[Home]

Summary:DAHLIN-00135: [patch] Echo with dahdi-linux 2.2.0 and wcte12xp w/VPMADT032
Reporter:Isaac Gal (isaacgal)Labels:
Date Opened:2009-08-15 17:03:08Date Closed:2010-05-21 13:08:30
Priority:MajorRegression?No
Status:Closed/CompleteComponents:dahdi (the module)
Versions:2.2.0.2 Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) mantis-15724-2.patch
( 1) putty.log
Description:Note: I've changed the name of this issue since the original issue was a duplicate of 0015498 and this issue has become more about echo, and not about HDLC aborts. -sruffell

Original description

going back to 2.1.0.4 and the problem is gone.

this issue was reported by alecdavid in 0015498 and it is NOT fixed and should not have been merged into 2.2.0.2.


****** ADDITIONAL INFORMATION ******

the console log is full of this:

Aug  3 15:01:13 pbx last message repeated 2 times

Aug  3 15:01:14 pbx kernel: Unable to set SW Companding on channel 29 (reason 4)

Aug  3 15:01:23 pbx last message repeated 7 times

Aug  3 15:01:23 pbx asterisk[26687]: NOTICE[26726]: chan_dahdi.c:9041 in pri_dchannel: PRI got event: HDLC Abort (6) on Prim$

Aug  3 15:01:23 pbx asterisk[26687]: NOTICE[26726]: chan_dahdi.c:9041 in pri_dchannel: PRI got event: HDLC Abort (6) on Prim$

Comments:By: Alec Davis (alecdavis) 2009-08-16 16:54:28

isaacgal:

The developers are going to need to know which hardware this is related to, as the issue in our case ASTERISK-1532498 at that site was fixed in the wcte12xp driver.<pre>
console output:
chch1*CLI> dahdi show status
Description                              Alarms  IRQ    bpviol CRC4   Fra Codi Options  LBO
Wildcard TE121 Card 0                    OK      22     0      0      CCS HDB3 CRC4     0 db (CSU)/0-133 feet (DSX-1)
chch1*CLI>
</pre>
By: Isaac Gal (isaacgal) 2009-08-16 23:16:08

pbx*CLI> dahdi show status
Description                              Alarms     IRQ        bpviol     CRC4
Wildcard TE121 Card 0                    OK         1          0          0

[root@pbx ~]# cat /proc/interrupts
          CPU0       CPU1
 0:  264241332  283085498    IO-APIC-edge  timer
 1:          3          2    IO-APIC-edge  i8042
 8:          0          1    IO-APIC-edge  rtc
 9:          1          0   IO-APIC-level  acpi
58:     483771          0         PCI-MSI  eth0
169:  341716886  322870984   IO-APIC-level  wcte12xp0
225:     881849          0         PCI-MSI  ahci
233:    2607675    2609185   IO-APIC-level  libata
NMI:          0          0
LOC:  538857926  538857814
ERR:          0
MIS:          0


dahdi: Telephony Interface Registered on major 196
dahdi: Version: 2.2.0.2
ACPI: PCI Interrupt 0000:02:08.0[A] -> GSI 16 (level, low) -> IRQ 169
wcte12xp: VPMADT032 Loading firmware... <6>wcte12xp: Success
wcte12xp: VPM present and operational (Firmware version 107)
wcte12xp: Setting up global serial parameters for E1
wcte12xp: Found a Wildcard TE121
dahdi_transcode: Loaded.
dahdi: Registered tone zone 19 (Israel)
wcte12xp: Span configured for CCS/HDB3/CRC4

server is HP DL120, Card TE121B.

hope it helps, i know it is an issue with 2.2.0.2 since going back to 2.1.0.4 with the exact same hardware totaly eliminate the problem and the symptoms.

regards

isaac
By: Shaun Ruffell (sruffell) 2009-08-17 06:47:49

isaacgal:  Something is not right.  If you're using DAHDI version 2.2.0.2, you should see the VPM firmware version 1.17.

Could you reinstall 2.2.0.2, power off your computer, power it back on, and then post the dmesg output from when the drivers have loaded?
By: Jason Parker (jparker) 2009-08-20 12:37:15

Is this system from a pre-packaged distro?  How are you installing DAHDI?  The distro may not be installing the firmware properly.
By: Trey Blancher (tblancher) 2009-09-18 10:46:15

I have a customer with the same problem, but with wctdm24xxp.  dmesg is as follows:

dahdi: Telephony Interface Registered on major 196
dahdi: Version: 2.2.0.2
ACPI: PCI Interrupt 0000:03:0b.0[A] -> GSI 23 (level, low) -> IRQ 177
Port 1: Installed -- AUTO FXO (FCC mode)
Port 2: Installed -- AUTO FXO (FCC mode)
Port 3: Installed -- AUTO FXO (FCC mode)
Port 4: Installed -- AUTO FXO (FCC mode)
Port 5: Installed -- AUTO FXO (FCC mode)
Port 6: Installed -- AUTO FXO (FCC mode)
Port 7: Installed -- AUTO FXO (FCC mode)
Port 8: Installed -- AUTO FXO (FCC mode)
VPM100: Not Present
dahdi_vpmadt032_loader: module license 'Digium Commercial' taints kernel.
eth0: no IPv6 routers present
Booting VPMADT032
VPMADT032: Present and operational (Firmware version 117)
Found a Wildcard TDM: Wildcard TDM800P (8 modules)
dahdi: Registered tone zone 0 (United States / North America)
dahdi_echocan_mg2: Registered echo canceler 'MG2'
-- Setting echo registers:
-- Set echo registers successfully
-- Setting echo registers:
-- Set echo registers successfully
-- Setting echo registers:
-- Set echo registers successfully
-- Setting echo registers:
-- Set echo registers successfully
-- Setting echo registers:
-- Set echo registers successfully
-- Setting echo registers:
-- Set echo registers successfully
-- Setting echo registers:
-- Set echo registers successfully
-- Setting echo registers:
-- Set echo registers successfully
Bluetooth: HIDP (Human Interface Emulation) ver 1.1
Unable to set SW Companding on channel 4 (reason 4)  # Repeated several hundred times

Edit:  The original problem was that inbound calls always fail, and outbound calls had one way audio.  Disabling the VPM (vpmsupport=0) works around the problem.


By: Shaun Ruffell (sruffell) 2009-09-18 11:27:43

tblancher:  Your issue is different since you're getting an error about being unable to set SW Companding.  This is when I say you probably want to contact Digium Technical support for assistance with this issue.
By: Isaac Gal (isaacgal) 2009-09-22 16:05:56

sorry, my mistake, i took it output from the wrong log file.
i just reinstalled the 2.2.0.2 and this is the dmesg output:

wcte12xp: VPM present and operational (Firmware version 117)
wcte12xp: Setting up global serial parameters for E1
ACPI: PCI interrupt for device 0000:02:08.0 disabled
PCI: Enabling device 0000:02:08.0 (0350 -> 0353)
ACPI: PCI Interrupt 0000:02:08.0[A] -> GSI 16 (level, low) -> IRQ 169
wcte12xp: VPM present and operational (Firmware version 117)
wcte12xp: Setting up global serial parameters for E1
wcte12xp: Found a Wildcard TE121
dahdi_transcode: Loaded.

the distro is based on elastix but i removed all rpms and built asterisk/dahdi from source since elastix repos wasnt updated enough at the time i installed the box. needed a few issues resolved urgently.

anyway,. the systems looks stable now with no HDLC errors. last time it happend about 12 hours after restarting asterisk, it begins with horrible noise on all channels, inbound calls dropped and outbound calls are 50/50 chance of droping.

i enabled intense logging so lets wait till it happens again and i will post the logs.

thanks
By: Isaac Gal (isaacgal) 2009-09-23 05:44:58

ok. about 10 hours later the calls started dropping, incoming audio was totaly distoreted while outgoing was fine.

i attached a cli log with the pri issues. its a bit big but it has all the details you need.

again, going back to 2.1.0.4 and all issues are gone.
By: Shaun Ruffell (sruffell) 2009-09-23 10:09:19

isaacgal: Was there anything in the kernel that correlates to when the calls started dropping and the audio was corrupted?
By: Isaac Gal (isaacgal) 2009-09-23 11:14:58

nope. /var/log/messeges and dmesg got nothing speciel in them.
this isnt a kernel issue in my opinion, going back to 2.1.0.4 and everything is back to normal.

all version were built from source.

thanks
By: Shaun Ruffell (sruffell) 2009-09-23 11:21:01

isaacgal:  I agree that I don't think it was a general kernel version....I was just wondering if there was output from the drivers at the time the audio went bad (bumps in latency, failures to SW companding, etc..).  My recommendation for you too now is to contact Digium Technical support.
By: Isaac Gal (isaacgal) 2009-09-23 11:38:07

sruffell: this isnt just audio going bad. the whole E1 is down, lots of HDLC erros and making outbound/inbound calls is almost impossible.

btw: removing the VPMADT032 and the system is running great. i had the same issue in two differnt servers, one was DELL R200 and the current one is HP DL120. so i'm almost sure there is an issue with the VPMADT032 + E1 and dahdi 2.2.0.2.
By: Shaun Ruffell (sruffell) 2009-09-23 11:45:05

isaacgal:  Yes...for some reason it appears that after your system is running the VPMADT032 starts companding on the dchannel as if it were audio.  That is why you get the aborts and the E1 goes down.   The question is why is this happening on your system.

VPMADT032 + E1 and 2.2.0.2 does work.  The 2.2.0.2 release was specifically made for this (see issue DAHLIN-126).  So it isn't a problem in principal (at least all indications are that it isn't).


By: Isaac Gal (isaacgal) 2009-09-23 11:55:01

sruffell: I have a feeling DAHLIN-126 is very similar to my case, since everything was great before moving to 2.2.0.2 and all errors disapear when going back to 2.1.04 all indications show that there are still issues with VPMADT032 + E1 and 2.2.0.2 that needs digiums attention.

what do you suggest i do now?
By: Shaun Ruffell (sruffell) 2009-09-23 12:06:28

isaacgal: DAHLIN-126 was resolved with the 2.2.0.2 release of dahdi-linux.  It was broken when that user went to 2.2.0 from 2.1.0.4.

My suggestion for you at this point is to open a case with Digium Technical support.


By: viniciusfontes (viniciusfontes) 2009-10-21 14:16:15

I'm also having the exact same problem. My setup is almost identical: TE122P + VPMADT032, DAHDI-Linux 2.2.0.2, E1, PRI.

Already contacted tech support and they asked me to update libpri and check cabling/grouding. They also recommended running patlooptest, which returned a lot of errors. Tried replacing the TE122P board, same thing.

I was running DAHDI-Linux 2.2.0.1 and had fewer problems after upgrading to 2.2.0.2, but the only way to completely solve it is removing the VPMADT032.
By: Darren Franks (ghomf) 2009-10-29 12:17:29

I had the same problem with dropped calls. I had to replace the server hardware twice. Now I have a TE121P and I have lots of echo. I think the latest DAHDI is definately broken.
By: Shaun Ruffell (sruffell) 2009-10-29 12:27:58

Ghomf:  What was the asterisk/libpri/dahdi(or zaptel version) that previously worked in your setup?
By: Darren Franks (ghomf) 2009-10-29 12:31:10

When I first started getting dropped calls , I was using the latest zaptel . Now I am using Asterisk 1.4.26.2 and Dahdi Version 2.2.0.2+2.2.0 and Libpri 1.4.10.1. Only once I changed the actual server hardware did I fix the HDLC Abort issue. But now I get lots of echo.I have several digium case id's : Case ID: UBD-432482


By: Shaun Ruffell (sruffell) 2009-10-29 12:33:44

Ghomf:  And you have a hardware echo canceller on your TE121P?
By: Darren Franks (ghomf) 2009-10-29 12:43:00

Yes ,i have a hardware echo canceller on my TE121P. And set to 1024 taps and still echo.
By: Alec Davis (alecdavis) 2009-11-09 02:39:58.000-0600

We also have echo with the TE121B (with VPMADT032):
  dahdi Linux 2.2.0.2 (version released to fix DAHLIN-126).
  asterisk (SVN-branch-1.6.1-r226977)
  libpri (SVN-branch-1.4-r1249)

We also have 2 other sites running TE220B (with VPMOCT064) (wct2xxp driver) with no echo problems at all, running;
  asterisk (SVN-branch-1.6.1-r226977)
  libpri (SVN-branch-1.4-r1249)
  dahdi 2.2.0 (R6707).

What is repeatable with the site with the TE121B is that after a dahdi stop and start (including asterisk), all calls out of the TE121 are fine, but as soon as 1 call comes into the TE121, echo becomes apparent on every further outbound call.

The test I have setup is to a telecom voice mail service, where it asks you for your number.
Sequence of events:
1). ensure you have the line to yourself, after hours helps.
2). unload asterisk, dahdi, load dahdi then asterisk.
3). Call telecom voice mail (multiple times), dialing DTMF digits when required. All will sound fine, as long as only outbound calls have been made since restart.
4). Using Cellphone, ring into the TE121, unimportant if echo (afterhours no-one there)
5). Call telecom voice mail (again), dialing DTMF digits when required, Echo will be there (digits are echo'd back), for each and every subsequent call!
6). unload asterisk, dahdi, load dahdi then asterisk.
7). Call telecom voice mail service, echo has gone.

I've repeated the above sequence the versions below, with the results as above:
DAHDI trunk (R7533)
DAHDI Branch 2.2 (R7533)
DAHDI 2.2.0.2 (R7032)
DAHDI 2.2.0 (R6707) (with patch for compander)
By: Alec Davis (alecdavis) 2009-11-09 04:52:47.000-0600

Going back to dahdi 2.1.0.4

Asterisk SVN-branch-1.6.1-r228695
libpri version: SVN-branch-1.4-r1261
DAHDI Version: 2.1.0.4 Echo Canceller:

Echo is gone.

Obviously, not a configuration we want to stay with.
By: Shaun Ruffell (sruffell) 2009-11-09 13:54:03.000-0600

alecdavis:  We're talking about a different issue than what this is about, but do you think you could try the head of the 2.2 branch now?  There is an updated firmware image for the VPM module there. [EDIT: disregard....I see that you tried revision 7533 in the trunk]


By: Alec Davis (alecdavis) 2009-11-09 15:18:11.000-0600

What version of firmware image should I be expecting /var/log/syslog below.
This was only a dahdi unload then reload.
Echo is now present again :-(

Nov 10 10:07:33 chch1 kernel: dahdi: Telephony Interface Registered on major 196
Nov 10 10:07:33 chch1 kernel: dahdi: Version: SVN-branch-2.2-r7541
Nov 10 10:07:33 chch1 kernel: PCI: Enabling device 0000:03:08.0 (0000 -> 0003)
Nov 10 10:07:33 chch1 kernel: ACPI: PCI Interrupt 0000:03:08.0[A] -> GSI 16 (level, low) -> IRQ 169
Nov 10 10:07:41 chch1 kernel: wcte12xp 0000:03:08.0: Booting VPMADT032
Nov 10 10:07:50 chch1 kernel: wcte12xp: VPM present and operational (Firmware version 117)
Nov 10 10:07:50 chch1 kernel: wcte12xp: Setting up global serial parameters for E1
Nov 10 10:07:50 chch1 kernel: wcte12xp: Found a Wildcard TE121
Nov 10 10:07:50 chch1 kernel: dahdi: Registered tone zone 10 (New Zealand)
Nov 10 10:07:50 chch1 kernel: wcte12xp: Span configured for CCS/HDB3/CRC4
By: Shaun Ruffell (sruffell) 2009-11-09 15:25:24.000-0600

It's 1.20 is what you should see....but I see that you already were trying the latest version when you tried the revision of trunk you did.  I'm going to try and setup your test case in the lab here.
By: Alec Davis (alecdavis) 2009-11-09 15:28:21.000-0600

for the record from syslog, 3 different Firmwares have been loaded.

Nov  9 17:11:39 chch1 kernel: wcte12xp: VPM present and operational (Firmware version 117)
Nov  9 17:24:13 chch1 kernel: wcte12xp: VPM present and operational (Firmware version 120)
Nov  9 17:34:37 chch1 kernel: wcte12xp: VPM present and operational (Firmware version 117)
Nov  9 23:17:21 chch1 kernel: wcte12xp: VPM present and operational (Firmware version 107)

Firmware 107, which got loaded last night, is from DAHDI 2.1.0.4
By: Isaac Gal (isaacgal) 2009-11-09 15:34:40.000-0600

just for the record as well:

something has defiantly changed between 2.1.0.4 and 2.2.0.2 when dealing with E1 trunks.

the only way i can run a stable system with 2.2.0.2 is to remove the VPM Module.and rely on SW EC to do the job.

kernel log has nothing special, i changed hardware as well. no improvement.
By: Darren Franks (ghomf) 2009-11-09 16:40:57.000-0600

I agree . I have disabled the VPM module and SW EC works great , with the VPM module enabled , lots of echo.
By: Shaun Ruffell (sruffell) 2009-11-10 03:30:08.000-0600

isaacgal:  Are you still having the abort messages, or just problems with echo?

others:  If you're having echo problems with 2.2.0.2...could you try the trunk of dahdi-linux and apply the mantis-15724-1.patch file?
By: Shaun Ruffell (sruffell) 2009-11-10 04:36:06.000-0600

deleted mantis-15724-1.patch and replaced it with mantis-15724-2.patch.  The first one potentially could lock your system up due to double freeing memory.
By: Alec Davis (alecdavis) 2009-11-10 05:02:21.000-0600

mantis-15724-2.patch looks safe, it's very similar to wctdm24xxp/base.c echocan_create

mantis-15724-1.patch worked for me, except for reload, will try mantis-15724-2.patch when box is back up.

Thanks Shaun, huge effort.
By: Digium Subversion (svnbot) 2009-11-10 09:52:13.000-0600

Repository: dahdi
Revision: 7549

U   linux/trunk/drivers/dahdi/wcte12xp/base.c

------------------------------------------------------------------------
r7549 | sruffell | 2009-11-10 09:52:12 -0600 (Tue, 10 Nov 2009) | 9 lines

wcte12xp: Export features and operations for VPMADT032.

Fixes a regression in dahdi-linux 2.2.0 where it was impossible for userspace to
reset the state of a channel in the VPM.

(issue DAHLIN-135)
Patches:
     mantis-15724-2.patch uploaded by sruffell (license 456)
     Tested by: alecdavis
------------------------------------------------------------------------

http://svn.digium.com/view/dahdi?view=rev&revision=7549
By: Digium Subversion (svnbot) 2009-11-10 09:52:56.000-0600

Repository: dahdi
Revision: 7550

_U  linux/branches/2.2/
U   linux/branches/2.2/drivers/dahdi/wcte12xp/base.c

------------------------------------------------------------------------
r7550 | sruffell | 2009-11-10 09:52:55 -0600 (Tue, 10 Nov 2009) | 17 lines

Merged revisions 7549 via svnmerge from
https://origsvn.digium.com/svn/dahdi/linux/trunk

........
 r7549 | sruffell | 2009-11-10 09:46:18 -0600 (Tue, 10 Nov 2009) | 9 lines
 
 wcte12xp: Export features and operations for VPMADT032.
 
 Fixes a regression in dahdi-linux 2.2.0 where it was impossible for userspace to
 reset the state of a channel in the VPM.
 
 (issue DAHLIN-135)
 Patches:
       mantis-15724-2.patch uploaded by sruffell (license 456)
       Tested by: alecdavis
........

------------------------------------------------------------------------

http://svn.digium.com/view/dahdi?view=rev&revision=7550
By: viniciusfontes (viniciusfontes) 2009-11-10 09:54:34.000-0600

Just for clarification: this bugfix was for the VPM cancelling echo on the B-channel (and bringing the entire E1 down) or those echo issues?
By: Shaun Ruffell (sruffell) 2009-11-10 09:55:17.000-0600

The patch reloaded all night here, and definitely looked warranted so I committed it to trunk and merged into the 2.2 branch.  I left this issue open while I wait for more responses....but feel free to test either trunk or the head of the 2.2 branch instead of applying the patch.


By: Shaun Ruffell (sruffell) 2009-11-10 09:56:40.000-0600

viniciusfontes: The echo issues.  My belief is that the B-channel issues are resolved and I haven't (yet...) been able to reproduce them.
By: Shaun Ruffell (sruffell) 2009-11-10 10:10:26.000-0600

I changed the name of this issue to reflect what this has become.  Any discussion about D-channel problems should be on 0015498.  Sorry...I realize I should have just opened a new issue before attaching the patch.
By: viniciusfontes (viniciusfontes) 2009-11-10 10:13:34.000-0600

0015498 is closed. I would like to report that the problem still happen on 2.2.0.2. Is there a way to reopen it or should I open a new bug report?
By: Shaun Ruffell (sruffell) 2009-11-10 10:27:09.000-0600

Ok...I reopened DAHLIN-126.


By: Isaac Gal (isaacgal) 2009-11-10 10:43:46.000-0600

Hallelujah.......Hallelujah .........

Hallelujah

i guess you needed more victims before reopening the investigation.

i also want to report that the dahdi faxdetect does not work with 2.2.0.2 as well. when dahdi detects fax tone it does not turn of EC.

ie "FaxHandled=yes but EC currently on" when doing core show channel XX, both with VPM and with OSLEC.

i also noticed the /var/log/messages does not include "dahdi detected a fax tone (XX) on channel XX and turned of echo cancellation on channel..." was this kernel message removed on purpose?

Isaac Gal
By: Shaun Ruffell (sruffell) 2009-11-10 12:29:15.000-0600

isaacgal:  same problem with fax in the head of trunk currently?  Most likely are related.
By: Alec Davis (alecdavis) 2009-11-10 22:05:42.000-0600

Although the patch was obviously needed, we still have echo :(

Going back again to Dahdi 2.1.0.4 echo is non existant.
By: Shaun Ruffell (sruffell) 2009-11-11 10:42:17.000-0600

:( indeed.  What may be necessary to really pin down where/what the problem is between 2.1.0.4 and 2.2.0.2 is to make a patch against the current trunk to allow use of the 1.07 version of the firmware.

alecdavis:  Are you getting the echo regularly like you were before?  Is there something else you can suggest that will allow me to simulate your environment?

I'll update this issue (hopefully within a day or two) when there is something new to test.
By: Alec Davis (alecdavis) 2009-11-11 15:08:43.000-0600

With trunk, it's with every call, and can still get it to do it when calling the Telecom Voice mail IVR (DTMF keys echo'd back).

Our setup at the TE121 site is relativly simple, but not simplist, a Jtec ISDN switch is between TE121 and Telecom ISDN connection.
SIP <-> ASTERISK <-> TE121 <-> [E1 - Jtec ISDN Switch - BRA] <-> BRA ISDN Telecom

The sites with the TE220B are similar.
SIP <-> ASTERISK <-> TE220B <-> [E1 - Jtec ISDN Switch - E1] <-> PRI ISDN Telecom

For me to test though, its over an IAX trunk between sites.
Needless to say, users at the TE121 site, dialing out to their local area also have echo, so IAX isn't involved.
By: Digium Subversion (svnbot) 2009-11-12 13:28:11.000-0600

Repository: dahdi
Revision: 7565

U   linux/trunk/drivers/dahdi/voicebus/GpakCust.c
U   linux/trunk/drivers/dahdi/voicebus/GpakCust.h
U   linux/trunk/drivers/dahdi/wctdm24xxp/base.c
U   linux/trunk/drivers/dahdi/wcte12xp/base.c

------------------------------------------------------------------------
r7565 | sruffell | 2009-11-12 13:28:10 -0600 (Thu, 12 Nov 2009) | 7 lines

voicebus: Fix race when enabling/disabling hardware echocan.

This closes a race condition where it was possible for the driver to
believe it has enabled the VPMADT032 when in fact, it really has not.
This fixes a regression introduced in dahdi-linux 2.2.0.

(issue DAHLIN-135)
------------------------------------------------------------------------

http://svn.digium.com/view/dahdi?view=rev&revision=7565
By: Digium Subversion (svnbot) 2009-11-12 14:08:28.000-0600

Repository: dahdi
Revision: 7572

_U  linux/branches/2.2/
U   linux/branches/2.2/drivers/dahdi/adt_lec.c
U   linux/branches/2.2/drivers/dahdi/adt_lec.h
U   linux/branches/2.2/drivers/dahdi/voicebus/GpakCust.c
U   linux/branches/2.2/drivers/dahdi/voicebus/GpakCust.h
U   linux/branches/2.2/drivers/dahdi/wctdm24xxp/base.c
U   linux/branches/2.2/drivers/dahdi/wcte12xp/base.c
U   linux/branches/2.2/include/dahdi/kernel.h

------------------------------------------------------------------------
r7572 | sruffell | 2009-11-12 14:08:27 -0600 (Thu, 12 Nov 2009) | 52 lines

Merged revisions 7309,7348,7565-7571 via svnmerge from
https://origsvn.digium.com/svn/dahdi/linux/trunk

........
 r7309 | mattf | 2009-10-02 11:31:58 -0500 (Fri, 02 Oct 2009) | 1 line
 
 Implement API update to do per-channel companding selection for VPMADT032
........
 r7348 | mattf | 2009-10-07 16:26:08 -0500 (Wed, 07 Oct 2009) | 1 line
 
 Fix a logic error in the companding check.  Duh....
........
 r7565 | sruffell | 2009-11-12 13:22:06 -0600 (Thu, 12 Nov 2009) | 7 lines
 
 voicebus: Fix race when enabling/disabling hardware echocan.
 
 This closes a race condition where it was possible for the driver to
 believe it has enabled the VPMADT032 when in fact, it really has not.
 This fixes a regression introduced in dahdi-linux 2.2.0.
 
 (issue DAHLIN-135)
........
 r7566 | sruffell | 2009-11-12 13:22:06 -0600 (Thu, 12 Nov 2009) | 1 line
 
 wctdm24xxp, wcte12xp: We no longer have any DTMF events to check for.
........
 r7567 | sruffell | 2009-11-12 13:22:07 -0600 (Thu, 12 Nov 2009) | 1 line
 
 voicebus: Remove unused curtone from 'struct vpmadt032'
........
 r7568 | sruffell | 2009-11-12 13:22:07 -0600 (Thu, 12 Nov 2009) | 1 line
 
 voicebus: Remove redundant MAX_CHANNELS_FROM_SPAN
........
 r7569 | sruffell | 2009-11-12 13:22:08 -0600 (Thu, 12 Nov 2009) | 3 lines
 
 voicebus: Use dev_xxx macro when printing vpm messages.
 
 We also do not need the unused context member of the vpmadt032 structure.
........
 r7570 | sruffell | 2009-11-12 13:22:08 -0600 (Thu, 12 Nov 2009) | 4 lines
 
 wcte12xp: Change serial port configuration setting for hw echocan.
 
 The wcte12xp, like the wctdm24xpp, should have the PcmOutPortA set to
 SerialPortNull.
........
 r7571 | sruffell | 2009-11-12 13:56:49 -0600 (Thu, 12 Nov 2009) | 1 line
 
 kernel.h: Define 'list_replace' for kernels < 2.6.18
........

------------------------------------------------------------------------

http://svn.digium.com/view/dahdi?view=rev&revision=7572
By: Shaun Ruffell (sruffell) 2009-11-12 14:11:20.000-0600

alecdavis, et. al:  I've stressed the above changes last night and today, and haven't been able to get them into a condition where the vpm isn't enabled anymore.  So....if anyone would be willing to try revision 7572 of the 2.2 branch of dahdi-linux, I would appreciate it.
By: Darren Franks (ghomf) 2009-11-12 14:38:51.000-0600

I have install 7571 and will let you know how the echo is.I used the command
svn co http://svn.asterisk.org/svn/dahdi/linux-complete/trunk dahdi-linux-complete-trunk

After installing trunk 7571 of dahdi , dahdi_tool reports 30 irq misses as opposed to 23 irq misses with dahdi current. I can see that the vpm firmware is 120 in var/log/messages.
I tested faxing and the vpm turns off and detects fax

Nov 12 15:40:11 asterisk kernel: dahdi: Telephony Interface Registered on major 196
Nov 12 15:40:11 asterisk kernel: dahdi: Version: SVN-trunk-r7571
Nov 12 15:40:11 asterisk kernel: ACPI: PCI Interrupt 0000:04:08.0[A] -> GSI 17 (level, low) -> IRQ 177
Nov 12 15:40:14 asterisk kernel: dahdi_vpmadt032_loader: module license 'Digium Commercial' taints kernel.
Nov 12 15:40:18 asterisk kernel: wcte12xp 0000:04:08.0: Booting VPMADT032
Nov 12 15:40:27 asterisk kernel: wcte12xp: VPM present and operational (Firmware version 120)
Nov 12 15:40:27 asterisk kernel: wcte12xp: Setting up global serial parameters for T1
Nov 12 15:40:27 asterisk kernel: wcte12xp: Found a Wildcard TE121
Nov 12 15:40:28 asterisk kernel: dahdi_echocan_mg2: Registered echo canceler 'MG2'
Nov 12 15:40:28 asterisk kernel: dahdi: Registered tone zone 0 (United States / North America)
Nov 12 15:40:28 asterisk kernel: wcte12xp: Span configured for ESF/B8ZS


By: Alec Davis (alecdavis) 2009-11-13 03:57:10.000-0600

Working well, we are E1 connected, not T1.

Using the following versions:
DAHDI Version: SVN-trunk-r7582 Echo Canceller:
libpri version: SVN-branch-1.4-r1261
Asterisk SVN-branch-1.6.1-r228695


By: Isaac Gal (isaacgal) 2009-11-14 11:50:12.000-0600

sruffell: is the fax detect issue merged to 7582 as well?

since i'm running a production system i rather wait till all issues resolved before forcing the client to hate us more (-:

running zaptel for now
By: Shaun Ruffell (sruffell) 2010-05-21 13:08:29

Cleaning up stale issues that are still open.  I'm not aware of any fax detect issues on trunk, the 2.2 or 2.3 branches with the VPMADT032.  If there are, please open a new issue.