- OpenBSC - lists.osmocom.org

[PATCH openggsn 0/4] add and update gtp-kernel support to openggsn
by Andreas Schultz 21 Dec '15

21 Dec '15

This series imports Pablo Neira Ayuso original gtp-kernel support patch unmodified to master and applies some fixes ontop of it. The last patch in the series then add network namespace support to it. This state is in sync with gtp-kernel patch I posted a few minutes ago. Andreas -- Andreas Schultz (3): ggsn: update gpt-kernel logging to libosmocore ggsn: fix autotool pkg-config invokation ggsn: add network namespace support Pablo Neira Ayuso (1): ggsn: add support for GTP kernel data encapsulation configure.ac | 15 ++++ ggsn/Makefile.am | 11 ++- ggsn/cmdline.c | 48 +++++++++- ggsn/cmdline.ggo | 4 + ggsn/cmdline.h | 8 ++ ggsn/ggsn.c | 52 ++++++++++- ggsn/gtp-kernel.c | 258 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ ggsn/gtp-kernel.h | 53 +++++++++++ lib/Makefile.am | 4 +- lib/netns.c | 135 ++++++++++++++++++++++++++++ lib/netns.h | 26 ++++++ 11 files changed, 607 insertions(+), 7 deletions(-) create mode 100644 ggsn/gtp-kernel.c create mode 100644 ggsn/gtp-kernel.h create mode 100644 lib/netns.c create mode 100644 lib/netns.h -- 2.5.0

4 10

Re: PySIM: SW match failed
by Florian 15 Dec '15

15 Dec '15

On Thu Nov 5 09:05:15 UTC 2015, Harald Welte wrote: >> RuntimeError: SW match failed ! Expected 9000 and got 6a86. > According to ISO 7816-4, this 6a86 means 'incorrect P1 or P2 > parameter'. >> At this point I don't have any more ideas what to try, if anyone would >> have any suggestions I would apreciate it. > Please activate (or hack some code for) tracing the actual APDUs that > pySim excahnges with the card. IIRC, pySim already has that option. > Once you see the raw APDUs, you can compare their encoding > (particularly > P1/P2) with those described in the relevant ETSI/3GPP (U)SIM > specifications. I enhanced my local copy of pySim (zecke/tmp) to show the raw PDUs. And I added a method to check the Status of the PIN and ADM register. On one card I messed up the ADM, so pySim-prog.py shows whey query for SIM 1: > send_apdu_raw -> 00200001 < received status word 63c3 So that is SW1='63' with SW2='CX': Counter (verification failed: 'X' indicates the number of further allowed retries Which means I have three more attempts for the PIN1, but > send_apdu_raw -> 0020000A < received status word 63c0 Ahhrg, I do not have any additional attempt to verify the ADM-Key. Trying to verify ADM results in > send_apdu_raw -> 0020000A083132333435363738 < received status word 6983 an ugly SW1='69' with SW2='83': Authentication method blocked. So I can no longer verify the ADM-Key on that card. --> Is there any way to unblock the card? On the second card I was able to successfully verify the ADM and change the IMSI... Cheers, Flo

4 4

[PATCH] gtphub: add to debian build
by Neels Hofmeyr 14 Dec '15

14 Dec '15

By the example of osmo-sgsn, package osmo-gtphub for debian. Sponsored-by: On-Waves ehi --- debian/control | 14 +++ debian/osmocom-gtphub.default | 2 + debian/osmocom-gtphub.examples | 1 + debian/osmocom-gtphub.init | 150 +++++++++++++++++++++++ debian/osmocom-gtphub.install | 1 + debian/rules | 1 + openbsc/doc/examples/osmo-gtphub/osmo-gtphub.cfg | 23 ++++ 7 files changed, 192 insertions(+) create mode 100644 debian/osmocom-gtphub.default create mode 100644 debian/osmocom-gtphub.examples create mode 100755 debian/osmocom-gtphub.init create mode 100644 debian/osmocom-gtphub.install create mode 100644 openbsc/doc/examples/osmo-gtphub/osmo-gtphub.cfg diff --git a/debian/control b/debian/control index 2447d29..53b6908 100644 --- a/debian/control +++ b/debian/control @@ -50,6 +50,12 @@ Depends: ${shlibs:Depends}, ${misc:Depends} Description: Osmocom Base Station Controller Network Address Translation Network address translation for BSC. +Package: osmocom-gtphub +Architecture: any +Depends: ${shlibs:Depends}, ${misc:Depends} +Description: Osmocom GTP hub + Proxy for comms between SGSN and GGSN. + Package: osmocom-bsc-dbg Architecture: any Section: debug @@ -105,3 +111,11 @@ Priority: extra Depends: osmocom-bsc-nat (= ${binary:Version}), ${misc:Depends} Description: Debug symbols for the OpenBSC Network Address Translation Make debugging possible + +Package: osmocom-gtphub-dbg +Architecture: any +Section: debug +Priority: extra +Depends: osmocom-gtphub (= ${binary:Version}), ${misc:Depends} +Description: Debug symbols for Osmocom GTP hub + Make debugging possible diff --git a/debian/osmocom-gtphub.default b/debian/osmocom-gtphub.default new file mode 100644 index 0000000..6af82da --- /dev/null +++ b/debian/osmocom-gtphub.default @@ -0,0 +1,2 @@ +CONFIG_FILE="/etc/osmocom/osmo-gtphub.cfg" + diff --git a/debian/osmocom-gtphub.examples b/debian/osmocom-gtphub.examples new file mode 100644 index 0000000..48c2dc0 --- /dev/null +++ b/debian/osmocom-gtphub.examples @@ -0,0 +1 @@ +openbsc/doc/examples/osmo-gtphub diff --git a/debian/osmocom-gtphub.init b/debian/osmocom-gtphub.init new file mode 100755 index 0000000..4dc67b3 --- /dev/null +++ b/debian/osmocom-gtphub.init @@ -0,0 +1,150 @@ +#!/bin/sh +### BEGIN INIT INFO +# Provides: osmo-gtphub +# Required-Start: $network $local_fs +# Required-Stop: +# Default-Start: 2 3 4 5 +# Default-Stop: 0 1 6 +# Short-Description: Osmocom GTP hub +# Description: Osmocom GTP hub +### END INIT INFO + +# Author: Neels Hofmeyr <nhofmeyr(a)sysmocom.de> + +# PATH should only include /usr/* if it runs after the mountnfs.sh script +PATH=/sbin:/usr/sbin:/bin:/usr/bin +NAME=osmo-gtphub # Introduce the short server's name here +DESC="Osmocom GTP hub" # Introduce a short description here +DAEMON=/usr/bin/osmo-gtphub # Introduce the server's location here +SCRIPTNAME=/etc/init.d/osmocom-gtphub + +# Exit if the package is not installed +[ -x $DAEMON ] || exit 0 + +# Read configuration variable file if it is present +[ -r /etc/default/osmocom-gtphub ] && . /etc/default/osmocom-gtphub + +# Load the VERBOSE setting and other rcS variables +. /lib/init/vars.sh + +# Define LSB log_* functions. +# Depend on lsb-base (>= 3.0-6) to ensure that this file is present. +. /lib/lsb/init-functions + +DAEMON_ARGS="$DAEMON_ARGS -D -c $CONFIG_FILE" + +# +# Function that starts the daemon/service +# +do_start() +{ + # Return + # 0 if daemon has been started + # 1 if daemon was already running + # 2 if daemon could not be started + start-stop-daemon --start --quiet --exec $DAEMON --test > /dev/null \ + || return 1 + start-stop-daemon --start --quiet --exec $DAEMON -- \ + $DAEMON_ARGS \ + || return 2 + # Add code here, if necessary, that waits for the process to be ready + # to handle requests from services started subsequently which depend + # on this one. As a last resort, sleep for some time. +} + +# +# Function that stops the daemon/service +# +do_stop() +{ + # Return + # 0 if daemon has been stopped + # 1 if daemon was already stopped + # 2 if daemon could not be stopped + # other if a failure occurred + start-stop-daemon --stop --quiet --retry=TERM/30/KILL/5 --name $NAME + RETVAL="$?" + [ "$RETVAL" = 2 ] && return 2 + # Wait for children to finish too if this is a daemon that forks + # and if the daemon is only ever run from this initscript. + # If the above conditions are not satisfied then add some other code + # that waits for the process to drop all resources that could be + # needed by services started subsequently. A last resort is to + # sleep for some time. + start-stop-daemon --stop --quiet --oknodo --retry=0/30/KILL/5 --exec $DAEMON + [ "$?" = 2 ] && return 2 + return "$RETVAL" +} + +# +# Function that sends a SIGHUP to the daemon/service +# +do_reload() { + # + # If the daemon can reload its configuration without + # restarting (for example, when it is sent a SIGHUP), + # then implement that here. + # + start-stop-daemon --stop --signal 1 --quiet $PIDFILE --name $NAME + return 0 +} + +case "$1" in + start) + [ "$VERBOSE" != no ] && log_daemon_msg "Starting $DESC " "$NAME" + do_start + case "$?" in + 0|1) [ "$VERBOSE" != no ] && log_end_msg 0 ;; + 2) [ "$VERBOSE" != no ] && log_end_msg 1 ;; + esac + ;; + stop) + [ "$VERBOSE" != no ] && log_daemon_msg "Stopping $DESC" "$NAME" + do_stop + case "$?" in + 0|1) [ "$VERBOSE" != no ] && log_end_msg 0 ;; + 2) [ "$VERBOSE" != no ] && log_end_msg 1 ;; + esac + ;; + status) + status_of_proc "$DAEMON" "$NAME" && exit 0 || exit $? + ;; + #reload|force-reload) + # + # If do_reload() is not implemented then leave this commented out + # and leave 'force-reload' as an alias for 'restart'. + # + #log_daemon_msg "Reloading $DESC" "$NAME" + #do_reload + #log_end_msg $? + #;; + restart|force-reload) + # + # If the "reload" option is implemented then remove the + # 'force-reload' alias + # + log_daemon_msg "Restarting $DESC" "$NAME" + do_stop + case "$?" in + 0|1) + do_start + case "$?" in + 0) log_end_msg 0 ;; + 1) log_end_msg 1 ;; # Old process is still running + *) log_end_msg 1 ;; # Failed to start + esac + ;; + *) + # Failed to stop + log_end_msg 1 + ;; + esac + ;; + *) + #echo "Usage: $SCRIPTNAME {start|stop|restart|reload|force-reload}" >&2 + echo "Usage: $SCRIPTNAME {start|stop|status|restart|force-reload}" >&2 + exit 3 + ;; +esac + +: diff --git a/debian/osmocom-gtphub.install b/debian/osmocom-gtphub.install new file mode 100644 index 0000000..908c1a5 --- /dev/null +++ b/debian/osmocom-gtphub.install @@ -0,0 +1 @@ +/usr/bin/osmo-gtphub diff --git a/debian/rules b/debian/rules index 8047b79..62518d9 100755 --- a/debian/rules +++ b/debian/rules @@ -34,6 +34,7 @@ override_dh_strip: dh_strip -posmocom-sgsn --dbg-package=osmocom-sgsn-dbg dh_strip -posmocom-gbproxy --dbg-package=osmocom-gbproxy-dbg dh_strip -posmocom-bsc-nat --dbg-package=osmocom-bsc-nat-dbg + dh_strip -posmocom-gtphub --dbg-package=osmocom-gtphub-dbg override_dh_auto_configure: echo $(VERSION) > openbsc/.tarball-version diff --git a/openbsc/doc/examples/osmo-gtphub/osmo-gtphub.cfg b/openbsc/doc/examples/osmo-gtphub/osmo-gtphub.cfg new file mode 100644 index 0000000..c9bb4cf --- /dev/null +++ b/openbsc/doc/examples/osmo-gtphub/osmo-gtphub.cfg @@ -0,0 +1,23 @@ +! +! Osmocom gtphub configuration +! + +line vty + no login + +gtphub + ! Local addresses to listen on and send from, each on standard ports + ! 2123 and 2152. Setting these addresses is mandatory. + bind-to-sgsns 127.0.0.1 + bind-to-ggsns 127.0.0.2 + + ! Local nonstandard ports or separate IPs: + !bind-to-sgsns ctrl 127.0.0.1 2342 user 127.0.0.1 4223 + + ! Proxy: unconditionally direct all traffic to... + !ggsn-proxy 127.0.0.3 + !sgsn-proxy 127.0.0.4 + + ! Proxy with nonstandard ports or separate IPs: + !ggsn-proxy ctrl 127.0.0.3 2123 user 127.0.0.5 2152 + -- 2.1.4

2 7

[PATCH 1/3] mncc: Implement helper functions to convert MNCC cause to a string.
by Alexander Chemeris 13 Dec '15

13 Dec '15

--- openbsc/include/openbsc/mncc.h | 5 +- openbsc/src/libmsc/mncc.c | 103 ++++++++++++++++++++++++++++++++++++++--- 2 files changed, 100 insertions(+), 8 deletions(-) diff --git a/openbsc/include/openbsc/mncc.h b/openbsc/include/openbsc/mncc.h index 46ba237..92389c2 100644 --- a/openbsc/include/openbsc/mncc.h +++ b/openbsc/include/openbsc/mncc.h @@ -191,8 +191,11 @@ struct gsm_mncc_rtp { uint32_t payload_msg_type; }; -char *get_mncc_name(int value); +const char *get_mncc_name(int value); void mncc_set_cause(struct gsm_mncc *data, int loc, int val); +int mncc_has_cause(struct gsm_mncc *data); +const char *get_mncc_location(int value); +const char *get_mncc_cause(int value); void cc_tx_to_mncc(struct gsm_network *net, struct msgb *msg); /* input from CC code into mncc_builtin */ diff --git a/openbsc/src/libmsc/mncc.c b/openbsc/src/libmsc/mncc.c index 2767faa..e8bd9ba 100644 --- a/openbsc/src/libmsc/mncc.c +++ b/openbsc/src/libmsc/mncc.c @@ -34,10 +34,12 @@ #include <openbsc/transaction.h> #include <openbsc/rtp_proxy.h> -static struct mncc_names { +struct name_value { char *name; int value; -} mncc_names[] = { +}; + +static struct name_value mncc_names[] = { {"MNCC_SETUP_REQ", 0x0101}, {"MNCC_SETUP_IND", 0x0102}, {"MNCC_SETUP_RSP", 0x0103}, @@ -95,16 +97,103 @@ static struct mncc_names { {NULL, 0} }; -char *get_mncc_name(int value) +static struct name_value mncc_locations[] = { + {"GSM48_CAUSE_LOC_USER", 0x00}, + {"GSM48_CAUSE_LOC_PRN_S_LU", 0x01}, + {"GSM48_CAUSE_LOC_PUN_S_LU", 0x02}, + {"GSM48_CAUSE_LOC_TRANS_NET", 0x03}, + {"GSM48_CAUSE_LOC_PUN_S_RU", 0x04}, + {"GSM48_CAUSE_LOC_PRN_S_RU", 0x05}, + /* not defined */ + {"GSM48_CAUSE_LOC_INN_NET", 0x07}, + {"GSM48_CAUSE_LOC_NET_BEYOND", 0x0a}, + + {NULL, 0} }; + +static struct name_value mncc_causes[] = { + {"GSM48_CC_CAUSE_UNASSIGNED_NR", 1}, + {"GSM48_CC_CAUSE_NO_ROUTE", 3}, + {"GSM48_CC_CAUSE_CHAN_UNACCEPT", 6}, + {"GSM48_CC_CAUSE_OP_DET_BARRING", 8}, + {"GSM48_CC_CAUSE_NORM_CALL_CLEAR", 16}, + {"GSM48_CC_CAUSE_USER_BUSY", 17}, + {"GSM48_CC_CAUSE_USER_NOTRESPOND", 18}, + {"GSM48_CC_CAUSE_USER_ALERTING_NA", 19}, + {"GSM48_CC_CAUSE_CALL_REJECTED", 21}, + {"GSM48_CC_CAUSE_NUMBER_CHANGED", 22}, + {"GSM48_CC_CAUSE_PRE_EMPTION", 25}, + {"GSM48_CC_CAUSE_NONSE_USER_CLR", 26}, + {"GSM48_CC_CAUSE_DEST_OOO", 27}, + {"GSM48_CC_CAUSE_INV_NR_FORMAT", 28}, + {"GSM48_CC_CAUSE_FACILITY_REJ", 29}, + {"GSM48_CC_CAUSE_RESP_STATUS_INQ", 30}, + {"GSM48_CC_CAUSE_NORMAL_UNSPEC", 31}, + {"GSM48_CC_CAUSE_NO_CIRCUIT_CHAN", 34}, + {"GSM48_CC_CAUSE_NETWORK_OOO", 38}, + {"GSM48_CC_CAUSE_TEMP_FAILURE", 41}, + {"GSM48_CC_CAUSE_SWITCH_CONG", 42}, + {"GSM48_CC_CAUSE_ACC_INF_DISCARD", 43}, + {"GSM48_CC_CAUSE_REQ_CHAN_UNAVAIL", 44}, + {"GSM48_CC_CAUSE_RESOURCE_UNAVAIL", 47}, + {"GSM48_CC_CAUSE_QOS_UNAVAIL", 49}, + {"GSM48_CC_CAUSE_REQ_FAC_NOT_SUBSC", 50}, + {"GSM48_CC_CAUSE_INC_BARRED_CUG", 55}, + {"GSM48_CC_CAUSE_BEARER_CAP_UNAUTH", 57}, + {"GSM48_CC_CAUSE_BEARER_CA_UNAVAIL", 58}, + {"GSM48_CC_CAUSE_SERV_OPT_UNAVAIL", 63}, + {"GSM48_CC_CAUSE_BEARERSERV_UNIMPL", 65}, + {"GSM48_CC_CAUSE_ACM_GE_ACM_MAX", 68}, + {"GSM48_CC_CAUSE_REQ_FAC_NOTIMPL", 69}, + {"GSM48_CC_CAUSE_RESTR_BCAP_AVAIL", 70}, + {"GSM48_CC_CAUSE_SERV_OPT_UNIMPL", 79}, + {"GSM48_CC_CAUSE_INVAL_TRANS_ID", 81}, + {"GSM48_CC_CAUSE_USER_NOT_IN_CUG", 87}, + {"GSM48_CC_CAUSE_INCOMPAT_DEST", 88}, + {"GSM48_CC_CAUSE_INVAL_TRANS_NET", 91}, + {"GSM48_CC_CAUSE_SEMANTIC_INCORR", 95}, + {"GSM48_CC_CAUSE_INVAL_MAND_INF", 96}, + {"GSM48_CC_CAUSE_MSGTYPE_NOTEXIST", 97}, + {"GSM48_CC_CAUSE_MSGTYPE_INCOMPAT", 98}, + {"GSM48_CC_CAUSE_IE_NOTEXIST", 99}, + {"GSM48_CC_CAUSE_COND_IE_ERR", 100}, + {"GSM48_CC_CAUSE_MSG_INCOMP_STATE", 101}, + {"GSM48_CC_CAUSE_RECOVERY_TIMER", 102}, + {"GSM48_CC_CAUSE_PROTO_ERR", 111}, + {"GSM48_CC_CAUSE_INTERWORKING", 127}, + + {NULL, 0} }; + +const char *get_name_by_value(struct name_value *pairs, int value, const char *default_val) { int i; - for (i = 0; mncc_names[i].name; i++) { - if (mncc_names[i].value == value) - return mncc_names[i].name; + for (i = 0; pairs[i].name; i++) { + if (pairs[i].value == value) + return pairs[i].name; } - return "MNCC_Unknown"; + return default_val; +} + + +const char *get_mncc_name(int value) +{ + return get_name_by_value(mncc_names, value, "MNCC_Unknown"); +} + +const char *get_mncc_location(int value) +{ + return get_name_by_value(mncc_locations, value, "GSM48_CAUSE_LOC_Unknown"); +} + +const char *get_mncc_cause(int value) +{ + return get_name_by_value(mncc_causes, value, "GSM48_CC_CAUSE_Unknown"); +} + +int mncc_has_cause(struct gsm_mncc *data) +{ + return data->fields & MNCC_F_CAUSE; } void mncc_set_cause(struct gsm_mncc *data, int loc, int val) -- 1.9.1

3 6

mncc-python tool for talking to MNCC interface
by Harald Welte 13 Dec '15

13 Dec '15

Hi all, I've been working on a small python tool that can be used to attach to the MNCC interface of OsmoNITB. It implements the 04.08 CC state machine with our MNCC primitives, including support for RTP bridge mode of the voice streams. The immediate first use case for this was to be able to automatically, reproducibly and quickly generate MT calls to a set of known MSISDNs and load all 14 TCH/H channels of a single-TRX BTS. It will connect the MT calls in pairs, so you end up with 7 MS-to-MS calls. Other use cases are expected to be added shortly. The first working version of the tool is available from http://git.osmocom.org/mncc-python/ or git clone git://git.osmocom.org/mncc-python The code is pretty hacky in some places. That's partially due to the fact that I'm much more familiar in the C, Perl and Erlang world than in python. Still I thought it's a good idea to do it in python to enable more people to use/edit/contribute to it. I'm happy for review / cleanup suggestion by people with more Python-foo than I have. Architecturally, I decided to do things a bit erlang-like, where we have finite state machines in an actor models, and message passing between the actors. This is what happens with the GsmCallFsm()'s, which are created by the GsmCallConnector() representing both legs of a call and the MnccActor() that wraps the MNCC socket towards OsmoNITB. The actual encoding/decodng of MNCC messages is auto-generated from the mncc header file #defines, enums and c-structures by means of ctypes code generation. mncc_test.py currently drops you into a python shell where you can e.g. start more / new calls by calling functions like connect_call("7839", "3802") from that shell. Exiting the shell by quit() or Ctrl+C will terminate all call FSMs and terminate. -- - Harald Welte <laforge(a)gnumonks.org> http://laforge.gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)

2 3

Mergin meas_json to master
by Sipos Csaba 12 Dec '15

12 Dec '15

Dear Alexander, I started to write the Wiki page for Meas_web. During this I just realized that because of some outstanding merges the installation of this tool is painfully complicated. In order to make things easier, I suggest to merge the meas_json part into OpenBSC master: http://cgit.osmocom.org/openbsc/log/?h=achemeris/meas_json Regards, Csaba

2 3

Re: The test suite for libosmocore fails on big-endian architectures
by Ruben Undheim 11 Dec '15

11 Dec '15

>> +#ifndef OSMO_IS_LITTLE_ENDIAN >> + #define OSMO_IS_LITTLE_ENDIAN 0 >> +#endif > > I am surprised by this hunk. Our endian.h should define both macros. I think I used > the vim search to search for a typo in one of the branches (BSD vs. Linux, LE vs. BE) > but it looks good. Do you remember why this was needed? It isn't needed. It was just added out of habit to make sure it is defined when it is being checked further down. You can remove it, it won't make any difference. Ruben

1 0

Re: OpenBSC on CalypsoBTS without voice
by Вадим Яницкий 11 Dec '15

11 Dec '15

Hello, Harald, thanks for the answer. first of all, CalypsoBTS is not an active / maintained / stable solution > but more lik an (mindbogglingly amazing!) Hack. Yes, I know that CalypsoBTS is not stable solution. But currently I am a student and I have no so much money to buy anything more powerful. Furthermore, I am interested in this for research purposes only, not for creating public land mobile networks. As you are the first person to report using the code to this list for > quite some time, it would be great if you had some time to re-base, test > and submit it. As I previously said, I also tested master branches. Here are the results: * When I start OsmoBTS I often see this message in OpenBSC log: "Lost some E1 TEI link: X X" from bsc_init.c. In this case my MS can not connect to the test network and I have to restart OsmoBTS until this message will no hide. * When the Location Update process I always see this error: "input/ipaccess.c:277 Bad signalling message,sign_link returned error". But a phone successfully registers in the network. * SMS, USSD and Silent SMS works fine. * When TS1 assigned TCH/H I can call to another ms or LCR. DTMF works. * When TS1 assigned TCH/F I can use FR codec and call to LCR. * PDCH does not work on this BTS. * RRLP works. * ISMI detach works but as in case of Location Update I see "Bad signalling message..." error. * It seems, the network works more stable when battery capacity is 100% and backlight is disabled. If I understood correctly, osmobts-trx was just moved from obsolete repository to the master. Now I am reading the source code and trying to fix this bugs. I have one thing in my mind for a long time. Is it possible to serve a PDCH lchan on TS1 instead of TCH (in case of this "powerful" BTS)? I have tested OsmoSGSN+OpenGGSN with one PDCH channel on TS1, but it did not work. With best regards, Яницкий Вадим. 2015-12-06 18:14 GMT+06:00 Harald Welte <laforge(a)gnumonks.org>: > Hi Vadim, >> >> first of all, CalypsoBTS is not an active / maintained / stable solution >> but more lik an (mindbogglingly amazing!) Hack. >> >> On Sat, Dec 05, 2015 at 06:32:45PM +0600, Вадим Яницкий wrote: >> > When I use jolly/multi-trx branch of libosmo-abis, jolly/testing branch >> of >> > OpenBSC and jolly/trx branch of OsmoBTS my network works very unstable. >> >> I'm sorry to say that none of the branches you refer to are maintained >> by anyone at this point. It would be great if somebody interested in >> that code (i.e. using it) could forward-port it onto current master. >> Now tat l1sap and osmo-bts-trx is in osmo-bts master, this should be >> relatively straight forward. >> >> It's really sad for us to see that people are continuing to use >> old/outdated non-master, rather than rebasing + submitting their changes >> for inclusion. It always makes me tempte to remove some of those >> branches from the repo, ort at least rename them to something like >> 'outdated'. >> >> As you are the first person to report using the code to this list for >> quite some time, it would be great if you had some time to re-base, test >> and submit it. >> >> -- >> - Harald Welte <laforge(a)gnumonks.org> >> http://laforge.gnumonks.org/ >> >> ============================================================================ >> "Privacy in residential applications is a desirable marketing option." >> (ETSI EN 300 175-7 Ch. >> A6) >> > >

1 0

The test suite for libosmocore fails on big-endian architectures
by Ruben Undheim 11 Dec '15

11 Dec '15

Hi, While building the package for Debian, apparently there is a problem related to big-endian architectures. Please see the logs here: https://buildd.debian.org/status/fetch.php?pkg=libosmocore&arch=powerpc&ver… and https://buildd.debian.org/status/package.php?p=libosmocore It is related to the test case for smscb and the struct gsm341_ms_message: +++ /«PKGBUILDDIR»/tests/testsuite.dir/at-groups/7/stdout 2015-12-05 15:51:53.463182812 +0000 @@ -1,4 +1,4 @@ -(srl) GS: 1 MSG_CODE: 1 UPDATE: 0 +(srl) GS: 0 MSG_CODE: 256 UPDATE: 1 (msg) msg_id: 1293 -(dcs) group: 1 language: 0 +(dcs) group: 0 language: 1 (pge) page total: 1 current: 1 7. testsuite.at:45: 7. smscb (testsuite.at:45): FAILED (testsuite.at:48) Could you please suggest how we should fix this so that the package also can build for powerpc and other big-endian architectures? Thank you very much in advance. Ruben

5 15

osmo-bts / osmo-pcu profiling results
by Harald Welte 10 Dec '15

10 Dec '15

Hi all, I've been doing some profiling on osmo-bts recently (on sysmobts hardware, which has only a relatively slow ARM926 CPU core), and the two things that show up most are: * msgb_alloc() -> talloc_zero() -> malloc this can be alleviated somewhat by using talloc pools. For some reason the pools don't remove all of the malloc() calls. * vfprintf() and friends, from logp() statements. The sad part is that calls like gsm_lchan_name() are of course executed beefore the call into logp(), at which point the vfprintf/sprintf/... for arguments has already been executed, and only the last/final one hasn't happened yet. Here we can do two things: Calls like gsm_lchan_name() don't need to happen all the time, as the lchan name is static and can be generated once at the time gsm_lchan is created. I implemented that in osmo-bts (and openbsc, as it's from gsm_data_shared). The second idea would be to expand the LOGP() macro a bit in a way to ensure the the checking whether the log line is enabled _before_ the arguments (and thus associated function calls) are evaluated. Any ideas on that? After a brief look at osmo-pcu profiling, it looks like in the attached picture. We cannot do much about the __copy_to_user_std, do_select and core_sys_select, as those are kernel side. However, there again we see vfprintf and friends, mostly via gprs_rlcmac_tbf::name() - and of course the msgb_alloc() and msgb_free() going through talloc and finally malloc. So the same strategies as above could (and probably should) be applied to osmo-pcu. Regards, Harald -- - Harald Welte <laforge(a)gnumonks.org> http://laforge.gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)

4 4