Huawei E220 and E1550 can't connect on Ubuntu 11.10/12.04

[Akila Wajirasena]

[SRU JUSTIFICATION]
This changes reverts the rework of probing methods for Huawei devices which broke proper support for the Huawei E220, E1550 and some other specific models of Huawei 3G modems. It's low risk as it is known to work properly with most Huawei devices already; and only affects users of such devices. The E220, among others, remains one of the most common 3G modems available.

[REGRESSION POTENTIAL]
Low; this patch changes probing for Huawei devices only, and has been tested successfully with good support for the affected devices after the patch is installed, as well as unchanged behavior for many other devices (tested by cyphermox with a E1831 too).

[TESTING]
Requires Huawei 3G modem; needs to be the E220 or E1550 model for successfully testing the resolution of this issue. Testing with other Huawei devices is useful as control. Devices from other manufacturers are unaffected.

=======

I'm using Ubuntu in my HP Pavilion dv6745tx laptop. I cannot connect the Huawei E220 usb modem since 11.10 beta1 (now I'm in beta2). There is no problem in natty.

I have found this in syslog

Oct 5 07:10:51 ubuntu NetworkManager[1681]: <info> Activation (ttyUSB0) starting connection 'Dialog GSM Postpaid'
Oct 5 07:10:51 ubuntu NetworkManager[1681]: <info> (ttyUSB0): device state change: disconnected -> prepare (reason 'none') [30 40 0]
Oct 5 07:10:51 ubuntu NetworkManager[1681]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) scheduled...
Oct 5 07:10:51 ubuntu NetworkManager[1681]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) started...
Oct 5 07:10:51 ubuntu NetworkManager[1681]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) complete.
Oct 5 07:10:58 ubuntu NetworkManager[1681]: <warn> GSM connection failed: (32) Sending command failed: 'Resource temporarily unavailable'
Oct 5 07:10:58 ubuntu NetworkManager[1681]: <info> (ttyUSB0): device state change: prepare -> failed (reason 'unknown') [40 120 1]
Oct 5 07:10:58 ubuntu NetworkManager[1681]: <warn> Activation (ttyUSB0) failed.
Oct 5 07:10:58 ubuntu NetworkManager[1681]: <info> (ttyUSB0): device state change: failed -> disconnected (reason 'none') [120 30 0]
Oct 5 07:10:58 ubuntu NetworkManager[1681]: <info> (ttyUSB0): deactivating device (reason: 0).
Oct 5 07:10:58 ubuntu NetworkManager[1681]: nm_system_iface_flush_routes: assertion `ifindex > 0' failed
Oct 5 07:10:58 ubuntu NetworkManager[1681]: nm_system_iface_flush_addresses: assertion `ifindex > 0' failed

[Akila Wajirasena (akila-wajirasena)]

I have attached the network-manager and modem- manager logs

[girts (girtsz)]

I have the same problem.

I found such solution in the Internet. Please confirm if this work!!!!!

the recommend change should be applied to "usb_modeswitch.conf" ( if any) . try edit the

/lib/udev/rules.d/40-usb_modeswitch.rules

to disable the rule place a # at start of the line , example

Quote:
# Huawei E220, E230, E270, E870
# ATTRS{idVendor}=="12d1", ATTRS{idProduct}=="1003", RUN+="usb_modeswitch '%b/%k'"

this should allow the use of modems which require switching.

[Mathieu Trudel-Lapierre (cyphermox)]

Akila,

Please try the above-mentioned fix; let me know if this works for you :)

[Akila Wajirasena (akila-wajirasena)]

I tried that. But it doesn't work. The modem is not detected when I comment that lines.
girts,
Can you tell me the exact steps you did? and whether it works?

BTW there is thread on this issue in ubuntu forums.
http://ubuntuforums.org/showthread.php?t=1857506

[Chamila Mihiripenna (chamila-mihiripenna)]

I'm also facing the same issue. Huawei E220 worked in 11.04 after upgrade not working at all. Tried with Huawei E153 was able to connect Internet. So its E220 incompatibility issue. Appreciate if E220 can be fixed since I dont onw the E153. Ubuntu 11.10 is not usable until this is fixed and this force me to use window7.

[Chamila Mihiripenna (chamila-mihiripenna)]

Tried girts (girtsz-gmail) 's fix in 40-usb_modeswitch.rules . Not working ... Issue still exists.

[girts (girtsz)]

Everyone can try Mobile Partner for Linux until bug will be fixed. You must configure Mobile Partner for Linux for your ISP!!
http://www.lmt.lv/files/modemu_atjauninajumi/E620_LMT_internet_Linux.zip

[Chamila Mihiripenna (chamila-mihiripenna)]

Mobile partner for linux works with Huawei E220 on ubuntu 11.10

[Tomas (omas)]

I'm not sure if something was done about this issue but from time to time I get my E220 working (M1330, Ubuntu 11.10 32bit), however usually I have to create a new connection to be able to do that. I didn't have this problem in a previous version.

[ggonzar (ggonzar)]

This is my case:

- I use Edubuntu 11.10 amd64
- when I plug the modem se system Load grows up (plus 4,0) and the laptop work in "flashes"...
- the modem work sometimes. When work the system load is OK.
- I try to comment the line of /lib/udev/rules.d/40-usb_modeswitch.rules and nothing happend.

It is like the system go in a loop when detecting de modem and this causes de system load grows up. But I can't tell how sometimes works.

[Sleepy John (john-eyre)]

I am experiencing similar problems after upgrading my Acer Aspire One to 11.10 from 11.04. Once connected, my E220 will disconnect and reconnect quite reliably provided it continues to be powered on via its USB connection. Howver if I unplug it or depower it by swirching off, I cannot get it to reconnect on 11.10, and I have to go either to Natty or to Windows to get it to work again. From there, a direct restart to 11.10 will deliver it it back in a condition from which it will connect again. More details on ubintuforums.org/showthread.php?t=1868014

[Sleepy John (john-eyre)]

For what it's worth, I've found this inelegant workaround will get me on-line again after my E220 has been depowered:

CODE
sudo usb_modeswitch -v 0x12d1 -p 1003 -V 0x12d1 -P 1003 -R

Then unplug and replug the dongle, and now it will connect.

An lsusb check shows that the Device # for the E220 has been increased by 2

[Roman Stingler (coolx67)]

@Sleepy John

Thank you very much, this works perfect.
first
Bus 005 Device 003: ID 12d1:1003 Huawei Technologies Co., Ltd. E220 HSDPA Modem
doesn't work
after
Bus 005 Device 005: ID 12d1:1003 Huawei Technologies Co., Ltd. E220 HSDPA Modem
works perfect.

you are great.

[Per Ronny Westin (vaxxipooh)]

@Sleepy John

Well that:
--------------------------------------------------------------------------------------
CODE
sudo usb_modeswitch -v 0x12d1 -p 1003 -V 0x12d1 -P 1003 -R
-------------------------------------------------------------------------------------

looks nice and make sense - how do we implement it to be sortof an steady fix ? Is it possible to put the line: "usb_modeswitch -v 0x12d1 -p 1003 -V 0x12d1 -P 1003 -R" in /etc/rc.local" or something ??

Anyway thanx for your great "Sherlock Holmes" work in this issue :-)

[Sleepy John (john-eyre)]

Thanks for two votes of confidence above, but seriously I don't see this code as anything more than a temporary workaround until a properly designed fix is found. In my setup, it only worked after a manual USB unplug/replug, and I'm sure there must be a better way of simulating this automatcally. Basically I would have thought looking into the differences in the way Ocelot handles an E220 compared to how Natty does it would be the best way forward, since it all works fine in Natty.

[Fredrik Hultkrantz (fjutt)]

I can also confirm that the workaround posted by Sleepy John fixes the problem for me. Thank you! Not the final solution perhaps, but at least I can use my 3g modem again from Ubuntu, which is a good thing.

[Damjan Jovanovic (damjan-jov)]

Sleepy John you're the man! This is what Ubuntu needs more of: working hardware.

Later on I'm going to dig up the workarounds for the old regression and see if they also work now. I see that udisks is polling /dev/sr0 - this did play a role before.

[Indika Kodagoda (indikaimk)]

@Sleepy John
Your solution worked for me too.

[Mikhail Solovyev (mikhail-solovyev)]

@Sleepy John
This didn't work for me :(

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Sleepy John (john-eyre)]

Sorry, I'm not so familiar with the methods used to determine status and importance here, and am curious to know what factors lie behind the decision to classify this bug with importance "Low". From my own personal standpoint as someone who relies on an E220 for my regular daily internet access, I see this bug as a high nuisance value, and the inelegant workaround I've posted above as nothing more than a very short-term temporary alternative to being forced to regress to Natty where my E220 logged in without any difficulty. It would be interesting to hear from other E220 users about how important and how seriously this bug affects them.

[Fredrik Hultkrantz (fjutt)]

I can only agree with Sleepy John, for me the 3g modem is my the only way to get internet access for me and that it doesn't work, or forces me to do the kind of workarounds posted above is not for me a problem with low importance. Using my comp with no internet access is not very useful. So it feels like it could be upgraded :)

[girts (girtsz)]

Bug is super important!!!!
It must be fixed. In such times when there is critical bugs need after fixing new iso images!!!
I am asking others for new iso images after bug will be fixed.
Canonical must have one distribution Ubuntu with one repository which is contributed from all people which creates packages for Ubuntu. KDE, WindowMaker, XFCE, KDE... must be in repositories.
By using such strategy will be faster to create and test bugs and solve them. No need Kubuntu, Lubuntu, Xubuntu... Eastest way how to install Ubuntu and choose what to use!

[Per Ronny Westin (vaxxipooh)]

I agree to the full with girts (girtsz-gmail) - it must be fixed . . . i have no other connection than USB Modem too (Huawei E220) - by the way i have seen people have issues with other models of Huawei GSM modems too . . .

[Sleepy John (john-eyre)]

Per Ronny - It might be helpful if you could do some checking and post on here the model numbers of any other GSM modems that are giving other people trouble on Ocelot. This could help to clarify how widespread the problem generally is, and aid developers in pinpointing the cause. If we can show that the problem is not only confined to E220 users, it will make it all the more urgent to try and get it fixed.

[Akila Wajirasena (akila-wajirasena)]

One of my friend said adding "blacklist usb_storage" to /etc/modprobe.d/blacklist.conf solved the issue

Can any one confirm it? I think several other Huawei modems also doesn't work.

There is also another method of fixing this using old network manager packages

http://www.technewsdaddy.com/2011/ubuntu-11-10-and-huawei-e220/

[Sleepy John (john-eyre)]

Yes!! Adding that "blacklist usb_storage" to /etc/modprobe.d/blacklist.conf has worked in my case. So long as it doesn't upset any other USB iunctions, that looks like a good fix to me! Thanks Akila.

[Sleepy John (john-eyre)]

I might have spoken too soon because I can't get that addition to blacklist.conf fix, which seemed to be working perfectly for me yesterday, to work at all today. Still checking, though. Is anyone else seeing it either working or failing?

[Shrijit Nair (shrijith1)]

I too can't get it to work with the blacklist.conf fix. Please note, I have a ZTE modem. Is this bug specific to Huawei modems ?

[Sleepy John (john-eyre)]

Shrijit - This bug thread was opened specifically to talk about about the problems that have appeared with the Huawei E220 on Ubuntu 11.10 that were not there on Ubuntu 11.04. It does appear from what some posters here say, that 11.10 is also causing the same or similar bugs to show up on other modems. The same fixes don't always apply across different modems though, and it's not safe to assume the fixes are universal. Some work, some don't, but it's worth giving them a try.

I still remain puzzled about that blacklist.conf fix because I'm pretty sure the tests I did with it yesterday proved that it was working OK, and yet today it's not working at all. Still interested to hear the experience of others on it.

[quiricada (raraland)]

same problem here, network-manager tries to connect and just hangs. everything was working fine in 11.04.

so am back to using terminal and wvdial; openvpn in terminal also when i need it.

[Josua Dietze (digidietze)]

To determine if the problem is caused by usb_modeswitch, please enable logging in
/etc/usb_modeswitch.conf.

Post the largest file (name starting with "usb_modeswitch") that is created in /var/log after you plugged the device.

DON'T comment out any entry in the "rules"file !!

[Josua Dietze (digidietze)]

Furthermore I don't recommend to blacklist the "usb-storage" module. Several modems want to be initialized before the mode switch can successfully happen. The storage driver should not be able to keep any device from switching.

Again, please enable usb_modeswitch's logging.

[Mark Robbie (mark-robbie)]

I too have this problem with a Huawei E220 USB modem. I had no problems with Ubuntu versions 9.10 thru 11.04 but am seeing it now that I have installed 11.10. The problem also occurs if booting from the 11.10 ISO disk, so it is not due to a corrupt installation. I can use the sakis3g script to make a connection so there is no problem with the modem or ISP.

I have enabled logging in /etc/usb_modeswitch.conf which produces a single log file (/var/log/usb_modeswitch.log - attached) that is written at the time that I plug in the USB modem.

What looks interesting to me is that there is a list of USB devices in this log, but the modem is not listed (there is only one connected device which lsusb reveals to be the built-in webcam) and therefore usb_modeswitch concludes that there is nothing to do.

So despite the fact that the running of usb_modeswitch appears to get triggered by my plugging in the modem, the modem does not appear in the list of USB devices. Could it be that it takes some small time for the device to initialise and that usb_modeswitch is running before this has happened?

There are two other bugs/poor behaviour related to this issue but they don't bother me sufficiently to raise separate bug reports:
1. usb_modeswitch does not appear to run (there is no log file) if the modem is plugged-in before booting
2. If I bring up the link using some other method (e.g. using the sakis3g script) network-manager does not realise that a connection has been made so it grays-out the menu options to make my VPN connections unavailable.

Mark.

[Josua Dietze (digidietze)]

If udev starts usb_modeswitch, but usb_modeswitch can't find the device when searching the busses, it means that something else is affecting the device in the meantime. This should be visible when calling "dmesg" after plugging in.

Either it's a driver thing (I'll check the source code of usb-storage in a moment) or there is annother udev rule or hotplug mechanism kicking in. I'll have a look in my 11.10 installation.

Regarding the other problems you mention, no. 1 is a known problem fixed in version 1.2.0; about no. 2 I have no idea ...

[Josua Dietze (digidietze)]

It's the kernel.

I have forgotten about it - there is a switching function in usb-storage from the time when usb_modeswitch wasn't wide-spread. The affected devices are listed in "unusual_devs.h".

The bottom line is that usb_modeswitch can't be responsible for the reported bug. It is not needed for this device, but does not interfere if present.

I therefore remove usb-modeswitch-data from the list of affected packages.

[Roosweltcr (roosweltcr)]

Thanks for Sleepy John (john-eyre) . it works fine with the command but time to time it will give headache, is there is a permanent solution for this??

[Brad Figg (brad-figg)]

This bug is missing log files that will aid in diagnosing the problem. From a terminal window please run:

apport-collect 868034

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Sleepy John (john-eyre)]

AlsaVersion: Advanced Linux Sound Architecture Driver Version 1.0.24.
ApportVersion: 1.23-0ubuntu4
Architecture: i386
ArecordDevices:
 **** List of CAPTURE Hardware Devices ****
 card 1: SB [HDA ATI SB], device 0: CONEXANT Analog [CONEXANT Analog]
   Subdevices: 1/1
   Subdevice #0: subdevice #0
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC1: john 1557 F.... pulseaudio
CRDA: Error: [Errno 2] No such file or directory
Card0.Amixer.info:
 Card hw:0 'Generic'/'HD-Audio Generic at 0x90544000 irq 40'
   Mixer name : 'ATI R6xx HDMI'
   Components : 'HDA:1002aa01,00aa0100,00100200'
   Controls : 4
   Simple ctrls : 1
Card0.Amixer.values:
 Simple mixer control 'IEC958',0
   Capabilities: pswitch pswitch-joined penum
   Playback channels: Mono
   Mono: Playback [on]
Card1.Amixer.info:
 Card hw:1 'SB'/'HDA ATI SB at 0x90540000 irq 16'
   Mixer name : 'Conexant CX20584'
   Components : 'HDA:14f15068,10250543,00100302'
   Controls : 16
   Simple ctrls : 10
DistroRelease: Ubuntu 11.10
HibernationDevice: RESUME=UUID=6ec19fda-71f6-4532-8d87-e62648372fcf
IfupdownConfig:
 auto lo
 iface lo inet loopback
InstallationMedia: Ubuntu 11.10 "Oneiric Ocelot" - Release i386 (20111012)
IpRoute:
 default via 10.64.64.64 dev ppp0 proto static
 10.64.64.64 dev ppp0 proto kernel scope link src 10.220.36.209
 169.254.0.0/16 dev ppp0 scope link metric 1000
MachineType: Acer Aspire One 522
NetworkManager.state:
 [main]
 NetworkingEnabled=true
 WirelessEnabled=true
 WWANEnabled=true
 WimaxEnabled=true
Package: network-manager 0.9.1.90-0ubuntu5
PackageArchitecture: i386
ProcEnviron:
 PATH=(custom, no user)
 LANG=en_GB.UTF-8
 SHELL=/bin/bash
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-3.0.0-12-generic root=UUID=19fb8d43-4848-49bf-94b9-a08eec919875 ro quiet splash vt.handoff=7
ProcVersionSignature: Ubuntu 3.0.0-12.20-generic 3.0.4
RelatedPackageVersions:
 linux-restricted-modules-3.0.0-12-generic N/A
 linux-backports-modules-3.0.0-12-generic N/A
 linux-firmware 1.60
RfKill:
 0: phy0: Wireless LAN
  Soft blocked: no
  Hard blocked: no
Tags: oneiric oneiric
Uname: Linux 3.0.0-12-generic i686
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare
dmi.bios.date: 03/21/2011
dmi.bios.vendor: Acer
dmi.bios.version: V1.09
dmi.board.name: Aspire One 522
dmi.board.vendor: Acer
dmi.board.version: V1.09
dmi.chassis.type: 10
dmi.chassis.vendor: Chassis Manufacturer
dmi.chassis.version: Chassis Version
dmi.modalias: dmi:bvnAcer:bvrV1.09:bd03/21/2011:svnAcer:pnAspireOne522:pvrV1.09:rvnAcer:rnAspireOne522:rvrV1.09:cvnChassisManufacturer:ct10:cvrChassisVersion:
dmi.product.name: Aspire One 522
dmi.product.version: V1.09
dmi.sys.vendor: Acer
http_proxy: http://213.196.2.97:8080/

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Sleepy John (john-eyre)]

apport information

[Mark Robbie (mark-robbie)]

If this problem is really due to a kernel bug (as suggested by Josua Dietze in comment #38) then wouldn't it go away if grub were used to boot an older kernel? I'm sure this was one of the things I tried when I first noticed the problem.

[Josua Dietze (digidietze)]

I think I have not been clear enough; the built-in switching function in the kernel is not related to the bug reported here. It has not changed since the 2.6.2x versions, AFAIK.
It just confirmed the fact that usb_modeswitch could not be involved with this bug in any way.

My assumption (rather gut-oriented) is that either the probing of modem-manager or some change in the usb-storage driver may cause a firmware crash or a device reset, probably due to a previously unexposed problem with the firmware itself.

[Fredrik Hultkrantz (fjutt)]

Hmm, but if that was the case why would the work around posted above (comment #13) work then? Would not the same problem occur even after that then, if it was a problem with the firmware? Clearly something changed between 11.10 and 11.04 that made this stop working, but I guess we already know that. I use comment 13's workaround every day, and it works, but it is slightly inconvenient to have to go through that to connect to the network...

[Josua Dietze (digidietze)]

The command from comment #13 will do nothing but send a standard device reset control message to the modem. Any tool using libusb could do that, so it's not a genuine property of usb_modeswitch. It was added just for convenience.

And the workaround is quite logical: if something is screwing up the device, a clean reset will make it work again. This is *not* a mode switch, the device stays in modem mode obviously.

BTW, there is no need to provide a target ID in the command since this is no mode switch with success check.
Just run "sudo usb_modeswitch -v 0x12d1 -p 1003 -R".

[Mark Robbie (mark-robbie)]

Attached is the output from dmesg at the time of plugging-in the modem. It doesn't mean a whole lot to me though the disconnection of device number 2 followed 0.7 seconds later by a connection of device number 3 seems odd.

[Josua Dietze (digidietze)]

The dmesg output from comment #73 does not indicate a problem. This is exactly what will happen every time during the device mode switch. The first device will 'vanish' and then return as a second one with a very different setup.

[Sleepy John (john-eyre)]

Just thinking aloud to try and understand this problem better, we know that the E220 can appear as two different USB devices, depending on which way it's switched. In one mode it looks like a storage device, and in the other it looks like a modem. Do we know whether these two modes are bistable, in other words can it equally remain in either of its two states, or is there a "default" state that it always reverts to when it's depowered? I ask this because please notice in my post #13 fix I find I have to unplug and replug my USB connection to get it to work. Is it possible for an E220 to be left in the "wrong" state when it's switched off?

I'm noticing that after being on-line with my E220 on Ocelot, if I then restart and go over to Windows and try to go on-line with it again using Huawei's Mobile Connect application, quite often it won't connect at the first attempt, and I have to unplug and replug its USB connection before it will connect properl.y. This leads me to suspect that when I close an on-line session on Ocelot, my E220 is being left in a different (wrong) state to what happens when I close an on-line session on Huawei's Mobile Connect.

Before I came up with post #13 fix, I was having to first prime my E220 by going on-line with it using either Huawei's Mobile Connect or with Natty, and from there doing a powered-on restart to Ocelot where it would then connect correctly. I guess this priming was effectively doing the same thing as usb_modeswitch does. There must be a clue in the fact that I need to do a USB unplug-replug after running "sudo usb_modeswitch -v 0x12d1 -p 1003 -R".

One other observation I'd like to mention is that when I first switch on Ocelot from cold and plug my E220 in, I can see my 3G mobile connection being displayed as available, so that leads me to believe my E220 has been switched into modem mode at that stage. However, if I try clicking on that connection, it fails to connect. I then run "sudo usb_modeswitch -v 0x12d1 -p 1003 -R". and that causes my 3G mobile connection available display to disappear for several seconds. When it becomes available again (i.e.switched back to modem mode) it still won't connect until I do a usb unplug replug. Only then, after another few seconds of it's availability disappearing, will it successfully connect.

There must be clues to what's happening hidden in these observations, but all I can offer at this stage are these observations as a brainstorm excercise.

[Josua Dietze (digidietze)]

> Is it possible for an E220 to be left in the "wrong"
> state when it's switched off?

Not that I know of. Power gone = back to install mode. Always. No known exceptions.

This would not make sense anyway: the point of the whole process is to have the (Windows) drivers available at *every* plugin. The device does not know if the machine it is plugged into has the drivers installed already.

The problem of this bug lies somewhere in the interaction between software (Network Manager) and modem device. The mode switch will happen all the time, but maybe the modem registration of Network/Modem Manager will persist during the re-plug and the probing of the port will be skipped the second time.

You can try to find out if there are differences between your first plug and the second (plus the reset) by looking at dmesg/modem-manager/networkmanager logs.

The way I see it, the bug appears first in the networkmanager.log from comment #2:

NetworkManager[5781]: <info> Activation (ttyUSB0) starting connection 'Dialog GSM Postpaid'
NetworkManager[5781]: <info> (ttyUSB0): device state change: disconnected -> prepare (reason 'none') [30 40 0]
NetworkManager[5781]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) scheduled...
NetworkManager[5781]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) started...
NetworkManager[5781]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) complete.
NetworkManager[5781]: <info> WWAN now enabled by management service
NetworkManager[5781]: <warn> GSM connection failed: (32) Serial command timed out
NetworkManager[5781]: <info> (ttyUSB0): device state change: prepare -> failed (reason 'unknown') [40 120 1]
NetworkManager[5781]: <warn> Activation (ttyUSB0) failed.

There is no indication of the problem in modem-manager.log (comment #1), but it shows the communication between the system and the modem.

Now, if we had a respective log from Natty, we could see any differences in this communication.

Annother thing to try is if access to storage interfaces is still possible after the modem problem becomes apparent (MicroSD card slot). This would rule out that the storage driver is the culprit by making the device hang somehow.

[Sleepy John (john-eyre)]

Josua Dietze (digidietze) wrote on 2011-12-01: #76

>> Is it possible for an E220 to be left in the "wrong"
>> state when it's switched off?
>
>Not that I know of. Power gone = back to install mode. Always. No known exceptions.
>
>This would not make sense anyway: the point of the whole process is to have the (Windows) drivers available at *every* plugin. >The device does not know if the machine it is plugged into has the drivers installed already.

Yes, I agree with that, but what about the modem mode in the dongle? The dongle has flash memory, so Isn't it possible for the status within the modem mode to be changed in such a way that the change survives an unplug?

 I'm still seeing inconsistent behaviour, but some of it suggests that when my dongle switches into modem mode, my system is not always seeing the same condition. It always gives an indication of the mobile service being available, but sometimes it won't connect and sometimes it will connect. How can we explain why an unplug-replug sometimes makes it connect, and sometimes it doesn't??

Today my Ocelot left my dongle in such a condition that when I tried to use it with my Windows Huawei Mobile Connect software, it wouldn't connect there either. No amount of unplugging-replugging or closing and reopening Mobile Connect would persuade it to connect, even though it was al.ways indicating that my mobile provider's service had been detected and was available. Only when I went back on Ocelot and performed a usb-modeswitch command on it did it finally connect again.

As you may gather, I'm still stuck in brainstorm mode !

[Mark Robbie (mark-robbie)]

Hmm. It seems I can only attach one file per comment.
So attached here is an annotated syslog from 11.04 where the modem works just fine. In order to rule out any local changes being the source of the problem, the system was booted from the live CD.

Subsequent post will have a syslog from 11.10 where the modem fails to work. There are significant differences between the two, but sadly I am not sufficiently familiar with this to determine the problem.

[Mark Robbie (mark-robbie)]

Attached here is the syslog from 11.10 (see previous comment #78)

[Josua Dietze (digidietze)]

Can you extract the modem-manager.log from 11.04?

[Mark Robbie (mark-robbie)]

Probably I am confused, but my belief is that by default modem-manager writes to syslog rather than a separate log file. A lsof of the modem-manager does not show an opened log file. Isn't it possible to extract the modem-manager log from either of the syslogs I attacked by simply running a command like:
  grep modem-manager syslog-1104.txt

Anyway, I have run these tests again (I'm using the live CDs so anything I don't think to maker a copy of gets lost) and over the next few posts will attach the modem-manager and NetworkManager logs from 11.04 and 11.10.

This time I have run them in debug mode with these commands (after first having stopped the existing instances):
modem-manager --debug --log-level=debug &>modem-manager.log
NM_PPP_DEBUG=1 /usr/sbin/NetworkManager --log-level=debug --no-daemon &>NetworkManager.log

[Mark Robbie (mark-robbie)]

Additional log - see comment #81

[Mark Robbie (mark-robbie)]

Additional log - see comment #81

[Mark Robbie (mark-robbie)]

Additional log - see comment #81

[Josua Dietze (digidietze)]

One striking difference between the two modem-manager.logs is that in 11.10 there is communication with ttyUSB1 in parallel, whereas in 11.04 it is "type ignored" after the probing.

The commands that are going to ttyUSB0 seem not to be different in both logs, but the response from there stops at the moment the command "AT+CGREG=2" is sent to ttyUSB1 (log from 11.10). This is supposed to switch on unsolicited status output on that port, to monitor the signal and other parameters during a connection.

This seems to be a fundamental difference in the modem handling between 11.04 and 11.10. I'm not familiar with the NM/MM code though, so I can't give further assistance here.

[Fiona Klute (fiona-klute)]

I've been using my E220 with wvdial since the upgrade to Oneiric, attached is the configuration I use (except the PIN). I tried to reproduce the "try to connect -> reset -> replug -> works with network-manager" workaround with limited success (worked sometimes, but not reliably).

I have some additional observations:

1. After plugging in the modem, I need to wait until usb_modeswitch is done before starting wvdial or it won't work. Network-manager recognizes the modem before the switch is complete and asks for the PIN, if I provide it the device won't work.

2. wvdial does not work after a failed network-manager connection attempt, no matter if with or without reset. I have to replug the device.

3. Network-manager does not recognize the modem after reset without replugging, even though the kernel log shows the disconnect/connect events.

[Chamila Mihiripenna (chamila-mihiripenna)]

can someone tell, when this issue going to be fixed ? in which release ?

[Bartosz Kosiorek (gang65)]

On Ubuntu 11.10 the NetworkManager is in version 0.9.1.90-0ubuntu5.1
On Ubunto 11.04 there is version 0.8.4~git.20110319t175609.d14809b-0ubuntu3

Is it possible to go back to the previous version of networkManager and check if connection is working?

[girts (girtsz)]

Question:
Why there is still beta packages under Oneiric? NetworkManager update contains not stable vesion. Please it is strongly recommend and others will ask I think too update packages to stable versions. Latest NetworkManager is 0.9.2. Canonical please update to latest stable version.

This bug must be critical but not Low!!!!

Sorry, but problems are stil which must be solved!!!
No beta, pre-realease packages in latest versions, Oneiric too.

[Bartosz Kosiorek (gang65)]

@girts
Try to install latest stable NetworkManager (0.9.2) and check if it solved problem with connection.
For Debian package is available at:
http://packages.debian.org/pl/sid/network-manager
All other:
http://ftp.gnome.org/pub/GNOME/sources/NetworkManager/0.9/

[Mathieu Trudel-Lapierre (cyphermox)]

Doesn't seem to be a kernel issue, but more of some kind of interaction between network-manager and modemmanager. No need to keep the bug task open for linux.

Bartosz; that's not very helpful. Debian packages are very different that those we use for Ubuntu, at least for NetworkManager.

The package available in 11.10 is one that was thoroughly tested with various different scenarios. While it may happen that things slip and fail to work; in this particular case it appears to be very specific to that device. If not specific to all Huawei E220s, then specific to that particular dongle/firmware.

[Mathieu Trudel-Lapierre (cyphermox)]

For the purpose of testing, could someone having this issue please confirm the exact version of network-manager and modemmanager you have? This can be achieved with the following command:

dpkg -l network-manager modemmanager

Also, it may be helpful to test with a later version of NetworkManager from git; fortunately, we do have daily builds of NetworkManager, but they require libnl3 3.2.3; so I'll prepare a different package without that new requirement for Oneiric and put it in my own PPA.

[Fiona Klute (fiona-klute)]

@mathieu-tl
I have the following versions:

||/ Name Version Description
+++-===========================-===========================-===========================================
ii modemmanager 0.5-1ubuntu1 D-Bus service for managing modems
ii network-manager 0.9.1.90-0ubuntu5.1 network management framework (daemon and userspace tools)

[Fredrik Hultkrantz (fjutt)]

I have almost the same, just not .1 on the network-manager.

$ dpkg -l network-manager modemmanager
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version Description
+++-======================================-======================================-============================================================================================
ii modemmanager 0.5-1ubuntu1 D-Bus service for managing modems
ii network-manager 0.9.1.90-0ubuntu5 network management framework (daemon and userspace tools)

[Mathieu Trudel-Lapierre (cyphermox)]

Updating to link to the most likely upstream bug report. (https://bugzilla.gnome.org/show_bug.cgi?id=662113)

[Christopher Forster (christopherforster)]

There should be drivers built in to support the Huawei modems. I have and I'm still having some problems with the 3g connection since Ubuntu 9.10 with the Huawei E220 and Huawei E1550 modems. There may be problems getting the dongles to work from Ubuntu. I sometimes solved and got the modems working by unplug and plug back in workaround. Kubuntu 11.10 has sometimes connected with the Huawei E220 modem.

@Sleepy John (john-eyre) comment #13

"CODE
sudo usb_modeswitch -v 0x12d1 -p 1003 -V 0x12d1 -P 1003 -R

Then unplug and replug the dongle, and now it will connect."

Solved the problem, this works perfect. Kubuntu 11.10 now connects with my Huawei E220: "Connected to Tele2 - Mobilt Bredband". Thanks Sleepy John (john-eyre).

[Sleepy John (john-eyre)]

Almost identical in my case, too:

dpkg -l network-manager modemmanager
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name Version Description
+++-==============-==============-============================================
ii modemmanager 0.5-1ubuntu1 D-Bus service for managing modems
ii network-manage 0.9.1.90-0ubun network management framework (daemon and use

[Akila Wajirasena (akila-wajirasena)]

Some similar bugs
https://bugzilla.redhat.com/show_bug.cgi?id=717727
https://bugs.launchpad.net/ubuntu/+source/modemmanager/+bug/875455

https://bugzilla.redhat.com/show_bug.cgi?id=720634
https://bugzilla.redhat.com/show_bug.cgi?id=767439
https://bugzilla.redhat.com/show_bug.cgi?id=746518

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Bartosz Kosiorek (gang65)]

ModemManager repository is available at:
http://cgit.freedesktop.org/ModemManager/ModemManager/

[Sleepy John (john-eyre)]

As one of the mere users who rely on a Huawei E220 for my internet access, and as one who doesn't fully undertand what's going on with getting this bug fixed, I'm finding it frustrating to see the status keeps changing and no apparent progress being made towards correcting it. Maybe I misunderstand, but how can the importance of this bug suddenly be changed to low without consulting all the Huawei users who have subscribed to this bug and have confirmed the difficulties they are having as a result of it. In simple terms our modems were working just fine in Natty and earlier versions, but as soon as we ugraded to Ocelot they have ceased to work unless we go through awkward lengthy and unreliable workarounds to log-in. I for one am becoming weary of this situation, especially when I see the priority for getting it fixed appears to have been downgraded apparently without regard for us the Huawei users.

I would like to exhort any other Huawei users on here who are experiencing similar difficulities to voice their opinion about the priority recently being downgraded to low again.

[Aleksander Morgado (aleksander-m)]

Upstream devs know about the issue, and I was supposed to try to reproduce this with my Huawei 220. I just don't have time to test it myself whenever I remember that I have to test it. Can someone provide modem-manager debug logs with the previous version, the one that worked?

[Mark Robbie (mark-robbie)]

I had previously attached modem-manager debug logs from a successful (ubuntu 11.04) and unsuccessful (ubuntu 11.10) system at comments #81 and #83 respectively. However when I look at this issue now, I can only see those logs if I click on the individual comment (i.e. my attachments do not show in the view that lists all comments - and in fact comments 83 and 84 do not show at all). So I hope you can see the logs at the following addresses:

https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/868034/comments/81
https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/868034/comments/83

If not, please let me know as I still have the logs.

[Aleksander Morgado (aleksander-m)]

Mark: Thanks for the logs

I can find two main differences in the MM 0.4 (the one which worked) and MM 0.5 (the one which failed) log.

 1) In the MM 0.5 log, the secondary ttyUSB1 gets probed and grabbed as secondary port; while in the MM 0.4 log capability probing in that port is not even launched (just grep for "USB1" in both logs, and compare the results)

In the MM 0.4 log:
<debug> [mm-modem-base.c:155] mm_modem_base_add_port(): (ttyUSB1) type ignored claimed by /sys/devices/pci0000:00/0000:00:1a.1/usb4/4-2

In the MM 0.5 log:
<debug> [mm-modem-base.c:155] mm_modem_base_add_port(): (ttyUSB1) type secondary claimed by /sys/devices/pci0000:00/0000:00:1a.1/usb4/4-2

This is already pretty weird, as in MM 0.4 the ttyUSB1 port is grabbed but wasn't probed at all. I wonder if there is some kind of issue with the kernel driver which exposes two ports with usbif=0. Don't know if that is even possible.

2) MM 0.5 has some changes in the Huawei plugin. We now use "AT^CURC", "AT^GETPORTMODE" commands when it gets initialized. AT^CURC=0 when the first port is tried will disable unsolicited messages in the secondary port, and AT^GETPORTMODE tries to query

In any case, the problem seems to be that the modem gets rebooted at some point. This happens after we enable unsolicited messages in the secondary port with AT^CURC=1, and after we send the first commands to that secondary port. Last logs from the modem show this BOOT message.
<debug> [mm-at-serial-port.c:298] debug_log(): (ttyUSB1): <-- '<CR><LF>^BOOT:92988511,0,0,0,6<CR><LF>'

If someone can do more tests, I would now try:

A) Try the MM 0.4 version from Ubuntu 11.04 directly in Ubuntu 11.10. You should be able to dpkg --purge modem-manager (no package depends on it in Ubuntu) ad then dpkg -i the specific deb retrieved from the Ubuntu 11.04 archive. This would help to see if we do get ttyUSB1 grabbed also as secondary port in the newer kernel.

B) Hack the Huawei plugin in a custom MM 0.5 branch so that it ignores all ports except for the primary one grabbed. This can be done by changing the whole last else {} switch in grab_port() within plugins/mm-plugin-huawei.c, to just mm_modem_grab_port() passing ptype = MM_PORT_TYPE_IGNORED. This would help to see if we really are managing wrong the secondary port, and is just better to skip it in this specific modem, or try to refix it if we know how.

[Sleepy John (john-eyre)]

I'm pleased to see the recent progress being made by Mark and Aleksander (comments #102 - #104), though not happy that the priority still remains listed as "low".
Meanwhile last night I managed to locate syslog file entries from both Maverick (ver 0.4) and from Ocelot (ver 0.5) for plugging in and trying to log-in on my Huawei E220, and I'm attaching the Maverick one here. I'll attach the Ocelot one for comparison in the next post. These differ from Mark's examples in that they show kernel entries as well as MM entries, so perhaps there could be an additional clue there.
I'd love to try MM 0.4 in Ocelot and see what happens, but sorry I don't feel I'm quite up-to-speed enough myself to do that without risking an error that cuts me off completely!

[Sleepy John (john-eyre)]

OK, I've now posted two attachments:
mmMavsyslog - shows both kernel and MM entries on Maverick (mm ver 0.4)
mmOcesyslog - shows both kernel and MM entries on Ocelot (mm ver 0.5)

[Josua Dietze (digidietze)]

If I may refer to my comment #85 - I had a look at the "working" and "nonworking" log of modem-manager and pointed out the obvious differences.

I strongly suspect that the newer version of MM is attempting to make use of the diagnostic port, which can give status information (like signal strength) even during a connection.
In this case though, it affects the modem in some bad way; I doubt the log will help much to pinpoint the underlying problem ...

[Aleksander Morgado (aleksander-m)]

Sleepy John: syslog log is pretty useless if ModemManager wasn't run with --debug, :-/ And regarding the priority, I really don't give a s*** about it :-) I personally do this in my free time, so at least for me, it doesn't matter if it was low or super-critical, I'll try to help if I have time.

Josua Dietze: The first test I suggest is to ensure that the newer kernel has no effect on this issue. If MM 0.4 works ok with the newer kernel, it's probably MM's fault. The second test I suggest is to completely avoid using the secondary port. If that test succeeds we can probably force MM to avoid using the secondary port for this specific Huawei device, at least until we can figure out why it failed that way. So logs (in debug mode) for these tests will really help.

[Aleksander Morgado (aleksander-m)]

I'll follow up the issue in GB#662113

[Mark Robbie (mark-robbie)]

Aleksander: I have installed modem-manager 0.4 on my 11.10 system and it works ok - a debug log is attached (it contains the sequence connect/disconnect/connect)

It was quite painless for me to install the older release, so this might be a suitable interim solution for some other people who are having this problem.

To install the version from Natty you need to:

Step 1:
Fetch the old .deb file. You can get this from:
http://packages.ubuntu.com/natty/modemmanager
In the section "Download modemmanager" click on either amd64 or i386 as appropriate for your machine and save the file somewhere.

I'm using 64-bit so the file I retrieved is called:
modemmanager_0.4+git.20110124t203624.00b6cce-2ubuntu1_amd64.deb

Step 2:
Remove the current (broken) version of modemmanager from your machine.
sudo dpkg --purge modemmanager

Step 3:
Install the older version that you downloaded in Step 1.
sudo dpkg -i modemmanager_0.4+git.20110124t203624.00b6cce-2ubuntu1_amd64.deb

Step 4:
(not 100% sure this is needed but won't hurt)
Bounce the NetworkManager
sudo stop network-manager
sudo start network-manager

The modem should now work as it did in 11.04.

You might want to perform an additional step to prevent the update manager from reinstalling the broken version:

Optional Step 5:
Start the Synaptic Package Manager
Find modemmanager in the packages list. It should look something like:
S | | Package | Installed Version | Latest Version | Description
! modemmanager | 0.4+git.20110124t203624.00b6cce-2ubuntu1| 0.5-1ubuntu1 | D-Bus service for managing modems

Click on the modemmanager package, then from the menu select Package -> Lock Version

Obviously at some future time when this bug is fixed you will want to unlock this version by again selecting Package -> Lock Version

[Shrijit Nair (shrijith1)]

Mark, can not thank you enough. Works just perfect with the work around you have explained. Thanks, a ton. Hope it gets fixed soon.

[Aleksander Morgado (aleksander-m)]

Great, so anyone requiring to use MM with the Huawei modem can fallback to 0.4 until the proper fix arrives in a 0.5 update.

Mark: step 4 not really needed; just "killall modem-manager" should do it after installing the new .deb; and even that may not be needed depending on the post install packaging rules.

[Bartosz Kosiorek (gang65)]

After install modemmanager_0.4+git.20110124t203624.00b6cce, the problem was gone for me.
Thanks.

[Bartosz Kosiorek (gang65)]

Here is all modemmanager version for Oneric:
https://launchpad.net/ubuntu/oneiric/+source/modemmanager

Working version for Oneric:
http://launchpadlibrarian.net/68965258/modemmanager_0.4%2Bgit.20110124t203624.00b6cce-2ubuntu1_i386.deb
Not working version for Oneric:
http://launchpadlibrarian.net/71813592/modemmanager_0.4%2Bgit.20110429t103114.863dbca-1_i386.deb

And here is the diff between this packages:
http://launchpadlibrarian.net/71813411/modemmanager_0.4%2Bgit.20110124t203624.00b6cce-2ubuntu1_0.4%2Bgit.20110429t103114.863dbca-1.diff.gz
https://launchpad.net/ubuntu/+source/modemmanager/0.4+git.20110429t103114.863dbca-1

Hope it helps to track the bug.

[Sleepy John (john-eyre)]

I have also installed the workaround Mark describes in post #110. In my case modemmanager_0.4+git.20110124t203624.00b6cce-2ubuntu1_i386.deb

I find now that if my Huawei is plugged in at bootup, it will connect correctly first time when I request it, but if I don't make the USB connection until after bootup, it will not connect unless I go through a sudo usb_modeswitch -v 0x12d1 -p 1003 -R followed by an unplug-replug routine like I had to do all the time with mm 0.5.

So this is a vast improvement that will make life much easier than it was with mm 0.5, but it still leaves a shortcoming (at least in my case) compared to Maverick and Natty that either a system boot or a usb_modeswitch and unplug-replug is necessary before the Huawei will connect.

Anyway, let me also add my grateful thanks to Aleksander and to Mark for this solution.

[antero_h__ (antero-halminen)]

Hand building ModemManager from modemmanager_0.5-1ubuntu1 sources replacing plugins/mm-modem-huawei-gsm.c and plugins/mm-plugin-huawei.c with same from commit 1e332a but with dc89c0 reverted as suggested by http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=626712 also mostly fixed broken internet with a Huawei E169.

[Aleksander Morgado (aleksander-m)]

Antero: cool, so we can possibly revert dc89c0 until we get a better solution; will suggest that upstream.

[Bartosz Kosiorek (gang65)]

Hi Aleksander.

What is the status of this bug?

[Bartosz Kosiorek (gang65)]

Commented the bug in the upstream ML:
http://mail.gnome.org/archives/networkmanager-list/2012-January/msg00020.html

Need to decide whether we revert the commit temporarily or try to find a better
solution.

[Aleksander Morgado (aleksander-m)]

If anyone willing to help with this, please checkout the 'huawei-fix-guessing' branch from the following git repo: git://gitorious.org/aleksander/modemmanager.git and tell me if it worked. If it didn't work, please drop me an email with the debug logs.

[Bartosz Kosiorek (gang65)]

Thanks Aleksander.

I made following steps:
1. git clone git://gitorious.org/aleksander/modemmanager.git
2. cd modemmanager/
3. git checkout huawei-fix-guessing
4. ./autogen.sh --prefix=/usr
5. Install missing packages, and try again to configure
6. make
7. sudo make install

Unfortunately after restart, the modem (Huawei E220) was not detected (in network manager).
I restarted several times, and unplug/plug modem, without success.

[Bartosz Kosiorek (gang65)]

Maybe someone else will follow this instruction and check if it solve problem...

[Aleksander Morgado (aleksander-m)]

Bartosz: could you get the debug log with the following steps?

1) sudo mv /usr/bin/modem-manager /
2) sudo killall -9 modem-manager
3) sudo /modem-manager --debug

plug your modem here

4) Ctrl+C
5) sudo mv /modem-manager /usr/bin/

[operator494D (operator494d)]

hi people

just installed 11.10 and have this problem from start.
i had this same problem on maverick also ,random disconnections often with complete system crash(keyboard leds flashing and system not responding).dont know which modem manager i was using then but it was all the same with all maverick kernels and updates.

i am using huawei E153 ,and i noticed this was mostly happening when switching from 2g to 3g .when i set up connection to use only 3g crashes went down significantly.
When i used gnome ppp there was no crashes

now with 11.10 i was using its own mm 0.5 and it was horror.
after disconnection modem was visible with lsusb and usb-devices but i couldnt make it work even with manually running usb-modeswitch .

a note for gnome ppp in 11.10 ,for some reason it gave me error saying it cannot start pppd.
after runing it as root it was ok

now i followed above advice in #110 and tried mm 0.4 .so far so good(only one disconnect but w/o long term side effects :)

[Bartosz Kosiorek (gang65)]

One more question before I proceed:
Where is stored modemmanager logs?

[Bartosz Kosiorek (gang65)]

Hi Aleksander.
It works perfectly!
I don't know why it wasn't work yesterday. I have sucessfully connected to internet.

I modified your steps (the modem-manager is stored in /usr/sbin directory):
1) sudo mv /usr/sbin/modem-manager /
2) sudo killall -9 modem-manager
3) sudo /modem-manager --debug

plug your modem here

4) Ctrl+C
5) sudo mv /modem-manager /usr/sbin/

[Bartosz Kosiorek (gang65)]

In attachment you have modem detection:
patched-modem-manager-detection.log

and modem detection and connection to network:
patched-modem-manager-connection.log

[Aleksander Morgado (aleksander-m)]

Hum... not really fixed.

The whole point of the patch was to avoid sending any command to the secondary port, while still configuring the handlers to get unsolicited messages for signal quality and registration updates and such. But I just checked the logs again and the modem returns an error to AT+CIND=?, which means we will probably not get any CIEV unsolicited messages, and therefore we do need to periodically query for signal quality/registration status. And to do that while connected, we need to be able to send commands to the secondary port.

I'll try to prepare a new patch with the ideas Dan gave in the ML:
http://mail.gnome.org/archives/networkmanager-list/2012-January/msg00123.html

[Aleksander Morgado (aleksander-m)]

Bartosz:

Wait, could you please retest with the same patch, and leave it connected for some time, to check if we do get any unsolicited message in the secondary port? The huawei modems do not work with CIND/CIEV, they have their own RSSI based signal quality update notifications, so it may be that we do get them after all.

[Bartosz Kosiorek (gang65)]

I have restarted Ubuntu and follow the instruction.
After plug modem, I tried to connect.... WIthout success.
I tried several time... Still without luck

Log in attachment.

[Bartosz Kosiorek (gang65)]

Next I unplug modem, restart modem-manager, and plugin modem.
The modem successfully connected to network.

It is really strange why it sometimes connect, and sometimes not.
Log is in "huaweie220_connection_successfull_after_replug.log" file.

[Bartosz Kosiorek (gang65)]

How I could check the firmware of my E220 modem.
 $ lsusb -vvv ??

[Aleksander Morgado (aleksander-m)]

The reply to AT+GMR/CGMR would give you the revision id of the firmware.

[Bartosz Kosiorek (gang65)]

@Aleksander
Do you need some more logs, or it is enough to solve this bug?

[Ubuntu Foundations Team Bug Bot (crichton)]

The attachment "huawei_fix.patch" of this bug report has been identified as being a patch. The ubuntu-reviewers team has been subscribed to the bug report so that they can review the patch. In the event that this is in fact not a patch you can resolve this situation by removing the tag 'patch' from the bug report and editing the attachment so that it is not flagged as a patch. Additionally, if you are member of the ubuntu-reviewers team please also unsubscribe the team from this bug report.

[This is an automated message performed by a Launchpad user owned by Brian Murray. Please contact him regarding any issues with the action taken in this bug report.]

[Aleksander Morgado (aleksander-m)]

Bartosz: we want to try another approach to fix it. As suggested by Dan in the ML, we will try to use the primary port for connection only, and the secondary port for all AT commands. Currently I just need some free time to hack that, and after that I will push a new branch for testing.

Having such a modem for the fix development would be the best option, so if anyone is willing to send me one of those, I'll happily accept it :-)

[Kent Lin (kent-jclin)]

Following 2 3G dongles also have same issues:
Huawei E220 & Huawei E169u

[Aleksander Morgado (aleksander-m)]

Can anyone try the 'ports' branch from the upstream MM repo:
  git://anongit.freedesktop.org/ModemManager/ModemManager

Ref: http://mail.gnome.org/archives/networkmanager-list/2012-February/msg00011.html

[linctus (yongjin-liu)]

try to clone by

git clone git://anongit.freedesktop.org/ModemManager/ModemManager.git

but failed to access the repository.
And the error message is:

anongit.freedesktop.org[0: 131.252.210.161]: errno=Connection timed out
fatal: unable to connect a socket (Connection timed out)

[Bartosz Kosiorek (gang65)]

I have the same problem as linctus.

[linctus (yongjin-liu)]

Hi there,

Seems the project adress should be git://anongit.freedesktop.org/git/ModemManager/ModemManager.git ruther than git://anongit.freedesktop.org/ModemManager/ModemManager.git

But I cannot find a 'ports' branch in git://anongit.freedesktop.org/git/ModemManager/ModemManager.git

regards.

[Aleksander Morgado (aleksander-m)]

The freedesktop.org git server had some problems during the last days it seems, please retry with the URL I originally gave:
  $> git clone git://anongit.freedesktop.org/ModemManager/ModemManager

And then:
  $> cd ModemManager
  $> git checkout -b ports origin/ports

[Bartosz Kosiorek (gang65)]

I tested "ports" branch on Ubuntu 11.10 Oneiric and it seems work.
I will test it further, and attach debug logs.

By the way: I cannot build on Ubuntu 12.04 Precise. It display warnings/errors:

CC modem_manager-mm-modem.o
mm-modem.c: In function ‘value_array_add_uint’:
mm-modem.c:289:5: error: ‘g_value_array_append’ is deprecated (declared at /usr/include/glib-2.0/gobject/gvaluearray.h:71): Use 'g_array_append_vals' instead [-Werror=deprecated-declarations]
mm-modem.c: In function ‘get_ip4_done’:
mm-modem.c:310:9: error: ‘g_value_array_new’ is deprecated (declared at /usr/include/glib-2.0/gobject/gvaluearray.h:58): Use 'g_array_new' instead [-Werror=deprecated-declarations]
mm-modem.c: In function ‘info_call_done’:
mm-modem.c:395:9: error: ‘g_value_array_new’ is deprecated (declared at /usr/include/glib-2.0/gobject/gvaluearray.h:58): Use 'g_array_new' instead [-Werror=deprecated-declarations]
mm-modem.c:400:9: error: ‘g_value_array_append’ is deprecated (declared at /usr/include/glib-2.0/gobject/gvaluearray.h:71): Use 'g_array_append_vals' instead [-Werror=deprecated-declarations]
mm-modem.c:406:9: error: ‘g_value_array_append’ is deprecated (declared at /usr/include/glib-2.0/gobject/gvaluearray.h:71): Use 'g_array_append_vals' instead [-Werror=deprecated-declarations]
mm-modem.c:412:9: error: ‘g_value_array_append’ is deprecated (declared at /usr/include/glib-2.0/gobject/gvaluearray.h:71): Use 'g_array_append_vals' instead [-Werror=deprecated-declarations]
cc1: all warnings being treated as errors

Is it possible to fix this warnings/errors?

[Aleksander Morgado (aleksander-m)]

Bartosz: sure, I will remove the deprecated declarations myself

[Bartosz Kosiorek (gang65)]

In attachment there is two logs from successfuly connect via Huawei E220 on Oneiric.

[Bartosz Kosiorek (gang65)]

Hi Aleksander.
Is it possible to push this patch to master?
It is works perfectly for me.

[Aleksander Morgado (aleksander-m)]

Next MM 0.5 stable release will come with commit dc89c0 reverted; the latest patch with the ports improvements will come later in master/0.6.

[Joakim Lagergren (jb-lager)]

Hi Aleksander.

I've had the same problem since I updated my Huawei E220 to firmware revision 11.117.10.02.00 and 11.313.02.00.01 as detailed in duplicate bug #932824 <https://bugs.launchpad.net/ubuntu/+source/modemmanager/+bug/932824/>. Before firmware update (possibly with firmware 11.110.xx?) everything was working fine.

I can confirm that the "ports" branch works for me. Debug log of successful connection is attached.

[SuperEngineer (suprengr)]

This is *obviously* an issue with rhythmbox / banshee / system equivalent search looking for removeable media. It has been a similar problem on last release of rhythmbox on 10.04 - except on 10.04 by asking for a reconnect following disaconnect - all is well. On 11.10 - try as many times as you dare - no reconnect. Only way round is given here - https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/868034 - surely that one plain fact can help???

[SuperEngineer (suprengr)]

[& putting the bug severity status as "medium" [though definitely not critical] is worrying... how do Ubuntu users update on slow connections when WiFi is a poor option if their alternative [only?] connection is killed?] This is being posted from 10.04 via the modem in question. Couldn't do it from 11.10 :( ... there is surely reason to up the status?

[Дмитрий (pakirava-datsuma)]

I'm thirty, who encountered this problem on my E220. I solved this problem as follows:
$ sudo echo "AT^U2DIAG=0" >> /dev/ttyUSB2
This command takes my modem into "Modem only" forever. And the manager is no longer will have to switch the modem from the CD-Mode
And I always can do next thing to turn on "CD-first-mode"
$ sudo echo "AT^U2DIAG=1" >> /dev/ttyUSB2

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Launchpad Janitor (janitor)]

This bug was fixed in the package modemmanager - 0.5.1.97-0ubuntu1

---------------
modemmanager (0.5.1.97-0ubuntu1) precise; urgency=low

  * New upstream bugfix release.
    - release: update NEWS
    - cdma: fix crash on NULL error (bgo #670145)
    - Revert "huawei: rework probing and detection" (LP: #868034)
    - core: fix loop limits in echo removal
    - at-serial-port: implement built-in echo/garbage removal (LP: #916038)
 -- Mathieu Trudel-Lapierre <email address hidden> Wed, 07 Mar 2012 16:22:00 -0500

[Aleksander Morgado (aleksander-m)]

Can anyone test this last 0.5.1.97~0ubuntu1 version with the specific device?

[Bartosz Kosiorek (gang65)]

@Aleksander
I will test on my Huawei E220 and let you know about results

[Bartosz Kosiorek (gang65)]

@Aleksander
The modemmanager 0.5.1.97~0ubuntu1 works much better for me on Ubuntu Precise (12.04).

[Aleksander Morgado (aleksander-m)]

@Bartosz: great! :-)

This change is *not* in Ubuntu 11.10 yet, packagers should probably upgrade to the next stable in 11.10 as well.

[Mathieu Trudel-Lapierre (cyphermox)]

We'll cherry-pick that patch separately to provide as SRU on 11.10.

[Mathieu Trudel-Lapierre (cyphermox)]

Can't cherry-pick this just yet, as there was a regression identified for HSO devices.

[Aleksander Morgado (aleksander-m)]

The regression for HSO devices was due to the echo removal patch, not because of the huawei-specific revert.

[Mathieu Trudel-Lapierre (cyphermox)]

Ah, you're right :)

[Mathieu Trudel-Lapierre (cyphermox)]

Uploaded to the oneiric-proposed queue. Now this will still have to wait to be accepted by a member of the ubuntu-sru team, which is subscribed.

[Jonathan Riddell (jr)]

ubuntu-sru: ping?

[Martin Pitt (pitti)]

Hello අකිල, or anyone else affected,

Accepted modemmanager into oneiric-proposed. The package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

[Ian Gibbs (flash666)]

The updated package is not in oneiric-proposed. Has something gone wrong in the release process?

[Fiona Klute (fiona-klute)]

After using modemmanager 0.5-1ubuntu1.1 for about a week, I can say that it works reliably for me.

[Launchpad Janitor (janitor)]

This bug was fixed in the package modemmanager - 0.5-1ubuntu1.1

---------------
modemmanager (0.5-1ubuntu1.1) oneiric-proposed; urgency=low

  * debian/patches/git_revert_huawei_probing_rework_7e406c2.patch: revert the
    rework for probing Huawei devices that was introduced in 0.5; it appears to
    break the Huawei E220 and as such constitues a regression. The revert
    appears to let all devices properly work (even the newer ones) and is
    included in version 0.5.2. (LP: #868034)
 -- Mathieu Trudel-Lapierre <email address hidden> Wed, 14 Mar 2012 15:35:28 -0400

[quiricada (raraland)]

ubuntu 12.04 beta2 with all updates
3.2.0-23-generic-pae
modemmanager 0.5.2.0-0ubuntu2

huawei e220 cannot connect

[garybrlow (garybrlow)]

Confirm here: https://bugs.launchpad.net/ubuntu/+source/modemmanager/+bug/991680

[Pablo (pjferra)]

I have the same problem with ubuntu 12.04 :(

[Sleepy John (john-eyre)]

Confirm I'm also seeing the same original problem in 12.04 when I test it from a live CD

[Akila Wajirasena (akila-wajirasena)]

I can connect usb modem from modem manager without the workaround. But still auto connect is not working.

[Pablo (pjferra)]

In 12.04 the "workaround" sometimes works, sometimes not.
Besides that, when I connect the dongle, many times the system freezes for a while.
Several times I've got a crash message due to modem-manager. I've submitted the crash report, but I don't know how to get it to put it on this thread.
It's really very annoying. I had the problem with 11.10 and for this reason I went back to 10.04, and I don't want to go back again.

[Aleksander Morgado (aleksander-m)]

Can anyone get MM debug logs?

[Дмитрий (pakirava-datsuma)]

Tell us how to do this. And what version of Ubuntu needed.

[Aleksander Morgado (aleksander-m)]

https://wiki.ubuntu.com/DebuggingModemmanager and any ubuntu release with MM 5.2

[Akila Wajirasena (akila-wajirasena)]

Here's the logs I've taken. and the steps I have taken;

1. Restart the computer (with modem disconnected)
2. Kill network manager and modem manager
3. Connect the modem
4. Modem didn't auto connect and nothing happened. (I have attached the log up to this point)
5. Click to connect the modem after waiting about a minute or two
6. Then it started trying to connect but didn't connect (I waited for 4, 5 tries and click disconnect. The logs up to this point is also attached.)

[Akila Wajirasena (akila-wajirasena)]

Modem did not connect only when I'm doing the test. In normal conditions even if I connect the modem after logging in it will connect, if I click to connect. But connect automatically does not work in any condition.

[Josua Dietze (digidietze)]

Looking at Akila's MM log, I see two possible issues.

1. There may be an offending command here:

[1317801779.197484] [mm-at-serial-port.c:298] debug_log(): (ttyUSB0): --> 'AT+COPS=3,0;+COPS?<CR>'

After that command, there is no more answer from ttyUSB0.

2. Very close to that point there is an error on ttyUSB1 after the failed command:

[1317801779.161136] [mm-at-serial-port.c:298] debug_log(): (ttyUSB1): --> 'AT+CREG=2<CR>'

The response on ttyUSB1 is "+CME ERROR: 3". It might have messed up the modem somehow.

[Дмитрий (pakirava-datsuma)]

Here are my logs.
Ubuntu 12.04 32bit + Huawei E1550

[Дмитрий (pakirava-datsuma)]

Here are my logs.
Ubuntu 12.04 32bit + Huawei E1550

[Chamila Mihiripenna (chamila-mihiripenna)]

I'm also seeing the same original problem in 12.04, thought this was fixed.
Why such a simple fix is dragging.... I wonder...

[Aleksander Morgado (aleksander-m)]

Reverting the huawei rework patch seemed to have the bug fixed or at least minimized. The proper fix is already in the MM_06 and git master branches, but it cannot/shouldn't be backported to MM_05, as it is quite a change. A hardcore workaround would be to force the huawei plugin to grab only the primary AT port, no secondary. I do not have such a modem, so if anyone wants to try that... patches welcome :-)

[Wolfgang (wt-lists)]

I have exactly the same problem with 12.04 and Huawei E220

[Pablo (pjferra)]

Is there any progress?
I cannot use my Huawei e220 modem but just sometimes (when it wants to work)
With ubuntu 10.04 no problem. Upgrading to 11.10 o 12.04 breaks the recognizing of the modem
I've tried all the "solutions" proposed in this thread:
        1) sudo usb_modeswitch -v 0x12d1 -p 1003 -V 0x12d1 -P 1003 -R
        2) sudo echo "AT^U2DIAG=0" >> /dev/ttyUSB1
        3) modify the /lib/udev/rules.d/40-usb_modeswitch.rules file
        4) downgrade modemmanager to 0.4 version
None of them worked for me. Some times the modem is recognized and many many times not.
It's really annoying.
Thanks in advance

[Sleepy John (john-eyre)]

I have to go along with Pablo's comment above. Where exactly are we now with this bug?

I'm aware of lots of good investigative work having been done, some problems and solutions having been found, and I can see reports of fixes having been released in both Oneric and Precise, but I'm still lost and confused. Have these fixes actually come out as updates now? I've been watching the updates in Oneric but haven't noticed any. Then if and when updates are issued, how will I know, and am I at that point expected to revert to the latest modem manager?

But as a mere E220 user who has been suffering the effects of this bug since Oct 2011, and who hasn't been able to follow all the technical details here, I'm still lost. Considering my Oneric is fully updated, should I now be able to plug my E220 in, click on it under Mobile Broadband, and log-in? Or do I still have to plug it in before booting my Oneric in order to get it to work?

I would very much appreciate (and I expect Pablo and others would too) a review in simple language of where we think we are with this bug now, and what we should be doing to get our E220s to log-in easily.

I would also like to ask for clarification on the status of this bug in Precise. Is it the same as in Oneric? If I were to install Precise now and apply all the latest updates, do we think my E220 would log-in properly like it used to do in 10.04?

[Pablo (pjferra)]

Sleepy John,

I'm sorry to tell you that I'm using Precise nowadays, and I'm suffering exactly the same issue :(

[Sleepy John (john-eyre)]

Thanks for that info, Pablo. I think it reinforces my view that a review is needed. I haven't installed Precise myself yet because I was afraid what you now confirm might be the case despite a Precise fix having been released.

[Shrijit Nair (shrijith1)]

I too have been following this bug quite closely for sometime. I am on 12.04 now and still using some crazy work-arounds to get my modem working. I use a ZTE modem and I think that this issue might not be specific to Huawei modem. The work-arounds I use are:
1) Remove the default modemmanager and install modemmanager_0.4+git.20110124t203624.00b6cce-2ubuntu1
2) Insert the USB modem before booting up Ubuntu.

[Pablo (pjferra)]

I've just (10 minutes) updated my ubuntu 12.04 with the last actualizations. One of them updated modemmanager to 0.5.2 and promised to solve this bug (with an explicit reference).
Restarted my system and everything goes on as it was yesterday. My Huawei modem doesn't work with ubuntu 12.04. It just is not recognized by networkmanager. lsusb shows the huawei modem, but I can't see it to connect my computer to the web.
By the way, 2 months ago I was using 10.04 and it worked perfectly, and I've got another computer with 10.04 and goes on working. So is not a problem of the modem.
The fix released simply doesn't fix anything :(

[Mathieu Trudel-Lapierre (cyphermox)]

SleepyJohn: Oneiric should have a similar patch applied from what is in Precise, if you have all the updates installed and your modem is working, then this tells me the patch itself is not the issue; there is something else that needs to be done in Precise to fix these devices.

For testing purposes, I can prepare a testing package in a PPA for those who would like to verify if their Huawei E220 modem works properly in ModemManager 0.6. Unfortunately, as it was mentioned by Aleksander, not all the changes in ModemManager 0.6 can be provided as a stable update.

What I would very much like, is that someone having this issue on Precise would file a new separate bug, because this with comments about Oneiric, about other releases, and with all kinds of workarounds is getting very difficult to follow and debug. As far as I know we fixed the issue in Oneiric, there may be a different one in Precise and as such it needs to be debugged separately.

[Mathieu Trudel-Lapierre (cyphermox)]

I have a E220 now, so I should be able to do some amounts of testing with it.

[Mathieu Trudel-Lapierre (cyphermox)]

Something else that is *very* relevant; if you're running into the issue and filing a new bug, please follow the steps listed above for getting debug logs (see http://wiki.ubuntu.com/DebuggingNetworkManager); and look for the firmware version for your device. The lines should be similar to this:

modem-manager[5955]: <debug> [1337321031.098461] [mm-at-serial-port.c:333] debug_log(): (ttyUSB0): --> 'AT+CGMR<CR>'
modem-manager[5955]: <debug> [1337321031.108098] [mm-at-serial-port.c:333] debug_log(): (ttyUSB0): <-- '<CR><LF>11.117.09.04.00<CR><LF><CR><LF>OK<CR><LF>'

11.117.09.04.00 might not be a recent enough firmware to work properly. Seems like the version on my own device: 11.310.16.03.149 might not be good enough or at least seems to crash and reconnect to USB frequently. There are newer revisions available on the Internet, I'll try to upgrade and see how well that works.

[Sleepy John (john-eyre)]

In response to Mathiew-tl's [QUOTE] Oneiric should have a similar patch applied from what is in Precise, if you have all the updates installed and your modem is working, then this tells me the patch itself is not the issue;[/QUOTE] I should like to make it clear that my E220 is still *not* working correctly in my fully updated Oneiric.

In order to log-in on my E220 on Oneiric, I am still following shrijith1's workarounds [QUOTE]1) Remove the default modemmanager and install modemmanager_0.4+git.20110124t203624.00b6cce-2ubuntu1
2) Insert the USB modem before booting up Ubuntu.[/QUOTE]

Therefore I don't see it as necessarily being conclusive that the patch itself is not the issue. The patch could still be the common issue on both Oneiric and Precise. How can I actually verify whether updates have applied the patch in my Oneiric or not? Simple instructions, please!

Any other Oneiric users with an E220 care to comment?

[Pablo (pjferra)]

I've just (10 minutes) updated my ubuntu 12.04 with the last actualizations. One of them updated modemmanager to 0.5.2 and promised to solve this bug (with an explicit reference).
Restarted my system and everything goes on as it was yesterday. My Huawei modem doesn't work with ubuntu 12.04. It just is not recognized by networkmanager. lsusb shows the huawei modem, but I can't see it to connect my computer to the web.
By the way, 2 months ago I was using 10.04 and it worked perfectly, and I've got another computer with 10.04 and goes on working. So is not a problem of the modem.
The fix released simply doesn't fix anything :(

[Pablo (pjferra)]

Sorry, my comment went away again by itself

[Pablo (pjferra)]

I'm an idiot. When I run lsusb I got the line
      Bus 002 Device 004: ID 12d1:1003 Huawei Technologies Co., Ltd. E220 HSDPA Modem / E230/E270/E870 HSDPA/HSUPA Modem
Due to this I've been all the time thinking I have an E220 Huawei modem, but watching the backside of the modem I've discovered that it is en E160 modem.

So, besides the issues are very similar I'm migrating to another bug thread: #982283

Mathieu T-L: I saw you've been there. Please help us too :)

[madbiologist (me-again)]

@Pablo - historically there has been a lot of confusion with the various Huawei model names. It usually better to trust the output of lsusb, particularly the ISB ID part (12d1:1003 in your case) rather than the labelling on the exterior of the device.

@Sleepy John and Pablo - Mathieu's comment #201 seems relevant given and old bug report - see comment #75 in bug 430011

[Pablo (pjferra)]

Hey!! I'm glad to tell you that I've been able to make my huawei modem work with Mobile Partner.
The system still freezes when I plug the dongle, but I've tried several times and allways got a connection although network manager didn't recognizes the device.
I'll go on trying, and I'll try upgrading the firmware. (thanks madbiologist for your suggestion)

[Pablo (pjferra)]

What should I to update the firmware?
Please, post steps and where to download it

[javierpedros (javierpedros)]

Hello:
I solved this problem but a way not elegant.
In the Application disc ulitity I umount and safely disassemble the unit CD Huawei.
One minute after the 3G modem is recognized properly.

It works for me.
I think the problem is the CD-Huawei not the usb mass storage.

I have Ubuntu 12.04 and Huawei E156G (I think this is like E220)

[Sleepy John (john-eyre)]

I would like to second Pablo's request for step by step instructions.

I know there are some workarounds that will allow me to log-in with Oneiric and my E220 so I'm not totally cut off from the internet, but I am still confused about this bug supposedly now being "fixed", and yet I still have to use these workarounds. If updates incorporating the fix have already been issued, then why aren't I seeing the benefit of them?

I would like to know if the updates that have been issued will automatically replace my deliberately reverted modem manager with a latest fixed default version, or if I should be trying to do that manually?

My aim is to have a standard fully updated Oneiric that works with an E220 without having to adopt workaround procedures. Is that a realistic aim, or am I expecting too much?

Regarding Mathieu's #201 comment about the firmware maybe not being recent enough, isn't this ignoring the point that our E220s were all working perfectly satisfactorily in Maverick and Natty, and it was only when Oneiric came along that all this trouble started?

If I thought that an upgrade to Precise would make my E220 log-in any easier, I would upgrade and forget about trying to fix Oneiric, but is anyone seeing evidence that Precise works better?

[madbiologist (me-again)]

The fixed modemmanager 0.5-1ubuntu1.1 package from oneiric-proposed which is described in comment #171 did make it into oneiric-updates. You can view this at https://launchpad.net/ubuntu/+source/modemmanager
Comment #171 also states that this fix is in modemmanager 0.5.2. modemmanager 0.5.2.0-0ubuntu2 was released in precise and is the current precise version - you can view this a the same link I gave above for the oneiric version. To check your own installation you can run dpkg -l <packagename> or look in Synaptic Package manager.

Now we just have to figure out what else is preventing the device from working normally for you. I am reluctant to take the time to provide further information given your apparent preconception that it won't work, but here it goes anyway. I can't fully explain why a firmware update would work given that the E220's worked in Maverick and Natty, other than that some of the workarounds here remind me of some of the workarounds used in the old bugs. Feel free to ignore my suggestions if you like.

I'm pretty sure that in comment #199, Mathieu meant to say "modem is not working" instead of "modem is working". He seems to have vanished since your reply though. Regarding your question about whether the updates will automatically replace the modemmanager_0.4+git.20110124t203624.00b6cce-2ubuntu1 which you downgraded to, that will depend on whether or not you used apt (or it's front-end Synaptic) to perform the downgrade or not, and whether you have locked the version or not. If necessary you can just use apt-get install (or Synaptic) to install the new version and then you will get any later update from then on. See above for how to check your current version.

Obviously if you didn't have the fixed package you should try with that before upgrading the firmware. The following steps for upgrading the firmware are taken from comments #15 and #16 in bug #449394

- it is highly recommended to work on a notebook or a PC with UPS, to be protected from power failure or you may loose your E220 dongle
- on a Windows based machine, download archive from http://www.maxis.com.my/maxisbb/download/E220SoftwareUpgrade.zip
- unpack ZIP file in a separate directory, like c:/tmp/E220
- plug-in E220 and wait about 1 minute
- close any software that may autorun when E220 is plugged
- start the exe file with Firmware inside the name and follow on-screen instructions - upgrade will take several minutes, be patient
- unplug your E220
- this is how I did it on a XP Windows!

- It may happen that deviation could appear, based on Windows you have installed. You have to deal with this by yourself .

Unfortunately there is the major inconvenience of needing a Windows or MacOS X installation/machine to perform the firmware upgrade. All of the other people who upgraded their modem's firmware seem to have used Windows to do so. I have looked hard on the internet for a Linux-based update tool, but it seems that neither Huawei nor their vendors (mobile communications companies) have released one.

Other than the download link provided in the steps above, other possible download locations are http://www.vodafone.de/hilfe-support/technischer-s...

[madbiologist (me-again)]

Actually, apt-cache policy <packagename> is better than dpkg -l <packagename> for checking what version you have installed. It doesn't truncate the package names the way that dpkg -l does. Sorry about that.

[Mathieu Trudel-Lapierre (cyphermox)]

@madbiologist: the point is certainly not to have everyone upgrade their firmware, but that this course of action needs testing regardless.

The problem is, the E220 responds differently based on different firmware versions, and we can only make work the ones that we have to test with.

If someone can test the modemmanager package from Quantal as well (using a live CD or installing the package directly) that could also be very useful.

I'm resetting the states for this bug to Triaged for the Oneiric, Precise and floating tasks; it's evidently not fixed and the state should have been reset long ago.

[Sleepy John (john-eyre)]

Let me first express my thanks and appreciation to madbiologist for his #211 and #212 review and comments.
Sorry to be so slow in responding; the truth is that I lost my focus on this topic while I had access to an alternative landline broadband, but now that I'm trying to get back on my E220, I'm having to catch up.

Second, let me apologise for the negative tone of what follows, since my E220 still isn't working properly.

OK, more of my details in-line below:

[QUOTE]The fixed modemmanager 0.5-1ubuntu1.1 package from oneiric-proposed
which is described in comment #171 did make it into oneiric-updates. You
can view this at https://launchpad.net/ubuntu/+source/modemmanager
<https://launchpad.net/ubuntu/+source/modemmanager>[/QUOTE]

  Yes, I had previously regressed to modemmanager_0.4, but
  when I check it now, it's showing as 0.5-1buntu1.1, so
  it must have updated itself at some stage without me noticing.

[QUOTE]Comment #171 also states that this fix is in modemmanager 0.5.2.
modemmanager 0.5.2.0-0ubuntu2 was released in precise and is the current
precise version -[/QUOTE]

   Some people (including myself) have remarked that Precise is
   suffering from exactly the same symptoms as Ocelot, so that
   leaves me a little skeptical about the effectiveness of these
   fixes. However the timing of these releases and whether they
   made it in time to be incorporated in Precise's Live issue,
   would be critical.

[QUOTE]Now we just have to figure out what else is preventing the device
from working normally for you. I am reluctant to take the time to provide
further information given your apparent preconception that it won't
work, but here it goes anyway. I can't fully explain why a firmware
update would work given that the E220's worked in Maverick and Natty,
other than that some of the workarounds here remind me of some of the
workarounds used in the old bugs. Feel free to ignore my suggestions if
you like.[/QUOTE]

    Many of the original complainants on this bug were seeing it
    as something new that was happening with Ocelot. Everything
    had been fine with their E220s in Maverick and Natty, so I
    have a kind of gut feeling that we could be barking up the
    wrong tree, but I don't understand enough about the actual
    fixes to be adamant about that. Reversion to modemmanager_0.4
    made it easier to connect, but it wasn't a total fix. There
    was and I believe there still is something else.
 .

[QUOTE] Regarding your question about whether the updates will automatically
replace the modemmanager_0.4+git.20110124t203624.00b6cce-2ubuntu1 which you
downgraded to, that will depend on whether or not you used apt (or it's
front-end Synaptic) to perform the downgrade or not, and whether you
have locked the version or not. If necessary you can just use apt-get
install (or Synaptic) to install the new version and then you will get
any later update from then on. See above for how to check your current
version.[/QUOTE]

    I'm not sure how or when it happened, but I'm on MM 0.5 -1ubuntu1.1
    now

[QUOTE]Obviously if you didn't h...

[Joakim Lagergren (jb-lager)]

Hi Sleepy John and others,

Since people still seem to have problems I thought I should chime in and share some of my experiences.

I had problems with connecting my E220 on 11.10 oneiric using MM 0.5-1ubuntu1 after upgrading the modem firmware to 11.117.10.02.00. (Don't remember what firmware version I had before the upgrade.)

I found a simple workaround that allowed me to connect reliably with ModemManager (detailed in duplicate bug #932824):
Prevent MM to probe /dev/ttyUSB1 by starting the command 'cat /dev/ttyUSB1' before MM starts its probing.

The updated version of MM currently in oneiric (which is supposed to fix this bug?), 0.5-1ubuntu1.1, does not work for me, but my workaround still works.

To test if you have the same problem as I have, you can try the following:
- Make sure the modem is physically disconnected.
- Open a terminal and run the command 'while true; do cat /dev/ttyUSB1; sleep 1; done'
(Before the modem is connected you will see messages like 'cat: /dev/ttyUSB1: No such file or directory' once every second. After the modem is connected and initialized you will see messages from the modem, such as ^BOOT:59866241,0,0,0,6 ^MODE:5,5 ^RSSI:27 and others.)
- Connect the modem to your computer.
- Connect using modem manager as usual.
Note: After ModemManager is done with its probing you can abort the cat /dev/ttyUSB1 command.

If this workaround works for you, you might want to try the "ports" branch of ModemManager which Aleksander Morgado mentioned in comment #145 (with more info available in the comments before and after). I've been using the ports branch since February and it has been working perfectly all the time.

[Aleksander Morgado (aleksander-m)]

IIRC, the MM_06 branch already comes with the previous 'ports' branch merged, you may want to try that one. git master also has the 'ports' functionality included, but lots of work to do there before being able to fully use it... :-/ help with the development welcome :-)

[Marius B. Kotsbak (mariusko)]

Maybe this patch will solve this problem?:

http://<email address hidden>/msg20874.html

[Sleepy John (john-eyre)]

Thanks Marius and Aleksander. I'd like to try this patch, but since I'm entirely dependent on my E220 and MM for my inernet access just now, I'm worried in case it doesn't work as planned and I can't get back on-line at all. What steps can I take to pre- save my existing MM and restore back to it in the event of problems?

[Marius B. Kotsbak (mariusko)]

You can just build a recent modemmanager (0.6 branch probably) with the patch and run its modem-manager before starting network-manager and you have restored your old configuration by just killing the new modem-manager.

[Aleksander Morgado (aleksander-m)]

No, that last patch is for the case where usbif 0 doesn't show up.

The problem in this bugreport is that the PCUI and modem ports get swapped when MM grabs them, and the proper fix for this should already be in MM_06 and git master.

[Marius B. Kotsbak (mariusko)]

Has anyone tested if this still is unresolved in Ubuntu 12.04?

[Sleepy John (john-eyre)]

Or is there an easy way I could download and test MM_06 on my Oneiric in such a way that I could revert to my MM_05 in case MM_6 failed and I completely lost my internet access as a result?

[madbiologist (me-again)]

modemmanager 0.6~git201206221719.8289a64-0ubuntu1 is in Ubuntu 12.10 "Quantal Quetzal" development repositories if you want to burn a LiveCD and test with that.

According to https://launchpad.net/ubuntu/+source/modemmanager there doesn't seem to be any modemmanager PPA's for Oneiric.

[Mathieu Trudel-Lapierre (cyphermox)]

Right there isn't. I can very well provide MM 0.6 in a PPA for precise, but I'm not sure it really qualifies so well for an official updates package, since it's multiple patches, and we don't know exactly which ones.

[Marius B. Kotsbak (mariusko)]

Mathieu: that would be very useful anyway for people that has problems that is solved upstream and are willing to risk stability for solving those problems, and to check if issues are solved upstream for future Ubuntu packages.

[Nikos Georgakis (ubuntu-ngtech.gr)]

I am using the following workaround sucessfully on two different 12.04 desktops & one linux Mint R13 laptop to connect with a huawei E220 Modem on 3G / 2G networks

#Running the following AS ROOT after the first unsuccesfulll connection attempt allows the modem to connect without problems until it is removed from the USB port:

   echo -e "AT+CNMI=2,1,0,2,0\r\nAT\r\n" > /dev/ttyUSB1

Without it I can't get a connection.

   I picked the commad from the C:\Program Files\Mobile Partner\Log\ATRecord.txt file on a Windows XP VM, after noticing that the modem would work properly in ubuntu after connecting first one time inside the XP VM, until the modem gets unplugged. Other AT command combinations don't have the same effect - it needs to be the exact above sequence.

[Marius B. Kotsbak (mariusko)]

There is a PPA available of MM_06, but it does not build. Seems like the patches that are upstream must be excluded from the build.

[Mark Fraser (launchpad-mfraz)]

I'm running Kubuntu 12.10 on a netbook and I'm still unable to connect using a E220 modem.

[Naresh(OpenERP) (nch-openerp)]

I have the same problem I am not able to connect my Huawei E1550 on latest ubuntu 12.10...Network manager detects perfectly the modem but when I click to connect its just says "Modem Disconnected" you are now offline.

[LinaPorras (linaporras)]

I have the same problem but with a Huawei E353, I have tried so many things, some of the errors are that is that the device is not dettected in TTY, also all the solutions that you have propposed didn't work....
I have Ubuntu 12.10 64Bit in a ASUS N56vj laptop, but I have another PC computer with Ubuntu 12.10 64 BIt and in there I don't have the problem... it's a little weird...

Sorry about the english... I Dont'speak it perfectly

[Sleepy John (john-eyre)]

No need to apologise for your English, KinaPorras, it's fine!
But sorry to read this trouble is still ongoing, now on an E353 as well as E220. I've been holding off on an upgrade to 12.04 from my 11.10 in the hope that a real solution to this bug will be found.
This whole problem started in 11.10. E220 worked fine in Natty, and it doesn't seem as if the basic change that was introduced in Oneiric and which is responsible for this failure has been adequately addressed in any of the fixes since then . All the new releases, it appears, are coming out with the same basic error that started with 11.10. I wonder if there's any chance that 12.04.2 (just out I believe) is compatible with E220 and E353. Anyone been able to test it??

[Marius B. Kotsbak (mariusko)]

You could just try it yourself using a live CD/USB stick. It would also ve interesting to know if a firmware upgrade would solve it.

[Aleksander Morgado (aleksander-m)]

I bought a E220 just for the sake of testing it; and it worked fine both with 0.6.x and git master.

[madbiologist (me-again)]

Thanks Aleksander. Which version of Ubuntu are you having success on? If 12.10, did you use the default shipped MM 0.6.x or a version you compiled yourself? What firmware version does your E220 have? Can you please post the output of lsusb -v

Do you boot with the E220 connected or disconnected?

[Aleksander Morgado (aleksander-m)]

I don't use Ubuntu; I'm using Fedora and a self-compiled MM 06/master.

[madbiologist (me-again)]

What version of Fedora? What firmware version does your E220 have? Can you please post the output of lsusb -v (I'm not sure if Fedora has such a command, but I assume there is something similar).

Do you boot with the E220 connected or disconnected?

[Aleksander Morgado (aleksander-m)]

My E220 has the following revision ID: 11.117.10.03.99

I usually test the modems plugging them, not with them already connected.

[Pablo (pjferra)]

Try this (it worked for me with ubuntu 12.04 64 bits):
1) using gedit create a text file with the following commands:
          modprobe -r option
          modprobe -r usb_wwan
          modprobe -r usbserial
          modprobe usbserial vendor=0x12d1 product=0x1003
2) save it anywhere (I've saved it as /opt/modem.sh)
3) asign execut permission
4) in a terminal type: sudo gedit /etc/rc.local
5) add the following lines at the end of the file
           /opt/modem.sh ## or the path and name you choosed for your file
           exit 0
6) save and exit
7) reboot
8) try your modem

[Sleepy John (john-eyre)]

Marius B K wrote [QUOTE]You could just try it yourself using a live CD/USB stick. It would also ve interesting to know if a firmware upgrade would solve it.[/QUOTE] Yes, reasonable suggestion, and I do now have a working live 12.04.2-desktop-i386 USB stick. but sorry, no valid data-enabled simcard at the moment. I only use my E220 for 3 or 4 months of the year when I'm travelling, but of course during those months it becomes an essential lifeline.

In post #234 madbiologist asks [QUOTE] What firmware version does your E220 have?[/QUOTE]. In June 2012 I tried an upgrade to 11.117.10.03.99, the same as Aleksander is now using, and I also changed to Maxis Broadband as recommended in post #211 but neither of these made any improvement.

Although the upgrade to modemmanager 0.5 and subsequent patches to it have made it a little easier to get a successful connection, and this applies particularly in cases where I've left my E220 and simcard plugged in between subsequent Oneiric boots, I'm still left with a gut feeling or hunch that there's still something basic that changed between Natty and Oneiric that we've missed. I don't know what it is, but it appears to have carried forward and still be causing the same problems in Precise and Quantal. If I plug my E220 in to an already booted Oneiric, or if I use my E220 or my simcard in some other device between Oneiric boots, then I can expect trouble and I have to go through some usb_modeswitch commands and some unplugging and re-plugging sequence before I can connect. My post #115 refers.

This discrepancy in how successful attempts to connect are, depending on the recent past history of my E220 and simcard, takes me back to Joshua Dietze's reply in post #76 saying he doesn't think there can be anything retained in flash memory when the device is powered down. My evidence points to the contrary; past history of what I've been doing with the device between Ocelot boots *does* affect my results. This is leading me to speculate that Ocelot can't be re-initialising the E220 (or maybe even the simcard??) to exactly the same state as Natty was. Maybe I should point out that I sometimes use my E220 to connect on a Windows partition between Ocelot boots, and at other times I remove the simcard from my E220 and use it quite independently in a smartphone. On the other hand if I leave my E220 and simcard plugged in and depowered between Ocelot boots, I've a much better chance of a successful connection next time.

Could we perhaps take a closer look at the differences between the early part of the re-initialisation sequence that Maverick and Natty used to perform, compared to what Ocelot and subsequents are now doing?

[Marius B. Kotsbak (mariusko)]

Aleksander, both Precise and Oneiric use MM 0.5. Could you please test that too? And also compare between 0.5.0 and 0.5.2, as that was the change between those distro releases:

http://packages.ubuntu.com/search?keywords=modemmanager&searchon=names&suite=all&section=all

[Marius B. Kotsbak (mariusko)]

Other people having issues, please post your firmware versions.

[Aleksander Morgado (aleksander-m)]

In 0.5.0 we introduced a Huawei-specific patch which ended up not being good enough for the E220; we reverted the patch for 0.5.2, though. See:

http://cgit.freedesktop.org/ModemManager/ModemManager/commit/?h=MM_05&id=7e406c27af9a3d2fb7266daa55d12cd0ba00f6fc

In 0.6.x, the Huawei-specific behaviour is handled by the new ports reorganization logic, same as in git master.

I'll try to gather older MM logs later.

[madbiologist (me-again)]

Sleepy John wrote:-

> Could we perhaps take a closer look at the differences between the early part of the re-initialisation sequence that Maverick and Natty used to perform, compared to what Ocelot and subsequents are now doing?

Sure. Can you please attach the log file/s from Natty and the log file/s from Oneiric after plugging in the modem? (See http://wiki.ubuntu.com/DebuggingNetworkManager) My apologies if you have provided this information before, but this bug is getting rather long and it would be easier if you could provide them again at this point.

> I'm still left with a gut feeling or hunch that there's still something basic that changed between Natty and Oneiric that we've missed. I don't know what it is, but it appears to have carried forward and still be causing the same problems in Precise and Quantal.

This has indeed been a protracted and frustrating issue. Many networking things were changed in Oneiric, some in the kernel, some in usb-modeswitch, some in modemmanger and some in network-manager. Reverting them en masse is likely to break other things. Some of them may not be able to be reverted individually/separately as they may be inter-related to others. However focusing on the past might be counterproductive given that the problem seems to be finally fixed in a newer version. More on that in a moment.

Is there any particular reason that you want to continue using Oneiric? Also, official Canonical support for Ubuntu 11.10 "Oneiric Ocelot" ends in April 2013 - see https://wiki.ubuntu.com/Releases

This problem might have been fixed in Precise by the patch reversion mentioned in comment #244. However comments #203, #226 and #240 say otherwise. It is almost certainly fixed in Quantal's modemmanager 0.6.0.0.really-0ubuntu1, as stated in comments #220 and #233. I realise that comments #228, #229 and #230 say otherwise, but two of them are different modems (E1550 and E353) and we don't have their USB product IDs to see if they are the same as the one that you, Pablo and Aleksander have (1003). The E220 in comment #228 is more of a concern but we don't have it's USB product ID either. It has been shown previously that some E220's are in fact a rebadged E156/E169/E1550 (I can't remember which one anymore).

Mark Fraser, Naresh and LinaPorras - can you please attach the output of lsusb -v

If we can get some more people testing E220s on Quantal that would be good. When doing so please attach the output of lsusb -v and the firmware version of your modem.

[Pablo (pjferra)]

Have tried adding

modprobe usbserial vendor=0x12d1 product=0x1003

to the /etc/rc.local file?

[Mark Robbie (mark-robbie)]

This problem has completely cleared up for me since installing 14.04 about two months ago. Previously the problem would be seen almost every day and I had to either run usb_modeswitch to reset the modem or bounce modem-manager, so two months without a problem is great.

Thanks to whomever finally fixed this.

[Rolf Leggewie (r0lf)]

oneiric has seen the end of its life and is no longer receiving any updates. Marking the oneiric task for this ticket as "Won't Fix".

[Steve Langasek (vorlon)]

The Precise Pangolin has reached end of life, so this bug will not be fixed for that release