RFC says SIP headers look like:
HCOLON = *( SP / HTAB ) ":" SWS
SWS = [LWS] ; sep whitespace
LWS = [*WSP CRLF] 1*WSP ; linear whitespace
WSP = SP / HTAB ; from rfc2234
chan_sip implemented this:
HCOLON = *( LOWCTL / SP ) ":" SWS
LOWCTL = %x00-1F ; CTL without DEL
This discrepancy meant that SIP proxies in front of Asterisk with
chan_sip could pass on unknown headers with \x00-\x1F in them, which
would be treated by Asterisk as a different (known) header. For
example, the "To\x01:" header would gladly be forwarded by some proxies
as irrelevant, but chan_sip would treat it as the relevant "To:" header.
Those relying on a SIP proxy to scrub certain headers could mistakenly
get unexpected and unvalidated data fed to Asterisk.
This change fixes so chan_sip only considers SP/HTAB as valid tokens
before the colon, making it agree on the headers with other speakers of
SIP.
ASTERISK-26433 #close
AST-2016-009
Change-Id: I78086fbc524ac733b8f7f78cb423c91075fd489b
When an opus offer or answer was received that contained an
fmtp line with spaces between the attributes the module would
fail to properly parse it and crash due to recursion.
This change makes the module handle the space properly and
also removes the recursion requirement.
ASTERISK-26579
Change-Id: I01f53e5d9fa9f1925a7365f8d25071b5b3ac2dc3
The PJSIPShowRegistrationsInbound AMI command was just dumping out
all AORs which was pretty useless and resource heavy since it had
to get all endpoints, then all aors for each endpoint, then all
contacts for each aor.
PJSIPShowRegistrationInboundContactStatuses sends ContactStatusDetail
events which meets the intended purpose of the other command and has
significantly less overhead. Also, some additional fields that were
added to Contact since the original creation of the ContactStatusDetail
event have been added to the end of the event.
For compatibility purposes, PJSIPShowRegistrationsInbound is left
intact.
ASTERISK-26644 #close
Change-Id: I326f12c9ecb52bf37ba03f0748749de4da01490a
Fix the tests for DNS to use older style nameser.h as
in ASTERISK-26608.
Tested on: OpenBSD 6.0, Debian 8
ASTERISK-26647 #close
Change-Id: I285913c44202537c04b3ed09c015efa6e5f9052d
Occasionally SIP message transactions are not found when they should be.
In the particular case an incoming INVITE transaction is CANCELed but the
INVITE transaction cannot be found so a 481 response is returned for the
CANCEL. The problematic calls have a '_' character in the Via branch
parameter.
The problem is in the pjproject PJ_HASH_USE_OWN_TOLOWER feature's code.
The problem with the "own tolower" code is that it does not calculate the
same hash value as when the pj_tolower() function is used. The "own
tolower" code will erroneously modify the ASCII characters '@', '[', '\\',
']', '^', and '_'. Calls to pj_hash_calc_tolower() can use the
PJ_HASH_USE_OWN_TOLOWER substitute algorithm when enabled. Calls to
pj_hash_get_lower(), pj_hash_set_lower(), and pj_hash_set_np_lower() call
find_entry() which never uses the PJ_HASH_USE_OWN_TOLOWER algorithm. As a
result you may not be able to find a hash tabled entry because the
calculated hash values would differ.
* Simply disable PJ_HASH_USE_OWN_TOLOWER.
ASTERISK-26490 #close
Change-Id: If89bfdb5f301b8b685881a9a2a6e0c3c5af32253
This is a semi-regression caused by the iostreams change. Prior to
iostreams, HTTP headers were written to a FILE handle using fprintf.
Then the body was written using a call to fwrite(). Because of internal
buffering, the result was that the HTTP headers and body would be sent
out in a single write to the socket.
With the change to iostreams, the HTTP headers are written using
ast_iostream_printf(), which under the hood calls write(). The HTTP body
calls ast_iostream_write(), which also calls write() under the hood.
This results in two separate writes to the socket.
Most HTTP client libraries out there will handle this change just fine.
However, a few of our testsuite tests started failing because of the
change. As a result, in order to reduce frustration for users, this
change alters the HTTP code to write the headers and body in a single
write operation.
ASTERISK-26629 #close
Reported by Joshua Colp
Change-Id: Idc2d2fb3d9b3db14b8631a1e302244fa18b0e518
ast_iostream_printf() attempts first to use a fixed-size buffer to
perform its printf-like operation. If the fixed-size buffer is too
small, then a heap allocation is used instead. The heap allocation in
this case was exactly the length of the string to print. The issue here
is that the ensuing call to vsnprintf() will print a NULL byte in the
final space of the string. This meant that the final character was being
chopped off the string and replaced with a NULL byte. For HTTP in
particular, this caused problems because HTTP publishes the expected
Contact-Length. This meant HTTP was publishing a length one character
larger than what was actually present in the message.
This patch corrects the issue by adding one to the allocation length.
ASTERISK-26629
Reported by Joshua Colp
Change-Id: Ib3c5f41e96833d0415cf000656ac368168add639
Added back in a -g3, and an -O3 when DONT_OPTIMIZE is not set, to
the CFLAGS. Not sure how they went missing.
Also fixed an uninstall problem where we weren't removing the
symlink from libasteriskpj.so.2 to libasteriskpj.so. While I was
there, I fixed it for libasteriskssl as well.
Change-Id: I9e00873b1e9082d05b5549d974534b48a2142556
Increasing the testsuite shutdown timeout before forcibly killing
Asterisk allowed more events to be sent out. Some tests failed as
a result. The tests/channels/pjsip/statsd/registrations failed
because we now get the statsd events that a comment in the test
configuration stated couldn't be intercepted. Unfortunately, we
get a variable number of events because of internal status state
transition races generating redundant statsd events.
We were reporting redundant statsd PJSIP.registrations.state changes
for internal state changes that equated to the same thing publicly.
* Made update_client_state_status() filter out redundant statsd
updates.
ASTERISK-26527
Change-Id: If851c7d514bb530d9226e4941ba97dcf52000646
OpenSSL 1.1.0 includes some major changes in the interface. See
https://wiki.openssl.org/index.php/1.1_API_Changes .
Status: Right now there are still a few deprecation notes with OpenSSL
1.1.0. But it's a start.
Changes:
* CRYPTO_LOCK is no longer available. Replace it with its value for now.
I don't completely understand what it is used for there.
* Remove several functions from libasteriskssl that seem to no longer be
needed.
* Structures have become opaque and are accesses with accessors.
* ERR_remove_thread_state() no longer needed.
* SSLv2 code now could no longer be used in 1.1.
ASTERISK-26109 #close
Change-Id: I5e29d477d486ca29b6aae0dc2f5dff960c1cb82b
The latest Release candidate fails to create RTP streams when IPv6
is not available. Due to the changes made in September the ast_sockaddr
structure passed around to create these streams is always of AF_INET6
type, causing failure when used for IPv4. This patch adds a utility
function to check for availability of IPv6 and applies such check
at startup to determine how to create the ast_sockaddr structures.
ASTERISK-26617 #close
Change-Id: I627a4e91795e821111e1cda523f083a40d0e0c3e
Use of the new logging is as simple as issuing the new CLI command or
setting the new pjproject.conf option.
Other options that can affect the logging are how you have the pjproject
log levels mapped to Asterisk log types in pjproject.conf and if you have
configured Asterisk to log the DEBUG type messages. Altering the
pjproject.conf level mapping shouldn't be necessary for most installations
as the default mapping is sensible. Configuring Asterisk to log the DEBUG
message type is standard practice for collecting debug information.
* Added CLI "pjproject set log level" command to dynamically adjust the
maximum pjproject log message level.
* Added CLI "pjproject show log level" command to see the currently set
maximum pjproject log message level.
* Added pjproject.conf startup section "log_level" option to set the
initial maximum pjproject log message level so all messages could be
captured from initialization.
* Set PJ_LOG_MAX_LEVEL to 6 to compile in all defined logging levels into
bundled pjproject. Pjproject will use the currently set run time log
level to determine if a log message is generated just like Asterisk
verbose and debug logging levels.
* In log_forwarder(), made always log enabled and mapped pjproject log
messages. DEBUG mapped log messages are no longer gated by the current
Asterisk debug logging level.
* Removed RAII_VAR() from res_pjproject.c:get_log_level().
ASTERISK-26630 #close
Change-Id: I6dca12979f482ffb0450aaf58db0fe0f6d2e5389
The recent change that made frame deferral into an API had a behavior
change to it. When frame deferral was completed, we would take all of
the deferred frames and queue them all onto the channel in one call to
ast_queue_frame_head(). Before frame deferral was API-ized, places that
performed manual frame deferral would actually take each deferred frame
and queue them onto the channel.
This change in behavior caused the confbridge_recording test to start
failing consistently. Without going too crazily deep into the details,
a channel was getting "stuck" in an ast_safe_sleep(). An AMI redirect
was attempting to break it out of the sleep, but because there were more
frames in the channel read queue than expected, the channel ended up
being unable to break from its sleep loop.
By restoring the behavior of individual frame queuing after deferral,
the test starts passing again.
Note, this points to a potential underlying issue pointing to an
"unbalance" that can occur when queuing multiple frames at once,
and so a follow-up issue is being created to investigate that
possibility.
Change-Id: Ied5dacacda06d343dea751ed5814a03364fe5a7d
The sending codec is switched to the receiving codec and then
is switched back to the best native codec on EVERY receiving RTP packets.
This is because after call of ast_channel_set_rawwriteformat there is call
of ast_set_write_format which calls set_format which sets rawwriteformat
to the best native format.
This patch adds a new function ast_set_write_format_path which set
specific write path on channel and uses this function to switch
the sending codec.
ASTERISK-26603 #close
Change-Id: I5b7d098f8b254ce8f45546e6c36e5d324737f71d
Issue/patch ASTERISK-26587 was inspired by issue ASTERISK-22992
that requested ability to add callerid into app_originate.
Comments in that issue suggested that it was better solved by
adding an option to gosub prior to originating the call. The
attached patch implements this much like app_dial with two
options one to gosub on the originating channel and one to gosub
on the newly created channel and behaves just like app_dial.
I have tested this patch by adding callerid info to the new
channel and also SIPAddHeader (to e.g. add header to force auto
answer) and confirmed it works. Have also tested both 'exten'
and 'app' versions of app_originate.
Opened by: dkerr
Patch by: dkerr
Change-Id: I36abc39b58567ffcab4a636ea196ef48be234c57
The response from gmail calendar includes the string name
"caldav:calendar-data". res_calendar_caldav implements
the example included in RFC 4791: string "C:calendar-data".
When reading the calendar, res_calendar_caldav compare the
string and if does not match just discards the event.
This commit compares the response to both strings,
successfully loading gmail calendar events.
Writing to gmail calendar is working prior to this fix.
ASTERISK-26624
Reported by: Eduardo S. Libardi
Change-Id: Ia1eef10552ae616efb645d390f5ffe81260d7d4a
Per RFC 7118 5.2, the SIP URI 'transport' parameter should advertise
'ws' when WebSockets are to be used as the transport. This applies to
both secure and insecure WebSockets.
There were two bugs in Asterisk with respect to this:
(1) The most egregious occurs in res_pjsip. There, we advertise 'ws' for
insecure websockets and 'wss' for secure websockets. While this
would seem to make sense - since 'WS' and 'WSS' are used for the Via
Transport parameter - this is not the case for the SIP URI. This
patch corrects that by registering the secure websockets with
pjproject using the shorthand 'WS', and by returning 'ws' when asked
for the transport parameter. Note that in pjproject, it is perfectly
valid to have multiple transports use the same shorthand.
(2) In chan_sip, we return an upper-case version of the transport 'WS'
instead of 'ws'. Since we should be strict in what we send and
liberal in what we accept (within reason), this patch lower-cases
the transport before appending it to the parameter.
ASTERISK-24330 #close
Reported by: cervajs, Inaki Baz Castillo
Change-Id: Iff77b645f8cc3b7cd35168a6676c26b147f22f42
The asterisk.h header file needs to be included first or else
some things go awry, such as:
implicit declaration of function 'vasprintf'
Change-Id: I981dc2a77a1ba791888e4f1726644d4656c0407c
If a TCP/TLS connection was pending (not accepted and not timed out) during
unload of chan_sip, Asterisk would segfault when trying to send a signal to
a thread whose thread ID hadn't been recorded yet. This commit fixes that by
recording the thread ID before calling the blocking connect() syscall.
This was a regression introduced by 776a14386a.
The above wasn't enough to fix the segfault, which was now delayed to the
point where connect() timed out. Therefore, it was necessary to also remove
the SA_RESTART flag from the SIGURG sigaction so that pthread_kill() could be
used to interruput the connect() syscall.
This was a regression introduced by 5d313f51b9.
ASTERISK-26586 #close
Change-Id: I76fd9d47d56e4264e2629bce8ec15fecba673e7b
Switched from AST_MODULE_LOAD_DECLINE to AST_MODULE_LOAD_FAILURE.
Therefore, if pbx_lua fails to load and pbx_lua is marked as required,
Asterisk exits as expected.
If extensions.lua cannot be opened, AST_MODULE_LOAD_DECLINE is reported.
Change-Id: I8e5a0037e69b41743db60c568541ebb2f52a7a8f
When retrieving RTCP stats for PJSIP channels, RTT values are unreliable.
RTT calculation is correct, but the data representation isn't. RTT is
represented by a 32-bit fixed-point number with the integer part in the
first 16 bits and the fractional part in the last 16 bits. In order to
get the RTT value, the fractional part is miscalculated, there is an
unnecessary 16 bit shift that causes overflow. Besides this there is
another mistake, when transforming the integer value to the fixed point
fractional part via bitwise operation, that loses precision.
* RTT fractional part is no longer shifted, avoiding overflow.
* RTT fractional part is transformed to its fixed-point value more
precisely.
* Fixed timeval2ntp() and ntp2timeval() second fraction conversions.
* Fixed NTP timestamp report logging. The usec was inexplicably
multiplied by 4096.
ASTERISK-26566 #close
Reported by Hector Royo Concepcion
Change-Id: Ie09bdabfee75afb3f1b8ddfd963e5219ada3b96f
Previously, a TLS server socket would only be restarted upon sip reload if the
bind address had changed. This commit adds checking for changes to TLS
parameters like certificate, ciphers, etc. so they get picked up without
requiring a reload of the entire chan_sip module. This does not affect open
connections in any way, but new connections will use the new TLS parameters.
The changes also apply to HTTP and Manager.
ASTERISK-26604 #close
Change-Id: I169e86cefc6dcd627c915134015a6a1ab1aadbe6