ndiswrapper-general

[Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Paulo da S. <psd...@es...>]

I am successfully using ndiswrapper with a asus wl-167g. It worked fine
except for a few ndiswrapper versions.

When in the very few times the connection breaks, in general caused by
the laptop moving, a script just shuts down the net (ifconfig) and
remove the module, reinsert it again and relaunch the net. This never
failed.

Recently my card got a HW malfunction that may occur after hours,
sometimes days. In these cases the card stops working (the led is off).
Unfortunately, if I try to remove the mdule, the rm command stays
undefinitely in execution. Another rm just says the module is busy. To
recover I need to reboot the PC.

Is there anything the developers can do to enforce the possibility of
the module removal?  I think this would be a nice improvement to
ndiswrapper.

Thanks.
Paulo

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Pavel R. <pr...@gn...>]

Hello!

On Thu, 2007-05-03 at 00:19 +0100, Paulo da Silva wrote:

> Recently my card got a HW malfunction that may occur after hours,
> sometimes days. In these cases the card stops working (the led is off).
> Unfortunately, if I try to remove the mdule, the rm command stays
> undefinitely in execution. Another rm just says the module is busy. To
> recover I need to reboot the PC.

It's very likely that you have some kind of kernel oops.  See dmesg
output.

> Is there anything the developers can do to enforce the possibility of
> the module removal?  I think this would be a nice improvement to
> ndiswrapper.

There is an option in the kernel called CONFIG_MODULE_FORCE_UNLOAD.  But
I think you have a bigger problem than just a stuck module, and the need
for reboot is justified.

-- 
Regards,
Pavel Roskin

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Paulo da S. <psd...@es...>]

Pavel Roskin escreveu:
> Hello!
>
> On Thu, 2007-05-03 at 00:19 +0100, Paulo da Silva wrote:
>
>   
>> Recently my card got a HW malfunction that may occur after hours,
>> sometimes days. In these cases the card stops working (the led is off).
>> Unfortunately, if I try to remove the mdule, the rm command stays
>> undefinitely in execution. Another rm just says the module is busy. To
>> recover I need to reboot the PC.
>>     
>
> It's very likely that you have some kind of kernel oops.  See dmesg
> output.
>   

I replaced my card for a new one and the problem didn't occur anymore.
At the time I looked at several logs, "messages" included, for all
ndiswrapper lines but nothing strange was there. I didn't look for
"oops" however.
I just posted this here in the hope that some part of the code could
cause this and be somehow identifiable.

I am using my laptop as a X terminal and these crashes/reboots are a
headache. Besides they are unpredictable and very sparse in time. So, I
am not able, at least by now (too busy), to return to my damaged card.
Sorry. May be I could find a way to use it in an old computer I have
without disturbing my work too much.

Thanks anyway for answer.
Regards.
Paulo

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Paulo da S. <psd...@es...>]

Paulo da Silva escreveu:
> Pavel Roskin escreveu:
>   
>> Hello!
>>
>> On Thu, 2007-05-03 at 00:19 +0100, Paulo da Silva wrote:
>>
>>   
>>     
>>> Recently my card got a HW malfunction that may occur after hours,
>>> sometimes days. 
Unfortunately the problem seems to have nothing to do with the card, but
with some other, probabilly HW, thing.
>>> In these cases the card stops working (the led is off).
>>> Unfortunately, if I try to remove the module, the rm command stays
>>> undefinitely in execution. Another rm just says the module is busy. To
>>> recover I need to reboot the PC.
>>>     
>>>       
>> It's very likely that you have some kind of kernel oops.  See dmesg
>> output.
>>   

>> I replaced my card for a new one and the problem didn't occur anymore.
>>     
That's what I first thought of ...
...

OK! I have reproduced the situation ... Here is the output of dmesg.

BTW, it seems the problem has nothing to do with the card! The new card
also hangs!

It may take hours (rarely 1 to 2 days) however.


If any further info is needed pls. let me know.
_________________________________________________________________________________

ndiswrapper: driver rt2500usb (Ralink,10/17/2005, 2.01.00.0000) loaded
wlan0: ethernet device ......................... using NDIS driver:
rt2500usb, version: 0x20000, NDIS version: 0x500, vendor: 'Ralink
Technology Inc.', ................conf
wlan0: encryption modes supported: WEP; TKIP with WPA, WPA2, WPA2PSK;
AES/CCMP with WPA, WPA2, WPA2PSK
usbcore: registered new interface driver ndiswrapper
usb 3-3: USB disconnect, address 2
ndiswrapper (set_essid:59): setting essid failed (00010003)
ndiswrapper (set_auth_mode:661): setting auth mode to 1 failed (00010003)
ndiswrapper (iw_set_wep:915): setting authentication mode failed (FFFFFFA1)
ndiswrapper (set_essid:59): setting essid failed (00010003)
ndiswrapper (set_auth_mode:661): setting auth mode to 1 failed (00010003)
ndiswrapper (iw_set_wep:915): setting authentication mode failed (FFFFFFA1)
ndiswrapper (set_essid:59): setting essid failed (00010003)
ndiswrapper (set_auth_mode:661): setting auth mode to 1 failed (00010003)
ndiswrapper (iw_set_wep:915): setting authentication mode failed (FFFFFFA1)
ndiswrapper (set_essid:59): setting essid failed (00010003)
ndiswrapper (set_auth_mode:661): setting auth mode to 1 failed (00010003)
ndiswrapper (iw_set_wep:915): setting authentication mode failed (FFFFFFA1)
ndiswrapper (set_essid:59): setting essid failed (00010003)
ndiswrapper (set_auth_mode:661): setting auth mode to 1 failed (00010003)
ndiswrapper (iw_set_wep:915): setting authentication mode failed (FFFFFFA1)
usbcore: deregistering interface driver ndiswrapper

_________________________________________________________________________________
Here rmmod ndiswrapper hangs until I reboot!
Is there a way to fix this, so than rmmod could remove ndiswrapper?

Thanks.
Paulo

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Pavel R. <pr...@gn...>]

On Mon, 2007-05-28 at 23:04 +0100, Paulo da Silva wrote:
> usbcore: deregistering interface driver ndiswrapper
> 
> _________________________________________________________________________________
> Here rmmod ndiswrapper hangs until I reboot!
> Is there a way to fix this, so than rmmod could remove ndiswrapper?

It should be possible.  Hanging is not a correct behavior.

There is a file called oops-tracing.txt in Linux sources under
Documentation.  It describes how to get the error messages from the
kernel.

-- 
Regards,
Pavel Roskin

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Paulo da S. <psd...@es...>]

Pavel Roskin escreveu:
> On Mon, 2007-05-28 at 23:04 +0100, Paulo da Silva wrote:
>   
>> usbcore: deregistering interface driver ndiswrapper
>>
>> _________________________________________________________________________________
>> Here rmmod ndiswrapper hangs until I reboot!
>> Is there a way to fix this, so than rmmod could remove ndiswrapper?
>>     
>
> It should be possible.  Hanging is not a correct behavior.
>
> There is a file called oops-tracing.txt in Linux sources under
> Documentation.  It describes how to get the error messages from the
> kernel.
>
>   
None of my log files report any oops! I forgot to say this in my last
post. Is there anything else I can do to help?


Regards.
Paulo

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Pavel R. <pr...@gn...>]

On Wed, 2007-05-30 at 19:06 +0100, Paulo da Silva wrote:
> Pavel Roskin escreveu:
> > There is a file called oops-tracing.txt in Linux sources under
> > Documentation.  It describes how to get the error messages from the
> > kernel.
> >   
> None of my log files report any oops! I forgot to say this in my last
> post. Is there anything else I can do to help?

The instructions in that file also apply to kernel panics, which would
not appear in the logs because the kernel won't write to any files, only
to the console.

Debugging kernel issues is hard and requires more than writing two-line
replies.

A good start would be if you post a complete and detailed description of
the problem with enough information for others to reproduce it.  It
would also help if you actually check the kernel messages on the console
when the system stops working.

-- 
Regards,
Pavel Roskin

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[James S. <Sha...@ea...>]

> I am successfully using ndiswrapper with a asus wl-167g. It worked fine
> except for a few ndiswrapper versions.
> 
> 
> Recently my card got a HW malfunction that may occur after hours,
> sometimes days. In these cases the card stops working (the led is off).
> Unfortunately, if I try to remove the mdule, the rm command stays
> undefinitely in execution. Another rm just says the module is busy. To
> recover I need to reboot the PC.

For my BCM4318, I have a button, which doubles as the led.  I actually tell my mom to push that button(the blue one) rather the power(the red one) to disconnect me from the internet and free up the bandwidth(dialup) for her to use.

When the blue button is pushed the led goes off and I have to do the following to bring it back up.

1) push the button again (the led comes back on)
2) # /etc/init.d/networking stop
3) # rmmod ndiswrapper
4) # modprobe ndiswrapper
5) # /etc/init.d/networking start

And it works again.  Alternatives include using ifconfig instead of the networking (distro specific) script.  And/or modprobe -r ndiswrapper, instead of rmmod.  Anyone know if there's a way to automate the above?  Just bringing the network down and up is not enough.

HTH

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Paulo da S. <psd...@es...>]

James Shatto escreveu:
...

> For my BCM4318, I have a button, which doubles as the led.  I actually tell my mom to push that button(the blue one) rather the power(the red one) to disconnect me from the internet and free up the bandwidth(dialup) for her to use.
>
> When the blue button is pushed the led goes off and I have to do the following to bring it back up.
>
>   
No way :-)
My card is a USB. No buttons or whatever.
I tried everything. Removing it for example. Nothing works.
The blue led remains always off after the problem occurs.
Since I cannot remove the module, I need to reboot to get the connection
back.
Whenever, in very very rare cases, the module can be removed its
reinsertion causes the card to be operational again. There is some SW
problem that is causing the module not being removable.
About the causes why the card is crashing its a very strange situation!
I have been working withount any problem for more than a year! A couple
of months ago this problem began to occur ... I thought it was a problem
with the card but after I replaced it the problem din't go away.

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[James S. <Sha...@ea...>]

> No way :-)
> My card is a USB. No buttons or whatever.
> I tried everything. Removing it for example. Nothing works.
> The blue led remains always off after the problem occurs.

Perhaps you also need to remove the usb modules.  Which you may not be able to unload if you can't unload ndiswrapper.  Or if you have need of the usb modules in such a way that they can't be unloaded.  Such as your root(/) filesystem being on a usb drive.  Just load them in the reverse order that you unloaded them in, and vise versa.

If this started after you upgraded your kernel, distro, and/or ndiswrapper, you may try downgrading one of them and see if that changes and/or fixes your problem.  If you think it's hardware related, you may try using a different usb port.  Or see if that usb device has the same issue on another machine.  You may also check /var/log/dmesg, and messages, and others for any indication as to what may be going on.

Re: [Ndiswrapper-general] Problem removing ndiswrapper-1.42

[Paulo da S. <psd...@es...>]

Pavel Roskin escreveu:
> The instructions in that file also apply to kernel panics, which would
> not appear in the logs because the kernel won't write to any files, only
> to the console.
>
> Debugging kernel issues is hard and requires more than writing two-line
> replies.
>
> A good start would be if you post a complete and detailed description of
> the problem with enough information for others to reproduce it.  It
> would also help if you actually check the kernel messages on the console
> when the system stops working.
>
>   
I have posted here everything I got!
Let me summarize ...

1. Symptoms:
    a) The USB card led goes off. The card seems dead.
    b) Trying to remove the module causes the "rm" (not the system) to
hang. The task remains active. Issuing another rm (the previous still
"running") says module is busy.
    c) Need to reboot to get back the connection.

2. Causes or ways to reproduce the situation:
    a) Unknown.
    b) Most likely: poor quality of the signal. Router malfunction (?).
Neverthless, today for example, the situation occured. After rebooting,
the quality of the signal was 60/100 (windows considers this level
"excelent"). I don't know what was the quality of the signal when the
situation occured. May be there was a temporary lost of quality.

3. Outputs/Logs:
    Only var/log/messages (no other, console or Logs messages). No oops
messages.

If you think I can provide more info, please say how.

This is the only relevant output I can get.

Jun  4 14:39:42 HAL0002 kernel: ndiswrapper: driver rt2500usb
(Ralink,10/17/2005, 2.01.00.0000) loaded
Jun  4 14:39:43 HAL0002 kernel: wlan0: ethernet device
.................... using NDIS driver: rt2500usb, version: 0x20000,
NDIS version: 0x500, vendor: 'Ralink Technology Inc.', .........F.conf
Jun  4 14:39:43 HAL0002 kernel: wlan0: encryption modes supported: WEP;
TKIP with WPA, WPA2, WPA2PSK; AES/CCMP with WPA, WPA2, WPA2PSK
Jun  4 14:39:43 HAL0002 kernel: usbcore: registered new interface driver
ndiswrapper
Jun  4 16:18:03 HAL0002 kernel: usb 3-3: USB disconnect, address 2
Jun  4 16:18:25 HAL0002 kernel: ndiswrapper (set_essid:59): setting
essid failed (00010003)
Jun  4 16:18:25 HAL0002 kernel: ndiswrapper (set_auth_mode:661): setting
auth mode to 1 failed (00010003)
Jun  4 16:18:25 HAL0002 kernel: ndiswrapper (iw_set_wep:915): setting
authentication mode failed (FFFFFFA1)
Jun  4 16:18:49 HAL0002 kernel: ndiswrapper (set_essid:59): setting
essid failed (00010003)
Jun  4 16:18:49 HAL0002 kernel: ndiswrapper (set_auth_mode:661): setting
auth mode to 1 failed (00010003)
Jun  4 16:18:49 HAL0002 kernel: ndiswrapper (iw_set_wep:915): setting
authentication mode failed (FFFFFFA1)
Jun  4 16:19:12 HAL0002 kernel: ndiswrapper (set_essid:59): setting
essid failed (00010003)
Jun  4 16:19:12 HAL0002 kernel: ndiswrapper (set_auth_mode:661): setting
auth mode to 1 failed (00010003)
Jun  4 16:19:12 HAL0002 kernel: ndiswrapper (iw_set_wep:915): setting
authentication mode failed (FFFFFFA1)
Jun  4 16:19:36 HAL0002 kernel: ndiswrapper (set_essid:59): setting
essid failed (00010003)
Jun  4 16:19:36 HAL0002 kernel: ndiswrapper (set_auth_mode:661): setting
auth mode to 1 failed (00010003)
Jun  4 16:19:36 HAL0002 kernel: ndiswrapper (iw_set_wep:915): setting
authentication mode failed (FFFFFFA1)
Jun  4 16:20:00 HAL0002 kernel: ndiswrapper (set_essid:59): setting
essid failed (00010003)
Jun  4 16:20:00 HAL0002 kernel: ndiswrapper (set_auth_mode:661): setting
auth mode to 1 failed (00010003)
Jun  4 16:20:00 HAL0002 kernel: ndiswrapper (iw_set_wep:915): setting
authentication mode failed (FFFFFFA1)
Jun  4 16:20:18 HAL0002 kernel: usbcore: deregistering interface driver
ndiswrapper

Re: [Ndiswrapper-general] [NEWS] Problem removing ndiswrapper-1.42

[Paulo da S. <psd...@es...>]

Paulo da Silva escreveu:

I think I found the reason for these failures although not sure yet.

Today, having my wifi card off, I was reading a pendrive and got the
same symptoms.
The pendrive turned its led off and stopped reading.
Sutting down my system caused the script that stops the wifi card to be
executed. It tried to unload ndiswrapper which I already had removed.
Nevertheless the rmmod hanged.

Later the card failed as usual and I saw that every rmmod to the modules
dependent on usbcore hangs.

These lead to the conclusion that the problem seems to reside on the
USB. Either my USB HW is broken or so is my k2.16.19.

PS:
In my last post I wrote "rm" instead of "rmmod". I only saw this later.
I think this was understandable as a mistype error. Sorry anyway.

Regards.
Paulo