George,
Your criticisms may need some moderation.
What is a CODEC? See http://en.wikipedia.org/wiki/AC%2797
This kind of chip is not limited to modems, you will find it in many
audio/video systems.
And what is CODEC 0x31 ? an MS codec for GSM.
(http://www.mplayerhq.hu/DOCS/codecs-status.html#ac)
Is your modem a GSM modem at all?
AC'97 (and HDA) modems, which need a CODEC to exist, are very widely
spread counter examples of your screams about using winmodems under Linux.
One and the same driver, slmodem, in its slamr and its ALSA versions,
serves all brands of AC'97 modems except Conexant, thanks to the good
will of SmartLink which allowed using their proprietary Linux core for
their AC'97 chip with any competitor chip, all of which have disclosed
their codec information except Conexant. You see, this property of AC'97
architecture was discovered by one of our former students, Sasha
Khapyorski, in my office, when he was trying to explain what AC'97
means. Thanks to him, at least tens if not hundreds of thousands of
AC'97 modems are working under Linux.
Sure enough, we have a maintenance problem for new versions of the Linux
OS which pour in faster than a handful of volunteers can maintain the
drivers. Joining the team is usually much more effective than
deliberately killing it.
And there is nothing against linking any modem driver to /dev/modem
except when one forgets that the link only lives as long as the driver
is in memory, About the Agere modules such as AGS3 (???, strange, I
thought AGRSM,..) this is a story over 10 years old. Agere was once
Lucent. They supplied a code named LTmodem, naming the driver device
ttyLT0 . Not a bad idea, except that we extensively use a tool
wvdial which searches modems with a pattern ttyS* in souvenir with
real serial devices ttyS0 , ttyS1 , ... ttyS99 or more, Thus was the
need for a link born.
Jacques
George H wrote
It's weird.
All your suggestions are not working and they go against the actual
installation scripts that come with the driver. I am not trying to be
rude here but saying:
On your system, /dev/modem is directly linked to /dev/ttyAGS3, which
seems wrong - as if you're bypassing the softmodem application.
is miss-leading and perhaps wrong.
Thanks for you're help anyways, it helped me rule out certain options
that I did not think of from the start. I still can't find out what
the message "codecType=0x31" means. I think this might be a good clue
to help solving this.
Either way, as the docs in the kernel serial driver says "If you
intend to use a software modem (also called winmodem) under Linux,
then forget it. These modems are crippled and require proprietary
drivers which are only available under Windows."
Further more I been reading a lot of content on
http://tldp.org/HOWTO/Modem-HOWTO.html which goes into so much detail.
It seems winmodems are just a big headache. I'm scrapping out this
winmodem and going pure-serial modem.
--
George H
george.dma@xxxxxxxxx
On Wed, Mar 10, 2010 at 9:52 AM, Yair Elharrar <freezer.spam@xxxxxxxxx> wrote:
Hi George,
If the softmodem driver is up then you should be getting responses to
AT-commands, regardless of the actual state of the hardware.
I'm afraid I'm not familiar with this specific softmodem's
installation procedure, but it seems we're missing something.
The kernel module is the one connected to /dev/ttyAGS3.
The softmodem application should connect to it, and present a
/dev/modem interface to other applications.
On your system, /dev/modem is directly linked to /dev/ttyAGS3, which
seems wrong - as if you're bypassing the softmodem application.
On Wed, Mar 10, 2010 at 9:40 AM, George H <george.dma@xxxxxxxxx> wrote:
Did it.
I'm afraid I have no luck.
There is something odd that is happening. Each time I send a command
to the modem I see the line
"==> codecType = 0x31" in my /var/log/kern.log file
I wonder what that means.
I also tried using minicom while using hardware flow control and
software. Both do nothing.
Any other ideas?
--
George H
george.dma@xxxxxxxxx
On Wed, Mar 10, 2010 at 9:16 AM, Yair Elharrar <freezer.spam@xxxxxxxxx> wrote:
OK, that's a match. I would double-check /dev after removing the
serial.conf line and restarting, just to make sure the serial driver
didn't change anything; then run wvdialconf and see if it detects the
soft modem.
-Yair
On Wed, Mar 10, 2010 at 9:08 AM, George H <george.dma@xxxxxxxxx> wrote:
Ah ok.
The output of "cat /proc/devices" is
62 ttyAGS
The output of "ls -l /dev/ | grep AG" is
lrwxrwxrwx 1 root root 12 Mar 10 05:54 mmodem -> /dev/ttyAGS3
crw-rw---- 1 root uucp 62, 67 Mar 10 05:54 ttyAGS3
lrwxrwxrwx 1 root root 12 Mar 10 05:55 ttySAGR -> /dev/ttyAGS3
--
George H
george.dma@xxxxxxxxx
On Wed, Mar 10, 2010 at 9:02 AM, Yair Elharrar <freezer.spam@xxxxxxxxx> wrote:
Hi George,
That's true, but the softmodem driver is actually a fake serial port.
wvdial has to think it's a real serial port, when it isn't.
This is why I recommended to remove the line from /etc/serial.conf,
and reboot - to make sure the real serial driver doesn't attempt to
access the output port.
I asked for ls -l again because I want to make sure the device numbers
from the ls -l output match the numbers in /proc/devices.
-Yair
On Wed, Mar 10, 2010 at 8:54 AM, George H <george.dma@xxxxxxxxx> wrote:
Hi,
I don't see how this would change anything.
It says in the wvdialconf messages to see if the setserial is setting
the right parameters on the serial port. I have tried wvdial with and
without setserial.
The output of ls -l /dev | grep AG is already mentioned in my previous email.
--
George H
george.dma@xxxxxxxxx
On Wed, Mar 10, 2010 at 8:43 AM, Yair Elharrar <freezer.spam@xxxxxxxxx> wrote:
Hi George,
Oh, OK...
Anyway I would remove the /dev/ttySAGR line from /etc/serial.conf and
reboot, since we don't want the serial driver to handle this device.
Check the output of "ls -l /dev/ | grep AG" and "cat /proc/devices".
-Yair
On Wed, Mar 10, 2010 at 8:27 AM, George H <george.dma@xxxxxxxxx> wrote:
Hi Yair,
That extra 'm' was a character I forgot to erase when editing the
script log. You know the program "script" it's like a key logger
except it also logs all the colors and keyboard escape codes. So that
extra 'm' is just a left over debris of the script after editing.
Main point is., the modem does not get detected.
Thank you.
On Wed, Mar 10, 2010 at 8:24 AM, Yair Elharrar <freezer.spam@xxxxxxxxx> wrote:
George,
You seem to have an extra "m" in the /dev/modem link.
Try: ln -sf /dev/ttySAGR /dev/modem
Good luck.
-Yair
On Wed, Mar 10, 2010 at 8:16 AM, George H <george.dma@xxxxxxxxx> wrote:
Hi Marvin,
Thanks for your reply.
In the agrsm06pci-2.1.80~20100106/agrsoftmodem.c
file there are indeed recognition lines for your 047e:2892 chip set in
{ USB_DEVICE(USB_USS_VENDOR_ID, USB_USS_PRODUCT_ID_2828) },
{ USB_DEVICE(USB_USS_VENDOR_ID, USB_USS_PRODUCT_ID_2892
That's right, and I am confused as to why it still does not work.
Using the AgereMdmDaemon is sometimes troublesome. Set a symbolic link:
$ sudo modprobe /dev/ttySAGR /dev/ttyAGS3
and check if the modem is detected by:
$ sudo wvdial
Please see the script below of all the commands i've executed. I hope
it helps to explain what it is I am getting.
Script started on Wed Mar 10 05:54:10 2010
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 #./agr_softmodem start
Starting Agere Soft Modem
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # lsmod
Module Size Used by
agrserial 10096 0
agrmodem 1331844 1
fglrx 1947356 28
intel_agp 22932 0
agpgart 26564 2 fglrx,intel_agp
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # tail /var/log/kern.log
Mar 10 05:54:21 toshi kernel: * Hello Agere Driver PCIe **
Mar 10 05:54:21 toshi kernel: Loading module Agere Modem Controller
driver version 2.1.80-Red Flag NE 5.0 (2007-10-1)
Mar 10 05:54:21 toshi kernel: ttyAGS3 at I/O 0x400 (irq = 3) is a AgereModem
Mar 10 05:54:21 toshi kernel: Loading module Agere Modem Interface
driver version 2.1.80.0Red Flag 5.0 NE (2007-10-01)
Mar 10 05:54:21 toshi kernel: ==> codecType = 0x31
Mar 10 05:54:21 toshi kernel: App PID = 5068
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # ln -s
/dev/ttyAGS3 /dev/ttySAGR
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # ls -l /dev | grep AG
lrwxrwxrwx 1 root root 12 Mar 10 05:54 mmodem -> /dev/ttyAGS3
crw-rw---- 1 root uucp 62, 67 Mar 10 05:54 ttyAGS3
lrwxrwxrwx 1 root root 12 Mar 10 05:55 ttySAGR -> /dev/ttyAGS3
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # cat/etc/serial.conf
# These are the standard COM1 through COM4 devices
#
/dev/ttyS0 uart 16450 port 0x3F8 irq 4
/dev/ttyS1 uart 16450 port 0x2F8 irq 3
/dev/ttyS2 uart 16450 port 0x3E8 irq 4
/dev/ttyS3 uart 16450 port 0x2E8 irq 3
/dev/ttySAGR uart 16550A port 0x0400 irq 3
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # /etc/init.d/serial start
* Setting /dev/ttyS0 to uart 16450 port 0x3F8 irq 4 ... ok
* Setting /dev/ttyS1 to uart 16450 port 0x2F8 irq 3 ... ok
* Setting /dev/ttyS2 to uart 16450 port 0x3E8 irq 4 ... ok
* Setting /dev/ttyS3 to uart 16450 port 0x2E8 irq 3 ... ok
* Setting /dev/ttySAGR to uart 16550A port 0x0400 irq 3 ... ok
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # wvdialconf
/etc/wvdial.conf
Editing `/etc/wvdial.conf'.
Scanning your serial ports for a modem.
Modem Port Scan<*1>: S0 S1 S2 S3
ttySAGR<*1>: ATQ0 V1 E1 -- failed with 2400 baud, next try: 9600 baud
ttySAGR<*1>: ATQ0 V1 E1 -- failed with 9600 baud, next try: 115200 baud
ttySAGR<*1>: ATQ0 V1 E1 -- and failed too at 115200, giving up.
Sorry, no modem was detected! Is it in use by another program?
Did you configure it properly with setserial?
Please read the FAQ at http://open.nit.ca/wiki/?WvDial
If you still have problems, send mail to <wvdial-list@xxxxxxxxxxxx>.
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # wvdial
--> WvDial: Internet dialer version 1.60
--> Warning: section [Dialer Defaults] does not exist in wvdial.conf.
--> Initializing modem.
--> Sending: ATZ
--> Sending: ATQ0
--> Re-Sending: ATZ
--> Modem not responding.
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # tail /var/log/kern.log
Mar 10 06:00:50 toshi kernel: ==> codecType = 0x31
Mar 10 06:01:06 toshi kernel: ttyS0: LSR safety check engaged!
Mar 10 06:01:06 toshi kernel: ttyS1: LSR safety check engaged!
Mar 10 06:01:06 toshi kernel: ttyS2: LSR safety check engaged!
Mar 10 06:01:06 toshi kernel: ttyS3: LSR safety check engaged!
Mar 10 06:01:06 toshi kernel: ==> codecType = 0x31
Mar 10 06:01:11 toshi kernel: ==> codecType = 0x31
root@toshi:~/bins/linmodem/agrsm06pci-2.1.80-20100106 # exit
Script done on Wed Mar 10 06:01:32 2010
