Re: USB-Modem, slusb and kernels later than 2.6.24

Dear Bernd,


I am not trying to criticize what you do, nor do I wish to start a quarrel.

I want to make my best to help you and find where your problem is located.

This is exceedingly easy while being "hands on" the computer, and 
exceedingly difficult while being limited to the information supplied by 
the person in trouble,

Here is your log with my comments flagged by leading >>>

bernd@LRK:~/src/slmodem-2.9.11-20110321> sudo winmodusb
Running modprobe ungrab-winmodem
DONE

 >>> because of bash  syntax error ( = in a test must have a space on 
its left and a space on its right, both missing in your winusb, the 
 >>>output of the test is always TRUE thus your DONE does not prove that 
this modprobe was "done".

Running modprobe slusb

 >>>this one too ens with a misleading DONE (below, near the end) for 
the same reason. No PROOF that slusb was loaded

 >>>the following command was probably typed by you while waiting for 
your script to end, unfortunately you did not quote who >>>introduced it 
(it is NOT in the script, which is why I risk a GUESS that it was 
introduced by you.

insmod /lib/modules/2.6.37.1-1.2-default/extra/slusb.ko

 >>>and you killed your script winmodusb  while it was active at line 
17, but clearly modprobe -v slusb was never completed'

/home/bernd/bin/winmodusb: Zeile 17:  7953 Getötet                
modprobe -v slusb

 >>> A few line below you get a comforting DONE (due to a programming 
error in the script) which makes you believe that slusb was >>> loaded 
into memory ... , unfortunately you did not check its presence in memory 
after your insmod command

 >>> as you should do by     lsmod|grep slsub    instead of a line in a 
script.


 >>> now plenty of system messages some but not necessarily all 
occurring during the failing modprobe

Message from syslogd@LRK at May 20 20:24:29 ...

  kernel:[  441.892465] Oops: 0000 [#1] SMP


Message from syslogd@LRK at May 20 20:24:29 ...
  kernel:[  441.892472] last sysfs file: /sys/module/usbcore/initstate

Message from syslogd@LRK at May 20 20:24:29 ...
  kernel:[  441.892655] Process modprobe (pid: 7953, ti=ea64a000 
task=e54bee30 task.ti=ea64a000)

Message from syslogd@LRK at May 20 20:24:29 ...
  kernel:[  441.892661] Stack:

Message from syslogd@LRK at May 20 20:24:29 ...
  kernel:[  441.892704] Call Trace:

Message from syslogd@LRK at May 20 20:24:29 ...
  kernel:[  441.893068] Code: 28 00 00 00 00 0f 88 9e 03 00 00 8b 44 24 
14 03 44 24 18 89 44 24 10 73 12 8b 54 24 14 c7 44 24 10 ff ff ff ff f7 
d2 89 54 24 18 <0f> b6 06 8b 5c 24 14 84 c0 74 7c 8d 54 24 24 89 f0 e8 
51 ea ff

Message from syslogd@LRK at May 20 20:24:29 ...
  kernel:[  441.893123] EIP: [<c03f8fc9>] vsnprintf+0x49/0x400 SS:ESP 
0068:ea64bd54

Message from syslogd@LRK at May 20 20:24:29 ...
  kernel:[  441.893135] CR2: 0000000000000000

 >>>the second misleading DONE from the script

DONE

Making node /dev/slusb0

 >>>The node correctly created

Running slmodemd -c GERMANY /dev/slusb0 &
bernd@LRK:~/src/slmodem-2.9.11-20110321> error: mdm setup: cannot open 
dev `/dev/slusb0': No such device or address
error: cannot setup device `/dev/slusb0'
 >>>and the final message that slmodem cannot start because slusb has 
not been loaded in the memory.
bernd@LRK:~/src/slmodem-2.9.11-20110321>

 >>>>


Please always remember that Linux is a true multitasking system, several 
commands can run simultaneously, and the order in which output messages 
appear is independent of the order in which the commands were submitted, 
unless you carefully wait for the completion of each command before 
giving the next. Specially when sleep's are introduced in a script.


Which is why I humbly tried to discourage you from using scripts for 
debugging purposes.


I am sorry for these bad news but you are still before slusb has 
successfully been loaded.

I have a bet (educated bet, from the System messages) but it is no more 
than a bet.

You  have a wrong set of USB libraries in your machine, older or newer 
than what the code in slusb assumes.


Jacques


On 05/22/2011 03:21 PM, Bernd Weber wrote:

> Hi Jacques,
>
> this script worked fine for the other Linux-Versions. So I saw no
> necessity to debug it.
> Besides, the error is caused by the "modprobe -v slusb" - command,
> executed by the script.
>
> Of course I had read the README-file in the tarball. It is of little
> use. There is nothing said about the option USB=1, for example.
> I don't simply follow the instructions in the README, because they don't
> just work.
> I saw the mistakes in the script now. They are of little importance for
> the failure I described, because the essential commands as ungrab and
> modprobe slusb are executed nevertheless.
> In the script:
> modprobe ungrab-winmodem
> modprobe -v slusb
> mknod -m 600 /dev/slusb0 c 242 0
>
> As you can see, in the essential lines you don't have a confusion
> between node, or module. The insertion of the compiled module just
> throws an exception. After that the USV-Port was blocked, so I had to
> reboot my computer tp get my external USB-Drive to be recognized again
> at that Port.
>
>
> As I remember I got the script from the technion site some years ago.
> Because it worked I saw no need to debug it.
>
> Greetings
>
> Bernd
>
> Jacques Goldberg schrieb:
>> Bernd,
>>
>>
>> Whoever the author of the script which you used, winmodusb, you should
>> not use it. Instead of debugging your modem driver, this turns into
>> debugging a script which does anything else than what the author
>> intended it to do. The script as you sent it is full of mistake, such as
>> quotes where back quotes are meant, slusb0 (a device) confused with
>> slusb (a program), and more.
>>
>> I know it is tiring to type again and again a sequence of commands.
>>
>> Why would you not simply follow the instructions in the README file of
>> the tarball which you used?
>>
>> I did it on an 32 bits laptop which unfortunately does not have the
>> right Smartllink or compatible modem but a Conexant, and everything went
>> smooth until slmodemd tried to start the wrong modem.
>>
>>
>> By the way at least while debugging, running slmodemd in the background
>> may become a cause of headache as the simple way to stop a session is by
>> killing slmodemd with CTRL-C.
>>
>>
>> My former remark about powering the modem is not in conflict with your
>> previous usage of the modem with an other computer and/or another
>> version of Linux. Simply, we have seen here more than a handful of users
>> not understanding why the driver would not start, because they had
>> forgotten to power up their external modem. The pseudodevice such as
>> /dev/slusb0 will not exist if the modem is not connected and powered,
>> because the system then does not know about its existence. I have made
>> the mistake enough times myself to have thought of that as a possible
>> cause of your troubles.
>>
>>
>> Jacques
>>
>>
>> On 05/20/2011 10:26 PM, Bernd Weber wrote:
>>
>>> Hi Antoni, hi Jacquea,
>>>
>>> thank you very much. I forgot that depmod thing. This was the best way
>>> to come to the next errors. Please find attached what the syslogd said
>>> and the shell-script I used to activate the modem.
>>>
>>> Besides: the modem was powered and is O.K. I know this, because I use it
>>> on another machine with SuSe 10.2.
>>>
>>> Regards
>>>
>>> Bernd
>>>
>>> Jacques Goldberg schrieb:
>>>> On 05/20/2011 04:56 PM, Antonio Olivares wrote:
>>>>
>>>>> Bernd,
>>>>>
>>>>> On Fri, May 20, 2011 at 7:13 AM, Bernd Weber<[email protected]>
>>>>> wrote:
>>>>>> Hi Antonio,
>>>>>>
>>>>>> thank you for the link. The slusb.ko is compiled now. But a "make
>>>>>> install" only installs the slamr.ko into the
>>>>>> /lib/modules/kernel/extra-directory. Even when I copy the module
>>>>>> manually, modprobe doesen't find the module:
>>>>>> modprobe slusb
>>>>>>
>>>>>> Do you have any ideas?
>>>>> run
>>>>>>>>>> of course with the modem connected and powered up, something
>>>> which might have been skipped while fighting with the software...
>>>>> # depmod -a
>>>>> or
>>>>> $ sudo depmod -a
>>>>> then
>>>>> $ sudo modprobe slusb
>>>>> or
>>>>> # modprobe slusb
>>>>>
>>>>> Regards,
>>>>>
>>>>>
>>>>> Antonio
>>>>>> Regards
>>>>>>
>>>>>> Bernd
>>>>>>
>>>>>> Antonio Olivares schrieb:
>>>>>>> Bernd,
>>>>>>>
>>>>>>> Please check the following pages out:
>>>>>>>
>>>>>>> http://www.mail-archive.com/[email protected]/msg883918.html
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> http://www.mail-archive.com/[email protected]/msg883985.html
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> http://www.mail-archive.com/[email protected]/msg884183.html
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Check them out and see if you can get it working on your system?
>>>>>>>
>>>>>>> Regards,
>>>>>>>
>>>>>>> Antonio
>>>>>>>
>>>>>>> On Thu, May 19, 2011 at 3:51 PM, Bernd Weber<[email protected]>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Hi Antonio,
>>>>>>>>
>>>>>>>> thanks a lot. This option carried me a step further - to the next
>>>>>>>> error.
>>>>>>>> Please find enclosed the compilation and the error-log as well.
>>>>>>>> In the tar-ball enclosed are different patch-file. But there is no
>>>>>>>> documentation to what file or files these shold be applied and in
>>>>>>>> which
>>>>>>>> cases.
>>>>>>>>
>>>>>>>> Regards
>>>>>>>>
>>>>>>>> Bernd
>>>>>>>>
>>>>>>>> Antonio Olivares schrieb:
>>>>>>>>
>>>>>>>>> Bernd,
>>>>>>>>>
>>>>>>>>> Try to use a special command make USB=1
>>>>>>>>>
>>>>>>>>> $ sudo make USB=1
>>>>>>>>> or
>>>>>>>>> $ make USB=1
>>>>>>>>> the slusb as well as the slamr driver will be compiled.  Try that
>>>>>>>>> and
>>>>>>>>> report back.
>>>>>>>>>
>>>>>>>>> Regards,
>>>>>>>>>
>>>>>>>>> Antonio
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> On Thu, May 19, 2011 at 10:22 AM, Bernd
>>>>>>>>> Weber<[email protected]>    wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> Jacques,
>>>>>>>>>>
>>>>>>>>>> I am using a 32-bits machine.
>>>>>>>>>>
>>>>>>>>>> the URL:
>>>>>>>>>> http://linmodems.technion.ac.il/packages/smartlink/slmodem-2.9.11-20110321.tar.gz
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> My steps:
>>>>>>>>>>
>>>>>>>>>> tar xzvf slmodem-2.9.11-20110321.tar.gz
>>>>>>>>>>
>>>>>>>>>> make
>>>>>>>>>> make install
>>>>>>>>>>
>>>>>>>>>> And then I learned that slusb.ko wasn't compiled. I couldn't
>>>>>>>>>> find the
>>>>>>>>>> reason. On some pages I got hints that there might be problems
>>>>>>>>>> with
>>>>>>>>>> kernels later than 2.6.24.
>>>>>>>>>> May be I should have done further configuration efforts for
>>>>>>>>>> slusb. But I
>>>>>>>>>> didn't find hints what I should do. So I have the choice to
>>>>>>>>>> experiment a
>>>>>>>>>> few days or ask questions ;)
>>>>>>>>>>
>>>>>>>>>> Greetings Bernd
>>>>>>>>>>
>>>>>>>>>> Jacques Goldberg schrieb:
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>> Bernd,
>>>>>>>>>>>
>>>>>>>>>>> Please give us the URL of the file which compiled slamr and
>>>>>>>>>>> failed
>>>>>>>>>>> slusb - I just would like to give it a try on my machine
>>>>>>>>>>> (because I had the necessary hardware in my office, I happend
>>>>>>>>>>> to watch
>>>>>>>>>>> Sacha, the author of the drivers, compiling and testing them ,
>>>>>>>>>>> and I
>>>>>>>>>>> am curious to see what fails and where)
>>>>>>>>>>> Also, are you using a 32 bits or a 64 bits machine?
>>>>>>>>>>>
>>>>>>>>>>> Jacques
>>>>>>>>>>>
>>>>>>>>>>> On 05/19/2011 05:50 PM, Bernd Weber wrote:
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>> Hi,
>>>>>>>>>>>>
>>>>>>>>>>>> the last kernel slusb worked with was 2.6.18.2-34-default,
>>>>>>>>>>>> Suse 10.2
>>>>>>>>>>>> with slmodem-2.9.10.
>>>>>>>>>>>>
>>>>>>>>>>>> I compiled the latest version on SuSe 11.4. But slusb wasn't
>>>>>>>>>>>> compiled,
>>>>>>>>>>>> only slamr. I tediously searched the site of linmodem, but
>>>>>>>>>>>> didn't find
>>>>>>>>>>>> an answer whats to do. SuSe 11.4 comes with kernel 2.6.37
>>>>>>>>>>>>
>>>>>>>>>>>> The Modem in question is a Fujitsu Siemens USB-Modem.
>>>>>>>>>>>> As I said, it works fine with older kernels and older
>>>>>>>>>>>> Versions of
>>>>>>>>>>>> slmodem.
>>>>>>>>>>>>
>>>>>>>>>>>> If there is some programming to be done, please let me know.
>>>>>>>>>>>> Though I am
>>>>>>>>>>>> much better in Perl than in C/C++, if necessary I would try to
>>>>>>>>>>>> solve the
>>>>>>>>>>>> problems.
>>>>>>>>>>>>
>>>>>>>>>>>> Could somebody give me a short briefing?
>>>>>>>>>>>>
>>>>>>>>>>>> Greetings
>>>>>>>>>>>>
>>>>>>>>>>>> Bernd
>>>>>>>>>>>>
>>>>>>>>>>>>