These changes fix up some dubious code that I came across while auditing what
happens in the autoservice thread when there are no channels currently in
autoservice.
1) Change it so that autoservice thread doesn't keep looping around calling
ast_waitfor_n() on 0 channels twice a second. Instead, use a thread condition
so that the thread properly goes to sleep and does not wake up until a
channel is put into autoservice.
This actually fixes an interesting bug, as well. If the autoservice thread
is already running (almost always is the case), then when the thread goes
from having 0 channels to have 1 channel to autoservice, that channel would
have to wait for up to 1/2 of a second to have the first frame read from it.
2) Fix up the code in ast_waitfor_nandfds() for when it gets called with no
channels and no fds to poll() on, such as was the case with the previous code
for the autoservice thread. In this case, the code would call alloca(0), and
pass the result as the first argument to poll(). In this case, the 2nd
argument to poll() specified that there were no fds, so this invalid pointer
shouldn't actually get dereferenced, but, this code makes it explicit and
ensures the pointers are NULL unless we have valid data to put there.
(related to issue #12116)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@105563 65c4cc65-6c06-0410-ace0-fbb531ad65f3
the list of channels in autoservice. The problem was that it was possible for
a channel to get removed from autoservice and destroyed, while the autoservice
thread was still messing with the channel. This led to memory corruption, and
caused crashes. This explains multiple backtraces I have seen that have
references to autoservice, but do to the nature of the issue (memory corruption),
could cause crashes in a number of areas.
(fixes the crash in BE-386)
(closes issue #11694)
(closes issue #11940)
The following issues could be related. If you are the reporter of one of these,
please update to include this fix and try again.
(potentially fixes issue #11189)
(potentially fixes issue #12107)
(potentially fixes issue #11573)
(potentially fixes issue #12008)
(potentially fixes issue #11189)
(potentially fixes issue #11993)
(potentially fixes issue #11791)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@105409 65c4cc65-6c06-0410-ace0-fbb531ad65f3
is that if the lock history array was full, then the functions to mark a lock as
acquired or not would adjust the stats for whatever lock is at the end of the array,
which may not be itself. So, do a sanity check to make sure that we're updating
lock info for the proper lock.
(This explains the bizarre stats on lock #63 in BE-396, thanks Mark!)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@105116 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This protects against possible segfaults in applications that may try
to use data before checking length (ast_strdupa'ing it, for example)
(closes issue #12100)
Reported by: foxfire
Patches:
12100-nullappargs.diff uploaded by qwell (license 4)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@105005 65c4cc65-6c06-0410-ace0-fbb531ad65f3
1. Make the list of ast_dial_channels a lockable list. This is because in some cases,
the ast_dial may exist in multiple threads due to asynchronous execution of its application, and
I found some cases where race conditions could exist.
2. Check in ast_dial_join to be sure that the channel still exists before attempting to lock it, since
it could have gotten hung up but the is_running_app flag on the ast_dial_channel may not have been
cleared yet.
(closes issue #12038)
Reported by: jvandal
Patches:
12038v2.patch uploaded by putnopvut (license 60)
Tested by: jvandal
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104841 65c4cc65-6c06-0410-ace0-fbb531ad65f3
(closes issue #11831)
Reported by: IgorG
Patches:
fallbacken.v1.diff uploaded by IgorG (license 20) (modified by me to improve code and conform rest of function to coding guidelines)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104593 65c4cc65-6c06-0410-ace0-fbb531ad65f3
set to make sure that when we come back out of alarm, it gets reported in the log
and manager interface (after discussion with tzafrir on the -dev list)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104591 65c4cc65-6c06-0410-ace0-fbb531ad65f3
does not know what to do with these alarms. Only Asterisk 1.6 cares about it.
So, if we get an unknown alarm in chan_zap, don't generate confusing log messages
about it.
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104332 65c4cc65-6c06-0410-ace0-fbb531ad65f3
wasn't an absolute path (such as DESTDIR=~/asterisk-1.4).
Apparently what was happening, was that some of the targets were being expanded to
the full path, so $@ ended up being /root/asterisk-1.4/[...]/ rather than ~/asterisk-1.4/[...]/
It appears that this may be a new "feature" in GNU make.
(*cough* http://en.wikipedia.org/wiki/Principle_of_least_surprise *cough*)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104132 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This commit brings in a significant set of changes to the SMDI support in Asterisk.
There were a number of bugs in the current implementation, most notably being that
it was very likely on busy systems to pop off the wrong message from the SMDI message
queue. So, this set of changes fixes the issues discovered as well as introducing
some new ways to use the SMDI support which are required to avoid the bugs with
grabbing the wrong message off of the queue.
This code introduces a new interface to SMDI, with two dialplan functions. First,
you get an SMDI message in the dialplan using SMDI_MSG_RETRIEVE() and then you access
details in the message using the SMDI_MSG() function. A side benefit of this is that
it now supports more than just chan_zap.
For example, with this implementation, you can have some FXO lines being terminated
on a SIP gateway, but the SMDI link in Asterisk.
Another issue with the current implementation is that it is quite common that the
station ID that comes in on the SMDI link is not necessarily the same as the Asterisk
voicemail box. There are now additional directives in the smdi.conf configuration
file which let you map SMDI station IDs to Asterisk voicemail boxes.
Yet another issue with the current SMDI support was related to MWI reporting over
the SMDI link. The current code could only report a MWI change when the change
was made by someone calling into voicemail. If the change was made by some other
entity (such as with IMAP storage, or with a web interface of some kind), then the
MWI change would never be sent. The SMDI module can now poll for MWI changes if
configured to do so.
This work was inspired by and primarily done for the University of Pennsylvania.
(also related to issue #9260)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104119 65c4cc65-6c06-0410-ace0-fbb531ad65f3
pointers to channels that are being spied upon. It was very likely that a
crash would occur if the channel being spied upon hung up. This was because
the current ast_channel handling _requires_ that the object is locked or else
it could disappear at any time (except in the owning channel thread). So, this
patch uses some channel datastore magic on the spied upon channel to be able to
detect if and when the channel goes away.
(closes issue #11877)
(patch written by me, but thanks to kpfleming for the idea, and to file for review)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104106 65c4cc65-6c06-0410-ace0-fbb531ad65f3
failed to lock don't sit around in the history. When a lock is first locked,
this checks to see if the last lock in the list was one that was failed to be
locked. If it is, then that was a lock that we're no longer sitting in a trylock
loop trying to lock, so just remove it.
(inspired by issue #11712)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104102 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Specifically, this fixes using #include and #exec in extconfig.conf.
This was basically caused because the config file itself raises the include level to 1.
I opted not to raise the include limit, because recursion here could cause very bizarre behavior.
Pointed out, and tested by jmls
(closes issue #12064)
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104092 65c4cc65-6c06-0410-ace0-fbb531ad65f3
which event was received. What actually was happening was that it was reporting the number of bytes returned
from a call to read().
Thanks to Jared Smith for bringing the issue up on IRC
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@104026 65c4cc65-6c06-0410-ace0-fbb531ad65f3
"dead" realtime queue. Since from the user's perspective, the queue
does exist, we shouldn't tell them we couldn't find the queue. Instead
since it is a dead queue, report a 0 waiting count
This issue was brought up on IRC by jmls
git-svn-id: https://origsvn.digium.com/svn/asterisk/branches/1.4@103956 65c4cc65-6c06-0410-ace0-fbb531ad65f3
