In the previous implementation, the processing time was not being taken
into account, which was implicitly added for each new packet to be sent,
which caused a steady incremental drift in the clock clearly visible
when analysing a RTP stream.
As it uses timespecsub, it depends on libosmocore Change-Id
I45fc993b9bb0a343763238bf463c8640f47b00f1.
Change-Id: I11cb9a63e16e829ccd4af1096b9f473c802d822f
Introduce osmo_stream_srv_set_flush_and_destroy() which marks a
stream to be 'flushed and destroyed'. No new messages will be
received on this stream, and no new messages can be queued.
Once the Tx queue has been drained, the connection is destroyed.
The API user is given a chance to perform cleanup operations
in the closed_cb() callback for the connection.
The same mechanism will be added for client-side connections
in a follow-up patch.
Change-Id: I8ed78fe39c463e9018756700d13ee5ebe003b57f
Related: OS#2789
Suggested-by: Harald Welte
On destroying a client or server stream, deallocate any msgbs that are still
pending in the queue.
In libosmo-sccp, the ss7_test.c in test_as(), messages are queued and were,
before this, left floating after the stream was destroyed, causing a sanitizer
memory leak. This patch fixes the leak.
Depends: Ia291832ca445d4071f0ed9a01730d945ff691cf7 (libosmocore)
Change-Id: Iaad35f03e3bdfabf3ba82b16e563c0a5d1f03639
In previous implementation, if no reconfiguring is needed, a new socket
would be created without closing the old one, leaking the previous
socket. Instead, if we don't need reconfiguring, we return 0 as no
operation is required.
Change-Id: I6c1a7fff63e44840fb5e2bc7ace5e9a61e304987
See osmo-ci change I2409b2928b4d7ebbd6c005097d4ad7337307dd93 for rationale.
Depends: I2409b2928b4d7ebbd6c005097d4ad7337307dd93
Change-Id: I6bdbd22b5f05ab75de2c6c777667bf1a63eb0b86
This fixes:
dpkg-gencontrol: warning: Depends field of package libosmo-netif-dev: unknown substitution variable ${shlibs:Depends}
Change-Id: Ic57cbc8a1876e6e5449e7dd4a73f269444dd67da
Short changelog:
* Add Doxygen documentation
* SCTP support in stream.c
* new udp-test-client and udp-test-server programs
* better / more verbose error handling in examples
* new osmo_dgram_tx_set_local_{addr,port}() functions
* use IPA definitions from libosmogsm, rather than repeating them
* encode RTP header M field of RFC3550/4867 in OSMUX header
* new osmo_stream_srv_link_set_nodelay()
* new osmo_stream_srv_link_set_proto()
* new osmo_stream_cli_set_nodelay()
* new osmo_stream_cli_set_proto()
* new osmo_stream_cli_set_local_addr()
* new osmo_stream_cli_set_local_port()
* new osmo_stream_cli_reconnect()
* new osmo_stream_cli_open2() with reconnect argument
* more vrebose osmux_snprintf()
* remove mistaken reference to AGPL in rs232.c
* fix memory leak in osmo_stream_srv_link_set_addr()
* add osmo-pcap-test for SLL and Ethernet
* extend osmux-test
Change-Id: Ibf75fcd6643351ce3946faa155ae1db8c33a5e35
Previous implementation handled all types as if they were Osmux AMR
frames. For Dummy frames, we account the padding but we don't care about
the padding content. For Signalling ones, as they are not in the
specification yet, it is better avoid using unespecified fields and
return an error because it's still not known how extra data will be
handled in the input msgb.
Change-Id: I48565472b47c2a0e5db50881fbb005537af8c70d
The current code still expects to parse only AMR osmux frames, but that
will be fixed in following patches.
Change-Id: Ic2f4d1d3cc88af912bb43c8ecd90eacc6ff7190f
Previously recv() returning 0 for UDP socket was considered as error
although it's legitimate return value for empty UDP packets. Relax the
error check to avoid confusing error messages.
The function behavior is the same:
* msg is not altered while receiving 0-length UDP packet
* return value is 0
The only result of the relaxed error check is the absense of error log
message for 0-length UDP packet.
Change-Id: I32e5fcbf5ed92cc923660ac59e6a37fd3f0703a7
Fixes: OS#2219
In addition to showing basic UDP send/receive example, it helps to test
corner-case when dealing with 0-length UDP packets.
Change-Id: I08c0adf1cf9b6a6f1f7090b237d0497c2ec13cdf
Related: OS#2219
There's no need for the rs232 code to include a files from
libosmoabis. The only users of libosmoabis left now are the LAPD
examples:
examples/lapd-over-datagram-network.c
examples/lapd-over-datagram-user.c
Change-Id: Ie1bc0dd811362cec546486edc41d632740ed19cd
This patch inconditionally initializes the buffer we get to
nul-terminate it, whenever possible. It's a very simple solution to
catch three overly corner cases:
1) snprintf() returns -1, very much unlikely in practise.
2) msg->len == 0: In such case, I would expect this function is never
called with an empty message, but let's be safe in this case too.
3) If your buffer is empty, it doesn't nul-terminate the buffer.
Change-Id: I97e517f2d98e83894ea707c63489559302ff6bd2
SNPRINTF_BUFFER_SIZE() looks too complex, previous version maintains two
different variables to account for the remaining space in the buffer,
one of them is always decremented based on what snprintf() returns,
which may result in underflow. These variables are swapped - not used
consistently - all over this code.
Replace this macro by a simplified version, with one single parameter to
account for remaining space. This macro also deals with two corner
cases:
1) snprintf() fails, actually never happens in practise, but
documentation indicates it may return -1, so let's catch this case
from here to stick to specs.
2) There is not enough space in the buffer, in that case, keep
increasing offset, so we know how much would have been printed, just
like snprintf() does.
Thanks to Pau Espin for reporting, and Holger for clues on this.
I have run osmux_test and, at quick glance, it looks good.
Change-Id: I5b5d6ec57a02f57c23b1ae86dbd894bad28ea797
The buffer for osmux_test is increased as the former doesn't seem to be
able to cope with the whole output.
Change-Id: Ic838dd9d7ad89b4510ccfa58c0390c69a075b616
Before this patch, ETH was assumed and other types were not supported.
This patch also adds Linux cooked packet support for L2.
Change-Id: Ie62fa0a8e45e1e141edb64b116dad185ad9c7a5f
This flag, when set, allows to unconditionally disable doxygen
documentation generation, even if doxygen command is found.
Change-Id: I16b3502be3e32274f548da6a2a0b0363dd3bfe5d
We didn't check for cases where setsockopt_nodelay() fails. Let's check
for that and bail out + close the socket.
Change-Id: I0adbc4cec35be7c36bdf01d4d8fefd6097e9be5d
Fixes: coverity CID#166970
We couldn't figure out what "crx" as supposed to mean and assume
it is a typo. Fix the code and call it ctx, this is fixing the
API documentation as well.
Change-Id: I27ed1178fdbbcf3fc0e1070dc19b4ecf9a327a04