Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken

Bug #554099 reported by Viktors Petrovs
478
This bug affects 84 people
Affects Status Importance Assigned to Milestone
Linux
Fix Released
Medium
Ubuntu
Fix Released
High
Unassigned
Lucid
Won't Fix
High
Unassigned
Maverick
Fix Released
High
noarz
gobi-loader (Ubuntu)
Fix Released
Undecided
Unassigned
Lucid
Invalid
Undecided
Unassigned
Maverick
Fix Released
Undecided
Unassigned
linux-backports-modules-2.6.32 (Ubuntu)
Invalid
Undecided
Unassigned
Lucid
Fix Released
Medium
Stefan Bader
Maverick
Invalid
Undecided
Unassigned

Bug Description

SRU justification (for LBM in Lucid):

Impact: Certain models of 3G hardware (Gobi) which apparently are quite
common, require several changes to work. First, the kernel module needs
to be updated, which requires larger modifications. Second, a new firm-
ware loader is required together with a matching udev rule.

Fix: In order to prevent regressions, the driver backport was done in
LBM and also adds a new binary subpackage (lbm-wwan). This package also
contains the loader and the udev rule.
NOTE! The drivers also requires firmware files which need to be manually
gathered as they are not redistributable.

Testcase: Without this backport no 3G connections are possible. Sevral
users in the bug report had successfully used the driver from the test
package after collecting the required firmware files.

---

Binary package hint: linux-image-generic

$ uname -a
Linux phoenix 2.6.32-19-generic #28-Ubuntu SMP Wed Mar 31 17:46:20 UTC 2010 i686 GNU/Linux

$ cat /proc/version_signature
Ubuntu 2.6.32-19.28-generic 2.6.32.10+drm33.1

$ lsusb
03f0:201d Hewlett-Packard <-- HP un2400 Mobile Broadband Module (QDL interface)

HP un2400 mobile broadband modem stopped working after i upgraded to lucid.
I can not load firmware anymore to my modem's Qualcomm Gobi USB chip using gobi_loader.
http://www.codon.org.uk/~mjg59/gobi_loader/

There was no problems on previous Ubuntu version (Karmic, kernel 2.6.31*). As i understand all devices based on this chip will not work on lucid, if this will not be fixed.

HP un2400 is recognized by kernel's qcserial module. qcserial.c was not changed from 2.6.31 to 2.6.32.
After searching internet i found that this is related to changes in the kernel's usbserial module:
"The qcserial driver is currently broken in 2.6.32 and later. It's due to the switch to using kfifo for usb serial, but we haven't been able to work out the actual cause. I'm looking at alternative approaches."
Source: http://mjg59.livejournal.com/120577.html

Contact me if You need additional information.

Revision history for this message
Jeremy Foshee (jeremyfoshee) wrote :

[This is an automated message. Apologies if it has reached you inappropriately.]

This bug was reported against the linux-meta package when it likely should have been reported against the linux package instead. We are automatically transitioning this to the linux kernel package so that the appropriate teams are notified and made aware of this issue.

If this bug really is a bug in the linux-meta package you can move it back to linux-meta and tag it kj-linux-meta, or contact us on the #ubuntu-kernel channel on the FreeNode IRC server. Thanks.

affects: linux-meta (Ubuntu) → linux (Ubuntu)
Revision history for this message
lordlaurin (lordlaurin) wrote :

same here ... please fix this in final lucid!

Revision history for this message
Tom Inglis (tominglis) wrote :

Please get the Qualcomm Gobi 2000 drivers working again in the release version of 10.04! It would be great to be able to use this modem in my Lenovo ThinkPad T510.

Revision history for this message
darko (darko2) wrote :

I have the same problem with HP EliteBook 6930p. It works if I activate it in Karmic and reboot to Lucid. But after shutdown gobi_loader hangs ... Please fix this. There are many HP laptops with UN2400!
Thank!

Darko

Changed in linux (Ubuntu):
status: New → Confirmed
Revision history for this message
Michał Walecki (michal-walecki) wrote :

I have the same problem with Lenovo ThinkPad T410. Please fix the problem in final release of Ubuntu Lucid Lynx...

Revision history for this message
Tom Inglis (tominglis) wrote :

In Ubuntu 9.10, it was possible to update the qcserial driver with the correct ID for your device, to compile and install gobi_loader 0.4, and copy the correct firmware files to /lib/firmware/gobi. There were issues with coming out of suspension / hibernation, but with a regular boot this worked.

This method no longer works in Ubuntu 10.04, presumably because of the changes mentioned above in the qcserial driver et al.

It would be great if the Gobi and Gobi 2000 devices were supported natively in the kernel by a combined driver / firmware loader with the unlocked firmware as the default (even if these have to be "restricted").

Revision history for this message
Motin (motin) wrote :

Seems like this netbook (Compaq Mini 110c) is not going to meet Lucid anytime soon... :(

Awaiting meerkat...

Revision history for this message
John Smith (jn-smith) wrote :

Same problem with Compaq mini 733ez. Any news?

Revision history for this message
AmenophisIII (amenophisiii) wrote :

there are kernel patches available. see http://thread.gmane.org/gmane.linux.usb.general/29707

Revision history for this message
Jesse Palser (jessepalser) wrote :

I have this problem too on my netbook. The only method for me to use my Qualcomm 9212 3G modem is to load the firmware using Gobi Loader. This worked OK on Ubuntu 9.10, but it does not work on 10.04. Please fix this before 10.04 release or I will have no method to use my 3G modem in my netbook. Thanks.

Revision history for this message
Viktors Petrovs (viks77) wrote :

Thank You, AmenophisIII!

I used patch parts 1/4 & 2/4 to compile only qcserial.ko & usb-wwan.ko for current kernel (2.6.32-21-generic) and now my HP un2400 Mobile Broadband Module is working again!!
But beware there are some errors in the patch! usb-wwan files sometimes is referenced as usb-wwan and sometimes as usb_wwan ! Must be changed everywhere uniformly!

Revision history for this message
Paul Stadig (pjstadig) wrote :

It's been too long since I did a kernel compile. I installed the linux-source package, expanded it, configured the kernel (I just accepted the defaults), patched with the patch from http://www.codon.org.uk/~mjg59/gobi_loader/kernel_patches/, compiled, and copied options.ko, qcserial.ko, and usb_wwan.ko to /lib/modules/`uname -r`/kernel/drivers/usb/serial, and when I do `sudo modprobe qcserial` I get an error about unknown symbols from usb_wwan. I tried `sudo modprobe usb_wwan` but I get an error that usb_wwan is an unknown module. What steps am I missing? Could someone kindly post step-by-step instructions?

Revision history for this message
Viktors Petrovs (viks77) wrote :

After You copied modules try:
sudo /sbin/depmod -a

Revision history for this message
Viktors Petrovs (viks77) wrote :

Sometimes i need to compile just one kernel module for current kernel without compiling full kernel tree.
This is how i get it done:

1) get attached patch
2) install and extract kernel source

sudo apt-get install linux-source
tar xjf /usr/src/linux-source-2.6.32.tar.bz2

3) patching and compilation (only drivers/usb/serial subdirectory)

mkdir temp

cd ./linux-source-2.6.32
patch -p1 < ../usb-wwan-2.6.32.diff

cp /lib/modules/`uname -r`/build/.config ../temp
cp /lib/modules/`uname -r`/build/Module.symvers ../temp
cp /lib/modules/`uname -r`/build/Makefile .

make O=../temp outputmakefile
make O=../temp archprepare
make O=../temp prepare
make O=../temp modules SUBDIRS=scripts
make O=../temp modules SUBDIRS=drivers/usb/serial/

4) copy resulting modules to kernel modules directory

sudo cp ../temp/drivers/usb/serial/usb_wwan.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo cp ../temp/drivers/usb/serial/qcserial.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo cp ../temp/drivers/usb/serial/option.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo /sbin/depmod -a

Revision history for this message
Viktors Petrovs (viks77) wrote :

There is patch i forgot to add to previous post!

Revision history for this message
Tom Inglis (tominglis) wrote :

Is there a fix for this issue that enables ordinary users of Ubuntu to get their Qualcomm Gobi 2000 chip working without having to compile a new kernel, patch it, install a firmware loader, and copy the correct firmware from their Windows installation!?

Revision history for this message
Paul Stadig (pjstadig) wrote :

Thanks you very much, Viktors! I was missing the depmod step, and by following your instructions I was able to build the modules without having to build the entire kernel, which saved a bunch of time.

I agree with, Tom, even though there is a workaround, this should be fixed, especially for users who may not be able to build a kernel.

Revision history for this message
Tom Inglis (tominglis) wrote :

If we can get this chip working for the ordinary user, without them having to do anything extra, then that would be fantastic, particularly if it can be patched into 10.04?

Otherwise, the least we could do is to provide a really detailed step by step explanation of how to get it to work on the Ubuntu wiki?

These chips are probably going to get a lot more popular over the next year or two, as people use 3G / 4G / GPS more in their laptops...

Revision history for this message
Nick vd Merwe (nickvdmer) wrote :

I am running the Released Lucid and it was not fixed. I have a HP Elitebook 8530w.
Thus I have a Not-Functioning 3G. To access the Internet I have to borrow the wife's Huawei Vodafone card.

Please can someone fix the kernel and make it available through Update-Manager?

Changed in linux (Ubuntu):
status: Confirmed → Triaged
importance: Undecided → Medium
tags: added: patch
Revision history for this message
Boris Bopp (boris.bopp) wrote :

It should be possible to create a package for ordinary users, if there is a patched Kernel.

The steps are:
1. Download firmware. I don't know the legal situation.
wget ftp://ftp.hp.com/pub/softpaq/sp44501-45000/sp44916.exe
2. cabextract 44916.exe
3. Dialog that asks which firmware to use (http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&objectID=c01738839&jumpid=reg_R1002_USEN)
4. move firmware to /lib/firmware/gobi
5. install gobi_loader
6. insert qcserial and hp-wmi into /etc/modules.

A rfkill power switch somewhere would be nice , too.

Revision history for this message
Tom Inglis (tominglis) wrote :

NB - This is not just a problem for HP laptops, it affects anything with the Qualcomm Gobi series of chips (including in my case the Lenovo ThinkPad series).

Revision history for this message
Paul Stadig (pjstadig) wrote :

Thanks, Tom. I was actually going to post the same thing. I have a Lenovo ThinkPad. It's anyone with the Gobi 2000, not just HP.

summary: - HP un2400 mobile broadband module not working after upgrade to lucid
+ qualcomm gobi (2000) mobile broadband module not working after upgrade
+ to lucid
Revision history for this message
Fume (fume) wrote : Re: qualcomm gobi (2000) mobile broadband module not working after upgrade to lucid

The Problem is also on Acer Aspire one with Qualcomm Gobi modem. I think very much Laptops use this
Modem so please try to fix it.

Revision history for this message
jay (htl-rudi) wrote :

I have the same problem (HP 6930p; Gobi 1000)
Patch worked for me without any problems.

Revision history for this message
Danie Thom (danie-thom-gmail) wrote :

I have the same problem with the Fujitsu Lifebook S710. They ship with the Sierra Wireless modules using the Qualcomm firmware. I received the following SDK file that was supposed to be the driver for the card, but it is only the SDK. Can anyone help with a How To that can help me with the installation? I have never done this, so I don't know where to start.?field.comment=I have the same problem with the Fujitsu Lifebook S710. They ship with the Sierra Wireless modules using the Qualcomm firmware. I received the following SDK file that was supposed to be the driver for the card, but it is only the SDK. Can anyone help with a How To that can help me with the installation? I have never done this, so I don't know where to start.

Revision history for this message
Tapani Tarvainen (ubuntu-tapani) wrote :

Same issue with HP ProBook 6440B, but patch worked without problems.

description: updated
Revision history for this message
tnf (tnf-ars-solvendi) wrote :

I tried to follow the instructions on comments 14 and 20, but unfortunately this seems to work only partly. I get the following in /var/log/messages:

Jun 1 08:24:33 tptfeubu kernel: [ 4.805522] USB Serial support registered for Qualcomm USB modem
Jun 1 08:24:33 tptfeubu kernel: [ 4.805771] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
Jun 1 08:24:33 tptfeubu kernel: [ 4.805958] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
Jun 1 08:24:33 tptfeubu kernel: [ 4.806006] usbcore: registered new interface driver qcserial

I also have /dev/ttyUSB0, but NetworkManager and umtsmon don't recognize the device. lsusb shows

Bus 002 Device 004: ID 05c6:9204 Qualcomm, Inc.

Any ideas where I might start looking?

Revision history for this message
Viktors Petrovs (viks77) wrote :

Hello, tnf!
05c6:9204 is a Gobi 2000 QDL device. You need to download firmware.

http://www.codon.org.uk/~mjg59/gobi_loader/download/

Revision history for this message
tnf (tnf-ars-solvendi) wrote :

I installed gobi_loader and put the firmware in /lib/gobi/formware as suggested in comment 20. This seemed to work. Anything I might have missed? I did a make and make install, nothing else

Revision history for this message
Tom Inglis (tominglis) wrote :

Is there any chance of making a package for ordinary users so that they can grab the firmware and the package and get their modem working?

If not, is it possible to create some instructions for ordinary users to explain how to get these modems working in the Ubuntu wiki?

Will the 10.04 kernel be patched with this update, or will this only be something that makes its way into Maverick?

Revision history for this message
tomx (tom-xitio) wrote :

I'd be grateful for some clarification, too.
So, the hardware is used in 2 different aspects (9204/9205 - QDL/Modem Device) AND the qcserial has to be compiled manually?
I use a Thinkpad T410s on Ubuntu 10.04 2.6.32-22-generic.

lsusb gives
Bus 002 Device 003: ID 05c6:9204 Qualcomm, Inc.
and no ttyUSB0 shows up in /dev.

Revision history for this message
tnf (tnf-ars-solvendi) wrote :

@Viktors: thanks for the hint, I didn't realize I had the wrong firmware

Revision history for this message
tnf (tnf-ars-solvendi) wrote :
Download full text (3.3 KiB)

@Tom Inglis and tomx: following are the steps I did to get the device working. I didn't create a wiki page because I couldn't find the right place and ubuntuguide.org doesn't let me edit. And maybe some day this bug will be fixed and getting the device to work will be easier :)

## Download the firmware and copy it to the right place
I have a ThinkPad T510, so I downloaded the driver from http://www-307.ibm.com/pc/support/site.wss/document.do?lndocid=MIGR-72938 and installed it on Windows. Then I copied the files from c:\Programme\QUALCOMM\Images\Lenovo\6 and c:\Programme\QUALCOMM\Images\Lenovo\UMTS to /lib/firmware/gobi (you'll have to create that folder). Check http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&objectID=c01738839&jumpid=reg_R1002_USEN to find out which folders are right for you

## Install the kernel patch and gobi_loader
Open a terminal
mkdir umts
cd umts
wget http://launchpadlibrarian.net/46353288/usb-wwan-2.6.32.diff
sudo apt-get install linux-source
tar xjf /usr/src/linux-source-2.6.32.tar.bz2
mkdir temp
cd ./linux-source-2.6.32
patch -p1 < ../usb-wwan-2.6.32.diff
cp /lib/modules/`uname -r`/build/.config ../temp
cp /lib/modules/`uname -r`/build/Module.symvers ../temp
cp /lib/modules/`uname -r`/build/Makefile .
make O=../temp outputmakefile
make O=../temp archprepare
make O=../temp prepare
make O=../temp modules SUBDIRS=scripts
make O=../temp modules SUBDIRS=drivers/usb/serial/
sudo cp ../temp/drivers/usb/serial/usb_wwan.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo cp ../temp/drivers/usb/serial/qcserial.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo cp ../temp/drivers/usb/serial/option.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo /sbin/depmod -a
cd ..
wget http://www.codon.org.uk/~mjg59/gobi_loader/download/gobi_loader-0.5.tar.gz
tar -xzf gobi_loader-0.5.tar.gz
cd gobi_loader-0.5/
make
sudo make install
cd ../..
rm -r umts

Now reboot the system and the NetworkManager applet should have an "activate mobile broadband" option
lsusb shows "Bus 002 Device 004: ID 05c6:9205 Qualcomm, Inc."
In /var/log/messages I see
Jun 2 09:08:04 tptfeubu kernel: [ 4.324342] USB Serial support registered for Qualcomm USB modem
Jun 2 09:08:04 tptfeubu kernel: [ 4.324378] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
Jun 2 09:08:04 tptfeubu kernel: [ 4.327814] ip_tables: (C) 2000-2006 Netfilter Core Team
Jun 2 09:08:04 tptfeubu kernel: [ 4.328086] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
Jun 2 09:08:04 tptfeubu kernel: [ 4.328105] usbcore: registered new interface driver qcserial
and later
Jun 2 09:08:15 tptfeubu kernel: [ 14.546188] usb 2-1.4: USB disconnect, address 3
Jun 2 09:08:15 tptfeubu kernel: [ 14.546406] qcserial ttyUSB0: Qualcomm USB modem converter now disconnected from ttyUSB0
Jun 2 09:08:15 tptfeubu kernel: [ 14.546414] qcserial 2-1.4:1.1: device disconnected
Jun 2 09:08:15 tptfeubu kernel: [ 14.777742] usb 2-1.4: new high speed USB device using ehci_hcd and address 4
Jun 2 09:08:15 tptfeubu kernel: [ 14.897398] usb 2-1.4: configuration #1 chosen from 1 choice
Jun 2 09:08:15 tptfeubu kernel: [ ...

Read more...

Revision history for this message
Tom Inglis (tominglis) wrote :

TNF that's great thank you very much. I'll try that now and will look into creating a wiki page.

Is this something that needs to be done each time the kernel is updated, and will the USB-WWAN patch be included in a future kernel update in 10.04?

Revision history for this message
Michał Walecki (michal-walecki) wrote :

I have Lenovo Thinkpad T410 with Gobi 2000 and I used tnf's instructions. However it does not work completely, because I have to run Windows first and next Ubuntu and then I see 05c6:9205 Qualcomm, Inc. in lsusb. If not I run Windows first I see 05c6:9204 Qualcomm, Inc.

How to run GPS module of Gobi 2000?

Revision history for this message
Martijn vdS (martijn) wrote :

My Gobi 2000 works fine with 2.3.35-rc1 mainline (from the kernel-team PPA) + the gobi-loader tool from http://www.codon.org.uk/~mjg59/gobi_loader/ + the firmware I copied from my Windows partition.

In 2.3.34 /dev/ttyUSB0 doesn't show up, so firmware uploading can't be done.

Revision history for this message
tnf (tnf-ars-solvendi) wrote :

@Tom Inglis: not sure if this will be the same for other updates, but it still works for me after the last 2.6.32-22 fix. I guess it won't work after a 2.6.32-23 update, but I don't now for sure

Revision history for this message
Marco (bulletxt) wrote :

Hi Martijn vdS, where can I get that kernel 2.6.35RC1? Did you compile it yourself or did you get it from PPA? If from PPA, which one? The only one I can find for Ubuntu 10.04 is a kernel ppa with linux kernel 2.6.34! Thanks

Revision history for this message
Tobias Bradtke (webwurst) wrote :
Revision history for this message
Marco (bulletxt) wrote :

Thanks. About the gobi loader, which version did you use? Version 0.6? Also, did you apply a patch to it?

Revision history for this message
Brian Candler (b-candler) wrote :

On my thinkpad X100e with gobi 2000 and Lucid amd64, it works with either mainline linux-image-2.6.35-020635rc1-generic,
or 2.6.32-22-generic patched as above, plus gobi-loader 0.6

However, even though gobi-loader installed a udev rules file (/lib/udev/rules.d/60-gobi.rules), the firmware isn't auto-loaded. I have to do the following:

* /lib/udev/gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi
* run lsusb until device id changes from 05c6:9204 to 05c6:9205
* sudo killall modem-manager

After that, network-manager detects it fine, and it can connect.

Revision history for this message
Simon Hirscher (codethief) wrote :

On my Thinkpad W510 I just installed http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.35-rc3-maverick/, downloaded the gobi_loader and copied the firmware from C:\Program Files (x86)\QUALCOMM\Images\Lenovo\UMTS to /lib/firmware/gobi and it magically worked after a reboot.
Two things I noticed:
- lsusb displayed "Bus 002 Device 003: ID 05c6:9205 Qualcomm, Inc." before and after the installation / reboot.
- I received a hell of a lot of error messages during start up though everything seems to work.

Thank you guys!

Revision history for this message
Simon Hirscher (codethief) wrote :

I was wrong - after another reboot it doesn't work anymore. Surprisingly, Brian Candler's solution doesn't work for me either:

$ sudo /lib/udev/gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi
Failed to open tertiary firmware: : No such file or directory
usage: /lib/udev/gobi_loader [-2000] serial_device firmware_dir

According to the source code the file "uqcn.mbn" is missing. But there are only 2 files in the C:\Program Files (x86)\QUALCOMM\Images\Lenovo\UMTS directory, namely amss.mbn and apps.mbn. Anyone here who knows why it was working with only those two files after the first reboot but isn't anymore?

Revision history for this message
Simon Hirscher (codethief) wrote :

For the record: I found the "uqcn.mbn" file in several of the numbered directories in C:\Program Files (x86)\QUALCOMM\Images\Lenovo\ and took the one which appeared the most up-to-date (that is, the one with the latest last-modified date). Rebooted. Works! (Without having to explicitly load the firmware after start up.)
I still don't know why it had been working with only two firmware files after the first reboot but as long as it keeps working with all three of them I think I won't investigate this further.

Revision history for this message
Brian Candler (b-candler) wrote :

It's important to have the three firmware files, and choose the ones with the right md5 checksums (as there were various versions in the Windows drivers under C:\Program Files\QUALCOMM\Images\Lenovo\*)

I used the md5sums from here:
http://www.thinkwiki.org/wiki/Qualcomm_Gobi_2000

$ md5sum /lib/firmware/gobi/*
84d002b0ef003cde6c95826bfbf067fe /lib/firmware/gobi/amss.mbn
d7496085f1af3d1bfdf0fa60c3222766 /lib/firmware/gobi/apps.mbn
1aa5727b034dd1f371a3412d5800c1a3 /lib/firmware/gobi/UQCN.mbn

They work fine with Vodafone UK.

Revision history for this message
Simon Hirscher (codethief) wrote :
Download full text (5.3 KiB)

Apparently, I was wrong - it doesn't work flawlessly as I thought with kernel http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.35-rc3-maverick/. During video playback I noticed several hiccups some of which were more or less severe and caused the system to finally hang.
/var/log/syslog gave me the following:

Jun 24 04:16:46 philosoph modem-manager: last message repeated 2 times
Jun 24 04:16:46 philosoph modem-manager: (ttyUSB0) closing serial device...
Jun 24 04:16:46 philosoph modem-manager: (Gobi): GSM modem /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.4 claimed port ttyUSB0
Jun 24 04:16:46 philosoph modem-manager: Added modem /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.4
Jun 24 04:16:46 philosoph modem-manager: Exported modem /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.4 as /org/freedesktop/ModemManager/Modems/35
Jun 24 04:16:46 philosoph NetworkManager: <info> (ttyUSB0): new GSM device (driver: 'qcserial')
Jun 24 04:16:46 philosoph NetworkManager: <info> (ttyUSB0): exported as /org/freedesktop/NetworkManager/Devices/411
Jun 24 04:16:46 philosoph NetworkManager: <info> (ttyUSB0): now managed
Jun 24 04:16:46 philosoph NetworkManager: <info> (ttyUSB0): device state change: 1 -> 2 (reason 2)
Jun 24 04:16:46 philosoph NetworkManager: <info> (ttyUSB0): deactivating device (reason: 2).
Jun 24 04:16:46 philosoph NetworkManager: <info> (ttyUSB0): device state change: 2 -> 3 (reason 0)
Jun 24 04:16:52 philosoph modem-manager: Removed modem /sys/devices/pci0000:00/0000:00:1d.0/usb2/2-1/2-1.4
Jun 24 04:16:52 philosoph NetworkManager: <info> (ttyUSB0): now unmanaged
Jun 24 04:16:52 philosoph NetworkManager: <info> (ttyUSB0): device state change: 3 -> 1 (reason 36)
Jun 24 04:16:52 philosoph NetworkManager: <info> (ttyUSB0): cleaning up...
Jun 24 04:16:52 philosoph NetworkManager: <info> (ttyUSB0): taking down device.
Jun 24 04:16:52 philosoph kernel: [ 5974.643537] usb 2-1.4: USB disconnect, address 79
Jun 24 04:16:52 philosoph kernel: [ 5974.643892] qcserial ttyUSB0: Qualcomm USB modem converter now disconnected from ttyUSB0
Jun 24 04:16:52 philosoph kernel: [ 5974.643929] qcserial 2-1.4:1.2: device disconnected
Jun 24 04:16:52 philosoph kernel: [ 5974.878313] usb 2-1.4: new high speed USB device using ehci_hcd and address 80
Jun 24 04:16:52 philosoph kernel: [ 5974.990211] usb 2-1.4: config 1 has an invalid interface number: 1 but max is 0
Jun 24 04:16:52 philosoph kernel: [ 5974.990218] usb 2-1.4: config 1 has no interface number 0
Jun 24 04:16:52 philosoph kernel: [ 5974.993614] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
Jun 24 04:16:52 philosoph kernel: [ 5974.993722] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
Jun 24 04:16:53 philosoph modem-manager: (ttyUSB0) opening serial device...
Jun 24 04:16:53 philosoph modem-manager: (ttyUSB0): probe requested by plugin 'Gobi'
Jun 24 04:16:57 philosoph modem-manager: (ttyUSB0) closing serial device...
Jun 24 04:16:57 philosoph kernel: [ 5980.014722] usb 2-1.4: USB disconnect, address 80
Jun 24 04:16:57 philosoph kernel: [ 5980.014929] qcserial ttyUSB0: Qualcomm USB modem converter now disconnected from ttyUSB0
Jun 24 04:16:57 philosoph kern...

Read more...

Revision history for this message
Brian Candler (b-candler) wrote :

I haven't tried video streaming over 3G as it would be too expensive on my PAYG SIM. But for basic browsing and E-mail it works fine; it hasn't locked up the whole machine.

However I'm using the standard Lucid kernel with self-compiled modules for qcserial and thinkpad_acpi, not the mainline kernel.

Have you checked the output of
    md5sum /lib/firmware/gobi/*
?

Revision history for this message
Simon Hirscher (codethief) wrote :

I'm sorry, my previous post was misleading. I wasn't talking about video streaming but simple video playback (thus, with the video file already on my hard drive) which tended to lag and freeze. Constantly adding and removing the 3G device therefore seemed to have kind of an impact on the overall system performance.

Didn't you mention earlier that you also tried out the 2.6.35-rc1 mainline kernel? In case you still have it installed could you please check if you can confirm my issues? (As I'm experiencing them on both, rc1 and rc3.)

Anyway, those are my md5 checksums:

$ md5sum /lib/firmware/gobi/*
84d002b0ef003cde6c95826bfbf067fe /lib/firmware/gobi/amss.mbn
d7496085f1af3d1bfdf0fa60c3222766 /lib/firmware/gobi/apps.mbn
68e46b936d5400862a0381b101af3972 /lib/firmware/gobi/UQCN.mbn

I don't even have the UQCN.mbn version you are using in one of my C:\Program Files\QUALCOMM\Images\Lenovo\* subdirectories... :\

$ find -type f -exec md5sum {} \; | sort -k 2
918ecc998ca0b151f01111ff8cca39d2 ./05c69205.did
c3d6fd93ae2e52775ef9cd8fccbc20be ./0/UQCN.mbn
4d6203bf9fe8ae1af439d4d163e91596 ./12/amss.mbn
385a22740f80c0d00f8acdd9ad637032 ./12/apps.mbn
e868df00bfa88596d588a52f872ff703 ./12/UQCN.mbn
48191d62dc4091fb9d23e0b947611c58 ./1/amss.mbn
4f46a856fcceb197943d0cf3257c3621 ./1/apps.mbn
bb05d912e27b822348d4de93fc3ac911 ./1/UQCN.mbn
ff935ec4da6f6711b48caaddcf737566 ./2/UQCN.mbn
4f74d79ed8df027db3b75e36feb2bd27 ./3/amss.mbn
d9ff7b27c491332650b1efbbee64ee31 ./3/apps.mbn
6f6462fe0cc185cc3dc24720a28770fb ./3/UQCN.mbn
46418504e70971bda7b2ee34cc80b5b7 ./4/UQCN.mbn
68e46b936d5400862a0381b101af3972 ./6/UQCN.mbn
9f6773625d59098f9360c24263d99615 ./7/UQCN.mbn
e012286117c0c7708369f8dbc7906e64 ./8/UQCN.mbn
c65246da47695b4ecaaf098a02718c9d ./9/UQCN.mbn
84d002b0ef003cde6c95826bfbf067fe ./UMTS/amss.mbn
d7496085f1af3d1bfdf0fa60c3222766 ./UMTS/apps.mbn

According to http://www.thinkwiki.org/wiki/Talk:Qualcomm_Gobi_2000 the 7/UQCN.mbn is for Telefónica (which o2 belongs to) so I guess I'm going to try that, now.

Revision history for this message
Brian Candler (b-candler) wrote :

For comparison, here's what I have:

# cd '/mnt/Program Files/QUALCOMM/Images/Lenovo' && find . -name '*.mbn' -print0 | xargs -0 md5sum
c3d6fd93ae2e52775ef9cd8fccbc20be ./0/UQCN.mbn
48191d62dc4091fb9d23e0b947611c58 ./1/amss.mbn
4f46a856fcceb197943d0cf3257c3621 ./1/apps.mbn
bb05d912e27b822348d4de93fc3ac911 ./1/UQCN.mbn
4d6203bf9fe8ae1af439d4d163e91596 ./12/amss.mbn
385a22740f80c0d00f8acdd9ad637032 ./12/apps.mbn
e868df00bfa88596d588a52f872ff703 ./12/UQCN.mbn
ff935ec4da6f6711b48caaddcf737566 ./2/UQCN.mbn
4f74d79ed8df027db3b75e36feb2bd27 ./3/amss.mbn
d9ff7b27c491332650b1efbbee64ee31 ./3/apps.mbn
6f6462fe0cc185cc3dc24720a28770fb ./3/UQCN.mbn
46418504e70971bda7b2ee34cc80b5b7 ./4/UQCN.mbn
1aa5727b034dd1f371a3412d5800c1a3 ./6/UQCN.mbn
9f6773625d59098f9360c24263d99615 ./7/UQCN.mbn
e012286117c0c7708369f8dbc7906e64 ./8/UQCN.mbn
c65246da47695b4ecaaf098a02718c9d ./9/UQCN.mbn
84d002b0ef003cde6c95826bfbf067fe ./UMTS/amss.mbn
d7496085f1af3d1bfdf0fa60c3222766 ./UMTS/apps.mbn

I'm not sure I'm going to have time to try crashing kernels for a while. I've checked my logs and don't see the repeated disconnections and reconnections that you see. Having said that, I'm only using 3G when I have no ethernet or wifi, which isn't very often.

Revision history for this message
Simon Hirscher (codethief) wrote :

Strangely enough, now, none of the different firmware versions produces the "error loop" anymore which I described above. Instead, the device is always available in the network connections manager with every firmware version but trying to connect results in the following /var/log/syslog entries:

Jun 24 18:42:53 philosoph NetworkManager: <info> Activation (ttyUSB0) starting connection 'o2 (Mobile Flat L)'
Jun 24 18:42:53 philosoph NetworkManager: <info> (ttyUSB0): device state change: 3 -> 4 (reason 0)
Jun 24 18:42:53 philosoph NetworkManager: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) scheduled...
Jun 24 18:42:53 philosoph NetworkManager: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) started...
Jun 24 18:42:53 philosoph NetworkManager: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) complete.
Jun 24 18:42:53 philosoph modem-manager: (ttyUSB0) opening serial device...
Jun 24 18:42:53 philosoph modem-manager: Modem /org/freedesktop/ModemManager/Modems/0: state changed (disabled -> enabling)
Jun 24 18:42:54 philosoph modem-manager: Got failure code 100: Unknown error
Jun 24 18:42:54 philosoph modem-manager: Modem /org/freedesktop/ModemManager/Modems/0: state changed (enabling -> enabled)
Jun 24 18:43:54 philosoph NetworkManager: <WARN> stage1_prepare_done(): GSM modem connection failed: (32) Network timeout
Jun 24 18:43:54 philosoph NetworkManager: <info> (ttyUSB0): device state change: 4 -> 9 (reason 1)
Jun 24 18:43:54 philosoph NetworkManager: <info> Marking connection 'o2 (Mobile Flat L)' invalid.
Jun 24 18:43:54 philosoph NetworkManager: <info> Activation (ttyUSB0) failed.
Jun 24 18:43:54 philosoph NetworkManager: <info> (ttyUSB0): device state change: 9 -> 3 (reason 0)
Jun 24 18:43:54 philosoph NetworkManager: <info> (ttyUSB0): deactivating device (reason: 0).
Jun 24 18:43:54 philosoph NetworkManager: <info> Policy set 'Auto keksdose' (wlan0) as default for routing and DNS.

I don't get it. What did I do to change the behavior which occured previously? Any ideas on how to "reset" it?

Revision history for this message
Simon Hirscher (codethief) wrote :

I'm about to give up. I've tried to get it working, now, with the kernel patch (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/554099/comments/33), and the 2.6.35 rc1 and rc3 mainline kernels. The result stays the same:
The device is either constantly being added and removed, which results in /var/log/syslog getting several MiB big (see my previous post: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/554099/comments/46), or it is recognized and appears stable but it's impossible to connect (see https://bugs.launchpad.net/ubuntu/+source/linux/+bug/554099/comments/50). Which one of the two results occur seems to be rather random. In my case it changed every few reboots or so.

Revision history for this message
Motin (motin) wrote :

Thanks tnf for the summary.

Don't forget this part:
http://sanitarium.se/wiki/HP_un2400_3G_modem_in_Linux#Holding_back_updates_for_the_kernel

Holding back updates for the kernel

If you've installed a patched kernel you may not want to recompile every time Ubuntu updates the kernel...

# echo "linux-image-2.6.32-21-generic hold" | dpkg --set-selections

Thanks all for finding the work-arounds necessary, I finally could upgrade this Compaq Mini 110c to Lucid :)

Revision history for this message
Brian (x-brian) wrote :

I got this to work on my Dell E6410 with the "Dell Wireless 5620" Gobi 2000 over Verizon Wireless.

I found a pretty good walk-through I could adapt to Dell/Verizon at:
http://www.nulldevice.de/2010/06/wwan-lenovo-thinkpad-x100e-ubuntu-10-04/

I updated to the 2.6.35rc1 kernel, and built gobi_loader. The Dell device is in qcserial based on extrapolation from:
http://cateee.net/lkddb/web-lkddb/USB_SERIAL_QUALCOMM.html
(Dell device is: ID 413c:8186, or 413c:8185 before firmware is loaded.)

I got gobi_loader working with what I thought were the Verizon firmware files, but after the card was activated with Verizon it still would not connect.

Finally I imaged my primary partition to my /home partition and wiped it, installed Windows 7 64b and the Dell drivers, then "activated" the card. It then connected in Windows fine. (I think the process of "activating" the card in Windows is just choosing the correct firmware to load.) From my Windows 7 install, I copied the firmware files to a USB stick and restored my Ubuntu installation. A warm reboot resulted in the card working from Ubuntu. However, a cold reboot and the card stopped working, and the Gobi 2000 acted like it did before running Windows. I pulled the firmware files off of the memory stick and looked at the md5sum values for the files. The firmware I got from the Dell install was different from the Lenovo firmware. So I copied the Dell firmware for Verizon (from the 1/ directory) to /lib/firmware/gobi/ and did a cold reboot. Now the Gobi 2000 is working!

Here are the md5sum values for the "Dell Wireless 5620" firmware:

$ find -name "*.mbn" -exec md5sum {} \; | sort -k 2
1aa40f2f7e790243e1a0e8c6145c9d12 ./0/UQCN.mbn
9b0526b4f1cd016e9c2103db63536e4b ./15/amss.mbn
c28c5f749e5481b472f6cfb7c04806cb ./15/apps.mbn
68d2f4f631d2c72957d15a8b23fed3be ./15/UQCN.mbn
51aa00c43b7eab331e6da1a422e37ff3 ./1/amss.mbn
4f46a856fcceb197943d0cf3257c3621 ./1/apps.mbn
6e05582f7997f9385e0edab431a4814c ./1/UQCN.mbn
820e2ad056bce0450b9769e9ca7635ea ./2/UQCN.mbn
86d9d0287b5fd8c0f1dae329c36be3c5 ./3/amss.mbn
d9ff7b27c491332650b1efbbee64ee31 ./3/apps.mbn
d3a317a0eb91b19cd17671fafb1ac35c ./3/UQCN.mbn
9a85dfae619ef5e6df8d3ebc8efe81d7 ./6/UQCN.mbn
920000c8fd170956b436461203b38c97 ./UMTS/amss.mbn
d7496085f1af3d1bfdf0fa60c3222766 ./UMTS/apps.mbn

The directory structure seems to be the same--Verizon firmware is in the ./1/ directory. Although the firmware that is available is different.

Revision history for this message
Simon Hirscher (codethief) wrote :

@Brian: Actually that sounds similar to the problem I am experiencing at the moment: It worked, first, but after a reboot it didn't. Put another firmware in place. Worked. Reboot: Failure. Yet another reboot: Works again... (For a more comprehensive description look above.)
I noticed that if it doesn't work a warm reboot will usually result in the device denying service again yet a cold reboot often does the trick (though not always). I suppose the Gobi device doesn't lose state on a warm reboot (as it doesn't lose power).

Ok, I might be totally wrong here but could you check if it works constantly (i.e. after every boot?) and didn't just this time? Also, do remember the error message you received on Linux (/var/log/syslog)?

Revision history for this message
Brian (x-brian) wrote :

With the correct firmware and kernel v2.6.35rc1 it is so far working every time. I have done multiple cold and warm reboots. This kernel has some issues, I think, but nothing that's a show stopper. Looks like the SD card reader might not work (never used it) and it makes a strange sound on shutdown. Also VirtualBox DKMS won't install with this kernel, but none of those things are show stoppers for me.

Your problems seem related to qcserial. My device is supported--it is listed on the page at:
http://cateee.net/lkddb/web-lkddb/USB_SERIAL_QUALCOMM.html

If your device isn't listed there that might be why you're having troubles.

Revision history for this message
Simon Hirscher (codethief) wrote :

It is listed there, and as I said the device sometimes works perfectly (4 Mbit/s down!) and then again it doesn't at all. Also, it's not just Linux but also Windows which means my problems are either related to the device itself or the firmware which both operating systems use…

Revision history for this message
Brian (x-brian) wrote :

Just got an update notification for kernel v2.6.32-23. Didn't install that--just wondering if anyone uses it and will it work? Maybe I can try it in a while but not now; I will need my WWAN in the near future.

Revision history for this message
tnf (tnf-ars-solvendi) wrote :

I installed the update and had to patch the kernel again as described in comment 33. No need to reinstall gobi_loader. Works fine for me

Revision history for this message
Brian Candler (b-candler) wrote :

Works for me too, after v2.6.32-23 update then rebuilding the qualcomm and thinkpad_acpi modules as done originally. I couldn't load the previously-built modules.

Revision history for this message
Brian (x-brian) wrote :

Went ahead and (up)graded my kernel from v2.6.35-rc1 to v2.6.32-23 and patched as described in https://bugs.launchpad.net/ubuntu/+source/linux/+bug/554099/comments/33

Works now, though another bug is present in this kernel too (Intel i3/i5/i7 fails to resume from suspend).

So now I am going to try to apply two patches, at some point.

Revision history for this message
Tapani Tarvainen (ubuntu-tapani) wrote :

Brian, this does not relate directly to this bug, but suspend-resume works just fine even after upgrading kernel to 2.6.32-23 (and applying the patch) in this machine, which has i5 M520 CPU.

Paul Sladen (sladen)
summary: - qualcomm gobi (2000) mobile broadband module not working after upgrade
- to lucid
+ [PATCH] Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broke in 10.04
tags: added: regression
Paul Sladen (sladen)
summary: - [PATCH] Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broke in 10.04
+ [PATCH] Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken in 10.04
Revision history for this message
Lance Albertson (lance-osuosl) wrote : Re: [PATCH] Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken in 10.04

I tried the backports package listed in the duplicate bug #592046 in comment #7 and was able to successfully get my Gobi 2000 to function on Verizon.

Revision history for this message
Simon Hirscher (codethief) wrote :

In case someone's experiencing the same issues as I do:
Over at http://forums.lenovo.com/t5/W-Series-ThinkPad-Laptops/w510-BSOD/td-p/202086 it's becoming more and more obvious that it's due to the Windows drivers / firmware that we also use on Linux. Let's just hope that there's going to be a fix (or that some Linux guru manages to program a native driver? :)).

Revision history for this message
Tom Inglis (tominglis) wrote :
Revision history for this message
Brian (x-brian) wrote :

Hmmm. If you look at:
https://www.codeaurora.org/patches/quic/gobi/

there are a bunch of patches from various dates, with the latest being 06182010. The files are specific to certain hardware--for example, VU936.Gobi2000Drivers_06182010.tar.gz contains Readme.txt files that have information about "Gobi2000 network driver for Dell 1.0.110" and "Gobi2000 Serial driver for Dell 1.0.80."

Right now I'm assuming that means the "Dell 5620" Gobi 2000 device, which is what I have.

Another file, VR305.Gobi2000Drivers_06182010.tar.gz, contains Readme.txt files stating it is for Asus. Another is for Sony, etc.

What these files do not have, is firmware. They are "Dual BSD/GPL" licensed, so that would be incompatible with the proprietary firmware.

---Actually, i am sure this is for the "Dell 5620" Gobi 2000 device. From the QCSerial2kDell.c source:

static struct usb_device_id QCVIDPIDTable[] =
{
   { USB_DEVICE(0x413c, 0x8185) }, // Dell Gobi 2000 QDL device
   { USB_DEVICE(0x413c, 0x8186) }, // Dell Gobi 2000 Modem Device
   { } // Terminating entry
};

Those device IDs are the device IDs for my Dell 5620 device.

I am guessing this is a replacement for gobi_loader? I've poked around a tiny bit in the source but I'm no coder. I couldn't figure out how it applies firmware.

Revision history for this message
Simon Hirscher (codethief) wrote :

Having looked at the source, I assume that those drivers / kernel modules are actually supposed to be a replacement for the firmware.
However, I'm apparently too stupid to get them working. I compiled both modules for my ThinkPad W510 as described in the README files in N0383.Gobi2000Drivers_06182010.tar.gz and also have a /dev/ttyUSB0 device, now, but lsusb still yields:

Bus 002 Device 003: ID 05c6:9204 Qualcomm, Inc.

… instead of 9205. And the network-manager doesn't find any WWAN device.

Unfortunately, there isn't any contact information on codeaurora.org, either, so we're on our own I guess…

Revision history for this message
Andreas Schildbach (schildbach) wrote :

Just like to note that this bug seems to be fixed with Maverick Merkaat Alpha 2 (kernel 2.6.35-6).

UMTS WWAN works out of the box using my Lenovo T410s, and the PIN request dialog explicitely mentioned "Qualcomm Gobi 2000".

Revision history for this message
Tom Inglis (tominglis) wrote :

Andreas, that's great. Do you mean that it just works without doing anything (!?), or that it works if you use gobi_loader and the Windows firmware?

Is there any chance of getting this change (whatever it is) patched / backported to the LTS?

Revision history for this message
Andreas Schildbach (schildbach) wrote :

With Maverick, the only thing you need to do is use Network Manager to set up a Mobile Broadband connection, select your carrier and then enter the SIM PIN. No copying of firmware or fiddling with gobi_loader involved. Extremely painless, I'd say...

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

This is an extremely common 3G device, used in many netbooks. Please can we fix in Maverick and backport + SRU the fix for Lucid.

Changed in linux (Ubuntu Lucid):
status: New → Confirmed
Changed in linux (Ubuntu Maverick):
status: Triaged → Confirmed
importance: Medium → High
Changed in linux (Ubuntu Lucid):
importance: Undecided → High
tags: added: kernel-needs-review kernel-net regression-release
removed: regression
Revision history for this message
Jeremy Foshee (jeremyfoshee) wrote :

From Bug# 592046(duplicate of this bug):

"Stefan Bader wrote on 2010-06-22: #7
An initial version of the LBM package has been uploaded to http://people.canonical.com/~smb/lp592046/. Can someone please do some testing? Due to licensing the package cannot provide the required firmware files. Please see http://www.codon.org.uk/~mjg59/gobi_loader/ for more information."

Would those of you affected by this mind testing the mentioned LBM package?

Thanks in advance!

~JFo

Revision history for this message
Andreas Schildbach (schildbach) wrote :

I'm afraid I could not reproduce my positive experiences in comments #67 and #69. I suspect that at that particular moment, the firmware was already loaded by booting into Windows 7 first. Did not know that firmwares survive reboots. Sorry about the confusion.

Revision history for this message
Andreas Schildbach (schildbach) wrote :

In order to test, I would need the firmware files. Does anyone know where to get them from *without* any Windows installation or setup media?

Revision history for this message
Pete Graner (pgraner) wrote :

@Andreas the firmware instructions are in comment #20

Revision history for this message
Andreas Schildbach (schildbach) wrote :

@Pete the problem is, the md5sums of the files there nowhere match the md5s mentioned in the later comments. Also, UQCN.mbn is missing at all. Looks like comment #20 is of no help.

Revision history for this message
Andreas Schildbach (schildbach) wrote :

Answering my own question: Here is how to get the firmware files...

1) Download 7xwc44ww.exe from http://www-307.ibm.com/pc/support/site.wss/MIGR-72938.html

2) Install wine and run "wine 7xwc44ww.exe". Follow the install until it will page fault.

3) Install cabextract and run "cabextract ~/.wine/drive_c/DRIVERS/WWANQL/Driver/GobiInstaller.msi".

4) md5sum _* | grep <md5 to look for> to select the correct file (filename is mangled)

Revision history for this message
Andreas Schildbach (schildbach) wrote :

Ok, with the firmwares in /lib/firmware/gobi and the LBM package from comment #71 installed, I got my o2 Germany UMTS connection working on Lucid on a Thinkpad T410s. Thank you very much!

Revision history for this message
Flurin Rindisbacher (flurin-deactivatedaccount) wrote :

Im writing this on a Thinkpad T410s over a Swisscom 3g connection. I can therefore also confirm that the patch is working.

Revision history for this message
Tom Inglis (tominglis) wrote :

I have this working on my ThinkPad T510 with Vodafone UK, however, it only works intermittently. Sometimes when I start up from cold, the option is not available in NetworkManager and requires another shutdown / startup cycle?

Revision history for this message
Tom Inglis (tominglis) wrote :

PS - I am using the firmware that shipped with the laptop in January, I will try downloading the latest driver as per Andreas' instructions in #76.

Revision history for this message
Brian Candler (b-candler) wrote :

> I have this working on my ThinkPad T510 with Vodafone UK, however, it
> only works intermittently. Sometimes when I start up from cold, the
> option is not available in NetworkManager and requires another shutdown /
> startup cycle?

Have a look at the output of "lsusb", because the device changes its ID pre-firmware load and after firmware load. For me it goes from 05c6:9204 to 05c6:9205

If it shows the former, then it means the firmware hasn't loaded. For me, I can fix as follows:

(1) Run gobi-loader manually
 /lib/udev/gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi

(2) Wait until lsusb shows the expected second device ID

(3) sudo killall modem-manager

HTH,

Brian.

Revision history for this message
Simon Hirscher (codethief) wrote :

@Tom Inglis: Is it constantly working with the latest firmware? If not it might be due to a bug somewhere in the firmware which causes crashes (see http://forums.lenovo.com/t5/W-Series-ThinkPad-Laptops/w510-BSOD/td-p/202086). Also, could you please cite /var/log/syslog?

Revision history for this message
lapistano (lapis-bastian-feder) wrote :

Hey,
#33
I am writing ths via swisscom UMTS from my T410 to confirm the partly success I have following the instructions from comment#33. Thank you very much tnf.
Unfortunately the modem does not appear every time I reboot the machine. It's a kind of a gambling game, if the network manager will show the modem or not, unless it is listed in 'dmesg' and 'lsusb' (see output at the end of the comment).
I use Ubuntu 10.04.

Thx for your supprt.
Lapis

OUTOUT:
$ dmesg
<---snip -->
[ 12.906167] USB Serial support registered for Qualcomm USB modem
[ 12.906323] Registered led device: tpacpi::thinkvantage
[ 12.907578] qcserial 2-1.4:1.2: Qualcomm USB modem converter detected
[ 12.907738] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
[ 12.907761] usbcore: registered new interface driver qcserial
<---snip -->

$ uname -a
Linux trillian 2.6.32-23-generic-pae #37-Ubuntu SMP Fri Jun 11 09:26:55 UTC 2010 i686 GNU/Linux

$ lsusb
Bus 002 Device 003: ID 05c6:9205 Qualcomm, Inc.
Bus 002 Device 002: ID 8087:0020
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 001 Device 004: ID 17ef:480f Lenovo
Bus 001 Device 003: ID 147e:2016 Upek Biometric Touchchip/Touchstrip Fingerprint Sensor
Bus 001 Device 002: ID 8087:0020
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

Revision history for this message
Simon Hirscher (codethief) wrote :

@lapistano: Is there anything in /var/log/syslog worth reporting? (Take a look at my previous messages to get an idea of what I mean.) If so, go on, please. :)

Revision history for this message
Lubo Ivanovic (livanovic) wrote :

Hi,
  I followed instructions from #33, however I end up without any progress. My gobi device does not get mapped to ttyUSB so I cannot use gobi_loader. I am a newbie, so I do not really know, what details to post to give proper basis for diagnosing the issue. Now the questions:
- What should I inspect, to diagnose the issue?
- How do I resolve this?

Thanx in advance

Revision history for this message
lapistano (lapis-bastian-feder) wrote :

Hey Simon,
just installed the recent kernel update (2.6.32-24-generic-pae #38-Ubuntu SMP Mon Jul 5 10:54:21 UTC 2010 i686 GNU/Linux) and followed again the instructions in comment #33. Unfortunately the modem does not appear anymore, unless it is (imho) recongized. ;(

The following log is an extract of the syslog as you requested. As far as I understand it the device gets recognized and initialized. Hope this will help to get the issue with the Qualcomm device resolved.

thx
Lapis

/var/log/syslog
<--snip-->
Jul 25 13:20:40 trillian kernel: [ 12.218342] usbcore: registered new interface driver usbserial
Jul 25 13:20:40 trillian kernel: [ 12.218351] USB Serial support registered for generic
Jul 25 13:20:40 trillian kernel: [ 12.218400] usbcore: registered new interface driver usbserial_generic
Jul 25 13:20:40 trillian kernel: [ 12.218402] usbserial: USB Serial Driver core
Jul 25 13:20:40 trillian kernel: [ 12.243783] USB Serial support registered for Qualcomm USB modem
Jul 25 13:20:40 trillian kernel: [ 12.243797] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
Jul 25 13:20:40 trillian kernel: [ 12.243907] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
Jul 25 13:20:40 trillian kernel: [ 12.243928] usbcore: registered new interface driver qcserial
<---snip--->
Jul 25 13:20:40 trillian NetworkManager: <info> WWAN enabled by radio killswitch; enabled by state file
Jul 25 13:20:40 trillian NetworkManager: SCPlugin-Ifupdown: (161050352) ... get_connections.
Jul 25 13:20:40 trillian NetworkManager: SCPlugin-Ifupdown: (161050352) ... get_connections (managed=false): return empty list.
Jul 25 13:20:40 trillian modem-manager: Loaded plugin AnyData
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Ericsson MBM
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Novatel
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Generic
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Option High-Speed
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Longcheer
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Option
Jul 25 13:20:40 trillian modem-manager: Loaded plugin MotoC
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Nokia
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Huawei
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Sierra
Jul 25 13:20:40 trillian modem-manager: Loaded plugin Gobi
Jul 25 13:20:40 trillian modem-manager: Loaded plugin ZTE
Jul 25 13:20:40 trillian modem-manager: (ttyS0) opening serial device...
Jul 25 13:20:40 trillian modem-manager: (ttyS0): probe requested by plugin 'Generic'
Jul 25 13:20:40 trillian modem-manager: (ttyUSB0) opening serial device...
Jul 25 13:20:40 trillian modem-manager: (ttyUSB0): probe requested by plugin 'Gobi'
Jul 25 13:20:40 trillian NetworkManager: Ifupdown: get unmanaged devices count: 0
<--snip-->

Revision history for this message
lapistano (lapis-bastian-feder) wrote :
Download full text (12.0 KiB)

@Simon: turned out I needed to reload the qcserial module and 'killall modem-manager' to get the modem recognized by the network manager. Once it is listed as an interface in the network manager I can connect to the network using it. As I am now.
I attached the syslog of establishing the connection. I can see the 'unknown error 100' lines as you experiencened in you log, but I do have a stable link with no hickups.

hope this helps
Lapis

$: dmesg
[ 132.972808] usbcore: deregistering interface driver qcserial
[ 132.973059] qcserial ttyUSB0: Qualcomm USB modem converter now disconnected from ttyUSB0
[ 132.973149] qcserial 2-1.4:1.1: device disconnected
[ 132.973247] USB Serial deregistering driver Qualcomm USB modem
[ 148.603023] USB Serial support registered for Qualcomm USB modem
[ 148.603216] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
[ 148.603360] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
[ 148.603376] usbcore: registered new interface driver qcserial
[ 152.813457] usb 2-1.4: USB disconnect, address 3
[ 152.813636] qcserial ttyUSB0: Qualcomm USB modem converter now disconnected from ttyUSB0
[ 152.813660] qcserial 2-1.4:1.1: device disconnected
[ 153.012339] usb 2-1.4: new high speed USB device using ehci_hcd and address 4
[ 153.115551] usb 2-1.4: configuration #1 chosen from 1 choice
[ 153.120055] qcserial 2-1.4:1.2: Qualcomm USB modem converter detected
[ 153.120176] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0

$: less /var/log/syslog
Jul 25 14:16:11 trillian kernel: [ 132.972808] usbcore: deregistering interface driver qcserial
Jul 25 14:16:11 trillian kernel: [ 132.973059] qcserial ttyUSB0: Qualcomm USB modem converter now di
sconnected from ttyUSB0
Jul 25 14:16:11 trillian kernel: [ 132.973149] qcserial 2-1.4:1.1: device disconnected
Jul 25 14:16:11 trillian kernel: [ 132.973247] USB Serial deregistering driver Qualcomm USB modem
Jul 25 14:16:27 trillian kernel: [ 148.603023] USB Serial support registered for Qualcomm USB modem
Jul 25 14:16:27 trillian kernel: [ 148.603216] qcserial 2-1.4:1.1: Qualcomm USB modem converter dete
cted
Jul 25 14:16:27 trillian kernel: [ 148.603360] usb 2-1.4: Qualcomm USB modem converter now attached
to ttyUSB0
Jul 25 14:16:27 trillian kernel: [ 148.603376] usbcore: registered new interface driver qcserial
Jul 25 14:16:31 trillian kernel: [ 152.813457] usb 2-1.4: USB disconnect, address 3
Jul 25 14:16:31 trillian kernel: [ 152.813636] qcserial ttyUSB0: Qualcomm USB modem converter now di
sconnected from ttyUSB0
Jul 25 14:16:31 trillian kernel: [ 152.813660] qcserial 2-1.4:1.1: device disconnected
Jul 25 14:16:31 trillian kernel: [ 153.012339] usb 2-1.4: new high speed USB device using ehci_hcd a
nd address 4
Jul 25 14:16:31 trillian kernel: [ 153.115551] usb 2-1.4: configuration #1 chosen from 1 choice
Jul 25 14:16:31 trillian kernel: [ 153.120055] qcserial 2-1.4:1.2: Qualcomm USB modem converter dete
cted
Jul 25 14:16:31 trillian kernel: [ 153.120176] usb 2-1.4: Qualcomm USB modem converter now attached
to ttyUSB0
Jul 25 14:16:46 trillian modem-manager: Caught signal 15, shutting down...
Jul 25 14:16:46 trillian Netw...

Revision history for this message
lapistano (lapis-bastian-feder) wrote :

btw using the Vodafone (0) firmware if you are conntecting to a swisscom unlimited account does fix some issues with DNS lookups, which take like minutes or sometimes will timeout using the general (6) firmware.

fyi.
Lapis

papukaija (papukaija)
tags: added: lucid maverick
summary: - [PATCH] Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken in 10.04
+ Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken
Revision history for this message
lapistano (lapis-bastian-feder) wrote :

just an update on which firmware to use if you are connecting to swisscom Mobile unlimited. After I experienced a massive DNS Problem using the gerneric firmware, I switched to the Vodafone fw. This 1stly did the trick in resolving the DNS problem. But after serveral reboots and hibernation wakeups the connection speed seemed on the very minimum. SoI monitored the speed of the connection using the System monitor of Ubuntu. Guess what for minutes there were neither a download nor an upload registered even I just requested a webpage with a browser. So I did some trial and error and found out that the fw in dir 7 (Telefonica) does not only provide me with a stable connection (avg. 120KiB/s) but also does not face the DMS problem I had. bozinga! ;)

Just to let you know my findings to let you know I apreciate your work.
Lapis

Revision history for this message
Stefan Bader (smb) wrote :

I am marking the Maverick part as fix released. The kernel parts would have been there before and the loader is now provided by an additional package (gobi-loader).

Changed in linux (Ubuntu Maverick):
status: Confirmed → Fix Released
Changed in linux-backports-modules-2.6.32 (Ubuntu Maverick):
status: New → Invalid
Revision history for this message
Stefan Bader (smb) wrote :

For Lucid we won't fix this in the kernel package (to limit regression potential) but will have a linux-backports-module-wwan binary package added in the next update to lbm there.

Changed in linux (Ubuntu Lucid):
status: Confirmed → Won't Fix
Changed in linux-backports-modules-2.6.32 (Ubuntu Lucid):
assignee: nobody → Stefan Bader (stefan-bader-canonical)
importance: Undecided → Medium
status: New → In Progress
Stefan Bader (smb)
description: updated
Changed in linux-backports-modules-2.6.32 (Ubuntu Lucid):
status: In Progress → Fix Committed
Revision history for this message
Stanislaw Waszkiewicz (tuwxyz) wrote :

OK where can I find fixed deb package? I have already installed linux-backports-modules-wwan-2.6.32-23-generic.
Do I have to update it for new kernel 2.6.32-24?

Revision history for this message
Stefan Bader (smb) wrote : Re: [Bug 554099] Re: Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken

On 08/06/2010 06:59 PM, dev wrote:
> OK where can I find fixed deb package? I have already installed
> linux-backports-modules-wwan-2.6.32-23-generic. Do I have to update it for
> new kernel 2.6.32-24?
>
For now (because LBM with that patch has not been uploaded) I think I had those
rebuild for -24 and added them to the people page mentioned earlier.
When all is completed there should be a
linux-backports-modules-wwan-lucid-<flavour> meta package which can be installed
and will automatically pull the latest lbm-wwan package.

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

Stefan, will gobi-loader be installed by Jockey if it detects the
relevant hardware?

Mark

Revision history for this message
Данило Шеган (danilo) wrote :

Not sure if you need more input, the modules-backports-wwan package works for me on Lucid with gobi-loader 0.7 for a Gobi 1000 card (05c6:9221).

Revision history for this message
Stefan Bader (smb) wrote :

On 08/06/2010 09:40 PM, Mark Shuttleworth wrote:
> Stefan, will gobi-loader be installed by Jockey if it detects the
> relevant hardware?

I am not sure there will be enough time for this to happen in Maverick. But I
put this down to things that should be thought of on UDS-N. There we should also
evaluate whether there is a better way to acquire the firmware files.

Revision history for this message
Martin Pitt (pitti) wrote :

Teaching Jockey about gobi-loader is relatively easy, gobi-loader's udev rules provide a nice vendor-product list already. However, this won't help at all with retrieving the firmware, of course. Incidentally I got a new laptop last Friday (ThinkPad x201), which also has one of those gobi modems, so I now can play around with this as well. (Unfortunately the modem still doesn't work even with gobi-loader and the firmware both installed.)

Revision history for this message
Stefan Bader (smb) wrote :

The rough but not very much developed idea would have been to try get the firmware by fw-cutter. As I don't have the hw I can only guess from descriptions, but it seemed to be not trivial to get the right files. HW vendor FW vs. generic FW.

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

I'll find out if we have anyone working on the firmware issue with Qualcomm.

Mark

Revision history for this message
Tom Inglis (tominglis) wrote :

I updated my firmware from the latest Lenovo driver download using the instructions at #76 and the MD5 checksums here:

http://www.thinkwiki.org/wiki/Talk:Qualcomm_Gobi_2000 (0 for the Vodafone firmware)

I updated my linux-backports-modules-wwan package to the latest version shared here:

http://people.canonical.com/~smb/lp592046/

Now instead of loading the firmware intermittently, it doesn't seem to do it all? Have you any suggestions!? :-)

Revision history for this message
Tom Inglis (tominglis) wrote :

I followed the instructions at #81, and that worked - I am now writing this on my Vodafone connection, but it is jolly slow, despite having the maximum 3G signal? I don't know whether that's a firmware or a Vodafone problem?

It would be cool to have this load automagically at startup though!

Revision history for this message
Marcin (czarlji) wrote :

hi
did anyone now how to run this modem on hp mini 210-1005sa?
i try all options and still nothing:(.it's possible to replace for other modem:)?
or maybe Maverick resolve that problem?

Revision history for this message
Simon Hirscher (codethief) wrote :

@Tom:
> It would be cool to have this load automagically at startup though!

Actually, gobi_loader should do just that.

Revision history for this message
Mark W. Ogden (mwogden) wrote :

Thank you for working on this. using "old os" to access this
internal modem(Qualcomm gobi 2000)
in a Toshiba NB200. Can d/l any patches to my Karmic install.
please direct me to newbie instructions when this is fixed.

DocHofa (zifotronic)
Changed in linux-backports-modules-2.6.32 (Ubuntu Lucid):
status: Fix Committed → Fix Released
status: Fix Released → Fix Committed
Revision history for this message
Simon Hirscher (codethief) wrote :

I'd like to apologize for the very late reply, I only had the chance to have a quick look at you replies from time to time. I was on holidays.

@lapistano:
I am surprised that you could connect despite the "unknown" errors. But how has it been going lately? Can you still establish a connection with no issues at all? Or are there times when you can't?
Also, having read that you own a T410 I suspect that my issues might be related to the fact that my machine is a W510, as 13 pages of complaints over at the Lenovo forums suggest (http://forums.lenovo.com/t5/W-Series-ThinkPad-Laptops/w510-BSOD/td-p/202086/page/13). So, in case everything still works beautifully for you I think we don't have to dig any further.

Revision history for this message
Simon Hirscher (codethief) wrote :

@Lubo Ivanovic: Actually, gobi_loader should do the mapping to /dev/ttyUSB0. Or did I get you wrong?

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

 @DocHofa

Thanks for marking it fix-committed. In order to verify it as
fix-released, could you post the from-scratch steps needed to get this
device working under Lucid, please?

Thanks,
Mark

Revision history for this message
lapistano (lapis-bastian-feder) wrote :
Download full text (3.1 KiB)

@Simon I just upgraded the kernel as Ubunutu 10.04 asked me to and redid the kernel module patching and installation as tnf described in comment #33. If the modem is not recognized I have to

$>sudo rmmod qcserial
$>sudo modprobe qcserial
$>sudo killall modem-manager

to get it recognized and listed in the Networkmanager. Sometimes I am not able to unload the qcserial module, because of somekind of a circular reference between the the 3 (qcserial,usb_wwan, usbserial) modules.
Also a

$>killall modem-manager

does not bring it up, but a reboot sometimes does the trick. But once recognized the connection is just fine (after I found out to use the Telefonica firmware) and responsive. I can establish a PPTP VPN connection to my server with out any problems and it even survives hibernation and an on/off cycle using the 'power switch' to turn on/off the wireless hardware (the switch on the right hand side right next to the eSata port).
About the error you highlighted: I do not experience this behavior on my machine.

For further investigation here are the stats of my machine. If there is anything else you need just let me know.

$ ll 7xwc44ww.exe
-rw-r--r-- 1 lapis lapis 31243392 2010-07-05 20:26 7xwc44ww.exe

$ md5sum 7xwc44ww.exe
5d59b037c74559a2276aeb21cdf5b692 7xwc44ww.exe

$ uname -a
Linux trillian 2.6.32-24-generic-pae #39-Ubuntu SMP Wed Jul 28 07:39:26 UTC 2010 i686 GNU/Linux
$ lsusb
Bus 002 Device 003: ID 05c6:9204 Qualcomm, Inc.
[...]

* dmesg, if modem gets not recognized at startup:
<---snip--->
[ 11.793011] USB Serial support registered for Qualcomm USB modem
[ 11.793150] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
[ 11.793259] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
[ 11.793287] usbcore: registered new interface driver qcserial
<---snip--->

Seems the modem gets recognized by the system, but not attached to a tty to be able to be used. So I reloaded the qcserial and killed the modem-manager stuff. And violà:

* dmesg after reloading the kernel module:
<---snip--->
[ 69.866433] usbcore: deregistering interface driver qcserial
[ 69.866546] qcserial ttyUSB0: Qualcomm USB modem converter now disconnected from ttyUSB0
[ 69.866560] qcserial 2-1.4:1.1: device disconnected
[ 69.866572] USB Serial deregistering driver Qualcomm USB modem
[ 110.872535] USB Serial support registered for Qualcomm USB modem
[ 110.872560] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
[ 110.872641] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
[ 110.872667] usbcore: registered new interface driver qcserial
[ 114.899703] usb 2-1.4: USB disconnect, address 3
[ 114.899891] qcserial ttyUSB0: Qualcomm USB modem converter now disconnected from ttyUSB0
[ 114.899915] qcserial 2-1.4:1.1: device disconnected
[ 115.130953] usb 2-1.4: new high speed USB device using ehci_hcd and address 4
[ 115.233937] usb 2-1.4: configuration #1 chosen from 1 choice
[ 115.238163] qcserial 2-1.4:1.2: Qualcomm USB modem converter detected
[ 115.238288] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0
<---snip--->

the last line is what you wanna see to be sure the modem is ready to...

Read more...

Revision history for this message
DocHofa (zifotronic) wrote :
Download full text (3.6 KiB)

I'm sorry, but sometimes it is better to not work on things while doing some
other things alongside ... I did the wrong click!

LG

Boris

On 19 August 2010 07:56, Mark Shuttleworth <email address hidden>wrote:

> @DocHofa
>
> Thanks for marking it fix-committed. In order to verify it as
> fix-released, could you post the from-scratch steps needed to get this
> device working under Lucid, please?
>
> Thanks,
> Mark
>
> --
> Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken
> https://bugs.launchpad.net/bugs/554099
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in The Linux Kernel: Unknown
> Status in “linux” package in Ubuntu: Fix Released
> Status in “linux-backports-modules-2.6.32” package in Ubuntu: Invalid
> Status in “linux” source package in Lucid: Won't Fix
> Status in “linux-backports-modules-2.6.32” source package in Lucid: Fix
> Committed
> Status in “linux” source package in Maverick: Fix Released
> Status in “linux-backports-modules-2.6.32” source package in Maverick:
> Invalid
>
> Bug description:
> SRU justification (for LBM in Lucid):
>
> Impact: Certain models of 3G hardware (Gobi) which apparently are quite
> common, require several changes to work. First, the kernel module needs
> to be updated, which requires larger modifications. Second, a new firm-
> ware loader is required together with a matching udev rule.
>
> Fix: In order to prevent regressions, the driver backport was done in
> LBM and also adds a new binary subpackage (lbm-wwan). This package also
> contains the loader and the udev rule.
> NOTE! The drivers also requires firmware files which need to be manually
> gathered as they are not redistributable.
>
> Testcase: Without this backport no 3G connections are possible. Sevral
> users in the bug report had successfully used the driver from the test
> package after collecting the required firmware files.
>
> ---
>
> Binary package hint: linux-image-generic
>
> $ uname -a
> Linux phoenix 2.6.32-19-generic #28-Ubuntu SMP Wed Mar 31 17:46:20 UTC 2010
> i686 GNU/Linux
>
> $ cat /proc/version_signature
> Ubuntu 2.6.32-19.28-generic 2.6.32.10+drm33.1
>
> $ lsusb
> 03f0:201d Hewlett-Packard <-- HP un2400 Mobile Broadband Module (QDL
> interface)
>
> HP un2400 mobile broadband modem stopped working after i upgraded to lucid.
> I can not load firmware anymore to my modem's Qualcomm Gobi USB chip using
> gobi_loader.
> http://www.codon.org.uk/~mjg59/gobi_loader/<http://www.codon.org.uk/%7Emjg59/gobi_loader/>
>
> There was no problems on previous Ubuntu version (Karmic, kernel 2.6.31*).
> As i understand all devices based on this chip will not work on lucid, if
> this will not be fixed.
>
> HP un2400 is recognized by kernel's qcserial module. qcserial.c was not
> changed from 2.6.31 to 2.6.32.
> After searching internet i found that this is related to changes in the
> kernel's usbserial module:
> "The qcserial driver is currently broken in 2.6.32 and later. It's due to
> the switch to using kfifo for usb serial, but we haven't been able to work
> out the actual cause. I'm looking at alternative approaches."
> Source: http://mjg59.livejournal.com/120577.html
>
> Contact...

Read more...

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

Ok, resetting to Confirmed.

Changed in linux (Ubuntu Lucid):
status: Won't Fix → Confirmed
Changed in linux-backports-modules-2.6.32 (Ubuntu Lucid):
status: Fix Committed → Confirmed
Revision history for this message
Steve Langasek (vorlon) wrote :

Won't fix in the linux package for lucid, per Stefan in comment #91.

Changed in linux (Ubuntu Lucid):
status: Confirmed → Won't Fix
Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

I'd like to review regression potential once we actually have a fix in
hand. This is a very widely deployed part.

Mark

Revision history for this message
Brian Candler (b-candler) wrote :

Sorry if I'm being dense, but where do I find linux-backports-modules-wwan-lucid[-generic]?

I have uncommented the backports from /etc/apt/sources.list, re-run apt-get update, but still can't find it. Do I have to get it from a private PPA?

This would be easier than recompiling modules every time a patched kernel comes out...

Thanks, Brian.

Revision history for this message
Stefan Bader (smb) wrote :

Just reflecting that for Maverick the user-space part is provided as a separate package.

Changed in gobi-loader (Ubuntu Lucid):
status: New → Invalid
Changed in gobi-loader (Ubuntu Maverick):
status: New → Fix Released
Revision history for this message
Stefan Bader (smb) wrote :

This has been committed to the repo. I am working on an upload right now.

Changed in linux-backports-modules-2.6.32 (Ubuntu Lucid):
status: Confirmed → Fix Committed
Revision history for this message
tnf (tnf-ars-solvendi) wrote :

I have the same question as Brian: which package needs to be installed in lucid? I only get linux-backports-modules-wireless-<kernel>-<flavor> but nothing called ...-wwan-...

Revision history for this message
Pete Graner (pgraner) wrote :

@tnf,

The package is not ready yet, it has been uploaded but not accepted. It should appear shortly, it will be called -wwan

Revision history for this message
Justin Seyster (justin-seyster) wrote :

I tested this on a Samsung N150 netbook that comes with a Gobi 2000 HD 3G modem but without success.

The firmware and hp_wmi kernel driver load successfully, and the wireless broadband appears in the Network Manager applet, but attempting to connect fails immediately without an error message.

It's also worth noting that the firmware that comes with the N150 is different from the firmware in the drivers that HP an Lenovo supply. Unfortunately, nobody at Samsung or Verizon took the trouble to post the driver online, so the only way to get the firmware is off the Windows partition that comes loaded with the netbook.

Revision history for this message
Martin Pitt (pitti) wrote : Please test proposed package

Accepted linux-backports-modules-2.6.32 into lucid-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!

tags: added: verification-needed
Revision history for this message
Martin Pitt (pitti) wrote : Re: [Bug 554099] Re: Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken

Brian Candler [2010-08-20 7:39 -0000]:
> Sorry if I'm being dense, but where do I find linux-backports-modules-
> wwan-lucid[-generic]?

They are not built yet, will still take some hours.

For the initiated, I just binary NEWed the kernel packages, so lbm
should start to build soon. Tomorrow morning I'll NEW them as well,
and then also accept the linux-meta packages. Then all packages will
hit the archive and can be tested.

Revision history for this message
Stanislaw Waszkiewicz (tuwxyz) wrote :

I've installed linux-backports-modules-wwan-lucid-generic, version 2.6.32.25.27 from the proposed repository.
I had some problems. I couldn't upgrade previous version, I had to remove it and install the new one from scratch (aptitude said that it can't overwrite /lib/udev/gobi_loader which already exists in linux-backport-modules-wwan-2.6.32.24-generic).

New version works for me. Qualcomm Incorporated Qualcomm Gobi 2000 appears in Network Manager.

Revision history for this message
Andreas Schildbach (schildbach) wrote :

@dev I had this problem as well, when upgrading from
linux-backports-modules-wwan-2.6.32-23-generic 2.6.32-23.17~wwan1 to
linux-backports-modules-wwan-2.6.32-24-generic 2.6.32-24.18+wwan1

Looks like this issue could affect seamless upgrading of wwan package when kernel packages will be upgraded in the future. Maybe gobi_loader should get its own package?

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

 On 03/09/10 14:55, Andreas Schildbach wrote:
> Maybe gobi_loader
> should get its own package?

Stefan can comment on packaging strategy. A dedicated gobi_loader
package might be easiest for Jockey integration, too.

Mark

Revision history for this message
Stanislaw Waszkiewicz (tuwxyz) wrote :

@Andreas Schildbach Of course the same problem occurred with previous versions too :] But I didn't touch gobi at all. It's the problem with wwan package itself... I think.

Revision history for this message
Remeez (remeez-gmail) wrote :

Hi, I am a newbie in the linux space. I used Synaptic Package Manager and searched for the linux-backports-modules-wwan-lucid-generic, version 2.6.32.25.27 in the list and installed it. After rebooting I can now connect to my 3G connection. I am from South Africa and using VodacomSA. My speeds are 1500 kbps download and 447 kbs upload.

Revision history for this message
Jesse Palser (jessepalser) wrote :

OK, I am totally confused now. Running Ubuntu 10.04 32bit Lucid (fully updated). Can someone post some simple instructions on what to do to get my QualComm 3G modem to work? Thanks!

Revision history for this message
Kaspar Metz (kap) wrote :

The new linux-backports-modules-wwan-2.6.32-25-generic package in lucid-proposed works fine my machine (Thinkpad T410s), my Gobi 2000 shows up under nm and is perfectly usable (3G, o2 Germany).

Unfortunately, linux-backports-modules-wwan-2.6.32-25-generic-pae (with kernel 2.6.32-25-generic-pae) is broken for me.
I get a fatal error when trying to modprobe qcserial by hand:

root@t410s:~# modprobe qcserial
WARNING: All config files need .conf: /etc/modprobe.d/thinkpad_acpi.modprobe, it will be ignored in a future release.
FATAL: Error inserting qcserial (/lib/modules/2.6.32-25-generic-pae/updates/wwan/qcserial.ko): Invalid module format
root@t410s:~#

Revision history for this message
darren (darrenm) wrote :

Still a bit of a pfaff to get this working.

Just installed Lucid on a Lenovo x100e netbook with the Gobi 2000 WWAN device and followed these steps:

Enable all backports and proposed repos, do all updates - not working
Copy the following firmware files:
    c:\program files\QUALCOMM\Images\Lenovo\UMTS\apps.mbn
    c:\program files\QUALCOMM\Images\Lenovo\UMTS\amss.mbn
    c:\program files\QUALCOMM\Images\Lenovo\6\UQCN.mbn
to /lib/firmware/gobi - not working
Install linux-backports-modules-wwan-lucid-generic and reboot - not working
Add my user to dialout (and every other group in Users and Groups) - not working
Run /lib/udev/gobi_loader -2000 /dev/ttyUSB0/ /lib/firmware/gobi and killall modem-manager - working after I right click on network-manager and enable mobile broadband.

I now have to do the final step every time I want to use 3G.

Added to the issues of the laptop locking solid with the default radeon driver, no wireless by default, it's a bit of a sorry state with this laptop running Ubuntu at the moment.

Revision history for this message
darren (darrenm) wrote :

Sorry, should have checked this before commenting - it seems the /lib/udev/rules.d/60-gobi.rules contain the following:

ATTRS{idVendor}=="16d8", ATTRS{idProduct}=="8001", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="1199", ATTRS{idProduct}=="9000", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="03f0", ATTRS{idProduct}=="241d", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9204", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9214", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9224", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9234", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9244", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9264", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9274", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"
ATTRS{idVendor}=="413c", ATTRS{idProduct}=="8185", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"

whereas my device is:

Bus 001 Device 003: ID 05c6:9205 Qualcomm, Inc.
Device Descriptor:
  bLength 18
  bDescriptorType 1
  bcdUSB 2.00
  bDeviceClass 0 (Defined at Interface level)
  bDeviceSubClass 0
  bDeviceProtocol 0
  bMaxPacketSize0 64
  idVendor 0x05c6 Qualcomm, Inc.
  idProduct 0x9205
  bcdDevice 0.02
  iManufacturer 4 Qualcomm Incorporated
  iProduct 3 Qualcomm Gobi 2000

So adding this in there:

ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9205", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"

will make it start on boot.

Revision history for this message
Stefan Bader (smb) wrote :

For Maverick it is its own package but for Lucid I rather want to stay with it
being packaged with LBM. I "just" need to make the rules and the loader specific
to the ABI and flavour.
I am on vacation right now, but I will look into it next week. People that need
different PCI IDs supported should open a new bug. Likewise I think we should
open a new separate bug for the packaging problem in Lucid.

Revision history for this message
darko (darko2) wrote :

On EliteBook 6930p I haven't seen Gobi device with lsusb until I used "echo 1 > /sys/devices/platform/hp-wmi/rfkill3/state". After that network manager recognized it and it works excellent.

Revision history for this message
bekirserifoglu@gmail.com (bekirserifoglu) wrote :

the fix does not work on dell latitude 2110. lsusb reports:

Bus 003 Device 059: ID 413c:8187 Dell Computer Corp.

for my qualcomm device. I guess it's not included in the fix. Can you please add it?

Revision history for this message
Brian (x-brian) wrote :

@bekirserifoglu #132:

You seem to have the firmware loaded on that device--the "Dell 5620" Gobi 2000 device should show up in lsusb as:
  413c:8185 (Dell Gobi 2000 QDL device)
-and-
  413c:8186 (Dell Gobi 2000 Modem Device)
before the firmware is loaded. Once the firmware is loaded, the device will come up as 413c:8186 and 413c:8187.

In #129, Darren states that /lib/udev/rules.d/60-gobi.rules does contain a rule for the Dell device:
  ATTRS{idVendor}=="413c", ATTRS{idProduct}=="8185", RUN+="gobi_loader -2000 $env{DEVNAME} /lib/firmware/gobi"

Not sure what your problem is here. Do you have correct firmware? Do you need to do:
  rfkill unblock wwan
maybe?

Revision history for this message
tnf (tnf-ars-solvendi) wrote :

Works for me: 10.04 64 with kernel 2.6.32-25-generic and linux-backports-modules-wwan-2.6.32-25-generic

Revision history for this message
darren (darrenm) wrote :

Had no luck with Lucid so tried the latest Maverick beta which still doesn't work.

I was incorrect in stating that the device in the Lenovo x100e required 05c6:9205 as this is the id when initialised by the firmware. When not initialised it is 05c6:9204 which is already in /lib/udev/rules.d/60-gobi.rules.

Udev logs that it has run gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi and exited 0 but checking lsusb the device still shows as the uninitialised id. If I run the command myself it shows as 9205 and is initialised.

So it seems something is unloading the firmware after udev has loaded it. I'll post the hardware info for diagnosis.

Revision history for this message
Isa (ibaboo) wrote :

I have a lenovo W510 I have manged to load the firmware successfully, Network-Manger does pick up the modem but It just does not connect.
 NetworkManager: <info> (ttyUSB0): device state change: 3 -> 4 (reason 0)
 NetworkManager: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) scheduled...
 NetworkManager: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) started...
 NetworkManager: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) complete.
 lycid-lynx modem-manager: (ttyUSB0) opening serial device...
 modem-manager: Modem /org/freedesktop/ModemManager/Modems/0: state changed (disabled -> enabling)
 modem-manager: Got failure code 100: Unknown error

Revision history for this message
Simon Hirscher (codethief) wrote :

@lsa: I'm having the same issue on my W510. Take a look at the 19
pages of complaints over here:
http://forums.lenovo.com/t5/W-Series-ThinkPad-Laptops/w510-BSOD/td-p/202086
Seems to be a bug in the firmware which is flashed onto the Gobi device.

Revision history for this message
Isa (ibaboo) wrote :

Hey guys I my 3g is working. The thing that was wrong was (don't laugh) I did not know that the 3g and wireless radio were enable with the same button on the laptop. So I just enabled the the wireless on the side of my laptop and its working like charm. :-)

Martin Pitt (pitti)
tags: added: verification-done
removed: verification-needed
Revision history for this message
blah (billgay) wrote :

I'm facing the same problem Kasper Metz had.

Unfortunately, linux-backports-modules-wwan-2.6.32-25-generic-pae (with kernel 2.6.32-25-generic-pae) is broken for me.
I get a fatal error when trying to modprobe qcserial by hand:

root@TP410s:~# modprobe qcserial
FATAL: Error inserting qcserial (/lib/modules/2.6.32-25-generic-pae/updates/wwan/qcserial.ko): Invalid module format
root@TP410s:~#

syslog says: usb_wwan: disagrees about version of symbol module_layout

when I try loading the module.

Bill

Revision history for this message
Remeez (remeez-gmail) wrote :

Hi Guys,

I have this fix in place and have been using without fail for about 2 weeks now. My system is prompting me to do a partial upgrade, but I am afraid of losing my 3G connection. I can see that the linux-backports-modules-wwan-2.6.32-25-generic is displayed as a proposed update. Should I do the upgrade???

Regards
Remeez

Revision history for this message
RossH (hendo2515) wrote :

I installed the backport and restarted the computer. I got and NVIDIA error decided to use the reduced graphics and was able to connect fine. Restarted the laptop reconfigured the drivers on the drivers error screen resetting to default and now cant see the driver in NM.

When I look in the log i see:

Sep 20 09:50:18 mel-prg01 kernel: [ 13.979322] USB Serial support registered for Qualcomm USB modem
Sep 20 09:50:18 mel-prg01 kernel: [ 13.979348] qcserial 2-1.4:1.1: Qualcomm USB modem converter detected
Sep 20 09:50:18 mel-prg01 kernel: [ 13.979421] usb 2-1.4: Qualcomm USB modem converter now attached to ttyUSB0

rhenderson@mel-prg01:~$ lsusb
Bus 002 Device 003: ID 05c6:9204 Qualcomm, Inc.
Bus 002 Device 002: ID 8087:0020
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 001 Device 005: ID 17ef:480f Lenovo
Bus 001 Device 004: ID 0a5c:217f Broadcom Corp.
Bus 001 Device 003: ID 147e:2016 Upek Biometric Touchchip/Touchstrip Fingerprint Sensor
Bus 001 Device 002: ID 8087:0020
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

Laptop is a Lenovo T510 running 64bit lucid. Any ideas?

Revision history for this message
RossH (hendo2515) wrote :

Ok found I needed to add the three driver files to the /etc/firmware/gobi directory

Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package linux-backports-modules-2.6.32 - 2.6.32-25.23

---------------
linux-backports-modules-2.6.32 (2.6.32-25.23) lucid-proposed; urgency=low

  [ Brad Figg ]

  * SAUCE: ALSA: HDA: Enable internal mic on Dell E6410 and Dell E6510
    - LP: #605047, #628961

linux-backports-modules-2.6.32 (2.6.32-25.22) lucid-proposed; urgency=low

  [ Stefan Bader ]

  * Bump ABI to -25 for proposed upload.

linux-backports-modules-2.6.32 (2.6.32-24.21) lucid-proposed; urgency=low

  [ Tim Gardner ]

  * Add Conflicts/Replaces for compat-wireless
    - LP: #621185

linux-backports-modules-2.6.32 (2.6.32-24.20) lucid-proposed; urgency=low

  [ Tim Gardner ]

  * Added patch infrastructure
    - LP: #621185
  * Update atl1c to 1.0.1.0-NAPI from 2.6.36-rc2
    - LP: #621185

linux-backports-modules-2.6.32 (2.6.32-24.19) lucid-proposed; urgency=low

  [ Stefan Bader ]

  * Add WWAN sub-package and backported qcserial
    - LP: #554099

  [ Tim Gardner ]

  * Make udev firmware helper file name unique by ABI and flavour
    - LP: #535132
  * Added compat-wireless-2.6.35-1
    - LP: #621185
  * Package multiple versions of compat-wireless
    - LP: #621185
  * Update to compat-wireless-2.6.34.1
    - LP: #618578
  * Use POCKET in flavour stub to correctly reference meta files
    - LP: #621185
 -- Brad Figg <email address hidden> Thu, 09 Sep 2010 10:44:36 -0700

Changed in linux-backports-modules-2.6.32 (Ubuntu Lucid):
status: Fix Committed → Fix Released
Revision history for this message
lapistano (lapis-bastian-feder) wrote :

Hey folks,
I recently installed the kernal update to 2.6.32.25. Unfortunately this broke the working solution with the 2.6.32.24 kernel.
Meaning I cannot patch the kernel modules like described in #32, which btw always worked fine for me :|
So I went to install the backports. This unfortunately does also not work. Even after I recreated the missing /etc/firmware/gobi dir and copying the necessary files into it.
Looking into the log I find:

usb_wwan: disagrees about version of symbol module_layout

Furthermore loading the qcserial module manually returns:

$ sudo modprobe qcserial
FATAL: Error inserting qcserial (/lib/modules/2.6.32-25-generic-pae/updates/wwan/qcserial.ko): Invalid module format

Could anyone give me a heads up what this means?

thx in advance
Lapis

btw: my stats:

$ uname -a
Linux trillian 2.6.32-25-generic-pae #44-Ubuntu SMP Fri Sep 17 21:57:48 UTC 2010 i686 GNU/Linux
$ dpkg --get-selections '*wwan*'
linux-backports-modules-wwan-2.6.32-25-generic deinstall
linux-backports-modules-wwan-2.6.32-25-generic-pae install
linux-backports-modules-wwan-lucid-generic-pae install

Revision history for this message
Stefan Bader (smb) wrote :

Usually it means that one kernel module has been compiled against different
headers than the kernel (or another kernel module providing some functions).
In your case there seems to be some mismatch in the usb_wwan module. The invalid
module format is just a misleading result of failing to load usb_wwan as a
dependency.

Doing a quick test there seems to be no problem for me to probe it after
updating the kernel and linux-backports-module to the currently latest version.
I am not sure what you patch and why as the gobi_loader and the updated kernel
modules are part of the backports package in Lucid.

Revision history for this message
lapistano (lapis-bastian-feder) wrote :

Hi Stefan,

thx for the fast reply.
As I recently got a new SSD.. I am going to reinstall the system anyway.
So if I end up with being able to use it then .. I am fine.

What I did:
I installed and used the gobi-loader manually and patched the kernel
modules as described by tnf. in comment #32.

As of the kernal update to revision *.25 the patch does not apply to the
ne sources. So I skipped the manual patching and used apt-get to
retrieve and install the latest wwan backports matching the linux kernel
on my machine.
This is when I ended up with my system not recognising the modem anymore.

As far as I know I could always go back to the latest working kernel
(since grub keeps it somewhere) and use the modem there. But this would
be a not so fine workaround ;)

hf and thx
Bastian (aka Lapistano)

Revision history for this message
Anders Østerholt (diebels) wrote :

My thinkpad x201 with id 05c6:9204/05c6:9205 is working fine with gprs and 3g. Needs reloading of qcserial sometimes. Is there any chance to get in this patch in: https://patchwork.kernel.org/patch/204812/ to get gps working, or should this be a separate bug?

Revision history for this message
AmenophisIII (amenophisiii) wrote :

i think this bug (report) is messy enough. i opened #653126 for this matter.

Revision history for this message
Brian Candler (b-candler) wrote :

The patch from comments #14 and #15 no longer applies cleanly to latest Lucid kernel (2.6.32-25.44). An updated patch is attached.

Since I need to rebuild the modules every time a kernel update comes out, I do it using the following script. It's based on the instructions from comment #14 but also rebuilds thinkpad_acpi.

#!/bin/sh -ex
rm -rf linux-source-2.6.32 || true
tar -xvjf /usr/src/linux-source-2.6.32.tar.bz2
cd linux-source-2.6.32
patch -p1 < ../usb-wwan-2.6.32.diff
cp ../thinkpad_acpi.c drivers/platform/x86/

cp /lib/modules/`uname -r`/build/.config ../temp
cp /lib/modules/`uname -r`/build/Module.symvers ../temp
cp /lib/modules/`uname -r`/build/Makefile .

make O=../temp outputmakefile
make O=../temp archprepare
make O=../temp prepare
make O=../temp modules SUBDIRS=scripts
make O=../temp modules SUBDIRS=drivers/usb/serial/
make O=../temp modules SUBDIRS=drivers/platform/x86/

sudo cp ../temp/drivers/usb/serial/usb_wwan.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo cp ../temp/drivers/usb/serial/qcserial.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo cp ../temp/drivers/usb/serial/option.ko /lib/modules/`uname -r`/kernel/drivers/usb/serial/
sudo cp ../temp/drivers/platform/x86/thinkpad_acpi.ko /lib/modules/`uname -r`/kernel/drivers/platform/x86/
sudo /sbin/depmod -a

Revision history for this message
lapistano (lapis-bastian-feder) wrote :

Hi Brian,
thx for the shell script. Unless the backports do not work for me, I wanted to give it a try. But as I as scanning what the script does and running the single commands it expresses by hand, I realized the 'thinkpad_acpi.c' is not available in the parent dir. Could you please give me a hint how to get it or where to find it else.

Thanks for your help.
Bastian

Revision history for this message
Brian (x-brian) wrote :

My WWAN stopped working today, probably due to installation of a new kernel on 9/30/2010, followed by my first cold boot today. Warm reboots/suspend-resume kept my firmware loaded for about a week.
$ uname -srvmo
Linux 2.6.32-25-generic #44-Ubuntu SMP Fri Sep 17 20:05:27 UTC 2010 x86_64 GNU/Linux

So I installed linux-backports-modules-wwan-lucid-generic from Synaptic. Then I copied my firmware files to /etc/firmware/gobi:
$ sudo mkdir -p /etc/firmware/gobi
$ sudo cp /lib/firmware/gobi/* /etc/firmware/gobi/

I had my firmware in /lib/firmware/gobi, but I don't think that worked. Since someone above mentioned /etc/firmware/gobi I just put it in both places. Now it works great!

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

Are there clear instructions for where to find the firmware in the first
place?

Revision history for this message
Brian (x-brian) wrote :

@ Mark #152

I don't think there are clear instructions--part of that is because firmware has to be obtained from the hardware manufacturer, and they are only provided in Windows driver packages. Mine came from Dell, so I had to pry the firmware out of the Dell Windows drivers. Others have to get it from Lenovo, HP, etc. The md5sums are available for at least a couple of firmware sources above, but there's probably a need for some kind of HowTo or something.

Revision history for this message
Jelle de Jong (jelledejong) wrote :

Is this problem already fixed in the mainstream kernel. The WLAN modules are broken for quite a while now.... A fix should be this hard? An status report would be appreciated.

Revision history for this message
Stefan Bader (smb) wrote :

The kernel modules in Maverick and onward are fine. gobi_loader is a separate
package there.

Revision history for this message
darren (darrenm) wrote :

So if it doesn't work we should file bugs against gobi_loader?

Revision history for this message
Stefan Bader (smb) wrote :

I could think of two cases. If the usb id for your modem is in the udev rules
file provided by the package and the firmware files get loaded but it does not
work, the problem would rather be the kernel (linux) package.
If the udev rules miss the id and there is no indication that the firmware is
tried to load, then the problem should be reported against the loader.
Any case it helps to have problems reported separately as things work for some
people and this hints that this is rather a problem very specific to a certain
make of hardware.

Revision history for this message
Anders Østerholt (diebels) wrote :

Howto for Maverick, generic UMTS on Thinkpads:
Download and extract windows driver install file from Lenovo. http://download.lenovo.com/ibmdl/pub/pc/pccbbs/mobiles/7xwc44ww.exe or get the latest from http://www-307.ibm.com/pc/support/site.wss/MIGR-72938.html
Install a recent wine as 1.2 crashes during msi extraction.
Run the install file with wine.
Extract firmware from GobiInstaller.msi. mkdir ~/.wine/drive_c/DRIVERS/GOBI; wine msiexec /a ~/.wine/drive_c/DRIVERS/WWANQL/Driver/GobiInstaller.msi /qb TARGETDIR=C:\\DRIVERS\\GOBI

install "gobi-loader"

Copy firmware into /lib/firmware/gobi. sudo mkdir /lib/firmware/gobi; cd ~/.wine/drive_c/DRIVERS/GOBI/Images/Lenovo; sudo cp 6/* /lib/firmware/gobi/; sudo cp UMTS/* /lib/firmware/gobi/

Do a reboot or modem-manger restart and qcserial unload load. sudo pkill modem-manager; sudo rmmod qcserial; sudo modprobe qcserial

You should now have a "New Mobile Broadband (GSM) connection..." in the nm-applet. If not, do the thing in the previous paragraph or have a look at syslog

Revision history for this message
Giuseppe Scarciglia (gscarcio) wrote :

@Andres
thanks for your guide.
I've a thinkpad edge 13 with qualcomm gobi 2000 modem. I've done your instructions with Maverick release.
Now i can see a "New Mobile Broadband (GSM) connection..." in the nm-applet but it's impossibile to connect.
The message is always the same "GSM network, now it's offline"..
I'don't undrestand because:
* /lib/udev/60-gobi.rules contain my ID modem
* when i reboot in /var/log/messages i can see: Qualcomm USB modem converter now attached to ttyUSB0
* run lsusb and id changes from 05c6:9204 to 05c6:9205 (firmware loaded)
...is there anything else I can do?

(Thinkpad EDGE 13" Ubuntu maverick 32bit)
Thanks a lot!

Revision history for this message
Anders Østerholt (diebels) wrote :

@Giuseppe
you could try ~/.wine/drive_c/DRIVERS/GOBI/Images/Lenovo/8/UQCN.mbn

There are different providers listed at http://www.thinkwiki.org/wiki/Talk:Qualcomm_Gobi_2000

Do a tail -f /var/log/syslog, and modem-manager should give some output when it tries to connect.

Revision history for this message
Konstantin (jakshi) wrote :

I also have problem with autostarting.
udev rule with firmware loading executed but nothing happend.
But when I try to do the same in console, all work well. Firmware uploads to device, 3G appears.
May be gobi loader broken and exit without do anything in certain circumstance?

Revision history for this message
Konstantin (jakshi) wrote :

My workaround is:
Create file ~/bin/gobi1000:

#! /bin/sh
sudo /lib/udev/gobi_loader-25-generic /dev/ttyUSB0 /lib/firmware/gobi
sudo pkill modem-manager

Set execution bit for this file:

chmod a+x ~/bin/gobi1000

place gobi1000.desktop in ~/.config/autostart:

[Desktop Entry]
Type=Application
Exec=/home/jakshi/bin/gobi1000
Hidden=false
NoDisplay=false
X-GNOME-Autostart-enabled=true
Name[ru_UA]=Gobi1000 3G activate
Name=Gobi1000 3G activate
Comment[ru_UA]=upload firmware and restart modem-manager
Comment=upload firmware and restart modem-manager

add to /etc/sudoers:

%admin ALL=NOPASSWD: /lib/udev/gobi_loader-25-generic
%admin ALL=NOPASSWD: /usr/bin/pkill

Revision history for this message
Simon Hirscher (codethief) wrote :

For those of you who own a W510 and couldn't get the gobi device working in the past (due to it constantly crashing) I have very good news:
Lenovo finally released (it took them more than half a year) a new firmware version which should fix this issue: http://forums.lenovo.com/t5/W-Series-ThinkPad-Laptops/w510-BSOD/m-p/309878#M11808

Revision history for this message
darren (darrenm) wrote :

Can someone change this from fix released please? It's not fixed, it still doesn't work unless you manually load the driver yourself using gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi and then killall modem-manager after every boot.

Revision history for this message
Mark Shuttleworth (sabdfl) wrote :

Yes, please leave the bug open until the fix is Dad-proof:

 - enable backports
 - install a package
 - go

Mark

Revision history for this message
Brian Candler (b-candler) wrote :

@lapistano: to get thinkpad_acpi.c, see #557736 comment #23

The one I was using is:

$ grep TPACPI_VERSION /v/build/thinkpad_acpi.c | head -1
#define TPACPI_VERSION "0.24-20100516"

$ sha1sum /v/build/thinkpad_acpi.c
029c81cb90b0e0b8d3775913e4ddb5820b50683b /v/build/thinkpad_acpi.c

However I've now moved to Maverick and the default thinkpad_acpi module loads without problems.

Revision history for this message
J Boyan (ubuntu-boyan) wrote :

The workaround is *not* working for me, I cannot activate the device from Ubuntu. lsusb shows
     Bus 002 Device 003: ID 05c6:9204 Qualcomm, Inc.
and when I run
     sudo gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi
it just hangs. :-(

However, if I boot into Windows and initiate the Verizon connection there, and then reboot back into Ubuntu, the setup DOES work - I see "Verizon connection" in the network manager dropdown; the lsusb shows "05c6:9205"; and I can connect, disconnect, reconnect successfully. Unfortunately, after suspend and resume, or after shutdown and powering back on, the device is deactivated (05c6:9204) and I can't reactivate it from Ubuntu.

Any ideas? What additional info can I provide to help debug this?

I'm running Maverick 10.10 with kernel 2.6.35-23 on a Lenovo Thinkpad X201 with Gobi 2000 / Verizon CDMA.
$ md5sum /lib/firmware/gobi/*
06f76ed398458dad7b91c2d99a85a0a7 /lib/firmware/gobi/amss.mbn
88a60ed745d75fb1b92c539574ecc972 /lib/firmware/gobi/apps.mbn
2dccbd125ddd2cb327309ba75c6054d2 /lib/firmware/gobi/UQCN.mbn
(these were copied from the Qualcomm 1/ subdirectory in Windows)

Revision history for this message
Jochen Gruse (lpnet) wrote :

@JBoyan:
On my T410 with the same modem (but different firmware), I have had the same problem. I've opened a new bug - including a bugfix/workaround - for modem-manager, which I found to be the culprit.

Have a look here: https://bugs.launchpad.net/ubuntu/+source/modemmanager/+bug/686418

Short version: Add the line

ATTRS{idVendor}=="05c6", ATTRS{idProduct}=="9204", ENV{ID_MM_DEVICE_IGNORE}="1"

to the file "/lib/udev/rules.d/77-mm-usb-device-blacklist.rules", then reboot. This should prevent the modem-manager from corrupting the firmware download.

Hope that helped,
Jochen Gruse

Revision history for this message
Aleksander Morgado (aleksander-m) wrote :
Download full text (3.2 KiB)

Just followed the steps given by Anders (comment 158) and now I can see the new Mobile Broadband connection in the NetworkManager applet. But, when I try to use it, 3G won't work, the connection only sets up in GSM mode (not even GPRS mode). Of course, I can use that connection with 3G or GPRS when using the external usb dongle.

Any idea why this would happen?

When I configure the mobile broadband connection to force using only 3G, syslog says:
ec 25 21:02:09 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) complete.
Dec 25 21:02:09 piticli-ubuntu NetworkManager[1213]: <warn> GSM connection failed: (32) Operation not supported
Dec 25 21:02:09 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): device state change: 4 -> 9 (reason 1)
Dec 25 21:02:09 piticli-ubuntu NetworkManager[1213]: <info> Marking connection 'Orange' invalid.
Dec 25 21:02:09 piticli-ubuntu NetworkManager[1213]: <warn> Activation (ttyUSB0) failed.
Dec 25 21:02:09 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): device state change: 9 -> 3 (reason 0)
Dec 25 21:02:09 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): deactivating device (reason: 0).
Dec 25 21:02:10 piticli-ubuntu modem-manager: CS registration state changed: 1
Dec 25 21:02:10 piticli-ubuntu modem-manager: Modem /org/freedesktop/ModemManager/Modems/0: state changed (searching -> registered)
Dec 25 21:02:10 piticli-ubuntu modem-manager: PS registration state changed: 1
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) starting connection 'Orange'
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): device state change: 3 -> 4 (reason 0)
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) scheduled...
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) started...
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): device state change: 4 -> 6 (reason 0)
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) complete.
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) scheduled...
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) started...
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): device state change: 6 -> 4 (reason 0)
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Activation (ttyUSB0) Stage 1 of 5 (Device Prepare) complete.
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <warn> GSM connection failed: (32) Operation not supported
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): device state change: 4 -> 9 (reason 1)
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> Marking connection 'Orange' invalid.
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <warn> Activation (ttyUSB0) failed.
Dec 25 21:02:20 piticli-ubuntu NetworkManager[1213]: <info> (ttyUSB0): device state change: 9 -> 3 (reason 0)
Dec 25 21:02:20 piticli-u...

Read more...

Revision history for this message
Simon Hirscher (codethief) wrote :

@Aleksander: I suppose you have a ThinkPad. Which one exactly? Also, are you using the latest Gobi firmware available?

Revision history for this message
Aleksander Morgado (aleksander-m) wrote :

@Simon: Yes, it's a thinkpad T510i bought a couple of months ago. The gobi firmware was freshly downloaded from the Lenovo Support webpage (1.1.170/2.0.7.3, released on 2010/10/28)

Revision history for this message
Simon Hirscher (codethief) wrote :

Hm, the only thing I know is that the firmware still has some bugs. At least I experience the one or other every now and then. However, I've got a W510 where it's known that the firmware is extremely buggy. I'm not sure about your problem… Does it always occur?

Revision history for this message
Aleksander Morgado (aleksander-m) wrote :

@Simon: At least all the 5-6 times I've tried, always connected in GSM mode and if forcing GPRS or 3G it didn't work. So, quite useless the SIM connector right now... :-/

Revision history for this message
Michael Rauch (michael-rauch) wrote :

Post 158 works, but i had to install also the following package for ubuntu 10.04.

linux-backports-modules-wwan-2.6.32-27-generic

You have to choose the package accordingly to your running kernel.

michal@linux:~$ uname -r
2.6.32-27-generic

Testet with a lenovo T510

michal@linux:~$ lsusb
Bus 002 Device 004: ID 05c6:9205 Qualcomm, Inc.

Revision history for this message
AIAMUZZ (muzafsh-113) wrote :

Hi,

All-in-all i have been able to find my Gobi 2000 device on my Ubuntu 10.10, Network Manager only 3 times.
(the first time as a GSM device, the second as a CDMA device and third as a GSM device again.)

ever since i am just not able to get it running ...

When i try to run the following command ...

"sudo /lib/udev/gobi_loader /dev/ttyUSB0 /lib/firmware/gobi"
or
"sudo /lib/udev/gobi_loader -2000 /dev/ttyUSB0 /lib/firmware/gobi"

the terminal hangs...

Could this be because the modem manager may be preventing the gobi_loader from loading the firmware ?
Or
Could it be because the patch (usb-wwan-2.6.34.diff from - http://www.codon.org.uk/~mjg59/gobi_loader/) spoken about in Viktors Petrov's post 14 was not added ?
" https://bugs.launchpad.net/ubuntu/maverick/+source/linux/+bug/554099/comments/14 "

Detail of My Machine :
----------------------

Acer Aspire One 8.9
hosting a 'HP Qualcomm Gobi 2000 card'
with no luck on Windows 7 with this card, i am trying to have it running on my Ubuntu.

I am using Ubuntu : 10.10
on kernel : 2.6.35-24-generic

I have uninstalled 'gobi_loader' as well as 'qcserial' completely

I did the following...

1. firmware files : 3 firmware files copied to /lib/firmware/gobi/ Checked

2. Gobi_loader : 60-gobi-rules file carries my device ID's Checked
   downloaded-extracted-make-make install-

3. qcserial : qcserial.c file carries my device ID's Checked
   downloaded-extracted-make-make install-

post installation i verified the installation ...

1. 'gobi_loader' executable installed to location ... - /lib/udev/...
2. '60-gobi-rules' file installed to location ... - /etc/udev/rules.d/...
3. 'qcserial.ko' installed to location ... - /lib/modules/.../kernel/drivers/usb/serial...

the command 'lsmod' shows the following activity ...
"
qcserial 2620 1
bluetooth 50500 9 rfcomm,sco,bnep,l2cap,btusb
snd_seq_device 5744 3 snd_seq_midi,snd_rawmidi,snd_seq
psmouse 59033 0
usbserial 33100 3 qcserial
"

command 'dmesg' shows the following activity ...
"
[ 13.955839] USB Serial support registered for Qualcomm USB modem
[ 13.960673] qcserial 1-6:1.1: Qualcomm USB modem converter detected
[ 13.961281] usb 1-6: Qualcomm USB modem converter now attached to ttyUSB0
[ 13.961485] usbcore: registered new interface driver qcserial
"

Please let me know what i should try...

to apply the patch ' usb-wwan-2.6.34.diff ' or a different direction all together...

PS : I am not a programmer just a brave novice ubuntu user :).
So any instructions bearing this fact in mind would be most helpful.

thanks
Muzafsh.

Revision history for this message
AIAMUZZ (muzafsh-113) wrote :

Well I tried the procedure listed by "Viktors Petrov's post 14 "

Can somebody who understands the output/result of those steps go through the attached file and confirm if my effort to patch " usb-wwan-2.6.34.diff " was a success or a failure ?

However after running the above My Gobi 2000 device is still as elusive as before :(

the 'dmesg' and 'lsmod' commands record the following activity ...

"
[ 14.071236] USB Serial support registered for Qualcomm USB modem
[ 14.071305] qcserial 1-6:1.1: Qualcomm USB modem converter detected
[ 14.071564] usb 1-6: Qualcomm USB modem converter now attached to ttyUSB0
[ 14.071655] usbcore: registered new interface driver qcserial
"

&

"
usb_wwan 9953 1 qcserial
acerhdf 6895 0
cfg80211 144470 3 ath5k,mac80211,ath
videodev 43098 1 uvcvideo
psmouse 59033 0
usbserial 33100 4 qcserial,usb_wwan
"

Please help ...

thanks,

Revision history for this message
AIAMUZZ (muzafsh-113) wrote :

Well I tried the procedure listed by "Viktors Petrov's post 14 "

Can somebody who understands the output/result of those steps go through the attached file and confirm if my effort to patch " usb-wwan-2.6.34.diff " was a success or a failure ?

However after running the above My Gobi 2000 device is still as elusive as before :(

the 'dmesg' and 'lsmod' commands record the following activity ...

"
[ 14.071236] USB Serial support registered for Qualcomm USB modem
[ 14.071305] qcserial 1-6:1.1: Qualcomm USB modem converter detected
[ 14.071564] usb 1-6: Qualcomm USB modem converter now attached to ttyUSB0
[ 14.071655] usbcore: registered new interface driver qcserial
"

&

"
usb_wwan 9953 1 qcserial
acerhdf 6895 0
cfg80211 144470 3 ath5k,mac80211,ath
videodev 43098 1 uvcvideo
psmouse 59033 0
usbserial 33100 4 qcserial,usb_wwan
"

Please help ...

thanks,

Revision history for this message
AIAMUZZ (muzafsh-113) wrote :

Hi,
Without any explanation my Gobi 2000 show's up in the network manager. subsequent restarts don't result in a consistent firmware load. so its is occassional.

When the firmware got loaded, i tried to connect through my DOCOMO(indian SIM Card) 3G network.
But no success in establishing an effective connection.

I have attached the syslog file.
Can somebody help me diagnose the reason why my 3G account is not able to connect.

thanks.

Changed in linux:
status: Unknown → Confirmed
Changed in linux:
importance: Unknown → Medium
status: Confirmed → Fix Released
Revision history for this message
darren (darrenm) wrote :

Fix released? Awesome!

Anyone know where I can get it?

Revision history for this message
Alex Denvir (coldfff) wrote :

Done some testing with both lucid and maverick on a Dell Latitude E5410 here -

With maverick, installing gobi-loader, dropping the firmware into /lig/firmware/gobi and rebooting works first time, nice and easy.

With lucid, I had absolutely no success at all. Installed gobi-loader, backports etc. and network-manager would never show any options under mobile broadband.

Ideally I'd like to get this working on lucid as I have about 4 of these laptops in the company, and would preferably like to stick to the LTS, but I guess I can cope with shifting to maverick for this.

Revision history for this message
Simon Hirscher (codethief) wrote :
Download full text (4.1 KiB)

@Alex: I'm also running Lucid and Gobi works fine here (at least most
of the time but that might be related to my ThinkPad W510). I didn't
install any backports packages, though, but patched the kernel by hand
as described in one of messages at the very beginning of this bug
discussion. (Also 2.6.35-020635rc4-generic is my kernel version in
case you need it but any newer version should do, too.)

On Wed, Mar 2, 2011 at 17:32, Alex Denvir <email address hidden> wrote:
> Done some testing with both lucid and maverick on a Dell Latitude E5410
> here -
>
> With maverick, installing gobi-loader, dropping the firmware into
> /lig/firmware/gobi and rebooting works first time, nice and easy.
>
> With lucid, I had absolutely no success at all. Installed gobi-loader,
> backports etc. and network-manager would never show any options under
> mobile broadband.
>
> Ideally I'd like to get this working on lucid as I have about 4 of these
> laptops in the company, and would preferably like to stick to the LTS,
> but I guess I can cope with shifting to maverick for this.
>
> --
> You received this bug notification because you are a direct subscriber
> of the bug.
> https://bugs.launchpad.net/bugs/554099
>
> Title:
>  Qualcomm Gobi 2000 3G (gobi_loader/qcserial) broken
>
> Status in The Linux Kernel:
>  Fix Released
> Status in “gobi-loader” package in Ubuntu:
>  Fix Released
> Status in “linux” package in Ubuntu:
>  Fix Released
> Status in “linux-backports-modules-2.6.32” package in Ubuntu:
>  Invalid
> Status in “gobi-loader” source package in Lucid:
>  Invalid
> Status in “linux” source package in Lucid:
>  Won't Fix
> Status in “linux-backports-modules-2.6.32” source package in Lucid:
>  Fix Released
> Status in “gobi-loader” source package in Maverick:
>  Fix Released
> Status in “linux” source package in Maverick:
>  Fix Released
> Status in “linux-backports-modules-2.6.32” source package in Maverick:
>  Invalid
>
> Bug description:
>  SRU justification (for LBM in Lucid):
>
>  Impact: Certain models of 3G hardware (Gobi) which apparently are quite
>  common, require several changes to work. First, the kernel module needs
>  to be updated, which requires larger modifications. Second, a new firm-
>  ware loader is required together with a matching udev rule.
>
>  Fix: In order to prevent regressions, the driver backport was done in
>  LBM and also adds a new binary subpackage (lbm-wwan). This package also
>  contains the loader and the udev rule.
>  NOTE! The drivers also requires firmware files which need to be manually
>  gathered as they are not redistributable.
>
>  Testcase: Without this backport no 3G connections are possible. Sevral
>  users in the bug report had successfully used the driver from the test
>  package after collecting the required firmware files.
>
>  ---
>
>  Binary package hint: linux-image-generic
>
>  $ uname -a
>  Linux phoenix 2.6.32-19-generic #28-Ubuntu SMP Wed Mar 31 17:46:20 UTC 2010 i686 GNU/Linux
>
>  $ cat /proc/version_signature
>  Ubuntu 2.6.32-19.28-generic 2.6.32.10+drm33.1
>
>  $ lsusb
>  03f0:201d Hewlett-Packard <-- HP un2400 Mobile Broadband Module (QDL interface)
>
>  HP un2400 mobile broadband modem stoppe...

Read more...

noarz (noar-adv)
Changed in linux (Ubuntu Maverick):
assignee: nobody → noarz (noar-adv)
Revision history for this message
Benjamin Tayehanpour (konaya) wrote :

I wish I could reproduce Alex's success on my own laptop, but I can't. I'm running Maverick, I've installed gobi-loader, extracted the firmware etc etc, but it just won't work. lsof shows that gobi_loader is stuck in the background trying to upload the firmware to /dev/ttyUSB0, which shows that it at least recognizes the device...

Oh, and I've tried with both 2.6.35-28 and 2.6.36-020636. What to do what to do?

Revision history for this message
Tom Inglis (tominglis) wrote :

I've got Ubuntu 10.04. I recently reformatted my computer, and since then have not been able to get my Gobi device to work. I installed the correct backport and put the firmware in /lib/firmware/gobi. The firmware doesn't load automatically or manually.

I have tried loading the firmware manually using "sudo /lib/udev/gobi_loader-31-generic -2000 /dev/ttyUSB0 /lib/firmware/gobi", but I get:

"Failed to open serial device: : No such file or directory
usage: /lib/udev/gobi_loader-31-generic [-2000] serial_device firmware_dir"

Revision history for this message
Kenrick Bingham (loxo) wrote :

This still seems to be broken on Ubuntu 10.04.2 LTS x86_64. This worked about a year ago, but stopped working with some kernel upgrade, probably somewhere near 2.6.32-25. The device does not show up in Network Manager, and "sudo modprobe qcserial" gives the error:

FATAL: Error inserting qcserial (/lib/modules/2.6.32-33-generic/updates/wwan/qcserial.ko): Invalid module format

Is this perhaps a packaging problem? I think it would be important to restore the functionality of this device.

Revision history for this message
Stefan Bader (smb) wrote :

The module failing to load is indeed a packaging problem. I opened bug #814186 to track that getting fixed.

steve coleman (scoleman)
affects: linux (Ubuntu Maverick) → Ubuntu Maverick
Revision history for this message
Sugunakar (sugunakar2-2) wrote :

@Tom Inglis (tominglis) wrote on 2011-05-11:
It is working for me with "sudo /lib/udev/gobi_loader /dev/ttyUSB0 /lib/firmware/gobi" after copying firmware files in /lib/firmware/gobi

thanks. after lot of struggle it is working.
How to send and recieve the messages, through network manager or other application in ubuntu to check the balance?

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.