asterisk-tools:sruffells-dahdi-linux.git
8 years agoPIC 4 rev. 7381: fix T1 returning signaling register in non-CAS
Tzafrir Cohen [Wed, 14 Jul 2010 11:50:59 +0000 (11:50 +0000)]
PIC 4 rev. 7381: fix T1 returning signaling register in non-CAS

This is the second part of the partial fix in r8872, though each of the
two should work independently.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8931 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoAllow reading caps when twinstar watchdog set
Tzafrir Cohen [Wed, 14 Jul 2010 11:42:00 +0000 (11:42 +0000)]
Allow reading caps when twinstar watchdog set

Fixes reading capabilities when the watchdog is set. In that case most
of the data, as seen, e.g. in output of 'astribank_allow', was 0xFF.

Xorcom rev: 7809

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8930 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoSet Astribank PRI before initialization
Tzafrir Cohen [Wed, 14 Jul 2010 10:28:10 +0000 (10:28 +0000)]
Set Astribank PRI before initialization

This avoids most cases of sending garbage at startup by setting each port
to tristate mode at init time (in the init script) and enable it back at
DAHDI_STARTUO (end of dahdi_cfg).

Upgrade note: if you have upgraded the files (including init script) but old
module is still loaded, you may end up with the port not getting ever
enabled and the line practically dead. In such a case, reload the new module.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8922 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agostart migrating from xbus_num() to get_xbus()/put_xbus()
Tzafrir Cohen [Tue, 13 Jul 2010 19:38:13 +0000 (19:38 +0000)]
start migrating from xbus_num() to get_xbus()/put_xbus()

* Now get_xbus() receive and xbus number (not poiner) and uses xbus_num()
  internally to map it to an xbus pointer + refcount increment. (this is
  atomic)
* Migrate all obvious places that used xbus_num() to map bus number into a
  pointer, so now they use get_xbus() + put_xbus() to aquire and release an
  xbus.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8913 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoPRI Astribanks always sync AB (and independent)
Tzafrir Cohen [Tue, 13 Jul 2010 18:35:20 +0000 (18:35 +0000)]
PRI Astribanks always sync AB (and independent)

* PRI devices always get SYNC-AB (never PLL, not even for NT)
* Fix Timing priority calculation accordingly.
* On PRI layer1 changes, elect_syncer() is called, so we have
  re-election.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8909 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoprevent handling of duplicate Astribanks
Tzafrir Cohen [Tue, 13 Jul 2010 17:25:06 +0000 (17:25 +0000)]
prevent handling of duplicate Astribanks

Don't let AB which are unplugged/replugged create duplicate XBUS's (if
asterisk don't close the original channels).

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8905 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoCreate /sys/devices/astribanks:
Tzafrir Cohen [Tue, 13 Jul 2010 16:28:37 +0000 (16:28 +0000)]
Create /sys/devices/astribanks:

* Astribank devices now reside under their own top-level device.
* The USB device is still used as the 'transport'.
* This means the astribanks parents are NOT the USB devices.
* As a result, even after a USB disconnect, we have valid sysfs representation.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8904 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoKeep SYSFS objects after disconnect
Tzafrir Cohen [Tue, 13 Jul 2010 15:10:30 +0000 (15:10 +0000)]
Keep SYSFS objects after disconnect

When the USB device disconnects, we keep references to them to make sure
they don't disappear.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8900 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoDon't reset span flag RBS on pri module
Tzafrir Cohen [Tue, 13 Jul 2010 13:59:24 +0000 (13:59 +0000)]
Don't reset span flag RBS on pri module

Don't attempt to reset the span flag DAHDI_FLAG_RBS.
At pre-registration we don't yet know if we're CAS or not.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8896 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodon't send "duplicates" in E1 as in D4
Tzafrir Cohen [Tue, 13 Jul 2010 12:05:58 +0000 (12:05 +0000)]
don't send "duplicates" in E1 as in D4

Make sure we only consider T1-s to be potentially D4 (which requires
sending the bits twice). This fixes occasional command-queue floods in
E1 CAS.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8892 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoPut a space where it belongs
Tzafrir Cohen [Tue, 13 Jul 2010 11:13:01 +0000 (11:13 +0000)]
Put a space where it belongs

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8888 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoBasic support for DAHDI_VMWI_CONFIG
Tzafrir Cohen [Tue, 13 Jul 2010 11:07:57 +0000 (11:07 +0000)]
Basic support for DAHDI_VMWI_CONFIG

* Supports only our current NEON notification.
* No more global 'vmwineon' parameter:
  - We expect asterisk to tell us by channel.
  - If 'vmwi_ioctl' is not set (it is by default), that we revert
    to FSK searching (expensive) and call directly our NEON code
    (as before).

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8887 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoSolve race xbus_populate
Tzafrir Cohen [Tue, 13 Jul 2010 10:31:55 +0000 (10:31 +0000)]
Solve race xbus_populate

Fixes a crash resulting from a race between disconnecting and connecting
Astribanks (on multi-core systems)

* Use get_xbus()/put_xbus() arround xbus_populate(), so a disconnect
  in the middle won't release the xbus too early.
* Aquire all XPDs before starting initialization and release them
  after it finishes (so we don't have up/down races among XPDs)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8886 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoput xbus->worker in xbus; wake_up_interruptible_all()
Tzafrir Cohen [Tue, 13 Jul 2010 09:42:45 +0000 (09:42 +0000)]
put xbus->worker in xbus; wake_up_interruptible_all()

* The worker member of 'struct xbus' is now an embedded struct xbus_workqueue.
* Replace wake_up() to wake_up_interruptible_all().

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8882 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoCleanup: refcount_xpd()
Tzafrir Cohen [Tue, 13 Jul 2010 08:44:36 +0000 (08:44 +0000)]
Cleanup: refcount_xpd()

* Change refcount_xpd(): instead of returning a pointer to the atomic
  counter, return directly its value (just like refcount_xbus())
* Add a refcount_xbus and refcount_xpd attributes to sysfs (for debugging)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8878 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoFix building without CONFIG_PROC_FS defined
Tzafrir Cohen [Tue, 13 Jul 2010 08:33:00 +0000 (08:33 +0000)]
Fix building without CONFIG_PROC_FS defined

This changeset should have no functional changes.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8877 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoDon't assume hardwired pri_protocol is E1
Tzafrir Cohen [Tue, 13 Jul 2010 07:41:45 +0000 (07:41 +0000)]
Don't assume hardwired pri_protocol is E1

When initializing registers for the "PRI" (E1/T1) module, in the case the
user has not explicitly asked for E1/T1 settings in xpp.conf (pri_protocol)
fall back to the current value, which is the build-time fall-back.

The value is read from sysfs.

Previously we implicitly assumed this value is E1.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8873 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agopartial fix of T1 returning signaling register in non-CAS mode
Tzafrir Cohen [Tue, 13 Jul 2010 07:22:02 +0000 (07:22 +0000)]
partial fix of T1 returning signaling register in non-CAS mode

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8872 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi: Explicitly ensure we don't schedule a timer for the current tick.
Shaun Ruffell [Mon, 12 Jul 2010 18:45:03 +0000 (18:45 +0000)]
dahdi: Explicitly ensure we don't schedule a timer for the current tick.

As best as I can tell, when CONFIG_NO_HZ is set along with CONFIG_HZ < 250, it
is possible for the system timer to exceed MAX_SOFTIRQ_RESTART. Tony
Mountifield alluded that this might be a problem in the below mailing list
posting, but when I was originally testing, I wasn't using CONFIG_NO_HZ and HZ
< 250.

http://www.mail-archive.com/asterisk-dev@lists.digium.com/msg37384.html

(closes issue #17620)
Reported by: seanbright

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8868 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoExtra PCI ID for Junghanns PCI-E cards.
Tzafrir Cohen [Tue, 6 Jul 2010 17:49:51 +0000 (17:49 +0000)]
Extra PCI ID for Junghanns PCI-E cards.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8853 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoAdd support for max_latency module parameter in wct4xxp
Matthew Fredrickson [Fri, 2 Jul 2010 18:58:32 +0000 (18:58 +0000)]
Add support for max_latency module parameter in wct4xxp

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8849 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoFix for circumstances where the framer needs extra help to reset itself.
Matthew Fredrickson [Wed, 30 Jun 2010 20:23:24 +0000 (20:23 +0000)]
Fix for circumstances where the framer needs extra help to reset itself.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8841 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp, wcte12xp: Fix "operation on may be undefined" warning.
Shaun Ruffell [Tue, 29 Jun 2010 20:20:20 +0000 (20:20 +0000)]
wctdm24xxp, wcte12xp: Fix "operation on may be undefined" warning.

gcc 4.5.0 generates a warning on the changed lines and
http://gcc.gnu.org/ml/gcc/2004-10/msg00032.html explains why.

Essentially, the only thing guaranteed with the preincrement operator is that
the value will be incremented before the assignment.  It's undefined where in
the sequence the mask will be applied.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8832 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctc4xxp: Fix "operation on may be undefined" warning.
Shaun Ruffell [Tue, 29 Jun 2010 20:20:19 +0000 (20:20 +0000)]
wctc4xxp: Fix "operation on may be undefined" warning.

gcc 4.5.0 was generating a warning on the changed line, and the
discussion at http://gcc.gnu.org/ml/gcc/2004-10/msg00024.html explains
why.

Essentially, the only thing guaranteed with the preincrement operator is
that the value will be incremented before the assignment.  It's
undefined where in the sequence the mask will be applied.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8831 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoDocument DAHDI 2.3 and 2.2 incompatibilities
Tzafrir Cohen [Tue, 29 Jun 2010 12:39:35 +0000 (12:39 +0000)]
Document DAHDI 2.3 and 2.2 incompatibilities

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8823 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoFix the README "_V1" ioctls examples
Tzafrir Cohen [Tue, 29 Jun 2010 12:26:59 +0000 (12:26 +0000)]
Fix the README "_V1" ioctls examples

zap->dahdi and such.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8822 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agolive_dahdi: properly unload EC modules
Tzafrir Cohen [Tue, 29 Jun 2010 09:41:22 +0000 (09:41 +0000)]
live_dahdi: properly unload EC modules

Copy sruffel's fix to the init.d script from -tools.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8818 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agolive_dahdi: regen config or restart asterisk without full reload
Tzafrir Cohen [Mon, 28 Jun 2010 10:25:34 +0000 (10:25 +0000)]
live_dahdi: regen config or restart asterisk without full reload

Adds commands 'genconf' (for basically: dahdi_genconf; dahdi_cfg) and
'asterisk' (start asterisk) so doing those should not require a full
restart.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8814 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agochan_notice() now works
Tzafrir Cohen [Sun, 27 Jun 2010 15:36:17 +0000 (15:36 +0000)]
chan_notice() now works

Make chan_notice() and does not complain about undefined 'chan'.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8809 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoFix trailing whitespace and make last change more readible. (sruffell suggestion)
Mike Spiceland [Thu, 24 Jun 2010 21:02:56 +0000 (21:02 +0000)]
Fix trailing whitespace and make last change more readible. (sruffell suggestion)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8805 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoImpliments a workaround for an errata in the qfalc v3.1 chip which caused
Mike Spiceland [Thu, 24 Jun 2010 16:13:44 +0000 (16:13 +0000)]
Impliments a workaround for an errata in the qfalc v3.1 chip which caused
RBS modes to be broken when using AMI coding.  DAHDI-647.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8801 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoDisable DAHDI_AUDIO_NOTIFY by default
Tzafrir Cohen [Fri, 18 Jun 2010 10:57:22 +0000 (10:57 +0000)]
Disable DAHDI_AUDIO_NOTIFY by default

It's still not tested well-enough, if one enables OPTIMIZE_CHANMUTE.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8784 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoExpose some dynamic latency options as module parameters for the wct4xxp driver
Mike Spiceland [Mon, 14 Jun 2010 21:34:54 +0000 (21:34 +0000)]
Expose some dynamic latency options as module parameters for the wct4xxp driver

latency - the default number of ms of buffering to start off with
ms_per_irq - how often the card interrupts

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8768 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoUse usb_{alloc,free}_coherent instead of usb_buffer_{alloc,free}.
Tzafrir Cohen [Tue, 8 Jun 2010 19:46:15 +0000 (19:46 +0000)]
Use usb_{alloc,free}_coherent instead of usb_buffer_{alloc,free}.

usb_buffer_alloc has been replaced with usb_alloc_coherent , and likewise
usb_buffer_free has been replaced with usb_free_coherent. in kernels 2.6.34
the older names are proxies for the new ones, and in 2.6.35 the older names are gone.

This patch defines the new names for the old ones, for older kernel versions.

(closes issue #17383)
Reported by: Chainsaw

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8762 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoinclude semaphore.h explicitly for newer kernels
Tzafrir Cohen [Tue, 8 Jun 2010 19:09:09 +0000 (19:09 +0000)]
include semaphore.h explicitly for newer kernels

With kernel version >= 2.6.35-rc1, building failed with:

  drivers/dahdi/voicebus/GpakCust.h:114: error: field ‘sem’ has incomplete type

(closes issue #17382)
Reported by: Chainsaw
Patches:
      07-semaphore-include-not-optional.diff uploaded by Chainsaw (license 723)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8758 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agolive_dahdi: load firmware_class if required
Tzafrir Cohen [Tue, 8 Jun 2010 14:44:37 +0000 (14:44 +0000)]
live_dahdi: load firmware_class if required

Modprobe firmware_class for modules that may need it (and that we
insmod later)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8751 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp: Do not access memory after we free it.
Shaun Ruffell [Tue, 25 May 2010 19:23:40 +0000 (19:23 +0000)]
wctdm24xxp: Do not access memory after we free it.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8689 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agohpec: #include <linux/slab.h>
Shaun Ruffell [Mon, 24 May 2010 16:55:50 +0000 (16:55 +0000)]
hpec: #include <linux/slab.h>

Fix the same issue as in r8550 , for dahdi_echocan_hpec.c:

With kernel 2.6.34 an explicit '#include <slab.h>' is required for
using kzalloc() and friends.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8680 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agomissing #include: fixes building with 2.6.34-rc5
Tzafrir Cohen [Sun, 23 May 2010 12:17:52 +0000 (12:17 +0000)]
missing #include: fixes building with 2.6.34-rc5

Fix the same issue as in r8550 , for dahdi_echocan_oslec.c:

With kernel 2.6.34-rc5 an explicit '#include <slab.h>' is required for using
kzalloc() and friends.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8674 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoA typo (lintian)
Tzafrir Cohen [Sun, 23 May 2010 12:15:21 +0000 (12:15 +0000)]
A typo (lintian)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8673 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoMerge in pulse mask improvement patch for B410P.
Matthew Fredrickson [Mon, 17 May 2010 21:44:44 +0000 (21:44 +0000)]
Merge in pulse mask improvement patch for B410P.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8653 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp: dev_notice -> dev_info
Shaun Ruffell [Mon, 17 May 2010 14:45:13 +0000 (14:45 +0000)]
wctdm24xxp: dev_notice -> dev_info

Modules that aren't installed should show at the same level as
modules that are installed.  It could be confusing if the console
log level is set to 3 and only messages about which ports do not
have any modules installed show up on the console.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8642 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi: Check if wait_for_completion_timeout is backported.
Shaun Ruffell [Mon, 17 May 2010 14:45:12 +0000 (14:45 +0000)]
dahdi: Check if wait_for_completion_timeout is backported.

CentOS4 has this backported in their 2.6.9 kernel.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8641 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Removed unnecessary instrumentation
Russ Meyerriecks [Mon, 3 May 2010 18:23:09 +0000 (18:23 +0000)]
wcte12xp: Removed unnecessary instrumentation

Over-verbose instrumentation snuck in, on commit r8594

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8600 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Fixes RBS/Clear mode not set correctly bug
Russ Meyerriecks [Mon, 3 May 2010 14:50:59 +0000 (14:50 +0000)]
wcte12xp: Fixes RBS/Clear mode not set correctly bug

(part 3) Fixes DAHDI-449 where dahdi_cfg would need to be run multiple times
in order to properly set the rbs or clear mode of a channel. The prior
logic was calling set_clear in the context of setting all channels to
clear mode, even if the channel was intended to be in bit robbed mode.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8594 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agofix kernel-cp download test
Tzafrir Cohen [Mon, 3 May 2010 07:48:12 +0000 (07:48 +0000)]
fix kernel-cp download test

'wc -c <f$ile' returns the name of the file as well.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8585 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wctdm24xxp: Retry if the VPMADT032 reports not ready.
Shaun Ruffell [Mon, 26 Apr 2010 23:29:17 +0000 (23:29 +0000)]
wcte12xp, wctdm24xxp: Retry if the VPMADT032 reports not ready.

CheckDspReset can return -1 if the DSP is not ready to process any new
commands. In this case we should retry a few times to give the DSP a chance
to become ready.  While I'm not ready to say this definitely fixes recently
reported cases when the wcte12xp driver constantly resets, it eliminated
communication failures to the DSP module when under stress (via the
vpm_firmware_version sysfs attribute).  However, I haven't let it run long
enough to say that the issue is resolved.  DAHDI-603.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8576 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wctdm24xxp: Do not allow interruptible sleep on VPM lock.
Shaun Ruffell [Mon, 26 Apr 2010 23:29:16 +0000 (23:29 +0000)]
wcte12xp, wctdm24xxp: Do not allow interruptible sleep on VPM lock.

If the sleep is ever interrupted, 'up' will still be called in the GpakApi,
essentially making the lock useless after that point.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8575 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wctdm24xxp: Add vpm_firmware_version sysfs attribute.
Shaun Ruffell [Mon, 26 Apr 2010 23:29:16 +0000 (23:29 +0000)]
wcte12xp, wctdm24xxp: Add vpm_firmware_version sysfs attribute.

Currently only exported if CONFIG_VOICEBUS_SYSFS is defined in
drivers/dahdi/voicebus.h. Reading from the 'vpm_firmware_version' attribute
will poll the firmware on the VPMADT032 for it's current version.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8574 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Add the voicebus_current_latency sysfs node.
Shaun Ruffell [Mon, 26 Apr 2010 23:29:15 +0000 (23:29 +0000)]
wcte12xp: Add the voicebus_current_latency sysfs node.

'voicebus_current_latency' is only exported when CONFIG_VOICEBUS_SYSFS is
defined in voicebus.h.  This is a debugging aide which enables determing the
board specific latency without parsing through the kernel logs.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8573 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi-base: Reverted a change from last patch
Russ Meyerriecks [Mon, 26 Apr 2010 21:49:55 +0000 (21:49 +0000)]
dahdi-base: Reverted a change from last patch

Removed a change to dahdi-base from last patch which might have caused
compatibility with drivers other than the wcte12xp.

wcte12xp: The channel clear/rbs function no longer reads the register first.
It now uses the span's channel flags to determine each channels clear state.
Also added various minor readability improvements.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8569 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi-base: wcte12xp: Fixed Clear/RBS channel mode issue
Russ Meyerriecks [Fri, 23 Apr 2010 21:54:30 +0000 (21:54 +0000)]
dahdi-base: wcte12xp: Fixed Clear/RBS channel mode issue

Fixes DAHDI-449 where chanconfig was failing on the first call. It needed
to be called twice in a row. This was due to the channel configuration using
a non-relative channel number in its loop.

Also re-added the register dumping ioctl for inspecting the framer's state.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8564 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wctdm24xxp: Ensure writes to I/O registers are flushed.
Shaun Ruffell [Thu, 22 Apr 2010 21:36:16 +0000 (21:36 +0000)]
wcte12xp, wctdm24xxp: Ensure writes to I/O registers are flushed.

In revision 8176 I changed register access from I/O space to memory mapped
registers.  Unfortunately, when I made that change, I didn't account for
posted writes.  This change makes sure all the registers are read back to
ensure that they are posted through any intermediate bridges.

The most readily observable symptom were cards that were taking 2000
interrupts/second. The card reported that it handled an interrupt but the
write to silence the card wasn't flushed through until the second time the
interrupt handler run. DAHDI-602.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8560 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoConfigurable debounce to support old rotary phones
Tzafrir Cohen [Thu, 22 Apr 2010 13:28:43 +0000 (13:28 +0000)]
Configurable debounce to support old rotary phones

(closes issue #16339)
Reported by: alecdavis
Patches:
      20091129__issue16339.diff.txt uploaded by tilghman (license 14)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8556 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agomissing #include: fixes building with 2.6.34-rc5
Tzafrir Cohen [Thu, 22 Apr 2010 10:39:03 +0000 (10:39 +0000)]
missing #include: fixes building with 2.6.34-rc5

With kernel 2.6.34-rc5 an explicit '#include <slab.h>' is required for using
kzalloc() and friends.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8550 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agotypos, mostly unkown
Tzafrir Cohen [Mon, 19 Apr 2010 19:13:32 +0000 (19:13 +0000)]
typos, mostly unkown

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8539 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoRemove generation of kernel-doc API docs
Tzafrir Cohen [Mon, 19 Apr 2010 18:57:11 +0000 (18:57 +0000)]
Remove generation of kernel-doc API docs

It's not really useful, and it breaks building 'docs' without a kernel tree.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8538 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agovoicebus: Change WARN_ON_ONCE to BUG_ON.
Shaun Ruffell [Tue, 6 Apr 2010 23:19:56 +0000 (23:19 +0000)]
voicebus: Change WARN_ON_ONCE to BUG_ON.

WARN_ON_ONCE is not defined in 2.6.9, and this condition would be catastrophic
anyway if it were to occur.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8494 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Load the VPM in startup, not spanconfig
Shaun Ruffell [Tue, 6 Apr 2010 22:36:10 +0000 (22:36 +0000)]
wcte12xp: Load the VPM in startup, not spanconfig

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8489 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Check and dequeue/decipher the VPM commands atomically
Shaun Ruffell [Tue, 6 Apr 2010 22:36:09 +0000 (22:36 +0000)]
wcte12xp: Check and dequeue/decipher the VPM commands atomically

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8488 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Make sure the interrupt is stopped before freeing the vpmadt032
Shaun Ruffell [Tue, 6 Apr 2010 22:36:09 +0000 (22:36 +0000)]
wcte12xp: Make sure the interrupt is stopped before freeing the vpmadt032

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8487 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Eliminate a trailing whitespace.
Shaun Ruffell [Tue, 6 Apr 2010 19:06:22 +0000 (19:06 +0000)]
wcte12xp: Eliminate a trailing whitespace.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8482 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Do not reconfigure the VPMADT032 if shutting down.
Shaun Ruffell [Tue, 6 Apr 2010 19:05:29 +0000 (19:05 +0000)]
wcte12xp: Do not reconfigure the VPMADT032 if shutting down.

If we try to unload the driver soon after a high latency event, it is possible
to get stuck for several seconds reloading the firmware. DAHDI-573.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8481 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Do not restart the timer on alarm poll if shutting down.
Shaun Ruffell [Tue, 6 Apr 2010 19:05:28 +0000 (19:05 +0000)]
wcte12xp: Do not restart the timer on alarm poll if shutting down.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8480 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Expose the maximum latency as a module parameter
Shaun Ruffell [Tue, 6 Apr 2010 19:05:28 +0000 (19:05 +0000)]
wcte12xp: Expose the maximum latency as a module parameter

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8479 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Reprogram the VPM only when neccessary.
Shaun Ruffell [Tue, 6 Apr 2010 19:05:27 +0000 (19:05 +0000)]
wcte12xp: Reprogram the VPM only when neccessary.

Only program the VPM in spanconfig if it either is not setup, or if it fails a
ping test.  Also, if the channel config fails (but ping would otherwise work),
force a reset / reconfiguration if the VPM module. DAHDI-573.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8478 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Try to reconfigure the VPM if we fail to configure the channels.
Shaun Ruffell [Mon, 5 Apr 2010 19:15:59 +0000 (19:15 +0000)]
wcte12xp: Try to reconfigure the VPM if we fail to configure the channels.

Also, fall back to any software echocan configured for this channel for new
calls while we're in the middle of a recovery.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8473 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Poll the VPM and reset it if necessary.
Shaun Ruffell [Sun, 4 Apr 2010 16:14:59 +0000 (16:14 +0000)]
wcte12xp:  Poll the VPM and reset it if necessary.

When the transmit descriptor runs out completely, there (appears to be) a
chance for a random command to be sent that results in the VPMADT032 to no
longer respond, typically resulting in one way audio.  This change introduces
a poll of the VPM.  If it fails the poll, it will be bypassed temporarily
while the driver resets and reprograms it.  Also, the VPM is initially
programmed in the spanconfig callback instead of at driver load.  This moves
the potential for underruns until later in the boot process. DAHDI-573.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8468 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wctdm24xxp: If we're only one packet behind, just exit.
Shaun Ruffell [Sun, 4 Apr 2010 16:14:56 +0000 (16:14 +0000)]
wcte12xp, wctdm24xxp: If we're only one packet behind, just exit.

This will make us two behind, which is fine, and eliminates a busy loop in
atomic context.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8467 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Use ARRAY_SIZE in a couple more places
Shaun Ruffell [Sun, 4 Apr 2010 16:14:53 +0000 (16:14 +0000)]
wcte12xp: Use ARRAY_SIZE in a couple more places

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8466 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wtdm24xxp: Make sure the writes are retried.
Shaun Ruffell [Sat, 3 Apr 2010 18:10:27 +0000 (18:10 +0000)]
wcte12xp, wtdm24xxp: Make sure the writes are retried.

r8454 and r8460 introduced a change where writes are not retried when other
module/framer commands are retried.  This was an error and wasn't what was
actually under test.  This commit restores the behavior in wctdm24xxp and
makes sure the vpm writes are retried in the wcte12xp.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8461 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wctdm24xxp: Close a memory leak when processing VPM commands.
Shaun Ruffell [Sat, 3 Apr 2010 17:50:28 +0000 (17:50 +0000)]
wcte12xp, wctdm24xxp: Close a memory leak when processing VPM commands.

Closes a memory leak when processing the VPM write commands introduced in
r8454.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8460 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Use our own workqueue for the timer and maint events.
Shaun Ruffell [Sat, 3 Apr 2010 00:03:58 +0000 (00:03 +0000)]
wcte12xp: Use our own workqueue for the timer and maint events.

When the latency is large and register reads can take 100s of milliseconds, the
alarm polling function could tie up one of the global workqueue threads long
enough to interfere with other system operations.  Most noticeably the console.
DAHDI-573

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8455 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Resend both framer and VPM commands when reads may have been lost.
Shaun Ruffell [Sat, 3 Apr 2010 00:03:58 +0000 (00:03 +0000)]
wcte12xp: Resend both framer and VPM commands when reads may have been lost.

Latency conditions could cause the driver to misconfigure the VPM which
would result in one way audio. DAHDI-572.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8454 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi-base: Fix a bug preventing clobbering maint state
Russ Meyerriecks [Mon, 29 Mar 2010 20:43:33 +0000 (20:43 +0000)]
dahdi-base: Fix a bug preventing clobbering maint state

Now the errror injection code prevents placing the driver into a loopback
state in a less forceful way.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8443 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp: Add an 'hx8' mode for buffer processing.
Shaun Ruffell [Thu, 25 Mar 2010 16:23:53 +0000 (16:23 +0000)]
wctdm24xxp: Add an 'hx8' mode for buffer processing.

The Hx8 series cards do not need any idle buffers and idle_buffers complicate
processing when using the timing_cable.  This change adds another mode of
operation for the voicebus layer for the Hx8 cards that operates without the
idle buffers.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8431 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcb4xxp - add Swyx 4xS0 SX2 QuadBri PCI IDs
Tzafrir Cohen [Thu, 25 Mar 2010 15:13:46 +0000 (15:13 +0000)]
wcb4xxp - add Swyx 4xS0 SX2 QuadBri PCI IDs

(closes issue #16493)
Reported by: nic_bellamy
Patches:
      wcb4xxp_swyx_sx2_quadbri_pci_ids.patch uploaded by nic bellamy (license 299)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8423 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoSome more wcb4xxp PCI IDs for Junghanns cards
Tzafrir Cohen [Thu, 25 Mar 2010 14:43:28 +0000 (14:43 +0000)]
Some more wcb4xxp PCI IDs for Junghanns cards

Resolve two open issues. One of them that accidentally wasn't closed.

(closes issue #15446)
(closes issue #16447)

Reported by: lpistone
Tested by: okrief

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8421 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Fix typo in last commit.
Shaun Ruffell [Wed, 24 Mar 2010 15:56:51 +0000 (15:56 +0000)]
wcte12xp: Fix typo in last commit.

Wouldn't compile on kernels >= 2.6.20.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8416 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: t1xxp_maint is also called with interrupts disabled from dahdi_ioctl
Shaun Ruffell [Wed, 24 Mar 2010 15:39:05 +0000 (15:39 +0000)]
wcte12xp: t1xxp_maint is also called with interrupts disabled from dahdi_ioctl

So just push all the maintenance mode processing off to the workqueue.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8415 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: t1xxp_maint can be called from interrupt context.
Shaun Ruffell [Wed, 24 Mar 2010 14:32:23 +0000 (14:32 +0000)]
wcte12xp: t1xxp_maint can be called from interrupt context.

Since t1xxp_maint can be called from interrupt context with the
DAHDI_MAINT_LOOPSTOP cmd, push the processing of that command to a workqueue
since it may sleep in the t1_getreg call. DAHDI-560.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8410 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodocument the header line of /proc/dahdi/NN
Tzafrir Cohen [Wed, 24 Mar 2010 12:07:32 +0000 (12:07 +0000)]
document the header line of /proc/dahdi/NN

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8405 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp: Fix potential race on command handling.
Shaun Ruffell [Tue, 23 Mar 2010 21:56:02 +0000 (21:56 +0000)]
wcte12xp: Fix potential race on command handling.

If we timeout a command, don't free it right away unless we are the one who
removed it from whatever list it was on.  Also, increase the timeout (2 seconds
wasn't enough when the firmware for the VPMOCT was being loaded on a system with
4K stacks) and check the return value from t1_getreg in case there were timeouts.
DAHDI-560.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8400 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi: 'clamp' may already be backported.
Shaun Ruffell [Tue, 23 Mar 2010 21:56:01 +0000 (21:56 +0000)]
dahdi: 'clamp' may already be backported.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8399 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoone-liner asciidoc syntax fix
Tzafrir Cohen [Tue, 23 Mar 2010 09:30:01 +0000 (09:30 +0000)]
one-liner asciidoc syntax fix

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8387 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp, wcte12xp: Fix several problems with buffer processing.
Shaun Ruffell [Mon, 22 Mar 2010 09:20:59 +0000 (09:20 +0000)]
wctdm24xxp, wcte12xp: Fix several problems with buffer processing.

* If the receive packet isn't the correct size, we can't just drop it, we have
  to resubmit it to the receive queue immediately so it will get cleaned up
  properly.

* It isn't safe to leave the buffers on any lists while calling
  the function to resubmit it to the descriptor ring, since it may be freed
  which potentially results in a corrupted list.

* If the card is held up in a state where it is waiting for receive data, there
  is a chance that when the transmit process is started up again it could DMA
  data into a buffer that has already been freed.  By using our own freelist,
  there isn't a chance that the board will DMA data into memory that has been
  recycled for another purpose.

* Disable the tasklet when processing hard underruns in order to not corrupt the
  descriptor rings.

* Make sure voicebus_stop and voicebus_release aren't running at the same time
  as the hard underrun handler.  This too can result in corrupted descriptor
  lists.

DAHDI-560.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8379 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp, wcte12xp: Remove the completion on shutdown.
Shaun Ruffell [Mon, 22 Mar 2010 09:20:54 +0000 (09:20 +0000)]
wctdm24xxp, wcte12xp: Remove the completion on shutdown.

Oddities in the receive processor state make this completion not so useful
anymore.  It is more straightforward to simply poll the state on shutdown.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8378 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp, wcte12xp: Remove locking macros.
Shaun Ruffell [Mon, 22 Mar 2010 09:20:48 +0000 (09:20 +0000)]
wctdm24xxp, wcte12xp: Remove locking macros.

Since the locking will always take place in a bottom half now (either a timer or
tasklet), we do not need the special locking macros anymore.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8377 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowctdm24xxp: Make sure we're really stopped before deleting the channels.
Shaun Ruffell [Mon, 22 Mar 2010 09:20:44 +0000 (09:20 +0000)]
wctdm24xxp: Make sure we're really stopped before deleting the channels.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8376 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowcte12xp, wctdm24xxp: Do not keep individual free lists for VPM commands.
Shaun Ruffell [Mon, 22 Mar 2010 09:20:37 +0000 (09:20 +0000)]
wcte12xp, wctdm24xxp: Do not keep individual free lists for VPM commands.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8375 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agofix two cases of missing __user annotation (sparse)
Tzafrir Cohen [Sun, 21 Mar 2010 20:33:27 +0000 (20:33 +0000)]
fix two cases of missing __user annotation (sparse)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8370 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoTwo more printk macros: NOTICE
Tzafrir Cohen [Sun, 21 Mar 2010 17:14:43 +0000 (17:14 +0000)]
Two more printk macros: NOTICE

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8360 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agoprintk wrapper macros - first shot
Tzafrir Cohen [Sat, 20 Mar 2010 19:56:34 +0000 (19:56 +0000)]
printk wrapper macros - first shot

Add a number of printk-like macros to print messages for span and
channel. I found them useful in the sysfs branch.

The _dbg ones use the magical variable debug, and hence require the code
to acknowledge that explicitly by defining DAHDI_PRINK_MACROS_USE_debug
explicitly.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8354 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agolive_dahdi: fixes for rsync mode
Tzafrir Cohen [Thu, 18 Mar 2010 12:38:53 +0000 (12:38 +0000)]
live_dahdi: fixes for rsync mode

* Run programs from live tree rather than tools dir (not copied)
* Calculate MODULES dir after reading config file (you'll still need to
  override it from the command line for building, but not for reload)

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8347 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi-base: Minor bugfix for alarm state indication
Russ Meyerriecks [Tue, 9 Mar 2010 02:38:27 +0000 (02:38 +0000)]
dahdi-base: Minor bugfix for alarm state indication

Fixed a minor issue where stateless DAHDI_MAINT ioctl commands, such
as error insertion and clearing, were indicating the span going into
loopback alarm state when it really was not.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8298 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowct4xxp, wcte12xp: Fixed broken E1 and unsupported function bugs
Russ Meyerriecks [Mon, 8 Mar 2010 23:22:54 +0000 (23:22 +0000)]
wct4xxp, wcte12xp: Fixed broken E1 and unsupported function bugs

* Fixed the bug where the card could not be brought out of local
  loopback in E1 mode.

* Fixed a bunch of issues where the drivers didn't report unsupported
  maintenance functions correctly.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8291 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agodahdi_maint: Added local loopback capability in E1 modes
Russ Meyerriecks [Mon, 8 Mar 2010 18:37:00 +0000 (18:37 +0000)]
dahdi_maint: Added local loopback capability in E1 modes

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8279 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowct4xxp: wcte12xp: Revised maint clearing
Russ Meyerriecks [Mon, 8 Mar 2010 16:49:50 +0000 (16:49 +0000)]
wct4xxp: wcte12xp: Revised maint clearing

Looping modes are now mutually exclusive. If two looping modes are enabled
simultaneously it tends to hose up our framer chip. Now, all looping modes are
cleared in the driver before any are set.

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8274 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowct4xxp: Fixed inability to disable network payload looping
Russ Meyerriecks [Fri, 5 Mar 2010 19:37:14 +0000 (19:37 +0000)]
wct4xxp: Fixed inability to disable network payload looping

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8255 a0bf4364-ded3-4de4-8d8a-66a801d63aff

8 years agowct4xxp: Removed over-verbose instrumentation
Russ Meyerriecks [Fri, 5 Mar 2010 19:08:05 +0000 (19:08 +0000)]
wct4xxp: Removed over-verbose instrumentation

git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@8250 a0bf4364-ded3-4de4-8d8a-66a801d63aff