Age | Commit message (Collapse) | Author |
|
DAHDI-681.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9145 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Add compile time option to improve the reference signal provided to
software echo cancelers. The intent here is for this functionality to
become the default behavior but more testing and work on the edge cases
is needed. It's being brought in now as a compile time option since there
have been reports that it helps in some environments.
Instead of using two buffers, which means that at best we're two
milliseconds behind, use a circular buffer where audio data is written
on the transmit side and read on the receive path. In this way high
latency values will not interfere with the operation of software
echo cancelers. DAHDI-291. DAHDI-387.
This work was originally on:
http://svn.asterisk.org/svn/dahdi/linux/team/sruffell/improved_ecreference@9143
and includes a generic kfifo replacement by Matt Fredrickson.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9144 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Fixes a regression from 9101 'wctdm24xxp: Added "auto" companding option'
where the analog modules were not defaulted to alaw properly when a
digital module is on an Hx8. This could result in very poor audio since the
modules were providing ulaw data, but dahdi-base believed the audio was in
alaw when converting to signed linear.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9143 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
More work to kill the BKL (Big Kernel Lock)
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=10041d2d14688e207d0d829095147aa82c1f211b
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9142 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
From http://svn.digium.com/svn/octasic_api/oct612x/tags/PR49-03/software@44
This is only currently maintained as part of DAHDI linux so it makes sense to
have it directly in DAHDI linux.
This obliterates any chance of having 0 checkpatch.pl errors between the
2.3.0 and 2.4.0 releases. Oh well...
Signed-off-by: Shaun Ruffell <sruffell@digium.com>
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9138 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Defaults to "TE" but can be set by software to "NT" by dahdi_cfg.
Signed-off-by: Shaun Ruffell <sruffell@digium.com>
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9132 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Signed-off-by: Shaun Ruffell <sruffell@digium.com>
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9131 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
The linecompat member can be used by a span to identify the
global-to-the-span signalling types supported. Analog spans do not
support any span-global signalling and therefore linecompat should
always be 0.
Signed-off-by: Shaun Ruffell <sruffell@digium.com>
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9130 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
-1 in cases where automatic timing sync was specified.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9127 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
The previous commit was making all chan configs fail on a span that did not
have a dacs callback. Also add some documentation for 'dahdi_chan_dacs' and
make it more explicit when we're disabling cross connect.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9124 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Proper sentences are always preferred. Thanks kpfleming :)
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9121 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9120 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
RHEL/Centos 5.4 kernels broke after the latest strcasecmp backport. So skip
them.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9117 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9105 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
For kernels >= 2.6.18, each individual card has it's local
timing hung off the pci device in the sysfs tree. dahdi-626
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9102 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Added the "auto" option for the module parameter "companding". In auto
mode its left up to each card to decide what companding mode it should
be in. If a BRI module is installed, it set everything to alaw; ulaw
is selected if no BRI module is installed.
Also fixed a bug where forcing companding one way or the other wasn't
affecting the BRI modules. dahdi-673
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9101 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Added the module parameter "companding" to eventually replace
the current alawoverride parameter. Added some deprecated messages
so current users will know to move over before we remove it entirely.
dahdi-673
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9100 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Renamed the module parameter "alawoverride" to "companding". The valid
values for "companding" are now the strings "alaw" or "ulaw", instead
of integers. dahdi-673
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9099 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Due to the very small number of affected customers we have
removed the module parameter "extendedreset" in favor of a
compile time option "CONFIG_EXTENDED_RESET". dahdi-673
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9098 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
strcasecmp was brought into the kernel at v2.6.22
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9097 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9094 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
This is a regression from r8985 . The temporary 'module' was there because
the pointer 'chan' may become invalid after the close function.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9090 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
We cannot also embed it in struct b400m_span structure and expect container_of
to work. This fixes a regression introduced in r8984.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9087 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9082 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
during call
Since the 3 second click after answering the call on an FXS port has been fixed issue# 15352,
the side effect is that the Proslic during a call, still has the audio signal paths still powered on.
Reading the Si3215 specs it reads that an extra ~20mA is consumed while in OHT mode.
(issue #17764)
Reported by: alecdavis
Patches:
wctdm_fxs_offhook.diff.txt uploaded by alecdavis (license 585)
Tested by: alecdavis
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9073 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
during call [One-liner summary of changes]
Now that the 3 second click after answering the call on an FXS port has been fixed, the side effect is that the Proslic during a call, still has the audio signal paths still powered on.
Reading the Si3215 specs it reads that an extra ~20mA is consumed while in OHT mode.
(closes issue #17764)
Reported by: alecdavis
Patches:
wctdm24xxp_fxs_offhook.diff.txt uploaded by alecdavis (license 585)
Tested by: alecdavis
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9070 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
set_lasttxhook_interruptible()
Problem:
'wctdm_proslic_recheck_sanity()' is executed (4 times??) every ~256ms which monitors and clears the SLIC_LF_OPPENDING on fxs->lasttxhook when the LF state has been reached.
'set_lasttxhook_interruptible()' times out after 100ms waiting from the previous write to the SLIC_LF register, by waiting for SLIC_LF_OPPENDING to be cleared, thus the write will fail.
Solution:
adds wctdm_proslic_check_oppending which does the monitoring and clearing of SLIC_LF_OPPENDING, which is execute every 1ms. It also if the correct state is not reached within 100ms, resends the lasttxhook state.
(closes issue #17724)
Reported by: alecdavis
Patches:
wctdm24xxp_fxs_answer.diff4.txt uploaded by alecdavis (license 585)
Tested by: alecdavis, sruffell
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9063 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Against Kernel Coding Guidelines
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9060 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9059 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
works
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9056 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
description more accurate
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9055 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9054 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
spans and analog channels after detection
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9053 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9052 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
dahdi_maint
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9051 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
when flashing firmware
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9050 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9049 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9046 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Mainly to quiet a sparse warning, but this could be a real problem on
some non x86 platforms.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9045 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Thanks sparse.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9044 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Quiets sparse.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9043 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Make setup_chunks static, eliminate a dynamic array in the interrupt handler,
don't cast away a memory region specifier, and don't initialize statics to 0.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9042 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Fixing formatting errors that have snuck in since the 2.3.0 release.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9041 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Fixing up some of the formatting errors that crept in since the 2.3.0 release.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9040 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9039 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Fixes the following sparse warning "warning: context imbalance in
'dahdi_ctl_ioctl' - different lock contexts for basic block". Thank you
sparse!
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9038 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Added as a workaround for a system which was not routing interrupts
properly and therefore is off by default. Candidate for reversion in a
couple of years. DAHDI-627.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9037 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Certain platforms have trouble with the registers mapped
from I/O space. DAHDI-627.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9036 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
Closes a small window of opportunity where system conditions prevent the
driver from servicing it's receive ring within the timeout period of a
request, and then in the middle of retrying the request (after the
request was already removed from the "waiting_for_response" list) the
driver process the response, it's possible for the driver to ACK the
response without ever pairing it up with the original request. The
result being that the DTE will then ignore our attempts to retry the
original request. DAHDI-430.
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9035 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|
|
As of this writing codec_dahdi (Asterisk module) will not call 'read' on
a transcoder channel more often than it calls 'write'. When a
translation path is setup that will transcode from g723 to g729, write
is called every 30ms and each 'read' returns only 20ms of data. The end
result is audio that slowly becomes increasing delayed.
Since codec_dahdi calls the system read function with a buffer large
enough to hold more than one packet this can prevents packets from
backing up on the channel. DAHDI-582
git-svn-id: http://svn.asterisk.org/svn/dahdi/linux/trunk@9034 a0bf4364-ded3-4de4-8d8a-66a801d63aff
|