Elmar Stellnberger 2022-09-17 00:18:26 CEST
When I try to use NetworkManager or the mcc-app to connect, even to the unencrypted Pareus network it doesn´t work: can´t connnect. I have to quit both and connect manually to the unencrypted Pareus network with iwlist wlp58s0 scanning, iwconfig essid & ap set. However it should also be possible to connect using wpa_supplicant with the supported ciphers. However I get "no suitable network found" although WPA2-PSK/TKIP is claimed to be supported by both sides. Possibly it could be a problem that I am using -Dnl80211 since -Dbsd as recommended for Atheros Wifi chips is not supported:

$ wpa_supplicant -Dbsd -iwlp58s0 -cwpa_pareus.conf
Successfully initialized wpa_supplicant
wlp58s0: Unsupported driver 'bsd'

Now see the -dd logs with -Dnl80211
Comment 1 Elmar Stellnberger 2022-09-17 00:19:25 CEST
Created attachment 13372 [details]
iwlist wlp58s0 scanning
Comment 2 Elmar Stellnberger 2022-09-17 00:19:59 CEST
Created attachment 13373 [details]
lspci: Atheros wireless adapter
Comment 3 Elmar Stellnberger 2022-09-17 00:20:36 CEST
Created attachment 13374 [details]
wpa_pareus2.conf
Comment 4 Elmar Stellnberger 2022-09-17 00:21:04 CEST
Created attachment 13375 [details]
wpa_supplicant2.msg
Comment 5 Elmar Stellnberger 2022-09-17 00:21:29 CEST
Created attachment 13376 [details]
wpa_pareus3.conf
Comment 6 Elmar Stellnberger 2022-09-17 00:21:55 CEST
Created attachment 13377 [details]
wpa_supplicant3.msg
Comment 7 Elmar Stellnberger 2022-09-17 00:24:28 CEST
Can it be that there are differences between WPA2-PSK/TKIP and WPA2-PSK/TKIP or why does it not find a suitable network. Output is -dd so all debug info should be there. Also it is somewhat shocking that you can not assign an ESSID manually if either NetworkManager is running or the MCC Networking app is in the system tray. You would think that the wifi driver is at fault if you don´t know that.
Comment 8 Elmar Stellnberger 2022-09-17 00:28:52 CEST
I wonder how to use the other ciphers with wpa_supplicant like GMAC-256, GMAC-128 or CMAC - https://www.wikieduonline.com/wiki/Supported_Ciphers .
Comment 9 Marja Van Waes 2022-09-17 13:41:40 CEST
This is beyond my comprehension, CC'ing all packagers collectively for help.

Note that iwlist is deprecated, but we still use it, see bug 17706

CC: (none) => marja11, pkg-bugs

Comment 10 Elmar Stellnberger 2022-09-17 18:42:04 CEST
I think this bug is different than bug 17706, since iwlist reports all access points correctly. Why is iwlist considered deprecated and what replacement shall there be for it? Wpa_supplicant fails to set the ESSID while iwconfig does that correctly. Without iwlist and iwconfig I wouldn´t have internet access by now. See also bug 30859.
Comment 11 Elmar Stellnberger 2022-09-17 18:52:03 CEST
note: I can test for this Wifi until Saturday/Sunday next week. Then I will return home from Caorle/Italy and I won´t have the possibility to test for it until in about half a year. You could test for it on your own if there was someone of you here who tried to connect in Altanea/BaiaBlu.
Comment 12 Elmar Stellnberger 2022-09-17 22:34:54 CEST
The encrypted Pareus wlan is available at Baiablu, at the Pareus apartements the free-unencrypted and the encrypted wifi access points are present.
Comment 14 Elmar Stellnberger 2022-09-18 22:27:38 CEST
Created attachment 13387 [details]
iw dev wlp58s0 scan (new non-deprecated scan tool)
Comment 15 Elmar Stellnberger 2022-09-19 10:26:56 CEST
Created attachment 13391 [details]
wpa_supplicant messages for empty essid and wpa_pareus2.conf

I have now tried it with an empty ssid since it was my guess that wpa_supplicant only cares about the real essid (empty for encrypted wlans) and not the OWE transition mode essid (non-empty for encrypted wlans). However that did not work either. As it seems the problem may also not be related to essids.
Comment 16 Elmar Stellnberger 2022-09-19 10:28:17 CEST
Created attachment 13392 [details]
wpa_pareus1.conf
Comment 17 Elmar Stellnberger 2022-09-19 10:29:06 CEST
Created attachment 13393 [details]
wpa_supplicant messages for empty essid and wpa_pareus1.conf
Comment 18 Lewis Smith 2022-09-20 09:48:55 CEST
I agree with Marja, we need some comment from more expert people.

> the unencrypted Pareus network
> who tried to connect in Altanea/BaiaBlu
> The encrypted Pareus wlan is available at Baiablu, at the Pareus apartements
All this is cryptic. What is 'Pareus network'? What are 'Altanea' and 'BaiaBlu'?

Are you saying that your connection problems relate to just one specific network, or are more general?

CC: (none) => lewyssmith

Comment 19 sturmvogel 2022-09-20 10:37:48 CEST
(In reply to Lewis Smith from comment #18)
> Are you saying that your connection problems relate to just one specific
> network, or are more general?

"Pareus" is the name of a messed up wifi in a beach ressort in italy where Elmar makes vacation atm.
Comment 20 Elmar Stellnberger 2022-09-20 10:52:35 CEST
The Wifi works with iPhones and Android Phones, with MacOS computers (tested) and  very very likely also with Windows computers. Only the Linux desktop seems to make problems here. I can not exclude that there are other Wifis like this one around the world since to me it makes sense to have an encrypted Wifi and an open fallback Wifi under the same effective SSID. Pareus can tell the people their SSID and it will work for both - the encrypted and the unencrypted WLAN. Altanea is a region in front of the beach near Caorle/Italy (There is Caorle, Porto Santa Margarita and Altanea with Pareus and PraDelleTorre, likely you will find them on an internet map). Baiablu is the operator at the freely accessible beach bar in front of the Pareus resort.
Comment 21 Elmar Stellnberger 2022-09-21 09:05:44 CEST
As it seems I will now stay until Sunday the 2nd October and I can test until at most one day before. If that should fail I will be here again by next spring. I have also tried to contact Jouni Malinen, but he seems to be too occupied to answer.
Comment 22 Elmar Stellnberger 2022-09-22 18:15:25 CEST
  Could it be that wpa_supplicant makes problems if you want to connect with WPA2/TKIP instead of WPA/TKIP or WPA2/CCMP? At least the official site does not give a sample configuration for that: https://hostap.epitest.fi/wpa_supplicant/ .
   What about GMAC-256, GMAC-128 or CMAC? Anybody has a sample configuration using that?
Comment 23 Elmar Stellnberger 2022-09-23 11:57:46 CEST
  It turned out that the website url referenced in the mg8 package header was stale: https://bugs.mageia.org/show_bug.cgi?id=30881
  Using this I have now asked for help upstreams: http://lists.infradead.org/pipermail/hostap/2022-September/040823.html
Comment 24 Lewis Smith 2022-09-23 20:55:37 CEST
(In reply to Elmar Stellnberger from comment #23)
> It turned out that the website url referenced in the mg8 package header
> was stale: https://bugs.mageia.org/show_bug.cgi?id=30881
This now raised as bug 30881.

CC'ing Thomas in the hope he can make some sense of this.

CC: (none) => tmb

Comment 25 Elmar Stellnberger 2022-09-28 17:02:14 CEST
If I wanna ask the Wifi provider to change something about his setup I will need to know precisely on why it does not work: Is it because of the combination of WPA2 and TKIP or because of the ESSID configuration? If I can know this by Friday or a bit later, maybe they will change it for me.
Comment 26 Elmar Stellnberger 2023-05-18 08:52:58 CEST
Created attachment 13846 [details]
iwlist wlp2s0 scanning, 2023-05-18, Atombook

Seems resolved with wpa_supplicant-2.10-3.mga9, 6.3.2-desktop-3.mga9 (So 14 Mai 2023 17:23:13, Key ID b742fa8b80420f66). 'iwlist  wlp2s0 scanning' returns list with the given as well as empty essids, as it did last year. Nonetheless wpa_supplicant connects now without any problems and I am currently pasting this update via that wifi-connection.
Comment 27 Elmar Stellnberger 2023-05-18 08:54:50 CEST
Created attachment 13847 [details]
output of wpa_supplicant & iwconfig: access point associated, IP given
Comment 28 Elmar Stellnberger 2023-05-18 08:55:47 CEST
Created attachment 13848 [details]
dmesg of successful connection
Comment 29 Elmar Stellnberger 2023-05-18 08:57:38 CEST
Created attachment 13849 [details]
wpa_pareus1.conf I used for setting up the successful connection
Comment 30 Elmar Stellnberger 2023-05-18 08:58:13 CEST
.

Status: NEW => RESOLVED
Resolution: (none) => FIXED

Comment 31 Elmar Stellnberger 2023-05-18 09:11:20 CEST
No, it can´t be said that anything would have changed by the software. If you compare the iwlist scan results, there was an encrypted as well as an unencrypted Wifi with the same effective ssid last year. Now the effective essid of the unencrypted Wifis has been changed to 'apartement'. The fact that multiple ssdis get condensed into one by iwlist can be revealed by 'iw dev wlp2s0 scan'. I remember having posted and analysed this last year but I can neither find my scan results of last year in this report, nor can I find any program used for it on my hdd in the respective folder.

Ever confirmed: 1 => 0
Resolution: FIXED => (none)
Status: RESOLVED => UNCONFIRMED

Comment 32 Elmar Stellnberger 2023-05-18 09:12:37 CEST
Created attachment 13850 [details]
iw dev wlp2s0 scan, 2023-05-18
Comment 33 Elmar Stellnberger 2023-05-18 09:15:21 CEST
Ahh, here is the scan result of last year. Anyone interested in this issue can compare the scan results and find out whether the problem was really just due to two or three real essids being mapped onto one 'effective' essids. I guess that nothing else about the setup here would have changed.
Comment 34 Marja Van Waes 2024-06-22 19:37:49 CEST
This bug report was filed against Mageia 8, which is no longer supported https://blog.mageia.org/en/2023/12/30/mageia-8-end-of-life/

Therefore closing this report. 

Feel free to reopen and change Version: to 9, if the issue is still present in that maintained version.

Status: UNCONFIRMED => RESOLVED
Resolution: (none) => OLD


Note You need to log in before you can comment on or make changes to this bug.