Are there WinModem drivers around for a Agere Systems AC'97?

Message
Author
koolie
Posts: 518
Joined: Mon 12 Mar 2007, 06:38

#21 Post by koolie »

Hi Richard,
Thanks for taking the time to look into this.

Puppy detects ttySL0 at boot.

CarrierCheck=unchecked.
Using StupidMode.
Using the Probe option of PupDial = detects OK.
Clicking ttySL0 button = not detected.
Only using Puppy 4.00.
Tried to get 8 or 10 Linux distros to run on this lappie,
but only Puppy 2.17.1, Puppy 4.00 and SliTaz would run.
Booted with the parameter "loglevel=7" .. results attached.

I dont have Puppy 3.01 to try.

Hope you can make some sense out of all this.



Cheers.
Attachments
logwvdialprobe.gz
(339 Bytes) Downloaded 608 times
bootsysinit.log.gz
(757 Bytes) Downloaded 611 times
messages.gz
(6.28 KiB) Downloaded 616 times
pupdial.log.gz
(123 Bytes) Downloaded 599 times

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

#22 Post by rerwin »

koolie,
Thanks for the data. I used to google for more insight. The plot is beginning to thicken, now.

I found a 2004 message from the person responsible for the slmodem package, which gives us some hope:
From: Sasha Khapyorsky sashak@smlink.com
Date: Wed, 4 Feb 2004 19:04:05 +0200
To: marvstod@rcn.com
Subject: Re: [Fwd: Re: [M5457 AC-Link Modem]]


On Wed, 04 Feb 2004 09:25:15 -0500
Marvin Stodolsky <marvstod@rcn.com> wrote:

> =======
> From below, can you interpret:
>
>
> > slamr: mc97 codec is ADS74
> > slamr: cannot init card.
> > slamr: module license 'Smart Link Ltd.' taints kernel.
> > slamr: SmartLink AMRMO modem.
> > slamr: probe 10b9:5457 SL1800 card...
> > slamr: mc97 codec is ADS74
> > slamr: cannot init card.
>
> what is ADS74 ???

ADS74 is audio codec. Modem codec is not detected at all - known slamr' behaviour with 10b9:5457.

Actual status of this device is 'almost unsupported', at the moment I have very different reports:
- works (reason why I don't remove it from pci_table)
- doesn't work at all (sometime with crash on startup - ATD,ATA)
- works only after warm reboot from win

Main problem is that 'slamr' was not developed to support this device, but similar SL1800 (10b9:5459) - single AC97 codec modem pci card. Some non-trivial efforts are needed for 'slamr'. Also currently I don't have such machine (actually one very old exists in company, but it is unavailable for next one or two months).

Hope I will find some time when machine will be available and add support for this modem to 'intel8x0m' ALSA driver (more or the less as it is done for audio).
If you may find anybody who is ready to do it - excellent (I've tried, but without success). Also bad thing here is that this is almost undocumented hardware - spec draft exists under NDA.

Regards,
Sasha.
This suggests to me that we are using the wrong driver module. If Sasha followed through and updated the ALSA driver snd_intel8x0m, it should work for this modem.

First, would you verify that driver snd_intel8x0m is loaded (lsmod) already -- or what sound drivers (snd_something) are loaded. If it is loaded, that should be all we need.

Instead of adding to /etc/rc.d/MODULESCONFIG the line:

Code: Select all

slamr 0x000010b9 0x00005457 #v3.98 Agere Systems AC97 Modem
be sure it is removed from the file. If you are running on your old pup_save environment, just reboot to see what happens. (We need the /etc/init.d/slmodem (or zzslmodem) modem script to run, and it should already be in that directory.) Alternatively, if you run with a new environment (pfix=ram), then in a console, enter:

Code: Select all

modprobe slamr
rmmod slamr
That should copy slmodem into /etc/init.d, for next boot; reboot and save, to test this out.

As a last resort, if the ALSA driver is not already loaded, you could try changing the added MODULESCONFIG entry to:

Code: Select all

snd_intel8x0m  0x000010b9 0x00005457 #Agere Systems AC97 ALSA Modem
If Sasha never made the update he mentioned, then we are probably out of luck. I saw many, many reports of people having the same or similar problems, but Sasha's message was the only one that answered their requests for help.
Richard

koolie
Posts: 518
Joined: Mon 12 Mar 2007, 06:38

#23 Post by koolie »

@Richard,

I followed your suggestions to the letter, with no positive outcome.
However, I detected one thing which seems to be hindering this.
The native sound module, snd_ali5451 keeps loading even when blacklisted,
and it won't remove with rmmod.
This, I presume, is preventing snd_intel8x0m (which is loaded) from functioning.

Any suggestions?

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

#24 Post by rerwin »

koolie,
That other module, snd_ali5451, is also considered an SL modem. We could modify the modem script to check it before checking for snd_intel8x0m. Maybe Sasha put the fix in ali5451 (fingers crossed).

But before doing anything, check out what I just found about this modem, dated 2006_April_26:
== Checking PCI IDs through modem chip suppliers ==

Vendor 10b9 is Acer Labs, producing highly integrated motherboards and Ali components.
The tight integration unfortunately ofter blocks identification of the modem chipset.
Desired information may be gained by using a COMM console under MS Windows,
and using ATI commands to elicit chipset and driver information.
10b9:5450 ALI 5450 and 10b9:5451 ALI 5451 are controllers for unsupported "sound modems"


Providing detail for device at 0000:00:09.0
with vendor-ID:device-ID
----:----
Class 0703: 10b9:5457 Modem: ALi Corporation M5457 AC'97 Modem Controller (prog-if 00 [Generic])
SubSystem 161f:2027 Rioworks: Unknown device 2027
Flags: medium devsel, IRQ 5
Checking for IRQ 5 sharing with modem.
XT-PIC ide1


-----PCI_IDs------- --CompilerVer-
Feature List: Primary Subsystem Distr KernelVer kernel default CPU
./scanModem test 10b9:5457 161f:2027 debian_version 2.6.12-9-386 3.4.5 4.0.2 i686
The following two Root commands should set up the modem.
sudo modprobe snd-intel8x0m
sudo slmodemd --alsa -c YOUR_COUNTRY modem:1
Get the SLMODEMD.gcc3.tar.gz from http://linmodems.technion.ac.il/packages/smartlink/

The controller: 10b9:5457 ALI 5457 AC-Link
is capable of supporting soft modem chips from AT LEAST manufacturers:
Pctel
Conexant
Intel
Smartlink
The Subsystem PCI id does not itself identify the modem Codec.
Checking for autoloaded ALSA modem drivers
snd_intel8x0m 16836 0
snd_ac97_codec 72188 2 snd_intel8x0m,snd_ali5451
snd_pcm 78344 4 snd_intel8x0m,snd_ali5451,snd_ac97_codec,snd_pcm_oss
snd 48644 11 snd_intel8x0m,snd_ali5451,snd_ac97_codec,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer
snd_page_alloc 10120 2 snd_intel8x0m,snd_pcm
snd_ali5451 22596 2
snd_ac97_codec 72188 2 snd_intel8x0m,snd_ali5451
snd_pcm 78344 4 snd_intel8x0m,snd_ali5451,snd_ac97_codec,snd_pcm_oss
snd 48644 11 snd_intel8x0m,snd_ali5451,snd_ac97_codec,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer

Driver snd-intel8x0m may enable codec acquisition
=== Begin mc97 codec query ===

Files under /proc/asound/ do not include modem codec information.
If more decisive information is not given below,
this is a tentative signature of a Conexant hsfmodem. Download a hsfmodem package from
http://www.linuxant.com/drivers or your installation CDs. For SuSE/Novell, install both
km_hsfmodem and hsfmodem packages, for drivers and configuration tools respectively.

=== End mc97 codec query ===

Beginning check for older ac97_codec modems.
An older ac97_modem codec was not detected.


Information on several modem chipset providers is provided below,
because ambiguities remain on the correct choice of supporting software.

== Checking PCI IDs through modem chip suppliers ==
== Checking PCI IDs through modem chip suppliers ==

Vendor 10b9 is Acer Labs, producing highly integrated motherboards and Ali components.
The tight integration unfortunately ofter blocks identification of the modem chipset.
Desired information may be gained by using a COMM console under MS Windows,
and using ATI commands to elicit chipset and driver information.
10b9:5450 ALI 5450 and 10b9:5451 ALI 5451 are controllers for unsupported "sound modems"


These messages may aid setup of soft modems under 10b9:M5457 controllers:
http://linmodems.technion.ac.il/archive ... 02518.html
http://linmodems.technion.ac.il/archive ... 02100.html
The slmodem-2.9.9 support was developed for 10b9:5459,
but there a range of reports the related 10b9:5457 modemd controllers:
fully functional;
functional only after a power on reboot from Microsoft windows;
hang/crash upon initiation of modem usage.

10b9:5457 Modem: ALi Corporation [M5457 AC-Link Modem]
SubSystem 1179:0001 Toshiba America Info Systems: Unknown device 0001
has an AgereSoftModem chip which may be supported by the Smartlink slmodem-2.9.9 driver


======= PCI_ID checking completed ======
This suggests that the presence of both snd drivers is correct. Before changing the script, let's gather more info. Since loglevel=7 lets puppy boot up, use it and capture the /var/log/messages and /tmp/bootsysinit.log, and attach them. Also, see if (1) slmodemd is running as a process (System>Pprocess process manager), which would show that the init script worked; (2) see if the /dev/modem and /dev/ttySL0 links are set up. Maybe that absence of /dev/ttySL0 is causing a problem (since it may not be created when snd-intel8x0m are loaded); if not there, try creating an empty file /dev/ttySL0 and rebooting..
Richard

koolie
Posts: 518
Joined: Mon 12 Mar 2007, 06:38

#25 Post by koolie »

Requested files attached.

slmodemd is not running.
/dev/modem and /dev/ttySL0 are there as broken links to /dev/pts/1.

Running slmodemd manually, it shows that it needs slamr loaded to create /dev/slamr0,
so I would presume that slamr should be loaded.
Attachments
bootsysinit.log.gz
(818 Bytes) Downloaded 616 times
messages.gz
(5.98 KiB) Downloaded 606 times

User avatar
rerwin
Posts: 2017
Joined: Wed 24 Aug 2005, 22:50
Location: Maine, USA

#26 Post by rerwin »

koolie,
Thanks for the data. What can you tell me about the attempt to load "#slamr"?
loading module #slamr
snd_intel8x0m
FATAL: Module #slamr not found.
That appears to be a bug, unless you entered that somewhere. There is no "#" in the module name. If puppy is picking that up internally, we may have something to work on!

I detect that you were not running puppy 4.00. To minimize the confounding factors in our troubleshooting, could you test with a P4.00 (kernel 2.6.21.7) system without all the added hardware I see in the data? Can you remove all PCcards (assuming the modem is built-in)? Can you do without the server package that I assume is associated with the error messages:
/etc/init.d/wwwoffled: line 15: [: =: unary operator expected
/etc/init.d/wwwoffled: line 20: gprintf: command not found
/etc/init.d/wwwoffled: line 21: daemon: command not found
Who knows whether that affects our results.

Just to be sure we're in sync, leave out any blacklisted or added modules through the BootManager. We are depending on the fact that the modem-initialization script /etc/init.d/slmodem exists, which should detect the intel8x0m driver and act appropriately. There might be a bug in that code, since I have never been able to test with an ALSA modem; now is our chance! I have done extensive work on the 4.00 modem support, so know it best. Someday maybe it can be back-ported it to 3.01+.

So far, I think, the script has been there as a a result of the loading of slamr, however that got done. On a new, clean 4.00 system it won't be there automatically. To deal with that, when you boot up the clean system, type in a console:

Code: Select all

modprobe slamr
rmmod slamr
When you reboot, the script should be in place and run. When booted up, the Connection Wizard should show the modem as ttySL0, if things work right, which I doubt. At that point, capture /tmp/bootsysinit.log and /var/log/messages.

After that, try the PupDial Probe option, just to see what happens. Then capture /tmp/logwvdialprobe. I don't expect much from the probe, but you never know.

Remember, even when we resolve this stuff, you may still have to use Windows to initialize the modem, But we need to be sure we have the correct driver, before giving up. Thanks for your continued help.
Richard

koolie
Posts: 518
Joined: Mon 12 Mar 2007, 06:38

#27 Post by koolie »

Richard,
I had noticed the "#slamr" and wondered where that came from.
There is something odd there if you were detecting other tha P4.00 running.
I have only been running P4.00(k2.6.21.7) for the last 2 months. Maybe that came from some recycled script from P2.17.1.

I have now done a fresh clean installation which I will keep for testing this.
When you said to leave out any blacklisted or added modules through the boot manager, I couldn't work out whether you meant
to remove the default ones, or not to add new ones. I left the default ones.

Did modprobe slamr and rmmod slamr, rebooted and /etc/init.d/slmodem exists.
Captured /tmp/bootsysinit.log and /var/log/messages..

Ran modem-wizard..

# modem-wizard
error: mdm setup: cannot open dev `/dev/slamr0': No such device
error: cannot setup device `/dev/slamr0'
modem-stats: Can't open /dev/ttySL0 for reading and writing.
Usage: modem-stats [-c command] [-e end_string] device
killall: modem-stats: no process killed

PupDial probe detects ttySL0. Clicking "ttySL0" button.. can't find modem.
Captured /tmp/logwvdialprobe.

There is undoubtedly a better chance of sorting this out now with a clean setup.

Here's hoping.

Cheers.

Edit: I must have been having a blonde spasm when I wrote that,
I have corrected it so it makes sense.
Attachments
messages.gz
(5.79 KiB) Downloaded 611 times
logwvdialprobe.gz
(337 Bytes) Downloaded 618 times
bootsysinit.log.gz
(437 Bytes) Downloaded 616 times

Post Reply