|
Bugzilla – Full Text Bug Listing |
| Summary: | wpa_supplicant packaged incorrectly results in madwifi (ath_pci) support missing | ||
|---|---|---|---|
| Product: | [openSUSE] openSUSE 11.0 | Reporter: | Felix Möller <felix> |
| Component: | Network | Assignee: | Helmut Schaa <hschaa> |
| Status: | RESOLVED FIXED | QA Contact: | E-mail List <qa-bugs> |
| Severity: | Major | ||
| Priority: | P5 - None | CC: | ro |
| Version: | Alpha 2 | ||
| Target Milestone: | --- | ||
| Hardware: | Other | ||
| OS: | Other | ||
| Whiteboard: | |||
| Found By: | --- | Services Priority: | |
| Business Priority: | Blocker: | --- | |
| Marketing QA Status: | --- | IT Deployment: | --- |
| Attachments: |
here the log of a successfull connect
here the log with newest wpa_supplicant patch to enable madwifi support in wpa_supplicant again |
||
|
Description
Felix Möller
2007-11-19 10:37:10 UTC
Created attachment 183874 [details]
here the log of a successfull connect
Created attachment 183875 [details]
here the log with newest wpa_supplicant
wpa_supplicant in Factory has a bunch of patches applied for the new NetworkManager to come. I don't know whether the old NM is supposed to still work with the updated wpa_supplicant, but I suppose it will continue to work as soon as the NM update is checked in. However, reassigning to Tambet in case I am wrong. There should be no reason for the new wpa_supplicant to not work with older NetworkManager (or any other software using it). These new patches only implement new interface for external programs to communicate with wpa_supplicant, they do not change how wpa_supplicant works. If something broke, it has to be a bug with these patches and it's time to use the new debugging ability also provided by these patches. :) It seems, the problem is that the wpa_supplicant in Factory does not support madwifi driver back-end anymore. I did not disable it, it seems a patch or a flaw in the spec file breaks it. But that's not really a problem as it is obsolete anyway. I suppose NM will use wext backend for any wireless driver, right? Yep. I will close it once I can submit the new NetworkManager (still waiting for a required dhcp-client change to get submitted). *** Bug 343249 has been marked as a duplicate of this bug. *** Ok network manager is updated now but that gives way more problem than before ... <info> Activation (wlan0/wireless): connection 'Auto freakshow' has security, and secrets exist. No new secrets needed. <info> Config: added 'ssid' value 'freakshow' <info> Config: added 'key_mgmt' value 'WPA-PSK' <info> Config: added 'psk' value '<omitted>' <info> Config: added 'proto' value 'WPA RSN' <info> Config: added 'pairwise' value 'TKIP CCMP' <info> Config: added 'group' value 'WEP40 WEP104 TKIP CCMP' nm_supplicant_interface_set_config: assertion `NM_IS_SUPPLICANT_INTERFACE (self)' failed <WARN> real_act_stage2_config(): Activation (wlan0/wireless): couldn't send wireless configuration to the supplicant. <info> Trying to start the supplicant... last message repeated 4 times <info> Trying to start the supplicant... # rpm -qf /usr/bin/nm-tool NetworkManager-0.7.0-2 # LANG=CC nm-tool /usr/bin/nm-tool: line 101: cd: /usr/src/packages/BUILD/NetworkManager-0.7.0/test: No such file or directory gcc: nm-tool.o: No such file or directory gcc: ../libnm-glib/.libs/libnm_glib.so: No such file or directory gcc: /usr/src/packages/BUILD/NetworkManager-0.7.0/libnm-util/.libs/libnm-util.so: No such file or directory gcc: ../libnm-util/.libs/libnm-util.so: No such file or directory You'll need to upgrade the wpa_supplicant and dbus packages as well to get NM working. The nm-tool issue looks like a packaging bug. I am running factory as of today with just wpa_supplicant downgraded. Factory seems to be downgraded to nm 0.6.5 again. See ftp://ftp.gwdg.de/pub/opensuse/distribution/SL-OSS-factory/inst-source/suse/i586/ Latest factory broke my internet again and I investigated some more. With wpa_supplicant-0.5.8-33 (10.3) I am able to connect with wpa_supplicant: # wpa_supplicant -Dmadwifi -iwlan0 -c/etc/wpa_supplicant/wpa_supplicant.conf.rpmsave Trying to associate with 00:18:f8:b8:b6:24 (SSID='freakshow' freq=2437 MHz) Associated with 00:18:f8:b8:b6:24 WPA: Key negotiation completed with 00:18:f8:b8:b6:24 [PTK=TKIP GTK=TKIP] CTRL-EVENT-CONNECTED - Connection to 00:18:f8:b8:b6:24 completed (auth) [id=0 id_str=] With wpa_supplicant-0.5.8-72 I have to use -Dwext and I am not able to connect. (-Dwext fails with wpa_supplicant-0.5.8-33 too) # wpa_supplicant -Dwext -iwlan0 -c/etc/wpa_supplicant/wpa_supplicant.conf.rpmsave Trying to associate with 00:18:f8:b8:b6:24 (SSID='freakshow' freq=2437 MHz) Authentication with 00:00:00:00:00:00 timed out. Trying to associate with 00:18:f8:b8:b6:24 (SSID='freakshow' freq=2437 MHz) Authentication with 00:00:00:00:00:00 timed out. Trying to associate with 00:18:f8:b8:b6:24 (SSID='freakshow' freq=2437 MHz) Authentication with 00:00:00:00:00:00 timed out. Trying to associate with 00:18:f8:b8:b6:24 (SSID='freakshow' freq=2437 MHz) Therefore the madwifi driver seems not to work with -Dwext. I am using SVN r3122 of madwifi. (In reply to comment #5 from Joachim Gleissner) > It seems, the problem is that the wpa_supplicant in Factory does not support > madwifi driver back-end anymore. I did not disable it, it seems a patch or a > flaw in the spec file breaks it. It is the SOURCES/config file which breaks it. It does not have a line break at the end and therefore the "echo "CONFIG_DRIVER_MADWIFI=y" >>.config" adds it to the last line. @jg could you please fix this, because otherwise wpa_supplicant does not work for me. (see last comment) Ok as knetworkmanager does nothing at all I tried with nm-applet, but got the following: <info> Activation (wlan0) Stage 1 of 5 (Device Prepare) scheduled... <info> Activation (wlan0) Stage 1 of 5 (Device Prepare) started... <info> Activation (wlan0) Stage 2 of 5 (Device Configure) scheduled... <info> Activation (wlan0) Stage 1 of 5 (Device Prepare) complete. <info> Activation (wlan0) Stage 2 of 5 (Device Configure) starting... <info> Activation (wlan0/wireless): connection 'Auto freakshow' has security, and secrets exist. No new secrets needed. <info> Config: added 'ssid' value 'freakshow' <info> Config: added 'key_mgmt' value 'WPA-PSK' <info> Config: added 'psk' value '<omitted>' <info> Config: added 'proto' value 'WPA RSN' <info> Config: added 'pairwise' value 'TKIP CCMP' <info> Config: added 'group' value 'WEP40 WEP104 TKIP CCMP' <info> Activation (wlan0) Stage 2 of 5 (Device Configure) complete. <info> Config: set interface ap_scan to 1 <info> (wlan0) Supplicant interface state change: 0 -> 2 <info> (wlan0) Supplicant interface state change: 2 -> 0 <info> (wlan0) Supplicant interface state change: 0 -> 2 <info> (wlan0) Supplicant interface state change: 2 -> 3 <info> (wlan0) Supplicant interface state change: 3 -> 0 <info> (wlan0) Supplicant interface state change: 0 -> 2 <info> (wlan0) Supplicant interface state change: 2 -> 3 <info> Activation (wlan0/wireless): association took too long, asking for new key. <info> (wlan0) Supplicant interface state change: 3 -> 0 What are the state changes? Is this expected? I can connect with the wpa_gui without a problem. Should I open another bugreport for the spec-file issue of comment #12 or can somebody here just add the missing return, please? Created attachment 191031 [details]
patch to enable madwifi support in wpa_supplicant again
a patch to enable madwifi support in wpa_supplicant working again. Just to make the issue really clear.
Re-assigning to wpa_supplicant maintainer. With the following patch (and wpa_supplicant patched as mentioned above) I am able to connect with nm-applet:
--- nm-supplicant-interface.c.orig 2008-01-20 19:52:26.000000000 +0100
+++ nm-supplicant-interface.c 2008-01-20 19:52:50.000000000 +0100
@@ -767,7 +767,7 @@
driver = g_new0 (GValue, 1);
g_value_init (driver, G_TYPE_STRING);
- g_value_set_string (driver, priv->is_wireless ? "wext" : "wired");
+ g_value_set_string (driver, priv->is_wireless ? "madwifi" : "wired");
g_hash_table_insert (hash, "driver", driver);
call = dbus_g_proxy_begin_call (proxy,
This gives the followig in /var/log/NetworkManager:
<info> Activation (wlan0) Stage 2 of 5 (Device Configure) complete.
<info> Config: set interface ap_scan to 1
<info> (wlan0) Supplicant interface state change: 1 -> 2
<info> (wlan0) Supplicant interface state change: 2 -> 3
<info> (wlan0) Supplicant interface state change: 3 -> 4
<info> (wlan0) Supplicant interface state change: 4 -> 5
<info> (wlan0) Supplicant interface state change: 5 -> 6
<info> (wlan0) Supplicant interface state change: 6 -> 7
<info> Activation (wlan0/wireless) Stage 2 of 5 (Device Configure) successful. Connected to wireless network 'freakshow'.
<info> Activation (wlan0) Stage 3 of 5 (IP Configure Start) scheduled.
<info> Activation (wlan0) Stage 3 of 5 (IP Configure Start) started...
<info> Activation (wlan0) Beginning DHCP transaction.
<info> dhclient started with pid 25176
I will rename this report to to get the wpa_supplicant issue fixed and open a bug later for NetworkManager not working with madwifi (ath_pci) driver any longer. Helmut the solution is in comment #14. Thanks Felix :) Submitted to STABLE. |