- baseband-devel - lists.osmocom.org

27 Jan '12

Hi list, hi Jolly, I just tested the DTMF support on a real network and discovered that the timeout of 70ms for receiving the Start DTMF ACK is too short. The consequences are many repeated Start DTMF messages, never received ACKs and no tone is played. Using a timer value of 170ms in mnccms.c:759 fixes this for that network. Unfortunately I didn't found - or overlooked - a mandatory timeout value for Ta in TS 23.014 which the MSC has to meet. Is that undefined in the standard and where came the 70ms from? Regards, Mad

2 2

Re: Linux + u-boot port to MT6235
by Victor Rojas 26 Jan '12

26 Jan '12

Nacesito temas y confaguracion

1 0

Creating unique features for a real OsmocomBB phone
by alphaflux 26 Jan '12

26 Jan '12

A novel feature would be direct connections to other OsmocomBB handsets that are in range. So not only handsets communicate with towers but a handset also listens to announcements of compatible handsets within range. Reason is to have free communication between different buildings, floors, rooms.

6 5

Re: Creating unique features for a real OsmocomBB phone
by Dave Schmidt 25 Jan '12

25 Jan '12

> why should two phones use uplink frequency to communicate directly > together? why not using downlink frequency? Because in most if not all counties transmission on downlink frequency requires a license? (Sorry and please correct me if this is some obvious nonsense... to me, the original proposal was looking quite nonsensial because of this, but I'd love to learn if I'm wrong...) Dave.

3 2

Asterisk GSM channel driver?
by Tim Ehlers 22 Jan '12

22 Jan '12

Hi, I have seen, that there are channel driver for the DECT project, where you can connect your DECT handset to the Asterisk PBX. There is also somehow a possibility to connect a mobile GSM phone via openBSC and asterisk. But has somebody thought about just connecting the osmocombb mobile phone (which is connected to a public mobile network [or openBSC]) to asterisk? I would think of making a call through a channel driver, osmocombb is doing a call like "call ms number", but instead of using mike and speaker asterisk would send the already in GSM-codec modulated voice data through the serial port (maybe it would needed to use the non-standard USB-TTL version to handle full duplex GSM speech). And layer1 of osmocombb could send that nearly untouched on the air. For incoming calls this could be similar + signalling the call through the channel driver. Has somebody already thought of that? Would that be a achieveable task or is it technically impossible? Thanks Tim

2 2

read Kc with osmocombb
by Tim Ehlers 19 Jan '12

19 Jan '12

Hi, today I seached for a possibility to read the temporary session key "Kc" from my phone in osmocombb. I did not find a function. So I used the function, which writes SMS into sms.txt to write the Kc. Since I never used the Kc to decode something I would like to ask if I did it correct this way: --- old/host/layer23/src/mobile/subscriber.c 2012-01-10 17:06:40.000000000 +0100 +++ new/host/layer23/src/mobile/subscriber.c 2012-01-19 23:52:36.000000000 +0100 @@ -60,6 +60,39 @@ return NULL; } +static int gsm_kc(struct osmocom_ms *ms, uint8_t *kc) +{ + const char osmocomkc[] = ".osmocom/bb/kc.txt"; + int len; + const char *home; + char *kc_file; + FILE *fp; + + home = getenv("HOME"); + if (!home) { +fail: + fprintf(stderr, "Can't deliver SMS, be sure to create '%s' in " + "your home directory.\n", osmocomkc); + return NULL; + } + len = strlen(home) + 1 + sizeof(osmocomkc); + kc_file = talloc_size(l23_ctx, len); + if (!kc_file) + goto fail; + snprintf(kc_file, len, "%s/%s", home, osmocomkc); + + fp = fopen(kc_file, "a"); + if (!fp) + goto fail; + fprintf(fp, "%s %02X%02X%02X%02X%02X%02X%02X%02X\n", ms->name, kc[0], kc[1], kc[2], kc[3], kc[4], kc[5], kc[6], kc[7], kc[8]); + fclose(fp); + + talloc_free(kc_file); + + return 0; +} + + static char *sim_decode_bcd(uint8_t *data, uint8_t length) { int i, j = 0; @@ -377,6 +410,7 @@ subscr->key_seq = data[8] & 0x07; LOGP(DMM, LOGL_INFO, "received KEY from SIM\n"); + gsm_kc(subscr->ms, subscr->key); return 0; } Maybe it would be a better way to implement a function in VTY. BTW, am I wrong or why did nobody do that until present? Thanks Tim

2 2

[PATCH] Switch to new gpsd API
by Paul Wise 18 Jan '12

18 Jan '12

--- src/host/layer23/src/common/gps.c | 26 +++++++++++++------------- 1 files changed, 13 insertions(+), 13 deletions(-) diff --git a/src/host/layer23/src/common/gps.c b/src/host/layer23/src/common/gps.c index 38aae2c..2bdfb97 100644 --- a/src/host/layer23/src/common/gps.c +++ b/src/host/layer23/src/common/gps.c @@ -56,7 +56,7 @@ static struct osmo_fd gps_bfd; #ifdef _HAVE_GPSD -static struct gps_data_t* gdata; +static struct gps_data_t gdata; int osmo_gpsd_cb(struct osmo_fd *bfd, unsigned int what) { @@ -66,25 +66,25 @@ int osmo_gpsd_cb(struct osmo_fd *bfd, unsigned int what) g.valid = 0; /* gps is offline */ - if (gdata->online) + if (gdata.online) goto gps_not_ready; /* gps has no data */ - if (gps_waiting(gdata)) + if (gps_waiting(&gdata, 500)) goto gps_not_ready; /* polling returned an error */ - if (gps_poll(gdata)) + if (gps_read(&gdata)) goto gps_not_ready; /* data are valid */ - if (gdata->set & LATLON_SET) { + if (gdata.set & LATLON_SET) { g.valid = 1; - g.gmt = gdata->fix.time; + g.gmt = gdata.fix.time; tm = localtime(&g.gmt); diff = time(NULL) - g.gmt; - g.latitude = gdata->fix.latitude; - g.longitude = gdata->fix.longitude; + g.latitude = gdata.fix.latitude; + g.longitude = gdata.fix.longitude; LOGP(DGPS, LOGL_INFO, " time=%02d:%02d:%02d %04d-%02d-%02d, " "diff-to-host=%d, latitude=%do%.4f, longitude=%do%.4f\n", @@ -111,16 +111,15 @@ int osmo_gpsd_open(void) gps_bfd.when = BSC_FD_READ; gps_bfd.cb = osmo_gpsd_cb; - gdata = gps_open(g.gpsd_host, g.gpsd_port); - if (gdata == NULL) { + if (gps_open(g.gpsd_host, g.gpsd_port, &gdata) == -1) { LOGP(DGPS, LOGL_ERROR, "Can't connect to gpsd\n"); return -1; } - gps_bfd.fd = gdata->gps_fd; + gps_bfd.fd = gdata.gps_fd; if (gps_bfd.fd < 0) return gps_bfd.fd; - if (gps_stream(gdata, WATCH_ENABLE, NULL) == -1) { + if (gps_stream(&gdata, WATCH_ENABLE, NULL) == -1) { LOGP(DGPS, LOGL_ERROR, "Error in gps_stream()\n"); return -1; } @@ -139,7 +138,8 @@ void osmo_gpsd_close(void) osmo_fd_unregister(&gps_bfd); - gps_close(gdata); + gps_stream(&gdata, WATCH_DISABLE, NULL); + gps_close(&gdata); gps_bfd.fd = -1; /* -1 or 0 indicates: 'close' */ } -- 1.7.7.3

6 9

Strange issues booting layer1
by EMP 18 Jan '12

18 Jan '12

Hi all I have compiled osmocombb and everything seems ok. Now i'm having a strange issue which I can replicate across a number of phones. Using a C118 I could boot everything and run the mobile app. Using a C139 I could do the same. However, now the phones dont seem to want to boot? I start everything up (after running the mobile app a few times, stopping the phone and PC etc) and then when I press the Power Off key, nothing happens. It just refuses to download anything. Am wondering if i'm damaging the phones somehow as its now two phones that will no longer start via the USB cable. The phones boot the normal phone firmware fine but wont respond to osmocom. Guess i'm just curious as to others thoughts as i'm unsure where its breaking?

1 0

GSM 7bit encoding/decoding
by Tim Ehlers 14 Jan '12

14 Jan '12

Hello, since I had problems with german umlauts when receiving (and probably also when sending) SMS, I looked today in the gsm translating part in libosmocore. I first noticed a problem, when sending "NETZ Number" to "4636" in O2 network, you get a response SMS saying in which network the number is located. Using osmocombb, I get: [SMS from 66399] Sehr geehrter Kunde, die angefragte Nummer ist im Netz von o2 aktiv (Angabe ohne Gew§hr). So the umlaut "ae" is converted to "§". The problem seems to be in the convert table in src/shared/libosmocore/src/gsm/gsm_utils.c, which is not bijective. "Ae" in gsm7 is "0x7b" and must be converted to latin1 "0xe4". Looking in gsm_7bit_alphabet at place "0xe4", the is "0x7b". But there is also an "0x7b" at place "0xa7", which comes first and is indeed "§". So the question is, are there mistakes in the table or is there a reason for the double entries? Thanks Tim

1 1

Paging Request3 - not standard compliant? o2-germany
by Stefan Bauer 14 Jan '12

14 Jan '12

Good evening, according to EN 300 940 - V7.7.1 - Digital cellular telecommunications system (Phase 2+); Mobile radio interface layer 3 specification (GSM 04.08 version 7.7.1 Release 1998) 9.1.24 Paging request type 3 such requests shall only address MSs by T/MSIs but i notice Type3-Requests with IMSI in o2-germany network. <0001> app_ccch_scan.c:378 Paging3: Normal paging chan n/a to imsi M(2261022118380XX) What do i miss? Stefan

2 1