Bug 164394

Created attachment 77135 [details]
/var/log/NetworkManager logs

I believe this is a duplicate.

*** This bug has been marked as a duplicate of 158892 ***

This bug is not a duplicate : along with ndiswrapper, NM does not see this kind of network (RC2 release) and does not show it in the list.

Can you confirm with the latest NM package and with the nm-applet (GNOME version), which is in NetworkManager-gnome (you can run it fine in KDE), instead of KNetworkManager?  We need to see if the problem is in the daemon or in the applet.

With NetworkManager-gnome-0.6.2-21 which is the latest package for SUSE 10.1 RC3, I have NO network connection after launching /usr/bin/nm-applet and enabling wireless.

Did you also update NetworkManager itself? The version that wrote the log file you attached used ap_scan 1, which can't work with ndiswrapper and hidden SSID. If yes, could you attach your /var/log/NetworkManager file again?

Gilles, ping?

Can you retest with the latest NM?  You can grab packages from here, if you desire:

    http://primates.ximian.com/~rml/misc/NetworkManager/

As jg said, the "AP_SCAN 1" that is in use in your logs will not work.  We should use "AP_SCAN 2", which NM has used for awhile.

Robert? here you are?

Good job! K-NM is working! I have successfully configured my Broadcom BCM4306 and now, I am able to switch between my Centrino 2100 and my BCM4306.

But, Gnome-NM is now broken! Whenever I tried to launch it in a terminal,
I get the following error message (under KDE): 

gilles@gilles-portable:~> /usr/bin/nm-applet

** (nm-applet:14954): WARNING **: Icon nm-no-connection missing: L'icône « nm-no-connection » n'est pas présente dans le thème

(nm-applet:14954): GLib-GObject-CRITICAL **: g_object_unref: assertion `G_IS_OBJECT (object)' failed

and, in a box, I can see the message :

The NetworkManager applet could not find some required resources.  It cannot continue.

I do not know very well the gnome environment, but many gnome packages are already installed and I have upgraded NetworkManager-glib and NetworkManager-gnome; here's the list :
gnome-icon-theme-2.12.1-25
gnome-mime-data-2.4.2-22
libgsf-gnome-1.13.99-13
gnome-spell2-1.0.6-22
gnome-filesystem-0.1-259
gnome-utils-2.12.2-32
gnome-vfs2-2.12.2-56
libgnomeprintui-2.12.1-20
gnome-audio-2.0.0-132
gnome-keyring-0.4.6-15
libgnomecanvas-2.12.0-19
libgnomesu-1.0.0-29
gnome-desktop-2.12.2-23
gnome2-user-docs-2.8.1-21
gnome-doc-utils-0.5.3-15
gnome-themes-2.12.1-33
NetworkManager-gnome-0.6.2-40
libgnomecups-0.2.2-21
libgnomeui-2.12.0-28
gnome-panel-2.12.2-40
libgnomeprint-2.12.1-20
libgnome-2.12.0.1-26
gnome-menus-2.12.0-29

Great news that your driver is fixed!

Let's try to get nm-applet running, now.

Try running

    gtk-update-icon-cache -f /opt/gnome/share/icons/Tango/
    gtk-update-icon-cache -f /opt/gnome/share/icons/hicolor/

As root.  Or, alternatively, run /sbin/SuSEconfig (also as root).

And then rerun nm-applet.

Ok, I have no more problem with gnome's icon, but Gnome-NM does not work : I have no connection, and I cannot find any log.

Can you expand on "does not work" ?

Can you restart the daemon (`/etc/init.d/network restart` as root) and make sure only one GNOME applet is running, and that KNM is not running?

If KNM works, nm-applet should work fine.  It might be that your GNOME setup is hosed..

Anyhow, your driver works, so I am happy of that :)

The driver works, but not after I have restarted linux. Here's the log

Created attachment 83975 [details]
Knetworkmanager logs after linux have restarted

So it no longer works, at all?

The log is incomplete / truncated.

The log is complete : there is nothing more after this dbus error

For Gnome-NM, I caught this in /var/log/messages :

May 18 00:17:01 gilles-portable network: Starting the NetworkManagerDispatcher
May 18 00:19:11 gilles-portable gconfd (gilles-10286): démarrage (version 2.12.1), pid 10286 utilisateur « gilles »
May 18 00:19:11 gilles-portable gconfd (gilles-10286): Adresse « xml:readonly:/etc/opt/gnome/gconf/gconf.xml.mandatory » résolue vers une source de configuration en lecture seule à la position 0
May 18 00:19:11 gilles-portable gconfd (gilles-10286): Adresse « xml:readwrite:/home/gilles/.gconf » résolue vers une source de configuration accessible en écriture à la position 1
May 18 00:19:11 gilles-portable gconfd (gilles-10286): Adresse « xml:readonly:/etc/opt/gnome/gconf/gconf.xml.defaults » résolue vers une source de configuration en lecture seule à la position 2
May 18 00:19:40 gilles-portable kernel: ADDRCONF(NETDEV_UP): eth0: link is not ready
May 18 00:20:11 gilles-portable gconfd (gilles-10286): GConf server is not in use, shutting down.
May 18 00:20:11 gilles-portable gconfd (gilles-10286): Exiting

Ugh you have some sort of DBUS problem, almost assuredly unrelated..

Now, I give again wireless parameters for BCM4306 interface to reactivate it and my centrino 2100 is no more working : No wireless network found.

Yah, the DBUS problem is preventing NM from seeing any devices, I suspect.

I don't know what went wrong.

(In reply to comment #14)

> Created an attachment (id=83975) [edit]
> Knetworkmanager logs after linux have restarted

If you were using KNetworkManager for this log you probably have a screwed KWallet configuration.  Please make sure that the KWallet subsystem is enabled (start kwalletmanager -> Settings -> Wallet Preferences -> Enable the KDE wallet subsystem).  Then restart KNetworkManager.  Please report if this makes the D-BUS error go away.

Other than that your system configuration looks quite groggy.

KDE wallet subsystem was already enabled, and I have to re-enable it again, and there is no more dbud error.

Now, whenever I restart Linux and after I give my wallet password, NM connects only with centrino 2100 interface. I can't see my network on BCM4306 side, even with AP_SCAN 2, and so I am not able to switch the interface.

Created attachment 84201 [details]
NM logs after re-enabled wallet

This has been fixed in SLES/SLED 10.

I don't know yet about a YOU fix for SUSE 10.1, but I built packages and put them up here:

    http://primates.ximian.com/~rml/misc/NetworkManager/


*** This bug has been marked as a duplicate of 178210 ***

Robert, I encounter the same behaviour than before and have almost the same logs : my hidden network is only seen by centrino ipw2100 interface, not with ndiswrapper/BCM4306 interface.

however, I am able to connect with BCM4306 interface whenever I choose " connect to another wireless network " menu and I give again informations to connect with ( SSID name, WPA type, WPA cyphering type ).

Thie bug is NOT a duplicate of bug #178210 but more a specific case with hidden wireless network and ndiswrapper

Created attachment 85377 [details]
Logs after may, 25 NetworkManager upgrade

If your card does not see the network, there is not much that NM can do, and your observed behavior is expected.

Robert, I have no problem with ifup method and ndiswrapper / wpa_supplicant, with same configuration parameters. Hence, this is not a hardware problem. Here's an output of iwlist command :

gilles-portable:~ # iwlist wlan0 scan
wlan0     Scan completed :
          Cell 01 - Address: 00:13:10:27:EC:CC
                    ESSID:"Gilles-17-Leclerc"
                    Protocol:IEEE 802.11b
                    Mode:Managed
                    Frequency:2.462 GHz (Channel 11)
                    Quality:0/100  Signal level:-55 dBm  Noise level:-256 dBm
                    Encryption key:on
                    Bit Rates:1 Mb/s; 2 Mb/s; 5.5 Mb/s; 6 Mb/s; 9 Mb/s
                              11 Mb/s; 12 Mb/s; 18 Mb/s; 24 Mb/s; 36 Mb/s
                              48 Mb/s; 54 Mb/s
                    Extra:bcn_int=100
                    Extra:atim=0
                    Extra:wpa_ie=dd180050f20101000050f20201000050f20201000050f2020000

NM and ifup do not do the same things.  NM exercises the driver quite heavily.

Because NM basis decisions on scanning, if the driver is broken w.r.t. scanning or more advanced wireless extensions, it will fail with NM but not necessarily ifup.

Did this work any better in 10.2?

Exactly same behaviour for my Linksys WPC54GS network card (that relies on ndiswrapper & Wds drivers) and my LAN 2100 centrino (that is supported directly by INTEL firmware) : NM takes 60s at 28% "Configuring device" and after that, it gives up and displays the not-connected icon.

I have to say that I am currently using ndiswrapper v1.37 and it works well for both Linksys and INTEL.

I am surprised as there has been no new version for Networkmanager between openSuse 10.1 and 10.2

... But I am not surprised about NetworkManager behaviour since its core still embeds older ndiswrapper version.

I am getting this in 10.3 also.. 

Networkmanager 0.7 has been completly rewritten and shipped with fedora 8. Maybe we will get this one for next release, in openSUSE 11 ?

2 months after opensuse 10.3 was out, I am surprised to not be able to find networkmanager v0.7 in any official or non official repository.

Anyway, nidswrapper is useless on 64 bits architectures, because windows drivers are 32 bits drivers. This solution has no future since 32 bits laptops have disappeared.

Actually, there are few mature open source solutions since GNU drivers are limited to no crypted or WEP crypted wireless networks (I.E. BCM43xx, RTL818X ...)

Just for the record: One may argue about the usefulness of ndiswrapper, but it isn't especially useless on 64 bit platforms, as there are 64 bit Windows drivers, although not that many. And with the introduction of the new wireless LAN stack mac80211 (for which ports of the bcm43xx and rtl818x you mentioned already exist), drivers are improving quickly and finally offer full WPA support.

*** Bug 261588 has been marked as a duplicate of this bug. ***

Is this bug still relevant? NetworkManager 0.7 has been part of opensuse since 11.0, bcm43xx driver has been in linux kernel since 2.6.17-rc2.

NetworkManager 0.7 still exhibits this behaviour with ndiswrapper 1.52 / kernel 2.6.25.18

I am using b43 module (kernel 2.6.25.18) to get wireless connection, but this module is instable : sometimes, he's eating much CPU for half a second and often disconnect. However NetworkManager gets automatically the connection back.

I'll have a look at new openSUSE 11.1 when it'll be released.

Same situation on openSUSE 11.1, on same wireless environment [however, I have removed my Linksys router - WRT54GS to use only a Freebox v5, a internet provider network gateway, configured both as a router and wireless acces point].

- My centrino LAN 2100 3B (ipw2100) does not work out of box. It needs "wireless" configuration file to tell wpa_supplicant with WIRELESS_AP_SCANMODE=2 and works only with "ifup" / kinternet managers, not with NetworkManager.

- My Linksys WPS54GS (ssb / b43) works only with NetworkManager, not with ifup / kinternet managers.

NetworkManager does not automatically connect, because it tries eth0 (ipw2100) first, and fails. It seems to process my wlan0 (b32) connection, but also fails: I have to relaunch manually on wlan0 to get connected.

I got also many messages reported by kernel, like :

Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:34 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:41 linux-hpkv kernel: __ratelimit: 123 callbacks suppressed
Jan  4 09:48:41 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:41 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:41 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:41 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:41 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:41 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error
Jan  4 09:48:41 linux-hpkv kernel: b43-phy0 ERROR: PHY transmission error

But the wireless connection with b43 is more stable using videolan vlc (no connectivity loss after fiew seconds). Streaming Stops after 10/20 minutes, and tries to restart but vlc is out of sync. Sometimes streaming stutters. I have to manually reconnect to the network after 20 minutes.

Is this solved in current versions? 

(just shaking old bugs during openSUSE Bug Day)

set to needinfo

I am currently using openSUSE 11.3 and have switched a long time ago to ethernet cable network, since wireless connections were slow and used to eat cpu resources.

Today, i give a try to my wireless network with bcm4306 (linksys wpc54g hardware) and a freebox v5 from my ISP, and have configured a masked network name.

This behaves now nicely! So, you can close this bug report at the end.

(In reply to comment #44)
...
> This behaves now nicely! So, you can close this bug report at the end.

Thanks for the feedback Gilles.