RNDIS over USB doesn't work because of driver Bug

Bug #192411 reported by Jean-Christophe Dubois
26
This bug affects 1 person
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Fix Released
Medium
Gerold Jens Wucherpfennig
Hardy
Fix Released
Medium
Colin Ian King
Intrepid
Fix Released
Medium
Colin Ian King

Bug Description

This bug was first reported as part of bug 138583 and now is moved into its own bug.

Basically, when a WM5 smartphone was plugged into its USB cradle, the establishment of the RNDIS session to communicate with the phone would fail because of a driver bug negotiating an improper frame size with the device. the dccm daemon would therefore not detect the phone and no dialog (pls or other) was possible.

The synce project proposed a patch that would make things work (http://synce.svn.sourceforge.net/svnroot/synce/trunk/patches/linux-2.6.22-rndis_host-wm5.patch) but it was considered inadequate by the linux people. As this patch is unofficial and not part of any distribution, it would therefore require the user to recompile its kernel which is not really friendly for the average desktop user.

I submitted a first proposal to fix the problem in a more appropriate way (http://bugzilla.kernel.org/show_bug.cgi?id=8094#c37) but it was considered too much intrusive.

I then submitted a second proposal (http://www.mail-archive.com/linux-usb%40vger.kernel.org/msg01608.html) that should be integrated in the Linux kernel tree very soon (I hope).

Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

Thanks Jean,

Care to post a comment to this report when you see the patch you submitted has officially been pulled into upstream. Hardy development just entered feature freeze so this may not get fixed for Hardy. But assuming the patch makes it upstream it should get into Hardy + 1 automatically. Thanks.

Changed in linux:
status: New → Incomplete
Revision history for this message
Jean-Christophe Dubois (jcd) wrote :

The patch has now been submitted to the "netdev" mailing list for upstream inclusion (hopefully).

http://www.spinics.net/lists/netdev/msg56069.html

Revision history for this message
Laurent Bonnaud (laurent-bonnaud) wrote :

Another approach to fix this bug would be to replace the current rndis subsystem in the kernel by the rndis-lite subsystem that can be found there:

  http://www.synce.org/moin/SynceInstallation/Source

I tested both rndis modules with my HTC Touch Cruise PDAPhone that runs under Windows Mobile 6: the current one in hardy does not work but the rndis-lite does work. Since most Ubuntu users will use rndis to connect to a WM5/6 PDAPhone, making the switch would greatly improve their experience.

Revision history for this message
Laurent Bonnaud (laurent-bonnaud) wrote :

Changing the status to prevent this bug from expiring.

Changed in linux:
status: Incomplete → Confirmed
Revision history for this message
Jean-Christophe Dubois (jcd) wrote :

Laurent,

Thanks for your comment. I have not tried the rndis-lite driver myself so I can't really comment on it.

Now I believe the current RNDIS kernel driver will work for WM5/6 if my patch is applied.

The point is that the Ubuntu policy does not seem to be very found of drivers that are not integrated upstream. So far I don't think the rndis-lite driver is candidate for such thing, so it seems easier to fix the current driver to get it accepted/backported to Gusty/Hardy.

If you can try your WM6 device with the proposed patch that would be great.

Changed in linux:
assignee: nobody → leannogasawara
importance: Undecided → Medium
status: Confirmed → Triaged
Revision history for this message
Jean-Christophe Dubois (jcd) wrote :

The patch has officialy been merged into Linus 2.6 tree.

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=10d0f27c1baa4a094b4965708a15f2b0c4d65f5e

Time to backport?

JC

Revision history for this message
Luís Silva (luis) wrote :

I second this request. This would really be an advantage over windows...

Right now, besides some games (I admit... :$) and work (even though SQLDeveloper is a great tool to work with Oracle Database, SQLNavigator is still one of the best) the only thing that keeps my windows install is my windows mobile PDA... From my experience rndis devices don't work as well in windows as they do in linux with this patch applied... And this would allow me to sync my PDA with evolution wich is a great "new" feature to ubuntu!

Revision history for this message
exactt (giesbert) wrote :

@Leann

what is the status of this bug? i am asking because this bug really keeps one from using WM devices. lengthy howtos starting with "patch your kernel and rebuild it" are definitely a show-stopper. also the latest mandriva distribution is supposed to have better support for WM devices and ubuntu should be able to catch up. i would love to see a backport to hardy!

Thx

Revision history for this message
Luís Silva (luis) wrote :

This should be a release critical bug... The patch is trivial and went mainstream in 2.6.25. Without the patch, WM6 or WM6 devices don't work at all... If the patch was merged it would allow at least network sharing...

Revision history for this message
flying_rhino (wat3rgat3) wrote :

Then will this bug be fixed? Can I help with some logs - when i tried to connect my HTC WM 5 device with Hardy?

Revision history for this message
Ximenes Silva (xa-silva) wrote : Re:[Bug 192411] Re: RNDIS over USB doesn't work because of driver Bug
  • unnamed Edit (1.8 KiB, text/html; charset=iso-8859-1)

> Then will this bug be fixed? Can I help with some logs - when i tried to
> connect my HTC WM 5 device with Hardy?
>
> --
> RNDIS over USB doesn't work because of driver Bug
> https://bugs.launchpad.net/bugs/192411
> You received this bug notification because you are subscribed to Ubuntu.
>
> Status in Source Package "linux" in Ubuntu: Triaged
>
> Bug description:
> This bug was first reported as part of bug 138583 and now is moved into its own bug.
>
> Basically, when a WM5 smartphone was plugged into its USB cradle, the establishment of the RNDIS session to communicate with the phone would fail because of a driver bug negotiating an improper frame size with the device. the dccm daemon would therefore not detect the phone and no dialog (pls or other) was possible.
>
> The synce project proposed a patch that would make things work (http://synce.svn.sourceforge.net/svnroot/synce/trunk/patches/linux-2.6.22-rndis_host-wm5.patch) but it was considered inadequate by the linux people. As this patch is unofficial and not part of any distribution, it would therefore require the user to recompile its kernel which is not really friendly for the average desktop user.
>
> I submitted a first proposal to fix the problem in a more appropriate way (http://bugzilla.kernel.org/show_bug.cgi?id=8094#c37) but it was considered too much intrusive.
>
> I then submitted a second proposal (http://www.mail-archive.com/linux-usb%40vger.kernel.org/msg01608.html) that should be integrated in the Linux kernel tree very soon (I hope).
>

Changed in linux:
assignee: leannogasawara → ubuntu-kernel-team
milestone: none → ubuntu-8.04.1
Revision history for this message
flying_rhino (wat3rgat3) wrote :

Leann Ogasawara schrieb:
> ** Changed in: linux (Ubuntu)
> Assignee: Leann Ogasawara (leannogasawara) => Ubuntu Kernel Team (ubuntu-kernel-team)
> Target: None => ubuntu-8.04.1
>
>
So my WindowsMobile 5.0 device - will sync ?

Revision history for this message
Jean-Christophe Dubois (jcd) wrote :

Well flying_rhino this fix is only one piece of the equation.

Now assuming the rest of the user space/daemon is up to date on hardy, your mobile should sync.

Revision history for this message
Gavin Hamill (gdh) wrote :

Palm support in Ubuntu is good. Unfortunately Palm's market share is rapidly declining.

Support for Windows Mobile devices is currently 'good' and rapidly improving (a gvfs backend to enable nautilus browsing of WM devices now exists, for example)

The kernel patch is an utter killer for new users - and is especially annoying to have to rebuild modules every time the kernel is upgraded.

I can't give enough support to this request to backport this 6 line patch!

Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

Hi Guys,

Just a quick update. I did indeed milestone this for 8.04.1 but the final decision to backport the patches will be up to the kernel team. I just didn't want this to fall off their radar for consideration.

However, as an alternate solution, care to test the Intrepid Ibex 8.10 kernel which was most recently rebased with the upstream 2.6.25 kernel and also contains the patch referenced here. It is currently available in the following PPA:

https://edge.launchpad.net/~kernel-ppa/+archive

If you are not familiar with how to install packages from a PPA basically do the following:

Create the file /etc/apt/sources.list.d/kernel-ppa.list to include the following two lines:

deb http://ppa.launchpad.net/kernel-ppa/ubuntu hardy main
deb-src http://ppa.launchpad.net/kernel-ppa/ubuntu hardy main

Then run the command:

sudo apt-get update

You should then be able to install the linux-image-2.6.25 kernel package. Please let us know your results. Thanks.

Revision history for this message
Gavin Hamill (gdh) wrote :

I've elected not to test this, since it provides a 'quick fix' that is not in the interests of Ubuntu as a whole. It will most likely 'get it working' but does a disservice to the project.

The thought of having Ubuntu say 'no' to Windows Mobile out-of-the-box until 2011 on an LTS desktop makes me shudder.

Changed in linux:
assignee: ubuntu-kernel-team → colin-king
status: Triaged → In Progress
Revision history for this message
Colin Ian King (colin-king) wrote :

Hi,

I've put up some kernel packages (linux-2.6.24-17.32ckingppa1 with a linux-image that contains the
RNDIS fixes) in my PPA at https://launchpad.net/~colin-king/+archive

Please try this kernel and report any success/regressions. If this fixes the problem I can add the fix to Hardy 8.04.1. Thanks.

Revision history for this message
Gavin Hamill (gdh) wrote :

Hi Colin,

Thanks greatly for that. I've been running it for a while now whilst plugging/unplugging various other USB devices (printers/cameras/scanners...) with no noticeable negative effects. All seems well and obviously the WM device is working just fine. :)

Revision history for this message
Mark Ellis (mark-mpellis) wrote :

Yes thanks for that Colin, I too can confirm it's working wonderfully..

And many thanks Jean-Christophe.

Changed in linux:
status: In Progress → Fix Committed
Revision history for this message
Colin Ian King (colin-king) wrote :

SRU justification:

Impact:

When a WM5 smartphone is plugged into its USB cradle, the establishment
of the RNDIS session to communicate with the phone fails because of a
driver bug negotiating an improper frame size with the device. The dccm
daemon would therefore not detect the phone and no dialog is possible.

The attached patch is relatively simple and been backported from
upstream:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=10d0f27c1baa4a094b4965708a15f2b0c4d65f5e

Testing: User has tested against patched kernel, plugging and
unplugging various other USB devices (printers/cameras/scanners...) with
no noticeable negative effects. See above comment.

Changed in linux:
assignee: nobody → colin-king
importance: Undecided → Medium
milestone: none → ubuntu-8.04.1
status: New → Triaged
Steve Langasek (vorlon)
Changed in linux:
milestone: ubuntu-8.04.1 → none
Changed in linux:
status: Triaged → Fix Committed
Revision history for this message
Martin Pitt (pitti) wrote :

Accepted into -proposed, please test and give feedback here

Revision history for this message
Steve Langasek (vorlon) wrote :

Can folks please confirm that the 2.6.24-19 kernel package in hardy-proposed works with their hardware?

Revision history for this message
Gavin Hamill (gdh) wrote :

Sorry - yes -19 is working great here - happily sync'ing away with no sign of any badness from other devices

Revision history for this message
exactt (giesbert) wrote :

so far everything seems to work fine here with -19 kernel. thx a lot!

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

Copied to hardy-updates.

Changed in linux:
status: Fix Committed → Fix Released
Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

Marking this "Fix Released" against Intrepid kernel. Thanks.

Changed in linux:
status: Fix Committed → Fix Released
Revision history for this message
Gerold Jens Wucherpfennig (gjwucherpfennig) wrote :

I get this error with Ubuntu 10.10 when plugging in my Windows Mobile 6.0 device
on Linux Kernel 2.6.35-26-generic:

[211505.040318] usb 1-3: new high speed USB device using ehci_hcd and address 14
[211555.590082] rndis_host 1-3:1.0: RNDIS init failed, -110
[211555.590120] rndis_host: probe of 1-3:1.0 failed with error -110
[211560.592797] rndis_wlan 1-3:1.0: RNDIS init failed, -110
[211560.592854] rndis_wlan: probe of 1-3:1.0 failed with error -110

Changed in linux (Ubuntu):
assignee: Colin King (colin-king) → Gerold Jens Wucherpfennig (gjwucherpfennig)
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

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