dial up modem problems

Message
Author
User avatar
peebee
Posts: 4370
Joined: Sun 21 Sep 2008, 12:31
Location: Worcestershire, UK
Contact:

Re: Agere PCI-SV92PP freeze

#21 Post by peebee »

session wrote:Wary 5 has worked with every dial-up modem I've tried but one, an Agere PCI-SV92PP.
Wary is supposed to work with this one, and indeed the connection wizard detects it,
Hi session

Which version of wary5xx are you using? Barry posted on his blog that he had problems compiling the agere drivers under k2.6.32.28 - http://bkhome.org/blog/?viewDetailed=02134

Have you tried the k2.6.31.14 version?

Cheers
Peter
ImageLxPup = Puppy + LXDE
Main version used daily: LxPupSc; Assembler of UPups, ScPup & ScPup64, LxPup, LxPupSc & LxPupSc64

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

#22 Post by rerwin »

session,
Thank you very much for posting your pmodemdiag file for me to analyze. The strange thing I see in it is that the size of the variant of the agrmodem module that was loaded does not match that of the correct variant. It appears that the variant for HDA modems is loaded, instead of the variant for the Agere modems with device ID "0620" and "0630". The question is, then, why did that happen?

It turns out that you may actually have two different Agere modems that require the agrmodem driver, but different variants of it! The loaded variant is for the built-in, HDA sound-card modem, in case there is one. If there is no modem jack for a built-in modem, then we might have a problem. Even so, one would hope that the PCI "0620" modem would prevail. But apparently not. Here is what I think is happening.

The built-in modem is found by linux first, then the PCI modem later. The "firmware" that determines the appropriate driver variant then runs and apparently does not see the PCI modem as found by linux, so defaults to the HDA variant. When the PCI modem is eventually found and its driver loaded, the "firmware" does not run again, so leaves the variant as it is.

We need a way to immobilize the processing of the HDA card, as if it had a modem component, so that the PCI modem can operate. The direct way is to deactivate the udev rule that causes the HDA modem driver to be loaded. A more self-sufficient way might be to delay the "firmware" execution a few seconds to allow linux to find the PCI modem and so indicate, so the firmware knows it is present. The first method is something you can try easily; the other is mine to figure out, with the help of you or someone with this situation.

Here is what you can try:
- Open the following file in the text editor: /etc/udev/rules.d/60-dialup-modem.rules
- Find line 110:

Code: Select all

ATTR{vendor}=="0x8086", ATTR{device}=="0x266[d8]", RUN+="/sbin/pup_event_backend_modprobe agrmodem"
- Comment it out by inserting a number sign (#) at the beginning of the line.
- Save the file.
- In pupdial, click on CHOOSE and then ERASE ( & Okay)
- Reboot and expect the modem to work.

I hope you can help me work out a general solution, but I understand if you have other priorities. To me, the problem seems to be that the PCI modem should be used, even if a built-in HDA possible-modem exists. Thanks.

Peebee, I think Barry could not compile the variant for the "048x" modems, so that should not be a factor here.

Anyone else who suspect they have a similar problem, try the above technique, but choose the udev rule that matches the hardware IDs of your HDA sound device. After entering the command:
lspci -n
look in last half of the 60...rules file for any of the IDs listed by the command and comment out that rule.
Richard

User avatar
session
Posts: 89
Joined: Mon 07 Feb 2011, 23:11
Location: Valley of the Sun

#23 Post by session »

Thank you all for the quick responses.

rerwin, thanks for the (as always) mindful reply; I will try this tomorrow and post results.

User avatar
session
Posts: 89
Joined: Mon 07 Feb 2011, 23:11
Location: Valley of the Sun

...results...

#24 Post by session »

*update* Stupid mode fixed login problem;
yet another fully functional winmodem under Puppy.*

Well, the modem dials out.

rerwin, it's a good thing I got a glance at your original post, because
your newer suggestion of commenting out the udev rule had no effect.
I hid/unhid the agrmodem modules, ran depmod and the PCI modem could then be probed.
The modem dialed out, then I got a "Name, password pair incorrect" message in wvdial
after the login prompt, an issue discussed elsewhere in the forum.
Unfortunately I can't test any of that right now.
Last edited by session on Sat 12 Mar 2011, 06:06, edited 1 time in total.

dialeddown
Posts: 20
Joined: Mon 26 Apr 2010, 00:39

#25 Post by dialeddown »

Try stupid mode.

User avatar
nitehawk
Posts: 658
Joined: Sun 13 Apr 2008, 22:30
Location: West Central Florida

#26 Post by nitehawk »

I bought a brand new USRobotics modem the other day for 20.00 and that included the USB to serial adapter
...and don't forget that if all else fails,....try Ebay! I got several used USRobotics off of there,...(OK,..one didn't work, but the others did). I think I only paid a few bucks for them. And I found a "TrendNet" at a local thrift store for $1......!!! It works fine.

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

#27 Post by rerwin »

session,
rerwin, it's a good thing I got a glance at your original post, because
your newer suggestion of commenting out the udev rule had no effect.
That is good to know. I felt the second was less intrusive and should prevent the built-in from triggering the firmware script that makes the choice. So there is more for me to think about.

I am unable to see some of the diag data I need to better understand the sequence of events during startup, because that logging is removed now. I am finishing up (I hope) a package that would restore the logging by the old "protect" function that not only sequences certain events, but also logs what happens. I hope you can try it when it is ready, so I might better automate the handling of dual potential Agere modems.
Richard

dialeddown
Posts: 20
Joined: Mon 26 Apr 2010, 00:39

#28 Post by dialeddown »

Richard, while your attention is here, in case you missed my recent thread, pupdial failed to detect the agere in qemu even though it worked outside of qemu.

Qemu has a redirect but I am not sure what port ID would work for an hda modem that is a sub-device on a sound card.

The question probably will occur again as linmodem solutions spread and more people virtualize.

Thank you

dragonxxx
Posts: 2
Joined: Thu 16 Sep 2010, 23:17

Agere PCI-SV92PP canno't reconnect

#29 Post by dragonxxx »

Hi everybody,

I am using that Agere SV92PP PCI modem with Wary 5.1.2 . Everything works fine until I disconnect for whatever reason. Then there is no way to reconnect except rebooting the computer.

Here is the message from PupDial:
--> WvDial: Internet dialer version 1.53
--> Cannot open /dev/ttyS_AGS3: Device or resource busy
--> Cannot open /dev/ttyS_AGS3: Device or resource busy
--> Cannot open /dev/ttyS_AGS3: Device or resource busy

Here are some lines from /var/log/messages.
Jun 29 22:09:21 puppypc daemon.info pppd[15352]: Hangup (SIGHUP)
Jun 29 22:09:21 puppypc daemon.notice pppd[15352]: Modem hangup
Jun 29 22:09:21 puppypc daemon.info pppd[15352]: Connect time 212.3 minutes.
Jun 29 22:09:21 puppypc daemon.info pppd[15352]: Sent 1466188 bytes, received 26150975 bytes.
Jun 29 22:09:21 puppypc daemon.notice pppd[15352]: Connection terminated.
Jun 29 22:09:21 puppypc daemon.info pppd[15352]: Exit.
Jun 29 22:09:21 puppypc user.warn kernel: ==> codecType = 0x33
Jun 29 22:09:26 puppypc user.warn kernel: ==> codecType = 0x33
Jun 29 22:09:26 puppypc user.emerg kernel: ------------[ cut here ]------------
Jun 29 22:09:26 puppypc user.crit kernel: kernel BUG at kernel/timer.c:791!
Jun 29 22:09:26 puppypc user.emerg kernel: invalid opcode: 0000 [#1] SMP
Jun 29 22:09:26 puppypc user.emerg kernel: last sysfs file: /sys/devices/pci0000:00/0000:00:04.0/net/eth0/dormant
Jun 29 22:09:26 puppypc user.warn kernel: Modules linked in: usb_storage ppp_deflate zlib_deflate bsd_comp ppp_async crc_ccitt ppp_generic slhc usbhid agrserial agrmodem snd_hda_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_seq_dummy snd_seq_oss snd_seq_midi snd_rawmidi
Jun 29 22:09:26 puppypc user.warn kernel: Pid: 15037, comm: wvdial Tainted: P (2.6.32 #1) K7SOM+
Jun 29 22:09:26 puppypc user.warn kernel: EIP: 0060:[<c102b917>] EFLAGS: 00010286 CPU: 0
Jun 29 22:09:26 puppypc user.warn kernel: EIP is at add_timer+0x5/0x11
Jun 29 22:09:26 puppypc user.warn kernel: EAX: f85857b0 EBX: f85f58d8 ECX: 00000000 EDX: 00000000
Jun 29 22:09:26 puppypc user.warn kernel: ESI: f291e998 EDI: 00000000 EBP: f29abdd4 ESP: f29abdc8
Jun 29 22:09:26 puppypc user.warn kernel: DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
Jun 29 22:09:26 puppypc user.emerg kernel: Process wvdial (pid: 15037, ti=f29aa000 task=f282a7a0 task.ti=f29aa000)
Jun 29 22:09:26 puppypc user.emerg kernel: Stack:

The remaining is contained in a file "AgereSV92PP_ModemBug.txt" that I could PM if requested.

I thought to kill some process using Pprocess but cannot locate a suitable process in that list as I am still a newbie.

If someone could tell me what process/module to kill/initialize it would be fine for me as long as I don't have to reboot each time.

On another subject, but still 56k modem, I have another PCI modem: Intel/Ambient MD5628D-L-B wich don't work in Wary 5.1.2, i.e. isn't recognized. I read somewhere it could be an hardware modem, maybe not. I found a WinXP driver but I don't have a Windows machine right now. What is the status of that modem in Wary?

Thanks everybody for all that good work,
dragomxxx

Post Reply