Ian,
RE: am I supposed to be doing something at the moment?
Following Jacques suggestion:
re: >>>> Once only, sudo ln -sf /dev/ttyAGS3 /dev/ttyS16
>>>> S16 exists and is never used, so choose S16 and link with a force (the f
>>>> in -sf) to override its current definition
>>>> Try wvdialconf again to see if before hanging (or not) it will find your
>>>> modem as /dev/ttyS16 or not.
====
The test we would like your to do is:
$ sudo ./unloading
$ sudo modprobe agrserial
to load the drivers.
$ sudo ln -sf /dev/ttyAGS3 /dev/ttyS16
so that wvdialconf will follow /dev/ttyS16 --> /dev/ttyAGS3 , the real port.
Then once more
$ sudo wvdialconf wvtest.txt
The possibility Jacques raises is that something other than the query
to /dev/ttyAGS3
is causing your System to hang. If it still hangs under this test, we
know modem is indeed the culprit.
MarvS
On Wed, Aug 31, 2011 at 10:01 AM, Ian Smith <ismith@xxxxxxxxxxxxxx> wrote:
> Jacques,
>
> My apologies, but am I supposed to be doing something at the moment? I
> would hate to think through my oversight I am causing a delay!
>
> Kind Regards,
>
>
> Ian
>
> On 29/08/11 16:50, Jacques Goldberg wrote:
>> Thanks Marv for the clarification.
>> Please look at this relevant section circa lines 485 ssq of file
>> wvmodemscan.cc in wvdial 1.61:
>> static int fileselect(const struct dirent *e)
>> {
>> return !strncmp(e->d_name, "ttyS", 4) // serial
>> || !strncmp(e->d_name, "ttyLT", 5) // Lucent WinModem
>> || !strncmp(e->d_name, "ttyACM", 6) // USB acm Modems
>> || !strncmp(e->d_name, "ttyUSB", 6) // Modems on USB RS232
>> || !strncmp(e->d_name, "ircomm", 6) // Handys over IrDA
>> || !strncmp(e->d_name, "ttySL", 5); // SmartLink WinModem
>>
>> // (no internal ISDN support) || !strncmp(e->d_name, "ttyI", 4);
>> }
>>
>> ttyS is the first accepted pattern.
>> My suspicion about the need in this version for the end of ttyS to be
>> completed with digits, hence the suggestion to try S16, comes from the
>> fifth pattern with ttySL. Who needs it if ttyS followed by letter L is
>> treated with the first pattern?
>>
>> Jacques
>>
>>
>> On 08/29/2011 05:13 PM, Marvin Stodolsky wrote:
>>> Just to clarify issues
>>>
>>> 1) loading agrmodem is typically trouble free
>>>
>>> 2) loading agrserial dynamically creates /dev/ttyAGS3 (so terribly
>>> named by Agere)
>>>
>>> 3) the scripts provided by my agrsm-tools package
>>> a) preload agrmodem, upon
>>> $ sudo modprobe agrserial
>>> b) create the /dev/ttySAGR --> /dev/ttyAGS3 symbolic link, for later
>>> use in the wvdialconf test.
>>>
>>> It is when modem port usage is actually begun, for example by
>>> wvdialconf, that some systems have a conflict. This seems to be
>>> happening on Ian's system,
>>> as all the /dev/ttySn (n a number) are first queried. /dev/ttySAGR
>>> would be next in line,
>>> and than is when the hang seems to occur.
>>>
>>> Certain doing Jacques suggested test with the /devttyS16 -->
>>> /dev/ttyAGS3 symbolic link is worthwhile, though I expect this is
>>> Resource Conflict case that is difficult to resolve,
>>>
>>> MarvS
>>>
>>>
>>>
>>> On Mon, Aug 29, 2011 at 9:44 AM, Jacques Goldberg
>>> <Jacques.Goldberg@xxxxxxx> wrote:
>>>> Ian,
>>>>
>>>> I am getting old. I just suddenly remember that I got mad, long ago, trying
>>>> to help somebody, with wvdialconf hanging up on my own computer with no
>>>> modem installed. I got rid of the problem using an older version of
>>>> wvdialconf. The difference was that the older would only scan /dev/ttyS*
>>>> addresses while the newer one also looked for USB, ACM, ISDN and perhaps
>>>> other devices.
>>>> Remembering that, I just downloaded the source code for wvdialconf.
>>>> Analyzing it I already have three trivial questions to ask you, please
>>>> forgive me for such stupid questions:
>>>> 1-Which version are you using? The command wvdialconf --version
>>>> gives the answer
>>>> 2-When the computer hangs, just where you at least tried to give a carriage
>>>> return, have you tried to give CTRL C (that is hit C while holding down
>>>> the CTRL key)?
>>>> 3-This one comes from an other user: have you tried to wait, doing nothing
>>>> after having started wvdialconf , for at least one timed minute, a feeling
>>>> of the time elapsed is not enough, just to make sure that it does not come
>>>> back to life without intervention? That user had observed that the additions
>>>> beyond the basic /dev/ttyS* pattern cause wvdialconf to silently work for
>>>> quite a bit of time before reporting null results.
>>>>
>>>> >From the photo there is something puzzling. It shows a device named
>>>> /dev/ttyAGS3 (yellow), probably created when loading the driver, linked to
>>>> an object /dev/ttySAGR for the obvious and unique purpose of respecting the
>>>> old convention of naming serial ports according to
>>>> /dev/ttyS<something> . If my hint that wvdialconf hangs AFTER having
>>>> completed scanning for /dev/ttyS* , and therefore having failed to find
>>>> /dev/ttySAGR, I can see at least two reasons: the driver does not correctly
>>>> create /dev/ttyAGS3 , or, the version of wvdialconf which you use insists
>>>> that /dev/ttyS<something> requires a number withing some range for
>>>> <something>.
>>>>
>>>> Elder readers who remember the early days of the Lucent driver LTMODEM
>>>> from which is some way your driver is derived will understand why in your
>>>> place I would give this a try:
>>>> Once only, sudo ln -sf /dev/ttyAGS3 /dev/ttyS16
>>>> S16 exists and is never used, so choose S16 and link with a force (the f
>>>> in -sf) to override its current definition
>>>> Try wvdialconf again to see if before hanging (or not) it will find your
>>>> modem as /dev/ttyS16 or not.
>>>>
>>>> Jacques
>>>>
>>>>
>>>> On 08/29/2011 11:35 AM, Ian Smith wrote
>>>>
>>>> Marvin,
>>>>
>>>> Up to the wvdialconf command, the text below was from the console -
>>>> after the wvdialconf command, the laptop froze - and I hav
>>>>
>>
>>
>
