Extra symbols returned by check-exports when building 0.8.4 on Lucid.
Hi, As discussed on IRC; I'm finding some extra symbols when I try to build 0.8.4 on Lucid. I suspect these are due to the older kernel and older GCC, but I admit I haven't really dug into the exact cause so much. The extra symbols are returned by the objdump -t $so | grep [.]hidden.* command run in tools/check-exports.sh: 00040ff4 l O *ABS* .hidden _GLOBAL_OFFSET_TABLE_ 00041380 l O .data .hidden __dso_handle 000406e8 l O .dtors .hidden __DTOR_END__ fbe4 l F .text .hidden __i686.get_pc_thunk.cx 0002f980 l F .text 0014 .hidden __stack_chk_fail_local b1e7 l F .text .hidden __i686.get_pc_thunk.bx 00040e60 l O *ABS* .hidden _DYNAMIC This was returned by the same objdump command and grep on i386. The symbols on amd64 are similar, with a slight difference: ../tools/check-exports.sh ./.libs/libnm-util.so ./libnm-util.ver ./.libs/libnm-util.so: checking exported symbols against ./libnm-util.ver --- ./libnm-util.ver2011-05-04 13:34:48.0 + +++ - 2011-05-04 13:37:42.782033236 + @@ -1,5 +1,10 @@ { global: + _DYNAMIC; + _GLOBAL_OFFSET_TABLE_; + __DTOR_END__; + __dso_handle; + atexit; nm_connection_add_setting; nm_connection_clear_secrets; nm_connection_compare; I suspect atexit also comes from the hidden symbols, but I haven't been able to verify this yet (time to build a x86_64 VM). Is there any way to properly exclude them so tests succeed? Regards, Mathieu Trudel-Lapierre mathieu...@gmail.com Freenode: cyphermox, Jabber: mathieu...@gmail.com 4096R/EE018C93 1967 8F7D 03A1 8F38 732E FF82 C126 33E1 EE01 8C93 ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Question
Hi people, I have a network that I want to connect to. It is WEP encrypted. I am not sure but the key is FF:AA:DD:BB:00. How should I type in this key in the dialog box? Thank you all! ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: Auto-connect OFF option.
On Mon, 2011-05-02 at 20:44 +0200, Noes1s wrote: Im using this script to disable the autoconnect function on all my stored networks: ### Disables auto-connect on all saved networks netlist=$(ls .gconf/system/networking/connections/ | grep -v %) for l in ${netlist[*]}; { gconftool-2 --type bool --set /system/networking/connections/$l/connection/autoconnect false ;} 1) Can you add a button with this option? Probably not; I'm not sure the functionality would be widely used enough to justify one. That doesn't mean your needs are not valid, just that this is the first time I've heard of anyone requesting this, leading me to conclude that it's not a widely requested behavior at this time. 2) How can i detect the latest network added to disable the autoconnect on it instead of do it on all of them? (This is implemented on the program but dont know what code is used) NM emits D-Bus signals when new connections are added to system settings, and (for 0.8 and earlier) the user settings service also does this. You are probably interested in listening to the user settings service on D-Bus for the NewConnection signal. 3) How can i make this run everytime i connect to a network? (Tried using a new script on /etc/NetworkManager/dispatcher.d without luck) Dispatcher scripts run in root context, while the connections you're probably interested in are stored in the users session. Thus the dispatcher scripts simply cannot access that user information. The best bet is probably to run a small program in the user's session that listens for new network connections, and then runs your script whenever a new one shows up. Dan ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: bluetooth DUN silently discarding invalid APNs
On Tue, 2011-05-03 at 12:33 +0100, Marc Herbert wrote: Hi, I wasted a number of hours when trying to tether using bluetooth... it seems any APN containing an underscore _ causes the DUN configuration entered into the gnome bluetooth wizard to be *SILENTLY* discarded. APNs are defined by GSM 03.03 section 14.9 which says: http://www.3gpp.org/ftp/Specs/html-info/0303.htm = The syntax of the APN shall follow the Name Syntax defined in RFC 2181 [14] and RFC 1035 [15]. The APN consists of one or more labels. Each label is coded as one octet length field followed by that number of octets coded as 8 bit ASCII characters. Following RFC 1035 [15] the labels should consist only of the alphabetic characters (A-Z and a-z), digits (0-9) and the dash (-). The case of alphabetic characters is not significant. The APN is not terminated by a length byte of zero. = Note that none of the APNs in the mobile broadband provider database contain an underscore. But the specification does use the word should, which implies that APNs may deviate from the suggestion. Many APNs already use '.' (which the specification does not suggest) and perhaps we should allow _ too. Can anyone reproduce this? Only a bluetooth phone is needed, plus deleting the bluetooth configuration for this phone if you already have one (sorry), so you can run the wizard on it again. You do not even need a valid network subscription to reproduce this problem. I am using NetworkManager 0.8.4 in Fedora 14. Since the APN is the hostname of the GGSN or PDN gateway, I guess this validation tries to apply the restrictions of RFC 1123 concerning hostnames (note that, as opposed to a common misconception, the DNS itself does not have any such restriction, see section 11 in RFC 2181. DNS is not just for hostnames.) I see extremely little value in this validation. There are millions of other and more likely typos that it will never catch. Since it does not even issue an error message but silently discard the user input instead, the little value that ever was intended is completely gone. This validation feature has now become a severe bug since it hides the next and proper error message (i.e., connection failed, check your settings). And wastes hours. Even worse, wvdial is perfectly able to get me online using an APN that includes an underscore. So whatever the standards say, this validation prevents some configurations to work. By the way it is not possible to enter a blank APN either (asking the network use the default APN). This again works perfectly with wvdial. And this is valid. Yes, it's valid, but note that the default APN is stored in the *device*, not the SIM card, and has no relation to the SIM card at all. So if you ever swap SIM cards, or use a different provider, then the APN is surely going to be wrong and the dialing will fail. However, I've been thinking of ways to enable using the default APN since that works for some phones that don't allow setting the APN at all via AT commands, but where dialing works fine. Dan ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: Howto change port used for 3g modem?
On Tue, 2011-05-03 at 15:21 +0100, pera...@portugalmail.pt wrote: Dan, Seems as though I was trying to build from very old source. I checked out current git and built it. I had a package dependency problem with the applet wanting gtk3 2.91.4 or better whereas fedora 14 has 2.90.5. So I used gtk3 from fedora15 which allowed be to build it but this update and all its dependencies messed up the rest of my fedora 14 system. Attached is the system log messages I receive now when plugging the modem in. The modem is from Orange with brand Alcatel. Inside back cover it says: TCT mobile X220D-2DRGIL1 Ok, so it really is an X220. There's a plugin for that already, the x22x plugin, which was added on Wed Sep 22 2010. I think what's going on here though is that the Longcheer plugin (which supports the Alcatel X030s and X060s) needs to be more careful about what modems it tries to grab. I'll fix that up in git, and then I'd expect your device to be handled by the x22x plugin. Dan You will see the Product and Vendor IDs in the log before and after modeswitch. I notice now that there are two lines in the log tagged with Longcheer which I think is another type of modem - it is listed in the plugins. At this point I am not sure it is being correctly detected but I can connect using wvdial on ttyUSB4. I will now try and get you the windows driver files. Thanks for help. Perazim May 3 16:58:23 travelmate kernel: [114816.488048] usb 2-1: new high speed USB device using ehci_hcd and address 17 May 3 16:58:24 travelmate kernel: [114816.604820] usb 2-1: New USB device found, idVendor=1bbb, idProduct=f000 May 3 16:58:24 travelmate kernel: [114816.604826] usb 2-1: New USB device strings: Mfr=3, Product=2, SerialNumber=4 May 3 16:58:24 travelmate kernel: [114816.604829] usb 2-1: Product: HSPA Data Card May 3 16:58:24 travelmate kernel: [114816.604831] usb 2-1: Manufacturer: USBModem May 3 16:58:24 travelmate kernel: [114816.604834] usb 2-1: SerialNumber: 1234567890ABCDEF May 3 16:58:24 travelmate kernel: [114816.633179] scsi13 : usb-storage 2-1:1.0 May 3 16:58:24 travelmate usb_modeswitch: switching 1bbb:f000 (USBModem: HSPA Data Card) May 3 16:58:24 travelmate kernel: [114817.531273] usb 2-1: USB disconnect, address 17 May 3 16:58:25 travelmate kernel: [114817.842053] usb 2-1: new high speed USB device using ehci_hcd and address 18 May 3 16:58:25 travelmate kernel: [114817.959065] usb 2-1: New USB device found, idVendor=1bbb, idProduct=0017 May 3 16:58:25 travelmate kernel: [114817.959070] usb 2-1: New USB device strings: Mfr=3, Product=2, SerialNumber=4 May 3 16:58:25 travelmate kernel: [114817.959073] usb 2-1: Product: HSPA Data Card May 3 16:58:25 travelmate kernel: [114817.959076] usb 2-1: Manufacturer: USBModem May 3 16:58:25 travelmate kernel: [114817.959078] usb 2-1: SerialNumber: 1234567890ABCDEF May 3 16:58:25 travelmate kernel: [114817.962067] option 2-1:1.0: GSM modem (1-port) converter detected May 3 16:58:25 travelmate kernel: [114817.962235] usb 2-1: GSM modem (1-port) converter now attached to ttyUSB0 May 3 16:58:25 travelmate kernel: [114817.962344] option 2-1:1.1: GSM modem (1-port) converter detected May 3 16:58:25 travelmate kernel: [114817.962456] usb 2-1: GSM modem (1-port) converter now attached to ttyUSB1 May 3 16:58:25 travelmate kernel: [114817.962562] option 2-1:1.2: GSM modem (1-port) converter detected May 3 16:58:25 travelmate kernel: [114817.962668] usb 2-1: GSM modem (1-port) converter now attached to ttyUSB2 May 3 16:58:25 travelmate kernel: [114817.962774] option 2-1:1.3: GSM modem (1-port) converter detected May 3 16:58:25 travelmate kernel: [114817.962873] usb 2-1: GSM modem (1-port) converter now attached to ttyUSB3 May 3 16:58:25 travelmate kernel: [114817.965974] scsi14 : usb-storage 2-1:1.4 May 3 16:58:25 travelmate kernel: [114817.966484] option 2-1:1.5: GSM modem (1-port) converter detected May 3 16:58:25 travelmate kernel: [114817.966693] usb 2-1: GSM modem (1-port) converter now attached to ttyUSB4 May 3 16:58:25 travelmate modem-manager: (ttyUSB1) opening serial device... May 3 16:58:25 travelmate modem-manager: (ttyUSB3) opening serial device... May 3 16:58:25 travelmate modem-manager: (ttyUSB2) opening serial device... May 3 16:58:25 travelmate modem-manager: (ttyUSB4) opening serial device... May 3 16:58:25 travelmate modem-manager: (ttyUSB0) opening serial device... May 3 16:58:26 travelmate usb_modeswitch: switched to 1bbb:0017 (USBModem: HSPA Data Card) May 3 16:58:26 travelmate kernel: [114818.984732] scsi 14:0:0:0: Direct-Access ALCATEL Mass Storage 2.31 PQ: 0 ANSI: 2 May 3 16:58:26 travelmate kernel: [114818.988811] sd 14:0:0:0: Attached scsi generic sg4 type 0 May 3 16:58:26 travelmate kernel: [114819.006841] sd 14:0:0:0: [sdd] Attached SCSI
Re: Support for ADSL modems
On Tue, 2011-05-03 at 19:09 +0300, Pantelis Koukousoulas wrote: Hi, I just wanted to announce that I 'm going to be working on adding support for ADSL modems in NetworkManager for a while, as a nice project to learn NM internals and Great! in order to finish what we started 2 years ago at Chania LUG's coding camp. It also makes NetworkManager more complete (and if support for old dialup modems is also added some time in the future by whatever means, NetworkManager can finally become Linux one stop shop for client-side network configuration, at least for me). Why ADSL modem support is useful? Users have requested it (e.g., http://brainstorm.ubuntu.com/item/3853/) and it is also among the items in NM TODO file. What are you going to work on? The initial effort will be for the only related device I already have (Sagem FAST 800, a USB ADSL modem) and PPPoA, I hope that PCI ADSL devices and PPPoE will be supported as well in the future. I know that there is already some PPPoE support, it just remains to be seen if it can be reused mostly as-is by using the nas0 interface one gets from br2684 as just another ordinary ethernet device. I personally have some random Zoom ueagle III based device, so I can at least help test the carrier stuff and device detection. Initially I 'm going to work in NetworkManager and nm-applet. KDE support can be written by someone else with my help, or I can write it myself as well in the future. Existing work I have found https://github.com/hicham-haouari/NetworkManager-ADSL-Support that has some initial libnm-util work (NMSettingAdsl), unfortunately only after I implemented this myself as well ( :/ ) so I will try to merge the 2 versions. Yeah, I talked extensively with hicham and made suggestions on what to do, so his work is mostly a result of our discussions and where i thought stuff should go. I also wrote up those thoughts on that page which you reference just below here. There is also http://live.gnome.org/NetworkManager/ADSL which I 'm using as reference wrt suggested class names etc. Is there any other documentation or preliminary implementation? Is anyone else working on this so that we can collaborate? I don't think hicham got as far as an NMDeviceAdsl class or anything, just the libnm-util/ configuration class bits. So feel free to run with it there. Code My initial experiments are implemented on top of whatever version of NetworkManager (0.8.x) is in ubuntu maverick since this made it possible to not care at all about build dependencies etc. Are current git tips of NetworkManager/nm-applet stable enough to do the actual development on, or are they in a turbulence state right now? They are actually pretty stable at this point, though the change in the D-Bus API may cause you some issues. But as code churn goes we're almost ready for a release of 0.9 so it's not going to change much. That said, I'd expect the work on either 0.8.x or 0.9 to port back and forth without much of a problem since those bits of the code haven't changed much. Are there instructions for how to install the git versions of NetworkManager/nm-applet in parallel with the system-provided ones (e.g., under /opt) so that one can experiment with them without breaking existing networking functionality? (I will try to do this myself anyway, just asking in case I can save the 15 minutes required :P) Not really, you can certainly install the binaries and libraries into a different prefix (configure with --prefix, --libdir, --localstatedir, as appropriate) but of course you cannot run old NM + new NM at the same time because they would try to manage the same devices, plus the bus name is the same. But it may work to just kill the old one and start the new one. Let me know how things go, great to hear you're working on this. Dan ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: preferring the DNS server of one interface over another
On Tue, 2011-05-03 at 12:59 -0400, Laine Stump wrote: I have a laptop (currently running Fedora 13, with NetworkManager-0.8.4-1.fc13.x86_64) that I sometimes use with a single ethernet, sometimes with that ethernet + a VPN connection, and sometimes plug in a 2nd ethernet onto a physical private network that provides the same connectivity as the VPN. When the 2nd ethernet is plugged in, sometimes its configured DNS servers are listed first in /etc/resolv.conf, and sometimes the others are (and it can periodically change, I'm assuming this happens when the DHCP lease is refreshed on one interface or the other). To avoid disruptions in service, I need to be able to force the DNS server on the 2nd ethernet to always takes precedence over the DNS server on the 1st ethernet (ie, that it be listed first in resolv.conf), but I haven't found a way to specify that. Am I missing something, or does this functionality not exist? I don't think this functionality exists yet, as there isn't a way to say that a specific interface or connection of the same device type as another is always preferred over another yet. Not sure what the best way to handle this yet is... suggestions welcome. One that I know would come up are priorities with the device's class so that you could do exactly this and prefer one ethernet connection over another ethernet connection. Dan ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: Question
On Tue, 2011-05-03 at 19:57 -0300, Lucas Mura wrote: Hi people, I have a network that I want to connect to. It is WEP encrypted. I am not sure but the key is FF:AA:DD:BB:00. How should I type in this key in the dialog box? wait for dialog appears , and use keyboard Thank you all! ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list -- Sérgio M. B. smime.p7s Description: S/MIME cryptographic signature ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: Question
On Wed, 2011-05-04 at 21:22 +0100, Sergio Monteiro Basto wrote: On Tue, 2011-05-03 at 19:57 -0300, Lucas Mura wrote: Hi people, I have a network that I want to connect to. It is WEP encrypted. I am not sure but the key is FF:AA:DD:BB:00. How should I type in this key in the dialog box? wait for dialog appears , and use keyboard Note that the key should be entered as FFAADDBB00 since : is not a valid character for WEP keys (and never has been, I assume its there for readability). Otherwise it looks like a standard WEP-40/64 10-character hex key. Dan Thank you all! ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: Support for ADSL modems
On Wed, May 4, 2011 at 11:18 PM, Dan Williams d...@redhat.com wrote: I personally have some random Zoom ueagle III based device, so I can at least help test the carrier stuff and device detection. Cool, thanks :) For now I have a problem in the device detection path, it seems NetworkManager needs an interface index otherwise it refuses to enable the device. I can't find an interface index number for my ueagle-atm0 interface so for now I have a hack that returns a fake one, but this will need to be fixed in a better way. They are actually pretty stable at this point, though the change in the D-Bus API may cause you some issues. But as code churn goes we're almost ready for a release of 0.9 so it's not going to change much. That said, I'd expect the work on either 0.8.x or 0.9 to port back and forth without much of a problem since those bits of the code haven't changed much. Great, this has been my assessment as well :) Are there instructions for how to install the git versions of NetworkManager/nm-applet in parallel with the system-provided ones (e.g., under /opt) so that one can experiment with them without breaking existing networking functionality? Not really, you can certainly install the binaries and libraries into a different prefix (configure with --prefix, --libdir, --localstatedir, as appropriate) but of course you cannot run old NM + new NM at the same time because they would try to manage the same devices, plus the bus name is the same. But it may work to just kill the old one and start the new one. Ok, I made a blog post on the procedure I used myself under ubuntu (it may look a little complex with the dpkg-diverts and all but this way I can develop on my regular laptop, the security updates can modify any of the files in the system-version of NM without problem and I can script switching between the 2 versions of NM and the applet without problem. http://polytechnitis.blogspot.com/2011/05/compiling-latest-versions-of.html Let me know how things go, great to hear you're working on this. Great, I will try to push code somewhere soonish. Cheers, Pantelis ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
SetLogging on a GSM Modem
How can I set SetLogging(DEBUG,PPP) for a gsm device? I understand the command but I can't find the method in the device. Also is there a way to read the NM_DEVICE_STATE_REASON apart from using a signal? Would anyone have a PERL/Python script example? Below are the methods, properties I can see for the device. $ qdbus --system org.freedesktop.NetworkManagerSystemSettings /org/freedesktop/NetworkManager/Devices/3 method QString org.freedesktop.DBus.Introspectable.Introspect() method QDBusVariant org.freedesktop.DBus.Properties.Get(QString interface, QString propname) method void org.freedesktop.DBus.Properties.Set(QString interface, QString propname, QDBusVariant value) signal void org.freedesktop.NetworkManager.Device.Serial.PppStats(uint, uint) property read uint org.freedesktop.NetworkManager.Device.Capabilities property read uint org.freedesktop.NetworkManager.Device.DeviceType property readwrite QDBusObjectPath org.freedesktop.NetworkManager.Device.Dhcp4Config property read QString org.freedesktop.NetworkManager.Device.Driver property readwrite QString org.freedesktop.NetworkManager.Device.Interface property readwrite uint org.freedesktop.NetworkManager.Device.Ip4Address property readwrite QDBusObjectPath org.freedesktop.NetworkManager.Device.Ip4Config property readwrite QDBusObjectPath org.freedesktop.NetworkManager.Device.Ip6Config property read bool org.freedesktop.NetworkManager.Device.Managed property read uint org.freedesktop.NetworkManager.Device.State property read QString org.freedesktop.NetworkManager.Device.Udi method void org.freedesktop.NetworkManager.Device.Disconnect() signal void org.freedesktop.NetworkManager.Device.StateChanged(uint, uint, uint) Version is 0.8 running on Ubuntu ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: Support for ADSL modems
On Wed, 2011-05-04 at 23:39 +0300, Pantelis Koukousoulas wrote: On Wed, May 4, 2011 at 11:18 PM, Dan Williams d...@redhat.com wrote: I personally have some random Zoom ueagle III based device, so I can at least help test the carrier stuff and device detection. Cool, thanks :) For now I have a problem in the device detection path, it seems NetworkManager needs an interface index otherwise it refuses to enable the device. I can't find an interface index number for my ueagle-atm0 interface so for now I have a hack that returns a fake one, but this will need to be fixed in a better way. Hmm, it shouldn't since modems don't have them either, nor do bluetooth devices. I'm pretty sure you shouldn't need them. What's the code that you're trying to use that wants one? We might need a slightly different approach. They are actually pretty stable at this point, though the change in the D-Bus API may cause you some issues. But as code churn goes we're almost ready for a release of 0.9 so it's not going to change much. That said, I'd expect the work on either 0.8.x or 0.9 to port back and forth without much of a problem since those bits of the code haven't changed much. Great, this has been my assessment as well :) Are there instructions for how to install the git versions of NetworkManager/nm-applet in parallel with the system-provided ones (e.g., under /opt) so that one can experiment with them without breaking existing networking functionality? Not really, you can certainly install the binaries and libraries into a different prefix (configure with --prefix, --libdir, --localstatedir, as appropriate) but of course you cannot run old NM + new NM at the same time because they would try to manage the same devices, plus the bus name is the same. But it may work to just kill the old one and start the new one. Ok, I made a blog post on the procedure I used myself under ubuntu (it may look a little complex with the dpkg-diverts and all but this way I can develop on my regular laptop, the security updates can modify any of the files in the system-version of NM without problem and I can script switching between the 2 versions of NM and the applet without problem. http://polytechnitis.blogspot.com/2011/05/compiling-latest-versions-of.html Let me know how things go, great to hear you're working on this. Great, I will try to push code somewhere soonish. I'm happy to review and suggest, thanks! Dan ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: SetLogging on a GSM Modem
On Wed, 2011-05-04 at 22:10 +0100, Andy Maginnis wrote: How can I set SetLogging(DEBUG,PPP) for a gsm device? I understand the command but I can't find the method in the device. It should be a global method for all of NM, so you'd use the org.freedesktop.NetworkManager interface with the /org/freedesktop/NetworkManager object path. LIke so with dbus-send: sudo dbus-send --system --print-reply --dest=org.freedesktop.NetworkManager /org/freedesktop/NetworkMager org.freedesktop.NetworkManager.SetLogging string:DEBUG string: will cause logging to be set to DEBUG for all current logging domains. Dan Also is there a way to read the NM_DEVICE_STATE_REASON apart from using a signal? Would anyone have a PERL/Python script example? Below are the methods, properties I can see for the device. $ qdbus --system org.freedesktop.NetworkManagerSystemSettings /org/freedesktop/NetworkManager/Devices/3 method QString org.freedesktop.DBus.Introspectable.Introspect() method QDBusVariant org.freedesktop.DBus.Properties.Get(QString interface, QString propname) method void org.freedesktop.DBus.Properties.Set(QString interface, QString propname, QDBusVariant value) signal void org.freedesktop.NetworkManager.Device.Serial.PppStats(uint, uint) property read uint org.freedesktop.NetworkManager.Device.Capabilities property read uint org.freedesktop.NetworkManager.Device.DeviceType property readwrite QDBusObjectPath org.freedesktop.NetworkManager.Device.Dhcp4Config property read QString org.freedesktop.NetworkManager.Device.Driver property readwrite QString org.freedesktop.NetworkManager.Device.Interface property readwrite uint org.freedesktop.NetworkManager.Device.Ip4Address property readwrite QDBusObjectPath org.freedesktop.NetworkManager.Device.Ip4Config property readwrite QDBusObjectPath org.freedesktop.NetworkManager.Device.Ip6Config property read bool org.freedesktop.NetworkManager.Device.Managed property read uint org.freedesktop.NetworkManager.Device.State property read QString org.freedesktop.NetworkManager.Device.Udi method void org.freedesktop.NetworkManager.Device.Disconnect() signal void org.freedesktop.NetworkManager.Device.StateChanged(uint, uint, uint) Version is 0.8 running on Ubuntu ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
How to debug 3G USB modem problems?
I understand that this a upstream list and most readers are interested in the bleeding edge. However, I have to support some systems for naive users and I prefer to use a plain vanilla distribution as long as possible. So please bear with my question regarding versions straight from the museum... On Ubuntu 10.04 (long time support) they have nm version 0.8-0ubuntu3 The USB modem in use is a Nokia CS-17. (I know that if you search for trouble, you select a 3G USB modem. Unfortunately this variant is really by orders of magnitude cheaper in this case than anything more reliable.) The good news is that the CS-17 works even with that old software. The bad news is that it's completely unreliable. Some 30%-50% of the connection attempts just fail. Unfortunately the failures seem to come in waves. If it fails the first time, it will also fail the 2nd, 3rd etc. Removing the modem does not help. But 1 or 2 hours later it might just work again. Alternatively if I want to repeat the problem, I can make 15 connections in a row without any failure. The other problem is that while the connection typically stays open for hours during normal web browsing, it will disconnect with quite high probability when downloading bigger files. Today I needed 4 attempts to download a 20 MB file. I would exclude network-side problems or weak signal, because I use 3G data on the same network nearly all day long on my smartphone without such problems. Additionlly, when the USB modem fails to connect I can use my phone as the modem using the data cable and everything works. So how could I debug this? Is it a Linux-side problem or a firmware problem in the CS-17? (Haven't had a chance to test it with M$ yet) From searching in the mailing list archives I found 2 debugging hints: 1. starting nm from command line as # NM_SERIAL_DEBUG=1 NM_PPP_DEBUG=1 NetworkManager --no-daemon | tee log.txt That works nicely. When the connection attempt fails I see how PPP negotiation starts, but it doesn't seem to get any reply. sent [LCP ConfReq id=0x1 asyncmap 0x0 magic 0x5c8b3860 pcomp accomp] After 5 repetitions pppd seems to give up and nm complains that the connection failed. (Not sure whether NM_SERIAL_DEBUG has any effect, I don't think I see that variable in the source code of my version) The problem is that using the command line (and root) is OK for me on my test machine, but not for the user in question on the real target machine. So I'd like to configure the real service such that logging is always on. I tried to do so by adding the 2 environment variables to /etc/init/network-manager.conf (Ubuntu uses upstart) env NM_PPP_DEBUG=1 env NM_SERIAL_DEBUG=1 exec NetworkManager Unfortunately that seems to have no effect to the pppd logging. The pppd stuff that I can see when running with --no-daemon does just not appear in syslog, neither in successful nor in failing cases. (I have checked from /proc/nnn/environ that the environment variables really end up in the network-manager process) Also the debug parameter appears on the pppd command line as shown in syslog. How can I get pppd logging when running as a daemon? 2. The other hint I found in the mailing archive was a config file entry [logging] level=WARN I changed that to level=DEBUG, but I don't think it made a difference. Could not find such parameter in my source code. Has it possibly been added only in a later version? Yesterday I read in some forum that killing modem-manager after a failed connection attempt helps. Today I had only very few failed attempts in my testing and killing modem-manager helped each time. Not yet sure whether this is really a reliable work-around. If yes, I could of course script it. Does the problem description ring any bells? If you remember specific fixes that solve these issues, I might try to backport them to my version. Or just install a newer version manually. Regards, Uwe P.S. Yes, I am aware of the modeswitching. I left that out from the description above, I'm sure that the modem was always in the right mode. ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: How to debug 3G USB modem problems?
On Thu, 2011-05-05 at 01:15 +0300, Uwe Geuder wrote: I understand that this a upstream list and most readers are interested in the bleeding edge. However, I have to support some systems for naive users and I prefer to use a plain vanilla distribution as long as possible. So please bear with my question regarding versions straight from the museum... On Ubuntu 10.04 (long time support) they have nm version 0.8-0ubuntu3 So what you want to do here is also grab modem-manager logs as described here: http://live.gnome.org/NetworkManager/Debugging under the Debugging NetworkManager 0.8.x 3G connections section. The USB modem in use is a Nokia CS-17. (I know that if you search for trouble, you select a 3G USB modem. Unfortunately this variant is really by orders of magnitude cheaper in this case than anything more reliable.) Huh, because they are $120 on eBay which isn't cheap at all :) Otherwise I'd buy one. The good news is that the CS-17 works even with that old software. The bad news is that it's completely unreliable. Some 30%-50% of the connection attempts just fail. Unfortunately the failures seem to come in waves. If it fails the first time, it will also fail the 2nd, 3rd etc. Removing the modem does not help. But 1 or 2 hours later it might just work again. Alternatively if I want to repeat the problem, I can make 15 connections in a row without any failure. So in this case, more logs from NetworkManager's PPP debugging would help, as would the ModemManager logs as described above. The other problem is that while the connection typically stays open for hours during normal web browsing, it will disconnect with quite high probability when downloading bigger files. Today I needed 4 attempts to download a 20 MB file. What does the PPP debugging say when this happens? NM wont' terminate the connection unless PPP says it's down. Dan I would exclude network-side problems or weak signal, because I use 3G data on the same network nearly all day long on my smartphone without such problems. Additionlly, when the USB modem fails to connect I can use my phone as the modem using the data cable and everything works. So how could I debug this? Is it a Linux-side problem or a firmware problem in the CS-17? (Haven't had a chance to test it with M$ yet) From searching in the mailing list archives I found 2 debugging hints: 1. starting nm from command line as # NM_SERIAL_DEBUG=1 NM_PPP_DEBUG=1 NetworkManager --no-daemon | tee log.txt That works nicely. When the connection attempt fails I see how PPP negotiation starts, but it doesn't seem to get any reply. sent [LCP ConfReq id=0x1 asyncmap 0x0 magic 0x5c8b3860 pcomp accomp] After 5 repetitions pppd seems to give up and nm complains that the connection failed. (Not sure whether NM_SERIAL_DEBUG has any effect, I don't think I see that variable in the source code of my version) The problem is that using the command line (and root) is OK for me on my test machine, but not for the user in question on the real target machine. So I'd like to configure the real service such that logging is always on. I tried to do so by adding the 2 environment variables to /etc/init/network-manager.conf (Ubuntu uses upstart) env NM_PPP_DEBUG=1 env NM_SERIAL_DEBUG=1 exec NetworkManager Unfortunately that seems to have no effect to the pppd logging. The pppd stuff that I can see when running with --no-daemon does just not appear in syslog, neither in successful nor in failing cases. (I have checked from /proc/nnn/environ that the environment variables really end up in the network-manager process) Also the debug parameter appears on the pppd command line as shown in syslog. How can I get pppd logging when running as a daemon? 2. The other hint I found in the mailing archive was a config file entry [logging] level=WARN I changed that to level=DEBUG, but I don't think it made a difference. Could not find such parameter in my source code. Has it possibly been added only in a later version? Yesterday I read in some forum that killing modem-manager after a failed connection attempt helps. Today I had only very few failed attempts in my testing and killing modem-manager helped each time. Not yet sure whether this is really a reliable work-around. If yes, I could of course script it. Does the problem description ring any bells? If you remember specific fixes that solve these issues, I might try to backport them to my version. Or just install a newer version manually. Regards, Uwe P.S. Yes, I am aware of the modeswitching. I left that out from the description above, I'm sure that the modem was always in the right mode. ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list ___ networkmanager-list mailing list
ANN: Release of NetworkManager 0.8.999 (0.9-rc2)
Hi, I've tagged and uploaded 0.8.999 which has quite a few fixes from the last release, including: - IPv6 compliance and RDNSS lifetime fixes - systemd cooperation fixes - API fixes for clients creating wifi connections - API additions to make client code simpler - startup efficiency fixes by not parsing the ConsoleKit database more than required - Fixes for IBM s390 CTC-type network devices - Fixes for WWAN enable/disable status - Fixes for ifcfg-rh configuration plugin handling of IP addresses - Support for Easytether Android handset tethering - Better handling of rfkill for WiFi and WiMAX interfaces - Addition of IPv6 support for dispatcher scripts - Fixed handling of DER-format certificates - Give modems more time to unlock themselves after PIN entry - Fix creation of 802.1x-enabled connections in the applet - Don't crash gnome-bluetooth on some failures when creating new DUN connections - Much better handling of IP addresses in the editor Thanks to everyone who contributed to this release: Jiří Klimeš, Lennart Poettering, Karsten Hopp, Torsten Spindler, Eckhart Wörner, and anyone I may have forgotten. Downloads in the usual places; new to the list is a plugin for OpenSWAN that's existed in Fedora and RHEL for a while, but we figured it should get upstream so everyone could try it out. That in particular needs testing. http://ftp.gnome.org/pub/GNOME/sources/NetworkManager/0.8/NetworkManager-0.8.1.999.tar.bz2 http://ftp.gnome.org/pub/GNOME/sources/network-manager-applet/0.8/network-manager-applet-0.8.1.999.tar.bz2 http://ftp.gnome.org/pub/GNOME/sources/NetworkManager-vpnc/0.8/NetworkManager-vpnc-0.8.1.999.tar.bz2 http://ftp.gnome.org/pub/GNOME/sources/NetworkManager-pptp/0.8/NetworkManager-pptp-0.8.1.999.tar.bz2 http://ftp.gnome.org/pub/GNOME/sources/NetworkManager-openvpn/0.8/NetworkManager-openvpn-0.8.1.999.tar.bz2 http://ftp.gnome.org/pub/GNOME/sources/NetworkManager-openconnect/0.8/NetworkManager-openconnect-0.8.1.999.tar.bz2 http://ftp.gnome.org/pub/GNOME/sources/NetworkManager-openswan/0.8/NetworkManager-openswan-0.8.999.tar.bz2 Cheers, Dan ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list
Re: How to debug 3G USB modem problems?
Thanks Dan for your quick reply. It's too late here for taking more logs so let me just reply quickly on your other questions. Huh, because they are $120 on eBay which isn't cheap at all :) Otherwise I'd buy one. This operator http://saunalahti.fi/ gives you one for free even you just shop for a cheap fixed-line ADSL. Sounds unlikely that they pay very much for them, because their prices are on the cheaper side compared to the competitors even if you don't count the 3G modem at all. Might not be an option for you though :( Maybe they get cheaper on ebay.fi when many people end up with one without intending to use it. What does the PPP debugging say when this happens? NM wont' terminate the connection unless PPP says it's down. I would have included the log if I had it :( I have seen it when using command line, but I did not store it. My point and main question was how to get the logging to work when network-manager runs as a service. Because I can only test occasionally, the real user is not familiar with the command line and she has no root access. Hmm, actually I just find a copy on another machine. (And weird enough the pppd logging I was complaining about *IS* in that syslog) May 4 14:11:39 geuder-u10 pppd[1783]: Using interface ppp0 May 4 14:11:39 geuder-u10 pppd[1783]: Connect: ppp0 -- /dev/ttyACM1 May 4 14:11:39 geuder-u10 pppd[1783]: sent [LCP ConfReq id=0x1 asyncmap 0x0 magic 0x5a37c2dc pcomp accomp] May 4 14:11:59 geuder-u10 pppd[1783]: last message repeated 6 times May 4 14:11:59 geuder-u10 NetworkManager: WARN pppd_timed_out(): Looks like pppd didn't initialize our dbus module May 4 14:11:59 geuder-u10 NetworkManager: info (ttyACM1): device state change: 7 - 9 (reason 14) May 4 14:11:59 geuder-u10 NetworkManager: info Marking connection 'Saunalahti Postpaid (contract) 1' invalid. May 4 14:11:59 geuder-u10 pppd[1783]: Terminating on signal 15 May 4 14:11:59 geuder-u10 pppd[1783]: sent [LCP TermReq id=0x2 User request] To me it looks that pppd does not get or not recognize any responses (rcvd lines) although it sends its first message uplink 7 times altogether. So NetworkManager decides to kill it after 20 seconds. I'll try to capture full logs including modem-manager later. Regards, Uwe ___ networkmanager-list mailing list networkmanager-list@gnome.org http://mail.gnome.org/mailman/listinfo/networkmanager-list