Age | Commit message (Collapse) | Author | Files | Lines |
|
Depends: I5021c64a787b63314e0f2f1cba0b8fc7bff4f09b
Change-Id: I0fc8142756040c98c1862ddb18a77a032f2a7143
|
|
Currently the connect is blocking leading to one bad connection
(e.g. to a black hole) blocking all other connections leading to
bursty traffic.
Change-Id: Idfca8acbce09176055da3e577566386e07d7a348
|
|
Change-Id: Id3c41595475dac06dcdf3a104570640349c8f8bf
|
|
It seems that the Ubuntu linker is quite picky in where it is looking
for symbols. Link libosmocore (and as such talloc) last of the Osmocom
libraries. This could fix:
gcc -Wall -I/usr/include/ -I/usr/include/ -I/usr/include/ -I/usr/include/ -DNO_UNIPORTE -g -O2 -fdebug-prefix-map=/usr/src/packages/BUILD=. -fstack-protector-strong -Wformat -Werror=format-security -Wl,-Bsymbolic-functions -Wl,-z,relro -o cellmgr_ng main.o mtp_layer3.o thread.o ipaccess.o pcap.o bss_patch.o bssap_sccp.o bsc_sccp.o bsc_ussd.o links.o msc_conn.o link_udp.o snmp_mtp.o debug.o isup.o mtp_link.o counter.o sccp_state.o bsc.o ss7_application.o vty_interface_legacy.o vty_interface_cmds.o mgcp_patch.o mgcp_callagent.o isup_filter.o -ltalloc -losmocore -ltalloc -losmogsm -losmocore -ltalloc -losmovty -losmocore -losmosccp -losmoxua -lpthread -lnetsnmp -lcrypto
[ 79s] /usr/lib/gcc/x86_64-linux-gnu/6/../../../x86_64-linux-gnu/libosmosccp.so: undefined reference to `talloc_free'
Change-Id: I94c87496ebffd41a6217d860b0ca4bed6ff2d24c
|
|
There seems to be issues with the write handling of the osmo-stp
and so far we did not enable vty tests here. Make it possible to
enable the tests, fix the VTY strings, change string to OsmoSTP.
Change-Id: I547fd4840d86ce16e8589fb63802dd7099781194
|
|
Change-Id: I4ed058aba15d9fdc22dde99d60df0aabd6585b75
|
|
mgcp/mgcp_protocol.c: In function ‘mgcp_handle_message’:
/home/ich/install/openbsc/include/osmocom/core/logging.h:59:4: warning: ‘save’ may be used uninitialized in this function [-Wmaybe-uninitialized]
logp2(ss, level, __BASE_FILE__, __LINE__, 0, fmt, ##args); \
^
mgcp/mgcp_protocol.c:368:15: note: ‘save’ was declared here
char *elem, *save;
Change-Id: I28152373b043c8a5b55a83ac09dd7bafce4d2d66
|
|
The return type changed from enum to int and it is time to catch
up with it.
vty_interface.c:108:2: warning: initialization from incompatible pointer type
.go_parent_cb = ss7_go_parent,
^
vty_interface.c:108:2: warning: (near initialization for ‘vty_info.go_parent_cb’)
mgcp_ss7_vty.c:36:2: warning: initialization from incompatible pointer type
.go_parent_cb = mgcp_go_parent,
^
mgcp_ss7_vty.c:36:2: warning: (near initialization for ‘vty_info.go_parent_cb’)
Change-Id: I8061df697daf79fa57b8313c15f6567ff152f4a5
|
|
The introduction of a 'if + do {} while' lead to not being able to
return the result of the logging anymore. Just return.
|
|
For M3UA we already skip the link-test handling and there is no
point in waiting for management messages to collect routing info.
We already know where to send things. At first I tried to overload
the skip handling but it is better to just set the timer to 0 and
then move on.
|
|
Make it possible to set it. 10s seems to be a very reasonable
timeout. It should not really take much longer than that.
|
|
For M3UA we should have one time-out for operation we want and
then re-transmit it. As this is too much work right now create
a single timer that waits that after a connect the ASPAC_ACK
state will be reached.
|
|
|
|
In case of back pressure we can not send messages quick enough
and will drop messages. Add a counter so we can count how often
this happened per link/linkset. The SCTP M2UA code is not changed
as we don't have a link in the hands and adding it is too risky
without tests (not all M2UA have a link index so some might not
have a link).
|
|
Try to make it clear that this is only for the IPA bridging
mode and not a general setting.
|
|
For the "bridging" to IPA mode we can force the SCTP/M3UA connection
down. This way the remote STP will do a proper link fail-over
procedure instead of the STP throwing data away.
This is not configurable yet.
|
|
This can be used to force one connection down while the other
is down. E.g. to force SCTP down when the IPA connection is
down as well.
|
|
Print the errno and RC to give some kind of idea about
the error.
|
|
|
|
We don't have regression testing here but this could be
the correct fix.
|
|
|
|
RFC4666:
The DATA message MUST NOT be sent on stream 0
|
|
|
|
Make the code work for both ISUP and SCCP and let the address
handling still be part of the linkset/mtp3 code.
|
|
the opc/dpc for ISUP is currently wrong. The send/recv path
for actual SCCP data is not tested.
|
|
|
|
Implement a basic M3UA state machine on a SCTP connection.
|
|
For M3UA links we can't send SLTM messages but we can
assume that the link is fine. Add a flag that a link
can set in case SLTM should be skipped. The linkset will
then directly wait for routable DPCs and bring the
SCCP part up.
|
|
|
|
|
|
|
|
Thanks to mnhauke for the patch
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
|
|
The ASP has not been looking at this value at all, fix the byte
order. On the other hand my ASP doesn't care about this either.
|
|
Currently this is only true for the same connection.
sctp_m2ua.c:195 Two active SCTP conns with 1.2.3.4 on 0x9c21cf0, 0x9c21cf0
|
|
Make it possible to configure the transmit power. These settings
will be in effect immediately (there is no lock between the two
threads but it is a read only).
|
|
We get a trap for detected tones and then need to query them. Print
the detected characters for now and do not forward them yet.
|
|
When an endpoint is allocated we do not have enough resources to
connect it again and this situation is permanent as long as the
endpoint is connected.
Returning 502 might help some switches to release stale audio resources.
|
|
Playing tones should be printable ascii characters, begin by checking
for NULL (as it happened before) as the first illegal tone. This is
an attempt to make the API more robust and detect errors more early.
|
|
Because of the issue parsing the MGCP request a '\0' was added to
the queue. This lead to the dtmf scheduler believing that a play
was in progress while the playing code didn't play anything. This
lead to the queue to be stuck and nothing being played at all.
Return the number of tones that should be played and stop using
strlen on the tones.
|
|
Print the number of log messages that are queued in the error message.
|
|
The RQNT message might not contain a 'S:' line with the actual tone
to play. Instead of calling the callback with the 0 as tone just leave
early.
Example:
X: 6B9519B88F0
R: D/[0-9#*](N), G/ft, fxr/t38
Merged from the OpenBSC development tree
|
|
For some equipment it is the easiest to patch the assignment
complete message transported to the MSC. Add a VTY config to
enable this patching, create a testcase that tests that the
original message is truncated. The setting of the VTY option
has been manually tested. The entire system has not been end
to end tested.
|
|
|
|
We need to create a copy of the messages before we forward it
to MSC. This is already done in the other code path.
|
|
This code could be shared with the STP setup code.
|
|
This change was either never implemented or has been lost at
some point in time. Allow to pass SCCP messages un-modified through
the A-link.
|
|
|
|
* Patch all slots with the FR2 codec. This is the easiest way to
patch the entire message.
|
|
* Chosen Channel and Speech Version are both optional. Patch them
separately if they are present.
* Add testcase for the assignment request patching.
|
|
|