Dapper - RT61 with wep key doesn't work

Bug #35474 reported by coffee on 2006-03-18
42
Affects Status Importance Assigned to Milestone
linux-source-2.6.15 (Ubuntu)
High
Unassigned
linux-source-2.6.17 (Ubuntu)
Undecided
Ben Collins
Dapper
Undecided
Unassigned
Edgy
Undecided
Unassigned

Bug Description

I have bought a D-Link DWL-G630. Dapper detects it and use the RT61 module.
I have configured it but when I put a wep key, I can't ping anything. Without, everything seems ok.
I am in managed mode.

if you need more information, inform me.

Gaëtan Petit (gaetanp) wrote :

i can confirm this
and now i'v no wifi T_T

Gaëtan Petit (gaetanp) wrote :

mor infos
if i set the wep key in hexa in network-admin
All seems to be ok, but connection is broken (no ping etc.)
When rebooting, it crash while GNOME is starting (before displaying the bootsplash)

If i set the key with the other mode it crash the system

NetworkManager don't recognize ra0 a an network interface

I can't use my card with ndiswrapper

Very bad ...

coffee (kaizoku-alucard) wrote :

Strange, my card is recognized by dapper as wlan0 not ra0.

I haven't test ndiswrapper because it is recognized.

I have found this: http://lists.debian.org/debian-kernel/2006/02/msg00229.html
http://forums.gentoo.org/viewtopic-t-436655-highlight-lynx.html?sid=887db4e02e9cda85b1121a5665818beb

It seems that the driver itself doesn't use wep/wpa but nothing is said in http://www.ralinktech.com/drivers/Linux/ReleaseNote-20051230

While looking for a working module, I found this:
http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=924&postdays=0&postorder=asc&highlight=rt61&start=0
and this
http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?t=999&highlight=rt61
But it seems to be only for pci card

Gaëtan Petit (gaetanp) wrote :

ok this:
http://lists.debian.org/debian-kernel/2006/02/msg00229.html

seems to be the exact descripion of my problem.
My card is pci

If i could help for anything
ask me!

Tenshu

gregrs (gregrs) wrote :

I have the same problem, except that the card won't work even without WEP. It works fine with no WEP in Breezy using the source drivers from the Ralink website. The card is detected but I cannot ping anything when configured using the Gnome networking GUI.

Thanks.

Gaëtan Petit (gaetanp) wrote :

okay nothing seems to move here
And all of us still have no wifi.

I don't know if this bug report mentionned the right package and his assign to the right people.

I must admit that i underestimated the severity of this,
It is acting like everything work well ... but not!
It make my GNOME crash on reboot

It won't work with:
default drivers
compiled drivers from ralink
compiled drivers from serial-monkey SF project
ndiswrapper

So i think this is a kernel problem,
assuming the module isn't working correctly in dapper
(compiled drivers are working in breezy so this is a regression).

I am taking the risk to increase this bug severity, and change the Package fiel from ubuntu to the actual kernel.

If i am wrong feel free to correct it,
If you want of any kind to fix this bug ASK!
I'm ready to ship my card worldwild if it could help to make it working under Ubuntu.

Tenshu

Alphager (florian-haas) wrote :

When asked about this bug, a maintainer of the rt2x00-project said that it is a bug of the legacy drivers released by Ralink. THe bug should be fixed in the rt2x00-version of the driver. (http://rt2x00.serialmonkey.com/phpBB2/viewtopic.php?p=10738#10738)

Why is that rt2x00-driver not in Ubuntu ?

Ben Collins (ben-collins) wrote :

> Why is that rt2x00-driver not in Ubuntu ?

Because it is still beta, and from first hand experience, it didn't work on my rt2500 PCI card. Even the current CVS (which I am trying to merge into 2.6.17 for edgy) is broken on my rt2500.

Peter Parkkali (peter-parkkali) wrote :

AFAIK, the driver from Ralink cannot be configured to use WEP or WPA by standard means (iwconfig etc). Instead, it's necessary to enter the key, authentication mode etc to a text file in /etc/Wireless/RT61STA/rt61sta.dat *before* loading the module. Instrucions for this can be found in the driver's readme file,

2005_1230_RT61_Linux_STA_Drv1.0.3.0.tar.gz
 -> RT61_Linux_STA_Drv1.0.3.0_200512230/Module/readme

I hope I'm not the only one who finds this a bit "unique"....

I can also confirm what was previously commented in this bug: the driver doesn't work *at all* in Dapper. The card[*] is detected, rt61 module is loaded, and it's possible to set parameters[**]. However, the "Link" light never lights up on the card, dhclient doesn't get an address, ping w/ static IPs doesn't work....

[*] D-Link DWL-G630, P/N EWLG630EU.E1, H/W ver. E1, F/W ver. 5.00
[**] Within limits of Bug #48087

I'm seeing the very same behaviour that others are seeing with the kernels from Dapper/Final.

I remember that once (say, during flight-4 or so), I could get it working which made me quite happy. It seems that the problems appeared during revisions 22 or 23 of the kernel from the ubuntu repository. :-(

I've been getting a consistent Oops whenever the card is plugged in into my Armada V300 (yes, I isolated the problem and it indeed occours during execution of functions inside the rt61 module).

Is this Oops from dmesg of any interest?

I would love to get this working as soon as possible and I even tried to look into the pool directory of many ubuntu repositories, seeing if I could get any package older to contrast with what I am seeing right now.

Oh, BTW, the problem occours with both the -386 and the -686 kernels. :-(

Again, if any extra information is needed to get this working, I would love to help fix this as fast as possible, even without encryption or anything fancy.

Thanks in advance for any help, Rogério.

Hi.

On Jun 05 2006, Rogério Brito wrote:
> I'm seeing the very same behaviour that others are seeing with the
> kernels from Dapper/Final.

I have managed to workaround this bug by using a hint from another
bugreport:

In my /etc/network/interfaces file, I use, before anything dealing with
the wireless stuff, a pre-up that just brings up the interface (the
contents of my pre-up are simply "ip link set ra0 up", minus the
quotes).

Of course, this is a dirty workaround, but it does avoid the Oops that I
have been seeing.

Hope this helps, Rogério Brito.

--
Rogério Brito : <email address hidden> : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat: http://freshmeat.net/projects/algorithms/

Ben Collins (ben-collins) wrote :

This bug may be fixed in the current edgy developement system. In an effort to track this bug to our latest distribution, it is being targeted for edgy.

Please confirm whether this bug exists in edgy. If not, then please re-attach all related output (e.g. dmesg, oops output) while under edgy.

Latest Edgy CD's can be downloaded from:

http://cdimage.ubuntu.com/releases/edgy/knot-1/

If this image does not boot from you, you can also download a current daily build from:

http://cdimage.ubuntu.com/daily-live/current/

Note that it may only be necessary to boot the LiveCD to see if your bug is fixed. In this case, you do not need to upgrade your installed system at all to confirm it. Certain bugs may require that you do an actual installation.

Changed in linux-source-2.6.17:
status: Unconfirmed → Needs Info
Alphager (florian-haas) wrote :

Installing with the image from http://cdimage.ubuntulinux.org/dvd/20060829/edgy-dvd-i386.iso
results in a lockup after the Message "Now booting system".
Using the single-user-mode, it freezes after the line
[17179573.79600] CPU0: Intel(R) Pentium(R) 4 CPU 2.60GHz stepping 07

James_pic (james-pickering) wrote :

I may have found a solution for this problem in dapper. I was having the same problem, so I downloaded the newest ralink drivers (1.0.4.0). As well as the newest kernel module, they include a driver for wpa supplicant. Using the newer module, as well as the ralink wpa supplicant driver, I was able to use WPA. I've attached an archive with the compiled module, an updated wpasupplicant package with the new driver, as well as the modified wpa supplicant source, and instructions. I'm not a programmer, so it's kinda scrappy - most of you could probably do the same much more tidily - but it seems to work.

Bruce Miller (brm0423) wrote :

I tested with the Edgy Eft knot 2 LiveCD using an IBM ThinkPad A21m (from about 2000) and an AirLink 101 AWLC2036T PCMCIA adapter which uses the RaLink RT61 driver.

Although the adapter is PCMCIA / CardBus, lsmod reports that the driver is rt61pci.

Commands which touched the interface while it was down (e.g. iwlist wlan0 scan) did not hang the terminal as has been confirmed in Dapper. They brought up the interface instead.

I was, however, unable to test for WEP functionality. I was unable to get the adapter to associate to the access point barely a metre away. Both a more modern laptop and this system running off the installed Dapper associate instantly. The error was "NO DHCPOFFERS RECEIVED".

This particular card does not require installing and editing a firmware file, a process described by reharpernc on ubuntforums.org in order to run under Dapper.

I'd just like to add a comment here.

I've experienced the same, and recompiling the source provided above and replacing the rt61.ko with the generated one has solved the problem, but there's one thing I'm unable to understand:

Booting Dapper in "Live CD mode" (without installing) WORKS PERFECTLY, no problem at all, but when installing and after rebooting, the problem raises as described. I've tested the kernel version and the module file (rt61.ko) at setup time and I find they're exactly the same. This fact has led me that the driver provided by the kernel was good, and try to find a solution somewhere in the installation.

Just for learning and my mind's health, can anybody explain this?

Gaëtan Petit (gaetanp) wrote :

According to Oscar RT61 still don't work untill you go and compile the drivers from ralink site.

I didn't try dapper in "Live cd mode" but i just try edgy Knot3 and this time it get worse my RT61 wifi card isn't recognized properly.
I get an wlan0 & wlanmaster interface and nothing comes out when i do "iwlist scan" like Bruce Mille described.

I will attach a dmesg and oops ASAP for you Ben Collins

Tenshu

Ben Collins (ben-collins) wrote :

Newer rt2x00 is going into dapper. Hopefully this will improve rt61.

Changed in linux-source-2.6.17:
assignee: nobody → ben-collins
status: Needs Info → Fix Committed
Gaëtan Petit (gaetanp) wrote :

great news
i'll test it ASAP

Thanks again Ben

André Cruz (andrefcruz) wrote :

"Newer rt2x00 is going into dapper."

Didn't you mean edgy?

Alphager (florian-haas) wrote :

I really hope it is dapper, because it is a show-stopper bug in dapper.

Simon Ruggier (simon80) wrote :

Based on the version of linux-source referenced, I think he meant edgy. I
agree though, it should go into Dapper, the rt2x00 driver is recognized by
networkmanager, whereas the first party driver is not.

On 10/4/06, Alphager <email address hidden> wrote:
>
> I really hope it is dapper, because it is a show-stopper bug in dapper.
>

Gaëtan Petit (gaetanp) wrote :

yeah i think he meant edgy too

but actually i can't see no change.
Compiling the official module work well following the method describe by the ubuntu.com wiki

Hi there, people.

Well, it seems that the situation with the rt61 cards is worse in edgy
than it was in dapper. :-(

I upgraded my laptop to edgy and the default kernel doesn't work at all
with the module rt61pci. :-( I kept the dapper kernel for now, with the
workaround that I mentioned earlier. :-(

In fact, I tried to install it from scratch from the alternernate CD for
i386 and, during the insertion of the module, a stack backtrace is
generated (I don't have the dmesg here at hand, but I can send it here,
if desired). :-(

It seems that a lot of other people have the same problem, according to
what I see in <http://ubuntuforums.org/showthread.php?t=285072>.

The problem that I seen when trying to install edgy from scratch is
exactly what is described in the first post of the forum: two interfaces
wlan0 and wmaster0, but none work. :-(

Ben, please let me know if anything else is needed.

Regards, Rogério Brito.

P.S.: I also have a wireless USB card that would be driven by the ralink
drivers. I can also test it, if desired.
--
Rogério Brito : <email address hidden> : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat: http://freshmeat.net/projects/algorithms/

lukaso (launchpad-dby) wrote :

After a lot of blood sweat and tears, I've got the RT61 card working, kinda. I downgraded to the older drivers, blacklisted, etc. However, the connection drops within an hour or so and then I have to reboot to get the connection back.

Editorial: Wireless on linux doesn't work.
1) Because it's almost impossible to find a card that will work on linux (the listing I used was the Ubuntu one which promised this card would work out of the box). Well it didn't. How many cards should I try before I find one that works.
2) Because the networks don't properly reconnect once the damn thing is working. Any glitch, movement of the machine, etc, see you later.

Sigh.

Matt Grice (mattgrice) wrote :

I can confirm that the rt61pci module is borked in edgy.

As above, two interfaces are created(wlan0 and wmaster0), the purpose of the latter is unclear.

'iwlist wlan0 scan' will find APs, but will not accept a WEP key (by 'iwconfig wlan0 enc XXXXXXXXXXXXXXX' - iwconfig reports WEP: None)

I have heard reports that compiling source modules from scratch (ie rt61.ko as opposed to rt61pci.ko) is a solution. However, The box with the RT61 has no net connection so I am having difficulty testing it ;).

Hi, Matt.

On Jan 26 2007, Matt Grice wrote:
> I can confirm that the rt61pci module is borked in edgy.

Can you even use the rt* modules successfully, even without encryption?

> As above, two interfaces are created(wlan0 and wmaster0), the purpose
> of the latter is unclear.
>
> 'iwlist wlan0 scan' will find APs, but will not accept a WEP key (by
> 'iwconfig wlan0 enc XXXXXXXXXXXXXXX' - iwconfig reports WEP: None)

Does this mean that you can actually use the card, even if it is without
encryption?

I couldn't manage to get it working (only with a dirty trick and
dapper's kernel, despite the whole system being upgraded to feisty).

--
Rogério Brito : <email address hidden> : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat: http://freshmeat.net/projects/algorithms/

Matt Grice (mattgrice) wrote :

On 27/01/07, Rogério Brito <email address hidden> wrote:
> Hi, Matt.
>
> On Jan 26 2007, Matt Grice wrote:
> > I can confirm that the rt61pci module is borked in edgy.
>
> Can you even use the rt* modules successfully, even without encryption?
>
I wasn't willing to turn off the encryption on my router to test it. I
have to be very security conscious in the area I live in :)

> > As above, two interfaces are created(wlan0 and wmaster0), the purpose
> > of the latter is unclear.
> >
> > 'iwlist wlan0 scan' will find APs, but will not accept a WEP key (by
> > 'iwconfig wlan0 enc XXXXXXXXXXXXXXX' - iwconfig reports WEP: None)
>
> Does this mean that you can actually use the card, even if it is without
> encryption?
>
The only thing I could get working is the active scan (iwlist wlan0
scan). Wouldn't associate with any APs.

> I couldn't manage to get it working (only with a dirty trick and
> dapper's kernel, despite the whole system being upgraded to feisty).
>

I'm currently recompiling my kernel so I can use the official (and
latest) drivers from RaLink. I couldn't get the rt61.ko module to
modprobe - probably because of the gcc version I'm using. So I'm
compiling them all with the same compiler - fingers crossed.

Matt

>
> Dapper - RT61 with wep key doesn't work
> https://launchpad.net/bugs/35474
>

--
Science is to computer science as hydrodynamics is to plumbing.

Should array indices start at 0 or 1? My compromise of 0.5 was
rejected without, I thought, proper consideration. (Stan Kelly-Bootle)

Hi again, Matt.

On Jan 27 2007, Matt Grice wrote:
> On 27/01/07, Rogério Brito <email address hidden> wrote:
> > Can you even use the rt* modules successfully, even without
> > encryption?
>
> I wasn't willing to turn off the encryption on my router to test it. I
> have to be very security conscious in the area I live in :)

Can't you turn off your encryption just for a second, please, to see if
the basic functionality is OK? I would love to know if others can make
it work with the basics, before going further to test the encryption.

> I'm currently recompiling my kernel so I can use the official (and
> latest) drivers from RaLink. I couldn't get the rt61.ko module to
> modprobe - probably because of the gcc version I'm using. So I'm
> compiling them all with the same compiler - fingers crossed.

Thanks. I am curious to know if you can get things working. And, if you
get, would you please describe the exact problems that you saw and that
you didn't? I think that Ben Collins would appreciate this.

> Science is to computer science as hydrodynamics is to plumbing.

I wouldn't say this about Theoretical Computer Science, which is the,
ahem, best field of Computer Science. :-)

Rogério.

--
Rogério Brito : <email address hidden> : http://www.ime.usp.br/~rbrito
Homepage of the algorithms package : http://algorithms.berlios.de
Homepage on freshmeat: http://freshmeat.net/projects/algorithms/

Matt Grice (mattgrice) wrote :

On 28/01/07, Rogério Brito <email address hidden> wrote:
> Hi again, Matt.
>
> On Jan 27 2007, Matt Grice wrote:
> > On 27/01/07, Rogério Brito <email address hidden> wrote:
> > > Can you even use the rt* modules successfully, even without
> > > encryption?
> >
> > I wasn't willing to turn off the encryption on my router to test it. I
> > have to be very security conscious in the area I live in :)
>
> Can't you turn off your encryption just for a second, please, to see if
> the basic functionality is OK? I would love to know if others can make
> it work with the basics, before going further to test the encryption.
>

I can, though it will be later on in the week.

I recompiled my kernel, and the latest official driver from raLink
with the official firmware, and it's working well with WEP. Doesn't
like starting up as the PC boots at the moment - I might have to write
a script to let it settle after initialisation before configuring the
interface though.

I wonder if it's possible to bump the new driver into Feisty?

> Thanks. I am curious to know if you can get things working. And, if you
> get, would you please describe the exact problems that you saw and that
> you didn't? I think that Ben Collins would appreciate this.

With the standard Edgy driver and kernel, two interfaces were
generated for one card - wmaster0 and wlan0 (using the rt61pci kernel
module). The commands I issued were:

iwconfig wlan0 essid "MYSSID" - this worked. ESSID was set but no
association.
iwlist wlan0 scan - this also worked and spotted my AP.
iwconfig wlan0 enc XXXXXXXXXXXXXXXXXXX - this command completed
without complaint, though a further iwconfig showed encryption was
off.

After that, I took the decision to recompile with the new driver.

>
> > Science is to computer science as hydrodynamics is to plumbing.
>
> I wouldn't say this about Theoretical Computer Science, which is the,
> ahem, best field of Computer Science. :-)

I'm currently doing a Computer Science degree, though it's more
business and IT strategy than anything else. I have yet to find the
Science part. :D

I dream of doing theoretical Computer Science... Algorithmics, P=NP?
Travelling Salesmen, Djikstra, Turing machines... That's what I'm
talking about :D
Matt

--
Science is to computer science as hydrodynamics is to plumbing.

Should array indices start at 0 or 1? My compromise of 0.5 was
rejected without, I thought, proper consideration. (Stan Kelly-Bootle)

Download full text (3.4 KiB)

Hi, Matt (once again).

On 2007-01-28, at 21:00, Matt Grice wrote:

> On 28/01/07, Rogério Brito <email address hidden> wrote:
>> Can't you turn off your encryption just for a second, please, to
>> see if
>> the basic functionality is OK? I would love to know if others can
>> make
>> it work with the basics, before going further to test the encryption.
>
> I can, though it will be later on in the week.
>
> I recompiled my kernel, and the latest official driver from raLink
> with the official firmware, and it's working well with WEP.

Nice to know that, but could you please ellaborate a little please on
the firmware part, for documentation purposes? I'll try to compile a
minimal kernel for PowerPC and I will be interested in that (and in
preparation for the ra*usb* modules too).

I do think that Ben Collins would like to know about this, as I
imagine that he would like to make things "just work" in a default
install of future versions of Ubuntu.

> Doesn't like starting up as the PC boots at the moment - I might
> have to write
> a script to let it settle after initialisation before configuring the
> interface though.

I did this with dapper's driver. I had to put, as a pre-up command,
the following:

# ip link dev ra0 up

(I don't remember the syntax correctly and I'm not using my own
computer right now to check it, even though it may be a good idea to
leave this documented, as others may "fall into the same trap").

> I wonder if it's possible to bump the new driver into Feisty?

I will try this as soon as I have some spare time (see below) and
report back.

> With the standard Edgy driver and kernel, two interfaces were
> generated for one card - wmaster0 and wlan0 (using the rt61pci kernel
> module). The commands I issued were:
>
> iwconfig wlan0 essid "MYSSID" - this worked. ESSID was set but no
> association.

Ok.

> iwlist wlan0 scan - this also worked and spotted my AP.

I think that this shouldn't be needed (under normal circusmstances,
that is).

> iwconfig wlan0 enc XXXXXXXXXXXXXXXXXXX - this command completed
> without complaint, though a further iwconfig showed encryption was
> off.

Right.

> After that, I took the decision to recompile with the new driver.

Nice. Did you use the module-assistant package or not?

>>> Science is to computer science as hydrodynamics is to plumbing.
>>
>> I wouldn't say this about Theoretical Computer Science, which is the,
>> ahem, best field of Computer Science. :-)
>
> I'm currently doing a Computer Science degree, though it's more
> business and IT strategy than anything else. I have yet to find the
> Science part. :D

Here in Brazil, we have a distinction between Computer Science and
IT. Computer Science being the thing that "makes everything work" and
IT being the "business part". I teach for both courses and, having
lots of students of mine, I see the quite different focus that they
seek.

> I dream of doing theoretical Computer Science... Algorithmics, P=NP?
> Travelling Salesmen, Djikstra, Turing machines... That's what I'm
> talking about :D

I am, obviously, quite biased towards Theoretical Computer Science,
as Graph Theory and its interactions with...

Read more...

This report was marked 'Fix Committed' a while ago and there hasn't been any activity in it recently. I'm curious if this is still and issue? If this is no longer an issue, we can go ahead and mark the status to 'Fix Released'. If this is not the case, please retest against the latest Hardy Alpha release and report back your results. Thanks in advance.

Changed in linux-source-2.6.17:
status: New → Invalid
status: New → Incomplete
Simon Ruggier (simon80) wrote :

I've seen this working properly on a fresh install of gutsy on a machine with rt61, and two others with rt2500.

Simon Ruggier (simon80) wrote :

Actually, with the rt61, I've seen it work with no encryption and with WPA, but not with WEP, so perhaps someone else should confirm that it works with WEP enabled.

Simon, thanks for the feedback. I'm just going to go ahead and mark this 'Fix Released' as there was no other comments that this still is an issue. Thanks!

Changed in linux-source-2.6.17:
status: Fix Committed → Fix Released
status: Incomplete → Fix Released
Changed in linux-source-2.6.15:
status: Confirmed → Won't Fix
Curtis Hovey (sinzui) on 2012-05-29
Changed in linux-source-2.6.15 (Ubuntu):
assignee: Registry Administrators (registry) → nobody
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Duplicates of this bug

Other bug subscribers