bs11
6 weeks agoadd 'ciph_state' member to BTS-side lchan definition master
Harald Welte [Sun, 8 Apr 2012 18:23:28 +0000]
add 'ciph_state' member to BTS-side lchan definition

6 weeks agointroduce HAVE_TM_GMTOFF_IN_TM
Harald Welte [Sun, 8 Apr 2012 14:59:24 +0000]
introduce HAVE_TM_GMTOFF_IN_TM

Not all architectures have the tm.tm_gmtoff member.  This fixes cygwin
builds.

6 weeks agomake sure we still build on systems that don't have IPPROTO_GRE
Harald Welte [Sun, 8 Apr 2012 13:56:00 +0000]
make sure we still build on systems that don't have IPPROTO_GRE

7 weeks agonat: Add another test case for TP-SRR clearing and number rewrite
Holger Hans Peter Freyther [Mon, 26 Mar 2012 14:21:42 +0000]
nat: Add another test case for TP-SRR clearing and number rewrite

2 months agomisc: Use other size modifiers to fix compiler warnings
Holger Hans Peter Freyther [Fri, 16 Mar 2012 11:18:39 +0000]
misc: Use other size modifiers to fix compiler warnings

control_if.c:521:2: warning: format ‘%lu’ expects argument of type ‘long unsigned int’, but argument 3 has type ‘uint64_t’ [-Wformat]
osmo_bsc_bssap.c:473:3: warning: format ‘%lu’ expects argument of type ‘long unsigned int’, but argument 7 has type ‘unsigned int’ [-Wformat]
mgcp_main.c:162:4: warning: format ‘%lu’ expects argument of type ‘long unsigned int’, but argument 4 has type ‘unsigned int’ [-Wformatt]

2 months agobsc: Prepare to have multiple MSC connections
Holger Hans Peter Freyther [Sat, 4 Jun 2011 17:58:26 +0000]
bsc: Prepare to have multiple MSC connections

We now have a list of MSCs but in the code we will
try to access the MSC with the nr 0.

2 months agobsc: Use the right connection for outgoing packets
Holger Hans Peter Freyther [Sat, 4 Jun 2011 15:34:02 +0000]
bsc: Use the right connection for outgoing packets

This is needed for simple UDT messages where we do not have
a SCCP connection.

2 months agobsc: Move more things to use osmo_msc_data* directly
Holger Hans Peter Freyther [Sat, 4 Jun 2011 13:34:37 +0000]
bsc: Move more things to use osmo_msc_data* directly

2 months agobsc: Move the bsc_filter to use the osmo_bsc_sccp_con
Holger Hans Peter Freyther [Sat, 4 Jun 2011 13:12:57 +0000]
bsc: Move the bsc_filter to use the osmo_bsc_sccp_con

2 months agobsc: Move away from ->bsc.msc to use the selected MSC
Holger Hans Peter Freyther [Sat, 4 Jun 2011 12:51:51 +0000]
bsc: Move away from ->bsc.msc to use the selected MSC

For multiple MSCs we should only have one place where the MSC
is selected and the rest will extract it from somewhere.

2 months agobsc: Stop using net->bsc->msc and get the right msc from somewhere else
Holger Hans Peter Freyther [Sat, 4 Jun 2011 11:20:04 +0000]
bsc: Stop using net->bsc->msc and get the right msc from somewhere else

2 months agobsc: Create a osmo_bsc_data and embed osmo_msc_data
Holger Hans Peter Freyther [Mon, 15 Aug 2011 13:53:00 +0000]
bsc: Create a osmo_bsc_data and embed osmo_msc_data

We want to have multiple MSCs but we also have some data
that is only present on a per BSC basis. Right now the
MSC data is not allocated with talloc, so we have some
change in the talloc contexts.

2 months agobsc: Address warnings about implicit declarations
Holger Hans Peter Freyther [Fri, 16 Mar 2012 10:54:55 +0000]
bsc: Address warnings about implicit declarations

osmo_bsc_main.c: In function ‘main’:
osmo_bsc_main.c:398:2: warning: implicit declaration of function ‘bts_init’ [-Wimplicit-function-declaration]
osmo_bsc_main.c:399:2: warning: implicit declaration of function ‘libosmo_abis_init’ [-Wimplicit-function-declaration]
osmo_bsc_main.c:418:2: warning: implicit declaration of function ‘bsc_bootstrap_network’ [-Wimplicit-function-declaration]

2 months agobsc: Fix compiler warning about wrong return and unused variable
Holger Hans Peter Freyther [Fri, 16 Mar 2012 10:49:46 +0000]
bsc: Fix compiler warning about wrong return and unused variable

osmo_bsc_api.c: In function ‘bsc_cm_update’:
osmo_bsc_api.c:195:2: warning: ‘return’ with a value, in function returning void [enabled by default]
osmo_bsc_api.c:193:28: warning: unused variable ‘sccp’ [-Wunused-variable]

2 months agomgcp: implement a more tolerant parser based on strtok_r()
Harald Welte [Fri, 27 Jan 2012 11:31:36 +0000]
mgcp: implement a more tolerant parser based on strtok_r()

Instead of building complex manual byte-wise parsers, we simply use two
strtok_r loops:  one iterating over all the lines, the next one
iterating over the invididual space-separated elements in the first line.

The benefit is that we now accept \r, \n or \r\n, or any multiple of
them as line ending.  This works around incompliant MGCP implementations
like that of Zynetix MSC.

Addition: mgcp_analyze_header returns 0 when all out parameters have
been set.

Signed-off-by: Holger Hans Peter Freyther <zecke@selfish.org>

2 months agomgcp: make sure all command line options are documented
Harald Welte [Fri, 27 Jan 2012 12:21:46 +0000]
mgcp: make sure all command line options are documented

Signed-off-by: Holger Hans Peter Freyther <zecke@selfish.org>

2 months agomgcp: consistently name bsc_mgcp osmo-bsc_mgcp
Harald Welte [Fri, 27 Jan 2012 12:18:07 +0000]
mgcp: consistently name bsc_mgcp osmo-bsc_mgcp

Signed-off-by: Holger Hans Peter Freyther <zecke@selfish.org>

2 months agomgcp_test: Add wrong messages as input and check what happens
Holger Hans Peter Freyther [Fri, 16 Mar 2012 08:53:49 +0000]
mgcp_test: Add wrong messages as input and check what happens

Test parsing some invalid messages and see if we start to crash.

2 months agomgcp_test: Rename method as it now operates on an array of messages
Holger Hans Peter Freyther [Fri, 16 Mar 2012 08:45:13 +0000]
mgcp_test: Rename method as it now operates on an array of messages

This code was cleaned up, catch up and rename the method.

2 months agomisc: Deal with smatch warnings about the bitfields
Holger Hans Peter Freyther [Fri, 2 Mar 2012 13:14:33 +0000]
misc: Deal with smatch warnings about the bitfields

Use unsigned ints for the bitfield.

2 months agomisc: Use msgb_free for freeing the messages
Holger Hans Peter Freyther [Thu, 1 Mar 2012 19:30:32 +0000]
misc: Use msgb_free for freeing the messages

2 months agoImproved generation of SYSTEM INFORMATIONs 2* and 5* for neighbour bands
Andreas Eversberg [Wed, 29 Feb 2012 08:04:07 +0000]
Improved generation of SYSTEM INFORMATIONs 2* and 5* for neighbour bands

In addition to SI 2 and SI 5, the SI 2ter and 2bis is generated, if
neighbour cells in other bands exist. Also it is indicated in the rest
octets of SI3, that SI 2ter is used. If no neighbour cell in a different
band exists, the SI 2ter and SI 5ter is omitted.

A special case is P-GSM range (channels 1-124). To be compatible with
older phones, SI 2bis and SI 5bis is used. If the BCCH lays inside the
P-GSM band, only neighbour cells of the P-GSM range are included in
SI 2 and SI 5. If neighbour cells exist in the same band (900), but lay
outside the P-GSM range, the SI 2bis and SI 5bis is used to extend the
list of neighbour cells. The extension is also indicated in SI 2 and
SI 5. If the BCCH lays inside the P-GSM range, but no neighbour cell
exists in the same band outside the P-GSM range, the SI 2bis ans SI 5bis
are omitted.

2 months agoMake sure USSD code only responds *#100# for retrieving own number
Tobias Engel [Fri, 24 Feb 2012 21:23:17 +0000]
Make sure USSD code only responds *#100# for retrieving own number

strstr() was used with wrong argument order, causing it to always match,
and causing an invalid response to a variety of different SS and USSD
requests.

This has apparently caused havoc among a number of HTC phones which
issue SS requests without user interaction upon boot, and then trip over
our inappropriate response.

3 months agofix generation of frequency list (E-GSM vs. P-GSM 900)
Harald Welte [Fri, 17 Feb 2012 14:58:23 +0000]
fix generation of frequency list (E-GSM vs. P-GSM 900)

3 months agobts_ipaccess_nanobts.c: Remove unused variable.
Holger Hans Peter Freyther [Fri, 3 Feb 2012 19:31:05 +0000]
bts_ipaccess_nanobts.c: Remove unused variable.

This must have been obsoleted by the move to libosmo-abis.

GCC warning:
bts_ipaccess_nanobts.c: In function ‘ipaccess_drop_oml’:
bts_ipaccess_nanobts.c:509:21: warning: variable ‘line’ set but not used [-Wunused-but-set-variable]

3 months agoabis_rsl.c: Remove unused variable net.
Holger Hans Peter Freyther [Fri, 3 Feb 2012 19:29:00 +0000]
abis_rsl.c: Remove unused variable net.

The old BSC code had code to override the payload type, this has
been removed, remove the variable accessing it.

GCC warning:
abis_rsl.c: In function ‘ipa_rtp_pt_for_lchan’:
abis_rsl.c:1590:22: warning: unused variable ‘net’ [-Wunused-variable]

3 months agoabis_rsl: The rach information was not used, introduce a signal
Holger Hans Peter Freyther [Fri, 3 Feb 2012 19:26:25 +0000]
abis_rsl: The rach information was not used, introduce a signal

Introduce a SS_CCCH for the paging and the rach load. The paging
code could now start using the signal.

GCC warning:
abis_rsl.c: In function ‘rsl_rx_ccch_load’:
abis_rsl.c:1371:11: warning: variable ‘rach_access_count’ set but not used [-Wunused-but-set-variable]
abis_rsl.c:1370:11: warning: variable ‘rach_busy_count’ set but not used [-Wunused-but-set-variable]
abis_rsl.c:1369:11: warning: variable ‘rach_slot_count’ set but not used [-Wunused-but-set-variable]

3 months agoabis_rsl.c: Remove unused variable ts_number
Holger Hans Peter Freyther [Fri, 3 Feb 2012 19:10:13 +0000]
abis_rsl.c: Remove unused variable ts_number

GCC warning:
abis_rsl.c: In function ‘rsl_rx_chan_rqd’:
abis_rsl.c:1245:10: warning: variable ‘ts_number’ set but not used [-Wunused-but-set-variable]

3 months agoabis_om2000.c: Remove unused variable to_trx_oml.
Holger Hans Peter Freyther [Fri, 3 Feb 2012 19:01:32 +0000]
abis_om2000.c: Remove unused variable to_trx_oml.

GCC warning:
abis_om2000.c: In function ‘abis_om2k_sendmsg’:
abis_om2000.c:804:6: warning: unused variable ‘to_trx_oml’ [-Wunused-variable]

3 months agovty: Remove two unimplemented OML functions
Holger Hans Peter Freyther [Fri, 3 Feb 2012 18:59:50 +0000]
vty: Remove two unimplemented OML functions

attribute get|set <0-255> (.HEX) was never implemented and the
output about the unused attributes clutter the build output, remove
them.

GCC warning:
abis_nm_vty.c: In function ‘oml_attrib_get’:
abis_nm_vty.c:141:25: warning: unused variable ‘oms’ [-Wunused-variable]
abis_nm_vty.c: In function ‘oml_attrib_set’:
abis_nm_vty.c:152:25: warning: unused variable ‘oms’ [-Wunused-variable]

3 months agomisc: abis_nm.c Use the result of ret.
Holger Hans Peter Freyther [Fri, 3 Feb 2012 18:48:30 +0000]
misc: abis_nm.c Use the result of ret.

Use LOGP(DNM, LOGL_ERROR, ...); for errors in the
abis_nm_rx_sw_act_req method.

GCC warning:
abis_nm.c: In function ‘abis_nm_rx_sw_act_req’:
abis_nm.c:412:6: warning: variable ‘ret’ set but not used [-Wunused-but-set-variable]

3 months agomisc: gsm_data.c address compiler warning
Holger Hans Peter Freyther [Fri, 3 Feb 2012 18:44:37 +0000]
misc: gsm_data.c address compiler warning

Assume nothing special needs to be done for the Nokia
*site hardware.

GCC warning:
gsm_data.c: In function ‘gsm_set_bts_type’:
gsm_data.c:342:2: warning: enumeration value ‘GSM_BTS_TYPE_NOKIA_SITE’ not handled in switch [-Wswitch]

3 months agoARFCNs are in the range of 0..1023, not 0..1024
Harald Welte [Sun, 29 Jan 2012 12:24:12 +0000]
ARFCNs are in the range of 0..1023, not 0..1024

I'm sure I read somewhere that it actually was 0..1024, as I kept
wondering how stupid it was to use 10bit+1.  However, that source
was incorrect, as GSM TS 05.05 quite clearly states 0..1023

3 months agomgcp: add missing prototypes for send_mgcp_reset_*
Harald Welte [Thu, 26 Jan 2012 23:46:49 +0000]
mgcp: add missing prototypes for send_mgcp_reset_*

3 months agoosmo-bsc: Remove hard-coded check for codec preferecne in ASS CMD
Harald Welte [Thu, 26 Jan 2012 23:43:06 +0000]
osmo-bsc: Remove hard-coded check for codec preferecne in ASS CMD

When we get an assignment command from the MSC, we no longer have to
check statically for certain codecs, as we have a vty-configured list of
codecs that is checked just below.

3 months agoMGCP: Add VTY commands to reset (RSIP) one or all endpoints
Harald Welte [Thu, 26 Jan 2012 23:41:39 +0000]
MGCP: Add VTY commands to reset (RSIP) one or all endpoints

3 months agobsc_mgcp: optionally connect the MGCP socket to the MGW
Harald Welte [Thu, 26 Jan 2012 23:38:59 +0000]
bsc_mgcp: optionally connect the MGCP socket to the MGW

This allows the use of the existing "call-agent ip A.B.C.D" command in
libmgcp in order to set a default destination address of the call agent.

This is required as a pre-condition for certain call agents (like the
zynetix MSC) that require a RSIP from the BSC to the MSC as the initial
MGCP message.

3 months agoMGCP: Add CRCX test cases
Harald Welte [Thu, 26 Jan 2012 23:37:23 +0000]
MGCP: Add CRCX test cases

The second test case (CRCX_ZYN) fails as the message is not compliant
with the spec.  However, our parer should be more tolerant, just in
case...

3 months agofurther generalize mgcp testing code
Harald Welte [Thu, 26 Jan 2012 23:36:56 +0000]
further generalize mgcp testing code

use a data-structure driven approach

3 months agomgcp_test: remove duplicate code
Harald Welte [Thu, 26 Jan 2012 23:00:44 +0000]
mgcp_test: remove duplicate code

3 months agoMove the bulk of RR processing from MSC into BSC
Harald Welte [Mon, 23 Jan 2012 15:40:24 +0000]
Move the bulk of RR processing from MSC into BSC

RR Messages like STATUS, GPRS SUSPEND, HANDOVER COMPLETE/FAIL, ...
should be processed on the BSC side of things, not on the MSC side.

This is among other things required in preparation of intra-BSC
hand-over support in osmo-bsc.

3 months agoMove processing of CLASSMARK CHANGE behind bsc_api
Harald Welte [Mon, 23 Jan 2012 09:28:35 +0000]
Move processing of CLASSMARK CHANGE behind bsc_api

This prevents osmo-bsc from sending RR messages to the MSC and rather
process them inside the BSC and turn them into BSSAP CM UPDATE.

3 months agoosmo-bsc: correctly parse the TMSI in BSSAP paging from MSC
Harald Welte [Fri, 20 Jan 2012 16:20:51 +0000]
osmo-bsc: correctly parse the TMSI in BSSAP paging from MSC

the TMSI is a uint32_t big-endian value and not a gsm 04.08 mobile
identity like the IMSI.

4 months agonat: Add VTY command for the TP-DA rewriting.
Holger Hans Peter Freyther [Wed, 18 Jan 2012 19:40:01 +0000]
nat: Add VTY command for the TP-DA rewriting.

4 months agonat: Fix the no sms-clear-tp-srr command
Holger Hans Peter Freyther [Wed, 18 Jan 2012 19:34:08 +0000]
nat: Fix the no sms-clear-tp-srr command

4 months agonat: Internationalize the TP-DA if it starts with a 00
Holger Hans Peter Freyther [Wed, 18 Jan 2012 19:15:29 +0000]
nat: Internationalize the TP-DA if it starts with a 00

Internationalize the number like we do it for Call Control. Update
the test result to match this new behavior.

4 months agonat: Prepare to rewrite the TP-DA number of a SMS submit.
Holger Hans Peter Freyther [Wed, 18 Jan 2012 19:00:28 +0000]
nat: Prepare to rewrite the TP-DA number of a SMS submit.

Introduce number rewriting of SMS-SUBMIT. Introduce a new list,
move code around to help with finding a new number, somehow the
number encoding for TP-DA is borked, 03.40 references 04.11 but
the length appears to be strlen(number) without taken the type
field into account.

4 months agonat: Print the sizes and the offending data in case of a failure
Holger Hans Peter Freyther [Wed, 18 Jan 2012 16:20:23 +0000]
nat: Print the sizes and the offending data in case of a failure

4 months agorsl: This is provided by libosmo-abis, remove it here
Holger Hans Peter Freyther [Tue, 17 Jan 2012 15:14:23 +0000]
rsl: This is provided by libosmo-abis, remove it here

4 months agomgcp: Fix warning about unused variable
Holger Hans Peter Freyther [Tue, 17 Jan 2012 15:00:08 +0000]
mgcp: Fix warning about unused variable

mgcp_network.c:335:22: warning: variable ‘cfg’ set but not used [-Wunused-but-set-variable]

4 months agomgcp: Provide documentation for the method, check sscanf return
Holger Hans Peter Freyther [Tue, 17 Jan 2012 14:58:48 +0000]
mgcp: Provide documentation for the method, check sscanf return

4 months agobts: Convert the rsl_link to be a struct ipabis_link
Holger Hans Peter Freyther [Tue, 17 Jan 2012 14:52:45 +0000]
bts: Convert the rsl_link to be a struct ipabis_link

4 months agolchan: Close a possible still open RTP Socket
Holger Hans Peter Freyther [Tue, 17 Jan 2012 14:27:33 +0000]
lchan: Close a possible still open RTP Socket

In case of a reset (loss of the BTS) close down all remaining
RTP Proxy sockets. In case of a lchan_free shout if the rtp
proxy is still open. I would prefer if the proxy code sits inside
the gsm subscriber connection.

4 months agoho: Remove rtp_proxy.h include, no proxy code in here
Holger Hans Peter Freyther [Tue, 17 Jan 2012 14:27:13 +0000]
ho: Remove rtp_proxy.h include, no proxy code in here

4 months agonat: Add the bsc_data.c file to the EXTRA_DIST
Holger Hans Peter Freyther [Sun, 15 Jan 2012 16:55:50 +0000]
nat: Add the bsc_data.c file to the EXTRA_DIST

4 months agoabis: gsm_data_shared.h includes libosmo-abis header, add abis to cflags
Holger Hans Peter Freyther [Sun, 15 Jan 2012 16:52:02 +0000]
abis: gsm_data_shared.h includes libosmo-abis header, add abis to cflags

gsm_data_shared.h includes e1_input.h of libosmo-abis, add the
LIBOSMOABIS_CFLAGS to Makefile.am, remove AM_LDFLAGS at the same
time as we only build .a archives.

4 months agomncc: Include size and offsets of struct gsm_mncc in the hello
Holger Hans Peter Freyther [Sat, 14 Jan 2012 23:38:35 +0000]
mncc: Include size and offsets of struct gsm_mncc in the hello

Increase the version number and send the sizeof of the gsm_mncc
and the gsm_data_frame structure. Include the offsets of some
members as well to maybe identify ABI problems.

The reason we go through this is that we want to benefit from
the native ABI and don't want to force packed structs.

4 months agochannel: Review lchan_free calls and the state the channel is in
Holger Hans Peter Freyther [Thu, 29 Dec 2011 22:33:04 +0000]
channel: Review lchan_free calls and the state the channel is in

In case of a memory allocation failure in rsl_rx_chan_rqd we would
have left the channel in the LCHAN_S_ACT_REQ state. Move the state
change below the allocation.

4 months agolchan: Rename release_reason to release_mode
Holger Hans Peter Freyther [Tue, 27 Dec 2011 19:49:53 +0000]
lchan: Rename release_reason to release_mode

Use the name of GSM 08.58 for the release mode.

4 months agochannel: Fix a typo in the lchan state documentation
Holger Hans Peter Freyther [Sat, 14 Jan 2012 23:02:17 +0000]
channel: Fix a typo in the lchan state documentation

4 months agochannel: The error timer and the lchan_deact_tmr are both running
Holger Hans Peter Freyther [Sat, 14 Jan 2012 20:56:37 +0000]
channel: The error timer and the lchan_deact_tmr are both running

In case a BTS does not send the RF Channel Release ACK and we are
closing the channel because of an error two timers are running to
set the state back to none.

Make lchan_deact_tmr_cb and rsl_rx_rf_chan_rel_ack behave the same
in regard to changing the state of the lchan. For the other direction
we are save, the error path will set the state to NONE and the timeout
will call lchan_free to set the channel type back to NONE, only then
the lchan may be allocated again.

The channel release procedure requires some more tweaking, some part
was started in the zecke/28c3 branch and needs to be tested/integrated
with the goal of having one common release path.

4 months agoRevert "libbsc: Log if a channel is freed with lchan->state != LCHAN_S_NONE"
Holger Hans Peter Freyther [Thu, 29 Dec 2011 22:27:11 +0000]
Revert "libbsc: Log if a channel is freed with lchan->state != LCHAN_S_NONE"

In case of a failed channel we still want the channel to not be
re-allocated right away but keep it closed/unused for (T3109 + T311).

rsl_rx_rf_chan_rel_ack has a check to not set the channel to S_NONE in
case the channel is in the error state. Add the camp Harald added a timer
to set the channel back to the none state in case the RF Channel Release
is not acked.

This reverts commit fc462dd59ee4772f06c9f2477b36b3f166a8e74b.

4 months agobsc: Fix "show lchan" for partially provided information
Holger Hans Peter Freyther [Tue, 27 Dec 2011 15:34:12 +0000]
bsc: Fix "show lchan" for partially provided information

show lchan should be capable of showing all allocated lchans,
all of a given bts, a given trx, a given ts. This feature was
broken when I added the ability to show a more simple summary.

Restore the initial behavior by splitting out the for loops
for the bts/trx/ts and check if we have parsed all parameters
and then call and return the subroutine.

4 months agonat: Add the line break after the '.
Holger Hans Peter Freyther [Wed, 11 Jan 2012 10:23:25 +0000]
nat: Add the line break after the '.

4 months agonat: Implement clearing of TP-SRR flags from TPDUs
Holger Hans Peter Freyther [Tue, 10 Jan 2012 21:39:07 +0000]
nat: Implement clearing of TP-SRR flags from TPDUs

Match IMSI and destination address against a set of entries, if it
is matching the header will be modified and no sender report will be
requested. Change the test case to request the sender report and then
verify that this bit is reset to 0.

4 months agonat: Reduce copy and paste in the test, create a verify_msg
Holger Hans Peter Freyther [Tue, 10 Jan 2012 21:31:39 +0000]
nat: Reduce copy and paste in the test, create a verify_msg

4 months agonat: Use the link_id from the original message in the new DTAP message
Holger Hans Peter Freyther [Tue, 10 Jan 2012 21:29:35 +0000]
nat: Use the link_id from the original message in the new DTAP message

SMS went from SAPI=3 to SAPI=0 and nobody notices on the NAT->MSC direction.

4 months agonat: Compare the DTAP L3 size with the size of the msgb - dtap header
Holger Hans Peter Freyther [Tue, 10 Jan 2012 21:27:43 +0000]
nat: Compare the DTAP L3 size with the size of the msgb - dtap header

4 months agonat: Fix comment and mask for looking at TP-MTI
Holger Hans Peter Freyther [Tue, 10 Jan 2012 13:33:27 +0000]
nat: Fix comment and mask for looking at TP-MTI

4 months agonat: SMS rewrite matching, empty list means match everything
Holger Hans Peter Freyther [Tue, 10 Jan 2012 13:22:55 +0000]
nat: SMS rewrite matching, empty list means match everything

4 months agonat: SMS comment updates
Holger Hans Peter Freyther [Tue, 10 Jan 2012 13:13:43 +0000]
nat: SMS comment updates

4 months agonat: Move SMS creation into a new method, no functional change
Holger Hans Peter Freyther [Tue, 10 Jan 2012 12:44:05 +0000]
nat: Move SMS creation into a new method, no functional change

4 months agonat: Access data_ptr after we have checked the length
Holger Hans Peter Freyther [Tue, 10 Jan 2012 12:04:40 +0000]
nat: Access data_ptr after we have checked the length

4 months agonat: Move the finding of the number to a new method
Holger Hans Peter Freyther [Tue, 10 Jan 2012 11:05:52 +0000]
nat: Move the finding of the number to a new method

4 months agotests: Use atlocal/atlocal.in to enable/disable the NAT test
Holger Hans Peter Freyther [Mon, 9 Jan 2012 21:53:04 +0000]
tests: Use atlocal/atlocal.in to enable/disable the NAT test

Kill the hacking, use atlocal to remember if the NAT test should
be enabled and then skip the test (exit with 77).

4 months agoRevert "tests: Make the test setup more abusive to fix the build"
Holger Hans Peter Freyther [Mon, 9 Jan 2012 20:49:47 +0000]
Revert "tests: Make the test setup more abusive to fix the build"

This reverts commit 41b42c72294d181217bae59e37ddf50c490df17f.

4 months agotests: Make the test setup more abusive to fix the build
Holger Hans Peter Freyther [Mon, 9 Jan 2012 20:13:15 +0000]
tests: Make the test setup more abusive to fix the build

Paolo says I will need to use atconfig/atlocal, need to read up on
it first though.

4 months agogprs: Honor GSM 04.64 8.4.2 Receipt of unacknowledged information
Holger Hans Peter Freyther [Thu, 23 Jun 2011 21:53:27 +0000]
gprs: Honor GSM 04.64 8.4.2 Receipt of unacknowledged information

GSM 04.64 8.4.2 asks to ignore UI frames if the DLCI is not known,
or if the "(V(UR)- 32) <= N(U) < V(UR)". E.g. if we want to have
V(UR) == 511 and this frame is dropped, we would ignore N(U)'s
0 to 510. Calculate the delta.

The code is based on Jonathan Santos's "LLC UI window" fix but the
issue was discovered independly.

4 months agogprs_sgsn.h: Include timer.h for the timer struct
Holger Hans Peter Freyther [Fri, 6 Jan 2012 16:29:06 +0000]
gprs_sgsn.h: Include timer.h for the timer struct

4 months agotests: Introduce running tests with GNU autotest in OpenBSC
Holger Hans Peter Freyther [Fri, 6 Jan 2012 14:16:12 +0000]
tests: Introduce running tests with GNU autotest in OpenBSC

The quality of the tests is of different value but it is good to
get started and improve from here.

4 months agomgcp: Print message when the test is finished
Holger Hans Peter Freyther [Fri, 6 Jan 2012 14:04:43 +0000]
mgcp: Print message when the test is finished

4 months agogsm0408: Print message when the test is done
Holger Hans Peter Freyther [Fri, 6 Jan 2012 14:03:28 +0000]
gsm0408: Print message when the test is done

4 months agodb: Fix the test code
Holger Hans Peter Freyther [Fri, 6 Jan 2012 13:27:08 +0000]
db: Fix the test code

* Initialize logging
* use a dummy net before calling subcr_put as the keep in ram
  option is going to be set.

4 months agochannel: Make the test not crash again.
Holger Hans Peter Freyther [Fri, 6 Jan 2012 13:06:56 +0000]
channel: Make the test not crash again.

* Initialize logging
* The callback expects NULL or a subscriber connection

4 months agonat: Log test output stdout using printf in prep for GNU autotest
Holger Hans Peter Freyther [Fri, 6 Jan 2012 12:56:12 +0000]
nat: Log test output stdout using printf in prep for GNU autotest

4 months agobsc: Assume assignment_complete/assignment_fail is set
Holger Hans Peter Freyther [Tue, 27 Dec 2011 11:31:02 +0000]
bsc: Assume assignment_complete/assignment_fail is set

The osmo-nitb application sometimes crashes because the BSC API
is doing an assignment underneath which is not handled by the code,
add dumy handlers to not crash, the right thing to do is to change
MNCC to have an assignment that can succeed/fail.

The keyword to look for is MNCC_LCHAN_MODIFY and mncc_sock should
wait for an ack/nack but right now the call just continues.

4 months agoRSL: Nokia BTS reports TA in quarter-bits, not full bits.
Andreas Eversberg [Sat, 24 Dec 2011 10:49:05 +0000]
RSL: Nokia BTS reports TA in quarter-bits, not full bits.

We need to compensate accordingly...

5 months agoRSL: BS-11 reports TA in quarter-bits, not full bits.
Andreas Eversberg [Fri, 16 Dec 2011 16:45:37 +0000]
RSL: BS-11 reports TA in quarter-bits, not full bits.

We need to compensate accordingly...

5 months agoAdd NITZ (timezone) support as part of MM INFO
Gus Bourg [Fri, 2 Dec 2011 09:18:17 +0000]
Add NITZ (timezone) support as part of MM INFO

The UTC offset from the operating system will be used by default to
calculate the NITZ in MM INFO.  However,  a "timezone" vty command is
added at the BTS level, allowing BTS-specific overrides, e.g. in case
BTSs are distributed accross multiple timezones.

5 months agofix two mistakes in AM_LDFLAGS
Alexander Huemer [Sun, 20 Nov 2011 20:53:15 +0000]
fix two mistakes in AM_LDFLAGS

5 months agoBSSGP: Add functions required for true BSS-side BSSGP implementation
Harald Welte [Fri, 25 Nov 2011 07:58:40 +0000]
BSSGP: Add functions required for true BSS-side BSSGP implementation

The BSS-side of BSSGP requires quite a number of additional functions
for sending unidirectional messages that a SGSN never sends.

This is a first step into completing the BSSGP implementation and making
it ready to be used from osmo-bts and other BTS-side GPRS
implementations.

5 months agoBSSGP: add more doxygen comments
Harald Welte [Fri, 25 Nov 2011 07:15:42 +0000]
BSSGP: add more doxygen comments

5 months agoGPRS: add new bssgp_create_cell_id(), opposite of bssgp_parse_cell_id()
Harald Welte [Thu, 24 Nov 2011 20:32:07 +0000]
GPRS: add new bssgp_create_cell_id(), opposite of bssgp_parse_cell_id()

5 months agoGPRS NS: Add some doxygen API documentation
Harald Welte [Wed, 23 Nov 2011 14:01:31 +0000]
GPRS NS: Add some doxygen API documentation

6 months agobsc_vty: Fix usage of deprecated osmo_osmo_hexdump_nospc
Sylvain Munaut [Sun, 13 Nov 2011 22:05:23 +0000]
bsc_vty: Fix usage of deprecated osmo_osmo_hexdump_nospc

The typo was fixed upstream

Signed-off-by: Sylvain Munaut <tnt@246tNt.com>

6 months agoosmo_nitb: introduce --no-dbcounter command line option
Harald Welte [Sun, 13 Nov 2011 18:02:06 +0000]
osmo_nitb: introduce --no-dbcounter command line option

this allows us to disable writing the current rate counters to the
sqlite database every 60 seconds.

6 months agogsm0408: Print a message when the LU is timing out.
Holger Hans Peter Freyther [Mon, 7 Nov 2011 11:26:29 +0000]
gsm0408: Print a message when the LU is timing out.

6 months agomncc: Introduce a hello packet that is sent to the client.
Holger Hans Peter Freyther [Fri, 21 Oct 2011 12:12:46 +0000]
mncc: Introduce a hello packet that is sent to the client.

Send a hello packet down to the client with the version number
of the MNCC interface. The hello structure might be extended to
include the endianes, size of each structure, etc.

6 months agosgsn: Make sure each parameter has some sort of documentation
Holger Hans Peter Freyther [Sat, 5 Nov 2011 14:21:16 +0000]
sgsn: Make sure each parameter has some sort of documentation

6 months agogb_proxy: Make sure each parameter has some sort of documentation
Holger Hans Peter Freyther [Sat, 5 Nov 2011 14:14:59 +0000]
gb_proxy: Make sure each parameter has some sort of documentation