Modify chan_sip's packet generation api to automatically calculate the Content-Length. This is done by storing packet content in a buffer until it is actually time to send the packet, at which time the size of the packet is calculated. This change was made to ensure that the Content-Length is always correct.
This change also adds an ast_str_copy_string() function (similar to ast_copy_string), that copies one ast_str into another, properly handling embedded nulls.
Fix crash when parsing some heavily nested statements in AEL on reload.
Due to the recursion used when compiling AEL in gen_prios, all the stack space
was being consumed when parsing some AEL that contained nesting 13 levels deep.
Changing a few large buffers to be heap allocated fixed the crash, although I
did not test how many more levels can now be safely used.
(This is a backport of 269307, committed to trunk by rmudgett.)
Calling dahdi_indicate() when the channel private lock is already
held can cause a deadlock if the PRI lock is needed because
dahdi_indicate() will also get the channel private lock. The pri_grab()
function assumes that the channel private lock is held once to avoid
deadlock.
David Vossel [Thu, 17 Jun 2010 18:45:32 +0000 (18:45 +0000)]
retransmit response to BYE requests until timer J expires
According to RFC 3261 section 17.2.2, which describes non-INVITE server
transaction, when a dialog enters the Completed state it must destroy
the dialog after Timer J (T1*64) fires. For a BYE transaction Asterisk
terminates the dialog immediately during sip_hangup() when it should be
waiting T1*64 ms. This results in some odd behavior. For instance if
Asterisk receives a BYE and transmits a 200ok in response, if the endpoint
never receives the 200ok it will retransmit the BYE to which Asterisk
responds with a "481 Call leg/transaction does not exist" because the
dialog is already gone.
To resolve this I made a function called sip_scheddestroy_final(). This
differs slightly from sip_schedestroy() in that it enables a flag that
will prevent the destruction from ever being rescheduled or canceled
afterwards. It also prevents the pvt's needdestroy flag from being set
which triggers the destruction of the dialog within the do_monitor thread().
By using this function we are guaranteed destruction will not occur
until the scheduled time. This allows Asterisk to respond to any possible
retransmits for a dialog after we process the initial BYE request for T1*64 ms.
Other changes: I removed two instances where sip_cancel_destroy is used
right before calling sip_scheddestroy. sip_scheddestroy always calls
sip_cancel_destroy before scheduling the new destruction so it is completely
unnecessary.
There is code in chan_iax2.c that attempts to guarantee that only a single
active thread will handle a call number at a time. This code works once
the thread is added to an active_list of threads, but we are not currently
guaranteed that a newly activated thread will enter the active_list immediately
because it is left up to the thread to add itself after frames have been
queued to it. This means that if two frames come in for the same call number
at the same time, it is possible for them to grab two separate threads because
the first thread did not add itself to the active_list fast enough. This
causes some pretty complex problems.
This patch resolves this race condition by immediately adding an activated
thread to the active_list within the network thread and only depending on
the thread to remove itself once it is done processing the frames queued to
it. By doing this we are guaranteed that if another frame for the same call
number comes in at the same time, that this thread will immediately be found
in the active_list of threads.
Terry Wilson [Tue, 15 Jun 2010 21:42:33 +0000 (21:42 +0000)]
Don't continue sending the file when there has been an error
If there is a problem with a firmware file, Polycom phones will close the
connection. We were continuing to send the file anyway. There should be no
reason to continue sending a file if there is an error writing it.
Terry Wilson [Tue, 15 Jun 2010 21:10:15 +0000 (21:10 +0000)]
Don't send files twice and remove extra \r\n from header
After the manager http auth changes, we forgot to remove the manual
sending of the file. Also, ast_http_send adds two \r\n to the header that
is passed to it, so a trailing \r\n is removed from the Content-type
header. It might be better to change ast_http_send, but I don't like changing
the behavior of an API function.
Terry Wilson [Tue, 15 Jun 2010 20:18:04 +0000 (20:18 +0000)]
Make contactdeny apply to src ip when nat=yes
chan_sip's "contactdeny" feature screens the "to be registered contact".
In case of nat=yes it should not use the address information from the
Contact header (which is not used at all for routing), but the source
IP address of the request.
Thus, if nat=yes and a client sends a request from a denied IP address
(e.g. by spoofing the src-IP address) it can bypass the screening.
This commit makes contactdeny apply to the src ip when nat=yes instead.
Jason Parker [Mon, 14 Jun 2010 19:41:43 +0000 (19:41 +0000)]
Add option to get untruncated channel name from AGENT function.
The "channel" option would chop the channel name at the last '-', which made
it useless for something like a channel transfer from the dialplan. The
"fullchannel" option will return the channel name as-is.
Tzafrir Cohen [Sun, 13 Jun 2010 09:16:25 +0000 (09:16 +0000)]
bashism in configure script
Theoretically the ./configure script is a pure bourne-shell script.
Practically it may be run by bash if /bin/sh is not good enough. But we should not count on it. See bug report for the gory details.
(closes issue #17485)
Patches:
0001-remove-bashism-from-ast_check_pwlib.m4.patch uploaded by tzafrir (license 46)
Fix potential crash when writing raw SLIN audio on a PLC-enabled channel.
The issue here was that the frame created when adjusting for PLC had no offset
to its audio data. If this frame were translated to another format prior to
being sent out an RTP socket, all went well because the translation code would
put an appropriate offset into the frame. However, if the SLIN audio were not
translated before being sent out the RTP socket, bad things would happen.
Specifically, the ast_rtp_raw_write makes the assumption that the frame has
at least enough of an offset that it can accommodate an RTP header. This was
not the case. As such, data was being written prior to the allocation, likely
corrupting the data the memory allocator had written. Thus when the time came
to free the data, all hell broke loose. ....Well, Asterisk crashed at least.
The fix was just what one would expect. Offset the data in the frame by a reasonable
amount. The method I used is a bit odd since the data in the frame is 16 bit integers
and not bytes. I left a big ol' comment about it. This can be improved on if someone
is interested. I was more interested in getting the crash resolved.
........
Russell Bryant [Wed, 9 Jun 2010 21:11:43 +0000 (21:11 +0000)]
Resolve an invalid memory read on an event.
Valgrind pointed out that attempting to get an IE value from an event that has
no IEs produces an invalid memory read past the end of the event. Thanks to
mmichelson for pointing the problem out to me and then testing the fix.
When using the init script as-is currently, it could cause issues on Debian
such as high CPU usage. This fix has worked for several people so I'm
implementing the change. We now handle color displays properly.
Richard Mudgett [Wed, 9 Jun 2010 16:54:38 +0000 (16:54 +0000)]
Eliminate deadlock potential in dahdi_fixup().
Calling dahdi_indicate() within dahdi_fixup() while the owner pointers are
in a potentially inconsistent state is a potentially bad thing in
principle.
However, calling dahdi_indicate() when the channel private lock is already
held can cause a deadlock if the PRI lock is needed because
dahdi_indicate() will also get the channel private lock. The pri_grab()
function assumes that the channel private lock is held once to avoid
deadlock.
David Vossel [Wed, 9 Jun 2010 15:09:25 +0000 (15:09 +0000)]
fixes crash in moh when cachertclasses flag is used
The result for moh_register was not verified to guarantee
the mohclass as added to the container.
(closes issue #16993)
Reported by: dmitri
Patches:
res_musiconhold_rtclass2.patch uploaded by dmitri (license 1001)
moh_crash2.diff uploaded by dvossel (license 671)
Tested by: dmitri
Tzafrir Cohen [Wed, 9 Jun 2010 13:17:43 +0000 (13:17 +0000)]
dial by name in chan_dahdi
* chan_dahdi supports dialing configuring and dialing by device file name.
DAHDI/span-name!local!1 will use /dev/dahdi/span-name/local/1 . Likewise
it may appear in chan_dahdi.conf as 'channel => span-name!local!1'.
* A new options for chan_dahdi.conf: 'ignore_failed_channels'. Boolean.
False by default. If set, chan_dahdi will ignore failed 'channel' entries.
Handy for the above name-based syntax as it does not depend on
initialization order.
* have my_pri_make_cc_dialstring() only manupulate dial-strings of group
(gGrR) dialing, which make it lsightly more complicated.
Bradley Latus [Tue, 8 Jun 2010 23:48:17 +0000 (23:48 +0000)]
Add High Resolution Times to CDRs for Asterisk
People expressed an interest in having access to the exact length of calls to a finer degree than seconds. See the CHANGES and UPGRADE.txt for usage also updated the sample configs to note the change.
Terry Wilson [Tue, 8 Jun 2010 05:29:08 +0000 (05:29 +0000)]
Add SRTP support for Asterisk
After 5 years in mantis and over a year on reviewboard, SRTP support is finally
being comitted. This includes generic CHANNEL dialplan functions that work for
getting the status of whether a call has secure media or signaling as defined
by the underlying channel technology and for setting whether or not a new
channel being bridged to a calling channel should have secure signaling or
media. See doc/tex/secure-calls.tex for examples.
Original patch by mikma, updated for trunk and revised by me.
Tilghman Lesher [Mon, 7 Jun 2010 19:52:39 +0000 (19:52 +0000)]
Seems strange (and the code backs up) that if the max and min of a statistic is expressed as a double, the last value would not also need to be a double.
Richard Mudgett [Mon, 7 Jun 2010 15:51:39 +0000 (15:51 +0000)]
Suppress warning in waitstream_core().
Suppress the warning about unexpected control subclass frames for
AST_CONTROL_CONNECTED_LINE, AST_CONTROL_REDIRECTING, and AST_CONTROL_AOC
in file.c:waitstream_core().
Tilghman Lesher [Sat, 5 Jun 2010 17:55:28 +0000 (17:55 +0000)]
Fix crash in DTMF detection.
What I did not originally see in my previous commit was that even though the
next digit could be detected before the previous was considered ended, the
detection of the next digit effectively ends the detection of the previous.
Therefore, the length moves in lockstep with the digit, and no separate counter
is needed for the length alone.
projects / thirdparty / asterisk.git / log
